qed: FW 8.42.2.0 HSI changes

This patch contains several HSI changes. The changes are part of features like RDMA VF and OVS, the patch also contains a fix to how the init code determines if the dmae is ready to be used. Signed-off-by: Ariel Elior <ariel.elior@marvell.com> Signed-off-by: Michal Kalderon <michal.kalderon@marvell.com> Signed-off-by: David S. Miller <davem@davemloft.net>
author: Michal Kalderon <michal.kalderon@marvell.com> 2020-01-27 15:26:15 +0200
committer: David S. Miller <davem@davemloft.net> 2020-01-27 14:35:32 +0100
commit: 0500a70d6e071040ffdaadebb966986afa83c5e9 (patch)
tree: 0905d51352b4a10aa351092ad5c55ead9b481991 /tools/perf/scripts/python/netdev-times.py
parent: 6459d93619b5bc21f775e7eb12bc4d051743d7aa (diff)
0 files changed, 0 insertions, 0 deletions
diff --git a/drivers/gpu/Makefile b/drivers/gpu/Makefile
index 835c88318cec..36a54d456630 100644
--- a/drivers/gpu/Makefile
+++ b/drivers/gpu/Makefile
@@ -2,7 +2,7 @@
 # drm/tegra depends on host1x, so if both drivers are built-in care must be
 # taken to initialize them in the correct order. Link order is the only way
 # to ensure this currently.
-obj-$(CONFIG_TEGRA_HOST1X)	+= host1x/
-obj-y			+= drm/ vga/
+obj-y			+= host1x/ drm/ vga/
 obj-$(CONFIG_IMX_IPUV3_CORE)	+= ipu-v3/
 obj-$(CONFIG_TRACE_GPU_MEM)		+= trace/
+obj-$(CONFIG_NOVA_CORE)		+= nova-core/
diff --git a/drivers/gpu/drm/Kconfig b/drivers/gpu/drm/Kconfig
index 0039df26854b..7e6bc0b3a589 100644
--- a/drivers/gpu/drm/Kconfig
+++ b/drivers/gpu/drm/Kconfig
@@ -10,14 +10,13 @@ menuconfig DRM
 	depends on (AGP || AGP=n) && !EMULATED_CMPXCHG && HAS_DMA
 	select DRM_PANEL_ORIENTATION_QUIRKS
 	select HDMI
-	select FB_CMDLINE
 	select I2C
-	select I2C_ALGOBIT
 	select DMA_SHARED_BUFFER
 	select SYNC_FILE
 # gallium uses SYS_kcmp for os_same_file_description() to de-duplicate
 # device and dmabuf fd. Let's make sure that is available for our userspace.
 	select KCMP
+	select VIDEO
 	help
 	  Kernel-level support for the Direct Rendering Infrastructure (DRI)
 	  introduced in XFree86 4.0. If you say Y here, you need to select
@@ -27,69 +26,115 @@ menuconfig DRM
 	  details.  You should also select and configure AGP
 	  (/dev/agpgart) support if it is available for your platform.
 
+menu "DRM debugging options"
+depends on DRM
+source "drivers/gpu/drm/Kconfig.debug"
+endmenu
+
+if DRM
+
 config DRM_MIPI_DBI
 	tristate
 	depends on DRM
+	select DRM_KMS_HELPER
 
 config DRM_MIPI_DSI
 	bool
 	depends on DRM
 
-config DRM_DP_AUX_BUS
+config DRM_KMS_HELPER
 	tristate
 	depends on DRM
-	depends on OF
-
-config DRM_DP_AUX_CHARDEV
-	bool "DRM DP AUX Interface"
-	depends on DRM
-	help
-	  Choose this option to enable a /dev/drm_dp_auxN node that allows to
-	  read and write values to arbitrary DPCD registers on the DP aux
-	  channel.
-
-config DRM_DEBUG_MM
-	bool "Insert extra checks and debug info into the DRM range managers"
-	default n
-	depends on DRM=y
-	depends on STACKTRACE_SUPPORT
-	select STACKDEPOT
+	select FB_CORE if DRM_FBDEV_EMULATION
 	help
-	  Enable allocation tracking of memory manager and leak detection on
-	  shutdown.
-
-	  Recommended for driver developers only.
+	  CRTC helpers for KMS drivers.
 
-	  If in doubt, say "N".
+config DRM_DRAW
+	bool
+	depends on DRM
 
-config DRM_DEBUG_SELFTEST
-	tristate "kselftests for DRM"
+config DRM_PANIC
+	bool "Display a user-friendly message when a kernel panic occurs"
 	depends on DRM
-	depends on DEBUG_KERNEL
-	select PRIME_NUMBERS
-	select DRM_LIB_RANDOM
-	select DRM_KMS_HELPER
-	select DRM_EXPORT_FOR_TESTS if m
-	default n
+	select FONT_SUPPORT
+	select DRM_DRAW
 	help
-	  This option provides kernel modules that can be used to run
-	  various selftests on parts of the DRM api. This option is not
-	  useful for distributions or general kernels, but only for kernel
-	  developers working on DRM and associated drivers.
-
+	  Enable a drm panic handler, which will display a user-friendly message
+	  when a kernel panic occurs. It's useful when using a user-space
+	  console instead of fbcon.
+	  It will only work if your graphic driver supports this feature.
+	  To support Hi-DPI Display, you can enable bigger fonts like
+	  FONT_TER16x32
+
+config DRM_PANIC_FOREGROUND_COLOR
+	hex "Drm panic screen foreground color, in RGB"
+	depends on DRM_PANIC
+	default 0xffffff
+
+config DRM_PANIC_BACKGROUND_COLOR
+	hex "Drm panic screen background color, in RGB"
+	depends on DRM_PANIC
+	default 0x000000
+
+config DRM_PANIC_DEBUG
+	bool "Add a debug fs entry to trigger drm_panic"
+	depends on DRM_PANIC && DEBUG_FS
+	help
+	  Add dri/[device]/drm_panic_plane_x in the kernel debugfs, to force the
+	  panic handler to write the panic message to this plane scanout buffer.
+	  This is unsafe and should not be enabled on a production build.
 	  If in doubt, say "N".
 
-config DRM_KMS_HELPER
-	tristate
-	depends on DRM
+config DRM_PANIC_SCREEN
+	string "Panic screen formatter"
+	default "user"
+	depends on DRM_PANIC
 	help
-	  CRTC helpers for KMS drivers.
+	  This option enable to choose what will be displayed when a kernel
+	  panic occurs. You can choose between "user", a short message telling
+	  the user to reboot the system, or "kmsg" which will display the last
+	  lines of kmsg.
+	  This can also be overridden by drm.panic_screen=xxxx kernel parameter
+	  or by writing to /sys/module/drm/parameters/panic_screen sysfs entry
+	  Default is "user"
+
+config DRM_PANIC_SCREEN_QR_CODE
+	bool "Add a panic screen with a QR code"
+	depends on DRM_PANIC && RUST
+	select ZLIB_DEFLATE
+	help
+	  This option adds a QR code generator, and a panic screen with a QR
+	  code. The QR code will contain the last lines of kmsg and other debug
+	  information. This should be easier for the user to report a kernel
+	  panic, with all debug information available.
+	  To use this panic screen, also set DRM_PANIC_SCREEN to "qr_code"
+
+config DRM_PANIC_SCREEN_QR_CODE_URL
+	string "Base URL of the QR code in the panic screen"
+	depends on DRM_PANIC_SCREEN_QR_CODE
+	help
+	  This option sets the base URL to report the kernel panic. If it's set
+	  the QR code will contain the URL and the kmsg compressed with zlib as
+	  a URL parameter. If it's empty, the QR code will contain the kmsg as
+	  uncompressed text only.
+	  There is a demo code in javascript, to decode and uncompress the kmsg
+	  data from the URL parameter at https://github.com/kdj0c/panic_report
+
+config DRM_PANIC_SCREEN_QR_VERSION
+	int "Maximum version (size) of the QR code."
+	depends on DRM_PANIC_SCREEN_QR_CODE
+	default 40
+	help
+	  This option limits the version (or size) of the QR code. QR code
+	  version ranges from Version 1 (21x21) to Version 40 (177x177).
+	  Smaller QR code are easier to read, but will contain less debugging
+	  data. Default is 40.
 
 config DRM_DEBUG_DP_MST_TOPOLOGY_REFS
         bool "Enable refcount backtrace history in the DP MST helpers"
 	depends on STACKTRACE_SUPPORT
         select STACKDEPOT
-        depends on DRM_KMS_HELPER
+        select DRM_KMS_HELPER
         depends on DEBUG_KERNEL
         depends on EXPERT
         help
@@ -115,56 +160,16 @@ config DRM_DEBUG_MODESET_LOCK
 
 	  If in doubt, say "N".
 
-config DRM_FBDEV_EMULATION
-	bool "Enable legacy fbdev support for your modesetting driver"
-	depends on DRM_KMS_HELPER
-	depends on FB=y || FB=DRM_KMS_HELPER
-	select FB_CFB_FILLRECT
-	select FB_CFB_COPYAREA
-	select FB_CFB_IMAGEBLIT
-	select FB_DEFERRED_IO
-	select FB_SYS_FOPS
-	select FB_SYS_FILLRECT
-	select FB_SYS_COPYAREA
-	select FB_SYS_IMAGEBLIT
-	select FRAMEBUFFER_CONSOLE if !EXPERT
-	select FRAMEBUFFER_CONSOLE_DETECT_PRIMARY if FRAMEBUFFER_CONSOLE
-	default y
-	help
-	  Choose this option if you have a need for the legacy fbdev
-	  support. Note that this support also provides the linux console
-	  support on top of your modesetting driver.
-
-	  If in doubt, say "Y".
-
-config DRM_FBDEV_OVERALLOC
-	int "Overallocation of the fbdev buffer"
-	depends on DRM_FBDEV_EMULATION
-	default 100
-	help
-	  Defines the fbdev buffer overallocation in percent. Default
-	  is 100. Typical values for double buffering will be 200,
-	  triple buffering 300.
-
-config DRM_FBDEV_LEAK_PHYS_SMEM
-	bool "Shamelessly allow leaking of fbdev physical address (DANGEROUS)"
-	depends on DRM_FBDEV_EMULATION && EXPERT
-	default n
+config DRM_CLIENT
+	bool
+	depends on DRM
 	help
-	  In order to keep user-space compatibility, we want in certain
-	  use-cases to keep leaking the fbdev physical address to the
-	  user-space program handling the fbdev buffer.
-	  This affects, not only, Amlogic, Allwinner or Rockchip devices
-	  with ARM Mali GPUs using an userspace Blob.
-	  This option is not supported by upstream developers and should be
-	  removed as soon as possible and be considered as a broken and
-	  legacy behaviour from a modern fbdev device driver.
+	  Enables support for DRM clients. DRM drivers that need
+	  struct drm_client_dev and its interfaces should select this
+	  option. Drivers that support the default clients should
+	  select DRM_CLIENT_SELECTION instead.
 
-	  Please send any bug reports when using this to your proprietary
-	  software vendor that requires this.
-
-	  If in doubt, say "N" or spread the word to your closed source
-	  library vendor.
+source "drivers/gpu/drm/clients/Kconfig"
 
 config DRM_LOAD_EDID_FIRMWARE
 	bool "Allow to specify an EDID data set instead of probing for it"
@@ -178,25 +183,46 @@ config DRM_LOAD_EDID_FIRMWARE
 	  default case is N. Details and instructions how to build your own
 	  EDID data are given in Documentation/admin-guide/edid.rst.
 
-config DRM_DP_CEC
-	bool "Enable DisplayPort CEC-Tunneling-over-AUX HDMI support"
-	depends on DRM
-	select CEC_CORE
-	help
-	  Choose this option if you want to enable HDMI CEC support for
-	  DisplayPort/USB-C to HDMI adapters.
-
-	  Note: not all adapters support this feature, and even for those
-	  that do support this they often do not hook up the CEC pin.
+source "drivers/gpu/drm/display/Kconfig"
 
 config DRM_TTM
 	tristate
 	depends on DRM && MMU
+	select SHMEM
 	help
 	  GPU memory management subsystem for devices with multiple
 	  GPU memory types. Will be enabled automatically if a device driver
 	  uses it.
 
+config DRM_EXEC
+	tristate
+	depends on DRM
+	help
+	  Execution context for command submissions
+
+config DRM_GPUVM
+	tristate
+	depends on DRM
+	select DRM_EXEC
+	help
+	  GPU-VM representation providing helpers to manage a GPUs virtual
+	  address space
+
+config DRM_GPUSVM
+	tristate
+	depends on DRM && DEVICE_PRIVATE
+	select HMM_MIRROR
+	select MMU_NOTIFIER
+	help
+	  GPU-SVM representation providing helpers to manage a GPUs shared
+	  virtual memory
+
+config DRM_BUDDY
+	tristate
+	depends on DRM
+	help
+	  A page based buddy allocator
+
 config DRM_VRAM_HELPER
 	tristate
 	depends on DRM
@@ -207,80 +233,54 @@ config DRM_TTM_HELPER
 	tristate
 	depends on DRM
 	select DRM_TTM
+	select DRM_KMS_HELPER if DRM_FBDEV_EMULATION
+	select FB_CORE if DRM_FBDEV_EMULATION
+	select FB_SYSMEM_HELPERS_DEFERRED if DRM_FBDEV_EMULATION
 	help
 	  Helpers for ttm-based gem objects
 
-config DRM_GEM_CMA_HELPER
-	bool
-	depends on DRM
-	help
-	  Choose this if you need the GEM CMA helper functions
-
-config DRM_KMS_CMA_HELPER
-	bool
+config DRM_GEM_DMA_HELPER
+	tristate
 	depends on DRM
-	select DRM_GEM_CMA_HELPER
+	select DRM_KMS_HELPER if DRM_FBDEV_EMULATION
+	select FB_CORE if DRM_FBDEV_EMULATION
+	select FB_DMAMEM_HELPERS_DEFERRED if DRM_FBDEV_EMULATION
 	help
-	  Choose this if you need the KMS CMA helper functions
+	  Choose this if you need the GEM DMA helper functions
 
 config DRM_GEM_SHMEM_HELPER
-	bool
+	tristate
 	depends on DRM && MMU
+	select DRM_KMS_HELPER if DRM_FBDEV_EMULATION
+	select FB_CORE if DRM_FBDEV_EMULATION
+	select FB_SYSMEM_HELPERS_DEFERRED if DRM_FBDEV_EMULATION
 	help
 	  Choose this if you need the GEM shmem helper functions
 
+config DRM_SUBALLOC_HELPER
+	tristate
+	depends on DRM
+
 config DRM_SCHED
 	tristate
 	depends on DRM
 
-source "drivers/gpu/drm/i2c/Kconfig"
+source "drivers/gpu/drm/sysfb/Kconfig"
 
 source "drivers/gpu/drm/arm/Kconfig"
 
-config DRM_RADEON
-	tristate "ATI Radeon"
-	depends on DRM && PCI && MMU
-	depends on AGP || !AGP
-	select FW_LOADER
-        select DRM_KMS_HELPER
-        select DRM_TTM
-	select DRM_TTM_HELPER
-	select POWER_SUPPLY
-	select HWMON
-	select BACKLIGHT_CLASS_DEVICE
-	select INTERVAL_TREE
-	help
-	  Choose this option if you have an ATI Radeon graphics card.  There
-	  are both PCI and AGP versions.  You don't need to choose this to
-	  run the Radeon in plain VGA mode.
-
-	  If M is selected, the module will be called radeon.
-
 source "drivers/gpu/drm/radeon/Kconfig"
 
-config DRM_AMDGPU
-	tristate "AMD GPU"
-	depends on DRM && PCI && MMU
-	select FW_LOADER
-	select DRM_KMS_HELPER
-	select DRM_SCHED
-	select DRM_TTM
-	select DRM_TTM_HELPER
-	select POWER_SUPPLY
-	select HWMON
-	select BACKLIGHT_CLASS_DEVICE
-	select INTERVAL_TREE
-	help
-	  Choose this option if you have a recent AMD Radeon graphics card.
-
-	  If M is selected, the module will be called amdgpu.
-
 source "drivers/gpu/drm/amd/amdgpu/Kconfig"
 
 source "drivers/gpu/drm/nouveau/Kconfig"
 
+source "drivers/gpu/drm/nova/Kconfig"
+
 source "drivers/gpu/drm/i915/Kconfig"
 
+source "drivers/gpu/drm/xe/Kconfig"
+
 source "drivers/gpu/drm/kmb/Kconfig"
 
 config DRM_VGEM
@@ -292,19 +292,7 @@ config DRM_VGEM
 	  as used by Mesa's software renderer for enhanced performance.
 	  If M is selected the module will be called vgem.
 
-config DRM_VKMS
-	tristate "Virtual KMS (EXPERIMENTAL)"
-	depends on DRM && MMU
-	select DRM_KMS_HELPER
-	select DRM_GEM_SHMEM_HELPER
-	select CRC32
-	default n
-	help
-	  Virtual Kernel Mode-Setting (VKMS) is used for testing or for
-	  running GPU in a headless machines. Choose this option to get
-	  a VKMS.
-
-	  If M is selected the module will be called vkms.
+source "drivers/gpu/drm/vkms/Kconfig"
 
 source "drivers/gpu/drm/exynos/Kconfig"
 
@@ -324,9 +312,7 @@ source "drivers/gpu/drm/armada/Kconfig"
 
 source "drivers/gpu/drm/atmel-hlcdc/Kconfig"
 
-source "drivers/gpu/drm/rcar-du/Kconfig"
-
-source "drivers/gpu/drm/shmobile/Kconfig"
+source "drivers/gpu/drm/renesas/Kconfig"
 
 source "drivers/gpu/drm/sun4i/Kconfig"
 
@@ -360,10 +346,14 @@ source "drivers/gpu/drm/v3d/Kconfig"
 
 source "drivers/gpu/drm/vc4/Kconfig"
 
+source "drivers/gpu/drm/loongson/Kconfig"
+
 source "drivers/gpu/drm/etnaviv/Kconfig"
 
 source "drivers/gpu/drm/hisilicon/Kconfig"
 
+source "drivers/gpu/drm/logicvc/Kconfig"
+
 source "drivers/gpu/drm/mediatek/Kconfig"
 
 source "drivers/gpu/drm/mxsfb/Kconfig"
@@ -384,19 +374,34 @@ source "drivers/gpu/drm/lima/Kconfig"
 
 source "drivers/gpu/drm/panfrost/Kconfig"
 
+source "drivers/gpu/drm/panthor/Kconfig"
+
 source "drivers/gpu/drm/aspeed/Kconfig"
 
 source "drivers/gpu/drm/mcde/Kconfig"
 
 source "drivers/gpu/drm/tidss/Kconfig"
 
+source "drivers/gpu/drm/adp/Kconfig"
+
 source "drivers/gpu/drm/xlnx/Kconfig"
 
 source "drivers/gpu/drm/gud/Kconfig"
 
+source "drivers/gpu/drm/sitronix/Kconfig"
+
+source "drivers/gpu/drm/solomon/Kconfig"
+
+source "drivers/gpu/drm/sprd/Kconfig"
+
+source "drivers/gpu/drm/imagination/Kconfig"
+
+source "drivers/gpu/drm/tyr/Kconfig"
+
 config DRM_HYPERV
 	tristate "DRM Support for Hyper-V synthetic video device"
-	depends on DRM && PCI && MMU && HYPERV
+	depends on DRM && PCI && HYPERV_VMBUS
+	select DRM_CLIENT_SELECTION
 	select DRM_KMS_HELPER
 	select DRM_GEM_SHMEM_HELPER
 	help
@@ -407,91 +412,20 @@ config DRM_HYPERV
 
 	 If M is selected the module will be called hyperv_drm.
 
-# Keep legacy drivers last
-
-menuconfig DRM_LEGACY
-	bool "Enable legacy drivers (DANGEROUS)"
-	depends on DRM && MMU
-	help
-	  Enable legacy DRI1 drivers. Those drivers expose unsafe and dangerous
-	  APIs to user-space, which can be used to circumvent access
-	  restrictions and other security measures. For backwards compatibility
-	  those drivers are still available, but their use is highly
-	  inadvisable and might harm your system.
-
-	  You are recommended to use the safe modeset-only drivers instead, and
-	  perform 3D emulation in user-space.
-
-	  Unless you have strong reasons to go rogue, say "N".
-
-if DRM_LEGACY
-
-config DRM_TDFX
-	tristate "3dfx Banshee/Voodoo3+"
-	depends on DRM && PCI
-	help
-	  Choose this option if you have a 3dfx Banshee or Voodoo3 (or later),
-	  graphics card.  If M is selected, the module will be called tdfx.
-
-config DRM_R128
-	tristate "ATI Rage 128"
-	depends on DRM && PCI
-	select FW_LOADER
-	help
-	  Choose this option if you have an ATI Rage 128 graphics card.  If M
-	  is selected, the module will be called r128.  AGP support for
-	  this card is strongly suggested (unless you have a PCI version).
-
-config DRM_I810
-	tristate "Intel I810"
-	# !PREEMPTION because of missing ioctl locking
-	depends on DRM && AGP && AGP_INTEL && (!PREEMPTION || BROKEN)
-	help
-	  Choose this option if you have an Intel I810 graphics card.  If M is
-	  selected, the module will be called i810.  AGP support is required
-	  for this driver to work.
-
-config DRM_MGA
-	tristate "Matrox g200/g400"
-	depends on DRM && PCI
-	select FW_LOADER
-	help
-	  Choose this option if you have a Matrox G200, G400 or G450 graphics
-	  card.  If M is selected, the module will be called mga.  AGP
-	  support is required for this driver to work.
-
-config DRM_SIS
-	tristate "SiS video cards"
-	depends on DRM && AGP
-	depends on FB_SIS || FB_SIS=n
-	help
-	  Choose this option if you have a SiS 630 or compatible video
-	  chipset. If M is selected the module will be called sis. AGP
-	  support is required for this driver to work.
-
-config DRM_VIA
-	tristate "Via unichrome video cards"
-	depends on DRM && PCI
-	help
-	  Choose this option if you have a Via unichrome or compatible video
-	  chipset. If M is selected the module will be called via.
-
-config DRM_SAVAGE
-	tristate "Savage video cards"
-	depends on DRM && PCI
-	help
-	  Choose this option if you have a Savage3D/4/SuperSavage/Pro/Twister
-	  chipset. If M is selected the module will be called savage.
+# Separate option as not all DRM drivers use it
+config DRM_PANEL_BACKLIGHT_QUIRKS
+	tristate
 
-endif # DRM_LEGACY
+config DRM_LIB_RANDOM
+	bool
+	default n
 
-config DRM_EXPORT_FOR_TESTS
+config DRM_PRIVACY_SCREEN
 	bool
+	default n
+
+endif
 
 # Separate option because drm_panel_orientation_quirks.c is shared with fbdev
 config DRM_PANEL_ORIENTATION_QUIRKS
 	tristate
-
-config DRM_LIB_RANDOM
-	bool
-	default n
diff --git a/drivers/gpu/drm/Kconfig.debug b/drivers/gpu/drm/Kconfig.debug
new file mode 100644
index 000000000000..05dc43c0b8c5
--- /dev/null
+++ b/drivers/gpu/drm/Kconfig.debug
@@ -0,0 +1,117 @@
+config DRM_USE_DYNAMIC_DEBUG
+	bool "use dynamic debug to implement drm.debug"
+	default n
+	depends on BROKEN
+	depends on DRM
+	depends on DYNAMIC_DEBUG || DYNAMIC_DEBUG_CORE
+	depends on JUMP_LABEL
+	help
+	 Use dynamic-debug to avoid drm_debug_enabled() runtime overheads.
+	 Due to callsite counts in DRM drivers (~4k in amdgpu) and 56
+	 bytes per callsite, the .data costs can be substantial, and
+	 are therefore configurable.
+
+config DRM_WERROR
+	bool "Compile the drm subsystem with warnings as errors"
+	depends on DRM && EXPERT
+	depends on !WERROR
+	default n
+	help
+	  A kernel build should not cause any compiler warnings, and this
+	  enables the '-Werror' flag to enforce that rule in the drm subsystem.
+
+	  The drm subsystem enables more warnings than the kernel default, so
+	  this config option is disabled by default.
+
+	  If in doubt, say N.
+
+config DRM_HEADER_TEST
+	bool "Ensure DRM headers are self-contained and pass kernel-doc"
+	depends on DRM && EXPERT && BROKEN
+	default n
+	help
+	  Ensure the DRM subsystem headers both under drivers/gpu/drm and
+	  include/drm compile, are self-contained, have header guards, and have
+	  no kernel-doc warnings.
+
+	  If in doubt, say N.
+
+config DRM_DEBUG_MM
+	bool "Insert extra checks and debug info into the DRM range managers"
+	default n
+	depends on DRM
+	depends on STACKTRACE_SUPPORT
+	select STACKDEPOT
+	help
+	  Enable allocation tracking of memory manager and leak detection on
+	  shutdown.
+
+	  Recommended for driver developers only.
+
+	  If in doubt, say "N".
+
+config DRM_KUNIT_TEST_HELPERS
+	tristate
+	depends on DRM && KUNIT
+	select DRM_KMS_HELPER
+	help
+	  KUnit Helpers for KMS drivers.
+
+config DRM_KUNIT_TEST
+	tristate "KUnit tests for DRM" if !KUNIT_ALL_TESTS
+	depends on DRM && KUNIT && MMU
+	select DRM_BRIDGE_CONNECTOR
+	select DRM_BUDDY
+	select DRM_DISPLAY_DP_HELPER
+	select DRM_DISPLAY_HDMI_STATE_HELPER
+	select DRM_DISPLAY_HELPER
+	select DRM_EXEC
+	select DRM_EXPORT_FOR_TESTS if m
+	select DRM_GEM_SHMEM_HELPER
+	select DRM_KUNIT_TEST_HELPERS
+	select DRM_LIB_RANDOM
+	select DRM_SYSFB_HELPER
+	select PRIME_NUMBERS
+	default KUNIT_ALL_TESTS
+	help
+	  This builds unit tests for DRM. This option is not useful for
+	  distributions or general kernels, but only for kernel
+	  developers working on DRM and associated drivers.
+
+	  For more information on KUnit and unit tests in general,
+	  please refer to the KUnit documentation in
+	  Documentation/dev-tools/kunit/.
+
+	  If in doubt, say "N".
+
+config DRM_TTM_KUNIT_TEST
+	tristate "KUnit tests for TTM" if !KUNIT_ALL_TESTS
+	default n
+	depends on DRM && KUNIT && MMU && (UML || COMPILE_TEST)
+	select DRM_TTM
+	select DRM_BUDDY
+	select DRM_EXPORT_FOR_TESTS if m
+	select DRM_KUNIT_TEST_HELPERS
+	default KUNIT_ALL_TESTS
+	help
+	  Enables unit tests for TTM, a GPU memory manager subsystem used
+	  to manage memory buffers. This option is mostly useful for kernel
+	  developers. It depends on (UML || COMPILE_TEST) since no other driver
+	  which uses TTM can be loaded while running the tests.
+
+	  If in doubt, say "N".
+
+config DRM_SCHED_KUNIT_TEST
+	tristate "KUnit tests for the DRM scheduler" if !KUNIT_ALL_TESTS
+	select DRM_SCHED
+	depends on DRM && KUNIT
+	default KUNIT_ALL_TESTS
+	help
+	  Choose this option to build unit tests for the DRM scheduler.
+
+	  Recommended for driver developers only.
+
+	  If in doubt, say "N".
+
+config DRM_EXPORT_FOR_TESTS
+	bool
diff --git a/drivers/gpu/drm/Makefile b/drivers/gpu/drm/Makefile
index 0dff40bb863c..0e1c668b46d2 100644
--- a/drivers/gpu/drm/Makefile
+++ b/drivers/gpu/drm/Makefile
@@ -3,88 +3,188 @@
 # Makefile for the drm device driver.  This driver provides support for the
 # Direct Rendering Infrastructure (DRI) in XFree86 4.1.0 and higher.
 
-drm-y       :=	drm_aperture.o drm_auth.o drm_cache.o \
-		drm_file.o drm_gem.o drm_ioctl.o drm_irq.o \
-		drm_drv.o \
-		drm_sysfs.o drm_hashtab.o drm_mm.o \
-		drm_crtc.o drm_fourcc.o drm_modes.o drm_edid.o drm_displayid.o \
-		drm_encoder_slave.o \
-		drm_trace_points.o drm_prime.o \
-		drm_rect.o drm_vma_manager.o drm_flip_work.o \
-		drm_modeset_lock.o drm_atomic.o drm_bridge.o \
-		drm_framebuffer.o drm_connector.o drm_blend.o \
-		drm_encoder.o drm_mode_object.o drm_property.o \
-		drm_plane.o drm_color_mgmt.o drm_print.o \
-		drm_dumb_buffers.o drm_mode_config.o drm_vblank.o \
-		drm_syncobj.o drm_lease.o drm_writeback.o drm_client.o \
-		drm_client_modeset.o drm_atomic_uapi.o drm_hdcp.o \
-		drm_managed.o drm_vblank_work.o
-
-drm-$(CONFIG_DRM_LEGACY) += drm_agpsupport.o drm_bufs.o drm_context.o drm_dma.o \
-			    drm_legacy_misc.o drm_lock.o drm_memory.o drm_scatter.o \
-			    drm_vm.o
+CFLAGS-$(CONFIG_DRM_USE_DYNAMIC_DEBUG)	+= -DDYNAMIC_DEBUG_MODULE
+
+# Unconditionally enable W=1 warnings locally
+# --- begin copy-paste W=1 warnings from scripts/Makefile.warn
+subdir-ccflags-y += -Wextra -Wunused -Wno-unused-parameter
+subdir-ccflags-y += $(call cc-option, -Wrestrict)
+subdir-ccflags-y += -Wmissing-format-attribute
+subdir-ccflags-y += -Wold-style-definition
+subdir-ccflags-y += -Wmissing-include-dirs
+subdir-ccflags-y += $(call cc-option, -Wunused-but-set-variable)
+subdir-ccflags-y += $(call cc-option, -Wunused-const-variable)
+subdir-ccflags-y += $(call cc-option, -Wpacked-not-aligned)
+subdir-ccflags-y += $(call cc-option, -Wformat-overflow)
+# FIXME: fix -Wformat-truncation warnings and uncomment
+#subdir-ccflags-y += $(call cc-option, -Wformat-truncation)
+subdir-ccflags-y += $(call cc-option, -Wstringop-truncation)
+# The following turn off the warnings enabled by -Wextra
+ifeq ($(findstring 2, $(KBUILD_EXTRA_WARN)),)
+subdir-ccflags-y += -Wno-missing-field-initializers
+subdir-ccflags-y += -Wno-type-limits
+subdir-ccflags-y += -Wno-shift-negative-value
+endif
+ifeq ($(findstring 3, $(KBUILD_EXTRA_WARN)),)
+subdir-ccflags-y += -Wno-sign-compare
+endif
+# --- end copy-paste
+
+# Enable -Werror in CI and development
+subdir-ccflags-$(CONFIG_DRM_WERROR) += -Werror
+
+drm-y := \
+	drm_atomic.o \
+	drm_atomic_uapi.o \
+	drm_auth.o \
+	drm_blend.o \
+	drm_bridge.o \
+	drm_cache.o \
+	drm_color_mgmt.o \
+	drm_colorop.o \
+	drm_connector.o \
+	drm_crtc.o \
+	drm_displayid.o \
+	drm_drv.o \
+	drm_dumb_buffers.o \
+	drm_edid.o \
+	drm_eld.o \
+	drm_encoder.o \
+	drm_file.o \
+	drm_fourcc.o \
+	drm_framebuffer.o \
+	drm_gem.o \
+	drm_ioctl.o \
+	drm_lease.o \
+	drm_managed.o \
+	drm_mm.o \
+	drm_mode_config.o \
+	drm_mode_object.o \
+	drm_modes.o \
+	drm_modeset_lock.o \
+	drm_plane.o \
+	drm_prime.o \
+	drm_print.o \
+	drm_property.o \
+	drm_rect.o \
+	drm_syncobj.o \
+	drm_sysfs.o \
+	drm_trace_points.o \
+	drm_vblank.o \
+	drm_vblank_work.o \
+	drm_vma_manager.o \
+	drm_writeback.o
+drm-$(CONFIG_DRM_CLIENT) += \
+	drm_client.o \
+	drm_client_event.o \
+	drm_client_modeset.o \
+	drm_client_sysrq.o
 drm-$(CONFIG_DRM_LIB_RANDOM) += lib/drm_random.o
 drm-$(CONFIG_COMPAT) += drm_ioc32.o
-drm-$(CONFIG_DRM_GEM_CMA_HELPER) += drm_gem_cma_helper.o
-drm-$(CONFIG_DRM_GEM_SHMEM_HELPER) += drm_gem_shmem_helper.o
 drm-$(CONFIG_DRM_PANEL) += drm_panel.o
 drm-$(CONFIG_OF) += drm_of.o
 drm-$(CONFIG_PCI) += drm_pci.o
-drm-$(CONFIG_DEBUG_FS) += drm_debugfs.o drm_debugfs_crc.o
+drm-$(CONFIG_DEBUG_FS) += \
+	drm_debugfs.o \
+	drm_debugfs_crc.o
 drm-$(CONFIG_DRM_LOAD_EDID_FIRMWARE) += drm_edid_load.o
+drm-$(CONFIG_DRM_PRIVACY_SCREEN) += \
+	drm_privacy_screen.o \
+	drm_privacy_screen_x86.o
+drm-$(CONFIG_DRM_ACCEL) += ../../accel/drm_accel.o
+drm-$(CONFIG_DRM_PANIC) += drm_panic.o
+drm-$(CONFIG_DRM_DRAW) += drm_draw.o
+drm-$(CONFIG_DRM_PANIC_SCREEN_QR_CODE) += drm_panic_qr.o
+obj-$(CONFIG_DRM)	+= drm.o
+
+obj-$(CONFIG_DRM_PANEL_ORIENTATION_QUIRKS) += drm_panel_orientation_quirks.o
+obj-$(CONFIG_DRM_PANEL_BACKLIGHT_QUIRKS) += drm_panel_backlight_quirks.o
+
+#
+# Memory-management helpers
+#
+#
+obj-$(CONFIG_DRM_EXEC) += drm_exec.o
+obj-$(CONFIG_DRM_GPUVM) += drm_gpuvm.o
+
+drm_gpusvm_helper-y := \
+	drm_gpusvm.o\
+	drm_pagemap.o
+obj-$(CONFIG_DRM_GPUSVM) += drm_gpusvm_helper.o
+
+obj-$(CONFIG_DRM_BUDDY) += drm_buddy.o
+
+drm_dma_helper-y := drm_gem_dma_helper.o
+drm_dma_helper-$(CONFIG_DRM_FBDEV_EMULATION) += drm_fbdev_dma.o
+drm_dma_helper-$(CONFIG_DRM_KMS_HELPER) += drm_fb_dma_helper.o
+obj-$(CONFIG_DRM_GEM_DMA_HELPER) += drm_dma_helper.o
 
-obj-$(CONFIG_DRM_DP_AUX_BUS) += drm_dp_aux_bus.o
+drm_shmem_helper-y := drm_gem_shmem_helper.o
+drm_shmem_helper-$(CONFIG_DRM_FBDEV_EMULATION) += drm_fbdev_shmem.o
+obj-$(CONFIG_DRM_GEM_SHMEM_HELPER) += drm_shmem_helper.o
+
+drm_suballoc_helper-y := drm_suballoc.o
+obj-$(CONFIG_DRM_SUBALLOC_HELPER) += drm_suballoc_helper.o
 
 drm_vram_helper-y := drm_gem_vram_helper.o
 obj-$(CONFIG_DRM_VRAM_HELPER) += drm_vram_helper.o
 
 drm_ttm_helper-y := drm_gem_ttm_helper.o
+drm_ttm_helper-$(CONFIG_DRM_FBDEV_EMULATION) += drm_fbdev_ttm.o
 obj-$(CONFIG_DRM_TTM_HELPER) += drm_ttm_helper.o
 
-drm_kms_helper-y := drm_bridge_connector.o drm_crtc_helper.o drm_dp_helper.o \
-		drm_dsc.o drm_probe_helper.o \
-		drm_plane_helper.o drm_dp_mst_topology.o drm_atomic_helper.o \
-		drm_kms_helper_common.o drm_dp_dual_mode_helper.o \
-		drm_simple_kms_helper.o drm_modeset_helper.o \
-		drm_scdc_helper.o drm_gem_atomic_helper.o \
-		drm_gem_framebuffer_helper.o \
-		drm_atomic_state_helper.o drm_damage_helper.o \
-		drm_format_helper.o drm_self_refresh_helper.o
+#
+# Modesetting helpers
+#
 
+drm_kms_helper-y := \
+	drm_atomic_helper.o \
+	drm_atomic_state_helper.o \
+	drm_bridge_helper.o \
+	drm_crtc_helper.o \
+	drm_damage_helper.o \
+	drm_flip_work.o \
+	drm_format_helper.o \
+	drm_gem_atomic_helper.o \
+	drm_gem_framebuffer_helper.o \
+	drm_kms_helper_common.o \
+	drm_modeset_helper.o \
+	drm_plane_helper.o \
+	drm_probe_helper.o \
+	drm_self_refresh_helper.o \
+	drm_simple_kms_helper.o \
+	drm_vblank_helper.o
 drm_kms_helper-$(CONFIG_DRM_PANEL_BRIDGE) += bridge/panel.o
 drm_kms_helper-$(CONFIG_DRM_FBDEV_EMULATION) += drm_fb_helper.o
-drm_kms_helper-$(CONFIG_DRM_KMS_CMA_HELPER) += drm_fb_cma_helper.o
-drm_kms_helper-$(CONFIG_DRM_DP_AUX_CHARDEV) += drm_dp_aux_dev.o
-drm_kms_helper-$(CONFIG_DRM_DP_CEC) += drm_dp_cec.o
-
 obj-$(CONFIG_DRM_KMS_HELPER) += drm_kms_helper.o
-obj-$(CONFIG_DRM_DEBUG_SELFTEST) += selftests/
 
-obj-$(CONFIG_DRM)	+= drm.o
+#
+# Drivers and the rest
+#
+
+obj-y			+= tests/
+
 obj-$(CONFIG_DRM_MIPI_DBI) += drm_mipi_dbi.o
 obj-$(CONFIG_DRM_MIPI_DSI) += drm_mipi_dsi.o
-obj-$(CONFIG_DRM_PANEL_ORIENTATION_QUIRKS) += drm_panel_orientation_quirks.o
 obj-y			+= arm/
+obj-y			+= clients/
+obj-y			+= display/
 obj-$(CONFIG_DRM_TTM)	+= ttm/
 obj-$(CONFIG_DRM_SCHED)	+= scheduler/
-obj-$(CONFIG_DRM_TDFX)	+= tdfx/
-obj-$(CONFIG_DRM_R128)	+= r128/
 obj-$(CONFIG_DRM_RADEON)+= radeon/
 obj-$(CONFIG_DRM_AMDGPU)+= amd/amdgpu/
-obj-$(CONFIG_DRM_MGA)	+= mga/
-obj-$(CONFIG_DRM_I810)	+= i810/
+obj-$(CONFIG_DRM_AMDGPU)+= amd/amdxcp/
 obj-$(CONFIG_DRM_I915)	+= i915/
+obj-$(CONFIG_DRM_XE)	+= xe/
 obj-$(CONFIG_DRM_KMB_DISPLAY)  += kmb/
 obj-$(CONFIG_DRM_MGAG200) += mgag200/
 obj-$(CONFIG_DRM_V3D)  += v3d/
 obj-$(CONFIG_DRM_VC4)  += vc4/
-obj-$(CONFIG_DRM_SIS)   += sis/
-obj-$(CONFIG_DRM_SAVAGE)+= savage/
 obj-$(CONFIG_DRM_VMWGFX)+= vmwgfx/
-obj-$(CONFIG_DRM_VIA)	+=via/
 obj-$(CONFIG_DRM_VGEM)	+= vgem/
 obj-$(CONFIG_DRM_VKMS)	+= vkms/
 obj-$(CONFIG_DRM_NOUVEAU) +=nouveau/
+obj-$(CONFIG_DRM_NOVA) += nova/
 obj-$(CONFIG_DRM_EXYNOS) +=exynos/
 obj-$(CONFIG_DRM_ROCKCHIP) +=rockchip/
 obj-$(CONFIG_DRM_GMA500) += gma500/
@@ -92,8 +192,7 @@ obj-$(CONFIG_DRM_UDL) += udl/
 obj-$(CONFIG_DRM_AST) += ast/
 obj-$(CONFIG_DRM_ARMADA) += armada/
 obj-$(CONFIG_DRM_ATMEL_HLCDC)	+= atmel-hlcdc/
-obj-y			+= rcar-du/
-obj-$(CONFIG_DRM_SHMOBILE) +=shmobile/
+obj-y			+= renesas/
 obj-y			+= omapdrm/
 obj-$(CONFIG_DRM_SUN4I) += sun4i/
 obj-y			+= tilcdc/
@@ -105,25 +204,52 @@ obj-$(CONFIG_DRM_STM) += stm/
 obj-$(CONFIG_DRM_STI) += sti/
 obj-y 			+= imx/
 obj-$(CONFIG_DRM_INGENIC) += ingenic/
+obj-$(CONFIG_DRM_LOGICVC) += logicvc/
 obj-$(CONFIG_DRM_MEDIATEK) += mediatek/
 obj-$(CONFIG_DRM_MESON)	+= meson/
-obj-y			+= i2c/
 obj-y			+= panel/
 obj-y			+= bridge/
 obj-$(CONFIG_DRM_FSL_DCU) += fsl-dcu/
 obj-$(CONFIG_DRM_ETNAVIV) += etnaviv/
 obj-y			+= hisilicon/
-obj-$(CONFIG_DRM_MXSFB)	+= mxsfb/
+obj-y			+= mxsfb/
+obj-y			+= sysfb/
 obj-y			+= tiny/
 obj-$(CONFIG_DRM_PL111) += pl111/
 obj-$(CONFIG_DRM_TVE200) += tve200/
+obj-$(CONFIG_DRM_ADP) += adp/
 obj-$(CONFIG_DRM_XEN) += xen/
 obj-$(CONFIG_DRM_VBOXVIDEO) += vboxvideo/
 obj-$(CONFIG_DRM_LIMA)  += lima/
 obj-$(CONFIG_DRM_PANFROST) += panfrost/
+obj-$(CONFIG_DRM_PANTHOR) += panthor/
+obj-$(CONFIG_DRM_TYR) += tyr/
 obj-$(CONFIG_DRM_ASPEED_GFX) += aspeed/
 obj-$(CONFIG_DRM_MCDE) += mcde/
 obj-$(CONFIG_DRM_TIDSS) += tidss/
 obj-y			+= xlnx/
 obj-y			+= gud/
 obj-$(CONFIG_DRM_HYPERV) += hyperv/
+obj-y			+= sitronix/
+obj-y			+= solomon/
+obj-$(CONFIG_DRM_SPRD) += sprd/
+obj-$(CONFIG_DRM_LOONGSON) += loongson/
+obj-$(CONFIG_DRM_POWERVR) += imagination/
+
+# Ensure drm headers are self-contained and pass kernel-doc
+hdrtest-files := \
+	$(shell cd $(src) && find . -maxdepth 1 -name 'drm_*.h') \
+	$(shell cd $(src) && find display lib -name '*.h')
+
+always-$(CONFIG_DRM_HEADER_TEST) += \
+	$(patsubst %.h,%.hdrtest, $(hdrtest-files))
+
+# Include the header twice to detect missing include guard.
+quiet_cmd_hdrtest = HDRTEST $(patsubst %.hdrtest,%.h,$@)
+      cmd_hdrtest = \
+		$(CC) $(c_flags) -fsyntax-only -x c /dev/null -include $< -include $<; \
+		 PYTHONDONTWRITEBYTECODE=1 $(PYTHON3) $(KERNELDOC) -none $(if $(CONFIG_WERROR)$(CONFIG_DRM_WERROR),-Werror) $<; \
+		touch $@
+
+$(obj)/%.hdrtest: $(src)/%.h FORCE
+	$(call if_changed_dep,hdrtest)
diff --git a/drivers/gpu/drm/adp/Kconfig b/drivers/gpu/drm/adp/Kconfig
new file mode 100644
index 000000000000..9fcc27eb200d
--- /dev/null
+++ b/drivers/gpu/drm/adp/Kconfig
@@ -0,0 +1,17 @@
+# SPDX-License-Identifier: GPL-2.0-only OR MIT
+config DRM_ADP
+	tristate "DRM Support for pre-DCP Apple display controllers"
+	depends on DRM && OF && ARM64
+	depends on ARCH_APPLE || COMPILE_TEST
+	select DRM_KMS_HELPER
+	select DRM_BRIDGE_CONNECTOR
+	select DRM_DISPLAY_HELPER
+	select DRM_KMS_DMA_HELPER
+	select DRM_GEM_DMA_HELPER
+	select DRM_PANEL_BRIDGE
+	select VIDEOMODE_HELPERS
+	select DRM_MIPI_DSI
+	help
+	  Chose this option if you have an Apple Arm laptop with a touchbar.
+
+	  If M is selected, this module will be called adpdrm.
diff --git a/drivers/gpu/drm/adp/Makefile b/drivers/gpu/drm/adp/Makefile
new file mode 100644
index 000000000000..8e7b618edd35
--- /dev/null
+++ b/drivers/gpu/drm/adp/Makefile
@@ -0,0 +1,5 @@
+# SPDX-License-Identifier: GPL-2.0-only OR MIT
+
+adpdrm-y := adp_drv.o
+adpdrm-mipi-y := adp-mipi.o
+obj-$(CONFIG_DRM_ADP) += adpdrm.o adpdrm-mipi.o
diff --git a/drivers/gpu/drm/adp/adp-mipi.c b/drivers/gpu/drm/adp/adp-mipi.c
new file mode 100644
index 000000000000..cba7d32150a9
--- /dev/null
+++ b/drivers/gpu/drm/adp/adp-mipi.c
@@ -0,0 +1,277 @@
+// SPDX-License-Identifier: GPL-2.0-only
+
+#include <linux/component.h>
+#include <linux/iopoll.h>
+#include <linux/of.h>
+#include <linux/platform_device.h>
+
+#include <drm/drm_bridge.h>
+#include <drm/drm_mipi_dsi.h>
+
+#define DSI_GEN_HDR 0x6c
+#define DSI_GEN_PLD_DATA 0x70
+
+#define DSI_CMD_PKT_STATUS 0x74
+
+#define GEN_PLD_R_EMPTY BIT(4)
+#define GEN_PLD_W_FULL BIT(3)
+#define GEN_PLD_W_EMPTY BIT(2)
+#define GEN_CMD_FULL BIT(1)
+#define GEN_CMD_EMPTY BIT(0)
+#define GEN_RD_CMD_BUSY BIT(6)
+#define CMD_PKT_STATUS_TIMEOUT_US 20000
+
+struct adp_mipi_drv_private {
+	struct mipi_dsi_host dsi;
+	struct drm_bridge bridge;
+	struct drm_bridge *next_bridge;
+	void __iomem *mipi;
+};
+
+#define mipi_to_adp(x) container_of(x, struct adp_mipi_drv_private, dsi)
+
+static int adp_dsi_gen_pkt_hdr_write(struct adp_mipi_drv_private *adp, u32 hdr_val)
+{
+	int ret;
+	u32 val, mask;
+
+	ret = readl_poll_timeout(adp->mipi + DSI_CMD_PKT_STATUS,
+				 val, !(val & GEN_CMD_FULL), 1000,
+				 CMD_PKT_STATUS_TIMEOUT_US);
+	if (ret) {
+		dev_err(adp->dsi.dev, "failed to get available command FIFO\n");
+		return ret;
+	}
+
+	writel(hdr_val, adp->mipi + DSI_GEN_HDR);
+
+	mask = GEN_CMD_EMPTY | GEN_PLD_W_EMPTY;
+	ret = readl_poll_timeout(adp->mipi + DSI_CMD_PKT_STATUS,
+				 val, (val & mask) == mask,
+				 1000, CMD_PKT_STATUS_TIMEOUT_US);
+	if (ret) {
+		dev_err(adp->dsi.dev, "failed to write command FIFO\n");
+		return ret;
+	}
+
+	return 0;
+}
+
+static int adp_dsi_write(struct adp_mipi_drv_private *adp,
+			 const struct mipi_dsi_packet *packet)
+{
+	const u8 *tx_buf = packet->payload;
+	int len = packet->payload_length, pld_data_bytes = sizeof(u32), ret;
+	__le32 word;
+	u32 val;
+
+	while (len) {
+		if (len < pld_data_bytes) {
+			word = 0;
+			memcpy(&word, tx_buf, len);
+			writel(le32_to_cpu(word), adp->mipi + DSI_GEN_PLD_DATA);
+			len = 0;
+		} else {
+			memcpy(&word, tx_buf, pld_data_bytes);
+			writel(le32_to_cpu(word), adp->mipi + DSI_GEN_PLD_DATA);
+			tx_buf += pld_data_bytes;
+			len -= pld_data_bytes;
+		}
+
+		ret = readl_poll_timeout(adp->mipi + DSI_CMD_PKT_STATUS,
+					 val, !(val & GEN_PLD_W_FULL), 1000,
+					 CMD_PKT_STATUS_TIMEOUT_US);
+		if (ret) {
+			dev_err(adp->dsi.dev,
+				"failed to get available write payload FIFO\n");
+			return ret;
+		}
+	}
+
+	word = 0;
+	memcpy(&word, packet->header, sizeof(packet->header));
+	return adp_dsi_gen_pkt_hdr_write(adp, le32_to_cpu(word));
+}
+
+static int adp_dsi_read(struct adp_mipi_drv_private *adp,
+			const struct mipi_dsi_msg *msg)
+{
+	int i, j, ret, len = msg->rx_len;
+	u8 *buf = msg->rx_buf;
+	u32 val;
+
+	/* Wait end of the read operation */
+	ret = readl_poll_timeout(adp->mipi + DSI_CMD_PKT_STATUS,
+				 val, !(val & GEN_RD_CMD_BUSY),
+				 1000, CMD_PKT_STATUS_TIMEOUT_US);
+	if (ret) {
+		dev_err(adp->dsi.dev, "Timeout during read operation\n");
+		return ret;
+	}
+
+	for (i = 0; i < len; i += 4) {
+		/* Read fifo must not be empty before all bytes are read */
+		ret = readl_poll_timeout(adp->mipi + DSI_CMD_PKT_STATUS,
+					 val, !(val & GEN_PLD_R_EMPTY),
+					 1000, CMD_PKT_STATUS_TIMEOUT_US);
+		if (ret) {
+			dev_err(adp->dsi.dev, "Read payload FIFO is empty\n");
+			return ret;
+		}
+
+		val = readl(adp->mipi + DSI_GEN_PLD_DATA);
+		for (j = 0; j < 4 && j + i < len; j++)
+			buf[i + j] = val >> (8 * j);
+	}
+
+	return ret;
+}
+
+static ssize_t adp_dsi_host_transfer(struct mipi_dsi_host *host,
+				     const struct mipi_dsi_msg *msg)
+{
+	struct adp_mipi_drv_private *adp = mipi_to_adp(host);
+	struct mipi_dsi_packet packet;
+	int ret, nb_bytes;
+
+	ret = mipi_dsi_create_packet(&packet, msg);
+	if (ret) {
+		dev_err(adp->dsi.dev, "failed to create packet: %d\n", ret);
+		return ret;
+	}
+
+	ret = adp_dsi_write(adp, &packet);
+	if (ret)
+		return ret;
+
+	if (msg->rx_buf && msg->rx_len) {
+		ret = adp_dsi_read(adp, msg);
+		if (ret)
+			return ret;
+		nb_bytes = msg->rx_len;
+	} else {
+		nb_bytes = packet.size;
+	}
+
+	return nb_bytes;
+}
+
+static int adp_dsi_bind(struct device *dev, struct device *master, void *data)
+{
+	return 0;
+}
+
+static void adp_dsi_unbind(struct device *dev, struct device *master, void *data)
+{
+}
+
+static const struct component_ops adp_dsi_component_ops = {
+	.bind	= adp_dsi_bind,
+	.unbind	= adp_dsi_unbind,
+};
+
+static int adp_dsi_host_attach(struct mipi_dsi_host *host,
+			       struct mipi_dsi_device *dev)
+{
+	struct adp_mipi_drv_private *adp = mipi_to_adp(host);
+	struct drm_bridge *next;
+	int ret;
+
+	next = devm_drm_of_get_bridge(adp->dsi.dev, adp->dsi.dev->of_node, 1, 0);
+	if (IS_ERR(next))
+		return PTR_ERR(next);
+
+	adp->next_bridge = next;
+
+	drm_bridge_add(&adp->bridge);
+
+	ret = component_add(host->dev, &adp_dsi_component_ops);
+	if (ret) {
+		pr_err("failed to add dsi_host component: %d\n", ret);
+		drm_bridge_remove(&adp->bridge);
+		return ret;
+	}
+
+	return 0;
+}
+
+static int adp_dsi_host_detach(struct mipi_dsi_host *host,
+			       struct mipi_dsi_device *dev)
+{
+	struct adp_mipi_drv_private *adp = mipi_to_adp(host);
+
+	component_del(host->dev, &adp_dsi_component_ops);
+	drm_bridge_remove(&adp->bridge);
+	return 0;
+}
+
+static const struct mipi_dsi_host_ops adp_dsi_host_ops = {
+	.transfer = adp_dsi_host_transfer,
+	.attach = adp_dsi_host_attach,
+	.detach = adp_dsi_host_detach,
+};
+
+static int adp_dsi_bridge_attach(struct drm_bridge *bridge,
+				 struct drm_encoder *encoder,
+				 enum drm_bridge_attach_flags flags)
+{
+	struct adp_mipi_drv_private *adp =
+		container_of(bridge, struct adp_mipi_drv_private, bridge);
+
+	return drm_bridge_attach(encoder, adp->next_bridge, bridge, flags);
+}
+
+static const struct drm_bridge_funcs adp_dsi_bridge_funcs = {
+	.attach	= adp_dsi_bridge_attach,
+};
+
+static int adp_mipi_probe(struct platform_device *pdev)
+{
+	struct adp_mipi_drv_private *adp;
+
+	adp = devm_drm_bridge_alloc(&pdev->dev, struct adp_mipi_drv_private,
+				    bridge, &adp_dsi_bridge_funcs);
+	if (IS_ERR(adp))
+		return PTR_ERR(adp);
+
+	adp->mipi = devm_platform_ioremap_resource(pdev, 0);
+	if (IS_ERR(adp->mipi)) {
+		dev_err(&pdev->dev, "failed to map mipi mmio");
+		return PTR_ERR(adp->mipi);
+	}
+
+	adp->dsi.dev = &pdev->dev;
+	adp->dsi.ops = &adp_dsi_host_ops;
+	adp->bridge.of_node = pdev->dev.of_node;
+	adp->bridge.type = DRM_MODE_CONNECTOR_DSI;
+	dev_set_drvdata(&pdev->dev, adp);
+	return mipi_dsi_host_register(&adp->dsi);
+}
+
+static void adp_mipi_remove(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct adp_mipi_drv_private *adp = dev_get_drvdata(dev);
+
+	mipi_dsi_host_unregister(&adp->dsi);
+}
+
+static const struct of_device_id adp_mipi_of_match[] = {
+	{ .compatible = "apple,h7-display-pipe-mipi", },
+	{ },
+};
+MODULE_DEVICE_TABLE(of, adp_mipi_of_match);
+
+static struct platform_driver adp_mipi_platform_driver = {
+	.driver = {
+		.name = "adp-mipi",
+		.of_match_table = adp_mipi_of_match,
+	},
+	.probe = adp_mipi_probe,
+	.remove = adp_mipi_remove,
+};
+
+module_platform_driver(adp_mipi_platform_driver);
+
+MODULE_DESCRIPTION("Apple Display Pipe MIPI driver");
+MODULE_LICENSE("GPL");
diff --git a/drivers/gpu/drm/adp/adp_drv.c b/drivers/gpu/drm/adp/adp_drv.c
new file mode 100644
index 000000000000..4554cf75565e
--- /dev/null
+++ b/drivers/gpu/drm/adp/adp_drv.c
@@ -0,0 +1,614 @@
+// SPDX-License-Identifier: GPL-2.0-only
+
+#include <linux/component.h>
+#include <linux/iopoll.h>
+#include <linux/of.h>
+#include <linux/platform_device.h>
+
+#include <drm/drm_atomic.h>
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_bridge.h>
+#include <drm/drm_bridge_connector.h>
+#include <drm/drm_drv.h>
+#include <drm/drm_fb_dma_helper.h>
+#include <drm/drm_framebuffer.h>
+#include <drm/drm_gem_atomic_helper.h>
+#include <drm/drm_gem_dma_helper.h>
+#include <drm/drm_gem_framebuffer_helper.h>
+#include <drm/drm_of.h>
+#include <drm/drm_print.h>
+#include <drm/drm_probe_helper.h>
+#include <drm/drm_vblank.h>
+
+#define ADP_INT_STATUS 0x34
+#define ADP_INT_STATUS_INT_MASK 0x7
+#define ADP_INT_STATUS_VBLANK 0x1
+#define ADP_CTRL 0x100
+#define ADP_CTRL_VBLANK_ON 0x12
+#define ADP_CTRL_FIFO_ON 0x601
+#define ADP_SCREEN_SIZE 0x0c
+#define ADP_SCREEN_HSIZE GENMASK(15, 0)
+#define ADP_SCREEN_VSIZE GENMASK(31, 16)
+
+#define ADBE_FIFO 0x10c0
+#define ADBE_FIFO_SYNC 0xc0000000
+
+#define ADBE_BLEND_BYPASS 0x2020
+#define ADBE_BLEND_EN1 0x2028
+#define ADBE_BLEND_EN2 0x2074
+#define ADBE_BLEND_EN3 0x202c
+#define ADBE_BLEND_EN4 0x2034
+#define ADBE_MASK_BUF 0x2200
+
+#define ADBE_SRC_START 0x4040
+#define ADBE_SRC_SIZE 0x4048
+#define ADBE_DST_START 0x4050
+#define ADBE_DST_SIZE 0x4054
+#define ADBE_STRIDE 0x4038
+#define ADBE_FB_BASE 0x4030
+
+#define ADBE_LAYER_EN1 0x4020
+#define ADBE_LAYER_EN2 0x4068
+#define ADBE_LAYER_EN3 0x40b4
+#define ADBE_LAYER_EN4 0x40f4
+#define ADBE_SCALE_CTL 0x40ac
+#define ADBE_SCALE_CTL_BYPASS 0x100000
+
+#define ADBE_LAYER_CTL 0x1038
+#define ADBE_LAYER_CTL_ENABLE 0x10000
+
+#define ADBE_PIX_FMT 0x402c
+#define ADBE_PIX_FMT_XRGB32 0x53e4001
+
+static int adp_open(struct inode *inode, struct file *filp)
+{
+	/*
+	 * The modesetting driver does not check the non-desktop connector
+	 * property and keeps the device open and locked. If the touchbar daemon
+	 * opens the device first, modesetting breaks the whole X session.
+	 * Simply refuse to open the device for X11 server processes as
+	 * workaround.
+	 */
+	if (current->comm[0] == 'X')
+		return -EBUSY;
+
+	return drm_open(inode, filp);
+}
+
+static const struct file_operations adp_fops = {
+	.owner          = THIS_MODULE,
+	.open           = adp_open,
+	.release        = drm_release,
+	.unlocked_ioctl = drm_ioctl,
+	.compat_ioctl   = drm_compat_ioctl,
+	.poll           = drm_poll,
+	.read           = drm_read,
+	.llseek         = noop_llseek,
+	.mmap           = drm_gem_mmap,
+	.fop_flags      = FOP_UNSIGNED_OFFSET,
+	DRM_GEM_DMA_UNMAPPED_AREA_FOPS
+};
+
+static int adp_drm_gem_dumb_create(struct drm_file *file_priv,
+					struct drm_device *drm,
+					struct drm_mode_create_dumb *args)
+{
+	args->height = ALIGN(args->height, 64);
+	args->size = args->pitch * args->height;
+
+	return drm_gem_dma_dumb_create_internal(file_priv, drm, args);
+}
+
+static const struct drm_driver adp_driver = {
+	.driver_features = DRIVER_GEM | DRIVER_MODESET | DRIVER_ATOMIC,
+	.fops = &adp_fops,
+	DRM_GEM_DMA_DRIVER_OPS_VMAP_WITH_DUMB_CREATE(adp_drm_gem_dumb_create),
+	.name = "adp",
+	.desc = "Apple Display Pipe DRM Driver",
+	.major = 0,
+	.minor = 1,
+};
+
+struct adp_drv_private {
+	struct drm_device drm;
+	struct drm_crtc crtc;
+	struct drm_encoder *encoder;
+	struct drm_connector *connector;
+	struct drm_bridge *next_bridge;
+	void __iomem *be;
+	void __iomem *fe;
+	u32 *mask_buf;
+	u64 mask_buf_size;
+	dma_addr_t mask_iova;
+	int be_irq;
+	int fe_irq;
+	struct drm_pending_vblank_event *event;
+};
+
+#define to_adp(x) container_of(x, struct adp_drv_private, drm)
+#define crtc_to_adp(x) container_of(x, struct adp_drv_private, crtc)
+
+static int adp_plane_atomic_check(struct drm_plane *plane,
+				    struct drm_atomic_state *state)
+{
+	struct drm_plane_state *new_plane_state;
+	struct drm_crtc_state *crtc_state;
+
+	new_plane_state = drm_atomic_get_new_plane_state(state, plane);
+
+	if (!new_plane_state->crtc)
+		return 0;
+
+	crtc_state = drm_atomic_get_crtc_state(state, new_plane_state->crtc);
+	if (IS_ERR(crtc_state))
+		return PTR_ERR(crtc_state);
+
+	return drm_atomic_helper_check_plane_state(new_plane_state,
+						   crtc_state,
+						   DRM_PLANE_NO_SCALING,
+						   DRM_PLANE_NO_SCALING,
+						   true, true);
+}
+
+static void adp_plane_atomic_update(struct drm_plane *plane,
+				    struct drm_atomic_state *state)
+{
+	struct adp_drv_private *adp;
+	struct drm_rect src_rect;
+	struct drm_gem_dma_object *obj;
+	struct drm_framebuffer *fb;
+	struct drm_plane_state *new_state = drm_atomic_get_new_plane_state(state, plane);
+	u32 src_pos, src_size, dst_pos, dst_size;
+
+	if (!plane || !new_state)
+		return;
+
+	fb = new_state->fb;
+	if (!fb)
+		return;
+	adp = to_adp(plane->dev);
+
+	drm_rect_fp_to_int(&src_rect, &new_state->src);
+	src_pos = src_rect.x1 << 16 | src_rect.y1;
+	dst_pos = new_state->dst.x1 << 16 | new_state->dst.y1;
+	src_size = drm_rect_width(&src_rect) << 16 | drm_rect_height(&src_rect);
+	dst_size = drm_rect_width(&new_state->dst) << 16 |
+		drm_rect_height(&new_state->dst);
+	writel(src_pos, adp->be + ADBE_SRC_START);
+	writel(src_size, adp->be + ADBE_SRC_SIZE);
+	writel(dst_pos, adp->be + ADBE_DST_START);
+	writel(dst_size, adp->be + ADBE_DST_SIZE);
+	writel(fb->pitches[0], adp->be + ADBE_STRIDE);
+	obj = drm_fb_dma_get_gem_obj(fb, 0);
+	if (obj)
+		writel(obj->dma_addr + fb->offsets[0], adp->be + ADBE_FB_BASE);
+
+	writel(BIT(0), adp->be + ADBE_LAYER_EN1);
+	writel(BIT(0), adp->be + ADBE_LAYER_EN2);
+	writel(BIT(0), adp->be + ADBE_LAYER_EN3);
+	writel(BIT(0), adp->be + ADBE_LAYER_EN4);
+	writel(ADBE_SCALE_CTL_BYPASS, adp->be + ADBE_SCALE_CTL);
+	writel(ADBE_LAYER_CTL_ENABLE | BIT(0), adp->be + ADBE_LAYER_CTL);
+	writel(ADBE_PIX_FMT_XRGB32, adp->be + ADBE_PIX_FMT);
+}
+
+static void adp_plane_atomic_disable(struct drm_plane *plane,
+				     struct drm_atomic_state *state)
+{
+	struct adp_drv_private *adp = to_adp(plane->dev);
+
+	writel(0x0, adp->be + ADBE_LAYER_EN1);
+	writel(0x0, adp->be + ADBE_LAYER_EN2);
+	writel(0x0, adp->be + ADBE_LAYER_EN3);
+	writel(0x0, adp->be + ADBE_LAYER_EN4);
+	writel(ADBE_LAYER_CTL_ENABLE, adp->be + ADBE_LAYER_CTL);
+}
+
+static const struct drm_plane_helper_funcs adp_plane_helper_funcs = {
+	.atomic_check = adp_plane_atomic_check,
+	.atomic_update = adp_plane_atomic_update,
+	.atomic_disable = adp_plane_atomic_disable,
+	DRM_GEM_SHADOW_PLANE_HELPER_FUNCS
+};
+
+static const struct drm_plane_funcs adp_plane_funcs = {
+	.update_plane = drm_atomic_helper_update_plane,
+	.disable_plane = drm_atomic_helper_disable_plane,
+	DRM_GEM_SHADOW_PLANE_FUNCS
+};
+
+static const u32 plane_formats[] = {
+	DRM_FORMAT_XRGB8888,
+};
+
+#define ALL_CRTCS 1
+
+static struct drm_plane *adp_plane_new(struct adp_drv_private *adp)
+{
+	struct drm_device *drm = &adp->drm;
+	struct drm_plane *plane;
+
+	plane = __drmm_universal_plane_alloc(drm, sizeof(struct drm_plane), 0,
+					     ALL_CRTCS, &adp_plane_funcs,
+					     plane_formats, ARRAY_SIZE(plane_formats),
+					     NULL, DRM_PLANE_TYPE_PRIMARY, "plane");
+	if (IS_ERR(plane)) {
+		drm_err(drm, "failed to allocate plane");
+		return plane;
+	}
+
+	drm_plane_helper_add(plane, &adp_plane_helper_funcs);
+	return plane;
+}
+
+static void adp_enable_vblank(struct adp_drv_private *adp)
+{
+	u32 cur_ctrl;
+
+	writel(ADP_INT_STATUS_INT_MASK, adp->fe + ADP_INT_STATUS);
+
+	cur_ctrl = readl(adp->fe + ADP_CTRL);
+	writel(cur_ctrl | ADP_CTRL_VBLANK_ON, adp->fe + ADP_CTRL);
+}
+
+static int adp_crtc_enable_vblank(struct drm_crtc *crtc)
+{
+	struct drm_device *dev = crtc->dev;
+	struct adp_drv_private *adp = to_adp(dev);
+
+	adp_enable_vblank(adp);
+
+	return 0;
+}
+
+static void adp_disable_vblank(struct adp_drv_private *adp)
+{
+	u32 cur_ctrl;
+
+	cur_ctrl = readl(adp->fe + ADP_CTRL);
+	writel(cur_ctrl & ~ADP_CTRL_VBLANK_ON, adp->fe + ADP_CTRL);
+	writel(ADP_INT_STATUS_INT_MASK, adp->fe + ADP_INT_STATUS);
+}
+
+static void adp_crtc_disable_vblank(struct drm_crtc *crtc)
+{
+	struct drm_device *dev = crtc->dev;
+	struct adp_drv_private *adp = to_adp(dev);
+
+	adp_disable_vblank(adp);
+}
+
+static void adp_crtc_atomic_enable(struct drm_crtc *crtc,
+				   struct drm_atomic_state *state)
+{
+	struct adp_drv_private *adp = crtc_to_adp(crtc);
+
+	writel(BIT(0), adp->be + ADBE_BLEND_EN2);
+	writel(BIT(4), adp->be + ADBE_BLEND_EN1);
+	writel(BIT(0), adp->be + ADBE_BLEND_EN3);
+	writel(BIT(0), adp->be + ADBE_BLEND_BYPASS);
+	writel(BIT(0), adp->be + ADBE_BLEND_EN4);
+	drm_crtc_vblank_on(crtc);
+}
+
+static void adp_crtc_atomic_disable(struct drm_crtc *crtc,
+				    struct drm_atomic_state *state)
+{
+	struct adp_drv_private *adp = crtc_to_adp(crtc);
+	struct drm_crtc_state *old_state = drm_atomic_get_old_crtc_state(state, crtc);
+
+	drm_atomic_helper_disable_planes_on_crtc(old_state, false);
+
+	writel(0x0, adp->be + ADBE_BLEND_EN2);
+	writel(0x0, adp->be + ADBE_BLEND_EN1);
+	writel(0x0, adp->be + ADBE_BLEND_EN3);
+	writel(0x0, adp->be + ADBE_BLEND_BYPASS);
+	writel(0x0, adp->be + ADBE_BLEND_EN4);
+	drm_crtc_vblank_off(crtc);
+}
+
+static void adp_crtc_atomic_flush(struct drm_crtc *crtc,
+				  struct drm_atomic_state *state)
+{
+	u32 frame_num = 1;
+	unsigned long flags;
+	struct adp_drv_private *adp = crtc_to_adp(crtc);
+	struct drm_crtc_state *new_state = drm_atomic_get_new_crtc_state(state, crtc);
+	u64 new_size = ALIGN(new_state->mode.hdisplay *
+			     new_state->mode.vdisplay * 4, PAGE_SIZE);
+
+	if (new_size != adp->mask_buf_size) {
+		if (adp->mask_buf)
+			dma_free_coherent(crtc->dev->dev, adp->mask_buf_size,
+					  adp->mask_buf, adp->mask_iova);
+		adp->mask_buf = NULL;
+		if (new_size != 0) {
+			adp->mask_buf = dma_alloc_coherent(crtc->dev->dev, new_size,
+							   &adp->mask_iova, GFP_KERNEL);
+			memset(adp->mask_buf, 0xFF, new_size);
+			writel(adp->mask_iova, adp->be + ADBE_MASK_BUF);
+		}
+		adp->mask_buf_size = new_size;
+	}
+	writel(ADBE_FIFO_SYNC | frame_num, adp->be + ADBE_FIFO);
+	//FIXME: use adbe flush interrupt
+	if (crtc->state->event) {
+		struct drm_pending_vblank_event *event = crtc->state->event;
+
+		crtc->state->event = NULL;
+		spin_lock_irqsave(&crtc->dev->event_lock, flags);
+
+		if (drm_crtc_vblank_get(crtc) != 0)
+			drm_crtc_send_vblank_event(crtc, event);
+		else
+			adp->event = event;
+
+		spin_unlock_irqrestore(&crtc->dev->event_lock, flags);
+	}
+}
+
+static const struct drm_crtc_funcs adp_crtc_funcs = {
+	.destroy = drm_crtc_cleanup,
+	.set_config = drm_atomic_helper_set_config,
+	.page_flip = drm_atomic_helper_page_flip,
+	.reset = drm_atomic_helper_crtc_reset,
+	.atomic_duplicate_state = drm_atomic_helper_crtc_duplicate_state,
+	.atomic_destroy_state = drm_atomic_helper_crtc_destroy_state,
+	.enable_vblank = adp_crtc_enable_vblank,
+	.disable_vblank = adp_crtc_disable_vblank,
+};
+
+
+static const struct drm_crtc_helper_funcs adp_crtc_helper_funcs = {
+	.atomic_enable = adp_crtc_atomic_enable,
+	.atomic_disable = adp_crtc_atomic_disable,
+	.atomic_flush = adp_crtc_atomic_flush,
+};
+
+static int adp_setup_crtc(struct adp_drv_private *adp)
+{
+	struct drm_device *drm = &adp->drm;
+	struct drm_plane *primary;
+	int ret;
+
+	primary = adp_plane_new(adp);
+	if (IS_ERR(primary))
+		return PTR_ERR(primary);
+
+	ret = drm_crtc_init_with_planes(drm, &adp->crtc, primary,
+					NULL, &adp_crtc_funcs, NULL);
+	if (ret)
+		return ret;
+
+	drm_crtc_helper_add(&adp->crtc, &adp_crtc_helper_funcs);
+	return 0;
+}
+
+static const struct drm_mode_config_funcs adp_mode_config_funcs = {
+	.fb_create = drm_gem_fb_create_with_dirty,
+	.atomic_check = drm_atomic_helper_check,
+	.atomic_commit = drm_atomic_helper_commit,
+};
+
+static int adp_setup_mode_config(struct adp_drv_private *adp)
+{
+	struct drm_device *drm = &adp->drm;
+	int ret;
+	u32 size;
+
+	ret = drmm_mode_config_init(drm);
+	if (ret)
+		return ret;
+
+	/*
+	 * Query screen size restrict the frame buffer size to the screen size
+	 * aligned to the next multiple of 64. This is not necessary but can be
+	 * used as simple check for non-desktop devices.
+	 * Xorg's modesetting driver does not care about the connector
+	 * "non-desktop" property. The max frame buffer width or height can be
+	 * easily checked and a device can be reject if the max width/height is
+	 * smaller than 120 for example.
+	 * Any touchbar daemon is not limited by this small framebuffer size.
+	 */
+	size = readl(adp->fe + ADP_SCREEN_SIZE);
+
+	drm->mode_config.min_width = 32;
+	drm->mode_config.min_height = 32;
+	drm->mode_config.max_width = ALIGN(FIELD_GET(ADP_SCREEN_HSIZE, size), 64);
+	drm->mode_config.max_height = ALIGN(FIELD_GET(ADP_SCREEN_VSIZE, size), 64);
+	drm->mode_config.preferred_depth = 24;
+	drm->mode_config.prefer_shadow = 0;
+	drm->mode_config.funcs = &adp_mode_config_funcs;
+
+	ret = adp_setup_crtc(adp);
+	if (ret) {
+		drm_err(drm, "failed to create crtc");
+		return ret;
+	}
+
+	adp->encoder = drmm_plain_encoder_alloc(drm, NULL, DRM_MODE_ENCODER_DSI, NULL);
+	if (IS_ERR(adp->encoder)) {
+		drm_err(drm, "failed to init encoder");
+		return PTR_ERR(adp->encoder);
+	}
+	adp->encoder->possible_crtcs = ALL_CRTCS;
+
+	ret = drm_bridge_attach(adp->encoder, adp->next_bridge, NULL,
+				DRM_BRIDGE_ATTACH_NO_CONNECTOR);
+	if (ret) {
+		drm_err(drm, "failed to init bridge chain");
+		return ret;
+	}
+
+	adp->connector = drm_bridge_connector_init(drm, adp->encoder);
+	if (IS_ERR(adp->connector))
+		return PTR_ERR(adp->connector);
+
+	drm_connector_attach_encoder(adp->connector, adp->encoder);
+
+	ret = drm_vblank_init(drm, drm->mode_config.num_crtc);
+	if (ret < 0) {
+		drm_err(drm, "failed to initialize vblank");
+		return ret;
+	}
+
+	drm_mode_config_reset(drm);
+
+	return 0;
+}
+
+static int adp_parse_of(struct platform_device *pdev, struct adp_drv_private *adp)
+{
+	struct device *dev = &pdev->dev;
+
+	adp->be = devm_platform_ioremap_resource_byname(pdev, "be");
+	if (IS_ERR(adp->be)) {
+		dev_err(dev, "failed to map display backend mmio");
+		return PTR_ERR(adp->be);
+	}
+
+	adp->fe = devm_platform_ioremap_resource_byname(pdev, "fe");
+	if (IS_ERR(adp->fe)) {
+		dev_err(dev, "failed to map display pipe mmio");
+		return PTR_ERR(adp->fe);
+	}
+
+	adp->be_irq = platform_get_irq_byname(pdev, "be");
+	if (adp->be_irq < 0)
+		return adp->be_irq;
+
+	adp->fe_irq = platform_get_irq_byname(pdev, "fe");
+	if (adp->fe_irq < 0)
+		return adp->fe_irq;
+
+	return 0;
+}
+
+static irqreturn_t adp_fe_irq(int irq, void *arg)
+{
+	struct adp_drv_private *adp = (struct adp_drv_private *)arg;
+	u32 int_status;
+	u32 int_ctl;
+
+	int_status = readl(adp->fe + ADP_INT_STATUS);
+	if (int_status & ADP_INT_STATUS_VBLANK) {
+		drm_crtc_handle_vblank(&adp->crtc);
+		spin_lock(&adp->crtc.dev->event_lock);
+		if (adp->event) {
+			int_ctl = readl(adp->fe + ADP_CTRL);
+			if ((int_ctl & 0xF00) == 0x600) {
+				drm_crtc_send_vblank_event(&adp->crtc, adp->event);
+				adp->event = NULL;
+				drm_crtc_vblank_put(&adp->crtc);
+			}
+		}
+		spin_unlock(&adp->crtc.dev->event_lock);
+	}
+
+	writel(int_status, adp->fe + ADP_INT_STATUS);
+
+
+	return IRQ_HANDLED;
+}
+
+static int adp_drm_bind(struct device *dev)
+{
+	struct drm_device *drm = dev_get_drvdata(dev);
+	struct adp_drv_private *adp = to_adp(drm);
+	int err;
+
+	writel(ADP_CTRL_FIFO_ON, adp->fe + ADP_CTRL);
+
+	adp->next_bridge = drmm_of_get_bridge(&adp->drm, dev->of_node, 0, 0);
+	if (IS_ERR(adp->next_bridge)) {
+		dev_err(dev, "failed to find next bridge");
+		return PTR_ERR(adp->next_bridge);
+	}
+
+	err = adp_setup_mode_config(adp);
+	if (err < 0)
+		return err;
+
+	err = request_irq(adp->fe_irq, adp_fe_irq, 0, "adp-fe", adp);
+	if (err)
+		return err;
+
+	err = drm_dev_register(&adp->drm, 0);
+	if (err)
+		return err;
+
+	return 0;
+}
+
+static void adp_drm_unbind(struct device *dev)
+{
+	struct drm_device *drm = dev_get_drvdata(dev);
+	struct adp_drv_private *adp = to_adp(drm);
+
+	drm_dev_unregister(drm);
+	drm_atomic_helper_shutdown(drm);
+	free_irq(adp->fe_irq, adp);
+}
+
+static const struct component_master_ops adp_master_ops = {
+	.bind	= adp_drm_bind,
+	.unbind = adp_drm_unbind,
+};
+
+static int compare_dev(struct device *dev, void *data)
+{
+	return dev->of_node == data;
+}
+
+static int adp_probe(struct platform_device *pdev)
+{
+	struct device_node *port;
+	struct component_match *match = NULL;
+	struct adp_drv_private *adp;
+	int err;
+
+	adp = devm_drm_dev_alloc(&pdev->dev, &adp_driver, struct adp_drv_private, drm);
+	if (IS_ERR(adp))
+		return PTR_ERR(adp);
+
+	dev_set_drvdata(&pdev->dev, &adp->drm);
+
+	err = adp_parse_of(pdev, adp);
+	if (err < 0)
+		return err;
+
+	port = of_graph_get_remote_node(pdev->dev.of_node, 0, 0);
+	if (!port)
+		return -ENODEV;
+
+	drm_of_component_match_add(&pdev->dev, &match, compare_dev, port);
+	of_node_put(port);
+
+	return component_master_add_with_match(&pdev->dev, &adp_master_ops, match);
+}
+
+static void adp_remove(struct platform_device *pdev)
+{
+	component_master_del(&pdev->dev, &adp_master_ops);
+	dev_set_drvdata(&pdev->dev, NULL);
+}
+
+static const struct of_device_id adp_of_match[] = {
+	{ .compatible = "apple,h7-display-pipe", },
+	{ },
+};
+MODULE_DEVICE_TABLE(of, adp_of_match);
+
+static struct platform_driver adp_platform_driver = {
+	.driver = {
+		.name = "adp",
+		.of_match_table = adp_of_match,
+	},
+	.probe = adp_probe,
+	.remove = adp_remove,
+};
+
+module_platform_driver(adp_platform_driver);
+
+MODULE_DESCRIPTION("Apple Display Pipe DRM driver");
+MODULE_LICENSE("GPL");
diff --git a/drivers/gpu/drm/amd/acp/include/acp_gfx_if.h b/drivers/gpu/drm/amd/acp/include/acp_gfx_if.h
index feab8eb7f2a8..b26710cae801 100644
--- a/drivers/gpu/drm/amd/acp/include/acp_gfx_if.h
+++ b/drivers/gpu/drm/amd/acp/include/acp_gfx_if.h
@@ -19,7 +19,7 @@
  * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
  * OTHER DEALINGS IN THE SOFTWARE.
  *
-*/
+ */
 
 #ifndef _ACP_GFX_IF_H
 #define _ACP_GFX_IF_H
diff --git a/drivers/gpu/drm/amd/amdgpu/Kconfig b/drivers/gpu/drm/amd/amdgpu/Kconfig
index 74a8105fd2c0..7f515be5185d 100644
--- a/drivers/gpu/drm/amd/amdgpu/Kconfig
+++ b/drivers/gpu/drm/amd/amdgpu/Kconfig
@@ -1,39 +1,108 @@
 # SPDX-License-Identifier: MIT
+
+config DRM_AMDGPU
+	tristate "AMD GPU"
+	depends on DRM && PCI
+	depends on !UML
+	select FW_LOADER
+	select DRM_CLIENT
+	select DRM_CLIENT_SELECTION
+	select DRM_DISPLAY_DP_HELPER
+	select DRM_DISPLAY_DSC_HELPER
+	select DRM_DISPLAY_HDMI_HELPER
+	select DRM_DISPLAY_HDCP_HELPER
+	select DRM_DISPLAY_HELPER
+	select DRM_KMS_HELPER
+	select DRM_SCHED
+	select DRM_TTM
+	select DRM_TTM_HELPER
+	select POWER_SUPPLY
+	select HWMON
+	select I2C
+	select I2C_ALGOBIT
+	select CRC16
+	select BACKLIGHT_CLASS_DEVICE
+	select INTERVAL_TREE
+	select DRM_BUDDY
+	select DRM_SUBALLOC_HELPER
+	select DRM_EXEC
+	select DRM_PANEL_BACKLIGHT_QUIRKS
+	# amdgpu depends on ACPI_VIDEO when ACPI is enabled, for select to work
+	# ACPI_VIDEO's dependencies must also be selected.
+	select INPUT if ACPI
+	select ACPI_VIDEO if ACPI
+	# On x86 ACPI_VIDEO also needs ACPI_WMI
+	select X86_PLATFORM_DEVICES if ACPI && X86
+	select ACPI_WMI if ACPI && X86
+	help
+	  Choose this option if you have a recent AMD Radeon graphics card.
+
+	  If M is selected, the module will be called amdgpu.
+
 config DRM_AMDGPU_SI
 	bool "Enable amdgpu support for SI parts"
 	depends on DRM_AMDGPU
 	help
-	  Choose this option if you want to enable experimental support
-	  for SI asics.
+	  Choose this option if you want to enable support
+	  for SI (Southern Islands) asics.
 
-	  SI is already supported in radeon. Experimental support for SI
-	  in amdgpu will be disabled by default and is still provided by
-	  radeon. Use module options to override this:
+	  SI (Southern Islands) are first generation GCN GPUs,
+	  supported by both drivers: radeon (old) and amdgpu (new).
+	  By default, SI dedicated GPUs are supported by amdgpu.
 
-	  radeon.si_support=0 amdgpu.si_support=1
+	  Use module options to override this:
+	  To use radeon for SI,
+	  radeon.si_support=1 amdgpu.si_support=0
 
 config DRM_AMDGPU_CIK
 	bool "Enable amdgpu support for CIK parts"
 	depends on DRM_AMDGPU
 	help
-	  Choose this option if you want to enable support for CIK asics.
+	  Choose this option if you want to enable support for CIK (Sea
+	  Islands) asics.
 
-	  CIK is already supported in radeon. Support for CIK in amdgpu
-	  will be disabled by default and is still provided by radeon.
-	  Use module options to override this:
+	  CIK (Sea Islands) are second generation GCN GPUs,
+	  supported by both drivers: radeon (old) and amdgpu (new).
+	  By default,
+	  CIK dedicated GPUs are supported by amdgpu
+	  CIK APUs are supported by radeon
 
+	  Use module options to override this:
+	  To use amdgpu for CIK,
 	  radeon.cik_support=0 amdgpu.cik_support=1
+	  To use radeon for CIK,
+	  radeon.cik_support=1 amdgpu.cik_support=0
 
 config DRM_AMDGPU_USERPTR
 	bool "Always enable userptr write support"
 	depends on DRM_AMDGPU
-	depends on MMU
 	select HMM_MIRROR
 	select MMU_NOTIFIER
 	help
 	  This option selects CONFIG_HMM and CONFIG_HMM_MIRROR if it
 	  isn't already selected to enabled full userptr support.
 
+config DRM_AMD_ISP
+	bool "Enable AMD Image Signal Processor IP support"
+	depends on DRM_AMDGPU && ACPI
+	select MFD_CORE
+	select PM_GENERIC_DOMAINS if PM
+	help
+	Choose this option to enable ISP IP support for AMD SOCs.
+	This adds the ISP (Image Signal Processor) IP driver and wires
+	it up into the amdgpu driver.  It is required for cameras
+	on APUs which utilize mipi cameras.
+
+config DRM_AMDGPU_WERROR
+	bool "Force the compiler to throw an error instead of a warning when compiling"
+	depends on DRM_AMDGPU
+	depends on EXPERT
+	depends on !COMPILE_TEST
+	default n
+	help
+	  Add -Werror to the build flags for amdgpu.ko.
+	  Only enable this if you are warning code for amdgpu.ko.
+
 source "drivers/gpu/drm/amd/acp/Kconfig"
 source "drivers/gpu/drm/amd/display/Kconfig"
 source "drivers/gpu/drm/amd/amdkfd/Kconfig"
diff --git a/drivers/gpu/drm/amd/amdgpu/Makefile b/drivers/gpu/drm/amd/amdgpu/Makefile
index 653726588956..c88760fb52ea 100644
--- a/drivers/gpu/drm/amd/amdgpu/Makefile
+++ b/drivers/gpu/drm/amd/amdgpu/Makefile
@@ -1,5 +1,5 @@
 #
-# Copyright 2017 Advanced Micro Devices, Inc.
+# Copyright 2017-2024 Advanced Micro Devices, Inc. All rights reserved.
 #
 # Permission is hereby granted, free of charge, to any person obtaining a
 # copy of this software and associated documentation files (the "Software"),
@@ -23,7 +23,7 @@
 # Makefile for the drm device driver.  This driver provides support for the
 # Direct Rendering Infrastructure (DRI) in XFree86 4.1.0 and higher.
 
-FULL_AMD_PATH=$(srctree)/$(src)/..
+FULL_AMD_PATH=$(src)/..
 DISPLAY_FOLDER_NAME=display
 FULL_AMD_DISPLAY_PATH = $(FULL_AMD_PATH)/$(DISPLAY_FOLDER_NAME)
 
@@ -34,31 +34,40 @@ ccflags-y := -I$(FULL_AMD_PATH)/include/asic_reg \
 	-I$(FULL_AMD_PATH)/acp/include \
 	-I$(FULL_AMD_DISPLAY_PATH) \
 	-I$(FULL_AMD_DISPLAY_PATH)/include \
+	-I$(FULL_AMD_DISPLAY_PATH)/modules/inc \
 	-I$(FULL_AMD_DISPLAY_PATH)/dc \
 	-I$(FULL_AMD_DISPLAY_PATH)/amdgpu_dm \
-	-I$(FULL_AMD_PATH)/amdkfd
+	-I$(FULL_AMD_PATH)/amdkfd \
+	-I$(FULL_AMD_PATH)/ras/ras_mgr
+
+# Locally disable W=1 warnings enabled in drm subsystem Makefile
+subdir-ccflags-y += -Wno-override-init
+subdir-ccflags-$(CONFIG_DRM_AMDGPU_WERROR) += -Werror
 
 amdgpu-y := amdgpu_drv.o
 
 # add KMS driver
-amdgpu-y += amdgpu_device.o amdgpu_kms.o \
+amdgpu-y += amdgpu_device.o amdgpu_doorbell_mgr.o amdgpu_kms.o \
 	amdgpu_atombios.o atombios_crtc.o amdgpu_connectors.o \
 	atom.o amdgpu_fence.o amdgpu_ttm.o amdgpu_object.o amdgpu_gart.o \
 	amdgpu_encoders.o amdgpu_display.o amdgpu_i2c.o \
-	amdgpu_fb.o amdgpu_gem.o amdgpu_ring.o \
-	amdgpu_cs.o amdgpu_bios.o amdgpu_benchmark.o amdgpu_test.o \
+	amdgpu_gem.o amdgpu_ring.o \
+	amdgpu_cs.o amdgpu_bios.o amdgpu_benchmark.o \
 	atombios_dp.o amdgpu_afmt.o amdgpu_trace_points.o \
 	atombios_encoders.o amdgpu_sa.o atombios_i2c.o \
-	amdgpu_dma_buf.o amdgpu_vm.o amdgpu_ib.o amdgpu_pll.o \
+	amdgpu_dma_buf.o amdgpu_vm.o amdgpu_vm_pt.o amdgpu_vm_tlb_fence.o \
+	amdgpu_ib.o amdgpu_pll.o \
 	amdgpu_ucode.o amdgpu_bo_list.o amdgpu_ctx.o amdgpu_sync.o \
 	amdgpu_gtt_mgr.o amdgpu_preempt_mgr.o amdgpu_vram_mgr.o amdgpu_virt.o \
 	amdgpu_atomfirmware.o amdgpu_vf_error.o amdgpu_sched.o \
-	amdgpu_debugfs.o amdgpu_ids.o amdgpu_gmc.o amdgpu_mmhub.o \
+	amdgpu_debugfs.o amdgpu_ids.o amdgpu_gmc.o amdgpu_mmhub.o amdgpu_hdp.o \
 	amdgpu_xgmi.o amdgpu_csa.o amdgpu_ras.o amdgpu_vm_cpu.o \
 	amdgpu_vm_sdma.o amdgpu_discovery.o amdgpu_ras_eeprom.o amdgpu_nbio.o \
 	amdgpu_umc.o smu_v11_0_i2c.o amdgpu_fru_eeprom.o amdgpu_rap.o \
-	amdgpu_fw_attestation.o amdgpu_securedisplay.o amdgpu_hdp.o \
-	amdgpu_eeprom.o amdgpu_mca.o
+	amdgpu_fw_attestation.o amdgpu_securedisplay.o \
+	amdgpu_eeprom.o amdgpu_mca.o amdgpu_psp_ta.o amdgpu_lsdma.o \
+	amdgpu_ring_mux.o amdgpu_xcp.o amdgpu_seq64.o amdgpu_aca.o amdgpu_dev_coredump.o \
+	amdgpu_cper.o amdgpu_userq_fence.o amdgpu_eviction_fence.o amdgpu_ip.o
 
 amdgpu-$(CONFIG_PROC_FS) += amdgpu_fdinfo.o
 
@@ -69,17 +78,23 @@ amdgpu-$(CONFIG_DRM_AMDGPU_CIK)+= cik.o cik_ih.o \
 	dce_v8_0.o gfx_v7_0.o cik_sdma.o uvd_v4_2.o vce_v2_0.o
 
 amdgpu-$(CONFIG_DRM_AMDGPU_SI)+= si.o gmc_v6_0.o gfx_v6_0.o si_ih.o si_dma.o dce_v6_0.o \
-	uvd_v3_1.o
+	uvd_v3_1.o vce_v1_0.o
 
 amdgpu-y += \
 	vi.o mxgpu_vi.o nbio_v6_1.o soc15.o emu_soc.o mxgpu_ai.o nbio_v7_0.o vega10_reg_init.o \
 	vega20_reg_init.o nbio_v7_4.o nbio_v2_3.o nv.o arct_reg_init.o mxgpu_nv.o \
-	nbio_v7_2.o hdp_v4_0.o hdp_v5_0.o aldebaran_reg_init.o aldebaran.o
+	nbio_v7_2.o hdp_v4_0.o hdp_v5_0.o aldebaran_reg_init.o aldebaran.o soc21.o soc24.o \
+	sienna_cichlid.o smu_v13_0_10.o nbio_v4_3.o hdp_v6_0.o nbio_v7_7.o hdp_v5_2.o lsdma_v6_0.o \
+	nbio_v7_9.o aqua_vanjaram.o nbio_v7_11.o lsdma_v7_0.o hdp_v7_0.o nbif_v6_3_1.o \
+	cyan_skillfish_reg_init.o
 
 # add DF block
 amdgpu-y += \
 	df_v1_7.o \
-	df_v3_6.o
+	df_v3_6.o \
+	df_v4_3.o \
+	df_v4_6_2.o \
+	df_v4_15.o
 
 # add GMC block
 amdgpu-y += \
@@ -87,11 +102,13 @@ amdgpu-y += \
 	gmc_v8_0.o \
 	gfxhub_v1_0.o mmhub_v1_0.o gmc_v9_0.o gfxhub_v1_1.o mmhub_v9_4.o \
 	gfxhub_v2_0.o mmhub_v2_0.o gmc_v10_0.o gfxhub_v2_1.o mmhub_v2_3.o \
-	mmhub_v1_7.o
+	mmhub_v1_7.o gfxhub_v3_0.o mmhub_v3_0.o mmhub_v3_0_2.o gmc_v11_0.o \
+	mmhub_v3_0_1.o gfxhub_v3_0_3.o gfxhub_v1_2.o mmhub_v1_8.o mmhub_v3_3.o \
+	gfxhub_v11_5_0.o mmhub_v4_1_0.o gfxhub_v12_0.o gmc_v12_0.o
 
 # add UMC block
 amdgpu-y += \
-	umc_v6_0.o umc_v6_1.o umc_v6_7.o umc_v8_7.o
+	umc_v6_0.o umc_v6_1.o umc_v6_7.o umc_v8_7.o umc_v8_10.o umc_v12_0.o umc_v8_14.o
 
 # add IH block
 amdgpu-y += \
@@ -102,7 +119,10 @@ amdgpu-y += \
 	cz_ih.o \
 	vega10_ih.o \
 	vega20_ih.o \
-	navi10_ih.o
+	navi10_ih.o \
+	ih_v6_0.o \
+	ih_v6_1.o \
+	ih_v7_0.o
 
 # add PSP block
 amdgpu-y += \
@@ -112,12 +132,13 @@ amdgpu-y += \
 	psp_v11_0.o \
 	psp_v11_0_8.o \
 	psp_v12_0.o \
-	psp_v13_0.o
+	psp_v13_0.o \
+	psp_v13_0_4.o \
+	psp_v14_0.o
 
 # add DCE block
 amdgpu-y += \
 	dce_v10_0.o \
-	dce_v11_0.o \
 	amdgpu_vkms.o
 
 # add GFX block
@@ -128,7 +149,14 @@ amdgpu-y += \
 	gfx_v9_0.o \
 	gfx_v9_4.o \
 	gfx_v9_4_2.o \
-	gfx_v10_0.o
+	gfx_v9_4_3.o \
+	gfx_v10_0.o \
+	imu_v11_0.o \
+	gfx_v11_0.o \
+	gfx_v11_0_3.o \
+	imu_v11_0_3.o \
+	gfx_v12_0.o \
+	imu_v12_0.o
 
 # add async DMA block
 amdgpu-y += \
@@ -137,12 +165,20 @@ amdgpu-y += \
 	sdma_v3_0.o \
 	sdma_v4_0.o \
 	sdma_v4_4.o \
+	sdma_v4_4_2.o \
 	sdma_v5_0.o \
-	sdma_v5_2.o
+	sdma_v5_2.o \
+	sdma_v6_0.o \
+	sdma_v7_0.o
 
 # add MES block
 amdgpu-y += \
-	mes_v10_1.o
+	amdgpu_mes.o \
+	mes_v11_0.o \
+	mes_v12_0.o \
+
+# add GFX userqueue support
+amdgpu-y += mes_userqueue.o
 
 # add UVD block
 amdgpu-y += \
@@ -160,28 +196,55 @@ amdgpu-y += \
 # add VCN and JPEG block
 amdgpu-y += \
 	amdgpu_vcn.o \
+	vcn_sw_ring.o \
 	vcn_v1_0.o \
 	vcn_v2_0.o \
 	vcn_v2_5.o \
 	vcn_v3_0.o \
+	vcn_v4_0.o \
+	vcn_v4_0_3.o \
+	vcn_v4_0_5.o \
+	vcn_v5_0_0.o \
+	vcn_v5_0_1.o \
 	amdgpu_jpeg.o \
 	jpeg_v1_0.o \
 	jpeg_v2_0.o \
 	jpeg_v2_5.o \
-	jpeg_v3_0.o
+	jpeg_v3_0.o \
+	jpeg_v4_0.o \
+	jpeg_v4_0_3.o \
+	jpeg_v4_0_5.o \
+	jpeg_v5_0_0.o \
+	jpeg_v5_0_1.o
+
+# add VPE block
+amdgpu-y += \
+	amdgpu_vpe.o \
+	vpe_v6_1.o
+
+# add UMSCH block
+amdgpu-y += \
+	amdgpu_umsch_mm.o \
+	umsch_mm_v4_0.o
 
+#
 # add ATHUB block
 amdgpu-y += \
 	athub_v1_0.o \
 	athub_v2_0.o \
-	athub_v2_1.o
+	athub_v2_1.o \
+	athub_v3_0.o \
+	athub_v4_1_0.o
 
 # add SMUIO block
 amdgpu-y += \
 	smuio_v9_0.o \
 	smuio_v11_0.o \
 	smuio_v11_0_6.o \
-	smuio_v13_0.o
+	smuio_v13_0.o \
+	smuio_v13_0_3.o \
+	smuio_v13_0_6.o \
+	smuio_v14_0_2.o
 
 # add reset block
 amdgpu-y += \
@@ -194,6 +257,8 @@ amdgpu-y += \
 # add amdkfd interfaces
 amdgpu-y += amdgpu_amdkfd.o
 
+# add gfx usermode queue
+amdgpu-y += amdgpu_userq.o
 
 ifneq ($(CONFIG_HSA_AMD),)
 AMDKFD_PATH := ../amdkfd
@@ -206,8 +271,11 @@ amdgpu-y += \
 	amdgpu_amdkfd_gfx_v9.o \
 	amdgpu_amdkfd_arcturus.o \
 	amdgpu_amdkfd_aldebaran.o \
+	amdgpu_amdkfd_gc_9_4_3.o \
 	amdgpu_amdkfd_gfx_v10.o \
-	amdgpu_amdkfd_gfx_v10_3.o
+	amdgpu_amdkfd_gfx_v10_3.o \
+	amdgpu_amdkfd_gfx_v11.o \
+	amdgpu_amdkfd_gfx_v12.o
 
 ifneq ($(CONFIG_DRM_AMDGPU_CIK),)
 amdgpu-y += amdgpu_amdkfd_gfx_v7.o
@@ -234,7 +302,7 @@ endif
 amdgpu-$(CONFIG_COMPAT) += amdgpu_ioc32.o
 amdgpu-$(CONFIG_VGA_SWITCHEROO) += amdgpu_atpx_handler.o
 amdgpu-$(CONFIG_ACPI) += amdgpu_acpi.o
-amdgpu-$(CONFIG_HMM_MIRROR) += amdgpu_mn.o
+amdgpu-$(CONFIG_HMM_MIRROR) += amdgpu_hmm.o
 
 include $(FULL_AMD_PATH)/pm/Makefile
 
@@ -249,4 +317,17 @@ amdgpu-y += $(AMD_DISPLAY_FILES)
 
 endif
 
+# add isp block
+ifneq ($(CONFIG_DRM_AMD_ISP),)
+amdgpu-y += \
+	amdgpu_isp.o \
+	isp_v4_1_0.o \
+	isp_v4_1_1.o
+endif
+
+AMD_GPU_RAS_PATH := ../ras
+AMD_GPU_RAS_FULL_PATH := $(FULL_AMD_PATH)/ras
+include $(AMD_GPU_RAS_FULL_PATH)/Makefile
+amdgpu-y += $(AMD_GPU_RAS_FILES)
+
 obj-$(CONFIG_DRM_AMDGPU)+= amdgpu.o
diff --git a/drivers/gpu/drm/amd/amdgpu/ObjectID.h b/drivers/gpu/drm/amd/amdgpu/ObjectID.h
index 5b393622f592..a0f0a17e224f 100644
--- a/drivers/gpu/drm/amd/amdgpu/ObjectID.h
+++ b/drivers/gpu/drm/amd/amdgpu/ObjectID.h
@@ -119,6 +119,7 @@
 #define CONNECTOR_OBJECT_ID_eDP                   0x14
 #define CONNECTOR_OBJECT_ID_MXM                   0x15
 #define CONNECTOR_OBJECT_ID_LVDS_eDP              0x16
+#define CONNECTOR_OBJECT_ID_USBC                  0x17
 
 /* deleted */
 
diff --git a/drivers/gpu/drm/amd/amdgpu/aldebaran.c b/drivers/gpu/drm/amd/amdgpu/aldebaran.c
index bcfdb63b1d42..daa7b23bc775 100644
--- a/drivers/gpu/drm/amd/amdgpu/aldebaran.c
+++ b/drivers/gpu/drm/amd/amdgpu/aldebaran.c
@@ -31,65 +31,81 @@
 #include "amdgpu_psp.h"
 #include "amdgpu_xgmi.h"
 
+static bool aldebaran_is_mode2_default(struct amdgpu_reset_control *reset_ctl)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)reset_ctl->handle;
+
+	if ((amdgpu_ip_version(adev, MP1_HWIP, 0) == IP_VERSION(13, 0, 2) &&
+	     adev->gmc.xgmi.connected_to_cpu))
+		return true;
+
+	return false;
+}
+
 static struct amdgpu_reset_handler *
 aldebaran_get_reset_handler(struct amdgpu_reset_control *reset_ctl,
 			    struct amdgpu_reset_context *reset_context)
 {
 	struct amdgpu_reset_handler *handler;
 	struct amdgpu_device *adev = (struct amdgpu_device *)reset_ctl->handle;
+	int i;
+
+	if (reset_context->method == AMD_RESET_METHOD_NONE) {
+		if (aldebaran_is_mode2_default(reset_ctl))
+			reset_context->method = AMD_RESET_METHOD_MODE2;
+		else
+			reset_context->method = amdgpu_asic_reset_method(adev);
+	}
 
 	if (reset_context->method != AMD_RESET_METHOD_NONE) {
 		dev_dbg(adev->dev, "Getting reset handler for method %d\n",
 			reset_context->method);
-		list_for_each_entry(handler, &reset_ctl->reset_handlers,
-				     handler_list) {
+		for_each_handler(i, handler, reset_ctl) {
 			if (handler->reset_method == reset_context->method)
 				return handler;
 		}
 	}
 
-	if (adev->gmc.xgmi.connected_to_cpu) {
-		list_for_each_entry(handler, &reset_ctl->reset_handlers,
-				     handler_list) {
-			if (handler->reset_method == AMD_RESET_METHOD_MODE2) {
-				reset_context->method = AMD_RESET_METHOD_MODE2;
-				return handler;
-			}
-		}
-	}
-
 	dev_dbg(adev->dev, "Reset handler not found!\n");
 
 	return NULL;
 }
 
+static inline uint32_t aldebaran_get_ip_block_mask(struct amdgpu_device *adev)
+{
+	uint32_t ip_block_mask = BIT(AMD_IP_BLOCK_TYPE_GFX) |
+				 BIT(AMD_IP_BLOCK_TYPE_SDMA);
+
+	if (adev->aid_mask)
+		ip_block_mask |= BIT(AMD_IP_BLOCK_TYPE_IH);
+
+	return ip_block_mask;
+}
+
 static int aldebaran_mode2_suspend_ip(struct amdgpu_device *adev)
 {
+	uint32_t ip_block_mask = aldebaran_get_ip_block_mask(adev);
+	uint32_t ip_block;
 	int r, i;
 
+	/* Skip suspend of SDMA IP versions >= 4.4.2. They are multi-aid */
+	if (adev->aid_mask)
+		ip_block_mask &= ~BIT(AMD_IP_BLOCK_TYPE_SDMA);
+
 	amdgpu_device_set_pg_state(adev, AMD_PG_STATE_UNGATE);
 	amdgpu_device_set_cg_state(adev, AMD_CG_STATE_UNGATE);
 
 	for (i = adev->num_ip_blocks - 1; i >= 0; i--) {
-		if (!(adev->ip_blocks[i].version->type ==
-			      AMD_IP_BLOCK_TYPE_GFX ||
-		      adev->ip_blocks[i].version->type ==
-			      AMD_IP_BLOCK_TYPE_SDMA))
+		ip_block = BIT(adev->ip_blocks[i].version->type);
+		if (!(ip_block_mask & ip_block))
 			continue;
 
-		r = adev->ip_blocks[i].version->funcs->suspend(adev);
-
-		if (r) {
-			dev_err(adev->dev,
-				"suspend of IP block <%s> failed %d\n",
-				adev->ip_blocks[i].version->funcs->name, r);
+		r = amdgpu_ip_block_suspend(&adev->ip_blocks[i]);
+		if (r)
 			return r;
-		}
-
-		adev->ip_blocks[i].status.hw = false;
 	}
 
-	return r;
+	return 0;
 }
 
 static int
@@ -113,9 +129,9 @@ static void aldebaran_async_reset(struct work_struct *work)
 	struct amdgpu_reset_control *reset_ctl =
 		container_of(work, struct amdgpu_reset_control, reset_work);
 	struct amdgpu_device *adev = (struct amdgpu_device *)reset_ctl->handle;
+	int i;
 
-	list_for_each_entry(handler, &reset_ctl->reset_handlers,
-			     handler_list) {
+	for_each_handler(i, handler, reset_ctl)	{
 		if (handler->reset_method == reset_ctl->active_reset) {
 			dev_dbg(adev->dev, "Resetting device\n");
 			handler->do_reset(adev);
@@ -136,18 +152,23 @@ static int
 aldebaran_mode2_perform_reset(struct amdgpu_reset_control *reset_ctl,
 			      struct amdgpu_reset_context *reset_context)
 {
-	struct amdgpu_device *tmp_adev = NULL;
 	struct amdgpu_device *adev = (struct amdgpu_device *)reset_ctl->handle;
+	struct list_head *reset_device_list = reset_context->reset_device_list;
+	struct amdgpu_device *tmp_adev = NULL;
 	int r = 0;
 
 	dev_dbg(adev->dev, "aldebaran perform hw reset\n");
-	if (reset_context->hive == NULL) {
+
+	if (reset_device_list == NULL)
+		return -EINVAL;
+
+	if (amdgpu_ip_version(adev, MP1_HWIP, 0) == IP_VERSION(13, 0, 2) &&
+	    reset_context->hive == NULL) {
 		/* Wrong context, return error */
 		return -EINVAL;
 	}
 
-	list_for_each_entry(tmp_adev, &reset_context->hive->device_list,
-			     gmc.xgmi.head) {
+	list_for_each_entry(tmp_adev, reset_device_list, reset_list) {
 		mutex_lock(&tmp_adev->reset_cntl->reset_lock);
 		tmp_adev->reset_cntl->active_reset = AMD_RESET_METHOD_MODE2;
 	}
@@ -155,8 +176,7 @@ aldebaran_mode2_perform_reset(struct amdgpu_reset_control *reset_ctl,
 	 * Mode2 reset doesn't need any sync between nodes in XGMI hive, instead launch
 	 * them together so that they can be completed asynchronously on multiple nodes
 	 */
-	list_for_each_entry(tmp_adev, &reset_context->hive->device_list,
-			     gmc.xgmi.head) {
+	list_for_each_entry(tmp_adev, reset_device_list, reset_list) {
 		/* For XGMI run all resets in parallel to speed up the process */
 		if (tmp_adev->gmc.xgmi.num_physical_nodes > 1) {
 			if (!queue_work(system_unbound_wq,
@@ -174,9 +194,7 @@ aldebaran_mode2_perform_reset(struct amdgpu_reset_control *reset_ctl,
 
 	/* For XGMI wait for all resets to complete before proceed */
 	if (!r) {
-		list_for_each_entry(tmp_adev,
-				     &reset_context->hive->device_list,
-				     gmc.xgmi.head) {
+		list_for_each_entry(tmp_adev, reset_device_list, reset_list) {
 			if (tmp_adev->gmc.xgmi.num_physical_nodes > 1) {
 				flush_work(&tmp_adev->reset_cntl->reset_work);
 				r = tmp_adev->asic_reset_res;
@@ -186,8 +204,7 @@ aldebaran_mode2_perform_reset(struct amdgpu_reset_control *reset_ctl,
 		}
 	}
 
-	list_for_each_entry(tmp_adev, &reset_context->hive->device_list,
-			     gmc.xgmi.head) {
+	list_for_each_entry(tmp_adev, reset_device_list, reset_list) {
 		mutex_unlock(&tmp_adev->reset_cntl->reset_lock);
 		tmp_adev->reset_cntl->active_reset = AMD_RESET_METHOD_NONE;
 	}
@@ -198,8 +215,10 @@ aldebaran_mode2_perform_reset(struct amdgpu_reset_control *reset_ctl,
 static int aldebaran_mode2_restore_ip(struct amdgpu_device *adev)
 {
 	struct amdgpu_firmware_info *ucode_list[AMDGPU_UCODE_ID_MAXIMUM];
+	uint32_t ip_block_mask = aldebaran_get_ip_block_mask(adev);
 	struct amdgpu_firmware_info *ucode;
 	struct amdgpu_ip_block *cmn_block;
+	struct amdgpu_ip_block *ih_block;
 	int ucode_count = 0;
 	int i, r;
 
@@ -237,10 +256,22 @@ static int aldebaran_mode2_restore_ip(struct amdgpu_device *adev)
 		dev_err(adev->dev, "Failed to get BIF handle\n");
 		return -EINVAL;
 	}
-	r = cmn_block->version->funcs->resume(adev);
+	r = amdgpu_ip_block_resume(cmn_block);
 	if (r)
 		return r;
 
+	if (ip_block_mask & BIT(AMD_IP_BLOCK_TYPE_IH)) {
+		ih_block = amdgpu_device_ip_get_ip_block(adev,
+							 AMD_IP_BLOCK_TYPE_IH);
+		if (unlikely(!ih_block)) {
+			dev_err(adev->dev, "Failed to get IH handle\n");
+			return -EINVAL;
+		}
+		r = amdgpu_ip_block_resume(ih_block);
+		if (r)
+			return r;
+	}
+
 	/* Reinit GFXHUB */
 	adev->gfxhub.funcs->init(adev);
 	r = adev->gfxhub.funcs->gart_enable(adev);
@@ -260,7 +291,7 @@ static int aldebaran_mode2_restore_ip(struct amdgpu_device *adev)
 	adev->gfx.rlc.funcs->resume(adev);
 
 	/* Wait for FW reset event complete */
-	r = smu_wait_for_event(adev, SMU_EVENT_RESET_COMPLETE, 0);
+	r = amdgpu_dpm_wait_for_event(adev, SMU_EVENT_RESET_COMPLETE, 0);
 	if (r) {
 		dev_err(adev->dev,
 			"Failed to get response from firmware after reset\n");
@@ -273,15 +304,10 @@ static int aldebaran_mode2_restore_ip(struct amdgpu_device *adev)
 		      adev->ip_blocks[i].version->type ==
 			      AMD_IP_BLOCK_TYPE_SDMA))
 			continue;
-		r = adev->ip_blocks[i].version->funcs->resume(adev);
-		if (r) {
-			dev_err(adev->dev,
-				"resume of IP block <%s> failed %d\n",
-				adev->ip_blocks[i].version->funcs->name, r);
-			return r;
-		}
 
-		adev->ip_blocks[i].status.hw = true;
+		r = amdgpu_ip_block_resume(&adev->ip_blocks[i]);
+		if (r)
+			return r;
 	}
 
 	for (i = 0; i < adev->num_ip_blocks; i++) {
@@ -295,7 +321,7 @@ static int aldebaran_mode2_restore_ip(struct amdgpu_device *adev)
 
 		if (adev->ip_blocks[i].version->funcs->late_init) {
 			r = adev->ip_blocks[i].version->funcs->late_init(
-				(void *)adev);
+				&adev->ip_blocks[i]);
 			if (r) {
 				dev_err(adev->dev,
 					"late_init of IP block <%s> failed %d after reset\n",
@@ -307,8 +333,6 @@ static int aldebaran_mode2_restore_ip(struct amdgpu_device *adev)
 		adev->ip_blocks[i].status.late_initialized = true;
 	}
 
-	amdgpu_ras_set_error_query_ready(adev, true);
-
 	amdgpu_device_set_cg_state(adev, AMD_CG_STATE_GATE);
 	amdgpu_device_set_pg_state(adev, AMD_PG_STATE_GATE);
 
@@ -319,18 +343,28 @@ static int
 aldebaran_mode2_restore_hwcontext(struct amdgpu_reset_control *reset_ctl,
 				  struct amdgpu_reset_context *reset_context)
 {
-	int r;
+	struct list_head *reset_device_list = reset_context->reset_device_list;
 	struct amdgpu_device *tmp_adev = NULL;
+	struct amdgpu_ras *con;
+	int r;
+
+	if (reset_device_list == NULL)
+		return -EINVAL;
 
-	if (reset_context->hive == NULL) {
+	if (amdgpu_ip_version(reset_context->reset_req_dev, MP1_HWIP, 0) ==
+		    IP_VERSION(13, 0, 2) &&
+	    reset_context->hive == NULL) {
 		/* Wrong context, return error */
 		return -EINVAL;
 	}
 
-	list_for_each_entry(tmp_adev, &reset_context->hive->device_list,
-			     gmc.xgmi.head) {
+	list_for_each_entry(tmp_adev, reset_device_list, reset_list) {
+		amdgpu_set_init_level(tmp_adev,
+				AMDGPU_INIT_LEVEL_RESET_RECOVERY);
 		dev_info(tmp_adev->dev,
 			 "GPU reset succeeded, trying to resume\n");
+		/*TBD: Ideally should clear only GFX, SDMA blocks*/
+		amdgpu_ras_clear_err_state(tmp_adev);
 		r = aldebaran_mode2_restore_ip(tmp_adev);
 		if (r)
 			goto end;
@@ -341,7 +375,30 @@ aldebaran_mode2_restore_hwcontext(struct amdgpu_reset_control *reset_ctl,
 		 */
 		amdgpu_register_gpu_instance(tmp_adev);
 
-		/* Resume RAS */
+		/* Resume RAS, ecc_irq */
+		con = amdgpu_ras_get_context(tmp_adev);
+		if (!amdgpu_sriov_vf(tmp_adev) && con) {
+			if (tmp_adev->sdma.ras &&
+				tmp_adev->sdma.ras->ras_block.ras_late_init) {
+				r = tmp_adev->sdma.ras->ras_block.ras_late_init(tmp_adev,
+						&tmp_adev->sdma.ras->ras_block.ras_comm);
+				if (r) {
+					dev_err(tmp_adev->dev, "SDMA failed to execute ras_late_init! ret:%d\n", r);
+					goto end;
+				}
+			}
+
+			if (tmp_adev->gfx.ras &&
+				tmp_adev->gfx.ras->ras_block.ras_late_init) {
+				r = tmp_adev->gfx.ras->ras_block.ras_late_init(tmp_adev,
+						&tmp_adev->gfx.ras->ras_block.ras_comm);
+				if (r) {
+					dev_err(tmp_adev->dev, "GFX failed to execute ras_late_init! ret:%d\n", r);
+					goto end;
+				}
+			}
+		}
+
 		amdgpu_ras_resume(tmp_adev);
 
 		/* Update PSP FW topology after reset */
@@ -351,6 +408,8 @@ aldebaran_mode2_restore_hwcontext(struct amdgpu_reset_control *reset_ctl,
 							tmp_adev);
 
 		if (!r) {
+			amdgpu_set_init_level(tmp_adev,
+					      AMDGPU_INIT_LEVEL_DEFAULT);
 			amdgpu_irq_gpu_reset_resume_helper(tmp_adev);
 
 			r = amdgpu_ib_ring_tests(tmp_adev);
@@ -378,6 +437,12 @@ static struct amdgpu_reset_handler aldebaran_mode2_handler = {
 	.do_reset		= aldebaran_mode2_reset,
 };
 
+static struct amdgpu_reset_handler
+	*aldebaran_rst_handlers[AMDGPU_RESET_MAX_HANDLERS] = {
+		&aldebaran_mode2_handler,
+		&xgmi_reset_on_init_handler,
+	};
+
 int aldebaran_reset_init(struct amdgpu_device *adev)
 {
 	struct amdgpu_reset_control *reset_ctl;
@@ -391,10 +456,9 @@ int aldebaran_reset_init(struct amdgpu_device *adev)
 	reset_ctl->active_reset = AMD_RESET_METHOD_NONE;
 	reset_ctl->get_reset_handler = aldebaran_get_reset_handler;
 
-	INIT_LIST_HEAD(&reset_ctl->reset_handlers);
 	INIT_WORK(&reset_ctl->reset_work, reset_ctl->async_reset);
 	/* Only mode2 is handled through reset control now */
-	amdgpu_reset_add_handler(reset_ctl, &aldebaran_mode2_handler);
+	reset_ctl->reset_handlers = &aldebaran_rst_handlers;
 
 	adev->reset_cntl = reset_ctl;
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index b85b67a88a3d..9f9774f58ce1 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -50,23 +50,20 @@
 #include <linux/hashtable.h>
 #include <linux/dma-fence.h>
 #include <linux/pci.h>
-#include <linux/aer.h>
 
-#include <drm/ttm/ttm_bo_api.h>
-#include <drm/ttm/ttm_bo_driver.h>
+#include <drm/ttm/ttm_bo.h>
 #include <drm/ttm/ttm_placement.h>
-#include <drm/ttm/ttm_execbuf_util.h>
 
 #include <drm/amdgpu_drm.h>
 #include <drm/drm_gem.h>
 #include <drm/drm_ioctl.h>
-#include <drm/gpu_scheduler.h>
 
 #include <kgd_kfd_interface.h>
 #include "dm_pp_interface.h"
 #include "kgd_pp_interface.h"
 
 #include "amd_shared.h"
+#include "amdgpu_utils.h"
 #include "amdgpu_mode.h"
 #include "amdgpu_ih.h"
 #include "amdgpu_irq.h"
@@ -83,15 +80,18 @@
 #include "amdgpu_vce.h"
 #include "amdgpu_vcn.h"
 #include "amdgpu_jpeg.h"
-#include "amdgpu_mn.h"
+#include "amdgpu_vpe.h"
+#include "amdgpu_umsch_mm.h"
 #include "amdgpu_gmc.h"
 #include "amdgpu_gfx.h"
 #include "amdgpu_sdma.h"
+#include "amdgpu_lsdma.h"
 #include "amdgpu_nbio.h"
 #include "amdgpu_hdp.h"
 #include "amdgpu_dm.h"
 #include "amdgpu_virt.h"
 #include "amdgpu_csa.h"
+#include "amdgpu_mes_ctx.h"
 #include "amdgpu_gart.h"
 #include "amdgpu_debugfs.h"
 #include "amdgpu_job.h"
@@ -99,7 +99,6 @@
 #include "amdgpu_gem.h"
 #include "amdgpu_doorbell.h"
 #include "amdgpu_amdkfd.h"
-#include "amdgpu_smu.h"
 #include "amdgpu_discovery.h"
 #include "amdgpu_mes.h"
 #include "amdgpu_umc.h"
@@ -109,26 +108,33 @@
 #include "amdgpu_smuio.h"
 #include "amdgpu_fdinfo.h"
 #include "amdgpu_mca.h"
+#include "amdgpu_aca.h"
+#include "amdgpu_ras.h"
+#include "amdgpu_cper.h"
+#include "amdgpu_xcp.h"
+#include "amdgpu_seq64.h"
+#include "amdgpu_reg_state.h"
+#include "amdgpu_userq.h"
+#include "amdgpu_eviction_fence.h"
+#if defined(CONFIG_DRM_AMD_ISP)
+#include "amdgpu_isp.h"
+#endif
 
-#define MAX_GPU_INSTANCE		16
+#define MAX_GPU_INSTANCE		64
 
-struct amdgpu_gpu_instance
-{
+#define GFX_SLICE_PERIOD_MS		250
+
+struct amdgpu_gpu_instance {
 	struct amdgpu_device		*adev;
 	int				mgpu_fan_enabled;
 };
 
-struct amdgpu_mgpu_info
-{
+struct amdgpu_mgpu_info {
 	struct amdgpu_gpu_instance	gpu_ins[MAX_GPU_INSTANCE];
 	struct mutex			mutex;
 	uint32_t			num_gpu;
 	uint32_t			num_dgpu;
 	uint32_t			num_apu;
-
-	/* delayed reset_func for XGMI configuration if necessary */
-	struct delayed_work		delayed_reset_work;
-	bool				pending_reset;
 };
 
 enum amdgpu_ss {
@@ -138,8 +144,15 @@ enum amdgpu_ss {
 	AMDGPU_SS_DRV_UNLOAD
 };
 
-struct amdgpu_watchdog_timer
-{
+struct amdgpu_hwip_reg_entry {
+	u32		hwip;
+	u32		inst;
+	u32		seg;
+	u32		reg_offset;
+	const char	*reg_name;
+};
+
+struct amdgpu_watchdog_timer {
 	bool timeout_fatal_disable;
 	uint32_t period; /* maxCycles = (1 << period), the number of cycles before a timeout */
 };
@@ -150,13 +163,11 @@ struct amdgpu_watchdog_timer
  * Modules parameters.
  */
 extern int amdgpu_modeset;
-extern int amdgpu_vram_limit;
+extern unsigned int amdgpu_vram_limit;
 extern int amdgpu_vis_vram_limit;
 extern int amdgpu_gart_size;
 extern int amdgpu_gtt_size;
 extern int amdgpu_moverate;
-extern int amdgpu_benchmarking;
-extern int amdgpu_testing;
 extern int amdgpu_audio;
 extern int amdgpu_disp_priority;
 extern int amdgpu_hw_i2c;
@@ -182,14 +193,13 @@ extern int amdgpu_sched_jobs;
 extern int amdgpu_sched_hw_submission;
 extern uint amdgpu_pcie_gen_cap;
 extern uint amdgpu_pcie_lane_cap;
-extern uint amdgpu_cg_mask;
+extern u64 amdgpu_cg_mask;
 extern uint amdgpu_pg_mask;
 extern uint amdgpu_sdma_phase_quantum;
 extern char *amdgpu_disable_cu;
 extern char *amdgpu_virtual_display;
 extern uint amdgpu_pp_feature_mask;
 extern uint amdgpu_force_long_training;
-extern int amdgpu_job_hang_limit;
 extern int amdgpu_lbpw;
 extern int amdgpu_compute_multipipe;
 extern int amdgpu_gpu_recovery;
@@ -199,8 +209,10 @@ extern int amdgpu_smu_pptable_id;
 extern uint amdgpu_dc_feature_mask;
 extern uint amdgpu_freesync_vid_mode;
 extern uint amdgpu_dc_debug_mask;
-extern uint amdgpu_dm_abm_level;
+extern uint amdgpu_dc_visual_confirm;
+extern int amdgpu_dm_abm_level;
 extern int amdgpu_backlight;
+extern int amdgpu_damage_clips;
 extern struct amdgpu_mgpu_info mgpu_info;
 extern int amdgpu_ras_enable;
 extern uint amdgpu_ras_mask;
@@ -211,17 +223,29 @@ extern int amdgpu_async_gfx_ring;
 extern int amdgpu_mcbp;
 extern int amdgpu_discovery;
 extern int amdgpu_mes;
+extern int amdgpu_mes_log_enable;
+extern int amdgpu_mes_kiq;
+extern int amdgpu_uni_mes;
 extern int amdgpu_noretry;
 extern int amdgpu_force_asic_type;
 extern int amdgpu_smartshift_bias;
+extern int amdgpu_use_xgmi_p2p;
+extern int amdgpu_mtype_local;
+extern int amdgpu_enforce_isolation;
 #ifdef CONFIG_HSA_AMD
 extern int sched_policy;
 extern bool debug_evictions;
 extern bool no_system_mem_limit;
+extern int halt_if_hws_hang;
+extern uint amdgpu_svm_default_granularity;
 #else
 static const int __maybe_unused sched_policy = KFD_SCHED_POLICY_HWS;
 static const bool __maybe_unused debug_evictions; /* = false */
 static const bool __maybe_unused no_system_mem_limit;
+static const int __maybe_unused halt_if_hws_hang;
+#endif
+#ifdef CONFIG_HSA_AMD_P2P
+extern bool pcie_p2p;
 #endif
 
 extern int amdgpu_tmz;
@@ -235,9 +259,23 @@ extern int amdgpu_cik_support;
 #endif
 extern int amdgpu_num_kcq;
 
+#define AMDGPU_VCNFW_LOG_SIZE (32 * 1024)
+#define AMDGPU_UMSCHFW_LOG_SIZE (32 * 1024)
+extern int amdgpu_vcnfw_log;
+extern int amdgpu_sg_display;
+extern int amdgpu_umsch_mm;
+extern int amdgpu_seamless;
+extern int amdgpu_umsch_mm_fwlog;
+
+extern int amdgpu_user_partt_mode;
+extern int amdgpu_agp;
+extern int amdgpu_rebar;
+
+extern int amdgpu_wbrf;
+extern int amdgpu_user_queue;
+
 #define AMDGPU_VM_MAX_NUM_CTX			4096
 #define AMDGPU_SG_THRESHOLD			(256*1024*1024)
-#define AMDGPU_DEFAULT_GTT_SIZE_MB		3072ULL /* 3GB by default */
 #define AMDGPU_WAIT_IDLE_TIMEOUT_IN_MS	        3000
 #define AMDGPU_MAX_USEC_TIMEOUT			100000	/* 100 ms */
 #define AMDGPU_FENCE_JIFFIES_TIMEOUT		(HZ / 2)
@@ -267,18 +305,25 @@ extern int amdgpu_num_kcq;
 #define AMDGPU_RESET_VCE			(1 << 13)
 #define AMDGPU_RESET_VCE1			(1 << 14)
 
+/* reset mask */
+#define AMDGPU_RESET_TYPE_FULL (1 << 0) /* full adapter reset, mode1/mode2/BACO/etc. */
+#define AMDGPU_RESET_TYPE_SOFT_RESET (1 << 1) /* IP level soft reset */
+#define AMDGPU_RESET_TYPE_PER_QUEUE (1 << 2) /* per queue */
+#define AMDGPU_RESET_TYPE_PER_PIPE (1 << 3) /* per pipe */
+
 /* max cursor sizes (in pixels) */
 #define CIK_CURSOR_WIDTH 128
 #define CIK_CURSOR_HEIGHT 128
 
-/* smasrt shift bias level limits */
+/* smart shift bias level limits */
 #define AMDGPU_SMARTSHIFT_MAX_BIAS (100)
 #define AMDGPU_SMARTSHIFT_MIN_BIAS (-100)
 
+/* Extra time delay(in ms) to eliminate the influence of temperature momentary fluctuation */
+#define AMDGPU_SWCTF_EXTRA_DELAY		50
+
+struct amdgpu_xcp_mgr;
 struct amdgpu_device;
-struct amdgpu_ib;
-struct amdgpu_cs_parser;
-struct amdgpu_job;
 struct amdgpu_irq_src;
 struct amdgpu_fpriv;
 struct amdgpu_bo_va_mapping;
@@ -313,7 +358,6 @@ enum amdgpu_kiq_irq {
 	AMDGPU_CP_KIQ_IRQ_DRIVER0 = 0,
 	AMDGPU_CP_KIQ_IRQ_LAST
 };
-
 #define MAX_KIQ_REG_WAIT       5000 /* in usecs, 5ms */
 #define MAX_KIQ_REG_BAILOUT_INTERVAL   5 /* in msecs, 5ms */
 #define MAX_KIQ_REG_TRY 1000
@@ -325,13 +369,18 @@ int amdgpu_device_ip_set_powergating_state(void *dev,
 					   enum amd_ip_block_type block_type,
 					   enum amd_powergating_state state);
 void amdgpu_device_ip_get_clockgating_state(struct amdgpu_device *adev,
-					    u32 *flags);
+					    u64 *flags);
 int amdgpu_device_ip_wait_for_idle(struct amdgpu_device *adev,
 				   enum amd_ip_block_type block_type);
-bool amdgpu_device_ip_is_idle(struct amdgpu_device *adev,
+bool amdgpu_device_ip_is_hw(struct amdgpu_device *adev,
+			    enum amd_ip_block_type block_type);
+bool amdgpu_device_ip_is_valid(struct amdgpu_device *adev,
 			      enum amd_ip_block_type block_type);
+int amdgpu_ip_block_suspend(struct amdgpu_ip_block *ip_block);
 
-#define AMDGPU_MAX_IP_NUM 16
+int amdgpu_ip_block_resume(struct amdgpu_ip_block *ip_block);
+
+#define AMDGPU_MAX_IP_NUM AMD_IP_BLOCK_TYPE_NUM
 
 struct amdgpu_ip_block_status {
 	bool valid;
@@ -349,12 +398,10 @@ struct amdgpu_ip_block_version {
 	const struct amd_ip_funcs *funcs;
 };
 
-#define HW_REV(_Major, _Minor, _Rev) \
-	((((uint32_t) (_Major)) << 16) | ((uint32_t) (_Minor) << 8) | ((uint32_t) (_Rev)))
-
 struct amdgpu_ip_block {
 	struct amdgpu_ip_block_status status;
 	const struct amdgpu_ip_block_version *version;
+	struct amdgpu_device *adev;
 };
 
 int amdgpu_device_ip_block_version_cmp(struct amdgpu_device *adev,
@@ -373,7 +420,9 @@ int amdgpu_device_ip_block_add(struct amdgpu_device *adev,
  */
 bool amdgpu_get_bios(struct amdgpu_device *adev);
 bool amdgpu_read_bios(struct amdgpu_device *adev);
-
+bool amdgpu_soc15_read_bios_from_rom(struct amdgpu_device *adev,
+				     u8 *bios, u32 length_bytes);
+void amdgpu_bios_release(struct amdgpu_device *adev);
 /*
  * Clocks
  */
@@ -388,7 +437,6 @@ struct amdgpu_clock {
 	uint32_t default_mclk;
 	uint32_t default_sclk;
 	uint32_t default_dispclk;
-	uint32_t current_dispclk;
 	uint32_t dp_extclk;
 	uint32_t max_pixel_clock;
 };
@@ -417,34 +465,13 @@ struct amdgpu_clock {
  * alignment).
  */
 
-#define AMDGPU_SA_NUM_FENCE_LISTS	32
-
 struct amdgpu_sa_manager {
-	wait_queue_head_t	wq;
-	struct amdgpu_bo	*bo;
-	struct list_head	*hole;
-	struct list_head	flist[AMDGPU_SA_NUM_FENCE_LISTS];
-	struct list_head	olist;
-	unsigned		size;
-	uint64_t		gpu_addr;
-	void			*cpu_ptr;
-	uint32_t		domain;
-	uint32_t		align;
-};
-
-/* sub-allocation buffer */
-struct amdgpu_sa_bo {
-	struct list_head		olist;
-	struct list_head		flist;
-	struct amdgpu_sa_manager	*manager;
-	unsigned			soffset;
-	unsigned			eoffset;
-	struct dma_fence	        *fence;
+	struct drm_suballoc_manager	base;
+	struct amdgpu_bo		*bo;
+	uint64_t			gpu_addr;
+	void				*cpu_ptr;
 };
 
-int amdgpu_fence_slab_init(void);
-void amdgpu_fence_slab_fini(void);
-
 /*
  * IRQS.
  */
@@ -458,28 +485,12 @@ struct amdgpu_flip_work {
 	uint64_t			base;
 	struct drm_pending_vblank_event *event;
 	struct amdgpu_bo		*old_abo;
-	struct dma_fence		*excl;
 	unsigned			shared_count;
 	struct dma_fence		**shared;
 	struct dma_fence_cb		cb;
 	bool				async;
 };
 
-
-/*
- * CP & rings.
- */
-
-struct amdgpu_ib {
-	struct amdgpu_sa_bo		*sa_bo;
-	uint32_t			length_dw;
-	uint64_t			gpu_addr;
-	uint32_t			*ptr;
-	uint32_t			flags;
-};
-
-extern const struct drm_sched_backend_ops amdgpu_sched_ops;
-
 /*
  * file private structure
  */
@@ -488,97 +499,83 @@ struct amdgpu_fpriv {
 	struct amdgpu_vm	vm;
 	struct amdgpu_bo_va	*prt_va;
 	struct amdgpu_bo_va	*csa_va;
+	struct amdgpu_bo_va	*seq64_va;
 	struct mutex		bo_list_lock;
 	struct idr		bo_list_handles;
 	struct amdgpu_ctx_mgr	ctx_mgr;
-};
+	struct amdgpu_userq_mgr	userq_mgr;
 
-int amdgpu_file_to_fpriv(struct file *filp, struct amdgpu_fpriv **fpriv);
+	/* Eviction fence infra */
+	struct amdgpu_eviction_fence_mgr evf_mgr;
 
-int amdgpu_ib_get(struct amdgpu_device *adev, struct amdgpu_vm *vm,
-		  unsigned size,
-		  enum amdgpu_ib_pool_type pool,
-		  struct amdgpu_ib *ib);
-void amdgpu_ib_free(struct amdgpu_device *adev, struct amdgpu_ib *ib,
-		    struct dma_fence *f);
-int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs,
-		       struct amdgpu_ib *ibs, struct amdgpu_job *job,
-		       struct dma_fence **f);
-int amdgpu_ib_pool_init(struct amdgpu_device *adev);
-void amdgpu_ib_pool_fini(struct amdgpu_device *adev);
-int amdgpu_ib_ring_tests(struct amdgpu_device *adev);
-
-/*
- * CS.
- */
-struct amdgpu_cs_chunk {
-	uint32_t		chunk_id;
-	uint32_t		length_dw;
-	void			*kdata;
+	/** GPU partition selection */
+	uint32_t		xcp_id;
 };
 
-struct amdgpu_cs_post_dep {
-	struct drm_syncobj *syncobj;
-	struct dma_fence_chain *chain;
-	u64 point;
-};
-
-struct amdgpu_cs_parser {
-	struct amdgpu_device	*adev;
-	struct drm_file		*filp;
-	struct amdgpu_ctx	*ctx;
-
-	/* chunks */
-	unsigned		nchunks;
-	struct amdgpu_cs_chunk	*chunks;
-
-	/* scheduler job object */
-	struct amdgpu_job	*job;
-	struct drm_sched_entity	*entity;
-
-	/* buffer objects */
-	struct ww_acquire_ctx		ticket;
-	struct amdgpu_bo_list		*bo_list;
-	struct amdgpu_mn		*mn;
-	struct amdgpu_bo_list_entry	vm_pd;
-	struct list_head		validated;
-	struct dma_fence		*fence;
-	uint64_t			bytes_moved_threshold;
-	uint64_t			bytes_moved_vis_threshold;
-	uint64_t			bytes_moved;
-	uint64_t			bytes_moved_vis;
-
-	/* user fence */
-	struct amdgpu_bo_list_entry	uf_entry;
-
-	unsigned			num_post_deps;
-	struct amdgpu_cs_post_dep	*post_deps;
-};
-
-static inline u32 amdgpu_get_ib_value(struct amdgpu_cs_parser *p,
-				      uint32_t ib_idx, int idx)
-{
-	return p->job->ibs[ib_idx].ptr[idx];
-}
-
-static inline void amdgpu_set_ib_value(struct amdgpu_cs_parser *p,
-				       uint32_t ib_idx, int idx,
-				       uint32_t value)
-{
-	p->job->ibs[ib_idx].ptr[idx] = value;
-}
+int amdgpu_file_to_fpriv(struct file *filp, struct amdgpu_fpriv **fpriv);
 
 /*
  * Writeback
  */
-#define AMDGPU_MAX_WB 256	/* Reserve at most 256 WB slots for amdgpu-owned rings. */
+#define AMDGPU_MAX_WB 1024	/* Reserve at most 1024 WB slots for amdgpu-owned rings. */
 
+/**
+ * amdgpu_wb - This struct is used for small GPU memory allocation.
+ *
+ * This struct is used to allocate a small amount of GPU memory that can be
+ * used to shadow certain states into the memory. This is especially useful for
+ * providing easy CPU access to some states without requiring register access
+ * (e.g., if some block is power gated, reading register may be problematic).
+ *
+ * Note: the term writeback was initially used because many of the amdgpu
+ * components had some level of writeback memory, and this struct initially
+ * described those components.
+ */
 struct amdgpu_wb {
+
+	/**
+	 * @wb_obj:
+	 *
+	 * Buffer Object used for the writeback memory.
+	 */
 	struct amdgpu_bo	*wb_obj;
-	volatile uint32_t	*wb;
+
+	/**
+	 * @wb:
+	 *
+	 * Pointer to the first writeback slot. In terms of CPU address
+	 * this value can be accessed directly by using the offset as an index.
+	 * For the GPU address, it is necessary to use gpu_addr and the offset.
+	 */
+	uint32_t		*wb;
+
+	/**
+	 * @gpu_addr:
+	 *
+	 * Writeback base address in the GPU.
+	 */
 	uint64_t		gpu_addr;
-	u32			num_wb;	/* Number of wb slots actually reserved for amdgpu. */
+
+	/**
+	 * @num_wb:
+	 *
+	 * Number of writeback slots reserved for amdgpu.
+	 */
+	u32			num_wb;
+
+	/**
+	 * @used:
+	 *
+	 * Track the writeback slot already used.
+	 */
 	unsigned long		used[DIV_ROUND_UP(AMDGPU_MAX_WB, BITS_PER_LONG)];
+
+	/**
+	 * @lock:
+	 *
+	 * Protects read and write of the used field array.
+	 */
+	spinlock_t		lock;
 };
 
 int amdgpu_device_wb_get(struct amdgpu_device *adev, u32 *wb);
@@ -587,13 +584,7 @@ void amdgpu_device_wb_free(struct amdgpu_device *adev, u32 wb);
 /*
  * Benchmarking
  */
-void amdgpu_benchmark(struct amdgpu_device *adev, int test_number);
-
-
-/*
- * Testing
- */
-void amdgpu_test_moves(struct amdgpu_device *adev);
+int amdgpu_benchmark(struct amdgpu_device *adev, int test_number);
 
 /*
  * ASIC specific register table accessible by UMD
@@ -603,14 +594,42 @@ struct amdgpu_allowed_register_entry {
 	bool grbm_indexed;
 };
 
+/**
+ * enum amd_reset_method - Methods for resetting AMD GPU devices
+ *
+ * @AMD_RESET_METHOD_NONE: The device will not be reset.
+ * @AMD_RESET_LEGACY: Method reserved for SI, CIK and VI ASICs.
+ * @AMD_RESET_MODE0: Reset the entire ASIC. Not currently available for the
+ *                   any device.
+ * @AMD_RESET_MODE1: Resets all IP blocks on the ASIC (SDMA, GFX, VCN, etc.)
+ *                   individually. Suitable only for some discrete GPU, not
+ *                   available for all ASICs.
+ * @AMD_RESET_MODE2: Resets a lesser level of IPs compared to MODE1. Which IPs
+ *                   are reset depends on the ASIC. Notably doesn't reset IPs
+ *                   shared with the CPU on APUs or the memory controllers (so
+ *                   VRAM is not lost). Not available on all ASICs.
+ * @AMD_RESET_LINK: Triggers SW-UP link reset on other GPUs
+ * @AMD_RESET_BACO: BACO (Bus Alive, Chip Off) method powers off and on the card
+ *                  but without powering off the PCI bus. Suitable only for
+ *                  discrete GPUs.
+ * @AMD_RESET_PCI: Does a full bus reset using core Linux subsystem PCI reset
+ *                 and does a secondary bus reset or FLR, depending on what the
+ *                 underlying hardware supports.
+ *
+ * Methods available for AMD GPU driver for resetting the device. Not all
+ * methods are suitable for every device. User can override the method using
+ * module parameter `reset_method`.
+ */
 enum amd_reset_method {
 	AMD_RESET_METHOD_NONE = -1,
 	AMD_RESET_METHOD_LEGACY = 0,
 	AMD_RESET_METHOD_MODE0,
 	AMD_RESET_METHOD_MODE1,
 	AMD_RESET_METHOD_MODE2,
+	AMD_RESET_METHOD_LINK,
 	AMD_RESET_METHOD_BACO,
 	AMD_RESET_METHOD_PCI,
+	AMD_RESET_METHOD_ON_INIT,
 };
 
 struct amdgpu_video_codec_info {
@@ -672,7 +691,7 @@ struct amdgpu_asic_funcs {
 	/* PCIe replay counter */
 	uint64_t (*get_pcie_replay_count)(struct amdgpu_device *adev);
 	/* device supports BACO */
-	bool (*supports_baco)(struct amdgpu_device *adev);
+	int (*supports_baco)(struct amdgpu_device *adev);
 	/* pre asic_init quirks */
 	void (*pre_asic_init)(struct amdgpu_device *adev);
 	/* enter/exit umd stable pstate */
@@ -680,6 +699,12 @@ struct amdgpu_asic_funcs {
 	/* query video codecs */
 	int (*query_video_codecs)(struct amdgpu_device *adev, bool encode,
 				  const struct amdgpu_video_codecs **codecs);
+	/* encode "> 32bits" smn addressing */
+	u64 (*encode_ext_smn_addressing)(int ext_id);
+
+	ssize_t (*get_reg_state)(struct amdgpu_device *adev,
+				 enum amdgpu_reg_state reg_state, void *buf,
+				 size_t max_size);
 };
 
 /*
@@ -696,9 +721,9 @@ int amdgpu_cs_wait_fences_ioctl(struct drm_device *dev, void *data,
 				struct drm_file *filp);
 
 /* VRAM scratch page for HDP bug, default vram page */
-struct amdgpu_vram_scratch {
+struct amdgpu_mem_scratch {
 	struct amdgpu_bo		*robj;
-	volatile uint32_t		*ptr;
+	uint32_t			*ptr;
 	u64				gpu_addr;
 };
 
@@ -714,15 +739,22 @@ void amdgpu_cgs_destroy_device(struct cgs_device *cgs_device);
 typedef uint32_t (*amdgpu_rreg_t)(struct amdgpu_device*, uint32_t);
 typedef void (*amdgpu_wreg_t)(struct amdgpu_device*, uint32_t, uint32_t);
 
+typedef uint32_t (*amdgpu_rreg_ext_t)(struct amdgpu_device*, uint64_t);
+typedef void (*amdgpu_wreg_ext_t)(struct amdgpu_device*, uint64_t, uint32_t);
+
 typedef uint64_t (*amdgpu_rreg64_t)(struct amdgpu_device*, uint32_t);
 typedef void (*amdgpu_wreg64_t)(struct amdgpu_device*, uint32_t, uint64_t);
 
+typedef uint64_t (*amdgpu_rreg64_ext_t)(struct amdgpu_device*, uint64_t);
+typedef void (*amdgpu_wreg64_ext_t)(struct amdgpu_device*, uint64_t, uint64_t);
+
 typedef uint32_t (*amdgpu_block_rreg_t)(struct amdgpu_device*, uint32_t, uint32_t);
 typedef void (*amdgpu_block_wreg_t)(struct amdgpu_device*, uint32_t, uint32_t, uint32_t);
 
 struct amdgpu_mmio_remap {
 	u32 reg_offset;
 	resource_size_t bus_addr;
+	struct amdgpu_bo *bo;
 };
 
 /* Define the HW IP blocks will be used in driver , add more if necessary */
@@ -737,6 +769,7 @@ enum amd_hw_ip_block_type {
 	SDMA5_HWIP,
 	SDMA6_HWIP,
 	SDMA7_HWIP,
+	LSDMA_HWIP,
 	MMHUB_HWIP,
 	ATHUB_HWIP,
 	NBIO_HWIP,
@@ -747,6 +780,7 @@ enum amd_hw_ip_block_type {
 	JPEG_HWIP = VCN_HWIP,
 	VCN1_HWIP,
 	VCE_HWIP,
+	VPE_HWIP,
 	DF_HWIP,
 	DCE_HWIP,
 	OSSSYS_HWIP,
@@ -759,13 +793,48 @@ enum amd_hw_ip_block_type {
 	RSMU_HWIP,
 	XGMI_HWIP,
 	DCI_HWIP,
+	PCIE_HWIP,
+	ISP_HWIP,
 	MAX_HWIP
 };
 
-#define HWIP_MAX_INSTANCE	10
+#define HWIP_MAX_INSTANCE	44
 
 #define HW_ID_MAX		300
-#define IP_VERSION(mj, mn, rv) (((mj) << 16) | ((mn) << 8) | (rv))
+#define IP_VERSION_FULL(mj, mn, rv, var, srev) \
+	(((mj) << 24) | ((mn) << 16) | ((rv) << 8) | ((var) << 4) | (srev))
+#define IP_VERSION(mj, mn, rv)		IP_VERSION_FULL(mj, mn, rv, 0, 0)
+#define IP_VERSION_MAJ(ver)		((ver) >> 24)
+#define IP_VERSION_MIN(ver)		(((ver) >> 16) & 0xFF)
+#define IP_VERSION_REV(ver)		(((ver) >> 8) & 0xFF)
+#define IP_VERSION_VARIANT(ver)		(((ver) >> 4) & 0xF)
+#define IP_VERSION_SUBREV(ver)		((ver) & 0xF)
+#define IP_VERSION_MAJ_MIN_REV(ver)	((ver) >> 8)
+
+struct amdgpu_ip_map_info {
+	/* Map of logical to actual dev instances/mask */
+	uint32_t 		dev_inst[MAX_HWIP][HWIP_MAX_INSTANCE];
+	int8_t (*logical_to_dev_inst)(struct amdgpu_device *adev,
+				      enum amd_hw_ip_block_type block,
+				      int8_t inst);
+	uint32_t (*logical_to_dev_mask)(struct amdgpu_device *adev,
+					enum amd_hw_ip_block_type block,
+					uint32_t mask);
+};
+
+enum amdgpu_uid_type {
+	AMDGPU_UID_TYPE_XCD,
+	AMDGPU_UID_TYPE_AID,
+	AMDGPU_UID_TYPE_SOC,
+	AMDGPU_UID_TYPE_MAX
+};
+
+#define AMDGPU_UID_INST_MAX 8 /* max number of instances for each UID type */
+
+struct amdgpu_uid {
+	uint64_t uid[AMDGPU_UID_TYPE_MAX][AMDGPU_UID_INST_MAX];
+	struct amdgpu_device *adev;
+};
 
 struct amd_powerplay {
 	void *pp_handle;
@@ -811,8 +880,73 @@ struct amd_powerplay {
 					  (rid == 0x01) || \
 					  (rid == 0x10))))
 
+struct amdgpu_mqd_prop {
+	uint64_t mqd_gpu_addr;
+	uint64_t hqd_base_gpu_addr;
+	uint64_t rptr_gpu_addr;
+	uint64_t wptr_gpu_addr;
+	uint32_t queue_size;
+	bool use_doorbell;
+	uint32_t doorbell_index;
+	uint64_t eop_gpu_addr;
+	uint32_t hqd_pipe_priority;
+	uint32_t hqd_queue_priority;
+	bool allow_tunneling;
+	bool hqd_active;
+	uint64_t shadow_addr;
+	uint64_t gds_bkup_addr;
+	uint64_t csa_addr;
+	uint64_t fence_address;
+	bool tmz_queue;
+	bool kernel_queue;
+};
+
+struct amdgpu_mqd {
+	unsigned mqd_size;
+	int (*init_mqd)(struct amdgpu_device *adev, void *mqd,
+			struct amdgpu_mqd_prop *p);
+};
+
+struct amdgpu_pcie_reset_ctx {
+	bool in_link_reset;
+	bool occurs_dpc;
+	bool audio_suspended;
+	struct pci_dev *swus;
+	struct pci_saved_state *swus_pcistate;
+	struct pci_saved_state *swds_pcistate;
+};
+
+/*
+ * Custom Init levels could be defined for different situations where a full
+ * initialization of all hardware blocks are not expected. Sample cases are
+ * custom init sequences after resume after S0i3/S3, reset on initialization,
+ * partial reset of blocks etc. Presently, this defines only two levels. Levels
+ * are described in corresponding struct definitions - amdgpu_init_default,
+ * amdgpu_init_minimal_xgmi.
+ */
+enum amdgpu_init_lvl_id {
+	AMDGPU_INIT_LEVEL_DEFAULT,
+	AMDGPU_INIT_LEVEL_MINIMAL_XGMI,
+	AMDGPU_INIT_LEVEL_RESET_RECOVERY,
+};
+
+struct amdgpu_init_level {
+	enum amdgpu_init_lvl_id level;
+	uint32_t hwini_ip_block_mask;
+};
+
 #define AMDGPU_RESET_MAGIC_NUM 64
 #define AMDGPU_MAX_DF_PERFMONS 4
+struct amdgpu_reset_domain;
+struct amdgpu_fru_info;
+
+enum amdgpu_enforce_isolation_mode {
+	AMDGPU_ENFORCE_ISOLATION_DISABLE = 0,
+	AMDGPU_ENFORCE_ISOLATION_ENABLE = 1,
+	AMDGPU_ENFORCE_ISOLATION_ENABLE_LEGACY = 2,
+	AMDGPU_ENFORCE_ISOLATION_NO_CLEANER_SHADER = 3,
+};
+
 struct amdgpu_device {
 	struct device			*dev;
 	struct pci_dev			*pdev;
@@ -822,6 +956,7 @@ struct amdgpu_device {
 	struct amdgpu_acp		acp;
 #endif
 	struct amdgpu_hive_info *hive;
+	struct amdgpu_xcp_mgr *xcp_mgr;
 	/* ASIC */
 	enum amd_asic_type		asic_type;
 	uint32_t			family;
@@ -835,9 +970,9 @@ struct amdgpu_device {
 	bool				need_swiotlb;
 	bool				accel_working;
 	struct notifier_block		acpi_nb;
+	struct notifier_block		pm_nb;
 	struct amdgpu_i2c_chan		*i2c_bus[AMDGPU_MAX_I2C_BUS];
-	struct debugfs_blob_wrapper     debugfs_vbios_blob;
-	struct debugfs_blob_wrapper     debugfs_discovery_blob;
+	struct debugfs_blob_wrapper debugfs_vbios_blob;
 	struct mutex			srbm_mutex;
 	/* GRBM index mutex. Protects concurrent access to GRBM index */
 	struct mutex                    grbm_idx_mutex;
@@ -869,8 +1004,12 @@ struct amdgpu_device {
 	amdgpu_wreg_t			pcie_wreg;
 	amdgpu_rreg_t			pciep_rreg;
 	amdgpu_wreg_t			pciep_wreg;
+	amdgpu_rreg_ext_t		pcie_rreg_ext;
+	amdgpu_wreg_ext_t		pcie_wreg_ext;
 	amdgpu_rreg64_t			pcie_rreg64;
 	amdgpu_wreg64_t			pcie_wreg64;
+	amdgpu_rreg64_ext_t			pcie_rreg64_ext;
+	amdgpu_wreg64_ext_t			pcie_wreg64_ext;
 	/* protects concurrent UVD register access */
 	spinlock_t uvd_ctx_idx_lock;
 	amdgpu_rreg_t			uvd_ctx_rreg;
@@ -902,11 +1041,11 @@ struct amdgpu_device {
 	dma_addr_t			dummy_page_addr;
 	struct amdgpu_vm_manager	vm_manager;
 	struct amdgpu_vmhub             vmhub[AMDGPU_MAX_VMHUBS];
-	unsigned			num_vmhubs;
+	DECLARE_BITMAP(vmhubs_mask, AMDGPU_MAX_VMHUBS);
 
 	/* memory management */
 	struct amdgpu_mman		mman;
-	struct amdgpu_vram_scratch	vram_scratch;
+	struct amdgpu_mem_scratch	mem_scratch;
 	struct amdgpu_wb		wb;
 	atomic64_t			num_bytes_moved;
 	atomic64_t			num_evictions;
@@ -923,12 +1062,15 @@ struct amdgpu_device {
 		u32			log2_max_MBps;
 	} mm_stats;
 
+	/* discovery*/
+	struct amdgpu_discovery_info discovery;
+
 	/* display */
 	bool				enable_virtual_display;
 	struct amdgpu_vkms_output       *amdgpu_vkms_output;
 	struct amdgpu_mode_info		mode_info;
 	/* For pre-DCE11. DCE11 and later are in "struct amdgpu_device->dm" */
-	struct work_struct		hotplug_work;
+	struct delayed_work         hotplug_work;
 	struct amdgpu_irq_src		crtc_irq;
 	struct amdgpu_irq_src		vline0_irq;
 	struct amdgpu_irq_src		vupdate_irq;
@@ -941,6 +1083,7 @@ struct amdgpu_device {
 	u64				fence_context;
 	unsigned			num_rings;
 	struct amdgpu_ring		*rings[AMDGPU_MAX_RINGS];
+	struct dma_fence __rcu		*gang_submit;
 	bool				ib_pool_ready;
 	struct amdgpu_sa_manager	ib_pools[AMDGPU_IB_POOL_MAX];
 	struct amdgpu_sched		gpu_sched[AMDGPU_HW_IP_NUM][AMDGPU_RING_PRIO_MAX];
@@ -950,14 +1093,8 @@ struct amdgpu_device {
 
 	/* powerplay */
 	struct amd_powerplay		powerplay;
-	bool				pp_force_state_enabled;
-
-	/* smu */
-	struct smu_context		smu;
-
-	/* dpm */
 	struct amdgpu_pm		pm;
-	u32				cg_flags;
+	u64				cg_flags;
 	u32				pg_flags;
 
 	/* nbio */
@@ -981,6 +1118,9 @@ struct amdgpu_device {
 	/* sdma */
 	struct amdgpu_sdma		sdma;
 
+	/* lsdma */
+	struct amdgpu_lsdma		lsdma;
+
 	/* uvd */
 	struct amdgpu_uvd		uvd;
 
@@ -993,6 +1133,13 @@ struct amdgpu_device {
 	/* jpeg */
 	struct amdgpu_jpeg		jpeg;
 
+	/* vpe */
+	struct amdgpu_vpe		vpe;
+
+	/* umsch */
+	struct amdgpu_umsch_mm		umsch_mm;
+	bool				enable_umsch_mm;
+
 	/* firmwares */
 	struct amdgpu_firmware		firmware;
 
@@ -1002,8 +1149,8 @@ struct amdgpu_device {
 	/* GDS */
 	struct amdgpu_gds		gds;
 
-	/* KFD */
-	struct amdgpu_kfd_dev		kfd;
+	/* for userq and VM fences */
+	struct amdgpu_seq64		seq64;
 
 	/* UMC */
 	struct amdgpu_umc		umc;
@@ -1011,9 +1158,30 @@ struct amdgpu_device {
 	/* display related functionality */
 	struct amdgpu_display_manager dm;
 
+#if defined(CONFIG_DRM_AMD_ISP)
+	/* isp */
+	struct amdgpu_isp		isp;
+#endif
+
 	/* mes */
 	bool                            enable_mes;
+	bool                            enable_mes_kiq;
+	bool                            enable_uni_mes;
 	struct amdgpu_mes               mes;
+	struct amdgpu_mqd               mqds[AMDGPU_HW_IP_NUM];
+	const struct amdgpu_userq_funcs *userq_funcs[AMDGPU_HW_IP_NUM];
+
+	/* xarray used to retrieve the user queue fence driver reference
+	 * in the EOP interrupt handler to signal the particular user
+	 * queue fence.
+	 */
+	struct xarray			userq_xa;
+	/**
+	 * @userq_doorbell_xa: Global user queue map (doorbell index → queue)
+	 * Key: doorbell_index (unique global identifier for the queue)
+	 * Value: struct amdgpu_usermode_queue
+	 */
+	struct xarray userq_doorbell_xa;
 
 	/* df */
 	struct amdgpu_df                df;
@@ -1021,6 +1189,12 @@ struct amdgpu_device {
 	/* MCA */
 	struct amdgpu_mca               mca;
 
+	/* ACA */
+	struct amdgpu_aca		aca;
+
+	/* CPER */
+	struct amdgpu_cper		cper;
+
 	struct amdgpu_ip_block          ip_blocks[AMDGPU_MAX_IP_NUM];
 	uint32_t		        harvest_ip_mask;
 	int				num_ip_blocks;
@@ -1034,16 +1208,13 @@ struct amdgpu_device {
 
 	/* soc15 register offset based on ip, instance and  segment */
 	uint32_t		*reg_offset[MAX_HWIP][HWIP_MAX_INSTANCE];
+	struct amdgpu_ip_map_info	ip_map;
 
 	/* delayed work_func for deferring clockgating during resume */
 	struct delayed_work     delayed_init_work;
 
 	struct amdgpu_virt	virt;
 
-	/* link all shadow bo */
-	struct list_head                shadow_list;
-	struct mutex                    shadow_list_lock;
-
 	/* record hw reset is performed */
 	bool has_hw_reset;
 	u8				reset_magic[AMDGPU_RESET_MAGIC_NUM];
@@ -1053,10 +1224,9 @@ struct amdgpu_device {
 	bool				in_s3;
 	bool				in_s4;
 	bool				in_s0ix;
+	suspend_state_t			last_suspend_state;
 
-	atomic_t 			in_gpu_reset;
 	enum pp_mp1_state               mp1_state;
-	struct rw_semaphore reset_sem;
 	struct amdgpu_doorbell_index doorbell_index;
 
 	struct mutex			notifier_lock;
@@ -1069,36 +1239,108 @@ struct amdgpu_device {
 	long				sdma_timeout;
 	long				video_timeout;
 	long				compute_timeout;
+	long				psp_timeout;
 
 	uint64_t			unique_id;
 	uint64_t	df_perfmon_config_assign_mask[AMDGPU_MAX_DF_PERFMONS];
 
 	/* enable runtime pm on the device */
-	bool                            runpm;
 	bool                            in_runpm;
 	bool                            has_pr3;
 
-	bool                            pm_sysfs_en;
 	bool                            ucode_sysfs_en;
 
-	/* Chip product information */
-	char				product_number[16];
-	char				product_name[32];
-	char				serial[20];
-
+	struct amdgpu_fru_info		*fru_info;
 	atomic_t			throttling_logging_enabled;
 	struct ratelimit_state		throttling_logging_rs;
 	uint32_t                        ras_hw_enabled;
 	uint32_t                        ras_enabled;
+	bool                            ras_default_ecc_enabled;
 
 	bool                            no_hw_access;
 	struct pci_saved_state          *pci_state;
 	pci_channel_state_t		pci_channel_state;
 
+	struct amdgpu_pcie_reset_ctx	pcie_reset_ctx;
+
+	/* Track auto wait count on s_barrier settings */
+	bool				barrier_has_auto_waitcnt;
+
 	struct amdgpu_reset_control     *reset_cntl;
-	uint32_t                        ip_versions[HW_ID_MAX][HWIP_MAX_INSTANCE];
+	uint32_t                        ip_versions[MAX_HWIP][HWIP_MAX_INSTANCE];
+
+	bool				ram_is_direct_mapped;
+
+	struct list_head                ras_list;
+
+	struct amdgpu_reset_domain	*reset_domain;
+
+	struct mutex			benchmark_mutex;
+
+	bool                            scpm_enabled;
+	uint32_t                        scpm_status;
+
+	struct work_struct		reset_work;
+
+	bool                            dc_enabled;
+	/* Mask of active clusters */
+	uint32_t			aid_mask;
+
+	/* Debug */
+	bool                            debug_vm;
+	bool                            debug_largebar;
+	bool                            debug_disable_soft_recovery;
+	bool                            debug_use_vram_fw_buf;
+	bool                            debug_enable_ras_aca;
+	bool                            debug_exp_resets;
+	bool                            debug_disable_gpu_ring_reset;
+	bool                            debug_vm_userptr;
+	bool                            debug_disable_ce_logs;
+	bool                            debug_enable_ce_cs;
+
+	/* Protection for the following isolation structure */
+	struct mutex                    enforce_isolation_mutex;
+	enum amdgpu_enforce_isolation_mode	enforce_isolation[MAX_XCP];
+	struct amdgpu_isolation {
+		void			*owner;
+		struct dma_fence	*spearhead;
+		struct amdgpu_sync	active;
+		struct amdgpu_sync	prev;
+	} isolation[MAX_XCP];
+
+	struct amdgpu_init_level *init_lvl;
+
+	/* This flag is used to determine how VRAM allocations are handled for APUs
+	 * in KFD: VRAM or GTT.
+	 */
+	bool                            apu_prefer_gtt;
+
+	bool                            userq_halt_for_enforce_isolation;
+	struct work_struct              userq_reset_work;
+	struct amdgpu_uid *uid_info;
+
+	/* KFD
+	 * Must be last --ends in a flexible-array member.
+	 */
+	struct amdgpu_kfd_dev		kfd;
 };
 
+static inline uint32_t amdgpu_ip_version(const struct amdgpu_device *adev,
+					 uint8_t ip, uint8_t inst)
+{
+	/* This considers only major/minor/rev and ignores
+	 * subrevision/variant fields.
+	 */
+	return adev->ip_versions[ip][inst] & ~0xFFU;
+}
+
+static inline uint32_t amdgpu_ip_version_full(const struct amdgpu_device *adev,
+					      uint8_t ip, uint8_t inst)
+{
+	/* This returns full version - major/minor/rev/variant/subrevision */
+	return adev->ip_versions[ip][inst];
+}
+
 static inline struct amdgpu_device *drm_to_adev(struct drm_device *ddev)
 {
 	return container_of(ddev, struct amdgpu_device, ddev);
@@ -1114,6 +1356,11 @@ static inline struct amdgpu_device *amdgpu_ttm_adev(struct ttm_device *bdev)
 	return container_of(bdev, struct amdgpu_device, mman.bdev);
 }
 
+static inline bool amdgpu_is_multi_aid(struct amdgpu_device *adev)
+{
+	return !!adev->aid_mask;
+}
+
 int amdgpu_device_init(struct amdgpu_device *adev,
 		       uint32_t flags);
 void amdgpu_device_fini_hw(struct amdgpu_device *adev);
@@ -1128,38 +1375,57 @@ size_t amdgpu_device_aper_access(struct amdgpu_device *adev, loff_t pos,
 
 void amdgpu_device_vram_access(struct amdgpu_device *adev, loff_t pos,
 			       void *buf, size_t size, bool write);
+uint32_t amdgpu_device_wait_on_rreg(struct amdgpu_device *adev,
+			    uint32_t inst, uint32_t reg_addr, char reg_name[],
+			    uint32_t expected_value, uint32_t mask);
 uint32_t amdgpu_device_rreg(struct amdgpu_device *adev,
 			    uint32_t reg, uint32_t acc_flags);
+u32 amdgpu_device_indirect_rreg_ext(struct amdgpu_device *adev,
+				    u64 reg_addr);
+uint32_t amdgpu_device_xcc_rreg(struct amdgpu_device *adev,
+				uint32_t reg, uint32_t acc_flags,
+				uint32_t xcc_id);
 void amdgpu_device_wreg(struct amdgpu_device *adev,
 			uint32_t reg, uint32_t v,
 			uint32_t acc_flags);
+void amdgpu_device_indirect_wreg_ext(struct amdgpu_device *adev,
+				     u64 reg_addr, u32 reg_data);
+void amdgpu_device_xcc_wreg(struct amdgpu_device *adev,
+			    uint32_t reg, uint32_t v,
+			    uint32_t acc_flags,
+			    uint32_t xcc_id);
 void amdgpu_mm_wreg_mmio_rlc(struct amdgpu_device *adev,
-			     uint32_t reg, uint32_t v);
+			     uint32_t reg, uint32_t v, uint32_t xcc_id);
 void amdgpu_mm_wreg8(struct amdgpu_device *adev, uint32_t offset, uint8_t value);
 uint8_t amdgpu_mm_rreg8(struct amdgpu_device *adev, uint32_t offset);
 
 u32 amdgpu_device_indirect_rreg(struct amdgpu_device *adev,
-				u32 pcie_index, u32 pcie_data,
 				u32 reg_addr);
 u64 amdgpu_device_indirect_rreg64(struct amdgpu_device *adev,
-				  u32 pcie_index, u32 pcie_data,
 				  u32 reg_addr);
+u64 amdgpu_device_indirect_rreg64_ext(struct amdgpu_device *adev,
+				  u64 reg_addr);
 void amdgpu_device_indirect_wreg(struct amdgpu_device *adev,
-				 u32 pcie_index, u32 pcie_data,
 				 u32 reg_addr, u32 reg_data);
 void amdgpu_device_indirect_wreg64(struct amdgpu_device *adev,
-				   u32 pcie_index, u32 pcie_data,
 				   u32 reg_addr, u64 reg_data);
-
-bool amdgpu_device_asic_has_dc_support(enum amd_asic_type asic_type);
+void amdgpu_device_indirect_wreg64_ext(struct amdgpu_device *adev,
+				   u64 reg_addr, u64 reg_data);
+u32 amdgpu_device_get_rev_id(struct amdgpu_device *adev);
+bool amdgpu_device_asic_has_dc_support(struct pci_dev *pdev,
+				       enum amd_asic_type asic_type);
 bool amdgpu_device_has_dc_support(struct amdgpu_device *adev);
 
+void amdgpu_device_set_sriov_virtual_display(struct amdgpu_device *adev);
+
 int amdgpu_device_pre_asic_reset(struct amdgpu_device *adev,
 				 struct amdgpu_reset_context *reset_context);
 
 int amdgpu_do_asic_reset(struct list_head *device_list_handle,
 			 struct amdgpu_reset_context *reset_context);
 
+int amdgpu_device_reinit_after_reset(struct amdgpu_reset_context *reset_context);
+
 int emu_soc_asic_init(struct amdgpu_device *adev);
 
 /*
@@ -1171,8 +1437,8 @@ int emu_soc_asic_init(struct amdgpu_device *adev);
 #define RREG32_NO_KIQ(reg) amdgpu_device_rreg(adev, (reg), AMDGPU_REGS_NO_KIQ)
 #define WREG32_NO_KIQ(reg, v) amdgpu_device_wreg(adev, (reg), (v), AMDGPU_REGS_NO_KIQ)
 
-#define RREG32_KIQ(reg) amdgpu_kiq_rreg(adev, (reg))
-#define WREG32_KIQ(reg, v) amdgpu_kiq_wreg(adev, (reg), (v))
+#define RREG32_KIQ(reg) amdgpu_kiq_rreg(adev, (reg), 0)
+#define WREG32_KIQ(reg, v) amdgpu_kiq_wreg(adev, (reg), (v), 0)
 
 #define RREG8(reg) amdgpu_mm_rreg8(adev, (reg))
 #define WREG8(reg, v) amdgpu_mm_wreg8(adev, (reg), (v))
@@ -1182,12 +1448,18 @@ int emu_soc_asic_init(struct amdgpu_device *adev);
 #define WREG32(reg, v) amdgpu_device_wreg(adev, (reg), (v), 0)
 #define REG_SET(FIELD, v) (((v) << FIELD##_SHIFT) & FIELD##_MASK)
 #define REG_GET(FIELD, v) (((v) << FIELD##_SHIFT) & FIELD##_MASK)
+#define RREG32_XCC(reg, inst) amdgpu_device_xcc_rreg(adev, (reg), 0, inst)
+#define WREG32_XCC(reg, v, inst) amdgpu_device_xcc_wreg(adev, (reg), (v), 0, inst)
 #define RREG32_PCIE(reg) adev->pcie_rreg(adev, (reg))
 #define WREG32_PCIE(reg, v) adev->pcie_wreg(adev, (reg), (v))
 #define RREG32_PCIE_PORT(reg) adev->pciep_rreg(adev, (reg))
 #define WREG32_PCIE_PORT(reg, v) adev->pciep_wreg(adev, (reg), (v))
+#define RREG32_PCIE_EXT(reg) adev->pcie_rreg_ext(adev, (reg))
+#define WREG32_PCIE_EXT(reg, v) adev->pcie_wreg_ext(adev, (reg), (v))
 #define RREG64_PCIE(reg) adev->pcie_rreg64(adev, (reg))
 #define WREG64_PCIE(reg, v) adev->pcie_wreg64(adev, (reg), (v))
+#define RREG64_PCIE_EXT(reg) adev->pcie_rreg64_ext(adev, (reg))
+#define WREG64_PCIE_EXT(reg, v) adev->pcie_wreg64_ext(adev, (reg), (v))
 #define RREG32_SMC(reg) adev->smc_rreg(adev, (reg))
 #define WREG32_SMC(reg, v) adev->smc_wreg(adev, (reg), (v))
 #define RREG32_UVD_CTX(reg) adev->uvd_ctx_rreg(adev, (reg))
@@ -1243,6 +1515,7 @@ int emu_soc_asic_init(struct amdgpu_device *adev);
 #define WREG32_FIELD_OFFSET(reg, offset, field, val)	\
 	WREG32(mm##reg + offset, (RREG32(mm##reg + offset) & ~REG_FIELD_MASK(reg, field)) | (val) << REG_FIELD_SHIFT(reg, field))
 
+#define AMDGPU_GET_REG_FIELD(x, h, l) (((x) & GENMASK_ULL(h, l)) >> (l))
 /*
  * BIOS helpers.
  */
@@ -1253,7 +1526,8 @@ int emu_soc_asic_init(struct amdgpu_device *adev);
 /*
  * ASICs macro.
  */
-#define amdgpu_asic_set_vga_state(adev, state) (adev)->asic_funcs->set_vga_state((adev), (state))
+#define amdgpu_asic_set_vga_state(adev, state) \
+    ((adev)->asic_funcs->set_vga_state ? (adev)->asic_funcs->set_vga_state((adev), (state)) : 0)
 #define amdgpu_asic_reset(adev) (adev)->asic_funcs->reset((adev))
 #define amdgpu_asic_reset_method(adev) (adev)->asic_funcs->reset_method((adev))
 #define amdgpu_asic_get_xclk(adev) (adev)->asic_funcs->get_xclk((adev))
@@ -1266,10 +1540,6 @@ int emu_soc_asic_init(struct amdgpu_device *adev);
 #define amdgpu_asic_read_bios_from_rom(adev, b, l) (adev)->asic_funcs->read_bios_from_rom((adev), (b), (l))
 #define amdgpu_asic_read_register(adev, se, sh, offset, v)((adev)->asic_funcs->read_register((adev), (se), (sh), (offset), (v)))
 #define amdgpu_asic_get_config_memsize(adev) (adev)->asic_funcs->get_config_memsize((adev))
-#define amdgpu_asic_flush_hdp(adev, r) \
-	((adev)->asic_funcs->flush_hdp ? (adev)->asic_funcs->flush_hdp((adev), (r)) : (adev)->hdp.funcs->flush_hdp((adev), (r)))
-#define amdgpu_asic_invalidate_hdp(adev, r) \
-	((adev)->asic_funcs->invalidate_hdp ? (adev)->asic_funcs->invalidate_hdp((adev), (r)) : (adev)->hdp.funcs->invalidate_hdp((adev), (r)))
 #define amdgpu_asic_need_full_reset(adev) (adev)->asic_funcs->need_full_reset((adev))
 #define amdgpu_asic_init_doorbell_index(adev) (adev)->asic_funcs->init_doorbell_index((adev))
 #define amdgpu_asic_get_pcie_usage(adev, cnt0, cnt1) ((adev)->asic_funcs->get_pcie_usage((adev), (cnt0), (cnt1)))
@@ -1281,18 +1551,24 @@ int emu_soc_asic_init(struct amdgpu_device *adev);
 	((adev)->asic_funcs->update_umd_stable_pstate ? (adev)->asic_funcs->update_umd_stable_pstate((adev), (enter)) : 0)
 #define amdgpu_asic_query_video_codecs(adev, e, c) (adev)->asic_funcs->query_video_codecs((adev), (e), (c))
 
-#define amdgpu_inc_vram_lost(adev) atomic_inc(&((adev)->vram_lost_counter));
+#define amdgpu_inc_vram_lost(adev) atomic_inc(&((adev)->vram_lost_counter))
 
-#define MIN(X, Y) ((X) < (Y) ? (X) : (Y))
+#define BIT_MASK_UPPER(i) ((i) >= BITS_PER_LONG ? 0 : ~0UL << (i))
+#define for_each_inst(i, inst_mask)        \
+	for (i = ffs(inst_mask); i-- != 0; \
+	     i = ffs(inst_mask & BIT_MASK_UPPER(i + 1)))
 
 /* Common functions */
 bool amdgpu_device_has_job_running(struct amdgpu_device *adev);
 bool amdgpu_device_should_recover_gpu(struct amdgpu_device *adev);
 int amdgpu_device_gpu_recover(struct amdgpu_device *adev,
-			      struct amdgpu_job* job);
+			      struct amdgpu_job *job,
+			      struct amdgpu_reset_context *reset_context);
 void amdgpu_device_pci_config_reset(struct amdgpu_device *adev);
 int amdgpu_device_pci_reset(struct amdgpu_device *adev);
 bool amdgpu_device_need_post(struct amdgpu_device *adev);
+bool amdgpu_device_seamless_boot_supported(struct amdgpu_device *adev);
+bool amdgpu_device_should_use_aspm(struct amdgpu_device *adev);
 
 void amdgpu_cs_report_moved_bytes(struct amdgpu_device *adev, u64 num_bytes,
 				  u64 num_vis_bytes);
@@ -1302,44 +1578,53 @@ void amdgpu_device_program_register_sequence(struct amdgpu_device *adev,
 					     const u32 array_size);
 
 int amdgpu_device_mode1_reset(struct amdgpu_device *adev);
-bool amdgpu_device_supports_atpx(struct drm_device *dev);
-bool amdgpu_device_supports_px(struct drm_device *dev);
-bool amdgpu_device_supports_boco(struct drm_device *dev);
-bool amdgpu_device_supports_smart_shift(struct drm_device *dev);
-bool amdgpu_device_supports_baco(struct drm_device *dev);
+int amdgpu_device_link_reset(struct amdgpu_device *adev);
+bool amdgpu_device_supports_atpx(struct amdgpu_device *adev);
+bool amdgpu_device_supports_px(struct amdgpu_device *adev);
+bool amdgpu_device_supports_boco(struct amdgpu_device *adev);
+bool amdgpu_device_supports_smart_shift(struct amdgpu_device *adev);
+int amdgpu_device_supports_baco(struct amdgpu_device *adev);
+void amdgpu_device_detect_runtime_pm_mode(struct amdgpu_device *adev);
 bool amdgpu_device_is_peer_accessible(struct amdgpu_device *adev,
 				      struct amdgpu_device *peer_adev);
-int amdgpu_device_baco_enter(struct drm_device *dev);
-int amdgpu_device_baco_exit(struct drm_device *dev);
+int amdgpu_device_baco_enter(struct amdgpu_device *adev);
+int amdgpu_device_baco_exit(struct amdgpu_device *adev);
 
 void amdgpu_device_flush_hdp(struct amdgpu_device *adev,
 		struct amdgpu_ring *ring);
 void amdgpu_device_invalidate_hdp(struct amdgpu_device *adev,
 		struct amdgpu_ring *ring);
 
+void amdgpu_device_halt(struct amdgpu_device *adev);
+u32 amdgpu_device_pcie_port_rreg(struct amdgpu_device *adev,
+				u32 reg);
+void amdgpu_device_pcie_port_wreg(struct amdgpu_device *adev,
+				u32 reg, u32 v);
+struct dma_fence *amdgpu_device_get_gang(struct amdgpu_device *adev);
+struct dma_fence *amdgpu_device_switch_gang(struct amdgpu_device *adev,
+					    struct dma_fence *gang);
+struct dma_fence *amdgpu_device_enforce_isolation(struct amdgpu_device *adev,
+						  struct amdgpu_ring *ring,
+						  struct amdgpu_job *job);
+bool amdgpu_device_has_display_hardware(struct amdgpu_device *adev);
+ssize_t amdgpu_get_soft_full_reset_mask(struct amdgpu_ring *ring);
+ssize_t amdgpu_show_reset_mask(char *buf, uint32_t supported_reset);
+
 /* atpx handler */
 #if defined(CONFIG_VGA_SWITCHEROO)
 void amdgpu_register_atpx_handler(void);
 void amdgpu_unregister_atpx_handler(void);
 bool amdgpu_has_atpx_dgpu_power_cntl(void);
 bool amdgpu_is_atpx_hybrid(void);
-bool amdgpu_atpx_dgpu_req_power_for_displays(void);
 bool amdgpu_has_atpx(void);
 #else
 static inline void amdgpu_register_atpx_handler(void) {}
 static inline void amdgpu_unregister_atpx_handler(void) {}
 static inline bool amdgpu_has_atpx_dgpu_power_cntl(void) { return false; }
 static inline bool amdgpu_is_atpx_hybrid(void) { return false; }
-static inline bool amdgpu_atpx_dgpu_req_power_for_displays(void) { return false; }
 static inline bool amdgpu_has_atpx(void) { return false; }
 #endif
 
-#if defined(CONFIG_VGA_SWITCHEROO) && defined(CONFIG_ACPI)
-void *amdgpu_atpx_get_dhandle(void);
-#else
-static inline void *amdgpu_atpx_get_dhandle(void) { return NULL; }
-#endif
-
 /*
  * KMS
  */
@@ -1348,20 +1633,18 @@ extern const int amdgpu_max_kms_ioctl;
 
 int amdgpu_driver_load_kms(struct amdgpu_device *adev, unsigned long flags);
 void amdgpu_driver_unload_kms(struct drm_device *dev);
-void amdgpu_driver_lastclose_kms(struct drm_device *dev);
 int amdgpu_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv);
 void amdgpu_driver_postclose_kms(struct drm_device *dev,
 				 struct drm_file *file_priv);
 void amdgpu_driver_release_kms(struct drm_device *dev);
 
-int amdgpu_device_ip_suspend(struct amdgpu_device *adev);
+int amdgpu_device_prepare(struct drm_device *dev);
+void amdgpu_device_complete(struct drm_device *dev);
 int amdgpu_device_suspend(struct drm_device *dev, bool fbcon);
 int amdgpu_device_resume(struct drm_device *dev, bool fbcon);
 u32 amdgpu_get_vblank_counter_kms(struct drm_crtc *crtc);
 int amdgpu_enable_vblank_kms(struct drm_crtc *crtc);
 void amdgpu_disable_vblank_kms(struct drm_crtc *crtc);
-long amdgpu_kms_compat_ioctl(struct file *filp, unsigned int cmd,
-			     unsigned long arg);
 int amdgpu_info_ioctl(struct drm_device *dev, void *data,
 		      struct drm_file *filp);
 
@@ -1386,6 +1669,12 @@ struct amdgpu_afmt_acr amdgpu_afmt_acr(uint32_t clock);
 
 /* amdgpu_acpi.c */
 
+struct amdgpu_numa_info {
+	uint64_t size;
+	int pxm;
+	int nid;
+};
+
 /* ATCS Device/Driver State */
 #define AMDGPU_ATCS_PSC_DEV_STATE_D0		0
 #define AMDGPU_ATCS_PSC_DEV_STATE_D3_HOT	3
@@ -1401,34 +1690,57 @@ int amdgpu_acpi_pcie_performance_request(struct amdgpu_device *adev,
 						u8 perf_req, bool advertise);
 int amdgpu_acpi_power_shift_control(struct amdgpu_device *adev,
 				    u8 dev_state, bool drv_state);
-int amdgpu_acpi_smart_shift_update(struct drm_device *dev, enum amdgpu_ss ss_state);
+int amdgpu_acpi_smart_shift_update(struct amdgpu_device *adev,
+				   enum amdgpu_ss ss_state);
 int amdgpu_acpi_pcie_notify_device_ready(struct amdgpu_device *adev);
+int amdgpu_acpi_get_tmr_info(struct amdgpu_device *adev, u64 *tmr_offset,
+			     u64 *tmr_size);
+int amdgpu_acpi_get_mem_info(struct amdgpu_device *adev, int xcc_id,
+			     struct amdgpu_numa_info *numa_info);
 
 void amdgpu_acpi_get_backlight_caps(struct amdgpu_dm_backlight_caps *caps);
-bool amdgpu_acpi_is_s0ix_active(struct amdgpu_device *adev);
+bool amdgpu_acpi_should_gpu_reset(struct amdgpu_device *adev);
 void amdgpu_acpi_detect(void);
+void amdgpu_acpi_release(void);
 #else
 static inline int amdgpu_acpi_init(struct amdgpu_device *adev) { return 0; }
+static inline int amdgpu_acpi_get_tmr_info(struct amdgpu_device *adev,
+					   u64 *tmr_offset, u64 *tmr_size)
+{
+	return -EINVAL;
+}
+static inline int amdgpu_acpi_get_mem_info(struct amdgpu_device *adev,
+					   int xcc_id,
+					   struct amdgpu_numa_info *numa_info)
+{
+	return -EINVAL;
+}
 static inline void amdgpu_acpi_fini(struct amdgpu_device *adev) { }
-static inline bool amdgpu_acpi_is_s0ix_active(struct amdgpu_device *adev) { return false; }
+static inline bool amdgpu_acpi_should_gpu_reset(struct amdgpu_device *adev) { return false; }
 static inline void amdgpu_acpi_detect(void) { }
+static inline void amdgpu_acpi_release(void) { }
 static inline bool amdgpu_acpi_is_power_shift_control_supported(void) { return false; }
 static inline int amdgpu_acpi_power_shift_control(struct amdgpu_device *adev,
 						  u8 dev_state, bool drv_state) { return 0; }
-static inline int amdgpu_acpi_smart_shift_update(struct drm_device *dev,
-						 enum amdgpu_ss ss_state) { return 0; }
+static inline int amdgpu_acpi_smart_shift_update(struct amdgpu_device *adev,
+						 enum amdgpu_ss ss_state)
+{
+	return 0;
+}
+static inline void amdgpu_acpi_get_backlight_caps(struct amdgpu_dm_backlight_caps *caps) { }
 #endif
 
-int amdgpu_cs_find_mapping(struct amdgpu_cs_parser *parser,
-			   uint64_t addr, struct amdgpu_bo **bo,
-			   struct amdgpu_bo_va_mapping **mapping);
-
-#if defined(CONFIG_DRM_AMD_DC)
-int amdgpu_dm_display_resume(struct amdgpu_device *adev );
+#if defined(CONFIG_ACPI) && defined(CONFIG_SUSPEND)
+bool amdgpu_acpi_is_s3_active(struct amdgpu_device *adev);
+bool amdgpu_acpi_is_s0ix_active(struct amdgpu_device *adev);
 #else
-static inline int amdgpu_dm_display_resume(struct amdgpu_device *adev) { return 0; }
+static inline bool amdgpu_acpi_is_s0ix_active(struct amdgpu_device *adev) { return false; }
+static inline bool amdgpu_acpi_is_s3_active(struct amdgpu_device *adev) { return false; }
 #endif
 
+#if defined(CONFIG_DRM_AMD_ISP)
+int amdgpu_acpi_get_isp4_dev(struct acpi_device **dev);
+#endif
 
 void amdgpu_register_gpu_instance(struct amdgpu_device *adev);
 void amdgpu_unregister_gpu_instance(struct amdgpu_device *adev);
@@ -1449,6 +1761,15 @@ int amdgpu_device_set_cg_state(struct amdgpu_device *adev,
 int amdgpu_device_set_pg_state(struct amdgpu_device *adev,
 			       enum amd_powergating_state state);
 
+static inline bool amdgpu_device_has_timeouts_enabled(struct amdgpu_device *adev)
+{
+	return amdgpu_gpu_recovery != 0 &&
+		adev->gfx_timeout != MAX_SCHEDULE_TIMEOUT &&
+		adev->compute_timeout != MAX_SCHEDULE_TIMEOUT &&
+		adev->sdma_timeout != MAX_SCHEDULE_TIMEOUT &&
+		adev->video_timeout != MAX_SCHEDULE_TIMEOUT;
+}
+
 #include "amdgpu_object.h"
 
 static inline bool amdgpu_is_tmz(struct amdgpu_device *adev)
@@ -1456,8 +1777,32 @@ static inline bool amdgpu_is_tmz(struct amdgpu_device *adev)
        return adev->gmc.tmz_enabled;
 }
 
-static inline int amdgpu_in_reset(struct amdgpu_device *adev)
+int amdgpu_in_reset(struct amdgpu_device *adev);
+
+extern const struct attribute_group amdgpu_vram_mgr_attr_group;
+extern const struct attribute_group amdgpu_gtt_mgr_attr_group;
+extern const struct attribute_group amdgpu_flash_attr_group;
+
+void amdgpu_set_init_level(struct amdgpu_device *adev,
+			   enum amdgpu_init_lvl_id lvl);
+
+static inline int amdgpu_device_bus_status_check(struct amdgpu_device *adev)
 {
-	return atomic_read(&adev->in_gpu_reset);
+       u32 status;
+       int r;
+
+       r = pci_read_config_dword(adev->pdev, PCI_COMMAND, &status);
+       if (r || PCI_POSSIBLE_ERROR(status)) {
+		dev_err(adev->dev, "device lost from bus!");
+		return -ENODEV;
+       }
+
+       return 0;
 }
+
+void amdgpu_device_set_uid(struct amdgpu_uid *uid_info,
+			   enum amdgpu_uid_type type, uint8_t inst,
+			   uint64_t uid);
+uint64_t amdgpu_device_get_uid(struct amdgpu_uid *uid_info,
+			       enum amdgpu_uid_type type, uint8_t inst);
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_aca.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_aca.c
new file mode 100644
index 000000000000..9b3180449150
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_aca.c
@@ -0,0 +1,984 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/list.h>
+#include "amdgpu.h"
+#include "amdgpu_aca.h"
+#include "amdgpu_ras.h"
+
+#define ACA_BANK_HWID(type, hwid, mcatype) [ACA_HWIP_TYPE_##type] = {hwid, mcatype}
+
+typedef int bank_handler_t(struct aca_handle *handle, struct aca_bank *bank, enum aca_smu_type type, void *data);
+
+static struct aca_hwip aca_hwid_mcatypes[ACA_HWIP_TYPE_COUNT] = {
+	ACA_BANK_HWID(SMU,	0x01,	0x01),
+	ACA_BANK_HWID(PCS_XGMI, 0x50,	0x00),
+	ACA_BANK_HWID(UMC,	0x96,	0x00),
+};
+
+static void aca_banks_init(struct aca_banks *banks)
+{
+	if (!banks)
+		return;
+
+	memset(banks, 0, sizeof(*banks));
+	INIT_LIST_HEAD(&banks->list);
+}
+
+static int aca_banks_add_bank(struct aca_banks *banks, struct aca_bank *bank)
+{
+	struct aca_bank_node *node;
+
+	if (!bank)
+		return -EINVAL;
+
+	node = kvzalloc(sizeof(*node), GFP_KERNEL);
+	if (!node)
+		return -ENOMEM;
+
+	memcpy(&node->bank, bank, sizeof(*bank));
+
+	INIT_LIST_HEAD(&node->node);
+	list_add_tail(&node->node, &banks->list);
+
+	banks->nr_banks++;
+
+	return 0;
+}
+
+static void aca_banks_release(struct aca_banks *banks)
+{
+	struct aca_bank_node *node, *tmp;
+
+	if (list_empty(&banks->list))
+		return;
+
+	list_for_each_entry_safe(node, tmp, &banks->list, node) {
+		list_del(&node->node);
+		kvfree(node);
+		banks->nr_banks--;
+	}
+}
+
+static int aca_smu_get_valid_aca_count(struct amdgpu_device *adev, enum aca_smu_type type, u32 *count)
+{
+	struct amdgpu_aca *aca = &adev->aca;
+	const struct aca_smu_funcs *smu_funcs = aca->smu_funcs;
+
+	if (!count)
+		return -EINVAL;
+
+	if (!smu_funcs || !smu_funcs->get_valid_aca_count)
+		return -EOPNOTSUPP;
+
+	return smu_funcs->get_valid_aca_count(adev, type, count);
+}
+
+static struct aca_regs_dump {
+	const char *name;
+	int reg_idx;
+} aca_regs[] = {
+	{"CONTROL",		ACA_REG_IDX_CTL},
+	{"STATUS",		ACA_REG_IDX_STATUS},
+	{"ADDR",		ACA_REG_IDX_ADDR},
+	{"MISC",		ACA_REG_IDX_MISC0},
+	{"CONFIG",		ACA_REG_IDX_CONFIG},
+	{"IPID",		ACA_REG_IDX_IPID},
+	{"SYND",		ACA_REG_IDX_SYND},
+	{"DESTAT",		ACA_REG_IDX_DESTAT},
+	{"DEADDR",		ACA_REG_IDX_DEADDR},
+	{"CONTROL_MASK",	ACA_REG_IDX_CTL_MASK},
+};
+
+static void aca_smu_bank_dump(struct amdgpu_device *adev, int idx, int total, struct aca_bank *bank,
+			      struct ras_query_context *qctx)
+{
+	u64 event_id = qctx ? qctx->evid.event_id : RAS_EVENT_INVALID_ID;
+	int i;
+
+	if (adev->debug_disable_ce_logs &&
+	    bank->smu_err_type == ACA_SMU_TYPE_CE &&
+	    !ACA_BANK_ERR_IS_DEFFERED(bank))
+		return;
+
+	RAS_EVENT_LOG(adev, event_id, HW_ERR "Accelerator Check Architecture events logged\n");
+	/* plus 1 for output format, e.g: ACA[08/08]: xxxx */
+	for (i = 0; i < ARRAY_SIZE(aca_regs); i++)
+		RAS_EVENT_LOG(adev, event_id, HW_ERR "ACA[%02d/%02d].%s=0x%016llx\n",
+			      idx + 1, total, aca_regs[i].name, bank->regs[aca_regs[i].reg_idx]);
+
+	if (ACA_REG__STATUS__SCRUB(bank->regs[ACA_REG_IDX_STATUS]))
+		RAS_EVENT_LOG(adev, event_id, HW_ERR "hardware error logged by the scrubber\n");
+}
+
+static bool aca_bank_hwip_is_matched(struct aca_bank *bank, enum aca_hwip_type type)
+{
+
+	struct aca_hwip *hwip;
+	int hwid, mcatype;
+	u64 ipid;
+
+	if (!bank || type == ACA_HWIP_TYPE_UNKNOW)
+		return false;
+
+	hwip = &aca_hwid_mcatypes[type];
+	if (!hwip->hwid)
+		return false;
+
+	ipid = bank->regs[ACA_REG_IDX_IPID];
+	hwid = ACA_REG__IPID__HARDWAREID(ipid);
+	mcatype = ACA_REG__IPID__MCATYPE(ipid);
+
+	return hwip->hwid == hwid && hwip->mcatype == mcatype;
+}
+
+static int aca_smu_get_valid_aca_banks(struct amdgpu_device *adev, enum aca_smu_type type,
+				       int start, int count,
+				       struct aca_banks *banks, struct ras_query_context *qctx)
+{
+	struct amdgpu_aca *aca = &adev->aca;
+	const struct aca_smu_funcs *smu_funcs = aca->smu_funcs;
+	struct aca_bank bank;
+	int i, max_count, ret;
+
+	if (!count)
+		return 0;
+
+	if (!smu_funcs || !smu_funcs->get_valid_aca_bank)
+		return -EOPNOTSUPP;
+
+	switch (type) {
+	case ACA_SMU_TYPE_UE:
+		max_count = smu_funcs->max_ue_bank_count;
+		break;
+	case ACA_SMU_TYPE_CE:
+		max_count = smu_funcs->max_ce_bank_count;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	if (start + count > max_count)
+		return -EINVAL;
+
+	count = min_t(int, count, max_count);
+	for (i = 0; i < count; i++) {
+		memset(&bank, 0, sizeof(bank));
+		ret = smu_funcs->get_valid_aca_bank(adev, type, start + i, &bank);
+		if (ret)
+			return ret;
+
+		bank.smu_err_type = type;
+
+		/*
+		 * Poison being consumed when injecting a UE while running background workloads,
+		 * which are unexpected.
+		 */
+		if (type == ACA_SMU_TYPE_UE &&
+		    ACA_REG__STATUS__POISON(bank.regs[ACA_REG_IDX_STATUS]) &&
+		    !aca_bank_hwip_is_matched(&bank, ACA_HWIP_TYPE_UMC))
+			continue;
+
+		aca_smu_bank_dump(adev, i, count, &bank, qctx);
+
+		ret = aca_banks_add_bank(banks, &bank);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+static bool aca_bank_is_valid(struct aca_handle *handle, struct aca_bank *bank, enum aca_smu_type type)
+{
+	const struct aca_bank_ops *bank_ops = handle->bank_ops;
+
+	/* Parse all deferred errors with UMC aca handle */
+	if (ACA_BANK_ERR_IS_DEFFERED(bank))
+		return handle->hwip == ACA_HWIP_TYPE_UMC;
+
+	if (!aca_bank_hwip_is_matched(bank, handle->hwip))
+		return false;
+
+	if (!bank_ops->aca_bank_is_valid)
+		return true;
+
+	return bank_ops->aca_bank_is_valid(handle, bank, type, handle->data);
+}
+
+static struct aca_bank_error *new_bank_error(struct aca_error *aerr, struct aca_bank_info *info)
+{
+	struct aca_bank_error *bank_error;
+
+	bank_error = kvzalloc(sizeof(*bank_error), GFP_KERNEL);
+	if (!bank_error)
+		return NULL;
+
+	INIT_LIST_HEAD(&bank_error->node);
+	memcpy(&bank_error->info, info, sizeof(*info));
+
+	mutex_lock(&aerr->lock);
+	list_add_tail(&bank_error->node, &aerr->list);
+	aerr->nr_errors++;
+	mutex_unlock(&aerr->lock);
+
+	return bank_error;
+}
+
+static struct aca_bank_error *find_bank_error(struct aca_error *aerr, struct aca_bank_info *info)
+{
+	struct aca_bank_error *bank_error = NULL;
+	struct aca_bank_info *tmp_info;
+	bool found = false;
+
+	mutex_lock(&aerr->lock);
+	list_for_each_entry(bank_error, &aerr->list, node) {
+		tmp_info = &bank_error->info;
+		if (tmp_info->socket_id == info->socket_id &&
+		    tmp_info->die_id == info->die_id) {
+			found = true;
+			goto out_unlock;
+		}
+	}
+
+out_unlock:
+	mutex_unlock(&aerr->lock);
+
+	return found ? bank_error : NULL;
+}
+
+static void aca_bank_error_remove(struct aca_error *aerr, struct aca_bank_error *bank_error)
+{
+	if (!aerr || !bank_error)
+		return;
+
+	list_del(&bank_error->node);
+	aerr->nr_errors--;
+
+	kvfree(bank_error);
+}
+
+static struct aca_bank_error *get_bank_error(struct aca_error *aerr, struct aca_bank_info *info)
+{
+	struct aca_bank_error *bank_error;
+
+	if (!aerr || !info)
+		return NULL;
+
+	bank_error = find_bank_error(aerr, info);
+	if (bank_error)
+		return bank_error;
+
+	return new_bank_error(aerr, info);
+}
+
+int aca_error_cache_log_bank_error(struct aca_handle *handle, struct aca_bank_info *info,
+				   enum aca_error_type type, u64 count)
+{
+	struct aca_error_cache *error_cache = &handle->error_cache;
+	struct aca_bank_error *bank_error;
+	struct aca_error *aerr;
+
+	if (!handle || !info || type >= ACA_ERROR_TYPE_COUNT)
+		return -EINVAL;
+
+	if (!count)
+		return 0;
+
+	aerr = &error_cache->errors[type];
+	bank_error = get_bank_error(aerr, info);
+	if (!bank_error)
+		return -ENOMEM;
+
+	bank_error->count += count;
+
+	return 0;
+}
+
+static int aca_bank_parser(struct aca_handle *handle, struct aca_bank *bank, enum aca_smu_type type)
+{
+	const struct aca_bank_ops *bank_ops = handle->bank_ops;
+
+	if (!bank)
+		return -EINVAL;
+
+	if (!bank_ops->aca_bank_parser)
+		return -EOPNOTSUPP;
+
+	return bank_ops->aca_bank_parser(handle, bank, type,
+					 handle->data);
+}
+
+static int handler_aca_log_bank_error(struct aca_handle *handle, struct aca_bank *bank,
+				      enum aca_smu_type type, void *data)
+{
+	int ret;
+
+	ret = aca_bank_parser(handle, bank, type);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+static int aca_dispatch_bank(struct aca_handle_manager *mgr, struct aca_bank *bank,
+			     enum aca_smu_type type, bank_handler_t handler, void *data)
+{
+	struct aca_handle *handle;
+	int ret;
+
+	if (list_empty(&mgr->list))
+		return 0;
+
+	list_for_each_entry(handle, &mgr->list, node) {
+		if (!aca_bank_is_valid(handle, bank, type))
+			continue;
+
+		ret = handler(handle, bank, type, data);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+static int aca_dispatch_banks(struct aca_handle_manager *mgr, struct aca_banks *banks,
+			      enum aca_smu_type type, bank_handler_t handler, void *data)
+{
+	struct aca_bank_node *node;
+	struct aca_bank *bank;
+	int ret;
+
+	if (!mgr || !banks)
+		return -EINVAL;
+
+	/* pre check to avoid unnecessary operations */
+	if (list_empty(&mgr->list) || list_empty(&banks->list))
+		return 0;
+
+	list_for_each_entry(node, &banks->list, node) {
+		bank = &node->bank;
+
+		ret = aca_dispatch_bank(mgr, bank, type, handler, data);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+static bool aca_bank_should_update(struct amdgpu_device *adev, enum aca_smu_type type)
+{
+	struct amdgpu_aca *aca = &adev->aca;
+	bool ret = true;
+
+	/*
+	 * Because the UE Valid MCA count will only be cleared after reset,
+	 * in order to avoid repeated counting of the error count,
+	 * the aca bank is only updated once during the gpu recovery stage.
+	 */
+	if (type == ACA_SMU_TYPE_UE) {
+		if (amdgpu_ras_intr_triggered())
+			ret = atomic_cmpxchg(&aca->ue_update_flag, 0, 1) == 0;
+		else
+			atomic_set(&aca->ue_update_flag, 0);
+	}
+
+	return ret;
+}
+
+static void aca_banks_generate_cper(struct amdgpu_device *adev,
+				    enum aca_smu_type type,
+				    struct aca_banks *banks,
+				    int count)
+{
+	struct aca_bank_node *node;
+	struct aca_bank *bank;
+	int r;
+
+	if (!adev->cper.enabled)
+		return;
+
+	if (!banks || !count) {
+		dev_warn(adev->dev, "fail to generate cper records\n");
+		return;
+	}
+
+	/* UEs must be encoded into separate CPER entries */
+	if (type == ACA_SMU_TYPE_UE) {
+		struct aca_banks de_banks;
+
+		aca_banks_init(&de_banks);
+		list_for_each_entry(node, &banks->list, node) {
+			bank = &node->bank;
+			if (bank->aca_err_type == ACA_ERROR_TYPE_DEFERRED) {
+				r = aca_banks_add_bank(&de_banks, bank);
+				if (r)
+					dev_warn(adev->dev, "fail to add de banks, ret = %d\n", r);
+			} else {
+				if (amdgpu_cper_generate_ue_record(adev, bank))
+					dev_warn(adev->dev, "fail to generate ue cper records\n");
+			}
+		}
+
+		if (!list_empty(&de_banks.list)) {
+			if (amdgpu_cper_generate_ce_records(adev, &de_banks, de_banks.nr_banks))
+				dev_warn(adev->dev, "fail to generate de cper records\n");
+		}
+
+		aca_banks_release(&de_banks);
+	} else {
+		/*
+		 * SMU_TYPE_CE banks are combined into 1 CPER entries,
+		 * they could be CEs or DEs or both
+		 */
+		if (amdgpu_cper_generate_ce_records(adev, banks, count))
+			dev_warn(adev->dev, "fail to generate ce cper records\n");
+	}
+}
+
+static int aca_banks_update(struct amdgpu_device *adev, enum aca_smu_type type,
+			    bank_handler_t handler, struct ras_query_context *qctx, void *data)
+{
+	struct amdgpu_aca *aca = &adev->aca;
+	struct aca_banks banks;
+	u32 count = 0;
+	int ret;
+
+	if (list_empty(&aca->mgr.list))
+		return 0;
+
+	if (!aca_bank_should_update(adev, type))
+		return 0;
+
+	ret = aca_smu_get_valid_aca_count(adev, type, &count);
+	if (ret)
+		return ret;
+
+	if (!count)
+		return 0;
+
+	aca_banks_init(&banks);
+
+	ret = aca_smu_get_valid_aca_banks(adev, type, 0, count, &banks, qctx);
+	if (ret)
+		goto err_release_banks;
+
+	if (list_empty(&banks.list)) {
+		ret = 0;
+		goto err_release_banks;
+	}
+
+	ret = aca_dispatch_banks(&aca->mgr, &banks, type,
+				 handler, data);
+	if (ret)
+		goto err_release_banks;
+
+	aca_banks_generate_cper(adev, type, &banks, count);
+
+err_release_banks:
+	aca_banks_release(&banks);
+
+	return ret;
+}
+
+static int aca_log_aca_error_data(struct aca_bank_error *bank_error, enum aca_error_type type, struct ras_err_data *err_data)
+{
+	struct aca_bank_info *info;
+	struct amdgpu_smuio_mcm_config_info mcm_info;
+	u64 count;
+
+	if (type >= ACA_ERROR_TYPE_COUNT)
+		return -EINVAL;
+
+	count = bank_error->count;
+	if (!count)
+		return 0;
+
+	info = &bank_error->info;
+	mcm_info.die_id = info->die_id;
+	mcm_info.socket_id = info->socket_id;
+
+	switch (type) {
+	case ACA_ERROR_TYPE_UE:
+		amdgpu_ras_error_statistic_ue_count(err_data, &mcm_info, count);
+		break;
+	case ACA_ERROR_TYPE_CE:
+		amdgpu_ras_error_statistic_ce_count(err_data, &mcm_info, count);
+		break;
+	case ACA_ERROR_TYPE_DEFERRED:
+		amdgpu_ras_error_statistic_de_count(err_data, &mcm_info, count);
+		break;
+	default:
+		break;
+	}
+
+	return 0;
+}
+
+static int aca_log_aca_error(struct aca_handle *handle, enum aca_error_type type, struct ras_err_data *err_data)
+{
+	struct aca_error_cache *error_cache = &handle->error_cache;
+	struct aca_error *aerr = &error_cache->errors[type];
+	struct aca_bank_error *bank_error, *tmp;
+
+	mutex_lock(&aerr->lock);
+
+	if (list_empty(&aerr->list))
+		goto out_unlock;
+
+	list_for_each_entry_safe(bank_error, tmp, &aerr->list, node) {
+		aca_log_aca_error_data(bank_error, type, err_data);
+		aca_bank_error_remove(aerr, bank_error);
+	}
+
+out_unlock:
+	mutex_unlock(&aerr->lock);
+
+	return 0;
+}
+
+static int __aca_get_error_data(struct amdgpu_device *adev, struct aca_handle *handle, enum aca_error_type type,
+				struct ras_err_data *err_data, struct ras_query_context *qctx)
+{
+	enum aca_smu_type smu_type;
+	int ret;
+
+	switch (type) {
+	case ACA_ERROR_TYPE_UE:
+		smu_type = ACA_SMU_TYPE_UE;
+		break;
+	case ACA_ERROR_TYPE_CE:
+	case ACA_ERROR_TYPE_DEFERRED:
+		smu_type = ACA_SMU_TYPE_CE;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	/* update aca bank to aca source error_cache first */
+	ret = aca_banks_update(adev, smu_type, handler_aca_log_bank_error, qctx, NULL);
+	if (ret)
+		return ret;
+
+	/* DEs may contain in CEs or UEs */
+	if (type != ACA_ERROR_TYPE_DEFERRED)
+		aca_log_aca_error(handle, ACA_ERROR_TYPE_DEFERRED, err_data);
+
+	return aca_log_aca_error(handle, type, err_data);
+}
+
+static bool aca_handle_is_valid(struct aca_handle *handle)
+{
+	if (!handle->mask || !list_empty(&handle->node))
+		return false;
+
+	return true;
+}
+
+int amdgpu_aca_get_error_data(struct amdgpu_device *adev, struct aca_handle *handle,
+			      enum aca_error_type type, struct ras_err_data *err_data,
+			      struct ras_query_context *qctx)
+{
+	if (!handle || !err_data)
+		return -EINVAL;
+
+	if (aca_handle_is_valid(handle))
+		return -EOPNOTSUPP;
+
+	if ((type < 0) || (!(BIT(type) & handle->mask)))
+		return  0;
+
+	return __aca_get_error_data(adev, handle, type, err_data, qctx);
+}
+
+static void aca_error_init(struct aca_error *aerr, enum aca_error_type type)
+{
+	mutex_init(&aerr->lock);
+	INIT_LIST_HEAD(&aerr->list);
+	aerr->type = type;
+	aerr->nr_errors = 0;
+}
+
+static void aca_init_error_cache(struct aca_handle *handle)
+{
+	struct aca_error_cache *error_cache = &handle->error_cache;
+	int type;
+
+	for (type = ACA_ERROR_TYPE_UE; type < ACA_ERROR_TYPE_COUNT; type++)
+		aca_error_init(&error_cache->errors[type], type);
+}
+
+static void aca_error_fini(struct aca_error *aerr)
+{
+	struct aca_bank_error *bank_error, *tmp;
+
+	mutex_lock(&aerr->lock);
+	if (list_empty(&aerr->list))
+		goto out_unlock;
+
+	list_for_each_entry_safe(bank_error, tmp, &aerr->list, node)
+		aca_bank_error_remove(aerr, bank_error);
+
+out_unlock:
+	mutex_destroy(&aerr->lock);
+}
+
+static void aca_fini_error_cache(struct aca_handle *handle)
+{
+	struct aca_error_cache *error_cache = &handle->error_cache;
+	int type;
+
+	for (type = ACA_ERROR_TYPE_UE; type < ACA_ERROR_TYPE_COUNT; type++)
+		aca_error_fini(&error_cache->errors[type]);
+}
+
+static int add_aca_handle(struct amdgpu_device *adev, struct aca_handle_manager *mgr, struct aca_handle *handle,
+			  const char *name, const struct aca_info *ras_info, void *data)
+{
+	memset(handle, 0, sizeof(*handle));
+
+	handle->adev = adev;
+	handle->mgr = mgr;
+	handle->name = name;
+	handle->hwip = ras_info->hwip;
+	handle->mask = ras_info->mask;
+	handle->bank_ops = ras_info->bank_ops;
+	handle->data = data;
+	aca_init_error_cache(handle);
+
+	INIT_LIST_HEAD(&handle->node);
+	list_add_tail(&handle->node, &mgr->list);
+	mgr->nr_handles++;
+
+	return 0;
+}
+
+static ssize_t aca_sysfs_read(struct device *dev,
+			      struct device_attribute *attr, char *buf)
+{
+	struct aca_handle *handle = container_of(attr, struct aca_handle, aca_attr);
+
+	/* NOTE: the aca cache will be auto cleared once read,
+	 * So the driver should unify the query entry point, forward request to ras query interface directly */
+	return amdgpu_ras_aca_sysfs_read(dev, attr, handle, buf, handle->data);
+}
+
+static int add_aca_sysfs(struct amdgpu_device *adev, struct aca_handle *handle)
+{
+	struct device_attribute *aca_attr = &handle->aca_attr;
+
+	snprintf(handle->attr_name, sizeof(handle->attr_name) - 1, "aca_%s", handle->name);
+	aca_attr->show = aca_sysfs_read;
+	aca_attr->attr.name = handle->attr_name;
+	aca_attr->attr.mode = S_IRUGO;
+	sysfs_attr_init(&aca_attr->attr);
+
+	return sysfs_add_file_to_group(&adev->dev->kobj,
+				       &aca_attr->attr,
+				       "ras");
+}
+
+int amdgpu_aca_add_handle(struct amdgpu_device *adev, struct aca_handle *handle,
+			  const char *name, const struct aca_info *ras_info, void *data)
+{
+	struct amdgpu_aca *aca = &adev->aca;
+	int ret;
+
+	if (!amdgpu_aca_is_enabled(adev))
+		return 0;
+
+	ret = add_aca_handle(adev, &aca->mgr, handle, name, ras_info, data);
+	if (ret)
+		return ret;
+
+	return add_aca_sysfs(adev, handle);
+}
+
+static void remove_aca_handle(struct aca_handle *handle)
+{
+	struct aca_handle_manager *mgr = handle->mgr;
+
+	aca_fini_error_cache(handle);
+	list_del(&handle->node);
+	mgr->nr_handles--;
+}
+
+static void remove_aca_sysfs(struct aca_handle *handle)
+{
+	struct amdgpu_device *adev = handle->adev;
+	struct device_attribute *aca_attr = &handle->aca_attr;
+
+	if (adev->dev->kobj.sd)
+		sysfs_remove_file_from_group(&adev->dev->kobj,
+					     &aca_attr->attr,
+					     "ras");
+}
+
+void amdgpu_aca_remove_handle(struct aca_handle *handle)
+{
+	if (!handle || list_empty(&handle->node))
+		return;
+
+	remove_aca_sysfs(handle);
+	remove_aca_handle(handle);
+}
+
+static int aca_manager_init(struct aca_handle_manager *mgr)
+{
+	INIT_LIST_HEAD(&mgr->list);
+	mgr->nr_handles = 0;
+
+	return 0;
+}
+
+static void aca_manager_fini(struct aca_handle_manager *mgr)
+{
+	struct aca_handle *handle, *tmp;
+
+	if (list_empty(&mgr->list))
+		return;
+
+	list_for_each_entry_safe(handle, tmp, &mgr->list, node)
+		amdgpu_aca_remove_handle(handle);
+}
+
+bool amdgpu_aca_is_enabled(struct amdgpu_device *adev)
+{
+	return (adev->aca.is_enabled ||
+		adev->debug_enable_ras_aca);
+}
+
+int amdgpu_aca_init(struct amdgpu_device *adev)
+{
+	struct amdgpu_aca *aca = &adev->aca;
+	int ret;
+
+	atomic_set(&aca->ue_update_flag, 0);
+
+	ret = aca_manager_init(&aca->mgr);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+void amdgpu_aca_fini(struct amdgpu_device *adev)
+{
+	struct amdgpu_aca *aca = &adev->aca;
+
+	aca_manager_fini(&aca->mgr);
+
+	atomic_set(&aca->ue_update_flag, 0);
+}
+
+int amdgpu_aca_reset(struct amdgpu_device *adev)
+{
+	struct amdgpu_aca *aca = &adev->aca;
+
+	atomic_set(&aca->ue_update_flag, 0);
+
+	return 0;
+}
+
+void amdgpu_aca_set_smu_funcs(struct amdgpu_device *adev, const struct aca_smu_funcs *smu_funcs)
+{
+	struct amdgpu_aca *aca = &adev->aca;
+
+	WARN_ON(aca->smu_funcs);
+	aca->smu_funcs = smu_funcs;
+}
+
+int aca_bank_info_decode(struct aca_bank *bank, struct aca_bank_info *info)
+{
+	u64 ipid;
+	u32 instidhi, instidlo;
+
+	if (!bank || !info)
+		return -EINVAL;
+
+	ipid = bank->regs[ACA_REG_IDX_IPID];
+	info->hwid = ACA_REG__IPID__HARDWAREID(ipid);
+	info->mcatype = ACA_REG__IPID__MCATYPE(ipid);
+	/*
+	 * Unfied DieID Format: SAASS. A:AID, S:Socket.
+	 * Unfied DieID[4:4] = InstanceId[0:0]
+	 * Unfied DieID[0:3] = InstanceIdHi[0:3]
+	 */
+	instidhi = ACA_REG__IPID__INSTANCEIDHI(ipid);
+	instidlo = ACA_REG__IPID__INSTANCEIDLO(ipid);
+	info->die_id = ((instidhi >> 2) & 0x03);
+	info->socket_id = ((instidlo & 0x1) << 2) | (instidhi & 0x03);
+
+	return 0;
+}
+
+static int aca_bank_get_error_code(struct amdgpu_device *adev, struct aca_bank *bank)
+{
+	struct amdgpu_aca *aca = &adev->aca;
+	const struct aca_smu_funcs *smu_funcs = aca->smu_funcs;
+
+	if (!smu_funcs || !smu_funcs->parse_error_code)
+		return -EOPNOTSUPP;
+
+	return smu_funcs->parse_error_code(adev, bank);
+}
+
+int aca_bank_check_error_codes(struct amdgpu_device *adev, struct aca_bank *bank, int *err_codes, int size)
+{
+	int i, error_code;
+
+	if (!bank || !err_codes)
+		return -EINVAL;
+
+	error_code = aca_bank_get_error_code(adev, bank);
+	if (error_code < 0)
+		return error_code;
+
+	for (i = 0; i < size; i++) {
+		if (err_codes[i] == error_code)
+			return 0;
+	}
+
+	return -EINVAL;
+}
+
+int amdgpu_aca_smu_set_debug_mode(struct amdgpu_device *adev, bool en)
+{
+	struct amdgpu_aca *aca = &adev->aca;
+	const struct aca_smu_funcs *smu_funcs = aca->smu_funcs;
+
+	if (!smu_funcs || !smu_funcs->set_debug_mode)
+		return -EOPNOTSUPP;
+
+	return smu_funcs->set_debug_mode(adev, en);
+}
+
+#if defined(CONFIG_DEBUG_FS)
+static int amdgpu_aca_smu_debug_mode_set(void *data, u64 val)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)data;
+	int ret;
+
+	ret = amdgpu_ras_set_aca_debug_mode(adev, val ? true : false);
+	if (ret)
+		return ret;
+
+	dev_info(adev->dev, "amdgpu set smu aca debug mode %s success\n", val ? "on" : "off");
+
+	return 0;
+}
+
+static void aca_dump_entry(struct seq_file *m, struct aca_bank *bank, enum aca_smu_type type, int idx)
+{
+	struct aca_bank_info info;
+	int i, ret;
+
+	ret = aca_bank_info_decode(bank, &info);
+	if (ret)
+		return;
+
+	seq_printf(m, "aca entry[%d].type: %s\n", idx, type ==  ACA_SMU_TYPE_UE ? "UE" : "CE");
+	seq_printf(m, "aca entry[%d].info: socketid:%d aid:%d hwid:0x%03x mcatype:0x%04x\n",
+		   idx, info.socket_id, info.die_id, info.hwid, info.mcatype);
+
+	for (i = 0; i < ARRAY_SIZE(aca_regs); i++)
+		seq_printf(m, "aca entry[%d].regs[%d]: 0x%016llx\n", idx, aca_regs[i].reg_idx, bank->regs[aca_regs[i].reg_idx]);
+}
+
+struct aca_dump_context {
+	struct seq_file *m;
+	int idx;
+};
+
+static int handler_aca_bank_dump(struct aca_handle *handle, struct aca_bank *bank,
+				 enum aca_smu_type type, void *data)
+{
+	struct aca_dump_context *ctx = (struct aca_dump_context *)data;
+
+	aca_dump_entry(ctx->m, bank, type, ctx->idx++);
+
+	return handler_aca_log_bank_error(handle, bank, type, NULL);
+}
+
+static int aca_dump_show(struct seq_file *m, enum aca_smu_type type)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)m->private;
+	struct aca_dump_context context = {
+		.m = m,
+		.idx = 0,
+	};
+
+	return aca_banks_update(adev, type, handler_aca_bank_dump, NULL, (void *)&context);
+}
+
+static int aca_dump_ce_show(struct seq_file *m, void *unused)
+{
+	return aca_dump_show(m, ACA_SMU_TYPE_CE);
+}
+
+static int aca_dump_ce_open(struct inode *inode, struct file *file)
+{
+	return single_open(file, aca_dump_ce_show, inode->i_private);
+}
+
+static const struct file_operations aca_ce_dump_debug_fops = {
+	.owner = THIS_MODULE,
+	.open = aca_dump_ce_open,
+	.read = seq_read,
+	.llseek = seq_lseek,
+	.release = single_release,
+};
+
+static int aca_dump_ue_show(struct seq_file *m, void *unused)
+{
+	return aca_dump_show(m, ACA_SMU_TYPE_UE);
+}
+
+static int aca_dump_ue_open(struct inode *inode, struct file *file)
+{
+	return single_open(file, aca_dump_ue_show, inode->i_private);
+}
+
+static const struct file_operations aca_ue_dump_debug_fops = {
+	.owner = THIS_MODULE,
+	.open = aca_dump_ue_open,
+	.read = seq_read,
+	.llseek = seq_lseek,
+	.release = single_release,
+};
+
+DEFINE_DEBUGFS_ATTRIBUTE(aca_debug_mode_fops, NULL, amdgpu_aca_smu_debug_mode_set, "%llu\n");
+#endif
+
+void amdgpu_aca_smu_debugfs_init(struct amdgpu_device *adev, struct dentry *root)
+{
+#if defined(CONFIG_DEBUG_FS)
+	if (!root)
+		return;
+
+	debugfs_create_file("aca_debug_mode", 0200, root, adev, &aca_debug_mode_fops);
+	debugfs_create_file("aca_ue_dump", 0400, root, adev, &aca_ue_dump_debug_fops);
+	debugfs_create_file("aca_ce_dump", 0400, root, adev, &aca_ce_dump_debug_fops);
+#endif
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_aca.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_aca.h
new file mode 100644
index 000000000000..38c88897e1ec
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_aca.h
@@ -0,0 +1,232 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __AMDGPU_ACA_H__
+#define __AMDGPU_ACA_H__
+
+#include <linux/list.h>
+
+struct ras_err_data;
+struct ras_query_context;
+
+#define ACA_MAX_REGS_COUNT	(16)
+
+#define ACA_REG_FIELD(x, h, l)			(((x) & GENMASK_ULL(h, l)) >> l)
+#define ACA_REG__STATUS__VAL(x)			ACA_REG_FIELD(x, 63, 63)
+#define ACA_REG__STATUS__OVERFLOW(x)		ACA_REG_FIELD(x, 62, 62)
+#define ACA_REG__STATUS__UC(x)			ACA_REG_FIELD(x, 61, 61)
+#define ACA_REG__STATUS__EN(x)			ACA_REG_FIELD(x, 60, 60)
+#define ACA_REG__STATUS__MISCV(x)		ACA_REG_FIELD(x, 59, 59)
+#define ACA_REG__STATUS__ADDRV(x)		ACA_REG_FIELD(x, 58, 58)
+#define ACA_REG__STATUS__PCC(x)			ACA_REG_FIELD(x, 57, 57)
+#define ACA_REG__STATUS__ERRCOREIDVAL(x)	ACA_REG_FIELD(x, 56, 56)
+#define ACA_REG__STATUS__TCC(x)			ACA_REG_FIELD(x, 55, 55)
+#define ACA_REG__STATUS__SYNDV(x)		ACA_REG_FIELD(x, 53, 53)
+#define ACA_REG__STATUS__CECC(x)		ACA_REG_FIELD(x, 46, 46)
+#define ACA_REG__STATUS__UECC(x)		ACA_REG_FIELD(x, 45, 45)
+#define ACA_REG__STATUS__DEFERRED(x)		ACA_REG_FIELD(x, 44, 44)
+#define ACA_REG__STATUS__POISON(x)		ACA_REG_FIELD(x, 43, 43)
+#define ACA_REG__STATUS__SCRUB(x)		ACA_REG_FIELD(x, 40, 40)
+#define ACA_REG__STATUS__ERRCOREID(x)		ACA_REG_FIELD(x, 37, 32)
+#define ACA_REG__STATUS__ADDRLSB(x)		ACA_REG_FIELD(x, 29, 24)
+#define ACA_REG__STATUS__ERRORCODEEXT(x)	ACA_REG_FIELD(x, 21, 16)
+#define ACA_REG__STATUS__ERRORCODE(x)		ACA_REG_FIELD(x, 15, 0)
+
+#define ACA_REG__IPID__MCATYPE(x)		ACA_REG_FIELD(x, 63, 48)
+#define ACA_REG__IPID__INSTANCEIDHI(x)		ACA_REG_FIELD(x, 47, 44)
+#define ACA_REG__IPID__HARDWAREID(x)		ACA_REG_FIELD(x, 43, 32)
+#define ACA_REG__IPID__INSTANCEIDLO(x)		ACA_REG_FIELD(x, 31, 0)
+
+#define ACA_REG__MISC0__VALID(x)		ACA_REG_FIELD(x, 63, 63)
+#define ACA_REG__MISC0__OVRFLW(x)		ACA_REG_FIELD(x, 48, 48)
+#define ACA_REG__MISC0__ERRCNT(x)		ACA_REG_FIELD(x, 43, 32)
+
+#define ACA_REG__SYND__ERRORINFORMATION(x)	ACA_REG_FIELD(x, 17, 0)
+
+/* NOTE: The following codes refers to the smu header file */
+#define ACA_EXTERROR_CODE_CE			0x3a
+#define ACA_EXTERROR_CODE_FAULT			0x3b
+
+#define ACA_ERROR_UE_MASK		BIT_MASK(ACA_ERROR_TYPE_UE)
+#define ACA_ERROR_CE_MASK		BIT_MASK(ACA_ERROR_TYPE_CE)
+#define ACA_ERROR_DEFERRED_MASK		BIT_MASK(ACA_ERROR_TYPE_DEFERRED)
+
+#define mmSMNAID_AID0_MCA_SMU		0x03b30400	/* SMN AID AID0 */
+#define mmSMNAID_XCD0_MCA_SMU		0x36430400	/* SMN AID XCD0 */
+#define mmSMNAID_XCD1_MCA_SMU		0x38430400	/* SMN AID XCD1 */
+#define mmSMNXCD_XCD0_MCA_SMU		0x40430400	/* SMN XCD XCD0 */
+
+#define ACA_BANK_ERR_IS_DEFFERED(bank)                                \
+	(ACA_REG__STATUS__POISON((bank)->regs[ACA_REG_IDX_STATUS]) || \
+	 ACA_REG__STATUS__DEFERRED((bank)->regs[ACA_REG_IDX_STATUS]))
+
+enum aca_reg_idx {
+	ACA_REG_IDX_CTL			= 0,
+	ACA_REG_IDX_STATUS		= 1,
+	ACA_REG_IDX_ADDR		= 2,
+	ACA_REG_IDX_MISC0		= 3,
+	ACA_REG_IDX_CONFIG		= 4,
+	ACA_REG_IDX_IPID		= 5,
+	ACA_REG_IDX_SYND		= 6,
+	ACA_REG_IDX_DESTAT		= 8,
+	ACA_REG_IDX_DEADDR		= 9,
+	ACA_REG_IDX_CTL_MASK		= 10,
+	ACA_REG_IDX_COUNT		= 16,
+};
+
+enum aca_hwip_type {
+	ACA_HWIP_TYPE_UNKNOW = -1,
+	ACA_HWIP_TYPE_PSP = 0,
+	ACA_HWIP_TYPE_UMC,
+	ACA_HWIP_TYPE_SMU,
+	ACA_HWIP_TYPE_PCS_XGMI,
+	ACA_HWIP_TYPE_COUNT,
+};
+
+enum aca_error_type {
+	ACA_ERROR_TYPE_INVALID = -1,
+	ACA_ERROR_TYPE_UE = 0,
+	ACA_ERROR_TYPE_CE,
+	ACA_ERROR_TYPE_DEFERRED,
+	ACA_ERROR_TYPE_COUNT
+};
+
+enum aca_smu_type {
+	ACA_SMU_TYPE_INVALID = -1,
+	ACA_SMU_TYPE_UE = 0,
+	ACA_SMU_TYPE_CE,
+	ACA_SMU_TYPE_COUNT,
+};
+
+struct aca_hwip {
+	int hwid;
+	int mcatype;
+};
+
+struct aca_bank {
+	enum aca_error_type aca_err_type;
+	enum aca_smu_type smu_err_type;
+	u64 regs[ACA_MAX_REGS_COUNT];
+};
+
+struct aca_bank_node {
+	struct aca_bank bank;
+	struct list_head node;
+};
+
+struct aca_banks {
+	int nr_banks;
+	struct list_head list;
+};
+
+struct aca_bank_info {
+	int die_id;
+	int socket_id;
+	int hwid;
+	int mcatype;
+};
+
+struct aca_bank_error {
+	struct list_head node;
+	struct aca_bank_info info;
+	u64 count;
+};
+
+struct aca_error {
+	struct list_head list;
+	struct mutex lock;
+	enum aca_error_type type;
+	int nr_errors;
+};
+
+struct aca_handle_manager {
+	struct list_head list;
+	int nr_handles;
+};
+
+struct aca_error_cache {
+	struct aca_error errors[ACA_ERROR_TYPE_COUNT];
+};
+
+struct aca_handle {
+	struct list_head node;
+	enum aca_hwip_type hwip;
+	struct amdgpu_device *adev;
+	struct aca_handle_manager *mgr;
+	struct aca_error_cache error_cache;
+	const struct aca_bank_ops *bank_ops;
+	struct device_attribute aca_attr;
+	char attr_name[64];
+	const char *name;
+	u32 mask;
+	void *data;
+};
+
+struct aca_bank_ops {
+	int (*aca_bank_parser)(struct aca_handle *handle, struct aca_bank *bank, enum aca_smu_type type, void *data);
+	bool (*aca_bank_is_valid)(struct aca_handle *handle, struct aca_bank *bank, enum aca_smu_type type,
+				  void *data);
+};
+
+struct aca_smu_funcs {
+	int max_ue_bank_count;
+	int max_ce_bank_count;
+	int (*set_debug_mode)(struct amdgpu_device *adev, bool enable);
+	int (*get_valid_aca_count)(struct amdgpu_device *adev, enum aca_smu_type type, u32 *count);
+	int (*get_valid_aca_bank)(struct amdgpu_device *adev, enum aca_smu_type type, int idx, struct aca_bank *bank);
+	int (*parse_error_code)(struct amdgpu_device *adev, struct aca_bank *bank);
+};
+
+struct amdgpu_aca {
+	struct aca_handle_manager mgr;
+	const struct aca_smu_funcs *smu_funcs;
+	atomic_t ue_update_flag;
+	bool is_enabled;
+};
+
+struct aca_info {
+	enum aca_hwip_type hwip;
+	const struct aca_bank_ops *bank_ops;
+	u32 mask;
+};
+
+int amdgpu_aca_init(struct amdgpu_device *adev);
+void amdgpu_aca_fini(struct amdgpu_device *adev);
+int amdgpu_aca_reset(struct amdgpu_device *adev);
+void amdgpu_aca_set_smu_funcs(struct amdgpu_device *adev, const struct aca_smu_funcs *smu_funcs);
+bool amdgpu_aca_is_enabled(struct amdgpu_device *adev);
+
+int aca_bank_info_decode(struct aca_bank *bank, struct aca_bank_info *info);
+int aca_bank_check_error_codes(struct amdgpu_device *adev, struct aca_bank *bank, int *err_codes, int size);
+
+int amdgpu_aca_add_handle(struct amdgpu_device *adev, struct aca_handle *handle,
+			  const char *name, const struct aca_info *aca_info, void *data);
+void amdgpu_aca_remove_handle(struct aca_handle *handle);
+int amdgpu_aca_get_error_data(struct amdgpu_device *adev, struct aca_handle *handle,
+			      enum aca_error_type type, struct ras_err_data *err_data,
+			      struct ras_query_context *qctx);
+int amdgpu_aca_smu_set_debug_mode(struct amdgpu_device *adev, bool en);
+void amdgpu_aca_smu_debugfs_init(struct amdgpu_device *adev, struct dentry *root);
+int aca_error_cache_log_bank_error(struct aca_handle *handle, struct aca_bank_info *info,
+				   enum aca_error_type type, u64 count);
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_acp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_acp.c
index cc9c9f8b23b2..381ef205b0df 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_acp.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_acp.c
@@ -29,6 +29,8 @@
 #include <linux/platform_device.h>
 #include <sound/designware_i2s.h>
 #include <sound/pcm.h>
+#include <linux/acpi.h>
+#include <linux/dmi.h>
 
 #include "amdgpu.h"
 #include "atom.h"
@@ -36,6 +38,7 @@
 
 #include "acp_gfx_if.h"
 
+#define ST_JADEITE 1
 #define ACP_TILE_ON_MASK			0x03
 #define ACP_TILE_OFF_MASK			0x02
 #define ACP_TILE_ON_RETAIN_REG_MASK		0x1f
@@ -85,6 +88,8 @@
 #define ACP_DEVS				4
 #define ACP_SRC_ID				162
 
+static unsigned long acp_machine_id;
+
 enum {
 	ACP_TILE_P1 = 0,
 	ACP_TILE_P2,
@@ -93,9 +98,9 @@ enum {
 	ACP_TILE_DSP2,
 };
 
-static int acp_sw_init(void *handle)
+static int acp_sw_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	adev->acp.parent = adev->dev;
 
@@ -107,9 +112,9 @@ static int acp_sw_init(void *handle)
 	return 0;
 }
 
-static int acp_sw_fini(void *handle)
+static int acp_sw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (adev->acp.cgs_device)
 		amdgpu_cgs_destroy_device(adev->acp.cgs_device);
@@ -128,16 +133,14 @@ static int acp_poweroff(struct generic_pm_domain *genpd)
 	struct amdgpu_device *adev;
 
 	apd = container_of(genpd, struct acp_pm_domain, gpd);
-	if (apd != NULL) {
-		adev = apd->adev;
+	adev = apd->adev;
 	/* call smu to POWER GATE ACP block
 	 * smu will
 	 * 1. turn off the acp clock
 	 * 2. power off the acp tiles
 	 * 3. check and enter ulv state
 	 */
-		amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_ACP, true);
-	}
+	amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_ACP, true, 0);
 	return 0;
 }
 
@@ -147,16 +150,14 @@ static int acp_poweron(struct generic_pm_domain *genpd)
 	struct amdgpu_device *adev;
 
 	apd = container_of(genpd, struct acp_pm_domain, gpd);
-	if (apd != NULL) {
-		adev = apd->adev;
+	adev = apd->adev;
 	/* call smu to UNGATE ACP block
 	 * smu will
 	 * 1. exit ulv
 	 * 2. turn on acp clock
 	 * 3. power on acp tiles
 	 */
-		amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_ACP, false);
-	}
+	amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_ACP, false, 0);
 	return 0;
 }
 
@@ -184,33 +185,58 @@ static int acp_genpd_remove_device(struct device *dev, void *data)
 	return 0;
 }
 
+static int acp_quirk_cb(const struct dmi_system_id *id)
+{
+	acp_machine_id = ST_JADEITE;
+	return 1;
+}
+
+static const struct dmi_system_id acp_quirk_table[] = {
+	{
+		.callback = acp_quirk_cb,
+		.matches = {
+			DMI_EXACT_MATCH(DMI_BOARD_VENDOR, "AMD"),
+			DMI_EXACT_MATCH(DMI_PRODUCT_NAME, "Jadeite"),
+		}
+	},
+	{
+		.callback = acp_quirk_cb,
+		.matches = {
+			DMI_EXACT_MATCH(DMI_BOARD_VENDOR, "IP3 Technology CO.,Ltd."),
+			DMI_EXACT_MATCH(DMI_PRODUCT_NAME, "ASN1D"),
+		},
+	},
+	{
+		.callback = acp_quirk_cb,
+		.matches = {
+			DMI_EXACT_MATCH(DMI_BOARD_VENDOR, "Standard"),
+			DMI_EXACT_MATCH(DMI_PRODUCT_NAME, "ASN10"),
+		},
+	},
+	{}
+};
+
 /**
  * acp_hw_init - start and test ACP block
  *
- * @handle: handle used to pass amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  */
-static int acp_hw_init(void *handle)
+static int acp_hw_init(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	uint64_t acp_base;
+	u64 acp_base;
 	u32 val = 0;
 	u32 count = 0;
 	struct i2s_platform_data *i2s_pdata = NULL;
 
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	const struct amdgpu_ip_block *ip_block =
-		amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_ACP);
-
-	if (!ip_block)
-		return -EINVAL;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	r = amd_acp_hw_init(adev->acp.cgs_device,
 			    ip_block->version->major, ip_block->version->minor);
 	/* -ENODEV means board uses AZ rather than ACP */
 	if (r == -ENODEV) {
-		amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_ACP, true);
+		amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_ACP, true, 0);
 		return 0;
 	} else if (r) {
 		return r;
@@ -220,141 +246,208 @@ static int acp_hw_init(void *handle)
 		return -EINVAL;
 
 	acp_base = adev->rmmio_base;
-
-
 	adev->acp.acp_genpd = kzalloc(sizeof(struct acp_pm_domain), GFP_KERNEL);
-	if (adev->acp.acp_genpd == NULL)
+	if (!adev->acp.acp_genpd)
 		return -ENOMEM;
 
 	adev->acp.acp_genpd->gpd.name = "ACP_AUDIO";
 	adev->acp.acp_genpd->gpd.power_off = acp_poweroff;
 	adev->acp.acp_genpd->gpd.power_on = acp_poweron;
-
-
 	adev->acp.acp_genpd->adev = adev;
 
 	pm_genpd_init(&adev->acp.acp_genpd->gpd, NULL, false);
+	dmi_check_system(acp_quirk_table);
+	switch (acp_machine_id) {
+	case ST_JADEITE:
+	{
+		adev->acp.acp_cell = kcalloc(2, sizeof(struct mfd_cell),
+					     GFP_KERNEL);
+		if (!adev->acp.acp_cell) {
+			r = -ENOMEM;
+			goto failure;
+		}
 
-	adev->acp.acp_cell = kcalloc(ACP_DEVS, sizeof(struct mfd_cell),
-							GFP_KERNEL);
-
-	if (adev->acp.acp_cell == NULL) {
-		r = -ENOMEM;
-		goto failure;
-	}
-
-	adev->acp.acp_res = kcalloc(5, sizeof(struct resource), GFP_KERNEL);
-	if (adev->acp.acp_res == NULL) {
-		r = -ENOMEM;
-		goto failure;
-	}
+		adev->acp.acp_res = kcalloc(3, sizeof(struct resource), GFP_KERNEL);
+		if (!adev->acp.acp_res) {
+			r = -ENOMEM;
+			goto failure;
+		}
 
-	i2s_pdata = kcalloc(3, sizeof(struct i2s_platform_data), GFP_KERNEL);
-	if (i2s_pdata == NULL) {
-		r = -ENOMEM;
-		goto failure;
-	}
+		i2s_pdata = kcalloc(1, sizeof(struct i2s_platform_data), GFP_KERNEL);
+		if (!i2s_pdata) {
+			r = -ENOMEM;
+			goto failure;
+		}
 
-	switch (adev->asic_type) {
-	case CHIP_STONEY:
 		i2s_pdata[0].quirks = DW_I2S_QUIRK_COMP_REG_OFFSET |
-			DW_I2S_QUIRK_16BIT_IDX_OVERRIDE;
+				      DW_I2S_QUIRK_16BIT_IDX_OVERRIDE;
+		i2s_pdata[0].cap = DWC_I2S_PLAY | DWC_I2S_RECORD;
+		i2s_pdata[0].snd_rates = SNDRV_PCM_RATE_8000_96000;
+		i2s_pdata[0].i2s_reg_comp1 = ACP_I2S_COMP1_CAP_REG_OFFSET;
+		i2s_pdata[0].i2s_reg_comp2 = ACP_I2S_COMP2_CAP_REG_OFFSET;
+
+		adev->acp.acp_res[0].name = "acp2x_dma";
+		adev->acp.acp_res[0].flags = IORESOURCE_MEM;
+		adev->acp.acp_res[0].start = acp_base;
+		adev->acp.acp_res[0].end = acp_base + ACP_DMA_REGS_END;
+
+		adev->acp.acp_res[1].name = "acp2x_dw_i2s_play_cap";
+		adev->acp.acp_res[1].flags = IORESOURCE_MEM;
+		adev->acp.acp_res[1].start = acp_base + ACP_I2S_CAP_REGS_START;
+		adev->acp.acp_res[1].end = acp_base + ACP_I2S_CAP_REGS_END;
+
+		adev->acp.acp_res[2].name = "acp2x_dma_irq";
+		adev->acp.acp_res[2].flags = IORESOURCE_IRQ;
+		adev->acp.acp_res[2].start = amdgpu_irq_create_mapping(adev, 162);
+		adev->acp.acp_res[2].end = adev->acp.acp_res[2].start;
+
+		adev->acp.acp_cell[0].name = "acp_audio_dma";
+		adev->acp.acp_cell[0].id = 0;
+		adev->acp.acp_cell[0].num_resources = 3;
+		adev->acp.acp_cell[0].resources = &adev->acp.acp_res[0];
+		adev->acp.acp_cell[0].platform_data = &adev->asic_type;
+		adev->acp.acp_cell[0].pdata_size = sizeof(adev->asic_type);
+
+		adev->acp.acp_cell[1].name = "designware-i2s";
+		adev->acp.acp_cell[1].id = 1;
+		adev->acp.acp_cell[1].num_resources = 1;
+		adev->acp.acp_cell[1].resources = &adev->acp.acp_res[1];
+		adev->acp.acp_cell[1].platform_data = &i2s_pdata[0];
+		adev->acp.acp_cell[1].pdata_size = sizeof(struct i2s_platform_data);
+		r = mfd_add_devices(adev->acp.parent, 0, adev->acp.acp_cell, 2, NULL, 0, NULL);
+		if (r)
+			goto failure;
+		r = device_for_each_child(adev->acp.parent, &adev->acp.acp_genpd->gpd,
+					  acp_genpd_add_device);
+		if (r)
+			goto failure;
 		break;
-	default:
-		i2s_pdata[0].quirks = DW_I2S_QUIRK_COMP_REG_OFFSET;
 	}
-	i2s_pdata[0].cap = DWC_I2S_PLAY;
-	i2s_pdata[0].snd_rates = SNDRV_PCM_RATE_8000_96000;
-	i2s_pdata[0].i2s_reg_comp1 = ACP_I2S_COMP1_PLAY_REG_OFFSET;
-	i2s_pdata[0].i2s_reg_comp2 = ACP_I2S_COMP2_PLAY_REG_OFFSET;
-	switch (adev->asic_type) {
-	case CHIP_STONEY:
-		i2s_pdata[1].quirks = DW_I2S_QUIRK_COMP_REG_OFFSET |
-			DW_I2S_QUIRK_COMP_PARAM1 |
-			DW_I2S_QUIRK_16BIT_IDX_OVERRIDE;
-		break;
 	default:
-		i2s_pdata[1].quirks = DW_I2S_QUIRK_COMP_REG_OFFSET |
-			DW_I2S_QUIRK_COMP_PARAM1;
-	}
+		adev->acp.acp_cell = kcalloc(ACP_DEVS, sizeof(struct mfd_cell),
+					     GFP_KERNEL);
 
-	i2s_pdata[1].cap = DWC_I2S_RECORD;
-	i2s_pdata[1].snd_rates = SNDRV_PCM_RATE_8000_96000;
-	i2s_pdata[1].i2s_reg_comp1 = ACP_I2S_COMP1_CAP_REG_OFFSET;
-	i2s_pdata[1].i2s_reg_comp2 = ACP_I2S_COMP2_CAP_REG_OFFSET;
+		if (!adev->acp.acp_cell) {
+			r = -ENOMEM;
+			goto failure;
+		}
 
-	i2s_pdata[2].quirks = DW_I2S_QUIRK_COMP_REG_OFFSET;
-	switch (adev->asic_type) {
-	case CHIP_STONEY:
-		i2s_pdata[2].quirks |= DW_I2S_QUIRK_16BIT_IDX_OVERRIDE;
-		break;
-	default:
-		break;
-	}
+		adev->acp.acp_res = kcalloc(5, sizeof(struct resource), GFP_KERNEL);
+		if (!adev->acp.acp_res) {
+			r = -ENOMEM;
+			goto failure;
+		}
+
+		i2s_pdata = kcalloc(3, sizeof(struct i2s_platform_data), GFP_KERNEL);
+		if (!i2s_pdata) {
+			r = -ENOMEM;
+			goto failure;
+		}
+
+		switch (adev->asic_type) {
+		case CHIP_STONEY:
+			i2s_pdata[0].quirks = DW_I2S_QUIRK_COMP_REG_OFFSET |
+				DW_I2S_QUIRK_16BIT_IDX_OVERRIDE;
+			break;
+		default:
+			i2s_pdata[0].quirks = DW_I2S_QUIRK_COMP_REG_OFFSET;
+		}
+		i2s_pdata[0].cap = DWC_I2S_PLAY;
+		i2s_pdata[0].snd_rates = SNDRV_PCM_RATE_8000_96000;
+		i2s_pdata[0].i2s_reg_comp1 = ACP_I2S_COMP1_PLAY_REG_OFFSET;
+		i2s_pdata[0].i2s_reg_comp2 = ACP_I2S_COMP2_PLAY_REG_OFFSET;
+		switch (adev->asic_type) {
+		case CHIP_STONEY:
+			i2s_pdata[1].quirks = DW_I2S_QUIRK_COMP_REG_OFFSET |
+				DW_I2S_QUIRK_COMP_PARAM1 |
+				DW_I2S_QUIRK_16BIT_IDX_OVERRIDE;
+			break;
+		default:
+			i2s_pdata[1].quirks = DW_I2S_QUIRK_COMP_REG_OFFSET |
+				DW_I2S_QUIRK_COMP_PARAM1;
+		}
 
-	i2s_pdata[2].cap = DWC_I2S_PLAY | DWC_I2S_RECORD;
-	i2s_pdata[2].snd_rates = SNDRV_PCM_RATE_8000_96000;
-	i2s_pdata[2].i2s_reg_comp1 = ACP_BT_COMP1_REG_OFFSET;
-	i2s_pdata[2].i2s_reg_comp2 = ACP_BT_COMP2_REG_OFFSET;
-
-	adev->acp.acp_res[0].name = "acp2x_dma";
-	adev->acp.acp_res[0].flags = IORESOURCE_MEM;
-	adev->acp.acp_res[0].start = acp_base;
-	adev->acp.acp_res[0].end = acp_base + ACP_DMA_REGS_END;
-
-	adev->acp.acp_res[1].name = "acp2x_dw_i2s_play";
-	adev->acp.acp_res[1].flags = IORESOURCE_MEM;
-	adev->acp.acp_res[1].start = acp_base + ACP_I2S_PLAY_REGS_START;
-	adev->acp.acp_res[1].end = acp_base + ACP_I2S_PLAY_REGS_END;
-
-	adev->acp.acp_res[2].name = "acp2x_dw_i2s_cap";
-	adev->acp.acp_res[2].flags = IORESOURCE_MEM;
-	adev->acp.acp_res[2].start = acp_base + ACP_I2S_CAP_REGS_START;
-	adev->acp.acp_res[2].end = acp_base + ACP_I2S_CAP_REGS_END;
-
-	adev->acp.acp_res[3].name = "acp2x_dw_bt_i2s_play_cap";
-	adev->acp.acp_res[3].flags = IORESOURCE_MEM;
-	adev->acp.acp_res[3].start = acp_base + ACP_BT_PLAY_REGS_START;
-	adev->acp.acp_res[3].end = acp_base + ACP_BT_PLAY_REGS_END;
-
-	adev->acp.acp_res[4].name = "acp2x_dma_irq";
-	adev->acp.acp_res[4].flags = IORESOURCE_IRQ;
-	adev->acp.acp_res[4].start = amdgpu_irq_create_mapping(adev, 162);
-	adev->acp.acp_res[4].end = adev->acp.acp_res[4].start;
-
-	adev->acp.acp_cell[0].name = "acp_audio_dma";
-	adev->acp.acp_cell[0].num_resources = 5;
-	adev->acp.acp_cell[0].resources = &adev->acp.acp_res[0];
-	adev->acp.acp_cell[0].platform_data = &adev->asic_type;
-	adev->acp.acp_cell[0].pdata_size = sizeof(adev->asic_type);
-
-	adev->acp.acp_cell[1].name = "designware-i2s";
-	adev->acp.acp_cell[1].num_resources = 1;
-	adev->acp.acp_cell[1].resources = &adev->acp.acp_res[1];
-	adev->acp.acp_cell[1].platform_data = &i2s_pdata[0];
-	adev->acp.acp_cell[1].pdata_size = sizeof(struct i2s_platform_data);
-
-	adev->acp.acp_cell[2].name = "designware-i2s";
-	adev->acp.acp_cell[2].num_resources = 1;
-	adev->acp.acp_cell[2].resources = &adev->acp.acp_res[2];
-	adev->acp.acp_cell[2].platform_data = &i2s_pdata[1];
-	adev->acp.acp_cell[2].pdata_size = sizeof(struct i2s_platform_data);
-
-	adev->acp.acp_cell[3].name = "designware-i2s";
-	adev->acp.acp_cell[3].num_resources = 1;
-	adev->acp.acp_cell[3].resources = &adev->acp.acp_res[3];
-	adev->acp.acp_cell[3].platform_data = &i2s_pdata[2];
-	adev->acp.acp_cell[3].pdata_size = sizeof(struct i2s_platform_data);
-
-	r = mfd_add_hotplug_devices(adev->acp.parent, adev->acp.acp_cell,
-								ACP_DEVS);
-	if (r)
-		goto failure;
-
-	r = device_for_each_child(adev->acp.parent, &adev->acp.acp_genpd->gpd,
-				  acp_genpd_add_device);
-	if (r)
-		goto failure;
+		i2s_pdata[1].cap = DWC_I2S_RECORD;
+		i2s_pdata[1].snd_rates = SNDRV_PCM_RATE_8000_96000;
+		i2s_pdata[1].i2s_reg_comp1 = ACP_I2S_COMP1_CAP_REG_OFFSET;
+		i2s_pdata[1].i2s_reg_comp2 = ACP_I2S_COMP2_CAP_REG_OFFSET;
+
+		i2s_pdata[2].quirks = DW_I2S_QUIRK_COMP_REG_OFFSET;
+		switch (adev->asic_type) {
+		case CHIP_STONEY:
+			i2s_pdata[2].quirks |= DW_I2S_QUIRK_16BIT_IDX_OVERRIDE;
+			break;
+		default:
+			break;
+		}
+
+		i2s_pdata[2].cap = DWC_I2S_PLAY | DWC_I2S_RECORD;
+		i2s_pdata[2].snd_rates = SNDRV_PCM_RATE_8000_96000;
+		i2s_pdata[2].i2s_reg_comp1 = ACP_BT_COMP1_REG_OFFSET;
+		i2s_pdata[2].i2s_reg_comp2 = ACP_BT_COMP2_REG_OFFSET;
+
+		adev->acp.acp_res[0].name = "acp2x_dma";
+		adev->acp.acp_res[0].flags = IORESOURCE_MEM;
+		adev->acp.acp_res[0].start = acp_base;
+		adev->acp.acp_res[0].end = acp_base + ACP_DMA_REGS_END;
+
+		adev->acp.acp_res[1].name = "acp2x_dw_i2s_play";
+		adev->acp.acp_res[1].flags = IORESOURCE_MEM;
+		adev->acp.acp_res[1].start = acp_base + ACP_I2S_PLAY_REGS_START;
+		adev->acp.acp_res[1].end = acp_base + ACP_I2S_PLAY_REGS_END;
+
+		adev->acp.acp_res[2].name = "acp2x_dw_i2s_cap";
+		adev->acp.acp_res[2].flags = IORESOURCE_MEM;
+		adev->acp.acp_res[2].start = acp_base + ACP_I2S_CAP_REGS_START;
+		adev->acp.acp_res[2].end = acp_base + ACP_I2S_CAP_REGS_END;
+
+		adev->acp.acp_res[3].name = "acp2x_dw_bt_i2s_play_cap";
+		adev->acp.acp_res[3].flags = IORESOURCE_MEM;
+		adev->acp.acp_res[3].start = acp_base + ACP_BT_PLAY_REGS_START;
+		adev->acp.acp_res[3].end = acp_base + ACP_BT_PLAY_REGS_END;
+
+		adev->acp.acp_res[4].name = "acp2x_dma_irq";
+		adev->acp.acp_res[4].flags = IORESOURCE_IRQ;
+		adev->acp.acp_res[4].start = amdgpu_irq_create_mapping(adev, 162);
+		adev->acp.acp_res[4].end = adev->acp.acp_res[4].start;
+
+		adev->acp.acp_cell[0].name = "acp_audio_dma";
+		adev->acp.acp_cell[0].id = 0;
+		adev->acp.acp_cell[0].num_resources = 5;
+		adev->acp.acp_cell[0].resources = &adev->acp.acp_res[0];
+		adev->acp.acp_cell[0].platform_data = &adev->asic_type;
+		adev->acp.acp_cell[0].pdata_size = sizeof(adev->asic_type);
+
+		adev->acp.acp_cell[1].name = "designware-i2s";
+		adev->acp.acp_cell[1].id = 1;
+		adev->acp.acp_cell[1].num_resources = 1;
+		adev->acp.acp_cell[1].resources = &adev->acp.acp_res[1];
+		adev->acp.acp_cell[1].platform_data = &i2s_pdata[0];
+		adev->acp.acp_cell[1].pdata_size = sizeof(struct i2s_platform_data);
+
+		adev->acp.acp_cell[2].name = "designware-i2s";
+		adev->acp.acp_cell[2].id = 2;
+		adev->acp.acp_cell[2].num_resources = 1;
+		adev->acp.acp_cell[2].resources = &adev->acp.acp_res[2];
+		adev->acp.acp_cell[2].platform_data = &i2s_pdata[1];
+		adev->acp.acp_cell[2].pdata_size = sizeof(struct i2s_platform_data);
+
+		adev->acp.acp_cell[3].name = "designware-i2s";
+		adev->acp.acp_cell[3].id = 3;
+		adev->acp.acp_cell[3].num_resources = 1;
+		adev->acp.acp_cell[3].resources = &adev->acp.acp_res[3];
+		adev->acp.acp_cell[3].platform_data = &i2s_pdata[2];
+		adev->acp.acp_cell[3].pdata_size = sizeof(struct i2s_platform_data);
+
+		r = mfd_add_devices(adev->acp.parent, 0, adev->acp.acp_cell, ACP_DEVS, NULL, 0, NULL);
+		if (r)
+			goto failure;
+
+		r = device_for_each_child(adev->acp.parent, &adev->acp.acp_genpd->gpd,
+					  acp_genpd_add_device);
+		if (r)
+			goto failure;
+	}
 
 	/* Assert Soft reset of ACP */
 	val = cgs_read_register(adev->acp.cgs_device, mmACP_SOFT_RESET);
@@ -410,18 +503,18 @@ failure:
 /**
  * acp_hw_fini - stop the hardware block
  *
- * @handle: handle used to pass amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  */
-static int acp_hw_fini(void *handle)
+static int acp_hw_fini(struct amdgpu_ip_block *ip_block)
 {
 	u32 val = 0;
 	u32 count = 0;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	/* return early if no ACP */
 	if (!adev->acp.acp_genpd) {
-		amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_ACP, false);
+		amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_ACP, false, 0);
 		return 0;
 	}
 
@@ -472,67 +565,50 @@ static int acp_hw_fini(void *handle)
 	return 0;
 }
 
-static int acp_suspend(void *handle)
+static int acp_suspend(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	/* power up on suspend */
 	if (!adev->acp.acp_cell)
-		amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_ACP, false);
+		amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_ACP, false, 0);
 	return 0;
 }
 
-static int acp_resume(void *handle)
+static int acp_resume(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	/* power down again on resume */
 	if (!adev->acp.acp_cell)
-		amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_ACP, true);
+		amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_ACP, true, 0);
 	return 0;
 }
 
-static int acp_early_init(void *handle)
-{
-	return 0;
-}
-
-static bool acp_is_idle(void *handle)
+static bool acp_is_idle(struct amdgpu_ip_block *ip_block)
 {
 	return true;
 }
 
-static int acp_wait_for_idle(void *handle)
-{
-	return 0;
-}
-
-static int acp_soft_reset(void *handle)
-{
-	return 0;
-}
-
-static int acp_set_clockgating_state(void *handle,
+static int acp_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 				     enum amd_clockgating_state state)
 {
 	return 0;
 }
 
-static int acp_set_powergating_state(void *handle,
+static int acp_set_powergating_state(struct amdgpu_ip_block *ip_block,
 				     enum amd_powergating_state state)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	bool enable = (state == AMD_PG_STATE_GATE);
 
-	amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_ACP, enable);
+	amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_ACP, enable, 0);
 
 	return 0;
 }
 
 static const struct amd_ip_funcs acp_ip_funcs = {
 	.name = "acp_ip",
-	.early_init = acp_early_init,
-	.late_init = NULL,
 	.sw_init = acp_sw_init,
 	.sw_fini = acp_sw_fini,
 	.hw_init = acp_hw_init,
@@ -540,14 +616,11 @@ static const struct amd_ip_funcs acp_ip_funcs = {
 	.suspend = acp_suspend,
 	.resume = acp_resume,
 	.is_idle = acp_is_idle,
-	.wait_for_idle = acp_wait_for_idle,
-	.soft_reset = acp_soft_reset,
 	.set_clockgating_state = acp_set_clockgating_state,
 	.set_powergating_state = acp_set_powergating_state,
 };
 
-const struct amdgpu_ip_block_version acp_ip_block =
-{
+const struct amdgpu_ip_block_version acp_ip_block = {
 	.type = AMD_IP_BLOCK_TYPE_ACP,
 	.major = 2,
 	.minor = 2,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c
index 4811b0faafd9..d31460a9e958 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c
@@ -1,3 +1,4 @@
+// SPDX-License-Identifier: MIT
 /*
  * Copyright 2012 Advanced Micro Devices, Inc.
  *
@@ -23,20 +24,60 @@
 
 #include <linux/pci.h>
 #include <linux/acpi.h>
+#include <linux/backlight.h>
 #include <linux/slab.h>
+#include <linux/xarray.h>
 #include <linux/power_supply.h>
 #include <linux/pm_runtime.h>
 #include <linux/suspend.h>
 #include <acpi/video.h>
 #include <acpi/actbl.h>
 
-#include <drm/drm_crtc_helper.h>
 #include "amdgpu.h"
 #include "amdgpu_pm.h"
 #include "amdgpu_display.h"
 #include "amd_acpi.h"
 #include "atom.h"
 
+/* Declare GUID for AMD _DSM method for XCCs */
+static const guid_t amd_xcc_dsm_guid = GUID_INIT(0x8267f5d5, 0xa556, 0x44f2,
+						 0xb8, 0xb4, 0x45, 0x56, 0x2e,
+						 0x8c, 0x5b, 0xec);
+
+#define AMD_XCC_HID_START 3000
+#define AMD_XCC_DSM_GET_NUM_FUNCS 0
+#define AMD_XCC_DSM_GET_SUPP_MODE 1
+#define AMD_XCC_DSM_GET_XCP_MODE 2
+#define AMD_XCC_DSM_GET_VF_XCC_MAPPING 4
+#define AMD_XCC_DSM_GET_TMR_INFO 5
+#define AMD_XCC_DSM_NUM_FUNCS 5
+
+#define AMD_XCC_MAX_HID 24
+
+struct xarray numa_info_xa;
+
+/* Encapsulates the XCD acpi object information */
+struct amdgpu_acpi_xcc_info {
+	struct list_head list;
+	struct amdgpu_numa_info *numa_info;
+	uint8_t xcp_node;
+	uint8_t phy_id;
+	acpi_handle handle;
+};
+
+struct amdgpu_acpi_dev_info {
+	struct list_head list;
+	struct list_head xcc_list;
+	uint32_t sbdf;
+	uint16_t supp_xcp_mode;
+	uint16_t xcp_mode;
+	uint16_t mem_mode;
+	uint64_t tmr_base;
+	uint64_t tmr_size;
+};
+
+struct list_head amdgpu_acpi_dev_list;
+
 struct amdgpu_atif_notification_cfg {
 	bool enabled;
 	int command_code;
@@ -66,9 +107,7 @@ struct amdgpu_atif {
 	struct amdgpu_atif_notifications notifications;
 	struct amdgpu_atif_functions functions;
 	struct amdgpu_atif_notification_cfg notification_cfg;
-#if defined(CONFIG_BACKLIGHT_CLASS_DEVICE) || defined(CONFIG_BACKLIGHT_CLASS_DEVICE_MODULE)
 	struct backlight_device *bd;
-#endif
 	struct amdgpu_dm_backlight_caps backlight_caps;
 };
 
@@ -108,6 +147,7 @@ static union acpi_object *amdgpu_atif_call(struct amdgpu_atif *atif,
 					   struct acpi_buffer *params)
 {
 	acpi_status status;
+	union acpi_object *obj;
 	union acpi_object atif_arg_elements[2];
 	struct acpi_object_list atif_arg;
 	struct acpi_buffer buffer = { ACPI_ALLOCATE_BUFFER, NULL };
@@ -130,16 +170,24 @@ static union acpi_object *amdgpu_atif_call(struct amdgpu_atif *atif,
 
 	status = acpi_evaluate_object(atif->handle, NULL, &atif_arg,
 				      &buffer);
+	obj = (union acpi_object *)buffer.pointer;
 
-	/* Fail only if calling the method fails and ATIF is supported */
-	if (ACPI_FAILURE(status) && status != AE_NOT_FOUND) {
+	/* Fail if calling the method fails */
+	if (ACPI_FAILURE(status)) {
 		DRM_DEBUG_DRIVER("failed to evaluate ATIF got %s\n",
 				 acpi_format_exception(status));
-		kfree(buffer.pointer);
+		kfree(obj);
 		return NULL;
 	}
 
-	return buffer.pointer;
+	if (obj->type != ACPI_TYPE_BUFFER) {
+		DRM_DEBUG_DRIVER("bad object returned from ATIF: %d\n",
+				 obj->type);
+		kfree(obj);
+		return NULL;
+	}
+
+	return obj;
 }
 
 /**
@@ -344,6 +392,12 @@ static int amdgpu_atif_query_backlight_caps(struct amdgpu_atif *atif)
 			characteristics.min_input_signal;
 	atif->backlight_caps.max_input_signal =
 			characteristics.max_input_signal;
+	atif->backlight_caps.ac_level = characteristics.ac_level;
+	atif->backlight_caps.dc_level = characteristics.dc_level;
+	atif->backlight_caps.data_points = characteristics.number_of_points;
+	memcpy(atif->backlight_caps.luminance_data,
+	       characteristics.data_points,
+	       sizeof(atif->backlight_caps.luminance_data));
 out:
 	kfree(info);
 	return err;
@@ -436,7 +490,6 @@ static int amdgpu_atif_handler(struct amdgpu_device *adev,
 		DRM_DEBUG_DRIVER("ATIF: %d pending SBIOS requests\n", count);
 
 		if (req.pending & ATIF_PANEL_BRIGHTNESS_CHANGE_REQUEST) {
-#if defined(CONFIG_BACKLIGHT_CLASS_DEVICE) || defined(CONFIG_BACKLIGHT_CLASS_DEVICE_MODULE)
 			if (atif->bd) {
 				DRM_DEBUG_DRIVER("Changing brightness to %d\n",
 						 req.backlight_level);
@@ -447,7 +500,6 @@ static int amdgpu_atif_handler(struct amdgpu_device *adev,
 				 */
 				backlight_device_set_brightness(atif->bd, req.backlight_level);
 			}
-#endif
 		}
 
 		if (req.pending & ATIF_DGPU_DISPLAY_EVENT) {
@@ -455,7 +507,6 @@ static int amdgpu_atif_handler(struct amdgpu_device *adev,
 				pm_runtime_get_sync(adev_to_drm(adev)->dev);
 				/* Just fire off a uevent and let userspace tell us what to do */
 				drm_helper_hpd_irq_event(adev_to_drm(adev));
-				pm_runtime_mark_last_busy(adev_to_drm(adev)->dev);
 				pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
 			}
 		}
@@ -669,7 +720,7 @@ int amdgpu_acpi_pcie_performance_request(struct amdgpu_device *adev,
 
 	atcs_input.size = sizeof(struct atcs_pref_req_input);
 	/* client id (bit 2-0: func num, 7-3: dev num, 15-8: bus num) */
-	atcs_input.client_id = adev->pdev->devfn | (adev->pdev->bus->number << 8);
+	atcs_input.client_id = pci_dev_id(adev->pdev);
 	atcs_input.valid_flags_mask = ATCS_VALID_FLAGS_MASK;
 	atcs_input.flags = ATCS_WAIT_FOR_COMPLETION;
 	if (advertise)
@@ -739,7 +790,7 @@ int amdgpu_acpi_power_shift_control(struct amdgpu_device *adev,
 
 	atcs_input.size = sizeof(struct atcs_pwr_shift_input);
 	/* dGPU id (bit 2-0: func num, 7-3: dev num, 15-8: bus num) */
-	atcs_input.dgpu_id = adev->pdev->devfn | (adev->pdev->bus->number << 8);
+	atcs_input.dgpu_id = pci_dev_id(adev->pdev);
 	atcs_input.dev_acpi_state = dev_state;
 	atcs_input.drv_state = drv_state;
 
@@ -752,24 +803,25 @@ int amdgpu_acpi_power_shift_control(struct amdgpu_device *adev,
 		return -EIO;
 	}
 
+	kfree(info);
 	return 0;
 }
 
 /**
  * amdgpu_acpi_smart_shift_update - update dGPU device state to SBIOS
  *
- * @dev: drm_device pointer
+ * @adev: amdgpu device pointer
  * @ss_state: current smart shift event
  *
  * returns 0 on success,
  * otherwise return error number.
  */
-int amdgpu_acpi_smart_shift_update(struct drm_device *dev, enum amdgpu_ss ss_state)
+int amdgpu_acpi_smart_shift_update(struct amdgpu_device *adev,
+				   enum amdgpu_ss ss_state)
 {
-	struct amdgpu_device *adev = drm_to_adev(dev);
 	int r;
 
-	if (!amdgpu_device_supports_smart_shift(dev))
+	if (!amdgpu_device_supports_smart_shift(adev))
 		return 0;
 
 	switch (ss_state) {
@@ -804,6 +856,347 @@ int amdgpu_acpi_smart_shift_update(struct drm_device *dev, enum amdgpu_ss ss_sta
 	return r;
 }
 
+#ifdef CONFIG_ACPI_NUMA
+static inline uint64_t amdgpu_acpi_get_numa_size(int nid)
+{
+	/* This is directly using si_meminfo_node implementation as the
+	 * function is not exported.
+	 */
+	int zone_type;
+	uint64_t managed_pages = 0;
+
+	pg_data_t *pgdat = NODE_DATA(nid);
+
+	for (zone_type = 0; zone_type < MAX_NR_ZONES; zone_type++)
+		managed_pages +=
+			zone_managed_pages(&pgdat->node_zones[zone_type]);
+	return managed_pages * PAGE_SIZE;
+}
+
+static struct amdgpu_numa_info *amdgpu_acpi_get_numa_info(uint32_t pxm)
+{
+	struct amdgpu_numa_info *numa_info;
+	int nid;
+
+	numa_info = xa_load(&numa_info_xa, pxm);
+
+	if (!numa_info) {
+		struct sysinfo info;
+
+		numa_info = kzalloc(sizeof(*numa_info), GFP_KERNEL);
+		if (!numa_info)
+			return NULL;
+
+		nid = pxm_to_node(pxm);
+		numa_info->pxm = pxm;
+		numa_info->nid = nid;
+
+		if (numa_info->nid == NUMA_NO_NODE) {
+			si_meminfo(&info);
+			numa_info->size = info.totalram * info.mem_unit;
+		} else {
+			numa_info->size = amdgpu_acpi_get_numa_size(nid);
+		}
+		xa_store(&numa_info_xa, numa_info->pxm, numa_info, GFP_KERNEL);
+	}
+
+	return numa_info;
+}
+#endif
+
+/**
+ * amdgpu_acpi_get_node_id - obtain the NUMA node id for corresponding amdgpu
+ * acpi device handle
+ *
+ * @handle: acpi handle
+ * @numa_info: amdgpu_numa_info structure holding numa information
+ *
+ * Queries the ACPI interface to fetch the corresponding NUMA Node ID for a
+ * given amdgpu acpi device.
+ *
+ * Returns ACPI STATUS OK with Node ID on success or the corresponding failure reason
+ */
+static acpi_status amdgpu_acpi_get_node_id(acpi_handle handle,
+				    struct amdgpu_numa_info **numa_info)
+{
+#ifdef CONFIG_ACPI_NUMA
+	u64 pxm;
+	acpi_status status;
+
+	if (!numa_info)
+		return_ACPI_STATUS(AE_ERROR);
+
+	status = acpi_evaluate_integer(handle, "_PXM", NULL, &pxm);
+
+	if (ACPI_FAILURE(status))
+		return status;
+
+	*numa_info = amdgpu_acpi_get_numa_info(pxm);
+
+	if (!*numa_info)
+		return_ACPI_STATUS(AE_ERROR);
+
+	return_ACPI_STATUS(AE_OK);
+#else
+	return_ACPI_STATUS(AE_NOT_EXIST);
+#endif
+}
+
+static struct amdgpu_acpi_dev_info *amdgpu_acpi_get_dev(u32 sbdf)
+{
+	struct amdgpu_acpi_dev_info *acpi_dev;
+
+	if (list_empty(&amdgpu_acpi_dev_list))
+		return NULL;
+
+	list_for_each_entry(acpi_dev, &amdgpu_acpi_dev_list, list)
+		if (acpi_dev->sbdf == sbdf)
+			return acpi_dev;
+
+	return NULL;
+}
+
+static int amdgpu_acpi_dev_init(struct amdgpu_acpi_dev_info **dev_info,
+				struct amdgpu_acpi_xcc_info *xcc_info, u32 sbdf)
+{
+	struct amdgpu_acpi_dev_info *tmp;
+	union acpi_object *obj;
+	int ret = -ENOENT;
+
+	*dev_info = NULL;
+	tmp = kzalloc(sizeof(struct amdgpu_acpi_dev_info), GFP_KERNEL);
+	if (!tmp)
+		return -ENOMEM;
+
+	INIT_LIST_HEAD(&tmp->xcc_list);
+	INIT_LIST_HEAD(&tmp->list);
+	tmp->sbdf = sbdf;
+
+	obj = acpi_evaluate_dsm_typed(xcc_info->handle, &amd_xcc_dsm_guid, 0,
+				      AMD_XCC_DSM_GET_SUPP_MODE, NULL,
+				      ACPI_TYPE_INTEGER);
+
+	if (!obj) {
+		acpi_handle_debug(xcc_info->handle,
+				  "_DSM function %d evaluation failed",
+				  AMD_XCC_DSM_GET_SUPP_MODE);
+		ret = -ENOENT;
+		goto out;
+	}
+
+	tmp->supp_xcp_mode = obj->integer.value & 0xFFFF;
+	ACPI_FREE(obj);
+
+	obj = acpi_evaluate_dsm_typed(xcc_info->handle, &amd_xcc_dsm_guid, 0,
+				      AMD_XCC_DSM_GET_XCP_MODE, NULL,
+				      ACPI_TYPE_INTEGER);
+
+	if (!obj) {
+		acpi_handle_debug(xcc_info->handle,
+				  "_DSM function %d evaluation failed",
+				  AMD_XCC_DSM_GET_XCP_MODE);
+		ret = -ENOENT;
+		goto out;
+	}
+
+	tmp->xcp_mode = obj->integer.value & 0xFFFF;
+	tmp->mem_mode = (obj->integer.value >> 32) & 0xFFFF;
+	ACPI_FREE(obj);
+
+	/* Evaluate DSMs and fill XCC information */
+	obj = acpi_evaluate_dsm_typed(xcc_info->handle, &amd_xcc_dsm_guid, 0,
+				      AMD_XCC_DSM_GET_TMR_INFO, NULL,
+				      ACPI_TYPE_PACKAGE);
+
+	if (!obj || obj->package.count < 2) {
+		acpi_handle_debug(xcc_info->handle,
+				  "_DSM function %d evaluation failed",
+				  AMD_XCC_DSM_GET_TMR_INFO);
+		ret = -ENOENT;
+		goto out;
+	}
+
+	tmp->tmr_base = obj->package.elements[0].integer.value;
+	tmp->tmr_size = obj->package.elements[1].integer.value;
+	ACPI_FREE(obj);
+
+	DRM_DEBUG_DRIVER(
+		"New dev(%x): Supported xcp mode: %x curr xcp_mode : %x mem mode : %x, tmr base: %llx tmr size: %llx  ",
+		tmp->sbdf, tmp->supp_xcp_mode, tmp->xcp_mode, tmp->mem_mode,
+		tmp->tmr_base, tmp->tmr_size);
+	list_add_tail(&tmp->list, &amdgpu_acpi_dev_list);
+	*dev_info = tmp;
+
+	return 0;
+
+out:
+	if (obj)
+		ACPI_FREE(obj);
+	kfree(tmp);
+
+	return ret;
+}
+
+static int amdgpu_acpi_get_xcc_info(struct amdgpu_acpi_xcc_info *xcc_info,
+				    u32 *sbdf)
+{
+	union acpi_object *obj;
+	acpi_status status;
+	int ret = -ENOENT;
+
+	obj = acpi_evaluate_dsm_typed(xcc_info->handle, &amd_xcc_dsm_guid, 0,
+				      AMD_XCC_DSM_GET_NUM_FUNCS, NULL,
+				      ACPI_TYPE_INTEGER);
+
+	if (!obj || obj->integer.value != AMD_XCC_DSM_NUM_FUNCS)
+		goto out;
+	ACPI_FREE(obj);
+
+	/* Evaluate DSMs and fill XCC information */
+	obj = acpi_evaluate_dsm_typed(xcc_info->handle, &amd_xcc_dsm_guid, 0,
+				      AMD_XCC_DSM_GET_VF_XCC_MAPPING, NULL,
+				      ACPI_TYPE_INTEGER);
+
+	if (!obj) {
+		acpi_handle_debug(xcc_info->handle,
+				  "_DSM function %d evaluation failed",
+				  AMD_XCC_DSM_GET_VF_XCC_MAPPING);
+		ret = -EINVAL;
+		goto out;
+	}
+
+	/* PF xcc id [39:32] */
+	xcc_info->phy_id = (obj->integer.value >> 32) & 0xFF;
+	/* xcp node of this xcc [47:40] */
+	xcc_info->xcp_node = (obj->integer.value >> 40) & 0xFF;
+	/* PF domain of this xcc [31:16] */
+	*sbdf = (obj->integer.value) & 0xFFFF0000;
+	/* PF bus/dev/fn of this xcc [63:48] */
+	*sbdf |= (obj->integer.value >> 48) & 0xFFFF;
+	ACPI_FREE(obj);
+	obj = NULL;
+
+	status =
+		amdgpu_acpi_get_node_id(xcc_info->handle, &xcc_info->numa_info);
+
+	/* TODO: check if this check is required */
+	if (ACPI_SUCCESS(status))
+		ret = 0;
+out:
+	if (obj)
+		ACPI_FREE(obj);
+
+	return ret;
+}
+
+static int amdgpu_acpi_enumerate_xcc(void)
+{
+	struct amdgpu_acpi_dev_info *dev_info = NULL;
+	struct amdgpu_acpi_xcc_info *xcc_info;
+	struct acpi_device *acpi_dev;
+	char hid[ACPI_ID_LEN];
+	int ret, id;
+	u32 sbdf;
+
+	INIT_LIST_HEAD(&amdgpu_acpi_dev_list);
+	xa_init(&numa_info_xa);
+
+	for (id = 0; id < AMD_XCC_MAX_HID; id++) {
+		sprintf(hid, "%s%d", "AMD", AMD_XCC_HID_START + id);
+		acpi_dev = acpi_dev_get_first_match_dev(hid, NULL, -1);
+		/* These ACPI objects are expected to be in sequential order. If
+		 * one is not found, no need to check the rest.
+		 */
+		if (!acpi_dev) {
+			DRM_DEBUG_DRIVER("No matching acpi device found for %s",
+					 hid);
+			break;
+		}
+
+		xcc_info = kzalloc(sizeof(struct amdgpu_acpi_xcc_info),
+				   GFP_KERNEL);
+		if (!xcc_info) {
+			DRM_ERROR("Failed to allocate memory for xcc info\n");
+			return -ENOMEM;
+		}
+
+		INIT_LIST_HEAD(&xcc_info->list);
+		xcc_info->handle = acpi_device_handle(acpi_dev);
+		acpi_dev_put(acpi_dev);
+
+		ret = amdgpu_acpi_get_xcc_info(xcc_info, &sbdf);
+		if (ret) {
+			kfree(xcc_info);
+			continue;
+		}
+
+		dev_info = amdgpu_acpi_get_dev(sbdf);
+
+		if (!dev_info)
+			ret = amdgpu_acpi_dev_init(&dev_info, xcc_info, sbdf);
+
+		if (ret == -ENOMEM)
+			return ret;
+
+		if (!dev_info) {
+			kfree(xcc_info);
+			continue;
+		}
+
+		list_add_tail(&xcc_info->list, &dev_info->xcc_list);
+	}
+
+	return 0;
+}
+
+int amdgpu_acpi_get_tmr_info(struct amdgpu_device *adev, u64 *tmr_offset,
+			     u64 *tmr_size)
+{
+	struct amdgpu_acpi_dev_info *dev_info;
+	u32 sbdf;
+
+	if (!tmr_offset || !tmr_size)
+		return -EINVAL;
+
+	sbdf = (pci_domain_nr(adev->pdev->bus) << 16);
+	sbdf |= pci_dev_id(adev->pdev);
+	dev_info = amdgpu_acpi_get_dev(sbdf);
+	if (!dev_info)
+		return -ENOENT;
+
+	*tmr_offset = dev_info->tmr_base;
+	*tmr_size = dev_info->tmr_size;
+
+	return 0;
+}
+
+int amdgpu_acpi_get_mem_info(struct amdgpu_device *adev, int xcc_id,
+			     struct amdgpu_numa_info *numa_info)
+{
+	struct amdgpu_acpi_dev_info *dev_info;
+	struct amdgpu_acpi_xcc_info *xcc_info;
+	u32 sbdf;
+
+	if (!numa_info)
+		return -EINVAL;
+
+	sbdf = (pci_domain_nr(adev->pdev->bus) << 16);
+	sbdf |= pci_dev_id(adev->pdev);
+	dev_info = amdgpu_acpi_get_dev(sbdf);
+	if (!dev_info)
+		return -ENOENT;
+
+	list_for_each_entry(xcc_info, &dev_info->xcc_list, list) {
+		if (xcc_info->phy_id == xcc_id) {
+			memcpy(numa_info, xcc_info->numa_info,
+			       sizeof(*numa_info));
+			return 0;
+		}
+	}
+
+	return -ENOENT;
+}
+
 /**
  * amdgpu_acpi_event - handle notify events
  *
@@ -849,11 +1242,11 @@ int amdgpu_acpi_init(struct amdgpu_device *adev)
 {
 	struct amdgpu_atif *atif = &amdgpu_acpi_priv.atif;
 
-#if defined(CONFIG_BACKLIGHT_CLASS_DEVICE) || defined(CONFIG_BACKLIGHT_CLASS_DEVICE_MODULE)
 	if (atif->notifications.brightness_change) {
-		if (amdgpu_device_has_dc_support(adev)) {
+		if (adev->dc_enabled) {
 #if defined(CONFIG_DRM_AMD_DC)
 			struct amdgpu_display_manager *dm = &adev->dm;
+
 			if (dm->backlight_dev[0])
 				atif->bd = dm->backlight_dev[0];
 #endif
@@ -868,6 +1261,7 @@ int amdgpu_acpi_init(struct amdgpu_device *adev)
 				if ((enc->devices & (ATOM_DEVICE_LCD_SUPPORT)) &&
 				    enc->enc_priv) {
 					struct amdgpu_encoder_atom_dig *dig = enc->enc_priv;
+
 					if (dig->bl_dev) {
 						atif->bd = dig->bl_dev;
 						break;
@@ -876,7 +1270,6 @@ int amdgpu_acpi_init(struct amdgpu_device *adev)
 			}
 		}
 	}
-#endif
 	adev->acpi_nb.notifier_call = amdgpu_acpi_event;
 	register_acpi_notifier(&adev->acpi_nb);
 
@@ -887,9 +1280,7 @@ void amdgpu_acpi_get_backlight_caps(struct amdgpu_dm_backlight_caps *caps)
 {
 	struct amdgpu_atif *atif = &amdgpu_acpi_priv.atif;
 
-	caps->caps_valid = atif->backlight_caps.caps_valid;
-	caps->min_input_signal = atif->backlight_caps.min_input_signal;
-	caps->max_input_signal = atif->backlight_caps.max_input_signal;
+	memcpy(caps, &atif->backlight_caps, sizeof(*caps));
 }
 
 /**
@@ -925,9 +1316,9 @@ static bool amdgpu_atif_pci_probe_handle(struct pci_dev *pdev)
 		return false;
 
 	status = acpi_get_handle(dhandle, "ATIF", &atif_handle);
-	if (ACPI_FAILURE(status)) {
+	if (ACPI_FAILURE(status))
 		return false;
-	}
+
 	amdgpu_acpi_priv.atif.handle = atif_handle;
 	acpi_get_name(amdgpu_acpi_priv.atif.handle, ACPI_FULL_PATHNAME, &buffer);
 	DRM_DEBUG_DRIVER("Found ATIF handle %s\n", acpi_method_name);
@@ -960,9 +1351,9 @@ static bool amdgpu_atcs_pci_probe_handle(struct pci_dev *pdev)
 		return false;
 
 	status = acpi_get_handle(dhandle, "ATCS", &atcs_handle);
-	if (ACPI_FAILURE(status)) {
+	if (ACPI_FAILURE(status))
 		return false;
-	}
+
 	amdgpu_acpi_priv.atcs.handle = atcs_handle;
 	acpi_get_name(amdgpu_acpi_priv.atcs.handle, ACPI_FULL_PATHNAME, &buffer);
 	DRM_DEBUG_DRIVER("Found ATCS handle %s\n", acpi_method_name);
@@ -974,6 +1365,34 @@ static bool amdgpu_atcs_pci_probe_handle(struct pci_dev *pdev)
 	return true;
 }
 
+
+/**
+ * amdgpu_acpi_should_gpu_reset
+ *
+ * @adev: amdgpu_device_pointer
+ *
+ * returns true if should reset GPU, false if not
+ */
+bool amdgpu_acpi_should_gpu_reset(struct amdgpu_device *adev)
+{
+	if ((adev->flags & AMD_IS_APU) &&
+	    adev->gfx.imu.funcs) /* Not need to do mode2 reset for IMU enabled APUs */
+		return false;
+
+	if ((adev->flags & AMD_IS_APU) &&
+	    amdgpu_acpi_is_s3_active(adev))
+		return false;
+
+	if (amdgpu_sriov_vf(adev))
+		return false;
+
+#if IS_ENABLED(CONFIG_SUSPEND)
+	return pm_suspend_target_state != PM_SUSPEND_TO_IDLE;
+#else
+	return true;
+#endif
+}
+
 /*
  * amdgpu_acpi_detect - detect ACPI ATIF/ATCS methods
  *
@@ -987,14 +1406,11 @@ void amdgpu_acpi_detect(void)
 	struct pci_dev *pdev = NULL;
 	int ret;
 
-	while ((pdev = pci_get_class(PCI_CLASS_DISPLAY_VGA << 8, pdev)) != NULL) {
-		if (!atif->handle)
-			amdgpu_atif_pci_probe_handle(pdev);
-		if (!atcs->handle)
-			amdgpu_atcs_pci_probe_handle(pdev);
-	}
+	while ((pdev = pci_get_base_class(PCI_BASE_CLASS_DISPLAY, pdev))) {
+		if ((pdev->class != PCI_CLASS_DISPLAY_VGA << 8) &&
+		    (pdev->class != PCI_CLASS_DISPLAY_OTHER << 8))
+			continue;
 
-	while ((pdev = pci_get_class(PCI_CLASS_DISPLAY_OTHER << 8, pdev)) != NULL) {
 		if (!atif->handle)
 			amdgpu_atif_pci_probe_handle(pdev);
 		if (!atcs->handle)
@@ -1029,6 +1445,50 @@ void amdgpu_acpi_detect(void)
 	} else {
 		atif->backlight_caps.caps_valid = false;
 	}
+
+	amdgpu_acpi_enumerate_xcc();
+}
+
+void amdgpu_acpi_release(void)
+{
+	struct amdgpu_acpi_dev_info *dev_info, *dev_tmp;
+	struct amdgpu_acpi_xcc_info *xcc_info, *xcc_tmp;
+	struct amdgpu_numa_info *numa_info;
+	unsigned long index;
+
+	xa_for_each(&numa_info_xa, index, numa_info) {
+		kfree(numa_info);
+		xa_erase(&numa_info_xa, index);
+	}
+
+	if (list_empty(&amdgpu_acpi_dev_list))
+		return;
+
+	list_for_each_entry_safe(dev_info, dev_tmp, &amdgpu_acpi_dev_list,
+				 list) {
+		list_for_each_entry_safe(xcc_info, xcc_tmp, &dev_info->xcc_list,
+					 list) {
+			list_del(&xcc_info->list);
+			kfree(xcc_info);
+		}
+
+		list_del(&dev_info->list);
+		kfree(dev_info);
+	}
+}
+
+#if IS_ENABLED(CONFIG_SUSPEND)
+/**
+ * amdgpu_acpi_is_s3_active
+ *
+ * @adev: amdgpu_device_pointer
+ *
+ * returns true if supported, false if not.
+ */
+bool amdgpu_acpi_is_s3_active(struct amdgpu_device *adev)
+{
+	return !(adev->flags & AMD_IS_APU) ||
+		(pm_suspend_target_state == PM_SUSPEND_MEM);
 }
 
 /**
@@ -1040,11 +1500,66 @@ void amdgpu_acpi_detect(void)
  */
 bool amdgpu_acpi_is_s0ix_active(struct amdgpu_device *adev)
 {
-#if IS_ENABLED(CONFIG_AMD_PMC) && IS_ENABLED(CONFIG_SUSPEND)
-	if (acpi_gbl_FADT.flags & ACPI_FADT_LOW_POWER_S0) {
-		if (adev->flags & AMD_IS_APU)
-			return pm_suspend_target_state == PM_SUSPEND_TO_IDLE;
+	if (!(adev->flags & AMD_IS_APU) ||
+	    (pm_suspend_target_state != PM_SUSPEND_TO_IDLE))
+		return false;
+
+	if (adev->asic_type < CHIP_RAVEN)
+		return false;
+
+	if (!(adev->pm.pp_feature & PP_GFXOFF_MASK))
+		return false;
+
+	/*
+	 * If ACPI_FADT_LOW_POWER_S0 is not set in the FADT, it is generally
+	 * risky to do any special firmware-related preparations for entering
+	 * S0ix even though the system is suspending to idle, so return false
+	 * in that case.
+	 */
+	if (!(acpi_gbl_FADT.flags & ACPI_FADT_LOW_POWER_S0)) {
+		dev_err_once(adev->dev,
+			      "Power consumption will be higher as BIOS has not been configured for suspend-to-idle.\n"
+			      "To use suspend-to-idle change the sleep mode in BIOS setup.\n");
+		return false;
 	}
-#endif
+
+#if !IS_ENABLED(CONFIG_AMD_PMC)
+	dev_err_once(adev->dev,
+		      "Power consumption will be higher as the kernel has not been compiled with CONFIG_AMD_PMC.\n");
 	return false;
+#else
+	return true;
+#endif /* CONFIG_AMD_PMC */
+}
+#endif /* CONFIG_SUSPEND */
+
+#if IS_ENABLED(CONFIG_DRM_AMD_ISP)
+static const struct acpi_device_id isp_sensor_ids[] = {
+	{ "OMNI5C10" },
+	{ }
+};
+
+static int isp_match_acpi_device_ids(struct device *dev, const void *data)
+{
+	return acpi_match_device(data, dev) ? 1 : 0;
+}
+
+int amdgpu_acpi_get_isp4_dev(struct acpi_device **dev)
+{
+	struct device *pdev __free(put_device) = NULL;
+	struct acpi_device *acpi_pdev;
+
+	pdev = bus_find_device(&platform_bus_type, NULL, isp_sensor_ids,
+			       isp_match_acpi_device_ids);
+	if (!pdev)
+		return -EINVAL;
+
+	acpi_pdev = ACPI_COMPANION(pdev);
+	if (!acpi_pdev)
+		return -ENODEV;
+
+	*dev = acpi_pdev;
+
+	return 0;
 }
+#endif /* CONFIG_DRM_AMD_ISP */
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_afmt.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_afmt.c
index a4d65973bf7c..80771b1480ff 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_afmt.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_afmt.c
@@ -100,6 +100,7 @@ struct amdgpu_afmt_acr amdgpu_afmt_acr(uint32_t clock)
 	amdgpu_afmt_calc_cts(clock, &res.cts_32khz, &res.n_32khz, 32000);
 	amdgpu_afmt_calc_cts(clock, &res.cts_44_1khz, &res.n_44_1khz, 44100);
 	amdgpu_afmt_calc_cts(clock, &res.cts_48khz, &res.n_48khz, 48000);
+	res.clock = clock;
 
 	return res;
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c
index 7077f21f0021..a2879d2b7c8e 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c
@@ -1,3 +1,4 @@
+// SPDX-License-Identifier: MIT
 /*
  * Copyright 2014 Advanced Micro Devices, Inc.
  *
@@ -27,12 +28,14 @@
 #include "amdgpu.h"
 #include "amdgpu_gfx.h"
 #include "amdgpu_dma_buf.h"
+#include <drm/ttm/ttm_tt.h>
 #include <linux/module.h>
 #include <linux/dma-buf.h>
 #include "amdgpu_xgmi.h"
 #include <uapi/linux/kfd_ioctl.h>
 #include "amdgpu_ras.h"
 #include "amdgpu_umc.h"
+#include "amdgpu_reset.h"
 
 /* Total memory size in system memory and all GPU VRAM. Used to
  * estimate worst case amount of memory to reserve for page tables
@@ -51,7 +54,6 @@ int amdgpu_amdkfd_init(void)
 	amdgpu_amdkfd_total_mem_size *= si.mem_unit;
 
 	ret = kgd2kfd_init();
-	amdgpu_amdkfd_gpuvm_init_mem_limits();
 	kfd_initialized = !ret;
 
 	return ret;
@@ -72,10 +74,7 @@ void amdgpu_amdkfd_device_probe(struct amdgpu_device *adev)
 	if (!kfd_initialized)
 		return;
 
-	adev->kfd.dev = kgd2kfd_probe((struct kgd_dev *)adev, vf);
-
-	if (adev->kfd.dev)
-		amdgpu_amdkfd_total_mem_size += adev->gmc.real_vram_size;
+	adev->kfd.dev = kgd2kfd_probe(adev, vf);
 }
 
 /**
@@ -97,13 +96,24 @@ static void amdgpu_doorbell_get_kfd_info(struct amdgpu_device *adev,
 					 size_t *start_offset)
 {
 	/*
-	 * The first num_doorbells are used by amdgpu.
+	 * The first num_kernel_doorbells are used by amdgpu.
 	 * amdkfd takes whatever's left in the aperture.
 	 */
-	if (adev->doorbell.size > adev->doorbell.num_doorbells * sizeof(u32)) {
+	if (adev->enable_mes) {
+		/*
+		 * With MES enabled, we only need to initialize
+		 * the base address. The size and offset are
+		 * not initialized as AMDGPU manages the whole
+		 * doorbell space.
+		 */
+		*aperture_base = adev->doorbell.base;
+		*aperture_size = 0;
+		*start_offset = 0;
+	} else if (adev->doorbell.size > adev->doorbell.num_kernel_doorbells *
+						sizeof(u32)) {
 		*aperture_base = adev->doorbell.base;
 		*aperture_size = adev->doorbell.size;
-		*start_offset = adev->doorbell.num_doorbells * sizeof(u32);
+		*start_offset = adev->doorbell.num_kernel_doorbells * sizeof(u32);
 	} else {
 		*aperture_base = 0;
 		*aperture_size = 0;
@@ -111,11 +121,57 @@ static void amdgpu_doorbell_get_kfd_info(struct amdgpu_device *adev,
 	}
 }
 
+
+static void amdgpu_amdkfd_reset_work(struct work_struct *work)
+{
+	struct amdgpu_device *adev = container_of(work, struct amdgpu_device,
+						  kfd.reset_work);
+
+	struct amdgpu_reset_context reset_context;
+
+	memset(&reset_context, 0, sizeof(reset_context));
+
+	reset_context.method = AMD_RESET_METHOD_NONE;
+	reset_context.reset_req_dev = adev;
+	reset_context.src = adev->enable_mes ?
+			    AMDGPU_RESET_SRC_MES :
+			    AMDGPU_RESET_SRC_HWS;
+	clear_bit(AMDGPU_NEED_FULL_RESET, &reset_context.flags);
+
+	amdgpu_device_gpu_recover(adev, NULL, &reset_context);
+}
+
+static const struct drm_client_funcs kfd_client_funcs = {
+	.unregister	= drm_client_release,
+};
+
+int amdgpu_amdkfd_drm_client_create(struct amdgpu_device *adev)
+{
+	int ret;
+
+	if (!adev->kfd.init_complete || adev->kfd.client.dev)
+		return 0;
+
+	ret = drm_client_init(&adev->ddev, &adev->kfd.client, "kfd",
+			      &kfd_client_funcs);
+	if (ret) {
+		dev_err(adev->dev, "Failed to init DRM client: %d\n",
+			ret);
+		return ret;
+	}
+
+	drm_client_register(&adev->kfd.client);
+
+	return 0;
+}
+
 void amdgpu_amdkfd_device_init(struct amdgpu_device *adev)
 {
 	int i;
 	int last_valid_bit;
 
+	amdgpu_amdkfd_gpuvm_init_mem_limits();
+
 	if (adev->kfd.dev) {
 		struct kgd2kfd_shared_resources gpu_resources = {
 			.compute_vmid_bitmap =
@@ -128,15 +184,15 @@ void amdgpu_amdkfd_device_init(struct amdgpu_device *adev)
 					  AMDGPU_GMC_HOLE_START),
 			.drm_render_minor = adev_to_drm(adev)->render->index,
 			.sdma_doorbell_idx = adev->doorbell_index.sdma_engine,
-
+			.enable_mes = adev->enable_mes,
 		};
 
 		/* this is going to have a few of the MSBs set that we need to
 		 * clear
 		 */
 		bitmap_complement(gpu_resources.cp_queue_bitmap,
-				  adev->gfx.mec.queue_bitmap,
-				  KGD_MAX_QUEUES);
+				  adev->gfx.mec_bitmap[0].queue_bitmap,
+				  AMDGPU_MAX_QUEUES);
 
 		/* According to linux/bitmap.h we shouldn't use bitmap_clear if
 		 * nbits is not compile time constant
@@ -144,7 +200,7 @@ void amdgpu_amdkfd_device_init(struct amdgpu_device *adev)
 		last_valid_bit = 1 /* only first MEC can have compute queues */
 				* adev->gfx.mec.num_pipe_per_mec
 				* adev->gfx.mec.num_queue_per_pipe;
-		for (i = last_valid_bit; i < KGD_MAX_QUEUES; ++i)
+		for (i = last_valid_bit; i < AMDGPU_MAX_QUEUES; ++i)
 			clear_bit(i, gpu_resources.cp_queue_bitmap);
 
 		amdgpu_doorbell_get_kfd_info(adev,
@@ -168,7 +224,11 @@ void amdgpu_amdkfd_device_init(struct amdgpu_device *adev)
 		}
 
 		adev->kfd.init_complete = kgd2kfd_device_init(adev->kfd.dev,
-						adev_to_drm(adev), &gpu_resources);
+							&gpu_resources);
+
+		amdgpu_amdkfd_total_mem_size += adev->gmc.real_vram_size;
+
+		INIT_WORK(&adev->kfd.reset_work, amdgpu_amdkfd_reset_work);
 	}
 }
 
@@ -177,6 +237,7 @@ void amdgpu_amdkfd_device_fini_sw(struct amdgpu_device *adev)
 	if (adev->kfd.dev) {
 		kgd2kfd_device_exit(adev->kfd.dev);
 		adev->kfd.dev = NULL;
+		amdgpu_amdkfd_total_mem_size -= adev->gmc.real_vram_size;
 	}
 }
 
@@ -187,38 +248,53 @@ void amdgpu_amdkfd_interrupt(struct amdgpu_device *adev,
 		kgd2kfd_interrupt(adev->kfd.dev, ih_ring_entry);
 }
 
-void amdgpu_amdkfd_suspend(struct amdgpu_device *adev, bool run_pm)
+void amdgpu_amdkfd_suspend(struct amdgpu_device *adev, bool suspend_proc)
 {
-	if (adev->kfd.dev)
-		kgd2kfd_suspend(adev->kfd.dev, run_pm);
+	if (adev->kfd.dev) {
+		if (adev->in_s0ix)
+			kgd2kfd_stop_sched_all_nodes(adev->kfd.dev);
+		else
+			kgd2kfd_suspend(adev->kfd.dev, suspend_proc);
+	}
 }
 
-int amdgpu_amdkfd_resume_iommu(struct amdgpu_device *adev)
+int amdgpu_amdkfd_resume(struct amdgpu_device *adev, bool resume_proc)
 {
 	int r = 0;
 
-	if (adev->kfd.dev)
-		r = kgd2kfd_resume_iommu(adev->kfd.dev);
+	if (adev->kfd.dev) {
+		if (adev->in_s0ix)
+			r = kgd2kfd_start_sched_all_nodes(adev->kfd.dev);
+		else
+			r = kgd2kfd_resume(adev->kfd.dev, resume_proc);
+	}
 
 	return r;
 }
 
-int amdgpu_amdkfd_resume(struct amdgpu_device *adev, bool run_pm)
+void amdgpu_amdkfd_suspend_process(struct amdgpu_device *adev)
+{
+	if (adev->kfd.dev)
+		kgd2kfd_suspend_process(adev->kfd.dev);
+}
+
+int amdgpu_amdkfd_resume_process(struct amdgpu_device *adev)
 {
 	int r = 0;
 
 	if (adev->kfd.dev)
-		r = kgd2kfd_resume(adev->kfd.dev, run_pm);
+		r = kgd2kfd_resume_process(adev->kfd.dev);
 
 	return r;
 }
 
-int amdgpu_amdkfd_pre_reset(struct amdgpu_device *adev)
+int amdgpu_amdkfd_pre_reset(struct amdgpu_device *adev,
+			    struct amdgpu_reset_context *reset_context)
 {
 	int r = 0;
 
 	if (adev->kfd.dev)
-		r = kgd2kfd_pre_reset(adev->kfd.dev);
+		r = kgd2kfd_pre_reset(adev->kfd.dev, reset_context);
 
 	return r;
 }
@@ -233,19 +309,17 @@ int amdgpu_amdkfd_post_reset(struct amdgpu_device *adev)
 	return r;
 }
 
-void amdgpu_amdkfd_gpu_reset(struct kgd_dev *kgd)
+void amdgpu_amdkfd_gpu_reset(struct amdgpu_device *adev)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
-
 	if (amdgpu_device_should_recover_gpu(adev))
-		amdgpu_device_gpu_recover(adev, NULL);
+		amdgpu_reset_domain_schedule(adev->reset_domain,
+					     &adev->kfd.reset_work);
 }
 
-int amdgpu_amdkfd_alloc_gtt_mem(struct kgd_dev *kgd, size_t size,
+int amdgpu_amdkfd_alloc_gtt_mem(struct amdgpu_device *adev, size_t size,
 				void **mem_obj, uint64_t *gpu_addr,
 				void **cpu_ptr, bool cp_mqd_gfx9)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
 	struct amdgpu_bo *bo = NULL;
 	struct amdgpu_bo_param bp;
 	int r;
@@ -314,21 +388,23 @@ allocate_mem_reserve_bo_failed:
 	return r;
 }
 
-void amdgpu_amdkfd_free_gtt_mem(struct kgd_dev *kgd, void *mem_obj)
+void amdgpu_amdkfd_free_gtt_mem(struct amdgpu_device *adev, void **mem_obj)
 {
-	struct amdgpu_bo *bo = (struct amdgpu_bo *) mem_obj;
+	struct amdgpu_bo **bo = (struct amdgpu_bo **) mem_obj;
 
-	amdgpu_bo_reserve(bo, true);
-	amdgpu_bo_kunmap(bo);
-	amdgpu_bo_unpin(bo);
-	amdgpu_bo_unreserve(bo);
-	amdgpu_bo_unref(&(bo));
+	if (!bo || !*bo)
+		return;
+
+	(void)amdgpu_bo_reserve(*bo, true);
+	amdgpu_bo_kunmap(*bo);
+	amdgpu_bo_unpin(*bo);
+	amdgpu_bo_unreserve(*bo);
+	amdgpu_bo_unref(bo);
 }
 
-int amdgpu_amdkfd_alloc_gws(struct kgd_dev *kgd, size_t size,
+int amdgpu_amdkfd_alloc_gws(struct amdgpu_device *adev, size_t size,
 				void **mem_obj)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
 	struct amdgpu_bo *bo = NULL;
 	struct amdgpu_bo_user *ubo;
 	struct amdgpu_bo_param bp;
@@ -355,18 +431,16 @@ int amdgpu_amdkfd_alloc_gws(struct kgd_dev *kgd, size_t size,
 	return 0;
 }
 
-void amdgpu_amdkfd_free_gws(struct kgd_dev *kgd, void *mem_obj)
+void amdgpu_amdkfd_free_gws(struct amdgpu_device *adev, void *mem_obj)
 {
 	struct amdgpu_bo *bo = (struct amdgpu_bo *)mem_obj;
 
 	amdgpu_bo_unref(&bo);
 }
 
-uint32_t amdgpu_amdkfd_get_fw_version(struct kgd_dev *kgd,
+uint32_t amdgpu_amdkfd_get_fw_version(struct amdgpu_device *adev,
 				      enum kgd_engine_type type)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
-
 	switch (type) {
 	case KGD_ENGINE_PFP:
 		return adev->gfx.pfp_fw_version;
@@ -399,17 +473,27 @@ uint32_t amdgpu_amdkfd_get_fw_version(struct kgd_dev *kgd,
 	return 0;
 }
 
-void amdgpu_amdkfd_get_local_mem_info(struct kgd_dev *kgd,
-				      struct kfd_local_mem_info *mem_info)
+void amdgpu_amdkfd_get_local_mem_info(struct amdgpu_device *adev,
+				      struct kfd_local_mem_info *mem_info,
+				      struct amdgpu_xcp *xcp)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
-
 	memset(mem_info, 0, sizeof(*mem_info));
 
-	mem_info->local_mem_size_public = adev->gmc.visible_vram_size;
-	mem_info->local_mem_size_private = adev->gmc.real_vram_size -
+	if (xcp) {
+		if (adev->gmc.real_vram_size == adev->gmc.visible_vram_size)
+			mem_info->local_mem_size_public =
+					KFD_XCP_MEMORY_SIZE(adev, xcp->id);
+		else
+			mem_info->local_mem_size_private =
+					KFD_XCP_MEMORY_SIZE(adev, xcp->id);
+	} else if (adev->apu_prefer_gtt) {
+		mem_info->local_mem_size_public = (ttm_tt_pages_limit() << PAGE_SHIFT);
+		mem_info->local_mem_size_private = 0;
+	} else {
+		mem_info->local_mem_size_public = adev->gmc.visible_vram_size;
+		mem_info->local_mem_size_private = adev->gmc.real_vram_size -
 						adev->gmc.visible_vram_size;
-
+	}
 	mem_info->vram_width = adev->gmc.vram_width;
 
 	pr_debug("Address base: %pap public 0x%llx private 0x%llx\n",
@@ -417,9 +501,7 @@ void amdgpu_amdkfd_get_local_mem_info(struct kgd_dev *kgd,
 			mem_info->local_mem_size_public,
 			mem_info->local_mem_size_private);
 
-	if (amdgpu_sriov_vf(adev))
-		mem_info->mem_clk_max = adev->clock.default_mclk / 100;
-	else if (adev->pm.dpm_enabled) {
+	if (adev->pm.dpm_enabled) {
 		if (amdgpu_emu_mode == 1)
 			mem_info->mem_clk_max = 0;
 		else
@@ -428,58 +510,28 @@ void amdgpu_amdkfd_get_local_mem_info(struct kgd_dev *kgd,
 		mem_info->mem_clk_max = 100;
 }
 
-uint64_t amdgpu_amdkfd_get_gpu_clock_counter(struct kgd_dev *kgd)
+uint64_t amdgpu_amdkfd_get_gpu_clock_counter(struct amdgpu_device *adev)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
-
 	if (adev->gfx.funcs->get_gpu_clock_counter)
 		return adev->gfx.funcs->get_gpu_clock_counter(adev);
 	return 0;
 }
 
-uint32_t amdgpu_amdkfd_get_max_engine_clock_in_mhz(struct kgd_dev *kgd)
+uint32_t amdgpu_amdkfd_get_max_engine_clock_in_mhz(struct amdgpu_device *adev)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
-
 	/* the sclk is in quantas of 10kHz */
-	if (amdgpu_sriov_vf(adev))
-		return adev->clock.default_sclk / 100;
-	else if (adev->pm.dpm_enabled)
+	if (adev->pm.dpm_enabled)
 		return amdgpu_dpm_get_sclk(adev, false) / 100;
 	else
 		return 100;
 }
 
-void amdgpu_amdkfd_get_cu_info(struct kgd_dev *kgd, struct kfd_cu_info *cu_info)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
-	struct amdgpu_cu_info acu_info = adev->gfx.cu_info;
-
-	memset(cu_info, 0, sizeof(*cu_info));
-	if (sizeof(cu_info->cu_bitmap) != sizeof(acu_info.bitmap))
-		return;
-
-	cu_info->cu_active_number = acu_info.number;
-	cu_info->cu_ao_mask = acu_info.ao_cu_mask;
-	memcpy(&cu_info->cu_bitmap[0], &acu_info.bitmap[0],
-	       sizeof(acu_info.bitmap));
-	cu_info->num_shader_engines = adev->gfx.config.max_shader_engines;
-	cu_info->num_shader_arrays_per_engine = adev->gfx.config.max_sh_per_se;
-	cu_info->num_cu_per_sh = adev->gfx.config.max_cu_per_sh;
-	cu_info->simd_per_cu = acu_info.simd_per_cu;
-	cu_info->max_waves_per_simd = acu_info.max_waves_per_simd;
-	cu_info->wave_front_size = acu_info.wave_front_size;
-	cu_info->max_scratch_slots_per_cu = acu_info.max_scratch_slots_per_cu;
-	cu_info->lds_size = acu_info.lds_size;
-}
-
-int amdgpu_amdkfd_get_dmabuf_info(struct kgd_dev *kgd, int dma_buf_fd,
-				  struct kgd_dev **dma_buf_kgd,
+int amdgpu_amdkfd_get_dmabuf_info(struct amdgpu_device *adev, int dma_buf_fd,
+				  struct amdgpu_device **dmabuf_adev,
 				  uint64_t *bo_size, void *metadata_buffer,
 				  size_t buffer_size, uint32_t *metadata_size,
-				  uint32_t *flags)
+				  uint32_t *flags, int8_t *xcp_id)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
 	struct dma_buf *dma_buf;
 	struct drm_gem_object *obj;
 	struct amdgpu_bo *bo;
@@ -507,8 +559,8 @@ int amdgpu_amdkfd_get_dmabuf_info(struct kgd_dev *kgd, int dma_buf_fd,
 		goto out_put;
 
 	r = 0;
-	if (dma_buf_kgd)
-		*dma_buf_kgd = (struct kgd_dev *)adev;
+	if (dmabuf_adev)
+		*dmabuf_adev = adev;
 	if (bo_size)
 		*bo_size = amdgpu_bo_size(bo);
 	if (metadata_buffer)
@@ -522,76 +574,16 @@ int amdgpu_amdkfd_get_dmabuf_info(struct kgd_dev *kgd, int dma_buf_fd,
 		if (bo->flags & AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED)
 			*flags |= KFD_IOC_ALLOC_MEM_FLAGS_PUBLIC;
 	}
+	if (xcp_id)
+		*xcp_id = bo->xcp_id;
 
 out_put:
 	dma_buf_put(dma_buf);
 	return r;
 }
 
-uint64_t amdgpu_amdkfd_get_vram_usage(struct kgd_dev *kgd)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
-	struct ttm_resource_manager *vram_man = ttm_manager_type(&adev->mman.bdev, TTM_PL_VRAM);
-
-	return amdgpu_vram_mgr_usage(vram_man);
-}
-
-uint64_t amdgpu_amdkfd_get_hive_id(struct kgd_dev *kgd)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
-
-	return adev->gmc.xgmi.hive_id;
-}
-
-uint64_t amdgpu_amdkfd_get_unique_id(struct kgd_dev *kgd)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
-
-	return adev->unique_id;
-}
-
-uint8_t amdgpu_amdkfd_get_xgmi_hops_count(struct kgd_dev *dst, struct kgd_dev *src)
-{
-	struct amdgpu_device *peer_adev = (struct amdgpu_device *)src;
-	struct amdgpu_device *adev = (struct amdgpu_device *)dst;
-	int ret = amdgpu_xgmi_get_hops_count(adev, peer_adev);
-
-	if (ret < 0) {
-		DRM_ERROR("amdgpu: failed to get  xgmi hops count between node %d and %d. ret = %d\n",
-			adev->gmc.xgmi.physical_node_id,
-			peer_adev->gmc.xgmi.physical_node_id, ret);
-		ret = 0;
-	}
-	return  (uint8_t)ret;
-}
-
-int amdgpu_amdkfd_get_xgmi_bandwidth_mbytes(struct kgd_dev *dst, struct kgd_dev *src, bool is_min)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)dst, *peer_adev;
-	int num_links;
-
-	if (adev->asic_type != CHIP_ALDEBARAN)
-		return 0;
-
-	if (src)
-		peer_adev = (struct amdgpu_device *)src;
-
-	/* num links returns 0 for indirect peers since indirect route is unknown. */
-	num_links = is_min ? 1 : amdgpu_xgmi_get_num_links(adev, peer_adev);
-	if (num_links < 0) {
-		DRM_ERROR("amdgpu: failed to get xgmi num links between node %d and %d. ret = %d\n",
-			adev->gmc.xgmi.physical_node_id,
-			peer_adev->gmc.xgmi.physical_node_id, num_links);
-		num_links = 0;
-	}
-
-	/* Aldebaran xGMI DPM is defeatured so assume x16 x 25Gbps for bandwidth. */
-	return (num_links * 16 * 25000)/BITS_PER_BYTE;
-}
-
-int amdgpu_amdkfd_get_pcie_bandwidth_mbytes(struct kgd_dev *dev, bool is_min)
+int amdgpu_amdkfd_get_pcie_bandwidth_mbytes(struct amdgpu_device *adev, bool is_min)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)dev;
 	int num_lanes_shift = (is_min ? ffs(adev->pm.pcie_mlw_mask) :
 							fls(adev->pm.pcie_mlw_mask)) - 1;
 	int gen_speed_shift = (is_min ? ffs(adev->pm.pcie_gen_mask &
@@ -647,39 +639,11 @@ int amdgpu_amdkfd_get_pcie_bandwidth_mbytes(struct kgd_dev *dev, bool is_min)
 	return (num_lanes_factor * gen_speed_mbits_factor)/BITS_PER_BYTE;
 }
 
-uint64_t amdgpu_amdkfd_get_mmio_remap_phys_addr(struct kgd_dev *kgd)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
-
-	return adev->rmmio_remap.bus_addr;
-}
-
-uint32_t amdgpu_amdkfd_get_num_gws(struct kgd_dev *kgd)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
-
-	return adev->gds.gws_size;
-}
-
-uint32_t amdgpu_amdkfd_get_asic_rev_id(struct kgd_dev *kgd)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
-
-	return adev->rev_id;
-}
-
-int amdgpu_amdkfd_get_noretry(struct kgd_dev *kgd)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
-
-	return adev->gmc.noretry;
-}
-
-int amdgpu_amdkfd_submit_ib(struct kgd_dev *kgd, enum kgd_engine_type engine,
+int amdgpu_amdkfd_submit_ib(struct amdgpu_device *adev,
+				enum kgd_engine_type engine,
 				uint32_t vmid, uint64_t gpu_addr,
 				uint32_t *ib_cmd, uint32_t ib_len)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
 	struct amdgpu_job *job;
 	struct amdgpu_ib *ib;
 	struct amdgpu_ring *ring;
@@ -702,7 +666,7 @@ int amdgpu_amdkfd_submit_ib(struct kgd_dev *kgd, enum kgd_engine_type engine,
 		goto err;
 	}
 
-	ret = amdgpu_job_alloc(adev, 1, &job, NULL);
+	ret = amdgpu_job_alloc(adev, NULL, NULL, NULL, 1, &job, 0);
 	if (ret)
 		goto err;
 
@@ -714,6 +678,7 @@ int amdgpu_amdkfd_submit_ib(struct kgd_dev *kgd, enum kgd_engine_type engine,
 	ib->length_dw = ib_len;
 	/* This works for NO_HWS. TODO: need to handle without knowing VMID */
 	job->vmid = vmid;
+	job->num_ibs = 1;
 
 	ret = amdgpu_ib_schedule(ring, 1, ib, job, &f);
 
@@ -722,6 +687,8 @@ int amdgpu_amdkfd_submit_ib(struct kgd_dev *kgd, enum kgd_engine_type engine,
 		goto err_ib_sched;
 	}
 
+	/* Drop the initial kref_init count (see drm_sched_main as example) */
+	dma_fence_put(f);
 	ret = dma_fence_wait(f, false);
 
 err_ib_sched:
@@ -730,10 +697,23 @@ err:
 	return ret;
 }
 
-void amdgpu_amdkfd_set_compute_idle(struct kgd_dev *kgd, bool idle)
+void amdgpu_amdkfd_set_compute_idle(struct amdgpu_device *adev, bool idle)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
-
+	enum amd_powergating_state state = idle ? AMD_PG_STATE_GATE : AMD_PG_STATE_UNGATE;
+	if ((IP_VERSION_MAJ(amdgpu_ip_version(adev, GC_HWIP, 0)) == 11 &&
+	    ((adev->mes.kiq_version & AMDGPU_MES_VERSION_MASK) <= 64)) ||
+		(IP_VERSION_MAJ(amdgpu_ip_version(adev, GC_HWIP, 0)) == 12)) {
+		pr_debug("GFXOFF is %s\n", idle ? "enabled" : "disabled");
+		amdgpu_gfx_off_ctrl(adev, idle);
+	} else if ((IP_VERSION_MAJ(amdgpu_ip_version(adev, GC_HWIP, 0)) == 9) &&
+		(adev->flags & AMD_IS_APU)) {
+		/* Disable GFXOFF and PG. Temporary workaround
+		 * to fix some compute applications issue on GFX9.
+		 */
+		struct amdgpu_ip_block *gfx_block = amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_GFX);
+		if (gfx_block != NULL)
+			gfx_block->version->funcs->set_powergating_state((void *)gfx_block, state);
+	}
 	amdgpu_dpm_switch_power_profile(adev,
 					PP_SMC_POWER_PROFILE_COMPUTE,
 					!idle);
@@ -747,49 +727,186 @@ bool amdgpu_amdkfd_is_kfd_vmid(struct amdgpu_device *adev, u32 vmid)
 	return false;
 }
 
-int amdgpu_amdkfd_flush_gpu_tlb_vmid(struct kgd_dev *kgd, uint16_t vmid)
+bool amdgpu_amdkfd_have_atomics_support(struct amdgpu_device *adev)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
+	return adev->have_atomics_support;
+}
 
-	if (adev->family == AMDGPU_FAMILY_AI) {
-		int i;
+void amdgpu_amdkfd_debug_mem_fence(struct amdgpu_device *adev)
+{
+	amdgpu_device_flush_hdp(adev, NULL);
+}
+
+bool amdgpu_amdkfd_is_fed(struct amdgpu_device *adev)
+{
+	return amdgpu_ras_get_fed_status(adev);
+}
+
+void amdgpu_amdkfd_ras_pasid_poison_consumption_handler(struct amdgpu_device *adev,
+				enum amdgpu_ras_block block, uint16_t pasid,
+				pasid_notify pasid_fn, void *data, uint32_t reset)
+{
+	amdgpu_umc_pasid_poison_handler(adev, block, pasid, pasid_fn, data, reset);
+}
+
+void amdgpu_amdkfd_ras_poison_consumption_handler(struct amdgpu_device *adev,
+	enum amdgpu_ras_block block, uint32_t reset)
+{
+	amdgpu_umc_pasid_poison_handler(adev, block, 0, NULL, NULL, reset);
+}
+
+int amdgpu_amdkfd_send_close_event_drain_irq(struct amdgpu_device *adev,
+					uint32_t *payload)
+{
+	int ret;
+
+	/* Device or IH ring is not ready so bail. */
+	ret = amdgpu_ih_wait_on_checkpoint_process_ts(adev, &adev->irq.ih);
+	if (ret)
+		return ret;
+
+	/* Send payload to fence KFD interrupts */
+	amdgpu_amdkfd_interrupt(adev, payload);
+
+	return 0;
+}
+
+int amdgpu_amdkfd_check_and_lock_kfd(struct amdgpu_device *adev)
+{
+	return kgd2kfd_check_and_lock_kfd(adev->kfd.dev);
+}
 
-		for (i = 0; i < adev->num_vmhubs; i++)
-			amdgpu_gmc_flush_gpu_tlb(adev, vmid, i, 0);
+void amdgpu_amdkfd_unlock_kfd(struct amdgpu_device *adev)
+{
+	kgd2kfd_unlock_kfd(adev->kfd.dev);
+}
+
+
+u64 amdgpu_amdkfd_xcp_memory_size(struct amdgpu_device *adev, int xcp_id)
+{
+	s8 mem_id = KFD_XCP_MEM_ID(adev, xcp_id);
+	u64 tmp;
+
+	if (adev->gmc.num_mem_partitions && xcp_id >= 0 && mem_id >= 0) {
+		if (adev->gmc.is_app_apu && adev->gmc.num_mem_partitions == 1) {
+			/* In NPS1 mode, we should restrict the vram reporting
+			 * tied to the ttm_pages_limit which is 1/2 of the system
+			 * memory. For other partition modes, the HBM is uniformly
+			 * divided already per numa node reported. If user wants to
+			 * go beyond the default ttm limit and maximize the ROCm
+			 * allocations, they can go up to max ttm and sysmem limits.
+			 */
+
+			tmp = (ttm_tt_pages_limit() << PAGE_SHIFT) / num_online_nodes();
+		} else {
+			tmp = adev->gmc.mem_partitions[mem_id].size;
+		}
+		do_div(tmp, adev->xcp_mgr->num_xcp_per_mem_partition);
+		return ALIGN_DOWN(tmp, PAGE_SIZE);
+	} else if (adev->apu_prefer_gtt) {
+		return (ttm_tt_pages_limit() << PAGE_SHIFT);
 	} else {
-		amdgpu_gmc_flush_gpu_tlb(adev, vmid, AMDGPU_GFXHUB_0, 0);
+		return adev->gmc.real_vram_size;
 	}
+}
 
-	return 0;
+int amdgpu_amdkfd_unmap_hiq(struct amdgpu_device *adev, u32 doorbell_off,
+			    u32 inst)
+{
+	struct amdgpu_kiq *kiq = &adev->gfx.kiq[inst];
+	struct amdgpu_ring *kiq_ring = &kiq->ring;
+	struct amdgpu_ring_funcs *ring_funcs;
+	struct amdgpu_ring *ring;
+	int r = 0;
+
+	if (!kiq->pmf || !kiq->pmf->kiq_unmap_queues)
+		return -EINVAL;
+
+	if (!kiq_ring->sched.ready || amdgpu_in_reset(adev))
+		return 0;
+
+	ring_funcs = kzalloc(sizeof(*ring_funcs), GFP_KERNEL);
+	if (!ring_funcs)
+		return -ENOMEM;
+
+	ring = kzalloc(sizeof(*ring), GFP_KERNEL);
+	if (!ring) {
+		r = -ENOMEM;
+		goto free_ring_funcs;
+	}
+
+	ring_funcs->type = AMDGPU_RING_TYPE_COMPUTE;
+	ring->doorbell_index = doorbell_off;
+	ring->funcs = ring_funcs;
+
+	spin_lock(&kiq->ring_lock);
+
+	if (amdgpu_ring_alloc(kiq_ring, kiq->pmf->unmap_queues_size)) {
+		spin_unlock(&kiq->ring_lock);
+		r = -ENOMEM;
+		goto free_ring;
+	}
+
+	kiq->pmf->kiq_unmap_queues(kiq_ring, ring, RESET_QUEUES, 0, 0);
+
+	/* Submit unmap queue packet */
+	amdgpu_ring_commit(kiq_ring);
+	/*
+	 * Ring test will do a basic scratch register change check. Just run
+	 * this to ensure that unmap queues that is submitted before got
+	 * processed successfully before returning.
+	 */
+	r = amdgpu_ring_test_helper(kiq_ring);
+
+	spin_unlock(&kiq->ring_lock);
+
+free_ring:
+	kfree(ring);
+
+free_ring_funcs:
+	kfree(ring_funcs);
+
+	return r;
 }
 
-int amdgpu_amdkfd_flush_gpu_tlb_pasid(struct kgd_dev *kgd, uint16_t pasid,
-				      enum TLB_FLUSH_TYPE flush_type)
+/* Stop scheduling on KFD */
+int amdgpu_amdkfd_stop_sched(struct amdgpu_device *adev, uint32_t node_id)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
-	bool all_hub = false;
+	if (!adev->kfd.init_complete)
+		return 0;
 
-	if (adev->family == AMDGPU_FAMILY_AI)
-		all_hub = true;
+	return kgd2kfd_stop_sched(adev->kfd.dev, node_id);
+}
+
+/* Start scheduling on KFD */
+int amdgpu_amdkfd_start_sched(struct amdgpu_device *adev, uint32_t node_id)
+{
+	if (!adev->kfd.init_complete)
+		return 0;
 
-	return amdgpu_gmc_flush_gpu_tlb_pasid(adev, pasid, flush_type, all_hub);
+	return kgd2kfd_start_sched(adev->kfd.dev, node_id);
 }
 
-bool amdgpu_amdkfd_have_atomics_support(struct kgd_dev *kgd)
+/* check if there are KFD queues active */
+bool amdgpu_amdkfd_compute_active(struct amdgpu_device *adev, uint32_t node_id)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
+	if (!adev->kfd.init_complete)
+		return false;
 
-	return adev->have_atomics_support;
+	return kgd2kfd_compute_active(adev->kfd.dev, node_id);
 }
 
-void amdgpu_amdkfd_ras_poison_consumption_handler(struct kgd_dev *kgd)
+/* Config CGTT_SQ_CLK_CTRL */
+int amdgpu_amdkfd_config_sq_perfmon(struct amdgpu_device *adev, uint32_t xcp_id,
+	bool core_override_enable, bool reg_override_enable, bool perfmon_override_enable)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
-	struct ras_err_data err_data = {0, 0, 0, NULL};
+	int r;
 
-	/* CPU MCA will handle page retirement if connected_to_cpu is 1 */
-	if (!adev->gmc.xgmi.connected_to_cpu)
-		amdgpu_umc_process_ras_data_cb(adev, &err_data, NULL);
-	else
-		amdgpu_amdkfd_gpu_reset(kgd);
+	if (!adev->kfd.init_complete)
+		return 0;
+
+	r = psp_config_sq_perfmon(&adev->psp, xcp_id, core_override_enable,
+					reg_override_enable, perfmon_override_enable);
+
+	return r;
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h
index a15a4787c7ee..8bdfcde2029b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h
@@ -25,14 +25,18 @@
 #ifndef AMDGPU_AMDKFD_H_INCLUDED
 #define AMDGPU_AMDKFD_H_INCLUDED
 
+#include <linux/list.h>
 #include <linux/types.h>
 #include <linux/mm.h>
 #include <linux/kthread.h>
 #include <linux/workqueue.h>
+#include <linux/mmu_notifier.h>
+#include <linux/memremap.h>
 #include <kgd_kfd_interface.h>
-#include <drm/ttm/ttm_execbuf_util.h>
+#include <drm/drm_client.h>
 #include "amdgpu_sync.h"
 #include "amdgpu_vm.h"
+#include "amdgpu_xcp.h"
 
 extern uint64_t amdgpu_amdkfd_total_mem_size;
 
@@ -43,11 +47,14 @@ enum TLB_FLUSH_TYPE {
 };
 
 struct amdgpu_device;
+struct kfd_process_device;
+struct amdgpu_reset_context;
 
 enum kfd_mem_attachment_type {
 	KFD_MEM_ATT_SHARED,	/* Share kgd_mem->bo or another attachment's */
 	KFD_MEM_ATT_USERPTR,	/* SG bo to DMA map pages from a userptr bo */
 	KFD_MEM_ATT_DMABUF,	/* DMAbuf to DMA map TTM BOs */
+	KFD_MEM_ATT_SG		/* Tag to DMA map SG BOs */
 };
 
 struct kfd_mem_attachment {
@@ -64,21 +71,22 @@ struct kgd_mem {
 	struct mutex lock;
 	struct amdgpu_bo *bo;
 	struct dma_buf *dmabuf;
+	struct amdgpu_hmm_range *range;
 	struct list_head attachments;
 	/* protected by amdkfd_process_info.lock */
-	struct ttm_validate_buffer validate_list;
-	struct ttm_validate_buffer resv_list;
+	struct list_head validate_list;
 	uint32_t domain;
 	unsigned int mapped_to_gpu_memory;
 	uint64_t va;
 
 	uint32_t alloc_flags;
 
-	atomic_t invalid;
+	uint32_t invalid;
 	struct amdkfd_process_info *process_info;
 
 	struct amdgpu_sync sync;
 
+	uint32_t gem_handle;
 	bool aql_queue;
 	bool is_imported;
 };
@@ -94,8 +102,18 @@ struct amdgpu_amdkfd_fence {
 
 struct amdgpu_kfd_dev {
 	struct kfd_dev *dev;
-	uint64_t vram_used;
+	int64_t vram_used[MAX_XCP];
+	uint64_t vram_used_aligned[MAX_XCP];
 	bool init_complete;
+	struct work_struct reset_work;
+
+	/* Client for KFD BO GEM handle allocations */
+	struct drm_client_dev client;
+
+	/* HMM page migration MEMORY_DEVICE_PRIVATE mapping
+	 * Must be last --ends in a flexible-array member.
+	 */
+	struct dev_pagemap pgmap;
 };
 
 enum kgd_engine_type {
@@ -128,38 +146,42 @@ struct amdkfd_process_info {
 	struct amdgpu_amdkfd_fence *eviction_fence;
 
 	/* MMU-notifier related fields */
-	atomic_t evicted_bos;
+	struct mutex notifier_lock;
+	uint32_t evicted_bos;
 	struct delayed_work restore_userptr_work;
 	struct pid *pid;
+	bool block_mmu_notifications;
 };
 
 int amdgpu_amdkfd_init(void);
 void amdgpu_amdkfd_fini(void);
 
-void amdgpu_amdkfd_suspend(struct amdgpu_device *adev, bool run_pm);
-int amdgpu_amdkfd_resume_iommu(struct amdgpu_device *adev);
-int amdgpu_amdkfd_resume(struct amdgpu_device *adev, bool run_pm);
+void amdgpu_amdkfd_suspend(struct amdgpu_device *adev, bool suspend_proc);
+int amdgpu_amdkfd_resume(struct amdgpu_device *adev, bool resume_proc);
+void amdgpu_amdkfd_suspend_process(struct amdgpu_device *adev);
+int amdgpu_amdkfd_resume_process(struct amdgpu_device *adev);
 void amdgpu_amdkfd_interrupt(struct amdgpu_device *adev,
 			const void *ih_ring_entry);
 void amdgpu_amdkfd_device_probe(struct amdgpu_device *adev);
 void amdgpu_amdkfd_device_init(struct amdgpu_device *adev);
 void amdgpu_amdkfd_device_fini_sw(struct amdgpu_device *adev);
-int amdgpu_amdkfd_submit_ib(struct kgd_dev *kgd, enum kgd_engine_type engine,
+int amdgpu_amdkfd_check_and_lock_kfd(struct amdgpu_device *adev);
+void amdgpu_amdkfd_unlock_kfd(struct amdgpu_device *adev);
+int amdgpu_amdkfd_submit_ib(struct amdgpu_device *adev,
+				enum kgd_engine_type engine,
 				uint32_t vmid, uint64_t gpu_addr,
 				uint32_t *ib_cmd, uint32_t ib_len);
-void amdgpu_amdkfd_set_compute_idle(struct kgd_dev *kgd, bool idle);
-bool amdgpu_amdkfd_have_atomics_support(struct kgd_dev *kgd);
-int amdgpu_amdkfd_flush_gpu_tlb_vmid(struct kgd_dev *kgd, uint16_t vmid);
-int amdgpu_amdkfd_flush_gpu_tlb_pasid(struct kgd_dev *kgd, uint16_t pasid,
-				      enum TLB_FLUSH_TYPE flush_type);
+void amdgpu_amdkfd_set_compute_idle(struct amdgpu_device *adev, bool idle);
+bool amdgpu_amdkfd_have_atomics_support(struct amdgpu_device *adev);
 
 bool amdgpu_amdkfd_is_kfd_vmid(struct amdgpu_device *adev, u32 vmid);
 
-int amdgpu_amdkfd_pre_reset(struct amdgpu_device *adev);
+int amdgpu_amdkfd_pre_reset(struct amdgpu_device *adev,
+			    struct amdgpu_reset_context *reset_context);
 
 int amdgpu_amdkfd_post_reset(struct amdgpu_device *adev);
 
-void amdgpu_amdkfd_gpu_reset(struct kgd_dev *kgd);
+void amdgpu_amdkfd_gpu_reset(struct amdgpu_device *adev);
 
 int amdgpu_queue_mask_bit_to_set_resource_bit(struct amdgpu_device *adev,
 					int queue_bit);
@@ -167,11 +189,20 @@ int amdgpu_queue_mask_bit_to_set_resource_bit(struct amdgpu_device *adev,
 struct amdgpu_amdkfd_fence *amdgpu_amdkfd_fence_create(u64 context,
 				struct mm_struct *mm,
 				struct svm_range_bo *svm_bo);
+
+int amdgpu_amdkfd_drm_client_create(struct amdgpu_device *adev);
+#if defined(CONFIG_DEBUG_FS)
+int kfd_debugfs_kfd_mem_limits(struct seq_file *m, void *data);
+#endif
 #if IS_ENABLED(CONFIG_HSA_AMD)
 bool amdkfd_fence_check_mm(struct dma_fence *f, struct mm_struct *mm);
 struct amdgpu_amdkfd_fence *to_amdgpu_amdkfd_fence(struct dma_fence *f);
-int amdgpu_amdkfd_remove_fence_on_pt_pd_bos(struct amdgpu_bo *bo);
-int amdgpu_amdkfd_evict_userptr(struct kgd_mem *mem, struct mm_struct *mm);
+void amdgpu_amdkfd_remove_all_eviction_fences(struct amdgpu_bo *bo);
+int amdgpu_amdkfd_evict_userptr(struct mmu_interval_notifier *mni,
+				unsigned long cur_seq, struct kgd_mem *mem);
+int amdgpu_amdkfd_bo_validate_and_fence(struct amdgpu_bo *bo,
+					uint32_t domain,
+					struct dma_fence *fence);
 #else
 static inline
 bool amdkfd_fence_check_mm(struct dma_fence *f, struct mm_struct *mm)
@@ -186,49 +217,58 @@ struct amdgpu_amdkfd_fence *to_amdgpu_amdkfd_fence(struct dma_fence *f)
 }
 
 static inline
-int amdgpu_amdkfd_remove_fence_on_pt_pd_bos(struct amdgpu_bo *bo)
+void amdgpu_amdkfd_remove_all_eviction_fences(struct amdgpu_bo *bo)
 {
-	return 0;
 }
 
 static inline
-int amdgpu_amdkfd_evict_userptr(struct kgd_mem *mem, struct mm_struct *mm)
+int amdgpu_amdkfd_evict_userptr(struct mmu_interval_notifier *mni,
+				unsigned long cur_seq, struct kgd_mem *mem)
+{
+	return 0;
+}
+static inline
+int amdgpu_amdkfd_bo_validate_and_fence(struct amdgpu_bo *bo,
+					uint32_t domain,
+					struct dma_fence *fence)
 {
 	return 0;
 }
 #endif
 /* Shared API */
-int amdgpu_amdkfd_alloc_gtt_mem(struct kgd_dev *kgd, size_t size,
+int amdgpu_amdkfd_alloc_gtt_mem(struct amdgpu_device *adev, size_t size,
 				void **mem_obj, uint64_t *gpu_addr,
 				void **cpu_ptr, bool mqd_gfx9);
-void amdgpu_amdkfd_free_gtt_mem(struct kgd_dev *kgd, void *mem_obj);
-int amdgpu_amdkfd_alloc_gws(struct kgd_dev *kgd, size_t size, void **mem_obj);
-void amdgpu_amdkfd_free_gws(struct kgd_dev *kgd, void *mem_obj);
+void amdgpu_amdkfd_free_gtt_mem(struct amdgpu_device *adev, void **mem_obj);
+int amdgpu_amdkfd_alloc_gws(struct amdgpu_device *adev, size_t size,
+				void **mem_obj);
+void amdgpu_amdkfd_free_gws(struct amdgpu_device *adev, void *mem_obj);
 int amdgpu_amdkfd_add_gws_to_process(void *info, void *gws, struct kgd_mem **mem);
 int amdgpu_amdkfd_remove_gws_from_process(void *info, void *mem);
-uint32_t amdgpu_amdkfd_get_fw_version(struct kgd_dev *kgd,
+uint32_t amdgpu_amdkfd_get_fw_version(struct amdgpu_device *adev,
 				      enum kgd_engine_type type);
-void amdgpu_amdkfd_get_local_mem_info(struct kgd_dev *kgd,
-				      struct kfd_local_mem_info *mem_info);
-uint64_t amdgpu_amdkfd_get_gpu_clock_counter(struct kgd_dev *kgd);
-
-uint32_t amdgpu_amdkfd_get_max_engine_clock_in_mhz(struct kgd_dev *kgd);
-void amdgpu_amdkfd_get_cu_info(struct kgd_dev *kgd, struct kfd_cu_info *cu_info);
-int amdgpu_amdkfd_get_dmabuf_info(struct kgd_dev *kgd, int dma_buf_fd,
-				  struct kgd_dev **dmabuf_kgd,
+void amdgpu_amdkfd_get_local_mem_info(struct amdgpu_device *adev,
+				      struct kfd_local_mem_info *mem_info,
+				      struct amdgpu_xcp *xcp);
+uint64_t amdgpu_amdkfd_get_gpu_clock_counter(struct amdgpu_device *adev);
+
+uint32_t amdgpu_amdkfd_get_max_engine_clock_in_mhz(struct amdgpu_device *adev);
+int amdgpu_amdkfd_get_dmabuf_info(struct amdgpu_device *adev, int dma_buf_fd,
+				  struct amdgpu_device **dmabuf_adev,
 				  uint64_t *bo_size, void *metadata_buffer,
 				  size_t buffer_size, uint32_t *metadata_size,
-				  uint32_t *flags);
-uint64_t amdgpu_amdkfd_get_vram_usage(struct kgd_dev *kgd);
-uint64_t amdgpu_amdkfd_get_hive_id(struct kgd_dev *kgd);
-uint64_t amdgpu_amdkfd_get_unique_id(struct kgd_dev *kgd);
-uint64_t amdgpu_amdkfd_get_mmio_remap_phys_addr(struct kgd_dev *kgd);
-uint32_t amdgpu_amdkfd_get_num_gws(struct kgd_dev *kgd);
-uint32_t amdgpu_amdkfd_get_asic_rev_id(struct kgd_dev *kgd);
-int amdgpu_amdkfd_get_noretry(struct kgd_dev *kgd);
-uint8_t amdgpu_amdkfd_get_xgmi_hops_count(struct kgd_dev *dst, struct kgd_dev *src);
-int amdgpu_amdkfd_get_xgmi_bandwidth_mbytes(struct kgd_dev *dst, struct kgd_dev *src, bool is_min);
-int amdgpu_amdkfd_get_pcie_bandwidth_mbytes(struct kgd_dev *dev, bool is_min);
+				  uint32_t *flags, int8_t *xcp_id);
+int amdgpu_amdkfd_get_pcie_bandwidth_mbytes(struct amdgpu_device *adev, bool is_min);
+int amdgpu_amdkfd_send_close_event_drain_irq(struct amdgpu_device *adev,
+					uint32_t *payload);
+int amdgpu_amdkfd_unmap_hiq(struct amdgpu_device *adev, u32 doorbell_off,
+				u32 inst);
+int amdgpu_amdkfd_start_sched(struct amdgpu_device *adev, uint32_t node_id);
+int amdgpu_amdkfd_stop_sched(struct amdgpu_device *adev, uint32_t node_id);
+int amdgpu_amdkfd_config_sq_perfmon(struct amdgpu_device *adev, uint32_t xcp_id,
+	bool core_override_enable, bool reg_override_enable, bool perfmon_override_enable);
+bool amdgpu_amdkfd_compute_active(struct amdgpu_device *adev, uint32_t node_id);
+
 
 /* Read user wptr from a specified user address space with page fault
  * disabled. The memory must be pinned and mapped to the hardware when
@@ -258,45 +298,81 @@ int amdgpu_amdkfd_get_pcie_bandwidth_mbytes(struct kgd_dev *dev, bool is_min);
 	(&((struct amdgpu_fpriv *)					\
 		((struct drm_file *)(drm_priv))->driver_priv)->vm)
 
-int amdgpu_amdkfd_gpuvm_acquire_process_vm(struct kgd_dev *kgd,
-					struct file *filp, u32 pasid,
+int amdgpu_amdkfd_gpuvm_acquire_process_vm(struct amdgpu_device *adev,
+					struct amdgpu_vm *avm,
 					void **process_info,
 					struct dma_fence **ef);
-void amdgpu_amdkfd_gpuvm_release_process_vm(struct kgd_dev *kgd, void *drm_priv);
 uint64_t amdgpu_amdkfd_gpuvm_get_process_page_dir(void *drm_priv);
+size_t amdgpu_amdkfd_get_available_memory(struct amdgpu_device *adev,
+					uint8_t xcp_id);
 int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu(
-		struct kgd_dev *kgd, uint64_t va, uint64_t size,
+		struct amdgpu_device *adev, uint64_t va, uint64_t size,
 		void *drm_priv, struct kgd_mem **mem,
-		uint64_t *offset, uint32_t flags);
+		uint64_t *offset, uint32_t flags, bool criu_resume);
 int amdgpu_amdkfd_gpuvm_free_memory_of_gpu(
-		struct kgd_dev *kgd, struct kgd_mem *mem, void *drm_priv,
+		struct amdgpu_device *adev, struct kgd_mem *mem, void *drm_priv,
 		uint64_t *size);
-int amdgpu_amdkfd_gpuvm_map_memory_to_gpu(
-		struct kgd_dev *kgd, struct kgd_mem *mem, void *drm_priv, bool *table_freed);
+int amdgpu_amdkfd_gpuvm_map_memory_to_gpu(struct amdgpu_device *adev,
+					  struct kgd_mem *mem, void *drm_priv);
 int amdgpu_amdkfd_gpuvm_unmap_memory_from_gpu(
-		struct kgd_dev *kgd, struct kgd_mem *mem, void *drm_priv);
+		struct amdgpu_device *adev, struct kgd_mem *mem, void *drm_priv);
+int amdgpu_amdkfd_gpuvm_dmaunmap_mem(struct kgd_mem *mem, void *drm_priv);
 int amdgpu_amdkfd_gpuvm_sync_memory(
-		struct kgd_dev *kgd, struct kgd_mem *mem, bool intr);
-int amdgpu_amdkfd_gpuvm_map_gtt_bo_to_kernel(struct kgd_dev *kgd,
-		struct kgd_mem *mem, void **kptr, uint64_t *size);
-void amdgpu_amdkfd_gpuvm_unmap_gtt_bo_from_kernel(struct kgd_dev *kgd, struct kgd_mem *mem);
+		struct amdgpu_device *adev, struct kgd_mem *mem, bool intr);
+int amdgpu_amdkfd_gpuvm_map_gtt_bo_to_kernel(struct kgd_mem *mem,
+					     void **kptr, uint64_t *size);
+void amdgpu_amdkfd_gpuvm_unmap_gtt_bo_from_kernel(struct kgd_mem *mem);
+
+int amdgpu_amdkfd_map_gtt_bo_to_gart(struct amdgpu_bo *bo, struct amdgpu_bo **bo_gart);
 
 int amdgpu_amdkfd_gpuvm_restore_process_bos(void *process_info,
-					    struct dma_fence **ef);
-int amdgpu_amdkfd_gpuvm_get_vm_fault_info(struct kgd_dev *kgd,
+					    struct dma_fence __rcu **ef);
+int amdgpu_amdkfd_gpuvm_get_vm_fault_info(struct amdgpu_device *adev,
 					      struct kfd_vm_fault_info *info);
-int amdgpu_amdkfd_gpuvm_import_dmabuf(struct kgd_dev *kgd,
-				      struct dma_buf *dmabuf,
-				      uint64_t va, void *drm_priv,
-				      struct kgd_mem **mem, uint64_t *size,
-				      uint64_t *mmap_offset);
-int amdgpu_amdkfd_get_tile_config(struct kgd_dev *kgd,
+int amdgpu_amdkfd_gpuvm_import_dmabuf_fd(struct amdgpu_device *adev, int fd,
+					 uint64_t va, void *drm_priv,
+					 struct kgd_mem **mem, uint64_t *size,
+					 uint64_t *mmap_offset);
+int amdgpu_amdkfd_gpuvm_export_dmabuf(struct kgd_mem *mem,
+				      struct dma_buf **dmabuf);
+void amdgpu_amdkfd_debug_mem_fence(struct amdgpu_device *adev);
+int amdgpu_amdkfd_get_tile_config(struct amdgpu_device *adev,
 				struct tile_config *config);
-void amdgpu_amdkfd_ras_poison_consumption_handler(struct kgd_dev *kgd);
+void amdgpu_amdkfd_ras_poison_consumption_handler(struct amdgpu_device *adev,
+			enum amdgpu_ras_block block, uint32_t reset);
+
+void amdgpu_amdkfd_ras_pasid_poison_consumption_handler(struct amdgpu_device *adev,
+			enum amdgpu_ras_block block, uint16_t pasid,
+			pasid_notify pasid_fn, void *data, uint32_t reset);
+
+bool amdgpu_amdkfd_is_fed(struct amdgpu_device *adev);
+bool amdgpu_amdkfd_bo_mapped_to_dev(void *drm_priv, struct kgd_mem *mem);
+void amdgpu_amdkfd_block_mmu_notifications(void *p);
+int amdgpu_amdkfd_criu_resume(void *p);
+int amdgpu_amdkfd_reserve_mem_limit(struct amdgpu_device *adev,
+		uint64_t size, u32 alloc_flag, int8_t xcp_id);
+void amdgpu_amdkfd_unreserve_mem_limit(struct amdgpu_device *adev,
+		uint64_t size, u32 alloc_flag, int8_t xcp_id);
+
+u64 amdgpu_amdkfd_xcp_memory_size(struct amdgpu_device *adev, int xcp_id);
+
+#define KFD_XCP_MEM_ID(adev, xcp_id) \
+		((adev)->xcp_mgr && (xcp_id) >= 0 ?\
+		(adev)->xcp_mgr->xcp[(xcp_id)].mem_id : -1)
+
+#define KFD_XCP_MEMORY_SIZE(adev, xcp_id) amdgpu_amdkfd_xcp_memory_size((adev), (xcp_id))
+
+
 #if IS_ENABLED(CONFIG_HSA_AMD)
 void amdgpu_amdkfd_gpuvm_init_mem_limits(void);
 void amdgpu_amdkfd_gpuvm_destroy_cb(struct amdgpu_device *adev,
 				struct amdgpu_vm *vm);
+
+/**
+ * @amdgpu_amdkfd_release_notify() - Notify KFD when GEM object is released
+ *
+ * Allows KFD to release its resources associated with the GEM object.
+ */
 void amdgpu_amdkfd_release_notify(struct amdgpu_bo *bo);
 void amdgpu_amdkfd_reserve_system_mem(uint64_t size);
 #else
@@ -316,27 +392,49 @@ void amdgpu_amdkfd_release_notify(struct amdgpu_bo *bo)
 {
 }
 #endif
+
+#if IS_ENABLED(CONFIG_HSA_AMD_SVM)
+int kgd2kfd_init_zone_device(struct amdgpu_device *adev);
+#else
+static inline
+int kgd2kfd_init_zone_device(struct amdgpu_device *adev)
+{
+	return 0;
+}
+#endif
+
 /* KGD2KFD callbacks */
-int kgd2kfd_quiesce_mm(struct mm_struct *mm);
+int kgd2kfd_quiesce_mm(struct mm_struct *mm, uint32_t trigger);
 int kgd2kfd_resume_mm(struct mm_struct *mm);
 int kgd2kfd_schedule_evict_and_restore_process(struct mm_struct *mm,
 						struct dma_fence *fence);
 #if IS_ENABLED(CONFIG_HSA_AMD)
 int kgd2kfd_init(void);
 void kgd2kfd_exit(void);
-struct kfd_dev *kgd2kfd_probe(struct kgd_dev *kgd, bool vf);
+struct kfd_dev *kgd2kfd_probe(struct amdgpu_device *adev, bool vf);
 bool kgd2kfd_device_init(struct kfd_dev *kfd,
-			 struct drm_device *ddev,
 			 const struct kgd2kfd_shared_resources *gpu_resources);
 void kgd2kfd_device_exit(struct kfd_dev *kfd);
-void kgd2kfd_suspend(struct kfd_dev *kfd, bool run_pm);
-int kgd2kfd_resume_iommu(struct kfd_dev *kfd);
-int kgd2kfd_resume(struct kfd_dev *kfd, bool run_pm);
-int kgd2kfd_pre_reset(struct kfd_dev *kfd);
+void kgd2kfd_suspend(struct kfd_dev *kfd, bool suspend_proc);
+int kgd2kfd_resume(struct kfd_dev *kfd, bool resume_proc);
+void kgd2kfd_suspend_process(struct kfd_dev *kfd);
+int kgd2kfd_resume_process(struct kfd_dev *kfd);
+int kgd2kfd_pre_reset(struct kfd_dev *kfd,
+		      struct amdgpu_reset_context *reset_context);
 int kgd2kfd_post_reset(struct kfd_dev *kfd);
 void kgd2kfd_interrupt(struct kfd_dev *kfd, const void *ih_ring_entry);
 void kgd2kfd_set_sram_ecc_flag(struct kfd_dev *kfd);
 void kgd2kfd_smi_event_throttle(struct kfd_dev *kfd, uint64_t throttle_bitmask);
+int kgd2kfd_check_and_lock_kfd(struct kfd_dev *kfd);
+void kgd2kfd_unlock_kfd(struct kfd_dev *kfd);
+int kgd2kfd_start_sched(struct kfd_dev *kfd, uint32_t node_id);
+int kgd2kfd_start_sched_all_nodes(struct kfd_dev *kfd);
+int kgd2kfd_stop_sched(struct kfd_dev *kfd, uint32_t node_id);
+int kgd2kfd_stop_sched_all_nodes(struct kfd_dev *kfd);
+bool kgd2kfd_compute_active(struct kfd_dev *kfd, uint32_t node_id);
+bool kgd2kfd_vmfault_fast_path(struct amdgpu_device *adev, struct amdgpu_iv_entry *entry,
+			       bool retry_fault);
+
 #else
 static inline int kgd2kfd_init(void)
 {
@@ -348,13 +446,13 @@ static inline void kgd2kfd_exit(void)
 }
 
 static inline
-struct kfd_dev *kgd2kfd_probe(struct kgd_dev *kgd, bool vf)
+struct kfd_dev *kgd2kfd_probe(struct amdgpu_device *adev, bool vf)
 {
 	return NULL;
 }
 
 static inline
-bool kgd2kfd_device_init(struct kfd_dev *kfd, struct drm_device *ddev,
+bool kgd2kfd_device_init(struct kfd_dev *kfd,
 				const struct kgd2kfd_shared_resources *gpu_resources)
 {
 	return false;
@@ -364,21 +462,26 @@ static inline void kgd2kfd_device_exit(struct kfd_dev *kfd)
 {
 }
 
-static inline void kgd2kfd_suspend(struct kfd_dev *kfd, bool run_pm)
+static inline void kgd2kfd_suspend(struct kfd_dev *kfd, bool suspend_proc)
 {
 }
 
-static int __maybe_unused kgd2kfd_resume_iommu(struct kfd_dev *kfd)
+static inline int kgd2kfd_resume(struct kfd_dev *kfd, bool resume_proc)
 {
 	return 0;
 }
 
-static inline int kgd2kfd_resume(struct kfd_dev *kfd, bool run_pm)
+static inline void kgd2kfd_suspend_process(struct kfd_dev *kfd)
+{
+}
+
+static inline int kgd2kfd_resume_process(struct kfd_dev *kfd)
 {
 	return 0;
 }
 
-static inline int kgd2kfd_pre_reset(struct kfd_dev *kfd)
+static inline int kgd2kfd_pre_reset(struct kfd_dev *kfd,
+				    struct amdgpu_reset_context *reset_context)
 {
 	return 0;
 }
@@ -402,5 +505,46 @@ static inline
 void kgd2kfd_smi_event_throttle(struct kfd_dev *kfd, uint64_t throttle_bitmask)
 {
 }
+
+static inline int kgd2kfd_check_and_lock_kfd(struct kfd_dev *kfd)
+{
+	return 0;
+}
+
+static inline void kgd2kfd_unlock_kfd(struct kfd_dev *kfd)
+{
+}
+
+static inline int kgd2kfd_start_sched(struct kfd_dev *kfd, uint32_t node_id)
+{
+	return 0;
+}
+
+static inline int kgd2kfd_start_sched_all_nodes(struct kfd_dev *kfd)
+{
+	return 0;
+}
+
+static inline int kgd2kfd_stop_sched(struct kfd_dev *kfd, uint32_t node_id)
+{
+	return 0;
+}
+
+static inline int kgd2kfd_stop_sched_all_nodes(struct kfd_dev *kfd)
+{
+	return 0;
+}
+
+static inline bool kgd2kfd_compute_active(struct kfd_dev *kfd, uint32_t node_id)
+{
+	return false;
+}
+
+static inline bool kgd2kfd_vmfault_fast_path(struct amdgpu_device *adev, struct amdgpu_iv_entry *entry,
+				      bool retry_fault)
+{
+	return false;
+}
+
 #endif
 #endif /* AMDGPU_AMDKFD_H_INCLUDED */
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_aldebaran.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_aldebaran.c
index 46cd4ee6bafb..7e9f7a280c1b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_aldebaran.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_aldebaran.c
@@ -23,6 +23,145 @@
 #include "amdgpu_amdkfd.h"
 #include "amdgpu_amdkfd_arcturus.h"
 #include "amdgpu_amdkfd_gfx_v9.h"
+#include "amdgpu_amdkfd_aldebaran.h"
+#include "gc/gc_9_4_2_offset.h"
+#include "gc/gc_9_4_2_sh_mask.h"
+#include <uapi/linux/kfd_ioctl.h>
+
+/*
+ * Returns TRAP_EN, EXCP_EN and EXCP_REPLACE.
+ *
+ * restore_dbg_registers is ignored here but is a general interface requirement
+ * for devices that support GFXOFF and where the RLC save/restore list
+ * does not support hw registers for debugging i.e. the driver has to manually
+ * initialize the debug mode registers after it has disabled GFX off during the
+ * debug session.
+ */
+uint32_t kgd_aldebaran_enable_debug_trap(struct amdgpu_device *adev,
+					    bool restore_dbg_registers,
+					    uint32_t vmid)
+{
+	uint32_t data = 0;
+
+	data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, TRAP_EN, 1);
+	data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, EXCP_EN, 0);
+	data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, EXCP_REPLACE, 0);
+
+	return data;
+}
+
+/* returns TRAP_EN, EXCP_EN and EXCP_REPLACE. */
+static uint32_t kgd_aldebaran_disable_debug_trap(struct amdgpu_device *adev,
+						bool keep_trap_enabled,
+						uint32_t vmid)
+{
+	uint32_t data = 0;
+
+	data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, TRAP_EN, keep_trap_enabled);
+	data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, EXCP_EN, 0);
+	data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, EXCP_REPLACE, 0);
+
+	return data;
+}
+
+static int kgd_aldebaran_validate_trap_override_request(struct amdgpu_device *adev,
+							uint32_t trap_override,
+							uint32_t *trap_mask_supported)
+{
+	*trap_mask_supported &= KFD_DBG_TRAP_MASK_FP_INVALID |
+				KFD_DBG_TRAP_MASK_FP_INPUT_DENORMAL |
+				KFD_DBG_TRAP_MASK_FP_DIVIDE_BY_ZERO |
+				KFD_DBG_TRAP_MASK_FP_OVERFLOW |
+				KFD_DBG_TRAP_MASK_FP_UNDERFLOW |
+				KFD_DBG_TRAP_MASK_FP_INEXACT |
+				KFD_DBG_TRAP_MASK_INT_DIVIDE_BY_ZERO |
+				KFD_DBG_TRAP_MASK_DBG_ADDRESS_WATCH |
+				KFD_DBG_TRAP_MASK_DBG_MEMORY_VIOLATION;
+
+	if (trap_override != KFD_DBG_TRAP_OVERRIDE_OR &&
+			trap_override != KFD_DBG_TRAP_OVERRIDE_REPLACE)
+		return -EPERM;
+
+	return 0;
+}
+
+/* returns TRAP_EN, EXCP_EN and EXCP_RPLACE. */
+static uint32_t kgd_aldebaran_set_wave_launch_trap_override(struct amdgpu_device *adev,
+					uint32_t vmid,
+					uint32_t trap_override,
+					uint32_t trap_mask_bits,
+					uint32_t trap_mask_request,
+					uint32_t *trap_mask_prev,
+					uint32_t kfd_dbg_trap_cntl_prev)
+
+{
+	uint32_t data = 0;
+
+	*trap_mask_prev = REG_GET_FIELD(kfd_dbg_trap_cntl_prev, SPI_GDBG_PER_VMID_CNTL, EXCP_EN);
+	trap_mask_bits = (trap_mask_bits & trap_mask_request) |
+		(*trap_mask_prev & ~trap_mask_request);
+
+	data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, TRAP_EN, 1);
+	data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, EXCP_EN, trap_mask_bits);
+	data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, EXCP_REPLACE, trap_override);
+
+	return data;
+}
+
+uint32_t kgd_aldebaran_set_wave_launch_mode(struct amdgpu_device *adev,
+					uint8_t wave_launch_mode,
+					uint32_t vmid)
+{
+	uint32_t data = 0;
+
+	data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, LAUNCH_MODE, wave_launch_mode);
+
+	return data;
+}
+
+#define TCP_WATCH_STRIDE (regTCP_WATCH1_ADDR_H - regTCP_WATCH0_ADDR_H)
+static uint32_t kgd_gfx_aldebaran_set_address_watch(
+					struct amdgpu_device *adev,
+					uint64_t watch_address,
+					uint32_t watch_address_mask,
+					uint32_t watch_id,
+					uint32_t watch_mode,
+					uint32_t debug_vmid,
+					uint32_t inst)
+{
+	uint32_t watch_address_high;
+	uint32_t watch_address_low;
+	uint32_t watch_address_cntl;
+
+	watch_address_cntl = 0;
+	watch_address_low = lower_32_bits(watch_address);
+	watch_address_high = upper_32_bits(watch_address) & 0xffff;
+
+	watch_address_cntl = REG_SET_FIELD(watch_address_cntl,
+			TCP_WATCH0_CNTL,
+			MODE,
+			watch_mode);
+
+	watch_address_cntl = REG_SET_FIELD(watch_address_cntl,
+			TCP_WATCH0_CNTL,
+			MASK,
+			watch_address_mask >> 6);
+
+	watch_address_cntl = REG_SET_FIELD(watch_address_cntl,
+			TCP_WATCH0_CNTL,
+			VALID,
+			1);
+
+	WREG32_RLC((SOC15_REG_OFFSET(GC, 0, regTCP_WATCH0_ADDR_H) +
+			(watch_id * TCP_WATCH_STRIDE)),
+			watch_address_high);
+
+	WREG32_RLC((SOC15_REG_OFFSET(GC, 0, regTCP_WATCH0_ADDR_L) +
+			(watch_id * TCP_WATCH_STRIDE)),
+			watch_address_low);
+
+	return watch_address_cntl;
+}
 
 const struct kfd2kgd_calls aldebaran_kfd2kgd = {
 	.program_sh_mem_settings = kgd_gfx_v9_program_sh_mem_settings,
@@ -37,12 +176,22 @@ const struct kfd2kgd_calls aldebaran_kfd2kgd = {
 	.hqd_sdma_is_occupied = kgd_arcturus_hqd_sdma_is_occupied,
 	.hqd_destroy = kgd_gfx_v9_hqd_destroy,
 	.hqd_sdma_destroy = kgd_arcturus_hqd_sdma_destroy,
-	.address_watch_disable = kgd_gfx_v9_address_watch_disable,
-	.address_watch_execute = kgd_gfx_v9_address_watch_execute,
 	.wave_control_execute = kgd_gfx_v9_wave_control_execute,
-	.address_watch_get_offset = kgd_gfx_v9_address_watch_get_offset,
 	.get_atc_vmid_pasid_mapping_info =
 				kgd_gfx_v9_get_atc_vmid_pasid_mapping_info,
 	.set_vm_context_page_table_base = kgd_gfx_v9_set_vm_context_page_table_base,
-	.program_trap_handler_settings = kgd_gfx_v9_program_trap_handler_settings
+	.get_cu_occupancy = kgd_gfx_v9_get_cu_occupancy,
+	.enable_debug_trap = kgd_aldebaran_enable_debug_trap,
+	.disable_debug_trap = kgd_aldebaran_disable_debug_trap,
+	.validate_trap_override_request = kgd_aldebaran_validate_trap_override_request,
+	.set_wave_launch_trap_override = kgd_aldebaran_set_wave_launch_trap_override,
+	.set_wave_launch_mode = kgd_aldebaran_set_wave_launch_mode,
+	.set_address_watch = kgd_gfx_aldebaran_set_address_watch,
+	.clear_address_watch = kgd_gfx_v9_clear_address_watch,
+	.get_iq_wait_times = kgd_gfx_v9_get_iq_wait_times,
+	.build_dequeue_wait_counts_packet_info = kgd_gfx_v9_build_dequeue_wait_counts_packet_info,
+	.program_trap_handler_settings = kgd_gfx_v9_program_trap_handler_settings,
+	.hqd_get_pq_addr = kgd_gfx_v9_hqd_get_pq_addr,
+	.hqd_reset = kgd_gfx_v9_hqd_reset,
+	.hqd_sdma_get_doorbell = kgd_gfx_v9_hqd_sdma_get_doorbell
 };
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_aldebaran.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_aldebaran.h
new file mode 100644
index 000000000000..a7bdaf8d82dd
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_aldebaran.h
@@ -0,0 +1,27 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+uint32_t kgd_aldebaran_enable_debug_trap(struct amdgpu_device *adev,
+					bool restore_dbg_registers,
+					uint32_t vmid);
+uint32_t kgd_aldebaran_set_wave_launch_mode(struct amdgpu_device *adev,
+					uint8_t wave_launch_mode,
+					uint32_t vmid);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.c
index 5a7f680bcb3f..1105a09e55dc 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.c
@@ -20,12 +20,12 @@
  * OTHER DEALINGS IN THE SOFTWARE.
  */
 #include <linux/module.h>
-#include <linux/fdtable.h>
 #include <linux/uaccess.h>
 #include <linux/firmware.h>
 #include "amdgpu.h"
 #include "amdgpu_amdkfd.h"
 #include "amdgpu_amdkfd_arcturus.h"
+#include "amdgpu_reset.h"
 #include "sdma0/sdma0_4_2_2_offset.h"
 #include "sdma0/sdma0_4_2_2_sh_mask.h"
 #include "sdma1/sdma1_4_2_2_offset.h"
@@ -48,6 +48,8 @@
 #include "amdgpu_amdkfd_gfx_v9.h"
 #include "gfxhub_v1_0.h"
 #include "mmhub_v9_4.h"
+#include "gc/gc_9_0_offset.h"
+#include "gc/gc_9_0_sh_mask.h"
 
 #define HQD_N_REGS 56
 #define DUMP_REG(addr) do {				\
@@ -57,11 +59,6 @@
 		(*dump)[i++][1] = RREG32(addr);		\
 	} while (0)
 
-static inline struct amdgpu_device *get_amdgpu_device(struct kgd_dev *kgd)
-{
-	return (struct amdgpu_device *)kgd;
-}
-
 static inline struct v9_sdma_mqd *get_sdma_mqd(void *mqd)
 {
 	return (struct v9_sdma_mqd *)mqd;
@@ -123,10 +120,9 @@ static uint32_t get_sdma_rlc_reg_offset(struct amdgpu_device *adev,
 	return sdma_rlc_reg_offset;
 }
 
-int kgd_arcturus_hqd_sdma_load(struct kgd_dev *kgd, void *mqd,
+int kgd_arcturus_hqd_sdma_load(struct amdgpu_device *adev, void *mqd,
 			     uint32_t __user *wptr, struct mm_struct *mm)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct v9_sdma_mqd *m;
 	uint32_t sdma_rlc_reg_offset;
 	unsigned long end_jiffies;
@@ -193,18 +189,17 @@ int kgd_arcturus_hqd_sdma_load(struct kgd_dev *kgd, void *mqd,
 	return 0;
 }
 
-int kgd_arcturus_hqd_sdma_dump(struct kgd_dev *kgd,
+int kgd_arcturus_hqd_sdma_dump(struct amdgpu_device *adev,
 			     uint32_t engine_id, uint32_t queue_id,
 			     uint32_t (**dump)[2], uint32_t *n_regs)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t sdma_rlc_reg_offset = get_sdma_rlc_reg_offset(adev,
 			engine_id, queue_id);
 	uint32_t i = 0, reg;
 #undef HQD_N_REGS
 #define HQD_N_REGS (19+6+7+10)
 
-	*dump = kmalloc_array(HQD_N_REGS * 2, sizeof(uint32_t), GFP_KERNEL);
+	*dump = kmalloc_array(HQD_N_REGS, sizeof(**dump), GFP_KERNEL);
 	if (*dump == NULL)
 		return -ENOMEM;
 
@@ -225,9 +220,9 @@ int kgd_arcturus_hqd_sdma_dump(struct kgd_dev *kgd,
 	return 0;
 }
 
-bool kgd_arcturus_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd)
+bool kgd_arcturus_hqd_sdma_is_occupied(struct amdgpu_device *adev,
+				void *mqd)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct v9_sdma_mqd *m;
 	uint32_t sdma_rlc_reg_offset;
 	uint32_t sdma_rlc_rb_cntl;
@@ -244,10 +239,9 @@ bool kgd_arcturus_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd)
 	return false;
 }
 
-int kgd_arcturus_hqd_sdma_destroy(struct kgd_dev *kgd, void *mqd,
+int kgd_arcturus_hqd_sdma_destroy(struct amdgpu_device *adev, void *mqd,
 				unsigned int utimeout)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct v9_sdma_mqd *m;
 	uint32_t sdma_rlc_reg_offset;
 	uint32_t temp;
@@ -284,6 +278,117 @@ int kgd_arcturus_hqd_sdma_destroy(struct kgd_dev *kgd, void *mqd,
 	return 0;
 }
 
+/*
+ * Helper used to suspend/resume gfx pipe for image post process work to set
+ * barrier behaviour.
+ */
+static int suspend_resume_compute_scheduler(struct amdgpu_device *adev, bool suspend)
+{
+	int i, r = 0;
+
+	for (i = 0; i < adev->gfx.num_compute_rings; i++) {
+		struct amdgpu_ring *ring = &adev->gfx.compute_ring[i];
+
+		if (!amdgpu_ring_sched_ready(ring))
+			continue;
+
+		/* stop secheduler and drain ring. */
+		if (suspend) {
+			drm_sched_stop(&ring->sched, NULL);
+			r = amdgpu_fence_wait_empty(ring);
+			if (r)
+				goto out;
+		} else {
+			drm_sched_start(&ring->sched, 0);
+		}
+	}
+
+out:
+	/* return on resume or failure to drain rings. */
+	if (!suspend || r)
+		return r;
+
+	return amdgpu_device_ip_wait_for_idle(adev, AMD_IP_BLOCK_TYPE_GFX);
+}
+
+static void set_barrier_auto_waitcnt(struct amdgpu_device *adev, bool enable_waitcnt)
+{
+	uint32_t data;
+
+	WRITE_ONCE(adev->barrier_has_auto_waitcnt, enable_waitcnt);
+
+	if (!down_read_trylock(&adev->reset_domain->sem))
+		return;
+
+	amdgpu_amdkfd_suspend(adev, true);
+
+	if (suspend_resume_compute_scheduler(adev, true))
+		goto out;
+
+	data = RREG32(SOC15_REG_OFFSET(GC, 0, mmSQ_CONFIG));
+	data = REG_SET_FIELD(data, SQ_CONFIG, DISABLE_BARRIER_WAITCNT,
+						!enable_waitcnt);
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmSQ_CONFIG), data);
+
+out:
+	suspend_resume_compute_scheduler(adev, false);
+
+	amdgpu_amdkfd_resume(adev, true);
+
+	up_read(&adev->reset_domain->sem);
+}
+
+/*
+ * restore_dbg_registers is ignored here but is a general interface requirement
+ * for devices that support GFXOFF and where the RLC save/restore list
+ * does not support hw registers for debugging i.e. the driver has to manually
+ * initialize the debug mode registers after it has disabled GFX off during the
+ * debug session.
+ */
+static uint32_t kgd_arcturus_enable_debug_trap(struct amdgpu_device *adev,
+				bool restore_dbg_registers,
+				uint32_t vmid)
+{
+	mutex_lock(&adev->grbm_idx_mutex);
+
+	kgd_gfx_v9_set_wave_launch_stall(adev, vmid, true);
+
+	set_barrier_auto_waitcnt(adev, true);
+
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_MASK), 0);
+
+	kgd_gfx_v9_set_wave_launch_stall(adev, vmid, false);
+
+	mutex_unlock(&adev->grbm_idx_mutex);
+
+	return 0;
+}
+
+/*
+ * keep_trap_enabled is ignored here but is a general interface requirement
+ * for devices that support multi-process debugging where the performance
+ * overhead from trap temporary setup needs to be bypassed when the debug
+ * session has ended.
+ */
+static uint32_t kgd_arcturus_disable_debug_trap(struct amdgpu_device *adev,
+					bool keep_trap_enabled,
+					uint32_t vmid)
+{
+
+	mutex_lock(&adev->grbm_idx_mutex);
+
+	kgd_gfx_v9_set_wave_launch_stall(adev, vmid, true);
+
+	set_barrier_auto_waitcnt(adev, false);
+
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_MASK), 0);
+
+	kgd_gfx_v9_set_wave_launch_stall(adev, vmid, false);
+
+	mutex_unlock(&adev->grbm_idx_mutex);
+
+	return 0;
+}
 const struct kfd2kgd_calls arcturus_kfd2kgd = {
 	.program_sh_mem_settings = kgd_gfx_v9_program_sh_mem_settings,
 	.set_pasid_vmid_mapping = kgd_gfx_v9_set_pasid_vmid_mapping,
@@ -297,14 +402,23 @@ const struct kfd2kgd_calls arcturus_kfd2kgd = {
 	.hqd_sdma_is_occupied = kgd_arcturus_hqd_sdma_is_occupied,
 	.hqd_destroy = kgd_gfx_v9_hqd_destroy,
 	.hqd_sdma_destroy = kgd_arcturus_hqd_sdma_destroy,
-	.address_watch_disable = kgd_gfx_v9_address_watch_disable,
-	.address_watch_execute = kgd_gfx_v9_address_watch_execute,
 	.wave_control_execute = kgd_gfx_v9_wave_control_execute,
-	.address_watch_get_offset = kgd_gfx_v9_address_watch_get_offset,
 	.get_atc_vmid_pasid_mapping_info =
 				kgd_gfx_v9_get_atc_vmid_pasid_mapping_info,
 	.set_vm_context_page_table_base =
 				kgd_gfx_v9_set_vm_context_page_table_base,
+	.enable_debug_trap = kgd_arcturus_enable_debug_trap,
+	.disable_debug_trap = kgd_arcturus_disable_debug_trap,
+	.validate_trap_override_request = kgd_gfx_v9_validate_trap_override_request,
+	.set_wave_launch_trap_override = kgd_gfx_v9_set_wave_launch_trap_override,
+	.set_wave_launch_mode = kgd_gfx_v9_set_wave_launch_mode,
+	.set_address_watch = kgd_gfx_v9_set_address_watch,
+	.clear_address_watch = kgd_gfx_v9_clear_address_watch,
+	.get_iq_wait_times = kgd_gfx_v9_get_iq_wait_times,
+	.build_dequeue_wait_counts_packet_info = kgd_gfx_v9_build_dequeue_wait_counts_packet_info,
 	.get_cu_occupancy = kgd_gfx_v9_get_cu_occupancy,
-	.program_trap_handler_settings = kgd_gfx_v9_program_trap_handler_settings
+	.program_trap_handler_settings = kgd_gfx_v9_program_trap_handler_settings,
+	.hqd_get_pq_addr = kgd_gfx_v9_hqd_get_pq_addr,
+	.hqd_reset = kgd_gfx_v9_hqd_reset,
+	.hqd_sdma_get_doorbell = kgd_gfx_v9_hqd_sdma_get_doorbell
 };
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.h
index ce08131b7b5f..756c1a5679c0 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.h
@@ -20,11 +20,12 @@
  * OTHER DEALINGS IN THE SOFTWARE.
  */
 
-int kgd_arcturus_hqd_sdma_load(struct kgd_dev *kgd, void *mqd,
+int kgd_arcturus_hqd_sdma_load(struct amdgpu_device *adev, void *mqd,
 			     uint32_t __user *wptr, struct mm_struct *mm);
-int kgd_arcturus_hqd_sdma_dump(struct kgd_dev *kgd,
+int kgd_arcturus_hqd_sdma_dump(struct amdgpu_device *adev,
 			     uint32_t engine_id, uint32_t queue_id,
 			     uint32_t (**dump)[2], uint32_t *n_regs);
-bool kgd_arcturus_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd);
-int kgd_arcturus_hqd_sdma_destroy(struct kgd_dev *kgd, void *mqd,
+bool kgd_arcturus_hqd_sdma_is_occupied(struct amdgpu_device *adev,
+				void *mqd);
+int kgd_arcturus_hqd_sdma_destroy(struct amdgpu_device *adev, void *mqd,
 				unsigned int utimeout);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_fence.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_fence.c
index 1d0dbff87d3f..1ef758ac5076 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_fence.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_fence.c
@@ -90,7 +90,7 @@ struct amdgpu_amdkfd_fence *to_amdgpu_amdkfd_fence(struct dma_fence *f)
 		return NULL;
 
 	fence = container_of(f, struct amdgpu_amdkfd_fence, base);
-	if (fence && f->ops == &amdkfd_fence_ops)
+	if (f->ops == &amdkfd_fence_ops)
 		return fence;
 
 	return NULL;
@@ -159,11 +159,14 @@ static void amdkfd_fence_release(struct dma_fence *f)
 }
 
 /**
- * amdkfd_fence_check_mm - Check if @mm is same as that of the fence @f
- *  if same return TRUE else return FALSE.
+ * amdkfd_fence_check_mm - Check whether to prevent eviction of @f by @mm
  *
  * @f: [IN] fence
  * @mm: [IN] mm that needs to be verified
+ *
+ * Check if @mm is same as that of the fence @f, if same return TRUE else
+ * return FALSE.
+ * For svm bo, which support vram overcommitment, always return FALSE.
  */
 bool amdkfd_fence_check_mm(struct dma_fence *f, struct mm_struct *mm)
 {
@@ -171,7 +174,7 @@ bool amdkfd_fence_check_mm(struct dma_fence *f, struct mm_struct *mm)
 
 	if (!fence)
 		return false;
-	else if (fence->mm == mm)
+	else if (fence->mm == mm  && !fence->svm_bo)
 		return true;
 
 	return false;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gc_9_4_3.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gc_9_4_3.c
new file mode 100644
index 000000000000..89a45a9218f3
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gc_9_4_3.c
@@ -0,0 +1,559 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+#include "amdgpu.h"
+#include "amdgpu_amdkfd.h"
+#include "amdgpu_amdkfd_gfx_v9.h"
+#include "amdgpu_amdkfd_aldebaran.h"
+#include "gc/gc_9_4_3_offset.h"
+#include "gc/gc_9_4_3_sh_mask.h"
+#include "athub/athub_1_8_0_offset.h"
+#include "athub/athub_1_8_0_sh_mask.h"
+#include "oss/osssys_4_4_2_offset.h"
+#include "oss/osssys_4_4_2_sh_mask.h"
+#include "v9_structs.h"
+#include "soc15.h"
+#include "sdma/sdma_4_4_2_offset.h"
+#include "sdma/sdma_4_4_2_sh_mask.h"
+#include <uapi/linux/kfd_ioctl.h>
+
+static inline struct v9_sdma_mqd *get_sdma_mqd(void *mqd)
+{
+	return (struct v9_sdma_mqd *)mqd;
+}
+
+static uint32_t get_sdma_rlc_reg_offset(struct amdgpu_device *adev,
+					unsigned int engine_id,
+					unsigned int queue_id)
+{
+	uint32_t sdma_engine_reg_base =
+		SOC15_REG_OFFSET(SDMA0, GET_INST(SDMA0, engine_id),
+				 regSDMA_RLC0_RB_CNTL) -
+		regSDMA_RLC0_RB_CNTL;
+	uint32_t retval = sdma_engine_reg_base +
+		  queue_id * (regSDMA_RLC1_RB_CNTL - regSDMA_RLC0_RB_CNTL);
+
+	pr_debug("RLC register offset for SDMA%d RLC%d: 0x%x\n", engine_id,
+							queue_id, retval);
+	return retval;
+}
+
+static int kgd_gfx_v9_4_3_hqd_sdma_load(struct amdgpu_device *adev, void *mqd,
+				 uint32_t __user *wptr, struct mm_struct *mm)
+{
+	struct v9_sdma_mqd *m;
+	uint32_t sdma_rlc_reg_offset;
+	unsigned long end_jiffies;
+	uint32_t data;
+	uint64_t data64;
+	uint64_t __user *wptr64 = (uint64_t __user *)wptr;
+
+	m = get_sdma_mqd(mqd);
+	sdma_rlc_reg_offset = get_sdma_rlc_reg_offset(adev, m->sdma_engine_id,
+							m->sdma_queue_id);
+
+	WREG32(sdma_rlc_reg_offset + regSDMA_RLC0_RB_CNTL,
+		m->sdmax_rlcx_rb_cntl & (~SDMA_RLC0_RB_CNTL__RB_ENABLE_MASK));
+
+	end_jiffies = msecs_to_jiffies(2000) + jiffies;
+	while (true) {
+		data = RREG32(sdma_rlc_reg_offset + regSDMA_RLC0_CONTEXT_STATUS);
+		if (data & SDMA_RLC0_CONTEXT_STATUS__IDLE_MASK)
+			break;
+		if (time_after(jiffies, end_jiffies)) {
+			pr_err("SDMA RLC not idle in %s\n", __func__);
+			return -ETIME;
+		}
+		usleep_range(500, 1000);
+	}
+
+	WREG32(sdma_rlc_reg_offset + regSDMA_RLC0_DOORBELL_OFFSET,
+		m->sdmax_rlcx_doorbell_offset);
+
+	data = REG_SET_FIELD(m->sdmax_rlcx_doorbell, SDMA_RLC0_DOORBELL,
+				ENABLE, 1);
+	WREG32(sdma_rlc_reg_offset + regSDMA_RLC0_DOORBELL, data);
+	WREG32(sdma_rlc_reg_offset + regSDMA_RLC0_RB_RPTR,
+					m->sdmax_rlcx_rb_rptr);
+	WREG32(sdma_rlc_reg_offset + regSDMA_RLC0_RB_RPTR_HI,
+					m->sdmax_rlcx_rb_rptr_hi);
+
+	WREG32(sdma_rlc_reg_offset + regSDMA_RLC0_MINOR_PTR_UPDATE, 1);
+	if (read_user_wptr(mm, wptr64, data64)) {
+		WREG32(sdma_rlc_reg_offset + regSDMA_RLC0_RB_WPTR,
+			lower_32_bits(data64));
+		WREG32(sdma_rlc_reg_offset + regSDMA_RLC0_RB_WPTR_HI,
+			upper_32_bits(data64));
+	} else {
+		WREG32(sdma_rlc_reg_offset + regSDMA_RLC0_RB_WPTR,
+			m->sdmax_rlcx_rb_rptr);
+		WREG32(sdma_rlc_reg_offset + regSDMA_RLC0_RB_WPTR_HI,
+			m->sdmax_rlcx_rb_rptr_hi);
+	}
+	WREG32(sdma_rlc_reg_offset + regSDMA_RLC0_MINOR_PTR_UPDATE, 0);
+
+	WREG32(sdma_rlc_reg_offset + regSDMA_RLC0_RB_BASE, m->sdmax_rlcx_rb_base);
+	WREG32(sdma_rlc_reg_offset + regSDMA_RLC0_RB_BASE_HI,
+			m->sdmax_rlcx_rb_base_hi);
+	WREG32(sdma_rlc_reg_offset + regSDMA_RLC0_RB_RPTR_ADDR_LO,
+			m->sdmax_rlcx_rb_rptr_addr_lo);
+	WREG32(sdma_rlc_reg_offset + regSDMA_RLC0_RB_RPTR_ADDR_HI,
+			m->sdmax_rlcx_rb_rptr_addr_hi);
+
+	data = REG_SET_FIELD(m->sdmax_rlcx_rb_cntl, SDMA_RLC0_RB_CNTL,
+				RB_ENABLE, 1);
+	WREG32(sdma_rlc_reg_offset + regSDMA_RLC0_RB_CNTL, data);
+
+	return 0;
+}
+
+static int kgd_gfx_v9_4_3_hqd_sdma_dump(struct amdgpu_device *adev,
+				 uint32_t engine_id, uint32_t queue_id,
+				 uint32_t (**dump)[2], uint32_t *n_regs)
+{
+	uint32_t sdma_rlc_reg_offset = get_sdma_rlc_reg_offset(adev,
+							engine_id, queue_id);
+	uint32_t i = 0, reg;
+#undef HQD_N_REGS
+#define HQD_N_REGS (19+6+7+12)
+#define DUMP_REG(addr) do {				\
+		if (WARN_ON_ONCE(i >= HQD_N_REGS))      \
+			break;				\
+		(*dump)[i][0] = (addr) << 2;            \
+		(*dump)[i++][1] = RREG32(addr);         \
+	} while (0)
+
+	*dump = kmalloc_array(HQD_N_REGS, sizeof(**dump), GFP_KERNEL);
+	if (*dump == NULL)
+		return -ENOMEM;
+
+	for (reg = regSDMA_RLC0_RB_CNTL; reg <= regSDMA_RLC0_DOORBELL; reg++)
+		DUMP_REG(sdma_rlc_reg_offset + reg);
+	for (reg = regSDMA_RLC0_STATUS; reg <= regSDMA_RLC0_CSA_ADDR_HI; reg++)
+		DUMP_REG(sdma_rlc_reg_offset + reg);
+	for (reg = regSDMA_RLC0_IB_SUB_REMAIN;
+	     reg <= regSDMA_RLC0_MINOR_PTR_UPDATE; reg++)
+		DUMP_REG(sdma_rlc_reg_offset + reg);
+	for (reg = regSDMA_RLC0_MIDCMD_DATA0;
+	     reg <= regSDMA_RLC0_MIDCMD_CNTL; reg++)
+		DUMP_REG(sdma_rlc_reg_offset + reg);
+
+	WARN_ON_ONCE(i != HQD_N_REGS);
+	*n_regs = i;
+
+	return 0;
+}
+
+static bool kgd_gfx_v9_4_3_hqd_sdma_is_occupied(struct amdgpu_device *adev, void *mqd)
+{
+	struct v9_sdma_mqd *m;
+	uint32_t sdma_rlc_reg_offset;
+	uint32_t sdma_rlc_rb_cntl;
+
+	m = get_sdma_mqd(mqd);
+	sdma_rlc_reg_offset = get_sdma_rlc_reg_offset(adev, m->sdma_engine_id,
+							m->sdma_queue_id);
+
+	sdma_rlc_rb_cntl = RREG32(sdma_rlc_reg_offset + regSDMA_RLC0_RB_CNTL);
+
+	if (sdma_rlc_rb_cntl & SDMA_RLC0_RB_CNTL__RB_ENABLE_MASK)
+		return true;
+
+	return false;
+}
+
+static int kgd_gfx_v9_4_3_hqd_sdma_destroy(struct amdgpu_device *adev, void *mqd,
+				    unsigned int utimeout)
+{
+	struct v9_sdma_mqd *m;
+	uint32_t sdma_rlc_reg_offset;
+	uint32_t temp;
+	unsigned long end_jiffies = (utimeout * HZ / 1000) + jiffies;
+
+	m = get_sdma_mqd(mqd);
+	sdma_rlc_reg_offset = get_sdma_rlc_reg_offset(adev, m->sdma_engine_id,
+							m->sdma_queue_id);
+
+	temp = RREG32(sdma_rlc_reg_offset + regSDMA_RLC0_RB_CNTL);
+	temp = temp & ~SDMA_RLC0_RB_CNTL__RB_ENABLE_MASK;
+	WREG32(sdma_rlc_reg_offset + regSDMA_RLC0_RB_CNTL, temp);
+
+	while (true) {
+		temp = RREG32(sdma_rlc_reg_offset + regSDMA_RLC0_CONTEXT_STATUS);
+		if (temp & SDMA_RLC0_CONTEXT_STATUS__IDLE_MASK)
+			break;
+		if (time_after(jiffies, end_jiffies)) {
+			pr_err("SDMA RLC not idle in %s\n", __func__);
+			return -ETIME;
+		}
+		usleep_range(500, 1000);
+	}
+
+	WREG32(sdma_rlc_reg_offset + regSDMA_RLC0_DOORBELL, 0);
+	WREG32(sdma_rlc_reg_offset + regSDMA_RLC0_RB_CNTL,
+		RREG32(sdma_rlc_reg_offset + regSDMA_RLC0_RB_CNTL) |
+		SDMA_RLC0_RB_CNTL__RB_ENABLE_MASK);
+
+	m->sdmax_rlcx_rb_rptr =
+			RREG32(sdma_rlc_reg_offset + regSDMA_RLC0_RB_RPTR);
+	m->sdmax_rlcx_rb_rptr_hi =
+			RREG32(sdma_rlc_reg_offset + regSDMA_RLC0_RB_RPTR_HI);
+
+	return 0;
+}
+
+static int kgd_gfx_v9_4_3_set_pasid_vmid_mapping(struct amdgpu_device *adev,
+			u32 pasid, unsigned int vmid, uint32_t xcc_inst)
+{
+	unsigned long timeout;
+	unsigned int reg;
+	unsigned int phy_inst = GET_INST(GC, xcc_inst);
+	/* Every two XCCs share one AID */
+	unsigned int aid = phy_inst / 2;
+
+	/*
+	 * We have to assume that there is no outstanding mapping.
+	 * The ATC_VMID_PASID_MAPPING_UPDATE_STATUS bit could be 0 because
+	 * a mapping is in progress or because a mapping finished
+	 * and the SW cleared it.
+	 * So the protocol is to always wait & clear.
+	 */
+	uint32_t pasid_mapping = (pasid == 0) ? 0 : (uint32_t)pasid |
+			ATC_VMID0_PASID_MAPPING__VALID_MASK;
+
+	WREG32(SOC15_REG_OFFSET(ATHUB, 0,
+		regATC_VMID0_PASID_MAPPING) + vmid, pasid_mapping);
+
+	timeout = jiffies + msecs_to_jiffies(10);
+	while (!(RREG32(SOC15_REG_OFFSET(ATHUB, 0,
+			regATC_VMID_PASID_MAPPING_UPDATE_STATUS)) &
+			(1U << vmid))) {
+		if (time_after(jiffies, timeout)) {
+			pr_err("Fail to program VMID-PASID mapping\n");
+			return -ETIME;
+		}
+		cpu_relax();
+	}
+
+	WREG32(SOC15_REG_OFFSET(ATHUB, 0,
+		regATC_VMID_PASID_MAPPING_UPDATE_STATUS),
+		1U << vmid);
+
+	reg = RREG32(SOC15_REG_OFFSET(OSSSYS, 0, regIH_VMID_LUT_INDEX));
+	/* Every 4 numbers is a cycle. 1st is AID, 2nd and 3rd are XCDs,
+	 * and the 4th is reserved. Therefore "aid * 4 + (xcc_inst % 2) + 1"
+	 * programs _LUT for XCC and "aid * 4" for AID where the XCC connects
+	 * to.
+	 */
+	WREG32(SOC15_REG_OFFSET(OSSSYS, 0, regIH_VMID_LUT_INDEX),
+		aid * 4 + (phy_inst % 2) + 1);
+	WREG32(SOC15_REG_OFFSET(OSSSYS, 0, regIH_VMID_0_LUT) + vmid,
+		pasid_mapping);
+	WREG32(SOC15_REG_OFFSET(OSSSYS, 0, regIH_VMID_LUT_INDEX),
+		aid * 4);
+	WREG32(SOC15_REG_OFFSET(OSSSYS, 0, regIH_VMID_0_LUT_MM) + vmid,
+		pasid_mapping);
+	WREG32(SOC15_REG_OFFSET(OSSSYS, 0, regIH_VMID_LUT_INDEX), reg);
+
+	return 0;
+}
+
+static inline struct v9_mqd *get_mqd(void *mqd)
+{
+	return (struct v9_mqd *)mqd;
+}
+
+static int kgd_gfx_v9_4_3_hqd_load(struct amdgpu_device *adev, void *mqd,
+			uint32_t pipe_id, uint32_t queue_id,
+			uint32_t __user *wptr, uint32_t wptr_shift,
+			uint32_t wptr_mask, struct mm_struct *mm, uint32_t inst)
+{
+	struct v9_mqd *m;
+	uint32_t *mqd_hqd;
+	uint32_t reg, hqd_base, hqd_end, data;
+
+	m = get_mqd(mqd);
+
+	kgd_gfx_v9_acquire_queue(adev, pipe_id, queue_id, inst);
+
+	/* HQD registers extend to CP_HQD_AQL_DISPATCH_ID_HI */
+	mqd_hqd = &m->cp_mqd_base_addr_lo;
+	hqd_base = SOC15_REG_OFFSET(GC, GET_INST(GC, inst), regCP_MQD_BASE_ADDR);
+	hqd_end = SOC15_REG_OFFSET(GC, GET_INST(GC, inst), regCP_HQD_AQL_DISPATCH_ID_HI);
+
+	for (reg = hqd_base; reg <= hqd_end; reg++)
+		WREG32_XCC(reg, mqd_hqd[reg - hqd_base], inst);
+
+
+	/* Activate doorbell logic before triggering WPTR poll. */
+	data = REG_SET_FIELD(m->cp_hqd_pq_doorbell_control,
+			     CP_HQD_PQ_DOORBELL_CONTROL, DOORBELL_EN, 1);
+	WREG32_SOC15_RLC(GC, GET_INST(GC, inst), regCP_HQD_PQ_DOORBELL_CONTROL, data);
+
+	if (wptr) {
+		/* Don't read wptr with get_user because the user
+		 * context may not be accessible (if this function
+		 * runs in a work queue). Instead trigger a one-shot
+		 * polling read from memory in the CP. This assumes
+		 * that wptr is GPU-accessible in the queue's VMID via
+		 * ATC or SVM. WPTR==RPTR before starting the poll so
+		 * the CP starts fetching new commands from the right
+		 * place.
+		 *
+		 * Guessing a 64-bit WPTR from a 32-bit RPTR is a bit
+		 * tricky. Assume that the queue didn't overflow. The
+		 * number of valid bits in the 32-bit RPTR depends on
+		 * the queue size. The remaining bits are taken from
+		 * the saved 64-bit WPTR. If the WPTR wrapped, add the
+		 * queue size.
+		 */
+		uint32_t queue_size =
+			2 << REG_GET_FIELD(m->cp_hqd_pq_control,
+					   CP_HQD_PQ_CONTROL, QUEUE_SIZE);
+		uint64_t guessed_wptr = m->cp_hqd_pq_rptr & (queue_size - 1);
+
+		if ((m->cp_hqd_pq_wptr_lo & (queue_size - 1)) < guessed_wptr)
+			guessed_wptr += queue_size;
+		guessed_wptr += m->cp_hqd_pq_wptr_lo & ~(queue_size - 1);
+		guessed_wptr += (uint64_t)m->cp_hqd_pq_wptr_hi << 32;
+
+		WREG32_SOC15_RLC(GC, GET_INST(GC, inst), regCP_HQD_PQ_WPTR_LO,
+			lower_32_bits(guessed_wptr));
+		WREG32_SOC15_RLC(GC, GET_INST(GC, inst), regCP_HQD_PQ_WPTR_HI,
+			upper_32_bits(guessed_wptr));
+		WREG32_SOC15_RLC(GC, GET_INST(GC, inst), regCP_HQD_PQ_WPTR_POLL_ADDR,
+			lower_32_bits((uintptr_t)wptr));
+		WREG32_SOC15_RLC(GC, GET_INST(GC, inst), regCP_HQD_PQ_WPTR_POLL_ADDR_HI,
+			upper_32_bits((uintptr_t)wptr));
+		WREG32_SOC15_RLC(GC, GET_INST(GC, inst), regCP_PQ_WPTR_POLL_CNTL1,
+			(uint32_t)kgd_gfx_v9_get_queue_mask(adev, pipe_id, queue_id));
+	}
+
+	/* Start the EOP fetcher */
+	WREG32_SOC15_RLC(GC, GET_INST(GC, inst), regCP_HQD_EOP_RPTR,
+	       REG_SET_FIELD(m->cp_hqd_eop_rptr, CP_HQD_EOP_RPTR, INIT_FETCHER, 1));
+
+	data = REG_SET_FIELD(m->cp_hqd_active, CP_HQD_ACTIVE, ACTIVE, 1);
+	WREG32_SOC15_RLC(GC, GET_INST(GC, inst), regCP_HQD_ACTIVE, data);
+
+	kgd_gfx_v9_release_queue(adev, inst);
+
+	return 0;
+}
+
+/* returns TRAP_EN, EXCP_EN and EXCP_REPLACE. */
+static uint32_t kgd_gfx_v9_4_3_disable_debug_trap(struct amdgpu_device *adev,
+						bool keep_trap_enabled,
+						uint32_t vmid)
+{
+	uint32_t data = 0;
+
+	data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, TRAP_EN, 1);
+	data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, EXCP_EN, 0);
+	data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, EXCP_REPLACE, 0);
+
+	return data;
+}
+
+static int kgd_gfx_v9_4_3_validate_trap_override_request(
+				struct amdgpu_device *adev,
+				uint32_t trap_override,
+				uint32_t *trap_mask_supported)
+{
+	*trap_mask_supported &= KFD_DBG_TRAP_MASK_FP_INVALID |
+				KFD_DBG_TRAP_MASK_FP_INPUT_DENORMAL |
+				KFD_DBG_TRAP_MASK_FP_DIVIDE_BY_ZERO |
+				KFD_DBG_TRAP_MASK_FP_OVERFLOW |
+				KFD_DBG_TRAP_MASK_FP_UNDERFLOW |
+				KFD_DBG_TRAP_MASK_FP_INEXACT |
+				KFD_DBG_TRAP_MASK_INT_DIVIDE_BY_ZERO |
+				KFD_DBG_TRAP_MASK_DBG_ADDRESS_WATCH |
+				KFD_DBG_TRAP_MASK_DBG_MEMORY_VIOLATION |
+				KFD_DBG_TRAP_MASK_TRAP_ON_WAVE_START |
+				KFD_DBG_TRAP_MASK_TRAP_ON_WAVE_END;
+
+	if (trap_override != KFD_DBG_TRAP_OVERRIDE_OR &&
+			trap_override != KFD_DBG_TRAP_OVERRIDE_REPLACE)
+		return -EPERM;
+
+	return 0;
+}
+
+static uint32_t trap_mask_map_sw_to_hw(uint32_t mask)
+{
+	uint32_t trap_on_start = (mask & KFD_DBG_TRAP_MASK_TRAP_ON_WAVE_START) ? 1 : 0;
+	uint32_t trap_on_end = (mask & KFD_DBG_TRAP_MASK_TRAP_ON_WAVE_END) ? 1 : 0;
+	uint32_t excp_en = mask & (KFD_DBG_TRAP_MASK_FP_INVALID |
+				KFD_DBG_TRAP_MASK_FP_INPUT_DENORMAL |
+				KFD_DBG_TRAP_MASK_FP_DIVIDE_BY_ZERO |
+				KFD_DBG_TRAP_MASK_FP_OVERFLOW |
+				KFD_DBG_TRAP_MASK_FP_UNDERFLOW |
+				KFD_DBG_TRAP_MASK_FP_INEXACT |
+				KFD_DBG_TRAP_MASK_INT_DIVIDE_BY_ZERO |
+				KFD_DBG_TRAP_MASK_DBG_ADDRESS_WATCH |
+				KFD_DBG_TRAP_MASK_DBG_MEMORY_VIOLATION);
+	uint32_t ret;
+
+	ret = REG_SET_FIELD(0, SPI_GDBG_PER_VMID_CNTL, EXCP_EN, excp_en);
+	ret = REG_SET_FIELD(ret, SPI_GDBG_PER_VMID_CNTL, TRAP_ON_START, trap_on_start);
+	ret = REG_SET_FIELD(ret, SPI_GDBG_PER_VMID_CNTL, TRAP_ON_END, trap_on_end);
+
+	return ret;
+}
+
+static uint32_t trap_mask_map_hw_to_sw(uint32_t mask)
+{
+	uint32_t ret = REG_GET_FIELD(mask, SPI_GDBG_PER_VMID_CNTL, EXCP_EN);
+
+	if (REG_GET_FIELD(mask, SPI_GDBG_PER_VMID_CNTL, TRAP_ON_START))
+		ret |= KFD_DBG_TRAP_MASK_TRAP_ON_WAVE_START;
+
+	if (REG_GET_FIELD(mask, SPI_GDBG_PER_VMID_CNTL, TRAP_ON_END))
+		ret |= KFD_DBG_TRAP_MASK_TRAP_ON_WAVE_END;
+
+	return ret;
+}
+
+/* returns TRAP_EN, EXCP_EN and EXCP_REPLACE. */
+static uint32_t kgd_gfx_v9_4_3_set_wave_launch_trap_override(
+				struct amdgpu_device *adev,
+				uint32_t vmid,
+				uint32_t trap_override,
+				uint32_t trap_mask_bits,
+				uint32_t trap_mask_request,
+				uint32_t *trap_mask_prev,
+				uint32_t kfd_dbg_trap_cntl_prev)
+
+{
+	uint32_t data = 0;
+
+	*trap_mask_prev = trap_mask_map_hw_to_sw(kfd_dbg_trap_cntl_prev);
+
+	data = (trap_mask_bits & trap_mask_request) |
+	       (*trap_mask_prev & ~trap_mask_request);
+	data = trap_mask_map_sw_to_hw(data);
+
+	data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, TRAP_EN, 1);
+	data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, EXCP_REPLACE, trap_override);
+
+	return data;
+}
+
+#define TCP_WATCH_STRIDE (regTCP_WATCH1_ADDR_H - regTCP_WATCH0_ADDR_H)
+static uint32_t kgd_gfx_v9_4_3_set_address_watch(
+				struct amdgpu_device *adev,
+				uint64_t watch_address,
+				uint32_t watch_address_mask,
+				uint32_t watch_id,
+				uint32_t watch_mode,
+				uint32_t debug_vmid,
+				uint32_t inst)
+{
+	uint32_t watch_address_high;
+	uint32_t watch_address_low;
+	uint32_t watch_address_cntl;
+
+	watch_address_cntl = 0;
+	watch_address_low = lower_32_bits(watch_address);
+	watch_address_high = upper_32_bits(watch_address) & 0xffff;
+
+	watch_address_cntl = REG_SET_FIELD(watch_address_cntl,
+			TCP_WATCH0_CNTL,
+			MODE,
+			watch_mode);
+
+	watch_address_cntl = REG_SET_FIELD(watch_address_cntl,
+			TCP_WATCH0_CNTL,
+			MASK,
+			watch_address_mask >> 7);
+
+	watch_address_cntl = REG_SET_FIELD(watch_address_cntl,
+			TCP_WATCH0_CNTL,
+			VALID,
+			1);
+
+	WREG32_XCC((SOC15_REG_OFFSET(GC, GET_INST(GC, inst),
+			regTCP_WATCH0_ADDR_H) +
+			(watch_id * TCP_WATCH_STRIDE)),
+			watch_address_high, inst);
+
+	WREG32_XCC((SOC15_REG_OFFSET(GC, GET_INST(GC, inst),
+			regTCP_WATCH0_ADDR_L) +
+			(watch_id * TCP_WATCH_STRIDE)),
+			watch_address_low, inst);
+
+	return watch_address_cntl;
+}
+
+static uint32_t kgd_gfx_v9_4_3_clear_address_watch(struct amdgpu_device *adev,
+				uint32_t watch_id)
+{
+	return 0;
+}
+
+static uint32_t kgd_gfx_v9_4_3_hqd_sdma_get_doorbell(struct amdgpu_device *adev,
+						     int engine, int queue)
+{
+	uint32_t reg_offset = get_sdma_rlc_reg_offset(adev, engine, queue);
+	uint32_t status = RREG32(regSDMA_RLC0_CONTEXT_STATUS + reg_offset);
+	uint32_t doorbell_off = RREG32(regSDMA_RLC0_DOORBELL_OFFSET + reg_offset);
+	bool is_active = !!REG_GET_FIELD(status, SDMA_RLC0_CONTEXT_STATUS, SELECTED);
+
+	return is_active ? doorbell_off >> 2 : 0;
+}
+
+const struct kfd2kgd_calls gc_9_4_3_kfd2kgd = {
+	.program_sh_mem_settings = kgd_gfx_v9_program_sh_mem_settings,
+	.set_pasid_vmid_mapping = kgd_gfx_v9_4_3_set_pasid_vmid_mapping,
+	.init_interrupts = kgd_gfx_v9_init_interrupts,
+	.hqd_load = kgd_gfx_v9_4_3_hqd_load,
+	.hiq_mqd_load = kgd_gfx_v9_hiq_mqd_load,
+	.hqd_sdma_load = kgd_gfx_v9_4_3_hqd_sdma_load,
+	.hqd_dump = kgd_gfx_v9_hqd_dump,
+	.hqd_sdma_dump = kgd_gfx_v9_4_3_hqd_sdma_dump,
+	.hqd_is_occupied = kgd_gfx_v9_hqd_is_occupied,
+	.hqd_sdma_is_occupied = kgd_gfx_v9_4_3_hqd_sdma_is_occupied,
+	.hqd_destroy = kgd_gfx_v9_hqd_destroy,
+	.hqd_sdma_destroy = kgd_gfx_v9_4_3_hqd_sdma_destroy,
+	.wave_control_execute = kgd_gfx_v9_wave_control_execute,
+	.get_atc_vmid_pasid_mapping_info =
+				kgd_gfx_v9_get_atc_vmid_pasid_mapping_info,
+	.set_vm_context_page_table_base =
+				kgd_gfx_v9_set_vm_context_page_table_base,
+	.get_cu_occupancy = kgd_gfx_v9_get_cu_occupancy,
+	.program_trap_handler_settings =
+				kgd_gfx_v9_program_trap_handler_settings,
+	.build_dequeue_wait_counts_packet_info =
+				kgd_gfx_v9_build_dequeue_wait_counts_packet_info,
+	.get_iq_wait_times = kgd_gfx_v9_get_iq_wait_times,
+	.enable_debug_trap = kgd_aldebaran_enable_debug_trap,
+	.disable_debug_trap = kgd_gfx_v9_4_3_disable_debug_trap,
+	.validate_trap_override_request =
+			kgd_gfx_v9_4_3_validate_trap_override_request,
+	.set_wave_launch_trap_override =
+			kgd_gfx_v9_4_3_set_wave_launch_trap_override,
+	.set_wave_launch_mode = kgd_aldebaran_set_wave_launch_mode,
+	.set_address_watch = kgd_gfx_v9_4_3_set_address_watch,
+	.clear_address_watch = kgd_gfx_v9_4_3_clear_address_watch,
+	.hqd_get_pq_addr = kgd_gfx_v9_hqd_get_pq_addr,
+	.hqd_reset = kgd_gfx_v9_hqd_reset,
+	.hqd_sdma_get_doorbell = kgd_gfx_v9_4_3_hqd_sdma_get_doorbell
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10.c
index 960acf68150a..0239114fb6c4 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10.c
@@ -21,6 +21,7 @@
  */
 #include "amdgpu.h"
 #include "amdgpu_amdkfd.h"
+#include "amdgpu_amdkfd_gfx_v10.h"
 #include "gc/gc_10_1_0_offset.h"
 #include "gc/gc_10_1_0_sh_mask.h"
 #include "athub/athub_2_0_0_offset.h"
@@ -31,6 +32,7 @@
 #include "v10_structs.h"
 #include "nv.h"
 #include "nvd.h"
+#include <uapi/linux/kfd_ioctl.h>
 
 enum hqd_dequeue_request_type {
 	NO_ACTION = 0,
@@ -39,37 +41,26 @@ enum hqd_dequeue_request_type {
 	SAVE_WAVES
 };
 
-static inline struct amdgpu_device *get_amdgpu_device(struct kgd_dev *kgd)
-{
-	return (struct amdgpu_device *)kgd;
-}
-
-static void lock_srbm(struct kgd_dev *kgd, uint32_t mec, uint32_t pipe,
+static void lock_srbm(struct amdgpu_device *adev, uint32_t mec, uint32_t pipe,
 			uint32_t queue, uint32_t vmid)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	mutex_lock(&adev->srbm_mutex);
 	nv_grbm_select(adev, mec, pipe, queue, vmid);
 }
 
-static void unlock_srbm(struct kgd_dev *kgd)
+static void unlock_srbm(struct amdgpu_device *adev)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	nv_grbm_select(adev, 0, 0, 0, 0);
 	mutex_unlock(&adev->srbm_mutex);
 }
 
-static void acquire_queue(struct kgd_dev *kgd, uint32_t pipe_id,
+static void acquire_queue(struct amdgpu_device *adev, uint32_t pipe_id,
 				uint32_t queue_id)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	uint32_t mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
 	uint32_t pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);
 
-	lock_srbm(kgd, mec, pipe, queue_id, 0);
+	lock_srbm(adev, mec, pipe, queue_id, 0);
 }
 
 static uint64_t get_queue_mask(struct amdgpu_device *adev,
@@ -81,33 +72,29 @@ static uint64_t get_queue_mask(struct amdgpu_device *adev,
 	return 1ull << bit;
 }
 
-static void release_queue(struct kgd_dev *kgd)
+static void release_queue(struct amdgpu_device *adev)
 {
-	unlock_srbm(kgd);
+	unlock_srbm(adev);
 }
 
-static void kgd_program_sh_mem_settings(struct kgd_dev *kgd, uint32_t vmid,
+static void kgd_program_sh_mem_settings(struct amdgpu_device *adev, uint32_t vmid,
 					uint32_t sh_mem_config,
 					uint32_t sh_mem_ape1_base,
 					uint32_t sh_mem_ape1_limit,
-					uint32_t sh_mem_bases)
+					uint32_t sh_mem_bases, uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
-	lock_srbm(kgd, 0, 0, 0, vmid);
+	lock_srbm(adev, 0, 0, 0, vmid);
 
 	WREG32_SOC15(GC, 0, mmSH_MEM_CONFIG, sh_mem_config);
 	WREG32_SOC15(GC, 0, mmSH_MEM_BASES, sh_mem_bases);
 	/* APE1 no longer exists on GFX9 */
 
-	unlock_srbm(kgd);
+	unlock_srbm(adev);
 }
 
-static int kgd_set_pasid_vmid_mapping(struct kgd_dev *kgd, u32 pasid,
-					unsigned int vmid)
+static int kgd_set_pasid_vmid_mapping(struct amdgpu_device *adev, u32 pasid,
+					unsigned int vmid, uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	/*
 	 * We have to assume that there is no outstanding mapping.
 	 * The ATC_VMID_PASID_MAPPING_UPDATE_STATUS bit could be 0 because
@@ -150,22 +137,22 @@ static int kgd_set_pasid_vmid_mapping(struct kgd_dev *kgd, u32 pasid,
  * but still works
  */
 
-static int kgd_init_interrupts(struct kgd_dev *kgd, uint32_t pipe_id)
+static int kgd_init_interrupts(struct amdgpu_device *adev, uint32_t pipe_id,
+				uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t mec;
 	uint32_t pipe;
 
 	mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
 	pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);
 
-	lock_srbm(kgd, mec, pipe, 0, 0);
+	lock_srbm(adev, mec, pipe, 0, 0);
 
 	WREG32_SOC15(GC, 0, mmCPC_INT_CNTL,
 		CP_INT_CNTL_RING0__TIME_STAMP_INT_ENABLE_MASK |
 		CP_INT_CNTL_RING0__OPCODE_ERROR_INT_ENABLE_MASK);
 
-	unlock_srbm(kgd);
+	unlock_srbm(adev);
 
 	return 0;
 }
@@ -218,12 +205,11 @@ static inline struct v10_sdma_mqd *get_sdma_mqd(void *mqd)
 	return (struct v10_sdma_mqd *)mqd;
 }
 
-static int kgd_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
-			uint32_t queue_id, uint32_t __user *wptr,
-			uint32_t wptr_shift, uint32_t wptr_mask,
-			struct mm_struct *mm)
+static int kgd_hqd_load(struct amdgpu_device *adev, void *mqd,
+			uint32_t pipe_id, uint32_t queue_id,
+			uint32_t __user *wptr, uint32_t wptr_shift,
+			uint32_t wptr_mask, struct mm_struct *mm, uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct v10_compute_mqd *m;
 	uint32_t *mqd_hqd;
 	uint32_t reg, hqd_base, data;
@@ -231,7 +217,7 @@ static int kgd_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
 	m = get_mqd(mqd);
 
 	pr_debug("Load hqd of pipe %d queue %d\n", pipe_id, queue_id);
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 
 	/* HQD registers extend from CP_MQD_BASE_ADDR to CP_HQD_EOP_WPTR_MEM. */
 	mqd_hqd = &m->cp_mqd_base_addr_lo;
@@ -296,24 +282,23 @@ static int kgd_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
 	data = REG_SET_FIELD(m->cp_hqd_active, CP_HQD_ACTIVE, ACTIVE, 1);
 	WREG32_SOC15(GC, 0, mmCP_HQD_ACTIVE, data);
 
-	release_queue(kgd);
+	release_queue(adev);
 
 	return 0;
 }
 
-static int kgd_hiq_mqd_load(struct kgd_dev *kgd, void *mqd,
+static int kgd_hiq_mqd_load(struct amdgpu_device *adev, void *mqd,
 			    uint32_t pipe_id, uint32_t queue_id,
-			    uint32_t doorbell_off)
+			    uint32_t doorbell_off, uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-	struct amdgpu_ring *kiq_ring = &adev->gfx.kiq.ring;
+	struct amdgpu_ring *kiq_ring = &adev->gfx.kiq[0].ring;
 	struct v10_compute_mqd *m;
 	uint32_t mec, pipe;
 	int r;
 
 	m = get_mqd(mqd);
 
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 
 	mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
 	pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);
@@ -321,7 +306,7 @@ static int kgd_hiq_mqd_load(struct kgd_dev *kgd, void *mqd,
 	pr_debug("kfd: set HIQ, mec:%d, pipe:%d, queue:%d.\n",
 		 mec, pipe, queue_id);
 
-	spin_lock(&adev->gfx.kiq.ring_lock);
+	spin_lock(&adev->gfx.kiq[0].ring_lock);
 	r = amdgpu_ring_alloc(kiq_ring, 7);
 	if (r) {
 		pr_err("Failed to alloc KIQ (%d).\n", r);
@@ -348,17 +333,16 @@ static int kgd_hiq_mqd_load(struct kgd_dev *kgd, void *mqd,
 	amdgpu_ring_commit(kiq_ring);
 
 out_unlock:
-	spin_unlock(&adev->gfx.kiq.ring_lock);
-	release_queue(kgd);
+	spin_unlock(&adev->gfx.kiq[0].ring_lock);
+	release_queue(adev);
 
 	return r;
 }
 
-static int kgd_hqd_dump(struct kgd_dev *kgd,
+static int kgd_hqd_dump(struct amdgpu_device *adev,
 			uint32_t pipe_id, uint32_t queue_id,
-			uint32_t (**dump)[2], uint32_t *n_regs)
+			uint32_t (**dump)[2], uint32_t *n_regs, uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t i = 0, reg;
 #define HQD_N_REGS 56
 #define DUMP_REG(addr) do {				\
@@ -368,17 +352,17 @@ static int kgd_hqd_dump(struct kgd_dev *kgd,
 		(*dump)[i++][1] = RREG32_SOC15_IP(GC, addr);		\
 	} while (0)
 
-	*dump = kmalloc(HQD_N_REGS*2*sizeof(uint32_t), GFP_KERNEL);
+	*dump = kmalloc_array(HQD_N_REGS, sizeof(**dump), GFP_KERNEL);
 	if (*dump == NULL)
 		return -ENOMEM;
 
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 
 	for (reg = SOC15_REG_OFFSET(GC, 0, mmCP_MQD_BASE_ADDR);
 	     reg <= SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_HI); reg++)
 		DUMP_REG(reg);
 
-	release_queue(kgd);
+	release_queue(adev);
 
 	WARN_ON_ONCE(i != HQD_N_REGS);
 	*n_regs = i;
@@ -386,10 +370,9 @@ static int kgd_hqd_dump(struct kgd_dev *kgd,
 	return 0;
 }
 
-static int kgd_hqd_sdma_load(struct kgd_dev *kgd, void *mqd,
+static int kgd_hqd_sdma_load(struct amdgpu_device *adev, void *mqd,
 			     uint32_t __user *wptr, struct mm_struct *mm)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct v10_sdma_mqd *m;
 	uint32_t sdma_rlc_reg_offset;
 	unsigned long end_jiffies;
@@ -456,18 +439,17 @@ static int kgd_hqd_sdma_load(struct kgd_dev *kgd, void *mqd,
 	return 0;
 }
 
-static int kgd_hqd_sdma_dump(struct kgd_dev *kgd,
+static int kgd_hqd_sdma_dump(struct amdgpu_device *adev,
 			     uint32_t engine_id, uint32_t queue_id,
 			     uint32_t (**dump)[2], uint32_t *n_regs)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t sdma_rlc_reg_offset = get_sdma_rlc_reg_offset(adev,
 			engine_id, queue_id);
 	uint32_t i = 0, reg;
 #undef HQD_N_REGS
 #define HQD_N_REGS (19+6+7+10)
 
-	*dump = kmalloc(HQD_N_REGS*2*sizeof(uint32_t), GFP_KERNEL);
+	*dump = kmalloc_array(HQD_N_REGS, sizeof(**dump), GFP_KERNEL);
 	if (*dump == NULL)
 		return -ENOMEM;
 
@@ -488,15 +470,15 @@ static int kgd_hqd_sdma_dump(struct kgd_dev *kgd,
 	return 0;
 }
 
-static bool kgd_hqd_is_occupied(struct kgd_dev *kgd, uint64_t queue_address,
-				uint32_t pipe_id, uint32_t queue_id)
+static bool kgd_hqd_is_occupied(struct amdgpu_device *adev,
+				uint64_t queue_address, uint32_t pipe_id,
+				uint32_t queue_id, uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t act;
 	bool retval = false;
 	uint32_t low, high;
 
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 	act = RREG32_SOC15(GC, 0, mmCP_HQD_ACTIVE);
 	if (act) {
 		low = lower_32_bits(queue_address >> 8);
@@ -506,13 +488,12 @@ static bool kgd_hqd_is_occupied(struct kgd_dev *kgd, uint64_t queue_address,
 		   high == RREG32_SOC15(GC, 0, mmCP_HQD_PQ_BASE_HI))
 			retval = true;
 	}
-	release_queue(kgd);
+	release_queue(adev);
 	return retval;
 }
 
-static bool kgd_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd)
+static bool kgd_hqd_sdma_is_occupied(struct amdgpu_device *adev, void *mqd)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct v10_sdma_mqd *m;
 	uint32_t sdma_rlc_reg_offset;
 	uint32_t sdma_rlc_rb_cntl;
@@ -529,12 +510,11 @@ static bool kgd_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd)
 	return false;
 }
 
-static int kgd_hqd_destroy(struct kgd_dev *kgd, void *mqd,
+static int kgd_hqd_destroy(struct amdgpu_device *adev, void *mqd,
 				enum kfd_preempt_type reset_type,
 				unsigned int utimeout, uint32_t pipe_id,
-				uint32_t queue_id)
+				uint32_t queue_id, uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	enum hqd_dequeue_request_type type;
 	unsigned long end_jiffies;
 	uint32_t temp;
@@ -548,7 +528,7 @@ static int kgd_hqd_destroy(struct kgd_dev *kgd, void *mqd,
 	int retry;
 #endif
 
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 
 	if (m->cp_hqd_vmid == 0)
 		WREG32_FIELD15(GC, 0, RLC_CP_SCHEDULERS, scheduler1, 0);
@@ -633,20 +613,19 @@ loop:
 			break;
 		if (time_after(jiffies, end_jiffies)) {
 			pr_err("cp queue preemption time out.\n");
-			release_queue(kgd);
+			release_queue(adev);
 			return -ETIME;
 		}
 		usleep_range(500, 1000);
 	}
 
-	release_queue(kgd);
+	release_queue(adev);
 	return 0;
 }
 
-static int kgd_hqd_sdma_destroy(struct kgd_dev *kgd, void *mqd,
+static int kgd_hqd_sdma_destroy(struct amdgpu_device *adev, void *mqd,
 				unsigned int utimeout)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct v10_sdma_mqd *m;
 	uint32_t sdma_rlc_reg_offset;
 	uint32_t temp;
@@ -683,11 +662,10 @@ static int kgd_hqd_sdma_destroy(struct kgd_dev *kgd, void *mqd,
 	return 0;
 }
 
-static bool get_atc_vmid_pasid_mapping_info(struct kgd_dev *kgd,
+static bool get_atc_vmid_pasid_mapping_info(struct amdgpu_device *adev,
 					uint8_t vmid, uint16_t *p_pasid)
 {
 	uint32_t value;
-	struct amdgpu_device *adev = (struct amdgpu_device *) kgd;
 
 	value = RREG32(SOC15_REG_OFFSET(ATHUB, 0, mmATC_VMID0_PASID_MAPPING)
 		     + vmid);
@@ -696,25 +674,10 @@ static bool get_atc_vmid_pasid_mapping_info(struct kgd_dev *kgd,
 	return !!(value & ATC_VMID0_PASID_MAPPING__VALID_MASK);
 }
 
-static int kgd_address_watch_disable(struct kgd_dev *kgd)
-{
-	return 0;
-}
-
-static int kgd_address_watch_execute(struct kgd_dev *kgd,
-					unsigned int watch_point_id,
-					uint32_t cntl_val,
-					uint32_t addr_hi,
-					uint32_t addr_lo)
-{
-	return 0;
-}
-
-static int kgd_wave_control_execute(struct kgd_dev *kgd,
+static int kgd_wave_control_execute(struct amdgpu_device *adev,
 					uint32_t gfx_index_val,
-					uint32_t sq_cmd)
+					uint32_t sq_cmd, uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t data = 0;
 
 	mutex_lock(&adev->grbm_idx_mutex);
@@ -735,18 +698,9 @@ static int kgd_wave_control_execute(struct kgd_dev *kgd,
 	return 0;
 }
 
-static uint32_t kgd_address_watch_get_offset(struct kgd_dev *kgd,
-					unsigned int watch_point_id,
-					unsigned int reg_offset)
-{
-	return 0;
-}
-
-static void set_vm_context_page_table_base(struct kgd_dev *kgd, uint32_t vmid,
-		uint64_t page_table_base)
+static void set_vm_context_page_table_base(struct amdgpu_device *adev,
+		uint32_t vmid, uint64_t page_table_base)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	if (!amdgpu_amdkfd_is_kfd_vmid(adev, vmid)) {
 		pr_err("trying to set page table base for wrong VMID %u\n",
 		       vmid);
@@ -757,12 +711,344 @@ static void set_vm_context_page_table_base(struct kgd_dev *kgd, uint32_t vmid,
 	adev->gfxhub.funcs->setup_vm_pt_regs(adev, vmid, page_table_base);
 }
 
-static void program_trap_handler_settings(struct kgd_dev *kgd,
-		uint32_t vmid, uint64_t tba_addr, uint64_t tma_addr)
+/*
+ * GFX10 helper for wave launch stall requirements on debug trap setting.
+ *
+ * vmid:
+ *   Target VMID to stall/unstall.
+ *
+ * stall:
+ *   0-unstall wave launch (enable), 1-stall wave launch (disable).
+ *   After wavefront launch has been stalled, allocated waves must drain from
+ *   SPI in order for debug trap settings to take effect on those waves.
+ *   This is roughly a ~3500 clock cycle wait on SPI where a read on
+ *   SPI_GDBG_WAVE_CNTL translates to ~32 clock cycles.
+ *   KGD_GFX_V10_WAVE_LAUNCH_SPI_DRAIN_LATENCY indicates the number of reads required.
+ *
+ *   NOTE: We can afford to clear the entire STALL_VMID field on unstall
+ *   because current GFX10 chips cannot support multi-process debugging due to
+ *   trap configuration and masking being limited to global scope.  Always
+ *   assume single process conditions.
+ *
+ */
+
+#define KGD_GFX_V10_WAVE_LAUNCH_SPI_DRAIN_LATENCY	110
+static void kgd_gfx_v10_set_wave_launch_stall(struct amdgpu_device *adev, uint32_t vmid, bool stall)
+{
+	uint32_t data = RREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL));
+	int i;
+
+	data = REG_SET_FIELD(data, SPI_GDBG_WAVE_CNTL, STALL_VMID,
+							stall ? 1 << vmid : 0);
+
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL), data);
+
+	if (!stall)
+		return;
+
+	for (i = 0; i < KGD_GFX_V10_WAVE_LAUNCH_SPI_DRAIN_LATENCY; i++)
+		RREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL));
+}
+
+uint32_t kgd_gfx_v10_enable_debug_trap(struct amdgpu_device *adev,
+				bool restore_dbg_registers,
+				uint32_t vmid)
+{
+
+	mutex_lock(&adev->grbm_idx_mutex);
+
+	kgd_gfx_v10_set_wave_launch_stall(adev, vmid, true);
+
+	/* assume gfx off is disabled for the debug session if rlc restore not supported. */
+	if (restore_dbg_registers) {
+		uint32_t data = 0;
+
+		data = REG_SET_FIELD(data, SPI_GDBG_TRAP_CONFIG,
+				VMID_SEL, 1 << vmid);
+		data = REG_SET_FIELD(data, SPI_GDBG_TRAP_CONFIG,
+				TRAP_EN, 1);
+		WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_CONFIG), data);
+		WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_DATA0), 0);
+		WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_DATA1), 0);
+
+		kgd_gfx_v10_set_wave_launch_stall(adev, vmid, false);
+
+		mutex_unlock(&adev->grbm_idx_mutex);
+
+		return 0;
+	}
+
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_MASK), 0);
+
+	kgd_gfx_v10_set_wave_launch_stall(adev, vmid, false);
+
+	mutex_unlock(&adev->grbm_idx_mutex);
+
+	return 0;
+}
+
+uint32_t kgd_gfx_v10_disable_debug_trap(struct amdgpu_device *adev,
+					bool keep_trap_enabled,
+					uint32_t vmid)
+{
+	mutex_lock(&adev->grbm_idx_mutex);
+
+	kgd_gfx_v10_set_wave_launch_stall(adev, vmid, true);
+
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_MASK), 0);
+
+	kgd_gfx_v10_set_wave_launch_stall(adev, vmid, false);
+
+	mutex_unlock(&adev->grbm_idx_mutex);
+
+	return 0;
+}
+
+int kgd_gfx_v10_validate_trap_override_request(struct amdgpu_device *adev,
+					      uint32_t trap_override,
+					      uint32_t *trap_mask_supported)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
+	*trap_mask_supported &= KFD_DBG_TRAP_MASK_DBG_ADDRESS_WATCH;
 
-	lock_srbm(kgd, 0, 0, 0, vmid);
+	/* The SPI_GDBG_TRAP_MASK register is global and affects all
+	 * processes. Only allow OR-ing the address-watch bit, since
+	 * this only affects processes under the debugger. Other bits
+	 * should stay 0 to avoid the debugger interfering with other
+	 * processes.
+	 */
+	if (trap_override != KFD_DBG_TRAP_OVERRIDE_OR)
+		return -EINVAL;
+
+	return 0;
+}
+
+uint32_t kgd_gfx_v10_set_wave_launch_trap_override(struct amdgpu_device *adev,
+					      uint32_t vmid,
+					      uint32_t trap_override,
+					      uint32_t trap_mask_bits,
+					      uint32_t trap_mask_request,
+					      uint32_t *trap_mask_prev,
+					      uint32_t kfd_dbg_trap_cntl_prev)
+{
+	uint32_t data, wave_cntl_prev;
+
+	mutex_lock(&adev->grbm_idx_mutex);
+
+	wave_cntl_prev = RREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL));
+
+	kgd_gfx_v10_set_wave_launch_stall(adev, vmid, true);
+
+	data = RREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_MASK));
+	*trap_mask_prev = REG_GET_FIELD(data, SPI_GDBG_TRAP_MASK, EXCP_EN);
+
+	trap_mask_bits = (trap_mask_bits & trap_mask_request) |
+		(*trap_mask_prev & ~trap_mask_request);
+
+	data = REG_SET_FIELD(data, SPI_GDBG_TRAP_MASK, EXCP_EN, trap_mask_bits);
+	data = REG_SET_FIELD(data, SPI_GDBG_TRAP_MASK, REPLACE, trap_override);
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_MASK), data);
+
+	/* We need to preserve wave launch mode stall settings. */
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL), wave_cntl_prev);
+
+	mutex_unlock(&adev->grbm_idx_mutex);
+
+	return 0;
+}
+
+uint32_t kgd_gfx_v10_set_wave_launch_mode(struct amdgpu_device *adev,
+					uint8_t wave_launch_mode,
+					uint32_t vmid)
+{
+	uint32_t data = 0;
+	bool is_mode_set = !!wave_launch_mode;
+
+	mutex_lock(&adev->grbm_idx_mutex);
+
+	kgd_gfx_v10_set_wave_launch_stall(adev, vmid, true);
+
+	data = REG_SET_FIELD(data, SPI_GDBG_WAVE_CNTL2,
+			VMID_MASK, is_mode_set ? 1 << vmid : 0);
+	data = REG_SET_FIELD(data, SPI_GDBG_WAVE_CNTL2,
+			MODE, is_mode_set ? wave_launch_mode : 0);
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL2), data);
+
+	kgd_gfx_v10_set_wave_launch_stall(adev, vmid, false);
+
+	mutex_unlock(&adev->grbm_idx_mutex);
+
+	return 0;
+}
+
+#define TCP_WATCH_STRIDE (mmTCP_WATCH1_ADDR_H - mmTCP_WATCH0_ADDR_H)
+#define SQ_WATCH_STRIDE (mmSQ_WATCH1_ADDR_H - mmSQ_WATCH0_ADDR_H)
+uint32_t kgd_gfx_v10_set_address_watch(struct amdgpu_device *adev,
+					uint64_t watch_address,
+					uint32_t watch_address_mask,
+					uint32_t watch_id,
+					uint32_t watch_mode,
+					uint32_t debug_vmid,
+					uint32_t inst)
+{
+	/* SQ_WATCH?_ADDR_* and TCP_WATCH?_ADDR_* are programmed with the
+	 * same values.
+	 */
+	uint32_t watch_address_high;
+	uint32_t watch_address_low;
+	uint32_t tcp_watch_address_cntl;
+	uint32_t sq_watch_address_cntl;
+
+	watch_address_low = lower_32_bits(watch_address);
+	watch_address_high = upper_32_bits(watch_address) & 0xffff;
+
+	tcp_watch_address_cntl = 0;
+	tcp_watch_address_cntl = REG_SET_FIELD(tcp_watch_address_cntl,
+			TCP_WATCH0_CNTL,
+			VMID,
+			debug_vmid);
+	tcp_watch_address_cntl = REG_SET_FIELD(tcp_watch_address_cntl,
+			TCP_WATCH0_CNTL,
+			MODE,
+			watch_mode);
+	tcp_watch_address_cntl = REG_SET_FIELD(tcp_watch_address_cntl,
+			TCP_WATCH0_CNTL,
+			MASK,
+			watch_address_mask >> 7);
+
+	sq_watch_address_cntl = 0;
+	sq_watch_address_cntl = REG_SET_FIELD(sq_watch_address_cntl,
+			SQ_WATCH0_CNTL,
+			VMID,
+			debug_vmid);
+	sq_watch_address_cntl = REG_SET_FIELD(sq_watch_address_cntl,
+			SQ_WATCH0_CNTL,
+			MODE,
+			watch_mode);
+	sq_watch_address_cntl = REG_SET_FIELD(sq_watch_address_cntl,
+			SQ_WATCH0_CNTL,
+			MASK,
+			watch_address_mask >> 6);
+
+	/* Turning off this watch point until we set all the registers */
+	tcp_watch_address_cntl = REG_SET_FIELD(tcp_watch_address_cntl,
+			TCP_WATCH0_CNTL,
+			VALID,
+			0);
+	WREG32((SOC15_REG_OFFSET(GC, 0, mmTCP_WATCH0_CNTL) +
+			(watch_id * TCP_WATCH_STRIDE)),
+			tcp_watch_address_cntl);
+
+	sq_watch_address_cntl = REG_SET_FIELD(sq_watch_address_cntl,
+			SQ_WATCH0_CNTL,
+			VALID,
+			0);
+	WREG32((SOC15_REG_OFFSET(GC, 0, mmSQ_WATCH0_CNTL) +
+			(watch_id * SQ_WATCH_STRIDE)),
+			sq_watch_address_cntl);
+
+	/* Program {TCP,SQ}_WATCH?_ADDR* */
+	WREG32((SOC15_REG_OFFSET(GC, 0, mmTCP_WATCH0_ADDR_H) +
+			(watch_id * TCP_WATCH_STRIDE)),
+			watch_address_high);
+	WREG32((SOC15_REG_OFFSET(GC, 0, mmTCP_WATCH0_ADDR_L) +
+			(watch_id * TCP_WATCH_STRIDE)),
+			watch_address_low);
+
+	WREG32((SOC15_REG_OFFSET(GC, 0, mmSQ_WATCH0_ADDR_H) +
+			(watch_id * SQ_WATCH_STRIDE)),
+			watch_address_high);
+	WREG32((SOC15_REG_OFFSET(GC, 0, mmSQ_WATCH0_ADDR_L) +
+			(watch_id * SQ_WATCH_STRIDE)),
+			watch_address_low);
+
+	/* Enable the watch point */
+	tcp_watch_address_cntl = REG_SET_FIELD(tcp_watch_address_cntl,
+			TCP_WATCH0_CNTL,
+			VALID,
+			1);
+	WREG32((SOC15_REG_OFFSET(GC, 0, mmTCP_WATCH0_CNTL) +
+			(watch_id * TCP_WATCH_STRIDE)),
+			tcp_watch_address_cntl);
+
+	sq_watch_address_cntl = REG_SET_FIELD(sq_watch_address_cntl,
+			SQ_WATCH0_CNTL,
+			VALID,
+			1);
+	WREG32((SOC15_REG_OFFSET(GC, 0, mmSQ_WATCH0_CNTL) +
+			(watch_id * SQ_WATCH_STRIDE)),
+			sq_watch_address_cntl);
+
+	return 0;
+}
+
+uint32_t kgd_gfx_v10_clear_address_watch(struct amdgpu_device *adev,
+					uint32_t watch_id)
+{
+	uint32_t watch_address_cntl;
+
+	watch_address_cntl = 0;
+
+	WREG32((SOC15_REG_OFFSET(GC, 0, mmTCP_WATCH0_CNTL) +
+			(watch_id * TCP_WATCH_STRIDE)),
+			watch_address_cntl);
+
+	WREG32((SOC15_REG_OFFSET(GC, 0, mmSQ_WATCH0_CNTL) +
+			(watch_id * SQ_WATCH_STRIDE)),
+			watch_address_cntl);
+
+	return 0;
+}
+#undef TCP_WATCH_STRIDE
+#undef SQ_WATCH_STRIDE
+
+
+/* kgd_gfx_v10_get_iq_wait_times: Returns the mmCP_IQ_WAIT_TIME1/2 values
+ * The values read are:
+ *     ib_offload_wait_time     -- Wait Count for Indirect Buffer Offloads.
+ *     atomic_offload_wait_time -- Wait Count for L2 and GDS Atomics Offloads.
+ *     wrm_offload_wait_time    -- Wait Count for WAIT_REG_MEM Offloads.
+ *     gws_wait_time            -- Wait Count for Global Wave Syncs.
+ *     que_sleep_wait_time      -- Wait Count for Dequeue Retry.
+ *     sch_wave_wait_time       -- Wait Count for Scheduling Wave Message.
+ *     sem_rearm_wait_time      -- Wait Count for Semaphore re-arm.
+ *     deq_retry_wait_time      -- Wait Count for Global Wave Syncs.
+ */
+void kgd_gfx_v10_get_iq_wait_times(struct amdgpu_device *adev,
+					uint32_t *wait_times,
+					uint32_t inst)
+
+{
+	*wait_times = RREG32(SOC15_REG_OFFSET(GC, 0, mmCP_IQ_WAIT_TIME2));
+}
+
+void kgd_gfx_v10_build_dequeue_wait_counts_packet_info(struct amdgpu_device *adev,
+						uint32_t wait_times,
+						uint32_t sch_wave,
+						uint32_t que_sleep,
+						uint32_t *reg_offset,
+						uint32_t *reg_data)
+{
+	*reg_data = wait_times;
+
+	if (sch_wave)
+		*reg_data = REG_SET_FIELD(*reg_data,
+				CP_IQ_WAIT_TIME2,
+				SCH_WAVE,
+				sch_wave);
+	if (que_sleep)
+		*reg_data = REG_SET_FIELD(*reg_data,
+				CP_IQ_WAIT_TIME2,
+				QUE_SLEEP,
+				que_sleep);
+
+	*reg_offset = SOC15_REG_OFFSET(GC, 0, mmCP_IQ_WAIT_TIME2);
+}
+
+static void program_trap_handler_settings(struct amdgpu_device *adev,
+		uint32_t vmid, uint64_t tba_addr, uint64_t tma_addr,
+		uint32_t inst)
+{
+	lock_srbm(adev, 0, 0, 0, vmid);
 
 	/*
 	 * Program TBA registers
@@ -781,7 +1067,27 @@ static void program_trap_handler_settings(struct kgd_dev *kgd,
 	WREG32(SOC15_REG_OFFSET(GC, 0, mmSQ_SHADER_TMA_HI),
 			upper_32_bits(tma_addr >> 8));
 
-	unlock_srbm(kgd);
+	unlock_srbm(adev);
+}
+
+uint64_t kgd_gfx_v10_hqd_get_pq_addr(struct amdgpu_device *adev,
+				     uint32_t pipe_id, uint32_t queue_id,
+				     uint32_t inst)
+{
+	return 0;
+}
+
+uint64_t kgd_gfx_v10_hqd_reset(struct amdgpu_device *adev,
+			       uint32_t pipe_id, uint32_t queue_id,
+			       uint32_t inst, unsigned int utimeout)
+{
+	return 0;
+}
+
+uint32_t kgd_gfx_v10_hqd_sdma_get_doorbell(struct amdgpu_device *adev,
+					   int engine, int queue)
+{
+	return 0;
 }
 
 const struct kfd2kgd_calls gfx_v10_kfd2kgd = {
@@ -797,12 +1103,21 @@ const struct kfd2kgd_calls gfx_v10_kfd2kgd = {
 	.hqd_sdma_is_occupied = kgd_hqd_sdma_is_occupied,
 	.hqd_destroy = kgd_hqd_destroy,
 	.hqd_sdma_destroy = kgd_hqd_sdma_destroy,
-	.address_watch_disable = kgd_address_watch_disable,
-	.address_watch_execute = kgd_address_watch_execute,
 	.wave_control_execute = kgd_wave_control_execute,
-	.address_watch_get_offset = kgd_address_watch_get_offset,
 	.get_atc_vmid_pasid_mapping_info =
 			get_atc_vmid_pasid_mapping_info,
 	.set_vm_context_page_table_base = set_vm_context_page_table_base,
+	.enable_debug_trap = kgd_gfx_v10_enable_debug_trap,
+	.disable_debug_trap = kgd_gfx_v10_disable_debug_trap,
+	.validate_trap_override_request = kgd_gfx_v10_validate_trap_override_request,
+	.set_wave_launch_trap_override = kgd_gfx_v10_set_wave_launch_trap_override,
+	.set_wave_launch_mode = kgd_gfx_v10_set_wave_launch_mode,
+	.set_address_watch = kgd_gfx_v10_set_address_watch,
+	.clear_address_watch = kgd_gfx_v10_clear_address_watch,
+	.get_iq_wait_times = kgd_gfx_v10_get_iq_wait_times,
+	.build_dequeue_wait_counts_packet_info = kgd_gfx_v10_build_dequeue_wait_counts_packet_info,
 	.program_trap_handler_settings = program_trap_handler_settings,
+	.hqd_get_pq_addr = kgd_gfx_v10_hqd_get_pq_addr,
+	.hqd_reset = kgd_gfx_v10_hqd_reset,
+	.hqd_sdma_get_doorbell = kgd_gfx_v10_hqd_sdma_get_doorbell
 };
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10.h
new file mode 100644
index 000000000000..a4c607c88178
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10.h
@@ -0,0 +1,70 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+uint32_t kgd_gfx_v10_enable_debug_trap(struct amdgpu_device *adev,
+				      bool restore_dbg_registers,
+				      uint32_t vmid);
+uint32_t kgd_gfx_v10_disable_debug_trap(struct amdgpu_device *adev,
+					bool keep_trap_enabled,
+					uint32_t vmid);
+int kgd_gfx_v10_validate_trap_override_request(struct amdgpu_device *adev,
+					     uint32_t trap_override,
+					     uint32_t *trap_mask_supported);
+uint32_t kgd_gfx_v10_set_wave_launch_trap_override(struct amdgpu_device *adev,
+					     uint32_t vmid,
+					     uint32_t trap_override,
+					     uint32_t trap_mask_bits,
+					     uint32_t trap_mask_request,
+					     uint32_t *trap_mask_prev,
+					     uint32_t kfd_dbg_trap_cntl_prev);
+uint32_t kgd_gfx_v10_set_wave_launch_mode(struct amdgpu_device *adev,
+					 uint8_t wave_launch_mode,
+					 uint32_t vmid);
+uint32_t kgd_gfx_v10_set_address_watch(struct amdgpu_device *adev,
+					uint64_t watch_address,
+					uint32_t watch_address_mask,
+					uint32_t watch_id,
+					uint32_t watch_mode,
+					uint32_t debug_vmid,
+					uint32_t inst);
+uint32_t kgd_gfx_v10_clear_address_watch(struct amdgpu_device *adev,
+					uint32_t watch_id);
+void kgd_gfx_v10_get_iq_wait_times(struct amdgpu_device *adev,
+				uint32_t *wait_times,
+				uint32_t inst);
+void kgd_gfx_v10_build_dequeue_wait_counts_packet_info(struct amdgpu_device *adev,
+					       uint32_t wait_times,
+					       uint32_t sch_wave,
+					       uint32_t que_sleep,
+					       uint32_t *reg_offset,
+					       uint32_t *reg_data);
+uint64_t kgd_gfx_v10_hqd_get_pq_addr(struct amdgpu_device *adev,
+				    uint32_t pipe_id,
+				    uint32_t queue_id,
+				    uint32_t inst);
+uint64_t kgd_gfx_v10_hqd_reset(struct amdgpu_device *adev,
+			      uint32_t pipe_id,
+			      uint32_t queue_id,
+			      uint32_t inst,
+			      unsigned int utimeout);
+uint32_t kgd_gfx_v10_hqd_sdma_get_doorbell(struct amdgpu_device *adev,
+					   int engine, int queue);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10_3.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10_3.c
index dac0d751d5af..f2278a0937ff 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10_3.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10_3.c
@@ -22,10 +22,13 @@
 #include <linux/mmu_context.h>
 #include "amdgpu.h"
 #include "amdgpu_amdkfd.h"
+#include "amdgpu_amdkfd_gfx_v10.h"
 #include "gc/gc_10_3_0_offset.h"
 #include "gc/gc_10_3_0_sh_mask.h"
 #include "oss/osssys_5_0_0_offset.h"
 #include "oss/osssys_5_0_0_sh_mask.h"
+#include "athub/athub_2_1_0_offset.h"
+#include "athub/athub_2_1_0_sh_mask.h"
 #include "soc15_common.h"
 #include "v10_structs.h"
 #include "nv.h"
@@ -38,37 +41,26 @@ enum hqd_dequeue_request_type {
 	SAVE_WAVES
 };
 
-static inline struct amdgpu_device *get_amdgpu_device(struct kgd_dev *kgd)
-{
-	return (struct amdgpu_device *)kgd;
-}
-
-static void lock_srbm(struct kgd_dev *kgd, uint32_t mec, uint32_t pipe,
+static void lock_srbm(struct amdgpu_device *adev, uint32_t mec, uint32_t pipe,
 			uint32_t queue, uint32_t vmid)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	mutex_lock(&adev->srbm_mutex);
 	nv_grbm_select(adev, mec, pipe, queue, vmid);
 }
 
-static void unlock_srbm(struct kgd_dev *kgd)
+static void unlock_srbm(struct amdgpu_device *adev)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	nv_grbm_select(adev, 0, 0, 0, 0);
 	mutex_unlock(&adev->srbm_mutex);
 }
 
-static void acquire_queue(struct kgd_dev *kgd, uint32_t pipe_id,
+static void acquire_queue(struct amdgpu_device *adev, uint32_t pipe_id,
 				uint32_t queue_id)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	uint32_t mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
 	uint32_t pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);
 
-	lock_srbm(kgd, mec, pipe, queue_id, 0);
+	lock_srbm(adev, mec, pipe, queue_id, 0);
 }
 
 static uint64_t get_queue_mask(struct amdgpu_device *adev,
@@ -80,34 +72,30 @@ static uint64_t get_queue_mask(struct amdgpu_device *adev,
 	return 1ull << bit;
 }
 
-static void release_queue(struct kgd_dev *kgd)
+static void release_queue(struct amdgpu_device *adev)
 {
-	unlock_srbm(kgd);
+	unlock_srbm(adev);
 }
 
-static void program_sh_mem_settings_v10_3(struct kgd_dev *kgd, uint32_t vmid,
+static void program_sh_mem_settings_v10_3(struct amdgpu_device *adev, uint32_t vmid,
 					uint32_t sh_mem_config,
 					uint32_t sh_mem_ape1_base,
 					uint32_t sh_mem_ape1_limit,
-					uint32_t sh_mem_bases)
+					uint32_t sh_mem_bases, uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
-	lock_srbm(kgd, 0, 0, 0, vmid);
+	lock_srbm(adev, 0, 0, 0, vmid);
 
 	WREG32_SOC15(GC, 0, mmSH_MEM_CONFIG, sh_mem_config);
 	WREG32_SOC15(GC, 0, mmSH_MEM_BASES, sh_mem_bases);
 	/* APE1 no longer exists on GFX9 */
 
-	unlock_srbm(kgd);
+	unlock_srbm(adev);
 }
 
 /* ATC is defeatured on Sienna_Cichlid */
-static int set_pasid_vmid_mapping_v10_3(struct kgd_dev *kgd, unsigned int pasid,
-					unsigned int vmid)
+static int set_pasid_vmid_mapping_v10_3(struct amdgpu_device *adev, unsigned int pasid,
+					unsigned int vmid, uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	uint32_t value = pasid << IH_VMID_0_LUT__PASID__SHIFT;
 
 	/* Mapping vmid to pasid also for IH block */
@@ -118,22 +106,22 @@ static int set_pasid_vmid_mapping_v10_3(struct kgd_dev *kgd, unsigned int pasid,
 	return 0;
 }
 
-static int init_interrupts_v10_3(struct kgd_dev *kgd, uint32_t pipe_id)
+static int init_interrupts_v10_3(struct amdgpu_device *adev, uint32_t pipe_id,
+				uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t mec;
 	uint32_t pipe;
 
 	mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
 	pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);
 
-	lock_srbm(kgd, mec, pipe, 0, 0);
+	lock_srbm(adev, mec, pipe, 0, 0);
 
 	WREG32_SOC15(GC, 0, mmCPC_INT_CNTL,
 		CP_INT_CNTL_RING0__TIME_STAMP_INT_ENABLE_MASK |
 		CP_INT_CNTL_RING0__OPCODE_ERROR_INT_ENABLE_MASK);
 
-	unlock_srbm(kgd);
+	unlock_srbm(adev);
 
 	return 0;
 }
@@ -188,12 +176,11 @@ static inline struct v10_sdma_mqd *get_sdma_mqd(void *mqd)
 	return (struct v10_sdma_mqd *)mqd;
 }
 
-static int hqd_load_v10_3(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
-			uint32_t queue_id, uint32_t __user *wptr,
-			uint32_t wptr_shift, uint32_t wptr_mask,
-			struct mm_struct *mm)
+static int hqd_load_v10_3(struct amdgpu_device *adev, void *mqd,
+			uint32_t pipe_id, uint32_t queue_id,
+			uint32_t __user *wptr, uint32_t wptr_shift,
+			uint32_t wptr_mask, struct mm_struct *mm, uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct v10_compute_mqd *m;
 	uint32_t *mqd_hqd;
 	uint32_t reg, hqd_base, data;
@@ -201,7 +188,7 @@ static int hqd_load_v10_3(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
 	m = get_mqd(mqd);
 
 	pr_debug("Load hqd of pipe %d queue %d\n", pipe_id, queue_id);
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 
 	/* HIQ is set during driver init period with vmid set to 0*/
 	if (m->cp_hqd_vmid == 0) {
@@ -281,24 +268,23 @@ static int hqd_load_v10_3(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
 	data = REG_SET_FIELD(m->cp_hqd_active, CP_HQD_ACTIVE, ACTIVE, 1);
 	WREG32_SOC15(GC, 0, mmCP_HQD_ACTIVE, data);
 
-	release_queue(kgd);
+	release_queue(adev);
 
 	return 0;
 }
 
-static int hiq_mqd_load_v10_3(struct kgd_dev *kgd, void *mqd,
+static int hiq_mqd_load_v10_3(struct amdgpu_device *adev, void *mqd,
 			    uint32_t pipe_id, uint32_t queue_id,
-			    uint32_t doorbell_off)
+			    uint32_t doorbell_off, uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-	struct amdgpu_ring *kiq_ring = &adev->gfx.kiq.ring;
+	struct amdgpu_ring *kiq_ring = &adev->gfx.kiq[0].ring;
 	struct v10_compute_mqd *m;
 	uint32_t mec, pipe;
 	int r;
 
 	m = get_mqd(mqd);
 
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 
 	mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
 	pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);
@@ -306,7 +292,7 @@ static int hiq_mqd_load_v10_3(struct kgd_dev *kgd, void *mqd,
 	pr_debug("kfd: set HIQ, mec:%d, pipe:%d, queue:%d.\n",
 		 mec, pipe, queue_id);
 
-	spin_lock(&adev->gfx.kiq.ring_lock);
+	spin_lock(&adev->gfx.kiq[0].ring_lock);
 	r = amdgpu_ring_alloc(kiq_ring, 7);
 	if (r) {
 		pr_err("Failed to alloc KIQ (%d).\n", r);
@@ -333,17 +319,16 @@ static int hiq_mqd_load_v10_3(struct kgd_dev *kgd, void *mqd,
 	amdgpu_ring_commit(kiq_ring);
 
 out_unlock:
-	spin_unlock(&adev->gfx.kiq.ring_lock);
-	release_queue(kgd);
+	spin_unlock(&adev->gfx.kiq[0].ring_lock);
+	release_queue(adev);
 
 	return r;
 }
 
-static int hqd_dump_v10_3(struct kgd_dev *kgd,
+static int hqd_dump_v10_3(struct amdgpu_device *adev,
 			uint32_t pipe_id, uint32_t queue_id,
-			uint32_t (**dump)[2], uint32_t *n_regs)
+			uint32_t (**dump)[2], uint32_t *n_regs, uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t i = 0, reg;
 #define HQD_N_REGS 56
 #define DUMP_REG(addr) do {				\
@@ -353,17 +338,17 @@ static int hqd_dump_v10_3(struct kgd_dev *kgd,
 		(*dump)[i++][1] = RREG32_SOC15_IP(GC, addr);		\
 	} while (0)
 
-	*dump = kmalloc(HQD_N_REGS*2*sizeof(uint32_t), GFP_KERNEL);
+	*dump = kmalloc_array(HQD_N_REGS, sizeof(**dump), GFP_KERNEL);
 	if (*dump == NULL)
 		return -ENOMEM;
 
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 
 	for (reg = SOC15_REG_OFFSET(GC, 0, mmCP_MQD_BASE_ADDR);
 	     reg <= SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_HI); reg++)
 		DUMP_REG(reg);
 
-	release_queue(kgd);
+	release_queue(adev);
 
 	WARN_ON_ONCE(i != HQD_N_REGS);
 	*n_regs = i;
@@ -371,10 +356,9 @@ static int hqd_dump_v10_3(struct kgd_dev *kgd,
 	return 0;
 }
 
-static int hqd_sdma_load_v10_3(struct kgd_dev *kgd, void *mqd,
+static int hqd_sdma_load_v10_3(struct amdgpu_device *adev, void *mqd,
 			     uint32_t __user *wptr, struct mm_struct *mm)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct v10_sdma_mqd *m;
 	uint32_t sdma_rlc_reg_offset;
 	unsigned long end_jiffies;
@@ -441,18 +425,17 @@ static int hqd_sdma_load_v10_3(struct kgd_dev *kgd, void *mqd,
 	return 0;
 }
 
-static int hqd_sdma_dump_v10_3(struct kgd_dev *kgd,
+static int hqd_sdma_dump_v10_3(struct amdgpu_device *adev,
 			     uint32_t engine_id, uint32_t queue_id,
 			     uint32_t (**dump)[2], uint32_t *n_regs)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t sdma_rlc_reg_offset = get_sdma_rlc_reg_offset(adev,
 			engine_id, queue_id);
 	uint32_t i = 0, reg;
 #undef HQD_N_REGS
 #define HQD_N_REGS (19+6+7+12)
 
-	*dump = kmalloc(HQD_N_REGS*2*sizeof(uint32_t), GFP_KERNEL);
+	*dump = kmalloc_array(HQD_N_REGS, sizeof(**dump), GFP_KERNEL);
 	if (*dump == NULL)
 		return -ENOMEM;
 
@@ -473,15 +456,15 @@ static int hqd_sdma_dump_v10_3(struct kgd_dev *kgd,
 	return 0;
 }
 
-static bool hqd_is_occupied_v10_3(struct kgd_dev *kgd, uint64_t queue_address,
-				uint32_t pipe_id, uint32_t queue_id)
+static bool hqd_is_occupied_v10_3(struct amdgpu_device *adev,
+				uint64_t queue_address, uint32_t pipe_id,
+				uint32_t queue_id, uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t act;
 	bool retval = false;
 	uint32_t low, high;
 
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 	act = RREG32_SOC15(GC, 0, mmCP_HQD_ACTIVE);
 	if (act) {
 		low = lower_32_bits(queue_address >> 8);
@@ -491,13 +474,13 @@ static bool hqd_is_occupied_v10_3(struct kgd_dev *kgd, uint64_t queue_address,
 		   high == RREG32_SOC15(GC, 0, mmCP_HQD_PQ_BASE_HI))
 			retval = true;
 	}
-	release_queue(kgd);
+	release_queue(adev);
 	return retval;
 }
 
-static bool hqd_sdma_is_occupied_v10_3(struct kgd_dev *kgd, void *mqd)
+static bool hqd_sdma_is_occupied_v10_3(struct amdgpu_device *adev,
+				void *mqd)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct v10_sdma_mqd *m;
 	uint32_t sdma_rlc_reg_offset;
 	uint32_t sdma_rlc_rb_cntl;
@@ -514,18 +497,17 @@ static bool hqd_sdma_is_occupied_v10_3(struct kgd_dev *kgd, void *mqd)
 	return false;
 }
 
-static int hqd_destroy_v10_3(struct kgd_dev *kgd, void *mqd,
+static int hqd_destroy_v10_3(struct amdgpu_device *adev, void *mqd,
 				enum kfd_preempt_type reset_type,
 				unsigned int utimeout, uint32_t pipe_id,
-				uint32_t queue_id)
+				uint32_t queue_id, uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	enum hqd_dequeue_request_type type;
 	unsigned long end_jiffies;
 	uint32_t temp;
 	struct v10_compute_mqd *m = get_mqd(mqd);
 
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 
 	if (m->cp_hqd_vmid == 0)
 		WREG32_FIELD15(GC, 0, RLC_CP_SCHEDULERS, scheduler1, 0);
@@ -555,20 +537,19 @@ static int hqd_destroy_v10_3(struct kgd_dev *kgd, void *mqd,
 		if (time_after(jiffies, end_jiffies)) {
 			pr_err("cp queue pipe %d queue %d preemption failed\n",
 					pipe_id, queue_id);
-			release_queue(kgd);
+			release_queue(adev);
 			return -ETIME;
 		}
 		usleep_range(500, 1000);
 	}
 
-	release_queue(kgd);
+	release_queue(adev);
 	return 0;
 }
 
-static int hqd_sdma_destroy_v10_3(struct kgd_dev *kgd, void *mqd,
+static int hqd_sdma_destroy_v10_3(struct amdgpu_device *adev, void *mqd,
 				unsigned int utimeout)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct v10_sdma_mqd *m;
 	uint32_t sdma_rlc_reg_offset;
 	uint32_t temp;
@@ -605,26 +586,10 @@ static int hqd_sdma_destroy_v10_3(struct kgd_dev *kgd, void *mqd,
 	return 0;
 }
 
-
-static int address_watch_disable_v10_3(struct kgd_dev *kgd)
-{
-	return 0;
-}
-
-static int address_watch_execute_v10_3(struct kgd_dev *kgd,
-					unsigned int watch_point_id,
-					uint32_t cntl_val,
-					uint32_t addr_hi,
-					uint32_t addr_lo)
-{
-	return 0;
-}
-
-static int wave_control_execute_v10_3(struct kgd_dev *kgd,
+static int wave_control_execute_v10_3(struct amdgpu_device *adev,
 					uint32_t gfx_index_val,
-					uint32_t sq_cmd)
+					uint32_t sq_cmd, uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t data = 0;
 
 	mutex_lock(&adev->grbm_idx_mutex);
@@ -645,28 +610,30 @@ static int wave_control_execute_v10_3(struct kgd_dev *kgd,
 	return 0;
 }
 
-static uint32_t address_watch_get_offset_v10_3(struct kgd_dev *kgd,
-					unsigned int watch_point_id,
-					unsigned int reg_offset)
+static bool get_atc_vmid_pasid_mapping_info_v10_3(struct amdgpu_device *adev,
+					uint8_t vmid, uint16_t *p_pasid)
 {
-	return 0;
+	uint32_t value;
+
+	value = RREG32(SOC15_REG_OFFSET(ATHUB, 0, mmATC_VMID0_PASID_MAPPING)
+		     + vmid);
+	*p_pasid = value & ATC_VMID0_PASID_MAPPING__PASID_MASK;
+
+	return !!(value & ATC_VMID0_PASID_MAPPING__VALID_MASK);
 }
 
-static void set_vm_context_page_table_base_v10_3(struct kgd_dev *kgd, uint32_t vmid,
-		uint64_t page_table_base)
+static void set_vm_context_page_table_base_v10_3(struct amdgpu_device *adev,
+		uint32_t vmid, uint64_t page_table_base)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	/* SDMA is on gfxhub as well for Navi1* series */
 	adev->gfxhub.funcs->setup_vm_pt_regs(adev, vmid, page_table_base);
 }
 
-static void program_trap_handler_settings_v10_3(struct kgd_dev *kgd,
-			uint32_t vmid, uint64_t tba_addr, uint64_t tma_addr)
+static void program_trap_handler_settings_v10_3(struct amdgpu_device *adev,
+			uint32_t vmid, uint64_t tba_addr, uint64_t tma_addr,
+			uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
-	lock_srbm(kgd, 0, 0, 0, vmid);
+	lock_srbm(adev, 0, 0, 0, vmid);
 
 	/*
 	 * Program TBA registers
@@ -685,151 +652,8 @@ static void program_trap_handler_settings_v10_3(struct kgd_dev *kgd,
 	WREG32(SOC15_REG_OFFSET(GC, 0, mmSQ_SHADER_TMA_HI),
 			 upper_32_bits(tma_addr >> 8));
 
-	unlock_srbm(kgd);
-}
-
-#if 0
-uint32_t enable_debug_trap_v10_3(struct kgd_dev *kgd,
-				uint32_t trap_debug_wave_launch_mode,
-				uint32_t vmid)
-{
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-	uint32_t data = 0;
-	uint32_t orig_wave_cntl_value;
-	uint32_t orig_stall_vmid;
-
-	mutex_lock(&adev->grbm_idx_mutex);
-
-	orig_wave_cntl_value = RREG32(SOC15_REG_OFFSET(GC,
-				0,
-				mmSPI_GDBG_WAVE_CNTL));
-	orig_stall_vmid = REG_GET_FIELD(orig_wave_cntl_value,
-			SPI_GDBG_WAVE_CNTL,
-			STALL_VMID);
-
-	data = REG_SET_FIELD(data, SPI_GDBG_WAVE_CNTL, STALL_RA, 1);
-	WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL), data);
-
-	data = 0;
-	WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_MASK), data);
-
-	WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL), orig_stall_vmid);
-
-	mutex_unlock(&adev->grbm_idx_mutex);
-
-	return 0;
-}
-
-uint32_t disable_debug_trap_v10_3(struct kgd_dev *kgd)
-{
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
-	mutex_lock(&adev->grbm_idx_mutex);
-
-	WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_MASK), 0);
-
-	mutex_unlock(&adev->grbm_idx_mutex);
-
-	return 0;
-}
-
-uint32_t set_wave_launch_trap_override_v10_3(struct kgd_dev *kgd,
-						uint32_t trap_override,
-						uint32_t trap_mask)
-{
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-	uint32_t data = 0;
-
-	mutex_lock(&adev->grbm_idx_mutex);
-
-	data = RREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL));
-	data = REG_SET_FIELD(data, SPI_GDBG_WAVE_CNTL, STALL_RA, 1);
-	WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL), data);
-
-	data = 0;
-	data = REG_SET_FIELD(data, SPI_GDBG_TRAP_MASK,
-			EXCP_EN, trap_mask);
-	data = REG_SET_FIELD(data, SPI_GDBG_TRAP_MASK,
-			REPLACE, trap_override);
-	WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_MASK), data);
-
-	data = RREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL));
-	data = REG_SET_FIELD(data, SPI_GDBG_WAVE_CNTL, STALL_RA, 0);
-	WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL), data);
-
-	mutex_unlock(&adev->grbm_idx_mutex);
-
-	return 0;
-}
-
-uint32_t set_wave_launch_mode_v10_3(struct kgd_dev *kgd,
-					uint8_t wave_launch_mode,
-					uint32_t vmid)
-{
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-	uint32_t data = 0;
-	bool is_stall_mode;
-	bool is_mode_set;
-
-	is_stall_mode = (wave_launch_mode == 4);
-	is_mode_set = (wave_launch_mode != 0 && wave_launch_mode != 4);
-
-	mutex_lock(&adev->grbm_idx_mutex);
-
-	data = REG_SET_FIELD(data, SPI_GDBG_WAVE_CNTL2,
-			VMID_MASK, is_mode_set ? 1 << vmid : 0);
-	data = REG_SET_FIELD(data, SPI_GDBG_WAVE_CNTL2,
-			MODE, is_mode_set ? wave_launch_mode : 0);
-	WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL2), data);
-
-	data = RREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL));
-	data = REG_SET_FIELD(data, SPI_GDBG_WAVE_CNTL,
-			STALL_VMID, is_stall_mode ? 1 << vmid : 0);
-	data = REG_SET_FIELD(data, SPI_GDBG_WAVE_CNTL,
-			STALL_RA, is_stall_mode ? 1 : 0);
-	WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL), data);
-
-	mutex_unlock(&adev->grbm_idx_mutex);
-
-	return 0;
-}
-
-/* kgd_get_iq_wait_times: Returns the mmCP_IQ_WAIT_TIME1/2 values
- * The values read are:
- *	ib_offload_wait_time     -- Wait Count for Indirect Buffer Offloads.
- *	atomic_offload_wait_time -- Wait Count for L2 and GDS Atomics Offloads.
- *	wrm_offload_wait_time    -- Wait Count for WAIT_REG_MEM Offloads.
- *	gws_wait_time            -- Wait Count for Global Wave Syncs.
- *	que_sleep_wait_time      -- Wait Count for Dequeue Retry.
- *	sch_wave_wait_time       -- Wait Count for Scheduling Wave Message.
- *	sem_rearm_wait_time      -- Wait Count for Semaphore re-arm.
- *	deq_retry_wait_time      -- Wait Count for Global Wave Syncs.
- */
-void get_iq_wait_times_v10_3(struct kgd_dev *kgd,
-					uint32_t *wait_times)
-
-{
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
-	*wait_times = RREG32(SOC15_REG_OFFSET(GC, 0, mmCP_IQ_WAIT_TIME2));
-}
-
-void build_grace_period_packet_info_v10_3(struct kgd_dev *kgd,
-						uint32_t wait_times,
-						uint32_t grace_period,
-						uint32_t *reg_offset,
-						uint32_t *reg_data)
-{
-	*reg_data = wait_times;
-
-	*reg_data = REG_SET_FIELD(*reg_data,
-			CP_IQ_WAIT_TIME2,
-			SCH_WAVE,
-			grace_period);
-
-	*reg_offset = mmCP_IQ_WAIT_TIME2;
+	unlock_srbm(adev);
 }
-#endif
 
 const struct kfd2kgd_calls gfx_v10_3_kfd2kgd = {
 	.program_sh_mem_settings = program_sh_mem_settings_v10_3,
@@ -844,19 +668,20 @@ const struct kfd2kgd_calls gfx_v10_3_kfd2kgd = {
 	.hqd_sdma_is_occupied = hqd_sdma_is_occupied_v10_3,
 	.hqd_destroy = hqd_destroy_v10_3,
 	.hqd_sdma_destroy = hqd_sdma_destroy_v10_3,
-	.address_watch_disable = address_watch_disable_v10_3,
-	.address_watch_execute = address_watch_execute_v10_3,
 	.wave_control_execute = wave_control_execute_v10_3,
-	.address_watch_get_offset = address_watch_get_offset_v10_3,
-	.get_atc_vmid_pasid_mapping_info = NULL,
+	.get_atc_vmid_pasid_mapping_info = get_atc_vmid_pasid_mapping_info_v10_3,
 	.set_vm_context_page_table_base = set_vm_context_page_table_base_v10_3,
 	.program_trap_handler_settings = program_trap_handler_settings_v10_3,
-#if 0
-	.enable_debug_trap = enable_debug_trap_v10_3,
-	.disable_debug_trap = disable_debug_trap_v10_3,
-	.set_wave_launch_trap_override = set_wave_launch_trap_override_v10_3,
-	.set_wave_launch_mode = set_wave_launch_mode_v10_3,
-	.get_iq_wait_times = get_iq_wait_times_v10_3,
-	.build_grace_period_packet_info = build_grace_period_packet_info_v10_3,
-#endif
+	.get_iq_wait_times = kgd_gfx_v10_get_iq_wait_times,
+	.build_dequeue_wait_counts_packet_info = kgd_gfx_v10_build_dequeue_wait_counts_packet_info,
+	.enable_debug_trap = kgd_gfx_v10_enable_debug_trap,
+	.disable_debug_trap = kgd_gfx_v10_disable_debug_trap,
+	.validate_trap_override_request = kgd_gfx_v10_validate_trap_override_request,
+	.set_wave_launch_trap_override = kgd_gfx_v10_set_wave_launch_trap_override,
+	.set_wave_launch_mode = kgd_gfx_v10_set_wave_launch_mode,
+	.set_address_watch = kgd_gfx_v10_set_address_watch,
+	.clear_address_watch = kgd_gfx_v10_clear_address_watch,
+	.hqd_get_pq_addr = kgd_gfx_v10_hqd_get_pq_addr,
+	.hqd_reset = kgd_gfx_v10_hqd_reset,
+	.hqd_sdma_get_doorbell = kgd_gfx_v10_hqd_sdma_get_doorbell
 };
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v11.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v11.c
new file mode 100644
index 000000000000..aaccf0b9947d
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v11.c
@@ -0,0 +1,835 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+#include <linux/mmu_context.h>
+#include "amdgpu.h"
+#include "amdgpu_amdkfd.h"
+#include "gc/gc_11_0_0_offset.h"
+#include "gc/gc_11_0_0_sh_mask.h"
+#include "oss/osssys_6_0_0_offset.h"
+#include "oss/osssys_6_0_0_sh_mask.h"
+#include "soc15_common.h"
+#include "soc15d.h"
+#include "v11_structs.h"
+#include "soc21.h"
+#include <uapi/linux/kfd_ioctl.h>
+
+enum hqd_dequeue_request_type {
+	NO_ACTION = 0,
+	DRAIN_PIPE,
+	RESET_WAVES,
+	SAVE_WAVES
+};
+
+static void lock_srbm(struct amdgpu_device *adev, uint32_t mec, uint32_t pipe,
+			uint32_t queue, uint32_t vmid)
+{
+	mutex_lock(&adev->srbm_mutex);
+	soc21_grbm_select(adev, mec, pipe, queue, vmid);
+}
+
+static void unlock_srbm(struct amdgpu_device *adev)
+{
+	soc21_grbm_select(adev, 0, 0, 0, 0);
+	mutex_unlock(&adev->srbm_mutex);
+}
+
+static void acquire_queue(struct amdgpu_device *adev, uint32_t pipe_id,
+				uint32_t queue_id)
+{
+	uint32_t mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
+	uint32_t pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);
+
+	lock_srbm(adev, mec, pipe, queue_id, 0);
+}
+
+static uint64_t get_queue_mask(struct amdgpu_device *adev,
+			       uint32_t pipe_id, uint32_t queue_id)
+{
+	unsigned int bit = pipe_id * adev->gfx.mec.num_queue_per_pipe +
+			queue_id;
+
+	return 1ull << bit;
+}
+
+static void release_queue(struct amdgpu_device *adev)
+{
+	unlock_srbm(adev);
+}
+
+static void program_sh_mem_settings_v11(struct amdgpu_device *adev, uint32_t vmid,
+					uint32_t sh_mem_config,
+					uint32_t sh_mem_ape1_base,
+					uint32_t sh_mem_ape1_limit,
+					uint32_t sh_mem_bases, uint32_t inst)
+{
+	lock_srbm(adev, 0, 0, 0, vmid);
+
+	WREG32(SOC15_REG_OFFSET(GC, 0, regSH_MEM_CONFIG), sh_mem_config);
+	WREG32(SOC15_REG_OFFSET(GC, 0, regSH_MEM_BASES), sh_mem_bases);
+
+	unlock_srbm(adev);
+}
+
+static int set_pasid_vmid_mapping_v11(struct amdgpu_device *adev, unsigned int pasid,
+					unsigned int vmid, uint32_t inst)
+{
+	uint32_t value = pasid << IH_VMID_0_LUT__PASID__SHIFT;
+
+	/* Mapping vmid to pasid also for IH block */
+	pr_debug("mapping vmid %d -> pasid %d in IH block for GFX client\n",
+			vmid, pasid);
+	WREG32(SOC15_REG_OFFSET(OSSSYS, 0, regIH_VMID_0_LUT) + vmid, value);
+
+	return 0;
+}
+
+static int init_interrupts_v11(struct amdgpu_device *adev, uint32_t pipe_id,
+				uint32_t inst)
+{
+	uint32_t mec;
+	uint32_t pipe;
+
+	mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
+	pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);
+
+	lock_srbm(adev, mec, pipe, 0, 0);
+
+	WREG32_SOC15(GC, 0, regCPC_INT_CNTL,
+		CP_INT_CNTL_RING0__TIME_STAMP_INT_ENABLE_MASK |
+		CP_INT_CNTL_RING0__OPCODE_ERROR_INT_ENABLE_MASK);
+
+	unlock_srbm(adev);
+
+	return 0;
+}
+
+static uint32_t get_sdma_rlc_reg_offset(struct amdgpu_device *adev,
+				unsigned int engine_id,
+				unsigned int queue_id)
+{
+	uint32_t sdma_engine_reg_base = 0;
+	uint32_t sdma_rlc_reg_offset;
+
+	switch (engine_id) {
+	case 0:
+		sdma_engine_reg_base = SOC15_REG_OFFSET(SDMA0, 0,
+				regSDMA0_QUEUE0_RB_CNTL) - regSDMA0_QUEUE0_RB_CNTL;
+		break;
+	case 1:
+		sdma_engine_reg_base = SOC15_REG_OFFSET(SDMA1, 0,
+				regSDMA1_QUEUE0_RB_CNTL) - regSDMA0_QUEUE0_RB_CNTL;
+		break;
+	default:
+		BUG();
+	}
+
+	sdma_rlc_reg_offset = sdma_engine_reg_base
+		+ queue_id * (regSDMA0_QUEUE1_RB_CNTL - regSDMA0_QUEUE0_RB_CNTL);
+
+	pr_debug("RLC register offset for SDMA%d RLC%d: 0x%x\n", engine_id,
+			queue_id, sdma_rlc_reg_offset);
+
+	return sdma_rlc_reg_offset;
+}
+
+static inline struct v11_compute_mqd *get_mqd(void *mqd)
+{
+	return (struct v11_compute_mqd *)mqd;
+}
+
+static inline struct v11_sdma_mqd *get_sdma_mqd(void *mqd)
+{
+	return (struct v11_sdma_mqd *)mqd;
+}
+
+static int hqd_load_v11(struct amdgpu_device *adev, void *mqd, uint32_t pipe_id,
+			uint32_t queue_id, uint32_t __user *wptr,
+			uint32_t wptr_shift, uint32_t wptr_mask,
+			struct mm_struct *mm, uint32_t inst)
+{
+	struct v11_compute_mqd *m;
+	uint32_t *mqd_hqd;
+	uint32_t reg, hqd_base, data;
+
+	m = get_mqd(mqd);
+
+	pr_debug("Load hqd of pipe %d queue %d\n", pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
+
+	/* HIQ is set during driver init period with vmid set to 0*/
+	if (m->cp_hqd_vmid == 0) {
+		uint32_t value, mec, pipe;
+
+		mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
+		pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);
+
+		pr_debug("kfd: set HIQ, mec:%d, pipe:%d, queue:%d.\n",
+			mec, pipe, queue_id);
+		value = RREG32(SOC15_REG_OFFSET(GC, 0, regRLC_CP_SCHEDULERS));
+		value = REG_SET_FIELD(value, RLC_CP_SCHEDULERS, scheduler1,
+			((mec << 5) | (pipe << 3) | queue_id | 0x80));
+		WREG32(SOC15_REG_OFFSET(GC, 0, regRLC_CP_SCHEDULERS), value);
+	}
+
+	/* HQD registers extend from CP_MQD_BASE_ADDR to CP_HQD_EOP_WPTR_MEM. */
+	mqd_hqd = &m->cp_mqd_base_addr_lo;
+	hqd_base = SOC15_REG_OFFSET(GC, 0, regCP_MQD_BASE_ADDR);
+
+	for (reg = hqd_base;
+	     reg <= SOC15_REG_OFFSET(GC, 0, regCP_HQD_PQ_WPTR_HI); reg++)
+		WREG32(reg, mqd_hqd[reg - hqd_base]);
+
+
+	/* Activate doorbell logic before triggering WPTR poll. */
+	data = REG_SET_FIELD(m->cp_hqd_pq_doorbell_control,
+			     CP_HQD_PQ_DOORBELL_CONTROL, DOORBELL_EN, 1);
+	WREG32(SOC15_REG_OFFSET(GC, 0, regCP_HQD_PQ_DOORBELL_CONTROL), data);
+
+	if (wptr) {
+		/* Don't read wptr with get_user because the user
+		 * context may not be accessible (if this function
+		 * runs in a work queue). Instead trigger a one-shot
+		 * polling read from memory in the CP. This assumes
+		 * that wptr is GPU-accessible in the queue's VMID via
+		 * ATC or SVM. WPTR==RPTR before starting the poll so
+		 * the CP starts fetching new commands from the right
+		 * place.
+		 *
+		 * Guessing a 64-bit WPTR from a 32-bit RPTR is a bit
+		 * tricky. Assume that the queue didn't overflow. The
+		 * number of valid bits in the 32-bit RPTR depends on
+		 * the queue size. The remaining bits are taken from
+		 * the saved 64-bit WPTR. If the WPTR wrapped, add the
+		 * queue size.
+		 */
+		uint32_t queue_size =
+			2 << REG_GET_FIELD(m->cp_hqd_pq_control,
+					   CP_HQD_PQ_CONTROL, QUEUE_SIZE);
+		uint64_t guessed_wptr = m->cp_hqd_pq_rptr & (queue_size - 1);
+
+		if ((m->cp_hqd_pq_wptr_lo & (queue_size - 1)) < guessed_wptr)
+			guessed_wptr += queue_size;
+		guessed_wptr += m->cp_hqd_pq_wptr_lo & ~(queue_size - 1);
+		guessed_wptr += (uint64_t)m->cp_hqd_pq_wptr_hi << 32;
+
+		WREG32(SOC15_REG_OFFSET(GC, 0, regCP_HQD_PQ_WPTR_LO),
+		       lower_32_bits(guessed_wptr));
+		WREG32(SOC15_REG_OFFSET(GC, 0, regCP_HQD_PQ_WPTR_HI),
+		       upper_32_bits(guessed_wptr));
+		WREG32(SOC15_REG_OFFSET(GC, 0, regCP_HQD_PQ_WPTR_POLL_ADDR),
+		       lower_32_bits((uint64_t)wptr));
+		WREG32(SOC15_REG_OFFSET(GC, 0, regCP_HQD_PQ_WPTR_POLL_ADDR_HI),
+		       upper_32_bits((uint64_t)wptr));
+		pr_debug("%s setting CP_PQ_WPTR_POLL_CNTL1 to %x\n", __func__,
+			 (uint32_t)get_queue_mask(adev, pipe_id, queue_id));
+		WREG32(SOC15_REG_OFFSET(GC, 0, regCP_PQ_WPTR_POLL_CNTL1),
+		       (uint32_t)get_queue_mask(adev, pipe_id, queue_id));
+	}
+
+	/* Start the EOP fetcher */
+	WREG32(SOC15_REG_OFFSET(GC, 0, regCP_HQD_EOP_RPTR),
+	       REG_SET_FIELD(m->cp_hqd_eop_rptr,
+			     CP_HQD_EOP_RPTR, INIT_FETCHER, 1));
+
+	data = REG_SET_FIELD(m->cp_hqd_active, CP_HQD_ACTIVE, ACTIVE, 1);
+	WREG32(SOC15_REG_OFFSET(GC, 0, regCP_HQD_ACTIVE), data);
+
+	release_queue(adev);
+
+	return 0;
+}
+
+static int hiq_mqd_load_v11(struct amdgpu_device *adev, void *mqd,
+			      uint32_t pipe_id, uint32_t queue_id,
+			      uint32_t doorbell_off, uint32_t inst)
+{
+	struct amdgpu_ring *kiq_ring = &adev->gfx.kiq[0].ring;
+	struct v11_compute_mqd *m;
+	uint32_t mec, pipe;
+	int r;
+
+	m = get_mqd(mqd);
+
+	acquire_queue(adev, pipe_id, queue_id);
+
+	mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
+	pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);
+
+	pr_debug("kfd: set HIQ, mec:%d, pipe:%d, queue:%d.\n",
+		 mec, pipe, queue_id);
+
+	spin_lock(&adev->gfx.kiq[0].ring_lock);
+	r = amdgpu_ring_alloc(kiq_ring, 7);
+	if (r) {
+		pr_err("Failed to alloc KIQ (%d).\n", r);
+		goto out_unlock;
+	}
+
+	amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_MAP_QUEUES, 5));
+	amdgpu_ring_write(kiq_ring,
+			  PACKET3_MAP_QUEUES_QUEUE_SEL(0) | /* Queue_Sel */
+			  PACKET3_MAP_QUEUES_VMID(m->cp_hqd_vmid) | /* VMID */
+			  PACKET3_MAP_QUEUES_QUEUE(queue_id) |
+			  PACKET3_MAP_QUEUES_PIPE(pipe) |
+			  PACKET3_MAP_QUEUES_ME((mec - 1)) |
+			  PACKET3_MAP_QUEUES_QUEUE_TYPE(0) | /*queue_type: normal compute queue */
+			  PACKET3_MAP_QUEUES_ALLOC_FORMAT(0) | /* alloc format: all_on_one_pipe */
+			  PACKET3_MAP_QUEUES_ENGINE_SEL(1) | /* engine_sel: hiq */
+			  PACKET3_MAP_QUEUES_NUM_QUEUES(1)); /* num_queues: must be 1 */
+	amdgpu_ring_write(kiq_ring,
+			PACKET3_MAP_QUEUES_DOORBELL_OFFSET(doorbell_off));
+	amdgpu_ring_write(kiq_ring, m->cp_mqd_base_addr_lo);
+	amdgpu_ring_write(kiq_ring, m->cp_mqd_base_addr_hi);
+	amdgpu_ring_write(kiq_ring, m->cp_hqd_pq_wptr_poll_addr_lo);
+	amdgpu_ring_write(kiq_ring, m->cp_hqd_pq_wptr_poll_addr_hi);
+	amdgpu_ring_commit(kiq_ring);
+
+out_unlock:
+	spin_unlock(&adev->gfx.kiq[0].ring_lock);
+	release_queue(adev);
+
+	return r;
+}
+
+static int hqd_dump_v11(struct amdgpu_device *adev,
+			uint32_t pipe_id, uint32_t queue_id,
+			uint32_t (**dump)[2], uint32_t *n_regs, uint32_t inst)
+{
+	uint32_t i = 0, reg;
+#define HQD_N_REGS 56
+#define DUMP_REG(addr) do {				\
+		if (WARN_ON_ONCE(i >= HQD_N_REGS))	\
+			break;				\
+		(*dump)[i][0] = (addr) << 2;		\
+		(*dump)[i++][1] = RREG32(addr);		\
+	} while (0)
+
+	*dump = kmalloc_array(HQD_N_REGS, sizeof(**dump), GFP_KERNEL);
+	if (*dump == NULL)
+		return -ENOMEM;
+
+	acquire_queue(adev, pipe_id, queue_id);
+
+	for (reg = SOC15_REG_OFFSET(GC, 0, regCP_MQD_BASE_ADDR);
+	     reg <= SOC15_REG_OFFSET(GC, 0, regCP_HQD_PQ_WPTR_HI); reg++)
+		DUMP_REG(reg);
+
+	release_queue(adev);
+
+	WARN_ON_ONCE(i != HQD_N_REGS);
+	*n_regs = i;
+
+	return 0;
+}
+
+static int hqd_sdma_load_v11(struct amdgpu_device *adev, void *mqd,
+			     uint32_t __user *wptr, struct mm_struct *mm)
+{
+	struct v11_sdma_mqd *m;
+	uint32_t sdma_rlc_reg_offset;
+	unsigned long end_jiffies;
+	uint32_t data;
+	uint64_t data64;
+	uint64_t __user *wptr64 = (uint64_t __user *)wptr;
+
+	m = get_sdma_mqd(mqd);
+	sdma_rlc_reg_offset = get_sdma_rlc_reg_offset(adev, m->sdma_engine_id,
+					    m->sdma_queue_id);
+
+	WREG32(sdma_rlc_reg_offset + regSDMA0_QUEUE0_RB_CNTL,
+		m->sdmax_rlcx_rb_cntl & (~SDMA0_QUEUE0_RB_CNTL__RB_ENABLE_MASK));
+
+	end_jiffies = msecs_to_jiffies(2000) + jiffies;
+	while (true) {
+		data = RREG32(sdma_rlc_reg_offset + regSDMA0_QUEUE0_CONTEXT_STATUS);
+		if (data & SDMA0_QUEUE0_CONTEXT_STATUS__IDLE_MASK)
+			break;
+		if (time_after(jiffies, end_jiffies)) {
+			pr_err("SDMA RLC not idle in %s\n", __func__);
+			return -ETIME;
+		}
+		usleep_range(500, 1000);
+	}
+
+	WREG32(sdma_rlc_reg_offset + regSDMA0_QUEUE0_DOORBELL_OFFSET,
+	       m->sdmax_rlcx_doorbell_offset);
+
+	data = REG_SET_FIELD(m->sdmax_rlcx_doorbell, SDMA0_QUEUE0_DOORBELL,
+			     ENABLE, 1);
+	WREG32(sdma_rlc_reg_offset + regSDMA0_QUEUE0_DOORBELL, data);
+	WREG32(sdma_rlc_reg_offset + regSDMA0_QUEUE0_RB_RPTR,
+				m->sdmax_rlcx_rb_rptr);
+	WREG32(sdma_rlc_reg_offset + regSDMA0_QUEUE0_RB_RPTR_HI,
+				m->sdmax_rlcx_rb_rptr_hi);
+
+	WREG32(sdma_rlc_reg_offset + regSDMA0_QUEUE0_MINOR_PTR_UPDATE, 1);
+	if (read_user_wptr(mm, wptr64, data64)) {
+		WREG32(sdma_rlc_reg_offset + regSDMA0_QUEUE0_RB_WPTR,
+		       lower_32_bits(data64));
+		WREG32(sdma_rlc_reg_offset + regSDMA0_QUEUE0_RB_WPTR_HI,
+		       upper_32_bits(data64));
+	} else {
+		WREG32(sdma_rlc_reg_offset + regSDMA0_QUEUE0_RB_WPTR,
+		       m->sdmax_rlcx_rb_rptr);
+		WREG32(sdma_rlc_reg_offset + regSDMA0_QUEUE0_RB_WPTR_HI,
+		       m->sdmax_rlcx_rb_rptr_hi);
+	}
+	WREG32(sdma_rlc_reg_offset + regSDMA0_QUEUE0_MINOR_PTR_UPDATE, 0);
+
+	WREG32(sdma_rlc_reg_offset + regSDMA0_QUEUE0_RB_BASE, m->sdmax_rlcx_rb_base);
+	WREG32(sdma_rlc_reg_offset + regSDMA0_QUEUE0_RB_BASE_HI,
+			m->sdmax_rlcx_rb_base_hi);
+	WREG32(sdma_rlc_reg_offset + regSDMA0_QUEUE0_RB_RPTR_ADDR_LO,
+			m->sdmax_rlcx_rb_rptr_addr_lo);
+	WREG32(sdma_rlc_reg_offset + regSDMA0_QUEUE0_RB_RPTR_ADDR_HI,
+			m->sdmax_rlcx_rb_rptr_addr_hi);
+
+	data = REG_SET_FIELD(m->sdmax_rlcx_rb_cntl, SDMA0_QUEUE0_RB_CNTL,
+			     RB_ENABLE, 1);
+	WREG32(sdma_rlc_reg_offset + regSDMA0_QUEUE0_RB_CNTL, data);
+
+	return 0;
+}
+
+static int hqd_sdma_dump_v11(struct amdgpu_device *adev,
+			     uint32_t engine_id, uint32_t queue_id,
+			     uint32_t (**dump)[2], uint32_t *n_regs)
+{
+	uint32_t sdma_rlc_reg_offset = get_sdma_rlc_reg_offset(adev,
+			engine_id, queue_id);
+	uint32_t i = 0, reg;
+#undef HQD_N_REGS
+#define HQD_N_REGS (7+11+1+12+12)
+
+	*dump = kmalloc_array(HQD_N_REGS, sizeof(**dump), GFP_KERNEL);
+	if (*dump == NULL)
+		return -ENOMEM;
+
+	for (reg = regSDMA0_QUEUE0_RB_CNTL;
+	     reg <= regSDMA0_QUEUE0_RB_WPTR_HI; reg++)
+		DUMP_REG(sdma_rlc_reg_offset + reg);
+	for (reg = regSDMA0_QUEUE0_RB_RPTR_ADDR_HI;
+	     reg <= regSDMA0_QUEUE0_DOORBELL; reg++)
+		DUMP_REG(sdma_rlc_reg_offset + reg);
+	for (reg = regSDMA0_QUEUE0_DOORBELL_LOG;
+	     reg <= regSDMA0_QUEUE0_DOORBELL_LOG; reg++)
+		DUMP_REG(sdma_rlc_reg_offset + reg);
+	for (reg = regSDMA0_QUEUE0_DOORBELL_OFFSET;
+	     reg <= regSDMA0_QUEUE0_RB_PREEMPT; reg++)
+		DUMP_REG(sdma_rlc_reg_offset + reg);
+	for (reg = regSDMA0_QUEUE0_MIDCMD_DATA0;
+	     reg <= regSDMA0_QUEUE0_MIDCMD_CNTL; reg++)
+		DUMP_REG(sdma_rlc_reg_offset + reg);
+
+	WARN_ON_ONCE(i != HQD_N_REGS);
+	*n_regs = i;
+
+	return 0;
+}
+
+static bool hqd_is_occupied_v11(struct amdgpu_device *adev, uint64_t queue_address,
+				uint32_t pipe_id, uint32_t queue_id, uint32_t inst)
+{
+	uint32_t act;
+	bool retval = false;
+	uint32_t low, high;
+
+	acquire_queue(adev, pipe_id, queue_id);
+	act = RREG32(SOC15_REG_OFFSET(GC, 0, regCP_HQD_ACTIVE));
+	if (act) {
+		low = lower_32_bits(queue_address >> 8);
+		high = upper_32_bits(queue_address >> 8);
+
+		if (low == RREG32(SOC15_REG_OFFSET(GC, 0, regCP_HQD_PQ_BASE)) &&
+		   high == RREG32(SOC15_REG_OFFSET(GC, 0, regCP_HQD_PQ_BASE_HI)))
+			retval = true;
+	}
+	release_queue(adev);
+	return retval;
+}
+
+static bool hqd_sdma_is_occupied_v11(struct amdgpu_device *adev, void *mqd)
+{
+	struct v11_sdma_mqd *m;
+	uint32_t sdma_rlc_reg_offset;
+	uint32_t sdma_rlc_rb_cntl;
+
+	m = get_sdma_mqd(mqd);
+	sdma_rlc_reg_offset = get_sdma_rlc_reg_offset(adev, m->sdma_engine_id,
+					    m->sdma_queue_id);
+
+	sdma_rlc_rb_cntl = RREG32(sdma_rlc_reg_offset + regSDMA0_QUEUE0_RB_CNTL);
+
+	if (sdma_rlc_rb_cntl & SDMA0_QUEUE0_RB_CNTL__RB_ENABLE_MASK)
+		return true;
+
+	return false;
+}
+
+static int hqd_destroy_v11(struct amdgpu_device *adev, void *mqd,
+				enum kfd_preempt_type reset_type,
+				unsigned int utimeout, uint32_t pipe_id,
+				uint32_t queue_id, uint32_t inst)
+{
+	enum hqd_dequeue_request_type type;
+	unsigned long end_jiffies;
+	uint32_t temp;
+	struct v11_compute_mqd *m = get_mqd(mqd);
+
+	acquire_queue(adev, pipe_id, queue_id);
+
+	if (m->cp_hqd_vmid == 0)
+		WREG32_FIELD15_PREREG(GC, 0, RLC_CP_SCHEDULERS, scheduler1, 0);
+
+	switch (reset_type) {
+	case KFD_PREEMPT_TYPE_WAVEFRONT_DRAIN:
+		type = DRAIN_PIPE;
+		break;
+	case KFD_PREEMPT_TYPE_WAVEFRONT_RESET:
+		type = RESET_WAVES;
+		break;
+	default:
+		type = DRAIN_PIPE;
+		break;
+	}
+
+	WREG32(SOC15_REG_OFFSET(GC, 0, regCP_HQD_DEQUEUE_REQUEST), type);
+
+	end_jiffies = (utimeout * HZ / 1000) + jiffies;
+	while (true) {
+		temp = RREG32(SOC15_REG_OFFSET(GC, 0, regCP_HQD_ACTIVE));
+		if (!(temp & CP_HQD_ACTIVE__ACTIVE_MASK))
+			break;
+		if (time_after(jiffies, end_jiffies)) {
+			pr_err("cp queue pipe %d queue %d preemption failed\n",
+					pipe_id, queue_id);
+			release_queue(adev);
+			return -ETIME;
+		}
+		usleep_range(500, 1000);
+	}
+
+	release_queue(adev);
+	return 0;
+}
+
+static int hqd_sdma_destroy_v11(struct amdgpu_device *adev, void *mqd,
+				unsigned int utimeout)
+{
+	struct v11_sdma_mqd *m;
+	uint32_t sdma_rlc_reg_offset;
+	uint32_t temp;
+	unsigned long end_jiffies = (utimeout * HZ / 1000) + jiffies;
+
+	m = get_sdma_mqd(mqd);
+	sdma_rlc_reg_offset = get_sdma_rlc_reg_offset(adev, m->sdma_engine_id,
+					    m->sdma_queue_id);
+
+	temp = RREG32(sdma_rlc_reg_offset + regSDMA0_QUEUE0_RB_CNTL);
+	temp = temp & ~SDMA0_QUEUE0_RB_CNTL__RB_ENABLE_MASK;
+	WREG32(sdma_rlc_reg_offset + regSDMA0_QUEUE0_RB_CNTL, temp);
+
+	while (true) {
+		temp = RREG32(sdma_rlc_reg_offset + regSDMA0_QUEUE0_CONTEXT_STATUS);
+		if (temp & SDMA0_QUEUE0_CONTEXT_STATUS__IDLE_MASK)
+			break;
+		if (time_after(jiffies, end_jiffies)) {
+			pr_err("SDMA RLC not idle in %s\n", __func__);
+			return -ETIME;
+		}
+		usleep_range(500, 1000);
+	}
+
+	WREG32(sdma_rlc_reg_offset + regSDMA0_QUEUE0_DOORBELL, 0);
+	WREG32(sdma_rlc_reg_offset + regSDMA0_QUEUE0_RB_CNTL,
+		RREG32(sdma_rlc_reg_offset + regSDMA0_QUEUE0_RB_CNTL) |
+		SDMA0_QUEUE0_RB_CNTL__RB_ENABLE_MASK);
+
+	m->sdmax_rlcx_rb_rptr = RREG32(sdma_rlc_reg_offset + regSDMA0_QUEUE0_RB_RPTR);
+	m->sdmax_rlcx_rb_rptr_hi =
+		RREG32(sdma_rlc_reg_offset + regSDMA0_QUEUE0_RB_RPTR_HI);
+
+	return 0;
+}
+
+static int wave_control_execute_v11(struct amdgpu_device *adev,
+					uint32_t gfx_index_val,
+					uint32_t sq_cmd, uint32_t inst)
+{
+	uint32_t data = 0;
+
+	mutex_lock(&adev->grbm_idx_mutex);
+
+	WREG32(SOC15_REG_OFFSET(GC, 0, regGRBM_GFX_INDEX), gfx_index_val);
+	WREG32(SOC15_REG_OFFSET(GC, 0, regSQ_CMD), sq_cmd);
+
+	data = REG_SET_FIELD(data, GRBM_GFX_INDEX,
+		INSTANCE_BROADCAST_WRITES, 1);
+	data = REG_SET_FIELD(data, GRBM_GFX_INDEX,
+		SA_BROADCAST_WRITES, 1);
+	data = REG_SET_FIELD(data, GRBM_GFX_INDEX,
+		SE_BROADCAST_WRITES, 1);
+
+	WREG32(SOC15_REG_OFFSET(GC, 0, regGRBM_GFX_INDEX), data);
+	mutex_unlock(&adev->grbm_idx_mutex);
+
+	return 0;
+}
+
+static void set_vm_context_page_table_base_v11(struct amdgpu_device *adev,
+		uint32_t vmid, uint64_t page_table_base)
+{
+	if (!amdgpu_amdkfd_is_kfd_vmid(adev, vmid)) {
+		pr_err("trying to set page table base for wrong VMID %u\n",
+		       vmid);
+		return;
+	}
+
+	/* SDMA is on gfxhub as well for gfx11 adapters */
+	adev->gfxhub.funcs->setup_vm_pt_regs(adev, vmid, page_table_base);
+}
+
+/*
+ * Returns TRAP_EN, EXCP_EN and EXCP_REPLACE.
+ *
+ * restore_dbg_registers is ignored here but is a general interface requirement
+ * for devices that support GFXOFF and where the RLC save/restore list
+ * does not support hw registers for debugging i.e. the driver has to manually
+ * initialize the debug mode registers after it has disabled GFX off during the
+ * debug session.
+ */
+static uint32_t kgd_gfx_v11_enable_debug_trap(struct amdgpu_device *adev,
+					    bool restore_dbg_registers,
+					    uint32_t vmid)
+{
+	uint32_t data = 0;
+
+	data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, TRAP_EN, 1);
+	data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, EXCP_EN, 0);
+	data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, EXCP_REPLACE, 0);
+
+	return data;
+}
+
+/* Returns TRAP_EN, EXCP_EN and EXCP_REPLACE. */
+static uint32_t kgd_gfx_v11_disable_debug_trap(struct amdgpu_device *adev,
+						bool keep_trap_enabled,
+						uint32_t vmid)
+{
+	uint32_t data = 0;
+
+	data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, TRAP_EN, 1);
+	data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, EXCP_EN, 0);
+	data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, EXCP_REPLACE, 0);
+
+	return data;
+}
+
+static int kgd_gfx_v11_validate_trap_override_request(struct amdgpu_device *adev,
+							uint32_t trap_override,
+							uint32_t *trap_mask_supported)
+{
+	*trap_mask_supported &= KFD_DBG_TRAP_MASK_FP_INVALID |
+				KFD_DBG_TRAP_MASK_FP_INPUT_DENORMAL |
+				KFD_DBG_TRAP_MASK_FP_DIVIDE_BY_ZERO |
+				KFD_DBG_TRAP_MASK_FP_OVERFLOW |
+				KFD_DBG_TRAP_MASK_FP_UNDERFLOW |
+				KFD_DBG_TRAP_MASK_FP_INEXACT |
+				KFD_DBG_TRAP_MASK_INT_DIVIDE_BY_ZERO |
+				KFD_DBG_TRAP_MASK_DBG_ADDRESS_WATCH |
+				KFD_DBG_TRAP_MASK_DBG_MEMORY_VIOLATION;
+
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) >= IP_VERSION(11, 0, 4))
+		*trap_mask_supported |= KFD_DBG_TRAP_MASK_TRAP_ON_WAVE_START |
+					KFD_DBG_TRAP_MASK_TRAP_ON_WAVE_END;
+
+	if (trap_override != KFD_DBG_TRAP_OVERRIDE_OR &&
+			trap_override != KFD_DBG_TRAP_OVERRIDE_REPLACE)
+		return -EPERM;
+
+	return 0;
+}
+
+static uint32_t trap_mask_map_sw_to_hw(uint32_t mask)
+{
+	uint32_t trap_on_start = (mask & KFD_DBG_TRAP_MASK_TRAP_ON_WAVE_START) ? 1 : 0;
+	uint32_t trap_on_end = (mask & KFD_DBG_TRAP_MASK_TRAP_ON_WAVE_END) ? 1 : 0;
+	uint32_t excp_en = mask & (KFD_DBG_TRAP_MASK_FP_INVALID |
+			KFD_DBG_TRAP_MASK_FP_INPUT_DENORMAL |
+			KFD_DBG_TRAP_MASK_FP_DIVIDE_BY_ZERO |
+			KFD_DBG_TRAP_MASK_FP_OVERFLOW |
+			KFD_DBG_TRAP_MASK_FP_UNDERFLOW |
+			KFD_DBG_TRAP_MASK_FP_INEXACT |
+			KFD_DBG_TRAP_MASK_INT_DIVIDE_BY_ZERO |
+			KFD_DBG_TRAP_MASK_DBG_ADDRESS_WATCH |
+			KFD_DBG_TRAP_MASK_DBG_MEMORY_VIOLATION);
+	uint32_t ret;
+
+	ret = REG_SET_FIELD(0, SPI_GDBG_PER_VMID_CNTL, EXCP_EN, excp_en);
+	ret = REG_SET_FIELD(ret, SPI_GDBG_PER_VMID_CNTL, TRAP_ON_START, trap_on_start);
+	ret = REG_SET_FIELD(ret, SPI_GDBG_PER_VMID_CNTL, TRAP_ON_END, trap_on_end);
+
+	return ret;
+}
+
+static uint32_t trap_mask_map_hw_to_sw(uint32_t mask)
+{
+	uint32_t ret = REG_GET_FIELD(mask, SPI_GDBG_PER_VMID_CNTL, EXCP_EN);
+
+	if (REG_GET_FIELD(mask, SPI_GDBG_PER_VMID_CNTL, TRAP_ON_START))
+		ret |= KFD_DBG_TRAP_MASK_TRAP_ON_WAVE_START;
+
+	if (REG_GET_FIELD(mask, SPI_GDBG_PER_VMID_CNTL, TRAP_ON_END))
+		ret |= KFD_DBG_TRAP_MASK_TRAP_ON_WAVE_END;
+
+	return ret;
+}
+
+/* Returns TRAP_EN, EXCP_EN and EXCP_REPLACE. */
+static uint32_t kgd_gfx_v11_set_wave_launch_trap_override(struct amdgpu_device *adev,
+					uint32_t vmid,
+					uint32_t trap_override,
+					uint32_t trap_mask_bits,
+					uint32_t trap_mask_request,
+					uint32_t *trap_mask_prev,
+					uint32_t kfd_dbg_trap_cntl_prev)
+{
+	uint32_t data = 0;
+
+	*trap_mask_prev = trap_mask_map_hw_to_sw(kfd_dbg_trap_cntl_prev);
+
+	data = (trap_mask_bits & trap_mask_request) | (*trap_mask_prev & ~trap_mask_request);
+	data = trap_mask_map_sw_to_hw(data);
+
+	data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, TRAP_EN, 1);
+	data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, EXCP_REPLACE, trap_override);
+
+	return data;
+}
+
+static uint32_t kgd_gfx_v11_set_wave_launch_mode(struct amdgpu_device *adev,
+					uint8_t wave_launch_mode,
+					uint32_t vmid)
+{
+	uint32_t data = 0;
+
+	data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, LAUNCH_MODE, wave_launch_mode);
+
+	return data;
+}
+
+#define TCP_WATCH_STRIDE (regTCP_WATCH1_ADDR_H - regTCP_WATCH0_ADDR_H)
+static uint32_t kgd_gfx_v11_set_address_watch(struct amdgpu_device *adev,
+					uint64_t watch_address,
+					uint32_t watch_address_mask,
+					uint32_t watch_id,
+					uint32_t watch_mode,
+					uint32_t debug_vmid,
+					uint32_t inst)
+{
+	uint32_t watch_address_high;
+	uint32_t watch_address_low;
+	uint32_t watch_address_cntl;
+
+	watch_address_cntl = 0;
+	watch_address_low = lower_32_bits(watch_address);
+	watch_address_high = upper_32_bits(watch_address) & 0xffff;
+
+	watch_address_cntl = REG_SET_FIELD(watch_address_cntl,
+			TCP_WATCH0_CNTL,
+			MODE,
+			watch_mode);
+
+	watch_address_cntl = REG_SET_FIELD(watch_address_cntl,
+			TCP_WATCH0_CNTL,
+			MASK,
+			watch_address_mask >> 7);
+
+	watch_address_cntl = REG_SET_FIELD(watch_address_cntl,
+			TCP_WATCH0_CNTL,
+			VALID,
+			1);
+
+	WREG32_RLC((SOC15_REG_OFFSET(GC, 0, regTCP_WATCH0_ADDR_H) +
+			(watch_id * TCP_WATCH_STRIDE)),
+			watch_address_high);
+
+	WREG32_RLC((SOC15_REG_OFFSET(GC, 0, regTCP_WATCH0_ADDR_L) +
+			(watch_id * TCP_WATCH_STRIDE)),
+			watch_address_low);
+
+	return watch_address_cntl;
+}
+
+static uint32_t kgd_gfx_v11_clear_address_watch(struct amdgpu_device *adev,
+						uint32_t watch_id)
+{
+	return 0;
+}
+
+static uint64_t kgd_gfx_v11_hqd_get_pq_addr(struct amdgpu_device *adev,
+					    uint32_t pipe_id, uint32_t queue_id,
+					    uint32_t inst)
+{
+	return 0;
+}
+
+static uint64_t kgd_gfx_v11_hqd_reset(struct amdgpu_device *adev,
+				      uint32_t pipe_id, uint32_t queue_id,
+				      uint32_t inst, unsigned int utimeout)
+{
+	return 0;
+}
+
+static uint32_t kgd_gfx_v11_hqd_sdma_get_doorbell(struct amdgpu_device *adev,
+						  int engine, int queue)
+{
+	return 0;
+}
+
+const struct kfd2kgd_calls gfx_v11_kfd2kgd = {
+	.program_sh_mem_settings = program_sh_mem_settings_v11,
+	.set_pasid_vmid_mapping = set_pasid_vmid_mapping_v11,
+	.init_interrupts = init_interrupts_v11,
+	.hqd_load = hqd_load_v11,
+	.hiq_mqd_load = hiq_mqd_load_v11,
+	.hqd_sdma_load = hqd_sdma_load_v11,
+	.hqd_dump = hqd_dump_v11,
+	.hqd_sdma_dump = hqd_sdma_dump_v11,
+	.hqd_is_occupied = hqd_is_occupied_v11,
+	.hqd_sdma_is_occupied = hqd_sdma_is_occupied_v11,
+	.hqd_destroy = hqd_destroy_v11,
+	.hqd_sdma_destroy = hqd_sdma_destroy_v11,
+	.wave_control_execute = wave_control_execute_v11,
+	.get_atc_vmid_pasid_mapping_info = NULL,
+	.set_vm_context_page_table_base = set_vm_context_page_table_base_v11,
+	.enable_debug_trap = kgd_gfx_v11_enable_debug_trap,
+	.disable_debug_trap = kgd_gfx_v11_disable_debug_trap,
+	.validate_trap_override_request = kgd_gfx_v11_validate_trap_override_request,
+	.set_wave_launch_trap_override = kgd_gfx_v11_set_wave_launch_trap_override,
+	.set_wave_launch_mode = kgd_gfx_v11_set_wave_launch_mode,
+	.set_address_watch = kgd_gfx_v11_set_address_watch,
+	.clear_address_watch = kgd_gfx_v11_clear_address_watch,
+	.hqd_get_pq_addr = kgd_gfx_v11_hqd_get_pq_addr,
+	.hqd_reset = kgd_gfx_v11_hqd_reset,
+	.hqd_sdma_get_doorbell = kgd_gfx_v11_hqd_sdma_get_doorbell
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v12.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v12.c
new file mode 100644
index 000000000000..e0ceab400b2d
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v12.c
@@ -0,0 +1,384 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#include "amdgpu.h"
+#include "amdgpu_amdkfd.h"
+#include "gc/gc_12_0_0_offset.h"
+#include "gc/gc_12_0_0_sh_mask.h"
+#include "soc24.h"
+#include <uapi/linux/kfd_ioctl.h>
+
+static void lock_srbm(struct amdgpu_device *adev, uint32_t mec, uint32_t pipe,
+			uint32_t queue, uint32_t vmid)
+{
+	mutex_lock(&adev->srbm_mutex);
+	soc24_grbm_select(adev, mec, pipe, queue, vmid);
+}
+
+static void unlock_srbm(struct amdgpu_device *adev)
+{
+	soc24_grbm_select(adev, 0, 0, 0, 0);
+	mutex_unlock(&adev->srbm_mutex);
+}
+
+static void acquire_queue(struct amdgpu_device *adev, uint32_t pipe_id,
+				uint32_t queue_id)
+{
+	uint32_t mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
+	uint32_t pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);
+
+	lock_srbm(adev, mec, pipe, queue_id, 0);
+}
+
+static void release_queue(struct amdgpu_device *adev)
+{
+	unlock_srbm(adev);
+}
+
+static int init_interrupts_v12(struct amdgpu_device *adev, uint32_t pipe_id, uint32_t inst)
+{
+	uint32_t mec;
+	uint32_t pipe;
+
+	mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
+	pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);
+
+	lock_srbm(adev, mec, pipe, 0, 0);
+
+	WREG32_SOC15(GC, 0, regCPC_INT_CNTL,
+		CP_INT_CNTL_RING0__TIME_STAMP_INT_ENABLE_MASK |
+		CP_INT_CNTL_RING0__OPCODE_ERROR_INT_ENABLE_MASK);
+
+	unlock_srbm(adev);
+
+	return 0;
+}
+
+static uint32_t get_sdma_rlc_reg_offset(struct amdgpu_device *adev,
+				unsigned int engine_id,
+				unsigned int queue_id)
+{
+	uint32_t sdma_engine_reg_base = 0;
+	uint32_t sdma_rlc_reg_offset;
+
+	switch (engine_id) {
+	case 0:
+		sdma_engine_reg_base = SOC15_REG_OFFSET(SDMA0, 0,
+				regSDMA0_QUEUE0_RB_CNTL) - regSDMA0_QUEUE0_RB_CNTL;
+		break;
+	case 1:
+		sdma_engine_reg_base = SOC15_REG_OFFSET(SDMA1, 0,
+				regSDMA1_QUEUE0_RB_CNTL) - regSDMA0_QUEUE0_RB_CNTL;
+		break;
+	default:
+		BUG();
+	}
+
+	sdma_rlc_reg_offset = sdma_engine_reg_base
+		+ queue_id * (regSDMA0_QUEUE1_RB_CNTL - regSDMA0_QUEUE0_RB_CNTL);
+
+	pr_debug("RLC register offset for SDMA%d RLC%d: 0x%x\n", engine_id,
+			queue_id, sdma_rlc_reg_offset);
+
+	return sdma_rlc_reg_offset;
+}
+
+static int hqd_dump_v12(struct amdgpu_device *adev,
+			uint32_t pipe_id, uint32_t queue_id,
+			uint32_t (**dump)[2], uint32_t *n_regs, uint32_t inst)
+{
+	uint32_t i = 0, reg;
+#define HQD_N_REGS 56
+#define DUMP_REG(addr) do {				\
+		if (WARN_ON_ONCE(i >= HQD_N_REGS))	\
+			break;				\
+		(*dump)[i][0] = (addr) << 2;		\
+		(*dump)[i++][1] = RREG32(addr);		\
+	} while (0)
+
+	*dump = kmalloc_array(HQD_N_REGS, sizeof(**dump), GFP_KERNEL);
+	if (*dump == NULL)
+		return -ENOMEM;
+
+	acquire_queue(adev, pipe_id, queue_id);
+
+	for (reg = SOC15_REG_OFFSET(GC, 0, regCP_MQD_BASE_ADDR);
+	     reg <= SOC15_REG_OFFSET(GC, 0, regCP_HQD_PQ_WPTR_HI); reg++)
+		DUMP_REG(reg);
+
+	release_queue(adev);
+
+	WARN_ON_ONCE(i != HQD_N_REGS);
+	*n_regs = i;
+
+	return 0;
+}
+
+static int hqd_sdma_dump_v12(struct amdgpu_device *adev,
+			     uint32_t engine_id, uint32_t queue_id,
+			     uint32_t (**dump)[2], uint32_t *n_regs)
+{
+	uint32_t sdma_rlc_reg_offset = get_sdma_rlc_reg_offset(adev,
+			engine_id, queue_id);
+	uint32_t i = 0, reg;
+
+	const uint32_t first_reg = regSDMA0_QUEUE0_RB_CNTL;
+	const uint32_t last_reg = regSDMA0_QUEUE0_CONTEXT_STATUS;
+#undef HQD_N_REGS
+#define HQD_N_REGS (last_reg - first_reg + 1)
+
+	*dump = kmalloc_array(HQD_N_REGS, sizeof(**dump), GFP_KERNEL);
+	if (*dump == NULL)
+		return -ENOMEM;
+
+	for (reg = first_reg;
+	     reg <= last_reg; reg++)
+		DUMP_REG(sdma_rlc_reg_offset + reg);
+
+	WARN_ON_ONCE(i != HQD_N_REGS);
+	*n_regs = i;
+
+	return 0;
+}
+
+static int wave_control_execute_v12(struct amdgpu_device *adev,
+					uint32_t gfx_index_val,
+					uint32_t sq_cmd, uint32_t inst)
+{
+	uint32_t data = 0;
+
+	mutex_lock(&adev->grbm_idx_mutex);
+
+	WREG32(SOC15_REG_OFFSET(GC, 0, regGRBM_GFX_INDEX), gfx_index_val);
+	WREG32(SOC15_REG_OFFSET(GC, 0, regSQ_CMD), sq_cmd);
+
+	data = REG_SET_FIELD(data, GRBM_GFX_INDEX,
+		INSTANCE_BROADCAST_WRITES, 1);
+	data = REG_SET_FIELD(data, GRBM_GFX_INDEX,
+		SA_BROADCAST_WRITES, 1);
+	data = REG_SET_FIELD(data, GRBM_GFX_INDEX,
+		SE_BROADCAST_WRITES, 1);
+
+	WREG32(SOC15_REG_OFFSET(GC, 0, regGRBM_GFX_INDEX), data);
+	mutex_unlock(&adev->grbm_idx_mutex);
+
+	return 0;
+}
+
+/* returns TRAP_EN, EXCP_EN and EXCP_REPLACE. */
+static uint32_t kgd_gfx_v12_enable_debug_trap(struct amdgpu_device *adev,
+					    bool restore_dbg_registers,
+					    uint32_t vmid)
+{
+	uint32_t data = 0;
+
+	data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, TRAP_EN, 1);
+	data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, EXCP_EN, 0);
+	data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, EXCP_REPLACE, 0);
+
+	return data;
+}
+
+/* returns TRAP_EN, EXCP_EN and EXCP_REPLACE. */
+static uint32_t kgd_gfx_v12_disable_debug_trap(struct amdgpu_device *adev,
+						bool keep_trap_enabled,
+						uint32_t vmid)
+{
+	uint32_t data = 0;
+
+	data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, TRAP_EN, 1);
+	data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, EXCP_EN, 0);
+	data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, EXCP_REPLACE, 0);
+
+	return data;
+}
+
+static int kgd_gfx_v12_validate_trap_override_request(struct amdgpu_device *adev,
+							uint32_t trap_override,
+							uint32_t *trap_mask_supported)
+{
+	*trap_mask_supported &= KFD_DBG_TRAP_MASK_FP_INVALID |
+				KFD_DBG_TRAP_MASK_FP_INPUT_DENORMAL |
+				KFD_DBG_TRAP_MASK_FP_DIVIDE_BY_ZERO |
+				KFD_DBG_TRAP_MASK_FP_OVERFLOW |
+				KFD_DBG_TRAP_MASK_FP_UNDERFLOW |
+				KFD_DBG_TRAP_MASK_FP_INEXACT |
+				KFD_DBG_TRAP_MASK_INT_DIVIDE_BY_ZERO |
+				KFD_DBG_TRAP_MASK_DBG_ADDRESS_WATCH |
+				KFD_DBG_TRAP_MASK_DBG_MEMORY_VIOLATION |
+				KFD_DBG_TRAP_MASK_TRAP_ON_WAVE_START |
+				KFD_DBG_TRAP_MASK_TRAP_ON_WAVE_END;
+
+
+	if (trap_override != KFD_DBG_TRAP_OVERRIDE_OR &&
+			trap_override != KFD_DBG_TRAP_OVERRIDE_REPLACE)
+		return -EPERM;
+
+	return 0;
+}
+
+static uint32_t trap_mask_map_sw_to_hw(uint32_t mask)
+{
+	uint32_t trap_on_start = (mask & KFD_DBG_TRAP_MASK_TRAP_ON_WAVE_START) ? 1 : 0;
+	uint32_t trap_on_end = (mask & KFD_DBG_TRAP_MASK_TRAP_ON_WAVE_END) ? 1 : 0;
+	uint32_t excp_en = mask & (KFD_DBG_TRAP_MASK_FP_INVALID |
+			KFD_DBG_TRAP_MASK_FP_INPUT_DENORMAL |
+			KFD_DBG_TRAP_MASK_FP_DIVIDE_BY_ZERO |
+			KFD_DBG_TRAP_MASK_FP_OVERFLOW |
+			KFD_DBG_TRAP_MASK_FP_UNDERFLOW |
+			KFD_DBG_TRAP_MASK_FP_INEXACT |
+			KFD_DBG_TRAP_MASK_INT_DIVIDE_BY_ZERO |
+			KFD_DBG_TRAP_MASK_DBG_ADDRESS_WATCH |
+			KFD_DBG_TRAP_MASK_DBG_MEMORY_VIOLATION);
+	uint32_t ret;
+
+	ret = REG_SET_FIELD(0, SPI_GDBG_PER_VMID_CNTL, EXCP_EN, excp_en);
+	ret = REG_SET_FIELD(ret, SPI_GDBG_PER_VMID_CNTL, TRAP_ON_START, trap_on_start);
+	ret = REG_SET_FIELD(ret, SPI_GDBG_PER_VMID_CNTL, TRAP_ON_END, trap_on_end);
+
+	return ret;
+}
+
+static uint32_t trap_mask_map_hw_to_sw(uint32_t mask)
+{
+	uint32_t ret = REG_GET_FIELD(mask, SPI_GDBG_PER_VMID_CNTL, EXCP_EN);
+
+	if (REG_GET_FIELD(mask, SPI_GDBG_PER_VMID_CNTL, TRAP_ON_START))
+		ret |= KFD_DBG_TRAP_MASK_TRAP_ON_WAVE_START;
+
+	if (REG_GET_FIELD(mask, SPI_GDBG_PER_VMID_CNTL, TRAP_ON_END))
+		ret |= KFD_DBG_TRAP_MASK_TRAP_ON_WAVE_END;
+
+	return ret;
+}
+
+/* returns TRAP_EN, EXCP_EN and EXCP_REPLACE. */
+static uint32_t kgd_gfx_v12_set_wave_launch_trap_override(struct amdgpu_device *adev,
+					uint32_t vmid,
+					uint32_t trap_override,
+					uint32_t trap_mask_bits,
+					uint32_t trap_mask_request,
+					uint32_t *trap_mask_prev,
+					uint32_t kfd_dbg_trap_cntl_prev)
+
+{
+	uint32_t data = 0;
+
+	*trap_mask_prev = trap_mask_map_hw_to_sw(kfd_dbg_trap_cntl_prev);
+
+	data = (trap_mask_bits & trap_mask_request) | (*trap_mask_prev & ~trap_mask_request);
+	data = trap_mask_map_sw_to_hw(data);
+
+	data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, TRAP_EN, 1);
+	data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, EXCP_REPLACE, trap_override);
+
+	return data;
+}
+
+/* returns STALL_VMID or LAUNCH_MODE. */
+static uint32_t kgd_gfx_v12_set_wave_launch_mode(struct amdgpu_device *adev,
+					uint8_t wave_launch_mode,
+					uint32_t vmid)
+{
+	uint32_t data = 0;
+	bool is_stall_mode = wave_launch_mode == 4;
+
+	if (is_stall_mode)
+		data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, STALL_VMID,
+									1);
+	else
+		data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, LAUNCH_MODE,
+							wave_launch_mode);
+
+	return data;
+}
+
+#define TCP_WATCH_STRIDE (regTCP_WATCH1_ADDR_H - regTCP_WATCH0_ADDR_H)
+static uint32_t kgd_gfx_v12_set_address_watch(struct amdgpu_device *adev,
+					uint64_t watch_address,
+					uint32_t watch_address_mask,
+					uint32_t watch_id,
+					uint32_t watch_mode,
+					uint32_t debug_vmid,
+					uint32_t inst)
+{
+	uint32_t watch_address_high;
+	uint32_t watch_address_low;
+	uint32_t watch_address_cntl;
+
+	watch_address_cntl = 0;
+	watch_address_low = lower_32_bits(watch_address);
+	watch_address_high = upper_32_bits(watch_address) & 0xffff;
+
+	watch_address_cntl = REG_SET_FIELD(watch_address_cntl,
+			TCP_WATCH0_CNTL,
+			MODE,
+			watch_mode);
+
+	watch_address_cntl = REG_SET_FIELD(watch_address_cntl,
+			TCP_WATCH0_CNTL,
+			MASK,
+			watch_address_mask >> 7);
+
+	watch_address_cntl = REG_SET_FIELD(watch_address_cntl,
+			TCP_WATCH0_CNTL,
+			VALID,
+			1);
+
+	WREG32_RLC((SOC15_REG_OFFSET(GC, 0, regTCP_WATCH0_ADDR_H) +
+			(watch_id * TCP_WATCH_STRIDE)),
+			watch_address_high);
+
+	WREG32_RLC((SOC15_REG_OFFSET(GC, 0, regTCP_WATCH0_ADDR_L) +
+			(watch_id * TCP_WATCH_STRIDE)),
+			watch_address_low);
+
+	return watch_address_cntl;
+}
+
+static uint32_t kgd_gfx_v12_clear_address_watch(struct amdgpu_device *adev,
+					uint32_t watch_id)
+{
+	return 0;
+}
+
+static uint32_t kgd_gfx_v12_hqd_sdma_get_doorbell(struct amdgpu_device *adev,
+						 int engine, int queue)
+{
+	return 0;
+}
+
+const struct kfd2kgd_calls gfx_v12_kfd2kgd = {
+	.init_interrupts = init_interrupts_v12,
+	.hqd_dump = hqd_dump_v12,
+	.hqd_sdma_dump = hqd_sdma_dump_v12,
+	.wave_control_execute = wave_control_execute_v12,
+	.get_atc_vmid_pasid_mapping_info = NULL,
+	.enable_debug_trap = kgd_gfx_v12_enable_debug_trap,
+	.disable_debug_trap = kgd_gfx_v12_disable_debug_trap,
+	.validate_trap_override_request = kgd_gfx_v12_validate_trap_override_request,
+	.set_wave_launch_trap_override = kgd_gfx_v12_set_wave_launch_trap_override,
+	.set_wave_launch_mode = kgd_gfx_v12_set_wave_launch_mode,
+	.set_address_watch = kgd_gfx_v12_set_address_watch,
+	.clear_address_watch = kgd_gfx_v12_clear_address_watch,
+	.hqd_sdma_get_doorbell = kgd_gfx_v12_hqd_sdma_get_doorbell
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v7.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v7.c
index b91d27e39bad..df77558e03ef 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v7.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v7.c
@@ -45,105 +45,54 @@ enum {
 	MAX_WATCH_ADDRESSES = 4
 };
 
-enum {
-	ADDRESS_WATCH_REG_ADDR_HI = 0,
-	ADDRESS_WATCH_REG_ADDR_LO,
-	ADDRESS_WATCH_REG_CNTL,
-	ADDRESS_WATCH_REG_MAX
-};
-
-/*  not defined in the CI/KV reg file  */
-enum {
-	ADDRESS_WATCH_REG_CNTL_ATC_BIT = 0x10000000UL,
-	ADDRESS_WATCH_REG_CNTL_DEFAULT_MASK = 0x00FFFFFF,
-	ADDRESS_WATCH_REG_ADDLOW_MASK_EXTENSION = 0x03000000,
-	/* extend the mask to 26 bits to match the low address field */
-	ADDRESS_WATCH_REG_ADDLOW_SHIFT = 6,
-	ADDRESS_WATCH_REG_ADDHIGH_MASK = 0xFFFF
-};
-
-static const uint32_t watchRegs[MAX_WATCH_ADDRESSES * ADDRESS_WATCH_REG_MAX] = {
-	mmTCP_WATCH0_ADDR_H, mmTCP_WATCH0_ADDR_L, mmTCP_WATCH0_CNTL,
-	mmTCP_WATCH1_ADDR_H, mmTCP_WATCH1_ADDR_L, mmTCP_WATCH1_CNTL,
-	mmTCP_WATCH2_ADDR_H, mmTCP_WATCH2_ADDR_L, mmTCP_WATCH2_CNTL,
-	mmTCP_WATCH3_ADDR_H, mmTCP_WATCH3_ADDR_L, mmTCP_WATCH3_CNTL
-};
-
-union TCP_WATCH_CNTL_BITS {
-	struct {
-		uint32_t mask:24;
-		uint32_t vmid:4;
-		uint32_t atc:1;
-		uint32_t mode:2;
-		uint32_t valid:1;
-	} bitfields, bits;
-	uint32_t u32All;
-	signed int i32All;
-	float f32All;
-};
-
-static inline struct amdgpu_device *get_amdgpu_device(struct kgd_dev *kgd)
-{
-	return (struct amdgpu_device *)kgd;
-}
-
-static void lock_srbm(struct kgd_dev *kgd, uint32_t mec, uint32_t pipe,
+static void lock_srbm(struct amdgpu_device *adev, uint32_t mec, uint32_t pipe,
 			uint32_t queue, uint32_t vmid)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t value = PIPEID(pipe) | MEID(mec) | VMID(vmid) | QUEUEID(queue);
 
 	mutex_lock(&adev->srbm_mutex);
 	WREG32(mmSRBM_GFX_CNTL, value);
 }
 
-static void unlock_srbm(struct kgd_dev *kgd)
+static void unlock_srbm(struct amdgpu_device *adev)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	WREG32(mmSRBM_GFX_CNTL, 0);
 	mutex_unlock(&adev->srbm_mutex);
 }
 
-static void acquire_queue(struct kgd_dev *kgd, uint32_t pipe_id,
+static void acquire_queue(struct amdgpu_device *adev, uint32_t pipe_id,
 				uint32_t queue_id)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	uint32_t mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
 	uint32_t pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);
 
-	lock_srbm(kgd, mec, pipe, queue_id, 0);
+	lock_srbm(adev, mec, pipe, queue_id, 0);
 }
 
-static void release_queue(struct kgd_dev *kgd)
+static void release_queue(struct amdgpu_device *adev)
 {
-	unlock_srbm(kgd);
+	unlock_srbm(adev);
 }
 
-static void kgd_program_sh_mem_settings(struct kgd_dev *kgd, uint32_t vmid,
+static void kgd_program_sh_mem_settings(struct amdgpu_device *adev, uint32_t vmid,
 					uint32_t sh_mem_config,
 					uint32_t sh_mem_ape1_base,
 					uint32_t sh_mem_ape1_limit,
-					uint32_t sh_mem_bases)
+					uint32_t sh_mem_bases, uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
-	lock_srbm(kgd, 0, 0, 0, vmid);
+	lock_srbm(adev, 0, 0, 0, vmid);
 
 	WREG32(mmSH_MEM_CONFIG, sh_mem_config);
 	WREG32(mmSH_MEM_APE1_BASE, sh_mem_ape1_base);
 	WREG32(mmSH_MEM_APE1_LIMIT, sh_mem_ape1_limit);
 	WREG32(mmSH_MEM_BASES, sh_mem_bases);
 
-	unlock_srbm(kgd);
+	unlock_srbm(adev);
 }
 
-static int kgd_set_pasid_vmid_mapping(struct kgd_dev *kgd, u32 pasid,
-					unsigned int vmid)
+static int kgd_set_pasid_vmid_mapping(struct amdgpu_device *adev, u32 pasid,
+					unsigned int vmid, uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	/*
 	 * We have to assume that there is no outstanding mapping.
 	 * The ATC_VMID_PASID_MAPPING_UPDATE_STATUS bit could be 0 because
@@ -165,21 +114,21 @@ static int kgd_set_pasid_vmid_mapping(struct kgd_dev *kgd, u32 pasid,
 	return 0;
 }
 
-static int kgd_init_interrupts(struct kgd_dev *kgd, uint32_t pipe_id)
+static int kgd_init_interrupts(struct amdgpu_device *adev, uint32_t pipe_id,
+				uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t mec;
 	uint32_t pipe;
 
 	mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
 	pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);
 
-	lock_srbm(kgd, mec, pipe, 0, 0);
+	lock_srbm(adev, mec, pipe, 0, 0);
 
 	WREG32(mmCPC_INT_CNTL, CP_INT_CNTL_RING0__TIME_STAMP_INT_ENABLE_MASK |
 			CP_INT_CNTL_RING0__OPCODE_ERROR_INT_ENABLE_MASK);
 
-	unlock_srbm(kgd);
+	unlock_srbm(adev);
 
 	return 0;
 }
@@ -207,12 +156,11 @@ static inline struct cik_sdma_rlc_registers *get_sdma_mqd(void *mqd)
 	return (struct cik_sdma_rlc_registers *)mqd;
 }
 
-static int kgd_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
-			uint32_t queue_id, uint32_t __user *wptr,
-			uint32_t wptr_shift, uint32_t wptr_mask,
-			struct mm_struct *mm)
+static int kgd_hqd_load(struct amdgpu_device *adev, void *mqd,
+			uint32_t pipe_id, uint32_t queue_id,
+			uint32_t __user *wptr, uint32_t wptr_shift,
+			uint32_t wptr_mask, struct mm_struct *mm, uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct cik_mqd *m;
 	uint32_t *mqd_hqd;
 	uint32_t reg, wptr_val, data;
@@ -220,7 +168,7 @@ static int kgd_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
 
 	m = get_mqd(mqd);
 
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 
 	/* HQD registers extend from CP_MQD_BASE_ADDR to CP_MQD_CONTROL. */
 	mqd_hqd = &m->cp_mqd_base_addr_lo;
@@ -237,27 +185,26 @@ static int kgd_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
 
 	/* read_user_ptr may take the mm->mmap_lock.
 	 * release srbm_mutex to avoid circular dependency between
-	 * srbm_mutex->mm_sem->reservation_ww_class_mutex->srbm_mutex.
+	 * srbm_mutex->mmap_lock->reservation_ww_class_mutex->srbm_mutex.
 	 */
-	release_queue(kgd);
+	release_queue(adev);
 	valid_wptr = read_user_wptr(mm, wptr, wptr_val);
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 	if (valid_wptr)
 		WREG32(mmCP_HQD_PQ_WPTR, (wptr_val << wptr_shift) & wptr_mask);
 
 	data = REG_SET_FIELD(m->cp_hqd_active, CP_HQD_ACTIVE, ACTIVE, 1);
 	WREG32(mmCP_HQD_ACTIVE, data);
 
-	release_queue(kgd);
+	release_queue(adev);
 
 	return 0;
 }
 
-static int kgd_hqd_dump(struct kgd_dev *kgd,
+static int kgd_hqd_dump(struct amdgpu_device *adev,
 			uint32_t pipe_id, uint32_t queue_id,
-			uint32_t (**dump)[2], uint32_t *n_regs)
+			uint32_t (**dump)[2], uint32_t *n_regs, uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t i = 0, reg;
 #define HQD_N_REGS (35+4)
 #define DUMP_REG(addr) do {				\
@@ -267,11 +214,11 @@ static int kgd_hqd_dump(struct kgd_dev *kgd,
 		(*dump)[i++][1] = RREG32(addr);		\
 	} while (0)
 
-	*dump = kmalloc_array(HQD_N_REGS * 2, sizeof(uint32_t), GFP_KERNEL);
+	*dump = kmalloc_array(HQD_N_REGS, sizeof(**dump), GFP_KERNEL);
 	if (*dump == NULL)
 		return -ENOMEM;
 
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 
 	DUMP_REG(mmCOMPUTE_STATIC_THREAD_MGMT_SE0);
 	DUMP_REG(mmCOMPUTE_STATIC_THREAD_MGMT_SE1);
@@ -281,7 +228,7 @@ static int kgd_hqd_dump(struct kgd_dev *kgd,
 	for (reg = mmCP_MQD_BASE_ADDR; reg <= mmCP_MQD_CONTROL; reg++)
 		DUMP_REG(reg);
 
-	release_queue(kgd);
+	release_queue(adev);
 
 	WARN_ON_ONCE(i != HQD_N_REGS);
 	*n_regs = i;
@@ -289,10 +236,9 @@ static int kgd_hqd_dump(struct kgd_dev *kgd,
 	return 0;
 }
 
-static int kgd_hqd_sdma_load(struct kgd_dev *kgd, void *mqd,
+static int kgd_hqd_sdma_load(struct amdgpu_device *adev, void *mqd,
 			     uint32_t __user *wptr, struct mm_struct *mm)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct cik_sdma_rlc_registers *m;
 	unsigned long end_jiffies;
 	uint32_t sdma_rlc_reg_offset;
@@ -345,18 +291,17 @@ static int kgd_hqd_sdma_load(struct kgd_dev *kgd, void *mqd,
 	return 0;
 }
 
-static int kgd_hqd_sdma_dump(struct kgd_dev *kgd,
+static int kgd_hqd_sdma_dump(struct amdgpu_device *adev,
 			     uint32_t engine_id, uint32_t queue_id,
 			     uint32_t (**dump)[2], uint32_t *n_regs)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t sdma_offset = engine_id * SDMA1_REGISTER_OFFSET +
 		queue_id * KFD_CIK_SDMA_QUEUE_OFFSET;
 	uint32_t i = 0, reg;
 #undef HQD_N_REGS
 #define HQD_N_REGS (19+4)
 
-	*dump = kmalloc_array(HQD_N_REGS * 2, sizeof(uint32_t), GFP_KERNEL);
+	*dump = kmalloc_array(HQD_N_REGS, sizeof(**dump), GFP_KERNEL);
 	if (*dump == NULL)
 		return -ENOMEM;
 
@@ -372,15 +317,15 @@ static int kgd_hqd_sdma_dump(struct kgd_dev *kgd,
 	return 0;
 }
 
-static bool kgd_hqd_is_occupied(struct kgd_dev *kgd, uint64_t queue_address,
-				uint32_t pipe_id, uint32_t queue_id)
+static bool kgd_hqd_is_occupied(struct amdgpu_device *adev,
+				uint64_t queue_address, uint32_t pipe_id,
+				uint32_t queue_id, uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t act;
 	bool retval = false;
 	uint32_t low, high;
 
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 	act = RREG32(mmCP_HQD_ACTIVE);
 	if (act) {
 		low = lower_32_bits(queue_address >> 8);
@@ -390,13 +335,12 @@ static bool kgd_hqd_is_occupied(struct kgd_dev *kgd, uint64_t queue_address,
 				high == RREG32(mmCP_HQD_PQ_BASE_HI))
 			retval = true;
 	}
-	release_queue(kgd);
+	release_queue(adev);
 	return retval;
 }
 
-static bool kgd_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd)
+static bool kgd_hqd_sdma_is_occupied(struct amdgpu_device *adev, void *mqd)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct cik_sdma_rlc_registers *m;
 	uint32_t sdma_rlc_reg_offset;
 	uint32_t sdma_rlc_rb_cntl;
@@ -412,12 +356,11 @@ static bool kgd_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd)
 	return false;
 }
 
-static int kgd_hqd_destroy(struct kgd_dev *kgd, void *mqd,
+static int kgd_hqd_destroy(struct amdgpu_device *adev, void *mqd,
 				enum kfd_preempt_type reset_type,
 				unsigned int utimeout, uint32_t pipe_id,
-				uint32_t queue_id)
+				uint32_t queue_id, uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t temp;
 	enum hqd_dequeue_request_type type;
 	unsigned long flags, end_jiffies;
@@ -426,7 +369,7 @@ static int kgd_hqd_destroy(struct kgd_dev *kgd, void *mqd,
 	if (amdgpu_in_reset(adev))
 		return -EIO;
 
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 	WREG32(mmCP_HQD_PQ_DOORBELL_CONTROL, 0);
 
 	switch (reset_type) {
@@ -504,20 +447,19 @@ loop:
 			break;
 		if (time_after(jiffies, end_jiffies)) {
 			pr_err("cp queue preemption time out\n");
-			release_queue(kgd);
+			release_queue(adev);
 			return -ETIME;
 		}
 		usleep_range(500, 1000);
 	}
 
-	release_queue(kgd);
+	release_queue(adev);
 	return 0;
 }
 
-static int kgd_hqd_sdma_destroy(struct kgd_dev *kgd, void *mqd,
+static int kgd_hqd_sdma_destroy(struct amdgpu_device *adev, void *mqd,
 				unsigned int utimeout)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct cik_sdma_rlc_registers *m;
 	uint32_t sdma_rlc_reg_offset;
 	uint32_t temp;
@@ -551,62 +493,10 @@ static int kgd_hqd_sdma_destroy(struct kgd_dev *kgd, void *mqd,
 	return 0;
 }
 
-static int kgd_address_watch_disable(struct kgd_dev *kgd)
-{
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-	union TCP_WATCH_CNTL_BITS cntl;
-	unsigned int i;
-
-	cntl.u32All = 0;
-
-	cntl.bitfields.valid = 0;
-	cntl.bitfields.mask = ADDRESS_WATCH_REG_CNTL_DEFAULT_MASK;
-	cntl.bitfields.atc = 1;
-
-	/* Turning off this address until we set all the registers */
-	for (i = 0; i < MAX_WATCH_ADDRESSES; i++)
-		WREG32(watchRegs[i * ADDRESS_WATCH_REG_MAX +
-			ADDRESS_WATCH_REG_CNTL], cntl.u32All);
-
-	return 0;
-}
-
-static int kgd_address_watch_execute(struct kgd_dev *kgd,
-					unsigned int watch_point_id,
-					uint32_t cntl_val,
-					uint32_t addr_hi,
-					uint32_t addr_lo)
-{
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-	union TCP_WATCH_CNTL_BITS cntl;
-
-	cntl.u32All = cntl_val;
-
-	/* Turning off this watch point until we set all the registers */
-	cntl.bitfields.valid = 0;
-	WREG32(watchRegs[watch_point_id * ADDRESS_WATCH_REG_MAX +
-		ADDRESS_WATCH_REG_CNTL], cntl.u32All);
-
-	WREG32(watchRegs[watch_point_id * ADDRESS_WATCH_REG_MAX +
-		ADDRESS_WATCH_REG_ADDR_HI], addr_hi);
-
-	WREG32(watchRegs[watch_point_id * ADDRESS_WATCH_REG_MAX +
-		ADDRESS_WATCH_REG_ADDR_LO], addr_lo);
-
-	/* Enable the watch point */
-	cntl.bitfields.valid = 1;
-
-	WREG32(watchRegs[watch_point_id * ADDRESS_WATCH_REG_MAX +
-		ADDRESS_WATCH_REG_CNTL], cntl.u32All);
-
-	return 0;
-}
-
-static int kgd_wave_control_execute(struct kgd_dev *kgd,
+static int kgd_wave_control_execute(struct amdgpu_device *adev,
 					uint32_t gfx_index_val,
-					uint32_t sq_cmd)
+					uint32_t sq_cmd, uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t data;
 
 	mutex_lock(&adev->grbm_idx_mutex);
@@ -627,18 +517,10 @@ static int kgd_wave_control_execute(struct kgd_dev *kgd,
 	return 0;
 }
 
-static uint32_t kgd_address_watch_get_offset(struct kgd_dev *kgd,
-					unsigned int watch_point_id,
-					unsigned int reg_offset)
-{
-	return watchRegs[watch_point_id * ADDRESS_WATCH_REG_MAX + reg_offset];
-}
-
-static bool get_atc_vmid_pasid_mapping_info(struct kgd_dev *kgd,
+static bool get_atc_vmid_pasid_mapping_info(struct amdgpu_device *adev,
 					uint8_t vmid, uint16_t *p_pasid)
 {
 	uint32_t value;
-	struct amdgpu_device *adev = (struct amdgpu_device *) kgd;
 
 	value = RREG32(mmATC_VMID0_PASID_MAPPING + vmid);
 	*p_pasid = value & ATC_VMID0_PASID_MAPPING__PASID_MASK;
@@ -646,21 +528,17 @@ static bool get_atc_vmid_pasid_mapping_info(struct kgd_dev *kgd,
 	return !!(value & ATC_VMID0_PASID_MAPPING__VALID_MASK);
 }
 
-static void set_scratch_backing_va(struct kgd_dev *kgd,
+static void set_scratch_backing_va(struct amdgpu_device *adev,
 					uint64_t va, uint32_t vmid)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *) kgd;
-
-	lock_srbm(kgd, 0, 0, 0, vmid);
+	lock_srbm(adev, 0, 0, 0, vmid);
 	WREG32(mmSH_HIDDEN_PRIVATE_BASE_VMID, va);
-	unlock_srbm(kgd);
+	unlock_srbm(adev);
 }
 
-static void set_vm_context_page_table_base(struct kgd_dev *kgd, uint32_t vmid,
-			uint64_t page_table_base)
+static void set_vm_context_page_table_base(struct amdgpu_device *adev,
+			uint32_t vmid, uint64_t page_table_base)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	if (!amdgpu_amdkfd_is_kfd_vmid(adev, vmid)) {
 		pr_err("trying to set page table base for wrong VMID\n");
 		return;
@@ -676,15 +554,20 @@ static void set_vm_context_page_table_base(struct kgd_dev *kgd, uint32_t vmid,
   * @vmid: vmid pointer
   * read vmid from register (CIK).
   */
-static uint32_t read_vmid_from_vmfault_reg(struct kgd_dev *kgd)
+static uint32_t read_vmid_from_vmfault_reg(struct amdgpu_device *adev)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	uint32_t status = RREG32(mmVM_CONTEXT1_PROTECTION_FAULT_STATUS);
 
 	return REG_GET_FIELD(status, VM_CONTEXT1_PROTECTION_FAULT_STATUS, VMID);
 }
 
+static uint32_t kgd_hqd_sdma_get_doorbell(struct amdgpu_device *adev,
+					  int engine, int queue)
+
+{
+	return 0;
+}
+
 const struct kfd2kgd_calls gfx_v7_kfd2kgd = {
 	.program_sh_mem_settings = kgd_program_sh_mem_settings,
 	.set_pasid_vmid_mapping = kgd_set_pasid_vmid_mapping,
@@ -697,12 +580,10 @@ const struct kfd2kgd_calls gfx_v7_kfd2kgd = {
 	.hqd_sdma_is_occupied = kgd_hqd_sdma_is_occupied,
 	.hqd_destroy = kgd_hqd_destroy,
 	.hqd_sdma_destroy = kgd_hqd_sdma_destroy,
-	.address_watch_disable = kgd_address_watch_disable,
-	.address_watch_execute = kgd_address_watch_execute,
 	.wave_control_execute = kgd_wave_control_execute,
-	.address_watch_get_offset = kgd_address_watch_get_offset,
 	.get_atc_vmid_pasid_mapping_info = get_atc_vmid_pasid_mapping_info,
 	.set_scratch_backing_va = set_scratch_backing_va,
 	.set_vm_context_page_table_base = set_vm_context_page_table_base,
 	.read_vmid_from_vmfault_reg = read_vmid_from_vmfault_reg,
+	.hqd_sdma_get_doorbell = kgd_hqd_sdma_get_doorbell,
 };
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v8.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v8.c
index 5ce0ce704a21..e68c0fa8d751 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v8.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v8.c
@@ -39,68 +39,54 @@ enum hqd_dequeue_request_type {
 	RESET_WAVES
 };
 
-static inline struct amdgpu_device *get_amdgpu_device(struct kgd_dev *kgd)
-{
-	return (struct amdgpu_device *)kgd;
-}
-
-static void lock_srbm(struct kgd_dev *kgd, uint32_t mec, uint32_t pipe,
+static void lock_srbm(struct amdgpu_device *adev, uint32_t mec, uint32_t pipe,
 			uint32_t queue, uint32_t vmid)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t value = PIPEID(pipe) | MEID(mec) | VMID(vmid) | QUEUEID(queue);
 
 	mutex_lock(&adev->srbm_mutex);
 	WREG32(mmSRBM_GFX_CNTL, value);
 }
 
-static void unlock_srbm(struct kgd_dev *kgd)
+static void unlock_srbm(struct amdgpu_device *adev)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	WREG32(mmSRBM_GFX_CNTL, 0);
 	mutex_unlock(&adev->srbm_mutex);
 }
 
-static void acquire_queue(struct kgd_dev *kgd, uint32_t pipe_id,
+static void acquire_queue(struct amdgpu_device *adev, uint32_t pipe_id,
 				uint32_t queue_id)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	uint32_t mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
 	uint32_t pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);
 
-	lock_srbm(kgd, mec, pipe, queue_id, 0);
+	lock_srbm(adev, mec, pipe, queue_id, 0);
 }
 
-static void release_queue(struct kgd_dev *kgd)
+static void release_queue(struct amdgpu_device *adev)
 {
-	unlock_srbm(kgd);
+	unlock_srbm(adev);
 }
 
-static void kgd_program_sh_mem_settings(struct kgd_dev *kgd, uint32_t vmid,
+static void kgd_program_sh_mem_settings(struct amdgpu_device *adev, uint32_t vmid,
 					uint32_t sh_mem_config,
 					uint32_t sh_mem_ape1_base,
 					uint32_t sh_mem_ape1_limit,
-					uint32_t sh_mem_bases)
+					uint32_t sh_mem_bases, uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
-	lock_srbm(kgd, 0, 0, 0, vmid);
+	lock_srbm(adev, 0, 0, 0, vmid);
 
 	WREG32(mmSH_MEM_CONFIG, sh_mem_config);
 	WREG32(mmSH_MEM_APE1_BASE, sh_mem_ape1_base);
 	WREG32(mmSH_MEM_APE1_LIMIT, sh_mem_ape1_limit);
 	WREG32(mmSH_MEM_BASES, sh_mem_bases);
 
-	unlock_srbm(kgd);
+	unlock_srbm(adev);
 }
 
-static int kgd_set_pasid_vmid_mapping(struct kgd_dev *kgd, u32 pasid,
-					unsigned int vmid)
+static int kgd_set_pasid_vmid_mapping(struct amdgpu_device *adev, u32 pasid,
+					unsigned int vmid, uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	/*
 	 * We have to assume that there is no outstanding mapping.
 	 * The ATC_VMID_PASID_MAPPING_UPDATE_STATUS bit could be 0 because
@@ -123,21 +109,21 @@ static int kgd_set_pasid_vmid_mapping(struct kgd_dev *kgd, u32 pasid,
 	return 0;
 }
 
-static int kgd_init_interrupts(struct kgd_dev *kgd, uint32_t pipe_id)
+static int kgd_init_interrupts(struct amdgpu_device *adev, uint32_t pipe_id,
+				uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t mec;
 	uint32_t pipe;
 
 	mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
 	pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);
 
-	lock_srbm(kgd, mec, pipe, 0, 0);
+	lock_srbm(adev, mec, pipe, 0, 0);
 
 	WREG32(mmCPC_INT_CNTL, CP_INT_CNTL_RING0__TIME_STAMP_INT_ENABLE_MASK |
 			CP_INT_CNTL_RING0__OPCODE_ERROR_INT_ENABLE_MASK);
 
-	unlock_srbm(kgd);
+	unlock_srbm(adev);
 
 	return 0;
 }
@@ -165,12 +151,11 @@ static inline struct vi_sdma_mqd *get_sdma_mqd(void *mqd)
 	return (struct vi_sdma_mqd *)mqd;
 }
 
-static int kgd_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
-			uint32_t queue_id, uint32_t __user *wptr,
-			uint32_t wptr_shift, uint32_t wptr_mask,
-			struct mm_struct *mm)
+static int kgd_hqd_load(struct amdgpu_device *adev, void *mqd,
+			uint32_t pipe_id, uint32_t queue_id,
+			uint32_t __user *wptr, uint32_t wptr_shift,
+			uint32_t wptr_mask, struct mm_struct *mm, uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct vi_mqd *m;
 	uint32_t *mqd_hqd;
 	uint32_t reg, wptr_val, data;
@@ -178,7 +163,7 @@ static int kgd_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
 
 	m = get_mqd(mqd);
 
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 
 	/* HIQ is set during driver init period with vmid set to 0*/
 	if (m->cp_hqd_vmid == 0) {
@@ -206,7 +191,7 @@ static int kgd_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
 	 * on ASICs that do not support context-save.
 	 * EOP writes/reads can start anywhere in the ring.
 	 */
-	if (get_amdgpu_device(kgd)->asic_type != CHIP_TONGA) {
+	if (adev->asic_type != CHIP_TONGA) {
 		WREG32(mmCP_HQD_EOP_RPTR, m->cp_hqd_eop_rptr);
 		WREG32(mmCP_HQD_EOP_WPTR, m->cp_hqd_eop_wptr);
 		WREG32(mmCP_HQD_EOP_WPTR_MEM, m->cp_hqd_eop_wptr_mem);
@@ -224,27 +209,26 @@ static int kgd_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
 
 	/* read_user_ptr may take the mm->mmap_lock.
 	 * release srbm_mutex to avoid circular dependency between
-	 * srbm_mutex->mm_sem->reservation_ww_class_mutex->srbm_mutex.
+	 * srbm_mutex->mmap_lock->reservation_ww_class_mutex->srbm_mutex.
 	 */
-	release_queue(kgd);
+	release_queue(adev);
 	valid_wptr = read_user_wptr(mm, wptr, wptr_val);
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 	if (valid_wptr)
 		WREG32(mmCP_HQD_PQ_WPTR, (wptr_val << wptr_shift) & wptr_mask);
 
 	data = REG_SET_FIELD(m->cp_hqd_active, CP_HQD_ACTIVE, ACTIVE, 1);
 	WREG32(mmCP_HQD_ACTIVE, data);
 
-	release_queue(kgd);
+	release_queue(adev);
 
 	return 0;
 }
 
-static int kgd_hqd_dump(struct kgd_dev *kgd,
+static int kgd_hqd_dump(struct amdgpu_device *adev,
 			uint32_t pipe_id, uint32_t queue_id,
-			uint32_t (**dump)[2], uint32_t *n_regs)
+			uint32_t (**dump)[2], uint32_t *n_regs, uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t i = 0, reg;
 #define HQD_N_REGS (54+4)
 #define DUMP_REG(addr) do {				\
@@ -254,11 +238,11 @@ static int kgd_hqd_dump(struct kgd_dev *kgd,
 		(*dump)[i++][1] = RREG32(addr);		\
 	} while (0)
 
-	*dump = kmalloc_array(HQD_N_REGS * 2, sizeof(uint32_t), GFP_KERNEL);
+	*dump = kmalloc_array(HQD_N_REGS, sizeof(**dump), GFP_KERNEL);
 	if (*dump == NULL)
 		return -ENOMEM;
 
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 
 	DUMP_REG(mmCOMPUTE_STATIC_THREAD_MGMT_SE0);
 	DUMP_REG(mmCOMPUTE_STATIC_THREAD_MGMT_SE1);
@@ -268,7 +252,7 @@ static int kgd_hqd_dump(struct kgd_dev *kgd,
 	for (reg = mmCP_MQD_BASE_ADDR; reg <= mmCP_HQD_EOP_DONES; reg++)
 		DUMP_REG(reg);
 
-	release_queue(kgd);
+	release_queue(adev);
 
 	WARN_ON_ONCE(i != HQD_N_REGS);
 	*n_regs = i;
@@ -276,10 +260,9 @@ static int kgd_hqd_dump(struct kgd_dev *kgd,
 	return 0;
 }
 
-static int kgd_hqd_sdma_load(struct kgd_dev *kgd, void *mqd,
+static int kgd_hqd_sdma_load(struct amdgpu_device *adev, void *mqd,
 			     uint32_t __user *wptr, struct mm_struct *mm)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct vi_sdma_mqd *m;
 	unsigned long end_jiffies;
 	uint32_t sdma_rlc_reg_offset;
@@ -331,18 +314,17 @@ static int kgd_hqd_sdma_load(struct kgd_dev *kgd, void *mqd,
 	return 0;
 }
 
-static int kgd_hqd_sdma_dump(struct kgd_dev *kgd,
+static int kgd_hqd_sdma_dump(struct amdgpu_device *adev,
 			     uint32_t engine_id, uint32_t queue_id,
 			     uint32_t (**dump)[2], uint32_t *n_regs)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t sdma_offset = engine_id * SDMA1_REGISTER_OFFSET +
 		queue_id * KFD_VI_SDMA_QUEUE_OFFSET;
 	uint32_t i = 0, reg;
 #undef HQD_N_REGS
 #define HQD_N_REGS (19+4+2+3+7)
 
-	*dump = kmalloc_array(HQD_N_REGS * 2, sizeof(uint32_t), GFP_KERNEL);
+	*dump = kmalloc_array(HQD_N_REGS, sizeof(**dump), GFP_KERNEL);
 	if (*dump == NULL)
 		return -ENOMEM;
 
@@ -367,15 +349,15 @@ static int kgd_hqd_sdma_dump(struct kgd_dev *kgd,
 	return 0;
 }
 
-static bool kgd_hqd_is_occupied(struct kgd_dev *kgd, uint64_t queue_address,
-				uint32_t pipe_id, uint32_t queue_id)
+static bool kgd_hqd_is_occupied(struct amdgpu_device *adev,
+				uint64_t queue_address, uint32_t pipe_id,
+				uint32_t queue_id, uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t act;
 	bool retval = false;
 	uint32_t low, high;
 
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 	act = RREG32(mmCP_HQD_ACTIVE);
 	if (act) {
 		low = lower_32_bits(queue_address >> 8);
@@ -385,13 +367,12 @@ static bool kgd_hqd_is_occupied(struct kgd_dev *kgd, uint64_t queue_address,
 				high == RREG32(mmCP_HQD_PQ_BASE_HI))
 			retval = true;
 	}
-	release_queue(kgd);
+	release_queue(adev);
 	return retval;
 }
 
-static bool kgd_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd)
+static bool kgd_hqd_sdma_is_occupied(struct amdgpu_device *adev, void *mqd)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct vi_sdma_mqd *m;
 	uint32_t sdma_rlc_reg_offset;
 	uint32_t sdma_rlc_rb_cntl;
@@ -407,12 +388,11 @@ static bool kgd_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd)
 	return false;
 }
 
-static int kgd_hqd_destroy(struct kgd_dev *kgd, void *mqd,
+static int kgd_hqd_destroy(struct amdgpu_device *adev, void *mqd,
 				enum kfd_preempt_type reset_type,
 				unsigned int utimeout, uint32_t pipe_id,
-				uint32_t queue_id)
+				uint32_t queue_id, uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t temp;
 	enum hqd_dequeue_request_type type;
 	unsigned long flags, end_jiffies;
@@ -422,7 +402,7 @@ static int kgd_hqd_destroy(struct kgd_dev *kgd, void *mqd,
 	if (amdgpu_in_reset(adev))
 		return -EIO;
 
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 
 	if (m->cp_hqd_vmid == 0)
 		WREG32_FIELD(RLC_CP_SCHEDULERS, scheduler1, 0);
@@ -502,20 +482,19 @@ loop:
 			break;
 		if (time_after(jiffies, end_jiffies)) {
 			pr_err("cp queue preemption time out.\n");
-			release_queue(kgd);
+			release_queue(adev);
 			return -ETIME;
 		}
 		usleep_range(500, 1000);
 	}
 
-	release_queue(kgd);
+	release_queue(adev);
 	return 0;
 }
 
-static int kgd_hqd_sdma_destroy(struct kgd_dev *kgd, void *mqd,
+static int kgd_hqd_sdma_destroy(struct amdgpu_device *adev, void *mqd,
 				unsigned int utimeout)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct vi_sdma_mqd *m;
 	uint32_t sdma_rlc_reg_offset;
 	uint32_t temp;
@@ -549,11 +528,10 @@ static int kgd_hqd_sdma_destroy(struct kgd_dev *kgd, void *mqd,
 	return 0;
 }
 
-static bool get_atc_vmid_pasid_mapping_info(struct kgd_dev *kgd,
+static bool get_atc_vmid_pasid_mapping_info(struct amdgpu_device *adev,
 					uint8_t vmid, uint16_t *p_pasid)
 {
 	uint32_t value;
-	struct amdgpu_device *adev = (struct amdgpu_device *) kgd;
 
 	value = RREG32(mmATC_VMID0_PASID_MAPPING + vmid);
 	*p_pasid = value & ATC_VMID0_PASID_MAPPING__PASID_MASK;
@@ -561,25 +539,10 @@ static bool get_atc_vmid_pasid_mapping_info(struct kgd_dev *kgd,
 	return !!(value & ATC_VMID0_PASID_MAPPING__VALID_MASK);
 }
 
-static int kgd_address_watch_disable(struct kgd_dev *kgd)
-{
-	return 0;
-}
-
-static int kgd_address_watch_execute(struct kgd_dev *kgd,
-					unsigned int watch_point_id,
-					uint32_t cntl_val,
-					uint32_t addr_hi,
-					uint32_t addr_lo)
-{
-	return 0;
-}
-
-static int kgd_wave_control_execute(struct kgd_dev *kgd,
+static int kgd_wave_control_execute(struct amdgpu_device *adev,
 					uint32_t gfx_index_val,
-					uint32_t sq_cmd)
+					uint32_t sq_cmd, uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t data = 0;
 
 	mutex_lock(&adev->grbm_idx_mutex);
@@ -600,28 +563,17 @@ static int kgd_wave_control_execute(struct kgd_dev *kgd,
 	return 0;
 }
 
-static uint32_t kgd_address_watch_get_offset(struct kgd_dev *kgd,
-					unsigned int watch_point_id,
-					unsigned int reg_offset)
-{
-	return 0;
-}
-
-static void set_scratch_backing_va(struct kgd_dev *kgd,
+static void set_scratch_backing_va(struct amdgpu_device *adev,
 					uint64_t va, uint32_t vmid)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *) kgd;
-
-	lock_srbm(kgd, 0, 0, 0, vmid);
+	lock_srbm(adev, 0, 0, 0, vmid);
 	WREG32(mmSH_HIDDEN_PRIVATE_BASE_VMID, va);
-	unlock_srbm(kgd);
+	unlock_srbm(adev);
 }
 
-static void set_vm_context_page_table_base(struct kgd_dev *kgd, uint32_t vmid,
-		uint64_t page_table_base)
+static void set_vm_context_page_table_base(struct amdgpu_device *adev,
+		uint32_t vmid, uint64_t page_table_base)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	if (!amdgpu_amdkfd_is_kfd_vmid(adev, vmid)) {
 		pr_err("trying to set page table base for wrong VMID\n");
 		return;
@@ -630,6 +582,13 @@ static void set_vm_context_page_table_base(struct kgd_dev *kgd, uint32_t vmid,
 			lower_32_bits(page_table_base));
 }
 
+static uint32_t kgd_hqd_sdma_get_doorbell(struct amdgpu_device *adev,
+					  int engine, int queue)
+
+{
+	return 0;
+}
+
 const struct kfd2kgd_calls gfx_v8_kfd2kgd = {
 	.program_sh_mem_settings = kgd_program_sh_mem_settings,
 	.set_pasid_vmid_mapping = kgd_set_pasid_vmid_mapping,
@@ -642,12 +601,10 @@ const struct kfd2kgd_calls gfx_v8_kfd2kgd = {
 	.hqd_sdma_is_occupied = kgd_hqd_sdma_is_occupied,
 	.hqd_destroy = kgd_hqd_destroy,
 	.hqd_sdma_destroy = kgd_hqd_sdma_destroy,
-	.address_watch_disable = kgd_address_watch_disable,
-	.address_watch_execute = kgd_address_watch_execute,
 	.wave_control_execute = kgd_wave_control_execute,
-	.address_watch_get_offset = kgd_address_watch_get_offset,
 	.get_atc_vmid_pasid_mapping_info =
 			get_atc_vmid_pasid_mapping_info,
 	.set_scratch_backing_va = set_scratch_backing_va,
 	.set_vm_context_page_table_base = set_vm_context_page_table_base,
+	.hqd_sdma_get_doorbell = kgd_hqd_sdma_get_doorbell,
 };
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c
index bcc1cbeb8799..088d09cc7a72 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c
@@ -38,6 +38,7 @@
 #include "soc15d.h"
 #include "gfx_v9_0.h"
 #include "amdgpu_amdkfd_gfx_v9.h"
+#include <uapi/linux/kfd_ioctl.h>
 
 enum hqd_dequeue_request_type {
 	NO_ACTION = 0,
@@ -46,40 +47,29 @@ enum hqd_dequeue_request_type {
 	SAVE_WAVES
 };
 
-static inline struct amdgpu_device *get_amdgpu_device(struct kgd_dev *kgd)
+static void kgd_gfx_v9_lock_srbm(struct amdgpu_device *adev, uint32_t mec, uint32_t pipe,
+			uint32_t queue, uint32_t vmid, uint32_t inst)
 {
-	return (struct amdgpu_device *)kgd;
-}
-
-static void lock_srbm(struct kgd_dev *kgd, uint32_t mec, uint32_t pipe,
-			uint32_t queue, uint32_t vmid)
-{
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	mutex_lock(&adev->srbm_mutex);
-	soc15_grbm_select(adev, mec, pipe, queue, vmid);
+	soc15_grbm_select(adev, mec, pipe, queue, vmid, GET_INST(GC, inst));
 }
 
-static void unlock_srbm(struct kgd_dev *kgd)
+static void kgd_gfx_v9_unlock_srbm(struct amdgpu_device *adev, uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
-	soc15_grbm_select(adev, 0, 0, 0, 0);
+	soc15_grbm_select(adev, 0, 0, 0, 0, GET_INST(GC, inst));
 	mutex_unlock(&adev->srbm_mutex);
 }
 
-static void acquire_queue(struct kgd_dev *kgd, uint32_t pipe_id,
-				uint32_t queue_id)
+void kgd_gfx_v9_acquire_queue(struct amdgpu_device *adev, uint32_t pipe_id,
+				uint32_t queue_id, uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	uint32_t mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
 	uint32_t pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);
 
-	lock_srbm(kgd, mec, pipe, queue_id, 0);
+	kgd_gfx_v9_lock_srbm(adev, mec, pipe, queue_id, 0, inst);
 }
 
-static uint64_t get_queue_mask(struct amdgpu_device *adev,
+uint64_t kgd_gfx_v9_get_queue_mask(struct amdgpu_device *adev,
 			       uint32_t pipe_id, uint32_t queue_id)
 {
 	unsigned int bit = pipe_id * adev->gfx.mec.num_queue_per_pipe +
@@ -88,33 +78,29 @@ static uint64_t get_queue_mask(struct amdgpu_device *adev,
 	return 1ull << bit;
 }
 
-static void release_queue(struct kgd_dev *kgd)
+void kgd_gfx_v9_release_queue(struct amdgpu_device *adev, uint32_t inst)
 {
-	unlock_srbm(kgd);
+	kgd_gfx_v9_unlock_srbm(adev, inst);
 }
 
-void kgd_gfx_v9_program_sh_mem_settings(struct kgd_dev *kgd, uint32_t vmid,
+void kgd_gfx_v9_program_sh_mem_settings(struct amdgpu_device *adev, uint32_t vmid,
 					uint32_t sh_mem_config,
 					uint32_t sh_mem_ape1_base,
 					uint32_t sh_mem_ape1_limit,
-					uint32_t sh_mem_bases)
+					uint32_t sh_mem_bases, uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
-	lock_srbm(kgd, 0, 0, 0, vmid);
+	kgd_gfx_v9_lock_srbm(adev, 0, 0, 0, vmid, inst);
 
-	WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmSH_MEM_CONFIG), sh_mem_config);
-	WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmSH_MEM_BASES), sh_mem_bases);
+	WREG32_SOC15_RLC(GC, GET_INST(GC, inst), mmSH_MEM_CONFIG, sh_mem_config);
+	WREG32_SOC15_RLC(GC, GET_INST(GC, inst), mmSH_MEM_BASES, sh_mem_bases);
 	/* APE1 no longer exists on GFX9 */
 
-	unlock_srbm(kgd);
+	kgd_gfx_v9_unlock_srbm(adev, inst);
 }
 
-int kgd_gfx_v9_set_pasid_vmid_mapping(struct kgd_dev *kgd, u32 pasid,
-					unsigned int vmid)
+int kgd_gfx_v9_set_pasid_vmid_mapping(struct amdgpu_device *adev, u32 pasid,
+					unsigned int vmid, uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	/*
 	 * We have to assume that there is no outstanding mapping.
 	 * The ATC_VMID_PASID_MAPPING_UPDATE_STATUS bit could be 0 because
@@ -171,22 +157,22 @@ int kgd_gfx_v9_set_pasid_vmid_mapping(struct kgd_dev *kgd, u32 pasid,
  * but still works
  */
 
-int kgd_gfx_v9_init_interrupts(struct kgd_dev *kgd, uint32_t pipe_id)
+int kgd_gfx_v9_init_interrupts(struct amdgpu_device *adev, uint32_t pipe_id,
+				uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t mec;
 	uint32_t pipe;
 
 	mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
 	pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);
 
-	lock_srbm(kgd, mec, pipe, 0, 0);
+	kgd_gfx_v9_lock_srbm(adev, mec, pipe, 0, 0, inst);
 
-	WREG32(SOC15_REG_OFFSET(GC, 0, mmCPC_INT_CNTL),
+	WREG32_SOC15(GC, GET_INST(GC, inst), mmCPC_INT_CNTL,
 		CP_INT_CNTL_RING0__TIME_STAMP_INT_ENABLE_MASK |
 		CP_INT_CNTL_RING0__OPCODE_ERROR_INT_ENABLE_MASK);
 
-	unlock_srbm(kgd);
+	kgd_gfx_v9_unlock_srbm(adev, inst);
 
 	return 0;
 }
@@ -233,33 +219,33 @@ static inline struct v9_sdma_mqd *get_sdma_mqd(void *mqd)
 	return (struct v9_sdma_mqd *)mqd;
 }
 
-int kgd_gfx_v9_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
-			uint32_t queue_id, uint32_t __user *wptr,
-			uint32_t wptr_shift, uint32_t wptr_mask,
-			struct mm_struct *mm)
+int kgd_gfx_v9_hqd_load(struct amdgpu_device *adev, void *mqd,
+			uint32_t pipe_id, uint32_t queue_id,
+			uint32_t __user *wptr, uint32_t wptr_shift,
+			uint32_t wptr_mask, struct mm_struct *mm,
+			uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct v9_mqd *m;
 	uint32_t *mqd_hqd;
 	uint32_t reg, hqd_base, data;
 
 	m = get_mqd(mqd);
 
-	acquire_queue(kgd, pipe_id, queue_id);
+	kgd_gfx_v9_acquire_queue(adev, pipe_id, queue_id, inst);
 
 	/* HQD registers extend from CP_MQD_BASE_ADDR to CP_HQD_EOP_WPTR_MEM. */
 	mqd_hqd = &m->cp_mqd_base_addr_lo;
-	hqd_base = SOC15_REG_OFFSET(GC, 0, mmCP_MQD_BASE_ADDR);
+	hqd_base = SOC15_REG_OFFSET(GC, GET_INST(GC, inst), mmCP_MQD_BASE_ADDR);
 
 	for (reg = hqd_base;
-	     reg <= SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_HI); reg++)
-		WREG32_RLC(reg, mqd_hqd[reg - hqd_base]);
+	     reg <= SOC15_REG_OFFSET(GC, GET_INST(GC, inst), mmCP_HQD_PQ_WPTR_HI); reg++)
+		WREG32_XCC(reg, mqd_hqd[reg - hqd_base], inst);
 
 
 	/* Activate doorbell logic before triggering WPTR poll. */
 	data = REG_SET_FIELD(m->cp_hqd_pq_doorbell_control,
 			     CP_HQD_PQ_DOORBELL_CONTROL, DOORBELL_EN, 1);
-	WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_DOORBELL_CONTROL), data);
+	WREG32_SOC15_RLC(GC, GET_INST(GC, inst), mmCP_HQD_PQ_DOORBELL_CONTROL, data);
 
 	if (wptr) {
 		/* Don't read wptr with get_user because the user
@@ -288,44 +274,42 @@ int kgd_gfx_v9_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
 		guessed_wptr += m->cp_hqd_pq_wptr_lo & ~(queue_size - 1);
 		guessed_wptr += (uint64_t)m->cp_hqd_pq_wptr_hi << 32;
 
-		WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_LO),
-		       lower_32_bits(guessed_wptr));
-		WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_HI),
-		       upper_32_bits(guessed_wptr));
-		WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_POLL_ADDR),
-		       lower_32_bits((uintptr_t)wptr));
-		WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_POLL_ADDR_HI),
-		       upper_32_bits((uintptr_t)wptr));
-		WREG32(SOC15_REG_OFFSET(GC, 0, mmCP_PQ_WPTR_POLL_CNTL1),
-		       (uint32_t)get_queue_mask(adev, pipe_id, queue_id));
+		WREG32_SOC15_RLC(GC, GET_INST(GC, inst), mmCP_HQD_PQ_WPTR_LO,
+			lower_32_bits(guessed_wptr));
+		WREG32_SOC15_RLC(GC, GET_INST(GC, inst), mmCP_HQD_PQ_WPTR_HI,
+			upper_32_bits(guessed_wptr));
+		WREG32_SOC15_RLC(GC, GET_INST(GC, inst), mmCP_HQD_PQ_WPTR_POLL_ADDR,
+			lower_32_bits((uintptr_t)wptr));
+		WREG32_SOC15_RLC(GC, GET_INST(GC, inst), mmCP_HQD_PQ_WPTR_POLL_ADDR_HI,
+			upper_32_bits((uintptr_t)wptr));
+		WREG32_SOC15_RLC(GC, GET_INST(GC, inst), mmCP_PQ_WPTR_POLL_CNTL1,
+			(uint32_t)kgd_gfx_v9_get_queue_mask(adev, pipe_id, queue_id));
 	}
 
 	/* Start the EOP fetcher */
-	WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_EOP_RPTR),
-	       REG_SET_FIELD(m->cp_hqd_eop_rptr,
-			     CP_HQD_EOP_RPTR, INIT_FETCHER, 1));
+	WREG32_SOC15_RLC(GC, GET_INST(GC, inst), mmCP_HQD_EOP_RPTR,
+	       REG_SET_FIELD(m->cp_hqd_eop_rptr, CP_HQD_EOP_RPTR, INIT_FETCHER, 1));
 
 	data = REG_SET_FIELD(m->cp_hqd_active, CP_HQD_ACTIVE, ACTIVE, 1);
-	WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_ACTIVE), data);
+	WREG32_SOC15_RLC(GC, GET_INST(GC, inst), mmCP_HQD_ACTIVE, data);
 
-	release_queue(kgd);
+	kgd_gfx_v9_release_queue(adev, inst);
 
 	return 0;
 }
 
-int kgd_gfx_v9_hiq_mqd_load(struct kgd_dev *kgd, void *mqd,
+int kgd_gfx_v9_hiq_mqd_load(struct amdgpu_device *adev, void *mqd,
 			    uint32_t pipe_id, uint32_t queue_id,
-			    uint32_t doorbell_off)
+			    uint32_t doorbell_off, uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-	struct amdgpu_ring *kiq_ring = &adev->gfx.kiq.ring;
+	struct amdgpu_ring *kiq_ring = &adev->gfx.kiq[inst].ring;
 	struct v9_mqd *m;
 	uint32_t mec, pipe;
 	int r;
 
 	m = get_mqd(mqd);
 
-	acquire_queue(kgd, pipe_id, queue_id);
+	kgd_gfx_v9_acquire_queue(adev, pipe_id, queue_id, inst);
 
 	mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
 	pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);
@@ -333,7 +317,7 @@ int kgd_gfx_v9_hiq_mqd_load(struct kgd_dev *kgd, void *mqd,
 	pr_debug("kfd: set HIQ, mec:%d, pipe:%d, queue:%d.\n",
 		 mec, pipe, queue_id);
 
-	spin_lock(&adev->gfx.kiq.ring_lock);
+	spin_lock(&adev->gfx.kiq[inst].ring_lock);
 	r = amdgpu_ring_alloc(kiq_ring, 7);
 	if (r) {
 		pr_err("Failed to alloc KIQ (%d).\n", r);
@@ -360,17 +344,16 @@ int kgd_gfx_v9_hiq_mqd_load(struct kgd_dev *kgd, void *mqd,
 	amdgpu_ring_commit(kiq_ring);
 
 out_unlock:
-	spin_unlock(&adev->gfx.kiq.ring_lock);
-	release_queue(kgd);
+	spin_unlock(&adev->gfx.kiq[inst].ring_lock);
+	kgd_gfx_v9_release_queue(adev, inst);
 
 	return r;
 }
 
-int kgd_gfx_v9_hqd_dump(struct kgd_dev *kgd,
+int kgd_gfx_v9_hqd_dump(struct amdgpu_device *adev,
 			uint32_t pipe_id, uint32_t queue_id,
-			uint32_t (**dump)[2], uint32_t *n_regs)
+			uint32_t (**dump)[2], uint32_t *n_regs, uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t i = 0, reg;
 #define HQD_N_REGS 56
 #define DUMP_REG(addr) do {				\
@@ -380,17 +363,17 @@ int kgd_gfx_v9_hqd_dump(struct kgd_dev *kgd,
 		(*dump)[i++][1] = RREG32(addr);		\
 	} while (0)
 
-	*dump = kmalloc_array(HQD_N_REGS * 2, sizeof(uint32_t), GFP_KERNEL);
+	*dump = kmalloc_array(HQD_N_REGS, sizeof(**dump), GFP_KERNEL);
 	if (*dump == NULL)
 		return -ENOMEM;
 
-	acquire_queue(kgd, pipe_id, queue_id);
+	kgd_gfx_v9_acquire_queue(adev, pipe_id, queue_id, inst);
 
-	for (reg = SOC15_REG_OFFSET(GC, 0, mmCP_MQD_BASE_ADDR);
-	     reg <= SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_HI); reg++)
+	for (reg = SOC15_REG_OFFSET(GC, GET_INST(GC, inst), mmCP_MQD_BASE_ADDR);
+	     reg <= SOC15_REG_OFFSET(GC, GET_INST(GC, inst), mmCP_HQD_PQ_WPTR_HI); reg++)
 		DUMP_REG(reg);
 
-	release_queue(kgd);
+	kgd_gfx_v9_release_queue(adev, inst);
 
 	WARN_ON_ONCE(i != HQD_N_REGS);
 	*n_regs = i;
@@ -398,10 +381,9 @@ int kgd_gfx_v9_hqd_dump(struct kgd_dev *kgd,
 	return 0;
 }
 
-static int kgd_hqd_sdma_load(struct kgd_dev *kgd, void *mqd,
+static int kgd_hqd_sdma_load(struct amdgpu_device *adev, void *mqd,
 			     uint32_t __user *wptr, struct mm_struct *mm)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct v9_sdma_mqd *m;
 	uint32_t sdma_rlc_reg_offset;
 	unsigned long end_jiffies;
@@ -468,18 +450,17 @@ static int kgd_hqd_sdma_load(struct kgd_dev *kgd, void *mqd,
 	return 0;
 }
 
-static int kgd_hqd_sdma_dump(struct kgd_dev *kgd,
+static int kgd_hqd_sdma_dump(struct amdgpu_device *adev,
 			     uint32_t engine_id, uint32_t queue_id,
 			     uint32_t (**dump)[2], uint32_t *n_regs)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t sdma_rlc_reg_offset = get_sdma_rlc_reg_offset(adev,
 			engine_id, queue_id);
 	uint32_t i = 0, reg;
 #undef HQD_N_REGS
 #define HQD_N_REGS (19+6+7+10)
 
-	*dump = kmalloc_array(HQD_N_REGS * 2, sizeof(uint32_t), GFP_KERNEL);
+	*dump = kmalloc_array(HQD_N_REGS, sizeof(**dump), GFP_KERNEL);
 	if (*dump == NULL)
 		return -ENOMEM;
 
@@ -500,31 +481,30 @@ static int kgd_hqd_sdma_dump(struct kgd_dev *kgd,
 	return 0;
 }
 
-bool kgd_gfx_v9_hqd_is_occupied(struct kgd_dev *kgd, uint64_t queue_address,
-				uint32_t pipe_id, uint32_t queue_id)
+bool kgd_gfx_v9_hqd_is_occupied(struct amdgpu_device *adev,
+				uint64_t queue_address, uint32_t pipe_id,
+				uint32_t queue_id, uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t act;
 	bool retval = false;
 	uint32_t low, high;
 
-	acquire_queue(kgd, pipe_id, queue_id);
-	act = RREG32(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_ACTIVE));
+	kgd_gfx_v9_acquire_queue(adev, pipe_id, queue_id, inst);
+	act = RREG32_SOC15(GC, GET_INST(GC, inst), mmCP_HQD_ACTIVE);
 	if (act) {
 		low = lower_32_bits(queue_address >> 8);
 		high = upper_32_bits(queue_address >> 8);
 
-		if (low == RREG32(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_BASE)) &&
-		   high == RREG32(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_BASE_HI)))
+		if (low == RREG32_SOC15(GC, GET_INST(GC, inst), mmCP_HQD_PQ_BASE) &&
+		   high == RREG32_SOC15(GC, GET_INST(GC, inst), mmCP_HQD_PQ_BASE_HI))
 			retval = true;
 	}
-	release_queue(kgd);
+	kgd_gfx_v9_release_queue(adev, inst);
 	return retval;
 }
 
-static bool kgd_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd)
+static bool kgd_hqd_sdma_is_occupied(struct amdgpu_device *adev, void *mqd)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct v9_sdma_mqd *m;
 	uint32_t sdma_rlc_reg_offset;
 	uint32_t sdma_rlc_rb_cntl;
@@ -541,12 +521,11 @@ static bool kgd_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd)
 	return false;
 }
 
-int kgd_gfx_v9_hqd_destroy(struct kgd_dev *kgd, void *mqd,
+int kgd_gfx_v9_hqd_destroy(struct amdgpu_device *adev, void *mqd,
 				enum kfd_preempt_type reset_type,
 				unsigned int utimeout, uint32_t pipe_id,
-				uint32_t queue_id)
+				uint32_t queue_id, uint32_t inst)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	enum hqd_dequeue_request_type type;
 	unsigned long end_jiffies;
 	uint32_t temp;
@@ -555,10 +534,10 @@ int kgd_gfx_v9_hqd_destroy(struct kgd_dev *kgd, void *mqd,
 	if (amdgpu_in_reset(adev))
 		return -EIO;
 
-	acquire_queue(kgd, pipe_id, queue_id);
+	kgd_gfx_v9_acquire_queue(adev, pipe_id, queue_id, inst);
 
 	if (m->cp_hqd_vmid == 0)
-		WREG32_FIELD15_RLC(GC, 0, RLC_CP_SCHEDULERS, scheduler1, 0);
+		WREG32_FIELD15_RLC(GC, GET_INST(GC, inst), RLC_CP_SCHEDULERS, scheduler1, 0);
 
 	switch (reset_type) {
 	case KFD_PREEMPT_TYPE_WAVEFRONT_DRAIN:
@@ -575,29 +554,28 @@ int kgd_gfx_v9_hqd_destroy(struct kgd_dev *kgd, void *mqd,
 		break;
 	}
 
-	WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_DEQUEUE_REQUEST), type);
+	WREG32_SOC15_RLC(GC, GET_INST(GC, inst), mmCP_HQD_DEQUEUE_REQUEST, type);
 
 	end_jiffies = (utimeout * HZ / 1000) + jiffies;
 	while (true) {
-		temp = RREG32(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_ACTIVE));
+		temp = RREG32_SOC15(GC, GET_INST(GC, inst), mmCP_HQD_ACTIVE);
 		if (!(temp & CP_HQD_ACTIVE__ACTIVE_MASK))
 			break;
 		if (time_after(jiffies, end_jiffies)) {
 			pr_err("cp queue preemption time out.\n");
-			release_queue(kgd);
+			kgd_gfx_v9_release_queue(adev, inst);
 			return -ETIME;
 		}
 		usleep_range(500, 1000);
 	}
 
-	release_queue(kgd);
+	kgd_gfx_v9_release_queue(adev, inst);
 	return 0;
 }
 
-static int kgd_hqd_sdma_destroy(struct kgd_dev *kgd, void *mqd,
+static int kgd_hqd_sdma_destroy(struct amdgpu_device *adev, void *mqd,
 				unsigned int utimeout)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct v9_sdma_mqd *m;
 	uint32_t sdma_rlc_reg_offset;
 	uint32_t temp;
@@ -634,11 +612,10 @@ static int kgd_hqd_sdma_destroy(struct kgd_dev *kgd, void *mqd,
 	return 0;
 }
 
-bool kgd_gfx_v9_get_atc_vmid_pasid_mapping_info(struct kgd_dev *kgd,
+bool kgd_gfx_v9_get_atc_vmid_pasid_mapping_info(struct amdgpu_device *adev,
 					uint8_t vmid, uint16_t *p_pasid)
 {
 	uint32_t value;
-	struct amdgpu_device *adev = (struct amdgpu_device *) kgd;
 
 	value = RREG32(SOC15_REG_OFFSET(ATHUB, 0, mmATC_VMID0_PASID_MAPPING)
 		     + vmid);
@@ -647,57 +624,295 @@ bool kgd_gfx_v9_get_atc_vmid_pasid_mapping_info(struct kgd_dev *kgd,
 	return !!(value & ATC_VMID0_PASID_MAPPING__VALID_MASK);
 }
 
-int kgd_gfx_v9_address_watch_disable(struct kgd_dev *kgd)
+int kgd_gfx_v9_wave_control_execute(struct amdgpu_device *adev,
+					uint32_t gfx_index_val,
+					uint32_t sq_cmd, uint32_t inst)
+{
+	uint32_t data = 0;
+
+	mutex_lock(&adev->grbm_idx_mutex);
+
+	WREG32_SOC15_RLC_SHADOW(GC, GET_INST(GC, inst), mmGRBM_GFX_INDEX, gfx_index_val);
+	WREG32_SOC15(GC, GET_INST(GC, inst), mmSQ_CMD, sq_cmd);
+
+	data = REG_SET_FIELD(data, GRBM_GFX_INDEX,
+		INSTANCE_BROADCAST_WRITES, 1);
+	data = REG_SET_FIELD(data, GRBM_GFX_INDEX,
+		SH_BROADCAST_WRITES, 1);
+	data = REG_SET_FIELD(data, GRBM_GFX_INDEX,
+		SE_BROADCAST_WRITES, 1);
+
+	WREG32_SOC15_RLC_SHADOW(GC, GET_INST(GC, inst), mmGRBM_GFX_INDEX, data);
+	mutex_unlock(&adev->grbm_idx_mutex);
+
+	return 0;
+}
+
+/*
+ * GFX9 helper for wave launch stall requirements on debug trap setting.
+ *
+ * vmid:
+ *   Target VMID to stall/unstall.
+ *
+ * stall:
+ *   0-unstall wave launch (enable), 1-stall wave launch (disable).
+ *   After wavefront launch has been stalled, allocated waves must drain from
+ *   SPI in order for debug trap settings to take effect on those waves.
+ *   This is roughly a ~96 clock cycle wait on SPI where a read on
+ *   SPI_GDBG_WAVE_CNTL translates to ~32 clock cycles.
+ *   KGD_GFX_V9_WAVE_LAUNCH_SPI_DRAIN_LATENCY indicates the number of reads required.
+ *
+ *   NOTE: We can afford to clear the entire STALL_VMID field on unstall
+ *   because GFX9.4.1 cannot support multi-process debugging due to trap
+ *   configuration and masking being limited to global scope.  Always assume
+ *   single process conditions.
+ */
+#define KGD_GFX_V9_WAVE_LAUNCH_SPI_DRAIN_LATENCY	3
+void kgd_gfx_v9_set_wave_launch_stall(struct amdgpu_device *adev,
+					uint32_t vmid,
+					bool stall)
+{
+	int i;
+	uint32_t data = RREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL));
+
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 1))
+		data = REG_SET_FIELD(data, SPI_GDBG_WAVE_CNTL, STALL_VMID,
+							stall ? 1 << vmid : 0);
+	else
+		data = REG_SET_FIELD(data, SPI_GDBG_WAVE_CNTL, STALL_RA,
+							stall ? 1 : 0);
+
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL), data);
+
+	if (!stall)
+		return;
+
+	for (i = 0; i < KGD_GFX_V9_WAVE_LAUNCH_SPI_DRAIN_LATENCY; i++)
+		RREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL));
+}
+
+/*
+ * restore_dbg_registers is ignored here but is a general interface requirement
+ * for devices that support GFXOFF and where the RLC save/restore list
+ * does not support hw registers for debugging i.e. the driver has to manually
+ * initialize the debug mode registers after it has disabled GFX off during the
+ * debug session.
+ */
+uint32_t kgd_gfx_v9_enable_debug_trap(struct amdgpu_device *adev,
+				bool restore_dbg_registers,
+				uint32_t vmid)
 {
+	mutex_lock(&adev->grbm_idx_mutex);
+
+	kgd_gfx_v9_set_wave_launch_stall(adev, vmid, true);
+
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_MASK), 0);
+
+	kgd_gfx_v9_set_wave_launch_stall(adev, vmid, false);
+
+	mutex_unlock(&adev->grbm_idx_mutex);
+
 	return 0;
 }
 
-int kgd_gfx_v9_address_watch_execute(struct kgd_dev *kgd,
-					unsigned int watch_point_id,
-					uint32_t cntl_val,
-					uint32_t addr_hi,
-					uint32_t addr_lo)
+/*
+ * keep_trap_enabled is ignored here but is a general interface requirement
+ * for devices that support multi-process debugging where the performance
+ * overhead from trap temporary setup needs to be bypassed when the debug
+ * session has ended.
+ */
+uint32_t kgd_gfx_v9_disable_debug_trap(struct amdgpu_device *adev,
+					bool keep_trap_enabled,
+					uint32_t vmid)
 {
+	mutex_lock(&adev->grbm_idx_mutex);
+
+	kgd_gfx_v9_set_wave_launch_stall(adev, vmid, true);
+
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_MASK), 0);
+
+	kgd_gfx_v9_set_wave_launch_stall(adev, vmid, false);
+
+	mutex_unlock(&adev->grbm_idx_mutex);
+
 	return 0;
 }
 
-int kgd_gfx_v9_wave_control_execute(struct kgd_dev *kgd,
-					uint32_t gfx_index_val,
-					uint32_t sq_cmd)
+int kgd_gfx_v9_validate_trap_override_request(struct amdgpu_device *adev,
+					uint32_t trap_override,
+					uint32_t *trap_mask_supported)
+{
+	*trap_mask_supported &= KFD_DBG_TRAP_MASK_DBG_ADDRESS_WATCH;
+
+	/* The SPI_GDBG_TRAP_MASK register is global and affects all
+	 * processes. Only allow OR-ing the address-watch bit, since
+	 * this only affects processes under the debugger. Other bits
+	 * should stay 0 to avoid the debugger interfering with other
+	 * processes.
+	 */
+	if (trap_override != KFD_DBG_TRAP_OVERRIDE_OR)
+		return -EINVAL;
+
+	return 0;
+}
+
+uint32_t kgd_gfx_v9_set_wave_launch_trap_override(struct amdgpu_device *adev,
+					     uint32_t vmid,
+					     uint32_t trap_override,
+					     uint32_t trap_mask_bits,
+					     uint32_t trap_mask_request,
+					     uint32_t *trap_mask_prev,
+					     uint32_t kfd_dbg_cntl_prev)
+{
+	uint32_t data, wave_cntl_prev;
+
+	mutex_lock(&adev->grbm_idx_mutex);
+
+	wave_cntl_prev = RREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL));
+
+	kgd_gfx_v9_set_wave_launch_stall(adev, vmid, true);
+
+	data = RREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_MASK));
+	*trap_mask_prev = REG_GET_FIELD(data, SPI_GDBG_TRAP_MASK, EXCP_EN);
+
+	trap_mask_bits = (trap_mask_bits & trap_mask_request) |
+		(*trap_mask_prev & ~trap_mask_request);
+
+	data = REG_SET_FIELD(data, SPI_GDBG_TRAP_MASK, EXCP_EN, trap_mask_bits);
+	data = REG_SET_FIELD(data, SPI_GDBG_TRAP_MASK, REPLACE, trap_override);
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_MASK), data);
+
+	/* We need to preserve wave launch mode stall settings. */
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL), wave_cntl_prev);
+
+	mutex_unlock(&adev->grbm_idx_mutex);
+
+	return 0;
+}
+
+uint32_t kgd_gfx_v9_set_wave_launch_mode(struct amdgpu_device *adev,
+					uint8_t wave_launch_mode,
+					uint32_t vmid)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t data = 0;
+	bool is_mode_set = !!wave_launch_mode;
 
 	mutex_lock(&adev->grbm_idx_mutex);
 
-	WREG32_SOC15_RLC_SHADOW(GC, 0, mmGRBM_GFX_INDEX, gfx_index_val);
-	WREG32(SOC15_REG_OFFSET(GC, 0, mmSQ_CMD), sq_cmd);
+	kgd_gfx_v9_set_wave_launch_stall(adev, vmid, true);
 
-	data = REG_SET_FIELD(data, GRBM_GFX_INDEX,
-		INSTANCE_BROADCAST_WRITES, 1);
-	data = REG_SET_FIELD(data, GRBM_GFX_INDEX,
-		SH_BROADCAST_WRITES, 1);
-	data = REG_SET_FIELD(data, GRBM_GFX_INDEX,
-		SE_BROADCAST_WRITES, 1);
+	data = REG_SET_FIELD(data, SPI_GDBG_WAVE_CNTL2,
+		VMID_MASK, is_mode_set ? 1 << vmid : 0);
+	data = REG_SET_FIELD(data, SPI_GDBG_WAVE_CNTL2,
+		MODE, is_mode_set ? wave_launch_mode : 0);
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_WAVE_CNTL2), data);
+
+	kgd_gfx_v9_set_wave_launch_stall(adev, vmid, false);
 
-	WREG32_SOC15_RLC_SHADOW(GC, 0, mmGRBM_GFX_INDEX, data);
 	mutex_unlock(&adev->grbm_idx_mutex);
 
 	return 0;
 }
 
-uint32_t kgd_gfx_v9_address_watch_get_offset(struct kgd_dev *kgd,
-					unsigned int watch_point_id,
-					unsigned int reg_offset)
+#define TCP_WATCH_STRIDE (mmTCP_WATCH1_ADDR_H - mmTCP_WATCH0_ADDR_H)
+uint32_t kgd_gfx_v9_set_address_watch(struct amdgpu_device *adev,
+					uint64_t watch_address,
+					uint32_t watch_address_mask,
+					uint32_t watch_id,
+					uint32_t watch_mode,
+					uint32_t debug_vmid,
+					uint32_t inst)
 {
+	uint32_t watch_address_high;
+	uint32_t watch_address_low;
+	uint32_t watch_address_cntl;
+
+	watch_address_cntl = 0;
+
+	watch_address_low = lower_32_bits(watch_address);
+	watch_address_high = upper_32_bits(watch_address) & 0xffff;
+
+	watch_address_cntl = REG_SET_FIELD(watch_address_cntl,
+			TCP_WATCH0_CNTL,
+			VMID,
+			debug_vmid);
+	watch_address_cntl = REG_SET_FIELD(watch_address_cntl,
+			TCP_WATCH0_CNTL,
+			MODE,
+			watch_mode);
+	watch_address_cntl = REG_SET_FIELD(watch_address_cntl,
+			TCP_WATCH0_CNTL,
+			MASK,
+			watch_address_mask >> 6);
+
+	/* Turning off this watch point until we set all the registers */
+	watch_address_cntl = REG_SET_FIELD(watch_address_cntl,
+			TCP_WATCH0_CNTL,
+			VALID,
+			0);
+
+	WREG32_RLC((SOC15_REG_OFFSET(GC, 0, mmTCP_WATCH0_CNTL) +
+			(watch_id * TCP_WATCH_STRIDE)),
+			watch_address_cntl);
+
+	WREG32_RLC((SOC15_REG_OFFSET(GC, 0, mmTCP_WATCH0_ADDR_H) +
+			(watch_id * TCP_WATCH_STRIDE)),
+			watch_address_high);
+
+	WREG32_RLC((SOC15_REG_OFFSET(GC, 0, mmTCP_WATCH0_ADDR_L) +
+			(watch_id * TCP_WATCH_STRIDE)),
+			watch_address_low);
+
+	/* Enable the watch point */
+	watch_address_cntl = REG_SET_FIELD(watch_address_cntl,
+			TCP_WATCH0_CNTL,
+			VALID,
+			1);
+
+	WREG32_RLC((SOC15_REG_OFFSET(GC, 0, mmTCP_WATCH0_CNTL) +
+			(watch_id * TCP_WATCH_STRIDE)),
+			watch_address_cntl);
+
 	return 0;
 }
 
-void kgd_gfx_v9_set_vm_context_page_table_base(struct kgd_dev *kgd,
-			uint32_t vmid, uint64_t page_table_base)
+uint32_t kgd_gfx_v9_clear_address_watch(struct amdgpu_device *adev,
+					uint32_t watch_id)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
+	uint32_t watch_address_cntl;
+
+	watch_address_cntl = 0;
+
+	WREG32_RLC((SOC15_REG_OFFSET(GC, 0, mmTCP_WATCH0_CNTL) +
+			(watch_id * TCP_WATCH_STRIDE)),
+			watch_address_cntl);
+
+	return 0;
+}
+
+/* kgd_gfx_v9_get_iq_wait_times: Returns the mmCP_IQ_WAIT_TIME1/2 values
+ * The values read are:
+ *     ib_offload_wait_time     -- Wait Count for Indirect Buffer Offloads.
+ *     atomic_offload_wait_time -- Wait Count for L2 and GDS Atomics Offloads.
+ *     wrm_offload_wait_time    -- Wait Count for WAIT_REG_MEM Offloads.
+ *     gws_wait_time            -- Wait Count for Global Wave Syncs.
+ *     que_sleep_wait_time      -- Wait Count for Dequeue Retry.
+ *     sch_wave_wait_time       -- Wait Count for Scheduling Wave Message.
+ *     sem_rearm_wait_time      -- Wait Count for Semaphore re-arm.
+ *     deq_retry_wait_time      -- Wait Count for Global Wave Syncs.
+ */
+void kgd_gfx_v9_get_iq_wait_times(struct amdgpu_device *adev,
+					uint32_t *wait_times,
+					uint32_t inst)
 
+{
+	*wait_times = RREG32_SOC15_RLC(GC, GET_INST(GC, inst),
+			mmCP_IQ_WAIT_TIME2);
+}
+
+void kgd_gfx_v9_set_vm_context_page_table_base(struct amdgpu_device *adev,
+			uint32_t vmid, uint64_t page_table_base)
+{
 	if (!amdgpu_amdkfd_is_kfd_vmid(adev, vmid)) {
 		pr_err("trying to set page table base for wrong VMID %u\n",
 		       vmid);
@@ -729,33 +944,34 @@ static void unlock_spi_csq_mutexes(struct amdgpu_device *adev)
  *
  * @adev: Handle of device whose registers are to be read
  * @queue_idx: Index of queue in the queue-map bit-field
- * @wave_cnt: Output parameter updated with number of waves in flight
- * @vmid: Output parameter updated with VMID of queue whose wave count
- * is being collected
+ * @queue_cnt: Stores the wave count and doorbell offset for an active queue
+ * @inst: xcc's instance number on a multi-XCC setup
  */
 static void get_wave_count(struct amdgpu_device *adev, int queue_idx,
-		int *wave_cnt, int *vmid)
+		struct kfd_cu_occupancy *queue_cnt, uint32_t inst)
 {
 	int pipe_idx;
 	int queue_slot;
 	unsigned int reg_val;
-
+	unsigned int wave_cnt;
 	/*
 	 * Program GRBM with appropriate MEID, PIPEID, QUEUEID and VMID
 	 * parameters to read out waves in flight. Get VMID if there are
 	 * non-zero waves in flight.
 	 */
-	*vmid = 0xFF;
-	*wave_cnt = 0;
 	pipe_idx = queue_idx / adev->gfx.mec.num_queue_per_pipe;
 	queue_slot = queue_idx % adev->gfx.mec.num_queue_per_pipe;
-	soc15_grbm_select(adev, 1, pipe_idx, queue_slot, 0);
-	reg_val = RREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_CSQ_WF_ACTIVE_COUNT_0) +
-			 queue_slot);
-	*wave_cnt = reg_val & SPI_CSQ_WF_ACTIVE_COUNT_0__COUNT_MASK;
-	if (*wave_cnt != 0)
-		*vmid = (RREG32_SOC15(GC, 0, mmCP_HQD_VMID) &
-			 CP_HQD_VMID__VMID_MASK) >> CP_HQD_VMID__VMID__SHIFT;
+	soc15_grbm_select(adev, 1, pipe_idx, queue_slot, 0, GET_INST(GC, inst));
+	reg_val = RREG32_SOC15_IP(GC, SOC15_REG_OFFSET(GC, GET_INST(GC, inst),
+				  mmSPI_CSQ_WF_ACTIVE_COUNT_0) + queue_slot);
+	wave_cnt = reg_val & SPI_CSQ_WF_ACTIVE_COUNT_0__COUNT_MASK;
+	if (wave_cnt != 0) {
+		queue_cnt->wave_cnt += wave_cnt;
+		queue_cnt->doorbell_off =
+			(RREG32_SOC15(GC, GET_INST(GC, inst), mmCP_HQD_PQ_DOORBELL_CONTROL) &
+			 CP_HQD_PQ_DOORBELL_CONTROL__DOORBELL_OFFSET_MASK) >>
+			 CP_HQD_PQ_DOORBELL_CONTROL__DOORBELL_OFFSET__SHIFT;
+	}
 }
 
 /**
@@ -764,16 +980,16 @@ static void get_wave_count(struct amdgpu_device *adev, int queue_idx,
  * process whose pasid is provided as a parameter. The process could have ZERO
  * or more queues running and submitting waves to compute units.
  *
- * @kgd: Handle of device from which to get number of waves in flight
- * @pasid: Identifies the process for which this query call is invoked
- * @pasid_wave_cnt: Output parameter updated with number of waves in flight that
- * belong to process with given pasid
+ * @adev: Handle of device from which to get number of waves in flight
+ * @cu_occupancy: Array that gets filled with wave_cnt and doorbell offset
+ *		  for comparison later.
  * @max_waves_per_cu: Output parameter updated with maximum number of waves
- * possible per Compute Unit
+ *                    possible per Compute Unit
+ * @inst: xcc's instance number on a multi-XCC setup
  *
  * Note: It's possible that the device has too many queues (oversubscription)
  * in which case a VMID could be remapped to a different PASID. This could lead
- * to an iaccurate wave count. Following is a high-level sequence:
+ * to an inaccurate wave count. Following is a high-level sequence:
  *    Time T1: vmid = getVmid(); vmid is associated with Pasid P1
  *    Time T2: passId = getPasId(vmid); vmid is associated with Pasid P2
  * In the sequence above wave count obtained from time T1 will be incorrectly
@@ -794,118 +1010,224 @@ static void get_wave_count(struct amdgpu_device *adev, int queue_idx,
  *    number of waves that are in flight for the queue at specified index. The
  *    index ranges from 0 to 7.
  *
- *    If non-zero waves are in flight, read CP_HQD_VMID register to obtain VMID
- *    of the wave(s).
+ *    If non-zero waves are in flight, store the corresponding doorbell offset
+ *    of the queue, along with the wave count.
  *
- *    Determine if VMID from above step maps to pasid provided as parameter. If
- *    it matches agrregate the wave count. That the VMID will not match pasid is
- *    a normal condition i.e. a device is expected to support multiple queues
- *    from multiple proceses.
+ *    Determine if the queue belongs to the process by comparing the doorbell
+ *    offset against the process's queues. If it matches, aggregate the wave
+ *    count for the process.
  *
  *  Reading registers referenced above involves programming GRBM appropriately
  */
-void kgd_gfx_v9_get_cu_occupancy(struct kgd_dev *kgd, int pasid,
-		int *pasid_wave_cnt, int *max_waves_per_cu)
+void kgd_gfx_v9_get_cu_occupancy(struct amdgpu_device *adev,
+				 struct kfd_cu_occupancy *cu_occupancy,
+				 int *max_waves_per_cu, uint32_t inst)
 {
 	int qidx;
-	int vmid;
 	int se_idx;
-	int sh_idx;
 	int se_cnt;
-	int sh_cnt;
-	int wave_cnt;
 	int queue_map;
-	int pasid_tmp;
 	int max_queue_cnt;
-	int vmid_wave_cnt = 0;
-	struct amdgpu_device *adev;
-	DECLARE_BITMAP(cp_queue_bitmap, KGD_MAX_QUEUES);
+	DECLARE_BITMAP(cp_queue_bitmap, AMDGPU_MAX_QUEUES);
 
-	adev = get_amdgpu_device(kgd);
 	lock_spi_csq_mutexes(adev);
-	soc15_grbm_select(adev, 1, 0, 0, 0);
+	soc15_grbm_select(adev, 1, 0, 0, 0, GET_INST(GC, inst));
 
 	/*
 	 * Iterate through the shader engines and arrays of the device
 	 * to get number of waves in flight
 	 */
-	bitmap_complement(cp_queue_bitmap, adev->gfx.mec.queue_bitmap,
-			  KGD_MAX_QUEUES);
+	bitmap_complement(cp_queue_bitmap, adev->gfx.mec_bitmap[0].queue_bitmap,
+			  AMDGPU_MAX_QUEUES);
 	max_queue_cnt = adev->gfx.mec.num_pipe_per_mec *
 			adev->gfx.mec.num_queue_per_pipe;
-	sh_cnt = adev->gfx.config.max_sh_per_se;
 	se_cnt = adev->gfx.config.max_shader_engines;
 	for (se_idx = 0; se_idx < se_cnt; se_idx++) {
-		for (sh_idx = 0; sh_idx < sh_cnt; sh_idx++) {
+		amdgpu_gfx_select_se_sh(adev, se_idx, 0, 0xffffffff, inst);
+		queue_map = RREG32_SOC15(GC, GET_INST(GC, inst), mmSPI_CSQ_WF_ACTIVE_STATUS);
+
+		/*
+		 * Assumption: queue map encodes following schema: four
+		 * pipes per each micro-engine, with each pipe mapping
+		 * eight queues. This schema is true for GFX9 devices
+		 * and must be verified for newer device families
+		 */
+		for (qidx = 0; qidx < max_queue_cnt; qidx++) {
+			/* Skip qeueus that are not associated with
+			 * compute functions
+			 */
+			if (!test_bit(qidx, cp_queue_bitmap))
+				continue;
 
-			gfx_v9_0_select_se_sh(adev, se_idx, sh_idx, 0xffffffff);
-			queue_map = RREG32(SOC15_REG_OFFSET(GC, 0,
-					   mmSPI_CSQ_WF_ACTIVE_STATUS));
+			if (!(queue_map & (1 << qidx)))
+				continue;
 
-			/*
-			 * Assumption: queue map encodes following schema: four
-			 * pipes per each micro-engine, with each pipe mapping
-			 * eight queues. This schema is true for GFX9 devices
-			 * and must be verified for newer device families
-			 */
-			for (qidx = 0; qidx < max_queue_cnt; qidx++) {
-
-				/* Skip qeueus that are not associated with
-				 * compute functions
-				 */
-				if (!test_bit(qidx, cp_queue_bitmap))
-					continue;
-
-				if (!(queue_map & (1 << qidx)))
-					continue;
-
-				/* Get number of waves in flight and aggregate them */
-				get_wave_count(adev, qidx, &wave_cnt, &vmid);
-				if (wave_cnt != 0) {
-					pasid_tmp =
-					  RREG32(SOC15_REG_OFFSET(OSSSYS, 0,
-						 mmIH_VMID_0_LUT) + vmid);
-					if (pasid_tmp == pasid)
-						vmid_wave_cnt += wave_cnt;
-				}
-			}
+			/* Get number of waves in flight and aggregate them */
+			get_wave_count(adev, qidx, &cu_occupancy[qidx],
+					inst);
 		}
 	}
 
-	gfx_v9_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
-	soc15_grbm_select(adev, 0, 0, 0, 0);
+	amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, inst);
+	soc15_grbm_select(adev, 0, 0, 0, 0, GET_INST(GC, inst));
 	unlock_spi_csq_mutexes(adev);
 
 	/* Update the output parameters and return */
-	*pasid_wave_cnt = vmid_wave_cnt;
 	*max_waves_per_cu = adev->gfx.cu_info.simd_per_cu *
 				adev->gfx.cu_info.max_waves_per_simd;
 }
 
-void kgd_gfx_v9_program_trap_handler_settings(struct kgd_dev *kgd,
-                        uint32_t vmid, uint64_t tba_addr, uint64_t tma_addr)
+void kgd_gfx_v9_build_dequeue_wait_counts_packet_info(struct amdgpu_device *adev,
+		uint32_t wait_times,
+		uint32_t sch_wave,
+		uint32_t que_sleep,
+		uint32_t *reg_offset,
+		uint32_t *reg_data)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
+	*reg_data = wait_times;
+
+	if (sch_wave)
+		*reg_data = REG_SET_FIELD(*reg_data,
+				CP_IQ_WAIT_TIME2,
+				SCH_WAVE,
+				sch_wave);
+	if (que_sleep)
+		*reg_data = REG_SET_FIELD(*reg_data,
+				CP_IQ_WAIT_TIME2,
+				QUE_SLEEP,
+				que_sleep);
+
+	*reg_offset = SOC15_REG_OFFSET(GC, 0, mmCP_IQ_WAIT_TIME2);
+}
 
-	lock_srbm(kgd, 0, 0, 0, vmid);
+void kgd_gfx_v9_program_trap_handler_settings(struct amdgpu_device *adev,
+		uint32_t vmid, uint64_t tba_addr, uint64_t tma_addr, uint32_t inst)
+{
+	kgd_gfx_v9_lock_srbm(adev, 0, 0, 0, vmid, inst);
 
 	/*
 	 * Program TBA registers
 	 */
-	WREG32(SOC15_REG_OFFSET(GC, 0, mmSQ_SHADER_TBA_LO),
-                        lower_32_bits(tba_addr >> 8));
-	WREG32(SOC15_REG_OFFSET(GC, 0, mmSQ_SHADER_TBA_HI),
-                        upper_32_bits(tba_addr >> 8));
+	WREG32_SOC15(GC, GET_INST(GC, inst), mmSQ_SHADER_TBA_LO,
+			lower_32_bits(tba_addr >> 8));
+	WREG32_SOC15(GC, GET_INST(GC, inst), mmSQ_SHADER_TBA_HI,
+			upper_32_bits(tba_addr >> 8));
 
 	/*
 	 * Program TMA registers
 	 */
-	WREG32(SOC15_REG_OFFSET(GC, 0, mmSQ_SHADER_TMA_LO),
+	WREG32_SOC15(GC, GET_INST(GC, inst), mmSQ_SHADER_TMA_LO,
 			lower_32_bits(tma_addr >> 8));
-	WREG32(SOC15_REG_OFFSET(GC, 0, mmSQ_SHADER_TMA_HI),
+	WREG32_SOC15(GC, GET_INST(GC, inst), mmSQ_SHADER_TMA_HI,
 			upper_32_bits(tma_addr >> 8));
 
-	unlock_srbm(kgd);
+	kgd_gfx_v9_unlock_srbm(adev, inst);
+}
+
+uint64_t kgd_gfx_v9_hqd_get_pq_addr(struct amdgpu_device *adev,
+				    uint32_t pipe_id, uint32_t queue_id,
+				    uint32_t inst)
+{
+	uint32_t low, high;
+	uint64_t queue_addr = 0;
+
+	kgd_gfx_v9_acquire_queue(adev, pipe_id, queue_id, inst);
+	amdgpu_gfx_rlc_enter_safe_mode(adev, inst);
+
+	if (!RREG32_SOC15(GC, GET_INST(GC, inst), mmCP_HQD_ACTIVE))
+		goto unlock_out;
+
+	low = RREG32_SOC15(GC, GET_INST(GC, inst), mmCP_HQD_PQ_BASE);
+	high = RREG32_SOC15(GC, GET_INST(GC, inst), mmCP_HQD_PQ_BASE_HI);
+
+	/* only concerned with user queues. */
+	if (!high)
+		goto unlock_out;
+
+	queue_addr = (((queue_addr | high) << 32) | low) << 8;
+
+unlock_out:
+	amdgpu_gfx_rlc_exit_safe_mode(adev, inst);
+	kgd_gfx_v9_release_queue(adev, inst);
+
+	return queue_addr;
+}
+
+/* assume queue acquired  */
+static int kgd_gfx_v9_hqd_dequeue_wait(struct amdgpu_device *adev, uint32_t inst,
+				       unsigned int utimeout)
+{
+	unsigned long end_jiffies = (utimeout * HZ / 1000) + jiffies;
+
+	while (true) {
+		uint32_t temp = RREG32_SOC15(GC, GET_INST(GC, inst), mmCP_HQD_ACTIVE);
+
+		if (!(temp & CP_HQD_ACTIVE__ACTIVE_MASK))
+			return 0;
+
+		if (time_after(jiffies, end_jiffies))
+			return -ETIME;
+
+		usleep_range(500, 1000);
+	}
+}
+
+uint64_t kgd_gfx_v9_hqd_reset(struct amdgpu_device *adev,
+			      uint32_t pipe_id, uint32_t queue_id,
+			      uint32_t inst, unsigned int utimeout)
+{
+	uint32_t low, high, pipe_reset_data = 0;
+	uint64_t queue_addr = 0;
+
+	kgd_gfx_v9_acquire_queue(adev, pipe_id, queue_id, inst);
+	amdgpu_gfx_rlc_enter_safe_mode(adev, inst);
+
+	if (!RREG32_SOC15(GC, GET_INST(GC, inst), mmCP_HQD_ACTIVE))
+		goto unlock_out;
+
+	low = RREG32_SOC15(GC, GET_INST(GC, inst), mmCP_HQD_PQ_BASE);
+	high = RREG32_SOC15(GC, GET_INST(GC, inst), mmCP_HQD_PQ_BASE_HI);
+
+	/* only concerned with user queues. */
+	if (!high)
+		goto unlock_out;
+
+	queue_addr = (((queue_addr | high) << 32) | low) << 8;
+
+	pr_debug("Attempting queue reset on XCC %i pipe id %i queue id %i\n",
+		 inst, pipe_id, queue_id);
+
+	/* assume previous dequeue request issued will take affect after reset */
+	WREG32_SOC15(GC, GET_INST(GC, inst), mmSPI_COMPUTE_QUEUE_RESET, 0x1);
+
+	if (!kgd_gfx_v9_hqd_dequeue_wait(adev, inst, utimeout))
+		goto unlock_out;
+
+	pr_debug("Attempting pipe reset on XCC %i pipe id %i\n", inst, pipe_id);
+
+	pipe_reset_data = REG_SET_FIELD(pipe_reset_data, CP_MEC_CNTL, MEC_ME1_PIPE0_RESET, 1);
+	pipe_reset_data = pipe_reset_data << pipe_id;
+
+	WREG32_SOC15(GC, GET_INST(GC, inst), mmCP_MEC_CNTL, pipe_reset_data);
+	WREG32_SOC15(GC, GET_INST(GC, inst), mmCP_MEC_CNTL, 0);
+
+	if (kgd_gfx_v9_hqd_dequeue_wait(adev, inst, utimeout))
+		queue_addr = 0;
+
+unlock_out:
+	pr_debug("queue reset on XCC %i pipe id %i queue id %i %s\n",
+		 inst, pipe_id, queue_id, !!queue_addr ? "succeeded!" : "failed!");
+	amdgpu_gfx_rlc_exit_safe_mode(adev, inst);
+	kgd_gfx_v9_release_queue(adev, inst);
+
+	return queue_addr;
+}
+
+uint32_t kgd_gfx_v9_hqd_sdma_get_doorbell(struct amdgpu_device *adev,
+					  int engine, int queue)
+
+{
+	return 0;
 }
 
 const struct kfd2kgd_calls gfx_v9_kfd2kgd = {
@@ -921,13 +1243,22 @@ const struct kfd2kgd_calls gfx_v9_kfd2kgd = {
 	.hqd_sdma_is_occupied = kgd_hqd_sdma_is_occupied,
 	.hqd_destroy = kgd_gfx_v9_hqd_destroy,
 	.hqd_sdma_destroy = kgd_hqd_sdma_destroy,
-	.address_watch_disable = kgd_gfx_v9_address_watch_disable,
-	.address_watch_execute = kgd_gfx_v9_address_watch_execute,
 	.wave_control_execute = kgd_gfx_v9_wave_control_execute,
-	.address_watch_get_offset = kgd_gfx_v9_address_watch_get_offset,
 	.get_atc_vmid_pasid_mapping_info =
 			kgd_gfx_v9_get_atc_vmid_pasid_mapping_info,
 	.set_vm_context_page_table_base = kgd_gfx_v9_set_vm_context_page_table_base,
+	.enable_debug_trap = kgd_gfx_v9_enable_debug_trap,
+	.disable_debug_trap = kgd_gfx_v9_disable_debug_trap,
+	.validate_trap_override_request = kgd_gfx_v9_validate_trap_override_request,
+	.set_wave_launch_trap_override = kgd_gfx_v9_set_wave_launch_trap_override,
+	.set_wave_launch_mode = kgd_gfx_v9_set_wave_launch_mode,
+	.set_address_watch = kgd_gfx_v9_set_address_watch,
+	.clear_address_watch = kgd_gfx_v9_clear_address_watch,
+	.get_iq_wait_times = kgd_gfx_v9_get_iq_wait_times,
+	.build_dequeue_wait_counts_packet_info = kgd_gfx_v9_build_dequeue_wait_counts_packet_info,
 	.get_cu_occupancy = kgd_gfx_v9_get_cu_occupancy,
 	.program_trap_handler_settings = kgd_gfx_v9_program_trap_handler_settings,
+	.hqd_get_pq_addr = kgd_gfx_v9_hqd_get_pq_addr,
+	.hqd_reset = kgd_gfx_v9_hqd_reset,
+	.hqd_sdma_get_doorbell = kgd_gfx_v9_hqd_sdma_get_doorbell
 };
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.h
index c63591106879..704452ca62f8 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.h
@@ -20,50 +20,97 @@
  * OTHER DEALINGS IN THE SOFTWARE.
  */
 
-
-
-void kgd_gfx_v9_program_sh_mem_settings(struct kgd_dev *kgd, uint32_t vmid,
+void kgd_gfx_v9_program_sh_mem_settings(struct amdgpu_device *adev, uint32_t vmid,
 		uint32_t sh_mem_config,
 		uint32_t sh_mem_ape1_base, uint32_t sh_mem_ape1_limit,
-		uint32_t sh_mem_bases);
-int kgd_gfx_v9_set_pasid_vmid_mapping(struct kgd_dev *kgd, u32 pasid,
-		unsigned int vmid);
-int kgd_gfx_v9_init_interrupts(struct kgd_dev *kgd, uint32_t pipe_id);
-int kgd_gfx_v9_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
+		uint32_t sh_mem_bases, uint32_t inst);
+int kgd_gfx_v9_set_pasid_vmid_mapping(struct amdgpu_device *adev, u32 pasid,
+		unsigned int vmid, uint32_t inst);
+int kgd_gfx_v9_init_interrupts(struct amdgpu_device *adev, uint32_t pipe_id,
+				uint32_t inst);
+int kgd_gfx_v9_hqd_load(struct amdgpu_device *adev, void *mqd, uint32_t pipe_id,
 			uint32_t queue_id, uint32_t __user *wptr,
 			uint32_t wptr_shift, uint32_t wptr_mask,
-			struct mm_struct *mm);
-int kgd_gfx_v9_hiq_mqd_load(struct kgd_dev *kgd, void *mqd,
+			struct mm_struct *mm, uint32_t inst);
+int kgd_gfx_v9_hiq_mqd_load(struct amdgpu_device *adev, void *mqd,
 			    uint32_t pipe_id, uint32_t queue_id,
-			    uint32_t doorbell_off);
-int kgd_gfx_v9_hqd_dump(struct kgd_dev *kgd,
+			    uint32_t doorbell_off, uint32_t inst);
+int kgd_gfx_v9_hqd_dump(struct amdgpu_device *adev,
 			uint32_t pipe_id, uint32_t queue_id,
-			uint32_t (**dump)[2], uint32_t *n_regs);
-bool kgd_gfx_v9_hqd_is_occupied(struct kgd_dev *kgd, uint64_t queue_address,
-		uint32_t pipe_id, uint32_t queue_id);
-int kgd_gfx_v9_hqd_destroy(struct kgd_dev *kgd, void *mqd,
+			uint32_t (**dump)[2], uint32_t *n_regs, uint32_t inst);
+bool kgd_gfx_v9_hqd_is_occupied(struct amdgpu_device *adev,
+			uint64_t queue_address, uint32_t pipe_id,
+			uint32_t queue_id, uint32_t inst);
+int kgd_gfx_v9_hqd_destroy(struct amdgpu_device *adev, void *mqd,
 				enum kfd_preempt_type reset_type,
 				unsigned int utimeout, uint32_t pipe_id,
-				uint32_t queue_id);
-int kgd_gfx_v9_address_watch_disable(struct kgd_dev *kgd);
-int kgd_gfx_v9_address_watch_execute(struct kgd_dev *kgd,
-					unsigned int watch_point_id,
-					uint32_t cntl_val,
-					uint32_t addr_hi,
-					uint32_t addr_lo);
-int kgd_gfx_v9_wave_control_execute(struct kgd_dev *kgd,
+				uint32_t queue_id, uint32_t inst);
+int kgd_gfx_v9_wave_control_execute(struct amdgpu_device *adev,
 					uint32_t gfx_index_val,
-					uint32_t sq_cmd);
-uint32_t kgd_gfx_v9_address_watch_get_offset(struct kgd_dev *kgd,
-					unsigned int watch_point_id,
-					unsigned int reg_offset);
-
-bool kgd_gfx_v9_get_atc_vmid_pasid_mapping_info(struct kgd_dev *kgd,
+					uint32_t sq_cmd, uint32_t inst);
+bool kgd_gfx_v9_get_atc_vmid_pasid_mapping_info(struct amdgpu_device *adev,
 					uint8_t vmid, uint16_t *p_pasid);
-
-void kgd_gfx_v9_set_vm_context_page_table_base(struct kgd_dev *kgd,
+void kgd_gfx_v9_set_vm_context_page_table_base(struct amdgpu_device *adev,
 			uint32_t vmid, uint64_t page_table_base);
-void kgd_gfx_v9_get_cu_occupancy(struct kgd_dev *kgd, int pasid,
-		int *pasid_wave_cnt, int *max_waves_per_cu);
-void kgd_gfx_v9_program_trap_handler_settings(struct kgd_dev *kgd,
-		uint32_t vmid, uint64_t tba_addr, uint64_t tma_addr);
+void kgd_gfx_v9_get_cu_occupancy(struct amdgpu_device *adev,
+				 struct kfd_cu_occupancy *cu_occupancy,
+				 int *max_waves_per_cu, uint32_t inst);
+void kgd_gfx_v9_program_trap_handler_settings(struct amdgpu_device *adev,
+		uint32_t vmid, uint64_t tba_addr, uint64_t tma_addr,
+		uint32_t inst);
+void kgd_gfx_v9_acquire_queue(struct amdgpu_device *adev, uint32_t pipe_id,
+				uint32_t queue_id, uint32_t inst);
+uint64_t kgd_gfx_v9_get_queue_mask(struct amdgpu_device *adev,
+				uint32_t pipe_id, uint32_t queue_id);
+void kgd_gfx_v9_release_queue(struct amdgpu_device *adev, uint32_t inst);
+void kgd_gfx_v9_set_wave_launch_stall(struct amdgpu_device *adev,
+					uint32_t vmid,
+					bool stall);
+uint32_t kgd_gfx_v9_enable_debug_trap(struct amdgpu_device *adev,
+				      bool restore_dbg_registers,
+				      uint32_t vmid);
+uint32_t kgd_gfx_v9_disable_debug_trap(struct amdgpu_device *adev,
+					bool keep_trap_enabled,
+					uint32_t vmid);
+int kgd_gfx_v9_validate_trap_override_request(struct amdgpu_device *adev,
+					     uint32_t trap_override,
+					     uint32_t *trap_mask_supported);
+uint32_t kgd_gfx_v9_set_wave_launch_mode(struct amdgpu_device *adev,
+					uint8_t wave_launch_mode,
+					uint32_t vmid);
+uint32_t kgd_gfx_v9_set_wave_launch_trap_override(struct amdgpu_device *adev,
+					     uint32_t vmid,
+					     uint32_t trap_override,
+					     uint32_t trap_mask_bits,
+					     uint32_t trap_mask_request,
+					     uint32_t *trap_mask_prev,
+					     uint32_t kfd_dbg_trap_cntl_prev);
+uint32_t kgd_gfx_v9_set_address_watch(struct amdgpu_device *adev,
+					uint64_t watch_address,
+					uint32_t watch_address_mask,
+					uint32_t watch_id,
+					uint32_t watch_mode,
+					uint32_t debug_vmid,
+					uint32_t inst);
+uint32_t kgd_gfx_v9_clear_address_watch(struct amdgpu_device *adev,
+					uint32_t watch_id);
+void kgd_gfx_v9_get_iq_wait_times(struct amdgpu_device *adev,
+				uint32_t *wait_times,
+				uint32_t inst);
+void kgd_gfx_v9_build_dequeue_wait_counts_packet_info(struct amdgpu_device *adev,
+					       uint32_t wait_times,
+					       uint32_t sch_wave,
+					       uint32_t que_sleep,
+					       uint32_t *reg_offset,
+					       uint32_t *reg_data);
+uint64_t kgd_gfx_v9_hqd_get_pq_addr(struct amdgpu_device *adev,
+				    uint32_t pipe_id,
+				    uint32_t queue_id,
+				    uint32_t inst);
+uint64_t kgd_gfx_v9_hqd_reset(struct amdgpu_device *adev,
+			      uint32_t pipe_id,
+			      uint32_t queue_id,
+			      uint32_t inst,
+			      unsigned int utimeout);
+uint32_t kgd_gfx_v9_hqd_sdma_get_doorbell(struct amdgpu_device *adev,
+					  int engine, int queue);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
index 6348559608ce..b1c24c8fa686 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
@@ -1,3 +1,4 @@
+// SPDX-License-Identifier: MIT
 /*
  * Copyright 2014-2018 Advanced Micro Devices, Inc.
  *
@@ -24,19 +25,32 @@
 #include <linux/pagemap.h>
 #include <linux/sched/mm.h>
 #include <linux/sched/task.h>
+#include <drm/ttm/ttm_tt.h>
+
+#include <drm/drm_exec.h>
 
 #include "amdgpu_object.h"
 #include "amdgpu_gem.h"
 #include "amdgpu_vm.h"
+#include "amdgpu_hmm.h"
 #include "amdgpu_amdkfd.h"
 #include "amdgpu_dma_buf.h"
 #include <uapi/linux/kfd_ioctl.h>
 #include "amdgpu_xgmi.h"
+#include "kfd_priv.h"
+#include "kfd_smi_events.h"
 
 /* Userptr restore delay, just long enough to allow consecutive VM
  * changes to accumulate
  */
 #define AMDGPU_USERPTR_RESTORE_DELAY_MS 1
+#define AMDGPU_RESERVE_MEM_LIMIT			(3UL << 29)
+
+/*
+ * Align VRAM availability to 2MB to avoid fragmentation caused by 4K allocations in the tail 2MB
+ * BO chunk
+ */
+#define VRAM_AVAILABLITY_ALIGN (1 << 21)
 
 /* Impose limit on how much memory KFD can use */
 static struct {
@@ -60,12 +74,6 @@ static const char * const domain_bit_to_string[] = {
 
 static void amdgpu_amdkfd_restore_userptr_worker(struct work_struct *work);
 
-
-static inline struct amdgpu_device *get_amdgpu_device(struct kgd_dev *kgd)
-{
-	return (struct amdgpu_device *)kgd;
-}
-
 static bool kfd_mem_is_attached(struct amdgpu_vm *avm,
 		struct kgd_mem *mem)
 {
@@ -78,6 +86,25 @@ static bool kfd_mem_is_attached(struct amdgpu_vm *avm,
 	return false;
 }
 
+/**
+ * reuse_dmamap() - Check whether adev can share the original
+ * userptr BO
+ *
+ * If both adev and bo_adev are in direct mapping or
+ * in the same iommu group, they can share the original BO.
+ *
+ * @adev: Device to which can or cannot share the original BO
+ * @bo_adev: Device to which allocated BO belongs to
+ *
+ * Return: returns true if adev can share original userptr BO,
+ * false otherwise.
+ */
+static bool reuse_dmamap(struct amdgpu_device *adev, struct amdgpu_device *bo_adev)
+{
+	return (adev->ram_is_direct_mapped && bo_adev->ram_is_direct_mapped) ||
+			(adev->dev->iommu_group == bo_adev->dev->iommu_group);
+}
+
 /* Set memory usage limits. Current, limits are
  *  System (TTM + userptr) memory - 15/16th System RAM
  *  TTM memory - 3/8th System RAM
@@ -87,13 +114,21 @@ void amdgpu_amdkfd_gpuvm_init_mem_limits(void)
 	struct sysinfo si;
 	uint64_t mem;
 
+	if (kfd_mem_limit.max_system_mem_limit)
+		return;
+
 	si_meminfo(&si);
-	mem = si.freeram - si.freehigh;
+	mem = si.totalram - si.totalhigh;
 	mem *= si.mem_unit;
 
 	spin_lock_init(&kfd_mem_limit.mem_limit_lock);
-	kfd_mem_limit.max_system_mem_limit = mem - (mem >> 4);
-	kfd_mem_limit.max_ttm_mem_limit = (mem >> 1) - (mem >> 3);
+	kfd_mem_limit.max_system_mem_limit = mem - (mem >> 6);
+	if (kfd_mem_limit.max_system_mem_limit < 2 * AMDGPU_RESERVE_MEM_LIMIT)
+		kfd_mem_limit.max_system_mem_limit >>= 1;
+	else
+		kfd_mem_limit.max_system_mem_limit -= AMDGPU_RESERVE_MEM_LIMIT;
+
+	kfd_mem_limit.max_ttm_mem_limit = ttm_tt_pages_limit() << PAGE_SHIFT;
 	pr_debug("Kernel memory limit %lluM, TTM limit %lluM\n",
 		(kfd_mem_limit.max_system_mem_limit >> 20),
 		(kfd_mem_limit.max_ttm_mem_limit >> 20));
@@ -114,115 +149,215 @@ void amdgpu_amdkfd_reserve_system_mem(uint64_t size)
  * compromise that should work in most cases without reserving too
  * much memory for page tables unnecessarily (factor 16K, >> 14).
  */
-#define ESTIMATE_PT_SIZE(mem_size) ((mem_size) >> 14)
-
-static size_t amdgpu_amdkfd_acc_size(uint64_t size)
-{
-	size >>= PAGE_SHIFT;
-	size *= sizeof(dma_addr_t) + sizeof(void *);
 
-	return __roundup_pow_of_two(sizeof(struct amdgpu_bo)) +
-		__roundup_pow_of_two(sizeof(struct ttm_tt)) +
-		PAGE_ALIGN(size);
-}
+#define ESTIMATE_PT_SIZE(mem_size) max(((mem_size) >> 14), AMDGPU_VM_RESERVED_VRAM)
 
-static int amdgpu_amdkfd_reserve_mem_limit(struct amdgpu_device *adev,
-		uint64_t size, u32 domain, bool sg)
+/**
+ * amdgpu_amdkfd_reserve_mem_limit() - Decrease available memory by size
+ * of buffer.
+ *
+ * @adev: Device to which allocated BO belongs to
+ * @size: Size of buffer, in bytes, encapsulated by B0. This should be
+ * equivalent to amdgpu_bo_size(BO)
+ * @alloc_flag: Flag used in allocating a BO as noted above
+ * @xcp_id: xcp_id is used to get xcp from xcp manager, one xcp is
+ * managed as one compute node in driver for app
+ *
+ * Return:
+ *	returns -ENOMEM in case of error, ZERO otherwise
+ */
+int amdgpu_amdkfd_reserve_mem_limit(struct amdgpu_device *adev,
+		uint64_t size, u32 alloc_flag, int8_t xcp_id)
 {
 	uint64_t reserved_for_pt =
 		ESTIMATE_PT_SIZE(amdgpu_amdkfd_total_mem_size);
-	size_t acc_size, system_mem_needed, ttm_mem_needed, vram_needed;
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+	uint64_t reserved_for_ras = (con ? con->reserved_pages_in_bytes : 0);
+	size_t system_mem_needed, ttm_mem_needed, vram_needed;
 	int ret = 0;
+	uint64_t vram_size = 0;
 
-	acc_size = amdgpu_amdkfd_acc_size(size);
-
+	system_mem_needed = 0;
+	ttm_mem_needed = 0;
 	vram_needed = 0;
-	if (domain == AMDGPU_GEM_DOMAIN_GTT) {
-		/* TTM GTT memory */
-		system_mem_needed = acc_size + size;
-		ttm_mem_needed = acc_size + size;
-	} else if (domain == AMDGPU_GEM_DOMAIN_CPU && !sg) {
-		/* Userptr */
-		system_mem_needed = acc_size + size;
-		ttm_mem_needed = acc_size;
-	} else {
-		/* VRAM and SG */
-		system_mem_needed = acc_size;
-		ttm_mem_needed = acc_size;
-		if (domain == AMDGPU_GEM_DOMAIN_VRAM)
-			vram_needed = size;
+	if (alloc_flag & KFD_IOC_ALLOC_MEM_FLAGS_GTT) {
+		system_mem_needed = size;
+		ttm_mem_needed = size;
+	} else if (alloc_flag & KFD_IOC_ALLOC_MEM_FLAGS_VRAM) {
+		/*
+		 * Conservatively round up the allocation requirement to 2 MB
+		 * to avoid fragmentation caused by 4K allocations in the tail
+		 * 2M BO chunk.
+		 */
+		vram_needed = size;
+		/*
+		 * For GFX 9.4.3, get the VRAM size from XCP structs
+		 */
+		if (WARN_ONCE(xcp_id < 0, "invalid XCP ID %d", xcp_id))
+			return -EINVAL;
+
+		vram_size = KFD_XCP_MEMORY_SIZE(adev, xcp_id);
+		if (adev->apu_prefer_gtt) {
+			system_mem_needed = size;
+			ttm_mem_needed = size;
+		}
+	} else if (alloc_flag & KFD_IOC_ALLOC_MEM_FLAGS_USERPTR) {
+		system_mem_needed = size;
+	} else if (!(alloc_flag &
+				(KFD_IOC_ALLOC_MEM_FLAGS_DOORBELL |
+				 KFD_IOC_ALLOC_MEM_FLAGS_MMIO_REMAP))) {
+		pr_err("%s: Invalid BO type %#x\n", __func__, alloc_flag);
+		return -ENOMEM;
 	}
 
 	spin_lock(&kfd_mem_limit.mem_limit_lock);
 
 	if (kfd_mem_limit.system_mem_used + system_mem_needed >
-	    kfd_mem_limit.max_system_mem_limit)
+	    kfd_mem_limit.max_system_mem_limit) {
 		pr_debug("Set no_system_mem_limit=1 if using shared memory\n");
+		if (!no_system_mem_limit) {
+			ret = -ENOMEM;
+			goto release;
+		}
+	}
 
-	if ((kfd_mem_limit.system_mem_used + system_mem_needed >
-	     kfd_mem_limit.max_system_mem_limit && !no_system_mem_limit) ||
-	    (kfd_mem_limit.ttm_mem_used + ttm_mem_needed >
-	     kfd_mem_limit.max_ttm_mem_limit) ||
-	    (adev->kfd.vram_used + vram_needed >
-	     adev->gmc.real_vram_size - reserved_for_pt)) {
+	if (kfd_mem_limit.ttm_mem_used + ttm_mem_needed >
+		kfd_mem_limit.max_ttm_mem_limit) {
 		ret = -ENOMEM;
-	} else {
-		kfd_mem_limit.system_mem_used += system_mem_needed;
-		kfd_mem_limit.ttm_mem_used += ttm_mem_needed;
-		adev->kfd.vram_used += vram_needed;
+		goto release;
+	}
+
+	/*if is_app_apu is false and apu_prefer_gtt is true, it is an APU with
+	 * carve out < gtt. In that case, VRAM allocation will go to gtt domain, skip
+	 * VRAM check since ttm_mem_limit check already cover this allocation
+	 */
+
+	if (adev && xcp_id >= 0 && (!adev->apu_prefer_gtt || adev->gmc.is_app_apu)) {
+		uint64_t vram_available =
+			vram_size - reserved_for_pt - reserved_for_ras -
+			atomic64_read(&adev->vram_pin_size);
+		if (adev->kfd.vram_used[xcp_id] + vram_needed > vram_available) {
+			ret = -ENOMEM;
+			goto release;
+		}
 	}
 
+	/* Update memory accounting by decreasing available system
+	 * memory, TTM memory and GPU memory as computed above
+	 */
+	WARN_ONCE(vram_needed && !adev,
+		  "adev reference can't be null when vram is used");
+	if (adev && xcp_id >= 0) {
+		adev->kfd.vram_used[xcp_id] += vram_needed;
+		adev->kfd.vram_used_aligned[xcp_id] +=
+				adev->apu_prefer_gtt ?
+				vram_needed :
+				ALIGN(vram_needed, VRAM_AVAILABLITY_ALIGN);
+	}
+	kfd_mem_limit.system_mem_used += system_mem_needed;
+	kfd_mem_limit.ttm_mem_used += ttm_mem_needed;
+
+release:
 	spin_unlock(&kfd_mem_limit.mem_limit_lock);
 	return ret;
 }
 
-static void unreserve_mem_limit(struct amdgpu_device *adev,
-		uint64_t size, u32 domain, bool sg)
+void amdgpu_amdkfd_unreserve_mem_limit(struct amdgpu_device *adev,
+		uint64_t size, u32 alloc_flag, int8_t xcp_id)
 {
-	size_t acc_size;
-
-	acc_size = amdgpu_amdkfd_acc_size(size);
-
 	spin_lock(&kfd_mem_limit.mem_limit_lock);
-	if (domain == AMDGPU_GEM_DOMAIN_GTT) {
-		kfd_mem_limit.system_mem_used -= (acc_size + size);
-		kfd_mem_limit.ttm_mem_used -= (acc_size + size);
-	} else if (domain == AMDGPU_GEM_DOMAIN_CPU && !sg) {
-		kfd_mem_limit.system_mem_used -= (acc_size + size);
-		kfd_mem_limit.ttm_mem_used -= acc_size;
-	} else {
-		kfd_mem_limit.system_mem_used -= acc_size;
-		kfd_mem_limit.ttm_mem_used -= acc_size;
-		if (domain == AMDGPU_GEM_DOMAIN_VRAM) {
-			adev->kfd.vram_used -= size;
-			WARN_ONCE(adev->kfd.vram_used < 0,
-				  "kfd VRAM memory accounting unbalanced");
+
+	if (alloc_flag & KFD_IOC_ALLOC_MEM_FLAGS_GTT) {
+		kfd_mem_limit.system_mem_used -= size;
+		kfd_mem_limit.ttm_mem_used -= size;
+	} else if (alloc_flag & KFD_IOC_ALLOC_MEM_FLAGS_VRAM) {
+		WARN_ONCE(!adev,
+			  "adev reference can't be null when alloc mem flags vram is set");
+		if (WARN_ONCE(xcp_id < 0, "invalid XCP ID %d", xcp_id))
+			goto release;
+
+		if (adev) {
+			adev->kfd.vram_used[xcp_id] -= size;
+			if (adev->apu_prefer_gtt) {
+				adev->kfd.vram_used_aligned[xcp_id] -= size;
+				kfd_mem_limit.system_mem_used -= size;
+				kfd_mem_limit.ttm_mem_used -= size;
+			} else {
+				adev->kfd.vram_used_aligned[xcp_id] -=
+					ALIGN(size, VRAM_AVAILABLITY_ALIGN);
+			}
 		}
-	}
-	WARN_ONCE(kfd_mem_limit.system_mem_used < 0,
-		  "kfd system memory accounting unbalanced");
+	} else if (alloc_flag & KFD_IOC_ALLOC_MEM_FLAGS_USERPTR) {
+		kfd_mem_limit.system_mem_used -= size;
+	} else if (!(alloc_flag &
+				(KFD_IOC_ALLOC_MEM_FLAGS_DOORBELL |
+				 KFD_IOC_ALLOC_MEM_FLAGS_MMIO_REMAP))) {
+		pr_err("%s: Invalid BO type %#x\n", __func__, alloc_flag);
+		goto release;
+	}
+	WARN_ONCE(adev && xcp_id >= 0 && adev->kfd.vram_used[xcp_id] < 0,
+		  "KFD VRAM memory accounting unbalanced for xcp: %d", xcp_id);
 	WARN_ONCE(kfd_mem_limit.ttm_mem_used < 0,
-		  "kfd TTM memory accounting unbalanced");
+		  "KFD TTM memory accounting unbalanced");
+	WARN_ONCE(kfd_mem_limit.system_mem_used < 0,
+		  "KFD system memory accounting unbalanced");
 
+release:
 	spin_unlock(&kfd_mem_limit.mem_limit_lock);
 }
 
 void amdgpu_amdkfd_release_notify(struct amdgpu_bo *bo)
 {
 	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
-	u32 domain = bo->preferred_domains;
-	bool sg = (bo->preferred_domains == AMDGPU_GEM_DOMAIN_CPU);
+	u32 alloc_flags = bo->kfd_bo->alloc_flags;
+	u64 size = amdgpu_bo_size(bo);
 
-	if (bo->flags & AMDGPU_AMDKFD_CREATE_USERPTR_BO) {
-		domain = AMDGPU_GEM_DOMAIN_CPU;
-		sg = false;
-	}
-
-	unreserve_mem_limit(adev, amdgpu_bo_size(bo), domain, sg);
+	amdgpu_amdkfd_unreserve_mem_limit(adev, size, alloc_flags,
+					  bo->xcp_id);
 
 	kfree(bo->kfd_bo);
 }
 
+/**
+ * create_dmamap_sg_bo() - Creates a amdgpu_bo object to reflect information
+ * about USERPTR or DOOREBELL or MMIO BO.
+ *
+ * @adev: Device for which dmamap BO is being created
+ * @mem: BO of peer device that is being DMA mapped. Provides parameters
+ *	 in building the dmamap BO
+ * @bo_out: Output parameter updated with handle of dmamap BO
+ */
+static int
+create_dmamap_sg_bo(struct amdgpu_device *adev,
+		 struct kgd_mem *mem, struct amdgpu_bo **bo_out)
+{
+	struct drm_gem_object *gem_obj;
+	int ret;
+	uint64_t flags = 0;
+
+	ret = amdgpu_bo_reserve(mem->bo, false);
+	if (ret)
+		return ret;
+
+	if (mem->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_USERPTR)
+		flags |= mem->bo->flags & (AMDGPU_GEM_CREATE_COHERENT |
+					AMDGPU_GEM_CREATE_UNCACHED);
+
+	ret = amdgpu_gem_object_create(adev, mem->bo->tbo.base.size, 1,
+			AMDGPU_GEM_DOMAIN_CPU, AMDGPU_GEM_CREATE_PREEMPTIBLE | flags,
+			ttm_bo_type_sg, mem->bo->tbo.base.resv, &gem_obj, 0);
+
+	amdgpu_bo_unreserve(mem->bo);
+
+	if (ret) {
+		pr_err("Error in creating DMA mappable SG BO on domain: %d\n", ret);
+		return -EINVAL;
+	}
+
+	*bo_out = gem_to_amdgpu_bo(gem_obj);
+	(*bo_out)->parent = amdgpu_bo_ref(mem->bo);
+	return ret;
+}
 
 /* amdgpu_amdkfd_remove_eviction_fence - Removes eviction fence from BO's
  *  reservation object.
@@ -236,90 +371,47 @@ void amdgpu_amdkfd_release_notify(struct amdgpu_bo *bo)
 static int amdgpu_amdkfd_remove_eviction_fence(struct amdgpu_bo *bo,
 					struct amdgpu_amdkfd_fence *ef)
 {
-	struct dma_resv *resv = bo->tbo.base.resv;
-	struct dma_resv_list *old, *new;
-	unsigned int i, j, k;
+	struct dma_fence *replacement;
 
 	if (!ef)
 		return -EINVAL;
 
-	old = dma_resv_shared_list(resv);
-	if (!old)
-		return 0;
-
-	new = kmalloc(struct_size(new, shared, old->shared_max), GFP_KERNEL);
-	if (!new)
-		return -ENOMEM;
-
-	/* Go through all the shared fences in the resevation object and sort
-	 * the interesting ones to the end of the list.
+	/* TODO: Instead of block before we should use the fence of the page
+	 * table update and TLB flush here directly.
 	 */
-	for (i = 0, j = old->shared_count, k = 0; i < old->shared_count; ++i) {
-		struct dma_fence *f;
-
-		f = rcu_dereference_protected(old->shared[i],
-					      dma_resv_held(resv));
-
-		if (f->context == ef->base.context)
-			RCU_INIT_POINTER(new->shared[--j], f);
-		else
-			RCU_INIT_POINTER(new->shared[k++], f);
-	}
-	new->shared_max = old->shared_max;
-	new->shared_count = k;
-
-	/* Install the new fence list, seqcount provides the barriers */
-	write_seqcount_begin(&resv->seq);
-	RCU_INIT_POINTER(resv->fence, new);
-	write_seqcount_end(&resv->seq);
-
-	/* Drop the references to the removed fences or move them to ef_list */
-	for (i = j; i < old->shared_count; ++i) {
-		struct dma_fence *f;
-
-		f = rcu_dereference_protected(new->shared[i],
-					      dma_resv_held(resv));
-		dma_fence_put(f);
-	}
-	kfree_rcu(old, rcu);
-
+	replacement = dma_fence_get_stub();
+	dma_resv_replace_fences(bo->tbo.base.resv, ef->base.context,
+				replacement, DMA_RESV_USAGE_BOOKKEEP);
+	dma_fence_put(replacement);
 	return 0;
 }
 
-int amdgpu_amdkfd_remove_fence_on_pt_pd_bos(struct amdgpu_bo *bo)
+/**
+ * amdgpu_amdkfd_remove_all_eviction_fences - Remove all eviction fences
+ * @bo: the BO where to remove the evictions fences from.
+ *
+ * This functions should only be used on release when all references to the BO
+ * are already dropped. We remove the eviction fence from the private copy of
+ * the dma_resv object here since that is what is used during release to
+ * determine of the BO is idle or not.
+ */
+void amdgpu_amdkfd_remove_all_eviction_fences(struct amdgpu_bo *bo)
 {
-	struct amdgpu_bo *root = bo;
-	struct amdgpu_vm_bo_base *vm_bo;
-	struct amdgpu_vm *vm;
-	struct amdkfd_process_info *info;
-	struct amdgpu_amdkfd_fence *ef;
-	int ret;
-
-	/* we can always get vm_bo from root PD bo.*/
-	while (root->parent)
-		root = root->parent;
+	struct dma_resv *resv = &bo->tbo.base._resv;
+	struct dma_fence *fence, *stub;
+	struct dma_resv_iter cursor;
 
-	vm_bo = root->vm_bo;
-	if (!vm_bo)
-		return 0;
+	dma_resv_assert_held(resv);
 
-	vm = vm_bo->vm;
-	if (!vm)
-		return 0;
-
-	info = vm->process_info;
-	if (!info || !info->eviction_fence)
-		return 0;
-
-	ef = container_of(dma_fence_get(&info->eviction_fence->base),
-			struct amdgpu_amdkfd_fence, base);
-
-	BUG_ON(!dma_resv_trylock(bo->tbo.base.resv));
-	ret = amdgpu_amdkfd_remove_eviction_fence(bo, ef);
-	dma_resv_unlock(bo->tbo.base.resv);
+	stub = dma_fence_get_stub();
+	dma_resv_for_each_fence(&cursor, resv, DMA_RESV_USAGE_BOOKKEEP, fence) {
+		if (!to_amdgpu_amdkfd_fence(fence))
+			continue;
 
-	dma_fence_put(&ef->base);
-	return ret;
+		dma_resv_replace_fences(resv, fence->context, stub,
+					DMA_RESV_USAGE_BOOKKEEP);
+	}
+	dma_fence_put(stub);
 }
 
 static int amdgpu_amdkfd_bo_validate(struct amdgpu_bo *bo, uint32_t domain,
@@ -332,6 +424,10 @@ static int amdgpu_amdkfd_bo_validate(struct amdgpu_bo *bo, uint32_t domain,
 		 "Called with userptr BO"))
 		return -EINVAL;
 
+	/* bo has been pinned, not need validate it */
+	if (bo->tbo.pin_count)
+		return 0;
+
 	amdgpu_bo_placement_from_domain(bo, domain);
 
 	ret = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
@@ -344,6 +440,32 @@ validate_fail:
 	return ret;
 }
 
+int amdgpu_amdkfd_bo_validate_and_fence(struct amdgpu_bo *bo,
+					uint32_t domain,
+					struct dma_fence *fence)
+{
+	int ret = amdgpu_bo_reserve(bo, false);
+
+	if (ret)
+		return ret;
+
+	ret = amdgpu_amdkfd_bo_validate(bo, domain, true);
+	if (ret)
+		goto unreserve_out;
+
+	ret = dma_resv_reserve_fences(bo->tbo.base.resv, 1);
+	if (ret)
+		goto unreserve_out;
+
+	dma_resv_add_fence(bo->tbo.base.resv, fence,
+			   DMA_RESV_USAGE_BOOKKEEP);
+
+unreserve_out:
+	amdgpu_bo_unreserve(bo);
+
+	return ret;
+}
+
 static int amdgpu_amdkfd_validate_vm_bo(void *_unused, struct amdgpu_bo *bo)
 {
 	return amdgpu_amdkfd_bo_validate(bo, bo->allowed_domains, false);
@@ -356,34 +478,22 @@ static int amdgpu_amdkfd_validate_vm_bo(void *_unused, struct amdgpu_bo *bo)
  * again. Page directories are only updated after updating page
  * tables.
  */
-static int vm_validate_pt_pd_bos(struct amdgpu_vm *vm)
+static int vm_validate_pt_pd_bos(struct amdgpu_vm *vm,
+				 struct ww_acquire_ctx *ticket)
 {
 	struct amdgpu_bo *pd = vm->root.bo;
 	struct amdgpu_device *adev = amdgpu_ttm_adev(pd->tbo.bdev);
 	int ret;
 
-	ret = amdgpu_vm_validate_pt_bos(adev, vm, amdgpu_amdkfd_validate_vm_bo, NULL);
+	ret = amdgpu_vm_validate(adev, vm, ticket,
+				 amdgpu_amdkfd_validate_vm_bo, NULL);
 	if (ret) {
 		pr_err("failed to validate PT BOs\n");
 		return ret;
 	}
 
-	ret = amdgpu_amdkfd_validate_vm_bo(NULL, pd);
-	if (ret) {
-		pr_err("failed to validate PD\n");
-		return ret;
-	}
-
 	vm->pd_phys_addr = amdgpu_gmc_pd_addr(vm->root.bo);
 
-	if (vm->use_cpu_for_update) {
-		ret = amdgpu_bo_kmap(pd, NULL);
-		if (ret) {
-			pr_err("failed to kmap PD, ret=%d\n", ret);
-			return ret;
-		}
-	}
-
 	return 0;
 }
 
@@ -397,71 +507,53 @@ static int vm_update_pds(struct amdgpu_vm *vm, struct amdgpu_sync *sync)
 	if (ret)
 		return ret;
 
-	return amdgpu_sync_fence(sync, vm->last_update);
+	return amdgpu_sync_fence(sync, vm->last_update, GFP_KERNEL);
 }
 
-static uint64_t get_pte_flags(struct amdgpu_device *adev, struct kgd_mem *mem)
+static uint64_t get_pte_flags(struct amdgpu_device *adev, struct amdgpu_vm *vm,
+			      struct kgd_mem *mem)
 {
-	struct amdgpu_device *bo_adev = amdgpu_ttm_adev(mem->bo->tbo.bdev);
-	bool coherent = mem->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_COHERENT;
-	bool uncached = mem->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_UNCACHED;
-	uint32_t mapping_flags;
-	uint64_t pte_flags;
-	bool snoop = false;
+	uint32_t mapping_flags = AMDGPU_VM_PAGE_READABLE |
+				 AMDGPU_VM_MTYPE_DEFAULT;
 
-	mapping_flags = AMDGPU_VM_PAGE_READABLE;
 	if (mem->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_WRITABLE)
 		mapping_flags |= AMDGPU_VM_PAGE_WRITEABLE;
 	if (mem->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_EXECUTABLE)
 		mapping_flags |= AMDGPU_VM_PAGE_EXECUTABLE;
 
-	switch (adev->asic_type) {
-	case CHIP_ARCTURUS:
-		if (mem->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_VRAM) {
-			if (bo_adev == adev)
-				mapping_flags |= coherent ?
-					AMDGPU_VM_MTYPE_CC : AMDGPU_VM_MTYPE_RW;
-			else
-				mapping_flags |= coherent ?
-					AMDGPU_VM_MTYPE_UC : AMDGPU_VM_MTYPE_NC;
-		} else {
-			mapping_flags |= coherent ?
-				AMDGPU_VM_MTYPE_UC : AMDGPU_VM_MTYPE_NC;
-		}
-		break;
-	case CHIP_ALDEBARAN:
-		if (coherent && uncached) {
-			if (adev->gmc.xgmi.connected_to_cpu ||
-				!(mem->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_VRAM))
-				snoop = true;
-			mapping_flags |= AMDGPU_VM_MTYPE_UC;
-		} else if (mem->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_VRAM) {
-			if (bo_adev == adev) {
-				mapping_flags |= coherent ?
-					AMDGPU_VM_MTYPE_CC : AMDGPU_VM_MTYPE_RW;
-				if (adev->gmc.xgmi.connected_to_cpu)
-					snoop = true;
-			} else {
-				mapping_flags |= coherent ?
-					AMDGPU_VM_MTYPE_UC : AMDGPU_VM_MTYPE_NC;
-				if (amdgpu_xgmi_same_hive(adev, bo_adev))
-					snoop = true;
-			}
-		} else {
-			snoop = true;
-			mapping_flags |= coherent ?
-				AMDGPU_VM_MTYPE_UC : AMDGPU_VM_MTYPE_NC;
-		}
-		break;
-	default:
-		mapping_flags |= coherent ?
-			AMDGPU_VM_MTYPE_UC : AMDGPU_VM_MTYPE_NC;
-	}
+	return mapping_flags;
+}
 
-	pte_flags = amdgpu_gem_va_map_flags(adev, mapping_flags);
-	pte_flags |= snoop ? AMDGPU_PTE_SNOOPED : 0;
+/**
+ * create_sg_table() - Create an sg_table for a contiguous DMA addr range
+ * @addr: The starting address to point to
+ * @size: Size of memory area in bytes being pointed to
+ *
+ * Allocates an instance of sg_table and initializes it to point to memory
+ * area specified by input parameters. The address used to build is assumed
+ * to be DMA mapped, if needed.
+ *
+ * DOORBELL or MMIO BOs use only one scatterlist node in their sg_table
+ * because they are physically contiguous.
+ *
+ * Return: Initialized instance of SG Table or NULL
+ */
+static struct sg_table *create_sg_table(uint64_t addr, uint32_t size)
+{
+	struct sg_table *sg = kmalloc(sizeof(*sg), GFP_KERNEL);
 
-	return pte_flags;
+	if (!sg)
+		return NULL;
+	if (sg_alloc_table(sg, 1, GFP_KERNEL)) {
+		kfree(sg);
+		return NULL;
+	}
+	sg_dma_address(sg->sgl) = addr;
+	sg->sgl->length = size;
+#ifdef CONFIG_NEED_SG_DMA_LENGTH
+	sg->sgl->dma_length = size;
+#endif
+	return sg;
 }
 
 static int
@@ -478,13 +570,13 @@ kfd_mem_dmamap_userptr(struct kgd_mem *mem,
 	struct ttm_tt *ttm = bo->tbo.ttm;
 	int ret;
 
+	if (WARN_ON(ttm->num_pages != src_ttm->num_pages))
+		return -EINVAL;
+
 	ttm->sg = kmalloc(sizeof(*ttm->sg), GFP_KERNEL);
 	if (unlikely(!ttm->sg))
 		return -ENOMEM;
 
-	if (WARN_ON(ttm->num_pages != src_ttm->num_pages))
-		return -EINVAL;
-
 	/* Same sequence as in amdgpu_ttm_tt_pin_userptr */
 	ret = sg_alloc_table_from_pages(ttm->sg, src_ttm->pages,
 					ttm->num_pages, 0,
@@ -497,9 +589,6 @@ kfd_mem_dmamap_userptr(struct kgd_mem *mem,
 	if (unlikely(ret))
 		goto release_sg;
 
-	drm_prime_sg_to_dma_addr_array(ttm->sg, ttm->dma_address,
-				       ttm->num_pages);
-
 	amdgpu_bo_placement_from_domain(bo, AMDGPU_GEM_DOMAIN_GTT);
 	ret = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
 	if (ret)
@@ -528,6 +617,87 @@ kfd_mem_dmamap_dmabuf(struct kfd_mem_attachment *attachment)
 	return ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
 }
 
+/**
+ * kfd_mem_dmamap_sg_bo() - Create DMA mapped sg_table to access DOORBELL or MMIO BO
+ * @mem: SG BO of the DOORBELL or MMIO resource on the owning device
+ * @attachment: Virtual address attachment of the BO on accessing device
+ *
+ * An access request from the device that owns DOORBELL does not require DMA mapping.
+ * This is because the request doesn't go through PCIe root complex i.e. it instead
+ * loops back. The need to DMA map arises only when accessing peer device's DOORBELL
+ *
+ * In contrast, all access requests for MMIO need to be DMA mapped without regard to
+ * device ownership. This is because access requests for MMIO go through PCIe root
+ * complex.
+ *
+ * This is accomplished in two steps:
+ *   - Obtain DMA mapped address of DOORBELL or MMIO memory that could be used
+ *         in updating requesting device's page table
+ *   - Signal TTM to mark memory pointed to by requesting device's BO as GPU
+ *         accessible. This allows an update of requesting device's page table
+ *         with entries associated with DOOREBELL or MMIO memory
+ *
+ * This method is invoked in the following contexts:
+ *   - Mapping of DOORBELL or MMIO BO of same or peer device
+ *   - Validating an evicted DOOREBELL or MMIO BO on device seeking access
+ *
+ * Return: ZERO if successful, NON-ZERO otherwise
+ */
+static int
+kfd_mem_dmamap_sg_bo(struct kgd_mem *mem,
+		     struct kfd_mem_attachment *attachment)
+{
+	struct ttm_operation_ctx ctx = {.interruptible = true};
+	struct amdgpu_bo *bo = attachment->bo_va->base.bo;
+	struct amdgpu_device *adev = attachment->adev;
+	struct ttm_tt *ttm = bo->tbo.ttm;
+	enum dma_data_direction dir;
+	dma_addr_t dma_addr;
+	bool mmio;
+	int ret;
+
+	/* Expect SG Table of dmapmap BO to be NULL */
+	mmio = (mem->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_MMIO_REMAP);
+	if (unlikely(ttm->sg)) {
+		pr_err("SG Table of %d BO for peer device is UNEXPECTEDLY NON-NULL", mmio);
+		return -EINVAL;
+	}
+
+	dir = mem->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_WRITABLE ?
+			DMA_BIDIRECTIONAL : DMA_TO_DEVICE;
+	dma_addr = mem->bo->tbo.sg->sgl->dma_address;
+	pr_debug("%d BO size: %d\n", mmio, mem->bo->tbo.sg->sgl->length);
+	pr_debug("%d BO address before DMA mapping: %llx\n", mmio, dma_addr);
+	dma_addr = dma_map_resource(adev->dev, dma_addr,
+			mem->bo->tbo.sg->sgl->length, dir, DMA_ATTR_SKIP_CPU_SYNC);
+	ret = dma_mapping_error(adev->dev, dma_addr);
+	if (unlikely(ret))
+		return ret;
+	pr_debug("%d BO address after DMA mapping: %llx\n", mmio, dma_addr);
+
+	ttm->sg = create_sg_table(dma_addr, mem->bo->tbo.sg->sgl->length);
+	if (unlikely(!ttm->sg)) {
+		ret = -ENOMEM;
+		goto unmap_sg;
+	}
+
+	amdgpu_bo_placement_from_domain(bo, AMDGPU_GEM_DOMAIN_GTT);
+	ret = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
+	if (unlikely(ret))
+		goto free_sg;
+
+	return ret;
+
+free_sg:
+	sg_free_table(ttm->sg);
+	kfree(ttm->sg);
+	ttm->sg = NULL;
+unmap_sg:
+	dma_unmap_resource(adev->dev, dma_addr, mem->bo->tbo.sg->sgl->length,
+			   dir, DMA_ATTR_SKIP_CPU_SYNC);
+	return ret;
+}
+
 static int
 kfd_mem_dmamap_attachment(struct kgd_mem *mem,
 			  struct kfd_mem_attachment *attachment)
@@ -539,6 +709,8 @@ kfd_mem_dmamap_attachment(struct kgd_mem *mem,
 		return kfd_mem_dmamap_userptr(mem, attachment);
 	case KFD_MEM_ATT_DMABUF:
 		return kfd_mem_dmamap_dmabuf(attachment);
+	case KFD_MEM_ATT_SG:
+		return kfd_mem_dmamap_sg_bo(mem, attachment);
 	default:
 		WARN_ON_ONCE(1);
 	}
@@ -561,7 +733,7 @@ kfd_mem_dmaunmap_userptr(struct kgd_mem *mem,
 		return;
 
 	amdgpu_bo_placement_from_domain(bo, AMDGPU_GEM_DOMAIN_CPU);
-	ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
+	(void)ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
 
 	dma_unmap_sgtable(adev->dev, ttm->sg, direction, 0);
 	sg_free_table(ttm->sg);
@@ -572,11 +744,54 @@ kfd_mem_dmaunmap_userptr(struct kgd_mem *mem,
 static void
 kfd_mem_dmaunmap_dmabuf(struct kfd_mem_attachment *attachment)
 {
+	/* This is a no-op. We don't want to trigger eviction fences when
+	 * unmapping DMABufs. Therefore the invalidation (moving to system
+	 * domain) is done in kfd_mem_dmamap_dmabuf.
+	 */
+}
+
+/**
+ * kfd_mem_dmaunmap_sg_bo() - Free DMA mapped sg_table of DOORBELL or MMIO BO
+ * @mem: SG BO of the DOORBELL or MMIO resource on the owning device
+ * @attachment: Virtual address attachment of the BO on accessing device
+ *
+ * The method performs following steps:
+ *   - Signal TTM to mark memory pointed to by BO as GPU inaccessible
+ *   - Free SG Table that is used to encapsulate DMA mapped memory of
+ *          peer device's DOORBELL or MMIO memory
+ *
+ * This method is invoked in the following contexts:
+ *     UNMapping of DOORBELL or MMIO BO on a device having access to its memory
+ *     Eviction of DOOREBELL or MMIO BO on device having access to its memory
+ *
+ * Return: void
+ */
+static void
+kfd_mem_dmaunmap_sg_bo(struct kgd_mem *mem,
+		       struct kfd_mem_attachment *attachment)
+{
 	struct ttm_operation_ctx ctx = {.interruptible = true};
 	struct amdgpu_bo *bo = attachment->bo_va->base.bo;
+	struct amdgpu_device *adev = attachment->adev;
+	struct ttm_tt *ttm = bo->tbo.ttm;
+	enum dma_data_direction dir;
+
+	if (unlikely(!ttm->sg)) {
+		pr_debug("SG Table of BO is NULL");
+		return;
+	}
 
 	amdgpu_bo_placement_from_domain(bo, AMDGPU_GEM_DOMAIN_CPU);
-	ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
+	(void)ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
+
+	dir = mem->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_WRITABLE ?
+				DMA_BIDIRECTIONAL : DMA_TO_DEVICE;
+	dma_unmap_resource(adev->dev, ttm->sg->sgl->dma_address,
+			ttm->sg->sgl->length, dir, DMA_ATTR_SKIP_CPU_SYNC);
+	sg_free_table(ttm->sg);
+	kfree(ttm->sg);
+	ttm->sg = NULL;
+	bo->tbo.sg = NULL;
 }
 
 static void
@@ -592,34 +807,29 @@ kfd_mem_dmaunmap_attachment(struct kgd_mem *mem,
 	case KFD_MEM_ATT_DMABUF:
 		kfd_mem_dmaunmap_dmabuf(attachment);
 		break;
+	case KFD_MEM_ATT_SG:
+		kfd_mem_dmaunmap_sg_bo(mem, attachment);
+		break;
 	default:
 		WARN_ON_ONCE(1);
 	}
 }
 
-static int
-kfd_mem_attach_userptr(struct amdgpu_device *adev, struct kgd_mem *mem,
-		       struct amdgpu_bo **bo)
+static int kfd_mem_export_dmabuf(struct kgd_mem *mem)
 {
-	unsigned long bo_size = mem->bo->tbo.base.size;
-	struct drm_gem_object *gobj;
-	int ret;
-
-	ret = amdgpu_bo_reserve(mem->bo, false);
-	if (ret)
-		return ret;
-
-	ret = amdgpu_gem_object_create(adev, bo_size, 1,
-				       AMDGPU_GEM_DOMAIN_CPU,
-				       AMDGPU_GEM_CREATE_PREEMPTIBLE,
-				       ttm_bo_type_sg, mem->bo->tbo.base.resv,
-				       &gobj);
-	amdgpu_bo_unreserve(mem->bo);
-	if (ret)
-		return ret;
+	if (!mem->dmabuf) {
+		struct amdgpu_device *bo_adev;
+		struct dma_buf *dmabuf;
 
-	*bo = gem_to_amdgpu_bo(gobj);
-	(*bo)->parent = amdgpu_bo_ref(mem->bo);
+		bo_adev = amdgpu_ttm_adev(mem->bo->tbo.bdev);
+		dmabuf = drm_gem_prime_handle_to_dmabuf(&bo_adev->ddev, bo_adev->kfd.client.file,
+					       mem->gem_handle,
+			mem->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_WRITABLE ?
+					       DRM_RDWR : 0);
+		if (IS_ERR(dmabuf))
+			return PTR_ERR(dmabuf);
+		mem->dmabuf = dmabuf;
+	}
 
 	return 0;
 }
@@ -631,16 +841,9 @@ kfd_mem_attach_dmabuf(struct amdgpu_device *adev, struct kgd_mem *mem,
 	struct drm_gem_object *gobj;
 	int ret;
 
-	if (!mem->dmabuf) {
-		mem->dmabuf = amdgpu_gem_prime_export(&mem->bo->tbo.base,
-			mem->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_WRITABLE ?
-				DRM_RDWR : 0);
-		if (IS_ERR(mem->dmabuf)) {
-			ret = PTR_ERR(mem->dmabuf);
-			mem->dmabuf = NULL;
-			return ret;
-		}
-	}
+	ret = kfd_mem_export_dmabuf(mem);
+	if (ret)
+		return ret;
 
 	gobj = amdgpu_gem_prime_import(adev_to_drm(adev), mem->dmabuf);
 	if (IS_ERR(gobj))
@@ -648,7 +851,6 @@ kfd_mem_attach_dmabuf(struct amdgpu_device *adev, struct kgd_mem *mem,
 
 	*bo = gem_to_amdgpu_bo(gobj);
 	(*bo)->flags |= AMDGPU_GEM_CREATE_PREEMPTIBLE;
-	(*bo)->parent = amdgpu_bo_ref(mem->bo);
 
 	return 0;
 }
@@ -674,6 +876,8 @@ static int kfd_mem_attach(struct amdgpu_device *adev, struct kgd_mem *mem,
 	uint64_t va = mem->va;
 	struct kfd_mem_attachment *attachment[2] = {NULL, NULL};
 	struct amdgpu_bo *bo[2] = {NULL, NULL};
+	struct amdgpu_bo_va *bo_va;
+	bool same_hive = false;
 	int i, ret;
 
 	if (!va) {
@@ -681,6 +885,24 @@ static int kfd_mem_attach(struct amdgpu_device *adev, struct kgd_mem *mem,
 		return -EINVAL;
 	}
 
+	/* Determine access to VRAM, MMIO and DOORBELL BOs of peer devices
+	 *
+	 * The access path of MMIO and DOORBELL BOs of is always over PCIe.
+	 * In contrast the access path of VRAM BOs depens upon the type of
+	 * link that connects the peer device. Access over PCIe is allowed
+	 * if peer device has large BAR. In contrast, access over xGMI is
+	 * allowed for both small and large BAR configurations of peer device
+	 */
+	if ((adev != bo_adev && !adev->apu_prefer_gtt) &&
+	    ((mem->domain == AMDGPU_GEM_DOMAIN_VRAM) ||
+	     (mem->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_DOORBELL) ||
+	     (mem->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_MMIO_REMAP))) {
+		if (mem->domain == AMDGPU_GEM_DOMAIN_VRAM)
+			same_hive = amdgpu_xgmi_same_hive(adev, bo_adev);
+		if (!same_hive && !amdgpu_device_is_peer_accessible(bo_adev, adev))
+			return -EINVAL;
+	}
+
 	for (i = 0; i <= is_aql; i++) {
 		attachment[i] = kzalloc(sizeof(*attachment[i]), GFP_KERNEL);
 		if (unlikely(!attachment[i])) {
@@ -691,10 +913,13 @@ static int kfd_mem_attach(struct amdgpu_device *adev, struct kgd_mem *mem,
 		pr_debug("\t add VA 0x%llx - 0x%llx to vm %p\n", va,
 			 va + bo_size, vm);
 
-		if (adev == bo_adev || (mem->domain == AMDGPU_GEM_DOMAIN_VRAM &&
-					amdgpu_xgmi_same_hive(adev, bo_adev))) {
-			/* Mappings on the local GPU and VRAM mappings in the
-			 * local hive share the original BO
+		if ((adev == bo_adev && !(mem->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_MMIO_REMAP)) ||
+		    (amdgpu_ttm_tt_get_usermm(mem->bo->tbo.ttm) && reuse_dmamap(adev, bo_adev)) ||
+		    (mem->domain == AMDGPU_GEM_DOMAIN_GTT && reuse_dmamap(adev, bo_adev)) ||
+		    same_hive) {
+			/* Mappings on the local GPU, or VRAM mappings in the
+			 * local hive, or userptr, or GTT mapping can reuse dma map
+			 * address space share the original BO
 			 */
 			attachment[i]->type = KFD_MEM_ATT_SHARED;
 			bo[i] = mem->bo;
@@ -707,26 +932,30 @@ static int kfd_mem_attach(struct amdgpu_device *adev, struct kgd_mem *mem,
 		} else if (amdgpu_ttm_tt_get_usermm(mem->bo->tbo.ttm)) {
 			/* Create an SG BO to DMA-map userptrs on other GPUs */
 			attachment[i]->type = KFD_MEM_ATT_USERPTR;
-			ret = kfd_mem_attach_userptr(adev, mem, &bo[i]);
+			ret = create_dmamap_sg_bo(adev, mem, &bo[i]);
 			if (ret)
 				goto unwind;
-		} else if (mem->domain == AMDGPU_GEM_DOMAIN_GTT &&
-			   mem->bo->tbo.type != ttm_bo_type_sg) {
-			/* GTT BOs use DMA-mapping ability of dynamic-attach
-			 * DMA bufs. TODO: The same should work for VRAM on
-			 * large-BAR GPUs.
-			 */
+		/* Handle DOORBELL BOs of peer devices and MMIO BOs of local and peer devices */
+		} else if (mem->bo->tbo.type == ttm_bo_type_sg) {
+			WARN_ONCE(!(mem->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_DOORBELL ||
+				    mem->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_MMIO_REMAP),
+				  "Handing invalid SG BO in ATTACH request");
+			attachment[i]->type = KFD_MEM_ATT_SG;
+			ret = create_dmamap_sg_bo(adev, mem, &bo[i]);
+			if (ret)
+				goto unwind;
+		/* Enable acces to GTT and VRAM BOs of peer devices */
+		} else if (mem->domain == AMDGPU_GEM_DOMAIN_GTT ||
+			   mem->domain == AMDGPU_GEM_DOMAIN_VRAM) {
 			attachment[i]->type = KFD_MEM_ATT_DMABUF;
 			ret = kfd_mem_attach_dmabuf(adev, mem, &bo[i]);
 			if (ret)
 				goto unwind;
+			pr_debug("Employ DMABUF mechanism to enable peer GPU access\n");
 		} else {
-			/* FIXME: Need to DMA-map other BO types:
-			 * large-BAR VRAM, doorbells, MMIO remap
-			 */
-			attachment[i]->type = KFD_MEM_ATT_SHARED;
-			bo[i] = mem->bo;
-			drm_gem_object_get(&bo[i]->tbo.base);
+			WARN_ONCE(true, "Handling invalid ATTACH request");
+			ret = -EINVAL;
+			goto unwind;
 		}
 
 		/* Add BO to VM internal data structures */
@@ -735,7 +964,12 @@ static int kfd_mem_attach(struct amdgpu_device *adev, struct kgd_mem *mem,
 			pr_debug("Unable to reserve BO during memory attach");
 			goto unwind;
 		}
-		attachment[i]->bo_va = amdgpu_vm_bo_add(adev, vm, bo[i]);
+		bo_va = amdgpu_vm_bo_find(vm, bo[i]);
+		if (!bo_va)
+			bo_va = amdgpu_vm_bo_add(adev, vm, bo[i]);
+		else
+			++bo_va->ref_count;
+		attachment[i]->bo_va = bo_va;
 		amdgpu_bo_unreserve(bo[i]);
 		if (unlikely(!attachment[i]->bo_va)) {
 			ret = -ENOMEM;
@@ -744,7 +978,7 @@ static int kfd_mem_attach(struct amdgpu_device *adev, struct kgd_mem *mem,
 			goto unwind;
 		}
 		attachment[i]->va = va;
-		attachment[i]->pte_flags = get_pte_flags(adev, mem);
+		attachment[i]->pte_flags = get_pte_flags(adev, vm, mem);
 		attachment[i]->adev = adev;
 		list_add(&attachment[i]->list, &mem->attachments);
 
@@ -758,8 +992,9 @@ unwind:
 		if (!attachment[i])
 			continue;
 		if (attachment[i]->bo_va) {
-			amdgpu_bo_reserve(bo[i], true);
-			amdgpu_vm_bo_rmv(adev, attachment[i]->bo_va);
+			(void)amdgpu_bo_reserve(bo[i], true);
+			if (--attachment[i]->bo_va->ref_count == 0)
+				amdgpu_vm_bo_del(adev, attachment[i]->bo_va);
 			amdgpu_bo_unreserve(bo[i]);
 			list_del(&attachment[i]->list);
 		}
@@ -776,7 +1011,8 @@ static void kfd_mem_detach(struct kfd_mem_attachment *attachment)
 
 	pr_debug("\t remove VA 0x%llx in entry %p\n",
 			attachment->va, attachment);
-	amdgpu_vm_bo_rmv(attachment->adev, attachment->bo_va);
+	if (--attachment->bo_va->ref_count == 0)
+		amdgpu_vm_bo_del(attachment->adev, attachment->bo_va);
 	drm_gem_object_put(&bo->tbo.base);
 	list_del(&attachment->list);
 	kfree(attachment);
@@ -786,28 +1022,20 @@ static void add_kgd_mem_to_kfd_bo_list(struct kgd_mem *mem,
 				struct amdkfd_process_info *process_info,
 				bool userptr)
 {
-	struct ttm_validate_buffer *entry = &mem->validate_list;
-	struct amdgpu_bo *bo = mem->bo;
-
-	INIT_LIST_HEAD(&entry->head);
-	entry->num_shared = 1;
-	entry->bo = &bo->tbo;
 	mutex_lock(&process_info->lock);
 	if (userptr)
-		list_add_tail(&entry->head, &process_info->userptr_valid_list);
+		list_add_tail(&mem->validate_list,
+			      &process_info->userptr_valid_list);
 	else
-		list_add_tail(&entry->head, &process_info->kfd_bo_list);
+		list_add_tail(&mem->validate_list, &process_info->kfd_bo_list);
 	mutex_unlock(&process_info->lock);
 }
 
 static void remove_kgd_mem_from_kfd_bo_list(struct kgd_mem *mem,
 		struct amdkfd_process_info *process_info)
 {
-	struct ttm_validate_buffer *bo_list_entry;
-
-	bo_list_entry = &mem->validate_list;
 	mutex_lock(&process_info->lock);
-	list_del(&bo_list_entry->head);
+	list_del(&mem->validate_list);
 	mutex_unlock(&process_info->lock);
 }
 
@@ -823,11 +1051,13 @@ static void remove_kgd_mem_from_kfd_bo_list(struct kgd_mem *mem,
  *
  * Returns 0 for success, negative errno for errors.
  */
-static int init_user_pages(struct kgd_mem *mem, uint64_t user_addr)
+static int init_user_pages(struct kgd_mem *mem, uint64_t user_addr,
+			   bool criu_resume)
 {
 	struct amdkfd_process_info *process_info = mem->process_info;
 	struct amdgpu_bo *bo = mem->bo;
 	struct ttm_operation_ctx ctx = { true, false };
+	struct amdgpu_hmm_range *range;
 	int ret = 0;
 
 	mutex_lock(&process_info->lock);
@@ -838,16 +1068,40 @@ static int init_user_pages(struct kgd_mem *mem, uint64_t user_addr)
 		goto out;
 	}
 
-	ret = amdgpu_mn_register(bo, user_addr);
+	ret = amdgpu_hmm_register(bo, user_addr);
 	if (ret) {
 		pr_err("%s: Failed to register MMU notifier: %d\n",
 		       __func__, ret);
 		goto out;
 	}
 
-	ret = amdgpu_ttm_tt_get_user_pages(bo, bo->tbo.ttm->pages);
+	if (criu_resume) {
+		/*
+		 * During a CRIU restore operation, the userptr buffer objects
+		 * will be validated in the restore_userptr_work worker at a
+		 * later stage when it is scheduled by another ioctl called by
+		 * CRIU master process for the target pid for restore.
+		 */
+		mutex_lock(&process_info->notifier_lock);
+		mem->invalid++;
+		mutex_unlock(&process_info->notifier_lock);
+		mutex_unlock(&process_info->lock);
+		return 0;
+	}
+
+	range = amdgpu_hmm_range_alloc(NULL);
+	if (unlikely(!range)) {
+		ret = -ENOMEM;
+		goto unregister_out;
+	}
+
+	ret = amdgpu_ttm_tt_get_user_pages(bo, range);
 	if (ret) {
-		pr_err("%s: Failed to get user pages: %d\n", __func__, ret);
+		amdgpu_hmm_range_free(range);
+		if (ret == -EAGAIN)
+			pr_debug("Failed to get user pages, try again\n");
+		else
+			pr_err("%s: Failed to get user pages: %d\n", __func__, ret);
 		goto unregister_out;
 	}
 
@@ -856,6 +1110,9 @@ static int init_user_pages(struct kgd_mem *mem, uint64_t user_addr)
 		pr_err("%s: Failed to reserve BO\n", __func__);
 		goto release_out;
 	}
+
+	amdgpu_ttm_tt_set_user_pages(bo->tbo.ttm, range);
+
 	amdgpu_bo_placement_from_domain(bo, mem->domain);
 	ret = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
 	if (ret)
@@ -863,10 +1120,10 @@ static int init_user_pages(struct kgd_mem *mem, uint64_t user_addr)
 	amdgpu_bo_unreserve(bo);
 
 release_out:
-	amdgpu_ttm_tt_get_user_pages_done(bo->tbo.ttm);
+	amdgpu_hmm_range_free(range);
 unregister_out:
 	if (ret)
-		amdgpu_mn_unregister(bo);
+		amdgpu_hmm_unregister(bo);
 out:
 	mutex_unlock(&process_info->lock);
 	return ret;
@@ -878,13 +1135,12 @@ out:
  * object can track VM updates.
  */
 struct bo_vm_reservation_context {
-	struct amdgpu_bo_list_entry kfd_bo; /* BO list entry for the KFD BO */
-	unsigned int n_vms;		    /* Number of VMs reserved	    */
-	struct amdgpu_bo_list_entry *vm_pd; /* Array of VM BO list entries  */
-	struct ww_acquire_ctx ticket;	    /* Reservation ticket	    */
-	struct list_head list, duplicates;  /* BO lists			    */
-	struct amdgpu_sync *sync;	    /* Pointer to sync object	    */
-	bool reserved;			    /* Whether BOs are reserved	    */
+	/* DRM execution context for the reservation */
+	struct drm_exec exec;
+	/* Number of VMs reserved */
+	unsigned int n_vms;
+	/* Pointer to sync object */
+	struct amdgpu_sync *sync;
 };
 
 enum bo_vm_match {
@@ -908,35 +1164,26 @@ static int reserve_bo_and_vm(struct kgd_mem *mem,
 
 	WARN_ON(!vm);
 
-	ctx->reserved = false;
 	ctx->n_vms = 1;
 	ctx->sync = &mem->sync;
-
-	INIT_LIST_HEAD(&ctx->list);
-	INIT_LIST_HEAD(&ctx->duplicates);
-
-	ctx->vm_pd = kcalloc(ctx->n_vms, sizeof(*ctx->vm_pd), GFP_KERNEL);
-	if (!ctx->vm_pd)
-		return -ENOMEM;
-
-	ctx->kfd_bo.priority = 0;
-	ctx->kfd_bo.tv.bo = &bo->tbo;
-	ctx->kfd_bo.tv.num_shared = 1;
-	list_add(&ctx->kfd_bo.tv.head, &ctx->list);
-
-	amdgpu_vm_get_pd_bo(vm, &ctx->list, &ctx->vm_pd[0]);
-
-	ret = ttm_eu_reserve_buffers(&ctx->ticket, &ctx->list,
-				     false, &ctx->duplicates);
-	if (ret) {
-		pr_err("Failed to reserve buffers in ttm.\n");
-		kfree(ctx->vm_pd);
-		ctx->vm_pd = NULL;
-		return ret;
+	drm_exec_init(&ctx->exec, DRM_EXEC_INTERRUPTIBLE_WAIT, 0);
+	drm_exec_until_all_locked(&ctx->exec) {
+		ret = amdgpu_vm_lock_pd(vm, &ctx->exec, 2);
+		drm_exec_retry_on_contention(&ctx->exec);
+		if (unlikely(ret))
+			goto error;
+
+		ret = drm_exec_prepare_obj(&ctx->exec, &bo->tbo.base, 1);
+		drm_exec_retry_on_contention(&ctx->exec);
+		if (unlikely(ret))
+			goto error;
 	}
-
-	ctx->reserved = true;
 	return 0;
+
+error:
+	pr_err("Failed to reserve buffers in ttm.\n");
+	drm_exec_fini(&ctx->exec);
+	return ret;
 }
 
 /**
@@ -953,63 +1200,40 @@ static int reserve_bo_and_cond_vms(struct kgd_mem *mem,
 				struct amdgpu_vm *vm, enum bo_vm_match map_type,
 				struct bo_vm_reservation_context *ctx)
 {
-	struct amdgpu_bo *bo = mem->bo;
 	struct kfd_mem_attachment *entry;
-	unsigned int i;
+	struct amdgpu_bo *bo = mem->bo;
 	int ret;
 
-	ctx->reserved = false;
-	ctx->n_vms = 0;
-	ctx->vm_pd = NULL;
 	ctx->sync = &mem->sync;
+	drm_exec_init(&ctx->exec, DRM_EXEC_INTERRUPTIBLE_WAIT |
+		      DRM_EXEC_IGNORE_DUPLICATES, 0);
+	drm_exec_until_all_locked(&ctx->exec) {
+		ctx->n_vms = 0;
+		list_for_each_entry(entry, &mem->attachments, list) {
+			if ((vm && vm != entry->bo_va->base.vm) ||
+				(entry->is_mapped != map_type
+				&& map_type != BO_VM_ALL))
+				continue;
 
-	INIT_LIST_HEAD(&ctx->list);
-	INIT_LIST_HEAD(&ctx->duplicates);
-
-	list_for_each_entry(entry, &mem->attachments, list) {
-		if ((vm && vm != entry->bo_va->base.vm) ||
-			(entry->is_mapped != map_type
-			&& map_type != BO_VM_ALL))
-			continue;
-
-		ctx->n_vms++;
-	}
-
-	if (ctx->n_vms != 0) {
-		ctx->vm_pd = kcalloc(ctx->n_vms, sizeof(*ctx->vm_pd),
-				     GFP_KERNEL);
-		if (!ctx->vm_pd)
-			return -ENOMEM;
-	}
-
-	ctx->kfd_bo.priority = 0;
-	ctx->kfd_bo.tv.bo = &bo->tbo;
-	ctx->kfd_bo.tv.num_shared = 1;
-	list_add(&ctx->kfd_bo.tv.head, &ctx->list);
-
-	i = 0;
-	list_for_each_entry(entry, &mem->attachments, list) {
-		if ((vm && vm != entry->bo_va->base.vm) ||
-			(entry->is_mapped != map_type
-			&& map_type != BO_VM_ALL))
-			continue;
-
-		amdgpu_vm_get_pd_bo(entry->bo_va->base.vm, &ctx->list,
-				&ctx->vm_pd[i]);
-		i++;
-	}
+			ret = amdgpu_vm_lock_pd(entry->bo_va->base.vm,
+						&ctx->exec, 2);
+			drm_exec_retry_on_contention(&ctx->exec);
+			if (unlikely(ret))
+				goto error;
+			++ctx->n_vms;
+		}
 
-	ret = ttm_eu_reserve_buffers(&ctx->ticket, &ctx->list,
-				     false, &ctx->duplicates);
-	if (ret) {
-		pr_err("Failed to reserve buffers in ttm.\n");
-		kfree(ctx->vm_pd);
-		ctx->vm_pd = NULL;
-		return ret;
+		ret = drm_exec_prepare_obj(&ctx->exec, &bo->tbo.base, 1);
+		drm_exec_retry_on_contention(&ctx->exec);
+		if (unlikely(ret))
+			goto error;
 	}
-
-	ctx->reserved = true;
 	return 0;
+
+error:
+	pr_err("Failed to reserve buffers in ttm.\n");
+	drm_exec_fini(&ctx->exec);
+	return ret;
 }
 
 /**
@@ -1030,19 +1254,12 @@ static int unreserve_bo_and_vms(struct bo_vm_reservation_context *ctx,
 	if (wait)
 		ret = amdgpu_sync_wait(ctx->sync, intr);
 
-	if (ctx->reserved)
-		ttm_eu_backoff_reservation(&ctx->ticket, &ctx->list);
-	kfree(ctx->vm_pd);
-
+	drm_exec_fini(&ctx->exec);
 	ctx->sync = NULL;
-
-	ctx->reserved = false;
-	ctx->vm_pd = NULL;
-
 	return ret;
 }
 
-static void unmap_bo_from_gpuvm(struct kgd_mem *mem,
+static int unmap_bo_from_gpuvm(struct kgd_mem *mem,
 				struct kfd_mem_attachment *entry,
 				struct amdgpu_sync *sync)
 {
@@ -1050,19 +1267,27 @@ static void unmap_bo_from_gpuvm(struct kgd_mem *mem,
 	struct amdgpu_device *adev = entry->adev;
 	struct amdgpu_vm *vm = bo_va->base.vm;
 
-	amdgpu_vm_bo_unmap(adev, bo_va, entry->va);
+	if (bo_va->queue_refcount) {
+		pr_debug("bo_va->queue_refcount %d\n", bo_va->queue_refcount);
+		return -EBUSY;
+	}
 
-	amdgpu_vm_clear_freed(adev, vm, &bo_va->last_pt_update);
+	(void)amdgpu_vm_bo_unmap(adev, bo_va, entry->va);
 
-	amdgpu_sync_fence(sync, bo_va->last_pt_update);
+	/* VM entity stopped if process killed, don't clear freed pt bo */
+	if (!amdgpu_vm_ready(vm))
+		return 0;
 
-	kfd_mem_dmaunmap_attachment(mem, entry);
+	(void)amdgpu_vm_clear_freed(adev, vm, &bo_va->last_pt_update);
+
+	(void)amdgpu_sync_fence(sync, bo_va->last_pt_update, GFP_KERNEL);
+
+	return 0;
 }
 
 static int update_gpuvm_pte(struct kgd_mem *mem,
 			    struct kfd_mem_attachment *entry,
-			    struct amdgpu_sync *sync,
-			    bool *table_freed)
+			    struct amdgpu_sync *sync)
 {
 	struct amdgpu_bo_va *bo_va = entry->bo_va;
 	struct amdgpu_device *adev = entry->adev;
@@ -1073,20 +1298,19 @@ static int update_gpuvm_pte(struct kgd_mem *mem,
 		return ret;
 
 	/* Update the page tables  */
-	ret = amdgpu_vm_bo_update(adev, bo_va, false, table_freed);
+	ret = amdgpu_vm_bo_update(adev, bo_va, false);
 	if (ret) {
 		pr_err("amdgpu_vm_bo_update failed\n");
 		return ret;
 	}
 
-	return amdgpu_sync_fence(sync, bo_va->last_pt_update);
+	return amdgpu_sync_fence(sync, bo_va->last_pt_update, GFP_KERNEL);
 }
 
 static int map_bo_to_gpuvm(struct kgd_mem *mem,
 			   struct kfd_mem_attachment *entry,
 			   struct amdgpu_sync *sync,
-			   bool no_update_pte,
-			   bool *table_freed)
+			   bool no_update_pte)
 {
 	int ret;
 
@@ -1103,7 +1327,7 @@ static int map_bo_to_gpuvm(struct kgd_mem *mem,
 	if (no_update_pte)
 		return 0;
 
-	ret = update_gpuvm_pte(mem, entry, sync, table_freed);
+	ret = update_gpuvm_pte(mem, entry, sync);
 	if (ret) {
 		pr_err("update_gpuvm_pte() failed\n");
 		goto update_gpuvm_pte_failed;
@@ -1113,35 +1337,19 @@ static int map_bo_to_gpuvm(struct kgd_mem *mem,
 
 update_gpuvm_pte_failed:
 	unmap_bo_from_gpuvm(mem, entry, sync);
+	kfd_mem_dmaunmap_attachment(mem, entry);
 	return ret;
 }
 
-static struct sg_table *create_doorbell_sg(uint64_t addr, uint32_t size)
-{
-	struct sg_table *sg = kmalloc(sizeof(*sg), GFP_KERNEL);
-
-	if (!sg)
-		return NULL;
-	if (sg_alloc_table(sg, 1, GFP_KERNEL)) {
-		kfree(sg);
-		return NULL;
-	}
-	sg->sgl->dma_address = addr;
-	sg->sgl->length = size;
-#ifdef CONFIG_NEED_SG_DMA_LENGTH
-	sg->sgl->dma_length = size;
-#endif
-	return sg;
-}
-
-static int process_validate_vms(struct amdkfd_process_info *process_info)
+static int process_validate_vms(struct amdkfd_process_info *process_info,
+				struct ww_acquire_ctx *ticket)
 {
 	struct amdgpu_vm *peer_vm;
 	int ret;
 
 	list_for_each_entry(peer_vm, &process_info->vm_list_head,
 			    vm_list_node) {
-		ret = vm_validate_pt_pd_bos(peer_vm);
+		ret = vm_validate_pt_pd_bos(peer_vm, ticket);
 		if (ret)
 			return ret;
 	}
@@ -1197,6 +1405,7 @@ static int init_kfd_vm(struct amdgpu_vm *vm, void **process_info,
 			return -ENOMEM;
 
 		mutex_init(&info->lock);
+		mutex_init(&info->notifier_lock);
 		INIT_LIST_HEAD(&info->vm_list_head);
 		INIT_LIST_HEAD(&info->kfd_bo_list);
 		INIT_LIST_HEAD(&info->userptr_valid_list);
@@ -1213,12 +1422,10 @@ static int init_kfd_vm(struct amdgpu_vm *vm, void **process_info,
 		}
 
 		info->pid = get_task_pid(current->group_leader, PIDTYPE_PID);
-		atomic_set(&info->evicted_bos, 0);
 		INIT_DELAYED_WORK(&info->restore_userptr_work,
 				  amdgpu_amdkfd_restore_userptr_worker);
 
 		*process_info = info;
-		*ef = dma_fence_get(&info->eviction_fence->base);
 	}
 
 	vm->process_info = *process_info;
@@ -1227,7 +1434,7 @@ static int init_kfd_vm(struct amdgpu_vm *vm, void **process_info,
 	ret = amdgpu_bo_reserve(vm->root.bo, true);
 	if (ret)
 		goto reserve_pd_fail;
-	ret = vm_validate_pt_pd_bos(vm);
+	ret = vm_validate_pt_pd_bos(vm, NULL);
 	if (ret) {
 		pr_err("validate_pt_pd_bos() failed\n");
 		goto validate_pd_fail;
@@ -1236,11 +1443,12 @@ static int init_kfd_vm(struct amdgpu_vm *vm, void **process_info,
 				  AMDGPU_FENCE_OWNER_KFD, false);
 	if (ret)
 		goto wait_pd_fail;
-	ret = dma_resv_reserve_shared(vm->root.bo->tbo.base.resv, 1);
+	ret = dma_resv_reserve_fences(vm->root.bo->tbo.base.resv, 1);
 	if (ret)
 		goto reserve_shared_fail;
-	amdgpu_bo_fence(vm->root.bo,
-			&vm->process_info->eviction_fence->base, true);
+	dma_resv_add_fence(vm->root.bo->tbo.base.resv,
+			   &vm->process_info->eviction_fence->base,
+			   DMA_RESV_USAGE_BOOKKEEP);
 	amdgpu_bo_unreserve(vm->root.bo);
 
 	/* Update process info */
@@ -1248,6 +1456,8 @@ static int init_kfd_vm(struct amdgpu_vm *vm, void **process_info,
 	list_add_tail(&vm->vm_list_node,
 			&(vm->process_info->vm_list_head));
 	vm->process_info->n_vms++;
+	if (ef)
+		*ef = dma_fence_get(&vm->process_info->eviction_fence->base);
 	mutex_unlock(&vm->process_info->lock);
 
 	return 0;
@@ -1259,54 +1469,99 @@ validate_pd_fail:
 reserve_pd_fail:
 	vm->process_info = NULL;
 	if (info) {
-		/* Two fence references: one in info and one in *ef */
 		dma_fence_put(&info->eviction_fence->base);
-		dma_fence_put(*ef);
-		*ef = NULL;
 		*process_info = NULL;
 		put_pid(info->pid);
 create_evict_fence_fail:
 		mutex_destroy(&info->lock);
+		mutex_destroy(&info->notifier_lock);
 		kfree(info);
 	}
 	return ret;
 }
 
-int amdgpu_amdkfd_gpuvm_acquire_process_vm(struct kgd_dev *kgd,
-					   struct file *filp, u32 pasid,
+/**
+ * amdgpu_amdkfd_gpuvm_pin_bo() - Pins a BO using following criteria
+ * @bo: Handle of buffer object being pinned
+ * @domain: Domain into which BO should be pinned
+ *
+ *   - USERPTR BOs are UNPINNABLE and will return error
+ *   - All other BO types (GTT, VRAM, MMIO and DOORBELL) will have their
+ *     PIN count incremented. It is valid to PIN a BO multiple times
+ *
+ * Return: ZERO if successful in pinning, Non-Zero in case of error.
+ */
+static int amdgpu_amdkfd_gpuvm_pin_bo(struct amdgpu_bo *bo, u32 domain)
+{
+	int ret = 0;
+
+	ret = amdgpu_bo_reserve(bo, false);
+	if (unlikely(ret))
+		return ret;
+
+	if (bo->flags & AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS) {
+		/*
+		 * If bo is not contiguous on VRAM, move to system memory first to ensure
+		 * we can get contiguous VRAM space after evicting other BOs.
+		 */
+		if (!(bo->tbo.resource->placement & TTM_PL_FLAG_CONTIGUOUS)) {
+			struct ttm_operation_ctx ctx = { true, false };
+
+			amdgpu_bo_placement_from_domain(bo, AMDGPU_GEM_DOMAIN_GTT);
+			ret = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
+			if (unlikely(ret)) {
+				pr_debug("validate bo 0x%p to GTT failed %d\n", &bo->tbo, ret);
+				goto out;
+			}
+		}
+	}
+
+	ret = amdgpu_bo_pin(bo, domain);
+	if (ret)
+		pr_err("Error in Pinning BO to domain: %d\n", domain);
+
+	amdgpu_bo_sync_wait(bo, AMDGPU_FENCE_OWNER_KFD, false);
+out:
+	amdgpu_bo_unreserve(bo);
+	return ret;
+}
+
+/**
+ * amdgpu_amdkfd_gpuvm_unpin_bo() - Unpins BO using following criteria
+ * @bo: Handle of buffer object being unpinned
+ *
+ *   - Is a illegal request for USERPTR BOs and is ignored
+ *   - All other BO types (GTT, VRAM, MMIO and DOORBELL) will have their
+ *     PIN count decremented. Calls to UNPIN must balance calls to PIN
+ */
+static void amdgpu_amdkfd_gpuvm_unpin_bo(struct amdgpu_bo *bo)
+{
+	int ret = 0;
+
+	ret = amdgpu_bo_reserve(bo, false);
+	if (unlikely(ret))
+		return;
+
+	amdgpu_bo_unpin(bo);
+	amdgpu_bo_unreserve(bo);
+}
+
+int amdgpu_amdkfd_gpuvm_acquire_process_vm(struct amdgpu_device *adev,
+					   struct amdgpu_vm *avm,
 					   void **process_info,
 					   struct dma_fence **ef)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-	struct amdgpu_fpriv *drv_priv;
-	struct amdgpu_vm *avm;
 	int ret;
 
-	ret = amdgpu_file_to_fpriv(filp, &drv_priv);
-	if (ret)
-		return ret;
-	avm = &drv_priv->vm;
-
 	/* Already a compute VM? */
 	if (avm->process_info)
 		return -EINVAL;
 
-	/* Free the original amdgpu allocated pasid,
-	 * will be replaced with kfd allocated pasid.
-	 */
-	if (avm->pasid) {
-		amdgpu_pasid_free(avm->pasid);
-		amdgpu_vm_set_pasid(adev, avm, 0);
-	}
-
 	/* Convert VM into a compute VM */
 	ret = amdgpu_vm_make_compute(adev, avm);
 	if (ret)
 		return ret;
 
-	ret = amdgpu_vm_set_pasid(adev, avm, pasid);
-	if (ret)
-		return ret;
 	/* Initialize KFD part of the VM and process info */
 	ret = init_kfd_vm(avm, process_info, ef);
 	if (ret)
@@ -1321,16 +1576,10 @@ void amdgpu_amdkfd_gpuvm_destroy_cb(struct amdgpu_device *adev,
 				    struct amdgpu_vm *vm)
 {
 	struct amdkfd_process_info *process_info = vm->process_info;
-	struct amdgpu_bo *pd = vm->root.bo;
 
 	if (!process_info)
 		return;
 
-	/* Release eviction fence from PD */
-	amdgpu_bo_reserve(pd, false);
-	amdgpu_bo_fence(pd, NULL, false);
-	amdgpu_bo_unreserve(pd);
-
 	/* Update process info */
 	mutex_lock(&process_info->lock);
 	process_info->n_vms--;
@@ -1349,31 +1598,11 @@ void amdgpu_amdkfd_gpuvm_destroy_cb(struct amdgpu_device *adev,
 		cancel_delayed_work_sync(&process_info->restore_userptr_work);
 		put_pid(process_info->pid);
 		mutex_destroy(&process_info->lock);
+		mutex_destroy(&process_info->notifier_lock);
 		kfree(process_info);
 	}
 }
 
-void amdgpu_amdkfd_gpuvm_release_process_vm(struct kgd_dev *kgd, void *drm_priv)
-{
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-	struct amdgpu_vm *avm;
-
-	if (WARN_ON(!kgd || !drm_priv))
-		return;
-
-	avm = drm_priv_to_vm(drm_priv);
-
-	pr_debug("Releasing process vm %p\n", avm);
-
-	/* The original pasid of amdgpu vm has already been
-	 * released during making a amdgpu vm to a compute vm
-	 * The current pasid is managed by kfd and will be
-	 * released on kfd process destroy. Set amdgpu pasid
-	 * to 0 to avoid duplicate release.
-	 */
-	amdgpu_vm_release_compute(adev, avm);
-}
-
 uint64_t amdgpu_amdkfd_gpuvm_get_process_page_dir(void *drm_priv)
 {
 	struct amdgpu_vm *avm = drm_priv_to_vm(drm_priv);
@@ -1385,19 +1614,98 @@ uint64_t amdgpu_amdkfd_gpuvm_get_process_page_dir(void *drm_priv)
 	return avm->pd_phys_addr;
 }
 
+void amdgpu_amdkfd_block_mmu_notifications(void *p)
+{
+	struct amdkfd_process_info *pinfo = (struct amdkfd_process_info *)p;
+
+	mutex_lock(&pinfo->lock);
+	WRITE_ONCE(pinfo->block_mmu_notifications, true);
+	mutex_unlock(&pinfo->lock);
+}
+
+int amdgpu_amdkfd_criu_resume(void *p)
+{
+	int ret = 0;
+	struct amdkfd_process_info *pinfo = (struct amdkfd_process_info *)p;
+
+	mutex_lock(&pinfo->lock);
+	pr_debug("scheduling work\n");
+	mutex_lock(&pinfo->notifier_lock);
+	pinfo->evicted_bos++;
+	mutex_unlock(&pinfo->notifier_lock);
+	if (!READ_ONCE(pinfo->block_mmu_notifications)) {
+		ret = -EINVAL;
+		goto out_unlock;
+	}
+	WRITE_ONCE(pinfo->block_mmu_notifications, false);
+	queue_delayed_work(system_freezable_wq,
+			   &pinfo->restore_userptr_work, 0);
+
+out_unlock:
+	mutex_unlock(&pinfo->lock);
+	return ret;
+}
+
+size_t amdgpu_amdkfd_get_available_memory(struct amdgpu_device *adev,
+					  uint8_t xcp_id)
+{
+	uint64_t reserved_for_pt =
+		ESTIMATE_PT_SIZE(amdgpu_amdkfd_total_mem_size);
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+	uint64_t reserved_for_ras = (con ? con->reserved_pages_in_bytes : 0);
+	ssize_t available;
+	uint64_t vram_available, system_mem_available, ttm_mem_available;
+
+	spin_lock(&kfd_mem_limit.mem_limit_lock);
+	if (adev->apu_prefer_gtt && !adev->gmc.is_app_apu)
+		vram_available = KFD_XCP_MEMORY_SIZE(adev, xcp_id)
+			- adev->kfd.vram_used_aligned[xcp_id];
+	else
+		vram_available = KFD_XCP_MEMORY_SIZE(adev, xcp_id)
+			- adev->kfd.vram_used_aligned[xcp_id]
+			- atomic64_read(&adev->vram_pin_size)
+			- reserved_for_pt
+			- reserved_for_ras;
+
+	if (adev->apu_prefer_gtt) {
+		system_mem_available = no_system_mem_limit ?
+					kfd_mem_limit.max_system_mem_limit :
+					kfd_mem_limit.max_system_mem_limit -
+					kfd_mem_limit.system_mem_used;
+
+		ttm_mem_available = kfd_mem_limit.max_ttm_mem_limit -
+				kfd_mem_limit.ttm_mem_used;
+
+		available = min3(system_mem_available, ttm_mem_available,
+				 vram_available);
+		available = ALIGN_DOWN(available, PAGE_SIZE);
+	} else {
+		available = ALIGN_DOWN(vram_available, VRAM_AVAILABLITY_ALIGN);
+	}
+
+	spin_unlock(&kfd_mem_limit.mem_limit_lock);
+
+	if (available < 0)
+		available = 0;
+
+	return available;
+}
+
 int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu(
-		struct kgd_dev *kgd, uint64_t va, uint64_t size,
+		struct amdgpu_device *adev, uint64_t va, uint64_t size,
 		void *drm_priv, struct kgd_mem **mem,
-		uint64_t *offset, uint32_t flags)
+		uint64_t *offset, uint32_t flags, bool criu_resume)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct amdgpu_vm *avm = drm_priv_to_vm(drm_priv);
+	struct amdgpu_fpriv *fpriv = container_of(avm, struct amdgpu_fpriv, vm);
 	enum ttm_bo_type bo_type = ttm_bo_type_device;
 	struct sg_table *sg = NULL;
 	uint64_t user_addr = 0;
 	struct amdgpu_bo *bo;
 	struct drm_gem_object *gobj = NULL;
 	u32 domain, alloc_domain;
+	uint64_t aligned_size;
+	int8_t xcp_id = -1;
 	u64 alloc_flags;
 	int ret;
 
@@ -1406,34 +1714,54 @@ int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu(
 	 */
 	if (flags & KFD_IOC_ALLOC_MEM_FLAGS_VRAM) {
 		domain = alloc_domain = AMDGPU_GEM_DOMAIN_VRAM;
-		alloc_flags = AMDGPU_GEM_CREATE_VRAM_WIPE_ON_RELEASE;
-		alloc_flags |= (flags & KFD_IOC_ALLOC_MEM_FLAGS_PUBLIC) ?
+
+		if (adev->apu_prefer_gtt) {
+			domain = AMDGPU_GEM_DOMAIN_GTT;
+			alloc_domain = AMDGPU_GEM_DOMAIN_GTT;
+			alloc_flags = 0;
+		} else {
+			alloc_flags = AMDGPU_GEM_CREATE_VRAM_WIPE_ON_RELEASE;
+			alloc_flags |= (flags & KFD_IOC_ALLOC_MEM_FLAGS_PUBLIC) ?
 			AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED : 0;
+
+			/* For contiguous VRAM allocation */
+			if (flags & KFD_IOC_ALLOC_MEM_FLAGS_CONTIGUOUS)
+				alloc_flags |= AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS;
+		}
+		xcp_id = fpriv->xcp_id == AMDGPU_XCP_NO_PARTITION ?
+					0 : fpriv->xcp_id;
 	} else if (flags & KFD_IOC_ALLOC_MEM_FLAGS_GTT) {
 		domain = alloc_domain = AMDGPU_GEM_DOMAIN_GTT;
 		alloc_flags = 0;
-	} else if (flags & KFD_IOC_ALLOC_MEM_FLAGS_USERPTR) {
+	} else {
 		domain = AMDGPU_GEM_DOMAIN_GTT;
 		alloc_domain = AMDGPU_GEM_DOMAIN_CPU;
 		alloc_flags = AMDGPU_GEM_CREATE_PREEMPTIBLE;
-		if (!offset || !*offset)
-			return -EINVAL;
-		user_addr = untagged_addr(*offset);
-	} else if (flags & (KFD_IOC_ALLOC_MEM_FLAGS_DOORBELL |
-			KFD_IOC_ALLOC_MEM_FLAGS_MMIO_REMAP)) {
-		domain = AMDGPU_GEM_DOMAIN_GTT;
-		alloc_domain = AMDGPU_GEM_DOMAIN_CPU;
-		bo_type = ttm_bo_type_sg;
-		alloc_flags = 0;
-		if (size > UINT_MAX)
+
+		if (flags & KFD_IOC_ALLOC_MEM_FLAGS_USERPTR) {
+			if (!offset || !*offset)
+				return -EINVAL;
+			user_addr = untagged_addr(*offset);
+		} else if (flags & (KFD_IOC_ALLOC_MEM_FLAGS_DOORBELL |
+				    KFD_IOC_ALLOC_MEM_FLAGS_MMIO_REMAP)) {
+			bo_type = ttm_bo_type_sg;
+			if (size > UINT_MAX)
+				return -EINVAL;
+			sg = create_sg_table(*offset, size);
+			if (!sg)
+				return -ENOMEM;
+		} else {
 			return -EINVAL;
-		sg = create_doorbell_sg(*offset, size);
-		if (!sg)
-			return -ENOMEM;
-	} else {
-		return -EINVAL;
+		}
 	}
 
+	if (flags & KFD_IOC_ALLOC_MEM_FLAGS_COHERENT)
+		alloc_flags |= AMDGPU_GEM_CREATE_COHERENT;
+	if (flags & KFD_IOC_ALLOC_MEM_FLAGS_EXT_COHERENT)
+		alloc_flags |= AMDGPU_GEM_CREATE_EXT_COHERENT;
+	if (flags & KFD_IOC_ALLOC_MEM_FLAGS_UNCACHED)
+		alloc_flags |= AMDGPU_GEM_CREATE_UNCACHED;
+
 	*mem = kzalloc(sizeof(struct kgd_mem), GFP_KERNEL);
 	if (!*mem) {
 		ret = -ENOMEM;
@@ -1448,23 +1776,26 @@ int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu(
 	 * the memory.
 	 */
 	if ((*mem)->aql_queue)
-		size = size >> 1;
+		size >>= 1;
+	aligned_size = PAGE_ALIGN(size);
 
 	(*mem)->alloc_flags = flags;
 
 	amdgpu_sync_create(&(*mem)->sync);
 
-	ret = amdgpu_amdkfd_reserve_mem_limit(adev, size, alloc_domain, !!sg);
+	ret = amdgpu_amdkfd_reserve_mem_limit(adev, aligned_size, flags,
+					      xcp_id);
 	if (ret) {
 		pr_debug("Insufficient memory\n");
 		goto err_reserve_limit;
 	}
 
-	pr_debug("\tcreate BO VA 0x%llx size 0x%llx domain %s\n",
-			va, size, domain_string(alloc_domain));
+	pr_debug("\tcreate BO VA 0x%llx size 0x%llx domain %s xcp_id %d\n",
+		 va, (*mem)->aql_queue ? size << 1 : size,
+		 domain_string(alloc_domain), xcp_id);
 
-	ret = amdgpu_gem_object_create(adev, size, 1, alloc_domain, alloc_flags,
-				       bo_type, NULL, &gobj);
+	ret = amdgpu_gem_object_create(adev, aligned_size, 1, alloc_domain, alloc_flags,
+				       bo_type, NULL, &gobj, xcp_id + 1);
 	if (ret) {
 		pr_debug("Failed to create BO on domain %s. ret %d\n",
 			 domain_string(alloc_domain), ret);
@@ -1475,6 +1806,9 @@ int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu(
 		pr_debug("Failed to allow vma node access. ret %d\n", ret);
 		goto err_node_allow;
 	}
+	ret = drm_gem_handle_create(adev->kfd.client.file, gobj, &(*mem)->gem_handle);
+	if (ret)
+		goto err_gem_handle_create;
 	bo = gem_to_amdgpu_bo(gobj);
 	if (bo_type == ttm_bo_type_sg) {
 		bo->tbo.sg = sg;
@@ -1489,12 +1823,32 @@ int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu(
 	(*mem)->domain = domain;
 	(*mem)->mapped_to_gpu_memory = 0;
 	(*mem)->process_info = avm->process_info;
+
 	add_kgd_mem_to_kfd_bo_list(*mem, avm->process_info, user_addr);
 
 	if (user_addr) {
-		ret = init_user_pages(*mem, user_addr);
+		pr_debug("creating userptr BO for user_addr = %llx\n", user_addr);
+		ret = init_user_pages(*mem, user_addr, criu_resume);
 		if (ret)
 			goto allocate_init_user_pages_failed;
+	} else  if (flags & (KFD_IOC_ALLOC_MEM_FLAGS_DOORBELL |
+				KFD_IOC_ALLOC_MEM_FLAGS_MMIO_REMAP)) {
+		ret = amdgpu_amdkfd_gpuvm_pin_bo(bo, AMDGPU_GEM_DOMAIN_GTT);
+		if (ret) {
+			pr_err("Pinning MMIO/DOORBELL BO during ALLOC FAILED\n");
+			goto err_pin_bo;
+		}
+		bo->allowed_domains = AMDGPU_GEM_DOMAIN_GTT;
+		bo->preferred_domains = AMDGPU_GEM_DOMAIN_GTT;
+	} else {
+		mutex_lock(&avm->process_info->lock);
+		if (avm->process_info->eviction_fence &&
+		    !dma_fence_is_signaled(&avm->process_info->eviction_fence->base))
+			ret = amdgpu_amdkfd_bo_validate_and_fence(bo, domain,
+				&avm->process_info->eviction_fence->base);
+		mutex_unlock(&avm->process_info->lock);
+		if (ret)
+			goto err_validate_bo;
 	}
 
 	if (offset)
@@ -1503,14 +1857,19 @@ int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu(
 	return 0;
 
 allocate_init_user_pages_failed:
+err_pin_bo:
+err_validate_bo:
 	remove_kgd_mem_from_kfd_bo_list(*mem, avm->process_info);
+	drm_gem_handle_delete(adev->kfd.client.file, (*mem)->gem_handle);
+err_gem_handle_create:
 	drm_vma_node_revoke(&gobj->vma_node, drm_priv);
 err_node_allow:
 	/* Don't unreserve system mem limit twice */
 	goto err_reserve_limit;
 err_bo_create:
-	unreserve_mem_limit(adev, size, alloc_domain, !!sg);
+	amdgpu_amdkfd_unreserve_mem_limit(adev, aligned_size, flags, xcp_id);
 err_reserve_limit:
+	amdgpu_sync_free(&(*mem)->sync);
 	mutex_destroy(&(*mem)->lock);
 	if (gobj)
 		drm_gem_object_put(gobj);
@@ -1525,19 +1884,27 @@ err:
 }
 
 int amdgpu_amdkfd_gpuvm_free_memory_of_gpu(
-		struct kgd_dev *kgd, struct kgd_mem *mem, void *drm_priv,
+		struct amdgpu_device *adev, struct kgd_mem *mem, void *drm_priv,
 		uint64_t *size)
 {
 	struct amdkfd_process_info *process_info = mem->process_info;
 	unsigned long bo_size = mem->bo->tbo.base.size;
+	bool use_release_notifier = (mem->bo->kfd_bo == mem);
 	struct kfd_mem_attachment *entry, *tmp;
 	struct bo_vm_reservation_context ctx;
-	struct ttm_validate_buffer *bo_list_entry;
 	unsigned int mapped_to_gpu_memory;
 	int ret;
 	bool is_imported = false;
 
 	mutex_lock(&mem->lock);
+
+	/* Unpin MMIO/DOORBELL BO's that were pinned during allocation */
+	if (mem->alloc_flags &
+	    (KFD_IOC_ALLOC_MEM_FLAGS_DOORBELL |
+	     KFD_IOC_ALLOC_MEM_FLAGS_MMIO_REMAP)) {
+		amdgpu_amdkfd_gpuvm_unpin_bo(mem->bo);
+	}
+
 	mapped_to_gpu_memory = mem->mapped_to_gpu_memory;
 	is_imported = mem->is_imported;
 	mutex_unlock(&mem->lock);
@@ -1552,30 +1919,32 @@ int amdgpu_amdkfd_gpuvm_free_memory_of_gpu(
 	}
 
 	/* Make sure restore workers don't access the BO any more */
-	bo_list_entry = &mem->validate_list;
 	mutex_lock(&process_info->lock);
-	list_del(&bo_list_entry->head);
+	list_del(&mem->validate_list);
 	mutex_unlock(&process_info->lock);
 
-	/* No more MMU notifiers */
-	amdgpu_mn_unregister(mem->bo);
+	/* Cleanup user pages and MMU notifiers */
+	if (amdgpu_ttm_tt_get_usermm(mem->bo->tbo.ttm)) {
+		amdgpu_hmm_unregister(mem->bo);
+		mutex_lock(&process_info->notifier_lock);
+		amdgpu_hmm_range_free(mem->range);
+		mutex_unlock(&process_info->notifier_lock);
+	}
 
 	ret = reserve_bo_and_cond_vms(mem, NULL, BO_VM_ALL, &ctx);
 	if (unlikely(ret))
 		return ret;
 
-	/* The eviction fence should be removed by the last unmap.
-	 * TODO: Log an error condition if the bo still has the eviction fence
-	 * attached
-	 */
 	amdgpu_amdkfd_remove_eviction_fence(mem->bo,
 					process_info->eviction_fence);
 	pr_debug("Release VA 0x%llx - 0x%llx\n", mem->va,
 		mem->va + bo_size * (1 + mem->aql_queue));
 
 	/* Remove from VM internal data structures */
-	list_for_each_entry_safe(entry, tmp, &mem->attachments, list)
+	list_for_each_entry_safe(entry, tmp, &mem->attachments, list) {
+		kfd_mem_dmaunmap_attachment(mem, entry);
 		kfd_mem_detach(entry);
+	}
 
 	ret = unreserve_bo_and_vms(&ctx, false, false);
 
@@ -1591,11 +1960,12 @@ int amdgpu_amdkfd_gpuvm_free_memory_of_gpu(
 	}
 
 	/* Update the size of the BO being freed if it was allocated from
-	 * VRAM and is not imported.
+	 * VRAM and is not imported. For APP APU VRAM allocations are done
+	 * in GTT domain
 	 */
 	if (size) {
-		if ((mem->bo->preferred_domains == AMDGPU_GEM_DOMAIN_VRAM) &&
-		    (!is_imported))
+		if (!is_imported &&
+		   mem->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_VRAM)
 			*size = bo_size;
 		else
 			*size = 0;
@@ -1603,8 +1973,11 @@ int amdgpu_amdkfd_gpuvm_free_memory_of_gpu(
 
 	/* Free the BO*/
 	drm_vma_node_revoke(&mem->bo->tbo.base.vma_node, drm_priv);
-	if (mem->dmabuf)
+	drm_gem_handle_delete(adev->kfd.client.file, mem->gem_handle);
+	if (mem->dmabuf) {
 		dma_buf_put(mem->dmabuf);
+		mem->dmabuf = NULL;
+	}
 	mutex_destroy(&mem->lock);
 
 	/* If this releases the last reference, it will end up calling
@@ -1613,14 +1986,20 @@ int amdgpu_amdkfd_gpuvm_free_memory_of_gpu(
 	 */
 	drm_gem_object_put(&mem->bo->tbo.base);
 
+	/*
+	 * For kgd_mem allocated in amdgpu_amdkfd_gpuvm_import_dmabuf(),
+	 * explicitly free it here.
+	 */
+	if (!use_release_notifier)
+		kfree(mem);
+
 	return ret;
 }
 
 int amdgpu_amdkfd_gpuvm_map_memory_to_gpu(
-		struct kgd_dev *kgd, struct kgd_mem *mem,
-		void *drm_priv, bool *table_freed)
+		struct amdgpu_device *adev, struct kgd_mem *mem,
+		void *drm_priv)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct amdgpu_vm *avm = drm_priv_to_vm(drm_priv);
 	int ret;
 	struct amdgpu_bo *bo;
@@ -1642,14 +2021,14 @@ int amdgpu_amdkfd_gpuvm_map_memory_to_gpu(
 	 */
 	mutex_lock(&mem->process_info->lock);
 
-	/* Lock mmap-sem. If we find an invalid userptr BO, we can be
+	/* Lock notifier lock. If we find an invalid userptr BO, we can be
 	 * sure that the MMU notifier is no longer running
 	 * concurrently and the queues are actually stopped
 	 */
 	if (amdgpu_ttm_tt_get_usermm(bo->tbo.ttm)) {
-		mmap_write_lock(current->mm);
-		is_invalid_userptr = atomic_read(&mem->invalid);
-		mmap_write_unlock(current->mm);
+		mutex_lock(&mem->process_info->notifier_lock);
+		is_invalid_userptr = !!mem->invalid;
+		mutex_unlock(&mem->process_info->notifier_lock);
 	}
 
 	mutex_lock(&mem->lock);
@@ -1681,23 +2060,10 @@ int amdgpu_amdkfd_gpuvm_map_memory_to_gpu(
 	    bo->tbo.resource->mem_type == TTM_PL_SYSTEM)
 		is_invalid_userptr = true;
 
-	ret = vm_validate_pt_pd_bos(avm);
+	ret = vm_validate_pt_pd_bos(avm, NULL);
 	if (unlikely(ret))
 		goto out_unreserve;
 
-	if (mem->mapped_to_gpu_memory == 0 &&
-	    !amdgpu_ttm_tt_get_usermm(bo->tbo.ttm)) {
-		/* Validate BO only once. The eviction fence gets added to BO
-		 * the first time it is mapped. Validate will wait for all
-		 * background evictions to complete.
-		 */
-		ret = amdgpu_amdkfd_bo_validate(bo, domain, true);
-		if (ret) {
-			pr_debug("Validate failed\n");
-			goto out_unreserve;
-		}
-	}
-
 	list_for_each_entry(entry, &mem->attachments, list) {
 		if (entry->bo_va->base.vm != avm || entry->is_mapped)
 			continue;
@@ -1706,7 +2072,7 @@ int amdgpu_amdkfd_gpuvm_map_memory_to_gpu(
 			 entry->va, entry->va + bo_size, entry);
 
 		ret = map_bo_to_gpuvm(mem, entry, ctx.sync,
-				      is_invalid_userptr, table_freed);
+				      is_invalid_userptr);
 		if (ret) {
 			pr_err("Failed to map bo to gpuvm\n");
 			goto out_unreserve;
@@ -1724,18 +2090,8 @@ int amdgpu_amdkfd_gpuvm_map_memory_to_gpu(
 			 mem->mapped_to_gpu_memory);
 	}
 
-	if (!amdgpu_ttm_tt_get_usermm(bo->tbo.ttm) && !bo->tbo.pin_count)
-		amdgpu_bo_fence(bo,
-				&avm->process_info->eviction_fence->base,
-				true);
 	ret = unreserve_bo_and_vms(&ctx, false, false);
 
-	/* Only apply no TLB flush on Aldebaran to
-	 * workaround regressions on other Asics.
-	 */
-	if (table_freed && (adev->asic_type != CHIP_ALDEBARAN))
-		*table_freed = true;
-
 	goto out;
 
 out_unreserve:
@@ -1746,11 +2102,41 @@ out:
 	return ret;
 }
 
+int amdgpu_amdkfd_gpuvm_dmaunmap_mem(struct kgd_mem *mem, void *drm_priv)
+{
+	struct kfd_mem_attachment *entry;
+	struct amdgpu_vm *vm;
+	int ret;
+
+	vm = drm_priv_to_vm(drm_priv);
+
+	mutex_lock(&mem->lock);
+
+	ret = amdgpu_bo_reserve(mem->bo, true);
+	if (ret)
+		goto out;
+
+	list_for_each_entry(entry, &mem->attachments, list) {
+		if (entry->bo_va->base.vm != vm)
+			continue;
+		if (entry->bo_va->base.bo->tbo.ttm &&
+		    !entry->bo_va->base.bo->tbo.ttm->sg)
+			continue;
+
+		kfd_mem_dmaunmap_attachment(mem, entry);
+	}
+
+	amdgpu_bo_unreserve(mem->bo);
+out:
+	mutex_unlock(&mem->lock);
+
+	return ret;
+}
+
 int amdgpu_amdkfd_gpuvm_unmap_memory_from_gpu(
-		struct kgd_dev *kgd, struct kgd_mem *mem, void *drm_priv)
+		struct amdgpu_device *adev, struct kgd_mem *mem, void *drm_priv)
 {
 	struct amdgpu_vm *avm = drm_priv_to_vm(drm_priv);
-	struct amdkfd_process_info *process_info = avm->process_info;
 	unsigned long bo_size = mem->bo->tbo.base.size;
 	struct kfd_mem_attachment *entry;
 	struct bo_vm_reservation_context ctx;
@@ -1767,7 +2153,7 @@ int amdgpu_amdkfd_gpuvm_unmap_memory_from_gpu(
 		goto unreserve_out;
 	}
 
-	ret = vm_validate_pt_pd_bos(avm);
+	ret = vm_validate_pt_pd_bos(avm, NULL);
 	if (unlikely(ret))
 		goto unreserve_out;
 
@@ -1783,7 +2169,10 @@ int amdgpu_amdkfd_gpuvm_unmap_memory_from_gpu(
 		pr_debug("\t unmap VA 0x%llx - 0x%llx from entry %p\n",
 			 entry->va, entry->va + bo_size, entry);
 
-		unmap_bo_from_gpuvm(mem, entry, ctx.sync);
+		ret = unmap_bo_from_gpuvm(mem, entry, ctx.sync);
+		if (ret)
+			goto unreserve_out;
+
 		entry->is_mapped = false;
 
 		mem->mapped_to_gpu_memory--;
@@ -1791,15 +2180,6 @@ int amdgpu_amdkfd_gpuvm_unmap_memory_from_gpu(
 			 mem->mapped_to_gpu_memory);
 	}
 
-	/* If BO is unmapped from all VMs, unfence it. It can be evicted if
-	 * required.
-	 */
-	if (mem->mapped_to_gpu_memory == 0 &&
-	    !amdgpu_ttm_tt_get_usermm(mem->bo->tbo.ttm) &&
-	    !mem->bo->tbo.pin_count)
-		amdgpu_amdkfd_remove_eviction_fence(mem->bo,
-						process_info->eviction_fence);
-
 unreserve_out:
 	unreserve_bo_and_vms(&ctx, false, false);
 out:
@@ -1808,7 +2188,7 @@ out:
 }
 
 int amdgpu_amdkfd_gpuvm_sync_memory(
-		struct kgd_dev *kgd, struct kgd_mem *mem, bool intr)
+		struct amdgpu_device *adev, struct kgd_mem *mem, bool intr)
 {
 	struct amdgpu_sync sync;
 	int ret;
@@ -1824,8 +2204,69 @@ int amdgpu_amdkfd_gpuvm_sync_memory(
 	return ret;
 }
 
-int amdgpu_amdkfd_gpuvm_map_gtt_bo_to_kernel(struct kgd_dev *kgd,
-		struct kgd_mem *mem, void **kptr, uint64_t *size)
+/**
+ * amdgpu_amdkfd_map_gtt_bo_to_gart - Map BO to GART and increment reference count
+ * @bo: Buffer object to be mapped
+ * @bo_gart: Return bo reference
+ *
+ * Before return, bo reference count is incremented. To release the reference and unpin/
+ * unmap the BO, call amdgpu_amdkfd_free_gtt_mem.
+ */
+int amdgpu_amdkfd_map_gtt_bo_to_gart(struct amdgpu_bo *bo, struct amdgpu_bo **bo_gart)
+{
+	int ret;
+
+	ret = amdgpu_bo_reserve(bo, true);
+	if (ret) {
+		pr_err("Failed to reserve bo. ret %d\n", ret);
+		goto err_reserve_bo_failed;
+	}
+
+	ret = amdgpu_bo_pin(bo, AMDGPU_GEM_DOMAIN_GTT);
+	if (ret) {
+		pr_err("Failed to pin bo. ret %d\n", ret);
+		goto err_pin_bo_failed;
+	}
+
+	ret = amdgpu_ttm_alloc_gart(&bo->tbo);
+	if (ret) {
+		pr_err("Failed to bind bo to GART. ret %d\n", ret);
+		goto err_map_bo_gart_failed;
+	}
+
+	amdgpu_amdkfd_remove_eviction_fence(
+		bo, bo->vm_bo->vm->process_info->eviction_fence);
+
+	amdgpu_bo_unreserve(bo);
+
+	*bo_gart = amdgpu_bo_ref(bo);
+
+	return 0;
+
+err_map_bo_gart_failed:
+	amdgpu_bo_unpin(bo);
+err_pin_bo_failed:
+	amdgpu_bo_unreserve(bo);
+err_reserve_bo_failed:
+
+	return ret;
+}
+
+/** amdgpu_amdkfd_gpuvm_map_gtt_bo_to_kernel() - Map a GTT BO for kernel CPU access
+ *
+ * @mem: Buffer object to be mapped for CPU access
+ * @kptr[out]: pointer in kernel CPU address space
+ * @size[out]: size of the buffer
+ *
+ * Pins the BO and maps it for kernel CPU access. The eviction fence is removed
+ * from the BO, since pinned BOs cannot be evicted. The bo must remain on the
+ * validate_list, so the GPU mapping can be restored after a page table was
+ * evicted.
+ *
+ * Return: 0 on success, error code on failure
+ */
+int amdgpu_amdkfd_gpuvm_map_gtt_bo_to_kernel(struct kgd_mem *mem,
+					     void **kptr, uint64_t *size)
 {
 	int ret;
 	struct amdgpu_bo *bo = mem->bo;
@@ -1835,9 +2276,6 @@ int amdgpu_amdkfd_gpuvm_map_gtt_bo_to_kernel(struct kgd_dev *kgd,
 		return -EINVAL;
 	}
 
-	/* delete kgd_mem from kfd_bo_list to avoid re-validating
-	 * this BO in BO's restoring after eviction.
-	 */
 	mutex_lock(&mem->process_info->lock);
 
 	ret = amdgpu_bo_reserve(bo, true);
@@ -1860,7 +2298,6 @@ int amdgpu_amdkfd_gpuvm_map_gtt_bo_to_kernel(struct kgd_dev *kgd,
 
 	amdgpu_amdkfd_remove_eviction_fence(
 		bo, mem->process_info->eviction_fence);
-	list_del_init(&mem->validate_list.head);
 
 	if (size)
 		*size = amdgpu_bo_size(bo);
@@ -1880,51 +2317,45 @@ bo_reserve_failed:
 	return ret;
 }
 
-void amdgpu_amdkfd_gpuvm_unmap_gtt_bo_from_kernel(struct kgd_dev *kgd, struct kgd_mem *mem)
+/** amdgpu_amdkfd_gpuvm_map_gtt_bo_to_kernel() - Unmap a GTT BO for kernel CPU access
+ *
+ * @mem: Buffer object to be unmapped for CPU access
+ *
+ * Removes the kernel CPU mapping and unpins the BO. It does not restore the
+ * eviction fence, so this function should only be used for cleanup before the
+ * BO is destroyed.
+ */
+void amdgpu_amdkfd_gpuvm_unmap_gtt_bo_from_kernel(struct kgd_mem *mem)
 {
 	struct amdgpu_bo *bo = mem->bo;
 
-	amdgpu_bo_reserve(bo, true);
+	(void)amdgpu_bo_reserve(bo, true);
 	amdgpu_bo_kunmap(bo);
 	amdgpu_bo_unpin(bo);
 	amdgpu_bo_unreserve(bo);
 }
 
-int amdgpu_amdkfd_gpuvm_get_vm_fault_info(struct kgd_dev *kgd,
-					      struct kfd_vm_fault_info *mem)
+int amdgpu_amdkfd_gpuvm_get_vm_fault_info(struct amdgpu_device *adev,
+					  struct kfd_vm_fault_info *mem)
 {
-	struct amdgpu_device *adev;
-
-	adev = (struct amdgpu_device *)kgd;
-	if (atomic_read(&adev->gmc.vm_fault_info_updated) == 1) {
+	if (atomic_read_acquire(&adev->gmc.vm_fault_info_updated) == 1) {
 		*mem = *adev->gmc.vm_fault_info;
-		mb();
-		atomic_set(&adev->gmc.vm_fault_info_updated, 0);
+		atomic_set_release(&adev->gmc.vm_fault_info_updated, 0);
 	}
 	return 0;
 }
 
-int amdgpu_amdkfd_gpuvm_import_dmabuf(struct kgd_dev *kgd,
-				      struct dma_buf *dma_buf,
-				      uint64_t va, void *drm_priv,
-				      struct kgd_mem **mem, uint64_t *size,
-				      uint64_t *mmap_offset)
+static int import_obj_create(struct amdgpu_device *adev,
+			     struct dma_buf *dma_buf,
+			     struct drm_gem_object *obj,
+			     uint64_t va, void *drm_priv,
+			     struct kgd_mem **mem, uint64_t *size,
+			     uint64_t *mmap_offset)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
 	struct amdgpu_vm *avm = drm_priv_to_vm(drm_priv);
-	struct drm_gem_object *obj;
 	struct amdgpu_bo *bo;
 	int ret;
 
-	if (dma_buf->ops != &amdgpu_dmabuf_ops)
-		/* Can't handle non-graphics buffers */
-		return -EINVAL;
-
-	obj = dma_buf->priv;
-	if (drm_to_adev(obj->dev) != adev)
-		/* Can't handle buffers from other devices */
-		return -EINVAL;
-
 	bo = gem_to_amdgpu_bo(obj);
 	if (!(bo->preferred_domains & (AMDGPU_GEM_DOMAIN_VRAM |
 				    AMDGPU_GEM_DOMAIN_GTT)))
@@ -1936,10 +2367,8 @@ int amdgpu_amdkfd_gpuvm_import_dmabuf(struct kgd_dev *kgd,
 		return -ENOMEM;
 
 	ret = drm_vma_node_allow(&obj->vma_node, drm_priv);
-	if (ret) {
-		kfree(mem);
-		return ret;
-	}
+	if (ret)
+		goto err_free_mem;
 
 	if (size)
 		*size = amdgpu_bo_size(bo);
@@ -1956,48 +2385,131 @@ int amdgpu_amdkfd_gpuvm_import_dmabuf(struct kgd_dev *kgd,
 		| KFD_IOC_ALLOC_MEM_FLAGS_WRITABLE
 		| KFD_IOC_ALLOC_MEM_FLAGS_EXECUTABLE;
 
-	drm_gem_object_get(&bo->tbo.base);
+	get_dma_buf(dma_buf);
+	(*mem)->dmabuf = dma_buf;
 	(*mem)->bo = bo;
 	(*mem)->va = va;
-	(*mem)->domain = (bo->preferred_domains & AMDGPU_GEM_DOMAIN_VRAM) ?
-		AMDGPU_GEM_DOMAIN_VRAM : AMDGPU_GEM_DOMAIN_GTT;
+	(*mem)->domain = (bo->preferred_domains & AMDGPU_GEM_DOMAIN_VRAM) &&
+			 !adev->apu_prefer_gtt ?
+			 AMDGPU_GEM_DOMAIN_VRAM : AMDGPU_GEM_DOMAIN_GTT;
+
 	(*mem)->mapped_to_gpu_memory = 0;
 	(*mem)->process_info = avm->process_info;
 	add_kgd_mem_to_kfd_bo_list(*mem, avm->process_info, false);
 	amdgpu_sync_create(&(*mem)->sync);
 	(*mem)->is_imported = true;
 
+	mutex_lock(&avm->process_info->lock);
+	if (avm->process_info->eviction_fence &&
+	    !dma_fence_is_signaled(&avm->process_info->eviction_fence->base))
+		ret = amdgpu_amdkfd_bo_validate_and_fence(bo, (*mem)->domain,
+				&avm->process_info->eviction_fence->base);
+	mutex_unlock(&avm->process_info->lock);
+	if (ret)
+		goto err_remove_mem;
+
+	return 0;
+
+err_remove_mem:
+	remove_kgd_mem_from_kfd_bo_list(*mem, avm->process_info);
+	drm_vma_node_revoke(&obj->vma_node, drm_priv);
+err_free_mem:
+	kfree(*mem);
+	return ret;
+}
+
+int amdgpu_amdkfd_gpuvm_import_dmabuf_fd(struct amdgpu_device *adev, int fd,
+					 uint64_t va, void *drm_priv,
+					 struct kgd_mem **mem, uint64_t *size,
+					 uint64_t *mmap_offset)
+{
+	struct drm_gem_object *obj;
+	uint32_t handle;
+	int ret;
+
+	ret = drm_gem_prime_fd_to_handle(&adev->ddev, adev->kfd.client.file, fd,
+					 &handle);
+	if (ret)
+		return ret;
+	obj = drm_gem_object_lookup(adev->kfd.client.file, handle);
+	if (!obj) {
+		ret = -EINVAL;
+		goto err_release_handle;
+	}
+
+	ret = import_obj_create(adev, obj->dma_buf, obj, va, drm_priv, mem, size,
+				mmap_offset);
+	if (ret)
+		goto err_put_obj;
+
+	(*mem)->gem_handle = handle;
+
 	return 0;
+
+err_put_obj:
+	drm_gem_object_put(obj);
+err_release_handle:
+	drm_gem_handle_delete(adev->kfd.client.file, handle);
+	return ret;
+}
+
+int amdgpu_amdkfd_gpuvm_export_dmabuf(struct kgd_mem *mem,
+				      struct dma_buf **dma_buf)
+{
+	int ret;
+
+	mutex_lock(&mem->lock);
+	ret = kfd_mem_export_dmabuf(mem);
+	if (ret)
+		goto out;
+
+	get_dma_buf(mem->dmabuf);
+	*dma_buf = mem->dmabuf;
+out:
+	mutex_unlock(&mem->lock);
+	return ret;
 }
 
 /* Evict a userptr BO by stopping the queues if necessary
  *
  * Runs in MMU notifier, may be in RECLAIM_FS context. This means it
  * cannot do any memory allocations, and cannot take any locks that
- * are held elsewhere while allocating memory. Therefore this is as
- * simple as possible, using atomic counters.
+ * are held elsewhere while allocating memory.
  *
  * It doesn't do anything to the BO itself. The real work happens in
  * restore, where we get updated page addresses. This function only
  * ensures that GPU access to the BO is stopped.
  */
-int amdgpu_amdkfd_evict_userptr(struct kgd_mem *mem,
-				struct mm_struct *mm)
+int amdgpu_amdkfd_evict_userptr(struct mmu_interval_notifier *mni,
+				unsigned long cur_seq, struct kgd_mem *mem)
 {
 	struct amdkfd_process_info *process_info = mem->process_info;
-	int evicted_bos;
 	int r = 0;
 
-	atomic_inc(&mem->invalid);
-	evicted_bos = atomic_inc_return(&process_info->evicted_bos);
-	if (evicted_bos == 1) {
+	/* Do not process MMU notifications during CRIU restore until
+	 * KFD_CRIU_OP_RESUME IOCTL is received
+	 */
+	if (READ_ONCE(process_info->block_mmu_notifications))
+		return 0;
+
+	mutex_lock(&process_info->notifier_lock);
+	mmu_interval_set_seq(mni, cur_seq);
+
+	mem->invalid++;
+	if (++process_info->evicted_bos == 1) {
 		/* First eviction, stop the queues */
-		r = kgd2kfd_quiesce_mm(mm);
-		if (r)
+		r = kgd2kfd_quiesce_mm(mni->mm,
+				       KFD_QUEUE_EVICTION_TRIGGER_USERPTR);
+
+		if (r && r != -ESRCH)
 			pr_err("Failed to quiesce KFD\n");
-		schedule_delayed_work(&process_info->restore_userptr_work,
-			msecs_to_jiffies(AMDGPU_USERPTR_RESTORE_DELAY_MS));
+
+		if (r != -ESRCH)
+			queue_delayed_work(system_freezable_wq,
+				&process_info->restore_userptr_work,
+				msecs_to_jiffies(AMDGPU_USERPTR_RESTORE_DELAY_MS));
 	}
+	mutex_unlock(&process_info->notifier_lock);
 
 	return r;
 }
@@ -2014,52 +2526,63 @@ static int update_invalid_user_pages(struct amdkfd_process_info *process_info,
 	struct kgd_mem *mem, *tmp_mem;
 	struct amdgpu_bo *bo;
 	struct ttm_operation_ctx ctx = { false, false };
-	int invalid, ret;
+	uint32_t invalid;
+	int ret = 0;
 
-	/* Move all invalidated BOs to the userptr_inval_list and
-	 * release their user pages by migration to the CPU domain
-	 */
+	mutex_lock(&process_info->notifier_lock);
+
+	/* Move all invalidated BOs to the userptr_inval_list */
 	list_for_each_entry_safe(mem, tmp_mem,
 				 &process_info->userptr_valid_list,
-				 validate_list.head) {
-		if (!atomic_read(&mem->invalid))
-			continue; /* BO is still valid */
-
-		bo = mem->bo;
-
-		if (amdgpu_bo_reserve(bo, true))
-			return -EAGAIN;
-		amdgpu_bo_placement_from_domain(bo, AMDGPU_GEM_DOMAIN_CPU);
-		ret = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
-		amdgpu_bo_unreserve(bo);
-		if (ret) {
-			pr_err("%s: Failed to invalidate userptr BO\n",
-			       __func__);
-			return -EAGAIN;
-		}
-
-		list_move_tail(&mem->validate_list.head,
-			       &process_info->userptr_inval_list);
-	}
-
-	if (list_empty(&process_info->userptr_inval_list))
-		return 0; /* All evicted userptr BOs were freed */
+				 validate_list)
+		if (mem->invalid)
+			list_move_tail(&mem->validate_list,
+				       &process_info->userptr_inval_list);
 
 	/* Go through userptr_inval_list and update any invalid user_pages */
 	list_for_each_entry(mem, &process_info->userptr_inval_list,
-			    validate_list.head) {
-		invalid = atomic_read(&mem->invalid);
+			    validate_list) {
+		invalid = mem->invalid;
 		if (!invalid)
 			/* BO hasn't been invalidated since the last
-			 * revalidation attempt. Keep its BO list.
+			 * revalidation attempt. Keep its page list.
 			 */
 			continue;
 
 		bo = mem->bo;
 
+		amdgpu_hmm_range_free(mem->range);
+		mem->range = NULL;
+
+		/* BO reservations and getting user pages (hmm_range_fault)
+		 * must happen outside the notifier lock
+		 */
+		mutex_unlock(&process_info->notifier_lock);
+
+		/* Move the BO to system (CPU) domain if necessary to unmap
+		 * and free the SG table
+		 */
+		if (bo->tbo.resource->mem_type != TTM_PL_SYSTEM) {
+			if (amdgpu_bo_reserve(bo, true))
+				return -EAGAIN;
+			amdgpu_bo_placement_from_domain(bo, AMDGPU_GEM_DOMAIN_CPU);
+			ret = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
+			amdgpu_bo_unreserve(bo);
+			if (ret) {
+				pr_err("%s: Failed to invalidate userptr BO\n",
+				       __func__);
+				return -EAGAIN;
+			}
+		}
+
+		mem->range = amdgpu_hmm_range_alloc(NULL);
+		if (unlikely(!mem->range))
+			return -ENOMEM;
 		/* Get updated user pages */
-		ret = amdgpu_ttm_tt_get_user_pages(bo, bo->tbo.ttm->pages);
+		ret = amdgpu_ttm_tt_get_user_pages(bo, mem->range);
 		if (ret) {
+			amdgpu_hmm_range_free(mem->range);
+			mem->range = NULL;
 			pr_debug("Failed %d to get user pages\n", ret);
 
 			/* Return -EFAULT bad address error as success. It will
@@ -2071,86 +2594,100 @@ static int update_invalid_user_pages(struct amdkfd_process_info *process_info,
 			 */
 			if (ret != -EFAULT)
 				return ret;
-		} else {
 
-			/*
-			 * FIXME: Cannot ignore the return code, must hold
-			 * notifier_lock
+			/* If applications unmap memory before destroying the userptr
+			 * from the KFD, trigger a segmentation fault in VM debug mode.
 			 */
-			amdgpu_ttm_tt_get_user_pages_done(bo->tbo.ttm);
+			if (amdgpu_ttm_adev(bo->tbo.bdev)->debug_vm_userptr) {
+				struct kfd_process *p;
+
+				pr_err("Pid %d unmapped memory before destroying userptr at GPU addr 0x%llx\n",
+								pid_nr(process_info->pid), mem->va);
+
+				// Send GPU VM fault to user space
+				p = kfd_lookup_process_by_pid(process_info->pid);
+				if (p) {
+					kfd_signal_vm_fault_event_with_userptr(p, mem->va);
+					kfd_unref_process(p);
+				}
+			}
+
+			ret = 0;
 		}
 
+		amdgpu_ttm_tt_set_user_pages(bo->tbo.ttm, mem->range);
+
+		mutex_lock(&process_info->notifier_lock);
+
 		/* Mark the BO as valid unless it was invalidated
 		 * again concurrently.
 		 */
-		if (atomic_cmpxchg(&mem->invalid, invalid, 0) != invalid)
-			return -EAGAIN;
+		if (mem->invalid != invalid) {
+			ret = -EAGAIN;
+			goto unlock_out;
+		}
+		 /* set mem valid if mem has hmm range associated */
+		if (mem->range)
+			mem->invalid = 0;
 	}
 
-	return 0;
+unlock_out:
+	mutex_unlock(&process_info->notifier_lock);
+
+	return ret;
 }
 
 /* Validate invalid userptr BOs
  *
- * Validates BOs on the userptr_inval_list, and moves them back to the
- * userptr_valid_list. Also updates GPUVM page tables with new page
- * addresses and waits for the page table updates to complete.
+ * Validates BOs on the userptr_inval_list. Also updates GPUVM page tables
+ * with new page addresses and waits for the page table updates to complete.
  */
 static int validate_invalid_user_pages(struct amdkfd_process_info *process_info)
 {
-	struct amdgpu_bo_list_entry *pd_bo_list_entries;
-	struct list_head resv_list, duplicates;
-	struct ww_acquire_ctx ticket;
+	struct ttm_operation_ctx ctx = { false, false };
 	struct amdgpu_sync sync;
+	struct drm_exec exec;
 
 	struct amdgpu_vm *peer_vm;
 	struct kgd_mem *mem, *tmp_mem;
 	struct amdgpu_bo *bo;
-	struct ttm_operation_ctx ctx = { false, false };
-	int i, ret;
-
-	pd_bo_list_entries = kcalloc(process_info->n_vms,
-				     sizeof(struct amdgpu_bo_list_entry),
-				     GFP_KERNEL);
-	if (!pd_bo_list_entries) {
-		pr_err("%s: Failed to allocate PD BO list entries\n", __func__);
-		ret = -ENOMEM;
-		goto out_no_mem;
-	}
-
-	INIT_LIST_HEAD(&resv_list);
-	INIT_LIST_HEAD(&duplicates);
+	int ret;
 
-	/* Get all the page directory BOs that need to be reserved */
-	i = 0;
-	list_for_each_entry(peer_vm, &process_info->vm_list_head,
-			    vm_list_node)
-		amdgpu_vm_get_pd_bo(peer_vm, &resv_list,
-				    &pd_bo_list_entries[i++]);
-	/* Add the userptr_inval_list entries to resv_list */
-	list_for_each_entry(mem, &process_info->userptr_inval_list,
-			    validate_list.head) {
-		list_add_tail(&mem->resv_list.head, &resv_list);
-		mem->resv_list.bo = mem->validate_list.bo;
-		mem->resv_list.num_shared = mem->validate_list.num_shared;
-	}
+	amdgpu_sync_create(&sync);
 
+	drm_exec_init(&exec, 0, 0);
 	/* Reserve all BOs and page tables for validation */
-	ret = ttm_eu_reserve_buffers(&ticket, &resv_list, false, &duplicates);
-	WARN(!list_empty(&duplicates), "Duplicates should be empty");
-	if (ret)
-		goto out_free;
+	drm_exec_until_all_locked(&exec) {
+		/* Reserve all the page directories */
+		list_for_each_entry(peer_vm, &process_info->vm_list_head,
+				    vm_list_node) {
+			ret = amdgpu_vm_lock_pd(peer_vm, &exec, 2);
+			drm_exec_retry_on_contention(&exec);
+			if (unlikely(ret))
+				goto unreserve_out;
+		}
 
-	amdgpu_sync_create(&sync);
+		/* Reserve the userptr_inval_list entries to resv_list */
+		list_for_each_entry(mem, &process_info->userptr_inval_list,
+				    validate_list) {
+			struct drm_gem_object *gobj;
 
-	ret = process_validate_vms(process_info);
+			gobj = &mem->bo->tbo.base;
+			ret = drm_exec_prepare_obj(&exec, gobj, 1);
+			drm_exec_retry_on_contention(&exec);
+			if (unlikely(ret))
+				goto unreserve_out;
+		}
+	}
+
+	ret = process_validate_vms(process_info, NULL);
 	if (ret)
 		goto unreserve_out;
 
 	/* Validate BOs and update GPUVM page tables */
 	list_for_each_entry_safe(mem, tmp_mem,
 				 &process_info->userptr_inval_list,
-				 validate_list.head) {
+				 validate_list) {
 		struct kfd_mem_attachment *attachment;
 
 		bo = mem->bo;
@@ -2165,9 +2702,6 @@ static int validate_invalid_user_pages(struct amdkfd_process_info *process_info)
 			}
 		}
 
-		list_move_tail(&mem->validate_list.head,
-			       &process_info->userptr_valid_list);
-
 		/* Update mapping. If the BO was not validated
 		 * (because we couldn't get user pages), this will
 		 * clear the page table entries, which will result in
@@ -2179,11 +2713,13 @@ static int validate_invalid_user_pages(struct amdkfd_process_info *process_info)
 				continue;
 
 			kfd_mem_dmaunmap_attachment(mem, attachment);
-			ret = update_gpuvm_pte(mem, attachment, &sync, NULL);
+			ret = update_gpuvm_pte(mem, attachment, &sync);
 			if (ret) {
 				pr_err("%s: update PTE failed\n", __func__);
 				/* make sure this gets validated again */
-				atomic_inc(&mem->invalid);
+				mutex_lock(&process_info->notifier_lock);
+				mem->invalid++;
+				mutex_unlock(&process_info->notifier_lock);
 				goto unreserve_out;
 			}
 		}
@@ -2193,12 +2729,51 @@ static int validate_invalid_user_pages(struct amdkfd_process_info *process_info)
 	ret = process_update_pds(process_info, &sync);
 
 unreserve_out:
-	ttm_eu_backoff_reservation(&ticket, &resv_list);
+	drm_exec_fini(&exec);
 	amdgpu_sync_wait(&sync, false);
 	amdgpu_sync_free(&sync);
-out_free:
-	kfree(pd_bo_list_entries);
-out_no_mem:
+
+	return ret;
+}
+
+/* Confirm that all user pages are valid while holding the notifier lock
+ *
+ * Moves valid BOs from the userptr_inval_list back to userptr_val_list.
+ */
+static int confirm_valid_user_pages_locked(struct amdkfd_process_info *process_info)
+{
+	struct kgd_mem *mem, *tmp_mem;
+	int ret = 0;
+
+	list_for_each_entry_safe(mem, tmp_mem,
+				 &process_info->userptr_inval_list,
+				 validate_list) {
+		bool valid;
+
+		/* keep mem without hmm range at userptr_inval_list */
+		if (!mem->range)
+			continue;
+
+		/* Only check mem with hmm range associated */
+		valid = amdgpu_hmm_range_valid(mem->range);
+		amdgpu_hmm_range_free(mem->range);
+
+		mem->range = NULL;
+		if (!valid) {
+			WARN(!mem->invalid, "Invalid BO not marked invalid");
+			ret = -EAGAIN;
+			continue;
+		}
+
+		if (mem->invalid) {
+			WARN(1, "Valid BO is marked invalid");
+			ret = -EAGAIN;
+			continue;
+		}
+
+		list_move_tail(&mem->validate_list,
+			       &process_info->userptr_valid_list);
+	}
 
 	return ret;
 }
@@ -2217,9 +2792,11 @@ static void amdgpu_amdkfd_restore_userptr_worker(struct work_struct *work)
 			     restore_userptr_work);
 	struct task_struct *usertask;
 	struct mm_struct *mm;
-	int evicted_bos;
+	uint32_t evicted_bos;
 
-	evicted_bos = atomic_read(&process_info->evicted_bos);
+	mutex_lock(&process_info->notifier_lock);
+	evicted_bos = process_info->evicted_bos;
+	mutex_unlock(&process_info->notifier_lock);
 	if (!evicted_bos)
 		return;
 
@@ -2242,9 +2819,6 @@ static void amdgpu_amdkfd_restore_userptr_worker(struct work_struct *work)
 	 * and we can just restart the queues.
 	 */
 	if (!list_empty(&process_info->userptr_inval_list)) {
-		if (atomic_read(&process_info->evicted_bos) != evicted_bos)
-			goto unlock_out; /* Concurrent eviction, try again */
-
 		if (validate_invalid_user_pages(process_info))
 			goto unlock_out;
 	}
@@ -2253,10 +2827,17 @@ static void amdgpu_amdkfd_restore_userptr_worker(struct work_struct *work)
 	 * be a first eviction that calls quiesce_mm. The eviction
 	 * reference counting inside KFD will handle this case.
 	 */
-	if (atomic_cmpxchg(&process_info->evicted_bos, evicted_bos, 0) !=
-	    evicted_bos)
-		goto unlock_out;
-	evicted_bos = 0;
+	mutex_lock(&process_info->notifier_lock);
+	if (process_info->evicted_bos != evicted_bos)
+		goto unlock_notifier_out;
+
+	if (confirm_valid_user_pages_locked(process_info)) {
+		WARN(1, "User pages unexpectedly invalid");
+		goto unlock_notifier_out;
+	}
+
+	process_info->evicted_bos = evicted_bos = 0;
+
 	if (kgd2kfd_resume_mm(mm)) {
 		pr_err("%s: Failed to resume KFD\n", __func__);
 		/* No recovery from this failure. Probably the CP is
@@ -2264,15 +2845,38 @@ static void amdgpu_amdkfd_restore_userptr_worker(struct work_struct *work)
 		 */
 	}
 
+unlock_notifier_out:
+	mutex_unlock(&process_info->notifier_lock);
 unlock_out:
 	mutex_unlock(&process_info->lock);
-	mmput(mm);
-	put_task_struct(usertask);
 
 	/* If validation failed, reschedule another attempt */
-	if (evicted_bos)
-		schedule_delayed_work(&process_info->restore_userptr_work,
+	if (evicted_bos) {
+		queue_delayed_work(system_freezable_wq,
+			&process_info->restore_userptr_work,
 			msecs_to_jiffies(AMDGPU_USERPTR_RESTORE_DELAY_MS));
+
+		kfd_smi_event_queue_restore_rescheduled(mm);
+	}
+	mmput(mm);
+	put_task_struct(usertask);
+}
+
+static void replace_eviction_fence(struct dma_fence __rcu **ef,
+				   struct dma_fence *new_ef)
+{
+	struct dma_fence *old_ef = rcu_replace_pointer(*ef, new_ef, true
+		/* protected by process_info->lock */);
+
+	/* If we're replacing an unsignaled eviction fence, that fence will
+	 * never be signaled, and if anyone is still waiting on that fence,
+	 * they will hang forever. This should never happen. We should only
+	 * replace the fence in restore_work that only gets scheduled after
+	 * eviction work signaled the fence.
+	 */
+	WARN_ONCE(!dma_fence_is_signaled(old_ef),
+		  "Replacing unsignaled eviction fence");
+	dma_fence_put(old_ef);
 }
 
 /** amdgpu_amdkfd_gpuvm_restore_process_bos - Restore all BOs for the given
@@ -2293,74 +2897,61 @@ unlock_out:
  * 7.  Add fence to all PD and PT BOs.
  * 8.  Unreserve all BOs
  */
-int amdgpu_amdkfd_gpuvm_restore_process_bos(void *info, struct dma_fence **ef)
+int amdgpu_amdkfd_gpuvm_restore_process_bos(void *info, struct dma_fence __rcu **ef)
 {
-	struct amdgpu_bo_list_entry *pd_bo_list;
 	struct amdkfd_process_info *process_info = info;
 	struct amdgpu_vm *peer_vm;
 	struct kgd_mem *mem;
-	struct bo_vm_reservation_context ctx;
-	struct amdgpu_amdkfd_fence *new_fence;
-	int ret = 0, i;
 	struct list_head duplicate_save;
 	struct amdgpu_sync sync_obj;
 	unsigned long failed_size = 0;
 	unsigned long total_size = 0;
+	struct drm_exec exec;
+	int ret;
 
 	INIT_LIST_HEAD(&duplicate_save);
-	INIT_LIST_HEAD(&ctx.list);
-	INIT_LIST_HEAD(&ctx.duplicates);
 
-	pd_bo_list = kcalloc(process_info->n_vms,
-			     sizeof(struct amdgpu_bo_list_entry),
-			     GFP_KERNEL);
-	if (!pd_bo_list)
-		return -ENOMEM;
-
-	i = 0;
 	mutex_lock(&process_info->lock);
-	list_for_each_entry(peer_vm, &process_info->vm_list_head,
-			vm_list_node)
-		amdgpu_vm_get_pd_bo(peer_vm, &ctx.list, &pd_bo_list[i++]);
 
-	/* Reserve all BOs and page tables/directory. Add all BOs from
-	 * kfd_bo_list to ctx.list
-	 */
-	list_for_each_entry(mem, &process_info->kfd_bo_list,
-			    validate_list.head) {
-
-		list_add_tail(&mem->resv_list.head, &ctx.list);
-		mem->resv_list.bo = mem->validate_list.bo;
-		mem->resv_list.num_shared = mem->validate_list.num_shared;
-	}
+	drm_exec_init(&exec, DRM_EXEC_IGNORE_DUPLICATES, 0);
+	drm_exec_until_all_locked(&exec) {
+		list_for_each_entry(peer_vm, &process_info->vm_list_head,
+				    vm_list_node) {
+			ret = amdgpu_vm_lock_pd(peer_vm, &exec, 2);
+			drm_exec_retry_on_contention(&exec);
+			if (unlikely(ret)) {
+				pr_err("Locking VM PD failed, ret: %d\n", ret);
+				goto ttm_reserve_fail;
+			}
+		}
 
-	ret = ttm_eu_reserve_buffers(&ctx.ticket, &ctx.list,
-				     false, &duplicate_save);
-	if (ret) {
-		pr_debug("Memory eviction: TTM Reserve Failed. Try again\n");
-		goto ttm_reserve_fail;
+		/* Reserve all BOs and page tables/directory. Add all BOs from
+		 * kfd_bo_list to ctx.list
+		 */
+		list_for_each_entry(mem, &process_info->kfd_bo_list,
+				    validate_list) {
+			struct drm_gem_object *gobj;
+
+			gobj = &mem->bo->tbo.base;
+			ret = drm_exec_prepare_obj(&exec, gobj, 1);
+			drm_exec_retry_on_contention(&exec);
+			if (unlikely(ret)) {
+				pr_err("drm_exec_prepare_obj failed, ret: %d\n", ret);
+				goto ttm_reserve_fail;
+			}
+		}
 	}
 
 	amdgpu_sync_create(&sync_obj);
 
-	/* Validate PDs and PTs */
-	ret = process_validate_vms(process_info);
-	if (ret)
-		goto validate_map_fail;
-
-	ret = process_sync_pds_resv(process_info, &sync_obj);
-	if (ret) {
-		pr_debug("Memory eviction: Failed to sync to PD BO moving fence. Try again\n");
-		goto validate_map_fail;
-	}
-
-	/* Validate BOs and map them to GPUVM (update VM page tables). */
+	/* Validate BOs managed by KFD */
 	list_for_each_entry(mem, &process_info->kfd_bo_list,
-			    validate_list.head) {
+			    validate_list) {
 
 		struct amdgpu_bo *bo = mem->bo;
 		uint32_t domain = mem->domain;
-		struct kfd_mem_attachment *attachment;
+		struct dma_resv_iter cursor;
+		struct dma_fence *fence;
 
 		total_size += amdgpu_bo_size(bo);
 
@@ -2375,17 +2966,39 @@ int amdgpu_amdkfd_gpuvm_restore_process_bos(void *info, struct dma_fence **ef)
 				goto validate_map_fail;
 			}
 		}
-		ret = amdgpu_sync_fence(&sync_obj, bo->tbo.moving);
-		if (ret) {
-			pr_debug("Memory eviction: Sync BO fence failed. Try again\n");
-			goto validate_map_fail;
+		dma_resv_for_each_fence(&cursor, bo->tbo.base.resv,
+					DMA_RESV_USAGE_KERNEL, fence) {
+			ret = amdgpu_sync_fence(&sync_obj, fence, GFP_KERNEL);
+			if (ret) {
+				pr_debug("Memory eviction: Sync BO fence failed. Try again\n");
+				goto validate_map_fail;
+			}
 		}
+	}
+
+	if (failed_size)
+		pr_debug("0x%lx/0x%lx in system\n", failed_size, total_size);
+
+	/* Validate PDs, PTs and evicted DMABuf imports last. Otherwise BO
+	 * validations above would invalidate DMABuf imports again.
+	 */
+	ret = process_validate_vms(process_info, &exec.ticket);
+	if (ret) {
+		pr_debug("Validating VMs failed, ret: %d\n", ret);
+		goto validate_map_fail;
+	}
+
+	/* Update mappings managed by KFD. */
+	list_for_each_entry(mem, &process_info->kfd_bo_list,
+			    validate_list) {
+		struct kfd_mem_attachment *attachment;
+
 		list_for_each_entry(attachment, &mem->attachments, list) {
 			if (!attachment->is_mapped)
 				continue;
 
 			kfd_mem_dmaunmap_attachment(mem, attachment);
-			ret = update_gpuvm_pte(mem, attachment, &sync_obj, NULL);
+			ret = update_gpuvm_pte(mem, attachment, &sync_obj);
 			if (ret) {
 				pr_debug("Memory eviction: update PTE failed. Try again\n");
 				goto validate_map_fail;
@@ -2393,8 +3006,31 @@ int amdgpu_amdkfd_gpuvm_restore_process_bos(void *info, struct dma_fence **ef)
 		}
 	}
 
-	if (failed_size)
-		pr_debug("0x%lx/0x%lx in system\n", failed_size, total_size);
+	/* Update mappings not managed by KFD */
+	list_for_each_entry(peer_vm, &process_info->vm_list_head,
+			vm_list_node) {
+		struct amdgpu_device *adev = amdgpu_ttm_adev(
+			peer_vm->root.bo->tbo.bdev);
+
+		struct amdgpu_fpriv *fpriv =
+			container_of(peer_vm, struct amdgpu_fpriv, vm);
+
+		ret = amdgpu_vm_bo_update(adev, fpriv->prt_va, false);
+		if (ret) {
+			dev_dbg(adev->dev,
+				"Memory eviction: handle PRT moved failed, pid %8d. Try again.\n",
+				pid_nr(process_info->pid));
+			goto validate_map_fail;
+		}
+
+		ret = amdgpu_vm_handle_moved(adev, peer_vm, &exec.ticket);
+		if (ret) {
+			dev_dbg(adev->dev,
+				"Memory eviction: handle moved failed, pid %8d. Try again.\n",
+				pid_nr(process_info->pid));
+			goto validate_map_fail;
+		}
+	}
 
 	/* Update page directories */
 	ret = process_update_pds(process_info, &sync_obj);
@@ -2403,46 +3039,72 @@ int amdgpu_amdkfd_gpuvm_restore_process_bos(void *info, struct dma_fence **ef)
 		goto validate_map_fail;
 	}
 
+	/* Sync with fences on all the page tables. They implicitly depend on any
+	 * move fences from amdgpu_vm_handle_moved above.
+	 */
+	ret = process_sync_pds_resv(process_info, &sync_obj);
+	if (ret) {
+		pr_debug("Memory eviction: Failed to sync to PD BO moving fence. Try again\n");
+		goto validate_map_fail;
+	}
+
 	/* Wait for validate and PT updates to finish */
 	amdgpu_sync_wait(&sync_obj, false);
 
-	/* Release old eviction fence and create new one, because fence only
-	 * goes from unsignaled to signaled, fence cannot be reused.
-	 * Use context and mm from the old fence.
+	/* The old eviction fence may be unsignaled if restore happens
+	 * after a GPU reset or suspend/resume. Keep the old fence in that
+	 * case. Otherwise release the old eviction fence and create new
+	 * one, because fence only goes from unsignaled to signaled once
+	 * and cannot be reused. Use context and mm from the old fence.
+	 *
+	 * If an old eviction fence signals after this check, that's OK.
+	 * Anyone signaling an eviction fence must stop the queues first
+	 * and schedule another restore worker.
 	 */
-	new_fence = amdgpu_amdkfd_fence_create(
+	if (dma_fence_is_signaled(&process_info->eviction_fence->base)) {
+		struct amdgpu_amdkfd_fence *new_fence =
+			amdgpu_amdkfd_fence_create(
 				process_info->eviction_fence->base.context,
 				process_info->eviction_fence->mm,
 				NULL);
-	if (!new_fence) {
-		pr_err("Failed to create eviction fence\n");
-		ret = -ENOMEM;
-		goto validate_map_fail;
+
+		if (!new_fence) {
+			pr_err("Failed to create eviction fence\n");
+			ret = -ENOMEM;
+			goto validate_map_fail;
+		}
+		dma_fence_put(&process_info->eviction_fence->base);
+		process_info->eviction_fence = new_fence;
+		replace_eviction_fence(ef, dma_fence_get(&new_fence->base));
+	} else {
+		WARN_ONCE(*ef != &process_info->eviction_fence->base,
+			  "KFD eviction fence doesn't match KGD process_info");
 	}
-	dma_fence_put(&process_info->eviction_fence->base);
-	process_info->eviction_fence = new_fence;
-	*ef = dma_fence_get(&new_fence->base);
 
-	/* Attach new eviction fence to all BOs */
-	list_for_each_entry(mem, &process_info->kfd_bo_list,
-		validate_list.head)
-		amdgpu_bo_fence(mem->bo,
-			&process_info->eviction_fence->base, true);
+	/* Attach new eviction fence to all BOs except pinned ones */
+	list_for_each_entry(mem, &process_info->kfd_bo_list, validate_list) {
+		if (mem->bo->tbo.pin_count)
+			continue;
 
-	/* Attach eviction fence to PD / PT BOs */
+		dma_resv_add_fence(mem->bo->tbo.base.resv,
+				   &process_info->eviction_fence->base,
+				   DMA_RESV_USAGE_BOOKKEEP);
+	}
+	/* Attach eviction fence to PD / PT BOs and DMABuf imports */
 	list_for_each_entry(peer_vm, &process_info->vm_list_head,
 			    vm_list_node) {
 		struct amdgpu_bo *bo = peer_vm->root.bo;
 
-		amdgpu_bo_fence(bo, &process_info->eviction_fence->base, true);
+		dma_resv_add_fence(bo->tbo.base.resv,
+				   &process_info->eviction_fence->base,
+				   DMA_RESV_USAGE_BOOKKEEP);
 	}
 
 validate_map_fail:
-	ttm_eu_backoff_reservation(&ctx.ticket, &ctx.list);
 	amdgpu_sync_free(&sync_obj);
 ttm_reserve_fail:
+	drm_exec_fini(&exec);
 	mutex_unlock(&process_info->lock);
-	kfree(pd_bo_list);
 	return ret;
 }
 
@@ -2485,10 +3147,12 @@ int amdgpu_amdkfd_add_gws_to_process(void *info, void *gws, struct kgd_mem **mem
 	 * Add process eviction fence to bo so they can
 	 * evict each other.
 	 */
-	ret = dma_resv_reserve_shared(gws_bo->tbo.base.resv, 1);
+	ret = dma_resv_reserve_fences(gws_bo->tbo.base.resv, 1);
 	if (ret)
 		goto reserve_shared_fail;
-	amdgpu_bo_fence(gws_bo, &process_info->eviction_fence->base, true);
+	dma_resv_add_fence(gws_bo->tbo.base.resv,
+			   &process_info->eviction_fence->base,
+			   DMA_RESV_USAGE_BOOKKEEP);
 	amdgpu_bo_unreserve(gws_bo);
 	mutex_unlock(&(*mem)->process_info->lock);
 
@@ -2537,11 +3201,9 @@ int amdgpu_amdkfd_remove_gws_from_process(void *info, void *mem)
 }
 
 /* Returns GPU-specific tiling mode information */
-int amdgpu_amdkfd_get_tile_config(struct kgd_dev *kgd,
+int amdgpu_amdkfd_get_tile_config(struct amdgpu_device *adev,
 				struct tile_config *config)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
-
 	config->gb_addr_config = adev->gfx.config.gb_addr_config;
 	config->tile_config_ptr = adev->gfx.config.tile_mode_array;
 	config->num_tile_configs =
@@ -2557,3 +3219,34 @@ int amdgpu_amdkfd_get_tile_config(struct kgd_dev *kgd,
 
 	return 0;
 }
+
+bool amdgpu_amdkfd_bo_mapped_to_dev(void *drm_priv, struct kgd_mem *mem)
+{
+	struct amdgpu_vm *vm = drm_priv_to_vm(drm_priv);
+	struct kfd_mem_attachment *entry;
+
+	list_for_each_entry(entry, &mem->attachments, list) {
+		if (entry->is_mapped && entry->bo_va->base.vm == vm)
+			return true;
+	}
+	return false;
+}
+
+#if defined(CONFIG_DEBUG_FS)
+
+int kfd_debugfs_kfd_mem_limits(struct seq_file *m, void *data)
+{
+
+	spin_lock(&kfd_mem_limit.mem_limit_lock);
+	seq_printf(m, "System mem used %lldM out of %lluM\n",
+		  (kfd_mem_limit.system_mem_used >> 20),
+		  (kfd_mem_limit.max_system_mem_limit >> 20));
+	seq_printf(m, "TTM mem used %lldM out of %lluM\n",
+		  (kfd_mem_limit.ttm_mem_used >> 20),
+		  (kfd_mem_limit.max_ttm_mem_limit >> 20));
+	spin_unlock(&kfd_mem_limit.mem_limit_lock);
+
+	return 0;
+}
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c
index 12a6b1c99c93..763f2b8dcf13 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c
@@ -36,13 +36,6 @@
 #include "atombios_encoders.h"
 #include "bif/bif_4_1_d.h"
 
-static void amdgpu_atombios_lookup_i2c_gpio_quirks(struct amdgpu_device *adev,
-					  ATOM_GPIO_I2C_ASSIGMENT *gpio,
-					  u8 index)
-{
-
-}
-
 static struct amdgpu_i2c_bus_rec amdgpu_atombios_get_bus_rec_for_i2c_gpio(ATOM_GPIO_I2C_ASSIGMENT *gpio)
 {
 	struct amdgpu_i2c_bus_rec i2c;
@@ -108,9 +101,6 @@ struct amdgpu_i2c_bus_rec amdgpu_atombios_lookup_i2c_gpio(struct amdgpu_device *
 
 		gpio = &i2c_info->asGPIO_Info[0];
 		for (i = 0; i < num_indices; i++) {
-
-			amdgpu_atombios_lookup_i2c_gpio_quirks(adev, gpio, i);
-
 			if (gpio->sucI2cId.ucAccess == id) {
 				i2c = amdgpu_atombios_get_bus_rec_for_i2c_gpio(gpio);
 				break;
@@ -142,8 +132,6 @@ void amdgpu_atombios_i2c_init(struct amdgpu_device *adev)
 
 		gpio = &i2c_info->asGPIO_Info[0];
 		for (i = 0; i < num_indices; i++) {
-			amdgpu_atombios_lookup_i2c_gpio_quirks(adev, gpio, i);
-
 			i2c = amdgpu_atombios_get_bus_rec_for_i2c_gpio(gpio);
 
 			if (i2c.valid) {
@@ -156,6 +144,38 @@ void amdgpu_atombios_i2c_init(struct amdgpu_device *adev)
 	}
 }
 
+void amdgpu_atombios_oem_i2c_init(struct amdgpu_device *adev, u8 i2c_id)
+{
+	struct atom_context *ctx = adev->mode_info.atom_context;
+	ATOM_GPIO_I2C_ASSIGMENT *gpio;
+	struct amdgpu_i2c_bus_rec i2c;
+	int index = GetIndexIntoMasterTable(DATA, GPIO_I2C_Info);
+	struct _ATOM_GPIO_I2C_INFO *i2c_info;
+	uint16_t data_offset, size;
+	int i, num_indices;
+	char stmp[32];
+
+	if (amdgpu_atom_parse_data_header(ctx, index, &size, NULL, NULL, &data_offset)) {
+		i2c_info = (struct _ATOM_GPIO_I2C_INFO *)(ctx->bios + data_offset);
+
+		num_indices = (size - sizeof(ATOM_COMMON_TABLE_HEADER)) /
+			sizeof(ATOM_GPIO_I2C_ASSIGMENT);
+
+		gpio = &i2c_info->asGPIO_Info[0];
+		for (i = 0; i < num_indices; i++) {
+			i2c = amdgpu_atombios_get_bus_rec_for_i2c_gpio(gpio);
+
+			if (i2c.valid && i2c.i2c_id == i2c_id) {
+				sprintf(stmp, "OEM 0x%x", i2c.i2c_id);
+				adev->i2c_bus[i] = amdgpu_i2c_create(adev_to_drm(adev), &i2c, stmp);
+				break;
+			}
+			gpio = (ATOM_GPIO_I2C_ASSIGMENT *)
+				((u8 *)gpio + sizeof(ATOM_GPIO_I2C_ASSIGMENT));
+		}
+	}
+}
+
 struct amdgpu_gpio_rec
 amdgpu_atombios_lookup_gpio(struct amdgpu_device *adev,
 			    u8 id)
@@ -686,7 +706,6 @@ int amdgpu_atombios_get_clock_info(struct amdgpu_device *adev)
 		}
 		adev->clock.dp_extclk =
 			le16_to_cpu(firmware_info->info_21.usUniphyDPModeExtClkFreq);
-		adev->clock.current_dispclk = adev->clock.default_dispclk;
 
 		adev->clock.max_pixel_clock = le16_to_cpu(firmware_info->info.usMaxPixelClock);
 		if (adev->clock.max_pixel_clock == 0)
@@ -1018,7 +1037,9 @@ int amdgpu_atombios_get_clock_dividers(struct amdgpu_device *adev,
 		if (clock_type == COMPUTE_ENGINE_PLL_PARAM) {
 			args.v3.ulClockParams = cpu_to_le32((clock_type << 24) | clock);
 
-			amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+			if (amdgpu_atom_execute_table(adev->mode_info.atom_context,
+			    index, (uint32_t *)&args, sizeof(args)))
+				return -EINVAL;
 
 			dividers->post_div = args.v3.ucPostDiv;
 			dividers->enable_post_div = (args.v3.ucCntlFlag &
@@ -1038,7 +1059,9 @@ int amdgpu_atombios_get_clock_dividers(struct amdgpu_device *adev,
 			if (strobe_mode)
 				args.v5.ucInputFlag = ATOM_PLL_INPUT_FLAG_PLL_STROBE_MODE_EN;
 
-			amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+			if (amdgpu_atom_execute_table(adev->mode_info.atom_context,
+			    index, (uint32_t *)&args, sizeof(args)))
+				return -EINVAL;
 
 			dividers->post_div = args.v5.ucPostDiv;
 			dividers->enable_post_div = (args.v5.ucCntlFlag &
@@ -1056,7 +1079,9 @@ int amdgpu_atombios_get_clock_dividers(struct amdgpu_device *adev,
 		/* fusion */
 		args.v4.ulClock = cpu_to_le32(clock);	/* 10 khz */
 
-		amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+		if (amdgpu_atom_execute_table(adev->mode_info.atom_context,
+		    index, (uint32_t *)&args, sizeof(args)))
+			return -EINVAL;
 
 		dividers->post_divider = dividers->post_div = args.v4.ucPostDiv;
 		dividers->real_clock = le32_to_cpu(args.v4.ulClock);
@@ -1067,7 +1092,9 @@ int amdgpu_atombios_get_clock_dividers(struct amdgpu_device *adev,
 		args.v6_in.ulClock.ulComputeClockFlag = clock_type;
 		args.v6_in.ulClock.ulClockFreq = cpu_to_le32(clock);	/* 10 khz */
 
-		amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+		if (amdgpu_atom_execute_table(adev->mode_info.atom_context,
+		    index, (uint32_t *)&args, sizeof(args)))
+			return -EINVAL;
 
 		dividers->whole_fb_div = le16_to_cpu(args.v6_out.ulFbDiv.usFbDiv);
 		dividers->frac_fb_div = le16_to_cpu(args.v6_out.ulFbDiv.usFbDivFrac);
@@ -1083,6 +1110,7 @@ int amdgpu_atombios_get_clock_dividers(struct amdgpu_device *adev,
 	return 0;
 }
 
+#ifdef CONFIG_DRM_AMDGPU_SI
 int amdgpu_atombios_get_memory_pll_dividers(struct amdgpu_device *adev,
 					    u32 clock,
 					    bool strobe_mode,
@@ -1108,7 +1136,9 @@ int amdgpu_atombios_get_memory_pll_dividers(struct amdgpu_device *adev,
 			if (strobe_mode)
 				args.ucInputFlag |= MPLL_INPUT_FLAG_STROBE_MODE_EN;
 
-			amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+			if (amdgpu_atom_execute_table(adev->mode_info.atom_context,
+			    index, (uint32_t *)&args, sizeof(args)))
+				return -EINVAL;
 
 			mpll_param->clkfrac = le16_to_cpu(args.ulFbDiv.usFbDivFrac);
 			mpll_param->clkf = le16_to_cpu(args.ulFbDiv.usFbDiv);
@@ -1134,8 +1164,8 @@ int amdgpu_atombios_get_memory_pll_dividers(struct amdgpu_device *adev,
 	return 0;
 }
 
-void amdgpu_atombios_set_engine_dram_timings(struct amdgpu_device *adev,
-					     u32 eng_clock, u32 mem_clock)
+int amdgpu_atombios_set_engine_dram_timings(struct amdgpu_device *adev,
+					    u32 eng_clock, u32 mem_clock)
 {
 	SET_ENGINE_CLOCK_PS_ALLOCATION args;
 	int index = GetIndexIntoMasterTable(COMMAND, DynamicMemorySettings);
@@ -1150,7 +1180,8 @@ void amdgpu_atombios_set_engine_dram_timings(struct amdgpu_device *adev,
 	if (mem_clock)
 		args.sReserved.ulClock = cpu_to_le32(mem_clock & SET_CLOCK_FREQ_MASK);
 
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	return amdgpu_atom_execute_table(adev->mode_info.atom_context, index,
+					 (uint32_t *)&args, sizeof(args));
 }
 
 void amdgpu_atombios_get_default_voltages(struct amdgpu_device *adev,
@@ -1204,7 +1235,9 @@ int amdgpu_atombios_get_max_vddc(struct amdgpu_device *adev, u8 voltage_type,
 		args.v2.ucVoltageMode = 0;
 		args.v2.usVoltageLevel = 0;
 
-		amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+		if (amdgpu_atom_execute_table(adev->mode_info.atom_context,
+		    index, (uint32_t *)&args, sizeof(args)))
+			return -EINVAL;
 
 		*voltage = le16_to_cpu(args.v2.usVoltageLevel);
 		break;
@@ -1213,7 +1246,9 @@ int amdgpu_atombios_get_max_vddc(struct amdgpu_device *adev, u8 voltage_type,
 		args.v3.ucVoltageMode = ATOM_GET_VOLTAGE_LEVEL;
 		args.v3.usVoltageLevel = cpu_to_le16(voltage_id);
 
-		amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+		if (amdgpu_atom_execute_table(adev->mode_info.atom_context,
+		    index, (uint32_t *)&args, sizeof(args)))
+			return -EINVAL;
 
 		*voltage = le16_to_cpu(args.v3.usVoltageLevel);
 		break;
@@ -1475,6 +1510,8 @@ int amdgpu_atombios_init_mc_reg_table(struct amdgpu_device *adev,
 										(u32)le32_to_cpu(*((u32 *)reg_data + j));
 									j++;
 								} else if ((reg_table->mc_reg_address[i].pre_reg_data & LOW_NIBBLE_MASK) == DATA_EQU_PREV) {
+									if (i == 0)
+										continue;
 									reg_table->mc_reg_table_entry[num_ranges].mc_data[i] =
 										reg_table->mc_reg_table_entry[num_ranges].mc_data[i - 1];
 								}
@@ -1503,6 +1540,7 @@ int amdgpu_atombios_init_mc_reg_table(struct amdgpu_device *adev,
 	}
 	return -EINVAL;
 }
+#endif
 
 bool amdgpu_atombios_has_gpu_virtualization_table(struct amdgpu_device *adev)
 {
@@ -1774,21 +1812,57 @@ static ssize_t amdgpu_atombios_get_vbios_version(struct device *dev,
 	struct amdgpu_device *adev = drm_to_adev(ddev);
 	struct atom_context *ctx = adev->mode_info.atom_context;
 
-	return sysfs_emit(buf, "%s\n", ctx->vbios_version);
+	return sysfs_emit(buf, "%s\n", ctx->vbios_pn);
+}
+
+static ssize_t amdgpu_atombios_get_vbios_build(struct device *dev,
+					       struct device_attribute *attr,
+					       char *buf)
+{
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = drm_to_adev(ddev);
+	struct atom_context *ctx = adev->mode_info.atom_context;
+
+	return sysfs_emit(buf, "%s\n", ctx->build_num);
 }
 
 static DEVICE_ATTR(vbios_version, 0444, amdgpu_atombios_get_vbios_version,
 		   NULL);
+static DEVICE_ATTR(vbios_build, 0444, amdgpu_atombios_get_vbios_build, NULL);
 
 static struct attribute *amdgpu_vbios_version_attrs[] = {
-	&dev_attr_vbios_version.attr,
-	NULL
+	&dev_attr_vbios_version.attr, &dev_attr_vbios_build.attr, NULL
 };
 
+static umode_t amdgpu_vbios_version_attrs_is_visible(struct kobject *kobj,
+						     struct attribute *attr,
+						     int index)
+{
+	struct device *dev = kobj_to_dev(kobj);
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = drm_to_adev(ddev);
+	struct atom_context *ctx = adev->mode_info.atom_context;
+
+	if (attr == &dev_attr_vbios_build.attr && !strlen(ctx->build_num))
+		return 0;
+
+	return attr->mode;
+}
+
 const struct attribute_group amdgpu_vbios_version_attr_group = {
-	.attrs = amdgpu_vbios_version_attrs
+	.attrs = amdgpu_vbios_version_attrs,
+	.is_visible = amdgpu_vbios_version_attrs_is_visible,
 };
 
+int amdgpu_atombios_sysfs_init(struct amdgpu_device *adev)
+{
+	if (adev->mode_info.atom_context)
+		return devm_device_add_group(adev->dev,
+					     &amdgpu_vbios_version_attr_group);
+
+	return 0;
+}
+
 /**
  * amdgpu_atombios_fini - free the driver info and callbacks for atombios
  *
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.h
index 27e74b1fc260..867bc5c5ce67 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.h
@@ -89,8 +89,7 @@ struct atom_memory_info {
 
 #define MAX_AC_TIMING_ENTRIES 16
 
-struct atom_memory_clock_range_table
-{
+struct atom_memory_clock_range_table {
 	u8 num_entries;
 	u8 rsv[3];
 	u32 mclk[MAX_AC_TIMING_ENTRIES];
@@ -118,14 +117,12 @@ struct atom_mc_reg_table {
 
 #define MAX_VOLTAGE_ENTRIES 32
 
-struct atom_voltage_table_entry
-{
+struct atom_voltage_table_entry {
 	u16 value;
 	u32 smio_low;
 };
 
-struct atom_voltage_table
-{
+struct atom_voltage_table {
 	u32 count;
 	u32 mask_low;
 	u32 phase_delay;
@@ -139,6 +136,7 @@ amdgpu_atombios_lookup_gpio(struct amdgpu_device *adev,
 struct amdgpu_i2c_bus_rec amdgpu_atombios_lookup_i2c_gpio(struct amdgpu_device *adev,
 							  uint8_t id);
 void amdgpu_atombios_i2c_init(struct amdgpu_device *adev);
+void amdgpu_atombios_oem_i2c_init(struct amdgpu_device *adev, u8 i2c_id);
 
 bool amdgpu_atombios_has_dce_engine_info(struct amdgpu_device *adev);
 
@@ -160,13 +158,14 @@ int amdgpu_atombios_get_clock_dividers(struct amdgpu_device *adev,
 				       bool strobe_mode,
 				       struct atom_clock_dividers *dividers);
 
+#ifdef CONFIG_DRM_AMDGPU_SI
 int amdgpu_atombios_get_memory_pll_dividers(struct amdgpu_device *adev,
 					    u32 clock,
 					    bool strobe_mode,
 					    struct atom_mpll_param *mpll_param);
 
-void amdgpu_atombios_set_engine_dram_timings(struct amdgpu_device *adev,
-					     u32 eng_clock, u32 mem_clock);
+int amdgpu_atombios_set_engine_dram_timings(struct amdgpu_device *adev,
+					    u32 eng_clock, u32 mem_clock);
 
 bool
 amdgpu_atombios_is_voltage_gpio(struct amdgpu_device *adev,
@@ -179,6 +178,17 @@ int amdgpu_atombios_get_voltage_table(struct amdgpu_device *adev,
 int amdgpu_atombios_init_mc_reg_table(struct amdgpu_device *adev,
 				      u8 module_index,
 				      struct atom_mc_reg_table *reg_table);
+int amdgpu_atombios_get_max_vddc(struct amdgpu_device *adev, u8 voltage_type,
+			     u16 voltage_id, u16 *voltage);
+int amdgpu_atombios_get_leakage_vddc_based_on_leakage_idx(struct amdgpu_device *adev,
+						      u16 *voltage,
+						      u16 leakage_idx);
+void amdgpu_atombios_get_default_voltages(struct amdgpu_device *adev,
+					  u16 *vddc, u16 *vddci, u16 *mvdd);
+int amdgpu_atombios_get_svi2_info(struct amdgpu_device *adev,
+			      u8 voltage_type,
+			      u8 *svd_gpio_id, u8 *svc_gpio_id);
+#endif
 
 bool amdgpu_atombios_has_gpu_virtualization_table(struct amdgpu_device *adev);
 
@@ -190,21 +200,11 @@ void amdgpu_atombios_scratch_regs_set_backlight_level(struct amdgpu_device *adev
 bool amdgpu_atombios_scratch_need_asic_init(struct amdgpu_device *adev);
 
 void amdgpu_atombios_copy_swap(u8 *dst, u8 *src, u8 num_bytes, bool to_le);
-int amdgpu_atombios_get_max_vddc(struct amdgpu_device *adev, u8 voltage_type,
-			     u16 voltage_id, u16 *voltage);
-int amdgpu_atombios_get_leakage_vddc_based_on_leakage_idx(struct amdgpu_device *adev,
-						      u16 *voltage,
-						      u16 leakage_idx);
-void amdgpu_atombios_get_default_voltages(struct amdgpu_device *adev,
-					  u16 *vddc, u16 *vddci, u16 *mvdd);
 int amdgpu_atombios_get_clock_dividers(struct amdgpu_device *adev,
 				       u8 clock_type,
 				       u32 clock,
 				       bool strobe_mode,
 				       struct atom_clock_dividers *dividers);
-int amdgpu_atombios_get_svi2_info(struct amdgpu_device *adev,
-			      u8 voltage_type,
-			      u8 *svd_gpio_id, u8 *svc_gpio_id);
 
 int amdgpu_atombios_get_data_table(struct amdgpu_device *adev,
 				   uint32_t table,
@@ -215,5 +215,6 @@ int amdgpu_atombios_get_data_table(struct amdgpu_device *adev,
 
 void amdgpu_atombios_fini(struct amdgpu_device *adev);
 int amdgpu_atombios_init(struct amdgpu_device *adev);
+int amdgpu_atombios_sysfs_init(struct amdgpu_device *adev);
 
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c
index 97178b307ed6..636385c80f64 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c
@@ -34,6 +34,7 @@ union firmware_info {
 	struct atom_firmware_info_v3_2 v32;
 	struct atom_firmware_info_v3_3 v33;
 	struct atom_firmware_info_v3_4 v34;
+	struct atom_firmware_info_v3_5 v35;
 };
 
 /*
@@ -58,7 +59,7 @@ uint32_t amdgpu_atomfirmware_query_firmware_capability(struct amdgpu_device *ade
 	if (amdgpu_atom_parse_data_header(adev->mode_info.atom_context,
 				index, &size, &frev, &crev, &data_offset)) {
 		/* support firmware_info 3.1 + */
-		if ((frev == 3 && crev >=1) || (frev > 3)) {
+		if ((frev == 3 && crev >= 1) || (frev > 3)) {
 			firmware_info = (union firmware_info *)
 				(mode_info->atom_context->bios + data_offset);
 			fw_cap = le32_to_cpu(firmware_info->v31.firmware_capability);
@@ -101,39 +102,104 @@ void amdgpu_atomfirmware_scratch_regs_init(struct amdgpu_device *adev)
 	}
 }
 
+static int amdgpu_atomfirmware_allocate_fb_v2_1(struct amdgpu_device *adev,
+	struct vram_usagebyfirmware_v2_1 *fw_usage, int *usage_bytes)
+{
+	u32 start_addr, fw_size, drv_size;
+
+	start_addr = le32_to_cpu(fw_usage->start_address_in_kb);
+	fw_size = le16_to_cpu(fw_usage->used_by_firmware_in_kb);
+	drv_size = le16_to_cpu(fw_usage->used_by_driver_in_kb);
+
+	DRM_DEBUG("atom firmware v2_1 requested %08x %dkb fw %dkb drv\n",
+			  start_addr,
+			  fw_size,
+			  drv_size);
+
+	if ((start_addr & ATOM_VRAM_OPERATION_FLAGS_MASK) ==
+		(u32)(ATOM_VRAM_BLOCK_SRIOV_MSG_SHARE_RESERVATION <<
+		ATOM_VRAM_OPERATION_FLAGS_SHIFT)) {
+		/* Firmware request VRAM reservation for SR-IOV */
+		adev->mman.fw_vram_usage_start_offset = (start_addr &
+			(~ATOM_VRAM_OPERATION_FLAGS_MASK)) << 10;
+		adev->mman.fw_vram_usage_size = fw_size << 10;
+		/* Use the default scratch size */
+		*usage_bytes = 0;
+	} else {
+		*usage_bytes = drv_size << 10;
+	}
+	return 0;
+}
+
+static int amdgpu_atomfirmware_allocate_fb_v2_2(struct amdgpu_device *adev,
+		struct vram_usagebyfirmware_v2_2 *fw_usage, int *usage_bytes)
+{
+	u32 fw_start_addr, fw_size, drv_start_addr, drv_size;
+
+	fw_start_addr = le32_to_cpu(fw_usage->fw_region_start_address_in_kb);
+	fw_size = le16_to_cpu(fw_usage->used_by_firmware_in_kb);
+
+	drv_start_addr = le32_to_cpu(fw_usage->driver_region0_start_address_in_kb);
+	drv_size = le32_to_cpu(fw_usage->used_by_driver_region0_in_kb);
+
+	DRM_DEBUG("atom requested fw start at %08x %dkb and drv start at %08x %dkb\n",
+			  fw_start_addr,
+			  fw_size,
+			  drv_start_addr,
+			  drv_size);
+
+	if (amdgpu_sriov_vf(adev) &&
+	    ((fw_start_addr & (ATOM_VRAM_BLOCK_NEEDS_NO_RESERVATION <<
+		ATOM_VRAM_OPERATION_FLAGS_SHIFT)) == 0)) {
+		/* Firmware request VRAM reservation for SR-IOV */
+		adev->mman.fw_vram_usage_start_offset = (fw_start_addr &
+			(~ATOM_VRAM_OPERATION_FLAGS_MASK)) << 10;
+		adev->mman.fw_vram_usage_size = fw_size << 10;
+	}
+
+	if (amdgpu_sriov_vf(adev) &&
+	    ((drv_start_addr & (ATOM_VRAM_BLOCK_NEEDS_NO_RESERVATION <<
+		ATOM_VRAM_OPERATION_FLAGS_SHIFT)) == 0)) {
+		/* driver request VRAM reservation for SR-IOV */
+		adev->mman.drv_vram_usage_start_offset = (drv_start_addr &
+			(~ATOM_VRAM_OPERATION_FLAGS_MASK)) << 10;
+		adev->mman.drv_vram_usage_size = drv_size << 10;
+	}
+
+	*usage_bytes = 0;
+	return 0;
+}
+
 int amdgpu_atomfirmware_allocate_fb_scratch(struct amdgpu_device *adev)
 {
 	struct atom_context *ctx = adev->mode_info.atom_context;
 	int index = get_index_into_master_table(atom_master_list_of_data_tables_v2_1,
 						vram_usagebyfirmware);
-	struct vram_usagebyfirmware_v2_1 *firmware_usage;
-	uint32_t start_addr, size;
-	uint16_t data_offset;
+	struct vram_usagebyfirmware_v2_1 *fw_usage_v2_1;
+	struct vram_usagebyfirmware_v2_2 *fw_usage_v2_2;
+	u16 data_offset;
+	u8 frev, crev;
 	int usage_bytes = 0;
 
-	if (amdgpu_atom_parse_data_header(ctx, index, NULL, NULL, NULL, &data_offset)) {
-		firmware_usage = (struct vram_usagebyfirmware_v2_1 *)(ctx->bios + data_offset);
-		DRM_DEBUG("atom firmware requested %08x %dkb fw %dkb drv\n",
-			  le32_to_cpu(firmware_usage->start_address_in_kb),
-			  le16_to_cpu(firmware_usage->used_by_firmware_in_kb),
-			  le16_to_cpu(firmware_usage->used_by_driver_in_kb));
-
-		start_addr = le32_to_cpu(firmware_usage->start_address_in_kb);
-		size = le16_to_cpu(firmware_usage->used_by_firmware_in_kb);
-
-		if ((uint32_t)(start_addr & ATOM_VRAM_OPERATION_FLAGS_MASK) ==
-			(uint32_t)(ATOM_VRAM_BLOCK_SRIOV_MSG_SHARE_RESERVATION <<
-			ATOM_VRAM_OPERATION_FLAGS_SHIFT)) {
-			/* Firmware request VRAM reservation for SR-IOV */
-			adev->mman.fw_vram_usage_start_offset = (start_addr &
-				(~ATOM_VRAM_OPERATION_FLAGS_MASK)) << 10;
-			adev->mman.fw_vram_usage_size = size << 10;
-			/* Use the default scratch size */
-			usage_bytes = 0;
-		} else {
-			usage_bytes = le16_to_cpu(firmware_usage->used_by_driver_in_kb) << 10;
+	/* Skip atomfirmware allocation for SRIOV VFs when dynamic crit regn is enabled */
+	if (!(amdgpu_sriov_vf(adev) && adev->virt.is_dynamic_crit_regn_enabled)) {
+		if (amdgpu_atom_parse_data_header(ctx, index, NULL, &frev, &crev, &data_offset)) {
+			if (frev == 2 && crev == 1) {
+				fw_usage_v2_1 =
+					(struct vram_usagebyfirmware_v2_1 *)(ctx->bios + data_offset);
+				amdgpu_atomfirmware_allocate_fb_v2_1(adev,
+						fw_usage_v2_1,
+						&usage_bytes);
+			} else if (frev >= 2 && crev >= 2) {
+				fw_usage_v2_2 =
+					(struct vram_usagebyfirmware_v2_2 *)(ctx->bios + data_offset);
+				amdgpu_atomfirmware_allocate_fb_v2_2(adev,
+						fw_usage_v2_2,
+						&usage_bytes);
+			}
 		}
 	}
+
 	ctx->scratch_size_bytes = 0;
 	if (usage_bytes == 0)
 		usage_bytes = 20 * 1024;
@@ -149,12 +215,14 @@ union igp_info {
 	struct atom_integrated_system_info_v1_11 v11;
 	struct atom_integrated_system_info_v1_12 v12;
 	struct atom_integrated_system_info_v2_1 v21;
+	struct atom_integrated_system_info_v2_3 v23;
 };
 
 union umc_info {
 	struct atom_umc_info_v3_1 v31;
 	struct atom_umc_info_v3_2 v32;
 	struct atom_umc_info_v3_3 v33;
+	struct atom_umc_info_v4_0 v40;
 };
 
 union vram_info {
@@ -162,12 +230,14 @@ union vram_info {
 	struct atom_vram_info_header_v2_4 v24;
 	struct atom_vram_info_header_v2_5 v25;
 	struct atom_vram_info_header_v2_6 v26;
+	struct atom_vram_info_header_v3_0 v30;
 };
 
 union vram_module {
 	struct atom_vram_module_v9 v9;
 	struct atom_vram_module_v10 v10;
 	struct atom_vram_module_v11 v11;
+	struct atom_vram_module_v3_0 v30;
 };
 
 static int convert_atom_mem_type_to_vram_type(struct amdgpu_device *adev,
@@ -186,13 +256,17 @@ static int convert_atom_mem_type_to_vram_type(struct amdgpu_device *adev,
 			vram_type = AMDGPU_VRAM_TYPE_DDR3;
 			break;
 		case Ddr4MemType:
-		case LpDdr4MemType:
 			vram_type = AMDGPU_VRAM_TYPE_DDR4;
 			break;
+		case LpDdr4MemType:
+			vram_type = AMDGPU_VRAM_TYPE_LPDDR4;
+			break;
 		case Ddr5MemType:
-		case LpDdr5MemType:
 			vram_type = AMDGPU_VRAM_TYPE_DDR5;
 			break;
+		case LpDdr5MemType:
+			vram_type = AMDGPU_VRAM_TYPE_LPDDR5;
+			break;
 		default:
 			vram_type = AMDGPU_VRAM_TYPE_UNKNOWN;
 			break;
@@ -204,11 +278,15 @@ static int convert_atom_mem_type_to_vram_type(struct amdgpu_device *adev,
 			break;
 		case ATOM_DGPU_VRAM_TYPE_HBM2:
 		case ATOM_DGPU_VRAM_TYPE_HBM2E:
+		case ATOM_DGPU_VRAM_TYPE_HBM3:
 			vram_type = AMDGPU_VRAM_TYPE_HBM;
 			break;
 		case ATOM_DGPU_VRAM_TYPE_GDDR6:
 			vram_type = AMDGPU_VRAM_TYPE_GDDR6;
 			break;
+		case ATOM_DGPU_VRAM_TYPE_HBM3E:
+			vram_type = AMDGPU_VRAM_TYPE_HBM3E;
+			break;
 		default:
 			vram_type = AMDGPU_VRAM_TYPE_UNKNOWN;
 			break;
@@ -218,7 +296,6 @@ static int convert_atom_mem_type_to_vram_type(struct amdgpu_device *adev,
 	return vram_type;
 }
 
-
 int
 amdgpu_atomfirmware_get_vram_info(struct amdgpu_device *adev,
 				  int *vram_width, int *vram_type,
@@ -229,6 +306,7 @@ amdgpu_atomfirmware_get_vram_info(struct amdgpu_device *adev,
 	u16 data_offset, size;
 	union igp_info *igp_info;
 	union vram_info *vram_info;
+	union umc_info *umc_info;
 	union vram_module *vram_module;
 	u8 frev, crev;
 	u8 mem_type;
@@ -240,10 +318,16 @@ amdgpu_atomfirmware_get_vram_info(struct amdgpu_device *adev,
 	if (adev->flags & AMD_IS_APU)
 		index = get_index_into_master_table(atom_master_list_of_data_tables_v2_1,
 						    integratedsysteminfo);
-	else
-		index = get_index_into_master_table(atom_master_list_of_data_tables_v2_1,
-						    vram_info);
-
+	else {
+		switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+		case IP_VERSION(12, 0, 0):
+		case IP_VERSION(12, 0, 1):
+			index = get_index_into_master_table(atom_master_list_of_data_tables_v2_1, umc_info);
+			break;
+		default:
+			index = get_index_into_master_table(atom_master_list_of_data_tables_v2_1, vram_info);
+		}
+	}
 	if (amdgpu_atom_parse_data_header(mode_info->atom_context,
 					  index, &size,
 					  &frev, &crev, &data_offset)) {
@@ -258,10 +342,13 @@ amdgpu_atomfirmware_get_vram_info(struct amdgpu_device *adev,
 					mem_channel_number = igp_info->v11.umachannelnumber;
 					if (!mem_channel_number)
 						mem_channel_number = 1;
-					/* channel width is 64 */
-					if (vram_width)
-						*vram_width = mem_channel_number * 64;
 					mem_type = igp_info->v11.memorytype;
+					if (mem_type == LpDdr5MemType)
+						mem_channel_width = 32;
+					else
+						mem_channel_width = 64;
+					if (vram_width)
+						*vram_width = mem_channel_number * mem_channel_width;
 					if (vram_type)
 						*vram_type = convert_atom_mem_type_to_vram_type(adev, mem_type);
 					break;
@@ -276,10 +363,27 @@ amdgpu_atomfirmware_get_vram_info(struct amdgpu_device *adev,
 					mem_channel_number = igp_info->v21.umachannelnumber;
 					if (!mem_channel_number)
 						mem_channel_number = 1;
-					/* channel width is 64 */
-					if (vram_width)
-						*vram_width = mem_channel_number * 64;
 					mem_type = igp_info->v21.memorytype;
+					if (mem_type == LpDdr5MemType)
+						mem_channel_width = 32;
+					else
+						mem_channel_width = 64;
+					if (vram_width)
+						*vram_width = mem_channel_number * mem_channel_width;
+					if (vram_type)
+						*vram_type = convert_atom_mem_type_to_vram_type(adev, mem_type);
+					break;
+				case 3:
+					mem_channel_number = igp_info->v23.umachannelnumber;
+					if (!mem_channel_number)
+						mem_channel_number = 1;
+					mem_type = igp_info->v23.memorytype;
+					if (mem_type == LpDdr5MemType)
+						mem_channel_width = 32;
+					else
+						mem_channel_width = 64;
+					if (vram_width)
+						*vram_width = mem_channel_number * mem_channel_width;
 					if (vram_type)
 						*vram_type = convert_atom_mem_type_to_vram_type(adev, mem_type);
 					break;
@@ -291,95 +395,150 @@ amdgpu_atomfirmware_get_vram_info(struct amdgpu_device *adev,
 				return -EINVAL;
 			}
 		} else {
-			vram_info = (union vram_info *)
-				(mode_info->atom_context->bios + data_offset);
-			module_id = (RREG32(adev->bios_scratch_reg_offset + 4) & 0x00ff0000) >> 16;
-			switch (crev) {
-			case 3:
-				if (module_id > vram_info->v23.vram_module_num)
-					module_id = 0;
-				vram_module = (union vram_module *)vram_info->v23.vram_module;
-				while (i < module_id) {
-					vram_module = (union vram_module *)
-						((u8 *)vram_module + vram_module->v9.vram_module_size);
-					i++;
-				}
-				mem_type = vram_module->v9.memory_type;
-				if (vram_type)
-					*vram_type = convert_atom_mem_type_to_vram_type(adev, mem_type);
-				mem_channel_number = vram_module->v9.channel_num;
-				mem_channel_width = vram_module->v9.channel_width;
-				if (vram_width)
-					*vram_width = mem_channel_number * (1 << mem_channel_width);
-				mem_vendor = (vram_module->v9.vender_rev_id) & 0xF;
-				if (vram_vendor)
-					*vram_vendor = mem_vendor;
-				break;
-			case 4:
-				if (module_id > vram_info->v24.vram_module_num)
-					module_id = 0;
-				vram_module = (union vram_module *)vram_info->v24.vram_module;
-				while (i < module_id) {
-					vram_module = (union vram_module *)
-						((u8 *)vram_module + vram_module->v10.vram_module_size);
-					i++;
-				}
-				mem_type = vram_module->v10.memory_type;
-				if (vram_type)
-					*vram_type = convert_atom_mem_type_to_vram_type(adev, mem_type);
-				mem_channel_number = vram_module->v10.channel_num;
-				mem_channel_width = vram_module->v10.channel_width;
-				if (vram_width)
-					*vram_width = mem_channel_number * (1 << mem_channel_width);
-				mem_vendor = (vram_module->v10.vender_rev_id) & 0xF;
-				if (vram_vendor)
-					*vram_vendor = mem_vendor;
-				break;
-			case 5:
-				if (module_id > vram_info->v25.vram_module_num)
-					module_id = 0;
-				vram_module = (union vram_module *)vram_info->v25.vram_module;
-				while (i < module_id) {
-					vram_module = (union vram_module *)
-						((u8 *)vram_module + vram_module->v11.vram_module_size);
-					i++;
-				}
-				mem_type = vram_module->v11.memory_type;
-				if (vram_type)
-					*vram_type = convert_atom_mem_type_to_vram_type(adev, mem_type);
-				mem_channel_number = vram_module->v11.channel_num;
-				mem_channel_width = vram_module->v11.channel_width;
-				if (vram_width)
-					*vram_width = mem_channel_number * (1 << mem_channel_width);
-				mem_vendor = (vram_module->v11.vender_rev_id) & 0xF;
-				if (vram_vendor)
-					*vram_vendor = mem_vendor;
-				break;
-			case 6:
-				if (module_id > vram_info->v26.vram_module_num)
-					module_id = 0;
-				vram_module = (union vram_module *)vram_info->v26.vram_module;
-				while (i < module_id) {
-					vram_module = (union vram_module *)
-						((u8 *)vram_module + vram_module->v9.vram_module_size);
-					i++;
-				}
-				mem_type = vram_module->v9.memory_type;
-				if (vram_type)
-					*vram_type = convert_atom_mem_type_to_vram_type(adev, mem_type);
-				mem_channel_number = vram_module->v9.channel_num;
-				mem_channel_width = vram_module->v9.channel_width;
-				if (vram_width)
-					*vram_width = mem_channel_number * (1 << mem_channel_width);
-				mem_vendor = (vram_module->v9.vender_rev_id) & 0xF;
-				if (vram_vendor)
-					*vram_vendor = mem_vendor;
+			switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+			case IP_VERSION(12, 0, 0):
+			case IP_VERSION(12, 0, 1):
+				umc_info = (union umc_info *)(mode_info->atom_context->bios + data_offset);
+
+				if (frev == 4) {
+					switch (crev) {
+					case 0:
+						mem_channel_number = le32_to_cpu(umc_info->v40.channel_num);
+						mem_type = le32_to_cpu(umc_info->v40.vram_type);
+						mem_channel_width = le32_to_cpu(umc_info->v40.channel_width);
+						mem_vendor = RREG32(adev->bios_scratch_reg_offset + 4) & 0xF;
+						if (vram_vendor)
+							*vram_vendor = mem_vendor;
+						if (vram_type)
+							*vram_type = convert_atom_mem_type_to_vram_type(adev, mem_type);
+						if (vram_width)
+							*vram_width = mem_channel_number * (1 << mem_channel_width);
+						break;
+					default:
+						return -EINVAL;
+					}
+				} else
+					return -EINVAL;
 				break;
 			default:
-				return -EINVAL;
+				vram_info = (union vram_info *)
+					(mode_info->atom_context->bios + data_offset);
+
+				module_id = (RREG32(adev->bios_scratch_reg_offset + 4) & 0x00ff0000) >> 16;
+				if (frev == 3) {
+					switch (crev) {
+					/* v30 */
+					case 0:
+						vram_module = (union vram_module *)vram_info->v30.vram_module;
+						mem_vendor = (vram_module->v30.dram_vendor_id) & 0xF;
+						if (vram_vendor)
+							*vram_vendor = mem_vendor;
+						mem_type = vram_info->v30.memory_type;
+						if (vram_type)
+							*vram_type = convert_atom_mem_type_to_vram_type(adev, mem_type);
+						mem_channel_number = vram_info->v30.channel_num;
+						mem_channel_width = vram_info->v30.channel_width;
+						if (vram_width)
+							*vram_width = mem_channel_number * 16;
+						break;
+					default:
+						return -EINVAL;
+					}
+				} else if (frev == 2) {
+					switch (crev) {
+					/* v23 */
+					case 3:
+						if (module_id > vram_info->v23.vram_module_num)
+							module_id = 0;
+						vram_module = (union vram_module *)vram_info->v23.vram_module;
+						while (i < module_id) {
+							vram_module = (union vram_module *)
+								((u8 *)vram_module + vram_module->v9.vram_module_size);
+							i++;
+						}
+						mem_type = vram_module->v9.memory_type;
+						if (vram_type)
+							*vram_type = convert_atom_mem_type_to_vram_type(adev, mem_type);
+						mem_channel_number = vram_module->v9.channel_num;
+						mem_channel_width = vram_module->v9.channel_width;
+						if (vram_width)
+							*vram_width = mem_channel_number * (1 << mem_channel_width);
+						mem_vendor = (vram_module->v9.vender_rev_id) & 0xF;
+						if (vram_vendor)
+							*vram_vendor = mem_vendor;
+						break;
+					/* v24 */
+					case 4:
+						if (module_id > vram_info->v24.vram_module_num)
+							module_id = 0;
+						vram_module = (union vram_module *)vram_info->v24.vram_module;
+						while (i < module_id) {
+							vram_module = (union vram_module *)
+								((u8 *)vram_module + vram_module->v10.vram_module_size);
+							i++;
+						}
+						mem_type = vram_module->v10.memory_type;
+						if (vram_type)
+							*vram_type = convert_atom_mem_type_to_vram_type(adev, mem_type);
+						mem_channel_number = vram_module->v10.channel_num;
+						mem_channel_width = vram_module->v10.channel_width;
+						if (vram_width)
+							*vram_width = mem_channel_number * (1 << mem_channel_width);
+						mem_vendor = (vram_module->v10.vender_rev_id) & 0xF;
+						if (vram_vendor)
+							*vram_vendor = mem_vendor;
+						break;
+					/* v25 */
+					case 5:
+						if (module_id > vram_info->v25.vram_module_num)
+							module_id = 0;
+						vram_module = (union vram_module *)vram_info->v25.vram_module;
+						while (i < module_id) {
+							vram_module = (union vram_module *)
+								((u8 *)vram_module + vram_module->v11.vram_module_size);
+							i++;
+						}
+						mem_type = vram_module->v11.memory_type;
+						if (vram_type)
+							*vram_type = convert_atom_mem_type_to_vram_type(adev, mem_type);
+						mem_channel_number = vram_module->v11.channel_num;
+						mem_channel_width = vram_module->v11.channel_width;
+						if (vram_width)
+							*vram_width = mem_channel_number * (1 << mem_channel_width);
+						mem_vendor = (vram_module->v11.vender_rev_id) & 0xF;
+						if (vram_vendor)
+							*vram_vendor = mem_vendor;
+						break;
+					/* v26 */
+					case 6:
+						if (module_id > vram_info->v26.vram_module_num)
+							module_id = 0;
+						vram_module = (union vram_module *)vram_info->v26.vram_module;
+						while (i < module_id) {
+							vram_module = (union vram_module *)
+								((u8 *)vram_module + vram_module->v9.vram_module_size);
+							i++;
+						}
+						mem_type = vram_module->v9.memory_type;
+						if (vram_type)
+							*vram_type = convert_atom_mem_type_to_vram_type(adev, mem_type);
+						mem_channel_number = vram_module->v9.channel_num;
+						mem_channel_width = vram_module->v9.channel_width;
+						if (vram_width)
+							*vram_width = mem_channel_number * (1 << mem_channel_width);
+						mem_vendor = (vram_module->v9.vender_rev_id) & 0xF;
+						if (vram_vendor)
+							*vram_vendor = mem_vendor;
+						break;
+					default:
+						return -EINVAL;
+					}
+				} else {
+					/* invalid frev */
+					return -EINVAL;
+				}
 			}
 		}
-
 	}
 
 	return 0;
@@ -396,44 +555,63 @@ bool amdgpu_atomfirmware_mem_ecc_supported(struct amdgpu_device *adev)
 	u16 data_offset, size;
 	union umc_info *umc_info;
 	u8 frev, crev;
-	bool ecc_default_enabled = false;
+	bool mem_ecc_enabled = false;
 	u8 umc_config;
 	u32 umc_config1;
+	adev->ras_default_ecc_enabled = false;
 
 	index = get_index_into_master_table(atom_master_list_of_data_tables_v2_1,
 			umc_info);
 
 	if (amdgpu_atom_parse_data_header(mode_info->atom_context,
 				index, &size, &frev, &crev, &data_offset)) {
+		umc_info = (union umc_info *)(mode_info->atom_context->bios + data_offset);
 		if (frev == 3) {
-			umc_info = (union umc_info *)
-				(mode_info->atom_context->bios + data_offset);
 			switch (crev) {
 			case 1:
 				umc_config = le32_to_cpu(umc_info->v31.umc_config);
-				ecc_default_enabled =
+				mem_ecc_enabled =
 					(umc_config & UMC_CONFIG__DEFAULT_MEM_ECC_ENABLE) ? true : false;
 				break;
 			case 2:
 				umc_config = le32_to_cpu(umc_info->v32.umc_config);
-				ecc_default_enabled =
+				mem_ecc_enabled =
 					(umc_config & UMC_CONFIG__DEFAULT_MEM_ECC_ENABLE) ? true : false;
 				break;
 			case 3:
 				umc_config = le32_to_cpu(umc_info->v33.umc_config);
 				umc_config1 = le32_to_cpu(umc_info->v33.umc_config1);
-				ecc_default_enabled =
+				mem_ecc_enabled =
 					((umc_config & UMC_CONFIG__DEFAULT_MEM_ECC_ENABLE) ||
 					 (umc_config1 & UMC_CONFIG1__ENABLE_ECC_CAPABLE)) ? true : false;
+				adev->ras_default_ecc_enabled =
+					(umc_config & UMC_CONFIG__DEFAULT_MEM_ECC_ENABLE) ? true : false;
+				break;
+			default:
+				/* unsupported crev */
+				return false;
+			}
+		} else if (frev == 4) {
+			switch (crev) {
+			case 0:
+				umc_config = le32_to_cpu(umc_info->v40.umc_config);
+				umc_config1 = le32_to_cpu(umc_info->v40.umc_config1);
+				mem_ecc_enabled =
+					(umc_config1 & UMC_CONFIG1__ENABLE_ECC_CAPABLE) ? true : false;
+				adev->ras_default_ecc_enabled =
+					(umc_config & UMC_CONFIG__DEFAULT_MEM_ECC_ENABLE) ? true : false;
 				break;
 			default:
 				/* unsupported crev */
 				return false;
 			}
+		} else {
+			/* unsupported frev */
+			return false;
 		}
 	}
 
-	return ecc_default_enabled;
+	return mem_ecc_enabled;
 }
 
 /*
@@ -470,8 +648,8 @@ bool amdgpu_atomfirmware_dynamic_boot_config_supported(struct amdgpu_device *ade
 
 /**
  * amdgpu_atomfirmware_ras_rom_addr -- Get the RAS EEPROM addr from VBIOS
- * adev: amdgpu_device pointer
- * i2c_address: pointer to u8; if not NULL, will contain
+ * @adev: amdgpu_device pointer
+ * @i2c_address: pointer to u8; if not NULL, will contain
  *    the RAS EEPROM address if the function returns true
  *
  * Return true if VBIOS supports RAS EEPROM address reporting,
@@ -494,7 +672,7 @@ bool amdgpu_atomfirmware_ras_rom_addr(struct amdgpu_device *adev,
 					  index, &size, &frev, &crev,
 					  &data_offset)) {
 		/* support firmware_info 3.4 + */
-		if ((frev == 3 && crev >=4) || (frev > 3)) {
+		if ((frev == 3 && crev >= 4) || (frev > 3)) {
 			firmware_info = (union firmware_info *)
 				(mode_info->atom_context->bios + data_offset);
 			/* The ras_rom_i2c_slave_addr should ideally
@@ -526,6 +704,14 @@ bool amdgpu_atomfirmware_ras_rom_addr(struct amdgpu_device *adev,
 
 union smu_info {
 	struct atom_smu_info_v3_1 v31;
+	struct atom_smu_info_v4_0 v40;
+};
+
+union gfx_info {
+	struct atom_gfx_info_v2_2 v22;
+	struct atom_gfx_info_v2_4 v24;
+	struct atom_gfx_info_v2_7 v27;
+	struct atom_gfx_info_v3_0 v30;
 };
 
 int amdgpu_atomfirmware_get_clock_info(struct amdgpu_device *adev)
@@ -565,7 +751,10 @@ int amdgpu_atomfirmware_get_clock_info(struct amdgpu_device *adev)
 					   data_offset);
 
 		/* system clock */
-		spll->reference_freq = le32_to_cpu(smu_info->v31.core_refclk_10khz);
+		if (frev == 3)
+			spll->reference_freq = le32_to_cpu(smu_info->v31.core_refclk_10khz);
+		else if (frev == 4)
+			spll->reference_freq = le32_to_cpu(smu_info->v40.core_refclk_10khz);
 
 		spll->reference_div = 0;
 		spll->min_post_div = 1;
@@ -609,22 +798,26 @@ int amdgpu_atomfirmware_get_clock_info(struct amdgpu_device *adev)
 						   gfx_info);
 		if (amdgpu_atom_parse_data_header(mode_info->atom_context, index, NULL,
 					  &frev, &crev, &data_offset)) {
-			struct atom_gfx_info_v2_2 *gfx_info = (struct atom_gfx_info_v2_2*)
+			union gfx_info *gfx_info = (union gfx_info *)
 				(mode_info->atom_context->bios + data_offset);
-			if ((frev == 2) && (crev >= 2))
-				spll->reference_freq = le32_to_cpu(gfx_info->rlc_gpu_timer_refclk);
-			ret = 0;
+			if ((frev == 3) ||
+			    (frev == 2 && crev == 6)) {
+				spll->reference_freq = le32_to_cpu(gfx_info->v30.golden_tsc_count_lower_refclk);
+				ret = 0;
+			} else if ((frev == 2) &&
+				   (crev >= 2) &&
+				   (crev != 6)) {
+				spll->reference_freq = le32_to_cpu(gfx_info->v22.rlc_gpu_timer_refclk);
+				ret = 0;
+			} else {
+				BUG();
+			}
 		}
 	}
 
 	return ret;
 }
 
-union gfx_info {
-	struct atom_gfx_info_v2_4 v24;
-	struct atom_gfx_info_v2_7 v27;
-};
-
 int amdgpu_atomfirmware_get_gfx_info(struct amdgpu_device *adev)
 {
 	struct amdgpu_mode_info *mode_info = &adev->mode_info;
@@ -638,42 +831,58 @@ int amdgpu_atomfirmware_get_gfx_info(struct amdgpu_device *adev)
 				   &frev, &crev, &data_offset)) {
 		union gfx_info *gfx_info = (union gfx_info *)
 			(mode_info->atom_context->bios + data_offset);
-		switch (crev) {
-		case 4:
-			adev->gfx.config.max_shader_engines = gfx_info->v24.max_shader_engines;
-			adev->gfx.config.max_cu_per_sh = gfx_info->v24.max_cu_per_sh;
-			adev->gfx.config.max_sh_per_se = gfx_info->v24.max_sh_per_se;
-			adev->gfx.config.max_backends_per_se = gfx_info->v24.max_backends_per_se;
-			adev->gfx.config.max_texture_channel_caches = gfx_info->v24.max_texture_channel_caches;
-			adev->gfx.config.max_gprs = le16_to_cpu(gfx_info->v24.gc_num_gprs);
-			adev->gfx.config.max_gs_threads = gfx_info->v24.gc_num_max_gs_thds;
-			adev->gfx.config.gs_vgt_table_depth = gfx_info->v24.gc_gs_table_depth;
-			adev->gfx.config.gs_prim_buffer_depth =
-				le16_to_cpu(gfx_info->v24.gc_gsprim_buff_depth);
-			adev->gfx.config.double_offchip_lds_buf =
-				gfx_info->v24.gc_double_offchip_lds_buffer;
-			adev->gfx.cu_info.wave_front_size = le16_to_cpu(gfx_info->v24.gc_wave_size);
-			adev->gfx.cu_info.max_waves_per_simd = le16_to_cpu(gfx_info->v24.gc_max_waves_per_simd);
-			adev->gfx.cu_info.max_scratch_slots_per_cu = gfx_info->v24.gc_max_scratch_slots_per_cu;
-			adev->gfx.cu_info.lds_size = le16_to_cpu(gfx_info->v24.gc_lds_size);
-			return 0;
-		case 7:
-			adev->gfx.config.max_shader_engines = gfx_info->v27.max_shader_engines;
-			adev->gfx.config.max_cu_per_sh = gfx_info->v27.max_cu_per_sh;
-			adev->gfx.config.max_sh_per_se = gfx_info->v27.max_sh_per_se;
-			adev->gfx.config.max_backends_per_se = gfx_info->v27.max_backends_per_se;
-			adev->gfx.config.max_texture_channel_caches = gfx_info->v27.max_texture_channel_caches;
-			adev->gfx.config.max_gprs = le16_to_cpu(gfx_info->v27.gc_num_gprs);
-			adev->gfx.config.max_gs_threads = gfx_info->v27.gc_num_max_gs_thds;
-			adev->gfx.config.gs_vgt_table_depth = gfx_info->v27.gc_gs_table_depth;
-			adev->gfx.config.gs_prim_buffer_depth = le16_to_cpu(gfx_info->v27.gc_gsprim_buff_depth);
-			adev->gfx.config.double_offchip_lds_buf = gfx_info->v27.gc_double_offchip_lds_buffer;
-			adev->gfx.cu_info.wave_front_size = le16_to_cpu(gfx_info->v27.gc_wave_size);
-			adev->gfx.cu_info.max_waves_per_simd = le16_to_cpu(gfx_info->v27.gc_max_waves_per_simd);
-			adev->gfx.cu_info.max_scratch_slots_per_cu = gfx_info->v27.gc_max_scratch_slots_per_cu;
-			adev->gfx.cu_info.lds_size = le16_to_cpu(gfx_info->v27.gc_lds_size);
-			return 0;
-		default:
+		if (frev == 2) {
+			switch (crev) {
+			case 4:
+				adev->gfx.config.max_shader_engines = gfx_info->v24.max_shader_engines;
+				adev->gfx.config.max_cu_per_sh = gfx_info->v24.max_cu_per_sh;
+				adev->gfx.config.max_sh_per_se = gfx_info->v24.max_sh_per_se;
+				adev->gfx.config.max_backends_per_se = gfx_info->v24.max_backends_per_se;
+				adev->gfx.config.max_texture_channel_caches = gfx_info->v24.max_texture_channel_caches;
+				adev->gfx.config.max_gprs = le16_to_cpu(gfx_info->v24.gc_num_gprs);
+				adev->gfx.config.max_gs_threads = gfx_info->v24.gc_num_max_gs_thds;
+				adev->gfx.config.gs_vgt_table_depth = gfx_info->v24.gc_gs_table_depth;
+				adev->gfx.config.gs_prim_buffer_depth =
+					le16_to_cpu(gfx_info->v24.gc_gsprim_buff_depth);
+				adev->gfx.config.double_offchip_lds_buf =
+					gfx_info->v24.gc_double_offchip_lds_buffer;
+				adev->gfx.cu_info.wave_front_size = le16_to_cpu(gfx_info->v24.gc_wave_size);
+				adev->gfx.cu_info.max_waves_per_simd = le16_to_cpu(gfx_info->v24.gc_max_waves_per_simd);
+				adev->gfx.cu_info.max_scratch_slots_per_cu = gfx_info->v24.gc_max_scratch_slots_per_cu;
+				adev->gfx.cu_info.lds_size = le16_to_cpu(gfx_info->v24.gc_lds_size);
+				return 0;
+			case 7:
+				adev->gfx.config.max_shader_engines = gfx_info->v27.max_shader_engines;
+				adev->gfx.config.max_cu_per_sh = gfx_info->v27.max_cu_per_sh;
+				adev->gfx.config.max_sh_per_se = gfx_info->v27.max_sh_per_se;
+				adev->gfx.config.max_backends_per_se = gfx_info->v27.max_backends_per_se;
+				adev->gfx.config.max_texture_channel_caches = gfx_info->v27.max_texture_channel_caches;
+				adev->gfx.config.max_gprs = le16_to_cpu(gfx_info->v27.gc_num_gprs);
+				adev->gfx.config.max_gs_threads = gfx_info->v27.gc_num_max_gs_thds;
+				adev->gfx.config.gs_vgt_table_depth = gfx_info->v27.gc_gs_table_depth;
+				adev->gfx.config.gs_prim_buffer_depth = le16_to_cpu(gfx_info->v27.gc_gsprim_buff_depth);
+				adev->gfx.config.double_offchip_lds_buf = gfx_info->v27.gc_double_offchip_lds_buffer;
+				adev->gfx.cu_info.wave_front_size = le16_to_cpu(gfx_info->v27.gc_wave_size);
+				adev->gfx.cu_info.max_waves_per_simd = le16_to_cpu(gfx_info->v27.gc_max_waves_per_simd);
+				adev->gfx.cu_info.max_scratch_slots_per_cu = gfx_info->v27.gc_max_scratch_slots_per_cu;
+				adev->gfx.cu_info.lds_size = le16_to_cpu(gfx_info->v27.gc_lds_size);
+				return 0;
+			default:
+				return -EINVAL;
+			}
+		} else if (frev == 3) {
+			switch (crev) {
+			case 0:
+				adev->gfx.config.max_shader_engines = gfx_info->v30.max_shader_engines;
+				adev->gfx.config.max_cu_per_sh = gfx_info->v30.max_cu_per_sh;
+				adev->gfx.config.max_sh_per_se = gfx_info->v30.max_sh_per_se;
+				adev->gfx.config.max_backends_per_se = gfx_info->v30.max_backends_per_se;
+				adev->gfx.config.max_texture_channel_caches = gfx_info->v30.max_texture_channel_caches;
+				return 0;
+			default:
+				return -EINVAL;
+			}
+		} else {
 			return -EINVAL;
 		}
 
@@ -716,7 +925,7 @@ int amdgpu_atomfirmware_get_fw_reserved_fb_size(struct amdgpu_device *adev)
 
 	firmware_info = (union firmware_info *)(ctx->bios + data_offset);
 
-	if (frev !=3)
+	if (frev != 3)
 		return -EINVAL;
 
 	switch (crev) {
@@ -724,6 +933,10 @@ int amdgpu_atomfirmware_get_fw_reserved_fb_size(struct amdgpu_device *adev)
 		fw_reserved_fb_size =
 			(firmware_info->v34.fw_reserved_size_in_kb << 10);
 		break;
+	case 5:
+		fw_reserved_fb_size =
+			(firmware_info->v35.fw_reserved_size_in_kb << 10);
+		break;
 	default:
 		fw_reserved_fb_size = 0;
 		break;
@@ -731,3 +944,68 @@ int amdgpu_atomfirmware_get_fw_reserved_fb_size(struct amdgpu_device *adev)
 
 	return fw_reserved_fb_size;
 }
+
+/*
+ * Helper function to execute asic_init table
+ *
+ * @adev: amdgpu_device pointer
+ * @fb_reset: flag to indicate whether fb is reset or not
+ *
+ * Return 0 if succeed, otherwise failed
+ */
+int amdgpu_atomfirmware_asic_init(struct amdgpu_device *adev, bool fb_reset)
+{
+	struct amdgpu_mode_info *mode_info = &adev->mode_info;
+	struct atom_context *ctx;
+	uint8_t frev, crev;
+	uint16_t data_offset;
+	uint32_t bootup_sclk_in10khz, bootup_mclk_in10khz;
+	struct asic_init_ps_allocation_v2_1 asic_init_ps_v2_1;
+	int index;
+
+	if (!mode_info)
+		return -EINVAL;
+
+	ctx = mode_info->atom_context;
+	if (!ctx)
+		return -EINVAL;
+
+	/* query bootup sclk/mclk from firmware_info table */
+	index = get_index_into_master_table(atom_master_list_of_data_tables_v2_1,
+					    firmwareinfo);
+	if (amdgpu_atom_parse_data_header(ctx, index, NULL,
+				&frev, &crev, &data_offset)) {
+		union firmware_info *firmware_info =
+			(union firmware_info *)(ctx->bios +
+						data_offset);
+
+		bootup_sclk_in10khz =
+			le32_to_cpu(firmware_info->v31.bootup_sclk_in10khz);
+		bootup_mclk_in10khz =
+			le32_to_cpu(firmware_info->v31.bootup_mclk_in10khz);
+	} else {
+		return -EINVAL;
+	}
+
+	index = get_index_into_master_table(atom_master_list_of_command_functions_v2_1,
+					asic_init);
+	if (amdgpu_atom_parse_cmd_header(mode_info->atom_context, index, &frev, &crev)) {
+		if (frev == 2 && crev >= 1) {
+			memset(&asic_init_ps_v2_1, 0, sizeof(asic_init_ps_v2_1));
+			asic_init_ps_v2_1.param.engineparam.sclkfreqin10khz = bootup_sclk_in10khz;
+			asic_init_ps_v2_1.param.memparam.mclkfreqin10khz = bootup_mclk_in10khz;
+			asic_init_ps_v2_1.param.engineparam.engineflag = b3NORMAL_ENGINE_INIT;
+			if (!fb_reset)
+				asic_init_ps_v2_1.param.memparam.memflag = b3DRAM_SELF_REFRESH_EXIT;
+			else
+				asic_init_ps_v2_1.param.memparam.memflag = 0;
+		} else {
+			return -EINVAL;
+		}
+	} else {
+		return -EINVAL;
+	}
+
+	return amdgpu_atom_execute_table(ctx, ATOM_CMD_INIT, (uint32_t *)&asic_init_ps_v2_1,
+		sizeof(asic_init_ps_v2_1));
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.h
index 751248b253de..649b5530d8ae 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.h
@@ -36,9 +36,10 @@ int amdgpu_atomfirmware_get_clock_info(struct amdgpu_device *adev);
 int amdgpu_atomfirmware_get_gfx_info(struct amdgpu_device *adev);
 bool amdgpu_atomfirmware_mem_ecc_supported(struct amdgpu_device *adev);
 bool amdgpu_atomfirmware_sram_ecc_supported(struct amdgpu_device *adev);
-bool amdgpu_atomfirmware_ras_rom_addr(struct amdgpu_device *adev, uint8_t* i2c_address);
+bool amdgpu_atomfirmware_ras_rom_addr(struct amdgpu_device *adev, uint8_t *i2c_address);
 bool amdgpu_atomfirmware_mem_training_supported(struct amdgpu_device *adev);
 bool amdgpu_atomfirmware_dynamic_boot_config_supported(struct amdgpu_device *adev);
 int amdgpu_atomfirmware_get_fw_reserved_fb_size(struct amdgpu_device *adev);
+int amdgpu_atomfirmware_asic_init(struct amdgpu_device *adev, bool fb_reset);
 
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atpx_handler.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_atpx_handler.c
index 7abe9500c0c6..3893e6fc2f03 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atpx_handler.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_atpx_handler.c
@@ -11,6 +11,7 @@
 #include <linux/pci.h>
 #include <linux/delay.h>
 
+#include "amdgpu.h"
 #include "amd_acpi.h"
 
 #define AMDGPU_PX_QUIRK_FORCE_ATPX  (1 << 0)
@@ -73,28 +74,21 @@ struct atpx_mux {
 	u16 mux;
 } __packed;
 
-bool amdgpu_has_atpx(void) {
+bool amdgpu_has_atpx(void)
+{
 	return amdgpu_atpx_priv.atpx_detected;
 }
 
-bool amdgpu_has_atpx_dgpu_power_cntl(void) {
+bool amdgpu_has_atpx_dgpu_power_cntl(void)
+{
 	return amdgpu_atpx_priv.atpx.functions.power_cntl;
 }
 
-bool amdgpu_is_atpx_hybrid(void) {
+bool amdgpu_is_atpx_hybrid(void)
+{
 	return amdgpu_atpx_priv.atpx.is_hybrid;
 }
 
-bool amdgpu_atpx_dgpu_req_power_for_displays(void) {
-	return amdgpu_atpx_priv.atpx.dgpu_req_power_for_displays;
-}
-
-#if defined(CONFIG_ACPI)
-void *amdgpu_atpx_get_dhandle(void) {
-	return amdgpu_atpx_priv.dhandle;
-}
-#endif
-
 /**
  * amdgpu_atpx_call - call an ATPX method
  *
@@ -133,7 +127,7 @@ static union acpi_object *amdgpu_atpx_call(acpi_handle handle, int function,
 
 	/* Fail only if calling the method fails and ATPX is supported */
 	if (ACPI_FAILURE(status) && status != AE_NOT_FOUND) {
-		printk("failed to evaluate ATPX got %s\n",
+		pr_err("failed to evaluate ATPX got %s\n",
 		       acpi_format_exception(status));
 		kfree(buffer.pointer);
 		return NULL;
@@ -165,7 +159,7 @@ static void amdgpu_atpx_parse_functions(struct amdgpu_atpx_functions *f, u32 mas
 }
 
 /**
- * amdgpu_atpx_validate_functions - validate ATPX functions
+ * amdgpu_atpx_validate - validate ATPX functions
  *
  * @atpx: amdgpu atpx struct
  *
@@ -189,7 +183,7 @@ static int amdgpu_atpx_validate(struct amdgpu_atpx *atpx)
 
 		size = *(u16 *) info->buffer.pointer;
 		if (size < 10) {
-			printk("ATPX buffer is too small: %zu\n", size);
+			pr_err("ATPX buffer is too small: %zu\n", size);
 			kfree(info);
 			return -EINVAL;
 		}
@@ -222,11 +216,11 @@ static int amdgpu_atpx_validate(struct amdgpu_atpx *atpx)
 	atpx->is_hybrid = false;
 	if (valid_bits & ATPX_MS_HYBRID_GFX_SUPPORTED) {
 		if (amdgpu_atpx_priv.quirks & AMDGPU_PX_QUIRK_FORCE_ATPX) {
-			printk("ATPX Hybrid Graphics, forcing to ATPX\n");
+			pr_warn("ATPX Hybrid Graphics, forcing to ATPX\n");
 			atpx->functions.power_cntl = true;
 			atpx->is_hybrid = false;
 		} else {
-			printk("ATPX Hybrid Graphics\n");
+			pr_notice("ATPX Hybrid Graphics\n");
 			/*
 			 * Disable legacy PM methods only when pcie port PM is usable,
 			 * otherwise the device might fail to power off or power on.
@@ -268,7 +262,7 @@ static int amdgpu_atpx_verify_interface(struct amdgpu_atpx *atpx)
 
 	size = *(u16 *) info->buffer.pointer;
 	if (size < 8) {
-		printk("ATPX buffer is too small: %zu\n", size);
+		pr_err("ATPX buffer is too small: %zu\n", size);
 		err = -EINVAL;
 		goto out;
 	}
@@ -277,8 +271,8 @@ static int amdgpu_atpx_verify_interface(struct amdgpu_atpx *atpx)
 	memcpy(&output, info->buffer.pointer, size);
 
 	/* TODO: check version? */
-	printk("ATPX version %u, functions 0x%08x\n",
-	       output.version, output.function_bits);
+	pr_notice("ATPX version %u, functions 0x%08x\n",
+		  output.version, output.function_bits);
 
 	amdgpu_atpx_parse_functions(&atpx->functions, output.function_bits);
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_benchmark.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_benchmark.c
index 313517f7cf10..199693369c7c 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_benchmark.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_benchmark.c
@@ -29,18 +29,17 @@
 #define AMDGPU_BENCHMARK_COMMON_MODES_N 17
 
 static int amdgpu_benchmark_do_move(struct amdgpu_device *adev, unsigned size,
-				    uint64_t saddr, uint64_t daddr, int n)
+				    uint64_t saddr, uint64_t daddr, int n, s64 *time_ms)
 {
-	unsigned long start_jiffies;
-	unsigned long end_jiffies;
+	ktime_t stime, etime;
 	struct dma_fence *fence;
 	int i, r;
 
-	start_jiffies = jiffies;
+	stime = ktime_get();
 	for (i = 0; i < n; i++) {
 		struct amdgpu_ring *ring = adev->mman.buffer_funcs_ring;
 		r = amdgpu_copy_buffer(ring, saddr, daddr, size, NULL, &fence,
-				       false, false, false);
+				       false, false, 0);
 		if (r)
 			goto exit_do_move;
 		r = dma_fence_wait(fence, false);
@@ -48,120 +47,81 @@ static int amdgpu_benchmark_do_move(struct amdgpu_device *adev, unsigned size,
 		if (r)
 			goto exit_do_move;
 	}
-	end_jiffies = jiffies;
-	r = jiffies_to_msecs(end_jiffies - start_jiffies);
 
 exit_do_move:
+	etime = ktime_get();
+	*time_ms = ktime_ms_delta(etime, stime);
+
 	return r;
 }
 
 
-static void amdgpu_benchmark_log_results(int n, unsigned size,
-					 unsigned int time,
+static void amdgpu_benchmark_log_results(struct amdgpu_device *adev,
+					 int n, unsigned size,
+					 s64 time_ms,
 					 unsigned sdomain, unsigned ddomain,
 					 char *kind)
 {
-	unsigned int throughput = (n * (size >> 10)) / time;
-	DRM_INFO("amdgpu: %s %u bo moves of %u kB from"
-		 " %d to %d in %u ms, throughput: %u Mb/s or %u MB/s\n",
-		 kind, n, size >> 10, sdomain, ddomain, time,
+	s64 throughput = (n * (size >> 10));
+
+	throughput = div64_s64(throughput, time_ms);
+
+	dev_info(adev->dev, "amdgpu: %s %u bo moves of %u kB from"
+		 " %d to %d in %lld ms, throughput: %lld Mb/s or %lld MB/s\n",
+		 kind, n, size >> 10, sdomain, ddomain, time_ms,
 		 throughput * 8, throughput);
 }
 
-static void amdgpu_benchmark_move(struct amdgpu_device *adev, unsigned size,
-				  unsigned sdomain, unsigned ddomain)
+static int amdgpu_benchmark_move(struct amdgpu_device *adev, unsigned size,
+				 unsigned sdomain, unsigned ddomain)
 {
 	struct amdgpu_bo *dobj = NULL;
 	struct amdgpu_bo *sobj = NULL;
-	struct amdgpu_bo_param bp;
 	uint64_t saddr, daddr;
+	s64 time_ms;
 	int r, n;
-	int time;
-
-	memset(&bp, 0, sizeof(bp));
-	bp.size = size;
-	bp.byte_align = PAGE_SIZE;
-	bp.domain = sdomain;
-	bp.flags = 0;
-	bp.type = ttm_bo_type_kernel;
-	bp.resv = NULL;
-	bp.bo_ptr_size = sizeof(struct amdgpu_bo);
 
 	n = AMDGPU_BENCHMARK_ITERATIONS;
-	r = amdgpu_bo_create(adev, &bp, &sobj);
-	if (r) {
-		goto out_cleanup;
-	}
-	r = amdgpu_bo_reserve(sobj, false);
-	if (unlikely(r != 0))
-		goto out_cleanup;
-	r = amdgpu_bo_pin(sobj, sdomain);
-	if (r) {
-		amdgpu_bo_unreserve(sobj);
-		goto out_cleanup;
-	}
-	r = amdgpu_ttm_alloc_gart(&sobj->tbo);
-	amdgpu_bo_unreserve(sobj);
-	if (r) {
-		goto out_cleanup;
-	}
-	saddr = amdgpu_bo_gpu_offset(sobj);
-	bp.domain = ddomain;
-	r = amdgpu_bo_create(adev, &bp, &dobj);
-	if (r) {
-		goto out_cleanup;
-	}
-	r = amdgpu_bo_reserve(dobj, false);
-	if (unlikely(r != 0))
-		goto out_cleanup;
-	r = amdgpu_bo_pin(dobj, ddomain);
-	if (r) {
-		amdgpu_bo_unreserve(sobj);
+
+	r = amdgpu_bo_create_kernel(adev, size,
+				    PAGE_SIZE, sdomain,
+				    &sobj,
+				    &saddr,
+				    NULL);
+	if (r)
 		goto out_cleanup;
-	}
-	r = amdgpu_ttm_alloc_gart(&dobj->tbo);
-	amdgpu_bo_unreserve(dobj);
-	if (r) {
+	r = amdgpu_bo_create_kernel(adev, size,
+				    PAGE_SIZE, ddomain,
+				    &dobj,
+				    &daddr,
+				    NULL);
+	if (r)
 		goto out_cleanup;
-	}
-	daddr = amdgpu_bo_gpu_offset(dobj);
 
 	if (adev->mman.buffer_funcs) {
-		time = amdgpu_benchmark_do_move(adev, size, saddr, daddr, n);
-		if (time < 0)
+		r = amdgpu_benchmark_do_move(adev, size, saddr, daddr, n, &time_ms);
+		if (r)
 			goto out_cleanup;
-		if (time > 0)
-			amdgpu_benchmark_log_results(n, size, time,
+		else
+			amdgpu_benchmark_log_results(adev, n, size, time_ms,
 						     sdomain, ddomain, "dma");
 	}
 
 out_cleanup:
 	/* Check error value now. The value can be overwritten when clean up.*/
-	if (r) {
-		DRM_ERROR("Error while benchmarking BO move.\n");
-	}
+	if (r < 0)
+		dev_info(adev->dev, "Error while benchmarking BO move.\n");
 
-	if (sobj) {
-		r = amdgpu_bo_reserve(sobj, true);
-		if (likely(r == 0)) {
-			amdgpu_bo_unpin(sobj);
-			amdgpu_bo_unreserve(sobj);
-		}
-		amdgpu_bo_unref(&sobj);
-	}
-	if (dobj) {
-		r = amdgpu_bo_reserve(dobj, true);
-		if (likely(r == 0)) {
-			amdgpu_bo_unpin(dobj);
-			amdgpu_bo_unreserve(dobj);
-		}
-		amdgpu_bo_unref(&dobj);
-	}
+	if (sobj)
+		amdgpu_bo_free_kernel(&sobj, &saddr, NULL);
+	if (dobj)
+		amdgpu_bo_free_kernel(&dobj, &daddr, NULL);
+	return r;
 }
 
-void amdgpu_benchmark(struct amdgpu_device *adev, int test_number)
+int amdgpu_benchmark(struct amdgpu_device *adev, int test_number)
 {
-	int i;
+	int i, r;
 	static const int common_modes[AMDGPU_BENCHMARK_COMMON_MODES_N] = {
 		640 * 480 * 4,
 		720 * 480 * 4,
@@ -182,63 +142,119 @@ void amdgpu_benchmark(struct amdgpu_device *adev, int test_number)
 		1920 * 1200 * 4
 	};
 
+	mutex_lock(&adev->benchmark_mutex);
 	switch (test_number) {
 	case 1:
+		dev_info(adev->dev,
+			 "benchmark test: %d (simple test, VRAM to GTT and GTT to VRAM)\n",
+			 test_number);
 		/* simple test, VRAM to GTT and GTT to VRAM */
-		amdgpu_benchmark_move(adev, 1024*1024, AMDGPU_GEM_DOMAIN_GTT,
-				      AMDGPU_GEM_DOMAIN_VRAM);
-		amdgpu_benchmark_move(adev, 1024*1024, AMDGPU_GEM_DOMAIN_VRAM,
-				      AMDGPU_GEM_DOMAIN_GTT);
+		r = amdgpu_benchmark_move(adev, 1024*1024, AMDGPU_GEM_DOMAIN_GTT,
+					  AMDGPU_GEM_DOMAIN_VRAM);
+		if (r)
+			goto done;
+		r = amdgpu_benchmark_move(adev, 1024*1024, AMDGPU_GEM_DOMAIN_VRAM,
+					  AMDGPU_GEM_DOMAIN_GTT);
+		if (r)
+			goto done;
 		break;
 	case 2:
+		dev_info(adev->dev,
+			 "benchmark test: %d (simple test, VRAM to VRAM)\n",
+			 test_number);
 		/* simple test, VRAM to VRAM */
-		amdgpu_benchmark_move(adev, 1024*1024, AMDGPU_GEM_DOMAIN_VRAM,
-				      AMDGPU_GEM_DOMAIN_VRAM);
+		r = amdgpu_benchmark_move(adev, 1024*1024, AMDGPU_GEM_DOMAIN_VRAM,
+					  AMDGPU_GEM_DOMAIN_VRAM);
+		if (r)
+			goto done;
 		break;
 	case 3:
+		dev_info(adev->dev,
+			 "benchmark test: %d (GTT to VRAM, buffer size sweep, powers of 2)\n",
+			 test_number);
 		/* GTT to VRAM, buffer size sweep, powers of 2 */
-		for (i = 1; i <= 16384; i <<= 1)
-			amdgpu_benchmark_move(adev, i * AMDGPU_GPU_PAGE_SIZE,
-					      AMDGPU_GEM_DOMAIN_GTT,
-					      AMDGPU_GEM_DOMAIN_VRAM);
+		for (i = 1; i <= 16384; i <<= 1) {
+			r = amdgpu_benchmark_move(adev, i * AMDGPU_GPU_PAGE_SIZE,
+						  AMDGPU_GEM_DOMAIN_GTT,
+						  AMDGPU_GEM_DOMAIN_VRAM);
+			if (r)
+				goto done;
+		}
 		break;
 	case 4:
+		dev_info(adev->dev,
+			 "benchmark test: %d (VRAM to GTT, buffer size sweep, powers of 2)\n",
+			 test_number);
 		/* VRAM to GTT, buffer size sweep, powers of 2 */
-		for (i = 1; i <= 16384; i <<= 1)
-			amdgpu_benchmark_move(adev, i * AMDGPU_GPU_PAGE_SIZE,
-					      AMDGPU_GEM_DOMAIN_VRAM,
-					      AMDGPU_GEM_DOMAIN_GTT);
+		for (i = 1; i <= 16384; i <<= 1) {
+			r = amdgpu_benchmark_move(adev, i * AMDGPU_GPU_PAGE_SIZE,
+						  AMDGPU_GEM_DOMAIN_VRAM,
+						  AMDGPU_GEM_DOMAIN_GTT);
+			if (r)
+				goto done;
+		}
 		break;
 	case 5:
+		dev_info(adev->dev,
+			 "benchmark test: %d (VRAM to VRAM, buffer size sweep, powers of 2)\n",
+			 test_number);
 		/* VRAM to VRAM, buffer size sweep, powers of 2 */
-		for (i = 1; i <= 16384; i <<= 1)
-			amdgpu_benchmark_move(adev, i * AMDGPU_GPU_PAGE_SIZE,
-					      AMDGPU_GEM_DOMAIN_VRAM,
-					      AMDGPU_GEM_DOMAIN_VRAM);
+		for (i = 1; i <= 16384; i <<= 1) {
+			r = amdgpu_benchmark_move(adev, i * AMDGPU_GPU_PAGE_SIZE,
+						  AMDGPU_GEM_DOMAIN_VRAM,
+						  AMDGPU_GEM_DOMAIN_VRAM);
+			if (r)
+				goto done;
+		}
 		break;
 	case 6:
+		dev_info(adev->dev,
+			 "benchmark test: %d (GTT to VRAM, buffer size sweep, common modes)\n",
+			 test_number);
 		/* GTT to VRAM, buffer size sweep, common modes */
-		for (i = 0; i < AMDGPU_BENCHMARK_COMMON_MODES_N; i++)
-			amdgpu_benchmark_move(adev, common_modes[i],
-					      AMDGPU_GEM_DOMAIN_GTT,
-					      AMDGPU_GEM_DOMAIN_VRAM);
+		for (i = 0; i < AMDGPU_BENCHMARK_COMMON_MODES_N; i++) {
+			r = amdgpu_benchmark_move(adev, common_modes[i],
+						  AMDGPU_GEM_DOMAIN_GTT,
+						  AMDGPU_GEM_DOMAIN_VRAM);
+			if (r)
+				goto done;
+		}
 		break;
 	case 7:
+		dev_info(adev->dev,
+			 "benchmark test: %d (VRAM to GTT, buffer size sweep, common modes)\n",
+			 test_number);
 		/* VRAM to GTT, buffer size sweep, common modes */
-		for (i = 0; i < AMDGPU_BENCHMARK_COMMON_MODES_N; i++)
-			amdgpu_benchmark_move(adev, common_modes[i],
-					      AMDGPU_GEM_DOMAIN_VRAM,
-					      AMDGPU_GEM_DOMAIN_GTT);
+		for (i = 0; i < AMDGPU_BENCHMARK_COMMON_MODES_N; i++) {
+			r = amdgpu_benchmark_move(adev, common_modes[i],
+						  AMDGPU_GEM_DOMAIN_VRAM,
+						  AMDGPU_GEM_DOMAIN_GTT);
+			if (r)
+				goto done;
+		}
 		break;
 	case 8:
+		dev_info(adev->dev,
+			 "benchmark test: %d (VRAM to VRAM, buffer size sweep, common modes)\n",
+			 test_number);
 		/* VRAM to VRAM, buffer size sweep, common modes */
-		for (i = 0; i < AMDGPU_BENCHMARK_COMMON_MODES_N; i++)
-			amdgpu_benchmark_move(adev, common_modes[i],
+		for (i = 0; i < AMDGPU_BENCHMARK_COMMON_MODES_N; i++) {
+			r = amdgpu_benchmark_move(adev, common_modes[i],
 					      AMDGPU_GEM_DOMAIN_VRAM,
 					      AMDGPU_GEM_DOMAIN_VRAM);
+			if (r)
+				goto done;
+		}
 		break;
 
 	default:
-		DRM_ERROR("Unknown benchmark\n");
+		dev_info(adev->dev, "Unknown benchmark %d\n", test_number);
+		r = -EINVAL;
+		break;
 	}
+
+done:
+	mutex_unlock(&adev->benchmark_mutex);
+
+	return r;
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_bios.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_bios.c
index 27b19503773b..35d04e69aec0 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_bios.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_bios.c
@@ -29,6 +29,7 @@
 #include "amdgpu.h"
 #include "atom.h"
 
+#include <linux/device.h>
 #include <linux/pci.h>
 #include <linux/slab.h>
 #include <linux/acpi.h>
@@ -46,52 +47,63 @@
 /* Check if current bios is an ATOM BIOS.
  * Return true if it is ATOM BIOS. Otherwise, return false.
  */
-static bool check_atom_bios(uint8_t *bios, size_t size)
+static bool check_atom_bios(struct amdgpu_device *adev, size_t size)
 {
 	uint16_t tmp, bios_header_start;
+	uint8_t *bios = adev->bios;
 
 	if (!bios || size < 0x49) {
-		DRM_INFO("vbios mem is null or mem size is wrong\n");
+		dev_dbg(adev->dev, "VBIOS mem is null or mem size is wrong\n");
 		return false;
 	}
 
 	if (!AMD_IS_VALID_VBIOS(bios)) {
-		DRM_INFO("BIOS signature incorrect %x %x\n", bios[0], bios[1]);
+		dev_dbg(adev->dev, "VBIOS signature incorrect %x %x\n", bios[0],
+			bios[1]);
 		return false;
 	}
 
 	bios_header_start = bios[0x48] | (bios[0x49] << 8);
 	if (!bios_header_start) {
-		DRM_INFO("Can't locate bios header\n");
+		dev_dbg(adev->dev, "Can't locate VBIOS header\n");
 		return false;
 	}
 
 	tmp = bios_header_start + 4;
 	if (size < tmp) {
-		DRM_INFO("BIOS header is broken\n");
+		dev_dbg(adev->dev, "VBIOS header is broken\n");
 		return false;
 	}
 
 	if (!memcmp(bios + tmp, "ATOM", 4) ||
 	    !memcmp(bios + tmp, "MOTA", 4)) {
-		DRM_DEBUG("ATOMBIOS detected\n");
+		dev_dbg(adev->dev, "ATOMBIOS detected\n");
 		return true;
 	}
 
 	return false;
 }
 
+void amdgpu_bios_release(struct amdgpu_device *adev)
+{
+	kfree(adev->bios);
+	adev->bios = NULL;
+	adev->bios_size = 0;
+}
+
 /* If you boot an IGP board with a discrete card as the primary,
  * the IGP rom is not accessible via the rom bar as the IGP rom is
  * part of the system bios.  On boot, the system bios puts a
  * copy of the igp rom at the start of vram if a discrete card is
  * present.
+ * For SR-IOV, if dynamic critical region is not enabled,
+ * the vbios image is also put at the start of VRAM in the VF.
  */
-static bool igp_read_bios_from_vram(struct amdgpu_device *adev)
+static bool amdgpu_read_bios_from_vram(struct amdgpu_device *adev)
 {
-	uint8_t __iomem *bios;
+	uint8_t __iomem *bios = NULL;
 	resource_size_t vram_base;
-	resource_size_t size = 256 * 1024; /* ??? */
+	u32 size = 256U * 1024U; /* ??? */
 
 	if (!(adev->flags & AMD_IS_APU))
 		if (amdgpu_device_need_post(adev))
@@ -103,22 +115,36 @@ static bool igp_read_bios_from_vram(struct amdgpu_device *adev)
 
 	adev->bios = NULL;
 	vram_base = pci_resource_start(adev->pdev, 0);
-	bios = ioremap_wc(vram_base, size);
-	if (!bios) {
-		return false;
-	}
 
 	adev->bios = kmalloc(size, GFP_KERNEL);
-	if (!adev->bios) {
-		iounmap(bios);
+	if (!adev->bios)
 		return false;
+
+	/* For SRIOV with dynamic critical region is enabled,
+	 * the vbios image is put at a dynamic offset of VRAM in the VF.
+	 * If dynamic critical region is disabled, follow the existing logic as on baremetal.
+	 */
+	if (amdgpu_sriov_vf(adev) && adev->virt.is_dynamic_crit_regn_enabled) {
+		if (amdgpu_virt_get_dynamic_data_info(adev,
+				AMD_SRIOV_MSG_VBIOS_IMG_TABLE_ID, adev->bios, &size)) {
+			amdgpu_bios_release(adev);
+			return false;
+		}
+	} else {
+		bios = ioremap_wc(vram_base, size);
+		if (!bios) {
+			amdgpu_bios_release(adev);
+			return false;
+		}
+
+		memcpy_fromio(adev->bios, bios, size);
+		iounmap(bios);
 	}
+
 	adev->bios_size = size;
-	memcpy_fromio(adev->bios, bios, size);
-	iounmap(bios);
 
-	if (!check_atom_bios(adev->bios, size)) {
-		kfree(adev->bios);
+	if (!check_atom_bios(adev, size)) {
+		amdgpu_bios_release(adev);
 		return false;
 	}
 
@@ -133,9 +159,8 @@ bool amdgpu_read_bios(struct amdgpu_device *adev)
 	adev->bios = NULL;
 	/* XXX: some cards may return 0 for rom size? ddx has a workaround */
 	bios = pci_map_rom(adev->pdev, &size);
-	if (!bios) {
+	if (!bios)
 		return false;
-	}
 
 	adev->bios = kzalloc(size, GFP_KERNEL);
 	if (adev->bios == NULL) {
@@ -146,8 +171,8 @@ bool amdgpu_read_bios(struct amdgpu_device *adev)
 	memcpy_fromio(adev->bios, bios, size);
 	pci_unmap_rom(adev->pdev, bios);
 
-	if (!check_atom_bios(adev->bios, size)) {
-		kfree(adev->bios);
+	if (!check_atom_bios(adev, size)) {
+		amdgpu_bios_release(adev);
 		return false;
 	}
 
@@ -168,9 +193,9 @@ static bool amdgpu_read_bios_from_rom(struct amdgpu_device *adev)
 	header[AMD_VBIOS_SIGNATURE_END] = 0;
 
 	if ((!AMD_IS_VALID_VBIOS(header)) ||
-	    0 != memcmp((char *)&header[AMD_VBIOS_SIGNATURE_OFFSET],
-			AMD_VBIOS_SIGNATURE,
-			strlen(AMD_VBIOS_SIGNATURE)))
+		memcmp((char *)&header[AMD_VBIOS_SIGNATURE_OFFSET],
+		       AMD_VBIOS_SIGNATURE,
+		       strlen(AMD_VBIOS_SIGNATURE)) != 0)
 		return false;
 
 	/* valid vbios, go on */
@@ -186,8 +211,8 @@ static bool amdgpu_read_bios_from_rom(struct amdgpu_device *adev)
 	/* read complete BIOS */
 	amdgpu_asic_read_bios_from_rom(adev, adev->bios, len);
 
-	if (!check_atom_bios(adev->bios, len)) {
-		kfree(adev->bios);
+	if (!check_atom_bios(adev, len)) {
+		amdgpu_bios_release(adev);
 		return false;
 	}
 
@@ -216,14 +241,15 @@ static bool amdgpu_read_platform_bios(struct amdgpu_device *adev)
 	memcpy_fromio(adev->bios, bios, romlen);
 	iounmap(bios);
 
-	if (!check_atom_bios(adev->bios, romlen))
+	if (!check_atom_bios(adev, romlen))
 		goto free_bios;
 
 	adev->bios_size = romlen;
 
 	return true;
 free_bios:
-	kfree(adev->bios);
+	amdgpu_bios_release(adev);
+
 	return false;
 }
 
@@ -264,7 +290,7 @@ static int amdgpu_atrm_call(acpi_handle atrm_handle, uint8_t *bios,
 
 	status = acpi_evaluate_object(atrm_handle, NULL, &atrm_arg, &buffer);
 	if (ACPI_FAILURE(status)) {
-		printk("failed to evaluate ATRM got %s\n", acpi_format_exception(status));
+		DRM_ERROR("failed to evaluate ATRM got %s\n", acpi_format_exception(status));
 		return -ENODEV;
 	}
 
@@ -285,11 +311,15 @@ static bool amdgpu_atrm_get_bios(struct amdgpu_device *adev)
 	acpi_status status;
 	bool found = false;
 
-	/* ATRM is for the discrete card only */
-	if (adev->flags & AMD_IS_APU)
+	/* ATRM is for on-platform devices only */
+	if (dev_is_removable(&adev->pdev->dev))
 		return false;
 
-	while ((pdev = pci_get_class(PCI_CLASS_DISPLAY_VGA << 8, pdev)) != NULL) {
+	while ((pdev = pci_get_base_class(PCI_BASE_CLASS_DISPLAY, pdev))) {
+		if ((pdev->class != PCI_CLASS_DISPLAY_VGA << 8) &&
+		    (pdev->class != PCI_CLASS_DISPLAY_OTHER << 8))
+			continue;
+
 		dhandle = ACPI_HANDLE(&pdev->dev);
 		if (!dhandle)
 			continue;
@@ -301,22 +331,9 @@ static bool amdgpu_atrm_get_bios(struct amdgpu_device *adev)
 		}
 	}
 
-	if (!found) {
-		while ((pdev = pci_get_class(PCI_CLASS_DISPLAY_OTHER << 8, pdev)) != NULL) {
-			dhandle = ACPI_HANDLE(&pdev->dev);
-			if (!dhandle)
-				continue;
-
-			status = acpi_get_handle(dhandle, "ATRM", &atrm_handle);
-			if (ACPI_SUCCESS(status)) {
-				found = true;
-				break;
-			}
-		}
-	}
-
 	if (!found)
 		return false;
+	pci_dev_put(pdev);
 
 	adev->bios = kmalloc(size, GFP_KERNEL);
 	if (!adev->bios) {
@@ -333,8 +350,8 @@ static bool amdgpu_atrm_get_bios(struct amdgpu_device *adev)
 			break;
 	}
 
-	if (!check_atom_bios(adev->bios, size)) {
-		kfree(adev->bios);
+	if (!check_atom_bios(adev, size)) {
+		amdgpu_bios_release(adev);
 		return false;
 	}
 	adev->bios_size = size;
@@ -349,11 +366,8 @@ static inline bool amdgpu_atrm_get_bios(struct amdgpu_device *adev)
 
 static bool amdgpu_read_disabled_bios(struct amdgpu_device *adev)
 {
-	if (adev->flags & AMD_IS_APU)
-		return igp_read_bios_from_vram(adev);
-	else
-		return (!adev->asic_funcs || !adev->asic_funcs->read_disabled_bios) ?
-			false : amdgpu_asic_read_disabled_bios(adev);
+	return (!adev->asic_funcs || !adev->asic_funcs->read_disabled_bios) ?
+		false : amdgpu_asic_read_disabled_bios(adev);
 }
 
 #ifdef CONFIG_ACPI
@@ -362,7 +376,7 @@ static bool amdgpu_acpi_vfct_bios(struct amdgpu_device *adev)
 	struct acpi_table_header *hdr;
 	acpi_size tbl_size;
 	UEFI_ACPI_VFCT *vfct;
-	unsigned offset;
+	unsigned int offset;
 
 	if (!ACPI_SUCCESS(acpi_get_table("VFCT", 1, &hdr)))
 		return false;
@@ -401,8 +415,8 @@ static bool amdgpu_acpi_vfct_bios(struct amdgpu_device *adev)
 					     vhdr->ImageLength,
 					     GFP_KERNEL);
 
-			if (!check_atom_bios(adev->bios, vhdr->ImageLength)) {
-				kfree(adev->bios);
+			if (!check_atom_bios(adev, vhdr->ImageLength)) {
+				amdgpu_bios_release(adev);
 				return false;
 			}
 			adev->bios_size = vhdr->ImageLength;
@@ -420,7 +434,43 @@ static inline bool amdgpu_acpi_vfct_bios(struct amdgpu_device *adev)
 }
 #endif
 
-bool amdgpu_get_bios(struct amdgpu_device *adev)
+static bool amdgpu_get_bios_apu(struct amdgpu_device *adev)
+{
+	if (amdgpu_acpi_vfct_bios(adev)) {
+		dev_info(adev->dev, "Fetched VBIOS from VFCT\n");
+		goto success;
+	}
+
+	if (amdgpu_read_bios_from_vram(adev)) {
+		dev_info(adev->dev, "Fetched VBIOS from VRAM BAR\n");
+		goto success;
+	}
+
+	if (amdgpu_read_bios(adev)) {
+		dev_info(adev->dev, "Fetched VBIOS from ROM BAR\n");
+		goto success;
+	}
+
+	if (amdgpu_read_platform_bios(adev)) {
+		dev_info(adev->dev, "Fetched VBIOS from platform\n");
+		goto success;
+	}
+
+	dev_err(adev->dev, "Unable to locate a BIOS ROM\n");
+	return false;
+
+success:
+	return true;
+}
+
+static bool amdgpu_prefer_rom_resource(struct amdgpu_device *adev)
+{
+	struct resource *res = &adev->pdev->resource[PCI_ROM_RESOURCE];
+
+	return (res->flags & IORESOURCE_ROM_SHADOW);
+}
+
+static bool amdgpu_get_bios_dgpu(struct amdgpu_device *adev)
 {
 	if (amdgpu_atrm_get_bios(adev)) {
 		dev_info(adev->dev, "Fetched VBIOS from ATRM\n");
@@ -432,14 +482,33 @@ bool amdgpu_get_bios(struct amdgpu_device *adev)
 		goto success;
 	}
 
-	if (igp_read_bios_from_vram(adev)) {
+	/* this is required for SR-IOV */
+	if (amdgpu_read_bios_from_vram(adev)) {
 		dev_info(adev->dev, "Fetched VBIOS from VRAM BAR\n");
 		goto success;
 	}
 
-	if (amdgpu_read_bios(adev)) {
-		dev_info(adev->dev, "Fetched VBIOS from ROM BAR\n");
-		goto success;
+	if (amdgpu_prefer_rom_resource(adev)) {
+		if (amdgpu_read_bios(adev)) {
+			dev_info(adev->dev, "Fetched VBIOS from ROM BAR\n");
+			goto success;
+		}
+
+		if (amdgpu_read_platform_bios(adev)) {
+			dev_info(adev->dev, "Fetched VBIOS from platform\n");
+			goto success;
+		}
+
+	} else {
+		if (amdgpu_read_platform_bios(adev)) {
+			dev_info(adev->dev, "Fetched VBIOS from platform\n");
+			goto success;
+		}
+
+		if (amdgpu_read_bios(adev)) {
+			dev_info(adev->dev, "Fetched VBIOS from ROM BAR\n");
+			goto success;
+		}
 	}
 
 	if (amdgpu_read_bios_from_rom(adev)) {
@@ -452,15 +521,71 @@ bool amdgpu_get_bios(struct amdgpu_device *adev)
 		goto success;
 	}
 
-	if (amdgpu_read_platform_bios(adev)) {
-		dev_info(adev->dev, "Fetched VBIOS from platform\n");
-		goto success;
-	}
-
 	dev_err(adev->dev, "Unable to locate a BIOS ROM\n");
 	return false;
 
 success:
-	adev->is_atom_fw = (adev->asic_type >= CHIP_VEGA10) ? true : false;
+	return true;
+}
+
+bool amdgpu_get_bios(struct amdgpu_device *adev)
+{
+	bool found;
+
+	if (adev->flags & AMD_IS_APU)
+		found = amdgpu_get_bios_apu(adev);
+	else
+		found = amdgpu_get_bios_dgpu(adev);
+
+	if (found)
+		adev->is_atom_fw = adev->asic_type >= CHIP_VEGA10;
+
+	return found;
+}
+
+/* helper function for soc15 and onwards to read bios from rom */
+bool amdgpu_soc15_read_bios_from_rom(struct amdgpu_device *adev,
+				     u8 *bios, u32 length_bytes)
+{
+	u32 *dw_ptr;
+	u32 i, length_dw;
+	u32 rom_offset;
+	u32 rom_index_offset;
+	u32 rom_data_offset;
+
+	if (bios == NULL)
+		return false;
+	if (length_bytes == 0)
+		return false;
+	/* APU vbios image is part of sbios image */
+	if (adev->flags & AMD_IS_APU)
+		return false;
+	if (!adev->smuio.funcs ||
+	    !adev->smuio.funcs->get_rom_index_offset ||
+	    !adev->smuio.funcs->get_rom_data_offset)
+		return false;
+
+	dw_ptr = (u32 *)bios;
+	length_dw = ALIGN(length_bytes, 4) / 4;
+
+	rom_index_offset =
+		adev->smuio.funcs->get_rom_index_offset(adev);
+	rom_data_offset =
+		adev->smuio.funcs->get_rom_data_offset(adev);
+
+	if (adev->nbio.funcs &&
+	    adev->nbio.funcs->get_rom_offset) {
+		rom_offset = adev->nbio.funcs->get_rom_offset(adev);
+		rom_offset = rom_offset << 17;
+	} else {
+		rom_offset = 0;
+	}
+
+	/* set rom index to rom_offset */
+	WREG32(rom_index_offset, rom_offset);
+	/* read out the rom data */
+	for (i = 0; i < length_dw; i++)
+		dw_ptr[i] = RREG32(rom_data_offset);
+
 	return true;
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c
index 714178f1b6c6..66fb37b64388 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c
@@ -28,6 +28,7 @@
  *    Christian König <deathsimple@vodafone.de>
  */
 
+#include <linux/sort.h>
 #include <linux/uaccess.h>
 
 #include "amdgpu.h"
@@ -40,7 +41,7 @@ static void amdgpu_bo_list_free_rcu(struct rcu_head *rcu)
 {
 	struct amdgpu_bo_list *list = container_of(rcu, struct amdgpu_bo_list,
 						   rhead);
-
+	mutex_destroy(&list->bo_list_mutex);
 	kvfree(list);
 }
 
@@ -50,13 +51,20 @@ static void amdgpu_bo_list_free(struct kref *ref)
 						   refcount);
 	struct amdgpu_bo_list_entry *e;
 
-	amdgpu_bo_list_for_each_entry(e, list) {
-		struct amdgpu_bo *bo = ttm_to_amdgpu_bo(e->tv.bo);
+	amdgpu_bo_list_for_each_entry(e, list)
+		amdgpu_bo_unref(&e->bo);
+	call_rcu(&list->rhead, amdgpu_bo_list_free_rcu);
+}
 
-		amdgpu_bo_unref(&bo);
-	}
+static int amdgpu_bo_list_entry_cmp(const void *_a, const void *_b)
+{
+	const struct amdgpu_bo_list_entry *a = _a, *b = _b;
 
-	call_rcu(&list->rhead, amdgpu_bo_list_free_rcu);
+	if (a->priority > b->priority)
+		return 1;
+	if (a->priority < b->priority)
+		return -1;
+	return 0;
 }
 
 int amdgpu_bo_list_create(struct amdgpu_device *adev, struct drm_file *filp,
@@ -67,27 +75,17 @@ int amdgpu_bo_list_create(struct amdgpu_device *adev, struct drm_file *filp,
 	struct amdgpu_bo_list_entry *array;
 	struct amdgpu_bo_list *list;
 	uint64_t total_size = 0;
-	size_t size;
 	unsigned i;
 	int r;
 
-	if (num_entries > (SIZE_MAX - sizeof(struct amdgpu_bo_list))
-				/ sizeof(struct amdgpu_bo_list_entry))
-		return -EINVAL;
-
-	size = sizeof(struct amdgpu_bo_list);
-	size += num_entries * sizeof(struct amdgpu_bo_list_entry);
-	list = kvmalloc(size, GFP_KERNEL);
+	list = kvzalloc(struct_size(list, entries, num_entries), GFP_KERNEL);
 	if (!list)
 		return -ENOMEM;
 
 	kref_init(&list->refcount);
-	list->gds_obj = NULL;
-	list->gws_obj = NULL;
-	list->oa_obj = NULL;
 
-	array = amdgpu_bo_list_array_entry(list, 0);
-	memset(array, 0, num_entries * sizeof(struct amdgpu_bo_list_entry));
+	list->num_entries = num_entries;
+	array = list->entries;
 
 	for (i = 0; i < num_entries; ++i) {
 		struct amdgpu_bo_list_entry *entry;
@@ -118,7 +116,7 @@ int amdgpu_bo_list_create(struct amdgpu_device *adev, struct drm_file *filp,
 
 		entry->priority = min(info[i].bo_priority,
 				      AMDGPU_BO_LIST_MAX_PRIORITY);
-		entry->tv.bo = &bo->tbo;
+		entry->bo = bo;
 
 		if (bo->preferred_domains == AMDGPU_GEM_DOMAIN_GDS)
 			list->gds_obj = bo;
@@ -132,24 +130,20 @@ int amdgpu_bo_list_create(struct amdgpu_device *adev, struct drm_file *filp,
 	}
 
 	list->first_userptr = first_userptr;
-	list->num_entries = num_entries;
+	sort(array, last_entry, sizeof(struct amdgpu_bo_list_entry),
+	     amdgpu_bo_list_entry_cmp, NULL);
 
 	trace_amdgpu_cs_bo_status(list->num_entries, total_size);
 
+	mutex_init(&list->bo_list_mutex);
 	*result = list;
 	return 0;
 
 error_free:
-	for (i = 0; i < last_entry; ++i) {
-		struct amdgpu_bo *bo = ttm_to_amdgpu_bo(array[i].tv.bo);
-
-		amdgpu_bo_unref(&bo);
-	}
-	for (i = first_userptr; i < num_entries; ++i) {
-		struct amdgpu_bo *bo = ttm_to_amdgpu_bo(array[i].tv.bo);
-
-		amdgpu_bo_unref(&bo);
-	}
+	for (i = 0; i < last_entry; ++i)
+		amdgpu_bo_unref(&array[i].bo);
+	for (i = first_userptr; i < num_entries; ++i)
+		amdgpu_bo_unref(&array[i].bo);
 	kvfree(list);
 	return r;
 
@@ -178,43 +172,10 @@ int amdgpu_bo_list_get(struct amdgpu_fpriv *fpriv, int id,
 	}
 
 	rcu_read_unlock();
+	*result = NULL;
 	return -ENOENT;
 }
 
-void amdgpu_bo_list_get_list(struct amdgpu_bo_list *list,
-			     struct list_head *validated)
-{
-	/* This is based on the bucket sort with O(n) time complexity.
-	 * An item with priority "i" is added to bucket[i]. The lists are then
-	 * concatenated in descending order.
-	 */
-	struct list_head bucket[AMDGPU_BO_LIST_NUM_BUCKETS];
-	struct amdgpu_bo_list_entry *e;
-	unsigned i;
-
-	for (i = 0; i < AMDGPU_BO_LIST_NUM_BUCKETS; i++)
-		INIT_LIST_HEAD(&bucket[i]);
-
-	/* Since buffers which appear sooner in the relocation list are
-	 * likely to be used more often than buffers which appear later
-	 * in the list, the sort mustn't change the ordering of buffers
-	 * with the same priority, i.e. it must be stable.
-	 */
-	amdgpu_bo_list_for_each_entry(e, list) {
-		struct amdgpu_bo *bo = ttm_to_amdgpu_bo(e->tv.bo);
-		unsigned priority = e->priority;
-
-		if (!bo->parent)
-			list_add_tail(&e->tv.head, &bucket[priority]);
-
-		e->user_pages = NULL;
-	}
-
-	/* Connect the sorted buckets in the output list. */
-	for (i = 0; i < AMDGPU_BO_LIST_NUM_BUCKETS; i++)
-		list_splice(&bucket[i], validated);
-}
-
 void amdgpu_bo_list_put(struct amdgpu_bo_list *list)
 {
 	kref_put(&list->refcount, amdgpu_bo_list_free);
@@ -223,43 +184,36 @@ void amdgpu_bo_list_put(struct amdgpu_bo_list *list)
 int amdgpu_bo_create_list_entry_array(struct drm_amdgpu_bo_list_in *in,
 				      struct drm_amdgpu_bo_list_entry **info_param)
 {
-	const void __user *uptr = u64_to_user_ptr(in->bo_info_ptr);
 	const uint32_t info_size = sizeof(struct drm_amdgpu_bo_list_entry);
+	const void __user *uptr = u64_to_user_ptr(in->bo_info_ptr);
+	const uint32_t bo_info_size = in->bo_info_size;
+	const uint32_t bo_number = in->bo_number;
 	struct drm_amdgpu_bo_list_entry *info;
-	int r;
-
-	info = kvmalloc_array(in->bo_number, info_size, GFP_KERNEL);
-	if (!info)
-		return -ENOMEM;
 
 	/* copy the handle array from userspace to a kernel buffer */
-	r = -EFAULT;
-	if (likely(info_size == in->bo_info_size)) {
-		unsigned long bytes = in->bo_number *
-			in->bo_info_size;
-
-		if (copy_from_user(info, uptr, bytes))
-			goto error_free;
-
+	if (likely(info_size == bo_info_size)) {
+		info = vmemdup_array_user(uptr, bo_number, info_size);
+		if (IS_ERR(info))
+			return PTR_ERR(info);
 	} else {
-		unsigned long bytes = min(in->bo_info_size, info_size);
+		const uint32_t bytes = min(bo_info_size, info_size);
 		unsigned i;
 
-		memset(info, 0, in->bo_number * info_size);
-		for (i = 0; i < in->bo_number; ++i) {
-			if (copy_from_user(&info[i], uptr, bytes))
-				goto error_free;
+		info = kvmalloc_array(bo_number, info_size, GFP_KERNEL);
+		if (!info)
+			return -ENOMEM;
 
-			uptr += in->bo_info_size;
+		memset(info, 0, bo_number * info_size);
+		for (i = 0; i < bo_number; ++i, uptr += bo_info_size) {
+			if (copy_from_user(&info[i], uptr, bytes)) {
+				kvfree(info);
+				return -EFAULT;
+			}
 		}
 	}
 
 	*info_param = info;
 	return 0;
-
-error_free:
-	kvfree(info);
-	return r;
 }
 
 int amdgpu_bo_list_ioctl(struct drm_device *dev, void *data,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.h
index 044b41f0bfd9..2b5e7c46a39d 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.h
@@ -23,20 +23,22 @@
 #ifndef __AMDGPU_BO_LIST_H__
 #define __AMDGPU_BO_LIST_H__
 
-#include <drm/ttm/ttm_execbuf_util.h>
 #include <drm/amdgpu_drm.h>
 
+struct hmm_range;
+
+struct drm_file;
+
 struct amdgpu_device;
 struct amdgpu_bo;
 struct amdgpu_bo_va;
 struct amdgpu_fpriv;
 
 struct amdgpu_bo_list_entry {
-	struct ttm_validate_buffer	tv;
+	struct amdgpu_bo		*bo;
 	struct amdgpu_bo_va		*bo_va;
-	struct dma_fence_chain		*chain;
 	uint32_t			priority;
-	struct page			**user_pages;
+	struct amdgpu_hmm_range		*range;
 	bool				user_invalidated;
 };
 
@@ -48,12 +50,16 @@ struct amdgpu_bo_list {
 	struct amdgpu_bo *oa_obj;
 	unsigned first_userptr;
 	unsigned num_entries;
+
+	/* Protect access during command submission.
+	 */
+	struct mutex bo_list_mutex;
+
+	struct amdgpu_bo_list_entry entries[] __counted_by(num_entries);
 };
 
 int amdgpu_bo_list_get(struct amdgpu_fpriv *fpriv, int id,
 		       struct amdgpu_bo_list **result);
-void amdgpu_bo_list_get_list(struct amdgpu_bo_list *list,
-			     struct list_head *validated);
 void amdgpu_bo_list_put(struct amdgpu_bo_list *list);
 int amdgpu_bo_create_list_entry_array(struct drm_amdgpu_bo_list_in *in,
 				      struct drm_amdgpu_bo_list_entry **info_param);
@@ -64,22 +70,14 @@ int amdgpu_bo_list_create(struct amdgpu_device *adev,
 				 size_t num_entries,
 				 struct amdgpu_bo_list **list);
 
-static inline struct amdgpu_bo_list_entry *
-amdgpu_bo_list_array_entry(struct amdgpu_bo_list *list, unsigned index)
-{
-	struct amdgpu_bo_list_entry *array = (void *)&list[1];
-
-	return &array[index];
-}
-
 #define amdgpu_bo_list_for_each_entry(e, list) \
-	for (e = amdgpu_bo_list_array_entry(list, 0); \
-	     e != amdgpu_bo_list_array_entry(list, (list)->num_entries); \
+	for (e = list->entries; \
+	     e != &list->entries[list->num_entries]; \
 	     ++e)
 
 #define amdgpu_bo_list_for_each_userptr_entry(e, list) \
-	for (e = amdgpu_bo_list_array_entry(list, (list)->first_userptr); \
-	     e != amdgpu_bo_list_array_entry(list, (list)->num_entries); \
+	for (e = &list->entries[list->first_userptr]; \
+	     e != &list->entries[list->num_entries]; \
 	     ++e)
 
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c
index f1a050379190..004a6a9d6b9f 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c
@@ -41,13 +41,13 @@ struct amdgpu_cgs_device {
 		((struct amdgpu_cgs_device *)cgs_device)->adev
 
 
-static uint32_t amdgpu_cgs_read_register(struct cgs_device *cgs_device, unsigned offset)
+static uint32_t amdgpu_cgs_read_register(struct cgs_device *cgs_device, unsigned int offset)
 {
 	CGS_FUNC_ADEV;
 	return RREG32(offset);
 }
 
-static void amdgpu_cgs_write_register(struct cgs_device *cgs_device, unsigned offset,
+static void amdgpu_cgs_write_register(struct cgs_device *cgs_device, unsigned int offset,
 				      uint32_t value)
 {
 	CGS_FUNC_ADEV;
@@ -56,7 +56,7 @@ static void amdgpu_cgs_write_register(struct cgs_device *cgs_device, unsigned of
 
 static uint32_t amdgpu_cgs_read_ind_register(struct cgs_device *cgs_device,
 					     enum cgs_ind_reg space,
-					     unsigned index)
+					     unsigned int index)
 {
 	CGS_FUNC_ADEV;
 	switch (space) {
@@ -84,7 +84,7 @@ static uint32_t amdgpu_cgs_read_ind_register(struct cgs_device *cgs_device,
 
 static void amdgpu_cgs_write_ind_register(struct cgs_device *cgs_device,
 					  enum cgs_ind_reg space,
-					  unsigned index, uint32_t value)
+					  unsigned int index, uint32_t value)
 {
 	CGS_FUNC_ADEV;
 	switch (space) {
@@ -163,38 +163,38 @@ static uint16_t amdgpu_get_firmware_version(struct cgs_device *cgs_device,
 	uint16_t fw_version = 0;
 
 	switch (type) {
-		case CGS_UCODE_ID_SDMA0:
-			fw_version = adev->sdma.instance[0].fw_version;
-			break;
-		case CGS_UCODE_ID_SDMA1:
-			fw_version = adev->sdma.instance[1].fw_version;
-			break;
-		case CGS_UCODE_ID_CP_CE:
-			fw_version = adev->gfx.ce_fw_version;
-			break;
-		case CGS_UCODE_ID_CP_PFP:
-			fw_version = adev->gfx.pfp_fw_version;
-			break;
-		case CGS_UCODE_ID_CP_ME:
-			fw_version = adev->gfx.me_fw_version;
-			break;
-		case CGS_UCODE_ID_CP_MEC:
-			fw_version = adev->gfx.mec_fw_version;
-			break;
-		case CGS_UCODE_ID_CP_MEC_JT1:
-			fw_version = adev->gfx.mec_fw_version;
-			break;
-		case CGS_UCODE_ID_CP_MEC_JT2:
-			fw_version = adev->gfx.mec_fw_version;
-			break;
-		case CGS_UCODE_ID_RLC_G:
-			fw_version = adev->gfx.rlc_fw_version;
-			break;
-		case CGS_UCODE_ID_STORAGE:
-			break;
-		default:
-			DRM_ERROR("firmware type %d do not have version\n", type);
-			break;
+	case CGS_UCODE_ID_SDMA0:
+		fw_version = adev->sdma.instance[0].fw_version;
+		break;
+	case CGS_UCODE_ID_SDMA1:
+		fw_version = adev->sdma.instance[1].fw_version;
+		break;
+	case CGS_UCODE_ID_CP_CE:
+		fw_version = adev->gfx.ce_fw_version;
+		break;
+	case CGS_UCODE_ID_CP_PFP:
+		fw_version = adev->gfx.pfp_fw_version;
+		break;
+	case CGS_UCODE_ID_CP_ME:
+		fw_version = adev->gfx.me_fw_version;
+		break;
+	case CGS_UCODE_ID_CP_MEC:
+		fw_version = adev->gfx.mec_fw_version;
+		break;
+	case CGS_UCODE_ID_CP_MEC_JT1:
+		fw_version = adev->gfx.mec_fw_version;
+		break;
+	case CGS_UCODE_ID_CP_MEC_JT2:
+		fw_version = adev->gfx.mec_fw_version;
+		break;
+	case CGS_UCODE_ID_RLC_G:
+		fw_version = adev->gfx.rlc_fw_version;
+		break;
+	case CGS_UCODE_ID_STORAGE:
+		break;
+	default:
+		DRM_ERROR("firmware type %d do not have version\n", type);
+		break;
 	}
 	return fw_version;
 }
@@ -205,7 +205,7 @@ static int amdgpu_cgs_get_firmware_info(struct cgs_device *cgs_device,
 {
 	CGS_FUNC_ADEV;
 
-	if ((CGS_UCODE_ID_SMU != type) && (CGS_UCODE_ID_SMU_SK != type)) {
+	if (type != CGS_UCODE_ID_SMU && type != CGS_UCODE_ID_SMU_SK) {
 		uint64_t gpu_addr;
 		uint32_t data_size;
 		const struct gfx_firmware_header_v1_0 *header;
@@ -213,6 +213,9 @@ static int amdgpu_cgs_get_firmware_info(struct cgs_device *cgs_device,
 		struct amdgpu_firmware_info *ucode;
 
 		id = fw_type_convert(cgs_device, type);
+		if (id >= AMDGPU_UCODE_ID_MAXIMUM)
+			return -EINVAL;
+
 		ucode = &adev->firmware.ucode[id];
 		if (ucode->fw == NULL)
 			return -EINVAL;
@@ -232,7 +235,7 @@ static int amdgpu_cgs_get_firmware_info(struct cgs_device *cgs_device,
 		info->mc_addr = gpu_addr;
 		info->version = (uint16_t)le32_to_cpu(header->header.ucode_version);
 
-		if (CGS_UCODE_ID_CP_MEC == type)
+		if (type == CGS_UCODE_ID_CP_MEC)
 			info->image_size = le32_to_cpu(header->jt_offset) << 2;
 
 		info->fw_version = amdgpu_get_firmware_version(cgs_device, type);
@@ -249,83 +252,22 @@ static int amdgpu_cgs_get_firmware_info(struct cgs_device *cgs_device,
 
 		if (!adev->pm.fw) {
 			switch (adev->asic_type) {
-			case CHIP_TAHITI:
-				strcpy(fw_name, "radeon/tahiti_smc.bin");
-				break;
-			case CHIP_PITCAIRN:
-				if ((adev->pdev->revision == 0x81) &&
-				    ((adev->pdev->device == 0x6810) ||
-				    (adev->pdev->device == 0x6811))) {
-					info->is_kicker = true;
-					strcpy(fw_name, "radeon/pitcairn_k_smc.bin");
-				} else {
-					strcpy(fw_name, "radeon/pitcairn_smc.bin");
-				}
-				break;
-			case CHIP_VERDE:
-				if (((adev->pdev->device == 0x6820) &&
-					((adev->pdev->revision == 0x81) ||
-					(adev->pdev->revision == 0x83))) ||
-				    ((adev->pdev->device == 0x6821) &&
-					((adev->pdev->revision == 0x83) ||
-					(adev->pdev->revision == 0x87))) ||
-				    ((adev->pdev->revision == 0x87) &&
-					((adev->pdev->device == 0x6823) ||
-					(adev->pdev->device == 0x682b)))) {
-					info->is_kicker = true;
-					strcpy(fw_name, "radeon/verde_k_smc.bin");
-				} else {
-					strcpy(fw_name, "radeon/verde_smc.bin");
-				}
-				break;
-			case CHIP_OLAND:
-				if (((adev->pdev->revision == 0x81) &&
-					((adev->pdev->device == 0x6600) ||
-					(adev->pdev->device == 0x6604) ||
-					(adev->pdev->device == 0x6605) ||
-					(adev->pdev->device == 0x6610))) ||
-				    ((adev->pdev->revision == 0x83) &&
-					(adev->pdev->device == 0x6610))) {
-					info->is_kicker = true;
-					strcpy(fw_name, "radeon/oland_k_smc.bin");
-				} else {
-					strcpy(fw_name, "radeon/oland_smc.bin");
-				}
-				break;
-			case CHIP_HAINAN:
-				if (((adev->pdev->revision == 0x81) &&
-					(adev->pdev->device == 0x6660)) ||
-				    ((adev->pdev->revision == 0x83) &&
-					((adev->pdev->device == 0x6660) ||
-					(adev->pdev->device == 0x6663) ||
-					(adev->pdev->device == 0x6665) ||
-					 (adev->pdev->device == 0x6667)))) {
-					info->is_kicker = true;
-					strcpy(fw_name, "radeon/hainan_k_smc.bin");
-				} else if ((adev->pdev->revision == 0xc3) &&
-					 (adev->pdev->device == 0x6665)) {
-					info->is_kicker = true;
-					strcpy(fw_name, "radeon/banks_k_2_smc.bin");
-				} else {
-					strcpy(fw_name, "radeon/hainan_smc.bin");
-				}
-				break;
 			case CHIP_BONAIRE:
 				if ((adev->pdev->revision == 0x80) ||
 					(adev->pdev->revision == 0x81) ||
 					(adev->pdev->device == 0x665f)) {
 					info->is_kicker = true;
-					strcpy(fw_name, "amdgpu/bonaire_k_smc.bin");
+					strscpy(fw_name, "amdgpu/bonaire_k_smc.bin");
 				} else {
-					strcpy(fw_name, "amdgpu/bonaire_smc.bin");
+					strscpy(fw_name, "amdgpu/bonaire_smc.bin");
 				}
 				break;
 			case CHIP_HAWAII:
 				if (adev->pdev->revision == 0x80) {
 					info->is_kicker = true;
-					strcpy(fw_name, "amdgpu/hawaii_k_smc.bin");
+					strscpy(fw_name, "amdgpu/hawaii_k_smc.bin");
 				} else {
-					strcpy(fw_name, "amdgpu/hawaii_smc.bin");
+					strscpy(fw_name, "amdgpu/hawaii_smc.bin");
 				}
 				break;
 			case CHIP_TOPAZ:
@@ -335,93 +277,88 @@ static int amdgpu_cgs_get_firmware_info(struct cgs_device *cgs_device,
 				    ((adev->pdev->device == 0x6900) && (adev->pdev->revision == 0xD1)) ||
 				    ((adev->pdev->device == 0x6900) && (adev->pdev->revision == 0xD3))) {
 					info->is_kicker = true;
-					strcpy(fw_name, "amdgpu/topaz_k_smc.bin");
+					strscpy(fw_name, "amdgpu/topaz_k_smc.bin");
 				} else
-					strcpy(fw_name, "amdgpu/topaz_smc.bin");
+					strscpy(fw_name, "amdgpu/topaz_smc.bin");
 				break;
 			case CHIP_TONGA:
 				if (((adev->pdev->device == 0x6939) && (adev->pdev->revision == 0xf1)) ||
 				    ((adev->pdev->device == 0x6938) && (adev->pdev->revision == 0xf1))) {
 					info->is_kicker = true;
-					strcpy(fw_name, "amdgpu/tonga_k_smc.bin");
+					strscpy(fw_name, "amdgpu/tonga_k_smc.bin");
 				} else
-					strcpy(fw_name, "amdgpu/tonga_smc.bin");
+					strscpy(fw_name, "amdgpu/tonga_smc.bin");
 				break;
 			case CHIP_FIJI:
-				strcpy(fw_name, "amdgpu/fiji_smc.bin");
+				strscpy(fw_name, "amdgpu/fiji_smc.bin");
 				break;
 			case CHIP_POLARIS11:
 				if (type == CGS_UCODE_ID_SMU) {
 					if (ASICID_IS_P21(adev->pdev->device, adev->pdev->revision)) {
 						info->is_kicker = true;
-						strcpy(fw_name, "amdgpu/polaris11_k_smc.bin");
+						strscpy(fw_name, "amdgpu/polaris11_k_smc.bin");
 					} else if (ASICID_IS_P31(adev->pdev->device, adev->pdev->revision)) {
 						info->is_kicker = true;
-						strcpy(fw_name, "amdgpu/polaris11_k2_smc.bin");
+						strscpy(fw_name, "amdgpu/polaris11_k2_smc.bin");
 					} else {
-						strcpy(fw_name, "amdgpu/polaris11_smc.bin");
+						strscpy(fw_name, "amdgpu/polaris11_smc.bin");
 					}
 				} else if (type == CGS_UCODE_ID_SMU_SK) {
-					strcpy(fw_name, "amdgpu/polaris11_smc_sk.bin");
+					strscpy(fw_name, "amdgpu/polaris11_smc_sk.bin");
 				}
 				break;
 			case CHIP_POLARIS10:
 				if (type == CGS_UCODE_ID_SMU) {
 					if (ASICID_IS_P20(adev->pdev->device, adev->pdev->revision)) {
 						info->is_kicker = true;
-						strcpy(fw_name, "amdgpu/polaris10_k_smc.bin");
+						strscpy(fw_name, "amdgpu/polaris10_k_smc.bin");
 					} else if (ASICID_IS_P30(adev->pdev->device, adev->pdev->revision)) {
 						info->is_kicker = true;
-						strcpy(fw_name, "amdgpu/polaris10_k2_smc.bin");
+						strscpy(fw_name, "amdgpu/polaris10_k2_smc.bin");
 					} else {
-						strcpy(fw_name, "amdgpu/polaris10_smc.bin");
+						strscpy(fw_name, "amdgpu/polaris10_smc.bin");
 					}
 				} else if (type == CGS_UCODE_ID_SMU_SK) {
-					strcpy(fw_name, "amdgpu/polaris10_smc_sk.bin");
+					strscpy(fw_name, "amdgpu/polaris10_smc_sk.bin");
 				}
 				break;
 			case CHIP_POLARIS12:
 				if (ASICID_IS_P23(adev->pdev->device, adev->pdev->revision)) {
 					info->is_kicker = true;
-					strcpy(fw_name, "amdgpu/polaris12_k_smc.bin");
+					strscpy(fw_name, "amdgpu/polaris12_k_smc.bin");
 				} else {
-					strcpy(fw_name, "amdgpu/polaris12_smc.bin");
+					strscpy(fw_name, "amdgpu/polaris12_smc.bin");
 				}
 				break;
 			case CHIP_VEGAM:
-				strcpy(fw_name, "amdgpu/vegam_smc.bin");
+				strscpy(fw_name, "amdgpu/vegam_smc.bin");
 				break;
 			case CHIP_VEGA10:
 				if ((adev->pdev->device == 0x687f) &&
 					((adev->pdev->revision == 0xc0) ||
 					(adev->pdev->revision == 0xc1) ||
 					(adev->pdev->revision == 0xc3)))
-					strcpy(fw_name, "amdgpu/vega10_acg_smc.bin");
+					strscpy(fw_name, "amdgpu/vega10_acg_smc.bin");
 				else
-					strcpy(fw_name, "amdgpu/vega10_smc.bin");
+					strscpy(fw_name, "amdgpu/vega10_smc.bin");
 				break;
 			case CHIP_VEGA12:
-				strcpy(fw_name, "amdgpu/vega12_smc.bin");
+				strscpy(fw_name, "amdgpu/vega12_smc.bin");
 				break;
 			case CHIP_VEGA20:
-				strcpy(fw_name, "amdgpu/vega20_smc.bin");
+				strscpy(fw_name, "amdgpu/vega20_smc.bin");
 				break;
 			default:
 				DRM_ERROR("SMC firmware not supported\n");
 				return -EINVAL;
 			}
 
-			err = request_firmware(&adev->pm.fw, fw_name, adev->dev);
-			if (err) {
-				DRM_ERROR("Failed to request firmware\n");
-				return err;
-			}
-
-			err = amdgpu_ucode_validate(adev->pm.fw);
+			err = amdgpu_ucode_request(adev, &adev->pm.fw,
+						   AMDGPU_UCODE_REQUIRED,
+						   "%s", fw_name);
 			if (err) {
 				DRM_ERROR("Failed to load firmware \"%s\"", fw_name);
-				release_firmware(adev->pm.fw);
-				adev->pm.fw = NULL;
+				amdgpu_ucode_release(&adev->pm.fw);
 				return err;
 			}
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.c
index 0de66f59adb8..9f96d568acf2 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.c
@@ -24,9 +24,10 @@
  *          Alex Deucher
  */
 
+#include <drm/display/drm_dp_helper.h>
+#include <drm/drm_crtc_helper.h>
 #include <drm/drm_edid.h>
-#include <drm/drm_fb_helper.h>
-#include <drm/drm_dp_helper.h>
+#include <drm/drm_modeset_helper_vtables.h>
 #include <drm/drm_probe_helper.h>
 #include <drm/amdgpu_drm.h>
 #include "amdgpu.h"
@@ -102,13 +103,13 @@ int amdgpu_connector_get_monitor_bpc(struct drm_connector *connector)
 	struct amdgpu_connector *amdgpu_connector = to_amdgpu_connector(connector);
 	struct amdgpu_connector_atom_dig *dig_connector;
 	int bpc = 8;
-	unsigned mode_clock, max_tmds_clock;
+	unsigned int mode_clock, max_tmds_clock;
 
 	switch (connector->connector_type) {
 	case DRM_MODE_CONNECTOR_DVII:
 	case DRM_MODE_CONNECTOR_HDMIB:
 		if (amdgpu_connector->use_digital) {
-			if (drm_detect_hdmi_monitor(amdgpu_connector_edid(connector))) {
+			if (connector->display_info.is_hdmi) {
 				if (connector->display_info.bpc)
 					bpc = connector->display_info.bpc;
 			}
@@ -116,7 +117,7 @@ int amdgpu_connector_get_monitor_bpc(struct drm_connector *connector)
 		break;
 	case DRM_MODE_CONNECTOR_DVID:
 	case DRM_MODE_CONNECTOR_HDMIA:
-		if (drm_detect_hdmi_monitor(amdgpu_connector_edid(connector))) {
+		if (connector->display_info.is_hdmi) {
 			if (connector->display_info.bpc)
 				bpc = connector->display_info.bpc;
 		}
@@ -125,7 +126,7 @@ int amdgpu_connector_get_monitor_bpc(struct drm_connector *connector)
 		dig_connector = amdgpu_connector->con_priv;
 		if ((dig_connector->dp_sink_type == CONNECTOR_OBJECT_ID_DISPLAYPORT) ||
 		    (dig_connector->dp_sink_type == CONNECTOR_OBJECT_ID_eDP) ||
-		    drm_detect_hdmi_monitor(amdgpu_connector_edid(connector))) {
+		    connector->display_info.is_hdmi) {
 			if (connector->display_info.bpc)
 				bpc = connector->display_info.bpc;
 		}
@@ -149,7 +150,7 @@ int amdgpu_connector_get_monitor_bpc(struct drm_connector *connector)
 		break;
 	}
 
-	if (drm_detect_hdmi_monitor(amdgpu_connector_edid(connector))) {
+	if (connector->display_info.is_hdmi) {
 		/*
 		 * Pre DCE-8 hw can't handle > 12 bpc, and more than 12 bpc doesn't make
 		 * much sense without support for > 12 bpc framebuffers. RGB 4:4:4 at
@@ -175,7 +176,7 @@ int amdgpu_connector_get_monitor_bpc(struct drm_connector *connector)
 
 			/* Check if bpc is within clock limit. Try to degrade gracefully otherwise */
 			if ((bpc == 12) && (mode_clock * 3/2 > max_tmds_clock)) {
-				if ((connector->display_info.edid_hdmi_dc_modes & DRM_EDID_HDMI_DC_30) &&
+				if ((connector->display_info.edid_hdmi_rgb444_dc_modes & DRM_EDID_HDMI_DC_30) &&
 				    (mode_clock * 5/4 <= max_tmds_clock))
 					bpc = 10;
 				else
@@ -245,36 +246,10 @@ amdgpu_connector_find_encoder(struct drm_connector *connector,
 	return NULL;
 }
 
-struct edid *amdgpu_connector_edid(struct drm_connector *connector)
-{
-	struct amdgpu_connector *amdgpu_connector = to_amdgpu_connector(connector);
-	struct drm_property_blob *edid_blob = connector->edid_blob_ptr;
-
-	if (amdgpu_connector->edid) {
-		return amdgpu_connector->edid;
-	} else if (edid_blob) {
-		struct edid *edid = kmemdup(edid_blob->data, edid_blob->length, GFP_KERNEL);
-		if (edid)
-			amdgpu_connector->edid = edid;
-	}
-	return amdgpu_connector->edid;
-}
-
 static struct edid *
 amdgpu_connector_get_hardcoded_edid(struct amdgpu_device *adev)
 {
-	struct edid *edid;
-
-	if (adev->mode_info.bios_hardcoded_edid) {
-		edid = kmalloc(adev->mode_info.bios_hardcoded_edid_size, GFP_KERNEL);
-		if (edid) {
-			memcpy((unsigned char *)edid,
-			       (unsigned char *)adev->mode_info.bios_hardcoded_edid,
-			       adev->mode_info.bios_hardcoded_edid_size);
-			return edid;
-		}
-	}
-	return NULL;
+	return drm_edid_duplicate(drm_edid_raw(adev->mode_info.bios_hardcoded_edid));
 }
 
 static void amdgpu_connector_get_edid(struct drm_connector *connector)
@@ -315,8 +290,10 @@ static void amdgpu_connector_get_edid(struct drm_connector *connector)
 	if (!amdgpu_connector->edid) {
 		/* some laptops provide a hardcoded edid in rom for LCDs */
 		if (((connector->connector_type == DRM_MODE_CONNECTOR_LVDS) ||
-		     (connector->connector_type == DRM_MODE_CONNECTOR_eDP)))
+		     (connector->connector_type == DRM_MODE_CONNECTOR_eDP))) {
 			amdgpu_connector->edid = amdgpu_connector_get_hardcoded_edid(adev);
+			drm_connector_update_edid_property(connector, amdgpu_connector->edid);
+		}
 	}
 }
 
@@ -387,6 +364,9 @@ amdgpu_connector_lcd_native_mode(struct drm_encoder *encoder)
 	    native_mode->vdisplay != 0 &&
 	    native_mode->clock != 0) {
 		mode = drm_mode_duplicate(dev, native_mode);
+		if (!mode)
+			return NULL;
+
 		mode->type = DRM_MODE_TYPE_PREFERRED | DRM_MODE_TYPE_DRIVER;
 		drm_mode_set_name(mode);
 
@@ -401,6 +381,9 @@ amdgpu_connector_lcd_native_mode(struct drm_encoder *encoder)
 		 * simpler.
 		 */
 		mode = drm_cvt_mode(dev, native_mode->hdisplay, native_mode->vdisplay, 60, true, false, false);
+		if (!mode)
+			return NULL;
+
 		mode->type = DRM_MODE_TYPE_PREFERRED | DRM_MODE_TYPE_DRIVER;
 		DRM_DEBUG_KMS("Adding cvt approximation of native panel mode %s\n", mode->name);
 	}
@@ -415,30 +398,28 @@ static void amdgpu_connector_add_common_modes(struct drm_encoder *encoder,
 	struct drm_display_mode *mode = NULL;
 	struct drm_display_mode *native_mode = &amdgpu_encoder->native_mode;
 	int i;
-	static const struct mode_size {
+	int n;
+	struct mode_size {
+		char name[DRM_DISPLAY_MODE_LEN];
 		int w;
 		int h;
-	} common_modes[17] = {
-		{ 640,  480},
-		{ 720,  480},
-		{ 800,  600},
-		{ 848,  480},
-		{1024,  768},
-		{1152,  768},
-		{1280,  720},
-		{1280,  800},
-		{1280,  854},
-		{1280,  960},
-		{1280, 1024},
-		{1440,  900},
-		{1400, 1050},
-		{1680, 1050},
-		{1600, 1200},
-		{1920, 1080},
-		{1920, 1200}
+	} common_modes[] = {
+		{  "640x480",  640,  480},
+		{  "800x600",  800,  600},
+		{ "1024x768", 1024,  768},
+		{ "1280x720", 1280,  720},
+		{ "1280x800", 1280,  800},
+		{"1280x1024", 1280, 1024},
+		{ "1440x900", 1440,  900},
+		{"1680x1050", 1680, 1050},
+		{"1600x1200", 1600, 1200},
+		{"1920x1080", 1920, 1080},
+		{"1920x1200", 1920, 1200}
 	};
 
-	for (i = 0; i < 17; i++) {
+	n = ARRAY_SIZE(common_modes);
+
+	for (i = 0; i < n; i++) {
 		if (amdgpu_encoder->devices & (ATOM_DEVICE_TV_SUPPORT)) {
 			if (common_modes[i].w > 1024 ||
 			    common_modes[i].h > 768)
@@ -451,10 +432,12 @@ static void amdgpu_connector_add_common_modes(struct drm_encoder *encoder,
 			     common_modes[i].h == native_mode->vdisplay))
 				continue;
 		}
-		if (common_modes[i].w < 320 || common_modes[i].h < 200)
-			continue;
 
 		mode = drm_cvt_mode(dev, common_modes[i].w, common_modes[i].h, 60, false, false, false);
+		if (!mode)
+			return;
+		strscpy(mode->name, common_modes[i].name, DRM_DISPLAY_MODE_LEN);
+
 		drm_mode_probed_add(connector, mode);
 	}
 }
@@ -579,16 +562,26 @@ static int amdgpu_connector_set_property(struct drm_connector *connector,
 			amdgpu_encoder = to_amdgpu_encoder(connector->encoder);
 		} else {
 			const struct drm_connector_helper_funcs *connector_funcs = connector->helper_private;
+
 			amdgpu_encoder = to_amdgpu_encoder(connector_funcs->best_encoder(connector));
 		}
 
 		switch (val) {
 		default:
-		case DRM_MODE_SCALE_NONE: rmx_type = RMX_OFF; break;
-		case DRM_MODE_SCALE_CENTER: rmx_type = RMX_CENTER; break;
-		case DRM_MODE_SCALE_ASPECT: rmx_type = RMX_ASPECT; break;
-		case DRM_MODE_SCALE_FULLSCREEN: rmx_type = RMX_FULL; break;
+		case DRM_MODE_SCALE_NONE:
+			rmx_type = RMX_OFF;
+			break;
+		case DRM_MODE_SCALE_CENTER:
+			rmx_type = RMX_CENTER;
+			break;
+		case DRM_MODE_SCALE_ASPECT:
+			rmx_type = RMX_ASPECT;
+			break;
+		case DRM_MODE_SCALE_FULLSCREEN:
+			rmx_type = RMX_FULL;
+			break;
 		}
+
 		if (amdgpu_encoder->rmx_type == rmx_type)
 			return 0;
 
@@ -617,7 +610,7 @@ amdgpu_connector_fixup_lcd_native_mode(struct drm_encoder *encoder,
 		if (mode->type & DRM_MODE_TYPE_PREFERRED) {
 			if (mode->hdisplay != native_mode->hdisplay ||
 			    mode->vdisplay != native_mode->vdisplay)
-				memcpy(native_mode, mode, sizeof(*mode));
+				drm_mode_copy(native_mode, mode);
 		}
 	}
 
@@ -626,7 +619,7 @@ amdgpu_connector_fixup_lcd_native_mode(struct drm_encoder *encoder,
 		list_for_each_entry_safe(mode, t, &connector->probed_modes, head) {
 			if (mode->hdisplay == native_mode->hdisplay &&
 			    mode->vdisplay == native_mode->vdisplay) {
-				*native_mode = *mode;
+				drm_mode_copy(native_mode, mode);
 				drm_mode_set_crtcinfo(native_mode, CRTC_INTERLACE_HALVE_V);
 				DRM_DEBUG_KMS("Determined LVDS native mode details from EDID\n");
 				break;
@@ -678,7 +671,7 @@ static int amdgpu_connector_lvds_get_modes(struct drm_connector *connector)
 }
 
 static enum drm_mode_status amdgpu_connector_lvds_mode_valid(struct drm_connector *connector,
-					     struct drm_display_mode *mode)
+					     const struct drm_display_mode *mode)
 {
 	struct drm_encoder *encoder = amdgpu_connector_best_single_encoder(connector);
 
@@ -741,10 +734,8 @@ amdgpu_connector_lvds_detect(struct drm_connector *connector, bool force)
 
 	amdgpu_connector_update_scratch_regs(connector, ret);
 
-	if (!drm_kms_helper_is_poll_worker()) {
-		pm_runtime_mark_last_busy(connector->dev->dev);
+	if (!drm_kms_helper_is_poll_worker())
 		pm_runtime_put_autosuspend(connector->dev->dev);
-	}
 
 	return ret;
 }
@@ -786,16 +777,26 @@ static int amdgpu_connector_set_lcd_property(struct drm_connector *connector,
 		amdgpu_encoder = to_amdgpu_encoder(connector->encoder);
 	else {
 		const struct drm_connector_helper_funcs *connector_funcs = connector->helper_private;
+
 		amdgpu_encoder = to_amdgpu_encoder(connector_funcs->best_encoder(connector));
 	}
 
 	switch (value) {
-	case DRM_MODE_SCALE_NONE: rmx_type = RMX_OFF; break;
-	case DRM_MODE_SCALE_CENTER: rmx_type = RMX_CENTER; break;
-	case DRM_MODE_SCALE_ASPECT: rmx_type = RMX_ASPECT; break;
+	case DRM_MODE_SCALE_NONE:
+		rmx_type = RMX_OFF;
+		break;
+	case DRM_MODE_SCALE_CENTER:
+		rmx_type = RMX_CENTER;
+		break;
+	case DRM_MODE_SCALE_ASPECT:
+		rmx_type = RMX_ASPECT;
+		break;
 	default:
-	case DRM_MODE_SCALE_FULLSCREEN: rmx_type = RMX_FULL; break;
+	case DRM_MODE_SCALE_FULLSCREEN:
+		rmx_type = RMX_FULL;
+		break;
 	}
+
 	if (amdgpu_encoder->rmx_type == rmx_type)
 		return 0;
 
@@ -833,7 +834,7 @@ static int amdgpu_connector_vga_get_modes(struct drm_connector *connector)
 }
 
 static enum drm_mode_status amdgpu_connector_vga_mode_valid(struct drm_connector *connector,
-					    struct drm_display_mode *mode)
+					    const struct drm_display_mode *mode)
 {
 	struct drm_device *dev = connector->dev;
 	struct amdgpu_device *adev = drm_to_adev(dev);
@@ -916,10 +917,8 @@ amdgpu_connector_vga_detect(struct drm_connector *connector, bool force)
 	amdgpu_connector_update_scratch_regs(connector, ret);
 
 out:
-	if (!drm_kms_helper_is_poll_worker()) {
-		pm_runtime_mark_last_busy(connector->dev->dev);
+	if (!drm_kms_helper_is_poll_worker())
 		pm_runtime_put_autosuspend(connector->dev->dev);
-	}
 
 	return ret;
 }
@@ -959,6 +958,41 @@ amdgpu_connector_check_hpd_status_unchanged(struct drm_connector *connector)
 	return false;
 }
 
+static void amdgpu_connector_shared_ddc(enum drm_connector_status *status,
+					struct drm_connector *connector,
+					struct amdgpu_connector *amdgpu_connector)
+{
+	struct drm_connector *list_connector;
+	struct drm_connector_list_iter iter;
+	struct amdgpu_connector *list_amdgpu_connector;
+	struct drm_device *dev = connector->dev;
+	struct amdgpu_device *adev = drm_to_adev(dev);
+
+	if (amdgpu_connector->shared_ddc && *status == connector_status_connected) {
+		drm_connector_list_iter_begin(dev, &iter);
+		drm_for_each_connector_iter(list_connector,
+					    &iter) {
+			if (connector == list_connector)
+				continue;
+			list_amdgpu_connector = to_amdgpu_connector(list_connector);
+			if (list_amdgpu_connector->shared_ddc &&
+			    list_amdgpu_connector->ddc_bus->rec.i2c_id ==
+			     amdgpu_connector->ddc_bus->rec.i2c_id) {
+				/* cases where both connectors are digital */
+				if (list_connector->connector_type != DRM_MODE_CONNECTOR_VGA) {
+					/* hpd is our only option in this case */
+					if (!amdgpu_display_hpd_sense(adev,
+								      amdgpu_connector->hpd.hpd)) {
+						amdgpu_connector_free_edid(connector);
+						*status = connector_status_disconnected;
+					}
+				}
+			}
+		}
+		drm_connector_list_iter_end(&iter);
+	}
+}
+
 /*
  * DVI is complicated
  * Do a DDC probe, if DDC probe passes, get the full EDID so
@@ -989,13 +1023,33 @@ amdgpu_connector_dvi_detect(struct drm_connector *connector, bool force)
 		}
 	}
 
+	if (amdgpu_connector->detected_hpd_without_ddc) {
+		force = true;
+		amdgpu_connector->detected_hpd_without_ddc = false;
+	}
+
 	if (!force && amdgpu_connector_check_hpd_status_unchanged(connector)) {
 		ret = connector->status;
 		goto exit;
 	}
 
-	if (amdgpu_connector->ddc_bus)
+	if (amdgpu_connector->ddc_bus) {
 		dret = amdgpu_display_ddc_probe(amdgpu_connector, false);
+
+		/* Sometimes the pins required for the DDC probe on DVI
+		 * connectors don't make contact at the same time that the ones
+		 * for HPD do. If the DDC probe fails even though we had an HPD
+		 * signal, try again later
+		 */
+		if (!dret && !force &&
+		    amdgpu_display_hpd_sense(adev, amdgpu_connector->hpd.hpd)) {
+			DRM_DEBUG_KMS("hpd detected without ddc, retrying in 1 second\n");
+			amdgpu_connector->detected_hpd_without_ddc = true;
+			schedule_delayed_work(&adev->hotplug_work,
+					      msecs_to_jiffies(1000));
+			goto exit;
+		}
+	}
 	if (dret) {
 		amdgpu_connector->detected_by_load = false;
 		amdgpu_connector_free_edid(connector);
@@ -1025,32 +1079,7 @@ amdgpu_connector_dvi_detect(struct drm_connector *connector, bool force)
 			 * DDC line.  The latter is more complex because with DVI<->HDMI adapters
 			 * you don't really know what's connected to which port as both are digital.
 			 */
-			if (amdgpu_connector->shared_ddc && (ret == connector_status_connected)) {
-				struct drm_connector *list_connector;
-				struct drm_connector_list_iter iter;
-				struct amdgpu_connector *list_amdgpu_connector;
-
-				drm_connector_list_iter_begin(dev, &iter);
-				drm_for_each_connector_iter(list_connector,
-							    &iter) {
-					if (connector == list_connector)
-						continue;
-					list_amdgpu_connector = to_amdgpu_connector(list_connector);
-					if (list_amdgpu_connector->shared_ddc &&
-					    (list_amdgpu_connector->ddc_bus->rec.i2c_id ==
-					     amdgpu_connector->ddc_bus->rec.i2c_id)) {
-						/* cases where both connectors are digital */
-						if (list_connector->connector_type != DRM_MODE_CONNECTOR_VGA) {
-							/* hpd is our only option in this case */
-							if (!amdgpu_display_hpd_sense(adev, amdgpu_connector->hpd.hpd)) {
-								amdgpu_connector_free_edid(connector);
-								ret = connector_status_disconnected;
-							}
-						}
-					}
-				}
-				drm_connector_list_iter_end(&iter);
-			}
+			amdgpu_connector_shared_ddc(&ret, connector, amdgpu_connector);
 		}
 	}
 
@@ -1098,7 +1127,8 @@ amdgpu_connector_dvi_detect(struct drm_connector *connector, bool force)
 					/* assume digital unless load detected otherwise */
 					amdgpu_connector->use_digital = true;
 					lret = encoder_funcs->detect(encoder, connector);
-					DRM_DEBUG_KMS("load_detect %x returned: %x\n",encoder->encoder_type,lret);
+					DRM_DEBUG_KMS("load_detect %x returned: %x\n",
+						      encoder->encoder_type, lret);
 					if (lret == connector_status_connected)
 						amdgpu_connector->use_digital = false;
 				}
@@ -1112,10 +1142,8 @@ out:
 	amdgpu_connector_update_scratch_regs(connector, ret);
 
 exit:
-	if (!drm_kms_helper_is_poll_worker()) {
-		pm_runtime_mark_last_busy(connector->dev->dev);
+	if (!drm_kms_helper_is_poll_worker())
 		pm_runtime_put_autosuspend(connector->dev->dev);
-	}
 
 	return ret;
 }
@@ -1151,35 +1179,76 @@ amdgpu_connector_dvi_encoder(struct drm_connector *connector)
 static void amdgpu_connector_dvi_force(struct drm_connector *connector)
 {
 	struct amdgpu_connector *amdgpu_connector = to_amdgpu_connector(connector);
+
 	if (connector->force == DRM_FORCE_ON)
 		amdgpu_connector->use_digital = false;
 	if (connector->force == DRM_FORCE_ON_DIGITAL)
 		amdgpu_connector->use_digital = true;
 }
 
+/**
+ * amdgpu_max_hdmi_pixel_clock - Return max supported HDMI (TMDS) pixel clock
+ * @adev: pointer to amdgpu_device
+ *
+ * Return: maximum supported HDMI (TMDS) pixel clock in KHz.
+ */
+static int amdgpu_max_hdmi_pixel_clock(const struct amdgpu_device *adev)
+{
+	if (adev->asic_type >= CHIP_POLARIS10)
+		return 600000;
+	else if (adev->asic_type >= CHIP_TONGA)
+		return 300000;
+	else
+		return 297000;
+}
+
+/**
+ * amdgpu_connector_dvi_mode_valid - Validate a mode on DVI/HDMI connectors
+ * @connector: DRM connector to validate the mode on
+ * @mode: display mode to validate
+ *
+ * Validate the given display mode on DVI and HDMI connectors, including
+ * analog signals on DVI-I.
+ *
+ * Return: drm_mode_status indicating whether the mode is valid.
+ */
 static enum drm_mode_status amdgpu_connector_dvi_mode_valid(struct drm_connector *connector,
-					    struct drm_display_mode *mode)
+					    const struct drm_display_mode *mode)
 {
 	struct drm_device *dev = connector->dev;
 	struct amdgpu_device *adev = drm_to_adev(dev);
 	struct amdgpu_connector *amdgpu_connector = to_amdgpu_connector(connector);
+	const int max_hdmi_pixel_clock = amdgpu_max_hdmi_pixel_clock(adev);
+	const int max_dvi_single_link_pixel_clock = 165000;
+	int max_digital_pixel_clock_khz;
 
 	/* XXX check mode bandwidth */
 
-	if (amdgpu_connector->use_digital && (mode->clock > 165000)) {
-		if ((amdgpu_connector->connector_object_id == CONNECTOR_OBJECT_ID_DUAL_LINK_DVI_I) ||
-		    (amdgpu_connector->connector_object_id == CONNECTOR_OBJECT_ID_DUAL_LINK_DVI_D) ||
-		    (amdgpu_connector->connector_object_id == CONNECTOR_OBJECT_ID_HDMI_TYPE_B)) {
-			return MODE_OK;
-		} else if (drm_detect_hdmi_monitor(amdgpu_connector_edid(connector))) {
-			/* HDMI 1.3+ supports max clock of 340 Mhz */
-			if (mode->clock > 340000)
-				return MODE_CLOCK_HIGH;
-			else
-				return MODE_OK;
-		} else {
-			return MODE_CLOCK_HIGH;
+	if (amdgpu_connector->use_digital) {
+		switch (amdgpu_connector->connector_object_id) {
+		case CONNECTOR_OBJECT_ID_HDMI_TYPE_A:
+			max_digital_pixel_clock_khz = max_hdmi_pixel_clock;
+			break;
+		case CONNECTOR_OBJECT_ID_SINGLE_LINK_DVI_I:
+		case CONNECTOR_OBJECT_ID_SINGLE_LINK_DVI_D:
+			max_digital_pixel_clock_khz = max_dvi_single_link_pixel_clock;
+			break;
+		case CONNECTOR_OBJECT_ID_DUAL_LINK_DVI_I:
+		case CONNECTOR_OBJECT_ID_DUAL_LINK_DVI_D:
+		case CONNECTOR_OBJECT_ID_HDMI_TYPE_B:
+			max_digital_pixel_clock_khz = max_dvi_single_link_pixel_clock * 2;
+			break;
 		}
+
+		/* When the display EDID claims that it's an HDMI display,
+		 * we use the HDMI encoder mode of the display HW,
+		 * so we should verify against the max HDMI clock here.
+		 */
+		if (connector->display_info.is_hdmi)
+			max_digital_pixel_clock_khz = max_hdmi_pixel_clock;
+
+		if (mode->clock > max_digital_pixel_clock_khz)
+			return MODE_CLOCK_HIGH;
 	}
 
 	/* check against the max pixel clock */
@@ -1385,6 +1454,7 @@ amdgpu_connector_dp_detect(struct drm_connector *connector, bool force)
 				ret = connector_status_connected;
 			else if (amdgpu_connector->dac_load_detect) { /* try load detection */
 				const struct drm_encoder_helper_funcs *encoder_funcs = encoder->helper_private;
+
 				ret = encoder_funcs->detect(encoder, connector);
 			}
 		}
@@ -1410,10 +1480,8 @@ amdgpu_connector_dp_detect(struct drm_connector *connector, bool force)
 
 	amdgpu_connector_update_scratch_regs(connector, ret);
 out:
-	if (!drm_kms_helper_is_poll_worker()) {
-		pm_runtime_mark_last_busy(connector->dev->dev);
+	if (!drm_kms_helper_is_poll_worker())
 		pm_runtime_put_autosuspend(connector->dev->dev);
-	}
 
 	if (connector->connector_type == DRM_MODE_CONNECTOR_DisplayPort ||
 	    connector->connector_type == DRM_MODE_CONNECTOR_eDP)
@@ -1425,7 +1493,7 @@ out:
 }
 
 static enum drm_mode_status amdgpu_connector_dp_mode_valid(struct drm_connector *connector,
-					   struct drm_display_mode *mode)
+					   const struct drm_display_mode *mode)
 {
 	struct amdgpu_connector *amdgpu_connector = to_amdgpu_connector(connector);
 	struct amdgpu_connector_atom_dig *amdgpu_dig_connector = amdgpu_connector->con_priv;
@@ -1463,7 +1531,7 @@ static enum drm_mode_status amdgpu_connector_dp_mode_valid(struct drm_connector
 		    (amdgpu_dig_connector->dp_sink_type == CONNECTOR_OBJECT_ID_eDP)) {
 			return amdgpu_atombios_dp_mode_valid_helper(connector, mode);
 		} else {
-			if (drm_detect_hdmi_monitor(amdgpu_connector_edid(connector))) {
+			if (connector->display_info.is_hdmi) {
 				/* HDMI 1.3+ supports max clock of 340 Mhz */
 				if (mode->clock > 340000)
 					return MODE_CLOCK_HIGH;
@@ -1665,10 +1733,12 @@ amdgpu_connector_add(struct amdgpu_device *adev,
 						   adev->mode_info.dither_property,
 						   AMDGPU_FMT_DITHER_DISABLE);
 
-			if (amdgpu_audio != 0)
+			if (amdgpu_audio != 0) {
 				drm_object_attach_property(&amdgpu_connector->base.base,
 							   adev->mode_info.audio_property,
 							   AMDGPU_AUDIO_AUTO);
+				amdgpu_connector->audio = AMDGPU_AUDIO_AUTO;
+			}
 
 			subpixel_order = SubPixelHorizontalRGB;
 			connector->interlace_allowed = true;
@@ -1790,6 +1860,7 @@ amdgpu_connector_add(struct amdgpu_device *adev,
 				drm_object_attach_property(&amdgpu_connector->base.base,
 							   adev->mode_info.audio_property,
 							   AMDGPU_AUDIO_AUTO);
+				amdgpu_connector->audio = AMDGPU_AUDIO_AUTO;
 			}
 			drm_object_attach_property(&amdgpu_connector->base.base,
 						   adev->mode_info.dither_property,
@@ -1843,6 +1914,7 @@ amdgpu_connector_add(struct amdgpu_device *adev,
 				drm_object_attach_property(&amdgpu_connector->base.base,
 							   adev->mode_info.audio_property,
 							   AMDGPU_AUDIO_AUTO);
+				amdgpu_connector->audio = AMDGPU_AUDIO_AUTO;
 			}
 			drm_object_attach_property(&amdgpu_connector->base.base,
 						   adev->mode_info.dither_property,
@@ -1893,6 +1965,7 @@ amdgpu_connector_add(struct amdgpu_device *adev,
 				drm_object_attach_property(&amdgpu_connector->base.base,
 							   adev->mode_info.audio_property,
 							   AMDGPU_AUDIO_AUTO);
+				amdgpu_connector->audio = AMDGPU_AUDIO_AUTO;
 			}
 			drm_object_attach_property(&amdgpu_connector->base.base,
 						   adev->mode_info.dither_property,
@@ -1957,7 +2030,7 @@ amdgpu_connector_add(struct amdgpu_device *adev,
 	if (amdgpu_connector->hpd.hpd == AMDGPU_HPD_NONE) {
 		if (i2c_bus->valid) {
 			connector->polled = DRM_CONNECTOR_POLL_CONNECT |
-			                    DRM_CONNECTOR_POLL_DISCONNECT;
+						DRM_CONNECTOR_POLL_DISCONNECT;
 		}
 	} else
 		connector->polled = DRM_CONNECTOR_POLL_HPD;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.h
index 61fcef15ad72..eff833b6ed31 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.h
@@ -24,7 +24,6 @@
 #ifndef __AMDGPU_CONNECTORS_H__
 #define __AMDGPU_CONNECTORS_H__
 
-struct edid *amdgpu_connector_edid(struct drm_connector *connector);
 void amdgpu_connector_hotplug(struct drm_connector *connector);
 int amdgpu_connector_get_monitor_bpc(struct drm_connector *connector);
 u16 amdgpu_connector_encoder_get_dp_bridge_encoder_id(struct drm_connector *connector);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cper.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cper.c
new file mode 100644
index 000000000000..425a3e564360
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cper.c
@@ -0,0 +1,591 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright 2025 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include <linux/list.h>
+#include "amdgpu.h"
+
+static const guid_t MCE			= CPER_NOTIFY_MCE;
+static const guid_t CMC			= CPER_NOTIFY_CMC;
+static const guid_t BOOT		= BOOT_TYPE;
+
+static const guid_t CRASHDUMP		= AMD_CRASHDUMP;
+static const guid_t RUNTIME		= AMD_GPU_NONSTANDARD_ERROR;
+
+static void __inc_entry_length(struct cper_hdr *hdr, uint32_t size)
+{
+	hdr->record_length += size;
+}
+
+static void amdgpu_cper_get_timestamp(struct cper_timestamp *timestamp)
+{
+	struct tm tm;
+	time64_t now = ktime_get_real_seconds();
+
+	time64_to_tm(now, 0, &tm);
+	timestamp->seconds = tm.tm_sec;
+	timestamp->minutes = tm.tm_min;
+	timestamp->hours = tm.tm_hour;
+	timestamp->flag = 0;
+	timestamp->day = tm.tm_mday;
+	timestamp->month = 1 + tm.tm_mon;
+	timestamp->year = (1900 + tm.tm_year) % 100;
+	timestamp->century = (1900 + tm.tm_year) / 100;
+}
+
+void amdgpu_cper_entry_fill_hdr(struct amdgpu_device *adev,
+				struct cper_hdr *hdr,
+				enum amdgpu_cper_type type,
+				enum cper_error_severity sev)
+{
+	char record_id[16];
+
+	hdr->signature[0]		= 'C';
+	hdr->signature[1]		= 'P';
+	hdr->signature[2]		= 'E';
+	hdr->signature[3]		= 'R';
+	hdr->revision			= CPER_HDR_REV_1;
+	hdr->signature_end		= 0xFFFFFFFF;
+	hdr->error_severity		= sev;
+
+	hdr->valid_bits.platform_id	= 1;
+	hdr->valid_bits.timestamp	= 1;
+
+	amdgpu_cper_get_timestamp(&hdr->timestamp);
+
+	snprintf(record_id, 9, "%d:%X",
+		 (adev->smuio.funcs && adev->smuio.funcs->get_socket_id) ?
+			 adev->smuio.funcs->get_socket_id(adev) :
+			 0,
+		 atomic_inc_return(&adev->cper.unique_id));
+	memcpy(hdr->record_id, record_id, 8);
+
+	snprintf(hdr->platform_id, 16, "0x%04X:0x%04X",
+		 adev->pdev->vendor, adev->pdev->device);
+	/* pmfw version should be part of creator_id according to CPER spec */
+	snprintf(hdr->creator_id, 16, "%s", CPER_CREATOR_ID_AMDGPU);
+
+	switch (type) {
+	case AMDGPU_CPER_TYPE_BOOT:
+		hdr->notify_type = BOOT;
+		break;
+	case AMDGPU_CPER_TYPE_FATAL:
+	case AMDGPU_CPER_TYPE_BP_THRESHOLD:
+		hdr->notify_type = MCE;
+		break;
+	case AMDGPU_CPER_TYPE_RUNTIME:
+		if (sev == CPER_SEV_NON_FATAL_CORRECTED)
+			hdr->notify_type = CMC;
+		else
+			hdr->notify_type = MCE;
+		break;
+	default:
+		dev_err(adev->dev, "Unknown CPER Type\n");
+		break;
+	}
+
+	__inc_entry_length(hdr, HDR_LEN);
+}
+
+static int amdgpu_cper_entry_fill_section_desc(struct amdgpu_device *adev,
+					       struct cper_sec_desc *section_desc,
+					       bool bp_threshold,
+					       bool poison,
+					       enum cper_error_severity sev,
+					       guid_t sec_type,
+					       uint32_t section_length,
+					       uint32_t section_offset)
+{
+	section_desc->revision_minor		= CPER_SEC_MINOR_REV_1;
+	section_desc->revision_major		= CPER_SEC_MAJOR_REV_22;
+	section_desc->sec_offset		= section_offset;
+	section_desc->sec_length		= section_length;
+	section_desc->valid_bits.fru_text	= 1;
+	section_desc->flag_bits.primary		= 1;
+	section_desc->severity			= sev;
+	section_desc->sec_type			= sec_type;
+
+	snprintf(section_desc->fru_text, 20, "OAM%d",
+		 (adev->smuio.funcs && adev->smuio.funcs->get_socket_id) ?
+			 adev->smuio.funcs->get_socket_id(adev) :
+			 0);
+
+	if (bp_threshold)
+		section_desc->flag_bits.exceed_err_threshold = 1;
+	if (poison)
+		section_desc->flag_bits.latent_err = 1;
+
+	return 0;
+}
+
+int amdgpu_cper_entry_fill_fatal_section(struct amdgpu_device *adev,
+					 struct cper_hdr *hdr,
+					 uint32_t idx,
+					 struct cper_sec_crashdump_reg_data reg_data)
+{
+	struct cper_sec_desc *section_desc;
+	struct cper_sec_crashdump_fatal *section;
+
+	section_desc = (struct cper_sec_desc *)((uint8_t *)hdr + SEC_DESC_OFFSET(idx));
+	section = (struct cper_sec_crashdump_fatal *)((uint8_t *)hdr +
+		   FATAL_SEC_OFFSET(hdr->sec_cnt, idx));
+
+	amdgpu_cper_entry_fill_section_desc(adev, section_desc, false, false,
+					    CPER_SEV_FATAL, CRASHDUMP, FATAL_SEC_LEN,
+					    FATAL_SEC_OFFSET(hdr->sec_cnt, idx));
+
+	section->body.reg_ctx_type = CPER_CTX_TYPE_CRASH;
+	section->body.reg_arr_size = sizeof(reg_data);
+	section->body.data = reg_data;
+
+	__inc_entry_length(hdr, SEC_DESC_LEN + FATAL_SEC_LEN);
+
+	return 0;
+}
+
+int amdgpu_cper_entry_fill_runtime_section(struct amdgpu_device *adev,
+					   struct cper_hdr *hdr,
+					   uint32_t idx,
+					   enum cper_error_severity sev,
+					   uint32_t *reg_dump,
+					   uint32_t reg_count)
+{
+	struct cper_sec_desc *section_desc;
+	struct cper_sec_nonstd_err *section;
+	bool poison;
+
+	poison = sev != CPER_SEV_NON_FATAL_CORRECTED;
+	section_desc = (struct cper_sec_desc *)((uint8_t *)hdr + SEC_DESC_OFFSET(idx));
+	section = (struct cper_sec_nonstd_err *)((uint8_t *)hdr +
+		   NONSTD_SEC_OFFSET(hdr->sec_cnt, idx));
+
+	amdgpu_cper_entry_fill_section_desc(adev, section_desc, false, poison,
+					    sev, RUNTIME, NONSTD_SEC_LEN,
+					    NONSTD_SEC_OFFSET(hdr->sec_cnt, idx));
+
+	reg_count = umin(reg_count, CPER_ACA_REG_COUNT);
+
+	section->hdr.valid_bits.err_info_cnt = 1;
+	section->hdr.valid_bits.err_context_cnt = 1;
+
+	section->info.error_type = RUNTIME;
+	section->info.ms_chk_bits.err_type_valid = 1;
+	section->ctx.reg_ctx_type = CPER_CTX_TYPE_CRASH;
+	section->ctx.reg_arr_size = sizeof(section->ctx.reg_dump);
+
+	memcpy(section->ctx.reg_dump, reg_dump, reg_count * sizeof(uint32_t));
+
+	__inc_entry_length(hdr, SEC_DESC_LEN + NONSTD_SEC_LEN);
+
+	return 0;
+}
+
+int amdgpu_cper_entry_fill_bad_page_threshold_section(struct amdgpu_device *adev,
+						      struct cper_hdr *hdr,
+						      uint32_t idx)
+{
+	struct cper_sec_desc *section_desc;
+	struct cper_sec_nonstd_err *section;
+	uint32_t socket_id;
+
+	section_desc = (struct cper_sec_desc *)((uint8_t *)hdr + SEC_DESC_OFFSET(idx));
+	section = (struct cper_sec_nonstd_err *)((uint8_t *)hdr +
+		   NONSTD_SEC_OFFSET(hdr->sec_cnt, idx));
+
+	amdgpu_cper_entry_fill_section_desc(adev, section_desc, true, false,
+					    CPER_SEV_FATAL, RUNTIME, NONSTD_SEC_LEN,
+					    NONSTD_SEC_OFFSET(hdr->sec_cnt, idx));
+
+	section->hdr.valid_bits.err_info_cnt = 1;
+	section->hdr.valid_bits.err_context_cnt = 1;
+
+	section->info.error_type = RUNTIME;
+	section->info.valid_bits.ms_chk = 1;
+	section->info.ms_chk_bits.err_type_valid = 1;
+	section->info.ms_chk_bits.err_type = 1;
+	section->info.ms_chk_bits.pcc = 1;
+	section->ctx.reg_ctx_type = CPER_CTX_TYPE_CRASH;
+	section->ctx.reg_arr_size = sizeof(section->ctx.reg_dump);
+
+	/* Hardcoded Reg dump for bad page threshold CPER */
+	socket_id = (adev->smuio.funcs && adev->smuio.funcs->get_socket_id) ?
+				adev->smuio.funcs->get_socket_id(adev) :
+				0;
+	section->ctx.reg_dump[CPER_ACA_REG_CTL_LO]    = 0x1;
+	section->ctx.reg_dump[CPER_ACA_REG_CTL_HI]    = 0x0;
+	section->ctx.reg_dump[CPER_ACA_REG_STATUS_LO] = 0x137;
+	section->ctx.reg_dump[CPER_ACA_REG_STATUS_HI] = 0xB0000000;
+	section->ctx.reg_dump[CPER_ACA_REG_ADDR_LO]   = 0x0;
+	section->ctx.reg_dump[CPER_ACA_REG_ADDR_HI]   = 0x0;
+	section->ctx.reg_dump[CPER_ACA_REG_MISC0_LO]  = 0x0;
+	section->ctx.reg_dump[CPER_ACA_REG_MISC0_HI]  = 0x0;
+	section->ctx.reg_dump[CPER_ACA_REG_CONFIG_LO] = 0x2;
+	section->ctx.reg_dump[CPER_ACA_REG_CONFIG_HI] = 0x1ff;
+	section->ctx.reg_dump[CPER_ACA_REG_IPID_LO]   = (socket_id / 4) & 0x01;
+	section->ctx.reg_dump[CPER_ACA_REG_IPID_HI]   = 0x096 | (((socket_id % 4) & 0x3) << 12);
+	section->ctx.reg_dump[CPER_ACA_REG_SYND_LO]   = 0x0;
+	section->ctx.reg_dump[CPER_ACA_REG_SYND_HI]   = 0x0;
+
+	__inc_entry_length(hdr, SEC_DESC_LEN + NONSTD_SEC_LEN);
+
+	return 0;
+}
+
+struct cper_hdr *amdgpu_cper_alloc_entry(struct amdgpu_device *adev,
+					 enum amdgpu_cper_type type,
+					 uint16_t section_count)
+{
+	struct cper_hdr *hdr;
+	uint32_t size = 0;
+
+	size += HDR_LEN;
+	size += (SEC_DESC_LEN * section_count);
+
+	switch (type) {
+	case AMDGPU_CPER_TYPE_RUNTIME:
+	case AMDGPU_CPER_TYPE_BP_THRESHOLD:
+		size += (NONSTD_SEC_LEN * section_count);
+		break;
+	case AMDGPU_CPER_TYPE_FATAL:
+		size += (FATAL_SEC_LEN * section_count);
+		break;
+	case AMDGPU_CPER_TYPE_BOOT:
+		size += (BOOT_SEC_LEN * section_count);
+		break;
+	default:
+		dev_err(adev->dev, "Unknown CPER Type!\n");
+		return NULL;
+	}
+
+	hdr = kzalloc(size, GFP_KERNEL);
+	if (!hdr)
+		return NULL;
+
+	/* Save this early */
+	hdr->sec_cnt = section_count;
+
+	return hdr;
+}
+
+int amdgpu_cper_generate_ue_record(struct amdgpu_device *adev,
+				   struct aca_bank *bank)
+{
+	struct cper_hdr *fatal = NULL;
+	struct cper_sec_crashdump_reg_data reg_data = { 0 };
+	struct amdgpu_ring *ring = &adev->cper.ring_buf;
+	int ret;
+
+	fatal = amdgpu_cper_alloc_entry(adev, AMDGPU_CPER_TYPE_FATAL, 1);
+	if (!fatal) {
+		dev_err(adev->dev, "fail to alloc cper entry for ue record\n");
+		return -ENOMEM;
+	}
+
+	reg_data.status_lo = lower_32_bits(bank->regs[ACA_REG_IDX_STATUS]);
+	reg_data.status_hi = upper_32_bits(bank->regs[ACA_REG_IDX_STATUS]);
+	reg_data.addr_lo   = lower_32_bits(bank->regs[ACA_REG_IDX_ADDR]);
+	reg_data.addr_hi   = upper_32_bits(bank->regs[ACA_REG_IDX_ADDR]);
+	reg_data.ipid_lo   = lower_32_bits(bank->regs[ACA_REG_IDX_IPID]);
+	reg_data.ipid_hi   = upper_32_bits(bank->regs[ACA_REG_IDX_IPID]);
+	reg_data.synd_lo   = lower_32_bits(bank->regs[ACA_REG_IDX_SYND]);
+	reg_data.synd_hi   = upper_32_bits(bank->regs[ACA_REG_IDX_SYND]);
+
+	amdgpu_cper_entry_fill_hdr(adev, fatal, AMDGPU_CPER_TYPE_FATAL, CPER_SEV_FATAL);
+	ret = amdgpu_cper_entry_fill_fatal_section(adev, fatal, 0, reg_data);
+	if (ret)
+		return ret;
+
+	amdgpu_cper_ring_write(ring, fatal, fatal->record_length);
+	kfree(fatal);
+
+	return 0;
+}
+
+int amdgpu_cper_generate_bp_threshold_record(struct amdgpu_device *adev)
+{
+	struct cper_hdr *bp_threshold = NULL;
+	struct amdgpu_ring *ring = &adev->cper.ring_buf;
+	int ret;
+
+	bp_threshold = amdgpu_cper_alloc_entry(adev, AMDGPU_CPER_TYPE_BP_THRESHOLD, 1);
+	if (!bp_threshold) {
+		dev_err(adev->dev, "fail to alloc cper entry for bad page threshold record\n");
+		return -ENOMEM;
+	}
+
+	amdgpu_cper_entry_fill_hdr(adev, bp_threshold,
+				   AMDGPU_CPER_TYPE_BP_THRESHOLD,
+				   CPER_SEV_FATAL);
+	ret = amdgpu_cper_entry_fill_bad_page_threshold_section(adev, bp_threshold, 0);
+	if (ret)
+		return ret;
+
+	amdgpu_cper_ring_write(ring, bp_threshold, bp_threshold->record_length);
+	kfree(bp_threshold);
+
+	return 0;
+}
+
+static enum cper_error_severity amdgpu_aca_err_type_to_cper_sev(struct amdgpu_device *adev,
+								enum aca_error_type aca_err_type)
+{
+	switch (aca_err_type) {
+	case ACA_ERROR_TYPE_UE:
+		return CPER_SEV_FATAL;
+	case ACA_ERROR_TYPE_CE:
+		return CPER_SEV_NON_FATAL_CORRECTED;
+	case ACA_ERROR_TYPE_DEFERRED:
+		return CPER_SEV_NON_FATAL_UNCORRECTED;
+	default:
+		dev_err(adev->dev, "Unknown ACA error type!\n");
+		return CPER_SEV_FATAL;
+	}
+}
+
+int amdgpu_cper_generate_ce_records(struct amdgpu_device *adev,
+				    struct aca_banks *banks,
+				    uint16_t bank_count)
+{
+	struct cper_hdr *corrected = NULL;
+	enum cper_error_severity sev = CPER_SEV_NON_FATAL_CORRECTED;
+	struct amdgpu_ring *ring = &adev->cper.ring_buf;
+	uint32_t reg_data[CPER_ACA_REG_COUNT] = { 0 };
+	struct aca_bank_node *node;
+	struct aca_bank *bank;
+	uint32_t i = 0;
+	int ret;
+
+	corrected = amdgpu_cper_alloc_entry(adev, AMDGPU_CPER_TYPE_RUNTIME, bank_count);
+	if (!corrected) {
+		dev_err(adev->dev, "fail to allocate cper entry for ce records\n");
+		return -ENOMEM;
+	}
+
+	/* Raise severity if any DE is detected in the ACA bank list */
+	list_for_each_entry(node, &banks->list, node) {
+		bank = &node->bank;
+		if (bank->aca_err_type == ACA_ERROR_TYPE_DEFERRED) {
+			sev = CPER_SEV_NON_FATAL_UNCORRECTED;
+			break;
+		}
+	}
+
+	amdgpu_cper_entry_fill_hdr(adev, corrected, AMDGPU_CPER_TYPE_RUNTIME, sev);
+
+	/* Combine CE and DE in cper record */
+	list_for_each_entry(node, &banks->list, node) {
+		bank = &node->bank;
+		reg_data[CPER_ACA_REG_CTL_LO]    = lower_32_bits(bank->regs[ACA_REG_IDX_CTL]);
+		reg_data[CPER_ACA_REG_CTL_HI]    = upper_32_bits(bank->regs[ACA_REG_IDX_CTL]);
+		reg_data[CPER_ACA_REG_STATUS_LO] = lower_32_bits(bank->regs[ACA_REG_IDX_STATUS]);
+		reg_data[CPER_ACA_REG_STATUS_HI] = upper_32_bits(bank->regs[ACA_REG_IDX_STATUS]);
+		reg_data[CPER_ACA_REG_ADDR_LO]   = lower_32_bits(bank->regs[ACA_REG_IDX_ADDR]);
+		reg_data[CPER_ACA_REG_ADDR_HI]   = upper_32_bits(bank->regs[ACA_REG_IDX_ADDR]);
+		reg_data[CPER_ACA_REG_MISC0_LO]  = lower_32_bits(bank->regs[ACA_REG_IDX_MISC0]);
+		reg_data[CPER_ACA_REG_MISC0_HI]  = upper_32_bits(bank->regs[ACA_REG_IDX_MISC0]);
+		reg_data[CPER_ACA_REG_CONFIG_LO] = lower_32_bits(bank->regs[ACA_REG_IDX_CONFIG]);
+		reg_data[CPER_ACA_REG_CONFIG_HI] = upper_32_bits(bank->regs[ACA_REG_IDX_CONFIG]);
+		reg_data[CPER_ACA_REG_IPID_LO]   = lower_32_bits(bank->regs[ACA_REG_IDX_IPID]);
+		reg_data[CPER_ACA_REG_IPID_HI]   = upper_32_bits(bank->regs[ACA_REG_IDX_IPID]);
+		reg_data[CPER_ACA_REG_SYND_LO]   = lower_32_bits(bank->regs[ACA_REG_IDX_SYND]);
+		reg_data[CPER_ACA_REG_SYND_HI]   = upper_32_bits(bank->regs[ACA_REG_IDX_SYND]);
+
+		ret = amdgpu_cper_entry_fill_runtime_section(adev, corrected, i++,
+				amdgpu_aca_err_type_to_cper_sev(adev, bank->aca_err_type),
+				reg_data, CPER_ACA_REG_COUNT);
+		if (ret)
+			return ret;
+	}
+
+	amdgpu_cper_ring_write(ring, corrected, corrected->record_length);
+	kfree(corrected);
+
+	return 0;
+}
+
+static bool amdgpu_cper_is_hdr(struct amdgpu_ring *ring, u64 pos)
+{
+	struct cper_hdr *chdr;
+
+	chdr = (struct cper_hdr *)&(ring->ring[pos]);
+	return strcmp(chdr->signature, "CPER") ? false : true;
+}
+
+static u32 amdgpu_cper_ring_get_ent_sz(struct amdgpu_ring *ring, u64 pos)
+{
+	struct cper_hdr *chdr;
+	u64 p;
+	u32 chunk, rec_len = 0;
+
+	chdr = (struct cper_hdr *)&(ring->ring[pos]);
+	chunk = ring->ring_size - (pos << 2);
+
+	if (!strcmp(chdr->signature, "CPER")) {
+		rec_len = chdr->record_length;
+		goto calc;
+	}
+
+	/* ring buffer is not full, no cper data after ring->wptr */
+	if (ring->count_dw)
+		goto calc;
+
+	for (p = pos + 1; p <= ring->buf_mask; p++) {
+		chdr = (struct cper_hdr *)&(ring->ring[p]);
+		if (!strcmp(chdr->signature, "CPER")) {
+			rec_len = (p - pos) << 2;
+			goto calc;
+		}
+	}
+
+calc:
+	if (!rec_len)
+		return chunk;
+	else
+		return umin(rec_len, chunk);
+}
+
+void amdgpu_cper_ring_write(struct amdgpu_ring *ring, void *src, int count)
+{
+	u64 pos, wptr_old, rptr;
+	int rec_cnt_dw = count >> 2;
+	u32 chunk, ent_sz;
+	u8 *s = (u8 *)src;
+
+	if (count >= ring->ring_size - 4) {
+		dev_err(ring->adev->dev,
+			"CPER data size(%d) is larger than ring size(%d)\n",
+			count, ring->ring_size - 4);
+
+		return;
+	}
+
+	mutex_lock(&ring->adev->cper.ring_lock);
+
+	wptr_old = ring->wptr;
+	rptr = *ring->rptr_cpu_addr & ring->ptr_mask;
+
+	while (count) {
+		ent_sz = amdgpu_cper_ring_get_ent_sz(ring, ring->wptr);
+		chunk = umin(ent_sz, count);
+
+		memcpy(&ring->ring[ring->wptr], s, chunk);
+
+		ring->wptr += (chunk >> 2);
+		ring->wptr &= ring->ptr_mask;
+		count -= chunk;
+		s += chunk;
+	}
+
+	if (ring->count_dw < rec_cnt_dw)
+		ring->count_dw = 0;
+
+	/* the buffer is overflow, adjust rptr */
+	if (((wptr_old < rptr) && (rptr <= ring->wptr)) ||
+	    ((ring->wptr < wptr_old) && (wptr_old < rptr)) ||
+	    ((rptr <= ring->wptr) && (ring->wptr < wptr_old))) {
+		pos = (ring->wptr + 1) & ring->ptr_mask;
+
+		do {
+			ent_sz = amdgpu_cper_ring_get_ent_sz(ring, pos);
+
+			rptr += (ent_sz >> 2);
+			rptr &= ring->ptr_mask;
+			*ring->rptr_cpu_addr = rptr;
+
+			pos = rptr;
+		} while (!amdgpu_cper_is_hdr(ring, rptr));
+	}
+
+	if (ring->count_dw >= rec_cnt_dw)
+		ring->count_dw -= rec_cnt_dw;
+	mutex_unlock(&ring->adev->cper.ring_lock);
+}
+
+static u64 amdgpu_cper_ring_get_rptr(struct amdgpu_ring *ring)
+{
+	return *(ring->rptr_cpu_addr);
+}
+
+static u64 amdgpu_cper_ring_get_wptr(struct amdgpu_ring *ring)
+{
+	return ring->wptr;
+}
+
+static const struct amdgpu_ring_funcs cper_ring_funcs = {
+	.type = AMDGPU_RING_TYPE_CPER,
+	.align_mask = 0xff,
+	.support_64bit_ptrs = false,
+	.get_rptr = amdgpu_cper_ring_get_rptr,
+	.get_wptr = amdgpu_cper_ring_get_wptr,
+};
+
+static int amdgpu_cper_ring_init(struct amdgpu_device *adev)
+{
+	struct amdgpu_ring *ring = &(adev->cper.ring_buf);
+
+	mutex_init(&adev->cper.ring_lock);
+
+	ring->adev = NULL;
+	ring->ring_obj = NULL;
+	ring->use_doorbell = false;
+	ring->no_scheduler = true;
+	ring->funcs = &cper_ring_funcs;
+
+	sprintf(ring->name, "cper");
+	return amdgpu_ring_init(adev, ring, CPER_MAX_RING_SIZE, NULL, 0,
+				AMDGPU_RING_PRIO_DEFAULT, NULL);
+}
+
+int amdgpu_cper_init(struct amdgpu_device *adev)
+{
+	int r;
+
+	if (!amdgpu_aca_is_enabled(adev) && !amdgpu_sriov_ras_cper_en(adev))
+		return 0;
+
+	r = amdgpu_cper_ring_init(adev);
+	if (r) {
+		dev_err(adev->dev, "failed to initialize cper ring, r = %d\n", r);
+		return r;
+	}
+
+	mutex_init(&adev->cper.cper_lock);
+
+	adev->cper.enabled = true;
+	adev->cper.max_count = CPER_MAX_ALLOWED_COUNT;
+
+	return 0;
+}
+
+int amdgpu_cper_fini(struct amdgpu_device *adev)
+{
+	if (!amdgpu_aca_is_enabled(adev) && !amdgpu_sriov_ras_cper_en(adev))
+		return 0;
+
+	adev->cper.enabled = false;
+
+	amdgpu_ring_fini(&(adev->cper.ring_buf));
+	adev->cper.count = 0;
+	adev->cper.wptr = 0;
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cper.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_cper.h
new file mode 100644
index 000000000000..353421807387
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cper.h
@@ -0,0 +1,105 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2025 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __AMDGPU_CPER_H__
+#define __AMDGPU_CPER_H__
+
+#include "amd_cper.h"
+#include "amdgpu_aca.h"
+
+#define CPER_MAX_ALLOWED_COUNT		0x1000
+#define CPER_MAX_RING_SIZE		0X100000
+#define HDR_LEN				(sizeof(struct cper_hdr))
+#define SEC_DESC_LEN			(sizeof(struct cper_sec_desc))
+
+#define BOOT_SEC_LEN			(sizeof(struct cper_sec_crashdump_boot))
+#define FATAL_SEC_LEN			(sizeof(struct cper_sec_crashdump_fatal))
+#define NONSTD_SEC_LEN			(sizeof(struct cper_sec_nonstd_err))
+
+#define SEC_DESC_OFFSET(idx)		(HDR_LEN + (SEC_DESC_LEN * idx))
+
+#define BOOT_SEC_OFFSET(count, idx)	(HDR_LEN + (SEC_DESC_LEN * count) + (BOOT_SEC_LEN * idx))
+#define FATAL_SEC_OFFSET(count, idx)	(HDR_LEN + (SEC_DESC_LEN * count) + (FATAL_SEC_LEN * idx))
+#define NONSTD_SEC_OFFSET(count, idx)	(HDR_LEN + (SEC_DESC_LEN * count) + (NONSTD_SEC_LEN * idx))
+
+enum amdgpu_cper_type {
+	AMDGPU_CPER_TYPE_RUNTIME,
+	AMDGPU_CPER_TYPE_FATAL,
+	AMDGPU_CPER_TYPE_BOOT,
+	AMDGPU_CPER_TYPE_BP_THRESHOLD,
+};
+
+struct amdgpu_cper {
+	bool enabled;
+
+	atomic_t unique_id;
+	struct mutex cper_lock;
+
+	/* Lifetime CPERs generated */
+	uint32_t count;
+	uint32_t max_count;
+
+	uint32_t wptr;
+
+	void *ring[CPER_MAX_ALLOWED_COUNT];
+	struct amdgpu_ring ring_buf;
+	struct mutex ring_lock;
+};
+
+void amdgpu_cper_entry_fill_hdr(struct amdgpu_device *adev,
+				struct cper_hdr *hdr,
+				enum amdgpu_cper_type type,
+				enum cper_error_severity sev);
+int amdgpu_cper_entry_fill_fatal_section(struct amdgpu_device *adev,
+					 struct cper_hdr *hdr,
+					 uint32_t idx,
+					 struct cper_sec_crashdump_reg_data reg_data);
+int amdgpu_cper_entry_fill_runtime_section(struct amdgpu_device *adev,
+					   struct cper_hdr *hdr,
+					   uint32_t idx,
+					   enum cper_error_severity sev,
+					   uint32_t *reg_dump,
+					   uint32_t reg_count);
+int amdgpu_cper_entry_fill_bad_page_threshold_section(struct amdgpu_device *adev,
+						      struct cper_hdr *hdr,
+						      uint32_t section_idx);
+
+struct cper_hdr *amdgpu_cper_alloc_entry(struct amdgpu_device *adev,
+					 enum amdgpu_cper_type type,
+					 uint16_t section_count);
+/* UE must be encoded into separated cper entries, 1 UE 1 cper */
+int amdgpu_cper_generate_ue_record(struct amdgpu_device *adev,
+				   struct aca_bank *bank);
+/* CEs and DEs are combined into 1 cper entry */
+int amdgpu_cper_generate_ce_records(struct amdgpu_device *adev,
+				    struct aca_banks *banks,
+				    uint16_t bank_count);
+/* Bad page threshold is encoded into separated cper entry */
+int amdgpu_cper_generate_bp_threshold_record(struct amdgpu_device *adev);
+void amdgpu_cper_ring_write(struct amdgpu_ring *ring,
+			void *src, int count);
+int amdgpu_cper_init(struct amdgpu_device *adev);
+int amdgpu_cper_fini(struct amdgpu_device *adev);
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
index 0311d799a010..ecdfe6cb36cc 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
@@ -32,58 +32,127 @@
 
 #include <drm/amdgpu_drm.h>
 #include <drm/drm_syncobj.h>
+#include <drm/ttm/ttm_tt.h>
+
+#include "amdgpu_cs.h"
 #include "amdgpu.h"
 #include "amdgpu_trace.h"
 #include "amdgpu_gmc.h"
 #include "amdgpu_gem.h"
 #include "amdgpu_ras.h"
+#include "amdgpu_hmm.h"
+
+static int amdgpu_cs_parser_init(struct amdgpu_cs_parser *p,
+				 struct amdgpu_device *adev,
+				 struct drm_file *filp,
+				 union drm_amdgpu_cs *cs)
+{
+	struct amdgpu_fpriv *fpriv = filp->driver_priv;
+
+	if (cs->in.num_chunks == 0)
+		return -EINVAL;
+
+	memset(p, 0, sizeof(*p));
+	p->adev = adev;
+	p->filp = filp;
+
+	p->ctx = amdgpu_ctx_get(fpriv, cs->in.ctx_id);
+	if (!p->ctx)
+		return -EINVAL;
+
+	if (atomic_read(&p->ctx->guilty)) {
+		amdgpu_ctx_put(p->ctx);
+		return -ECANCELED;
+	}
+
+	amdgpu_sync_create(&p->sync);
+	drm_exec_init(&p->exec, DRM_EXEC_INTERRUPTIBLE_WAIT |
+		      DRM_EXEC_IGNORE_DUPLICATES, 0);
+	return 0;
+}
+
+static int amdgpu_cs_job_idx(struct amdgpu_cs_parser *p,
+			     struct drm_amdgpu_cs_chunk_ib *chunk_ib)
+{
+	struct drm_sched_entity *entity;
+	unsigned int i;
+	int r;
+
+	r = amdgpu_ctx_get_entity(p->ctx, chunk_ib->ip_type,
+				  chunk_ib->ip_instance,
+				  chunk_ib->ring, &entity);
+	if (r)
+		return r;
+
+	/*
+	 * Abort if there is no run queue associated with this entity.
+	 * Possibly because of disabled HW IP.
+	 */
+	if (entity->rq == NULL)
+		return -EINVAL;
+
+	/* Check if we can add this IB to some existing job */
+	for (i = 0; i < p->gang_size; ++i)
+		if (p->entities[i] == entity)
+			return i;
+
+	/* If not increase the gang size if possible */
+	if (i == AMDGPU_CS_GANG_SIZE)
+		return -EINVAL;
 
-static int amdgpu_cs_user_fence_chunk(struct amdgpu_cs_parser *p,
-				      struct drm_amdgpu_cs_chunk_fence *data,
-				      uint32_t *offset)
+	p->entities[i] = entity;
+	p->gang_size = i + 1;
+	return i;
+}
+
+static int amdgpu_cs_p1_ib(struct amdgpu_cs_parser *p,
+			   struct drm_amdgpu_cs_chunk_ib *chunk_ib,
+			   unsigned int *num_ibs)
+{
+	int r;
+
+	r = amdgpu_cs_job_idx(p, chunk_ib);
+	if (r < 0)
+		return r;
+
+	if (num_ibs[r] >= amdgpu_ring_max_ibs(chunk_ib->ip_type))
+		return -EINVAL;
+
+	++(num_ibs[r]);
+	p->gang_leader_idx = r;
+	return 0;
+}
+
+static int amdgpu_cs_p1_user_fence(struct amdgpu_cs_parser *p,
+				   struct drm_amdgpu_cs_chunk_fence *data,
+				   uint32_t *offset)
 {
 	struct drm_gem_object *gobj;
-	struct amdgpu_bo *bo;
 	unsigned long size;
-	int r;
 
 	gobj = drm_gem_object_lookup(p->filp, data->handle);
 	if (gobj == NULL)
 		return -EINVAL;
 
-	bo = amdgpu_bo_ref(gem_to_amdgpu_bo(gobj));
-	p->uf_entry.priority = 0;
-	p->uf_entry.tv.bo = &bo->tbo;
-	/* One for TTM and one for the CS job */
-	p->uf_entry.tv.num_shared = 2;
-
+	p->uf_bo = amdgpu_bo_ref(gem_to_amdgpu_bo(gobj));
 	drm_gem_object_put(gobj);
 
-	size = amdgpu_bo_size(bo);
-	if (size != PAGE_SIZE || (data->offset + 8) > size) {
-		r = -EINVAL;
-		goto error_unref;
-	}
+	size = amdgpu_bo_size(p->uf_bo);
+	if (size != PAGE_SIZE || data->offset > (size - 8))
+		return -EINVAL;
 
-	if (amdgpu_ttm_tt_get_usermm(bo->tbo.ttm)) {
-		r = -EINVAL;
-		goto error_unref;
-	}
+	if (amdgpu_ttm_tt_get_usermm(p->uf_bo->tbo.ttm))
+		return -EINVAL;
 
 	*offset = data->offset;
-
 	return 0;
-
-error_unref:
-	amdgpu_bo_unref(&bo);
-	return r;
 }
 
-static int amdgpu_cs_bo_handles_chunk(struct amdgpu_cs_parser *p,
-				      struct drm_amdgpu_bo_list_in *data)
+static int amdgpu_cs_p1_bo_handles(struct amdgpu_cs_parser *p,
+				   struct drm_amdgpu_bo_list_in *data)
 {
+	struct drm_amdgpu_bo_list_entry *info;
 	int r;
-	struct drm_amdgpu_bo_list_entry *info = NULL;
 
 	r = amdgpu_bo_create_list_entry_array(data, &info);
 	if (r)
@@ -103,45 +172,24 @@ error_free:
 	return r;
 }
 
-static int amdgpu_cs_parser_init(struct amdgpu_cs_parser *p, union drm_amdgpu_cs *cs)
+/* Copy the data from userspace and go over it the first time */
+static int amdgpu_cs_pass1(struct amdgpu_cs_parser *p,
+			   union drm_amdgpu_cs *cs)
 {
 	struct amdgpu_fpriv *fpriv = p->filp->driver_priv;
+	unsigned int num_ibs[AMDGPU_CS_GANG_SIZE] = { };
 	struct amdgpu_vm *vm = &fpriv->vm;
-	uint64_t *chunk_array_user;
 	uint64_t *chunk_array;
-	unsigned size, num_ibs = 0;
 	uint32_t uf_offset = 0;
-	int i;
+	size_t size;
 	int ret;
+	int i;
 
-	if (cs->in.num_chunks == 0)
-		return 0;
-
-	chunk_array = kvmalloc_array(cs->in.num_chunks, sizeof(uint64_t), GFP_KERNEL);
-	if (!chunk_array)
-		return -ENOMEM;
-
-	p->ctx = amdgpu_ctx_get(fpriv, cs->in.ctx_id);
-	if (!p->ctx) {
-		ret = -EINVAL;
-		goto free_chunk;
-	}
-
-	mutex_lock(&p->ctx->lock);
-
-	/* skip guilty context job */
-	if (atomic_read(&p->ctx->guilty) == 1) {
-		ret = -ECANCELED;
-		goto free_chunk;
-	}
-
-	/* get chunks */
-	chunk_array_user = u64_to_user_ptr(cs->in.chunks);
-	if (copy_from_user(chunk_array, chunk_array_user,
-			   sizeof(uint64_t)*cs->in.num_chunks)) {
-		ret = -EFAULT;
-		goto free_chunk;
-	}
+	chunk_array = memdup_array_user(u64_to_user_ptr(cs->in.chunks),
+					cs->in.num_chunks,
+					sizeof(uint64_t));
+	if (IS_ERR(chunk_array))
+		return PTR_ERR(chunk_array);
 
 	p->nchunks = cs->in.num_chunks;
 	p->chunks = kvmalloc_array(p->nchunks, sizeof(struct amdgpu_cs_chunk),
@@ -152,9 +200,8 @@ static int amdgpu_cs_parser_init(struct amdgpu_cs_parser *p, union drm_amdgpu_cs
 	}
 
 	for (i = 0; i < p->nchunks; i++) {
-		struct drm_amdgpu_cs_chunk __user **chunk_ptr = NULL;
+		struct drm_amdgpu_cs_chunk __user *chunk_ptr = NULL;
 		struct drm_amdgpu_cs_chunk user_chunk;
-		uint32_t __user *cdata;
 
 		chunk_ptr = u64_to_user_ptr(chunk_array[i]);
 		if (copy_from_user(&user_chunk, chunk_ptr,
@@ -167,50 +214,50 @@ static int amdgpu_cs_parser_init(struct amdgpu_cs_parser *p, union drm_amdgpu_cs
 		p->chunks[i].length_dw = user_chunk.length_dw;
 
 		size = p->chunks[i].length_dw;
-		cdata = u64_to_user_ptr(user_chunk.chunk_data);
 
-		p->chunks[i].kdata = kvmalloc_array(size, sizeof(uint32_t), GFP_KERNEL);
-		if (p->chunks[i].kdata == NULL) {
-			ret = -ENOMEM;
+		p->chunks[i].kdata = vmemdup_array_user(u64_to_user_ptr(user_chunk.chunk_data),
+							size,
+							sizeof(uint32_t));
+		if (IS_ERR(p->chunks[i].kdata)) {
+			ret = PTR_ERR(p->chunks[i].kdata);
 			i--;
 			goto free_partial_kdata;
 		}
 		size *= sizeof(uint32_t);
-		if (copy_from_user(p->chunks[i].kdata, cdata, size)) {
-			ret = -EFAULT;
-			goto free_partial_kdata;
-		}
 
+		/* Assume the worst on the following checks */
+		ret = -EINVAL;
 		switch (p->chunks[i].chunk_id) {
 		case AMDGPU_CHUNK_ID_IB:
-			++num_ibs;
+			if (size < sizeof(struct drm_amdgpu_cs_chunk_ib))
+				goto free_partial_kdata;
+
+			ret = amdgpu_cs_p1_ib(p, p->chunks[i].kdata, num_ibs);
+			if (ret)
+				goto free_partial_kdata;
 			break;
 
 		case AMDGPU_CHUNK_ID_FENCE:
-			size = sizeof(struct drm_amdgpu_cs_chunk_fence);
-			if (p->chunks[i].length_dw * sizeof(uint32_t) < size) {
-				ret = -EINVAL;
+			if (size < sizeof(struct drm_amdgpu_cs_chunk_fence))
 				goto free_partial_kdata;
-			}
 
-			ret = amdgpu_cs_user_fence_chunk(p, p->chunks[i].kdata,
-							 &uf_offset);
+			ret = amdgpu_cs_p1_user_fence(p, p->chunks[i].kdata,
+						      &uf_offset);
 			if (ret)
 				goto free_partial_kdata;
-
 			break;
 
 		case AMDGPU_CHUNK_ID_BO_HANDLES:
-			size = sizeof(struct drm_amdgpu_bo_list_in);
-			if (p->chunks[i].length_dw * sizeof(uint32_t) < size) {
-				ret = -EINVAL;
+			if (size < sizeof(struct drm_amdgpu_bo_list_in))
 				goto free_partial_kdata;
-			}
 
-			ret = amdgpu_cs_bo_handles_chunk(p, p->chunks[i].kdata);
-			if (ret)
+			/* Only a single BO list is allowed to simplify handling. */
+			if (p->bo_list)
 				goto free_partial_kdata;
 
+			ret = amdgpu_cs_p1_bo_handles(p, p->chunks[i].kdata);
+			if (ret)
+				goto free_partial_kdata;
 			break;
 
 		case AMDGPU_CHUNK_ID_DEPENDENCIES:
@@ -219,25 +266,54 @@ static int amdgpu_cs_parser_init(struct amdgpu_cs_parser *p, union drm_amdgpu_cs
 		case AMDGPU_CHUNK_ID_SCHEDULED_DEPENDENCIES:
 		case AMDGPU_CHUNK_ID_SYNCOBJ_TIMELINE_WAIT:
 		case AMDGPU_CHUNK_ID_SYNCOBJ_TIMELINE_SIGNAL:
+		case AMDGPU_CHUNK_ID_CP_GFX_SHADOW:
 			break;
 
 		default:
-			ret = -EINVAL;
 			goto free_partial_kdata;
 		}
 	}
 
-	ret = amdgpu_job_alloc(p->adev, num_ibs, &p->job, vm);
-	if (ret)
+	if (!p->gang_size || (amdgpu_sriov_vf(p->adev) && p->gang_size > 1)) {
+		ret = -EINVAL;
 		goto free_all_kdata;
+	}
+
+	for (i = 0; i < p->gang_size; ++i) {
+		ret = amdgpu_job_alloc(p->adev, vm, p->entities[i], vm,
+				       num_ibs[i], &p->jobs[i],
+				       p->filp->client_id);
+		if (ret)
+			goto free_all_kdata;
+		switch (p->adev->enforce_isolation[fpriv->xcp_id]) {
+		case AMDGPU_ENFORCE_ISOLATION_DISABLE:
+		default:
+			p->jobs[i]->enforce_isolation = false;
+			p->jobs[i]->run_cleaner_shader = false;
+			break;
+		case AMDGPU_ENFORCE_ISOLATION_ENABLE:
+			p->jobs[i]->enforce_isolation = true;
+			p->jobs[i]->run_cleaner_shader = true;
+			break;
+		case AMDGPU_ENFORCE_ISOLATION_ENABLE_LEGACY:
+			p->jobs[i]->enforce_isolation = true;
+			p->jobs[i]->run_cleaner_shader = false;
+			break;
+		case AMDGPU_ENFORCE_ISOLATION_NO_CLEANER_SHADER:
+			p->jobs[i]->enforce_isolation = true;
+			p->jobs[i]->run_cleaner_shader = false;
+			break;
+		}
+	}
+	p->gang_leader = p->jobs[p->gang_leader_idx];
 
-	if (p->ctx->vram_lost_counter != p->job->vram_lost_counter) {
+	if (p->ctx->generation != p->gang_leader->generation) {
 		ret = -ECANCELED;
 		goto free_all_kdata;
 	}
 
-	if (p->uf_entry.tv.bo)
-		p->job->uf_addr = uf_offset;
+	if (p->uf_bo)
+		p->gang_leader->uf_addr = uf_offset;
 	kvfree(chunk_array);
 
 	/* Use this opportunity to fill in task info for the vm */
@@ -259,6 +335,331 @@ free_chunk:
 	return ret;
 }
 
+static int amdgpu_cs_p2_ib(struct amdgpu_cs_parser *p,
+			   struct amdgpu_cs_chunk *chunk,
+			   unsigned int *ce_preempt,
+			   unsigned int *de_preempt)
+{
+	struct drm_amdgpu_cs_chunk_ib *chunk_ib = chunk->kdata;
+	struct amdgpu_fpriv *fpriv = p->filp->driver_priv;
+	struct amdgpu_vm *vm = &fpriv->vm;
+	struct amdgpu_ring *ring;
+	struct amdgpu_job *job;
+	struct amdgpu_ib *ib;
+	int r;
+
+	r = amdgpu_cs_job_idx(p, chunk_ib);
+	if (r < 0)
+		return r;
+
+	job = p->jobs[r];
+	ring = amdgpu_job_ring(job);
+	ib = &job->ibs[job->num_ibs++];
+
+	/* submissions to kernel queues are disabled */
+	if (ring->no_user_submission)
+		return -EINVAL;
+
+	/* MM engine doesn't support user fences */
+	if (p->uf_bo && ring->funcs->no_user_fence)
+		return -EINVAL;
+
+	if (!p->adev->debug_enable_ce_cs &&
+	    chunk_ib->flags & AMDGPU_IB_FLAG_CE) {
+		dev_err_ratelimited(p->adev->dev, "CE CS is blocked, use debug=0x400 to override\n");
+		return -EINVAL;
+	}
+
+	if (chunk_ib->ip_type == AMDGPU_HW_IP_GFX &&
+	    chunk_ib->flags & AMDGPU_IB_FLAG_PREEMPT) {
+		if (chunk_ib->flags & AMDGPU_IB_FLAG_CE)
+			(*ce_preempt)++;
+		else
+			(*de_preempt)++;
+
+		/* Each GFX command submit allows only 1 IB max
+		 * preemptible for CE & DE */
+		if (*ce_preempt > 1 || *de_preempt > 1)
+			return -EINVAL;
+	}
+
+	if (chunk_ib->flags & AMDGPU_IB_FLAG_PREAMBLE)
+		job->preamble_status |= AMDGPU_PREAMBLE_IB_PRESENT;
+
+	r =  amdgpu_ib_get(p->adev, vm, ring->funcs->parse_cs ?
+			   chunk_ib->ib_bytes : 0,
+			   AMDGPU_IB_POOL_DELAYED, ib);
+	if (r) {
+		drm_err(adev_to_drm(p->adev), "Failed to get ib !\n");
+		return r;
+	}
+
+	ib->gpu_addr = chunk_ib->va_start;
+	ib->length_dw = chunk_ib->ib_bytes / 4;
+	ib->flags = chunk_ib->flags;
+	return 0;
+}
+
+static int amdgpu_cs_p2_dependencies(struct amdgpu_cs_parser *p,
+				     struct amdgpu_cs_chunk *chunk)
+{
+	struct drm_amdgpu_cs_chunk_dep *deps = chunk->kdata;
+	struct amdgpu_fpriv *fpriv = p->filp->driver_priv;
+	unsigned int num_deps;
+	int i, r;
+
+	num_deps = chunk->length_dw * 4 /
+		sizeof(struct drm_amdgpu_cs_chunk_dep);
+
+	for (i = 0; i < num_deps; ++i) {
+		struct amdgpu_ctx *ctx;
+		struct drm_sched_entity *entity;
+		struct dma_fence *fence;
+
+		ctx = amdgpu_ctx_get(fpriv, deps[i].ctx_id);
+		if (ctx == NULL)
+			return -EINVAL;
+
+		r = amdgpu_ctx_get_entity(ctx, deps[i].ip_type,
+					  deps[i].ip_instance,
+					  deps[i].ring, &entity);
+		if (r) {
+			amdgpu_ctx_put(ctx);
+			return r;
+		}
+
+		fence = amdgpu_ctx_get_fence(ctx, entity, deps[i].handle);
+		amdgpu_ctx_put(ctx);
+
+		if (IS_ERR(fence))
+			return PTR_ERR(fence);
+		else if (!fence)
+			continue;
+
+		if (chunk->chunk_id == AMDGPU_CHUNK_ID_SCHEDULED_DEPENDENCIES) {
+			struct drm_sched_fence *s_fence;
+			struct dma_fence *old = fence;
+
+			s_fence = to_drm_sched_fence(fence);
+			fence = dma_fence_get(&s_fence->scheduled);
+			dma_fence_put(old);
+		}
+
+		r = amdgpu_sync_fence(&p->sync, fence, GFP_KERNEL);
+		dma_fence_put(fence);
+		if (r)
+			return r;
+	}
+	return 0;
+}
+
+static int amdgpu_syncobj_lookup_and_add(struct amdgpu_cs_parser *p,
+					 uint32_t handle, u64 point,
+					 u64 flags)
+{
+	struct dma_fence *fence;
+	int r;
+
+	r = drm_syncobj_find_fence(p->filp, handle, point, flags, &fence);
+	if (r) {
+		drm_err(adev_to_drm(p->adev), "syncobj %u failed to find fence @ %llu (%d)!\n",
+			  handle, point, r);
+		return r;
+	}
+
+	r = amdgpu_sync_fence(&p->sync, fence, GFP_KERNEL);
+	dma_fence_put(fence);
+	return r;
+}
+
+static int amdgpu_cs_p2_syncobj_in(struct amdgpu_cs_parser *p,
+				   struct amdgpu_cs_chunk *chunk)
+{
+	struct drm_amdgpu_cs_chunk_sem *deps = chunk->kdata;
+	unsigned int num_deps;
+	int i, r;
+
+	num_deps = chunk->length_dw * 4 /
+		sizeof(struct drm_amdgpu_cs_chunk_sem);
+	for (i = 0; i < num_deps; ++i) {
+		r = amdgpu_syncobj_lookup_and_add(p, deps[i].handle, 0, 0);
+		if (r)
+			return r;
+	}
+
+	return 0;
+}
+
+static int amdgpu_cs_p2_syncobj_timeline_wait(struct amdgpu_cs_parser *p,
+					      struct amdgpu_cs_chunk *chunk)
+{
+	struct drm_amdgpu_cs_chunk_syncobj *syncobj_deps = chunk->kdata;
+	unsigned int num_deps;
+	int i, r;
+
+	num_deps = chunk->length_dw * 4 /
+		sizeof(struct drm_amdgpu_cs_chunk_syncobj);
+	for (i = 0; i < num_deps; ++i) {
+		r = amdgpu_syncobj_lookup_and_add(p, syncobj_deps[i].handle,
+						  syncobj_deps[i].point,
+						  syncobj_deps[i].flags);
+		if (r)
+			return r;
+	}
+
+	return 0;
+}
+
+static int amdgpu_cs_p2_syncobj_out(struct amdgpu_cs_parser *p,
+				    struct amdgpu_cs_chunk *chunk)
+{
+	struct drm_amdgpu_cs_chunk_sem *deps = chunk->kdata;
+	unsigned int num_deps;
+	int i;
+
+	num_deps = chunk->length_dw * 4 /
+		sizeof(struct drm_amdgpu_cs_chunk_sem);
+
+	if (p->post_deps)
+		return -EINVAL;
+
+	p->post_deps = kmalloc_array(num_deps, sizeof(*p->post_deps),
+				     GFP_KERNEL);
+	p->num_post_deps = 0;
+
+	if (!p->post_deps)
+		return -ENOMEM;
+
+
+	for (i = 0; i < num_deps; ++i) {
+		p->post_deps[i].syncobj =
+			drm_syncobj_find(p->filp, deps[i].handle);
+		if (!p->post_deps[i].syncobj)
+			return -EINVAL;
+		p->post_deps[i].chain = NULL;
+		p->post_deps[i].point = 0;
+		p->num_post_deps++;
+	}
+
+	return 0;
+}
+
+static int amdgpu_cs_p2_syncobj_timeline_signal(struct amdgpu_cs_parser *p,
+						struct amdgpu_cs_chunk *chunk)
+{
+	struct drm_amdgpu_cs_chunk_syncobj *syncobj_deps = chunk->kdata;
+	unsigned int num_deps;
+	int i;
+
+	num_deps = chunk->length_dw * 4 /
+		sizeof(struct drm_amdgpu_cs_chunk_syncobj);
+
+	if (p->post_deps)
+		return -EINVAL;
+
+	p->post_deps = kmalloc_array(num_deps, sizeof(*p->post_deps),
+				     GFP_KERNEL);
+	p->num_post_deps = 0;
+
+	if (!p->post_deps)
+		return -ENOMEM;
+
+	for (i = 0; i < num_deps; ++i) {
+		struct amdgpu_cs_post_dep *dep = &p->post_deps[i];
+
+		dep->chain = NULL;
+		if (syncobj_deps[i].point) {
+			dep->chain = dma_fence_chain_alloc();
+			if (!dep->chain)
+				return -ENOMEM;
+		}
+
+		dep->syncobj = drm_syncobj_find(p->filp,
+						syncobj_deps[i].handle);
+		if (!dep->syncobj) {
+			dma_fence_chain_free(dep->chain);
+			return -EINVAL;
+		}
+		dep->point = syncobj_deps[i].point;
+		p->num_post_deps++;
+	}
+
+	return 0;
+}
+
+static int amdgpu_cs_p2_shadow(struct amdgpu_cs_parser *p,
+			       struct amdgpu_cs_chunk *chunk)
+{
+	struct drm_amdgpu_cs_chunk_cp_gfx_shadow *shadow = chunk->kdata;
+	int i;
+
+	if (shadow->flags & ~AMDGPU_CS_CHUNK_CP_GFX_SHADOW_FLAGS_INIT_SHADOW)
+		return -EINVAL;
+
+	for (i = 0; i < p->gang_size; ++i) {
+		p->jobs[i]->shadow_va = shadow->shadow_va;
+		p->jobs[i]->csa_va = shadow->csa_va;
+		p->jobs[i]->gds_va = shadow->gds_va;
+		p->jobs[i]->init_shadow =
+			shadow->flags & AMDGPU_CS_CHUNK_CP_GFX_SHADOW_FLAGS_INIT_SHADOW;
+	}
+
+	return 0;
+}
+
+static int amdgpu_cs_pass2(struct amdgpu_cs_parser *p)
+{
+	unsigned int ce_preempt = 0, de_preempt = 0;
+	int i, r;
+
+	for (i = 0; i < p->nchunks; ++i) {
+		struct amdgpu_cs_chunk *chunk;
+
+		chunk = &p->chunks[i];
+
+		switch (chunk->chunk_id) {
+		case AMDGPU_CHUNK_ID_IB:
+			r = amdgpu_cs_p2_ib(p, chunk, &ce_preempt, &de_preempt);
+			if (r)
+				return r;
+			break;
+		case AMDGPU_CHUNK_ID_DEPENDENCIES:
+		case AMDGPU_CHUNK_ID_SCHEDULED_DEPENDENCIES:
+			r = amdgpu_cs_p2_dependencies(p, chunk);
+			if (r)
+				return r;
+			break;
+		case AMDGPU_CHUNK_ID_SYNCOBJ_IN:
+			r = amdgpu_cs_p2_syncobj_in(p, chunk);
+			if (r)
+				return r;
+			break;
+		case AMDGPU_CHUNK_ID_SYNCOBJ_OUT:
+			r = amdgpu_cs_p2_syncobj_out(p, chunk);
+			if (r)
+				return r;
+			break;
+		case AMDGPU_CHUNK_ID_SYNCOBJ_TIMELINE_WAIT:
+			r = amdgpu_cs_p2_syncobj_timeline_wait(p, chunk);
+			if (r)
+				return r;
+			break;
+		case AMDGPU_CHUNK_ID_SYNCOBJ_TIMELINE_SIGNAL:
+			r = amdgpu_cs_p2_syncobj_timeline_signal(p, chunk);
+			if (r)
+				return r;
+			break;
+		case AMDGPU_CHUNK_ID_CP_GFX_SHADOW:
+			r = amdgpu_cs_p2_shadow(p, chunk);
+			if (r)
+				return r;
+			break;
+		}
+	}
+
+	return 0;
+}
+
 /* Convert microseconds to bytes. */
 static u64 us_to_bytes(struct amdgpu_device *adev, s64 us)
 {
@@ -298,7 +699,6 @@ static void amdgpu_cs_get_threshold_for_moves(struct amdgpu_device *adev,
 {
 	s64 time_us, increment_us;
 	u64 free_vram, total_vram, used_vram;
-	struct ttm_resource_manager *vram_man = ttm_manager_type(&adev->mman.bdev, TTM_PL_VRAM);
 	/* Allow a maximum of 200 accumulated ms. This is basically per-IB
 	 * throttling.
 	 *
@@ -308,14 +708,14 @@ static void amdgpu_cs_get_threshold_for_moves(struct amdgpu_device *adev,
 	 */
 	const s64 us_upper_bound = 200000;
 
-	if (!adev->mm_stats.log2_max_MBps) {
+	if ((!adev->mm_stats.log2_max_MBps) || !ttm_resource_manager_used(&adev->mman.vram_mgr.manager)) {
 		*max_bytes = 0;
 		*max_vis_bytes = 0;
 		return;
 	}
 
 	total_vram = adev->gmc.real_vram_size - atomic64_read(&adev->vram_pin_size);
-	used_vram = amdgpu_vram_mgr_usage(vram_man);
+	used_vram = ttm_resource_manager_usage(&adev->mman.vram_mgr.manager);
 	free_vram = used_vram >= total_vram ? 0 : total_vram - used_vram;
 
 	spin_lock(&adev->mm_stats.lock);
@@ -342,7 +742,7 @@ static void amdgpu_cs_get_threshold_for_moves(struct amdgpu_device *adev,
 	if (free_vram >= 128 * 1024 * 1024 || free_vram >= total_vram / 8) {
 		s64 min_us;
 
-		/* Be more aggresive on dGPUs. Try to fill a portion of free
+		/* Be more aggressive on dGPUs. Try to fill a portion of free
 		 * VRAM now.
 		 */
 		if (!(adev->flags & AMD_IS_APU))
@@ -362,10 +762,11 @@ static void amdgpu_cs_get_threshold_for_moves(struct amdgpu_device *adev,
 	if (!amdgpu_gmc_vram_full_visible(&adev->gmc)) {
 		u64 total_vis_vram = adev->gmc.visible_vram_size;
 		u64 used_vis_vram =
-		  amdgpu_vram_mgr_vis_usage(vram_man);
+		  amdgpu_vram_mgr_vis_usage(&adev->mman.vram_mgr);
 
 		if (used_vis_vram < total_vis_vram) {
 			u64 free_vis_vram = total_vis_vram - used_vis_vram;
+
 			adev->mm_stats.accum_us_vis = min(adev->mm_stats.accum_us_vis +
 							  increment_us, us_upper_bound);
 
@@ -440,7 +841,7 @@ retry:
 
 	p->bytes_moved += ctx.bytes_moved;
 	if (!amdgpu_gmc_vram_full_visible(&adev->gmc) &&
-	    amdgpu_bo_in_cpu_visible_vram(bo))
+	    amdgpu_res_cpu_visible(adev, bo->tbo.resource))
 		p->bytes_moved_vis += ctx.bytes_moved;
 
 	if (unlikely(r == -ENOMEM) && domain != bo->allowed_domains) {
@@ -451,57 +852,18 @@ retry:
 	return r;
 }
 
-static int amdgpu_cs_list_validate(struct amdgpu_cs_parser *p,
-			    struct list_head *validated)
-{
-	struct ttm_operation_ctx ctx = { true, false };
-	struct amdgpu_bo_list_entry *lobj;
-	int r;
-
-	list_for_each_entry(lobj, validated, tv.head) {
-		struct amdgpu_bo *bo = ttm_to_amdgpu_bo(lobj->tv.bo);
-		struct mm_struct *usermm;
-
-		usermm = amdgpu_ttm_tt_get_usermm(bo->tbo.ttm);
-		if (usermm && usermm != current->mm)
-			return -EPERM;
-
-		if (amdgpu_ttm_tt_is_userptr(bo->tbo.ttm) &&
-		    lobj->user_invalidated && lobj->user_pages) {
-			amdgpu_bo_placement_from_domain(bo,
-							AMDGPU_GEM_DOMAIN_CPU);
-			r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
-			if (r)
-				return r;
-
-			amdgpu_ttm_tt_set_user_pages(bo->tbo.ttm,
-						     lobj->user_pages);
-		}
-
-		r = amdgpu_cs_bo_validate(p, bo);
-		if (r)
-			return r;
-
-		kvfree(lobj->user_pages);
-		lobj->user_pages = NULL;
-	}
-	return 0;
-}
-
 static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p,
 				union drm_amdgpu_cs *cs)
 {
 	struct amdgpu_fpriv *fpriv = p->filp->driver_priv;
+	struct ttm_operation_ctx ctx = { true, false };
 	struct amdgpu_vm *vm = &fpriv->vm;
 	struct amdgpu_bo_list_entry *e;
-	struct list_head duplicates;
-	struct amdgpu_bo *gds;
-	struct amdgpu_bo *gws;
-	struct amdgpu_bo *oa;
+	struct drm_gem_object *obj;
+	unsigned long index;
+	unsigned int i;
 	int r;
 
-	INIT_LIST_HEAD(&p->validated);
-
 	/* p->bo_list could already be assigned if AMDGPU_CHUNK_ID_BO_HANDLES is present */
 	if (cs->in.bo_list_handle) {
 		if (p->bo_list)
@@ -519,44 +881,27 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p,
 			return r;
 	}
 
-	/* One for TTM and one for the CS job */
-	amdgpu_bo_list_for_each_entry(e, p->bo_list)
-		e->tv.num_shared = 2;
-
-	amdgpu_bo_list_get_list(p->bo_list, &p->validated);
-
-	INIT_LIST_HEAD(&duplicates);
-	amdgpu_vm_get_pd_bo(&fpriv->vm, &p->validated, &p->vm_pd);
-
-	if (p->uf_entry.tv.bo && !ttm_to_amdgpu_bo(p->uf_entry.tv.bo)->parent)
-		list_add(&p->uf_entry.tv.head, &p->validated);
+	mutex_lock(&p->bo_list->bo_list_mutex);
 
 	/* Get userptr backing pages. If pages are updated after registered
 	 * in amdgpu_gem_userptr_ioctl(), amdgpu_cs_list_validate() will do
 	 * amdgpu_ttm_backend_bind() to flush and invalidate new pages
 	 */
 	amdgpu_bo_list_for_each_userptr_entry(e, p->bo_list) {
-		struct amdgpu_bo *bo = ttm_to_amdgpu_bo(e->tv.bo);
 		bool userpage_invalidated = false;
-		int i;
+		struct amdgpu_bo *bo = e->bo;
 
-		e->user_pages = kvmalloc_array(bo->tbo.ttm->num_pages,
-					sizeof(struct page *),
-					GFP_KERNEL | __GFP_ZERO);
-		if (!e->user_pages) {
-			DRM_ERROR("kvmalloc_array failure\n");
+		e->range = amdgpu_hmm_range_alloc(NULL);
+		if (unlikely(!e->range))
 			return -ENOMEM;
-		}
 
-		r = amdgpu_ttm_tt_get_user_pages(bo, e->user_pages);
-		if (r) {
-			kvfree(e->user_pages);
-			e->user_pages = NULL;
-			return r;
-		}
+		r = amdgpu_ttm_tt_get_user_pages(bo, e->range);
+		if (r)
+			goto out_free_user_pages;
 
 		for (i = 0; i < bo->tbo.ttm->num_pages; i++) {
-			if (bo->tbo.ttm->pages[i] != e->user_pages[i]) {
+			if (bo->tbo.ttm->pages[i] !=
+				hmm_pfn_to_page(e->range->hmm_range.hmm_pfns[i])) {
 				userpage_invalidated = true;
 				break;
 			}
@@ -564,25 +909,52 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p,
 		e->user_invalidated = userpage_invalidated;
 	}
 
-	r = ttm_eu_reserve_buffers(&p->ticket, &p->validated, true,
-				   &duplicates);
-	if (unlikely(r != 0)) {
-		if (r != -ERESTARTSYS)
-			DRM_ERROR("ttm_eu_reserve_buffers failed.\n");
-		goto out;
+	drm_exec_until_all_locked(&p->exec) {
+		r = amdgpu_vm_lock_pd(&fpriv->vm, &p->exec, 1 + p->gang_size);
+		drm_exec_retry_on_contention(&p->exec);
+		if (unlikely(r))
+			goto out_free_user_pages;
+
+		amdgpu_bo_list_for_each_entry(e, p->bo_list) {
+			/* One fence for TTM and one for each CS job */
+			r = drm_exec_prepare_obj(&p->exec, &e->bo->tbo.base,
+						 1 + p->gang_size);
+			drm_exec_retry_on_contention(&p->exec);
+			if (unlikely(r))
+				goto out_free_user_pages;
+
+			e->bo_va = amdgpu_vm_bo_find(vm, e->bo);
+		}
+
+		if (p->uf_bo) {
+			r = drm_exec_prepare_obj(&p->exec, &p->uf_bo->tbo.base,
+						 1 + p->gang_size);
+			drm_exec_retry_on_contention(&p->exec);
+			if (unlikely(r))
+				goto out_free_user_pages;
+		}
 	}
 
-	amdgpu_bo_list_for_each_entry(e, p->bo_list) {
-		struct amdgpu_bo *bo = ttm_to_amdgpu_bo(e->tv.bo);
+	amdgpu_bo_list_for_each_userptr_entry(e, p->bo_list) {
+		struct mm_struct *usermm;
+
+		usermm = amdgpu_ttm_tt_get_usermm(e->bo->tbo.ttm);
+		if (usermm && usermm != current->mm) {
+			r = -EPERM;
+			goto out_free_user_pages;
+		}
 
-		e->bo_va = amdgpu_vm_bo_find(vm, bo);
+		if (amdgpu_ttm_tt_is_userptr(e->bo->tbo.ttm) &&
+		    e->user_invalidated) {
+			amdgpu_bo_placement_from_domain(e->bo,
+							AMDGPU_GEM_DOMAIN_CPU);
+			r = ttm_bo_validate(&e->bo->tbo, &e->bo->placement,
+					    &ctx);
+			if (r)
+				goto out_free_user_pages;
 
-		if (bo->tbo.base.dma_buf && !amdgpu_bo_explicit_sync(bo)) {
-			e->chain = dma_fence_chain_alloc();
-			if (!e->chain) {
-				r = -ENOMEM;
-				goto error_validate;
-			}
+			amdgpu_ttm_tt_set_user_pages(e->bo->tbo.ttm,
+						     e->range);
 		}
 	}
 
@@ -591,254 +963,209 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p,
 	p->bytes_moved = 0;
 	p->bytes_moved_vis = 0;
 
-	r = amdgpu_vm_validate_pt_bos(p->adev, &fpriv->vm,
-				      amdgpu_cs_bo_validate, p);
+	r = amdgpu_vm_validate(p->adev, &fpriv->vm, NULL,
+			       amdgpu_cs_bo_validate, p);
 	if (r) {
-		DRM_ERROR("amdgpu_vm_validate_pt_bos() failed.\n");
-		goto error_validate;
+		drm_err(adev_to_drm(p->adev), "amdgpu_vm_validate() failed.\n");
+		goto out_free_user_pages;
 	}
 
-	r = amdgpu_cs_list_validate(p, &duplicates);
-	if (r)
-		goto error_validate;
-
-	r = amdgpu_cs_list_validate(p, &p->validated);
-	if (r)
-		goto error_validate;
-
-	amdgpu_cs_report_moved_bytes(p->adev, p->bytes_moved,
-				     p->bytes_moved_vis);
+	drm_exec_for_each_locked_object(&p->exec, index, obj) {
+		r = amdgpu_cs_bo_validate(p, gem_to_amdgpu_bo(obj));
+		if (unlikely(r))
+			goto out_free_user_pages;
+	}
 
-	gds = p->bo_list->gds_obj;
-	gws = p->bo_list->gws_obj;
-	oa = p->bo_list->oa_obj;
+	if (p->uf_bo) {
+		r = amdgpu_ttm_alloc_gart(&p->uf_bo->tbo);
+		if (unlikely(r))
+			goto out_free_user_pages;
 
-	if (gds) {
-		p->job->gds_base = amdgpu_bo_gpu_offset(gds) >> PAGE_SHIFT;
-		p->job->gds_size = amdgpu_bo_size(gds) >> PAGE_SHIFT;
-	}
-	if (gws) {
-		p->job->gws_base = amdgpu_bo_gpu_offset(gws) >> PAGE_SHIFT;
-		p->job->gws_size = amdgpu_bo_size(gws) >> PAGE_SHIFT;
-	}
-	if (oa) {
-		p->job->oa_base = amdgpu_bo_gpu_offset(oa) >> PAGE_SHIFT;
-		p->job->oa_size = amdgpu_bo_size(oa) >> PAGE_SHIFT;
+		p->gang_leader->uf_addr += amdgpu_bo_gpu_offset(p->uf_bo);
 	}
 
-	if (!r && p->uf_entry.tv.bo) {
-		struct amdgpu_bo *uf = ttm_to_amdgpu_bo(p->uf_entry.tv.bo);
+	amdgpu_cs_report_moved_bytes(p->adev, p->bytes_moved,
+				     p->bytes_moved_vis);
 
-		r = amdgpu_ttm_alloc_gart(&uf->tbo);
-		p->job->uf_addr += amdgpu_bo_gpu_offset(uf);
-	}
+	for (i = 0; i < p->gang_size; ++i)
+		amdgpu_job_set_resources(p->jobs[i], p->bo_list->gds_obj,
+					 p->bo_list->gws_obj,
+					 p->bo_list->oa_obj);
+	return 0;
 
-error_validate:
-	if (r) {
-		amdgpu_bo_list_for_each_entry(e, p->bo_list) {
-			dma_fence_chain_free(e->chain);
-			e->chain = NULL;
-		}
-		ttm_eu_backoff_reservation(&p->ticket, &p->validated);
+out_free_user_pages:
+	amdgpu_bo_list_for_each_userptr_entry(e, p->bo_list) {
+		amdgpu_hmm_range_free(e->range);
+		e->range = NULL;
 	}
-out:
+	mutex_unlock(&p->bo_list->bo_list_mutex);
 	return r;
 }
 
-static int amdgpu_cs_sync_rings(struct amdgpu_cs_parser *p)
+static void trace_amdgpu_cs_ibs(struct amdgpu_cs_parser *p)
 {
-	struct amdgpu_fpriv *fpriv = p->filp->driver_priv;
-	struct amdgpu_bo_list_entry *e;
-	int r;
+	int i, j;
 
-	list_for_each_entry(e, &p->validated, tv.head) {
-		struct amdgpu_bo *bo = ttm_to_amdgpu_bo(e->tv.bo);
-		struct dma_resv *resv = bo->tbo.base.resv;
-		enum amdgpu_sync_mode sync_mode;
+	if (!trace_amdgpu_cs_enabled())
+		return;
 
-		sync_mode = amdgpu_bo_explicit_sync(bo) ?
-			AMDGPU_SYNC_EXPLICIT : AMDGPU_SYNC_NE_OWNER;
-		r = amdgpu_sync_resv(p->adev, &p->job->sync, resv, sync_mode,
-				     &fpriv->vm);
-		if (r)
-			return r;
+	for (i = 0; i < p->gang_size; ++i) {
+		struct amdgpu_job *job = p->jobs[i];
+
+		for (j = 0; j < job->num_ibs; ++j)
+			trace_amdgpu_cs(p, job, &job->ibs[j]);
 	}
-	return 0;
 }
 
-/**
- * amdgpu_cs_parser_fini() - clean parser states
- * @parser:	parser structure holding parsing context.
- * @error:	error number
- * @backoff:	indicator to backoff the reservation
- *
- * If error is set then unvalidate buffer, otherwise just free memory
- * used by parsing context.
- **/
-static void amdgpu_cs_parser_fini(struct amdgpu_cs_parser *parser, int error,
-				  bool backoff)
+static int amdgpu_cs_patch_ibs(struct amdgpu_cs_parser *p,
+			       struct amdgpu_job *job)
 {
-	unsigned i;
+	struct amdgpu_ring *ring = amdgpu_job_ring(job);
+	unsigned int i;
+	int r;
+
+	/* Only for UVD/VCE VM emulation */
+	if (!ring->funcs->parse_cs && !ring->funcs->patch_cs_in_place)
+		return 0;
 
-	if (error && backoff) {
-		struct amdgpu_bo_list_entry *e;
+	for (i = 0; i < job->num_ibs; ++i) {
+		struct amdgpu_ib *ib = &job->ibs[i];
+		struct amdgpu_bo_va_mapping *m;
+		struct amdgpu_bo *aobj;
+		uint64_t va_start;
+		uint8_t *kptr;
 
-		amdgpu_bo_list_for_each_entry(e, parser->bo_list) {
-			dma_fence_chain_free(e->chain);
-			e->chain = NULL;
+		va_start = ib->gpu_addr & AMDGPU_GMC_HOLE_MASK;
+		r = amdgpu_cs_find_mapping(p, va_start, &aobj, &m);
+		if (r) {
+			drm_err(adev_to_drm(p->adev), "IB va_start is invalid\n");
+			return r;
 		}
 
-		ttm_eu_backoff_reservation(&parser->ticket,
-					   &parser->validated);
-	}
+		if ((va_start + ib->length_dw * 4) >
+		    (m->last + 1) * AMDGPU_GPU_PAGE_SIZE) {
+			drm_err(adev_to_drm(p->adev), "IB va_start+ib_bytes is invalid\n");
+			return -EINVAL;
+		}
 
-	for (i = 0; i < parser->num_post_deps; i++) {
-		drm_syncobj_put(parser->post_deps[i].syncobj);
-		kfree(parser->post_deps[i].chain);
-	}
-	kfree(parser->post_deps);
+		/* the IB should be reserved at this point */
+		r = amdgpu_bo_kmap(aobj, (void **)&kptr);
+		if (r)
+			return r;
 
-	dma_fence_put(parser->fence);
+		kptr += va_start - (m->start * AMDGPU_GPU_PAGE_SIZE);
 
-	if (parser->ctx) {
-		mutex_unlock(&parser->ctx->lock);
-		amdgpu_ctx_put(parser->ctx);
+		if (ring->funcs->parse_cs) {
+			memcpy(ib->ptr, kptr, ib->length_dw * 4);
+			amdgpu_bo_kunmap(aobj);
+
+			r = amdgpu_ring_parse_cs(ring, p, job, ib);
+			if (r)
+				return r;
+
+			if (ib->sa_bo)
+				ib->gpu_addr =  amdgpu_sa_bo_gpu_addr(ib->sa_bo);
+		} else {
+			ib->ptr = (uint32_t *)kptr;
+			r = amdgpu_ring_patch_cs_in_place(ring, p, job, ib);
+			amdgpu_bo_kunmap(aobj);
+			if (r)
+				return r;
+		}
 	}
-	if (parser->bo_list)
-		amdgpu_bo_list_put(parser->bo_list);
 
-	for (i = 0; i < parser->nchunks; i++)
-		kvfree(parser->chunks[i].kdata);
-	kvfree(parser->chunks);
-	if (parser->job)
-		amdgpu_job_free(parser->job);
-	if (parser->uf_entry.tv.bo) {
-		struct amdgpu_bo *uf = ttm_to_amdgpu_bo(parser->uf_entry.tv.bo);
+	return 0;
+}
 
-		amdgpu_bo_unref(&uf);
+static int amdgpu_cs_patch_jobs(struct amdgpu_cs_parser *p)
+{
+	unsigned int i;
+	int r;
+
+	for (i = 0; i < p->gang_size; ++i) {
+		r = amdgpu_cs_patch_ibs(p, p->jobs[i]);
+		if (r)
+			return r;
 	}
+	return 0;
 }
 
 static int amdgpu_cs_vm_handling(struct amdgpu_cs_parser *p)
 {
-	struct amdgpu_ring *ring = to_amdgpu_ring(p->entity->rq->sched);
 	struct amdgpu_fpriv *fpriv = p->filp->driver_priv;
+	struct amdgpu_job *job = p->gang_leader;
 	struct amdgpu_device *adev = p->adev;
 	struct amdgpu_vm *vm = &fpriv->vm;
 	struct amdgpu_bo_list_entry *e;
 	struct amdgpu_bo_va *bo_va;
-	struct amdgpu_bo *bo;
+	unsigned int i;
 	int r;
 
-	/* Only for UVD/VCE VM emulation */
-	if (ring->funcs->parse_cs || ring->funcs->patch_cs_in_place) {
-		unsigned i, j;
-
-		for (i = 0, j = 0; i < p->nchunks && j < p->job->num_ibs; i++) {
-			struct drm_amdgpu_cs_chunk_ib *chunk_ib;
-			struct amdgpu_bo_va_mapping *m;
-			struct amdgpu_bo *aobj = NULL;
-			struct amdgpu_cs_chunk *chunk;
-			uint64_t offset, va_start;
-			struct amdgpu_ib *ib;
-			uint8_t *kptr;
-
-			chunk = &p->chunks[i];
-			ib = &p->job->ibs[j];
-			chunk_ib = chunk->kdata;
-
-			if (chunk->chunk_id != AMDGPU_CHUNK_ID_IB)
-				continue;
-
-			va_start = chunk_ib->va_start & AMDGPU_GMC_HOLE_MASK;
-			r = amdgpu_cs_find_mapping(p, va_start, &aobj, &m);
-			if (r) {
-				DRM_ERROR("IB va_start is invalid\n");
-				return r;
-			}
+	/*
+	 * We can't use gang submit on with reserved VMIDs when the VM changes
+	 * can't be invalidated by more than one engine at the same time.
+	 */
+	if (p->gang_size > 1 && !adev->vm_manager.concurrent_flush) {
+		for (i = 0; i < p->gang_size; ++i) {
+			struct drm_sched_entity *entity = p->entities[i];
+			struct drm_gpu_scheduler *sched = entity->rq->sched;
+			struct amdgpu_ring *ring = to_amdgpu_ring(sched);
 
-			if ((va_start + chunk_ib->ib_bytes) >
-			    (m->last + 1) * AMDGPU_GPU_PAGE_SIZE) {
-				DRM_ERROR("IB va_start+ib_bytes is invalid\n");
+			if (amdgpu_vmid_uses_reserved(vm, ring->vm_hub))
 				return -EINVAL;
-			}
-
-			/* the IB should be reserved at this point */
-			r = amdgpu_bo_kmap(aobj, (void **)&kptr);
-			if (r) {
-				return r;
-			}
-
-			offset = m->start * AMDGPU_GPU_PAGE_SIZE;
-			kptr += va_start - offset;
-
-			if (ring->funcs->parse_cs) {
-				memcpy(ib->ptr, kptr, chunk_ib->ib_bytes);
-				amdgpu_bo_kunmap(aobj);
-
-				r = amdgpu_ring_parse_cs(ring, p, j);
-				if (r)
-					return r;
-			} else {
-				ib->ptr = (uint32_t *)kptr;
-				r = amdgpu_ring_patch_cs_in_place(ring, p, j);
-				amdgpu_bo_kunmap(aobj);
-				if (r)
-					return r;
-			}
-
-			j++;
 		}
 	}
 
-	if (!p->job->vm)
-		return amdgpu_cs_sync_rings(p);
-
+	if (!amdgpu_vm_ready(vm))
+		return -EINVAL;
 
 	r = amdgpu_vm_clear_freed(adev, vm, NULL);
 	if (r)
 		return r;
 
-	r = amdgpu_vm_bo_update(adev, fpriv->prt_va, false, NULL);
+	r = amdgpu_vm_bo_update(adev, fpriv->prt_va, false);
 	if (r)
 		return r;
 
-	r = amdgpu_sync_vm_fence(&p->job->sync, fpriv->prt_va->last_pt_update);
+	r = amdgpu_sync_fence(&p->sync, fpriv->prt_va->last_pt_update,
+			      GFP_KERNEL);
 	if (r)
 		return r;
 
-	if (amdgpu_mcbp || amdgpu_sriov_vf(adev)) {
+	if (fpriv->csa_va) {
 		bo_va = fpriv->csa_va;
 		BUG_ON(!bo_va);
-		r = amdgpu_vm_bo_update(adev, bo_va, false, NULL);
+		r = amdgpu_vm_bo_update(adev, bo_va, false);
 		if (r)
 			return r;
 
-		r = amdgpu_sync_vm_fence(&p->job->sync, bo_va->last_pt_update);
+		r = amdgpu_sync_fence(&p->sync, bo_va->last_pt_update,
+				      GFP_KERNEL);
 		if (r)
 			return r;
 	}
 
+	/* FIXME: In theory this loop shouldn't be needed any more when
+	 * amdgpu_vm_handle_moved handles all moved BOs that are reserved
+	 * with p->ticket. But removing it caused test regressions, so I'm
+	 * leaving it here for now.
+	 */
 	amdgpu_bo_list_for_each_entry(e, p->bo_list) {
-		/* ignore duplicates */
-		bo = ttm_to_amdgpu_bo(e->tv.bo);
-		if (!bo)
-			continue;
-
 		bo_va = e->bo_va;
 		if (bo_va == NULL)
 			continue;
 
-		r = amdgpu_vm_bo_update(adev, bo_va, false, NULL);
+		r = amdgpu_vm_bo_update(adev, bo_va, false);
 		if (r)
 			return r;
 
-		r = amdgpu_sync_vm_fence(&p->job->sync, bo_va->last_pt_update);
+		r = amdgpu_sync_fence(&p->sync, bo_va->last_pt_update,
+				      GFP_KERNEL);
 		if (r)
 			return r;
 	}
 
-	r = amdgpu_vm_handle_moved(adev, vm);
+	r = amdgpu_vm_handle_moved(adev, vm, &p->exec.ticket);
 	if (r)
 		return r;
 
@@ -846,345 +1173,93 @@ static int amdgpu_cs_vm_handling(struct amdgpu_cs_parser *p)
 	if (r)
 		return r;
 
-	r = amdgpu_sync_vm_fence(&p->job->sync, vm->last_update);
+	r = amdgpu_sync_fence(&p->sync, vm->last_update, GFP_KERNEL);
 	if (r)
 		return r;
 
-	p->job->vm_pd_addr = amdgpu_gmc_pd_addr(vm->root.bo);
+	for (i = 0; i < p->gang_size; ++i) {
+		job = p->jobs[i];
+
+		if (!job->vm)
+			continue;
+
+		job->vm_pd_addr = amdgpu_gmc_pd_addr(vm->root.bo);
+	}
 
-	if (amdgpu_vm_debug) {
+	if (adev->debug_vm) {
 		/* Invalidate all BOs to test for userspace bugs */
 		amdgpu_bo_list_for_each_entry(e, p->bo_list) {
-			struct amdgpu_bo *bo = ttm_to_amdgpu_bo(e->tv.bo);
+			struct amdgpu_bo *bo = e->bo;
 
 			/* ignore duplicates */
 			if (!bo)
 				continue;
 
-			amdgpu_vm_bo_invalidate(adev, bo, false);
-		}
-	}
-
-	return amdgpu_cs_sync_rings(p);
-}
-
-static int amdgpu_cs_ib_fill(struct amdgpu_device *adev,
-			     struct amdgpu_cs_parser *parser)
-{
-	struct amdgpu_fpriv *fpriv = parser->filp->driver_priv;
-	struct amdgpu_vm *vm = &fpriv->vm;
-	int r, ce_preempt = 0, de_preempt = 0;
-	struct amdgpu_ring *ring;
-	int i, j;
-
-	for (i = 0, j = 0; i < parser->nchunks && j < parser->job->num_ibs; i++) {
-		struct amdgpu_cs_chunk *chunk;
-		struct amdgpu_ib *ib;
-		struct drm_amdgpu_cs_chunk_ib *chunk_ib;
-		struct drm_sched_entity *entity;
-
-		chunk = &parser->chunks[i];
-		ib = &parser->job->ibs[j];
-		chunk_ib = (struct drm_amdgpu_cs_chunk_ib *)chunk->kdata;
-
-		if (chunk->chunk_id != AMDGPU_CHUNK_ID_IB)
-			continue;
-
-		if (chunk_ib->ip_type == AMDGPU_HW_IP_GFX &&
-		    (amdgpu_mcbp || amdgpu_sriov_vf(adev))) {
-			if (chunk_ib->flags & AMDGPU_IB_FLAG_PREEMPT) {
-				if (chunk_ib->flags & AMDGPU_IB_FLAG_CE)
-					ce_preempt++;
-				else
-					de_preempt++;
-			}
-
-			/* each GFX command submit allows 0 or 1 IB preemptible for CE & DE */
-			if (ce_preempt > 1 || de_preempt > 1)
-				return -EINVAL;
+			amdgpu_vm_bo_invalidate(bo, false);
 		}
-
-		r = amdgpu_ctx_get_entity(parser->ctx, chunk_ib->ip_type,
-					  chunk_ib->ip_instance, chunk_ib->ring,
-					  &entity);
-		if (r)
-			return r;
-
-		if (chunk_ib->flags & AMDGPU_IB_FLAG_PREAMBLE)
-			parser->job->preamble_status |=
-				AMDGPU_PREAMBLE_IB_PRESENT;
-
-		if (parser->entity && parser->entity != entity)
-			return -EINVAL;
-
-		/* Return if there is no run queue associated with this entity.
-		 * Possibly because of disabled HW IP*/
-		if (entity->rq == NULL)
-			return -EINVAL;
-
-		parser->entity = entity;
-
-		ring = to_amdgpu_ring(entity->rq->sched);
-		r =  amdgpu_ib_get(adev, vm, ring->funcs->parse_cs ?
-				   chunk_ib->ib_bytes : 0,
-				   AMDGPU_IB_POOL_DELAYED, ib);
-		if (r) {
-			DRM_ERROR("Failed to get ib !\n");
-			return r;
-		}
-
-		ib->gpu_addr = chunk_ib->va_start;
-		ib->length_dw = chunk_ib->ib_bytes / 4;
-		ib->flags = chunk_ib->flags;
-
-		j++;
 	}
 
-	/* MM engine doesn't support user fences */
-	ring = to_amdgpu_ring(parser->entity->rq->sched);
-	if (parser->job->uf_addr && ring->funcs->no_user_fence)
-		return -EINVAL;
-
-	return amdgpu_ctx_wait_prev_fence(parser->ctx, parser->entity);
-}
-
-static int amdgpu_cs_process_fence_dep(struct amdgpu_cs_parser *p,
-				       struct amdgpu_cs_chunk *chunk)
-{
-	struct amdgpu_fpriv *fpriv = p->filp->driver_priv;
-	unsigned num_deps;
-	int i, r;
-	struct drm_amdgpu_cs_chunk_dep *deps;
-
-	deps = (struct drm_amdgpu_cs_chunk_dep *)chunk->kdata;
-	num_deps = chunk->length_dw * 4 /
-		sizeof(struct drm_amdgpu_cs_chunk_dep);
-
-	for (i = 0; i < num_deps; ++i) {
-		struct amdgpu_ctx *ctx;
-		struct drm_sched_entity *entity;
-		struct dma_fence *fence;
-
-		ctx = amdgpu_ctx_get(fpriv, deps[i].ctx_id);
-		if (ctx == NULL)
-			return -EINVAL;
-
-		r = amdgpu_ctx_get_entity(ctx, deps[i].ip_type,
-					  deps[i].ip_instance,
-					  deps[i].ring, &entity);
-		if (r) {
-			amdgpu_ctx_put(ctx);
-			return r;
-		}
-
-		fence = amdgpu_ctx_get_fence(ctx, entity, deps[i].handle);
-		amdgpu_ctx_put(ctx);
-
-		if (IS_ERR(fence))
-			return PTR_ERR(fence);
-		else if (!fence)
-			continue;
-
-		if (chunk->chunk_id == AMDGPU_CHUNK_ID_SCHEDULED_DEPENDENCIES) {
-			struct drm_sched_fence *s_fence;
-			struct dma_fence *old = fence;
-
-			s_fence = to_drm_sched_fence(fence);
-			fence = dma_fence_get(&s_fence->scheduled);
-			dma_fence_put(old);
-		}
-
-		r = amdgpu_sync_fence(&p->job->sync, fence);
-		dma_fence_put(fence);
-		if (r)
-			return r;
-	}
 	return 0;
 }
 
-static int amdgpu_syncobj_lookup_and_add_to_sync(struct amdgpu_cs_parser *p,
-						 uint32_t handle, u64 point,
-						 u64 flags)
+static int amdgpu_cs_sync_rings(struct amdgpu_cs_parser *p)
 {
+	struct amdgpu_fpriv *fpriv = p->filp->driver_priv;
+	struct drm_gpu_scheduler *sched;
+	struct drm_gem_object *obj;
 	struct dma_fence *fence;
+	unsigned long index;
+	unsigned int i;
 	int r;
 
-	r = drm_syncobj_find_fence(p->filp, handle, point, flags, &fence);
+	r = amdgpu_ctx_wait_prev_fence(p->ctx, p->entities[p->gang_leader_idx]);
 	if (r) {
-		DRM_ERROR("syncobj %u failed to find fence @ %llu (%d)!\n",
-			  handle, point, r);
+		if (r != -ERESTARTSYS)
+			drm_err(adev_to_drm(p->adev), "amdgpu_ctx_wait_prev_fence failed.\n");
 		return r;
 	}
 
-	r = amdgpu_sync_fence(&p->job->sync, fence);
-	dma_fence_put(fence);
-
-	return r;
-}
+	drm_exec_for_each_locked_object(&p->exec, index, obj) {
+		struct amdgpu_bo *bo = gem_to_amdgpu_bo(obj);
 
-static int amdgpu_cs_process_syncobj_in_dep(struct amdgpu_cs_parser *p,
-					    struct amdgpu_cs_chunk *chunk)
-{
-	struct drm_amdgpu_cs_chunk_sem *deps;
-	unsigned num_deps;
-	int i, r;
+		struct dma_resv *resv = bo->tbo.base.resv;
+		enum amdgpu_sync_mode sync_mode;
 
-	deps = (struct drm_amdgpu_cs_chunk_sem *)chunk->kdata;
-	num_deps = chunk->length_dw * 4 /
-		sizeof(struct drm_amdgpu_cs_chunk_sem);
-	for (i = 0; i < num_deps; ++i) {
-		r = amdgpu_syncobj_lookup_and_add_to_sync(p, deps[i].handle,
-							  0, 0);
+		sync_mode = amdgpu_bo_explicit_sync(bo) ?
+			AMDGPU_SYNC_EXPLICIT : AMDGPU_SYNC_NE_OWNER;
+		r = amdgpu_sync_resv(p->adev, &p->sync, resv, sync_mode,
+				     &fpriv->vm);
 		if (r)
 			return r;
 	}
 
-	return 0;
-}
-
-
-static int amdgpu_cs_process_syncobj_timeline_in_dep(struct amdgpu_cs_parser *p,
-						     struct amdgpu_cs_chunk *chunk)
-{
-	struct drm_amdgpu_cs_chunk_syncobj *syncobj_deps;
-	unsigned num_deps;
-	int i, r;
-
-	syncobj_deps = (struct drm_amdgpu_cs_chunk_syncobj *)chunk->kdata;
-	num_deps = chunk->length_dw * 4 /
-		sizeof(struct drm_amdgpu_cs_chunk_syncobj);
-	for (i = 0; i < num_deps; ++i) {
-		r = amdgpu_syncobj_lookup_and_add_to_sync(p,
-							  syncobj_deps[i].handle,
-							  syncobj_deps[i].point,
-							  syncobj_deps[i].flags);
+	for (i = 0; i < p->gang_size; ++i) {
+		r = amdgpu_sync_push_to_job(&p->sync, p->jobs[i]);
 		if (r)
 			return r;
 	}
 
-	return 0;
-}
-
-static int amdgpu_cs_process_syncobj_out_dep(struct amdgpu_cs_parser *p,
-					     struct amdgpu_cs_chunk *chunk)
-{
-	struct drm_amdgpu_cs_chunk_sem *deps;
-	unsigned num_deps;
-	int i;
-
-	deps = (struct drm_amdgpu_cs_chunk_sem *)chunk->kdata;
-	num_deps = chunk->length_dw * 4 /
-		sizeof(struct drm_amdgpu_cs_chunk_sem);
-
-	if (p->post_deps)
-		return -EINVAL;
-
-	p->post_deps = kmalloc_array(num_deps, sizeof(*p->post_deps),
-				     GFP_KERNEL);
-	p->num_post_deps = 0;
-
-	if (!p->post_deps)
-		return -ENOMEM;
-
-
-	for (i = 0; i < num_deps; ++i) {
-		p->post_deps[i].syncobj =
-			drm_syncobj_find(p->filp, deps[i].handle);
-		if (!p->post_deps[i].syncobj)
-			return -EINVAL;
-		p->post_deps[i].chain = NULL;
-		p->post_deps[i].point = 0;
-		p->num_post_deps++;
-	}
-
-	return 0;
-}
-
+	sched = p->gang_leader->base.entity->rq->sched;
+	while ((fence = amdgpu_sync_get_fence(&p->sync))) {
+		struct drm_sched_fence *s_fence = to_drm_sched_fence(fence);
 
-static int amdgpu_cs_process_syncobj_timeline_out_dep(struct amdgpu_cs_parser *p,
-						      struct amdgpu_cs_chunk *chunk)
-{
-	struct drm_amdgpu_cs_chunk_syncobj *syncobj_deps;
-	unsigned num_deps;
-	int i;
-
-	syncobj_deps = (struct drm_amdgpu_cs_chunk_syncobj *)chunk->kdata;
-	num_deps = chunk->length_dw * 4 /
-		sizeof(struct drm_amdgpu_cs_chunk_syncobj);
-
-	if (p->post_deps)
-		return -EINVAL;
-
-	p->post_deps = kmalloc_array(num_deps, sizeof(*p->post_deps),
-				     GFP_KERNEL);
-	p->num_post_deps = 0;
-
-	if (!p->post_deps)
-		return -ENOMEM;
-
-	for (i = 0; i < num_deps; ++i) {
-		struct amdgpu_cs_post_dep *dep = &p->post_deps[i];
-
-		dep->chain = NULL;
-		if (syncobj_deps[i].point) {
-			dep->chain = dma_fence_chain_alloc();
-			if (!dep->chain)
-				return -ENOMEM;
-		}
-
-		dep->syncobj = drm_syncobj_find(p->filp,
-						syncobj_deps[i].handle);
-		if (!dep->syncobj) {
-			dma_fence_chain_free(dep->chain);
-			return -EINVAL;
+		/*
+		 * When we have an dependency it might be necessary to insert a
+		 * pipeline sync to make sure that all caches etc are flushed and the
+		 * next job actually sees the results from the previous one
+		 * before we start executing on the same scheduler ring.
+		 */
+		if (!s_fence || s_fence->sched != sched) {
+			dma_fence_put(fence);
+			continue;
 		}
-		dep->point = syncobj_deps[i].point;
-		p->num_post_deps++;
-	}
 
-	return 0;
-}
-
-static int amdgpu_cs_dependencies(struct amdgpu_device *adev,
-				  struct amdgpu_cs_parser *p)
-{
-	int i, r;
-
-	for (i = 0; i < p->nchunks; ++i) {
-		struct amdgpu_cs_chunk *chunk;
-
-		chunk = &p->chunks[i];
-
-		switch (chunk->chunk_id) {
-		case AMDGPU_CHUNK_ID_DEPENDENCIES:
-		case AMDGPU_CHUNK_ID_SCHEDULED_DEPENDENCIES:
-			r = amdgpu_cs_process_fence_dep(p, chunk);
-			if (r)
-				return r;
-			break;
-		case AMDGPU_CHUNK_ID_SYNCOBJ_IN:
-			r = amdgpu_cs_process_syncobj_in_dep(p, chunk);
-			if (r)
-				return r;
-			break;
-		case AMDGPU_CHUNK_ID_SYNCOBJ_OUT:
-			r = amdgpu_cs_process_syncobj_out_dep(p, chunk);
-			if (r)
-				return r;
-			break;
-		case AMDGPU_CHUNK_ID_SYNCOBJ_TIMELINE_WAIT:
-			r = amdgpu_cs_process_syncobj_timeline_in_dep(p, chunk);
-			if (r)
-				return r;
-			break;
-		case AMDGPU_CHUNK_ID_SYNCOBJ_TIMELINE_SIGNAL:
-			r = amdgpu_cs_process_syncobj_timeline_out_dep(p, chunk);
-			if (r)
-				return r;
-			break;
-		}
+		r = amdgpu_sync_fence(&p->gang_leader->explicit_sync, fence,
+				      GFP_KERNEL);
+		dma_fence_put(fence);
+		if (r)
+			return r;
 	}
-
 	return 0;
 }
 
@@ -1209,20 +1284,36 @@ static int amdgpu_cs_submit(struct amdgpu_cs_parser *p,
 			    union drm_amdgpu_cs *cs)
 {
 	struct amdgpu_fpriv *fpriv = p->filp->driver_priv;
-	struct drm_sched_entity *entity = p->entity;
+	struct amdgpu_job *leader = p->gang_leader;
 	struct amdgpu_bo_list_entry *e;
-	struct amdgpu_job *job;
+	struct drm_gem_object *gobj;
+	unsigned long index;
+	unsigned int i;
 	uint64_t seq;
 	int r;
 
-	job = p->job;
-	p->job = NULL;
+	for (i = 0; i < p->gang_size; ++i)
+		drm_sched_job_arm(&p->jobs[i]->base);
 
-	r = drm_sched_job_init(&job->base, entity, &fpriv->vm);
-	if (r)
-		goto error_unlock;
+	for (i = 0; i < p->gang_size; ++i) {
+		struct dma_fence *fence;
+
+		if (p->jobs[i] == leader)
+			continue;
+
+		fence = &p->jobs[i]->base.s_fence->scheduled;
+		dma_fence_get(fence);
+		r = drm_sched_job_add_dependency(&leader->base, fence);
+		if (r) {
+			dma_fence_put(fence);
+			return r;
+		}
+	}
 
-	drm_sched_job_arm(&job->base);
+	if (p->gang_size > 1) {
+		for (i = 0; i < p->gang_size; ++i)
+			amdgpu_job_set_gang_leader(p->jobs[i], leader);
+	}
 
 	/* No memory allocation is allowed while holding the notifier lock.
 	 * The lock is held until amdgpu_cs_submit is finished and fence is
@@ -1233,91 +1324,100 @@ static int amdgpu_cs_submit(struct amdgpu_cs_parser *p,
 	/* If userptr are invalidated after amdgpu_cs_parser_bos(), return
 	 * -EAGAIN, drmIoctl in libdrm will restart the amdgpu_cs_ioctl.
 	 */
+	r = 0;
 	amdgpu_bo_list_for_each_userptr_entry(e, p->bo_list) {
-		struct amdgpu_bo *bo = ttm_to_amdgpu_bo(e->tv.bo);
-
-		r |= !amdgpu_ttm_tt_get_user_pages_done(bo->tbo.ttm);
+		r |= !amdgpu_hmm_range_valid(e->range);
+		amdgpu_hmm_range_free(e->range);
+		e->range = NULL;
 	}
 	if (r) {
 		r = -EAGAIN;
-		goto error_abort;
+		mutex_unlock(&p->adev->notifier_lock);
+		return r;
 	}
 
-	p->fence = dma_fence_get(&job->base.s_fence->finished);
+	p->fence = dma_fence_get(&leader->base.s_fence->finished);
+	drm_exec_for_each_locked_object(&p->exec, index, gobj) {
+
+		ttm_bo_move_to_lru_tail_unlocked(&gem_to_amdgpu_bo(gobj)->tbo);
+
+		/* Everybody except for the gang leader uses READ */
+		for (i = 0; i < p->gang_size; ++i) {
+			if (p->jobs[i] == leader)
+				continue;
 
-	amdgpu_ctx_add_fence(p->ctx, entity, p->fence, &seq);
+			dma_resv_add_fence(gobj->resv,
+					   &p->jobs[i]->base.s_fence->finished,
+					   DMA_RESV_USAGE_READ);
+		}
+
+		/* The gang leader as remembered as writer */
+		dma_resv_add_fence(gobj->resv, p->fence, DMA_RESV_USAGE_WRITE);
+	}
+
+	seq = amdgpu_ctx_add_fence(p->ctx, p->entities[p->gang_leader_idx],
+				   p->fence);
 	amdgpu_cs_post_dependencies(p);
 
-	if ((job->preamble_status & AMDGPU_PREAMBLE_IB_PRESENT) &&
+	if ((leader->preamble_status & AMDGPU_PREAMBLE_IB_PRESENT) &&
 	    !p->ctx->preamble_presented) {
-		job->preamble_status |= AMDGPU_PREAMBLE_IB_PRESENT_FIRST;
+		leader->preamble_status |= AMDGPU_PREAMBLE_IB_PRESENT_FIRST;
 		p->ctx->preamble_presented = true;
 	}
 
 	cs->out.handle = seq;
-	job->uf_sequence = seq;
-
-	amdgpu_job_free_resources(job);
-
-	trace_amdgpu_cs_ioctl(job);
-	amdgpu_vm_bo_trace_cs(&fpriv->vm, &p->ticket);
-	drm_sched_entity_push_job(&job->base);
+	leader->uf_sequence = seq;
+
+	amdgpu_vm_bo_trace_cs(&fpriv->vm, &p->exec.ticket);
+	for (i = 0; i < p->gang_size; ++i) {
+		amdgpu_job_free_resources(p->jobs[i]);
+		trace_amdgpu_cs_ioctl(p->jobs[i]);
+		drm_sched_entity_push_job(&p->jobs[i]->base);
+		p->jobs[i] = NULL;
+	}
 
 	amdgpu_vm_move_to_lru_tail(p->adev, &fpriv->vm);
 
-	amdgpu_bo_list_for_each_entry(e, p->bo_list) {
-		struct dma_resv *resv = e->tv.bo->base.resv;
-		struct dma_fence_chain *chain = e->chain;
-
-		if (!chain)
-			continue;
-
-		/*
-		 * Work around dma_resv shortcommings by wrapping up the
-		 * submission in a dma_fence_chain and add it as exclusive
-		 * fence, but first add the submission as shared fence to make
-		 * sure that shared fences never signal before the exclusive
-		 * one.
-		 */
-		dma_fence_chain_init(chain, dma_resv_excl_fence(resv),
-				     dma_fence_get(p->fence), 1);
-
-		dma_resv_add_shared_fence(resv, p->fence);
-		rcu_assign_pointer(resv->fence_excl, &chain->base);
-		e->chain = NULL;
-	}
-
-	ttm_eu_fence_buffer_objects(&p->ticket, &p->validated, p->fence);
 	mutex_unlock(&p->adev->notifier_lock);
-
+	mutex_unlock(&p->bo_list->bo_list_mutex);
 	return 0;
-
-error_abort:
-	drm_sched_job_cleanup(&job->base);
-	mutex_unlock(&p->adev->notifier_lock);
-
-error_unlock:
-	amdgpu_job_free(job);
-	return r;
 }
 
-static void trace_amdgpu_cs_ibs(struct amdgpu_cs_parser *parser)
+/* Cleanup the parser structure */
+static void amdgpu_cs_parser_fini(struct amdgpu_cs_parser *parser)
 {
-	int i;
+	unsigned int i;
 
-	if (!trace_amdgpu_cs_enabled())
-		return;
+	amdgpu_sync_free(&parser->sync);
+	drm_exec_fini(&parser->exec);
+
+	for (i = 0; i < parser->num_post_deps; i++) {
+		drm_syncobj_put(parser->post_deps[i].syncobj);
+		kfree(parser->post_deps[i].chain);
+	}
+	kfree(parser->post_deps);
 
-	for (i = 0; i < parser->job->num_ibs; i++)
-		trace_amdgpu_cs(parser, i);
+	dma_fence_put(parser->fence);
+
+	if (parser->ctx)
+		amdgpu_ctx_put(parser->ctx);
+	if (parser->bo_list)
+		amdgpu_bo_list_put(parser->bo_list);
+
+	for (i = 0; i < parser->nchunks; i++)
+		kvfree(parser->chunks[i].kdata);
+	kvfree(parser->chunks);
+	for (i = 0; i < parser->gang_size; ++i) {
+		if (parser->jobs[i])
+			amdgpu_job_free(parser->jobs[i]);
+	}
+	amdgpu_bo_unref(&parser->uf_bo);
 }
 
 int amdgpu_cs_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 {
 	struct amdgpu_device *adev = drm_to_adev(dev);
-	union drm_amdgpu_cs *cs = data;
-	struct amdgpu_cs_parser parser = {};
-	bool reserved_buffers = false;
+	struct amdgpu_cs_parser parser;
 	int r;
 
 	if (amdgpu_ras_intr_triggered())
@@ -1326,48 +1426,55 @@ int amdgpu_cs_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 	if (!adev->accel_working)
 		return -EBUSY;
 
-	parser.adev = adev;
-	parser.filp = filp;
-
-	r = amdgpu_cs_parser_init(&parser, data);
+	r = amdgpu_cs_parser_init(&parser, adev, filp, data);
 	if (r) {
-		if (printk_ratelimit())
-			DRM_ERROR("Failed to initialize parser %d!\n", r);
-		goto out;
+		drm_err_ratelimited(dev, "Failed to initialize parser %d!\n", r);
+		return r;
 	}
 
-	r = amdgpu_cs_ib_fill(adev, &parser);
+	r = amdgpu_cs_pass1(&parser, data);
 	if (r)
-		goto out;
+		goto error_fini;
 
-	r = amdgpu_cs_dependencies(adev, &parser);
-	if (r) {
-		DRM_ERROR("Failed in the dependencies handling %d!\n", r);
-		goto out;
-	}
+	r = amdgpu_cs_pass2(&parser);
+	if (r)
+		goto error_fini;
 
 	r = amdgpu_cs_parser_bos(&parser, data);
 	if (r) {
 		if (r == -ENOMEM)
-			DRM_ERROR("Not enough memory for command submission!\n");
+			drm_err(dev, "Not enough memory for command submission!\n");
 		else if (r != -ERESTARTSYS && r != -EAGAIN)
-			DRM_ERROR("Failed to process the buffer list %d!\n", r);
-		goto out;
+			drm_dbg(dev, "Failed to process the buffer list %d!\n", r);
+		goto error_fini;
 	}
 
-	reserved_buffers = true;
+	r = amdgpu_cs_patch_jobs(&parser);
+	if (r)
+		goto error_backoff;
+
+	r = amdgpu_cs_vm_handling(&parser);
+	if (r)
+		goto error_backoff;
+
+	r = amdgpu_cs_sync_rings(&parser);
+	if (r)
+		goto error_backoff;
 
 	trace_amdgpu_cs_ibs(&parser);
 
-	r = amdgpu_cs_vm_handling(&parser);
+	r = amdgpu_cs_submit(&parser, data);
 	if (r)
-		goto out;
+		goto error_backoff;
 
-	r = amdgpu_cs_submit(&parser, cs);
+	amdgpu_cs_parser_fini(&parser);
+	return 0;
 
-out:
-	amdgpu_cs_parser_fini(&parser, r, reserved_buffers);
+error_backoff:
+	mutex_unlock(&parser.bo_list->bo_list_mutex);
 
+error_fini:
+	amdgpu_cs_parser_fini(&parser);
 	return r;
 }
 
@@ -1510,6 +1617,7 @@ int amdgpu_cs_fence_to_handle_ioctl(struct drm_device *dev, void *data,
 		return 0;
 
 	default:
+		dma_fence_put(fence);
 		return -EINVAL;
 	}
 }
@@ -1542,15 +1650,15 @@ static int amdgpu_cs_wait_all_fences(struct amdgpu_device *adev,
 			continue;
 
 		r = dma_fence_wait_timeout(fence, true, timeout);
+		if (r > 0 && fence->error)
+			r = fence->error;
+
 		dma_fence_put(fence);
 		if (r < 0)
 			return r;
 
 		if (r == 0)
 			break;
-
-		if (fence->error)
-			return fence->error;
 	}
 
 	memset(wait, 0, sizeof(*wait));
@@ -1636,30 +1744,21 @@ int amdgpu_cs_wait_fences_ioctl(struct drm_device *dev, void *data,
 {
 	struct amdgpu_device *adev = drm_to_adev(dev);
 	union drm_amdgpu_wait_fences *wait = data;
-	uint32_t fence_count = wait->in.fence_count;
-	struct drm_amdgpu_fence *fences_user;
 	struct drm_amdgpu_fence *fences;
 	int r;
 
 	/* Get the fences from userspace */
-	fences = kmalloc_array(fence_count, sizeof(struct drm_amdgpu_fence),
-			GFP_KERNEL);
-	if (fences == NULL)
-		return -ENOMEM;
-
-	fences_user = u64_to_user_ptr(wait->in.fences);
-	if (copy_from_user(fences, fences_user,
-		sizeof(struct drm_amdgpu_fence) * fence_count)) {
-		r = -EFAULT;
-		goto err_free_fences;
-	}
+	fences = memdup_array_user(u64_to_user_ptr(wait->in.fences),
+				   wait->in.fence_count,
+				   sizeof(struct drm_amdgpu_fence));
+	if (IS_ERR(fences))
+		return PTR_ERR(fences);
 
 	if (wait->in.wait_all)
 		r = amdgpu_cs_wait_all_fences(adev, filp, wait, fences);
 	else
 		r = amdgpu_cs_wait_any_fence(adev, filp, wait, fences);
 
-err_free_fences:
 	kfree(fences);
 
 	return r;
@@ -1685,7 +1784,7 @@ int amdgpu_cs_find_mapping(struct amdgpu_cs_parser *parser,
 	struct ttm_operation_ctx ctx = { false, false };
 	struct amdgpu_vm *vm = &fpriv->vm;
 	struct amdgpu_bo_va_mapping *mapping;
-	int r;
+	int i, r;
 
 	addr /= AMDGPU_GPU_PAGE_SIZE;
 
@@ -1697,12 +1796,17 @@ int amdgpu_cs_find_mapping(struct amdgpu_cs_parser *parser,
 	*map = mapping;
 
 	/* Double check that the BO is reserved by this CS */
-	if (dma_resv_locking_ctx((*bo)->tbo.base.resv) != &parser->ticket)
+	if (dma_resv_locking_ctx((*bo)->tbo.base.resv) != &parser->exec.ticket)
 		return -EINVAL;
 
-	if (!((*bo)->flags & AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS)) {
-		(*bo)->flags |= AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS;
+	/* Make sure VRAM is allocated contigiously */
+	(*bo)->flags |= AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS;
+	if ((*bo)->tbo.resource->mem_type == TTM_PL_VRAM &&
+	    !((*bo)->tbo.resource->placement & TTM_PL_FLAG_CONTIGUOUS)) {
+
 		amdgpu_bo_placement_from_domain(*bo, (*bo)->allowed_domains);
+		for (i = 0; i < (*bo)->placement.num_placement; i++)
+			(*bo)->placements[i].flags |= TTM_PL_FLAG_CONTIGUOUS;
 		r = ttm_bo_validate(&(*bo)->tbo, &(*bo)->placement, &ctx);
 		if (r)
 			return r;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.h
new file mode 100644
index 000000000000..39c33ad100cb
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.h
@@ -0,0 +1,88 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __AMDGPU_CS_H__
+#define __AMDGPU_CS_H__
+
+#include <linux/ww_mutex.h>
+#include <drm/drm_exec.h>
+
+#include "amdgpu_job.h"
+#include "amdgpu_bo_list.h"
+#include "amdgpu_ring.h"
+
+#define AMDGPU_CS_GANG_SIZE	4
+
+struct amdgpu_bo_va_mapping;
+
+struct amdgpu_cs_chunk {
+	uint32_t		chunk_id;
+	uint32_t		length_dw;
+	void			*kdata;
+};
+
+struct amdgpu_cs_post_dep {
+	struct drm_syncobj *syncobj;
+	struct dma_fence_chain *chain;
+	u64 point;
+};
+
+struct amdgpu_cs_parser {
+	struct amdgpu_device	*adev;
+	struct drm_file		*filp;
+	struct amdgpu_ctx	*ctx;
+
+	/* chunks */
+	unsigned		nchunks;
+	struct amdgpu_cs_chunk	*chunks;
+
+	/* scheduler job objects */
+	unsigned int		gang_size;
+	unsigned int		gang_leader_idx;
+	struct drm_sched_entity	*entities[AMDGPU_CS_GANG_SIZE];
+	struct amdgpu_job	*jobs[AMDGPU_CS_GANG_SIZE];
+	struct amdgpu_job	*gang_leader;
+
+	/* buffer objects */
+	struct drm_exec			exec;
+	struct amdgpu_bo_list		*bo_list;
+	struct amdgpu_mn		*mn;
+	struct dma_fence		*fence;
+	uint64_t			bytes_moved_threshold;
+	uint64_t			bytes_moved_vis_threshold;
+	uint64_t			bytes_moved;
+	uint64_t			bytes_moved_vis;
+
+	/* user fence */
+	struct amdgpu_bo		*uf_bo;
+
+	unsigned			num_post_deps;
+	struct amdgpu_cs_post_dep	*post_deps;
+
+	struct amdgpu_sync		sync;
+};
+
+int amdgpu_cs_find_mapping(struct amdgpu_cs_parser *parser,
+			   uint64_t addr, struct amdgpu_bo **bo,
+			   struct amdgpu_bo_va_mapping **mapping);
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_csa.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_csa.c
index da21e60bb827..02138aa55793 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_csa.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_csa.c
@@ -22,13 +22,14 @@
  * * Author: Monk.liu@amd.com
  */
 
+#include <drm/drm_exec.h>
+
 #include "amdgpu.h"
 
 uint64_t amdgpu_csa_vaddr(struct amdgpu_device *adev)
 {
-	uint64_t addr = adev->vm_manager.max_pfn << AMDGPU_GPU_PAGE_SHIFT;
+	uint64_t addr = AMDGPU_VA_RESERVED_CSA_START(adev);
 
-	addr -= AMDGPU_VA_RESERVED_SIZE;
 	addr = amdgpu_gmc_sign_extend(addr);
 
 	return addr;
@@ -65,31 +66,25 @@ int amdgpu_map_static_csa(struct amdgpu_device *adev, struct amdgpu_vm *vm,
 			  struct amdgpu_bo *bo, struct amdgpu_bo_va **bo_va,
 			  uint64_t csa_addr, uint32_t size)
 {
-	struct ww_acquire_ctx ticket;
-	struct list_head list;
-	struct amdgpu_bo_list_entry pd;
-	struct ttm_validate_buffer csa_tv;
+	struct drm_exec exec;
 	int r;
 
-	INIT_LIST_HEAD(&list);
-	INIT_LIST_HEAD(&csa_tv.head);
-	csa_tv.bo = &bo->tbo;
-	csa_tv.num_shared = 1;
-
-	list_add(&csa_tv.head, &list);
-	amdgpu_vm_get_pd_bo(vm, &list, &pd);
-
-	r = ttm_eu_reserve_buffers(&ticket, &list, true, NULL);
-	if (r) {
-		DRM_ERROR("failed to reserve CSA,PD BOs: err=%d\n", r);
-		return r;
+	drm_exec_init(&exec, DRM_EXEC_INTERRUPTIBLE_WAIT, 0);
+	drm_exec_until_all_locked(&exec) {
+		r = amdgpu_vm_lock_pd(vm, &exec, 0);
+		if (likely(!r))
+			r = drm_exec_lock_obj(&exec, &bo->tbo.base);
+		drm_exec_retry_on_contention(&exec);
+		if (unlikely(r)) {
+			DRM_ERROR("failed to reserve CSA,PD BOs: err=%d\n", r);
+			goto error;
+		}
 	}
 
 	*bo_va = amdgpu_vm_bo_add(adev, vm, bo);
 	if (!*bo_va) {
-		ttm_eu_backoff_reservation(&ticket, &list);
-		DRM_ERROR("failed to create bo_va for static CSA\n");
-		return -ENOMEM;
+		r = -ENOMEM;
+		goto error;
 	}
 
 	r = amdgpu_vm_bo_map(adev, *bo_va, csa_addr, 0, size,
@@ -98,11 +93,43 @@ int amdgpu_map_static_csa(struct amdgpu_device *adev, struct amdgpu_vm *vm,
 
 	if (r) {
 		DRM_ERROR("failed to do bo_map on static CSA, err=%d\n", r);
-		amdgpu_vm_bo_rmv(adev, *bo_va);
-		ttm_eu_backoff_reservation(&ticket, &list);
-		return r;
+		amdgpu_vm_bo_del(adev, *bo_va);
+		goto error;
 	}
 
-	ttm_eu_backoff_reservation(&ticket, &list);
-	return 0;
+error:
+	drm_exec_fini(&exec);
+	return r;
+}
+
+int amdgpu_unmap_static_csa(struct amdgpu_device *adev, struct amdgpu_vm *vm,
+			    struct amdgpu_bo *bo, struct amdgpu_bo_va *bo_va,
+			    uint64_t csa_addr)
+{
+	struct drm_exec exec;
+	int r;
+
+	drm_exec_init(&exec, 0, 0);
+	drm_exec_until_all_locked(&exec) {
+		r = amdgpu_vm_lock_pd(vm, &exec, 0);
+		if (likely(!r))
+			r = drm_exec_lock_obj(&exec, &bo->tbo.base);
+		drm_exec_retry_on_contention(&exec);
+		if (unlikely(r)) {
+			DRM_ERROR("failed to reserve CSA,PD BOs: err=%d\n", r);
+			goto error;
+		}
+	}
+
+	r = amdgpu_vm_bo_unmap(adev, bo_va, csa_addr);
+	if (r) {
+		DRM_ERROR("failed to do bo_unmap on static CSA, err=%d\n", r);
+		goto error;
+	}
+
+	amdgpu_vm_bo_del(adev, bo_va);
+
+error:
+	drm_exec_fini(&exec);
+	return r;
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_csa.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_csa.h
index 524b4437a021..7dfc1f2012eb 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_csa.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_csa.h
@@ -34,6 +34,9 @@ int amdgpu_allocate_static_csa(struct amdgpu_device *adev, struct amdgpu_bo **bo
 int amdgpu_map_static_csa(struct amdgpu_device *adev, struct amdgpu_vm *vm,
 			  struct amdgpu_bo *bo, struct amdgpu_bo_va **bo_va,
 			  uint64_t csa_addr, uint32_t size);
+int amdgpu_unmap_static_csa(struct amdgpu_device *adev, struct amdgpu_vm *vm,
+			    struct amdgpu_bo *bo, struct amdgpu_bo_va *bo_va,
+			    uint64_t csa_addr);
 void amdgpu_free_static_csa(struct amdgpu_bo **bo);
 
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c
index 468003583b2a..afedea02188d 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c
@@ -23,6 +23,7 @@
  */
 
 #include <drm/drm_auth.h>
+#include <drm/drm_drv.h>
 #include "amdgpu.h"
 #include "amdgpu_sched.h"
 #include "amdgpu_ras.h"
@@ -41,12 +42,12 @@ const unsigned int amdgpu_ctx_num_entities[AMDGPU_HW_IP_NUM] = {
 	[AMDGPU_HW_IP_VCN_DEC]	=	1,
 	[AMDGPU_HW_IP_VCN_ENC]	=	1,
 	[AMDGPU_HW_IP_VCN_JPEG]	=	1,
+	[AMDGPU_HW_IP_VPE]	=	1,
 };
 
 bool amdgpu_ctx_priority_is_valid(int32_t ctx_prio)
 {
 	switch (ctx_prio) {
-	case AMDGPU_CTX_PRIORITY_UNSET:
 	case AMDGPU_CTX_PRIORITY_VERY_LOW:
 	case AMDGPU_CTX_PRIORITY_LOW:
 	case AMDGPU_CTX_PRIORITY_NORMAL:
@@ -54,6 +55,11 @@ bool amdgpu_ctx_priority_is_valid(int32_t ctx_prio)
 	case AMDGPU_CTX_PRIORITY_VERY_HIGH:
 		return true;
 	default:
+	case AMDGPU_CTX_PRIORITY_UNSET:
+		/* UNSET priority is not valid and we don't carry that
+		 * around, but set it to NORMAL in the only place this
+		 * function is called, amdgpu_ctx_ioctl().
+		 */
 		return false;
 	}
 }
@@ -63,13 +69,14 @@ amdgpu_ctx_to_drm_sched_prio(int32_t ctx_prio)
 {
 	switch (ctx_prio) {
 	case AMDGPU_CTX_PRIORITY_UNSET:
-		return DRM_SCHED_PRIORITY_UNSET;
+		pr_warn_once("AMD-->DRM context priority value UNSET-->NORMAL");
+		return DRM_SCHED_PRIORITY_NORMAL;
 
 	case AMDGPU_CTX_PRIORITY_VERY_LOW:
-		return DRM_SCHED_PRIORITY_MIN;
+		return DRM_SCHED_PRIORITY_LOW;
 
 	case AMDGPU_CTX_PRIORITY_LOW:
-		return DRM_SCHED_PRIORITY_MIN;
+		return DRM_SCHED_PRIORITY_LOW;
 
 	case AMDGPU_CTX_PRIORITY_NORMAL:
 		return DRM_SCHED_PRIORITY_NORMAL;
@@ -93,9 +100,6 @@ amdgpu_ctx_to_drm_sched_prio(int32_t ctx_prio)
 static int amdgpu_ctx_priority_permit(struct drm_file *filp,
 				      int32_t priority)
 {
-	if (!amdgpu_ctx_priority_is_valid(priority))
-		return -EINVAL;
-
 	/* NORMAL and below are accessible by everyone */
 	if (priority <= AMDGPU_CTX_PRIORITY_NORMAL)
 		return 0;
@@ -109,7 +113,7 @@ static int amdgpu_ctx_priority_permit(struct drm_file *filp,
 	return -EACCES;
 }
 
-static enum amdgpu_gfx_pipe_priority amdgpu_ctx_prio_to_compute_prio(int32_t prio)
+static enum amdgpu_gfx_pipe_priority amdgpu_ctx_prio_to_gfx_pipe_prio(int32_t prio)
 {
 	switch (prio) {
 	case AMDGPU_CTX_PRIORITY_HIGH:
@@ -134,16 +138,17 @@ static enum amdgpu_ring_priority_level amdgpu_ctx_sched_prio_to_ring_prio(int32_
 
 static unsigned int amdgpu_ctx_get_hw_prio(struct amdgpu_ctx *ctx, u32 hw_ip)
 {
-	struct amdgpu_device *adev = ctx->adev;
-	int32_t ctx_prio;
+	struct amdgpu_device *adev = ctx->mgr->adev;
 	unsigned int hw_prio;
+	int32_t ctx_prio;
 
 	ctx_prio = (ctx->override_priority == AMDGPU_CTX_PRIORITY_UNSET) ?
 			ctx->init_priority : ctx->override_priority;
 
 	switch (hw_ip) {
+	case AMDGPU_HW_IP_GFX:
 	case AMDGPU_HW_IP_COMPUTE:
-		hw_prio = amdgpu_ctx_prio_to_compute_prio(ctx_prio);
+		hw_prio = amdgpu_ctx_prio_to_gfx_pipe_prio(ctx_prio);
 		break;
 	case AMDGPU_HW_IP_VCE:
 	case AMDGPU_HW_IP_VCN_ENC:
@@ -161,17 +166,50 @@ static unsigned int amdgpu_ctx_get_hw_prio(struct amdgpu_ctx *ctx, u32 hw_ip)
 	return hw_prio;
 }
 
+/* Calculate the time spend on the hw */
+static ktime_t amdgpu_ctx_fence_time(struct dma_fence *fence)
+{
+	struct drm_sched_fence *s_fence;
+
+	if (!fence)
+		return ns_to_ktime(0);
+
+	/* When the fence is not even scheduled it can't have spend time */
+	s_fence = to_drm_sched_fence(fence);
+	if (!test_bit(DMA_FENCE_FLAG_TIMESTAMP_BIT, &s_fence->scheduled.flags))
+		return ns_to_ktime(0);
+
+	/* When it is still running account how much already spend */
+	if (!test_bit(DMA_FENCE_FLAG_TIMESTAMP_BIT, &s_fence->finished.flags))
+		return ktime_sub(ktime_get(), s_fence->scheduled.timestamp);
+
+	return ktime_sub(s_fence->finished.timestamp,
+			 s_fence->scheduled.timestamp);
+}
+
+static ktime_t amdgpu_ctx_entity_time(struct amdgpu_ctx *ctx,
+				      struct amdgpu_ctx_entity *centity)
+{
+	ktime_t res = ns_to_ktime(0);
+	uint32_t i;
+
+	spin_lock(&ctx->ring_lock);
+	for (i = 0; i < amdgpu_sched_jobs; i++) {
+		res = ktime_add(res, amdgpu_ctx_fence_time(centity->fences[i]));
+	}
+	spin_unlock(&ctx->ring_lock);
+	return res;
+}
 
 static int amdgpu_ctx_init_entity(struct amdgpu_ctx *ctx, u32 hw_ip,
 				  const u32 ring)
 {
-	struct amdgpu_device *adev = ctx->adev;
-	struct amdgpu_ctx_entity *entity;
 	struct drm_gpu_scheduler **scheds = NULL, *sched = NULL;
-	unsigned num_scheds = 0;
-	int32_t ctx_prio;
-	unsigned int hw_prio;
+	struct amdgpu_device *adev = ctx->mgr->adev;
+	struct amdgpu_ctx_entity *entity;
 	enum drm_sched_priority drm_prio;
+	unsigned int hw_prio, num_scheds;
+	int32_t ctx_prio;
 	int r;
 
 	entity = kzalloc(struct_size(entity, fences, amdgpu_sched_jobs),
@@ -181,13 +219,25 @@ static int amdgpu_ctx_init_entity(struct amdgpu_ctx *ctx, u32 hw_ip,
 
 	ctx_prio = (ctx->override_priority == AMDGPU_CTX_PRIORITY_UNSET) ?
 			ctx->init_priority : ctx->override_priority;
+	entity->hw_ip = hw_ip;
 	entity->sequence = 1;
 	hw_prio = amdgpu_ctx_get_hw_prio(ctx, hw_ip);
 	drm_prio = amdgpu_ctx_to_drm_sched_prio(ctx_prio);
 
 	hw_ip = array_index_nospec(hw_ip, AMDGPU_HW_IP_NUM);
-	scheds = adev->gpu_sched[hw_ip][hw_prio].sched;
-	num_scheds = adev->gpu_sched[hw_ip][hw_prio].num_scheds;
+
+	if (!(adev)->xcp_mgr) {
+		scheds = adev->gpu_sched[hw_ip][hw_prio].sched;
+		num_scheds = adev->gpu_sched[hw_ip][hw_prio].num_scheds;
+	} else {
+		struct amdgpu_fpriv *fpriv;
+
+		fpriv = container_of(ctx->ctx_mgr, struct amdgpu_fpriv, ctx_mgr);
+		r = amdgpu_xcp_select_scheds(adev, hw_ip, hw_prio, fpriv,
+						&num_scheds, &scheds);
+		if (r)
+			goto error_free_entity;
+	}
 
 	/* disable load balance if the hw engine retains context among dependent jobs */
 	if (hw_ip == AMDGPU_HW_IP_VCN_ENC ||
@@ -204,20 +254,74 @@ static int amdgpu_ctx_init_entity(struct amdgpu_ctx *ctx, u32 hw_ip,
 	if (r)
 		goto error_free_entity;
 
-	ctx->entities[hw_ip][ring] = entity;
+	/* It's not an error if we fail to install the new entity */
+	if (cmpxchg(&ctx->entities[hw_ip][ring], NULL, entity))
+		goto cleanup_entity;
+
 	return 0;
 
+cleanup_entity:
+	drm_sched_entity_fini(&entity->entity);
+
 error_free_entity:
 	kfree(entity);
 
 	return r;
 }
 
-static int amdgpu_ctx_init(struct amdgpu_device *adev,
-			   int32_t priority,
-			   struct drm_file *filp,
-			   struct amdgpu_ctx *ctx)
+static ktime_t amdgpu_ctx_fini_entity(struct amdgpu_device *adev,
+				  struct amdgpu_ctx_entity *entity)
 {
+	ktime_t res = ns_to_ktime(0);
+	int i;
+
+	if (!entity)
+		return res;
+
+	for (i = 0; i < amdgpu_sched_jobs; ++i) {
+		res = ktime_add(res, amdgpu_ctx_fence_time(entity->fences[i]));
+		dma_fence_put(entity->fences[i]);
+	}
+
+	amdgpu_xcp_release_sched(adev, entity);
+
+	kfree(entity);
+	return res;
+}
+
+static int amdgpu_ctx_get_stable_pstate(struct amdgpu_ctx *ctx,
+					u32 *stable_pstate)
+{
+	struct amdgpu_device *adev = ctx->mgr->adev;
+	enum amd_dpm_forced_level current_level;
+
+	current_level = amdgpu_dpm_get_performance_level(adev);
+
+	switch (current_level) {
+	case AMD_DPM_FORCED_LEVEL_PROFILE_STANDARD:
+		*stable_pstate = AMDGPU_CTX_STABLE_PSTATE_STANDARD;
+		break;
+	case AMD_DPM_FORCED_LEVEL_PROFILE_MIN_SCLK:
+		*stable_pstate = AMDGPU_CTX_STABLE_PSTATE_MIN_SCLK;
+		break;
+	case AMD_DPM_FORCED_LEVEL_PROFILE_MIN_MCLK:
+		*stable_pstate = AMDGPU_CTX_STABLE_PSTATE_MIN_MCLK;
+		break;
+	case AMD_DPM_FORCED_LEVEL_PROFILE_PEAK:
+		*stable_pstate = AMDGPU_CTX_STABLE_PSTATE_PEAK;
+		break;
+	default:
+		*stable_pstate = AMDGPU_CTX_STABLE_PSTATE_NONE;
+		break;
+	}
+	return 0;
+}
+
+static int amdgpu_ctx_init(struct amdgpu_ctx_mgr *mgr, int32_t priority,
+			   struct drm_file *filp, struct amdgpu_ctx *ctx)
+{
+	struct amdgpu_fpriv *fpriv = filp->driver_priv;
+	u32 current_stable_pstate;
 	int r;
 
 	r = amdgpu_ctx_priority_permit(filp, priority);
@@ -226,52 +330,104 @@ static int amdgpu_ctx_init(struct amdgpu_device *adev,
 
 	memset(ctx, 0, sizeof(*ctx));
 
-	ctx->adev = adev;
-
 	kref_init(&ctx->refcount);
+	ctx->mgr = mgr;
 	spin_lock_init(&ctx->ring_lock);
-	mutex_init(&ctx->lock);
 
-	ctx->reset_counter = atomic_read(&adev->gpu_reset_counter);
+	ctx->reset_counter = atomic_read(&mgr->adev->gpu_reset_counter);
 	ctx->reset_counter_query = ctx->reset_counter;
-	ctx->vram_lost_counter = atomic_read(&adev->vram_lost_counter);
+	ctx->generation = amdgpu_vm_generation(mgr->adev, &fpriv->vm);
 	ctx->init_priority = priority;
 	ctx->override_priority = AMDGPU_CTX_PRIORITY_UNSET;
 
+	r = amdgpu_ctx_get_stable_pstate(ctx, &current_stable_pstate);
+	if (r)
+		return r;
+
+	if (mgr->adev->pm.stable_pstate_ctx)
+		ctx->stable_pstate = mgr->adev->pm.stable_pstate_ctx->stable_pstate;
+	else
+		ctx->stable_pstate = current_stable_pstate;
+
+	ctx->ctx_mgr = &(fpriv->ctx_mgr);
 	return 0;
 }
 
-static void amdgpu_ctx_fini_entity(struct amdgpu_ctx_entity *entity)
+static int amdgpu_ctx_set_stable_pstate(struct amdgpu_ctx *ctx,
+					u32 stable_pstate)
 {
+	struct amdgpu_device *adev = ctx->mgr->adev;
+	enum amd_dpm_forced_level level;
+	u32 current_stable_pstate;
+	int r;
 
-	int i;
+	mutex_lock(&adev->pm.stable_pstate_ctx_lock);
+	if (adev->pm.stable_pstate_ctx && adev->pm.stable_pstate_ctx != ctx) {
+		r = -EBUSY;
+		goto done;
+	}
 
-	if (!entity)
-		return;
+	r = amdgpu_ctx_get_stable_pstate(ctx, &current_stable_pstate);
+	if (r || (stable_pstate == current_stable_pstate))
+		goto done;
 
-	for (i = 0; i < amdgpu_sched_jobs; ++i)
-		dma_fence_put(entity->fences[i]);
+	switch (stable_pstate) {
+	case AMDGPU_CTX_STABLE_PSTATE_NONE:
+		level = AMD_DPM_FORCED_LEVEL_AUTO;
+		break;
+	case AMDGPU_CTX_STABLE_PSTATE_STANDARD:
+		level = AMD_DPM_FORCED_LEVEL_PROFILE_STANDARD;
+		break;
+	case AMDGPU_CTX_STABLE_PSTATE_MIN_SCLK:
+		level = AMD_DPM_FORCED_LEVEL_PROFILE_MIN_SCLK;
+		break;
+	case AMDGPU_CTX_STABLE_PSTATE_MIN_MCLK:
+		level = AMD_DPM_FORCED_LEVEL_PROFILE_MIN_MCLK;
+		break;
+	case AMDGPU_CTX_STABLE_PSTATE_PEAK:
+		level = AMD_DPM_FORCED_LEVEL_PROFILE_PEAK;
+		break;
+	default:
+		r = -EINVAL;
+		goto done;
+	}
 
-	kfree(entity);
+	r = amdgpu_dpm_force_performance_level(adev, level);
+
+	if (level == AMD_DPM_FORCED_LEVEL_AUTO)
+		adev->pm.stable_pstate_ctx = NULL;
+	else
+		adev->pm.stable_pstate_ctx = ctx;
+done:
+	mutex_unlock(&adev->pm.stable_pstate_ctx_lock);
+
+	return r;
 }
 
 static void amdgpu_ctx_fini(struct kref *ref)
 {
 	struct amdgpu_ctx *ctx = container_of(ref, struct amdgpu_ctx, refcount);
-	struct amdgpu_device *adev = ctx->adev;
-	unsigned i, j;
+	struct amdgpu_ctx_mgr *mgr = ctx->mgr;
+	struct amdgpu_device *adev = mgr->adev;
+	unsigned i, j, idx;
 
 	if (!adev)
 		return;
 
 	for (i = 0; i < AMDGPU_HW_IP_NUM; ++i) {
 		for (j = 0; j < AMDGPU_MAX_ENTITY_NUM; ++j) {
-			amdgpu_ctx_fini_entity(ctx->entities[i][j]);
-			ctx->entities[i][j] = NULL;
+			ktime_t spend;
+
+			spend = amdgpu_ctx_fini_entity(adev, ctx->entities[i][j]);
+			atomic64_add(ktime_to_ns(spend), &mgr->time_spend[i]);
 		}
 	}
 
-	mutex_destroy(&ctx->lock);
+	if (drm_dev_enter(adev_to_drm(adev), &idx)) {
+		amdgpu_ctx_set_stable_pstate(ctx, ctx->stable_pstate);
+		drm_dev_exit(idx);
+	}
+
 	kfree(ctx);
 }
 
@@ -279,6 +435,7 @@ int amdgpu_ctx_get_entity(struct amdgpu_ctx *ctx, u32 hw_ip, u32 instance,
 			  u32 ring, struct drm_sched_entity **entity)
 {
 	int r;
+	struct drm_sched_entity *ctx_entity;
 
 	if (hw_ip >= AMDGPU_HW_IP_NUM) {
 		DRM_ERROR("unknown HW IP type: %d\n", hw_ip);
@@ -302,7 +459,14 @@ int amdgpu_ctx_get_entity(struct amdgpu_ctx *ctx, u32 hw_ip, u32 instance,
 			return r;
 	}
 
-	*entity = &ctx->entities[hw_ip][ring]->entity;
+	ctx_entity = &ctx->entities[hw_ip][ring]->entity;
+	r = drm_sched_entity_error(ctx_entity);
+	if (r) {
+		DRM_DEBUG("error entity %p\n", ctx_entity);
+		return r;
+	}
+
+	*entity = ctx_entity;
 	return 0;
 }
 
@@ -329,7 +493,7 @@ static int amdgpu_ctx_alloc(struct amdgpu_device *adev,
 	}
 
 	*id = (uint32_t)r;
-	r = amdgpu_ctx_init(adev, priority, filp, ctx);
+	r = amdgpu_ctx_init(mgr, priority, filp, ctx);
 	if (r) {
 		idr_remove(&mgr->ctx_handles, *id);
 		*id = 0;
@@ -433,12 +597,15 @@ static int amdgpu_ctx_query2(struct amdgpu_device *adev,
 	if (ctx->reset_counter != atomic_read(&adev->gpu_reset_counter))
 		out->state.flags |= AMDGPU_CTX_QUERY2_FLAGS_RESET;
 
-	if (ctx->vram_lost_counter != atomic_read(&adev->vram_lost_counter))
+	if (ctx->generation != amdgpu_vm_generation(adev, &fpriv->vm))
 		out->state.flags |= AMDGPU_CTX_QUERY2_FLAGS_VRAMLOST;
 
 	if (atomic_read(&ctx->guilty))
 		out->state.flags |= AMDGPU_CTX_QUERY2_FLAGS_GUILTY;
 
+	if (amdgpu_in_reset(adev))
+		out->state.flags |= AMDGPU_CTX_QUERY2_FLAGS_RESET_IN_PROGRESS;
+
 	if (adev->ras_enabled && con) {
 		/* Return the cached values in O(1),
 		 * and schedule delayed work to cache
@@ -467,11 +634,39 @@ static int amdgpu_ctx_query2(struct amdgpu_device *adev,
 	return 0;
 }
 
+static int amdgpu_ctx_stable_pstate(struct amdgpu_device *adev,
+				    struct amdgpu_fpriv *fpriv, uint32_t id,
+				    bool set, u32 *stable_pstate)
+{
+	struct amdgpu_ctx *ctx;
+	struct amdgpu_ctx_mgr *mgr;
+	int r;
+
+	if (!fpriv)
+		return -EINVAL;
+
+	mgr = &fpriv->ctx_mgr;
+	mutex_lock(&mgr->lock);
+	ctx = idr_find(&mgr->ctx_handles, id);
+	if (!ctx) {
+		mutex_unlock(&mgr->lock);
+		return -EINVAL;
+	}
+
+	if (set)
+		r = amdgpu_ctx_set_stable_pstate(ctx, *stable_pstate);
+	else
+		r = amdgpu_ctx_get_stable_pstate(ctx, stable_pstate);
+
+	mutex_unlock(&mgr->lock);
+	return r;
+}
+
 int amdgpu_ctx_ioctl(struct drm_device *dev, void *data,
 		     struct drm_file *filp)
 {
 	int r;
-	uint32_t id;
+	uint32_t id, stable_pstate;
 	int32_t priority;
 
 	union drm_amdgpu_ctx *args = data;
@@ -481,25 +676,50 @@ int amdgpu_ctx_ioctl(struct drm_device *dev, void *data,
 	id = args->in.ctx_id;
 	priority = args->in.priority;
 
-	/* For backwards compatibility reasons, we need to accept
-	 * ioctls with garbage in the priority field */
+	/* For backwards compatibility, we need to accept ioctls with garbage
+	 * in the priority field. Garbage values in the priority field, result
+	 * in the priority being set to NORMAL.
+	 */
 	if (!amdgpu_ctx_priority_is_valid(priority))
 		priority = AMDGPU_CTX_PRIORITY_NORMAL;
 
 	switch (args->in.op) {
 	case AMDGPU_CTX_OP_ALLOC_CTX:
+		if (args->in.flags)
+			return -EINVAL;
 		r = amdgpu_ctx_alloc(adev, fpriv, filp, priority, &id);
 		args->out.alloc.ctx_id = id;
 		break;
 	case AMDGPU_CTX_OP_FREE_CTX:
+		if (args->in.flags)
+			return -EINVAL;
 		r = amdgpu_ctx_free(fpriv, id);
 		break;
 	case AMDGPU_CTX_OP_QUERY_STATE:
+		if (args->in.flags)
+			return -EINVAL;
 		r = amdgpu_ctx_query(adev, fpriv, id, &args->out);
 		break;
 	case AMDGPU_CTX_OP_QUERY_STATE2:
+		if (args->in.flags)
+			return -EINVAL;
 		r = amdgpu_ctx_query2(adev, fpriv, id, &args->out);
 		break;
+	case AMDGPU_CTX_OP_GET_STABLE_PSTATE:
+		if (args->in.flags)
+			return -EINVAL;
+		r = amdgpu_ctx_stable_pstate(adev, fpriv, id, false, &stable_pstate);
+		if (!r)
+			args->out.pstate.flags = stable_pstate;
+		break;
+	case AMDGPU_CTX_OP_SET_STABLE_PSTATE:
+		if (args->in.flags & ~AMDGPU_CTX_STABLE_PSTATE_FLAGS_MASK)
+			return -EINVAL;
+		stable_pstate = args->in.flags & AMDGPU_CTX_STABLE_PSTATE_FLAGS_MASK;
+		if (stable_pstate > AMDGPU_CTX_STABLE_PSTATE_PEAK)
+			return -EINVAL;
+		r = amdgpu_ctx_stable_pstate(adev, fpriv, id, true, &stable_pstate);
+		break;
 	default:
 		return -EINVAL;
 	}
@@ -534,9 +754,9 @@ int amdgpu_ctx_put(struct amdgpu_ctx *ctx)
 	return 0;
 }
 
-void amdgpu_ctx_add_fence(struct amdgpu_ctx *ctx,
-			  struct drm_sched_entity *entity,
-			  struct dma_fence *fence, uint64_t *handle)
+uint64_t amdgpu_ctx_add_fence(struct amdgpu_ctx *ctx,
+			      struct drm_sched_entity *entity,
+			      struct dma_fence *fence)
 {
 	struct amdgpu_ctx_entity *centity = to_amdgpu_ctx_entity(entity);
 	uint64_t seq = centity->sequence;
@@ -545,8 +765,7 @@ void amdgpu_ctx_add_fence(struct amdgpu_ctx *ctx,
 
 	idx = seq & (amdgpu_sched_jobs - 1);
 	other = centity->fences[idx];
-	if (other)
-		BUG_ON(!dma_fence_is_signaled(other));
+	WARN_ON(other && !dma_fence_is_signaled(other));
 
 	dma_fence_get(fence);
 
@@ -555,9 +774,11 @@ void amdgpu_ctx_add_fence(struct amdgpu_ctx *ctx,
 	centity->sequence++;
 	spin_unlock(&ctx->ring_lock);
 
+	atomic64_add(ktime_to_ns(amdgpu_ctx_fence_time(other)),
+		     &ctx->mgr->time_spend[centity->hw_ip]);
+
 	dma_fence_put(other);
-	if (handle)
-		*handle = seq;
+	return seq;
 }
 
 struct dma_fence *amdgpu_ctx_get_fence(struct amdgpu_ctx *ctx,
@@ -594,7 +815,7 @@ static void amdgpu_ctx_set_entity_priority(struct amdgpu_ctx *ctx,
 					   int hw_ip,
 					   int32_t priority)
 {
-	struct amdgpu_device *adev = ctx->adev;
+	struct amdgpu_device *adev = ctx->mgr->adev;
 	unsigned int hw_prio;
 	struct drm_gpu_scheduler **scheds = NULL;
 	unsigned num_scheds;
@@ -604,7 +825,7 @@ static void amdgpu_ctx_set_entity_priority(struct amdgpu_ctx *ctx,
 				      amdgpu_ctx_to_drm_sched_prio(priority));
 
 	/* set hw priority */
-	if (hw_ip == AMDGPU_HW_IP_COMPUTE) {
+	if (hw_ip == AMDGPU_HW_IP_COMPUTE || hw_ip == AMDGPU_HW_IP_GFX) {
 		hw_prio = amdgpu_ctx_get_hw_prio(ctx, hw_ip);
 		hw_prio = array_index_nospec(hw_prio, AMDGPU_RING_PRIO_MAX);
 		scheds = adev->gpu_sched[hw_ip][hw_prio].sched;
@@ -659,10 +880,17 @@ int amdgpu_ctx_wait_prev_fence(struct amdgpu_ctx *ctx,
 	return r;
 }
 
-void amdgpu_ctx_mgr_init(struct amdgpu_ctx_mgr *mgr)
+void amdgpu_ctx_mgr_init(struct amdgpu_ctx_mgr *mgr,
+			 struct amdgpu_device *adev)
 {
+	unsigned int i;
+
+	mgr->adev = adev;
 	mutex_init(&mgr->lock);
-	idr_init(&mgr->ctx_handles);
+	idr_init_base(&mgr->ctx_handles, 1);
+
+	for (i = 0; i < AMDGPU_HW_IP_NUM; ++i)
+		atomic64_set(&mgr->time_spend[i], 0);
 }
 
 long amdgpu_ctx_mgr_entity_flush(struct amdgpu_ctx_mgr *mgr, long timeout)
@@ -691,7 +919,7 @@ long amdgpu_ctx_mgr_entity_flush(struct amdgpu_ctx_mgr *mgr, long timeout)
 	return timeout;
 }
 
-void amdgpu_ctx_mgr_entity_fini(struct amdgpu_ctx_mgr *mgr)
+static void amdgpu_ctx_mgr_entity_fini(struct amdgpu_ctx_mgr *mgr)
 {
 	struct amdgpu_ctx *ctx;
 	struct idr *idp;
@@ -716,102 +944,49 @@ void amdgpu_ctx_mgr_entity_fini(struct amdgpu_ctx_mgr *mgr)
 				drm_sched_entity_fini(entity);
 			}
 		}
+		kref_put(&ctx->refcount, amdgpu_ctx_fini);
 	}
 }
 
 void amdgpu_ctx_mgr_fini(struct amdgpu_ctx_mgr *mgr)
 {
-	struct amdgpu_ctx *ctx;
-	struct idr *idp;
-	uint32_t id;
-
 	amdgpu_ctx_mgr_entity_fini(mgr);
-
-	idp = &mgr->ctx_handles;
-
-	idr_for_each_entry(idp, ctx, id) {
-		if (kref_put(&ctx->refcount, amdgpu_ctx_fini) != 1)
-			DRM_ERROR("ctx %p is still alive\n", ctx);
-	}
-
 	idr_destroy(&mgr->ctx_handles);
 	mutex_destroy(&mgr->lock);
 }
 
-static void amdgpu_ctx_fence_time(struct amdgpu_ctx *ctx,
-		struct amdgpu_ctx_entity *centity, ktime_t *total, ktime_t *max)
+void amdgpu_ctx_mgr_usage(struct amdgpu_ctx_mgr *mgr,
+			  ktime_t usage[AMDGPU_HW_IP_NUM])
 {
-	ktime_t now, t1;
-	uint32_t i;
-
-	*total = *max = 0;
-
-	now = ktime_get();
-	for (i = 0; i < amdgpu_sched_jobs; i++) {
-		struct dma_fence *fence;
-		struct drm_sched_fence *s_fence;
-
-		spin_lock(&ctx->ring_lock);
-		fence = dma_fence_get(centity->fences[i]);
-		spin_unlock(&ctx->ring_lock);
-		if (!fence)
-			continue;
-		s_fence = to_drm_sched_fence(fence);
-		if (!dma_fence_is_signaled(&s_fence->scheduled)) {
-			dma_fence_put(fence);
-			continue;
-		}
-		t1 = s_fence->scheduled.timestamp;
-		if (!ktime_before(t1, now)) {
-			dma_fence_put(fence);
-			continue;
-		}
-		if (dma_fence_is_signaled(&s_fence->finished) &&
-			s_fence->finished.timestamp < now)
-			*total += ktime_sub(s_fence->finished.timestamp, t1);
-		else
-			*total += ktime_sub(now, t1);
-		t1 = ktime_sub(now, t1);
-		dma_fence_put(fence);
-		*max = max(t1, *max);
-	}
-}
-
-ktime_t amdgpu_ctx_mgr_fence_usage(struct amdgpu_ctx_mgr *mgr, uint32_t hwip,
-		uint32_t idx, uint64_t *elapsed)
-{
-	struct idr *idp;
 	struct amdgpu_ctx *ctx;
+	unsigned int hw_ip, i;
 	uint32_t id;
-	struct amdgpu_ctx_entity *centity;
-	ktime_t total = 0, max = 0;
 
-	if (idx >= AMDGPU_MAX_ENTITY_NUM)
-		return 0;
-	idp = &mgr->ctx_handles;
+	/*
+	 * This is a little bit racy because it can be that a ctx or a fence are
+	 * destroyed just in the moment we try to account them. But that is ok
+	 * since exactly that case is explicitely allowed by the interface.
+	 */
 	mutex_lock(&mgr->lock);
-	idr_for_each_entry(idp, ctx, id) {
-		ktime_t ttotal, tmax;
+	for (hw_ip = 0; hw_ip < AMDGPU_HW_IP_NUM; ++hw_ip) {
+		uint64_t ns = atomic64_read(&mgr->time_spend[hw_ip]);
 
-		if (!ctx->entities[hwip][idx])
-			continue;
-
-		centity = ctx->entities[hwip][idx];
-		amdgpu_ctx_fence_time(ctx, centity, &ttotal, &tmax);
+		usage[hw_ip] = ns_to_ktime(ns);
+	}
 
-		/* Harmonic mean approximation diverges for very small
-		 * values. If ratio < 0.01% ignore
-		 */
-		if (AMDGPU_CTX_FENCE_USAGE_MIN_RATIO(tmax, ttotal))
-			continue;
+	idr_for_each_entry(&mgr->ctx_handles, ctx, id) {
+		for (hw_ip = 0; hw_ip < AMDGPU_HW_IP_NUM; ++hw_ip) {
+			for (i = 0; i < amdgpu_ctx_num_entities[hw_ip]; ++i) {
+				struct amdgpu_ctx_entity *centity;
+				ktime_t spend;
 
-		total = ktime_add(total, ttotal);
-		max = ktime_after(tmax, max) ? tmax : max;
+				centity = ctx->entities[hw_ip][i];
+				if (!centity)
+					continue;
+				spend = amdgpu_ctx_entity_time(ctx, centity);
+				usage[hw_ip] = ktime_add(usage[hw_ip], spend);
+			}
+		}
 	}
-
 	mutex_unlock(&mgr->lock);
-	if (elapsed)
-		*elapsed = max;
-
-	return total;
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.h
index a44b8b8ed39c..090dfe86f75b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.h
@@ -23,16 +23,20 @@
 #ifndef __AMDGPU_CTX_H__
 #define __AMDGPU_CTX_H__
 
+#include <linux/ktime.h>
+#include <linux/types.h>
+
 #include "amdgpu_ring.h"
 
 struct drm_device;
 struct drm_file;
 struct amdgpu_fpriv;
+struct amdgpu_ctx_mgr;
 
 #define AMDGPU_MAX_ENTITY_NUM 4
-#define AMDGPU_CTX_FENCE_USAGE_MIN_RATIO(max, total) ((max) > 16384ULL*(total))
 
 struct amdgpu_ctx_entity {
+	uint32_t		hw_ip;
 	uint64_t		sequence;
 	struct drm_sched_entity	entity;
 	struct dma_fence	*fences[];
@@ -40,19 +44,20 @@ struct amdgpu_ctx_entity {
 
 struct amdgpu_ctx {
 	struct kref			refcount;
-	struct amdgpu_device		*adev;
+	struct amdgpu_ctx_mgr		*mgr;
 	unsigned			reset_counter;
 	unsigned			reset_counter_query;
-	uint32_t			vram_lost_counter;
+	uint64_t			generation;
 	spinlock_t			ring_lock;
 	struct amdgpu_ctx_entity	*entities[AMDGPU_HW_IP_NUM][AMDGPU_MAX_ENTITY_NUM];
 	bool				preamble_presented;
 	int32_t				init_priority;
 	int32_t				override_priority;
-	struct mutex			lock;
 	atomic_t			guilty;
 	unsigned long			ras_counter_ce;
 	unsigned long			ras_counter_ue;
+	uint32_t			stable_pstate;
+	struct amdgpu_ctx_mgr		*ctx_mgr;
 };
 
 struct amdgpu_ctx_mgr {
@@ -60,6 +65,7 @@ struct amdgpu_ctx_mgr {
 	struct mutex		lock;
 	/* protected by lock */
 	struct idr		ctx_handles;
+	atomic64_t		time_spend[AMDGPU_HW_IP_NUM];
 };
 
 extern const unsigned int amdgpu_ctx_num_entities[AMDGPU_HW_IP_NUM];
@@ -69,9 +75,9 @@ int amdgpu_ctx_put(struct amdgpu_ctx *ctx);
 
 int amdgpu_ctx_get_entity(struct amdgpu_ctx *ctx, u32 hw_ip, u32 instance,
 			  u32 ring, struct drm_sched_entity **entity);
-void amdgpu_ctx_add_fence(struct amdgpu_ctx *ctx,
-			  struct drm_sched_entity *entity,
-			  struct dma_fence *fence, uint64_t *seq);
+uint64_t amdgpu_ctx_add_fence(struct amdgpu_ctx *ctx,
+			      struct drm_sched_entity *entity,
+			      struct dma_fence *fence);
 struct dma_fence *amdgpu_ctx_get_fence(struct amdgpu_ctx *ctx,
 				       struct drm_sched_entity *entity,
 				       uint64_t seq);
@@ -84,10 +90,11 @@ int amdgpu_ctx_ioctl(struct drm_device *dev, void *data,
 int amdgpu_ctx_wait_prev_fence(struct amdgpu_ctx *ctx,
 			       struct drm_sched_entity *entity);
 
-void amdgpu_ctx_mgr_init(struct amdgpu_ctx_mgr *mgr);
-void amdgpu_ctx_mgr_entity_fini(struct amdgpu_ctx_mgr *mgr);
+void amdgpu_ctx_mgr_init(struct amdgpu_ctx_mgr *mgr,
+			 struct amdgpu_device *adev);
 long amdgpu_ctx_mgr_entity_flush(struct amdgpu_ctx_mgr *mgr, long timeout);
 void amdgpu_ctx_mgr_fini(struct amdgpu_ctx_mgr *mgr);
-ktime_t amdgpu_ctx_mgr_fence_usage(struct amdgpu_ctx_mgr *mgr, uint32_t hwip,
-		uint32_t idx, uint64_t *elapsed);
+void amdgpu_ctx_mgr_usage(struct amdgpu_ctx_mgr *mgr,
+			  ktime_t usage[AMDGPU_HW_IP_NUM]);
+
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
index 164d6a9e9fbb..62d43b8cbe58 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
@@ -37,6 +37,9 @@
 #include "amdgpu_fw_attestation.h"
 #include "amdgpu_umr.h"
 
+#include "amdgpu_reset.h"
+#include "amdgpu_psp_ta.h"
+
 #if defined(CONFIG_DEBUG_FS)
 
 /**
@@ -53,14 +56,14 @@
  *
  * Bit 62:  Indicates a GRBM bank switch is needed
  * Bit 61:  Indicates a SRBM bank switch is needed (implies bit 62 is
- * 	    zero)
+ *	    zero)
  * Bits 24..33: The SE or ME selector if needed
  * Bits 34..43: The SH (or SA) or PIPE selector if needed
  * Bits 44..53: The INSTANCE (or CU/WGP) or QUEUE selector if needed
  *
  * Bit 23:  Indicates that the PM power gating lock should be held
- * 	    This is necessary to read registers that might be
- * 	    unreliable during a power gating transistion.
+ *	    This is necessary to read registers that might be
+ *	    unreliable during a power gating transistion.
  *
  * The lower bits are the BYTE offset of the register to read.  This
  * allows reading multiple registers in a single call and having
@@ -73,7 +76,7 @@ static int  amdgpu_debugfs_process_reg_op(bool read, struct file *f,
 	ssize_t result = 0;
 	int r;
 	bool pm_pg_lock, use_bank, use_ring;
-	unsigned instance_bank, sh_bank, se_bank, me, pipe, queue, vmid;
+	unsigned int instance_bank, sh_bank, se_bank, me, pipe, queue, vmid;
 
 	pm_pg_lock = use_bank = use_ring = false;
 	instance_bank = sh_bank = se_bank = me = pipe = queue = vmid = 0;
@@ -126,17 +129,16 @@ static int  amdgpu_debugfs_process_reg_op(bool read, struct file *f,
 	if (use_bank) {
 		if ((sh_bank != 0xFFFFFFFF && sh_bank >= adev->gfx.config.max_sh_per_se) ||
 		    (se_bank != 0xFFFFFFFF && se_bank >= adev->gfx.config.max_shader_engines)) {
-			pm_runtime_mark_last_busy(adev_to_drm(adev)->dev);
 			pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
 			amdgpu_virt_disable_access_debugfs(adev);
 			return -EINVAL;
 		}
 		mutex_lock(&adev->grbm_idx_mutex);
 		amdgpu_gfx_select_se_sh(adev, se_bank,
-					sh_bank, instance_bank);
+					sh_bank, instance_bank, 0);
 	} else if (use_ring) {
 		mutex_lock(&adev->srbm_mutex);
-		amdgpu_gfx_select_me_pipe_q(adev, me, pipe, queue, vmid);
+		amdgpu_gfx_select_me_pipe_q(adev, me, pipe, queue, vmid, 0);
 	}
 
 	if (pm_pg_lock)
@@ -151,7 +153,7 @@ static int  amdgpu_debugfs_process_reg_op(bool read, struct file *f,
 		} else {
 			r = get_user(value, (uint32_t *)buf);
 			if (!r)
-				amdgpu_mm_wreg_mmio_rlc(adev, *pos >> 2, value);
+				amdgpu_mm_wreg_mmio_rlc(adev, *pos >> 2, value, 0);
 		}
 		if (r) {
 			result = r;
@@ -166,17 +168,16 @@ static int  amdgpu_debugfs_process_reg_op(bool read, struct file *f,
 
 end:
 	if (use_bank) {
-		amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+		amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
 		mutex_unlock(&adev->grbm_idx_mutex);
 	} else if (use_ring) {
-		amdgpu_gfx_select_me_pipe_q(adev, 0, 0, 0, 0);
+		amdgpu_gfx_select_me_pipe_q(adev, 0, 0, 0, 0, 0);
 		mutex_unlock(&adev->srbm_mutex);
 	}
 
 	if (pm_pg_lock)
 		mutex_unlock(&adev->pm.mutex);
 
-	pm_runtime_mark_last_busy(adev_to_drm(adev)->dev);
 	pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
 
 	amdgpu_virt_disable_access_debugfs(adev);
@@ -205,7 +206,7 @@ static int amdgpu_debugfs_regs2_open(struct inode *inode, struct file *file)
 {
 	struct amdgpu_debugfs_regs2_data *rd;
 
-	rd = kzalloc(sizeof *rd, GFP_KERNEL);
+	rd = kzalloc(sizeof(*rd), GFP_KERNEL);
 	if (!rd)
 		return -ENOMEM;
 	rd->adev = file_inode(file)->i_private;
@@ -218,6 +219,7 @@ static int amdgpu_debugfs_regs2_open(struct inode *inode, struct file *file)
 static int amdgpu_debugfs_regs2_release(struct inode *inode, struct file *file)
 {
 	struct amdgpu_debugfs_regs2_data *rd = file->private_data;
+
 	mutex_destroy(&rd->lock);
 	kfree(file->private_data);
 	return 0;
@@ -251,7 +253,6 @@ static ssize_t amdgpu_debugfs_regs2_op(struct file *f, char __user *buf, u32 off
 	if (rd->id.use_grbm) {
 		if ((rd->id.grbm.sh != 0xFFFFFFFF && rd->id.grbm.sh >= adev->gfx.config.max_sh_per_se) ||
 		    (rd->id.grbm.se != 0xFFFFFFFF && rd->id.grbm.se >= adev->gfx.config.max_shader_engines)) {
-			pm_runtime_mark_last_busy(adev_to_drm(adev)->dev);
 			pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
 			amdgpu_virt_disable_access_debugfs(adev);
 			mutex_unlock(&rd->lock);
@@ -259,14 +260,14 @@ static ssize_t amdgpu_debugfs_regs2_op(struct file *f, char __user *buf, u32 off
 		}
 		mutex_lock(&adev->grbm_idx_mutex);
 		amdgpu_gfx_select_se_sh(adev, rd->id.grbm.se,
-								rd->id.grbm.sh,
-								rd->id.grbm.instance);
+						  rd->id.grbm.sh,
+						  rd->id.grbm.instance, rd->id.xcc_id);
 	}
 
 	if (rd->id.use_srbm) {
 		mutex_lock(&adev->srbm_mutex);
 		amdgpu_gfx_select_me_pipe_q(adev, rd->id.srbm.me, rd->id.srbm.pipe,
-									rd->id.srbm.queue, rd->id.srbm.vmid);
+					    rd->id.srbm.queue, rd->id.srbm.vmid, rd->id.xcc_id);
 	}
 
 	if (rd->id.pg_lock)
@@ -279,7 +280,7 @@ static ssize_t amdgpu_debugfs_regs2_op(struct file *f, char __user *buf, u32 off
 		} else {
 			r = get_user(value, (uint32_t *)buf);
 			if (!r)
-				amdgpu_mm_wreg_mmio_rlc(adev, offset >> 2, value);
+				amdgpu_mm_wreg_mmio_rlc(adev, offset >> 2, value, rd->id.xcc_id);
 		}
 		if (r) {
 			result = r;
@@ -292,12 +293,12 @@ static ssize_t amdgpu_debugfs_regs2_op(struct file *f, char __user *buf, u32 off
 	}
 end:
 	if (rd->id.use_grbm) {
-		amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+		amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, rd->id.xcc_id);
 		mutex_unlock(&adev->grbm_idx_mutex);
 	}
 
 	if (rd->id.use_srbm) {
-		amdgpu_gfx_select_me_pipe_q(adev, 0, 0, 0, 0);
+		amdgpu_gfx_select_me_pipe_q(adev, 0, 0, 0, 0, rd->id.xcc_id);
 		mutex_unlock(&adev->srbm_mutex);
 	}
 
@@ -306,7 +307,6 @@ end:
 
 	mutex_unlock(&rd->lock);
 
-	pm_runtime_mark_last_busy(adev_to_drm(adev)->dev);
 	pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
 
 	amdgpu_virt_disable_access_debugfs(adev);
@@ -316,18 +316,45 @@ end:
 static long amdgpu_debugfs_regs2_ioctl(struct file *f, unsigned int cmd, unsigned long data)
 {
 	struct amdgpu_debugfs_regs2_data *rd = f->private_data;
+	struct amdgpu_debugfs_regs2_iocdata v1_data;
 	int r;
 
+	mutex_lock(&rd->lock);
+
 	switch (cmd) {
+	case AMDGPU_DEBUGFS_REGS2_IOC_SET_STATE_V2:
+		r = copy_from_user(&rd->id, (struct amdgpu_debugfs_regs2_iocdata_v2 *)data,
+				   sizeof(rd->id));
+		if (r)
+			r = -EINVAL;
+		goto done;
 	case AMDGPU_DEBUGFS_REGS2_IOC_SET_STATE:
-		mutex_lock(&rd->lock);
-		r = copy_from_user(&rd->id, (struct amdgpu_debugfs_regs2_iocdata *)data, sizeof rd->id);
-		mutex_unlock(&rd->lock);
-		return r ? -EINVAL : 0;
+		r = copy_from_user(&v1_data, (struct amdgpu_debugfs_regs2_iocdata *)data,
+				   sizeof(v1_data));
+		if (r) {
+			r = -EINVAL;
+			goto done;
+		}
+		goto v1_copy;
 	default:
-		return -EINVAL;
-	}
-	return 0;
+		r = -EINVAL;
+		goto done;
+	}
+
+v1_copy:
+	rd->id.use_srbm = v1_data.use_srbm;
+	rd->id.use_grbm = v1_data.use_grbm;
+	rd->id.pg_lock = v1_data.pg_lock;
+	rd->id.grbm.se = v1_data.grbm.se;
+	rd->id.grbm.sh = v1_data.grbm.sh;
+	rd->id.grbm.instance = v1_data.grbm.instance;
+	rd->id.srbm.me = v1_data.srbm.me;
+	rd->id.srbm.pipe = v1_data.srbm.pipe;
+	rd->id.srbm.queue = v1_data.srbm.queue;
+	rd->id.xcc_id = 0;
+done:
+	mutex_unlock(&rd->lock);
+	return r;
 }
 
 static ssize_t amdgpu_debugfs_regs2_read(struct file *f, char __user *buf, size_t size, loff_t *pos)
@@ -340,6 +367,136 @@ static ssize_t amdgpu_debugfs_regs2_write(struct file *f, const char __user *buf
 	return amdgpu_debugfs_regs2_op(f, (char __user *)buf, *pos, size, 1);
 }
 
+static int amdgpu_debugfs_gprwave_open(struct inode *inode, struct file *file)
+{
+	struct amdgpu_debugfs_gprwave_data *rd;
+
+	rd = kzalloc(sizeof(*rd), GFP_KERNEL);
+	if (!rd)
+		return -ENOMEM;
+	rd->adev = file_inode(file)->i_private;
+	file->private_data = rd;
+	mutex_init(&rd->lock);
+
+	return 0;
+}
+
+static int amdgpu_debugfs_gprwave_release(struct inode *inode, struct file *file)
+{
+	struct amdgpu_debugfs_gprwave_data *rd = file->private_data;
+
+	mutex_destroy(&rd->lock);
+	kfree(file->private_data);
+	return 0;
+}
+
+static ssize_t amdgpu_debugfs_gprwave_read(struct file *f, char __user *buf, size_t size, loff_t *pos)
+{
+	struct amdgpu_debugfs_gprwave_data *rd = f->private_data;
+	struct amdgpu_device *adev = rd->adev;
+	ssize_t result = 0;
+	int r;
+	uint32_t *data, x;
+
+	if (size > 4096 || size & 0x3 || *pos & 0x3)
+		return -EINVAL;
+
+	r = pm_runtime_get_sync(adev_to_drm(adev)->dev);
+	if (r < 0) {
+		pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
+		return r;
+	}
+
+	r = amdgpu_virt_enable_access_debugfs(adev);
+	if (r < 0) {
+		pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
+		return r;
+	}
+
+	data = kcalloc(1024, sizeof(*data), GFP_KERNEL);
+	if (!data) {
+		pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
+		amdgpu_virt_disable_access_debugfs(adev);
+		return -ENOMEM;
+	}
+
+	/* switch to the specific se/sh/cu */
+	mutex_lock(&adev->grbm_idx_mutex);
+	amdgpu_gfx_select_se_sh(adev, rd->id.se, rd->id.sh, rd->id.cu, rd->id.xcc_id);
+
+	if (!rd->id.gpr_or_wave) {
+		x = 0;
+		if (adev->gfx.funcs->read_wave_data)
+			adev->gfx.funcs->read_wave_data(adev, rd->id.xcc_id, rd->id.simd, rd->id.wave, data, &x);
+	} else {
+		x = size >> 2;
+		if (rd->id.gpr.vpgr_or_sgpr) {
+			if (adev->gfx.funcs->read_wave_vgprs)
+				adev->gfx.funcs->read_wave_vgprs(adev, rd->id.xcc_id, rd->id.simd, rd->id.wave, rd->id.gpr.thread, *pos, size>>2, data);
+		} else {
+			if (adev->gfx.funcs->read_wave_sgprs)
+				adev->gfx.funcs->read_wave_sgprs(adev, rd->id.xcc_id, rd->id.simd, rd->id.wave, *pos, size>>2, data);
+		}
+	}
+
+	amdgpu_gfx_select_se_sh(adev, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, rd->id.xcc_id);
+	mutex_unlock(&adev->grbm_idx_mutex);
+
+	pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
+
+	if (!x) {
+		result = -EINVAL;
+		goto done;
+	}
+
+	while (size && (*pos < x * 4)) {
+		uint32_t value;
+
+		value = data[*pos >> 2];
+		r = put_user(value, (uint32_t *)buf);
+		if (r) {
+			result = r;
+			goto done;
+		}
+
+		result += 4;
+		buf += 4;
+		*pos += 4;
+		size -= 4;
+	}
+
+done:
+	amdgpu_virt_disable_access_debugfs(adev);
+	kfree(data);
+	return result;
+}
+
+static long amdgpu_debugfs_gprwave_ioctl(struct file *f, unsigned int cmd, unsigned long data)
+{
+	struct amdgpu_debugfs_gprwave_data *rd = f->private_data;
+	int r = 0;
+
+	mutex_lock(&rd->lock);
+
+	switch (cmd) {
+	case AMDGPU_DEBUGFS_GPRWAVE_IOC_SET_STATE:
+		if (copy_from_user(&rd->id,
+				   (struct amdgpu_debugfs_gprwave_iocdata *)data,
+				   sizeof(rd->id)))
+			r = -EFAULT;
+		goto done;
+	default:
+		r = -EINVAL;
+		goto done;
+	}
+
+done:
+	mutex_unlock(&rd->lock);
+	return r;
+}
+
+
+
 
 /**
  * amdgpu_debugfs_regs_pcie_read - Read from a PCIE register
@@ -378,14 +535,14 @@ static ssize_t amdgpu_debugfs_regs_pcie_read(struct file *f, char __user *buf,
 	while (size) {
 		uint32_t value;
 
-		value = RREG32_PCIE(*pos);
+		if (upper_32_bits(*pos))
+			value = RREG32_PCIE_EXT(*pos);
+		else
+			value = RREG32_PCIE(*pos);
+
 		r = put_user(value, (uint32_t *)buf);
-		if (r) {
-			pm_runtime_mark_last_busy(adev_to_drm(adev)->dev);
-			pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
-			amdgpu_virt_disable_access_debugfs(adev);
-			return r;
-		}
+		if (r)
+			goto out;
 
 		result += 4;
 		buf += 4;
@@ -393,11 +550,11 @@ static ssize_t amdgpu_debugfs_regs_pcie_read(struct file *f, char __user *buf,
 		size -= 4;
 	}
 
-	pm_runtime_mark_last_busy(adev_to_drm(adev)->dev);
+	r = result;
+out:
 	pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
-
 	amdgpu_virt_disable_access_debugfs(adev);
-	return result;
+	return r;
 }
 
 /**
@@ -438,14 +595,13 @@ static ssize_t amdgpu_debugfs_regs_pcie_write(struct file *f, const char __user
 		uint32_t value;
 
 		r = get_user(value, (uint32_t *)buf);
-		if (r) {
-			pm_runtime_mark_last_busy(adev_to_drm(adev)->dev);
-			pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
-			amdgpu_virt_disable_access_debugfs(adev);
-			return r;
-		}
+		if (r)
+			goto out;
 
-		WREG32_PCIE(*pos, value);
+		if (upper_32_bits(*pos))
+			WREG32_PCIE_EXT(*pos, value);
+		else
+			WREG32_PCIE(*pos, value);
 
 		result += 4;
 		buf += 4;
@@ -453,11 +609,11 @@ static ssize_t amdgpu_debugfs_regs_pcie_write(struct file *f, const char __user
 		size -= 4;
 	}
 
-	pm_runtime_mark_last_busy(adev_to_drm(adev)->dev);
+	r = result;
+out:
 	pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
-
 	amdgpu_virt_disable_access_debugfs(adev);
-	return result;
+	return r;
 }
 
 /**
@@ -482,6 +638,9 @@ static ssize_t amdgpu_debugfs_regs_didt_read(struct file *f, char __user *buf,
 	if (size & 0x3 || *pos & 0x3)
 		return -EINVAL;
 
+	if (!adev->didt_rreg)
+		return -EOPNOTSUPP;
+
 	r = pm_runtime_get_sync(adev_to_drm(adev)->dev);
 	if (r < 0) {
 		pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
@@ -499,12 +658,8 @@ static ssize_t amdgpu_debugfs_regs_didt_read(struct file *f, char __user *buf,
 
 		value = RREG32_DIDT(*pos >> 2);
 		r = put_user(value, (uint32_t *)buf);
-		if (r) {
-			pm_runtime_mark_last_busy(adev_to_drm(adev)->dev);
-			pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
-			amdgpu_virt_disable_access_debugfs(adev);
-			return r;
-		}
+		if (r)
+			goto out;
 
 		result += 4;
 		buf += 4;
@@ -512,11 +667,11 @@ static ssize_t amdgpu_debugfs_regs_didt_read(struct file *f, char __user *buf,
 		size -= 4;
 	}
 
-	pm_runtime_mark_last_busy(adev_to_drm(adev)->dev);
+	r = result;
+out:
 	pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
-
 	amdgpu_virt_disable_access_debugfs(adev);
-	return result;
+	return r;
 }
 
 /**
@@ -541,6 +696,9 @@ static ssize_t amdgpu_debugfs_regs_didt_write(struct file *f, const char __user
 	if (size & 0x3 || *pos & 0x3)
 		return -EINVAL;
 
+	if (!adev->didt_wreg)
+		return -EOPNOTSUPP;
+
 	r = pm_runtime_get_sync(adev_to_drm(adev)->dev);
 	if (r < 0) {
 		pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
@@ -557,12 +715,8 @@ static ssize_t amdgpu_debugfs_regs_didt_write(struct file *f, const char __user
 		uint32_t value;
 
 		r = get_user(value, (uint32_t *)buf);
-		if (r) {
-			pm_runtime_mark_last_busy(adev_to_drm(adev)->dev);
-			pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
-			amdgpu_virt_disable_access_debugfs(adev);
-			return r;
-		}
+		if (r)
+			goto out;
 
 		WREG32_DIDT(*pos >> 2, value);
 
@@ -572,11 +726,11 @@ static ssize_t amdgpu_debugfs_regs_didt_write(struct file *f, const char __user
 		size -= 4;
 	}
 
-	pm_runtime_mark_last_busy(adev_to_drm(adev)->dev);
+	r = result;
+out:
 	pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
-
 	amdgpu_virt_disable_access_debugfs(adev);
-	return result;
+	return r;
 }
 
 /**
@@ -598,6 +752,9 @@ static ssize_t amdgpu_debugfs_regs_smc_read(struct file *f, char __user *buf,
 	ssize_t result = 0;
 	int r;
 
+	if (!adev->smc_rreg)
+		return -EOPNOTSUPP;
+
 	if (size & 0x3 || *pos & 0x3)
 		return -EINVAL;
 
@@ -618,12 +775,8 @@ static ssize_t amdgpu_debugfs_regs_smc_read(struct file *f, char __user *buf,
 
 		value = RREG32_SMC(*pos);
 		r = put_user(value, (uint32_t *)buf);
-		if (r) {
-			pm_runtime_mark_last_busy(adev_to_drm(adev)->dev);
-			pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
-			amdgpu_virt_disable_access_debugfs(adev);
-			return r;
-		}
+		if (r)
+			goto out;
 
 		result += 4;
 		buf += 4;
@@ -631,11 +784,11 @@ static ssize_t amdgpu_debugfs_regs_smc_read(struct file *f, char __user *buf,
 		size -= 4;
 	}
 
-	pm_runtime_mark_last_busy(adev_to_drm(adev)->dev);
+	r = result;
+out:
 	pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
-
 	amdgpu_virt_disable_access_debugfs(adev);
-	return result;
+	return r;
 }
 
 /**
@@ -657,6 +810,9 @@ static ssize_t amdgpu_debugfs_regs_smc_write(struct file *f, const char __user *
 	ssize_t result = 0;
 	int r;
 
+	if (!adev->smc_wreg)
+		return -EOPNOTSUPP;
+
 	if (size & 0x3 || *pos & 0x3)
 		return -EINVAL;
 
@@ -676,12 +832,8 @@ static ssize_t amdgpu_debugfs_regs_smc_write(struct file *f, const char __user *
 		uint32_t value;
 
 		r = get_user(value, (uint32_t *)buf);
-		if (r) {
-			pm_runtime_mark_last_busy(adev_to_drm(adev)->dev);
-			pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
-			amdgpu_virt_disable_access_debugfs(adev);
-			return r;
-		}
+		if (r)
+			goto out;
 
 		WREG32_SMC(*pos, value);
 
@@ -691,11 +843,11 @@ static ssize_t amdgpu_debugfs_regs_smc_write(struct file *f, const char __user *
 		size -= 4;
 	}
 
-	pm_runtime_mark_last_busy(adev_to_drm(adev)->dev);
+	r = result;
+out:
 	pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
-
 	amdgpu_virt_disable_access_debugfs(adev);
-	return result;
+	return r;
 }
 
 /**
@@ -728,7 +880,7 @@ static ssize_t amdgpu_debugfs_gca_config_read(struct file *f, char __user *buf,
 		return -ENOMEM;
 
 	/* version, increment each time something is added */
-	config[no_regs++] = 3;
+	config[no_regs++] = 5;
 	config[no_regs++] = adev->gfx.config.max_shader_engines;
 	config[no_regs++] = adev->gfx.config.max_tile_pipes;
 	config[no_regs++] = adev->gfx.config.max_cu_per_sh;
@@ -756,7 +908,7 @@ static ssize_t amdgpu_debugfs_gca_config_read(struct file *f, char __user *buf,
 	/* rev==1 */
 	config[no_regs++] = adev->rev_id;
 	config[no_regs++] = adev->pg_flags;
-	config[no_regs++] = adev->cg_flags;
+	config[no_regs++] = lower_32_bits(adev->cg_flags);
 
 	/* rev==2 */
 	config[no_regs++] = adev->family;
@@ -768,6 +920,13 @@ static ssize_t amdgpu_debugfs_gca_config_read(struct file *f, char __user *buf,
 	config[no_regs++] = adev->pdev->subsystem_device;
 	config[no_regs++] = adev->pdev->subsystem_vendor;
 
+	/* rev==4 APU flag */
+	config[no_regs++] = adev->flags & AMD_IS_APU ? 1 : 0;
+
+	/* rev==5 PG/CG flag upper 32bit */
+	config[no_regs++] = 0;
+	config[no_regs++] = upper_32_bits(adev->cg_flags);
+
 	while (size && (*pos < no_regs * 4)) {
 		uint32_t value;
 
@@ -833,7 +992,6 @@ static ssize_t amdgpu_debugfs_sensor_read(struct file *f, char __user *buf,
 
 	r = amdgpu_dpm_read_sensor(adev, idx, &values[0], &valuesize);
 
-	pm_runtime_mark_last_busy(adev_to_drm(adev)->dev);
 	pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
 
 	if (r) {
@@ -871,7 +1029,7 @@ static ssize_t amdgpu_debugfs_sensor_read(struct file *f, char __user *buf,
  * The offset being sought changes which wave that the status data
  * will be returned for.  The bits are used as follows:
  *
- * Bits 0..6: 	Byte offset into data
+ * Bits 0..6:	Byte offset into data
  * Bits 7..14:	SE selector
  * Bits 15..22:	SH/SA selector
  * Bits 23..30: CU/{WGP+SIMD} selector
@@ -915,16 +1073,15 @@ static ssize_t amdgpu_debugfs_wave_read(struct file *f, char __user *buf,
 
 	/* switch to the specific se/sh/cu */
 	mutex_lock(&adev->grbm_idx_mutex);
-	amdgpu_gfx_select_se_sh(adev, se, sh, cu);
+	amdgpu_gfx_select_se_sh(adev, se, sh, cu, 0);
 
 	x = 0;
 	if (adev->gfx.funcs->read_wave_data)
-		adev->gfx.funcs->read_wave_data(adev, simd, wave, data, &x);
+		adev->gfx.funcs->read_wave_data(adev, 0, simd, wave, data, &x);
 
-	amdgpu_gfx_select_se_sh(adev, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF);
+	amdgpu_gfx_select_se_sh(adev, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0);
 	mutex_unlock(&adev->grbm_idx_mutex);
 
-	pm_runtime_mark_last_busy(adev_to_drm(adev)->dev);
 	pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
 
 	if (!x) {
@@ -1009,20 +1166,19 @@ static ssize_t amdgpu_debugfs_gpr_read(struct file *f, char __user *buf,
 
 	/* switch to the specific se/sh/cu */
 	mutex_lock(&adev->grbm_idx_mutex);
-	amdgpu_gfx_select_se_sh(adev, se, sh, cu);
+	amdgpu_gfx_select_se_sh(adev, se, sh, cu, 0);
 
 	if (bank == 0) {
 		if (adev->gfx.funcs->read_wave_vgprs)
-			adev->gfx.funcs->read_wave_vgprs(adev, simd, wave, thread, offset, size>>2, data);
+			adev->gfx.funcs->read_wave_vgprs(adev, 0, simd, wave, thread, offset, size>>2, data);
 	} else {
 		if (adev->gfx.funcs->read_wave_sgprs)
-			adev->gfx.funcs->read_wave_sgprs(adev, simd, wave, offset, size>>2, data);
+			adev->gfx.funcs->read_wave_sgprs(adev, 0, simd, wave, offset, size>>2, data);
 	}
 
-	amdgpu_gfx_select_se_sh(adev, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF);
+	amdgpu_gfx_select_se_sh(adev, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0);
 	mutex_unlock(&adev->grbm_idx_mutex);
 
-	pm_runtime_mark_last_busy(adev_to_drm(adev)->dev);
 	pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
 
 	while (size) {
@@ -1051,6 +1207,154 @@ err:
 }
 
 /**
+ * amdgpu_debugfs_gfxoff_residency_read - Read GFXOFF residency
+ *
+ * @f: open file handle
+ * @buf: User buffer to store read data in
+ * @size: Number of bytes to read
+ * @pos:  Offset to seek to
+ *
+ * Read the last residency value logged. It doesn't auto update, one needs to
+ * stop logging before getting the current value.
+ */
+static ssize_t amdgpu_debugfs_gfxoff_residency_read(struct file *f, char __user *buf,
+						    size_t size, loff_t *pos)
+{
+	struct amdgpu_device *adev = file_inode(f)->i_private;
+	ssize_t result = 0;
+	int r;
+
+	if (size & 0x3 || *pos & 0x3)
+		return -EINVAL;
+
+	r = pm_runtime_get_sync(adev_to_drm(adev)->dev);
+	if (r < 0) {
+		pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
+		return r;
+	}
+
+	while (size) {
+		uint32_t value;
+
+		r = amdgpu_get_gfx_off_residency(adev, &value);
+		if (r)
+			goto out;
+
+		r = put_user(value, (uint32_t *)buf);
+		if (r)
+			goto out;
+
+		result += 4;
+		buf += 4;
+		*pos += 4;
+		size -= 4;
+	}
+
+	r = result;
+out:
+	pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
+
+	return r;
+}
+
+/**
+ * amdgpu_debugfs_gfxoff_residency_write - Log GFXOFF Residency
+ *
+ * @f: open file handle
+ * @buf: User buffer to write data from
+ * @size: Number of bytes to write
+ * @pos:  Offset to seek to
+ *
+ * Write a 32-bit non-zero to start logging; write a 32-bit zero to stop
+ */
+static ssize_t amdgpu_debugfs_gfxoff_residency_write(struct file *f, const char __user *buf,
+						     size_t size, loff_t *pos)
+{
+	struct amdgpu_device *adev = file_inode(f)->i_private;
+	ssize_t result = 0;
+	int r;
+
+	if (size & 0x3 || *pos & 0x3)
+		return -EINVAL;
+
+	r = pm_runtime_get_sync(adev_to_drm(adev)->dev);
+	if (r < 0) {
+		pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
+		return r;
+	}
+
+	while (size) {
+		u32 value;
+
+		r = get_user(value, (uint32_t *)buf);
+		if (r)
+			goto out;
+
+		amdgpu_set_gfx_off_residency(adev, value ? true : false);
+
+		result += 4;
+		buf += 4;
+		*pos += 4;
+		size -= 4;
+	}
+
+	r = result;
+out:
+	pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
+
+	return r;
+}
+
+
+/**
+ * amdgpu_debugfs_gfxoff_count_read - Read GFXOFF entry count
+ *
+ * @f: open file handle
+ * @buf: User buffer to store read data in
+ * @size: Number of bytes to read
+ * @pos:  Offset to seek to
+ */
+static ssize_t amdgpu_debugfs_gfxoff_count_read(struct file *f, char __user *buf,
+						size_t size, loff_t *pos)
+{
+	struct amdgpu_device *adev = file_inode(f)->i_private;
+	ssize_t result = 0;
+	int r;
+
+	if (size & 0x3 || *pos & 0x3)
+		return -EINVAL;
+
+	r = pm_runtime_get_sync(adev_to_drm(adev)->dev);
+	if (r < 0) {
+		pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
+		return r;
+	}
+
+	while (size) {
+		u64 value = 0;
+
+		r = amdgpu_get_gfx_off_entrycount(adev, &value);
+		if (r)
+			goto out;
+
+		r = put_user(value, (u64 *)buf);
+		if (r)
+			goto out;
+
+		result += 4;
+		buf += 4;
+		*pos += 4;
+		size -= 4;
+	}
+
+	r = result;
+out:
+	pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
+
+	return r;
+}
+
+/**
  * amdgpu_debugfs_gfxoff_write - Enable/disable GFXOFF
  *
  * @f: open file handle
@@ -1080,11 +1384,8 @@ static ssize_t amdgpu_debugfs_gfxoff_write(struct file *f, const char __user *bu
 		uint32_t value;
 
 		r = get_user(value, (uint32_t *)buf);
-		if (r) {
-			pm_runtime_mark_last_busy(adev_to_drm(adev)->dev);
-			pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
-			return r;
-		}
+		if (r)
+			goto out;
 
 		amdgpu_gfx_off_ctrl(adev, value ? true : false);
 
@@ -1094,10 +1395,11 @@ static ssize_t amdgpu_debugfs_gfxoff_write(struct file *f, const char __user *bu
 		size -= 4;
 	}
 
-	pm_runtime_mark_last_busy(adev_to_drm(adev)->dev);
+	r = result;
+out:
 	pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
 
-	return result;
+	return r;
 }
 
 
@@ -1120,25 +1422,57 @@ static ssize_t amdgpu_debugfs_gfxoff_read(struct file *f, char __user *buf,
 		return -EINVAL;
 
 	r = pm_runtime_get_sync(adev_to_drm(adev)->dev);
-	if (r < 0)
+	if (r < 0) {
+		pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
 		return r;
+	}
 
 	while (size) {
-		uint32_t value;
+		u32 value = adev->gfx.gfx_off_state;
+
+		r = put_user(value, (u32 *)buf);
+		if (r)
+			goto out;
+
+		result += 4;
+		buf += 4;
+		*pos += 4;
+		size -= 4;
+	}
+
+	r = result;
+out:
+	pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
+
+	return r;
+}
+
+static ssize_t amdgpu_debugfs_gfxoff_status_read(struct file *f, char __user *buf,
+						 size_t size, loff_t *pos)
+{
+	struct amdgpu_device *adev = file_inode(f)->i_private;
+	ssize_t result = 0;
+	int r;
+
+	if (size & 0x3 || *pos & 0x3)
+		return -EINVAL;
+
+	r = pm_runtime_get_sync(adev_to_drm(adev)->dev);
+	if (r < 0) {
+		pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
+		return r;
+	}
+
+	while (size) {
+		u32 value;
 
 		r = amdgpu_get_gfx_off_status(adev, &value);
-		if (r) {
-			pm_runtime_mark_last_busy(adev_to_drm(adev)->dev);
-			pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
-			return r;
-		}
+		if (r)
+			goto out;
 
-		r = put_user(value, (uint32_t *)buf);
-		if (r) {
-			pm_runtime_mark_last_busy(adev_to_drm(adev)->dev);
-			pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
-			return r;
-		}
+		r = put_user(value, (u32 *)buf);
+		if (r)
+			goto out;
 
 		result += 4;
 		buf += 4;
@@ -1146,10 +1480,11 @@ static ssize_t amdgpu_debugfs_gfxoff_read(struct file *f, char __user *buf,
 		size -= 4;
 	}
 
-	pm_runtime_mark_last_busy(adev_to_drm(adev)->dev);
+	r = result;
+out:
 	pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
 
-	return result;
+	return r;
 }
 
 static const struct file_operations amdgpu_debugfs_regs2_fops = {
@@ -1162,6 +1497,15 @@ static const struct file_operations amdgpu_debugfs_regs2_fops = {
 	.llseek = default_llseek
 };
 
+static const struct file_operations amdgpu_debugfs_gprwave_fops = {
+	.owner = THIS_MODULE,
+	.unlocked_ioctl = amdgpu_debugfs_gprwave_ioctl,
+	.read = amdgpu_debugfs_gprwave_read,
+	.open = amdgpu_debugfs_gprwave_open,
+	.release = amdgpu_debugfs_gprwave_release,
+	.llseek = default_llseek
+};
+
 static const struct file_operations amdgpu_debugfs_regs_fops = {
 	.owner = THIS_MODULE,
 	.read = amdgpu_debugfs_regs_read,
@@ -1217,9 +1561,29 @@ static const struct file_operations amdgpu_debugfs_gfxoff_fops = {
 	.llseek = default_llseek
 };
 
+static const struct file_operations amdgpu_debugfs_gfxoff_status_fops = {
+	.owner = THIS_MODULE,
+	.read = amdgpu_debugfs_gfxoff_status_read,
+	.llseek = default_llseek
+};
+
+static const struct file_operations amdgpu_debugfs_gfxoff_count_fops = {
+	.owner = THIS_MODULE,
+	.read = amdgpu_debugfs_gfxoff_count_read,
+	.llseek = default_llseek
+};
+
+static const struct file_operations amdgpu_debugfs_gfxoff_residency_fops = {
+	.owner = THIS_MODULE,
+	.read = amdgpu_debugfs_gfxoff_residency_read,
+	.write = amdgpu_debugfs_gfxoff_residency_write,
+	.llseek = default_llseek
+};
+
 static const struct file_operations *debugfs_regs[] = {
 	&amdgpu_debugfs_regs_fops,
 	&amdgpu_debugfs_regs2_fops,
+	&amdgpu_debugfs_gprwave_fops,
 	&amdgpu_debugfs_regs_didt_fops,
 	&amdgpu_debugfs_regs_pcie_fops,
 	&amdgpu_debugfs_regs_smc_fops,
@@ -1228,11 +1592,15 @@ static const struct file_operations *debugfs_regs[] = {
 	&amdgpu_debugfs_wave_fops,
 	&amdgpu_debugfs_gpr_fops,
 	&amdgpu_debugfs_gfxoff_fops,
+	&amdgpu_debugfs_gfxoff_status_fops,
+	&amdgpu_debugfs_gfxoff_count_fops,
+	&amdgpu_debugfs_gfxoff_residency_fops,
 };
 
-static const char *debugfs_regs_names[] = {
+static const char * const debugfs_regs_names[] = {
 	"amdgpu_regs",
 	"amdgpu_regs2",
+	"amdgpu_gprwave",
 	"amdgpu_regs_didt",
 	"amdgpu_regs_pcie",
 	"amdgpu_regs_smc",
@@ -1241,11 +1609,14 @@ static const char *debugfs_regs_names[] = {
 	"amdgpu_wave",
 	"amdgpu_gpr",
 	"amdgpu_gfxoff",
+	"amdgpu_gfxoff_status",
+	"amdgpu_gfxoff_count",
+	"amdgpu_gfxoff_residency",
 };
 
 /**
  * amdgpu_debugfs_regs_init -	Initialize debugfs entries that provide
- * 				register access.
+ *				register access.
  *
  * @adev: The device to attach the debugfs entries to
  */
@@ -1257,7 +1628,7 @@ int amdgpu_debugfs_regs_init(struct amdgpu_device *adev)
 
 	for (i = 0; i < ARRAY_SIZE(debugfs_regs); i++) {
 		ent = debugfs_create_file(debugfs_regs_names[i],
-					  S_IFREG | S_IRUGO, root,
+					  S_IFREG | 0400, root,
 					  adev, debugfs_regs[i]);
 		if (!i && !IS_ERR_OR_NULL(ent))
 			i_size_write(ent->d_inode, adev->rmmio_size);
@@ -1268,7 +1639,7 @@ int amdgpu_debugfs_regs_init(struct amdgpu_device *adev)
 
 static int amdgpu_debugfs_test_ib_show(struct seq_file *m, void *unused)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)m->private;
+	struct amdgpu_device *adev = m->private;
 	struct drm_device *dev = adev_to_drm(adev);
 	int r = 0, i;
 
@@ -1279,7 +1650,7 @@ static int amdgpu_debugfs_test_ib_show(struct seq_file *m, void *unused)
 	}
 
 	/* Avoid accidently unparking the sched thread during GPU reset */
-	r = down_write_killable(&adev->reset_sem);
+	r = down_write_killable(&adev->reset_domain->sem);
 	if (r)
 		return r;
 
@@ -1287,30 +1658,29 @@ static int amdgpu_debugfs_test_ib_show(struct seq_file *m, void *unused)
 	for (i = 0; i < AMDGPU_MAX_RINGS; i++) {
 		struct amdgpu_ring *ring = adev->rings[i];
 
-		if (!ring || !ring->sched.thread)
+		if (!amdgpu_ring_sched_ready(ring))
 			continue;
-		kthread_park(ring->sched.thread);
+		drm_sched_wqueue_stop(&ring->sched);
 	}
 
-	seq_printf(m, "run ib test:\n");
+	seq_puts(m, "run ib test:\n");
 	r = amdgpu_ib_ring_tests(adev);
 	if (r)
 		seq_printf(m, "ib ring tests failed (%d).\n", r);
 	else
-		seq_printf(m, "ib ring tests passed.\n");
+		seq_puts(m, "ib ring tests passed.\n");
 
 	/* go on the scheduler */
 	for (i = 0; i < AMDGPU_MAX_RINGS; i++) {
 		struct amdgpu_ring *ring = adev->rings[i];
 
-		if (!ring || !ring->sched.thread)
+		if (!amdgpu_ring_sched_ready(ring))
 			continue;
-		kthread_unpark(ring->sched.thread);
+		drm_sched_wqueue_start(&ring->sched);
 	}
 
-	up_write(&adev->reset_sem);
+	up_write(&adev->reset_domain->sem);
 
-	pm_runtime_mark_last_busy(dev->dev);
 	pm_runtime_put_autosuspend(dev->dev);
 
 	return 0;
@@ -1330,7 +1700,6 @@ static int amdgpu_debugfs_evict_vram(void *data, u64 *val)
 
 	*val = amdgpu_ttm_evict_resources(adev, TTM_PL_VRAM);
 
-	pm_runtime_mark_last_busy(dev->dev);
 	pm_runtime_put_autosuspend(dev->dev);
 
 	return 0;
@@ -1351,16 +1720,33 @@ static int amdgpu_debugfs_evict_gtt(void *data, u64 *val)
 
 	*val = amdgpu_ttm_evict_resources(adev, TTM_PL_TT);
 
-	pm_runtime_mark_last_busy(dev->dev);
 	pm_runtime_put_autosuspend(dev->dev);
 
 	return 0;
 }
 
+static int amdgpu_debugfs_benchmark(void *data, u64 val)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)data;
+	struct drm_device *dev = adev_to_drm(adev);
+	int r;
+
+	r = pm_runtime_get_sync(dev->dev);
+	if (r < 0) {
+		pm_runtime_put_autosuspend(dev->dev);
+		return r;
+	}
+
+	r = amdgpu_benchmark(adev, val);
+
+	pm_runtime_put_autosuspend(dev->dev);
+
+	return r;
+}
 
 static int amdgpu_debugfs_vm_info_show(struct seq_file *m, void *unused)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)m->private;
+	struct amdgpu_device *adev = m->private;
 	struct drm_device *dev = adev_to_drm(adev);
 	struct drm_file *file;
 	int r;
@@ -1372,9 +1758,14 @@ static int amdgpu_debugfs_vm_info_show(struct seq_file *m, void *unused)
 	list_for_each_entry(file, &dev->filelist, lhead) {
 		struct amdgpu_fpriv *fpriv = file->driver_priv;
 		struct amdgpu_vm *vm = &fpriv->vm;
+		struct amdgpu_task_info *ti;
+
+		ti = amdgpu_vm_get_task_info_vm(vm);
+		if (ti) {
+			seq_printf(m, "pid:%d\tProcess:%s ----------\n", ti->task.pid, ti->process_name);
+			amdgpu_vm_put_task_info(ti);
+		}
 
-		seq_printf(m, "pid:%d\tProcess:%s ----------\n",
-				vm->task_info.pid, vm->task_info.process_name);
 		r = amdgpu_bo_reserve(vm->root.bo, true);
 		if (r)
 			break;
@@ -1393,6 +1784,8 @@ DEFINE_DEBUGFS_ATTRIBUTE(amdgpu_evict_vram_fops, amdgpu_debugfs_evict_vram,
 			 NULL, "%lld\n");
 DEFINE_DEBUGFS_ATTRIBUTE(amdgpu_evict_gtt_fops, amdgpu_debugfs_evict_gtt,
 			 NULL, "%lld\n");
+DEFINE_DEBUGFS_ATTRIBUTE(amdgpu_benchmark_fops, NULL, amdgpu_debugfs_benchmark,
+			 "%lld\n");
 
 static void amdgpu_ib_preempt_fences_swap(struct amdgpu_ring *ring,
 					  struct dma_fence **fences)
@@ -1485,7 +1878,7 @@ no_preempt:
 			continue;
 		}
 		job = to_amdgpu_job(s_job);
-		if (preempted && (&job->hw_fence) == fence)
+		if (preempted && (&job->hw_fence->base) == fence)
 			/* mark the job as preempted */
 			job->preemption_status |= AMDGPU_IB_PREEMPTED;
 	}
@@ -1494,7 +1887,7 @@ no_preempt:
 
 static int amdgpu_debugfs_ib_preempt(void *data, u64 val)
 {
-	int r, resched, length;
+	int r, length;
 	struct amdgpu_ring *ring;
 	struct dma_fence **fences = NULL;
 	struct amdgpu_device *adev = (struct amdgpu_device *)data;
@@ -1504,7 +1897,8 @@ static int amdgpu_debugfs_ib_preempt(void *data, u64 val)
 
 	ring = adev->rings[val];
 
-	if (!ring || !ring->funcs->preempt_ib || !ring->sched.thread)
+	if (!amdgpu_ring_sched_ready(ring) ||
+	    !ring->funcs->preempt_ib)
 		return -EINVAL;
 
 	/* the last preemption failed */
@@ -1517,14 +1911,12 @@ static int amdgpu_debugfs_ib_preempt(void *data, u64 val)
 		return -ENOMEM;
 
 	/* Avoid accidently unparking the sched thread during GPU reset */
-	r = down_read_killable(&adev->reset_sem);
+	r = down_read_killable(&adev->reset_domain->sem);
 	if (r)
 		goto pro_end;
 
 	/* stop the scheduler */
-	kthread_park(ring->sched.thread);
-
-	resched = ttm_bo_lock_delayed_workqueue(&adev->mman.bdev);
+	drm_sched_wqueue_stop(&ring->sched);
 
 	/* preempt the IB */
 	r = amdgpu_ring_preempt_ib(ring);
@@ -1558,11 +1950,9 @@ static int amdgpu_debugfs_ib_preempt(void *data, u64 val)
 
 failure:
 	/* restart the scheduler */
-	kthread_unpark(ring->sched.thread);
+	drm_sched_wqueue_start(&ring->sched);
 
-	up_read(&adev->reset_sem);
-
-	ttm_bo_unlock_delayed_workqueue(&adev->mman.bdev, resched);
+	up_read(&adev->reset_domain->sem);
 
 pro_end:
 	kfree(fences);
@@ -1576,7 +1966,7 @@ static int amdgpu_debugfs_sclk_set(void *data, u64 val)
 	uint32_t max_freq, min_freq;
 	struct amdgpu_device *adev = (struct amdgpu_device *)data;
 
-	if (amdgpu_sriov_vf(adev) && !amdgpu_sriov_is_pp_one_vf(adev))
+	if (amdgpu_sriov_multi_vf_mode(adev))
 		return -EINVAL;
 
 	ret = pm_runtime_get_sync(adev_to_drm(adev)->dev);
@@ -1585,22 +1975,24 @@ static int amdgpu_debugfs_sclk_set(void *data, u64 val)
 		return ret;
 	}
 
-	if (is_support_sw_smu(adev)) {
-		ret = smu_get_dpm_freq_range(&adev->smu, SMU_SCLK, &min_freq, &max_freq);
-		if (ret || val > max_freq || val < min_freq)
-			return -EINVAL;
-		ret = smu_set_soft_freq_range(&adev->smu, SMU_SCLK, (uint32_t)val, (uint32_t)val);
-	} else {
-		return 0;
+	ret = amdgpu_dpm_get_dpm_freq_range(adev, PP_SCLK, &min_freq, &max_freq);
+	if (ret == -EOPNOTSUPP) {
+		ret = 0;
+		goto out;
+	}
+	if (ret || val > max_freq || val < min_freq) {
+		ret = -EINVAL;
+		goto out;
 	}
 
-	pm_runtime_mark_last_busy(adev_to_drm(adev)->dev);
-	pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
-
+	ret = amdgpu_dpm_set_soft_freq_range(adev, PP_SCLK, (uint32_t)val, (uint32_t)val);
 	if (ret)
-		return -EINVAL;
+		ret = -EINVAL;
 
-	return 0;
+out:
+	pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
+
+	return ret;
 }
 
 DEFINE_DEBUGFS_ATTRIBUTE(fops_ib_preempt, NULL,
@@ -1618,6 +2010,9 @@ int amdgpu_debugfs_init(struct amdgpu_device *adev)
 	if (!debugfs_initialized())
 		return 0;
 
+	debugfs_create_x32("amdgpu_smu_debug", 0600, root,
+			   &adev->pm.smu_debug_mask);
+
 	ent = debugfs_create_file("amdgpu_preempt_ib", 0600, root, adev,
 				  &fops_ib_preempt);
 	if (IS_ERR(ent)) {
@@ -1644,9 +2039,12 @@ int amdgpu_debugfs_init(struct amdgpu_device *adev)
 		DRM_ERROR("registering register debugfs failed (%d).\n", r);
 
 	amdgpu_debugfs_firmware_init(adev);
+	amdgpu_ta_if_debugfs_init(adev);
+
+	amdgpu_debugfs_mes_event_log_init(adev);
 
 #if defined(CONFIG_DRM_AMD_DC)
-	if (amdgpu_device_has_dc_support(adev))
+	if (adev->dc_enabled)
 		dtn_debugfs_init(adev);
 #endif
 
@@ -1659,33 +2057,109 @@ int amdgpu_debugfs_init(struct amdgpu_device *adev)
 		amdgpu_debugfs_ring_init(adev, ring);
 	}
 
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+		if (!amdgpu_vcnfw_log)
+			break;
+
+		if (adev->vcn.harvest_config & (1 << i))
+			continue;
+
+		amdgpu_debugfs_vcn_fwlog_init(adev, i, &adev->vcn.inst[i]);
+	}
+
+	if (amdgpu_umsch_mm & amdgpu_umsch_mm_fwlog)
+		amdgpu_debugfs_umsch_fwlog_init(adev, &adev->umsch_mm);
+
+	amdgpu_debugfs_vcn_sched_mask_init(adev);
+	amdgpu_debugfs_jpeg_sched_mask_init(adev);
+	amdgpu_debugfs_gfx_sched_mask_init(adev);
+	amdgpu_debugfs_compute_sched_mask_init(adev);
+	amdgpu_debugfs_sdma_sched_mask_init(adev);
+
 	amdgpu_ras_debugfs_create_all(adev);
 	amdgpu_rap_debugfs_init(adev);
 	amdgpu_securedisplay_debugfs_init(adev);
 	amdgpu_fw_attestation_debugfs_init(adev);
+	amdgpu_psp_debugfs_init(adev);
 
-	debugfs_create_file("amdgpu_evict_vram", 0444, root, adev,
+	debugfs_create_file("amdgpu_evict_vram", 0400, root, adev,
 			    &amdgpu_evict_vram_fops);
-	debugfs_create_file("amdgpu_evict_gtt", 0444, root, adev,
+	debugfs_create_file("amdgpu_evict_gtt", 0400, root, adev,
 			    &amdgpu_evict_gtt_fops);
-	debugfs_create_file("amdgpu_test_ib", 0444, root, adev,
+	debugfs_create_file("amdgpu_test_ib", 0400, root, adev,
 			    &amdgpu_debugfs_test_ib_fops);
 	debugfs_create_file("amdgpu_vm_info", 0444, root, adev,
 			    &amdgpu_debugfs_vm_info_fops);
+	debugfs_create_file("amdgpu_benchmark", 0200, root, adev,
+			    &amdgpu_benchmark_fops);
 
 	adev->debugfs_vbios_blob.data = adev->bios;
 	adev->debugfs_vbios_blob.size = adev->bios_size;
 	debugfs_create_blob("amdgpu_vbios", 0444, root,
 			    &adev->debugfs_vbios_blob);
 
-	adev->debugfs_discovery_blob.data = adev->mman.discovery_bin;
-	adev->debugfs_discovery_blob.size = adev->mman.discovery_tmr_size;
-	debugfs_create_blob("amdgpu_discovery", 0444, root,
-			    &adev->debugfs_discovery_blob);
+	if (adev->discovery.debugfs_blob.size)
+		debugfs_create_blob("amdgpu_discovery", 0444, root,
+				    &adev->discovery.debugfs_blob);
+
+	return 0;
+}
+
+static int amdgpu_pt_info_read(struct seq_file *m, void *unused)
+{
+	struct drm_file *file;
+	struct amdgpu_fpriv *fpriv;
+	struct amdgpu_bo *root_bo;
+	struct amdgpu_device *adev;
+	int r;
+
+	file = m->private;
+	if (!file)
+		return -EINVAL;
+
+	adev = drm_to_adev(file->minor->dev);
+	fpriv = file->driver_priv;
+	if (!fpriv || !fpriv->vm.root.bo)
+		return -ENODEV;
+
+	root_bo = amdgpu_bo_ref(fpriv->vm.root.bo);
+	r = amdgpu_bo_reserve(root_bo, true);
+	if (r) {
+		amdgpu_bo_unref(&root_bo);
+		return -EINVAL;
+	}
+
+	seq_printf(m, "pd_address: 0x%llx\n", amdgpu_gmc_pd_addr(fpriv->vm.root.bo));
+	seq_printf(m, "max_pfn: 0x%llx\n", adev->vm_manager.max_pfn);
+	seq_printf(m, "num_level: 0x%x\n", adev->vm_manager.num_level);
+	seq_printf(m, "block_size: 0x%x\n", adev->vm_manager.block_size);
+	seq_printf(m, "fragment_size: 0x%x\n", adev->vm_manager.fragment_size);
+
+	amdgpu_bo_unreserve(root_bo);
+	amdgpu_bo_unref(&root_bo);
 
 	return 0;
 }
 
+static int amdgpu_pt_info_open(struct inode *inode, struct file *file)
+{
+	return single_open(file, amdgpu_pt_info_read, inode->i_private);
+}
+
+static const struct file_operations amdgpu_pt_info_fops = {
+	.owner = THIS_MODULE,
+	.open = amdgpu_pt_info_open,
+	.read = seq_read,
+	.llseek = seq_lseek,
+	.release = single_release,
+};
+
+void amdgpu_debugfs_vm_init(struct drm_file *file)
+{
+	debugfs_create_file("vm_pagetable_info", 0444, file->debugfs_client, file,
+			    &amdgpu_pt_info_fops);
+}
+
 #else
 int amdgpu_debugfs_init(struct amdgpu_device *adev)
 {
@@ -1695,4 +2169,7 @@ int amdgpu_debugfs_regs_init(struct amdgpu_device *adev)
 {
 	return 0;
 }
+void amdgpu_debugfs_vm_init(struct drm_file *file)
+{
+}
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.h
index 371a6f0deb29..e7b3c38e5186 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.h
@@ -32,3 +32,6 @@ void amdgpu_debugfs_fini(struct amdgpu_device *adev);
 void amdgpu_debugfs_fence_init(struct amdgpu_device *adev);
 void amdgpu_debugfs_firmware_init(struct amdgpu_device *adev);
 void amdgpu_debugfs_gem_init(struct amdgpu_device *adev);
+void amdgpu_debugfs_mes_event_log_init(struct amdgpu_device *adev);
+void amdgpu_debugfs_vm_init(struct drm_file *file);
+
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_dev_coredump.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_dev_coredump.c
new file mode 100644
index 000000000000..4e2fe6674db8
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_dev_coredump.c
@@ -0,0 +1,371 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright 2024 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include <generated/utsrelease.h>
+#include <linux/devcoredump.h>
+#include "amdgpu_dev_coredump.h"
+#include "atom.h"
+
+#ifndef CONFIG_DEV_COREDUMP
+void amdgpu_coredump(struct amdgpu_device *adev, bool skip_vram_check,
+		     bool vram_lost, struct amdgpu_job *job)
+{
+}
+#else
+
+const char *hw_ip_names[MAX_HWIP] = {
+	[GC_HWIP]		= "GC",
+	[HDP_HWIP]		= "HDP",
+	[SDMA0_HWIP]		= "SDMA0",
+	[SDMA1_HWIP]		= "SDMA1",
+	[SDMA2_HWIP]		= "SDMA2",
+	[SDMA3_HWIP]		= "SDMA3",
+	[SDMA4_HWIP]		= "SDMA4",
+	[SDMA5_HWIP]		= "SDMA5",
+	[SDMA6_HWIP]		= "SDMA6",
+	[SDMA7_HWIP]		= "SDMA7",
+	[LSDMA_HWIP]		= "LSDMA",
+	[MMHUB_HWIP]		= "MMHUB",
+	[ATHUB_HWIP]		= "ATHUB",
+	[NBIO_HWIP]		= "NBIO",
+	[MP0_HWIP]		= "MP0",
+	[MP1_HWIP]		= "MP1",
+	[UVD_HWIP]		= "UVD/JPEG/VCN",
+	[VCN1_HWIP]		= "VCN1",
+	[VCE_HWIP]		= "VCE",
+	[VPE_HWIP]		= "VPE",
+	[DF_HWIP]		= "DF",
+	[DCE_HWIP]		= "DCE",
+	[OSSSYS_HWIP]		= "OSSSYS",
+	[SMUIO_HWIP]		= "SMUIO",
+	[PWR_HWIP]		= "PWR",
+	[NBIF_HWIP]		= "NBIF",
+	[THM_HWIP]		= "THM",
+	[CLK_HWIP]		= "CLK",
+	[UMC_HWIP]		= "UMC",
+	[RSMU_HWIP]		= "RSMU",
+	[XGMI_HWIP]		= "XGMI",
+	[DCI_HWIP]		= "DCI",
+	[PCIE_HWIP]		= "PCIE",
+};
+
+static void amdgpu_devcoredump_fw_info(struct amdgpu_device *adev,
+				       struct drm_printer *p)
+{
+	uint32_t version;
+	uint32_t feature;
+	uint8_t smu_program, smu_major, smu_minor, smu_debug;
+	struct atom_context *ctx = adev->mode_info.atom_context;
+
+	drm_printf(p, "VCE feature version: %u, fw version: 0x%08x\n",
+		   adev->vce.fb_version, adev->vce.fw_version);
+	drm_printf(p, "UVD feature version: %u, fw version: 0x%08x\n", 0,
+		   adev->uvd.fw_version);
+	drm_printf(p, "GMC feature version: %u, fw version: 0x%08x\n", 0,
+		   adev->gmc.fw_version);
+	drm_printf(p, "ME feature version: %u, fw version: 0x%08x\n",
+		   adev->gfx.me_feature_version, adev->gfx.me_fw_version);
+	drm_printf(p, "PFP feature version: %u, fw version: 0x%08x\n",
+		   adev->gfx.pfp_feature_version, adev->gfx.pfp_fw_version);
+	drm_printf(p, "CE feature version: %u, fw version: 0x%08x\n",
+		   adev->gfx.ce_feature_version, adev->gfx.ce_fw_version);
+	drm_printf(p, "RLC feature version: %u, fw version: 0x%08x\n",
+		   adev->gfx.rlc_feature_version, adev->gfx.rlc_fw_version);
+
+	drm_printf(p, "RLC SRLC feature version: %u, fw version: 0x%08x\n",
+		   adev->gfx.rlc_srlc_feature_version,
+		   adev->gfx.rlc_srlc_fw_version);
+	drm_printf(p, "RLC SRLG feature version: %u, fw version: 0x%08x\n",
+		   adev->gfx.rlc_srlg_feature_version,
+		   adev->gfx.rlc_srlg_fw_version);
+	drm_printf(p, "RLC SRLS feature version: %u, fw version: 0x%08x\n",
+		   adev->gfx.rlc_srls_feature_version,
+		   adev->gfx.rlc_srls_fw_version);
+	drm_printf(p, "RLCP feature version: %u, fw version: 0x%08x\n",
+		   adev->gfx.rlcp_ucode_feature_version,
+		   adev->gfx.rlcp_ucode_version);
+	drm_printf(p, "RLCV feature version: %u, fw version: 0x%08x\n",
+		   adev->gfx.rlcv_ucode_feature_version,
+		   adev->gfx.rlcv_ucode_version);
+	drm_printf(p, "MEC feature version: %u, fw version: 0x%08x\n",
+		   adev->gfx.mec_feature_version, adev->gfx.mec_fw_version);
+
+	if (adev->gfx.mec2_fw)
+		drm_printf(p, "MEC2 feature version: %u, fw version: 0x%08x\n",
+			   adev->gfx.mec2_feature_version,
+			   adev->gfx.mec2_fw_version);
+
+	drm_printf(p, "IMU feature version: %u, fw version: 0x%08x\n", 0,
+		   adev->gfx.imu_fw_version);
+	drm_printf(p, "PSP SOS feature version: %u, fw version: 0x%08x\n",
+		   adev->psp.sos.feature_version, adev->psp.sos.fw_version);
+	drm_printf(p, "PSP ASD feature version: %u, fw version: 0x%08x\n",
+		   adev->psp.asd_context.bin_desc.feature_version,
+		   adev->psp.asd_context.bin_desc.fw_version);
+
+	drm_printf(p, "TA XGMI feature version: 0x%08x, fw version: 0x%08x\n",
+		   adev->psp.xgmi_context.context.bin_desc.feature_version,
+		   adev->psp.xgmi_context.context.bin_desc.fw_version);
+	drm_printf(p, "TA RAS feature version: 0x%08x, fw version: 0x%08x\n",
+		   adev->psp.ras_context.context.bin_desc.feature_version,
+		   adev->psp.ras_context.context.bin_desc.fw_version);
+	drm_printf(p, "TA HDCP feature version: 0x%08x, fw version: 0x%08x\n",
+		   adev->psp.hdcp_context.context.bin_desc.feature_version,
+		   adev->psp.hdcp_context.context.bin_desc.fw_version);
+	drm_printf(p, "TA DTM feature version: 0x%08x, fw version: 0x%08x\n",
+		   adev->psp.dtm_context.context.bin_desc.feature_version,
+		   adev->psp.dtm_context.context.bin_desc.fw_version);
+	drm_printf(p, "TA RAP feature version: 0x%08x, fw version: 0x%08x\n",
+		   adev->psp.rap_context.context.bin_desc.feature_version,
+		   adev->psp.rap_context.context.bin_desc.fw_version);
+	drm_printf(p,
+		   "TA SECURE DISPLAY feature version: 0x%08x, fw version: 0x%08x\n",
+		   adev->psp.securedisplay_context.context.bin_desc.feature_version,
+		   adev->psp.securedisplay_context.context.bin_desc.fw_version);
+
+	/* SMC firmware */
+	version = adev->pm.fw_version;
+
+	smu_program = (version >> 24) & 0xff;
+	smu_major = (version >> 16) & 0xff;
+	smu_minor = (version >> 8) & 0xff;
+	smu_debug = (version >> 0) & 0xff;
+	drm_printf(p,
+		   "SMC feature version: %u, program: %d, fw version: 0x%08x (%d.%d.%d)\n",
+		   0, smu_program, version, smu_major, smu_minor, smu_debug);
+
+	/* SDMA firmware */
+	for (int i = 0; i < adev->sdma.num_instances; i++) {
+		drm_printf(p,
+			   "SDMA%d feature version: %u, firmware version: 0x%08x\n",
+			   i, adev->sdma.instance[i].feature_version,
+			   adev->sdma.instance[i].fw_version);
+	}
+
+	drm_printf(p, "VCN feature version: %u, fw version: 0x%08x\n", 0,
+		   adev->vcn.fw_version);
+	drm_printf(p, "DMCU feature version: %u, fw version: 0x%08x\n", 0,
+		   adev->dm.dmcu_fw_version);
+	drm_printf(p, "DMCUB feature version: %u, fw version: 0x%08x\n", 0,
+		   adev->dm.dmcub_fw_version);
+	drm_printf(p, "PSP TOC feature version: %u, fw version: 0x%08x\n",
+		   adev->psp.toc.feature_version, adev->psp.toc.fw_version);
+
+	version = adev->mes.kiq_version & AMDGPU_MES_VERSION_MASK;
+	feature = (adev->mes.kiq_version & AMDGPU_MES_FEAT_VERSION_MASK) >>
+		  AMDGPU_MES_FEAT_VERSION_SHIFT;
+	drm_printf(p, "MES_KIQ feature version: %u, fw version: 0x%08x\n",
+		   feature, version);
+
+	version = adev->mes.sched_version & AMDGPU_MES_VERSION_MASK;
+	feature = (adev->mes.sched_version & AMDGPU_MES_FEAT_VERSION_MASK) >>
+		  AMDGPU_MES_FEAT_VERSION_SHIFT;
+	drm_printf(p, "MES feature version: %u, fw version: 0x%08x\n", feature,
+		   version);
+
+	drm_printf(p, "VPE feature version: %u, fw version: 0x%08x\n",
+		   adev->vpe.feature_version, adev->vpe.fw_version);
+
+	drm_printf(p, "\nVBIOS Information\n");
+	drm_printf(p, "vbios name       : %s\n", ctx->name);
+	drm_printf(p, "vbios pn         : %s\n", ctx->vbios_pn);
+	drm_printf(p, "vbios version    : %d\n", ctx->version);
+	drm_printf(p, "vbios ver_str    : %s\n", ctx->vbios_ver_str);
+	drm_printf(p, "vbios date       : %s\n", ctx->date);
+}
+
+static ssize_t
+amdgpu_devcoredump_read(char *buffer, loff_t offset, size_t count,
+			void *data, size_t datalen)
+{
+	struct drm_printer p;
+	struct amdgpu_coredump_info *coredump = data;
+	struct drm_print_iterator iter;
+	struct amdgpu_vm_fault_info *fault_info;
+	struct amdgpu_ip_block *ip_block;
+	int ver;
+
+	iter.data = buffer;
+	iter.offset = 0;
+	iter.start = offset;
+	iter.remain = count;
+
+	p = drm_coredump_printer(&iter);
+
+	drm_printf(&p, "**** AMDGPU Device Coredump ****\n");
+	drm_printf(&p, "version: " AMDGPU_COREDUMP_VERSION "\n");
+	drm_printf(&p, "kernel: " UTS_RELEASE "\n");
+	drm_printf(&p, "module: " KBUILD_MODNAME "\n");
+	drm_printf(&p, "time: %ptSp\n", &coredump->reset_time);
+
+	if (coredump->reset_task_info.task.pid)
+		drm_printf(&p, "process_name: %s PID: %d\n",
+			   coredump->reset_task_info.process_name,
+			   coredump->reset_task_info.task.pid);
+
+	/* SOC Information */
+	drm_printf(&p, "\nSOC Information\n");
+	drm_printf(&p, "SOC Device id: %d\n", coredump->adev->pdev->device);
+	drm_printf(&p, "SOC PCI Revision id: %d\n", coredump->adev->pdev->revision);
+	drm_printf(&p, "SOC Family: %d\n", coredump->adev->family);
+	drm_printf(&p, "SOC Revision id: %d\n", coredump->adev->rev_id);
+	drm_printf(&p, "SOC External Revision id: %d\n", coredump->adev->external_rev_id);
+
+	/* Memory Information */
+	drm_printf(&p, "\nSOC Memory Information\n");
+	drm_printf(&p, "real vram size: %llu\n", coredump->adev->gmc.real_vram_size);
+	drm_printf(&p, "visible vram size: %llu\n", coredump->adev->gmc.visible_vram_size);
+	drm_printf(&p, "gtt size: %llu\n", coredump->adev->mman.gtt_mgr.manager.size);
+
+	/* GDS Config */
+	drm_printf(&p, "\nGDS Config\n");
+	drm_printf(&p, "gds: total size: %d\n", coredump->adev->gds.gds_size);
+	drm_printf(&p, "gds: compute partition size: %d\n", coredump->adev->gds.gds_size);
+	drm_printf(&p, "gds: gws per compute partition: %d\n", coredump->adev->gds.gws_size);
+	drm_printf(&p, "gds: os per compute partition: %d\n", coredump->adev->gds.oa_size);
+
+	/* HWIP Version Information */
+	drm_printf(&p, "\nHW IP Version Information\n");
+	for (int i = 1; i < MAX_HWIP; i++) {
+		for (int j = 0; j < HWIP_MAX_INSTANCE; j++) {
+			ver = coredump->adev->ip_versions[i][j];
+			if (ver)
+				drm_printf(&p, "HWIP: %s[%d][%d]: v%d.%d.%d.%d.%d\n",
+					   hw_ip_names[i], i, j,
+					   IP_VERSION_MAJ(ver),
+					   IP_VERSION_MIN(ver),
+					   IP_VERSION_REV(ver),
+					   IP_VERSION_VARIANT(ver),
+					   IP_VERSION_SUBREV(ver));
+		}
+	}
+
+	/* IP firmware information */
+	drm_printf(&p, "\nIP Firmwares\n");
+	amdgpu_devcoredump_fw_info(coredump->adev, &p);
+
+	if (coredump->ring) {
+		drm_printf(&p, "\nRing timed out details\n");
+		drm_printf(&p, "IP Type: %d Ring Name: %s\n",
+			   coredump->ring->funcs->type,
+			   coredump->ring->name);
+	}
+
+	/* Add page fault information */
+	fault_info = &coredump->adev->vm_manager.fault_info;
+	drm_printf(&p, "\n[%s] Page fault observed\n",
+		   fault_info->vmhub ? "mmhub" : "gfxhub");
+	drm_printf(&p, "Faulty page starting at address: 0x%016llx\n", fault_info->addr);
+	drm_printf(&p, "Protection fault status register: 0x%x\n\n", fault_info->status);
+
+	/* dump the ip state for each ip */
+	drm_printf(&p, "IP Dump\n");
+	for (int i = 0; i < coredump->adev->num_ip_blocks; i++) {
+		ip_block = &coredump->adev->ip_blocks[i];
+		if (ip_block->version->funcs->print_ip_state) {
+			drm_printf(&p, "IP: %s\n", ip_block->version->funcs->name);
+			ip_block->version->funcs->print_ip_state(ip_block, &p);
+			drm_printf(&p, "\n");
+		}
+	}
+
+	/* Add ring buffer information */
+	drm_printf(&p, "Ring buffer information\n");
+	for (int i = 0; i < coredump->adev->num_rings; i++) {
+		int j = 0;
+		struct amdgpu_ring *ring = coredump->adev->rings[i];
+
+		drm_printf(&p, "ring name: %s\n", ring->name);
+		drm_printf(&p, "Rptr: 0x%llx Wptr: 0x%llx RB mask: %x\n",
+			   amdgpu_ring_get_rptr(ring),
+			   amdgpu_ring_get_wptr(ring),
+			   ring->buf_mask);
+		drm_printf(&p, "Ring size in dwords: %d\n",
+			   ring->ring_size / 4);
+		drm_printf(&p, "Ring contents\n");
+		drm_printf(&p, "Offset \t Value\n");
+
+		while (j < ring->ring_size) {
+			drm_printf(&p, "0x%x \t 0x%x\n", j, ring->ring[j / 4]);
+			j += 4;
+		}
+	}
+
+	if (coredump->skip_vram_check)
+		drm_printf(&p, "VRAM lost check is skipped!\n");
+	else if (coredump->reset_vram_lost)
+		drm_printf(&p, "VRAM is lost due to GPU reset!\n");
+
+	return count - iter.remain;
+}
+
+static void amdgpu_devcoredump_free(void *data)
+{
+	kfree(data);
+}
+
+void amdgpu_coredump(struct amdgpu_device *adev, bool skip_vram_check,
+		     bool vram_lost, struct amdgpu_job *job)
+{
+	struct drm_device *dev = adev_to_drm(adev);
+	struct amdgpu_coredump_info *coredump;
+	struct drm_sched_job *s_job;
+
+	coredump = kzalloc(sizeof(*coredump), GFP_NOWAIT);
+
+	if (!coredump) {
+		DRM_ERROR("%s: failed to allocate memory for coredump\n", __func__);
+		return;
+	}
+
+	coredump->skip_vram_check = skip_vram_check;
+	coredump->reset_vram_lost = vram_lost;
+
+	if (job && job->pasid) {
+		struct amdgpu_task_info *ti;
+
+		ti = amdgpu_vm_get_task_info_pasid(adev, job->pasid);
+		if (ti) {
+			coredump->reset_task_info = *ti;
+			amdgpu_vm_put_task_info(ti);
+		}
+	}
+
+	if (job) {
+		s_job = &job->base;
+		coredump->ring = to_amdgpu_ring(s_job->sched);
+	}
+
+	coredump->adev = adev;
+
+	ktime_get_ts64(&coredump->reset_time);
+
+	dev_coredumpm(dev->dev, THIS_MODULE, coredump, 0, GFP_NOWAIT,
+		      amdgpu_devcoredump_read, amdgpu_devcoredump_free);
+
+	drm_info(dev, "AMDGPU device coredump file has been created\n");
+	drm_info(dev, "Check your /sys/class/drm/card%d/device/devcoredump/data\n",
+		 dev->primary->index);
+}
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_dev_coredump.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_dev_coredump.h
new file mode 100644
index 000000000000..ef9772c6bcc9
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_dev_coredump.h
@@ -0,0 +1,46 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2024 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __AMDGPU_DEV_COREDUMP_H__
+#define __AMDGPU_DEV_COREDUMP_H__
+
+#include "amdgpu.h"
+
+#ifdef CONFIG_DEV_COREDUMP
+
+#define AMDGPU_COREDUMP_VERSION "1"
+
+struct amdgpu_coredump_info {
+	struct amdgpu_device            *adev;
+	struct amdgpu_task_info         reset_task_info;
+	struct timespec64               reset_time;
+	bool                            skip_vram_check;
+	bool                            reset_vram_lost;
+	struct amdgpu_ring              *ring;
+};
+#endif
+
+void amdgpu_coredump(struct amdgpu_device *adev, bool skip_vram_check,
+		     bool vram_lost, struct amdgpu_job *job);
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index 1e651b959141..58c3ffe707d1 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -25,15 +25,24 @@
  *          Alex Deucher
  *          Jerome Glisse
  */
+
+#include <linux/aperture.h>
 #include <linux/power_supply.h>
 #include <linux/kthread.h>
 #include <linux/module.h>
 #include <linux/console.h>
 #include <linux/slab.h>
+#include <linux/iommu.h>
+#include <linux/pci.h>
+#include <linux/pci-p2pdma.h>
+#include <linux/apple-gmux.h>
 
 #include <drm/drm_atomic_helper.h>
+#include <drm/drm_client_event.h>
+#include <drm/drm_crtc_helper.h>
 #include <drm/drm_probe_helper.h>
 #include <drm/amdgpu_drm.h>
+#include <linux/device.h>
 #include <linux/vgaarb.h>
 #include <linux/vga_switcheroo.h>
 #include <linux/efi.h>
@@ -54,7 +63,6 @@
 #include "soc15.h"
 #include "nv.h"
 #include "bif/bif_4_1_d.h"
-#include <linux/pci.h>
 #include <linux/firmware.h>
 #include "amdgpu_vf_error.h"
 
@@ -63,9 +71,12 @@
 
 #include "amdgpu_xgmi.h"
 #include "amdgpu_ras.h"
+#include "amdgpu_ras_mgr.h"
 #include "amdgpu_pmu.h"
 #include "amdgpu_fru_eeprom.h"
 #include "amdgpu_reset.h"
+#include "amdgpu_virt.h"
+#include "amdgpu_dev_coredump.h"
 
 #include <linux/suspend.h>
 #include <drm/task_barrier.h>
@@ -73,20 +84,31 @@
 
 #include <drm/drm_drv.h>
 
+#if IS_ENABLED(CONFIG_X86)
+#include <asm/intel-family.h>
+#include <asm/cpu_device_id.h>
+#endif
+
 MODULE_FIRMWARE("amdgpu/vega10_gpu_info.bin");
 MODULE_FIRMWARE("amdgpu/vega12_gpu_info.bin");
 MODULE_FIRMWARE("amdgpu/raven_gpu_info.bin");
 MODULE_FIRMWARE("amdgpu/picasso_gpu_info.bin");
 MODULE_FIRMWARE("amdgpu/raven2_gpu_info.bin");
 MODULE_FIRMWARE("amdgpu/arcturus_gpu_info.bin");
-MODULE_FIRMWARE("amdgpu/renoir_gpu_info.bin");
-MODULE_FIRMWARE("amdgpu/navi10_gpu_info.bin");
-MODULE_FIRMWARE("amdgpu/navi14_gpu_info.bin");
 MODULE_FIRMWARE("amdgpu/navi12_gpu_info.bin");
-MODULE_FIRMWARE("amdgpu/vangogh_gpu_info.bin");
-MODULE_FIRMWARE("amdgpu/yellow_carp_gpu_info.bin");
+MODULE_FIRMWARE("amdgpu/cyan_skillfish_gpu_info.bin");
 
 #define AMDGPU_RESUME_MS		2000
+#define AMDGPU_MAX_RETRY_LIMIT		2
+#define AMDGPU_RETRY_SRIOV_RESET(r) ((r) == -EBUSY || (r) == -ETIMEDOUT || (r) == -EINVAL)
+#define AMDGPU_PCIE_INDEX_FALLBACK (0x38 >> 2)
+#define AMDGPU_PCIE_INDEX_HI_FALLBACK (0x44 >> 2)
+#define AMDGPU_PCIE_DATA_FALLBACK (0x3C >> 2)
+
+#define AMDGPU_VBIOS_SKIP (1U << 0)
+#define AMDGPU_VBIOS_OPTIONAL (1U << 1)
+
+static const struct drm_driver amdgpu_kms_driver;
 
 const char *amdgpu_asic_name[] = {
 	"TAHITI",
@@ -129,13 +151,76 @@ const char *amdgpu_asic_name[] = {
 	"LAST",
 };
 
+#define AMDGPU_IP_BLK_MASK_ALL GENMASK(AMD_IP_BLOCK_TYPE_NUM  - 1, 0)
+/*
+ * Default init level where all blocks are expected to be initialized. This is
+ * the level of initialization expected by default and also after a full reset
+ * of the device.
+ */
+struct amdgpu_init_level amdgpu_init_default = {
+	.level = AMDGPU_INIT_LEVEL_DEFAULT,
+	.hwini_ip_block_mask = AMDGPU_IP_BLK_MASK_ALL,
+};
+
+struct amdgpu_init_level amdgpu_init_recovery = {
+	.level = AMDGPU_INIT_LEVEL_RESET_RECOVERY,
+	.hwini_ip_block_mask = AMDGPU_IP_BLK_MASK_ALL,
+};
+
+/*
+ * Minimal blocks needed to be initialized before a XGMI hive can be reset. This
+ * is used for cases like reset on initialization where the entire hive needs to
+ * be reset before first use.
+ */
+struct amdgpu_init_level amdgpu_init_minimal_xgmi = {
+	.level = AMDGPU_INIT_LEVEL_MINIMAL_XGMI,
+	.hwini_ip_block_mask =
+		BIT(AMD_IP_BLOCK_TYPE_GMC) | BIT(AMD_IP_BLOCK_TYPE_SMC) |
+		BIT(AMD_IP_BLOCK_TYPE_COMMON) | BIT(AMD_IP_BLOCK_TYPE_IH) |
+		BIT(AMD_IP_BLOCK_TYPE_PSP)
+};
+
+static int amdgpu_device_ip_resume_phase1(struct amdgpu_device *adev);
+static int amdgpu_device_ip_resume_phase2(struct amdgpu_device *adev);
+static int amdgpu_device_ip_resume_phase3(struct amdgpu_device *adev);
+
+static void amdgpu_device_load_switch_state(struct amdgpu_device *adev);
+
+static inline bool amdgpu_ip_member_of_hwini(struct amdgpu_device *adev,
+					     enum amd_ip_block_type block)
+{
+	return (adev->init_lvl->hwini_ip_block_mask & (1U << block)) != 0;
+}
+
+void amdgpu_set_init_level(struct amdgpu_device *adev,
+			   enum amdgpu_init_lvl_id lvl)
+{
+	switch (lvl) {
+	case AMDGPU_INIT_LEVEL_MINIMAL_XGMI:
+		adev->init_lvl = &amdgpu_init_minimal_xgmi;
+		break;
+	case AMDGPU_INIT_LEVEL_RESET_RECOVERY:
+		adev->init_lvl = &amdgpu_init_recovery;
+		break;
+	case AMDGPU_INIT_LEVEL_DEFAULT:
+		fallthrough;
+	default:
+		adev->init_lvl = &amdgpu_init_default;
+		break;
+	}
+}
+
+static inline void amdgpu_device_stop_pending_resets(struct amdgpu_device *adev);
+static int amdgpu_device_pm_notifier(struct notifier_block *nb, unsigned long mode,
+				     void *data);
+
 /**
  * DOC: pcie_replay_count
  *
  * The amdgpu driver provides a sysfs API for reporting the total number
- * of PCIe replays (NAKs)
+ * of PCIe replays (NAKs).
  * The file pcie_replay_count is used for this and returns the total
- * number of replays as a sum of the NAKs generated and NAKs received
+ * number of replays as a sum of the NAKs generated and NAKs received.
  */
 
 static ssize_t amdgpu_device_get_pcie_replay_count(struct device *dev,
@@ -148,89 +233,202 @@ static ssize_t amdgpu_device_get_pcie_replay_count(struct device *dev,
 	return sysfs_emit(buf, "%llu\n", cnt);
 }
 
-static DEVICE_ATTR(pcie_replay_count, S_IRUGO,
+static DEVICE_ATTR(pcie_replay_count, 0444,
 		amdgpu_device_get_pcie_replay_count, NULL);
 
-static void amdgpu_device_get_pcie_info(struct amdgpu_device *adev);
+static int amdgpu_device_attr_sysfs_init(struct amdgpu_device *adev)
+{
+	int ret = 0;
 
-/**
- * DOC: product_name
- *
- * The amdgpu driver provides a sysfs API for reporting the product name
- * for the device
- * The file serial_number is used for this and returns the product name
- * as returned from the FRU.
- * NOTE: This is only available for certain server cards
- */
+	if (amdgpu_nbio_is_replay_cnt_supported(adev))
+		ret = sysfs_create_file(&adev->dev->kobj,
+					&dev_attr_pcie_replay_count.attr);
 
-static ssize_t amdgpu_device_get_product_name(struct device *dev,
-		struct device_attribute *attr, char *buf)
+	return ret;
+}
+
+static void amdgpu_device_attr_sysfs_fini(struct amdgpu_device *adev)
+{
+	if (amdgpu_nbio_is_replay_cnt_supported(adev))
+		sysfs_remove_file(&adev->dev->kobj,
+				  &dev_attr_pcie_replay_count.attr);
+}
+
+static ssize_t amdgpu_sysfs_reg_state_get(struct file *f, struct kobject *kobj,
+					  const struct bin_attribute *attr, char *buf,
+					  loff_t ppos, size_t count)
 {
+	struct device *dev = kobj_to_dev(kobj);
 	struct drm_device *ddev = dev_get_drvdata(dev);
 	struct amdgpu_device *adev = drm_to_adev(ddev);
+	ssize_t bytes_read;
+
+	switch (ppos) {
+	case AMDGPU_SYS_REG_STATE_XGMI:
+		bytes_read = amdgpu_asic_get_reg_state(
+			adev, AMDGPU_REG_STATE_TYPE_XGMI, buf, count);
+		break;
+	case AMDGPU_SYS_REG_STATE_WAFL:
+		bytes_read = amdgpu_asic_get_reg_state(
+			adev, AMDGPU_REG_STATE_TYPE_WAFL, buf, count);
+		break;
+	case AMDGPU_SYS_REG_STATE_PCIE:
+		bytes_read = amdgpu_asic_get_reg_state(
+			adev, AMDGPU_REG_STATE_TYPE_PCIE, buf, count);
+		break;
+	case AMDGPU_SYS_REG_STATE_USR:
+		bytes_read = amdgpu_asic_get_reg_state(
+			adev, AMDGPU_REG_STATE_TYPE_USR, buf, count);
+		break;
+	case AMDGPU_SYS_REG_STATE_USR_1:
+		bytes_read = amdgpu_asic_get_reg_state(
+			adev, AMDGPU_REG_STATE_TYPE_USR_1, buf, count);
+		break;
+	default:
+		return -EINVAL;
+	}
 
-	return sysfs_emit(buf, "%s\n", adev->product_name);
+	return bytes_read;
 }
 
-static DEVICE_ATTR(product_name, S_IRUGO,
-		amdgpu_device_get_product_name, NULL);
+static const BIN_ATTR(reg_state, 0444, amdgpu_sysfs_reg_state_get, NULL,
+		      AMDGPU_SYS_REG_STATE_END);
+
+int amdgpu_reg_state_sysfs_init(struct amdgpu_device *adev)
+{
+	int ret;
+
+	if (!amdgpu_asic_get_reg_state_supported(adev))
+		return 0;
+
+	ret = sysfs_create_bin_file(&adev->dev->kobj, &bin_attr_reg_state);
+
+	return ret;
+}
+
+void amdgpu_reg_state_sysfs_fini(struct amdgpu_device *adev)
+{
+	if (!amdgpu_asic_get_reg_state_supported(adev))
+		return;
+	sysfs_remove_bin_file(&adev->dev->kobj, &bin_attr_reg_state);
+}
+
+int amdgpu_ip_block_suspend(struct amdgpu_ip_block *ip_block)
+{
+	int r;
+
+	if (ip_block->version->funcs->suspend) {
+		r = ip_block->version->funcs->suspend(ip_block);
+		if (r) {
+			dev_err(ip_block->adev->dev,
+				"suspend of IP block <%s> failed %d\n",
+				ip_block->version->funcs->name, r);
+			return r;
+		}
+	}
+
+	ip_block->status.hw = false;
+	return 0;
+}
+
+int amdgpu_ip_block_resume(struct amdgpu_ip_block *ip_block)
+{
+	int r;
+
+	if (ip_block->version->funcs->resume) {
+		r = ip_block->version->funcs->resume(ip_block);
+		if (r) {
+			dev_err(ip_block->adev->dev,
+				"resume of IP block <%s> failed %d\n",
+				ip_block->version->funcs->name, r);
+			return r;
+		}
+	}
+
+	ip_block->status.hw = true;
+	return 0;
+}
 
 /**
- * DOC: product_number
+ * DOC: board_info
+ *
+ * The amdgpu driver provides a sysfs API for giving board related information.
+ * It provides the form factor information in the format
+ *
+ *   type : form factor
+ *
+ * Possible form factor values
+ *
+ * - "cem"		- PCIE CEM card
+ * - "oam"		- Open Compute Accelerator Module
+ * - "unknown"	- Not known
  *
- * The amdgpu driver provides a sysfs API for reporting the part number
- * for the device
- * The file serial_number is used for this and returns the part number
- * as returned from the FRU.
- * NOTE: This is only available for certain server cards
  */
 
-static ssize_t amdgpu_device_get_product_number(struct device *dev,
-		struct device_attribute *attr, char *buf)
+static ssize_t amdgpu_device_get_board_info(struct device *dev,
+					    struct device_attribute *attr,
+					    char *buf)
 {
 	struct drm_device *ddev = dev_get_drvdata(dev);
 	struct amdgpu_device *adev = drm_to_adev(ddev);
+	enum amdgpu_pkg_type pkg_type = AMDGPU_PKG_TYPE_CEM;
+	const char *pkg;
+
+	if (adev->smuio.funcs && adev->smuio.funcs->get_pkg_type)
+		pkg_type = adev->smuio.funcs->get_pkg_type(adev);
 
-	return sysfs_emit(buf, "%s\n", adev->product_number);
+	switch (pkg_type) {
+	case AMDGPU_PKG_TYPE_CEM:
+		pkg = "cem";
+		break;
+	case AMDGPU_PKG_TYPE_OAM:
+		pkg = "oam";
+		break;
+	default:
+		pkg = "unknown";
+		break;
+	}
+
+	return sysfs_emit(buf, "%s : %s\n", "type", pkg);
 }
 
-static DEVICE_ATTR(product_number, S_IRUGO,
-		amdgpu_device_get_product_number, NULL);
+static DEVICE_ATTR(board_info, 0444, amdgpu_device_get_board_info, NULL);
 
-/**
- * DOC: serial_number
- *
- * The amdgpu driver provides a sysfs API for reporting the serial number
- * for the device
- * The file serial_number is used for this and returns the serial number
- * as returned from the FRU.
- * NOTE: This is only available for certain server cards
- */
+static struct attribute *amdgpu_board_attrs[] = {
+	&dev_attr_board_info.attr,
+	NULL,
+};
 
-static ssize_t amdgpu_device_get_serial_number(struct device *dev,
-		struct device_attribute *attr, char *buf)
+static umode_t amdgpu_board_attrs_is_visible(struct kobject *kobj,
+					     struct attribute *attr, int n)
 {
+	struct device *dev = kobj_to_dev(kobj);
 	struct drm_device *ddev = dev_get_drvdata(dev);
 	struct amdgpu_device *adev = drm_to_adev(ddev);
 
-	return sysfs_emit(buf, "%s\n", adev->serial);
+	if (adev->flags & AMD_IS_APU)
+		return 0;
+
+	return attr->mode;
 }
 
-static DEVICE_ATTR(serial_number, S_IRUGO,
-		amdgpu_device_get_serial_number, NULL);
+static const struct attribute_group amdgpu_board_attrs_group = {
+	.attrs = amdgpu_board_attrs,
+	.is_visible = amdgpu_board_attrs_is_visible
+};
+
+static void amdgpu_device_get_pcie_info(struct amdgpu_device *adev);
 
 /**
  * amdgpu_device_supports_px - Is the device a dGPU with ATPX power control
  *
- * @dev: drm_device pointer
+ * @adev: amdgpu device pointer
  *
  * Returns true if the device is a dGPU with ATPX power control,
  * otherwise return false.
  */
-bool amdgpu_device_supports_px(struct drm_device *dev)
+bool amdgpu_device_supports_px(struct amdgpu_device *adev)
 {
-	struct amdgpu_device *adev = drm_to_adev(dev);
-
 	if ((adev->flags & AMD_IS_PX) && !amdgpu_is_atpx_hybrid())
 		return true;
 	return false;
@@ -239,14 +437,15 @@ bool amdgpu_device_supports_px(struct drm_device *dev)
 /**
  * amdgpu_device_supports_boco - Is the device a dGPU with ACPI power resources
  *
- * @dev: drm_device pointer
+ * @adev: amdgpu device pointer
  *
  * Returns true if the device is a dGPU with ACPI power control,
  * otherwise return false.
  */
-bool amdgpu_device_supports_boco(struct drm_device *dev)
+bool amdgpu_device_supports_boco(struct amdgpu_device *adev)
 {
-	struct amdgpu_device *adev = drm_to_adev(dev);
+	if (!IS_ENABLED(CONFIG_HOTPLUG_PCI_PCIE))
+		return false;
 
 	if (adev->has_pr3 ||
 	    ((adev->flags & AMD_IS_PX) && amdgpu_is_atpx_hybrid()))
@@ -257,30 +456,105 @@ bool amdgpu_device_supports_boco(struct drm_device *dev)
 /**
  * amdgpu_device_supports_baco - Does the device support BACO
  *
- * @dev: drm_device pointer
+ * @adev: amdgpu device pointer
  *
- * Returns true if the device supporte BACO,
- * otherwise return false.
+ * Return:
+ * 1 if the device supports BACO;
+ * 3 if the device supports MACO (only works if BACO is supported)
+ * otherwise return 0.
  */
-bool amdgpu_device_supports_baco(struct drm_device *dev)
+int amdgpu_device_supports_baco(struct amdgpu_device *adev)
 {
-	struct amdgpu_device *adev = drm_to_adev(dev);
-
 	return amdgpu_asic_supports_baco(adev);
 }
 
+void amdgpu_device_detect_runtime_pm_mode(struct amdgpu_device *adev)
+{
+	int bamaco_support;
+
+	adev->pm.rpm_mode = AMDGPU_RUNPM_NONE;
+	bamaco_support = amdgpu_device_supports_baco(adev);
+
+	switch (amdgpu_runtime_pm) {
+	case 2:
+		if (bamaco_support & MACO_SUPPORT) {
+			adev->pm.rpm_mode = AMDGPU_RUNPM_BAMACO;
+			dev_info(adev->dev, "Forcing BAMACO for runtime pm\n");
+		} else if (bamaco_support == BACO_SUPPORT) {
+			adev->pm.rpm_mode = AMDGPU_RUNPM_BACO;
+			dev_info(adev->dev, "Requested mode BAMACO not available,fallback to use BACO\n");
+		}
+		break;
+	case 1:
+		if (bamaco_support & BACO_SUPPORT) {
+			adev->pm.rpm_mode = AMDGPU_RUNPM_BACO;
+			dev_info(adev->dev, "Forcing BACO for runtime pm\n");
+		}
+		break;
+	case -1:
+	case -2:
+		if (amdgpu_device_supports_px(adev)) {
+			/* enable PX as runtime mode */
+			adev->pm.rpm_mode = AMDGPU_RUNPM_PX;
+			dev_info(adev->dev, "Using ATPX for runtime pm\n");
+		} else if (amdgpu_device_supports_boco(adev)) {
+			/* enable boco as runtime mode */
+			adev->pm.rpm_mode = AMDGPU_RUNPM_BOCO;
+			dev_info(adev->dev, "Using BOCO for runtime pm\n");
+		} else {
+			if (!bamaco_support)
+				goto no_runtime_pm;
+
+			switch (adev->asic_type) {
+			case CHIP_VEGA20:
+			case CHIP_ARCTURUS:
+				/* BACO are not supported on vega20 and arctrus */
+				break;
+			case CHIP_VEGA10:
+				/* enable BACO as runpm mode if noretry=0 */
+				if (!adev->gmc.noretry && !amdgpu_passthrough(adev))
+					adev->pm.rpm_mode = AMDGPU_RUNPM_BACO;
+				break;
+			default:
+				/* enable BACO as runpm mode on CI+ */
+				if (!amdgpu_passthrough(adev))
+					adev->pm.rpm_mode = AMDGPU_RUNPM_BACO;
+				break;
+			}
+
+			if (adev->pm.rpm_mode == AMDGPU_RUNPM_BACO) {
+				if (bamaco_support & MACO_SUPPORT) {
+					adev->pm.rpm_mode = AMDGPU_RUNPM_BAMACO;
+					dev_info(adev->dev, "Using BAMACO for runtime pm\n");
+				} else {
+					dev_info(adev->dev, "Using BACO for runtime pm\n");
+				}
+			}
+		}
+		break;
+	case 0:
+		dev_info(adev->dev, "runtime pm is manually disabled\n");
+		break;
+	default:
+		break;
+	}
+
+no_runtime_pm:
+	if (adev->pm.rpm_mode == AMDGPU_RUNPM_NONE)
+		dev_info(adev->dev, "Runtime PM not available\n");
+}
 /**
  * amdgpu_device_supports_smart_shift - Is the device dGPU with
  * smart shift support
  *
- * @dev: drm_device pointer
+ * @adev: amdgpu device pointer
  *
  * Returns true if the device is a dGPU with Smart Shift support,
  * otherwise returns false.
  */
-bool amdgpu_device_supports_smart_shift(struct drm_device *dev)
+bool amdgpu_device_supports_smart_shift(struct amdgpu_device *adev)
 {
-	return (amdgpu_device_supports_boco(dev) &&
+	return (amdgpu_device_supports_boco(adev) &&
 		amdgpu_acpi_is_power_shift_control_supported());
 }
 
@@ -331,7 +605,7 @@ void amdgpu_device_mm_access(struct amdgpu_device *adev, loff_t pos,
 }
 
 /**
- * amdgpu_device_vram_access - access vram by vram aperature
+ * amdgpu_device_aper_access - access vram by vram aperture
  *
  * @adev: amdgpu_device pointer
  * @pos: offset of the buffer in vram
@@ -359,10 +633,16 @@ size_t amdgpu_device_aper_access(struct amdgpu_device *adev, loff_t pos,
 
 		if (write) {
 			memcpy_toio(addr, buf, count);
+			/* Make sure HDP write cache flush happens without any reordering
+			 * after the system memory contents are sent over PCIe device
+			 */
 			mb();
 			amdgpu_device_flush_hdp(adev, NULL);
 		} else {
 			amdgpu_device_invalidate_hdp(adev, NULL);
+			/* Make sure HDP read cache is invalidated before issuing a read
+			 * to the PCIe device
+			 */
 			mb();
 			memcpy_fromio(buf, addr, count);
 		}
@@ -416,17 +696,17 @@ bool amdgpu_device_skip_hw_access(struct amdgpu_device *adev)
 	 * here is that the GPU reset is not running on another thread in parallel.
 	 *
 	 * For this we trylock the read side of the reset semaphore, if that succeeds
-	 * we know that the reset is not running in paralell.
+	 * we know that the reset is not running in parallel.
 	 *
 	 * If the trylock fails we assert that we are either already holding the read
 	 * side of the lock or are the reset thread itself and hold the write side of
 	 * the lock.
 	 */
 	if (in_task()) {
-		if (down_read_trylock(&adev->reset_sem))
-			up_read(&adev->reset_sem);
+		if (down_read_trylock(&adev->reset_domain->sem))
+			up_read(&adev->reset_domain->sem);
 		else
-			lockdep_assert_held(&adev->reset_sem);
+			lockdep_assert_held(&adev->reset_domain->sem);
 	}
 #endif
 	return false;
@@ -452,9 +732,9 @@ uint32_t amdgpu_device_rreg(struct amdgpu_device *adev,
 	if ((reg * 4) < adev->rmmio_size) {
 		if (!(acc_flags & AMDGPU_REGS_NO_KIQ) &&
 		    amdgpu_sriov_runtime(adev) &&
-		    down_read_trylock(&adev->reset_sem)) {
-			ret = amdgpu_kiq_rreg(adev, reg);
-			up_read(&adev->reset_sem);
+		    down_read_trylock(&adev->reset_domain->sem)) {
+			ret = amdgpu_kiq_rreg(adev, reg, 0);
+			up_read(&adev->reset_domain->sem);
 		} else {
 			ret = readl(((void __iomem *)adev->rmmio) + (reg * 4));
 		}
@@ -470,8 +750,7 @@ uint32_t amdgpu_device_rreg(struct amdgpu_device *adev,
 /*
  * MMIO register read with bytes helper functions
  * @offset:bytes offset from MMIO start
- *
-*/
+ */
 
 /**
  * amdgpu_mm_rreg8 - read a memory mapped IO register
@@ -491,12 +770,55 @@ uint8_t amdgpu_mm_rreg8(struct amdgpu_device *adev, uint32_t offset)
 	BUG();
 }
 
+
+/**
+ * amdgpu_device_xcc_rreg - read a memory mapped IO or indirect register with specific XCC
+ *
+ * @adev: amdgpu_device pointer
+ * @reg: dword aligned register offset
+ * @acc_flags: access flags which require special behavior
+ * @xcc_id: xcc accelerated compute core id
+ *
+ * Returns the 32 bit value from the offset specified.
+ */
+uint32_t amdgpu_device_xcc_rreg(struct amdgpu_device *adev,
+				uint32_t reg, uint32_t acc_flags,
+				uint32_t xcc_id)
+{
+	uint32_t ret, rlcg_flag;
+
+	if (amdgpu_device_skip_hw_access(adev))
+		return 0;
+
+	if ((reg * 4) < adev->rmmio_size) {
+		if (amdgpu_sriov_vf(adev) &&
+		    !amdgpu_sriov_runtime(adev) &&
+		    adev->gfx.rlc.rlcg_reg_access_supported &&
+		    amdgpu_virt_get_rlcg_reg_access_flag(adev, acc_flags,
+							 GC_HWIP, false,
+							 &rlcg_flag)) {
+			ret = amdgpu_virt_rlcg_reg_rw(adev, reg, 0, rlcg_flag, GET_INST(GC, xcc_id));
+		} else if (!(acc_flags & AMDGPU_REGS_NO_KIQ) &&
+		    amdgpu_sriov_runtime(adev) &&
+		    down_read_trylock(&adev->reset_domain->sem)) {
+			ret = amdgpu_kiq_rreg(adev, reg, xcc_id);
+			up_read(&adev->reset_domain->sem);
+		} else {
+			ret = readl(((void __iomem *)adev->rmmio) + (reg * 4));
+		}
+	} else {
+		ret = adev->pcie_rreg(adev, reg * 4);
+	}
+
+	return ret;
+}
+
 /*
  * MMIO register write with bytes helper functions
  * @offset:bytes offset from MMIO start
  * @value: the value want to be written to the register
- *
-*/
+ */
+
 /**
  * amdgpu_mm_wreg8 - read a memory mapped IO register
  *
@@ -537,9 +859,9 @@ void amdgpu_device_wreg(struct amdgpu_device *adev,
 	if ((reg * 4) < adev->rmmio_size) {
 		if (!(acc_flags & AMDGPU_REGS_NO_KIQ) &&
 		    amdgpu_sriov_runtime(adev) &&
-		    down_read_trylock(&adev->reset_sem)) {
-			amdgpu_kiq_wreg(adev, reg, v);
-			up_read(&adev->reset_sem);
+		    down_read_trylock(&adev->reset_domain->sem)) {
+			amdgpu_kiq_wreg(adev, reg, v, 0);
+			up_read(&adev->reset_domain->sem);
 		} else {
 			writel(v, ((void __iomem *)adev->rmmio) + (reg * 4));
 		}
@@ -550,13 +872,19 @@ void amdgpu_device_wreg(struct amdgpu_device *adev,
 	trace_amdgpu_device_wreg(adev->pdev->device, reg, v);
 }
 
-/*
- * amdgpu_mm_wreg_mmio_rlc -  write register either with mmio or with RLC path if in range
+/**
+ * amdgpu_mm_wreg_mmio_rlc -  write register either with direct/indirect mmio or with RLC path if in range
  *
- * this function is invoked only the debugfs register access
- * */
+ * @adev: amdgpu_device pointer
+ * @reg: mmio/rlc register
+ * @v: value to write
+ * @xcc_id: xcc accelerated compute core id
+ *
+ * this function is invoked only for the debugfs register access
+ */
 void amdgpu_mm_wreg_mmio_rlc(struct amdgpu_device *adev,
-			     uint32_t reg, uint32_t v)
+			     uint32_t reg, uint32_t v,
+			     uint32_t xcc_id)
 {
 	if (amdgpu_device_skip_hw_access(adev))
 		return;
@@ -565,126 +893,133 @@ void amdgpu_mm_wreg_mmio_rlc(struct amdgpu_device *adev,
 	    adev->gfx.rlc.funcs &&
 	    adev->gfx.rlc.funcs->is_rlcg_access_range) {
 		if (adev->gfx.rlc.funcs->is_rlcg_access_range(adev, reg))
-			return adev->gfx.rlc.funcs->sriov_wreg(adev, reg, v, 0, 0);
+			return amdgpu_sriov_wreg(adev, reg, v, 0, 0, xcc_id);
+	} else if ((reg * 4) >= adev->rmmio_size) {
+		adev->pcie_wreg(adev, reg * 4, v);
 	} else {
 		writel(v, ((void __iomem *)adev->rmmio) + (reg * 4));
 	}
 }
 
 /**
- * amdgpu_mm_rdoorbell - read a doorbell dword
+ * amdgpu_device_xcc_wreg - write to a memory mapped IO or indirect register with specific XCC
  *
  * @adev: amdgpu_device pointer
- * @index: doorbell index
+ * @reg: dword aligned register offset
+ * @v: 32 bit value to write to the register
+ * @acc_flags: access flags which require special behavior
+ * @xcc_id: xcc accelerated compute core id
  *
- * Returns the value in the doorbell aperture at the
- * requested doorbell index (CIK).
+ * Writes the value specified to the offset specified.
  */
-u32 amdgpu_mm_rdoorbell(struct amdgpu_device *adev, u32 index)
+void amdgpu_device_xcc_wreg(struct amdgpu_device *adev,
+			uint32_t reg, uint32_t v,
+			uint32_t acc_flags, uint32_t xcc_id)
 {
-	if (amdgpu_device_skip_hw_access(adev))
-		return 0;
-
-	if (index < adev->doorbell.num_doorbells) {
-		return readl(adev->doorbell.ptr + index);
-	} else {
-		DRM_ERROR("reading beyond doorbell aperture: 0x%08x!\n", index);
-		return 0;
-	}
-}
+	uint32_t rlcg_flag;
 
-/**
- * amdgpu_mm_wdoorbell - write a doorbell dword
- *
- * @adev: amdgpu_device pointer
- * @index: doorbell index
- * @v: value to write
- *
- * Writes @v to the doorbell aperture at the
- * requested doorbell index (CIK).
- */
-void amdgpu_mm_wdoorbell(struct amdgpu_device *adev, u32 index, u32 v)
-{
 	if (amdgpu_device_skip_hw_access(adev))
 		return;
 
-	if (index < adev->doorbell.num_doorbells) {
-		writel(v, adev->doorbell.ptr + index);
+	if ((reg * 4) < adev->rmmio_size) {
+		if (amdgpu_sriov_vf(adev) &&
+		    !amdgpu_sriov_runtime(adev) &&
+		    adev->gfx.rlc.rlcg_reg_access_supported &&
+		    amdgpu_virt_get_rlcg_reg_access_flag(adev, acc_flags,
+							 GC_HWIP, true,
+							 &rlcg_flag)) {
+			amdgpu_virt_rlcg_reg_rw(adev, reg, v, rlcg_flag, GET_INST(GC, xcc_id));
+		} else if (!(acc_flags & AMDGPU_REGS_NO_KIQ) &&
+		    amdgpu_sriov_runtime(adev) &&
+		    down_read_trylock(&adev->reset_domain->sem)) {
+			amdgpu_kiq_wreg(adev, reg, v, xcc_id);
+			up_read(&adev->reset_domain->sem);
+		} else {
+			writel(v, ((void __iomem *)adev->rmmio) + (reg * 4));
+		}
 	} else {
-		DRM_ERROR("writing beyond doorbell aperture: 0x%08x!\n", index);
+		adev->pcie_wreg(adev, reg * 4, v);
 	}
 }
 
 /**
- * amdgpu_mm_rdoorbell64 - read a doorbell Qword
+ * amdgpu_device_indirect_rreg - read an indirect register
  *
  * @adev: amdgpu_device pointer
- * @index: doorbell index
+ * @reg_addr: indirect register address to read from
  *
- * Returns the value in the doorbell aperture at the
- * requested doorbell index (VEGA10+).
+ * Returns the value of indirect register @reg_addr
  */
-u64 amdgpu_mm_rdoorbell64(struct amdgpu_device *adev, u32 index)
+u32 amdgpu_device_indirect_rreg(struct amdgpu_device *adev,
+				u32 reg_addr)
 {
-	if (amdgpu_device_skip_hw_access(adev))
-		return 0;
+	unsigned long flags, pcie_index, pcie_data;
+	void __iomem *pcie_index_offset;
+	void __iomem *pcie_data_offset;
+	u32 r;
 
-	if (index < adev->doorbell.num_doorbells) {
-		return atomic64_read((atomic64_t *)(adev->doorbell.ptr + index));
-	} else {
-		DRM_ERROR("reading beyond doorbell aperture: 0x%08x!\n", index);
-		return 0;
-	}
-}
+	pcie_index = adev->nbio.funcs->get_pcie_index_offset(adev);
+	pcie_data = adev->nbio.funcs->get_pcie_data_offset(adev);
 
-/**
- * amdgpu_mm_wdoorbell64 - write a doorbell Qword
- *
- * @adev: amdgpu_device pointer
- * @index: doorbell index
- * @v: value to write
- *
- * Writes @v to the doorbell aperture at the
- * requested doorbell index (VEGA10+).
- */
-void amdgpu_mm_wdoorbell64(struct amdgpu_device *adev, u32 index, u64 v)
-{
-	if (amdgpu_device_skip_hw_access(adev))
-		return;
+	spin_lock_irqsave(&adev->pcie_idx_lock, flags);
+	pcie_index_offset = (void __iomem *)adev->rmmio + pcie_index * 4;
+	pcie_data_offset = (void __iomem *)adev->rmmio + pcie_data * 4;
 
-	if (index < adev->doorbell.num_doorbells) {
-		atomic64_set((atomic64_t *)(adev->doorbell.ptr + index), v);
-	} else {
-		DRM_ERROR("writing beyond doorbell aperture: 0x%08x!\n", index);
-	}
+	writel(reg_addr, pcie_index_offset);
+	readl(pcie_index_offset);
+	r = readl(pcie_data_offset);
+	spin_unlock_irqrestore(&adev->pcie_idx_lock, flags);
+
+	return r;
 }
 
-/**
- * amdgpu_device_indirect_rreg - read an indirect register
- *
- * @adev: amdgpu_device pointer
- * @pcie_index: mmio register offset
- * @pcie_data: mmio register offset
- * @reg_addr: indirect register address to read from
- *
- * Returns the value of indirect register @reg_addr
- */
-u32 amdgpu_device_indirect_rreg(struct amdgpu_device *adev,
-				u32 pcie_index, u32 pcie_data,
-				u32 reg_addr)
+u32 amdgpu_device_indirect_rreg_ext(struct amdgpu_device *adev,
+				    u64 reg_addr)
 {
-	unsigned long flags;
+	unsigned long flags, pcie_index, pcie_index_hi, pcie_data;
 	u32 r;
 	void __iomem *pcie_index_offset;
+	void __iomem *pcie_index_hi_offset;
 	void __iomem *pcie_data_offset;
 
+	if (unlikely(!adev->nbio.funcs)) {
+		pcie_index = AMDGPU_PCIE_INDEX_FALLBACK;
+		pcie_data = AMDGPU_PCIE_DATA_FALLBACK;
+	} else {
+		pcie_index = adev->nbio.funcs->get_pcie_index_offset(adev);
+		pcie_data = adev->nbio.funcs->get_pcie_data_offset(adev);
+	}
+
+	if (reg_addr >> 32) {
+		if (unlikely(!adev->nbio.funcs))
+			pcie_index_hi = AMDGPU_PCIE_INDEX_HI_FALLBACK;
+		else
+			pcie_index_hi = adev->nbio.funcs->get_pcie_index_hi_offset(adev);
+	} else {
+		pcie_index_hi = 0;
+	}
+
 	spin_lock_irqsave(&adev->pcie_idx_lock, flags);
 	pcie_index_offset = (void __iomem *)adev->rmmio + pcie_index * 4;
 	pcie_data_offset = (void __iomem *)adev->rmmio + pcie_data * 4;
+	if (pcie_index_hi != 0)
+		pcie_index_hi_offset = (void __iomem *)adev->rmmio +
+				pcie_index_hi * 4;
 
 	writel(reg_addr, pcie_index_offset);
 	readl(pcie_index_offset);
+	if (pcie_index_hi != 0) {
+		writel((reg_addr >> 32) & 0xff, pcie_index_hi_offset);
+		readl(pcie_index_hi_offset);
+	}
 	r = readl(pcie_data_offset);
+
+	/* clear the high bits */
+	if (pcie_index_hi != 0) {
+		writel(0, pcie_index_hi_offset);
+		readl(pcie_index_hi_offset);
+	}
+
 	spin_unlock_irqrestore(&adev->pcie_idx_lock, flags);
 
 	return r;
@@ -694,33 +1029,83 @@ u32 amdgpu_device_indirect_rreg(struct amdgpu_device *adev,
  * amdgpu_device_indirect_rreg64 - read a 64bits indirect register
  *
  * @adev: amdgpu_device pointer
- * @pcie_index: mmio register offset
- * @pcie_data: mmio register offset
  * @reg_addr: indirect register address to read from
  *
  * Returns the value of indirect register @reg_addr
  */
 u64 amdgpu_device_indirect_rreg64(struct amdgpu_device *adev,
-				  u32 pcie_index, u32 pcie_data,
 				  u32 reg_addr)
 {
-	unsigned long flags;
+	unsigned long flags, pcie_index, pcie_data;
+	void __iomem *pcie_index_offset;
+	void __iomem *pcie_data_offset;
 	u64 r;
+
+	pcie_index = adev->nbio.funcs->get_pcie_index_offset(adev);
+	pcie_data = adev->nbio.funcs->get_pcie_data_offset(adev);
+
+	spin_lock_irqsave(&adev->pcie_idx_lock, flags);
+	pcie_index_offset = (void __iomem *)adev->rmmio + pcie_index * 4;
+	pcie_data_offset = (void __iomem *)adev->rmmio + pcie_data * 4;
+
+	/* read low 32 bits */
+	writel(reg_addr, pcie_index_offset);
+	readl(pcie_index_offset);
+	r = readl(pcie_data_offset);
+	/* read high 32 bits */
+	writel(reg_addr + 4, pcie_index_offset);
+	readl(pcie_index_offset);
+	r |= ((u64)readl(pcie_data_offset) << 32);
+	spin_unlock_irqrestore(&adev->pcie_idx_lock, flags);
+
+	return r;
+}
+
+u64 amdgpu_device_indirect_rreg64_ext(struct amdgpu_device *adev,
+				  u64 reg_addr)
+{
+	unsigned long flags, pcie_index, pcie_data;
+	unsigned long pcie_index_hi = 0;
 	void __iomem *pcie_index_offset;
+	void __iomem *pcie_index_hi_offset;
 	void __iomem *pcie_data_offset;
+	u64 r;
+
+	pcie_index = adev->nbio.funcs->get_pcie_index_offset(adev);
+	pcie_data = adev->nbio.funcs->get_pcie_data_offset(adev);
+	if ((reg_addr >> 32) && (adev->nbio.funcs->get_pcie_index_hi_offset))
+		pcie_index_hi = adev->nbio.funcs->get_pcie_index_hi_offset(adev);
 
 	spin_lock_irqsave(&adev->pcie_idx_lock, flags);
 	pcie_index_offset = (void __iomem *)adev->rmmio + pcie_index * 4;
 	pcie_data_offset = (void __iomem *)adev->rmmio + pcie_data * 4;
+	if (pcie_index_hi != 0)
+		pcie_index_hi_offset = (void __iomem *)adev->rmmio +
+			pcie_index_hi * 4;
 
 	/* read low 32 bits */
 	writel(reg_addr, pcie_index_offset);
 	readl(pcie_index_offset);
+	if (pcie_index_hi != 0) {
+		writel((reg_addr >> 32) & 0xff, pcie_index_hi_offset);
+		readl(pcie_index_hi_offset);
+	}
 	r = readl(pcie_data_offset);
 	/* read high 32 bits */
 	writel(reg_addr + 4, pcie_index_offset);
 	readl(pcie_index_offset);
+	if (pcie_index_hi != 0) {
+		writel((reg_addr >> 32) & 0xff, pcie_index_hi_offset);
+		readl(pcie_index_hi_offset);
+	}
 	r |= ((u64)readl(pcie_data_offset) << 32);
+
+	/* clear the high bits */
+	if (pcie_index_hi != 0) {
+		writel(0, pcie_index_hi_offset);
+		readl(pcie_index_hi_offset);
+	}
+
 	spin_unlock_irqrestore(&adev->pcie_idx_lock, flags);
 
 	return r;
@@ -730,20 +1115,20 @@ u64 amdgpu_device_indirect_rreg64(struct amdgpu_device *adev,
  * amdgpu_device_indirect_wreg - write an indirect register address
  *
  * @adev: amdgpu_device pointer
- * @pcie_index: mmio register offset
- * @pcie_data: mmio register offset
  * @reg_addr: indirect register offset
  * @reg_data: indirect register data
  *
  */
 void amdgpu_device_indirect_wreg(struct amdgpu_device *adev,
-				 u32 pcie_index, u32 pcie_data,
 				 u32 reg_addr, u32 reg_data)
 {
-	unsigned long flags;
+	unsigned long flags, pcie_index, pcie_data;
 	void __iomem *pcie_index_offset;
 	void __iomem *pcie_data_offset;
 
+	pcie_index = adev->nbio.funcs->get_pcie_index_offset(adev);
+	pcie_data = adev->nbio.funcs->get_pcie_data_offset(adev);
+
 	spin_lock_irqsave(&adev->pcie_idx_lock, flags);
 	pcie_index_offset = (void __iomem *)adev->rmmio + pcie_index * 4;
 	pcie_data_offset = (void __iomem *)adev->rmmio + pcie_data * 4;
@@ -755,24 +1140,64 @@ void amdgpu_device_indirect_wreg(struct amdgpu_device *adev,
 	spin_unlock_irqrestore(&adev->pcie_idx_lock, flags);
 }
 
+void amdgpu_device_indirect_wreg_ext(struct amdgpu_device *adev,
+				     u64 reg_addr, u32 reg_data)
+{
+	unsigned long flags, pcie_index, pcie_index_hi, pcie_data;
+	void __iomem *pcie_index_offset;
+	void __iomem *pcie_index_hi_offset;
+	void __iomem *pcie_data_offset;
+
+	pcie_index = adev->nbio.funcs->get_pcie_index_offset(adev);
+	pcie_data = adev->nbio.funcs->get_pcie_data_offset(adev);
+	if ((reg_addr >> 32) && (adev->nbio.funcs->get_pcie_index_hi_offset))
+		pcie_index_hi = adev->nbio.funcs->get_pcie_index_hi_offset(adev);
+	else
+		pcie_index_hi = 0;
+
+	spin_lock_irqsave(&adev->pcie_idx_lock, flags);
+	pcie_index_offset = (void __iomem *)adev->rmmio + pcie_index * 4;
+	pcie_data_offset = (void __iomem *)adev->rmmio + pcie_data * 4;
+	if (pcie_index_hi != 0)
+		pcie_index_hi_offset = (void __iomem *)adev->rmmio +
+				pcie_index_hi * 4;
+
+	writel(reg_addr, pcie_index_offset);
+	readl(pcie_index_offset);
+	if (pcie_index_hi != 0) {
+		writel((reg_addr >> 32) & 0xff, pcie_index_hi_offset);
+		readl(pcie_index_hi_offset);
+	}
+	writel(reg_data, pcie_data_offset);
+	readl(pcie_data_offset);
+
+	/* clear the high bits */
+	if (pcie_index_hi != 0) {
+		writel(0, pcie_index_hi_offset);
+		readl(pcie_index_hi_offset);
+	}
+
+	spin_unlock_irqrestore(&adev->pcie_idx_lock, flags);
+}
+
 /**
  * amdgpu_device_indirect_wreg64 - write a 64bits indirect register address
  *
  * @adev: amdgpu_device pointer
- * @pcie_index: mmio register offset
- * @pcie_data: mmio register offset
  * @reg_addr: indirect register offset
  * @reg_data: indirect register data
  *
  */
 void amdgpu_device_indirect_wreg64(struct amdgpu_device *adev,
-				   u32 pcie_index, u32 pcie_data,
 				   u32 reg_addr, u64 reg_data)
 {
-	unsigned long flags;
+	unsigned long flags, pcie_index, pcie_data;
 	void __iomem *pcie_index_offset;
 	void __iomem *pcie_data_offset;
 
+	pcie_index = adev->nbio.funcs->get_pcie_index_offset(adev);
+	pcie_data = adev->nbio.funcs->get_pcie_data_offset(adev);
+
 	spin_lock_irqsave(&adev->pcie_idx_lock, flags);
 	pcie_index_offset = (void __iomem *)adev->rmmio + pcie_index * 4;
 	pcie_data_offset = (void __iomem *)adev->rmmio + pcie_data * 4;
@@ -790,6 +1215,67 @@ void amdgpu_device_indirect_wreg64(struct amdgpu_device *adev,
 	spin_unlock_irqrestore(&adev->pcie_idx_lock, flags);
 }
 
+void amdgpu_device_indirect_wreg64_ext(struct amdgpu_device *adev,
+				   u64 reg_addr, u64 reg_data)
+{
+	unsigned long flags, pcie_index, pcie_data;
+	unsigned long pcie_index_hi = 0;
+	void __iomem *pcie_index_offset;
+	void __iomem *pcie_index_hi_offset;
+	void __iomem *pcie_data_offset;
+
+	pcie_index = adev->nbio.funcs->get_pcie_index_offset(adev);
+	pcie_data = adev->nbio.funcs->get_pcie_data_offset(adev);
+	if ((reg_addr >> 32) && (adev->nbio.funcs->get_pcie_index_hi_offset))
+		pcie_index_hi = adev->nbio.funcs->get_pcie_index_hi_offset(adev);
+
+	spin_lock_irqsave(&adev->pcie_idx_lock, flags);
+	pcie_index_offset = (void __iomem *)adev->rmmio + pcie_index * 4;
+	pcie_data_offset = (void __iomem *)adev->rmmio + pcie_data * 4;
+	if (pcie_index_hi != 0)
+		pcie_index_hi_offset = (void __iomem *)adev->rmmio +
+				pcie_index_hi * 4;
+
+	/* write low 32 bits */
+	writel(reg_addr, pcie_index_offset);
+	readl(pcie_index_offset);
+	if (pcie_index_hi != 0) {
+		writel((reg_addr >> 32) & 0xff, pcie_index_hi_offset);
+		readl(pcie_index_hi_offset);
+	}
+	writel((u32)(reg_data & 0xffffffffULL), pcie_data_offset);
+	readl(pcie_data_offset);
+	/* write high 32 bits */
+	writel(reg_addr + 4, pcie_index_offset);
+	readl(pcie_index_offset);
+	if (pcie_index_hi != 0) {
+		writel((reg_addr >> 32) & 0xff, pcie_index_hi_offset);
+		readl(pcie_index_hi_offset);
+	}
+	writel((u32)(reg_data >> 32), pcie_data_offset);
+	readl(pcie_data_offset);
+
+	/* clear the high bits */
+	if (pcie_index_hi != 0) {
+		writel(0, pcie_index_hi_offset);
+		readl(pcie_index_hi_offset);
+	}
+
+	spin_unlock_irqrestore(&adev->pcie_idx_lock, flags);
+}
+
+/**
+ * amdgpu_device_get_rev_id - query device rev_id
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Return device rev_id
+ */
+u32 amdgpu_device_get_rev_id(struct amdgpu_device *adev)
+{
+	return adev->nbio.funcs->get_rev_id(adev);
+}
+
 /**
  * amdgpu_invalid_rreg - dummy reg read function
  *
@@ -802,7 +1288,14 @@ void amdgpu_device_indirect_wreg64(struct amdgpu_device *adev,
  */
 static uint32_t amdgpu_invalid_rreg(struct amdgpu_device *adev, uint32_t reg)
 {
-	DRM_ERROR("Invalid callback to read register 0x%04X\n", reg);
+	dev_err(adev->dev, "Invalid callback to read register 0x%04X\n", reg);
+	BUG();
+	return 0;
+}
+
+static uint32_t amdgpu_invalid_rreg_ext(struct amdgpu_device *adev, uint64_t reg)
+{
+	dev_err(adev->dev, "Invalid callback to read register 0x%llX\n", reg);
 	BUG();
 	return 0;
 }
@@ -819,8 +1312,17 @@ static uint32_t amdgpu_invalid_rreg(struct amdgpu_device *adev, uint32_t reg)
  */
 static void amdgpu_invalid_wreg(struct amdgpu_device *adev, uint32_t reg, uint32_t v)
 {
-	DRM_ERROR("Invalid callback to write register 0x%04X with 0x%08X\n",
-		  reg, v);
+	dev_err(adev->dev,
+		"Invalid callback to write register 0x%04X with 0x%08X\n", reg,
+		v);
+	BUG();
+}
+
+static void amdgpu_invalid_wreg_ext(struct amdgpu_device *adev, uint64_t reg, uint32_t v)
+{
+	dev_err(adev->dev,
+		"Invalid callback to write register 0x%llX with 0x%08X\n", reg,
+		v);
 	BUG();
 }
 
@@ -836,7 +1338,15 @@ static void amdgpu_invalid_wreg(struct amdgpu_device *adev, uint32_t reg, uint32
  */
 static uint64_t amdgpu_invalid_rreg64(struct amdgpu_device *adev, uint32_t reg)
 {
-	DRM_ERROR("Invalid callback to read 64 bit register 0x%04X\n", reg);
+	dev_err(adev->dev, "Invalid callback to read 64 bit register 0x%04X\n",
+		reg);
+	BUG();
+	return 0;
+}
+
+static uint64_t amdgpu_invalid_rreg64_ext(struct amdgpu_device *adev, uint64_t reg)
+{
+	dev_err(adev->dev, "Invalid callback to read register 0x%llX\n", reg);
 	BUG();
 	return 0;
 }
@@ -853,8 +1363,17 @@ static uint64_t amdgpu_invalid_rreg64(struct amdgpu_device *adev, uint32_t reg)
  */
 static void amdgpu_invalid_wreg64(struct amdgpu_device *adev, uint32_t reg, uint64_t v)
 {
-	DRM_ERROR("Invalid callback to write 64 bit register 0x%04X with 0x%08llX\n",
-		  reg, v);
+	dev_err(adev->dev,
+		"Invalid callback to write 64 bit register 0x%04X with 0x%08llX\n",
+		reg, v);
+	BUG();
+}
+
+static void amdgpu_invalid_wreg64_ext(struct amdgpu_device *adev, uint64_t reg, uint64_t v)
+{
+	dev_err(adev->dev,
+		"Invalid callback to write 64 bit register 0x%llX with 0x%08llX\n",
+		reg, v);
 	BUG();
 }
 
@@ -872,8 +1391,9 @@ static void amdgpu_invalid_wreg64(struct amdgpu_device *adev, uint32_t reg, uint
 static uint32_t amdgpu_block_invalid_rreg(struct amdgpu_device *adev,
 					  uint32_t block, uint32_t reg)
 {
-	DRM_ERROR("Invalid callback to read register 0x%04X in block 0x%04X\n",
-		  reg, block);
+	dev_err(adev->dev,
+		"Invalid callback to read register 0x%04X in block 0x%04X\n",
+		reg, block);
 	BUG();
 	return 0;
 }
@@ -893,11 +1413,23 @@ static void amdgpu_block_invalid_wreg(struct amdgpu_device *adev,
 				      uint32_t block,
 				      uint32_t reg, uint32_t v)
 {
-	DRM_ERROR("Invalid block callback to write register 0x%04X in block 0x%04X with 0x%08X\n",
-		  reg, block, v);
+	dev_err(adev->dev,
+		"Invalid block callback to write register 0x%04X in block 0x%04X with 0x%08X\n",
+		reg, block, v);
 	BUG();
 }
 
+static uint32_t amdgpu_device_get_vbios_flags(struct amdgpu_device *adev)
+{
+	if (hweight32(adev->aid_mask) && (adev->flags & AMD_IS_APU))
+		return AMDGPU_VBIOS_SKIP;
+
+	if (hweight32(adev->aid_mask) && amdgpu_passthrough(adev))
+		return AMDGPU_VBIOS_OPTIONAL;
+
+	return 0;
+}
+
 /**
  * amdgpu_device_asic_init - Wrapper for atom asic_init
  *
@@ -907,38 +1439,62 @@ static void amdgpu_block_invalid_wreg(struct amdgpu_device *adev,
  */
 static int amdgpu_device_asic_init(struct amdgpu_device *adev)
 {
+	uint32_t flags;
+	bool optional;
+	int ret;
+
 	amdgpu_asic_pre_asic_init(adev);
+	flags = amdgpu_device_get_vbios_flags(adev);
+	optional = !!(flags & (AMDGPU_VBIOS_OPTIONAL | AMDGPU_VBIOS_SKIP));
+
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3) ||
+	    amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4) ||
+	    amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 5, 0) ||
+	    amdgpu_ip_version(adev, GC_HWIP, 0) >= IP_VERSION(11, 0, 0)) {
+		amdgpu_psp_wait_for_bootloader(adev);
+		if (optional && !adev->bios)
+			return 0;
 
-	return amdgpu_atom_asic_init(adev->mode_info.atom_context);
+		ret = amdgpu_atomfirmware_asic_init(adev, true);
+		return ret;
+	} else {
+		if (optional && !adev->bios)
+			return 0;
+
+		return amdgpu_atom_asic_init(adev->mode_info.atom_context);
+	}
+
+	return 0;
 }
 
 /**
- * amdgpu_device_vram_scratch_init - allocate the VRAM scratch page
+ * amdgpu_device_mem_scratch_init - allocate the VRAM scratch page
  *
  * @adev: amdgpu_device pointer
  *
  * Allocates a scratch page of VRAM for use by various things in the
  * driver.
  */
-static int amdgpu_device_vram_scratch_init(struct amdgpu_device *adev)
+static int amdgpu_device_mem_scratch_init(struct amdgpu_device *adev)
 {
-	return amdgpu_bo_create_kernel(adev, AMDGPU_GPU_PAGE_SIZE,
-				       PAGE_SIZE, AMDGPU_GEM_DOMAIN_VRAM,
-				       &adev->vram_scratch.robj,
-				       &adev->vram_scratch.gpu_addr,
-				       (void **)&adev->vram_scratch.ptr);
+	return amdgpu_bo_create_kernel(adev, AMDGPU_GPU_PAGE_SIZE, PAGE_SIZE,
+				       AMDGPU_GEM_DOMAIN_VRAM |
+				       AMDGPU_GEM_DOMAIN_GTT,
+				       &adev->mem_scratch.robj,
+				       &adev->mem_scratch.gpu_addr,
+				       (void **)&adev->mem_scratch.ptr);
 }
 
 /**
- * amdgpu_device_vram_scratch_fini - Free the VRAM scratch page
+ * amdgpu_device_mem_scratch_fini - Free the VRAM scratch page
  *
  * @adev: amdgpu_device pointer
  *
  * Frees the VRAM scratch page.
  */
-static void amdgpu_device_vram_scratch_fini(struct amdgpu_device *adev)
+static void amdgpu_device_mem_scratch_fini(struct amdgpu_device *adev)
 {
-	amdgpu_bo_free_kernel(&adev->vram_scratch.robj, NULL, NULL);
+	amdgpu_bo_free_kernel(&adev->mem_scratch.robj, NULL, NULL);
 }
 
 /**
@@ -948,7 +1504,7 @@ static void amdgpu_device_vram_scratch_fini(struct amdgpu_device *adev)
  * @registers: pointer to the register array
  * @array_size: size of the register array
  *
- * Programs an array or registers with and and or masks.
+ * Programs an array or registers with and or masks.
  * This is a helper for setting golden registers.
  */
 void amdgpu_device_program_register_sequence(struct amdgpu_device *adev,
@@ -961,7 +1517,7 @@ void amdgpu_device_program_register_sequence(struct amdgpu_device *adev,
 	if (array_size % 3)
 		return;
 
-	for (i = 0; i < array_size; i +=3) {
+	for (i = 0; i < array_size; i += 3) {
 		reg = registers[i + 0];
 		and_mask = registers[i + 1];
 		or_mask = registers[i + 2];
@@ -1006,76 +1562,6 @@ int amdgpu_device_pci_reset(struct amdgpu_device *adev)
 }
 
 /*
- * GPU doorbell aperture helpers function.
- */
-/**
- * amdgpu_device_doorbell_init - Init doorbell driver information.
- *
- * @adev: amdgpu_device pointer
- *
- * Init doorbell driver information (CIK)
- * Returns 0 on success, error on failure.
- */
-static int amdgpu_device_doorbell_init(struct amdgpu_device *adev)
-{
-
-	/* No doorbell on SI hardware generation */
-	if (adev->asic_type < CHIP_BONAIRE) {
-		adev->doorbell.base = 0;
-		adev->doorbell.size = 0;
-		adev->doorbell.num_doorbells = 0;
-		adev->doorbell.ptr = NULL;
-		return 0;
-	}
-
-	if (pci_resource_flags(adev->pdev, 2) & IORESOURCE_UNSET)
-		return -EINVAL;
-
-	amdgpu_asic_init_doorbell_index(adev);
-
-	/* doorbell bar mapping */
-	adev->doorbell.base = pci_resource_start(adev->pdev, 2);
-	adev->doorbell.size = pci_resource_len(adev->pdev, 2);
-
-	adev->doorbell.num_doorbells = min_t(u32, adev->doorbell.size / sizeof(u32),
-					     adev->doorbell_index.max_assignment+1);
-	if (adev->doorbell.num_doorbells == 0)
-		return -EINVAL;
-
-	/* For Vega, reserve and map two pages on doorbell BAR since SDMA
-	 * paging queue doorbell use the second page. The
-	 * AMDGPU_DOORBELL64_MAX_ASSIGNMENT definition assumes all the
-	 * doorbells are in the first page. So with paging queue enabled,
-	 * the max num_doorbells should + 1 page (0x400 in dword)
-	 */
-	if (adev->asic_type >= CHIP_VEGA10)
-		adev->doorbell.num_doorbells += 0x400;
-
-	adev->doorbell.ptr = ioremap(adev->doorbell.base,
-				     adev->doorbell.num_doorbells *
-				     sizeof(u32));
-	if (adev->doorbell.ptr == NULL)
-		return -ENOMEM;
-
-	return 0;
-}
-
-/**
- * amdgpu_device_doorbell_fini - Tear down doorbell driver information.
- *
- * @adev: amdgpu_device pointer
- *
- * Tear down doorbell driver information (CIK)
- */
-static void amdgpu_device_doorbell_fini(struct amdgpu_device *adev)
-{
-	iounmap(adev->doorbell.ptr);
-	adev->doorbell.ptr = NULL;
-}
-
-
-
-/*
  * amdgpu_device_wb_*()
  * Writeback is the method by which the GPU updates special pages in memory
  * with the status of certain GPU events (fences, ring pointers,etc.).
@@ -1100,7 +1586,7 @@ static void amdgpu_device_wb_fini(struct amdgpu_device *adev)
 }
 
 /**
- * amdgpu_device_wb_init- Init Writeback driver info and allocate memory
+ * amdgpu_device_wb_init - Init Writeback driver info and allocate memory
  *
  * @adev: amdgpu_device pointer
  *
@@ -1144,13 +1630,17 @@ static int amdgpu_device_wb_init(struct amdgpu_device *adev)
  */
 int amdgpu_device_wb_get(struct amdgpu_device *adev, u32 *wb)
 {
-	unsigned long offset = find_first_zero_bit(adev->wb.used, adev->wb.num_wb);
+	unsigned long flags, offset;
 
+	spin_lock_irqsave(&adev->wb.lock, flags);
+	offset = find_first_zero_bit(adev->wb.used, adev->wb.num_wb);
 	if (offset < adev->wb.num_wb) {
 		__set_bit(offset, adev->wb.used);
+		spin_unlock_irqrestore(&adev->wb.lock, flags);
 		*wb = offset << 3; /* convert to dw offset */
 		return 0;
 	} else {
+		spin_unlock_irqrestore(&adev->wb.lock, flags);
 		return -EINVAL;
 	}
 }
@@ -1165,9 +1655,13 @@ int amdgpu_device_wb_get(struct amdgpu_device *adev, u32 *wb)
  */
 void amdgpu_device_wb_free(struct amdgpu_device *adev, u32 wb)
 {
+	unsigned long flags;
+
 	wb >>= 3;
+	spin_lock_irqsave(&adev->wb.lock, flags);
 	if (wb < adev->wb.num_wb)
 		__clear_bit(wb, adev->wb.used);
+	spin_unlock_irqrestore(&adev->wb.lock, flags);
 }
 
 /**
@@ -1184,14 +1678,33 @@ int amdgpu_device_resize_fb_bar(struct amdgpu_device *adev)
 	int rbar_size = pci_rebar_bytes_to_size(adev->gmc.real_vram_size);
 	struct pci_bus *root;
 	struct resource *res;
-	unsigned i;
+	int max_size, r;
+	unsigned int i;
 	u16 cmd;
-	int r;
+
+	if (!IS_ENABLED(CONFIG_PHYS_ADDR_T_64BIT))
+		return 0;
 
 	/* Bypass for VF */
 	if (amdgpu_sriov_vf(adev))
 		return 0;
 
+	if (!amdgpu_rebar)
+		return 0;
+
+	/* resizing on Dell G5 SE platforms causes problems with runtime pm */
+	if ((amdgpu_runtime_pm != 0) &&
+	    adev->pdev->vendor == PCI_VENDOR_ID_ATI &&
+	    adev->pdev->device == 0x731f &&
+	    adev->pdev->subsystem_vendor == PCI_VENDOR_ID_DELL)
+		return 0;
+
+	/* PCI_EXT_CAP_ID_VNDR extended capability is located at 0x100 */
+	if (!pci_find_ext_capability(adev->pdev, PCI_EXT_CAP_ID_VNDR))
+		dev_warn(
+			adev->dev,
+			"System can't access extended configuration space, please check!!\n");
+
 	/* skip if the bios has already enabled large BAR */
 	if (adev->gmc.real_vram_size &&
 	    (pci_resource_len(adev->pdev, 0) >= adev->gmc.real_vram_size))
@@ -1213,33 +1726,32 @@ int amdgpu_device_resize_fb_bar(struct amdgpu_device *adev)
 		return 0;
 
 	/* Limit the BAR size to what is available */
-	rbar_size = min(fls(pci_rebar_get_possible_sizes(adev->pdev, 0)) - 1,
-			rbar_size);
+	max_size = pci_rebar_get_max_size(adev->pdev, 0);
+	if (max_size < 0)
+		return 0;
+	rbar_size = min(max_size, rbar_size);
 
 	/* Disable memory decoding while we change the BAR addresses and size */
 	pci_read_config_word(adev->pdev, PCI_COMMAND, &cmd);
 	pci_write_config_word(adev->pdev, PCI_COMMAND,
 			      cmd & ~PCI_COMMAND_MEMORY);
 
-	/* Free the VRAM and doorbell BAR, we most likely need to move both. */
-	amdgpu_device_doorbell_fini(adev);
-	if (adev->asic_type >= CHIP_BONAIRE)
-		pci_release_resource(adev->pdev, 2);
-
-	pci_release_resource(adev->pdev, 0);
+	/* Tear down doorbell as resizing will release BARs */
+	amdgpu_doorbell_fini(adev);
 
-	r = pci_resize_resource(adev->pdev, 0, rbar_size);
+	r = pci_resize_resource(adev->pdev, 0, rbar_size,
+				(adev->asic_type >= CHIP_BONAIRE) ? 1 << 5
+								  : 1 << 2);
 	if (r == -ENOSPC)
-		DRM_INFO("Not enough PCI address space for a large BAR.");
+		dev_info(adev->dev,
+			 "Not enough PCI address space for a large BAR.");
 	else if (r && r != -ENOTSUPP)
-		DRM_ERROR("Problem resizing BAR0 (%d).", r);
-
-	pci_assign_unassigned_bus_resources(adev->pdev->bus);
+		dev_err(adev->dev, "Problem resizing BAR0 (%d).", r);
 
 	/* When the doorbell or fb BAR isn't available we have no chance of
 	 * using the device.
 	 */
-	r = amdgpu_device_doorbell_init(adev);
+	r = amdgpu_doorbell_init(adev);
 	if (r || (pci_resource_flags(adev->pdev, 0) & IORESOURCE_UNSET))
 		return -ENODEV;
 
@@ -1262,11 +1774,17 @@ int amdgpu_device_resize_fb_bar(struct amdgpu_device *adev)
  */
 bool amdgpu_device_need_post(struct amdgpu_device *adev)
 {
-	uint32_t reg;
+	uint32_t reg, flags;
 
 	if (amdgpu_sriov_vf(adev))
 		return false;
 
+	flags = amdgpu_device_get_vbios_flags(adev);
+	if (flags & AMDGPU_VBIOS_SKIP)
+		return false;
+	if ((flags & AMDGPU_VBIOS_OPTIONAL) && !adev->bios)
+		return false;
+
 	if (amdgpu_passthrough(adev)) {
 		/* for FIJI: In whole GPU pass-through virtualization case, after VM reboot
 		 * some old smc fw still need driver do vPost otherwise gpu hang, while
@@ -1276,19 +1794,21 @@ bool amdgpu_device_need_post(struct amdgpu_device *adev)
 		if (adev->asic_type == CHIP_FIJI) {
 			int err;
 			uint32_t fw_ver;
+
 			err = request_firmware(&adev->pm.fw, "amdgpu/fiji_smc.bin", adev->dev);
-			/* force vPost if error occured */
+			/* force vPost if error occurred */
 			if (err)
 				return true;
 
 			fw_ver = *((uint32_t *)adev->pm.fw->data + 69);
+			release_firmware(adev->pm.fw);
 			if (fw_ver < 0x00160e00)
 				return true;
 		}
 	}
 
 	/* Don't post if we need to reset whole hive on init */
-	if (adev->gmc.xgmi.pending_reset)
+	if (adev->init_lvl->level == AMDGPU_INIT_LEVEL_MINIMAL_XGMI)
 		return false;
 
 	if (adev->has_hw_reset) {
@@ -1309,6 +1829,125 @@ bool amdgpu_device_need_post(struct amdgpu_device *adev)
 	return true;
 }
 
+/*
+ * Check whether seamless boot is supported.
+ *
+ * So far we only support seamless boot on DCE 3.0 or later.
+ * If users report that it works on older ASICS as well, we may
+ * loosen this.
+ */
+bool amdgpu_device_seamless_boot_supported(struct amdgpu_device *adev)
+{
+	switch (amdgpu_seamless) {
+	case -1:
+		break;
+	case 1:
+		return true;
+	case 0:
+		return false;
+	default:
+		dev_err(adev->dev, "Invalid value for amdgpu.seamless: %d\n",
+			amdgpu_seamless);
+		return false;
+	}
+
+	if (!(adev->flags & AMD_IS_APU))
+		return false;
+
+	if (adev->mman.keep_stolen_vga_memory)
+		return false;
+
+	return amdgpu_ip_version(adev, DCE_HWIP, 0) >= IP_VERSION(3, 0, 0);
+}
+
+/*
+ * Intel hosts such as Rocket Lake, Alder Lake, Raptor Lake and Sapphire Rapids
+ * don't support dynamic speed switching. Until we have confirmation from Intel
+ * that a specific host supports it, it's safer that we keep it disabled for all.
+ *
+ * https://edc.intel.com/content/www/us/en/design/products/platforms/details/raptor-lake-s/13th-generation-core-processors-datasheet-volume-1-of-2/005/pci-express-support/
+ * https://gitlab.freedesktop.org/drm/amd/-/issues/2663
+ */
+static bool amdgpu_device_pcie_dynamic_switching_supported(struct amdgpu_device *adev)
+{
+#if IS_ENABLED(CONFIG_X86)
+	struct cpuinfo_x86 *c = &cpu_data(0);
+
+	/* eGPU change speeds based on USB4 fabric conditions */
+	if (dev_is_removable(adev->dev))
+		return true;
+
+	if (c->x86_vendor == X86_VENDOR_INTEL)
+		return false;
+#endif
+	return true;
+}
+
+static bool amdgpu_device_aspm_support_quirk(struct amdgpu_device *adev)
+{
+	/* Enabling ASPM causes randoms hangs on Tahiti and Oland on Zen4.
+	 * It's unclear if this is a platform-specific or GPU-specific issue.
+	 * Disable ASPM on SI for the time being.
+	 */
+	if (adev->family == AMDGPU_FAMILY_SI)
+		return true;
+
+#if IS_ENABLED(CONFIG_X86)
+	struct cpuinfo_x86 *c = &cpu_data(0);
+
+	if (!(amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(12, 0, 0) ||
+		  amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(12, 0, 1)))
+		return false;
+
+	if (c->x86 == 6 &&
+		adev->pm.pcie_gen_mask & CAIL_PCIE_LINK_SPEED_SUPPORT_GEN5) {
+		switch (c->x86_model) {
+		case VFM_MODEL(INTEL_ALDERLAKE):
+		case VFM_MODEL(INTEL_ALDERLAKE_L):
+		case VFM_MODEL(INTEL_RAPTORLAKE):
+		case VFM_MODEL(INTEL_RAPTORLAKE_P):
+		case VFM_MODEL(INTEL_RAPTORLAKE_S):
+			return true;
+		default:
+			return false;
+		}
+	} else {
+		return false;
+	}
+#else
+	return false;
+#endif
+}
+
+/**
+ * amdgpu_device_should_use_aspm - check if the device should program ASPM
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Confirm whether the module parameter and pcie bridge agree that ASPM should
+ * be set for this device.
+ *
+ * Returns true if it should be used or false if not.
+ */
+bool amdgpu_device_should_use_aspm(struct amdgpu_device *adev)
+{
+	switch (amdgpu_aspm) {
+	case -1:
+		break;
+	case 0:
+		return false;
+	case 1:
+		return true;
+	default:
+		return false;
+	}
+	if (adev->flags & AMD_IS_APU)
+		return false;
+	if (amdgpu_device_aspm_support_quirk(adev))
+		return false;
+	return pcie_aspm_enabled(adev->pdev);
+}
+
 /* if we get transitioned to only one device, take VGA back */
 /**
  * amdgpu_device_vga_set_decode - enable/disable vga decode
@@ -1323,6 +1962,7 @@ static unsigned int amdgpu_device_vga_set_decode(struct pci_dev *pdev,
 		bool state)
 {
 	struct amdgpu_device *adev = drm_to_adev(pci_get_drvdata(pdev));
+
 	amdgpu_asic_set_vga_state(adev, state);
 	if (state)
 		return VGA_RSRC_LEGACY_IO | VGA_RSRC_LEGACY_MEM |
@@ -1345,7 +1985,8 @@ static void amdgpu_device_check_block_size(struct amdgpu_device *adev)
 {
 	/* defines number of bits in page table versus page directory,
 	 * a page is 4KB so we have 12 bits offset, minimum 9 bits in the
-	 * page table and the remaining bits are in the page directory */
+	 * page table and the remaining bits are in the page directory
+	 */
 	if (amdgpu_vm_block_size == -1)
 		return;
 
@@ -1389,7 +2030,7 @@ static void amdgpu_device_check_smu_prv_buffer_size(struct amdgpu_device *adev)
 		return;
 
 	if (!is_os_64) {
-		DRM_WARN("Not 64-bit OS, feature not supported\n");
+		dev_warn(adev->dev, "Not 64-bit OS, feature not supported\n");
 		goto def_value;
 	}
 	si_meminfo(&si);
@@ -1404,7 +2045,7 @@ static void amdgpu_device_check_smu_prv_buffer_size(struct amdgpu_device *adev)
 		if (total_memory < dram_size_seven_GB)
 			goto def_value1;
 	} else {
-		DRM_WARN("Smu memory pool size not supported\n");
+		dev_warn(adev->dev, "Smu memory pool size not supported\n");
 		goto def_value;
 	}
 	adev->pm.smu_prv_buffer_size = amdgpu_smu_memory_pool_size << 28;
@@ -1412,7 +2053,7 @@ static void amdgpu_device_check_smu_prv_buffer_size(struct amdgpu_device *adev)
 	return;
 
 def_value1:
-	DRM_WARN("No enough system memory\n");
+	dev_warn(adev->dev, "No enough system memory\n");
 def_value:
 	adev->pm.smu_prv_buffer_size = 0;
 }
@@ -1443,11 +2084,12 @@ static int amdgpu_device_init_apu_flags(struct amdgpu_device *adev)
 	case CHIP_YELLOW_CARP:
 		break;
 	case CHIP_CYAN_SKILLFISH:
-		if (adev->pdev->device == 0x13FE)
+		if ((adev->pdev->device == 0x13FE) ||
+		    (adev->pdev->device == 0x143F))
 			adev->apu_flags |= AMD_APU_IS_CYAN_SKILLFISH2;
 		break;
 	default:
-		return -EINVAL;
+		break;
 	}
 
 	return 0;
@@ -1463,11 +2105,13 @@ static int amdgpu_device_init_apu_flags(struct amdgpu_device *adev)
  */
 static int amdgpu_device_check_arguments(struct amdgpu_device *adev)
 {
+	int i;
+
 	if (amdgpu_sched_jobs < 4) {
 		dev_warn(adev->dev, "sched jobs (%d) must be at least 4\n",
 			 amdgpu_sched_jobs);
 		amdgpu_sched_jobs = 4;
-	} else if (!is_power_of_2(amdgpu_sched_jobs)){
+	} else if (!is_power_of_2(amdgpu_sched_jobs)) {
 		dev_warn(adev->dev, "sched jobs (%d) must be a power of 2\n",
 			 amdgpu_sched_jobs);
 		amdgpu_sched_jobs = roundup_pow_of_two(amdgpu_sched_jobs);
@@ -1504,6 +2148,11 @@ static int amdgpu_device_check_arguments(struct amdgpu_device *adev)
 		amdgpu_sched_hw_submission = roundup_pow_of_two(amdgpu_sched_hw_submission);
 	}
 
+	if (amdgpu_reset_method < -1 || amdgpu_reset_method > 4) {
+		dev_warn(adev->dev, "invalid option for reset method, reverting to default\n");
+		amdgpu_reset_method = -1;
+	}
+
 	amdgpu_device_check_smu_prv_buffer_size(adev);
 
 	amdgpu_device_check_vm_size(adev);
@@ -1512,9 +2161,31 @@ static int amdgpu_device_check_arguments(struct amdgpu_device *adev)
 
 	adev->firmware.load_type = amdgpu_ucode_get_load_type(adev, amdgpu_fw_load_type);
 
-	amdgpu_gmc_tmz_set(adev);
-
-	amdgpu_gmc_noretry_set(adev);
+	for (i = 0; i < MAX_XCP; i++) {
+		switch (amdgpu_enforce_isolation) {
+		case -1:
+		case 0:
+		default:
+			/* disable */
+			adev->enforce_isolation[i] = AMDGPU_ENFORCE_ISOLATION_DISABLE;
+			break;
+		case 1:
+			/* enable */
+			adev->enforce_isolation[i] =
+				AMDGPU_ENFORCE_ISOLATION_ENABLE;
+			break;
+		case 2:
+			/* enable legacy mode */
+			adev->enforce_isolation[i] =
+				AMDGPU_ENFORCE_ISOLATION_ENABLE_LEGACY;
+			break;
+		case 3:
+			/* enable only process isolation without submitting cleaner shader */
+			adev->enforce_isolation[i] =
+				AMDGPU_ENFORCE_ISOLATION_NO_CLEANER_SHADER;
+			break;
+		}
+	}
 
 	return 0;
 }
@@ -1525,7 +2196,7 @@ static int amdgpu_device_check_arguments(struct amdgpu_device *adev)
  * @pdev: pci dev pointer
  * @state: vga_switcheroo state
  *
- * Callback for the switcheroo driver.  Suspends or resumes the
+ * Callback for the switcheroo driver.  Suspends or resumes
  * the asics before or after it is powered up using ACPI methods.
  */
 static void amdgpu_switcheroo_set_state(struct pci_dev *pdev,
@@ -1534,7 +2205,8 @@ static void amdgpu_switcheroo_set_state(struct pci_dev *pdev,
 	struct drm_device *dev = pci_get_drvdata(pdev);
 	int r;
 
-	if (amdgpu_device_supports_px(dev) && state == VGA_SWITCHEROO_OFF)
+	if (amdgpu_device_supports_px(drm_to_adev(dev)) &&
+	    state == VGA_SWITCHEROO_OFF)
 		return;
 
 	if (state == VGA_SWITCHEROO_ON) {
@@ -1546,13 +2218,15 @@ static void amdgpu_switcheroo_set_state(struct pci_dev *pdev,
 		amdgpu_device_load_pci_state(pdev);
 		r = pci_enable_device(pdev);
 		if (r)
-			DRM_WARN("pci_enable_device failed (%d)\n", r);
+			dev_warn(&pdev->dev, "pci_enable_device failed (%d)\n",
+				 r);
 		amdgpu_device_resume(dev, true);
 
 		dev->switch_power_state = DRM_SWITCH_POWER_ON;
 	} else {
-		pr_info("switched off\n");
+		dev_info(&pdev->dev, "switched off\n");
 		dev->switch_power_state = DRM_SWITCH_POWER_CHANGING;
+		amdgpu_device_prepare(dev);
 		amdgpu_device_suspend(dev, true);
 		amdgpu_device_cache_pci_state(pdev);
 		/* Shut down the device */
@@ -1575,7 +2249,7 @@ static bool amdgpu_switcheroo_can_switch(struct pci_dev *pdev)
 {
 	struct drm_device *dev = pci_get_drvdata(pdev);
 
-	/*
+       /*
 	* FIXME: open_count is protected by drm_global_mutex but that would lead to
 	* locking inversion with the driver load path. And the access here is
 	* completely racy anyway. So don't bother with locking for now.
@@ -1615,10 +2289,11 @@ int amdgpu_device_ip_set_clockgating_state(void *dev,
 		if (!adev->ip_blocks[i].version->funcs->set_clockgating_state)
 			continue;
 		r = adev->ip_blocks[i].version->funcs->set_clockgating_state(
-			(void *)adev, state);
+			&adev->ip_blocks[i], state);
 		if (r)
-			DRM_ERROR("set_clockgating_state of IP block <%s> failed %d\n",
-				  adev->ip_blocks[i].version->funcs->name, r);
+			dev_err(adev->dev,
+				"set_clockgating_state of IP block <%s> failed %d\n",
+				adev->ip_blocks[i].version->funcs->name, r);
 	}
 	return r;
 }
@@ -1649,10 +2324,11 @@ int amdgpu_device_ip_set_powergating_state(void *dev,
 		if (!adev->ip_blocks[i].version->funcs->set_powergating_state)
 			continue;
 		r = adev->ip_blocks[i].version->funcs->set_powergating_state(
-			(void *)adev, state);
+			&adev->ip_blocks[i], state);
 		if (r)
-			DRM_ERROR("set_powergating_state of IP block <%s> failed %d\n",
-				  adev->ip_blocks[i].version->funcs->name, r);
+			dev_err(adev->dev,
+				"set_powergating_state of IP block <%s> failed %d\n",
+				adev->ip_blocks[i].version->funcs->name, r);
 	}
 	return r;
 }
@@ -1669,7 +2345,7 @@ int amdgpu_device_ip_set_powergating_state(void *dev,
  * clockgating is enabled.
  */
 void amdgpu_device_ip_get_clockgating_state(struct amdgpu_device *adev,
-					    u32 *flags)
+					    u64 *flags)
 {
 	int i;
 
@@ -1677,7 +2353,8 @@ void amdgpu_device_ip_get_clockgating_state(struct amdgpu_device *adev,
 		if (!adev->ip_blocks[i].status.valid)
 			continue;
 		if (adev->ip_blocks[i].version->funcs->get_clockgating_state)
-			adev->ip_blocks[i].version->funcs->get_clockgating_state((void *)adev, flags);
+			adev->ip_blocks[i].version->funcs->get_clockgating_state(
+				&adev->ip_blocks[i], flags);
 	}
 }
 
@@ -1699,9 +2376,12 @@ int amdgpu_device_ip_wait_for_idle(struct amdgpu_device *adev,
 		if (!adev->ip_blocks[i].status.valid)
 			continue;
 		if (adev->ip_blocks[i].version->type == block_type) {
-			r = adev->ip_blocks[i].version->funcs->wait_for_idle((void *)adev);
-			if (r)
-				return r;
+			if (adev->ip_blocks[i].version->funcs->wait_for_idle) {
+				r = adev->ip_blocks[i].version->funcs->wait_for_idle(
+								&adev->ip_blocks[i]);
+				if (r)
+					return r;
+			}
 			break;
 		}
 	}
@@ -1710,26 +2390,45 @@ int amdgpu_device_ip_wait_for_idle(struct amdgpu_device *adev,
 }
 
 /**
- * amdgpu_device_ip_is_idle - is the hardware IP idle
+ * amdgpu_device_ip_is_hw - is the hardware IP enabled
  *
  * @adev: amdgpu_device pointer
  * @block_type: Type of hardware IP (SMU, GFX, UVD, etc.)
  *
- * Check if the hardware IP is idle or not.
- * Returns true if it the IP is idle, false if not.
+ * Check if the hardware IP is enable or not.
+ * Returns true if it the IP is enable, false if not.
  */
-bool amdgpu_device_ip_is_idle(struct amdgpu_device *adev,
-			      enum amd_ip_block_type block_type)
+bool amdgpu_device_ip_is_hw(struct amdgpu_device *adev,
+			    enum amd_ip_block_type block_type)
 {
 	int i;
 
 	for (i = 0; i < adev->num_ip_blocks; i++) {
-		if (!adev->ip_blocks[i].status.valid)
-			continue;
 		if (adev->ip_blocks[i].version->type == block_type)
-			return adev->ip_blocks[i].version->funcs->is_idle((void *)adev);
+			return adev->ip_blocks[i].status.hw;
 	}
-	return true;
+	return false;
+}
+
+/**
+ * amdgpu_device_ip_is_valid - is the hardware IP valid
+ *
+ * @adev: amdgpu_device pointer
+ * @block_type: Type of hardware IP (SMU, GFX, UVD, etc.)
+ *
+ * Check if the hardware IP is valid or not.
+ * Returns true if it the IP is valid, false if not.
+ */
+bool amdgpu_device_ip_is_valid(struct amdgpu_device *adev,
+			       enum amd_ip_block_type block_type)
+{
+	int i;
+
+	for (i = 0; i < adev->num_ip_blocks; i++) {
+		if (adev->ip_blocks[i].version->type == block_type)
+			return adev->ip_blocks[i].status.valid;
+	}
+	return false;
 
 }
 
@@ -1780,6 +2479,34 @@ int amdgpu_device_ip_block_version_cmp(struct amdgpu_device *adev,
 	return 1;
 }
 
+static const char *ip_block_names[] = {
+	[AMD_IP_BLOCK_TYPE_COMMON] = "common",
+	[AMD_IP_BLOCK_TYPE_GMC] = "gmc",
+	[AMD_IP_BLOCK_TYPE_IH] = "ih",
+	[AMD_IP_BLOCK_TYPE_SMC] = "smu",
+	[AMD_IP_BLOCK_TYPE_PSP] = "psp",
+	[AMD_IP_BLOCK_TYPE_DCE] = "dce",
+	[AMD_IP_BLOCK_TYPE_GFX] = "gfx",
+	[AMD_IP_BLOCK_TYPE_SDMA] = "sdma",
+	[AMD_IP_BLOCK_TYPE_UVD] = "uvd",
+	[AMD_IP_BLOCK_TYPE_VCE] = "vce",
+	[AMD_IP_BLOCK_TYPE_ACP] = "acp",
+	[AMD_IP_BLOCK_TYPE_VCN] = "vcn",
+	[AMD_IP_BLOCK_TYPE_MES] = "mes",
+	[AMD_IP_BLOCK_TYPE_JPEG] = "jpeg",
+	[AMD_IP_BLOCK_TYPE_VPE] = "vpe",
+	[AMD_IP_BLOCK_TYPE_UMSCH_MM] = "umsch_mm",
+	[AMD_IP_BLOCK_TYPE_ISP] = "isp",
+	[AMD_IP_BLOCK_TYPE_RAS] = "ras",
+};
+
+static const char *ip_block_name(struct amdgpu_device *adev, enum amd_ip_block_type type)
+{
+	int idx = (int)type;
+
+	return idx < ARRAY_SIZE(ip_block_names) ? ip_block_names[idx] : "unknown";
+}
+
 /**
  * amdgpu_device_ip_block_add
  *
@@ -1808,8 +2535,15 @@ int amdgpu_device_ip_block_add(struct amdgpu_device *adev,
 		break;
 	}
 
-	DRM_INFO("add ip block number %d <%s>\n", adev->num_ip_blocks,
-		  ip_block_version->funcs->name);
+	dev_info(adev->dev, "detected ip block number %d <%s_v%d_%d_%d> (%s)\n",
+		 adev->num_ip_blocks,
+		 ip_block_name(adev, ip_block_version->type),
+		 ip_block_version->major,
+		 ip_block_version->minor,
+		 ip_block_version->rev,
+		 ip_block_version->funcs->name);
+
+	adev->ip_blocks[adev->num_ip_blocks].adev = adev;
 
 	adev->ip_blocks[adev->num_ip_blocks++].version = ip_block_version;
 
@@ -1825,7 +2559,7 @@ int amdgpu_device_ip_block_add(struct amdgpu_device *adev,
  * the module parameter virtual_display.  This feature provides a virtual
  * display hardware on headless boards or in virtualized environments.
  * This function parses and validates the configuration string specified by
- * the user and configues the virtual display configuration (number of
+ * the user and configures the virtual display configuration (number of
  * virtual connectors, crtcs, etc.) specified.
  */
 static void amdgpu_device_enable_virtual_display(struct amdgpu_device *adev)
@@ -1864,75 +2598,46 @@ static void amdgpu_device_enable_virtual_display(struct amdgpu_device *adev)
 			}
 		}
 
-		DRM_INFO("virtual display string:%s, %s:virtual_display:%d, num_crtc:%d\n",
-			 amdgpu_virtual_display, pci_address_name,
-			 adev->enable_virtual_display, adev->mode_info.num_crtc);
+		dev_info(
+			adev->dev,
+			"virtual display string:%s, %s:virtual_display:%d, num_crtc:%d\n",
+			amdgpu_virtual_display, pci_address_name,
+			adev->enable_virtual_display, adev->mode_info.num_crtc);
 
 		kfree(pciaddstr);
 	}
 }
 
+void amdgpu_device_set_sriov_virtual_display(struct amdgpu_device *adev)
+{
+	if (amdgpu_sriov_vf(adev) && !adev->enable_virtual_display) {
+		adev->mode_info.num_crtc = 1;
+		adev->enable_virtual_display = true;
+		dev_info(adev->dev, "virtual_display:%d, num_crtc:%d\n",
+			 adev->enable_virtual_display,
+			 adev->mode_info.num_crtc);
+	}
+}
+
 /**
  * amdgpu_device_parse_gpu_info_fw - parse gpu info firmware
  *
  * @adev: amdgpu_device pointer
  *
  * Parses the asic configuration parameters specified in the gpu info
- * firmware and makes them availale to the driver for use in configuring
+ * firmware and makes them available to the driver for use in configuring
  * the asic.
  * Returns 0 on success, -EINVAL on failure.
  */
 static int amdgpu_device_parse_gpu_info_fw(struct amdgpu_device *adev)
 {
 	const char *chip_name;
-	char fw_name[40];
 	int err;
 	const struct gpu_info_firmware_header_v1_0 *hdr;
 
 	adev->firmware.gpu_info_fw = NULL;
 
-	if (adev->mman.discovery_bin) {
-		amdgpu_discovery_get_gfx_info(adev);
-
-		/*
-		 * FIXME: The bounding box is still needed by Navi12, so
-		 * temporarily read it from gpu_info firmware. Should be droped
-		 * when DAL no longer needs it.
-		 */
-		if (adev->asic_type != CHIP_NAVI12)
-			return 0;
-	}
-
 	switch (adev->asic_type) {
-#ifdef CONFIG_DRM_AMDGPU_SI
-	case CHIP_VERDE:
-	case CHIP_TAHITI:
-	case CHIP_PITCAIRN:
-	case CHIP_OLAND:
-	case CHIP_HAINAN:
-#endif
-#ifdef CONFIG_DRM_AMDGPU_CIK
-	case CHIP_BONAIRE:
-	case CHIP_HAWAII:
-	case CHIP_KAVERI:
-	case CHIP_KABINI:
-	case CHIP_MULLINS:
-#endif
-	case CHIP_TOPAZ:
-	case CHIP_TONGA:
-	case CHIP_FIJI:
-	case CHIP_POLARIS10:
-	case CHIP_POLARIS11:
-	case CHIP_POLARIS12:
-	case CHIP_VEGAM:
-	case CHIP_CARRIZO:
-	case CHIP_STONEY:
-	case CHIP_VEGA20:
-	case CHIP_ALDEBARAN:
-	case CHIP_SIENNA_CICHLID:
-	case CHIP_NAVY_FLOUNDER:
-	case CHIP_DIMGREY_CAVEFISH:
-	case CHIP_BEIGE_GOBY:
 	default:
 		return 0;
 	case CHIP_VEGA10:
@@ -1952,42 +2657,25 @@ static int amdgpu_device_parse_gpu_info_fw(struct amdgpu_device *adev)
 	case CHIP_ARCTURUS:
 		chip_name = "arcturus";
 		break;
-	case CHIP_RENOIR:
-		if (adev->apu_flags & AMD_APU_IS_RENOIR)
-			chip_name = "renoir";
-		else
-			chip_name = "green_sardine";
-		break;
-	case CHIP_NAVI10:
-		chip_name = "navi10";
-		break;
-	case CHIP_NAVI14:
-		chip_name = "navi14";
-		break;
 	case CHIP_NAVI12:
+		if (adev->discovery.bin)
+			return 0;
 		chip_name = "navi12";
 		break;
-	case CHIP_VANGOGH:
-		chip_name = "vangogh";
-		break;
-	case CHIP_YELLOW_CARP:
-		chip_name = "yellow_carp";
+	case CHIP_CYAN_SKILLFISH:
+		if (adev->discovery.bin)
+			return 0;
+		chip_name = "cyan_skillfish";
 		break;
 	}
 
-	snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_gpu_info.bin", chip_name);
-	err = request_firmware(&adev->firmware.gpu_info_fw, fw_name, adev->dev);
+	err = amdgpu_ucode_request(adev, &adev->firmware.gpu_info_fw,
+				   AMDGPU_UCODE_OPTIONAL,
+				   "amdgpu/%s_gpu_info.bin", chip_name);
 	if (err) {
 		dev_err(adev->dev,
-			"Failed to load gpu_info firmware \"%s\"\n",
-			fw_name);
-		goto out;
-	}
-	err = amdgpu_ucode_validate(adev->firmware.gpu_info_fw);
-	if (err) {
-		dev_err(adev->dev,
-			"Failed to validate gpu_info firmware \"%s\"\n",
-			fw_name);
+			"Failed to get gpu_info firmware \"%s_gpu_info.bin\"\n",
+			chip_name);
 		goto out;
 	}
 
@@ -2002,7 +2690,7 @@ static int amdgpu_device_parse_gpu_info_fw(struct amdgpu_device *adev)
 								le32_to_cpu(hdr->header.ucode_array_offset_bytes));
 
 		/*
-		 * Should be droped when DAL no longer needs it.
+		 * Should be dropped when DAL no longer needs it.
 		 */
 		if (adev->asic_type == CHIP_NAVI12)
 			goto parse_soc_bounding_box;
@@ -2058,6 +2746,24 @@ out:
 	return err;
 }
 
+static void amdgpu_uid_init(struct amdgpu_device *adev)
+{
+	/* Initialize the UID for the device */
+	adev->uid_info = kzalloc(sizeof(struct amdgpu_uid), GFP_KERNEL);
+	if (!adev->uid_info) {
+		dev_warn(adev->dev, "Failed to allocate memory for UID\n");
+		return;
+	}
+	adev->uid_info->adev = adev;
+}
+
+static void amdgpu_uid_fini(struct amdgpu_device *adev)
+{
+	/* Free the UID memory */
+	kfree(adev->uid_info);
+	adev->uid_info = NULL;
+}
+
 /**
  * amdgpu_device_ip_early_init - run early init for hardware IPs
  *
@@ -2070,6 +2776,10 @@ out:
  */
 static int amdgpu_device_ip_early_init(struct amdgpu_device *adev)
 {
+	struct amdgpu_ip_block *ip_block;
+	struct pci_dev *parent;
+	bool total, skip_bios;
+	uint32_t bios_flags;
 	int i, r;
 
 	amdgpu_device_enable_virtual_display(adev);
@@ -2078,6 +2788,10 @@ static int amdgpu_device_ip_early_init(struct amdgpu_device *adev)
 		r = amdgpu_virt_request_full_gpu(adev, true);
 		if (r)
 			return r;
+
+		r = amdgpu_virt_init_critical_region(adev);
+		if (r)
+			return r;
 	}
 
 	switch (adev->asic_type) {
@@ -2134,34 +2848,62 @@ static int amdgpu_device_ip_early_init(struct amdgpu_device *adev)
 		break;
 	}
 
-	amdgpu_amdkfd_device_probe(adev);
+	/* Check for IP version 9.4.3 with A0 hardware */
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3) &&
+	    !amdgpu_device_get_rev_id(adev)) {
+		dev_err(adev->dev, "Unsupported A0 hardware\n");
+		return -ENODEV;	/* device unsupported - no device error */
+	}
+
+	if (amdgpu_has_atpx() &&
+	    (amdgpu_is_atpx_hybrid() ||
+	     amdgpu_has_atpx_dgpu_power_cntl()) &&
+	    ((adev->flags & AMD_IS_APU) == 0) &&
+	    !dev_is_removable(&adev->pdev->dev))
+		adev->flags |= AMD_IS_PX;
+
+	if (!(adev->flags & AMD_IS_APU)) {
+		parent = pcie_find_root_port(adev->pdev);
+		adev->has_pr3 = parent ? pci_pr3_present(parent) : false;
+	}
 
 	adev->pm.pp_feature = amdgpu_pp_feature_mask;
 	if (amdgpu_sriov_vf(adev) || sched_policy == KFD_SCHED_POLICY_NO_HWS)
 		adev->pm.pp_feature &= ~PP_GFXOFF_MASK;
 	if (amdgpu_sriov_vf(adev) && adev->asic_type == CHIP_SIENNA_CICHLID)
 		adev->pm.pp_feature &= ~PP_OVERDRIVE_MASK;
+	if (!amdgpu_device_pcie_dynamic_switching_supported(adev))
+		adev->pm.pp_feature &= ~PP_PCIE_DPM_MASK;
+
+	adev->virt.is_xgmi_node_migrate_enabled = false;
+	if (amdgpu_sriov_vf(adev)) {
+		adev->virt.is_xgmi_node_migrate_enabled =
+			amdgpu_ip_version((adev), GC_HWIP, 0) == IP_VERSION(9, 4, 4);
+	}
 
+	total = true;
 	for (i = 0; i < adev->num_ip_blocks; i++) {
+		ip_block = &adev->ip_blocks[i];
+
 		if ((amdgpu_ip_block_mask & (1 << i)) == 0) {
-			DRM_ERROR("disabled ip block: %d <%s>\n",
-				  i, adev->ip_blocks[i].version->funcs->name);
+			dev_warn(adev->dev, "disabled ip block: %d <%s>\n", i,
+				 adev->ip_blocks[i].version->funcs->name);
 			adev->ip_blocks[i].status.valid = false;
-		} else {
-			if (adev->ip_blocks[i].version->funcs->early_init) {
-				r = adev->ip_blocks[i].version->funcs->early_init((void *)adev);
-				if (r == -ENOENT) {
-					adev->ip_blocks[i].status.valid = false;
-				} else if (r) {
-					DRM_ERROR("early_init of IP block <%s> failed %d\n",
-						  adev->ip_blocks[i].version->funcs->name, r);
-					return r;
-				} else {
-					adev->ip_blocks[i].status.valid = true;
-				}
+		} else if (ip_block->version->funcs->early_init) {
+			r = ip_block->version->funcs->early_init(ip_block);
+			if (r == -ENOENT) {
+				adev->ip_blocks[i].status.valid = false;
+			} else if (r) {
+				dev_err(adev->dev,
+					"early_init of IP block <%s> failed %d\n",
+					adev->ip_blocks[i].version->funcs->name,
+					r);
+				total = false;
 			} else {
 				adev->ip_blocks[i].status.valid = true;
 			}
+		} else {
+			adev->ip_blocks[i].status.valid = true;
 		}
 		/* get the vbios after the asic_funcs are set up */
 		if (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_COMMON) {
@@ -2169,15 +2911,32 @@ static int amdgpu_device_ip_early_init(struct amdgpu_device *adev)
 			if (r)
 				return r;
 
+			bios_flags = amdgpu_device_get_vbios_flags(adev);
+			skip_bios = !!(bios_flags & AMDGPU_VBIOS_SKIP);
 			/* Read BIOS */
-			if (!amdgpu_get_bios(adev))
-				return -EINVAL;
-
-			r = amdgpu_atombios_init(adev);
-			if (r) {
-				dev_err(adev->dev, "amdgpu_atombios_init failed\n");
-				amdgpu_vf_error_put(adev, AMDGIM_ERROR_VF_ATOMBIOS_INIT_FAIL, 0, 0);
-				return r;
+			if (!skip_bios) {
+				bool optional =
+					!!(bios_flags & AMDGPU_VBIOS_OPTIONAL);
+				if (!amdgpu_get_bios(adev) && !optional)
+					return -EINVAL;
+
+				if (optional && !adev->bios)
+					dev_info(
+						adev->dev,
+						"VBIOS image optional, proceeding without VBIOS image");
+
+				if (adev->bios) {
+					r = amdgpu_atombios_init(adev);
+					if (r) {
+						dev_err(adev->dev,
+							"amdgpu_atombios_init failed\n");
+						amdgpu_vf_error_put(
+							adev,
+							AMDGIM_ERROR_VF_ATOMBIOS_INIT_FAIL,
+							0, 0);
+						return r;
+					}
+				}
 			}
 
 			/*get pf2vf msg info at it's earliest time*/
@@ -2186,6 +2945,17 @@ static int amdgpu_device_ip_early_init(struct amdgpu_device *adev)
 
 		}
 	}
+	if (!total)
+		return -ENODEV;
+
+	if (adev->gmc.xgmi.supported)
+		amdgpu_xgmi_early_init(adev);
+
+	if (amdgpu_is_multi_aid(adev))
+		amdgpu_uid_init(adev);
+	ip_block = amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_GFX);
+	if (ip_block->status.valid != false)
+		amdgpu_amdkfd_device_probe(adev);
 
 	adev->cg_flags &= amdgpu_cg_mask;
 	adev->pg_flags &= amdgpu_pg_mask;
@@ -2202,13 +2972,18 @@ static int amdgpu_device_ip_hw_init_phase1(struct amdgpu_device *adev)
 			continue;
 		if (adev->ip_blocks[i].status.hw)
 			continue;
+		if (!amdgpu_ip_member_of_hwini(
+			    adev, adev->ip_blocks[i].version->type))
+			continue;
 		if (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_COMMON ||
 		    (amdgpu_sriov_vf(adev) && (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_PSP)) ||
 		    adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_IH) {
-			r = adev->ip_blocks[i].version->funcs->hw_init(adev);
+			r = adev->ip_blocks[i].version->funcs->hw_init(&adev->ip_blocks[i]);
 			if (r) {
-				DRM_ERROR("hw_init of IP block <%s> failed %d\n",
-					  adev->ip_blocks[i].version->funcs->name, r);
+				dev_err(adev->dev,
+					"hw_init of IP block <%s> failed %d\n",
+					adev->ip_blocks[i].version->funcs->name,
+					r);
 				return r;
 			}
 			adev->ip_blocks[i].status.hw = true;
@@ -2227,10 +3002,14 @@ static int amdgpu_device_ip_hw_init_phase2(struct amdgpu_device *adev)
 			continue;
 		if (adev->ip_blocks[i].status.hw)
 			continue;
-		r = adev->ip_blocks[i].version->funcs->hw_init(adev);
+		if (!amdgpu_ip_member_of_hwini(
+			    adev, adev->ip_blocks[i].version->type))
+			continue;
+		r = adev->ip_blocks[i].version->funcs->hw_init(&adev->ip_blocks[i]);
 		if (r) {
-			DRM_ERROR("hw_init of IP block <%s> failed %d\n",
-				  adev->ip_blocks[i].version->funcs->name, r);
+			dev_err(adev->dev,
+				"hw_init of IP block <%s> failed %d\n",
+				adev->ip_blocks[i].version->funcs->name, r);
 			return r;
 		}
 		adev->ip_blocks[i].status.hw = true;
@@ -2250,6 +3029,10 @@ static int amdgpu_device_fw_loading(struct amdgpu_device *adev)
 			if (adev->ip_blocks[i].version->type != AMD_IP_BLOCK_TYPE_PSP)
 				continue;
 
+			if (!amdgpu_ip_member_of_hwini(adev,
+						       AMD_IP_BLOCK_TYPE_PSP))
+				break;
+
 			if (!adev->ip_blocks[i].status.sw)
 				continue;
 
@@ -2258,22 +3041,21 @@ static int amdgpu_device_fw_loading(struct amdgpu_device *adev)
 				break;
 
 			if (amdgpu_in_reset(adev) || adev->in_suspend) {
-				r = adev->ip_blocks[i].version->funcs->resume(adev);
-				if (r) {
-					DRM_ERROR("resume of IP block <%s> failed %d\n",
-							  adev->ip_blocks[i].version->funcs->name, r);
+				r = amdgpu_ip_block_resume(&adev->ip_blocks[i]);
+				if (r)
 					return r;
-				}
 			} else {
-				r = adev->ip_blocks[i].version->funcs->hw_init(adev);
+				r = adev->ip_blocks[i].version->funcs->hw_init(&adev->ip_blocks[i]);
 				if (r) {
-					DRM_ERROR("hw_init of IP block <%s> failed %d\n",
-							  adev->ip_blocks[i].version->funcs->name, r);
+					dev_err(adev->dev,
+						"hw_init of IP block <%s> failed %d\n",
+						adev->ip_blocks[i]
+							.version->funcs->name,
+						r);
 					return r;
 				}
+				adev->ip_blocks[i].status.hw = true;
 			}
-
-			adev->ip_blocks[i].status.hw = true;
 			break;
 		}
 	}
@@ -2284,6 +3066,74 @@ static int amdgpu_device_fw_loading(struct amdgpu_device *adev)
 	return r;
 }
 
+static int amdgpu_device_init_schedulers(struct amdgpu_device *adev)
+{
+	struct drm_sched_init_args args = {
+		.ops = &amdgpu_sched_ops,
+		.num_rqs = DRM_SCHED_PRIORITY_COUNT,
+		.timeout_wq = adev->reset_domain->wq,
+		.dev = adev->dev,
+	};
+	long timeout;
+	int r, i;
+
+	for (i = 0; i < AMDGPU_MAX_RINGS; ++i) {
+		struct amdgpu_ring *ring = adev->rings[i];
+
+		/* No need to setup the GPU scheduler for rings that don't need it */
+		if (!ring || ring->no_scheduler)
+			continue;
+
+		switch (ring->funcs->type) {
+		case AMDGPU_RING_TYPE_GFX:
+			timeout = adev->gfx_timeout;
+			break;
+		case AMDGPU_RING_TYPE_COMPUTE:
+			timeout = adev->compute_timeout;
+			break;
+		case AMDGPU_RING_TYPE_SDMA:
+			timeout = adev->sdma_timeout;
+			break;
+		default:
+			timeout = adev->video_timeout;
+			break;
+		}
+
+		args.timeout = timeout;
+		args.credit_limit = ring->num_hw_submission;
+		args.score = ring->sched_score;
+		args.name = ring->name;
+
+		r = drm_sched_init(&ring->sched, &args);
+		if (r) {
+			dev_err(adev->dev,
+				"Failed to create scheduler on ring %s.\n",
+				ring->name);
+			return r;
+		}
+		r = amdgpu_uvd_entity_init(adev, ring);
+		if (r) {
+			dev_err(adev->dev,
+				"Failed to create UVD scheduling entity on ring %s.\n",
+				ring->name);
+			return r;
+		}
+		r = amdgpu_vce_entity_init(adev, ring);
+		if (r) {
+			dev_err(adev->dev,
+				"Failed to create VCE scheduling entity on ring %s.\n",
+				ring->name);
+			return r;
+		}
+	}
+
+	if (adev->xcp_mgr)
+		amdgpu_xcp_update_partition_sched_list(adev);
+
+	return 0;
+}
+
+
 /**
  * amdgpu_device_ip_init - run init for hardware IPs
  *
@@ -2297,6 +3147,7 @@ static int amdgpu_device_fw_loading(struct amdgpu_device *adev)
  */
 static int amdgpu_device_ip_init(struct amdgpu_device *adev)
 {
+	bool init_badpage;
 	int i, r;
 
 	r = amdgpu_ras_init(adev);
@@ -2306,43 +3157,77 @@ static int amdgpu_device_ip_init(struct amdgpu_device *adev)
 	for (i = 0; i < adev->num_ip_blocks; i++) {
 		if (!adev->ip_blocks[i].status.valid)
 			continue;
-		r = adev->ip_blocks[i].version->funcs->sw_init((void *)adev);
-		if (r) {
-			DRM_ERROR("sw_init of IP block <%s> failed %d\n",
-				  adev->ip_blocks[i].version->funcs->name, r);
-			goto init_failed;
+		if (adev->ip_blocks[i].version->funcs->sw_init) {
+			r = adev->ip_blocks[i].version->funcs->sw_init(&adev->ip_blocks[i]);
+			if (r) {
+				dev_err(adev->dev,
+					"sw_init of IP block <%s> failed %d\n",
+					adev->ip_blocks[i].version->funcs->name,
+					r);
+				goto init_failed;
+			}
 		}
 		adev->ip_blocks[i].status.sw = true;
 
-		/* need to do gmc hw init early so we can allocate gpu mem */
-		if (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_GMC) {
-			r = amdgpu_device_vram_scratch_init(adev);
+		if (!amdgpu_ip_member_of_hwini(
+			    adev, adev->ip_blocks[i].version->type))
+			continue;
+
+		if (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_COMMON) {
+			/* need to do common hw init early so everything is set up for gmc */
+			r = adev->ip_blocks[i].version->funcs->hw_init(&adev->ip_blocks[i]);
 			if (r) {
-				DRM_ERROR("amdgpu_vram_scratch_init failed %d\n", r);
+				dev_err(adev->dev, "hw_init %d failed %d\n", i,
+					r);
 				goto init_failed;
 			}
-			r = adev->ip_blocks[i].version->funcs->hw_init((void *)adev);
+			adev->ip_blocks[i].status.hw = true;
+		} else if (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_GMC) {
+			/* need to do gmc hw init early so we can allocate gpu mem */
+			/* Try to reserve bad pages early */
+			if (amdgpu_sriov_vf(adev))
+				amdgpu_virt_exchange_data(adev);
+
+			r = amdgpu_device_mem_scratch_init(adev);
+			if (r) {
+				dev_err(adev->dev,
+					"amdgpu_mem_scratch_init failed %d\n",
+					r);
+				goto init_failed;
+			}
+			r = adev->ip_blocks[i].version->funcs->hw_init(&adev->ip_blocks[i]);
 			if (r) {
-				DRM_ERROR("hw_init %d failed %d\n", i, r);
+				dev_err(adev->dev, "hw_init %d failed %d\n", i,
+					r);
 				goto init_failed;
 			}
 			r = amdgpu_device_wb_init(adev);
 			if (r) {
-				DRM_ERROR("amdgpu_device_wb_init failed %d\n", r);
+				dev_err(adev->dev,
+					"amdgpu_device_wb_init failed %d\n", r);
 				goto init_failed;
 			}
 			adev->ip_blocks[i].status.hw = true;
 
 			/* right after GMC hw init, we create CSA */
-			if (amdgpu_mcbp || amdgpu_sriov_vf(adev)) {
+			if (adev->gfx.mcbp) {
 				r = amdgpu_allocate_static_csa(adev, &adev->virt.csa_obj,
-								AMDGPU_GEM_DOMAIN_VRAM,
-								AMDGPU_CSA_SIZE);
+							       AMDGPU_GEM_DOMAIN_VRAM |
+							       AMDGPU_GEM_DOMAIN_GTT,
+							       AMDGPU_CSA_SIZE);
 				if (r) {
-					DRM_ERROR("allocate CSA failed %d\n", r);
+					dev_err(adev->dev,
+						"allocate CSA failed %d\n", r);
 					goto init_failed;
 				}
 			}
+
+			r = amdgpu_seq64_init(adev);
+			if (r) {
+				dev_err(adev->dev, "allocate seq64 failed %d\n",
+					r);
+				goto init_failed;
+			}
 		}
 	}
 
@@ -2387,22 +3272,58 @@ static int amdgpu_device_ip_init(struct amdgpu_device *adev)
 	 * Note: theoretically, this should be called before all vram allocations
 	 * to protect retired page from abusing
 	 */
-	r = amdgpu_ras_recovery_init(adev);
+	init_badpage = (adev->init_lvl->level != AMDGPU_INIT_LEVEL_MINIMAL_XGMI);
+	r = amdgpu_ras_recovery_init(adev, init_badpage);
 	if (r)
 		goto init_failed;
 
-	if (adev->gmc.xgmi.num_physical_nodes > 1)
-		amdgpu_xgmi_add_device(adev);
+	/**
+	 * In case of XGMI grab extra reference for reset domain for this device
+	 */
+	if (adev->gmc.xgmi.num_physical_nodes > 1) {
+		if (amdgpu_xgmi_add_device(adev) == 0) {
+			if (!amdgpu_sriov_vf(adev)) {
+				struct amdgpu_hive_info *hive = amdgpu_get_xgmi_hive(adev);
+
+				if (WARN_ON(!hive)) {
+					r = -ENOENT;
+					goto init_failed;
+				}
+
+				if (!hive->reset_domain ||
+				    !amdgpu_reset_get_reset_domain(hive->reset_domain)) {
+					r = -ENOENT;
+					amdgpu_put_xgmi_hive(hive);
+					goto init_failed;
+				}
+
+				/* Drop the early temporary reset domain we created for device */
+				amdgpu_reset_put_reset_domain(adev->reset_domain);
+				adev->reset_domain = hive->reset_domain;
+				amdgpu_put_xgmi_hive(hive);
+			}
+		}
+	}
+
+	r = amdgpu_device_init_schedulers(adev);
+	if (r)
+		goto init_failed;
+
+	if (adev->mman.buffer_funcs_ring->sched.ready)
+		amdgpu_ttm_set_buffer_funcs_status(adev, true);
 
 	/* Don't init kfd if whole hive need to be reset during init */
-	if (!adev->gmc.xgmi.pending_reset)
+	if (adev->init_lvl->level != AMDGPU_INIT_LEVEL_MINIMAL_XGMI) {
+		kgd2kfd_init_zone_device(adev);
 		amdgpu_amdkfd_device_init(adev);
+	}
 
 	amdgpu_fru_get_product_info(adev);
 
+	if (!amdgpu_sriov_vf(adev) || amdgpu_sriov_ras_cper_en(adev))
+		r = amdgpu_cper_init(adev);
+
 init_failed:
-	if (amdgpu_sriov_vf(adev))
-		amdgpu_virt_release_full_gpu(adev, true);
 
 	return r;
 }
@@ -2414,7 +3335,7 @@ init_failed:
  *
  * Writes a reset magic value to the gart pointer in VRAM.  The driver calls
  * this function before a GPU reset.  If the value is retained after a
- * GPU reset, VRAM has not been lost.  Some GPU resets may destry VRAM contents.
+ * GPU reset, VRAM has not been lost. Some GPU resets may destroy VRAM contents.
  */
 static void amdgpu_device_fill_reset_magic(struct amdgpu_device *adev)
 {
@@ -2445,6 +3366,8 @@ static bool amdgpu_device_check_vram_lost(struct amdgpu_device *adev)
 	 * always assumed to be lost.
 	 */
 	switch (amdgpu_asic_reset_method(adev)) {
+	case AMD_RESET_METHOD_LEGACY:
+	case AMD_RESET_METHOD_LINK:
 	case AMD_RESET_METHOD_BACO:
 	case AMD_RESET_METHOD_MODE1:
 		return true;
@@ -2478,9 +3401,10 @@ int amdgpu_device_set_cg_state(struct amdgpu_device *adev,
 		i = state == AMD_CG_STATE_GATE ? j : adev->num_ip_blocks - j - 1;
 		if (!adev->ip_blocks[i].status.late_initialized)
 			continue;
-		/* skip CG for GFX on S0ix */
+		/* skip CG for GFX, SDMA on S0ix */
 		if (adev->in_s0ix &&
-		    adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_GFX)
+		    (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_GFX ||
+		     adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_SDMA))
 			continue;
 		/* skip CG for VCE/UVD, it's handled specially */
 		if (adev->ip_blocks[i].version->type != AMD_IP_BLOCK_TYPE_UVD &&
@@ -2489,11 +3413,13 @@ int amdgpu_device_set_cg_state(struct amdgpu_device *adev,
 		    adev->ip_blocks[i].version->type != AMD_IP_BLOCK_TYPE_JPEG &&
 		    adev->ip_blocks[i].version->funcs->set_clockgating_state) {
 			/* enable clockgating to save power */
-			r = adev->ip_blocks[i].version->funcs->set_clockgating_state((void *)adev,
+			r = adev->ip_blocks[i].version->funcs->set_clockgating_state(&adev->ip_blocks[i],
 										     state);
 			if (r) {
-				DRM_ERROR("set_clockgating_state(gate) of IP block <%s> failed %d\n",
-					  adev->ip_blocks[i].version->funcs->name, r);
+				dev_err(adev->dev,
+					"set_clockgating_state(gate) of IP block <%s> failed %d\n",
+					adev->ip_blocks[i].version->funcs->name,
+					r);
 				return r;
 			}
 		}
@@ -2514,22 +3440,26 @@ int amdgpu_device_set_pg_state(struct amdgpu_device *adev,
 		i = state == AMD_PG_STATE_GATE ? j : adev->num_ip_blocks - j - 1;
 		if (!adev->ip_blocks[i].status.late_initialized)
 			continue;
-		/* skip PG for GFX on S0ix */
+		/* skip PG for GFX, SDMA on S0ix */
 		if (adev->in_s0ix &&
-		    adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_GFX)
+		    (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_GFX ||
+		     adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_SDMA))
 			continue;
-		/* skip CG for VCE/UVD, it's handled specially */
+		/* skip CG for VCE/UVD/VPE, it's handled specially */
 		if (adev->ip_blocks[i].version->type != AMD_IP_BLOCK_TYPE_UVD &&
 		    adev->ip_blocks[i].version->type != AMD_IP_BLOCK_TYPE_VCE &&
 		    adev->ip_blocks[i].version->type != AMD_IP_BLOCK_TYPE_VCN &&
+		    adev->ip_blocks[i].version->type != AMD_IP_BLOCK_TYPE_VPE &&
 		    adev->ip_blocks[i].version->type != AMD_IP_BLOCK_TYPE_JPEG &&
 		    adev->ip_blocks[i].version->funcs->set_powergating_state) {
 			/* enable powergating to save power */
-			r = adev->ip_blocks[i].version->funcs->set_powergating_state((void *)adev,
+			r = adev->ip_blocks[i].version->funcs->set_powergating_state(&adev->ip_blocks[i],
 											state);
 			if (r) {
-				DRM_ERROR("set_powergating_state(gate) of IP block <%s> failed %d\n",
-					  adev->ip_blocks[i].version->funcs->name, r);
+				dev_err(adev->dev,
+					"set_powergating_state(gate) of IP block <%s> failed %d\n",
+					adev->ip_blocks[i].version->funcs->name,
+					r);
 				return r;
 			}
 		}
@@ -2556,7 +3486,7 @@ static int amdgpu_device_enable_mgpu_fan_boost(void)
 	for (i = 0; i < mgpu_info.num_dgpu; i++) {
 		gpu_ins = &(mgpu_info.gpu_ins[i]);
 		adev = gpu_ins->adev;
-		if (!(adev->flags & AMD_IS_APU) &&
+		if (!(adev->flags & AMD_IS_APU || amdgpu_sriov_multi_vf_mode(adev)) &&
 		    !gpu_ins->mgpu_fan_enabled) {
 			ret = amdgpu_dpm_enable_mgpu_fan_boost(adev);
 			if (ret)
@@ -2593,17 +3523,26 @@ static int amdgpu_device_ip_late_init(struct amdgpu_device *adev)
 		if (!adev->ip_blocks[i].status.hw)
 			continue;
 		if (adev->ip_blocks[i].version->funcs->late_init) {
-			r = adev->ip_blocks[i].version->funcs->late_init((void *)adev);
+			r = adev->ip_blocks[i].version->funcs->late_init(&adev->ip_blocks[i]);
 			if (r) {
-				DRM_ERROR("late_init of IP block <%s> failed %d\n",
-					  adev->ip_blocks[i].version->funcs->name, r);
+				dev_err(adev->dev,
+					"late_init of IP block <%s> failed %d\n",
+					adev->ip_blocks[i].version->funcs->name,
+					r);
 				return r;
 			}
 		}
 		adev->ip_blocks[i].status.late_initialized = true;
 	}
 
-	amdgpu_ras_set_error_query_ready(adev, true);
+	r = amdgpu_ras_late_init(adev);
+	if (r) {
+		dev_err(adev->dev, "amdgpu_ras_late_init failed %d", r);
+		return r;
+	}
+
+	if (!amdgpu_reset_in_recovery(adev))
+		amdgpu_ras_set_error_query_ready(adev, true);
 
 	amdgpu_device_set_cg_state(adev, AMD_CG_STATE_GATE);
 	amdgpu_device_set_pg_state(adev, AMD_PG_STATE_GATE);
@@ -2612,13 +3551,13 @@ static int amdgpu_device_ip_late_init(struct amdgpu_device *adev)
 
 	r = amdgpu_device_enable_mgpu_fan_boost();
 	if (r)
-		DRM_ERROR("enable mgpu fan boost failed (%d).\n", r);
+		dev_err(adev->dev, "enable mgpu fan boost failed (%d).\n", r);
 
-	/* For XGMI + passthrough configuration on arcturus, enable light SBR */
-	if (adev->asic_type == CHIP_ARCTURUS &&
-	    amdgpu_passthrough(adev) &&
-	    adev->gmc.xgmi.num_physical_nodes > 1)
-		smu_set_light_sbr(&adev->smu, true);
+	/* For passthrough configuration on arcturus and aldebaran, enable special handling SBR */
+	if (amdgpu_passthrough(adev) &&
+	    ((adev->asic_type == CHIP_ARCTURUS && adev->gmc.xgmi.num_physical_nodes > 1) ||
+	     adev->asic_type == CHIP_ALDEBARAN))
+		amdgpu_dpm_handle_passthrough_sbr(adev, true);
 
 	if (adev->gmc.xgmi.num_physical_nodes > 1) {
 		mutex_lock(&mgpu_info.mutex);
@@ -2645,7 +3584,9 @@ static int amdgpu_device_ip_late_init(struct amdgpu_device *adev)
 				r = amdgpu_xgmi_set_pstate(gpu_instance->adev,
 						AMDGPU_XGMI_PSTATE_MIN);
 				if (r) {
-					DRM_ERROR("pstate setting failed (%d).\n", r);
+					dev_err(adev->dev,
+						"pstate setting failed (%d).\n",
+						r);
 					break;
 				}
 			}
@@ -2657,59 +3598,101 @@ static int amdgpu_device_ip_late_init(struct amdgpu_device *adev)
 	return 0;
 }
 
-static int amdgpu_device_ip_fini_early(struct amdgpu_device *adev)
+static void amdgpu_ip_block_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	int i, r;
-
-	for (i = 0; i < adev->num_ip_blocks; i++) {
-		if (!adev->ip_blocks[i].version->funcs->early_fini)
-			continue;
+	struct amdgpu_device *adev = ip_block->adev;
+	int r;
 
-		r = adev->ip_blocks[i].version->funcs->early_fini((void *)adev);
+	if (!ip_block->version->funcs->hw_fini) {
+		dev_err(adev->dev, "hw_fini of IP block <%s> not defined\n",
+			ip_block->version->funcs->name);
+	} else {
+		r = ip_block->version->funcs->hw_fini(ip_block);
+		/* XXX handle errors */
 		if (r) {
-			DRM_DEBUG("early_fini of IP block <%s> failed %d\n",
-				  adev->ip_blocks[i].version->funcs->name, r);
+			dev_dbg(adev->dev,
+				"hw_fini of IP block <%s> failed %d\n",
+				ip_block->version->funcs->name, r);
 		}
 	}
 
-	amdgpu_amdkfd_suspend(adev, false);
+	ip_block->status.hw = false;
+}
 
-	amdgpu_device_set_pg_state(adev, AMD_PG_STATE_UNGATE);
-	amdgpu_device_set_cg_state(adev, AMD_CG_STATE_UNGATE);
+/**
+ * amdgpu_device_smu_fini_early - smu hw_fini wrapper
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * For ASICs need to disable SMC first
+ */
+static void amdgpu_device_smu_fini_early(struct amdgpu_device *adev)
+{
+	int i;
+
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) > IP_VERSION(9, 0, 0))
+		return;
 
-	/* need to disable SMC first */
 	for (i = 0; i < adev->num_ip_blocks; i++) {
 		if (!adev->ip_blocks[i].status.hw)
 			continue;
 		if (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_SMC) {
-			r = adev->ip_blocks[i].version->funcs->hw_fini((void *)adev);
-			/* XXX handle errors */
-			if (r) {
-				DRM_DEBUG("hw_fini of IP block <%s> failed %d\n",
-					  adev->ip_blocks[i].version->funcs->name, r);
-			}
-			adev->ip_blocks[i].status.hw = false;
+			amdgpu_ip_block_hw_fini(&adev->ip_blocks[i]);
 			break;
 		}
 	}
+}
 
-	for (i = adev->num_ip_blocks - 1; i >= 0; i--) {
-		if (!adev->ip_blocks[i].status.hw)
+static int amdgpu_device_ip_fini_early(struct amdgpu_device *adev)
+{
+	int i, r;
+
+	for (i = 0; i < adev->num_ip_blocks; i++) {
+		if (!adev->ip_blocks[i].version->funcs->early_fini)
 			continue;
 
-		r = adev->ip_blocks[i].version->funcs->hw_fini((void *)adev);
-		/* XXX handle errors */
+		r = adev->ip_blocks[i].version->funcs->early_fini(&adev->ip_blocks[i]);
 		if (r) {
-			DRM_DEBUG("hw_fini of IP block <%s> failed %d\n",
-				  adev->ip_blocks[i].version->funcs->name, r);
+			dev_dbg(adev->dev,
+				"early_fini of IP block <%s> failed %d\n",
+				adev->ip_blocks[i].version->funcs->name, r);
 		}
+	}
+
+	amdgpu_device_set_pg_state(adev, AMD_PG_STATE_UNGATE);
+	amdgpu_device_set_cg_state(adev, AMD_CG_STATE_UNGATE);
 
-		adev->ip_blocks[i].status.hw = false;
+	amdgpu_amdkfd_suspend(adev, true);
+	amdgpu_userq_suspend(adev);
+
+	/* Workaround for ASICs need to disable SMC first */
+	amdgpu_device_smu_fini_early(adev);
+
+	for (i = adev->num_ip_blocks - 1; i >= 0; i--) {
+		if (!adev->ip_blocks[i].status.hw)
+			continue;
+
+		amdgpu_ip_block_hw_fini(&adev->ip_blocks[i]);
 	}
 
 	if (amdgpu_sriov_vf(adev)) {
 		if (amdgpu_virt_release_full_gpu(adev, false))
-			DRM_ERROR("failed to release exclusive mode on fini\n");
+			dev_err(adev->dev,
+				"failed to release exclusive mode on fini\n");
+	}
+
+	/*
+	 * Driver reload on the APU can fail due to firmware validation because
+	 * the PSP is always running, as it is shared across the whole SoC.
+	 * This same issue does not occur on dGPU because it has a mechanism
+	 * that checks whether the PSP is running. A solution for those issues
+	 * in the APU is to trigger a GPU reset, but this should be done during
+	 * the unload phase to avoid adding boot latency and screen flicker.
+	 */
+	if ((adev->flags & AMD_IS_APU) && !adev->gmc.is_app_apu) {
+		r = amdgpu_asic_reset(adev);
+		if (r)
+			dev_err(adev->dev, "asic reset on %s failed\n", __func__);
 	}
 
 	return 0;
@@ -2730,11 +3713,11 @@ static int amdgpu_device_ip_fini(struct amdgpu_device *adev)
 {
 	int i, r;
 
+	amdgpu_cper_fini(adev);
+
 	if (amdgpu_sriov_vf(adev) && adev->virt.ras_init_done)
 		amdgpu_virt_release_ras_err_handler_data(adev);
 
-	amdgpu_ras_pre_fini(adev);
-
 	if (adev->gmc.xgmi.num_physical_nodes > 1)
 		amdgpu_xgmi_remove_device(adev);
 
@@ -2748,15 +3731,20 @@ static int amdgpu_device_ip_fini(struct amdgpu_device *adev)
 			amdgpu_ucode_free_bo(adev);
 			amdgpu_free_static_csa(&adev->virt.csa_obj);
 			amdgpu_device_wb_fini(adev);
-			amdgpu_device_vram_scratch_fini(adev);
+			amdgpu_device_mem_scratch_fini(adev);
 			amdgpu_ib_pool_fini(adev);
+			amdgpu_seq64_fini(adev);
+			amdgpu_doorbell_fini(adev);
 		}
-
-		r = adev->ip_blocks[i].version->funcs->sw_fini((void *)adev);
-		/* XXX handle errors */
-		if (r) {
-			DRM_DEBUG("sw_fini of IP block <%s> failed %d\n",
-				  adev->ip_blocks[i].version->funcs->name, r);
+		if (adev->ip_blocks[i].version->funcs->sw_fini) {
+			r = adev->ip_blocks[i].version->funcs->sw_fini(&adev->ip_blocks[i]);
+			/* XXX handle errors */
+			if (r) {
+				dev_dbg(adev->dev,
+					"sw_fini of IP block <%s> failed %d\n",
+					adev->ip_blocks[i].version->funcs->name,
+					r);
+			}
 		}
 		adev->ip_blocks[i].status.sw = false;
 		adev->ip_blocks[i].status.valid = false;
@@ -2766,11 +3754,12 @@ static int amdgpu_device_ip_fini(struct amdgpu_device *adev)
 		if (!adev->ip_blocks[i].status.late_initialized)
 			continue;
 		if (adev->ip_blocks[i].version->funcs->late_fini)
-			adev->ip_blocks[i].version->funcs->late_fini((void *)adev);
+			adev->ip_blocks[i].version->funcs->late_fini(&adev->ip_blocks[i]);
 		adev->ip_blocks[i].status.late_initialized = false;
 	}
 
 	amdgpu_ras_fini(adev);
+	amdgpu_uid_fini(adev);
 
 	return 0;
 }
@@ -2788,7 +3777,7 @@ static void amdgpu_device_delayed_init_work_handler(struct work_struct *work)
 
 	r = amdgpu_ib_ring_tests(adev);
 	if (r)
-		DRM_ERROR("ib ring test failed (%d).\n", r);
+		dev_err(adev->dev, "ib ring test failed (%d).\n", r);
 }
 
 static void amdgpu_device_delay_enable_gfx_off(struct work_struct *work)
@@ -2799,7 +3788,7 @@ static void amdgpu_device_delay_enable_gfx_off(struct work_struct *work)
 	WARN_ON_ONCE(adev->gfx.gfx_off_state);
 	WARN_ON_ONCE(adev->gfx.gfx_off_req_count);
 
-	if (!amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_GFX, true))
+	if (!amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_GFX, true, 0))
 		adev->gfx.gfx_off_state = true;
 }
 
@@ -2816,11 +3805,19 @@ static void amdgpu_device_delay_enable_gfx_off(struct work_struct *work)
  */
 static int amdgpu_device_ip_suspend_phase1(struct amdgpu_device *adev)
 {
-	int i, r;
+	int i, r, rec;
 
 	amdgpu_device_set_pg_state(adev, AMD_PG_STATE_UNGATE);
 	amdgpu_device_set_cg_state(adev, AMD_CG_STATE_UNGATE);
 
+	/*
+	 * Per PMFW team's suggestion, driver needs to handle gfxoff
+	 * and df cstate features disablement for gpu reset(e.g. Mode1Reset)
+	 * scenario. Add the missing df cstate disablement here.
+	 */
+	if (amdgpu_dpm_set_df_cstate(adev, DF_CSTATE_DISALLOW))
+		dev_warn(adev->dev, "Failed to disallow df cstate");
+
 	for (i = adev->num_ip_blocks - 1; i >= 0; i--) {
 		if (!adev->ip_blocks[i].status.valid)
 			continue;
@@ -2829,19 +3826,25 @@ static int amdgpu_device_ip_suspend_phase1(struct amdgpu_device *adev)
 		if (adev->ip_blocks[i].version->type != AMD_IP_BLOCK_TYPE_DCE)
 			continue;
 
-		/* XXX handle errors */
-		r = adev->ip_blocks[i].version->funcs->suspend(adev);
-		/* XXX handle errors */
-		if (r) {
-			DRM_ERROR("suspend of IP block <%s> failed %d\n",
-				  adev->ip_blocks[i].version->funcs->name, r);
-			return r;
-		}
-
-		adev->ip_blocks[i].status.hw = false;
+		r = amdgpu_ip_block_suspend(&adev->ip_blocks[i]);
+		if (r)
+			goto unwind;
 	}
 
 	return 0;
+unwind:
+	rec = amdgpu_device_ip_resume_phase3(adev);
+	if (rec)
+		dev_err(adev->dev,
+			"amdgpu_device_ip_resume_phase3 failed during unwind: %d\n",
+			rec);
+
+	amdgpu_dpm_set_df_cstate(adev, DF_CSTATE_ALLOW);
+
+	amdgpu_device_set_pg_state(adev, AMD_PG_STATE_GATE);
+	amdgpu_device_set_cg_state(adev, AMD_CG_STATE_GATE);
+
+	return r;
 }
 
 /**
@@ -2857,10 +3860,10 @@ static int amdgpu_device_ip_suspend_phase1(struct amdgpu_device *adev)
  */
 static int amdgpu_device_ip_suspend_phase2(struct amdgpu_device *adev)
 {
-	int i, r;
+	int i, r, rec;
 
 	if (adev->in_s0ix)
-		amdgpu_gfx_state_change_set(adev, sGpuChangeState_D3Entry);
+		amdgpu_dpm_gfx_state_change(adev, sGpuChangeState_D3Entry);
 
 	for (i = adev->num_ip_blocks - 1; i >= 0; i--) {
 		if (!adev->ip_blocks[i].status.valid)
@@ -2876,47 +3879,94 @@ static int amdgpu_device_ip_suspend_phase2(struct amdgpu_device *adev)
 		}
 
 		/* skip unnecessary suspend if we do not initialize them yet */
-		if (adev->gmc.xgmi.pending_reset &&
-		    !(adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_GMC ||
-		      adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_SMC ||
-		      adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_COMMON ||
-		      adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_IH)) {
-			adev->ip_blocks[i].status.hw = false;
+		if (!amdgpu_ip_member_of_hwini(
+			    adev, adev->ip_blocks[i].version->type))
 			continue;
-		}
 
-		/* skip suspend of gfx and psp for S0ix
+		/* Since we skip suspend for S0i3, we need to cancel the delayed
+		 * idle work here as the suspend callback never gets called.
+		 */
+		if (adev->in_s0ix &&
+		    adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_GFX &&
+		    amdgpu_ip_version(adev, GC_HWIP, 0) >= IP_VERSION(10, 0, 0))
+			cancel_delayed_work_sync(&adev->gfx.idle_work);
+		/* skip suspend of gfx/mes and psp for S0ix
 		 * gfx is in gfxoff state, so on resume it will exit gfxoff just
 		 * like at runtime. PSP is also part of the always on hardware
 		 * so no need to suspend it.
 		 */
 		if (adev->in_s0ix &&
 		    (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_PSP ||
-		     adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_GFX))
+		     adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_GFX ||
+		     adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_MES))
 			continue;
 
-		/* XXX handle errors */
-		r = adev->ip_blocks[i].version->funcs->suspend(adev);
-		/* XXX handle errors */
-		if (r) {
-			DRM_ERROR("suspend of IP block <%s> failed %d\n",
-				  adev->ip_blocks[i].version->funcs->name, r);
-		}
-		adev->ip_blocks[i].status.hw = false;
+		/* SDMA 5.x+ is part of GFX power domain so it's covered by GFXOFF */
+		if (adev->in_s0ix &&
+		    (amdgpu_ip_version(adev, SDMA0_HWIP, 0) >=
+		     IP_VERSION(5, 0, 0)) &&
+		    (adev->ip_blocks[i].version->type ==
+		     AMD_IP_BLOCK_TYPE_SDMA))
+			continue;
+
+		/* Once swPSP provides the IMU, RLC FW binaries to TOS during cold-boot.
+		 * These are in TMR, hence are expected to be reused by PSP-TOS to reload
+		 * from this location and RLC Autoload automatically also gets loaded
+		 * from here based on PMFW -> PSP message during re-init sequence.
+		 * Therefore, the psp suspend & resume should be skipped to avoid destroy
+		 * the TMR and reload FWs again for IMU enabled APU ASICs.
+		 */
+		if (amdgpu_in_reset(adev) &&
+		    (adev->flags & AMD_IS_APU) && adev->gfx.imu.funcs &&
+		    adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_PSP)
+			continue;
+
+		r = amdgpu_ip_block_suspend(&adev->ip_blocks[i]);
+		if (r)
+			goto unwind;
+
 		/* handle putting the SMC in the appropriate state */
-		if(!amdgpu_sriov_vf(adev)){
+		if (!amdgpu_sriov_vf(adev)) {
 			if (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_SMC) {
 				r = amdgpu_dpm_set_mp1_state(adev, adev->mp1_state);
 				if (r) {
-					DRM_ERROR("SMC failed to set mp1 state %d, %d\n",
-							adev->mp1_state, r);
-					return r;
+					dev_err(adev->dev,
+						"SMC failed to set mp1 state %d, %d\n",
+						adev->mp1_state, r);
+					goto unwind;
 				}
 			}
 		}
 	}
 
 	return 0;
+unwind:
+	/* suspend phase 2 = resume phase 1 + resume phase 2 */
+	rec = amdgpu_device_ip_resume_phase1(adev);
+	if (rec) {
+		dev_err(adev->dev,
+			"amdgpu_device_ip_resume_phase1 failed during unwind: %d\n",
+			rec);
+		return r;
+	}
+
+	rec = amdgpu_device_fw_loading(adev);
+	if (rec) {
+		dev_err(adev->dev,
+			"amdgpu_device_fw_loading failed during unwind: %d\n",
+			rec);
+		return r;
+	}
+
+	rec = amdgpu_device_ip_resume_phase2(adev);
+	if (rec) {
+		dev_err(adev->dev,
+			"amdgpu_device_ip_resume_phase2 failed during unwind: %d\n",
+			rec);
+		return r;
+	}
+
+	return r;
 }
 
 /**
@@ -2930,7 +3980,7 @@ static int amdgpu_device_ip_suspend_phase2(struct amdgpu_device *adev)
  * in each IP into a state suitable for suspend.
  * Returns 0 on success, negative error code on failure.
  */
-int amdgpu_device_ip_suspend(struct amdgpu_device *adev)
+static int amdgpu_device_ip_suspend(struct amdgpu_device *adev)
 {
 	int r;
 
@@ -2939,6 +3989,8 @@ int amdgpu_device_ip_suspend(struct amdgpu_device *adev)
 		amdgpu_virt_request_full_gpu(adev, false);
 	}
 
+	amdgpu_ttm_set_buffer_funcs_status(adev, false);
+
 	r = amdgpu_device_ip_suspend_phase1(adev);
 	if (r)
 		return r;
@@ -2955,8 +4007,8 @@ static int amdgpu_device_ip_reinit_early_sriov(struct amdgpu_device *adev)
 	int i, r;
 
 	static enum amd_ip_block_type ip_order[] = {
-		AMD_IP_BLOCK_TYPE_GMC,
 		AMD_IP_BLOCK_TYPE_COMMON,
+		AMD_IP_BLOCK_TYPE_GMC,
 		AMD_IP_BLOCK_TYPE_PSP,
 		AMD_IP_BLOCK_TYPE_IH,
 	};
@@ -2974,10 +4026,12 @@ static int amdgpu_device_ip_reinit_early_sriov(struct amdgpu_device *adev)
 				!block->status.valid)
 				continue;
 
-			r = block->version->funcs->hw_init(adev);
-			DRM_INFO("RE-INIT-early: %s %s\n", block->version->funcs->name, r?"failed":"succeeded");
-			if (r)
+			r = block->version->funcs->hw_init(&adev->ip_blocks[i]);
+			if (r) {
+				dev_err(adev->dev, "RE-INIT-early: %s failed\n",
+					 block->version->funcs->name);
 				return r;
+			}
 			block->status.hw = true;
 		}
 	}
@@ -2987,43 +4041,44 @@ static int amdgpu_device_ip_reinit_early_sriov(struct amdgpu_device *adev)
 
 static int amdgpu_device_ip_reinit_late_sriov(struct amdgpu_device *adev)
 {
-	int i, r;
+	struct amdgpu_ip_block *block;
+	int i, r = 0;
 
 	static enum amd_ip_block_type ip_order[] = {
 		AMD_IP_BLOCK_TYPE_SMC,
 		AMD_IP_BLOCK_TYPE_DCE,
 		AMD_IP_BLOCK_TYPE_GFX,
 		AMD_IP_BLOCK_TYPE_SDMA,
+		AMD_IP_BLOCK_TYPE_MES,
 		AMD_IP_BLOCK_TYPE_UVD,
 		AMD_IP_BLOCK_TYPE_VCE,
-		AMD_IP_BLOCK_TYPE_VCN
+		AMD_IP_BLOCK_TYPE_VCN,
+		AMD_IP_BLOCK_TYPE_JPEG
 	};
 
 	for (i = 0; i < ARRAY_SIZE(ip_order); i++) {
-		int j;
-		struct amdgpu_ip_block *block;
+		block = amdgpu_device_ip_get_ip_block(adev, ip_order[i]);
 
-		for (j = 0; j < adev->num_ip_blocks; j++) {
-			block = &adev->ip_blocks[j];
-
-			if (block->version->type != ip_order[i] ||
-				!block->status.valid ||
-				block->status.hw)
-				continue;
+		if (!block)
+			continue;
 
-			if (block->version->type == AMD_IP_BLOCK_TYPE_SMC)
-				r = block->version->funcs->resume(adev);
-			else
-				r = block->version->funcs->hw_init(adev);
+		if (block->status.valid && !block->status.hw) {
+			if (block->version->type == AMD_IP_BLOCK_TYPE_SMC) {
+				r = amdgpu_ip_block_resume(block);
+			} else {
+				r = block->version->funcs->hw_init(block);
+			}
 
-			DRM_INFO("RE-INIT-late: %s %s\n", block->version->funcs->name, r?"failed":"succeeded");
-			if (r)
-				return r;
+			if (r) {
+				dev_err(adev->dev, "RE-INIT-late: %s failed\n",
+					 block->version->funcs->name);
+				break;
+			}
 			block->status.hw = true;
 		}
 	}
 
-	return 0;
+	return r;
 }
 
 /**
@@ -3047,15 +4102,12 @@ static int amdgpu_device_ip_resume_phase1(struct amdgpu_device *adev)
 			continue;
 		if (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_COMMON ||
 		    adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_GMC ||
-		    adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_IH) {
+		    adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_IH ||
+		    (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_PSP && amdgpu_sriov_vf(adev))) {
 
-			r = adev->ip_blocks[i].version->funcs->resume(adev);
-			if (r) {
-				DRM_ERROR("resume of IP block <%s> failed %d\n",
-					  adev->ip_blocks[i].version->funcs->name, r);
+			r = amdgpu_ip_block_resume(&adev->ip_blocks[i]);
+			if (r)
 				return r;
-			}
-			adev->ip_blocks[i].status.hw = true;
 		}
 	}
 
@@ -3067,7 +4119,7 @@ static int amdgpu_device_ip_resume_phase1(struct amdgpu_device *adev)
  *
  * @adev: amdgpu_device pointer
  *
- * First resume function for hardware IPs.  The list of all the hardware
+ * Second resume function for hardware IPs.  The list of all the hardware
  * IPs that make up the asic is walked and the resume callbacks are run for
  * all blocks except COMMON, GMC, and IH.  resume puts the hardware into a
  * functional state after a suspend and updates the software state as
@@ -3085,15 +4137,42 @@ static int amdgpu_device_ip_resume_phase2(struct amdgpu_device *adev)
 		if (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_COMMON ||
 		    adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_GMC ||
 		    adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_IH ||
+		    adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_DCE ||
 		    adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_PSP)
 			continue;
-		r = adev->ip_blocks[i].version->funcs->resume(adev);
-		if (r) {
-			DRM_ERROR("resume of IP block <%s> failed %d\n",
-				  adev->ip_blocks[i].version->funcs->name, r);
+		r = amdgpu_ip_block_resume(&adev->ip_blocks[i]);
+		if (r)
 			return r;
+	}
+
+	return 0;
+}
+
+/**
+ * amdgpu_device_ip_resume_phase3 - run resume for hardware IPs
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Third resume function for hardware IPs.  The list of all the hardware
+ * IPs that make up the asic is walked and the resume callbacks are run for
+ * all DCE.  resume puts the hardware into a functional state after a suspend
+ * and updates the software state as necessary.  This function is also used
+ * for restoring the GPU after a GPU reset.
+ *
+ * Returns 0 on success, negative error code on failure.
+ */
+static int amdgpu_device_ip_resume_phase3(struct amdgpu_device *adev)
+{
+	int i, r;
+
+	for (i = 0; i < adev->num_ip_blocks; i++) {
+		if (!adev->ip_blocks[i].status.valid || adev->ip_blocks[i].status.hw)
+			continue;
+		if (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_DCE) {
+			r = amdgpu_ip_block_resume(&adev->ip_blocks[i]);
+			if (r)
+				return r;
 		}
-		adev->ip_blocks[i].status.hw = true;
 	}
 
 	return 0;
@@ -3106,7 +4185,7 @@ static int amdgpu_device_ip_resume_phase2(struct amdgpu_device *adev)
  *
  * Main resume function for hardware IPs.  The hardware IPs
  * are split into two resume functions because they are
- * are also used in in recovering from a GPU reset and some additional
+ * also used in recovering from a GPU reset and some additional
  * steps need to be take between them.  In this case (S3/S4) they are
  * run sequentially.
  * Returns 0 on success, negative error code on failure.
@@ -3115,10 +4194,6 @@ static int amdgpu_device_ip_resume(struct amdgpu_device *adev)
 {
 	int r;
 
-	r = amdgpu_amdkfd_resume_iommu(adev);
-	if (r)
-		return r;
-
 	r = amdgpu_device_ip_resume_phase1(adev);
 	if (r)
 		return r;
@@ -3129,6 +4204,16 @@ static int amdgpu_device_ip_resume(struct amdgpu_device *adev)
 
 	r = amdgpu_device_ip_resume_phase2(adev);
 
+	if (adev->mman.buffer_funcs_ring->sched.ready)
+		amdgpu_ttm_set_buffer_funcs_status(adev, true);
+
+	if (r)
+		return r;
+
+	amdgpu_fence_driver_hw_init(adev);
+
+	r = amdgpu_device_ip_resume_phase3(adev);
+
 	return r;
 }
 
@@ -3158,76 +4243,47 @@ static void amdgpu_device_detect_sriov_bios(struct amdgpu_device *adev)
 /**
  * amdgpu_device_asic_has_dc_support - determine if DC supports the asic
  *
+ * @pdev : pci device context
  * @asic_type: AMD asic type
  *
  * Check if there is DC (new modesetting infrastructre) support for an asic.
  * returns true if DC has support, false if not.
  */
-bool amdgpu_device_asic_has_dc_support(enum amd_asic_type asic_type)
+bool amdgpu_device_asic_has_dc_support(struct pci_dev *pdev,
+				       enum amd_asic_type asic_type)
 {
 	switch (asic_type) {
+#ifdef CONFIG_DRM_AMDGPU_SI
+	case CHIP_HAINAN:
+#endif
+	case CHIP_TOPAZ:
+		/* chips with no display hardware */
+		return false;
 #if defined(CONFIG_DRM_AMD_DC)
 	case CHIP_TAHITI:
 	case CHIP_PITCAIRN:
 	case CHIP_VERDE:
 	case CHIP_OLAND:
-		/*
-		 * We have systems in the wild with these ASICs that require
-		 * LVDS and VGA support which is not supported with DC.
-		 *
-		 * Fallback to the non-DC driver here by default so as not to
-		 * cause regressions.
-		 */
-#if defined(CONFIG_DRM_AMD_DC_SI)
-		return amdgpu_dc > 0;
-#else
-		return false;
-#endif
-	case CHIP_BONAIRE:
+		return amdgpu_dc != 0 && IS_ENABLED(CONFIG_DRM_AMD_DC_SI);
 	case CHIP_KAVERI:
 	case CHIP_KABINI:
 	case CHIP_MULLINS:
 		/*
 		 * We have systems in the wild with these ASICs that require
-		 * LVDS and VGA support which is not supported with DC.
+		 * TRAVIS and NUTMEG support which is not supported with DC.
 		 *
 		 * Fallback to the non-DC driver here by default so as not to
 		 * cause regressions.
 		 */
 		return amdgpu_dc > 0;
-	case CHIP_HAWAII:
-	case CHIP_CARRIZO:
-	case CHIP_STONEY:
-	case CHIP_POLARIS10:
-	case CHIP_POLARIS11:
-	case CHIP_POLARIS12:
-	case CHIP_VEGAM:
-	case CHIP_TONGA:
-	case CHIP_FIJI:
-	case CHIP_VEGA10:
-	case CHIP_VEGA12:
-	case CHIP_VEGA20:
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	case CHIP_RAVEN:
-	case CHIP_NAVI10:
-	case CHIP_NAVI14:
-	case CHIP_NAVI12:
-	case CHIP_RENOIR:
-	case CHIP_CYAN_SKILLFISH:
-	case CHIP_SIENNA_CICHLID:
-	case CHIP_NAVY_FLOUNDER:
-	case CHIP_DIMGREY_CAVEFISH:
-	case CHIP_BEIGE_GOBY:
-	case CHIP_VANGOGH:
-	case CHIP_YELLOW_CARP:
-#endif
 	default:
 		return amdgpu_dc != 0;
 #else
 	default:
 		if (amdgpu_dc > 0)
-			DRM_INFO_ONCE("Display Core has been requested via kernel parameter "
-					 "but isn't supported by ASIC, ignoring\n");
+			dev_info_once(
+				&pdev->dev,
+				"Display Core has been requested via kernel parameter but isn't supported by ASIC, ignoring\n");
 		return false;
 #endif
 	}
@@ -3242,12 +4298,11 @@ bool amdgpu_device_asic_has_dc_support(enum amd_asic_type asic_type)
  */
 bool amdgpu_device_has_dc_support(struct amdgpu_device *adev)
 {
-	if (amdgpu_sriov_vf(adev) || 
-	    adev->enable_virtual_display ||
+	if (adev->enable_virtual_display ||
 	    (adev->harvest_ip_mask & AMD_HARVEST_IP_DMU_MASK))
 		return false;
 
-	return amdgpu_device_asic_has_dc_support(adev->asic_type);
+	return amdgpu_device_asic_has_dc_support(adev->pdev, adev->asic_type);
 }
 
 static void amdgpu_device_xgmi_reset_func(struct work_struct *__work)
@@ -3269,20 +4324,18 @@ static void amdgpu_device_xgmi_reset_func(struct work_struct *__work)
 	if (amdgpu_asic_reset_method(adev) == AMD_RESET_METHOD_BACO) {
 
 		task_barrier_enter(&hive->tb);
-		adev->asic_reset_res = amdgpu_device_baco_enter(adev_to_drm(adev));
+		adev->asic_reset_res = amdgpu_device_baco_enter(adev);
 
 		if (adev->asic_reset_res)
 			goto fail;
 
 		task_barrier_exit(&hive->tb);
-		adev->asic_reset_res = amdgpu_device_baco_exit(adev_to_drm(adev));
+		adev->asic_reset_res = amdgpu_device_baco_exit(adev);
 
 		if (adev->asic_reset_res)
 			goto fail;
 
-		if (adev->mmhub.ras_funcs &&
-		    adev->mmhub.ras_funcs->reset_ras_error_count)
-			adev->mmhub.ras_funcs->reset_ras_error_count(adev);
+		amdgpu_ras_reset_error_count(adev, AMDGPU_RAS_BLOCK__MMHUB);
 	} else {
 
 		task_barrier_full(&hive->tb);
@@ -3291,7 +4344,8 @@ static void amdgpu_device_xgmi_reset_func(struct work_struct *__work)
 
 fail:
 	if (adev->asic_reset_res)
-		DRM_WARN("ASIC reset failed with error, %d for drm dev, %s",
+		dev_warn(adev->dev,
+			 "ASIC reset failed with error, %d for drm dev, %s",
 			 adev->asic_reset_res, adev_to_drm(adev)->unique);
 	amdgpu_put_xgmi_hive(hive);
 }
@@ -3304,76 +4358,155 @@ static int amdgpu_device_get_job_timeout_settings(struct amdgpu_device *adev)
 	long timeout;
 	int ret = 0;
 
-	/*
-	 * By default timeout for non compute jobs is 10000
-	 * and 60000 for compute jobs.
-	 * In SR-IOV or passthrough mode, timeout for compute
-	 * jobs are 60000 by default.
-	 */
-	adev->gfx_timeout = msecs_to_jiffies(10000);
-	adev->sdma_timeout = adev->video_timeout = adev->gfx_timeout;
-	if (amdgpu_sriov_vf(adev))
-		adev->compute_timeout = amdgpu_sriov_is_pp_one_vf(adev) ?
-					msecs_to_jiffies(60000) : msecs_to_jiffies(10000);
-	else
-		adev->compute_timeout =  msecs_to_jiffies(60000);
+	/* By default timeout for all queues is 2 sec */
+	adev->gfx_timeout = adev->compute_timeout = adev->sdma_timeout =
+		adev->video_timeout = msecs_to_jiffies(2000);
 
-	if (strnlen(input, AMDGPU_MAX_TIMEOUT_PARAM_LENGTH)) {
-		while ((timeout_setting = strsep(&input, ",")) &&
-				strnlen(timeout_setting, AMDGPU_MAX_TIMEOUT_PARAM_LENGTH)) {
-			ret = kstrtol(timeout_setting, 0, &timeout);
-			if (ret)
-				return ret;
+	if (!strnlen(input, AMDGPU_MAX_TIMEOUT_PARAM_LENGTH))
+		return 0;
 
-			if (timeout == 0) {
-				index++;
-				continue;
-			} else if (timeout < 0) {
-				timeout = MAX_SCHEDULE_TIMEOUT;
-				dev_warn(adev->dev, "lockup timeout disabled");
-				add_taint(TAINT_SOFTLOCKUP, LOCKDEP_STILL_OK);
-			} else {
-				timeout = msecs_to_jiffies(timeout);
-			}
+	while ((timeout_setting = strsep(&input, ",")) &&
+	       strnlen(timeout_setting, AMDGPU_MAX_TIMEOUT_PARAM_LENGTH)) {
+		ret = kstrtol(timeout_setting, 0, &timeout);
+		if (ret)
+			return ret;
 
-			switch (index++) {
-			case 0:
-				adev->gfx_timeout = timeout;
-				break;
-			case 1:
-				adev->compute_timeout = timeout;
-				break;
-			case 2:
-				adev->sdma_timeout = timeout;
-				break;
-			case 3:
-				adev->video_timeout = timeout;
-				break;
-			default:
-				break;
-			}
+		if (timeout == 0) {
+			index++;
+			continue;
+		} else if (timeout < 0) {
+			timeout = MAX_SCHEDULE_TIMEOUT;
+			dev_warn(adev->dev, "lockup timeout disabled");
+			add_taint(TAINT_SOFTLOCKUP, LOCKDEP_STILL_OK);
+		} else {
+			timeout = msecs_to_jiffies(timeout);
 		}
-		/*
-		 * There is only one value specified and
-		 * it should apply to all non-compute jobs.
-		 */
-		if (index == 1) {
-			adev->sdma_timeout = adev->video_timeout = adev->gfx_timeout;
-			if (amdgpu_sriov_vf(adev) || amdgpu_passthrough(adev))
-				adev->compute_timeout = adev->gfx_timeout;
+
+		switch (index++) {
+		case 0:
+			adev->gfx_timeout = timeout;
+			break;
+		case 1:
+			adev->compute_timeout = timeout;
+			break;
+		case 2:
+			adev->sdma_timeout = timeout;
+			break;
+		case 3:
+			adev->video_timeout = timeout;
+			break;
+		default:
+			break;
 		}
 	}
 
+	/* When only one value specified apply it to all queues. */
+	if (index == 1)
+		adev->gfx_timeout = adev->compute_timeout = adev->sdma_timeout =
+			adev->video_timeout = timeout;
+
 	return ret;
 }
 
-static const struct attribute *amdgpu_dev_attributes[] = {
-	&dev_attr_product_name.attr,
-	&dev_attr_product_number.attr,
-	&dev_attr_serial_number.attr,
-	&dev_attr_pcie_replay_count.attr,
-	NULL
-};
+/**
+ * amdgpu_device_check_iommu_direct_map - check if RAM direct mapped to GPU
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * RAM direct mapped to GPU if IOMMU is not enabled or is pass through mode
+ */
+static void amdgpu_device_check_iommu_direct_map(struct amdgpu_device *adev)
+{
+	struct iommu_domain *domain;
+
+	domain = iommu_get_domain_for_dev(adev->dev);
+	if (!domain || domain->type == IOMMU_DOMAIN_IDENTITY)
+		adev->ram_is_direct_mapped = true;
+}
+
+#if defined(CONFIG_HSA_AMD_P2P)
+/**
+ * amdgpu_device_check_iommu_remap - Check if DMA remapping is enabled.
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * return if IOMMU remapping bar address
+ */
+static bool amdgpu_device_check_iommu_remap(struct amdgpu_device *adev)
+{
+	struct iommu_domain *domain;
+
+	domain = iommu_get_domain_for_dev(adev->dev);
+	if (domain && (domain->type == IOMMU_DOMAIN_DMA ||
+		domain->type ==	IOMMU_DOMAIN_DMA_FQ))
+		return true;
+
+	return false;
+}
+#endif
+
+static void amdgpu_device_set_mcbp(struct amdgpu_device *adev)
+{
+	if (amdgpu_mcbp == 1)
+		adev->gfx.mcbp = true;
+	else if (amdgpu_mcbp == 0)
+		adev->gfx.mcbp = false;
+
+	if (amdgpu_sriov_vf(adev))
+		adev->gfx.mcbp = true;
+
+	if (adev->gfx.mcbp)
+		dev_info(adev->dev, "MCBP is enabled\n");
+}
+
+static int amdgpu_device_sys_interface_init(struct amdgpu_device *adev)
+{
+	int r;
+
+	r = amdgpu_atombios_sysfs_init(adev);
+	if (r)
+		drm_err(&adev->ddev,
+			"registering atombios sysfs failed (%d).\n", r);
+
+	r = amdgpu_pm_sysfs_init(adev);
+	if (r)
+		dev_err(adev->dev, "registering pm sysfs failed (%d).\n", r);
+
+	r = amdgpu_ucode_sysfs_init(adev);
+	if (r) {
+		adev->ucode_sysfs_en = false;
+		dev_err(adev->dev, "Creating firmware sysfs failed (%d).\n", r);
+	} else
+		adev->ucode_sysfs_en = true;
+
+	r = amdgpu_device_attr_sysfs_init(adev);
+	if (r)
+		dev_err(adev->dev, "Could not create amdgpu device attr\n");
+
+	r = devm_device_add_group(adev->dev, &amdgpu_board_attrs_group);
+	if (r)
+		dev_err(adev->dev,
+			"Could not create amdgpu board attributes\n");
+
+	amdgpu_fru_sysfs_init(adev);
+	amdgpu_reg_state_sysfs_init(adev);
+	amdgpu_xcp_sysfs_init(adev);
+
+	return r;
+}
+
+static void amdgpu_device_sys_interface_fini(struct amdgpu_device *adev)
+{
+	if (adev->pm.sysfs_initialized)
+		amdgpu_pm_sysfs_fini(adev);
+	if (adev->ucode_sysfs_en)
+		amdgpu_ucode_sysfs_fini(adev);
+	amdgpu_device_attr_sysfs_fini(adev);
+	amdgpu_fru_sysfs_fini(adev);
+
+	amdgpu_reg_state_sysfs_fini(adev);
+	amdgpu_xcp_sysfs_fini(adev);
+}
 
 /**
  * amdgpu_device_init - initialize the driver
@@ -3388,11 +4521,11 @@ static const struct attribute *amdgpu_dev_attributes[] = {
 int amdgpu_device_init(struct amdgpu_device *adev,
 		       uint32_t flags)
 {
-	struct drm_device *ddev = adev_to_drm(adev);
 	struct pci_dev *pdev = adev->pdev;
 	int r, i;
 	bool px = false;
 	u32 max_MBps;
+	int tmp;
 
 	adev->shutdown = false;
 	adev->flags = flags;
@@ -3408,6 +4541,7 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 	adev->gmc.gart_size = 512 * 1024 * 1024;
 	adev->accel_working = false;
 	adev->num_rings = 0;
+	RCU_INIT_POINTER(adev->gang_submit, dma_fence_get_stub());
 	adev->mman.buffer_funcs = NULL;
 	adev->mman.buffer_funcs_ring = NULL;
 	adev->vm_manager.vm_pte_funcs = NULL;
@@ -3421,10 +4555,14 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 	adev->smc_wreg = &amdgpu_invalid_wreg;
 	adev->pcie_rreg = &amdgpu_invalid_rreg;
 	adev->pcie_wreg = &amdgpu_invalid_wreg;
+	adev->pcie_rreg_ext = &amdgpu_invalid_rreg_ext;
+	adev->pcie_wreg_ext = &amdgpu_invalid_wreg_ext;
 	adev->pciep_rreg = &amdgpu_invalid_rreg;
 	adev->pciep_wreg = &amdgpu_invalid_wreg;
 	adev->pcie_rreg64 = &amdgpu_invalid_rreg64;
 	adev->pcie_wreg64 = &amdgpu_invalid_wreg64;
+	adev->pcie_rreg64_ext = &amdgpu_invalid_rreg64_ext;
+	adev->pcie_wreg64_ext = &amdgpu_invalid_wreg64_ext;
 	adev->uvd_ctx_rreg = &amdgpu_invalid_rreg;
 	adev->uvd_ctx_wreg = &amdgpu_invalid_wreg;
 	adev->didt_rreg = &amdgpu_invalid_rreg;
@@ -3434,30 +4572,43 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 	adev->audio_endpt_rreg = &amdgpu_block_invalid_rreg;
 	adev->audio_endpt_wreg = &amdgpu_block_invalid_wreg;
 
-	DRM_INFO("initializing kernel modesetting (%s 0x%04X:0x%04X 0x%04X:0x%04X 0x%02X).\n",
-		 amdgpu_asic_name[adev->asic_type], pdev->vendor, pdev->device,
-		 pdev->subsystem_vendor, pdev->subsystem_device, pdev->revision);
+	dev_info(
+		adev->dev,
+		"initializing kernel modesetting (%s 0x%04X:0x%04X 0x%04X:0x%04X 0x%02X).\n",
+		amdgpu_asic_name[adev->asic_type], pdev->vendor, pdev->device,
+		pdev->subsystem_vendor, pdev->subsystem_device, pdev->revision);
 
 	/* mutex initialization are all done here so we
-	 * can recall function without having locking issues */
+	 * can recall function without having locking issues
+	 */
 	mutex_init(&adev->firmware.mutex);
 	mutex_init(&adev->pm.mutex);
 	mutex_init(&adev->gfx.gpu_clock_mutex);
 	mutex_init(&adev->srbm_mutex);
 	mutex_init(&adev->gfx.pipe_reserve_mutex);
 	mutex_init(&adev->gfx.gfx_off_mutex);
+	mutex_init(&adev->gfx.partition_mutex);
 	mutex_init(&adev->grbm_idx_mutex);
 	mutex_init(&adev->mn_lock);
 	mutex_init(&adev->virt.vf_errors.lock);
 	hash_init(adev->mn_hash);
-	atomic_set(&adev->in_gpu_reset, 0);
-	init_rwsem(&adev->reset_sem);
 	mutex_init(&adev->psp.mutex);
 	mutex_init(&adev->notifier_lock);
-
-	r = amdgpu_device_init_apu_flags(adev);
-	if (r)
-		return r;
+	mutex_init(&adev->pm.stable_pstate_ctx_lock);
+	mutex_init(&adev->benchmark_mutex);
+	mutex_init(&adev->gfx.reset_sem_mutex);
+	/* Initialize the mutex for cleaner shader isolation between GFX and compute processes */
+	mutex_init(&adev->enforce_isolation_mutex);
+	for (i = 0; i < MAX_XCP; ++i) {
+		adev->isolation[i].spearhead = dma_fence_get_stub();
+		amdgpu_sync_create(&adev->isolation[i].active);
+		amdgpu_sync_create(&adev->isolation[i].prev);
+	}
+	mutex_init(&adev->gfx.userq_sch_mutex);
+	mutex_init(&adev->gfx.workload_profile_mutex);
+	mutex_init(&adev->vcn.workload_profile_mutex);
+
+	amdgpu_device_init_apu_flags(adev);
 
 	r = amdgpu_device_check_arguments(adev);
 	if (r)
@@ -3472,20 +4623,45 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 	spin_lock_init(&adev->se_cac_idx_lock);
 	spin_lock_init(&adev->audio_endpt_idx_lock);
 	spin_lock_init(&adev->mm_stats.lock);
+	spin_lock_init(&adev->virt.rlcg_reg_lock);
+	spin_lock_init(&adev->wb.lock);
 
-	INIT_LIST_HEAD(&adev->shadow_list);
-	mutex_init(&adev->shadow_list_lock);
+	xa_init_flags(&adev->userq_xa, XA_FLAGS_LOCK_IRQ);
 
 	INIT_LIST_HEAD(&adev->reset_list);
 
+	INIT_LIST_HEAD(&adev->ras_list);
+
+	INIT_LIST_HEAD(&adev->pm.od_kobj_list);
+
+	xa_init(&adev->userq_doorbell_xa);
+
 	INIT_DELAYED_WORK(&adev->delayed_init_work,
 			  amdgpu_device_delayed_init_work_handler);
 	INIT_DELAYED_WORK(&adev->gfx.gfx_off_delay_work,
 			  amdgpu_device_delay_enable_gfx_off);
+	/*
+	 * Initialize the enforce_isolation work structures for each XCP
+	 * partition.  This work handler is responsible for enforcing shader
+	 * isolation on AMD GPUs.  It counts the number of emitted fences for
+	 * each GFX and compute ring.  If there are any fences, it schedules
+	 * the `enforce_isolation_work` to be run after a delay.  If there are
+	 * no fences, it signals the Kernel Fusion Driver (KFD) to resume the
+	 * runqueue.
+	 */
+	for (i = 0; i < MAX_XCP; i++) {
+		INIT_DELAYED_WORK(&adev->gfx.enforce_isolation[i].work,
+				  amdgpu_gfx_enforce_isolation_handler);
+		adev->gfx.enforce_isolation[i].adev = adev;
+		adev->gfx.enforce_isolation[i].xcp_id = i;
+	}
 
 	INIT_WORK(&adev->xgmi_reset_work, amdgpu_device_xgmi_reset_func);
+	INIT_WORK(&adev->userq_reset_work, amdgpu_userq_reset_work);
 
 	adev->gfx.gfx_off_req_count = 1;
+	adev->gfx.gfx_off_residency = 0;
+	adev->gfx.gfx_off_entrycount = 0;
 	adev->pm.ac_power = power_supply_is_system_supplied() > 0;
 
 	atomic_set(&adev->throttling_logging_enabled, 1);
@@ -3497,6 +4673,7 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 	 * for throttling interrupt) = 60 seconds.
 	 */
 	ratelimit_state_init(&adev->throttling_logging_rs, (60 - 1) * HZ, 1);
+
 	ratelimit_set_flags(&adev->throttling_logging_rs, RATELIMIT_MSG_ON_RELEASE);
 
 	/* Registers mapping */
@@ -3513,22 +4690,27 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 		atomic_set(&adev->pm.pwr_state[i], POWER_STATE_UNKNOWN);
 
 	adev->rmmio = ioremap(adev->rmmio_base, adev->rmmio_size);
-	if (adev->rmmio == NULL) {
+	if (!adev->rmmio)
 		return -ENOMEM;
-	}
-	DRM_INFO("register mmio base: 0x%08X\n", (uint32_t)adev->rmmio_base);
-	DRM_INFO("register mmio size: %u\n", (unsigned)adev->rmmio_size);
-
-	amdgpu_device_get_pcie_info(adev);
 
-	if (amdgpu_mcbp)
-		DRM_INFO("MCBP is enabled\n");
+	dev_info(adev->dev, "register mmio base: 0x%08X\n",
+		 (uint32_t)adev->rmmio_base);
+	dev_info(adev->dev, "register mmio size: %u\n",
+		 (unsigned int)adev->rmmio_size);
 
-	if (amdgpu_mes && adev->asic_type >= CHIP_NAVI10)
-		adev->enable_mes = true;
+	/*
+	 * Reset domain needs to be present early, before XGMI hive discovered
+	 * (if any) and initialized to use reset sem and in_gpu reset flag
+	 * early on during init and before calling to RREG32.
+	 */
+	adev->reset_domain = amdgpu_reset_create_reset_domain(SINGLE_DEVICE, "amdgpu-reset-dev");
+	if (!adev->reset_domain)
+		return -ENOMEM;
 
 	/* detect hw virtualization here */
-	amdgpu_detect_virtualization(adev);
+	amdgpu_virt_init(adev);
+
+	amdgpu_device_get_pcie_info(adev);
 
 	r = amdgpu_device_get_job_timeout_settings(adev);
 	if (r) {
@@ -3536,26 +4718,74 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 		return r;
 	}
 
+	amdgpu_device_set_mcbp(adev);
+
+	/*
+	 * By default, use default mode where all blocks are expected to be
+	 * initialized. At present a 'swinit' of blocks is required to be
+	 * completed before the need for a different level is detected.
+	 */
+	amdgpu_set_init_level(adev, AMDGPU_INIT_LEVEL_DEFAULT);
 	/* early init functions */
 	r = amdgpu_device_ip_early_init(adev);
 	if (r)
 		return r;
 
+	/*
+	 * No need to remove conflicting FBs for non-display class devices.
+	 * This prevents the sysfb from being freed accidently.
+	 */
+	if ((pdev->class >> 8) == PCI_CLASS_DISPLAY_VGA ||
+	    (pdev->class >> 8) == PCI_CLASS_DISPLAY_OTHER) {
+		/* Get rid of things like offb */
+		r = aperture_remove_conflicting_pci_devices(adev->pdev, amdgpu_kms_driver.name);
+		if (r)
+			return r;
+	}
+
+	/* Enable TMZ based on IP_VERSION */
+	amdgpu_gmc_tmz_set(adev);
+
+	if (amdgpu_sriov_vf(adev) &&
+	    amdgpu_ip_version(adev, GC_HWIP, 0) >= IP_VERSION(10, 3, 0))
+		/* VF MMIO access (except mailbox range) from CPU
+		 * will be blocked during sriov runtime
+		 */
+		adev->virt.caps |= AMDGPU_VF_MMIO_ACCESS_PROTECT;
+
+	amdgpu_gmc_noretry_set(adev);
+	/* Need to get xgmi info early to decide the reset behavior*/
+	if (adev->gmc.xgmi.supported) {
+		r = adev->gfxhub.funcs->get_xgmi_info(adev);
+		if (r)
+			return r;
+	}
+
 	/* enable PCIE atomic ops */
-	if (amdgpu_sriov_vf(adev))
-		adev->have_atomics_support = ((struct amd_sriov_msg_pf2vf_info *)
-			adev->virt.fw_reserve.p_pf2vf)->pcie_atomic_ops_enabled_flags ==
-			(PCI_EXP_DEVCAP2_ATOMIC_COMP32 | PCI_EXP_DEVCAP2_ATOMIC_COMP64);
-	else
+	if (amdgpu_sriov_vf(adev)) {
+		if (adev->virt.fw_reserve.p_pf2vf)
+			adev->have_atomics_support = ((struct amd_sriov_msg_pf2vf_info *)
+						      adev->virt.fw_reserve.p_pf2vf)->pcie_atomic_ops_support_flags ==
+				(PCI_EXP_DEVCAP2_ATOMIC_COMP32 | PCI_EXP_DEVCAP2_ATOMIC_COMP64);
+	/* APUs w/ gfx9 onwards doesn't reply on PCIe atomics, rather it is a
+	 * internal path natively support atomics, set have_atomics_support to true.
+	 */
+	} else if ((adev->flags & AMD_IS_APU) &&
+		   (amdgpu_ip_version(adev, GC_HWIP, 0) >
+		    IP_VERSION(9, 0, 0))) {
+		adev->have_atomics_support = true;
+	} else {
 		adev->have_atomics_support =
 			!pci_enable_atomic_ops_to_root(adev->pdev,
 					  PCI_EXP_DEVCAP2_ATOMIC_COMP32 |
 					  PCI_EXP_DEVCAP2_ATOMIC_COMP64);
+	}
+
 	if (!adev->have_atomics_support)
 		dev_info(adev->dev, "PCIE atomic ops is not supported\n");
 
 	/* doorbell bar mapping and doorbell index init*/
-	amdgpu_device_doorbell_init(adev);
+	amdgpu_doorbell_init(adev);
 
 	if (amdgpu_emu_mode == 1) {
 		/* post the asic on emulation mode */
@@ -3566,7 +4796,8 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 	amdgpu_reset_init(adev);
 
 	/* detect if we are with an SRIOV vbios */
-	amdgpu_device_detect_sriov_bios(adev);
+	if (adev->bios)
+		amdgpu_device_detect_sriov_bios(adev);
 
 	/* check if we need to reset the asic
 	 *  E.g., driver was not cleanly unloaded previously, etc.
@@ -3574,30 +4805,26 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 	if (!amdgpu_sriov_vf(adev) && amdgpu_asic_need_reset_on_init(adev)) {
 		if (adev->gmc.xgmi.num_physical_nodes) {
 			dev_info(adev->dev, "Pending hive reset.\n");
-			adev->gmc.xgmi.pending_reset = true;
-			/* Only need to init necessary block for SMU to handle the reset */
-			for (i = 0; i < adev->num_ip_blocks; i++) {
-				if (!adev->ip_blocks[i].status.valid)
-					continue;
-				if (!(adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_GMC ||
-				      adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_COMMON ||
-				      adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_IH ||
-				      adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_SMC)) {
-					DRM_DEBUG("IP %s disabled for hw_init.\n",
-						adev->ip_blocks[i].version->funcs->name);
-					adev->ip_blocks[i].status.hw = true;
-				}
-			}
+			amdgpu_set_init_level(adev,
+					      AMDGPU_INIT_LEVEL_MINIMAL_XGMI);
+		} else if (amdgpu_ip_version(adev, MP1_HWIP, 0) == IP_VERSION(13, 0, 10) &&
+				   !amdgpu_device_has_display_hardware(adev)) {
+					r = psp_gpu_reset(adev);
 		} else {
-			r = amdgpu_asic_reset(adev);
-			if (r) {
-				dev_err(adev->dev, "asic reset on init failed\n");
-				goto failed;
-			}
+				tmp = amdgpu_reset_method;
+				/* It should do a default reset when loading or reloading the driver,
+				 * regardless of the module parameter reset_method.
+				 */
+				amdgpu_reset_method = AMD_RESET_METHOD_NONE;
+				r = amdgpu_asic_reset(adev);
+				amdgpu_reset_method = tmp;
 		}
-	}
 
-	pci_enable_pcie_error_reporting(adev->pdev);
+		if (r) {
+		  dev_err(adev->dev, "asic reset on init failed\n");
+		  goto failed;
+		}
+	}
 
 	/* Post card if necessary */
 	if (amdgpu_device_need_post(adev)) {
@@ -3606,7 +4833,7 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 			r = -EINVAL;
 			goto failed;
 		}
-		DRM_INFO("GPU posting now...\n");
+		dev_info(adev->dev, "GPU posting now...\n");
 		r = amdgpu_device_asic_init(adev);
 		if (r) {
 			dev_err(adev->dev, "gpu post error!\n");
@@ -3614,25 +4841,26 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 		}
 	}
 
-	if (adev->is_atom_fw) {
-		/* Initialize clocks */
-		r = amdgpu_atomfirmware_get_clock_info(adev);
-		if (r) {
-			dev_err(adev->dev, "amdgpu_atomfirmware_get_clock_info failed\n");
-			amdgpu_vf_error_put(adev, AMDGIM_ERROR_VF_ATOMBIOS_GET_CLOCK_FAIL, 0, 0);
-			goto failed;
-		}
-	} else {
-		/* Initialize clocks */
-		r = amdgpu_atombios_get_clock_info(adev);
-		if (r) {
-			dev_err(adev->dev, "amdgpu_atombios_get_clock_info failed\n");
-			amdgpu_vf_error_put(adev, AMDGIM_ERROR_VF_ATOMBIOS_GET_CLOCK_FAIL, 0, 0);
-			goto failed;
+	if (adev->bios) {
+		if (adev->is_atom_fw) {
+			/* Initialize clocks */
+			r = amdgpu_atomfirmware_get_clock_info(adev);
+			if (r) {
+				dev_err(adev->dev, "amdgpu_atomfirmware_get_clock_info failed\n");
+				amdgpu_vf_error_put(adev, AMDGIM_ERROR_VF_ATOMBIOS_GET_CLOCK_FAIL, 0, 0);
+				goto failed;
+			}
+		} else {
+			/* Initialize clocks */
+			r = amdgpu_atombios_get_clock_info(adev);
+			if (r) {
+				dev_err(adev->dev, "amdgpu_atombios_get_clock_info failed\n");
+				amdgpu_vf_error_put(adev, AMDGIM_ERROR_VF_ATOMBIOS_GET_CLOCK_FAIL, 0, 0);
+				goto failed;
+			}
+			/* init i2c buses */
+			amdgpu_i2c_init(adev);
 		}
-		/* init i2c buses */
-		if (!amdgpu_device_has_dc_support(adev))
-			amdgpu_atombios_i2c_init(adev);
 	}
 
 fence_driver_init:
@@ -3649,18 +4877,6 @@ fence_driver_init:
 
 	r = amdgpu_device_ip_init(adev);
 	if (r) {
-		/* failed in exclusive mode due to timeout */
-		if (amdgpu_sriov_vf(adev) &&
-		    !amdgpu_sriov_runtime(adev) &&
-		    amdgpu_virt_mmio_blocked(adev) &&
-		    !amdgpu_virt_wait_reset(adev)) {
-			dev_err(adev->dev, "VF exclusive mode timeout\n");
-			/* Don't send request since VF is inactive. */
-			adev->virt.caps &= ~AMDGPU_SRIOV_CAPS_RUNTIME;
-			adev->virt.ops = NULL;
-			r = -EAGAIN;
-			goto release_ras_con;
-		}
 		dev_err(adev->dev, "amdgpu_device_ip_init failed\n");
 		amdgpu_vf_error_put(adev, AMDGIM_ERROR_VF_AMDGPU_INIT_FAIL, 0, 0);
 		goto release_ras_con;
@@ -3687,35 +4903,6 @@ fence_driver_init:
 	/* Get a log2 for easy divisions. */
 	adev->mm_stats.log2_max_MBps = ilog2(max(1u, max_MBps));
 
-	amdgpu_fbdev_init(adev);
-
-	r = amdgpu_pm_sysfs_init(adev);
-	if (r) {
-		adev->pm_sysfs_en = false;
-		DRM_ERROR("registering pm debugfs failed (%d).\n", r);
-	} else
-		adev->pm_sysfs_en = true;
-
-	r = amdgpu_ucode_sysfs_init(adev);
-	if (r) {
-		adev->ucode_sysfs_en = false;
-		DRM_ERROR("Creating firmware sysfs failed (%d).\n", r);
-	} else
-		adev->ucode_sysfs_en = true;
-
-	if ((amdgpu_testing & 1)) {
-		if (adev->accel_working)
-			amdgpu_test_moves(adev);
-		else
-			DRM_INFO("amdgpu: acceleration disabled, skipping move tests\n");
-	}
-	if (amdgpu_benchmarking) {
-		if (adev->accel_working)
-			amdgpu_benchmark(adev, amdgpu_benchmarking);
-		else
-			DRM_INFO("amdgpu: acceleration disabled, skipping benchmarks\n");
-	}
-
 	/*
 	 * Register gpu instance before amdgpu_device_enable_mgpu_fan_boost.
 	 * Otherwise the mgpu fan boost feature will be skipped due to the
@@ -3726,7 +4913,7 @@ fence_driver_init:
 	/* enable clockgating, etc. after ib tests, etc. since some blocks require
 	 * explicit gating rather than handling it automatically.
 	 */
-	if (!adev->gmc.xgmi.pending_reset) {
+	if (adev->init_lvl->level != AMDGPU_INIT_LEVEL_MINIMAL_XGMI) {
 		r = amdgpu_device_ip_late_init(adev);
 		if (r) {
 			dev_err(adev->dev, "amdgpu_device_ip_late_init failed\n");
@@ -3739,12 +4926,19 @@ fence_driver_init:
 				   msecs_to_jiffies(AMDGPU_RESUME_MS));
 	}
 
-	if (amdgpu_sriov_vf(adev))
+	if (amdgpu_sriov_vf(adev)) {
+		amdgpu_virt_release_full_gpu(adev, true);
 		flush_delayed_work(&adev->delayed_init_work);
+	}
 
-	r = sysfs_create_files(&adev->dev->kobj, amdgpu_dev_attributes);
-	if (r)
-		dev_err(adev->dev, "Could not create amdgpu device attr\n");
+	if (adev->init_lvl->level == AMDGPU_INIT_LEVEL_MINIMAL_XGMI)
+		amdgpu_xgmi_reset_on_init(adev);
+	/*
+	 * Place those sysfs registering after `late_init`. As some of those
+	 * operations performed in `late_init` might affect the sysfs
+	 * interfaces creating.
+	 */
+	r = amdgpu_device_sys_interface_init(adev);
 
 	if (IS_ENABLED(CONFIG_PERF_EVENTS))
 		r = amdgpu_pmu_init(adev);
@@ -3757,24 +4951,45 @@ fence_driver_init:
 
 	/* if we have > 1 VGA cards, then disable the amdgpu VGA resources */
 	/* this will fail for cards that aren't VGA class devices, just
-	 * ignore it */
+	 * ignore it
+	 */
 	if ((adev->pdev->class >> 8) == PCI_CLASS_DISPLAY_VGA)
 		vga_client_register(adev->pdev, amdgpu_device_vga_set_decode);
 
-	if (amdgpu_device_supports_px(ddev)) {
-		px = true;
+	px = amdgpu_device_supports_px(adev);
+
+	if (px || (!dev_is_removable(&adev->pdev->dev) &&
+				apple_gmux_detect(NULL, NULL)))
 		vga_switcheroo_register_client(adev->pdev,
 					       &amdgpu_switcheroo_ops, px);
+
+	if (px)
 		vga_switcheroo_init_domain_pm_ops(adev->dev, &adev->vga_pm_domain);
-	}
 
-	if (adev->gmc.xgmi.pending_reset)
-		queue_delayed_work(system_wq, &mgpu_info.delayed_reset_work,
-				   msecs_to_jiffies(AMDGPU_RESUME_MS));
+	amdgpu_device_check_iommu_direct_map(adev);
+
+	adev->pm_nb.notifier_call = amdgpu_device_pm_notifier;
+	r = register_pm_notifier(&adev->pm_nb);
+	if (r)
+		goto failed;
 
 	return 0;
 
 release_ras_con:
+	if (amdgpu_sriov_vf(adev))
+		amdgpu_virt_release_full_gpu(adev, true);
+
+	/* failed in exclusive mode due to timeout */
+	if (amdgpu_sriov_vf(adev) &&
+		!amdgpu_sriov_runtime(adev) &&
+		amdgpu_virt_mmio_blocked(adev) &&
+		!amdgpu_virt_wait_reset(adev)) {
+		dev_err(adev->dev, "VF exclusive mode timeout\n");
+		/* Don't send request since VF is inactive. */
+		adev->virt.caps &= ~AMDGPU_SRIOV_CAPS_RUNTIME;
+		adev->virt.ops = NULL;
+		r = -EAGAIN;
+	}
 	amdgpu_release_ras_context(adev);
 
 failed:
@@ -3785,11 +5000,12 @@ failed:
 
 static void amdgpu_device_unmap_mmio(struct amdgpu_device *adev)
 {
+
 	/* Clear all CPU mappings pointing to this device */
 	unmap_mapping_range(adev->ddev.anon_inode->i_mapping, 0, 0, 1);
 
 	/* Unmap all mapped bars - Doorbell, registers and VRAM */
-	amdgpu_device_doorbell_fini(adev);
+	amdgpu_doorbell_fini(adev);
 
 	iounmap(adev->rmmio);
 	adev->rmmio = NULL;
@@ -3798,14 +5014,14 @@ static void amdgpu_device_unmap_mmio(struct amdgpu_device *adev)
 	adev->mman.aper_base_kaddr = NULL;
 
 	/* Memory manager related */
-	if (!adev->gmc.xgmi.connected_to_cpu) {
+	if (!adev->gmc.xgmi.connected_to_cpu && !adev->gmc.is_app_apu) {
 		arch_phys_wc_del(adev->gmc.vram_mtrr);
 		arch_io_free_memtype_wc(adev->gmc.aper_base, adev->gmc.aper_size);
 	}
 }
 
 /**
- * amdgpu_device_fini - tear down the driver
+ * amdgpu_device_fini_hw - tear down the driver
  *
  * @adev: amdgpu_device pointer
  *
@@ -3816,15 +5032,16 @@ void amdgpu_device_fini_hw(struct amdgpu_device *adev)
 {
 	dev_info(adev->dev, "amdgpu: finishing device.\n");
 	flush_delayed_work(&adev->delayed_init_work);
-	if (adev->mman.initialized) {
-		flush_delayed_work(&adev->mman.bdev.wq);
-		ttm_bo_lock_delayed_workqueue(&adev->mman.bdev);
-	}
+
+	if (adev->mman.initialized)
+		drain_workqueue(adev->mman.bdev.wq);
 	adev->shutdown = true;
 
+	unregister_pm_notifier(&adev->pm_nb);
+
 	/* make sure IB test finished before entering exclusive mode
 	 * to avoid preemption on IB test
-	 * */
+	 */
 	if (amdgpu_sriov_vf(adev)) {
 		amdgpu_virt_request_full_gpu(adev, false);
 		amdgpu_virt_fini_data_exchange(adev);
@@ -3832,7 +5049,7 @@ void amdgpu_device_fini_hw(struct amdgpu_device *adev)
 
 	/* disable all interrupts */
 	amdgpu_irq_disable_all(adev);
-	if (adev->mode_info.mode_config_initialized){
+	if (adev->mode_info.mode_config_initialized) {
 		if (!drm_drv_uses_atomic_modeset(adev_to_drm(adev)))
 			drm_helper_force_disable_all(adev_to_drm(adev));
 		else
@@ -3840,58 +5057,90 @@ void amdgpu_device_fini_hw(struct amdgpu_device *adev)
 	}
 	amdgpu_fence_driver_hw_fini(adev);
 
-	if (adev->pm_sysfs_en)
-		amdgpu_pm_sysfs_fini(adev);
-	if (adev->ucode_sysfs_en)
-		amdgpu_ucode_sysfs_fini(adev);
-	sysfs_remove_files(&adev->dev->kobj, amdgpu_dev_attributes);
+	amdgpu_device_sys_interface_fini(adev);
+
+	/* disable ras feature must before hw fini */
+	amdgpu_ras_pre_fini(adev);
 
-	amdgpu_fbdev_fini(adev);
+	amdgpu_ttm_set_buffer_funcs_status(adev, false);
 
 	amdgpu_device_ip_fini_early(adev);
 
 	amdgpu_irq_fini_hw(adev);
 
-	ttm_device_clear_dma_mappings(&adev->mman.bdev);
+	if (adev->mman.initialized)
+		ttm_device_clear_dma_mappings(&adev->mman.bdev);
 
 	amdgpu_gart_dummy_page_fini(adev);
 
-	amdgpu_device_unmap_mmio(adev);
+	if (drm_dev_is_unplugged(adev_to_drm(adev)))
+		amdgpu_device_unmap_mmio(adev);
+
 }
 
 void amdgpu_device_fini_sw(struct amdgpu_device *adev)
 {
-	amdgpu_fence_driver_sw_fini(adev);
+	int i, idx;
+	bool px;
+
 	amdgpu_device_ip_fini(adev);
-	release_firmware(adev->firmware.gpu_info_fw);
-	adev->firmware.gpu_info_fw = NULL;
+	amdgpu_fence_driver_sw_fini(adev);
+	amdgpu_ucode_release(&adev->firmware.gpu_info_fw);
 	adev->accel_working = false;
+	dma_fence_put(rcu_dereference_protected(adev->gang_submit, true));
+	for (i = 0; i < MAX_XCP; ++i) {
+		dma_fence_put(adev->isolation[i].spearhead);
+		amdgpu_sync_free(&adev->isolation[i].active);
+		amdgpu_sync_free(&adev->isolation[i].prev);
+	}
 
 	amdgpu_reset_fini(adev);
 
 	/* free i2c buses */
-	if (!amdgpu_device_has_dc_support(adev))
-		amdgpu_i2c_fini(adev);
+	amdgpu_i2c_fini(adev);
+
+	if (adev->bios) {
+		if (amdgpu_emu_mode != 1)
+			amdgpu_atombios_fini(adev);
+		amdgpu_bios_release(adev);
+	}
+
+	kfree(adev->fru_info);
+	adev->fru_info = NULL;
 
-	if (amdgpu_emu_mode != 1)
-		amdgpu_atombios_fini(adev);
+	kfree(adev->xcp_mgr);
+	adev->xcp_mgr = NULL;
 
-	kfree(adev->bios);
-	adev->bios = NULL;
-	if (amdgpu_device_supports_px(adev_to_drm(adev))) {
+	px = amdgpu_device_supports_px(adev);
+
+	if (px || (!dev_is_removable(&adev->pdev->dev) &&
+				apple_gmux_detect(NULL, NULL)))
 		vga_switcheroo_unregister_client(adev->pdev);
+
+	if (px)
 		vga_switcheroo_fini_domain_pm_ops(adev->dev);
-	}
+
 	if ((adev->pdev->class >> 8) == PCI_CLASS_DISPLAY_VGA)
 		vga_client_unregister(adev->pdev);
 
+	if (drm_dev_enter(adev_to_drm(adev), &idx)) {
+
+		iounmap(adev->rmmio);
+		adev->rmmio = NULL;
+		drm_dev_exit(idx);
+	}
+
 	if (IS_ENABLED(CONFIG_PERF_EVENTS))
 		amdgpu_pmu_fini(adev);
-	if (adev->mman.discovery_bin)
+	if (adev->discovery.bin)
 		amdgpu_discovery_fini(adev);
 
-	kfree(adev->pci_state);
+	amdgpu_reset_put_reset_domain(adev->reset_domain);
+	adev->reset_domain = NULL;
 
+	kfree(adev->pci_state);
+	kfree(adev->pcie_reset_ctx.swds_pcistate);
+	kfree(adev->pcie_reset_ctx.swus_pcistate);
 }
 
 /**
@@ -3903,66 +5152,260 @@ void amdgpu_device_fini_sw(struct amdgpu_device *adev)
  * at suspend time.
  *
  */
-static void amdgpu_device_evict_resources(struct amdgpu_device *adev)
+static int amdgpu_device_evict_resources(struct amdgpu_device *adev)
 {
-	/* No need to evict vram on APUs for suspend to ram */
-	if (adev->in_s3 && (adev->flags & AMD_IS_APU))
-		return;
+	int ret;
 
-	if (amdgpu_ttm_evict_resources(adev, TTM_PL_VRAM))
-		DRM_WARN("evicting device resources failed\n");
+	/* No need to evict vram on APUs unless going to S4 */
+	if (!adev->in_s4 && (adev->flags & AMD_IS_APU))
+		return 0;
+
+	/* No need to evict when going to S5 through S4 callbacks */
+	if (system_state == SYSTEM_POWER_OFF)
+		return 0;
 
+	ret = amdgpu_ttm_evict_resources(adev, TTM_PL_VRAM);
+	if (ret) {
+		dev_warn(adev->dev, "evicting device resources failed\n");
+		return ret;
+	}
+
+	if (adev->in_s4) {
+		ret = ttm_device_prepare_hibernation(&adev->mman.bdev);
+		if (ret)
+			dev_err(adev->dev, "prepare hibernation failed, %d\n", ret);
+	}
+	return ret;
 }
 
 /*
  * Suspend & resume.
  */
 /**
+ * amdgpu_device_pm_notifier - Notification block for Suspend/Hibernate events
+ * @nb: notifier block
+ * @mode: suspend mode
+ * @data: data
+ *
+ * This function is called when the system is about to suspend or hibernate.
+ * It is used to set the appropriate flags so that eviction can be optimized
+ * in the pm prepare callback.
+ */
+static int amdgpu_device_pm_notifier(struct notifier_block *nb, unsigned long mode,
+				     void *data)
+{
+	struct amdgpu_device *adev = container_of(nb, struct amdgpu_device, pm_nb);
+
+	switch (mode) {
+	case PM_HIBERNATION_PREPARE:
+		adev->in_s4 = true;
+		break;
+	case PM_POST_HIBERNATION:
+		adev->in_s4 = false;
+		break;
+	}
+
+	return NOTIFY_DONE;
+}
+
+/**
+ * amdgpu_device_prepare - prepare for device suspend
+ *
+ * @dev: drm dev pointer
+ *
+ * Prepare to put the hw in the suspend state (all asics).
+ * Returns 0 for success or an error on failure.
+ * Called at driver suspend.
+ */
+int amdgpu_device_prepare(struct drm_device *dev)
+{
+	struct amdgpu_device *adev = drm_to_adev(dev);
+	int i, r;
+
+	if (dev->switch_power_state == DRM_SWITCH_POWER_OFF)
+		return 0;
+
+	/* Evict the majority of BOs before starting suspend sequence */
+	r = amdgpu_device_evict_resources(adev);
+	if (r)
+		return r;
+
+	flush_delayed_work(&adev->gfx.gfx_off_delay_work);
+
+	for (i = 0; i < adev->num_ip_blocks; i++) {
+		if (!adev->ip_blocks[i].status.valid)
+			continue;
+		if (!adev->ip_blocks[i].version->funcs->prepare_suspend)
+			continue;
+		r = adev->ip_blocks[i].version->funcs->prepare_suspend(&adev->ip_blocks[i]);
+		if (r)
+			return r;
+	}
+
+	return 0;
+}
+
+/**
+ * amdgpu_device_complete - complete power state transition
+ *
+ * @dev: drm dev pointer
+ *
+ * Undo the changes from amdgpu_device_prepare. This will be
+ * called on all resume transitions, including those that failed.
+ */
+void amdgpu_device_complete(struct drm_device *dev)
+{
+	struct amdgpu_device *adev = drm_to_adev(dev);
+	int i;
+
+	for (i = 0; i < adev->num_ip_blocks; i++) {
+		if (!adev->ip_blocks[i].status.valid)
+			continue;
+		if (!adev->ip_blocks[i].version->funcs->complete)
+			continue;
+		adev->ip_blocks[i].version->funcs->complete(&adev->ip_blocks[i]);
+	}
+}
+
+/**
  * amdgpu_device_suspend - initiate device suspend
  *
  * @dev: drm dev pointer
- * @fbcon : notify the fbdev of suspend
+ * @notify_clients: notify in-kernel DRM clients
  *
  * Puts the hw in the suspend state (all asics).
  * Returns 0 for success or an error on failure.
  * Called at driver suspend.
  */
-int amdgpu_device_suspend(struct drm_device *dev, bool fbcon)
+int amdgpu_device_suspend(struct drm_device *dev, bool notify_clients)
 {
 	struct amdgpu_device *adev = drm_to_adev(dev);
+	int r, rec;
 
 	if (dev->switch_power_state == DRM_SWITCH_POWER_OFF)
 		return 0;
 
 	adev->in_suspend = true;
 
-	if (amdgpu_acpi_smart_shift_update(dev, AMDGPU_SS_DEV_D3))
-		DRM_WARN("smart shift update failed\n");
+	if (amdgpu_sriov_vf(adev)) {
+		if (!adev->in_runpm)
+			amdgpu_amdkfd_suspend_process(adev);
+		amdgpu_virt_fini_data_exchange(adev);
+		r = amdgpu_virt_request_full_gpu(adev, false);
+		if (r)
+			return r;
+	}
 
-	drm_kms_helper_poll_disable(dev);
+	r = amdgpu_acpi_smart_shift_update(adev, AMDGPU_SS_DEV_D3);
+	if (r)
+		goto unwind_sriov;
 
-	if (fbcon)
-		amdgpu_fbdev_set_suspend(adev, 1);
+	if (notify_clients)
+		drm_client_dev_suspend(adev_to_drm(adev));
 
 	cancel_delayed_work_sync(&adev->delayed_init_work);
 
 	amdgpu_ras_suspend(adev);
 
-	amdgpu_device_ip_suspend_phase1(adev);
+	r = amdgpu_device_ip_suspend_phase1(adev);
+	if (r)
+		goto unwind_smartshift;
 
-	if (!adev->in_s0ix)
-		amdgpu_amdkfd_suspend(adev, adev->in_runpm);
+	amdgpu_amdkfd_suspend(adev, !amdgpu_sriov_vf(adev) && !adev->in_runpm);
+	r = amdgpu_userq_suspend(adev);
+	if (r)
+		goto unwind_ip_phase1;
+
+	r = amdgpu_device_evict_resources(adev);
+	if (r)
+		goto unwind_userq;
 
-	/* First evict vram memory */
-	amdgpu_device_evict_resources(adev);
+	amdgpu_ttm_set_buffer_funcs_status(adev, false);
 
 	amdgpu_fence_driver_hw_fini(adev);
 
-	amdgpu_device_ip_suspend_phase2(adev);
-	/* This second call to evict device resources is to evict
-	 * the gart page table using the CPU.
+	r = amdgpu_device_ip_suspend_phase2(adev);
+	if (r)
+		goto unwind_evict;
+
+	if (amdgpu_sriov_vf(adev))
+		amdgpu_virt_release_full_gpu(adev, false);
+
+	return 0;
+
+unwind_evict:
+	if (adev->mman.buffer_funcs_ring->sched.ready)
+		amdgpu_ttm_set_buffer_funcs_status(adev, true);
+	amdgpu_fence_driver_hw_init(adev);
+
+unwind_userq:
+	rec = amdgpu_userq_resume(adev);
+	if (rec) {
+		dev_warn(adev->dev, "failed to re-initialize user queues: %d\n", rec);
+		return r;
+	}
+	rec = amdgpu_amdkfd_resume(adev, !amdgpu_sriov_vf(adev) && !adev->in_runpm);
+	if (rec) {
+		dev_warn(adev->dev, "failed to re-initialize kfd: %d\n", rec);
+		return r;
+	}
+
+unwind_ip_phase1:
+	/* suspend phase 1 = resume phase 3 */
+	rec = amdgpu_device_ip_resume_phase3(adev);
+	if (rec) {
+		dev_warn(adev->dev, "failed to re-initialize IPs phase1: %d\n", rec);
+		return r;
+	}
+
+unwind_smartshift:
+	rec = amdgpu_acpi_smart_shift_update(adev, AMDGPU_SS_DEV_D0);
+	if (rec) {
+		dev_warn(adev->dev, "failed to re-update smart shift: %d\n", rec);
+		return r;
+	}
+
+	if (notify_clients)
+		drm_client_dev_resume(adev_to_drm(adev));
+
+	amdgpu_ras_resume(adev);
+
+unwind_sriov:
+	if (amdgpu_sriov_vf(adev)) {
+		rec = amdgpu_virt_request_full_gpu(adev, true);
+		if (rec) {
+			dev_warn(adev->dev, "failed to reinitialize sriov: %d\n", rec);
+			return r;
+		}
+	}
+
+	adev->in_suspend = adev->in_s0ix = adev->in_s3 = false;
+
+	return r;
+}
+
+static inline int amdgpu_virt_resume(struct amdgpu_device *adev)
+{
+	int r;
+	unsigned int prev_physical_node_id = adev->gmc.xgmi.physical_node_id;
+
+	/* During VM resume, QEMU programming of VF MSIX table (register GFXMSIX_VECT0_ADDR_LO)
+	 * may not work. The access could be blocked by nBIF protection as VF isn't in
+	 * exclusive access mode. Exclusive access is enabled now, disable/enable MSIX
+	 * so that QEMU reprograms MSIX table.
 	 */
-	amdgpu_device_evict_resources(adev);
+	amdgpu_restore_msix(adev);
+
+	r = adev->gfxhub.funcs->get_xgmi_info(adev);
+	if (r)
+		return r;
+
+	dev_info(adev->dev, "xgmi node, old id %d, new id %d\n",
+		prev_physical_node_id, adev->gmc.xgmi.physical_node_id);
+
+	adev->vm_manager.vram_base_offset = adev->gfxhub.funcs->get_mc_fb_offset(adev);
+	adev->vm_manager.vram_base_offset +=
+		adev->gmc.xgmi.physical_node_id * adev->gmc.xgmi.node_segment_size;
 
 	return 0;
 }
@@ -3971,22 +5414,34 @@ int amdgpu_device_suspend(struct drm_device *dev, bool fbcon)
  * amdgpu_device_resume - initiate device resume
  *
  * @dev: drm dev pointer
- * @fbcon : notify the fbdev of resume
+ * @notify_clients: notify in-kernel DRM clients
  *
  * Bring the hw back to operating state (all asics).
  * Returns 0 for success or an error on failure.
  * Called at driver resume.
  */
-int amdgpu_device_resume(struct drm_device *dev, bool fbcon)
+int amdgpu_device_resume(struct drm_device *dev, bool notify_clients)
 {
 	struct amdgpu_device *adev = drm_to_adev(dev);
 	int r = 0;
 
+	if (amdgpu_sriov_vf(adev)) {
+		r = amdgpu_virt_request_full_gpu(adev, true);
+		if (r)
+			return r;
+	}
+
+	if (amdgpu_virt_xgmi_migrate_enabled(adev)) {
+		r = amdgpu_virt_resume(adev);
+		if (r)
+			goto exit;
+	}
+
 	if (dev->switch_power_state == DRM_SWITCH_POWER_OFF)
 		return 0;
 
 	if (adev->in_s0ix)
-		amdgpu_gfx_state_change_set(adev, sGpuChangeState_D0Entry);
+		amdgpu_dpm_gfx_state_change(adev, sGpuChangeState_D0Entry);
 
 	/* post card */
 	if (amdgpu_device_need_post(adev)) {
@@ -3996,58 +5451,73 @@ int amdgpu_device_resume(struct drm_device *dev, bool fbcon)
 	}
 
 	r = amdgpu_device_ip_resume(adev);
+
 	if (r) {
 		dev_err(adev->dev, "amdgpu_device_ip_resume failed (%d).\n", r);
-		return r;
+		goto exit;
 	}
-	amdgpu_fence_driver_hw_init(adev);
+
+	r = amdgpu_amdkfd_resume(adev, !amdgpu_sriov_vf(adev) && !adev->in_runpm);
+	if (r)
+		goto exit;
+
+	r = amdgpu_userq_resume(adev);
+	if (r)
+		goto exit;
 
 	r = amdgpu_device_ip_late_init(adev);
 	if (r)
-		return r;
+		goto exit;
 
 	queue_delayed_work(system_wq, &adev->delayed_init_work,
 			   msecs_to_jiffies(AMDGPU_RESUME_MS));
+exit:
+	if (amdgpu_sriov_vf(adev)) {
+		amdgpu_virt_init_data_exchange(adev);
+		amdgpu_virt_release_full_gpu(adev, true);
 
-	if (!adev->in_s0ix) {
-		r = amdgpu_amdkfd_resume(adev, adev->in_runpm);
-		if (r)
-			return r;
+		if (!r && !adev->in_runpm)
+			r = amdgpu_amdkfd_resume_process(adev);
 	}
 
+	if (r)
+		return r;
+
 	/* Make sure IB tests flushed */
 	flush_delayed_work(&adev->delayed_init_work);
 
-	if (fbcon)
-		amdgpu_fbdev_set_suspend(adev, 0);
-
-	drm_kms_helper_poll_enable(dev);
+	if (notify_clients)
+		drm_client_dev_resume(adev_to_drm(adev));
 
 	amdgpu_ras_resume(adev);
 
-	/*
-	 * Most of the connector probing functions try to acquire runtime pm
-	 * refs to ensure that the GPU is powered on when connector polling is
-	 * performed. Since we're calling this from a runtime PM callback,
-	 * trying to acquire rpm refs will cause us to deadlock.
-	 *
-	 * Since we're guaranteed to be holding the rpm lock, it's safe to
-	 * temporarily disable the rpm helpers so this doesn't deadlock us.
-	 */
+	if (adev->mode_info.num_crtc) {
+		/*
+		 * Most of the connector probing functions try to acquire runtime pm
+		 * refs to ensure that the GPU is powered on when connector polling is
+		 * performed. Since we're calling this from a runtime PM callback,
+		 * trying to acquire rpm refs will cause us to deadlock.
+		 *
+		 * Since we're guaranteed to be holding the rpm lock, it's safe to
+		 * temporarily disable the rpm helpers so this doesn't deadlock us.
+		 */
 #ifdef CONFIG_PM
-	dev->dev->power.disable_depth++;
+		dev->dev->power.disable_depth++;
 #endif
-	if (!amdgpu_device_has_dc_support(adev))
-		drm_helper_hpd_irq_event(dev);
-	else
-		drm_kms_helper_hotplug_event(dev);
+		if (!adev->dc_enabled)
+			drm_helper_hpd_irq_event(dev);
+		else
+			drm_kms_helper_hotplug_event(dev);
 #ifdef CONFIG_PM
-	dev->dev->power.disable_depth--;
+		dev->dev->power.disable_depth--;
 #endif
+	}
+
+	amdgpu_vram_mgr_clear_reset_blocks(adev);
 	adev->in_suspend = false;
 
-	if (amdgpu_acpi_smart_shift_update(dev, AMDGPU_SS_DEV_D0))
-		DRM_WARN("smart shift update failed\n");
+	if (amdgpu_acpi_smart_shift_update(adev, AMDGPU_SS_DEV_D0))
+		dev_warn(adev->dev, "smart shift update failed\n");
 
 	return 0;
 }
@@ -4078,7 +5548,8 @@ static bool amdgpu_device_ip_check_soft_reset(struct amdgpu_device *adev)
 			continue;
 		if (adev->ip_blocks[i].version->funcs->check_soft_reset)
 			adev->ip_blocks[i].status.hang =
-				adev->ip_blocks[i].version->funcs->check_soft_reset(adev);
+				adev->ip_blocks[i].version->funcs->check_soft_reset(
+					&adev->ip_blocks[i]);
 		if (adev->ip_blocks[i].status.hang) {
 			dev_info(adev->dev, "IP block:%s is hung!\n", adev->ip_blocks[i].version->funcs->name);
 			asic_hang = true;
@@ -4107,7 +5578,7 @@ static int amdgpu_device_ip_pre_soft_reset(struct amdgpu_device *adev)
 			continue;
 		if (adev->ip_blocks[i].status.hang &&
 		    adev->ip_blocks[i].version->funcs->pre_soft_reset) {
-			r = adev->ip_blocks[i].version->funcs->pre_soft_reset(adev);
+			r = adev->ip_blocks[i].version->funcs->pre_soft_reset(&adev->ip_blocks[i]);
 			if (r)
 				return r;
 		}
@@ -4169,7 +5640,7 @@ static int amdgpu_device_ip_soft_reset(struct amdgpu_device *adev)
 			continue;
 		if (adev->ip_blocks[i].status.hang &&
 		    adev->ip_blocks[i].version->funcs->soft_reset) {
-			r = adev->ip_blocks[i].version->funcs->soft_reset(adev);
+			r = adev->ip_blocks[i].version->funcs->soft_reset(&adev->ip_blocks[i]);
 			if (r)
 				return r;
 		}
@@ -4198,7 +5669,7 @@ static int amdgpu_device_ip_post_soft_reset(struct amdgpu_device *adev)
 			continue;
 		if (adev->ip_blocks[i].status.hang &&
 		    adev->ip_blocks[i].version->funcs->post_soft_reset)
-			r = adev->ip_blocks[i].version->funcs->post_soft_reset(adev);
+			r = adev->ip_blocks[i].version->funcs->post_soft_reset(&adev->ip_blocks[i]);
 		if (r)
 			return r;
 	}
@@ -4207,105 +5678,44 @@ static int amdgpu_device_ip_post_soft_reset(struct amdgpu_device *adev)
 }
 
 /**
- * amdgpu_device_recover_vram - Recover some VRAM contents
- *
- * @adev: amdgpu_device pointer
- *
- * Restores the contents of VRAM buffers from the shadows in GTT.  Used to
- * restore things like GPUVM page tables after a GPU reset where
- * the contents of VRAM might be lost.
- *
- * Returns:
- * 0 on success, negative error code on failure.
- */
-static int amdgpu_device_recover_vram(struct amdgpu_device *adev)
-{
-	struct dma_fence *fence = NULL, *next = NULL;
-	struct amdgpu_bo *shadow;
-	struct amdgpu_bo_vm *vmbo;
-	long r = 1, tmo;
-
-	if (amdgpu_sriov_runtime(adev))
-		tmo = msecs_to_jiffies(8000);
-	else
-		tmo = msecs_to_jiffies(100);
-
-	dev_info(adev->dev, "recover vram bo from shadow start\n");
-	mutex_lock(&adev->shadow_list_lock);
-	list_for_each_entry(vmbo, &adev->shadow_list, shadow_list) {
-		shadow = &vmbo->bo;
-		/* No need to recover an evicted BO */
-		if (shadow->tbo.resource->mem_type != TTM_PL_TT ||
-		    shadow->tbo.resource->start == AMDGPU_BO_INVALID_OFFSET ||
-		    shadow->parent->tbo.resource->mem_type != TTM_PL_VRAM)
-			continue;
-
-		r = amdgpu_bo_restore_shadow(shadow, &next);
-		if (r)
-			break;
-
-		if (fence) {
-			tmo = dma_fence_wait_timeout(fence, false, tmo);
-			dma_fence_put(fence);
-			fence = next;
-			if (tmo == 0) {
-				r = -ETIMEDOUT;
-				break;
-			} else if (tmo < 0) {
-				r = tmo;
-				break;
-			}
-		} else {
-			fence = next;
-		}
-	}
-	mutex_unlock(&adev->shadow_list_lock);
-
-	if (fence)
-		tmo = dma_fence_wait_timeout(fence, false, tmo);
-	dma_fence_put(fence);
-
-	if (r < 0 || tmo <= 0) {
-		dev_err(adev->dev, "recover vram bo from shadow failed, r is %ld, tmo is %ld\n", r, tmo);
-		return -EIO;
-	}
-
-	dev_info(adev->dev, "recover vram bo from shadow done\n");
-	return 0;
-}
-
-
-/**
  * amdgpu_device_reset_sriov - reset ASIC for SR-IOV vf
  *
  * @adev: amdgpu_device pointer
- * @from_hypervisor: request from hypervisor
+ * @reset_context: amdgpu reset context pointer
  *
  * do VF FLR and reinitialize Asic
  * return 0 means succeeded otherwise failed
  */
 static int amdgpu_device_reset_sriov(struct amdgpu_device *adev,
-				     bool from_hypervisor)
+				     struct amdgpu_reset_context *reset_context)
 {
 	int r;
+	struct amdgpu_hive_info *hive = NULL;
 
-	amdgpu_amdkfd_pre_reset(adev);
-
-	if (from_hypervisor)
+	if (test_bit(AMDGPU_HOST_FLR, &reset_context->flags)) {
+		if (!amdgpu_ras_get_fed_status(adev))
+			amdgpu_virt_ready_to_reset(adev);
+		amdgpu_virt_wait_reset(adev);
+		clear_bit(AMDGPU_HOST_FLR, &reset_context->flags);
 		r = amdgpu_virt_request_full_gpu(adev, true);
-	else
+	} else {
 		r = amdgpu_virt_reset_gpu(adev);
+	}
 	if (r)
 		return r;
 
+	amdgpu_ras_clear_err_state(adev);
+	amdgpu_irq_gpu_reset_resume_helper(adev);
+
+	/* some sw clean up VF needs to do before recover */
+	amdgpu_virt_post_reset(adev);
+
 	/* Resume IP prior to SMC */
 	r = amdgpu_device_ip_reinit_early_sriov(adev);
 	if (r)
-		goto error;
+		return r;
 
 	amdgpu_virt_init_data_exchange(adev);
-	/* we need recover gart prior to run SMC/CP/SDMA resume */
-	amdgpu_gtt_mgr_recover(ttm_manager_type(&adev->mman.bdev, TTM_PL_TT));
 
 	r = amdgpu_device_fw_loading(adev);
 	if (r)
@@ -4314,45 +5724,64 @@ static int amdgpu_device_reset_sriov(struct amdgpu_device *adev,
 	/* now we are okay to resume SMC/CP/SDMA */
 	r = amdgpu_device_ip_reinit_late_sriov(adev);
 	if (r)
-		goto error;
+		return r;
+
+	hive = amdgpu_get_xgmi_hive(adev);
+	/* Update PSP FW topology after reset */
+	if (hive && adev->gmc.xgmi.num_physical_nodes > 1)
+		r = amdgpu_xgmi_update_topology(hive, adev);
+	if (hive)
+		amdgpu_put_xgmi_hive(hive);
+	if (r)
+		return r;
 
-	amdgpu_irq_gpu_reset_resume_helper(adev);
 	r = amdgpu_ib_ring_tests(adev);
-	amdgpu_amdkfd_post_reset(adev);
+	if (r)
+		return r;
 
-error:
-	if (!r && adev->virt.gim_feature & AMDGIM_FEATURE_GIM_FLR_VRAMLOST) {
+	if (adev->virt.gim_feature & AMDGIM_FEATURE_GIM_FLR_VRAMLOST)
 		amdgpu_inc_vram_lost(adev);
-		r = amdgpu_device_recover_vram(adev);
-	}
+
+	/* need to be called during full access so we can't do it later like
+	 * bare-metal does.
+	 */
+	amdgpu_amdkfd_post_reset(adev);
 	amdgpu_virt_release_full_gpu(adev, true);
 
-	return r;
+	/* Aldebaran and gfx_11_0_3 support ras in SRIOV, so need resume ras during reset */
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 2) ||
+	    amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3) ||
+	    amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4) ||
+	    amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 5, 0) ||
+	    amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(11, 0, 3))
+		amdgpu_ras_resume(adev);
+
+	amdgpu_virt_ras_telemetry_post_reset(adev);
+
+	return 0;
 }
 
 /**
- * amdgpu_device_has_job_running - check if there is any job in mirror list
+ * amdgpu_device_has_job_running - check if there is any unfinished job
  *
  * @adev: amdgpu_device pointer
  *
- * check if there is any job in mirror list
+ * check if there is any job running on the device when guest driver receives
+ * FLR notification from host driver. If there are still jobs running, then
+ * the guest driver will not respond the FLR reset. Instead, let the job hit
+ * the timeout and guest driver then issue the reset request.
  */
 bool amdgpu_device_has_job_running(struct amdgpu_device *adev)
 {
 	int i;
-	struct drm_sched_job *job;
 
 	for (i = 0; i < AMDGPU_MAX_RINGS; ++i) {
 		struct amdgpu_ring *ring = adev->rings[i];
 
-		if (!ring || !ring->sched.thread)
+		if (!amdgpu_ring_sched_ready(ring))
 			continue;
 
-		spin_lock(&ring->sched.job_list_lock);
-		job = list_first_entry_or_null(&ring->sched.pending_list,
-					       struct drm_sched_job, list);
-		spin_unlock(&ring->sched.job_list_lock);
-		if (job)
+		if (amdgpu_fence_count_emitted(ring))
 			return true;
 	}
 	return false;
@@ -4368,46 +5797,37 @@ bool amdgpu_device_has_job_running(struct amdgpu_device *adev)
  */
 bool amdgpu_device_should_recover_gpu(struct amdgpu_device *adev)
 {
-	if (!amdgpu_device_ip_check_soft_reset(adev)) {
-		dev_info(adev->dev, "Timeout, but no hardware hang detected.\n");
-		return false;
-	}
 
 	if (amdgpu_gpu_recovery == 0)
 		goto disabled;
 
+	/* Skip soft reset check in fatal error mode */
+	if (!amdgpu_ras_is_poison_mode_supported(adev))
+		return true;
+
 	if (amdgpu_sriov_vf(adev))
 		return true;
 
 	if (amdgpu_gpu_recovery == -1) {
 		switch (adev->asic_type) {
-		case CHIP_BONAIRE:
-		case CHIP_HAWAII:
-		case CHIP_TOPAZ:
-		case CHIP_TONGA:
-		case CHIP_FIJI:
-		case CHIP_POLARIS10:
-		case CHIP_POLARIS11:
-		case CHIP_POLARIS12:
-		case CHIP_VEGAM:
-		case CHIP_VEGA20:
-		case CHIP_VEGA10:
-		case CHIP_VEGA12:
-		case CHIP_RAVEN:
-		case CHIP_ARCTURUS:
-		case CHIP_RENOIR:
-		case CHIP_NAVI10:
-		case CHIP_NAVI14:
-		case CHIP_NAVI12:
-		case CHIP_SIENNA_CICHLID:
-		case CHIP_NAVY_FLOUNDER:
-		case CHIP_DIMGREY_CAVEFISH:
-		case CHIP_BEIGE_GOBY:
-		case CHIP_VANGOGH:
-		case CHIP_ALDEBARAN:
-			break;
-		default:
+#ifdef CONFIG_DRM_AMDGPU_SI
+		case CHIP_VERDE:
+		case CHIP_TAHITI:
+		case CHIP_PITCAIRN:
+		case CHIP_OLAND:
+		case CHIP_HAINAN:
+#endif
+#ifdef CONFIG_DRM_AMDGPU_CIK
+		case CHIP_KAVERI:
+		case CHIP_KABINI:
+		case CHIP_MULLINS:
+#endif
+		case CHIP_CARRIZO:
+		case CHIP_STONEY:
+		case CHIP_CYAN_SKILLFISH:
 			goto disabled;
+		default:
+			break;
 		}
 	}
 
@@ -4420,88 +5840,121 @@ disabled:
 
 int amdgpu_device_mode1_reset(struct amdgpu_device *adev)
 {
-        u32 i;
-        int ret = 0;
+	u32 i;
+	int ret = 0;
 
-        amdgpu_atombios_scratch_regs_engine_hung(adev, true);
+	if (adev->bios)
+		amdgpu_atombios_scratch_regs_engine_hung(adev, true);
 
-        dev_info(adev->dev, "GPU mode1 reset\n");
+	dev_info(adev->dev, "GPU mode1 reset\n");
 
-        /* disable BM */
-        pci_clear_master(adev->pdev);
+	/* Cache the state before bus master disable. The saved config space
+	 * values are used in other cases like restore after mode-2 reset.
+	 */
+	amdgpu_device_cache_pci_state(adev->pdev);
 
-        amdgpu_device_cache_pci_state(adev->pdev);
+	/* disable BM */
+	pci_clear_master(adev->pdev);
 
-        if (amdgpu_dpm_is_mode1_reset_supported(adev)) {
-                dev_info(adev->dev, "GPU smu mode1 reset\n");
-                ret = amdgpu_dpm_mode1_reset(adev);
-        } else {
-                dev_info(adev->dev, "GPU psp mode1 reset\n");
-                ret = psp_gpu_reset(adev);
-        }
+	if (amdgpu_dpm_is_mode1_reset_supported(adev)) {
+		dev_info(adev->dev, "GPU smu mode1 reset\n");
+		ret = amdgpu_dpm_mode1_reset(adev);
+	} else {
+		dev_info(adev->dev, "GPU psp mode1 reset\n");
+		ret = psp_gpu_reset(adev);
+	}
 
-        if (ret)
-                dev_err(adev->dev, "GPU mode1 reset failed\n");
+	if (ret)
+		goto mode1_reset_failed;
 
-        amdgpu_device_load_pci_state(adev->pdev);
+	amdgpu_device_load_pci_state(adev->pdev);
+	ret = amdgpu_psp_wait_for_bootloader(adev);
+	if (ret)
+		goto mode1_reset_failed;
 
-        /* wait for asic to come out of reset */
-        for (i = 0; i < adev->usec_timeout; i++) {
-                u32 memsize = adev->nbio.funcs->get_memsize(adev);
+	/* wait for asic to come out of reset */
+	for (i = 0; i < adev->usec_timeout; i++) {
+		u32 memsize = adev->nbio.funcs->get_memsize(adev);
 
-                if (memsize != 0xffffffff)
-                        break;
-                udelay(1);
-        }
+		if (memsize != 0xffffffff)
+			break;
+		udelay(1);
+	}
+
+	if (i >= adev->usec_timeout) {
+		ret = -ETIMEDOUT;
+		goto mode1_reset_failed;
+	}
+
+	if (adev->bios)
+		amdgpu_atombios_scratch_regs_engine_hung(adev, false);
+
+	return 0;
 
-        amdgpu_atombios_scratch_regs_engine_hung(adev, false);
-        return ret;
+mode1_reset_failed:
+	dev_err(adev->dev, "GPU mode1 reset failed\n");
+	return ret;
+}
+
+int amdgpu_device_link_reset(struct amdgpu_device *adev)
+{
+	int ret = 0;
+
+	dev_info(adev->dev, "GPU link reset\n");
+
+	if (!amdgpu_reset_in_dpc(adev))
+		ret = amdgpu_dpm_link_reset(adev);
+
+	if (ret)
+		goto link_reset_failed;
+
+	ret = amdgpu_psp_wait_for_bootloader(adev);
+	if (ret)
+		goto link_reset_failed;
+
+	return 0;
+
+link_reset_failed:
+	dev_err(adev->dev, "GPU link reset failed\n");
+	return ret;
 }
 
 int amdgpu_device_pre_asic_reset(struct amdgpu_device *adev,
 				 struct amdgpu_reset_context *reset_context)
 {
-	int i, j, r = 0;
+	int i, r = 0;
 	struct amdgpu_job *job = NULL;
+	struct amdgpu_device *tmp_adev = reset_context->reset_req_dev;
 	bool need_full_reset =
 		test_bit(AMDGPU_NEED_FULL_RESET, &reset_context->flags);
 
 	if (reset_context->reset_req_dev == adev)
 		job = reset_context->job;
 
-	if (amdgpu_sriov_vf(adev)) {
-		/* stop the data exchange thread */
-		amdgpu_virt_fini_data_exchange(adev);
-	}
+	if (amdgpu_sriov_vf(adev))
+		amdgpu_virt_pre_reset(adev);
+
+	amdgpu_fence_driver_isr_toggle(adev, true);
 
 	/* block all schedulers and reset given job's ring */
 	for (i = 0; i < AMDGPU_MAX_RINGS; ++i) {
 		struct amdgpu_ring *ring = adev->rings[i];
 
-		if (!ring || !ring->sched.thread)
+		if (!amdgpu_ring_sched_ready(ring))
 			continue;
 
-		/*clear job fence from fence drv to avoid force_completion
-		 *leave NULL and vm flush fence in fence drv */
-		for (j = 0; j <= ring->fence_drv.num_fences_mask; j++) {
-			struct dma_fence *old, **ptr;
-
-			ptr = &ring->fence_drv.fences[j];
-			old = rcu_dereference_protected(*ptr, 1);
-			if (old && test_bit(AMDGPU_FENCE_FLAG_EMBED_IN_JOB_BIT, &old->flags)) {
-				RCU_INIT_POINTER(*ptr, NULL);
-			}
-		}
 		/* after all hw jobs are reset, hw fence is meaningless, so force_completion */
 		amdgpu_fence_driver_force_completion(ring);
 	}
 
+	amdgpu_fence_driver_isr_toggle(adev, false);
+
 	if (job && job->vm)
 		drm_sched_increase_karma(&job->base);
 
 	r = amdgpu_reset_prepare_hwcontext(adev, reset_context);
 	/* If reset handler not implemented, continue; otherwise return */
-	if (r == -ENOSYS)
+	if (r == -EOPNOTSUPP)
 		r = 0;
 	else
 		return r;
@@ -4512,7 +5965,8 @@ int amdgpu_device_pre_asic_reset(struct amdgpu_device *adev,
 		if (!need_full_reset)
 			need_full_reset = amdgpu_device_ip_need_full_reset(adev);
 
-		if (!need_full_reset) {
+		if (!need_full_reset && amdgpu_gpu_recovery &&
+		    amdgpu_device_ip_check_soft_reset(adev)) {
 			amdgpu_device_ip_pre_soft_reset(adev);
 			r = amdgpu_device_ip_soft_reset(adev);
 			amdgpu_device_ip_post_soft_reset(adev);
@@ -4522,6 +5976,16 @@ int amdgpu_device_pre_asic_reset(struct amdgpu_device *adev,
 			}
 		}
 
+		if (!test_bit(AMDGPU_SKIP_COREDUMP, &reset_context->flags)) {
+			dev_info(tmp_adev->dev, "Dumping IP State\n");
+			/* Trigger ip dump before we reset the asic */
+			for (i = 0; i < tmp_adev->num_ip_blocks; i++)
+				if (tmp_adev->ip_blocks[i].version->funcs->dump_ip_state)
+					tmp_adev->ip_blocks[i].version->funcs
+						->dump_ip_state((void *)&tmp_adev->ip_blocks[i]);
+			dev_info(tmp_adev->dev, "Dumping IP State Completed\n");
+		}
+
 		if (need_full_reset)
 			r = amdgpu_device_ip_suspend(adev);
 		if (need_full_reset)
@@ -4534,106 +5998,78 @@ int amdgpu_device_pre_asic_reset(struct amdgpu_device *adev,
 	return r;
 }
 
-int amdgpu_do_asic_reset(struct list_head *device_list_handle,
-			 struct amdgpu_reset_context *reset_context)
+int amdgpu_device_reinit_after_reset(struct amdgpu_reset_context *reset_context)
 {
-	struct amdgpu_device *tmp_adev = NULL;
-	bool need_full_reset, skip_hw_reset, vram_lost = false;
-	int r = 0;
+	struct list_head *device_list_handle;
+	bool full_reset, vram_lost = false;
+	struct amdgpu_device *tmp_adev;
+	int r, init_level;
 
-	/* Try reset handler method first */
-	tmp_adev = list_first_entry(device_list_handle, struct amdgpu_device,
-				    reset_list);
-	r = amdgpu_reset_perform_reset(tmp_adev, reset_context);
-	/* If reset handler not implemented, continue; otherwise return */
-	if (r == -ENOSYS)
-		r = 0;
-	else
-		return r;
-
-	/* Reset handler not implemented, use the default method */
-	need_full_reset =
-		test_bit(AMDGPU_NEED_FULL_RESET, &reset_context->flags);
-	skip_hw_reset = test_bit(AMDGPU_SKIP_HW_RESET, &reset_context->flags);
-
-	/*
-	 * ASIC reset has to be done on all XGMI hive nodes ASAP
-	 * to allow proper links negotiation in FW (within 1 sec)
-	 */
-	if (!skip_hw_reset && need_full_reset) {
-		list_for_each_entry(tmp_adev, device_list_handle, reset_list) {
-			/* For XGMI run all resets in parallel to speed up the process */
-			if (tmp_adev->gmc.xgmi.num_physical_nodes > 1) {
-				tmp_adev->gmc.xgmi.pending_reset = false;
-				if (!queue_work(system_unbound_wq, &tmp_adev->xgmi_reset_work))
-					r = -EALREADY;
-			} else
-				r = amdgpu_asic_reset(tmp_adev);
+	device_list_handle = reset_context->reset_device_list;
 
-			if (r) {
-				dev_err(tmp_adev->dev, "ASIC reset failed with error, %d for drm dev, %s",
-					 r, adev_to_drm(tmp_adev)->unique);
-				break;
-			}
-		}
-
-		/* For XGMI wait for all resets to complete before proceed */
-		if (!r) {
-			list_for_each_entry(tmp_adev, device_list_handle, reset_list) {
-				if (tmp_adev->gmc.xgmi.num_physical_nodes > 1) {
-					flush_work(&tmp_adev->xgmi_reset_work);
-					r = tmp_adev->asic_reset_res;
-					if (r)
-						break;
-				}
-			}
-		}
-	}
+	if (!device_list_handle)
+		return -EINVAL;
 
-	if (!r && amdgpu_ras_intr_triggered()) {
-		list_for_each_entry(tmp_adev, device_list_handle, reset_list) {
-			if (tmp_adev->mmhub.ras_funcs &&
-			    tmp_adev->mmhub.ras_funcs->reset_ras_error_count)
-				tmp_adev->mmhub.ras_funcs->reset_ras_error_count(tmp_adev);
-		}
+	full_reset = test_bit(AMDGPU_NEED_FULL_RESET, &reset_context->flags);
 
-		amdgpu_ras_intr_cleared();
-	}
+	/**
+	 * If it's reset on init, it's default init level, otherwise keep level
+	 * as recovery level.
+	 */
+	if (reset_context->method == AMD_RESET_METHOD_ON_INIT)
+			init_level = AMDGPU_INIT_LEVEL_DEFAULT;
+	else
+			init_level = AMDGPU_INIT_LEVEL_RESET_RECOVERY;
 
+	r = 0;
 	list_for_each_entry(tmp_adev, device_list_handle, reset_list) {
-		if (need_full_reset) {
+		amdgpu_set_init_level(tmp_adev, init_level);
+		if (full_reset) {
 			/* post card */
+			amdgpu_reset_set_dpc_status(tmp_adev, false);
+			amdgpu_ras_clear_err_state(tmp_adev);
 			r = amdgpu_device_asic_init(tmp_adev);
 			if (r) {
 				dev_warn(tmp_adev->dev, "asic atom init failed!");
 			} else {
 				dev_info(tmp_adev->dev, "GPU reset succeeded, trying to resume\n");
-				r = amdgpu_amdkfd_resume_iommu(tmp_adev);
-				if (r)
-					goto out;
 
 				r = amdgpu_device_ip_resume_phase1(tmp_adev);
 				if (r)
 					goto out;
 
 				vram_lost = amdgpu_device_check_vram_lost(tmp_adev);
+
+				if (!test_bit(AMDGPU_SKIP_COREDUMP, &reset_context->flags))
+					amdgpu_coredump(tmp_adev, false, vram_lost, reset_context->job);
+
 				if (vram_lost) {
-					DRM_INFO("VRAM is lost due to GPU reset!\n");
+					dev_info(
+						tmp_adev->dev,
+						"VRAM is lost due to GPU reset!\n");
 					amdgpu_inc_vram_lost(tmp_adev);
 				}
 
-				r = amdgpu_gtt_mgr_recover(ttm_manager_type(&tmp_adev->mman.bdev, TTM_PL_TT));
-				if (r)
-					goto out;
-
 				r = amdgpu_device_fw_loading(tmp_adev);
 				if (r)
 					return r;
 
+				r = amdgpu_xcp_restore_partition_mode(
+					tmp_adev->xcp_mgr);
+				if (r)
+					goto out;
+
 				r = amdgpu_device_ip_resume_phase2(tmp_adev);
 				if (r)
 					goto out;
 
+				if (tmp_adev->mman.buffer_funcs_ring->sched.ready)
+					amdgpu_ttm_set_buffer_funcs_status(tmp_adev, true);
+
+				r = amdgpu_device_ip_resume_phase3(tmp_adev);
+				if (r)
+					goto out;
+
 				if (vram_lost)
 					amdgpu_device_fill_reset_magic(tmp_adev);
 
@@ -4651,7 +6087,11 @@ int amdgpu_do_asic_reset(struct list_head *device_list_handle,
 				if (r)
 					goto out;
 
-				amdgpu_fbdev_set_suspend(tmp_adev, 0);
+				r = amdgpu_userq_post_reset(tmp_adev, vram_lost);
+				if (r)
+					goto out;
+
+				drm_client_dev_resume(adev_to_drm(tmp_adev));
 
 				/*
 				 * The GPU enters bad state once faulty pages
@@ -4663,7 +6103,7 @@ int amdgpu_do_asic_reset(struct list_head *device_list_handle,
 				 * bad_page_threshold value to fix this once
 				 * probing driver again.
 				 */
-				if (!amdgpu_ras_eeprom_check_err_threshold(tmp_adev)) {
+				if (!amdgpu_ras_is_rma(tmp_adev)) {
 					/* must succeed. */
 					amdgpu_ras_resume(tmp_adev);
 				} else {
@@ -4681,44 +6121,111 @@ int amdgpu_do_asic_reset(struct list_head *device_list_handle,
 
 out:
 		if (!r) {
+			/* IP init is complete now, set level as default */
+			amdgpu_set_init_level(tmp_adev,
+					      AMDGPU_INIT_LEVEL_DEFAULT);
 			amdgpu_irq_gpu_reset_resume_helper(tmp_adev);
 			r = amdgpu_ib_ring_tests(tmp_adev);
 			if (r) {
 				dev_err(tmp_adev->dev, "ib ring test failed (%d).\n", r);
-				need_full_reset = true;
 				r = -EAGAIN;
 				goto end;
 			}
 		}
 
-		if (!r)
-			r = amdgpu_device_recover_vram(tmp_adev);
-		else
+		if (r)
 			tmp_adev->asic_reset_res = r;
 	}
 
 end:
-	if (need_full_reset)
+	return r;
+}
+
+int amdgpu_do_asic_reset(struct list_head *device_list_handle,
+			 struct amdgpu_reset_context *reset_context)
+{
+	struct amdgpu_device *tmp_adev = NULL;
+	bool need_full_reset, skip_hw_reset;
+	int r = 0;
+
+	/* Try reset handler method first */
+	tmp_adev = list_first_entry(device_list_handle, struct amdgpu_device,
+				    reset_list);
+
+	reset_context->reset_device_list = device_list_handle;
+	r = amdgpu_reset_perform_reset(tmp_adev, reset_context);
+	/* If reset handler not implemented, continue; otherwise return */
+	if (r == -EOPNOTSUPP)
+		r = 0;
+	else
+		return r;
+
+	/* Reset handler not implemented, use the default method */
+	need_full_reset =
+		test_bit(AMDGPU_NEED_FULL_RESET, &reset_context->flags);
+	skip_hw_reset = test_bit(AMDGPU_SKIP_HW_RESET, &reset_context->flags);
+
+	/*
+	 * ASIC reset has to be done on all XGMI hive nodes ASAP
+	 * to allow proper links negotiation in FW (within 1 sec)
+	 */
+	if (!skip_hw_reset && need_full_reset) {
+		list_for_each_entry(tmp_adev, device_list_handle, reset_list) {
+			/* For XGMI run all resets in parallel to speed up the process */
+			if (tmp_adev->gmc.xgmi.num_physical_nodes > 1) {
+				if (!queue_work(system_unbound_wq,
+						&tmp_adev->xgmi_reset_work))
+					r = -EALREADY;
+			} else
+				r = amdgpu_asic_reset(tmp_adev);
+
+			if (r) {
+				dev_err(tmp_adev->dev,
+					"ASIC reset failed with error, %d for drm dev, %s",
+					r, adev_to_drm(tmp_adev)->unique);
+				goto out;
+			}
+		}
+
+		/* For XGMI wait for all resets to complete before proceed */
+		if (!r) {
+			list_for_each_entry(tmp_adev, device_list_handle,
+					    reset_list) {
+				if (tmp_adev->gmc.xgmi.num_physical_nodes > 1) {
+					flush_work(&tmp_adev->xgmi_reset_work);
+					r = tmp_adev->asic_reset_res;
+					if (r)
+						break;
+				}
+			}
+		}
+	}
+
+	if (!r && amdgpu_ras_intr_triggered()) {
+		list_for_each_entry(tmp_adev, device_list_handle, reset_list) {
+			amdgpu_ras_reset_error_count(tmp_adev,
+						     AMDGPU_RAS_BLOCK__MMHUB);
+		}
+
+		amdgpu_ras_intr_cleared();
+	}
+
+	r = amdgpu_device_reinit_after_reset(reset_context);
+	if (r == -EAGAIN)
 		set_bit(AMDGPU_NEED_FULL_RESET, &reset_context->flags);
 	else
 		clear_bit(AMDGPU_NEED_FULL_RESET, &reset_context->flags);
+
+out:
 	return r;
 }
 
-static bool amdgpu_device_lock_adev(struct amdgpu_device *adev,
-				struct amdgpu_hive_info *hive)
+static void amdgpu_device_set_mp1_state(struct amdgpu_device *adev)
 {
-	if (atomic_cmpxchg(&adev->in_gpu_reset, 0, 1) != 0)
-		return false;
-
-	if (hive) {
-		down_write_nest_lock(&adev->reset_sem, &hive->hive_lock);
-	} else {
-		down_write(&adev->reset_sem);
-	}
 
 	switch (amdgpu_asic_reset_method(adev)) {
 	case AMD_RESET_METHOD_MODE1:
+	case AMD_RESET_METHOD_LINK:
 		adev->mp1_state = PP_MP1_STATE_SHUTDOWN;
 		break;
 	case AMD_RESET_METHOD_MODE2:
@@ -4728,56 +6235,12 @@ static bool amdgpu_device_lock_adev(struct amdgpu_device *adev,
 		adev->mp1_state = PP_MP1_STATE_NONE;
 		break;
 	}
-
-	return true;
 }
 
-static void amdgpu_device_unlock_adev(struct amdgpu_device *adev)
+static void amdgpu_device_unset_mp1_state(struct amdgpu_device *adev)
 {
 	amdgpu_vf_error_trans_all(adev);
 	adev->mp1_state = PP_MP1_STATE_NONE;
-	atomic_set(&adev->in_gpu_reset, 0);
-	up_write(&adev->reset_sem);
-}
-
-/*
- * to lockup a list of amdgpu devices in a hive safely, if not a hive
- * with multiple nodes, it will be similar as amdgpu_device_lock_adev.
- *
- * unlock won't require roll back.
- */
-static int amdgpu_device_lock_hive_adev(struct amdgpu_device *adev, struct amdgpu_hive_info *hive)
-{
-	struct amdgpu_device *tmp_adev = NULL;
-
-	if (adev->gmc.xgmi.num_physical_nodes > 1) {
-		if (!hive) {
-			dev_err(adev->dev, "Hive is NULL while device has multiple xgmi nodes");
-			return -ENODEV;
-		}
-		list_for_each_entry(tmp_adev, &hive->device_list, gmc.xgmi.head) {
-			if (!amdgpu_device_lock_adev(tmp_adev, hive))
-				goto roll_back;
-		}
-	} else if (!amdgpu_device_lock_adev(adev, hive))
-		return -EAGAIN;
-
-	return 0;
-roll_back:
-	if (!list_is_first(&tmp_adev->gmc.xgmi.head, &hive->device_list)) {
-		/*
-		 * if the lockup iteration break in the middle of a hive,
-		 * it may means there may has a race issue,
-		 * or a hive device locked up independently.
-		 * we may be in trouble and may not, so will try to roll back
-		 * the lock and give out a warnning.
-		 */
-		dev_warn(tmp_adev->dev, "Hive lock iteration broke in the middle. Rolling back to unlock");
-		list_for_each_entry_continue_reverse(tmp_adev, &hive->device_list, gmc.xgmi.head) {
-			amdgpu_device_unlock_adev(tmp_adev);
-		}
-	}
-	return -EAGAIN;
 }
 
 static void amdgpu_device_resume_display_audio(struct amdgpu_device *adev)
@@ -4790,6 +6253,8 @@ static void amdgpu_device_resume_display_audio(struct amdgpu_device *adev)
 		pm_runtime_enable(&(p->dev));
 		pm_runtime_resume(&(p->dev));
 	}
+
+	pci_dev_put(p);
 }
 
 static int amdgpu_device_suspend_display_audio(struct amdgpu_device *adev)
@@ -4828,6 +6293,7 @@ static int amdgpu_device_suspend_display_audio(struct amdgpu_device *adev)
 
 		if (expires < ktime_get_mono_fast_ns()) {
 			dev_warn(adev->dev, "failed to suspend display audio\n");
+			pci_dev_put(p);
 			/* TODO: abort the succeeding gpu reset? */
 			return -ETIMEDOUT;
 		}
@@ -4835,187 +6301,107 @@ static int amdgpu_device_suspend_display_audio(struct amdgpu_device *adev)
 
 	pm_runtime_disable(&(p->dev));
 
+	pci_dev_put(p);
 	return 0;
 }
 
-static void amdgpu_device_recheck_guilty_jobs(
-	struct amdgpu_device *adev, struct list_head *device_list_handle,
-	struct amdgpu_reset_context *reset_context)
+static inline void amdgpu_device_stop_pending_resets(struct amdgpu_device *adev)
 {
-	int i, r = 0;
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
 
-	for (i = 0; i < AMDGPU_MAX_RINGS; ++i) {
-		struct amdgpu_ring *ring = adev->rings[i];
-		int ret = 0;
-		struct drm_sched_job *s_job;
+#if defined(CONFIG_DEBUG_FS)
+	if (!amdgpu_sriov_vf(adev))
+		cancel_work(&adev->reset_work);
+#endif
+	cancel_work(&adev->userq_reset_work);
 
-		if (!ring || !ring->sched.thread)
-			continue;
+	if (adev->kfd.dev)
+		cancel_work(&adev->kfd.reset_work);
 
-		s_job = list_first_entry_or_null(&ring->sched.pending_list,
-				struct drm_sched_job, list);
-		if (s_job == NULL)
-			continue;
+	if (amdgpu_sriov_vf(adev))
+		cancel_work(&adev->virt.flr_work);
 
-		/* clear job's guilty and depend the folowing step to decide the real one */
-		drm_sched_reset_karma(s_job);
-		/* for the real bad job, it will be resubmitted twice, adding a dma_fence_get
-		 * to make sure fence is balanced */
-		dma_fence_get(s_job->s_fence->parent);
-		drm_sched_resubmit_jobs_ext(&ring->sched, 1);
-
-		ret = dma_fence_wait_timeout(s_job->s_fence->parent, false, ring->sched.timeout);
-		if (ret == 0) { /* timeout */
-			DRM_ERROR("Found the real bad job! ring:%s, job_id:%llx\n",
-						ring->sched.name, s_job->id);
-
-			/* set guilty */
-			drm_sched_increase_karma(s_job);
-retry:
-			/* do hw reset */
-			if (amdgpu_sriov_vf(adev)) {
-				amdgpu_virt_fini_data_exchange(adev);
-				r = amdgpu_device_reset_sriov(adev, false);
-				if (r)
-					adev->asic_reset_res = r;
-			} else {
-				clear_bit(AMDGPU_SKIP_HW_RESET,
-					  &reset_context->flags);
-				r = amdgpu_do_asic_reset(device_list_handle,
-							 reset_context);
-				if (r && r == -EAGAIN)
-					goto retry;
-			}
+	if (con && adev->ras_enabled)
+		cancel_work(&con->recovery_work);
 
-			/*
-			 * add reset counter so that the following
-			 * resubmitted job could flush vmid
-			 */
-			atomic_inc(&adev->gpu_reset_counter);
-			continue;
-		}
+}
 
-		/* got the hw fence, signal finished fence */
-		atomic_dec(ring->sched.score);
-		dma_fence_put(s_job->s_fence->parent);
-		dma_fence_get(&s_job->s_fence->finished);
-		dma_fence_signal(&s_job->s_fence->finished);
-		dma_fence_put(&s_job->s_fence->finished);
+static int amdgpu_device_health_check(struct list_head *device_list_handle)
+{
+	struct amdgpu_device *tmp_adev;
+	int ret = 0;
 
-		/* remove node from list and free the job */
-		spin_lock(&ring->sched.job_list_lock);
-		list_del_init(&s_job->list);
-		spin_unlock(&ring->sched.job_list_lock);
-		ring->sched.ops->free_job(s_job);
+	list_for_each_entry(tmp_adev, device_list_handle, reset_list) {
+		ret |= amdgpu_device_bus_status_check(tmp_adev);
 	}
-}
 
-/**
- * amdgpu_device_gpu_recover - reset the asic and recover scheduler
- *
- * @adev: amdgpu_device pointer
- * @job: which job trigger hang
- *
- * Attempt to reset the GPU if it has hung (all asics).
- * Attempt to do soft-reset or full-reset and reinitialize Asic
- * Returns 0 for success or an error on failure.
- */
+	return ret;
+}
 
-int amdgpu_device_gpu_recover(struct amdgpu_device *adev,
-			      struct amdgpu_job *job)
+static void amdgpu_device_recovery_prepare(struct amdgpu_device *adev,
+					  struct list_head *device_list,
+					  struct amdgpu_hive_info *hive)
 {
-	struct list_head device_list, *device_list_handle =  NULL;
-	bool job_signaled = false;
-	struct amdgpu_hive_info *hive = NULL;
 	struct amdgpu_device *tmp_adev = NULL;
-	int i, r = 0;
-	bool need_emergency_restart = false;
-	bool audio_suspended = false;
-	int tmp_vram_lost_counter;
-	struct amdgpu_reset_context reset_context;
-
-	memset(&reset_context, 0, sizeof(reset_context));
-
-	/*
-	 * Special case: RAS triggered and full reset isn't supported
-	 */
-	need_emergency_restart = amdgpu_ras_need_emergency_restart(adev);
-
-	/*
-	 * Flush RAM to disk so that after reboot
-	 * the user can read log and see why the system rebooted.
-	 */
-	if (need_emergency_restart && amdgpu_ras_get_context(adev)->reboot) {
-		DRM_WARN("Emergency reboot.");
-
-		ksys_sync_helper();
-		emergency_restart();
-	}
-
-	dev_info(adev->dev, "GPU %s begin!\n",
-		need_emergency_restart ? "jobs stop":"reset");
 
 	/*
-	 * Here we trylock to avoid chain of resets executing from
-	 * either trigger by jobs on different adevs in XGMI hive or jobs on
-	 * different schedulers for same device while this TO handler is running.
-	 * We always reset all schedulers for device and all devices for XGMI
-	 * hive so that should take care of them too.
+	 * Build list of devices to reset.
+	 * In case we are in XGMI hive mode, resort the device list
+	 * to put adev in the 1st position.
 	 */
-	hive = amdgpu_get_xgmi_hive(adev);
-	if (hive) {
-		if (atomic_cmpxchg(&hive->in_reset, 0, 1) != 0) {
-			DRM_INFO("Bailing on TDR for s_job:%llx, hive: %llx as another already in progress",
-				job ? job->base.id : -1, hive->hive_id);
-			amdgpu_put_xgmi_hive(hive);
-			if (job && job->vm)
-				drm_sched_increase_karma(&job->base);
-			return 0;
+	if (!amdgpu_sriov_vf(adev) && (adev->gmc.xgmi.num_physical_nodes > 1) && hive) {
+		list_for_each_entry(tmp_adev, &hive->device_list, gmc.xgmi.head) {
+			list_add_tail(&tmp_adev->reset_list, device_list);
+			if (adev->shutdown)
+				tmp_adev->shutdown = true;
+			if (amdgpu_reset_in_dpc(adev))
+				tmp_adev->pcie_reset_ctx.in_link_reset = true;
 		}
-		mutex_lock(&hive->hive_lock);
+		if (!list_is_first(&adev->reset_list, device_list))
+			list_rotate_to_front(&adev->reset_list, device_list);
+	} else {
+		list_add_tail(&adev->reset_list, device_list);
 	}
+}
 
-	reset_context.method = AMD_RESET_METHOD_NONE;
-	reset_context.reset_req_dev = adev;
-	reset_context.job = job;
-	reset_context.hive = hive;
-	clear_bit(AMDGPU_NEED_FULL_RESET, &reset_context.flags);
+static void amdgpu_device_recovery_get_reset_lock(struct amdgpu_device *adev,
+						  struct list_head *device_list)
+{
+	struct amdgpu_device *tmp_adev = NULL;
 
-	/*
-	 * lock the device before we try to operate the linked list
-	 * if didn't get the device lock, don't touch the linked list since
-	 * others may iterating it.
-	 */
-	r = amdgpu_device_lock_hive_adev(adev, hive);
-	if (r) {
-		dev_info(adev->dev, "Bailing on TDR for s_job:%llx, as another already in progress",
-					job ? job->base.id : -1);
+	if (list_empty(device_list))
+		return;
+	tmp_adev =
+		list_first_entry(device_list, struct amdgpu_device, reset_list);
+	amdgpu_device_lock_reset_domain(tmp_adev->reset_domain);
+}
 
-		/* even we skipped this reset, still need to set the job to guilty */
-		if (job && job->vm)
-			drm_sched_increase_karma(&job->base);
-		goto skip_recovery;
-	}
+static void amdgpu_device_recovery_put_reset_lock(struct amdgpu_device *adev,
+						  struct list_head *device_list)
+{
+	struct amdgpu_device *tmp_adev = NULL;
 
-	/*
-	 * Build list of devices to reset.
-	 * In case we are in XGMI hive mode, resort the device list
-	 * to put adev in the 1st position.
-	 */
-	INIT_LIST_HEAD(&device_list);
-	if (adev->gmc.xgmi.num_physical_nodes > 1) {
-		list_for_each_entry(tmp_adev, &hive->device_list, gmc.xgmi.head)
-			list_add_tail(&tmp_adev->reset_list, &device_list);
-		if (!list_is_first(&adev->reset_list, &device_list))
-			list_rotate_to_front(&adev->reset_list, &device_list);
-		device_list_handle = &device_list;
-	} else {
-		list_add_tail(&adev->reset_list, &device_list);
-		device_list_handle = &device_list;
-	}
+	if (list_empty(device_list))
+		return;
+	tmp_adev =
+		list_first_entry(device_list, struct amdgpu_device, reset_list);
+	amdgpu_device_unlock_reset_domain(tmp_adev->reset_domain);
+}
+
+static void amdgpu_device_halt_activities(struct amdgpu_device *adev,
+					  struct amdgpu_job *job,
+					  struct amdgpu_reset_context *reset_context,
+					  struct list_head *device_list,
+					  struct amdgpu_hive_info *hive,
+					  bool need_emergency_restart)
+{
+	struct amdgpu_device *tmp_adev = NULL;
+	int i;
 
 	/* block all schedulers and reset given job's ring */
-	list_for_each_entry(tmp_adev, device_list_handle, reset_list) {
+	list_for_each_entry(tmp_adev, device_list, reset_list) {
+		amdgpu_device_set_mp1_state(tmp_adev);
+
 		/*
 		 * Try to put the audio codec into suspend state
 		 * before gpu reset started.
@@ -5027,32 +6413,33 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev,
 		 * some audio codec errors.
 		 */
 		if (!amdgpu_device_suspend_display_audio(tmp_adev))
-			audio_suspended = true;
+			tmp_adev->pcie_reset_ctx.audio_suspended = true;
 
 		amdgpu_ras_set_error_query_ready(tmp_adev, false);
 
 		cancel_delayed_work_sync(&tmp_adev->delayed_init_work);
 
-		if (!amdgpu_sriov_vf(tmp_adev))
-			amdgpu_amdkfd_pre_reset(tmp_adev);
+		amdgpu_amdkfd_pre_reset(tmp_adev, reset_context);
 
 		/*
-		 * Mark these ASICs to be reseted as untracked first
+		 * Mark these ASICs to be reset as untracked first
 		 * And add them back after reset completed
 		 */
 		amdgpu_unregister_gpu_instance(tmp_adev);
 
-		amdgpu_fbdev_set_suspend(tmp_adev, 1);
+		drm_client_dev_suspend(adev_to_drm(tmp_adev));
 
 		/* disable ras on ALL IPs */
-		if (!need_emergency_restart &&
-		      amdgpu_device_ip_need_full_reset(tmp_adev))
+		if (!need_emergency_restart && !amdgpu_reset_in_dpc(adev) &&
+		    amdgpu_device_ip_need_full_reset(tmp_adev))
 			amdgpu_ras_suspend(tmp_adev);
 
+		amdgpu_userq_pre_reset(tmp_adev);
+
 		for (i = 0; i < AMDGPU_MAX_RINGS; ++i) {
 			struct amdgpu_ring *ring = tmp_adev->rings[i];
 
-			if (!ring || !ring->sched.thread)
+			if (!amdgpu_ring_sched_ready(ring))
 				continue;
 
 			drm_sched_stop(&ring->sched, job ? &job->base : NULL);
@@ -5062,26 +6449,19 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev,
 		}
 		atomic_inc(&tmp_adev->gpu_reset_counter);
 	}
+}
 
-	if (need_emergency_restart)
-		goto skip_sched_resume;
-
-	/*
-	 * Must check guilty signal here since after this point all old
-	 * HW fences are force signaled.
-	 *
-	 * job->base holds a reference to parent fence
-	 */
-	if (job && job->base.s_fence->parent &&
-	    dma_fence_is_signaled(job->base.s_fence->parent)) {
-		job_signaled = true;
-		dev_info(adev->dev, "Guilty job already signaled, skipping HW reset");
-		goto skip_hw_reset;
-	}
+static int amdgpu_device_asic_reset(struct amdgpu_device *adev,
+			      struct list_head *device_list,
+			      struct amdgpu_reset_context *reset_context)
+{
+	struct amdgpu_device *tmp_adev = NULL;
+	int retry_limit = AMDGPU_MAX_RETRY_LIMIT;
+	int r = 0;
 
 retry:	/* Rest of adevs pre asic reset from XGMI hive. */
-	list_for_each_entry(tmp_adev, device_list_handle, reset_list) {
-		r = amdgpu_device_pre_asic_reset(tmp_adev, &reset_context);
+	list_for_each_entry(tmp_adev, device_list, reset_list) {
+		r = amdgpu_device_pre_asic_reset(tmp_adev, reset_context);
 		/*TODO Should we stop ?*/
 		if (r) {
 			dev_err(tmp_adev->dev, "GPU pre asic reset failed with err, %d for drm dev, %s ",
@@ -5090,68 +6470,107 @@ retry:	/* Rest of adevs pre asic reset from XGMI hive. */
 		}
 	}
 
-	tmp_vram_lost_counter = atomic_read(&((adev)->vram_lost_counter));
 	/* Actual ASIC resets if needed.*/
 	/* Host driver will handle XGMI hive reset for SRIOV */
 	if (amdgpu_sriov_vf(adev)) {
-		r = amdgpu_device_reset_sriov(adev, job ? false : true);
+
+		/* Bail out of reset early */
+		if (amdgpu_ras_is_rma(adev))
+			return -ENODEV;
+
+		if (amdgpu_ras_get_fed_status(adev) || amdgpu_virt_rcvd_ras_interrupt(adev)) {
+			dev_dbg(adev->dev, "Detected RAS error, wait for FLR completion\n");
+			amdgpu_ras_set_fed(adev, true);
+			set_bit(AMDGPU_HOST_FLR, &reset_context->flags);
+		}
+
+		r = amdgpu_device_reset_sriov(adev, reset_context);
+		if (AMDGPU_RETRY_SRIOV_RESET(r) && (retry_limit--) > 0) {
+			amdgpu_virt_release_full_gpu(adev, true);
+			goto retry;
+		}
 		if (r)
 			adev->asic_reset_res = r;
 	} else {
-		r = amdgpu_do_asic_reset(device_list_handle, &reset_context);
+		r = amdgpu_do_asic_reset(device_list, reset_context);
 		if (r && r == -EAGAIN)
 			goto retry;
 	}
 
-skip_hw_reset:
-
-	/* Post ASIC reset for all devs .*/
-	list_for_each_entry(tmp_adev, device_list_handle, reset_list) {
-
+	list_for_each_entry(tmp_adev, device_list, reset_list) {
 		/*
-		 * Sometimes a later bad compute job can block a good gfx job as gfx
-		 * and compute ring share internal GC HW mutually. We add an additional
-		 * guilty jobs recheck step to find the real guilty job, it synchronously
-		 * submits and pends for the first job being signaled. If it gets timeout,
-		 * we identify it as a real guilty job.
+		 * Drop any pending non scheduler resets queued before reset is done.
+		 * Any reset scheduled after this point would be valid. Scheduler resets
+		 * were already dropped during drm_sched_stop and no new ones can come
+		 * in before drm_sched_start.
 		 */
-		if (amdgpu_gpu_recovery == 2 &&
-			!(tmp_vram_lost_counter < atomic_read(&adev->vram_lost_counter)))
-			amdgpu_device_recheck_guilty_jobs(
-				tmp_adev, device_list_handle, &reset_context);
+		amdgpu_device_stop_pending_resets(tmp_adev);
+	}
+
+	return r;
+}
+
+static int amdgpu_device_sched_resume(struct list_head *device_list,
+			      struct amdgpu_reset_context *reset_context,
+			      bool   job_signaled)
+{
+	struct amdgpu_device *tmp_adev = NULL;
+	int i, r = 0;
+
+	/* Post ASIC reset for all devs .*/
+	list_for_each_entry(tmp_adev, device_list, reset_list) {
 
 		for (i = 0; i < AMDGPU_MAX_RINGS; ++i) {
 			struct amdgpu_ring *ring = tmp_adev->rings[i];
 
-			if (!ring || !ring->sched.thread)
+			if (!amdgpu_ring_sched_ready(ring))
 				continue;
 
-			/* No point to resubmit jobs if we didn't HW reset*/
-			if (!tmp_adev->asic_reset_res && !job_signaled)
-				drm_sched_resubmit_jobs(&ring->sched);
-
-			drm_sched_start(&ring->sched, !tmp_adev->asic_reset_res);
+			drm_sched_start(&ring->sched, 0);
 		}
 
-		if (!drm_drv_uses_atomic_modeset(adev_to_drm(tmp_adev)) && !job_signaled) {
+		if (!drm_drv_uses_atomic_modeset(adev_to_drm(tmp_adev)) && !job_signaled)
 			drm_helper_resume_force_mode(adev_to_drm(tmp_adev));
-		}
-
-		tmp_adev->asic_reset_res = 0;
 
-		if (r) {
-			/* bad news, how to tell it to userspace ? */
-			dev_info(tmp_adev->dev, "GPU reset(%d) failed\n", atomic_read(&tmp_adev->gpu_reset_counter));
-			amdgpu_vf_error_put(tmp_adev, AMDGIM_ERROR_VF_GPU_RESET_FAIL, 0, r);
+		if (tmp_adev->asic_reset_res) {
+			/* bad news, how to tell it to userspace ?
+			 * for ras error, we should report GPU bad status instead of
+			 * reset failure
+			 */
+			if (reset_context->src != AMDGPU_RESET_SRC_RAS ||
+			    !amdgpu_ras_eeprom_check_err_threshold(tmp_adev))
+				dev_info(
+					tmp_adev->dev,
+					"GPU reset(%d) failed with error %d \n",
+					atomic_read(
+						&tmp_adev->gpu_reset_counter),
+					tmp_adev->asic_reset_res);
+			amdgpu_vf_error_put(tmp_adev,
+					    AMDGIM_ERROR_VF_GPU_RESET_FAIL, 0,
+					    tmp_adev->asic_reset_res);
+			if (!r)
+				r = tmp_adev->asic_reset_res;
+			tmp_adev->asic_reset_res = 0;
 		} else {
-			dev_info(tmp_adev->dev, "GPU reset(%d) succeeded!\n", atomic_read(&tmp_adev->gpu_reset_counter));
-			if (amdgpu_acpi_smart_shift_update(adev_to_drm(tmp_adev), AMDGPU_SS_DEV_D0))
-				DRM_WARN("smart shift update failed\n");
+			dev_info(tmp_adev->dev, "GPU reset(%d) succeeded!\n",
+				 atomic_read(&tmp_adev->gpu_reset_counter));
+			if (amdgpu_acpi_smart_shift_update(tmp_adev,
+							   AMDGPU_SS_DEV_D0))
+				dev_warn(tmp_adev->dev,
+					 "smart shift update failed\n");
 		}
 	}
 
-skip_sched_resume:
-	list_for_each_entry(tmp_adev, device_list_handle, reset_list) {
+	return r;
+}
+
+static void amdgpu_device_gpu_resume(struct amdgpu_device *adev,
+			      struct list_head *device_list,
+			      bool   need_emergency_restart)
+{
+	struct amdgpu_device *tmp_adev = NULL;
+
+	list_for_each_entry(tmp_adev, device_list, reset_list) {
 		/* unlock kfd: SRIOV would do it separately */
 		if (!need_emergency_restart && !amdgpu_sriov_vf(tmp_adev))
 			amdgpu_amdkfd_post_reset(tmp_adev);
@@ -5162,37 +6581,239 @@ skip_sched_resume:
 		if (!adev->kfd.init_complete)
 			amdgpu_amdkfd_device_init(adev);
 
-		if (audio_suspended)
+		if (tmp_adev->pcie_reset_ctx.audio_suspended)
 			amdgpu_device_resume_display_audio(tmp_adev);
-		amdgpu_device_unlock_adev(tmp_adev);
+
+		amdgpu_device_unset_mp1_state(tmp_adev);
+
+		amdgpu_ras_set_error_query_ready(tmp_adev, true);
+
+	}
+}
+
+
+/**
+ * amdgpu_device_gpu_recover - reset the asic and recover scheduler
+ *
+ * @adev: amdgpu_device pointer
+ * @job: which job trigger hang
+ * @reset_context: amdgpu reset context pointer
+ *
+ * Attempt to reset the GPU if it has hung (all asics).
+ * Attempt to do soft-reset or full-reset and reinitialize Asic
+ * Returns 0 for success or an error on failure.
+ */
+
+int amdgpu_device_gpu_recover(struct amdgpu_device *adev,
+			      struct amdgpu_job *job,
+			      struct amdgpu_reset_context *reset_context)
+{
+	struct list_head device_list;
+	bool job_signaled = false;
+	struct amdgpu_hive_info *hive = NULL;
+	int r = 0;
+	bool need_emergency_restart = false;
+
+	/*
+	 * If it reaches here because of hang/timeout and a RAS error is
+	 * detected at the same time, let RAS recovery take care of it.
+	 */
+	if (amdgpu_ras_is_err_state(adev, AMDGPU_RAS_BLOCK__ANY) &&
+	    !amdgpu_sriov_vf(adev) &&
+	    reset_context->src != AMDGPU_RESET_SRC_RAS) {
+		dev_dbg(adev->dev,
+			"Gpu recovery from source: %d yielding to RAS error recovery handling",
+			reset_context->src);
+		return 0;
 	}
 
-skip_recovery:
+	/*
+	 * Special case: RAS triggered and full reset isn't supported
+	 */
+	need_emergency_restart = amdgpu_ras_need_emergency_restart(adev);
+
+	/*
+	 * Flush RAM to disk so that after reboot
+	 * the user can read log and see why the system rebooted.
+	 */
+	if (need_emergency_restart && amdgpu_ras_get_context(adev) &&
+		amdgpu_ras_get_context(adev)->reboot) {
+		dev_warn(adev->dev, "Emergency reboot.");
+
+		ksys_sync_helper();
+		emergency_restart();
+	}
+
+	dev_info(adev->dev, "GPU %s begin!. Source:  %d\n",
+		 need_emergency_restart ? "jobs stop" : "reset",
+		 reset_context->src);
+
+	if (!amdgpu_sriov_vf(adev))
+		hive = amdgpu_get_xgmi_hive(adev);
+	if (hive)
+		mutex_lock(&hive->hive_lock);
+
+	reset_context->job = job;
+	reset_context->hive = hive;
+	INIT_LIST_HEAD(&device_list);
+
+	amdgpu_device_recovery_prepare(adev, &device_list, hive);
+
+	if (!amdgpu_sriov_vf(adev)) {
+		r = amdgpu_device_health_check(&device_list);
+		if (r)
+			goto end_reset;
+	}
+
+	/* Cannot be called after locking reset domain */
+	amdgpu_ras_pre_reset(adev, &device_list);
+
+	/* We need to lock reset domain only once both for XGMI and single device */
+	amdgpu_device_recovery_get_reset_lock(adev, &device_list);
+
+	amdgpu_device_halt_activities(adev, job, reset_context, &device_list,
+				      hive, need_emergency_restart);
+	if (need_emergency_restart)
+		goto skip_sched_resume;
+	/*
+	 * Must check guilty signal here since after this point all old
+	 * HW fences are force signaled.
+	 *
+	 * job->base holds a reference to parent fence
+	 */
+	if (job && dma_fence_is_signaled(&job->hw_fence->base)) {
+		job_signaled = true;
+		dev_info(adev->dev, "Guilty job already signaled, skipping HW reset");
+		goto skip_hw_reset;
+	}
+
+	r = amdgpu_device_asic_reset(adev, &device_list, reset_context);
+	if (r)
+		goto reset_unlock;
+skip_hw_reset:
+	r = amdgpu_device_sched_resume(&device_list, reset_context, job_signaled);
+	if (r)
+		goto reset_unlock;
+skip_sched_resume:
+	amdgpu_device_gpu_resume(adev, &device_list, need_emergency_restart);
+reset_unlock:
+	amdgpu_device_recovery_put_reset_lock(adev, &device_list);
+	amdgpu_ras_post_reset(adev, &device_list);
+end_reset:
 	if (hive) {
-		atomic_set(&hive->in_reset, 0);
 		mutex_unlock(&hive->hive_lock);
 		amdgpu_put_xgmi_hive(hive);
 	}
 
-	if (r && r != -EAGAIN)
+	if (r)
 		dev_info(adev->dev, "GPU reset end with ret = %d\n", r);
+
+	atomic_set(&adev->reset_domain->reset_res, r);
+
+	if (!r) {
+		struct amdgpu_task_info *ti = NULL;
+
+		if (job)
+			ti = amdgpu_vm_get_task_info_pasid(adev, job->pasid);
+
+		drm_dev_wedged_event(adev_to_drm(adev), DRM_WEDGE_RECOVERY_NONE,
+				     ti ? &ti->task : NULL);
+
+		amdgpu_vm_put_task_info(ti);
+	}
+
 	return r;
 }
 
 /**
+ * amdgpu_device_partner_bandwidth - find the bandwidth of appropriate partner
+ *
+ * @adev: amdgpu_device pointer
+ * @speed: pointer to the speed of the link
+ * @width: pointer to the width of the link
+ *
+ * Evaluate the hierarchy to find the speed and bandwidth capabilities of the
+ * first physical partner to an AMD dGPU.
+ * This will exclude any virtual switches and links.
+ */
+static void amdgpu_device_partner_bandwidth(struct amdgpu_device *adev,
+					    enum pci_bus_speed *speed,
+					    enum pcie_link_width *width)
+{
+	struct pci_dev *parent = adev->pdev;
+
+	if (!speed || !width)
+		return;
+
+	*speed = PCI_SPEED_UNKNOWN;
+	*width = PCIE_LNK_WIDTH_UNKNOWN;
+
+	if (amdgpu_device_pcie_dynamic_switching_supported(adev)) {
+		while ((parent = pci_upstream_bridge(parent))) {
+			/* skip upstream/downstream switches internal to dGPU*/
+			if (parent->vendor == PCI_VENDOR_ID_ATI)
+				continue;
+			*speed = pcie_get_speed_cap(parent);
+			*width = pcie_get_width_cap(parent);
+			break;
+		}
+	} else {
+		/* use the current speeds rather than max if switching is not supported */
+		pcie_bandwidth_available(adev->pdev, NULL, speed, width);
+	}
+}
+
+/**
+ * amdgpu_device_gpu_bandwidth - find the bandwidth of the GPU
+ *
+ * @adev: amdgpu_device pointer
+ * @speed: pointer to the speed of the link
+ * @width: pointer to the width of the link
+ *
+ * Evaluate the hierarchy to find the speed and bandwidth capabilities of the
+ * AMD dGPU which may be a virtual upstream bridge.
+ */
+static void amdgpu_device_gpu_bandwidth(struct amdgpu_device *adev,
+					enum pci_bus_speed *speed,
+					enum pcie_link_width *width)
+{
+	struct pci_dev *parent = adev->pdev;
+
+	if (!speed || !width)
+		return;
+
+	parent = pci_upstream_bridge(parent);
+	if (parent && parent->vendor == PCI_VENDOR_ID_ATI) {
+		/* use the upstream/downstream switches internal to dGPU */
+		*speed = pcie_get_speed_cap(parent);
+		*width = pcie_get_width_cap(parent);
+		while ((parent = pci_upstream_bridge(parent))) {
+			if (parent->vendor == PCI_VENDOR_ID_ATI) {
+				/* use the upstream/downstream switches internal to dGPU */
+				*speed = pcie_get_speed_cap(parent);
+				*width = pcie_get_width_cap(parent);
+			}
+		}
+	} else {
+		/* use the device itself */
+		*speed = pcie_get_speed_cap(adev->pdev);
+		*width = pcie_get_width_cap(adev->pdev);
+	}
+}
+
+/**
  * amdgpu_device_get_pcie_info - fence pcie info about the PCIE slot
  *
  * @adev: amdgpu_device pointer
  *
- * Fetchs and stores in the driver the PCIE capabilities (gen speed
+ * Fetches and stores in the driver the PCIE capabilities (gen speed
  * and lanes) of the slot the device is in. Handles APUs and
  * virtualized environments where PCIE config space may not be available.
  */
 static void amdgpu_device_get_pcie_info(struct amdgpu_device *adev)
 {
-	struct pci_dev *pdev;
 	enum pci_bus_speed speed_cap, platform_speed_cap;
-	enum pcie_link_width platform_link_width;
+	enum pcie_link_width platform_link_width, link_width;
 
 	if (amdgpu_pcie_gen_cap)
 		adev->pm.pcie_gen_mask = amdgpu_pcie_gen_cap;
@@ -5201,7 +6822,7 @@ static void amdgpu_device_get_pcie_info(struct amdgpu_device *adev)
 		adev->pm.pcie_mlw_mask = amdgpu_pcie_lane_cap;
 
 	/* covers APUs as well */
-	if (pci_is_root_bus(adev->pdev->bus)) {
+	if (pci_is_root_bus(adev->pdev->bus) && !amdgpu_passthrough(adev)) {
 		if (adev->pm.pcie_gen_mask == 0)
 			adev->pm.pcie_gen_mask = AMDGPU_DEFAULT_PCIE_GEN_MASK;
 		if (adev->pm.pcie_mlw_mask == 0)
@@ -5212,13 +6833,12 @@ static void amdgpu_device_get_pcie_info(struct amdgpu_device *adev)
 	if (adev->pm.pcie_gen_mask && adev->pm.pcie_mlw_mask)
 		return;
 
-	pcie_bandwidth_available(adev->pdev, NULL,
-				 &platform_speed_cap, &platform_link_width);
+	amdgpu_device_partner_bandwidth(adev, &platform_speed_cap,
+					&platform_link_width);
+	amdgpu_device_gpu_bandwidth(adev, &speed_cap, &link_width);
 
 	if (adev->pm.pcie_gen_mask == 0) {
 		/* asic caps */
-		pdev = adev->pdev;
-		speed_cap = pcie_get_speed_cap(pdev);
 		if (speed_cap == PCI_SPEED_UNKNOWN) {
 			adev->pm.pcie_gen_mask |= (CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_GEN1 |
 						  CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_GEN2 |
@@ -5274,51 +6894,103 @@ static void amdgpu_device_get_pcie_info(struct amdgpu_device *adev)
 		}
 	}
 	if (adev->pm.pcie_mlw_mask == 0) {
+		/* asic caps */
+		if (link_width == PCIE_LNK_WIDTH_UNKNOWN) {
+			adev->pm.pcie_mlw_mask |= AMDGPU_DEFAULT_ASIC_PCIE_MLW_MASK;
+		} else {
+			switch (link_width) {
+			case PCIE_LNK_X32:
+				adev->pm.pcie_mlw_mask |= (CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X32 |
+							   CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X16 |
+							   CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X12 |
+							   CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X8 |
+							   CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X4 |
+							   CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X2 |
+							   CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X1);
+				break;
+			case PCIE_LNK_X16:
+				adev->pm.pcie_mlw_mask |= (CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X16 |
+							   CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X12 |
+							   CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X8 |
+							   CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X4 |
+							   CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X2 |
+							   CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X1);
+				break;
+			case PCIE_LNK_X12:
+				adev->pm.pcie_mlw_mask |= (CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X12 |
+							   CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X8 |
+							   CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X4 |
+							   CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X2 |
+							   CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X1);
+				break;
+			case PCIE_LNK_X8:
+				adev->pm.pcie_mlw_mask |= (CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X8 |
+							   CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X4 |
+							   CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X2 |
+							   CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X1);
+				break;
+			case PCIE_LNK_X4:
+				adev->pm.pcie_mlw_mask |= (CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X4 |
+							   CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X2 |
+							   CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X1);
+				break;
+			case PCIE_LNK_X2:
+				adev->pm.pcie_mlw_mask |= (CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X2 |
+							   CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X1);
+				break;
+			case PCIE_LNK_X1:
+				adev->pm.pcie_mlw_mask |= CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X1;
+				break;
+			default:
+				break;
+			}
+		}
+		/* platform caps */
 		if (platform_link_width == PCIE_LNK_WIDTH_UNKNOWN) {
 			adev->pm.pcie_mlw_mask |= AMDGPU_DEFAULT_PCIE_MLW_MASK;
 		} else {
 			switch (platform_link_width) {
 			case PCIE_LNK_X32:
-				adev->pm.pcie_mlw_mask = (CAIL_PCIE_LINK_WIDTH_SUPPORT_X32 |
-							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X16 |
-							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X12 |
-							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X8 |
-							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X4 |
-							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X2 |
-							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X1);
+				adev->pm.pcie_mlw_mask |= (CAIL_PCIE_LINK_WIDTH_SUPPORT_X32 |
+							   CAIL_PCIE_LINK_WIDTH_SUPPORT_X16 |
+							   CAIL_PCIE_LINK_WIDTH_SUPPORT_X12 |
+							   CAIL_PCIE_LINK_WIDTH_SUPPORT_X8 |
+							   CAIL_PCIE_LINK_WIDTH_SUPPORT_X4 |
+							   CAIL_PCIE_LINK_WIDTH_SUPPORT_X2 |
+							   CAIL_PCIE_LINK_WIDTH_SUPPORT_X1);
 				break;
 			case PCIE_LNK_X16:
-				adev->pm.pcie_mlw_mask = (CAIL_PCIE_LINK_WIDTH_SUPPORT_X16 |
-							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X12 |
-							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X8 |
-							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X4 |
-							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X2 |
-							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X1);
+				adev->pm.pcie_mlw_mask |= (CAIL_PCIE_LINK_WIDTH_SUPPORT_X16 |
+							   CAIL_PCIE_LINK_WIDTH_SUPPORT_X12 |
+							   CAIL_PCIE_LINK_WIDTH_SUPPORT_X8 |
+							   CAIL_PCIE_LINK_WIDTH_SUPPORT_X4 |
+							   CAIL_PCIE_LINK_WIDTH_SUPPORT_X2 |
+							   CAIL_PCIE_LINK_WIDTH_SUPPORT_X1);
 				break;
 			case PCIE_LNK_X12:
-				adev->pm.pcie_mlw_mask = (CAIL_PCIE_LINK_WIDTH_SUPPORT_X12 |
-							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X8 |
-							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X4 |
-							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X2 |
-							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X1);
+				adev->pm.pcie_mlw_mask |= (CAIL_PCIE_LINK_WIDTH_SUPPORT_X12 |
+							   CAIL_PCIE_LINK_WIDTH_SUPPORT_X8 |
+							   CAIL_PCIE_LINK_WIDTH_SUPPORT_X4 |
+							   CAIL_PCIE_LINK_WIDTH_SUPPORT_X2 |
+							   CAIL_PCIE_LINK_WIDTH_SUPPORT_X1);
 				break;
 			case PCIE_LNK_X8:
-				adev->pm.pcie_mlw_mask = (CAIL_PCIE_LINK_WIDTH_SUPPORT_X8 |
-							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X4 |
-							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X2 |
-							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X1);
+				adev->pm.pcie_mlw_mask |= (CAIL_PCIE_LINK_WIDTH_SUPPORT_X8 |
+							   CAIL_PCIE_LINK_WIDTH_SUPPORT_X4 |
+							   CAIL_PCIE_LINK_WIDTH_SUPPORT_X2 |
+							   CAIL_PCIE_LINK_WIDTH_SUPPORT_X1);
 				break;
 			case PCIE_LNK_X4:
-				adev->pm.pcie_mlw_mask = (CAIL_PCIE_LINK_WIDTH_SUPPORT_X4 |
-							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X2 |
-							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X1);
+				adev->pm.pcie_mlw_mask |= (CAIL_PCIE_LINK_WIDTH_SUPPORT_X4 |
+							   CAIL_PCIE_LINK_WIDTH_SUPPORT_X2 |
+							   CAIL_PCIE_LINK_WIDTH_SUPPORT_X1);
 				break;
 			case PCIE_LNK_X2:
-				adev->pm.pcie_mlw_mask = (CAIL_PCIE_LINK_WIDTH_SUPPORT_X2 |
-							  CAIL_PCIE_LINK_WIDTH_SUPPORT_X1);
+				adev->pm.pcie_mlw_mask |= (CAIL_PCIE_LINK_WIDTH_SUPPORT_X2 |
+							   CAIL_PCIE_LINK_WIDTH_SUPPORT_X1);
 				break;
 			case PCIE_LNK_X1:
-				adev->pm.pcie_mlw_mask = CAIL_PCIE_LINK_WIDTH_SUPPORT_X1;
+				adev->pm.pcie_mlw_mask |= CAIL_PCIE_LINK_WIDTH_SUPPORT_X1;
 				break;
 			default:
 				break;
@@ -5327,12 +6999,51 @@ static void amdgpu_device_get_pcie_info(struct amdgpu_device *adev)
 	}
 }
 
-int amdgpu_device_baco_enter(struct drm_device *dev)
+/**
+ * amdgpu_device_is_peer_accessible - Check peer access through PCIe BAR
+ *
+ * @adev: amdgpu_device pointer
+ * @peer_adev: amdgpu_device pointer for peer device trying to access @adev
+ *
+ * Return true if @peer_adev can access (DMA) @adev through the PCIe
+ * BAR, i.e. @adev is "large BAR" and the BAR matches the DMA mask of
+ * @peer_adev.
+ */
+bool amdgpu_device_is_peer_accessible(struct amdgpu_device *adev,
+				      struct amdgpu_device *peer_adev)
+{
+#ifdef CONFIG_HSA_AMD_P2P
+	bool p2p_access =
+		!adev->gmc.xgmi.connected_to_cpu &&
+		!(pci_p2pdma_distance(adev->pdev, peer_adev->dev, false) < 0);
+	if (!p2p_access)
+		dev_info(adev->dev, "PCIe P2P access from peer device %s is not supported by the chipset\n",
+			pci_name(peer_adev->pdev));
+
+	bool is_large_bar = adev->gmc.visible_vram_size &&
+		adev->gmc.real_vram_size == adev->gmc.visible_vram_size;
+	bool p2p_addressable = amdgpu_device_check_iommu_remap(peer_adev);
+
+	if (!p2p_addressable) {
+		uint64_t address_mask = peer_adev->dev->dma_mask ?
+			~*peer_adev->dev->dma_mask : ~((1ULL << 32) - 1);
+		resource_size_t aper_limit =
+			adev->gmc.aper_base + adev->gmc.aper_size - 1;
+
+		p2p_addressable = !(adev->gmc.aper_base & address_mask ||
+				     aper_limit & address_mask);
+	}
+	return pcie_p2p && is_large_bar && p2p_access && p2p_addressable;
+#else
+	return false;
+#endif
+}
+
+int amdgpu_device_baco_enter(struct amdgpu_device *adev)
 {
-	struct amdgpu_device *adev = drm_to_adev(dev);
 	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
 
-	if (!amdgpu_device_supports_baco(adev_to_drm(adev)))
+	if (!amdgpu_device_supports_baco(adev))
 		return -ENOTSUPP;
 
 	if (ras && adev->ras_enabled &&
@@ -5342,13 +7053,12 @@ int amdgpu_device_baco_enter(struct drm_device *dev)
 	return amdgpu_dpm_baco_enter(adev);
 }
 
-int amdgpu_device_baco_exit(struct drm_device *dev)
+int amdgpu_device_baco_exit(struct amdgpu_device *adev)
 {
-	struct amdgpu_device *adev = drm_to_adev(dev);
 	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
 	int ret = 0;
 
-	if (!amdgpu_device_supports_baco(adev_to_drm(adev)))
+	if (!amdgpu_device_supports_baco(adev))
 		return -ENOTSUPP;
 
 	ret = amdgpu_dpm_baco_exit(adev);
@@ -5359,27 +7069,13 @@ int amdgpu_device_baco_exit(struct drm_device *dev)
 	    adev->nbio.funcs->enable_doorbell_interrupt)
 		adev->nbio.funcs->enable_doorbell_interrupt(adev, true);
 
-	if (amdgpu_passthrough(adev) &&
+	if (amdgpu_passthrough(adev) && adev->nbio.funcs &&
 	    adev->nbio.funcs->clear_doorbell_interrupt)
 		adev->nbio.funcs->clear_doorbell_interrupt(adev);
 
 	return 0;
 }
 
-static void amdgpu_cancel_all_tdr(struct amdgpu_device *adev)
-{
-	int i;
-
-	for (i = 0; i < AMDGPU_MAX_RINGS; ++i) {
-		struct amdgpu_ring *ring = adev->rings[i];
-
-		if (!ring || !ring->sched.thread)
-			continue;
-
-		cancel_delayed_work_sync(&ring->sched.work_tdr);
-	}
-}
-
 /**
  * amdgpu_pci_error_detected - Called when a PCI error is detected.
  * @pdev: PCI device struct
@@ -5393,48 +7089,52 @@ pci_ers_result_t amdgpu_pci_error_detected(struct pci_dev *pdev, pci_channel_sta
 {
 	struct drm_device *dev = pci_get_drvdata(pdev);
 	struct amdgpu_device *adev = drm_to_adev(dev);
-	int i;
-
-	DRM_INFO("PCI error: detected callback, state(%d)!!\n", state);
+	struct amdgpu_hive_info *hive __free(xgmi_put_hive) =
+		amdgpu_get_xgmi_hive(adev);
+	struct amdgpu_reset_context reset_context;
+	struct list_head device_list;
 
-	if (adev->gmc.xgmi.num_physical_nodes > 1) {
-		DRM_WARN("No support for XGMI hive yet...");
-		return PCI_ERS_RESULT_DISCONNECT;
-	}
+	dev_info(adev->dev, "PCI error: detected callback!!\n");
 
 	adev->pci_channel_state = state;
 
 	switch (state) {
 	case pci_channel_io_normal:
+		dev_info(adev->dev, "pci_channel_io_normal: state(%d)!!\n", state);
 		return PCI_ERS_RESULT_CAN_RECOVER;
-	/* Fatal error, prepare for slot reset */
 	case pci_channel_io_frozen:
-		/*
-		 * Cancel and wait for all TDRs in progress if failing to
-		 * set  adev->in_gpu_reset in amdgpu_device_lock_adev
-		 *
-		 * Locking adev->reset_sem will prevent any external access
-		 * to GPU during PCI error recovery
-		 */
-		while (!amdgpu_device_lock_adev(adev, NULL))
-			amdgpu_cancel_all_tdr(adev);
-
-		/*
-		 * Block any work scheduling as we do for regular GPU reset
-		 * for the duration of the recovery
-		 */
-		for (i = 0; i < AMDGPU_MAX_RINGS; ++i) {
-			struct amdgpu_ring *ring = adev->rings[i];
-
-			if (!ring || !ring->sched.thread)
-				continue;
+		/* Fatal error, prepare for slot reset */
+		dev_info(adev->dev, "pci_channel_io_frozen: state(%d)!!\n", state);
+		if (hive) {
+			/* Hive devices should be able to support FW based
+			 * link reset on other devices, if not return.
+			 */
+			if (!amdgpu_dpm_is_link_reset_supported(adev)) {
+				dev_warn(adev->dev,
+					 "No support for XGMI hive yet...\n");
+				return PCI_ERS_RESULT_DISCONNECT;
+			}
+			/* Set dpc status only if device is part of hive
+			 * Non-hive devices should be able to recover after
+			 * link reset.
+			 */
+			amdgpu_reset_set_dpc_status(adev, true);
 
-			drm_sched_stop(&ring->sched, NULL);
+			mutex_lock(&hive->hive_lock);
 		}
-		atomic_inc(&adev->gpu_reset_counter);
+		memset(&reset_context, 0, sizeof(reset_context));
+		INIT_LIST_HEAD(&device_list);
+
+		amdgpu_device_recovery_prepare(adev, &device_list, hive);
+		amdgpu_device_recovery_get_reset_lock(adev, &device_list);
+		amdgpu_device_halt_activities(adev, NULL, &reset_context, &device_list,
+					      hive, false);
+		if (hive)
+			mutex_unlock(&hive->hive_lock);
 		return PCI_ERS_RESULT_NEED_RESET;
 	case pci_channel_io_perm_failure:
 		/* Permanent error, prepare for device removal */
+		dev_info(adev->dev, "pci_channel_io_perm_failure: state(%d)!!\n", state);
 		return PCI_ERS_RESULT_DISCONNECT;
 	}
 
@@ -5447,8 +7147,10 @@ pci_ers_result_t amdgpu_pci_error_detected(struct pci_dev *pdev, pci_channel_sta
  */
 pci_ers_result_t amdgpu_pci_mmio_enabled(struct pci_dev *pdev)
 {
+	struct drm_device *dev = pci_get_drvdata(pdev);
+	struct amdgpu_device *adev = drm_to_adev(dev);
 
-	DRM_INFO("PCI error: mmio enabled callback!!\n");
+	dev_info(adev->dev, "PCI error: mmio enabled callback!!\n");
 
 	/* TODO - dump whatever for debugging purposes */
 
@@ -5472,21 +7174,38 @@ pci_ers_result_t amdgpu_pci_slot_reset(struct pci_dev *pdev)
 {
 	struct drm_device *dev = pci_get_drvdata(pdev);
 	struct amdgpu_device *adev = drm_to_adev(dev);
-	int r, i;
 	struct amdgpu_reset_context reset_context;
-	u32 memsize;
+	struct amdgpu_device *tmp_adev;
+	struct amdgpu_hive_info *hive;
 	struct list_head device_list;
+	struct pci_dev *link_dev;
+	int r = 0, i, timeout;
+	u32 memsize;
+	u16 status;
 
-	DRM_INFO("PCI error: slot reset callback!!\n");
+	dev_info(adev->dev, "PCI error: slot reset callback!!\n");
 
 	memset(&reset_context, 0, sizeof(reset_context));
 
-	INIT_LIST_HEAD(&device_list);
-	list_add_tail(&adev->reset_list, &device_list);
-
-	/* wait for asic to come out of reset */
-	msleep(500);
+	if (adev->pcie_reset_ctx.swus)
+		link_dev = adev->pcie_reset_ctx.swus;
+	else
+		link_dev = adev->pdev;
+	/* wait for asic to come out of reset, timeout = 10s */
+	timeout = 10000;
+	do {
+		usleep_range(10000, 10500);
+		r = pci_read_config_word(link_dev, PCI_VENDOR_ID, &status);
+		timeout -= 10;
+	} while (timeout > 0 && (status != PCI_VENDOR_ID_ATI) &&
+		 (status != PCI_VENDOR_ID_AMD));
+
+	if ((status != PCI_VENDOR_ID_ATI) && (status != PCI_VENDOR_ID_AMD)) {
+		r = -ETIME;
+		goto out;
+	}
 
+	amdgpu_device_load_switch_state(adev);
 	/* Restore PCI confspace */
 	amdgpu_device_load_pci_state(pdev);
 
@@ -5506,25 +7225,40 @@ pci_ers_result_t amdgpu_pci_slot_reset(struct pci_dev *pdev)
 	reset_context.method = AMD_RESET_METHOD_NONE;
 	reset_context.reset_req_dev = adev;
 	set_bit(AMDGPU_NEED_FULL_RESET, &reset_context.flags);
-	set_bit(AMDGPU_SKIP_HW_RESET, &reset_context.flags);
-
-	adev->no_hw_access = true;
-	r = amdgpu_device_pre_asic_reset(adev, &reset_context);
-	adev->no_hw_access = false;
-	if (r)
-		goto out;
+	set_bit(AMDGPU_SKIP_COREDUMP, &reset_context.flags);
+	INIT_LIST_HEAD(&device_list);
 
-	r = amdgpu_do_asic_reset(&device_list, &reset_context);
+	hive = amdgpu_get_xgmi_hive(adev);
+	if (hive) {
+		mutex_lock(&hive->hive_lock);
+		reset_context.hive = hive;
+		list_for_each_entry(tmp_adev, &hive->device_list, gmc.xgmi.head) {
+			tmp_adev->pcie_reset_ctx.in_link_reset = true;
+			list_add_tail(&tmp_adev->reset_list, &device_list);
+		}
+	} else {
+		set_bit(AMDGPU_SKIP_HW_RESET, &reset_context.flags);
+		list_add_tail(&adev->reset_list, &device_list);
+	}
 
+	r = amdgpu_device_asic_reset(adev, &device_list, &reset_context);
 out:
 	if (!r) {
 		if (amdgpu_device_cache_pci_state(adev->pdev))
 			pci_restore_state(adev->pdev);
-
-		DRM_INFO("PCIe error recovery succeeded\n");
+		dev_info(adev->dev, "PCIe error recovery succeeded\n");
 	} else {
-		DRM_ERROR("PCIe error recovery failed, err:%d", r);
-		amdgpu_device_unlock_adev(adev);
+		dev_err(adev->dev, "PCIe error recovery failed, err:%d\n", r);
+		if (hive) {
+			list_for_each_entry(tmp_adev, &device_list, reset_list)
+				amdgpu_device_unset_mp1_state(tmp_adev);
+		}
+		amdgpu_device_recovery_put_reset_lock(adev, &device_list);
+	}
+
+	if (hive) {
+		mutex_unlock(&hive->hive_lock);
+		amdgpu_put_xgmi_hive(hive);
 	}
 
 	return r ? PCI_ERS_RESULT_DISCONNECT : PCI_ERS_RESULT_RECOVERED;
@@ -5541,27 +7275,95 @@ void amdgpu_pci_resume(struct pci_dev *pdev)
 {
 	struct drm_device *dev = pci_get_drvdata(pdev);
 	struct amdgpu_device *adev = drm_to_adev(dev);
-	int i;
-
+	struct list_head device_list;
+	struct amdgpu_hive_info *hive = NULL;
+	struct amdgpu_device *tmp_adev = NULL;
 
-	DRM_INFO("PCI error: resume callback!!\n");
+	dev_info(adev->dev, "PCI error: resume callback!!\n");
 
 	/* Only continue execution for the case of pci_channel_io_frozen */
 	if (adev->pci_channel_state != pci_channel_io_frozen)
 		return;
 
-	for (i = 0; i < AMDGPU_MAX_RINGS; ++i) {
-		struct amdgpu_ring *ring = adev->rings[i];
+	INIT_LIST_HEAD(&device_list);
 
-		if (!ring || !ring->sched.thread)
-			continue;
+	hive = amdgpu_get_xgmi_hive(adev);
+	if (hive) {
+		mutex_lock(&hive->hive_lock);
+		list_for_each_entry(tmp_adev, &hive->device_list, gmc.xgmi.head) {
+			tmp_adev->pcie_reset_ctx.in_link_reset = false;
+			list_add_tail(&tmp_adev->reset_list, &device_list);
+		}
+	} else
+		list_add_tail(&adev->reset_list, &device_list);
 
+	amdgpu_device_sched_resume(&device_list, NULL, NULL);
+	amdgpu_device_gpu_resume(adev, &device_list, false);
+	amdgpu_device_recovery_put_reset_lock(adev, &device_list);
 
-		drm_sched_resubmit_jobs(&ring->sched);
-		drm_sched_start(&ring->sched, true);
+	if (hive) {
+		mutex_unlock(&hive->hive_lock);
+		amdgpu_put_xgmi_hive(hive);
 	}
+}
+
+static void amdgpu_device_cache_switch_state(struct amdgpu_device *adev)
+{
+	struct pci_dev *swus, *swds;
+	int r;
 
-	amdgpu_device_unlock_adev(adev);
+	swds = pci_upstream_bridge(adev->pdev);
+	if (!swds || swds->vendor != PCI_VENDOR_ID_ATI ||
+	    pci_pcie_type(swds) != PCI_EXP_TYPE_DOWNSTREAM)
+		return;
+	swus = pci_upstream_bridge(swds);
+	if (!swus ||
+	    (swus->vendor != PCI_VENDOR_ID_ATI &&
+	     swus->vendor != PCI_VENDOR_ID_AMD) ||
+	    pci_pcie_type(swus) != PCI_EXP_TYPE_UPSTREAM)
+		return;
+
+	/* If already saved, return */
+	if (adev->pcie_reset_ctx.swus)
+		return;
+	/* Upstream bridge is ATI, assume it's SWUS/DS architecture */
+	r = pci_save_state(swds);
+	if (r)
+		return;
+	adev->pcie_reset_ctx.swds_pcistate = pci_store_saved_state(swds);
+
+	r = pci_save_state(swus);
+	if (r)
+		return;
+	adev->pcie_reset_ctx.swus_pcistate = pci_store_saved_state(swus);
+
+	adev->pcie_reset_ctx.swus = swus;
+}
+
+static void amdgpu_device_load_switch_state(struct amdgpu_device *adev)
+{
+	struct pci_dev *pdev;
+	int r;
+
+	if (!adev->pcie_reset_ctx.swds_pcistate ||
+	    !adev->pcie_reset_ctx.swus_pcistate)
+		return;
+
+	pdev = adev->pcie_reset_ctx.swus;
+	r = pci_load_saved_state(pdev, adev->pcie_reset_ctx.swus_pcistate);
+	if (!r) {
+		pci_restore_state(pdev);
+	} else {
+		dev_warn(adev->dev, "Failed to load SWUS state, err:%d\n", r);
+		return;
+	}
+
+	pdev = pci_upstream_bridge(adev->pdev);
+	r = pci_load_saved_state(pdev, adev->pcie_reset_ctx.swds_pcistate);
+	if (!r)
+		pci_restore_state(pdev);
+	else
+		dev_warn(adev->dev, "Failed to load SWDS state, err:%d\n", r);
 }
 
 bool amdgpu_device_cache_pci_state(struct pci_dev *pdev)
@@ -5570,6 +7372,9 @@ bool amdgpu_device_cache_pci_state(struct pci_dev *pdev)
 	struct amdgpu_device *adev = drm_to_adev(dev);
 	int r;
 
+	if (amdgpu_sriov_vf(adev))
+		return false;
+
 	r = pci_save_state(pdev);
 	if (!r) {
 		kfree(adev->pci_state);
@@ -5577,14 +7382,16 @@ bool amdgpu_device_cache_pci_state(struct pci_dev *pdev)
 		adev->pci_state = pci_store_saved_state(pdev);
 
 		if (!adev->pci_state) {
-			DRM_ERROR("Failed to store PCI saved state");
+			dev_err(adev->dev, "Failed to store PCI saved state");
 			return false;
 		}
 	} else {
-		DRM_WARN("Failed to save PCI state, err:%d\n", r);
+		dev_warn(adev->dev, "Failed to save PCI state, err:%d\n", r);
 		return false;
 	}
 
+	amdgpu_device_cache_switch_state(adev);
+
 	return true;
 }
 
@@ -5602,7 +7409,7 @@ bool amdgpu_device_load_pci_state(struct pci_dev *pdev)
 	if (!r) {
 		pci_restore_state(pdev);
 	} else {
-		DRM_WARN("Failed to load PCI state, err:%d\n", r);
+		dev_warn(adev->dev, "Failed to load PCI state, err:%d\n", r);
 		return false;
 	}
 
@@ -5613,27 +7420,418 @@ void amdgpu_device_flush_hdp(struct amdgpu_device *adev,
 		struct amdgpu_ring *ring)
 {
 #ifdef CONFIG_X86_64
-	if (adev->flags & AMD_IS_APU)
+	if ((adev->flags & AMD_IS_APU) && !amdgpu_passthrough(adev))
 		return;
 #endif
 	if (adev->gmc.xgmi.connected_to_cpu)
 		return;
 
-	if (ring && ring->funcs->emit_hdp_flush)
+	if (ring && ring->funcs->emit_hdp_flush) {
 		amdgpu_ring_emit_hdp_flush(ring);
-	else
-		amdgpu_asic_flush_hdp(adev, ring);
+		return;
+	}
+
+	if (!ring && amdgpu_sriov_runtime(adev)) {
+		if (!amdgpu_kiq_hdp_flush(adev))
+			return;
+	}
+
+	amdgpu_hdp_flush(adev, ring);
 }
 
 void amdgpu_device_invalidate_hdp(struct amdgpu_device *adev,
 		struct amdgpu_ring *ring)
 {
 #ifdef CONFIG_X86_64
-	if (adev->flags & AMD_IS_APU)
+	if ((adev->flags & AMD_IS_APU) && !amdgpu_passthrough(adev))
 		return;
 #endif
 	if (adev->gmc.xgmi.connected_to_cpu)
 		return;
 
-	amdgpu_asic_invalidate_hdp(adev, ring);
+	amdgpu_hdp_invalidate(adev, ring);
+}
+
+int amdgpu_in_reset(struct amdgpu_device *adev)
+{
+	return atomic_read(&adev->reset_domain->in_gpu_reset);
+}
+
+/**
+ * amdgpu_device_halt() - bring hardware to some kind of halt state
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Bring hardware to some kind of halt state so that no one can touch it
+ * any more. It will help to maintain error context when error occurred.
+ * Compare to a simple hang, the system will keep stable at least for SSH
+ * access. Then it should be trivial to inspect the hardware state and
+ * see what's going on. Implemented as following:
+ *
+ * 1. drm_dev_unplug() makes device inaccessible to user space(IOCTLs, etc),
+ *    clears all CPU mappings to device, disallows remappings through page faults
+ * 2. amdgpu_irq_disable_all() disables all interrupts
+ * 3. amdgpu_fence_driver_hw_fini() signals all HW fences
+ * 4. set adev->no_hw_access to avoid potential crashes after setp 5
+ * 5. amdgpu_device_unmap_mmio() clears all MMIO mappings
+ * 6. pci_disable_device() and pci_wait_for_pending_transaction()
+ *    flush any in flight DMA operations
+ */
+void amdgpu_device_halt(struct amdgpu_device *adev)
+{
+	struct pci_dev *pdev = adev->pdev;
+	struct drm_device *ddev = adev_to_drm(adev);
+
+	amdgpu_xcp_dev_unplug(adev);
+	drm_dev_unplug(ddev);
+
+	amdgpu_irq_disable_all(adev);
+
+	amdgpu_fence_driver_hw_fini(adev);
+
+	adev->no_hw_access = true;
+
+	amdgpu_device_unmap_mmio(adev);
+
+	pci_disable_device(pdev);
+	pci_wait_for_pending_transaction(pdev);
+}
+
+u32 amdgpu_device_pcie_port_rreg(struct amdgpu_device *adev,
+				u32 reg)
+{
+	unsigned long flags, address, data;
+	u32 r;
+
+	address = adev->nbio.funcs->get_pcie_port_index_offset(adev);
+	data = adev->nbio.funcs->get_pcie_port_data_offset(adev);
+
+	spin_lock_irqsave(&adev->pcie_idx_lock, flags);
+	WREG32(address, reg * 4);
+	(void)RREG32(address);
+	r = RREG32(data);
+	spin_unlock_irqrestore(&adev->pcie_idx_lock, flags);
+	return r;
+}
+
+void amdgpu_device_pcie_port_wreg(struct amdgpu_device *adev,
+				u32 reg, u32 v)
+{
+	unsigned long flags, address, data;
+
+	address = adev->nbio.funcs->get_pcie_port_index_offset(adev);
+	data = adev->nbio.funcs->get_pcie_port_data_offset(adev);
+
+	spin_lock_irqsave(&adev->pcie_idx_lock, flags);
+	WREG32(address, reg * 4);
+	(void)RREG32(address);
+	WREG32(data, v);
+	(void)RREG32(data);
+	spin_unlock_irqrestore(&adev->pcie_idx_lock, flags);
+}
+
+/**
+ * amdgpu_device_get_gang - return a reference to the current gang
+ * @adev: amdgpu_device pointer
+ *
+ * Returns: A new reference to the current gang leader.
+ */
+struct dma_fence *amdgpu_device_get_gang(struct amdgpu_device *adev)
+{
+	struct dma_fence *fence;
+
+	rcu_read_lock();
+	fence = dma_fence_get_rcu_safe(&adev->gang_submit);
+	rcu_read_unlock();
+	return fence;
+}
+
+/**
+ * amdgpu_device_switch_gang - switch to a new gang
+ * @adev: amdgpu_device pointer
+ * @gang: the gang to switch to
+ *
+ * Try to switch to a new gang.
+ * Returns: NULL if we switched to the new gang or a reference to the current
+ * gang leader.
+ */
+struct dma_fence *amdgpu_device_switch_gang(struct amdgpu_device *adev,
+					    struct dma_fence *gang)
+{
+	struct dma_fence *old = NULL;
+
+	dma_fence_get(gang);
+	do {
+		dma_fence_put(old);
+		old = amdgpu_device_get_gang(adev);
+		if (old == gang)
+			break;
+
+		if (!dma_fence_is_signaled(old)) {
+			dma_fence_put(gang);
+			return old;
+		}
+
+	} while (cmpxchg((struct dma_fence __force **)&adev->gang_submit,
+			 old, gang) != old);
+
+	/*
+	 * Drop it once for the exchanged reference in adev and once for the
+	 * thread local reference acquired in amdgpu_device_get_gang().
+	 */
+	dma_fence_put(old);
+	dma_fence_put(old);
+	return NULL;
+}
+
+/**
+ * amdgpu_device_enforce_isolation - enforce HW isolation
+ * @adev: the amdgpu device pointer
+ * @ring: the HW ring the job is supposed to run on
+ * @job: the job which is about to be pushed to the HW ring
+ *
+ * Makes sure that only one client at a time can use the GFX block.
+ * Returns: The dependency to wait on before the job can be pushed to the HW.
+ * The function is called multiple times until NULL is returned.
+ */
+struct dma_fence *amdgpu_device_enforce_isolation(struct amdgpu_device *adev,
+						  struct amdgpu_ring *ring,
+						  struct amdgpu_job *job)
+{
+	struct amdgpu_isolation *isolation = &adev->isolation[ring->xcp_id];
+	struct drm_sched_fence *f = job->base.s_fence;
+	struct dma_fence *dep;
+	void *owner;
+	int r;
+
+	/*
+	 * For now enforce isolation only for the GFX block since we only need
+	 * the cleaner shader on those rings.
+	 */
+	if (ring->funcs->type != AMDGPU_RING_TYPE_GFX &&
+	    ring->funcs->type != AMDGPU_RING_TYPE_COMPUTE)
+		return NULL;
+
+	/*
+	 * All submissions where enforce isolation is false are handled as if
+	 * they come from a single client. Use ~0l as the owner to distinct it
+	 * from kernel submissions where the owner is NULL.
+	 */
+	owner = job->enforce_isolation ? f->owner : (void *)~0l;
+
+	mutex_lock(&adev->enforce_isolation_mutex);
+
+	/*
+	 * The "spearhead" submission is the first one which changes the
+	 * ownership to its client. We always need to wait for it to be
+	 * pushed to the HW before proceeding with anything.
+	 */
+	if (&f->scheduled != isolation->spearhead &&
+	    !dma_fence_is_signaled(isolation->spearhead)) {
+		dep = isolation->spearhead;
+		goto out_grab_ref;
+	}
+
+	if (isolation->owner != owner) {
+
+		/*
+		 * Wait for any gang to be assembled before switching to a
+		 * different owner or otherwise we could deadlock the
+		 * submissions.
+		 */
+		if (!job->gang_submit) {
+			dep = amdgpu_device_get_gang(adev);
+			if (!dma_fence_is_signaled(dep))
+				goto out_return_dep;
+			dma_fence_put(dep);
+		}
+
+		dma_fence_put(isolation->spearhead);
+		isolation->spearhead = dma_fence_get(&f->scheduled);
+		amdgpu_sync_move(&isolation->active, &isolation->prev);
+		trace_amdgpu_isolation(isolation->owner, owner);
+		isolation->owner = owner;
+	}
+
+	/*
+	 * Specifying the ring here helps to pipeline submissions even when
+	 * isolation is enabled. If that is not desired for testing NULL can be
+	 * used instead of the ring to enforce a CPU round trip while switching
+	 * between clients.
+	 */
+	dep = amdgpu_sync_peek_fence(&isolation->prev, ring);
+	r = amdgpu_sync_fence(&isolation->active, &f->finished, GFP_NOWAIT);
+	if (r)
+		dev_warn(adev->dev, "OOM tracking isolation\n");
+
+out_grab_ref:
+	dma_fence_get(dep);
+out_return_dep:
+	mutex_unlock(&adev->enforce_isolation_mutex);
+	return dep;
+}
+
+bool amdgpu_device_has_display_hardware(struct amdgpu_device *adev)
+{
+	switch (adev->asic_type) {
+#ifdef CONFIG_DRM_AMDGPU_SI
+	case CHIP_HAINAN:
+#endif
+	case CHIP_TOPAZ:
+		/* chips with no display hardware */
+		return false;
+#ifdef CONFIG_DRM_AMDGPU_SI
+	case CHIP_TAHITI:
+	case CHIP_PITCAIRN:
+	case CHIP_VERDE:
+	case CHIP_OLAND:
+#endif
+#ifdef CONFIG_DRM_AMDGPU_CIK
+	case CHIP_BONAIRE:
+	case CHIP_HAWAII:
+	case CHIP_KAVERI:
+	case CHIP_KABINI:
+	case CHIP_MULLINS:
+#endif
+	case CHIP_TONGA:
+	case CHIP_FIJI:
+	case CHIP_POLARIS10:
+	case CHIP_POLARIS11:
+	case CHIP_POLARIS12:
+	case CHIP_VEGAM:
+	case CHIP_CARRIZO:
+	case CHIP_STONEY:
+		/* chips with display hardware */
+		return true;
+	default:
+		/* IP discovery */
+		if (!amdgpu_ip_version(adev, DCE_HWIP, 0) ||
+		    (adev->harvest_ip_mask & AMD_HARVEST_IP_DMU_MASK))
+			return false;
+		return true;
+	}
+}
+
+uint32_t amdgpu_device_wait_on_rreg(struct amdgpu_device *adev,
+		uint32_t inst, uint32_t reg_addr, char reg_name[],
+		uint32_t expected_value, uint32_t mask)
+{
+	uint32_t ret = 0;
+	uint32_t old_ = 0;
+	uint32_t tmp_ = RREG32(reg_addr);
+	uint32_t loop = adev->usec_timeout;
+
+	while ((tmp_ & (mask)) != (expected_value)) {
+		if (old_ != tmp_) {
+			loop = adev->usec_timeout;
+			old_ = tmp_;
+		} else
+			udelay(1);
+		tmp_ = RREG32(reg_addr);
+		loop--;
+		if (!loop) {
+			dev_warn(
+				adev->dev,
+				"Register(%d) [%s] failed to reach value 0x%08x != 0x%08xn",
+				inst, reg_name, (uint32_t)expected_value,
+				(uint32_t)(tmp_ & (mask)));
+			ret = -ETIMEDOUT;
+			break;
+		}
+	}
+	return ret;
+}
+
+ssize_t amdgpu_get_soft_full_reset_mask(struct amdgpu_ring *ring)
+{
+	ssize_t size = 0;
+
+	if (!ring || !ring->adev)
+		return size;
+
+	if (amdgpu_device_should_recover_gpu(ring->adev))
+		size |= AMDGPU_RESET_TYPE_FULL;
+
+	if (unlikely(!ring->adev->debug_disable_soft_recovery) &&
+	    !amdgpu_sriov_vf(ring->adev) && ring->funcs->soft_recovery)
+		size |= AMDGPU_RESET_TYPE_SOFT_RESET;
+
+	return size;
+}
+
+ssize_t amdgpu_show_reset_mask(char *buf, uint32_t supported_reset)
+{
+	ssize_t size = 0;
+
+	if (supported_reset == 0) {
+		size += sysfs_emit_at(buf, size, "unsupported");
+		size += sysfs_emit_at(buf, size, "\n");
+		return size;
+
+	}
+
+	if (supported_reset & AMDGPU_RESET_TYPE_SOFT_RESET)
+		size += sysfs_emit_at(buf, size, "soft ");
+
+	if (supported_reset & AMDGPU_RESET_TYPE_PER_QUEUE)
+		size += sysfs_emit_at(buf, size, "queue ");
+
+	if (supported_reset & AMDGPU_RESET_TYPE_PER_PIPE)
+		size += sysfs_emit_at(buf, size, "pipe ");
+
+	if (supported_reset & AMDGPU_RESET_TYPE_FULL)
+		size += sysfs_emit_at(buf, size, "full ");
+
+	size += sysfs_emit_at(buf, size, "\n");
+	return size;
+}
+
+void amdgpu_device_set_uid(struct amdgpu_uid *uid_info,
+			   enum amdgpu_uid_type type, uint8_t inst,
+			   uint64_t uid)
+{
+	if (!uid_info)
+		return;
+
+	if (type >= AMDGPU_UID_TYPE_MAX) {
+		dev_err_once(uid_info->adev->dev, "Invalid UID type %d\n",
+			     type);
+		return;
+	}
+
+	if (inst >= AMDGPU_UID_INST_MAX) {
+		dev_err_once(uid_info->adev->dev, "Invalid UID instance %d\n",
+			     inst);
+		return;
+	}
+
+	if (uid_info->uid[type][inst] != 0) {
+		dev_warn_once(
+			uid_info->adev->dev,
+			"Overwriting existing UID %llu for type %d instance %d\n",
+			uid_info->uid[type][inst], type, inst);
+	}
+
+	uid_info->uid[type][inst] = uid;
+}
+
+u64 amdgpu_device_get_uid(struct amdgpu_uid *uid_info,
+			  enum amdgpu_uid_type type, uint8_t inst)
+{
+	if (!uid_info)
+		return 0;
+
+	if (type >= AMDGPU_UID_TYPE_MAX) {
+		dev_err_once(uid_info->adev->dev, "Invalid UID type %d\n",
+			     type);
+		return 0;
+	}
+
+	if (inst >= AMDGPU_UID_INST_MAX) {
+		dev_err_once(uid_info->adev->dev, "Invalid UID instance %d\n",
+			     inst);
+		return 0;
+	}
+
+	return uid_info->uid[type][inst];
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_df.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_df.h
index 6b25837955c4..eb605e79ae0e 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_df.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_df.h
@@ -33,6 +33,7 @@ struct amdgpu_df_hash_status {
 struct amdgpu_df_funcs {
 	void (*sw_init)(struct amdgpu_device *adev);
 	void (*sw_fini)(struct amdgpu_device *adev);
+	void (*hw_init)(struct amdgpu_device *adev);
 	void (*enable_broadcast_mode)(struct amdgpu_device *adev,
 				      bool enable);
 	u32 (*get_fb_channel_number)(struct amdgpu_device *adev);
@@ -40,7 +41,7 @@ struct amdgpu_df_funcs {
 	void (*update_medium_grain_clock_gating)(struct amdgpu_device *adev,
 						 bool enable);
 	void (*get_clockgating_state)(struct amdgpu_device *adev,
-				      u32 *flags);
+				      u64 *flags);
 	void (*enable_ecc_force_par_wr_rmw)(struct amdgpu_device *adev,
 					    bool enable);
 	int (*pmc_start)(struct amdgpu_device *adev, uint64_t config,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c
index ea00090b3fb3..fa2a22dfa048 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c
@@ -1,5 +1,5 @@
 /*
- * Copyright 2018 Advanced Micro Devices, Inc.
+ * Copyright 2018-2024 Advanced Micro Devices, Inc. All rights reserved.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -27,19 +27,27 @@
 #include "amdgpu_discovery.h"
 #include "soc15_hw_ip.h"
 #include "discovery.h"
+#include "amdgpu_ras.h"
 
 #include "soc15.h"
 #include "gfx_v9_0.h"
+#include "gfx_v9_4_3.h"
 #include "gmc_v9_0.h"
 #include "df_v1_7.h"
 #include "df_v3_6.h"
+#include "df_v4_3.h"
+#include "df_v4_6_2.h"
+#include "df_v4_15.h"
 #include "nbio_v6_1.h"
 #include "nbio_v7_0.h"
 #include "nbio_v7_4.h"
+#include "nbio_v7_9.h"
+#include "nbio_v7_11.h"
 #include "hdp_v4_0.h"
 #include "vega10_ih.h"
 #include "vega20_ih.h"
 #include "sdma_v4_0.h"
+#include "sdma_v4_4_2.h"
 #include "uvd_v7_0.h"
 #include "vce_v4_0.h"
 #include "vcn_v1_0.h"
@@ -47,29 +55,78 @@
 #include "jpeg_v2_5.h"
 #include "smuio_v9_0.h"
 #include "gmc_v10_0.h"
+#include "gmc_v11_0.h"
+#include "gmc_v12_0.h"
 #include "gfxhub_v2_0.h"
 #include "mmhub_v2_0.h"
 #include "nbio_v2_3.h"
+#include "nbio_v4_3.h"
 #include "nbio_v7_2.h"
+#include "nbio_v7_7.h"
+#include "nbif_v6_3_1.h"
 #include "hdp_v5_0.h"
+#include "hdp_v5_2.h"
+#include "hdp_v6_0.h"
+#include "hdp_v7_0.h"
 #include "nv.h"
+#include "soc21.h"
+#include "soc24.h"
 #include "navi10_ih.h"
+#include "ih_v6_0.h"
+#include "ih_v6_1.h"
+#include "ih_v7_0.h"
 #include "gfx_v10_0.h"
+#include "gfx_v11_0.h"
+#include "gfx_v12_0.h"
 #include "sdma_v5_0.h"
 #include "sdma_v5_2.h"
+#include "sdma_v6_0.h"
+#include "sdma_v7_0.h"
+#include "lsdma_v6_0.h"
+#include "lsdma_v7_0.h"
 #include "vcn_v2_0.h"
 #include "jpeg_v2_0.h"
 #include "vcn_v3_0.h"
 #include "jpeg_v3_0.h"
+#include "vcn_v4_0.h"
+#include "jpeg_v4_0.h"
+#include "vcn_v4_0_3.h"
+#include "jpeg_v4_0_3.h"
+#include "vcn_v4_0_5.h"
+#include "jpeg_v4_0_5.h"
 #include "amdgpu_vkms.h"
-#include "mes_v10_1.h"
+#include "mes_v11_0.h"
+#include "mes_v12_0.h"
 #include "smuio_v11_0.h"
 #include "smuio_v11_0_6.h"
 #include "smuio_v13_0.h"
+#include "smuio_v13_0_3.h"
+#include "smuio_v13_0_6.h"
+#include "smuio_v14_0_2.h"
+#include "vcn_v5_0_0.h"
+#include "vcn_v5_0_1.h"
+#include "jpeg_v5_0_0.h"
+#include "jpeg_v5_0_1.h"
+#include "amdgpu_ras_mgr.h"
+
+#include "amdgpu_vpe.h"
+#if defined(CONFIG_DRM_AMD_ISP)
+#include "amdgpu_isp.h"
+#endif
 
 MODULE_FIRMWARE("amdgpu/ip_discovery.bin");
-
+MODULE_FIRMWARE("amdgpu/vega10_ip_discovery.bin");
+MODULE_FIRMWARE("amdgpu/vega12_ip_discovery.bin");
+MODULE_FIRMWARE("amdgpu/vega20_ip_discovery.bin");
+MODULE_FIRMWARE("amdgpu/raven_ip_discovery.bin");
+MODULE_FIRMWARE("amdgpu/raven2_ip_discovery.bin");
+MODULE_FIRMWARE("amdgpu/picasso_ip_discovery.bin");
+MODULE_FIRMWARE("amdgpu/arcturus_ip_discovery.bin");
+MODULE_FIRMWARE("amdgpu/aldebaran_ip_discovery.bin");
+
+#define mmIP_DISCOVERY_VERSION  0x16A00
 #define mmRCC_CONFIG_MEMSIZE	0xde3
+#define mmMP0_SMN_C2PMSG_33	0x16061
 #define mmMM_INDEX		0x0
 #define mmMM_INDEX_HI		0x6
 #define mmMM_DATA		0x1
@@ -110,6 +167,7 @@ static const char *hw_id_names[HW_ID_MAX] = {
 	[SDMA1_HWID]		= "SDMA1",
 	[SDMA2_HWID]		= "SDMA2",
 	[SDMA3_HWID]		= "SDMA3",
+	[LSDMA_HWID]		= "LSDMA",
 	[ISP_HWID]		= "ISP",
 	[DBGU_IO_HWID]		= "DBGU_IO",
 	[DF_HWID]		= "DF",
@@ -150,6 +208,7 @@ static const char *hw_id_names[HW_ID_MAX] = {
 	[XGMI_HWID]		= "XGMI",
 	[XGBE_HWID]		= "XGBE",
 	[MP0_HWID]		= "MP0",
+	[VPE_HWID]		= "VPE",
 };
 
 static int hw_id_map[MAX_HWIP] = {
@@ -159,6 +218,7 @@ static int hw_id_map[MAX_HWIP] = {
 	[SDMA1_HWIP]	= SDMA1_HWID,
 	[SDMA2_HWIP]    = SDMA2_HWID,
 	[SDMA3_HWIP]    = SDMA3_HWID,
+	[LSDMA_HWIP]    = LSDMA_HWID,
 	[MMHUB_HWIP]	= MMHUB_HWID,
 	[ATHUB_HWIP]	= ATHUB_HWID,
 	[NBIO_HWIP]	= NBIF_HWID,
@@ -177,15 +237,124 @@ static int hw_id_map[MAX_HWIP] = {
 	[UMC_HWIP]	= UMC_HWID,
 	[XGMI_HWIP]	= XGMI_HWID,
 	[DCI_HWIP]	= DCI_HWID,
+	[PCIE_HWIP]	= PCIE_HWID,
+	[VPE_HWIP]	= VPE_HWID,
+	[ISP_HWIP]	= ISP_HWID,
 };
 
-static int amdgpu_discovery_read_binary(struct amdgpu_device *adev, uint8_t *binary)
+static int amdgpu_discovery_read_binary_from_sysmem(struct amdgpu_device *adev, uint8_t *binary)
+{
+	u64 tmr_offset, tmr_size, pos;
+	void *discv_regn;
+	int ret;
+
+	ret = amdgpu_acpi_get_tmr_info(adev, &tmr_offset, &tmr_size);
+	if (ret)
+		return ret;
+
+	pos = tmr_offset + tmr_size - DISCOVERY_TMR_OFFSET;
+
+	/* This region is read-only and reserved from system use */
+	discv_regn = memremap(pos, adev->discovery.size, MEMREMAP_WC);
+	if (discv_regn) {
+		memcpy(binary, discv_regn, adev->discovery.size);
+		memunmap(discv_regn);
+		return 0;
+	}
+
+	return -ENOENT;
+}
+
+#define IP_DISCOVERY_V2		2
+#define IP_DISCOVERY_V4		4
+
+static int amdgpu_discovery_read_binary_from_mem(struct amdgpu_device *adev,
+						 uint8_t *binary)
+{
+	bool sz_valid = true;
+	uint64_t vram_size;
+	int i, ret = 0;
+	u32 msg;
+
+	if (!amdgpu_sriov_vf(adev)) {
+		/* It can take up to two second for IFWI init to complete on some dGPUs,
+		 * but generally it should be in the 60-100ms range.  Normally this starts
+		 * as soon as the device gets power so by the time the OS loads this has long
+		 * completed.  However, when a card is hotplugged via e.g., USB4, we need to
+		 * wait for this to complete.  Once the C2PMSG is updated, we can
+		 * continue.
+		 */
+
+		for (i = 0; i < 2000; i++) {
+			msg = RREG32(mmMP0_SMN_C2PMSG_33);
+			if (msg & 0x80000000)
+				break;
+			msleep(1);
+		}
+	}
+
+	vram_size = RREG32(mmRCC_CONFIG_MEMSIZE);
+	if (!vram_size || vram_size == U32_MAX)
+		sz_valid = false;
+	else
+		vram_size <<= 20;
+
+	/*
+	 * If in VRAM, discovery TMR is marked for reservation. If it is in system mem,
+	 * then it is not required to be reserved.
+	 */
+	if (sz_valid) {
+		if (amdgpu_sriov_vf(adev) && adev->virt.is_dynamic_crit_regn_enabled) {
+			/* For SRIOV VFs with dynamic critical region enabled,
+			 * we will get the IPD binary via below call.
+			 * If dynamic critical is disabled, fall through to normal seq.
+			 */
+			if (amdgpu_virt_get_dynamic_data_info(adev,
+						AMD_SRIOV_MSG_IPD_TABLE_ID, binary,
+						&adev->discovery.size)) {
+				dev_err(adev->dev,
+						"failed to read discovery info from dynamic critical region.");
+				ret = -EINVAL;
+				goto exit;
+			}
+		} else {
+			uint64_t pos = vram_size - DISCOVERY_TMR_OFFSET;
+
+			amdgpu_device_vram_access(adev, pos, (uint32_t *)binary,
+					adev->discovery.size, false);
+			adev->discovery.reserve_tmr = true;
+		}
+	} else {
+		ret = amdgpu_discovery_read_binary_from_sysmem(adev, binary);
+	}
+
+	if (ret)
+		dev_err(adev->dev,
+			"failed to read discovery info from memory, vram size read: %llx",
+			vram_size);
+exit:
+	return ret;
+}
+
+static int amdgpu_discovery_read_binary_from_file(struct amdgpu_device *adev,
+							uint8_t *binary,
+							const char *fw_name)
 {
-	uint64_t vram_size = (uint64_t)RREG32(mmRCC_CONFIG_MEMSIZE) << 20;
-	uint64_t pos = vram_size - DISCOVERY_TMR_OFFSET;
+	const struct firmware *fw;
+	int r;
+
+	r = firmware_request_nowarn(&fw, fw_name, adev->dev);
+	if (r) {
+		if (amdgpu_discovery == 2)
+			dev_err(adev->dev, "can't load firmware \"%s\"\n", fw_name);
+		else
+			drm_info(&adev->ddev, "Optional firmware \"%s\" was not found\n", fw_name);
+		return r;
+	}
+
+	memcpy((u8 *)binary, (u8 *)fw->data, fw->size);
+	release_firmware(fw);
 
-	amdgpu_device_vram_access(adev, pos, (uint32_t *)binary,
-				  adev->mman.discovery_tmr_size, false);
 	return 0;
 }
 
@@ -206,56 +375,152 @@ static inline bool amdgpu_discovery_verify_checksum(uint8_t *data, uint32_t size
 	return !!(amdgpu_discovery_calculate_checksum(data, size) == expected);
 }
 
+static inline bool amdgpu_discovery_verify_binary_signature(uint8_t *binary)
+{
+	struct binary_header *bhdr;
+	bhdr = (struct binary_header *)binary;
+
+	return (le32_to_cpu(bhdr->binary_signature) == BINARY_SIGNATURE);
+}
+
+static void amdgpu_discovery_harvest_config_quirk(struct amdgpu_device *adev)
+{
+	/*
+	 * So far, apply this quirk only on those Navy Flounder boards which
+	 * have a bad harvest table of VCN config.
+	 */
+	if ((amdgpu_ip_version(adev, UVD_HWIP, 1) == IP_VERSION(3, 0, 1)) &&
+	    (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(10, 3, 2))) {
+		switch (adev->pdev->revision) {
+		case 0xC1:
+		case 0xC2:
+		case 0xC3:
+		case 0xC5:
+		case 0xC7:
+		case 0xCF:
+		case 0xDF:
+			adev->vcn.harvest_config |= AMDGPU_VCN_HARVEST_VCN1;
+			adev->vcn.inst_mask &= ~AMDGPU_VCN_HARVEST_VCN1;
+			break;
+		default:
+			break;
+		}
+	}
+}
+
+static int amdgpu_discovery_verify_npsinfo(struct amdgpu_device *adev,
+					   struct binary_header *bhdr)
+{
+	uint8_t *discovery_bin = adev->discovery.bin;
+	struct table_info *info;
+	uint16_t checksum;
+	uint16_t offset;
+
+	info = &bhdr->table_list[NPS_INFO];
+	offset = le16_to_cpu(info->offset);
+	checksum = le16_to_cpu(info->checksum);
+
+	struct nps_info_header *nhdr =
+		(struct nps_info_header *)(discovery_bin + offset);
+
+	if (le32_to_cpu(nhdr->table_id) != NPS_INFO_TABLE_ID) {
+		dev_dbg(adev->dev, "invalid ip discovery nps info table id\n");
+		return -EINVAL;
+	}
+
+	if (!amdgpu_discovery_verify_checksum(discovery_bin + offset,
+					      le32_to_cpu(nhdr->size_bytes),
+					      checksum)) {
+		dev_dbg(adev->dev, "invalid nps info data table checksum\n");
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static const char *amdgpu_discovery_get_fw_name(struct amdgpu_device *adev)
+{
+	if (amdgpu_discovery == 2) {
+		/* Assume there is valid discovery TMR in VRAM even if binary is sideloaded */
+		adev->discovery.reserve_tmr = true;
+		return "amdgpu/ip_discovery.bin";
+	}
+
+	switch (adev->asic_type) {
+	case CHIP_VEGA10:
+		return "amdgpu/vega10_ip_discovery.bin";
+	case CHIP_VEGA12:
+		return "amdgpu/vega12_ip_discovery.bin";
+	case CHIP_RAVEN:
+		if (adev->apu_flags & AMD_APU_IS_RAVEN2)
+			return "amdgpu/raven2_ip_discovery.bin";
+		else if (adev->apu_flags & AMD_APU_IS_PICASSO)
+			return "amdgpu/picasso_ip_discovery.bin";
+		else
+			return "amdgpu/raven_ip_discovery.bin";
+	case CHIP_VEGA20:
+		return "amdgpu/vega20_ip_discovery.bin";
+	case CHIP_ARCTURUS:
+		return "amdgpu/arcturus_ip_discovery.bin";
+	case CHIP_ALDEBARAN:
+		return "amdgpu/aldebaran_ip_discovery.bin";
+	default:
+		return NULL;
+	}
+}
+
 static int amdgpu_discovery_init(struct amdgpu_device *adev)
 {
 	struct table_info *info;
 	struct binary_header *bhdr;
-	struct ip_discovery_header *ihdr;
-	struct gpu_info_header *ghdr;
-	const struct firmware *fw;
+	uint8_t *discovery_bin;
+	const char *fw_name;
 	uint16_t offset;
 	uint16_t size;
 	uint16_t checksum;
 	int r;
 
-	adev->mman.discovery_tmr_size = DISCOVERY_TMR_SIZE;
-	adev->mman.discovery_bin = kzalloc(adev->mman.discovery_tmr_size, GFP_KERNEL);
-	if (!adev->mman.discovery_bin)
+	adev->discovery.bin = kzalloc(DISCOVERY_TMR_SIZE, GFP_KERNEL);
+	if (!adev->discovery.bin)
 		return -ENOMEM;
-
-	if (amdgpu_discovery == 2) {
-		r = request_firmware(&fw, "amdgpu/ip_discovery.bin", adev->dev);
+	adev->discovery.size = DISCOVERY_TMR_SIZE;
+	adev->discovery.debugfs_blob.data = adev->discovery.bin;
+	adev->discovery.debugfs_blob.size = adev->discovery.size;
+
+	discovery_bin = adev->discovery.bin;
+	/* Read from file if it is the preferred option */
+	fw_name = amdgpu_discovery_get_fw_name(adev);
+	if (fw_name != NULL) {
+		drm_dbg(&adev->ddev, "use ip discovery information from file");
+		r = amdgpu_discovery_read_binary_from_file(adev, discovery_bin,
+							   fw_name);
 		if (r)
-			goto get_from_vram;
-		dev_info(adev->dev, "Using IP discovery from file\n");
-		memcpy((u8 *)adev->mman.discovery_bin, (u8 *)fw->data,
-		       adev->mman.discovery_tmr_size);
-		release_firmware(fw);
+			goto out;
 	} else {
-get_from_vram:
-		r = amdgpu_discovery_read_binary(adev, adev->mman.discovery_bin);
-		if (r) {
-			DRM_ERROR("failed to read ip discovery binary\n");
+		drm_dbg(&adev->ddev, "use ip discovery information from memory");
+		r = amdgpu_discovery_read_binary_from_mem(adev, discovery_bin);
+		if (r)
 			goto out;
-		}
 	}
 
-	bhdr = (struct binary_header *)adev->mman.discovery_bin;
-
-	if (le32_to_cpu(bhdr->binary_signature) != BINARY_SIGNATURE) {
-		DRM_ERROR("invalid ip discovery binary signature\n");
+	/* check the ip discovery binary signature */
+	if (!amdgpu_discovery_verify_binary_signature(discovery_bin)) {
+		dev_err(adev->dev,
+			"get invalid ip discovery binary signature\n");
 		r = -EINVAL;
 		goto out;
 	}
 
+	bhdr = (struct binary_header *)discovery_bin;
+
 	offset = offsetof(struct binary_header, binary_checksum) +
 		sizeof(bhdr->binary_checksum);
 	size = le16_to_cpu(bhdr->binary_size) - offset;
 	checksum = le16_to_cpu(bhdr->binary_checksum);
 
-	if (!amdgpu_discovery_verify_checksum(adev->mman.discovery_bin + offset,
-					      size, checksum)) {
-		DRM_ERROR("invalid ip discovery binary checksum\n");
+	if (!amdgpu_discovery_verify_checksum(discovery_bin + offset, size,
+					      checksum)) {
+		dev_err(adev->dev, "invalid ip discovery binary checksum\n");
 		r = -EINVAL;
 		goto out;
 	}
@@ -263,95 +528,880 @@ get_from_vram:
 	info = &bhdr->table_list[IP_DISCOVERY];
 	offset = le16_to_cpu(info->offset);
 	checksum = le16_to_cpu(info->checksum);
-	ihdr = (struct ip_discovery_header *)(adev->mman.discovery_bin + offset);
 
-	if (le32_to_cpu(ihdr->signature) != DISCOVERY_TABLE_SIGNATURE) {
-		DRM_ERROR("invalid ip discovery data table signature\n");
-		r = -EINVAL;
-		goto out;
-	}
+	if (offset) {
+		struct ip_discovery_header *ihdr =
+			(struct ip_discovery_header *)(discovery_bin + offset);
+		if (le32_to_cpu(ihdr->signature) != DISCOVERY_TABLE_SIGNATURE) {
+			dev_err(adev->dev, "invalid ip discovery data table signature\n");
+			r = -EINVAL;
+			goto out;
+		}
 
-	if (!amdgpu_discovery_verify_checksum(adev->mman.discovery_bin + offset,
-					      le16_to_cpu(ihdr->size), checksum)) {
-		DRM_ERROR("invalid ip discovery data table checksum\n");
-		r = -EINVAL;
-		goto out;
+		if (!amdgpu_discovery_verify_checksum(discovery_bin + offset,
+						      le16_to_cpu(ihdr->size),
+						      checksum)) {
+			dev_err(adev->dev, "invalid ip discovery data table checksum\n");
+			r = -EINVAL;
+			goto out;
+		}
 	}
 
 	info = &bhdr->table_list[GC];
 	offset = le16_to_cpu(info->offset);
 	checksum = le16_to_cpu(info->checksum);
-	ghdr = (struct gpu_info_header *)(adev->mman.discovery_bin + offset);
 
-	if (!amdgpu_discovery_verify_checksum(adev->mman.discovery_bin + offset,
-				              le32_to_cpu(ghdr->size), checksum)) {
-		DRM_ERROR("invalid gc data table checksum\n");
-		r = -EINVAL;
-		goto out;
+	if (offset) {
+		struct gpu_info_header *ghdr =
+			(struct gpu_info_header *)(discovery_bin + offset);
+
+		if (le32_to_cpu(ghdr->table_id) != GC_TABLE_ID) {
+			dev_err(adev->dev, "invalid ip discovery gc table id\n");
+			r = -EINVAL;
+			goto out;
+		}
+
+		if (!amdgpu_discovery_verify_checksum(discovery_bin + offset,
+						      le32_to_cpu(ghdr->size),
+						      checksum)) {
+			dev_err(adev->dev, "invalid gc data table checksum\n");
+			r = -EINVAL;
+			goto out;
+		}
+	}
+
+	info = &bhdr->table_list[HARVEST_INFO];
+	offset = le16_to_cpu(info->offset);
+	checksum = le16_to_cpu(info->checksum);
+
+	if (offset) {
+		struct harvest_info_header *hhdr =
+			(struct harvest_info_header *)(discovery_bin + offset);
+
+		if (le32_to_cpu(hhdr->signature) != HARVEST_TABLE_SIGNATURE) {
+			dev_err(adev->dev, "invalid ip discovery harvest table signature\n");
+			r = -EINVAL;
+			goto out;
+		}
+
+		if (!amdgpu_discovery_verify_checksum(
+			    discovery_bin + offset,
+			    sizeof(struct harvest_table), checksum)) {
+			dev_err(adev->dev, "invalid harvest data table checksum\n");
+			r = -EINVAL;
+			goto out;
+		}
+	}
+
+	info = &bhdr->table_list[VCN_INFO];
+	offset = le16_to_cpu(info->offset);
+	checksum = le16_to_cpu(info->checksum);
+
+	if (offset) {
+		struct vcn_info_header *vhdr =
+			(struct vcn_info_header *)(discovery_bin + offset);
+
+		if (le32_to_cpu(vhdr->table_id) != VCN_INFO_TABLE_ID) {
+			dev_err(adev->dev, "invalid ip discovery vcn table id\n");
+			r = -EINVAL;
+			goto out;
+		}
+
+		if (!amdgpu_discovery_verify_checksum(
+			    discovery_bin + offset,
+			    le32_to_cpu(vhdr->size_bytes), checksum)) {
+			dev_err(adev->dev, "invalid vcn data table checksum\n");
+			r = -EINVAL;
+			goto out;
+		}
+	}
+
+	info = &bhdr->table_list[MALL_INFO];
+	offset = le16_to_cpu(info->offset);
+	checksum = le16_to_cpu(info->checksum);
+
+	if (0 && offset) {
+		struct mall_info_header *mhdr =
+			(struct mall_info_header *)(discovery_bin + offset);
+
+		if (le32_to_cpu(mhdr->table_id) != MALL_INFO_TABLE_ID) {
+			dev_err(adev->dev, "invalid ip discovery mall table id\n");
+			r = -EINVAL;
+			goto out;
+		}
+
+		if (!amdgpu_discovery_verify_checksum(
+			    discovery_bin + offset,
+			    le32_to_cpu(mhdr->size_bytes), checksum)) {
+			dev_err(adev->dev, "invalid mall data table checksum\n");
+			r = -EINVAL;
+			goto out;
+		}
 	}
 
 	return 0;
 
 out:
-	kfree(adev->mman.discovery_bin);
-	adev->mman.discovery_bin = NULL;
-
+	kfree(adev->discovery.bin);
+	adev->discovery.bin = NULL;
+	if ((amdgpu_discovery != 2) &&
+	    (RREG32(mmIP_DISCOVERY_VERSION) == 4))
+		amdgpu_ras_query_boot_status(adev, 4);
 	return r;
 }
 
+static void amdgpu_discovery_sysfs_fini(struct amdgpu_device *adev);
+
 void amdgpu_discovery_fini(struct amdgpu_device *adev)
 {
-	kfree(adev->mman.discovery_bin);
-	adev->mman.discovery_bin = NULL;
+	amdgpu_discovery_sysfs_fini(adev);
+	kfree(adev->discovery.bin);
+	adev->discovery.bin = NULL;
 }
 
-static int amdgpu_discovery_validate_ip(const struct ip *ip)
+static int amdgpu_discovery_validate_ip(struct amdgpu_device *adev,
+					uint8_t instance, uint16_t hw_id)
 {
-	if (ip->number_instance >= HWIP_MAX_INSTANCE) {
-		DRM_ERROR("Unexpected number_instance (%d) from ip discovery blob\n",
-			  ip->number_instance);
+	if (instance >= HWIP_MAX_INSTANCE) {
+		dev_err(adev->dev,
+			"Unexpected instance_number (%d) from ip discovery blob\n",
+			instance);
 		return -EINVAL;
 	}
-	if (le16_to_cpu(ip->hw_id) >= HW_ID_MAX) {
-		DRM_ERROR("Unexpected hw_id (%d) from ip discovery blob\n",
-			  le16_to_cpu(ip->hw_id));
+	if (hw_id >= HW_ID_MAX) {
+		dev_err(adev->dev,
+			"Unexpected hw_id (%d) from ip discovery blob\n",
+			hw_id);
 		return -EINVAL;
 	}
 
 	return 0;
 }
 
-int amdgpu_discovery_reg_base_init(struct amdgpu_device *adev)
+static void amdgpu_discovery_read_harvest_bit_per_ip(struct amdgpu_device *adev,
+						uint32_t *vcn_harvest_count)
 {
+	uint8_t *discovery_bin = adev->discovery.bin;
 	struct binary_header *bhdr;
 	struct ip_discovery_header *ihdr;
 	struct die_header *dhdr;
 	struct ip *ip;
+	uint16_t die_offset, ip_offset, num_dies, num_ips;
+	uint16_t hw_id;
+	uint8_t inst;
+	int i, j;
+
+	bhdr = (struct binary_header *)discovery_bin;
+	ihdr = (struct ip_discovery_header
+			*)(discovery_bin +
+			   le16_to_cpu(bhdr->table_list[IP_DISCOVERY].offset));
+	num_dies = le16_to_cpu(ihdr->num_dies);
+
+	/* scan harvest bit of all IP data structures */
+	for (i = 0; i < num_dies; i++) {
+		die_offset = le16_to_cpu(ihdr->die_info[i].die_offset);
+		dhdr = (struct die_header *)(discovery_bin + die_offset);
+		num_ips = le16_to_cpu(dhdr->num_ips);
+		ip_offset = die_offset + sizeof(*dhdr);
+
+		for (j = 0; j < num_ips; j++) {
+			ip = (struct ip *)(discovery_bin + ip_offset);
+			inst = ip->number_instance;
+			hw_id = le16_to_cpu(ip->hw_id);
+			if (amdgpu_discovery_validate_ip(adev, inst, hw_id))
+				goto next_ip;
+
+			if (ip->harvest == 1) {
+				switch (hw_id) {
+				case VCN_HWID:
+					(*vcn_harvest_count)++;
+					if (inst == 0) {
+						adev->vcn.harvest_config |= AMDGPU_VCN_HARVEST_VCN0;
+						adev->vcn.inst_mask &=
+							~AMDGPU_VCN_HARVEST_VCN0;
+						adev->jpeg.inst_mask &=
+							~AMDGPU_VCN_HARVEST_VCN0;
+					} else {
+						adev->vcn.harvest_config |= AMDGPU_VCN_HARVEST_VCN1;
+						adev->vcn.inst_mask &=
+							~AMDGPU_VCN_HARVEST_VCN1;
+						adev->jpeg.inst_mask &=
+							~AMDGPU_VCN_HARVEST_VCN1;
+					}
+					break;
+				case DMU_HWID:
+					adev->harvest_ip_mask |= AMD_HARVEST_IP_DMU_MASK;
+					break;
+				default:
+					break;
+				}
+			}
+next_ip:
+			ip_offset += struct_size(ip, base_address,
+						 ip->num_base_address);
+		}
+	}
+}
+
+static void amdgpu_discovery_read_from_harvest_table(struct amdgpu_device *adev,
+						     uint32_t *vcn_harvest_count,
+						     uint32_t *umc_harvest_count)
+{
+	uint8_t *discovery_bin = adev->discovery.bin;
+	struct binary_header *bhdr;
+	struct harvest_table *harvest_info;
+	u16 offset;
+	int i;
+	uint32_t umc_harvest_config = 0;
+
+	bhdr = (struct binary_header *)discovery_bin;
+	offset = le16_to_cpu(bhdr->table_list[HARVEST_INFO].offset);
+
+	if (!offset) {
+		dev_err(adev->dev, "invalid harvest table offset\n");
+		return;
+	}
+
+	harvest_info = (struct harvest_table *)(discovery_bin + offset);
+
+	for (i = 0; i < 32; i++) {
+		if (le16_to_cpu(harvest_info->list[i].hw_id) == 0)
+			break;
+
+		switch (le16_to_cpu(harvest_info->list[i].hw_id)) {
+		case VCN_HWID:
+			(*vcn_harvest_count)++;
+			adev->vcn.harvest_config |=
+				(1 << harvest_info->list[i].number_instance);
+			adev->jpeg.harvest_config |=
+				(1 << harvest_info->list[i].number_instance);
+
+			adev->vcn.inst_mask &=
+				~(1U << harvest_info->list[i].number_instance);
+			adev->jpeg.inst_mask &=
+				~(1U << harvest_info->list[i].number_instance);
+			break;
+		case DMU_HWID:
+			adev->harvest_ip_mask |= AMD_HARVEST_IP_DMU_MASK;
+			break;
+		case UMC_HWID:
+			umc_harvest_config |=
+				1 << (le16_to_cpu(harvest_info->list[i].number_instance));
+			(*umc_harvest_count)++;
+			break;
+		case GC_HWID:
+			adev->gfx.xcc_mask &=
+				~(1U << harvest_info->list[i].number_instance);
+			break;
+		case SDMA0_HWID:
+			adev->sdma.sdma_mask &=
+				~(1U << harvest_info->list[i].number_instance);
+			break;
+#if defined(CONFIG_DRM_AMD_ISP)
+		case ISP_HWID:
+			adev->isp.harvest_config |=
+				~(1U << harvest_info->list[i].number_instance);
+			break;
+#endif
+		default:
+			break;
+		}
+	}
+
+	adev->umc.active_mask = ((1 << adev->umc.node_inst_num) - 1) &
+				~umc_harvest_config;
+}
+
+/* ================================================== */
+
+struct ip_hw_instance {
+	struct kobject kobj; /* ip_discovery/die/#die/#hw_id/#instance/<attrs...> */
+
+	int hw_id;
+	u8  num_instance;
+	u8  major, minor, revision;
+	u8  harvest;
+
+	int num_base_addresses;
+	u32 base_addr[] __counted_by(num_base_addresses);
+};
+
+struct ip_hw_id {
+	struct kset hw_id_kset;  /* ip_discovery/die/#die/#hw_id/, contains ip_hw_instance */
+	int hw_id;
+};
+
+struct ip_die_entry {
+	struct kset ip_kset;     /* ip_discovery/die/#die/, contains ip_hw_id  */
+	u16 num_ips;
+};
+
+/* -------------------------------------------------- */
+
+struct ip_hw_instance_attr {
+	struct attribute attr;
+	ssize_t (*show)(struct ip_hw_instance *ip_hw_instance, char *buf);
+};
+
+static ssize_t hw_id_show(struct ip_hw_instance *ip_hw_instance, char *buf)
+{
+	return sysfs_emit(buf, "%d\n", ip_hw_instance->hw_id);
+}
+
+static ssize_t num_instance_show(struct ip_hw_instance *ip_hw_instance, char *buf)
+{
+	return sysfs_emit(buf, "%d\n", ip_hw_instance->num_instance);
+}
+
+static ssize_t major_show(struct ip_hw_instance *ip_hw_instance, char *buf)
+{
+	return sysfs_emit(buf, "%d\n", ip_hw_instance->major);
+}
+
+static ssize_t minor_show(struct ip_hw_instance *ip_hw_instance, char *buf)
+{
+	return sysfs_emit(buf, "%d\n", ip_hw_instance->minor);
+}
+
+static ssize_t revision_show(struct ip_hw_instance *ip_hw_instance, char *buf)
+{
+	return sysfs_emit(buf, "%d\n", ip_hw_instance->revision);
+}
+
+static ssize_t harvest_show(struct ip_hw_instance *ip_hw_instance, char *buf)
+{
+	return sysfs_emit(buf, "0x%01X\n", ip_hw_instance->harvest);
+}
+
+static ssize_t num_base_addresses_show(struct ip_hw_instance *ip_hw_instance, char *buf)
+{
+	return sysfs_emit(buf, "%d\n", ip_hw_instance->num_base_addresses);
+}
+
+static ssize_t base_addr_show(struct ip_hw_instance *ip_hw_instance, char *buf)
+{
+	ssize_t res, at;
+	int ii;
+
+	for (res = at = ii = 0; ii < ip_hw_instance->num_base_addresses; ii++) {
+		/* Here we satisfy the condition that, at + size <= PAGE_SIZE.
+		 */
+		if (at + 12 > PAGE_SIZE)
+			break;
+		res = sysfs_emit_at(buf, at, "0x%08X\n",
+				    ip_hw_instance->base_addr[ii]);
+		if (res <= 0)
+			break;
+		at += res;
+	}
+
+	return res < 0 ? res : at;
+}
+
+static struct ip_hw_instance_attr ip_hw_attr[] = {
+	__ATTR_RO(hw_id),
+	__ATTR_RO(num_instance),
+	__ATTR_RO(major),
+	__ATTR_RO(minor),
+	__ATTR_RO(revision),
+	__ATTR_RO(harvest),
+	__ATTR_RO(num_base_addresses),
+	__ATTR_RO(base_addr),
+};
+
+static struct attribute *ip_hw_instance_attrs[ARRAY_SIZE(ip_hw_attr) + 1];
+ATTRIBUTE_GROUPS(ip_hw_instance);
+
+#define to_ip_hw_instance(x) container_of(x, struct ip_hw_instance, kobj)
+#define to_ip_hw_instance_attr(x) container_of(x, struct ip_hw_instance_attr, attr)
+
+static ssize_t ip_hw_instance_attr_show(struct kobject *kobj,
+					struct attribute *attr,
+					char *buf)
+{
+	struct ip_hw_instance *ip_hw_instance = to_ip_hw_instance(kobj);
+	struct ip_hw_instance_attr *ip_hw_attr = to_ip_hw_instance_attr(attr);
+
+	if (!ip_hw_attr->show)
+		return -EIO;
+
+	return ip_hw_attr->show(ip_hw_instance, buf);
+}
+
+static const struct sysfs_ops ip_hw_instance_sysfs_ops = {
+	.show = ip_hw_instance_attr_show,
+};
+
+static void ip_hw_instance_release(struct kobject *kobj)
+{
+	struct ip_hw_instance *ip_hw_instance = to_ip_hw_instance(kobj);
+
+	kfree(ip_hw_instance);
+}
+
+static const struct kobj_type ip_hw_instance_ktype = {
+	.release = ip_hw_instance_release,
+	.sysfs_ops = &ip_hw_instance_sysfs_ops,
+	.default_groups = ip_hw_instance_groups,
+};
+
+/* -------------------------------------------------- */
+
+#define to_ip_hw_id(x)  container_of(to_kset(x), struct ip_hw_id, hw_id_kset)
+
+static void ip_hw_id_release(struct kobject *kobj)
+{
+	struct ip_hw_id *ip_hw_id = to_ip_hw_id(kobj);
+
+	if (!list_empty(&ip_hw_id->hw_id_kset.list))
+		DRM_ERROR("ip_hw_id->hw_id_kset is not empty");
+	kfree(ip_hw_id);
+}
+
+static const struct kobj_type ip_hw_id_ktype = {
+	.release = ip_hw_id_release,
+	.sysfs_ops = &kobj_sysfs_ops,
+};
+
+/* -------------------------------------------------- */
+
+static void die_kobj_release(struct kobject *kobj);
+static void ip_disc_release(struct kobject *kobj);
+
+struct ip_die_entry_attribute {
+	struct attribute attr;
+	ssize_t (*show)(struct ip_die_entry *ip_die_entry, char *buf);
+};
+
+#define to_ip_die_entry_attr(x)  container_of(x, struct ip_die_entry_attribute, attr)
+
+static ssize_t num_ips_show(struct ip_die_entry *ip_die_entry, char *buf)
+{
+	return sysfs_emit(buf, "%d\n", ip_die_entry->num_ips);
+}
+
+/* If there are more ip_die_entry attrs, other than the number of IPs,
+ * we can make this intro an array of attrs, and then initialize
+ * ip_die_entry_attrs in a loop.
+ */
+static struct ip_die_entry_attribute num_ips_attr =
+	__ATTR_RO(num_ips);
+
+static struct attribute *ip_die_entry_attrs[] = {
+	&num_ips_attr.attr,
+	NULL,
+};
+ATTRIBUTE_GROUPS(ip_die_entry); /* ip_die_entry_groups */
+
+#define to_ip_die_entry(x) container_of(to_kset(x), struct ip_die_entry, ip_kset)
+
+static ssize_t ip_die_entry_attr_show(struct kobject *kobj,
+				      struct attribute *attr,
+				      char *buf)
+{
+	struct ip_die_entry_attribute *ip_die_entry_attr = to_ip_die_entry_attr(attr);
+	struct ip_die_entry *ip_die_entry = to_ip_die_entry(kobj);
+
+	if (!ip_die_entry_attr->show)
+		return -EIO;
+
+	return ip_die_entry_attr->show(ip_die_entry, buf);
+}
+
+static void ip_die_entry_release(struct kobject *kobj)
+{
+	struct ip_die_entry *ip_die_entry = to_ip_die_entry(kobj);
+
+	if (!list_empty(&ip_die_entry->ip_kset.list))
+		DRM_ERROR("ip_die_entry->ip_kset is not empty");
+	kfree(ip_die_entry);
+}
+
+static const struct sysfs_ops ip_die_entry_sysfs_ops = {
+	.show = ip_die_entry_attr_show,
+};
+
+static const struct kobj_type ip_die_entry_ktype = {
+	.release = ip_die_entry_release,
+	.sysfs_ops = &ip_die_entry_sysfs_ops,
+	.default_groups = ip_die_entry_groups,
+};
+
+static const struct kobj_type die_kobj_ktype = {
+	.release = die_kobj_release,
+	.sysfs_ops = &kobj_sysfs_ops,
+};
+
+static const struct kobj_type ip_discovery_ktype = {
+	.release = ip_disc_release,
+	.sysfs_ops = &kobj_sysfs_ops,
+};
+
+struct ip_discovery_top {
+	struct kobject kobj;    /* ip_discovery/ */
+	struct kset die_kset;   /* ip_discovery/die/, contains ip_die_entry */
+	struct amdgpu_device *adev;
+};
+
+static void die_kobj_release(struct kobject *kobj)
+{
+	struct ip_discovery_top *ip_top = container_of(to_kset(kobj),
+						       struct ip_discovery_top,
+						       die_kset);
+	if (!list_empty(&ip_top->die_kset.list))
+		DRM_ERROR("ip_top->die_kset is not empty");
+}
+
+static void ip_disc_release(struct kobject *kobj)
+{
+	struct ip_discovery_top *ip_top = container_of(kobj, struct ip_discovery_top,
+						       kobj);
+	struct amdgpu_device *adev = ip_top->adev;
+
+	kfree(ip_top);
+	adev->discovery.ip_top = NULL;
+}
+
+static uint8_t amdgpu_discovery_get_harvest_info(struct amdgpu_device *adev,
+						 uint16_t hw_id, uint8_t inst)
+{
+	uint8_t harvest = 0;
+
+	/* Until a uniform way is figured, get mask based on hwid */
+	switch (hw_id) {
+	case VCN_HWID:
+		/* VCN vs UVD+VCE */
+		if (!amdgpu_ip_version(adev, VCE_HWIP, 0))
+			harvest = ((1 << inst) & adev->vcn.inst_mask) == 0;
+		break;
+	case DMU_HWID:
+		if (adev->harvest_ip_mask & AMD_HARVEST_IP_DMU_MASK)
+			harvest = 0x1;
+		break;
+	case UMC_HWID:
+		/* TODO: It needs another parsing; for now, ignore.*/
+		break;
+	case GC_HWID:
+		harvest = ((1 << inst) & adev->gfx.xcc_mask) == 0;
+		break;
+	case SDMA0_HWID:
+		harvest = ((1 << inst) & adev->sdma.sdma_mask) == 0;
+		break;
+	default:
+		break;
+	}
+
+	return harvest;
+}
+
+static int amdgpu_discovery_sysfs_ips(struct amdgpu_device *adev,
+				      struct ip_die_entry *ip_die_entry,
+				      const size_t _ip_offset, const int num_ips,
+				      bool reg_base_64)
+{
+	uint8_t *discovery_bin = adev->discovery.bin;
+	int ii, jj, kk, res;
+	uint16_t hw_id;
+	uint8_t inst;
+
+	DRM_DEBUG("num_ips:%d", num_ips);
+
+	/* Find all IPs of a given HW ID, and add their instance to
+	 * #die/#hw_id/#instance/<attributes>
+	 */
+	for (ii = 0; ii < HW_ID_MAX; ii++) {
+		struct ip_hw_id *ip_hw_id = NULL;
+		size_t ip_offset = _ip_offset;
+
+		for (jj = 0; jj < num_ips; jj++) {
+			struct ip_v4 *ip;
+			struct ip_hw_instance *ip_hw_instance;
+
+			ip = (struct ip_v4 *)(discovery_bin + ip_offset);
+			inst = ip->instance_number;
+			hw_id = le16_to_cpu(ip->hw_id);
+			if (amdgpu_discovery_validate_ip(adev, inst, hw_id) ||
+			    hw_id != ii)
+				goto next_ip;
+
+			DRM_DEBUG("match:%d @ ip_offset:%zu", ii, ip_offset);
+
+			/* We have a hw_id match; register the hw
+			 * block if not yet registered.
+			 */
+			if (!ip_hw_id) {
+				ip_hw_id = kzalloc(sizeof(*ip_hw_id), GFP_KERNEL);
+				if (!ip_hw_id)
+					return -ENOMEM;
+				ip_hw_id->hw_id = ii;
+
+				kobject_set_name(&ip_hw_id->hw_id_kset.kobj, "%d", ii);
+				ip_hw_id->hw_id_kset.kobj.kset = &ip_die_entry->ip_kset;
+				ip_hw_id->hw_id_kset.kobj.ktype = &ip_hw_id_ktype;
+				res = kset_register(&ip_hw_id->hw_id_kset);
+				if (res) {
+					DRM_ERROR("Couldn't register ip_hw_id kset");
+					kfree(ip_hw_id);
+					return res;
+				}
+				if (hw_id_names[ii]) {
+					res = sysfs_create_link(&ip_die_entry->ip_kset.kobj,
+								&ip_hw_id->hw_id_kset.kobj,
+								hw_id_names[ii]);
+					if (res) {
+						DRM_ERROR("Couldn't create IP link %s in IP Die:%s\n",
+							  hw_id_names[ii],
+							  kobject_name(&ip_die_entry->ip_kset.kobj));
+					}
+				}
+			}
+
+			/* Now register its instance.
+			 */
+			ip_hw_instance = kzalloc(struct_size(ip_hw_instance,
+							     base_addr,
+							     ip->num_base_address),
+						 GFP_KERNEL);
+			if (!ip_hw_instance) {
+				DRM_ERROR("no memory for ip_hw_instance");
+				return -ENOMEM;
+			}
+			ip_hw_instance->hw_id = le16_to_cpu(ip->hw_id); /* == ii */
+			ip_hw_instance->num_instance = ip->instance_number;
+			ip_hw_instance->major = ip->major;
+			ip_hw_instance->minor = ip->minor;
+			ip_hw_instance->revision = ip->revision;
+			ip_hw_instance->harvest =
+				amdgpu_discovery_get_harvest_info(
+					adev, ip_hw_instance->hw_id,
+					ip_hw_instance->num_instance);
+			ip_hw_instance->num_base_addresses = ip->num_base_address;
+
+			for (kk = 0; kk < ip_hw_instance->num_base_addresses; kk++) {
+				if (reg_base_64)
+					ip_hw_instance->base_addr[kk] =
+						lower_32_bits(le64_to_cpu(ip->base_address_64[kk])) & 0x3FFFFFFF;
+				else
+					ip_hw_instance->base_addr[kk] = ip->base_address[kk];
+			}
+
+			kobject_init(&ip_hw_instance->kobj, &ip_hw_instance_ktype);
+			ip_hw_instance->kobj.kset = &ip_hw_id->hw_id_kset;
+			res = kobject_add(&ip_hw_instance->kobj, NULL,
+					  "%d", ip_hw_instance->num_instance);
+next_ip:
+			if (reg_base_64)
+				ip_offset += struct_size(ip, base_address_64,
+							 ip->num_base_address);
+			else
+				ip_offset += struct_size(ip, base_address,
+							 ip->num_base_address);
+		}
+	}
+
+	return 0;
+}
+
+static int amdgpu_discovery_sysfs_recurse(struct amdgpu_device *adev)
+{
+	struct ip_discovery_top *ip_top = adev->discovery.ip_top;
+	uint8_t *discovery_bin = adev->discovery.bin;
+	struct binary_header *bhdr;
+	struct ip_discovery_header *ihdr;
+	struct die_header *dhdr;
+	struct kset *die_kset = &ip_top->die_kset;
+	u16 num_dies, die_offset, num_ips;
+	size_t ip_offset;
+	int ii, res;
+
+	bhdr = (struct binary_header *)discovery_bin;
+	ihdr = (struct ip_discovery_header
+			*)(discovery_bin +
+			   le16_to_cpu(bhdr->table_list[IP_DISCOVERY].offset));
+	num_dies = le16_to_cpu(ihdr->num_dies);
+
+	DRM_DEBUG("number of dies: %d\n", num_dies);
+
+	for (ii = 0; ii < num_dies; ii++) {
+		struct ip_die_entry *ip_die_entry;
+
+		die_offset = le16_to_cpu(ihdr->die_info[ii].die_offset);
+		dhdr = (struct die_header *)(discovery_bin + die_offset);
+		num_ips = le16_to_cpu(dhdr->num_ips);
+		ip_offset = die_offset + sizeof(*dhdr);
+
+		/* Add the die to the kset.
+		 *
+		 * dhdr->die_id == ii, which was checked in
+		 * amdgpu_discovery_reg_base_init().
+		 */
+
+		ip_die_entry = kzalloc(sizeof(*ip_die_entry), GFP_KERNEL);
+		if (!ip_die_entry)
+			return -ENOMEM;
+
+		ip_die_entry->num_ips = num_ips;
+
+		kobject_set_name(&ip_die_entry->ip_kset.kobj, "%d", le16_to_cpu(dhdr->die_id));
+		ip_die_entry->ip_kset.kobj.kset = die_kset;
+		ip_die_entry->ip_kset.kobj.ktype = &ip_die_entry_ktype;
+		res = kset_register(&ip_die_entry->ip_kset);
+		if (res) {
+			DRM_ERROR("Couldn't register ip_die_entry kset");
+			kfree(ip_die_entry);
+			return res;
+		}
+
+		amdgpu_discovery_sysfs_ips(adev, ip_die_entry, ip_offset, num_ips, !!ihdr->base_addr_64_bit);
+	}
+
+	return 0;
+}
+
+static int amdgpu_discovery_sysfs_init(struct amdgpu_device *adev)
+{
+	uint8_t *discovery_bin = adev->discovery.bin;
+	struct ip_discovery_top *ip_top;
+	struct kset *die_kset;
+	int res, ii;
+
+	if (!discovery_bin)
+		return -EINVAL;
+
+	ip_top = kzalloc(sizeof(*ip_top), GFP_KERNEL);
+	if (!ip_top)
+		return -ENOMEM;
+
+	ip_top->adev = adev;
+	adev->discovery.ip_top = ip_top;
+	res = kobject_init_and_add(&ip_top->kobj, &ip_discovery_ktype,
+				   &adev->dev->kobj, "ip_discovery");
+	if (res) {
+		DRM_ERROR("Couldn't init and add ip_discovery/");
+		goto Err;
+	}
+
+	die_kset = &ip_top->die_kset;
+	kobject_set_name(&die_kset->kobj, "%s", "die");
+	die_kset->kobj.parent = &ip_top->kobj;
+	die_kset->kobj.ktype = &die_kobj_ktype;
+	res = kset_register(&ip_top->die_kset);
+	if (res) {
+		DRM_ERROR("Couldn't register die_kset");
+		goto Err;
+	}
+
+	for (ii = 0; ii < ARRAY_SIZE(ip_hw_attr); ii++)
+		ip_hw_instance_attrs[ii] = &ip_hw_attr[ii].attr;
+	ip_hw_instance_attrs[ii] = NULL;
+
+	res = amdgpu_discovery_sysfs_recurse(adev);
+
+	return res;
+Err:
+	kobject_put(&ip_top->kobj);
+	return res;
+}
+
+/* -------------------------------------------------- */
+
+#define list_to_kobj(el) container_of(el, struct kobject, entry)
+
+static void amdgpu_discovery_sysfs_ip_hw_free(struct ip_hw_id *ip_hw_id)
+{
+	struct list_head *el, *tmp;
+	struct kset *hw_id_kset;
+
+	hw_id_kset = &ip_hw_id->hw_id_kset;
+	spin_lock(&hw_id_kset->list_lock);
+	list_for_each_prev_safe(el, tmp, &hw_id_kset->list) {
+		list_del_init(el);
+		spin_unlock(&hw_id_kset->list_lock);
+		/* kobject is embedded in ip_hw_instance */
+		kobject_put(list_to_kobj(el));
+		spin_lock(&hw_id_kset->list_lock);
+	}
+	spin_unlock(&hw_id_kset->list_lock);
+	kobject_put(&ip_hw_id->hw_id_kset.kobj);
+}
+
+static void amdgpu_discovery_sysfs_die_free(struct ip_die_entry *ip_die_entry)
+{
+	struct list_head *el, *tmp;
+	struct kset *ip_kset;
+
+	ip_kset = &ip_die_entry->ip_kset;
+	spin_lock(&ip_kset->list_lock);
+	list_for_each_prev_safe(el, tmp, &ip_kset->list) {
+		list_del_init(el);
+		spin_unlock(&ip_kset->list_lock);
+		amdgpu_discovery_sysfs_ip_hw_free(to_ip_hw_id(list_to_kobj(el)));
+		spin_lock(&ip_kset->list_lock);
+	}
+	spin_unlock(&ip_kset->list_lock);
+	kobject_put(&ip_die_entry->ip_kset.kobj);
+}
+
+static void amdgpu_discovery_sysfs_fini(struct amdgpu_device *adev)
+{
+	struct ip_discovery_top *ip_top = adev->discovery.ip_top;
+	struct list_head *el, *tmp;
+	struct kset *die_kset;
+
+	die_kset = &ip_top->die_kset;
+	spin_lock(&die_kset->list_lock);
+	list_for_each_prev_safe(el, tmp, &die_kset->list) {
+		list_del_init(el);
+		spin_unlock(&die_kset->list_lock);
+		amdgpu_discovery_sysfs_die_free(to_ip_die_entry(list_to_kobj(el)));
+		spin_lock(&die_kset->list_lock);
+	}
+	spin_unlock(&die_kset->list_lock);
+	kobject_put(&ip_top->die_kset.kobj);
+	kobject_put(&ip_top->kobj);
+}
+
+/* ================================================== */
+
+static int amdgpu_discovery_reg_base_init(struct amdgpu_device *adev)
+{
+	uint8_t num_base_address, subrev, variant;
+	struct binary_header *bhdr;
+	struct ip_discovery_header *ihdr;
+	struct die_header *dhdr;
+	uint8_t *discovery_bin;
+	struct ip_v4 *ip;
 	uint16_t die_offset;
 	uint16_t ip_offset;
 	uint16_t num_dies;
+	uint32_t wafl_ver;
 	uint16_t num_ips;
-	uint8_t num_base_address;
+	uint16_t hw_id;
+	uint8_t inst;
 	int hw_ip;
 	int i, j, k;
 	int r;
 
 	r = amdgpu_discovery_init(adev);
-	if (r) {
-		DRM_ERROR("amdgpu_discovery_init failed\n");
+	if (r)
 		return r;
-	}
-
-	bhdr = (struct binary_header *)adev->mman.discovery_bin;
-	ihdr = (struct ip_discovery_header *)(adev->mman.discovery_bin +
-			le16_to_cpu(bhdr->table_list[IP_DISCOVERY].offset));
+	discovery_bin = adev->discovery.bin;
+	wafl_ver = 0;
+	adev->gfx.xcc_mask = 0;
+	adev->sdma.sdma_mask = 0;
+	adev->vcn.inst_mask = 0;
+	adev->jpeg.inst_mask = 0;
+	bhdr = (struct binary_header *)discovery_bin;
+	ihdr = (struct ip_discovery_header
+			*)(discovery_bin +
+			   le16_to_cpu(bhdr->table_list[IP_DISCOVERY].offset));
 	num_dies = le16_to_cpu(ihdr->num_dies);
 
 	DRM_DEBUG("number of dies: %d\n", num_dies);
 
 	for (i = 0; i < num_dies; i++) {
 		die_offset = le16_to_cpu(ihdr->die_info[i].die_offset);
-		dhdr = (struct die_header *)(adev->mman.discovery_bin + die_offset);
+		dhdr = (struct die_header *)(discovery_bin + die_offset);
 		num_ips = le16_to_cpu(dhdr->num_ips);
 		ip_offset = die_offset + sizeof(*dhdr);
 
@@ -365,9 +1415,11 @@ int amdgpu_discovery_reg_base_init(struct amdgpu_device *adev)
 				le16_to_cpu(dhdr->die_id), num_ips);
 
 		for (j = 0; j < num_ips; j++) {
-			ip = (struct ip *)(adev->mman.discovery_bin + ip_offset);
+			ip = (struct ip_v4 *)(discovery_bin + ip_offset);
 
-			if (amdgpu_discovery_validate_ip(ip))
+			inst = ip->instance_number;
+			hw_id = le16_to_cpu(ip->hw_id);
+			if (amdgpu_discovery_validate_ip(adev, inst, hw_id))
 				goto next_ip;
 
 			num_base_address = ip->num_base_address;
@@ -375,32 +1427,99 @@ int amdgpu_discovery_reg_base_init(struct amdgpu_device *adev)
 			DRM_DEBUG("%s(%d) #%d v%d.%d.%d:\n",
 				  hw_id_names[le16_to_cpu(ip->hw_id)],
 				  le16_to_cpu(ip->hw_id),
-				  ip->number_instance,
+				  ip->instance_number,
 				  ip->major, ip->minor,
 				  ip->revision);
 
-			if (le16_to_cpu(ip->hw_id) == VCN_HWID)
-				adev->vcn.num_vcn_inst++;
+			if (le16_to_cpu(ip->hw_id) == VCN_HWID) {
+				/* Bit [5:0]: original revision value
+				 * Bit [7:6]: en/decode capability:
+				 *     0b00 : VCN function normally
+				 *     0b10 : encode is disabled
+				 *     0b01 : decode is disabled
+				 */
+				if (adev->vcn.num_vcn_inst <
+				    AMDGPU_MAX_VCN_INSTANCES) {
+					adev->vcn.inst[adev->vcn.num_vcn_inst].vcn_config =
+						ip->revision & 0xc0;
+					adev->vcn.num_vcn_inst++;
+					adev->vcn.inst_mask |=
+						(1U << ip->instance_number);
+					adev->jpeg.inst_mask |=
+						(1U << ip->instance_number);
+				} else {
+					dev_err(adev->dev, "Too many VCN instances: %d vs %d\n",
+						adev->vcn.num_vcn_inst + 1,
+						AMDGPU_MAX_VCN_INSTANCES);
+				}
+				ip->revision &= ~0xc0;
+			}
 			if (le16_to_cpu(ip->hw_id) == SDMA0_HWID ||
 			    le16_to_cpu(ip->hw_id) == SDMA1_HWID ||
 			    le16_to_cpu(ip->hw_id) == SDMA2_HWID ||
-			    le16_to_cpu(ip->hw_id) == SDMA3_HWID)
-				adev->sdma.num_instances++;
+			    le16_to_cpu(ip->hw_id) == SDMA3_HWID) {
+				if (adev->sdma.num_instances <
+				    AMDGPU_MAX_SDMA_INSTANCES) {
+					adev->sdma.num_instances++;
+					adev->sdma.sdma_mask |=
+						(1U << ip->instance_number);
+				} else {
+					dev_err(adev->dev, "Too many SDMA instances: %d vs %d\n",
+						adev->sdma.num_instances + 1,
+						AMDGPU_MAX_SDMA_INSTANCES);
+				}
+			}
+
+			if (le16_to_cpu(ip->hw_id) == VPE_HWID) {
+				if (adev->vpe.num_instances < AMDGPU_MAX_VPE_INSTANCES)
+					adev->vpe.num_instances++;
+				else
+					dev_err(adev->dev, "Too many VPE instances: %d vs %d\n",
+						adev->vpe.num_instances + 1,
+						AMDGPU_MAX_VPE_INSTANCES);
+			}
+
+			if (le16_to_cpu(ip->hw_id) == UMC_HWID) {
+				adev->gmc.num_umc++;
+				adev->umc.node_inst_num++;
+			}
+
+			if (le16_to_cpu(ip->hw_id) == GC_HWID)
+				adev->gfx.xcc_mask |=
+					(1U << ip->instance_number);
+
+			if (!wafl_ver && le16_to_cpu(ip->hw_id) == WAFLC_HWID)
+				wafl_ver = IP_VERSION_FULL(ip->major, ip->minor,
+							   ip->revision, 0, 0);
 
 			for (k = 0; k < num_base_address; k++) {
 				/*
 				 * convert the endianness of base addresses in place,
 				 * so that we don't need to convert them when accessing adev->reg_offset.
 				 */
-				ip->base_address[k] = le32_to_cpu(ip->base_address[k]);
+				if (ihdr->base_addr_64_bit)
+					/* Truncate the 64bit base address from ip discovery
+					 * and only store lower 32bit ip base in reg_offset[].
+					 * Bits > 32 follows ASIC specific format, thus just
+					 * discard them and handle it within specific ASIC.
+					 * By this way reg_offset[] and related helpers can
+					 * stay unchanged.
+					 * The base address is in dwords, thus clear the
+					 * highest 2 bits to store.
+					 */
+					ip->base_address[k] =
+						lower_32_bits(le64_to_cpu(ip->base_address_64[k])) & 0x3FFFFFFF;
+				else
+					ip->base_address[k] = le32_to_cpu(ip->base_address[k]);
 				DRM_DEBUG("\t0x%08x\n", ip->base_address[k]);
 			}
 
 			for (hw_ip = 0; hw_ip < MAX_HWIP; hw_ip++) {
-				if (hw_id_map[hw_ip] == le16_to_cpu(ip->hw_id)) {
+				if (hw_id_map[hw_ip] == le16_to_cpu(ip->hw_id) &&
+				    hw_id_map[hw_ip] != 0) {
 					DRM_DEBUG("set register base offset for %s\n",
 							hw_id_names[le16_to_cpu(ip->hw_id)]);
-					adev->reg_offset[hw_ip][ip->number_instance] =
+					adev->reg_offset[hw_ip][ip->instance_number] =
 						ip->base_address;
 					/* Instance support is somewhat inconsistent.
 					 * SDMA is a good example.  Sienna cichlid has 4 total
@@ -411,153 +1530,411 @@ int amdgpu_discovery_reg_base_init(struct amdgpu_device *adev)
 					 * example.  On most chips there are multiple instances
 					 * with the same HWID.
 					 */
-					adev->ip_versions[hw_ip][ip->number_instance] =
-						IP_VERSION(ip->major, ip->minor, ip->revision);
+
+					if (ihdr->version < 3) {
+						subrev = 0;
+						variant = 0;
+					} else {
+						subrev = ip->sub_revision;
+						variant = ip->variant;
+					}
+
+					adev->ip_versions[hw_ip]
+							 [ip->instance_number] =
+						IP_VERSION_FULL(ip->major,
+								ip->minor,
+								ip->revision,
+								variant,
+								subrev);
 				}
 			}
 
 next_ip:
-			ip_offset += sizeof(*ip) + 4 * (ip->num_base_address - 1);
+			if (ihdr->base_addr_64_bit)
+				ip_offset += struct_size(ip, base_address_64, ip->num_base_address);
+			else
+				ip_offset += struct_size(ip, base_address, ip->num_base_address);
 		}
 	}
 
+	if (wafl_ver && !adev->ip_versions[XGMI_HWIP][0])
+		adev->ip_versions[XGMI_HWIP][0] = wafl_ver;
+
 	return 0;
 }
 
-int amdgpu_discovery_get_ip_version(struct amdgpu_device *adev, int hw_id, int number_instance,
-				    int *major, int *minor, int *revision)
+static void amdgpu_discovery_harvest_ip(struct amdgpu_device *adev)
 {
-	struct binary_header *bhdr;
+	uint8_t *discovery_bin = adev->discovery.bin;
 	struct ip_discovery_header *ihdr;
-	struct die_header *dhdr;
-	struct ip *ip;
-	uint16_t die_offset;
-	uint16_t ip_offset;
-	uint16_t num_dies;
-	uint16_t num_ips;
-	int i, j;
+	struct binary_header *bhdr;
+	int vcn_harvest_count = 0;
+	int umc_harvest_count = 0;
+	uint16_t offset, ihdr_ver;
+
+	bhdr = (struct binary_header *)discovery_bin;
+	offset = le16_to_cpu(bhdr->table_list[IP_DISCOVERY].offset);
+	ihdr = (struct ip_discovery_header *)(discovery_bin + offset);
+	ihdr_ver = le16_to_cpu(ihdr->version);
+	/*
+	 * Harvest table does not fit Navi1x and legacy GPUs,
+	 * so read harvest bit per IP data structure to set
+	 * harvest configuration.
+	 */
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) < IP_VERSION(10, 2, 0) &&
+	    ihdr_ver <= 2) {
+		if ((adev->pdev->device == 0x731E &&
+			(adev->pdev->revision == 0xC6 ||
+			 adev->pdev->revision == 0xC7)) ||
+			(adev->pdev->device == 0x7340 &&
+			 adev->pdev->revision == 0xC9) ||
+			(adev->pdev->device == 0x7360 &&
+			 adev->pdev->revision == 0xC7))
+			amdgpu_discovery_read_harvest_bit_per_ip(adev,
+				&vcn_harvest_count);
+	} else {
+		amdgpu_discovery_read_from_harvest_table(adev,
+							 &vcn_harvest_count,
+							 &umc_harvest_count);
+	}
 
-	if (!adev->mman.discovery_bin) {
-		DRM_ERROR("ip discovery uninitialized\n");
-		return -EINVAL;
+	amdgpu_discovery_harvest_config_quirk(adev);
+
+	if (vcn_harvest_count == adev->vcn.num_vcn_inst) {
+		adev->harvest_ip_mask |= AMD_HARVEST_IP_VCN_MASK;
+		adev->harvest_ip_mask |= AMD_HARVEST_IP_JPEG_MASK;
 	}
 
-	bhdr = (struct binary_header *)adev->mman.discovery_bin;
-	ihdr = (struct ip_discovery_header *)(adev->mman.discovery_bin +
-			le16_to_cpu(bhdr->table_list[IP_DISCOVERY].offset));
-	num_dies = le16_to_cpu(ihdr->num_dies);
+	if (umc_harvest_count < adev->gmc.num_umc) {
+		adev->gmc.num_umc -= umc_harvest_count;
+	}
+}
 
-	for (i = 0; i < num_dies; i++) {
-		die_offset = le16_to_cpu(ihdr->die_info[i].die_offset);
-		dhdr = (struct die_header *)(adev->mman.discovery_bin + die_offset);
-		num_ips = le16_to_cpu(dhdr->num_ips);
-		ip_offset = die_offset + sizeof(*dhdr);
+union gc_info {
+	struct gc_info_v1_0 v1;
+	struct gc_info_v1_1 v1_1;
+	struct gc_info_v1_2 v1_2;
+	struct gc_info_v1_3 v1_3;
+	struct gc_info_v2_0 v2;
+	struct gc_info_v2_1 v2_1;
+};
 
-		for (j = 0; j < num_ips; j++) {
-			ip = (struct ip *)(adev->mman.discovery_bin + ip_offset);
-
-			if ((le16_to_cpu(ip->hw_id) == hw_id) && (ip->number_instance == number_instance)) {
-				if (major)
-					*major = ip->major;
-				if (minor)
-					*minor = ip->minor;
-				if (revision)
-					*revision = ip->revision;
-				return 0;
-			}
-			ip_offset += sizeof(*ip) + 4 * (ip->num_base_address - 1);
-		}
+static int amdgpu_discovery_get_gfx_info(struct amdgpu_device *adev)
+{
+	uint8_t *discovery_bin = adev->discovery.bin;
+	struct binary_header *bhdr;
+	union gc_info *gc_info;
+	u16 offset;
+
+	if (!discovery_bin) {
+		DRM_ERROR("ip discovery uninitialized\n");
+		return -EINVAL;
 	}
 
-	return -EINVAL;
+	bhdr = (struct binary_header *)discovery_bin;
+	offset = le16_to_cpu(bhdr->table_list[GC].offset);
+
+	if (!offset)
+		return 0;
+
+	gc_info = (union gc_info *)(discovery_bin + offset);
+
+	switch (le16_to_cpu(gc_info->v1.header.version_major)) {
+	case 1:
+		adev->gfx.config.max_shader_engines = le32_to_cpu(gc_info->v1.gc_num_se);
+		adev->gfx.config.max_cu_per_sh = 2 * (le32_to_cpu(gc_info->v1.gc_num_wgp0_per_sa) +
+						      le32_to_cpu(gc_info->v1.gc_num_wgp1_per_sa));
+		adev->gfx.config.max_sh_per_se = le32_to_cpu(gc_info->v1.gc_num_sa_per_se);
+		adev->gfx.config.max_backends_per_se = le32_to_cpu(gc_info->v1.gc_num_rb_per_se);
+		adev->gfx.config.max_texture_channel_caches = le32_to_cpu(gc_info->v1.gc_num_gl2c);
+		adev->gfx.config.max_gprs = le32_to_cpu(gc_info->v1.gc_num_gprs);
+		adev->gfx.config.max_gs_threads = le32_to_cpu(gc_info->v1.gc_num_max_gs_thds);
+		adev->gfx.config.gs_vgt_table_depth = le32_to_cpu(gc_info->v1.gc_gs_table_depth);
+		adev->gfx.config.gs_prim_buffer_depth = le32_to_cpu(gc_info->v1.gc_gsprim_buff_depth);
+		adev->gfx.config.double_offchip_lds_buf = le32_to_cpu(gc_info->v1.gc_double_offchip_lds_buffer);
+		adev->gfx.cu_info.wave_front_size = le32_to_cpu(gc_info->v1.gc_wave_size);
+		adev->gfx.cu_info.max_waves_per_simd = le32_to_cpu(gc_info->v1.gc_max_waves_per_simd);
+		adev->gfx.cu_info.max_scratch_slots_per_cu = le32_to_cpu(gc_info->v1.gc_max_scratch_slots_per_cu);
+		adev->gfx.cu_info.lds_size = le32_to_cpu(gc_info->v1.gc_lds_size);
+		adev->gfx.config.num_sc_per_sh = le32_to_cpu(gc_info->v1.gc_num_sc_per_se) /
+			le32_to_cpu(gc_info->v1.gc_num_sa_per_se);
+		adev->gfx.config.num_packer_per_sc = le32_to_cpu(gc_info->v1.gc_num_packer_per_sc);
+		if (le16_to_cpu(gc_info->v1.header.version_minor) >= 1) {
+			adev->gfx.config.gc_num_tcp_per_sa = le32_to_cpu(gc_info->v1_1.gc_num_tcp_per_sa);
+			adev->gfx.config.gc_num_sdp_interface = le32_to_cpu(gc_info->v1_1.gc_num_sdp_interface);
+			adev->gfx.config.gc_num_tcps = le32_to_cpu(gc_info->v1_1.gc_num_tcps);
+		}
+		if (le16_to_cpu(gc_info->v1.header.version_minor) >= 2) {
+			adev->gfx.config.gc_num_tcp_per_wpg = le32_to_cpu(gc_info->v1_2.gc_num_tcp_per_wpg);
+			adev->gfx.config.gc_tcp_l1_size = le32_to_cpu(gc_info->v1_2.gc_tcp_l1_size);
+			adev->gfx.config.gc_num_sqc_per_wgp = le32_to_cpu(gc_info->v1_2.gc_num_sqc_per_wgp);
+			adev->gfx.config.gc_l1_instruction_cache_size_per_sqc = le32_to_cpu(gc_info->v1_2.gc_l1_instruction_cache_size_per_sqc);
+			adev->gfx.config.gc_l1_data_cache_size_per_sqc = le32_to_cpu(gc_info->v1_2.gc_l1_data_cache_size_per_sqc);
+			adev->gfx.config.gc_gl1c_per_sa = le32_to_cpu(gc_info->v1_2.gc_gl1c_per_sa);
+			adev->gfx.config.gc_gl1c_size_per_instance = le32_to_cpu(gc_info->v1_2.gc_gl1c_size_per_instance);
+			adev->gfx.config.gc_gl2c_per_gpu = le32_to_cpu(gc_info->v1_2.gc_gl2c_per_gpu);
+		}
+		if (le16_to_cpu(gc_info->v1.header.version_minor) >= 3) {
+			adev->gfx.config.gc_tcp_size_per_cu = le32_to_cpu(gc_info->v1_3.gc_tcp_size_per_cu);
+			adev->gfx.config.gc_tcp_cache_line_size = le32_to_cpu(gc_info->v1_3.gc_tcp_cache_line_size);
+			adev->gfx.config.gc_instruction_cache_size_per_sqc = le32_to_cpu(gc_info->v1_3.gc_instruction_cache_size_per_sqc);
+			adev->gfx.config.gc_instruction_cache_line_size = le32_to_cpu(gc_info->v1_3.gc_instruction_cache_line_size);
+			adev->gfx.config.gc_scalar_data_cache_size_per_sqc = le32_to_cpu(gc_info->v1_3.gc_scalar_data_cache_size_per_sqc);
+			adev->gfx.config.gc_scalar_data_cache_line_size = le32_to_cpu(gc_info->v1_3.gc_scalar_data_cache_line_size);
+			adev->gfx.config.gc_tcc_size = le32_to_cpu(gc_info->v1_3.gc_tcc_size);
+			adev->gfx.config.gc_tcc_cache_line_size = le32_to_cpu(gc_info->v1_3.gc_tcc_cache_line_size);
+		}
+		break;
+	case 2:
+		adev->gfx.config.max_shader_engines = le32_to_cpu(gc_info->v2.gc_num_se);
+		adev->gfx.config.max_cu_per_sh = le32_to_cpu(gc_info->v2.gc_num_cu_per_sh);
+		adev->gfx.config.max_sh_per_se = le32_to_cpu(gc_info->v2.gc_num_sh_per_se);
+		adev->gfx.config.max_backends_per_se = le32_to_cpu(gc_info->v2.gc_num_rb_per_se);
+		adev->gfx.config.max_texture_channel_caches = le32_to_cpu(gc_info->v2.gc_num_tccs);
+		adev->gfx.config.max_gprs = le32_to_cpu(gc_info->v2.gc_num_gprs);
+		adev->gfx.config.max_gs_threads = le32_to_cpu(gc_info->v2.gc_num_max_gs_thds);
+		adev->gfx.config.gs_vgt_table_depth = le32_to_cpu(gc_info->v2.gc_gs_table_depth);
+		adev->gfx.config.gs_prim_buffer_depth = le32_to_cpu(gc_info->v2.gc_gsprim_buff_depth);
+		adev->gfx.config.double_offchip_lds_buf = le32_to_cpu(gc_info->v2.gc_double_offchip_lds_buffer);
+		adev->gfx.cu_info.wave_front_size = le32_to_cpu(gc_info->v2.gc_wave_size);
+		adev->gfx.cu_info.max_waves_per_simd = le32_to_cpu(gc_info->v2.gc_max_waves_per_simd);
+		adev->gfx.cu_info.max_scratch_slots_per_cu = le32_to_cpu(gc_info->v2.gc_max_scratch_slots_per_cu);
+		adev->gfx.cu_info.lds_size = le32_to_cpu(gc_info->v2.gc_lds_size);
+		adev->gfx.config.num_sc_per_sh = le32_to_cpu(gc_info->v2.gc_num_sc_per_se) /
+			le32_to_cpu(gc_info->v2.gc_num_sh_per_se);
+		adev->gfx.config.num_packer_per_sc = le32_to_cpu(gc_info->v2.gc_num_packer_per_sc);
+		if (le16_to_cpu(gc_info->v2.header.version_minor) == 1) {
+			adev->gfx.config.gc_num_tcp_per_sa = le32_to_cpu(gc_info->v2_1.gc_num_tcp_per_sh);
+			adev->gfx.config.gc_tcp_size_per_cu = le32_to_cpu(gc_info->v2_1.gc_tcp_size_per_cu);
+			adev->gfx.config.gc_num_sdp_interface = le32_to_cpu(gc_info->v2_1.gc_num_sdp_interface); /* per XCD */
+			adev->gfx.config.gc_num_cu_per_sqc = le32_to_cpu(gc_info->v2_1.gc_num_cu_per_sqc);
+			adev->gfx.config.gc_l1_instruction_cache_size_per_sqc = le32_to_cpu(gc_info->v2_1.gc_instruction_cache_size_per_sqc);
+			adev->gfx.config.gc_l1_data_cache_size_per_sqc = le32_to_cpu(gc_info->v2_1.gc_scalar_data_cache_size_per_sqc);
+			adev->gfx.config.gc_tcc_size = le32_to_cpu(gc_info->v2_1.gc_tcc_size); /* per XCD */
+		}
+		break;
+	default:
+		dev_err(adev->dev,
+			"Unhandled GC info table %d.%d\n",
+			le16_to_cpu(gc_info->v1.header.version_major),
+			le16_to_cpu(gc_info->v1.header.version_minor));
+		return -EINVAL;
+	}
+	return 0;
 }
 
+union mall_info {
+	struct mall_info_v1_0 v1;
+	struct mall_info_v2_0 v2;
+};
 
-int amdgpu_discovery_get_vcn_version(struct amdgpu_device *adev, int vcn_instance,
-				     int *major, int *minor, int *revision)
+static int amdgpu_discovery_get_mall_info(struct amdgpu_device *adev)
 {
-	return amdgpu_discovery_get_ip_version(adev, VCN_HWID,
-					       vcn_instance, major, minor, revision);
+	uint8_t *discovery_bin = adev->discovery.bin;
+	struct binary_header *bhdr;
+	union mall_info *mall_info;
+	u32 u, mall_size_per_umc, m_s_present, half_use;
+	u64 mall_size;
+	u16 offset;
+
+	if (!discovery_bin) {
+		DRM_ERROR("ip discovery uninitialized\n");
+		return -EINVAL;
+	}
+
+	bhdr = (struct binary_header *)discovery_bin;
+	offset = le16_to_cpu(bhdr->table_list[MALL_INFO].offset);
+
+	if (!offset)
+		return 0;
+
+	mall_info = (union mall_info *)(discovery_bin + offset);
+
+	switch (le16_to_cpu(mall_info->v1.header.version_major)) {
+	case 1:
+		mall_size = 0;
+		mall_size_per_umc = le32_to_cpu(mall_info->v1.mall_size_per_m);
+		m_s_present = le32_to_cpu(mall_info->v1.m_s_present);
+		half_use = le32_to_cpu(mall_info->v1.m_half_use);
+		for (u = 0; u < adev->gmc.num_umc; u++) {
+			if (m_s_present & (1 << u))
+				mall_size += mall_size_per_umc * 2;
+			else if (half_use & (1 << u))
+				mall_size += mall_size_per_umc / 2;
+			else
+				mall_size += mall_size_per_umc;
+		}
+		adev->gmc.mall_size = mall_size;
+		adev->gmc.m_half_use = half_use;
+		break;
+	case 2:
+		mall_size_per_umc = le32_to_cpu(mall_info->v2.mall_size_per_umc);
+		adev->gmc.mall_size = (uint64_t)mall_size_per_umc * adev->gmc.num_umc;
+		break;
+	default:
+		dev_err(adev->dev,
+			"Unhandled MALL info table %d.%d\n",
+			le16_to_cpu(mall_info->v1.header.version_major),
+			le16_to_cpu(mall_info->v1.header.version_minor));
+		return -EINVAL;
+	}
+	return 0;
 }
 
-void amdgpu_discovery_harvest_ip(struct amdgpu_device *adev)
+union vcn_info {
+	struct vcn_info_v1_0 v1;
+};
+
+static int amdgpu_discovery_get_vcn_info(struct amdgpu_device *adev)
 {
+	uint8_t *discovery_bin = adev->discovery.bin;
 	struct binary_header *bhdr;
-	struct harvest_table *harvest_info;
-	int i, vcn_harvest_count = 0;
+	union vcn_info *vcn_info;
+	u16 offset;
+	int v;
 
-	bhdr = (struct binary_header *)adev->mman.discovery_bin;
-	harvest_info = (struct harvest_table *)(adev->mman.discovery_bin +
-			le16_to_cpu(bhdr->table_list[HARVEST_INFO].offset));
+	if (!discovery_bin) {
+		DRM_ERROR("ip discovery uninitialized\n");
+		return -EINVAL;
+	}
 
-	for (i = 0; i < 32; i++) {
-		if (le16_to_cpu(harvest_info->list[i].hw_id) == 0)
-			break;
+	/* num_vcn_inst is currently limited to AMDGPU_MAX_VCN_INSTANCES
+	 * which is smaller than VCN_INFO_TABLE_MAX_NUM_INSTANCES
+	 * but that may change in the future with new GPUs so keep this
+	 * check for defensive purposes.
+	 */
+	if (adev->vcn.num_vcn_inst > VCN_INFO_TABLE_MAX_NUM_INSTANCES) {
+		dev_err(adev->dev, "invalid vcn instances\n");
+		return -EINVAL;
+	}
 
-		switch (le16_to_cpu(harvest_info->list[i].hw_id)) {
-		case VCN_HWID:
-			vcn_harvest_count++;
-			if (harvest_info->list[i].number_instance == 0)
-				adev->vcn.harvest_config |= AMDGPU_VCN_HARVEST_VCN0;
-			else
-				adev->vcn.harvest_config |= AMDGPU_VCN_HARVEST_VCN1;
-			break;
-		case DMU_HWID:
-			adev->harvest_ip_mask |= AMD_HARVEST_IP_DMU_MASK;
-			break;
-		default:
-			break;
+	bhdr = (struct binary_header *)discovery_bin;
+	offset = le16_to_cpu(bhdr->table_list[VCN_INFO].offset);
+
+	if (!offset)
+		return 0;
+
+	vcn_info = (union vcn_info *)(discovery_bin + offset);
+
+	switch (le16_to_cpu(vcn_info->v1.header.version_major)) {
+	case 1:
+		/* num_vcn_inst is currently limited to AMDGPU_MAX_VCN_INSTANCES
+		 * so this won't overflow.
+		 */
+		for (v = 0; v < adev->vcn.num_vcn_inst; v++) {
+			adev->vcn.inst[v].vcn_codec_disable_mask =
+				le32_to_cpu(vcn_info->v1.instance_info[v].fuse_data.all_bits);
 		}
+		break;
+	default:
+		dev_err(adev->dev,
+			"Unhandled VCN info table %d.%d\n",
+			le16_to_cpu(vcn_info->v1.header.version_major),
+			le16_to_cpu(vcn_info->v1.header.version_minor));
+		return -EINVAL;
 	}
-	/* some IP discovery tables on Navy Flounder don't have this set correctly */
-	if ((adev->ip_versions[UVD_HWIP][1] == IP_VERSION(3, 0, 1)) &&
-	    (adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 3, 2)))
-		adev->vcn.harvest_config |= AMDGPU_VCN_HARVEST_VCN1;
-	if (vcn_harvest_count == adev->vcn.num_vcn_inst) {
-		adev->harvest_ip_mask |= AMD_HARVEST_IP_VCN_MASK;
-		adev->harvest_ip_mask |= AMD_HARVEST_IP_JPEG_MASK;
-	}
-	if ((adev->pdev->device == 0x731E &&
-	     (adev->pdev->revision == 0xC6 || adev->pdev->revision == 0xC7)) ||
-	    (adev->pdev->device == 0x7340 && adev->pdev->revision == 0xC9)  ||
-	    (adev->pdev->device == 0x7360 && adev->pdev->revision == 0xC7)) {
-		adev->harvest_ip_mask |= AMD_HARVEST_IP_VCN_MASK;
-		adev->harvest_ip_mask |= AMD_HARVEST_IP_JPEG_MASK;
+	return 0;
+}
+
+union nps_info {
+	struct nps_info_v1_0 v1;
+};
+
+static int amdgpu_discovery_refresh_nps_info(struct amdgpu_device *adev,
+					     union nps_info *nps_data)
+{
+	uint64_t vram_size, pos, offset;
+	struct nps_info_header *nhdr;
+	struct binary_header bhdr;
+	uint16_t checksum;
+
+	vram_size = (uint64_t)RREG32(mmRCC_CONFIG_MEMSIZE) << 20;
+	pos = vram_size - DISCOVERY_TMR_OFFSET;
+	amdgpu_device_vram_access(adev, pos, &bhdr, sizeof(bhdr), false);
+
+	offset = le16_to_cpu(bhdr.table_list[NPS_INFO].offset);
+	checksum = le16_to_cpu(bhdr.table_list[NPS_INFO].checksum);
+
+	amdgpu_device_vram_access(adev, (pos + offset), nps_data,
+				  sizeof(*nps_data), false);
+
+	nhdr = (struct nps_info_header *)(nps_data);
+	if (!amdgpu_discovery_verify_checksum((uint8_t *)nps_data,
+					      le32_to_cpu(nhdr->size_bytes),
+					      checksum)) {
+		dev_err(adev->dev, "nps data refresh, checksum mismatch\n");
+		return -EINVAL;
 	}
+
+	return 0;
 }
 
-int amdgpu_discovery_get_gfx_info(struct amdgpu_device *adev)
+int amdgpu_discovery_get_nps_info(struct amdgpu_device *adev,
+				  uint32_t *nps_type,
+				  struct amdgpu_gmc_memrange **ranges,
+				  int *range_cnt, bool refresh)
 {
+	uint8_t *discovery_bin = adev->discovery.bin;
+	struct amdgpu_gmc_memrange *mem_ranges;
 	struct binary_header *bhdr;
-	struct gc_info_v1_0 *gc_info;
+	union nps_info *nps_info;
+	union nps_info nps_data;
+	u16 offset;
+	int i, r;
 
-	if (!adev->mman.discovery_bin) {
-		DRM_ERROR("ip discovery uninitialized\n");
+	if (!nps_type || !range_cnt || !ranges)
 		return -EINVAL;
+
+	if (refresh) {
+		r = amdgpu_discovery_refresh_nps_info(adev, &nps_data);
+		if (r)
+			return r;
+		nps_info = &nps_data;
+	} else {
+		if (!discovery_bin) {
+			dev_err(adev->dev,
+				"fetch mem range failed, ip discovery uninitialized\n");
+			return -EINVAL;
+		}
+
+		bhdr = (struct binary_header *)discovery_bin;
+		offset = le16_to_cpu(bhdr->table_list[NPS_INFO].offset);
+
+		if (!offset)
+			return -ENOENT;
+
+		/* If verification fails, return as if NPS table doesn't exist */
+		if (amdgpu_discovery_verify_npsinfo(adev, bhdr))
+			return -ENOENT;
+
+		nps_info = (union nps_info *)(discovery_bin + offset);
 	}
 
-	bhdr = (struct binary_header *)adev->mman.discovery_bin;
-	gc_info = (struct gc_info_v1_0 *)(adev->mman.discovery_bin +
-			le16_to_cpu(bhdr->table_list[GC].offset));
-
-	adev->gfx.config.max_shader_engines = le32_to_cpu(gc_info->gc_num_se);
-	adev->gfx.config.max_cu_per_sh = 2 * (le32_to_cpu(gc_info->gc_num_wgp0_per_sa) +
-					      le32_to_cpu(gc_info->gc_num_wgp1_per_sa));
-	adev->gfx.config.max_sh_per_se = le32_to_cpu(gc_info->gc_num_sa_per_se);
-	adev->gfx.config.max_backends_per_se = le32_to_cpu(gc_info->gc_num_rb_per_se);
-	adev->gfx.config.max_texture_channel_caches = le32_to_cpu(gc_info->gc_num_gl2c);
-	adev->gfx.config.max_gprs = le32_to_cpu(gc_info->gc_num_gprs);
-	adev->gfx.config.max_gs_threads = le32_to_cpu(gc_info->gc_num_max_gs_thds);
-	adev->gfx.config.gs_vgt_table_depth = le32_to_cpu(gc_info->gc_gs_table_depth);
-	adev->gfx.config.gs_prim_buffer_depth = le32_to_cpu(gc_info->gc_gsprim_buff_depth);
-	adev->gfx.config.double_offchip_lds_buf = le32_to_cpu(gc_info->gc_double_offchip_lds_buffer);
-	adev->gfx.cu_info.wave_front_size = le32_to_cpu(gc_info->gc_wave_size);
-	adev->gfx.cu_info.max_waves_per_simd = le32_to_cpu(gc_info->gc_max_waves_per_simd);
-	adev->gfx.cu_info.max_scratch_slots_per_cu = le32_to_cpu(gc_info->gc_max_scratch_slots_per_cu);
-	adev->gfx.cu_info.lds_size = le32_to_cpu(gc_info->gc_lds_size);
-	adev->gfx.config.num_sc_per_sh = le32_to_cpu(gc_info->gc_num_sc_per_se) /
-					 le32_to_cpu(gc_info->gc_num_sa_per_se);
-	adev->gfx.config.num_packer_per_sc = le32_to_cpu(gc_info->gc_num_packer_per_sc);
+	switch (le16_to_cpu(nps_info->v1.header.version_major)) {
+	case 1:
+		mem_ranges = kvcalloc(nps_info->v1.count,
+				      sizeof(*mem_ranges),
+				      GFP_KERNEL);
+		if (!mem_ranges)
+			return -ENOMEM;
+		*nps_type = nps_info->v1.nps_type;
+		*range_cnt = nps_info->v1.count;
+		for (i = 0; i < *range_cnt; i++) {
+			mem_ranges[i].base_address =
+				nps_info->v1.instance_info[i].base_address;
+			mem_ranges[i].limit_address =
+				nps_info->v1.instance_info[i].limit_address;
+			mem_ranges[i].nid_mask = -1;
+			mem_ranges[i].flags = 0;
+		}
+		*ranges = mem_ranges;
+		break;
+	default:
+		dev_err(adev->dev, "Unhandled NPS info table %d.%d\n",
+			le16_to_cpu(nps_info->v1.header.version_major),
+			le16_to_cpu(nps_info->v1.header.version_minor));
+		return -EINVAL;
+	}
 
 	return 0;
 }
@@ -565,7 +1942,7 @@ int amdgpu_discovery_get_gfx_info(struct amdgpu_device *adev)
 static int amdgpu_discovery_set_common_ip_blocks(struct amdgpu_device *adev)
 {
 	/* what IP to use for this? */
-	switch (adev->ip_versions[GC_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 	case IP_VERSION(9, 0, 1):
 	case IP_VERSION(9, 1, 0):
 	case IP_VERSION(9, 2, 1):
@@ -574,24 +1951,45 @@ static int amdgpu_discovery_set_common_ip_blocks(struct amdgpu_device *adev)
 	case IP_VERSION(9, 4, 0):
 	case IP_VERSION(9, 4, 1):
 	case IP_VERSION(9, 4, 2):
+	case IP_VERSION(9, 4, 3):
+	case IP_VERSION(9, 4, 4):
+	case IP_VERSION(9, 5, 0):
 		amdgpu_device_ip_block_add(adev, &vega10_common_ip_block);
 		break;
 	case IP_VERSION(10, 1, 10):
 	case IP_VERSION(10, 1, 1):
 	case IP_VERSION(10, 1, 2):
 	case IP_VERSION(10, 1, 3):
+	case IP_VERSION(10, 1, 4):
 	case IP_VERSION(10, 3, 0):
 	case IP_VERSION(10, 3, 1):
 	case IP_VERSION(10, 3, 2):
 	case IP_VERSION(10, 3, 3):
 	case IP_VERSION(10, 3, 4):
 	case IP_VERSION(10, 3, 5):
+	case IP_VERSION(10, 3, 6):
+	case IP_VERSION(10, 3, 7):
 		amdgpu_device_ip_block_add(adev, &nv_common_ip_block);
 		break;
+	case IP_VERSION(11, 0, 0):
+	case IP_VERSION(11, 0, 1):
+	case IP_VERSION(11, 0, 2):
+	case IP_VERSION(11, 0, 3):
+	case IP_VERSION(11, 0, 4):
+	case IP_VERSION(11, 5, 0):
+	case IP_VERSION(11, 5, 1):
+	case IP_VERSION(11, 5, 2):
+	case IP_VERSION(11, 5, 3):
+		amdgpu_device_ip_block_add(adev, &soc21_common_ip_block);
+		break;
+	case IP_VERSION(12, 0, 0):
+	case IP_VERSION(12, 0, 1):
+		amdgpu_device_ip_block_add(adev, &soc24_common_ip_block);
+		break;
 	default:
 		dev_err(adev->dev,
 			"Failed to add common ip block(GC_HWIP:0x%x)\n",
-			adev->ip_versions[GC_HWIP][0]);
+			amdgpu_ip_version(adev, GC_HWIP, 0));
 		return -EINVAL;
 	}
 	return 0;
@@ -600,7 +1998,7 @@ static int amdgpu_discovery_set_common_ip_blocks(struct amdgpu_device *adev)
 static int amdgpu_discovery_set_gmc_ip_blocks(struct amdgpu_device *adev)
 {
 	/* use GC or MMHUB IP version */
-	switch (adev->ip_versions[GC_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 	case IP_VERSION(9, 0, 1):
 	case IP_VERSION(9, 1, 0):
 	case IP_VERSION(9, 2, 1):
@@ -609,24 +2007,44 @@ static int amdgpu_discovery_set_gmc_ip_blocks(struct amdgpu_device *adev)
 	case IP_VERSION(9, 4, 0):
 	case IP_VERSION(9, 4, 1):
 	case IP_VERSION(9, 4, 2):
+	case IP_VERSION(9, 4, 3):
+	case IP_VERSION(9, 4, 4):
+	case IP_VERSION(9, 5, 0):
 		amdgpu_device_ip_block_add(adev, &gmc_v9_0_ip_block);
 		break;
 	case IP_VERSION(10, 1, 10):
 	case IP_VERSION(10, 1, 1):
 	case IP_VERSION(10, 1, 2):
 	case IP_VERSION(10, 1, 3):
+	case IP_VERSION(10, 1, 4):
 	case IP_VERSION(10, 3, 0):
 	case IP_VERSION(10, 3, 1):
 	case IP_VERSION(10, 3, 2):
 	case IP_VERSION(10, 3, 3):
 	case IP_VERSION(10, 3, 4):
 	case IP_VERSION(10, 3, 5):
+	case IP_VERSION(10, 3, 6):
+	case IP_VERSION(10, 3, 7):
 		amdgpu_device_ip_block_add(adev, &gmc_v10_0_ip_block);
 		break;
+	case IP_VERSION(11, 0, 0):
+	case IP_VERSION(11, 0, 1):
+	case IP_VERSION(11, 0, 2):
+	case IP_VERSION(11, 0, 3):
+	case IP_VERSION(11, 0, 4):
+	case IP_VERSION(11, 5, 0):
+	case IP_VERSION(11, 5, 1):
+	case IP_VERSION(11, 5, 2):
+	case IP_VERSION(11, 5, 3):
+		amdgpu_device_ip_block_add(adev, &gmc_v11_0_ip_block);
+		break;
+	case IP_VERSION(12, 0, 0):
+	case IP_VERSION(12, 0, 1):
+		amdgpu_device_ip_block_add(adev, &gmc_v12_0_ip_block);
+		break;
 	default:
-		dev_err(adev->dev,
-			"Failed to add gmc ip block(GC_HWIP:0x%x)\n",
-			adev->ip_versions[GC_HWIP][0]);
+		dev_err(adev->dev, "Failed to add gmc ip block(GC_HWIP:0x%x)\n",
+			amdgpu_ip_version(adev, GC_HWIP, 0));
 		return -EINVAL;
 	}
 	return 0;
@@ -634,7 +2052,7 @@ static int amdgpu_discovery_set_gmc_ip_blocks(struct amdgpu_device *adev)
 
 static int amdgpu_discovery_set_ih_ip_blocks(struct amdgpu_device *adev)
 {
-	switch (adev->ip_versions[OSSSYS_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, OSSSYS_HWIP, 0)) {
 	case IP_VERSION(4, 0, 0):
 	case IP_VERSION(4, 0, 1):
 	case IP_VERSION(4, 1, 0):
@@ -645,6 +2063,8 @@ static int amdgpu_discovery_set_ih_ip_blocks(struct amdgpu_device *adev)
 	case IP_VERSION(4, 2, 0):
 	case IP_VERSION(4, 2, 1):
 	case IP_VERSION(4, 4, 0):
+	case IP_VERSION(4, 4, 2):
+	case IP_VERSION(4, 4, 5):
 		amdgpu_device_ip_block_add(adev, &vega20_ih_ip_block);
 		break;
 	case IP_VERSION(5, 0, 0):
@@ -655,10 +2075,21 @@ static int amdgpu_discovery_set_ih_ip_blocks(struct amdgpu_device *adev)
 	case IP_VERSION(5, 2, 1):
 		amdgpu_device_ip_block_add(adev, &navi10_ih_ip_block);
 		break;
+	case IP_VERSION(6, 0, 0):
+	case IP_VERSION(6, 0, 1):
+	case IP_VERSION(6, 0, 2):
+		amdgpu_device_ip_block_add(adev, &ih_v6_0_ip_block);
+		break;
+	case IP_VERSION(6, 1, 0):
+		amdgpu_device_ip_block_add(adev, &ih_v6_1_ip_block);
+		break;
+	case IP_VERSION(7, 0, 0):
+		amdgpu_device_ip_block_add(adev, &ih_v7_0_ip_block);
+		break;
 	default:
 		dev_err(adev->dev,
 			"Failed to add ih ip block(OSSSYS_HWIP:0x%x)\n",
-			adev->ip_versions[OSSSYS_HWIP][0]);
+			amdgpu_ip_version(adev, OSSSYS_HWIP, 0));
 		return -EINVAL;
 	}
 	return 0;
@@ -666,7 +2097,7 @@ static int amdgpu_discovery_set_ih_ip_blocks(struct amdgpu_device *adev)
 
 static int amdgpu_discovery_set_psp_ip_blocks(struct amdgpu_device *adev)
 {
-	switch (adev->ip_versions[MP0_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, MP0_HWIP, 0)) {
 	case IP_VERSION(9, 0, 0):
 		amdgpu_device_ip_block_add(adev, &psp_v3_1_ip_block);
 		break;
@@ -684,6 +2115,7 @@ static int amdgpu_discovery_set_psp_ip_blocks(struct amdgpu_device *adev)
 	case IP_VERSION(11, 0, 12):
 	case IP_VERSION(11, 0, 13):
 	case IP_VERSION(11, 5, 0):
+	case IP_VERSION(11, 5, 2):
 		amdgpu_device_ip_block_add(adev, &psp_v11_0_ip_block);
 		break;
 	case IP_VERSION(11, 0, 8):
@@ -693,15 +2125,35 @@ static int amdgpu_discovery_set_psp_ip_blocks(struct amdgpu_device *adev)
 	case IP_VERSION(12, 0, 1):
 		amdgpu_device_ip_block_add(adev, &psp_v12_0_ip_block);
 		break;
+	case IP_VERSION(13, 0, 0):
 	case IP_VERSION(13, 0, 1):
 	case IP_VERSION(13, 0, 2):
 	case IP_VERSION(13, 0, 3):
+	case IP_VERSION(13, 0, 5):
+	case IP_VERSION(13, 0, 6):
+	case IP_VERSION(13, 0, 7):
+	case IP_VERSION(13, 0, 8):
+	case IP_VERSION(13, 0, 10):
+	case IP_VERSION(13, 0, 11):
+	case IP_VERSION(13, 0, 12):
+	case IP_VERSION(13, 0, 14):
+	case IP_VERSION(14, 0, 0):
+	case IP_VERSION(14, 0, 1):
+	case IP_VERSION(14, 0, 4):
 		amdgpu_device_ip_block_add(adev, &psp_v13_0_ip_block);
 		break;
+	case IP_VERSION(13, 0, 4):
+		amdgpu_device_ip_block_add(adev, &psp_v13_0_4_ip_block);
+		break;
+	case IP_VERSION(14, 0, 2):
+	case IP_VERSION(14, 0, 3):
+	case IP_VERSION(14, 0, 5):
+		amdgpu_device_ip_block_add(adev, &psp_v14_0_ip_block);
+		break;
 	default:
 		dev_err(adev->dev,
 			"Failed to add psp ip block(MP0_HWIP:0x%x)\n",
-			adev->ip_versions[MP0_HWIP][0]);
+			amdgpu_ip_version(adev, MP0_HWIP, 0));
 		return -EINVAL;
 	}
 	return 0;
@@ -709,7 +2161,7 @@ static int amdgpu_discovery_set_psp_ip_blocks(struct amdgpu_device *adev)
 
 static int amdgpu_discovery_set_smu_ip_blocks(struct amdgpu_device *adev)
 {
-	switch (adev->ip_versions[MP1_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, MP1_HWIP, 0)) {
 	case IP_VERSION(9, 0, 0):
 	case IP_VERSION(10, 0, 0):
 	case IP_VERSION(10, 0, 1):
@@ -723,38 +2175,74 @@ static int amdgpu_discovery_set_smu_ip_blocks(struct amdgpu_device *adev)
 	case IP_VERSION(11, 0, 5):
 	case IP_VERSION(11, 0, 9):
 	case IP_VERSION(11, 0, 7):
-	case IP_VERSION(11, 0, 8):
 	case IP_VERSION(11, 0, 11):
 	case IP_VERSION(11, 0, 12):
 	case IP_VERSION(11, 0, 13):
 	case IP_VERSION(11, 5, 0):
+	case IP_VERSION(11, 5, 2):
 		amdgpu_device_ip_block_add(adev, &smu_v11_0_ip_block);
 		break;
+	case IP_VERSION(11, 0, 8):
+		if (adev->apu_flags & AMD_APU_IS_CYAN_SKILLFISH2)
+			amdgpu_device_ip_block_add(adev, &smu_v11_0_ip_block);
+		break;
 	case IP_VERSION(12, 0, 0):
 	case IP_VERSION(12, 0, 1):
 		amdgpu_device_ip_block_add(adev, &smu_v12_0_ip_block);
 		break;
+	case IP_VERSION(13, 0, 0):
 	case IP_VERSION(13, 0, 1):
 	case IP_VERSION(13, 0, 2):
 	case IP_VERSION(13, 0, 3):
+	case IP_VERSION(13, 0, 4):
+	case IP_VERSION(13, 0, 5):
+	case IP_VERSION(13, 0, 6):
+	case IP_VERSION(13, 0, 7):
+	case IP_VERSION(13, 0, 8):
+	case IP_VERSION(13, 0, 10):
+	case IP_VERSION(13, 0, 11):
+	case IP_VERSION(13, 0, 14):
+	case IP_VERSION(13, 0, 12):
 		amdgpu_device_ip_block_add(adev, &smu_v13_0_ip_block);
 		break;
+	case IP_VERSION(14, 0, 0):
+	case IP_VERSION(14, 0, 1):
+	case IP_VERSION(14, 0, 2):
+	case IP_VERSION(14, 0, 3):
+	case IP_VERSION(14, 0, 4):
+	case IP_VERSION(14, 0, 5):
+		amdgpu_device_ip_block_add(adev, &smu_v14_0_ip_block);
+		break;
 	default:
 		dev_err(adev->dev,
 			"Failed to add smu ip block(MP1_HWIP:0x%x)\n",
-			adev->ip_versions[MP1_HWIP][0]);
+			amdgpu_ip_version(adev, MP1_HWIP, 0));
 		return -EINVAL;
 	}
 	return 0;
 }
 
+#if defined(CONFIG_DRM_AMD_DC)
+static void amdgpu_discovery_set_sriov_display(struct amdgpu_device *adev)
+{
+	amdgpu_device_set_sriov_virtual_display(adev);
+	amdgpu_device_ip_block_add(adev, &amdgpu_vkms_ip_block);
+}
+#endif
+
 static int amdgpu_discovery_set_display_ip_blocks(struct amdgpu_device *adev)
 {
-	if (adev->enable_virtual_display || amdgpu_sriov_vf(adev)) {
+	if (adev->enable_virtual_display) {
 		amdgpu_device_ip_block_add(adev, &amdgpu_vkms_ip_block);
+		return 0;
+	}
+
+	if (!amdgpu_device_has_dc_support(adev))
+		return 0;
+
 #if defined(CONFIG_DRM_AMD_DC)
-	} else if (adev->ip_versions[DCE_HWIP][0]) {
-		switch (adev->ip_versions[DCE_HWIP][0]) {
+	if (amdgpu_ip_version(adev, DCE_HWIP, 0)) {
+		switch (amdgpu_ip_version(adev, DCE_HWIP, 0)) {
 		case IP_VERSION(1, 0, 0):
 		case IP_VERSION(1, 0, 1):
 		case IP_VERSION(2, 0, 2):
@@ -767,35 +2255,54 @@ static int amdgpu_discovery_set_display_ip_blocks(struct amdgpu_device *adev)
 		case IP_VERSION(3, 0, 1):
 		case IP_VERSION(3, 1, 2):
 		case IP_VERSION(3, 1, 3):
-			amdgpu_device_ip_block_add(adev, &dm_ip_block);
+		case IP_VERSION(3, 1, 4):
+		case IP_VERSION(3, 1, 5):
+		case IP_VERSION(3, 1, 6):
+		case IP_VERSION(3, 2, 0):
+		case IP_VERSION(3, 2, 1):
+		case IP_VERSION(3, 5, 0):
+		case IP_VERSION(3, 5, 1):
+		case IP_VERSION(3, 6, 0):
+		case IP_VERSION(4, 1, 0):
+			/* TODO: Fix IP version. DC code expects version 4.0.1 */
+			if (adev->ip_versions[DCE_HWIP][0] == IP_VERSION(4, 1, 0))
+				adev->ip_versions[DCE_HWIP][0] = IP_VERSION(4, 0, 1);
+
+			if (amdgpu_sriov_vf(adev))
+				amdgpu_discovery_set_sriov_display(adev);
+			else
+				amdgpu_device_ip_block_add(adev, &dm_ip_block);
 			break;
 		default:
 			dev_err(adev->dev,
 				"Failed to add dm ip block(DCE_HWIP:0x%x)\n",
-				adev->ip_versions[DCE_HWIP][0]);
+				amdgpu_ip_version(adev, DCE_HWIP, 0));
 			return -EINVAL;
 		}
-	} else if (adev->ip_versions[DCI_HWIP][0]) {
-		switch (adev->ip_versions[DCI_HWIP][0]) {
+	} else if (amdgpu_ip_version(adev, DCI_HWIP, 0)) {
+		switch (amdgpu_ip_version(adev, DCI_HWIP, 0)) {
 		case IP_VERSION(12, 0, 0):
 		case IP_VERSION(12, 0, 1):
 		case IP_VERSION(12, 1, 0):
-			amdgpu_device_ip_block_add(adev, &dm_ip_block);
+			if (amdgpu_sriov_vf(adev))
+				amdgpu_discovery_set_sriov_display(adev);
+			else
+				amdgpu_device_ip_block_add(adev, &dm_ip_block);
 			break;
 		default:
 			dev_err(adev->dev,
 				"Failed to add dm ip block(DCI_HWIP:0x%x)\n",
-				adev->ip_versions[DCI_HWIP][0]);
+				amdgpu_ip_version(adev, DCI_HWIP, 0));
 			return -EINVAL;
 		}
-#endif
 	}
+#endif
 	return 0;
 }
 
 static int amdgpu_discovery_set_gc_ip_blocks(struct amdgpu_device *adev)
 {
-	switch (adev->ip_versions[GC_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 	case IP_VERSION(9, 0, 1):
 	case IP_VERSION(9, 1, 0):
 	case IP_VERSION(9, 2, 1):
@@ -806,22 +2313,44 @@ static int amdgpu_discovery_set_gc_ip_blocks(struct amdgpu_device *adev)
 	case IP_VERSION(9, 4, 2):
 		amdgpu_device_ip_block_add(adev, &gfx_v9_0_ip_block);
 		break;
+	case IP_VERSION(9, 4, 3):
+	case IP_VERSION(9, 4, 4):
+	case IP_VERSION(9, 5, 0):
+		amdgpu_device_ip_block_add(adev, &gfx_v9_4_3_ip_block);
+		break;
 	case IP_VERSION(10, 1, 10):
 	case IP_VERSION(10, 1, 2):
 	case IP_VERSION(10, 1, 1):
 	case IP_VERSION(10, 1, 3):
+	case IP_VERSION(10, 1, 4):
 	case IP_VERSION(10, 3, 0):
 	case IP_VERSION(10, 3, 2):
 	case IP_VERSION(10, 3, 1):
 	case IP_VERSION(10, 3, 4):
 	case IP_VERSION(10, 3, 5):
+	case IP_VERSION(10, 3, 6):
 	case IP_VERSION(10, 3, 3):
+	case IP_VERSION(10, 3, 7):
 		amdgpu_device_ip_block_add(adev, &gfx_v10_0_ip_block);
 		break;
+	case IP_VERSION(11, 0, 0):
+	case IP_VERSION(11, 0, 1):
+	case IP_VERSION(11, 0, 2):
+	case IP_VERSION(11, 0, 3):
+	case IP_VERSION(11, 0, 4):
+	case IP_VERSION(11, 5, 0):
+	case IP_VERSION(11, 5, 1):
+	case IP_VERSION(11, 5, 2):
+	case IP_VERSION(11, 5, 3):
+		amdgpu_device_ip_block_add(adev, &gfx_v11_0_ip_block);
+		break;
+	case IP_VERSION(12, 0, 0):
+	case IP_VERSION(12, 0, 1):
+		amdgpu_device_ip_block_add(adev, &gfx_v12_0_ip_block);
+		break;
 	default:
-		dev_err(adev->dev,
-			"Failed to add gfx ip block(GC_HWIP:0x%x)\n",
-			adev->ip_versions[GC_HWIP][0]);
+		dev_err(adev->dev, "Failed to add gfx ip block(GC_HWIP:0x%x)\n",
+			amdgpu_ip_version(adev, GC_HWIP, 0));
 		return -EINVAL;
 	}
 	return 0;
@@ -829,7 +2358,7 @@ static int amdgpu_discovery_set_gc_ip_blocks(struct amdgpu_device *adev)
 
 static int amdgpu_discovery_set_sdma_ip_blocks(struct amdgpu_device *adev)
 {
-	switch (adev->ip_versions[SDMA0_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, SDMA0_HWIP, 0)) {
 	case IP_VERSION(4, 0, 0):
 	case IP_VERSION(4, 0, 1):
 	case IP_VERSION(4, 1, 0):
@@ -840,6 +2369,11 @@ static int amdgpu_discovery_set_sdma_ip_blocks(struct amdgpu_device *adev)
 	case IP_VERSION(4, 4, 0):
 		amdgpu_device_ip_block_add(adev, &sdma_v4_0_ip_block);
 		break;
+	case IP_VERSION(4, 4, 2):
+	case IP_VERSION(4, 4, 5):
+	case IP_VERSION(4, 4, 4):
+		amdgpu_device_ip_block_add(adev, &sdma_v4_4_2_ip_block);
+		break;
 	case IP_VERSION(5, 0, 0):
 	case IP_VERSION(5, 0, 1):
 	case IP_VERSION(5, 0, 2):
@@ -850,23 +2384,54 @@ static int amdgpu_discovery_set_sdma_ip_blocks(struct amdgpu_device *adev)
 	case IP_VERSION(5, 2, 2):
 	case IP_VERSION(5, 2, 4):
 	case IP_VERSION(5, 2, 5):
+	case IP_VERSION(5, 2, 6):
 	case IP_VERSION(5, 2, 3):
 	case IP_VERSION(5, 2, 1):
+	case IP_VERSION(5, 2, 7):
 		amdgpu_device_ip_block_add(adev, &sdma_v5_2_ip_block);
 		break;
+	case IP_VERSION(6, 0, 0):
+	case IP_VERSION(6, 0, 1):
+	case IP_VERSION(6, 0, 2):
+	case IP_VERSION(6, 0, 3):
+	case IP_VERSION(6, 1, 0):
+	case IP_VERSION(6, 1, 1):
+	case IP_VERSION(6, 1, 2):
+	case IP_VERSION(6, 1, 3):
+		amdgpu_device_ip_block_add(adev, &sdma_v6_0_ip_block);
+		break;
+	case IP_VERSION(7, 0, 0):
+	case IP_VERSION(7, 0, 1):
+		amdgpu_device_ip_block_add(adev, &sdma_v7_0_ip_block);
+		break;
 	default:
 		dev_err(adev->dev,
 			"Failed to add sdma ip block(SDMA0_HWIP:0x%x)\n",
-			adev->ip_versions[SDMA0_HWIP][0]);
+			amdgpu_ip_version(adev, SDMA0_HWIP, 0));
 		return -EINVAL;
 	}
+
+	return 0;
+}
+
+static int amdgpu_discovery_set_ras_ip_blocks(struct amdgpu_device *adev)
+{
+	switch (amdgpu_ip_version(adev, MP0_HWIP, 0)) {
+	case IP_VERSION(13, 0, 6):
+	case IP_VERSION(13, 0, 12):
+	case IP_VERSION(13, 0, 14):
+		amdgpu_device_ip_block_add(adev, &ras_v1_0_ip_block);
+		break;
+	default:
+		break;
+	}
 	return 0;
 }
 
 static int amdgpu_discovery_set_mm_ip_blocks(struct amdgpu_device *adev)
 {
-	if (adev->ip_versions[VCE_HWIP][0]) {
-		switch (adev->ip_versions[UVD_HWIP][0]) {
+	if (amdgpu_ip_version(adev, VCE_HWIP, 0)) {
+		switch (amdgpu_ip_version(adev, UVD_HWIP, 0)) {
 		case IP_VERSION(7, 0, 0):
 		case IP_VERSION(7, 2, 0):
 			/* UVD is not supported on vega20 SR-IOV */
@@ -876,10 +2441,10 @@ static int amdgpu_discovery_set_mm_ip_blocks(struct amdgpu_device *adev)
 		default:
 			dev_err(adev->dev,
 				"Failed to add uvd v7 ip block(UVD_HWIP:0x%x)\n",
-				adev->ip_versions[UVD_HWIP][0]);
+				amdgpu_ip_version(adev, UVD_HWIP, 0));
 			return -EINVAL;
 		}
-		switch (adev->ip_versions[VCE_HWIP][0]) {
+		switch (amdgpu_ip_version(adev, VCE_HWIP, 0)) {
 		case IP_VERSION(4, 0, 0):
 		case IP_VERSION(4, 1, 0):
 			/* VCE is not supported on vega20 SR-IOV */
@@ -889,11 +2454,11 @@ static int amdgpu_discovery_set_mm_ip_blocks(struct amdgpu_device *adev)
 		default:
 			dev_err(adev->dev,
 				"Failed to add VCE v4 ip block(VCE_HWIP:0x%x)\n",
-				adev->ip_versions[VCE_HWIP][0]);
+				amdgpu_ip_version(adev, VCE_HWIP, 0));
 			return -EINVAL;
 		}
 	} else {
-		switch (adev->ip_versions[UVD_HWIP][0]) {
+		switch (amdgpu_ip_version(adev, UVD_HWIP, 0)) {
 		case IP_VERSION(1, 0, 0):
 		case IP_VERSION(1, 0, 1):
 			amdgpu_device_ip_block_add(adev, &vcn_v1_0_ip_block);
@@ -917,10 +2482,9 @@ static int amdgpu_discovery_set_mm_ip_blocks(struct amdgpu_device *adev)
 			break;
 		case IP_VERSION(3, 0, 0):
 		case IP_VERSION(3, 0, 16):
-		case IP_VERSION(3, 0, 64):
 		case IP_VERSION(3, 1, 1):
+		case IP_VERSION(3, 1, 2):
 		case IP_VERSION(3, 0, 2):
-		case IP_VERSION(3, 0, 192):
 			amdgpu_device_ip_block_add(adev, &vcn_v3_0_ip_block);
 			if (!amdgpu_sriov_vf(adev))
 				amdgpu_device_ip_block_add(adev, &jpeg_v3_0_ip_block);
@@ -928,10 +2492,33 @@ static int amdgpu_discovery_set_mm_ip_blocks(struct amdgpu_device *adev)
 		case IP_VERSION(3, 0, 33):
 			amdgpu_device_ip_block_add(adev, &vcn_v3_0_ip_block);
 			break;
+		case IP_VERSION(4, 0, 0):
+		case IP_VERSION(4, 0, 2):
+		case IP_VERSION(4, 0, 4):
+			amdgpu_device_ip_block_add(adev, &vcn_v4_0_ip_block);
+			amdgpu_device_ip_block_add(adev, &jpeg_v4_0_ip_block);
+			break;
+		case IP_VERSION(4, 0, 3):
+			amdgpu_device_ip_block_add(adev, &vcn_v4_0_3_ip_block);
+			amdgpu_device_ip_block_add(adev, &jpeg_v4_0_3_ip_block);
+			break;
+		case IP_VERSION(4, 0, 5):
+		case IP_VERSION(4, 0, 6):
+			amdgpu_device_ip_block_add(adev, &vcn_v4_0_5_ip_block);
+			amdgpu_device_ip_block_add(adev, &jpeg_v4_0_5_ip_block);
+			break;
+		case IP_VERSION(5, 0, 0):
+			amdgpu_device_ip_block_add(adev, &vcn_v5_0_0_ip_block);
+			amdgpu_device_ip_block_add(adev, &jpeg_v5_0_0_ip_block);
+			break;
+		case IP_VERSION(5, 0, 1):
+			amdgpu_device_ip_block_add(adev, &vcn_v5_0_1_ip_block);
+			amdgpu_device_ip_block_add(adev, &jpeg_v5_0_1_ip_block);
+			break;
 		default:
 			dev_err(adev->dev,
 				"Failed to add vcn/jpeg ip block(UVD_HWIP:0x%x)\n",
-				adev->ip_versions[UVD_HWIP][0]);
+				amdgpu_ip_version(adev, UVD_HWIP, 0));
 			return -EINVAL;
 		}
 	}
@@ -940,33 +2527,113 @@ static int amdgpu_discovery_set_mm_ip_blocks(struct amdgpu_device *adev)
 
 static int amdgpu_discovery_set_mes_ip_blocks(struct amdgpu_device *adev)
 {
-	switch (adev->ip_versions[GC_HWIP][0]) {
-	case IP_VERSION(10, 1, 10):
-	case IP_VERSION(10, 1, 1):
-	case IP_VERSION(10, 1, 2):
-	case IP_VERSION(10, 1, 3):
-	case IP_VERSION(10, 3, 0):
-	case IP_VERSION(10, 3, 1):
-	case IP_VERSION(10, 3, 2):
-	case IP_VERSION(10, 3, 3):
-	case IP_VERSION(10, 3, 4):
-	case IP_VERSION(10, 3, 5):
-		amdgpu_device_ip_block_add(adev, &mes_v10_1_ip_block);
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	case IP_VERSION(11, 0, 0):
+	case IP_VERSION(11, 0, 1):
+	case IP_VERSION(11, 0, 2):
+	case IP_VERSION(11, 0, 3):
+	case IP_VERSION(11, 0, 4):
+	case IP_VERSION(11, 5, 0):
+	case IP_VERSION(11, 5, 1):
+	case IP_VERSION(11, 5, 2):
+	case IP_VERSION(11, 5, 3):
+		amdgpu_device_ip_block_add(adev, &mes_v11_0_ip_block);
+		adev->enable_mes = true;
+		adev->enable_mes_kiq = true;
+		break;
+	case IP_VERSION(12, 0, 0):
+	case IP_VERSION(12, 0, 1):
+		amdgpu_device_ip_block_add(adev, &mes_v12_0_ip_block);
+		adev->enable_mes = true;
+		adev->enable_mes_kiq = true;
+		if (amdgpu_uni_mes)
+			adev->enable_uni_mes = true;
 		break;
 	default:
-		break;;
+		break;
 	}
 	return 0;
 }
 
+static void amdgpu_discovery_init_soc_config(struct amdgpu_device *adev)
+{
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	case IP_VERSION(9, 4, 3):
+	case IP_VERSION(9, 4, 4):
+	case IP_VERSION(9, 5, 0):
+		aqua_vanjaram_init_soc_config(adev);
+		break;
+	default:
+		break;
+	}
+}
+
+static int amdgpu_discovery_set_vpe_ip_blocks(struct amdgpu_device *adev)
+{
+	switch (amdgpu_ip_version(adev, VPE_HWIP, 0)) {
+	case IP_VERSION(6, 1, 0):
+	case IP_VERSION(6, 1, 1):
+	case IP_VERSION(6, 1, 3):
+		amdgpu_device_ip_block_add(adev, &vpe_v6_1_ip_block);
+		break;
+	default:
+		break;
+	}
+
+	return 0;
+}
+
+static int amdgpu_discovery_set_umsch_mm_ip_blocks(struct amdgpu_device *adev)
+{
+	switch (amdgpu_ip_version(adev, VCN_HWIP, 0)) {
+	case IP_VERSION(4, 0, 5):
+	case IP_VERSION(4, 0, 6):
+		if (amdgpu_umsch_mm & 0x1) {
+			amdgpu_device_ip_block_add(adev, &umsch_mm_v4_0_ip_block);
+			adev->enable_umsch_mm = true;
+		}
+		break;
+	default:
+		break;
+	}
+
+	return 0;
+}
+
+static int amdgpu_discovery_set_isp_ip_blocks(struct amdgpu_device *adev)
+{
+#if defined(CONFIG_DRM_AMD_ISP)
+	switch (amdgpu_ip_version(adev, ISP_HWIP, 0)) {
+	case IP_VERSION(4, 1, 0):
+		amdgpu_device_ip_block_add(adev, &isp_v4_1_0_ip_block);
+		break;
+	case IP_VERSION(4, 1, 1):
+		amdgpu_device_ip_block_add(adev, &isp_v4_1_1_ip_block);
+		break;
+	default:
+		break;
+	}
+#endif
+
+	return 0;
+}
+
 int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev)
 {
 	int r;
 
 	switch (adev->asic_type) {
 	case CHIP_VEGA10:
+		/* This is not fatal.  We only need the discovery
+		 * binary for sysfs.  We don't need it for a
+		 * functional system.
+		 */
+		amdgpu_discovery_init(adev);
 		vega10_reg_base_init(adev);
 		adev->sdma.num_instances = 2;
+		adev->sdma.sdma_mask = 3;
+		adev->gmc.num_umc = 4;
+		adev->gfx.xcc_mask = 1;
 		adev->ip_versions[MMHUB_HWIP][0] = IP_VERSION(9, 0, 0);
 		adev->ip_versions[ATHUB_HWIP][0] = IP_VERSION(9, 0, 0);
 		adev->ip_versions[OSSSYS_HWIP][0] = IP_VERSION(4, 0, 0);
@@ -986,8 +2653,16 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev)
 		adev->ip_versions[DCI_HWIP][0] = IP_VERSION(12, 0, 0);
 		break;
 	case CHIP_VEGA12:
+		/* This is not fatal.  We only need the discovery
+		 * binary for sysfs.  We don't need it for a
+		 * functional system.
+		 */
+		amdgpu_discovery_init(adev);
 		vega10_reg_base_init(adev);
 		adev->sdma.num_instances = 2;
+		adev->sdma.sdma_mask = 3;
+		adev->gmc.num_umc = 4;
+		adev->gfx.xcc_mask = 1;
 		adev->ip_versions[MMHUB_HWIP][0] = IP_VERSION(9, 3, 0);
 		adev->ip_versions[ATHUB_HWIP][0] = IP_VERSION(9, 3, 0);
 		adev->ip_versions[OSSSYS_HWIP][0] = IP_VERSION(4, 0, 1);
@@ -1007,9 +2682,17 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev)
 		adev->ip_versions[DCI_HWIP][0] = IP_VERSION(12, 0, 1);
 		break;
 	case CHIP_RAVEN:
+		/* This is not fatal.  We only need the discovery
+		 * binary for sysfs.  We don't need it for a
+		 * functional system.
+		 */
+		amdgpu_discovery_init(adev);
 		vega10_reg_base_init(adev);
 		adev->sdma.num_instances = 1;
+		adev->sdma.sdma_mask = 1;
 		adev->vcn.num_vcn_inst = 1;
+		adev->gmc.num_umc = 2;
+		adev->gfx.xcc_mask = 1;
 		if (adev->apu_flags & AMD_APU_IS_RAVEN2) {
 			adev->ip_versions[MMHUB_HWIP][0] = IP_VERSION(9, 2, 0);
 			adev->ip_versions[ATHUB_HWIP][0] = IP_VERSION(9, 2, 0);
@@ -1026,6 +2709,7 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev)
 			adev->ip_versions[GC_HWIP][0] = IP_VERSION(9, 2, 2);
 			adev->ip_versions[UVD_HWIP][0] = IP_VERSION(1, 0, 1);
 			adev->ip_versions[DCE_HWIP][0] = IP_VERSION(1, 0, 1);
+			adev->ip_versions[ISP_HWIP][0] = IP_VERSION(2, 0, 0);
 		} else {
 			adev->ip_versions[MMHUB_HWIP][0] = IP_VERSION(9, 1, 0);
 			adev->ip_versions[ATHUB_HWIP][0] = IP_VERSION(9, 1, 0);
@@ -1042,11 +2726,20 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev)
 			adev->ip_versions[GC_HWIP][0] = IP_VERSION(9, 1, 0);
 			adev->ip_versions[UVD_HWIP][0] = IP_VERSION(1, 0, 0);
 			adev->ip_versions[DCE_HWIP][0] = IP_VERSION(1, 0, 0);
+			adev->ip_versions[ISP_HWIP][0] = IP_VERSION(2, 0, 0);
 		}
 		break;
 	case CHIP_VEGA20:
+		/* This is not fatal.  We only need the discovery
+		 * binary for sysfs.  We don't need it for a
+		 * functional system.
+		 */
+		amdgpu_discovery_init(adev);
 		vega20_reg_base_init(adev);
 		adev->sdma.num_instances = 2;
+		adev->sdma.sdma_mask = 3;
+		adev->gmc.num_umc = 8;
+		adev->gfx.xcc_mask = 1;
 		adev->ip_versions[MMHUB_HWIP][0] = IP_VERSION(9, 4, 0);
 		adev->ip_versions[ATHUB_HWIP][0] = IP_VERSION(9, 4, 0);
 		adev->ip_versions[OSSSYS_HWIP][0] = IP_VERSION(4, 2, 0);
@@ -1067,9 +2760,17 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev)
 		adev->ip_versions[DCI_HWIP][0] = IP_VERSION(12, 1, 0);
 		break;
 	case CHIP_ARCTURUS:
+		/* This is not fatal.  We only need the discovery
+		 * binary for sysfs.  We don't need it for a
+		 * functional system.
+		 */
+		amdgpu_discovery_init(adev);
 		arct_reg_base_init(adev);
 		adev->sdma.num_instances = 8;
+		adev->sdma.sdma_mask = 0xff;
 		adev->vcn.num_vcn_inst = 2;
+		adev->gmc.num_umc = 8;
+		adev->gfx.xcc_mask = 1;
 		adev->ip_versions[MMHUB_HWIP][0] = IP_VERSION(9, 4, 1);
 		adev->ip_versions[ATHUB_HWIP][0] = IP_VERSION(9, 4, 1);
 		adev->ip_versions[OSSSYS_HWIP][0] = IP_VERSION(4, 2, 1);
@@ -1094,9 +2795,17 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev)
 		adev->ip_versions[UVD_HWIP][1] = IP_VERSION(2, 5, 0);
 		break;
 	case CHIP_ALDEBARAN:
+		/* This is not fatal.  We only need the discovery
+		 * binary for sysfs.  We don't need it for a
+		 * functional system.
+		 */
+		amdgpu_discovery_init(adev);
 		aldebaran_reg_base_init(adev);
 		adev->sdma.num_instances = 5;
+		adev->sdma.sdma_mask = 0x1f;
 		adev->vcn.num_vcn_inst = 2;
+		adev->gmc.num_umc = 4;
+		adev->gfx.xcc_mask = 1;
 		adev->ip_versions[MMHUB_HWIP][0] = IP_VERSION(9, 4, 2);
 		adev->ip_versions[ATHUB_HWIP][0] = IP_VERSION(9, 4, 2);
 		adev->ip_versions[OSSSYS_HWIP][0] = IP_VERSION(4, 4, 0);
@@ -1118,26 +2827,64 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev)
 		adev->ip_versions[UVD_HWIP][1] = IP_VERSION(2, 6, 0);
 		adev->ip_versions[XGMI_HWIP][0] = IP_VERSION(6, 1, 0);
 		break;
+	case CHIP_CYAN_SKILLFISH:
+		if (adev->apu_flags & AMD_APU_IS_CYAN_SKILLFISH2) {
+			r = amdgpu_discovery_reg_base_init(adev);
+			if (r)
+				return -EINVAL;
+
+			amdgpu_discovery_harvest_ip(adev);
+			amdgpu_discovery_get_gfx_info(adev);
+			amdgpu_discovery_get_mall_info(adev);
+			amdgpu_discovery_get_vcn_info(adev);
+		} else {
+			cyan_skillfish_reg_base_init(adev);
+			adev->sdma.num_instances = 2;
+			adev->sdma.sdma_mask = 3;
+			adev->gfx.xcc_mask = 1;
+			adev->ip_versions[MMHUB_HWIP][0] = IP_VERSION(2, 0, 3);
+			adev->ip_versions[ATHUB_HWIP][0] = IP_VERSION(2, 0, 3);
+			adev->ip_versions[OSSSYS_HWIP][0] = IP_VERSION(5, 0, 1);
+			adev->ip_versions[HDP_HWIP][0] = IP_VERSION(5, 0, 1);
+			adev->ip_versions[SDMA0_HWIP][0] = IP_VERSION(5, 0, 1);
+			adev->ip_versions[SDMA1_HWIP][1] = IP_VERSION(5, 0, 1);
+			adev->ip_versions[DF_HWIP][0] = IP_VERSION(3, 5, 0);
+			adev->ip_versions[NBIO_HWIP][0] = IP_VERSION(2, 1, 1);
+			adev->ip_versions[UMC_HWIP][0] = IP_VERSION(8, 1, 1);
+			adev->ip_versions[MP0_HWIP][0] = IP_VERSION(11, 0, 8);
+			adev->ip_versions[MP1_HWIP][0] = IP_VERSION(11, 0, 8);
+			adev->ip_versions[THM_HWIP][0] = IP_VERSION(11, 0, 1);
+			adev->ip_versions[SMUIO_HWIP][0] = IP_VERSION(11, 0, 8);
+			adev->ip_versions[GC_HWIP][0] = IP_VERSION(10, 1, 3);
+			adev->ip_versions[UVD_HWIP][0] = IP_VERSION(2, 0, 3);
+		}
+		break;
 	default:
 		r = amdgpu_discovery_reg_base_init(adev);
-		if (r)
-			return -EINVAL;
+		if (r) {
+			drm_err(&adev->ddev, "discovery failed: %d\n", r);
+			return r;
+		}
 
 		amdgpu_discovery_harvest_ip(adev);
-
-		if (!adev->mman.discovery_bin) {
-			DRM_ERROR("ip discovery uninitialized\n");
-			return -EINVAL;
-		}
+		amdgpu_discovery_get_gfx_info(adev);
+		amdgpu_discovery_get_mall_info(adev);
+		amdgpu_discovery_get_vcn_info(adev);
 		break;
 	}
 
-	switch (adev->ip_versions[GC_HWIP][0]) {
+	amdgpu_discovery_init_soc_config(adev);
+	amdgpu_discovery_sysfs_init(adev);
+
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 	case IP_VERSION(9, 0, 1):
 	case IP_VERSION(9, 2, 1):
 	case IP_VERSION(9, 4, 0):
 	case IP_VERSION(9, 4, 1):
 	case IP_VERSION(9, 4, 2):
+	case IP_VERSION(9, 4, 3):
+	case IP_VERSION(9, 4, 4):
+	case IP_VERSION(9, 5, 0):
 		adev->family = AMDGPU_FAMILY_AI;
 		break;
 	case IP_VERSION(9, 1, 0):
@@ -1149,6 +2896,7 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev)
 	case IP_VERSION(10, 1, 1):
 	case IP_VERSION(10, 1, 2):
 	case IP_VERSION(10, 1, 3):
+	case IP_VERSION(10, 1, 4):
 	case IP_VERSION(10, 3, 0):
 	case IP_VERSION(10, 3, 2):
 	case IP_VERSION(10, 3, 4):
@@ -1157,19 +2905,64 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev)
 		break;
 	case IP_VERSION(10, 3, 1):
 		adev->family = AMDGPU_FAMILY_VGH;
+		adev->apu_flags |= AMD_APU_IS_VANGOGH;
 		break;
 	case IP_VERSION(10, 3, 3):
 		adev->family = AMDGPU_FAMILY_YC;
 		break;
+	case IP_VERSION(10, 3, 6):
+		adev->family = AMDGPU_FAMILY_GC_10_3_6;
+		break;
+	case IP_VERSION(10, 3, 7):
+		adev->family = AMDGPU_FAMILY_GC_10_3_7;
+		break;
+	case IP_VERSION(11, 0, 0):
+	case IP_VERSION(11, 0, 2):
+	case IP_VERSION(11, 0, 3):
+		adev->family = AMDGPU_FAMILY_GC_11_0_0;
+		break;
+	case IP_VERSION(11, 0, 1):
+	case IP_VERSION(11, 0, 4):
+		adev->family = AMDGPU_FAMILY_GC_11_0_1;
+		break;
+	case IP_VERSION(11, 5, 0):
+	case IP_VERSION(11, 5, 1):
+	case IP_VERSION(11, 5, 2):
+	case IP_VERSION(11, 5, 3):
+		adev->family = AMDGPU_FAMILY_GC_11_5_0;
+		break;
+	case IP_VERSION(12, 0, 0):
+	case IP_VERSION(12, 0, 1):
+		adev->family = AMDGPU_FAMILY_GC_12_0_0;
+		break;
 	default:
 		return -EINVAL;
 	}
 
-	if (adev->ip_versions[XGMI_HWIP][0] == IP_VERSION(4, 8, 0))
-		adev->gmc.xgmi.supported = true;
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	case IP_VERSION(9, 1, 0):
+	case IP_VERSION(9, 2, 2):
+	case IP_VERSION(9, 3, 0):
+	case IP_VERSION(10, 1, 3):
+	case IP_VERSION(10, 1, 4):
+	case IP_VERSION(10, 3, 1):
+	case IP_VERSION(10, 3, 3):
+	case IP_VERSION(10, 3, 6):
+	case IP_VERSION(10, 3, 7):
+	case IP_VERSION(11, 0, 1):
+	case IP_VERSION(11, 0, 4):
+	case IP_VERSION(11, 5, 0):
+	case IP_VERSION(11, 5, 1):
+	case IP_VERSION(11, 5, 2):
+	case IP_VERSION(11, 5, 3):
+		adev->flags |= AMD_IS_APU;
+		break;
+	default:
+		break;
+	}
 
 	/* set NBIO version */
-	switch (adev->ip_versions[NBIO_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, NBIO_HWIP, 0)) {
 	case IP_VERSION(6, 1, 0):
 	case IP_VERSION(6, 2, 0):
 		adev->nbio.funcs = &nbio_v6_1_funcs;
@@ -1183,16 +2976,27 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev)
 		break;
 	case IP_VERSION(7, 4, 0):
 	case IP_VERSION(7, 4, 1):
+	case IP_VERSION(7, 4, 4):
 		adev->nbio.funcs = &nbio_v7_4_funcs;
 		adev->nbio.hdp_flush_reg = &nbio_v7_4_hdp_flush_reg;
 		break;
-	case IP_VERSION(7, 4, 4):
-		adev->nbio.funcs = &nbio_v7_4_funcs;
-		adev->nbio.hdp_flush_reg = &nbio_v7_4_hdp_flush_reg_ald;
+	case IP_VERSION(7, 9, 0):
+	case IP_VERSION(7, 9, 1):
+		adev->nbio.funcs = &nbio_v7_9_funcs;
+		adev->nbio.hdp_flush_reg = &nbio_v7_9_hdp_flush_reg;
+		break;
+	case IP_VERSION(7, 11, 0):
+	case IP_VERSION(7, 11, 1):
+	case IP_VERSION(7, 11, 2):
+	case IP_VERSION(7, 11, 3):
+		adev->nbio.funcs = &nbio_v7_11_funcs;
+		adev->nbio.hdp_flush_reg = &nbio_v7_11_hdp_flush_reg;
 		break;
 	case IP_VERSION(7, 2, 0):
 	case IP_VERSION(7, 2, 1):
+	case IP_VERSION(7, 3, 0):
 	case IP_VERSION(7, 5, 0):
+	case IP_VERSION(7, 5, 1):
 		adev->nbio.funcs = &nbio_v7_2_funcs;
 		adev->nbio.hdp_flush_reg = &nbio_v7_2_hdp_flush_reg;
 		break;
@@ -1200,21 +3004,35 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev)
 	case IP_VERSION(2, 3, 0):
 	case IP_VERSION(2, 3, 1):
 	case IP_VERSION(2, 3, 2):
-		adev->nbio.funcs = &nbio_v2_3_funcs;
-		adev->nbio.hdp_flush_reg = &nbio_v2_3_hdp_flush_reg;
-		break;
 	case IP_VERSION(3, 3, 0):
 	case IP_VERSION(3, 3, 1):
 	case IP_VERSION(3, 3, 2):
 	case IP_VERSION(3, 3, 3):
 		adev->nbio.funcs = &nbio_v2_3_funcs;
-		adev->nbio.hdp_flush_reg = &nbio_v2_3_hdp_flush_reg_sc;
+		adev->nbio.hdp_flush_reg = &nbio_v2_3_hdp_flush_reg;
+		break;
+	case IP_VERSION(4, 3, 0):
+	case IP_VERSION(4, 3, 1):
+		if (amdgpu_sriov_vf(adev))
+			adev->nbio.funcs = &nbio_v4_3_sriov_funcs;
+		else
+			adev->nbio.funcs = &nbio_v4_3_funcs;
+		adev->nbio.hdp_flush_reg = &nbio_v4_3_hdp_flush_reg;
+		break;
+	case IP_VERSION(7, 7, 0):
+	case IP_VERSION(7, 7, 1):
+		adev->nbio.funcs = &nbio_v7_7_funcs;
+		adev->nbio.hdp_flush_reg = &nbio_v7_7_hdp_flush_reg;
+		break;
+	case IP_VERSION(6, 3, 1):
+		adev->nbio.funcs = &nbif_v6_3_1_funcs;
+		adev->nbio.hdp_flush_reg = &nbif_v6_3_1_hdp_flush_reg;
 		break;
 	default:
 		break;
 	}
 
-	switch (adev->ip_versions[HDP_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, HDP_HWIP, 0)) {
 	case IP_VERSION(4, 0, 0):
 	case IP_VERSION(4, 0, 1):
 	case IP_VERSION(4, 1, 0):
@@ -1223,6 +3041,8 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev)
 	case IP_VERSION(4, 2, 0):
 	case IP_VERSION(4, 2, 1):
 	case IP_VERSION(4, 4, 0):
+	case IP_VERSION(4, 4, 2):
+	case IP_VERSION(4, 4, 5):
 		adev->hdp.funcs = &hdp_v4_0_funcs;
 		break;
 	case IP_VERSION(5, 0, 0):
@@ -1233,11 +3053,22 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev)
 	case IP_VERSION(5, 2, 0):
 		adev->hdp.funcs = &hdp_v5_0_funcs;
 		break;
+	case IP_VERSION(5, 2, 1):
+		adev->hdp.funcs = &hdp_v5_2_funcs;
+		break;
+	case IP_VERSION(6, 0, 0):
+	case IP_VERSION(6, 0, 1):
+	case IP_VERSION(6, 1, 0):
+		adev->hdp.funcs = &hdp_v6_0_funcs;
+		break;
+	case IP_VERSION(7, 0, 0):
+		adev->hdp.funcs = &hdp_v7_0_funcs;
+		break;
 	default:
 		break;
 	}
 
-	switch (adev->ip_versions[DF_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, DF_HWIP, 0)) {
 	case IP_VERSION(3, 6, 0):
 	case IP_VERSION(3, 6, 1):
 	case IP_VERSION(3, 6, 2):
@@ -1250,11 +3081,21 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev)
 	case IP_VERSION(3, 5, 2):
 		adev->df.funcs = &df_v1_7_funcs;
 		break;
+	case IP_VERSION(4, 3, 0):
+		adev->df.funcs = &df_v4_3_funcs;
+		break;
+	case IP_VERSION(4, 6, 2):
+		adev->df.funcs = &df_v4_6_2_funcs;
+		break;
+	case IP_VERSION(4, 15, 0):
+	case IP_VERSION(4, 15, 1):
+		adev->df.funcs = &df_v4_15_funcs;
+		break;
 	default:
 		break;
 	}
 
-	switch (adev->ip_versions[SMUIO_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, SMUIO_HWIP, 0)) {
 	case IP_VERSION(9, 0, 0):
 	case IP_VERSION(9, 0, 1):
 	case IP_VERSION(10, 0, 0):
@@ -1274,12 +3115,46 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev)
 	case IP_VERSION(11, 0, 10):
 	case IP_VERSION(11, 0, 11):
 	case IP_VERSION(11, 5, 0):
+	case IP_VERSION(11, 5, 2):
 	case IP_VERSION(13, 0, 1):
+	case IP_VERSION(13, 0, 9):
+	case IP_VERSION(13, 0, 10):
 		adev->smuio.funcs = &smuio_v11_0_6_funcs;
 		break;
 	case IP_VERSION(13, 0, 2):
 		adev->smuio.funcs = &smuio_v13_0_funcs;
 		break;
+	case IP_VERSION(13, 0, 3):
+	case IP_VERSION(13, 0, 11):
+		adev->smuio.funcs = &smuio_v13_0_3_funcs;
+		if (adev->smuio.funcs->get_pkg_type(adev) == AMDGPU_PKG_TYPE_APU) {
+			adev->flags |= AMD_IS_APU;
+		}
+		break;
+	case IP_VERSION(13, 0, 6):
+	case IP_VERSION(13, 0, 8):
+	case IP_VERSION(14, 0, 0):
+	case IP_VERSION(14, 0, 1):
+		adev->smuio.funcs = &smuio_v13_0_6_funcs;
+		break;
+	case IP_VERSION(14, 0, 2):
+		adev->smuio.funcs = &smuio_v14_0_2_funcs;
+		break;
+	default:
+		break;
+	}
+
+	switch (amdgpu_ip_version(adev, LSDMA_HWIP, 0)) {
+	case IP_VERSION(6, 0, 0):
+	case IP_VERSION(6, 0, 1):
+	case IP_VERSION(6, 0, 2):
+	case IP_VERSION(6, 0, 3):
+		adev->lsdma.funcs = &lsdma_v6_0_funcs;
+		break;
+	case IP_VERSION(7, 0, 0):
+	case IP_VERSION(7, 0, 1):
+		adev->lsdma.funcs = &lsdma_v7_0_funcs;
+		break;
 	default:
 		break;
 	}
@@ -1330,8 +3205,13 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev)
 	if (r)
 		return r;
 
-	if (adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT &&
-	    !amdgpu_sriov_vf(adev)) {
+	r = amdgpu_discovery_set_ras_ip_blocks(adev);
+	if (r)
+		return r;
+
+	if ((adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT &&
+	     !amdgpu_sriov_vf(adev)) ||
+	    (adev->firmware.load_type == AMDGPU_FW_LOAD_RLC_BACKDOOR_AUTO && amdgpu_dpm == 1)) {
 		r = amdgpu_discovery_set_smu_ip_blocks(adev);
 		if (r)
 			return r;
@@ -1341,12 +3221,21 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev)
 	if (r)
 		return r;
 
-	if (adev->enable_mes) {
-		r = amdgpu_discovery_set_mes_ip_blocks(adev);
-		if (r)
-			return r;
-	}
+	r = amdgpu_discovery_set_mes_ip_blocks(adev);
+	if (r)
+		return r;
+
+	r = amdgpu_discovery_set_vpe_ip_blocks(adev);
+	if (r)
+		return r;
 
+	r = amdgpu_discovery_set_umsch_mm_ip_blocks(adev);
+	if (r)
+		return r;
+
+	r = amdgpu_discovery_set_isp_ip_blocks(adev);
+	if (r)
+		return r;
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.h
index 0ea029e3b850..4ce04486cc31 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.h
@@ -24,18 +24,27 @@
 #ifndef __AMDGPU_DISCOVERY__
 #define __AMDGPU_DISCOVERY__
 
-#define DISCOVERY_TMR_SIZE      (4 << 10)
+#include <linux/debugfs.h>
+
+#define DISCOVERY_TMR_SIZE      (10 << 10)
 #define DISCOVERY_TMR_OFFSET    (64 << 10)
 
+struct ip_discovery_top;
+
+struct amdgpu_discovery_info {
+	struct debugfs_blob_wrapper debugfs_blob;
+	struct ip_discovery_top *ip_top;
+	uint32_t size;
+	uint8_t *bin;
+	bool reserve_tmr;
+};
+
 void amdgpu_discovery_fini(struct amdgpu_device *adev);
-int amdgpu_discovery_reg_base_init(struct amdgpu_device *adev);
-void amdgpu_discovery_harvest_ip(struct amdgpu_device *adev);
-int amdgpu_discovery_get_ip_version(struct amdgpu_device *adev, int hw_id, int number_instance,
-                                    int *major, int *minor, int *revision);
-
-int amdgpu_discovery_get_vcn_version(struct amdgpu_device *adev, int vcn_instance,
-				     int *major, int *minor, int *revision);
-int amdgpu_discovery_get_gfx_info(struct amdgpu_device *adev);
 int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev);
 
+int amdgpu_discovery_get_nps_info(struct amdgpu_device *adev,
+				  uint32_t *nps_type,
+				  struct amdgpu_gmc_memrange **ranges,
+				  int *range_cnt, bool refresh);
+
 #endif /* __AMDGPU_DISCOVERY__ */
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c
index dc50c05f23fc..b5d34797d606 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c
@@ -30,17 +30,64 @@
 #include "atom.h"
 #include "amdgpu_connectors.h"
 #include "amdgpu_display.h"
+#include "soc15_common.h"
+#include "gc/gc_11_0_0_offset.h"
+#include "gc/gc_11_0_0_sh_mask.h"
+#include "bif/bif_4_1_d.h"
 #include <asm/div64.h>
 
 #include <linux/pci.h>
 #include <linux/pm_runtime.h>
 #include <drm/drm_crtc_helper.h>
+#include <drm/drm_damage_helper.h>
+#include <drm/drm_drv.h>
 #include <drm/drm_edid.h>
-#include <drm/drm_gem_framebuffer_helper.h>
 #include <drm/drm_fb_helper.h>
+#include <drm/drm_gem_framebuffer_helper.h>
 #include <drm/drm_fourcc.h>
+#include <drm/drm_modeset_helper.h>
 #include <drm/drm_vblank.h>
 
+/**
+ * amdgpu_display_hotplug_work_func - work handler for display hotplug event
+ *
+ * @work: work struct pointer
+ *
+ * This is the hotplug event work handler (all ASICs).
+ * The work gets scheduled from the IRQ handler if there
+ * was a hotplug interrupt.  It walks through the connector table
+ * and calls hotplug handler for each connector. After this, it sends
+ * a DRM hotplug event to alert userspace.
+ *
+ * This design approach is required in order to defer hotplug event handling
+ * from the IRQ handler to a work handler because hotplug handler has to use
+ * mutexes which cannot be locked in an IRQ handler (since &mutex_lock may
+ * sleep).
+ */
+void amdgpu_display_hotplug_work_func(struct work_struct *work)
+{
+	struct amdgpu_device *adev = container_of(work, struct amdgpu_device,
+						  hotplug_work.work);
+	struct drm_device *dev = adev_to_drm(adev);
+	struct drm_mode_config *mode_config = &dev->mode_config;
+	struct drm_connector *connector;
+	struct drm_connector_list_iter iter;
+
+	mutex_lock(&mode_config->mutex);
+	drm_connector_list_iter_begin(dev, &iter);
+	drm_for_each_connector_iter(connector, &iter)
+		amdgpu_connector_hotplug(connector);
+	drm_connector_list_iter_end(&iter);
+	mutex_unlock(&mode_config->mutex);
+	/* Just fire off a uevent and let userspace tell us what to do */
+	drm_helper_hpd_irq_event(dev);
+}
+
+static int amdgpu_display_framebuffer_init(struct drm_device *dev,
+					   struct amdgpu_framebuffer *rfb,
+					   const struct drm_mode_fb_cmd2 *mode_cmd,
+					   struct drm_gem_object *obj);
+
 static void amdgpu_display_flip_callback(struct dma_fence *f,
 					 struct dma_fence_cb *cb)
 {
@@ -54,7 +101,7 @@ static void amdgpu_display_flip_callback(struct dma_fence *f,
 static bool amdgpu_display_flip_handle_fence(struct amdgpu_flip_work *work,
 					     struct dma_fence **f)
 {
-	struct dma_fence *fence= *f;
+	struct dma_fence *fence = *f;
 
 	if (fence == NULL)
 		return false;
@@ -80,12 +127,9 @@ static void amdgpu_display_flip_work_func(struct work_struct *__work)
 
 	struct drm_crtc *crtc = &amdgpu_crtc->base;
 	unsigned long flags;
-	unsigned i;
+	unsigned int i;
 	int vpos, hpos;
 
-	if (amdgpu_display_flip_handle_fence(work, &work->excl))
-		return;
-
 	for (i = 0; i < work->shared_count; ++i)
 		if (amdgpu_display_flip_handle_fence(work, &work->shared[i]))
 			return;
@@ -116,8 +160,9 @@ static void amdgpu_display_flip_work_func(struct work_struct *__work)
 	spin_unlock_irqrestore(&crtc->dev->event_lock, flags);
 
 
-	DRM_DEBUG_DRIVER("crtc:%d[%p], pflip_stat:AMDGPU_FLIP_SUBMITTED, work: %p,\n",
-					 amdgpu_crtc->crtc_id, amdgpu_crtc, work);
+	drm_dbg_vbl(adev_to_drm(adev),
+		    "crtc:%d[%p], pflip_stat:AMDGPU_FLIP_SUBMITTED, work: %p,\n",
+		    amdgpu_crtc->crtc_id, amdgpu_crtc, work);
 
 }
 
@@ -159,7 +204,7 @@ int amdgpu_display_crtc_page_flip_target(struct drm_crtc *crtc,
 	u64 tiling_flags;
 	int i, r;
 
-	work = kzalloc(sizeof *work, GFP_KERNEL);
+	work = kzalloc(sizeof(*work), GFP_KERNEL);
 	if (work == NULL)
 		return -ENOMEM;
 
@@ -189,6 +234,7 @@ int amdgpu_display_crtc_page_flip_target(struct drm_crtc *crtc,
 	}
 
 	if (!adev->enable_virtual_display) {
+		new_abo->flags |= AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS;
 		r = amdgpu_bo_pin(new_abo,
 				  amdgpu_display_supported_domains(adev, new_abo->flags));
 		if (unlikely(r != 0)) {
@@ -203,8 +249,9 @@ int amdgpu_display_crtc_page_flip_target(struct drm_crtc *crtc,
 		goto unpin;
 	}
 
-	r = dma_resv_get_fences(new_abo->tbo.base.resv, &work->excl,
-				&work->shared_count, &work->shared);
+	r = dma_resv_get_fences(new_abo->tbo.base.resv, DMA_RESV_USAGE_WRITE,
+				&work->shared_count,
+				&work->shared);
 	if (unlikely(r != 0)) {
 		DRM_ERROR("failed to get fences for buffer\n");
 		goto unpin;
@@ -253,7 +300,6 @@ unreserve:
 
 cleanup:
 	amdgpu_bo_unref(&work->old_abo);
-	dma_fence_put(work->excl);
 	for (i = 0; i < work->shared_count; ++i)
 		dma_fence_put(work->shared[i]);
 	kfree(work->shared);
@@ -286,22 +332,19 @@ int amdgpu_display_crtc_set_config(struct drm_mode_set *set,
 		if (crtc->enabled)
 			active = true;
 
-	pm_runtime_mark_last_busy(dev->dev);
-
 	adev = drm_to_adev(dev);
 	/* if we have active crtcs and we don't have a power ref,
-	   take the current one */
+	 * take the current one
+	 */
 	if (active && !adev->have_disp_power_ref) {
 		adev->have_disp_power_ref = true;
 		return ret;
 	}
-	/* if we have no active crtcs, then drop the power ref
-	   we got before */
-	if (!active && adev->have_disp_power_ref) {
-		pm_runtime_put_autosuspend(dev->dev);
+	/* if we have no active crtcs, then go to
+	 * drop the power ref we got before
+	 */
+	if (!active && adev->have_disp_power_ref)
 		adev->have_disp_power_ref = false;
-	}
-
 out:
 	/* drop the power reference we got coming in here */
 	pm_runtime_put_autosuspend(dev->dev);
@@ -465,11 +508,10 @@ bool amdgpu_display_ddc_probe(struct amdgpu_connector *amdgpu_connector,
 	if (amdgpu_connector->router.ddc_valid)
 		amdgpu_i2c_router_select_ddc_port(amdgpu_connector);
 
-	if (use_aux) {
+	if (use_aux)
 		ret = i2c_transfer(&amdgpu_connector->ddc_bus->aux.ddc, msgs, 2);
-	} else {
+	else
 		ret = i2c_transfer(&amdgpu_connector->ddc_bus->adapter, msgs, 2);
-	}
 
 	if (ret != 2)
 		/* Couldn't find an accessible DDC on this connector */
@@ -478,20 +520,40 @@ bool amdgpu_display_ddc_probe(struct amdgpu_connector *amdgpu_connector,
 	 * EDID header starts with:
 	 * 0x00,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0x00.
 	 * Only the first 6 bytes must be valid as
-	 * drm_edid_block_valid() can fix the last 2 bytes */
+	 * drm_edid_block_valid() can fix the last 2 bytes
+	 */
 	if (drm_edid_header_is_valid(buf) < 6) {
 		/* Couldn't find an accessible EDID on this
-		 * connector */
+		 * connector
+		 */
 		return false;
 	}
 	return true;
 }
 
+static int amdgpu_dirtyfb(struct drm_framebuffer *fb, struct drm_file *file,
+			  unsigned int flags, unsigned int color,
+			  struct drm_clip_rect *clips, unsigned int num_clips)
+{
+
+	if (file)
+		return -ENOSYS;
+
+	return drm_atomic_helper_dirtyfb(fb, file, flags, color, clips,
+					 num_clips);
+}
+
 static const struct drm_framebuffer_funcs amdgpu_fb_funcs = {
 	.destroy = drm_gem_fb_destroy,
 	.create_handle = drm_gem_fb_create_handle,
 };
 
+static const struct drm_framebuffer_funcs amdgpu_fb_funcs_atomic = {
+	.destroy = drm_gem_fb_destroy,
+	.create_handle = drm_gem_fb_create_handle,
+	.dirty = amdgpu_dirtyfb
+};
+
 uint32_t amdgpu_display_supported_domains(struct amdgpu_device *adev,
 					  uint64_t bo_flags)
 {
@@ -508,28 +570,9 @@ uint32_t amdgpu_display_supported_domains(struct amdgpu_device *adev,
 	 */
 	if ((bo_flags & AMDGPU_GEM_CREATE_CPU_GTT_USWC) &&
 	    amdgpu_bo_support_uswc(bo_flags) &&
-	    amdgpu_device_asic_has_dc_support(adev->asic_type)) {
-		switch (adev->asic_type) {
-		case CHIP_CARRIZO:
-		case CHIP_STONEY:
-			domain |= AMDGPU_GEM_DOMAIN_GTT;
-			break;
-		case CHIP_RAVEN:
-			/* enable S/G on PCO and RV2 */
-			if ((adev->apu_flags & AMD_APU_IS_RAVEN2) ||
-			    (adev->apu_flags & AMD_APU_IS_PICASSO))
-				domain |= AMDGPU_GEM_DOMAIN_GTT;
-			break;
-		case CHIP_RENOIR:
-		case CHIP_VANGOGH:
-		case CHIP_YELLOW_CARP:
-			domain |= AMDGPU_GEM_DOMAIN_GTT;
-			break;
-
-		default:
-			break;
-		}
-	}
+	    adev->dc_enabled &&
+	    adev->mode_info.gpu_vm_support)
+		domain |= AMDGPU_GEM_DOMAIN_GTT;
 #endif
 
 	return domain;
@@ -611,6 +654,10 @@ amdgpu_lookup_format_info(u32 format, uint64_t modifier)
 	if (!IS_AMD_FMT_MOD(modifier))
 		return NULL;
 
+	if (AMD_FMT_MOD_GET(TILE_VERSION, modifier) < AMD_FMT_MOD_TILE_VER_GFX9 ||
+	    AMD_FMT_MOD_GET(TILE_VERSION, modifier) >= AMD_FMT_MOD_TILE_VER_GFX12)
+		return NULL;
+
 	if (AMD_FMT_MOD_GET(DCC_RETILE, modifier))
 		return lookup_format_info(dcc_retile_formats,
 					  ARRAY_SIZE(dcc_retile_formats),
@@ -675,10 +722,39 @@ extract_render_dcc_offset(struct amdgpu_device *adev,
 	return 0;
 }
 
+static int convert_tiling_flags_to_modifier_gfx12(struct amdgpu_framebuffer *afb)
+{
+	u64 modifier = 0;
+	int swizzle_mode = AMDGPU_TILING_GET(afb->tiling_flags, GFX12_SWIZZLE_MODE);
+
+	if (!swizzle_mode) {
+		modifier = DRM_FORMAT_MOD_LINEAR;
+	} else {
+		int max_comp_block =
+			AMDGPU_TILING_GET(afb->tiling_flags, GFX12_DCC_MAX_COMPRESSED_BLOCK);
+
+		modifier =
+			AMD_FMT_MOD |
+			AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX12) |
+			AMD_FMT_MOD_SET(TILE, swizzle_mode) |
+			AMD_FMT_MOD_SET(DCC, afb->gfx12_dcc) |
+			AMD_FMT_MOD_SET(DCC_MAX_COMPRESSED_BLOCK, max_comp_block);
+	}
+
+	afb->base.modifier = modifier;
+	afb->base.flags |= DRM_MODE_FB_MODIFIERS;
+	return 0;
+}
+
 static int convert_tiling_flags_to_modifier(struct amdgpu_framebuffer *afb)
 {
 	struct amdgpu_device *adev = drm_to_adev(afb->base.dev);
 	uint64_t modifier = 0;
+	int num_pipes = 0;
+	int num_pkrs = 0;
+
+	num_pkrs = adev->gfx.config.gb_addr_config_fields.num_pkrs;
+	num_pipes = adev->gfx.config.gb_addr_config_fields.num_pipes;
 
 	if (!afb->tiling_flags || !AMDGPU_TILING_GET(afb->tiling_flags, SWIZZLE_MODE)) {
 		modifier = DRM_FORMAT_MOD_LINEAR;
@@ -691,7 +767,7 @@ static int convert_tiling_flags_to_modifier(struct amdgpu_framebuffer *afb)
 		int bank_xor_bits = 0;
 		int packers = 0;
 		int rb = 0;
-		int pipes = ilog2(adev->gfx.config.gb_addr_config_fields.num_pipes);
+		int pipes = ilog2(num_pipes);
 		uint32_t dcc_offset = AMDGPU_TILING_GET(afb->tiling_flags, DCC_OFFSET_256B);
 
 		switch (swizzle >> 2) {
@@ -707,14 +783,21 @@ static int convert_tiling_flags_to_modifier(struct amdgpu_framebuffer *afb)
 		case 6: /* 64 KiB _X */
 			block_size_bits = 16;
 			break;
+		case 7: /* 256 KiB */
+			block_size_bits = 18;
+			break;
 		default:
 			/* RESERVED or VAR */
 			return -EINVAL;
 		}
 
-		if (adev->asic_type >= CHIP_SIENNA_CICHLID)
+		if (amdgpu_ip_version(adev, GC_HWIP, 0) >= IP_VERSION(11, 0, 0))
+			version = AMD_FMT_MOD_TILE_VER_GFX11;
+		else if (amdgpu_ip_version(adev, GC_HWIP, 0) >=
+			 IP_VERSION(10, 3, 0))
 			version = AMD_FMT_MOD_TILE_VER_GFX10_RBPLUS;
-		else if (adev->family == AMDGPU_FAMILY_NV)
+		else if (amdgpu_ip_version(adev, GC_HWIP, 0) >=
+			 IP_VERSION(10, 0, 0))
 			version = AMD_FMT_MOD_TILE_VER_GFX10;
 		else
 			version = AMD_FMT_MOD_TILE_VER_GFX9;
@@ -723,19 +806,34 @@ static int convert_tiling_flags_to_modifier(struct amdgpu_framebuffer *afb)
 		case 0: /* Z microtiling */
 			return -EINVAL;
 		case 1: /* S microtiling */
-			if (!has_xor)
-				version = AMD_FMT_MOD_TILE_VER_GFX9;
+			if (amdgpu_ip_version(adev, GC_HWIP, 0) <
+			    IP_VERSION(11, 0, 0)) {
+				if (!has_xor)
+					version = AMD_FMT_MOD_TILE_VER_GFX9;
+			}
 			break;
 		case 2:
-			if (!has_xor && afb->base.format->cpp[0] != 4)
-				version = AMD_FMT_MOD_TILE_VER_GFX9;
+			if (amdgpu_ip_version(adev, GC_HWIP, 0) <
+			    IP_VERSION(11, 0, 0)) {
+				if (!has_xor && afb->base.format->cpp[0] != 4)
+					version = AMD_FMT_MOD_TILE_VER_GFX9;
+			}
 			break;
 		case 3:
 			break;
 		}
 
 		if (has_xor) {
+			if (num_pipes == num_pkrs && num_pkrs == 0) {
+				DRM_ERROR("invalid number of pipes and packers\n");
+				return -EINVAL;
+			}
+
 			switch (version) {
+			case AMD_FMT_MOD_TILE_VER_GFX11:
+				pipe_xor_bits = min(block_size_bits - 8, pipes);
+				packers = ilog2(adev->gfx.config.gb_addr_config_fields.num_pkrs);
+				break;
 			case AMD_FMT_MOD_TILE_VER_GFX10_RBPLUS:
 				pipe_xor_bits = min(block_size_bits - 8, pipes);
 				packers = min(block_size_bits - 8 - pipe_xor_bits,
@@ -769,9 +867,12 @@ static int convert_tiling_flags_to_modifier(struct amdgpu_framebuffer *afb)
 			u64 render_dcc_offset;
 
 			/* Enable constant encode on RAVEN2 and later. */
-			bool dcc_constant_encode = adev->asic_type > CHIP_RAVEN ||
-						   (adev->asic_type == CHIP_RAVEN &&
-						    adev->external_rev_id >= 0x81);
+			bool dcc_constant_encode =
+				(adev->asic_type > CHIP_RAVEN ||
+				 (adev->asic_type == CHIP_RAVEN &&
+				  adev->external_rev_id >= 0x81)) &&
+				amdgpu_ip_version(adev, GC_HWIP, 0) <
+					IP_VERSION(11, 0, 0);
 
 			int max_cblock_size = dcc_i64b ? AMD_FMT_MOD_DCC_BLOCK_64B :
 					      dcc_i128b ? AMD_FMT_MOD_DCC_BLOCK_128B :
@@ -808,7 +909,9 @@ static int convert_tiling_flags_to_modifier(struct amdgpu_framebuffer *afb)
 				if (adev->family >= AMDGPU_FAMILY_NV) {
 					int extra_pipe = 0;
 
-					if (adev->asic_type >= CHIP_SIENNA_CICHLID &&
+					if ((amdgpu_ip_version(adev, GC_HWIP,
+							       0) >=
+					     IP_VERSION(10, 3, 0)) &&
 					    pipes == packers && pipes > 1)
 						extra_pipe = 1;
 
@@ -842,8 +945,7 @@ static int check_tiling_flags_gfx6(struct amdgpu_framebuffer *afb)
 {
 	u64 micro_tile_mode;
 
-	/* Zero swizzle mode means linear */
-	if (AMDGPU_TILING_GET(afb->tiling_flags, SWIZZLE_MODE) == 0)
+	if (AMDGPU_TILING_GET(afb->tiling_flags, ARRAY_MODE) == 1) /* LINEAR_ALIGNED */
 		return 0;
 
 	micro_tile_mode = AMDGPU_TILING_GET(afb->tiling_flags, MICRO_TILE_MODE);
@@ -886,10 +988,11 @@ static unsigned int get_dcc_block_size(uint64_t modifier, bool rb_aligned,
 		return max(10 + (rb_aligned ? (int)AMD_FMT_MOD_GET(RB, modifier) : 0), 12);
 	}
 	case AMD_FMT_MOD_TILE_VER_GFX10:
-	case AMD_FMT_MOD_TILE_VER_GFX10_RBPLUS: {
+	case AMD_FMT_MOD_TILE_VER_GFX10_RBPLUS:
+	case AMD_FMT_MOD_TILE_VER_GFX11: {
 		int pipes_log2 = AMD_FMT_MOD_GET(PIPE_XOR_BITS, modifier);
 
-		if (ver == AMD_FMT_MOD_TILE_VER_GFX10_RBPLUS && pipes_log2 > 1 &&
+		if (ver >= AMD_FMT_MOD_TILE_VER_GFX10_RBPLUS && pipes_log2 > 1 &&
 		    AMD_FMT_MOD_GET(PACKERS, modifier) == pipes_log2)
 			++pipes_log2;
 
@@ -958,7 +1061,7 @@ static int amdgpu_display_verify_sizes(struct amdgpu_framebuffer *rfb)
 	int ret;
 	unsigned int i, block_width, block_height, block_size_log2;
 
-	if (!rfb->base.dev->mode_config.allow_fb_modifiers)
+	if (rfb->base.dev->mode_config.fb_modifiers_not_supported)
 		return 0;
 
 	for (i = 0; i < format_info->num_planes; ++i) {
@@ -966,6 +1069,30 @@ static int amdgpu_display_verify_sizes(struct amdgpu_framebuffer *rfb)
 			block_width = 256 / format_info->cpp[i];
 			block_height = 1;
 			block_size_log2 = 8;
+		} else if (AMD_FMT_MOD_GET(TILE_VERSION, modifier) >= AMD_FMT_MOD_TILE_VER_GFX12) {
+			int swizzle = AMD_FMT_MOD_GET(TILE, modifier);
+
+			switch (swizzle) {
+			case AMD_FMT_MOD_TILE_GFX12_256B_2D:
+				block_size_log2 = 8;
+				break;
+			case AMD_FMT_MOD_TILE_GFX12_4K_2D:
+				block_size_log2 = 12;
+				break;
+			case AMD_FMT_MOD_TILE_GFX12_64K_2D:
+				block_size_log2 = 16;
+				break;
+			case AMD_FMT_MOD_TILE_GFX12_256K_2D:
+				block_size_log2 = 18;
+				break;
+			default:
+				drm_dbg_kms(rfb->base.dev,
+					    "Gfx12 swizzle mode with unknown block size: %d\n", swizzle);
+				return -EINVAL;
+			}
+
+			get_block_dimensions(block_size_log2, format_info->cpp[i],
+					     &block_width, &block_height);
 		} else {
 			int swizzle = AMD_FMT_MOD_GET(TILE, modifier);
 
@@ -982,6 +1109,9 @@ static int amdgpu_display_verify_sizes(struct amdgpu_framebuffer *rfb)
 			case DC_SW_64KB_S_X:
 				block_size_log2 = 16;
 				break;
+			case DC_SW_VAR_S_X:
+				block_size_log2 = 18;
+				break;
 			default:
 				drm_dbg_kms(rfb->base.dev,
 					    "Swizzle mode with unknown block size: %d\n", swizzle);
@@ -998,7 +1128,8 @@ static int amdgpu_display_verify_sizes(struct amdgpu_framebuffer *rfb)
 			return ret;
 	}
 
-	if (AMD_FMT_MOD_GET(DCC, modifier)) {
+	if (AMD_FMT_MOD_GET(TILE_VERSION, modifier) <= AMD_FMT_MOD_TILE_VER_GFX11 &&
+	    AMD_FMT_MOD_GET(DCC, modifier)) {
 		if (AMD_FMT_MOD_GET(DCC_RETILE, modifier)) {
 			block_size_log2 = get_dcc_block_size(modifier, false, false);
 			get_block_dimensions(block_size_log2 + 8, format_info->cpp[0],
@@ -1028,7 +1159,8 @@ static int amdgpu_display_verify_sizes(struct amdgpu_framebuffer *rfb)
 }
 
 static int amdgpu_display_get_fb_info(const struct amdgpu_framebuffer *amdgpu_fb,
-				      uint64_t *tiling_flags, bool *tmz_surface)
+				      uint64_t *tiling_flags, bool *tmz_surface,
+				      bool *gfx12_dcc)
 {
 	struct amdgpu_bo *rbo;
 	int r;
@@ -1036,6 +1168,7 @@ static int amdgpu_display_get_fb_info(const struct amdgpu_framebuffer *amdgpu_fb
 	if (!amdgpu_fb) {
 		*tiling_flags = 0;
 		*tmz_surface = false;
+		*gfx12_dcc = false;
 		return 0;
 	}
 
@@ -1049,51 +1182,26 @@ static int amdgpu_display_get_fb_info(const struct amdgpu_framebuffer *amdgpu_fb
 		return r;
 	}
 
-	if (tiling_flags)
-		amdgpu_bo_get_tiling_flags(rbo, tiling_flags);
-
-	if (tmz_surface)
-		*tmz_surface = amdgpu_bo_encrypted(rbo);
+	amdgpu_bo_get_tiling_flags(rbo, tiling_flags);
+	*tmz_surface = amdgpu_bo_encrypted(rbo);
+	*gfx12_dcc = rbo->flags & AMDGPU_GEM_CREATE_GFX12_DCC;
 
 	amdgpu_bo_unreserve(rbo);
 
 	return r;
 }
 
-int amdgpu_display_gem_fb_init(struct drm_device *dev,
-			       struct amdgpu_framebuffer *rfb,
-			       const struct drm_mode_fb_cmd2 *mode_cmd,
-			       struct drm_gem_object *obj)
-{
-	int ret;
-
-	rfb->base.obj[0] = obj;
-	drm_helper_mode_fill_fb_struct(dev, &rfb->base, mode_cmd);
-
-	ret = amdgpu_display_framebuffer_init(dev, rfb, mode_cmd, obj);
-	if (ret)
-		goto err;
-
-	ret = drm_framebuffer_init(dev, &rfb->base, &amdgpu_fb_funcs);
-	if (ret)
-		goto err;
-
-	return 0;
-err:
-	drm_dbg_kms(dev, "Failed to init gem fb: %d\n", ret);
-	rfb->base.obj[0] = NULL;
-	return ret;
-}
-
-int amdgpu_display_gem_fb_verify_and_init(
-	struct drm_device *dev, struct amdgpu_framebuffer *rfb,
-	struct drm_file *file_priv, const struct drm_mode_fb_cmd2 *mode_cmd,
-	struct drm_gem_object *obj)
+static int amdgpu_display_gem_fb_verify_and_init(struct drm_device *dev,
+						 struct amdgpu_framebuffer *rfb,
+						 struct drm_file *file_priv,
+						 const struct drm_format_info *info,
+						 const struct drm_mode_fb_cmd2 *mode_cmd,
+						 struct drm_gem_object *obj)
 {
 	int ret;
 
 	rfb->base.obj[0] = obj;
-	drm_helper_mode_fill_fb_struct(dev, &rfb->base, mode_cmd);
+	drm_helper_mode_fill_fb_struct(dev, &rfb->base, info, mode_cmd);
 	/* Verify that the modifier is supported. */
 	if (!drm_any_plane_has_format(dev, mode_cmd->pixel_format,
 				      mode_cmd->modifier[0])) {
@@ -1109,7 +1217,12 @@ int amdgpu_display_gem_fb_verify_and_init(
 	if (ret)
 		goto err;
 
-	ret = drm_framebuffer_init(dev, &rfb->base, &amdgpu_fb_funcs);
+	if (drm_drv_uses_atomic_modeset(dev))
+		ret = drm_framebuffer_init(dev, &rfb->base,
+					   &amdgpu_fb_funcs_atomic);
+	else
+		ret = drm_framebuffer_init(dev, &rfb->base, &amdgpu_fb_funcs);
+
 	if (ret)
 		goto err;
 
@@ -1120,10 +1233,10 @@ err:
 	return ret;
 }
 
-int amdgpu_display_framebuffer_init(struct drm_device *dev,
-				    struct amdgpu_framebuffer *rfb,
-				    const struct drm_mode_fb_cmd2 *mode_cmd,
-				    struct drm_gem_object *obj)
+static int amdgpu_display_framebuffer_init(struct drm_device *dev,
+					   struct amdgpu_framebuffer *rfb,
+					   const struct drm_mode_fb_cmd2 *mode_cmd,
+					   struct drm_gem_object *obj)
 {
 	struct amdgpu_device *adev = drm_to_adev(dev);
 	int ret, i;
@@ -1141,11 +1254,12 @@ int amdgpu_display_framebuffer_init(struct drm_device *dev,
 		}
 	}
 
-	ret = amdgpu_display_get_fb_info(rfb, &rfb->tiling_flags, &rfb->tmz_surface);
+	ret = amdgpu_display_get_fb_info(rfb, &rfb->tiling_flags, &rfb->tmz_surface,
+					 &rfb->gfx12_dcc);
 	if (ret)
 		return ret;
 
-	if (!dev->mode_config.allow_fb_modifiers) {
+	if (dev->mode_config.fb_modifiers_not_supported && !adev->enable_virtual_display) {
 		drm_WARN_ONCE(dev, adev->family >= AMDGPU_FAMILY_AI,
 			      "GFX9+ requires FB check based on format modifier\n");
 		ret = check_tiling_flags_gfx6(rfb);
@@ -1153,9 +1267,13 @@ int amdgpu_display_framebuffer_init(struct drm_device *dev,
 			return ret;
 	}
 
-	if (dev->mode_config.allow_fb_modifiers &&
+	if (!dev->mode_config.fb_modifiers_not_supported &&
 	    !(rfb->base.flags & DRM_MODE_FB_MODIFIERS)) {
-		ret = convert_tiling_flags_to_modifier(rfb);
+		if (amdgpu_ip_version(adev, GC_HWIP, 0) >= IP_VERSION(12, 0, 0))
+			ret = convert_tiling_flags_to_modifier_gfx12(rfb);
+		else
+			ret = convert_tiling_flags_to_modifier(rfb);
+
 		if (ret) {
 			drm_dbg_kms(dev, "Failed to convert tiling flags 0x%llX to a modifier",
 				    rfb->tiling_flags);
@@ -1178,6 +1296,7 @@ int amdgpu_display_framebuffer_init(struct drm_device *dev,
 struct drm_framebuffer *
 amdgpu_display_user_framebuffer_create(struct drm_device *dev,
 				       struct drm_file *file_priv,
+				       const struct drm_format_info *info,
 				       const struct drm_mode_fb_cmd2 *mode_cmd)
 {
 	struct amdgpu_framebuffer *amdgpu_fb;
@@ -1188,15 +1307,17 @@ amdgpu_display_user_framebuffer_create(struct drm_device *dev,
 
 	obj = drm_gem_object_lookup(file_priv, mode_cmd->handles[0]);
 	if (obj ==  NULL) {
-		drm_dbg_kms(dev, "No GEM object associated to handle 0x%08X, "
-			    "can't create framebuffer\n", mode_cmd->handles[0]);
+		drm_dbg_kms(dev,
+			    "No GEM object associated to handle 0x%08X, can't create framebuffer\n",
+			    mode_cmd->handles[0]);
+
 		return ERR_PTR(-ENOENT);
 	}
 
 	/* Handle is imported dma-buf, so cannot be migrated to VRAM for scanout */
 	bo = gem_to_amdgpu_bo(obj);
 	domains = amdgpu_display_supported_domains(drm_to_adev(dev), bo->flags);
-	if (obj->import_attach && !(domains & AMDGPU_GEM_DOMAIN_GTT)) {
+	if (drm_gem_is_imported(obj) && !(domains & AMDGPU_GEM_DOMAIN_GTT)) {
 		drm_dbg_kms(dev, "Cannot create framebuffer from imported dma_buf\n");
 		drm_gem_object_put(obj);
 		return ERR_PTR(-EINVAL);
@@ -1209,7 +1330,7 @@ amdgpu_display_user_framebuffer_create(struct drm_device *dev,
 	}
 
 	ret = amdgpu_display_gem_fb_verify_and_init(dev, amdgpu_fb, file_priv,
-						    mode_cmd, obj);
+						    info, mode_cmd, obj);
 	if (ret) {
 		kfree(amdgpu_fb);
 		drm_gem_object_put(obj);
@@ -1222,27 +1343,84 @@ amdgpu_display_user_framebuffer_create(struct drm_device *dev,
 
 const struct drm_mode_config_funcs amdgpu_mode_funcs = {
 	.fb_create = amdgpu_display_user_framebuffer_create,
-	.output_poll_changed = drm_fb_helper_output_poll_changed,
 };
 
-static const struct drm_prop_enum_list amdgpu_underscan_enum_list[] =
-{	{ UNDERSCAN_OFF, "off" },
+static const struct drm_prop_enum_list amdgpu_underscan_enum_list[] = {
+	{ UNDERSCAN_OFF, "off" },
 	{ UNDERSCAN_ON, "on" },
 	{ UNDERSCAN_AUTO, "auto" },
 };
 
-static const struct drm_prop_enum_list amdgpu_audio_enum_list[] =
-{	{ AMDGPU_AUDIO_DISABLE, "off" },
+static const struct drm_prop_enum_list amdgpu_audio_enum_list[] = {
+	{ AMDGPU_AUDIO_DISABLE, "off" },
 	{ AMDGPU_AUDIO_ENABLE, "on" },
 	{ AMDGPU_AUDIO_AUTO, "auto" },
 };
 
 /* XXX support different dither options? spatial, temporal, both, etc. */
-static const struct drm_prop_enum_list amdgpu_dither_enum_list[] =
-{	{ AMDGPU_FMT_DITHER_DISABLE, "off" },
+static const struct drm_prop_enum_list amdgpu_dither_enum_list[] = {
+	{ AMDGPU_FMT_DITHER_DISABLE, "off" },
 	{ AMDGPU_FMT_DITHER_ENABLE, "on" },
 };
 
+/**
+ * DOC: property for adaptive backlight modulation
+ *
+ * The 'adaptive backlight modulation' property is used for the compositor to
+ * directly control the adaptive backlight modulation power savings feature
+ * that is part of DCN hardware.
+ *
+ * The property will be attached specifically to eDP panels that support it.
+ *
+ * The property is by default set to 'sysfs' to allow the sysfs file 'panel_power_savings'
+ * to be able to control it.
+ * If set to 'off' the compositor will ensure it stays off.
+ * The other values 'min', 'bias min', 'bias max', and 'max' will control the
+ * intensity of the power savings.
+ *
+ * Modifying this value can have implications on color accuracy, so tread
+ * carefully.
+ */
+static int amdgpu_display_setup_abm_prop(struct amdgpu_device *adev)
+{
+	const struct drm_prop_enum_list props[] = {
+		{ ABM_SYSFS_CONTROL, "sysfs" },
+		{ ABM_LEVEL_OFF, "off" },
+		{ ABM_LEVEL_MIN, "min" },
+		{ ABM_LEVEL_BIAS_MIN, "bias min" },
+		{ ABM_LEVEL_BIAS_MAX, "bias max" },
+		{ ABM_LEVEL_MAX, "max" },
+	};
+	struct drm_property *prop;
+	int i;
+
+	if (!adev->dc_enabled)
+		return 0;
+
+	prop = drm_property_create(adev_to_drm(adev), DRM_MODE_PROP_ENUM,
+				"adaptive backlight modulation",
+				6);
+	if (!prop)
+		return -ENOMEM;
+
+	for (i = 0; i < ARRAY_SIZE(props); i++) {
+		int ret;
+
+		ret = drm_property_add_enum(prop, props[i].type,
+						props[i].name);
+
+		if (ret) {
+			drm_property_destroy(adev_to_drm(adev), prop);
+
+			return ret;
+		}
+	}
+
+	adev->mode_info.abm_level_property = prop;
+
+	return 0;
+}
+
 int amdgpu_display_modeset_create_props(struct amdgpu_device *adev)
 {
 	int sz;
@@ -1289,15 +1467,7 @@ int amdgpu_display_modeset_create_props(struct amdgpu_device *adev)
 					 "dither",
 					 amdgpu_dither_enum_list, sz);
 
-	if (amdgpu_device_has_dc_support(adev)) {
-		adev->mode_info.abm_level_property =
-			drm_property_create_range(adev_to_drm(adev), 0,
-						  "abm level", 0, 4);
-		if (!adev->mode_info.abm_level_property)
-			return -ENOMEM;
-	}
-
-	return 0;
+	return amdgpu_display_setup_abm_prop(adev);
 }
 
 void amdgpu_display_update_priority(struct amdgpu_device *adev)
@@ -1364,7 +1534,7 @@ bool amdgpu_display_crtc_scaling_mode_fixup(struct drm_crtc *crtc,
 		if ((!(mode->flags & DRM_MODE_FLAG_INTERLACE)) &&
 		    ((amdgpu_encoder->underscan_type == UNDERSCAN_ON) ||
 		     ((amdgpu_encoder->underscan_type == UNDERSCAN_AUTO) &&
-		      drm_detect_hdmi_monitor(amdgpu_connector_edid(connector)) &&
+		      connector && connector->display_info.is_hdmi &&
 		      amdgpu_display_is_hdtv_mode(mode)))) {
 			if (amdgpu_encoder->underscan_hborder != 0)
 				amdgpu_crtc->h_border = amdgpu_encoder->underscan_hborder;
@@ -1383,6 +1553,7 @@ bool amdgpu_display_crtc_scaling_mode_fixup(struct drm_crtc *crtc,
 	}
 	if (amdgpu_crtc->rmx_type != RMX_OFF) {
 		fixed20_12 a, b;
+
 		a.full = dfixed_const(src_v);
 		b.full = dfixed_const(dst_v);
 		amdgpu_crtc->vsc.full = dfixed_div(a, b);
@@ -1402,7 +1573,7 @@ bool amdgpu_display_crtc_scaling_mode_fixup(struct drm_crtc *crtc,
  *
  * \param dev Device to query.
  * \param pipe Crtc to query.
- * \param flags Flags from caller (DRM_CALLED_FROM_VBLIRQ or 0).
+ * \param flags from caller (DRM_CALLED_FROM_VBLIRQ or 0).
  *              For driver internal use only also supports these flags:
  *
  *              USE_REAL_VBLANKSTART to use the real start of vblank instead
@@ -1469,8 +1640,7 @@ int amdgpu_display_get_crtc_scanoutpos(struct drm_device *dev,
 		ret |= DRM_SCANOUTPOS_ACCURATE;
 		vbl_start = vbl & 0x1fff;
 		vbl_end = (vbl >> 16) & 0x1fff;
-	}
-	else {
+	} else {
 		/* No: Fake something reasonable which gives at least ok results. */
 		vbl_start = mode->crtc_vdisplay;
 		vbl_end = 0;
@@ -1478,8 +1648,8 @@ int amdgpu_display_get_crtc_scanoutpos(struct drm_device *dev,
 
 	/* Called from driver internal vblank counter query code? */
 	if (flags & GET_DISTANCE_TO_VBLANKSTART) {
-	    /* Caller wants distance from real vbl_start in *hpos */
-	    *hpos = *vpos - vbl_start;
+		/* Caller wants distance from real vbl_start in *hpos */
+		*hpos = *vpos - vbl_start;
 	}
 
 	/* Fudge vblank to start a few scanlines earlier to handle the
@@ -1501,7 +1671,7 @@ int amdgpu_display_get_crtc_scanoutpos(struct drm_device *dev,
 
 	/* In vblank? */
 	if (in_vbl)
-	    ret |= DRM_SCANOUTPOS_IN_VBLANK;
+		ret |= DRM_SCANOUTPOS_IN_VBLANK;
 
 	/* Called from driver internal vblank counter query code? */
 	if (flags & GET_DISTANCE_TO_VBLANKSTART) {
@@ -1568,6 +1738,21 @@ bool amdgpu_crtc_get_scanout_position(struct drm_crtc *crtc,
 						  stime, etime, mode);
 }
 
+static bool
+amdgpu_display_robj_is_fb(struct amdgpu_device *adev, struct amdgpu_bo *robj)
+{
+	struct drm_device *dev = adev_to_drm(adev);
+	struct drm_fb_helper *fb_helper = dev->fb_helper;
+
+	if (!fb_helper || !fb_helper->buffer)
+		return false;
+
+	if (gem_to_amdgpu_bo(fb_helper->buffer->gem) != robj)
+		return false;
+
+	return true;
+}
+
 int amdgpu_display_suspend_helper(struct amdgpu_device *adev)
 {
 	struct drm_device *dev = adev_to_drm(adev);
@@ -1576,6 +1761,8 @@ int amdgpu_display_suspend_helper(struct amdgpu_device *adev)
 	struct drm_connector_list_iter iter;
 	int r;
 
+	drm_kms_helper_poll_disable(dev);
+
 	/* turn off display hw */
 	drm_modeset_lock_all(dev);
 	drm_connector_list_iter_begin(dev, &iter);
@@ -1592,6 +1779,7 @@ int amdgpu_display_suspend_helper(struct amdgpu_device *adev)
 
 		if (amdgpu_crtc->cursor_bo && !adev->enable_virtual_display) {
 			struct amdgpu_bo *aobj = gem_to_amdgpu_bo(amdgpu_crtc->cursor_bo);
+
 			r = amdgpu_bo_reserve(aobj, true);
 			if (r == 0) {
 				amdgpu_bo_unpin(aobj);
@@ -1599,12 +1787,11 @@ int amdgpu_display_suspend_helper(struct amdgpu_device *adev)
 			}
 		}
 
-		if (fb == NULL || fb->obj[0] == NULL) {
+		if (!fb || !fb->obj[0])
 			continue;
-		}
+
 		robj = gem_to_amdgpu_bo(fb->obj[0]);
-		/* don't unpin kernel fb objects */
-		if (!amdgpu_fbdev_robj_is_fb(adev, robj)) {
+		if (!amdgpu_display_robj_is_fb(adev, robj)) {
 			r = amdgpu_bo_reserve(robj, true);
 			if (r == 0) {
 				amdgpu_bo_unpin(robj);
@@ -1629,8 +1816,10 @@ int amdgpu_display_resume_helper(struct amdgpu_device *adev)
 
 		if (amdgpu_crtc->cursor_bo && !adev->enable_virtual_display) {
 			struct amdgpu_bo *aobj = gem_to_amdgpu_bo(amdgpu_crtc->cursor_bo);
+
 			r = amdgpu_bo_reserve(aobj, true);
 			if (r == 0) {
+				aobj->flags |= AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS;
 				r = amdgpu_bo_pin(aobj, AMDGPU_GEM_DOMAIN_VRAM);
 				if (r != 0)
 					dev_err(adev->dev, "Failed to pin cursor BO (%d)\n", r);
@@ -1653,6 +1842,87 @@ int amdgpu_display_resume_helper(struct amdgpu_device *adev)
 
 	drm_modeset_unlock_all(dev);
 
+	drm_kms_helper_poll_enable(dev);
+
 	return 0;
 }
 
+/* panic_bo is set in amdgpu_dm_plane_get_scanout_buffer() and only used in amdgpu_dm_set_pixel()
+ * they are called from the panic handler, and protected by the drm_panic spinlock.
+ */
+static struct amdgpu_bo *panic_abo;
+
+/* Use the indirect MMIO to write each pixel to the GPU VRAM,
+ * This is a simplified version of amdgpu_device_mm_access()
+ */
+static void amdgpu_display_set_pixel(struct drm_scanout_buffer *sb,
+				     unsigned int x,
+				     unsigned int y,
+				     u32 color)
+{
+	struct amdgpu_res_cursor cursor;
+	unsigned long offset;
+	struct amdgpu_bo *abo = panic_abo;
+	struct amdgpu_device *adev = amdgpu_ttm_adev(abo->tbo.bdev);
+	uint32_t tmp;
+
+	offset = x * 4 + y * sb->pitch[0];
+	amdgpu_res_first(abo->tbo.resource, offset, 4, &cursor);
+
+	tmp = cursor.start >> 31;
+	WREG32_NO_KIQ(mmMM_INDEX, ((uint32_t) cursor.start) | 0x80000000);
+	if (tmp != 0xffffffff)
+		WREG32_NO_KIQ(mmMM_INDEX_HI, tmp);
+	WREG32_NO_KIQ(mmMM_DATA, color);
+}
+
+int amdgpu_display_get_scanout_buffer(struct drm_plane *plane,
+				      struct drm_scanout_buffer *sb)
+{
+	struct amdgpu_bo *abo;
+	struct drm_framebuffer *fb = plane->state->fb;
+
+	if (!fb)
+		return -EINVAL;
+
+	DRM_DEBUG_KMS("Framebuffer %dx%d %p4cc\n", fb->width, fb->height, &fb->format->format);
+
+	abo = gem_to_amdgpu_bo(fb->obj[0]);
+	if (!abo)
+		return -EINVAL;
+
+	sb->width = fb->width;
+	sb->height = fb->height;
+	/* Use the generic linear format, because tiling will be disabled in panic_flush() */
+	sb->format = drm_format_info(fb->format->format);
+	if (!sb->format)
+		return -EINVAL;
+
+	sb->pitch[0] = fb->pitches[0];
+
+	if (abo->flags & AMDGPU_GEM_CREATE_NO_CPU_ACCESS) {
+		if (abo->tbo.resource->mem_type != TTM_PL_VRAM) {
+			drm_warn(plane->dev, "amdgpu panic, framebuffer not in VRAM\n");
+			return -EINVAL;
+		}
+		/* Only handle 32bits format, to simplify mmio access */
+		if (fb->format->cpp[0] != 4) {
+			drm_warn(plane->dev, "amdgpu panic, pixel format is not 32bits\n");
+			return -EINVAL;
+		}
+		sb->set_pixel = amdgpu_display_set_pixel;
+		panic_abo = abo;
+		return 0;
+	}
+	if (!abo->kmap.virtual &&
+	    ttm_bo_kmap(&abo->tbo, 0, PFN_UP(abo->tbo.base.size), &abo->kmap)) {
+		drm_warn(plane->dev, "amdgpu bo map failed, panic won't be displayed\n");
+		return -ENOMEM;
+	}
+	if (abo->kmap.bo_kmap_type & TTM_BO_MAP_IOMEM_MASK)
+		iosys_map_set_vaddr_iomem(&sb->map[0], abo->kmap.virtual);
+	else
+		iosys_map_set_vaddr(&sb->map[0], abo->kmap.virtual);
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_display.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_display.h
index 7b6d83e2b13c..49a29bf47a37 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_display.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_display.h
@@ -23,6 +23,8 @@
 #ifndef __AMDGPU_DISPLAY_H__
 #define __AMDGPU_DISPLAY_H__
 
+#include <drm/drm_panic.h>
+
 #define amdgpu_display_vblank_get_counter(adev, crtc) (adev)->mode_info.funcs->vblank_get_counter((adev), (crtc))
 #define amdgpu_display_backlight_set_level(adev, e, l) (adev)->mode_info.funcs->backlight_set_level((e), (l))
 #define amdgpu_display_backlight_get_level(adev, e) (adev)->mode_info.funcs->backlight_get_level((e))
@@ -35,14 +37,14 @@
 #define amdgpu_display_add_encoder(adev, e, s, c) (adev)->mode_info.funcs->add_encoder((adev), (e), (s), (c))
 #define amdgpu_display_add_connector(adev, ci, sd, ct, ib, coi, h, r) (adev)->mode_info.funcs->add_connector((adev), (ci), (sd), (ct), (ib), (coi), (h), (r))
 
-int amdgpu_display_freesync_ioctl(struct drm_device *dev, void *data,
-				  struct drm_file *filp);
+void amdgpu_display_hotplug_work_func(struct work_struct *work);
 void amdgpu_display_update_priority(struct amdgpu_device *adev);
 uint32_t amdgpu_display_supported_domains(struct amdgpu_device *adev,
 					  uint64_t bo_flags);
 struct drm_framebuffer *
 amdgpu_display_user_framebuffer_create(struct drm_device *dev,
 				       struct drm_file *file_priv,
+				       const struct drm_format_info *info,
 				       const struct drm_mode_fb_cmd2 *mode_cmd);
 const struct drm_format_info *
 amdgpu_lookup_format_info(u32 format, uint64_t modifier);
@@ -50,4 +52,14 @@ amdgpu_lookup_format_info(u32 format, uint64_t modifier);
 int amdgpu_display_suspend_helper(struct amdgpu_device *adev);
 int amdgpu_display_resume_helper(struct amdgpu_device *adev);
 
+int amdgpu_display_get_scanout_buffer(struct drm_plane *plane,
+				      struct drm_scanout_buffer *sb);
+
+#define ABM_SYSFS_CONTROL	-1
+#define ABM_LEVEL_OFF		0
+#define ABM_LEVEL_MIN		1
+#define ABM_LEVEL_BIAS_MIN	2
+#define ABM_LEVEL_BIAS_MAX	3
+#define ABM_LEVEL_MAX		4
+
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c
index ae6ab93c868b..e22cfa7c6d32 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c
@@ -36,11 +36,35 @@
 #include "amdgpu_gem.h"
 #include "amdgpu_dma_buf.h"
 #include "amdgpu_xgmi.h"
+#include "amdgpu_vm.h"
 #include <drm/amdgpu_drm.h>
+#include <drm/ttm/ttm_tt.h>
 #include <linux/dma-buf.h>
 #include <linux/dma-fence-array.h>
 #include <linux/pci-p2pdma.h>
-#include <linux/pm_runtime.h>
+
+static const struct dma_buf_attach_ops amdgpu_dma_buf_attach_ops;
+
+/**
+ * dma_buf_attach_adev - Helper to get adev of an attachment
+ *
+ * @attach: attachment
+ *
+ * Returns:
+ * A struct amdgpu_device * if the attaching device is an amdgpu device or
+ * partition, NULL otherwise.
+ */
+static struct amdgpu_device *dma_buf_attach_adev(struct dma_buf_attachment *attach)
+{
+	if (attach->importer_ops == &amdgpu_dma_buf_attach_ops) {
+		struct drm_gem_object *obj = attach->importer_priv;
+		struct amdgpu_bo *bo = gem_to_amdgpu_bo(obj);
+
+		return amdgpu_ttm_adev(bo->tbo.bdev);
+	}
+
+	return NULL;
+}
 
 /**
  * amdgpu_dma_buf_attach - &dma_buf_ops.attach implementation
@@ -53,45 +77,49 @@
 static int amdgpu_dma_buf_attach(struct dma_buf *dmabuf,
 				 struct dma_buf_attachment *attach)
 {
+	struct amdgpu_device *attach_adev = dma_buf_attach_adev(attach);
 	struct drm_gem_object *obj = dmabuf->priv;
 	struct amdgpu_bo *bo = gem_to_amdgpu_bo(obj);
 	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
 	int r;
 
-	if (pci_p2pdma_distance_many(adev->pdev, &attach->dev, 1, true) < 0)
+	/*
+	 * Disable peer-to-peer access for DCC-enabled VRAM surfaces on GFX12+.
+	 * Such buffers cannot be safely accessed over P2P due to device-local
+	 * compression metadata. Fallback to system-memory path instead.
+	 * Device supports GFX12 (GC 12.x or newer)
+	 * BO was created with the AMDGPU_GEM_CREATE_GFX12_DCC flag
+	 *
+	 */
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) >= IP_VERSION(12, 0, 0) &&
+	    bo->flags & AMDGPU_GEM_CREATE_GFX12_DCC)
 		attach->peer2peer = false;
 
-	if (attach->dev->driver == adev->dev->driver)
-		return 0;
+	/*
+	 * Disable peer-to-peer access for DCC-enabled VRAM surfaces on GFX12+.
+	 * Such buffers cannot be safely accessed over P2P due to device-local
+	 * compression metadata. Fallback to system-memory path instead.
+	 * Device supports GFX12 (GC 12.x or newer)
+	 * BO was created with the AMDGPU_GEM_CREATE_GFX12_DCC flag
+	 *
+	 */
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) >= IP_VERSION(12, 0, 0) &&
+	    bo->flags & AMDGPU_GEM_CREATE_GFX12_DCC)
+		attach->peer2peer = false;
 
-	r = pm_runtime_get_sync(adev_to_drm(adev)->dev);
-	if (r < 0)
-		goto out;
+	if (!amdgpu_dmabuf_is_xgmi_accessible(attach_adev, bo) &&
+	    pci_p2pdma_distance(adev->pdev, attach->dev, false) < 0)
+		attach->peer2peer = false;
 
-	return 0;
+	r = dma_resv_lock(bo->tbo.base.resv, NULL);
+	if (r)
+		return r;
 
-out:
-	pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
-	return r;
-}
+	amdgpu_vm_bo_update_shared(bo);
 
-/**
- * amdgpu_dma_buf_detach - &dma_buf_ops.detach implementation
- *
- * @dmabuf: DMA-buf where we remove the attachment from
- * @attach: the attachment to remove
- *
- * Called when an attachment is removed from the DMA-buf.
- */
-static void amdgpu_dma_buf_detach(struct dma_buf *dmabuf,
-				  struct dma_buf_attachment *attach)
-{
-	struct drm_gem_object *obj = dmabuf->priv;
-	struct amdgpu_bo *bo = gem_to_amdgpu_bo(obj);
-	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
+	dma_resv_unlock(bo->tbo.base.resv);
 
-	pm_runtime_mark_last_busy(adev_to_drm(adev)->dev);
-	pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
+	return 0;
 }
 
 /**
@@ -103,23 +131,35 @@ static void amdgpu_dma_buf_detach(struct dma_buf *dmabuf,
  */
 static int amdgpu_dma_buf_pin(struct dma_buf_attachment *attach)
 {
-	struct drm_gem_object *obj = attach->dmabuf->priv;
-	struct amdgpu_bo *bo = gem_to_amdgpu_bo(obj);
-	int r;
+	struct dma_buf *dmabuf = attach->dmabuf;
+	struct amdgpu_bo *bo = gem_to_amdgpu_bo(dmabuf->priv);
+	u32 domains = bo->allowed_domains;
+
+	dma_resv_assert_held(dmabuf->resv);
+
+	/* Try pinning into VRAM to allow P2P with RDMA NICs without ODP
+	 * support if all attachments can do P2P. If any attachment can't do
+	 * P2P just pin into GTT instead.
+	 *
+	 * To avoid with conflicting pinnings between GPUs and RDMA when move
+	 * notifiers are disabled, only allow pinning in VRAM when move
+	 * notiers are enabled.
+	 */
+	if (!IS_ENABLED(CONFIG_DMABUF_MOVE_NOTIFY)) {
+		domains &= ~AMDGPU_GEM_DOMAIN_VRAM;
+	} else {
+		list_for_each_entry(attach, &dmabuf->attachments, node)
+			if (!attach->peer2peer)
+				domains &= ~AMDGPU_GEM_DOMAIN_VRAM;
+	}
 
-	/* pin buffer into GTT */
-	r = amdgpu_bo_pin(bo, AMDGPU_GEM_DOMAIN_GTT);
-	if (r)
-		return r;
+	if (domains & AMDGPU_GEM_DOMAIN_VRAM)
+		bo->flags |= AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED;
 
-	if (bo->tbo.moving) {
-		r = dma_fence_wait(bo->tbo.moving, true);
-		if (r) {
-			amdgpu_bo_unpin(bo);
-			return r;
-		}
-	}
-	return 0;
+	if (WARN_ON(!domains))
+		return -EINVAL;
+
+	return amdgpu_bo_pin(bo, domains);
 }
 
 /**
@@ -163,7 +203,7 @@ static struct sg_table *amdgpu_dma_buf_map(struct dma_buf_attachment *attach,
 	if (!bo->tbo.pin_count) {
 		/* move buffer into GTT or VRAM */
 		struct ttm_operation_ctx ctx = { false, false };
-		unsigned domains = AMDGPU_GEM_DOMAIN_GTT;
+		unsigned int domains = AMDGPU_GEM_DOMAIN_GTT;
 
 		if (bo->preferred_domains & AMDGPU_GEM_DOMAIN_VRAM &&
 		    attach->peer2peer) {
@@ -174,10 +214,6 @@ static struct sg_table *amdgpu_dma_buf_map(struct dma_buf_attachment *attach,
 		r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
 		if (r)
 			return ERR_PTR(r);
-
-	} else if (!(amdgpu_mem_type_to_domain(bo->tbo.resource->mem_type) &
-		     AMDGPU_GEM_DOMAIN_GTT)) {
-		return ERR_PTR(-EBUSY);
 	}
 
 	switch (bo->tbo.resource->mem_type) {
@@ -194,6 +230,11 @@ static struct sg_table *amdgpu_dma_buf_map(struct dma_buf_attachment *attach,
 		break;
 
 	case TTM_PL_VRAM:
+		/* XGMI-accessible memory should never be DMA-mapped */
+		if (WARN_ON(amdgpu_dmabuf_is_xgmi_accessible(
+				dma_buf_attach_adev(attach), bo)))
+			return ERR_PTR(-EINVAL);
+
 		r = amdgpu_vram_mgr_alloc_sgt(adev, bo->tbo.resource, 0,
 					      bo->tbo.base.size, attach->dev,
 					      dir, &sgt);
@@ -225,7 +266,7 @@ static void amdgpu_dma_buf_unmap(struct dma_buf_attachment *attach,
 				 struct sg_table *sgt,
 				 enum dma_data_direction dir)
 {
-	if (sgt->sgl->page_link) {
+	if (sg_page(sgt->sgl)) {
 		dma_unmap_sgtable(attach->dev, sgt, dir, 0);
 		sg_free_table(sgt);
 		kfree(sgt);
@@ -275,9 +316,38 @@ static int amdgpu_dma_buf_begin_cpu_access(struct dma_buf *dma_buf,
 	return ret;
 }
 
+static int amdgpu_dma_buf_vmap(struct dma_buf *dma_buf, struct iosys_map *map)
+{
+	struct drm_gem_object *obj = dma_buf->priv;
+	struct amdgpu_bo *bo = gem_to_amdgpu_bo(obj);
+	int ret;
+
+	/*
+	 * Pin to keep buffer in place while it's vmap'ed. The actual
+	 * domain is not that important as long as it's mapable. Using
+	 * GTT and VRAM should be compatible with most use cases.
+	 */
+	ret = amdgpu_bo_pin(bo, AMDGPU_GEM_DOMAIN_GTT | AMDGPU_GEM_DOMAIN_VRAM);
+	if (ret)
+		return ret;
+	ret = drm_gem_dmabuf_vmap(dma_buf, map);
+	if (ret)
+		amdgpu_bo_unpin(bo);
+
+	return ret;
+}
+
+static void amdgpu_dma_buf_vunmap(struct dma_buf *dma_buf, struct iosys_map *map)
+{
+	struct drm_gem_object *obj = dma_buf->priv;
+	struct amdgpu_bo *bo = gem_to_amdgpu_bo(obj);
+
+	drm_gem_dmabuf_vunmap(dma_buf, map);
+	amdgpu_bo_unpin(bo);
+}
+
 const struct dma_buf_ops amdgpu_dmabuf_ops = {
 	.attach = amdgpu_dma_buf_attach,
-	.detach = amdgpu_dma_buf_detach,
 	.pin = amdgpu_dma_buf_pin,
 	.unpin = amdgpu_dma_buf_unpin,
 	.map_dma_buf = amdgpu_dma_buf_map,
@@ -285,8 +355,8 @@ const struct dma_buf_ops amdgpu_dmabuf_ops = {
 	.release = drm_gem_dmabuf_release,
 	.begin_cpu_access = amdgpu_dma_buf_begin_cpu_access,
 	.mmap = drm_gem_dmabuf_mmap,
-	.vmap = drm_gem_dmabuf_vmap,
-	.vunmap = drm_gem_dmabuf_vunmap,
+	.vmap = amdgpu_dma_buf_vmap,
+	.vunmap = amdgpu_dma_buf_vunmap,
 };
 
 /**
@@ -304,11 +374,23 @@ struct dma_buf *amdgpu_gem_prime_export(struct drm_gem_object *gobj,
 {
 	struct amdgpu_bo *bo = gem_to_amdgpu_bo(gobj);
 	struct dma_buf *buf;
+	struct ttm_operation_ctx ctx = {
+		.interruptible = true,
+		.no_wait_gpu = true,
+		/* We opt to avoid OOM on system pages allocations */
+		.gfp_retry_mayfail = true,
+		.allow_res_evict = false,
+	};
+	int ret;
 
 	if (amdgpu_ttm_tt_get_usermm(bo->tbo.ttm) ||
 	    bo->flags & AMDGPU_GEM_CREATE_VM_ALWAYS_VALID)
 		return ERR_PTR(-EPERM);
 
+	ret = ttm_bo_setup_export(&bo->tbo, &ctx);
+	if (ret)
+		return ERR_PTR(ret);
+
 	buf = drm_gem_prime_export(gobj, flags);
 	if (!IS_ERR(buf))
 		buf->ops = &amdgpu_dmabuf_ops;
@@ -343,12 +425,15 @@ amdgpu_dma_buf_create_obj(struct drm_device *dev, struct dma_buf *dma_buf)
 	if (dma_buf->ops == &amdgpu_dmabuf_ops) {
 		struct amdgpu_bo *other = gem_to_amdgpu_bo(dma_buf->priv);
 
-		flags |= other->flags & AMDGPU_GEM_CREATE_CPU_GTT_USWC;
+		flags |= other->flags & (AMDGPU_GEM_CREATE_CPU_GTT_USWC |
+					 AMDGPU_GEM_CREATE_COHERENT |
+					 AMDGPU_GEM_CREATE_EXT_COHERENT |
+					 AMDGPU_GEM_CREATE_UNCACHED);
 	}
 
 	ret = amdgpu_gem_object_create(adev, dma_buf->size, PAGE_SIZE,
 				       AMDGPU_GEM_DOMAIN_CPU, flags,
-				       ttm_bo_type_sg, resv, &gobj);
+				       ttm_bo_type_sg, resv, &gobj, 0);
 	if (ret)
 		goto error;
 
@@ -384,7 +469,11 @@ amdgpu_dma_buf_move_notify(struct dma_buf_attachment *attach)
 	struct amdgpu_vm_bo_base *bo_base;
 	int r;
 
-	if (bo->tbo.resource->mem_type == TTM_PL_SYSTEM)
+	/* FIXME: This should be after the "if", but needs a fix to make sure
+	 * DMABuf imports are initialized in the right VM list.
+	 */
+	amdgpu_vm_bo_invalidate(bo, false);
+	if (!bo->tbo.resource || bo->tbo.resource->mem_type == TTM_PL_SYSTEM)
 		return;
 
 	r = ttm_bo_validate(&bo->tbo, &placement, &ctx);
@@ -415,9 +504,12 @@ amdgpu_dma_buf_move_notify(struct dma_buf_attachment *attach)
 				continue;
 		}
 
-		r = amdgpu_vm_clear_freed(adev, vm, NULL);
+		/* Reserve fences for two SDMA page table updates */
+		r = dma_resv_reserve_fences(resv, 2);
 		if (!r)
-			r = amdgpu_vm_handle_moved(adev, vm);
+			r = amdgpu_vm_clear_freed(adev, vm, NULL);
+		if (!r)
+			r = amdgpu_vm_handle_moved(adev, vm, ticket);
 
 		if (r && r != -EBUSY)
 			DRM_ERROR("Failed to invalidate VM page tables (%d))\n",
@@ -491,7 +583,10 @@ bool amdgpu_dmabuf_is_xgmi_accessible(struct amdgpu_device *adev,
 	struct drm_gem_object *obj = &bo->tbo.base;
 	struct drm_gem_object *gobj;
 
-	if (obj->import_attach) {
+	if (!adev)
+		return false;
+
+	if (drm_gem_is_imported(obj)) {
 		struct dma_buf *dma_buf = obj->import_attach->dmabuf;
 
 		if (dma_buf->ops != &amdgpu_dmabuf_ops)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h
index 89e6ad30396f..2675689ef70f 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h
@@ -21,6 +21,9 @@
  *
  */
 
+#ifndef AMDGPU_DOORBELL_H
+#define AMDGPU_DOORBELL_H
+
 /*
  * GPU doorbell structures, functions & helpers
  */
@@ -28,8 +31,15 @@ struct amdgpu_doorbell {
 	/* doorbell mmio */
 	resource_size_t		base;
 	resource_size_t		size;
-	u32 __iomem		*ptr;
-	u32			num_doorbells;	/* Number of doorbells actually reserved for amdgpu. */
+
+	/* Number of doorbells reserved for amdgpu kernel driver */
+	u32 num_kernel_doorbells;
+
+	/* Kernel doorbells */
+	struct amdgpu_bo *kernel_doorbells;
+
+	/* For CPU access of doorbells */
+	uint32_t *cpu_addr;
 };
 
 /* Reserved doorbells for amdgpu (including multimedia).
@@ -52,8 +62,11 @@ struct amdgpu_doorbell_index {
 	uint32_t userqueue_end;
 	uint32_t gfx_ring0;
 	uint32_t gfx_ring1;
-	uint32_t sdma_engine[8];
-	uint32_t mes_ring;
+	uint32_t gfx_userqueue_start;
+	uint32_t gfx_userqueue_end;
+	uint32_t sdma_engine[16];
+	uint32_t mes_ring0;
+	uint32_t mes_ring1;
 	uint32_t ih;
 	union {
 		struct {
@@ -73,15 +86,17 @@ struct amdgpu_doorbell_index {
 			uint32_t vce_ring6_7;
 		} uvd_vce;
 	};
+	uint32_t vpe_ring;
 	uint32_t first_non_cp;
 	uint32_t last_non_cp;
 	uint32_t max_assignment;
 	/* Per engine SDMA doorbell size in dword */
 	uint32_t sdma_doorbell_range;
+	/* Per xcc doorbell size for KIQ/KCQ */
+	uint32_t xcc_doorbell_range;
 };
 
-typedef enum _AMDGPU_DOORBELL_ASSIGNMENT
-{
+enum AMDGPU_DOORBELL_ASSIGNMENT {
 	AMDGPU_DOORBELL_KIQ                     = 0x000,
 	AMDGPU_DOORBELL_HIQ                     = 0x001,
 	AMDGPU_DOORBELL_DIQ                     = 0x002,
@@ -99,10 +114,10 @@ typedef enum _AMDGPU_DOORBELL_ASSIGNMENT
 	AMDGPU_DOORBELL_IH                      = 0x1E8,
 	AMDGPU_DOORBELL_MAX_ASSIGNMENT          = 0x3FF,
 	AMDGPU_DOORBELL_INVALID                 = 0xFFFF
-} AMDGPU_DOORBELL_ASSIGNMENT;
+};
+
+enum AMDGPU_VEGA20_DOORBELL_ASSIGNMENT {
 
-typedef enum _AMDGPU_VEGA20_DOORBELL_ASSIGNMENT
-{
 	/* Compute + GFX: 0~255 */
 	AMDGPU_VEGA20_DOORBELL_KIQ                     = 0x000,
 	AMDGPU_VEGA20_DOORBELL_HIQ                     = 0x001,
@@ -156,12 +171,20 @@ typedef enum _AMDGPU_VEGA20_DOORBELL_ASSIGNMENT
 	AMDGPU_VEGA20_DOORBELL64_FIRST_NON_CP            = AMDGPU_VEGA20_DOORBELL_sDMA_ENGINE0,
 	AMDGPU_VEGA20_DOORBELL64_LAST_NON_CP             = AMDGPU_VEGA20_DOORBELL64_VCE_RING6_7,
 
-	AMDGPU_VEGA20_DOORBELL_MAX_ASSIGNMENT            = 0x18F,
+	/* kiq/kcq from second XCD. Max 8 XCDs */
+	AMDGPU_VEGA20_DOORBELL_XCC1_KIQ_START             = 0x190,
+	/* 8 compute rings per GC. Max to 0x1CE */
+	AMDGPU_VEGA20_DOORBELL_XCC1_MEC_RING0_START       = 0x197,
+
+	/* AID1 SDMA: 0x1D0 ~ 0x1F7 */
+	AMDGPU_VEGA20_DOORBELL_AID1_sDMA_START           = 0x1D0,
+
+	AMDGPU_VEGA20_DOORBELL_MAX_ASSIGNMENT            = 0x1F7,
 	AMDGPU_VEGA20_DOORBELL_INVALID                   = 0xFFFF
-} AMDGPU_VEGA20_DOORBELL_ASSIGNMENT;
+};
+
+enum AMDGPU_NAVI10_DOORBELL_ASSIGNMENT {
 
-typedef enum _AMDGPU_NAVI10_DOORBELL_ASSIGNMENT
-{
 	/* Compute + GFX: 0~255 */
 	AMDGPU_NAVI10_DOORBELL_KIQ			= 0x000,
 	AMDGPU_NAVI10_DOORBELL_HIQ			= 0x001,
@@ -174,11 +197,15 @@ typedef enum _AMDGPU_NAVI10_DOORBELL_ASSIGNMENT
 	AMDGPU_NAVI10_DOORBELL_MEC_RING5		= 0x008,
 	AMDGPU_NAVI10_DOORBELL_MEC_RING6		= 0x009,
 	AMDGPU_NAVI10_DOORBELL_MEC_RING7		= 0x00A,
-	AMDGPU_NAVI10_DOORBELL_USERQUEUE_START		= 0x00B,
+	AMDGPU_NAVI10_DOORBELL_MES_RING0	        = 0x00B,
+	AMDGPU_NAVI10_DOORBELL_MES_RING1		= 0x00C,
+	AMDGPU_NAVI10_DOORBELL_USERQUEUE_START		= 0x00D,
 	AMDGPU_NAVI10_DOORBELL_USERQUEUE_END		= 0x08A,
 	AMDGPU_NAVI10_DOORBELL_GFX_RING0		= 0x08B,
 	AMDGPU_NAVI10_DOORBELL_GFX_RING1		= 0x08C,
-	AMDGPU_NAVI10_DOORBELL_MES_RING		        = 0x090,
+	AMDGPU_NAVI10_DOORBELL_GFX_USERQUEUE_START	= 0x08D,
+	AMDGPU_NAVI10_DOORBELL_GFX_USERQUEUE_END	= 0x0FF,
+
 	/* SDMA:256~335*/
 	AMDGPU_NAVI10_DOORBELL_sDMA_ENGINE0		= 0x100,
 	AMDGPU_NAVI10_DOORBELL_sDMA_ENGINE1		= 0x10A,
@@ -200,18 +227,19 @@ typedef enum _AMDGPU_NAVI10_DOORBELL_ASSIGNMENT
 	AMDGPU_NAVI10_DOORBELL64_VCNc_d			= 0x18E,
 	AMDGPU_NAVI10_DOORBELL64_VCNe_f			= 0x18F,
 
+	AMDGPU_NAVI10_DOORBELL64_VPE			= 0x190,
+
 	AMDGPU_NAVI10_DOORBELL64_FIRST_NON_CP		= AMDGPU_NAVI10_DOORBELL_sDMA_ENGINE0,
-	AMDGPU_NAVI10_DOORBELL64_LAST_NON_CP		= AMDGPU_NAVI10_DOORBELL64_VCNe_f,
+	AMDGPU_NAVI10_DOORBELL64_LAST_NON_CP		= AMDGPU_NAVI10_DOORBELL64_VPE,
 
-	AMDGPU_NAVI10_DOORBELL_MAX_ASSIGNMENT		= 0x18F,
+	AMDGPU_NAVI10_DOORBELL_MAX_ASSIGNMENT		= AMDGPU_NAVI10_DOORBELL64_VPE,
 	AMDGPU_NAVI10_DOORBELL_INVALID			= 0xFFFF
-} AMDGPU_NAVI10_DOORBELL_ASSIGNMENT;
+};
 
 /*
  * 64bit doorbell, offset are in QWORD, occupy 2KB doorbell space
  */
-typedef enum _AMDGPU_DOORBELL64_ASSIGNMENT
-{
+enum AMDGPU_DOORBELL64_ASSIGNMENT {
 	/*
 	 * All compute related doorbells: kiq, hiq, diq, traditional compute queue, user queue, should locate in
 	 * a continues range so that programming CP_MEC_DOORBELL_RANGE_LOWER/UPPER can cover this range.
@@ -287,15 +315,58 @@ typedef enum _AMDGPU_DOORBELL64_ASSIGNMENT
 
 	AMDGPU_DOORBELL64_MAX_ASSIGNMENT          = 0xFF,
 	AMDGPU_DOORBELL64_INVALID                 = 0xFFFF
-} AMDGPU_DOORBELL64_ASSIGNMENT;
+};
+
+enum AMDGPU_DOORBELL_ASSIGNMENT_LAYOUT1 {
+
+	/* XCC0: 0x00 ~20, XCC1: 20 ~ 2F ... */
+
+	/* KIQ/HIQ/DIQ */
+	AMDGPU_DOORBELL_LAYOUT1_KIQ_START		= 0x000,
+	AMDGPU_DOORBELL_LAYOUT1_HIQ			= 0x001,
+	AMDGPU_DOORBELL_LAYOUT1_DIQ			= 0x002,
+	/* Compute: 0x08 ~ 0x20  */
+	AMDGPU_DOORBELL_LAYOUT1_MEC_RING_START		= 0x008,
+	AMDGPU_DOORBELL_LAYOUT1_MEC_RING_END		= 0x00F,
+	AMDGPU_DOORBELL_LAYOUT1_USERQUEUE_START		= 0x010,
+	AMDGPU_DOORBELL_LAYOUT1_USERQUEUE_END		= 0x01F,
+	AMDGPU_DOORBELL_LAYOUT1_XCC_RANGE		= 0x020,
+
+	/* SDMA: 0x100 ~ 0x19F */
+	AMDGPU_DOORBELL_LAYOUT1_sDMA_ENGINE_START	= 0x100,
+	AMDGPU_DOORBELL_LAYOUT1_sDMA_ENGINE_END		= 0x19F,
+	/* IH: 0x1A0 ~ 0x1AF */
+	AMDGPU_DOORBELL_LAYOUT1_IH                      = 0x1A0,
+	/* VCN: 0x1B0 ~ 0x1E8 */
+	AMDGPU_DOORBELL_LAYOUT1_VCN_START               = 0x1B0,
+	AMDGPU_DOORBELL_LAYOUT1_VCN_END                 = 0x1E8,
+
+	AMDGPU_DOORBELL_LAYOUT1_FIRST_NON_CP		= AMDGPU_DOORBELL_LAYOUT1_sDMA_ENGINE_START,
+	AMDGPU_DOORBELL_LAYOUT1_LAST_NON_CP		= AMDGPU_DOORBELL_LAYOUT1_VCN_END,
+
+	AMDGPU_DOORBELL_LAYOUT1_MAX_ASSIGNMENT          = 0x1E8,
+	AMDGPU_DOORBELL_LAYOUT1_INVALID                 = 0xFFFF
+};
 
 u32 amdgpu_mm_rdoorbell(struct amdgpu_device *adev, u32 index);
 void amdgpu_mm_wdoorbell(struct amdgpu_device *adev, u32 index, u32 v);
 u64 amdgpu_mm_rdoorbell64(struct amdgpu_device *adev, u32 index);
 void amdgpu_mm_wdoorbell64(struct amdgpu_device *adev, u32 index, u64 v);
 
+/*
+ * GPU doorbell aperture helpers function.
+ */
+int amdgpu_doorbell_init(struct amdgpu_device *adev);
+void amdgpu_doorbell_fini(struct amdgpu_device *adev);
+int amdgpu_doorbell_create_kernel_doorbells(struct amdgpu_device *adev);
+uint32_t amdgpu_doorbell_index_on_bar(struct amdgpu_device *adev,
+				      struct amdgpu_bo *db_bo,
+				      uint32_t doorbell_index,
+				      uint32_t db_size);
+
 #define RDOORBELL32(index) amdgpu_mm_rdoorbell(adev, (index))
 #define WDOORBELL32(index, v) amdgpu_mm_wdoorbell(adev, (index), (v))
 #define RDOORBELL64(index) amdgpu_mm_rdoorbell64(adev, (index))
 #define WDOORBELL64(index, v) amdgpu_mm_wdoorbell64(adev, (index), (v))
 
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell_mgr.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell_mgr.c
new file mode 100644
index 000000000000..3040437d99c2
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell_mgr.c
@@ -0,0 +1,244 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "amdgpu.h"
+
+/**
+ * amdgpu_mm_rdoorbell - read a doorbell dword
+ *
+ * @adev: amdgpu_device pointer
+ * @index: doorbell index
+ *
+ * Returns the value in the doorbell aperture at the
+ * requested doorbell index (CIK).
+ */
+u32 amdgpu_mm_rdoorbell(struct amdgpu_device *adev, u32 index)
+{
+	if (amdgpu_device_skip_hw_access(adev))
+		return 0;
+
+	if (index < adev->doorbell.num_kernel_doorbells)
+		return readl(adev->doorbell.cpu_addr + index);
+
+	dev_err(adev->dev, "reading beyond doorbell aperture: 0x%08x!\n",
+		index);
+	return 0;
+}
+
+/**
+ * amdgpu_mm_wdoorbell - write a doorbell dword
+ *
+ * @adev: amdgpu_device pointer
+ * @index: doorbell index
+ * @v: value to write
+ *
+ * Writes @v to the doorbell aperture at the
+ * requested doorbell index (CIK).
+ */
+void amdgpu_mm_wdoorbell(struct amdgpu_device *adev, u32 index, u32 v)
+{
+	if (amdgpu_device_skip_hw_access(adev))
+		return;
+
+	if (index < adev->doorbell.num_kernel_doorbells)
+		writel(v, adev->doorbell.cpu_addr + index);
+	else
+		dev_err(adev->dev,
+			"writing beyond doorbell aperture: 0x%08x!\n", index);
+}
+
+/**
+ * amdgpu_mm_rdoorbell64 - read a doorbell Qword
+ *
+ * @adev: amdgpu_device pointer
+ * @index: doorbell index
+ *
+ * Returns the value in the doorbell aperture at the
+ * requested doorbell index (VEGA10+).
+ */
+u64 amdgpu_mm_rdoorbell64(struct amdgpu_device *adev, u32 index)
+{
+	if (amdgpu_device_skip_hw_access(adev))
+		return 0;
+
+	if (index < adev->doorbell.num_kernel_doorbells)
+		return atomic64_read((atomic64_t *)(adev->doorbell.cpu_addr + index));
+
+	dev_err(adev->dev, "reading beyond doorbell aperture: 0x%08x!\n",
+		index);
+	return 0;
+}
+
+/**
+ * amdgpu_mm_wdoorbell64 - write a doorbell Qword
+ *
+ * @adev: amdgpu_device pointer
+ * @index: doorbell index
+ * @v: value to write
+ *
+ * Writes @v to the doorbell aperture at the
+ * requested doorbell index (VEGA10+).
+ */
+void amdgpu_mm_wdoorbell64(struct amdgpu_device *adev, u32 index, u64 v)
+{
+	if (amdgpu_device_skip_hw_access(adev))
+		return;
+
+	if (index < adev->doorbell.num_kernel_doorbells)
+		atomic64_set((atomic64_t *)(adev->doorbell.cpu_addr + index), v);
+	else
+		dev_err(adev->dev,
+			"writing beyond doorbell aperture: 0x%08x!\n", index);
+}
+
+/**
+ * amdgpu_doorbell_index_on_bar - Find doorbell's absolute offset in BAR
+ *
+ * @adev: amdgpu_device pointer
+ * @db_bo: doorbell object's bo
+ * @doorbell_index: doorbell relative index in this doorbell object
+ * @db_size: doorbell size is in byte
+ *
+ * returns doorbell's absolute index in BAR
+ */
+uint32_t amdgpu_doorbell_index_on_bar(struct amdgpu_device *adev,
+				      struct amdgpu_bo *db_bo,
+				      uint32_t doorbell_index,
+				      uint32_t db_size)
+{
+	int db_bo_offset;
+
+	db_bo_offset = amdgpu_bo_gpu_offset_no_check(db_bo);
+
+	/* doorbell index is 32 bit but doorbell's size can be 32 bit
+	 * or 64 bit, so *db_size(in byte)/4 for alignment.
+	 */
+	return db_bo_offset / sizeof(u32) + doorbell_index *
+	       DIV_ROUND_UP(db_size, 4);
+}
+
+/**
+ * amdgpu_doorbell_create_kernel_doorbells - Create kernel doorbells for graphics
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Creates doorbells for graphics driver usages.
+ * returns 0 on success, error otherwise.
+ */
+int amdgpu_doorbell_create_kernel_doorbells(struct amdgpu_device *adev)
+{
+	int r;
+	int size;
+
+	/* SI HW does not have doorbells, skip allocation */
+	if (adev->doorbell.num_kernel_doorbells == 0)
+		return 0;
+
+	/* Reserve first num_kernel_doorbells (page-aligned) for kernel ops */
+	size = ALIGN(adev->doorbell.num_kernel_doorbells * sizeof(u32), PAGE_SIZE);
+
+	/* Allocate an extra page for MES kernel usages (ring test) */
+	adev->mes.db_start_dw_offset = size / sizeof(u32);
+	size += PAGE_SIZE;
+
+	r = amdgpu_bo_create_kernel(adev,
+				    size,
+				    PAGE_SIZE,
+				    AMDGPU_GEM_DOMAIN_DOORBELL,
+				    &adev->doorbell.kernel_doorbells,
+				    NULL,
+				    (void **)&adev->doorbell.cpu_addr);
+	if (r) {
+		dev_err(adev->dev,
+			"Failed to allocate kernel doorbells, err=%d\n", r);
+		return r;
+	}
+
+	adev->doorbell.num_kernel_doorbells = size / sizeof(u32);
+	return 0;
+}
+
+/*
+ * GPU doorbell aperture helpers function.
+ */
+/**
+ * amdgpu_doorbell_init - Init doorbell driver information.
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Init doorbell driver information (CIK)
+ * Returns 0 on success, error on failure.
+ */
+int amdgpu_doorbell_init(struct amdgpu_device *adev)
+{
+
+	/* No doorbell on SI hardware generation */
+	if (adev->asic_type < CHIP_BONAIRE) {
+		adev->doorbell.base = 0;
+		adev->doorbell.size = 0;
+		adev->doorbell.num_kernel_doorbells = 0;
+		return 0;
+	}
+
+	if (pci_resource_flags(adev->pdev, 2) & IORESOURCE_UNSET)
+		return -EINVAL;
+
+	amdgpu_asic_init_doorbell_index(adev);
+
+	/* doorbell bar mapping */
+	adev->doorbell.base = pci_resource_start(adev->pdev, 2);
+	adev->doorbell.size = pci_resource_len(adev->pdev, 2);
+
+	adev->doorbell.num_kernel_doorbells =
+		min_t(u32, adev->doorbell.size / sizeof(u32),
+		      adev->doorbell_index.max_assignment + 1);
+	if (adev->doorbell.num_kernel_doorbells == 0)
+		return -EINVAL;
+
+	/*
+	 * For Vega, reserve and map two pages on doorbell BAR since SDMA
+	 * paging queue doorbell use the second page. The
+	 * AMDGPU_DOORBELL64_MAX_ASSIGNMENT definition assumes all the
+	 * doorbells are in the first page. So with paging queue enabled,
+	 * the max num_kernel_doorbells should + 1 page (0x400 in dword)
+	 */
+	if (adev->asic_type >= CHIP_VEGA10)
+		adev->doorbell.num_kernel_doorbells += 0x400;
+
+	return 0;
+}
+
+/**
+ * amdgpu_doorbell_fini - Tear down doorbell driver information.
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Tear down doorbell driver information (CIK)
+ */
+void amdgpu_doorbell_fini(struct amdgpu_device *adev)
+{
+	amdgpu_bo_free_kernel(&adev->doorbell.kernel_doorbells,
+			      NULL,
+			      (void **)&adev->doorbell.cpu_addr);
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
index ad95de6399af..2dfbddcef9ab 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
@@ -23,33 +23,37 @@
  */
 
 #include <drm/amdgpu_drm.h>
-#include <drm/drm_aperture.h>
+#include <drm/clients/drm_client_setup.h>
 #include <drm/drm_drv.h>
+#include <drm/drm_fbdev_ttm.h>
 #include <drm/drm_gem.h>
-#include <drm/drm_vblank.h>
 #include <drm/drm_managed.h>
-#include "amdgpu_drv.h"
-
 #include <drm/drm_pciids.h>
-#include <linux/console.h>
-#include <linux/module.h>
-#include <linux/pm_runtime.h>
-#include <linux/vga_switcheroo.h>
 #include <drm/drm_probe_helper.h>
+#include <drm/drm_vblank.h>
+
+#include <linux/cc_platform.h>
+#include <linux/dynamic_debug.h>
+#include <linux/module.h>
 #include <linux/mmu_notifier.h>
+#include <linux/pm_runtime.h>
 #include <linux/suspend.h>
-#include <linux/cc_platform.h>
+#include <linux/vga_switcheroo.h>
 
 #include "amdgpu.h"
-#include "amdgpu_irq.h"
+#include "amdgpu_amdkfd.h"
 #include "amdgpu_dma_buf.h"
-#include "amdgpu_sched.h"
+#include "amdgpu_drv.h"
 #include "amdgpu_fdinfo.h"
-#include "amdgpu_amdkfd.h"
-
+#include "amdgpu_irq.h"
+#include "amdgpu_psp.h"
 #include "amdgpu_ras.h"
-#include "amdgpu_xgmi.h"
 #include "amdgpu_reset.h"
+#include "amdgpu_sched.h"
+#include "amdgpu_xgmi.h"
+#include "amdgpu_userq.h"
+#include "amdgpu_userq_fence.h"
+#include "../amdxcp/amdgpu_xcp_drv.h"
 
 /*
  * KMS wrapper.
@@ -81,7 +85,7 @@
  * - 3.24.0 - Add high priority compute support for gfx9
  * - 3.25.0 - Add support for sensor query info (stable pstate sclk/mclk).
  * - 3.26.0 - GFX9: Process AMDGPU_IB_FLAG_TC_WB_NOT_INVALIDATE.
- * - 3.27.0 - Add new chunk to to AMDGPU_CS to enable BO_LIST creation.
+ * - 3.27.0 - Add new chunk to AMDGPU_CS to enable BO_LIST creation.
  * - 3.28.0 - Add AMDGPU_CHUNK_ID_SCHEDULED_DEPENDENCIES
  * - 3.29.0 - Add AMDGPU_IB_FLAG_RESET_GDS_MAX_WAVE_ID
  * - 3.30.0 - Add AMDGPU_SCHED_OP_CONTEXT_PRIORITY_OVERRIDE.
@@ -99,18 +103,56 @@
  * - 3.42.0 - Add 16bpc fixed point display support
  * - 3.43.0 - Add device hot plug/unplug support
  * - 3.44.0 - DCN3 supports DCC independent block settings: !64B && 128B, 64B && 128B
+ * - 3.45.0 - Add context ioctl stable pstate interface
+ * - 3.46.0 - To enable hot plug amdgpu tests in libdrm
+ * - 3.47.0 - Add AMDGPU_GEM_CREATE_DISCARDABLE and AMDGPU_VM_NOALLOC flags
+ * - 3.48.0 - Add IP discovery version info to HW INFO
+ * - 3.49.0 - Add gang submit into CS IOCTL
+ * - 3.50.0 - Update AMDGPU_INFO_DEV_INFO IOCTL for minimum engine and memory clock
+ *            Update AMDGPU_INFO_SENSOR IOCTL for PEAK_PSTATE engine and memory clock
+ *   3.51.0 - Return the PCIe gen and lanes from the INFO ioctl
+ *   3.52.0 - Add AMDGPU_IDS_FLAGS_CONFORMANT_TRUNC_COORD, add device_info fields:
+ *            tcp_cache_size, num_sqc_per_wgp, sqc_data_cache_size, sqc_inst_cache_size,
+ *            gl1c_cache_size, gl2c_cache_size, mall_size, enabled_rb_pipes_mask_hi
+ *   3.53.0 - Support for GFX11 CP GFX shadowing
+ *   3.54.0 - Add AMDGPU_CTX_QUERY2_FLAGS_RESET_IN_PROGRESS support
+ * - 3.55.0 - Add AMDGPU_INFO_GPUVM_FAULT query
+ * - 3.56.0 - Update IB start address and size alignment for decode and encode
+ * - 3.57.0 - Compute tunneling on GFX10+
+ * - 3.58.0 - Add GFX12 DCC support
+ * - 3.59.0 - Cleared VRAM
+ * - 3.60.0 - Add AMDGPU_TILING_GFX12_DCC_WRITE_COMPRESS_DISABLE (Vulkan requirement)
+ * - 3.61.0 - Contains fix for RV/PCO compute queues
+ * - 3.62.0 - Add AMDGPU_IDS_FLAGS_MODE_PF, AMDGPU_IDS_FLAGS_MODE_VF & AMDGPU_IDS_FLAGS_MODE_PT
+ * - 3.63.0 - GFX12 display DCC supports 256B max compressed block size
+ * - 3.64.0 - Userq IP support query
  */
 #define KMS_DRIVER_MAJOR	3
-#define KMS_DRIVER_MINOR	44
+#define KMS_DRIVER_MINOR	64
 #define KMS_DRIVER_PATCHLEVEL	0
 
-int amdgpu_vram_limit;
+/*
+ * amdgpu.debug module options. Are all disabled by default
+ */
+enum AMDGPU_DEBUG_MASK {
+	AMDGPU_DEBUG_VM = BIT(0),
+	AMDGPU_DEBUG_LARGEBAR = BIT(1),
+	AMDGPU_DEBUG_DISABLE_GPU_SOFT_RECOVERY = BIT(2),
+	AMDGPU_DEBUG_USE_VRAM_FW_BUF = BIT(3),
+	AMDGPU_DEBUG_ENABLE_RAS_ACA = BIT(4),
+	AMDGPU_DEBUG_ENABLE_EXP_RESETS = BIT(5),
+	AMDGPU_DEBUG_DISABLE_GPU_RING_RESET = BIT(6),
+	AMDGPU_DEBUG_SMU_POOL = BIT(7),
+	AMDGPU_DEBUG_VM_USERPTR = BIT(8),
+	AMDGPU_DEBUG_DISABLE_RAS_CE_LOG = BIT(9),
+	AMDGPU_DEBUG_ENABLE_CE_CS = BIT(10)
+};
+
+unsigned int amdgpu_vram_limit = UINT_MAX;
 int amdgpu_vis_vram_limit;
 int amdgpu_gart_size = -1; /* auto */
 int amdgpu_gtt_size = -1; /* auto */
 int amdgpu_moverate = -1; /* auto */
-int amdgpu_benchmarking;
-int amdgpu_testing;
 int amdgpu_audio = -1;
 int amdgpu_disp_priority;
 int amdgpu_hw_i2c;
@@ -128,7 +170,6 @@ int amdgpu_vm_size = -1;
 int amdgpu_vm_fragment_size = -1;
 int amdgpu_vm_block_size = -1;
 int amdgpu_vm_fault_stop;
-int amdgpu_vm_debug;
 int amdgpu_vm_update_mode = -1;
 int amdgpu_exp_hw_support;
 int amdgpu_dc = -1;
@@ -136,11 +177,22 @@ int amdgpu_sched_jobs = 32;
 int amdgpu_sched_hw_submission = 2;
 uint amdgpu_pcie_gen_cap;
 uint amdgpu_pcie_lane_cap;
-uint amdgpu_cg_mask = 0xffffffff;
+u64 amdgpu_cg_mask = 0xffffffffffffffff;
 uint amdgpu_pg_mask = 0xffffffff;
 uint amdgpu_sdma_phase_quantum = 32;
-char *amdgpu_disable_cu = NULL;
-char *amdgpu_virtual_display = NULL;
+char *amdgpu_disable_cu;
+char *amdgpu_virtual_display;
+int amdgpu_enforce_isolation = -1;
+int amdgpu_modeset = -1;
+
+/* Specifies the default granularity for SVM, used in buffer
+ * migration and restoration of backing memory when handling
+ * recoverable page faults.
+ *
+ * The value is given as log(numPages(buffer)); for a 2 MiB
+ * buffer it computes to be 9
+ */
+uint amdgpu_svm_default_granularity = 9;
 
 /*
  * OverDrive(bit 14) disabled by default
@@ -148,7 +200,6 @@ char *amdgpu_virtual_display = NULL;
  */
 uint amdgpu_pp_feature_mask = 0xfff7bfff;
 uint amdgpu_force_long_training;
-int amdgpu_job_hang_limit;
 int amdgpu_lbpw = -1;
 int amdgpu_compute_multipipe = -1;
 int amdgpu_gpu_recovery = -1; /* auto */
@@ -167,10 +218,14 @@ int amdgpu_smu_pptable_id = -1;
  */
 uint amdgpu_dc_feature_mask = 2;
 uint amdgpu_dc_debug_mask;
+uint amdgpu_dc_visual_confirm;
 int amdgpu_async_gfx_ring = 1;
-int amdgpu_mcbp;
+int amdgpu_mcbp = -1;
 int amdgpu_discovery = -1;
 int amdgpu_mes;
+int amdgpu_mes_log_enable = 0;
+int amdgpu_mes_kiq;
+int amdgpu_uni_mes = 1;
 int amdgpu_noretry = -1;
 int amdgpu_force_asic_type = -1;
 int amdgpu_tmz = -1; /* auto */
@@ -178,14 +233,34 @@ uint amdgpu_freesync_vid_mode;
 int amdgpu_reset_method = -1; /* auto */
 int amdgpu_num_kcq = -1;
 int amdgpu_smartshift_bias;
-
-static void amdgpu_drv_delayed_reset_work_handler(struct work_struct *work);
+int amdgpu_use_xgmi_p2p = 1;
+int amdgpu_vcnfw_log;
+int amdgpu_sg_display = -1; /* auto */
+int amdgpu_user_partt_mode = AMDGPU_AUTO_COMPUTE_PARTITION_MODE;
+int amdgpu_umsch_mm;
+int amdgpu_seamless = -1; /* auto */
+uint amdgpu_debug_mask;
+int amdgpu_agp = -1; /* auto */
+int amdgpu_wbrf = -1;
+int amdgpu_damage_clips = -1; /* auto */
+int amdgpu_umsch_mm_fwlog;
+int amdgpu_rebar = -1; /* auto */
+int amdgpu_user_queue = -1;
+
+DECLARE_DYNDBG_CLASSMAP(drm_debug_classes, DD_CLASS_TYPE_DISJOINT_BITS, 0,
+			"DRM_UT_CORE",
+			"DRM_UT_DRIVER",
+			"DRM_UT_KMS",
+			"DRM_UT_PRIME",
+			"DRM_UT_ATOMIC",
+			"DRM_UT_VBL",
+			"DRM_UT_STATE",
+			"DRM_UT_LEASE",
+			"DRM_UT_DP",
+			"DRM_UT_DRMRES");
 
 struct amdgpu_mgpu_info mgpu_info = {
 	.mutex = __MUTEX_INITIALIZER(mgpu_info.mutex),
-	.delayed_reset_work = __DELAYED_WORK_INITIALIZER(
-			mgpu_info.delayed_reset_work,
-			amdgpu_drv_delayed_reset_work_handler, 0),
 };
 int amdgpu_ras_enable = -1;
 uint amdgpu_ras_mask = 0xffffffff;
@@ -211,17 +286,19 @@ module_param_named(vis_vramlimit, amdgpu_vis_vram_limit, int, 0444);
 
 /**
  * DOC: gartsize (uint)
- * Restrict the size of GART in Mib (32, 64, etc.) for testing. The default is -1 (The size depends on asic).
+ * Restrict the size of GART (for kernel use) in Mib (32, 64, etc.) for testing.
+ * The default is -1 (The size depends on asic).
  */
-MODULE_PARM_DESC(gartsize, "Size of GART to setup in megabytes (32, 64, etc., -1=auto)");
+MODULE_PARM_DESC(gartsize, "Size of kernel GART to setup in megabytes (32, 64, etc., -1=auto)");
 module_param_named(gartsize, amdgpu_gart_size, uint, 0600);
 
 /**
  * DOC: gttsize (int)
- * Restrict the size of GTT domain in MiB for testing. The default is -1 (It's VRAM size if 3GB < VRAM < 3/4 RAM,
- * otherwise 3/4 RAM size).
+ * Restrict the size of GTT domain (for userspace use) in MiB for testing.
+ * The default is -1 (Use value specified by TTM).
+ * This parameter is deprecated and will be removed in the future.
  */
-MODULE_PARM_DESC(gttsize, "Size of the GTT domain in megabytes (-1 = auto)");
+MODULE_PARM_DESC(gttsize, "Size of the GTT userspace domain in megabytes (-1 = auto)");
 module_param_named(gttsize, amdgpu_gtt_size, int, 0600);
 
 /**
@@ -232,24 +309,10 @@ MODULE_PARM_DESC(moverate, "Maximum buffer migration rate in MB/s. (32, 64, etc.
 module_param_named(moverate, amdgpu_moverate, int, 0600);
 
 /**
- * DOC: benchmark (int)
- * Run benchmarks. The default is 0 (Skip benchmarks).
- */
-MODULE_PARM_DESC(benchmark, "Run benchmark");
-module_param_named(benchmark, amdgpu_benchmarking, int, 0444);
-
-/**
- * DOC: test (int)
- * Test BO GTT->VRAM and VRAM->GTT GPU copies. The default is 0 (Skip test, only set 1 to run test).
- */
-MODULE_PARM_DESC(test, "Run tests");
-module_param_named(test, amdgpu_testing, int, 0444);
-
-/**
  * DOC: audio (int)
  * Set HDMI/DPAudio. Only affects non-DC display handling. The default is -1 (Enabled), set 0 to disabled it.
  */
-MODULE_PARM_DESC(audio, "Audio enable (-1 = auto, 0 = disable, 1 = enable)");
+MODULE_PARM_DESC(audio, "HDMI/DP Audio enable for non DC displays (-1 = auto, 0 = disable, 1 = enable)");
 module_param_named(audio, amdgpu_audio, int, 0444);
 
 /**
@@ -281,27 +344,26 @@ MODULE_PARM_DESC(msi, "MSI support (1 = enable, 0 = disable, -1 = auto)");
 module_param_named(msi, amdgpu_msi, int, 0444);
 
 /**
+ * DOC: svm_default_granularity (uint)
+ * Used in buffer migration and handling of recoverable page faults
+ */
+MODULE_PARM_DESC(svm_default_granularity, "SVM's default granularity in log(2^Pages), default 9 = 2^9 = 2 MiB");
+module_param_named(svm_default_granularity, amdgpu_svm_default_granularity, uint, 0644);
+
+/**
  * DOC: lockup_timeout (string)
  * Set GPU scheduler timeout value in ms.
  *
- * The format can be [Non-Compute] or [GFX,Compute,SDMA,Video]. That is there can be one or
- * multiple values specified. 0 and negative values are invalidated. They will be adjusted
- * to the default timeout.
- *
- * - With one value specified, the setting will apply to all non-compute jobs.
- * - With multiple values specified, the first one will be for GFX.
- *   The second one is for Compute. The third and fourth ones are
- *   for SDMA and Video.
+ * The format can be [single value] for setting all timeouts at once or
+ * [GFX,Compute,SDMA,Video] to set individual timeouts.
+ * Negative values mean infinity.
  *
- * By default(with no lockup_timeout settings), the timeout for all non-compute(GFX, SDMA and Video)
- * jobs is 10000. The timeout for compute is 60000.
+ * By default(with no lockup_timeout settings), the timeout for all queues is 2000.
  */
-MODULE_PARM_DESC(lockup_timeout, "GPU lockup timeout in ms (default: for bare metal 10000 for non-compute jobs and 60000 for compute jobs; "
-		"for passthrough or sriov, 10000 for all jobs."
-		" 0: keep default value. negative: infinity timeout), "
-		"format: for bare metal [Non-Compute] or [GFX,Compute,SDMA,Video]; "
-		"for passthrough or sriov [all jobs] or [GFX,Compute,SDMA,Video].");
-module_param_string(lockup_timeout, amdgpu_lockup_timeout, sizeof(amdgpu_lockup_timeout), 0444);
+MODULE_PARM_DESC(lockup_timeout,
+		 "GPU lockup timeout in ms (default: 2000. 0: keep default value. negative: infinity timeout), format: [single value for all] or [GFX,Compute,SDMA,Video].");
+module_param_string(lockup_timeout, amdgpu_lockup_timeout,
+		    sizeof(amdgpu_lockup_timeout), 0444);
 
 /**
  * DOC: dpm (int)
@@ -314,9 +376,12 @@ module_param_named(dpm, amdgpu_dpm, int, 0444);
 
 /**
  * DOC: fw_load_type (int)
- * Set different firmware loading type for debugging (0 = direct, 1 = SMU, 2 = PSP). The default is -1 (auto).
+ * Set different firmware loading type for debugging, if supported.
+ * Set to 0 to force direct loading if supported by the ASIC.  Set
+ * to -1 to select the default loading mode for the ASIC, as defined
+ * by the driver.  The default is -1 (auto).
  */
-MODULE_PARM_DESC(fw_load_type, "firmware loading type (0 = direct, 1 = SMU, 2 = PSP, -1 = auto)");
+MODULE_PARM_DESC(fw_load_type, "firmware loading type (3 = rlc backdoor autoload if supported, 2 = smu load if supported, 1 = psp load, 0 = force direct if supported, -1 = auto)");
 module_param_named(fw_load_type, amdgpu_fw_load_type, int, 0444);
 
 /**
@@ -328,10 +393,12 @@ module_param_named(aspm, amdgpu_aspm, int, 0444);
 
 /**
  * DOC: runpm (int)
- * Override for runtime power management control for dGPUs in PX/HG laptops. The amdgpu driver can dynamically power down
- * the dGPU on PX/HG laptops when it is idle. The default is -1 (auto enable). Setting the value to 0 disables this functionality.
+ * Override for runtime power management control for dGPUs. The amdgpu driver can dynamically power down
+ * the dGPUs when they are idle if supported. The default is -1 (auto enable).
+ * Setting the value to 0 disables this functionality.
+ * Setting the value to -2 is auto enabled with power down when displays are attached.
  */
-MODULE_PARM_DESC(runpm, "PX runtime pm (2 = force enable with BAMACO, 1 = force enable with BACO, 0 = disable, -1 = PX only default)");
+MODULE_PARM_DESC(runpm, "PX runtime pm (2 = force enable with BAMACO, 1 = force enable with BACO, 0 = disable, -1 = auto, -2 = auto with displays)");
 module_param_named(runpm, amdgpu_runtime_pm, int, 0444);
 
 /**
@@ -342,7 +409,7 @@ module_param_named(runpm, amdgpu_runtime_pm, int, 0444);
  * the kernel log for the list of IPs on the asic. The default is 0xffffffff (enable all blocks on a device).
  */
 MODULE_PARM_DESC(ip_block_mask, "IP Block Mask (all blocks enabled (default))");
-module_param_named(ip_block_mask, amdgpu_ip_block_mask, uint, 0444);
+module_param_named_unsafe(ip_block_mask, amdgpu_ip_block_mask, uint, 0444);
 
 /**
  * DOC: bapm (int)
@@ -388,13 +455,6 @@ MODULE_PARM_DESC(vm_fault_stop, "Stop on VM fault (0 = never (default), 1 = prin
 module_param_named(vm_fault_stop, amdgpu_vm_fault_stop, int, 0444);
 
 /**
- * DOC: vm_debug (int)
- * Debug VM handling (0 = disabled, 1 = enabled). The default is 0 (Disabled).
- */
-MODULE_PARM_DESC(vm_debug, "Debug VM handling (0 = disabled (default), 1 = enabled)");
-module_param_named(vm_debug, amdgpu_vm_debug, int, 0644);
-
-/**
  * DOC: vm_update_mode (int)
  * Override VM update mode. VM updated by using CPU (0 = never, 1 = Graphics only, 2 = Compute only, 3 = Both). The default
  * is -1 (Only in large BAR(LB) systems Compute VM tables will be updated by CPU, otherwise 0, never).
@@ -407,7 +467,7 @@ module_param_named(vm_update_mode, amdgpu_vm_update_mode, int, 0444);
  * Enable experimental hw support (1 = enable). The default is 0 (disabled).
  */
 MODULE_PARM_DESC(exp_hw_support, "experimental hw support (1 = enable, 0 = disable (default))");
-module_param_named(exp_hw_support, amdgpu_exp_hw_support, int, 0444);
+module_param_named_unsafe(exp_hw_support, amdgpu_exp_hw_support, int, 0444);
 
 /**
  * DOC: dc (int)
@@ -463,12 +523,12 @@ MODULE_PARM_DESC(pcie_lane_cap, "PCIE Lane Caps (0: autodetect (default))");
 module_param_named(pcie_lane_cap, amdgpu_pcie_lane_cap, uint, 0444);
 
 /**
- * DOC: cg_mask (uint)
+ * DOC: cg_mask (ullong)
  * Override Clockgating features enabled on GPU (0 = disable clock gating). See the AMD_CG_SUPPORT flags in
- * drivers/gpu/drm/amd/include/amd_shared.h. The default is 0xffffffff (all enabled).
+ * drivers/gpu/drm/amd/include/amd_shared.h. The default is 0xffffffffffffffff (all enabled).
  */
 MODULE_PARM_DESC(cg_mask, "Clockgating flags mask (0 = disable clock gating)");
-module_param_named(cg_mask, amdgpu_cg_mask, uint, 0444);
+module_param_named(cg_mask, amdgpu_cg_mask, ullong, 0444);
 
 /**
  * DOC: pg_mask (uint)
@@ -504,13 +564,6 @@ MODULE_PARM_DESC(virtual_display,
 module_param_named(virtual_display, amdgpu_virtual_display, charp, 0444);
 
 /**
- * DOC: job_hang_limit (int)
- * Set how much time allow a job hang and not drop it. The default is 0.
- */
-MODULE_PARM_DESC(job_hang_limit, "how much time allow a job hang and not drop it (default 0)");
-module_param_named(job_hang_limit, amdgpu_job_hang_limit, int ,0444);
-
-/**
  * DOC: lbpw (int)
  * Override Load Balancing Per Watt (LBPW) support (1 = enable, 0 = disable). The default is -1 (auto, enabled).
  */
@@ -524,15 +577,15 @@ module_param_named(compute_multipipe, amdgpu_compute_multipipe, int, 0444);
  * DOC: gpu_recovery (int)
  * Set to enable GPU recovery mechanism (1 = enable, 0 = disable). The default is -1 (auto, disabled except SRIOV).
  */
-MODULE_PARM_DESC(gpu_recovery, "Enable GPU recovery mechanism, (2 = advanced tdr mode, 1 = enable, 0 = disable, -1 = auto)");
-module_param_named(gpu_recovery, amdgpu_gpu_recovery, int, 0444);
+MODULE_PARM_DESC(gpu_recovery, "Enable GPU recovery mechanism, (1 = enable, 0 = disable, -1 = auto)");
+module_param_named_unsafe(gpu_recovery, amdgpu_gpu_recovery, int, 0444);
 
 /**
  * DOC: emu_mode (int)
  * Set value 1 to enable emulation mode. This is only needed when running on an emulator. The default is 0 (disabled).
  */
 MODULE_PARM_DESC(emu_mode, "Emulation mode, (1 = enable, 0 = disable)");
-module_param_named(emu_mode, amdgpu_emu_mode, int, 0444);
+module_param_named_unsafe(emu_mode, amdgpu_emu_mode, int, 0444);
 
 /**
  * DOC: ras_enable (int)
@@ -565,39 +618,39 @@ module_param_named(timeout_period, amdgpu_watchdog_timer.period, uint, 0644);
 
 /**
  * DOC: si_support (int)
- * Set SI support driver. This parameter works after set config CONFIG_DRM_AMDGPU_SI. For SI asic, when radeon driver is enabled,
- * set value 0 to use radeon driver, while set value 1 to use amdgpu driver. The default is using radeon driver when it available,
- * otherwise using amdgpu driver.
- */
+ * 1 = enabled, 0 = disabled, -1 = default
+ *
+ * SI (Southern Islands) are first generation GCN GPUs, supported by both
+ * drivers: radeon (old) and amdgpu (new). This parameter controls whether
+ * amdgpu should support SI.
+ * By default, SI dedicated GPUs are supported by amdgpu.
+ * Only relevant when CONFIG_DRM_AMDGPU_SI is enabled to build SI support in amdgpu.
+ * See also radeon.si_support which should be disabled when amdgpu.si_support is
+ * enabled, and vice versa.
+ */
+int amdgpu_si_support = -1;
 #ifdef CONFIG_DRM_AMDGPU_SI
-
-#if defined(CONFIG_DRM_RADEON) || defined(CONFIG_DRM_RADEON_MODULE)
-int amdgpu_si_support = 0;
-MODULE_PARM_DESC(si_support, "SI support (1 = enabled, 0 = disabled (default))");
-#else
-int amdgpu_si_support = 1;
-MODULE_PARM_DESC(si_support, "SI support (1 = enabled (default), 0 = disabled)");
-#endif
-
+MODULE_PARM_DESC(si_support, "SI support (1 = enabled, 0 = disabled, -1 = default)");
 module_param_named(si_support, amdgpu_si_support, int, 0444);
 #endif
 
 /**
  * DOC: cik_support (int)
- * Set CIK support driver. This parameter works after set config CONFIG_DRM_AMDGPU_CIK. For CIK asic, when radeon driver is enabled,
- * set value 0 to use radeon driver, while set value 1 to use amdgpu driver. The default is using radeon driver when it available,
- * otherwise using amdgpu driver.
- */
+ * 1 = enabled, 0 = disabled, -1 = default
+ *
+ * CIK (Sea Islands) are second generation GCN GPUs, supported by both
+ * drivers: radeon (old) and amdgpu (new). This parameter controls whether
+ * amdgpu should support CIK.
+ * By default:
+ * - CIK dedicated GPUs are supported by amdgpu.
+ * - CIK APUs are supported by radeon (except when radeon is not built).
+ * Only relevant when CONFIG_DRM_AMDGPU_CIK is enabled to build CIK support in amdgpu.
+ * See also radeon.cik_support which should be disabled when amdgpu.cik_support is
+ * enabled, and vice versa.
+ */
+int amdgpu_cik_support = -1;
 #ifdef CONFIG_DRM_AMDGPU_CIK
-
-#if defined(CONFIG_DRM_RADEON) || defined(CONFIG_DRM_RADEON_MODULE)
-int amdgpu_cik_support = 0;
-MODULE_PARM_DESC(cik_support, "CIK support (1 = enabled, 0 = disabled (default))");
-#else
-int amdgpu_cik_support = 1;
-MODULE_PARM_DESC(cik_support, "CIK support (1 = enabled (default), 0 = disabled)");
-#endif
-
+MODULE_PARM_DESC(cik_support, "CIK support  (1 = enabled, 0 = disabled, -1 = default)");
 module_param_named(cik_support, amdgpu_cik_support, int, 0444);
 #endif
 
@@ -607,8 +660,7 @@ module_param_named(cik_support, amdgpu_cik_support, int, 0444);
  * E.g. 0x1 = 256Mbyte, 0x2 = 512Mbyte, 0x4 = 1 Gbyte, 0x8 = 2GByte. The default is 0 (disabled).
  */
 MODULE_PARM_DESC(smu_memory_pool_size,
-	"reserve gtt for smu debug usage, 0 = disable,"
-		"0x1 = 256Mbyte, 0x2 = 512Mbyte, 0x4 = 1 Gbyte, 0x8 = 2GByte");
+	"reserve gtt for smu debug usage, 0 = disable,0x1 = 256Mbyte, 0x2 = 512Mbyte, 0x4 = 1 Gbyte, 0x8 = 2GByte");
 module_param_named(smu_memory_pool_size, amdgpu_smu_memory_pool_size, uint, 0444);
 
 /**
@@ -621,10 +673,10 @@ module_param_named(async_gfx_ring, amdgpu_async_gfx_ring, int, 0444);
 
 /**
  * DOC: mcbp (int)
- * It is used to enable mid command buffer preemption. (0 = disabled (default), 1 = enabled)
+ * It is used to enable mid command buffer preemption. (0 = disabled, 1 = enabled, -1 auto (default))
  */
 MODULE_PARM_DESC(mcbp,
-	"Enable Mid-command buffer preemption (0 = disabled (default), 1 = enabled)");
+	"Enable Mid-command buffer preemption (0 = disabled, 1 = enabled), -1 = auto (default)");
 module_param_named(mcbp, amdgpu_mcbp, int, 0444);
 
 /**
@@ -646,6 +698,33 @@ MODULE_PARM_DESC(mes,
 module_param_named(mes, amdgpu_mes, int, 0444);
 
 /**
+ * DOC: mes_log_enable (int)
+ * Enable Micro Engine Scheduler log. This is used to enable/disable MES internal log.
+ * (0 = disabled (default), 1 = enabled)
+ */
+MODULE_PARM_DESC(mes_log_enable,
+	"Enable Micro Engine Scheduler log (0 = disabled (default), 1 = enabled)");
+module_param_named(mes_log_enable, amdgpu_mes_log_enable, int, 0444);
+
+/**
+ * DOC: mes_kiq (int)
+ * Enable Micro Engine Scheduler KIQ. This is a new engine pipe for kiq.
+ * (0 = disabled (default), 1 = enabled)
+ */
+MODULE_PARM_DESC(mes_kiq,
+	"Enable Micro Engine Scheduler KIQ (0 = disabled (default), 1 = enabled)");
+module_param_named(mes_kiq, amdgpu_mes_kiq, int, 0444);
+
+/**
+ * DOC: uni_mes (int)
+ * Enable Unified Micro Engine Scheduler. This is a new engine pipe for unified scheduler.
+ * (0 = disabled (default), 1 = enabled)
+ */
+MODULE_PARM_DESC(uni_mes,
+	"Enable Unified Micro Engine Scheduler (0 = disabled, 1 = enabled(default)");
+module_param_named(uni_mes, amdgpu_uni_mes, int, 0444);
+
+/**
  * DOC: noretry (int)
  * Disable XNACK retry in the SQ by default on GFXv9 hardware. On ASICs that
  * do not support per-process XNACK this also disables retry page faults.
@@ -661,8 +740,15 @@ module_param_named(noretry, amdgpu_noretry, int, 0644);
  */
 MODULE_PARM_DESC(force_asic_type,
 	"A non negative value used to specify the asic type for all supported GPUs");
-module_param_named(force_asic_type, amdgpu_force_asic_type, int, 0444);
+module_param_named_unsafe(force_asic_type, amdgpu_force_asic_type, int, 0444);
 
+/**
+ * DOC: use_xgmi_p2p (int)
+ * Enables/disables XGMI P2P interface (0 = disable, 1 = enable).
+ */
+MODULE_PARM_DESC(use_xgmi_p2p,
+	"Enable XGMI P2P interface (0 = disable; 1 = enable (default))");
+module_param_named(use_xgmi_p2p, amdgpu_use_xgmi_p2p, int, 0444);
 
 
 #ifdef CONFIG_HSA_AMD
@@ -673,7 +759,7 @@ module_param_named(force_asic_type, amdgpu_force_asic_type, int, 0444);
  * assigns queues to HQDs.
  */
 int sched_policy = KFD_SCHED_POLICY_HWS;
-module_param(sched_policy, int, 0444);
+module_param_unsafe(sched_policy, int, 0444);
 MODULE_PARM_DESC(sched_policy,
 	"Scheduling policy (0 = HWS (Default), 1 = HWS without over-subscription, 2 = Non-HWS (Used for debugging only)");
 
@@ -682,7 +768,7 @@ MODULE_PARM_DESC(sched_policy,
  * Maximum number of processes that HWS can schedule concurrently. The maximum is the
  * number of VMIDs assigned to the HWS, which is also the default.
  */
-int hws_max_conc_proc = 8;
+int hws_max_conc_proc = -1;
 module_param(hws_max_conc_proc, int, 0444);
 MODULE_PARM_DESC(hws_max_conc_proc,
 	"Max # processes HWS can execute concurrently when sched_policy=0 (0 = no concurrency, #VMIDs for KFD = Maximum(default))");
@@ -718,38 +804,12 @@ MODULE_PARM_DESC(send_sigterm,
 	"Send sigterm to HSA process on unhandled exception (0 = disable, 1 = enable)");
 
 /**
- * DOC: debug_largebar (int)
- * Set debug_largebar as 1 to enable simulating large-bar capability on non-large bar
- * system. This limits the VRAM size reported to ROCm applications to the visible
- * size, usually 256MB.
- * Default value is 0, diabled.
- */
-int debug_largebar;
-module_param(debug_largebar, int, 0444);
-MODULE_PARM_DESC(debug_largebar,
-	"Debug large-bar flag used to simulate large-bar capability on non-large bar machine (0 = disable, 1 = enable)");
-
-/**
- * DOC: ignore_crat (int)
- * Ignore CRAT table during KFD initialization. By default, KFD uses the ACPI CRAT
- * table to get information about AMD APUs. This option can serve as a workaround on
- * systems with a broken CRAT table.
- *
- * Default is auto (according to asic type, iommu_v2, and crat table, to decide
- * whehter use CRAT)
- */
-int ignore_crat;
-module_param(ignore_crat, int, 0444);
-MODULE_PARM_DESC(ignore_crat,
-	"Ignore CRAT table during KFD initialization (0 = auto (default), 1 = ignore CRAT)");
-
-/**
  * DOC: halt_if_hws_hang (int)
  * Halt if HWS hang is detected. Default value, 0, disables the halt on hang.
  * Setting 1 enables halt on hang.
  */
 int halt_if_hws_hang;
-module_param(halt_if_hws_hang, int, 0644);
+module_param_unsafe(halt_if_hws_hang, int, 0644);
 MODULE_PARM_DESC(halt_if_hws_hang, "Halt if HWS hang is detected (0 = off (default), 1 = on)");
 
 /**
@@ -758,13 +818,13 @@ MODULE_PARM_DESC(halt_if_hws_hang, "Halt if HWS hang is detected (0 = off (defau
  * check says. Default value: false (rely on MEC2 firmware version check).
  */
 bool hws_gws_support;
-module_param(hws_gws_support, bool, 0444);
+module_param_unsafe(hws_gws_support, bool, 0444);
 MODULE_PARM_DESC(hws_gws_support, "Assume MEC2 FW supports GWS barriers (false = rely on FW version check (Default), true = force supported)");
 
 /**
-  * DOC: queue_preemption_timeout_ms (int)
-  * queue preemption timeout in ms (1 = Minimum, 9000 = default)
-  */
+ * DOC: queue_preemption_timeout_ms (int)
+ * queue preemption timeout in ms (1 = Minimum, 9000 = default)
+ */
 int queue_preemption_timeout_ms = 9000;
 module_param(queue_preemption_timeout_ms, int, 0644);
 MODULE_PARM_DESC(queue_preemption_timeout_ms, "queue preemption timeout in ms (1 = Minimum, 9000 = default)");
@@ -789,9 +849,26 @@ MODULE_PARM_DESC(no_system_mem_limit, "disable system memory limit (false = defa
  * DOC: no_queue_eviction_on_vm_fault (int)
  * If set, process queues will not be evicted on gpuvm fault. This is to keep the wavefront context for debugging (0 = queue eviction, 1 = no queue eviction). The default is 0 (queue eviction).
  */
-int amdgpu_no_queue_eviction_on_vm_fault = 0;
+int amdgpu_no_queue_eviction_on_vm_fault;
 MODULE_PARM_DESC(no_queue_eviction_on_vm_fault, "No queue eviction on VM fault (0 = queue eviction, 1 = no queue eviction)");
-module_param_named(no_queue_eviction_on_vm_fault, amdgpu_no_queue_eviction_on_vm_fault, int, 0444);
+module_param_named_unsafe(no_queue_eviction_on_vm_fault, amdgpu_no_queue_eviction_on_vm_fault, int, 0444);
+#endif
+
+/**
+ * DOC: mtype_local (int)
+ */
+int amdgpu_mtype_local;
+MODULE_PARM_DESC(mtype_local, "MTYPE for local memory (0 = MTYPE_RW (default), 1 = MTYPE_NC, 2 = MTYPE_CC)");
+module_param_named_unsafe(mtype_local, amdgpu_mtype_local, int, 0444);
+
+/**
+ * DOC: pcie_p2p (bool)
+ * Enable PCIe P2P (requires large-BAR). Default value: true (on)
+ */
+#ifdef CONFIG_HSA_AMD_P2P
+bool pcie_p2p = true;
+module_param(pcie_p2p, bool, 0444);
+MODULE_PARM_DESC(pcie_p2p, "Enable PCIe P2P (requires large-BAR). (N = off, Y = on(default))");
 #endif
 
 /**
@@ -804,11 +881,14 @@ module_param_named(dcfeaturemask, amdgpu_dc_feature_mask, uint, 0444);
 
 /**
  * DOC: dcdebugmask (uint)
- * Override display features enabled. See enum DC_DEBUG_MASK in drivers/gpu/drm/amd/include/amd_shared.h.
+ * Display debug options. See enum DC_DEBUG_MASK in drivers/gpu/drm/amd/include/amd_shared.h.
  */
 MODULE_PARM_DESC(dcdebugmask, "all debug options disabled (default))");
 module_param_named(dcdebugmask, amdgpu_dc_debug_mask, uint, 0444);
 
+MODULE_PARM_DESC(visualconfirm, "Visual confirm (0 = off (default), 1 = MPO, 5 = PSR)");
+module_param_named(visualconfirm, amdgpu_dc_visual_confirm, uint, 0444);
+
 /**
  * DOC: abmlevel (uint)
  * Override the default ABM (Adaptive Backlight Management) level used for DC
@@ -818,18 +898,31 @@ module_param_named(dcdebugmask, amdgpu_dc_debug_mask, uint, 0444);
  * the ABM algorithm, with 1 being the least reduction and 4 being the most
  * reduction.
  *
- * Defaults to 0, or disabled. Userspace can still override this level later
- * after boot.
+ * Defaults to -1, or auto. Userspace can only override this level after
+ * boot if it's set to auto.
  */
-uint amdgpu_dm_abm_level;
-MODULE_PARM_DESC(abmlevel, "ABM level (0 = off (default), 1-4 = backlight reduction level) ");
-module_param_named(abmlevel, amdgpu_dm_abm_level, uint, 0444);
+int amdgpu_dm_abm_level = -1;
+MODULE_PARM_DESC(abmlevel,
+		 "ABM level (0 = off, 1-4 = backlight reduction level, -1 auto (default))");
+module_param_named(abmlevel, amdgpu_dm_abm_level, int, 0444);
 
 int amdgpu_backlight = -1;
 MODULE_PARM_DESC(backlight, "Backlight control (0 = pwm, 1 = aux, -1 auto (default))");
 module_param_named(backlight, amdgpu_backlight, bint, 0444);
 
 /**
+ * DOC: damageclips (int)
+ * Enable or disable damage clips support. If damage clips support is disabled,
+ * we will force full frame updates, irrespective of what user space sends to
+ * us.
+ *
+ * Defaults to -1 (where it is enabled unless a PSR-SU display is detected).
+ */
+MODULE_PARM_DESC(damageclips,
+		 "Damage clips support (0 = disable, 1 = enable, -1 auto (default))");
+module_param_named(damageclips, amdgpu_damage_clips, int, 0444);
+
+/**
  * DOC: tmz (int)
  * Trusted Memory Zone (TMZ) is a method to protect data being written
  * to or read from memory.
@@ -862,15 +955,15 @@ module_param_named(tmz, amdgpu_tmz, int, 0444);
  */
 MODULE_PARM_DESC(
 	freesync_video,
-	"Enable freesync modesetting optimization feature (0 = off (default), 1 = on)");
+	"Adds additional modes via VRR for refresh changes without a full modeset (0 = off (default), 1 = on)");
 module_param_named(freesync_video, amdgpu_freesync_vid_mode, uint, 0444);
 
 /**
  * DOC: reset_method (int)
- * GPU reset method (-1 = auto (default), 0 = legacy, 1 = mode0, 2 = mode1, 3 = mode2, 4 = baco, 5 = pci)
+ * GPU reset method (-1 = auto (default), 0 = legacy, 1 = mode0, 2 = mode1, 3 = mode2, 4 = baco)
  */
-MODULE_PARM_DESC(reset_method, "GPU reset method (-1 = auto (default), 0 = legacy, 1 = mode0, 2 = mode1, 3 = mode2, 4 = baco/bamaco, 5 = pci)");
-module_param_named(reset_method, amdgpu_reset_method, int, 0444);
+MODULE_PARM_DESC(reset_method, "GPU reset method (-1 = auto (default), 0 = legacy, 1 = mode0, 2 = mode1, 3 = mode2, 4 = baco/bamaco)");
+module_param_named_unsafe(reset_method, amdgpu_reset_method, int, 0644);
 
 /**
  * DOC: bad_page_threshold (int) Bad page threshold is specifies the
@@ -878,13 +971,46 @@ module_param_named(reset_method, amdgpu_reset_method, int, 0444);
  * result in the GPU entering bad status when the number of total
  * faulty pages by ECC exceeds the threshold value.
  */
-MODULE_PARM_DESC(bad_page_threshold, "Bad page threshold(-1 = auto(default value), 0 = disable bad page retirement, -2 = ignore bad page threshold)");
+MODULE_PARM_DESC(bad_page_threshold, "Bad page threshold(-1 = ignore threshold (default value), 0 = disable bad page retirement, -2 = threshold determined by a formula, 0 < threshold < max records, user-defined threshold)");
 module_param_named(bad_page_threshold, amdgpu_bad_page_threshold, int, 0444);
 
 MODULE_PARM_DESC(num_kcq, "number of kernel compute queue user want to setup (8 if set to greater than 8 or less than 0, only affect gfx 8+)");
 module_param_named(num_kcq, amdgpu_num_kcq, int, 0444);
 
 /**
+ * DOC: vcnfw_log (int)
+ * Enable vcnfw log output for debugging, the default is disabled.
+ */
+MODULE_PARM_DESC(vcnfw_log, "Enable vcnfw log(0 = disable (default value), 1 = enable)");
+module_param_named(vcnfw_log, amdgpu_vcnfw_log, int, 0444);
+
+/**
+ * DOC: sg_display (int)
+ * Disable S/G (scatter/gather) display (i.e., display from system memory).
+ * This option is only relevant on APUs.  Set this option to 0 to disable
+ * S/G display if you experience flickering or other issues under memory
+ * pressure and report the issue.
+ */
+MODULE_PARM_DESC(sg_display, "S/G Display (-1 = auto (default), 0 = disable)");
+module_param_named(sg_display, amdgpu_sg_display, int, 0444);
+
+/**
+ * DOC: umsch_mm (int)
+ * Enable Multi Media User Mode Scheduler. This is a HW scheduling engine for VCN and VPE.
+ * (0 = disabled (default), 1 = enabled)
+ */
+MODULE_PARM_DESC(umsch_mm,
+	"Enable Multi Media User Mode Scheduler (0 = disabled (default), 1 = enabled)");
+module_param_named(umsch_mm, amdgpu_umsch_mm, int, 0444);
+
+/**
+ * DOC: umsch_mm_fwlog (int)
+ * Enable umschfw log output for debugging, the default is disabled.
+ */
+MODULE_PARM_DESC(umsch_mm_fwlog, "Enable umschfw log(0 = disable (default value), 1 = enable)");
+module_param_named(umsch_mm_fwlog, amdgpu_umsch_mm_fwlog, int, 0444);
+
+/**
  * DOC: smu_pptable_id (int)
  * Used to override pptable id. id = 0 use VBIOS pptable.
  * id > 0 use the soft pptable with specicfied id.
@@ -893,6 +1019,109 @@ MODULE_PARM_DESC(smu_pptable_id,
 	"specify pptable id to be used (-1 = auto(default) value, 0 = use pptable from vbios, > 0 = soft pptable id)");
 module_param_named(smu_pptable_id, amdgpu_smu_pptable_id, int, 0444);
 
+/**
+ * DOC: partition_mode (int)
+ * Used to override the default SPX mode.
+ */
+MODULE_PARM_DESC(
+	user_partt_mode,
+	"specify partition mode to be used (-2 = AMDGPU_AUTO_COMPUTE_PARTITION_MODE(default value) \
+						0 = AMDGPU_SPX_PARTITION_MODE, \
+						1 = AMDGPU_DPX_PARTITION_MODE, \
+						2 = AMDGPU_TPX_PARTITION_MODE, \
+						3 = AMDGPU_QPX_PARTITION_MODE, \
+						4 = AMDGPU_CPX_PARTITION_MODE)");
+module_param_named(user_partt_mode, amdgpu_user_partt_mode, uint, 0444);
+
+
+/**
+ * DOC: enforce_isolation (int)
+ * enforce process isolation between graphics and compute.
+ * (-1 = auto, 0 = disable, 1 = enable, 2 = enable legacy mode, 3 = enable without cleaner shader)
+ */
+module_param_named(enforce_isolation, amdgpu_enforce_isolation, int, 0444);
+MODULE_PARM_DESC(enforce_isolation,
+"enforce process isolation between graphics and compute. (-1 = auto, 0 = disable, 1 = enable, 2 = enable legacy mode, 3 = enable without cleaner shader)");
+
+/**
+ * DOC: modeset (int)
+ * Override nomodeset (1 = override, -1 = auto). The default is -1 (auto).
+ */
+MODULE_PARM_DESC(modeset, "Override nomodeset (1 = enable, -1 = auto)");
+module_param_named(modeset, amdgpu_modeset, int, 0444);
+
+/**
+ * DOC: seamless (int)
+ * Seamless boot will keep the image on the screen during the boot process.
+ */
+MODULE_PARM_DESC(seamless, "Seamless boot (-1 = auto (default), 0 = disable, 1 = enable)");
+module_param_named(seamless, amdgpu_seamless, int, 0444);
+
+/**
+ * DOC: debug_mask (uint)
+ * Debug options for amdgpu, work as a binary mask with the following options:
+ *
+ * - 0x1: Debug VM handling
+ * - 0x2: Enable simulating large-bar capability on non-large bar system. This
+ *   limits the VRAM size reported to ROCm applications to the visible
+ *   size, usually 256MB.
+ * - 0x4: Disable GPU soft recovery, always do a full reset
+ * - 0x8: Use VRAM for firmware loading
+ * - 0x10: Enable ACA based RAS logging
+ * - 0x20: Enable experimental resets
+ * - 0x40: Disable ring resets
+ * - 0x80: Use VRAM for SMU pool
+ */
+MODULE_PARM_DESC(debug_mask, "debug options for amdgpu, disabled by default");
+module_param_named_unsafe(debug_mask, amdgpu_debug_mask, uint, 0444);
+
+/**
+ * DOC: agp (int)
+ * Enable the AGP aperture.  This provides an aperture in the GPU's internal
+ * address space for direct access to system memory.  Note that these accesses
+ * are non-snooped, so they are only used for access to uncached memory.
+ */
+MODULE_PARM_DESC(agp, "AGP (-1 = auto (default), 0 = disable, 1 = enable)");
+module_param_named(agp, amdgpu_agp, int, 0444);
+
+/**
+ * DOC: wbrf (int)
+ * Enable Wifi RFI interference mitigation feature.
+ * Due to electrical and mechanical constraints there may be likely interference of
+ * relatively high-powered harmonics of the (G-)DDR memory clocks with local radio
+ * module frequency bands used by Wifi 6/6e/7. To mitigate the possible RFI interference,
+ * with this feature enabled, PMFW will use either “shadowed P-State” or “P-State” based
+ * on active list of frequencies in-use (to be avoided) as part of initial setting or
+ * P-state transition. However, there may be potential performance impact with this
+ * feature enabled.
+ * (0 = disabled, 1 = enabled, -1 = auto (default setting, will be enabled if supported))
+ */
+MODULE_PARM_DESC(wbrf,
+	"Enable Wifi RFI interference mitigation (0 = disabled, 1 = enabled, -1 = auto(default)");
+module_param_named(wbrf, amdgpu_wbrf, int, 0444);
+
+/**
+ * DOC: rebar (int)
+ * Allow BAR resizing.  Disable this to prevent the driver from attempting
+ * to resize the BAR if the GPU supports it and there is available MMIO space.
+ * Note that this just prevents the driver from resizing the BAR.  The BIOS
+ * may have already resized the BAR at boot time.
+ */
+MODULE_PARM_DESC(rebar, "Resizable BAR (-1 = auto (default), 0 = disable, 1 = enable)");
+module_param_named(rebar, amdgpu_rebar, int, 0444);
+
+/**
+ * DOC: user_queue (int)
+ * Enable user queues on systems that support user queues. Possible values:
+ *
+ * - -1 = auto (ASIC specific default)
+ * -  0 = user queues disabled
+ * -  1 = user queues enabled and kernel queues enabled (if supported)
+ * -  2 = user queues enabled and kernel queues disabled
+ */
+MODULE_PARM_DESC(user_queue, "Enable user queues (-1 = auto (default), 0 = disable, 1 = enable, 2 = enable UQs and disable KQs)");
+module_param_named(user_queue, amdgpu_user_queue, int, 0444);
+
 /* These devices are not supported by amdgpu.
  * They are supported by the mach64, r128, radeon drivers
  */
@@ -1521,10 +1750,91 @@ static const u16 amdgpu_unsupported_pciidlist[] = {
 	0x99A0,
 	0x99A2,
 	0x99A4,
+	/* radeon secondary ids */
+	0x3171,
+	0x3e70,
+	0x4164,
+	0x4165,
+	0x4166,
+	0x4168,
+	0x4170,
+	0x4171,
+	0x4172,
+	0x4173,
+	0x496e,
+	0x4a69,
+	0x4a6a,
+	0x4a6b,
+	0x4a70,
+	0x4a74,
+	0x4b69,
+	0x4b6b,
+	0x4b6c,
+	0x4c6e,
+	0x4e64,
+	0x4e65,
+	0x4e66,
+	0x4e67,
+	0x4e68,
+	0x4e69,
+	0x4e6a,
+	0x4e71,
+	0x4f73,
+	0x5569,
+	0x556b,
+	0x556d,
+	0x556f,
+	0x5571,
+	0x5854,
+	0x5874,
+	0x5940,
+	0x5941,
+	0x5b70,
+	0x5b72,
+	0x5b73,
+	0x5b74,
+	0x5b75,
+	0x5d44,
+	0x5d45,
+	0x5d6d,
+	0x5d6f,
+	0x5d72,
+	0x5d77,
+	0x5e6b,
+	0x5e6d,
+	0x7120,
+	0x7124,
+	0x7129,
+	0x712e,
+	0x712f,
+	0x7162,
+	0x7163,
+	0x7166,
+	0x7167,
+	0x7172,
+	0x7173,
+	0x71a0,
+	0x71a1,
+	0x71a3,
+	0x71a7,
+	0x71bb,
+	0x71e0,
+	0x71e1,
+	0x71e2,
+	0x71e6,
+	0x71e7,
+	0x71f2,
+	0x7269,
+	0x726b,
+	0x726e,
+	0x72a0,
+	0x72a8,
+	0x72b1,
+	0x72b3,
+	0x793f,
 };
 
 static const struct pci_device_id pciidlist[] = {
-#ifdef  CONFIG_DRM_AMDGPU_SI
 	{0x1002, 0x6780, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TAHITI},
 	{0x1002, 0x6784, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TAHITI},
 	{0x1002, 0x6788, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TAHITI},
@@ -1597,8 +1907,6 @@ static const struct pci_device_id pciidlist[] = {
 	{0x1002, 0x6665, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_HAINAN|AMD_IS_MOBILITY},
 	{0x1002, 0x6667, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_HAINAN|AMD_IS_MOBILITY},
 	{0x1002, 0x666F, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_HAINAN|AMD_IS_MOBILITY},
-#endif
-#ifdef CONFIG_DRM_AMDGPU_CIK
 	/* Kaveri */
 	{0x1002, 0x1304, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_KAVERI|AMD_IS_MOBILITY|AMD_IS_APU},
 	{0x1002, 0x1305, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_KAVERI|AMD_IS_APU},
@@ -1681,7 +1989,6 @@ static const struct pci_device_id pciidlist[] = {
 	{0x1002, 0x985D, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_MULLINS|AMD_IS_MOBILITY|AMD_IS_APU},
 	{0x1002, 0x985E, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_MULLINS|AMD_IS_MOBILITY|AMD_IS_APU},
 	{0x1002, 0x985F, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_MULLINS|AMD_IS_MOBILITY|AMD_IS_APU},
-#endif
 	/* topaz */
 	{0x1002, 0x6900, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TOPAZ},
 	{0x1002, 0x6901, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TOPAZ},
@@ -1824,9 +2131,6 @@ static const struct pci_device_id pciidlist[] = {
 	{0x1002, 0x73AF, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_SIENNA_CICHLID},
 	{0x1002, 0x73BF, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_SIENNA_CICHLID},
 
-	/* Van Gogh */
-	{0x1002, 0x163F, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VANGOGH|AMD_IS_APU},
-
 	/* Yellow Carp */
 	{0x1002, 0x164D, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_YELLOW_CARP|AMD_IS_APU},
 	{0x1002, 0x1681, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_YELLOW_CARP|AMD_IS_APU},
@@ -1857,19 +2161,26 @@ static const struct pci_device_id pciidlist[] = {
 	{0x1002, 0x73FF, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_DIMGREY_CAVEFISH},
 
 	/* Aldebaran */
-	{0x1002, 0x7408, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_ALDEBARAN|AMD_EXP_HW_SUPPORT},
-	{0x1002, 0x740C, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_ALDEBARAN|AMD_EXP_HW_SUPPORT},
-	{0x1002, 0x740F, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_ALDEBARAN|AMD_EXP_HW_SUPPORT},
-	{0x1002, 0x7410, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_ALDEBARAN|AMD_EXP_HW_SUPPORT},
+	{0x1002, 0x7408, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_ALDEBARAN},
+	{0x1002, 0x740C, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_ALDEBARAN},
+	{0x1002, 0x740F, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_ALDEBARAN},
+	{0x1002, 0x7410, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_ALDEBARAN},
 
 	/* CYAN_SKILLFISH */
+	{0x1002, 0x13DB, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_CYAN_SKILLFISH|AMD_IS_APU},
+	{0x1002, 0x13F9, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_CYAN_SKILLFISH|AMD_IS_APU},
+	{0x1002, 0x13FA, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_CYAN_SKILLFISH|AMD_IS_APU},
+	{0x1002, 0x13FB, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_CYAN_SKILLFISH|AMD_IS_APU},
+	{0x1002, 0x13FC, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_CYAN_SKILLFISH|AMD_IS_APU},
 	{0x1002, 0x13FE, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_CYAN_SKILLFISH|AMD_IS_APU},
+	{0x1002, 0x143F, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_CYAN_SKILLFISH|AMD_IS_APU},
 
 	/* BEIGE_GOBY */
 	{0x1002, 0x7420, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_BEIGE_GOBY},
 	{0x1002, 0x7421, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_BEIGE_GOBY},
 	{0x1002, 0x7422, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_BEIGE_GOBY},
 	{0x1002, 0x7423, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_BEIGE_GOBY},
+	{0x1002, 0x7424, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_BEIGE_GOBY},
 	{0x1002, 0x743F, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_BEIGE_GOBY},
 
 	{ PCI_DEVICE(0x1002, PCI_ANY_ID),
@@ -1882,13 +2193,185 @@ static const struct pci_device_id pciidlist[] = {
 	  .class_mask = 0xffffff,
 	  .driver_data = CHIP_IP_DISCOVERY },
 
+	{ PCI_DEVICE(0x1002, PCI_ANY_ID),
+	  .class = PCI_CLASS_ACCELERATOR_PROCESSING << 8,
+	  .class_mask = 0xffffff,
+	  .driver_data = CHIP_IP_DISCOVERY },
+
 	{0, 0, 0}
 };
 
 MODULE_DEVICE_TABLE(pci, pciidlist);
 
+static const struct amdgpu_asic_type_quirk asic_type_quirks[] = {
+	/* differentiate between P10 and P11 asics with the same DID */
+	{0x67FF, 0xE3, CHIP_POLARIS10},
+	{0x67FF, 0xE7, CHIP_POLARIS10},
+	{0x67FF, 0xF3, CHIP_POLARIS10},
+	{0x67FF, 0xF7, CHIP_POLARIS10},
+};
+
 static const struct drm_driver amdgpu_kms_driver;
 
+static void amdgpu_get_secondary_funcs(struct amdgpu_device *adev)
+{
+	struct pci_dev *p = NULL;
+	int i;
+
+	/* 0 - GPU
+	 * 1 - audio
+	 * 2 - USB
+	 * 3 - UCSI
+	 */
+	for (i = 1; i < 4; i++) {
+		p = pci_get_domain_bus_and_slot(pci_domain_nr(adev->pdev->bus),
+						adev->pdev->bus->number, i);
+		if (p) {
+			pm_runtime_get_sync(&p->dev);
+			pm_runtime_put_autosuspend(&p->dev);
+			pci_dev_put(p);
+		}
+	}
+}
+
+static void amdgpu_init_debug_options(struct amdgpu_device *adev)
+{
+	if (amdgpu_debug_mask & AMDGPU_DEBUG_VM) {
+		pr_info("debug: VM handling debug enabled\n");
+		adev->debug_vm = true;
+	}
+
+	if (amdgpu_debug_mask & AMDGPU_DEBUG_LARGEBAR) {
+		pr_info("debug: enabled simulating large-bar capability on non-large bar system\n");
+		adev->debug_largebar = true;
+	}
+
+	if (amdgpu_debug_mask & AMDGPU_DEBUG_DISABLE_GPU_SOFT_RECOVERY) {
+		pr_info("debug: soft reset for GPU recovery disabled\n");
+		adev->debug_disable_soft_recovery = true;
+	}
+
+	if (amdgpu_debug_mask & AMDGPU_DEBUG_USE_VRAM_FW_BUF) {
+		pr_info("debug: place fw in vram for frontdoor loading\n");
+		adev->debug_use_vram_fw_buf = true;
+	}
+
+	if (amdgpu_debug_mask & AMDGPU_DEBUG_ENABLE_RAS_ACA) {
+		pr_info("debug: enable RAS ACA\n");
+		adev->debug_enable_ras_aca = true;
+	}
+
+	if (amdgpu_debug_mask & AMDGPU_DEBUG_ENABLE_EXP_RESETS) {
+		pr_info("debug: enable experimental reset features\n");
+		adev->debug_exp_resets = true;
+	}
+
+	if (amdgpu_debug_mask & AMDGPU_DEBUG_DISABLE_GPU_RING_RESET) {
+		pr_info("debug: ring reset disabled\n");
+		adev->debug_disable_gpu_ring_reset = true;
+	}
+	if (amdgpu_debug_mask & AMDGPU_DEBUG_SMU_POOL) {
+		pr_info("debug: use vram for smu pool\n");
+		adev->pm.smu_debug_mask |= SMU_DEBUG_POOL_USE_VRAM;
+	}
+	if (amdgpu_debug_mask & AMDGPU_DEBUG_VM_USERPTR) {
+		pr_info("debug: VM mode debug for userptr is enabled\n");
+		adev->debug_vm_userptr = true;
+	}
+
+	if (amdgpu_debug_mask & AMDGPU_DEBUG_DISABLE_RAS_CE_LOG) {
+		pr_info("debug: disable kernel logs of correctable errors\n");
+		adev->debug_disable_ce_logs = true;
+	}
+
+	if (amdgpu_debug_mask & AMDGPU_DEBUG_ENABLE_CE_CS) {
+		pr_info("debug: allowing command submission to CE engine\n");
+		adev->debug_enable_ce_cs = true;
+	}
+}
+
+static unsigned long amdgpu_fix_asic_type(struct pci_dev *pdev, unsigned long flags)
+{
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(asic_type_quirks); i++) {
+		if (pdev->device == asic_type_quirks[i].device &&
+			pdev->revision == asic_type_quirks[i].revision) {
+				flags &= ~AMD_ASIC_MASK;
+				flags |= asic_type_quirks[i].type;
+				break;
+			}
+	}
+
+	return flags;
+}
+
+static bool amdgpu_support_enabled(struct device *dev,
+				   const enum amd_asic_type family)
+{
+	const char *gen;
+	const char *param;
+	int module_param = -1;
+	bool radeon_support_built = IS_ENABLED(CONFIG_DRM_RADEON);
+	bool amdgpu_support_built = false;
+	bool support_by_default = false;
+
+	switch (family) {
+	case CHIP_TAHITI:
+	case CHIP_PITCAIRN:
+	case CHIP_VERDE:
+	case CHIP_OLAND:
+	case CHIP_HAINAN:
+		gen = "SI";
+		param = "si_support";
+		module_param = amdgpu_si_support;
+		amdgpu_support_built = IS_ENABLED(CONFIG_DRM_AMDGPU_SI);
+		support_by_default = true;
+		break;
+
+	case CHIP_BONAIRE:
+	case CHIP_HAWAII:
+		support_by_default = true;
+		fallthrough;
+	case CHIP_KAVERI:
+	case CHIP_KABINI:
+	case CHIP_MULLINS:
+		gen = "CIK";
+		param = "cik_support";
+		module_param = amdgpu_cik_support;
+		amdgpu_support_built = IS_ENABLED(CONFIG_DRM_AMDGPU_CIK);
+		break;
+
+	default:
+		/* All other chips are supported by amdgpu only */
+		return true;
+	}
+
+	if (!amdgpu_support_built) {
+		dev_info(dev, "amdgpu built without %s support\n", gen);
+		return false;
+	}
+
+	if ((module_param == -1 && (support_by_default || !radeon_support_built)) ||
+	    module_param == 1) {
+		if (radeon_support_built)
+			dev_info(dev, "%s support provided by amdgpu.\n"
+				 "Use radeon.%s=1 amdgpu.%s=0 to override.\n",
+				 gen, param, param);
+
+		return true;
+	}
+
+	if (radeon_support_built)
+		dev_info(dev, "%s support provided by radeon.\n"
+			 "Use radeon.%s=0 amdgpu.%s=1 to override.\n",
+			 gen, param, param);
+	else if (module_param == 0)
+		dev_info(dev, "%s support disabled by module param\n", gen);
+
+	return false;
+}
+
 static int amdgpu_pci_probe(struct pci_dev *pdev,
 			    const struct pci_device_id *ent)
 {
@@ -1898,19 +2381,23 @@ static int amdgpu_pci_probe(struct pci_dev *pdev,
 	int ret, retry = 0, i;
 	bool supports_atomic = false;
 
+	if ((pdev->class >> 8) == PCI_CLASS_DISPLAY_VGA ||
+	    (pdev->class >> 8) == PCI_CLASS_DISPLAY_OTHER) {
+		if (drm_firmware_drivers_only() && amdgpu_modeset == -1)
+			return -EINVAL;
+	}
+
 	/* skip devices which are owned by radeon */
 	for (i = 0; i < ARRAY_SIZE(amdgpu_unsupported_pciidlist); i++) {
 		if (amdgpu_unsupported_pciidlist[i] == pdev->device)
 			return -ENODEV;
 	}
 
-	if (flags == 0) {
-		DRM_INFO("Unsupported asic.  Remove me when IP discovery init is in place.\n");
-		return -ENODEV;
-	}
+	if (amdgpu_aspm == -1 && !pcie_aspm_enabled(pdev))
+		amdgpu_aspm = 0;
 
 	if (amdgpu_virtual_display ||
-	    amdgpu_device_asic_has_dc_support(flags & AMD_ASIC_MASK))
+	    amdgpu_device_asic_has_dc_support(pdev, flags & AMD_ASIC_MASK))
 		supports_atomic = true;
 
 	if ((flags & AMD_EXP_HW_SUPPORT) && !amdgpu_exp_hw_support) {
@@ -1919,6 +2406,8 @@ static int amdgpu_pci_probe(struct pci_dev *pdev,
 		return -ENODEV;
 	}
 
+	flags = amdgpu_fix_asic_type(pdev, flags);
+
 	/* Due to hardware bugs, S/G Display on raven requires a 1:1 IOMMU mapping,
 	 * however, SME requires an indirect IOMMU mapping because the encryption
 	 * bit is beyond the DMA mask of the chip.
@@ -1930,45 +2419,8 @@ static int amdgpu_pci_probe(struct pci_dev *pdev,
 		return -ENOTSUPP;
 	}
 
-#ifdef CONFIG_DRM_AMDGPU_SI
-	if (!amdgpu_si_support) {
-		switch (flags & AMD_ASIC_MASK) {
-		case CHIP_TAHITI:
-		case CHIP_PITCAIRN:
-		case CHIP_VERDE:
-		case CHIP_OLAND:
-		case CHIP_HAINAN:
-			dev_info(&pdev->dev,
-				 "SI support provided by radeon.\n");
-			dev_info(&pdev->dev,
-				 "Use radeon.si_support=0 amdgpu.si_support=1 to override.\n"
-				);
-			return -ENODEV;
-		}
-	}
-#endif
-#ifdef CONFIG_DRM_AMDGPU_CIK
-	if (!amdgpu_cik_support) {
-		switch (flags & AMD_ASIC_MASK) {
-		case CHIP_KAVERI:
-		case CHIP_BONAIRE:
-		case CHIP_HAWAII:
-		case CHIP_KABINI:
-		case CHIP_MULLINS:
-			dev_info(&pdev->dev,
-				 "CIK support provided by radeon.\n");
-			dev_info(&pdev->dev,
-				 "Use radeon.cik_support=0 amdgpu.cik_support=1 to override.\n"
-				);
-			return -ENODEV;
-		}
-	}
-#endif
-
-	/* Get rid of things like offb */
-	ret = drm_aperture_remove_conflicting_pci_framebuffers(pdev, &amdgpu_kms_driver);
-	if (ret)
-		return ret;
+	if (!amdgpu_support_enabled(&pdev->dev, flags & AMD_ASIC_MASK))
+		return -ENODEV;
 
 	adev = devm_drm_dev_alloc(&pdev->dev, &amdgpu_kms_driver, typeof(*adev), ddev);
 	if (IS_ERR(adev))
@@ -1987,12 +2439,14 @@ static int amdgpu_pci_probe(struct pci_dev *pdev,
 
 	pci_set_drvdata(pdev, ddev);
 
-	ret = amdgpu_driver_load_kms(adev, ent->driver_data);
+	amdgpu_init_debug_options(adev);
+
+	ret = amdgpu_driver_load_kms(adev, flags);
 	if (ret)
 		goto err_pci;
 
 retry_init:
-	ret = drm_dev_register(ddev, ent->driver_data);
+	ret = drm_dev_register(ddev, flags);
 	if (ret == -EAGAIN && ++retry <= 3) {
 		DRM_INFO("retry init %d\n", retry);
 		/* Don't request EX mode too frequently which is attacking */
@@ -2002,10 +2456,78 @@ retry_init:
 		goto err_pci;
 	}
 
+	ret = amdgpu_xcp_dev_register(adev, ent);
+	if (ret)
+		goto err_pci;
+
+	ret = amdgpu_amdkfd_drm_client_create(adev);
+	if (ret)
+		goto err_pci;
+
+	/*
+	 * 1. don't init fbdev on hw without DCE
+	 * 2. don't init fbdev if there are no connectors
+	 */
+	if (adev->mode_info.mode_config_initialized &&
+	    !list_empty(&adev_to_drm(adev)->mode_config.connector_list)) {
+		const struct drm_format_info *format;
+
+		/* select 8 bpp console on low vram cards */
+		if (adev->gmc.real_vram_size <= (32*1024*1024))
+			format = drm_format_info(DRM_FORMAT_C8);
+		else
+			format = NULL;
+
+		drm_client_setup(adev_to_drm(adev), format);
+	}
+
 	ret = amdgpu_debugfs_init(adev);
 	if (ret)
 		DRM_ERROR("Creating debugfs files failed (%d).\n", ret);
 
+	if (adev->pm.rpm_mode != AMDGPU_RUNPM_NONE) {
+		/* only need to skip on ATPX */
+		if (amdgpu_device_supports_px(adev))
+			dev_pm_set_driver_flags(ddev->dev, DPM_FLAG_NO_DIRECT_COMPLETE);
+		/* we want direct complete for BOCO */
+		if (amdgpu_device_supports_boco(adev))
+			dev_pm_set_driver_flags(ddev->dev, DPM_FLAG_SMART_PREPARE |
+						DPM_FLAG_SMART_SUSPEND |
+						DPM_FLAG_MAY_SKIP_RESUME);
+		pm_runtime_use_autosuspend(ddev->dev);
+		pm_runtime_set_autosuspend_delay(ddev->dev, 5000);
+
+		pm_runtime_allow(ddev->dev);
+
+		pm_runtime_put_autosuspend(ddev->dev);
+
+		pci_wake_from_d3(pdev, TRUE);
+
+		/*
+		 * For runpm implemented via BACO, PMFW will handle the
+		 * timing for BACO in and out:
+		 *   - put ASIC into BACO state only when both video and
+		 *     audio functions are in D3 state.
+		 *   - pull ASIC out of BACO state when either video or
+		 *     audio function is in D0 state.
+		 * Also, at startup, PMFW assumes both functions are in
+		 * D0 state.
+		 *
+		 * So if snd driver was loaded prior to amdgpu driver
+		 * and audio function was put into D3 state, there will
+		 * be no PMFW-aware D-state transition(D0->D3) on runpm
+		 * suspend. Thus the BACO will be not correctly kicked in.
+		 *
+		 * Via amdgpu_get_secondary_funcs(), the audio dev is put
+		 * into D0 state. Then there will be a PMFW-aware D-state
+		 * transition(D0->D3) on runpm suspend.
+		 */
+		if (amdgpu_device_supports_baco(adev) &&
+		    !(adev->flags & AMD_IS_APU) &&
+		    adev->asic_type >= CHIP_NAVI10)
+			amdgpu_get_secondary_funcs(adev);
+	}
+
 	return 0;
 
 err_pci:
@@ -2017,8 +2539,18 @@ static void
 amdgpu_pci_remove(struct pci_dev *pdev)
 {
 	struct drm_device *dev = pci_get_drvdata(pdev);
+	struct amdgpu_device *adev = drm_to_adev(dev);
 
+	amdgpu_ras_eeprom_check_and_recover(adev);
+	amdgpu_xcp_dev_unplug(adev);
+	amdgpu_gmc_prepare_nps_mode_change(adev);
 	drm_dev_unplug(dev);
+
+	if (adev->pm.rpm_mode != AMDGPU_RUNPM_NONE) {
+		pm_runtime_get_sync(dev->dev);
+		pm_runtime_forbid(dev->dev);
+	}
+
 	amdgpu_driver_unload_kms(dev);
 
 	/*
@@ -2039,6 +2571,10 @@ amdgpu_pci_shutdown(struct pci_dev *pdev)
 	if (amdgpu_ras_intr_triggered())
 		return;
 
+	/* device maybe not resumed here, return immediately in this case */
+	if (adev->in_s4 && adev->in_suspend)
+		return;
+
 	/* if we are running in a VM, make sure the device
 	 * torn down properly on reboot/shutdown.
 	 * unfortunately we can't detect certain
@@ -2046,115 +2582,85 @@ amdgpu_pci_shutdown(struct pci_dev *pdev)
 	 */
 	if (!amdgpu_passthrough(adev))
 		adev->mp1_state = PP_MP1_STATE_UNLOAD;
-	amdgpu_device_ip_suspend(adev);
+	amdgpu_device_prepare(dev);
+	amdgpu_device_suspend(dev, true);
 	adev->mp1_state = PP_MP1_STATE_NONE;
 }
 
-/**
- * amdgpu_drv_delayed_reset_work_handler - work handler for reset
- *
- * @work: work_struct.
- */
-static void amdgpu_drv_delayed_reset_work_handler(struct work_struct *work)
-{
-	struct list_head device_list;
-	struct amdgpu_device *adev;
-	int i, r;
-	struct amdgpu_reset_context reset_context;
-
-	memset(&reset_context, 0, sizeof(reset_context));
-
-	mutex_lock(&mgpu_info.mutex);
-	if (mgpu_info.pending_reset == true) {
-		mutex_unlock(&mgpu_info.mutex);
-		return;
-	}
-	mgpu_info.pending_reset = true;
-	mutex_unlock(&mgpu_info.mutex);
-
-	/* Use a common context, just need to make sure full reset is done */
-	reset_context.method = AMD_RESET_METHOD_NONE;
-	set_bit(AMDGPU_NEED_FULL_RESET, &reset_context.flags);
-
-	for (i = 0; i < mgpu_info.num_dgpu; i++) {
-		adev = mgpu_info.gpu_ins[i].adev;
-		reset_context.reset_req_dev = adev;
-		r = amdgpu_device_pre_asic_reset(adev, &reset_context);
-		if (r) {
-			dev_err(adev->dev, "GPU pre asic reset failed with err, %d for drm dev, %s ",
-				r, adev_to_drm(adev)->unique);
-		}
-		if (!queue_work(system_unbound_wq, &adev->xgmi_reset_work))
-			r = -EALREADY;
-	}
-	for (i = 0; i < mgpu_info.num_dgpu; i++) {
-		adev = mgpu_info.gpu_ins[i].adev;
-		flush_work(&adev->xgmi_reset_work);
-		adev->gmc.xgmi.pending_reset = false;
-	}
-
-	/* reset function will rebuild the xgmi hive info , clear it now */
-	for (i = 0; i < mgpu_info.num_dgpu; i++)
-		amdgpu_xgmi_remove_device(mgpu_info.gpu_ins[i].adev);
-
-	INIT_LIST_HEAD(&device_list);
-
-	for (i = 0; i < mgpu_info.num_dgpu; i++)
-		list_add_tail(&mgpu_info.gpu_ins[i].adev->reset_list, &device_list);
-
-	/* unregister the GPU first, reset function will add them back */
-	list_for_each_entry(adev, &device_list, reset_list)
-		amdgpu_unregister_gpu_instance(adev);
-
-	/* Use a common context, just need to make sure full reset is done */
-	set_bit(AMDGPU_SKIP_HW_RESET, &reset_context.flags);
-	r = amdgpu_do_asic_reset(&device_list, &reset_context);
-
-	if (r) {
-		DRM_ERROR("reinit gpus failure");
-		return;
-	}
-	for (i = 0; i < mgpu_info.num_dgpu; i++) {
-		adev = mgpu_info.gpu_ins[i].adev;
-		if (!adev->kfd.init_complete)
-			amdgpu_amdkfd_device_init(adev);
-		amdgpu_ttm_set_buffer_funcs_status(adev, true);
-	}
-	return;
-}
-
 static int amdgpu_pmops_prepare(struct device *dev)
 {
 	struct drm_device *drm_dev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = drm_to_adev(drm_dev);
+
+	/* device maybe not resumed here, return immediately in this case */
+	if (adev->in_s4 && adev->in_suspend)
+		return 0;
 
 	/* Return a positive number here so
 	 * DPM_FLAG_SMART_SUSPEND works properly
 	 */
-	if (amdgpu_device_supports_boco(drm_dev))
-		return pm_runtime_suspended(dev) &&
-			pm_suspend_via_firmware();
+	if (amdgpu_device_supports_boco(adev) && pm_runtime_suspended(dev))
+		return 1;
 
-	return 0;
+	/* if we will not support s3 or s2i for the device
+	 *  then skip suspend
+	 */
+	if (!amdgpu_acpi_is_s0ix_active(adev) &&
+	    !amdgpu_acpi_is_s3_active(adev))
+		return 1;
+
+	return amdgpu_device_prepare(drm_dev);
 }
 
 static void amdgpu_pmops_complete(struct device *dev)
 {
-	/* nothing to do */
+	amdgpu_device_complete(dev_get_drvdata(dev));
 }
 
 static int amdgpu_pmops_suspend(struct device *dev)
 {
 	struct drm_device *drm_dev = dev_get_drvdata(dev);
 	struct amdgpu_device *adev = drm_to_adev(drm_dev);
-	int r;
 
 	if (amdgpu_acpi_is_s0ix_active(adev))
 		adev->in_s0ix = true;
-	adev->in_s3 = true;
-	r = amdgpu_device_suspend(drm_dev, true);
-	adev->in_s3 = false;
+	else if (amdgpu_acpi_is_s3_active(adev))
+		adev->in_s3 = true;
+	if (!adev->in_s0ix && !adev->in_s3) {
+#if IS_ENABLED(CONFIG_SUSPEND)
+		/* don't allow going deep first time followed by s2idle the next time */
+		if (adev->last_suspend_state != PM_SUSPEND_ON &&
+		    adev->last_suspend_state != pm_suspend_target_state) {
+			drm_err_once(drm_dev, "Unsupported suspend state %d\n",
+				     pm_suspend_target_state);
+			return -EINVAL;
+		}
+#endif
+		return 0;
+	}
 
-	return r;
+#if IS_ENABLED(CONFIG_SUSPEND)
+	/* cache the state last used for suspend */
+	adev->last_suspend_state = pm_suspend_target_state;
+#endif
+
+	return amdgpu_device_suspend(drm_dev, true);
+}
+
+static int amdgpu_pmops_suspend_noirq(struct device *dev)
+{
+	struct drm_device *drm_dev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = drm_to_adev(drm_dev);
+	int r;
+
+	if (amdgpu_acpi_should_gpu_reset(adev)) {
+		amdgpu_device_lock_reset_domain(adev->reset_domain);
+		r = amdgpu_asic_reset(adev);
+		amdgpu_device_unlock_reset_domain(adev->reset_domain);
+		return r;
+	}
+
+	return 0;
 }
 
 static int amdgpu_pmops_resume(struct device *dev)
@@ -2163,6 +2669,9 @@ static int amdgpu_pmops_resume(struct device *dev)
 	struct amdgpu_device *adev = drm_to_adev(drm_dev);
 	int r;
 
+	if (!adev->in_s0ix && !adev->in_s3)
+		return 0;
+
 	/* Avoids registers access if device is physically gone */
 	if (!pci_device_is_present(adev->pdev))
 		adev->no_hw_access = true;
@@ -2170,6 +2679,8 @@ static int amdgpu_pmops_resume(struct device *dev)
 	r = amdgpu_device_resume(drm_dev, true);
 	if (amdgpu_acpi_is_s0ix_active(adev))
 		adev->in_s0ix = false;
+	else
+		adev->in_s3 = false;
 	return r;
 }
 
@@ -2179,24 +2690,34 @@ static int amdgpu_pmops_freeze(struct device *dev)
 	struct amdgpu_device *adev = drm_to_adev(drm_dev);
 	int r;
 
-	adev->in_s4 = true;
 	r = amdgpu_device_suspend(drm_dev, true);
-	adev->in_s4 = false;
 	if (r)
 		return r;
-	return amdgpu_asic_reset(adev);
+
+	if (amdgpu_acpi_should_gpu_reset(adev))
+		return amdgpu_asic_reset(adev);
+	return 0;
 }
 
 static int amdgpu_pmops_thaw(struct device *dev)
 {
 	struct drm_device *drm_dev = dev_get_drvdata(dev);
 
+	/* do not resume device if it's normal hibernation */
+	if (!pm_hibernate_is_recovering() && !pm_hibernation_mode_is_suspend())
+		return 0;
+
 	return amdgpu_device_resume(drm_dev, true);
 }
 
 static int amdgpu_pmops_poweroff(struct device *dev)
 {
 	struct drm_device *drm_dev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = drm_to_adev(drm_dev);
+
+	/* device maybe not resumed here, return immediately in this case */
+	if (adev->in_s4 && adev->in_suspend)
+		return 0;
 
 	return amdgpu_device_suspend(drm_dev, true);
 }
@@ -2208,6 +2729,82 @@ static int amdgpu_pmops_restore(struct device *dev)
 	return amdgpu_device_resume(drm_dev, true);
 }
 
+static int amdgpu_runtime_idle_check_display(struct device *dev)
+{
+	struct pci_dev *pdev = to_pci_dev(dev);
+	struct drm_device *drm_dev = pci_get_drvdata(pdev);
+	struct amdgpu_device *adev = drm_to_adev(drm_dev);
+
+	if (adev->mode_info.num_crtc) {
+		struct drm_connector *list_connector;
+		struct drm_connector_list_iter iter;
+		int ret = 0;
+
+		if (amdgpu_runtime_pm != -2) {
+			/* XXX: Return busy if any displays are connected to avoid
+			 * possible display wakeups after runtime resume due to
+			 * hotplug events in case any displays were connected while
+			 * the GPU was in suspend.  Remove this once that is fixed.
+			 */
+			mutex_lock(&drm_dev->mode_config.mutex);
+			drm_connector_list_iter_begin(drm_dev, &iter);
+			drm_for_each_connector_iter(list_connector, &iter) {
+				if (list_connector->status == connector_status_connected) {
+					ret = -EBUSY;
+					break;
+				}
+			}
+			drm_connector_list_iter_end(&iter);
+			mutex_unlock(&drm_dev->mode_config.mutex);
+
+			if (ret)
+				return ret;
+		}
+
+		if (adev->dc_enabled) {
+			struct drm_crtc *crtc;
+
+			drm_for_each_crtc(crtc, drm_dev) {
+				drm_modeset_lock(&crtc->mutex, NULL);
+				if (crtc->state->active)
+					ret = -EBUSY;
+				drm_modeset_unlock(&crtc->mutex);
+				if (ret < 0)
+					break;
+			}
+		} else {
+			mutex_lock(&drm_dev->mode_config.mutex);
+			drm_modeset_lock(&drm_dev->mode_config.connection_mutex, NULL);
+
+			drm_connector_list_iter_begin(drm_dev, &iter);
+			drm_for_each_connector_iter(list_connector, &iter) {
+				if (list_connector->dpms ==  DRM_MODE_DPMS_ON) {
+					ret = -EBUSY;
+					break;
+				}
+			}
+
+			drm_connector_list_iter_end(&iter);
+
+			drm_modeset_unlock(&drm_dev->mode_config.connection_mutex);
+			mutex_unlock(&drm_dev->mode_config.mutex);
+		}
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+static int amdgpu_runtime_idle_check_userq(struct device *dev)
+{
+	struct pci_dev *pdev = to_pci_dev(dev);
+	struct drm_device *drm_dev = pci_get_drvdata(pdev);
+	struct amdgpu_device *adev = drm_to_adev(drm_dev);
+
+	return xa_empty(&adev->userq_doorbell_xa) ? 0 : -EBUSY;
+}
+
 static int amdgpu_pmops_runtime_suspend(struct device *dev)
 {
 	struct pci_dev *pdev = to_pci_dev(dev);
@@ -2215,14 +2812,22 @@ static int amdgpu_pmops_runtime_suspend(struct device *dev)
 	struct amdgpu_device *adev = drm_to_adev(drm_dev);
 	int ret, i;
 
-	if (!adev->runpm) {
+	if (adev->pm.rpm_mode == AMDGPU_RUNPM_NONE) {
 		pm_runtime_forbid(dev);
 		return -EBUSY;
 	}
 
+	ret = amdgpu_runtime_idle_check_display(dev);
+	if (ret)
+		return ret;
+	ret = amdgpu_runtime_idle_check_userq(dev);
+	if (ret)
+		return ret;
+
 	/* wait for all rings to drain before suspending */
 	for (i = 0; i < AMDGPU_MAX_RINGS; i++) {
 		struct amdgpu_ring *ring = adev->rings[i];
+
 		if (ring && ring->sched.ready) {
 			ret = amdgpu_fence_wait_empty(ring);
 			if (ret)
@@ -2231,16 +2836,34 @@ static int amdgpu_pmops_runtime_suspend(struct device *dev)
 	}
 
 	adev->in_runpm = true;
-	if (amdgpu_device_supports_px(drm_dev))
+	if (adev->pm.rpm_mode == AMDGPU_RUNPM_PX)
 		drm_dev->switch_power_state = DRM_SWITCH_POWER_CHANGING;
 
+	/*
+	 * By setting mp1_state as PP_MP1_STATE_UNLOAD, MP1 will do some
+	 * proper cleanups and put itself into a state ready for PNP. That
+	 * can address some random resuming failure observed on BOCO capable
+	 * platforms.
+	 * TODO: this may be also needed for PX capable platform.
+	 */
+	if (adev->pm.rpm_mode == AMDGPU_RUNPM_BOCO)
+		adev->mp1_state = PP_MP1_STATE_UNLOAD;
+
+	ret = amdgpu_device_prepare(drm_dev);
+	if (ret)
+		return ret;
 	ret = amdgpu_device_suspend(drm_dev, false);
 	if (ret) {
 		adev->in_runpm = false;
+		if (adev->pm.rpm_mode == AMDGPU_RUNPM_BOCO)
+			adev->mp1_state = PP_MP1_STATE_NONE;
 		return ret;
 	}
 
-	if (amdgpu_device_supports_px(drm_dev)) {
+	if (adev->pm.rpm_mode == AMDGPU_RUNPM_BOCO)
+		adev->mp1_state = PP_MP1_STATE_NONE;
+
+	if (adev->pm.rpm_mode == AMDGPU_RUNPM_PX) {
 		/* Only need to handle PCI state in the driver for ATPX
 		 * PCI core handles it for _PR3.
 		 */
@@ -2249,12 +2872,15 @@ static int amdgpu_pmops_runtime_suspend(struct device *dev)
 		pci_ignore_hotplug(pdev);
 		pci_set_power_state(pdev, PCI_D3cold);
 		drm_dev->switch_power_state = DRM_SWITCH_POWER_DYNAMIC_OFF;
-	} else if (amdgpu_device_supports_boco(drm_dev)) {
+	} else if (adev->pm.rpm_mode == AMDGPU_RUNPM_BOCO) {
 		/* nothing to do */
-	} else if (amdgpu_device_supports_baco(drm_dev)) {
-		amdgpu_device_baco_enter(drm_dev);
+	} else if ((adev->pm.rpm_mode == AMDGPU_RUNPM_BACO) ||
+			(adev->pm.rpm_mode == AMDGPU_RUNPM_BAMACO)) {
+		amdgpu_device_baco_enter(adev);
 	}
 
+	dev_dbg(&pdev->dev, "asic/device is runtime suspended\n");
+
 	return 0;
 }
 
@@ -2265,14 +2891,14 @@ static int amdgpu_pmops_runtime_resume(struct device *dev)
 	struct amdgpu_device *adev = drm_to_adev(drm_dev);
 	int ret;
 
-	if (!adev->runpm)
+	if (adev->pm.rpm_mode == AMDGPU_RUNPM_NONE)
 		return -EINVAL;
 
 	/* Avoids registers access if device is physically gone */
 	if (!pci_device_is_present(adev->pdev))
 		adev->no_hw_access = true;
 
-	if (amdgpu_device_supports_px(drm_dev)) {
+	if (adev->pm.rpm_mode == AMDGPU_RUNPM_PX) {
 		drm_dev->switch_power_state = DRM_SWITCH_POWER_CHANGING;
 
 		/* Only need to handle PCI state in the driver for ATPX
@@ -2284,19 +2910,23 @@ static int amdgpu_pmops_runtime_resume(struct device *dev)
 		if (ret)
 			return ret;
 		pci_set_master(pdev);
-	} else if (amdgpu_device_supports_boco(drm_dev)) {
+	} else if (adev->pm.rpm_mode == AMDGPU_RUNPM_BOCO) {
 		/* Only need to handle PCI state in the driver for ATPX
 		 * PCI core handles it for _PR3.
 		 */
 		pci_set_master(pdev);
-	} else if (amdgpu_device_supports_baco(drm_dev)) {
-		amdgpu_device_baco_exit(drm_dev);
+	} else if ((adev->pm.rpm_mode == AMDGPU_RUNPM_BACO) ||
+			(adev->pm.rpm_mode == AMDGPU_RUNPM_BAMACO)) {
+		amdgpu_device_baco_exit(adev);
 	}
 	ret = amdgpu_device_resume(drm_dev, false);
-	if (ret)
+	if (ret) {
+		if (adev->pm.rpm_mode == AMDGPU_RUNPM_PX)
+			pci_disable_device(pdev);
 		return ret;
+	}
 
-	if (amdgpu_device_supports_px(drm_dev))
+	if (adev->pm.rpm_mode == AMDGPU_RUNPM_PX)
 		drm_dev->switch_power_state = DRM_SWITCH_POWER_ON;
 	adev->in_runpm = false;
 	return 0;
@@ -2306,53 +2936,38 @@ static int amdgpu_pmops_runtime_idle(struct device *dev)
 {
 	struct drm_device *drm_dev = dev_get_drvdata(dev);
 	struct amdgpu_device *adev = drm_to_adev(drm_dev);
-	/* we don't want the main rpm_idle to call suspend - we want to autosuspend */
-	int ret = 1;
+	int ret;
 
-	if (!adev->runpm) {
+	if (adev->pm.rpm_mode == AMDGPU_RUNPM_NONE) {
 		pm_runtime_forbid(dev);
 		return -EBUSY;
 	}
 
-	if (amdgpu_device_has_dc_support(adev)) {
-		struct drm_crtc *crtc;
-
-		drm_for_each_crtc(crtc, drm_dev) {
-			drm_modeset_lock(&crtc->mutex, NULL);
-			if (crtc->state->active)
-				ret = -EBUSY;
-			drm_modeset_unlock(&crtc->mutex);
-			if (ret < 0)
-				break;
-		}
-
-	} else {
-		struct drm_connector *list_connector;
-		struct drm_connector_list_iter iter;
-
-		mutex_lock(&drm_dev->mode_config.mutex);
-		drm_modeset_lock(&drm_dev->mode_config.connection_mutex, NULL);
-
-		drm_connector_list_iter_begin(drm_dev, &iter);
-		drm_for_each_connector_iter(list_connector, &iter) {
-			if (list_connector->dpms ==  DRM_MODE_DPMS_ON) {
-				ret = -EBUSY;
-				break;
-			}
-		}
+	ret = amdgpu_runtime_idle_check_display(dev);
+	if (ret)
+		goto done;
 
-		drm_connector_list_iter_end(&iter);
+	ret = amdgpu_runtime_idle_check_userq(dev);
+done:
+	pm_runtime_autosuspend(dev);
+	return ret;
+}
 
-		drm_modeset_unlock(&drm_dev->mode_config.connection_mutex);
-		mutex_unlock(&drm_dev->mode_config.mutex);
+static int amdgpu_drm_release(struct inode *inode, struct file *filp)
+{
+	struct drm_file *file_priv = filp->private_data;
+	struct amdgpu_fpriv *fpriv = file_priv->driver_priv;
+	struct drm_device *dev = file_priv->minor->dev;
+	int idx;
+
+	if (fpriv && drm_dev_enter(dev, &idx)) {
+		fpriv->evf_mgr.fd_closing = true;
+		amdgpu_eviction_fence_destroy(&fpriv->evf_mgr);
+		amdgpu_userq_mgr_fini(&fpriv->userq_mgr);
+		drm_dev_exit(idx);
 	}
 
-	if (ret == -EBUSY)
-		DRM_DEBUG_DRIVER("failing to power off - crtc active\n");
-
-	pm_runtime_mark_last_busy(dev);
-	pm_runtime_autosuspend(dev);
-	return ret;
+	return drm_release(inode, filp);
 }
 
 long amdgpu_drm_ioctl(struct file *filp,
@@ -2361,6 +2976,7 @@ long amdgpu_drm_ioctl(struct file *filp,
 	struct drm_file *file_priv = filp->private_data;
 	struct drm_device *dev;
 	long ret;
+
 	dev = file_priv->minor->dev;
 	ret = pm_runtime_get_sync(dev->dev);
 	if (ret < 0)
@@ -2368,21 +2984,21 @@ long amdgpu_drm_ioctl(struct file *filp,
 
 	ret = drm_ioctl(filp, cmd, arg);
 
-	pm_runtime_mark_last_busy(dev->dev);
 out:
 	pm_runtime_put_autosuspend(dev->dev);
 	return ret;
 }
 
 static const struct dev_pm_ops amdgpu_pm_ops = {
-	.prepare = amdgpu_pmops_prepare,
-	.complete = amdgpu_pmops_complete,
-	.suspend = amdgpu_pmops_suspend,
-	.resume = amdgpu_pmops_resume,
-	.freeze = amdgpu_pmops_freeze,
-	.thaw = amdgpu_pmops_thaw,
-	.poweroff = amdgpu_pmops_poweroff,
-	.restore = amdgpu_pmops_restore,
+	.prepare = pm_sleep_ptr(amdgpu_pmops_prepare),
+	.complete = pm_sleep_ptr(amdgpu_pmops_complete),
+	.suspend = pm_sleep_ptr(amdgpu_pmops_suspend),
+	.suspend_noirq = pm_sleep_ptr(amdgpu_pmops_suspend_noirq),
+	.resume = pm_sleep_ptr(amdgpu_pmops_resume),
+	.freeze = pm_sleep_ptr(amdgpu_pmops_freeze),
+	.thaw = pm_sleep_ptr(amdgpu_pmops_thaw),
+	.poweroff = pm_sleep_ptr(amdgpu_pmops_poweroff),
+	.restore = pm_sleep_ptr(amdgpu_pmops_restore),
 	.runtime_suspend = amdgpu_pmops_runtime_suspend,
 	.runtime_resume = amdgpu_pmops_runtime_resume,
 	.runtime_idle = amdgpu_pmops_runtime_idle,
@@ -2404,7 +3020,7 @@ static const struct file_operations amdgpu_driver_kms_fops = {
 	.owner = THIS_MODULE,
 	.open = drm_open,
 	.flush = amdgpu_flush,
-	.release = drm_release,
+	.release = amdgpu_drm_release,
 	.unlocked_ioctl = amdgpu_drm_ioctl,
 	.mmap = drm_gem_mmap,
 	.poll = drm_poll,
@@ -2413,8 +3029,9 @@ static const struct file_operations amdgpu_driver_kms_fops = {
 	.compat_ioctl = amdgpu_kms_compat_ioctl,
 #endif
 #ifdef CONFIG_PROC_FS
-	.show_fdinfo = amdgpu_show_fdinfo
+	.show_fdinfo = drm_show_fdinfo,
 #endif
+	.fop_flags = FOP_UNSIGNED_OFFSET,
 };
 
 int amdgpu_file_to_fpriv(struct file *filp, struct amdgpu_fpriv **fpriv)
@@ -2424,9 +3041,8 @@ int amdgpu_file_to_fpriv(struct file *filp, struct amdgpu_fpriv **fpriv)
 	if (!filp)
 		return -EINVAL;
 
-	if (filp->f_op != &amdgpu_driver_kms_fops) {
+	if (filp->f_op != &amdgpu_driver_kms_fops)
 		return -EINVAL;
-	}
 
 	file = filp->private_data;
 	*fpriv = file->driver_priv;
@@ -2451,6 +3067,10 @@ const struct drm_ioctl_desc amdgpu_ioctls_kms[] = {
 	DRM_IOCTL_DEF_DRV(AMDGPU_GEM_VA, amdgpu_gem_va_ioctl, DRM_AUTH|DRM_RENDER_ALLOW),
 	DRM_IOCTL_DEF_DRV(AMDGPU_GEM_OP, amdgpu_gem_op_ioctl, DRM_AUTH|DRM_RENDER_ALLOW),
 	DRM_IOCTL_DEF_DRV(AMDGPU_GEM_USERPTR, amdgpu_gem_userptr_ioctl, DRM_AUTH|DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(AMDGPU_USERQ, amdgpu_userq_ioctl, DRM_AUTH|DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(AMDGPU_USERQ_SIGNAL, amdgpu_userq_signal_ioctl, DRM_AUTH|DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(AMDGPU_USERQ_WAIT, amdgpu_userq_wait_ioctl, DRM_AUTH|DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(AMDGPU_GEM_LIST_HANDLES, amdgpu_gem_list_handles_ioctl, DRM_AUTH|DRM_RENDER_ALLOW),
 };
 
 static const struct drm_driver amdgpu_kms_driver = {
@@ -2461,22 +3081,44 @@ static const struct drm_driver amdgpu_kms_driver = {
 	    DRIVER_SYNCOBJ_TIMELINE,
 	.open = amdgpu_driver_open_kms,
 	.postclose = amdgpu_driver_postclose_kms,
-	.lastclose = amdgpu_driver_lastclose_kms,
 	.ioctls = amdgpu_ioctls_kms,
 	.num_ioctls = ARRAY_SIZE(amdgpu_ioctls_kms),
 	.dumb_create = amdgpu_mode_dumb_create,
 	.dumb_map_offset = amdgpu_mode_dumb_mmap,
+	DRM_FBDEV_TTM_DRIVER_OPS,
+	.fops = &amdgpu_driver_kms_fops,
+	.release = &amdgpu_driver_release_kms,
+#ifdef CONFIG_PROC_FS
+	.show_fdinfo = amdgpu_show_fdinfo,
+#endif
+
+	.gem_prime_import = amdgpu_gem_prime_import,
+
+	.name = DRIVER_NAME,
+	.desc = DRIVER_DESC,
+	.major = KMS_DRIVER_MAJOR,
+	.minor = KMS_DRIVER_MINOR,
+	.patchlevel = KMS_DRIVER_PATCHLEVEL,
+};
+
+const struct drm_driver amdgpu_partition_driver = {
+	.driver_features =
+	    DRIVER_GEM | DRIVER_RENDER | DRIVER_SYNCOBJ |
+	    DRIVER_SYNCOBJ_TIMELINE,
+	.open = amdgpu_driver_open_kms,
+	.postclose = amdgpu_driver_postclose_kms,
+	.ioctls = amdgpu_ioctls_kms,
+	.num_ioctls = ARRAY_SIZE(amdgpu_ioctls_kms),
+	.dumb_create = amdgpu_mode_dumb_create,
+	.dumb_map_offset = amdgpu_mode_dumb_mmap,
+	DRM_FBDEV_TTM_DRIVER_OPS,
 	.fops = &amdgpu_driver_kms_fops,
 	.release = &amdgpu_driver_release_kms,
 
-	.prime_handle_to_fd = drm_gem_prime_handle_to_fd,
-	.prime_fd_to_handle = drm_gem_prime_fd_to_handle,
 	.gem_prime_import = amdgpu_gem_prime_import,
-	.gem_prime_mmap = drm_gem_prime_mmap,
 
 	.name = DRIVER_NAME,
 	.desc = DRIVER_DESC,
-	.date = DRIVER_DATE,
 	.major = KMS_DRIVER_MAJOR,
 	.minor = KMS_DRIVER_MINOR,
 	.patchlevel = KMS_DRIVER_PATCHLEVEL,
@@ -2489,25 +3131,20 @@ static struct pci_error_handlers amdgpu_pci_err_handler = {
 	.resume		= amdgpu_pci_resume,
 };
 
-extern const struct attribute_group amdgpu_vram_mgr_attr_group;
-extern const struct attribute_group amdgpu_gtt_mgr_attr_group;
-extern const struct attribute_group amdgpu_vbios_version_attr_group;
-
 static const struct attribute_group *amdgpu_sysfs_groups[] = {
 	&amdgpu_vram_mgr_attr_group,
 	&amdgpu_gtt_mgr_attr_group,
-	&amdgpu_vbios_version_attr_group,
+	&amdgpu_flash_attr_group,
 	NULL,
 };
 
-
 static struct pci_driver amdgpu_kms_pci_driver = {
 	.name = DRIVER_NAME,
 	.id_table = pciidlist,
 	.probe = amdgpu_pci_probe,
 	.remove = amdgpu_pci_remove,
 	.shutdown = amdgpu_pci_shutdown,
-	.driver.pm = &amdgpu_pm_ops,
+	.driver.pm = pm_ptr(&amdgpu_pm_ops),
 	.err_handler = &amdgpu_pci_err_handler,
 	.dev_groups = amdgpu_sysfs_groups,
 };
@@ -2516,16 +3153,11 @@ static int __init amdgpu_init(void)
 {
 	int r;
 
-	if (vgacon_text_force()) {
-		DRM_ERROR("VGACON disables amdgpu kernel modesetting.\n");
-		return -EINVAL;
-	}
-
 	r = amdgpu_sync_init();
 	if (r)
 		goto error_sync;
 
-	r = amdgpu_fence_slab_init();
+	r = amdgpu_userq_fence_slab_init();
 	if (r)
 		goto error_fence;
 
@@ -2536,6 +3168,12 @@ static int __init amdgpu_init(void)
 	/* Ignore KFD init failures. Normal when CONFIG_HSA_AMD is not set. */
 	amdgpu_amdkfd_init();
 
+	if (amdgpu_pp_feature_mask & PP_OVERDRIVE_MASK) {
+		add_taint(TAINT_CPU_OUT_OF_SPEC, LOCKDEP_STILL_OK);
+		pr_crit("Overdrive is enabled, please disable it before "
+			"reporting any bugs unrelated to overdrive.\n");
+	}
+
 	/* let modprobe override vga console setting */
 	return pci_register_driver(&amdgpu_kms_pci_driver);
 
@@ -2551,9 +3189,11 @@ static void __exit amdgpu_exit(void)
 	amdgpu_amdkfd_fini();
 	pci_unregister_driver(&amdgpu_kms_pci_driver);
 	amdgpu_unregister_atpx_handler();
+	amdgpu_acpi_release();
 	amdgpu_sync_fini();
-	amdgpu_fence_slab_fini();
+	amdgpu_userq_fence_slab_fini();
 	mmu_notifier_synchronize();
+	amdgpu_xcp_drv_release();
 }
 
 module_init(amdgpu_init);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.h
index e3a4f7048042..2d86cc6f7f4d 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.h
@@ -40,9 +40,13 @@
 
 #define DRIVER_NAME		"amdgpu"
 #define DRIVER_DESC		"AMD GPU"
-#define DRIVER_DATE		"20150101"
+
+extern const struct drm_driver amdgpu_partition_driver;
 
 long amdgpu_drm_ioctl(struct file *filp,
 		      unsigned int cmd, unsigned long arg);
 
+long amdgpu_kms_compat_ioctl(struct file *filp,
+			     unsigned int cmd, unsigned long arg);
+
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_eeprom.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_eeprom.c
index 4d9eb0137f8c..8cd69836dd99 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_eeprom.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_eeprom.c
@@ -79,16 +79,18 @@
  * That is, for an I2C EEPROM driver everything is controlled by
  * the "eeprom_addr".
  *
+ * See also top of amdgpu_ras_eeprom.c.
+ *
  * P.S. If you need to write, lock and read the Identification Page,
  * (M24M02-DR device only, which we do not use), change the "7" to
  * "0xF" in the macro below, and let the client set bit 20 to 1 in
  * "eeprom_addr", and set A10 to 0 to write into it, and A10 and A1 to
  * 1 to lock it permanently.
  */
-#define MAKE_I2C_ADDR(_aa) ((0xA << 3) | (((_aa) >> 16) & 7))
+#define MAKE_I2C_ADDR(_aa) ((0xA << 3) | (((_aa) >> 16) & 0xF))
 
 static int __amdgpu_eeprom_xfer(struct i2c_adapter *i2c_adap, u32 eeprom_addr,
-				u8 *eeprom_buf, u16 buf_size, bool read)
+				u8 *eeprom_buf, u32 buf_size, bool read)
 {
 	u8 eeprom_offset_buf[EEPROM_OFFSET_SIZE];
 	struct i2c_msg msgs[] = {
@@ -131,15 +133,15 @@ static int __amdgpu_eeprom_xfer(struct i2c_adapter *i2c_adap, u32 eeprom_addr,
 			 * cycle begins. This is implied for the
 			 * "i2c_transfer()" abstraction.
 			 */
-			len = min(EEPROM_PAGE_SIZE - (eeprom_addr &
-						      EEPROM_PAGE_MASK),
-				  (u32)buf_size);
+			len = min(EEPROM_PAGE_SIZE - (eeprom_addr & EEPROM_PAGE_MASK),
+					buf_size);
 		} else {
 			/* Reading from the EEPROM has no limitation
 			 * on the number of bytes read from the EEPROM
 			 * device--they are simply sequenced out.
+			 * Keep in mind that i2c_msg.len is u16 type.
 			 */
-			len = buf_size;
+			len = min(U16_MAX, buf_size);
 		}
 		msgs[1].len = len;
 		msgs[1].buf = eeprom_buf;
@@ -177,10 +179,12 @@ static int __amdgpu_eeprom_xfer(struct i2c_adapter *i2c_adap, u32 eeprom_addr,
  * Returns the number of bytes read/written; -errno on error.
  */
 static int amdgpu_eeprom_xfer(struct i2c_adapter *i2c_adap, u32 eeprom_addr,
-			      u8 *eeprom_buf, u16 buf_size, bool read)
+			      u8 *eeprom_buf, u32 buf_size, bool read)
 {
 	const struct i2c_adapter_quirks *quirks = i2c_adap->quirks;
 	u16 limit;
+	u16 ps; /* Partial size */
+	int res = 0, r;
 
 	if (!quirks)
 		limit = 0;
@@ -196,35 +200,32 @@ static int amdgpu_eeprom_xfer(struct i2c_adapter *i2c_adap, u32 eeprom_addr,
 		dev_err_ratelimited(&i2c_adap->dev,
 				    "maddr:0x%04X size:0x%02X:quirk max_%s_len must be > %d",
 				    eeprom_addr, buf_size,
-				    read ? "read" : "write", EEPROM_OFFSET_SIZE);
+				    str_read_write(read), EEPROM_OFFSET_SIZE);
 		return -EINVAL;
-	} else {
-		u16 ps; /* Partial size */
-		int res = 0, r;
-
-		/* The "limit" includes all data bytes sent/received,
-		 * which would include the EEPROM_OFFSET_SIZE bytes.
-		 * Account for them here.
-		 */
-		limit -= EEPROM_OFFSET_SIZE;
-		for ( ; buf_size > 0;
-		      buf_size -= ps, eeprom_addr += ps, eeprom_buf += ps) {
-			ps = min(limit, buf_size);
-
-			r = __amdgpu_eeprom_xfer(i2c_adap, eeprom_addr,
-						 eeprom_buf, ps, read);
-			if (r < 0)
-				return r;
-			res += r;
-		}
+	}
 
-		return res;
+	/* The "limit" includes all data bytes sent/received,
+	 * which would include the EEPROM_OFFSET_SIZE bytes.
+	 * Account for them here.
+	 */
+	limit -= EEPROM_OFFSET_SIZE;
+	for ( ; buf_size > 0;
+	      buf_size -= ps, eeprom_addr += ps, eeprom_buf += ps) {
+		ps = min(limit, buf_size);
+
+		r = __amdgpu_eeprom_xfer(i2c_adap, eeprom_addr,
+					 eeprom_buf, ps, read);
+		if (r < 0)
+			return r;
+		res += r;
 	}
+
+	return res;
 }
 
 int amdgpu_eeprom_read(struct i2c_adapter *i2c_adap,
 		       u32 eeprom_addr, u8 *eeprom_buf,
-		       u16 bytes)
+		       u32 bytes)
 {
 	return amdgpu_eeprom_xfer(i2c_adap, eeprom_addr, eeprom_buf, bytes,
 				  true);
@@ -232,7 +233,7 @@ int amdgpu_eeprom_read(struct i2c_adapter *i2c_adap,
 
 int amdgpu_eeprom_write(struct i2c_adapter *i2c_adap,
 			u32 eeprom_addr, u8 *eeprom_buf,
-			u16 bytes)
+			u32 bytes)
 {
 	return amdgpu_eeprom_xfer(i2c_adap, eeprom_addr, eeprom_buf, bytes,
 				  false);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_eeprom.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_eeprom.h
index 6935adb2be1f..8083b8253ef4 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_eeprom.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_eeprom.h
@@ -28,10 +28,10 @@
 
 int amdgpu_eeprom_read(struct i2c_adapter *i2c_adap,
 		       u32 eeprom_addr, u8 *eeprom_buf,
-		       u16 bytes);
+		       u32 bytes);
 
 int amdgpu_eeprom_write(struct i2c_adapter *i2c_adap,
 			u32 eeprom_addr, u8 *eeprom_buf,
-			u16 bytes);
+			u32 bytes);
 
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_encoders.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_encoders.c
index af4ef84e27a7..3aaeed2d3562 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_encoders.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_encoders.c
@@ -24,7 +24,6 @@
  *          Alex Deucher
  */
 
-#include <drm/drm_crtc_helper.h>
 #include <drm/amdgpu_drm.h>
 #include "amdgpu.h"
 #include "amdgpu_connectors.h"
@@ -71,6 +70,7 @@ void amdgpu_encoder_set_active_device(struct drm_encoder *encoder)
 	drm_for_each_connector_iter(connector, &iter) {
 		if (connector->encoder == encoder) {
 			struct amdgpu_connector *amdgpu_connector = to_amdgpu_connector(connector);
+
 			amdgpu_encoder->active_device = amdgpu_encoder->devices & amdgpu_connector->devices;
 			DRM_DEBUG_KMS("setting active device to %08x from %08x %08x for encoder %d\n",
 				  amdgpu_encoder->active_device, amdgpu_encoder->devices,
@@ -166,12 +166,12 @@ void amdgpu_panel_mode_fixup(struct drm_encoder *encoder,
 {
 	struct amdgpu_encoder *amdgpu_encoder = to_amdgpu_encoder(encoder);
 	struct drm_display_mode *native_mode = &amdgpu_encoder->native_mode;
-	unsigned hblank = native_mode->htotal - native_mode->hdisplay;
-	unsigned vblank = native_mode->vtotal - native_mode->vdisplay;
-	unsigned hover = native_mode->hsync_start - native_mode->hdisplay;
-	unsigned vover = native_mode->vsync_start - native_mode->vdisplay;
-	unsigned hsync_width = native_mode->hsync_end - native_mode->hsync_start;
-	unsigned vsync_width = native_mode->vsync_end - native_mode->vsync_start;
+	unsigned int hblank = native_mode->htotal - native_mode->hdisplay;
+	unsigned int vblank = native_mode->vtotal - native_mode->vdisplay;
+	unsigned int hover = native_mode->hsync_start - native_mode->hdisplay;
+	unsigned int vover = native_mode->vsync_start - native_mode->vdisplay;
+	unsigned int hsync_width = native_mode->hsync_end - native_mode->hsync_start;
+	unsigned int vsync_width = native_mode->vsync_end - native_mode->vsync_start;
 
 	adjusted_mode->clock = native_mode->clock;
 	adjusted_mode->flags = native_mode->flags;
@@ -222,7 +222,7 @@ bool amdgpu_dig_monitor_is_duallink(struct drm_encoder *encoder,
 	case DRM_MODE_CONNECTOR_HDMIB:
 		if (amdgpu_connector->use_digital) {
 			/* HDMI 1.3 supports up to 340 Mhz over single link */
-			if (drm_detect_hdmi_monitor(amdgpu_connector_edid(connector))) {
+			if (connector->display_info.is_hdmi) {
 				if (pixel_clock > 340000)
 					return true;
 				else
@@ -244,7 +244,7 @@ bool amdgpu_dig_monitor_is_duallink(struct drm_encoder *encoder,
 			return false;
 		else {
 			/* HDMI 1.3 supports up to 340 Mhz over single link */
-			if (drm_detect_hdmi_monitor(amdgpu_connector_edid(connector))) {
+			if (connector->display_info.is_hdmi) {
 				if (pixel_clock > 340000)
 					return true;
 				else
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_eviction_fence.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_eviction_fence.c
new file mode 100644
index 000000000000..23d7d0b0d625
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_eviction_fence.c
@@ -0,0 +1,241 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright 2024 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include <linux/sched.h>
+#include <drm/drm_exec.h>
+#include "amdgpu.h"
+
+#define work_to_evf_mgr(w, name) container_of(w, struct amdgpu_eviction_fence_mgr, name)
+#define evf_mgr_to_fpriv(e) container_of(e, struct amdgpu_fpriv, evf_mgr)
+
+static const char *
+amdgpu_eviction_fence_get_driver_name(struct dma_fence *fence)
+{
+	return "amdgpu_eviction_fence";
+}
+
+static const char *
+amdgpu_eviction_fence_get_timeline_name(struct dma_fence *f)
+{
+	struct amdgpu_eviction_fence *ef;
+
+	ef = container_of(f, struct amdgpu_eviction_fence, base);
+	return ef->timeline_name;
+}
+
+int
+amdgpu_eviction_fence_replace_fence(struct amdgpu_eviction_fence_mgr *evf_mgr,
+				    struct drm_exec *exec)
+{
+	struct amdgpu_eviction_fence *old_ef, *new_ef;
+	struct drm_gem_object *obj;
+	unsigned long index;
+	int ret;
+
+	if (evf_mgr->ev_fence &&
+	    !dma_fence_is_signaled(&evf_mgr->ev_fence->base))
+		return 0;
+	/*
+	 * Steps to replace eviction fence:
+	 * * lock all objects in exec (caller)
+	 * * create a new eviction fence
+	 * * update new eviction fence in evf_mgr
+	 * * attach the new eviction fence to BOs
+	 * * release the old fence
+	 * * unlock the objects (caller)
+	 */
+	new_ef = amdgpu_eviction_fence_create(evf_mgr);
+	if (!new_ef) {
+		DRM_ERROR("Failed to create new eviction fence\n");
+		return -ENOMEM;
+	}
+
+	/* Update the eviction fence now */
+	spin_lock(&evf_mgr->ev_fence_lock);
+	old_ef = evf_mgr->ev_fence;
+	evf_mgr->ev_fence = new_ef;
+	spin_unlock(&evf_mgr->ev_fence_lock);
+
+	/* Attach the new fence */
+	drm_exec_for_each_locked_object(exec, index, obj) {
+		struct amdgpu_bo *bo = gem_to_amdgpu_bo(obj);
+
+		if (!bo)
+			continue;
+		ret = amdgpu_eviction_fence_attach(evf_mgr, bo);
+		if (ret) {
+			DRM_ERROR("Failed to attch new eviction fence\n");
+			goto free_err;
+		}
+	}
+
+	/* Free old fence */
+	if (old_ef)
+		dma_fence_put(&old_ef->base);
+	return 0;
+
+free_err:
+	kfree(new_ef);
+	return ret;
+}
+
+static void
+amdgpu_eviction_fence_suspend_worker(struct work_struct *work)
+{
+	struct amdgpu_eviction_fence_mgr *evf_mgr = work_to_evf_mgr(work, suspend_work.work);
+	struct amdgpu_fpriv *fpriv = evf_mgr_to_fpriv(evf_mgr);
+	struct amdgpu_userq_mgr *uq_mgr = &fpriv->userq_mgr;
+	struct amdgpu_eviction_fence *ev_fence;
+
+	mutex_lock(&uq_mgr->userq_mutex);
+	spin_lock(&evf_mgr->ev_fence_lock);
+	ev_fence = evf_mgr->ev_fence;
+	if (ev_fence)
+		dma_fence_get(&ev_fence->base);
+	else
+		goto unlock;
+	spin_unlock(&evf_mgr->ev_fence_lock);
+
+	amdgpu_userq_evict(uq_mgr, ev_fence);
+
+	mutex_unlock(&uq_mgr->userq_mutex);
+	dma_fence_put(&ev_fence->base);
+	return;
+
+unlock:
+	spin_unlock(&evf_mgr->ev_fence_lock);
+	mutex_unlock(&uq_mgr->userq_mutex);
+}
+
+static bool amdgpu_eviction_fence_enable_signaling(struct dma_fence *f)
+{
+	struct amdgpu_eviction_fence_mgr *evf_mgr;
+	struct amdgpu_eviction_fence *ev_fence;
+
+	if (!f)
+		return true;
+
+	ev_fence = to_ev_fence(f);
+	evf_mgr = ev_fence->evf_mgr;
+
+	schedule_delayed_work(&evf_mgr->suspend_work, 0);
+	return true;
+}
+
+static const struct dma_fence_ops amdgpu_eviction_fence_ops = {
+	.get_driver_name = amdgpu_eviction_fence_get_driver_name,
+	.get_timeline_name = amdgpu_eviction_fence_get_timeline_name,
+	.enable_signaling = amdgpu_eviction_fence_enable_signaling,
+};
+
+void amdgpu_eviction_fence_signal(struct amdgpu_eviction_fence_mgr *evf_mgr,
+				  struct amdgpu_eviction_fence *ev_fence)
+{
+	spin_lock(&evf_mgr->ev_fence_lock);
+	dma_fence_signal(&ev_fence->base);
+	spin_unlock(&evf_mgr->ev_fence_lock);
+}
+
+struct amdgpu_eviction_fence *
+amdgpu_eviction_fence_create(struct amdgpu_eviction_fence_mgr *evf_mgr)
+{
+	struct amdgpu_eviction_fence *ev_fence;
+
+	ev_fence = kzalloc(sizeof(*ev_fence), GFP_KERNEL);
+	if (!ev_fence)
+		return NULL;
+
+	ev_fence->evf_mgr = evf_mgr;
+	get_task_comm(ev_fence->timeline_name, current);
+	spin_lock_init(&ev_fence->lock);
+	dma_fence_init64(&ev_fence->base, &amdgpu_eviction_fence_ops,
+			 &ev_fence->lock, evf_mgr->ev_fence_ctx,
+			 atomic_inc_return(&evf_mgr->ev_fence_seq));
+	return ev_fence;
+}
+
+void amdgpu_eviction_fence_destroy(struct amdgpu_eviction_fence_mgr *evf_mgr)
+{
+	struct amdgpu_eviction_fence *ev_fence;
+
+	/* Wait for any pending work to execute */
+	flush_delayed_work(&evf_mgr->suspend_work);
+
+	spin_lock(&evf_mgr->ev_fence_lock);
+	ev_fence = evf_mgr->ev_fence;
+	spin_unlock(&evf_mgr->ev_fence_lock);
+
+	if (!ev_fence)
+		return;
+
+	dma_fence_wait(&ev_fence->base, false);
+
+	/* Last unref of ev_fence */
+	dma_fence_put(&ev_fence->base);
+}
+
+int amdgpu_eviction_fence_attach(struct amdgpu_eviction_fence_mgr *evf_mgr,
+				 struct amdgpu_bo *bo)
+{
+	struct amdgpu_eviction_fence *ev_fence;
+	struct dma_resv *resv = bo->tbo.base.resv;
+	int ret;
+
+	if (!resv)
+		return 0;
+
+	ret = dma_resv_reserve_fences(resv, 1);
+	if (ret) {
+		DRM_DEBUG_DRIVER("Failed to resv fence space\n");
+		return ret;
+	}
+
+	spin_lock(&evf_mgr->ev_fence_lock);
+	ev_fence = evf_mgr->ev_fence;
+	if (ev_fence)
+		dma_resv_add_fence(resv, &ev_fence->base, DMA_RESV_USAGE_BOOKKEEP);
+	spin_unlock(&evf_mgr->ev_fence_lock);
+
+	return 0;
+}
+
+void amdgpu_eviction_fence_detach(struct amdgpu_eviction_fence_mgr *evf_mgr,
+				  struct amdgpu_bo *bo)
+{
+	struct dma_fence *stub = dma_fence_get_stub();
+
+	dma_resv_replace_fences(bo->tbo.base.resv, evf_mgr->ev_fence_ctx,
+				stub, DMA_RESV_USAGE_BOOKKEEP);
+	dma_fence_put(stub);
+}
+
+int amdgpu_eviction_fence_init(struct amdgpu_eviction_fence_mgr *evf_mgr)
+{
+	/* This needs to be done one time per open */
+	atomic_set(&evf_mgr->ev_fence_seq, 0);
+	evf_mgr->ev_fence_ctx = dma_fence_context_alloc(1);
+	spin_lock_init(&evf_mgr->ev_fence_lock);
+
+	INIT_DELAYED_WORK(&evf_mgr->suspend_work, amdgpu_eviction_fence_suspend_worker);
+	return 0;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_eviction_fence.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_eviction_fence.h
new file mode 100644
index 000000000000..fcd867b7147d
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_eviction_fence.h
@@ -0,0 +1,69 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef AMDGPU_EV_FENCE_H_
+#define AMDGPU_EV_FENCE_H_
+
+struct amdgpu_eviction_fence {
+	struct dma_fence base;
+	spinlock_t	 lock;
+	char		 timeline_name[TASK_COMM_LEN];
+	struct amdgpu_eviction_fence_mgr *evf_mgr;
+};
+
+struct amdgpu_eviction_fence_mgr {
+	u64			ev_fence_ctx;
+	atomic_t		ev_fence_seq;
+	spinlock_t		ev_fence_lock;
+	struct amdgpu_eviction_fence *ev_fence;
+	struct delayed_work	suspend_work;
+	uint8_t fd_closing;
+};
+
+/* Eviction fence helper functions */
+struct amdgpu_eviction_fence *
+amdgpu_eviction_fence_create(struct amdgpu_eviction_fence_mgr *evf_mgr);
+
+void
+amdgpu_eviction_fence_destroy(struct amdgpu_eviction_fence_mgr *evf_mgr);
+
+int
+amdgpu_eviction_fence_attach(struct amdgpu_eviction_fence_mgr *evf_mgr,
+			     struct amdgpu_bo *bo);
+
+void
+amdgpu_eviction_fence_detach(struct amdgpu_eviction_fence_mgr *evf_mgr,
+			     struct amdgpu_bo *bo);
+
+int
+amdgpu_eviction_fence_init(struct amdgpu_eviction_fence_mgr *evf_mgr);
+
+void
+amdgpu_eviction_fence_signal(struct amdgpu_eviction_fence_mgr *evf_mgr,
+			     struct amdgpu_eviction_fence *ev_fence);
+
+int
+amdgpu_eviction_fence_replace_fence(struct amdgpu_eviction_fence_mgr *evf_mgr,
+				    struct drm_exec *exec);
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_fb.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_fb.c
deleted file mode 100644
index cd0acbea75da..000000000000
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_fb.c
+++ /dev/null
@@ -1,388 +0,0 @@
-/*
- * Copyright © 2007 David Airlie
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice (including the next
- * paragraph) shall be included in all copies or substantial portions of the
- * Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
- * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
- * DEALINGS IN THE SOFTWARE.
- *
- * Authors:
- *     David Airlie
- */
-
-#include <linux/module.h>
-#include <linux/pm_runtime.h>
-#include <linux/slab.h>
-#include <linux/vga_switcheroo.h>
-
-#include <drm/amdgpu_drm.h>
-#include <drm/drm_crtc.h>
-#include <drm/drm_crtc_helper.h>
-#include <drm/drm_fb_helper.h>
-#include <drm/drm_fourcc.h>
-
-#include "amdgpu.h"
-#include "cikd.h"
-#include "amdgpu_gem.h"
-
-#include "amdgpu_display.h"
-
-/* object hierarchy -
-   this contains a helper + a amdgpu fb
-   the helper contains a pointer to amdgpu framebuffer baseclass.
-*/
-
-static int
-amdgpufb_open(struct fb_info *info, int user)
-{
-	struct drm_fb_helper *fb_helper = info->par;
-	int ret = pm_runtime_get_sync(fb_helper->dev->dev);
-	if (ret < 0 && ret != -EACCES) {
-		pm_runtime_mark_last_busy(fb_helper->dev->dev);
-		pm_runtime_put_autosuspend(fb_helper->dev->dev);
-		return ret;
-	}
-	return 0;
-}
-
-static int
-amdgpufb_release(struct fb_info *info, int user)
-{
-	struct drm_fb_helper *fb_helper = info->par;
-
-	pm_runtime_mark_last_busy(fb_helper->dev->dev);
-	pm_runtime_put_autosuspend(fb_helper->dev->dev);
-	return 0;
-}
-
-static const struct fb_ops amdgpufb_ops = {
-	.owner = THIS_MODULE,
-	DRM_FB_HELPER_DEFAULT_OPS,
-	.fb_open = amdgpufb_open,
-	.fb_release = amdgpufb_release,
-	.fb_fillrect = drm_fb_helper_cfb_fillrect,
-	.fb_copyarea = drm_fb_helper_cfb_copyarea,
-	.fb_imageblit = drm_fb_helper_cfb_imageblit,
-};
-
-
-int amdgpu_align_pitch(struct amdgpu_device *adev, int width, int cpp, bool tiled)
-{
-	int aligned = width;
-	int pitch_mask = 0;
-
-	switch (cpp) {
-	case 1:
-		pitch_mask = 255;
-		break;
-	case 2:
-		pitch_mask = 127;
-		break;
-	case 3:
-	case 4:
-		pitch_mask = 63;
-		break;
-	}
-
-	aligned += pitch_mask;
-	aligned &= ~pitch_mask;
-	return aligned * cpp;
-}
-
-static void amdgpufb_destroy_pinned_object(struct drm_gem_object *gobj)
-{
-	struct amdgpu_bo *abo = gem_to_amdgpu_bo(gobj);
-	int ret;
-
-	ret = amdgpu_bo_reserve(abo, true);
-	if (likely(ret == 0)) {
-		amdgpu_bo_kunmap(abo);
-		amdgpu_bo_unpin(abo);
-		amdgpu_bo_unreserve(abo);
-	}
-	drm_gem_object_put(gobj);
-}
-
-static int amdgpufb_create_pinned_object(struct amdgpu_fbdev *rfbdev,
-					 struct drm_mode_fb_cmd2 *mode_cmd,
-					 struct drm_gem_object **gobj_p)
-{
-	const struct drm_format_info *info;
-	struct amdgpu_device *adev = rfbdev->adev;
-	struct drm_gem_object *gobj = NULL;
-	struct amdgpu_bo *abo = NULL;
-	bool fb_tiled = false; /* useful for testing */
-	u32 tiling_flags = 0, domain;
-	int ret;
-	int aligned_size, size;
-	int height = mode_cmd->height;
-	u32 cpp;
-	u64 flags = AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED |
-			       AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS     |
-			       AMDGPU_GEM_CREATE_VRAM_CLEARED;
-
-	info = drm_get_format_info(adev_to_drm(adev), mode_cmd);
-	cpp = info->cpp[0];
-
-	/* need to align pitch with crtc limits */
-	mode_cmd->pitches[0] = amdgpu_align_pitch(adev, mode_cmd->width, cpp,
-						  fb_tiled);
-	domain = amdgpu_display_supported_domains(adev, flags);
-	height = ALIGN(mode_cmd->height, 8);
-	size = mode_cmd->pitches[0] * height;
-	aligned_size = ALIGN(size, PAGE_SIZE);
-	ret = amdgpu_gem_object_create(adev, aligned_size, 0, domain, flags,
-				       ttm_bo_type_device, NULL, &gobj);
-	if (ret) {
-		pr_err("failed to allocate framebuffer (%d)\n", aligned_size);
-		return -ENOMEM;
-	}
-	abo = gem_to_amdgpu_bo(gobj);
-
-	if (fb_tiled)
-		tiling_flags = AMDGPU_TILING_SET(ARRAY_MODE, GRPH_ARRAY_2D_TILED_THIN1);
-
-	ret = amdgpu_bo_reserve(abo, false);
-	if (unlikely(ret != 0))
-		goto out_unref;
-
-	if (tiling_flags) {
-		ret = amdgpu_bo_set_tiling_flags(abo,
-						 tiling_flags);
-		if (ret)
-			dev_err(adev->dev, "FB failed to set tiling flags\n");
-	}
-
-	ret = amdgpu_bo_pin(abo, domain);
-	if (ret) {
-		amdgpu_bo_unreserve(abo);
-		goto out_unref;
-	}
-
-	ret = amdgpu_ttm_alloc_gart(&abo->tbo);
-	if (ret) {
-		amdgpu_bo_unreserve(abo);
-		dev_err(adev->dev, "%p bind failed\n", abo);
-		goto out_unref;
-	}
-
-	ret = amdgpu_bo_kmap(abo, NULL);
-	amdgpu_bo_unreserve(abo);
-	if (ret) {
-		goto out_unref;
-	}
-
-	*gobj_p = gobj;
-	return 0;
-out_unref:
-	amdgpufb_destroy_pinned_object(gobj);
-	*gobj_p = NULL;
-	return ret;
-}
-
-static int amdgpufb_create(struct drm_fb_helper *helper,
-			   struct drm_fb_helper_surface_size *sizes)
-{
-	struct amdgpu_fbdev *rfbdev = (struct amdgpu_fbdev *)helper;
-	struct amdgpu_device *adev = rfbdev->adev;
-	struct fb_info *info;
-	struct drm_framebuffer *fb = NULL;
-	struct drm_mode_fb_cmd2 mode_cmd;
-	struct drm_gem_object *gobj = NULL;
-	struct amdgpu_bo *abo = NULL;
-	int ret;
-
-	memset(&mode_cmd, 0, sizeof(mode_cmd));
-	mode_cmd.width = sizes->surface_width;
-	mode_cmd.height = sizes->surface_height;
-
-	if (sizes->surface_bpp == 24)
-		sizes->surface_bpp = 32;
-
-	mode_cmd.pixel_format = drm_mode_legacy_fb_format(sizes->surface_bpp,
-							  sizes->surface_depth);
-
-	ret = amdgpufb_create_pinned_object(rfbdev, &mode_cmd, &gobj);
-	if (ret) {
-		DRM_ERROR("failed to create fbcon object %d\n", ret);
-		return ret;
-	}
-
-	abo = gem_to_amdgpu_bo(gobj);
-
-	/* okay we have an object now allocate the framebuffer */
-	info = drm_fb_helper_alloc_fbi(helper);
-	if (IS_ERR(info)) {
-		ret = PTR_ERR(info);
-		goto out;
-	}
-
-	ret = amdgpu_display_gem_fb_init(adev_to_drm(adev), &rfbdev->rfb,
-					 &mode_cmd, gobj);
-	if (ret) {
-		DRM_ERROR("failed to initialize framebuffer %d\n", ret);
-		goto out;
-	}
-
-	fb = &rfbdev->rfb.base;
-
-	/* setup helper */
-	rfbdev->helper.fb = fb;
-
-	info->fbops = &amdgpufb_ops;
-
-	info->fix.smem_start = amdgpu_gmc_vram_cpu_pa(adev, abo);
-	info->fix.smem_len = amdgpu_bo_size(abo);
-	info->screen_base = amdgpu_bo_kptr(abo);
-	info->screen_size = amdgpu_bo_size(abo);
-
-	drm_fb_helper_fill_info(info, &rfbdev->helper, sizes);
-
-	/* setup aperture base/size for vesafb takeover */
-	info->apertures->ranges[0].base = adev_to_drm(adev)->mode_config.fb_base;
-	info->apertures->ranges[0].size = adev->gmc.aper_size;
-
-	/* Use default scratch pixmap (info->pixmap.flags = FB_PIXMAP_SYSTEM) */
-
-	if (info->screen_base == NULL) {
-		ret = -ENOSPC;
-		goto out;
-	}
-
-	DRM_INFO("fb mappable at 0x%lX\n",  info->fix.smem_start);
-	DRM_INFO("vram apper at 0x%lX\n",  (unsigned long)adev->gmc.aper_base);
-	DRM_INFO("size %lu\n", (unsigned long)amdgpu_bo_size(abo));
-	DRM_INFO("fb depth is %d\n", fb->format->depth);
-	DRM_INFO("   pitch is %d\n", fb->pitches[0]);
-
-	vga_switcheroo_client_fb_set(adev->pdev, info);
-	return 0;
-
-out:
-	if (fb && ret) {
-		drm_gem_object_put(gobj);
-		drm_framebuffer_unregister_private(fb);
-		drm_framebuffer_cleanup(fb);
-		kfree(fb);
-	}
-	return ret;
-}
-
-static int amdgpu_fbdev_destroy(struct drm_device *dev, struct amdgpu_fbdev *rfbdev)
-{
-	struct amdgpu_framebuffer *rfb = &rfbdev->rfb;
-	int i;
-
-	drm_fb_helper_unregister_fbi(&rfbdev->helper);
-
-	if (rfb->base.obj[0]) {
-		for (i = 0; i < rfb->base.format->num_planes; i++)
-			drm_gem_object_put(rfb->base.obj[0]);
-		amdgpufb_destroy_pinned_object(rfb->base.obj[0]);
-		rfb->base.obj[0] = NULL;
-		drm_framebuffer_unregister_private(&rfb->base);
-		drm_framebuffer_cleanup(&rfb->base);
-	}
-	drm_fb_helper_fini(&rfbdev->helper);
-
-	return 0;
-}
-
-static const struct drm_fb_helper_funcs amdgpu_fb_helper_funcs = {
-	.fb_probe = amdgpufb_create,
-};
-
-int amdgpu_fbdev_init(struct amdgpu_device *adev)
-{
-	struct amdgpu_fbdev *rfbdev;
-	int bpp_sel = 32;
-	int ret;
-
-	/* don't init fbdev on hw without DCE */
-	if (!adev->mode_info.mode_config_initialized)
-		return 0;
-
-	/* don't init fbdev if there are no connectors */
-	if (list_empty(&adev_to_drm(adev)->mode_config.connector_list))
-		return 0;
-
-	/* select 8 bpp console on low vram cards */
-	if (adev->gmc.real_vram_size <= (32*1024*1024))
-		bpp_sel = 8;
-
-	rfbdev = kzalloc(sizeof(struct amdgpu_fbdev), GFP_KERNEL);
-	if (!rfbdev)
-		return -ENOMEM;
-
-	rfbdev->adev = adev;
-	adev->mode_info.rfbdev = rfbdev;
-
-	drm_fb_helper_prepare(adev_to_drm(adev), &rfbdev->helper,
-			      &amdgpu_fb_helper_funcs);
-
-	ret = drm_fb_helper_init(adev_to_drm(adev), &rfbdev->helper);
-	if (ret) {
-		kfree(rfbdev);
-		return ret;
-	}
-
-	/* disable all the possible outputs/crtcs before entering KMS mode */
-	if (!amdgpu_device_has_dc_support(adev) && !amdgpu_virtual_display)
-		drm_helper_disable_unused_functions(adev_to_drm(adev));
-
-	drm_fb_helper_initial_config(&rfbdev->helper, bpp_sel);
-	return 0;
-}
-
-void amdgpu_fbdev_fini(struct amdgpu_device *adev)
-{
-	if (!adev->mode_info.rfbdev)
-		return;
-
-	amdgpu_fbdev_destroy(adev_to_drm(adev), adev->mode_info.rfbdev);
-	kfree(adev->mode_info.rfbdev);
-	adev->mode_info.rfbdev = NULL;
-}
-
-void amdgpu_fbdev_set_suspend(struct amdgpu_device *adev, int state)
-{
-	if (adev->mode_info.rfbdev)
-		drm_fb_helper_set_suspend_unlocked(&adev->mode_info.rfbdev->helper,
-						   state);
-}
-
-int amdgpu_fbdev_total_size(struct amdgpu_device *adev)
-{
-	struct amdgpu_bo *robj;
-	int size = 0;
-
-	if (!adev->mode_info.rfbdev)
-		return 0;
-
-	robj = gem_to_amdgpu_bo(adev->mode_info.rfbdev->rfb.base.obj[0]);
-	size += amdgpu_bo_size(robj);
-	return size;
-}
-
-bool amdgpu_fbdev_robj_is_fb(struct amdgpu_device *adev, struct amdgpu_bo *robj)
-{
-	if (!adev->mode_info.rfbdev)
-		return false;
-	if (robj == gem_to_amdgpu_bo(adev->mode_info.rfbdev->rfb.base.obj[0]))
-		return true;
-	return false;
-}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_fdinfo.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_fdinfo.c
index 5a6857c44bb6..b349bb3676d5 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_fdinfo.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_fdinfo.c
@@ -32,6 +32,8 @@
 
 #include <drm/amdgpu_drm.h>
 #include <drm/drm_debugfs.h>
+#include <drm/drm_drv.h>
+#include <drm/drm_file.h>
 
 #include "amdgpu.h"
 #include "amdgpu_vm.h"
@@ -50,62 +52,73 @@ static const char *amdgpu_ip_name[AMDGPU_HW_IP_NUM] = {
 	[AMDGPU_HW_IP_VCN_DEC]	=	"dec",
 	[AMDGPU_HW_IP_VCN_ENC]	=	"enc",
 	[AMDGPU_HW_IP_VCN_JPEG]	=	"jpeg",
+	[AMDGPU_HW_IP_VPE]	=	"vpe",
 };
 
-void amdgpu_show_fdinfo(struct seq_file *m, struct file *f)
+void amdgpu_show_fdinfo(struct drm_printer *p, struct drm_file *file)
 {
-	struct amdgpu_fpriv *fpriv;
-	uint32_t bus, dev, fn, i, domain;
-	uint64_t vram_mem = 0, gtt_mem = 0, cpu_mem = 0;
-	struct drm_file *file = f->private_data;
-	struct amdgpu_device *adev = drm_to_adev(file->minor->dev);
-	struct amdgpu_bo *root;
-	int ret;
-
-	ret = amdgpu_file_to_fpriv(f, &fpriv);
-	if (ret)
-		return;
-	bus = adev->pdev->bus->number;
-	domain = pci_domain_nr(adev->pdev->bus);
-	dev = PCI_SLOT(adev->pdev->devfn);
-	fn = PCI_FUNC(adev->pdev->devfn);
-
-	root = amdgpu_bo_ref(fpriv->vm.root.bo);
-	if (!root)
-		return;
-
-	ret = amdgpu_bo_reserve(root, false);
-	if (ret) {
-		DRM_ERROR("Fail to reserve bo\n");
-		return;
+	struct amdgpu_fpriv *fpriv = file->driver_priv;
+	struct amdgpu_vm *vm = &fpriv->vm;
+
+	struct amdgpu_mem_stats stats[__AMDGPU_PL_NUM];
+	ktime_t usage[AMDGPU_HW_IP_NUM];
+	const char *pl_name[] = {
+		[TTM_PL_VRAM] = "vram",
+		[TTM_PL_TT] = "gtt",
+		[TTM_PL_SYSTEM] = "cpu",
+		[AMDGPU_PL_GDS] = "gds",
+		[AMDGPU_PL_GWS] = "gws",
+		[AMDGPU_PL_OA] = "oa",
+		[AMDGPU_PL_DOORBELL] = "doorbell",
+		[AMDGPU_PL_MMIO_REMAP] = "mmioremap",
+	};
+	unsigned int hw_ip, i;
+
+	amdgpu_vm_get_memory(vm, stats);
+	amdgpu_ctx_mgr_usage(&fpriv->ctx_mgr, usage);
+
+	/*
+	 * ******************************************************************
+	 * For text output format description please see drm-usage-stats.rst!
+	 * ******************************************************************
+	 */
+
+	drm_printf(p, "pasid:\t%u\n", fpriv->vm.pasid);
+
+	for (i = 0; i < ARRAY_SIZE(pl_name); i++) {
+		if (!pl_name[i])
+			continue;
+
+		drm_print_memory_stats(p,
+				       &stats[i].drm,
+				       DRM_GEM_OBJECT_RESIDENT |
+				       DRM_GEM_OBJECT_PURGEABLE,
+				       pl_name[i]);
 	}
-	amdgpu_vm_get_memory(&fpriv->vm, &vram_mem, &gtt_mem, &cpu_mem);
-	amdgpu_bo_unreserve(root);
-	amdgpu_bo_unref(&root);
-
-	seq_printf(m, "pdev:\t%04x:%02x:%02x.%d\npasid:\t%u\n", domain, bus,
-			dev, fn, fpriv->vm.pasid);
-	seq_printf(m, "vram mem:\t%llu kB\n", vram_mem/1024UL);
-	seq_printf(m, "gtt mem:\t%llu kB\n", gtt_mem/1024UL);
-	seq_printf(m, "cpu mem:\t%llu kB\n", cpu_mem/1024UL);
-	for (i = 0; i < AMDGPU_HW_IP_NUM; i++) {
-		uint32_t count = amdgpu_ctx_num_entities[i];
-		int idx = 0;
-		uint64_t total = 0, min = 0;
-		uint32_t perc, frac;
-
-		for (idx = 0; idx < count; idx++) {
-			total = amdgpu_ctx_mgr_fence_usage(&fpriv->ctx_mgr,
-				i, idx, &min);
-			if ((total == 0) || (min == 0))
-				continue;
-
-			perc = div64_u64(10000 * total, min);
-			frac = perc % 100;
-
-			seq_printf(m, "%s%d:\t%d.%d%%\n",
-					amdgpu_ip_name[i],
-					idx, perc/100, frac);
-		}
+
+	/* Legacy amdgpu keys, alias to drm-resident-memory-: */
+	drm_printf(p, "drm-memory-vram:\t%llu KiB\n",
+		   stats[TTM_PL_VRAM].drm.resident/1024UL);
+	drm_printf(p, "drm-memory-gtt: \t%llu KiB\n",
+		   stats[TTM_PL_TT].drm.resident/1024UL);
+	drm_printf(p, "drm-memory-cpu: \t%llu KiB\n",
+		   stats[TTM_PL_SYSTEM].drm.resident/1024UL);
+
+	/* Amdgpu specific memory accounting keys: */
+	drm_printf(p, "amd-evicted-vram:\t%llu KiB\n",
+		   stats[TTM_PL_VRAM].evicted/1024UL);
+	drm_printf(p, "amd-requested-vram:\t%llu KiB\n",
+		   (stats[TTM_PL_VRAM].drm.shared +
+		    stats[TTM_PL_VRAM].drm.private) / 1024UL);
+	drm_printf(p, "amd-requested-gtt:\t%llu KiB\n",
+		   (stats[TTM_PL_TT].drm.shared +
+		    stats[TTM_PL_TT].drm.private) / 1024UL);
+
+	for (hw_ip = 0; hw_ip < AMDGPU_HW_IP_NUM; ++hw_ip) {
+		if (!usage[hw_ip])
+			continue;
+
+		drm_printf(p, "drm-engine-%s:\t%lld ns\n", amdgpu_ip_name[hw_ip],
+			   ktime_to_ns(usage[hw_ip]));
 	}
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_fdinfo.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_fdinfo.h
index 41a4c7056729..0398f5a159ef 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_fdinfo.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_fdinfo.h
@@ -30,7 +30,6 @@
 #include <linux/rbtree.h>
 #include <drm/gpu_scheduler.h>
 #include <drm/drm_file.h>
-#include <drm/ttm/ttm_bo_driver.h>
 #include <linux/sched/mm.h>
 
 #include "amdgpu_sync.h"
@@ -38,6 +37,6 @@
 #include "amdgpu_ids.h"
 
 uint32_t amdgpu_get_ip_count(struct amdgpu_device *adev, int id);
-void amdgpu_show_fdinfo(struct seq_file *m, struct file *f);
+void amdgpu_show_fdinfo(struct drm_printer *p, struct drm_file *file);
 
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c
index 3b7e86ea7167..c7843e336310 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c
@@ -39,41 +39,9 @@
 #include <drm/drm_drv.h>
 #include "amdgpu.h"
 #include "amdgpu_trace.h"
+#include "amdgpu_reset.h"
 
 /*
- * Fences
- * Fences mark an event in the GPUs pipeline and are used
- * for GPU/CPU synchronization.  When the fence is written,
- * it is expected that all buffers associated with that fence
- * are no longer in use by the associated ring on the GPU and
- * that the the relevant GPU caches have been flushed.
- */
-
-struct amdgpu_fence {
-	struct dma_fence base;
-
-	/* RB, DMA, etc. */
-	struct amdgpu_ring		*ring;
-};
-
-static struct kmem_cache *amdgpu_fence_slab;
-
-int amdgpu_fence_slab_init(void)
-{
-	amdgpu_fence_slab = kmem_cache_create(
-		"amdgpu_fence", sizeof(struct amdgpu_fence), 0,
-		SLAB_HWCACHE_ALIGN, NULL);
-	if (!amdgpu_fence_slab)
-		return -ENOMEM;
-	return 0;
-}
-
-void amdgpu_fence_slab_fini(void)
-{
-	rcu_barrier();
-	kmem_cache_destroy(amdgpu_fence_slab);
-}
-/*
  * Cast helper
  */
 static const struct dma_fence_ops amdgpu_fence_ops;
@@ -81,10 +49,7 @@ static inline struct amdgpu_fence *to_amdgpu_fence(struct dma_fence *f)
 {
 	struct amdgpu_fence *__f = container_of(f, struct amdgpu_fence, base);
 
-	if (__f->base.ops == &amdgpu_fence_ops)
-		return __f;
-
-	return NULL;
+	return __f;
 }
 
 /**
@@ -128,53 +93,32 @@ static u32 amdgpu_fence_read(struct amdgpu_ring *ring)
  * amdgpu_fence_emit - emit a fence on the requested ring
  *
  * @ring: ring the fence is associated with
- * @f: resulting fence object
- * @job: job the fence is embedded in
+ * @af: amdgpu fence input
  * @flags: flags to pass into the subordinate .emit_fence() call
  *
  * Emits a fence command on the requested ring (all asics).
  * Returns 0 on success, -ENOMEM on failure.
  */
-int amdgpu_fence_emit(struct amdgpu_ring *ring, struct dma_fence **f, struct amdgpu_job *job,
-		      unsigned flags)
+int amdgpu_fence_emit(struct amdgpu_ring *ring, struct amdgpu_fence *af,
+		      unsigned int flags)
 {
 	struct amdgpu_device *adev = ring->adev;
 	struct dma_fence *fence;
-	struct amdgpu_fence *am_fence;
 	struct dma_fence __rcu **ptr;
 	uint32_t seq;
 	int r;
 
-	if (job == NULL) {
-		/* create a sperate hw fence */
-		am_fence = kmem_cache_alloc(amdgpu_fence_slab, GFP_ATOMIC);
-		if (am_fence == NULL)
-			return -ENOMEM;
-		fence = &am_fence->base;
-		am_fence->ring = ring;
-	} else {
-		/* take use of job-embedded fence */
-		fence = &job->hw_fence;
-	}
+	fence = &af->base;
+	af->ring = ring;
 
 	seq = ++ring->fence_drv.sync_seq;
-	if (job != NULL && job->job_run_counter) {
-		/* reinit seq for resubmitted jobs */
-		fence->seqno = seq;
-	} else {
-		dma_fence_init(fence, &amdgpu_fence_ops,
-				&ring->fence_drv.lock,
-				adev->fence_context + ring->idx,
-				seq);
-	}
-
-	if (job != NULL) {
-		/* mark this fence has a parent job */
-		set_bit(AMDGPU_FENCE_FLAG_EMBED_IN_JOB_BIT, &fence->flags);
-	}
+	dma_fence_init(fence, &amdgpu_fence_ops,
+		       &ring->fence_drv.lock,
+		       adev->fence_context + ring->idx, seq);
 
 	amdgpu_ring_emit_fence(ring, ring->fence_drv.gpu_addr,
 			       seq, flags | AMDGPU_FENCE_FLAG_INT);
+	amdgpu_fence_save_wptr(af);
 	pm_runtime_get_noresume(adev_to_drm(adev)->dev);
 	ptr = &ring->fence_drv.fences[seq & ring->fence_drv.num_fences_mask];
 	if (unlikely(rcu_dereference_protected(*ptr, 1))) {
@@ -192,13 +136,13 @@ int amdgpu_fence_emit(struct amdgpu_ring *ring, struct dma_fence **f, struct amd
 		}
 	}
 
+	to_amdgpu_fence(fence)->start_timestamp = ktime_get();
+
 	/* This function can't be called concurrently anyway, otherwise
 	 * emitting the fence would mess up the hardware ring buffer.
 	 */
 	rcu_assign_pointer(*ptr, dma_fence_get(fence));
 
-	*f = fence;
-
 	return 0;
 }
 
@@ -273,7 +217,7 @@ bool amdgpu_fence_process(struct amdgpu_ring *ring)
 
 	} while (atomic_cmpxchg(&drv->last_seq, last_seq, seq) != last_seq);
 
-	if (del_timer(&ring->fence_drv.fallback_timer) &&
+	if (timer_delete(&ring->fence_drv.fallback_timer) &&
 	    seq != ring->fence_drv.sync_seq)
 		amdgpu_fence_schedule_fallback(ring);
 
@@ -285,6 +229,7 @@ bool amdgpu_fence_process(struct amdgpu_ring *ring)
 
 	do {
 		struct dma_fence *fence, **ptr;
+		struct amdgpu_fence *am_fence;
 
 		++last_seq;
 		last_seq &= drv->num_fences_mask;
@@ -297,9 +242,14 @@ bool amdgpu_fence_process(struct amdgpu_ring *ring)
 		if (!fence)
 			continue;
 
+		/* Save the wptr in the fence driver so we know what the last processed
+		 * wptr was.  This is required for re-emitting the ring state for
+		 * queues that are reset but are not guilty and thus have no guilty fence.
+		 */
+		am_fence = container_of(fence, struct amdgpu_fence, base);
+		drv->signalled_wptr = am_fence->wptr;
 		dma_fence_signal(fence);
 		dma_fence_put(fence);
-		pm_runtime_mark_last_busy(adev_to_drm(adev)->dev);
 		pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
 	} while (last_seq != seq);
 
@@ -315,11 +265,13 @@ bool amdgpu_fence_process(struct amdgpu_ring *ring)
  */
 static void amdgpu_fence_fallback(struct timer_list *t)
 {
-	struct amdgpu_ring *ring = from_timer(ring, t,
-					      fence_drv.fallback_timer);
+	struct amdgpu_ring *ring = timer_container_of(ring, t,
+						      fence_drv.fallback_timer);
 
 	if (amdgpu_fence_process(ring))
-		DRM_WARN("Fence fallback timer expired on ring %s\n", ring->name);
+		dev_warn(ring->adev->dev,
+			 "Fence fallback timer expired on ring %s\n",
+			 ring->name);
 }
 
 /**
@@ -367,14 +319,11 @@ signed long amdgpu_fence_wait_polling(struct amdgpu_ring *ring,
 				      uint32_t wait_seq,
 				      signed long timeout)
 {
-	uint32_t seq;
-
-	do {
-		seq = amdgpu_fence_read(ring);
-		udelay(5);
-		timeout -= 5;
-	} while ((int32_t)(wait_seq - seq) > 0 && timeout > 0);
 
+	while ((int32_t)(wait_seq - amdgpu_fence_read(ring)) > 0 && timeout > 0) {
+		udelay(2);
+		timeout -= 2;
+	}
 	return timeout > 0 ? timeout : 0;
 }
 /**
@@ -386,14 +335,13 @@ signed long amdgpu_fence_wait_polling(struct amdgpu_ring *ring,
  * Returns the number of emitted fences on the ring.  Used by the
  * dynpm code to ring track activity.
  */
-unsigned amdgpu_fence_count_emitted(struct amdgpu_ring *ring)
+unsigned int amdgpu_fence_count_emitted(struct amdgpu_ring *ring)
 {
 	uint64_t emitted;
 
 	/* We are not protected by ring lock when reading the last sequence
 	 * but it's ok to report slightly wrong fence count here.
 	 */
-	amdgpu_fence_process(ring);
 	emitted = 0x100000000ull;
 	emitted -= atomic_read(&ring->fence_drv.last_seq);
 	emitted += READ_ONCE(ring->fence_drv.sync_seq);
@@ -401,6 +349,57 @@ unsigned amdgpu_fence_count_emitted(struct amdgpu_ring *ring)
 }
 
 /**
+ * amdgpu_fence_last_unsignaled_time_us - the time fence emitted until now
+ * @ring: ring the fence is associated with
+ *
+ * Find the earliest fence unsignaled until now, calculate the time delta
+ * between the time fence emitted and now.
+ */
+u64 amdgpu_fence_last_unsignaled_time_us(struct amdgpu_ring *ring)
+{
+	struct amdgpu_fence_driver *drv = &ring->fence_drv;
+	struct dma_fence *fence;
+	uint32_t last_seq, sync_seq;
+
+	last_seq = atomic_read(&ring->fence_drv.last_seq);
+	sync_seq = READ_ONCE(ring->fence_drv.sync_seq);
+	if (last_seq == sync_seq)
+		return 0;
+
+	++last_seq;
+	last_seq &= drv->num_fences_mask;
+	fence = drv->fences[last_seq];
+	if (!fence)
+		return 0;
+
+	return ktime_us_delta(ktime_get(),
+		to_amdgpu_fence(fence)->start_timestamp);
+}
+
+/**
+ * amdgpu_fence_update_start_timestamp - update the timestamp of the fence
+ * @ring: ring the fence is associated with
+ * @seq: the fence seq number to update.
+ * @timestamp: the start timestamp to update.
+ *
+ * The function called at the time the fence and related ib is about to
+ * resubmit to gpu in MCBP scenario. Thus we do not consider race condition
+ * with amdgpu_fence_process to modify the same fence.
+ */
+void amdgpu_fence_update_start_timestamp(struct amdgpu_ring *ring, uint32_t seq, ktime_t timestamp)
+{
+	struct amdgpu_fence_driver *drv = &ring->fence_drv;
+	struct dma_fence *fence;
+
+	seq &= drv->num_fences_mask;
+	fence = drv->fences[seq];
+	if (!fence)
+		return;
+
+	to_amdgpu_fence(fence)->start_timestamp = timestamp;
+}
+
+/**
  * amdgpu_fence_driver_start_ring - make the fence driver
  * ready for use on the requested ring.
  *
@@ -415,14 +414,14 @@ unsigned amdgpu_fence_count_emitted(struct amdgpu_ring *ring)
  */
 int amdgpu_fence_driver_start_ring(struct amdgpu_ring *ring,
 				   struct amdgpu_irq_src *irq_src,
-				   unsigned irq_type)
+				   unsigned int irq_type)
 {
 	struct amdgpu_device *adev = ring->adev;
 	uint64_t index;
 
 	if (ring->funcs->type != AMDGPU_RING_TYPE_UVD) {
-		ring->fence_drv.cpu_addr = &adev->wb.wb[ring->fence_offs];
-		ring->fence_drv.gpu_addr = adev->wb.gpu_addr + (ring->fence_offs * 4);
+		ring->fence_drv.cpu_addr = ring->fence_cpu_addr;
+		ring->fence_drv.gpu_addr = ring->fence_gpu_addr;
 	} else {
 		/* put fence directly behind firmware */
 		index = ALIGN(adev->uvd.fw->size, 8);
@@ -445,24 +444,18 @@ int amdgpu_fence_driver_start_ring(struct amdgpu_ring *ring,
  * for the requested ring.
  *
  * @ring: ring to init the fence driver on
- * @num_hw_submission: number of entries on the hardware queue
- * @sched_score: optional score atomic shared with other schedulers
  *
  * Init the fence driver for the requested ring (all asics).
  * Helper function for amdgpu_fence_driver_init().
  */
-int amdgpu_fence_driver_init_ring(struct amdgpu_ring *ring,
-				  unsigned num_hw_submission,
-				  atomic_t *sched_score)
+int amdgpu_fence_driver_init_ring(struct amdgpu_ring *ring)
 {
 	struct amdgpu_device *adev = ring->adev;
-	long timeout;
-	int r;
 
 	if (!adev)
 		return -EINVAL;
 
-	if (!is_power_of_2(num_hw_submission))
+	if (!is_power_of_2(ring->num_hw_submission))
 		return -EINVAL;
 
 	ring->fence_drv.cpu_addr = NULL;
@@ -473,41 +466,14 @@ int amdgpu_fence_driver_init_ring(struct amdgpu_ring *ring,
 
 	timer_setup(&ring->fence_drv.fallback_timer, amdgpu_fence_fallback, 0);
 
-	ring->fence_drv.num_fences_mask = num_hw_submission * 2 - 1;
+	ring->fence_drv.num_fences_mask = ring->num_hw_submission * 2 - 1;
 	spin_lock_init(&ring->fence_drv.lock);
-	ring->fence_drv.fences = kcalloc(num_hw_submission * 2, sizeof(void *),
+	ring->fence_drv.fences = kcalloc(ring->num_hw_submission * 2, sizeof(void *),
 					 GFP_KERNEL);
+
 	if (!ring->fence_drv.fences)
 		return -ENOMEM;
 
-	/* No need to setup the GPU scheduler for rings that don't need it */
-	if (ring->no_scheduler)
-		return 0;
-
-	switch (ring->funcs->type) {
-	case AMDGPU_RING_TYPE_GFX:
-		timeout = adev->gfx_timeout;
-		break;
-	case AMDGPU_RING_TYPE_COMPUTE:
-		timeout = adev->compute_timeout;
-		break;
-	case AMDGPU_RING_TYPE_SDMA:
-		timeout = adev->sdma_timeout;
-		break;
-	default:
-		timeout = adev->video_timeout;
-		break;
-	}
-
-	r = drm_sched_init(&ring->sched, &amdgpu_sched_ops,
-			   num_hw_submission, amdgpu_job_hang_limit,
-			   timeout, NULL, sched_score, ring->name);
-	if (r) {
-		DRM_ERROR("Failed to create scheduler on ring %s.\n",
-			  ring->name);
-		return r;
-	}
-
 	return 0;
 }
 
@@ -529,6 +495,42 @@ int amdgpu_fence_driver_sw_init(struct amdgpu_device *adev)
 }
 
 /**
+ * amdgpu_fence_need_ring_interrupt_restore - helper function to check whether
+ * fence driver interrupts need to be restored.
+ *
+ * @ring: ring that to be checked
+ *
+ * Interrupts for rings that belong to GFX IP don't need to be restored
+ * when the target power state is s0ix.
+ *
+ * Return true if need to restore interrupts, false otherwise.
+ */
+static bool amdgpu_fence_need_ring_interrupt_restore(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	bool is_gfx_power_domain = false;
+
+	switch (ring->funcs->type) {
+	case AMDGPU_RING_TYPE_SDMA:
+	/* SDMA 5.x+ is part of GFX power domain so it's covered by GFXOFF */
+		if (amdgpu_ip_version(adev, SDMA0_HWIP, 0) >=
+		    IP_VERSION(5, 0, 0))
+			is_gfx_power_domain = true;
+		break;
+	case AMDGPU_RING_TYPE_GFX:
+	case AMDGPU_RING_TYPE_COMPUTE:
+	case AMDGPU_RING_TYPE_KIQ:
+	case AMDGPU_RING_TYPE_MES:
+		is_gfx_power_domain = true;
+		break;
+	default:
+		break;
+	}
+
+	return !(adev->in_s0ix && is_gfx_power_domain);
+}
+
+/**
  * amdgpu_fence_driver_hw_fini - tear down the fence driver
  * for all possible rings.
  *
@@ -546,9 +548,6 @@ void amdgpu_fence_driver_hw_fini(struct amdgpu_device *adev)
 		if (!ring || !ring->fence_drv.initialized)
 			continue;
 
-		if (!ring->no_scheduler)
-			drm_sched_stop(&ring->sched, NULL);
-
 		/* You can't wait for HW to signal if it's gone */
 		if (!drm_dev_is_unplugged(adev_to_drm(adev)))
 			r = amdgpu_fence_wait_empty(ring);
@@ -558,11 +557,31 @@ void amdgpu_fence_driver_hw_fini(struct amdgpu_device *adev)
 		if (r)
 			amdgpu_fence_driver_force_completion(ring);
 
-		if (ring->fence_drv.irq_src)
+		if (!drm_dev_is_unplugged(adev_to_drm(adev)) &&
+		    ring->fence_drv.irq_src &&
+		    amdgpu_fence_need_ring_interrupt_restore(ring))
 			amdgpu_irq_put(adev, ring->fence_drv.irq_src,
 				       ring->fence_drv.irq_type);
 
-		del_timer_sync(&ring->fence_drv.fallback_timer);
+		timer_delete_sync(&ring->fence_drv.fallback_timer);
+	}
+}
+
+/* Will either stop and flush handlers for amdgpu interrupt or reanble it */
+void amdgpu_fence_driver_isr_toggle(struct amdgpu_device *adev, bool stop)
+{
+	int i;
+
+	for (i = 0; i < AMDGPU_MAX_RINGS; i++) {
+		struct amdgpu_ring *ring = adev->rings[i];
+
+		if (!ring || !ring->fence_drv.initialized || !ring->fence_drv.irq_src)
+			continue;
+
+		if (stop)
+			disable_irq(adev->irq.irq);
+		else
+			enable_irq(adev->irq.irq);
 	}
 }
 
@@ -576,7 +595,13 @@ void amdgpu_fence_driver_sw_fini(struct amdgpu_device *adev)
 		if (!ring || !ring->fence_drv.initialized)
 			continue;
 
-		if (!ring->no_scheduler)
+		/*
+		 * Notice we check for sched.ops since there's some
+		 * override on the meaning of sched.ready by amdgpu.
+		 * The natural check would be sched.ready, which is
+		 * set as drm_sched_init() finishes...
+		 */
+		if (ring->sched.ops)
 			drm_sched_fini(&ring->sched);
 
 		for (j = 0; j <= ring->fence_drv.num_fences_mask; ++j)
@@ -605,22 +630,43 @@ void amdgpu_fence_driver_hw_init(struct amdgpu_device *adev)
 
 	for (i = 0; i < AMDGPU_MAX_RINGS; i++) {
 		struct amdgpu_ring *ring = adev->rings[i];
+
 		if (!ring || !ring->fence_drv.initialized)
 			continue;
 
-		if (!ring->no_scheduler) {
-			drm_sched_resubmit_jobs(&ring->sched);
-			drm_sched_start(&ring->sched, true);
-		}
-
 		/* enable the interrupt */
-		if (ring->fence_drv.irq_src)
+		if (ring->fence_drv.irq_src &&
+		    amdgpu_fence_need_ring_interrupt_restore(ring))
 			amdgpu_irq_get(adev, ring->fence_drv.irq_src,
 				       ring->fence_drv.irq_type);
 	}
 }
 
 /**
+ * amdgpu_fence_driver_set_error - set error code on fences
+ * @ring: the ring which contains the fences
+ * @error: the error code to set
+ *
+ * Set an error code to all the fences pending on the ring.
+ */
+void amdgpu_fence_driver_set_error(struct amdgpu_ring *ring, int error)
+{
+	struct amdgpu_fence_driver *drv = &ring->fence_drv;
+	unsigned long flags;
+
+	spin_lock_irqsave(&drv->lock, flags);
+	for (unsigned int i = 0; i <= drv->num_fences_mask; ++i) {
+		struct dma_fence *fence;
+
+		fence = rcu_dereference_protected(drv->fences[i],
+						  lockdep_is_held(&drv->lock));
+		if (fence && !dma_fence_is_signaled_locked(fence))
+			dma_fence_set_error(fence, error);
+	}
+	spin_unlock_irqrestore(&drv->lock, flags);
+}
+
+/**
  * amdgpu_fence_driver_force_completion - force signal latest fence of ring
  *
  * @ring: fence of the ring to signal
@@ -628,10 +674,125 @@ void amdgpu_fence_driver_hw_init(struct amdgpu_device *adev)
  */
 void amdgpu_fence_driver_force_completion(struct amdgpu_ring *ring)
 {
+	amdgpu_fence_driver_set_error(ring, -ECANCELED);
 	amdgpu_fence_write(ring, ring->fence_drv.sync_seq);
 	amdgpu_fence_process(ring);
 }
 
+
+/*
+ * Kernel queue reset handling
+ *
+ * The driver can reset individual queues for most engines, but those queues
+ * may contain work from multiple contexts.  Resetting the queue will reset
+ * lose all of that state.  In order to minimize the collateral damage, the
+ * driver will save the ring contents which are not associated with the guilty
+ * context prior to resetting the queue.  After resetting the queue the queue
+ * contents from the other contexts is re-emitted to the rings so that it can
+ * be processed by the engine.  To handle this, we save the queue's write
+ * pointer (wptr) in the fences associated with each context.  If we get a
+ * queue timeout, we can then use the wptrs from the fences to determine
+ * which data needs to be saved out of the queue's ring buffer.
+ */
+
+/**
+ * amdgpu_fence_driver_guilty_force_completion - force signal of specified sequence
+ *
+ * @af: fence of the ring to signal
+ *
+ */
+void amdgpu_fence_driver_guilty_force_completion(struct amdgpu_fence *af)
+{
+	struct dma_fence *unprocessed;
+	struct dma_fence __rcu **ptr;
+	struct amdgpu_fence *fence;
+	struct amdgpu_ring *ring = af->ring;
+	unsigned long flags;
+	u32 seq, last_seq;
+
+	last_seq = amdgpu_fence_read(ring) & ring->fence_drv.num_fences_mask;
+	seq = ring->fence_drv.sync_seq & ring->fence_drv.num_fences_mask;
+
+	/* mark all fences from the guilty context with an error */
+	spin_lock_irqsave(&ring->fence_drv.lock, flags);
+	do {
+		last_seq++;
+		last_seq &= ring->fence_drv.num_fences_mask;
+
+		ptr = &ring->fence_drv.fences[last_seq];
+		rcu_read_lock();
+		unprocessed = rcu_dereference(*ptr);
+
+		if (unprocessed && !dma_fence_is_signaled_locked(unprocessed)) {
+			fence = container_of(unprocessed, struct amdgpu_fence, base);
+
+			if (fence == af)
+				dma_fence_set_error(&fence->base, -ETIME);
+			else if (fence->context == af->context)
+				dma_fence_set_error(&fence->base, -ECANCELED);
+		}
+		rcu_read_unlock();
+	} while (last_seq != seq);
+	spin_unlock_irqrestore(&ring->fence_drv.lock, flags);
+	/* signal the guilty fence */
+	amdgpu_fence_write(ring, (u32)af->base.seqno);
+	amdgpu_fence_process(ring);
+}
+
+void amdgpu_fence_save_wptr(struct amdgpu_fence *af)
+{
+	af->wptr = af->ring->wptr;
+}
+
+static void amdgpu_ring_backup_unprocessed_command(struct amdgpu_ring *ring,
+						   u64 start_wptr, u32 end_wptr)
+{
+	unsigned int first_idx = start_wptr & ring->buf_mask;
+	unsigned int last_idx = end_wptr & ring->buf_mask;
+	unsigned int i;
+
+	/* Backup the contents of the ring buffer. */
+	for (i = first_idx; i != last_idx; ++i, i &= ring->buf_mask)
+		ring->ring_backup[ring->ring_backup_entries_to_copy++] = ring->ring[i];
+}
+
+void amdgpu_ring_backup_unprocessed_commands(struct amdgpu_ring *ring,
+					     struct amdgpu_fence *guilty_fence)
+{
+	struct dma_fence *unprocessed;
+	struct dma_fence __rcu **ptr;
+	struct amdgpu_fence *fence;
+	u64 wptr;
+	u32 seq, last_seq;
+
+	last_seq = amdgpu_fence_read(ring) & ring->fence_drv.num_fences_mask;
+	seq = ring->fence_drv.sync_seq & ring->fence_drv.num_fences_mask;
+	wptr = ring->fence_drv.signalled_wptr;
+	ring->ring_backup_entries_to_copy = 0;
+
+	do {
+		last_seq++;
+		last_seq &= ring->fence_drv.num_fences_mask;
+
+		ptr = &ring->fence_drv.fences[last_seq];
+		rcu_read_lock();
+		unprocessed = rcu_dereference(*ptr);
+
+		if (unprocessed && !dma_fence_is_signaled(unprocessed)) {
+			fence = container_of(unprocessed, struct amdgpu_fence, base);
+
+			/* save everything if the ring is not guilty, otherwise
+			 * just save the content from other contexts.
+			 */
+			if (!guilty_fence || (fence->context != guilty_fence->context))
+				amdgpu_ring_backup_unprocessed_command(ring, wptr,
+								       fence->wptr);
+			wptr = fence->wptr;
+		}
+		rcu_read_unlock();
+	} while (last_seq != seq);
+}
+
 /*
  * Common fence implementation
  */
@@ -643,16 +804,7 @@ static const char *amdgpu_fence_get_driver_name(struct dma_fence *fence)
 
 static const char *amdgpu_fence_get_timeline_name(struct dma_fence *f)
 {
-	struct amdgpu_ring *ring;
-
-	if (test_bit(AMDGPU_FENCE_FLAG_EMBED_IN_JOB_BIT, &f->flags)) {
-		struct amdgpu_job *job = container_of(f, struct amdgpu_job, hw_fence);
-
-		ring = to_amdgpu_ring(job->base.sched);
-	} else {
-		ring = to_amdgpu_fence(f)->ring;
-	}
-	return (const char *)ring->name;
+	return (const char *)to_amdgpu_fence(f)->ring->name;
 }
 
 /**
@@ -665,18 +817,8 @@ static const char *amdgpu_fence_get_timeline_name(struct dma_fence *f)
  */
 static bool amdgpu_fence_enable_signaling(struct dma_fence *f)
 {
-	struct amdgpu_ring *ring;
-
-	if (test_bit(AMDGPU_FENCE_FLAG_EMBED_IN_JOB_BIT, &f->flags)) {
-		struct amdgpu_job *job = container_of(f, struct amdgpu_job, hw_fence);
-
-		ring = to_amdgpu_ring(job->base.sched);
-	} else {
-		ring = to_amdgpu_fence(f)->ring;
-	}
-
-	if (!timer_pending(&ring->fence_drv.fallback_timer))
-		amdgpu_fence_schedule_fallback(ring);
+	if (!timer_pending(&to_amdgpu_fence(f)->ring->fence_drv.fallback_timer))
+		amdgpu_fence_schedule_fallback(to_amdgpu_fence(f)->ring);
 
 	return true;
 }
@@ -692,19 +834,8 @@ static void amdgpu_fence_free(struct rcu_head *rcu)
 {
 	struct dma_fence *f = container_of(rcu, struct dma_fence, rcu);
 
-	if (test_bit(AMDGPU_FENCE_FLAG_EMBED_IN_JOB_BIT, &f->flags)) {
-	/* free job if fence has a parent job */
-		struct amdgpu_job *job;
-
-		job = container_of(f, struct amdgpu_job, hw_fence);
-		kfree(job);
-	} else {
 	/* free fence_slab if it's separated fence*/
-		struct amdgpu_fence *fence;
-
-		fence = to_amdgpu_fence(f);
-		kmem_cache_free(amdgpu_fence_slab, fence);
-	}
+	kfree(to_amdgpu_fence(f));
 }
 
 /**
@@ -727,18 +858,18 @@ static const struct dma_fence_ops amdgpu_fence_ops = {
 	.release = amdgpu_fence_release,
 };
 
-
 /*
  * Fence debugfs
  */
 #if defined(CONFIG_DEBUG_FS)
 static int amdgpu_debugfs_fence_info_show(struct seq_file *m, void *unused)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)m->private;
+	struct amdgpu_device *adev = m->private;
 	int i;
 
 	for (i = 0; i < AMDGPU_MAX_RINGS; ++i) {
 		struct amdgpu_ring *ring = adev->rings[i];
+
 		if (!ring || !ring->fence_drv.initialized)
 			continue;
 
@@ -791,9 +922,11 @@ static int gpu_recover_get(void *data, u64 *val)
 		return 0;
 	}
 
-	*val = amdgpu_device_gpu_recover(adev, NULL);
+	if (amdgpu_reset_domain_schedule(adev->reset_domain, &adev->reset_work))
+		flush_work(&adev->reset_work);
+
+	*val = atomic_read(&adev->reset_domain->reset_res);
 
-	pm_runtime_mark_last_busy(dev->dev);
 	pm_runtime_put_autosuspend(dev->dev);
 
 	return 0;
@@ -803,6 +936,24 @@ DEFINE_SHOW_ATTRIBUTE(amdgpu_debugfs_fence_info);
 DEFINE_DEBUGFS_ATTRIBUTE(amdgpu_debugfs_gpu_recover_fops, gpu_recover_get, NULL,
 			 "%lld\n");
 
+static void amdgpu_debugfs_reset_work(struct work_struct *work)
+{
+	struct amdgpu_device *adev = container_of(work, struct amdgpu_device,
+						  reset_work);
+
+	struct amdgpu_reset_context reset_context;
+
+	memset(&reset_context, 0, sizeof(reset_context));
+
+	reset_context.method = AMD_RESET_METHOD_NONE;
+	reset_context.reset_req_dev = adev;
+	reset_context.src = AMDGPU_RESET_SRC_USER;
+	set_bit(AMDGPU_NEED_FULL_RESET, &reset_context.flags);
+	set_bit(AMDGPU_SKIP_COREDUMP, &reset_context.flags);
+
+	amdgpu_device_gpu_recover(adev, NULL, &reset_context);
+}
+
 #endif
 
 void amdgpu_debugfs_fence_init(struct amdgpu_device *adev)
@@ -814,9 +965,12 @@ void amdgpu_debugfs_fence_init(struct amdgpu_device *adev)
 	debugfs_create_file("amdgpu_fence_info", 0444, root, adev,
 			    &amdgpu_debugfs_fence_info_fops);
 
-	if (!amdgpu_sriov_vf(adev))
+	if (!amdgpu_sriov_vf(adev)) {
+
+		INIT_WORK(&adev->reset_work, amdgpu_debugfs_reset_work);
 		debugfs_create_file("amdgpu_gpu_recover", 0444, root, adev,
 				    &amdgpu_debugfs_gpu_recover_fops);
+	}
 #endif
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_fru_eeprom.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_fru_eeprom.c
index 7709caeb233d..b0082aa7f3c6 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_fru_eeprom.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_fru_eeprom.c
@@ -29,10 +29,11 @@
 #include "amdgpu_fru_eeprom.h"
 #include "amdgpu_eeprom.h"
 
-#define FRU_EEPROM_MADDR        0x60000
-#define I2C_PRODUCT_INFO_OFFSET 0xC0
+#define FRU_EEPROM_MADDR_6      0x60000
+#define FRU_EEPROM_MADDR_8      0x80000
+#define FRU_EEPROM_MADDR_INV    0xFFFFF
 
-static bool is_fru_eeprom_supported(struct amdgpu_device *adev)
+static bool is_fru_eeprom_supported(struct amdgpu_device *adev, u32 *fru_addr)
 {
 	/* Only server cards have the FRU EEPROM
 	 * TODO: See if we can figure this out dynamically instead of
@@ -40,152 +41,368 @@ static bool is_fru_eeprom_supported(struct amdgpu_device *adev)
 	 */
 	struct atom_context *atom_ctx = adev->mode_info.atom_context;
 
-	/* VBIOS is of the format ###-DXXXYY-##. For SKU identification,
+	/* The i2c access is blocked on VF
+	 * TODO: Need other way to get the info
+	 * Also, FRU not valid for APU devices.
+	 */
+	if (amdgpu_sriov_vf(adev) || (adev->flags & AMD_IS_APU))
+		return false;
+
+	/* The default I2C EEPROM address of the FRU.
+	 */
+	if (fru_addr)
+		*fru_addr = FRU_EEPROM_MADDR_8;
+
+	/* VBIOS is of the format ###-DXXXYYYY-##. For SKU identification,
 	 * we can use just the "DXXX" portion. If there were more models, we
 	 * could convert the 3 characters to a hex integer and use a switch
 	 * for ease/speed/readability. For now, 2 string comparisons are
 	 * reasonable and not too expensive
 	 */
-	switch (adev->asic_type) {
-	case CHIP_VEGA20:
-		/* D161 and D163 are the VG20 server SKUs */
-		if (strnstr(atom_ctx->vbios_version, "D161",
-			    sizeof(atom_ctx->vbios_version)) ||
-		    strnstr(atom_ctx->vbios_version, "D163",
-			    sizeof(atom_ctx->vbios_version)))
+	switch (amdgpu_ip_version(adev, MP1_HWIP, 0)) {
+	case IP_VERSION(11, 0, 2):
+		switch (adev->asic_type) {
+		case CHIP_VEGA20:
+			/* D161 and D163 are the VG20 server SKUs */
+			if (atom_ctx && (strnstr(atom_ctx->vbios_pn, "D161",
+						 sizeof(atom_ctx->vbios_pn)) ||
+					 strnstr(atom_ctx->vbios_pn, "D163",
+						 sizeof(atom_ctx->vbios_pn)))) {
+				if (fru_addr)
+					*fru_addr = FRU_EEPROM_MADDR_6;
+				return true;
+			} else {
+				return false;
+			}
+		case CHIP_ARCTURUS:
+		default:
+			return false;
+		}
+	case IP_VERSION(11, 0, 7):
+		if (atom_ctx && strnstr(atom_ctx->vbios_pn, "D603",
+					sizeof(atom_ctx->vbios_pn))) {
+			if (strnstr(atom_ctx->vbios_pn, "D603GLXE",
+				    sizeof(atom_ctx->vbios_pn))) {
+				return false;
+			}
+
+			if (fru_addr)
+				*fru_addr = FRU_EEPROM_MADDR_6;
 			return true;
-		else
+
+		} else {
 			return false;
+		}
+	case IP_VERSION(13, 0, 2):
+		/* All Aldebaran SKUs have an FRU */
+		if (atom_ctx && !strnstr(atom_ctx->vbios_pn, "D673",
+					 sizeof(atom_ctx->vbios_pn)))
+			if (fru_addr)
+				*fru_addr = FRU_EEPROM_MADDR_6;
+		return true;
+	case IP_VERSION(13, 0, 6):
+	case IP_VERSION(13, 0, 14):
+			if (fru_addr)
+				*fru_addr = FRU_EEPROM_MADDR_8;
+			return true;
+	case IP_VERSION(13, 0, 12):
+			if (fru_addr)
+				*fru_addr = FRU_EEPROM_MADDR_INV;
+			return true;
 	default:
 		return false;
 	}
 }
 
-static int amdgpu_fru_read_eeprom(struct amdgpu_device *adev, uint32_t addrptr,
-				  unsigned char *buff)
+int amdgpu_fru_get_product_info(struct amdgpu_device *adev)
 {
-	int ret, size;
+	struct amdgpu_fru_info *fru_info;
+	unsigned char buf[8], *pia;
+	u32 addr, fru_addr;
+	int size, len;
+	u8 csum;
 
-	ret = amdgpu_eeprom_read(&adev->pm.smu_i2c, addrptr, buff, 1);
-	if (ret < 1) {
-		DRM_WARN("FRU: Failed to get size field");
-		return ret;
-	}
+	if (!is_fru_eeprom_supported(adev, &fru_addr))
+		return 0;
 
-	/* The size returned by the i2c requires subtraction of 0xC0 since the
-	 * size apparently always reports as 0xC0+actual size.
-	 */
-	size = buff[0] - I2C_PRODUCT_INFO_OFFSET;
+	/* FRU data avaialble, but no direct EEPROM access */
+	if (fru_addr == FRU_EEPROM_MADDR_INV)
+		return 0;
 
-	ret = amdgpu_eeprom_read(&adev->pm.smu_i2c, addrptr + 1, buff, size);
-	if (ret < 1) {
-		DRM_WARN("FRU: Failed to get data field");
-		return ret;
+	if (!adev->fru_info) {
+		adev->fru_info = kzalloc(sizeof(*adev->fru_info), GFP_KERNEL);
+		if (!adev->fru_info)
+			return -ENOMEM;
 	}
 
-	return size;
-}
-
-int amdgpu_fru_get_product_info(struct amdgpu_device *adev)
-{
-	unsigned char buff[34];
-	u32 addrptr;
-	int size, len;
-
-	if (!is_fru_eeprom_supported(adev))
-		return 0;
+	fru_info = adev->fru_info;
+	/* For Arcturus-and-later, default value of serial_number is unique_id
+	 * so convert it to a 16-digit HEX string for convenience and
+	 * backwards-compatibility.
+	 */
+	sprintf(fru_info->serial, "%llx", adev->unique_id);
 
 	/* If algo exists, it means that the i2c_adapter's initialized */
-	if (!adev->pm.smu_i2c.algo) {
-		DRM_WARN("Cannot access FRU, EEPROM accessor not initialized");
+	if (!adev->pm.fru_eeprom_i2c_bus || !adev->pm.fru_eeprom_i2c_bus->algo) {
+		dev_warn(adev->dev,
+			 "Cannot access FRU, EEPROM accessor not initialized");
 		return -ENODEV;
 	}
 
-	/* There's a lot of repetition here. This is due to the FRU having
-	 * variable-length fields. To get the information, we have to find the
-	 * size of each field, and then keep reading along and reading along
-	 * until we get all of the data that we want. We use addrptr to track
-	 * the address as we go
-	 */
-
-	/* The first fields are all of size 1-byte, from 0-7 are offsets that
-	 * contain information that isn't useful to us.
-	 * Bytes 8-a are all 1-byte and refer to the size of the entire struct,
-	 * and the language field, so just start from 0xb, manufacturer size
-	 */
-	addrptr = FRU_EEPROM_MADDR + 0xb;
-	size = amdgpu_fru_read_eeprom(adev, addrptr, buff);
-	if (size < 1) {
-		DRM_ERROR("Failed to read FRU Manufacturer, ret:%d", size);
-		return -EINVAL;
+	/* Read the IPMI Common header */
+	len = amdgpu_eeprom_read(adev->pm.fru_eeprom_i2c_bus, fru_addr, buf,
+				 sizeof(buf));
+	if (len != 8) {
+		dev_err(adev->dev, "Couldn't read the IPMI Common Header: %d",
+			len);
+		return len < 0 ? len : -EIO;
 	}
 
-	/* Increment the addrptr by the size of the field, and 1 due to the
-	 * size field being 1 byte. This pattern continues below.
-	 */
-	addrptr += size + 1;
-	size = amdgpu_fru_read_eeprom(adev, addrptr, buff);
-	if (size < 1) {
-		DRM_ERROR("Failed to read FRU product name, ret:%d", size);
-		return -EINVAL;
+	if (buf[0] != 1) {
+		dev_err(adev->dev, "Bad IPMI Common Header version: 0x%02x",
+			buf[0]);
+		return -EIO;
 	}
 
-	len = size;
-	/* Product name should only be 32 characters. Any more,
-	 * and something could be wrong. Cap it at 32 to be safe
-	 */
-	if (len >= sizeof(adev->product_name)) {
-		DRM_WARN("FRU Product Number is larger than 32 characters. This is likely a mistake");
-		len = sizeof(adev->product_name) - 1;
+	for (csum = 0; len > 0; len--)
+		csum += buf[len - 1];
+	if (csum) {
+		dev_err(adev->dev, "Bad IPMI Common Header checksum: 0x%02x",
+			csum);
+		return -EIO;
 	}
-	/* Start at 2 due to buff using fields 0 and 1 for the address */
-	memcpy(adev->product_name, &buff[2], len);
-	adev->product_name[len] = '\0';
-
-	addrptr += size + 1;
-	size = amdgpu_fru_read_eeprom(adev, addrptr, buff);
-	if (size < 1) {
-		DRM_ERROR("Failed to read FRU product number, ret:%d", size);
-		return -EINVAL;
+
+	/* Get the offset to the Product Info Area (PIA). */
+	addr = buf[4] * 8;
+	if (!addr)
+		return 0;
+
+	/* Get the absolute address to the PIA. */
+	addr += fru_addr;
+
+	/* Read the header of the PIA. */
+	len = amdgpu_eeprom_read(adev->pm.fru_eeprom_i2c_bus, addr, buf, 3);
+	if (len != 3) {
+		dev_err(adev->dev,
+			"Couldn't read the Product Info Area header: %d", len);
+		return len < 0 ? len : -EIO;
 	}
 
-	len = size;
-	/* Product number should only be 16 characters. Any more,
-	 * and something could be wrong. Cap it at 16 to be safe
-	 */
-	if (len >= sizeof(adev->product_number)) {
-		DRM_WARN("FRU Product Number is larger than 16 characters. This is likely a mistake");
-		len = sizeof(adev->product_number) - 1;
+	if (buf[0] != 1) {
+		dev_err(adev->dev, "Bad IPMI Product Info Area version: 0x%02x",
+			buf[0]);
+		return -EIO;
 	}
-	memcpy(adev->product_number, &buff[2], len);
-	adev->product_number[len] = '\0';
 
-	addrptr += size + 1;
-	size = amdgpu_fru_read_eeprom(adev, addrptr, buff);
+	size = buf[1] * 8;
+	pia = kzalloc(size, GFP_KERNEL);
+	if (!pia)
+		return -ENOMEM;
 
-	if (size < 1) {
-		DRM_ERROR("Failed to read FRU product version, ret:%d", size);
-		return -EINVAL;
+	/* Read the whole PIA. */
+	len = amdgpu_eeprom_read(adev->pm.fru_eeprom_i2c_bus, addr, pia, size);
+	if (len != size) {
+		kfree(pia);
+		dev_err(adev->dev, "Couldn't read the Product Info Area: %d",
+			len);
+		return len < 0 ? len : -EIO;
 	}
 
-	addrptr += size + 1;
-	size = amdgpu_fru_read_eeprom(adev, addrptr, buff);
-
-	if (size < 1) {
-		DRM_ERROR("Failed to read FRU serial number, ret:%d", size);
-		return -EINVAL;
+	for (csum = 0; size > 0; size--)
+		csum += pia[size - 1];
+	if (csum) {
+		dev_err(adev->dev, "Bad Product Info Area checksum: 0x%02x",
+			csum);
+		kfree(pia);
+		return -EIO;
 	}
 
-	len = size;
-	/* Serial number should only be 16 characters. Any more,
-	 * and something could be wrong. Cap it at 16 to be safe
+	/* Now extract useful information from the PIA.
+	 *
+	 * Read Manufacturer Name field whose length is [3].
 	 */
-	if (len >= sizeof(adev->serial)) {
-		DRM_WARN("FRU Serial Number is larger than 16 characters. This is likely a mistake");
-		len = sizeof(adev->serial) - 1;
-	}
-	memcpy(adev->serial, &buff[2], len);
-	adev->serial[len] = '\0';
+	addr = 3;
+	if (addr + 1 >= len)
+		goto Out;
+	memcpy(fru_info->manufacturer_name, pia + addr + 1,
+	       min_t(size_t, sizeof(fru_info->manufacturer_name),
+		     pia[addr] & 0x3F));
+	fru_info->manufacturer_name[sizeof(fru_info->manufacturer_name) - 1] =
+		'\0';
+
+	/* Read Product Name field. */
+	addr += 1 + (pia[addr] & 0x3F);
+	if (addr + 1 >= len)
+		goto Out;
+	memcpy(fru_info->product_name, pia + addr + 1,
+	       min_t(size_t, sizeof(fru_info->product_name), pia[addr] & 0x3F));
+	fru_info->product_name[sizeof(fru_info->product_name) - 1] = '\0';
+
+	/* Go to the Product Part/Model Number field. */
+	addr += 1 + (pia[addr] & 0x3F);
+	if (addr + 1 >= len)
+		goto Out;
+	memcpy(fru_info->product_number, pia + addr + 1,
+	       min_t(size_t, sizeof(fru_info->product_number),
+		     pia[addr] & 0x3F));
+	fru_info->product_number[sizeof(fru_info->product_number) - 1] = '\0';
+
+	/* Go to the Product Version field. */
+	addr += 1 + (pia[addr] & 0x3F);
+
+	/* Go to the Product Serial Number field. */
+	addr += 1 + (pia[addr] & 0x3F);
+	if (addr + 1 >= len)
+		goto Out;
+	memcpy(fru_info->serial, pia + addr + 1,
+	       min_t(size_t, sizeof(fru_info->serial), pia[addr] & 0x3F));
+	fru_info->serial[sizeof(fru_info->serial) - 1] = '\0';
+
+	/* Asset Tag field */
+	addr += 1 + (pia[addr] & 0x3F);
 
+	/* FRU File Id field. This could be 'null'. */
+	addr += 1 + (pia[addr] & 0x3F);
+	if ((addr + 1 >= len) || !(pia[addr] & 0x3F))
+		goto Out;
+	memcpy(fru_info->fru_id, pia + addr + 1,
+	       min_t(size_t, sizeof(fru_info->fru_id), pia[addr] & 0x3F));
+	fru_info->fru_id[sizeof(fru_info->fru_id) - 1] = '\0';
+
+Out:
+	kfree(pia);
 	return 0;
 }
+
+/**
+ * DOC: product_name
+ *
+ * The amdgpu driver provides a sysfs API for reporting the product name
+ * for the device
+ * The file product_name is used for this and returns the product name
+ * as returned from the FRU.
+ * NOTE: This is only available for certain server cards
+ */
+
+static ssize_t amdgpu_fru_product_name_show(struct device *dev,
+					    struct device_attribute *attr,
+					    char *buf)
+{
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = drm_to_adev(ddev);
+
+	return sysfs_emit(buf, "%s\n", adev->fru_info->product_name);
+}
+
+static DEVICE_ATTR(product_name, 0444, amdgpu_fru_product_name_show, NULL);
+
+/**
+ * DOC: product_number
+ *
+ * The amdgpu driver provides a sysfs API for reporting the part number
+ * for the device
+ * The file product_number is used for this and returns the part number
+ * as returned from the FRU.
+ * NOTE: This is only available for certain server cards
+ */
+
+static ssize_t amdgpu_fru_product_number_show(struct device *dev,
+					      struct device_attribute *attr,
+					      char *buf)
+{
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = drm_to_adev(ddev);
+
+	return sysfs_emit(buf, "%s\n", adev->fru_info->product_number);
+}
+
+static DEVICE_ATTR(product_number, 0444, amdgpu_fru_product_number_show, NULL);
+
+/**
+ * DOC: serial_number
+ *
+ * The amdgpu driver provides a sysfs API for reporting the serial number
+ * for the device
+ * The file serial_number is used for this and returns the serial number
+ * as returned from the FRU.
+ * NOTE: This is only available for certain server cards
+ */
+
+static ssize_t amdgpu_fru_serial_number_show(struct device *dev,
+					     struct device_attribute *attr,
+					     char *buf)
+{
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = drm_to_adev(ddev);
+
+	return sysfs_emit(buf, "%s\n", adev->fru_info->serial);
+}
+
+static DEVICE_ATTR(serial_number, 0444, amdgpu_fru_serial_number_show, NULL);
+
+/**
+ * DOC: fru_id
+ *
+ * The amdgpu driver provides a sysfs API for reporting FRU File Id
+ * for the device.
+ * The file fru_id is used for this and returns the File Id value
+ * as returned from the FRU.
+ * NOTE: This is only available for certain server cards
+ */
+
+static ssize_t amdgpu_fru_id_show(struct device *dev,
+				  struct device_attribute *attr, char *buf)
+{
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = drm_to_adev(ddev);
+
+	return sysfs_emit(buf, "%s\n", adev->fru_info->fru_id);
+}
+
+static DEVICE_ATTR(fru_id, 0444, amdgpu_fru_id_show, NULL);
+
+/**
+ * DOC: manufacturer
+ *
+ * The amdgpu driver provides a sysfs API for reporting manufacturer name from
+ * FRU information.
+ * The file manufacturer returns the value as returned from the FRU.
+ * NOTE: This is only available for certain server cards
+ */
+
+static ssize_t amdgpu_fru_manufacturer_name_show(struct device *dev,
+						 struct device_attribute *attr,
+						 char *buf)
+{
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = drm_to_adev(ddev);
+
+	return sysfs_emit(buf, "%s\n", adev->fru_info->manufacturer_name);
+}
+
+static DEVICE_ATTR(manufacturer, 0444, amdgpu_fru_manufacturer_name_show, NULL);
+
+static const struct attribute *amdgpu_fru_attributes[] = {
+	&dev_attr_product_name.attr,
+	&dev_attr_product_number.attr,
+	&dev_attr_serial_number.attr,
+	&dev_attr_fru_id.attr,
+	&dev_attr_manufacturer.attr,
+	NULL
+};
+
+int amdgpu_fru_sysfs_init(struct amdgpu_device *adev)
+{
+	if (!is_fru_eeprom_supported(adev, NULL) || !adev->fru_info)
+		return 0;
+
+	return sysfs_create_files(&adev->dev->kobj, amdgpu_fru_attributes);
+}
+
+void amdgpu_fru_sysfs_fini(struct amdgpu_device *adev)
+{
+	if (!adev->fru_info)
+		return;
+
+	sysfs_remove_files(&adev->dev->kobj, amdgpu_fru_attributes);
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_fru_eeprom.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_fru_eeprom.h
index 1308d976d60e..98f3196599ef 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_fru_eeprom.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_fru_eeprom.h
@@ -24,6 +24,19 @@
 #ifndef __AMDGPU_FRU_EEPROM_H__
 #define __AMDGPU_FRU_EEPROM_H__
 
+#define AMDGPU_PRODUCT_NAME_LEN 64
+
+/* FRU product information */
+struct amdgpu_fru_info {
+	char				product_number[20];
+	char				product_name[AMDGPU_PRODUCT_NAME_LEN];
+	char				serial[20];
+	char				manufacturer_name[32];
+	char				fru_id[50];
+};
+
 int amdgpu_fru_get_product_info(struct amdgpu_device *adev);
+int amdgpu_fru_sysfs_init(struct amdgpu_device *adev);
+void amdgpu_fru_sysfs_fini(struct amdgpu_device *adev);
 
 #endif  // __AMDGPU_FRU_EEPROM_H__
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_fw_attestation.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_fw_attestation.c
index 2ca3c329de6d..328a1b963548 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_fw_attestation.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_fw_attestation.c
@@ -32,17 +32,15 @@
 #include "soc15_common.h"
 
 #define FW_ATTESTATION_DB_COOKIE        0x143b6a37
-#define FW_ATTESTATION_RECORD_VALID  	1
+#define FW_ATTESTATION_RECORD_VALID	1
 #define FW_ATTESTATION_MAX_SIZE		4096
 
-typedef struct FW_ATT_DB_HEADER
-{
+struct FW_ATT_DB_HEADER {
 	uint32_t AttDbVersion;           /* version of the fwar feature */
 	uint32_t AttDbCookie;            /* cookie as an extra check for corrupt data */
-} FW_ATT_DB_HEADER;
+};
 
-typedef struct FW_ATT_RECORD
-{
+struct FW_ATT_RECORD {
 	uint16_t AttFwIdV1;              /* Legacy FW Type field */
 	uint16_t AttFwIdV2;              /* V2 FW ID field */
 	uint32_t AttFWVersion;           /* FW Version */
@@ -50,7 +48,7 @@ typedef struct FW_ATT_RECORD
 	uint8_t  AttSource;              /* FW source indicator */
 	uint8_t  RecordValid;            /* Indicates whether the record is a valid entry */
 	uint32_t AttFwTaId;              /* Ta ID (only in TA Attestation Table) */
-} FW_ATT_RECORD;
+};
 
 static ssize_t amdgpu_fw_attestation_debugfs_read(struct file *f,
 						  char __user *buf,
@@ -60,15 +58,15 @@ static ssize_t amdgpu_fw_attestation_debugfs_read(struct file *f,
 	struct amdgpu_device *adev = (struct amdgpu_device *)file_inode(f)->i_private;
 	uint64_t records_addr = 0;
 	uint64_t vram_pos = 0;
-	FW_ATT_DB_HEADER fw_att_hdr = {0};
-	FW_ATT_RECORD fw_att_record = {0};
+	struct FW_ATT_DB_HEADER fw_att_hdr = {0};
+	struct FW_ATT_RECORD fw_att_record = {0};
 
-	if (size < sizeof(FW_ATT_RECORD)) {
+	if (size < sizeof(struct FW_ATT_RECORD)) {
 		DRM_WARN("FW attestation input buffer not enough memory");
 		return -EINVAL;
 	}
 
-	if ((*pos + sizeof(FW_ATT_DB_HEADER)) >= FW_ATTESTATION_MAX_SIZE) {
+	if ((*pos + sizeof(struct FW_ATT_DB_HEADER)) >= FW_ATTESTATION_MAX_SIZE) {
 		DRM_WARN("FW attestation out of bounds");
 		return 0;
 	}
@@ -83,8 +81,8 @@ static ssize_t amdgpu_fw_attestation_debugfs_read(struct file *f,
 	if (*pos == 0) {
 		amdgpu_device_vram_access(adev,
 					  vram_pos,
-					  (uint32_t*)&fw_att_hdr,
-					  sizeof(FW_ATT_DB_HEADER),
+					  (uint32_t *)&fw_att_hdr,
+					  sizeof(struct FW_ATT_DB_HEADER),
 					  false);
 
 		if (fw_att_hdr.AttDbCookie != FW_ATTESTATION_DB_COOKIE) {
@@ -96,20 +94,20 @@ static ssize_t amdgpu_fw_attestation_debugfs_read(struct file *f,
 	}
 
 	amdgpu_device_vram_access(adev,
-				  vram_pos + sizeof(FW_ATT_DB_HEADER) + *pos,
-				  (uint32_t*)&fw_att_record,
-				  sizeof(FW_ATT_RECORD),
+				  vram_pos + sizeof(struct FW_ATT_DB_HEADER) + *pos,
+				  (uint32_t *)&fw_att_record,
+				  sizeof(struct FW_ATT_RECORD),
 				  false);
 
 	if (fw_att_record.RecordValid != FW_ATTESTATION_RECORD_VALID)
 		return 0;
 
-	if (copy_to_user(buf, (void*)&fw_att_record, sizeof(FW_ATT_RECORD)))
+	if (copy_to_user(buf, (void *)&fw_att_record, sizeof(struct FW_ATT_RECORD)))
 		return -EINVAL;
 
-	*pos += sizeof(FW_ATT_RECORD);
+	*pos += sizeof(struct FW_ATT_RECORD);
 
-	return sizeof(FW_ATT_RECORD);
+	return sizeof(struct FW_ATT_RECORD);
 }
 
 static const struct file_operations amdgpu_fw_attestation_debugfs_ops = {
@@ -124,6 +122,10 @@ static int amdgpu_is_fw_attestation_supported(struct amdgpu_device *adev)
 	if (adev->flags & AMD_IS_APU)
 		return 0;
 
+	if (amdgpu_ip_version(adev, MP0_HWIP, 0) == IP_VERSION(14, 0, 2) ||
+	    amdgpu_ip_version(adev, MP0_HWIP, 0) == IP_VERSION(14, 0, 3))
+		return 0;
+
 	if (adev->asic_type >= CHIP_SIENNA_CICHLID)
 		return 1;
 
@@ -136,7 +138,7 @@ void amdgpu_fw_attestation_debugfs_init(struct amdgpu_device *adev)
 		return;
 
 	debugfs_create_file("amdgpu_fw_attestation",
-			    S_IRUSR,
+			    0400,
 			    adev_to_drm(adev)->primary->debugfs_root,
 			    adev,
 			    &amdgpu_fw_attestation_debugfs_ops);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c
index d3e4203f6217..d2237ce9da70 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c
@@ -34,7 +34,9 @@
 #include <asm/set_memory.h>
 #endif
 #include "amdgpu.h"
+#include "amdgpu_reset.h"
 #include <drm/drm_drv.h>
+#include <drm/ttm/ttm_tt.h>
 
 /*
  * GART
@@ -76,8 +78,9 @@ static int amdgpu_gart_dummy_page_init(struct amdgpu_device *adev)
 
 	if (adev->dummy_page_addr)
 		return 0;
-	adev->dummy_page_addr = dma_map_page(&adev->pdev->dev, dummy_page, 0,
-					     PAGE_SIZE, DMA_BIDIRECTIONAL);
+	adev->dummy_page_addr = dma_map_page_attrs(&adev->pdev->dev, dummy_page, 0,
+							PAGE_SIZE, DMA_BIDIRECTIONAL,
+							DMA_ATTR_SKIP_CPU_SYNC);
 	if (dma_mapping_error(&adev->pdev->dev, adev->dummy_page_addr)) {
 		dev_err(&adev->pdev->dev, "Failed to DMA MAP the dummy page\n");
 		adev->dummy_page_addr = 0;
@@ -97,97 +100,174 @@ void amdgpu_gart_dummy_page_fini(struct amdgpu_device *adev)
 {
 	if (!adev->dummy_page_addr)
 		return;
-	dma_unmap_page(&adev->pdev->dev, adev->dummy_page_addr, PAGE_SIZE,
-		       DMA_BIDIRECTIONAL);
+	dma_unmap_page_attrs(&adev->pdev->dev, adev->dummy_page_addr, PAGE_SIZE,
+				DMA_BIDIRECTIONAL,
+				DMA_ATTR_SKIP_CPU_SYNC);
 	adev->dummy_page_addr = 0;
 }
 
 /**
- * amdgpu_gart_table_vram_alloc - allocate vram for gart page table
+ * amdgpu_gart_table_ram_alloc - allocate system ram for gart page table
  *
  * @adev: amdgpu_device pointer
  *
- * Allocate video memory for GART page table
- * (pcie r4xx, r5xx+).  These asics require the
- * gart table to be in video memory.
+ * Allocate system memory for GART page table for ASICs that don't have
+ * dedicated VRAM.
  * Returns 0 for success, error for failure.
  */
-int amdgpu_gart_table_vram_alloc(struct amdgpu_device *adev)
+int amdgpu_gart_table_ram_alloc(struct amdgpu_device *adev)
 {
-	int r;
+	unsigned int order = get_order(adev->gart.table_size);
+	gfp_t gfp_flags = GFP_KERNEL | __GFP_ZERO;
+	struct amdgpu_bo *bo = NULL;
+	struct sg_table *sg = NULL;
+	struct amdgpu_bo_param bp;
+	dma_addr_t dma_addr;
+	struct page *p;
+	unsigned long x;
+	int ret;
+
+	if (adev->gart.bo != NULL)
+		return 0;
 
-	if (adev->gart.bo == NULL) {
-		struct amdgpu_bo_param bp;
-
-		memset(&bp, 0, sizeof(bp));
-		bp.size = adev->gart.table_size;
-		bp.byte_align = PAGE_SIZE;
-		bp.domain = AMDGPU_GEM_DOMAIN_VRAM;
-		bp.flags = AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED |
-			AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS;
-		bp.type = ttm_bo_type_kernel;
-		bp.resv = NULL;
-		bp.bo_ptr_size = sizeof(struct amdgpu_bo);
-
-		r = amdgpu_bo_create(adev, &bp, &adev->gart.bo);
-		if (r) {
-			return r;
-		}
+	p = alloc_pages(gfp_flags, order);
+	if (!p)
+		return -ENOMEM;
+
+	/* assign pages to this device */
+	for (x = 0; x < (1UL << order); x++)
+		p[x].mapping = adev->mman.bdev.dev_mapping;
+
+	/* If the hardware does not support UTCL2 snooping of the CPU caches
+	 * then set_memory_wc() could be used as a workaround to mark the pages
+	 * as write combine memory.
+	 */
+	dma_addr = dma_map_page(&adev->pdev->dev, p, 0, adev->gart.table_size,
+				DMA_BIDIRECTIONAL);
+	if (dma_mapping_error(&adev->pdev->dev, dma_addr)) {
+		dev_err(&adev->pdev->dev, "Failed to DMA MAP the GART BO page\n");
+		__free_pages(p, order);
+		p = NULL;
+		return -EFAULT;
+	}
+
+	dev_info(adev->dev, "%s dma_addr:%pad\n", __func__, &dma_addr);
+	/* Create SG table */
+	sg = kmalloc(sizeof(*sg), GFP_KERNEL);
+	if (!sg) {
+		ret = -ENOMEM;
+		goto error;
 	}
+	ret = sg_alloc_table(sg, 1, GFP_KERNEL);
+	if (ret)
+		goto error;
+
+	sg_dma_address(sg->sgl) = dma_addr;
+	sg->sgl->length = adev->gart.table_size;
+#ifdef CONFIG_NEED_SG_DMA_LENGTH
+	sg->sgl->dma_length = adev->gart.table_size;
+#endif
+	/* Create SG BO */
+	memset(&bp, 0, sizeof(bp));
+	bp.size = adev->gart.table_size;
+	bp.byte_align = PAGE_SIZE;
+	bp.domain = AMDGPU_GEM_DOMAIN_CPU;
+	bp.type = ttm_bo_type_sg;
+	bp.resv = NULL;
+	bp.bo_ptr_size = sizeof(struct amdgpu_bo);
+	bp.flags = 0;
+	ret = amdgpu_bo_create(adev, &bp, &bo);
+	if (ret)
+		goto error;
+
+	bo->tbo.sg = sg;
+	bo->tbo.ttm->sg = sg;
+	bo->allowed_domains = AMDGPU_GEM_DOMAIN_GTT;
+	bo->preferred_domains = AMDGPU_GEM_DOMAIN_GTT;
+
+	ret = amdgpu_bo_reserve(bo, true);
+	if (ret) {
+		dev_err(adev->dev, "(%d) failed to reserve bo for GART system bo\n", ret);
+		goto error;
+	}
+
+	ret = amdgpu_bo_pin(bo, AMDGPU_GEM_DOMAIN_GTT);
+	WARN(ret, "Pinning the GART table failed");
+	if (ret)
+		goto error_resv;
+
+	adev->gart.bo = bo;
+	adev->gart.ptr = page_to_virt(p);
+	/* Make GART table accessible in VMID0 */
+	ret = amdgpu_ttm_alloc_gart(&adev->gart.bo->tbo);
+	if (ret)
+		amdgpu_gart_table_ram_free(adev);
+	amdgpu_bo_unreserve(bo);
+
 	return 0;
+
+error_resv:
+	amdgpu_bo_unreserve(bo);
+error:
+	amdgpu_bo_unref(&bo);
+	if (sg) {
+		sg_free_table(sg);
+		kfree(sg);
+	}
+	__free_pages(p, order);
+	return ret;
 }
 
 /**
- * amdgpu_gart_table_vram_pin - pin gart page table in vram
+ * amdgpu_gart_table_ram_free - free gart page table system ram
  *
  * @adev: amdgpu_device pointer
  *
- * Pin the GART page table in vram so it will not be moved
- * by the memory manager (pcie r4xx, r5xx+).  These asics require the
- * gart table to be in video memory.
- * Returns 0 for success, error for failure.
+ * Free the system memory used for the GART page tableon ASICs that don't
+ * have dedicated VRAM.
  */
-int amdgpu_gart_table_vram_pin(struct amdgpu_device *adev)
+void amdgpu_gart_table_ram_free(struct amdgpu_device *adev)
 {
-	int r;
-
-	r = amdgpu_bo_reserve(adev->gart.bo, false);
-	if (unlikely(r != 0))
-		return r;
-	r = amdgpu_bo_pin(adev->gart.bo, AMDGPU_GEM_DOMAIN_VRAM);
-	if (r) {
+	unsigned int order = get_order(adev->gart.table_size);
+	struct sg_table *sg = adev->gart.bo->tbo.sg;
+	struct page *p;
+	unsigned long x;
+	int ret;
+
+	ret = amdgpu_bo_reserve(adev->gart.bo, false);
+	if (!ret) {
+		amdgpu_bo_unpin(adev->gart.bo);
 		amdgpu_bo_unreserve(adev->gart.bo);
-		return r;
 	}
-	r = amdgpu_bo_kmap(adev->gart.bo, &adev->gart.ptr);
-	if (r)
-		amdgpu_bo_unpin(adev->gart.bo);
-	amdgpu_bo_unreserve(adev->gart.bo);
-	return r;
+	amdgpu_bo_unref(&adev->gart.bo);
+	sg_free_table(sg);
+	kfree(sg);
+	p = virt_to_page(adev->gart.ptr);
+	for (x = 0; x < (1UL << order); x++)
+		p[x].mapping = NULL;
+	__free_pages(p, order);
+
+	adev->gart.ptr = NULL;
 }
 
 /**
- * amdgpu_gart_table_vram_unpin - unpin gart page table in vram
+ * amdgpu_gart_table_vram_alloc - allocate vram for gart page table
  *
  * @adev: amdgpu_device pointer
  *
- * Unpin the GART page table in vram (pcie r4xx, r5xx+).
- * These asics require the gart table to be in video memory.
+ * Allocate video memory for GART page table
+ * (pcie r4xx, r5xx+).  These asics require the
+ * gart table to be in video memory.
+ * Returns 0 for success, error for failure.
  */
-void amdgpu_gart_table_vram_unpin(struct amdgpu_device *adev)
+int amdgpu_gart_table_vram_alloc(struct amdgpu_device *adev)
 {
-	int r;
+	if (adev->gart.bo != NULL)
+		return 0;
 
-	if (adev->gart.bo == NULL) {
-		return;
-	}
-	r = amdgpu_bo_reserve(adev->gart.bo, true);
-	if (likely(r == 0)) {
-		amdgpu_bo_kunmap(adev->gart.bo);
-		amdgpu_bo_unpin(adev->gart.bo);
-		amdgpu_bo_unreserve(adev->gart.bo);
-		adev->gart.ptr = NULL;
-	}
+	return amdgpu_bo_create_kernel(adev,  adev->gart.table_size, PAGE_SIZE,
+				       AMDGPU_GEM_DOMAIN_VRAM, &adev->gart.bo,
+				       NULL, (void *)&adev->gart.ptr);
 }
 
 /**
@@ -201,11 +281,7 @@ void amdgpu_gart_table_vram_unpin(struct amdgpu_device *adev)
  */
 void amdgpu_gart_table_vram_free(struct amdgpu_device *adev)
 {
-	if (adev->gart.bo == NULL) {
-		return;
-	}
-	amdgpu_bo_unref(&adev->gart.bo);
-	adev->gart.ptr = NULL;
+	amdgpu_bo_free_kernel(&adev->gart.bo, NULL, (void *)&adev->gart.ptr);
 }
 
 /*
@@ -222,28 +298,24 @@ void amdgpu_gart_table_vram_free(struct amdgpu_device *adev)
  * replaces them with the dummy page (all asics).
  * Returns 0 for success, -EINVAL for failure.
  */
-int amdgpu_gart_unbind(struct amdgpu_device *adev, uint64_t offset,
+void amdgpu_gart_unbind(struct amdgpu_device *adev, uint64_t offset,
 			int pages)
 {
 	unsigned t;
-	unsigned p;
 	int i, j;
 	u64 page_base;
 	/* Starting from VEGA10, system bit must be 0 to mean invalid. */
 	uint64_t flags = 0;
 	int idx;
 
-	if (!adev->gart.ready) {
-		WARN(1, "trying to unbind memory from uninitialized GART !\n");
-		return -EINVAL;
-	}
+	if (!adev->gart.ptr)
+		return;
 
 	if (!drm_dev_enter(adev_to_drm(adev), &idx))
-		return 0;
+		return;
 
 	t = offset / AMDGPU_GPU_PAGE_SIZE;
-	p = t / AMDGPU_GPU_PAGES_IN_CPU_PAGE;
-	for (i = 0; i < pages; i++, p++) {
+	for (i = 0; i < pages; i++) {
 		page_base = adev->dummy_page_addr;
 		if (!adev->gart.ptr)
 			continue;
@@ -254,13 +326,9 @@ int amdgpu_gart_unbind(struct amdgpu_device *adev, uint64_t offset,
 			page_base += AMDGPU_GPU_PAGE_SIZE;
 		}
 	}
-	mb();
-	amdgpu_device_flush_hdp(adev, NULL);
-	for (i = 0; i < adev->num_vmhubs; i++)
-		amdgpu_gmc_flush_gpu_tlb(adev, 0, i, 0);
+	amdgpu_gart_invalidate_tlb(adev);
 
 	drm_dev_exit(idx);
-	return 0;
 }
 
 /**
@@ -276,7 +344,7 @@ int amdgpu_gart_unbind(struct amdgpu_device *adev, uint64_t offset,
  * Map the dma_addresses into GART entries (all asics).
  * Returns 0 for success, -EINVAL for failure.
  */
-int amdgpu_gart_map(struct amdgpu_device *adev, uint64_t offset,
+void amdgpu_gart_map(struct amdgpu_device *adev, uint64_t offset,
 		    int pages, dma_addr_t *dma_addr, uint64_t flags,
 		    void *dst)
 {
@@ -284,13 +352,8 @@ int amdgpu_gart_map(struct amdgpu_device *adev, uint64_t offset,
 	unsigned i, j, t;
 	int idx;
 
-	if (!adev->gart.ready) {
-		WARN(1, "trying to bind memory to uninitialized GART !\n");
-		return -EINVAL;
-	}
-
 	if (!drm_dev_enter(adev_to_drm(adev), &idx))
-		return 0;
+		return;
 
 	t = offset / AMDGPU_GPU_PAGE_SIZE;
 
@@ -302,7 +365,42 @@ int amdgpu_gart_map(struct amdgpu_device *adev, uint64_t offset,
 		}
 	}
 	drm_dev_exit(idx);
-	return 0;
+}
+
+/**
+ * amdgpu_gart_map_vram_range - map VRAM pages into the GART page table
+ *
+ * @adev: amdgpu_device pointer
+ * @pa: physical address of the first page to be mapped
+ * @start_page: first page to map in the GART aperture
+ * @num_pages: number of pages to be mapped
+ * @flags: page table entry flags
+ * @dst: CPU address of the GART table
+ *
+ * Binds a BO that is allocated in VRAM to the GART page table
+ * (all ASICs).
+ *
+ * Useful when a kernel BO is located in VRAM but
+ * needs to be accessed from the GART address space.
+ */
+void amdgpu_gart_map_vram_range(struct amdgpu_device *adev, uint64_t pa,
+				uint64_t start_page, uint64_t num_pages,
+				uint64_t flags, void *dst)
+{
+	u32 i, idx;
+
+	/* The SYSTEM flag indicates the pages aren't in VRAM. */
+	WARN_ON_ONCE(flags & AMDGPU_PTE_SYSTEM);
+
+	if (!drm_dev_enter(adev_to_drm(adev), &idx))
+		return;
+
+	for (i = 0; i < num_pages; ++i) {
+		amdgpu_gmc_set_pte_pde(adev, adev->gart.ptr,
+			start_page + i, pa + AMDGPU_GPU_PAGE_SIZE * i, flags);
+	}
+
+	drm_dev_exit(idx);
 }
 
 /**
@@ -318,20 +416,14 @@ int amdgpu_gart_map(struct amdgpu_device *adev, uint64_t offset,
  * (all asics).
  * Returns 0 for success, -EINVAL for failure.
  */
-int amdgpu_gart_bind(struct amdgpu_device *adev, uint64_t offset,
+void amdgpu_gart_bind(struct amdgpu_device *adev, uint64_t offset,
 		     int pages, dma_addr_t *dma_addr,
 		     uint64_t flags)
 {
-	if (!adev->gart.ready) {
-		WARN(1, "trying to bind memory to uninitialized GART !\n");
-		return -EINVAL;
-	}
-
 	if (!adev->gart.ptr)
-		return 0;
+		return;
 
-	return amdgpu_gart_map(adev, offset, pages, dma_addr, flags,
-			       adev->gart.ptr);
+	amdgpu_gart_map(adev, offset, pages, dma_addr, flags, adev->gart.ptr);
 }
 
 /**
@@ -346,9 +438,15 @@ void amdgpu_gart_invalidate_tlb(struct amdgpu_device *adev)
 {
 	int i;
 
+	if (!adev->gart.ptr)
+		return;
+
 	mb();
-	amdgpu_device_flush_hdp(adev, NULL);
-	for (i = 0; i < adev->num_vmhubs; i++)
+	if (down_read_trylock(&adev->reset_domain->sem)) {
+		amdgpu_device_flush_hdp(adev, NULL);
+		up_read(&adev->reset_domain->sem);
+	}
+	for_each_set_bit(i, adev->vmhubs_mask, AMDGPU_MAX_VMHUBS)
 		amdgpu_gmc_flush_gpu_tlb(adev, 0, i, 0);
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.h
index 78895413cf9f..d3118275ddae 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.h
@@ -46,24 +46,26 @@ struct amdgpu_gart {
 	unsigned			num_gpu_pages;
 	unsigned			num_cpu_pages;
 	unsigned			table_size;
-	bool				ready;
 
 	/* Asic default pte flags */
 	uint64_t			gart_pte_flags;
 };
 
+int amdgpu_gart_table_ram_alloc(struct amdgpu_device *adev);
+void amdgpu_gart_table_ram_free(struct amdgpu_device *adev);
 int amdgpu_gart_table_vram_alloc(struct amdgpu_device *adev);
 void amdgpu_gart_table_vram_free(struct amdgpu_device *adev);
-int amdgpu_gart_table_vram_pin(struct amdgpu_device *adev);
-void amdgpu_gart_table_vram_unpin(struct amdgpu_device *adev);
 int amdgpu_gart_init(struct amdgpu_device *adev);
 void amdgpu_gart_dummy_page_fini(struct amdgpu_device *adev);
-int amdgpu_gart_unbind(struct amdgpu_device *adev, uint64_t offset,
-		       int pages);
-int amdgpu_gart_map(struct amdgpu_device *adev, uint64_t offset,
-		    int pages, dma_addr_t *dma_addr, uint64_t flags,
-		    void *dst);
-int amdgpu_gart_bind(struct amdgpu_device *adev, uint64_t offset,
-		     int pages, dma_addr_t *dma_addr, uint64_t flags);
+void amdgpu_gart_unbind(struct amdgpu_device *adev, uint64_t offset,
+			int pages);
+void amdgpu_gart_map(struct amdgpu_device *adev, uint64_t offset,
+		     int pages, dma_addr_t *dma_addr, uint64_t flags,
+		     void *dst);
+void amdgpu_gart_bind(struct amdgpu_device *adev, uint64_t offset,
+		      int pages, dma_addr_t *dma_addr, uint64_t flags);
+void amdgpu_gart_map_vram_range(struct amdgpu_device *adev, uint64_t pa,
+				uint64_t start_page, uint64_t num_pages,
+				uint64_t flags, void *dst);
 void amdgpu_gart_invalidate_tlb(struct amdgpu_device *adev);
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
index a1e63ba4c54a..3e38c5db2987 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
@@ -33,14 +33,125 @@
 
 #include <drm/amdgpu_drm.h>
 #include <drm/drm_drv.h>
+#include <drm/drm_exec.h>
 #include <drm/drm_gem_ttm_helper.h>
+#include <drm/ttm/ttm_tt.h>
+#include <drm/drm_syncobj.h>
 
 #include "amdgpu.h"
 #include "amdgpu_display.h"
 #include "amdgpu_dma_buf.h"
+#include "amdgpu_hmm.h"
 #include "amdgpu_xgmi.h"
+#include "amdgpu_vm.h"
 
-static const struct drm_gem_object_funcs amdgpu_gem_object_funcs;
+static int
+amdgpu_gem_add_input_fence(struct drm_file *filp,
+			   uint64_t syncobj_handles_array,
+			   uint32_t num_syncobj_handles)
+{
+	struct dma_fence *fence;
+	uint32_t *syncobj_handles;
+	int ret, i;
+
+	if (!num_syncobj_handles)
+		return 0;
+
+	syncobj_handles = memdup_user(u64_to_user_ptr(syncobj_handles_array),
+				      size_mul(sizeof(uint32_t), num_syncobj_handles));
+	if (IS_ERR(syncobj_handles))
+		return PTR_ERR(syncobj_handles);
+
+	for (i = 0; i < num_syncobj_handles; i++) {
+
+		if (!syncobj_handles[i]) {
+			ret = -EINVAL;
+			goto free_memdup;
+		}
+
+		ret = drm_syncobj_find_fence(filp, syncobj_handles[i], 0, 0, &fence);
+		if (ret)
+			goto free_memdup;
+
+		dma_fence_wait(fence, false);
+
+		/* TODO: optimize async handling */
+		dma_fence_put(fence);
+	}
+
+free_memdup:
+	kfree(syncobj_handles);
+	return ret;
+}
+
+static int
+amdgpu_gem_update_timeline_node(struct drm_file *filp,
+				uint32_t syncobj_handle,
+				uint64_t point,
+				struct drm_syncobj **syncobj,
+				struct dma_fence_chain **chain)
+{
+	if (!syncobj_handle)
+		return 0;
+
+	/* Find the sync object */
+	*syncobj = drm_syncobj_find(filp, syncobj_handle);
+	if (!*syncobj)
+		return -ENOENT;
+
+	if (!point)
+		return 0;
+
+	/* Allocate the chain node */
+	*chain = dma_fence_chain_alloc();
+	if (!*chain) {
+		drm_syncobj_put(*syncobj);
+		return -ENOMEM;
+	}
+
+	return 0;
+}
+
+static void
+amdgpu_gem_update_bo_mapping(struct drm_file *filp,
+			     struct amdgpu_bo_va *bo_va,
+			     uint32_t operation,
+			     uint64_t point,
+			     struct dma_fence *fence,
+			     struct drm_syncobj *syncobj,
+			     struct dma_fence_chain *chain)
+{
+	struct amdgpu_bo *bo = bo_va ? bo_va->base.bo : NULL;
+	struct amdgpu_fpriv *fpriv = filp->driver_priv;
+	struct amdgpu_vm *vm = &fpriv->vm;
+	struct dma_fence *last_update;
+
+	if (!syncobj)
+		return;
+
+	/* Find the last update fence */
+	switch (operation) {
+	case AMDGPU_VA_OP_MAP:
+	case AMDGPU_VA_OP_REPLACE:
+		if (bo && (bo->tbo.base.resv == vm->root.bo->tbo.base.resv))
+			last_update = vm->last_update;
+		else
+			last_update = bo_va->last_pt_update;
+		break;
+	case AMDGPU_VA_OP_UNMAP:
+	case AMDGPU_VA_OP_CLEAR:
+		last_update = fence;
+		break;
+	default:
+		return;
+	}
+
+	/* Add fence to timeline */
+	if (!point)
+		drm_syncobj_replace_fence(syncobj, last_update);
+	else
+		drm_syncobj_add_point(syncobj, chain, last_update, point);
+}
 
 static vm_fault_t amdgpu_gem_fault(struct vm_fault *vmf)
 {
@@ -60,10 +171,10 @@ static vm_fault_t amdgpu_gem_fault(struct vm_fault *vmf)
 			goto unlock;
 		}
 
-		 ret = ttm_bo_vm_fault_reserved(vmf, vmf->vma->vm_page_prot,
-						TTM_BO_VM_NUM_PREFAULT);
+		ret = ttm_bo_vm_fault_reserved(vmf, vmf->vma->vm_page_prot,
+					       TTM_BO_VM_NUM_PREFAULT);
 
-		 drm_dev_exit(idx);
+		drm_dev_exit(idx);
 	} else {
 		ret = ttm_bo_vm_dummy_page(vmf, vmf->vma->vm_page_prot);
 	}
@@ -84,19 +195,17 @@ static const struct vm_operations_struct amdgpu_gem_vm_ops = {
 
 static void amdgpu_gem_object_free(struct drm_gem_object *gobj)
 {
-	struct amdgpu_bo *robj = gem_to_amdgpu_bo(gobj);
+	struct amdgpu_bo *aobj = gem_to_amdgpu_bo(gobj);
 
-	if (robj) {
-		amdgpu_mn_unregister(robj);
-		amdgpu_bo_unref(&robj);
-	}
+	amdgpu_hmm_unregister(aobj);
+	ttm_bo_fini(&aobj->tbo);
 }
 
 int amdgpu_gem_object_create(struct amdgpu_device *adev, unsigned long size,
 			     int alignment, u32 initial_domain,
 			     u64 flags, enum ttm_bo_type type,
 			     struct dma_resv *resv,
-			     struct drm_gem_object **obj)
+			     struct drm_gem_object **obj, int8_t xcp_id_plus1)
 {
 	struct amdgpu_bo *bo;
 	struct amdgpu_bo_user *ubo;
@@ -105,6 +214,7 @@ int amdgpu_gem_object_create(struct amdgpu_device *adev, unsigned long size,
 
 	memset(&bp, 0, sizeof(bp));
 	*obj = NULL;
+	flags |= AMDGPU_GEM_CREATE_VRAM_WIPE_ON_RELEASE;
 
 	bp.size = size;
 	bp.byte_align = alignment;
@@ -114,6 +224,7 @@ int amdgpu_gem_object_create(struct amdgpu_device *adev, unsigned long size,
 	bp.flags = flags;
 	bp.domain = initial_domain;
 	bp.bo_ptr_size = sizeof(struct amdgpu_bo);
+	bp.xcp_id_plus1 = xcp_id_plus1;
 
 	r = amdgpu_bo_create_user(adev, &bp, &ubo);
 	if (r)
@@ -121,7 +232,6 @@ int amdgpu_gem_object_create(struct amdgpu_device *adev, unsigned long size,
 
 	bo = &ubo->bo;
 	*obj = &bo->tbo.base;
-	(*obj)->funcs = &amdgpu_gem_object_funcs;
 
 	return 0;
 }
@@ -170,21 +280,63 @@ static int amdgpu_gem_object_open(struct drm_gem_object *obj,
 		return -EPERM;
 
 	if (abo->flags & AMDGPU_GEM_CREATE_VM_ALWAYS_VALID &&
-	    abo->tbo.base.resv != vm->root.bo->tbo.base.resv)
+	    !amdgpu_vm_is_bo_always_valid(vm, abo))
 		return -EPERM;
 
 	r = amdgpu_bo_reserve(abo, false);
 	if (r)
 		return r;
 
+	amdgpu_vm_bo_update_shared(abo);
 	bo_va = amdgpu_vm_bo_find(vm, abo);
-	if (!bo_va) {
+	if (!bo_va)
 		bo_va = amdgpu_vm_bo_add(adev, vm, abo);
-	} else {
+	else
 		++bo_va->ref_count;
+
+	/* attach gfx eviction fence */
+	r = amdgpu_eviction_fence_attach(&fpriv->evf_mgr, abo);
+	if (r) {
+		DRM_DEBUG_DRIVER("Failed to attach eviction fence to BO\n");
+		amdgpu_bo_unreserve(abo);
+		return r;
 	}
+
 	amdgpu_bo_unreserve(abo);
-	return 0;
+
+	/* Validate and add eviction fence to DMABuf imports with dynamic
+	 * attachment in compute VMs. Re-validation will be done by
+	 * amdgpu_vm_validate. Fences are on the reservation shared with the
+	 * export, which is currently required to be validated and fenced
+	 * already by amdgpu_amdkfd_gpuvm_restore_process_bos.
+	 *
+	 * Nested locking below for the case that a GEM object is opened in
+	 * kfd_mem_export_dmabuf. Since the lock below is only taken for imports,
+	 * but not for export, this is a different lock class that cannot lead to
+	 * circular lock dependencies.
+	 */
+	if (!vm->is_compute_context || !vm->process_info)
+		return 0;
+	if (!drm_gem_is_imported(obj) ||
+	    !dma_buf_is_dynamic(obj->import_attach->dmabuf))
+		return 0;
+	mutex_lock_nested(&vm->process_info->lock, 1);
+	if (!WARN_ON(!vm->process_info->eviction_fence)) {
+		r = amdgpu_amdkfd_bo_validate_and_fence(abo, AMDGPU_GEM_DOMAIN_GTT,
+							&vm->process_info->eviction_fence->base);
+		if (r) {
+			struct amdgpu_task_info *ti = amdgpu_vm_get_task_info_vm(vm);
+
+			dev_warn(adev->dev, "validate_and_fence failed: %d\n", r);
+			if (ti) {
+				dev_warn(adev->dev, "pid %d\n", ti->task.pid);
+				amdgpu_vm_put_task_info(ti);
+			}
+		}
+	}
+	mutex_unlock(&vm->process_info->lock);
+
+	return r;
 }
 
 static void amdgpu_gem_object_close(struct drm_gem_object *obj,
@@ -195,44 +347,40 @@ static void amdgpu_gem_object_close(struct drm_gem_object *obj,
 	struct amdgpu_fpriv *fpriv = file_priv->driver_priv;
 	struct amdgpu_vm *vm = &fpriv->vm;
 
-	struct amdgpu_bo_list_entry vm_pd;
-	struct list_head list, duplicates;
 	struct dma_fence *fence = NULL;
-	struct ttm_validate_buffer tv;
-	struct ww_acquire_ctx ticket;
 	struct amdgpu_bo_va *bo_va;
+	struct drm_exec exec;
 	long r;
 
-	INIT_LIST_HEAD(&list);
-	INIT_LIST_HEAD(&duplicates);
-
-	tv.bo = &bo->tbo;
-	tv.num_shared = 2;
-	list_add(&tv.head, &list);
+	drm_exec_init(&exec, DRM_EXEC_IGNORE_DUPLICATES, 0);
+	drm_exec_until_all_locked(&exec) {
+		r = drm_exec_prepare_obj(&exec, &bo->tbo.base, 1);
+		drm_exec_retry_on_contention(&exec);
+		if (unlikely(r))
+			goto out_unlock;
+
+		r = amdgpu_vm_lock_pd(vm, &exec, 0);
+		drm_exec_retry_on_contention(&exec);
+		if (unlikely(r))
+			goto out_unlock;
+	}
 
-	amdgpu_vm_get_pd_bo(vm, &list, &vm_pd);
+	if (!amdgpu_vm_is_bo_always_valid(vm, bo))
+		amdgpu_eviction_fence_detach(&fpriv->evf_mgr, bo);
 
-	r = ttm_eu_reserve_buffers(&ticket, &list, false, &duplicates);
-	if (r) {
-		dev_err(adev->dev, "leaking bo va because "
-			"we fail to reserve bo (%ld)\n", r);
-		return;
-	}
 	bo_va = amdgpu_vm_bo_find(vm, bo);
 	if (!bo_va || --bo_va->ref_count)
 		goto out_unlock;
 
-	amdgpu_vm_bo_rmv(adev, bo_va);
+	amdgpu_vm_bo_del(adev, bo_va);
+	amdgpu_vm_bo_update_shared(bo);
 	if (!amdgpu_vm_ready(vm))
 		goto out_unlock;
 
-	fence = dma_resv_excl_fence(bo->tbo.base.resv);
-	if (fence) {
-		amdgpu_bo_fence(bo, fence, true);
-		fence = NULL;
-	}
-
 	r = amdgpu_vm_clear_freed(adev, vm, &fence);
+	if (unlikely(r < 0))
+		dev_err(adev->dev, "failed to clear page "
+			"tables on GEM object close (%ld)\n", r);
 	if (r || !fence)
 		goto out_unlock;
 
@@ -240,10 +388,9 @@ static void amdgpu_gem_object_close(struct drm_gem_object *obj,
 	dma_fence_put(fence);
 
 out_unlock:
-	if (unlikely(r < 0))
-		dev_err(adev->dev, "failed to clear page "
-			"tables on GEM object close (%ld)\n", r);
-	ttm_eu_backoff_reservation(&ticket, &list);
+	if (r)
+		dev_err(adev->dev, "leaking bo va (%ld)\n", r);
+	drm_exec_fini(&exec);
 }
 
 static int amdgpu_gem_object_mmap(struct drm_gem_object *obj, struct vm_area_struct *vma)
@@ -261,13 +408,13 @@ static int amdgpu_gem_object_mmap(struct drm_gem_object *obj, struct vm_area_str
 	 * becoming writable and makes is_cow_mapping(vm_flags) false.
 	 */
 	if (is_cow_mapping(vma->vm_flags) &&
-	    !(vma->vm_flags & (VM_READ | VM_WRITE | VM_EXEC)))
-		vma->vm_flags &= ~VM_MAYWRITE;
+	    !(vma->vm_flags & VM_ACCESS_FLAGS))
+		vm_flags_clear(vma, VM_MAYWRITE);
 
 	return drm_gem_ttm_mmap(obj, vma);
 }
 
-static const struct drm_gem_object_funcs amdgpu_gem_object_funcs = {
+const struct drm_gem_object_funcs amdgpu_gem_object_funcs = {
 	.free = amdgpu_gem_object_free,
 	.open = amdgpu_gem_object_open,
 	.close = amdgpu_gem_object_close,
@@ -296,14 +443,7 @@ int amdgpu_gem_create_ioctl(struct drm_device *dev, void *data,
 	int r;
 
 	/* reject invalid gem flags */
-	if (flags & ~(AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED |
-		      AMDGPU_GEM_CREATE_NO_CPU_ACCESS |
-		      AMDGPU_GEM_CREATE_CPU_GTT_USWC |
-		      AMDGPU_GEM_CREATE_VRAM_CLEARED |
-		      AMDGPU_GEM_CREATE_VM_ALWAYS_VALID |
-		      AMDGPU_GEM_CREATE_EXPLICIT_SYNC |
-		      AMDGPU_GEM_CREATE_ENCRYPTED))
-
+	if (flags & ~AMDGPU_GEM_CREATE_SETTABLE_MASK)
 		return -EINVAL;
 
 	/* reject invalid gem domains */
@@ -315,6 +455,12 @@ int amdgpu_gem_create_ioctl(struct drm_device *dev, void *data,
 		return -EINVAL;
 	}
 
+	/* always clear VRAM */
+	flags |= AMDGPU_GEM_CREATE_VRAM_CLEARED;
+
+	if (args->in.domains & AMDGPU_GEM_DOMAIN_MMIO_REMAP)
+		return -EINVAL;
+
 	/* create a gem object to contain this object in */
 	if (args->in.domains & (AMDGPU_GEM_DOMAIN_GDS |
 	    AMDGPU_GEM_DOMAIN_GWS | AMDGPU_GEM_DOMAIN_OA)) {
@@ -340,7 +486,7 @@ int amdgpu_gem_create_ioctl(struct drm_device *dev, void *data,
 retry:
 	r = amdgpu_gem_object_create(adev, size, args->in.alignment,
 				     initial_domain,
-				     flags, ttm_bo_type_device, resv, &gobj);
+				     flags, ttm_bo_type_device, resv, &gobj, fpriv->xcp_id + 1);
 	if (r && r != -ERESTARTSYS) {
 		if (flags & AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED) {
 			flags &= ~AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED;
@@ -383,7 +529,9 @@ int amdgpu_gem_userptr_ioctl(struct drm_device *dev, void *data,
 	struct ttm_operation_ctx ctx = { true, false };
 	struct amdgpu_device *adev = drm_to_adev(dev);
 	struct drm_amdgpu_gem_userptr *args = data;
+	struct amdgpu_fpriv *fpriv = filp->driver_priv;
 	struct drm_gem_object *gobj;
+	struct amdgpu_hmm_range *range;
 	struct amdgpu_bo *bo;
 	uint32_t handle;
 	int r;
@@ -408,7 +556,7 @@ int amdgpu_gem_userptr_ioctl(struct drm_device *dev, void *data,
 
 	/* create a gem object to contain this object in */
 	r = amdgpu_gem_object_create(adev, args->size, 0, AMDGPU_GEM_DOMAIN_CPU,
-				     0, ttm_bo_type_device, NULL, &gobj);
+				     0, ttm_bo_type_device, NULL, &gobj, fpriv->xcp_id + 1);
 	if (r)
 		return r;
 
@@ -419,21 +567,25 @@ int amdgpu_gem_userptr_ioctl(struct drm_device *dev, void *data,
 	if (r)
 		goto release_object;
 
-	if (args->flags & AMDGPU_GEM_USERPTR_REGISTER) {
-		r = amdgpu_mn_register(bo, args->addr);
-		if (r)
-			goto release_object;
-	}
+	r = amdgpu_hmm_register(bo, args->addr);
+	if (r)
+		goto release_object;
 
 	if (args->flags & AMDGPU_GEM_USERPTR_VALIDATE) {
-		r = amdgpu_ttm_tt_get_user_pages(bo, bo->tbo.ttm->pages);
-		if (r)
+		range = amdgpu_hmm_range_alloc(NULL);
+		if (unlikely(!range))
+			return -ENOMEM;
+		r = amdgpu_ttm_tt_get_user_pages(bo, range);
+		if (r) {
+			amdgpu_hmm_range_free(range);
 			goto release_object;
-
+		}
 		r = amdgpu_bo_reserve(bo, true);
 		if (r)
 			goto user_pages_done;
 
+		amdgpu_ttm_tt_set_user_pages(bo->tbo.ttm, range);
+
 		amdgpu_bo_placement_from_domain(bo, AMDGPU_GEM_DOMAIN_GTT);
 		r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
 		amdgpu_bo_unreserve(bo);
@@ -449,8 +601,7 @@ int amdgpu_gem_userptr_ioctl(struct drm_device *dev, void *data,
 
 user_pages_done:
 	if (args->flags & AMDGPU_GEM_USERPTR_VALIDATE)
-		amdgpu_ttm_tt_get_user_pages_done(bo->tbo.ttm);
-
+		amdgpu_hmm_range_free(range);
 release_object:
 	drm_gem_object_put(gobj);
 
@@ -465,9 +616,9 @@ int amdgpu_mode_dumb_mmap(struct drm_file *filp,
 	struct amdgpu_bo *robj;
 
 	gobj = drm_gem_object_lookup(filp, handle);
-	if (gobj == NULL) {
+	if (!gobj)
 		return -ENOENT;
-	}
+
 	robj = gem_to_amdgpu_bo(gobj);
 	if (amdgpu_ttm_tt_get_usermm(robj->tbo.ttm) ||
 	    (robj->flags & AMDGPU_GEM_CREATE_NO_CPU_ACCESS)) {
@@ -484,6 +635,7 @@ int amdgpu_gem_mmap_ioctl(struct drm_device *dev, void *data,
 {
 	union drm_amdgpu_gem_mmap *args = data;
 	uint32_t handle = args->in.handle;
+
 	memset(args, 0, sizeof(*args));
 	return amdgpu_mode_dumb_mmap(filp, dev, handle, &args->out.addr_ptr);
 }
@@ -510,7 +662,7 @@ unsigned long amdgpu_gem_timeout(uint64_t timeout_ns)
 
 	timeout_jiffies = nsecs_to_jiffies(ktime_to_ns(timeout));
 	/*  clamp timeout to avoid unsigned-> signed overflow */
-	if (timeout_jiffies > MAX_SCHEDULE_TIMEOUT )
+	if (timeout_jiffies > MAX_SCHEDULE_TIMEOUT)
 		return MAX_SCHEDULE_TIMEOUT - 1;
 
 	return timeout_jiffies;
@@ -528,11 +680,12 @@ int amdgpu_gem_wait_idle_ioctl(struct drm_device *dev, void *data,
 	long ret;
 
 	gobj = drm_gem_object_lookup(filp, handle);
-	if (gobj == NULL) {
+	if (!gobj)
 		return -ENOENT;
-	}
+
 	robj = gem_to_amdgpu_bo(gobj);
-	ret = dma_resv_wait_timeout(robj->tbo.base.resv, true, true, timeout);
+	ret = dma_resv_wait_timeout(robj->tbo.base.resv, DMA_RESV_USAGE_READ,
+				    true, timeout);
 
 	/* ret == 0 means not signaled,
 	 * ret > 0 means signaled
@@ -556,7 +709,7 @@ int amdgpu_gem_metadata_ioctl(struct drm_device *dev, void *data,
 	struct amdgpu_bo *robj;
 	int r = -1;
 
-	DRM_DEBUG("%d \n", args->handle);
+	DRM_DEBUG("%d\n", args->handle);
 	gobj = drm_gem_object_lookup(filp, args->handle);
 	if (gobj == NULL)
 		return -ENOENT;
@@ -601,24 +754,29 @@ out:
  *
  * Update the bo_va directly after setting its address. Errors are not
  * vital here, so they are not reported back to userspace.
+ *
+ * Returns resulting fence if freed BO(s) got cleared from the PT.
+ * otherwise stub fence in case of error.
  */
-static void amdgpu_gem_va_update_vm(struct amdgpu_device *adev,
-				    struct amdgpu_vm *vm,
-				    struct amdgpu_bo_va *bo_va,
-				    uint32_t operation)
+static struct dma_fence *
+amdgpu_gem_va_update_vm(struct amdgpu_device *adev,
+			struct amdgpu_vm *vm,
+			struct amdgpu_bo_va *bo_va,
+			uint32_t operation)
 {
+	struct dma_fence *fence = dma_fence_get_stub();
 	int r;
 
 	if (!amdgpu_vm_ready(vm))
-		return;
+		return fence;
 
-	r = amdgpu_vm_clear_freed(adev, vm, NULL);
+	r = amdgpu_vm_clear_freed(adev, vm, &fence);
 	if (r)
 		goto error;
 
 	if (operation == AMDGPU_VA_OP_MAP ||
 	    operation == AMDGPU_VA_OP_REPLACE) {
-		r = amdgpu_vm_bo_update(adev, bo_va, false, NULL);
+		r = amdgpu_vm_bo_update(adev, bo_va, false);
 		if (r)
 			goto error;
 	}
@@ -628,34 +786,8 @@ static void amdgpu_gem_va_update_vm(struct amdgpu_device *adev,
 error:
 	if (r && r != -ERESTARTSYS)
 		DRM_ERROR("Couldn't update BO_VA (%d)\n", r);
-}
 
-/**
- * amdgpu_gem_va_map_flags - map GEM UAPI flags into hardware flags
- *
- * @adev: amdgpu_device pointer
- * @flags: GEM UAPI flags
- *
- * Returns the GEM UAPI flags mapped into hardware for the ASIC.
- */
-uint64_t amdgpu_gem_va_map_flags(struct amdgpu_device *adev, uint32_t flags)
-{
-	uint64_t pte_flag = 0;
-
-	if (flags & AMDGPU_VM_PAGE_EXECUTABLE)
-		pte_flag |= AMDGPU_PTE_EXECUTABLE;
-	if (flags & AMDGPU_VM_PAGE_READABLE)
-		pte_flag |= AMDGPU_PTE_READABLE;
-	if (flags & AMDGPU_VM_PAGE_WRITEABLE)
-		pte_flag |= AMDGPU_PTE_WRITEABLE;
-	if (flags & AMDGPU_VM_PAGE_PRT)
-		pte_flag |= AMDGPU_PTE_PRT;
-
-	if (adev->gmc.gmc_funcs->map_mtype)
-		pte_flag |= amdgpu_gmc_map_mtype(adev,
-						 flags & AMDGPU_VM_MTYPE_MASK);
-
-	return pte_flag;
+	return fence;
 }
 
 int amdgpu_gem_va_ioctl(struct drm_device *dev, void *data,
@@ -663,7 +795,8 @@ int amdgpu_gem_va_ioctl(struct drm_device *dev, void *data,
 {
 	const uint32_t valid_flags = AMDGPU_VM_DELAY_UPDATE |
 		AMDGPU_VM_PAGE_READABLE | AMDGPU_VM_PAGE_WRITEABLE |
-		AMDGPU_VM_PAGE_EXECUTABLE | AMDGPU_VM_MTYPE_MASK;
+		AMDGPU_VM_PAGE_EXECUTABLE | AMDGPU_VM_MTYPE_MASK |
+		AMDGPU_VM_PAGE_NOALLOC;
 	const uint32_t prt_flags = AMDGPU_VM_DELAY_UPDATE |
 		AMDGPU_VM_PAGE_PRT;
 
@@ -673,25 +806,24 @@ int amdgpu_gem_va_ioctl(struct drm_device *dev, void *data,
 	struct amdgpu_fpriv *fpriv = filp->driver_priv;
 	struct amdgpu_bo *abo;
 	struct amdgpu_bo_va *bo_va;
-	struct amdgpu_bo_list_entry vm_pd;
-	struct ttm_validate_buffer tv;
-	struct ww_acquire_ctx ticket;
-	struct list_head list, duplicates;
-	uint64_t va_flags;
+	struct drm_syncobj *timeline_syncobj = NULL;
+	struct dma_fence_chain *timeline_chain = NULL;
+	struct dma_fence *fence;
+	struct drm_exec exec;
 	uint64_t vm_size;
 	int r = 0;
 
-	if (args->va_address < AMDGPU_VA_RESERVED_SIZE) {
+	if (args->va_address < AMDGPU_VA_RESERVED_BOTTOM) {
 		dev_dbg(dev->dev,
-			"va_address 0x%LX is in reserved area 0x%LX\n",
-			args->va_address, AMDGPU_VA_RESERVED_SIZE);
+			"va_address 0x%llx is in reserved area 0x%llx\n",
+			args->va_address, AMDGPU_VA_RESERVED_BOTTOM);
 		return -EINVAL;
 	}
 
 	if (args->va_address >= AMDGPU_GMC_HOLE_START &&
 	    args->va_address < AMDGPU_GMC_HOLE_END) {
 		dev_dbg(dev->dev,
-			"va_address 0x%LX is in VA hole 0x%LX-0x%LX\n",
+			"va_address 0x%llx is in VA hole 0x%llx-0x%llx\n",
 			args->va_address, AMDGPU_GMC_HOLE_START,
 			AMDGPU_GMC_HOLE_END);
 		return -EINVAL;
@@ -700,7 +832,7 @@ int amdgpu_gem_va_ioctl(struct drm_device *dev, void *data,
 	args->va_address &= AMDGPU_GMC_HOLE_MASK;
 
 	vm_size = adev->vm_manager.max_pfn * AMDGPU_GPU_PAGE_SIZE;
-	vm_size -= AMDGPU_VA_RESERVED_SIZE;
+	vm_size -= AMDGPU_VA_RESERVED_TOP;
 	if (args->va_address + args->map_size > vm_size) {
 		dev_dbg(dev->dev,
 			"va_address 0x%llx is in top reserved area 0x%llx\n",
@@ -726,36 +858,44 @@ int amdgpu_gem_va_ioctl(struct drm_device *dev, void *data,
 		return -EINVAL;
 	}
 
-	INIT_LIST_HEAD(&list);
-	INIT_LIST_HEAD(&duplicates);
 	if ((args->operation != AMDGPU_VA_OP_CLEAR) &&
 	    !(args->flags & AMDGPU_VM_PAGE_PRT)) {
 		gobj = drm_gem_object_lookup(filp, args->handle);
 		if (gobj == NULL)
 			return -ENOENT;
 		abo = gem_to_amdgpu_bo(gobj);
-		tv.bo = &abo->tbo;
-		if (abo->flags & AMDGPU_GEM_CREATE_VM_ALWAYS_VALID)
-			tv.num_shared = 1;
-		else
-			tv.num_shared = 0;
-		list_add(&tv.head, &list);
 	} else {
 		gobj = NULL;
 		abo = NULL;
 	}
 
-	amdgpu_vm_get_pd_bo(&fpriv->vm, &list, &vm_pd);
-
-	r = ttm_eu_reserve_buffers(&ticket, &list, true, &duplicates);
+	r = amdgpu_gem_add_input_fence(filp,
+				       args->input_fence_syncobj_handles,
+				       args->num_syncobj_handles);
 	if (r)
-		goto error_unref;
+		goto error_put_gobj;
+
+	drm_exec_init(&exec, DRM_EXEC_INTERRUPTIBLE_WAIT |
+		      DRM_EXEC_IGNORE_DUPLICATES, 0);
+	drm_exec_until_all_locked(&exec) {
+		if (gobj) {
+			r = drm_exec_lock_obj(&exec, gobj);
+			drm_exec_retry_on_contention(&exec);
+			if (unlikely(r))
+				goto error;
+		}
+
+		r = amdgpu_vm_lock_pd(&fpriv->vm, &exec, 2);
+		drm_exec_retry_on_contention(&exec);
+		if (unlikely(r))
+			goto error;
+	}
 
 	if (abo) {
 		bo_va = amdgpu_vm_bo_find(&fpriv->vm, abo);
 		if (!bo_va) {
 			r = -ENOENT;
-			goto error_backoff;
+			goto error;
 		}
 	} else if (args->operation != AMDGPU_VA_OP_CLEAR) {
 		bo_va = fpriv->prt_va;
@@ -763,12 +903,19 @@ int amdgpu_gem_va_ioctl(struct drm_device *dev, void *data,
 		bo_va = NULL;
 	}
 
+	r = amdgpu_gem_update_timeline_node(filp,
+					    args->vm_timeline_syncobj_out,
+					    args->vm_timeline_point,
+					    &timeline_syncobj,
+					    &timeline_chain);
+	if (r)
+		goto error;
+
 	switch (args->operation) {
 	case AMDGPU_VA_OP_MAP:
-		va_flags = amdgpu_gem_va_map_flags(adev, args->flags);
 		r = amdgpu_vm_bo_map(adev, bo_va, args->va_address,
 				     args->offset_in_bo, args->map_size,
-				     va_flags);
+				     args->flags);
 		break;
 	case AMDGPU_VA_OP_UNMAP:
 		r = amdgpu_vm_bo_unmap(adev, bo_va, args->va_address);
@@ -780,22 +927,31 @@ int amdgpu_gem_va_ioctl(struct drm_device *dev, void *data,
 						args->map_size);
 		break;
 	case AMDGPU_VA_OP_REPLACE:
-		va_flags = amdgpu_gem_va_map_flags(adev, args->flags);
 		r = amdgpu_vm_bo_replace_map(adev, bo_va, args->va_address,
 					     args->offset_in_bo, args->map_size,
-					     va_flags);
+					     args->flags);
 		break;
 	default:
 		break;
 	}
-	if (!r && !(args->flags & AMDGPU_VM_DELAY_UPDATE) && !amdgpu_vm_debug)
-		amdgpu_gem_va_update_vm(adev, &fpriv->vm, bo_va,
-					args->operation);
+	if (!r && !(args->flags & AMDGPU_VM_DELAY_UPDATE) && !adev->debug_vm) {
+		fence = amdgpu_gem_va_update_vm(adev, &fpriv->vm, bo_va,
+						args->operation);
+
+		if (timeline_syncobj)
+			amdgpu_gem_update_bo_mapping(filp, bo_va,
+					     args->operation,
+					     args->vm_timeline_point,
+					     fence, timeline_syncobj,
+					     timeline_chain);
+		else
+			dma_fence_put(fence);
 
-error_backoff:
-	ttm_eu_backoff_reservation(&ticket, &list);
+	}
 
-error_unref:
+error:
+	drm_exec_fini(&exec);
+error_put_gobj:
 	drm_gem_object_put(gobj);
 	return r;
 }
@@ -803,22 +959,38 @@ error_unref:
 int amdgpu_gem_op_ioctl(struct drm_device *dev, void *data,
 			struct drm_file *filp)
 {
-	struct amdgpu_device *adev = drm_to_adev(dev);
 	struct drm_amdgpu_gem_op *args = data;
 	struct drm_gem_object *gobj;
 	struct amdgpu_vm_bo_base *base;
 	struct amdgpu_bo *robj;
+	struct drm_exec exec;
+	struct amdgpu_fpriv *fpriv = filp->driver_priv;
 	int r;
 
+	if (args->padding)
+		return -EINVAL;
+
 	gobj = drm_gem_object_lookup(filp, args->handle);
-	if (gobj == NULL) {
+	if (!gobj)
 		return -ENOENT;
-	}
+
 	robj = gem_to_amdgpu_bo(gobj);
 
-	r = amdgpu_bo_reserve(robj, false);
-	if (unlikely(r))
-		goto out;
+	drm_exec_init(&exec, DRM_EXEC_INTERRUPTIBLE_WAIT |
+			  DRM_EXEC_IGNORE_DUPLICATES, 0);
+	drm_exec_until_all_locked(&exec) {
+		r = drm_exec_lock_obj(&exec, gobj);
+		drm_exec_retry_on_contention(&exec);
+		if (r)
+			goto out_exec;
+
+		if (args->op == AMDGPU_GEM_OP_GET_MAPPING_INFO) {
+			r = amdgpu_vm_lock_pd(&fpriv->vm, &exec, 0);
+			drm_exec_retry_on_contention(&exec);
+			if (r)
+				goto out_exec;
+		}
+	}
 
 	switch (args->op) {
 	case AMDGPU_GEM_OP_GET_GEM_CREATE_INFO: {
@@ -829,29 +1001,26 @@ int amdgpu_gem_op_ioctl(struct drm_device *dev, void *data,
 		info.alignment = robj->tbo.page_alignment << PAGE_SHIFT;
 		info.domains = robj->preferred_domains;
 		info.domain_flags = robj->flags;
-		amdgpu_bo_unreserve(robj);
+		drm_exec_fini(&exec);
 		if (copy_to_user(out, &info, sizeof(info)))
 			r = -EFAULT;
 		break;
 	}
 	case AMDGPU_GEM_OP_SET_PLACEMENT:
-		if (robj->tbo.base.import_attach &&
+		if (drm_gem_is_imported(&robj->tbo.base) &&
 		    args->value & AMDGPU_GEM_DOMAIN_VRAM) {
 			r = -EINVAL;
-			amdgpu_bo_unreserve(robj);
-			break;
+			goto out_exec;
 		}
 		if (amdgpu_ttm_tt_get_usermm(robj->tbo.ttm)) {
 			r = -EPERM;
-			amdgpu_bo_unreserve(robj);
-			break;
+			goto out_exec;
 		}
 		for (base = robj->vm_bo; base; base = base->next)
 			if (amdgpu_xgmi_same_hive(amdgpu_ttm_adev(robj->tbo.bdev),
 				amdgpu_ttm_adev(base->vm->root.bo->tbo.bdev))) {
 				r = -EINVAL;
-				amdgpu_bo_unreserve(robj);
-				goto out;
+				goto out_exec;
 			}
 
 
@@ -863,18 +1032,173 @@ int amdgpu_gem_op_ioctl(struct drm_device *dev, void *data,
 			robj->allowed_domains |= AMDGPU_GEM_DOMAIN_GTT;
 
 		if (robj->flags & AMDGPU_GEM_CREATE_VM_ALWAYS_VALID)
-			amdgpu_vm_bo_invalidate(adev, robj, true);
+			amdgpu_vm_bo_invalidate(robj, true);
+		drm_exec_fini(&exec);
+		break;
+	case AMDGPU_GEM_OP_GET_MAPPING_INFO: {
+		struct amdgpu_bo_va *bo_va = amdgpu_vm_bo_find(&fpriv->vm, robj);
+		struct drm_amdgpu_gem_vm_entry *vm_entries;
+		struct amdgpu_bo_va_mapping *mapping;
+		int num_mappings = 0;
+		/*
+		 * num_entries is set as an input to the size of the user-allocated array of
+		 * drm_amdgpu_gem_vm_entry stored at args->value.
+		 * num_entries is sent back as output as the number of mappings the bo has.
+		 * If that number is larger than the size of the array, the ioctl must
+		 * be retried.
+		 */
+		vm_entries = kvcalloc(args->num_entries, sizeof(*vm_entries), GFP_KERNEL);
+		if (!vm_entries)
+			return -ENOMEM;
+
+		amdgpu_vm_bo_va_for_each_valid_mapping(bo_va, mapping) {
+			if (num_mappings < args->num_entries) {
+				vm_entries[num_mappings].addr = mapping->start * AMDGPU_GPU_PAGE_SIZE;
+				vm_entries[num_mappings].size = (mapping->last - mapping->start + 1) * AMDGPU_GPU_PAGE_SIZE;
+				vm_entries[num_mappings].offset = mapping->offset;
+				vm_entries[num_mappings].flags = mapping->flags;
+			}
+			num_mappings += 1;
+		}
+
+		amdgpu_vm_bo_va_for_each_invalid_mapping(bo_va, mapping) {
+			if (num_mappings < args->num_entries) {
+				vm_entries[num_mappings].addr = mapping->start * AMDGPU_GPU_PAGE_SIZE;
+				vm_entries[num_mappings].size = (mapping->last - mapping->start + 1) * AMDGPU_GPU_PAGE_SIZE;
+				vm_entries[num_mappings].offset = mapping->offset;
+				vm_entries[num_mappings].flags = mapping->flags;
+			}
+			num_mappings += 1;
+		}
+
+		drm_exec_fini(&exec);
+
+		if (num_mappings > 0 && num_mappings <= args->num_entries)
+			if (copy_to_user(u64_to_user_ptr(args->value), vm_entries, num_mappings * sizeof(*vm_entries)))
+				r = -EFAULT;
 
-		amdgpu_bo_unreserve(robj);
+		args->num_entries = num_mappings;
+
+		kvfree(vm_entries);
 		break;
+	}
 	default:
-		amdgpu_bo_unreserve(robj);
+		drm_exec_fini(&exec);
 		r = -EINVAL;
 	}
 
-out:
 	drm_gem_object_put(gobj);
 	return r;
+out_exec:
+	drm_exec_fini(&exec);
+	drm_gem_object_put(gobj);
+	return r;
+}
+
+/**
+ * amdgpu_gem_list_handles_ioctl - get information about a process' buffer objects
+ *
+ * @dev: drm device pointer
+ * @data: drm_amdgpu_gem_list_handles
+ * @filp: drm file pointer
+ *
+ * num_entries is set as an input to the size of the entries array.
+ * num_entries is sent back as output as the number of bos in the process.
+ * If that number is larger than the size of the array, the ioctl must
+ * be retried.
+ *
+ * Returns:
+ * 0 for success, -errno for errors.
+ */
+int amdgpu_gem_list_handles_ioctl(struct drm_device *dev, void *data,
+				  struct drm_file *filp)
+{
+	struct drm_amdgpu_gem_list_handles *args = data;
+	struct drm_amdgpu_gem_list_handles_entry *bo_entries;
+	struct drm_gem_object *gobj;
+	int id, ret = 0;
+	int bo_index = 0;
+	int num_bos = 0;
+
+	spin_lock(&filp->table_lock);
+	idr_for_each_entry(&filp->object_idr, gobj, id)
+		num_bos += 1;
+	spin_unlock(&filp->table_lock);
+
+	if (args->num_entries < num_bos) {
+		args->num_entries = num_bos;
+		return 0;
+	}
+
+	if (num_bos == 0) {
+		args->num_entries = 0;
+		return 0;
+	}
+
+	bo_entries = kvcalloc(num_bos, sizeof(*bo_entries), GFP_KERNEL);
+	if (!bo_entries)
+		return -ENOMEM;
+
+	spin_lock(&filp->table_lock);
+	idr_for_each_entry(&filp->object_idr, gobj, id) {
+		struct amdgpu_bo *bo = gem_to_amdgpu_bo(gobj);
+		struct drm_amdgpu_gem_list_handles_entry *bo_entry;
+
+		if (bo_index >= num_bos) {
+			ret = -EAGAIN;
+			break;
+		}
+
+		bo_entry = &bo_entries[bo_index];
+
+		bo_entry->size = amdgpu_bo_size(bo);
+		bo_entry->alloc_flags = bo->flags & AMDGPU_GEM_CREATE_SETTABLE_MASK;
+		bo_entry->preferred_domains = bo->preferred_domains;
+		bo_entry->gem_handle = id;
+		bo_entry->alignment = bo->tbo.page_alignment;
+
+		if (bo->tbo.base.import_attach)
+			bo_entry->flags |= AMDGPU_GEM_LIST_HANDLES_FLAG_IS_IMPORT;
+
+		bo_index += 1;
+	}
+	spin_unlock(&filp->table_lock);
+
+	args->num_entries = bo_index;
+
+	if (!ret)
+		if (copy_to_user(u64_to_user_ptr(args->entries), bo_entries, num_bos * sizeof(*bo_entries)))
+			ret = -EFAULT;
+
+	kvfree(bo_entries);
+
+	return ret;
+}
+
+static int amdgpu_gem_align_pitch(struct amdgpu_device *adev,
+				  int width,
+				  int cpp,
+				  bool tiled)
+{
+	int aligned = width;
+	int pitch_mask = 0;
+
+	switch (cpp) {
+	case 1:
+		pitch_mask = 255;
+		break;
+	case 2:
+		pitch_mask = 127;
+		break;
+	case 3:
+	case 4:
+		pitch_mask = 63;
+		break;
+	}
+
+	aligned += pitch_mask;
+	aligned &= ~pitch_mask;
+	return aligned * cpp;
 }
 
 int amdgpu_mode_dumb_create(struct drm_file *file_priv,
@@ -882,10 +1206,12 @@ int amdgpu_mode_dumb_create(struct drm_file *file_priv,
 			    struct drm_mode_create_dumb *args)
 {
 	struct amdgpu_device *adev = drm_to_adev(dev);
+	struct amdgpu_fpriv *fpriv = file_priv->driver_priv;
 	struct drm_gem_object *gobj;
 	uint32_t handle;
 	u64 flags = AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED |
-		    AMDGPU_GEM_CREATE_CPU_GTT_USWC;
+		    AMDGPU_GEM_CREATE_CPU_GTT_USWC |
+		    AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS;
 	u32 domain;
 	int r;
 
@@ -897,23 +1223,23 @@ int amdgpu_mode_dumb_create(struct drm_file *file_priv,
 	if (adev->mman.buffer_funcs_enabled)
 		flags |= AMDGPU_GEM_CREATE_VRAM_CLEARED;
 
-	args->pitch = amdgpu_align_pitch(adev, args->width,
-					 DIV_ROUND_UP(args->bpp, 8), 0);
+	args->pitch = amdgpu_gem_align_pitch(adev, args->width,
+					     DIV_ROUND_UP(args->bpp, 8), 0);
 	args->size = (u64)args->pitch * args->height;
 	args->size = ALIGN(args->size, PAGE_SIZE);
 	domain = amdgpu_bo_get_preferred_domain(adev,
 				amdgpu_display_supported_domains(adev, flags));
 	r = amdgpu_gem_object_create(adev, args->size, 0, domain, flags,
-				     ttm_bo_type_device, NULL, &gobj);
+				     ttm_bo_type_device, NULL, &gobj, fpriv->xcp_id + 1);
 	if (r)
 		return -ENOMEM;
 
 	r = drm_gem_handle_create(file_priv, gobj, &handle);
 	/* drop reference from allocate - handle holds it now */
 	drm_gem_object_put(gobj);
-	if (r) {
+	if (r)
 		return r;
-	}
+
 	args->handle = handle;
 	return 0;
 }
@@ -921,7 +1247,7 @@ int amdgpu_mode_dumb_create(struct drm_file *file_priv,
 #if defined(CONFIG_DEBUG_FS)
 static int amdgpu_debugfs_gem_info_show(struct seq_file *m, void *unused)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)m->private;
+	struct amdgpu_device *adev = m->private;
 	struct drm_device *dev = adev_to_drm(adev);
 	struct drm_file *file;
 	int r;
@@ -933,6 +1259,7 @@ static int amdgpu_debugfs_gem_info_show(struct seq_file *m, void *unused)
 	list_for_each_entry(file, &dev->filelist, lhead) {
 		struct task_struct *task;
 		struct drm_gem_object *gobj;
+		struct pid *pid;
 		int id;
 
 		/*
@@ -942,8 +1269,9 @@ static int amdgpu_debugfs_gem_info_show(struct seq_file *m, void *unused)
 		 * Therefore, we need to protect this ->comm access using RCU.
 		 */
 		rcu_read_lock();
-		task = pid_task(file->pid, PIDTYPE_PID);
-		seq_printf(m, "pid %8d command %s:\n", pid_nr(file->pid),
+		pid = rcu_dereference(file->pid);
+		task = pid_task(pid, PIDTYPE_TGID);
+		seq_printf(m, "pid %8d command %s:\n", pid_nr(pid),
 			   task ? task->comm : "<unknown>");
 		rcu_read_unlock();
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.h
index 637bf51dbf06..b558336bc4c6 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.h
@@ -33,6 +33,8 @@
 #define AMDGPU_GEM_DOMAIN_MAX		0x3
 #define gem_to_amdgpu_bo(gobj) container_of((gobj), struct amdgpu_bo, tbo.base)
 
+extern const struct drm_gem_object_funcs amdgpu_gem_object_funcs;
+
 unsigned long amdgpu_gem_timeout(uint64_t timeout_ns);
 
 /*
@@ -43,8 +45,7 @@ int amdgpu_gem_object_create(struct amdgpu_device *adev, unsigned long size,
 			     int alignment, u32 initial_domain,
 			     u64 flags, enum ttm_bo_type type,
 			     struct dma_resv *resv,
-			     struct drm_gem_object **obj);
-
+			     struct drm_gem_object **obj, int8_t xcp_id_plus1);
 int amdgpu_mode_dumb_create(struct drm_file *file_priv,
 			    struct drm_device *dev,
 			    struct drm_mode_create_dumb *args);
@@ -62,13 +63,28 @@ int amdgpu_gem_mmap_ioctl(struct drm_device *dev, void *data,
 			  struct drm_file *filp);
 int amdgpu_gem_wait_idle_ioctl(struct drm_device *dev, void *data,
 			      struct drm_file *filp);
-uint64_t amdgpu_gem_va_map_flags(struct amdgpu_device *adev, uint32_t flags);
 int amdgpu_gem_va_ioctl(struct drm_device *dev, void *data,
 			  struct drm_file *filp);
 int amdgpu_gem_op_ioctl(struct drm_device *dev, void *data,
 			struct drm_file *filp);
+int amdgpu_gem_list_handles_ioctl(struct drm_device *dev, void *data,
+				  struct drm_file *filp);
 
 int amdgpu_gem_metadata_ioctl(struct drm_device *dev, void *data,
 				struct drm_file *filp);
 
+#define AMDGPU_GEM_CREATE_SETTABLE_MASK	(AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED | \
+	AMDGPU_GEM_CREATE_NO_CPU_ACCESS | \
+	AMDGPU_GEM_CREATE_CPU_GTT_USWC | \
+	AMDGPU_GEM_CREATE_VRAM_CLEARED | \
+	AMDGPU_GEM_CREATE_VM_ALWAYS_VALID | \
+	AMDGPU_GEM_CREATE_EXPLICIT_SYNC | \
+	AMDGPU_GEM_CREATE_VRAM_WIPE_ON_RELEASE | \
+	AMDGPU_GEM_CREATE_ENCRYPTED | \
+	AMDGPU_GEM_CREATE_GFX12_DCC | \
+	AMDGPU_GEM_CREATE_DISCARDABLE | \
+	AMDGPU_GEM_CREATE_COHERENT | \
+	AMDGPU_GEM_CREATE_UNCACHED | \
+	AMDGPU_GEM_CREATE_EXT_COHERENT)
+
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
index 1916ec84dd71..8b118c53f351 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
@@ -23,10 +23,18 @@
  *
  */
 
+#include <linux/firmware.h>
+#include <linux/pm_runtime.h>
+
 #include "amdgpu.h"
 #include "amdgpu_gfx.h"
 #include "amdgpu_rlc.h"
 #include "amdgpu_ras.h"
+#include "amdgpu_reset.h"
+#include "amdgpu_xcp.h"
+#include "amdgpu_xgmi.h"
+#include "amdgpu_mes.h"
+#include "nvd.h"
 
 /* delay 0.1 second to enable gfx off feature */
 #define GFX_OFF_DELAY_ENABLE         msecs_to_jiffies(100)
@@ -62,35 +70,26 @@ void amdgpu_queue_mask_bit_to_mec_queue(struct amdgpu_device *adev, int bit,
 }
 
 bool amdgpu_gfx_is_mec_queue_enabled(struct amdgpu_device *adev,
-				     int mec, int pipe, int queue)
+				     int xcc_id, int mec, int pipe, int queue)
 {
 	return test_bit(amdgpu_gfx_mec_queue_to_bit(adev, mec, pipe, queue),
-			adev->gfx.mec.queue_bitmap);
+			adev->gfx.mec_bitmap[xcc_id].queue_bitmap);
 }
 
-int amdgpu_gfx_me_queue_to_bit(struct amdgpu_device *adev,
-			       int me, int pipe, int queue)
+static int amdgpu_gfx_me_queue_to_bit(struct amdgpu_device *adev,
+				      int me, int pipe, int queue)
 {
+	int num_queue_per_pipe = 1; /* we only enable 1 KGQ per pipe */
 	int bit = 0;
 
 	bit += me * adev->gfx.me.num_pipe_per_me
-		* adev->gfx.me.num_queue_per_pipe;
-	bit += pipe * adev->gfx.me.num_queue_per_pipe;
+		* num_queue_per_pipe;
+	bit += pipe * num_queue_per_pipe;
 	bit += queue;
 
 	return bit;
 }
 
-void amdgpu_gfx_bit_to_me_queue(struct amdgpu_device *adev, int bit,
-				int *me, int *pipe, int *queue)
-{
-	*queue = bit % adev->gfx.me.num_queue_per_pipe;
-	*pipe = (bit / adev->gfx.me.num_queue_per_pipe)
-		% adev->gfx.me.num_pipe_per_me;
-	*me = (bit / adev->gfx.me.num_queue_per_pipe)
-		/ adev->gfx.me.num_pipe_per_me;
-}
-
 bool amdgpu_gfx_is_me_queue_enabled(struct amdgpu_device *adev,
 				    int me, int pipe, int queue)
 {
@@ -99,42 +98,6 @@ bool amdgpu_gfx_is_me_queue_enabled(struct amdgpu_device *adev,
 }
 
 /**
- * amdgpu_gfx_scratch_get - Allocate a scratch register
- *
- * @adev: amdgpu_device pointer
- * @reg: scratch register mmio offset
- *
- * Allocate a CP scratch register for use by the driver (all asics).
- * Returns 0 on success or -EINVAL on failure.
- */
-int amdgpu_gfx_scratch_get(struct amdgpu_device *adev, uint32_t *reg)
-{
-	int i;
-
-	i = ffs(adev->gfx.scratch.free_mask);
-	if (i != 0 && i <= adev->gfx.scratch.num_reg) {
-		i--;
-		adev->gfx.scratch.free_mask &= ~(1u << i);
-		*reg = adev->gfx.scratch.reg_base + i;
-		return 0;
-	}
-	return -EINVAL;
-}
-
-/**
- * amdgpu_gfx_scratch_free - Free a scratch register
- *
- * @adev: amdgpu_device pointer
- * @reg: scratch register mmio offset
- *
- * Free a CP scratch register allocated for use by the driver (all asics)
- */
-void amdgpu_gfx_scratch_free(struct amdgpu_device *adev, uint32_t reg)
-{
-	adev->gfx.scratch.free_mask |= 1u << (reg - adev->gfx.scratch.reg_base);
-}
-
-/**
  * amdgpu_gfx_parse_disable_cu - Parse the disable_cu module parameter
  *
  * @mask: array in which the per-shader array disable masks will be stored
@@ -144,9 +107,9 @@ void amdgpu_gfx_scratch_free(struct amdgpu_device *adev, uint32_t reg)
  * The bitmask of CUs to be disabled in the shader array determined by se and
  * sh is stored in mask[se * max_sh + sh].
  */
-void amdgpu_gfx_parse_disable_cu(unsigned *mask, unsigned max_se, unsigned max_sh)
+void amdgpu_gfx_parse_disable_cu(unsigned int *mask, unsigned int max_se, unsigned int max_sh)
 {
-	unsigned se, sh, cu;
+	unsigned int se, sh, cu;
 	const char *p;
 
 	memset(mask, 0, sizeof(*mask) * max_se * max_sh);
@@ -158,6 +121,7 @@ void amdgpu_gfx_parse_disable_cu(unsigned *mask, unsigned max_se, unsigned max_s
 	for (;;) {
 		char *next;
 		int ret = sscanf(p, "%u.%u.%u", &se, &sh, &cu);
+
 		if (ret < 3) {
 			DRM_ERROR("amdgpu: could not parse disable_cu\n");
 			return;
@@ -178,14 +142,22 @@ void amdgpu_gfx_parse_disable_cu(unsigned *mask, unsigned max_se, unsigned max_s
 	}
 }
 
-static bool amdgpu_gfx_is_multipipe_capable(struct amdgpu_device *adev)
+static bool amdgpu_gfx_is_graphics_multipipe_capable(struct amdgpu_device *adev)
+{
+	return amdgpu_async_gfx_ring && adev->gfx.me.num_pipe_per_me > 1;
+}
+
+static bool amdgpu_gfx_is_compute_multipipe_capable(struct amdgpu_device *adev)
 {
 	if (amdgpu_compute_multipipe != -1) {
-		DRM_INFO("amdgpu: forcing compute pipe policy %d\n",
+		dev_info(adev->dev, "amdgpu: forcing compute pipe policy %d\n",
 			 amdgpu_compute_multipipe);
 		return amdgpu_compute_multipipe == 1;
 	}
 
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) > IP_VERSION(9, 0, 0))
+		return true;
+
 	/* FIXME: spreading the queues across pipes causes perf regressions
 	 * on POLARIS11 compute workloads */
 	if (adev->asic_type == CHIP_POLARIS11)
@@ -194,6 +166,28 @@ static bool amdgpu_gfx_is_multipipe_capable(struct amdgpu_device *adev)
 	return adev->gfx.mec.num_mec > 1;
 }
 
+bool amdgpu_gfx_is_high_priority_graphics_queue(struct amdgpu_device *adev,
+						struct amdgpu_ring *ring)
+{
+	int queue = ring->queue;
+	int pipe = ring->pipe;
+
+	/* Policy: use pipe1 queue0 as high priority graphics queue if we
+	 * have more than one gfx pipe.
+	 */
+	if (amdgpu_gfx_is_graphics_multipipe_capable(adev) &&
+	    adev->gfx.num_gfx_rings > 1 && pipe == 1 && queue == 0) {
+		int me = ring->me;
+		int bit;
+
+		bit = amdgpu_gfx_me_queue_to_bit(adev, me, pipe, queue);
+		if (ring == &adev->gfx.gfx_ring[bit])
+			return true;
+	}
+
+	return false;
+}
+
 bool amdgpu_gfx_is_high_priority_compute_queue(struct amdgpu_device *adev,
 					       struct amdgpu_ring *ring)
 {
@@ -209,55 +203,71 @@ bool amdgpu_gfx_is_high_priority_compute_queue(struct amdgpu_device *adev,
 
 void amdgpu_gfx_compute_queue_acquire(struct amdgpu_device *adev)
 {
-	int i, queue, pipe;
-	bool multipipe_policy = amdgpu_gfx_is_multipipe_capable(adev);
+	int i, j, queue, pipe;
+	bool multipipe_policy = amdgpu_gfx_is_compute_multipipe_capable(adev);
 	int max_queues_per_mec = min(adev->gfx.mec.num_pipe_per_mec *
 				     adev->gfx.mec.num_queue_per_pipe,
 				     adev->gfx.num_compute_rings);
+	int num_xcc = adev->gfx.xcc_mask ? NUM_XCC(adev->gfx.xcc_mask) : 1;
 
 	if (multipipe_policy) {
-		/* policy: make queues evenly cross all pipes on MEC1 only */
-		for (i = 0; i < max_queues_per_mec; i++) {
-			pipe = i % adev->gfx.mec.num_pipe_per_mec;
-			queue = (i / adev->gfx.mec.num_pipe_per_mec) %
-				adev->gfx.mec.num_queue_per_pipe;
-
-			set_bit(pipe * adev->gfx.mec.num_queue_per_pipe + queue,
-					adev->gfx.mec.queue_bitmap);
+		/* policy: make queues evenly cross all pipes on MEC1 only
+		 * for multiple xcc, just use the original policy for simplicity */
+		for (j = 0; j < num_xcc; j++) {
+			for (i = 0; i < max_queues_per_mec; i++) {
+				pipe = i % adev->gfx.mec.num_pipe_per_mec;
+				queue = (i / adev->gfx.mec.num_pipe_per_mec) %
+					 adev->gfx.mec.num_queue_per_pipe;
+
+				set_bit(pipe * adev->gfx.mec.num_queue_per_pipe + queue,
+					adev->gfx.mec_bitmap[j].queue_bitmap);
+			}
 		}
 	} else {
 		/* policy: amdgpu owns all queues in the given pipe */
-		for (i = 0; i < max_queues_per_mec; ++i)
-			set_bit(i, adev->gfx.mec.queue_bitmap);
+		for (j = 0; j < num_xcc; j++) {
+			for (i = 0; i < max_queues_per_mec; ++i)
+				set_bit(i, adev->gfx.mec_bitmap[j].queue_bitmap);
+		}
 	}
 
-	dev_dbg(adev->dev, "mec queue bitmap weight=%d\n", bitmap_weight(adev->gfx.mec.queue_bitmap, AMDGPU_MAX_COMPUTE_QUEUES));
+	for (j = 0; j < num_xcc; j++) {
+		dev_dbg(adev->dev, "mec queue bitmap weight=%d\n",
+			bitmap_weight(adev->gfx.mec_bitmap[j].queue_bitmap, AMDGPU_MAX_COMPUTE_QUEUES));
+	}
 }
 
 void amdgpu_gfx_graphics_queue_acquire(struct amdgpu_device *adev)
 {
-	int i, queue, me;
-
-	for (i = 0; i < AMDGPU_MAX_GFX_QUEUES; ++i) {
-		queue = i % adev->gfx.me.num_queue_per_pipe;
-		me = (i / adev->gfx.me.num_queue_per_pipe)
-		      / adev->gfx.me.num_pipe_per_me;
+	int i, queue, pipe;
+	bool multipipe_policy = amdgpu_gfx_is_graphics_multipipe_capable(adev);
+	int num_queue_per_pipe = 1; /* we only enable 1 KGQ per pipe */
+	int max_queues_per_me = adev->gfx.me.num_pipe_per_me * num_queue_per_pipe;
 
-		if (me >= adev->gfx.me.num_me)
-			break;
+	if (multipipe_policy) {
 		/* policy: amdgpu owns the first queue per pipe at this stage
 		 * will extend to mulitple queues per pipe later */
-		if (me == 0 && queue < 1)
+		for (i = 0; i < max_queues_per_me; i++) {
+			pipe = i % adev->gfx.me.num_pipe_per_me;
+			queue = (i / adev->gfx.me.num_pipe_per_me) %
+				num_queue_per_pipe;
+
+			set_bit(pipe * num_queue_per_pipe + queue,
+				adev->gfx.me.queue_bitmap);
+		}
+	} else {
+		for (i = 0; i < max_queues_per_me; ++i)
 			set_bit(i, adev->gfx.me.queue_bitmap);
 	}
 
 	/* update the number of active graphics rings */
-	adev->gfx.num_gfx_rings =
-		bitmap_weight(adev->gfx.me.queue_bitmap, AMDGPU_MAX_GFX_QUEUES);
+	if (adev->gfx.num_gfx_rings)
+		adev->gfx.num_gfx_rings =
+			bitmap_weight(adev->gfx.me.queue_bitmap, AMDGPU_MAX_GFX_QUEUES);
 }
 
 static int amdgpu_gfx_kiq_acquire(struct amdgpu_device *adev,
-				  struct amdgpu_ring *ring)
+				  struct amdgpu_ring *ring, int xcc_id)
 {
 	int queue_bit;
 	int mec, pipe, queue;
@@ -266,8 +276,8 @@ static int amdgpu_gfx_kiq_acquire(struct amdgpu_device *adev,
 		    * adev->gfx.mec.num_pipe_per_mec
 		    * adev->gfx.mec.num_queue_per_pipe;
 
-	while (queue_bit-- >= 0) {
-		if (test_bit(queue_bit, adev->gfx.mec.queue_bitmap))
+	while (--queue_bit >= 0) {
+		if (test_bit(queue_bit, adev->gfx.mec_bitmap[xcc_id].queue_bitmap))
 			continue;
 
 		amdgpu_queue_mask_bit_to_mec_queue(adev, queue_bit, &mec, &pipe, &queue);
@@ -291,11 +301,11 @@ static int amdgpu_gfx_kiq_acquire(struct amdgpu_device *adev,
 	return -EINVAL;
 }
 
-int amdgpu_gfx_kiq_init_ring(struct amdgpu_device *adev,
-			     struct amdgpu_ring *ring,
-			     struct amdgpu_irq_src *irq)
+int amdgpu_gfx_kiq_init_ring(struct amdgpu_device *adev, int xcc_id)
 {
-	struct amdgpu_kiq *kiq = &adev->gfx.kiq;
+	struct amdgpu_kiq *kiq = &adev->gfx.kiq[xcc_id];
+	struct amdgpu_irq_src *irq = &kiq->irq;
+	struct amdgpu_ring *ring = &kiq->ring;
 	int r = 0;
 
 	spin_lock_init(&kiq->ring_lock);
@@ -303,15 +313,22 @@ int amdgpu_gfx_kiq_init_ring(struct amdgpu_device *adev,
 	ring->adev = NULL;
 	ring->ring_obj = NULL;
 	ring->use_doorbell = true;
-	ring->doorbell_index = adev->doorbell_index.kiq;
-
-	r = amdgpu_gfx_kiq_acquire(adev, ring);
+	ring->xcc_id = xcc_id;
+	ring->vm_hub = AMDGPU_GFXHUB(xcc_id);
+	ring->doorbell_index =
+		(adev->doorbell_index.kiq +
+		 xcc_id * adev->doorbell_index.xcc_doorbell_range)
+		<< 1;
+
+	r = amdgpu_gfx_kiq_acquire(adev, ring, xcc_id);
 	if (r)
 		return r;
 
 	ring->eop_gpu_addr = kiq->eop_gpu_addr;
 	ring->no_scheduler = true;
-	sprintf(ring->name, "kiq_%d.%d.%d", ring->me, ring->pipe, ring->queue);
+	snprintf(ring->name, sizeof(ring->name), "kiq_%hhu.%hhu.%hhu.%hhu",
+		 (unsigned char)xcc_id, (unsigned char)ring->me,
+		 (unsigned char)ring->pipe, (unsigned char)ring->queue);
 	r = amdgpu_ring_init(adev, ring, 1024, irq, AMDGPU_CP_KIQ_IRQ_DRIVER0,
 			     AMDGPU_RING_PRIO_DEFAULT, NULL);
 	if (r)
@@ -325,19 +342,19 @@ void amdgpu_gfx_kiq_free_ring(struct amdgpu_ring *ring)
 	amdgpu_ring_fini(ring);
 }
 
-void amdgpu_gfx_kiq_fini(struct amdgpu_device *adev)
+void amdgpu_gfx_kiq_fini(struct amdgpu_device *adev, int xcc_id)
 {
-	struct amdgpu_kiq *kiq = &adev->gfx.kiq;
+	struct amdgpu_kiq *kiq = &adev->gfx.kiq[xcc_id];
 
 	amdgpu_bo_free_kernel(&kiq->eop_obj, &kiq->eop_gpu_addr, NULL);
 }
 
 int amdgpu_gfx_kiq_init(struct amdgpu_device *adev,
-			unsigned hpd_size)
+			unsigned int hpd_size, int xcc_id)
 {
 	int r;
 	u32 *hpd;
-	struct amdgpu_kiq *kiq = &adev->gfx.kiq;
+	struct amdgpu_kiq *kiq = &adev->gfx.kiq[xcc_id];
 
 	r = amdgpu_bo_create_kernel(adev, hpd_size, PAGE_SIZE,
 				    AMDGPU_GEM_DOMAIN_GTT, &kiq->eop_obj,
@@ -360,31 +377,44 @@ int amdgpu_gfx_kiq_init(struct amdgpu_device *adev,
 
 /* create MQD for each compute/gfx queue */
 int amdgpu_gfx_mqd_sw_init(struct amdgpu_device *adev,
-			   unsigned mqd_size)
+			   unsigned int mqd_size, int xcc_id)
 {
-	struct amdgpu_ring *ring = NULL;
-	int r, i;
+	int r, i, j;
+	struct amdgpu_kiq *kiq = &adev->gfx.kiq[xcc_id];
+	struct amdgpu_ring *ring = &kiq->ring;
+	u32 domain = AMDGPU_GEM_DOMAIN_GTT;
+
+#if !defined(CONFIG_ARM) && !defined(CONFIG_ARM64)
+	/* Only enable on gfx10 and 11 for now to avoid changing behavior on older chips */
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) >= IP_VERSION(10, 0, 0))
+		domain |= AMDGPU_GEM_DOMAIN_VRAM;
+#endif
 
 	/* create MQD for KIQ */
-	ring = &adev->gfx.kiq.ring;
-	if (!ring->mqd_obj) {
+	if (!adev->enable_mes_kiq && !ring->mqd_obj) {
 		/* originaly the KIQ MQD is put in GTT domain, but for SRIOV VRAM domain is a must
 		 * otherwise hypervisor trigger SAVE_VF fail after driver unloaded which mean MQD
 		 * deallocated and gart_unbind, to strict diverage we decide to use VRAM domain for
 		 * KIQ MQD no matter SRIOV or Bare-metal
 		 */
 		r = amdgpu_bo_create_kernel(adev, mqd_size, PAGE_SIZE,
-					    AMDGPU_GEM_DOMAIN_VRAM, &ring->mqd_obj,
-					    &ring->mqd_gpu_addr, &ring->mqd_ptr);
+					    AMDGPU_GEM_DOMAIN_VRAM |
+					    AMDGPU_GEM_DOMAIN_GTT,
+					    &ring->mqd_obj,
+					    &ring->mqd_gpu_addr,
+					    &ring->mqd_ptr);
 		if (r) {
 			dev_warn(adev->dev, "failed to create ring mqd ob (%d)", r);
 			return r;
 		}
 
 		/* prepare MQD backup */
-		adev->gfx.mec.mqd_backup[AMDGPU_MAX_COMPUTE_RINGS] = kmalloc(mqd_size, GFP_KERNEL);
-		if (!adev->gfx.mec.mqd_backup[AMDGPU_MAX_COMPUTE_RINGS])
-				dev_warn(adev->dev, "no memory to create MQD backup for ring %s\n", ring->name);
+		kiq->mqd_backup = kzalloc(mqd_size, GFP_KERNEL);
+		if (!kiq->mqd_backup) {
+			dev_warn(adev->dev,
+				 "no memory to create MQD backup for ring %s\n", ring->name);
+			return -ENOMEM;
+		}
 	}
 
 	if (adev->asic_type >= CHIP_NAVI10 && amdgpu_async_gfx_ring) {
@@ -393,47 +423,55 @@ int amdgpu_gfx_mqd_sw_init(struct amdgpu_device *adev,
 			ring = &adev->gfx.gfx_ring[i];
 			if (!ring->mqd_obj) {
 				r = amdgpu_bo_create_kernel(adev, mqd_size, PAGE_SIZE,
-							    AMDGPU_GEM_DOMAIN_GTT, &ring->mqd_obj,
+							    domain, &ring->mqd_obj,
 							    &ring->mqd_gpu_addr, &ring->mqd_ptr);
 				if (r) {
 					dev_warn(adev->dev, "failed to create ring mqd bo (%d)", r);
 					return r;
 				}
 
+				ring->mqd_size = mqd_size;
 				/* prepare MQD backup */
-				adev->gfx.me.mqd_backup[i] = kmalloc(mqd_size, GFP_KERNEL);
-				if (!adev->gfx.me.mqd_backup[i])
+				adev->gfx.me.mqd_backup[i] = kzalloc(mqd_size, GFP_KERNEL);
+				if (!adev->gfx.me.mqd_backup[i]) {
 					dev_warn(adev->dev, "no memory to create MQD backup for ring %s\n", ring->name);
+					return -ENOMEM;
+				}
 			}
 		}
 	}
 
 	/* create MQD for each KCQ */
 	for (i = 0; i < adev->gfx.num_compute_rings; i++) {
-		ring = &adev->gfx.compute_ring[i];
+		j = i + xcc_id * adev->gfx.num_compute_rings;
+		ring = &adev->gfx.compute_ring[j];
 		if (!ring->mqd_obj) {
 			r = amdgpu_bo_create_kernel(adev, mqd_size, PAGE_SIZE,
-						    AMDGPU_GEM_DOMAIN_GTT, &ring->mqd_obj,
+						    domain, &ring->mqd_obj,
 						    &ring->mqd_gpu_addr, &ring->mqd_ptr);
 			if (r) {
 				dev_warn(adev->dev, "failed to create ring mqd bo (%d)", r);
 				return r;
 			}
 
+			ring->mqd_size = mqd_size;
 			/* prepare MQD backup */
-			adev->gfx.mec.mqd_backup[i] = kmalloc(mqd_size, GFP_KERNEL);
-			if (!adev->gfx.mec.mqd_backup[i])
+			adev->gfx.mec.mqd_backup[j] = kzalloc(mqd_size, GFP_KERNEL);
+			if (!adev->gfx.mec.mqd_backup[j]) {
 				dev_warn(adev->dev, "no memory to create MQD backup for ring %s\n", ring->name);
+				return -ENOMEM;
+			}
 		}
 	}
 
 	return 0;
 }
 
-void amdgpu_gfx_mqd_sw_fini(struct amdgpu_device *adev)
+void amdgpu_gfx_mqd_sw_fini(struct amdgpu_device *adev, int xcc_id)
 {
 	struct amdgpu_ring *ring = NULL;
-	int i;
+	int i, j;
+	struct amdgpu_kiq *kiq = &adev->gfx.kiq[xcc_id];
 
 	if (adev->asic_type >= CHIP_NAVI10 && amdgpu_async_gfx_ring) {
 		for (i = 0; i < adev->gfx.num_gfx_rings; i++) {
@@ -446,41 +484,121 @@ void amdgpu_gfx_mqd_sw_fini(struct amdgpu_device *adev)
 	}
 
 	for (i = 0; i < adev->gfx.num_compute_rings; i++) {
-		ring = &adev->gfx.compute_ring[i];
-		kfree(adev->gfx.mec.mqd_backup[i]);
+		j = i + xcc_id * adev->gfx.num_compute_rings;
+		ring = &adev->gfx.compute_ring[j];
+		kfree(adev->gfx.mec.mqd_backup[j]);
 		amdgpu_bo_free_kernel(&ring->mqd_obj,
 				      &ring->mqd_gpu_addr,
 				      &ring->mqd_ptr);
 	}
 
-	ring = &adev->gfx.kiq.ring;
-	kfree(adev->gfx.mec.mqd_backup[AMDGPU_MAX_COMPUTE_RINGS]);
+	ring = &kiq->ring;
+	kfree(kiq->mqd_backup);
 	amdgpu_bo_free_kernel(&ring->mqd_obj,
 			      &ring->mqd_gpu_addr,
 			      &ring->mqd_ptr);
 }
 
-int amdgpu_gfx_disable_kcq(struct amdgpu_device *adev)
+int amdgpu_gfx_disable_kcq(struct amdgpu_device *adev, int xcc_id)
 {
-	struct amdgpu_kiq *kiq = &adev->gfx.kiq;
+	struct amdgpu_kiq *kiq = &adev->gfx.kiq[xcc_id];
 	struct amdgpu_ring *kiq_ring = &kiq->ring;
-	int i, r;
+	int i, r = 0;
+	int j;
+
+	if (adev->enable_mes) {
+		for (i = 0; i < adev->gfx.num_compute_rings; i++) {
+			j = i + xcc_id * adev->gfx.num_compute_rings;
+			amdgpu_mes_unmap_legacy_queue(adev,
+						   &adev->gfx.compute_ring[j],
+						   RESET_QUEUES, 0, 0);
+		}
+		return 0;
+	}
 
 	if (!kiq->pmf || !kiq->pmf->kiq_unmap_queues)
 		return -EINVAL;
 
-	spin_lock(&adev->gfx.kiq.ring_lock);
+	if (!kiq_ring->sched.ready || amdgpu_in_reset(adev))
+		return 0;
+
+	spin_lock(&kiq->ring_lock);
 	if (amdgpu_ring_alloc(kiq_ring, kiq->pmf->unmap_queues_size *
 					adev->gfx.num_compute_rings)) {
-		spin_unlock(&adev->gfx.kiq.ring_lock);
+		spin_unlock(&kiq->ring_lock);
 		return -ENOMEM;
 	}
 
-	for (i = 0; i < adev->gfx.num_compute_rings; i++)
-		kiq->pmf->kiq_unmap_queues(kiq_ring, &adev->gfx.compute_ring[i],
+	for (i = 0; i < adev->gfx.num_compute_rings; i++) {
+		j = i + xcc_id * adev->gfx.num_compute_rings;
+		kiq->pmf->kiq_unmap_queues(kiq_ring,
+					   &adev->gfx.compute_ring[j],
 					   RESET_QUEUES, 0, 0);
+	}
+	/* Submit unmap queue packet */
+	amdgpu_ring_commit(kiq_ring);
+	/*
+	 * Ring test will do a basic scratch register change check. Just run
+	 * this to ensure that unmap queues that is submitted before got
+	 * processed successfully before returning.
+	 */
 	r = amdgpu_ring_test_helper(kiq_ring);
-	spin_unlock(&adev->gfx.kiq.ring_lock);
+
+	spin_unlock(&kiq->ring_lock);
+
+	return r;
+}
+
+int amdgpu_gfx_disable_kgq(struct amdgpu_device *adev, int xcc_id)
+{
+	struct amdgpu_kiq *kiq = &adev->gfx.kiq[xcc_id];
+	struct amdgpu_ring *kiq_ring = &kiq->ring;
+	int i, r = 0;
+	int j;
+
+	if (adev->enable_mes) {
+		if (amdgpu_gfx_is_master_xcc(adev, xcc_id)) {
+			for (i = 0; i < adev->gfx.num_gfx_rings; i++) {
+				j = i + xcc_id * adev->gfx.num_gfx_rings;
+				amdgpu_mes_unmap_legacy_queue(adev,
+						      &adev->gfx.gfx_ring[j],
+						      PREEMPT_QUEUES, 0, 0);
+			}
+		}
+		return 0;
+	}
+
+	if (!kiq->pmf || !kiq->pmf->kiq_unmap_queues)
+		return -EINVAL;
+
+	if (!adev->gfx.kiq[0].ring.sched.ready || amdgpu_in_reset(adev))
+		return 0;
+
+	if (amdgpu_gfx_is_master_xcc(adev, xcc_id)) {
+		spin_lock(&kiq->ring_lock);
+		if (amdgpu_ring_alloc(kiq_ring, kiq->pmf->unmap_queues_size *
+						adev->gfx.num_gfx_rings)) {
+			spin_unlock(&kiq->ring_lock);
+			return -ENOMEM;
+		}
+
+		for (i = 0; i < adev->gfx.num_gfx_rings; i++) {
+			j = i + xcc_id * adev->gfx.num_gfx_rings;
+			kiq->pmf->kiq_unmap_queues(kiq_ring,
+						   &adev->gfx.gfx_ring[j],
+						   PREEMPT_QUEUES, 0, 0);
+		}
+		/* Submit unmap queue packet */
+		amdgpu_ring_commit(kiq_ring);
+
+		/*
+		 * Ring test will do a basic scratch register change check.
+		 * Just run this to ensure that unmap queues that is submitted
+		 * before got processed successfully before returning.
+		 */
+		r = amdgpu_ring_test_helper(kiq_ring);
+		spin_unlock(&kiq->ring_lock);
+	}
 
 	return r;
 }
@@ -498,67 +616,167 @@ int amdgpu_queue_mask_bit_to_set_resource_bit(struct amdgpu_device *adev,
 	return set_resource_bit;
 }
 
-int amdgpu_gfx_enable_kcq(struct amdgpu_device *adev)
+static int amdgpu_gfx_mes_enable_kcq(struct amdgpu_device *adev, int xcc_id)
 {
-	struct amdgpu_kiq *kiq = &adev->gfx.kiq;
-	struct amdgpu_ring *kiq_ring = &adev->gfx.kiq.ring;
+	struct amdgpu_kiq *kiq = &adev->gfx.kiq[xcc_id];
+	struct amdgpu_ring *kiq_ring = &kiq->ring;
+	uint64_t queue_mask = ~0ULL;
+	int r, i, j;
+
+	amdgpu_device_flush_hdp(adev, NULL);
+
+	if (!adev->enable_uni_mes) {
+		spin_lock(&kiq->ring_lock);
+		r = amdgpu_ring_alloc(kiq_ring, kiq->pmf->set_resources_size);
+		if (r) {
+			dev_err(adev->dev, "Failed to lock KIQ (%d).\n", r);
+			spin_unlock(&kiq->ring_lock);
+			return r;
+		}
+
+		kiq->pmf->kiq_set_resources(kiq_ring, queue_mask);
+		r = amdgpu_ring_test_helper(kiq_ring);
+		spin_unlock(&kiq->ring_lock);
+		if (r)
+			dev_err(adev->dev, "KIQ failed to set resources\n");
+	}
+
+	for (i = 0; i < adev->gfx.num_compute_rings; i++) {
+		j = i + xcc_id * adev->gfx.num_compute_rings;
+		r = amdgpu_mes_map_legacy_queue(adev,
+						&adev->gfx.compute_ring[j]);
+		if (r) {
+			dev_err(adev->dev, "failed to map compute queue\n");
+			return r;
+		}
+	}
+
+	return 0;
+}
+
+int amdgpu_gfx_enable_kcq(struct amdgpu_device *adev, int xcc_id)
+{
+	struct amdgpu_kiq *kiq = &adev->gfx.kiq[xcc_id];
+	struct amdgpu_ring *kiq_ring = &kiq->ring;
 	uint64_t queue_mask = 0;
-	int r, i;
+	int r, i, j;
+
+	if (adev->mes.enable_legacy_queue_map)
+		return amdgpu_gfx_mes_enable_kcq(adev, xcc_id);
 
 	if (!kiq->pmf || !kiq->pmf->kiq_map_queues || !kiq->pmf->kiq_set_resources)
 		return -EINVAL;
 
 	for (i = 0; i < AMDGPU_MAX_COMPUTE_QUEUES; ++i) {
-		if (!test_bit(i, adev->gfx.mec.queue_bitmap))
+		if (!test_bit(i, adev->gfx.mec_bitmap[xcc_id].queue_bitmap))
 			continue;
 
 		/* This situation may be hit in the future if a new HW
 		 * generation exposes more than 64 queues. If so, the
 		 * definition of queue_mask needs updating */
 		if (WARN_ON(i > (sizeof(queue_mask)*8))) {
-			DRM_ERROR("Invalid KCQ enabled: %d\n", i);
+			dev_err(adev->dev, "Invalid KCQ enabled: %d\n", i);
 			break;
 		}
 
 		queue_mask |= (1ull << amdgpu_queue_mask_bit_to_set_resource_bit(adev, i));
 	}
 
-	DRM_INFO("kiq ring mec %d pipe %d q %d\n", kiq_ring->me, kiq_ring->pipe,
-							kiq_ring->queue);
-	spin_lock(&adev->gfx.kiq.ring_lock);
+	amdgpu_device_flush_hdp(adev, NULL);
+
+	dev_info(adev->dev, "kiq ring mec %d pipe %d q %d\n", kiq_ring->me,
+		 kiq_ring->pipe, kiq_ring->queue);
+
+	spin_lock(&kiq->ring_lock);
 	r = amdgpu_ring_alloc(kiq_ring, kiq->pmf->map_queues_size *
 					adev->gfx.num_compute_rings +
 					kiq->pmf->set_resources_size);
 	if (r) {
-		DRM_ERROR("Failed to lock KIQ (%d).\n", r);
-		spin_unlock(&adev->gfx.kiq.ring_lock);
+		dev_err(adev->dev, "Failed to lock KIQ (%d).\n", r);
+		spin_unlock(&kiq->ring_lock);
 		return r;
 	}
 
 	kiq->pmf->kiq_set_resources(kiq_ring, queue_mask);
-	for (i = 0; i < adev->gfx.num_compute_rings; i++)
-		kiq->pmf->kiq_map_queues(kiq_ring, &adev->gfx.compute_ring[i]);
-
+	for (i = 0; i < adev->gfx.num_compute_rings; i++) {
+		j = i + xcc_id * adev->gfx.num_compute_rings;
+		kiq->pmf->kiq_map_queues(kiq_ring,
+					 &adev->gfx.compute_ring[j]);
+	}
+	/* Submit map queue packet */
+	amdgpu_ring_commit(kiq_ring);
+	/*
+	 * Ring test will do a basic scratch register change check. Just run
+	 * this to ensure that map queues that is submitted before got
+	 * processed successfully before returning.
+	 */
 	r = amdgpu_ring_test_helper(kiq_ring);
-	spin_unlock(&adev->gfx.kiq.ring_lock);
+	spin_unlock(&kiq->ring_lock);
 	if (r)
-		DRM_ERROR("KCQ enable failed\n");
+		dev_err(adev->dev, "KCQ enable failed\n");
 
 	return r;
 }
 
-/* amdgpu_gfx_off_ctrl - Handle gfx off feature enable/disable
- *
- * @adev: amdgpu_device pointer
- * @bool enable true: enable gfx off feature, false: disable gfx off feature
- *
- * 1. gfx off feature will be enabled by gfx ip after gfx cg gp enabled.
- * 2. other client can send request to disable gfx off feature, the request should be honored.
- * 3. other client can cancel their request of disable gfx off feature
- * 4. other client should not send request to enable gfx off feature before disable gfx off feature.
- */
+int amdgpu_gfx_enable_kgq(struct amdgpu_device *adev, int xcc_id)
+{
+	struct amdgpu_kiq *kiq = &adev->gfx.kiq[xcc_id];
+	struct amdgpu_ring *kiq_ring = &kiq->ring;
+	int r, i, j;
 
-void amdgpu_gfx_off_ctrl(struct amdgpu_device *adev, bool enable)
+	if (!kiq->pmf || !kiq->pmf->kiq_map_queues)
+		return -EINVAL;
+
+	amdgpu_device_flush_hdp(adev, NULL);
+
+	if (adev->mes.enable_legacy_queue_map) {
+		for (i = 0; i < adev->gfx.num_gfx_rings; i++) {
+			j = i + xcc_id * adev->gfx.num_gfx_rings;
+			r = amdgpu_mes_map_legacy_queue(adev,
+							&adev->gfx.gfx_ring[j]);
+			if (r) {
+				dev_err(adev->dev, "failed to map gfx queue\n");
+				return r;
+			}
+		}
+
+		return 0;
+	}
+
+	spin_lock(&kiq->ring_lock);
+	/* No need to map kcq on the slave */
+	if (amdgpu_gfx_is_master_xcc(adev, xcc_id)) {
+		r = amdgpu_ring_alloc(kiq_ring, kiq->pmf->map_queues_size *
+						adev->gfx.num_gfx_rings);
+		if (r) {
+			dev_err(adev->dev, "Failed to lock KIQ (%d).\n", r);
+			spin_unlock(&kiq->ring_lock);
+			return r;
+		}
+
+		for (i = 0; i < adev->gfx.num_gfx_rings; i++) {
+			j = i + xcc_id * adev->gfx.num_gfx_rings;
+			kiq->pmf->kiq_map_queues(kiq_ring,
+						 &adev->gfx.gfx_ring[j]);
+		}
+	}
+	/* Submit map queue packet */
+	amdgpu_ring_commit(kiq_ring);
+	/*
+	 * Ring test will do a basic scratch register change check. Just run
+	 * this to ensure that map queues that is submitted before got
+	 * processed successfully before returning.
+	 */
+	r = amdgpu_ring_test_helper(kiq_ring);
+	spin_unlock(&kiq->ring_lock);
+	if (r)
+		dev_err(adev->dev, "KGQ enable failed\n");
+
+	return r;
+}
+
+static void amdgpu_gfx_do_off_ctrl(struct amdgpu_device *adev, bool enable,
+				   bool no_delay)
 {
 	unsigned long delay = GFX_OFF_DELAY_ENABLE;
 
@@ -580,17 +798,21 @@ void amdgpu_gfx_off_ctrl(struct amdgpu_device *adev, bool enable)
 		if (adev->gfx.gfx_off_req_count == 0 &&
 		    !adev->gfx.gfx_off_state) {
 			/* If going to s2idle, no need to wait */
-			if (adev->in_s0ix)
-				delay = GFX_OFF_NO_DELAY;
-			schedule_delayed_work(&adev->gfx.gfx_off_delay_work,
+			if (no_delay) {
+				if (!amdgpu_dpm_set_powergating_by_smu(adev,
+						AMD_IP_BLOCK_TYPE_GFX, true, 0))
+					adev->gfx.gfx_off_state = true;
+			} else {
+				schedule_delayed_work(&adev->gfx.gfx_off_delay_work,
 					      delay);
+			}
 		}
 	} else {
 		if (adev->gfx.gfx_off_req_count == 0) {
 			cancel_delayed_work_sync(&adev->gfx.gfx_off_delay_work);
 
 			if (adev->gfx.gfx_off_state &&
-			    !amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_GFX, false)) {
+			    !amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_GFX, false, 0)) {
 				adev->gfx.gfx_off_state = false;
 
 				if (adev->gfx.funcs->init_spm_golden) {
@@ -608,6 +830,82 @@ unlock:
 	mutex_unlock(&adev->gfx.gfx_off_mutex);
 }
 
+/* amdgpu_gfx_off_ctrl - Handle gfx off feature enable/disable
+ *
+ * @adev: amdgpu_device pointer
+ * @bool enable true: enable gfx off feature, false: disable gfx off feature
+ *
+ * 1. gfx off feature will be enabled by gfx ip after gfx cg pg enabled.
+ * 2. other client can send request to disable gfx off feature, the request should be honored.
+ * 3. other client can cancel their request of disable gfx off feature
+ * 4. other client should not send request to enable gfx off feature before disable gfx off feature.
+ *
+ * gfx off allow will be delayed by GFX_OFF_DELAY_ENABLE ms.
+ */
+void amdgpu_gfx_off_ctrl(struct amdgpu_device *adev, bool enable)
+{
+	/* If going to s2idle, no need to wait */
+	bool no_delay = adev->in_s0ix ? true : false;
+
+	amdgpu_gfx_do_off_ctrl(adev, enable, no_delay);
+}
+
+/* amdgpu_gfx_off_ctrl_immediate - Handle gfx off feature enable/disable
+ *
+ * @adev: amdgpu_device pointer
+ * @bool enable true: enable gfx off feature, false: disable gfx off feature
+ *
+ * 1. gfx off feature will be enabled by gfx ip after gfx cg pg enabled.
+ * 2. other client can send request to disable gfx off feature, the request should be honored.
+ * 3. other client can cancel their request of disable gfx off feature
+ * 4. other client should not send request to enable gfx off feature before disable gfx off feature.
+ *
+ * gfx off allow will be issued immediately.
+ */
+void amdgpu_gfx_off_ctrl_immediate(struct amdgpu_device *adev, bool enable)
+{
+	amdgpu_gfx_do_off_ctrl(adev, enable, true);
+}
+
+int amdgpu_set_gfx_off_residency(struct amdgpu_device *adev, bool value)
+{
+	int r = 0;
+
+	mutex_lock(&adev->gfx.gfx_off_mutex);
+
+	r = amdgpu_dpm_set_residency_gfxoff(adev, value);
+
+	mutex_unlock(&adev->gfx.gfx_off_mutex);
+
+	return r;
+}
+
+int amdgpu_get_gfx_off_residency(struct amdgpu_device *adev, u32 *value)
+{
+	int r = 0;
+
+	mutex_lock(&adev->gfx.gfx_off_mutex);
+
+	r = amdgpu_dpm_get_residency_gfxoff(adev, value);
+
+	mutex_unlock(&adev->gfx.gfx_off_mutex);
+
+	return r;
+}
+
+int amdgpu_get_gfx_off_entrycount(struct amdgpu_device *adev, u64 *value)
+{
+	int r = 0;
+
+	mutex_lock(&adev->gfx.gfx_off_mutex);
+
+	r = amdgpu_dpm_get_entrycount_gfxoff(adev, value);
+
+	mutex_unlock(&adev->gfx.gfx_off_mutex);
+
+	return r;
+}
+
 int amdgpu_get_gfx_off_status(struct amdgpu_device *adev, uint32_t *value)
 {
 
@@ -615,72 +913,88 @@ int amdgpu_get_gfx_off_status(struct amdgpu_device *adev, uint32_t *value)
 
 	mutex_lock(&adev->gfx.gfx_off_mutex);
 
-	r = smu_get_status_gfxoff(adev, value);
+	r = amdgpu_dpm_get_status_gfxoff(adev, value);
 
 	mutex_unlock(&adev->gfx.gfx_off_mutex);
 
 	return r;
 }
 
-int amdgpu_gfx_ras_late_init(struct amdgpu_device *adev)
+int amdgpu_gfx_ras_late_init(struct amdgpu_device *adev, struct ras_common_if *ras_block)
 {
 	int r;
-	struct ras_fs_if fs_info = {
-		.sysfs_name = "gfx_err_count",
-	};
-	struct ras_ih_if ih_info = {
-		.cb = amdgpu_gfx_process_ras_data_cb,
-	};
 
-	if (!adev->gfx.ras_if) {
-		adev->gfx.ras_if = kmalloc(sizeof(struct ras_common_if), GFP_KERNEL);
-		if (!adev->gfx.ras_if)
-			return -ENOMEM;
-		adev->gfx.ras_if->block = AMDGPU_RAS_BLOCK__GFX;
-		adev->gfx.ras_if->type = AMDGPU_RAS_ERROR__MULTI_UNCORRECTABLE;
-		adev->gfx.ras_if->sub_block_index = 0;
-	}
-	fs_info.head = ih_info.head = *adev->gfx.ras_if;
-	r = amdgpu_ras_late_init(adev, adev->gfx.ras_if,
-				 &fs_info, &ih_info);
-	if (r)
-		goto free;
-
-	if (amdgpu_ras_is_supported(adev, adev->gfx.ras_if->block)) {
-		if (!amdgpu_persistent_edc_harvesting_supported(adev))
-			amdgpu_ras_reset_error_status(adev, AMDGPU_RAS_BLOCK__GFX);
+	if (amdgpu_ras_is_supported(adev, ras_block->block)) {
+		if (!amdgpu_persistent_edc_harvesting_supported(adev)) {
+			r = amdgpu_ras_reset_error_status(adev, AMDGPU_RAS_BLOCK__GFX);
+			if (r)
+				return r;
+		}
 
-		r = amdgpu_irq_get(adev, &adev->gfx.cp_ecc_error_irq, 0);
+		r = amdgpu_ras_block_late_init(adev, ras_block);
 		if (r)
-			goto late_fini;
+			return r;
+
+		if (amdgpu_sriov_vf(adev))
+			return r;
+
+		if (adev->gfx.cp_ecc_error_irq.funcs) {
+			r = amdgpu_irq_get(adev, &adev->gfx.cp_ecc_error_irq, 0);
+			if (r)
+				goto late_fini;
+		}
 	} else {
-		/* free gfx ras_if if ras is not supported */
-		r = 0;
-		goto free;
+		amdgpu_ras_feature_enable_on_boot(adev, ras_block, 0);
 	}
 
 	return 0;
 late_fini:
-	amdgpu_ras_late_fini(adev, adev->gfx.ras_if, &ih_info);
-free:
-	kfree(adev->gfx.ras_if);
-	adev->gfx.ras_if = NULL;
+	amdgpu_ras_block_late_fini(adev, ras_block);
 	return r;
 }
 
-void amdgpu_gfx_ras_fini(struct amdgpu_device *adev)
+int amdgpu_gfx_ras_sw_init(struct amdgpu_device *adev)
 {
-	if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__GFX) &&
-			adev->gfx.ras_if) {
-		struct ras_common_if *ras_if = adev->gfx.ras_if;
-		struct ras_ih_if ih_info = {
-			.head = *ras_if,
-			.cb = amdgpu_gfx_process_ras_data_cb,
-		};
+	int err = 0;
+	struct amdgpu_gfx_ras *ras = NULL;
+
+	/* adev->gfx.ras is NULL, which means gfx does not
+	 * support ras function, then do nothing here.
+	 */
+	if (!adev->gfx.ras)
+		return 0;
+
+	ras = adev->gfx.ras;
 
-		amdgpu_ras_late_fini(adev, ras_if, &ih_info);
-		kfree(ras_if);
+	err = amdgpu_ras_register_ras_block(adev, &ras->ras_block);
+	if (err) {
+		dev_err(adev->dev, "Failed to register gfx ras block!\n");
+		return err;
 	}
+
+	strcpy(ras->ras_block.ras_comm.name, "gfx");
+	ras->ras_block.ras_comm.block = AMDGPU_RAS_BLOCK__GFX;
+	ras->ras_block.ras_comm.type = AMDGPU_RAS_ERROR__MULTI_UNCORRECTABLE;
+	adev->gfx.ras_if = &ras->ras_block.ras_comm;
+
+	/* If not define special ras_late_init function, use gfx default ras_late_init */
+	if (!ras->ras_block.ras_late_init)
+		ras->ras_block.ras_late_init = amdgpu_gfx_ras_late_init;
+
+	/* If not defined special ras_cb function, use default ras_cb */
+	if (!ras->ras_block.ras_cb)
+		ras->ras_block.ras_cb = amdgpu_gfx_process_ras_data_cb;
+
+	return 0;
+}
+
+int amdgpu_gfx_poison_consumption_handler(struct amdgpu_device *adev,
+						struct amdgpu_iv_entry *entry)
+{
+	if (adev->gfx.ras && adev->gfx.ras->poison_consumption_handler)
+		return adev->gfx.ras->poison_consumption_handler(adev, entry);
+
+	return 0;
 }
 
 int amdgpu_gfx_process_ras_data_cb(struct amdgpu_device *adev,
@@ -695,9 +1009,9 @@ int amdgpu_gfx_process_ras_data_cb(struct amdgpu_device *adev,
 	 */
 	if (!amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__GFX)) {
 		kgd2kfd_set_sram_ecc_flag(adev->kfd.dev);
-		if (adev->gfx.ras_funcs &&
-		    adev->gfx.ras_funcs->query_ras_error_count)
-			adev->gfx.ras_funcs->query_ras_error_count(adev, err_data);
+		if (adev->gfx.ras && adev->gfx.ras->ras_block.hw_ops &&
+		    adev->gfx.ras->ras_block.hw_ops->query_ras_error_count)
+			adev->gfx.ras->ras_block.hw_ops->query_ras_error_count(adev, err_data);
 		amdgpu_ras_reset_gpu(adev);
 	}
 	return AMDGPU_RAS_SUCCESS;
@@ -717,22 +1031,44 @@ int amdgpu_gfx_cp_ecc_error_irq(struct amdgpu_device *adev,
 
 	ih_data.head = *ras_if;
 
-	DRM_ERROR("CP ECC ERROR IRQ\n");
+	dev_err(adev->dev, "CP ECC ERROR IRQ\n");
 	amdgpu_ras_interrupt_dispatch(adev, &ih_data);
 	return 0;
 }
 
-uint32_t amdgpu_kiq_rreg(struct amdgpu_device *adev, uint32_t reg)
+void amdgpu_gfx_ras_error_func(struct amdgpu_device *adev,
+		void *ras_error_status,
+		void (*func)(struct amdgpu_device *adev, void *ras_error_status,
+				int xcc_id))
+{
+	int i;
+	int num_xcc = adev->gfx.xcc_mask ? NUM_XCC(adev->gfx.xcc_mask) : 1;
+	uint32_t xcc_mask = GENMASK(num_xcc - 1, 0);
+	struct ras_err_data *err_data = (struct ras_err_data *)ras_error_status;
+
+	if (err_data) {
+		err_data->ue_count = 0;
+		err_data->ce_count = 0;
+	}
+
+	for_each_inst(i, xcc_mask)
+		func(adev, ras_error_status, i);
+}
+
+uint32_t amdgpu_kiq_rreg(struct amdgpu_device *adev, uint32_t reg, uint32_t xcc_id)
 {
 	signed long r, cnt = 0;
 	unsigned long flags;
 	uint32_t seq, reg_val_offs = 0, value = 0;
-	struct amdgpu_kiq *kiq = &adev->gfx.kiq;
+	struct amdgpu_kiq *kiq = &adev->gfx.kiq[xcc_id];
 	struct amdgpu_ring *ring = &kiq->ring;
 
 	if (amdgpu_device_skip_hw_access(adev))
 		return 0;
 
+	if (adev->mes.ring[0].sched.ready)
+		return amdgpu_mes_rreg(adev, reg);
+
 	BUG_ON(!ring->funcs->emit_rreg);
 
 	spin_lock_irqsave(&kiq->ring_lock, flags);
@@ -740,7 +1076,10 @@ uint32_t amdgpu_kiq_rreg(struct amdgpu_device *adev, uint32_t reg)
 		pr_err("critical bug! too many kiq readers\n");
 		goto failed_unlock;
 	}
-	amdgpu_ring_alloc(ring, 32);
+	r = amdgpu_ring_alloc(ring, 32);
+	if (r)
+		goto failed_unlock;
+
 	amdgpu_ring_emit_rreg(ring, reg, reg_val_offs);
 	r = amdgpu_fence_emit_polling(ring, &seq, MAX_KIQ_REG_WAIT);
 	if (r)
@@ -764,6 +1103,9 @@ uint32_t amdgpu_kiq_rreg(struct amdgpu_device *adev, uint32_t reg)
 
 	might_sleep();
 	while (r < 1 && cnt++ < MAX_KIQ_REG_TRY) {
+		if (amdgpu_in_reset(adev))
+			goto failed_kiq_read;
+
 		msleep(MAX_KIQ_REG_BAILOUT_INTERVAL);
 		r = amdgpu_fence_wait_polling(ring, seq, MAX_KIQ_REG_WAIT);
 	}
@@ -787,12 +1129,12 @@ failed_kiq_read:
 	return ~0;
 }
 
-void amdgpu_kiq_wreg(struct amdgpu_device *adev, uint32_t reg, uint32_t v)
+void amdgpu_kiq_wreg(struct amdgpu_device *adev, uint32_t reg, uint32_t v, uint32_t xcc_id)
 {
 	signed long r, cnt = 0;
 	unsigned long flags;
 	uint32_t seq;
-	struct amdgpu_kiq *kiq = &adev->gfx.kiq;
+	struct amdgpu_kiq *kiq = &adev->gfx.kiq[xcc_id];
 	struct amdgpu_ring *ring = &kiq->ring;
 
 	BUG_ON(!ring->funcs->emit_wreg);
@@ -800,8 +1142,16 @@ void amdgpu_kiq_wreg(struct amdgpu_device *adev, uint32_t reg, uint32_t v)
 	if (amdgpu_device_skip_hw_access(adev))
 		return;
 
+	if (adev->mes.ring[0].sched.ready) {
+		amdgpu_mes_wreg(adev, reg, v);
+		return;
+	}
+
 	spin_lock_irqsave(&kiq->ring_lock, flags);
-	amdgpu_ring_alloc(ring, 32);
+	r = amdgpu_ring_alloc(ring, 32);
+	if (r)
+		goto failed_unlock;
+
 	amdgpu_ring_emit_wreg(ring, reg, v);
 	r = amdgpu_fence_emit_polling(ring, &seq, MAX_KIQ_REG_WAIT);
 	if (r)
@@ -825,6 +1175,8 @@ void amdgpu_kiq_wreg(struct amdgpu_device *adev, uint32_t reg, uint32_t v)
 
 	might_sleep();
 	while (r < 1 && cnt++ < MAX_KIQ_REG_TRY) {
+		if (amdgpu_in_reset(adev))
+			goto failed_kiq_write;
 
 		msleep(MAX_KIQ_REG_BAILOUT_INTERVAL);
 		r = amdgpu_fence_wait_polling(ring, seq, MAX_KIQ_REG_WAIT);
@@ -837,11 +1189,81 @@ void amdgpu_kiq_wreg(struct amdgpu_device *adev, uint32_t reg, uint32_t v)
 
 failed_undo:
 	amdgpu_ring_undo(ring);
+failed_unlock:
 	spin_unlock_irqrestore(&kiq->ring_lock, flags);
 failed_kiq_write:
 	dev_err(adev->dev, "failed to write reg:%x\n", reg);
 }
 
+int amdgpu_kiq_hdp_flush(struct amdgpu_device *adev)
+{
+	signed long r, cnt = 0;
+	unsigned long flags;
+	uint32_t seq;
+	struct amdgpu_kiq *kiq = &adev->gfx.kiq[0];
+	struct amdgpu_ring *ring = &kiq->ring;
+
+	if (amdgpu_device_skip_hw_access(adev))
+		return 0;
+
+	if (adev->enable_mes_kiq && adev->mes.ring[0].sched.ready)
+		return amdgpu_mes_hdp_flush(adev);
+
+	if (!ring->funcs->emit_hdp_flush) {
+		return -EOPNOTSUPP;
+	}
+
+	spin_lock_irqsave(&kiq->ring_lock, flags);
+	r = amdgpu_ring_alloc(ring, 32);
+	if (r)
+		goto failed_unlock;
+
+	amdgpu_ring_emit_hdp_flush(ring);
+	r = amdgpu_fence_emit_polling(ring, &seq, MAX_KIQ_REG_WAIT);
+	if (r)
+		goto failed_undo;
+
+	amdgpu_ring_commit(ring);
+	spin_unlock_irqrestore(&kiq->ring_lock, flags);
+
+	r = amdgpu_fence_wait_polling(ring, seq, MAX_KIQ_REG_WAIT);
+
+	/* don't wait anymore for gpu reset case because this way may
+	 * block gpu_recover() routine forever, e.g. this virt_kiq_rreg
+	 * is triggered in TTM and ttm_bo_lock_delayed_workqueue() will
+	 * never return if we keep waiting in virt_kiq_rreg, which cause
+	 * gpu_recover() hang there.
+	 *
+	 * also don't wait anymore for IRQ context
+	 * */
+	if (r < 1 && (amdgpu_in_reset(adev) || in_interrupt()))
+		goto failed_kiq_hdp_flush;
+
+	might_sleep();
+	while (r < 1 && cnt++ < MAX_KIQ_REG_TRY) {
+		if (amdgpu_in_reset(adev))
+			goto failed_kiq_hdp_flush;
+
+		msleep(MAX_KIQ_REG_BAILOUT_INTERVAL);
+		r = amdgpu_fence_wait_polling(ring, seq, MAX_KIQ_REG_WAIT);
+	}
+
+	if (cnt > MAX_KIQ_REG_TRY) {
+		dev_err(adev->dev, "failed to flush HDP via KIQ timeout\n");
+		return -ETIMEDOUT;
+	}
+
+	return 0;
+
+failed_undo:
+	amdgpu_ring_undo(ring);
+failed_unlock:
+	spin_unlock_irqrestore(&kiq->ring_lock, flags);
+failed_kiq_hdp_flush:
+	dev_err(adev->dev, "failed to flush HDP via KIQ\n");
+	return r < 0 ? r : -EIO;
+}
+
 int amdgpu_gfx_get_num_kcq(struct amdgpu_device *adev)
 {
 	if (amdgpu_num_kcq == -1) {
@@ -853,18 +1275,1283 @@ int amdgpu_gfx_get_num_kcq(struct amdgpu_device *adev)
 	return amdgpu_num_kcq;
 }
 
-/* amdgpu_gfx_state_change_set - Handle gfx power state change set
+void amdgpu_gfx_cp_init_microcode(struct amdgpu_device *adev,
+				  uint32_t ucode_id)
+{
+	const struct gfx_firmware_header_v1_0 *cp_hdr;
+	const struct gfx_firmware_header_v2_0 *cp_hdr_v2_0;
+	struct amdgpu_firmware_info *info = NULL;
+	const struct firmware *ucode_fw;
+	unsigned int fw_size;
+
+	switch (ucode_id) {
+	case AMDGPU_UCODE_ID_CP_PFP:
+		cp_hdr = (const struct gfx_firmware_header_v1_0 *)
+			adev->gfx.pfp_fw->data;
+		adev->gfx.pfp_fw_version =
+			le32_to_cpu(cp_hdr->header.ucode_version);
+		adev->gfx.pfp_feature_version =
+			le32_to_cpu(cp_hdr->ucode_feature_version);
+		ucode_fw = adev->gfx.pfp_fw;
+		fw_size = le32_to_cpu(cp_hdr->header.ucode_size_bytes);
+		break;
+	case AMDGPU_UCODE_ID_CP_RS64_PFP:
+		cp_hdr_v2_0 = (const struct gfx_firmware_header_v2_0 *)
+			adev->gfx.pfp_fw->data;
+		adev->gfx.pfp_fw_version =
+			le32_to_cpu(cp_hdr_v2_0->header.ucode_version);
+		adev->gfx.pfp_feature_version =
+			le32_to_cpu(cp_hdr_v2_0->ucode_feature_version);
+		ucode_fw = adev->gfx.pfp_fw;
+		fw_size = le32_to_cpu(cp_hdr_v2_0->ucode_size_bytes);
+		break;
+	case AMDGPU_UCODE_ID_CP_RS64_PFP_P0_STACK:
+	case AMDGPU_UCODE_ID_CP_RS64_PFP_P1_STACK:
+		cp_hdr_v2_0 = (const struct gfx_firmware_header_v2_0 *)
+			adev->gfx.pfp_fw->data;
+		ucode_fw = adev->gfx.pfp_fw;
+		fw_size = le32_to_cpu(cp_hdr_v2_0->data_size_bytes);
+		break;
+	case AMDGPU_UCODE_ID_CP_ME:
+		cp_hdr = (const struct gfx_firmware_header_v1_0 *)
+			adev->gfx.me_fw->data;
+		adev->gfx.me_fw_version =
+			le32_to_cpu(cp_hdr->header.ucode_version);
+		adev->gfx.me_feature_version =
+			le32_to_cpu(cp_hdr->ucode_feature_version);
+		ucode_fw = adev->gfx.me_fw;
+		fw_size = le32_to_cpu(cp_hdr->header.ucode_size_bytes);
+		break;
+	case AMDGPU_UCODE_ID_CP_RS64_ME:
+		cp_hdr_v2_0 = (const struct gfx_firmware_header_v2_0 *)
+			adev->gfx.me_fw->data;
+		adev->gfx.me_fw_version =
+			le32_to_cpu(cp_hdr_v2_0->header.ucode_version);
+		adev->gfx.me_feature_version =
+			le32_to_cpu(cp_hdr_v2_0->ucode_feature_version);
+		ucode_fw = adev->gfx.me_fw;
+		fw_size = le32_to_cpu(cp_hdr_v2_0->ucode_size_bytes);
+		break;
+	case AMDGPU_UCODE_ID_CP_RS64_ME_P0_STACK:
+	case AMDGPU_UCODE_ID_CP_RS64_ME_P1_STACK:
+		cp_hdr_v2_0 = (const struct gfx_firmware_header_v2_0 *)
+			adev->gfx.me_fw->data;
+		ucode_fw = adev->gfx.me_fw;
+		fw_size = le32_to_cpu(cp_hdr_v2_0->data_size_bytes);
+		break;
+	case AMDGPU_UCODE_ID_CP_CE:
+		cp_hdr = (const struct gfx_firmware_header_v1_0 *)
+			adev->gfx.ce_fw->data;
+		adev->gfx.ce_fw_version =
+			le32_to_cpu(cp_hdr->header.ucode_version);
+		adev->gfx.ce_feature_version =
+			le32_to_cpu(cp_hdr->ucode_feature_version);
+		ucode_fw = adev->gfx.ce_fw;
+		fw_size = le32_to_cpu(cp_hdr->header.ucode_size_bytes);
+		break;
+	case AMDGPU_UCODE_ID_CP_MEC1:
+		cp_hdr = (const struct gfx_firmware_header_v1_0 *)
+			adev->gfx.mec_fw->data;
+		adev->gfx.mec_fw_version =
+			le32_to_cpu(cp_hdr->header.ucode_version);
+		adev->gfx.mec_feature_version =
+			le32_to_cpu(cp_hdr->ucode_feature_version);
+		ucode_fw = adev->gfx.mec_fw;
+		fw_size = le32_to_cpu(cp_hdr->header.ucode_size_bytes) -
+			  le32_to_cpu(cp_hdr->jt_size) * 4;
+		break;
+	case AMDGPU_UCODE_ID_CP_MEC1_JT:
+		cp_hdr = (const struct gfx_firmware_header_v1_0 *)
+			adev->gfx.mec_fw->data;
+		ucode_fw = adev->gfx.mec_fw;
+		fw_size = le32_to_cpu(cp_hdr->jt_size) * 4;
+		break;
+	case AMDGPU_UCODE_ID_CP_MEC2:
+		cp_hdr = (const struct gfx_firmware_header_v1_0 *)
+			adev->gfx.mec2_fw->data;
+		adev->gfx.mec2_fw_version =
+			le32_to_cpu(cp_hdr->header.ucode_version);
+		adev->gfx.mec2_feature_version =
+			le32_to_cpu(cp_hdr->ucode_feature_version);
+		ucode_fw = adev->gfx.mec2_fw;
+		fw_size = le32_to_cpu(cp_hdr->header.ucode_size_bytes) -
+			  le32_to_cpu(cp_hdr->jt_size) * 4;
+		break;
+	case AMDGPU_UCODE_ID_CP_MEC2_JT:
+		cp_hdr = (const struct gfx_firmware_header_v1_0 *)
+			adev->gfx.mec2_fw->data;
+		ucode_fw = adev->gfx.mec2_fw;
+		fw_size = le32_to_cpu(cp_hdr->jt_size) * 4;
+		break;
+	case AMDGPU_UCODE_ID_CP_RS64_MEC:
+		cp_hdr_v2_0 = (const struct gfx_firmware_header_v2_0 *)
+			adev->gfx.mec_fw->data;
+		adev->gfx.mec_fw_version =
+			le32_to_cpu(cp_hdr_v2_0->header.ucode_version);
+		adev->gfx.mec_feature_version =
+			le32_to_cpu(cp_hdr_v2_0->ucode_feature_version);
+		ucode_fw = adev->gfx.mec_fw;
+		fw_size = le32_to_cpu(cp_hdr_v2_0->ucode_size_bytes);
+		break;
+	case AMDGPU_UCODE_ID_CP_RS64_MEC_P0_STACK:
+	case AMDGPU_UCODE_ID_CP_RS64_MEC_P1_STACK:
+	case AMDGPU_UCODE_ID_CP_RS64_MEC_P2_STACK:
+	case AMDGPU_UCODE_ID_CP_RS64_MEC_P3_STACK:
+		cp_hdr_v2_0 = (const struct gfx_firmware_header_v2_0 *)
+			adev->gfx.mec_fw->data;
+		ucode_fw = adev->gfx.mec_fw;
+		fw_size = le32_to_cpu(cp_hdr_v2_0->data_size_bytes);
+		break;
+	default:
+		dev_err(adev->dev, "Invalid ucode id %u\n", ucode_id);
+		return;
+	}
+
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
+		info = &adev->firmware.ucode[ucode_id];
+		info->ucode_id = ucode_id;
+		info->fw = ucode_fw;
+		adev->firmware.fw_size += ALIGN(fw_size, PAGE_SIZE);
+	}
+}
+
+bool amdgpu_gfx_is_master_xcc(struct amdgpu_device *adev, int xcc_id)
+{
+	return !(xcc_id % (adev->gfx.num_xcc_per_xcp ?
+			adev->gfx.num_xcc_per_xcp : 1));
+}
+
+static ssize_t amdgpu_gfx_get_current_compute_partition(struct device *dev,
+						struct device_attribute *addr,
+						char *buf)
+{
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = drm_to_adev(ddev);
+	int mode;
+
+	/* Only minimal precaution taken to reject requests while in reset.*/
+	if (amdgpu_in_reset(adev))
+		return -EPERM;
+
+	mode = amdgpu_xcp_query_partition_mode(adev->xcp_mgr,
+					       AMDGPU_XCP_FL_NONE);
+
+	return sysfs_emit(buf, "%s\n", amdgpu_gfx_compute_mode_desc(mode));
+}
+
+static ssize_t amdgpu_gfx_set_compute_partition(struct device *dev,
+						struct device_attribute *addr,
+						const char *buf, size_t count)
+{
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = drm_to_adev(ddev);
+	enum amdgpu_gfx_partition mode;
+	int ret = 0, num_xcc;
+
+	num_xcc = NUM_XCC(adev->gfx.xcc_mask);
+	if (num_xcc % 2 != 0)
+		return -EINVAL;
+
+	if (!strncasecmp("SPX", buf, strlen("SPX"))) {
+		mode = AMDGPU_SPX_PARTITION_MODE;
+	} else if (!strncasecmp("DPX", buf, strlen("DPX"))) {
+		/*
+		 * DPX mode needs AIDs to be in multiple of 2.
+		 * Each AID connects 2 XCCs.
+		 */
+		if (num_xcc%4)
+			return -EINVAL;
+		mode = AMDGPU_DPX_PARTITION_MODE;
+	} else if (!strncasecmp("TPX", buf, strlen("TPX"))) {
+		if (num_xcc != 6)
+			return -EINVAL;
+		mode = AMDGPU_TPX_PARTITION_MODE;
+	} else if (!strncasecmp("QPX", buf, strlen("QPX"))) {
+		if (num_xcc != 8)
+			return -EINVAL;
+		mode = AMDGPU_QPX_PARTITION_MODE;
+	} else if (!strncasecmp("CPX", buf, strlen("CPX"))) {
+		mode = AMDGPU_CPX_PARTITION_MODE;
+	} else {
+		return -EINVAL;
+	}
+
+	/* Don't allow a switch while under reset */
+	if (!down_read_trylock(&adev->reset_domain->sem))
+		return -EPERM;
+
+	ret = amdgpu_xcp_switch_partition_mode(adev->xcp_mgr, mode);
+
+	up_read(&adev->reset_domain->sem);
+
+	if (ret)
+		return ret;
+
+	return count;
+}
+
+static const char *xcp_desc[] = {
+	[AMDGPU_SPX_PARTITION_MODE] = "SPX",
+	[AMDGPU_DPX_PARTITION_MODE] = "DPX",
+	[AMDGPU_TPX_PARTITION_MODE] = "TPX",
+	[AMDGPU_QPX_PARTITION_MODE] = "QPX",
+	[AMDGPU_CPX_PARTITION_MODE] = "CPX",
+};
+
+static ssize_t amdgpu_gfx_get_available_compute_partition(struct device *dev,
+						struct device_attribute *addr,
+						char *buf)
+{
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = drm_to_adev(ddev);
+	struct amdgpu_xcp_mgr *xcp_mgr = adev->xcp_mgr;
+	int size = 0, mode;
+	char *sep = "";
+
+	if (!xcp_mgr || !xcp_mgr->avail_xcp_modes)
+		return sysfs_emit(buf, "Not supported\n");
+
+	for_each_inst(mode, xcp_mgr->avail_xcp_modes) {
+		size += sysfs_emit_at(buf, size, "%s%s", sep, xcp_desc[mode]);
+		sep = ", ";
+	}
+
+	size += sysfs_emit_at(buf, size, "\n");
+
+	return size;
+}
+
+static int amdgpu_gfx_run_cleaner_shader_job(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct drm_gpu_scheduler *sched = &ring->sched;
+	struct drm_sched_entity entity;
+	static atomic_t counter;
+	struct dma_fence *f;
+	struct amdgpu_job *job;
+	struct amdgpu_ib *ib;
+	void *owner;
+	int i, r;
+
+	/* Initialize the scheduler entity */
+	r = drm_sched_entity_init(&entity, DRM_SCHED_PRIORITY_NORMAL,
+				  &sched, 1, NULL);
+	if (r) {
+		dev_err(adev->dev, "Failed setting up GFX kernel entity.\n");
+		goto err;
+	}
+
+	/*
+	 * Use some unique dummy value as the owner to make sure we execute
+	 * the cleaner shader on each submission. The value just need to change
+	 * for each submission and is otherwise meaningless.
+	 */
+	owner = (void *)(unsigned long)atomic_inc_return(&counter);
+
+	r = amdgpu_job_alloc_with_ib(ring->adev, &entity, owner,
+				     64, 0, &job,
+				     AMDGPU_KERNEL_JOB_ID_CLEANER_SHADER);
+	if (r)
+		goto err;
+
+	job->enforce_isolation = true;
+	/* always run the cleaner shader */
+	job->run_cleaner_shader = true;
+
+	ib = &job->ibs[0];
+	for (i = 0; i <= ring->funcs->align_mask; ++i)
+		ib->ptr[i] = ring->funcs->nop;
+	ib->length_dw = ring->funcs->align_mask + 1;
+
+	f = amdgpu_job_submit(job);
+
+	r = dma_fence_wait(f, false);
+	if (r)
+		goto err;
+
+	dma_fence_put(f);
+
+	/* Clean up the scheduler entity */
+	drm_sched_entity_destroy(&entity);
+	return 0;
+
+err:
+	return r;
+}
+
+static int amdgpu_gfx_run_cleaner_shader(struct amdgpu_device *adev, int xcp_id)
+{
+	int num_xcc = NUM_XCC(adev->gfx.xcc_mask);
+	struct amdgpu_ring *ring;
+	int num_xcc_to_clear;
+	int i, r, xcc_id;
+
+	if (adev->gfx.num_xcc_per_xcp)
+		num_xcc_to_clear = adev->gfx.num_xcc_per_xcp;
+	else
+		num_xcc_to_clear = 1;
+
+	for (xcc_id = 0; xcc_id < num_xcc; xcc_id++) {
+		for (i = 0; i < adev->gfx.num_compute_rings; i++) {
+			ring = &adev->gfx.compute_ring[i + xcc_id * adev->gfx.num_compute_rings];
+			if ((ring->xcp_id == xcp_id) && ring->sched.ready) {
+				r = amdgpu_gfx_run_cleaner_shader_job(ring);
+				if (r)
+					return r;
+				num_xcc_to_clear--;
+				break;
+			}
+		}
+	}
+
+	if (num_xcc_to_clear)
+		return -ENOENT;
+
+	return 0;
+}
+
+/**
+ * amdgpu_gfx_set_run_cleaner_shader - Execute the AMDGPU GFX Cleaner Shader
+ * @dev: The device structure
+ * @attr: The device attribute structure
+ * @buf: The buffer containing the input data
+ * @count: The size of the input data
+ *
+ * Provides the sysfs interface to manually run a cleaner shader, which is
+ * used to clear the GPU state between different tasks. Writing a value to the
+ * 'run_cleaner_shader' sysfs file triggers the cleaner shader execution.
+ * The value written corresponds to the partition index on multi-partition
+ * devices. On single-partition devices, the value should be '0'.
+ *
+ * The cleaner shader clears the Local Data Store (LDS) and General Purpose
+ * Registers (GPRs) to ensure data isolation between GPU workloads.
+ *
+ * Return: The number of bytes written to the sysfs file.
+ */
+static ssize_t amdgpu_gfx_set_run_cleaner_shader(struct device *dev,
+						 struct device_attribute *attr,
+						 const char *buf,
+						 size_t count)
+{
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = drm_to_adev(ddev);
+	int ret;
+	long value;
+
+	if (amdgpu_in_reset(adev))
+		return -EPERM;
+	if (adev->in_suspend && !adev->in_runpm)
+		return -EPERM;
+
+	if (adev->gfx.disable_kq)
+		return -EPERM;
+
+	ret = kstrtol(buf, 0, &value);
+
+	if (ret)
+		return -EINVAL;
+
+	if (value < 0)
+		return -EINVAL;
+
+	if (adev->xcp_mgr) {
+		if (value >= adev->xcp_mgr->num_xcps)
+			return -EINVAL;
+	} else {
+		if (value > 1)
+			return -EINVAL;
+	}
+
+	ret = pm_runtime_get_sync(ddev->dev);
+	if (ret < 0) {
+		pm_runtime_put_autosuspend(ddev->dev);
+		return ret;
+	}
+
+	ret = amdgpu_gfx_run_cleaner_shader(adev, value);
+
+	pm_runtime_put_autosuspend(ddev->dev);
+
+	if (ret)
+		return ret;
+
+	return count;
+}
+
+/**
+ * amdgpu_gfx_get_enforce_isolation - Query AMDGPU GFX Enforce Isolation Settings
+ * @dev: The device structure
+ * @attr: The device attribute structure
+ * @buf: The buffer to store the output data
+ *
+ * Provides the sysfs read interface to get the current settings of the 'enforce_isolation'
+ * feature for each GPU partition. Reading from the 'enforce_isolation'
+ * sysfs file returns the isolation settings for all partitions, where '0'
+ * indicates disabled, '1' indicates enabled, and '2' indicates enabled in legacy mode,
+ * and '3' indicates enabled without cleaner shader.
+ *
+ * Return: The number of bytes read from the sysfs file.
+ */
+static ssize_t amdgpu_gfx_get_enforce_isolation(struct device *dev,
+						struct device_attribute *attr,
+						char *buf)
+{
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = drm_to_adev(ddev);
+	int i;
+	ssize_t size = 0;
+
+	if (adev->xcp_mgr) {
+		for (i = 0; i < adev->xcp_mgr->num_xcps; i++) {
+			size += sysfs_emit_at(buf, size, "%u", adev->enforce_isolation[i]);
+			if (i < (adev->xcp_mgr->num_xcps - 1))
+				size += sysfs_emit_at(buf, size, " ");
+		}
+		buf[size++] = '\n';
+	} else {
+		size = sysfs_emit_at(buf, 0, "%u\n", adev->enforce_isolation[0]);
+	}
+
+	return size;
+}
+
+/**
+ * amdgpu_gfx_set_enforce_isolation - Control AMDGPU GFX Enforce Isolation
+ * @dev: The device structure
+ * @attr: The device attribute structure
+ * @buf: The buffer containing the input data
+ * @count: The size of the input data
+ *
+ * This function allows control over the 'enforce_isolation' feature, which
+ * serializes access to the graphics engine. Writing '0' to disable, '1' to
+ * enable isolation with cleaner shader, '2' to enable legacy isolation without
+ * cleaner shader, or '3' to enable process isolation without submitting the
+ * cleaner shader to the 'enforce_isolation' sysfs file sets the isolation mode
+ * for each partition. The input should specify the setting for all
+ * partitions.
+ *
+ * Return: The number of bytes written to the sysfs file.
+ */
+static ssize_t amdgpu_gfx_set_enforce_isolation(struct device *dev,
+						struct device_attribute *attr,
+						const char *buf, size_t count)
+{
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = drm_to_adev(ddev);
+	long partition_values[MAX_XCP] = {0};
+	int ret, i, num_partitions;
+	const char *input_buf = buf;
+
+	for (i = 0; i < (adev->xcp_mgr ? adev->xcp_mgr->num_xcps : 1); i++) {
+		ret = sscanf(input_buf, "%ld", &partition_values[i]);
+		if (ret <= 0)
+			break;
+
+		/* Move the pointer to the next value in the string */
+		input_buf = strchr(input_buf, ' ');
+		if (input_buf) {
+			input_buf++;
+		} else {
+			i++;
+			break;
+		}
+	}
+	num_partitions = i;
+
+	if (adev->xcp_mgr && num_partitions != adev->xcp_mgr->num_xcps)
+		return -EINVAL;
+
+	if (!adev->xcp_mgr && num_partitions != 1)
+		return -EINVAL;
+
+	for (i = 0; i < num_partitions; i++) {
+		if (partition_values[i] != 0 &&
+		    partition_values[i] != 1 &&
+		    partition_values[i] != 2 &&
+		    partition_values[i] != 3)
+			return -EINVAL;
+	}
+
+	mutex_lock(&adev->enforce_isolation_mutex);
+	for (i = 0; i < num_partitions; i++) {
+		switch (partition_values[i]) {
+		case 0:
+		default:
+			adev->enforce_isolation[i] = AMDGPU_ENFORCE_ISOLATION_DISABLE;
+			break;
+		case 1:
+			adev->enforce_isolation[i] =
+				AMDGPU_ENFORCE_ISOLATION_ENABLE;
+			break;
+		case 2:
+			adev->enforce_isolation[i] =
+				AMDGPU_ENFORCE_ISOLATION_ENABLE_LEGACY;
+			break;
+		case 3:
+			adev->enforce_isolation[i] =
+				AMDGPU_ENFORCE_ISOLATION_NO_CLEANER_SHADER;
+			break;
+		}
+	}
+	mutex_unlock(&adev->enforce_isolation_mutex);
+
+	amdgpu_mes_update_enforce_isolation(adev);
+
+	return count;
+}
+
+static ssize_t amdgpu_gfx_get_gfx_reset_mask(struct device *dev,
+						struct device_attribute *attr,
+						char *buf)
+{
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = drm_to_adev(ddev);
+
+	if (!adev)
+		return -ENODEV;
+
+	return amdgpu_show_reset_mask(buf, adev->gfx.gfx_supported_reset);
+}
+
+static ssize_t amdgpu_gfx_get_compute_reset_mask(struct device *dev,
+						struct device_attribute *attr,
+						char *buf)
+{
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = drm_to_adev(ddev);
+
+	if (!adev)
+		return -ENODEV;
+
+	return amdgpu_show_reset_mask(buf, adev->gfx.compute_supported_reset);
+}
+
+static DEVICE_ATTR(run_cleaner_shader, 0200,
+		   NULL, amdgpu_gfx_set_run_cleaner_shader);
+
+static DEVICE_ATTR(enforce_isolation, 0644,
+		   amdgpu_gfx_get_enforce_isolation,
+		   amdgpu_gfx_set_enforce_isolation);
+
+static DEVICE_ATTR(current_compute_partition, 0644,
+		   amdgpu_gfx_get_current_compute_partition,
+		   amdgpu_gfx_set_compute_partition);
+
+static DEVICE_ATTR(available_compute_partition, 0444,
+		   amdgpu_gfx_get_available_compute_partition, NULL);
+static DEVICE_ATTR(gfx_reset_mask, 0444,
+		   amdgpu_gfx_get_gfx_reset_mask, NULL);
+
+static DEVICE_ATTR(compute_reset_mask, 0444,
+		   amdgpu_gfx_get_compute_reset_mask, NULL);
+
+static int amdgpu_gfx_sysfs_xcp_init(struct amdgpu_device *adev)
+{
+	struct amdgpu_xcp_mgr *xcp_mgr = adev->xcp_mgr;
+	bool xcp_switch_supported;
+	int r;
+
+	if (!xcp_mgr)
+		return 0;
+
+	xcp_switch_supported =
+		(xcp_mgr->funcs && xcp_mgr->funcs->switch_partition_mode);
+
+	if (!xcp_switch_supported)
+		dev_attr_current_compute_partition.attr.mode &=
+			~(S_IWUSR | S_IWGRP | S_IWOTH);
+
+	r = device_create_file(adev->dev, &dev_attr_current_compute_partition);
+	if (r)
+		return r;
+
+	if (xcp_switch_supported)
+		r = device_create_file(adev->dev,
+				       &dev_attr_available_compute_partition);
+
+	return r;
+}
+
+static void amdgpu_gfx_sysfs_xcp_fini(struct amdgpu_device *adev)
+{
+	struct amdgpu_xcp_mgr *xcp_mgr = adev->xcp_mgr;
+	bool xcp_switch_supported;
+
+	if (!xcp_mgr)
+		return;
+
+	xcp_switch_supported =
+		(xcp_mgr->funcs && xcp_mgr->funcs->switch_partition_mode);
+	device_remove_file(adev->dev, &dev_attr_current_compute_partition);
+
+	if (xcp_switch_supported)
+		device_remove_file(adev->dev,
+				   &dev_attr_available_compute_partition);
+}
+
+static int amdgpu_gfx_sysfs_isolation_shader_init(struct amdgpu_device *adev)
+{
+	int r;
+
+	r = device_create_file(adev->dev, &dev_attr_enforce_isolation);
+	if (r)
+		return r;
+	if (adev->gfx.enable_cleaner_shader)
+		r = device_create_file(adev->dev, &dev_attr_run_cleaner_shader);
+
+	return r;
+}
+
+static void amdgpu_gfx_sysfs_isolation_shader_fini(struct amdgpu_device *adev)
+{
+	device_remove_file(adev->dev, &dev_attr_enforce_isolation);
+	if (adev->gfx.enable_cleaner_shader)
+		device_remove_file(adev->dev, &dev_attr_run_cleaner_shader);
+}
+
+static int amdgpu_gfx_sysfs_reset_mask_init(struct amdgpu_device *adev)
+{
+	int r = 0;
+
+	if (!amdgpu_gpu_recovery)
+		return r;
+
+	if (adev->gfx.num_gfx_rings) {
+		r = device_create_file(adev->dev, &dev_attr_gfx_reset_mask);
+		if (r)
+			return r;
+	}
+
+	if (adev->gfx.num_compute_rings) {
+		r = device_create_file(adev->dev, &dev_attr_compute_reset_mask);
+		if (r)
+			return r;
+	}
+
+	return r;
+}
+
+static void amdgpu_gfx_sysfs_reset_mask_fini(struct amdgpu_device *adev)
+{
+	if (!amdgpu_gpu_recovery)
+		return;
+
+	if (adev->gfx.num_gfx_rings)
+		device_remove_file(adev->dev, &dev_attr_gfx_reset_mask);
+
+	if (adev->gfx.num_compute_rings)
+		device_remove_file(adev->dev, &dev_attr_compute_reset_mask);
+}
+
+int amdgpu_gfx_sysfs_init(struct amdgpu_device *adev)
+{
+	int r;
+
+	r = amdgpu_gfx_sysfs_xcp_init(adev);
+	if (r) {
+		dev_err(adev->dev, "failed to create xcp sysfs files");
+		return r;
+	}
+
+	r = amdgpu_gfx_sysfs_isolation_shader_init(adev);
+	if (r)
+		dev_err(adev->dev, "failed to create isolation sysfs files");
+
+	r = amdgpu_gfx_sysfs_reset_mask_init(adev);
+	if (r)
+		dev_err(adev->dev, "failed to create reset mask sysfs files");
+
+	return r;
+}
+
+void amdgpu_gfx_sysfs_fini(struct amdgpu_device *adev)
+{
+	if (adev->dev->kobj.sd) {
+		amdgpu_gfx_sysfs_xcp_fini(adev);
+		amdgpu_gfx_sysfs_isolation_shader_fini(adev);
+		amdgpu_gfx_sysfs_reset_mask_fini(adev);
+	}
+}
+
+int amdgpu_gfx_cleaner_shader_sw_init(struct amdgpu_device *adev,
+				      unsigned int cleaner_shader_size)
+{
+	if (!adev->gfx.enable_cleaner_shader)
+		return -EOPNOTSUPP;
+
+	return amdgpu_bo_create_kernel(adev, cleaner_shader_size, PAGE_SIZE,
+				       AMDGPU_GEM_DOMAIN_VRAM | AMDGPU_GEM_DOMAIN_GTT,
+				       &adev->gfx.cleaner_shader_obj,
+				       &adev->gfx.cleaner_shader_gpu_addr,
+				       (void **)&adev->gfx.cleaner_shader_cpu_ptr);
+}
+
+void amdgpu_gfx_cleaner_shader_sw_fini(struct amdgpu_device *adev)
+{
+	if (!adev->gfx.enable_cleaner_shader)
+		return;
+
+	amdgpu_bo_free_kernel(&adev->gfx.cleaner_shader_obj,
+			      &adev->gfx.cleaner_shader_gpu_addr,
+			      (void **)&adev->gfx.cleaner_shader_cpu_ptr);
+}
+
+void amdgpu_gfx_cleaner_shader_init(struct amdgpu_device *adev,
+				    unsigned int cleaner_shader_size,
+				    const void *cleaner_shader_ptr)
+{
+	if (!adev->gfx.enable_cleaner_shader)
+		return;
+
+	if (adev->gfx.cleaner_shader_cpu_ptr && cleaner_shader_ptr)
+		memcpy_toio(adev->gfx.cleaner_shader_cpu_ptr, cleaner_shader_ptr,
+			    cleaner_shader_size);
+}
+
+/**
+ * amdgpu_gfx_kfd_sch_ctrl - Control the KFD scheduler from the KGD (Graphics Driver)
+ * @adev: amdgpu_device pointer
+ * @idx: Index of the scheduler to control
+ * @enable: Whether to enable or disable the KFD scheduler
+ *
+ * This function is used to control the KFD (Kernel Fusion Driver) scheduler
+ * from the KGD. It is part of the cleaner shader feature. This function plays
+ * a key role in enforcing process isolation on the GPU.
+ *
+ * The function uses a reference count mechanism (kfd_sch_req_count) to keep
+ * track of the number of requests to enable the KFD scheduler. When a request
+ * to enable the KFD scheduler is made, the reference count is decremented.
+ * When the reference count reaches zero, a delayed work is scheduled to
+ * enforce isolation after a delay of GFX_SLICE_PERIOD.
+ *
+ * When a request to disable the KFD scheduler is made, the function first
+ * checks if the reference count is zero. If it is, it cancels the delayed work
+ * for enforcing isolation and checks if the KFD scheduler is active. If the
+ * KFD scheduler is active, it sends a request to stop the KFD scheduler and
+ * sets the KFD scheduler state to inactive. Then, it increments the reference
+ * count.
+ *
+ * The function is synchronized using the kfd_sch_mutex to ensure that the KFD
+ * scheduler state and reference count are updated atomically.
+ *
+ * Note: If the reference count is already zero when a request to enable the
+ * KFD scheduler is made, it means there's an imbalance bug somewhere. The
+ * function triggers a warning in this case.
+ */
+static void amdgpu_gfx_kfd_sch_ctrl(struct amdgpu_device *adev, u32 idx,
+				    bool enable)
+{
+	mutex_lock(&adev->gfx.userq_sch_mutex);
+
+	if (enable) {
+		/* If the count is already 0, it means there's an imbalance bug somewhere.
+		 * Note that the bug may be in a different caller than the one which triggers the
+		 * WARN_ON_ONCE.
+		 */
+		if (WARN_ON_ONCE(adev->gfx.userq_sch_req_count[idx] == 0)) {
+			dev_err(adev->dev, "Attempted to enable KFD scheduler when reference count is already zero\n");
+			goto unlock;
+		}
+
+		adev->gfx.userq_sch_req_count[idx]--;
+
+		if (adev->gfx.userq_sch_req_count[idx] == 0 &&
+		    adev->gfx.userq_sch_inactive[idx]) {
+			schedule_delayed_work(&adev->gfx.enforce_isolation[idx].work,
+					      msecs_to_jiffies(adev->gfx.enforce_isolation_time[idx]));
+		}
+	} else {
+		if (adev->gfx.userq_sch_req_count[idx] == 0) {
+			cancel_delayed_work_sync(&adev->gfx.enforce_isolation[idx].work);
+			if (!adev->gfx.userq_sch_inactive[idx]) {
+				amdgpu_userq_stop_sched_for_enforce_isolation(adev, idx);
+				if (adev->kfd.init_complete)
+					amdgpu_amdkfd_stop_sched(adev, idx);
+				adev->gfx.userq_sch_inactive[idx] = true;
+			}
+		}
+
+		adev->gfx.userq_sch_req_count[idx]++;
+	}
+
+unlock:
+	mutex_unlock(&adev->gfx.userq_sch_mutex);
+}
+
+/**
+ * amdgpu_gfx_enforce_isolation_handler - work handler for enforcing shader isolation
+ *
+ * @work: work_struct.
+ *
+ * This function is the work handler for enforcing shader isolation on AMD GPUs.
+ * It counts the number of emitted fences for each GFX and compute ring. If there
+ * are any fences, it schedules the `enforce_isolation_work` to be run after a
+ * delay of `GFX_SLICE_PERIOD`. If there are no fences, it signals the Kernel Fusion
+ * Driver (KFD) to resume the runqueue. The function is synchronized using the
+ * `enforce_isolation_mutex`.
+ */
+void amdgpu_gfx_enforce_isolation_handler(struct work_struct *work)
+{
+	struct amdgpu_isolation_work *isolation_work =
+		container_of(work, struct amdgpu_isolation_work, work.work);
+	struct amdgpu_device *adev = isolation_work->adev;
+	u32 i, idx, fences = 0;
+
+	if (isolation_work->xcp_id == AMDGPU_XCP_NO_PARTITION)
+		idx = 0;
+	else
+		idx = isolation_work->xcp_id;
+
+	if (idx >= MAX_XCP)
+		return;
+
+	mutex_lock(&adev->enforce_isolation_mutex);
+	for (i = 0; i < AMDGPU_MAX_GFX_RINGS; ++i) {
+		if (isolation_work->xcp_id == adev->gfx.gfx_ring[i].xcp_id)
+			fences += amdgpu_fence_count_emitted(&adev->gfx.gfx_ring[i]);
+	}
+	for (i = 0; i < (AMDGPU_MAX_COMPUTE_RINGS * AMDGPU_MAX_GC_INSTANCES); ++i) {
+		if (isolation_work->xcp_id == adev->gfx.compute_ring[i].xcp_id)
+			fences += amdgpu_fence_count_emitted(&adev->gfx.compute_ring[i]);
+	}
+	if (fences) {
+		/* we've already had our timeslice, so let's wrap this up */
+		schedule_delayed_work(&adev->gfx.enforce_isolation[idx].work,
+				      msecs_to_jiffies(1));
+	} else {
+		/* Tell KFD to resume the runqueue */
+		WARN_ON_ONCE(!adev->gfx.userq_sch_inactive[idx]);
+		WARN_ON_ONCE(adev->gfx.userq_sch_req_count[idx]);
+
+		amdgpu_userq_start_sched_for_enforce_isolation(adev, idx);
+		if (adev->kfd.init_complete)
+			amdgpu_amdkfd_start_sched(adev, idx);
+		adev->gfx.userq_sch_inactive[idx] = false;
+	}
+	mutex_unlock(&adev->enforce_isolation_mutex);
+}
+
+/**
+ * amdgpu_gfx_enforce_isolation_wait_for_kfd - Manage KFD wait period for process isolation
  * @adev: amdgpu_device pointer
- * @state: gfx power state(1 -sGpuChangeState_D0Entry and 2 -sGpuChangeState_D3Entry)
+ * @idx: Index of the GPU partition
+ *
+ * When kernel submissions come in, the jobs are given a time slice and once
+ * that time slice is up, if there are KFD user queues active, kernel
+ * submissions are blocked until KFD has had its time slice. Once the KFD time
+ * slice is up, KFD user queues are preempted and kernel submissions are
+ * unblocked and allowed to run again.
+ */
+static void
+amdgpu_gfx_enforce_isolation_wait_for_kfd(struct amdgpu_device *adev,
+					  u32 idx)
+{
+	unsigned long cjiffies;
+	bool wait = false;
+
+	mutex_lock(&adev->enforce_isolation_mutex);
+	if (adev->enforce_isolation[idx] == AMDGPU_ENFORCE_ISOLATION_ENABLE) {
+		/* set the initial values if nothing is set */
+		if (!adev->gfx.enforce_isolation_jiffies[idx]) {
+			adev->gfx.enforce_isolation_jiffies[idx] = jiffies;
+			adev->gfx.enforce_isolation_time[idx] =	GFX_SLICE_PERIOD_MS;
+		}
+		/* Make sure KFD gets a chance to run */
+		if (amdgpu_amdkfd_compute_active(adev, idx)) {
+			cjiffies = jiffies;
+			if (time_after(cjiffies, adev->gfx.enforce_isolation_jiffies[idx])) {
+				cjiffies -= adev->gfx.enforce_isolation_jiffies[idx];
+				if ((jiffies_to_msecs(cjiffies) >= GFX_SLICE_PERIOD_MS)) {
+					/* if our time is up, let KGD work drain before scheduling more */
+					wait = true;
+					/* reset the timer period */
+					adev->gfx.enforce_isolation_time[idx] =	GFX_SLICE_PERIOD_MS;
+				} else {
+					/* set the timer period to what's left in our time slice */
+					adev->gfx.enforce_isolation_time[idx] =
+						GFX_SLICE_PERIOD_MS - jiffies_to_msecs(cjiffies);
+				}
+			} else {
+				/* if jiffies wrap around we will just wait a little longer */
+				adev->gfx.enforce_isolation_jiffies[idx] = jiffies;
+			}
+		} else {
+			/* if there is no KFD work, then set the full slice period */
+			adev->gfx.enforce_isolation_jiffies[idx] = jiffies;
+			adev->gfx.enforce_isolation_time[idx] = GFX_SLICE_PERIOD_MS;
+		}
+	}
+	mutex_unlock(&adev->enforce_isolation_mutex);
+
+	if (wait)
+		msleep(GFX_SLICE_PERIOD_MS);
+}
+
+/**
+ * amdgpu_gfx_enforce_isolation_ring_begin_use - Begin use of a ring with enforced isolation
+ * @ring: Pointer to the amdgpu_ring structure
  *
+ * Ring begin_use helper implementation for gfx which serializes access to the
+ * gfx IP between kernel submission IOCTLs and KFD user queues when isolation
+ * enforcement is enabled. The kernel submission IOCTLs and KFD user queues
+ * each get a time slice when both are active.
  */
+void amdgpu_gfx_enforce_isolation_ring_begin_use(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	u32 idx;
+	bool sched_work = false;
+
+	if (!adev->gfx.enable_cleaner_shader)
+		return;
+
+	if (ring->xcp_id == AMDGPU_XCP_NO_PARTITION)
+		idx = 0;
+	else
+		idx = ring->xcp_id;
 
-void amdgpu_gfx_state_change_set(struct amdgpu_device *adev, enum gfx_change_state state)
+	if (idx >= MAX_XCP)
+		return;
+
+	/* Don't submit more work until KFD has had some time */
+	amdgpu_gfx_enforce_isolation_wait_for_kfd(adev, idx);
+
+	mutex_lock(&adev->enforce_isolation_mutex);
+	if (adev->enforce_isolation[idx] == AMDGPU_ENFORCE_ISOLATION_ENABLE) {
+		if (adev->kfd.init_complete)
+			sched_work = true;
+	}
+	mutex_unlock(&adev->enforce_isolation_mutex);
+
+	if (sched_work)
+		amdgpu_gfx_kfd_sch_ctrl(adev, idx, false);
+}
+
+/**
+ * amdgpu_gfx_enforce_isolation_ring_end_use - End use of a ring with enforced isolation
+ * @ring: Pointer to the amdgpu_ring structure
+ *
+ * Ring end_use helper implementation for gfx which serializes access to the
+ * gfx IP between kernel submission IOCTLs and KFD user queues when isolation
+ * enforcement is enabled. The kernel submission IOCTLs and KFD user queues
+ * each get a time slice when both are active.
+ */
+void amdgpu_gfx_enforce_isolation_ring_end_use(struct amdgpu_ring *ring)
 {
-	mutex_lock(&adev->pm.mutex);
-	if (adev->powerplay.pp_funcs &&
-	    adev->powerplay.pp_funcs->gfx_state_change_set)
-		((adev)->powerplay.pp_funcs->gfx_state_change_set(
-			(adev)->powerplay.pp_handle, state));
-	mutex_unlock(&adev->pm.mutex);
+	struct amdgpu_device *adev = ring->adev;
+	u32 idx;
+	bool sched_work = false;
+
+	if (!adev->gfx.enable_cleaner_shader)
+		return;
+
+	if (ring->xcp_id == AMDGPU_XCP_NO_PARTITION)
+		idx = 0;
+	else
+		idx = ring->xcp_id;
+
+	if (idx >= MAX_XCP)
+		return;
+
+	mutex_lock(&adev->enforce_isolation_mutex);
+	if (adev->enforce_isolation[idx] == AMDGPU_ENFORCE_ISOLATION_ENABLE) {
+		if (adev->kfd.init_complete)
+			sched_work = true;
+	}
+	mutex_unlock(&adev->enforce_isolation_mutex);
+
+	if (sched_work)
+		amdgpu_gfx_kfd_sch_ctrl(adev, idx, true);
 }
+
+void amdgpu_gfx_profile_idle_work_handler(struct work_struct *work)
+{
+	struct amdgpu_device *adev =
+		container_of(work, struct amdgpu_device, gfx.idle_work.work);
+	enum PP_SMC_POWER_PROFILE profile;
+	u32 i, fences = 0;
+	int r;
+
+	if (adev->gfx.num_gfx_rings)
+		profile = PP_SMC_POWER_PROFILE_FULLSCREEN3D;
+	else
+		profile = PP_SMC_POWER_PROFILE_COMPUTE;
+
+	for (i = 0; i < AMDGPU_MAX_GFX_RINGS; ++i)
+		fences += amdgpu_fence_count_emitted(&adev->gfx.gfx_ring[i]);
+	for (i = 0; i < (AMDGPU_MAX_COMPUTE_RINGS * AMDGPU_MAX_GC_INSTANCES); ++i)
+		fences += amdgpu_fence_count_emitted(&adev->gfx.compute_ring[i]);
+	if (!fences && !atomic_read(&adev->gfx.total_submission_cnt)) {
+		mutex_lock(&adev->gfx.workload_profile_mutex);
+		if (adev->gfx.workload_profile_active) {
+			r = amdgpu_dpm_switch_power_profile(adev, profile, false);
+			if (r)
+				dev_warn(adev->dev, "(%d) failed to disable %s power profile mode\n", r,
+					 profile == PP_SMC_POWER_PROFILE_FULLSCREEN3D ?
+					 "fullscreen 3D" : "compute");
+			adev->gfx.workload_profile_active = false;
+		}
+		mutex_unlock(&adev->gfx.workload_profile_mutex);
+	} else {
+		schedule_delayed_work(&adev->gfx.idle_work, GFX_PROFILE_IDLE_TIMEOUT);
+	}
+}
+
+void amdgpu_gfx_profile_ring_begin_use(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	enum PP_SMC_POWER_PROFILE profile;
+	int r;
+
+	if (amdgpu_dpm_is_overdrive_enabled(adev))
+		return;
+
+	if (adev->gfx.num_gfx_rings)
+		profile = PP_SMC_POWER_PROFILE_FULLSCREEN3D;
+	else
+		profile = PP_SMC_POWER_PROFILE_COMPUTE;
+
+	atomic_inc(&adev->gfx.total_submission_cnt);
+
+	cancel_delayed_work_sync(&adev->gfx.idle_work);
+
+	/* We can safely return early here because we've cancelled the
+	 * the delayed work so there is no one else to set it to false
+	 * and we don't care if someone else sets it to true.
+	 */
+	if (adev->gfx.workload_profile_active)
+		return;
+
+	mutex_lock(&adev->gfx.workload_profile_mutex);
+	if (!adev->gfx.workload_profile_active) {
+		r = amdgpu_dpm_switch_power_profile(adev, profile, true);
+		if (r)
+			dev_warn(adev->dev, "(%d) failed to disable %s power profile mode\n", r,
+				 profile == PP_SMC_POWER_PROFILE_FULLSCREEN3D ?
+				 "fullscreen 3D" : "compute");
+		adev->gfx.workload_profile_active = true;
+	}
+	mutex_unlock(&adev->gfx.workload_profile_mutex);
+}
+
+void amdgpu_gfx_profile_ring_end_use(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	if (amdgpu_dpm_is_overdrive_enabled(adev))
+		return;
+
+	atomic_dec(&ring->adev->gfx.total_submission_cnt);
+
+	schedule_delayed_work(&ring->adev->gfx.idle_work, GFX_PROFILE_IDLE_TIMEOUT);
+}
+
+/**
+ * amdgpu_gfx_csb_preamble_start - Set CSB preamble start
+ *
+ * @buffer: This is an output variable that gets the PACKET3 preamble setup.
+ *
+ * Return:
+ * return the latest index.
+ */
+u32 amdgpu_gfx_csb_preamble_start(u32 *buffer)
+{
+	u32 count = 0;
+
+	buffer[count++] = cpu_to_le32(PACKET3(PACKET3_PREAMBLE_CNTL, 0));
+	buffer[count++] = cpu_to_le32(PACKET3_PREAMBLE_BEGIN_CLEAR_STATE);
+
+	buffer[count++] = cpu_to_le32(PACKET3(PACKET3_CONTEXT_CONTROL, 1));
+	buffer[count++] = cpu_to_le32(0x80000000);
+	buffer[count++] = cpu_to_le32(0x80000000);
+
+	return count;
+}
+
+/**
+ * amdgpu_gfx_csb_data_parser - Parser CS data
+ *
+ * @adev: amdgpu_device pointer used to get the CS data and other gfx info.
+ * @buffer: This is an output variable that gets the PACKET3 preamble end.
+ * @count: Index to start set the preemble end.
+ *
+ * Return:
+ * return the latest index.
+ */
+u32 amdgpu_gfx_csb_data_parser(struct amdgpu_device *adev, u32 *buffer, u32 count)
+{
+	const struct cs_section_def *sect = NULL;
+	const struct cs_extent_def *ext = NULL;
+	u32 i;
+
+	for (sect = adev->gfx.rlc.cs_data; sect->section != NULL; ++sect) {
+		for (ext = sect->section; ext->extent != NULL; ++ext) {
+			if (sect->id == SECT_CONTEXT) {
+				buffer[count++] = cpu_to_le32(PACKET3(PACKET3_SET_CONTEXT_REG, ext->reg_count));
+				buffer[count++] = cpu_to_le32(ext->reg_index - PACKET3_SET_CONTEXT_REG_START);
+
+				for (i = 0; i < ext->reg_count; i++)
+					buffer[count++] = cpu_to_le32(ext->extent[i]);
+			}
+		}
+	}
+
+	return count;
+}
+
+/**
+ * amdgpu_gfx_csb_preamble_end - Set CSB preamble end
+ *
+ * @buffer: This is an output variable that gets the PACKET3 preamble end.
+ * @count: Index to start set the preemble end.
+ */
+void amdgpu_gfx_csb_preamble_end(u32 *buffer, u32 count)
+{
+	buffer[count++] = cpu_to_le32(PACKET3(PACKET3_PREAMBLE_CNTL, 0));
+	buffer[count++] = cpu_to_le32(PACKET3_PREAMBLE_END_CLEAR_STATE);
+
+	buffer[count++] = cpu_to_le32(PACKET3(PACKET3_CLEAR_STATE, 0));
+	buffer[count++] = cpu_to_le32(0);
+}
+
+/*
+ * debugfs for to enable/disable gfx job submission to specific core.
+ */
+#if defined(CONFIG_DEBUG_FS)
+static int amdgpu_debugfs_gfx_sched_mask_set(void *data, u64 val)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)data;
+	u32 i;
+	u64 mask = 0;
+	struct amdgpu_ring *ring;
+
+	if (!adev)
+		return -ENODEV;
+
+	mask = (1ULL << adev->gfx.num_gfx_rings) - 1;
+	if ((val & mask) == 0)
+		return -EINVAL;
+
+	for (i = 0; i < adev->gfx.num_gfx_rings; ++i) {
+		ring = &adev->gfx.gfx_ring[i];
+		if (val & (1 << i))
+			ring->sched.ready = true;
+		else
+			ring->sched.ready = false;
+	}
+	/* publish sched.ready flag update effective immediately across smp */
+	smp_rmb();
+	return 0;
+}
+
+static int amdgpu_debugfs_gfx_sched_mask_get(void *data, u64 *val)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)data;
+	u32 i;
+	u64 mask = 0;
+	struct amdgpu_ring *ring;
+
+	if (!adev)
+		return -ENODEV;
+	for (i = 0; i < adev->gfx.num_gfx_rings; ++i) {
+		ring = &adev->gfx.gfx_ring[i];
+		if (ring->sched.ready)
+			mask |= 1ULL << i;
+	}
+
+	*val = mask;
+	return 0;
+}
+
+DEFINE_DEBUGFS_ATTRIBUTE(amdgpu_debugfs_gfx_sched_mask_fops,
+			 amdgpu_debugfs_gfx_sched_mask_get,
+			 amdgpu_debugfs_gfx_sched_mask_set, "%llx\n");
+
+#endif
+
+void amdgpu_debugfs_gfx_sched_mask_init(struct amdgpu_device *adev)
+{
+#if defined(CONFIG_DEBUG_FS)
+	struct drm_minor *minor = adev_to_drm(adev)->primary;
+	struct dentry *root = minor->debugfs_root;
+	char name[32];
+
+	if (!(adev->gfx.num_gfx_rings > 1))
+		return;
+	sprintf(name, "amdgpu_gfx_sched_mask");
+	debugfs_create_file(name, 0600, root, adev,
+			    &amdgpu_debugfs_gfx_sched_mask_fops);
+#endif
+}
+
+/*
+ * debugfs for to enable/disable compute job submission to specific core.
+ */
+#if defined(CONFIG_DEBUG_FS)
+static int amdgpu_debugfs_compute_sched_mask_set(void *data, u64 val)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)data;
+	u32 i;
+	u64 mask = 0;
+	struct amdgpu_ring *ring;
+
+	if (!adev)
+		return -ENODEV;
+
+	mask = (1ULL << adev->gfx.num_compute_rings) - 1;
+	if ((val & mask) == 0)
+		return -EINVAL;
+
+	for (i = 0; i < adev->gfx.num_compute_rings; ++i) {
+		ring = &adev->gfx.compute_ring[i];
+		if (val & (1 << i))
+			ring->sched.ready = true;
+		else
+			ring->sched.ready = false;
+	}
+
+	/* publish sched.ready flag update effective immediately across smp */
+	smp_rmb();
+	return 0;
+}
+
+static int amdgpu_debugfs_compute_sched_mask_get(void *data, u64 *val)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)data;
+	u32 i;
+	u64 mask = 0;
+	struct amdgpu_ring *ring;
+
+	if (!adev)
+		return -ENODEV;
+	for (i = 0; i < adev->gfx.num_compute_rings; ++i) {
+		ring = &adev->gfx.compute_ring[i];
+		if (ring->sched.ready)
+			mask |= 1ULL << i;
+	}
+
+	*val = mask;
+	return 0;
+}
+
+DEFINE_DEBUGFS_ATTRIBUTE(amdgpu_debugfs_compute_sched_mask_fops,
+			 amdgpu_debugfs_compute_sched_mask_get,
+			 amdgpu_debugfs_compute_sched_mask_set, "%llx\n");
+
+#endif
+
+void amdgpu_debugfs_compute_sched_mask_init(struct amdgpu_device *adev)
+{
+#if defined(CONFIG_DEBUG_FS)
+	struct drm_minor *minor = adev_to_drm(adev)->primary;
+	struct dentry *root = minor->debugfs_root;
+	char name[32];
+
+	if (!(adev->gfx.num_compute_rings > 1))
+		return;
+	sprintf(name, "amdgpu_compute_sched_mask");
+	debugfs_create_file(name, 0600, root, adev,
+			    &amdgpu_debugfs_compute_sched_mask_fops);
+#endif
+}
+
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h
index f851196c83a5..efd61a1ccc66 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h
@@ -30,7 +30,11 @@
 #include "clearstate_defs.h"
 #include "amdgpu_ring.h"
 #include "amdgpu_rlc.h"
+#include "amdgpu_imu.h"
 #include "soc15.h"
+#include "amdgpu_ras.h"
+#include "amdgpu_ring_mux.h"
+#include "amdgpu_xcp.h"
 
 /* GFX current status */
 #define AMDGPU_GFX_NORMAL_MODE			0x00000000L
@@ -39,33 +43,77 @@
 #define AMDGPU_GFX_CG_DISABLED_MODE		0x00000004L
 #define AMDGPU_GFX_LBPW_DISABLED_MODE		0x00000008L
 
-#define AMDGPU_MAX_GFX_QUEUES KGD_MAX_QUEUES
-#define AMDGPU_MAX_COMPUTE_QUEUES KGD_MAX_QUEUES
+#define AMDGPU_MAX_GC_INSTANCES		8
+#define AMDGPU_MAX_QUEUES		128
+
+#define AMDGPU_MAX_GFX_QUEUES AMDGPU_MAX_QUEUES
+#define AMDGPU_MAX_COMPUTE_QUEUES AMDGPU_MAX_QUEUES
 
 enum amdgpu_gfx_pipe_priority {
 	AMDGPU_GFX_PIPE_PRIO_NORMAL = AMDGPU_RING_PRIO_1,
 	AMDGPU_GFX_PIPE_PRIO_HIGH = AMDGPU_RING_PRIO_2
 };
 
-/* Argument for PPSMC_MSG_GpuChangeState */
-enum gfx_change_state {
-	sGpuChangeState_D0Entry = 1,
-	sGpuChangeState_D3Entry,
-};
-
 #define AMDGPU_GFX_QUEUE_PRIORITY_MINIMUM  0
 #define AMDGPU_GFX_QUEUE_PRIORITY_MAXIMUM  15
 
+/* 1 second timeout */
+#define GFX_PROFILE_IDLE_TIMEOUT	msecs_to_jiffies(1000)
+
+enum amdgpu_gfx_partition {
+	AMDGPU_SPX_PARTITION_MODE = 0,
+	AMDGPU_DPX_PARTITION_MODE = 1,
+	AMDGPU_TPX_PARTITION_MODE = 2,
+	AMDGPU_QPX_PARTITION_MODE = 3,
+	AMDGPU_CPX_PARTITION_MODE = 4,
+	AMDGPU_UNKNOWN_COMPUTE_PARTITION_MODE = -1,
+	/* Automatically choose the right mode */
+	AMDGPU_AUTO_COMPUTE_PARTITION_MODE = -2,
+};
+
+#define NUM_XCC(x) hweight16(x)
+
+enum amdgpu_gfx_ras_mem_id_type {
+	AMDGPU_GFX_CP_MEM = 0,
+	AMDGPU_GFX_GCEA_MEM,
+	AMDGPU_GFX_GC_CANE_MEM,
+	AMDGPU_GFX_GCUTCL2_MEM,
+	AMDGPU_GFX_GDS_MEM,
+	AMDGPU_GFX_LDS_MEM,
+	AMDGPU_GFX_RLC_MEM,
+	AMDGPU_GFX_SP_MEM,
+	AMDGPU_GFX_SPI_MEM,
+	AMDGPU_GFX_SQC_MEM,
+	AMDGPU_GFX_SQ_MEM,
+	AMDGPU_GFX_TA_MEM,
+	AMDGPU_GFX_TCC_MEM,
+	AMDGPU_GFX_TCA_MEM,
+	AMDGPU_GFX_TCI_MEM,
+	AMDGPU_GFX_TCP_MEM,
+	AMDGPU_GFX_TD_MEM,
+	AMDGPU_GFX_TCX_MEM,
+	AMDGPU_GFX_ATC_L2_MEM,
+	AMDGPU_GFX_UTCL2_MEM,
+	AMDGPU_GFX_VML2_MEM,
+	AMDGPU_GFX_VML2_WALKER_MEM,
+	AMDGPU_GFX_MEM_TYPE_NUM
+};
+
 struct amdgpu_mec {
 	struct amdgpu_bo	*hpd_eop_obj;
 	u64			hpd_eop_gpu_addr;
 	struct amdgpu_bo	*mec_fw_obj;
 	u64			mec_fw_gpu_addr;
+	struct amdgpu_bo	*mec_fw_data_obj;
+	u64			mec_fw_data_gpu_addr;
+
 	u32 num_mec;
 	u32 num_pipe_per_mec;
 	u32 num_queue_per_pipe;
-	void			*mqd_backup[AMDGPU_MAX_COMPUTE_RINGS + 1];
+	void			*mqd_backup[AMDGPU_MAX_COMPUTE_RINGS * AMDGPU_MAX_GC_INSTANCES];
+};
 
+struct amdgpu_mec_bitmap {
 	/* These are the resources for which amdgpu takes ownership */
 	DECLARE_BITMAP(queue_bitmap, AMDGPU_MAX_COMPUTE_QUEUES);
 };
@@ -94,6 +142,10 @@ struct kiq_pm4_funcs {
 	void (*kiq_invalidate_tlbs)(struct amdgpu_ring *kiq_ring,
 				uint16_t pasid, uint32_t flush_type,
 				bool all_hub);
+	void (*kiq_reset_hw_queue)(struct amdgpu_ring *kiq_ring,
+				   uint32_t queue_type, uint32_t me_id,
+				   uint32_t pipe_id, uint32_t queue_id,
+				   uint32_t xcc_id, uint32_t vmid);
 	/* Packet sizes */
 	int set_resources_size;
 	int map_queues_size;
@@ -109,15 +161,7 @@ struct amdgpu_kiq {
 	struct amdgpu_ring	ring;
 	struct amdgpu_irq_src	irq;
 	const struct kiq_pm4_funcs *pmf;
-};
-
-/*
- * GPU scratch registers structures, functions & helpers
- */
-struct amdgpu_scratch {
-	unsigned		num_reg;
-	uint32_t                reg_base;
-	uint32_t		free_mask;
+	void			*mqd_backup;
 };
 
 /*
@@ -126,10 +170,46 @@ struct amdgpu_scratch {
 #define AMDGPU_GFX_MAX_SE 4
 #define AMDGPU_GFX_MAX_SH_PER_SE 2
 
+/**
+ * amdgpu_rb_config - Configure a single Render Backend (RB)
+ *
+ * Bad RBs are fused off and there is a harvest register the driver reads to
+ * determine which RB(s) are fused off so that the driver can configure the
+ * hardware state so that nothing gets sent to them. There are also user
+ * harvest registers that the driver can program to disable additional RBs,
+ * etc., for testing purposes.
+ */
 struct amdgpu_rb_config {
+	/**
+	 * @rb_backend_disable:
+	 *
+	 * The value captured from register RB_BACKEND_DISABLE indicates if the
+	 * RB backend is disabled or not.
+	 */
 	uint32_t rb_backend_disable;
+
+	/**
+	 * @user_rb_backend_disable:
+	 *
+	 * The value captured from register USER_RB_BACKEND_DISABLE indicates
+	 * if the User RB backend is disabled or not.
+	 */
 	uint32_t user_rb_backend_disable;
+
+	/**
+	 * @raster_config:
+	 *
+	 * To set up all of the states, it is necessary to have two registers
+	 * to keep all of the states. This field holds the first register.
+	 */
 	uint32_t raster_config;
+
+	/**
+	 * @raster_config_1:
+	 *
+	 * To set up all of the states, it is necessary to have two registers
+	 * to keep all of the states. This field holds the second register.
+	 */
 	uint32_t raster_config_1;
 };
 
@@ -177,6 +257,13 @@ struct amdgpu_gfx_config {
 	uint32_t macrotile_mode_array[16];
 
 	struct gb_addr_config gb_addr_config_fields;
+
+	/**
+	 * @rb_config:
+	 *
+	 * Matrix that keeps all the Render Backend (color and depth buffer
+	 * handling) configuration on the 3D engine.
+	 */
 	struct amdgpu_rb_config rb_config[AMDGPU_GFX_MAX_SE][AMDGPU_GFX_MAX_SH_PER_SE];
 
 	/* gfx configure feature */
@@ -187,7 +274,29 @@ struct amdgpu_gfx_config {
 	uint32_t num_sc_per_sh;
 	uint32_t num_packer_per_sc;
 	uint32_t pa_sc_tile_steering_override;
+	/* Whether texture coordinate truncation is conformant. */
+	bool ta_cntl2_truncate_coord_mode;
 	uint64_t tcc_disabled_mask;
+	uint32_t gc_num_tcp_per_sa;
+	uint32_t gc_num_sdp_interface;
+	uint32_t gc_num_tcps;
+	uint32_t gc_num_tcp_per_wpg;
+	uint32_t gc_tcp_l1_size;
+	uint32_t gc_num_sqc_per_wgp;
+	uint32_t gc_l1_instruction_cache_size_per_sqc;
+	uint32_t gc_l1_data_cache_size_per_sqc;
+	uint32_t gc_gl1c_per_sa;
+	uint32_t gc_gl1c_size_per_instance;
+	uint32_t gc_gl2c_per_gpu;
+	uint32_t gc_tcp_size_per_cu;
+	uint32_t gc_num_cu_per_sqc;
+	uint32_t gc_tcc_size;
+	uint32_t gc_tcp_cache_line_size;
+	uint32_t gc_instruction_cache_size_per_sqc;
+	uint32_t gc_instruction_cache_line_size;
+	uint32_t gc_scalar_data_cache_size_per_sqc;
+	uint32_t gc_scalar_data_cache_line_size;
+	uint32_t gc_tcc_cache_line_size;
 };
 
 struct amdgpu_cu_info {
@@ -201,39 +310,52 @@ struct amdgpu_cu_info {
 	uint32_t number;
 	uint32_t ao_cu_mask;
 	uint32_t ao_cu_bitmap[4][4];
-	uint32_t bitmap[4][4];
+	uint32_t bitmap[AMDGPU_MAX_GC_INSTANCES][4][4];
 };
 
-struct amdgpu_gfx_ras_funcs {
-	int (*ras_late_init)(struct amdgpu_device *adev);
-	void (*ras_fini)(struct amdgpu_device *adev);
-	int (*ras_error_inject)(struct amdgpu_device *adev,
-				void *inject_if);
-	int (*query_ras_error_count)(struct amdgpu_device *adev,
-				     void *ras_error_status);
-	void (*reset_ras_error_count)(struct amdgpu_device *adev);
-	void (*query_ras_error_status)(struct amdgpu_device *adev);
-	void (*reset_ras_error_status)(struct amdgpu_device *adev);
+struct amdgpu_gfx_ras {
+	struct amdgpu_ras_block_object  ras_block;
 	void (*enable_watchdog_timer)(struct amdgpu_device *adev);
+	int (*rlc_gc_fed_irq)(struct amdgpu_device *adev,
+				struct amdgpu_irq_src *source,
+				struct amdgpu_iv_entry *entry);
+	int (*poison_consumption_handler)(struct amdgpu_device *adev,
+						struct amdgpu_iv_entry *entry);
+};
+
+struct amdgpu_gfx_shadow_info {
+	u32 shadow_size;
+	u32 shadow_alignment;
+	u32 csa_size;
+	u32 csa_alignment;
 };
 
 struct amdgpu_gfx_funcs {
 	/* get the gpu clock counter */
 	uint64_t (*get_gpu_clock_counter)(struct amdgpu_device *adev);
 	void (*select_se_sh)(struct amdgpu_device *adev, u32 se_num,
-			     u32 sh_num, u32 instance);
-	void (*read_wave_data)(struct amdgpu_device *adev, uint32_t simd,
+			     u32 sh_num, u32 instance, int xcc_id);
+	void (*read_wave_data)(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd,
 			       uint32_t wave, uint32_t *dst, int *no_fields);
-	void (*read_wave_vgprs)(struct amdgpu_device *adev, uint32_t simd,
+	void (*read_wave_vgprs)(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd,
 				uint32_t wave, uint32_t thread, uint32_t start,
 				uint32_t size, uint32_t *dst);
-	void (*read_wave_sgprs)(struct amdgpu_device *adev, uint32_t simd,
+	void (*read_wave_sgprs)(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd,
 				uint32_t wave, uint32_t start, uint32_t size,
 				uint32_t *dst);
 	void (*select_me_pipe_q)(struct amdgpu_device *adev, u32 me, u32 pipe,
-				 u32 queue, u32 vmid);
+				 u32 queue, u32 vmid, u32 xcc_id);
 	void (*init_spm_golden)(struct amdgpu_device *adev);
 	void (*update_perfmon_mgcg)(struct amdgpu_device *adev, bool enable);
+	int (*get_gfx_shadow_info)(struct amdgpu_device *adev,
+				   struct amdgpu_gfx_shadow_info *shadow_info,
+				   bool skip_check);
+	enum amdgpu_gfx_partition
+			(*query_partition_mode)(struct amdgpu_device *adev);
+	int (*switch_partition_mode)(struct amdgpu_device *adev,
+				     int num_xccs_per_xcp);
+	int (*ih_node_to_logical_xcc)(struct amdgpu_device *adev, int ih_node);
+	int (*get_xccs_per_xcp)(struct amdgpu_device *adev);
 };
 
 struct sq_work {
@@ -245,6 +367,10 @@ struct amdgpu_pfp {
 	struct amdgpu_bo		*pfp_fw_obj;
 	uint64_t			pfp_fw_gpu_addr;
 	uint32_t			*pfp_fw_ptr;
+
+	struct amdgpu_bo		*pfp_fw_data_obj;
+	uint64_t			pfp_fw_data_gpu_addr;
+	uint32_t			*pfp_fw_data_ptr;
 };
 
 struct amdgpu_ce {
@@ -257,6 +383,11 @@ struct amdgpu_me {
 	struct amdgpu_bo		*me_fw_obj;
 	uint64_t			me_fw_gpu_addr;
 	uint32_t			*me_fw_ptr;
+
+	struct amdgpu_bo		*me_fw_data_obj;
+	uint64_t			me_fw_data_gpu_addr;
+	uint32_t			*me_fw_data_ptr;
+
 	uint32_t			num_me;
 	uint32_t			num_pipe_per_me;
 	uint32_t			num_queue_per_pipe;
@@ -266,6 +397,12 @@ struct amdgpu_me {
 	DECLARE_BITMAP(queue_bitmap, AMDGPU_MAX_GFX_QUEUES);
 };
 
+struct amdgpu_isolation_work {
+	struct amdgpu_device		*adev;
+	u32				xcp_id;
+	struct delayed_work		work;
+};
+
 struct amdgpu_gfx {
 	struct mutex			gpu_clock_mutex;
 	struct amdgpu_gfx_config	config;
@@ -274,8 +411,10 @@ struct amdgpu_gfx {
 	struct amdgpu_ce		ce;
 	struct amdgpu_me		me;
 	struct amdgpu_mec		mec;
-	struct amdgpu_kiq		kiq;
-	struct amdgpu_scratch		scratch;
+	struct amdgpu_mec_bitmap	mec_bitmap[AMDGPU_MAX_GC_INSTANCES];
+	struct amdgpu_kiq		kiq[AMDGPU_MAX_GC_INSTANCES];
+	struct amdgpu_imu		imu;
+	bool				rs64_enable; /* firmware format */
 	const struct firmware		*me_fw;	/* ME firmware */
 	uint32_t			me_fw_version;
 	const struct firmware		*pfp_fw; /* PFP firmware */
@@ -288,6 +427,8 @@ struct amdgpu_gfx {
 	uint32_t			mec_fw_version;
 	const struct firmware		*mec2_fw; /* MEC2 firmware */
 	uint32_t			mec2_fw_version;
+	const struct firmware		*imu_fw; /* IMU firmware */
+	uint32_t			imu_fw_version;
 	uint32_t			me_feature_version;
 	uint32_t			ce_feature_version;
 	uint32_t			pfp_feature_version;
@@ -298,6 +439,10 @@ struct amdgpu_gfx {
 	uint32_t			rlc_srlg_feature_version;
 	uint32_t			rlc_srls_fw_version;
 	uint32_t			rlc_srls_feature_version;
+	uint32_t			rlcp_ucode_version;
+	uint32_t			rlcp_ucode_feature_version;
+	uint32_t			rlcv_ucode_version;
+	uint32_t			rlcv_ucode_feature_version;
 	uint32_t			mec_feature_version;
 	uint32_t			mec2_feature_version;
 	bool				mec_fw_write_wait;
@@ -305,13 +450,15 @@ struct amdgpu_gfx {
 	bool				cp_fw_write_wait;
 	struct amdgpu_ring		gfx_ring[AMDGPU_MAX_GFX_RINGS];
 	unsigned			num_gfx_rings;
-	struct amdgpu_ring		compute_ring[AMDGPU_MAX_COMPUTE_RINGS];
+	struct amdgpu_ring		compute_ring[AMDGPU_MAX_COMPUTE_RINGS * AMDGPU_MAX_GC_INSTANCES];
 	unsigned			num_compute_rings;
 	struct amdgpu_irq_src		eop_irq;
 	struct amdgpu_irq_src		priv_reg_irq;
 	struct amdgpu_irq_src		priv_inst_irq;
+	struct amdgpu_irq_src		bad_op_irq;
 	struct amdgpu_irq_src		cp_ecc_error_irq;
 	struct amdgpu_irq_src		sq_irq;
+	struct amdgpu_irq_src		rlc_gc_fed_irq;
 	struct sq_work			sq_work;
 
 	/* gfx status */
@@ -324,26 +471,86 @@ struct amdgpu_gfx {
 	/* reset mask */
 	uint32_t                        grbm_soft_reset;
 	uint32_t                        srbm_soft_reset;
+	uint32_t 			gfx_supported_reset;
+	uint32_t 			compute_supported_reset;
 
 	/* gfx off */
-	bool                            gfx_off_state; /* true: enabled, false: disabled */
-	struct mutex                    gfx_off_mutex;
-	uint32_t                        gfx_off_req_count; /* default 1, enable gfx off: dec 1, disable gfx off: add 1 */
-	struct delayed_work             gfx_off_delay_work;
+	bool                            gfx_off_state;      /* true: enabled, false: disabled */
+	struct mutex                    gfx_off_mutex;      /* mutex to change gfxoff state */
+	uint32_t                        gfx_off_req_count;  /* default 1, enable gfx off: dec 1, disable gfx off: add 1 */
+	struct delayed_work             gfx_off_delay_work; /* async work to set gfx block off */
+	uint32_t                        gfx_off_residency;  /* last logged residency */
+	uint64_t                        gfx_off_entrycount; /* count of times GPU has get into GFXOFF state */
 
 	/* pipe reservation */
 	struct mutex			pipe_reserve_mutex;
 	DECLARE_BITMAP			(pipe_reserve_bitmap, AMDGPU_MAX_COMPUTE_QUEUES);
 
 	/*ras */
-	struct ras_common_if			*ras_if;
-	const struct amdgpu_gfx_ras_funcs	*ras_funcs;
+	struct ras_common_if		*ras_if;
+	struct amdgpu_gfx_ras		*ras;
+
+	bool				is_poweron;
+
+	struct amdgpu_ring		sw_gfx_ring[AMDGPU_MAX_SW_GFX_RINGS];
+	struct amdgpu_ring_mux          muxer;
+
+	bool				cp_gfx_shadow; /* for gfx11 */
+
+	uint16_t 			xcc_mask;
+	uint32_t			num_xcc_per_xcp;
+	struct mutex			partition_mutex;
+	bool				mcbp; /* mid command buffer preemption */
+
+	/* IP reg dump */
+	uint32_t			*ip_dump_core;
+	uint32_t			*ip_dump_compute_queues;
+	uint32_t			*ip_dump_gfx_queues;
+
+	struct mutex			reset_sem_mutex;
+
+	/* cleaner shader */
+	struct amdgpu_bo		*cleaner_shader_obj;
+	unsigned int                    cleaner_shader_size;
+	u64				cleaner_shader_gpu_addr;
+	void				*cleaner_shader_cpu_ptr;
+	const void			*cleaner_shader_ptr;
+	bool				enable_cleaner_shader;
+	struct amdgpu_isolation_work	enforce_isolation[MAX_XCP];
+	/* Mutex for synchronizing KFD scheduler operations */
+	struct mutex                    userq_sch_mutex;
+	u64				userq_sch_req_count[MAX_XCP];
+	bool				userq_sch_inactive[MAX_XCP];
+	unsigned long			enforce_isolation_jiffies[MAX_XCP];
+	unsigned long			enforce_isolation_time[MAX_XCP];
+
+	atomic_t			total_submission_cnt;
+	struct delayed_work		idle_work;
+	bool				workload_profile_active;
+	struct mutex                    workload_profile_mutex;
+
+	bool				disable_kq;
+	bool				disable_uq;
 };
 
+struct amdgpu_gfx_ras_reg_entry {
+	struct amdgpu_ras_err_status_reg_entry reg_entry;
+	enum amdgpu_gfx_ras_mem_id_type mem_id_type;
+	uint32_t se_num;
+};
+
+struct amdgpu_gfx_ras_mem_id_entry {
+	const struct amdgpu_ras_memory_id_entry *mem_id_ent;
+	uint32_t size;
+};
+
+#define AMDGPU_GFX_MEMID_ENT(x) {(x), ARRAY_SIZE(x)},
+
 #define amdgpu_gfx_get_gpu_clock_counter(adev) (adev)->gfx.funcs->get_gpu_clock_counter((adev))
-#define amdgpu_gfx_select_se_sh(adev, se, sh, instance) (adev)->gfx.funcs->select_se_sh((adev), (se), (sh), (instance))
-#define amdgpu_gfx_select_me_pipe_q(adev, me, pipe, q, vmid) (adev)->gfx.funcs->select_me_pipe_q((adev), (me), (pipe), (q), (vmid))
+#define amdgpu_gfx_select_se_sh(adev, se, sh, instance, xcc_id) ((adev)->gfx.funcs->select_se_sh((adev), (se), (sh), (instance), (xcc_id)))
+#define amdgpu_gfx_select_me_pipe_q(adev, me, pipe, q, vmid, xcc_id) ((adev)->gfx.funcs->select_me_pipe_q((adev), (me), (pipe), (q), (vmid), (xcc_id)))
 #define amdgpu_gfx_init_spm_golden(adev) (adev)->gfx.funcs->init_spm_golden((adev))
+#define amdgpu_gfx_get_gfx_shadow_info(adev, si) ((adev)->gfx.funcs->get_gfx_shadow_info((adev), (si), false))
 
 /**
  * amdgpu_gfx_create_bitmask - create a bitmask
@@ -358,27 +565,24 @@ static inline u32 amdgpu_gfx_create_bitmask(u32 bit_width)
 	return (u32)((1ULL << bit_width) - 1);
 }
 
-int amdgpu_gfx_scratch_get(struct amdgpu_device *adev, uint32_t *reg);
-void amdgpu_gfx_scratch_free(struct amdgpu_device *adev, uint32_t reg);
-
 void amdgpu_gfx_parse_disable_cu(unsigned *mask, unsigned max_se,
 				 unsigned max_sh);
 
-int amdgpu_gfx_kiq_init_ring(struct amdgpu_device *adev,
-			     struct amdgpu_ring *ring,
-			     struct amdgpu_irq_src *irq);
+int amdgpu_gfx_kiq_init_ring(struct amdgpu_device *adev, int xcc_id);
 
 void amdgpu_gfx_kiq_free_ring(struct amdgpu_ring *ring);
 
-void amdgpu_gfx_kiq_fini(struct amdgpu_device *adev);
+void amdgpu_gfx_kiq_fini(struct amdgpu_device *adev, int xcc_id);
 int amdgpu_gfx_kiq_init(struct amdgpu_device *adev,
-			unsigned hpd_size);
+			unsigned hpd_size, int xcc_id);
 
 int amdgpu_gfx_mqd_sw_init(struct amdgpu_device *adev,
-			   unsigned mqd_size);
-void amdgpu_gfx_mqd_sw_fini(struct amdgpu_device *adev);
-int amdgpu_gfx_disable_kcq(struct amdgpu_device *adev);
-int amdgpu_gfx_enable_kcq(struct amdgpu_device *adev);
+			   unsigned mqd_size, int xcc_id);
+void amdgpu_gfx_mqd_sw_fini(struct amdgpu_device *adev, int xcc_id);
+int amdgpu_gfx_disable_kcq(struct amdgpu_device *adev, int xcc_id);
+int amdgpu_gfx_enable_kcq(struct amdgpu_device *adev, int xcc_id);
+int amdgpu_gfx_disable_kgq(struct amdgpu_device *adev, int xcc_id);
+int amdgpu_gfx_enable_kgq(struct amdgpu_device *adev, int xcc_id);
 
 void amdgpu_gfx_compute_queue_acquire(struct amdgpu_device *adev);
 void amdgpu_gfx_graphics_queue_acquire(struct amdgpu_device *adev);
@@ -387,28 +591,81 @@ int amdgpu_gfx_mec_queue_to_bit(struct amdgpu_device *adev, int mec,
 				int pipe, int queue);
 void amdgpu_queue_mask_bit_to_mec_queue(struct amdgpu_device *adev, int bit,
 				 int *mec, int *pipe, int *queue);
-bool amdgpu_gfx_is_mec_queue_enabled(struct amdgpu_device *adev, int mec,
-				     int pipe, int queue);
+bool amdgpu_gfx_is_mec_queue_enabled(struct amdgpu_device *adev, int xcc_id,
+				     int mec, int pipe, int queue);
 bool amdgpu_gfx_is_high_priority_compute_queue(struct amdgpu_device *adev,
 					       struct amdgpu_ring *ring);
-int amdgpu_gfx_me_queue_to_bit(struct amdgpu_device *adev, int me,
-			       int pipe, int queue);
-void amdgpu_gfx_bit_to_me_queue(struct amdgpu_device *adev, int bit,
-				int *me, int *pipe, int *queue);
+bool amdgpu_gfx_is_high_priority_graphics_queue(struct amdgpu_device *adev,
+						struct amdgpu_ring *ring);
 bool amdgpu_gfx_is_me_queue_enabled(struct amdgpu_device *adev, int me,
 				    int pipe, int queue);
 void amdgpu_gfx_off_ctrl(struct amdgpu_device *adev, bool enable);
+void amdgpu_gfx_off_ctrl_immediate(struct amdgpu_device *adev, bool enable);
 int amdgpu_get_gfx_off_status(struct amdgpu_device *adev, uint32_t *value);
-int amdgpu_gfx_ras_late_init(struct amdgpu_device *adev);
+int amdgpu_gfx_ras_late_init(struct amdgpu_device *adev, struct ras_common_if *ras_block);
 void amdgpu_gfx_ras_fini(struct amdgpu_device *adev);
+int amdgpu_get_gfx_off_entrycount(struct amdgpu_device *adev, u64 *value);
+int amdgpu_get_gfx_off_residency(struct amdgpu_device *adev, u32 *residency);
+int amdgpu_set_gfx_off_residency(struct amdgpu_device *adev, bool value);
 int amdgpu_gfx_process_ras_data_cb(struct amdgpu_device *adev,
 		void *err_data,
 		struct amdgpu_iv_entry *entry);
 int amdgpu_gfx_cp_ecc_error_irq(struct amdgpu_device *adev,
 				  struct amdgpu_irq_src *source,
 				  struct amdgpu_iv_entry *entry);
-uint32_t amdgpu_kiq_rreg(struct amdgpu_device *adev, uint32_t reg);
-void amdgpu_kiq_wreg(struct amdgpu_device *adev, uint32_t reg, uint32_t v);
+uint32_t amdgpu_kiq_rreg(struct amdgpu_device *adev, uint32_t reg, uint32_t xcc_id);
+void amdgpu_kiq_wreg(struct amdgpu_device *adev, uint32_t reg, uint32_t v, uint32_t xcc_id);
+int amdgpu_kiq_hdp_flush(struct amdgpu_device *adev);
 int amdgpu_gfx_get_num_kcq(struct amdgpu_device *adev);
-void amdgpu_gfx_state_change_set(struct amdgpu_device *adev, enum gfx_change_state state);
+void amdgpu_gfx_cp_init_microcode(struct amdgpu_device *adev, uint32_t ucode_id);
+
+int amdgpu_gfx_ras_sw_init(struct amdgpu_device *adev);
+int amdgpu_gfx_poison_consumption_handler(struct amdgpu_device *adev,
+						struct amdgpu_iv_entry *entry);
+
+bool amdgpu_gfx_is_master_xcc(struct amdgpu_device *adev, int xcc_id);
+int amdgpu_gfx_sysfs_init(struct amdgpu_device *adev);
+void amdgpu_gfx_sysfs_fini(struct amdgpu_device *adev);
+void amdgpu_gfx_ras_error_func(struct amdgpu_device *adev,
+		void *ras_error_status,
+		void (*func)(struct amdgpu_device *adev, void *ras_error_status,
+				int xcc_id));
+int amdgpu_gfx_cleaner_shader_sw_init(struct amdgpu_device *adev,
+				      unsigned int cleaner_shader_size);
+void amdgpu_gfx_cleaner_shader_sw_fini(struct amdgpu_device *adev);
+void amdgpu_gfx_cleaner_shader_init(struct amdgpu_device *adev,
+				    unsigned int cleaner_shader_size,
+				    const void *cleaner_shader_ptr);
+void amdgpu_gfx_enforce_isolation_handler(struct work_struct *work);
+void amdgpu_gfx_enforce_isolation_ring_begin_use(struct amdgpu_ring *ring);
+void amdgpu_gfx_enforce_isolation_ring_end_use(struct amdgpu_ring *ring);
+
+void amdgpu_gfx_profile_idle_work_handler(struct work_struct *work);
+void amdgpu_gfx_profile_ring_begin_use(struct amdgpu_ring *ring);
+void amdgpu_gfx_profile_ring_end_use(struct amdgpu_ring *ring);
+u32 amdgpu_gfx_csb_preamble_start(u32 *buffer);
+u32 amdgpu_gfx_csb_data_parser(struct amdgpu_device *adev, u32 *buffer, u32 count);
+void amdgpu_gfx_csb_preamble_end(u32 *buffer, u32 count);
+
+void amdgpu_debugfs_gfx_sched_mask_init(struct amdgpu_device *adev);
+void amdgpu_debugfs_compute_sched_mask_init(struct amdgpu_device *adev);
+
+static inline const char *amdgpu_gfx_compute_mode_desc(int mode)
+{
+	switch (mode) {
+	case AMDGPU_SPX_PARTITION_MODE:
+		return "SPX";
+	case AMDGPU_DPX_PARTITION_MODE:
+		return "DPX";
+	case AMDGPU_TPX_PARTITION_MODE:
+		return "TPX";
+	case AMDGPU_QPX_PARTITION_MODE:
+		return "QPX";
+	case AMDGPU_CPX_PARTITION_MODE:
+		return "CPX";
+	default:
+		return "UNKNOWN";
+	}
+}
+
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfxhub.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfxhub.h
index beabab515836..c7b44aeb671b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfxhub.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfxhub.h
@@ -35,6 +35,9 @@ struct amdgpu_gfxhub_funcs {
 	void (*init)(struct amdgpu_device *adev);
 	int (*get_xgmi_info)(struct amdgpu_device *adev);
 	void (*utcl2_harvest)(struct amdgpu_device *adev);
+	void (*mode2_save_regs)(struct amdgpu_device *adev);
+	void (*mode2_restore_regs)(struct amdgpu_device *adev);
+	void (*halt)(struct amdgpu_device *adev);
 };
 
 struct amdgpu_gfxhub {
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
index 08478fce00f2..869bceb0fe2c 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
@@ -25,13 +25,25 @@
  */
 
 #include <linux/io-64-nonatomic-lo-hi.h>
+#ifdef CONFIG_X86
+#include <asm/hypervisor.h>
+#endif
 
 #include "amdgpu.h"
 #include "amdgpu_gmc.h"
 #include "amdgpu_ras.h"
+#include "amdgpu_reset.h"
 #include "amdgpu_xgmi.h"
 
 #include <drm/drm_drv.h>
+#include <drm/ttm/ttm_tt.h>
+
+static const u64 four_gb = 0x100000000ULL;
+
+bool amdgpu_gmc_is_pdb0_enabled(struct amdgpu_device *adev)
+{
+	return adev->gmc.xgmi.connected_to_cpu || amdgpu_virt_xgmi_migrate_enabled(adev);
+}
 
 /**
  * amdgpu_gmc_pdb0_alloc - allocate vram for pdb0
@@ -47,7 +59,7 @@ int amdgpu_gmc_pdb0_alloc(struct amdgpu_device *adev)
 	struct amdgpu_bo_param bp;
 	u64 vram_size = adev->gmc.xgmi.node_segment_size * adev->gmc.xgmi.num_physical_nodes;
 	uint32_t pde0_page_shift = adev->gmc.vmid0_page_table_block_size + 21;
-	uint32_t npdes = (vram_size + (1ULL << pde0_page_shift) -1) >> pde0_page_shift;
+	uint32_t npdes = (vram_size + (1ULL << pde0_page_shift) - 1) >> pde0_page_shift;
 
 	memset(&bp, 0, sizeof(bp));
 	bp.size = PAGE_ALIGN((npdes + 1) * 8);
@@ -176,6 +188,9 @@ uint64_t amdgpu_gmc_agp_addr(struct ttm_buffer_object *bo)
 {
 	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->bdev);
 
+	if (!bo->ttm)
+		return AMDGPU_BO_INVALID_OFFSET;
+
 	if (bo->ttm->num_pages != 1 || bo->ttm->caching == ttm_cached)
 		return AMDGPU_BO_INVALID_OFFSET;
 
@@ -198,13 +213,20 @@ uint64_t amdgpu_gmc_agp_addr(struct ttm_buffer_object *bo)
 void amdgpu_gmc_vram_location(struct amdgpu_device *adev, struct amdgpu_gmc *mc,
 			      u64 base)
 {
+	uint64_t vis_limit = (uint64_t)amdgpu_vis_vram_limit << 20;
 	uint64_t limit = (uint64_t)amdgpu_vram_limit << 20;
 
 	mc->vram_start = base;
 	mc->vram_end = mc->vram_start + mc->mc_vram_size - 1;
-	if (limit && limit < mc->real_vram_size)
+	if (limit < mc->real_vram_size)
 		mc->real_vram_size = limit;
 
+	if (vis_limit && vis_limit < mc->visible_vram_size)
+		mc->visible_vram_size = vis_limit;
+
+	if (mc->real_vram_size < mc->visible_vram_size)
+		mc->visible_vram_size = mc->real_vram_size;
+
 	if (mc->xgmi.num_physical_nodes == 0) {
 		mc->fb_start = mc->vram_start;
 		mc->fb_end = mc->vram_end;
@@ -236,10 +258,20 @@ void amdgpu_gmc_sysvm_location(struct amdgpu_device *adev, struct amdgpu_gmc *mc
 	u64 hive_vram_end = mc->xgmi.node_segment_size * mc->xgmi.num_physical_nodes - 1;
 	mc->vram_start = mc->xgmi.node_segment_size * mc->xgmi.physical_node_id;
 	mc->vram_end = mc->vram_start + mc->xgmi.node_segment_size - 1;
-	mc->gart_start = hive_vram_end + 1;
+	/* node_segment_size may not 4GB aligned on SRIOV, align up is needed. */
+	mc->gart_start = ALIGN(hive_vram_end + 1, four_gb);
 	mc->gart_end = mc->gart_start + mc->gart_size - 1;
-	mc->fb_start = hive_vram_start;
-	mc->fb_end = hive_vram_end;
+	if (amdgpu_virt_xgmi_migrate_enabled(adev)) {
+		/* set mc->vram_start to 0 to switch the returned GPU address of
+		 * amdgpu_bo_create_reserved() from FB aperture to GART aperture.
+		 */
+		mc->vram_start = 0;
+		mc->vram_end = mc->vram_start + mc->mc_vram_size - 1;
+		mc->visible_vram_size = min(mc->visible_vram_size, mc->real_vram_size);
+	} else {
+		mc->fb_start = hive_vram_start;
+		mc->fb_end = hive_vram_end;
+	}
 	dev_info(adev->dev, "VRAM: %lluM 0x%016llX - 0x%016llX (%lluM used)\n",
 			mc->mc_vram_size >> 20, mc->vram_start,
 			mc->vram_end, mc->real_vram_size >> 20);
@@ -252,14 +284,15 @@ void amdgpu_gmc_sysvm_location(struct amdgpu_device *adev, struct amdgpu_gmc *mc
  *
  * @adev: amdgpu device structure holding all necessary information
  * @mc: memory controller structure holding memory information
+ * @gart_placement: GART placement policy with respect to VRAM
  *
- * Function will place try to place GART before or after VRAM.
+ * Function will try to place GART before or after VRAM.
  * If GART size is bigger than space left then we ajust GART size.
  * Thus function will never fails.
  */
-void amdgpu_gmc_gart_location(struct amdgpu_device *adev, struct amdgpu_gmc *mc)
+void amdgpu_gmc_gart_location(struct amdgpu_device *adev, struct amdgpu_gmc *mc,
+			      enum amdgpu_gart_placement gart_placement)
 {
-	const uint64_t four_gb = 0x100000000ULL;
 	u64 size_af, size_bf;
 	/*To avoid the hole, limit the max mc address to AMDGPU_GMC_HOLE_START*/
 	u64 max_mc_address = min(adev->gmc.mc_mask, AMDGPU_GMC_HOLE_START - 1);
@@ -275,11 +308,22 @@ void amdgpu_gmc_gart_location(struct amdgpu_device *adev, struct amdgpu_gmc *mc)
 		mc->gart_size = max(size_bf, size_af);
 	}
 
-	if ((size_bf >= mc->gart_size && size_bf < size_af) ||
-	    (size_af < mc->gart_size))
-		mc->gart_start = 0;
-	else
+	switch (gart_placement) {
+	case AMDGPU_GART_PLACEMENT_HIGH:
 		mc->gart_start = max_mc_address - mc->gart_size + 1;
+		break;
+	case AMDGPU_GART_PLACEMENT_LOW:
+		mc->gart_start = 0;
+		break;
+	case AMDGPU_GART_PLACEMENT_BEST_FIT:
+	default:
+		if ((size_bf >= mc->gart_size && size_bf < size_af) ||
+		    (size_af < mc->gart_size))
+			mc->gart_start = 0;
+		else
+			mc->gart_start = max_mc_address - mc->gart_size + 1;
+		break;
+	}
 
 	mc->gart_start &= ~(four_gb - 1);
 	mc->gart_end = mc->gart_start + mc->gart_size - 1;
@@ -304,14 +348,6 @@ void amdgpu_gmc_agp_location(struct amdgpu_device *adev, struct amdgpu_gmc *mc)
 	const uint64_t sixteen_gb_mask = ~(sixteen_gb - 1);
 	u64 size_af, size_bf;
 
-	if (amdgpu_sriov_vf(adev)) {
-		mc->agp_start = 0xffffffffffff;
-		mc->agp_end = 0x0;
-		mc->agp_size = 0;
-
-		return;
-	}
-
 	if (mc->fb_start > mc->gart_start) {
 		size_bf = (mc->fb_start & sixteen_gb_mask) -
 			ALIGN(mc->gart_end + 1, sixteen_gb);
@@ -336,6 +372,25 @@ void amdgpu_gmc_agp_location(struct amdgpu_device *adev, struct amdgpu_gmc *mc)
 }
 
 /**
+ * amdgpu_gmc_set_agp_default - Set the default AGP aperture value.
+ * @adev: amdgpu device structure holding all necessary information
+ * @mc: memory controller structure holding memory information
+ *
+ * To disable the AGP aperture, you need to set the start to a larger
+ * value than the end.  This function sets the default value which
+ * can then be overridden using amdgpu_gmc_agp_location() if you want
+ * to enable the AGP aperture on a specific chip.
+ *
+ */
+void amdgpu_gmc_set_agp_default(struct amdgpu_device *adev,
+				struct amdgpu_gmc *mc)
+{
+	mc->agp_start = 0xffffffffffff;
+	mc->agp_end = 0;
+	mc->agp_size = 0;
+}
+
+/**
  * amdgpu_gmc_fault_key - get hask key from vm fault address and pasid
  *
  * @addr: 48 bit physical address, page aligned (36 significant bits)
@@ -350,6 +405,7 @@ static inline uint64_t amdgpu_gmc_fault_key(uint64_t addr, uint16_t pasid)
  * amdgpu_gmc_filter_faults - filter VM faults
  *
  * @adev: amdgpu device structure
+ * @ih: interrupt ring that the fault received from
  * @addr: address of the VM fault
  * @pasid: PASID of the process causing the fault
  * @timestamp: timestamp of the fault
@@ -358,7 +414,8 @@ static inline uint64_t amdgpu_gmc_fault_key(uint64_t addr, uint16_t pasid)
  * True if the fault was filtered and should not be processed further.
  * False if the fault is a new one and needs to be handled.
  */
-bool amdgpu_gmc_filter_faults(struct amdgpu_device *adev, uint64_t addr,
+bool amdgpu_gmc_filter_faults(struct amdgpu_device *adev,
+			      struct amdgpu_ih_ring *ih, uint64_t addr,
 			      uint16_t pasid, uint64_t timestamp)
 {
 	struct amdgpu_gmc *gmc = &adev->gmc;
@@ -366,6 +423,10 @@ bool amdgpu_gmc_filter_faults(struct amdgpu_device *adev, uint64_t addr,
 	struct amdgpu_gmc_fault *fault;
 	uint32_t hash;
 
+	/* Stale retry fault if timestamp goes backward */
+	if (amdgpu_ih_ts_after(timestamp, ih->processed_timestamp))
+		return true;
+
 	/* If we don't have space left in the ring buffer return immediately */
 	stamp = max(timestamp, AMDGPU_GMC_FAULT_TIMEOUT + 1) -
 		AMDGPU_GMC_FAULT_TIMEOUT;
@@ -378,8 +439,21 @@ bool amdgpu_gmc_filter_faults(struct amdgpu_device *adev, uint64_t addr,
 	while (fault->timestamp >= stamp) {
 		uint64_t tmp;
 
-		if (atomic64_read(&fault->key) == key)
-			return true;
+		if (atomic64_read(&fault->key) == key) {
+			/*
+			 * if we get a fault which is already present in
+			 * the fault_ring and the timestamp of
+			 * the fault is after the expired timestamp,
+			 * then this is a new fault that needs to be added
+			 * into the fault ring.
+			 */
+			if (fault->timestamp_expiry != 0 &&
+			    amdgpu_ih_ts_after(fault->timestamp_expiry,
+					       timestamp))
+				break;
+			else
+				return true;
+		}
 
 		tmp = fault->timestamp;
 		fault = &gmc->fault_ring[fault->next];
@@ -415,97 +489,89 @@ void amdgpu_gmc_filter_faults_remove(struct amdgpu_device *adev, uint64_t addr,
 {
 	struct amdgpu_gmc *gmc = &adev->gmc;
 	uint64_t key = amdgpu_gmc_fault_key(addr, pasid);
+	struct amdgpu_ih_ring *ih;
 	struct amdgpu_gmc_fault *fault;
+	uint32_t last_wptr;
+	uint64_t last_ts;
 	uint32_t hash;
 	uint64_t tmp;
 
+	if (adev->irq.retry_cam_enabled)
+		return;
+
+	ih = &adev->irq.ih1;
+	/* Get the WPTR of the last entry in IH ring */
+	last_wptr = amdgpu_ih_get_wptr(adev, ih);
+	/* Order wptr with ring data. */
+	rmb();
+	/* Get the timetamp of the last entry in IH ring */
+	last_ts = amdgpu_ih_decode_iv_ts(adev, ih, last_wptr, -1);
+
 	hash = hash_64(key, AMDGPU_GMC_FAULT_HASH_ORDER);
 	fault = &gmc->fault_ring[gmc->fault_hash[hash].idx];
 	do {
-		if (atomic64_cmpxchg(&fault->key, key, 0) == key)
+		if (atomic64_read(&fault->key) == key) {
+			/*
+			 * Update the timestamp when this fault
+			 * expired.
+			 */
+			fault->timestamp_expiry = last_ts;
 			break;
+		}
 
 		tmp = fault->timestamp;
 		fault = &gmc->fault_ring[fault->next];
 	} while (fault->timestamp < tmp);
 }
 
-int amdgpu_gmc_ras_late_init(struct amdgpu_device *adev)
+int amdgpu_gmc_ras_sw_init(struct amdgpu_device *adev)
 {
 	int r;
 
-	if (adev->umc.ras_funcs &&
-	    adev->umc.ras_funcs->ras_late_init) {
-		r = adev->umc.ras_funcs->ras_late_init(adev);
-		if (r)
-			return r;
-	}
+	/* umc ras block */
+	r = amdgpu_umc_ras_sw_init(adev);
+	if (r)
+		return r;
 
-	if (adev->mmhub.ras_funcs &&
-	    adev->mmhub.ras_funcs->ras_late_init) {
-		r = adev->mmhub.ras_funcs->ras_late_init(adev);
-		if (r)
-			return r;
-	}
+	/* mmhub ras block */
+	r = amdgpu_mmhub_ras_sw_init(adev);
+	if (r)
+		return r;
 
-	if (!adev->gmc.xgmi.connected_to_cpu)
-		adev->gmc.xgmi.ras_funcs = &xgmi_ras_funcs;
+	/* hdp ras block */
+	r = amdgpu_hdp_ras_sw_init(adev);
+	if (r)
+		return r;
 
-	if (adev->gmc.xgmi.ras_funcs &&
-	    adev->gmc.xgmi.ras_funcs->ras_late_init) {
-		r = adev->gmc.xgmi.ras_funcs->ras_late_init(adev);
-		if (r)
-			return r;
-	}
+	/* mca.x ras block */
+	r = amdgpu_mca_mp0_ras_sw_init(adev);
+	if (r)
+		return r;
 
-	if (adev->hdp.ras_funcs &&
-	    adev->hdp.ras_funcs->ras_late_init) {
-		r = adev->hdp.ras_funcs->ras_late_init(adev);
-		if (r)
-			return r;
-	}
+	r = amdgpu_mca_mp1_ras_sw_init(adev);
+	if (r)
+		return r;
 
-	if (adev->mca.mp0.ras_funcs &&
-	    adev->mca.mp0.ras_funcs->ras_late_init) {
-		r = adev->mca.mp0.ras_funcs->ras_late_init(adev);
-		if (r)
-			return r;
-	}
+	r = amdgpu_mca_mpio_ras_sw_init(adev);
+	if (r)
+		return r;
 
-	if (adev->mca.mp1.ras_funcs &&
-	    adev->mca.mp1.ras_funcs->ras_late_init) {
-		r = adev->mca.mp1.ras_funcs->ras_late_init(adev);
-		if (r)
-			return r;
-	}
+	/* xgmi ras block */
+	r = amdgpu_xgmi_ras_sw_init(adev);
+	if (r)
+		return r;
 
-	if (adev->mca.mpio.ras_funcs &&
-	    adev->mca.mpio.ras_funcs->ras_late_init) {
-		r = adev->mca.mpio.ras_funcs->ras_late_init(adev);
-		if (r)
-			return r;
-	}
+	return 0;
+}
 
+int amdgpu_gmc_ras_late_init(struct amdgpu_device *adev)
+{
 	return 0;
 }
 
 void amdgpu_gmc_ras_fini(struct amdgpu_device *adev)
 {
-	if (adev->umc.ras_funcs &&
-	    adev->umc.ras_funcs->ras_fini)
-		adev->umc.ras_funcs->ras_fini(adev);
-
-	if (adev->mmhub.ras_funcs &&
-	    adev->mmhub.ras_funcs->ras_fini)
-		adev->mmhub.ras_funcs->ras_fini(adev);
-
-	if (adev->gmc.xgmi.ras_funcs &&
-	    adev->gmc.xgmi.ras_funcs->ras_fini)
-		adev->gmc.xgmi.ras_funcs->ras_fini(adev);
 
-	if (adev->hdp.ras_funcs &&
-	    adev->hdp.ras_funcs->ras_fini)
-		adev->hdp.ras_funcs->ras_fini(adev);
 }
 
 	/*
@@ -515,23 +581,42 @@ void amdgpu_gmc_ras_fini(struct amdgpu_device *adev)
 	 *                    subject to change when ring number changes
 	 * Engine 17: Gart flushes
 	 */
-#define GFXHUB_FREE_VM_INV_ENGS_BITMAP		0x1FFF3
-#define MMHUB_FREE_VM_INV_ENGS_BITMAP		0x1FFF3
+#define AMDGPU_VMHUB_INV_ENG_BITMAP		0x1FFF3
 
 int amdgpu_gmc_allocate_vm_inv_eng(struct amdgpu_device *adev)
 {
 	struct amdgpu_ring *ring;
-	unsigned vm_inv_engs[AMDGPU_MAX_VMHUBS] =
-		{GFXHUB_FREE_VM_INV_ENGS_BITMAP, MMHUB_FREE_VM_INV_ENGS_BITMAP,
-		GFXHUB_FREE_VM_INV_ENGS_BITMAP};
+	unsigned vm_inv_engs[AMDGPU_MAX_VMHUBS] = {0};
 	unsigned i;
 	unsigned vmhub, inv_eng;
+	struct amdgpu_ring *shared_ring;
+
+	/* init the vm inv eng for all vmhubs */
+	for_each_set_bit(i, adev->vmhubs_mask, AMDGPU_MAX_VMHUBS) {
+		vm_inv_engs[i] = AMDGPU_VMHUB_INV_ENG_BITMAP;
+		/* reserve engine 5 for firmware */
+		if (adev->enable_mes)
+			vm_inv_engs[i] &= ~(1 << 5);
+		/* reserve engine 6 for uni mes */
+		if (adev->enable_uni_mes)
+			vm_inv_engs[i] &= ~(1 << 6);
+		/* reserve mmhub engine 3 for firmware */
+		if (adev->enable_umsch_mm)
+			vm_inv_engs[i] &= ~(1 << 3);
+	}
 
 	for (i = 0; i < adev->num_rings; ++i) {
 		ring = adev->rings[i];
-		vmhub = ring->funcs->vmhub;
+		vmhub = ring->vm_hub;
+
+		if (ring == &adev->mes.ring[0] ||
+		    ring == &adev->mes.ring[1] ||
+		    ring == &adev->umsch_mm.ring ||
+		    ring == &adev->cper.ring_buf)
+			continue;
 
-		if (ring == &adev->mes.ring)
+		/* Skip if the ring is a shared ring */
+		if (amdgpu_sdma_is_shared_inv_eng(adev, ring))
 			continue;
 
 		inv_eng = ffs(vm_inv_engs[vmhub]);
@@ -545,12 +630,230 @@ int amdgpu_gmc_allocate_vm_inv_eng(struct amdgpu_device *adev)
 		vm_inv_engs[vmhub] &= ~(1 << ring->vm_inv_eng);
 
 		dev_info(adev->dev, "ring %s uses VM inv eng %u on hub %u\n",
-			 ring->name, ring->vm_inv_eng, ring->funcs->vmhub);
+			 ring->name, ring->vm_inv_eng, ring->vm_hub);
+		/* SDMA has a special packet which allows it to use the same
+		 * invalidation engine for all the rings in one instance.
+		 * Therefore, we do not allocate a separate VM invalidation engine
+		 * for SDMA page rings. Instead, they share the VM invalidation
+		 * engine with the SDMA gfx ring. This change ensures efficient
+		 * resource management and avoids the issue of insufficient VM
+		 * invalidation engines.
+		 */
+		shared_ring = amdgpu_sdma_get_shared_ring(adev, ring);
+		if (shared_ring) {
+			shared_ring->vm_inv_eng = ring->vm_inv_eng;
+			dev_info(adev->dev, "ring %s shares VM invalidation engine %u with ring %s on hub %u\n",
+					ring->name, ring->vm_inv_eng, shared_ring->name, ring->vm_hub);
+			continue;
+		}
 	}
 
 	return 0;
 }
 
+void amdgpu_gmc_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid,
+			      uint32_t vmhub, uint32_t flush_type)
+{
+	struct amdgpu_ring *ring = adev->mman.buffer_funcs_ring;
+	struct amdgpu_vmhub *hub = &adev->vmhub[vmhub];
+	struct dma_fence *fence;
+	struct amdgpu_job *job;
+	int r;
+
+	if (!hub->sdma_invalidation_workaround || vmid ||
+	    !adev->mman.buffer_funcs_enabled || !adev->ib_pool_ready ||
+	    !ring->sched.ready) {
+		/*
+		 * A GPU reset should flush all TLBs anyway, so no need to do
+		 * this while one is ongoing.
+		 */
+		if (!down_read_trylock(&adev->reset_domain->sem))
+			return;
+
+		if (adev->gmc.flush_tlb_needs_extra_type_2)
+			adev->gmc.gmc_funcs->flush_gpu_tlb(adev, vmid,
+							   vmhub, 2);
+
+		if (adev->gmc.flush_tlb_needs_extra_type_0 && flush_type == 2)
+			adev->gmc.gmc_funcs->flush_gpu_tlb(adev, vmid,
+							   vmhub, 0);
+
+		adev->gmc.gmc_funcs->flush_gpu_tlb(adev, vmid, vmhub,
+						   flush_type);
+		up_read(&adev->reset_domain->sem);
+		return;
+	}
+
+	/* The SDMA on Navi 1x has a bug which can theoretically result in memory
+	 * corruption if an invalidation happens at the same time as an VA
+	 * translation. Avoid this by doing the invalidation from the SDMA
+	 * itself at least for GART.
+	 */
+	mutex_lock(&adev->mman.gtt_window_lock);
+	r = amdgpu_job_alloc_with_ib(ring->adev, &adev->mman.high_pr,
+				     AMDGPU_FENCE_OWNER_UNDEFINED,
+				     16 * 4, AMDGPU_IB_POOL_IMMEDIATE,
+				     &job, AMDGPU_KERNEL_JOB_ID_FLUSH_GPU_TLB);
+	if (r)
+		goto error_alloc;
+
+	job->vm_pd_addr = amdgpu_gmc_pd_addr(adev->gart.bo);
+	job->vm_needs_flush = true;
+	job->ibs->ptr[job->ibs->length_dw++] = ring->funcs->nop;
+	amdgpu_ring_pad_ib(ring, &job->ibs[0]);
+	fence = amdgpu_job_submit(job);
+	mutex_unlock(&adev->mman.gtt_window_lock);
+
+	dma_fence_wait(fence, false);
+	dma_fence_put(fence);
+
+	return;
+
+error_alloc:
+	mutex_unlock(&adev->mman.gtt_window_lock);
+	dev_err(adev->dev, "Error flushing GPU TLB using the SDMA (%d)!\n", r);
+}
+
+int amdgpu_gmc_flush_gpu_tlb_pasid(struct amdgpu_device *adev, uint16_t pasid,
+				   uint32_t flush_type, bool all_hub,
+				   uint32_t inst)
+{
+	struct amdgpu_ring *ring = &adev->gfx.kiq[inst].ring;
+	struct amdgpu_kiq *kiq = &adev->gfx.kiq[inst];
+	unsigned int ndw;
+	int r, cnt = 0;
+	uint32_t seq;
+
+	/*
+	 * A GPU reset should flush all TLBs anyway, so no need to do
+	 * this while one is ongoing.
+	 */
+	if (!down_read_trylock(&adev->reset_domain->sem))
+		return 0;
+
+	if (!adev->gmc.flush_pasid_uses_kiq || !ring->sched.ready) {
+		if (adev->gmc.flush_tlb_needs_extra_type_2)
+			adev->gmc.gmc_funcs->flush_gpu_tlb_pasid(adev, pasid,
+								 2, all_hub,
+								 inst);
+
+		if (adev->gmc.flush_tlb_needs_extra_type_0 && flush_type == 2)
+			adev->gmc.gmc_funcs->flush_gpu_tlb_pasid(adev, pasid,
+								 0, all_hub,
+								 inst);
+
+		adev->gmc.gmc_funcs->flush_gpu_tlb_pasid(adev, pasid,
+							 flush_type, all_hub,
+							 inst);
+		r = 0;
+	} else {
+		/* 2 dwords flush + 8 dwords fence */
+		ndw = kiq->pmf->invalidate_tlbs_size + 8;
+
+		if (adev->gmc.flush_tlb_needs_extra_type_2)
+			ndw += kiq->pmf->invalidate_tlbs_size;
+
+		if (adev->gmc.flush_tlb_needs_extra_type_0)
+			ndw += kiq->pmf->invalidate_tlbs_size;
+
+		spin_lock(&adev->gfx.kiq[inst].ring_lock);
+		r = amdgpu_ring_alloc(ring, ndw);
+		if (r) {
+			spin_unlock(&adev->gfx.kiq[inst].ring_lock);
+			goto error_unlock_reset;
+		}
+		if (adev->gmc.flush_tlb_needs_extra_type_2)
+			kiq->pmf->kiq_invalidate_tlbs(ring, pasid, 2, all_hub);
+
+		if (flush_type == 2 && adev->gmc.flush_tlb_needs_extra_type_0)
+			kiq->pmf->kiq_invalidate_tlbs(ring, pasid, 0, all_hub);
+
+		kiq->pmf->kiq_invalidate_tlbs(ring, pasid, flush_type, all_hub);
+		r = amdgpu_fence_emit_polling(ring, &seq, MAX_KIQ_REG_WAIT);
+		if (r) {
+			amdgpu_ring_undo(ring);
+			spin_unlock(&adev->gfx.kiq[inst].ring_lock);
+			goto error_unlock_reset;
+		}
+
+		amdgpu_ring_commit(ring);
+		spin_unlock(&adev->gfx.kiq[inst].ring_lock);
+
+		r = amdgpu_fence_wait_polling(ring, seq, MAX_KIQ_REG_WAIT);
+
+		might_sleep();
+		while (r < 1 && cnt++ < MAX_KIQ_REG_TRY &&
+		       !amdgpu_reset_pending(adev->reset_domain)) {
+			msleep(MAX_KIQ_REG_BAILOUT_INTERVAL);
+			r = amdgpu_fence_wait_polling(ring, seq, MAX_KIQ_REG_WAIT);
+		}
+
+		if (cnt > MAX_KIQ_REG_TRY) {
+			dev_err(adev->dev, "timeout waiting for kiq fence\n");
+			r = -ETIME;
+		} else
+			r = 0;
+	}
+
+error_unlock_reset:
+	up_read(&adev->reset_domain->sem);
+	return r;
+}
+
+void amdgpu_gmc_fw_reg_write_reg_wait(struct amdgpu_device *adev,
+				      uint32_t reg0, uint32_t reg1,
+				      uint32_t ref, uint32_t mask,
+				      uint32_t xcc_inst)
+{
+	struct amdgpu_kiq *kiq = &adev->gfx.kiq[xcc_inst];
+	struct amdgpu_ring *ring = &kiq->ring;
+	signed long r, cnt = 0;
+	unsigned long flags;
+	uint32_t seq;
+
+	if (adev->mes.ring[0].sched.ready) {
+		amdgpu_mes_reg_write_reg_wait(adev, reg0, reg1,
+					      ref, mask);
+		return;
+	}
+
+	spin_lock_irqsave(&kiq->ring_lock, flags);
+	amdgpu_ring_alloc(ring, 32);
+	amdgpu_ring_emit_reg_write_reg_wait(ring, reg0, reg1,
+					    ref, mask);
+	r = amdgpu_fence_emit_polling(ring, &seq, MAX_KIQ_REG_WAIT);
+	if (r)
+		goto failed_undo;
+
+	amdgpu_ring_commit(ring);
+	spin_unlock_irqrestore(&kiq->ring_lock, flags);
+
+	r = amdgpu_fence_wait_polling(ring, seq, MAX_KIQ_REG_WAIT);
+
+	/* don't wait anymore for IRQ context */
+	if (r < 1 && in_interrupt())
+		goto failed_kiq;
+
+	might_sleep();
+	while (r < 1 && cnt++ < MAX_KIQ_REG_TRY &&
+	       !amdgpu_reset_pending(adev->reset_domain)) {
+
+		msleep(MAX_KIQ_REG_BAILOUT_INTERVAL);
+		r = amdgpu_fence_wait_polling(ring, seq, MAX_KIQ_REG_WAIT);
+	}
+
+	if (cnt > MAX_KIQ_REG_TRY)
+		goto failed_kiq;
+
+	return;
+
+failed_undo:
+	amdgpu_ring_undo(ring);
+	spin_unlock_irqrestore(&kiq->ring_lock, flags);
+failed_kiq:
+	dev_err(adev->dev, "failed to write reg %x wait reg %x\n", reg0, reg1);
+}
+
 /**
  * amdgpu_gmc_tmz_set -- check and set if a device supports TMZ
  * @adev: amdgpu_device pointer
@@ -560,9 +863,16 @@ int amdgpu_gmc_allocate_vm_inv_eng(struct amdgpu_device *adev)
  */
 void amdgpu_gmc_tmz_set(struct amdgpu_device *adev)
 {
-	switch (adev->asic_type) {
-	case CHIP_RAVEN:
-	case CHIP_RENOIR:
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	/* RAVEN */
+	case IP_VERSION(9, 2, 2):
+	case IP_VERSION(9, 1, 0):
+	/* RENOIR looks like RAVEN */
+	case IP_VERSION(9, 3, 0):
+	/* GC 10.3.7 */
+	case IP_VERSION(10, 3, 7):
+	/* GC 11.0.1 */
+	case IP_VERSION(11, 0, 1):
 		if (amdgpu_tmz == 0) {
 			adev->gmc.tmz_enabled = false;
 			dev_info(adev->dev,
@@ -573,11 +883,24 @@ void amdgpu_gmc_tmz_set(struct amdgpu_device *adev)
 				 "Trusted Memory Zone (TMZ) feature enabled\n");
 		}
 		break;
-	case CHIP_NAVI10:
-	case CHIP_NAVI14:
-	case CHIP_NAVI12:
-	case CHIP_VANGOGH:
-	case CHIP_YELLOW_CARP:
+	case IP_VERSION(10, 1, 10):
+	case IP_VERSION(10, 1, 1):
+	case IP_VERSION(10, 1, 2):
+	case IP_VERSION(10, 1, 3):
+	case IP_VERSION(10, 3, 0):
+	case IP_VERSION(10, 3, 2):
+	case IP_VERSION(10, 3, 4):
+	case IP_VERSION(10, 3, 5):
+	case IP_VERSION(10, 3, 6):
+	/* VANGOGH */
+	case IP_VERSION(10, 3, 1):
+	/* YELLOW_CARP*/
+	case IP_VERSION(10, 3, 3):
+	case IP_VERSION(11, 0, 4):
+	case IP_VERSION(11, 5, 0):
+	case IP_VERSION(11, 5, 1):
+	case IP_VERSION(11, 5, 2):
+	case IP_VERSION(11, 5, 3):
 		/* Don't enable it by default yet.
 		 */
 		if (amdgpu_tmz < 1) {
@@ -608,40 +931,20 @@ void amdgpu_gmc_tmz_set(struct amdgpu_device *adev)
 void amdgpu_gmc_noretry_set(struct amdgpu_device *adev)
 {
 	struct amdgpu_gmc *gmc = &adev->gmc;
-
-	switch (adev->asic_type) {
-	case CHIP_VEGA10:
-	case CHIP_VEGA20:
-	case CHIP_ARCTURUS:
-	case CHIP_ALDEBARAN:
-		/*
-		 * noretry = 0 will cause kfd page fault tests fail
-		 * for some ASICs, so set default to 1 for these ASICs.
-		 */
-		if (amdgpu_noretry == -1)
-			gmc->noretry = 1;
-		else
-			gmc->noretry = amdgpu_noretry;
-		break;
-	case CHIP_RAVEN:
-	default:
-		/* Raven currently has issues with noretry
-		 * regardless of what we decide for other
-		 * asics, we should leave raven with
-		 * noretry = 0 until we root cause the
-		 * issues.
-		 *
-		 * default this to 0 for now, but we may want
-		 * to change this in the future for certain
-		 * GPUs as it can increase performance in
-		 * certain cases.
-		 */
-		if (amdgpu_noretry == -1)
-			gmc->noretry = 0;
-		else
-			gmc->noretry = amdgpu_noretry;
-		break;
-	}
+	uint32_t gc_ver = amdgpu_ip_version(adev, GC_HWIP, 0);
+	bool noretry_default = (gc_ver == IP_VERSION(9, 0, 1) ||
+				gc_ver == IP_VERSION(9, 4, 0) ||
+				gc_ver == IP_VERSION(9, 4, 1) ||
+				gc_ver == IP_VERSION(9, 4, 2) ||
+				gc_ver == IP_VERSION(9, 4, 3) ||
+				gc_ver == IP_VERSION(9, 4, 4) ||
+				gc_ver == IP_VERSION(9, 5, 0) ||
+				gc_ver >= IP_VERSION(10, 3, 0));
+
+	if (!amdgpu_sriov_xnack_support(adev))
+		gmc->noretry = 1;
+	else
+		gmc->noretry = (amdgpu_noretry == -1) ? noretry_default : amdgpu_noretry;
 }
 
 void amdgpu_gmc_set_vm_fault_masks(struct amdgpu_device *adev, int hub_type,
@@ -654,7 +957,7 @@ void amdgpu_gmc_set_vm_fault_masks(struct amdgpu_device *adev, int hub_type,
 	for (i = 0; i < 16; i++) {
 		reg = hub->vm_context0_cntl + hub->ctx_distance * i;
 
-		tmp = (hub_type == AMDGPU_GFXHUB_0) ?
+		tmp = (hub_type == AMDGPU_GFXHUB(0)) ?
 			RREG32_SOC15_IP(GC, reg) :
 			RREG32_SOC15_IP(MMHUB, reg);
 
@@ -663,7 +966,7 @@ void amdgpu_gmc_set_vm_fault_masks(struct amdgpu_device *adev, int hub_type,
 		else
 			tmp &= ~hub->vm_cntx_cntl_vm_fault;
 
-		(hub_type == AMDGPU_GFXHUB_0) ?
+		(hub_type == AMDGPU_GFXHUB(0)) ?
 			WREG32_SOC15_IP(GC, reg, tmp) :
 			WREG32_SOC15_IP(MMHUB, reg, tmp);
 	}
@@ -674,6 +977,13 @@ void amdgpu_gmc_get_vbios_allocations(struct amdgpu_device *adev)
 	unsigned size;
 
 	/*
+	 * Some ASICs need to reserve a region of video memory to avoid access
+	 * from driver
+	 */
+	adev->mman.stolen_reserved_offset = 0;
+	adev->mman.stolen_reserved_size = 0;
+
+	/*
 	 * TODO:
 	 * Currently there is a bug where some memory client outside
 	 * of the driver writes to first 8M of VRAM on S3 resume,
@@ -683,6 +993,17 @@ void amdgpu_gmc_get_vbios_allocations(struct amdgpu_device *adev)
 	 */
 	switch (adev->asic_type) {
 	case CHIP_VEGA10:
+		adev->mman.keep_stolen_vga_memory = true;
+		/*
+		 * VEGA10 SRIOV VF with MS_HYPERV host needs some firmware reserved area.
+		 */
+#ifdef CONFIG_X86
+		if (amdgpu_sriov_vf(adev) && hypervisor_is_type(X86_HYPER_MS_HYPERV)) {
+			adev->mman.stolen_reserved_offset = 0x500000;
+			adev->mman.stolen_reserved_size = 0x200000;
+		}
+#endif
+		break;
 	case CHIP_RAVEN:
 	case CHIP_RENOIR:
 		adev->mman.keep_stolen_vga_memory = true;
@@ -693,7 +1014,7 @@ void amdgpu_gmc_get_vbios_allocations(struct amdgpu_device *adev)
 	}
 
 	if (amdgpu_sriov_vf(adev) ||
-	    !amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_DCE)) {
+	    !amdgpu_device_has_display_hardware(adev)) {
 		size = 0;
 	} else {
 		size = amdgpu_gmc_get_vbios_fb_size(adev);
@@ -739,9 +1060,7 @@ void amdgpu_gmc_init_pdb0(struct amdgpu_device *adev)
 	 */
 	u64 vram_size = adev->gmc.xgmi.node_segment_size * adev->gmc.xgmi.num_physical_nodes;
 	u64 pde0_page_size = (1ULL<<adev->gmc.vmid0_page_table_block_size)<<21;
-	u64 vram_addr = adev->vm_manager.vram_base_offset -
-		adev->gmc.xgmi.physical_node_id * adev->gmc.xgmi.node_segment_size;
-	u64 vram_end = vram_addr + vram_size;
+	u64 vram_addr, vram_end;
 	u64 gart_ptb_gpu_pa = amdgpu_gmc_vram_pa(adev, adev->gart.bo);
 	int idx;
 
@@ -752,7 +1071,12 @@ void amdgpu_gmc_init_pdb0(struct amdgpu_device *adev)
 	flags |= AMDGPU_PTE_WRITEABLE;
 	flags |= AMDGPU_PTE_SNOOPED;
 	flags |= AMDGPU_PTE_FRAG((adev->gmc.vmid0_page_table_block_size + 9*1));
-	flags |= AMDGPU_PDE_PTE;
+	flags |= AMDGPU_PDE_PTE_FLAG(adev);
+
+	vram_addr = adev->vm_manager.vram_base_offset;
+	if (!amdgpu_virt_xgmi_migrate_enabled(adev))
+		vram_addr -= adev->gmc.xgmi.physical_node_id * adev->gmc.xgmi.node_segment_size;
+	vram_end = vram_addr + vram_size;
 
 	/* The first n PDE0 entries are used as PTE,
 	 * pointing to vram
@@ -765,7 +1089,7 @@ void amdgpu_gmc_init_pdb0(struct amdgpu_device *adev)
 	 * pointing to a 4K system page
 	 */
 	flags = AMDGPU_PTE_VALID;
-	flags |= AMDGPU_PDE_BFS(0) | AMDGPU_PTE_SNOOPED;
+	flags |= AMDGPU_PTE_SNOOPED | AMDGPU_PDE_BFS_FLAG(adev, 0);
 	/* Requires gart_ptb_gpu_pa to be 4K aligned */
 	amdgpu_gmc_set_pte_pde(adev, adev->gmc.ptr_pdb0, i, gart_ptb_gpu_pa, flags);
 	drm_dev_exit(idx);
@@ -795,33 +1119,564 @@ uint64_t amdgpu_gmc_vram_pa(struct amdgpu_device *adev, struct amdgpu_bo *bo)
 	return amdgpu_gmc_vram_mc2pa(adev, amdgpu_bo_gpu_offset(bo));
 }
 
-/**
- * amdgpu_gmc_vram_cpu_pa - calculate vram buffer object's physical address
- * from CPU's view
- *
- * @adev: amdgpu_device pointer
- * @bo: amdgpu buffer object
- */
-uint64_t amdgpu_gmc_vram_cpu_pa(struct amdgpu_device *adev, struct amdgpu_bo *bo)
+int amdgpu_gmc_vram_checking(struct amdgpu_device *adev)
+{
+	struct amdgpu_bo *vram_bo = NULL;
+	uint64_t vram_gpu = 0;
+	void *vram_ptr = NULL;
+
+	int ret, size = 0x100000;
+	uint8_t cptr[10];
+
+	ret = amdgpu_bo_create_kernel(adev, size, PAGE_SIZE,
+				AMDGPU_GEM_DOMAIN_VRAM,
+				&vram_bo,
+				&vram_gpu,
+				&vram_ptr);
+	if (ret)
+		return ret;
+
+	memset(vram_ptr, 0x86, size);
+	memset(cptr, 0x86, 10);
+
+	/**
+	 * Check the start, the mid, and the end of the memory if the content of
+	 * each byte is the pattern "0x86". If yes, we suppose the vram bo is
+	 * workable.
+	 *
+	 * Note: If check the each byte of whole 1M bo, it will cost too many
+	 * seconds, so here, we just pick up three parts for emulation.
+	 */
+	ret = memcmp(vram_ptr, cptr, 10);
+	if (ret) {
+		ret = -EIO;
+		goto release_buffer;
+	}
+
+	ret = memcmp(vram_ptr + (size / 2), cptr, 10);
+	if (ret) {
+		ret = -EIO;
+		goto release_buffer;
+	}
+
+	ret = memcmp(vram_ptr + size - 10, cptr, 10);
+	if (ret) {
+		ret = -EIO;
+		goto release_buffer;
+	}
+
+release_buffer:
+	amdgpu_bo_free_kernel(&vram_bo, &vram_gpu,
+			&vram_ptr);
+
+	return ret;
+}
+
+static const char *nps_desc[] = {
+	[AMDGPU_NPS1_PARTITION_MODE] = "NPS1",
+	[AMDGPU_NPS2_PARTITION_MODE] = "NPS2",
+	[AMDGPU_NPS3_PARTITION_MODE] = "NPS3",
+	[AMDGPU_NPS4_PARTITION_MODE] = "NPS4",
+	[AMDGPU_NPS6_PARTITION_MODE] = "NPS6",
+	[AMDGPU_NPS8_PARTITION_MODE] = "NPS8",
+};
+
+static ssize_t available_memory_partition_show(struct device *dev,
+					       struct device_attribute *addr,
+					       char *buf)
 {
-	return amdgpu_bo_gpu_offset(bo) - adev->gmc.vram_start + adev->gmc.aper_base;
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = drm_to_adev(ddev);
+	int size = 0, mode;
+	char *sep = "";
+
+	for_each_inst(mode, adev->gmc.supported_nps_modes) {
+		size += sysfs_emit_at(buf, size, "%s%s", sep, nps_desc[mode]);
+		sep = ", ";
+	}
+	size += sysfs_emit_at(buf, size, "\n");
+
+	return size;
 }
 
-void amdgpu_gmc_get_reserved_allocation(struct amdgpu_device *adev)
+static ssize_t current_memory_partition_store(struct device *dev,
+					      struct device_attribute *attr,
+					      const char *buf, size_t count)
 {
-	/* Some ASICs need to reserve a region of video memory to avoid access
-	 * from driver */
-	adev->mman.stolen_reserved_offset = 0;
-	adev->mman.stolen_reserved_size = 0;
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = drm_to_adev(ddev);
+	enum amdgpu_memory_partition mode;
+	struct amdgpu_hive_info *hive;
+	int i;
 
-	switch (adev->asic_type) {
-	case CHIP_YELLOW_CARP:
-		if (amdgpu_discovery == 0) {
-			adev->mman.stolen_reserved_offset = 0x1ffb0000;
-			adev->mman.stolen_reserved_size = 64 * PAGE_SIZE;
+	mode = UNKNOWN_MEMORY_PARTITION_MODE;
+	for_each_inst(i, adev->gmc.supported_nps_modes) {
+		if (!strncasecmp(nps_desc[i], buf, strlen(nps_desc[i]))) {
+			mode = i;
+			break;
+		}
+	}
+
+	if (mode == UNKNOWN_MEMORY_PARTITION_MODE)
+		return -EINVAL;
+
+	if (mode == adev->gmc.gmc_funcs->query_mem_partition_mode(adev)) {
+		dev_info(
+			adev->dev,
+			"requested NPS mode is same as current NPS mode, skipping\n");
+		return count;
+	}
+
+	/* If device is part of hive, all devices in the hive should request the
+	 * same mode. Hence store the requested mode in hive.
+	 */
+	hive = amdgpu_get_xgmi_hive(adev);
+	if (hive) {
+		atomic_set(&hive->requested_nps_mode, mode);
+		amdgpu_put_xgmi_hive(hive);
+	} else {
+		adev->gmc.requested_nps_mode = mode;
+	}
+
+	dev_info(
+		adev->dev,
+		"NPS mode change requested, please remove and reload the driver\n");
+
+	return count;
+}
+
+static ssize_t current_memory_partition_show(
+	struct device *dev, struct device_attribute *addr, char *buf)
+{
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = drm_to_adev(ddev);
+	enum amdgpu_memory_partition mode;
+
+	/* Only minimal precaution taken to reject requests while in reset */
+	if (amdgpu_in_reset(adev))
+		return -EPERM;
+
+	mode = adev->gmc.gmc_funcs->query_mem_partition_mode(adev);
+	if ((mode >= ARRAY_SIZE(nps_desc)) ||
+	    (BIT(mode) & AMDGPU_ALL_NPS_MASK) != BIT(mode))
+		return sysfs_emit(buf, "UNKNOWN\n");
+
+	return sysfs_emit(buf, "%s\n", nps_desc[mode]);
+}
+
+static DEVICE_ATTR_RW(current_memory_partition);
+static DEVICE_ATTR_RO(available_memory_partition);
+
+int amdgpu_gmc_sysfs_init(struct amdgpu_device *adev)
+{
+	bool nps_switch_support;
+	int r = 0;
+
+	if (!adev->gmc.gmc_funcs->query_mem_partition_mode)
+		return 0;
+
+	nps_switch_support = (hweight32(adev->gmc.supported_nps_modes &
+					AMDGPU_ALL_NPS_MASK) > 1);
+	if (!nps_switch_support)
+		dev_attr_current_memory_partition.attr.mode &=
+			~(S_IWUSR | S_IWGRP | S_IWOTH);
+	else
+		r = device_create_file(adev->dev,
+				       &dev_attr_available_memory_partition);
+
+	if (r)
+		return r;
+
+	return device_create_file(adev->dev,
+				  &dev_attr_current_memory_partition);
+}
+
+void amdgpu_gmc_sysfs_fini(struct amdgpu_device *adev)
+{
+	if (!adev->gmc.gmc_funcs->query_mem_partition_mode)
+		return;
+
+	device_remove_file(adev->dev, &dev_attr_current_memory_partition);
+	device_remove_file(adev->dev, &dev_attr_available_memory_partition);
+}
+
+int amdgpu_gmc_get_nps_memranges(struct amdgpu_device *adev,
+				 struct amdgpu_mem_partition_info *mem_ranges,
+				 uint8_t *exp_ranges)
+{
+	struct amdgpu_gmc_memrange *ranges;
+	int range_cnt, ret, i, j;
+	uint32_t nps_type;
+	bool refresh;
+
+	if (!mem_ranges || !exp_ranges)
+		return -EINVAL;
+
+	refresh = (adev->init_lvl->level != AMDGPU_INIT_LEVEL_MINIMAL_XGMI) &&
+		  (adev->gmc.reset_flags & AMDGPU_GMC_INIT_RESET_NPS);
+	ret = amdgpu_discovery_get_nps_info(adev, &nps_type, &ranges,
+					    &range_cnt, refresh);
+
+	if (ret)
+		return ret;
+
+	/* TODO: For now, expect ranges and partition count to be the same.
+	 * Adjust if there are holes expected in any NPS domain.
+	 */
+	if (*exp_ranges && (range_cnt != *exp_ranges)) {
+		dev_warn(
+			adev->dev,
+			"NPS config mismatch - expected ranges: %d discovery - nps mode: %d, nps ranges: %d",
+			*exp_ranges, nps_type, range_cnt);
+		ret = -EINVAL;
+		goto err;
+	}
+
+	for (i = 0; i < range_cnt; ++i) {
+		if (ranges[i].base_address >= ranges[i].limit_address) {
+			dev_warn(
+				adev->dev,
+				"Invalid NPS range - nps mode: %d, range[%d]: base: %llx limit: %llx",
+				nps_type, i, ranges[i].base_address,
+				ranges[i].limit_address);
+			ret = -EINVAL;
+			goto err;
+		}
+
+		/* Check for overlaps, not expecting any now */
+		for (j = i - 1; j >= 0; j--) {
+			if (max(ranges[j].base_address,
+				ranges[i].base_address) <=
+			    min(ranges[j].limit_address,
+				ranges[i].limit_address)) {
+				dev_warn(
+					adev->dev,
+					"overlapping ranges detected [ %llx - %llx ] | [%llx - %llx]",
+					ranges[j].base_address,
+					ranges[j].limit_address,
+					ranges[i].base_address,
+					ranges[i].limit_address);
+				ret = -EINVAL;
+				goto err;
+			}
+		}
+
+		mem_ranges[i].range.fpfn =
+			(ranges[i].base_address -
+			 adev->vm_manager.vram_base_offset) >>
+			AMDGPU_GPU_PAGE_SHIFT;
+		mem_ranges[i].range.lpfn =
+			(ranges[i].limit_address -
+			 adev->vm_manager.vram_base_offset) >>
+			AMDGPU_GPU_PAGE_SHIFT;
+		mem_ranges[i].size =
+			ranges[i].limit_address - ranges[i].base_address + 1;
+	}
+
+	if (!*exp_ranges)
+		*exp_ranges = range_cnt;
+err:
+	kfree(ranges);
+
+	return ret;
+}
+
+int amdgpu_gmc_request_memory_partition(struct amdgpu_device *adev,
+					int nps_mode)
+{
+	/* Not supported on VF devices and APUs */
+	if (amdgpu_sriov_vf(adev) || (adev->flags & AMD_IS_APU))
+		return -EOPNOTSUPP;
+
+	if (!adev->psp.funcs) {
+		dev_err(adev->dev,
+			"PSP interface not available for nps mode change request");
+		return -EINVAL;
+	}
+
+	return psp_memory_partition(&adev->psp, nps_mode);
+}
+
+static inline bool amdgpu_gmc_need_nps_switch_req(struct amdgpu_device *adev,
+						  int req_nps_mode,
+						  int cur_nps_mode)
+{
+	return (((BIT(req_nps_mode) & adev->gmc.supported_nps_modes) ==
+			BIT(req_nps_mode)) &&
+		req_nps_mode != cur_nps_mode);
+}
+
+void amdgpu_gmc_prepare_nps_mode_change(struct amdgpu_device *adev)
+{
+	int req_nps_mode, cur_nps_mode, r;
+	struct amdgpu_hive_info *hive;
+
+	if (amdgpu_sriov_vf(adev) || !adev->gmc.supported_nps_modes ||
+	    !adev->gmc.gmc_funcs->request_mem_partition_mode)
+		return;
+
+	cur_nps_mode = adev->gmc.gmc_funcs->query_mem_partition_mode(adev);
+	hive = amdgpu_get_xgmi_hive(adev);
+	if (hive) {
+		req_nps_mode = atomic_read(&hive->requested_nps_mode);
+		if (!amdgpu_gmc_need_nps_switch_req(adev, req_nps_mode,
+						    cur_nps_mode)) {
+			amdgpu_put_xgmi_hive(hive);
+			return;
 		}
+		r = amdgpu_xgmi_request_nps_change(adev, hive, req_nps_mode);
+		amdgpu_put_xgmi_hive(hive);
+		goto out;
+	}
+
+	req_nps_mode = adev->gmc.requested_nps_mode;
+	if (!amdgpu_gmc_need_nps_switch_req(adev, req_nps_mode, cur_nps_mode))
+		return;
+
+	/* even if this fails, we should let driver unload w/o blocking */
+	r = adev->gmc.gmc_funcs->request_mem_partition_mode(adev, req_nps_mode);
+out:
+	if (r)
+		dev_err(adev->dev, "NPS mode change request failed\n");
+	else
+		dev_info(
+			adev->dev,
+			"NPS mode change request done, reload driver to complete the change\n");
+}
+
+bool amdgpu_gmc_need_reset_on_init(struct amdgpu_device *adev)
+{
+	if (adev->gmc.gmc_funcs->need_reset_on_init)
+		return adev->gmc.gmc_funcs->need_reset_on_init(adev);
+
+	return false;
+}
+
+enum amdgpu_memory_partition
+amdgpu_gmc_get_vf_memory_partition(struct amdgpu_device *adev)
+{
+	switch (adev->gmc.num_mem_partitions) {
+	case 0:
+		return UNKNOWN_MEMORY_PARTITION_MODE;
+	case 1:
+		return AMDGPU_NPS1_PARTITION_MODE;
+	case 2:
+		return AMDGPU_NPS2_PARTITION_MODE;
+	case 4:
+		return AMDGPU_NPS4_PARTITION_MODE;
+	case 8:
+		return AMDGPU_NPS8_PARTITION_MODE;
+	default:
+		return AMDGPU_NPS1_PARTITION_MODE;
+	}
+}
+
+enum amdgpu_memory_partition
+amdgpu_gmc_get_memory_partition(struct amdgpu_device *adev, u32 *supp_modes)
+{
+	enum amdgpu_memory_partition mode = UNKNOWN_MEMORY_PARTITION_MODE;
+
+	if (adev->nbio.funcs &&
+	    adev->nbio.funcs->get_memory_partition_mode)
+		mode = adev->nbio.funcs->get_memory_partition_mode(adev,
+								   supp_modes);
+	else
+		dev_warn(adev->dev, "memory partition mode query is not supported\n");
+
+	return mode;
+}
+
+enum amdgpu_memory_partition
+amdgpu_gmc_query_memory_partition(struct amdgpu_device *adev)
+{
+	if (amdgpu_sriov_vf(adev))
+		return amdgpu_gmc_get_vf_memory_partition(adev);
+	else
+		return amdgpu_gmc_get_memory_partition(adev, NULL);
+}
+
+static bool amdgpu_gmc_validate_partition_info(struct amdgpu_device *adev)
+{
+	enum amdgpu_memory_partition mode;
+	u32 supp_modes;
+	bool valid;
+
+	mode = amdgpu_gmc_get_memory_partition(adev, &supp_modes);
+
+	/* Mode detected by hardware not present in supported modes */
+	if ((mode != UNKNOWN_MEMORY_PARTITION_MODE) &&
+	    !(BIT(mode - 1) & supp_modes))
+		return false;
+
+	switch (mode) {
+	case UNKNOWN_MEMORY_PARTITION_MODE:
+	case AMDGPU_NPS1_PARTITION_MODE:
+		valid = (adev->gmc.num_mem_partitions == 1);
+		break;
+	case AMDGPU_NPS2_PARTITION_MODE:
+		valid = (adev->gmc.num_mem_partitions == 2);
+		break;
+	case AMDGPU_NPS4_PARTITION_MODE:
+		valid = (adev->gmc.num_mem_partitions == 3 ||
+			 adev->gmc.num_mem_partitions == 4);
+		break;
+	case AMDGPU_NPS8_PARTITION_MODE:
+		valid = (adev->gmc.num_mem_partitions == 8);
 		break;
 	default:
+		valid = false;
+	}
+
+	return valid;
+}
+
+static bool amdgpu_gmc_is_node_present(int *node_ids, int num_ids, int nid)
+{
+	int i;
+
+	/* Check if node with id 'nid' is present in 'node_ids' array */
+	for (i = 0; i < num_ids; ++i)
+		if (node_ids[i] == nid)
+			return true;
+
+	return false;
+}
+
+static void
+amdgpu_gmc_init_acpi_mem_ranges(struct amdgpu_device *adev,
+				struct amdgpu_mem_partition_info *mem_ranges)
+{
+	struct amdgpu_numa_info numa_info;
+	int node_ids[AMDGPU_MAX_MEM_RANGES];
+	int num_ranges = 0, ret;
+	int num_xcc, xcc_id;
+	uint32_t xcc_mask;
+
+	num_xcc = NUM_XCC(adev->gfx.xcc_mask);
+	xcc_mask = (1U << num_xcc) - 1;
+
+	for_each_inst(xcc_id, xcc_mask)	{
+		ret = amdgpu_acpi_get_mem_info(adev, xcc_id, &numa_info);
+		if (ret)
+			continue;
+
+		if (numa_info.nid == NUMA_NO_NODE) {
+			mem_ranges[0].size = numa_info.size;
+			mem_ranges[0].numa.node = numa_info.nid;
+			num_ranges = 1;
+			break;
+		}
+
+		if (amdgpu_gmc_is_node_present(node_ids, num_ranges,
+					     numa_info.nid))
+			continue;
+
+		node_ids[num_ranges] = numa_info.nid;
+		mem_ranges[num_ranges].numa.node = numa_info.nid;
+		mem_ranges[num_ranges].size = numa_info.size;
+		++num_ranges;
+	}
+
+	adev->gmc.num_mem_partitions = num_ranges;
+}
+
+void amdgpu_gmc_init_sw_mem_ranges(struct amdgpu_device *adev,
+				   struct amdgpu_mem_partition_info *mem_ranges)
+{
+	enum amdgpu_memory_partition mode;
+	u32 start_addr = 0, size;
+	int i, r, l;
+
+	mode = amdgpu_gmc_query_memory_partition(adev);
+
+	switch (mode) {
+	case UNKNOWN_MEMORY_PARTITION_MODE:
+		adev->gmc.num_mem_partitions = 0;
+		break;
+	case AMDGPU_NPS1_PARTITION_MODE:
+		adev->gmc.num_mem_partitions = 1;
+		break;
+	case AMDGPU_NPS2_PARTITION_MODE:
+		adev->gmc.num_mem_partitions = 2;
+		break;
+	case AMDGPU_NPS4_PARTITION_MODE:
+		if (adev->flags & AMD_IS_APU)
+			adev->gmc.num_mem_partitions = 3;
+		else
+			adev->gmc.num_mem_partitions = 4;
+		break;
+	case AMDGPU_NPS8_PARTITION_MODE:
+		adev->gmc.num_mem_partitions = 8;
+		break;
+	default:
+		adev->gmc.num_mem_partitions = 1;
 		break;
 	}
+
+	/* Use NPS range info, if populated */
+	r = amdgpu_gmc_get_nps_memranges(adev, mem_ranges,
+					 &adev->gmc.num_mem_partitions);
+	if (!r) {
+		l = 0;
+		for (i = 1; i < adev->gmc.num_mem_partitions; ++i) {
+			if (mem_ranges[i].range.lpfn >
+			    mem_ranges[i - 1].range.lpfn)
+				l = i;
+		}
+
+	} else {
+		if (!adev->gmc.num_mem_partitions) {
+			dev_warn(adev->dev,
+				 "Not able to detect NPS mode, fall back to NPS1\n");
+			adev->gmc.num_mem_partitions = 1;
+		}
+		/* Fallback to sw based calculation */
+		size = (adev->gmc.real_vram_size + SZ_16M) >> AMDGPU_GPU_PAGE_SHIFT;
+		size /= adev->gmc.num_mem_partitions;
+
+		for (i = 0; i < adev->gmc.num_mem_partitions; ++i) {
+			mem_ranges[i].range.fpfn = start_addr;
+			mem_ranges[i].size =
+				((u64)size << AMDGPU_GPU_PAGE_SHIFT);
+			mem_ranges[i].range.lpfn = start_addr + size - 1;
+			start_addr += size;
+		}
+
+		l = adev->gmc.num_mem_partitions - 1;
+	}
+
+	/* Adjust the last one */
+	mem_ranges[l].range.lpfn =
+		(adev->gmc.real_vram_size >> AMDGPU_GPU_PAGE_SHIFT) - 1;
+	mem_ranges[l].size =
+		adev->gmc.real_vram_size -
+		((u64)mem_ranges[l].range.fpfn << AMDGPU_GPU_PAGE_SHIFT);
+}
+
+int amdgpu_gmc_init_mem_ranges(struct amdgpu_device *adev)
+{
+	bool valid;
+
+	adev->gmc.mem_partitions = kcalloc(AMDGPU_MAX_MEM_RANGES,
+					   sizeof(struct amdgpu_mem_partition_info),
+					   GFP_KERNEL);
+	if (!adev->gmc.mem_partitions)
+		return -ENOMEM;
+
+	if (adev->gmc.is_app_apu)
+		amdgpu_gmc_init_acpi_mem_ranges(adev, adev->gmc.mem_partitions);
+	else
+		amdgpu_gmc_init_sw_mem_ranges(adev, adev->gmc.mem_partitions);
+
+	if (amdgpu_sriov_vf(adev))
+		valid = true;
+	else
+		valid = amdgpu_gmc_validate_partition_info(adev);
+	if (!valid) {
+		/* TODO: handle invalid case */
+		dev_warn(adev->dev,
+			 "Mem ranges not matching with hardware config\n");
+	}
+
+	return 0;
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.h
index e55201134a01..727342689d4b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.h
@@ -29,6 +29,8 @@
 #include <linux/types.h>
 
 #include "amdgpu_irq.h"
+#include "amdgpu_xgmi.h"
+#include "amdgpu_ras.h"
 
 /* VA hole for 48bit addresses on Vega10 */
 #define AMDGPU_GMC_HOLE_START	0x0000800000000000ULL
@@ -60,8 +62,35 @@
  */
 #define AMDGPU_GMC_FAULT_TIMEOUT	5000ULL
 
+/* XNACK flags */
+#define AMDGPU_GMC_XNACK_FLAG_CHAIN BIT(0)
+
 struct firmware;
 
+enum amdgpu_memory_partition {
+	UNKNOWN_MEMORY_PARTITION_MODE = 0,
+	AMDGPU_NPS1_PARTITION_MODE = 1,
+	AMDGPU_NPS2_PARTITION_MODE = 2,
+	AMDGPU_NPS3_PARTITION_MODE = 3,
+	AMDGPU_NPS4_PARTITION_MODE = 4,
+	AMDGPU_NPS6_PARTITION_MODE = 6,
+	AMDGPU_NPS8_PARTITION_MODE = 8,
+};
+
+#define AMDGPU_ALL_NPS_MASK                                                  \
+	(BIT(AMDGPU_NPS1_PARTITION_MODE) | BIT(AMDGPU_NPS2_PARTITION_MODE) | \
+	 BIT(AMDGPU_NPS3_PARTITION_MODE) | BIT(AMDGPU_NPS4_PARTITION_MODE) | \
+	 BIT(AMDGPU_NPS6_PARTITION_MODE) | BIT(AMDGPU_NPS8_PARTITION_MODE))
+
+#define AMDGPU_GMC_INIT_RESET_NPS  BIT(0)
+
+#define AMDGPU_MAX_MEM_RANGES 8
+
+#define AMDGPU_GMC9_FAULT_SOURCE_DATA_RETRY 0x80
+#define AMDGPU_GMC9_FAULT_SOURCE_DATA_READ  0x40
+#define AMDGPU_GMC9_FAULT_SOURCE_DATA_WRITE 0x20
+#define AMDGPU_GMC9_FAULT_SOURCE_DATA_EXE   0x10
+
 /*
  * GMC page fault information
  */
@@ -69,6 +98,7 @@ struct amdgpu_gmc_fault {
 	uint64_t	timestamp:48;
 	uint64_t	next:AMDGPU_GMC_FAULT_RING_ORDER;
 	atomic64_t	key;
+	uint64_t	timestamp_expiry:48;
 };
 
 /*
@@ -99,7 +129,13 @@ struct amdgpu_vmhub {
 	uint32_t	eng_distance;
 	uint32_t	eng_addr_distance; /* include LO32/HI32 */
 
+	uint32_t        vm_cntx_cntl;
 	uint32_t	vm_cntx_cntl_vm_fault;
+	uint32_t	vm_l2_bank_select_reserved_cid2;
+
+	uint32_t	vm_contexts_disable;
+
+	bool		sdma_invalidation_workaround;
 
 	const struct amdgpu_vmhub_funcs *vmhub_funcs;
 };
@@ -112,8 +148,9 @@ struct amdgpu_gmc_funcs {
 	void (*flush_gpu_tlb)(struct amdgpu_device *adev, uint32_t vmid,
 				uint32_t vmhub, uint32_t flush_type);
 	/* flush the vm tlb via pasid */
-	int (*flush_gpu_tlb_pasid)(struct amdgpu_device *adev, uint16_t pasid,
-					uint32_t flush_type, bool all_hub);
+	void (*flush_gpu_tlb_pasid)(struct amdgpu_device *adev, uint16_t pasid,
+				    uint32_t flush_type, bool all_hub,
+				    uint32_t inst);
 	/* flush the vm tlb via ring */
 	uint64_t (*emit_flush_gpu_tlb)(struct amdgpu_ring *ring, unsigned vmid,
 				       uint64_t pd_addr);
@@ -122,44 +159,58 @@ struct amdgpu_gmc_funcs {
 				   unsigned pasid);
 	/* enable/disable PRT support */
 	void (*set_prt)(struct amdgpu_device *adev, bool enable);
-	/* map mtype to hardware flags */
-	uint64_t (*map_mtype)(struct amdgpu_device *adev, uint32_t flags);
 	/* get the pde for a given mc addr */
 	void (*get_vm_pde)(struct amdgpu_device *adev, int level,
 			   u64 *dst, u64 *flags);
-	/* get the pte flags to use for a BO VA mapping */
+	/* get the pte flags to use for PTEs */
 	void (*get_vm_pte)(struct amdgpu_device *adev,
-			   struct amdgpu_bo_va_mapping *mapping,
-			   uint64_t *flags);
+			   struct amdgpu_vm *vm,
+			   struct amdgpu_bo *bo,
+			   uint32_t vm_flags,
+			   uint64_t *pte_flags);
+	/* override per-page pte flags */
+	void (*override_vm_pte_flags)(struct amdgpu_device *dev,
+				      struct amdgpu_vm *vm,
+				      uint64_t addr, uint64_t *flags);
 	/* get the amount of memory used by the vbios for pre-OS console */
 	unsigned int (*get_vbios_fb_size)(struct amdgpu_device *adev);
+	/* get the DCC buffer alignment */
+	unsigned int (*get_dcc_alignment)(struct amdgpu_device *adev);
+
+	enum amdgpu_memory_partition (*query_mem_partition_mode)(
+		struct amdgpu_device *adev);
+	/* Request NPS mode */
+	int (*request_mem_partition_mode)(struct amdgpu_device *adev,
+					  int nps_mode);
+	bool (*need_reset_on_init)(struct amdgpu_device *adev);
+};
+
+struct amdgpu_mem_partition_info {
+	union {
+		struct {
+			uint32_t fpfn;
+			uint32_t lpfn;
+		} range;
+		struct {
+			int node;
+		} numa;
+	};
+	uint64_t size;
 };
 
-struct amdgpu_xgmi_ras_funcs {
-	int (*ras_late_init)(struct amdgpu_device *adev);
-	void (*ras_fini)(struct amdgpu_device *adev);
-	int (*query_ras_error_count)(struct amdgpu_device *adev,
-				     void *ras_error_status);
-	void (*reset_ras_error_count)(struct amdgpu_device *adev);
+#define INVALID_PFN    -1
+
+struct amdgpu_gmc_memrange {
+	uint64_t base_address;
+	uint64_t limit_address;
+	uint32_t flags;
+	int nid_mask;
 };
 
-struct amdgpu_xgmi {
-	/* from psp */
-	u64 node_id;
-	u64 hive_id;
-	/* fixed per family */
-	u64 node_segment_size;
-	/* physical node (0-3) */
-	unsigned physical_node_id;
-	/* number of nodes (0-4) */
-	unsigned num_physical_nodes;
-	/* gpu list in the same hive */
-	struct list_head head;
-	bool supported;
-	struct ras_common_if *ras_if;
-	bool connected_to_cpu;
-	bool pending_reset;
-	const struct amdgpu_xgmi_ras_funcs *ras_funcs;
+enum amdgpu_gart_placement {
+	AMDGPU_GART_PLACEMENT_BEST_FIT = 0,
+	AMDGPU_GART_PLACEMENT_HIGH,
+	AMDGPU_GART_PLACEMENT_LOW,
 };
 
 struct amdgpu_gmc {
@@ -248,30 +299,80 @@ struct amdgpu_gmc {
 	uint64_t		last_fault:AMDGPU_GMC_FAULT_RING_ORDER;
 
 	bool tmz_enabled;
+	bool is_app_apu;
 
+	struct amdgpu_mem_partition_info *mem_partitions;
+	uint8_t num_mem_partitions;
 	const struct amdgpu_gmc_funcs	*gmc_funcs;
+	enum amdgpu_memory_partition	requested_nps_mode;
+	uint32_t supported_nps_modes;
+	uint32_t reset_flags;
 
 	struct amdgpu_xgmi xgmi;
 	struct amdgpu_irq_src	ecc_irq;
 	int noretry;
+	uint32_t xnack_flags;
 
 	uint32_t	vmid0_page_table_block_size;
 	uint32_t	vmid0_page_table_depth;
 	struct amdgpu_bo		*pdb0_bo;
 	/* CPU kmapped address of pdb0*/
 	void				*ptr_pdb0;
+
+	/* MALL size */
+	u64 mall_size;
+	uint32_t m_half_use;
+
+	/* number of UMC instances */
+	int num_umc;
+	/* mode2 save restore */
+	u64 VM_L2_CNTL;
+	u64 VM_L2_CNTL2;
+	u64 VM_DUMMY_PAGE_FAULT_CNTL;
+	u64 VM_DUMMY_PAGE_FAULT_ADDR_LO32;
+	u64 VM_DUMMY_PAGE_FAULT_ADDR_HI32;
+	u64 VM_L2_PROTECTION_FAULT_CNTL;
+	u64 VM_L2_PROTECTION_FAULT_CNTL2;
+	u64 VM_L2_PROTECTION_FAULT_MM_CNTL3;
+	u64 VM_L2_PROTECTION_FAULT_MM_CNTL4;
+	u64 VM_L2_PROTECTION_FAULT_ADDR_LO32;
+	u64 VM_L2_PROTECTION_FAULT_ADDR_HI32;
+	u64 VM_DEBUG;
+	u64 VM_L2_MM_GROUP_RT_CLASSES;
+	u64 VM_L2_BANK_SELECT_RESERVED_CID;
+	u64 VM_L2_BANK_SELECT_RESERVED_CID2;
+	u64 VM_L2_CACHE_PARITY_CNTL;
+	u64 VM_L2_IH_LOG_CNTL;
+	u64 VM_CONTEXT_CNTL[16];
+	u64 VM_CONTEXT_PAGE_TABLE_BASE_ADDR_LO32[16];
+	u64 VM_CONTEXT_PAGE_TABLE_BASE_ADDR_HI32[16];
+	u64 VM_CONTEXT_PAGE_TABLE_START_ADDR_LO32[16];
+	u64 VM_CONTEXT_PAGE_TABLE_START_ADDR_HI32[16];
+	u64 VM_CONTEXT_PAGE_TABLE_END_ADDR_LO32[16];
+	u64 VM_CONTEXT_PAGE_TABLE_END_ADDR_HI32[16];
+	u64 MC_VM_MX_L1_TLB_CNTL;
+
+	u64 noretry_flags;
+
+	bool flush_tlb_needs_extra_type_0;
+	bool flush_tlb_needs_extra_type_2;
+	bool flush_pasid_uses_kiq;
 };
 
-#define amdgpu_gmc_flush_gpu_tlb(adev, vmid, vmhub, type) ((adev)->gmc.gmc_funcs->flush_gpu_tlb((adev), (vmid), (vmhub), (type)))
-#define amdgpu_gmc_flush_gpu_tlb_pasid(adev, pasid, type, allhub) \
-	((adev)->gmc.gmc_funcs->flush_gpu_tlb_pasid \
-	((adev), (pasid), (type), (allhub)))
 #define amdgpu_gmc_emit_flush_gpu_tlb(r, vmid, addr) (r)->adev->gmc.gmc_funcs->emit_flush_gpu_tlb((r), (vmid), (addr))
 #define amdgpu_gmc_emit_pasid_mapping(r, vmid, pasid) (r)->adev->gmc.gmc_funcs->emit_pasid_mapping((r), (vmid), (pasid))
-#define amdgpu_gmc_map_mtype(adev, flags) (adev)->gmc.gmc_funcs->map_mtype((adev),(flags))
 #define amdgpu_gmc_get_vm_pde(adev, level, dst, flags) (adev)->gmc.gmc_funcs->get_vm_pde((adev), (level), (dst), (flags))
-#define amdgpu_gmc_get_vm_pte(adev, mapping, flags) (adev)->gmc.gmc_funcs->get_vm_pte((adev), (mapping), (flags))
+#define amdgpu_gmc_get_vm_pte(adev, vm, bo, vm_flags, pte_flags) \
+	((adev)->gmc.gmc_funcs->get_vm_pte((adev), (vm), (bo), (vm_flags), \
+					   (pte_flags)))
+#define amdgpu_gmc_override_vm_pte_flags(adev, vm, addr, pte_flags)	\
+	(adev)->gmc.gmc_funcs->override_vm_pte_flags			\
+		((adev), (vm), (addr), (pte_flags))
 #define amdgpu_gmc_get_vbios_fb_size(adev) (adev)->gmc.gmc_funcs->get_vbios_fb_size((adev))
+#define amdgpu_gmc_get_dcc_alignment(adev) ({			\
+	typeof(adev) _adev = (adev);				\
+	_adev->gmc.gmc_funcs->get_dcc_alignment(_adev);		\
+})
 
 /**
  * amdgpu_gmc_vram_full_visible - Check if full VRAM is visible through the BAR
@@ -301,6 +402,7 @@ static inline uint64_t amdgpu_gmc_sign_extend(uint64_t addr)
 	return addr;
 }
 
+bool amdgpu_gmc_is_pdb0_enabled(struct amdgpu_device *adev);
 int amdgpu_gmc_pdb0_alloc(struct amdgpu_device *adev);
 void amdgpu_gmc_get_pde_for_bo(struct amdgpu_bo *bo, int level,
 			       uint64_t *addr, uint64_t *flags);
@@ -313,16 +415,30 @@ void amdgpu_gmc_sysvm_location(struct amdgpu_device *adev, struct amdgpu_gmc *mc
 void amdgpu_gmc_vram_location(struct amdgpu_device *adev, struct amdgpu_gmc *mc,
 			      u64 base);
 void amdgpu_gmc_gart_location(struct amdgpu_device *adev,
-			      struct amdgpu_gmc *mc);
+			      struct amdgpu_gmc *mc,
+			      enum amdgpu_gart_placement gart_placement);
 void amdgpu_gmc_agp_location(struct amdgpu_device *adev,
 			     struct amdgpu_gmc *mc);
-bool amdgpu_gmc_filter_faults(struct amdgpu_device *adev, uint64_t addr,
+void amdgpu_gmc_set_agp_default(struct amdgpu_device *adev,
+				struct amdgpu_gmc *mc);
+bool amdgpu_gmc_filter_faults(struct amdgpu_device *adev,
+			      struct amdgpu_ih_ring *ih, uint64_t addr,
 			      uint16_t pasid, uint64_t timestamp);
 void amdgpu_gmc_filter_faults_remove(struct amdgpu_device *adev, uint64_t addr,
 				     uint16_t pasid);
+int amdgpu_gmc_ras_sw_init(struct amdgpu_device *adev);
 int amdgpu_gmc_ras_late_init(struct amdgpu_device *adev);
 void amdgpu_gmc_ras_fini(struct amdgpu_device *adev);
 int amdgpu_gmc_allocate_vm_inv_eng(struct amdgpu_device *adev);
+void amdgpu_gmc_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid,
+			      uint32_t vmhub, uint32_t flush_type);
+int amdgpu_gmc_flush_gpu_tlb_pasid(struct amdgpu_device *adev, uint16_t pasid,
+				   uint32_t flush_type, bool all_hub,
+				   uint32_t inst);
+void amdgpu_gmc_fw_reg_write_reg_wait(struct amdgpu_device *adev,
+				      uint32_t reg0, uint32_t reg1,
+				      uint32_t ref, uint32_t mask,
+				      uint32_t xcc_inst);
 
 extern void amdgpu_gmc_tmz_set(struct amdgpu_device *adev);
 extern void amdgpu_gmc_noretry_set(struct amdgpu_device *adev);
@@ -332,10 +448,29 @@ amdgpu_gmc_set_vm_fault_masks(struct amdgpu_device *adev, int hub_type,
 			      bool enable);
 
 void amdgpu_gmc_get_vbios_allocations(struct amdgpu_device *adev);
-void amdgpu_gmc_get_reserved_allocation(struct amdgpu_device *adev);
 
 void amdgpu_gmc_init_pdb0(struct amdgpu_device *adev);
 uint64_t amdgpu_gmc_vram_mc2pa(struct amdgpu_device *adev, uint64_t mc_addr);
 uint64_t amdgpu_gmc_vram_pa(struct amdgpu_device *adev, struct amdgpu_bo *bo);
-uint64_t amdgpu_gmc_vram_cpu_pa(struct amdgpu_device *adev, struct amdgpu_bo *bo);
+int amdgpu_gmc_vram_checking(struct amdgpu_device *adev);
+int amdgpu_gmc_sysfs_init(struct amdgpu_device *adev);
+void amdgpu_gmc_sysfs_fini(struct amdgpu_device *adev);
+
+int amdgpu_gmc_get_nps_memranges(struct amdgpu_device *adev,
+				 struct amdgpu_mem_partition_info *mem_ranges,
+				 uint8_t *exp_ranges);
+
+int amdgpu_gmc_request_memory_partition(struct amdgpu_device *adev,
+					int nps_mode);
+void amdgpu_gmc_prepare_nps_mode_change(struct amdgpu_device *adev);
+bool amdgpu_gmc_need_reset_on_init(struct amdgpu_device *adev);
+enum amdgpu_memory_partition
+amdgpu_gmc_get_vf_memory_partition(struct amdgpu_device *adev);
+enum amdgpu_memory_partition
+amdgpu_gmc_get_memory_partition(struct amdgpu_device *adev, u32 *supp_modes);
+enum amdgpu_memory_partition
+amdgpu_gmc_query_memory_partition(struct amdgpu_device *adev);
+int amdgpu_gmc_init_mem_ranges(struct amdgpu_device *adev);
+void amdgpu_gmc_init_sw_mem_ranges(struct amdgpu_device *adev,
+				   struct amdgpu_mem_partition_info *mem_ranges);
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gtt_mgr.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gtt_mgr.c
index 675a72ef305d..895c1e4c6747 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gtt_mgr.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gtt_mgr.c
@@ -26,23 +26,12 @@
 
 #include "amdgpu.h"
 
-struct amdgpu_gtt_node {
-	struct ttm_buffer_object *tbo;
-	struct ttm_range_mgr_node base;
-};
-
 static inline struct amdgpu_gtt_mgr *
 to_gtt_mgr(struct ttm_resource_manager *man)
 {
 	return container_of(man, struct amdgpu_gtt_mgr, manager);
 }
 
-static inline struct amdgpu_gtt_node *
-to_amdgpu_gtt_node(struct ttm_resource *res)
-{
-	return container_of(res, struct amdgpu_gtt_node, base.base);
-}
-
 /**
  * DOC: mem_info_gtt_total
  *
@@ -60,7 +49,7 @@ static ssize_t amdgpu_mem_info_gtt_total_show(struct device *dev,
 	struct ttm_resource_manager *man;
 
 	man = ttm_manager_type(&adev->mman.bdev, TTM_PL_TT);
-	return sysfs_emit(buf, "%llu\n", man->size * PAGE_SIZE);
+	return sysfs_emit(buf, "%llu\n", man->size);
 }
 
 /**
@@ -77,10 +66,9 @@ static ssize_t amdgpu_mem_info_gtt_used_show(struct device *dev,
 {
 	struct drm_device *ddev = dev_get_drvdata(dev);
 	struct amdgpu_device *adev = drm_to_adev(ddev);
-	struct ttm_resource_manager *man;
+	struct ttm_resource_manager *man = &adev->mman.gtt_mgr.manager;
 
-	man = ttm_manager_type(&adev->mman.bdev, TTM_PL_TT);
-	return sysfs_emit(buf, "%llu\n", amdgpu_gtt_mgr_usage(man));
+	return sysfs_emit(buf, "%llu\n", ttm_resource_manager_usage(man));
 }
 
 static DEVICE_ATTR(mem_info_gtt_total, S_IRUGO,
@@ -107,9 +95,9 @@ const struct attribute_group amdgpu_gtt_mgr_attr_group = {
  */
 bool amdgpu_gtt_mgr_has_gart_addr(struct ttm_resource *res)
 {
-	struct amdgpu_gtt_node *node = to_amdgpu_gtt_node(res);
+	struct ttm_range_mgr_node *node = to_ttm_range_mgr_node(res);
 
-	return drm_mm_node_allocated(&node->base.mm_nodes[0]);
+	return drm_mm_node_allocated(&node->mm_nodes[0]);
 }
 
 /**
@@ -129,28 +117,23 @@ static int amdgpu_gtt_mgr_new(struct ttm_resource_manager *man,
 {
 	struct amdgpu_gtt_mgr *mgr = to_gtt_mgr(man);
 	uint32_t num_pages = PFN_UP(tbo->base.size);
-	struct amdgpu_gtt_node *node;
+	struct ttm_range_mgr_node *node;
 	int r;
 
-	if (!(place->flags & TTM_PL_FLAG_TEMPORARY) &&
-	    atomic64_add_return(num_pages, &mgr->used) >  man->size) {
-		atomic64_sub(num_pages, &mgr->used);
-		return -ENOSPC;
-	}
+	node = kzalloc(struct_size(node, mm_nodes, 1), GFP_KERNEL);
+	if (!node)
+		return -ENOMEM;
 
-	node = kzalloc(struct_size(node, base.mm_nodes, 1), GFP_KERNEL);
-	if (!node) {
-		r = -ENOMEM;
-		goto err_out;
+	ttm_resource_init(tbo, place, &node->base);
+	if (!(place->flags & TTM_PL_FLAG_TEMPORARY) &&
+	    ttm_resource_manager_usage(man) > man->size) {
+		r = -ENOSPC;
+		goto err_free;
 	}
 
-	node->tbo = tbo;
-	ttm_resource_init(tbo, place, &node->base.base);
-
 	if (place->lpfn) {
 		spin_lock(&mgr->lock);
-		r = drm_mm_insert_node_in_range(&mgr->mm,
-						&node->base.mm_nodes[0],
+		r = drm_mm_insert_node_in_range(&mgr->mm, &node->mm_nodes[0],
 						num_pages, tbo->page_alignment,
 						0, place->fpfn, place->lpfn,
 						DRM_MM_INSERT_BEST);
@@ -158,23 +141,19 @@ static int amdgpu_gtt_mgr_new(struct ttm_resource_manager *man,
 		if (unlikely(r))
 			goto err_free;
 
-		node->base.base.start = node->base.mm_nodes[0].start;
+		node->base.start = node->mm_nodes[0].start;
 	} else {
-		node->base.mm_nodes[0].start = 0;
-		node->base.mm_nodes[0].size = node->base.base.num_pages;
-		node->base.base.start = AMDGPU_BO_INVALID_OFFSET;
+		node->mm_nodes[0].start = 0;
+		node->mm_nodes[0].size = PFN_UP(node->base.size);
+		node->base.start = AMDGPU_BO_INVALID_OFFSET;
 	}
 
-	*res = &node->base.base;
+	*res = &node->base;
 	return 0;
 
 err_free:
+	ttm_resource_fini(man, &node->base);
 	kfree(node);
-
-err_out:
-	if (!(place->flags & TTM_PL_FLAG_TEMPORARY))
-		atomic64_sub(num_pages, &mgr->used);
-
 	return r;
 }
 
@@ -189,62 +168,74 @@ err_out:
 static void amdgpu_gtt_mgr_del(struct ttm_resource_manager *man,
 			       struct ttm_resource *res)
 {
-	struct amdgpu_gtt_node *node = to_amdgpu_gtt_node(res);
+	struct ttm_range_mgr_node *node = to_ttm_range_mgr_node(res);
 	struct amdgpu_gtt_mgr *mgr = to_gtt_mgr(man);
 
 	spin_lock(&mgr->lock);
-	if (drm_mm_node_allocated(&node->base.mm_nodes[0]))
-		drm_mm_remove_node(&node->base.mm_nodes[0]);
+	if (drm_mm_node_allocated(&node->mm_nodes[0]))
+		drm_mm_remove_node(&node->mm_nodes[0]);
 	spin_unlock(&mgr->lock);
 
-	if (!(res->placement & TTM_PL_FLAG_TEMPORARY))
-		atomic64_sub(res->num_pages, &mgr->used);
-
+	ttm_resource_fini(man, res);
 	kfree(node);
 }
 
 /**
- * amdgpu_gtt_mgr_usage - return usage of GTT domain
- *
- * @man: TTM memory type manager
- *
- * Return how many bytes are used in the GTT domain
- */
-uint64_t amdgpu_gtt_mgr_usage(struct ttm_resource_manager *man)
-{
-	struct amdgpu_gtt_mgr *mgr = to_gtt_mgr(man);
-
-	return atomic64_read(&mgr->used) * PAGE_SIZE;
-}
-
-/**
  * amdgpu_gtt_mgr_recover - re-init gart
  *
- * @man: TTM memory type manager
+ * @mgr: amdgpu_gtt_mgr pointer
  *
  * Re-init the gart for each known BO in the GTT.
  */
-int amdgpu_gtt_mgr_recover(struct ttm_resource_manager *man)
+void amdgpu_gtt_mgr_recover(struct amdgpu_gtt_mgr *mgr)
 {
-	struct amdgpu_gtt_mgr *mgr = to_gtt_mgr(man);
-	struct amdgpu_device *adev;
-	struct amdgpu_gtt_node *node;
+	struct ttm_range_mgr_node *node;
 	struct drm_mm_node *mm_node;
-	int r = 0;
+	struct amdgpu_device *adev;
 
 	adev = container_of(mgr, typeof(*adev), mman.gtt_mgr);
 	spin_lock(&mgr->lock);
 	drm_mm_for_each_node(mm_node, &mgr->mm) {
-		node = container_of(mm_node, typeof(*node), base.mm_nodes[0]);
-		r = amdgpu_ttm_recover_gart(node->tbo);
-		if (r)
-			break;
+		node = container_of(mm_node, typeof(*node), mm_nodes[0]);
+		amdgpu_ttm_recover_gart(node->base.bo);
 	}
 	spin_unlock(&mgr->lock);
+}
 
-	amdgpu_gart_invalidate_tlb(adev);
+/**
+ * amdgpu_gtt_mgr_intersects - test for intersection
+ *
+ * @man: Our manager object
+ * @res: The resource to test
+ * @place: The place for the new allocation
+ * @size: The size of the new allocation
+ *
+ * Simplified intersection test, only interesting if we need GART or not.
+ */
+static bool amdgpu_gtt_mgr_intersects(struct ttm_resource_manager *man,
+				      struct ttm_resource *res,
+				      const struct ttm_place *place,
+				      size_t size)
+{
+	return !place->lpfn || amdgpu_gtt_mgr_has_gart_addr(res);
+}
 
-	return r;
+/**
+ * amdgpu_gtt_mgr_compatible - test for compatibility
+ *
+ * @man: Our manager object
+ * @res: The resource to test
+ * @place: The place for the new allocation
+ * @size: The size of the new allocation
+ *
+ * Simplified compatibility test.
+ */
+static bool amdgpu_gtt_mgr_compatible(struct ttm_resource_manager *man,
+				      struct ttm_resource *res,
+				      const struct ttm_place *place,
+				      size_t size)
+{
+	return !place->lpfn || amdgpu_gtt_mgr_has_gart_addr(res);
 }
 
 /**
@@ -263,14 +254,13 @@ static void amdgpu_gtt_mgr_debug(struct ttm_resource_manager *man,
 	spin_lock(&mgr->lock);
 	drm_mm_print(&mgr->mm, printer);
 	spin_unlock(&mgr->lock);
-
-	drm_printf(printer, "man size:%llu pages,  gtt used:%llu pages\n",
-		   man->size, atomic64_read(&mgr->used));
 }
 
 static const struct ttm_resource_manager_func amdgpu_gtt_mgr_func = {
 	.alloc = amdgpu_gtt_mgr_new,
 	.free = amdgpu_gtt_mgr_del,
+	.intersects = amdgpu_gtt_mgr_intersects,
+	.compatible = amdgpu_gtt_mgr_compatible,
 	.debug = amdgpu_gtt_mgr_debug
 };
 
@@ -291,13 +281,13 @@ int amdgpu_gtt_mgr_init(struct amdgpu_device *adev, uint64_t gtt_size)
 	man->use_tt = true;
 	man->func = &amdgpu_gtt_mgr_func;
 
-	ttm_resource_manager_init(man, gtt_size >> PAGE_SHIFT);
+	ttm_resource_manager_init(man, &adev->mman.bdev, gtt_size);
 
 	start = AMDGPU_GTT_MAX_TRANSFER_SIZE * AMDGPU_GTT_NUM_TRANSFER_WINDOWS;
+	start += amdgpu_vce_required_gart_pages(adev);
 	size = (adev->gmc.gart_size >> PAGE_SHIFT) - start;
 	drm_mm_init(&mgr->mm, start, size);
 	spin_lock_init(&mgr->lock);
-	atomic64_set(&mgr->used, 0);
 
 	ttm_set_driver_manager(&adev->mman.bdev, TTM_PL_TT, &mgr->manager);
 	ttm_resource_manager_set_used(man, true);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_hdp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_hdp.c
index a766e1aad2b9..5a60d69a3e1f 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_hdp.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_hdp.c
@@ -1,5 +1,5 @@
 /*
- * Copyright 2021 Advanced Micro Devices, Inc.
+ * Copyright 2023 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -20,49 +20,65 @@
  * OTHER DEALINGS IN THE SOFTWARE.
  *
  */
-
 #include "amdgpu.h"
 #include "amdgpu_ras.h"
+#include <uapi/linux/kfd_ioctl.h>
 
-int amdgpu_hdp_ras_late_init(struct amdgpu_device *adev)
+int amdgpu_hdp_ras_sw_init(struct amdgpu_device *adev)
 {
-	int r;
-	struct ras_ih_if ih_info = {
-		.cb = NULL,
-	};
-	struct ras_fs_if fs_info = {
-		.sysfs_name = "hdp_err_count",
-	};
+	int err;
+	struct amdgpu_hdp_ras *ras;
 
-	if (!adev->hdp.ras_if) {
-		adev->hdp.ras_if = kmalloc(sizeof(struct ras_common_if), GFP_KERNEL);
-		if (!adev->hdp.ras_if)
-			return -ENOMEM;
-		adev->hdp.ras_if->block = AMDGPU_RAS_BLOCK__HDP;
-		adev->hdp.ras_if->type = AMDGPU_RAS_ERROR__MULTI_UNCORRECTABLE;
-		adev->hdp.ras_if->sub_block_index = 0;
-	}
-	ih_info.head = fs_info.head = *adev->hdp.ras_if;
-	r = amdgpu_ras_late_init(adev, adev->hdp.ras_if,
-				 &fs_info, &ih_info);
-	if (r || !amdgpu_ras_is_supported(adev, adev->hdp.ras_if->block)) {
-		kfree(adev->hdp.ras_if);
-		adev->hdp.ras_if = NULL;
+	if (!adev->hdp.ras)
+		return 0;
+
+	ras = adev->hdp.ras;
+	err = amdgpu_ras_register_ras_block(adev, &ras->ras_block);
+	if (err) {
+		dev_err(adev->dev, "Failed to register hdp ras block!\n");
+		return err;
 	}
 
-	return r;
+	strcpy(ras->ras_block.ras_comm.name, "hdp");
+	ras->ras_block.ras_comm.block = AMDGPU_RAS_BLOCK__HDP;
+	ras->ras_block.ras_comm.type = AMDGPU_RAS_ERROR__MULTI_UNCORRECTABLE;
+	adev->hdp.ras_if = &ras->ras_block.ras_comm;
+
+	/* hdp ras follows amdgpu_ras_block_late_init_default for late init */
+	return 0;
 }
 
-void amdgpu_hdp_ras_fini(struct amdgpu_device *adev)
+void amdgpu_hdp_generic_flush(struct amdgpu_device *adev,
+			      struct amdgpu_ring *ring)
 {
-	if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__HDP) &&
-	    adev->hdp.ras_if) {
-		struct ras_common_if *ras_if = adev->hdp.ras_if;
-		struct ras_ih_if ih_info = {
-			.cb = NULL,
-		};
-
-		amdgpu_ras_late_fini(adev, ras_if, &ih_info);
-		kfree(ras_if);
+	if (!ring || !ring->funcs->emit_wreg) {
+		WREG32((adev->rmmio_remap.reg_offset +
+			KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >>
+			       2,
+		       0);
+		if (adev->nbio.funcs->get_memsize)
+			adev->nbio.funcs->get_memsize(adev);
+	} else {
+		amdgpu_ring_emit_wreg(ring,
+				      (adev->rmmio_remap.reg_offset +
+				       KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >>
+					      2,
+				      0);
 	}
 }
+
+void amdgpu_hdp_invalidate(struct amdgpu_device *adev, struct amdgpu_ring *ring)
+{
+	if (adev->asic_funcs && adev->asic_funcs->invalidate_hdp)
+		adev->asic_funcs->invalidate_hdp(adev, ring);
+	else if (adev->hdp.funcs && adev->hdp.funcs->invalidate_hdp)
+		adev->hdp.funcs->invalidate_hdp(adev, ring);
+}
+
+void amdgpu_hdp_flush(struct amdgpu_device *adev, struct amdgpu_ring *ring)
+{
+	if (adev->asic_funcs && adev->asic_funcs->flush_hdp)
+		adev->asic_funcs->flush_hdp(adev, ring);
+	else if (adev->hdp.funcs && adev->hdp.funcs->flush_hdp)
+		adev->hdp.funcs->flush_hdp(adev, ring);
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_hdp.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_hdp.h
index 7ec99d591584..d9f488fa76b9 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_hdp.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_hdp.h
@@ -22,13 +22,10 @@
  */
 #ifndef __AMDGPU_HDP_H__
 #define __AMDGPU_HDP_H__
+#include "amdgpu_ras.h"
 
-struct amdgpu_hdp_ras_funcs {
-	int (*ras_late_init)(struct amdgpu_device *adev);
-	void (*ras_fini)(struct amdgpu_device *adev);
-	void (*query_ras_error_count)(struct amdgpu_device *adev,
-				      void *ras_error_status);
-	void (*reset_ras_error_count)(struct amdgpu_device *adev);
+struct amdgpu_hdp_ras {
+	struct amdgpu_ras_block_object ras_block;
 };
 
 struct amdgpu_hdp_funcs {
@@ -36,16 +33,21 @@ struct amdgpu_hdp_funcs {
 	void (*invalidate_hdp)(struct amdgpu_device *adev,
 			       struct amdgpu_ring *ring);
 	void (*update_clock_gating)(struct amdgpu_device *adev, bool enable);
-	void (*get_clock_gating_state)(struct amdgpu_device *adev, u32 *flags);
+	void (*get_clock_gating_state)(struct amdgpu_device *adev, u64 *flags);
 	void (*init_registers)(struct amdgpu_device *adev);
 };
 
 struct amdgpu_hdp {
 	struct ras_common_if			*ras_if;
 	const struct amdgpu_hdp_funcs		*funcs;
-	const struct amdgpu_hdp_ras_funcs	*ras_funcs;
+	struct amdgpu_hdp_ras	*ras;
 };
 
-int amdgpu_hdp_ras_late_init(struct amdgpu_device *adev);
-void amdgpu_hdp_ras_fini(struct amdgpu_device *adev);
+int amdgpu_hdp_ras_sw_init(struct amdgpu_device *adev);
+void amdgpu_hdp_generic_flush(struct amdgpu_device *adev,
+			      struct amdgpu_ring *ring);
+void amdgpu_hdp_invalidate(struct amdgpu_device *adev,
+			   struct amdgpu_ring *ring);
+void amdgpu_hdp_flush(struct amdgpu_device *adev,
+		      struct amdgpu_ring *ring);
 #endif /* __AMDGPU_HDP_H__ */
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_hmm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_hmm.c
new file mode 100644
index 000000000000..90d26d820bac
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_hmm.c
@@ -0,0 +1,293 @@
+/*
+ * Copyright 2014 Advanced Micro Devices, Inc.
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
+ * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
+ * USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ */
+/*
+ * Authors:
+ *    Christian König <christian.koenig@amd.com>
+ */
+
+/**
+ * DOC: MMU Notifier
+ *
+ * For coherent userptr handling registers an MMU notifier to inform the driver
+ * about updates on the page tables of a process.
+ *
+ * When somebody tries to invalidate the page tables we block the update until
+ * all operations on the pages in question are completed, then those pages are
+ * marked as accessed and also dirty if it wasn't a read only access.
+ *
+ * New command submissions using the userptrs in question are delayed until all
+ * page table invalidation are completed and we once more see a coherent process
+ * address space.
+ */
+
+#include <linux/firmware.h>
+#include <linux/module.h>
+#include <drm/drm.h>
+
+#include "amdgpu.h"
+#include "amdgpu_amdkfd.h"
+#include "amdgpu_hmm.h"
+
+#define MAX_WALK_BYTE	(2UL << 30)
+
+/**
+ * amdgpu_hmm_invalidate_gfx - callback to notify about mm change
+ *
+ * @mni: the range (mm) is about to update
+ * @range: details on the invalidation
+ * @cur_seq: Value to pass to mmu_interval_set_seq()
+ *
+ * Block for operations on BOs to finish and mark pages as accessed and
+ * potentially dirty.
+ */
+static bool amdgpu_hmm_invalidate_gfx(struct mmu_interval_notifier *mni,
+				      const struct mmu_notifier_range *range,
+				      unsigned long cur_seq)
+{
+	struct amdgpu_bo *bo = container_of(mni, struct amdgpu_bo, notifier);
+	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
+	long r;
+
+	if (!mmu_notifier_range_blockable(range))
+		return false;
+
+	mutex_lock(&adev->notifier_lock);
+
+	mmu_interval_set_seq(mni, cur_seq);
+
+	r = dma_resv_wait_timeout(bo->tbo.base.resv, DMA_RESV_USAGE_BOOKKEEP,
+				  false, MAX_SCHEDULE_TIMEOUT);
+	mutex_unlock(&adev->notifier_lock);
+	if (r <= 0)
+		DRM_ERROR("(%ld) failed to wait for user bo\n", r);
+	return true;
+}
+
+static const struct mmu_interval_notifier_ops amdgpu_hmm_gfx_ops = {
+	.invalidate = amdgpu_hmm_invalidate_gfx,
+};
+
+/**
+ * amdgpu_hmm_invalidate_hsa - callback to notify about mm change
+ *
+ * @mni: the range (mm) is about to update
+ * @range: details on the invalidation
+ * @cur_seq: Value to pass to mmu_interval_set_seq()
+ *
+ * We temporarily evict the BO attached to this range. This necessitates
+ * evicting all user-mode queues of the process.
+ */
+static bool amdgpu_hmm_invalidate_hsa(struct mmu_interval_notifier *mni,
+				      const struct mmu_notifier_range *range,
+				      unsigned long cur_seq)
+{
+	struct amdgpu_bo *bo = container_of(mni, struct amdgpu_bo, notifier);
+
+	if (!mmu_notifier_range_blockable(range))
+		return false;
+
+	amdgpu_amdkfd_evict_userptr(mni, cur_seq, bo->kfd_bo);
+
+	return true;
+}
+
+static const struct mmu_interval_notifier_ops amdgpu_hmm_hsa_ops = {
+	.invalidate = amdgpu_hmm_invalidate_hsa,
+};
+
+/**
+ * amdgpu_hmm_register - register a BO for notifier updates
+ *
+ * @bo: amdgpu buffer object
+ * @addr: userptr addr we should monitor
+ *
+ * Registers a mmu_notifier for the given BO at the specified address.
+ * Returns 0 on success, -ERRNO if anything goes wrong.
+ */
+int amdgpu_hmm_register(struct amdgpu_bo *bo, unsigned long addr)
+{
+	int r;
+
+	if (bo->kfd_bo)
+		r = mmu_interval_notifier_insert(&bo->notifier, current->mm,
+						    addr, amdgpu_bo_size(bo),
+						    &amdgpu_hmm_hsa_ops);
+	else
+		r = mmu_interval_notifier_insert(&bo->notifier, current->mm, addr,
+							amdgpu_bo_size(bo),
+							&amdgpu_hmm_gfx_ops);
+	if (r)
+		/*
+		 * Make sure amdgpu_hmm_unregister() doesn't call
+		 * mmu_interval_notifier_remove() when the notifier isn't properly
+		 * initialized.
+		 */
+		bo->notifier.mm = NULL;
+
+	return r;
+}
+
+/**
+ * amdgpu_hmm_unregister - unregister a BO for notifier updates
+ *
+ * @bo: amdgpu buffer object
+ *
+ * Remove any registration of mmu notifier updates from the buffer object.
+ */
+void amdgpu_hmm_unregister(struct amdgpu_bo *bo)
+{
+	if (!bo->notifier.mm)
+		return;
+	mmu_interval_notifier_remove(&bo->notifier);
+	bo->notifier.mm = NULL;
+}
+
+int amdgpu_hmm_range_get_pages(struct mmu_interval_notifier *notifier,
+			       uint64_t start, uint64_t npages, bool readonly,
+			       void *owner,
+			       struct amdgpu_hmm_range *range)
+{
+	unsigned long end;
+	unsigned long timeout;
+	unsigned long *pfns;
+	int r = 0;
+	struct hmm_range *hmm_range = &range->hmm_range;
+
+	pfns = kvmalloc_array(npages, sizeof(*pfns), GFP_KERNEL);
+	if (unlikely(!pfns)) {
+		r = -ENOMEM;
+		goto out_free_range;
+	}
+
+	hmm_range->notifier = notifier;
+	hmm_range->default_flags = HMM_PFN_REQ_FAULT;
+	if (!readonly)
+		hmm_range->default_flags |= HMM_PFN_REQ_WRITE;
+	hmm_range->hmm_pfns = pfns;
+	hmm_range->start = start;
+	end = start + npages * PAGE_SIZE;
+	hmm_range->dev_private_owner = owner;
+
+	do {
+		hmm_range->end = min(hmm_range->start + MAX_WALK_BYTE, end);
+
+		pr_debug("hmm range: start = 0x%lx, end = 0x%lx",
+			hmm_range->start, hmm_range->end);
+
+		timeout = jiffies + msecs_to_jiffies(HMM_RANGE_DEFAULT_TIMEOUT);
+
+retry:
+		hmm_range->notifier_seq = mmu_interval_read_begin(notifier);
+		r = hmm_range_fault(hmm_range);
+		if (unlikely(r)) {
+			if (r == -EBUSY && !time_after(jiffies, timeout))
+				goto retry;
+			goto out_free_pfns;
+		}
+
+		if (hmm_range->end == end)
+			break;
+		hmm_range->hmm_pfns += MAX_WALK_BYTE >> PAGE_SHIFT;
+		hmm_range->start = hmm_range->end;
+	} while (hmm_range->end < end);
+
+	hmm_range->start = start;
+	hmm_range->hmm_pfns = pfns;
+
+	return 0;
+
+out_free_pfns:
+	kvfree(pfns);
+	hmm_range->hmm_pfns = NULL;
+out_free_range:
+	if (r == -EBUSY)
+		r = -EAGAIN;
+	return r;
+}
+
+/**
+ * amdgpu_hmm_range_valid - check if an HMM range is still valid
+ * @range: pointer to the &struct amdgpu_hmm_range to validate
+ *
+ * Determines whether the given HMM range @range is still valid by
+ * checking for invalidations via the MMU notifier sequence. This is
+ * typically used to verify that the range has not been invalidated
+ * by concurrent address space updates before it is accessed.
+ *
+ * Return:
+ * * true if @range is valid and can be used safely
+ * * false if @range is NULL or has been invalidated
+ */
+bool amdgpu_hmm_range_valid(struct amdgpu_hmm_range *range)
+{
+	if (!range)
+		return false;
+
+	return !mmu_interval_read_retry(range->hmm_range.notifier,
+					range->hmm_range.notifier_seq);
+}
+
+/**
+ * amdgpu_hmm_range_alloc - allocate and initialize an AMDGPU HMM range
+ * @bo: optional buffer object to associate with this HMM range
+ *
+ * Allocates memory for amdgpu_hmm_range and associates it with the @bo passed.
+ * The reference count of the @bo is incremented.
+ *
+ * Return:
+ * Pointer to a newly allocated struct amdgpu_hmm_range on success,
+ * or NULL if memory allocation fails.
+ */
+struct amdgpu_hmm_range *amdgpu_hmm_range_alloc(struct amdgpu_bo *bo)
+{
+	struct amdgpu_hmm_range *range;
+
+	range = kzalloc(sizeof(*range), GFP_KERNEL);
+	if (!range)
+		return NULL;
+
+	range->bo = amdgpu_bo_ref(bo);
+	return range;
+}
+
+/**
+ * amdgpu_hmm_range_free - release an AMDGPU HMM range
+ * @range: pointer to the range object to free
+ *
+ * Releases all resources held by @range, including the associated
+ * hmm_pfns and the dropping reference of associated bo if any.
+ *
+ * Return: void
+ */
+void amdgpu_hmm_range_free(struct amdgpu_hmm_range *range)
+{
+	if (!range)
+		return;
+
+	kvfree(range->hmm_range.hmm_pfns);
+	amdgpu_bo_unref(&range->bo);
+	kfree(range);
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_hmm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_hmm.h
new file mode 100644
index 000000000000..140bc9cd57b4
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_hmm.h
@@ -0,0 +1,73 @@
+/*
+ * Copyright 2017 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Christian König
+ */
+#ifndef __AMDGPU_MN_H__
+#define __AMDGPU_MN_H__
+
+#include <linux/types.h>
+#include <linux/hmm.h>
+#include <linux/rwsem.h>
+#include <linux/workqueue.h>
+#include <linux/interval_tree.h>
+#include <linux/mmu_notifier.h>
+
+struct amdgpu_hmm_range {
+	struct hmm_range hmm_range;
+	struct amdgpu_bo *bo;
+};
+
+int amdgpu_hmm_range_get_pages(struct mmu_interval_notifier *notifier,
+			       uint64_t start, uint64_t npages, bool readonly,
+			       void *owner,
+			       struct amdgpu_hmm_range *range);
+
+#if defined(CONFIG_HMM_MIRROR)
+bool amdgpu_hmm_range_valid(struct amdgpu_hmm_range *range);
+struct amdgpu_hmm_range *amdgpu_hmm_range_alloc(struct amdgpu_bo *bo);
+void amdgpu_hmm_range_free(struct amdgpu_hmm_range *range);
+int amdgpu_hmm_register(struct amdgpu_bo *bo, unsigned long addr);
+void amdgpu_hmm_unregister(struct amdgpu_bo *bo);
+#else
+static inline int amdgpu_hmm_register(struct amdgpu_bo *bo, unsigned long addr)
+{
+	DRM_WARN_ONCE("HMM_MIRROR kernel config option is not enabled, "
+		      "add CONFIG_ZONE_DEVICE=y in config file to fix this\n");
+	return -ENODEV;
+}
+
+static inline void amdgpu_hmm_unregister(struct amdgpu_bo *bo) {}
+
+static inline bool amdgpu_hmm_range_valid(struct amdgpu_hmm_range *range)
+{
+	return false;
+}
+
+static inline struct amdgpu_hmm_range *amdgpu_hmm_range_alloc(struct amdgpu_bo *bo)
+{
+	return NULL;
+}
+
+static inline void amdgpu_hmm_range_free(struct amdgpu_hmm_range *range) {}
+#endif
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_i2c.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_i2c.c
index 82608df43396..9cb72f0c5277 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_i2c.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_i2c.c
@@ -24,7 +24,6 @@
  *          Alex Deucher
  */
 
-#include <linux/export.h>
 #include <linux/pci.h>
 
 #include <drm/drm_edid.h>
@@ -175,7 +174,6 @@ struct amdgpu_i2c_chan *amdgpu_i2c_create(struct drm_device *dev,
 
 	i2c->rec = *rec;
 	i2c->adapter.owner = THIS_MODULE;
-	i2c->adapter.class = I2C_CLASS_DDC;
 	i2c->adapter.dev.parent = dev->dev;
 	i2c->dev = dev;
 	i2c_set_adapdata(&i2c->adapter, i2c);
@@ -186,7 +184,7 @@ struct amdgpu_i2c_chan *amdgpu_i2c_create(struct drm_device *dev,
 		snprintf(i2c->adapter.name, sizeof(i2c->adapter.name),
 			 "AMDGPU i2c hw bus %s", name);
 		i2c->adapter.algo = &amdgpu_atombios_i2c_algo;
-		ret = i2c_add_adapter(&i2c->adapter);
+		ret = devm_i2c_add_adapter(dev->dev, &i2c->adapter);
 		if (ret)
 			goto out_free;
 	} else {
@@ -217,22 +215,23 @@ out_free:
 
 }
 
-void amdgpu_i2c_destroy(struct amdgpu_i2c_chan *i2c)
-{
-	if (!i2c)
-		return;
-	WARN_ON(i2c->has_aux);
-	i2c_del_adapter(&i2c->adapter);
-	kfree(i2c);
-}
-
-/* Add the default buses */
 void amdgpu_i2c_init(struct amdgpu_device *adev)
 {
-	if (amdgpu_hw_i2c)
-		DRM_INFO("hw_i2c forced on, you may experience display detection problems!\n");
-
-	amdgpu_atombios_i2c_init(adev);
+	if (!adev->is_atom_fw) {
+		if (!amdgpu_device_has_dc_support(adev)) {
+			amdgpu_atombios_i2c_init(adev);
+		} else {
+			switch (adev->asic_type) {
+			case CHIP_POLARIS10:
+			case CHIP_POLARIS11:
+			case CHIP_POLARIS12:
+				amdgpu_atombios_oem_i2c_init(adev, 0x97);
+				break;
+			default:
+				break;
+			}
+		}
+	}
 }
 
 /* remove all the buses */
@@ -240,28 +239,9 @@ void amdgpu_i2c_fini(struct amdgpu_device *adev)
 {
 	int i;
 
-	for (i = 0; i < AMDGPU_MAX_I2C_BUS; i++) {
-		if (adev->i2c_bus[i]) {
-			amdgpu_i2c_destroy(adev->i2c_bus[i]);
+	for (i = 0; i < AMDGPU_MAX_I2C_BUS; i++)
+		if (adev->i2c_bus[i])
 			adev->i2c_bus[i] = NULL;
-		}
-	}
-}
-
-/* Add additional buses */
-void amdgpu_i2c_add(struct amdgpu_device *adev,
-		    const struct amdgpu_i2c_bus_rec *rec,
-		    const char *name)
-{
-	struct drm_device *dev = adev_to_drm(adev);
-	int i;
-
-	for (i = 0; i < AMDGPU_MAX_I2C_BUS; i++) {
-		if (!adev->i2c_bus[i]) {
-			adev->i2c_bus[i] = amdgpu_i2c_create(dev, rec, name);
-			return;
-		}
-	}
 }
 
 /* looks up bus based on id */
@@ -280,7 +260,7 @@ amdgpu_i2c_lookup(struct amdgpu_device *adev,
 	return NULL;
 }
 
-static void amdgpu_i2c_get_byte(struct amdgpu_i2c_chan *i2c_bus,
+static int amdgpu_i2c_get_byte(struct amdgpu_i2c_chan *i2c_bus,
 				 u8 slave_addr,
 				 u8 addr,
 				 u8 *val)
@@ -305,16 +285,18 @@ static void amdgpu_i2c_get_byte(struct amdgpu_i2c_chan *i2c_bus,
 	out_buf[0] = addr;
 	out_buf[1] = 0;
 
-	if (i2c_transfer(&i2c_bus->adapter, msgs, 2) == 2) {
-		*val = in_buf[0];
-		DRM_DEBUG("val = 0x%02x\n", *val);
-	} else {
-		DRM_DEBUG("i2c 0x%02x 0x%02x read failed\n",
-			  addr, *val);
+	if (i2c_transfer(&i2c_bus->adapter, msgs, 2) != 2) {
+		DRM_DEBUG("i2c 0x%02x read failed\n", addr);
+		return -EIO;
 	}
+
+	*val = in_buf[0];
+	DRM_DEBUG("val = 0x%02x\n", *val);
+
+	return 0;
 }
 
-static void amdgpu_i2c_put_byte(struct amdgpu_i2c_chan *i2c_bus,
+static int amdgpu_i2c_put_byte(struct amdgpu_i2c_chan *i2c_bus,
 				 u8 slave_addr,
 				 u8 addr,
 				 u8 val)
@@ -330,9 +312,12 @@ static void amdgpu_i2c_put_byte(struct amdgpu_i2c_chan *i2c_bus,
 	out_buf[0] = addr;
 	out_buf[1] = val;
 
-	if (i2c_transfer(&i2c_bus->adapter, &msg, 1) != 1)
-		DRM_DEBUG("i2c 0x%02x 0x%02x write failed\n",
-			  addr, val);
+	if (i2c_transfer(&i2c_bus->adapter, &msg, 1) != 1) {
+		DRM_DEBUG("i2c 0x%02x 0x%02x write failed\n", addr, val);
+		return -EIO;
+	}
+
+	return 0;
 }
 
 /* ddc router switching */
@@ -347,16 +332,18 @@ amdgpu_i2c_router_select_ddc_port(const struct amdgpu_connector *amdgpu_connecto
 	if (!amdgpu_connector->router_bus)
 		return;
 
-	amdgpu_i2c_get_byte(amdgpu_connector->router_bus,
+	if (amdgpu_i2c_get_byte(amdgpu_connector->router_bus,
 			    amdgpu_connector->router.i2c_addr,
-			    0x3, &val);
+			    0x3, &val))
+		return;
 	val &= ~amdgpu_connector->router.ddc_mux_control_pin;
 	amdgpu_i2c_put_byte(amdgpu_connector->router_bus,
 			    amdgpu_connector->router.i2c_addr,
 			    0x3, val);
-	amdgpu_i2c_get_byte(amdgpu_connector->router_bus,
+	if (amdgpu_i2c_get_byte(amdgpu_connector->router_bus,
 			    amdgpu_connector->router.i2c_addr,
-			    0x1, &val);
+			    0x1, &val))
+		return;
 	val &= ~amdgpu_connector->router.ddc_mux_control_pin;
 	val |= amdgpu_connector->router.ddc_mux_state;
 	amdgpu_i2c_put_byte(amdgpu_connector->router_bus,
@@ -376,16 +363,18 @@ amdgpu_i2c_router_select_cd_port(const struct amdgpu_connector *amdgpu_connector
 	if (!amdgpu_connector->router_bus)
 		return;
 
-	amdgpu_i2c_get_byte(amdgpu_connector->router_bus,
+	if (amdgpu_i2c_get_byte(amdgpu_connector->router_bus,
 			    amdgpu_connector->router.i2c_addr,
-			    0x3, &val);
+			    0x3, &val))
+		return;
 	val &= ~amdgpu_connector->router.cd_mux_control_pin;
 	amdgpu_i2c_put_byte(amdgpu_connector->router_bus,
 			    amdgpu_connector->router.i2c_addr,
 			    0x3, val);
-	amdgpu_i2c_get_byte(amdgpu_connector->router_bus,
+	if (amdgpu_i2c_get_byte(amdgpu_connector->router_bus,
 			    amdgpu_connector->router.i2c_addr,
-			    0x1, &val);
+			    0x1, &val))
+		return;
 	val &= ~amdgpu_connector->router.cd_mux_control_pin;
 	val |= amdgpu_connector->router.cd_mux_state;
 	amdgpu_i2c_put_byte(amdgpu_connector->router_bus,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_i2c.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_i2c.h
index 63c2ff7499e1..1d3d3806e0dd 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_i2c.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_i2c.h
@@ -30,9 +30,6 @@ struct amdgpu_i2c_chan *amdgpu_i2c_create(struct drm_device *dev,
 void amdgpu_i2c_destroy(struct amdgpu_i2c_chan *i2c);
 void amdgpu_i2c_init(struct amdgpu_device *adev);
 void amdgpu_i2c_fini(struct amdgpu_device *adev);
-void amdgpu_i2c_add(struct amdgpu_device *adev,
-		    const struct amdgpu_i2c_bus_rec *rec,
-		    const char *name);
 struct amdgpu_i2c_chan *
 amdgpu_i2c_lookup(struct amdgpu_device *adev,
 		  const struct amdgpu_i2c_bus_rec *i2c_bus);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
index bc1297dcdf97..586a58facca1 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
@@ -62,14 +62,14 @@
  * Returns 0 on success, error on failure.
  */
 int amdgpu_ib_get(struct amdgpu_device *adev, struct amdgpu_vm *vm,
-		  unsigned size, enum amdgpu_ib_pool_type pool_type,
+		  unsigned int size, enum amdgpu_ib_pool_type pool_type,
 		  struct amdgpu_ib *ib)
 {
 	int r;
 
 	if (size) {
 		r = amdgpu_sa_bo_new(&adev->ib_pools[pool_type],
-				      &ib->sa_bo, size, 256);
+				     &ib->sa_bo, size);
 		if (r) {
 			dev_err(adev->dev, "failed to get a new IB (%d)\n", r);
 			return r;
@@ -89,16 +89,14 @@ int amdgpu_ib_get(struct amdgpu_device *adev, struct amdgpu_vm *vm,
 /**
  * amdgpu_ib_free - free an IB (Indirect Buffer)
  *
- * @adev: amdgpu_device pointer
  * @ib: IB object to free
  * @f: the fence SA bo need wait on for the ib alloation
  *
  * Free an IB (all asics).
  */
-void amdgpu_ib_free(struct amdgpu_device *adev, struct amdgpu_ib *ib,
-		    struct dma_fence *f)
+void amdgpu_ib_free(struct amdgpu_ib *ib, struct dma_fence *f)
 {
-	amdgpu_sa_bo_free(adev, &ib->sa_bo, f);
+	amdgpu_sa_bo_free(&ib->sa_bo, f);
 }
 
 /**
@@ -123,24 +121,26 @@ void amdgpu_ib_free(struct amdgpu_device *adev, struct amdgpu_ib *ib,
  * a CONST_IB), it will be put on the ring prior to the DE IB.  Prior
  * to SI there was just a DE IB.
  */
-int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs,
+int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned int num_ibs,
 		       struct amdgpu_ib *ibs, struct amdgpu_job *job,
 		       struct dma_fence **f)
 {
 	struct amdgpu_device *adev = ring->adev;
 	struct amdgpu_ib *ib = &ibs[0];
 	struct dma_fence *tmp = NULL;
+	struct amdgpu_fence *af;
 	bool need_ctx_switch;
-	unsigned patch_offset = ~0;
 	struct amdgpu_vm *vm;
 	uint64_t fence_ctx;
 	uint32_t status = 0, alloc_size;
-	unsigned fence_flags = 0;
-	bool secure;
-
-	unsigned i;
-	int r = 0;
+	unsigned int fence_flags = 0;
+	bool secure, init_shadow;
+	u64 shadow_va, csa_va, gds_va;
+	int vmid = AMDGPU_JOB_GET_VMID(job);
 	bool need_pipe_sync = false;
+	unsigned int cond_exec;
+	unsigned int i;
+	int r = 0;
 
 	if (num_ibs == 0)
 		return -EINVAL;
@@ -149,26 +149,48 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs,
 	if (job) {
 		vm = job->vm;
 		fence_ctx = job->base.s_fence ?
-			job->base.s_fence->scheduled.context : 0;
+			job->base.s_fence->finished.context : 0;
+		shadow_va = job->shadow_va;
+		csa_va = job->csa_va;
+		gds_va = job->gds_va;
+		init_shadow = job->init_shadow;
+		af = job->hw_fence;
+		/* Save the context of the job for reset handling.
+		 * The driver needs this so it can skip the ring
+		 * contents for guilty contexts.
+		 */
+		af->context = fence_ctx;
+		/* the vm fence is also part of the job's context */
+		job->hw_vm_fence->context = fence_ctx;
 	} else {
 		vm = NULL;
 		fence_ctx = 0;
+		shadow_va = 0;
+		csa_va = 0;
+		gds_va = 0;
+		init_shadow = false;
+		af = kzalloc(sizeof(*af), GFP_ATOMIC);
+		if (!af)
+			return -ENOMEM;
 	}
 
 	if (!ring->sched.ready) {
 		dev_err(adev->dev, "couldn't schedule ib on ring <%s>\n", ring->name);
-		return -EINVAL;
+		r = -EINVAL;
+		goto free_fence;
 	}
 
 	if (vm && !job->vmid) {
 		dev_err(adev->dev, "VM IB without ID\n");
-		return -EINVAL;
+		r = -EINVAL;
+		goto free_fence;
 	}
 
 	if ((ib->flags & AMDGPU_IB_FLAGS_SECURE) &&
-	    (ring->funcs->type == AMDGPU_RING_TYPE_COMPUTE)) {
-		dev_err(adev->dev, "secure submissions not supported on compute rings\n");
-		return -EINVAL;
+	    (!ring->funcs->secure_submission_supported)) {
+		dev_err(adev->dev, "secure submissions not supported on ring <%s>\n", ring->name);
+		r = -EINVAL;
+		goto free_fence;
 	}
 
 	alloc_size = ring->funcs->emit_frame_size + num_ibs *
@@ -177,14 +199,14 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs,
 	r = amdgpu_ring_alloc(ring, alloc_size);
 	if (r) {
 		dev_err(adev->dev, "scheduling IB failed (%d).\n", r);
-		return r;
+		goto free_fence;
 	}
 
 	need_ctx_switch = ring->current_ctx != fence_ctx;
 	if (ring->funcs->emit_pipeline_sync && job &&
-	    ((tmp = amdgpu_sync_get_fence(&job->sched_sync)) ||
-	     (amdgpu_sriov_vf(adev) && need_ctx_switch) ||
-	     amdgpu_vm_need_pipeline_sync(ring, job))) {
+	    ((tmp = amdgpu_sync_get_fence(&job->explicit_sync)) ||
+	     need_ctx_switch || amdgpu_vm_need_pipeline_sync(ring, job))) {
+
 		need_pipe_sync = true;
 
 		if (tmp)
@@ -211,8 +233,15 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs,
 		}
 	}
 
-	if (job && ring->funcs->init_cond_exec)
-		patch_offset = amdgpu_ring_init_cond_exec(ring);
+	amdgpu_ring_ib_begin(ring);
+
+	if (ring->funcs->emit_gfx_shadow)
+		amdgpu_ring_emit_gfx_shadow(ring, shadow_va, csa_va, gds_va,
+					    init_shadow, vmid);
+
+	if (ring->funcs->init_cond_exec)
+		cond_exec = amdgpu_ring_init_cond_exec(ring,
+						       ring->cond_exe_gpu_addr);
 
 	amdgpu_device_flush_hdp(adev, ring);
 
@@ -262,31 +291,54 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs,
 				       fence_flags | AMDGPU_FENCE_FLAG_64BIT);
 	}
 
-	r = amdgpu_fence_emit(ring, f, job, fence_flags);
+	if (ring->funcs->emit_gfx_shadow && ring->funcs->init_cond_exec) {
+		amdgpu_ring_emit_gfx_shadow(ring, 0, 0, 0, false, 0);
+		amdgpu_ring_init_cond_exec(ring, ring->cond_exe_gpu_addr);
+	}
+
+	r = amdgpu_fence_emit(ring, af, fence_flags);
 	if (r) {
 		dev_err(adev->dev, "failed to emit fence (%d)\n", r);
 		if (job && job->vmid)
-			amdgpu_vmid_reset(adev, ring->funcs->vmhub, job->vmid);
+			amdgpu_vmid_reset(adev, ring->vm_hub, job->vmid);
 		amdgpu_ring_undo(ring);
 		return r;
 	}
+	*f = &af->base;
+	/* get a ref for the job */
+	if (job)
+		dma_fence_get(*f);
 
 	if (ring->funcs->insert_end)
 		ring->funcs->insert_end(ring);
 
-	if (patch_offset != ~0 && ring->funcs->patch_cond_exec)
-		amdgpu_ring_patch_cond_exec(ring, patch_offset);
+	amdgpu_ring_patch_cond_exec(ring, cond_exec);
 
 	ring->current_ctx = fence_ctx;
-	if (vm && ring->funcs->emit_switch_buffer)
+	if (job && ring->funcs->emit_switch_buffer)
 		amdgpu_ring_emit_switch_buffer(ring);
 
 	if (ring->funcs->emit_wave_limit &&
 	    ring->hw_prio == AMDGPU_GFX_PIPE_PRIO_HIGH)
 		ring->funcs->emit_wave_limit(ring, false);
 
+	/* Save the wptr associated with this fence.
+	 * This must be last for resets to work properly
+	 * as we need to save the wptr associated with this
+	 * fence so we know what rings contents to backup
+	 * after we reset the queue.
+	 */
+	amdgpu_fence_save_wptr(af);
+
+	amdgpu_ring_ib_end(ring);
 	amdgpu_ring_commit(ring);
+
 	return 0;
+
+free_fence:
+	if (!job)
+		kfree(af);
+	return r;
 }
 
 /**
@@ -307,8 +359,7 @@ int amdgpu_ib_pool_init(struct amdgpu_device *adev)
 
 	for (i = 0; i < AMDGPU_IB_POOL_MAX; i++) {
 		r = amdgpu_sa_bo_manager_init(adev, &adev->ib_pools[i],
-					      AMDGPU_IB_POOL_SIZE,
-					      AMDGPU_GPU_PAGE_SIZE,
+					      AMDGPU_IB_POOL_SIZE, 256,
 					      AMDGPU_GEM_DOMAIN_GTT);
 		if (r)
 			goto error;
@@ -357,7 +408,7 @@ int amdgpu_ib_ring_tests(struct amdgpu_device *adev)
 {
 	long tmo_gfx, tmo_mm;
 	int r, ret = 0;
-	unsigned i;
+	unsigned int i;
 
 	tmo_mm = tmo_gfx = AMDGPU_IB_TEST_TIMEOUT;
 	if (amdgpu_sriov_vf(adev)) {
@@ -374,7 +425,7 @@ int amdgpu_ib_ring_tests(struct amdgpu_device *adev)
 		/* for CP & SDMA engines since they are scheduled together so
 		 * need to make the timeout width enough to cover the time
 		 * cost waiting for it coming back under RUNTIME only
-		*/
+		 */
 		tmo_gfx = 8 * AMDGPU_IB_TEST_TIMEOUT;
 	} else if (adev->gmc.xgmi.hive_id) {
 		tmo_gfx = AMDGPU_IB_TEST_GFX_XGMI_TIMEOUT;
@@ -390,6 +441,10 @@ int amdgpu_ib_ring_tests(struct amdgpu_device *adev)
 		if (!ring->sched.ready || !ring->funcs->test_ib)
 			continue;
 
+		if (adev->enable_mes &&
+		    ring->funcs->type == AMDGPU_RING_TYPE_KIQ)
+			continue;
+
 		/* MM engine need more time */
 		if (ring->funcs->type == AMDGPU_RING_TYPE_UVD ||
 			ring->funcs->type == AMDGPU_RING_TYPE_VCE ||
@@ -431,15 +486,15 @@ int amdgpu_ib_ring_tests(struct amdgpu_device *adev)
 
 static int amdgpu_debugfs_sa_info_show(struct seq_file *m, void *unused)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)m->private;
+	struct amdgpu_device *adev = m->private;
 
-	seq_printf(m, "--------------------- DELAYED --------------------- \n");
+	seq_puts(m, "--------------------- DELAYED ---------------------\n");
 	amdgpu_sa_bo_dump_debug_info(&adev->ib_pools[AMDGPU_IB_POOL_DELAYED],
 				     m);
-	seq_printf(m, "-------------------- IMMEDIATE -------------------- \n");
+	seq_puts(m, "-------------------- IMMEDIATE --------------------\n");
 	amdgpu_sa_bo_dump_debug_info(&adev->ib_pools[AMDGPU_IB_POOL_IMMEDIATE],
 				     m);
-	seq_printf(m, "--------------------- DIRECT ---------------------- \n");
+	seq_puts(m, "--------------------- DIRECT ----------------------\n");
 	amdgpu_sa_bo_dump_debug_info(&adev->ib_pools[AMDGPU_IB_POOL_DIRECT], m);
 
 	return 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c
index b7fb72bff2c1..9cab36322c16 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c
@@ -62,9 +62,8 @@ int amdgpu_pasid_alloc(unsigned int bits)
 	int pasid = -EINVAL;
 
 	for (bits = min(bits, 31U); bits > 0; bits--) {
-		pasid = ida_simple_get(&amdgpu_pasid_ida,
-				       1U << (bits - 1), 1U << bits,
-				       GFP_KERNEL);
+		pasid = ida_alloc_range(&amdgpu_pasid_ida, 1U << (bits - 1),
+					(1U << bits) - 1, GFP_KERNEL);
 		if (pasid != -ENOSPC)
 			break;
 	}
@@ -82,7 +81,7 @@ int amdgpu_pasid_alloc(unsigned int bits)
 void amdgpu_pasid_free(u32 pasid)
 {
 	trace_amdgpu_pasid_freed(pasid);
-	ida_simple_remove(&amdgpu_pasid_ida, pasid);
+	ida_free(&amdgpu_pasid_ida, pasid);
 }
 
 static void amdgpu_pasid_free_cb(struct dma_fence *fence,
@@ -107,36 +106,19 @@ static void amdgpu_pasid_free_cb(struct dma_fence *fence,
 void amdgpu_pasid_free_delayed(struct dma_resv *resv,
 			       u32 pasid)
 {
-	struct dma_fence *fence, **fences;
 	struct amdgpu_pasid_cb *cb;
-	unsigned count;
+	struct dma_fence *fence;
 	int r;
 
-	r = dma_resv_get_fences(resv, NULL, &count, &fences);
+	r = dma_resv_get_singleton(resv, DMA_RESV_USAGE_BOOKKEEP, &fence);
 	if (r)
 		goto fallback;
 
-	if (count == 0) {
+	if (!fence) {
 		amdgpu_pasid_free(pasid);
 		return;
 	}
 
-	if (count == 1) {
-		fence = fences[0];
-		kfree(fences);
-	} else {
-		uint64_t context = dma_fence_context_alloc(1);
-		struct dma_fence_array *array;
-
-		array = dma_fence_array_create(count, fences, context,
-					       1, false);
-		if (!array) {
-			kfree(fences);
-			goto fallback;
-		}
-		fence = &array->base;
-	}
-
 	cb = kmalloc(sizeof(*cb), GFP_KERNEL);
 	if (!cb) {
 		/* Last resort when we are OOM */
@@ -156,7 +138,8 @@ fallback:
 	/* Not enough memory for the delayed delete, as last resort
 	 * block for all the fences to complete.
 	 */
-	dma_resv_wait_timeout(resv, true, false, MAX_SCHEDULE_TIMEOUT);
+	dma_resv_wait_timeout(resv, DMA_RESV_USAGE_BOOKKEEP,
+			      false, MAX_SCHEDULE_TIMEOUT);
 	amdgpu_pasid_free(pasid);
 }
 
@@ -181,76 +164,71 @@ bool amdgpu_vmid_had_gpu_reset(struct amdgpu_device *adev,
 		atomic_read(&adev->gpu_reset_counter);
 }
 
+/* Check if we need to switch to another set of resources */
+static bool amdgpu_vmid_gds_switch_needed(struct amdgpu_vmid *id,
+					  struct amdgpu_job *job)
+{
+	return id->gds_base != job->gds_base ||
+		id->gds_size != job->gds_size ||
+		id->gws_base != job->gws_base ||
+		id->gws_size != job->gws_size ||
+		id->oa_base != job->oa_base ||
+		id->oa_size != job->oa_size;
+}
+
+/* Check if the id is compatible with the job */
+static bool amdgpu_vmid_compatible(struct amdgpu_vmid *id,
+				   struct amdgpu_job *job)
+{
+	return  id->pd_gpu_addr == job->vm_pd_addr &&
+		!amdgpu_vmid_gds_switch_needed(id, job);
+}
+
 /**
  * amdgpu_vmid_grab_idle - grab idle VMID
  *
- * @vm: vm to allocate id for
  * @ring: ring we want to submit job to
- * @sync: sync object where we add dependencies
  * @idle: resulting idle VMID
+ * @fence: fence to wait for if no id could be grabbed
  *
  * Try to find an idle VMID, if none is idle add a fence to wait to the sync
  * object. Returns -ENOMEM when we are out of memory.
  */
-static int amdgpu_vmid_grab_idle(struct amdgpu_vm *vm,
-				 struct amdgpu_ring *ring,
-				 struct amdgpu_sync *sync,
-				 struct amdgpu_vmid **idle)
+static int amdgpu_vmid_grab_idle(struct amdgpu_ring *ring,
+				 struct amdgpu_vmid **idle,
+				 struct dma_fence **fence)
 {
 	struct amdgpu_device *adev = ring->adev;
-	unsigned vmhub = ring->funcs->vmhub;
+	unsigned vmhub = ring->vm_hub;
 	struct amdgpu_vmid_mgr *id_mgr = &adev->vm_manager.id_mgr[vmhub];
-	struct dma_fence **fences;
-	unsigned i;
-	int r;
-
-	if (ring->vmid_wait && !dma_fence_is_signaled(ring->vmid_wait))
-		return amdgpu_sync_fence(sync, ring->vmid_wait);
 
-	fences = kmalloc_array(id_mgr->num_ids, sizeof(void *), GFP_KERNEL);
-	if (!fences)
-		return -ENOMEM;
+	/* If anybody is waiting for a VMID let everybody wait for fairness */
+	if (!dma_fence_is_signaled(ring->vmid_wait)) {
+		*fence = dma_fence_get(ring->vmid_wait);
+		return 0;
+	}
 
 	/* Check if we have an idle VMID */
-	i = 0;
-	list_for_each_entry((*idle), &id_mgr->ids_lru, list) {
+	list_for_each_entry_reverse((*idle), &id_mgr->ids_lru, list) {
 		/* Don't use per engine and per process VMID at the same time */
 		struct amdgpu_ring *r = adev->vm_manager.concurrent_flush ?
 			NULL : ring;
 
-		fences[i] = amdgpu_sync_peek_fence(&(*idle)->active, r);
-		if (!fences[i])
-			break;
-		++i;
+		*fence = amdgpu_sync_peek_fence(&(*idle)->active, r);
+		if (!(*fence))
+			return 0;
 	}
 
-	/* If we can't find a idle VMID to use, wait till one becomes available */
-	if (&(*idle)->list == &id_mgr->ids_lru) {
-		u64 fence_context = adev->vm_manager.fence_context + ring->idx;
-		unsigned seqno = ++adev->vm_manager.seqno[ring->idx];
-		struct dma_fence_array *array;
-		unsigned j;
-
-		*idle = NULL;
-		for (j = 0; j < i; ++j)
-			dma_fence_get(fences[j]);
-
-		array = dma_fence_array_create(i, fences, fence_context,
-					       seqno, true);
-		if (!array) {
-			for (j = 0; j < i; ++j)
-				dma_fence_put(fences[j]);
-			kfree(fences);
-			return -ENOMEM;
-		}
-
-		r = amdgpu_sync_fence(sync, &array->base);
-		dma_fence_put(ring->vmid_wait);
-		ring->vmid_wait = &array->base;
-		return r;
-	}
-	kfree(fences);
+	/*
+	 * If we can't find a idle VMID to use, wait on a fence from the least
+	 * recently used in the hope that it will be available soon.
+	 */
+	*idle = NULL;
+	dma_fence_put(ring->vmid_wait);
+	ring->vmid_wait = dma_fence_get(*fence);
 
+	/* This is the reference we return */
+	dma_fence_get(*fence);
 	return 0;
 }
 
@@ -259,41 +237,41 @@ static int amdgpu_vmid_grab_idle(struct amdgpu_vm *vm,
  *
  * @vm: vm to allocate id for
  * @ring: ring we want to submit job to
- * @sync: sync object where we add dependencies
- * @fence: fence protecting ID from reuse
  * @job: job who wants to use the VMID
  * @id: resulting VMID
+ * @fence: fence to wait for if no id could be grabbed
  *
  * Try to assign a reserved VMID.
  */
 static int amdgpu_vmid_grab_reserved(struct amdgpu_vm *vm,
 				     struct amdgpu_ring *ring,
-				     struct amdgpu_sync *sync,
-				     struct dma_fence *fence,
 				     struct amdgpu_job *job,
-				     struct amdgpu_vmid **id)
+				     struct amdgpu_vmid **id,
+				     struct dma_fence **fence)
 {
 	struct amdgpu_device *adev = ring->adev;
-	unsigned vmhub = ring->funcs->vmhub;
+	unsigned vmhub = ring->vm_hub;
 	uint64_t fence_context = adev->fence_context + ring->idx;
-	struct dma_fence *updates = sync->last_vm_update;
 	bool needs_flush = vm->use_cpu_for_update;
-	int r = 0;
+	uint64_t updates = amdgpu_vm_tlb_seq(vm);
+	int r;
 
 	*id = vm->reserved_vmid[vmhub];
-	if (updates && (*id)->flushed_updates &&
-	    updates->context == (*id)->flushed_updates->context &&
-	    !dma_fence_is_later(updates, (*id)->flushed_updates))
-		updates = NULL;
-
 	if ((*id)->owner != vm->immediate.fence_context ||
-	    job->vm_pd_addr != (*id)->pd_gpu_addr ||
-	    updates || !(*id)->last_flush ||
+	    !amdgpu_vmid_compatible(*id, job) ||
+	    (*id)->flushed_updates < updates ||
+	    !(*id)->last_flush ||
 	    ((*id)->last_flush->context != fence_context &&
-	     !dma_fence_is_signaled((*id)->last_flush))) {
+	     !dma_fence_is_signaled((*id)->last_flush)))
+		needs_flush = true;
+
+	if ((*id)->owner != vm->immediate.fence_context ||
+	    (!adev->vm_manager.concurrent_flush && needs_flush)) {
 		struct dma_fence *tmp;
 
-		/* Don't use per engine and per process VMID at the same time */
+		/* Don't use per engine and per process VMID at the
+		 * same time
+		 */
 		if (adev->vm_manager.concurrent_flush)
 			ring = NULL;
 
@@ -302,24 +280,21 @@ static int amdgpu_vmid_grab_reserved(struct amdgpu_vm *vm,
 		tmp = amdgpu_sync_peek_fence(&(*id)->active, ring);
 		if (tmp) {
 			*id = NULL;
-			r = amdgpu_sync_fence(sync, tmp);
-			return r;
+			*fence = dma_fence_get(tmp);
+			return 0;
 		}
-		needs_flush = true;
 	}
 
 	/* Good we can use this VMID. Remember this submission as
 	* user of the VMID.
 	*/
-	r = amdgpu_sync_fence(&(*id)->active, fence);
+	r = amdgpu_sync_fence(&(*id)->active, &job->base.s_fence->finished,
+			      GFP_ATOMIC);
 	if (r)
 		return r;
 
-	if (updates) {
-		dma_fence_put((*id)->flushed_updates);
-		(*id)->flushed_updates = dma_fence_get(updates);
-	}
 	job->vm_needs_flush = needs_flush;
+	job->spm_update_needed = true;
 	return 0;
 }
 
@@ -328,8 +303,6 @@ static int amdgpu_vmid_grab_reserved(struct amdgpu_vm *vm,
  *
  * @vm: vm to allocate id for
  * @ring: ring we want to submit job to
- * @sync: sync object where we add dependencies
- * @fence: fence protecting ID from reuse
  * @job: job who wants to use the VMID
  * @id: resulting VMID
  *
@@ -337,16 +310,14 @@ static int amdgpu_vmid_grab_reserved(struct amdgpu_vm *vm,
  */
 static int amdgpu_vmid_grab_used(struct amdgpu_vm *vm,
 				 struct amdgpu_ring *ring,
-				 struct amdgpu_sync *sync,
-				 struct dma_fence *fence,
 				 struct amdgpu_job *job,
 				 struct amdgpu_vmid **id)
 {
 	struct amdgpu_device *adev = ring->adev;
-	unsigned vmhub = ring->funcs->vmhub;
+	unsigned vmhub = ring->vm_hub;
 	struct amdgpu_vmid_mgr *id_mgr = &adev->vm_manager.id_mgr[vmhub];
 	uint64_t fence_context = adev->fence_context + ring->idx;
-	struct dma_fence *updates = sync->last_vm_update;
+	uint64_t updates = amdgpu_vm_tlb_seq(vm);
 	int r;
 
 	job->vm_needs_flush = vm->use_cpu_for_update;
@@ -354,13 +325,12 @@ static int amdgpu_vmid_grab_used(struct amdgpu_vm *vm,
 	/* Check if we can use a VMID already assigned to this VM */
 	list_for_each_entry_reverse((*id), &id_mgr->ids_lru, list) {
 		bool needs_flush = vm->use_cpu_for_update;
-		struct dma_fence *flushed;
 
 		/* Check all the prerequisites to using this VMID */
 		if ((*id)->owner != vm->immediate.fence_context)
 			continue;
 
-		if ((*id)->pd_gpu_addr != job->vm_pd_addr)
+		if (!amdgpu_vmid_compatible(*id, job))
 			continue;
 
 		if (!(*id)->last_flush ||
@@ -368,8 +338,7 @@ static int amdgpu_vmid_grab_used(struct amdgpu_vm *vm,
 		     !dma_fence_is_signaled((*id)->last_flush)))
 			needs_flush = true;
 
-		flushed  = (*id)->flushed_updates;
-		if (updates && (!flushed || dma_fence_is_later(updates, flushed)))
+		if ((*id)->flushed_updates < updates)
 			needs_flush = true;
 
 		if (needs_flush && !adev->vm_manager.concurrent_flush)
@@ -378,15 +347,12 @@ static int amdgpu_vmid_grab_used(struct amdgpu_vm *vm,
 		/* Good, we can use this VMID. Remember this submission as
 		 * user of the VMID.
 		 */
-		r = amdgpu_sync_fence(&(*id)->active, fence);
+		r = amdgpu_sync_fence(&(*id)->active,
+				      &job->base.s_fence->finished,
+				      GFP_ATOMIC);
 		if (r)
 			return r;
 
-		if (updates && (!flushed || dma_fence_is_later(updates, flushed))) {
-			dma_fence_put((*id)->flushed_updates);
-			(*id)->flushed_updates = dma_fence_get(updates);
-		}
-
 		job->vm_needs_flush |= needs_flush;
 		return 0;
 	}
@@ -400,65 +366,70 @@ static int amdgpu_vmid_grab_used(struct amdgpu_vm *vm,
  *
  * @vm: vm to allocate id for
  * @ring: ring we want to submit job to
- * @sync: sync object where we add dependencies
- * @fence: fence protecting ID from reuse
  * @job: job who wants to use the VMID
+ * @fence: fence to wait for if no id could be grabbed
  *
  * Allocate an id for the vm, adding fences to the sync obj as necessary.
  */
 int amdgpu_vmid_grab(struct amdgpu_vm *vm, struct amdgpu_ring *ring,
-		     struct amdgpu_sync *sync, struct dma_fence *fence,
-		     struct amdgpu_job *job)
+		     struct amdgpu_job *job, struct dma_fence **fence)
 {
 	struct amdgpu_device *adev = ring->adev;
-	unsigned vmhub = ring->funcs->vmhub;
+	unsigned vmhub = ring->vm_hub;
 	struct amdgpu_vmid_mgr *id_mgr = &adev->vm_manager.id_mgr[vmhub];
 	struct amdgpu_vmid *idle = NULL;
 	struct amdgpu_vmid *id = NULL;
 	int r = 0;
 
 	mutex_lock(&id_mgr->lock);
-	r = amdgpu_vmid_grab_idle(vm, ring, sync, &idle);
+	r = amdgpu_vmid_grab_idle(ring, &idle, fence);
 	if (r || !idle)
 		goto error;
 
-	if (vm->reserved_vmid[vmhub]) {
-		r = amdgpu_vmid_grab_reserved(vm, ring, sync, fence, job, &id);
+	if (amdgpu_vmid_uses_reserved(vm, vmhub)) {
+		r = amdgpu_vmid_grab_reserved(vm, ring, job, &id, fence);
 		if (r || !id)
 			goto error;
 	} else {
-		r = amdgpu_vmid_grab_used(vm, ring, sync, fence, job, &id);
+		r = amdgpu_vmid_grab_used(vm, ring, job, &id);
 		if (r)
 			goto error;
 
 		if (!id) {
-			struct dma_fence *updates = sync->last_vm_update;
-
 			/* Still no ID to use? Then use the idle one found earlier */
 			id = idle;
 
 			/* Remember this submission as user of the VMID */
-			r = amdgpu_sync_fence(&id->active, fence);
+			r = amdgpu_sync_fence(&id->active,
+					      &job->base.s_fence->finished,
+					      GFP_ATOMIC);
 			if (r)
 				goto error;
 
-			dma_fence_put(id->flushed_updates);
-			id->flushed_updates = dma_fence_get(updates);
 			job->vm_needs_flush = true;
 		}
 
 		list_move_tail(&id->list, &id_mgr->ids_lru);
 	}
 
-	id->pd_gpu_addr = job->vm_pd_addr;
-	id->owner = vm->immediate.fence_context;
-
+	job->gds_switch_needed = amdgpu_vmid_gds_switch_needed(id, job);
 	if (job->vm_needs_flush) {
+		id->flushed_updates = amdgpu_vm_tlb_seq(vm);
 		dma_fence_put(id->last_flush);
 		id->last_flush = NULL;
 	}
 	job->vmid = id - id_mgr->ids;
 	job->pasid = vm->pasid;
+
+	id->gds_base = job->gds_base;
+	id->gds_size = job->gds_size;
+	id->gws_base = job->gws_base;
+	id->gws_size = job->gws_size;
+	id->oa_base = job->oa_base;
+	id->oa_size = job->oa_size;
+	id->pd_gpu_addr = job->vm_pd_addr;
+	id->owner = vm->immediate.fence_context;
+
 	trace_amdgpu_vm_grab_id(vm, ring, job);
 
 error:
@@ -466,29 +437,47 @@ error:
 	return r;
 }
 
-int amdgpu_vmid_alloc_reserved(struct amdgpu_device *adev,
-			       struct amdgpu_vm *vm,
+/*
+ * amdgpu_vmid_uses_reserved - check if a VM will use a reserved VMID
+ * @vm: the VM to check
+ * @vmhub: the VMHUB which will be used
+ *
+ * Returns: True if the VM will use a reserved VMID.
+ */
+bool amdgpu_vmid_uses_reserved(struct amdgpu_vm *vm, unsigned int vmhub)
+{
+	return vm->reserved_vmid[vmhub];
+}
+
+/*
+ * amdgpu_vmid_alloc_reserved - reserve a specific VMID for this vm
+ * @adev: amdgpu device structure
+ * @vm: the VM to reserve an ID for
+ * @vmhub: the VMHUB which should be used
+ *
+ * Mostly used to have a reserved VMID for debugging and SPM.
+ *
+ * Returns: 0 for success, -ENOENT if an ID is already reserved.
+ */
+int amdgpu_vmid_alloc_reserved(struct amdgpu_device *adev, struct amdgpu_vm *vm,
 			       unsigned vmhub)
 {
-	struct amdgpu_vmid_mgr *id_mgr;
-	struct amdgpu_vmid *idle;
+	struct amdgpu_vmid_mgr *id_mgr = &adev->vm_manager.id_mgr[vmhub];
+	struct amdgpu_vmid *id;
 	int r = 0;
 
-	id_mgr = &adev->vm_manager.id_mgr[vmhub];
 	mutex_lock(&id_mgr->lock);
 	if (vm->reserved_vmid[vmhub])
 		goto unlock;
-	if (atomic_inc_return(&id_mgr->reserved_vmid_num) >
-	    AMDGPU_VM_MAX_RESERVED_VMID) {
-		DRM_ERROR("Over limitation of reserved vmid\n");
-		atomic_dec(&id_mgr->reserved_vmid_num);
-		r = -EINVAL;
+	if (id_mgr->reserved_vmid) {
+		r = -ENOENT;
 		goto unlock;
 	}
-	/* Select the first entry VMID */
-	idle = list_first_entry(&id_mgr->ids_lru, struct amdgpu_vmid, list);
-	list_del_init(&idle->list);
-	vm->reserved_vmid[vmhub] = idle;
+	/* Remove from normal round robin handling */
+	id = list_first_entry(&id_mgr->ids_lru, struct amdgpu_vmid, list);
+	list_del_init(&id->list);
+	vm->reserved_vmid[vmhub] = id;
+	id_mgr->reserved_vmid = true;
 	mutex_unlock(&id_mgr->lock);
 
 	return 0;
@@ -497,8 +486,13 @@ unlock:
 	return r;
 }
 
-void amdgpu_vmid_free_reserved(struct amdgpu_device *adev,
-			       struct amdgpu_vm *vm,
+/*
+ * amdgpu_vmid_free_reserved - free up a reserved VMID again
+ * @adev: amdgpu device structure
+ * @vm: the VM with the reserved ID
+ * @vmhub: the VMHUB which should be used
+ */
+void amdgpu_vmid_free_reserved(struct amdgpu_device *adev, struct amdgpu_vm *vm,
 			       unsigned vmhub)
 {
 	struct amdgpu_vmid_mgr *id_mgr = &adev->vm_manager.id_mgr[vmhub];
@@ -508,7 +502,7 @@ void amdgpu_vmid_free_reserved(struct amdgpu_device *adev,
 		list_add(&vm->reserved_vmid[vmhub]->list,
 			&id_mgr->ids_lru);
 		vm->reserved_vmid[vmhub] = NULL;
-		atomic_dec(&id_mgr->reserved_vmid_num);
+		id_mgr->reserved_vmid = false;
 	}
 	mutex_unlock(&id_mgr->lock);
 }
@@ -576,10 +570,17 @@ void amdgpu_vmid_mgr_init(struct amdgpu_device *adev)
 
 		mutex_init(&id_mgr->lock);
 		INIT_LIST_HEAD(&id_mgr->ids_lru);
-		atomic_set(&id_mgr->reserved_vmid_num, 0);
 
-		/* manage only VMIDs not used by KFD */
-		id_mgr->num_ids = adev->vm_manager.first_kfd_vmid;
+		/* for GC <10, SDMA uses MMHUB so use first_kfd_vmid for both GC and MM */
+		if (amdgpu_ip_version(adev, GC_HWIP, 0) < IP_VERSION(10, 0, 0))
+			/* manage only VMIDs not used by KFD */
+			id_mgr->num_ids = adev->vm_manager.first_kfd_vmid;
+		else if (AMDGPU_IS_MMHUB0(i) ||
+			 AMDGPU_IS_MMHUB1(i))
+			id_mgr->num_ids = 16;
+		else
+			/* manage only VMIDs not used by KFD */
+			id_mgr->num_ids = adev->vm_manager.first_kfd_vmid;
 
 		/* skip over VMID 0, since it is the system VM */
 		for (j = 1; j < id_mgr->num_ids; ++j) {
@@ -610,7 +611,6 @@ void amdgpu_vmid_mgr_fini(struct amdgpu_device *adev)
 			struct amdgpu_vmid *id = &id_mgr->ids[j];
 
 			amdgpu_sync_free(&id->active);
-			dma_fence_put(id->flushed_updates);
 			dma_fence_put(id->last_flush);
 			dma_fence_put(id->pasid_mapping);
 		}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.h
index 0c3b4fa1f936..b3649cd3af56 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.h
@@ -47,7 +47,7 @@ struct amdgpu_vmid {
 
 	uint64_t		pd_gpu_addr;
 	/* last flushed PD/PT update */
-	struct dma_fence	*flushed_updates;
+	uint64_t		flushed_updates;
 
 	uint32_t                current_gpu_reset_count;
 
@@ -67,7 +67,7 @@ struct amdgpu_vmid_mgr {
 	unsigned		num_ids;
 	struct list_head	ids_lru;
 	struct amdgpu_vmid	ids[AMDGPU_NUM_VMID];
-	atomic_t		reserved_vmid_num;
+	bool			reserved_vmid;
 };
 
 int amdgpu_pasid_alloc(unsigned int bits);
@@ -77,15 +77,13 @@ void amdgpu_pasid_free_delayed(struct dma_resv *resv,
 
 bool amdgpu_vmid_had_gpu_reset(struct amdgpu_device *adev,
 			       struct amdgpu_vmid *id);
-int amdgpu_vmid_alloc_reserved(struct amdgpu_device *adev,
-			       struct amdgpu_vm *vm,
+bool amdgpu_vmid_uses_reserved(struct amdgpu_vm *vm, unsigned int vmhub);
+int amdgpu_vmid_alloc_reserved(struct amdgpu_device *adev, struct amdgpu_vm *vm,
 			       unsigned vmhub);
-void amdgpu_vmid_free_reserved(struct amdgpu_device *adev,
-			       struct amdgpu_vm *vm,
+void amdgpu_vmid_free_reserved(struct amdgpu_device *adev, struct amdgpu_vm *vm,
 			       unsigned vmhub);
 int amdgpu_vmid_grab(struct amdgpu_vm *vm, struct amdgpu_ring *ring,
-		     struct amdgpu_sync *sync, struct dma_fence *fence,
-		     struct amdgpu_job *job);
+		     struct amdgpu_job *job, struct dma_fence **fence);
 void amdgpu_vmid_reset(struct amdgpu_device *adev, unsigned vmhub,
 		       unsigned vmid);
 void amdgpu_vmid_reset_all(struct amdgpu_device *adev);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c
index 0c7963dfacad..a6419246e9c2 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c
@@ -25,6 +25,7 @@
 
 #include "amdgpu.h"
 #include "amdgpu_ih.h"
+#include "amdgpu_reset.h"
 
 /**
  * amdgpu_ih_ring_init - initialize the IH state
@@ -138,6 +139,7 @@ void amdgpu_ih_ring_fini(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih)
 /**
  * amdgpu_ih_ring_write - write IV to the ring buffer
  *
+ * @adev: amdgpu_device pointer
  * @ih: ih ring to write to
  * @iv: the iv to write
  * @num_dw: size of the iv in dw
@@ -145,8 +147,8 @@ void amdgpu_ih_ring_fini(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih)
  * Writes an IV to the ring buffer using the CPU and increment the wptr.
  * Used for testing and delegating IVs to a software ring.
  */
-void amdgpu_ih_ring_write(struct amdgpu_ih_ring *ih, const uint32_t *iv,
-			  unsigned int num_dw)
+void amdgpu_ih_ring_write(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih,
+			  const uint32_t *iv, unsigned int num_dw)
 {
 	uint32_t wptr = le32_to_cpu(*ih->wptr_cpu) >> 2;
 	unsigned int i;
@@ -161,55 +163,38 @@ void amdgpu_ih_ring_write(struct amdgpu_ih_ring *ih, const uint32_t *iv,
 	if (wptr != READ_ONCE(ih->rptr)) {
 		wmb();
 		WRITE_ONCE(*ih->wptr_cpu, cpu_to_le32(wptr));
+	} else if (adev->irq.retry_cam_enabled) {
+		dev_warn_once(adev->dev, "IH soft ring buffer overflow 0x%X, 0x%X\n",
+			      wptr, ih->rptr);
 	}
 }
 
-/* Waiter helper that checks current rptr matches or passes checkpoint wptr */
-static bool amdgpu_ih_has_checkpoint_processed(struct amdgpu_device *adev,
-					struct amdgpu_ih_ring *ih,
-					uint32_t checkpoint_wptr,
-					uint32_t *prev_rptr)
-{
-	uint32_t cur_rptr = ih->rptr | (*prev_rptr & ~ih->ptr_mask);
-
-	/* rptr has wrapped. */
-	if (cur_rptr < *prev_rptr)
-		cur_rptr += ih->ptr_mask + 1;
-	*prev_rptr = cur_rptr;
-
-	/* check ring is empty to workaround missing wptr overflow flag */
-	return cur_rptr >= checkpoint_wptr ||
-	       (cur_rptr & ih->ptr_mask) == amdgpu_ih_get_wptr(adev, ih);
-}
-
 /**
- * amdgpu_ih_wait_on_checkpoint_process - wait to process IVs up to checkpoint
+ * amdgpu_ih_wait_on_checkpoint_process_ts - wait to process IVs up to checkpoint
  *
  * @adev: amdgpu_device pointer
  * @ih: ih ring to process
  *
  * Used to ensure ring has processed IVs up to the checkpoint write pointer.
  */
-int amdgpu_ih_wait_on_checkpoint_process(struct amdgpu_device *adev,
+int amdgpu_ih_wait_on_checkpoint_process_ts(struct amdgpu_device *adev,
 					struct amdgpu_ih_ring *ih)
 {
-	uint32_t checkpoint_wptr, rptr;
+	uint32_t checkpoint_wptr;
+	uint64_t checkpoint_ts;
+	long timeout = HZ;
 
 	if (!ih->enabled || adev->shutdown)
 		return -ENODEV;
 
 	checkpoint_wptr = amdgpu_ih_get_wptr(adev, ih);
-	/* Order wptr with rptr. */
+	/* Order wptr with ring data. */
 	rmb();
-	rptr = READ_ONCE(ih->rptr);
-
-	/* wptr has wrapped. */
-	if (rptr > checkpoint_wptr)
-		checkpoint_wptr += ih->ptr_mask + 1;
+	checkpoint_ts = amdgpu_ih_decode_iv_ts(adev, ih, checkpoint_wptr, -1);
 
-	return wait_event_interruptible(ih->wait_process,
-				amdgpu_ih_has_checkpoint_processed(adev, ih,
-						checkpoint_wptr, &rptr));
+	return wait_event_interruptible_timeout(ih->wait_process,
+		    amdgpu_ih_ts_after(checkpoint_ts, ih->processed_timestamp) ||
+		    ih->rptr == amdgpu_ih_get_wptr(adev, ih), timeout);
 }
 
 /**
@@ -233,7 +218,7 @@ int amdgpu_ih_process(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih)
 
 restart_ih:
 	count  = AMDGPU_IH_MAX_NUM_IVS;
-	DRM_DEBUG("%s: rptr %d, wptr %d\n", __func__, ih->rptr, wptr);
+	dev_dbg(adev->dev, "%s: rptr %d, wptr %d\n", __func__, ih->rptr, wptr);
 
 	/* Order reading of wptr vs. reading of IH ring data */
 	rmb();
@@ -243,13 +228,23 @@ restart_ih:
 		ih->rptr &= ih->ptr_mask;
 	}
 
-	amdgpu_ih_set_rptr(adev, ih);
+	if (!ih->overflow)
+		amdgpu_ih_set_rptr(adev, ih);
+
 	wake_up_all(&ih->wait_process);
 
 	/* make sure wptr hasn't changed while processing */
 	wptr = amdgpu_ih_get_wptr(adev, ih);
 	if (wptr != ih->rptr)
-		goto restart_ih;
+		if (!ih->overflow)
+			goto restart_ih;
+
+	if (ih->overflow)
+		if (amdgpu_sriov_runtime(adev))
+			WARN_ONCE(!amdgpu_reset_domain_schedule(adev->reset_domain,
+				   &adev->virt.flr_work),
+				  "Failed to queue work! at %s",
+				  __func__);
 
 	return IRQ_HANDLED;
 }
@@ -262,7 +257,7 @@ restart_ih:
  * @entry: IV entry
  *
  * Decodes the interrupt vector at the current rptr
- * position and also advance the position for for Vega10
+ * position and also advance the position for Vega10
  * and later GPUs.
  */
 void amdgpu_ih_decode_iv_helper(struct amdgpu_device *adev,
@@ -290,7 +285,7 @@ void amdgpu_ih_decode_iv_helper(struct amdgpu_device *adev,
 	entry->timestamp = dw[1] | ((u64)(dw[2] & 0xffff) << 32);
 	entry->timestamp_src = dw[2] >> 31;
 	entry->pasid = dw[3] & 0xffff;
-	entry->pasid_src = dw[3] >> 31;
+	entry->node_id = (dw[3] >> 16) & 0xff;
 	entry->src_data[0] = dw[4];
 	entry->src_data[1] = dw[5];
 	entry->src_data[2] = dw[6];
@@ -299,3 +294,24 @@ void amdgpu_ih_decode_iv_helper(struct amdgpu_device *adev,
 	/* wptr/rptr are in bytes! */
 	ih->rptr += 32;
 }
+
+uint64_t amdgpu_ih_decode_iv_ts_helper(struct amdgpu_ih_ring *ih, u32 rptr,
+				       signed int offset)
+{
+	uint32_t iv_size = 32;
+	uint32_t ring_index;
+	uint32_t dw1, dw2;
+
+	rptr += iv_size * offset;
+	ring_index = (rptr & ih->ptr_mask) >> 2;
+
+	dw1 = le32_to_cpu(ih->ring[ring_index + 1]);
+	dw2 = le32_to_cpu(ih->ring[ring_index + 2]);
+	return dw1 | ((u64)(dw2 & 0xffff) << 32);
+}
+
+const char *amdgpu_ih_ring_name(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih)
+{
+	return ih == &adev->irq.ih ? "ih" : ih == &adev->irq.ih_soft ? "sw ih" :
+	       ih == &adev->irq.ih1 ? "ih1" : ih == &adev->irq.ih2 ? "ih2" : "unknown";
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h
index 0649b59830a5..f58b6be7fccc 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h
@@ -27,6 +27,9 @@
 /* Maximum number of IVs processed at once */
 #define AMDGPU_IH_MAX_NUM_IVS	32
 
+#define IH_RING_SIZE	(256 * 1024)
+#define IH_SW_RING_SIZE	(16 * 1024)	/* enough for 512 CAM entries */
+
 struct amdgpu_device;
 struct amdgpu_iv_entry;
 
@@ -53,14 +56,14 @@ struct amdgpu_ih_ring {
 	bool			use_bus_addr;
 
 	struct amdgpu_bo	*ring_obj;
-	volatile uint32_t	*ring;
+	uint32_t		*ring;
 	uint64_t		gpu_addr;
 
 	uint64_t		wptr_addr;
-	volatile uint32_t	*wptr_cpu;
+	uint32_t		*wptr_cpu;
 
 	uint64_t		rptr_addr;
-	volatile uint32_t	*rptr_cpu;
+	uint32_t		*rptr_cpu;
 
 	bool                    enabled;
 	unsigned		rptr;
@@ -68,31 +71,48 @@ struct amdgpu_ih_ring {
 
 	/* For waiting on IH processing at checkpoint. */
 	wait_queue_head_t wait_process;
+	uint64_t		processed_timestamp;
+	bool overflow;
 };
 
+/* return true if time stamp t2 is after t1 with 48bit wrap around */
+#define amdgpu_ih_ts_after(t1, t2) \
+		(((int64_t)((t2) << 16) - (int64_t)((t1) << 16)) > 0LL)
+
+#define amdgpu_ih_ts_after_or_equal(t1, t2) \
+		(((int64_t)((t2) << 16) - (int64_t)((t1) << 16)) >= 0LL)
+
 /* provided by the ih block */
 struct amdgpu_ih_funcs {
 	/* ring read/write ptr handling, called from interrupt context */
 	u32 (*get_wptr)(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih);
 	void (*decode_iv)(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih,
 			  struct amdgpu_iv_entry *entry);
+	uint64_t (*decode_iv_ts)(struct amdgpu_ih_ring *ih, u32 rptr,
+				 signed int offset);
 	void (*set_rptr)(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih);
 };
 
 #define amdgpu_ih_get_wptr(adev, ih) (adev)->irq.ih_funcs->get_wptr((adev), (ih))
 #define amdgpu_ih_decode_iv(adev, iv) \
 	(adev)->irq.ih_funcs->decode_iv((adev), (ih), (iv))
+#define amdgpu_ih_decode_iv_ts(adev, ih, rptr, offset) \
+	(WARN_ON_ONCE(!(adev)->irq.ih_funcs->decode_iv_ts) ? 0 : \
+	(adev)->irq.ih_funcs->decode_iv_ts((ih), (rptr), (offset)))
 #define amdgpu_ih_set_rptr(adev, ih) (adev)->irq.ih_funcs->set_rptr((adev), (ih))
 
 int amdgpu_ih_ring_init(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih,
 			unsigned ring_size, bool use_bus_addr);
 void amdgpu_ih_ring_fini(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih);
-void amdgpu_ih_ring_write(struct amdgpu_ih_ring *ih, const uint32_t *iv,
-			  unsigned int num_dw);
-int amdgpu_ih_wait_on_checkpoint_process(struct amdgpu_device *adev,
-					struct amdgpu_ih_ring *ih);
+void amdgpu_ih_ring_write(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih,
+			  const uint32_t *iv, unsigned int num_dw);
+int amdgpu_ih_wait_on_checkpoint_process_ts(struct amdgpu_device *adev,
+					    struct amdgpu_ih_ring *ih);
 int amdgpu_ih_process(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih);
 void amdgpu_ih_decode_iv_helper(struct amdgpu_device *adev,
 				struct amdgpu_ih_ring *ih,
 				struct amdgpu_iv_entry *entry);
+uint64_t amdgpu_ih_decode_iv_ts_helper(struct amdgpu_ih_ring *ih, u32 rptr,
+				       signed int offset);
+const char *amdgpu_ih_ring_name(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih);
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_imu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_imu.h
new file mode 100644
index 000000000000..484e936812e4
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_imu.h
@@ -0,0 +1,58 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __AMDGPU_IMU_H__
+#define __AMDGPU_IMU_H__
+
+enum imu_work_mode {
+	DEBUG_MODE,
+	MISSION_MODE
+};
+
+struct amdgpu_imu_funcs {
+    int (*init_microcode)(struct amdgpu_device *adev);
+    int (*load_microcode)(struct amdgpu_device *adev);
+    void (*setup_imu)(struct amdgpu_device *adev);
+    int (*start_imu)(struct amdgpu_device *adev);
+    void (*program_rlc_ram)(struct amdgpu_device *adev);
+    int (*wait_for_reset_status)(struct amdgpu_device *adev);
+};
+
+struct imu_rlc_ram_golden {
+    u32 hwip;
+    u32 instance;
+    u32 segment;
+    u32 reg;
+    u32 data;
+    u32 addr_mask;
+};
+
+#define IMU_RLC_RAM_GOLDEN_VALUE(ip, inst, reg, data, addr_mask) \
+    { ip##_HWIP, inst, reg##_BASE_IDX, reg, data, addr_mask }
+
+struct amdgpu_imu {
+    const struct amdgpu_imu_funcs *funcs;
+    enum imu_work_mode mode;
+};
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ioc32.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ioc32.c
index 5cf142e849bb..a1cbd7c3deb2 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ioc32.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ioc32.c
@@ -1,4 +1,4 @@
-/**
+/*
  * \file amdgpu_ioc32.c
  *
  * 32-bit ioctl compatibility routines for the AMDGPU DRM.
@@ -37,12 +37,9 @@
 long amdgpu_kms_compat_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
 {
 	unsigned int nr = DRM_IOCTL_NR(cmd);
-	int ret;
 
 	if (nr < DRM_COMMAND_BASE)
 		return drm_compat_ioctl(filp, cmd, arg);
 
-	ret = amdgpu_drm_ioctl(filp, cmd, arg);
-
-	return ret;
+	return amdgpu_drm_ioctl(filp, cmd, arg);
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ip.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ip.c
new file mode 100644
index 000000000000..99e1cf4fc955
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ip.c
@@ -0,0 +1,96 @@
+/*
+ * Copyright 2025 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "amdgpu.h"
+#include "amdgpu_ip.h"
+
+static int8_t amdgpu_logical_to_dev_inst(struct amdgpu_device *adev,
+					 enum amd_hw_ip_block_type block,
+					 int8_t inst)
+{
+	int8_t dev_inst;
+
+	switch (block) {
+	case GC_HWIP:
+	case SDMA0_HWIP:
+	/* Both JPEG and VCN as JPEG is only alias of VCN */
+	case VCN_HWIP:
+		dev_inst = adev->ip_map.dev_inst[block][inst];
+		break;
+	default:
+		/* For rest of the IPs, no look up required.
+		 * Assume 'logical instance == physical instance' for all configs. */
+		dev_inst = inst;
+		break;
+	}
+
+	return dev_inst;
+}
+
+static uint32_t amdgpu_logical_to_dev_mask(struct amdgpu_device *adev,
+					   enum amd_hw_ip_block_type block,
+					   uint32_t mask)
+{
+	uint32_t dev_mask = 0;
+	int8_t log_inst, dev_inst;
+
+	while (mask) {
+		log_inst = ffs(mask) - 1;
+		dev_inst = amdgpu_logical_to_dev_inst(adev, block, log_inst);
+		dev_mask |= (1 << dev_inst);
+		mask &= ~(1 << log_inst);
+	}
+
+	return dev_mask;
+}
+
+static void amdgpu_populate_ip_map(struct amdgpu_device *adev,
+				   enum amd_hw_ip_block_type ip_block,
+				   uint32_t inst_mask)
+{
+	int l = 0, i;
+
+	while (inst_mask) {
+		i = ffs(inst_mask) - 1;
+		adev->ip_map.dev_inst[ip_block][l++] = i;
+		inst_mask &= ~(1 << i);
+	}
+	for (; l < HWIP_MAX_INSTANCE; l++)
+		adev->ip_map.dev_inst[ip_block][l] = -1;
+}
+
+void amdgpu_ip_map_init(struct amdgpu_device *adev)
+{
+	u32 ip_map[][2] = {
+		{ GC_HWIP, adev->gfx.xcc_mask },
+		{ SDMA0_HWIP, adev->sdma.sdma_mask },
+		{ VCN_HWIP, adev->vcn.inst_mask },
+	};
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(ip_map); ++i)
+		amdgpu_populate_ip_map(adev, ip_map[i][0], ip_map[i][1]);
+
+	adev->ip_map.logical_to_dev_inst = amdgpu_logical_to_dev_inst;
+	adev->ip_map.logical_to_dev_mask = amdgpu_logical_to_dev_mask;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ip.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ip.h
new file mode 100644
index 000000000000..2490fd322aec
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ip.h
@@ -0,0 +1,29 @@
+/*
+ * Copyright 2025 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __AMDGPU_IP_H__
+#define __AMDGPU_IP_H__
+
+void amdgpu_ip_map_init(struct amdgpu_device *adev);
+
+#endif /* __AMDGPU_IP_H__ */
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c
index cc2e0c9cfe0a..8112ffc85995 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c
@@ -45,7 +45,6 @@
 #include <linux/irq.h>
 #include <linux/pci.h>
 
-#include <drm/drm_crtc_helper.h>
 #include <drm/drm_vblank.h>
 #include <drm/amdgpu_drm.h>
 #include <drm/drm_drv.h>
@@ -100,40 +99,20 @@ const char *soc15_ih_clientid_name[] = {
 	"MP1"
 };
 
-/**
- * amdgpu_hotplug_work_func - work handler for display hotplug event
- *
- * @work: work struct pointer
- *
- * This is the hotplug event work handler (all ASICs).
- * The work gets scheduled from the IRQ handler if there
- * was a hotplug interrupt.  It walks through the connector table
- * and calls hotplug handler for each connector. After this, it sends
- * a DRM hotplug event to alert userspace.
- *
- * This design approach is required in order to defer hotplug event handling
- * from the IRQ handler to a work handler because hotplug handler has to use
- * mutexes which cannot be locked in an IRQ handler (since &mutex_lock may
- * sleep).
- */
-static void amdgpu_hotplug_work_func(struct work_struct *work)
-{
-	struct amdgpu_device *adev = container_of(work, struct amdgpu_device,
-						  hotplug_work);
-	struct drm_device *dev = adev_to_drm(adev);
-	struct drm_mode_config *mode_config = &dev->mode_config;
-	struct drm_connector *connector;
-	struct drm_connector_list_iter iter;
-
-	mutex_lock(&mode_config->mutex);
-	drm_connector_list_iter_begin(dev, &iter);
-	drm_for_each_connector_iter(connector, &iter)
-		amdgpu_connector_hotplug(connector);
-	drm_connector_list_iter_end(&iter);
-	mutex_unlock(&mode_config->mutex);
-	/* Just fire off a uevent and let userspace tell us what to do */
-	drm_helper_hpd_irq_event(dev);
-}
+const int node_id_to_phys_map[NODEID_MAX] = {
+	[AID0_NODEID] = 0,
+	[XCD0_NODEID] = 0,
+	[XCD1_NODEID] = 1,
+	[AID1_NODEID] = 1,
+	[XCD2_NODEID] = 2,
+	[XCD3_NODEID] = 3,
+	[AID2_NODEID] = 2,
+	[XCD4_NODEID] = 4,
+	[XCD5_NODEID] = 5,
+	[AID3_NODEID] = 3,
+	[XCD6_NODEID] = 6,
+	[XCD7_NODEID] = 7,
+};
 
 /**
  * amdgpu_irq_disable_all - disable *all* interrupts
@@ -145,7 +124,7 @@ static void amdgpu_hotplug_work_func(struct work_struct *work)
 void amdgpu_irq_disable_all(struct amdgpu_device *adev)
 {
 	unsigned long irqflags;
-	unsigned i, j, k;
+	unsigned int i, j, k;
 	int r;
 
 	spin_lock_irqsave(&adev->irq.lock, irqflags);
@@ -160,12 +139,12 @@ void amdgpu_irq_disable_all(struct amdgpu_device *adev)
 				continue;
 
 			for (k = 0; k < src->num_types; ++k) {
-				atomic_set(&src->enabled_types[k], 0);
 				r = src->funcs->set(adev, src, k,
 						    AMDGPU_IRQ_STATE_DISABLE);
 				if (r)
-					DRM_ERROR("error disabling interrupt (%d)\n",
-						  r);
+					dev_err(adev->dev,
+						"error disabling interrupt (%d)\n",
+						r);
 			}
 		}
 	}
@@ -193,20 +172,7 @@ static irqreturn_t amdgpu_irq_handler(int irq, void *arg)
 	if (ret == IRQ_HANDLED)
 		pm_runtime_mark_last_busy(dev->dev);
 
-	/* For the hardware that cannot enable bif ring for both ras_controller_irq
-         * and ras_err_evnet_athub_irq ih cookies, the driver has to poll status
-	 * register to check whether the interrupt is triggered or not, and properly
-	 * ack the interrupt if it is there
-	 */
-	if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__PCIE_BIF)) {
-		if (adev->nbio.ras_funcs &&
-		    adev->nbio.ras_funcs->handle_ras_controller_intr_no_bifring)
-			adev->nbio.ras_funcs->handle_ras_controller_intr_no_bifring(adev);
-
-		if (adev->nbio.ras_funcs &&
-		    adev->nbio.ras_funcs->handle_ras_err_event_athub_intr_no_bifring)
-			adev->nbio.ras_funcs->handle_ras_err_event_athub_intr_no_bifring(adev);
-	}
+	amdgpu_ras_interrupt_fatal_error_handler(adev);
 
 	return ret;
 }
@@ -277,7 +243,7 @@ static bool amdgpu_msi_ok(struct amdgpu_device *adev)
 	return true;
 }
 
-static void amdgpu_restore_msix(struct amdgpu_device *adev)
+void amdgpu_restore_msix(struct amdgpu_device *adev)
 {
 	u16 ctrl;
 
@@ -305,44 +271,29 @@ static void amdgpu_restore_msix(struct amdgpu_device *adev)
  */
 int amdgpu_irq_init(struct amdgpu_device *adev)
 {
-	int r = 0;
-	unsigned int irq;
+	unsigned int irq, flags;
+	int r;
 
 	spin_lock_init(&adev->irq.lock);
 
 	/* Enable MSI if not disabled by module parameter */
 	adev->irq.msi_enabled = false;
 
-	if (amdgpu_msi_ok(adev)) {
-		int nvec = pci_msix_vec_count(adev->pdev);
-		unsigned int flags;
+	if (!amdgpu_msi_ok(adev))
+		flags = PCI_IRQ_INTX;
+	else
+		flags = PCI_IRQ_ALL_TYPES;
 
-		if (nvec <= 0) {
-			flags = PCI_IRQ_MSI;
-		} else {
-			flags = PCI_IRQ_MSI | PCI_IRQ_MSIX;
-		}
-		/* we only need one vector */
-		nvec = pci_alloc_irq_vectors(adev->pdev, 1, 1, flags);
-		if (nvec > 0) {
-			adev->irq.msi_enabled = true;
-			dev_dbg(adev->dev, "using MSI/MSI-X.\n");
-		}
+	/* we only need one vector */
+	r = pci_alloc_irq_vectors(adev->pdev, 1, 1, flags);
+	if (r < 0) {
+		dev_err(adev->dev, "Failed to alloc msi vectors\n");
+		return r;
 	}
 
-	if (!amdgpu_device_has_dc_support(adev)) {
-		if (!adev->enable_virtual_display)
-			/* Disable vblank IRQs aggressively for power-saving */
-			/* XXX: can this be enabled for DC? */
-			adev_to_drm(adev)->vblank_disable_immediate = true;
-
-		r = drm_vblank_init(adev_to_drm(adev), adev->mode_info.num_crtc);
-		if (r)
-			return r;
-
-		/* Pre-DCE11 */
-		INIT_WORK(&adev->hotplug_work,
-				amdgpu_hotplug_work_func);
+	if (amdgpu_msi_ok(adev)) {
+		adev->irq.msi_enabled = true;
+		dev_dbg(adev->dev, "using MSI/MSI-X.\n");
 	}
 
 	INIT_WORK(&adev->irq.ih1_work, amdgpu_irq_handle_ih1);
@@ -352,25 +303,29 @@ int amdgpu_irq_init(struct amdgpu_device *adev)
 	/* Use vector 0 for MSI-X. */
 	r = pci_irq_vector(adev->pdev, 0);
 	if (r < 0)
-		return r;
+		goto free_vectors;
 	irq = r;
 
 	/* PCI devices require shared interrupts. */
 	r = request_irq(irq, amdgpu_irq_handler, IRQF_SHARED, adev_to_drm(adev)->driver->name,
 			adev_to_drm(adev));
-	if (r) {
-		if (!amdgpu_device_has_dc_support(adev))
-			flush_work(&adev->hotplug_work);
-		return r;
-	}
+	if (r)
+		goto free_vectors;
+
 	adev->irq.installed = true;
 	adev->irq.irq = irq;
 	adev_to_drm(adev)->max_vblank_count = 0x00ffffff;
 
-	DRM_DEBUG("amdgpu: irq initialized.\n");
+	dev_dbg(adev->dev, "amdgpu: irq initialized.\n");
 	return 0;
-}
 
+free_vectors:
+	if (adev->irq.msi_enabled)
+		pci_free_irq_vectors(adev->pdev);
+
+	adev->irq.msi_enabled = false;
+	return r;
+}
 
 void amdgpu_irq_fini_hw(struct amdgpu_device *adev)
 {
@@ -379,9 +334,6 @@ void amdgpu_irq_fini_hw(struct amdgpu_device *adev)
 		adev->irq.installed = false;
 		if (adev->irq.msi_enabled)
 			pci_free_irq_vectors(adev->pdev);
-
-		if (!amdgpu_device_has_dc_support(adev))
-			flush_work(&adev->hotplug_work);
 	}
 
 	amdgpu_ih_ring_fini(adev, &adev->irq.ih_soft);
@@ -391,7 +343,7 @@ void amdgpu_irq_fini_hw(struct amdgpu_device *adev)
 }
 
 /**
- * amdgpu_irq_fini - shut down interrupt handling
+ * amdgpu_irq_fini_sw - shut down interrupt handling
  *
  * @adev: amdgpu device pointer
  *
@@ -401,7 +353,7 @@ void amdgpu_irq_fini_hw(struct amdgpu_device *adev)
  */
 void amdgpu_irq_fini_sw(struct amdgpu_device *adev)
 {
-	unsigned i, j;
+	unsigned int i, j;
 
 	for (i = 0; i < AMDGPU_IRQ_CLIENTID_MAX; ++i) {
 		if (!adev->irq.client[i].sources)
@@ -435,7 +387,7 @@ void amdgpu_irq_fini_sw(struct amdgpu_device *adev)
  * 0 on success or error code otherwise
  */
 int amdgpu_irq_add_id(struct amdgpu_device *adev,
-		      unsigned client_id, unsigned src_id,
+		      unsigned int client_id, unsigned int src_id,
 		      struct amdgpu_irq_src *source)
 {
 	if (client_id >= AMDGPU_IRQ_CLIENTID_MAX)
@@ -487,13 +439,21 @@ void amdgpu_irq_dispatch(struct amdgpu_device *adev,
 {
 	u32 ring_index = ih->rptr >> 2;
 	struct amdgpu_iv_entry entry;
-	unsigned client_id, src_id;
+	unsigned int client_id, src_id;
 	struct amdgpu_irq_src *src;
 	bool handled = false;
 	int r;
 
 	entry.ih = ih;
 	entry.iv_entry = (const uint32_t *)&ih->ring[ring_index];
+
+	/*
+	 * timestamp is not supported on some legacy SOCs (cik, cz, iceland,
+	 * si and tonga), so initialize timestamp and timestamp_src to 0
+	 */
+	entry.timestamp = 0;
+	entry.timestamp_src = 0;
+
 	amdgpu_ih_decode_iv(adev, &entry);
 
 	trace_amdgpu_iv(ih - &adev->irq.ih, &entry);
@@ -502,33 +462,41 @@ void amdgpu_irq_dispatch(struct amdgpu_device *adev,
 	src_id = entry.src_id;
 
 	if (client_id >= AMDGPU_IRQ_CLIENTID_MAX) {
-		DRM_DEBUG("Invalid client_id in IV: %d\n", client_id);
+		dev_dbg(adev->dev, "Invalid client_id in IV: %d\n", client_id);
 
 	} else	if (src_id >= AMDGPU_MAX_IRQ_SRC_ID) {
-		DRM_DEBUG("Invalid src_id in IV: %d\n", src_id);
+		dev_dbg(adev->dev, "Invalid src_id in IV: %d\n", src_id);
 
-	} else if ((client_id == AMDGPU_IRQ_CLIENTID_LEGACY) &&
+	} else if (((client_id == AMDGPU_IRQ_CLIENTID_LEGACY) ||
+		    (client_id == SOC15_IH_CLIENTID_ISP)) &&
 		   adev->irq.virq[src_id]) {
 		generic_handle_domain_irq(adev->irq.domain, src_id);
 
 	} else if (!adev->irq.client[client_id].sources) {
-		DRM_DEBUG("Unregistered interrupt client_id: %d src_id: %d\n",
-			  client_id, src_id);
+		dev_dbg(adev->dev,
+			"Unregistered interrupt client_id: %d src_id: %d\n",
+			client_id, src_id);
 
 	} else if ((src = adev->irq.client[client_id].sources[src_id])) {
 		r = src->funcs->process(adev, src, &entry);
 		if (r < 0)
-			DRM_ERROR("error processing interrupt (%d)\n", r);
+			dev_err(adev->dev, "error processing interrupt (%d)\n",
+				r);
 		else if (r)
 			handled = true;
 
 	} else {
-		DRM_DEBUG("Unhandled interrupt src_id: %d\n", src_id);
+		dev_dbg(adev->dev,
+			"Unregistered interrupt src_id: %d of client_id:%d\n",
+			src_id, client_id);
 	}
 
 	/* Send it to amdkfd as well if it isn't already handled */
 	if (!handled)
 		amdgpu_amdkfd_interrupt(adev, entry.iv_entry);
+
+	if (amdgpu_ih_ts_after(ih->processed_timestamp, entry.timestamp))
+		ih->processed_timestamp = entry.timestamp;
 }
 
 /**
@@ -545,7 +513,7 @@ void amdgpu_irq_delegate(struct amdgpu_device *adev,
 			 struct amdgpu_iv_entry *entry,
 			 unsigned int num_dw)
 {
-	amdgpu_ih_ring_write(&adev->irq.ih_soft, entry->iv_entry, num_dw);
+	amdgpu_ih_ring_write(adev, &adev->irq.ih_soft, entry->iv_entry, num_dw);
 	schedule_work(&adev->irq.ih_soft_work);
 }
 
@@ -559,7 +527,7 @@ void amdgpu_irq_delegate(struct amdgpu_device *adev,
  * Updates interrupt state for the specific source (all ASICs).
  */
 int amdgpu_irq_update(struct amdgpu_device *adev,
-			     struct amdgpu_irq_src *src, unsigned type)
+			     struct amdgpu_irq_src *src, unsigned int type)
 {
 	unsigned long irqflags;
 	enum amdgpu_interrupt_state state;
@@ -568,7 +536,8 @@ int amdgpu_irq_update(struct amdgpu_device *adev,
 	spin_lock_irqsave(&adev->irq.lock, irqflags);
 
 	/* We need to determine after taking the lock, otherwise
-	   we might disable just enabled interrupts again */
+	 * we might disable just enabled interrupts again
+	 */
 	if (amdgpu_irq_enabled(adev, src, type))
 		state = AMDGPU_IRQ_STATE_ENABLE;
 	else
@@ -622,7 +591,7 @@ void amdgpu_irq_gpu_reset_resume_helper(struct amdgpu_device *adev)
  * 0 on success or error code otherwise
  */
 int amdgpu_irq_get(struct amdgpu_device *adev, struct amdgpu_irq_src *src,
-		   unsigned type)
+		   unsigned int type)
 {
 	if (!adev->irq.installed)
 		return -ENOENT;
@@ -652,8 +621,12 @@ int amdgpu_irq_get(struct amdgpu_device *adev, struct amdgpu_irq_src *src,
  * 0 on success or error code otherwise
  */
 int amdgpu_irq_put(struct amdgpu_device *adev, struct amdgpu_irq_src *src,
-		   unsigned type)
+		   unsigned int type)
 {
+	/* When the threshold is reached,the interrupt source may not be enabled.return -EINVAL */
+	if (amdgpu_ras_is_rma(adev) && !amdgpu_irq_enabled(adev, src, type))
+		return -EINVAL;
+
 	if (!adev->irq.installed)
 		return -ENOENT;
 
@@ -663,6 +636,9 @@ int amdgpu_irq_put(struct amdgpu_device *adev, struct amdgpu_irq_src *src,
 	if (!src->enabled_types || !src->funcs->set)
 		return -EINVAL;
 
+	if (WARN_ON(!amdgpu_irq_enabled(adev, src, type)))
+		return -EINVAL;
+
 	if (atomic_dec_and_test(&src->enabled_types[type]))
 		return amdgpu_irq_update(adev, src, type);
 
@@ -683,7 +659,7 @@ int amdgpu_irq_put(struct amdgpu_device *adev, struct amdgpu_irq_src *src,
  * invalid parameters
  */
 bool amdgpu_irq_enabled(struct amdgpu_device *adev, struct amdgpu_irq_src *src,
-			unsigned type)
+			unsigned int type)
 {
 	if (!adev->irq.installed)
 		return false;
@@ -757,10 +733,10 @@ static const struct irq_domain_ops amdgpu_hw_irqdomain_ops = {
  */
 int amdgpu_irq_add_domain(struct amdgpu_device *adev)
 {
-	adev->irq.domain = irq_domain_add_linear(NULL, AMDGPU_MAX_IRQ_SRC_ID,
-						 &amdgpu_hw_irqdomain_ops, adev);
+	adev->irq.domain = irq_domain_create_linear(NULL, AMDGPU_MAX_IRQ_SRC_ID,
+						    &amdgpu_hw_irqdomain_ops, adev);
 	if (!adev->irq.domain) {
-		DRM_ERROR("GPU irq add domain failed\n");
+		dev_err(adev->dev, "GPU irq add domain failed\n");
 		return -ENODEV;
 	}
 
@@ -796,7 +772,7 @@ void amdgpu_irq_remove_domain(struct amdgpu_device *adev)
  * Returns:
  * Linux IRQ
  */
-unsigned amdgpu_irq_create_mapping(struct amdgpu_device *adev, unsigned src_id)
+unsigned int amdgpu_irq_create_mapping(struct amdgpu_device *adev, unsigned int src_id)
 {
 	adev->irq.virq[src_id] = irq_create_mapping(adev->irq.domain, src_id);
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.h
index e9f2c11ea416..9f0417456abd 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.h
@@ -53,7 +53,7 @@ struct amdgpu_iv_entry {
 	uint64_t timestamp;
 	unsigned timestamp_src;
 	unsigned pasid;
-	unsigned pasid_src;
+	unsigned node_id;
 	unsigned src_data[AMDGPU_IRQ_SRC_DATA_MAX_SIZE_DW];
 	const uint32_t *iv_entry;
 };
@@ -98,8 +98,28 @@ struct amdgpu_irq {
 	struct irq_domain		*domain; /* GPU irq controller domain */
 	unsigned			virq[AMDGPU_MAX_IRQ_SRC_ID];
 	uint32_t                        srbm_soft_reset;
+	u32                             retry_cam_doorbell_index;
+	bool                            retry_cam_enabled;
 };
 
+enum interrupt_node_id_per_aid {
+	AID0_NODEID = 0,
+	XCD0_NODEID = 1,
+	XCD1_NODEID = 2,
+	AID1_NODEID = 4,
+	XCD2_NODEID = 5,
+	XCD3_NODEID = 6,
+	AID2_NODEID = 8,
+	XCD4_NODEID = 9,
+	XCD5_NODEID = 10,
+	AID3_NODEID = 12,
+	XCD6_NODEID = 13,
+	XCD7_NODEID = 14,
+	NODEID_MAX,
+};
+
+extern const int node_id_to_phys_map[NODEID_MAX];
+
 void amdgpu_irq_disable_all(struct amdgpu_device *adev);
 
 int amdgpu_irq_init(struct amdgpu_device *adev);
@@ -126,5 +146,6 @@ void amdgpu_irq_gpu_reset_resume_helper(struct amdgpu_device *adev);
 int amdgpu_irq_add_domain(struct amdgpu_device *adev);
 void amdgpu_irq_remove_domain(struct amdgpu_device *adev);
 unsigned amdgpu_irq_create_mapping(struct amdgpu_device *adev, unsigned src_id);
+void amdgpu_restore_msix(struct amdgpu_device *adev);
 
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_isp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_isp.c
new file mode 100644
index 000000000000..37270c4dab8d
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_isp.c
@@ -0,0 +1,345 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright (C) 2024 Advanced Micro Devices, Inc. All rights reserved.
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
+ * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
+ * USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ */
+
+#include <linux/firmware.h>
+#include <linux/mfd/core.h>
+
+#include "amdgpu.h"
+#include "amdgpu_isp.h"
+#include "isp_v4_1_0.h"
+#include "isp_v4_1_1.h"
+
+#define ISP_MC_ADDR_ALIGN (1024 * 32)
+
+/**
+ * isp_hw_init - start and test isp block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ */
+static int isp_hw_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	struct amdgpu_isp *isp = &adev->isp;
+
+	if (isp->funcs->hw_init != NULL)
+		return isp->funcs->hw_init(isp);
+
+	return -ENODEV;
+}
+
+/**
+ * isp_hw_fini - stop the hardware block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ */
+static int isp_hw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_isp *isp = &ip_block->adev->isp;
+
+	if (isp->funcs->hw_fini != NULL)
+		return isp->funcs->hw_fini(isp);
+
+	return -ENODEV;
+}
+
+static int isp_load_fw_by_psp(struct amdgpu_device *adev)
+{
+	const struct common_firmware_header *hdr;
+	char ucode_prefix[10];
+	int r = 0;
+
+	/* get isp fw binary name and path */
+	amdgpu_ucode_ip_version_decode(adev, ISP_HWIP, ucode_prefix,
+				       sizeof(ucode_prefix));
+
+	/* read isp fw */
+	r = amdgpu_ucode_request(adev, &adev->isp.fw, AMDGPU_UCODE_OPTIONAL,
+				"amdgpu/%s.bin", ucode_prefix);
+	if (r) {
+		amdgpu_ucode_release(&adev->isp.fw);
+		return r;
+	}
+
+	hdr = (const struct common_firmware_header *)adev->isp.fw->data;
+
+	adev->firmware.ucode[AMDGPU_UCODE_ID_ISP].ucode_id =
+		AMDGPU_UCODE_ID_ISP;
+	adev->firmware.ucode[AMDGPU_UCODE_ID_ISP].fw = adev->isp.fw;
+
+	adev->firmware.fw_size +=
+		ALIGN(le32_to_cpu(hdr->ucode_size_bytes), PAGE_SIZE);
+
+	return r;
+}
+
+static int isp_early_init(struct amdgpu_ip_block *ip_block)
+{
+
+	struct amdgpu_device *adev = ip_block->adev;
+	struct amdgpu_isp *isp = &adev->isp;
+
+	switch (amdgpu_ip_version(adev, ISP_HWIP, 0)) {
+	case IP_VERSION(4, 1, 0):
+		isp_v4_1_0_set_isp_funcs(isp);
+		break;
+	case IP_VERSION(4, 1, 1):
+		isp_v4_1_1_set_isp_funcs(isp);
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	isp->adev = adev;
+	isp->parent = adev->dev;
+
+	if (isp_load_fw_by_psp(adev)) {
+		DRM_DEBUG_DRIVER("%s: isp fw load failed\n", __func__);
+		return -ENOENT;
+	}
+
+	return 0;
+}
+
+static bool isp_is_idle(struct amdgpu_ip_block *ip_block)
+{
+	return true;
+}
+
+static int isp_set_clockgating_state(struct amdgpu_ip_block *ip_block,
+				     enum amd_clockgating_state state)
+{
+	return 0;
+}
+
+static int isp_set_powergating_state(struct amdgpu_ip_block *ip_block,
+				     enum amd_powergating_state state)
+{
+	return 0;
+}
+
+static int is_valid_isp_device(struct device *isp_parent, struct device *amdgpu_dev)
+{
+	if (isp_parent != amdgpu_dev)
+		return -EINVAL;
+
+	return 0;
+}
+
+/**
+ * isp_user_buffer_alloc - create user buffer object (BO) for isp
+ *
+ * @dev: isp device handle
+ * @dmabuf: DMABUF handle for isp buffer allocated in system memory
+ * @buf_obj: GPU buffer object handle to initialize
+ * @buf_addr: GPU addr of the pinned BO to initialize
+ *
+ * Imports isp DMABUF to allocate and pin a user BO for isp internal use. It does
+ * GART alloc to generate GPU addr for BO to make it accessible through the
+ * GART aperture for ISP HW.
+ *
+ * This function is exported to allow the V4L2 isp device external to drm device
+ * to create and access the isp user BO.
+ *
+ * Returns:
+ * 0 on success, negative error code otherwise.
+ */
+int isp_user_buffer_alloc(struct device *dev, void *dmabuf,
+			  void **buf_obj, u64 *buf_addr)
+{
+	struct platform_device *ispdev = to_platform_device(dev);
+	const struct isp_platform_data *isp_pdata;
+	struct amdgpu_device *adev;
+	struct mfd_cell *mfd_cell;
+	struct amdgpu_bo *bo;
+	u64 gpu_addr;
+	int ret;
+
+	if (WARN_ON(!ispdev))
+		return -ENODEV;
+
+	if (WARN_ON(!buf_obj))
+		return -EINVAL;
+
+	if (WARN_ON(!buf_addr))
+		return -EINVAL;
+
+	mfd_cell = &ispdev->mfd_cell[0];
+	if (!mfd_cell)
+		return -ENODEV;
+
+	isp_pdata = mfd_cell->platform_data;
+	adev = isp_pdata->adev;
+
+	ret = is_valid_isp_device(ispdev->dev.parent, adev->dev);
+	if (ret)
+		return ret;
+
+	ret = amdgpu_bo_create_isp_user(adev, dmabuf,
+					AMDGPU_GEM_DOMAIN_GTT, &bo, &gpu_addr);
+	if (ret) {
+		drm_err(&adev->ddev, "failed to alloc gart user buffer (%d)", ret);
+		return ret;
+	}
+
+	*buf_obj = (void *)bo;
+	*buf_addr = gpu_addr;
+
+	return 0;
+}
+EXPORT_SYMBOL(isp_user_buffer_alloc);
+
+/**
+ * isp_user_buffer_free - free isp user buffer object (BO)
+ *
+ * @buf_obj: amdgpu isp user BO to free
+ *
+ * unpin and unref BO for isp internal use.
+ *
+ * This function is exported to allow the V4L2 isp device
+ * external to drm device to free the isp user BO.
+ */
+void isp_user_buffer_free(void *buf_obj)
+{
+	amdgpu_bo_free_isp_user(buf_obj);
+}
+EXPORT_SYMBOL(isp_user_buffer_free);
+
+/**
+ * isp_kernel_buffer_alloc - create kernel buffer object (BO) for isp
+ *
+ * @dev: isp device handle
+ * @size: size for the new BO
+ * @buf_obj: GPU BO handle to initialize
+ * @gpu_addr: GPU addr of the pinned BO
+ * @cpu_addr: CPU address mapping of BO
+ *
+ * Allocates and pins a kernel BO for internal isp firmware use.
+ *
+ * This function is exported to allow the V4L2 isp device
+ * external to drm device to create and access the kernel BO.
+ *
+ * Returns:
+ * 0 on success, negative error code otherwise.
+ */
+int isp_kernel_buffer_alloc(struct device *dev, u64 size,
+			    void **buf_obj, u64 *gpu_addr, void **cpu_addr)
+{
+	struct platform_device *ispdev = to_platform_device(dev);
+	struct amdgpu_bo **bo = (struct amdgpu_bo **)buf_obj;
+	const struct isp_platform_data *isp_pdata;
+	struct amdgpu_device *adev;
+	struct mfd_cell *mfd_cell;
+	int ret;
+
+	if (WARN_ON(!ispdev))
+		return -ENODEV;
+
+	if (WARN_ON(!buf_obj))
+		return -EINVAL;
+
+	if (WARN_ON(!gpu_addr))
+		return -EINVAL;
+
+	if (WARN_ON(!cpu_addr))
+		return -EINVAL;
+
+	mfd_cell = &ispdev->mfd_cell[0];
+	if (!mfd_cell)
+		return -ENODEV;
+
+	isp_pdata = mfd_cell->platform_data;
+	adev = isp_pdata->adev;
+
+	ret = is_valid_isp_device(ispdev->dev.parent, adev->dev);
+	if (ret)
+		return ret;
+
+	/* Ensure *bo is NULL so a new BO will be created */
+	*bo = NULL;
+	ret = amdgpu_bo_create_kernel(adev,
+				      size,
+				      ISP_MC_ADDR_ALIGN,
+				      AMDGPU_GEM_DOMAIN_GTT,
+				      bo,
+				      gpu_addr,
+				      cpu_addr);
+	if (!cpu_addr || ret) {
+		drm_err(&adev->ddev, "failed to alloc gart kernel buffer (%d)", ret);
+		return ret;
+	}
+
+	return 0;
+}
+EXPORT_SYMBOL(isp_kernel_buffer_alloc);
+
+/**
+ * isp_kernel_buffer_free - free isp kernel buffer object (BO)
+ *
+ * @buf_obj: amdgpu isp user BO to free
+ * @gpu_addr: GPU addr of isp kernel BO
+ * @cpu_addr: CPU addr of isp kernel BO
+ *
+ * unmaps and unpin a isp kernel BO.
+ *
+ * This function is exported to allow the V4L2 isp device
+ * external to drm device to free the kernel BO.
+ */
+void isp_kernel_buffer_free(void **buf_obj, u64 *gpu_addr, void **cpu_addr)
+{
+	struct amdgpu_bo **bo = (struct amdgpu_bo **)buf_obj;
+
+	amdgpu_bo_free_kernel(bo, gpu_addr, cpu_addr);
+}
+EXPORT_SYMBOL(isp_kernel_buffer_free);
+
+static const struct amd_ip_funcs isp_ip_funcs = {
+	.name = "isp_ip",
+	.early_init = isp_early_init,
+	.hw_init = isp_hw_init,
+	.hw_fini = isp_hw_fini,
+	.is_idle = isp_is_idle,
+	.set_clockgating_state = isp_set_clockgating_state,
+	.set_powergating_state = isp_set_powergating_state,
+};
+
+const struct amdgpu_ip_block_version isp_v4_1_0_ip_block = {
+	.type = AMD_IP_BLOCK_TYPE_ISP,
+	.major = 4,
+	.minor = 1,
+	.rev = 0,
+	.funcs = &isp_ip_funcs,
+};
+
+const struct amdgpu_ip_block_version isp_v4_1_1_ip_block = {
+	.type = AMD_IP_BLOCK_TYPE_ISP,
+	.major = 4,
+	.minor = 1,
+	.rev = 1,
+	.funcs = &isp_ip_funcs,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_isp.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_isp.h
new file mode 100644
index 000000000000..d6f4ffa4c97c
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_isp.h
@@ -0,0 +1,60 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright (C) 2024 Advanced Micro Devices, Inc. All rights reserved.
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
+ * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
+ * USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ */
+
+#ifndef __AMDGPU_ISP_H__
+#define __AMDGPU_ISP_H__
+
+#include <drm/amd/isp.h>
+#include <linux/pm_domain.h>
+
+#define ISP_REGS_OFFSET_END 0x629A4
+
+struct amdgpu_isp;
+
+struct isp_funcs {
+	int (*hw_init)(struct amdgpu_isp *isp);
+	int (*hw_fini)(struct amdgpu_isp *isp);
+};
+
+struct amdgpu_isp {
+	struct device *parent;
+	struct amdgpu_device	*adev;
+	const struct isp_funcs	*funcs;
+	struct mfd_cell *isp_cell;
+	struct resource *isp_res;
+	struct resource *isp_i2c_res;
+	struct resource *isp_gpio_res;
+	struct isp_platform_data *isp_pdata;
+	unsigned int harvest_config;
+	const struct firmware	*fw;
+	struct generic_pm_domain ispgpd;
+};
+
+extern const struct amdgpu_ip_block_version isp_v4_1_0_ip_block;
+extern const struct amdgpu_ip_block_version isp_v4_1_1_ip_block;
+
+#endif /* __AMDGPU_ISP_H__ */
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
index bfc47bea23db..0a0dcbf0798d 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
@@ -29,41 +29,144 @@
 
 #include "amdgpu.h"
 #include "amdgpu_trace.h"
+#include "amdgpu_reset.h"
+#include "amdgpu_dev_coredump.h"
+#include "amdgpu_xgmi.h"
+
+static void amdgpu_job_do_core_dump(struct amdgpu_device *adev,
+				    struct amdgpu_job *job)
+{
+	int i;
+
+	dev_info(adev->dev, "Dumping IP State\n");
+	for (i = 0; i < adev->num_ip_blocks; i++)
+		if (adev->ip_blocks[i].version->funcs->dump_ip_state)
+			adev->ip_blocks[i].version->funcs
+				->dump_ip_state((void *)&adev->ip_blocks[i]);
+	dev_info(adev->dev, "Dumping IP State Completed\n");
+
+	amdgpu_coredump(adev, true, false, job);
+}
+
+static void amdgpu_job_core_dump(struct amdgpu_device *adev,
+				 struct amdgpu_job *job)
+{
+	struct list_head device_list, *device_list_handle =  NULL;
+	struct amdgpu_device *tmp_adev = NULL;
+	struct amdgpu_hive_info *hive = NULL;
+
+	if (!amdgpu_sriov_vf(adev))
+		hive = amdgpu_get_xgmi_hive(adev);
+	if (hive)
+		mutex_lock(&hive->hive_lock);
+	/*
+	 * Reuse the logic in amdgpu_device_gpu_recover() to build list of
+	 * devices for code dump
+	 */
+	INIT_LIST_HEAD(&device_list);
+	if (!amdgpu_sriov_vf(adev) && (adev->gmc.xgmi.num_physical_nodes > 1) && hive) {
+		list_for_each_entry(tmp_adev, &hive->device_list, gmc.xgmi.head)
+			list_add_tail(&tmp_adev->reset_list, &device_list);
+		if (!list_is_first(&adev->reset_list, &device_list))
+			list_rotate_to_front(&adev->reset_list, &device_list);
+		device_list_handle = &device_list;
+	} else {
+		list_add_tail(&adev->reset_list, &device_list);
+		device_list_handle = &device_list;
+	}
+
+	/* Do the coredump for each device */
+	list_for_each_entry(tmp_adev, device_list_handle, reset_list)
+		amdgpu_job_do_core_dump(tmp_adev, job);
+
+	if (hive) {
+		mutex_unlock(&hive->hive_lock);
+		amdgpu_put_xgmi_hive(hive);
+	}
+}
 
 static enum drm_gpu_sched_stat amdgpu_job_timedout(struct drm_sched_job *s_job)
 {
 	struct amdgpu_ring *ring = to_amdgpu_ring(s_job->sched);
 	struct amdgpu_job *job = to_amdgpu_job(s_job);
-	struct amdgpu_task_info ti;
+	struct drm_wedge_task_info *info = NULL;
+	struct amdgpu_task_info *ti = NULL;
 	struct amdgpu_device *adev = ring->adev;
-	int idx;
+	int idx, r;
 
 	if (!drm_dev_enter(adev_to_drm(adev), &idx)) {
-		DRM_INFO("%s - device unplugged skipping recovery on scheduler:%s",
+		dev_info(adev->dev, "%s - device unplugged skipping recovery on scheduler:%s",
 			 __func__, s_job->sched->name);
 
 		/* Effectively the job is aborted as the device is gone */
 		return DRM_GPU_SCHED_STAT_ENODEV;
 	}
 
-	memset(&ti, 0, sizeof(struct amdgpu_task_info));
+	/*
+	 * Do the coredump immediately after a job timeout to get a very
+	 * close dump/snapshot/representation of GPU's current error status
+	 * Skip it for SRIOV, since VF FLR will be triggered by host driver
+	 * before job timeout
+	 */
+	if (!amdgpu_sriov_vf(adev))
+		amdgpu_job_core_dump(adev, job);
 
 	if (amdgpu_gpu_recovery &&
+	    amdgpu_ring_is_reset_type_supported(ring, AMDGPU_RESET_TYPE_SOFT_RESET) &&
 	    amdgpu_ring_soft_recovery(ring, job->vmid, s_job->s_fence->parent)) {
-		DRM_ERROR("ring %s timeout, but soft recovered\n",
-			  s_job->sched->name);
+		dev_err(adev->dev, "ring %s timeout, but soft recovered\n",
+			s_job->sched->name);
 		goto exit;
 	}
 
-	amdgpu_vm_get_task_info(ring->adev, job->pasid, &ti);
-	DRM_ERROR("ring %s timeout, signaled seq=%u, emitted seq=%u\n",
-		  job->base.sched->name, atomic_read(&ring->fence_drv.last_seq),
-		  ring->fence_drv.sync_seq);
-	DRM_ERROR("Process information: process %s pid %d thread %s pid %d\n",
-		  ti.process_name, ti.tgid, ti.task_name, ti.pid);
+	dev_err(adev->dev, "ring %s timeout, signaled seq=%u, emitted seq=%u\n",
+		job->base.sched->name, atomic_read(&ring->fence_drv.last_seq),
+		ring->fence_drv.sync_seq);
+
+	ti = amdgpu_vm_get_task_info_pasid(ring->adev, job->pasid);
+	if (ti) {
+		amdgpu_vm_print_task_info(adev, ti);
+		info = &ti->task;
+	}
+
+	/* attempt a per ring reset */
+	if (amdgpu_gpu_recovery &&
+	    amdgpu_ring_is_reset_type_supported(ring, AMDGPU_RESET_TYPE_PER_QUEUE) &&
+	    ring->funcs->reset) {
+		dev_err(adev->dev, "Starting %s ring reset\n",
+			s_job->sched->name);
+		r = amdgpu_ring_reset(ring, job->vmid, job->hw_fence);
+		if (!r) {
+			atomic_inc(&ring->adev->gpu_reset_counter);
+			dev_err(adev->dev, "Ring %s reset succeeded\n",
+				ring->sched.name);
+			drm_dev_wedged_event(adev_to_drm(adev),
+					     DRM_WEDGE_RECOVERY_NONE, info);
+			goto exit;
+		}
+		dev_err(adev->dev, "Ring %s reset failed\n", ring->sched.name);
+	}
+
+	dma_fence_set_error(&s_job->s_fence->finished, -ETIME);
 
 	if (amdgpu_device_should_recover_gpu(ring->adev)) {
-		amdgpu_device_gpu_recover(ring->adev, job);
+		struct amdgpu_reset_context reset_context;
+		memset(&reset_context, 0, sizeof(reset_context));
+
+		reset_context.method = AMD_RESET_METHOD_NONE;
+		reset_context.reset_req_dev = adev;
+		reset_context.src = AMDGPU_RESET_SRC_JOB;
+		clear_bit(AMDGPU_NEED_FULL_RESET, &reset_context.flags);
+
+		/*
+		 * To avoid an unnecessary extra coredump, as we have already
+		 * got the very close representation of GPU's error status
+		 */
+		set_bit(AMDGPU_SKIP_COREDUMP, &reset_context.flags);
+
+		r = amdgpu_device_gpu_recover(ring->adev, job, &reset_context);
+		if (r)
+			dev_err(adev->dev, "GPU Recovery Failed: %d\n", r);
 	} else {
 		drm_sched_suspend_timeout(&ring->sched);
 		if (amdgpu_sriov_vf(adev))
@@ -71,74 +174,119 @@ static enum drm_gpu_sched_stat amdgpu_job_timedout(struct drm_sched_job *s_job)
 	}
 
 exit:
+	amdgpu_vm_put_task_info(ti);
 	drm_dev_exit(idx);
-	return DRM_GPU_SCHED_STAT_NOMINAL;
+	return DRM_GPU_SCHED_STAT_RESET;
 }
 
-int amdgpu_job_alloc(struct amdgpu_device *adev, unsigned num_ibs,
-		     struct amdgpu_job **job, struct amdgpu_vm *vm)
+int amdgpu_job_alloc(struct amdgpu_device *adev, struct amdgpu_vm *vm,
+		     struct drm_sched_entity *entity, void *owner,
+		     unsigned int num_ibs, struct amdgpu_job **job,
+		     u64 drm_client_id)
 {
-	size_t size = sizeof(struct amdgpu_job);
+	struct amdgpu_fence *af;
+	int r;
 
 	if (num_ibs == 0)
 		return -EINVAL;
 
-	size += sizeof(struct amdgpu_ib) * num_ibs;
-
-	*job = kzalloc(size, GFP_KERNEL);
+	*job = kzalloc(struct_size(*job, ibs, num_ibs), GFP_KERNEL);
 	if (!*job)
 		return -ENOMEM;
 
-	/*
-	 * Initialize the scheduler to at least some ring so that we always
-	 * have a pointer to adev.
-	 */
-	(*job)->base.sched = &adev->rings[0]->sched;
+	af = kzalloc(sizeof(struct amdgpu_fence), GFP_KERNEL);
+	if (!af) {
+		r = -ENOMEM;
+		goto err_job;
+	}
+	(*job)->hw_fence = af;
+
+	af = kzalloc(sizeof(struct amdgpu_fence), GFP_KERNEL);
+	if (!af) {
+		r = -ENOMEM;
+		goto err_fence;
+	}
+	(*job)->hw_vm_fence = af;
+
 	(*job)->vm = vm;
-	(*job)->ibs = (void *)&(*job)[1];
-	(*job)->num_ibs = num_ibs;
 
-	amdgpu_sync_create(&(*job)->sync);
-	amdgpu_sync_create(&(*job)->sched_sync);
-	(*job)->vram_lost_counter = atomic_read(&adev->vram_lost_counter);
+	amdgpu_sync_create(&(*job)->explicit_sync);
+	(*job)->generation = amdgpu_vm_generation(adev, vm);
 	(*job)->vm_pd_addr = AMDGPU_BO_INVALID_OFFSET;
 
-	return 0;
+	if (!entity)
+		return 0;
+
+	return drm_sched_job_init(&(*job)->base, entity, 1, owner,
+				  drm_client_id);
+
+err_fence:
+	kfree((*job)->hw_fence);
+err_job:
+	kfree(*job);
+	*job = NULL;
+
+	return r;
 }
 
-int amdgpu_job_alloc_with_ib(struct amdgpu_device *adev, unsigned size,
-		enum amdgpu_ib_pool_type pool_type,
-		struct amdgpu_job **job)
+int amdgpu_job_alloc_with_ib(struct amdgpu_device *adev,
+			     struct drm_sched_entity *entity, void *owner,
+			     size_t size, enum amdgpu_ib_pool_type pool_type,
+			     struct amdgpu_job **job, u64 k_job_id)
 {
 	int r;
 
-	r = amdgpu_job_alloc(adev, 1, job, NULL);
+	r = amdgpu_job_alloc(adev, NULL, entity, owner, 1, job,
+			     k_job_id);
 	if (r)
 		return r;
 
+	(*job)->num_ibs = 1;
 	r = amdgpu_ib_get(adev, NULL, size, pool_type, &(*job)->ibs[0]);
-	if (r)
+	if (r) {
+		if (entity)
+			drm_sched_job_cleanup(&(*job)->base);
+		kfree((*job)->hw_vm_fence);
+		kfree((*job)->hw_fence);
 		kfree(*job);
+		*job = NULL;
+	}
 
 	return r;
 }
 
+void amdgpu_job_set_resources(struct amdgpu_job *job, struct amdgpu_bo *gds,
+			      struct amdgpu_bo *gws, struct amdgpu_bo *oa)
+{
+	if (gds) {
+		job->gds_base = amdgpu_bo_gpu_offset(gds) >> PAGE_SHIFT;
+		job->gds_size = amdgpu_bo_size(gds) >> PAGE_SHIFT;
+	}
+	if (gws) {
+		job->gws_base = amdgpu_bo_gpu_offset(gws) >> PAGE_SHIFT;
+		job->gws_size = amdgpu_bo_size(gws) >> PAGE_SHIFT;
+	}
+	if (oa) {
+		job->oa_base = amdgpu_bo_gpu_offset(oa) >> PAGE_SHIFT;
+		job->oa_size = amdgpu_bo_size(oa) >> PAGE_SHIFT;
+	}
+}
+
 void amdgpu_job_free_resources(struct amdgpu_job *job)
 {
-	struct amdgpu_ring *ring = to_amdgpu_ring(job->base.sched);
 	struct dma_fence *f;
-	struct dma_fence *hw_fence;
 	unsigned i;
 
-	if (job->hw_fence.ops == NULL)
-		hw_fence = job->external_hw_fence;
+	/* Check if any fences were initialized */
+	if (job->base.s_fence && job->base.s_fence->finished.ops)
+		f = &job->base.s_fence->finished;
+	else if (job->hw_fence && job->hw_fence->base.ops)
+		f = &job->hw_fence->base;
 	else
-		hw_fence = &job->hw_fence;
+		f = NULL;
 
-	/* use sched fence if available */
-	f = job->base.s_fence ? &job->base.s_fence->finished : hw_fence;
 	for (i = 0; i < job->num_ibs; ++i)
-		amdgpu_ib_free(ring->adev, &job->ibs[i], f);
+		amdgpu_ib_free(&job->ibs[i], f);
 }
 
 static void amdgpu_job_free_cb(struct drm_sched_job *s_job)
@@ -147,48 +295,68 @@ static void amdgpu_job_free_cb(struct drm_sched_job *s_job)
 
 	drm_sched_job_cleanup(s_job);
 
-	amdgpu_sync_free(&job->sync);
-	amdgpu_sync_free(&job->sched_sync);
+	amdgpu_sync_free(&job->explicit_sync);
 
-    /* only put the hw fence if has embedded fence */
-	if (job->hw_fence.ops != NULL)
-		dma_fence_put(&job->hw_fence);
+	if (job->hw_fence->base.ops)
+		dma_fence_put(&job->hw_fence->base);
 	else
-		kfree(job);
+		kfree(job->hw_fence);
+	if (job->hw_vm_fence->base.ops)
+		dma_fence_put(&job->hw_vm_fence->base);
+	else
+		kfree(job->hw_vm_fence);
+
+	kfree(job);
+}
+
+void amdgpu_job_set_gang_leader(struct amdgpu_job *job,
+				struct amdgpu_job *leader)
+{
+	struct dma_fence *fence = &leader->base.s_fence->scheduled;
+
+	WARN_ON(job->gang_submit);
+
+	/*
+	 * Don't add a reference when we are the gang leader to avoid circle
+	 * dependency.
+	 */
+	if (job != leader)
+		dma_fence_get(fence);
+	job->gang_submit = fence;
 }
 
 void amdgpu_job_free(struct amdgpu_job *job)
 {
+	if (job->base.entity)
+		drm_sched_job_cleanup(&job->base);
+
 	amdgpu_job_free_resources(job);
-	amdgpu_sync_free(&job->sync);
-	amdgpu_sync_free(&job->sched_sync);
+	amdgpu_sync_free(&job->explicit_sync);
+	if (job->gang_submit != &job->base.s_fence->scheduled)
+		dma_fence_put(job->gang_submit);
 
-	/* only put the hw fence if has embedded fence */
-	if (job->hw_fence.ops != NULL)
-		dma_fence_put(&job->hw_fence);
+	if (job->hw_fence->base.ops)
+		dma_fence_put(&job->hw_fence->base);
+	else
+		kfree(job->hw_fence);
+	if (job->hw_vm_fence->base.ops)
+		dma_fence_put(&job->hw_vm_fence->base);
 	else
-		kfree(job);
+		kfree(job->hw_vm_fence);
+
+	kfree(job);
 }
 
-int amdgpu_job_submit(struct amdgpu_job *job, struct drm_sched_entity *entity,
-		      void *owner, struct dma_fence **f)
+struct dma_fence *amdgpu_job_submit(struct amdgpu_job *job)
 {
-	int r;
-
-	if (!f)
-		return -EINVAL;
-
-	r = drm_sched_job_init(&job->base, entity, owner);
-	if (r)
-		return r;
+	struct dma_fence *f;
 
 	drm_sched_job_arm(&job->base);
-
-	*f = dma_fence_get(&job->base.s_fence->finished);
+	f = dma_fence_get(&job->base.s_fence->finished);
 	amdgpu_job_free_resources(job);
 	drm_sched_entity_push_job(&job->base);
 
-	return 0;
+	return f;
 }
 
 int amdgpu_job_submit_direct(struct amdgpu_job *job, struct amdgpu_ring *ring,
@@ -197,50 +365,58 @@ int amdgpu_job_submit_direct(struct amdgpu_job *job, struct amdgpu_ring *ring,
 	int r;
 
 	job->base.sched = &ring->sched;
-	r = amdgpu_ib_schedule(ring, job->num_ibs, job->ibs, NULL, fence);
-	/* record external_hw_fence for direct submit */
-	job->external_hw_fence = dma_fence_get(*fence);
+	r = amdgpu_ib_schedule(ring, job->num_ibs, job->ibs, job, fence);
+
 	if (r)
 		return r;
 
 	amdgpu_job_free(job);
-	dma_fence_put(*fence);
-
 	return 0;
 }
 
-static struct dma_fence *amdgpu_job_dependency(struct drm_sched_job *sched_job,
-					       struct drm_sched_entity *s_entity)
+static struct dma_fence *
+amdgpu_job_prepare_job(struct drm_sched_job *sched_job,
+		      struct drm_sched_entity *s_entity)
 {
 	struct amdgpu_ring *ring = to_amdgpu_ring(s_entity->rq->sched);
 	struct amdgpu_job *job = to_amdgpu_job(sched_job);
-	struct amdgpu_vm *vm = job->vm;
 	struct dma_fence *fence;
 	int r;
 
-	fence = amdgpu_sync_get_fence(&job->sync);
-	if (fence && drm_sched_dependency_optimized(fence, s_entity)) {
-		r = amdgpu_sync_fence(&job->sched_sync, fence);
-		if (r)
-			DRM_ERROR("Error adding fence (%d)\n", r);
+	r = drm_sched_entity_error(s_entity);
+	if (r)
+		goto error;
+
+	if (job->gang_submit) {
+		fence = amdgpu_device_switch_gang(ring->adev, job->gang_submit);
+		if (fence)
+			return fence;
 	}
 
-	while (fence == NULL && vm && !job->vmid) {
-		r = amdgpu_vmid_grab(vm, ring, &job->sync,
-				     &job->base.s_fence->finished,
-				     job);
-		if (r)
-			DRM_ERROR("Error getting VM ID (%d)\n", r);
+	fence = amdgpu_device_enforce_isolation(ring->adev, ring, job);
+	if (fence)
+		return fence;
 
-		fence = amdgpu_sync_get_fence(&job->sync);
+	if (job->vm && !job->vmid) {
+		r = amdgpu_vmid_grab(job->vm, ring, job, &fence);
+		if (r) {
+			dev_err(ring->adev->dev, "Error getting VM ID (%d)\n", r);
+			goto error;
+		}
+		return fence;
 	}
 
-	return fence;
+	return NULL;
+
+error:
+	dma_fence_set_error(&job->base.s_fence->finished, r);
+	return NULL;
 }
 
 static struct dma_fence *amdgpu_job_run(struct drm_sched_job *sched_job)
 {
 	struct amdgpu_ring *ring = to_amdgpu_ring(sched_job->sched);
+	struct amdgpu_device *adev = ring->adev;
 	struct dma_fence *fence = NULL, *finished;
 	struct amdgpu_job *job;
 	int r = 0;
@@ -248,26 +424,25 @@ static struct dma_fence *amdgpu_job_run(struct drm_sched_job *sched_job)
 	job = to_amdgpu_job(sched_job);
 	finished = &job->base.s_fence->finished;
 
-	BUG_ON(amdgpu_sync_peek_fence(&job->sync, NULL));
-
 	trace_amdgpu_sched_run_job(job);
 
-	if (job->vram_lost_counter != atomic_read(&ring->adev->vram_lost_counter))
-		dma_fence_set_error(finished, -ECANCELED);/* skip IB as well if VRAM lost */
+	/* Skip job if VRAM is lost and never resubmit gangs */
+	if (job->generation != amdgpu_vm_generation(adev, job->vm) ||
+	    (job->job_run_counter && job->gang_submit))
+		dma_fence_set_error(finished, -ECANCELED);
 
 	if (finished->error < 0) {
-		DRM_INFO("Skip scheduling IBs!\n");
+		dev_dbg(adev->dev, "Skip scheduling IBs in ring(%s)",
+			ring->name);
 	} else {
 		r = amdgpu_ib_schedule(ring, job->num_ibs, job->ibs, job,
 				       &fence);
 		if (r)
-			DRM_ERROR("Error scheduling IBs (%d)\n", r);
+			dev_err(adev->dev,
+				"Error scheduling IBs (%d) in ring(%s)", r,
+				ring->name);
 	}
 
-	if (!job->job_run_counter)
-		dma_fence_get(fence);
-	else if (finished->error < 0)
-		dma_fence_put(&job->hw_fence);
 	job->job_run_counter++;
 	amdgpu_job_free_resources(job);
 
@@ -275,8 +450,24 @@ static struct dma_fence *amdgpu_job_run(struct drm_sched_job *sched_job)
 	return fence;
 }
 
-#define to_drm_sched_job(sched_job)		\
-		container_of((sched_job), struct drm_sched_job, queue_node)
+/*
+ * This is a duplicate function from DRM scheduler sched_internal.h.
+ * Plan is to remove it when amdgpu_job_stop_all_jobs_on_sched is removed, due
+ * latter being incorrect and racy.
+ *
+ * See https://lore.kernel.org/amd-gfx/44edde63-7181-44fb-a4f7-94e50514f539@amd.com/
+ */
+static struct drm_sched_job *
+drm_sched_entity_queue_pop(struct drm_sched_entity *entity)
+{
+	struct spsc_node *node;
+
+	node = spsc_queue_pop(&entity->job_queue);
+	if (!node)
+		return NULL;
+
+	return container_of(node, struct drm_sched_job, queue_node);
+}
 
 void amdgpu_job_stop_all_jobs_on_sched(struct drm_gpu_scheduler *sched)
 {
@@ -285,15 +476,11 @@ void amdgpu_job_stop_all_jobs_on_sched(struct drm_gpu_scheduler *sched)
 	int i;
 
 	/* Signal all jobs not yet scheduled */
-	for (i = DRM_SCHED_PRIORITY_COUNT - 1; i >= DRM_SCHED_PRIORITY_MIN; i--) {
-		struct drm_sched_rq *rq = &sched->sched_rq[i];
-
-		if (!rq)
-			continue;
-
+	for (i = DRM_SCHED_PRIORITY_KERNEL; i < sched->num_rqs; i++) {
+		struct drm_sched_rq *rq = sched->sched_rq[i];
 		spin_lock(&rq->lock);
 		list_for_each_entry(s_entity, &rq->entities, list) {
-			while ((s_job = to_drm_sched_job(spsc_queue_pop(&s_entity->job_queue)))) {
+			while ((s_job = drm_sched_entity_queue_pop(s_entity))) {
 				struct drm_sched_fence *s_fence = s_job->s_fence;
 
 				dma_fence_signal(&s_fence->scheduled);
@@ -314,7 +501,7 @@ void amdgpu_job_stop_all_jobs_on_sched(struct drm_gpu_scheduler *sched)
 }
 
 const struct drm_sched_backend_ops amdgpu_sched_ops = {
-	.dependency = amdgpu_job_dependency,
+	.prepare_job = amdgpu_job_prepare_job,
 	.run_job = amdgpu_job_run,
 	.timedout_job = amdgpu_job_timedout,
 	.free_job = amdgpu_job_free_cb
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.h
index 9e65730193b8..7abf069d17d4 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.h
@@ -23,6 +23,10 @@
 #ifndef __AMDGPU_JOB_H__
 #define __AMDGPU_JOB_H__
 
+#include <drm/gpu_scheduler.h>
+#include "amdgpu_sync.h"
+#include "amdgpu_ring.h"
+
 /* bit set means command submit involves a preamble IB */
 #define AMDGPU_PREAMBLE_IB_PRESENT          (1 << 0)
 /* bit set means preamble IB is first presented in belonging context */
@@ -40,42 +44,84 @@
 struct amdgpu_fence;
 enum amdgpu_ib_pool_type;
 
+/* Internal kernel job ids. (decreasing values, starting from U64_MAX). */
+#define AMDGPU_KERNEL_JOB_ID_VM_UPDATE              (18446744073709551615ULL)
+#define AMDGPU_KERNEL_JOB_ID_VM_UPDATE_PDES         (18446744073709551614ULL)
+#define AMDGPU_KERNEL_JOB_ID_VM_UPDATE_RANGE        (18446744073709551613ULL)
+#define AMDGPU_KERNEL_JOB_ID_VM_PT_CLEAR            (18446744073709551612ULL)
+#define AMDGPU_KERNEL_JOB_ID_TTM_MAP_BUFFER         (18446744073709551611ULL)
+#define AMDGPU_KERNEL_JOB_ID_TTM_ACCESS_MEMORY_SDMA (18446744073709551610ULL)
+#define AMDGPU_KERNEL_JOB_ID_TTM_COPY_BUFFER        (18446744073709551609ULL)
+#define AMDGPU_KERNEL_JOB_ID_CLEAR_ON_RELEASE       (18446744073709551608ULL)
+#define AMDGPU_KERNEL_JOB_ID_MOVE_BLIT              (18446744073709551607ULL)
+#define AMDGPU_KERNEL_JOB_ID_TTM_CLEAR_BUFFER       (18446744073709551606ULL)
+#define AMDGPU_KERNEL_JOB_ID_CLEANER_SHADER         (18446744073709551605ULL)
+#define AMDGPU_KERNEL_JOB_ID_FLUSH_GPU_TLB          (18446744073709551604ULL)
+#define AMDGPU_KERNEL_JOB_ID_KFD_GART_MAP           (18446744073709551603ULL)
+#define AMDGPU_KERNEL_JOB_ID_VCN_RING_TEST          (18446744073709551602ULL)
+
 struct amdgpu_job {
 	struct drm_sched_job    base;
 	struct amdgpu_vm	*vm;
-	struct amdgpu_sync	sync;
-	struct amdgpu_sync	sched_sync;
-	struct amdgpu_ib	*ibs;
-	struct dma_fence	hw_fence;
-	struct dma_fence	*external_hw_fence;
+	struct amdgpu_sync	explicit_sync;
+	struct amdgpu_fence	*hw_fence;
+	struct amdgpu_fence	*hw_vm_fence;
+	struct dma_fence	*gang_submit;
 	uint32_t		preamble_status;
 	uint32_t                preemption_status;
-	uint32_t		num_ibs;
 	bool                    vm_needs_flush;
+	bool			gds_switch_needed;
+	bool			spm_update_needed;
 	uint64_t		vm_pd_addr;
 	unsigned		vmid;
 	unsigned		pasid;
 	uint32_t		gds_base, gds_size;
 	uint32_t		gws_base, gws_size;
 	uint32_t		oa_base, oa_size;
-	uint32_t		vram_lost_counter;
+	uint64_t		generation;
 
 	/* user fence handling */
 	uint64_t		uf_addr;
 	uint64_t		uf_sequence;
 
+	/* virtual addresses for shadow/GDS/CSA */
+	uint64_t		shadow_va;
+	uint64_t		csa_va;
+	uint64_t		gds_va;
+	bool			init_shadow;
+
 	/* job_run_counter >= 1 means a resubmit job */
 	uint32_t		job_run_counter;
+
+	/* enforce isolation */
+	bool			enforce_isolation;
+	bool			run_cleaner_shader;
+
+	uint32_t		num_ibs;
+	struct amdgpu_ib	ibs[];
 };
 
-int amdgpu_job_alloc(struct amdgpu_device *adev, unsigned num_ibs,
-		     struct amdgpu_job **job, struct amdgpu_vm *vm);
-int amdgpu_job_alloc_with_ib(struct amdgpu_device *adev, unsigned size,
-		enum amdgpu_ib_pool_type pool, struct amdgpu_job **job);
+static inline struct amdgpu_ring *amdgpu_job_ring(struct amdgpu_job *job)
+{
+	return to_amdgpu_ring(job->base.entity->rq->sched);
+}
+
+int amdgpu_job_alloc(struct amdgpu_device *adev, struct amdgpu_vm *vm,
+		     struct drm_sched_entity *entity, void *owner,
+		     unsigned int num_ibs, struct amdgpu_job **job,
+		     u64 drm_client_id);
+int amdgpu_job_alloc_with_ib(struct amdgpu_device *adev,
+			     struct drm_sched_entity *entity, void *owner,
+			     size_t size, enum amdgpu_ib_pool_type pool_type,
+			     struct amdgpu_job **job,
+			     u64 k_job_id);
+void amdgpu_job_set_resources(struct amdgpu_job *job, struct amdgpu_bo *gds,
+			      struct amdgpu_bo *gws, struct amdgpu_bo *oa);
 void amdgpu_job_free_resources(struct amdgpu_job *job);
+void amdgpu_job_set_gang_leader(struct amdgpu_job *job,
+				struct amdgpu_job *leader);
 void amdgpu_job_free(struct amdgpu_job *job);
-int amdgpu_job_submit(struct amdgpu_job *job, struct drm_sched_entity *entity,
-		      void *owner, struct dma_fence **f);
+struct dma_fence *amdgpu_job_submit(struct amdgpu_job *job);
 int amdgpu_job_submit_direct(struct amdgpu_job *job, struct amdgpu_ring *ring,
 			     struct dma_fence **fence);
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c
index 9342aa23ebd2..63ee6ba6a931 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c
@@ -33,27 +33,62 @@
 #define JPEG_IDLE_TIMEOUT	msecs_to_jiffies(1000)
 
 static void amdgpu_jpeg_idle_work_handler(struct work_struct *work);
+static void amdgpu_jpeg_reg_dump_fini(struct amdgpu_device *adev);
 
 int amdgpu_jpeg_sw_init(struct amdgpu_device *adev)
 {
+	int i, r;
+
 	INIT_DELAYED_WORK(&adev->jpeg.idle_work, amdgpu_jpeg_idle_work_handler);
 	mutex_init(&adev->jpeg.jpeg_pg_lock);
 	atomic_set(&adev->jpeg.total_submission_cnt, 0);
 
+	if ((adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) &&
+	    (adev->pg_flags & AMD_PG_SUPPORT_JPEG_DPG))
+		adev->jpeg.indirect_sram = true;
+
+	for (i = 0; i < adev->jpeg.num_jpeg_inst; i++) {
+		if (adev->jpeg.harvest_config & (1U << i))
+			continue;
+
+		if (adev->jpeg.indirect_sram) {
+			r = amdgpu_bo_create_kernel(adev, 64 * 2 * 4, PAGE_SIZE,
+					AMDGPU_GEM_DOMAIN_VRAM |
+					AMDGPU_GEM_DOMAIN_GTT,
+					&adev->jpeg.inst[i].dpg_sram_bo,
+					&adev->jpeg.inst[i].dpg_sram_gpu_addr,
+					&adev->jpeg.inst[i].dpg_sram_cpu_addr);
+			if (r) {
+				dev_err(adev->dev,
+				"JPEG %d (%d) failed to allocate DPG bo\n", i, r);
+				return r;
+			}
+		}
+	}
+
 	return 0;
 }
 
 int amdgpu_jpeg_sw_fini(struct amdgpu_device *adev)
 {
-	int i;
+	int i, j;
 
 	for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
-		if (adev->jpeg.harvest_config & (1 << i))
+		if (adev->jpeg.harvest_config & (1U << i))
 			continue;
 
-		amdgpu_ring_fini(&adev->jpeg.inst[i].ring_dec);
+		amdgpu_bo_free_kernel(
+			&adev->jpeg.inst[i].dpg_sram_bo,
+			&adev->jpeg.inst[i].dpg_sram_gpu_addr,
+			(void **)&adev->jpeg.inst[i].dpg_sram_cpu_addr);
+
+		for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j)
+			amdgpu_ring_fini(&adev->jpeg.inst[i].ring_dec[j]);
 	}
 
+	if (adev->jpeg.reg_list)
+		amdgpu_jpeg_reg_dump_fini(adev);
+
 	mutex_destroy(&adev->jpeg.jpeg_pg_lock);
 
 	return 0;
@@ -76,19 +111,22 @@ static void amdgpu_jpeg_idle_work_handler(struct work_struct *work)
 	struct amdgpu_device *adev =
 		container_of(work, struct amdgpu_device, jpeg.idle_work.work);
 	unsigned int fences = 0;
-	unsigned int i;
+	unsigned int i, j;
 
 	for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
-		if (adev->jpeg.harvest_config & (1 << i))
+		if (adev->jpeg.harvest_config & (1U << i))
 			continue;
 
-		fences += amdgpu_fence_count_emitted(&adev->jpeg.inst[i].ring_dec);
+		for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j)
+			fences += amdgpu_fence_count_emitted(&adev->jpeg.inst[i].ring_dec[j]);
 	}
 
-	if (!fences && !atomic_read(&adev->jpeg.total_submission_cnt))
+	if (!fences && !atomic_read(&adev->jpeg.total_submission_cnt)) {
+		mutex_lock(&adev->jpeg.jpeg_pg_lock);
 		amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_JPEG,
 						       AMD_PG_STATE_GATE);
-	else
+		mutex_unlock(&adev->jpeg.jpeg_pg_lock);
+	} else
 		schedule_delayed_work(&adev->jpeg.idle_work, JPEG_IDLE_TIMEOUT);
 }
 
@@ -118,18 +156,25 @@ int amdgpu_jpeg_dec_ring_test_ring(struct amdgpu_ring *ring)
 	unsigned i;
 	int r;
 
-	WREG32(adev->jpeg.inst[ring->me].external.jpeg_pitch, 0xCAFEDEAD);
+	/* JPEG in SRIOV does not support direct register read/write */
+	if (amdgpu_sriov_vf(adev))
+		return 0;
+
 	r = amdgpu_ring_alloc(ring, 3);
 	if (r)
 		return r;
 
-	amdgpu_ring_write(ring, PACKET0(adev->jpeg.internal.jpeg_pitch, 0));
-	amdgpu_ring_write(ring, 0xDEADBEEF);
+	WREG32(adev->jpeg.inst[ring->me].external.jpeg_pitch[ring->pipe], 0xCAFEDEAD);
+	/* Add a read register to make sure the write register is executed. */
+	RREG32(adev->jpeg.inst[ring->me].external.jpeg_pitch[ring->pipe]);
+
+	amdgpu_ring_write(ring, PACKET0(adev->jpeg.internal.jpeg_pitch[ring->pipe], 0));
+	amdgpu_ring_write(ring, 0xABADCAFE);
 	amdgpu_ring_commit(ring);
 
 	for (i = 0; i < adev->usec_timeout; i++) {
-		tmp = RREG32(adev->jpeg.inst[ring->me].external.jpeg_pitch);
-		if (tmp == 0xDEADBEEF)
+		tmp = RREG32(adev->jpeg.inst[ring->me].external.jpeg_pitch[ring->pipe]);
+		if (tmp == 0xABADCAFE)
 			break;
 		udelay(1);
 	}
@@ -150,14 +195,15 @@ static int amdgpu_jpeg_dec_set_reg(struct amdgpu_ring *ring, uint32_t handle,
 	const unsigned ib_size_dw = 16;
 	int i, r;
 
-	r = amdgpu_job_alloc_with_ib(ring->adev, ib_size_dw * 4,
-					AMDGPU_IB_POOL_DIRECT, &job);
+	r = amdgpu_job_alloc_with_ib(ring->adev, NULL, NULL, ib_size_dw * 4,
+				     AMDGPU_IB_POOL_DIRECT, &job,
+				     AMDGPU_KERNEL_JOB_ID_VCN_RING_TEST);
 	if (r)
 		return r;
 
 	ib = &job->ibs[0];
 
-	ib->ptr[0] = PACKETJ(adev->jpeg.internal.jpeg_pitch, 0, 0, PACKETJ_TYPE0);
+	ib->ptr[0] = PACKETJ(adev->jpeg.internal.jpeg_pitch[ring->pipe], 0, 0, PACKETJ_TYPE0);
 	ib->ptr[1] = 0xDEADBEEF;
 	for (i = 2; i < 16; i += 2) {
 		ib->ptr[i] = PACKETJ(0, 0, 0, PACKETJ_TYPE6);
@@ -202,17 +248,360 @@ int amdgpu_jpeg_dec_ring_test_ib(struct amdgpu_ring *ring, long timeout)
 		r = 0;
 	}
 
-	for (i = 0; i < adev->usec_timeout; i++) {
-		tmp = RREG32(adev->jpeg.inst[ring->me].external.jpeg_pitch);
-		if (tmp == 0xDEADBEEF)
-			break;
-		udelay(1);
+	if (!amdgpu_sriov_vf(adev)) {
+		for (i = 0; i < adev->usec_timeout; i++) {
+			tmp = RREG32(adev->jpeg.inst[ring->me].external.jpeg_pitch[ring->pipe]);
+			if (tmp == 0xDEADBEEF)
+				break;
+			udelay(1);
+			if (amdgpu_emu_mode == 1)
+				udelay(10);
+		}
+
+		if (i >= adev->usec_timeout)
+			r = -ETIMEDOUT;
 	}
 
-	if (i >= adev->usec_timeout)
-		r = -ETIMEDOUT;
-
 	dma_fence_put(fence);
 error:
 	return r;
 }
+
+int amdgpu_jpeg_process_poison_irq(struct amdgpu_device *adev,
+				struct amdgpu_irq_src *source,
+				struct amdgpu_iv_entry *entry)
+{
+	struct ras_common_if *ras_if = adev->jpeg.ras_if;
+	struct ras_dispatch_if ih_data = {
+		.entry = entry,
+	};
+
+	if (!ras_if)
+		return 0;
+
+	ih_data.head = *ras_if;
+	amdgpu_ras_interrupt_dispatch(adev, &ih_data);
+
+	return 0;
+}
+
+int amdgpu_jpeg_ras_late_init(struct amdgpu_device *adev, struct ras_common_if *ras_block)
+{
+	int r, i;
+
+	r = amdgpu_ras_block_late_init(adev, ras_block);
+	if (r)
+		return r;
+
+	if (amdgpu_ras_is_supported(adev, ras_block->block)) {
+		for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
+			if (adev->jpeg.harvest_config & (1 << i) ||
+			    !adev->jpeg.inst[i].ras_poison_irq.funcs)
+				continue;
+
+			r = amdgpu_irq_get(adev, &adev->jpeg.inst[i].ras_poison_irq, 0);
+			if (r)
+				goto late_fini;
+		}
+	}
+	return 0;
+
+late_fini:
+	amdgpu_ras_block_late_fini(adev, ras_block);
+	return r;
+}
+
+int amdgpu_jpeg_ras_sw_init(struct amdgpu_device *adev)
+{
+	int err;
+	struct amdgpu_jpeg_ras *ras;
+
+	if (!adev->jpeg.ras)
+		return 0;
+
+	ras = adev->jpeg.ras;
+	err = amdgpu_ras_register_ras_block(adev, &ras->ras_block);
+	if (err) {
+		dev_err(adev->dev, "Failed to register jpeg ras block!\n");
+		return err;
+	}
+
+	strcpy(ras->ras_block.ras_comm.name, "jpeg");
+	ras->ras_block.ras_comm.block = AMDGPU_RAS_BLOCK__JPEG;
+	ras->ras_block.ras_comm.type = AMDGPU_RAS_ERROR__POISON;
+	adev->jpeg.ras_if = &ras->ras_block.ras_comm;
+
+	if (!ras->ras_block.ras_late_init)
+		ras->ras_block.ras_late_init = amdgpu_jpeg_ras_late_init;
+
+	return 0;
+}
+
+int amdgpu_jpeg_psp_update_sram(struct amdgpu_device *adev, int inst_idx,
+			       enum AMDGPU_UCODE_ID ucode_id)
+{
+	struct amdgpu_firmware_info ucode = {
+		.ucode_id = AMDGPU_UCODE_ID_JPEG_RAM,
+		.mc_addr = adev->jpeg.inst[inst_idx].dpg_sram_gpu_addr,
+		.ucode_size = ((uintptr_t)adev->jpeg.inst[inst_idx].dpg_sram_curr_addr -
+			      (uintptr_t)adev->jpeg.inst[inst_idx].dpg_sram_cpu_addr),
+	};
+
+	return psp_execute_ip_fw_load(&adev->psp, &ucode);
+}
+
+/*
+ * debugfs for to enable/disable jpeg job submission to specific core.
+ */
+#if defined(CONFIG_DEBUG_FS)
+static int amdgpu_debugfs_jpeg_sched_mask_set(void *data, u64 val)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)data;
+	u32 i, j;
+	u64 mask = 0;
+	struct amdgpu_ring *ring;
+
+	if (!adev)
+		return -ENODEV;
+
+	mask = (1ULL << (adev->jpeg.num_jpeg_inst * adev->jpeg.num_jpeg_rings)) - 1;
+	if ((val & mask) == 0)
+		return -EINVAL;
+
+	for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
+		for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j) {
+			ring = &adev->jpeg.inst[i].ring_dec[j];
+			if (val & (BIT_ULL((i * adev->jpeg.num_jpeg_rings) + j)))
+				ring->sched.ready = true;
+			else
+				ring->sched.ready = false;
+		}
+	}
+	/* publish sched.ready flag update effective immediately across smp */
+	smp_rmb();
+	return 0;
+}
+
+static int amdgpu_debugfs_jpeg_sched_mask_get(void *data, u64 *val)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)data;
+	u32 i, j;
+	u64 mask = 0;
+	struct amdgpu_ring *ring;
+
+	if (!adev)
+		return -ENODEV;
+	for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
+		for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j) {
+			ring = &adev->jpeg.inst[i].ring_dec[j];
+			if (ring->sched.ready)
+				mask |= 1ULL << ((i * adev->jpeg.num_jpeg_rings) + j);
+		}
+	}
+	*val = mask;
+	return 0;
+}
+
+DEFINE_DEBUGFS_ATTRIBUTE(amdgpu_debugfs_jpeg_sched_mask_fops,
+			 amdgpu_debugfs_jpeg_sched_mask_get,
+			 amdgpu_debugfs_jpeg_sched_mask_set, "%llx\n");
+
+#endif
+
+void amdgpu_debugfs_jpeg_sched_mask_init(struct amdgpu_device *adev)
+{
+#if defined(CONFIG_DEBUG_FS)
+	struct drm_minor *minor = adev_to_drm(adev)->primary;
+	struct dentry *root = minor->debugfs_root;
+	char name[32];
+
+	if (!(adev->jpeg.num_jpeg_inst > 1) && !(adev->jpeg.num_jpeg_rings > 1))
+		return;
+	sprintf(name, "amdgpu_jpeg_sched_mask");
+	debugfs_create_file(name, 0600, root, adev,
+			    &amdgpu_debugfs_jpeg_sched_mask_fops);
+#endif
+}
+
+static ssize_t amdgpu_get_jpeg_reset_mask(struct device *dev,
+						struct device_attribute *attr,
+						char *buf)
+{
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = drm_to_adev(ddev);
+
+	if (!adev)
+		return -ENODEV;
+
+	return amdgpu_show_reset_mask(buf, adev->jpeg.supported_reset);
+}
+
+static DEVICE_ATTR(jpeg_reset_mask, 0444,
+		   amdgpu_get_jpeg_reset_mask, NULL);
+
+int amdgpu_jpeg_sysfs_reset_mask_init(struct amdgpu_device *adev)
+{
+	int r = 0;
+
+	if (adev->jpeg.num_jpeg_inst) {
+		r = device_create_file(adev->dev, &dev_attr_jpeg_reset_mask);
+		if (r)
+			return r;
+	}
+
+	return r;
+}
+
+void amdgpu_jpeg_sysfs_reset_mask_fini(struct amdgpu_device *adev)
+{
+	if (adev->dev->kobj.sd) {
+		if (adev->jpeg.num_jpeg_inst)
+			device_remove_file(adev->dev, &dev_attr_jpeg_reset_mask);
+	}
+}
+
+int amdgpu_jpeg_reg_dump_init(struct amdgpu_device *adev,
+			       const struct amdgpu_hwip_reg_entry *reg, u32 count)
+{
+	adev->jpeg.ip_dump = kcalloc(adev->jpeg.num_jpeg_inst * count,
+				     sizeof(uint32_t), GFP_KERNEL);
+	if (!adev->jpeg.ip_dump) {
+		dev_err(adev->dev,
+			"Failed to allocate memory for JPEG IP Dump\n");
+		return -ENOMEM;
+	}
+	adev->jpeg.reg_list = reg;
+	adev->jpeg.reg_count = count;
+
+	return 0;
+}
+
+static void amdgpu_jpeg_reg_dump_fini(struct amdgpu_device *adev)
+{
+	kfree(adev->jpeg.ip_dump);
+	adev->jpeg.reg_list = NULL;
+	adev->jpeg.reg_count = 0;
+}
+
+void amdgpu_jpeg_dump_ip_state(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	u32 inst_off, inst_id, is_powered;
+	int i, j;
+
+	if (!adev->jpeg.ip_dump)
+		return;
+
+	for (i = 0; i < adev->jpeg.num_jpeg_inst; i++) {
+		if (adev->jpeg.harvest_config & (1 << i))
+			continue;
+
+		inst_id = GET_INST(JPEG, i);
+		inst_off = i * adev->jpeg.reg_count;
+		/* check power status from UVD_JPEG_POWER_STATUS */
+		adev->jpeg.ip_dump[inst_off] =
+			RREG32(SOC15_REG_ENTRY_OFFSET_INST(adev->jpeg.reg_list[0],
+							   inst_id));
+		is_powered = ((adev->jpeg.ip_dump[inst_off] & 0x1) != 1);
+
+		if (is_powered)
+			for (j = 1; j < adev->jpeg.reg_count; j++)
+				adev->jpeg.ip_dump[inst_off + j] =
+					RREG32(SOC15_REG_ENTRY_OFFSET_INST(adev->jpeg.reg_list[j],
+									   inst_id));
+	}
+}
+
+void amdgpu_jpeg_print_ip_state(struct amdgpu_ip_block *ip_block, struct drm_printer *p)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	u32 inst_off, is_powered;
+	int i, j;
+
+	if (!adev->jpeg.ip_dump)
+		return;
+
+	drm_printf(p, "num_instances:%d\n", adev->jpeg.num_jpeg_inst);
+	for (i = 0; i < adev->jpeg.num_jpeg_inst; i++) {
+		if (adev->jpeg.harvest_config & (1 << i)) {
+			drm_printf(p, "\nHarvested Instance:JPEG%d Skipping dump\n", i);
+			continue;
+		}
+
+		inst_off = i * adev->jpeg.reg_count;
+		is_powered = ((adev->jpeg.ip_dump[inst_off] & 0x1) != 1);
+
+		if (is_powered) {
+			drm_printf(p, "Active Instance:JPEG%d\n", i);
+			for (j = 0; j < adev->jpeg.reg_count; j++)
+				drm_printf(p, "%-50s \t 0x%08x\n", adev->jpeg.reg_list[j].reg_name,
+					   adev->jpeg.ip_dump[inst_off + j]);
+		} else
+			drm_printf(p, "\nInactive Instance:JPEG%d\n", i);
+	}
+}
+
+static inline bool amdgpu_jpeg_reg_valid(u32 reg)
+{
+	if (reg < JPEG_REG_RANGE_START || reg > JPEG_REG_RANGE_END ||
+	    (reg >= JPEG_ATOMIC_RANGE_START && reg <= JPEG_ATOMIC_RANGE_END))
+		return false;
+	else
+		return true;
+}
+
+/**
+ * amdgpu_jpeg_dec_parse_cs - command submission parser
+ *
+ * @parser: Command submission parser context
+ * @job: the job to parse
+ * @ib: the IB to parse
+ *
+ * Parse the command stream, return -EINVAL for invalid packet,
+ * 0 otherwise
+ */
+
+int amdgpu_jpeg_dec_parse_cs(struct amdgpu_cs_parser *parser,
+			      struct amdgpu_job *job,
+			      struct amdgpu_ib *ib)
+{
+	u32 i, reg, res, cond, type;
+	struct amdgpu_device *adev = parser->adev;
+
+	for (i = 0; i < ib->length_dw ; i += 2) {
+		reg  = CP_PACKETJ_GET_REG(ib->ptr[i]);
+		res  = CP_PACKETJ_GET_RES(ib->ptr[i]);
+		cond = CP_PACKETJ_GET_COND(ib->ptr[i]);
+		type = CP_PACKETJ_GET_TYPE(ib->ptr[i]);
+
+		if (res) /* only support 0 at the moment */
+			return -EINVAL;
+
+		switch (type) {
+		case PACKETJ_TYPE0:
+			if (cond != PACKETJ_CONDITION_CHECK0 ||
+			    !amdgpu_jpeg_reg_valid(reg)) {
+				dev_err(adev->dev, "Invalid packet [0x%08x]!\n", ib->ptr[i]);
+				return -EINVAL;
+			}
+			break;
+		case PACKETJ_TYPE3:
+			if (cond != PACKETJ_CONDITION_CHECK3 ||
+			    !amdgpu_jpeg_reg_valid(reg)) {
+				dev_err(adev->dev, "Invalid packet [0x%08x]!\n", ib->ptr[i]);
+				return -EINVAL;
+			}
+			break;
+		case PACKETJ_TYPE6:
+			if (ib->ptr[i] == CP_PACKETJ_NOP)
+				continue;
+			dev_err(adev->dev, "Invalid packet [0x%08x]!\n", ib->ptr[i]);
+			return -EINVAL;
+		default:
+			dev_err(adev->dev, "Unknown packet type %d !\n", type);
+			return -EINVAL;
+		}
+	}
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h
index 55fbff2be761..346ae0ab09d3 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h
@@ -24,30 +24,131 @@
 #ifndef __AMDGPU_JPEG_H__
 #define __AMDGPU_JPEG_H__
 
-#define AMDGPU_MAX_JPEG_INSTANCES	2
+#include "amdgpu_ras.h"
+#include "amdgpu_cs.h"
+
+#define AMDGPU_MAX_JPEG_INSTANCES	4
+#define AMDGPU_MAX_JPEG_RINGS           10
+#define AMDGPU_MAX_JPEG_RINGS_4_0_3     8
+
+#define JPEG_REG_RANGE_START            0x4000
+#define JPEG_REG_RANGE_END              0x41c2
+#define JPEG_ATOMIC_RANGE_START         0x4120
+#define JPEG_ATOMIC_RANGE_END           0x412A
+
 
 #define AMDGPU_JPEG_HARVEST_JPEG0 (1 << 0)
 #define AMDGPU_JPEG_HARVEST_JPEG1 (1 << 1)
 
+#define WREG32_SOC15_JPEG_DPG_MODE(inst_idx, offset, value, indirect)			\
+	do {										\
+		if (!indirect) {							\
+			WREG32_SOC15(JPEG, GET_INST(JPEG, inst_idx),			\
+				     mmUVD_DPG_LMA_DATA, value);			\
+			WREG32_SOC15(							\
+				JPEG, GET_INST(JPEG, inst_idx),				\
+				mmUVD_DPG_LMA_CTL,					\
+				(UVD_DPG_LMA_CTL__READ_WRITE_MASK |			\
+				 offset << UVD_DPG_LMA_CTL__READ_WRITE_ADDR__SHIFT |	\
+				 indirect << UVD_DPG_LMA_CTL__SRAM_SEL__SHIFT));	\
+		} else {								\
+			*adev->jpeg.inst[inst_idx].dpg_sram_curr_addr++ =		\
+				offset;							\
+			*adev->jpeg.inst[inst_idx].dpg_sram_curr_addr++ =		\
+				value;							\
+		}									\
+	} while (0)
+
+#define RREG32_SOC15_JPEG_DPG_MODE(inst_idx, offset, mask_en)					\
+	({											\
+		WREG32_SOC15(JPEG, inst_idx, mmUVD_DPG_LMA_CTL,					\
+			(0x0 << UVD_DPG_LMA_CTL__READ_WRITE__SHIFT |				\
+			mask_en << UVD_DPG_LMA_CTL__MASK_EN__SHIFT |				\
+			offset << UVD_DPG_LMA_CTL__READ_WRITE_ADDR__SHIFT));			\
+		RREG32_SOC15(JPEG, inst_idx, mmUVD_DPG_LMA_DATA);				\
+	})
+
+#define WREG32_SOC24_JPEG_DPG_MODE(inst_idx, offset, value, indirect)		\
+	do {									\
+		WREG32_SOC15(JPEG, GET_INST(JPEG, inst_idx),			\
+			     regUVD_DPG_LMA_DATA, value);			\
+		WREG32_SOC15(JPEG, GET_INST(JPEG, inst_idx),			\
+			     regUVD_DPG_LMA_MASK, 0xFFFFFFFF);			\
+		WREG32_SOC15(							\
+			JPEG, GET_INST(JPEG, inst_idx),				\
+			regUVD_DPG_LMA_CTL,					\
+			(UVD_DPG_LMA_CTL__READ_WRITE_MASK |			\
+			 offset << UVD_DPG_LMA_CTL__READ_WRITE_ADDR__SHIFT |	\
+			 indirect << UVD_DPG_LMA_CTL__SRAM_SEL__SHIFT));	\
+	} while (0)
+
+#define RREG32_SOC24_JPEG_DPG_MODE(inst_idx, offset, mask_en)			\
+	do {									\
+		WREG32_SOC15(JPEG, GET_INST(JPEG, inst_idx),			\
+			regUVD_DPG_LMA_MASK, 0xFFFFFFFF);			\
+		WREG32_SOC15(JPEG, GET_INST(JPEG, inst_idx),			\
+			regUVD_DPG_LMA_CTL,					\
+			(UVD_DPG_LMA_CTL__MASK_EN_MASK |			\
+			offset << UVD_DPG_LMA_CTL__READ_WRITE_ADDR__SHIFT));	\
+		RREG32_SOC15(JPEG, inst_idx, regUVD_DPG_LMA_DATA);		\
+	} while (0)
+
+#define ADD_SOC24_JPEG_TO_DPG_SRAM(inst_idx, offset, value, indirect)		\
+	do {									\
+		*adev->jpeg.inst[inst_idx].dpg_sram_curr_addr++ = offset;	\
+		*adev->jpeg.inst[inst_idx].dpg_sram_curr_addr++ = value;	\
+	} while (0)
+
+struct amdgpu_hwip_reg_entry;
+
+enum amdgpu_jpeg_caps {
+	AMDGPU_JPEG_RRMT_ENABLED,
+};
+
+#define AMDGPU_JPEG_CAPS(caps) BIT(AMDGPU_JPEG_##caps)
+
 struct amdgpu_jpeg_reg{
-	unsigned jpeg_pitch;
+	unsigned jpeg_pitch[AMDGPU_MAX_JPEG_RINGS];
 };
 
 struct amdgpu_jpeg_inst {
-	struct amdgpu_ring ring_dec;
+	struct amdgpu_ring ring_dec[AMDGPU_MAX_JPEG_RINGS];
 	struct amdgpu_irq_src irq;
+	struct amdgpu_irq_src ras_poison_irq;
 	struct amdgpu_jpeg_reg external;
+	struct amdgpu_bo	*dpg_sram_bo;
+	struct dpg_pause_state	pause_state;
+	void			*dpg_sram_cpu_addr;
+	uint64_t		dpg_sram_gpu_addr;
+	uint32_t		*dpg_sram_curr_addr;
+	uint8_t aid_id;
+};
+
+struct amdgpu_jpeg_ras {
+	struct amdgpu_ras_block_object ras_block;
 };
 
 struct amdgpu_jpeg {
 	uint8_t	num_jpeg_inst;
 	struct amdgpu_jpeg_inst inst[AMDGPU_MAX_JPEG_INSTANCES];
+	unsigned num_jpeg_rings;
 	struct amdgpu_jpeg_reg internal;
 	unsigned harvest_config;
 	struct delayed_work idle_work;
 	enum amd_powergating_state cur_state;
 	struct mutex jpeg_pg_lock;
 	atomic_t total_submission_cnt;
+	struct ras_common_if	*ras_if;
+	struct amdgpu_jpeg_ras	*ras;
+
+	uint16_t inst_mask;
+	uint8_t num_inst_per_aid;
+	bool	indirect_sram;
+	uint32_t supported_reset;
+	uint32_t caps;
+	u32 *ip_dump;
+	u32 reg_count;
+	const struct amdgpu_hwip_reg_entry *reg_list;
 };
 
 int amdgpu_jpeg_sw_init(struct amdgpu_device *adev);
@@ -61,4 +162,23 @@ void amdgpu_jpeg_ring_end_use(struct amdgpu_ring *ring);
 int amdgpu_jpeg_dec_ring_test_ring(struct amdgpu_ring *ring);
 int amdgpu_jpeg_dec_ring_test_ib(struct amdgpu_ring *ring, long timeout);
 
+int amdgpu_jpeg_process_poison_irq(struct amdgpu_device *adev,
+				struct amdgpu_irq_src *source,
+				struct amdgpu_iv_entry *entry);
+int amdgpu_jpeg_ras_late_init(struct amdgpu_device *adev,
+				struct ras_common_if *ras_block);
+int amdgpu_jpeg_ras_sw_init(struct amdgpu_device *adev);
+int amdgpu_jpeg_psp_update_sram(struct amdgpu_device *adev, int inst_idx,
+			       enum AMDGPU_UCODE_ID ucode_id);
+void amdgpu_debugfs_jpeg_sched_mask_init(struct amdgpu_device *adev);
+int amdgpu_jpeg_sysfs_reset_mask_init(struct amdgpu_device *adev);
+void amdgpu_jpeg_sysfs_reset_mask_fini(struct amdgpu_device *adev);
+int amdgpu_jpeg_reg_dump_init(struct amdgpu_device *adev,
+			       const struct amdgpu_hwip_reg_entry *reg, u32 count);
+void amdgpu_jpeg_dump_ip_state(struct amdgpu_ip_block *ip_block);
+void amdgpu_jpeg_print_ip_state(struct amdgpu_ip_block *ip_block, struct drm_printer *p);
+int amdgpu_jpeg_dec_parse_cs(struct amdgpu_cs_parser *parser,
+			     struct amdgpu_job *job,
+			     struct amdgpu_ib *ib);
+
 #endif /*__AMDGPU_JPEG_H__*/
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
index 651c7abfde03..6ee77f431d56 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
@@ -29,6 +29,7 @@
 #include "amdgpu.h"
 #include <drm/amdgpu_drm.h>
 #include <drm/drm_drv.h>
+#include <drm/drm_fb_helper.h>
 #include "amdgpu_uvd.h"
 #include "amdgpu_vce.h"
 #include "atom.h"
@@ -42,6 +43,9 @@
 #include "amdgpu_gem.h"
 #include "amdgpu_display.h"
 #include "amdgpu_ras.h"
+#include "amdgpu_reset.h"
+#include "amd_pcie.h"
+#include "amdgpu_userq.h"
 
 void amdgpu_unregister_gpu_instance(struct amdgpu_device *adev)
 {
@@ -87,12 +91,7 @@ void amdgpu_driver_unload_kms(struct drm_device *dev)
 	if (adev->rmmio == NULL)
 		return;
 
-	if (adev->runpm) {
-		pm_runtime_get_sync(dev->dev);
-		pm_runtime_forbid(dev->dev);
-	}
-
-	if (amdgpu_acpi_smart_shift_update(dev, AMDGPU_SS_DRV_UNLOAD))
+	if (amdgpu_acpi_smart_shift_update(adev, AMDGPU_SS_DRV_UNLOAD))
 		DRM_WARN("smart shift update failed\n");
 
 	amdgpu_acpi_fini(adev);
@@ -124,22 +123,6 @@ void amdgpu_register_gpu_instance(struct amdgpu_device *adev)
 	mutex_unlock(&mgpu_info.mutex);
 }
 
-static void amdgpu_get_audio_func(struct amdgpu_device *adev)
-{
-	struct pci_dev *p = NULL;
-
-	p = pci_get_domain_bus_and_slot(pci_domain_nr(adev->pdev->bus),
-			adev->pdev->bus->number, 1);
-	if (p) {
-		pm_runtime_get_sync(&p->dev);
-
-		pm_runtime_mark_last_busy(&p->dev);
-		pm_runtime_put_autosuspend(&p->dev);
-
-		pci_dev_put(p);
-	}
-}
-
 /**
  * amdgpu_driver_load_kms - Main load function for KMS.
  *
@@ -152,21 +135,10 @@ static void amdgpu_get_audio_func(struct amdgpu_device *adev)
 int amdgpu_driver_load_kms(struct amdgpu_device *adev, unsigned long flags)
 {
 	struct drm_device *dev;
-	struct pci_dev *parent;
 	int r, acpi_status;
 
 	dev = adev_to_drm(adev);
 
-	if (amdgpu_has_atpx() &&
-	    (amdgpu_is_atpx_hybrid() ||
-	     amdgpu_has_atpx_dgpu_power_cntl()) &&
-	    ((flags & AMD_IS_APU) == 0) &&
-	    !pci_is_thunderbolt_attached(to_pci_dev(dev->dev)))
-		flags |= AMD_IS_PX;
-
-	parent = pci_upstream_bridge(adev->pdev);
-	adev->has_pr3 = parent ? pci_pr3_present(parent) : false;
-
 	/* amdgpu_device_init should report only fatal error
 	 * like memory allocation failure or iomapping failure,
 	 * or memory manager initialization failure, it must
@@ -179,36 +151,7 @@ int amdgpu_driver_load_kms(struct amdgpu_device *adev, unsigned long flags)
 		goto out;
 	}
 
-	if (amdgpu_device_supports_px(dev) &&
-	    (amdgpu_runtime_pm != 0)) { /* enable runpm by default for atpx */
-		adev->runpm = true;
-		dev_info(adev->dev, "Using ATPX for runtime pm\n");
-	} else if (amdgpu_device_supports_boco(dev) &&
-		   (amdgpu_runtime_pm != 0)) { /* enable runpm by default for boco */
-		adev->runpm = true;
-		dev_info(adev->dev, "Using BOCO for runtime pm\n");
-	} else if (amdgpu_device_supports_baco(dev) &&
-		   (amdgpu_runtime_pm != 0)) {
-		switch (adev->asic_type) {
-		case CHIP_VEGA20:
-		case CHIP_ARCTURUS:
-			/* enable runpm if runpm=1 */
-			if (amdgpu_runtime_pm > 0)
-				adev->runpm = true;
-			break;
-		case CHIP_VEGA10:
-			/* turn runpm on if noretry=0 */
-			if (!adev->gmc.noretry)
-				adev->runpm = true;
-			break;
-		default:
-			/* enable runpm on CI+ */
-			adev->runpm = true;
-			break;
-		}
-		if (adev->runpm)
-			dev_info(adev->dev, "Using BACO for runtime pm\n");
-	}
+	amdgpu_device_detect_runtime_pm_mode(adev);
 
 	/* Call ACPI methods: require modeset init
 	 * but failure is not fatal
@@ -218,62 +161,54 @@ int amdgpu_driver_load_kms(struct amdgpu_device *adev, unsigned long flags)
 	if (acpi_status)
 		dev_dbg(dev->dev, "Error during ACPI methods call\n");
 
-	if (adev->runpm) {
-		/* only need to skip on ATPX */
-		if (amdgpu_device_supports_px(dev))
-			dev_pm_set_driver_flags(dev->dev, DPM_FLAG_NO_DIRECT_COMPLETE);
-		/* we want direct complete for BOCO */
-		if (amdgpu_device_supports_boco(dev))
-			dev_pm_set_driver_flags(dev->dev, DPM_FLAG_SMART_PREPARE |
-						DPM_FLAG_SMART_SUSPEND |
-						DPM_FLAG_MAY_SKIP_RESUME);
-		pm_runtime_use_autosuspend(dev->dev);
-		pm_runtime_set_autosuspend_delay(dev->dev, 5000);
-
-		pm_runtime_allow(dev->dev);
-
-		pm_runtime_mark_last_busy(dev->dev);
-		pm_runtime_put_autosuspend(dev->dev);
-
-		/*
-		 * For runpm implemented via BACO, PMFW will handle the
-		 * timing for BACO in and out:
-		 *   - put ASIC into BACO state only when both video and
-		 *     audio functions are in D3 state.
-		 *   - pull ASIC out of BACO state when either video or
-		 *     audio function is in D0 state.
-		 * Also, at startup, PMFW assumes both functions are in
-		 * D0 state.
-		 *
-		 * So if snd driver was loaded prior to amdgpu driver
-		 * and audio function was put into D3 state, there will
-		 * be no PMFW-aware D-state transition(D0->D3) on runpm
-		 * suspend. Thus the BACO will be not correctly kicked in.
-		 *
-		 * Via amdgpu_get_audio_func(), the audio dev is put
-		 * into D0 state. Then there will be a PMFW-aware D-state
-		 * transition(D0->D3) on runpm suspend.
-		 */
-		if (amdgpu_device_supports_baco(dev) &&
-		    !(adev->flags & AMD_IS_APU) &&
-		    (adev->asic_type >= CHIP_NAVI10))
-			amdgpu_get_audio_func(adev);
-	}
-
-	if (amdgpu_acpi_smart_shift_update(dev, AMDGPU_SS_DRV_LOAD))
+	if (amdgpu_acpi_smart_shift_update(adev, AMDGPU_SS_DRV_LOAD))
 		DRM_WARN("smart shift update failed\n");
 
 out:
-	if (r) {
-		/* balance pm_runtime_get_sync in amdgpu_driver_unload_kms */
-		if (adev->rmmio && adev->runpm)
-			pm_runtime_put_noidle(dev->dev);
+	if (r)
 		amdgpu_driver_unload_kms(dev);
-	}
 
 	return r;
 }
 
+static enum amd_ip_block_type
+	amdgpu_ip_get_block_type(struct amdgpu_device *adev, uint32_t ip)
+{
+	enum amd_ip_block_type type;
+
+	switch (ip) {
+	case AMDGPU_HW_IP_GFX:
+		type = AMD_IP_BLOCK_TYPE_GFX;
+		break;
+	case AMDGPU_HW_IP_COMPUTE:
+		type = AMD_IP_BLOCK_TYPE_GFX;
+		break;
+	case AMDGPU_HW_IP_DMA:
+		type = AMD_IP_BLOCK_TYPE_SDMA;
+		break;
+	case AMDGPU_HW_IP_UVD:
+	case AMDGPU_HW_IP_UVD_ENC:
+		type = AMD_IP_BLOCK_TYPE_UVD;
+		break;
+	case AMDGPU_HW_IP_VCE:
+		type = AMD_IP_BLOCK_TYPE_VCE;
+		break;
+	case AMDGPU_HW_IP_VCN_DEC:
+	case AMDGPU_HW_IP_VCN_ENC:
+		type = AMD_IP_BLOCK_TYPE_VCN;
+		break;
+	case AMDGPU_HW_IP_VCN_JPEG:
+		type = (amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_JPEG)) ?
+				   AMD_IP_BLOCK_TYPE_JPEG : AMD_IP_BLOCK_TYPE_VCN;
+		break;
+	default:
+		type = AMD_IP_BLOCK_TYPE_NUM;
+		break;
+	}
+
+	return type;
+}
+
 static int amdgpu_firmware_info(struct drm_amdgpu_info_firmware *fw_info,
 				struct drm_amdgpu_query_fw *query_fw,
 				struct amdgpu_device *adev)
@@ -323,6 +258,14 @@ static int amdgpu_firmware_info(struct drm_amdgpu_info_firmware *fw_info,
 		fw_info->ver = adev->gfx.rlc_srls_fw_version;
 		fw_info->feature = adev->gfx.rlc_srls_feature_version;
 		break;
+	case AMDGPU_INFO_FW_GFX_RLCP:
+		fw_info->ver = adev->gfx.rlcp_ucode_version;
+		fw_info->feature = adev->gfx.rlcp_ucode_feature_version;
+		break;
+	case AMDGPU_INFO_FW_GFX_RLCV:
+		fw_info->ver = adev->gfx.rlcv_ucode_version;
+		fw_info->feature = adev->gfx.rlcv_ucode_feature_version;
+		break;
 	case AMDGPU_INFO_FW_GFX_MEC:
 		if (query_fw->index == 0) {
 			fw_info->ver = adev->gfx.mec_fw_version;
@@ -400,12 +343,54 @@ static int amdgpu_firmware_info(struct drm_amdgpu_info_firmware *fw_info,
 		fw_info->ver = adev->psp.toc.fw_version;
 		fw_info->feature = adev->psp.toc.feature_version;
 		break;
+	case AMDGPU_INFO_FW_CAP:
+		fw_info->ver = adev->psp.cap_fw_version;
+		fw_info->feature = adev->psp.cap_feature_version;
+		break;
+	case AMDGPU_INFO_FW_MES_KIQ:
+		fw_info->ver = adev->mes.kiq_version & AMDGPU_MES_VERSION_MASK;
+		fw_info->feature = (adev->mes.kiq_version & AMDGPU_MES_FEAT_VERSION_MASK)
+					>> AMDGPU_MES_FEAT_VERSION_SHIFT;
+		break;
+	case AMDGPU_INFO_FW_MES:
+		fw_info->ver = adev->mes.sched_version & AMDGPU_MES_VERSION_MASK;
+		fw_info->feature = (adev->mes.sched_version & AMDGPU_MES_FEAT_VERSION_MASK)
+					>> AMDGPU_MES_FEAT_VERSION_SHIFT;
+		break;
+	case AMDGPU_INFO_FW_IMU:
+		fw_info->ver = adev->gfx.imu_fw_version;
+		fw_info->feature = 0;
+		break;
+	case AMDGPU_INFO_FW_VPE:
+		fw_info->ver = adev->vpe.fw_version;
+		fw_info->feature = adev->vpe.feature_version;
+		break;
 	default:
 		return -EINVAL;
 	}
 	return 0;
 }
 
+static int amdgpu_userq_metadata_info_gfx(struct amdgpu_device *adev,
+					  struct drm_amdgpu_info *info,
+					  struct drm_amdgpu_info_uq_metadata_gfx *meta)
+{
+	int ret = -EOPNOTSUPP;
+
+	if (adev->gfx.funcs->get_gfx_shadow_info) {
+		struct amdgpu_gfx_shadow_info shadow = {};
+
+		adev->gfx.funcs->get_gfx_shadow_info(adev, &shadow, true);
+		meta->shadow_size = shadow.shadow_size;
+		meta->shadow_alignment = shadow.shadow_alignment;
+		meta->csa_size = shadow.csa_size;
+		meta->csa_alignment = shadow.csa_alignment;
+		ret = 0;
+	}
+
+	return ret;
+}
+
 static int amdgpu_hw_ip_info(struct amdgpu_device *adev,
 			     struct drm_amdgpu_info *info,
 			     struct drm_amdgpu_info_hw_ip *result)
@@ -414,6 +399,7 @@ static int amdgpu_hw_ip_info(struct amdgpu_device *adev,
 	uint32_t ib_size_alignment = 0;
 	enum amd_ip_block_type type;
 	unsigned int num_rings = 0;
+	uint32_t num_slots = 0;
 	unsigned int i, j;
 
 	if (info->query_hw_ip.ip_instance >= AMDGPU_HW_IP_INSTANCE_MAX_COUNT)
@@ -423,24 +409,45 @@ static int amdgpu_hw_ip_info(struct amdgpu_device *adev,
 	case AMDGPU_HW_IP_GFX:
 		type = AMD_IP_BLOCK_TYPE_GFX;
 		for (i = 0; i < adev->gfx.num_gfx_rings; i++)
-			if (adev->gfx.gfx_ring[i].sched.ready)
+			if (adev->gfx.gfx_ring[i].sched.ready &&
+			    !adev->gfx.gfx_ring[i].no_user_submission)
 				++num_rings;
+
+		if (!adev->gfx.disable_uq) {
+			for (i = 0; i < AMDGPU_MES_MAX_GFX_PIPES; i++)
+				num_slots += hweight32(adev->mes.gfx_hqd_mask[i]);
+		}
+
 		ib_start_alignment = 32;
 		ib_size_alignment = 32;
 		break;
 	case AMDGPU_HW_IP_COMPUTE:
 		type = AMD_IP_BLOCK_TYPE_GFX;
 		for (i = 0; i < adev->gfx.num_compute_rings; i++)
-			if (adev->gfx.compute_ring[i].sched.ready)
+			if (adev->gfx.compute_ring[i].sched.ready &&
+			    !adev->gfx.compute_ring[i].no_user_submission)
 				++num_rings;
+
+		if (!adev->sdma.disable_uq) {
+			for (i = 0; i < AMDGPU_MES_MAX_COMPUTE_PIPES; i++)
+				num_slots += hweight32(adev->mes.compute_hqd_mask[i]);
+		}
+
 		ib_start_alignment = 32;
 		ib_size_alignment = 32;
 		break;
 	case AMDGPU_HW_IP_DMA:
 		type = AMD_IP_BLOCK_TYPE_SDMA;
 		for (i = 0; i < adev->sdma.num_instances; i++)
-			if (adev->sdma.instance[i].ring.sched.ready)
+			if (adev->sdma.instance[i].ring.sched.ready &&
+			    !adev->sdma.instance[i].ring.no_user_submission)
 				++num_rings;
+
+		if (!adev->gfx.disable_uq) {
+			for (i = 0; i < AMDGPU_MES_MAX_SDMA_PIPES; i++)
+				num_slots += hweight32(adev->mes.sdma_hqd_mask[i]);
+		}
+
 		ib_start_alignment = 256;
 		ib_size_alignment = 4;
 		break;
@@ -450,19 +457,21 @@ static int amdgpu_hw_ip_info(struct amdgpu_device *adev,
 			if (adev->uvd.harvest_config & (1 << i))
 				continue;
 
-			if (adev->uvd.inst[i].ring.sched.ready)
+			if (adev->uvd.inst[i].ring.sched.ready &&
+			    !adev->uvd.inst[i].ring.no_user_submission)
 				++num_rings;
 		}
-		ib_start_alignment = 64;
+		ib_start_alignment = 256;
 		ib_size_alignment = 64;
 		break;
 	case AMDGPU_HW_IP_VCE:
 		type = AMD_IP_BLOCK_TYPE_VCE;
 		for (i = 0; i < adev->vce.num_rings; i++)
-			if (adev->vce.ring[i].sched.ready)
+			if (adev->vce.ring[i].sched.ready &&
+			    !adev->vce.ring[i].no_user_submission)
 				++num_rings;
-		ib_start_alignment = 4;
-		ib_size_alignment = 1;
+		ib_start_alignment = 256;
+		ib_size_alignment = 4;
 		break;
 	case AMDGPU_HW_IP_UVD_ENC:
 		type = AMD_IP_BLOCK_TYPE_UVD;
@@ -471,36 +480,39 @@ static int amdgpu_hw_ip_info(struct amdgpu_device *adev,
 				continue;
 
 			for (j = 0; j < adev->uvd.num_enc_rings; j++)
-				if (adev->uvd.inst[i].ring_enc[j].sched.ready)
+				if (adev->uvd.inst[i].ring_enc[j].sched.ready &&
+				    !adev->uvd.inst[i].ring_enc[j].no_user_submission)
 					++num_rings;
 		}
-		ib_start_alignment = 64;
-		ib_size_alignment = 64;
+		ib_start_alignment = 256;
+		ib_size_alignment = 4;
 		break;
 	case AMDGPU_HW_IP_VCN_DEC:
 		type = AMD_IP_BLOCK_TYPE_VCN;
 		for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
-			if (adev->uvd.harvest_config & (1 << i))
+			if (adev->vcn.harvest_config & (1 << i))
 				continue;
 
-			if (adev->vcn.inst[i].ring_dec.sched.ready)
+			if (adev->vcn.inst[i].ring_dec.sched.ready &&
+			    !adev->vcn.inst[i].ring_dec.no_user_submission)
 				++num_rings;
 		}
-		ib_start_alignment = 16;
-		ib_size_alignment = 16;
+		ib_start_alignment = 256;
+		ib_size_alignment = 64;
 		break;
 	case AMDGPU_HW_IP_VCN_ENC:
 		type = AMD_IP_BLOCK_TYPE_VCN;
 		for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
-			if (adev->uvd.harvest_config & (1 << i))
+			if (adev->vcn.harvest_config & (1 << i))
 				continue;
 
-			for (j = 0; j < adev->vcn.num_enc_rings; j++)
-				if (adev->vcn.inst[i].ring_enc[j].sched.ready)
+			for (j = 0; j < adev->vcn.inst[i].num_enc_rings; j++)
+				if (adev->vcn.inst[i].ring_enc[j].sched.ready &&
+				    !adev->vcn.inst[i].ring_enc[j].no_user_submission)
 					++num_rings;
 		}
-		ib_start_alignment = 64;
-		ib_size_alignment = 1;
+		ib_start_alignment = 256;
+		ib_size_alignment = 4;
 		break;
 	case AMDGPU_HW_IP_VCN_JPEG:
 		type = (amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_JPEG)) ?
@@ -510,11 +522,21 @@ static int amdgpu_hw_ip_info(struct amdgpu_device *adev,
 			if (adev->jpeg.harvest_config & (1 << i))
 				continue;
 
-			if (adev->jpeg.inst[i].ring_dec.sched.ready)
-				++num_rings;
+			for (j = 0; j < adev->jpeg.num_jpeg_rings; j++)
+				if (adev->jpeg.inst[i].ring_dec[j].sched.ready &&
+				    !adev->jpeg.inst[i].ring_dec[j].no_user_submission)
+					++num_rings;
 		}
-		ib_start_alignment = 16;
-		ib_size_alignment = 16;
+		ib_start_alignment = 256;
+		ib_size_alignment = 64;
+		break;
+	case AMDGPU_HW_IP_VPE:
+		type = AMD_IP_BLOCK_TYPE_VPE;
+		if (adev->vpe.ring.sched.ready &&
+		    !adev->vpe.ring.no_user_submission)
+			++num_rings;
+		ib_start_alignment = 256;
+		ib_size_alignment = 4;
 		break;
 	default:
 		return -EINVAL;
@@ -533,8 +555,41 @@ static int amdgpu_hw_ip_info(struct amdgpu_device *adev,
 
 	result->hw_ip_version_major = adev->ip_blocks[i].version->major;
 	result->hw_ip_version_minor = adev->ip_blocks[i].version->minor;
+
+	if (adev->asic_type >= CHIP_VEGA10) {
+		switch (type) {
+		case AMD_IP_BLOCK_TYPE_GFX:
+			result->ip_discovery_version =
+				IP_VERSION_MAJ_MIN_REV(amdgpu_ip_version(adev, GC_HWIP, 0));
+			break;
+		case AMD_IP_BLOCK_TYPE_SDMA:
+			result->ip_discovery_version =
+				IP_VERSION_MAJ_MIN_REV(amdgpu_ip_version(adev, SDMA0_HWIP, 0));
+			break;
+		case AMD_IP_BLOCK_TYPE_UVD:
+		case AMD_IP_BLOCK_TYPE_VCN:
+		case AMD_IP_BLOCK_TYPE_JPEG:
+			result->ip_discovery_version =
+				IP_VERSION_MAJ_MIN_REV(amdgpu_ip_version(adev, UVD_HWIP, 0));
+			break;
+		case AMD_IP_BLOCK_TYPE_VCE:
+			result->ip_discovery_version =
+				IP_VERSION_MAJ_MIN_REV(amdgpu_ip_version(adev, VCE_HWIP, 0));
+			break;
+		case AMD_IP_BLOCK_TYPE_VPE:
+			result->ip_discovery_version =
+				IP_VERSION_MAJ_MIN_REV(amdgpu_ip_version(adev, VPE_HWIP, 0));
+			break;
+		default:
+			result->ip_discovery_version = 0;
+			break;
+		}
+	} else {
+		result->ip_discovery_version = 0;
+	}
 	result->capabilities_flags = 0;
 	result->available_rings = (1 << num_rings) - 1;
+	result->userq_num_slots = num_slots;
 	result->ib_start_alignment = ib_start_alignment;
 	result->ib_size_alignment = ib_size_alignment;
 	return 0;
@@ -561,11 +616,16 @@ int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 	struct drm_amdgpu_info *info = data;
 	struct amdgpu_mode_info *minfo = &adev->mode_info;
 	void __user *out = (void __user *)(uintptr_t)info->return_pointer;
+	struct amdgpu_fpriv *fpriv;
+	struct amdgpu_ip_block *ip_block;
+	enum amd_ip_block_type type;
+	struct amdgpu_xcp *xcp;
+	u32 count, inst_mask;
 	uint32_t size = info->return_size;
 	struct drm_crtc *crtc;
 	uint32_t ui32 = 0;
 	uint64_t ui64 = 0;
-	int i, found;
+	int i, found, ret;
 	int ui32_size = sizeof(ui32);
 
 	if (!info->return_size || !info->return_pointer)
@@ -580,6 +640,7 @@ int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 			crtc = (struct drm_crtc *)minfo->crtcs[i];
 			if (crtc && crtc->base.id == info->mode_crtc.id) {
 				struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
+
 				ui32 = amdgpu_crtc->crtc_id;
 				found = 1;
 				break;
@@ -592,56 +653,82 @@ int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 		return copy_to_user(out, &ui32, min(size, 4u)) ? -EFAULT : 0;
 	case AMDGPU_INFO_HW_IP_INFO: {
 		struct drm_amdgpu_info_hw_ip ip = {};
-		int ret;
 
 		ret = amdgpu_hw_ip_info(adev, info, &ip);
 		if (ret)
 			return ret;
 
-		ret = copy_to_user(out, &ip, min((size_t)size, sizeof(ip)));
+		ret = copy_to_user(out, &ip, min_t(size_t, size, sizeof(ip)));
 		return ret ? -EFAULT : 0;
 	}
 	case AMDGPU_INFO_HW_IP_COUNT: {
-		enum amd_ip_block_type type;
-		uint32_t count = 0;
+		fpriv = (struct amdgpu_fpriv *)filp->driver_priv;
+		type = amdgpu_ip_get_block_type(adev, info->query_hw_ip.type);
+		ip_block = amdgpu_device_ip_get_ip_block(adev, type);
 
-		switch (info->query_hw_ip.type) {
-		case AMDGPU_HW_IP_GFX:
-			type = AMD_IP_BLOCK_TYPE_GFX;
-			break;
-		case AMDGPU_HW_IP_COMPUTE:
-			type = AMD_IP_BLOCK_TYPE_GFX;
-			break;
-		case AMDGPU_HW_IP_DMA:
-			type = AMD_IP_BLOCK_TYPE_SDMA;
-			break;
-		case AMDGPU_HW_IP_UVD:
-			type = AMD_IP_BLOCK_TYPE_UVD;
+		if (!ip_block || !ip_block->status.valid)
+			return -EINVAL;
+
+		if (adev->xcp_mgr && adev->xcp_mgr->num_xcps > 0 &&
+		    fpriv->xcp_id < adev->xcp_mgr->num_xcps) {
+			xcp = &adev->xcp_mgr->xcp[fpriv->xcp_id];
+			switch (type) {
+			case AMD_IP_BLOCK_TYPE_GFX:
+				ret = amdgpu_xcp_get_inst_details(xcp, AMDGPU_XCP_GFX, &inst_mask);
+				if (ret)
+					return ret;
+				count = hweight32(inst_mask);
+				break;
+			case AMD_IP_BLOCK_TYPE_SDMA:
+				ret = amdgpu_xcp_get_inst_details(xcp, AMDGPU_XCP_SDMA, &inst_mask);
+				if (ret)
+					return ret;
+				count = hweight32(inst_mask);
+				break;
+			case AMD_IP_BLOCK_TYPE_JPEG:
+				ret = amdgpu_xcp_get_inst_details(xcp, AMDGPU_XCP_VCN, &inst_mask);
+				if (ret)
+					return ret;
+				count = hweight32(inst_mask) * adev->jpeg.num_jpeg_rings;
+				break;
+			case AMD_IP_BLOCK_TYPE_VCN:
+				ret = amdgpu_xcp_get_inst_details(xcp, AMDGPU_XCP_VCN, &inst_mask);
+				if (ret)
+					return ret;
+				count = hweight32(inst_mask);
+				break;
+			default:
+				return -EINVAL;
+			}
+
+			return copy_to_user(out, &count, min(size, 4u)) ? -EFAULT : 0;
+		}
+
+		switch (type) {
+		case AMD_IP_BLOCK_TYPE_GFX:
+		case AMD_IP_BLOCK_TYPE_VCE:
+			count = 1;
 			break;
-		case AMDGPU_HW_IP_VCE:
-			type = AMD_IP_BLOCK_TYPE_VCE;
+		case AMD_IP_BLOCK_TYPE_SDMA:
+			count = adev->sdma.num_instances;
 			break;
-		case AMDGPU_HW_IP_UVD_ENC:
-			type = AMD_IP_BLOCK_TYPE_UVD;
+		case AMD_IP_BLOCK_TYPE_JPEG:
+			count = adev->jpeg.num_jpeg_inst * adev->jpeg.num_jpeg_rings;
 			break;
-		case AMDGPU_HW_IP_VCN_DEC:
-		case AMDGPU_HW_IP_VCN_ENC:
-			type = AMD_IP_BLOCK_TYPE_VCN;
+		case AMD_IP_BLOCK_TYPE_VCN:
+			count = adev->vcn.num_vcn_inst;
 			break;
-		case AMDGPU_HW_IP_VCN_JPEG:
-			type = (amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_JPEG)) ?
-				AMD_IP_BLOCK_TYPE_JPEG : AMD_IP_BLOCK_TYPE_VCN;
+		case AMD_IP_BLOCK_TYPE_UVD:
+			count = adev->uvd.num_uvd_inst;
 			break;
+		/* For all other IP block types not listed in the switch statement
+		 * the ip status is valid here and the instance count is one.
+		 */
 		default:
-			return -EINVAL;
+			count = 1;
+			break;
 		}
 
-		for (i = 0; i < adev->num_ip_blocks; i++)
-			if (adev->ip_blocks[i].version->type == type &&
-			    adev->ip_blocks[i].status.valid &&
-			    count < AMDGPU_HW_IP_INSTANCE_MAX_COUNT)
-				count++;
-
 		return copy_to_user(out, &count, min(size, 4u)) ? -EFAULT : 0;
 	}
 	case AMDGPU_INFO_TIMESTAMP:
@@ -649,7 +736,6 @@ int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 		return copy_to_user(out, &ui64, min(size, 8u)) ? -EFAULT : 0;
 	case AMDGPU_INFO_FW_VERSION: {
 		struct drm_amdgpu_info_firmware fw_info;
-		int ret;
 
 		/* We only support one instance of each IP block right now. */
 		if (info->query_fw.ip_instance != 0)
@@ -672,13 +758,14 @@ int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 		ui64 = atomic64_read(&adev->num_vram_cpu_page_faults);
 		return copy_to_user(out, &ui64, min(size, 8u)) ? -EFAULT : 0;
 	case AMDGPU_INFO_VRAM_USAGE:
-		ui64 = amdgpu_vram_mgr_usage(ttm_manager_type(&adev->mman.bdev, TTM_PL_VRAM));
+		ui64 = ttm_resource_manager_used(&adev->mman.vram_mgr.manager) ?
+			ttm_resource_manager_usage(&adev->mman.vram_mgr.manager) : 0;
 		return copy_to_user(out, &ui64, min(size, 8u)) ? -EFAULT : 0;
 	case AMDGPU_INFO_VIS_VRAM_USAGE:
-		ui64 = amdgpu_vram_mgr_vis_usage(ttm_manager_type(&adev->mman.bdev, TTM_PL_VRAM));
+		ui64 = amdgpu_vram_mgr_vis_usage(&adev->mman.vram_mgr);
 		return copy_to_user(out, &ui64, min(size, 8u)) ? -EFAULT : 0;
 	case AMDGPU_INFO_GTT_USAGE:
-		ui64 = amdgpu_gtt_mgr_usage(ttm_manager_type(&adev->mman.bdev, TTM_PL_TT));
+		ui64 = ttm_resource_manager_usage(&adev->mman.gtt_mgr.manager);
 		return copy_to_user(out, &ui64, min(size, 8u)) ? -EFAULT : 0;
 	case AMDGPU_INFO_GDS_CONFIG: {
 		struct drm_amdgpu_info_gds gds_info;
@@ -702,24 +789,24 @@ int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 			    atomic64_read(&adev->visible_pin_size),
 			    vram_gtt.vram_size);
 		vram_gtt.gtt_size = ttm_manager_type(&adev->mman.bdev, TTM_PL_TT)->size;
-		vram_gtt.gtt_size *= PAGE_SIZE;
 		vram_gtt.gtt_size -= atomic64_read(&adev->gart_pin_size);
 		return copy_to_user(out, &vram_gtt,
 				    min((size_t)size, sizeof(vram_gtt))) ? -EFAULT : 0;
 	}
 	case AMDGPU_INFO_MEMORY: {
 		struct drm_amdgpu_memory_info mem;
-		struct ttm_resource_manager *vram_man =
-			ttm_manager_type(&adev->mman.bdev, TTM_PL_VRAM);
 		struct ttm_resource_manager *gtt_man =
-			ttm_manager_type(&adev->mman.bdev, TTM_PL_TT);
+			&adev->mman.gtt_mgr.manager;
+		struct ttm_resource_manager *vram_man =
+			&adev->mman.vram_mgr.manager;
+
 		memset(&mem, 0, sizeof(mem));
 		mem.vram.total_heap_size = adev->gmc.real_vram_size;
 		mem.vram.usable_heap_size = adev->gmc.real_vram_size -
 			atomic64_read(&adev->vram_pin_size) -
 			AMDGPU_VM_RESERVED_VRAM;
-		mem.vram.heap_usage =
-			amdgpu_vram_mgr_usage(vram_man);
+		mem.vram.heap_usage = ttm_resource_manager_used(&adev->mman.vram_mgr.manager) ?
+				ttm_resource_manager_usage(vram_man) : 0;
 		mem.vram.max_allocation = mem.vram.usable_heap_size * 3 / 4;
 
 		mem.cpu_accessible_vram.total_heap_size =
@@ -729,16 +816,14 @@ int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 			    atomic64_read(&adev->visible_pin_size),
 			    mem.vram.usable_heap_size);
 		mem.cpu_accessible_vram.heap_usage =
-			amdgpu_vram_mgr_vis_usage(vram_man);
+			amdgpu_vram_mgr_vis_usage(&adev->mman.vram_mgr);
 		mem.cpu_accessible_vram.max_allocation =
 			mem.cpu_accessible_vram.usable_heap_size * 3 / 4;
 
 		mem.gtt.total_heap_size = gtt_man->size;
-		mem.gtt.total_heap_size *= PAGE_SIZE;
 		mem.gtt.usable_heap_size = mem.gtt.total_heap_size -
 			atomic64_read(&adev->gart_pin_size);
-		mem.gtt.heap_usage =
-			amdgpu_gtt_mgr_usage(gtt_man);
+		mem.gtt.heap_usage = ttm_resource_manager_usage(gtt_man);
 		mem.gtt.max_allocation = mem.gtt.usable_heap_size * 3 / 4;
 
 		return copy_to_user(out, &mem,
@@ -746,32 +831,47 @@ int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 				    ? -EFAULT : 0;
 	}
 	case AMDGPU_INFO_READ_MMR_REG: {
-		unsigned n, alloc_size;
+		int ret = 0;
+		unsigned int n, alloc_size;
 		uint32_t *regs;
-		unsigned se_num = (info->read_mmr_reg.instance >>
+		unsigned int se_num = (info->read_mmr_reg.instance >>
 				   AMDGPU_INFO_MMR_SE_INDEX_SHIFT) &
 				  AMDGPU_INFO_MMR_SE_INDEX_MASK;
-		unsigned sh_num = (info->read_mmr_reg.instance >>
+		unsigned int sh_num = (info->read_mmr_reg.instance >>
 				   AMDGPU_INFO_MMR_SH_INDEX_SHIFT) &
 				  AMDGPU_INFO_MMR_SH_INDEX_MASK;
 
+		if (!down_read_trylock(&adev->reset_domain->sem))
+			return -ENOENT;
+
 		/* set full masks if the userspace set all bits
-		 * in the bitfields */
-		if (se_num == AMDGPU_INFO_MMR_SE_INDEX_MASK)
+		 * in the bitfields
+		 */
+		if (se_num == AMDGPU_INFO_MMR_SE_INDEX_MASK) {
 			se_num = 0xffffffff;
-		else if (se_num >= AMDGPU_GFX_MAX_SE)
-			return -EINVAL;
-		if (sh_num == AMDGPU_INFO_MMR_SH_INDEX_MASK)
+		} else if (se_num >= AMDGPU_GFX_MAX_SE) {
+			ret = -EINVAL;
+			goto out;
+		}
+
+		if (sh_num == AMDGPU_INFO_MMR_SH_INDEX_MASK) {
 			sh_num = 0xffffffff;
-		else if (sh_num >= AMDGPU_GFX_MAX_SH_PER_SE)
-			return -EINVAL;
+		} else if (sh_num >= AMDGPU_GFX_MAX_SH_PER_SE) {
+			ret = -EINVAL;
+			goto out;
+		}
 
-		if (info->read_mmr_reg.count > 128)
-			return -EINVAL;
+		if (info->read_mmr_reg.count > 128) {
+			ret = -EINVAL;
+			goto out;
+		}
 
 		regs = kmalloc_array(info->read_mmr_reg.count, sizeof(*regs), GFP_KERNEL);
-		if (!regs)
-			return -ENOMEM;
+		if (!regs) {
+			ret = -ENOMEM;
+			goto out;
+		}
+
 		alloc_size = info->read_mmr_reg.count * sizeof(*regs);
 
 		amdgpu_gfx_off_ctrl(adev, false);
@@ -783,18 +883,22 @@ int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 					      info->read_mmr_reg.dword_offset + i);
 				kfree(regs);
 				amdgpu_gfx_off_ctrl(adev, true);
-				return -EFAULT;
+				ret = -EFAULT;
+				goto out;
 			}
 		}
 		amdgpu_gfx_off_ctrl(adev, true);
 		n = copy_to_user(out, regs, min(size, alloc_size));
 		kfree(regs);
-		return n ? -EFAULT : 0;
+		ret = (n ? -EFAULT : 0);
+out:
+		up_read(&adev->reset_domain->sem);
+		return ret;
 	}
 	case AMDGPU_INFO_DEV_INFO: {
 		struct drm_amdgpu_info_device *dev_info;
 		uint64_t vm_size;
-		int ret;
+		uint32_t pcie_gen_mask, pcie_width_mask;
 
 		dev_info = kzalloc(sizeof(*dev_info), GFP_KERNEL);
 		if (!dev_info)
@@ -812,32 +916,52 @@ int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 		if (adev->pm.dpm_enabled) {
 			dev_info->max_engine_clock = amdgpu_dpm_get_sclk(adev, false) * 10;
 			dev_info->max_memory_clock = amdgpu_dpm_get_mclk(adev, false) * 10;
+			dev_info->min_engine_clock = amdgpu_dpm_get_sclk(adev, true) * 10;
+			dev_info->min_memory_clock = amdgpu_dpm_get_mclk(adev, true) * 10;
 		} else {
-			dev_info->max_engine_clock = adev->clock.default_sclk * 10;
-			dev_info->max_memory_clock = adev->clock.default_mclk * 10;
+			dev_info->max_engine_clock =
+				dev_info->min_engine_clock =
+					adev->clock.default_sclk * 10;
+			dev_info->max_memory_clock =
+				dev_info->min_memory_clock =
+					adev->clock.default_mclk * 10;
 		}
 		dev_info->enabled_rb_pipes_mask = adev->gfx.config.backend_enable_mask;
 		dev_info->num_rb_pipes = adev->gfx.config.max_backends_per_se *
 			adev->gfx.config.max_shader_engines;
 		dev_info->num_hw_gfx_contexts = adev->gfx.config.max_hw_contexts;
-		dev_info->_pad = 0;
 		dev_info->ids_flags = 0;
 		if (adev->flags & AMD_IS_APU)
 			dev_info->ids_flags |= AMDGPU_IDS_FLAGS_FUSION;
-		if (amdgpu_mcbp || amdgpu_sriov_vf(adev))
+		if (adev->gfx.mcbp)
 			dev_info->ids_flags |= AMDGPU_IDS_FLAGS_PREEMPTION;
 		if (amdgpu_is_tmz(adev))
 			dev_info->ids_flags |= AMDGPU_IDS_FLAGS_TMZ;
+		if (adev->gfx.config.ta_cntl2_truncate_coord_mode)
+			dev_info->ids_flags |= AMDGPU_IDS_FLAGS_CONFORMANT_TRUNC_COORD;
+
+		/* Gang submit is not supported under SRIOV currently */
+		if (!amdgpu_sriov_vf(adev))
+			dev_info->ids_flags |= AMDGPU_IDS_FLAGS_GANG_SUBMIT;
+
+		if (amdgpu_passthrough(adev))
+			dev_info->ids_flags |= (AMDGPU_IDS_FLAGS_MODE_PT <<
+						AMDGPU_IDS_FLAGS_MODE_SHIFT) &
+						AMDGPU_IDS_FLAGS_MODE_MASK;
+		else if (amdgpu_sriov_vf(adev))
+			dev_info->ids_flags |= (AMDGPU_IDS_FLAGS_MODE_VF <<
+						AMDGPU_IDS_FLAGS_MODE_SHIFT) &
+						AMDGPU_IDS_FLAGS_MODE_MASK;
 
 		vm_size = adev->vm_manager.max_pfn * AMDGPU_GPU_PAGE_SIZE;
-		vm_size -= AMDGPU_VA_RESERVED_SIZE;
+		vm_size -= AMDGPU_VA_RESERVED_TOP;
 
 		/* Older VCE FW versions are buggy and can handle only 40bits */
 		if (adev->vce.fw_version &&
 		    adev->vce.fw_version < AMDGPU_VCE_FW_53_45)
 			vm_size = min(vm_size, 1ULL << 40);
 
-		dev_info->virtual_address_offset = AMDGPU_VA_RESERVED_SIZE;
+		dev_info->virtual_address_offset = AMDGPU_VA_RESERVED_BOTTOM;
 		dev_info->virtual_address_max =
 			min(vm_size, AMDGPU_GMC_HOLE_START);
 
@@ -854,7 +978,7 @@ int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 		memcpy(&dev_info->cu_ao_bitmap[0], &adev->gfx.cu_info.ao_cu_bitmap[0],
 		       sizeof(adev->gfx.cu_info.ao_cu_bitmap));
 		memcpy(&dev_info->cu_bitmap[0], &adev->gfx.cu_info.bitmap[0],
-		       sizeof(adev->gfx.cu_info.bitmap));
+		       sizeof(dev_info->cu_bitmap));
 		dev_info->vram_type = adev->gmc.vram_type;
 		dev_info->vram_bit_width = adev->gmc.vram_width;
 		dev_info->vce_harvest_config = adev->vce.harvest_config;
@@ -874,13 +998,51 @@ int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 
 		dev_info->tcc_disabled_mask = adev->gfx.config.tcc_disabled_mask;
 
+		/* Combine the chip gen mask with the platform (CPU/mobo) mask. */
+		pcie_gen_mask = adev->pm.pcie_gen_mask &
+			(adev->pm.pcie_gen_mask >> CAIL_PCIE_LINK_SPEED_SUPPORT_SHIFT);
+		pcie_width_mask = adev->pm.pcie_mlw_mask &
+			(adev->pm.pcie_mlw_mask >> CAIL_PCIE_LINK_WIDTH_SUPPORT_SHIFT);
+		dev_info->pcie_gen = fls(pcie_gen_mask);
+		dev_info->pcie_num_lanes =
+			pcie_width_mask & CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X32 ? 32 :
+			pcie_width_mask & CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X16 ? 16 :
+			pcie_width_mask & CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X12 ? 12 :
+			pcie_width_mask & CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X8 ? 8 :
+			pcie_width_mask & CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X4 ? 4 :
+			pcie_width_mask & CAIL_ASIC_PCIE_LINK_WIDTH_SUPPORT_X2 ? 2 : 1;
+
+		dev_info->tcp_cache_size = adev->gfx.config.gc_tcp_l1_size;
+		dev_info->num_sqc_per_wgp = adev->gfx.config.gc_num_sqc_per_wgp;
+		dev_info->sqc_data_cache_size = adev->gfx.config.gc_l1_data_cache_size_per_sqc;
+		dev_info->sqc_inst_cache_size = adev->gfx.config.gc_l1_instruction_cache_size_per_sqc;
+		dev_info->gl1c_cache_size = adev->gfx.config.gc_gl1c_size_per_instance *
+					    adev->gfx.config.gc_gl1c_per_sa;
+		dev_info->gl2c_cache_size = adev->gfx.config.gc_gl2c_per_gpu;
+		dev_info->mall_size = adev->gmc.mall_size;
+
+
+		if (adev->gfx.funcs->get_gfx_shadow_info) {
+			struct amdgpu_gfx_shadow_info shadow_info;
+
+			ret = amdgpu_gfx_get_gfx_shadow_info(adev, &shadow_info);
+			if (!ret) {
+				dev_info->shadow_size = shadow_info.shadow_size;
+				dev_info->shadow_alignment = shadow_info.shadow_alignment;
+				dev_info->csa_size = shadow_info.csa_size;
+				dev_info->csa_alignment = shadow_info.csa_alignment;
+			}
+		}
+
+		dev_info->userq_ip_mask = amdgpu_userq_get_supported_ip_mask(adev);
+
 		ret = copy_to_user(out, dev_info,
 				   min((size_t)size, sizeof(*dev_info))) ? -EFAULT : 0;
 		kfree(dev_info);
 		return ret;
 	}
 	case AMDGPU_INFO_VCE_CLOCK_TABLE: {
-		unsigned i;
+		unsigned int i;
 		struct drm_amdgpu_info_vce_clock_table vce_clk_table = {};
 		struct amd_vce_state *vce_state;
 
@@ -922,12 +1084,17 @@ int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 			struct atom_context *atom_context;
 
 			atom_context = adev->mode_info.atom_context;
-			memcpy(vbios_info.name, atom_context->name, sizeof(atom_context->name));
-			memcpy(vbios_info.vbios_pn, atom_context->vbios_pn, sizeof(atom_context->vbios_pn));
-			vbios_info.version = atom_context->version;
-			memcpy(vbios_info.vbios_ver_str, atom_context->vbios_ver_str,
-						sizeof(atom_context->vbios_ver_str));
-			memcpy(vbios_info.date, atom_context->date, sizeof(atom_context->date));
+			if (atom_context) {
+				memcpy(vbios_info.name, atom_context->name,
+				       sizeof(atom_context->name));
+				memcpy(vbios_info.vbios_pn, atom_context->vbios_pn,
+				       sizeof(atom_context->vbios_pn));
+				vbios_info.version = atom_context->version;
+				memcpy(vbios_info.vbios_ver_str, atom_context->vbios_ver_str,
+				       sizeof(atom_context->vbios_ver_str));
+				memcpy(vbios_info.date, atom_context->date,
+				       sizeof(atom_context->date));
+			}
 
 			return copy_to_user(out, &vbios_info,
 						min((size_t)size, sizeof(vbios_info))) ? -EFAULT : 0;
@@ -1001,7 +1168,21 @@ int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 		case AMDGPU_INFO_SENSOR_GPU_AVG_POWER:
 			/* get average GPU power */
 			if (amdgpu_dpm_read_sensor(adev,
-						   AMDGPU_PP_SENSOR_GPU_POWER,
+						   AMDGPU_PP_SENSOR_GPU_AVG_POWER,
+						   (void *)&ui32, &ui32_size)) {
+				/* fall back to input power for backwards compat */
+				if (amdgpu_dpm_read_sensor(adev,
+							   AMDGPU_PP_SENSOR_GPU_INPUT_POWER,
+							   (void *)&ui32, &ui32_size)) {
+					return -EINVAL;
+				}
+			}
+			ui32 >>= 8;
+			break;
+		case AMDGPU_INFO_SENSOR_GPU_INPUT_POWER:
+			/* get input GPU power */
+			if (amdgpu_dpm_read_sensor(adev,
+						   AMDGPU_PP_SENSOR_GPU_INPUT_POWER,
 						   (void *)&ui32, &ui32_size)) {
 				return -EINVAL;
 			}
@@ -1041,6 +1222,24 @@ int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 			}
 			ui32 /= 100;
 			break;
+		case AMDGPU_INFO_SENSOR_PEAK_PSTATE_GFX_SCLK:
+			/* get peak pstate sclk in Mhz */
+			if (amdgpu_dpm_read_sensor(adev,
+						   AMDGPU_PP_SENSOR_PEAK_PSTATE_SCLK,
+						   (void *)&ui32, &ui32_size)) {
+				return -EINVAL;
+			}
+			ui32 /= 100;
+			break;
+		case AMDGPU_INFO_SENSOR_PEAK_PSTATE_GFX_MCLK:
+			/* get peak pstate mclk in Mhz */
+			if (amdgpu_dpm_read_sensor(adev,
+						   AMDGPU_PP_SENSOR_PEAK_PSTATE_MCLK,
+						   (void *)&ui32, &ui32_size)) {
+				return -EINVAL;
+			}
+			ui32 /= 100;
+			break;
 		default:
 			DRM_DEBUG_KMS("Invalid request %d\n",
 				      info->sensor_info.type);
@@ -1068,6 +1267,9 @@ int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 		struct drm_amdgpu_info_video_caps *caps;
 		int r;
 
+		if (!adev->asic_funcs->query_video_codecs)
+			return -EINVAL;
+
 		switch (info->video_cap.type) {
 		case AMDGPU_INFO_VIDEO_CAPS_DECODE:
 			r = amdgpu_asic_query_video_codecs(adev, false, &codecs);
@@ -1120,6 +1322,51 @@ int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 		kfree(caps);
 		return r;
 	}
+	case AMDGPU_INFO_MAX_IBS: {
+		uint32_t max_ibs[AMDGPU_HW_IP_NUM];
+
+		for (i = 0; i < AMDGPU_HW_IP_NUM; ++i)
+			max_ibs[i] = amdgpu_ring_max_ibs(i);
+
+		return copy_to_user(out, max_ibs,
+				    min((size_t)size, sizeof(max_ibs))) ? -EFAULT : 0;
+	}
+	case AMDGPU_INFO_GPUVM_FAULT: {
+		struct amdgpu_fpriv *fpriv = filp->driver_priv;
+		struct amdgpu_vm *vm = &fpriv->vm;
+		struct drm_amdgpu_info_gpuvm_fault gpuvm_fault;
+		unsigned long flags;
+
+		if (!vm)
+			return -EINVAL;
+
+		memset(&gpuvm_fault, 0, sizeof(gpuvm_fault));
+
+		xa_lock_irqsave(&adev->vm_manager.pasids, flags);
+		gpuvm_fault.addr = vm->fault_info.addr;
+		gpuvm_fault.status = vm->fault_info.status;
+		gpuvm_fault.vmhub = vm->fault_info.vmhub;
+		xa_unlock_irqrestore(&adev->vm_manager.pasids, flags);
+
+		return copy_to_user(out, &gpuvm_fault,
+				    min((size_t)size, sizeof(gpuvm_fault))) ? -EFAULT : 0;
+	}
+	case AMDGPU_INFO_UQ_FW_AREAS: {
+		struct drm_amdgpu_info_uq_metadata meta_info = {};
+
+		switch (info->query_hw_ip.type) {
+		case AMDGPU_HW_IP_GFX:
+			ret = amdgpu_userq_metadata_info_gfx(adev, info, &meta_info.gfx);
+			if (ret)
+				return ret;
+
+			ret = copy_to_user(out, &meta_info,
+						min((size_t)size, sizeof(meta_info))) ? -EFAULT : 0;
+			return 0;
+		default:
+			return -EINVAL;
+		}
+	}
 	default:
 		DRM_DEBUG_KMS("Invalid request %d\n", info->query);
 		return -EINVAL;
@@ -1127,23 +1374,6 @@ int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 	return 0;
 }
 
-
-/*
- * Outdated mess for old drm with Xorg being in charge (void function now).
- */
-/**
- * amdgpu_driver_lastclose_kms - drm callback for last close
- *
- * @dev: drm dev pointer
- *
- * Switch vga_switcheroo state after last close (all asics).
- */
-void amdgpu_driver_lastclose_kms(struct drm_device *dev)
-{
-	drm_fb_helper_lastclose(dev);
-	vga_switcheroo_process_delayed_switch();
-}
-
 /**
  * amdgpu_driver_open_kms - drm callback for open
  *
@@ -1186,13 +1416,15 @@ int amdgpu_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv)
 		pasid = 0;
 	}
 
-	r = amdgpu_vm_init(adev, &fpriv->vm);
+	r = amdgpu_xcp_open_device(adev, fpriv, file_priv);
 	if (r)
 		goto error_pasid;
 
-	r = amdgpu_vm_set_pasid(adev, &fpriv->vm, pasid);
+	amdgpu_debugfs_vm_init(file_priv);
+
+	r = amdgpu_vm_init(adev, &fpriv->vm, fpriv->xcp_id, pasid);
 	if (r)
-		goto error_vm;
+		goto error_pasid;
 
 	fpriv->prt_va = amdgpu_vm_bo_add(adev, &fpriv->vm, NULL);
 	if (!fpriv->prt_va) {
@@ -1200,7 +1432,7 @@ int amdgpu_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv)
 		goto error_vm;
 	}
 
-	if (amdgpu_mcbp || amdgpu_sriov_vf(adev)) {
+	if (adev->gfx.mcbp) {
 		uint64_t csa_addr = amdgpu_csa_vaddr(adev) & AMDGPU_GMC_HOLE_MASK;
 
 		r = amdgpu_map_static_csa(adev, &fpriv->vm, adev->virt.csa_obj,
@@ -1209,10 +1441,22 @@ int amdgpu_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv)
 			goto error_vm;
 	}
 
+	r = amdgpu_seq64_map(adev, &fpriv->vm, &fpriv->seq64_va);
+	if (r)
+		goto error_vm;
+
 	mutex_init(&fpriv->bo_list_lock);
-	idr_init(&fpriv->bo_list_handles);
+	idr_init_base(&fpriv->bo_list_handles, 1);
+
+	r = amdgpu_userq_mgr_init(&fpriv->userq_mgr, file_priv, adev);
+	if (r)
+		DRM_WARN("Can't setup usermode queues, use legacy workload submission only\n");
 
-	amdgpu_ctx_mgr_init(&fpriv->ctx_mgr);
+	r = amdgpu_eviction_fence_init(&fpriv->evf_mgr);
+	if (r)
+		goto error_vm;
+
+	amdgpu_ctx_mgr_init(&fpriv->ctx_mgr, adev);
 
 	file_priv->driver_priv = fpriv;
 	goto out_suspend;
@@ -1221,15 +1465,12 @@ error_vm:
 	amdgpu_vm_fini(adev, &fpriv->vm);
 
 error_pasid:
-	if (pasid) {
+	if (pasid)
 		amdgpu_pasid_free(pasid);
-		amdgpu_vm_set_pasid(adev, &fpriv->vm, 0);
-	}
 
 	kfree(fpriv);
 
 out_suspend:
-	pm_runtime_mark_last_busy(dev->dev);
 pm_put:
 	pm_runtime_put_autosuspend(dev->dev);
 
@@ -1264,18 +1505,22 @@ void amdgpu_driver_postclose_kms(struct drm_device *dev,
 	if (amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_VCE) != NULL)
 		amdgpu_vce_free_handles(adev, file_priv);
 
-	amdgpu_vm_bo_rmv(adev, fpriv->prt_va);
+	if (fpriv->csa_va) {
+		uint64_t csa_addr = amdgpu_csa_vaddr(adev) & AMDGPU_GMC_HOLE_MASK;
 
-	if (amdgpu_mcbp || amdgpu_sriov_vf(adev)) {
-		/* TODO: how to handle reserve failure */
-		BUG_ON(amdgpu_bo_reserve(adev->virt.csa_obj, true));
-		amdgpu_vm_bo_rmv(adev, fpriv->csa_va);
+		WARN_ON(amdgpu_unmap_static_csa(adev, &fpriv->vm, adev->virt.csa_obj,
+						fpriv->csa_va, csa_addr));
 		fpriv->csa_va = NULL;
-		amdgpu_bo_unreserve(adev->virt.csa_obj);
 	}
 
+	amdgpu_seq64_unmap(adev, fpriv);
+
 	pasid = fpriv->vm.pasid;
 	pd = amdgpu_bo_ref(fpriv->vm.root.bo);
+	if (!WARN_ON(amdgpu_bo_reserve(pd, true))) {
+		amdgpu_vm_bo_del(adev, fpriv->prt_va);
+		amdgpu_bo_unreserve(pd);
+	}
 
 	amdgpu_ctx_mgr_fini(&fpriv->ctx_mgr);
 	amdgpu_vm_fini(adev, &fpriv->vm);
@@ -1293,7 +1538,6 @@ void amdgpu_driver_postclose_kms(struct drm_device *dev,
 	kfree(fpriv);
 	file_priv->driver_priv = NULL;
 
-	pm_runtime_mark_last_busy(dev->dev);
 	pm_runtime_put_autosuspend(dev->dev);
 }
 
@@ -1419,16 +1663,15 @@ void amdgpu_disable_vblank_kms(struct drm_crtc *crtc)
 
 static int amdgpu_debugfs_firmware_info_show(struct seq_file *m, void *unused)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)m->private;
+	struct amdgpu_device *adev = m->private;
 	struct drm_amdgpu_info_firmware fw_info;
 	struct drm_amdgpu_query_fw query_fw;
 	struct atom_context *ctx = adev->mode_info.atom_context;
-	uint8_t smu_minor, smu_debug;
-	uint16_t smu_major;
+	uint8_t smu_program, smu_major, smu_minor, smu_debug;
 	int ret, i;
 
 	static const char *ta_fw_name[TA_FW_TYPE_MAX_INDEX] = {
-#define TA_FW_NAME(type) [TA_FW_TYPE_PSP_##type] = #type
+#define TA_FW_NAME(type)[TA_FW_TYPE_PSP_##type] = #type
 		TA_FW_NAME(XGMI),
 		TA_FW_NAME(RAS),
 		TA_FW_NAME(HDCP),
@@ -1518,6 +1761,22 @@ static int amdgpu_debugfs_firmware_info_show(struct seq_file *m, void *unused)
 	seq_printf(m, "RLC SRLS feature version: %u, firmware version: 0x%08x\n",
 		   fw_info.feature, fw_info.ver);
 
+	/* RLCP */
+	query_fw.fw_type = AMDGPU_INFO_FW_GFX_RLCP;
+	ret = amdgpu_firmware_info(&fw_info, &query_fw, adev);
+	if (ret)
+		return ret;
+	seq_printf(m, "RLCP feature version: %u, firmware version: 0x%08x\n",
+		   fw_info.feature, fw_info.ver);
+
+	/* RLCV */
+	query_fw.fw_type = AMDGPU_INFO_FW_GFX_RLCV;
+	ret = amdgpu_firmware_info(&fw_info, &query_fw, adev);
+	if (ret)
+		return ret;
+	seq_printf(m, "RLCV feature version: %u, firmware version: 0x%08x\n",
+		   fw_info.feature, fw_info.ver);
+
 	/* MEC */
 	query_fw.fw_type = AMDGPU_INFO_FW_GFX_MEC;
 	query_fw.index = 0;
@@ -1537,6 +1796,15 @@ static int amdgpu_debugfs_firmware_info_show(struct seq_file *m, void *unused)
 			   fw_info.feature, fw_info.ver);
 	}
 
+	/* IMU */
+	query_fw.fw_type = AMDGPU_INFO_FW_IMU;
+	query_fw.index = 0;
+	ret = amdgpu_firmware_info(&fw_info, &query_fw, adev);
+	if (ret)
+		return ret;
+	seq_printf(m, "IMU feature version: %u, firmware version: 0x%08x\n",
+		   fw_info.feature, fw_info.ver);
+
 	/* PSP SOS */
 	query_fw.fw_type = AMDGPU_INFO_FW_SOS;
 	ret = amdgpu_firmware_info(&fw_info, &query_fw, adev);
@@ -1570,11 +1838,12 @@ static int amdgpu_debugfs_firmware_info_show(struct seq_file *m, void *unused)
 	ret = amdgpu_firmware_info(&fw_info, &query_fw, adev);
 	if (ret)
 		return ret;
-	smu_major = (fw_info.ver >> 16) & 0xffff;
+	smu_program = (fw_info.ver >> 24) & 0xff;
+	smu_major = (fw_info.ver >> 16) & 0xff;
 	smu_minor = (fw_info.ver >> 8) & 0xff;
 	smu_debug = (fw_info.ver >> 0) & 0xff;
-	seq_printf(m, "SMC feature version: %u, firmware version: 0x%08x (%d.%d.%d)\n",
-		   fw_info.feature, fw_info.ver, smu_major, smu_minor, smu_debug);
+	seq_printf(m, "SMC feature version: %u, program: %d, firmware version: 0x%08x (%d.%d.%d)\n",
+		   fw_info.feature, smu_program, fw_info.ver, smu_major, smu_minor, smu_debug);
 
 	/* SDMA */
 	query_fw.fw_type = AMDGPU_INFO_FW_SDMA;
@@ -1619,7 +1888,41 @@ static int amdgpu_debugfs_firmware_info_show(struct seq_file *m, void *unused)
 	seq_printf(m, "TOC feature version: %u, firmware version: 0x%08x\n",
 		   fw_info.feature, fw_info.ver);
 
-	seq_printf(m, "VBIOS version: %s\n", ctx->vbios_version);
+	/* CAP */
+	if (adev->psp.cap_fw) {
+		query_fw.fw_type = AMDGPU_INFO_FW_CAP;
+		ret = amdgpu_firmware_info(&fw_info, &query_fw, adev);
+		if (ret)
+			return ret;
+		seq_printf(m, "CAP feature version: %u, firmware version: 0x%08x\n",
+				fw_info.feature, fw_info.ver);
+	}
+
+	/* MES_KIQ */
+	query_fw.fw_type = AMDGPU_INFO_FW_MES_KIQ;
+	ret = amdgpu_firmware_info(&fw_info, &query_fw, adev);
+	if (ret)
+		return ret;
+	seq_printf(m, "MES_KIQ feature version: %u, firmware version: 0x%08x\n",
+		   fw_info.feature, fw_info.ver);
+
+	/* MES */
+	query_fw.fw_type = AMDGPU_INFO_FW_MES;
+	ret = amdgpu_firmware_info(&fw_info, &query_fw, adev);
+	if (ret)
+		return ret;
+	seq_printf(m, "MES feature version: %u, firmware version: 0x%08x\n",
+		   fw_info.feature, fw_info.ver);
+
+	/* VPE */
+	query_fw.fw_type = AMDGPU_INFO_FW_VPE;
+	ret = amdgpu_firmware_info(&fw_info, &query_fw, adev);
+	if (ret)
+		return ret;
+	seq_printf(m, "VPE feature version: %u, firmware version: 0x%08x\n",
+		   fw_info.feature, fw_info.ver);
+
+	seq_printf(m, "VBIOS version: %s\n", ctx->vbios_pn);
 
 	return 0;
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_lsdma.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_lsdma.c
new file mode 100644
index 000000000000..4d1d4994ea3f
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_lsdma.c
@@ -0,0 +1,91 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "amdgpu.h"
+#include "amdgpu_lsdma.h"
+
+#define AMDGPU_LSDMA_MAX_SIZE	0x2000000ULL
+
+int amdgpu_lsdma_wait_for(struct amdgpu_device *adev,
+			  uint32_t reg_index, uint32_t reg_val,
+			  uint32_t mask)
+{
+	uint32_t val;
+	int i;
+
+	for (i = 0; i < adev->usec_timeout; i++) {
+		val = RREG32(reg_index);
+		if ((val & mask) == reg_val)
+			return 0;
+		udelay(1);
+	}
+
+	return -ETIME;
+}
+
+int amdgpu_lsdma_copy_mem(struct amdgpu_device *adev,
+			  uint64_t src_addr,
+			  uint64_t dst_addr,
+			  uint64_t mem_size)
+{
+	int ret;
+
+	if (mem_size == 0)
+		return -EINVAL;
+
+	while (mem_size > 0) {
+		uint64_t current_copy_size = min(mem_size, AMDGPU_LSDMA_MAX_SIZE);
+
+		ret = adev->lsdma.funcs->copy_mem(adev, src_addr, dst_addr, current_copy_size);
+		if (ret)
+			return ret;
+		src_addr += current_copy_size;
+		dst_addr += current_copy_size;
+		mem_size -= current_copy_size;
+	}
+
+	return 0;
+}
+
+int amdgpu_lsdma_fill_mem(struct amdgpu_device *adev,
+			  uint64_t dst_addr,
+			  uint32_t data,
+			  uint64_t mem_size)
+{
+	int ret;
+
+	if (mem_size == 0)
+		return -EINVAL;
+
+	while (mem_size > 0) {
+		uint64_t current_fill_size = min(mem_size, AMDGPU_LSDMA_MAX_SIZE);
+
+		ret = adev->lsdma.funcs->fill_mem(adev, dst_addr, data, current_fill_size);
+		if (ret)
+			return ret;
+		dst_addr += current_fill_size;
+		mem_size -= current_fill_size;
+	}
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_lsdma.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_lsdma.h
new file mode 100644
index 000000000000..c61ba58c5ee0
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_lsdma.h
@@ -0,0 +1,46 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __AMDGPU_LSDMA_H__
+#define __AMDGPU_LSDMA_H__
+
+struct amdgpu_lsdma {
+	const struct amdgpu_lsdma_funcs      *funcs;
+};
+
+struct amdgpu_lsdma_funcs {
+	int (*copy_mem)(struct amdgpu_device *adev, uint64_t src_addr,
+			uint64_t dst_addr, uint64_t size);
+	int (*fill_mem)(struct amdgpu_device *adev, uint64_t dst_addr,
+			uint32_t data, uint64_t size);
+	void (*update_memory_power_gating)(struct amdgpu_device *adev, bool enable);
+};
+
+int amdgpu_lsdma_copy_mem(struct amdgpu_device *adev, uint64_t src_addr,
+			  uint64_t dst_addr, uint64_t mem_size);
+int amdgpu_lsdma_fill_mem(struct amdgpu_device *adev, uint64_t dst_addr,
+			  uint32_t data, uint64_t mem_size);
+int amdgpu_lsdma_wait_for(struct amdgpu_device *adev, uint32_t reg_index,
+			  uint32_t reg_val, uint32_t mask);
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mca.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_mca.c
index ce538f4819f9..3ca03b5e0f91 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mca.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mca.c
@@ -27,6 +27,16 @@
 #include "umc/umc_6_7_0_offset.h"
 #include "umc/umc_6_7_0_sh_mask.h"
 
+static bool amdgpu_mca_is_deferred_error(struct amdgpu_device *adev,
+					uint64_t mc_status)
+{
+	if (adev->umc.ras->check_ecc_err_status)
+		return adev->umc.ras->check_ecc_err_status(adev,
+				AMDGPU_MCA_ERROR_TYPE_DE, &mc_status);
+
+	return false;
+}
+
 void amdgpu_mca_query_correctable_error_count(struct amdgpu_device *adev,
 					      uint64_t mc_status_addr,
 					      unsigned long *error_count)
@@ -71,47 +81,550 @@ void amdgpu_mca_query_ras_error_count(struct amdgpu_device *adev,
 	amdgpu_mca_reset_error_count(adev, mc_status_addr);
 }
 
-int amdgpu_mca_ras_late_init(struct amdgpu_device *adev,
-			     struct amdgpu_mca_ras *mca_dev)
+int amdgpu_mca_mp0_ras_sw_init(struct amdgpu_device *adev)
 {
-	int r;
-	struct ras_ih_if ih_info = {
-		.cb = NULL,
-	};
-	struct ras_fs_if fs_info = {
-		.sysfs_name = mca_dev->ras_funcs->sysfs_name,
-	};
+	int err;
+	struct amdgpu_mca_ras_block *ras;
+
+	if (!adev->mca.mp0.ras)
+		return 0;
+
+	ras = adev->mca.mp0.ras;
+
+	err = amdgpu_ras_register_ras_block(adev, &ras->ras_block);
+	if (err) {
+		dev_err(adev->dev, "Failed to register mca.mp0 ras block!\n");
+		return err;
+	}
+
+	strcpy(ras->ras_block.ras_comm.name, "mca.mp0");
+	ras->ras_block.ras_comm.block = AMDGPU_RAS_BLOCK__MCA;
+	ras->ras_block.ras_comm.type = AMDGPU_RAS_ERROR__MULTI_UNCORRECTABLE;
+	adev->mca.mp0.ras_if = &ras->ras_block.ras_comm;
+
+	return 0;
+}
+
+int amdgpu_mca_mp1_ras_sw_init(struct amdgpu_device *adev)
+{
+	int err;
+	struct amdgpu_mca_ras_block *ras;
+
+	if (!adev->mca.mp1.ras)
+		return 0;
+
+	ras = adev->mca.mp1.ras;
+
+	err = amdgpu_ras_register_ras_block(adev, &ras->ras_block);
+	if (err) {
+		dev_err(adev->dev, "Failed to register mca.mp1 ras block!\n");
+		return err;
+	}
+
+	strcpy(ras->ras_block.ras_comm.name, "mca.mp1");
+	ras->ras_block.ras_comm.block = AMDGPU_RAS_BLOCK__MCA;
+	ras->ras_block.ras_comm.type = AMDGPU_RAS_ERROR__MULTI_UNCORRECTABLE;
+	adev->mca.mp1.ras_if = &ras->ras_block.ras_comm;
+
+	return 0;
+}
+
+int amdgpu_mca_mpio_ras_sw_init(struct amdgpu_device *adev)
+{
+	int err;
+	struct amdgpu_mca_ras_block *ras;
+
+	if (!adev->mca.mpio.ras)
+		return 0;
+
+	ras = adev->mca.mpio.ras;
+
+	err = amdgpu_ras_register_ras_block(adev, &ras->ras_block);
+	if (err) {
+		dev_err(adev->dev, "Failed to register mca.mpio ras block!\n");
+		return err;
+	}
+
+	strcpy(ras->ras_block.ras_comm.name, "mca.mpio");
+	ras->ras_block.ras_comm.block = AMDGPU_RAS_BLOCK__MCA;
+	ras->ras_block.ras_comm.type = AMDGPU_RAS_ERROR__MULTI_UNCORRECTABLE;
+	adev->mca.mpio.ras_if = &ras->ras_block.ras_comm;
+
+	return 0;
+}
+
+static void amdgpu_mca_bank_set_init(struct mca_bank_set *mca_set)
+{
+	if (!mca_set)
+		return;
+
+	memset(mca_set, 0, sizeof(*mca_set));
+	INIT_LIST_HEAD(&mca_set->list);
+}
+
+static int amdgpu_mca_bank_set_add_entry(struct mca_bank_set *mca_set, struct mca_bank_entry *entry)
+{
+	struct mca_bank_node *node;
+
+	if (!entry)
+		return -EINVAL;
+
+	node = kvzalloc(sizeof(*node), GFP_KERNEL);
+	if (!node)
+		return -ENOMEM;
+
+	memcpy(&node->entry, entry, sizeof(*entry));
+
+	INIT_LIST_HEAD(&node->node);
+	list_add_tail(&node->node, &mca_set->list);
+
+	mca_set->nr_entries++;
+
+	return 0;
+}
+
+static int amdgpu_mca_bank_set_merge(struct mca_bank_set *mca_set, struct mca_bank_set *new)
+{
+	struct mca_bank_node *node;
+
+	list_for_each_entry(node, &new->list, node)
+		amdgpu_mca_bank_set_add_entry(mca_set, &node->entry);
+
+	return 0;
+}
+
+static void amdgpu_mca_bank_set_remove_node(struct mca_bank_set *mca_set, struct mca_bank_node *node)
+{
+	if (!node)
+		return;
+
+	list_del(&node->node);
+	kvfree(node);
+
+	mca_set->nr_entries--;
+}
+
+static void amdgpu_mca_bank_set_release(struct mca_bank_set *mca_set)
+{
+	struct mca_bank_node *node, *tmp;
+
+	if (list_empty(&mca_set->list))
+		return;
+
+	list_for_each_entry_safe(node, tmp, &mca_set->list, node)
+		amdgpu_mca_bank_set_remove_node(mca_set, node);
+}
+
+void amdgpu_mca_smu_init_funcs(struct amdgpu_device *adev, const struct amdgpu_mca_smu_funcs *mca_funcs)
+{
+	struct amdgpu_mca *mca = &adev->mca;
+
+	mca->mca_funcs = mca_funcs;
+}
+
+int amdgpu_mca_init(struct amdgpu_device *adev)
+{
+	struct amdgpu_mca *mca = &adev->mca;
+	struct mca_bank_cache *mca_cache;
+	int i;
+
+	atomic_set(&mca->ue_update_flag, 0);
+
+	for (i = 0; i < ARRAY_SIZE(mca->mca_caches); i++) {
+		mca_cache = &mca->mca_caches[i];
+		mutex_init(&mca_cache->lock);
+		amdgpu_mca_bank_set_init(&mca_cache->mca_set);
+	}
+
+	return 0;
+}
+
+void amdgpu_mca_fini(struct amdgpu_device *adev)
+{
+	struct amdgpu_mca *mca = &adev->mca;
+	struct mca_bank_cache *mca_cache;
+	int i;
+
+	atomic_set(&mca->ue_update_flag, 0);
+
+	for (i = 0; i < ARRAY_SIZE(mca->mca_caches); i++) {
+		mca_cache = &mca->mca_caches[i];
+		amdgpu_mca_bank_set_release(&mca_cache->mca_set);
+		mutex_destroy(&mca_cache->lock);
+	}
+}
+
+int amdgpu_mca_reset(struct amdgpu_device *adev)
+{
+	amdgpu_mca_fini(adev);
+
+	return amdgpu_mca_init(adev);
+}
+
+int amdgpu_mca_smu_set_debug_mode(struct amdgpu_device *adev, bool enable)
+{
+	const struct amdgpu_mca_smu_funcs *mca_funcs = adev->mca.mca_funcs;
+
+	if (mca_funcs && mca_funcs->mca_set_debug_mode)
+		return mca_funcs->mca_set_debug_mode(adev, enable);
+
+	return -EOPNOTSUPP;
+}
+
+static void amdgpu_mca_smu_mca_bank_dump(struct amdgpu_device *adev, int idx, struct mca_bank_entry *entry,
+					 struct ras_query_context *qctx)
+{
+	u64 event_id = qctx ? qctx->evid.event_id : RAS_EVENT_INVALID_ID;
+
+	RAS_EVENT_LOG(adev, event_id, HW_ERR "Accelerator Check Architecture events logged\n");
+	RAS_EVENT_LOG(adev, event_id, HW_ERR "aca entry[%02d].STATUS=0x%016llx\n",
+		      idx, entry->regs[MCA_REG_IDX_STATUS]);
+	RAS_EVENT_LOG(adev, event_id, HW_ERR "aca entry[%02d].ADDR=0x%016llx\n",
+		      idx, entry->regs[MCA_REG_IDX_ADDR]);
+	RAS_EVENT_LOG(adev, event_id, HW_ERR "aca entry[%02d].MISC0=0x%016llx\n",
+		      idx, entry->regs[MCA_REG_IDX_MISC0]);
+	RAS_EVENT_LOG(adev, event_id, HW_ERR "aca entry[%02d].IPID=0x%016llx\n",
+		      idx, entry->regs[MCA_REG_IDX_IPID]);
+	RAS_EVENT_LOG(adev, event_id, HW_ERR "aca entry[%02d].SYND=0x%016llx\n",
+		      idx, entry->regs[MCA_REG_IDX_SYND]);
+}
+
+static int amdgpu_mca_smu_get_valid_mca_count(struct amdgpu_device *adev, enum amdgpu_mca_error_type type, uint32_t *count)
+{
+	const struct amdgpu_mca_smu_funcs *mca_funcs = adev->mca.mca_funcs;
+
+	if (!count)
+		return -EINVAL;
+
+	if (mca_funcs && mca_funcs->mca_get_valid_mca_count)
+		return mca_funcs->mca_get_valid_mca_count(adev, type, count);
+
+	return -EOPNOTSUPP;
+}
+
+static int amdgpu_mca_smu_get_mca_entry(struct amdgpu_device *adev, enum amdgpu_mca_error_type type,
+					int idx, struct mca_bank_entry *entry)
+{
+	const struct amdgpu_mca_smu_funcs *mca_funcs = adev->mca.mca_funcs;
+	int count;
+
+	if (!mca_funcs || !mca_funcs->mca_get_mca_entry)
+		return -EOPNOTSUPP;
+
+	switch (type) {
+	case AMDGPU_MCA_ERROR_TYPE_UE:
+		count = mca_funcs->max_ue_count;
+		break;
+	case AMDGPU_MCA_ERROR_TYPE_CE:
+		count = mca_funcs->max_ce_count;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	if (idx >= count)
+		return -EINVAL;
+
+	return mca_funcs->mca_get_mca_entry(adev, type, idx, entry);
+}
+
+static bool amdgpu_mca_bank_should_update(struct amdgpu_device *adev, enum amdgpu_mca_error_type type)
+{
+	struct amdgpu_mca *mca = &adev->mca;
+	bool ret = true;
+
+	/*
+	 * Because the UE Valid MCA count will only be cleared after reset,
+	 * in order to avoid repeated counting of the error count,
+	 * the aca bank is only updated once during the gpu recovery stage.
+	 */
+	if (type == AMDGPU_MCA_ERROR_TYPE_UE) {
+		if (amdgpu_ras_intr_triggered())
+			ret = atomic_cmpxchg(&mca->ue_update_flag, 0, 1) == 0;
+		else
+			atomic_set(&mca->ue_update_flag, 0);
+	}
+
+	return ret;
+}
+
+static bool amdgpu_mca_bank_should_dump(struct amdgpu_device *adev, enum amdgpu_mca_error_type type,
+					struct mca_bank_entry *entry)
+{
+	bool ret;
+
+	switch (type) {
+	case AMDGPU_MCA_ERROR_TYPE_CE:
+		ret = amdgpu_mca_is_deferred_error(adev, entry->regs[MCA_REG_IDX_STATUS]);
+		break;
+	case AMDGPU_MCA_ERROR_TYPE_UE:
+	default:
+		ret = true;
+		break;
+	}
+
+	return ret;
+}
+
+static int amdgpu_mca_smu_get_mca_set(struct amdgpu_device *adev, enum amdgpu_mca_error_type type, struct mca_bank_set *mca_set,
+				      struct ras_query_context *qctx)
+{
+	struct mca_bank_entry entry;
+	uint32_t count = 0, i;
+	int ret;
+
+	if (!mca_set)
+		return -EINVAL;
+
+	if (!amdgpu_mca_bank_should_update(adev, type))
+		return 0;
+
+	ret = amdgpu_mca_smu_get_valid_mca_count(adev, type, &count);
+	if (ret)
+		return ret;
+
+	for (i = 0; i < count; i++) {
+		memset(&entry, 0, sizeof(entry));
+		ret = amdgpu_mca_smu_get_mca_entry(adev, type, i, &entry);
+		if (ret)
+			return ret;
 
-	if (!mca_dev->ras_if) {
-		mca_dev->ras_if = kmalloc(sizeof(struct ras_common_if), GFP_KERNEL);
-		if (!mca_dev->ras_if)
-			return -ENOMEM;
-		mca_dev->ras_if->block = mca_dev->ras_funcs->ras_block;
-		mca_dev->ras_if->sub_block_index = mca_dev->ras_funcs->ras_sub_block;
-		mca_dev->ras_if->type = AMDGPU_RAS_ERROR__MULTI_UNCORRECTABLE;
+		amdgpu_mca_bank_set_add_entry(mca_set, &entry);
+
+		if (amdgpu_mca_bank_should_dump(adev, type, &entry))
+			amdgpu_mca_smu_mca_bank_dump(adev, i, &entry, qctx);
 	}
-	ih_info.head = fs_info.head = *mca_dev->ras_if;
-	r = amdgpu_ras_late_init(adev, mca_dev->ras_if,
-				 &fs_info, &ih_info);
-	if (r || !amdgpu_ras_is_supported(adev, mca_dev->ras_if->block)) {
-		kfree(mca_dev->ras_if);
-		mca_dev->ras_if = NULL;
+
+	return 0;
+}
+
+static int amdgpu_mca_smu_parse_mca_error_count(struct amdgpu_device *adev, enum amdgpu_ras_block blk,
+						enum amdgpu_mca_error_type type, struct mca_bank_entry *entry, uint32_t *count)
+{
+	const struct amdgpu_mca_smu_funcs *mca_funcs = adev->mca.mca_funcs;
+
+	if (!count || !entry)
+		return -EINVAL;
+
+	if (!mca_funcs || !mca_funcs->mca_parse_mca_error_count)
+		return -EOPNOTSUPP;
+
+	return mca_funcs->mca_parse_mca_error_count(adev, blk, type, entry, count);
+}
+
+static int amdgpu_mca_dispatch_mca_set(struct amdgpu_device *adev, enum amdgpu_ras_block blk, enum amdgpu_mca_error_type type,
+				       struct mca_bank_set *mca_set, struct ras_err_data *err_data)
+{
+	struct amdgpu_smuio_mcm_config_info mcm_info;
+	struct mca_bank_node *node, *tmp;
+	struct mca_bank_entry *entry;
+	uint32_t count;
+	int ret;
+
+	if (!mca_set)
+		return -EINVAL;
+
+	if (!mca_set->nr_entries)
+		return 0;
+
+	list_for_each_entry_safe(node, tmp, &mca_set->list, node) {
+		entry = &node->entry;
+
+		count = 0;
+		ret = amdgpu_mca_smu_parse_mca_error_count(adev, blk, type, entry, &count);
+		if (ret && ret != -EOPNOTSUPP)
+			return ret;
+
+		if (!count)
+			continue;
+
+		memset(&mcm_info, 0, sizeof(mcm_info));
+
+		mcm_info.socket_id = entry->info.socket_id;
+		mcm_info.die_id = entry->info.aid;
+
+		if (type == AMDGPU_MCA_ERROR_TYPE_UE) {
+			amdgpu_ras_error_statistic_ue_count(err_data,
+							    &mcm_info, (uint64_t)count);
+		} else {
+			if (amdgpu_mca_is_deferred_error(adev, entry->regs[MCA_REG_IDX_STATUS]))
+				amdgpu_ras_error_statistic_de_count(err_data,
+								    &mcm_info, (uint64_t)count);
+			else
+				amdgpu_ras_error_statistic_ce_count(err_data,
+								    &mcm_info, (uint64_t)count);
+		}
+
+		amdgpu_mca_bank_set_remove_node(mca_set, node);
 	}
 
-	return r;
+	return 0;
+}
+
+static int amdgpu_mca_add_mca_set_to_cache(struct amdgpu_device *adev, enum amdgpu_mca_error_type type, struct mca_bank_set *new)
+{
+	struct mca_bank_cache *mca_cache = &adev->mca.mca_caches[type];
+	int ret;
+
+	mutex_lock(&mca_cache->lock);
+	ret = amdgpu_mca_bank_set_merge(&mca_cache->mca_set, new);
+	mutex_unlock(&mca_cache->lock);
+
+	return ret;
 }
 
-void amdgpu_mca_ras_fini(struct amdgpu_device *adev,
-			 struct amdgpu_mca_ras *mca_dev)
+int amdgpu_mca_smu_log_ras_error(struct amdgpu_device *adev, enum amdgpu_ras_block blk, enum amdgpu_mca_error_type type,
+				 struct ras_err_data *err_data, struct ras_query_context *qctx)
 {
-	struct ras_ih_if ih_info = {
-		.cb = NULL,
+	struct mca_bank_set mca_set;
+	struct mca_bank_cache *mca_cache = &adev->mca.mca_caches[type];
+	int ret;
+
+	amdgpu_mca_bank_set_init(&mca_set);
+
+	ret = amdgpu_mca_smu_get_mca_set(adev, type, &mca_set, qctx);
+	if (ret)
+		goto out_mca_release;
+
+	ret = amdgpu_mca_dispatch_mca_set(adev, blk, type, &mca_set, err_data);
+	if (ret)
+		goto out_mca_release;
+
+	/* add remain mca bank to mca cache */
+	if (mca_set.nr_entries) {
+		ret = amdgpu_mca_add_mca_set_to_cache(adev, type, &mca_set);
+		if (ret)
+			goto out_mca_release;
+	}
+
+	/* dispatch mca set again if mca cache has valid data */
+	mutex_lock(&mca_cache->lock);
+	if (mca_cache->mca_set.nr_entries)
+		ret = amdgpu_mca_dispatch_mca_set(adev, blk, type, &mca_cache->mca_set, err_data);
+	mutex_unlock(&mca_cache->lock);
+
+out_mca_release:
+	amdgpu_mca_bank_set_release(&mca_set);
+
+	return ret;
+}
+
+#if defined(CONFIG_DEBUG_FS)
+static int amdgpu_mca_smu_debug_mode_set(void *data, u64 val)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)data;
+	int ret;
+
+	ret = amdgpu_ras_set_mca_debug_mode(adev, val ? true : false);
+	if (ret)
+		return ret;
+
+	dev_info(adev->dev, "amdgpu set smu mca debug mode %s success\n", val ? "on" : "off");
+
+	return 0;
+}
+
+static void mca_dump_entry(struct seq_file *m, struct mca_bank_entry *entry)
+{
+	int i, idx = entry->idx;
+	int reg_idx_array[] = {
+		MCA_REG_IDX_STATUS,
+		MCA_REG_IDX_ADDR,
+		MCA_REG_IDX_MISC0,
+		MCA_REG_IDX_IPID,
+		MCA_REG_IDX_SYND,
 	};
 
-	if (!mca_dev->ras_if)
+	seq_printf(m, "mca entry[%d].type: %s\n", idx, entry->type == AMDGPU_MCA_ERROR_TYPE_UE ? "UE" : "CE");
+	seq_printf(m, "mca entry[%d].ip: %d\n", idx, entry->ip);
+	seq_printf(m, "mca entry[%d].info: socketid:%d aid:%d hwid:0x%03x mcatype:0x%04x\n",
+		   idx, entry->info.socket_id, entry->info.aid, entry->info.hwid, entry->info.mcatype);
+
+	for (i = 0; i < ARRAY_SIZE(reg_idx_array); i++)
+		seq_printf(m, "mca entry[%d].regs[%d]: 0x%016llx\n", idx, reg_idx_array[i], entry->regs[reg_idx_array[i]]);
+}
+
+static int mca_dump_show(struct seq_file *m, enum amdgpu_mca_error_type type)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)m->private;
+	struct mca_bank_node *node;
+	struct mca_bank_set mca_set;
+	struct ras_query_context qctx;
+	int ret;
+
+	amdgpu_mca_bank_set_init(&mca_set);
+
+	qctx.evid.event_id = RAS_EVENT_INVALID_ID;
+	ret = amdgpu_mca_smu_get_mca_set(adev, type, &mca_set, &qctx);
+	if (ret)
+		goto err_free_mca_set;
+
+	seq_printf(m, "amdgpu smu %s valid mca count: %d\n",
+		   type == AMDGPU_MCA_ERROR_TYPE_UE ? "UE" : "CE", mca_set.nr_entries);
+
+	if (!mca_set.nr_entries)
+		goto err_free_mca_set;
+
+	list_for_each_entry(node, &mca_set.list, node)
+		mca_dump_entry(m, &node->entry);
+
+	/* add mca bank to mca bank cache */
+	ret = amdgpu_mca_add_mca_set_to_cache(adev, type, &mca_set);
+
+err_free_mca_set:
+	amdgpu_mca_bank_set_release(&mca_set);
+
+	return ret;
+}
+
+static int mca_dump_ce_show(struct seq_file *m, void *unused)
+{
+	return mca_dump_show(m, AMDGPU_MCA_ERROR_TYPE_CE);
+}
+
+static int mca_dump_ce_open(struct inode *inode, struct file *file)
+{
+	return single_open(file, mca_dump_ce_show, inode->i_private);
+}
+
+static const struct file_operations mca_ce_dump_debug_fops = {
+	.owner = THIS_MODULE,
+	.open = mca_dump_ce_open,
+	.read = seq_read,
+	.llseek = seq_lseek,
+	.release = single_release,
+};
+
+static int mca_dump_ue_show(struct seq_file *m, void *unused)
+{
+	return mca_dump_show(m, AMDGPU_MCA_ERROR_TYPE_UE);
+}
+
+static int mca_dump_ue_open(struct inode *inode, struct file *file)
+{
+	return single_open(file, mca_dump_ue_show, inode->i_private);
+}
+
+static const struct file_operations mca_ue_dump_debug_fops = {
+	.owner = THIS_MODULE,
+	.open = mca_dump_ue_open,
+	.read = seq_read,
+	.llseek = seq_lseek,
+	.release = single_release,
+};
+
+DEFINE_DEBUGFS_ATTRIBUTE(mca_debug_mode_fops, NULL, amdgpu_mca_smu_debug_mode_set, "%llu\n");
+#endif
+
+void amdgpu_mca_smu_debugfs_init(struct amdgpu_device *adev, struct dentry *root)
+{
+#if defined(CONFIG_DEBUG_FS)
+	if (!root)
 		return;
 
-	amdgpu_ras_late_fini(adev, mca_dev->ras_if, &ih_info);
-	kfree(mca_dev->ras_if);
-	mca_dev->ras_if = NULL;
-}
-\ No newline at end of file
+	debugfs_create_file("mca_debug_mode", 0200, root, adev, &mca_debug_mode_fops);
+	debugfs_create_file("mca_ue_dump", 0400, root, adev, &mca_ue_dump_debug_fops);
+	debugfs_create_file("mca_ce_dump", 0400, root, adev, &mca_ce_dump_debug_fops);
+#endif
+}
+
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mca.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_mca.h
index c74bc7177066..e80323ff90c1 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mca.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mca.h
@@ -21,32 +21,120 @@
 #ifndef __AMDGPU_MCA_H__
 #define __AMDGPU_MCA_H__
 
-struct amdgpu_mca_ras_funcs {
-	int (*ras_late_init)(struct amdgpu_device *adev);
-	void (*ras_fini)(struct amdgpu_device *adev);
-	void (*query_ras_error_count)(struct amdgpu_device *adev,
-				      void *ras_error_status);
-	void (*query_ras_error_address)(struct amdgpu_device *adev,
-					void *ras_error_status);
-	uint32_t ras_block;
-	uint32_t ras_sub_block;
-	const char* sysfs_name;
+#include "amdgpu_ras.h"
+
+#define MCA_MAX_REGS_COUNT	(16)
+
+#define MCA_REG_FIELD(x, h, l)			(((x) & GENMASK_ULL(h, l)) >> l)
+#define MCA_REG__STATUS__VAL(x)			MCA_REG_FIELD(x, 63, 63)
+#define MCA_REG__STATUS__OVERFLOW(x)		MCA_REG_FIELD(x, 62, 62)
+#define MCA_REG__STATUS__UC(x)			MCA_REG_FIELD(x, 61, 61)
+#define MCA_REG__STATUS__EN(x)			MCA_REG_FIELD(x, 60, 60)
+#define MCA_REG__STATUS__MISCV(x)		MCA_REG_FIELD(x, 59, 59)
+#define MCA_REG__STATUS__ADDRV(x)		MCA_REG_FIELD(x, 58, 58)
+#define MCA_REG__STATUS__PCC(x)			MCA_REG_FIELD(x, 57, 57)
+#define MCA_REG__STATUS__ERRCOREIDVAL(x)	MCA_REG_FIELD(x, 56, 56)
+#define MCA_REG__STATUS__TCC(x)			MCA_REG_FIELD(x, 55, 55)
+#define MCA_REG__STATUS__SYNDV(x)		MCA_REG_FIELD(x, 53, 53)
+#define MCA_REG__STATUS__CECC(x)		MCA_REG_FIELD(x, 46, 46)
+#define MCA_REG__STATUS__UECC(x)		MCA_REG_FIELD(x, 45, 45)
+#define MCA_REG__STATUS__DEFERRED(x)		MCA_REG_FIELD(x, 44, 44)
+#define MCA_REG__STATUS__POISON(x)		MCA_REG_FIELD(x, 43, 43)
+#define MCA_REG__STATUS__SCRUB(x)		MCA_REG_FIELD(x, 40, 40)
+#define MCA_REG__STATUS__ERRCOREID(x)		MCA_REG_FIELD(x, 37, 32)
+#define MCA_REG__STATUS__ADDRLSB(x)		MCA_REG_FIELD(x, 29, 24)
+#define MCA_REG__STATUS__ERRORCODEEXT(x)	MCA_REG_FIELD(x, 21, 16)
+#define MCA_REG__STATUS__ERRORCODE(x)		MCA_REG_FIELD(x, 15, 0)
+
+#define MCA_REG__MISC0__ERRCNT(x)		MCA_REG_FIELD(x, 43, 32)
+
+#define MCA_REG__SYND__ERRORINFORMATION(x)	MCA_REG_FIELD(x, 17, 0)
+
+enum amdgpu_mca_ip {
+	AMDGPU_MCA_IP_UNKNOW = -1,
+	AMDGPU_MCA_IP_PSP = 0,
+	AMDGPU_MCA_IP_SDMA,
+	AMDGPU_MCA_IP_GC,
+	AMDGPU_MCA_IP_SMU,
+	AMDGPU_MCA_IP_MP5,
+	AMDGPU_MCA_IP_UMC,
+	AMDGPU_MCA_IP_PCS_XGMI,
+	AMDGPU_MCA_IP_COUNT,
+};
+
+enum amdgpu_mca_error_type {
+	AMDGPU_MCA_ERROR_TYPE_UE = 0,
+	AMDGPU_MCA_ERROR_TYPE_CE,
+	AMDGPU_MCA_ERROR_TYPE_DE,
+};
+
+struct amdgpu_mca_ras_block {
+	struct amdgpu_ras_block_object ras_block;
 };
 
 struct amdgpu_mca_ras {
 	struct ras_common_if *ras_if;
-	const struct amdgpu_mca_ras_funcs *ras_funcs;
+	struct amdgpu_mca_ras_block *ras;
+};
+
+struct mca_bank_set {
+	int nr_entries;
+	struct list_head list;
 };
 
-struct amdgpu_mca_funcs {
-	void (*init)(struct amdgpu_device *adev);
+struct mca_bank_cache {
+	struct mca_bank_set mca_set;
+	struct mutex lock;
 };
 
 struct amdgpu_mca {
-	const struct amdgpu_mca_funcs *funcs;
 	struct amdgpu_mca_ras mp0;
 	struct amdgpu_mca_ras mp1;
 	struct amdgpu_mca_ras mpio;
+	const struct amdgpu_mca_smu_funcs *mca_funcs;
+	struct mca_bank_cache mca_caches[AMDGPU_MCA_ERROR_TYPE_DE];
+	atomic_t ue_update_flag;
+};
+
+enum mca_reg_idx {
+	MCA_REG_IDX_STATUS		= 1,
+	MCA_REG_IDX_ADDR		= 2,
+	MCA_REG_IDX_MISC0		= 3,
+	MCA_REG_IDX_IPID		= 5,
+	MCA_REG_IDX_SYND		= 6,
+	MCA_REG_IDX_COUNT		= 16,
+};
+
+struct mca_bank_info {
+	int socket_id;
+	int aid;
+	int hwid;
+	int mcatype;
+};
+
+struct mca_bank_entry {
+	int idx;
+	enum amdgpu_mca_error_type type;
+	enum amdgpu_mca_ip ip;
+	struct mca_bank_info info;
+	uint64_t regs[MCA_MAX_REGS_COUNT];
+};
+
+struct mca_bank_node {
+	struct mca_bank_entry entry;
+	struct list_head node;
+};
+
+struct amdgpu_mca_smu_funcs {
+	int max_ue_count;
+	int max_ce_count;
+	int (*mca_set_debug_mode)(struct amdgpu_device *adev, bool enable);
+	int (*mca_parse_mca_error_count)(struct amdgpu_device *adev, enum amdgpu_ras_block blk, enum amdgpu_mca_error_type type,
+					 struct mca_bank_entry *entry, uint32_t *count);
+	int (*mca_get_valid_mca_count)(struct amdgpu_device *adev, enum amdgpu_mca_error_type type,
+				       uint32_t *count);
+	int (*mca_get_mca_entry)(struct amdgpu_device *adev, enum amdgpu_mca_error_type type,
+				 int idx, struct mca_bank_entry *entry);
 };
 
 void amdgpu_mca_query_correctable_error_count(struct amdgpu_device *adev,
@@ -63,11 +151,19 @@ void amdgpu_mca_reset_error_count(struct amdgpu_device *adev,
 void amdgpu_mca_query_ras_error_count(struct amdgpu_device *adev,
 				      uint64_t mc_status_addr,
 				      void *ras_error_status);
+int amdgpu_mca_mp0_ras_sw_init(struct amdgpu_device *adev);
+int amdgpu_mca_mp1_ras_sw_init(struct amdgpu_device *adev);
+int amdgpu_mca_mpio_ras_sw_init(struct amdgpu_device *adev);
 
-int amdgpu_mca_ras_late_init(struct amdgpu_device *adev,
-			     struct amdgpu_mca_ras *mca_dev);
-
-void amdgpu_mca_ras_fini(struct amdgpu_device *adev,
-			 struct amdgpu_mca_ras *mca_dev);
+void amdgpu_mca_smu_init_funcs(struct amdgpu_device *adev, const struct amdgpu_mca_smu_funcs *mca_funcs);
+int amdgpu_mca_init(struct amdgpu_device *adev);
+void amdgpu_mca_fini(struct amdgpu_device *adev);
+int amdgpu_mca_reset(struct amdgpu_device *adev);
+int amdgpu_mca_smu_set_debug_mode(struct amdgpu_device *adev, bool enable);
+int amdgpu_mca_smu_get_mca_set_error_count(struct amdgpu_device *adev, enum amdgpu_ras_block blk,
+					   enum amdgpu_mca_error_type type, uint32_t *total);
+void amdgpu_mca_smu_debugfs_init(struct amdgpu_device *adev, struct dentry *root);
+int amdgpu_mca_smu_log_ras_error(struct amdgpu_device *adev, enum amdgpu_ras_block blk, enum amdgpu_mca_error_type type,
+				 struct ras_err_data *err_data, struct ras_query_context *qctx);
 
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c
new file mode 100644
index 000000000000..9c182ce501af
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c
@@ -0,0 +1,784 @@
+/*
+ * Copyright 2019 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/firmware.h>
+#include <drm/drm_exec.h>
+
+#include "amdgpu_mes.h"
+#include "amdgpu.h"
+#include "soc15_common.h"
+#include "amdgpu_mes_ctx.h"
+
+#define AMDGPU_MES_MAX_NUM_OF_QUEUES_PER_PROCESS 1024
+#define AMDGPU_ONE_DOORBELL_SIZE 8
+
+int amdgpu_mes_doorbell_process_slice(struct amdgpu_device *adev)
+{
+	return roundup(AMDGPU_ONE_DOORBELL_SIZE *
+		       AMDGPU_MES_MAX_NUM_OF_QUEUES_PER_PROCESS,
+		       PAGE_SIZE);
+}
+
+static int amdgpu_mes_doorbell_init(struct amdgpu_device *adev)
+{
+	int i;
+	struct amdgpu_mes *mes = &adev->mes;
+
+	/* Bitmap for dynamic allocation of kernel doorbells */
+	mes->doorbell_bitmap = bitmap_zalloc(PAGE_SIZE / sizeof(u32), GFP_KERNEL);
+	if (!mes->doorbell_bitmap) {
+		dev_err(adev->dev, "Failed to allocate MES doorbell bitmap\n");
+		return -ENOMEM;
+	}
+
+	mes->num_mes_dbs = PAGE_SIZE / AMDGPU_ONE_DOORBELL_SIZE;
+	for (i = 0; i < AMDGPU_MES_PRIORITY_NUM_LEVELS; i++) {
+		adev->mes.aggregated_doorbells[i] = mes->db_start_dw_offset + i * 2;
+		set_bit(i, mes->doorbell_bitmap);
+	}
+
+	return 0;
+}
+
+static int amdgpu_mes_event_log_init(struct amdgpu_device *adev)
+{
+	int r;
+
+	if (!amdgpu_mes_log_enable)
+		return 0;
+
+	r = amdgpu_bo_create_kernel(adev, adev->mes.event_log_size, PAGE_SIZE,
+				    AMDGPU_GEM_DOMAIN_VRAM,
+				    &adev->mes.event_log_gpu_obj,
+				    &adev->mes.event_log_gpu_addr,
+				    &adev->mes.event_log_cpu_addr);
+	if (r) {
+		dev_warn(adev->dev, "failed to create MES event log buffer (%d)", r);
+		return r;
+	}
+
+	memset(adev->mes.event_log_cpu_addr, 0, adev->mes.event_log_size);
+
+	return  0;
+
+}
+
+static void amdgpu_mes_doorbell_free(struct amdgpu_device *adev)
+{
+	bitmap_free(adev->mes.doorbell_bitmap);
+}
+
+int amdgpu_mes_init(struct amdgpu_device *adev)
+{
+	int i, r, num_pipes;
+
+	adev->mes.adev = adev;
+
+	idr_init(&adev->mes.pasid_idr);
+	idr_init(&adev->mes.gang_id_idr);
+	idr_init(&adev->mes.queue_id_idr);
+	ida_init(&adev->mes.doorbell_ida);
+	spin_lock_init(&adev->mes.queue_id_lock);
+	mutex_init(&adev->mes.mutex_hidden);
+
+	for (i = 0; i < AMDGPU_MAX_MES_PIPES; i++)
+		spin_lock_init(&adev->mes.ring_lock[i]);
+
+	adev->mes.total_max_queue = AMDGPU_FENCE_MES_QUEUE_ID_MASK;
+	adev->mes.vmid_mask_mmhub = 0xFF00;
+	adev->mes.vmid_mask_gfxhub = adev->gfx.disable_kq ? 0xFFFE : 0xFF00;
+
+	num_pipes = adev->gfx.me.num_pipe_per_me * adev->gfx.me.num_me;
+	if (num_pipes > AMDGPU_MES_MAX_GFX_PIPES)
+		dev_warn(adev->dev, "more gfx pipes than supported by MES! (%d vs %d)\n",
+			 num_pipes, AMDGPU_MES_MAX_GFX_PIPES);
+
+	for (i = 0; i < AMDGPU_MES_MAX_GFX_PIPES; i++) {
+		if (i >= num_pipes)
+			break;
+		if (amdgpu_ip_version(adev, GC_HWIP, 0) >=
+		    IP_VERSION(12, 0, 0))
+			/*
+			 * GFX V12 has only one GFX pipe, but 8 queues in it.
+			 * GFX pipe 0 queue 0 is being used by Kernel queue.
+			 * Set GFX pipe 0 queue 1-7 for MES scheduling
+			 * mask = 1111 1110b
+			 */
+			adev->mes.gfx_hqd_mask[i] = adev->gfx.disable_kq ? 0xFF : 0xFE;
+		else
+			/*
+			 * GFX pipe 0 queue 0 is being used by Kernel queue.
+			 * Set GFX pipe 0 queue 1 for MES scheduling
+			 * mask = 10b
+			 */
+			adev->mes.gfx_hqd_mask[i] = adev->gfx.disable_kq ? 0x3 : 0x2;
+	}
+
+	num_pipes = adev->gfx.mec.num_pipe_per_mec * adev->gfx.mec.num_mec;
+	if (num_pipes > AMDGPU_MES_MAX_COMPUTE_PIPES)
+		dev_warn(adev->dev, "more compute pipes than supported by MES! (%d vs %d)\n",
+			 num_pipes, AMDGPU_MES_MAX_COMPUTE_PIPES);
+
+	for (i = 0; i < AMDGPU_MES_MAX_COMPUTE_PIPES; i++) {
+		if (i >= num_pipes)
+			break;
+		adev->mes.compute_hqd_mask[i] = adev->gfx.disable_kq ? 0xF : 0xC;
+	}
+
+	num_pipes = adev->sdma.num_instances;
+	if (num_pipes > AMDGPU_MES_MAX_SDMA_PIPES)
+		dev_warn(adev->dev, "more SDMA pipes than supported by MES! (%d vs %d)\n",
+			 num_pipes, AMDGPU_MES_MAX_SDMA_PIPES);
+
+	for (i = 0; i < AMDGPU_MES_MAX_SDMA_PIPES; i++) {
+		if (i >= num_pipes)
+			break;
+		adev->mes.sdma_hqd_mask[i] = 0xfc;
+	}
+
+	for (i = 0; i < AMDGPU_MAX_MES_PIPES; i++) {
+		r = amdgpu_device_wb_get(adev, &adev->mes.sch_ctx_offs[i]);
+		if (r) {
+			dev_err(adev->dev,
+				"(%d) ring trail_fence_offs wb alloc failed\n",
+				r);
+			goto error;
+		}
+		adev->mes.sch_ctx_gpu_addr[i] =
+			adev->wb.gpu_addr + (adev->mes.sch_ctx_offs[i] * 4);
+		adev->mes.sch_ctx_ptr[i] =
+			(uint64_t *)&adev->wb.wb[adev->mes.sch_ctx_offs[i]];
+
+		r = amdgpu_device_wb_get(adev,
+				 &adev->mes.query_status_fence_offs[i]);
+		if (r) {
+			dev_err(adev->dev,
+			      "(%d) query_status_fence_offs wb alloc failed\n",
+			      r);
+			goto error;
+		}
+		adev->mes.query_status_fence_gpu_addr[i] = adev->wb.gpu_addr +
+			(adev->mes.query_status_fence_offs[i] * 4);
+		adev->mes.query_status_fence_ptr[i] =
+			(uint64_t *)&adev->wb.wb[adev->mes.query_status_fence_offs[i]];
+	}
+
+	r = amdgpu_mes_doorbell_init(adev);
+	if (r)
+		goto error;
+
+	r = amdgpu_mes_event_log_init(adev);
+	if (r)
+		goto error_doorbell;
+
+	if (adev->mes.hung_queue_db_array_size) {
+		r = amdgpu_bo_create_kernel(adev,
+					    adev->mes.hung_queue_db_array_size * sizeof(u32),
+					    PAGE_SIZE,
+					    AMDGPU_GEM_DOMAIN_GTT,
+					    &adev->mes.hung_queue_db_array_gpu_obj,
+					    &adev->mes.hung_queue_db_array_gpu_addr,
+					    &adev->mes.hung_queue_db_array_cpu_addr);
+		if (r) {
+			dev_warn(adev->dev, "failed to create MES hung db array buffer (%d)", r);
+			goto error_doorbell;
+		}
+	}
+
+	return 0;
+
+error_doorbell:
+	amdgpu_mes_doorbell_free(adev);
+error:
+	for (i = 0; i < AMDGPU_MAX_MES_PIPES; i++) {
+		if (adev->mes.sch_ctx_ptr[i])
+			amdgpu_device_wb_free(adev, adev->mes.sch_ctx_offs[i]);
+		if (adev->mes.query_status_fence_ptr[i])
+			amdgpu_device_wb_free(adev,
+				      adev->mes.query_status_fence_offs[i]);
+	}
+
+	idr_destroy(&adev->mes.pasid_idr);
+	idr_destroy(&adev->mes.gang_id_idr);
+	idr_destroy(&adev->mes.queue_id_idr);
+	ida_destroy(&adev->mes.doorbell_ida);
+	mutex_destroy(&adev->mes.mutex_hidden);
+	return r;
+}
+
+void amdgpu_mes_fini(struct amdgpu_device *adev)
+{
+	int i;
+
+	amdgpu_bo_free_kernel(&adev->mes.hung_queue_db_array_gpu_obj,
+			      &adev->mes.hung_queue_db_array_gpu_addr,
+			      &adev->mes.hung_queue_db_array_cpu_addr);
+
+	amdgpu_bo_free_kernel(&adev->mes.event_log_gpu_obj,
+			      &adev->mes.event_log_gpu_addr,
+			      &adev->mes.event_log_cpu_addr);
+
+	for (i = 0; i < AMDGPU_MAX_MES_PIPES; i++) {
+		if (adev->mes.sch_ctx_ptr[i])
+			amdgpu_device_wb_free(adev, adev->mes.sch_ctx_offs[i]);
+		if (adev->mes.query_status_fence_ptr[i])
+			amdgpu_device_wb_free(adev,
+				      adev->mes.query_status_fence_offs[i]);
+	}
+
+	amdgpu_mes_doorbell_free(adev);
+
+	idr_destroy(&adev->mes.pasid_idr);
+	idr_destroy(&adev->mes.gang_id_idr);
+	idr_destroy(&adev->mes.queue_id_idr);
+	ida_destroy(&adev->mes.doorbell_ida);
+	mutex_destroy(&adev->mes.mutex_hidden);
+}
+
+int amdgpu_mes_suspend(struct amdgpu_device *adev)
+{
+	struct mes_suspend_gang_input input;
+	int r;
+
+	if (!amdgpu_mes_suspend_resume_all_supported(adev))
+		return 0;
+
+	memset(&input, 0x0, sizeof(struct mes_suspend_gang_input));
+	input.suspend_all_gangs = 1;
+
+	/*
+	 * Avoid taking any other locks under MES lock to avoid circular
+	 * lock dependencies.
+	 */
+	amdgpu_mes_lock(&adev->mes);
+	r = adev->mes.funcs->suspend_gang(&adev->mes, &input);
+	amdgpu_mes_unlock(&adev->mes);
+	if (r)
+		dev_err(adev->dev, "failed to suspend all gangs");
+
+	return r;
+}
+
+int amdgpu_mes_resume(struct amdgpu_device *adev)
+{
+	struct mes_resume_gang_input input;
+	int r;
+
+	if (!amdgpu_mes_suspend_resume_all_supported(adev))
+		return 0;
+
+	memset(&input, 0x0, sizeof(struct mes_resume_gang_input));
+	input.resume_all_gangs = 1;
+
+	/*
+	 * Avoid taking any other locks under MES lock to avoid circular
+	 * lock dependencies.
+	 */
+	amdgpu_mes_lock(&adev->mes);
+	r = adev->mes.funcs->resume_gang(&adev->mes, &input);
+	amdgpu_mes_unlock(&adev->mes);
+	if (r)
+		dev_err(adev->dev, "failed to resume all gangs");
+
+	return r;
+}
+
+int amdgpu_mes_map_legacy_queue(struct amdgpu_device *adev,
+				struct amdgpu_ring *ring)
+{
+	struct mes_map_legacy_queue_input queue_input;
+	int r;
+
+	memset(&queue_input, 0, sizeof(queue_input));
+
+	queue_input.queue_type = ring->funcs->type;
+	queue_input.doorbell_offset = ring->doorbell_index;
+	queue_input.pipe_id = ring->pipe;
+	queue_input.queue_id = ring->queue;
+	queue_input.mqd_addr = amdgpu_bo_gpu_offset(ring->mqd_obj);
+	queue_input.wptr_addr = ring->wptr_gpu_addr;
+
+	amdgpu_mes_lock(&adev->mes);
+	r = adev->mes.funcs->map_legacy_queue(&adev->mes, &queue_input);
+	amdgpu_mes_unlock(&adev->mes);
+	if (r)
+		dev_err(adev->dev, "failed to map legacy queue\n");
+
+	return r;
+}
+
+int amdgpu_mes_unmap_legacy_queue(struct amdgpu_device *adev,
+				  struct amdgpu_ring *ring,
+				  enum amdgpu_unmap_queues_action action,
+				  u64 gpu_addr, u64 seq)
+{
+	struct mes_unmap_legacy_queue_input queue_input;
+	int r;
+
+	queue_input.action = action;
+	queue_input.queue_type = ring->funcs->type;
+	queue_input.doorbell_offset = ring->doorbell_index;
+	queue_input.pipe_id = ring->pipe;
+	queue_input.queue_id = ring->queue;
+	queue_input.trail_fence_addr = gpu_addr;
+	queue_input.trail_fence_data = seq;
+
+	amdgpu_mes_lock(&adev->mes);
+	r = adev->mes.funcs->unmap_legacy_queue(&adev->mes, &queue_input);
+	amdgpu_mes_unlock(&adev->mes);
+	if (r)
+		dev_err(adev->dev, "failed to unmap legacy queue\n");
+
+	return r;
+}
+
+int amdgpu_mes_reset_legacy_queue(struct amdgpu_device *adev,
+				  struct amdgpu_ring *ring,
+				  unsigned int vmid,
+				  bool use_mmio)
+{
+	struct mes_reset_queue_input queue_input;
+	int r;
+
+	memset(&queue_input, 0, sizeof(queue_input));
+
+	queue_input.queue_type = ring->funcs->type;
+	queue_input.doorbell_offset = ring->doorbell_index;
+	queue_input.me_id = ring->me;
+	queue_input.pipe_id = ring->pipe;
+	queue_input.queue_id = ring->queue;
+	queue_input.mqd_addr = ring->mqd_obj ? amdgpu_bo_gpu_offset(ring->mqd_obj) : 0;
+	queue_input.wptr_addr = ring->wptr_gpu_addr;
+	queue_input.vmid = vmid;
+	queue_input.use_mmio = use_mmio;
+	queue_input.is_kq = true;
+	if (ring->funcs->type == AMDGPU_RING_TYPE_GFX)
+		queue_input.legacy_gfx = true;
+
+	amdgpu_mes_lock(&adev->mes);
+	r = adev->mes.funcs->reset_hw_queue(&adev->mes, &queue_input);
+	amdgpu_mes_unlock(&adev->mes);
+	if (r)
+		dev_err(adev->dev, "failed to reset legacy queue\n");
+
+	return r;
+}
+
+int amdgpu_mes_get_hung_queue_db_array_size(struct amdgpu_device *adev)
+{
+	return adev->mes.hung_queue_db_array_size;
+}
+
+int amdgpu_mes_detect_and_reset_hung_queues(struct amdgpu_device *adev,
+					    int queue_type,
+					    bool detect_only,
+					    unsigned int *hung_db_num,
+					    u32 *hung_db_array)
+
+{
+	struct mes_detect_and_reset_queue_input input;
+	u32 *db_array = adev->mes.hung_queue_db_array_cpu_addr;
+	int r, i;
+
+	if (!hung_db_num || !hung_db_array)
+		return -EINVAL;
+
+	if ((queue_type != AMDGPU_RING_TYPE_GFX) &&
+	    (queue_type != AMDGPU_RING_TYPE_COMPUTE) &&
+	    (queue_type != AMDGPU_RING_TYPE_SDMA))
+		return -EINVAL;
+
+	/* Clear the doorbell array before detection */
+	memset(adev->mes.hung_queue_db_array_cpu_addr, AMDGPU_MES_INVALID_DB_OFFSET,
+		adev->mes.hung_queue_db_array_size * sizeof(u32));
+	input.queue_type = queue_type;
+	input.detect_only = detect_only;
+
+	r = adev->mes.funcs->detect_and_reset_hung_queues(&adev->mes,
+							  &input);
+	if (r) {
+		dev_err(adev->dev, "failed to detect and reset\n");
+	} else {
+		*hung_db_num = 0;
+		for (i = 0; i < adev->mes.hung_queue_hqd_info_offset; i++) {
+			if (db_array[i] != AMDGPU_MES_INVALID_DB_OFFSET) {
+				hung_db_array[i] = db_array[i];
+				*hung_db_num += 1;
+			}
+		}
+
+		/*
+		 * TODO: return HQD info for MES scheduled user compute queue reset cases
+		 * stored in hung_db_array hqd info offset to full array size
+		 */
+	}
+
+	return r;
+}
+
+uint32_t amdgpu_mes_rreg(struct amdgpu_device *adev, uint32_t reg)
+{
+	struct mes_misc_op_input op_input;
+	int r, val = 0;
+	uint32_t addr_offset = 0;
+	uint64_t read_val_gpu_addr;
+	uint32_t *read_val_ptr;
+
+	if (amdgpu_device_wb_get(adev, &addr_offset)) {
+		dev_err(adev->dev, "critical bug! too many mes readers\n");
+		goto error;
+	}
+	read_val_gpu_addr = adev->wb.gpu_addr + (addr_offset * 4);
+	read_val_ptr = (uint32_t *)&adev->wb.wb[addr_offset];
+	op_input.op = MES_MISC_OP_READ_REG;
+	op_input.read_reg.reg_offset = reg;
+	op_input.read_reg.buffer_addr = read_val_gpu_addr;
+
+	if (!adev->mes.funcs->misc_op) {
+		dev_err(adev->dev, "mes rreg is not supported!\n");
+		goto error;
+	}
+
+	amdgpu_mes_lock(&adev->mes);
+	r = adev->mes.funcs->misc_op(&adev->mes, &op_input);
+	amdgpu_mes_unlock(&adev->mes);
+	if (r)
+		dev_err(adev->dev, "failed to read reg (0x%x)\n", reg);
+	else
+		val = *(read_val_ptr);
+
+error:
+	if (addr_offset)
+		amdgpu_device_wb_free(adev, addr_offset);
+	return val;
+}
+
+int amdgpu_mes_wreg(struct amdgpu_device *adev,
+		    uint32_t reg, uint32_t val)
+{
+	struct mes_misc_op_input op_input;
+	int r;
+
+	op_input.op = MES_MISC_OP_WRITE_REG;
+	op_input.write_reg.reg_offset = reg;
+	op_input.write_reg.reg_value = val;
+
+	if (!adev->mes.funcs->misc_op) {
+		dev_err(adev->dev, "mes wreg is not supported!\n");
+		r = -EINVAL;
+		goto error;
+	}
+
+	amdgpu_mes_lock(&adev->mes);
+	r = adev->mes.funcs->misc_op(&adev->mes, &op_input);
+	amdgpu_mes_unlock(&adev->mes);
+	if (r)
+		dev_err(adev->dev, "failed to write reg (0x%x)\n", reg);
+
+error:
+	return r;
+}
+
+int amdgpu_mes_reg_write_reg_wait(struct amdgpu_device *adev,
+				  uint32_t reg0, uint32_t reg1,
+				  uint32_t ref, uint32_t mask)
+{
+	struct mes_misc_op_input op_input;
+	int r;
+
+	op_input.op = MES_MISC_OP_WRM_REG_WR_WAIT;
+	op_input.wrm_reg.reg0 = reg0;
+	op_input.wrm_reg.reg1 = reg1;
+	op_input.wrm_reg.ref = ref;
+	op_input.wrm_reg.mask = mask;
+
+	if (!adev->mes.funcs->misc_op) {
+		dev_err(adev->dev, "mes reg_write_reg_wait is not supported!\n");
+		r = -EINVAL;
+		goto error;
+	}
+
+	amdgpu_mes_lock(&adev->mes);
+	r = adev->mes.funcs->misc_op(&adev->mes, &op_input);
+	amdgpu_mes_unlock(&adev->mes);
+	if (r)
+		dev_err(adev->dev, "failed to reg_write_reg_wait\n");
+
+error:
+	return r;
+}
+
+int amdgpu_mes_hdp_flush(struct amdgpu_device *adev)
+{
+	uint32_t hdp_flush_req_offset, hdp_flush_done_offset, ref_and_mask;
+
+	hdp_flush_req_offset = adev->nbio.funcs->get_hdp_flush_req_offset(adev);
+	hdp_flush_done_offset = adev->nbio.funcs->get_hdp_flush_done_offset(adev);
+	ref_and_mask = adev->nbio.hdp_flush_reg->ref_and_mask_cp0;
+
+	return amdgpu_mes_reg_write_reg_wait(adev, hdp_flush_req_offset, hdp_flush_done_offset,
+					     ref_and_mask, ref_and_mask);
+}
+
+int amdgpu_mes_set_shader_debugger(struct amdgpu_device *adev,
+				uint64_t process_context_addr,
+				uint32_t spi_gdbg_per_vmid_cntl,
+				const uint32_t *tcp_watch_cntl,
+				uint32_t flags,
+				bool trap_en)
+{
+	struct mes_misc_op_input op_input = {0};
+	int r;
+
+	if (!adev->mes.funcs->misc_op) {
+		dev_err(adev->dev,
+			"mes set shader debugger is not supported!\n");
+		return -EINVAL;
+	}
+
+	op_input.op = MES_MISC_OP_SET_SHADER_DEBUGGER;
+	op_input.set_shader_debugger.process_context_addr = process_context_addr;
+	op_input.set_shader_debugger.flags.u32all = flags;
+
+	/* use amdgpu mes_flush_shader_debugger instead */
+	if (op_input.set_shader_debugger.flags.process_ctx_flush)
+		return -EINVAL;
+
+	op_input.set_shader_debugger.spi_gdbg_per_vmid_cntl = spi_gdbg_per_vmid_cntl;
+	memcpy(op_input.set_shader_debugger.tcp_watch_cntl, tcp_watch_cntl,
+			sizeof(op_input.set_shader_debugger.tcp_watch_cntl));
+
+	if (((adev->mes.sched_version & AMDGPU_MES_API_VERSION_MASK) >>
+			AMDGPU_MES_API_VERSION_SHIFT) >= 14)
+		op_input.set_shader_debugger.trap_en = trap_en;
+
+	amdgpu_mes_lock(&adev->mes);
+
+	r = adev->mes.funcs->misc_op(&adev->mes, &op_input);
+	if (r)
+		dev_err(adev->dev, "failed to set_shader_debugger\n");
+
+	amdgpu_mes_unlock(&adev->mes);
+
+	return r;
+}
+
+int amdgpu_mes_flush_shader_debugger(struct amdgpu_device *adev,
+				     uint64_t process_context_addr)
+{
+	struct mes_misc_op_input op_input = {0};
+	int r;
+
+	if (!adev->mes.funcs->misc_op) {
+		dev_err(adev->dev,
+			"mes flush shader debugger is not supported!\n");
+		return -EINVAL;
+	}
+
+	op_input.op = MES_MISC_OP_SET_SHADER_DEBUGGER;
+	op_input.set_shader_debugger.process_context_addr = process_context_addr;
+	op_input.set_shader_debugger.flags.process_ctx_flush = true;
+
+	amdgpu_mes_lock(&adev->mes);
+
+	r = adev->mes.funcs->misc_op(&adev->mes, &op_input);
+	if (r)
+		dev_err(adev->dev, "failed to set_shader_debugger\n");
+
+	amdgpu_mes_unlock(&adev->mes);
+
+	return r;
+}
+
+uint32_t amdgpu_mes_get_aggregated_doorbell_index(struct amdgpu_device *adev,
+						   enum amdgpu_mes_priority_level prio)
+{
+	return adev->mes.aggregated_doorbells[prio];
+}
+
+int amdgpu_mes_init_microcode(struct amdgpu_device *adev, int pipe)
+{
+	const struct mes_firmware_header_v1_0 *mes_hdr;
+	struct amdgpu_firmware_info *info;
+	char ucode_prefix[30];
+	char fw_name[50];
+	bool need_retry = false;
+	u32 *ucode_ptr;
+	int r;
+
+	amdgpu_ucode_ip_version_decode(adev, GC_HWIP, ucode_prefix,
+				       sizeof(ucode_prefix));
+	if (adev->enable_uni_mes) {
+		snprintf(fw_name, sizeof(fw_name),
+			 "amdgpu/%s_uni_mes.bin", ucode_prefix);
+	} else if (amdgpu_ip_version(adev, GC_HWIP, 0) >= IP_VERSION(11, 0, 0) &&
+	    amdgpu_ip_version(adev, GC_HWIP, 0) < IP_VERSION(12, 0, 0)) {
+		snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_mes%s.bin",
+			 ucode_prefix,
+			 pipe == AMDGPU_MES_SCHED_PIPE ? "_2" : "1");
+		need_retry = true;
+	} else {
+		snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_mes%s.bin",
+			 ucode_prefix,
+			 pipe == AMDGPU_MES_SCHED_PIPE ? "" : "1");
+	}
+
+	r = amdgpu_ucode_request(adev, &adev->mes.fw[pipe], AMDGPU_UCODE_REQUIRED,
+				 "%s", fw_name);
+	if (r && need_retry && pipe == AMDGPU_MES_SCHED_PIPE) {
+		dev_info(adev->dev, "try to fall back to %s_mes.bin\n", ucode_prefix);
+		r = amdgpu_ucode_request(adev, &adev->mes.fw[pipe],
+					 AMDGPU_UCODE_REQUIRED,
+					 "amdgpu/%s_mes.bin", ucode_prefix);
+	}
+
+	if (r)
+		goto out;
+
+	mes_hdr = (const struct mes_firmware_header_v1_0 *)
+		adev->mes.fw[pipe]->data;
+	adev->mes.uc_start_addr[pipe] =
+		le32_to_cpu(mes_hdr->mes_uc_start_addr_lo) |
+		((uint64_t)(le32_to_cpu(mes_hdr->mes_uc_start_addr_hi)) << 32);
+	adev->mes.data_start_addr[pipe] =
+		le32_to_cpu(mes_hdr->mes_data_start_addr_lo) |
+		((uint64_t)(le32_to_cpu(mes_hdr->mes_data_start_addr_hi)) << 32);
+	ucode_ptr = (u32 *)(adev->mes.fw[pipe]->data +
+			  sizeof(union amdgpu_firmware_header));
+	adev->mes.fw_version[pipe] =
+		le32_to_cpu(ucode_ptr[24]) & AMDGPU_MES_VERSION_MASK;
+
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
+		int ucode, ucode_data;
+
+		if (pipe == AMDGPU_MES_SCHED_PIPE) {
+			ucode = AMDGPU_UCODE_ID_CP_MES;
+			ucode_data = AMDGPU_UCODE_ID_CP_MES_DATA;
+		} else {
+			ucode = AMDGPU_UCODE_ID_CP_MES1;
+			ucode_data = AMDGPU_UCODE_ID_CP_MES1_DATA;
+		}
+
+		info = &adev->firmware.ucode[ucode];
+		info->ucode_id = ucode;
+		info->fw = adev->mes.fw[pipe];
+		adev->firmware.fw_size +=
+			ALIGN(le32_to_cpu(mes_hdr->mes_ucode_size_bytes),
+			      PAGE_SIZE);
+
+		info = &adev->firmware.ucode[ucode_data];
+		info->ucode_id = ucode_data;
+		info->fw = adev->mes.fw[pipe];
+		adev->firmware.fw_size +=
+			ALIGN(le32_to_cpu(mes_hdr->mes_ucode_data_size_bytes),
+			      PAGE_SIZE);
+	}
+
+	return 0;
+out:
+	amdgpu_ucode_release(&adev->mes.fw[pipe]);
+	return r;
+}
+
+bool amdgpu_mes_suspend_resume_all_supported(struct amdgpu_device *adev)
+{
+	uint32_t mes_rev = adev->mes.sched_version & AMDGPU_MES_VERSION_MASK;
+
+	return ((amdgpu_ip_version(adev, GC_HWIP, 0) >= IP_VERSION(11, 0, 0) &&
+		 amdgpu_ip_version(adev, GC_HWIP, 0) < IP_VERSION(12, 0, 0) &&
+		 mes_rev >= 0x63) ||
+		amdgpu_ip_version(adev, GC_HWIP, 0) >= IP_VERSION(12, 0, 0));
+}
+
+/* Fix me -- node_id is used to identify the correct MES instances in the future */
+static int amdgpu_mes_set_enforce_isolation(struct amdgpu_device *adev,
+					    uint32_t node_id, bool enable)
+{
+	struct mes_misc_op_input op_input = {0};
+	int r;
+
+	op_input.op = MES_MISC_OP_CHANGE_CONFIG;
+	op_input.change_config.option.limit_single_process = enable ? 1 : 0;
+
+	if (!adev->mes.funcs->misc_op) {
+		dev_err(adev->dev, "mes change config is not supported!\n");
+		r = -EINVAL;
+		goto error;
+	}
+
+	amdgpu_mes_lock(&adev->mes);
+	r = adev->mes.funcs->misc_op(&adev->mes, &op_input);
+	amdgpu_mes_unlock(&adev->mes);
+	if (r)
+		dev_err(adev->dev, "failed to change_config.\n");
+
+error:
+	return r;
+}
+
+int amdgpu_mes_update_enforce_isolation(struct amdgpu_device *adev)
+{
+	int i, r = 0;
+
+	if (adev->enable_mes && adev->gfx.enable_cleaner_shader) {
+		mutex_lock(&adev->enforce_isolation_mutex);
+		for (i = 0; i < (adev->xcp_mgr ? adev->xcp_mgr->num_xcps : 1); i++) {
+			if (adev->enforce_isolation[i] == AMDGPU_ENFORCE_ISOLATION_ENABLE)
+				r |= amdgpu_mes_set_enforce_isolation(adev, i, true);
+			else
+				r |= amdgpu_mes_set_enforce_isolation(adev, i, false);
+		}
+		mutex_unlock(&adev->enforce_isolation_mutex);
+	}
+	return r;
+}
+
+#if defined(CONFIG_DEBUG_FS)
+
+static int amdgpu_debugfs_mes_event_log_show(struct seq_file *m, void *unused)
+{
+	struct amdgpu_device *adev = m->private;
+	uint32_t *mem = (uint32_t *)(adev->mes.event_log_cpu_addr);
+
+	seq_hex_dump(m, "", DUMP_PREFIX_OFFSET, 32, 4,
+		     mem, adev->mes.event_log_size, false);
+
+	return 0;
+}
+
+DEFINE_SHOW_ATTRIBUTE(amdgpu_debugfs_mes_event_log);
+
+#endif
+
+void amdgpu_debugfs_mes_event_log_init(struct amdgpu_device *adev)
+{
+
+#if defined(CONFIG_DEBUG_FS)
+	struct drm_minor *minor = adev_to_drm(adev)->primary;
+	struct dentry *root = minor->debugfs_root;
+	if (adev->enable_mes && amdgpu_mes_log_enable)
+		debugfs_create_file("amdgpu_mes_event_log", 0444, root,
+				    adev, &amdgpu_debugfs_mes_event_log_fops);
+
+#endif
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.h
index 7334982ea702..e989225b354b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.h
@@ -24,10 +24,25 @@
 #ifndef __AMDGPU_MES_H__
 #define __AMDGPU_MES_H__
 
+#include "amdgpu_irq.h"
+#include "kgd_kfd_interface.h"
+#include "amdgpu_gfx.h"
+#include "amdgpu_doorbell.h"
+#include <linux/sched/mm.h>
+
 #define AMDGPU_MES_MAX_COMPUTE_PIPES        8
 #define AMDGPU_MES_MAX_GFX_PIPES            2
 #define AMDGPU_MES_MAX_SDMA_PIPES           2
 
+#define AMDGPU_MES_API_VERSION_SHIFT	12
+#define AMDGPU_MES_FEAT_VERSION_SHIFT	24
+
+#define AMDGPU_MES_VERSION_MASK		0x00000fff
+#define AMDGPU_MES_API_VERSION_MASK	0x00fff000
+#define AMDGPU_MES_FEAT_VERSION_MASK	0xff000000
+#define AMDGPU_MES_MSCRATCH_SIZE	0x40000
+#define AMDGPU_MES_INVALID_DB_OFFSET	0xffffffff
+
 enum amdgpu_mes_priority_level {
 	AMDGPU_MES_PRIORITY_LEVEL_LOW       = 0,
 	AMDGPU_MES_PRIORITY_LEVEL_NORMAL    = 1,
@@ -37,57 +52,159 @@ enum amdgpu_mes_priority_level {
 	AMDGPU_MES_PRIORITY_NUM_LEVELS
 };
 
+#define AMDGPU_MES_PROC_CTX_SIZE 0x1000 /* one page area */
+#define AMDGPU_MES_GANG_CTX_SIZE 0x1000 /* one page area */
+
 struct amdgpu_mes_funcs;
 
+enum amdgpu_mes_pipe {
+	AMDGPU_MES_SCHED_PIPE = 0,
+	AMDGPU_MES_KIQ_PIPE,
+	AMDGPU_MAX_MES_PIPES = 2,
+};
+
 struct amdgpu_mes {
 	struct amdgpu_device            *adev;
 
+	struct mutex                    mutex_hidden;
+
+	struct idr                      pasid_idr;
+	struct idr                      gang_id_idr;
+	struct idr                      queue_id_idr;
+	struct ida                      doorbell_ida;
+
+	spinlock_t                      queue_id_lock;
+
+	uint32_t			sched_version;
+	uint32_t			kiq_version;
+	uint32_t			fw_version[AMDGPU_MAX_MES_PIPES];
+	bool                            enable_legacy_queue_map;
+
 	uint32_t                        total_max_queue;
-	uint32_t                        doorbell_id_offset;
 	uint32_t                        max_doorbell_slices;
 
 	uint64_t                        default_process_quantum;
 	uint64_t                        default_gang_quantum;
 
-	struct amdgpu_ring              ring;
+	struct amdgpu_ring              ring[AMDGPU_MAX_MES_PIPES];
+	spinlock_t                      ring_lock[AMDGPU_MAX_MES_PIPES];
 
-	const struct firmware           *fw;
+	const struct firmware           *fw[AMDGPU_MAX_MES_PIPES];
 
 	/* mes ucode */
-	struct amdgpu_bo		*ucode_fw_obj;
-	uint64_t			ucode_fw_gpu_addr;
-	uint32_t			*ucode_fw_ptr;
-	uint32_t                        ucode_fw_version;
-	uint64_t                        uc_start_addr;
+	struct amdgpu_bo		*ucode_fw_obj[AMDGPU_MAX_MES_PIPES];
+	uint64_t			ucode_fw_gpu_addr[AMDGPU_MAX_MES_PIPES];
+	uint32_t			*ucode_fw_ptr[AMDGPU_MAX_MES_PIPES];
+	uint64_t                        uc_start_addr[AMDGPU_MAX_MES_PIPES];
 
 	/* mes ucode data */
-	struct amdgpu_bo		*data_fw_obj;
-	uint64_t			data_fw_gpu_addr;
-	uint32_t			*data_fw_ptr;
-	uint32_t                        data_fw_version;
-	uint64_t                        data_start_addr;
+	struct amdgpu_bo		*data_fw_obj[AMDGPU_MAX_MES_PIPES];
+	uint64_t			data_fw_gpu_addr[AMDGPU_MAX_MES_PIPES];
+	uint32_t			*data_fw_ptr[AMDGPU_MAX_MES_PIPES];
+	uint64_t                        data_start_addr[AMDGPU_MAX_MES_PIPES];
 
 	/* eop gpu obj */
-	struct amdgpu_bo		*eop_gpu_obj;
-	uint64_t                        eop_gpu_addr;
+	struct amdgpu_bo		*eop_gpu_obj[AMDGPU_MAX_MES_PIPES];
+	uint64_t                        eop_gpu_addr[AMDGPU_MAX_MES_PIPES];
 
-	void                            *mqd_backup;
+	void                            *mqd_backup[AMDGPU_MAX_MES_PIPES];
+	struct amdgpu_irq_src	        irq[AMDGPU_MAX_MES_PIPES];
 
 	uint32_t                        vmid_mask_gfxhub;
 	uint32_t                        vmid_mask_mmhub;
-	uint32_t                        compute_hqd_mask[AMDGPU_MES_MAX_COMPUTE_PIPES];
 	uint32_t                        gfx_hqd_mask[AMDGPU_MES_MAX_GFX_PIPES];
+	uint32_t                        compute_hqd_mask[AMDGPU_MES_MAX_COMPUTE_PIPES];
 	uint32_t                        sdma_hqd_mask[AMDGPU_MES_MAX_SDMA_PIPES];
-	uint32_t                        agreegated_doorbells[AMDGPU_MES_PRIORITY_NUM_LEVELS];
-	uint32_t                        sch_ctx_offs;
-	uint64_t			sch_ctx_gpu_addr;
-	uint64_t			*sch_ctx_ptr;
-	uint32_t			query_status_fence_offs;
-	uint64_t			query_status_fence_gpu_addr;
-	uint64_t			*query_status_fence_ptr;
+	uint32_t                        aggregated_doorbells[AMDGPU_MES_PRIORITY_NUM_LEVELS];
+	uint32_t                        sch_ctx_offs[AMDGPU_MAX_MES_PIPES];
+	uint64_t			sch_ctx_gpu_addr[AMDGPU_MAX_MES_PIPES];
+	uint64_t			*sch_ctx_ptr[AMDGPU_MAX_MES_PIPES];
+	uint32_t			query_status_fence_offs[AMDGPU_MAX_MES_PIPES];
+	uint64_t			query_status_fence_gpu_addr[AMDGPU_MAX_MES_PIPES];
+	uint64_t			*query_status_fence_ptr[AMDGPU_MAX_MES_PIPES];
+
+	uint32_t			saved_flags;
+
+	/* initialize kiq pipe */
+	int                             (*kiq_hw_init)(struct amdgpu_device *adev);
+	int                             (*kiq_hw_fini)(struct amdgpu_device *adev);
+
+	/* MES doorbells */
+	uint32_t			db_start_dw_offset;
+	uint32_t			num_mes_dbs;
+	unsigned long			*doorbell_bitmap;
+
+	/* MES event log buffer */
+	uint32_t			event_log_size;
+	struct amdgpu_bo	*event_log_gpu_obj;
+	uint64_t			event_log_gpu_addr;
+	void				*event_log_cpu_addr;
 
 	/* ip specific functions */
 	const struct amdgpu_mes_funcs   *funcs;
+
+	/* mes resource_1 bo*/
+	struct amdgpu_bo    *resource_1[AMDGPU_MAX_MES_PIPES];
+	uint64_t            resource_1_gpu_addr[AMDGPU_MAX_MES_PIPES];
+	void                *resource_1_addr[AMDGPU_MAX_MES_PIPES];
+
+	int				hung_queue_db_array_size;
+	int				hung_queue_hqd_info_offset;
+	struct amdgpu_bo		*hung_queue_db_array_gpu_obj;
+	uint64_t			hung_queue_db_array_gpu_addr;
+	void				*hung_queue_db_array_cpu_addr;
+};
+
+struct amdgpu_mes_gang {
+	int 				gang_id;
+	int 				priority;
+	int 				inprocess_gang_priority;
+	int 				global_priority_level;
+	struct list_head 		list;
+	struct amdgpu_mes_process 	*process;
+	struct amdgpu_bo 		*gang_ctx_bo;
+	uint64_t 			gang_ctx_gpu_addr;
+	void 				*gang_ctx_cpu_ptr;
+	uint64_t 			gang_quantum;
+	struct list_head 		queue_list;
+};
+
+struct amdgpu_mes_queue {
+	struct list_head 		list;
+	struct amdgpu_mes_gang 		*gang;
+	int 				queue_id;
+	uint64_t 			doorbell_off;
+	struct amdgpu_bo		*mqd_obj;
+	void				*mqd_cpu_ptr;
+	uint64_t 			mqd_gpu_addr;
+	uint64_t 			wptr_gpu_addr;
+	int 				queue_type;
+	int 				paging;
+	struct amdgpu_ring 		*ring;
+};
+
+struct amdgpu_mes_queue_properties {
+	int 			queue_type;
+	uint64_t                hqd_base_gpu_addr;
+	uint64_t                rptr_gpu_addr;
+	uint64_t                wptr_gpu_addr;
+	uint64_t                wptr_mc_addr;
+	uint32_t                queue_size;
+	uint64_t                eop_gpu_addr;
+	uint32_t                hqd_pipe_priority;
+	uint32_t                hqd_queue_priority;
+	bool 			paging;
+	struct amdgpu_ring 	*ring;
+	/* out */
+	uint64_t       		doorbell_off;
+};
+
+struct amdgpu_mes_gang_properties {
+	uint32_t 	priority;
+	uint32_t 	gang_quantum;
+	uint32_t 	inprocess_gang_priority;
+	uint32_t 	priority_level;
+	int 		global_priority_level;
 };
 
 struct mes_add_queue_input {
@@ -104,13 +221,44 @@ struct mes_add_queue_input {
 	uint32_t	doorbell_offset;
 	uint64_t	mqd_addr;
 	uint64_t	wptr_addr;
+	uint64_t	wptr_mc_addr;
 	uint32_t	queue_type;
 	uint32_t	paging;
+	uint32_t        gws_base;
+	uint32_t        gws_size;
+	uint64_t	tba_addr;
+	uint64_t	tma_addr;
+	uint32_t	trap_en;
+	uint32_t	skip_process_ctx_clear;
+	uint32_t	is_kfd_process;
+	uint32_t	is_aql_queue;
+	uint32_t	queue_size;
+	uint32_t	exclusively_scheduled;
 };
 
 struct mes_remove_queue_input {
 	uint32_t	doorbell_offset;
 	uint64_t	gang_context_addr;
+	bool		remove_queue_after_reset;
+};
+
+struct mes_map_legacy_queue_input {
+	uint32_t                           queue_type;
+	uint32_t                           doorbell_offset;
+	uint32_t                           pipe_id;
+	uint32_t                           queue_id;
+	uint64_t                           mqd_addr;
+	uint64_t                           wptr_addr;
+};
+
+struct mes_unmap_legacy_queue_input {
+	enum amdgpu_unmap_queues_action    action;
+	uint32_t                           queue_type;
+	uint32_t                           doorbell_offset;
+	uint32_t                           pipe_id;
+	uint32_t                           queue_id;
+	uint64_t                           trail_fence_addr;
+	uint64_t                           trail_fence_data;
 };
 
 struct mes_suspend_gang_input {
@@ -125,6 +273,95 @@ struct mes_resume_gang_input {
 	uint64_t	gang_context_addr;
 };
 
+struct mes_reset_queue_input {
+	uint32_t                           queue_type;
+	uint32_t                           doorbell_offset;
+	bool                               use_mmio;
+	uint32_t                           me_id;
+	uint32_t                           pipe_id;
+	uint32_t                           queue_id;
+	uint64_t                           mqd_addr;
+	uint64_t                           wptr_addr;
+	uint32_t                           vmid;
+	bool                               legacy_gfx;
+	bool                               is_kq;
+};
+
+struct mes_detect_and_reset_queue_input {
+	uint32_t                           queue_type;
+	bool                               detect_only;
+};
+
+struct mes_inv_tlbs_pasid_input {
+	uint32_t        xcc_id;
+	uint16_t        pasid;
+	uint8_t         hub_id;
+	uint8_t         flush_type;
+};
+
+enum mes_misc_opcode {
+	MES_MISC_OP_WRITE_REG,
+	MES_MISC_OP_READ_REG,
+	MES_MISC_OP_WRM_REG_WAIT,
+	MES_MISC_OP_WRM_REG_WR_WAIT,
+	MES_MISC_OP_SET_SHADER_DEBUGGER,
+	MES_MISC_OP_CHANGE_CONFIG,
+};
+
+struct mes_misc_op_input {
+	enum mes_misc_opcode op;
+
+	union {
+		struct {
+			uint32_t                  reg_offset;
+			uint64_t                  buffer_addr;
+		} read_reg;
+
+		struct {
+			uint32_t                  reg_offset;
+			uint32_t                  reg_value;
+		} write_reg;
+
+		struct {
+			uint32_t                   ref;
+			uint32_t                   mask;
+			uint32_t                   reg0;
+			uint32_t                   reg1;
+		} wrm_reg;
+
+		struct {
+			uint64_t process_context_addr;
+			union {
+				struct {
+					uint32_t single_memop : 1;
+					uint32_t single_alu_op : 1;
+					uint32_t reserved: 29;
+					uint32_t process_ctx_flush: 1;
+				};
+				uint32_t u32all;
+			} flags;
+			uint32_t spi_gdbg_per_vmid_cntl;
+			uint32_t tcp_watch_cntl[4];
+			uint32_t trap_en;
+		} set_shader_debugger;
+
+		struct {
+			union {
+				struct {
+					uint32_t limit_single_process : 1;
+					uint32_t enable_hws_logging_buffer : 1;
+					uint32_t reserved : 30;
+				};
+				uint32_t all;
+			} option;
+			struct {
+				uint32_t tdr_level;
+				uint32_t tdr_delay;
+			} tdr_config;
+		} change_config;
+	};
+};
+
 struct amdgpu_mes_funcs {
 	int (*add_hw_queue)(struct amdgpu_mes *mes,
 			    struct mes_add_queue_input *input);
@@ -132,11 +369,141 @@ struct amdgpu_mes_funcs {
 	int (*remove_hw_queue)(struct amdgpu_mes *mes,
 			       struct mes_remove_queue_input *input);
 
+	int (*map_legacy_queue)(struct amdgpu_mes *mes,
+				struct mes_map_legacy_queue_input *input);
+
+	int (*unmap_legacy_queue)(struct amdgpu_mes *mes,
+				  struct mes_unmap_legacy_queue_input *input);
+
 	int (*suspend_gang)(struct amdgpu_mes *mes,
 			    struct mes_suspend_gang_input *input);
 
 	int (*resume_gang)(struct amdgpu_mes *mes,
 			   struct mes_resume_gang_input *input);
+
+	int (*misc_op)(struct amdgpu_mes *mes,
+		       struct mes_misc_op_input *input);
+
+	int (*reset_hw_queue)(struct amdgpu_mes *mes,
+			      struct mes_reset_queue_input *input);
+
+	int (*detect_and_reset_hung_queues)(struct amdgpu_mes *mes,
+			      struct mes_detect_and_reset_queue_input *input);
+
+
+	int (*invalidate_tlbs_pasid)(struct amdgpu_mes *mes,
+			      struct mes_inv_tlbs_pasid_input *input);
 };
 
+#define amdgpu_mes_kiq_hw_init(adev) (adev)->mes.kiq_hw_init((adev))
+#define amdgpu_mes_kiq_hw_fini(adev) (adev)->mes.kiq_hw_fini((adev))
+
+int amdgpu_mes_init_microcode(struct amdgpu_device *adev, int pipe);
+int amdgpu_mes_init(struct amdgpu_device *adev);
+void amdgpu_mes_fini(struct amdgpu_device *adev);
+
+int amdgpu_mes_suspend(struct amdgpu_device *adev);
+int amdgpu_mes_resume(struct amdgpu_device *adev);
+
+int amdgpu_mes_map_legacy_queue(struct amdgpu_device *adev,
+				struct amdgpu_ring *ring);
+int amdgpu_mes_unmap_legacy_queue(struct amdgpu_device *adev,
+				  struct amdgpu_ring *ring,
+				  enum amdgpu_unmap_queues_action action,
+				  u64 gpu_addr, u64 seq);
+int amdgpu_mes_reset_legacy_queue(struct amdgpu_device *adev,
+				  struct amdgpu_ring *ring,
+				  unsigned int vmid,
+				  bool use_mmio);
+
+int amdgpu_mes_get_hung_queue_db_array_size(struct amdgpu_device *adev);
+int amdgpu_mes_detect_and_reset_hung_queues(struct amdgpu_device *adev,
+					    int queue_type,
+					    bool detect_only,
+					    unsigned int *hung_db_num,
+					    u32 *hung_db_array);
+
+uint32_t amdgpu_mes_rreg(struct amdgpu_device *adev, uint32_t reg);
+int amdgpu_mes_wreg(struct amdgpu_device *adev,
+		    uint32_t reg, uint32_t val);
+int amdgpu_mes_reg_write_reg_wait(struct amdgpu_device *adev,
+				  uint32_t reg0, uint32_t reg1,
+				  uint32_t ref, uint32_t mask);
+int amdgpu_mes_hdp_flush(struct amdgpu_device *adev);
+int amdgpu_mes_set_shader_debugger(struct amdgpu_device *adev,
+				uint64_t process_context_addr,
+				uint32_t spi_gdbg_per_vmid_cntl,
+				const uint32_t *tcp_watch_cntl,
+				uint32_t flags,
+				bool trap_en);
+int amdgpu_mes_flush_shader_debugger(struct amdgpu_device *adev,
+				uint64_t process_context_addr);
+
+uint32_t amdgpu_mes_get_aggregated_doorbell_index(struct amdgpu_device *adev,
+						   enum amdgpu_mes_priority_level prio);
+
+int amdgpu_mes_doorbell_process_slice(struct amdgpu_device *adev);
+
+/*
+ * MES lock can be taken in MMU notifiers.
+ *
+ * A bit more detail about why to set no-FS reclaim with MES lock:
+ *
+ * The purpose of the MMU notifier is to stop GPU access to memory so
+ * that the Linux VM subsystem can move pages around safely. This is
+ * done by preempting user mode queues for the affected process. When
+ * MES is used, MES lock needs to be taken to preempt the queues.
+ *
+ * The MMU notifier callback entry point in the driver is
+ * amdgpu_mn_invalidate_range_start_hsa. The relevant call chain from
+ * there is:
+ * amdgpu_amdkfd_evict_userptr -> kgd2kfd_quiesce_mm ->
+ * kfd_process_evict_queues -> pdd->dev->dqm->ops.evict_process_queues
+ *
+ * The last part of the chain is a function pointer where we take the
+ * MES lock.
+ *
+ * The problem with taking locks in the MMU notifier is, that MMU
+ * notifiers can be called in reclaim-FS context. That's where the
+ * kernel frees up pages to make room for new page allocations under
+ * memory pressure. While we are running in reclaim-FS context, we must
+ * not trigger another memory reclaim operation because that would
+ * recursively reenter the reclaim code and cause a deadlock. The
+ * memalloc_nofs_save/restore calls guarantee that.
+ *
+ * In addition we also need to avoid lock dependencies on other locks taken
+ * under the MES lock, for example reservation locks. Here is a possible
+ * scenario of a deadlock:
+ * Thread A: takes and holds reservation lock | triggers reclaim-FS |
+ * MMU notifier | blocks trying to take MES lock
+ * Thread B: takes and holds MES lock | blocks trying to take reservation lock
+ *
+ * In this scenario Thread B gets involved in a deadlock even without
+ * triggering a reclaim-FS operation itself.
+ * To fix this and break the lock dependency chain you'd need to either:
+ * 1. protect reservation locks with memalloc_nofs_save/restore, or
+ * 2. avoid taking reservation locks under the MES lock.
+ *
+ * Reservation locks are taken all over the kernel in different subsystems, we
+ * have no control over them and their lock dependencies.So the only workable
+ * solution is to avoid taking other locks under the MES lock.
+ * As a result, make sure no reclaim-FS happens while holding this lock anywhere
+ * to prevent deadlocks when an MMU notifier runs in reclaim-FS context.
+ */
+static inline void amdgpu_mes_lock(struct amdgpu_mes *mes)
+{
+	mutex_lock(&mes->mutex_hidden);
+	mes->saved_flags = memalloc_noreclaim_save();
+}
+
+static inline void amdgpu_mes_unlock(struct amdgpu_mes *mes)
+{
+	memalloc_noreclaim_restore(mes->saved_flags);
+	mutex_unlock(&mes->mutex_hidden);
+}
+
+bool amdgpu_mes_suspend_resume_all_supported(struct amdgpu_device *adev);
+
+int amdgpu_mes_update_enforce_isolation(struct amdgpu_device *adev);
+
 #endif /* __AMDGPU_MES_H__ */
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes_ctx.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes_ctx.h
new file mode 100644
index 000000000000..912a5be2ece6
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes_ctx.h
@@ -0,0 +1,122 @@
+/*
+ * Copyright 2019 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __AMDGPU_MES_CTX_H__
+#define __AMDGPU_MES_CTX_H__
+
+#include "v10_structs.h"
+
+enum {
+	AMDGPU_MES_CTX_RPTR_OFFS = 0,
+	AMDGPU_MES_CTX_WPTR_OFFS,
+	AMDGPU_MES_CTX_FENCE_OFFS,
+	AMDGPU_MES_CTX_COND_EXE_OFFS,
+	AMDGPU_MES_CTX_TRAIL_FENCE_OFFS,
+	AMDGPU_MES_CTX_MAX_OFFS,
+};
+
+enum {
+	AMDGPU_MES_CTX_RING_OFFS = AMDGPU_MES_CTX_MAX_OFFS,
+	AMDGPU_MES_CTX_IB_OFFS,
+	AMDGPU_MES_CTX_PADDING_OFFS,
+};
+
+#define AMDGPU_MES_CTX_MAX_GFX_RINGS            1
+#define AMDGPU_MES_CTX_MAX_COMPUTE_RINGS        4
+#define AMDGPU_MES_CTX_MAX_SDMA_RINGS           2
+#define AMDGPU_MES_CTX_MAX_RINGS					\
+	(AMDGPU_MES_CTX_MAX_GFX_RINGS +					\
+	 AMDGPU_MES_CTX_MAX_COMPUTE_RINGS +				\
+	 AMDGPU_MES_CTX_MAX_SDMA_RINGS)
+
+#define AMDGPU_CSA_SDMA_SIZE    64
+#define GFX10_MEC_HPD_SIZE	2048
+
+struct amdgpu_wb_slot {
+	uint32_t data[8];
+};
+
+struct amdgpu_mes_ctx_meta_data {
+	struct {
+		uint8_t ring[PAGE_SIZE * 4];
+
+		/* gfx csa */
+		struct v10_gfx_meta_data gfx_meta_data;
+
+		uint8_t gds_backup[64 * 1024];
+
+		struct amdgpu_wb_slot slots[AMDGPU_MES_CTX_MAX_OFFS];
+
+		/* only for ib test */
+		uint32_t ib[256] __aligned(256);
+
+		uint32_t padding[64];
+
+	} __aligned(PAGE_SIZE) gfx[AMDGPU_MES_CTX_MAX_GFX_RINGS];
+
+	struct {
+		uint8_t ring[PAGE_SIZE * 4];
+
+		uint8_t mec_hpd[GFX10_MEC_HPD_SIZE];
+
+		struct amdgpu_wb_slot slots[AMDGPU_MES_CTX_MAX_OFFS];
+
+		/* only for ib test */
+		uint32_t ib[256] __aligned(256);
+
+		uint32_t padding[64];
+
+	} __aligned(PAGE_SIZE) compute[AMDGPU_MES_CTX_MAX_COMPUTE_RINGS];
+
+	struct {
+		uint8_t ring[PAGE_SIZE * 4];
+
+		/* sdma csa for mcbp */
+		uint8_t sdma_meta_data[AMDGPU_CSA_SDMA_SIZE];
+
+		struct amdgpu_wb_slot slots[AMDGPU_MES_CTX_MAX_OFFS];
+
+		/* only for ib test */
+		uint32_t ib[256] __aligned(256);
+
+		uint32_t padding[64];
+
+	} __aligned(PAGE_SIZE) sdma[AMDGPU_MES_CTX_MAX_SDMA_RINGS];
+};
+
+struct amdgpu_mes_ctx_data {
+	struct amdgpu_bo	*meta_data_obj;
+	uint64_t                meta_data_gpu_addr;
+	uint64_t                meta_data_mc_addr;
+	struct amdgpu_bo_va	*meta_data_va;
+	void                    *meta_data_ptr;
+	uint32_t                gang_ids[AMDGPU_HW_IP_DMA+1];
+};
+
+#define AMDGPU_FENCE_MES_QUEUE_FLAG     0x1000000u
+#define AMDGPU_FENCE_MES_QUEUE_ID_MASK  (AMDGPU_FENCE_MES_QUEUE_FLAG - 1)
+
+#define AMDGPU_FENCE_MES_QUEUE_FLAG     0x1000000u
+#define AMDGPU_FENCE_MES_QUEUE_ID_MASK  (AMDGPU_FENCE_MES_QUEUE_FLAG - 1)
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mmhub.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_mmhub.c
index 24297dc51434..0f6b1021fef3 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mmhub.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mmhub.c
@@ -1,5 +1,5 @@
 /*
- * Copyright 2019 Advanced Micro Devices, Inc.
+ * Copyright (C) 2023  Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -8,61 +8,39 @@
  * and/or sell copies of the Software, and to permit persons to whom the
  * Software is furnished to do so, subject to the following conditions:
  *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
  *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
  * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
+ * THE COPYRIGHT HOLDER(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
+ * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-
 #include "amdgpu.h"
 #include "amdgpu_ras.h"
 
-int amdgpu_mmhub_ras_late_init(struct amdgpu_device *adev)
+int amdgpu_mmhub_ras_sw_init(struct amdgpu_device *adev)
 {
-	int r;
-	struct ras_ih_if ih_info = {
-		.cb = NULL,
-	};
-	struct ras_fs_if fs_info = {
-		.sysfs_name = "mmhub_err_count",
-	};
+	int err;
+	struct amdgpu_mmhub_ras *ras;
 
-	if (!adev->mmhub.ras_if) {
-		adev->mmhub.ras_if = kmalloc(sizeof(struct ras_common_if), GFP_KERNEL);
-		if (!adev->mmhub.ras_if)
-			return -ENOMEM;
-		adev->mmhub.ras_if->block = AMDGPU_RAS_BLOCK__MMHUB;
-		adev->mmhub.ras_if->type = AMDGPU_RAS_ERROR__MULTI_UNCORRECTABLE;
-		adev->mmhub.ras_if->sub_block_index = 0;
-	}
-	ih_info.head = fs_info.head = *adev->mmhub.ras_if;
-	r = amdgpu_ras_late_init(adev, adev->mmhub.ras_if,
-				 &fs_info, &ih_info);
-	if (r || !amdgpu_ras_is_supported(adev, adev->mmhub.ras_if->block)) {
-		kfree(adev->mmhub.ras_if);
-		adev->mmhub.ras_if = NULL;
-	}
+	if (!adev->mmhub.ras)
+		return 0;
 
-	return r;
-}
+	ras = adev->mmhub.ras;
+	err = amdgpu_ras_register_ras_block(adev, &ras->ras_block);
+	if (err) {
+		dev_err(adev->dev, "Failed to register mmhub ras block!\n");
+		return err;
+	}
 
-void amdgpu_mmhub_ras_fini(struct amdgpu_device *adev)
-{
-	if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__MMHUB) &&
-			adev->mmhub.ras_if) {
-		struct ras_common_if *ras_if = adev->mmhub.ras_if;
-		struct ras_ih_if ih_info = {
-			.cb = NULL,
-		};
+	strcpy(ras->ras_block.ras_comm.name, "mmhub");
+	ras->ras_block.ras_comm.block = AMDGPU_RAS_BLOCK__MMHUB;
+	ras->ras_block.ras_comm.type = AMDGPU_RAS_ERROR__MULTI_UNCORRECTABLE;
+	adev->mmhub.ras_if = &ras->ras_block.ras_comm;
 
-		amdgpu_ras_late_fini(adev, ras_if, &ih_info);
-		kfree(ras_if);
-	}
+	/* mmhub ras follows amdgpu_ras_block_late_init_default for late init */
+	return 0;
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mmhub.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_mmhub.h
index b27fcbccce2b..1ca9d4ed8063 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mmhub.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mmhub.h
@@ -21,18 +21,36 @@
 #ifndef __AMDGPU_MMHUB_H__
 #define __AMDGPU_MMHUB_H__
 
-struct amdgpu_mmhub_ras_funcs {
-	int (*ras_late_init)(struct amdgpu_device *adev);
-	void (*ras_fini)(struct amdgpu_device *adev);
-	void (*query_ras_error_count)(struct amdgpu_device *adev,
-				      void *ras_error_status);
-	void (*query_ras_error_status)(struct amdgpu_device *adev);
-	void (*reset_ras_error_count)(struct amdgpu_device *adev);
-	void (*reset_ras_error_status)(struct amdgpu_device *adev);
+enum amdgpu_mmhub_ras_memory_id {
+	AMDGPU_MMHUB_WGMI_PAGEMEM = 0,
+	AMDGPU_MMHUB_RGMI_PAGEMEM = 1,
+	AMDGPU_MMHUB_WDRAM_PAGEMEM = 2,
+	AMDGPU_MMHUB_RDRAM_PAGEMEM = 3,
+	AMDGPU_MMHUB_WIO_CMDMEM = 4,
+	AMDGPU_MMHUB_RIO_CMDMEM = 5,
+	AMDGPU_MMHUB_WGMI_CMDMEM = 6,
+	AMDGPU_MMHUB_RGMI_CMDMEM = 7,
+	AMDGPU_MMHUB_WDRAM_CMDMEM = 8,
+	AMDGPU_MMHUB_RDRAM_CMDMEM = 9,
+	AMDGPU_MMHUB_MAM_DMEM0 = 10,
+	AMDGPU_MMHUB_MAM_DMEM1 = 11,
+	AMDGPU_MMHUB_MAM_DMEM2 = 12,
+	AMDGPU_MMHUB_MAM_DMEM3 = 13,
+	AMDGPU_MMHUB_WRET_TAGMEM = 19,
+	AMDGPU_MMHUB_RRET_TAGMEM = 20,
+	AMDGPU_MMHUB_WIO_DATAMEM = 21,
+	AMDGPU_MMHUB_WGMI_DATAMEM = 22,
+	AMDGPU_MMHUB_WDRAM_DATAMEM = 23,
+	AMDGPU_MMHUB_MEMORY_BLOCK_LAST,
+};
+
+struct amdgpu_mmhub_ras {
+	struct amdgpu_ras_block_object ras_block;
 };
 
 struct amdgpu_mmhub_funcs {
 	u64 (*get_fb_location)(struct amdgpu_device *adev);
+	u64 (*get_mc_fb_offset)(struct amdgpu_device *adev);
 	void (*init)(struct amdgpu_device *adev);
 	int (*gart_enable)(struct amdgpu_device *adev);
 	void (*set_fault_enable_default)(struct amdgpu_device *adev,
@@ -40,7 +58,7 @@ struct amdgpu_mmhub_funcs {
 	void (*gart_disable)(struct amdgpu_device *adev);
 	int (*set_clockgating)(struct amdgpu_device *adev,
 			       enum amd_clockgating_state state);
-	void (*get_clockgating)(struct amdgpu_device *adev, u32 *flags);
+	void (*get_clockgating)(struct amdgpu_device *adev, u64 *flags);
 	void (*setup_vm_pt_regs)(struct amdgpu_device *adev, uint32_t vmid,
 				uint64_t page_table_base);
 	void (*update_power_gating)(struct amdgpu_device *adev,
@@ -50,10 +68,10 @@ struct amdgpu_mmhub_funcs {
 struct amdgpu_mmhub {
 	struct ras_common_if *ras_if;
 	const struct amdgpu_mmhub_funcs *funcs;
-	const struct amdgpu_mmhub_ras_funcs *ras_funcs;
+	struct amdgpu_mmhub_ras  *ras;
 };
 
-int amdgpu_mmhub_ras_late_init(struct amdgpu_device *adev);
-void amdgpu_mmhub_ras_fini(struct amdgpu_device *adev);
+int amdgpu_mmhub_ras_sw_init(struct amdgpu_device *adev);
+
 #endif
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mn.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_mn.c
deleted file mode 100644
index 4b153daf283d..000000000000
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mn.c
+++ /dev/null
@@ -1,244 +0,0 @@
-/*
- * Copyright 2014 Advanced Micro Devices, Inc.
- * All Rights Reserved.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the
- * "Software"), to deal in the Software without restriction, including
- * without limitation the rights to use, copy, modify, merge, publish,
- * distribute, sub license, and/or sell copies of the Software, and to
- * permit persons to whom the Software is furnished to do so, subject to
- * the following conditions:
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
- * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
- * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
- * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
- * USE OR OTHER DEALINGS IN THE SOFTWARE.
- *
- * The above copyright notice and this permission notice (including the
- * next paragraph) shall be included in all copies or substantial portions
- * of the Software.
- *
- */
-/*
- * Authors:
- *    Christian König <christian.koenig@amd.com>
- */
-
-/**
- * DOC: MMU Notifier
- *
- * For coherent userptr handling registers an MMU notifier to inform the driver
- * about updates on the page tables of a process.
- *
- * When somebody tries to invalidate the page tables we block the update until
- * all operations on the pages in question are completed, then those pages are
- * marked as accessed and also dirty if it wasn't a read only access.
- *
- * New command submissions using the userptrs in question are delayed until all
- * page table invalidation are completed and we once more see a coherent process
- * address space.
- */
-
-#include <linux/firmware.h>
-#include <linux/module.h>
-#include <drm/drm.h>
-
-#include "amdgpu.h"
-#include "amdgpu_amdkfd.h"
-
-/**
- * amdgpu_mn_invalidate_gfx - callback to notify about mm change
- *
- * @mni: the range (mm) is about to update
- * @range: details on the invalidation
- * @cur_seq: Value to pass to mmu_interval_set_seq()
- *
- * Block for operations on BOs to finish and mark pages as accessed and
- * potentially dirty.
- */
-static bool amdgpu_mn_invalidate_gfx(struct mmu_interval_notifier *mni,
-				     const struct mmu_notifier_range *range,
-				     unsigned long cur_seq)
-{
-	struct amdgpu_bo *bo = container_of(mni, struct amdgpu_bo, notifier);
-	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
-	long r;
-
-	if (!mmu_notifier_range_blockable(range))
-		return false;
-
-	mutex_lock(&adev->notifier_lock);
-
-	mmu_interval_set_seq(mni, cur_seq);
-
-	r = dma_resv_wait_timeout(bo->tbo.base.resv, true, false,
-				  MAX_SCHEDULE_TIMEOUT);
-	mutex_unlock(&adev->notifier_lock);
-	if (r <= 0)
-		DRM_ERROR("(%ld) failed to wait for user bo\n", r);
-	return true;
-}
-
-static const struct mmu_interval_notifier_ops amdgpu_mn_gfx_ops = {
-	.invalidate = amdgpu_mn_invalidate_gfx,
-};
-
-/**
- * amdgpu_mn_invalidate_hsa - callback to notify about mm change
- *
- * @mni: the range (mm) is about to update
- * @range: details on the invalidation
- * @cur_seq: Value to pass to mmu_interval_set_seq()
- *
- * We temporarily evict the BO attached to this range. This necessitates
- * evicting all user-mode queues of the process.
- */
-static bool amdgpu_mn_invalidate_hsa(struct mmu_interval_notifier *mni,
-				     const struct mmu_notifier_range *range,
-				     unsigned long cur_seq)
-{
-	struct amdgpu_bo *bo = container_of(mni, struct amdgpu_bo, notifier);
-	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
-
-	if (!mmu_notifier_range_blockable(range))
-		return false;
-
-	mutex_lock(&adev->notifier_lock);
-
-	mmu_interval_set_seq(mni, cur_seq);
-
-	amdgpu_amdkfd_evict_userptr(bo->kfd_bo, bo->notifier.mm);
-	mutex_unlock(&adev->notifier_lock);
-
-	return true;
-}
-
-static const struct mmu_interval_notifier_ops amdgpu_mn_hsa_ops = {
-	.invalidate = amdgpu_mn_invalidate_hsa,
-};
-
-/**
- * amdgpu_mn_register - register a BO for notifier updates
- *
- * @bo: amdgpu buffer object
- * @addr: userptr addr we should monitor
- *
- * Registers a mmu_notifier for the given BO at the specified address.
- * Returns 0 on success, -ERRNO if anything goes wrong.
- */
-int amdgpu_mn_register(struct amdgpu_bo *bo, unsigned long addr)
-{
-	if (bo->kfd_bo)
-		return mmu_interval_notifier_insert(&bo->notifier, current->mm,
-						    addr, amdgpu_bo_size(bo),
-						    &amdgpu_mn_hsa_ops);
-	return mmu_interval_notifier_insert(&bo->notifier, current->mm, addr,
-					    amdgpu_bo_size(bo),
-					    &amdgpu_mn_gfx_ops);
-}
-
-/**
- * amdgpu_mn_unregister - unregister a BO for notifier updates
- *
- * @bo: amdgpu buffer object
- *
- * Remove any registration of mmu notifier updates from the buffer object.
- */
-void amdgpu_mn_unregister(struct amdgpu_bo *bo)
-{
-	if (!bo->notifier.mm)
-		return;
-	mmu_interval_notifier_remove(&bo->notifier);
-	bo->notifier.mm = NULL;
-}
-
-int amdgpu_hmm_range_get_pages(struct mmu_interval_notifier *notifier,
-			       struct mm_struct *mm, struct page **pages,
-			       uint64_t start, uint64_t npages,
-			       struct hmm_range **phmm_range, bool readonly,
-			       bool mmap_locked, void *owner)
-{
-	struct hmm_range *hmm_range;
-	unsigned long timeout;
-	unsigned long i;
-	unsigned long *pfns;
-	int r = 0;
-
-	hmm_range = kzalloc(sizeof(*hmm_range), GFP_KERNEL);
-	if (unlikely(!hmm_range))
-		return -ENOMEM;
-
-	pfns = kvmalloc_array(npages, sizeof(*pfns), GFP_KERNEL);
-	if (unlikely(!pfns)) {
-		r = -ENOMEM;
-		goto out_free_range;
-	}
-
-	hmm_range->notifier = notifier;
-	hmm_range->default_flags = HMM_PFN_REQ_FAULT;
-	if (!readonly)
-		hmm_range->default_flags |= HMM_PFN_REQ_WRITE;
-	hmm_range->hmm_pfns = pfns;
-	hmm_range->start = start;
-	hmm_range->end = start + npages * PAGE_SIZE;
-	hmm_range->dev_private_owner = owner;
-
-	/* Assuming 512MB takes maxmium 1 second to fault page address */
-	timeout = max(npages >> 17, 1ULL) * HMM_RANGE_DEFAULT_TIMEOUT;
-	timeout = jiffies + msecs_to_jiffies(timeout);
-
-retry:
-	hmm_range->notifier_seq = mmu_interval_read_begin(notifier);
-
-	if (likely(!mmap_locked))
-		mmap_read_lock(mm);
-
-	r = hmm_range_fault(hmm_range);
-
-	if (likely(!mmap_locked))
-		mmap_read_unlock(mm);
-	if (unlikely(r)) {
-		/*
-		 * FIXME: This timeout should encompass the retry from
-		 * mmu_interval_read_retry() as well.
-		 */
-		if (r == -EBUSY && !time_after(jiffies, timeout))
-			goto retry;
-		goto out_free_pfns;
-	}
-
-	/*
-	 * Due to default_flags, all pages are HMM_PFN_VALID or
-	 * hmm_range_fault() fails. FIXME: The pages cannot be touched outside
-	 * the notifier_lock, and mmu_interval_read_retry() must be done first.
-	 */
-	for (i = 0; pages && i < npages; i++)
-		pages[i] = hmm_pfn_to_page(pfns[i]);
-
-	*phmm_range = hmm_range;
-
-	return 0;
-
-out_free_pfns:
-	kvfree(pfns);
-out_free_range:
-	kfree(hmm_range);
-
-	return r;
-}
-
-int amdgpu_hmm_range_get_pages_done(struct hmm_range *hmm_range)
-{
-	int r;
-
-	r = mmu_interval_read_retry(hmm_range->notifier,
-				    hmm_range->notifier_seq);
-	kvfree(hmm_range->hmm_pfns);
-	kfree(hmm_range);
-
-	return r;
-}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mn.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_mn.h
deleted file mode 100644
index 14a3c1864085..000000000000
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mn.h
+++ /dev/null
@@ -1,53 +0,0 @@
-/*
- * Copyright 2017 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: Christian König
- */
-#ifndef __AMDGPU_MN_H__
-#define __AMDGPU_MN_H__
-
-#include <linux/types.h>
-#include <linux/hmm.h>
-#include <linux/rwsem.h>
-#include <linux/workqueue.h>
-#include <linux/interval_tree.h>
-
-int amdgpu_hmm_range_get_pages(struct mmu_interval_notifier *notifier,
-			       struct mm_struct *mm, struct page **pages,
-			       uint64_t start, uint64_t npages,
-			       struct hmm_range **phmm_range, bool readonly,
-			       bool mmap_locked, void *owner);
-int amdgpu_hmm_range_get_pages_done(struct hmm_range *hmm_range);
-
-#if defined(CONFIG_HMM_MIRROR)
-int amdgpu_mn_register(struct amdgpu_bo *bo, unsigned long addr);
-void amdgpu_mn_unregister(struct amdgpu_bo *bo);
-#else
-static inline int amdgpu_mn_register(struct amdgpu_bo *bo, unsigned long addr)
-{
-	DRM_WARN_ONCE("HMM_MIRROR kernel config option is not enabled, "
-		      "add CONFIG_ZONE_DEVICE=y in config file to fix this\n");
-	return -ENODEV;
-}
-static inline void amdgpu_mn_unregister(struct amdgpu_bo *bo) {}
-#endif
-
-#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mode.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_mode.h
index 89fb372ed49c..dc8d2f52c7d6 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mode.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mode.h
@@ -30,21 +30,18 @@
 #ifndef AMDGPU_MODE_H
 #define AMDGPU_MODE_H
 
+#include <drm/display/drm_dp_helper.h>
 #include <drm/drm_crtc.h>
-#include <drm/drm_edid.h>
 #include <drm/drm_encoder.h>
-#include <drm/drm_dp_helper.h>
 #include <drm/drm_fixed.h>
-#include <drm/drm_crtc_helper.h>
-#include <drm/drm_fb_helper.h>
-#include <drm/drm_plane_helper.h>
+#include <drm/drm_framebuffer.h>
 #include <drm/drm_probe_helper.h>
 #include <linux/i2c.h>
 #include <linux/i2c-algo-bit.h>
 #include <linux/hrtimer.h>
 #include "amdgpu_irq.h"
 
-#include <drm/drm_dp_mst_helper.h>
+#include <drm/display/drm_dp_mst_helper.h>
 #include "modules/inc/mod_freesync.h"
 #include "amdgpu_dm_irq_params.h"
 
@@ -53,6 +50,8 @@ struct amdgpu_device;
 struct amdgpu_encoder;
 struct amdgpu_router;
 struct amdgpu_hpd;
+struct edid;
+struct drm_edid;
 
 #define to_amdgpu_crtc(x) container_of(x, struct amdgpu_crtc, base)
 #define to_amdgpu_connector(x) container_of(x, struct amdgpu_connector, base)
@@ -232,8 +231,6 @@ struct amdgpu_i2c_chan {
 	struct mutex mutex;
 };
 
-struct amdgpu_fbdev;
-
 struct amdgpu_afmt {
 	bool enabled;
 	int offset;
@@ -304,18 +301,12 @@ struct amdgpu_framebuffer {
 
 	uint64_t tiling_flags;
 	bool tmz_surface;
+	bool gfx12_dcc;
 
 	/* caching for later use */
 	uint64_t address;
 };
 
-struct amdgpu_fbdev {
-	struct drm_fb_helper helper;
-	struct amdgpu_framebuffer rfb;
-	struct list_head fbdev_list;
-	struct amdgpu_device *adev;
-};
-
 struct amdgpu_mode_info {
 	struct atom_context *atom_context;
 	struct card_info *atom_card_info;
@@ -338,11 +329,8 @@ struct amdgpu_mode_info {
 	/* Adaptive Backlight Modulation (power feature) */
 	struct drm_property *abm_level_property;
 	/* hardcoded DFP edid from BIOS */
-	struct edid *bios_hardcoded_edid;
-	int bios_hardcoded_edid_size;
+	const struct drm_edid *bios_hardcoded_edid;
 
-	/* pointer to fbdev info structure */
-	struct amdgpu_fbdev *rfbdev;
 	/* firmware flags */
 	u32 firmware_flags;
 	/* pointer to backlight encoder */
@@ -352,22 +340,110 @@ struct amdgpu_mode_info {
 	int			num_crtc; /* number of crtcs */
 	int			num_hpd; /* number of hpd pins */
 	int			num_dig; /* number of dig blocks */
+	bool			gpu_vm_support; /* supports display from GTT */
 	int			disp_priority;
 	const struct amdgpu_display_funcs *funcs;
 	const enum drm_plane_type *plane_type;
+
+	/* Driver-private color mgmt props */
+
+	/* @plane_degamma_lut_property: Plane property to set a degamma LUT to
+	 * convert encoded values to light linear values before sampling or
+	 * blending.
+	 */
+	struct drm_property *plane_degamma_lut_property;
+	/* @plane_degamma_lut_size_property: Plane property to define the max
+	 * size of degamma LUT as supported by the driver (read-only).
+	 */
+	struct drm_property *plane_degamma_lut_size_property;
+	/**
+	 * @plane_degamma_tf_property: Plane pre-defined transfer function to
+	 * to go from scanout/encoded values to linear values.
+	 */
+	struct drm_property *plane_degamma_tf_property;
+	/**
+	 * @plane_hdr_mult_property:
+	 */
+	struct drm_property *plane_hdr_mult_property;
+
+	struct drm_property *plane_ctm_property;
+	/**
+	 * @shaper_lut_property: Plane property to set pre-blending shaper LUT
+	 * that converts color content before 3D LUT. If
+	 * plane_shaper_tf_property != Identity TF, AMD color module will
+	 * combine the user LUT values with pre-defined TF into the LUT
+	 * parameters to be programmed.
+	 */
+	struct drm_property *plane_shaper_lut_property;
+	/**
+	 * @shaper_lut_size_property: Plane property for the size of
+	 * pre-blending shaper LUT as supported by the driver (read-only).
+	 */
+	struct drm_property *plane_shaper_lut_size_property;
+	/**
+	 * @plane_shaper_tf_property: Plane property to set a predefined
+	 * transfer function for pre-blending shaper (before applying 3D LUT)
+	 * with or without LUT. There is no shaper ROM, but we can use AMD
+	 * color modules to program LUT parameters from predefined TF (or
+	 * from a combination of pre-defined TF and the custom 1D LUT).
+	 */
+	struct drm_property *plane_shaper_tf_property;
+	/**
+	 * @plane_lut3d_property: Plane property for color transformation using
+	 * a 3D LUT (pre-blending), a three-dimensional array where each
+	 * element is an RGB triplet. Each dimension has the size of
+	 * lut3d_size. The array contains samples from the approximated
+	 * function. On AMD, values between samples are estimated by
+	 * tetrahedral interpolation. The array is accessed with three indices,
+	 * one for each input dimension (color channel), blue being the
+	 * outermost dimension, red the innermost.
+	 */
+	struct drm_property *plane_lut3d_property;
+	/**
+	 * @plane_degamma_lut_size_property: Plane property to define the max
+	 * size of 3D LUT as supported by the driver (read-only). The max size
+	 * is the max size of one dimension and, therefore, the max number of
+	 * entries for 3D LUT array is the 3D LUT size cubed;
+	 */
+	struct drm_property *plane_lut3d_size_property;
+	/**
+	 * @plane_blend_lut_property: Plane property for output gamma before
+	 * blending. Userspace set a blend LUT to convert colors after 3D LUT
+	 * conversion. It works as a post-3DLUT 1D LUT. With shaper LUT, they
+	 * are sandwiching 3D LUT with two 1D LUT. If plane_blend_tf_property
+	 * != Identity TF, AMD color module will combine the user LUT values
+	 * with pre-defined TF into the LUT parameters to be programmed.
+	 */
+	struct drm_property *plane_blend_lut_property;
+	/**
+	 * @plane_blend_lut_size_property: Plane property to define the max
+	 * size of blend LUT as supported by the driver (read-only).
+	 */
+	struct drm_property *plane_blend_lut_size_property;
+	/**
+	 * @plane_blend_tf_property: Plane property to set a predefined
+	 * transfer function for pre-blending blend/out_gamma (after applying
+	 * 3D LUT) with or without LUT. There is no blend ROM, but we can use
+	 * AMD color modules to program LUT parameters from predefined TF (or
+	 * from a combination of pre-defined TF and the custom 1D LUT).
+	 */
+	struct drm_property *plane_blend_tf_property;
+	/* @regamma_tf_property: Transfer function for CRTC regamma
+	 * (post-blending). Possible values are defined by `enum
+	 * amdgpu_transfer_function`. There is no regamma ROM, but we can use
+	 * AMD color modules to program LUT parameters from predefined TF (or
+	 * from a combination of pre-defined TF and the custom 1D LUT).
+	 */
+	struct drm_property *regamma_tf_property;
 };
 
 #define AMDGPU_MAX_BL_LEVEL 0xFF
 
-#if defined(CONFIG_BACKLIGHT_CLASS_DEVICE) || defined(CONFIG_BACKLIGHT_CLASS_DEVICE_MODULE)
-
 struct amdgpu_backlight_privdata {
 	struct amdgpu_encoder *encoder;
 	uint8_t negative;
 };
 
-#endif
-
 struct amdgpu_atom_ss {
 	uint16_t percentage;
 	uint16_t percentage_divider;
@@ -422,8 +498,6 @@ struct amdgpu_crtc {
 	struct drm_connector *connector;
 	/* for dpm */
 	u32 line_time;
-	u32 wm_low;
-	u32 wm_high;
 	u32 lb_vblank_lead_lines;
 	struct drm_display_mode hw_mode;
 	/* for virtual dce */
@@ -432,6 +506,10 @@ struct amdgpu_crtc {
 
 	int otg_inst;
 	struct drm_pending_vblank_event *event;
+
+	bool wb_pending;
+	bool wb_enabled;
+	struct drm_writeback_connector *wb_conn;
 };
 
 struct amdgpu_encoder_atom_dig {
@@ -531,6 +609,7 @@ struct amdgpu_i2c_adapter {
 	struct i2c_adapter base;
 
 	struct ddc_service *ddc_service;
+	bool oem;
 };
 
 #define TO_DM_AUX(x) container_of((x), struct amdgpu_dm_dp_aux, aux)
@@ -549,6 +628,7 @@ struct amdgpu_connector {
 	void *con_priv;
 	bool dac_load_detect;
 	bool detected_by_load; /* if the connection status was determined by load */
+	bool detected_hpd_without_ddc; /* if an HPD signal was detected on DVI, but ddc probing failed */
 	uint16_t connector_object_id;
 	struct amdgpu_hpd hpd;
 	struct amdgpu_router router;
@@ -564,8 +644,8 @@ struct amdgpu_mst_connector {
 
 	struct drm_dp_mst_topology_mgr mst_mgr;
 	struct amdgpu_dm_dp_aux dm_dp_aux;
-	struct drm_dp_mst_port *port;
-	struct amdgpu_connector *mst_port;
+	struct drm_dp_mst_port *mst_output_port;
+	struct amdgpu_connector *mst_root;
 	bool is_mst_connector;
 	struct amdgpu_encoder *mst_encoder;
 };
@@ -602,19 +682,6 @@ int amdgpu_display_get_crtc_scanoutpos(struct drm_device *dev,
 			int *hpos, ktime_t *stime, ktime_t *etime,
 			const struct drm_display_mode *mode);
 
-int amdgpu_display_gem_fb_init(struct drm_device *dev,
-			       struct amdgpu_framebuffer *rfb,
-			       const struct drm_mode_fb_cmd2 *mode_cmd,
-			       struct drm_gem_object *obj);
-int amdgpu_display_gem_fb_verify_and_init(
-	struct drm_device *dev, struct amdgpu_framebuffer *rfb,
-	struct drm_file *file_priv, const struct drm_mode_fb_cmd2 *mode_cmd,
-	struct drm_gem_object *obj);
-int amdgpu_display_framebuffer_init(struct drm_device *dev,
-				    struct amdgpu_framebuffer *rfb,
-				    const struct drm_mode_fb_cmd2 *mode_cmd,
-				    struct drm_gem_object *obj);
-
 int amdgpufb_remove(struct drm_device *dev, struct drm_framebuffer *fb);
 
 void amdgpu_enc_destroy(struct drm_encoder *encoder);
@@ -631,15 +698,6 @@ bool amdgpu_crtc_get_scanout_position(struct drm_crtc *crtc,
 			int *hpos, ktime_t *stime, ktime_t *etime,
 			const struct drm_display_mode *mode);
 
-/* fbdev layer */
-int amdgpu_fbdev_init(struct amdgpu_device *adev);
-void amdgpu_fbdev_fini(struct amdgpu_device *adev);
-void amdgpu_fbdev_set_suspend(struct amdgpu_device *adev, int state);
-int amdgpu_fbdev_total_size(struct amdgpu_device *adev);
-bool amdgpu_fbdev_robj_is_fb(struct amdgpu_device *adev, struct amdgpu_bo *robj);
-
-int amdgpu_align_pitch(struct amdgpu_device *adev, int width, int bpp, bool tiled);
-
 /* amdgpu_display.c */
 void amdgpu_display_print_display_setup(struct drm_device *dev);
 int amdgpu_display_modeset_create_props(struct amdgpu_device *adev);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_nbio.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_nbio.c
index 6afb02fef8cf..a974265837f0 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_nbio.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_nbio.c
@@ -22,61 +22,65 @@
 #include "amdgpu.h"
 #include "amdgpu_ras.h"
 
-int amdgpu_nbio_ras_late_init(struct amdgpu_device *adev)
+int amdgpu_nbio_ras_sw_init(struct amdgpu_device *adev)
 {
-	int r;
-	struct ras_ih_if ih_info = {
-		.cb = NULL,
-	};
-	struct ras_fs_if fs_info = {
-		.sysfs_name = "pcie_bif_err_count",
-	};
+	int err;
+	struct amdgpu_nbio_ras *ras;
+
+	if (!adev->nbio.ras)
+		return 0;
 
-	if (!adev->nbio.ras_if) {
-		adev->nbio.ras_if = kmalloc(sizeof(struct ras_common_if), GFP_KERNEL);
-		if (!adev->nbio.ras_if)
-			return -ENOMEM;
-		adev->nbio.ras_if->block = AMDGPU_RAS_BLOCK__PCIE_BIF;
-		adev->nbio.ras_if->type = AMDGPU_RAS_ERROR__MULTI_UNCORRECTABLE;
-		adev->nbio.ras_if->sub_block_index = 0;
+	ras = adev->nbio.ras;
+	err = amdgpu_ras_register_ras_block(adev, &ras->ras_block);
+	if (err) {
+		dev_err(adev->dev, "Failed to register pcie_bif ras block!\n");
+		return err;
 	}
-	ih_info.head = fs_info.head = *adev->nbio.ras_if;
-	r = amdgpu_ras_late_init(adev, adev->nbio.ras_if,
-				 &fs_info, &ih_info);
+
+	strcpy(ras->ras_block.ras_comm.name, "pcie_bif");
+	ras->ras_block.ras_comm.block = AMDGPU_RAS_BLOCK__PCIE_BIF;
+	ras->ras_block.ras_comm.type = AMDGPU_RAS_ERROR__MULTI_UNCORRECTABLE;
+	adev->nbio.ras_if = &ras->ras_block.ras_comm;
+
+	return 0;
+}
+
+u64 amdgpu_nbio_get_pcie_replay_count(struct amdgpu_device *adev)
+{
+	if (adev->nbio.funcs && adev->nbio.funcs->get_pcie_replay_count)
+		return adev->nbio.funcs->get_pcie_replay_count(adev);
+
+	return 0;
+}
+
+bool amdgpu_nbio_is_replay_cnt_supported(struct amdgpu_device *adev)
+{
+	if (amdgpu_sriov_vf(adev) || !adev->asic_funcs ||
+	    !adev->asic_funcs->get_pcie_replay_count ||
+	    (!adev->nbio.funcs || !adev->nbio.funcs->get_pcie_replay_count))
+		return false;
+
+	return true;
+}
+
+int amdgpu_nbio_ras_late_init(struct amdgpu_device *adev, struct ras_common_if *ras_block)
+{
+	int r;
+	r = amdgpu_ras_block_late_init(adev, ras_block);
 	if (r)
-		goto free;
+		return r;
 
-	if (amdgpu_ras_is_supported(adev, adev->nbio.ras_if->block)) {
+	if (amdgpu_ras_is_supported(adev, ras_block->block)) {
 		r = amdgpu_irq_get(adev, &adev->nbio.ras_controller_irq, 0);
 		if (r)
 			goto late_fini;
 		r = amdgpu_irq_get(adev, &adev->nbio.ras_err_event_athub_irq, 0);
 		if (r)
 			goto late_fini;
-	} else {
-		r = 0;
-		goto free;
 	}
 
 	return 0;
 late_fini:
-	amdgpu_ras_late_fini(adev, adev->nbio.ras_if, &ih_info);
-free:
-	kfree(adev->nbio.ras_if);
-	adev->nbio.ras_if = NULL;
+	amdgpu_ras_block_late_fini(adev, ras_block);
 	return r;
 }
-
-void amdgpu_nbio_ras_fini(struct amdgpu_device *adev)
-{
-	if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__PCIE_BIF) &&
-			adev->nbio.ras_if) {
-		struct ras_common_if *ras_if = adev->nbio.ras_if;
-		struct ras_ih_if ih_info = {
-			.cb = NULL,
-		};
-
-		amdgpu_ras_late_fini(adev, ras_if, &ih_info);
-		kfree(ras_if);
-	}
-}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_nbio.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_nbio.h
index 843052205bd5..b528de6a01f6 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_nbio.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_nbio.h
@@ -47,15 +47,12 @@ struct nbio_hdp_flush_reg {
 	u32 ref_and_mask_sdma7;
 };
 
-struct amdgpu_nbio_ras_funcs {
+struct amdgpu_nbio_ras {
+	struct amdgpu_ras_block_object ras_block;
 	void (*handle_ras_controller_intr_no_bifring)(struct amdgpu_device *adev);
 	void (*handle_ras_err_event_athub_intr_no_bifring)(struct amdgpu_device *adev);
 	int (*init_ras_controller_interrupt)(struct amdgpu_device *adev);
 	int (*init_ras_err_event_athub_interrupt)(struct amdgpu_device *adev);
-	void (*query_ras_error_count)(struct amdgpu_device *adev,
-				      void *ras_error_status);
-	int (*ras_late_init)(struct amdgpu_device *adev);
-	void (*ras_fini)(struct amdgpu_device *adev);
 };
 
 struct amdgpu_nbio_funcs {
@@ -64,6 +61,7 @@ struct amdgpu_nbio_funcs {
 	u32 (*get_hdp_flush_done_offset)(struct amdgpu_device *adev);
 	u32 (*get_pcie_index_offset)(struct amdgpu_device *adev);
 	u32 (*get_pcie_data_offset)(struct amdgpu_device *adev);
+	u32 (*get_pcie_index_hi_offset)(struct amdgpu_device *adev);
 	u32 (*get_pcie_port_index_offset)(struct amdgpu_device *adev);
 	u32 (*get_pcie_port_data_offset)(struct amdgpu_device *adev);
 	u32 (*get_rev_id)(struct amdgpu_device *adev);
@@ -71,8 +69,11 @@ struct amdgpu_nbio_funcs {
 	u32 (*get_memsize)(struct amdgpu_device *adev);
 	void (*sdma_doorbell_range)(struct amdgpu_device *adev, int instance,
 			bool use_doorbell, int doorbell_index, int doorbell_size);
+	void (*vpe_doorbell_range)(struct amdgpu_device *adev, int instance,
+			bool use_doorbell, int doorbell_index, int doorbell_size);
 	void (*vcn_doorbell_range)(struct amdgpu_device *adev, bool use_doorbell,
 				   int doorbell_index, int instance);
+	void (*gc_doorbell_init)(struct amdgpu_device *adev);
 	void (*enable_doorbell_aperture)(struct amdgpu_device *adev,
 					 bool enable);
 	void (*enable_doorbell_selfring_aperture)(struct amdgpu_device *adev,
@@ -86,7 +87,7 @@ struct amdgpu_nbio_funcs {
 	void (*update_medium_grain_light_sleep)(struct amdgpu_device *adev,
 						bool enable);
 	void (*get_clockgating_state)(struct amdgpu_device *adev,
-				      u32 *flags);
+				      u64 *flags);
 	void (*ih_control)(struct amdgpu_device *adev);
 	void (*init_registers)(struct amdgpu_device *adev);
 	void (*remap_hdp_registers)(struct amdgpu_device *adev);
@@ -96,6 +97,13 @@ struct amdgpu_nbio_funcs {
 	void (*apply_lc_spc_mode_wa)(struct amdgpu_device *adev);
 	void (*apply_l1_link_width_reconfig_wa)(struct amdgpu_device *adev);
 	void (*clear_doorbell_interrupt)(struct amdgpu_device *adev);
+	u32 (*get_rom_offset)(struct amdgpu_device *adev);
+	int (*get_compute_partition_mode)(struct amdgpu_device *adev);
+	u32 (*get_memory_partition_mode)(struct amdgpu_device *adev,
+					 u32 *supp_modes);
+	bool (*is_nps_switch_requested)(struct amdgpu_device *adev);
+	u64 (*get_pcie_replay_count)(struct amdgpu_device *adev);
+	void (*set_reg_remap)(struct amdgpu_device *adev);
 };
 
 struct amdgpu_nbio {
@@ -104,9 +112,13 @@ struct amdgpu_nbio {
 	struct amdgpu_irq_src ras_err_event_athub_irq;
 	struct ras_common_if *ras_if;
 	const struct amdgpu_nbio_funcs *funcs;
-	const struct amdgpu_nbio_ras_funcs *ras_funcs;
+	struct amdgpu_nbio_ras  *ras;
 };
 
-int amdgpu_nbio_ras_late_init(struct amdgpu_device *adev);
-void amdgpu_nbio_ras_fini(struct amdgpu_device *adev);
+int amdgpu_nbio_ras_sw_init(struct amdgpu_device *adev);
+int amdgpu_nbio_ras_late_init(struct amdgpu_device *adev, struct ras_common_if *ras_block);
+u64 amdgpu_nbio_get_pcie_replay_count(struct amdgpu_device *adev);
+
+bool amdgpu_nbio_is_replay_cnt_supported(struct amdgpu_device *adev);
+
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
index 4fcfc2313b8c..e08f58de4b17 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
@@ -32,12 +32,17 @@
 #include <linux/list.h>
 #include <linux/slab.h>
 #include <linux/dma-buf.h>
+#include <linux/export.h>
 
+#include <drm/drm_drv.h>
 #include <drm/amdgpu_drm.h>
 #include <drm/drm_cache.h>
 #include "amdgpu.h"
 #include "amdgpu_trace.h"
 #include "amdgpu_amdkfd.h"
+#include "amdgpu_vram_mgr.h"
+#include "amdgpu_vm.h"
+#include "amdgpu_dma_buf.h"
 
 /**
  * DOC: amdgpu_object
@@ -58,7 +63,7 @@ static void amdgpu_bo_destroy(struct ttm_buffer_object *tbo)
 
 	amdgpu_bo_kunmap(bo);
 
-	if (bo->tbo.base.import_attach)
+	if (drm_gem_is_imported(&bo->tbo.base))
 		drm_prime_gem_destroy(&bo->tbo.base, bo->tbo.sg);
 	drm_gem_object_release(&bo->tbo.base);
 	amdgpu_bo_unref(&bo->parent);
@@ -75,23 +80,6 @@ static void amdgpu_bo_user_destroy(struct ttm_buffer_object *tbo)
 	amdgpu_bo_destroy(tbo);
 }
 
-static void amdgpu_bo_vm_destroy(struct ttm_buffer_object *tbo)
-{
-	struct amdgpu_device *adev = amdgpu_ttm_adev(tbo->bdev);
-	struct amdgpu_bo *bo = ttm_to_amdgpu_bo(tbo);
-	struct amdgpu_bo_vm *vmbo;
-
-	vmbo = to_amdgpu_bo_vm(bo);
-	/* in case amdgpu_device_recover_vram got NULL of bo->parent */
-	if (!list_empty(&vmbo->shadow_list)) {
-		mutex_lock(&adev->shadow_list_lock);
-		list_del_init(&vmbo->shadow_list);
-		mutex_unlock(&adev->shadow_list_lock);
-	}
-
-	amdgpu_bo_destroy(tbo);
-}
-
 /**
  * amdgpu_bo_is_amdgpu_bo - check if the buffer object is an &amdgpu_bo
  * @bo: buffer object to be checked
@@ -105,8 +93,7 @@ static void amdgpu_bo_vm_destroy(struct ttm_buffer_object *tbo)
 bool amdgpu_bo_is_amdgpu_bo(struct ttm_buffer_object *bo)
 {
 	if (bo->destroy == &amdgpu_bo_destroy ||
-	    bo->destroy == &amdgpu_bo_user_destroy ||
-	    bo->destroy == &amdgpu_bo_vm_destroy)
+	    bo->destroy == &amdgpu_bo_user_destroy)
 		return true;
 
 	return false;
@@ -129,20 +116,48 @@ void amdgpu_bo_placement_from_domain(struct amdgpu_bo *abo, u32 domain)
 	u32 c = 0;
 
 	if (domain & AMDGPU_GEM_DOMAIN_VRAM) {
-		unsigned visible_pfn = adev->gmc.visible_vram_size >> PAGE_SHIFT;
-
-		places[c].fpfn = 0;
-		places[c].lpfn = 0;
+		unsigned int visible_pfn = adev->gmc.visible_vram_size >> PAGE_SHIFT;
+		int8_t mem_id = KFD_XCP_MEM_ID(adev, abo->xcp_id);
+
+		if (adev->gmc.mem_partitions && mem_id >= 0) {
+			places[c].fpfn = adev->gmc.mem_partitions[mem_id].range.fpfn;
+			/*
+			 * memory partition range lpfn is inclusive start + size - 1
+			 * TTM place lpfn is exclusive start + size
+			 */
+			places[c].lpfn = adev->gmc.mem_partitions[mem_id].range.lpfn + 1;
+		} else {
+			places[c].fpfn = 0;
+			places[c].lpfn = 0;
+		}
 		places[c].mem_type = TTM_PL_VRAM;
 		places[c].flags = 0;
 
 		if (flags & AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED)
-			places[c].lpfn = visible_pfn;
+			places[c].lpfn = min_not_zero(places[c].lpfn, visible_pfn);
 		else
 			places[c].flags |= TTM_PL_FLAG_TOPDOWN;
 
-		if (flags & AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS)
+		if (abo->tbo.type == ttm_bo_type_kernel &&
+		    flags & AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS)
 			places[c].flags |= TTM_PL_FLAG_CONTIGUOUS;
+
+		c++;
+	}
+
+	if (domain & AMDGPU_GEM_DOMAIN_DOORBELL) {
+		places[c].fpfn = 0;
+		places[c].lpfn = 0;
+		places[c].mem_type = AMDGPU_PL_DOORBELL;
+		places[c].flags = 0;
+		c++;
+	}
+
+	if (domain & AMDGPU_GEM_DOMAIN_MMIO_REMAP) {
+		places[c].fpfn = 0;
+		places[c].lpfn = 0;
+		places[c].mem_type = AMDGPU_PL_MMIO_REMAP;
+		places[c].flags = 0;
 		c++;
 	}
 
@@ -153,6 +168,13 @@ void amdgpu_bo_placement_from_domain(struct amdgpu_bo *abo, u32 domain)
 			abo->flags & AMDGPU_GEM_CREATE_PREEMPTIBLE ?
 			AMDGPU_PL_PREEMPT : TTM_PL_TT;
 		places[c].flags = 0;
+		/*
+		 * When GTT is just an alternative to VRAM make sure that we
+		 * only use it as fallback and still try to fill up VRAM first.
+		 */
+		if (abo->tbo.resource && !(adev->flags & AMD_IS_APU) &&
+		    domain & abo->preferred_domains & AMDGPU_GEM_DOMAIN_VRAM)
+			places[c].flags |= TTM_PL_FLAG_FALLBACK;
 		c++;
 	}
 
@@ -200,9 +222,6 @@ void amdgpu_bo_placement_from_domain(struct amdgpu_bo *abo, u32 domain)
 
 	placement->num_placement = c;
 	placement->placement = places;
-
-	placement->num_busy_placement = c;
-	placement->busy_placement = places;
 }
 
 /**
@@ -315,6 +334,9 @@ error_free:
  *
  * Allocates and pins a BO for kernel internal use.
  *
+ * This function is exported to allow the V4L2 isp device
+ * external to drm device to create and access the kernel BO.
+ *
  * Note: For bo_ptr new BO is only created if bo_ptr points to NULL.
  *
  * Returns:
@@ -340,22 +362,89 @@ int amdgpu_bo_create_kernel(struct amdgpu_device *adev,
 }
 
 /**
+ * amdgpu_bo_create_isp_user - create user BO for isp
+ *
+ * @adev: amdgpu device object
+ * @dma_buf: DMABUF handle for isp buffer
+ * @domain: where to place it
+ * @bo:  used to initialize BOs in structures
+ * @gpu_addr: GPU addr of the pinned BO
+ *
+ * Imports isp DMABUF to allocate and pin a user BO for isp internal use. It does
+ * GART alloc to generate gpu_addr for BO to make it accessible through the
+ * GART aperture for ISP HW.
+ *
+ * This function is exported to allow the V4L2 isp device external to drm device
+ * to create and access the isp user BO.
+ *
+ * Returns:
+ * 0 on success, negative error code otherwise.
+ */
+int amdgpu_bo_create_isp_user(struct amdgpu_device *adev,
+			   struct dma_buf *dma_buf, u32 domain, struct amdgpu_bo **bo,
+			   u64 *gpu_addr)
+
+{
+	struct drm_gem_object *gem_obj;
+	int r;
+
+	gem_obj = amdgpu_gem_prime_import(&adev->ddev, dma_buf);
+	*bo = gem_to_amdgpu_bo(gem_obj);
+	if (!(*bo)) {
+		dev_err(adev->dev, "failed to get valid isp user bo\n");
+		return -EINVAL;
+	}
+
+	r = amdgpu_bo_reserve(*bo, false);
+	if (r) {
+		dev_err(adev->dev, "(%d) failed to reserve isp user bo\n", r);
+		return r;
+	}
+
+	r = amdgpu_bo_pin(*bo, domain);
+	if (r) {
+		dev_err(adev->dev, "(%d) isp user bo pin failed\n", r);
+		goto error_unreserve;
+	}
+
+	r = amdgpu_ttm_alloc_gart(&(*bo)->tbo);
+	if (r) {
+		dev_err(adev->dev, "%p bind failed\n", *bo);
+		goto error_unpin;
+	}
+
+	if (!WARN_ON(!gpu_addr))
+		*gpu_addr = amdgpu_bo_gpu_offset(*bo);
+
+	amdgpu_bo_unreserve(*bo);
+
+	return 0;
+
+error_unpin:
+	amdgpu_bo_unpin(*bo);
+error_unreserve:
+	amdgpu_bo_unreserve(*bo);
+	amdgpu_bo_unref(bo);
+
+	return r;
+}
+
+/**
  * amdgpu_bo_create_kernel_at - create BO for kernel use at specific location
  *
  * @adev: amdgpu device object
  * @offset: offset of the BO
  * @size: size of the BO
- * @domain: where to place it
  * @bo_ptr:  used to initialize BOs in structures
  * @cpu_addr: optional CPU address mapping
  *
- * Creates a kernel BO at a specific offset in the address space of the domain.
+ * Creates a kernel BO at a specific offset in VRAM.
  *
  * Returns:
  * 0 on success, negative error code otherwise.
  */
 int amdgpu_bo_create_kernel_at(struct amdgpu_device *adev,
-			       uint64_t offset, uint64_t size, uint32_t domain,
+			       uint64_t offset, uint64_t size,
 			       struct amdgpu_bo **bo_ptr, void **cpu_addr)
 {
 	struct ttm_operation_ctx ctx = { false, false };
@@ -365,8 +454,9 @@ int amdgpu_bo_create_kernel_at(struct amdgpu_device *adev,
 	offset &= PAGE_MASK;
 	size = ALIGN(size, PAGE_SIZE);
 
-	r = amdgpu_bo_create_reserved(adev, size, PAGE_SIZE, domain, bo_ptr,
-				      NULL, cpu_addr);
+	r = amdgpu_bo_create_reserved(adev, size, PAGE_SIZE,
+				      AMDGPU_GEM_DOMAIN_VRAM, bo_ptr, NULL,
+				      cpu_addr);
 	if (r)
 		return r;
 
@@ -414,6 +504,9 @@ error:
  * @cpu_addr: pointer to where the BO's CPU memory space address was stored
  *
  * unmaps and unpin a BO for kernel internal use.
+ *
+ * This function is exported to allow the V4L2 isp device
+ * external to drm device to free the kernel BO.
  */
 void amdgpu_bo_free_kernel(struct amdgpu_bo **bo, u64 *gpu_addr,
 			   void **cpu_addr)
@@ -421,6 +514,8 @@ void amdgpu_bo_free_kernel(struct amdgpu_bo **bo, u64 *gpu_addr,
 	if (*bo == NULL)
 		return;
 
+	WARN_ON(amdgpu_ttm_adev((*bo)->tbo.bdev)->in_suspend);
+
 	if (likely(amdgpu_bo_reserve(*bo, true) == 0)) {
 		if (cpu_addr)
 			amdgpu_bo_kunmap(*bo);
@@ -437,7 +532,29 @@ void amdgpu_bo_free_kernel(struct amdgpu_bo **bo, u64 *gpu_addr,
 		*cpu_addr = NULL;
 }
 
-/* Validate bo size is bit bigger then the request domain */
+/**
+ * amdgpu_bo_free_isp_user - free BO for isp use
+ *
+ * @bo: amdgpu isp user BO to free
+ *
+ * unpin and unref BO for isp internal use.
+ *
+ * This function is exported to allow the V4L2 isp device
+ * external to drm device to free the isp user BO.
+ */
+void amdgpu_bo_free_isp_user(struct amdgpu_bo *bo)
+{
+	if (bo == NULL)
+		return;
+
+	if (amdgpu_bo_reserve(bo, true) == 0) {
+		amdgpu_bo_unpin(bo);
+		amdgpu_bo_unreserve(bo);
+	}
+	amdgpu_bo_unref(&bo);
+}
+
+/* Validate bo size is bit bigger than the request domain */
 static bool amdgpu_bo_validate_size(struct amdgpu_device *adev,
 					  unsigned long size, u32 domain)
 {
@@ -445,33 +562,26 @@ static bool amdgpu_bo_validate_size(struct amdgpu_device *adev,
 
 	/*
 	 * If GTT is part of requested domains the check must succeed to
-	 * allow fall back to GTT
+	 * allow fall back to GTT.
 	 */
-	if (domain & AMDGPU_GEM_DOMAIN_GTT) {
+	if (domain & AMDGPU_GEM_DOMAIN_GTT)
 		man = ttm_manager_type(&adev->mman.bdev, TTM_PL_TT);
-
-		if (size < (man->size << PAGE_SHIFT))
-			return true;
-		else
-			goto fail;
-	}
-
-	if (domain & AMDGPU_GEM_DOMAIN_VRAM) {
+	else if (domain & AMDGPU_GEM_DOMAIN_VRAM)
 		man = ttm_manager_type(&adev->mman.bdev, TTM_PL_VRAM);
+	else
+		return true;
 
-		if (size < (man->size << PAGE_SHIFT))
-			return true;
-		else
-			goto fail;
+	if (!man) {
+		if (domain & AMDGPU_GEM_DOMAIN_GTT)
+			WARN_ON_ONCE("GTT domain requested but GTT mem manager uninitialized");
+		return false;
 	}
 
+	/* TODO add more domains checks, such as AMDGPU_GEM_DOMAIN_CPU, _DOMAIN_DOORBELL */
+	if (size < man->size)
+		return true;
 
-	/* TODO add more domains checks, such as AMDGPU_GEM_DOMAIN_CPU */
-	return true;
-
-fail:
-	DRM_DEBUG("BO size %lu > total memory in domain: %llu\n", size,
-		  man->size << PAGE_SHIFT);
+	DRM_DEBUG("BO size %lu > total memory in domain: %llu\n", size, man->size);
 	return false;
 }
 
@@ -541,6 +651,7 @@ int amdgpu_bo_create(struct amdgpu_device *adev,
 		/* GWS and OA don't need any alignment. */
 		page_align = bp->byte_align;
 		size <<= PAGE_SHIFT;
+
 	} else if (bp->domain & AMDGPU_GEM_DOMAIN_GDS) {
 		/* Both size and alignment must be a multiple of 4. */
 		page_align = ALIGN(bp->byte_align, 4);
@@ -561,16 +672,25 @@ int amdgpu_bo_create(struct amdgpu_device *adev,
 	if (bo == NULL)
 		return -ENOMEM;
 	drm_gem_private_object_init(adev_to_drm(adev), &bo->tbo.base, size);
+	bo->tbo.base.funcs = &amdgpu_gem_object_funcs;
 	bo->vm_bo = NULL;
 	bo->preferred_domains = bp->preferred_domain ? bp->preferred_domain :
 		bp->domain;
 	bo->allowed_domains = bo->preferred_domains;
 	if (bp->type != ttm_bo_type_kernel &&
+	    !(bp->flags & AMDGPU_GEM_CREATE_DISCARDABLE) &&
 	    bo->allowed_domains == AMDGPU_GEM_DOMAIN_VRAM)
 		bo->allowed_domains |= AMDGPU_GEM_DOMAIN_GTT;
 
 	bo->flags = bp->flags;
 
+	if (adev->gmc.mem_partitions)
+		/* For GPUs with spatial partitioning, bo->xcp_id=-1 means any partition */
+		bo->xcp_id = bp->xcp_id_plus1 - 1;
+	else
+		/* For GPUs without spatial partitioning */
+		bo->xcp_id = 0;
+
 	if (!amdgpu_bo_support_uswc(bo->flags))
 		bo->flags &= ~AMDGPU_GEM_CREATE_CPU_GTT_USWC;
 
@@ -581,20 +701,21 @@ int amdgpu_bo_create(struct amdgpu_device *adev,
 	else
 		amdgpu_bo_placement_from_domain(bo, bp->domain);
 	if (bp->type == ttm_bo_type_kernel)
+		bo->tbo.priority = 2;
+	else if (!(bp->flags & AMDGPU_GEM_CREATE_DISCARDABLE))
 		bo->tbo.priority = 1;
 
 	if (!bp->destroy)
 		bp->destroy = &amdgpu_bo_destroy;
 
-	r = ttm_bo_init_reserved(&adev->mman.bdev, &bo->tbo, size, bp->type,
+	r = ttm_bo_init_reserved(&adev->mman.bdev, &bo->tbo, bp->type,
 				 &bo->placement, page_align, &ctx,  NULL,
 				 bp->resv, bp->destroy);
 	if (unlikely(r != 0))
 		return r;
 
 	if (!amdgpu_gmc_vram_full_visible(&adev->gmc) &&
-	    bo->tbo.resource->mem_type == TTM_PL_VRAM &&
-	    bo->tbo.resource->start < adev->gmc.visible_vram_size >> PAGE_SHIFT)
+	    amdgpu_res_cpu_visible(adev, bo->tbo.resource))
 		amdgpu_cs_report_moved_bytes(adev, ctx.bytes_moved,
 					     ctx.bytes_moved);
 	else
@@ -604,13 +725,12 @@ int amdgpu_bo_create(struct amdgpu_device *adev,
 	    bo->tbo.resource->mem_type == TTM_PL_VRAM) {
 		struct dma_fence *fence;
 
-		r = amdgpu_fill_buffer(bo, 0, bo->tbo.base.resv, &fence);
+		r = amdgpu_ttm_clear_buffer(bo, bo->tbo.base.resv, &fence);
 		if (unlikely(r))
 			goto fail_unreserve;
 
-		amdgpu_bo_fence(bo, fence, false);
-		dma_fence_put(bo->tbo.moving);
-		bo->tbo.moving = dma_fence_get(fence);
+		dma_resv_add_fence(bo->tbo.base.resv, fence,
+				   DMA_RESV_USAGE_KERNEL);
 		dma_fence_put(fence);
 	}
 	if (!bp->resv)
@@ -684,61 +804,15 @@ int amdgpu_bo_create_vm(struct amdgpu_device *adev,
 	 * num of amdgpu_vm_pt entries.
 	 */
 	BUG_ON(bp->bo_ptr_size < sizeof(struct amdgpu_bo_vm));
-	bp->destroy = &amdgpu_bo_vm_destroy;
 	r = amdgpu_bo_create(adev, bp, &bo_ptr);
 	if (r)
 		return r;
 
 	*vmbo_ptr = to_amdgpu_bo_vm(bo_ptr);
-	INIT_LIST_HEAD(&(*vmbo_ptr)->shadow_list);
 	return r;
 }
 
 /**
- * amdgpu_bo_add_to_shadow_list - add a BO to the shadow list
- *
- * @vmbo: BO that will be inserted into the shadow list
- *
- * Insert a BO to the shadow list.
- */
-void amdgpu_bo_add_to_shadow_list(struct amdgpu_bo_vm *vmbo)
-{
-	struct amdgpu_device *adev = amdgpu_ttm_adev(vmbo->bo.tbo.bdev);
-
-	mutex_lock(&adev->shadow_list_lock);
-	list_add_tail(&vmbo->shadow_list, &adev->shadow_list);
-	mutex_unlock(&adev->shadow_list_lock);
-}
-
-/**
- * amdgpu_bo_restore_shadow - restore an &amdgpu_bo shadow
- *
- * @shadow: &amdgpu_bo shadow to be restored
- * @fence: dma_fence associated with the operation
- *
- * Copies a buffer object's shadow content back to the object.
- * This is used for recovering a buffer from its shadow in case of a gpu
- * reset where vram context may be lost.
- *
- * Returns:
- * 0 for success or a negative error code on failure.
- */
-int amdgpu_bo_restore_shadow(struct amdgpu_bo *shadow, struct dma_fence **fence)
-
-{
-	struct amdgpu_device *adev = amdgpu_ttm_adev(shadow->tbo.bdev);
-	struct amdgpu_ring *ring = adev->mman.buffer_funcs_ring;
-	uint64_t shadow_addr, parent_addr;
-
-	shadow_addr = amdgpu_bo_gpu_offset(shadow);
-	parent_addr = amdgpu_bo_gpu_offset(shadow->parent);
-
-	return amdgpu_copy_buffer(ring, shadow_addr, parent_addr,
-				  amdgpu_bo_size(shadow), NULL, fence,
-				  true, false, false);
-}
-
-/**
  * amdgpu_bo_kmap - map an &amdgpu_bo buffer object
  * @bo: &amdgpu_bo buffer object to be mapped
  * @ptr: kernel virtual address to be returned
@@ -757,6 +831,11 @@ int amdgpu_bo_kmap(struct amdgpu_bo *bo, void **ptr)
 	if (bo->flags & AMDGPU_GEM_CREATE_NO_CPU_ACCESS)
 		return -EPERM;
 
+	r = dma_resv_wait_timeout(bo->tbo.base.resv, DMA_RESV_USAGE_KERNEL,
+				  false, MAX_SCHEDULE_TIMEOUT);
+	if (r < 0)
+		return r;
+
 	kptr = amdgpu_bo_kptr(bo);
 	if (kptr) {
 		if (ptr)
@@ -764,12 +843,7 @@ int amdgpu_bo_kmap(struct amdgpu_bo *bo, void **ptr)
 		return 0;
 	}
 
-	r = dma_resv_wait_timeout(bo->tbo.base.resv, false, false,
-				  MAX_SCHEDULE_TIMEOUT);
-	if (r < 0)
-		return r;
-
-	r = ttm_bo_kmap(&bo->tbo, 0, bo->tbo.resource->num_pages, &bo->kmap);
+	r = ttm_bo_kmap(&bo->tbo, 0, PFN_UP(bo->tbo.base.size), &bo->kmap);
 	if (r)
 		return r;
 
@@ -821,7 +895,7 @@ struct amdgpu_bo *amdgpu_bo_ref(struct amdgpu_bo *bo)
 	if (bo == NULL)
 		return NULL;
 
-	ttm_bo_get(&bo->tbo);
+	drm_gem_object_get(&bo->tbo.base);
 	return bo;
 }
 
@@ -833,40 +907,30 @@ struct amdgpu_bo *amdgpu_bo_ref(struct amdgpu_bo *bo)
  */
 void amdgpu_bo_unref(struct amdgpu_bo **bo)
 {
-	struct ttm_buffer_object *tbo;
-
 	if ((*bo) == NULL)
 		return;
 
-	tbo = &((*bo)->tbo);
-	ttm_bo_put(tbo);
+	drm_gem_object_put(&(*bo)->tbo.base);
 	*bo = NULL;
 }
 
 /**
- * amdgpu_bo_pin_restricted - pin an &amdgpu_bo buffer object
+ * amdgpu_bo_pin - pin an &amdgpu_bo buffer object
  * @bo: &amdgpu_bo buffer object to be pinned
  * @domain: domain to be pinned to
- * @min_offset: the start of requested address range
- * @max_offset: the end of requested address range
  *
- * Pins the buffer object according to requested domain and address range. If
- * the memory is unbound gart memory, binds the pages into gart table. Adjusts
- * pin_count and pin_size accordingly.
+ * Pins the buffer object according to requested domain. If the memory is
+ * unbound gart memory, binds the pages into gart table. Adjusts pin_count and
+ * pin_size accordingly.
  *
  * Pinning means to lock pages in memory along with keeping them at a fixed
  * offset. It is required when a buffer can not be moved, for example, when
  * a display buffer is being scanned out.
  *
- * Compared with amdgpu_bo_pin(), this function gives more flexibility on
- * where to pin a buffer if there are specific restrictions on where a buffer
- * must be located.
- *
  * Returns:
  * 0 for success or a negative error code on failure.
  */
-int amdgpu_bo_pin_restricted(struct amdgpu_bo *bo, u32 domain,
-			     u64 min_offset, u64 max_offset)
+int amdgpu_bo_pin(struct amdgpu_bo *bo, u32 domain)
 {
 	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
 	struct ttm_operation_ctx ctx = { false, false };
@@ -875,11 +939,12 @@ int amdgpu_bo_pin_restricted(struct amdgpu_bo *bo, u32 domain,
 	if (amdgpu_ttm_tt_get_usermm(bo->tbo.ttm))
 		return -EPERM;
 
-	if (WARN_ON_ONCE(min_offset > max_offset))
-		return -EINVAL;
+	/* Check domain to be pinned to against preferred domains */
+	if (bo->preferred_domains & domain)
+		domain = bo->preferred_domains & domain;
 
 	/* A shared bo cannot be migrated to VRAM */
-	if (bo->tbo.base.import_attach) {
+	if (drm_gem_is_imported(&bo->tbo.base)) {
 		if (domain & AMDGPU_GEM_DOMAIN_GTT)
 			domain = AMDGPU_GEM_DOMAIN_GTT;
 		else
@@ -899,14 +964,6 @@ int amdgpu_bo_pin_restricted(struct amdgpu_bo *bo, u32 domain,
 			return -EINVAL;
 
 		ttm_bo_pin(&bo->tbo);
-
-		if (max_offset != 0) {
-			u64 domain_start = amdgpu_ttm_domain_start(adev,
-								   mem_type);
-			WARN_ON_ONCE(max_offset <
-				     (amdgpu_bo_gpu_offset(bo) - domain_start));
-		}
-
 		return 0;
 	}
 
@@ -915,7 +972,7 @@ int amdgpu_bo_pin_restricted(struct amdgpu_bo *bo, u32 domain,
 	 */
 	domain = amdgpu_bo_get_preferred_domain(adev, domain);
 
-	if (bo->tbo.base.import_attach)
+	if (drm_gem_is_imported(&bo->tbo.base))
 		dma_buf_pin(bo->tbo.base.import_attach);
 
 	/* force to pin into visible video ram */
@@ -923,16 +980,9 @@ int amdgpu_bo_pin_restricted(struct amdgpu_bo *bo, u32 domain,
 		bo->flags |= AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED;
 	amdgpu_bo_placement_from_domain(bo, domain);
 	for (i = 0; i < bo->placement.num_placement; i++) {
-		unsigned fpfn, lpfn;
-
-		fpfn = min_offset >> PAGE_SHIFT;
-		lpfn = max_offset >> PAGE_SHIFT;
-
-		if (fpfn > bo->placements[i].fpfn)
-			bo->placements[i].fpfn = fpfn;
-		if (!bo->placements[i].lpfn ||
-		    (lpfn && lpfn < bo->placements[i].lpfn))
-			bo->placements[i].lpfn = lpfn;
+		if (bo->flags & AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS &&
+		    bo->placements[i].mem_type == TTM_PL_VRAM)
+			bo->placements[i].flags |= TTM_PL_FLAG_CONTIGUOUS;
 	}
 
 	r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
@@ -943,12 +993,11 @@ int amdgpu_bo_pin_restricted(struct amdgpu_bo *bo, u32 domain,
 
 	ttm_bo_pin(&bo->tbo);
 
-	domain = amdgpu_mem_type_to_domain(bo->tbo.resource->mem_type);
-	if (domain == AMDGPU_GEM_DOMAIN_VRAM) {
+	if (bo->tbo.resource->mem_type == TTM_PL_VRAM) {
 		atomic64_add(amdgpu_bo_size(bo), &adev->vram_pin_size);
 		atomic64_add(amdgpu_vram_mgr_bo_visible_size(bo),
 			     &adev->visible_pin_size);
-	} else if (domain == AMDGPU_GEM_DOMAIN_GTT) {
+	} else if (bo->tbo.resource->mem_type == TTM_PL_TT) {
 		atomic64_add(amdgpu_bo_size(bo), &adev->gart_pin_size);
 	}
 
@@ -957,24 +1006,6 @@ error:
 }
 
 /**
- * amdgpu_bo_pin - pin an &amdgpu_bo buffer object
- * @bo: &amdgpu_bo buffer object to be pinned
- * @domain: domain to be pinned to
- *
- * A simple wrapper to amdgpu_bo_pin_restricted().
- * Provides a simpler API for buffers that do not have any strict restrictions
- * on where a buffer must be located.
- *
- * Returns:
- * 0 for success or a negative error code on failure.
- */
-int amdgpu_bo_pin(struct amdgpu_bo *bo, u32 domain)
-{
-	bo->flags |= AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS;
-	return amdgpu_bo_pin_restricted(bo, domain, 0, 0);
-}
-
-/**
  * amdgpu_bo_unpin - unpin an &amdgpu_bo buffer object
  * @bo: &amdgpu_bo buffer object to be unpinned
  *
@@ -992,7 +1023,7 @@ void amdgpu_bo_unpin(struct amdgpu_bo *bo)
 	if (bo->tbo.pin_count)
 		return;
 
-	if (bo->tbo.base.import_attach)
+	if (drm_gem_is_imported(&bo->tbo.base))
 		dma_buf_unpin(bo->tbo.base.import_attach);
 
 	if (bo->tbo.resource->mem_type == TTM_PL_VRAM) {
@@ -1002,9 +1033,10 @@ void amdgpu_bo_unpin(struct amdgpu_bo *bo)
 	} else if (bo->tbo.resource->mem_type == TTM_PL_TT) {
 		atomic64_sub(amdgpu_bo_size(bo), &adev->gart_pin_size);
 	}
+
 }
 
-static const char *amdgpu_vram_names[] = {
+static const char * const amdgpu_vram_names[] = {
 	"UNKNOWN",
 	"GDDR1",
 	"DDR2",
@@ -1015,7 +1047,10 @@ static const char *amdgpu_vram_names[] = {
 	"DDR3",
 	"DDR4",
 	"GDDR6",
-	"DDR5"
+	"DDR5",
+	"LPDDR4",
+	"LPDDR5",
+	"HBM3E"
 };
 
 /**
@@ -1030,11 +1065,16 @@ static const char *amdgpu_vram_names[] = {
 int amdgpu_bo_init(struct amdgpu_device *adev)
 {
 	/* On A+A platform, VRAM can be mapped as WB */
-	if (!adev->gmc.xgmi.connected_to_cpu) {
+	if (!adev->gmc.xgmi.connected_to_cpu && !adev->gmc.is_app_apu) {
 		/* reserve PAT memory space to WC for VRAM */
-		arch_io_reserve_memtype_wc(adev->gmc.aper_base,
+		int r = arch_io_reserve_memtype_wc(adev->gmc.aper_base,
 				adev->gmc.aper_size);
 
+		if (r) {
+			DRM_ERROR("Unable to set WC memtype for the aperture base\n");
+			return r;
+		}
+
 		/* Add an MTRR for the VRAM */
 		adev->gmc.vram_mtrr = arch_phys_wc_add(adev->gmc.aper_base,
 				adev->gmc.aper_size);
@@ -1056,7 +1096,17 @@ int amdgpu_bo_init(struct amdgpu_device *adev)
  */
 void amdgpu_bo_fini(struct amdgpu_device *adev)
 {
+	int idx;
+
 	amdgpu_ttm_fini(adev);
+
+	if (drm_dev_enter(adev_to_drm(adev), &idx)) {
+		if (!adev->gmc.xgmi.connected_to_cpu && !adev->gmc.is_app_apu) {
+			arch_phys_wc_del(adev->gmc.vram_mtrr);
+			arch_io_free_memtype_wc(adev->gmc.aper_base, adev->gmc.aper_size);
+		}
+		drm_dev_exit(idx);
+	}
 }
 
 /**
@@ -1118,8 +1168,8 @@ void amdgpu_bo_get_tiling_flags(struct amdgpu_bo *bo, u64 *tiling_flags)
  * Returns:
  * 0 for success or a negative error code on failure.
  */
-int amdgpu_bo_set_metadata (struct amdgpu_bo *bo, void *metadata,
-			    uint32_t metadata_size, uint64_t flags)
+int amdgpu_bo_set_metadata(struct amdgpu_bo *bo, void *metadata,
+			   u32 metadata_size, uint64_t flags)
 {
 	struct amdgpu_bo_user *ubo;
 	void *buffer;
@@ -1197,7 +1247,7 @@ int amdgpu_bo_get_metadata(struct amdgpu_bo *bo, void *buffer,
  * amdgpu_bo_move_notify - notification about a memory move
  * @bo: pointer to a buffer object
  * @evict: if this move is evicting the buffer from the graphics address space
- * @new_mem: new information of the bufer object
+ * @new_mem: new resource for backing the BO
  *
  * Marks the corresponding &amdgpu_bo buffer object as invalid, also performs
  * bookkeeping.
@@ -1207,52 +1257,24 @@ void amdgpu_bo_move_notify(struct ttm_buffer_object *bo,
 			   bool evict,
 			   struct ttm_resource *new_mem)
 {
-	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->bdev);
-	struct amdgpu_bo *abo;
 	struct ttm_resource *old_mem = bo->resource;
+	struct amdgpu_bo *abo;
 
 	if (!amdgpu_bo_is_amdgpu_bo(bo))
 		return;
 
 	abo = ttm_to_amdgpu_bo(bo);
-	amdgpu_vm_bo_invalidate(adev, abo, evict);
+	amdgpu_vm_bo_move(abo, new_mem, evict);
 
 	amdgpu_bo_kunmap(abo);
 
-	if (abo->tbo.base.dma_buf && !abo->tbo.base.import_attach &&
-	    bo->resource->mem_type != TTM_PL_SYSTEM)
+	if (abo->tbo.base.dma_buf && !drm_gem_is_imported(&abo->tbo.base) &&
+	    old_mem && old_mem->mem_type != TTM_PL_SYSTEM)
 		dma_buf_move_notify(abo->tbo.base.dma_buf);
 
-	/* remember the eviction */
-	if (evict)
-		atomic64_inc(&adev->num_evictions);
-
-	/* update statistics */
-	if (!new_mem)
-		return;
-
 	/* move_notify is called before move happens */
-	trace_amdgpu_bo_move(abo, new_mem->mem_type, old_mem->mem_type);
-}
-
-void amdgpu_bo_get_memory(struct amdgpu_bo *bo, uint64_t *vram_mem,
-				uint64_t *gtt_mem, uint64_t *cpu_mem)
-{
-	unsigned int domain;
-
-	domain = amdgpu_mem_type_to_domain(bo->tbo.resource->mem_type);
-	switch (domain) {
-	case AMDGPU_GEM_DOMAIN_VRAM:
-		*vram_mem += amdgpu_bo_size(bo);
-		break;
-	case AMDGPU_GEM_DOMAIN_GTT:
-		*gtt_mem += amdgpu_bo_size(bo);
-		break;
-	case AMDGPU_GEM_DOMAIN_CPU:
-	default:
-		*cpu_mem += amdgpu_bo_size(bo);
-		break;
-	}
+	trace_amdgpu_bo_move(abo, new_mem ? new_mem->mem_type : -1,
+			     old_mem ? old_mem->mem_type : -1);
 }
 
 /**
@@ -1264,6 +1286,7 @@ void amdgpu_bo_get_memory(struct amdgpu_bo *bo, uint64_t *vram_mem,
  */
 void amdgpu_bo_release_notify(struct ttm_buffer_object *bo)
 {
+	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->bdev);
 	struct dma_fence *fence = NULL;
 	struct amdgpu_bo *abo;
 	int r;
@@ -1273,28 +1296,42 @@ void amdgpu_bo_release_notify(struct ttm_buffer_object *bo)
 
 	abo = ttm_to_amdgpu_bo(bo);
 
+	WARN_ON(abo->vm_bo);
+
 	if (abo->kfd_bo)
 		amdgpu_amdkfd_release_notify(abo);
 
-	/* We only remove the fence if the resv has individualized. */
-	WARN_ON_ONCE(bo->type == ttm_bo_type_kernel
-			&& bo->base.resv != &bo->base._resv);
-	if (bo->base.resv == &bo->base._resv)
-		amdgpu_amdkfd_remove_fence_on_pt_pd_bos(abo);
-
-	if (bo->resource->mem_type != TTM_PL_VRAM ||
-	    !(abo->flags & AMDGPU_GEM_CREATE_VRAM_WIPE_ON_RELEASE))
+	/*
+	 * We lock the private dma_resv object here and since the BO is about to
+	 * be released nobody else should have a pointer to it.
+	 * So when this locking here fails something is wrong with the reference
+	 * counting.
+	 */
+	if (WARN_ON_ONCE(!dma_resv_trylock(&bo->base._resv)))
 		return;
 
-	dma_resv_lock(bo->base.resv, NULL);
+	amdgpu_amdkfd_remove_all_eviction_fences(abo);
 
-	r = amdgpu_fill_buffer(abo, AMDGPU_POISON, bo->base.resv, &fence);
-	if (!WARN_ON(r)) {
-		amdgpu_bo_fence(abo, fence, false);
-		dma_fence_put(fence);
-	}
+	if (!bo->resource || bo->resource->mem_type != TTM_PL_VRAM ||
+	    !(abo->flags & AMDGPU_GEM_CREATE_VRAM_WIPE_ON_RELEASE) ||
+	    adev->in_suspend || drm_dev_is_unplugged(adev_to_drm(adev)))
+		goto out;
 
-	dma_resv_unlock(bo->base.resv);
+	r = dma_resv_reserve_fences(&bo->base._resv, 1);
+	if (r)
+		goto out;
+
+	r = amdgpu_fill_buffer(abo, 0, &bo->base._resv, &fence, true,
+			       AMDGPU_KERNEL_JOB_ID_CLEAR_ON_RELEASE);
+	if (WARN_ON(r))
+		goto out;
+
+	amdgpu_vram_mgr_set_cleared(bo->resource);
+	dma_resv_add_fence(&bo->base._resv, fence, DMA_RESV_USAGE_KERNEL);
+	dma_fence_put(fence);
+
+out:
+	dma_resv_unlock(&bo->base._resv);
 }
 
 /**
@@ -1313,17 +1350,12 @@ vm_fault_t amdgpu_bo_fault_reserve_notify(struct ttm_buffer_object *bo)
 	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->bdev);
 	struct ttm_operation_ctx ctx = { false, false };
 	struct amdgpu_bo *abo = ttm_to_amdgpu_bo(bo);
-	unsigned long offset;
 	int r;
 
 	/* Remember that this BO was accessed by the CPU */
 	abo->flags |= AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED;
 
-	if (bo->resource->mem_type != TTM_PL_VRAM)
-		return 0;
-
-	offset = bo->resource->start << PAGE_SHIFT;
-	if ((offset + bo->base.size) <= adev->gmc.visible_vram_size)
+	if (amdgpu_res_cpu_visible(adev, bo->resource))
 		return 0;
 
 	/* Can't move a pinned BO to visible VRAM */
@@ -1336,8 +1368,7 @@ vm_fault_t amdgpu_bo_fault_reserve_notify(struct ttm_buffer_object *bo)
 					AMDGPU_GEM_DOMAIN_GTT);
 
 	/* Avoid costly evictions; only set GTT as a busy placement */
-	abo->placement.num_busy_placement = 1;
-	abo->placement.busy_placement = &abo->placements[1];
+	abo->placements[0].flags |= TTM_PL_FLAG_DESIRED;
 
 	r = ttm_bo_validate(bo, &abo->placement, &ctx);
 	if (unlikely(r == -EBUSY || r == -ERESTARTSYS))
@@ -1345,10 +1376,9 @@ vm_fault_t amdgpu_bo_fault_reserve_notify(struct ttm_buffer_object *bo)
 	else if (unlikely(r))
 		return VM_FAULT_SIGBUS;
 
-	offset = bo->resource->start << PAGE_SHIFT;
 	/* this should never happen */
 	if (bo->resource->mem_type == TTM_PL_VRAM &&
-	    (offset + bo->base.size) > adev->gmc.visible_vram_size)
+	    !amdgpu_res_cpu_visible(adev, bo->resource))
 		return VM_FAULT_SIGBUS;
 
 	ttm_bo_move_to_lru_tail_unlocked(bo);
@@ -1367,11 +1397,17 @@ void amdgpu_bo_fence(struct amdgpu_bo *bo, struct dma_fence *fence,
 		     bool shared)
 {
 	struct dma_resv *resv = bo->tbo.base.resv;
+	int r;
 
-	if (shared)
-		dma_resv_add_shared_fence(resv, fence);
-	else
-		dma_resv_add_excl_fence(resv, fence);
+	r = dma_resv_reserve_fences(resv, 1);
+	if (r) {
+		/* As last resort on OOM we block for the fence */
+		dma_fence_wait(fence, false);
+		return;
+	}
+
+	dma_resv_add_fence(resv, fence, shared ? DMA_RESV_USAGE_READ :
+			   DMA_RESV_USAGE_WRITE);
 }
 
 /**
@@ -1443,6 +1479,26 @@ u64 amdgpu_bo_gpu_offset(struct amdgpu_bo *bo)
 }
 
 /**
+ * amdgpu_bo_fb_aper_addr - return FB aperture GPU offset of the VRAM bo
+ * @bo:	amdgpu VRAM buffer object for which we query the offset
+ *
+ * Returns:
+ * current FB aperture GPU offset of the object.
+ */
+u64 amdgpu_bo_fb_aper_addr(struct amdgpu_bo *bo)
+{
+	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
+	uint64_t offset, fb_base;
+
+	WARN_ON_ONCE(bo->tbo.resource->mem_type != TTM_PL_VRAM);
+
+	fb_base = adev->gmc.fb_start;
+	fb_base += adev->gmc.xgmi.physical_node_id * adev->gmc.xgmi.node_segment_size;
+	offset = (bo->tbo.resource->start << PAGE_SHIFT) + fb_base;
+	return amdgpu_gmc_sign_extend(offset);
+}
+
+/**
  * amdgpu_bo_gpu_offset_no_check - return GPU offset of bo
  * @bo:	amdgpu object for which we query the offset
  *
@@ -1452,15 +1508,60 @@ u64 amdgpu_bo_gpu_offset(struct amdgpu_bo *bo)
 u64 amdgpu_bo_gpu_offset_no_check(struct amdgpu_bo *bo)
 {
 	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
-	uint64_t offset;
+	uint64_t offset = AMDGPU_BO_INVALID_OFFSET;
 
-	offset = (bo->tbo.resource->start << PAGE_SHIFT) +
-		 amdgpu_ttm_domain_start(adev, bo->tbo.resource->mem_type);
+	if (bo->tbo.resource->mem_type == TTM_PL_TT)
+		offset = amdgpu_gmc_agp_addr(&bo->tbo);
+
+	if (offset == AMDGPU_BO_INVALID_OFFSET)
+		offset = (bo->tbo.resource->start << PAGE_SHIFT) +
+			amdgpu_ttm_domain_start(adev, bo->tbo.resource->mem_type);
 
 	return amdgpu_gmc_sign_extend(offset);
 }
 
 /**
+ * amdgpu_bo_mem_stats_placement - bo placement for memory accounting
+ * @bo:	the buffer object we should look at
+ *
+ * BO can have multiple preferred placements, to avoid double counting we want
+ * to file it under a single placement for memory stats.
+ * Luckily, if we take the highest set bit in preferred_domains the result is
+ * quite sensible.
+ *
+ * Returns:
+ * Which of the placements should the BO be accounted under.
+ */
+uint32_t amdgpu_bo_mem_stats_placement(struct amdgpu_bo *bo)
+{
+	uint32_t domain = bo->preferred_domains & AMDGPU_GEM_DOMAIN_MASK;
+
+	if (!domain)
+		return TTM_PL_SYSTEM;
+
+	switch (rounddown_pow_of_two(domain)) {
+	case AMDGPU_GEM_DOMAIN_CPU:
+		return TTM_PL_SYSTEM;
+	case AMDGPU_GEM_DOMAIN_GTT:
+		return TTM_PL_TT;
+	case AMDGPU_GEM_DOMAIN_VRAM:
+		return TTM_PL_VRAM;
+	case AMDGPU_GEM_DOMAIN_GDS:
+		return AMDGPU_PL_GDS;
+	case AMDGPU_GEM_DOMAIN_GWS:
+		return AMDGPU_PL_GWS;
+	case AMDGPU_GEM_DOMAIN_OA:
+		return AMDGPU_PL_OA;
+	case AMDGPU_GEM_DOMAIN_DOORBELL:
+		return AMDGPU_PL_DOORBELL;
+	case AMDGPU_GEM_DOMAIN_MMIO_REMAP:
+		return AMDGPU_PL_MMIO_REMAP;
+	default:
+		return TTM_PL_SYSTEM;
+	}
+}
+
+/**
  * amdgpu_bo_get_preferred_domain - get preferred domain
  * @adev: amdgpu device object
  * @domain: allowed :ref:`memory domains <amdgpu_memory_domains>`
@@ -1471,7 +1572,8 @@ u64 amdgpu_bo_gpu_offset_no_check(struct amdgpu_bo *bo)
 uint32_t amdgpu_bo_get_preferred_domain(struct amdgpu_device *adev,
 					    uint32_t domain)
 {
-	if (domain == (AMDGPU_GEM_DOMAIN_VRAM | AMDGPU_GEM_DOMAIN_GTT)) {
+	if ((domain == (AMDGPU_GEM_DOMAIN_VRAM | AMDGPU_GEM_DOMAIN_GTT)) &&
+	    ((adev->asic_type == CHIP_CARRIZO) || (adev->asic_type == CHIP_STONEY))) {
 		domain = AMDGPU_GEM_DOMAIN_VRAM;
 		if (adev->gmc.real_vram_size <= AMDGPU_SG_THRESHOLD)
 			domain = AMDGPU_GEM_DOMAIN_GTT;
@@ -1501,25 +1603,54 @@ uint32_t amdgpu_bo_get_preferred_domain(struct amdgpu_device *adev,
  */
 u64 amdgpu_bo_print_info(int id, struct amdgpu_bo *bo, struct seq_file *m)
 {
+	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
 	struct dma_buf_attachment *attachment;
 	struct dma_buf *dma_buf;
-	unsigned int domain;
 	const char *placement;
 	unsigned int pin_count;
 	u64 size;
 
-	domain = amdgpu_mem_type_to_domain(bo->tbo.resource->mem_type);
-	switch (domain) {
-	case AMDGPU_GEM_DOMAIN_VRAM:
-		placement = "VRAM";
-		break;
-	case AMDGPU_GEM_DOMAIN_GTT:
-		placement = " GTT";
-		break;
-	case AMDGPU_GEM_DOMAIN_CPU:
-	default:
-		placement = " CPU";
-		break;
+	if (dma_resv_trylock(bo->tbo.base.resv)) {
+		if (!bo->tbo.resource) {
+			placement = "NONE";
+		} else {
+			switch (bo->tbo.resource->mem_type) {
+			case TTM_PL_VRAM:
+				if (amdgpu_res_cpu_visible(adev, bo->tbo.resource))
+					placement = "VRAM VISIBLE";
+				else
+					placement = "VRAM";
+				break;
+			case TTM_PL_TT:
+				placement = "GTT";
+				break;
+			case AMDGPU_PL_GDS:
+				placement = "GDS";
+				break;
+			case AMDGPU_PL_GWS:
+				placement = "GWS";
+				break;
+			case AMDGPU_PL_OA:
+				placement = "OA";
+				break;
+			case AMDGPU_PL_PREEMPT:
+				placement = "PREEMPTIBLE";
+				break;
+			case AMDGPU_PL_DOORBELL:
+				placement = "DOORBELL";
+				break;
+			case AMDGPU_PL_MMIO_REMAP:
+				placement = "MMIO REMAP";
+				break;
+			case TTM_PL_SYSTEM:
+			default:
+				placement = "CPU";
+				break;
+			}
+		}
+		dma_resv_unlock(bo->tbo.base.resv);
+	} else {
+		placement = "UNKNOWN";
 	}
 
 	size = amdgpu_bo_size(bo);
@@ -1534,9 +1665,9 @@ u64 amdgpu_bo_print_info(int id, struct amdgpu_bo *bo, struct seq_file *m)
 	attachment = READ_ONCE(bo->tbo.base.import_attach);
 
 	if (attachment)
-		seq_printf(m, " imported from %p", dma_buf);
+		seq_printf(m, " imported from ino:%lu", file_inode(dma_buf->file)->i_ino);
 	else if (dma_buf)
-		seq_printf(m, " exported as %p", dma_buf);
+		seq_printf(m, " exported as ino:%lu", file_inode(dma_buf->file)->i_ino);
 
 	amdgpu_bo_print_flag(m, bo, CPU_ACCESS_REQUIRED);
 	amdgpu_bo_print_flag(m, bo, NO_CPU_ACCESS);
@@ -1545,7 +1676,11 @@ u64 amdgpu_bo_print_info(int id, struct amdgpu_bo *bo, struct seq_file *m)
 	amdgpu_bo_print_flag(m, bo, VRAM_CONTIGUOUS);
 	amdgpu_bo_print_flag(m, bo, VM_ALWAYS_VALID);
 	amdgpu_bo_print_flag(m, bo, EXPLICIT_SYNC);
-
+	/* Add the gem obj resv fence dump*/
+	if (dma_resv_trylock(bo->tbo.base.resv)) {
+		dma_resv_describe(bo->tbo.base.resv, m);
+		dma_resv_unlock(bo->tbo.base.resv);
+	}
 	seq_puts(m, "\n");
 
 	return size;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h
index 4c9cbdc66995..52c2d1731aab 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h
@@ -41,7 +41,6 @@
 
 /* BO flag to indicate a KFD userptr BO */
 #define AMDGPU_AMDKFD_CREATE_USERPTR_BO	(1ULL << 63)
-#define AMDGPU_AMDKFD_CREATE_SVM_BO	(1ULL << 62)
 
 #define to_amdgpu_bo_user(abo) container_of((abo), struct amdgpu_bo_user, bo)
 #define to_amdgpu_bo_vm(abo) container_of((abo), struct amdgpu_bo_vm, bo)
@@ -57,6 +56,8 @@ struct amdgpu_bo_param {
 	bool				no_wait_gpu;
 	struct dma_resv			*resv;
 	void				(*destroy)(struct ttm_buffer_object *bo);
+	/* xcp partition number plus 1, 0 means any partition */
+	int8_t				xcp_id_plus1;
 };
 
 /* bo virtual addresses in a vm */
@@ -68,7 +69,7 @@ struct amdgpu_bo_va_mapping {
 	uint64_t			last;
 	uint64_t			__subtree_last;
 	uint64_t			offset;
-	uint64_t			flags;
+	uint32_t			flags;
 };
 
 /* User space allocated BO in a VM */
@@ -89,6 +90,13 @@ struct amdgpu_bo_va {
 	bool				cleared;
 
 	bool				is_xgmi;
+
+	/*
+	 * protected by vm reservation lock
+	 * if non-zero, cannot unmap from GPU because user queues may still access it
+	 */
+	unsigned int			queue_refcount;
+	atomic_t			userq_va_mapped;
 };
 
 struct amdgpu_bo {
@@ -109,6 +117,13 @@ struct amdgpu_bo {
 	struct mmu_interval_notifier	notifier;
 #endif
 	struct kgd_mem                  *kfd_bo;
+
+	/*
+	 * For GPUs with spatial partitioning, xcp partition number, -1 means
+	 * any partition. For other ASICs without spatial partition, always 0
+	 * for memory accounting.
+	 */
+	int8_t				xcp_id;
 };
 
 struct amdgpu_bo_user {
@@ -122,8 +137,6 @@ struct amdgpu_bo_user {
 
 struct amdgpu_bo_vm {
 	struct amdgpu_bo		bo;
-	struct amdgpu_bo		*shadow;
-	struct list_head		shadow_list;
 	struct amdgpu_vm_bo_base        entries[];
 };
 
@@ -153,6 +166,10 @@ static inline unsigned amdgpu_mem_type_to_domain(u32 mem_type)
 		return AMDGPU_GEM_DOMAIN_GWS;
 	case AMDGPU_PL_OA:
 		return AMDGPU_GEM_DOMAIN_OA;
+	case AMDGPU_PL_DOORBELL:
+		return AMDGPU_GEM_DOMAIN_DOORBELL;
+	case AMDGPU_PL_MMIO_REMAP:
+		return AMDGPU_GEM_DOMAIN_MMIO_REMAP;
 	default:
 		break;
 	}
@@ -214,28 +231,6 @@ static inline u64 amdgpu_bo_mmap_offset(struct amdgpu_bo *bo)
 }
 
 /**
- * amdgpu_bo_in_cpu_visible_vram - check if BO is (partly) in visible VRAM
- */
-static inline bool amdgpu_bo_in_cpu_visible_vram(struct amdgpu_bo *bo)
-{
-	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
-	struct amdgpu_res_cursor cursor;
-
-	if (bo->tbo.resource->mem_type != TTM_PL_VRAM)
-		return false;
-
-	amdgpu_res_first(bo->tbo.resource, 0, amdgpu_bo_size(bo), &cursor);
-	while (cursor.remaining) {
-		if (cursor.start < adev->gmc.visible_vram_size)
-			return true;
-
-		amdgpu_res_next(&cursor, cursor.size);
-	}
-
-	return false;
-}
-
-/**
  * amdgpu_bo_explicit_sync - return whether the bo is explicitly synced
  */
 static inline bool amdgpu_bo_explicit_sync(struct amdgpu_bo *bo)
@@ -254,22 +249,6 @@ static inline bool amdgpu_bo_encrypted(struct amdgpu_bo *bo)
 	return bo->flags & AMDGPU_GEM_CREATE_ENCRYPTED;
 }
 
-/**
- * amdgpu_bo_shadowed - check if the BO is shadowed
- *
- * @bo: BO to be tested.
- *
- * Returns:
- * NULL if not shadowed or else return a BO pointer.
- */
-static inline struct amdgpu_bo *amdgpu_bo_shadowed(struct amdgpu_bo *bo)
-{
-	if (bo->tbo.type == ttm_bo_type_kernel)
-		return to_amdgpu_bo_vm(bo)->shadow;
-
-	return NULL;
-}
-
 bool amdgpu_bo_is_amdgpu_bo(struct ttm_buffer_object *bo);
 void amdgpu_bo_placement_from_domain(struct amdgpu_bo *abo, u32 domain);
 
@@ -284,8 +263,12 @@ int amdgpu_bo_create_kernel(struct amdgpu_device *adev,
 			    unsigned long size, int align,
 			    u32 domain, struct amdgpu_bo **bo_ptr,
 			    u64 *gpu_addr, void **cpu_addr);
+int amdgpu_bo_create_isp_user(struct amdgpu_device *adev,
+			   struct dma_buf *dbuf, u32 domain,
+			   struct amdgpu_bo **bo,
+			   u64 *gpu_addr);
 int amdgpu_bo_create_kernel_at(struct amdgpu_device *adev,
-			       uint64_t offset, uint64_t size, uint32_t domain,
+			       uint64_t offset, uint64_t size,
 			       struct amdgpu_bo **bo_ptr, void **cpu_addr);
 int amdgpu_bo_create_user(struct amdgpu_device *adev,
 			  struct amdgpu_bo_param *bp,
@@ -295,14 +278,13 @@ int amdgpu_bo_create_vm(struct amdgpu_device *adev,
 			struct amdgpu_bo_vm **ubo_ptr);
 void amdgpu_bo_free_kernel(struct amdgpu_bo **bo, u64 *gpu_addr,
 			   void **cpu_addr);
+void amdgpu_bo_free_isp_user(struct amdgpu_bo *bo);
 int amdgpu_bo_kmap(struct amdgpu_bo *bo, void **ptr);
 void *amdgpu_bo_kptr(struct amdgpu_bo *bo);
 void amdgpu_bo_kunmap(struct amdgpu_bo *bo);
 struct amdgpu_bo *amdgpu_bo_ref(struct amdgpu_bo *bo);
 void amdgpu_bo_unref(struct amdgpu_bo **bo);
 int amdgpu_bo_pin(struct amdgpu_bo *bo, u32 domain);
-int amdgpu_bo_pin_restricted(struct amdgpu_bo *bo, u32 domain,
-			     u64 min_offset, u64 max_offset);
 void amdgpu_bo_unpin(struct amdgpu_bo *bo);
 int amdgpu_bo_init(struct amdgpu_device *adev);
 void amdgpu_bo_fini(struct amdgpu_device *adev);
@@ -325,27 +307,31 @@ int amdgpu_bo_sync_wait_resv(struct amdgpu_device *adev, struct dma_resv *resv,
 			     bool intr);
 int amdgpu_bo_sync_wait(struct amdgpu_bo *bo, void *owner, bool intr);
 u64 amdgpu_bo_gpu_offset(struct amdgpu_bo *bo);
+u64 amdgpu_bo_fb_aper_addr(struct amdgpu_bo *bo);
 u64 amdgpu_bo_gpu_offset_no_check(struct amdgpu_bo *bo);
-void amdgpu_bo_get_memory(struct amdgpu_bo *bo, uint64_t *vram_mem,
-				uint64_t *gtt_mem, uint64_t *cpu_mem);
-void amdgpu_bo_add_to_shadow_list(struct amdgpu_bo_vm *vmbo);
-int amdgpu_bo_restore_shadow(struct amdgpu_bo *shadow,
-			     struct dma_fence **fence);
+uint32_t amdgpu_bo_mem_stats_placement(struct amdgpu_bo *bo);
 uint32_t amdgpu_bo_get_preferred_domain(struct amdgpu_device *adev,
 					    uint32_t domain);
 
 /*
  * sub allocation
  */
+static inline struct amdgpu_sa_manager *
+to_amdgpu_sa_manager(struct drm_suballoc_manager *manager)
+{
+	return container_of(manager, struct amdgpu_sa_manager, base);
+}
 
-static inline uint64_t amdgpu_sa_bo_gpu_addr(struct amdgpu_sa_bo *sa_bo)
+static inline uint64_t amdgpu_sa_bo_gpu_addr(struct drm_suballoc *sa_bo)
 {
-	return sa_bo->manager->gpu_addr + sa_bo->soffset;
+	return to_amdgpu_sa_manager(sa_bo->manager)->gpu_addr +
+		drm_suballoc_soffset(sa_bo);
 }
 
-static inline void * amdgpu_sa_bo_cpu_addr(struct amdgpu_sa_bo *sa_bo)
+static inline void *amdgpu_sa_bo_cpu_addr(struct drm_suballoc *sa_bo)
 {
-	return sa_bo->manager->cpu_ptr + sa_bo->soffset;
+	return to_amdgpu_sa_manager(sa_bo->manager)->cpu_ptr +
+		drm_suballoc_soffset(sa_bo);
 }
 
 int amdgpu_sa_bo_manager_init(struct amdgpu_device *adev,
@@ -356,11 +342,10 @@ void amdgpu_sa_bo_manager_fini(struct amdgpu_device *adev,
 int amdgpu_sa_bo_manager_start(struct amdgpu_device *adev,
 				      struct amdgpu_sa_manager *sa_manager);
 int amdgpu_sa_bo_new(struct amdgpu_sa_manager *sa_manager,
-		     struct amdgpu_sa_bo **sa_bo,
-		     unsigned size, unsigned align);
-void amdgpu_sa_bo_free(struct amdgpu_device *adev,
-			      struct amdgpu_sa_bo **sa_bo,
-			      struct dma_fence *fence);
+		     struct drm_suballoc **sa_bo,
+		     unsigned int size);
+void amdgpu_sa_bo_free(struct drm_suballoc **sa_bo,
+		       struct dma_fence *fence);
 #if defined(CONFIG_DEBUG_FS)
 void amdgpu_sa_bo_dump_debug_info(struct amdgpu_sa_manager *sa_manager,
 					 struct seq_file *m);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_pll.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_pll.c
index 4eaec446b49d..675aa138ea11 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_pll.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_pll.c
@@ -69,6 +69,7 @@ static void amdgpu_pll_reduce_ratio(unsigned *nom, unsigned *den,
 /**
  * amdgpu_pll_get_fb_ref_div - feedback and ref divider calculation
  *
+ * @adev: amdgpu_device pointer
  * @nom: nominator
  * @den: denominator
  * @post_div: post divider
@@ -93,7 +94,7 @@ static void amdgpu_pll_get_fb_ref_div(struct amdgpu_device *adev, unsigned int n
 		ref_div_max = min(128 / post_div, ref_div_max);
 
 	/* get matching reference and feedback divider */
-	*ref_div = min(max(DIV_ROUND_CLOSEST(den, post_div), 1u), ref_div_max);
+	*ref_div = clamp(DIV_ROUND_CLOSEST(den, post_div), 1u, ref_div_max);
 	*fb_div = DIV_ROUND_CLOSEST(nom * *ref_div * post_div, den);
 
 	/* limit fb divider to its maximum */
@@ -106,6 +107,7 @@ static void amdgpu_pll_get_fb_ref_div(struct amdgpu_device *adev, unsigned int n
 /**
  * amdgpu_pll_compute - compute PLL paramaters
  *
+ * @adev: amdgpu_device pointer
  * @pll: information about the PLL
  * @freq: requested frequency
  * @dot_clock_p: resulting pixel clock
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_pmu.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_pmu.c
index 82e9ecf84352..6e91ea1de5aa 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_pmu.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_pmu.c
@@ -233,6 +233,10 @@ static void amdgpu_perf_start(struct perf_event *event, int flags)
 	if (WARN_ON_ONCE(!(hwc->state & PERF_HES_STOPPED)))
 		return;
 
+	if ((!pe->adev->df.funcs) ||
+	    (!pe->adev->df.funcs->pmc_start))
+		return;
+
 	WARN_ON_ONCE(!(hwc->state & PERF_HES_UPTODATE));
 	hwc->state = 0;
 
@@ -268,9 +272,12 @@ static void amdgpu_perf_read(struct perf_event *event)
 						  pmu);
 	u64 count, prev;
 
-	do {
-		prev = local64_read(&hwc->prev_count);
+	if ((!pe->adev->df.funcs) ||
+	    (!pe->adev->df.funcs->pmc_get_count))
+		return;
 
+	prev = local64_read(&hwc->prev_count);
+	do {
 		switch (hwc->config_base) {
 		case AMDGPU_PMU_EVENT_CONFIG_TYPE_DF:
 		case AMDGPU_PMU_EVENT_CONFIG_TYPE_XGMI:
@@ -281,7 +288,7 @@ static void amdgpu_perf_read(struct perf_event *event)
 			count = 0;
 			break;
 		}
-	} while (local64_cmpxchg(&hwc->prev_count, prev, count) != prev);
+	} while (!local64_try_cmpxchg(&hwc->prev_count, &prev, count));
 
 	local64_add(count - prev, &event->count);
 }
@@ -297,6 +304,10 @@ static void amdgpu_perf_stop(struct perf_event *event, int flags)
 	if (hwc->state & PERF_HES_UPTODATE)
 		return;
 
+	if ((!pe->adev->df.funcs) ||
+	    (!pe->adev->df.funcs->pmc_stop))
+		return;
+
 	switch (hwc->config_base) {
 	case AMDGPU_PMU_EVENT_CONFIG_TYPE_DF:
 	case AMDGPU_PMU_EVENT_CONFIG_TYPE_XGMI:
@@ -326,6 +337,10 @@ static int amdgpu_perf_add(struct perf_event *event, int flags)
 						  struct amdgpu_pmu_entry,
 						  pmu);
 
+	if ((!pe->adev->df.funcs) ||
+	    (!pe->adev->df.funcs->pmc_start))
+		return -EINVAL;
+
 	switch (pe->pmu_perf_type) {
 	case AMDGPU_PMU_PERF_TYPE_DF:
 		hwc->config_base = AMDGPU_PMU_EVENT_CONFIG_TYPE_DF;
@@ -371,6 +386,9 @@ static void amdgpu_perf_del(struct perf_event *event, int flags)
 	struct amdgpu_pmu_entry *pe = container_of(event->pmu,
 						  struct amdgpu_pmu_entry,
 						  pmu);
+	if ((!pe->adev->df.funcs) ||
+	    (!pe->adev->df.funcs->pmc_stop))
+		return;
 
 	amdgpu_perf_stop(event, PERF_EF_UPDATE);
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_preempt_mgr.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_preempt_mgr.c
index d02c8637f909..34b5e22b44e5 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_preempt_mgr.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_preempt_mgr.c
@@ -25,12 +25,6 @@
 
 #include "amdgpu.h"
 
-static inline struct amdgpu_preempt_mgr *
-to_preempt_mgr(struct ttm_resource_manager *man)
-{
-	return container_of(man, struct amdgpu_preempt_mgr, manager);
-}
-
 /**
  * DOC: mem_info_preempt_used
  *
@@ -45,10 +39,9 @@ static ssize_t mem_info_preempt_used_show(struct device *dev,
 {
 	struct drm_device *ddev = dev_get_drvdata(dev);
 	struct amdgpu_device *adev = drm_to_adev(ddev);
-	struct ttm_resource_manager *man;
+	struct ttm_resource_manager *man = &adev->mman.preempt_mgr;
 
-	man = ttm_manager_type(&adev->mman.bdev, AMDGPU_PL_PREEMPT);
-	return sysfs_emit(buf, "%llu\n", amdgpu_preempt_mgr_usage(man));
+	return sysfs_emit(buf, "%llu\n", ttm_resource_manager_usage(man));
 }
 
 static DEVICE_ATTR_RO(mem_info_preempt_used);
@@ -59,7 +52,7 @@ static DEVICE_ATTR_RO(mem_info_preempt_used);
  * @man: TTM memory type manager
  * @tbo: TTM BO we need this range for
  * @place: placement flags and restrictions
- * @mem: the resulting mem object
+ * @res: TTM memory object
  *
  * Dummy, just count the space used without allocating resources or any limit.
  */
@@ -68,16 +61,12 @@ static int amdgpu_preempt_mgr_new(struct ttm_resource_manager *man,
 				  const struct ttm_place *place,
 				  struct ttm_resource **res)
 {
-	struct amdgpu_preempt_mgr *mgr = to_preempt_mgr(man);
-
 	*res = kzalloc(sizeof(**res), GFP_KERNEL);
 	if (!*res)
 		return -ENOMEM;
 
 	ttm_resource_init(tbo, place, *res);
 	(*res)->start = AMDGPU_BO_INVALID_OFFSET;
-
-	atomic64_add((*res)->num_pages, &mgr->used);
 	return 0;
 }
 
@@ -85,55 +74,20 @@ static int amdgpu_preempt_mgr_new(struct ttm_resource_manager *man,
  * amdgpu_preempt_mgr_del - free ranges
  *
  * @man: TTM memory type manager
- * @mem: TTM memory object
+ * @res: TTM memory object
  *
  * Free the allocated GTT again.
  */
 static void amdgpu_preempt_mgr_del(struct ttm_resource_manager *man,
 				   struct ttm_resource *res)
 {
-	struct amdgpu_preempt_mgr *mgr = to_preempt_mgr(man);
-
-	atomic64_sub(res->num_pages, &mgr->used);
+	ttm_resource_fini(man, res);
 	kfree(res);
 }
 
-/**
- * amdgpu_preempt_mgr_usage - return usage of PREEMPT domain
- *
- * @man: TTM memory type manager
- *
- * Return how many bytes are used in the GTT domain
- */
-uint64_t amdgpu_preempt_mgr_usage(struct ttm_resource_manager *man)
-{
-	struct amdgpu_preempt_mgr *mgr = to_preempt_mgr(man);
-	s64 result = atomic64_read(&mgr->used);
-
-	return (result > 0 ? result : 0) * PAGE_SIZE;
-}
-
-/**
- * amdgpu_preempt_mgr_debug - dump VRAM table
- *
- * @man: TTM memory type manager
- * @printer: DRM printer to use
- *
- * Dump the table content using printk.
- */
-static void amdgpu_preempt_mgr_debug(struct ttm_resource_manager *man,
-				     struct drm_printer *printer)
-{
-	struct amdgpu_preempt_mgr *mgr = to_preempt_mgr(man);
-
-	drm_printf(printer, "man size:%llu pages, preempt used:%lld pages\n",
-		   man->size, (u64)atomic64_read(&mgr->used));
-}
-
 static const struct ttm_resource_manager_func amdgpu_preempt_mgr_func = {
 	.alloc = amdgpu_preempt_mgr_new,
 	.free = amdgpu_preempt_mgr_del,
-	.debug = amdgpu_preempt_mgr_debug
 };
 
 /**
@@ -145,16 +99,13 @@ static const struct ttm_resource_manager_func amdgpu_preempt_mgr_func = {
  */
 int amdgpu_preempt_mgr_init(struct amdgpu_device *adev)
 {
-	struct amdgpu_preempt_mgr *mgr = &adev->mman.preempt_mgr;
-	struct ttm_resource_manager *man = &mgr->manager;
+	struct ttm_resource_manager *man = &adev->mman.preempt_mgr;
 	int ret;
 
 	man->use_tt = true;
 	man->func = &amdgpu_preempt_mgr_func;
 
-	ttm_resource_manager_init(man, (1 << 30));
-
-	atomic64_set(&mgr->used, 0);
+	ttm_resource_manager_init(man, &adev->mman.bdev, (1 << 30));
 
 	ret = device_create_file(adev->dev, &dev_attr_mem_info_preempt_used);
 	if (ret) {
@@ -162,8 +113,7 @@ int amdgpu_preempt_mgr_init(struct amdgpu_device *adev)
 		return ret;
 	}
 
-	ttm_set_driver_manager(&adev->mman.bdev, AMDGPU_PL_PREEMPT,
-			       &mgr->manager);
+	ttm_set_driver_manager(&adev->mman.bdev, AMDGPU_PL_PREEMPT, man);
 	ttm_resource_manager_set_used(man, true);
 	return 0;
 }
@@ -178,8 +128,7 @@ int amdgpu_preempt_mgr_init(struct amdgpu_device *adev)
  */
 void amdgpu_preempt_mgr_fini(struct amdgpu_device *adev)
 {
-	struct amdgpu_preempt_mgr *mgr = &adev->mman.preempt_mgr;
-	struct ttm_resource_manager *man = &mgr->manager;
+	struct ttm_resource_manager *man = &adev->mman.preempt_mgr;
 	int ret;
 
 	ttm_resource_manager_set_used(man, false);
@@ -188,7 +137,8 @@ void amdgpu_preempt_mgr_fini(struct amdgpu_device *adev)
 	if (ret)
 		return;
 
-	device_remove_file(adev->dev, &dev_attr_mem_info_preempt_used);
+	if (adev->dev->kobj.sd)
+		device_remove_file(adev->dev, &dev_attr_mem_info_preempt_used);
 
 	ttm_resource_manager_cleanup(man);
 	ttm_set_driver_manager(&adev->mman.bdev, AMDGPU_PL_PREEMPT, NULL);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
index c641f84649d6..0b10497d487c 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
@@ -37,20 +37,46 @@
 #include "psp_v11_0_8.h"
 #include "psp_v12_0.h"
 #include "psp_v13_0.h"
+#include "psp_v13_0_4.h"
+#include "psp_v14_0.h"
 
 #include "amdgpu_ras.h"
 #include "amdgpu_securedisplay.h"
 #include "amdgpu_atomfirmware.h"
 
-static int psp_sysfs_init(struct amdgpu_device *adev);
-static void psp_sysfs_fini(struct amdgpu_device *adev);
+#define AMD_VBIOS_FILE_MAX_SIZE_B      (1024*1024*16)
 
 static int psp_load_smu_fw(struct psp_context *psp);
-static int psp_ta_unload(struct psp_context *psp, struct ta_context *context);
-static int psp_ta_load(struct psp_context *psp, struct ta_context *context);
 static int psp_rap_terminate(struct psp_context *psp);
 static int psp_securedisplay_terminate(struct psp_context *psp);
 
+static int psp_ring_init(struct psp_context *psp,
+			 enum psp_ring_type ring_type)
+{
+	int ret = 0;
+	struct psp_ring *ring;
+	struct amdgpu_device *adev = psp->adev;
+
+	ring = &psp->km_ring;
+
+	ring->ring_type = ring_type;
+
+	/* allocate 4k Page of Local Frame Buffer memory for ring */
+	ring->ring_size = 0x1000;
+	ret = amdgpu_bo_create_kernel(adev, ring->ring_size, PAGE_SIZE,
+				      AMDGPU_GEM_DOMAIN_VRAM |
+				      AMDGPU_GEM_DOMAIN_GTT,
+				      &adev->firmware.rbuf,
+				      &ring->ring_mem_mc_addr,
+				      (void **)&ring->ring_mem);
+	if (ret) {
+		ring->ring_size = 0;
+		return ret;
+	}
+
+	return 0;
+}
+
 /*
  * Due to DF Cstate management centralized to PMFW, the firmware
  * loading sequence will be updated as below:
@@ -75,7 +101,7 @@ static void psp_check_pmfw_centralized_cstate_management(struct psp_context *psp
 		return;
 	}
 
-	switch (adev->ip_versions[MP0_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, MP0_HWIP, 0)) {
 	case IP_VERSION(11, 0, 0):
 	case IP_VERSION(11, 0, 4):
 	case IP_VERSION(11, 0, 5):
@@ -84,7 +110,9 @@ static void psp_check_pmfw_centralized_cstate_management(struct psp_context *psp
 	case IP_VERSION(11, 0, 11):
 	case IP_VERSION(11, 0, 12):
 	case IP_VERSION(11, 0, 13):
+	case IP_VERSION(13, 0, 0):
 	case IP_VERSION(13, 0, 2):
+	case IP_VERSION(13, 0, 7):
 		psp->pmfw_centralized_cstate_management = true;
 		break;
 	default:
@@ -93,54 +121,143 @@ static void psp_check_pmfw_centralized_cstate_management(struct psp_context *psp
 	}
 }
 
-static int psp_early_init(void *handle)
+static int psp_init_sriov_microcode(struct psp_context *psp)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = psp->adev;
+	char ucode_prefix[30];
+	int ret = 0;
+
+	amdgpu_ucode_ip_version_decode(adev, MP0_HWIP, ucode_prefix, sizeof(ucode_prefix));
+
+	switch (amdgpu_ip_version(adev, MP0_HWIP, 0)) {
+	case IP_VERSION(9, 0, 0):
+	case IP_VERSION(11, 0, 7):
+	case IP_VERSION(11, 0, 9):
+		adev->virt.autoload_ucode_id = AMDGPU_UCODE_ID_CP_MEC2;
+		ret = psp_init_cap_microcode(psp, ucode_prefix);
+		break;
+	case IP_VERSION(13, 0, 2):
+		adev->virt.autoload_ucode_id = AMDGPU_UCODE_ID_CP_MEC2;
+		ret = psp_init_cap_microcode(psp, ucode_prefix);
+		ret &= psp_init_ta_microcode(psp, ucode_prefix);
+		break;
+	case IP_VERSION(13, 0, 0):
+		adev->virt.autoload_ucode_id = 0;
+		break;
+	case IP_VERSION(13, 0, 6):
+	case IP_VERSION(13, 0, 14):
+		ret = psp_init_cap_microcode(psp, ucode_prefix);
+		ret &= psp_init_ta_microcode(psp, ucode_prefix);
+		break;
+	case IP_VERSION(13, 0, 10):
+		adev->virt.autoload_ucode_id = AMDGPU_UCODE_ID_CP_MES1_DATA;
+		ret = psp_init_cap_microcode(psp, ucode_prefix);
+		break;
+	case IP_VERSION(13, 0, 12):
+		ret = psp_init_ta_microcode(psp, ucode_prefix);
+		break;
+	default:
+		return -EINVAL;
+	}
+	return ret;
+}
+
+static int psp_early_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
 	struct psp_context *psp = &adev->psp;
 
-	switch (adev->ip_versions[MP0_HWIP][0]) {
+	psp->autoload_supported = true;
+	psp->boot_time_tmr = true;
+
+	switch (amdgpu_ip_version(adev, MP0_HWIP, 0)) {
 	case IP_VERSION(9, 0, 0):
 		psp_v3_1_set_psp_funcs(psp);
 		psp->autoload_supported = false;
+		psp->boot_time_tmr = false;
 		break;
 	case IP_VERSION(10, 0, 0):
 	case IP_VERSION(10, 0, 1):
 		psp_v10_0_set_psp_funcs(psp);
 		psp->autoload_supported = false;
+		psp->boot_time_tmr = false;
 		break;
 	case IP_VERSION(11, 0, 2):
 	case IP_VERSION(11, 0, 4):
 		psp_v11_0_set_psp_funcs(psp);
 		psp->autoload_supported = false;
+		psp->boot_time_tmr = false;
 		break;
 	case IP_VERSION(11, 0, 0):
+	case IP_VERSION(11, 0, 7):
+		adev->psp.sup_pd_fw_up = !amdgpu_sriov_vf(adev);
+		fallthrough;
 	case IP_VERSION(11, 0, 5):
 	case IP_VERSION(11, 0, 9):
-	case IP_VERSION(11, 0, 7):
 	case IP_VERSION(11, 0, 11):
 	case IP_VERSION(11, 5, 0):
+	case IP_VERSION(11, 5, 2):
 	case IP_VERSION(11, 0, 12):
 	case IP_VERSION(11, 0, 13):
 		psp_v11_0_set_psp_funcs(psp);
-		psp->autoload_supported = true;
+		psp->boot_time_tmr = false;
 		break;
 	case IP_VERSION(11, 0, 3):
 	case IP_VERSION(12, 0, 1):
 		psp_v12_0_set_psp_funcs(psp);
+		psp->autoload_supported = false;
+		psp->boot_time_tmr = false;
 		break;
 	case IP_VERSION(13, 0, 2):
+		psp->boot_time_tmr = false;
+		fallthrough;
+	case IP_VERSION(13, 0, 6):
+	case IP_VERSION(13, 0, 14):
+		psp_v13_0_set_psp_funcs(psp);
+		psp->autoload_supported = false;
+		break;
+	case IP_VERSION(13, 0, 12):
 		psp_v13_0_set_psp_funcs(psp);
+		psp->autoload_supported = false;
+		adev->psp.sup_ifwi_up = !amdgpu_sriov_vf(adev);
 		break;
 	case IP_VERSION(13, 0, 1):
 	case IP_VERSION(13, 0, 3):
+	case IP_VERSION(13, 0, 5):
+	case IP_VERSION(13, 0, 8):
+	case IP_VERSION(13, 0, 11):
+	case IP_VERSION(14, 0, 0):
+	case IP_VERSION(14, 0, 1):
+	case IP_VERSION(14, 0, 4):
 		psp_v13_0_set_psp_funcs(psp);
-		psp->autoload_supported = true;
+		psp->boot_time_tmr = false;
 		break;
 	case IP_VERSION(11, 0, 8):
 		if (adev->apu_flags & AMD_APU_IS_CYAN_SKILLFISH2) {
 			psp_v11_0_8_set_psp_funcs(psp);
-			psp->autoload_supported = false;
 		}
+		psp->autoload_supported = false;
+		psp->boot_time_tmr = false;
+		break;
+	case IP_VERSION(13, 0, 0):
+	case IP_VERSION(13, 0, 7):
+	case IP_VERSION(13, 0, 10):
+		psp_v13_0_set_psp_funcs(psp);
+		adev->psp.sup_ifwi_up = !amdgpu_sriov_vf(adev);
+		psp->boot_time_tmr = false;
+		break;
+	case IP_VERSION(13, 0, 4):
+		psp_v13_0_4_set_psp_funcs(psp);
+		psp->boot_time_tmr = false;
+		break;
+	case IP_VERSION(14, 0, 2):
+	case IP_VERSION(14, 0, 3):
+		adev->psp.sup_ifwi_up = !amdgpu_sriov_vf(adev);
+		psp_v14_0_set_psp_funcs(psp);
+		break;
+	case IP_VERSION(14, 0, 5):
+		psp_v14_0_set_psp_funcs(psp);
+		psp->boot_time_tmr = false;
 		break;
 	default:
 		return -EINVAL;
@@ -148,9 +265,52 @@ static int psp_early_init(void *handle)
 
 	psp->adev = adev;
 
+	adev->psp_timeout = 20000;
+
 	psp_check_pmfw_centralized_cstate_management(psp);
 
-	return 0;
+	if (amdgpu_sriov_vf(adev))
+		return psp_init_sriov_microcode(psp);
+	else
+		return psp_init_microcode(psp);
+}
+
+void psp_ta_free_shared_buf(struct ta_mem_context *mem_ctx)
+{
+	amdgpu_bo_free_kernel(&mem_ctx->shared_bo, &mem_ctx->shared_mc_addr,
+			      &mem_ctx->shared_buf);
+	mem_ctx->shared_bo = NULL;
+}
+
+static void psp_free_shared_bufs(struct psp_context *psp)
+{
+	void *tmr_buf;
+	void **pptr;
+
+	/* free TMR memory buffer */
+	pptr = amdgpu_sriov_vf(psp->adev) ? &tmr_buf : NULL;
+	amdgpu_bo_free_kernel(&psp->tmr_bo, &psp->tmr_mc_addr, pptr);
+	psp->tmr_bo = NULL;
+
+	/* free xgmi shared memory */
+	psp_ta_free_shared_buf(&psp->xgmi_context.context.mem_context);
+
+	/* free ras shared memory */
+	psp_ta_free_shared_buf(&psp->ras_context.context.mem_context);
+
+	/* free hdcp shared memory */
+	psp_ta_free_shared_buf(&psp->hdcp_context.context.mem_context);
+
+	/* free dtm shared memory */
+	psp_ta_free_shared_buf(&psp->dtm_context.context.mem_context);
+
+	/* free rap shared memory */
+	psp_ta_free_shared_buf(&psp->rap_context.context.mem_context);
+
+	/* free securedisplay shared memory */
+	psp_ta_free_shared_buf(&psp->securedisplay_context.context.mem_context);
+
+
 }
 
 static void psp_memory_training_fini(struct psp_context *psp)
@@ -168,21 +328,22 @@ static int psp_memory_training_init(struct psp_context *psp)
 	struct psp_memory_training_context *ctx = &psp->mem_train_ctx;
 
 	if (ctx->init != PSP_MEM_TRAIN_RESERVE_SUCCESS) {
-		DRM_DEBUG("memory training is not supported!\n");
+		dev_dbg(psp->adev->dev, "memory training is not supported!\n");
 		return 0;
 	}
 
 	ctx->sys_cache = kzalloc(ctx->train_data_size, GFP_KERNEL);
 	if (ctx->sys_cache == NULL) {
-		DRM_ERROR("alloc mem_train_ctx.sys_cache failed!\n");
+		dev_err(psp->adev->dev, "alloc mem_train_ctx.sys_cache failed!\n");
 		ret = -ENOMEM;
 		goto Err_out;
 	}
 
-	DRM_DEBUG("train_data_size:%llx,p2c_train_data_offset:%llx,c2p_train_data_offset:%llx.\n",
-		  ctx->train_data_size,
-		  ctx->p2c_train_data_offset,
-		  ctx->c2p_train_data_offset);
+	dev_dbg(psp->adev->dev,
+		"train_data_size:%llx,p2c_train_data_offset:%llx,c2p_train_data_offset:%llx.\n",
+		ctx->train_data_size,
+		ctx->p2c_train_data_offset,
+		ctx->c2p_train_data_offset);
 	ctx->init = PSP_MEM_TRAIN_INIT_SUCCESS;
 	return 0;
 
@@ -211,6 +372,11 @@ static bool psp_get_runtime_db_entry(struct amdgpu_device *adev,
 	bool ret = false;
 	int i;
 
+	if (amdgpu_ip_version(adev, MP0_HWIP, 0) == IP_VERSION(13, 0, 6) ||
+	    amdgpu_ip_version(adev, MP0_HWIP, 0) == IP_VERSION(13, 0, 12) ||
+	    amdgpu_ip_version(adev, MP0_HWIP, 0) == IP_VERSION(13, 0, 14))
+		return false;
+
 	db_header_pos = adev->gmc.mc_vram_size - PSP_RUNTIME_DB_OFFSET;
 	db_dir_pos = db_header_pos + sizeof(struct psp_runtime_data_header);
 
@@ -220,7 +386,7 @@ static bool psp_get_runtime_db_entry(struct amdgpu_device *adev,
 
 	if (db_header.cookie != PSP_RUNTIME_DB_COOKIE_ID) {
 		/* runtime db doesn't exist, exit */
-		dev_warn(adev->dev, "PSP runtime database doesn't exist\n");
+		dev_dbg(adev->dev, "PSP runtime database doesn't exist\n");
 		return false;
 	}
 
@@ -241,7 +407,7 @@ static bool psp_get_runtime_db_entry(struct amdgpu_device *adev,
 			case PSP_RUNTIME_ENTRY_TYPE_BOOT_CONFIG:
 				if (db_dir.entry_list[i].size < sizeof(struct psp_runtime_boot_cfg_entry)) {
 					/* invalid db entry size */
-					dev_warn(adev->dev, "Invalid PSP runtime database entry size\n");
+					dev_warn(adev->dev, "Invalid PSP runtime database boot cfg entry size\n");
 					return false;
 				}
 				/* read runtime database entry */
@@ -249,6 +415,17 @@ static bool psp_get_runtime_db_entry(struct amdgpu_device *adev,
 							  (uint32_t *)db_entry, sizeof(struct psp_runtime_boot_cfg_entry), false);
 				ret = true;
 				break;
+			case PSP_RUNTIME_ENTRY_TYPE_PPTABLE_ERR_STATUS:
+				if (db_dir.entry_list[i].size < sizeof(struct psp_runtime_scpm_entry)) {
+					/* invalid db entry size */
+					dev_warn(adev->dev, "Invalid PSP runtime database scpm entry size\n");
+					return false;
+				}
+				/* read runtime database entry */
+				amdgpu_device_vram_access(adev, db_header_pos + db_dir.entry_list[i].offset,
+							  (uint32_t *)db_entry, sizeof(struct psp_runtime_scpm_entry), false);
+				ret = true;
+				break;
 			default:
 				ret = false;
 				break;
@@ -259,35 +436,39 @@ static bool psp_get_runtime_db_entry(struct amdgpu_device *adev,
 	return ret;
 }
 
-static int psp_sw_init(void *handle)
+static int psp_sw_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	struct psp_context *psp = &adev->psp;
 	int ret;
 	struct psp_runtime_boot_cfg_entry boot_cfg_entry;
 	struct psp_memory_training_context *mem_training_ctx = &psp->mem_train_ctx;
+	struct psp_runtime_scpm_entry scpm_entry;
 
 	psp->cmd = kzalloc(sizeof(struct psp_gfx_cmd_resp), GFP_KERNEL);
 	if (!psp->cmd) {
-		DRM_ERROR("Failed to allocate memory to command buffer!\n");
-		ret = -ENOMEM;
+		dev_err(adev->dev, "Failed to allocate memory to command buffer!\n");
+		return -ENOMEM;
 	}
 
-	if (!amdgpu_sriov_vf(adev)) {
-		ret = psp_init_microcode(psp);
-		if (ret) {
-			DRM_ERROR("Failed to load psp firmware!\n");
-			return ret;
-		}
-	} else if (amdgpu_sriov_vf(adev) &&
-		   adev->ip_versions[MP0_HWIP][0] == IP_VERSION(13, 0, 2)) {
-		ret = psp_init_ta_microcode(psp, "aldebaran");
-		if (ret) {
-			DRM_ERROR("Failed to initialize ta microcode!\n");
-			return ret;
-		}
+	adev->psp.xgmi_context.supports_extended_data =
+		!adev->gmc.xgmi.connected_to_cpu &&
+		amdgpu_ip_version(adev, MP0_HWIP, 0) == IP_VERSION(13, 0, 2);
+
+	memset(&scpm_entry, 0, sizeof(scpm_entry));
+	if ((psp_get_runtime_db_entry(adev,
+				PSP_RUNTIME_ENTRY_TYPE_PPTABLE_ERR_STATUS,
+				&scpm_entry)) &&
+	    (scpm_entry.scpm_status != SCPM_DISABLE)) {
+		adev->scpm_enabled = true;
+		adev->scpm_status = scpm_entry.scpm_status;
+	} else {
+		adev->scpm_enabled = false;
+		adev->scpm_status = SCPM_DISABLE;
 	}
 
+	/* TODO: stop gpu driver services and print alarm if scpm is enabled with error status */
+
 	memset(&boot_cfg_entry, 0, sizeof(boot_cfg_entry));
 	if (psp_get_runtime_db_entry(adev,
 				PSP_RUNTIME_ENTRY_TYPE_BOOT_CONFIG,
@@ -298,75 +479,107 @@ static int psp_sw_init(void *handle)
 			/* If psp runtime database exists, then
 			 * only enable two stage memory training
 			 * when TWO_STAGE_DRAM_TRAINING bit is set
-			 * in runtime database */
+			 * in runtime database
+			 */
 			mem_training_ctx->enable_mem_training = true;
 		}
 
 	} else {
-		/* If psp runtime database doesn't exist or
-		 * is invalid, force enable two stage memory
-		 * training */
+		/* If psp runtime database doesn't exist or is
+		 * invalid, force enable two stage memory training
+		 */
 		mem_training_ctx->enable_mem_training = true;
 	}
 
 	if (mem_training_ctx->enable_mem_training) {
 		ret = psp_memory_training_init(psp);
 		if (ret) {
-			DRM_ERROR("Failed to initialize memory training!\n");
+			dev_err(adev->dev, "Failed to initialize memory training!\n");
 			return ret;
 		}
 
 		ret = psp_mem_training(psp, PSP_MEM_TRAIN_COLD_BOOT);
 		if (ret) {
-			DRM_ERROR("Failed to process memory training!\n");
+			dev_err(adev->dev, "Failed to process memory training!\n");
 			return ret;
 		}
 	}
 
-	if (adev->ip_versions[MP0_HWIP][0] == IP_VERSION(11, 0, 0) ||
-	    adev->ip_versions[MP0_HWIP][0] == IP_VERSION(11, 0, 7)) {
-		ret= psp_sysfs_init(adev);
-		if (ret) {
-			return ret;
-		}
-	}
+	ret = amdgpu_bo_create_kernel(adev, PSP_1_MEG, PSP_1_MEG,
+				      (amdgpu_sriov_vf(adev) || adev->debug_use_vram_fw_buf) ?
+				      AMDGPU_GEM_DOMAIN_VRAM : AMDGPU_GEM_DOMAIN_GTT,
+				      &psp->fw_pri_bo,
+				      &psp->fw_pri_mc_addr,
+				      &psp->fw_pri_buf);
+	if (ret)
+		return ret;
+
+	ret = amdgpu_bo_create_kernel(adev, PSP_FENCE_BUFFER_SIZE, PAGE_SIZE,
+				      AMDGPU_GEM_DOMAIN_VRAM |
+				      AMDGPU_GEM_DOMAIN_GTT,
+				      &psp->fence_buf_bo,
+				      &psp->fence_buf_mc_addr,
+				      &psp->fence_buf);
+	if (ret)
+		goto failed1;
+
+	ret = amdgpu_bo_create_kernel(adev, PSP_CMD_BUFFER_SIZE, PAGE_SIZE,
+				      AMDGPU_GEM_DOMAIN_VRAM |
+				      AMDGPU_GEM_DOMAIN_GTT,
+				      &psp->cmd_buf_bo, &psp->cmd_buf_mc_addr,
+				      (void **)&psp->cmd_buf_mem);
+	if (ret)
+		goto failed2;
 
 	return 0;
+
+failed2:
+	amdgpu_bo_free_kernel(&psp->fence_buf_bo,
+			      &psp->fence_buf_mc_addr, &psp->fence_buf);
+failed1:
+	amdgpu_bo_free_kernel(&psp->fw_pri_bo,
+			      &psp->fw_pri_mc_addr, &psp->fw_pri_buf);
+	return ret;
 }
 
-static int psp_sw_fini(void *handle)
+static int psp_sw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	struct psp_context *psp = &adev->psp;
-	struct psp_gfx_cmd_resp *cmd = psp->cmd;
 
 	psp_memory_training_fini(psp);
-	if (psp->sos_fw) {
-		release_firmware(psp->sos_fw);
-		psp->sos_fw = NULL;
-	}
-	if (psp->asd_fw) {
-		release_firmware(psp->asd_fw);
-		psp->asd_fw = NULL;
-	}
-	if (psp->ta_fw) {
-		release_firmware(psp->ta_fw);
-		psp->ta_fw = NULL;
-	}
 
-	if (adev->ip_versions[MP0_HWIP][0] == IP_VERSION(11, 0, 0) ||
-	    adev->ip_versions[MP0_HWIP][0] == IP_VERSION(11, 0, 7))
-		psp_sysfs_fini(adev);
+	amdgpu_ucode_release(&psp->sos_fw);
+	amdgpu_ucode_release(&psp->asd_fw);
+	amdgpu_ucode_release(&psp->ta_fw);
+	amdgpu_ucode_release(&psp->cap_fw);
+	amdgpu_ucode_release(&psp->toc_fw);
+
+	kfree(psp->cmd);
+	psp->cmd = NULL;
 
-	kfree(cmd);
-	cmd = NULL;
+	psp_free_shared_bufs(psp);
+
+	if (psp->km_ring.ring_mem)
+		amdgpu_bo_free_kernel(&adev->firmware.rbuf,
+				      &psp->km_ring.ring_mem_mc_addr,
+				      (void **)&psp->km_ring.ring_mem);
+
+	amdgpu_bo_free_kernel(&psp->fw_pri_bo,
+			      &psp->fw_pri_mc_addr, &psp->fw_pri_buf);
+	amdgpu_bo_free_kernel(&psp->fence_buf_bo,
+			      &psp->fence_buf_mc_addr, &psp->fence_buf);
+	amdgpu_bo_free_kernel(&psp->cmd_buf_bo, &psp->cmd_buf_mc_addr,
+			      (void **)&psp->cmd_buf_mem);
 
 	return 0;
 }
 
-int psp_wait_for(struct psp_context *psp, uint32_t reg_index,
-		 uint32_t reg_val, uint32_t mask, bool check_changed)
+int psp_wait_for(struct psp_context *psp, uint32_t reg_index, uint32_t reg_val,
+		 uint32_t mask, uint32_t flags)
 {
+	bool check_changed = flags & PSP_WAITREG_CHANGED;
+	bool verbose = !(flags & PSP_WAITREG_NOVERBOSE);
 	uint32_t val;
 	int i;
 	struct amdgpu_device *adev = psp->adev;
@@ -386,6 +599,31 @@ int psp_wait_for(struct psp_context *psp, uint32_t reg_index,
 		udelay(1);
 	}
 
+	if (verbose)
+		dev_err(adev->dev,
+			"psp reg (0x%x) wait timed out, mask: %x, read: %x exp: %x",
+			reg_index, mask, val, reg_val);
+
+	return -ETIME;
+}
+
+int psp_wait_for_spirom_update(struct psp_context *psp, uint32_t reg_index,
+			       uint32_t reg_val, uint32_t mask, uint32_t msec_timeout)
+{
+	uint32_t val;
+	int i;
+	struct amdgpu_device *adev = psp->adev;
+
+	if (psp->adev->no_hw_access)
+		return 0;
+
+	for (i = 0; i < msec_timeout; i++) {
+		val = RREG32(reg_index);
+		if ((val & mask) == reg_val)
+			return 0;
+		msleep(1);
+	}
+
 	return -ETIME;
 }
 
@@ -422,28 +660,49 @@ static const char *psp_gfx_cmd_name(enum psp_gfx_cmd_id cmd_id)
 		return "AUTOLOAD_RLC";
 	case GFX_CMD_ID_BOOT_CFG:
 		return "BOOT_CFG";
+	case GFX_CMD_ID_CONFIG_SQ_PERFMON:
+		return "CONFIG_SQ_PERFMON";
+	case GFX_CMD_ID_FB_FW_RESERV_ADDR:
+		return "FB_FW_RESERV_ADDR";
+	case GFX_CMD_ID_FB_FW_RESERV_EXT_ADDR:
+		return "FB_FW_RESERV_EXT_ADDR";
+	case GFX_CMD_ID_SRIOV_SPATIAL_PART:
+		return "SPATIAL_PARTITION";
+	case GFX_CMD_ID_FB_NPS_MODE:
+		return "NPS_MODE_CHANGE";
 	default:
 		return "UNKNOWN CMD";
 	}
 }
 
+static bool psp_err_warn(struct psp_context *psp)
+{
+	struct psp_gfx_cmd_resp *cmd = psp->cmd_buf_mem;
+
+	/* This response indicates reg list is already loaded */
+	if (amdgpu_ip_version(psp->adev, MP0_HWIP, 0) == IP_VERSION(13, 0, 2) &&
+	    cmd->cmd_id == GFX_CMD_ID_LOAD_IP_FW &&
+	    cmd->cmd.cmd_load_ip_fw.fw_type == GFX_FW_TYPE_REG_LIST &&
+	    cmd->resp.status == TEE_ERROR_CANCEL)
+		return false;
+
+	return true;
+}
+
 static int
 psp_cmd_submit_buf(struct psp_context *psp,
 		   struct amdgpu_firmware_info *ucode,
 		   struct psp_gfx_cmd_resp *cmd, uint64_t fence_mc_addr)
 {
 	int ret;
-	int index, idx;
-	int timeout = 20000;
+	int index;
+	int timeout = psp->adev->psp_timeout;
 	bool ras_intr = false;
 	bool skip_unsupport = false;
 
 	if (psp->adev->no_hw_access)
 		return 0;
 
-	if (!drm_dev_enter(adev_to_drm(psp->adev), &idx))
-		return 0;
-
 	memset(psp->cmd_buf_mem, 0, PSP_CMD_BUFFER_SIZE);
 
 	memcpy(psp->cmd_buf_mem, cmd, sizeof(struct psp_gfx_cmd_resp));
@@ -475,7 +734,7 @@ psp_cmd_submit_buf(struct psp_context *psp,
 	skip_unsupport = (psp->cmd_buf_mem->resp.status == TEE_ERROR_NOT_SUPPORTED ||
 		psp->cmd_buf_mem->resp.status == PSP_ERR_UNKNOWN_COMMAND) && amdgpu_sriov_vf(psp->adev);
 
-	memcpy((void*)&cmd->resp, (void*)&psp->cmd_buf_mem->resp, sizeof(struct psp_gfx_resp));
+	memcpy(&cmd->resp, &psp->cmd_buf_mem->resp, sizeof(struct psp_gfx_resp));
 
 	/* In some cases, psp response status is not 0 even there is no
 	 * problem while the command is submitted. Some version of PSP FW
@@ -486,12 +745,21 @@ psp_cmd_submit_buf(struct psp_context *psp,
 	 */
 	if (!skip_unsupport && (psp->cmd_buf_mem->resp.status || !timeout) && !ras_intr) {
 		if (ucode)
-			DRM_WARN("failed to load ucode %s(0x%X) ",
-				  amdgpu_ucode_name(ucode->ucode_id), ucode->ucode_id);
-		DRM_WARN("psp gfx command %s(0x%X) failed and response status is (0x%X)\n",
-			 psp_gfx_cmd_name(psp->cmd_buf_mem->cmd_id), psp->cmd_buf_mem->cmd_id,
-			 psp->cmd_buf_mem->resp.status);
-		if (!timeout) {
+			dev_warn(psp->adev->dev,
+				 "failed to load ucode %s(0x%X) ",
+				 amdgpu_ucode_name(ucode->ucode_id), ucode->ucode_id);
+		if (psp_err_warn(psp))
+			dev_warn(
+				psp->adev->dev,
+				"psp gfx command %s(0x%X) failed and response status is (0x%X)\n",
+				psp_gfx_cmd_name(psp->cmd_buf_mem->cmd_id),
+				psp->cmd_buf_mem->cmd_id,
+				psp->cmd_buf_mem->resp.status);
+		/* If any firmware (including CAP) load fails under SRIOV, it should
+		 * return failure to stop the VF from initializing.
+		 * Also return failure in case of timeout
+		 */
+		if ((ucode && amdgpu_sriov_vf(psp->adev)) || !timeout) {
 			ret = -EINVAL;
 			goto exit;
 		}
@@ -503,7 +771,6 @@ psp_cmd_submit_buf(struct psp_context *psp,
 	}
 
 exit:
-	drm_dev_exit(idx);
 	return ret;
 }
 
@@ -518,7 +785,7 @@ static struct psp_gfx_cmd_resp *acquire_psp_cmd_buf(struct psp_context *psp)
 	return cmd;
 }
 
-void release_psp_cmd_buf(struct psp_context *psp)
+static void release_psp_cmd_buf(struct psp_context *psp)
 {
 	mutex_unlock(&psp->mutex);
 }
@@ -528,8 +795,13 @@ static void psp_prep_tmr_cmd_buf(struct psp_context *psp,
 				 uint64_t tmr_mc, struct amdgpu_bo *tmr_bo)
 {
 	struct amdgpu_device *adev = psp->adev;
-	uint32_t size = amdgpu_bo_size(tmr_bo);
-	uint64_t tmr_pa = amdgpu_gmc_vram_pa(adev, tmr_bo);
+	uint32_t size = 0;
+	uint64_t tmr_pa = 0;
+
+	if (tmr_bo) {
+		size = amdgpu_bo_size(tmr_bo);
+		tmr_pa = amdgpu_gmc_vram_pa(adev, tmr_bo);
+	}
 
 	if (amdgpu_sriov_vf(psp->adev))
 		cmd->cmd_id = GFX_CMD_ID_SETUP_VMR;
@@ -577,7 +849,7 @@ static int psp_load_toc(struct psp_context *psp,
 /* Set up Trusted Memory Region */
 static int psp_tmr_init(struct psp_context *psp)
 {
-	int ret;
+	int ret = 0;
 	int tmr_size;
 	void *tmr_buf;
 	void **pptr;
@@ -592,32 +864,43 @@ static int psp_tmr_init(struct psp_context *psp)
 	tmr_size = PSP_TMR_SIZE(psp->adev);
 
 	/* For ASICs support RLC autoload, psp will parse the toc
-	 * and calculate the total size of TMR needed */
+	 * and calculate the total size of TMR needed
+	 */
 	if (!amdgpu_sriov_vf(psp->adev) &&
 	    psp->toc.start_addr &&
 	    psp->toc.size_bytes &&
 	    psp->fw_pri_buf) {
 		ret = psp_load_toc(psp, &tmr_size);
 		if (ret) {
-			DRM_ERROR("Failed to load toc\n");
+			dev_err(psp->adev->dev, "Failed to load toc\n");
 			return ret;
 		}
 	}
 
-	pptr = amdgpu_sriov_vf(psp->adev) ? &tmr_buf : NULL;
-	ret = amdgpu_bo_create_kernel(psp->adev, tmr_size, PSP_TMR_SIZE(psp->adev),
-				      AMDGPU_GEM_DOMAIN_VRAM,
-				      &psp->tmr_bo, &psp->tmr_mc_addr, pptr);
+	if (!psp->tmr_bo && !psp->boot_time_tmr) {
+		pptr = amdgpu_sriov_vf(psp->adev) ? &tmr_buf : NULL;
+		ret = amdgpu_bo_create_kernel(psp->adev, tmr_size,
+					      PSP_TMR_ALIGNMENT,
+					      AMDGPU_GEM_DOMAIN_GTT | AMDGPU_GEM_DOMAIN_VRAM,
+					      &psp->tmr_bo, &psp->tmr_mc_addr,
+					      pptr);
+	}
+	if (amdgpu_virt_xgmi_migrate_enabled(psp->adev) && psp->tmr_bo)
+		psp->tmr_mc_addr = amdgpu_bo_fb_aper_addr(psp->tmr_bo);
 
 	return ret;
 }
 
 static bool psp_skip_tmr(struct psp_context *psp)
 {
-	switch (psp->adev->ip_versions[MP0_HWIP][0]) {
+	switch (amdgpu_ip_version(psp->adev, MP0_HWIP, 0)) {
 	case IP_VERSION(11, 0, 9):
 	case IP_VERSION(11, 0, 7):
 	case IP_VERSION(13, 0, 2):
+	case IP_VERSION(13, 0, 6):
+	case IP_VERSION(13, 0, 10):
+	case IP_VERSION(13, 0, 12):
+	case IP_VERSION(13, 0, 14):
 		return true;
 	default:
 		return false;
@@ -638,8 +921,9 @@ static int psp_tmr_load(struct psp_context *psp)
 	cmd = acquire_psp_cmd_buf(psp);
 
 	psp_prep_tmr_cmd_buf(psp, cmd, psp->tmr_mc_addr, psp->tmr_bo);
-	DRM_INFO("reserve 0x%lx from 0x%llx for PSP TMR\n",
-		 amdgpu_bo_size(psp->tmr_bo), psp->tmr_mc_addr);
+	if (psp->tmr_bo)
+		dev_info(psp->adev->dev, "reserve 0x%lx from 0x%llx for PSP TMR\n",
+			 amdgpu_bo_size(psp->tmr_bo), psp->tmr_mc_addr);
 
 	ret = psp_cmd_submit_buf(psp, NULL, cmd,
 				 psp->fence_buf_mc_addr);
@@ -650,7 +934,7 @@ static int psp_tmr_load(struct psp_context *psp)
 }
 
 static void psp_prep_tmr_unload_cmd_buf(struct psp_context *psp,
-				        struct psp_gfx_cmd_resp *cmd)
+					struct psp_gfx_cmd_resp *cmd)
 {
 	if (amdgpu_sriov_vf(psp->adev))
 		cmd->cmd_id = GFX_CMD_ID_DESTROY_VMR;
@@ -661,10 +945,18 @@ static void psp_prep_tmr_unload_cmd_buf(struct psp_context *psp,
 static int psp_tmr_unload(struct psp_context *psp)
 {
 	int ret;
-	struct psp_gfx_cmd_resp *cmd = acquire_psp_cmd_buf(psp);
+	struct psp_gfx_cmd_resp *cmd;
+
+	/* skip TMR unload for Navi12 and CHIP_SIENNA_CICHLID SRIOV,
+	 * as TMR is not loaded at all
+	 */
+	if (amdgpu_sriov_vf(psp->adev) && psp_skip_tmr(psp))
+		return 0;
+
+	cmd = acquire_psp_cmd_buf(psp);
 
 	psp_prep_tmr_unload_cmd_buf(psp, cmd);
-	DRM_INFO("free PSP TMR buffer\n");
+	dev_dbg(psp->adev->dev, "free PSP TMR buffer\n");
 
 	ret = psp_cmd_submit_buf(psp, NULL, cmd,
 				 psp->fence_buf_mc_addr);
@@ -676,19 +968,7 @@ static int psp_tmr_unload(struct psp_context *psp)
 
 static int psp_tmr_terminate(struct psp_context *psp)
 {
-	int ret;
-	void *tmr_buf;
-	void **pptr;
-
-	ret = psp_tmr_unload(psp);
-	if (ret)
-		return ret;
-
-	/* free TMR memory buffer */
-	pptr = amdgpu_sriov_vf(psp->adev) ? &tmr_buf : NULL;
-	amdgpu_bo_free_kernel(&psp->tmr_bo, &psp->tmr_mc_addr, pptr);
-
-	return 0;
+	return psp_tmr_unload(psp);
 }
 
 int psp_get_fw_attestation_records_addr(struct psp_context *psp,
@@ -720,6 +1000,106 @@ int psp_get_fw_attestation_records_addr(struct psp_context *psp,
 	return ret;
 }
 
+static int psp_get_fw_reservation_info(struct psp_context *psp,
+						   uint32_t cmd_id,
+						   uint64_t *addr,
+						   uint32_t *size)
+{
+	int ret;
+	uint32_t status;
+	struct psp_gfx_cmd_resp *cmd;
+
+	cmd = acquire_psp_cmd_buf(psp);
+
+	cmd->cmd_id = cmd_id;
+
+	ret = psp_cmd_submit_buf(psp, NULL, cmd,
+				 psp->fence_buf_mc_addr);
+	if (ret) {
+		release_psp_cmd_buf(psp);
+		return ret;
+	}
+
+	status = cmd->resp.status;
+	if (status == PSP_ERR_UNKNOWN_COMMAND) {
+		release_psp_cmd_buf(psp);
+		*addr = 0;
+		*size = 0;
+		return 0;
+	}
+
+	*addr = (uint64_t)cmd->resp.uresp.fw_reserve_info.reserve_base_address_hi << 32 |
+		cmd->resp.uresp.fw_reserve_info.reserve_base_address_lo;
+	*size = cmd->resp.uresp.fw_reserve_info.reserve_size;
+
+	release_psp_cmd_buf(psp);
+
+	return 0;
+}
+
+int psp_update_fw_reservation(struct psp_context *psp)
+{
+	int ret;
+	uint64_t reserv_addr, reserv_addr_ext;
+	uint32_t reserv_size, reserv_size_ext, mp0_ip_ver;
+	struct amdgpu_device *adev = psp->adev;
+
+	mp0_ip_ver = amdgpu_ip_version(adev, MP0_HWIP, 0);
+
+	if (amdgpu_sriov_vf(psp->adev))
+		return 0;
+
+	switch (mp0_ip_ver) {
+	case IP_VERSION(14, 0, 2):
+		if (adev->psp.sos.fw_version < 0x3b0e0d)
+			return 0;
+		break;
+
+	case IP_VERSION(14, 0, 3):
+		if (adev->psp.sos.fw_version < 0x3a0e14)
+			return 0;
+		break;
+
+	default:
+		return 0;
+	}
+
+	ret = psp_get_fw_reservation_info(psp, GFX_CMD_ID_FB_FW_RESERV_ADDR, &reserv_addr, &reserv_size);
+	if (ret)
+		return ret;
+	ret = psp_get_fw_reservation_info(psp, GFX_CMD_ID_FB_FW_RESERV_EXT_ADDR, &reserv_addr_ext, &reserv_size_ext);
+	if (ret)
+		return ret;
+
+	if (reserv_addr != adev->gmc.real_vram_size - reserv_size) {
+		dev_warn(adev->dev, "reserve fw region is not valid!\n");
+		return 0;
+	}
+
+	amdgpu_bo_free_kernel(&adev->mman.fw_reserved_memory, NULL, NULL);
+
+	reserv_size = roundup(reserv_size, SZ_1M);
+
+	ret = amdgpu_bo_create_kernel_at(adev, reserv_addr, reserv_size, &adev->mman.fw_reserved_memory, NULL);
+	if (ret) {
+		dev_err(adev->dev, "reserve fw region failed(%d)!\n", ret);
+		amdgpu_bo_free_kernel(&adev->mman.fw_reserved_memory, NULL, NULL);
+		return ret;
+	}
+
+	reserv_size_ext = roundup(reserv_size_ext, SZ_1M);
+
+	ret = amdgpu_bo_create_kernel_at(adev, reserv_addr_ext, reserv_size_ext,
+					 &adev->mman.fw_reserved_memory_extend, NULL);
+	if (ret) {
+		dev_err(adev->dev, "reserve extend fw region failed(%d)!\n", ret);
+		amdgpu_bo_free_kernel(&adev->mman.fw_reserved_memory_extend, NULL, NULL);
+		return ret;
+	}
+
+	return 0;
+}
+
 static int psp_boot_config_get(struct amdgpu_device *adev, uint32_t *boot_cfg)
 {
 	struct psp_context *psp = &adev->psp;
@@ -795,9 +1175,50 @@ static int psp_rl_load(struct amdgpu_device *adev)
 	return ret;
 }
 
-static int psp_asd_load(struct psp_context *psp)
+int psp_memory_partition(struct psp_context *psp, int mode)
 {
-	return psp_ta_load(psp, &psp->asd_context);
+	struct psp_gfx_cmd_resp *cmd;
+	int ret;
+
+	if (amdgpu_sriov_vf(psp->adev))
+		return 0;
+
+	cmd = acquire_psp_cmd_buf(psp);
+
+	cmd->cmd_id = GFX_CMD_ID_FB_NPS_MODE;
+	cmd->cmd.cmd_memory_part.mode = mode;
+
+	dev_info(psp->adev->dev,
+		 "Requesting %d memory partition change through PSP", mode);
+	ret = psp_cmd_submit_buf(psp, NULL, cmd, psp->fence_buf_mc_addr);
+	if (ret)
+		dev_err(psp->adev->dev,
+			"PSP request failed to change to NPS%d mode\n", mode);
+
+	release_psp_cmd_buf(psp);
+
+	return ret;
+}
+
+int psp_spatial_partition(struct psp_context *psp, int mode)
+{
+	struct psp_gfx_cmd_resp *cmd;
+	int ret;
+
+	if (amdgpu_sriov_vf(psp->adev))
+		return 0;
+
+	cmd = acquire_psp_cmd_buf(psp);
+
+	cmd->cmd_id = GFX_CMD_ID_SRIOV_SPATIAL_PART;
+	cmd->cmd.cmd_spatial_part.mode = mode;
+
+	dev_info(psp->adev->dev, "Requesting %d partitions through PSP", mode);
+	ret = psp_cmd_submit_buf(psp, NULL, cmd, psp->fence_buf_mc_addr);
+
+	release_psp_cmd_buf(psp);
+
+	return ret;
 }
 
 static int psp_asd_initialize(struct psp_context *psp)
@@ -811,11 +1232,16 @@ static int psp_asd_initialize(struct psp_context *psp)
 	if (amdgpu_sriov_vf(psp->adev) || !psp->asd_context.bin_desc.size_bytes)
 		return 0;
 
+	/* bypass asd if display hardware is not available */
+	if (!amdgpu_device_has_display_hardware(psp->adev) &&
+	    amdgpu_ip_version(psp->adev, MP0_HWIP, 0) >= IP_VERSION(13, 0, 10))
+		return 0;
+
 	psp->asd_context.mem_context.shared_mc_addr  = 0;
 	psp->asd_context.mem_context.shared_mem_size = PSP_ASD_SHARED_MEM_SIZE;
 	psp->asd_context.ta_load_type                = GFX_CMD_ID_LOAD_ASD;
 
-	ret = psp_asd_load(psp);
+	ret = psp_ta_load(psp, &psp->asd_context);
 	if (!ret)
 		psp->asd_context.initialized = true;
 
@@ -829,7 +1255,7 @@ static void psp_prep_ta_unload_cmd_buf(struct psp_gfx_cmd_resp *cmd,
 	cmd->cmd.cmd_unload_ta.session_id = session_id;
 }
 
-static int psp_ta_unload(struct psp_context *psp, struct ta_context *context)
+int psp_ta_unload(struct psp_context *psp, struct ta_context *context)
 {
 	int ret;
 	struct psp_gfx_cmd_resp *cmd = acquire_psp_cmd_buf(psp);
@@ -838,16 +1264,13 @@ static int psp_ta_unload(struct psp_context *psp, struct ta_context *context)
 
 	ret = psp_cmd_submit_buf(psp, NULL, cmd, psp->fence_buf_mc_addr);
 
+	context->resp_status = cmd->resp.status;
+
 	release_psp_cmd_buf(psp);
 
 	return ret;
 }
 
-static int psp_asd_unload(struct psp_context *psp)
-{
-	return psp_ta_unload(psp, &psp->asd_context);
-}
-
 static int psp_asd_terminate(struct psp_context *psp)
 {
 	int ret;
@@ -858,8 +1281,7 @@ static int psp_asd_terminate(struct psp_context *psp)
 	if (!psp->asd_context.initialized)
 		return 0;
 
-	ret = psp_asd_unload(psp);
-
+	ret = psp_ta_unload(psp, &psp->asd_context);
 	if (!ret)
 		psp->asd_context.initialized = false;
 
@@ -888,7 +1310,7 @@ int psp_reg_program(struct psp_context *psp, enum psp_reg_prog_id reg,
 	psp_prep_reg_prog_cmd_buf(cmd, reg, value);
 	ret = psp_cmd_submit_buf(psp, NULL, cmd, psp->fence_buf_mc_addr);
 	if (ret)
-		DRM_ERROR("PSP failed to program reg id %d", reg);
+		dev_err(psp->adev->dev, "PSP failed to program reg id %d\n", reg);
 
 	release_psp_cmd_buf(psp);
 
@@ -900,7 +1322,7 @@ static void psp_prep_ta_load_cmd_buf(struct psp_gfx_cmd_resp *cmd,
 				     struct ta_context *context)
 {
 	cmd->cmd_id				= context->ta_load_type;
-	cmd->cmd.cmd_load_ta.app_phy_addr_lo 	= lower_32_bits(ta_bin_mc);
+	cmd->cmd.cmd_load_ta.app_phy_addr_lo	= lower_32_bits(ta_bin_mc);
 	cmd->cmd.cmd_load_ta.app_phy_addr_hi	= upper_32_bits(ta_bin_mc);
 	cmd->cmd.cmd_load_ta.app_len		= context->bin_desc.size_bytes;
 
@@ -911,33 +1333,19 @@ static void psp_prep_ta_load_cmd_buf(struct psp_gfx_cmd_resp *cmd,
 	cmd->cmd.cmd_load_ta.cmd_buf_len = context->mem_context.shared_mem_size;
 }
 
-static int psp_ta_init_shared_buf(struct psp_context *psp,
+int psp_ta_init_shared_buf(struct psp_context *psp,
 				  struct ta_mem_context *mem_ctx)
 {
-	int ret;
-
 	/*
-	* Allocate 16k memory aligned to 4k from Frame Buffer (local
-	* physical) for ta to host memory
-	*/
-	ret = amdgpu_bo_create_kernel(psp->adev, mem_ctx->shared_mem_size,
-				      PAGE_SIZE, AMDGPU_GEM_DOMAIN_VRAM,
+	 * Allocate 16k memory aligned to 4k from Frame Buffer (local
+	 * physical) for ta to host memory
+	 */
+	return amdgpu_bo_create_kernel(psp->adev, mem_ctx->shared_mem_size,
+				      PAGE_SIZE, AMDGPU_GEM_DOMAIN_VRAM |
+				      AMDGPU_GEM_DOMAIN_GTT,
 				      &mem_ctx->shared_bo,
 				      &mem_ctx->shared_mc_addr,
 				      &mem_ctx->shared_buf);
-
-	return ret;
-}
-
-static void psp_ta_free_shared_buf(struct ta_mem_context *mem_ctx)
-{
-	amdgpu_bo_free_kernel(&mem_ctx->shared_bo, &mem_ctx->shared_mc_addr,
-			      &mem_ctx->shared_buf);
-}
-
-static int psp_xgmi_init_shared_buf(struct psp_context *psp)
-{
-	return psp_ta_init_shared_buf(psp, &psp->xgmi_context.context.mem_context);
 }
 
 static void psp_prep_ta_invoke_cmd_buf(struct psp_gfx_cmd_resp *cmd,
@@ -949,7 +1357,7 @@ static void psp_prep_ta_invoke_cmd_buf(struct psp_gfx_cmd_resp *cmd,
 	cmd->cmd.cmd_invoke_cmd.ta_cmd_id	= ta_cmd_id;
 }
 
-static int psp_ta_invoke(struct psp_context *psp,
+int psp_ta_invoke(struct psp_context *psp,
 		  uint32_t ta_cmd_id,
 		  struct ta_context *context)
 {
@@ -961,12 +1369,14 @@ static int psp_ta_invoke(struct psp_context *psp,
 	ret = psp_cmd_submit_buf(psp, NULL, cmd,
 				 psp->fence_buf_mc_addr);
 
+	context->resp_status = cmd->resp.status;
+
 	release_psp_cmd_buf(psp);
 
 	return ret;
 }
 
-static int psp_ta_load(struct psp_context *psp, struct ta_context *context)
+int psp_ta_load(struct psp_context *psp, struct ta_context *context)
 {
 	int ret;
 	struct psp_gfx_cmd_resp *cmd;
@@ -976,30 +1386,26 @@ static int psp_ta_load(struct psp_context *psp, struct ta_context *context)
 	psp_copy_fw(psp, context->bin_desc.start_addr,
 		    context->bin_desc.size_bytes);
 
+	if (amdgpu_virt_xgmi_migrate_enabled(psp->adev) &&
+		context->mem_context.shared_bo)
+		context->mem_context.shared_mc_addr =
+			amdgpu_bo_fb_aper_addr(context->mem_context.shared_bo);
+
 	psp_prep_ta_load_cmd_buf(cmd, psp->fw_pri_mc_addr, context);
 
 	ret = psp_cmd_submit_buf(psp, NULL, cmd,
 				 psp->fence_buf_mc_addr);
 
-	if (!ret) {
+	context->resp_status = cmd->resp.status;
+
+	if (!ret)
 		context->session_id = cmd->resp.session_id;
-	}
 
 	release_psp_cmd_buf(psp);
 
 	return ret;
 }
 
-static int psp_xgmi_load(struct psp_context *psp)
-{
-	return psp_ta_load(psp, &psp->xgmi_context.context);
-}
-
-static int psp_xgmi_unload(struct psp_context *psp)
-{
-	return psp_ta_unload(psp, &psp->xgmi_context.context);
-}
-
 int psp_xgmi_invoke(struct psp_context *psp, uint32_t ta_cmd_id)
 {
 	return psp_ta_invoke(psp, ta_cmd_id, &psp->xgmi_context.context);
@@ -1011,24 +1417,19 @@ int psp_xgmi_terminate(struct psp_context *psp)
 	struct amdgpu_device *adev = psp->adev;
 
 	/* XGMI TA unload currently is not supported on Arcturus/Aldebaran A+A */
-	if (adev->ip_versions[MP0_HWIP][0] == IP_VERSION(11, 0, 4) ||
-	    (adev->ip_versions[MP0_HWIP][0] == IP_VERSION(13, 0, 2) &&
+	if (amdgpu_ip_version(adev, MP0_HWIP, 0) == IP_VERSION(11, 0, 4) ||
+	    (amdgpu_ip_version(adev, MP0_HWIP, 0) == IP_VERSION(13, 0, 2) &&
 	     adev->gmc.xgmi.connected_to_cpu))
 		return 0;
 
 	if (!psp->xgmi_context.context.initialized)
 		return 0;
 
-	ret = psp_xgmi_unload(psp);
-	if (ret)
-		return ret;
+	ret = psp_ta_unload(psp, &psp->xgmi_context.context);
 
 	psp->xgmi_context.context.initialized = false;
 
-	/* free xgmi shared memory */
-	psp_ta_free_shared_buf(&psp->xgmi_context.context.mem_context);
-
-	return 0;
+	return ret;
 }
 
 int psp_xgmi_initialize(struct psp_context *psp, bool set_extended_data, bool load_ta)
@@ -1047,14 +1448,14 @@ int psp_xgmi_initialize(struct psp_context *psp, bool set_extended_data, bool lo
 	psp->xgmi_context.context.mem_context.shared_mem_size = PSP_XGMI_SHARED_MEM_SIZE;
 	psp->xgmi_context.context.ta_load_type = GFX_CMD_ID_LOAD_TA;
 
-	if (!psp->xgmi_context.context.initialized) {
-		ret = psp_xgmi_init_shared_buf(psp);
+	if (!psp->xgmi_context.context.mem_context.shared_buf) {
+		ret = psp_ta_init_shared_buf(psp, &psp->xgmi_context.context.mem_context);
 		if (ret)
 			return ret;
 	}
 
 	/* Load XGMI TA */
-	ret = psp_xgmi_load(psp);
+	ret = psp_ta_load(psp, &psp->xgmi_context.context);
 	if (!ret)
 		psp->xgmi_context.context.initialized = true;
 	else
@@ -1068,6 +1469,8 @@ invoke:
 	xgmi_cmd->cmd_id = TA_COMMAND_XGMI__INITIALIZE;
 
 	ret = psp_xgmi_invoke(psp, xgmi_cmd->cmd_id);
+	/* note down the capbility flag for XGMI TA */
+	psp->xgmi_context.xgmi_ta_caps = xgmi_cmd->caps_flag;
 
 	return ret;
 }
@@ -1114,8 +1517,11 @@ int psp_xgmi_get_node_id(struct psp_context *psp, uint64_t *node_id)
 
 static bool psp_xgmi_peer_link_info_supported(struct psp_context *psp)
 {
-	return psp->adev->ip_versions[MP0_HWIP][0] == IP_VERSION(13, 0, 2) &&
-		psp->xgmi_context.context.bin_desc.fw_version >= 0x2000000b;
+	return (amdgpu_ip_version(psp->adev, MP0_HWIP, 0) ==
+			IP_VERSION(13, 0, 2) &&
+		psp->xgmi_context.context.bin_desc.fw_version >= 0x2000000b) ||
+	       amdgpu_ip_version(psp->adev, MP0_HWIP, 0) >=
+		       IP_VERSION(13, 0, 6);
 }
 
 /*
@@ -1133,9 +1539,13 @@ static void psp_xgmi_reflect_topology_info(struct psp_context *psp,
 	uint64_t src_node_id = psp->adev->gmc.xgmi.node_id;
 	uint64_t dst_node_id = node_info.node_id;
 	uint8_t dst_num_hops = node_info.num_hops;
+	uint8_t dst_is_sharing_enabled = node_info.is_sharing_enabled;
 	uint8_t dst_num_links = node_info.num_links;
 
 	hive = amdgpu_get_xgmi_hive(psp->adev);
+	if (WARN_ON(!hive))
+		return;
+
 	list_for_each_entry(mirror_adev, &hive->device_list, gmc.xgmi.head) {
 		struct psp_xgmi_topology_info *mirror_top_info;
 		int j;
@@ -1149,19 +1559,28 @@ static void psp_xgmi_reflect_topology_info(struct psp_context *psp,
 				continue;
 
 			mirror_top_info->nodes[j].num_hops = dst_num_hops;
-			/*
-			 * prevent 0 num_links value re-reflection since reflection
+			mirror_top_info->nodes[j].is_sharing_enabled = dst_is_sharing_enabled;
+			/* prevent 0 num_links value re-reflection since reflection
 			 * criteria is based on num_hops (direct or indirect).
-			 *
 			 */
-			if (dst_num_links)
+			if (dst_num_links) {
 				mirror_top_info->nodes[j].num_links = dst_num_links;
+				/* swap src and dst due to frame of reference */
+				for (int k = 0; k < dst_num_links; k++) {
+					mirror_top_info->nodes[j].port_num[k].src_xgmi_port_num =
+						node_info.port_num[k].dst_xgmi_port_num;
+					mirror_top_info->nodes[j].port_num[k].dst_xgmi_port_num =
+						node_info.port_num[k].src_xgmi_port_num;
+				}
+			}
 
 			break;
 		}
 
 		break;
 	}
+
+	amdgpu_put_xgmi_hive(hive);
 }
 
 int psp_xgmi_get_topology_info(struct psp_context *psp,
@@ -1184,7 +1603,7 @@ int psp_xgmi_get_topology_info(struct psp_context *psp,
 
 	/* Fill in the shared memory with topology information as input */
 	topology_info_input = &xgmi_cmd->xgmi_in_message.get_topology_info;
-	xgmi_cmd->cmd_id = TA_COMMAND_XGMI__GET_GET_TOPOLOGY_INFO;
+	xgmi_cmd->cmd_id = TA_COMMAND_XGMI__GET_TOPOLOGY_INFO;
 	topology_info_input->num_nodes = number_devices;
 
 	for (i = 0; i < topology_info_input->num_nodes; i++) {
@@ -1195,7 +1614,7 @@ int psp_xgmi_get_topology_info(struct psp_context *psp,
 	}
 
 	/* Invoke xgmi ta to get the topology information */
-	ret = psp_xgmi_invoke(psp, TA_COMMAND_XGMI__GET_GET_TOPOLOGY_INFO);
+	ret = psp_xgmi_invoke(psp, TA_COMMAND_XGMI__GET_TOPOLOGY_INFO);
 	if (ret)
 		return ret;
 
@@ -1220,26 +1639,64 @@ int psp_xgmi_get_topology_info(struct psp_context *psp,
 
 	/* Invoke xgmi ta again to get the link information */
 	if (psp_xgmi_peer_link_info_supported(psp)) {
-		struct ta_xgmi_cmd_get_peer_link_info_output *link_info_output;
+		struct ta_xgmi_cmd_get_peer_link_info *link_info_output;
+		struct ta_xgmi_cmd_get_extend_peer_link_info *link_extend_info_output;
+		bool requires_reflection =
+			(psp->xgmi_context.supports_extended_data &&
+			 get_extended_data) ||
+			amdgpu_ip_version(psp->adev, MP0_HWIP, 0) ==
+				IP_VERSION(13, 0, 6) ||
+			amdgpu_ip_version(psp->adev, MP0_HWIP, 0) ==
+				IP_VERSION(13, 0, 14) ||
+			amdgpu_sriov_vf(psp->adev);
+		bool ta_port_num_support = psp->xgmi_context.xgmi_ta_caps & EXTEND_PEER_LINK_INFO_CMD_FLAG ||
+			amdgpu_sriov_xgmi_ta_ext_peer_link_en(psp->adev);
+
+		/* popluate the shared output buffer rather than the cmd input buffer
+		 * with node_ids as the input for GET_PEER_LINKS command execution.
+		 * This is required for GET_PEER_LINKS per xgmi ta implementation.
+		 * The same requirement for GET_EXTEND_PEER_LINKS command.
+		 */
+		if (ta_port_num_support) {
+			link_extend_info_output = &xgmi_cmd->xgmi_out_message.get_extend_link_info;
 
-		xgmi_cmd->cmd_id = TA_COMMAND_XGMI__GET_PEER_LINKS;
+			for (i = 0; i < topology->num_nodes; i++)
+				link_extend_info_output->nodes[i].node_id = topology->nodes[i].node_id;
 
-		ret = psp_xgmi_invoke(psp, TA_COMMAND_XGMI__GET_PEER_LINKS);
+			link_extend_info_output->num_nodes = topology->num_nodes;
+			xgmi_cmd->cmd_id = TA_COMMAND_XGMI__GET_EXTEND_PEER_LINKS;
+		} else {
+			link_info_output = &xgmi_cmd->xgmi_out_message.get_link_info;
 
+			for (i = 0; i < topology->num_nodes; i++)
+				link_info_output->nodes[i].node_id = topology->nodes[i].node_id;
+
+			link_info_output->num_nodes = topology->num_nodes;
+			xgmi_cmd->cmd_id = TA_COMMAND_XGMI__GET_PEER_LINKS;
+		}
+
+		ret = psp_xgmi_invoke(psp, xgmi_cmd->cmd_id);
 		if (ret)
 			return ret;
 
-		link_info_output = &xgmi_cmd->xgmi_out_message.get_link_info;
 		for (i = 0; i < topology->num_nodes; i++) {
+			uint8_t node_num_links = ta_port_num_support ?
+				link_extend_info_output->nodes[i].num_links : link_info_output->nodes[i].num_links;
 			/* accumulate num_links on extended data */
-			topology->nodes[i].num_links = get_extended_data ?
-					topology->nodes[i].num_links +
-							link_info_output->nodes[i].num_links :
-					link_info_output->nodes[i].num_links;
+			if (get_extended_data) {
+				topology->nodes[i].num_links = topology->nodes[i].num_links + node_num_links;
+			} else {
+				topology->nodes[i].num_links = (requires_reflection && topology->nodes[i].num_links) ?
+								topology->nodes[i].num_links : node_num_links;
+			}
+			/* popluate the connected port num info if supported and available */
+			if (ta_port_num_support && topology->nodes[i].num_links) {
+				memcpy(topology->nodes[i].port_num, link_extend_info_output->nodes[i].port_num,
+				       sizeof(struct xgmi_connected_port_num) * TA_XGMI__MAX_PORT_NUM);
+			}
 
 			/* reflect the topology information for bi-directionality */
-			if (psp->xgmi_context.supports_extended_data &&
-					get_extended_data && topology->nodes[i].num_hops)
+			if (requires_reflection && topology->nodes[i].num_hops)
 				psp_xgmi_reflect_topology_info(psp, topology->nodes[i]);
 		}
 	}
@@ -1277,21 +1734,6 @@ int psp_xgmi_set_topology_info(struct psp_context *psp,
 }
 
 // ras begin
-static int psp_ras_init_shared_buf(struct psp_context *psp)
-{
-	return psp_ta_init_shared_buf(psp, &psp->ras_context.context.mem_context);
-}
-
-static int psp_ras_load(struct psp_context *psp)
-{
-	return psp_ta_load(psp, &psp->ras_context.context);
-}
-
-static int psp_ras_unload(struct psp_context *psp)
-{
-	return psp_ta_unload(psp, &psp->ras_context.context);
-}
-
 static void psp_ras_ta_check_status(struct psp_context *psp)
 {
 	struct ta_ras_shared_memory *ras_cmd =
@@ -1300,21 +1742,86 @@ static void psp_ras_ta_check_status(struct psp_context *psp)
 	switch (ras_cmd->ras_status) {
 	case TA_RAS_STATUS__ERROR_UNSUPPORTED_IP:
 		dev_warn(psp->adev->dev,
-				"RAS WARNING: cmd failed due to unsupported ip\n");
+			 "RAS WARNING: cmd failed due to unsupported ip\n");
 		break;
 	case TA_RAS_STATUS__ERROR_UNSUPPORTED_ERROR_INJ:
 		dev_warn(psp->adev->dev,
-				"RAS WARNING: cmd failed due to unsupported error injection\n");
+			 "RAS WARNING: cmd failed due to unsupported error injection\n");
 		break;
 	case TA_RAS_STATUS__SUCCESS:
 		break;
+	case TA_RAS_STATUS__TEE_ERROR_ACCESS_DENIED:
+		if (ras_cmd->cmd_id == TA_RAS_COMMAND__TRIGGER_ERROR)
+			dev_warn(psp->adev->dev,
+				 "RAS WARNING: Inject error to critical region is not allowed\n");
+		break;
 	default:
 		dev_warn(psp->adev->dev,
-				"RAS WARNING: ras status = 0x%X\n", ras_cmd->ras_status);
+			 "RAS WARNING: ras status = 0x%X\n", ras_cmd->ras_status);
 		break;
 	}
 }
 
+static int psp_ras_send_cmd(struct psp_context *psp,
+		enum ras_command cmd_id, void *in, void *out)
+{
+	struct ta_ras_shared_memory *ras_cmd;
+	uint32_t cmd = cmd_id;
+	int ret = 0;
+
+	if (!in)
+		return -EINVAL;
+
+	mutex_lock(&psp->ras_context.mutex);
+	ras_cmd = (struct ta_ras_shared_memory *)psp->ras_context.context.mem_context.shared_buf;
+	memset(ras_cmd, 0, sizeof(struct ta_ras_shared_memory));
+
+	switch (cmd) {
+	case TA_RAS_COMMAND__ENABLE_FEATURES:
+	case TA_RAS_COMMAND__DISABLE_FEATURES:
+		memcpy(&ras_cmd->ras_in_message,
+			in, sizeof(ras_cmd->ras_in_message));
+		break;
+	case TA_RAS_COMMAND__TRIGGER_ERROR:
+		memcpy(&ras_cmd->ras_in_message.trigger_error,
+			in, sizeof(ras_cmd->ras_in_message.trigger_error));
+		break;
+	case TA_RAS_COMMAND__QUERY_ADDRESS:
+		memcpy(&ras_cmd->ras_in_message.address,
+			in, sizeof(ras_cmd->ras_in_message.address));
+		break;
+	default:
+		dev_err(psp->adev->dev, "Invalid ras cmd id: %u\n", cmd);
+		ret = -EINVAL;
+		goto err_out;
+	}
+
+	ras_cmd->cmd_id = cmd;
+	ret = psp_ras_invoke(psp, ras_cmd->cmd_id);
+
+	switch (cmd) {
+	case TA_RAS_COMMAND__TRIGGER_ERROR:
+		if (!ret && out)
+			memcpy(out, &ras_cmd->ras_status, sizeof(ras_cmd->ras_status));
+		break;
+	case TA_RAS_COMMAND__QUERY_ADDRESS:
+		if (ret || ras_cmd->ras_status || psp->cmd_buf_mem->resp.status)
+			ret = -EINVAL;
+		else if (out)
+			memcpy(out,
+				&ras_cmd->ras_out_message.address,
+				sizeof(ras_cmd->ras_out_message.address));
+		break;
+	default:
+		break;
+	}
+
+err_out:
+	mutex_unlock(&psp->ras_context.mutex);
+
+	return ret;
+}
+
 int psp_ras_invoke(struct psp_context *psp, uint32_t ta_cmd_id)
 {
 	struct ta_ras_shared_memory *ras_cmd;
@@ -1333,9 +1840,8 @@ int psp_ras_invoke(struct psp_context *psp, uint32_t ta_cmd_id)
 	if (amdgpu_ras_intr_triggered())
 		return ret;
 
-	if (ras_cmd->if_version > RAS_TA_HOST_IF_VER)
-	{
-		DRM_WARN("RAS: Unsupported Interface");
+	if (ras_cmd->if_version > RAS_TA_HOST_IF_VER) {
+		dev_warn(psp->adev->dev, "RAS: Unsupported Interface\n");
 		return -EINVAL;
 	}
 
@@ -1344,8 +1850,7 @@ int psp_ras_invoke(struct psp_context *psp, uint32_t ta_cmd_id)
 			dev_warn(psp->adev->dev, "ECC switch disabled\n");
 
 			ras_cmd->ras_status = TA_RAS_STATUS__ERROR_RAS_NOT_AVAILABLE;
-		}
-		else if (ras_cmd->ras_out_message.flags.reg_access_failure_flag)
+		} else if (ras_cmd->ras_out_message.flags.reg_access_failure_flag)
 			dev_warn(psp->adev->dev,
 				 "RAS internal register access blocked\n");
 
@@ -1358,30 +1863,22 @@ int psp_ras_invoke(struct psp_context *psp, uint32_t ta_cmd_id)
 int psp_ras_enable_features(struct psp_context *psp,
 		union ta_ras_cmd_input *info, bool enable)
 {
-	struct ta_ras_shared_memory *ras_cmd;
+	enum ras_command cmd_id;
 	int ret;
 
-	if (!psp->ras_context.context.initialized)
+	if (!psp->ras_context.context.initialized || !info)
 		return -EINVAL;
 
-	ras_cmd = (struct ta_ras_shared_memory *)psp->ras_context.context.mem_context.shared_buf;
-	memset(ras_cmd, 0, sizeof(struct ta_ras_shared_memory));
-
-	if (enable)
-		ras_cmd->cmd_id = TA_RAS_COMMAND__ENABLE_FEATURES;
-	else
-		ras_cmd->cmd_id = TA_RAS_COMMAND__DISABLE_FEATURES;
-
-	ras_cmd->ras_in_message = *info;
-
-	ret = psp_ras_invoke(psp, ras_cmd->cmd_id);
+	cmd_id = enable ?
+		TA_RAS_COMMAND__ENABLE_FEATURES : TA_RAS_COMMAND__DISABLE_FEATURES;
+	ret = psp_ras_send_cmd(psp, cmd_id, info, NULL);
 	if (ret)
 		return -EINVAL;
 
 	return 0;
 }
 
-static int psp_ras_terminate(struct psp_context *psp)
+int psp_ras_terminate(struct psp_context *psp)
 {
 	int ret;
 
@@ -1394,19 +1891,16 @@ static int psp_ras_terminate(struct psp_context *psp)
 	if (!psp->ras_context.context.initialized)
 		return 0;
 
-	ret = psp_ras_unload(psp);
-	if (ret)
-		return ret;
+	ret = psp_ta_unload(psp, &psp->ras_context.context);
 
 	psp->ras_context.context.initialized = false;
 
-	/* free ras shared memory */
-	psp_ta_free_shared_buf(&psp->ras_context.context.mem_context);
+	mutex_destroy(&psp->ras_context.mutex);
 
-	return 0;
+	return ret;
 }
 
-static int psp_ras_initialize(struct psp_context *psp)
+int psp_ras_initialize(struct psp_context *psp)
 {
 	int ret;
 	uint32_t boot_cfg = 0xFF;
@@ -1433,35 +1927,47 @@ static int psp_ras_initialize(struct psp_context *psp)
 		if (ret)
 			dev_warn(adev->dev, "PSP get boot config failed\n");
 
-		if (!amdgpu_ras_is_supported(psp->adev, AMDGPU_RAS_BLOCK__UMC)) {
-			if (!boot_cfg) {
-				dev_info(adev->dev, "GECC is disabled\n");
-			} else {
-				/* disable GECC in next boot cycle if ras is
-				 * disabled by module parameter amdgpu_ras_enable
-				 * and/or amdgpu_ras_mask, or boot_config_get call
-				 * is failed
-				 */
-				ret = psp_boot_config_set(adev, 0);
-				if (ret)
-					dev_warn(adev->dev, "PSP set boot config failed\n");
-				else
-					dev_warn(adev->dev, "GECC will be disabled in next boot cycle "
-						 "if set amdgpu_ras_enable and/or amdgpu_ras_mask to 0x0\n");
-			}
+		if (boot_cfg == 1 && !adev->ras_default_ecc_enabled &&
+		    amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__UMC)) {
+			dev_warn(adev->dev, "GECC is currently enabled, which may affect performance\n");
+			dev_warn(adev->dev,
+				"To disable GECC, please reboot the system and load the amdgpu driver with the parameter amdgpu_ras_enable=0\n");
 		} else {
-			if (1 == boot_cfg) {
-				dev_info(adev->dev, "GECC is enabled\n");
+			if ((adev->ras_default_ecc_enabled || amdgpu_ras_enable == 1) &&
+				amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__UMC)) {
+				if (boot_cfg == 1) {
+					dev_info(adev->dev, "GECC is enabled\n");
+				} else {
+					/* enable GECC in next boot cycle if it is disabled
+					 * in boot config, or force enable GECC if failed to
+					 * get boot configuration
+					 */
+					ret = psp_boot_config_set(adev, BOOT_CONFIG_GECC);
+					if (ret)
+						dev_warn(adev->dev, "PSP set boot config failed\n");
+					else
+						dev_warn(adev->dev, "GECC will be enabled in next boot cycle\n");
+				}
 			} else {
-				/* enable GECC in next boot cycle if it is disabled
-				 * in boot config, or force enable GECC if failed to
-				 * get boot configuration
-				 */
-				ret = psp_boot_config_set(adev, BOOT_CONFIG_GECC);
-				if (ret)
-					dev_warn(adev->dev, "PSP set boot config failed\n");
-				else
-					dev_warn(adev->dev, "GECC will be enabled in next boot cycle\n");
+				if (!boot_cfg) {
+					if (!adev->ras_default_ecc_enabled &&
+					    amdgpu_ras_enable != 1 &&
+					    amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__UMC))
+						dev_warn(adev->dev, "GECC is disabled, set amdgpu_ras_enable=1 to enable GECC in next boot cycle if needed\n");
+					else
+						dev_info(adev->dev, "GECC is disabled\n");
+				} else {
+					/* disable GECC in next boot cycle if ras is
+					 * disabled by module parameter amdgpu_ras_enable
+					 * and/or amdgpu_ras_mask, or boot_config_get call
+					 * is failed
+					 */
+					ret = psp_boot_config_set(adev, 0);
+					if (ret)
+						dev_warn(adev->dev, "PSP set boot config failed\n");
+					else
+						dev_warn(adev->dev, "GECC will be disabled in next boot cycle if set amdgpu_ras_enable and/or amdgpu_ras_mask to 0x0\n");
+				}
 			}
 		}
 	}
@@ -1469,8 +1975,8 @@ static int psp_ras_initialize(struct psp_context *psp)
 	psp->ras_context.context.mem_context.shared_mem_size = PSP_RAS_SHARED_MEM_SIZE;
 	psp->ras_context.context.ta_load_type = GFX_CMD_ID_LOAD_TA;
 
-	if (!psp->ras_context.context.initialized) {
-		ret = psp_ras_init_shared_buf(psp);
+	if (!psp->ras_context.context.mem_context.shared_buf) {
+		ret = psp_ta_init_shared_buf(psp, &psp->ras_context.context.mem_context);
 		if (ret)
 			return ret;
 	}
@@ -1480,64 +1986,101 @@ static int psp_ras_initialize(struct psp_context *psp)
 
 	if (amdgpu_ras_is_poison_mode_supported(adev))
 		ras_cmd->ras_in_message.init_flags.poison_mode_en = 1;
-	if (!adev->gmc.xgmi.connected_to_cpu)
+	if (!adev->gmc.xgmi.connected_to_cpu && !adev->gmc.is_app_apu)
 		ras_cmd->ras_in_message.init_flags.dgpu_mode = 1;
+	ras_cmd->ras_in_message.init_flags.xcc_mask =
+		adev->gfx.xcc_mask;
+	ras_cmd->ras_in_message.init_flags.channel_dis_num = hweight32(adev->gmc.m_half_use) * 2;
+	if (adev->gmc.gmc_funcs->query_mem_partition_mode)
+		ras_cmd->ras_in_message.init_flags.nps_mode =
+			adev->gmc.gmc_funcs->query_mem_partition_mode(adev);
+	ras_cmd->ras_in_message.init_flags.active_umc_mask = adev->umc.active_mask;
 
-	ret = psp_ras_load(psp);
+	ret = psp_ta_load(psp, &psp->ras_context.context);
 
-	if (!ret && !ras_cmd->ras_status)
+	if (!ret && !ras_cmd->ras_status) {
 		psp->ras_context.context.initialized = true;
-	else {
+		mutex_init(&psp->ras_context.mutex);
+	} else {
 		if (ras_cmd->ras_status)
-			dev_warn(psp->adev->dev, "RAS Init Status: 0x%X\n", ras_cmd->ras_status);
-		amdgpu_ras_fini(psp->adev);
+			dev_warn(adev->dev, "RAS Init Status: 0x%X\n", ras_cmd->ras_status);
+
+		/* fail to load RAS TA */
+		psp->ras_context.context.initialized = false;
 	}
 
 	return ret;
 }
 
 int psp_ras_trigger_error(struct psp_context *psp,
-			  struct ta_ras_trigger_error_input *info)
+			  struct ta_ras_trigger_error_input *info, uint32_t instance_mask)
 {
-	struct ta_ras_shared_memory *ras_cmd;
+	struct amdgpu_device *adev = psp->adev;
 	int ret;
+	uint32_t dev_mask;
+	uint32_t ras_status = 0;
 
-	if (!psp->ras_context.context.initialized)
+	if (!psp->ras_context.context.initialized || !info)
 		return -EINVAL;
 
-	ras_cmd = (struct ta_ras_shared_memory *)psp->ras_context.context.mem_context.shared_buf;
-	memset(ras_cmd, 0, sizeof(struct ta_ras_shared_memory));
+	switch (info->block_id) {
+	case TA_RAS_BLOCK__GFX:
+		dev_mask = GET_MASK(GC, instance_mask);
+		break;
+	case TA_RAS_BLOCK__SDMA:
+		dev_mask = GET_MASK(SDMA0, instance_mask);
+		break;
+	case TA_RAS_BLOCK__VCN:
+	case TA_RAS_BLOCK__JPEG:
+		dev_mask = GET_MASK(VCN, instance_mask);
+		break;
+	default:
+		dev_mask = instance_mask;
+		break;
+	}
 
-	ras_cmd->cmd_id = TA_RAS_COMMAND__TRIGGER_ERROR;
-	ras_cmd->ras_in_message.trigger_error = *info;
+	/* reuse sub_block_index for backward compatibility */
+	dev_mask <<= AMDGPU_RAS_INST_SHIFT;
+	dev_mask &= AMDGPU_RAS_INST_MASK;
+	info->sub_block_index |= dev_mask;
 
-	ret = psp_ras_invoke(psp, ras_cmd->cmd_id);
+	ret = psp_ras_send_cmd(psp,
+			TA_RAS_COMMAND__TRIGGER_ERROR, info, &ras_status);
 	if (ret)
 		return -EINVAL;
 
 	/* If err_event_athub occurs error inject was successful, however
-	   return status from TA is no long reliable */
+	 *  return status from TA is no long reliable
+	 */
 	if (amdgpu_ras_intr_triggered())
 		return 0;
 
-	if (ras_cmd->ras_status)
+	if (ras_status == TA_RAS_STATUS__TEE_ERROR_ACCESS_DENIED)
+		return -EACCES;
+	else if (ras_status)
 		return -EINVAL;
 
 	return 0;
 }
-// ras end
 
-// HDCP start
-static int psp_hdcp_init_shared_buf(struct psp_context *psp)
+int psp_ras_query_address(struct psp_context *psp,
+			  struct ta_ras_query_address_input *addr_in,
+			  struct ta_ras_query_address_output *addr_out)
 {
-	return psp_ta_init_shared_buf(psp, &psp->hdcp_context.context.mem_context);
-}
+	int ret;
 
-static int psp_hdcp_load(struct psp_context *psp)
-{
-	return psp_ta_load(psp, &psp->hdcp_context.context);
+	if (!psp->ras_context.context.initialized ||
+		!addr_in || !addr_out)
+		return -EINVAL;
+
+	ret = psp_ras_send_cmd(psp,
+			TA_RAS_COMMAND__QUERY_ADDRESS, addr_in, addr_out);
+
+	return ret;
 }
+// ras end
 
+// HDCP start
 static int psp_hdcp_initialize(struct psp_context *psp)
 {
 	int ret;
@@ -1548,6 +2091,10 @@ static int psp_hdcp_initialize(struct psp_context *psp)
 	if (amdgpu_sriov_vf(psp->adev))
 		return 0;
 
+	/* bypass hdcp initialization if dmu is harvested */
+	if (!amdgpu_device_has_display_hardware(psp->adev))
+		return 0;
+
 	if (!psp->hdcp_context.context.bin_desc.size_bytes ||
 	    !psp->hdcp_context.context.bin_desc.start_addr) {
 		dev_info(psp->adev->dev, "HDCP: optional hdcp ta ucode is not available\n");
@@ -1557,13 +2104,13 @@ static int psp_hdcp_initialize(struct psp_context *psp)
 	psp->hdcp_context.context.mem_context.shared_mem_size = PSP_HDCP_SHARED_MEM_SIZE;
 	psp->hdcp_context.context.ta_load_type = GFX_CMD_ID_LOAD_TA;
 
-	if (!psp->hdcp_context.context.initialized) {
-		ret = psp_hdcp_init_shared_buf(psp);
+	if (!psp->hdcp_context.context.mem_context.shared_buf) {
+		ret = psp_ta_init_shared_buf(psp, &psp->hdcp_context.context.mem_context);
 		if (ret)
 			return ret;
 	}
 
-	ret = psp_hdcp_load(psp);
+	ret = psp_ta_load(psp, &psp->hdcp_context.context);
 	if (!ret) {
 		psp->hdcp_context.context.initialized = true;
 		mutex_init(&psp->hdcp_context.mutex);
@@ -1572,11 +2119,6 @@ static int psp_hdcp_initialize(struct psp_context *psp)
 	return ret;
 }
 
-static int psp_hdcp_unload(struct psp_context *psp)
-{
-	return psp_ta_unload(psp, &psp->hdcp_context.context);
-}
-
 int psp_hdcp_invoke(struct psp_context *psp, uint32_t ta_cmd_id)
 {
 	/*
@@ -1585,6 +2127,9 @@ int psp_hdcp_invoke(struct psp_context *psp, uint32_t ta_cmd_id)
 	if (amdgpu_sriov_vf(psp->adev))
 		return 0;
 
+	if (!psp->hdcp_context.context.initialized)
+		return 0;
+
 	return psp_ta_invoke(psp, ta_cmd_id, &psp->hdcp_context.context);
 }
 
@@ -1598,38 +2143,18 @@ static int psp_hdcp_terminate(struct psp_context *psp)
 	if (amdgpu_sriov_vf(psp->adev))
 		return 0;
 
-	if (!psp->hdcp_context.context.initialized) {
-		if (psp->hdcp_context.context.mem_context.shared_buf)
-			goto out;
-		else
-			return 0;
-	}
+	if (!psp->hdcp_context.context.initialized)
+		return 0;
 
-	ret = psp_hdcp_unload(psp);
-	if (ret)
-		return ret;
+	ret = psp_ta_unload(psp, &psp->hdcp_context.context);
 
 	psp->hdcp_context.context.initialized = false;
 
-out:
-	/* free hdcp shared memory */
-	psp_ta_free_shared_buf(&psp->hdcp_context.context.mem_context);
-
-	return 0;
+	return ret;
 }
 // HDCP end
 
 // DTM start
-static int psp_dtm_init_shared_buf(struct psp_context *psp)
-{
-	return psp_ta_init_shared_buf(psp, &psp->dtm_context.context.mem_context);
-}
-
-static int psp_dtm_load(struct psp_context *psp)
-{
-	return psp_ta_load(psp, &psp->dtm_context.context);
-}
-
 static int psp_dtm_initialize(struct psp_context *psp)
 {
 	int ret;
@@ -1640,6 +2165,10 @@ static int psp_dtm_initialize(struct psp_context *psp)
 	if (amdgpu_sriov_vf(psp->adev))
 		return 0;
 
+	/* bypass dtm initialization if dmu is harvested */
+	if (!amdgpu_device_has_display_hardware(psp->adev))
+		return 0;
+
 	if (!psp->dtm_context.context.bin_desc.size_bytes ||
 	    !psp->dtm_context.context.bin_desc.start_addr) {
 		dev_info(psp->adev->dev, "DTM: optional dtm ta ucode is not available\n");
@@ -1649,13 +2178,13 @@ static int psp_dtm_initialize(struct psp_context *psp)
 	psp->dtm_context.context.mem_context.shared_mem_size = PSP_DTM_SHARED_MEM_SIZE;
 	psp->dtm_context.context.ta_load_type = GFX_CMD_ID_LOAD_TA;
 
-	if (!psp->dtm_context.context.initialized) {
-		ret = psp_dtm_init_shared_buf(psp);
+	if (!psp->dtm_context.context.mem_context.shared_buf) {
+		ret = psp_ta_init_shared_buf(psp, &psp->dtm_context.context.mem_context);
 		if (ret)
 			return ret;
 	}
 
-	ret = psp_dtm_load(psp);
+	ret = psp_ta_load(psp, &psp->dtm_context.context);
 	if (!ret) {
 		psp->dtm_context.context.initialized = true;
 		mutex_init(&psp->dtm_context.mutex);
@@ -1664,11 +2193,6 @@ static int psp_dtm_initialize(struct psp_context *psp)
 	return ret;
 }
 
-static int psp_dtm_unload(struct psp_context *psp)
-{
-	return psp_ta_unload(psp, &psp->dtm_context.context);
-}
-
 int psp_dtm_invoke(struct psp_context *psp, uint32_t ta_cmd_id)
 {
 	/*
@@ -1677,6 +2201,9 @@ int psp_dtm_invoke(struct psp_context *psp, uint32_t ta_cmd_id)
 	if (amdgpu_sriov_vf(psp->adev))
 		return 0;
 
+	if (!psp->dtm_context.context.initialized)
+		return 0;
+
 	return psp_ta_invoke(psp, ta_cmd_id, &psp->dtm_context.context);
 }
 
@@ -1690,43 +2217,18 @@ static int psp_dtm_terminate(struct psp_context *psp)
 	if (amdgpu_sriov_vf(psp->adev))
 		return 0;
 
-	if (!psp->dtm_context.context.initialized) {
-		if (psp->dtm_context.context.mem_context.shared_buf)
-			goto out;
-		else
-			return 0;
-	}
+	if (!psp->dtm_context.context.initialized)
+		return 0;
 
-	ret = psp_dtm_unload(psp);
-	if (ret)
-		return ret;
+	ret = psp_ta_unload(psp, &psp->dtm_context.context);
 
 	psp->dtm_context.context.initialized = false;
 
-out:
-	/* free dtm shared memory */
-	psp_ta_free_shared_buf(&psp->dtm_context.context.mem_context);
-
-	return 0;
+	return ret;
 }
 // DTM end
 
 // RAP start
-static int psp_rap_init_shared_buf(struct psp_context *psp)
-{
-	return psp_ta_init_shared_buf(psp, &psp->rap_context.context.mem_context);
-}
-
-static int psp_rap_load(struct psp_context *psp)
-{
-	return psp_ta_load(psp, &psp->rap_context.context);
-}
-
-static int psp_rap_unload(struct psp_context *psp)
-{
-	return psp_ta_unload(psp, &psp->rap_context.context);
-}
-
 static int psp_rap_initialize(struct psp_context *psp)
 {
 	int ret;
@@ -1747,13 +2249,13 @@ static int psp_rap_initialize(struct psp_context *psp)
 	psp->rap_context.context.mem_context.shared_mem_size = PSP_RAP_SHARED_MEM_SIZE;
 	psp->rap_context.context.ta_load_type = GFX_CMD_ID_LOAD_TA;
 
-	if (!psp->rap_context.context.initialized) {
-		ret = psp_rap_init_shared_buf(psp);
+	if (!psp->rap_context.context.mem_context.shared_buf) {
+		ret = psp_ta_init_shared_buf(psp, &psp->rap_context.context.mem_context);
 		if (ret)
 			return ret;
 	}
 
-	ret = psp_rap_load(psp);
+	ret = psp_ta_load(psp, &psp->rap_context.context);
 	if (!ret) {
 		psp->rap_context.context.initialized = true;
 		mutex_init(&psp->rap_context.mutex);
@@ -1763,6 +2265,8 @@ static int psp_rap_initialize(struct psp_context *psp)
 	ret = psp_rap_invoke(psp, TA_CMD_RAP__INITIALIZE, &status);
 	if (ret || status != TA_RAP_STATUS__SUCCESS) {
 		psp_rap_terminate(psp);
+		/* free rap shared memory */
+		psp_ta_free_shared_buf(&psp->rap_context.context.mem_context);
 
 		dev_warn(psp->adev->dev, "RAP TA initialize fail (%d) status %d.\n",
 			 ret, status);
@@ -1780,13 +2284,10 @@ static int psp_rap_terminate(struct psp_context *psp)
 	if (!psp->rap_context.context.initialized)
 		return 0;
 
-	ret = psp_rap_unload(psp);
+	ret = psp_ta_unload(psp, &psp->rap_context.context);
 
 	psp->rap_context.context.initialized = false;
 
-	/* free rap shared memory */
-	psp_ta_free_shared_buf(&psp->rap_context.context.mem_context);
-
 	return ret;
 }
 
@@ -1826,26 +2327,10 @@ out_unlock:
 // RAP end
 
 /* securedisplay start */
-static int psp_securedisplay_init_shared_buf(struct psp_context *psp)
-{
-	return psp_ta_init_shared_buf(
-		psp, &psp->securedisplay_context.context.mem_context);
-}
-
-static int psp_securedisplay_load(struct psp_context *psp)
-{
-	return psp_ta_load(psp, &psp->securedisplay_context.context);
-}
-
-static int psp_securedisplay_unload(struct psp_context *psp)
-{
-	return psp_ta_unload(psp, &psp->securedisplay_context.context);
-}
-
 static int psp_securedisplay_initialize(struct psp_context *psp)
 {
 	int ret;
-	struct securedisplay_cmd *securedisplay_cmd;
+	struct ta_securedisplay_cmd *securedisplay_cmd;
 
 	/*
 	 * TODO: bypass the initialize in sriov for now
@@ -1853,9 +2338,14 @@ static int psp_securedisplay_initialize(struct psp_context *psp)
 	if (amdgpu_sriov_vf(psp->adev))
 		return 0;
 
+	/* bypass securedisplay initialization if dmu is harvested */
+	if (!amdgpu_device_has_display_hardware(psp->adev))
+		return 0;
+
 	if (!psp->securedisplay_context.context.bin_desc.size_bytes ||
 	    !psp->securedisplay_context.context.bin_desc.start_addr) {
-		dev_info(psp->adev->dev, "SECUREDISPLAY: securedisplay ta ucode is not available\n");
+		dev_info(psp->adev->dev,
+			 "SECUREDISPLAY: optional securedisplay ta ucode is not available\n");
 		return 0;
 	}
 
@@ -1864,24 +2354,35 @@ static int psp_securedisplay_initialize(struct psp_context *psp)
 	psp->securedisplay_context.context.ta_load_type = GFX_CMD_ID_LOAD_TA;
 
 	if (!psp->securedisplay_context.context.initialized) {
-		ret = psp_securedisplay_init_shared_buf(psp);
+		ret = psp_ta_init_shared_buf(psp,
+					     &psp->securedisplay_context.context.mem_context);
 		if (ret)
 			return ret;
 	}
 
-	ret = psp_securedisplay_load(psp);
-	if (!ret) {
+	ret = psp_ta_load(psp, &psp->securedisplay_context.context);
+	if (!ret && !psp->securedisplay_context.context.resp_status) {
 		psp->securedisplay_context.context.initialized = true;
 		mutex_init(&psp->securedisplay_context.mutex);
-	} else
+	} else {
+		/* don't try again */
+		psp->securedisplay_context.context.bin_desc.size_bytes = 0;
 		return ret;
+	}
+
+	mutex_lock(&psp->securedisplay_context.mutex);
 
 	psp_prep_securedisplay_cmd_buf(psp, &securedisplay_cmd,
 			TA_SECUREDISPLAY_COMMAND__QUERY_TA);
 
 	ret = psp_securedisplay_invoke(psp, TA_SECUREDISPLAY_COMMAND__QUERY_TA);
+
+	mutex_unlock(&psp->securedisplay_context.mutex);
+
 	if (ret) {
 		psp_securedisplay_terminate(psp);
+		/* free securedisplay shared memory */
+		psp_ta_free_shared_buf(&psp->securedisplay_context.context.mem_context);
 		dev_err(psp->adev->dev, "SECUREDISPLAY TA initialize fail.\n");
 		return -EINVAL;
 	}
@@ -1890,6 +2391,8 @@ static int psp_securedisplay_initialize(struct psp_context *psp)
 		psp_securedisplay_parse_resp_status(psp, securedisplay_cmd->status);
 		dev_err(psp->adev->dev, "SECUREDISPLAY: query securedisplay TA failed. ret 0x%x\n",
 			securedisplay_cmd->securedisplay_out_message.query_ta.query_cmd_ret);
+		/* don't try again */
+		psp->securedisplay_context.context.bin_desc.size_bytes = 0;
 	}
 
 	return 0;
@@ -1908,15 +2411,10 @@ static int psp_securedisplay_terminate(struct psp_context *psp)
 	if (!psp->securedisplay_context.context.initialized)
 		return 0;
 
-	ret = psp_securedisplay_unload(psp);
-	if (ret)
-		return ret;
+	ret = psp_ta_unload(psp, &psp->securedisplay_context.context);
 
 	psp->securedisplay_context.context.initialized = false;
 
-	/* free securedisplay shared memory */
-	psp_ta_free_shared_buf(&psp->securedisplay_context.context.mem_context);
-
 	return ret;
 }
 
@@ -1928,30 +2426,77 @@ int psp_securedisplay_invoke(struct psp_context *psp, uint32_t ta_cmd_id)
 		return -EINVAL;
 
 	if (ta_cmd_id != TA_SECUREDISPLAY_COMMAND__QUERY_TA &&
-	    ta_cmd_id != TA_SECUREDISPLAY_COMMAND__SEND_ROI_CRC)
+	    ta_cmd_id != TA_SECUREDISPLAY_COMMAND__SEND_ROI_CRC &&
+	    ta_cmd_id != TA_SECUREDISPLAY_COMMAND__SEND_ROI_CRC_V2)
 		return -EINVAL;
 
-	mutex_lock(&psp->securedisplay_context.mutex);
-
 	ret = psp_ta_invoke(psp, ta_cmd_id, &psp->securedisplay_context.context);
 
-	mutex_unlock(&psp->securedisplay_context.mutex);
-
 	return ret;
 }
 /* SECUREDISPLAY end */
 
+int amdgpu_psp_wait_for_bootloader(struct amdgpu_device *adev)
+{
+	struct psp_context *psp = &adev->psp;
+	int ret = 0;
+
+	if (!amdgpu_sriov_vf(adev) && psp->funcs && psp->funcs->wait_for_bootloader != NULL)
+		ret = psp->funcs->wait_for_bootloader(psp);
+
+	return ret;
+}
+
+bool amdgpu_psp_get_ras_capability(struct psp_context *psp)
+{
+	if (psp->funcs &&
+	    psp->funcs->get_ras_capability) {
+		return psp->funcs->get_ras_capability(psp);
+	} else {
+		return false;
+	}
+}
+
+bool amdgpu_psp_tos_reload_needed(struct amdgpu_device *adev)
+{
+	struct psp_context *psp = &adev->psp;
+
+	if (amdgpu_sriov_vf(adev) || (adev->flags & AMD_IS_APU))
+		return false;
+
+	if (psp->funcs && psp->funcs->is_reload_needed)
+		return psp->funcs->is_reload_needed(psp);
+
+	return false;
+}
+
+static void psp_update_gpu_addresses(struct amdgpu_device *adev)
+{
+	struct psp_context *psp = &adev->psp;
+
+	if (psp->cmd_buf_bo && psp->cmd_buf_mem) {
+		psp->fw_pri_mc_addr = amdgpu_bo_fb_aper_addr(psp->fw_pri_bo);
+		psp->fence_buf_mc_addr = amdgpu_bo_fb_aper_addr(psp->fence_buf_bo);
+		psp->cmd_buf_mc_addr = amdgpu_bo_fb_aper_addr(psp->cmd_buf_bo);
+	}
+	if (adev->firmware.rbuf && psp->km_ring.ring_mem)
+		psp->km_ring.ring_mem_mc_addr = amdgpu_bo_fb_aper_addr(adev->firmware.rbuf);
+}
+
 static int psp_hw_start(struct psp_context *psp)
 {
 	struct amdgpu_device *adev = psp->adev;
 	int ret;
 
+	if (amdgpu_virt_xgmi_migrate_enabled(adev))
+		psp_update_gpu_addresses(adev);
+
 	if (!amdgpu_sriov_vf(adev)) {
 		if ((is_psp_fw_valid(psp->kdb)) &&
 		    (psp->funcs->bootloader_load_kdb != NULL)) {
 			ret = psp_bootloader_load_kdb(psp);
 			if (ret) {
-				DRM_ERROR("PSP load kdb failed!\n");
+				dev_err(adev->dev, "PSP load kdb failed!\n");
 				return ret;
 			}
 		}
@@ -1960,7 +2505,7 @@ static int psp_hw_start(struct psp_context *psp)
 		    (psp->funcs->bootloader_load_spl != NULL)) {
 			ret = psp_bootloader_load_spl(psp);
 			if (ret) {
-				DRM_ERROR("PSP load spl failed!\n");
+				dev_err(adev->dev, "PSP load spl failed!\n");
 				return ret;
 			}
 		}
@@ -1969,7 +2514,7 @@ static int psp_hw_start(struct psp_context *psp)
 		    (psp->funcs->bootloader_load_sysdrv != NULL)) {
 			ret = psp_bootloader_load_sysdrv(psp);
 			if (ret) {
-				DRM_ERROR("PSP load sys drv failed!\n");
+				dev_err(adev->dev, "PSP load sys drv failed!\n");
 				return ret;
 			}
 		}
@@ -1978,7 +2523,7 @@ static int psp_hw_start(struct psp_context *psp)
 		    (psp->funcs->bootloader_load_soc_drv != NULL)) {
 			ret = psp_bootloader_load_soc_drv(psp);
 			if (ret) {
-				DRM_ERROR("PSP load soc drv failed!\n");
+				dev_err(adev->dev, "PSP load soc drv failed!\n");
 				return ret;
 			}
 		}
@@ -1987,7 +2532,7 @@ static int psp_hw_start(struct psp_context *psp)
 		    (psp->funcs->bootloader_load_intf_drv != NULL)) {
 			ret = psp_bootloader_load_intf_drv(psp);
 			if (ret) {
-				DRM_ERROR("PSP load intf drv failed!\n");
+				dev_err(adev->dev, "PSP load intf drv failed!\n");
 				return ret;
 			}
 		}
@@ -1996,7 +2541,34 @@ static int psp_hw_start(struct psp_context *psp)
 		    (psp->funcs->bootloader_load_dbg_drv != NULL)) {
 			ret = psp_bootloader_load_dbg_drv(psp);
 			if (ret) {
-				DRM_ERROR("PSP load dbg drv failed!\n");
+				dev_err(adev->dev, "PSP load dbg drv failed!\n");
+				return ret;
+			}
+		}
+
+		if ((is_psp_fw_valid(psp->ras_drv)) &&
+		    (psp->funcs->bootloader_load_ras_drv != NULL)) {
+			ret = psp_bootloader_load_ras_drv(psp);
+			if (ret) {
+				dev_err(adev->dev, "PSP load ras_drv failed!\n");
+				return ret;
+			}
+		}
+
+		if ((is_psp_fw_valid(psp->ipkeymgr_drv)) &&
+		    (psp->funcs->bootloader_load_ipkeymgr_drv != NULL)) {
+			ret = psp_bootloader_load_ipkeymgr_drv(psp);
+			if (ret) {
+				dev_err(adev->dev, "PSP load ipkeymgr_drv failed!\n");
+				return ret;
+			}
+		}
+
+		if ((is_psp_fw_valid(psp->spdm_drv)) &&
+		    (psp->funcs->bootloader_load_spdm_drv != NULL)) {
+			ret = psp_bootloader_load_spdm_drv(psp);
+			if (ret) {
+				dev_err(adev->dev, "PSP load spdm_drv failed!\n");
 				return ret;
 			}
 		}
@@ -2005,7 +2577,7 @@ static int psp_hw_start(struct psp_context *psp)
 		    (psp->funcs->bootloader_load_sos != NULL)) {
 			ret = psp_bootloader_load_sos(psp);
 			if (ret) {
-				DRM_ERROR("PSP load sos failed!\n");
+				dev_err(adev->dev, "PSP load sos failed!\n");
 				return ret;
 			}
 		}
@@ -2013,16 +2585,30 @@ static int psp_hw_start(struct psp_context *psp)
 
 	ret = psp_ring_create(psp, PSP_RING_TYPE__KM);
 	if (ret) {
-		DRM_ERROR("PSP create ring failed!\n");
+		dev_err(adev->dev, "PSP create ring failed!\n");
 		return ret;
 	}
 
-	ret = psp_tmr_init(psp);
-	if (ret) {
-		DRM_ERROR("PSP tmr init failed!\n");
-		return ret;
+	if (!amdgpu_in_reset(adev) && !adev->in_suspend) {
+		ret = psp_update_fw_reservation(psp);
+		if (ret) {
+			dev_err(adev->dev, "update fw reservation failed!\n");
+			return ret;
+		}
+	}
+
+	if (amdgpu_sriov_vf(adev) && amdgpu_in_reset(adev))
+		goto skip_pin_bo;
+
+	if (!psp->boot_time_tmr || psp->autoload_supported) {
+		ret = psp_tmr_init(psp);
+		if (ret) {
+			dev_err(adev->dev, "PSP tmr init failed!\n");
+			return ret;
+		}
 	}
 
+skip_pin_bo:
 	/*
 	 * For ASICs with DF Cstate management centralized
 	 * to PMFW, TMR setup should be performed after PMFW
@@ -2034,10 +2620,12 @@ static int psp_hw_start(struct psp_context *psp)
 			return ret;
 	}
 
-	ret = psp_tmr_load(psp);
-	if (ret) {
-		DRM_ERROR("PSP load tmr failed!\n");
-		return ret;
+	if (!psp->boot_time_tmr || !psp->autoload_supported) {
+		ret = psp_tmr_load(psp);
+		if (ret) {
+			dev_err(adev->dev, "PSP load tmr failed!\n");
+			return ret;
+		}
 	}
 
 	return 0;
@@ -2047,6 +2635,9 @@ static int psp_get_fw_type(struct amdgpu_firmware_info *ucode,
 			   enum psp_gfx_fw_type *type)
 {
 	switch (ucode->ucode_id) {
+	case AMDGPU_UCODE_ID_CAP:
+		*type = GFX_FW_TYPE_CAP;
+		break;
 	case AMDGPU_UCODE_ID_SDMA0:
 		*type = GFX_FW_TYPE_SDMA0;
 		break;
@@ -2077,6 +2668,12 @@ static int psp_get_fw_type(struct amdgpu_firmware_info *ucode,
 	case AMDGPU_UCODE_ID_CP_MES_DATA:
 		*type = GFX_FW_TYPE_MES_STACK;
 		break;
+	case AMDGPU_UCODE_ID_CP_MES1:
+		*type = GFX_FW_TYPE_CP_MES_KIQ;
+		break;
+	case AMDGPU_UCODE_ID_CP_MES1_DATA:
+		*type = GFX_FW_TYPE_MES_KIQ_STACK;
+		break;
 	case AMDGPU_UCODE_ID_CP_CE:
 		*type = GFX_FW_TYPE_CP_CE;
 		break;
@@ -2098,6 +2695,12 @@ static int psp_get_fw_type(struct amdgpu_firmware_info *ucode,
 	case AMDGPU_UCODE_ID_CP_MEC2_JT:
 		*type = GFX_FW_TYPE_CP_MEC_ME2;
 		break;
+	case AMDGPU_UCODE_ID_RLC_P:
+		*type = GFX_FW_TYPE_RLC_P;
+		break;
+	case AMDGPU_UCODE_ID_RLC_V:
+		*type = GFX_FW_TYPE_RLC_V;
+		break;
 	case AMDGPU_UCODE_ID_RLC_G:
 		*type = GFX_FW_TYPE_RLC_G;
 		break;
@@ -2116,9 +2719,27 @@ static int psp_get_fw_type(struct amdgpu_firmware_info *ucode,
 	case AMDGPU_UCODE_ID_RLC_DRAM:
 		*type = GFX_FW_TYPE_RLC_DRAM_BOOT;
 		break;
+	case AMDGPU_UCODE_ID_GLOBAL_TAP_DELAYS:
+		*type = GFX_FW_TYPE_GLOBAL_TAP_DELAYS;
+		break;
+	case AMDGPU_UCODE_ID_SE0_TAP_DELAYS:
+		*type = GFX_FW_TYPE_SE0_TAP_DELAYS;
+		break;
+	case AMDGPU_UCODE_ID_SE1_TAP_DELAYS:
+		*type = GFX_FW_TYPE_SE1_TAP_DELAYS;
+		break;
+	case AMDGPU_UCODE_ID_SE2_TAP_DELAYS:
+		*type = GFX_FW_TYPE_SE2_TAP_DELAYS;
+		break;
+	case AMDGPU_UCODE_ID_SE3_TAP_DELAYS:
+		*type = GFX_FW_TYPE_SE3_TAP_DELAYS;
+		break;
 	case AMDGPU_UCODE_ID_SMC:
 		*type = GFX_FW_TYPE_SMU;
 		break;
+	case AMDGPU_UCODE_ID_PPTABLE:
+		*type = GFX_FW_TYPE_PPTABLE;
+		break;
 	case AMDGPU_UCODE_ID_UVD:
 		*type = GFX_FW_TYPE_UVD;
 		break;
@@ -2149,6 +2770,79 @@ static int psp_get_fw_type(struct amdgpu_firmware_info *ucode,
 	case AMDGPU_UCODE_ID_DMCUB:
 		*type = GFX_FW_TYPE_DMUB;
 		break;
+	case AMDGPU_UCODE_ID_SDMA_UCODE_TH0:
+	case AMDGPU_UCODE_ID_SDMA_RS64:
+		*type = GFX_FW_TYPE_SDMA_UCODE_TH0;
+		break;
+	case AMDGPU_UCODE_ID_SDMA_UCODE_TH1:
+		*type = GFX_FW_TYPE_SDMA_UCODE_TH1;
+		break;
+	case AMDGPU_UCODE_ID_IMU_I:
+		*type = GFX_FW_TYPE_IMU_I;
+		break;
+	case AMDGPU_UCODE_ID_IMU_D:
+		*type = GFX_FW_TYPE_IMU_D;
+		break;
+	case AMDGPU_UCODE_ID_CP_RS64_PFP:
+		*type = GFX_FW_TYPE_RS64_PFP;
+		break;
+	case AMDGPU_UCODE_ID_CP_RS64_ME:
+		*type = GFX_FW_TYPE_RS64_ME;
+		break;
+	case AMDGPU_UCODE_ID_CP_RS64_MEC:
+		*type = GFX_FW_TYPE_RS64_MEC;
+		break;
+	case AMDGPU_UCODE_ID_CP_RS64_PFP_P0_STACK:
+		*type = GFX_FW_TYPE_RS64_PFP_P0_STACK;
+		break;
+	case AMDGPU_UCODE_ID_CP_RS64_PFP_P1_STACK:
+		*type = GFX_FW_TYPE_RS64_PFP_P1_STACK;
+		break;
+	case AMDGPU_UCODE_ID_CP_RS64_ME_P0_STACK:
+		*type = GFX_FW_TYPE_RS64_ME_P0_STACK;
+		break;
+	case AMDGPU_UCODE_ID_CP_RS64_ME_P1_STACK:
+		*type = GFX_FW_TYPE_RS64_ME_P1_STACK;
+		break;
+	case AMDGPU_UCODE_ID_CP_RS64_MEC_P0_STACK:
+		*type = GFX_FW_TYPE_RS64_MEC_P0_STACK;
+		break;
+	case AMDGPU_UCODE_ID_CP_RS64_MEC_P1_STACK:
+		*type = GFX_FW_TYPE_RS64_MEC_P1_STACK;
+		break;
+	case AMDGPU_UCODE_ID_CP_RS64_MEC_P2_STACK:
+		*type = GFX_FW_TYPE_RS64_MEC_P2_STACK;
+		break;
+	case AMDGPU_UCODE_ID_CP_RS64_MEC_P3_STACK:
+		*type = GFX_FW_TYPE_RS64_MEC_P3_STACK;
+		break;
+	case AMDGPU_UCODE_ID_VPE_CTX:
+		*type = GFX_FW_TYPE_VPEC_FW1;
+		break;
+	case AMDGPU_UCODE_ID_VPE_CTL:
+		*type = GFX_FW_TYPE_VPEC_FW2;
+		break;
+	case AMDGPU_UCODE_ID_VPE:
+		*type = GFX_FW_TYPE_VPE;
+		break;
+	case AMDGPU_UCODE_ID_UMSCH_MM_UCODE:
+		*type = GFX_FW_TYPE_UMSCH_UCODE;
+		break;
+	case AMDGPU_UCODE_ID_UMSCH_MM_DATA:
+		*type = GFX_FW_TYPE_UMSCH_DATA;
+		break;
+	case AMDGPU_UCODE_ID_UMSCH_MM_CMD_BUFFER:
+		*type = GFX_FW_TYPE_UMSCH_CMD_BUFFER;
+		break;
+	case AMDGPU_UCODE_ID_P2S_TABLE:
+		*type = GFX_FW_TYPE_P2S_TABLE;
+		break;
+	case AMDGPU_UCODE_ID_JPEG_RAM:
+		*type = GFX_FW_TYPE_JPEG_RAM;
+		break;
+	case AMDGPU_UCODE_ID_ISP:
+		*type = GFX_FW_TYPE_ISP;
+		break;
 	case AMDGPU_UCODE_ID_MAXIMUM:
 	default:
 		return -EINVAL;
@@ -2205,7 +2899,8 @@ static void psp_print_fw_hdr(struct psp_context *psp,
 	}
 }
 
-static int psp_prep_load_ip_fw_cmd_buf(struct amdgpu_firmware_info *ucode,
+static int psp_prep_load_ip_fw_cmd_buf(struct psp_context *psp,
+				       struct amdgpu_firmware_info *ucode,
 				       struct psp_gfx_cmd_resp *cmd)
 {
 	int ret;
@@ -2218,18 +2913,18 @@ static int psp_prep_load_ip_fw_cmd_buf(struct amdgpu_firmware_info *ucode,
 
 	ret = psp_get_fw_type(ucode, &cmd->cmd.cmd_load_ip_fw.fw_type);
 	if (ret)
-		DRM_ERROR("Unknown firmware type\n");
+		dev_err(psp->adev->dev, "Unknown firmware type\n");
 
 	return ret;
 }
 
-static int psp_execute_non_psp_fw_load(struct psp_context *psp,
-			          struct amdgpu_firmware_info *ucode)
+int psp_execute_ip_fw_load(struct psp_context *psp,
+			   struct amdgpu_firmware_info *ucode)
 {
 	int ret = 0;
 	struct psp_gfx_cmd_resp *cmd = acquire_psp_cmd_buf(psp);
 
-	ret = psp_prep_load_ip_fw_cmd_buf(ucode, cmd);
+	ret = psp_prep_load_ip_fw_cmd_buf(psp, ucode, cmd);
 	if (!ret) {
 		ret = psp_cmd_submit_buf(psp, ucode, cmd,
 					 psp->fence_buf_mc_addr);
@@ -2240,6 +2935,33 @@ static int psp_execute_non_psp_fw_load(struct psp_context *psp,
 	return ret;
 }
 
+static int psp_load_p2s_table(struct psp_context *psp)
+{
+	int ret;
+	struct amdgpu_device *adev = psp->adev;
+	struct amdgpu_firmware_info *ucode =
+		&adev->firmware.ucode[AMDGPU_UCODE_ID_P2S_TABLE];
+
+	if (adev->in_runpm && ((adev->pm.rpm_mode == AMDGPU_RUNPM_BACO) ||
+				(adev->pm.rpm_mode == AMDGPU_RUNPM_BAMACO)))
+		return 0;
+
+	if (amdgpu_ip_version(adev, MP0_HWIP, 0) == IP_VERSION(13, 0, 6) ||
+	    amdgpu_ip_version(adev, MP0_HWIP, 0) == IP_VERSION(13, 0, 14)) {
+		uint32_t supp_vers = adev->flags & AMD_IS_APU ? 0x0036013D :
+								0x0036003C;
+		if (psp->sos.fw_version < supp_vers)
+			return 0;
+	}
+
+	if (!ucode->fw || amdgpu_sriov_vf(psp->adev))
+		return 0;
+
+	ret = psp_execute_ip_fw_load(psp, ucode);
+
+	return ret;
+}
+
 static int psp_load_smu_fw(struct psp_context *psp)
 {
 	int ret;
@@ -2248,23 +2970,29 @@ static int psp_load_smu_fw(struct psp_context *psp)
 			&adev->firmware.ucode[AMDGPU_UCODE_ID_SMC];
 	struct amdgpu_ras *ras = psp->ras_context.ras;
 
+	/*
+	 * Skip SMU FW reloading in case of using BACO for runpm only,
+	 * as SMU is always alive.
+	 */
+	if (adev->in_runpm && ((adev->pm.rpm_mode == AMDGPU_RUNPM_BACO) ||
+				(adev->pm.rpm_mode == AMDGPU_RUNPM_BAMACO)))
+		return 0;
+
 	if (!ucode->fw || amdgpu_sriov_vf(psp->adev))
 		return 0;
 
-	if ((amdgpu_in_reset(adev) &&
-	     ras && adev->ras_enabled &&
-	     (adev->ip_versions[MP0_HWIP][0] == IP_VERSION(11, 0, 4) ||
-	      adev->ip_versions[MP0_HWIP][0] == IP_VERSION(11, 0, 2)))) {
+	if ((amdgpu_in_reset(adev) && ras && adev->ras_enabled &&
+	     (amdgpu_ip_version(adev, MP0_HWIP, 0) == IP_VERSION(11, 0, 4) ||
+	      amdgpu_ip_version(adev, MP0_HWIP, 0) == IP_VERSION(11, 0, 2)))) {
 		ret = amdgpu_dpm_set_mp1_state(adev, PP_MP1_STATE_UNLOAD);
-		if (ret) {
-			DRM_WARN("Failed to set MP1 state prepare for reload\n");
-		}
+		if (ret)
+			dev_err(adev->dev, "Failed to set MP1 state prepare for reload\n");
 	}
 
-	ret = psp_execute_non_psp_fw_load(psp, ucode);
+	ret = psp_execute_ip_fw_load(psp, ucode);
 
 	if (ret)
-		DRM_ERROR("PSP load smu failed!\n");
+		dev_err(adev->dev, "PSP load smu failed!\n");
 
 	return ret;
 }
@@ -2272,7 +3000,10 @@ static int psp_load_smu_fw(struct psp_context *psp)
 static bool fw_load_skip_check(struct psp_context *psp,
 			       struct amdgpu_firmware_info *ucode)
 {
-	if (!ucode->fw)
+	if (!ucode->fw || !ucode->ucode_size)
+		return true;
+
+	if (ucode->ucode_id == AMDGPU_UCODE_ID_P2S_TABLE)
 		return true;
 
 	if (ucode->ucode_id == AMDGPU_UCODE_ID_SMC &&
@@ -2282,20 +3013,7 @@ static bool fw_load_skip_check(struct psp_context *psp,
 		return true;
 
 	if (amdgpu_sriov_vf(psp->adev) &&
-	   (ucode->ucode_id == AMDGPU_UCODE_ID_SDMA0
-	    || ucode->ucode_id == AMDGPU_UCODE_ID_SDMA1
-	    || ucode->ucode_id == AMDGPU_UCODE_ID_SDMA2
-	    || ucode->ucode_id == AMDGPU_UCODE_ID_SDMA3
-	    || ucode->ucode_id == AMDGPU_UCODE_ID_SDMA4
-	    || ucode->ucode_id == AMDGPU_UCODE_ID_SDMA5
-	    || ucode->ucode_id == AMDGPU_UCODE_ID_SDMA6
-	    || ucode->ucode_id == AMDGPU_UCODE_ID_SDMA7
-	    || ucode->ucode_id == AMDGPU_UCODE_ID_RLC_G
-	    || ucode->ucode_id == AMDGPU_UCODE_ID_RLC_RESTORE_LIST_CNTL
-	    || ucode->ucode_id == AMDGPU_UCODE_ID_RLC_RESTORE_LIST_GPM_MEM
-	    || ucode->ucode_id == AMDGPU_UCODE_ID_RLC_RESTORE_LIST_SRM_MEM
-	    || ucode->ucode_id == AMDGPU_UCODE_ID_SMC))
-		/*skip ucode loading in SRIOV VF */
+	    amdgpu_virt_fw_load_skip_check(psp->adev, ucode->ucode_id))
 		return true;
 
 	if (psp->autoload_supported &&
@@ -2316,7 +3034,7 @@ int psp_load_fw_list(struct psp_context *psp,
 	for (i = 0; i < ucode_count; ++i) {
 		ucode = ucode_list[i];
 		psp_print_fw_hdr(psp, ucode);
-		ret = psp_execute_non_psp_fw_load(psp, ucode);
+		ret = psp_execute_ip_fw_load(psp, ucode);
 		if (ret)
 			return ret;
 	}
@@ -2336,6 +3054,9 @@ static int psp_load_non_psp_fw(struct psp_context *psp)
 			return ret;
 	}
 
+	/* Load P2S table first if it's available */
+	psp_load_p2s_table(psp);
+
 	for (i = 0; i < adev->firmware.max_ucodes; i++) {
 		ucode = &adev->firmware.ucode[i];
 
@@ -2351,28 +3072,32 @@ static int psp_load_non_psp_fw(struct psp_context *psp)
 			continue;
 
 		if (psp->autoload_supported &&
-		    (adev->ip_versions[MP0_HWIP][0] == IP_VERSION(11, 0, 7) ||
-		     adev->ip_versions[MP0_HWIP][0] == IP_VERSION(11, 0, 11) ||
-		     adev->ip_versions[MP0_HWIP][0] == IP_VERSION(11, 0, 12)) &&
+		    (amdgpu_ip_version(adev, MP0_HWIP, 0) ==
+			     IP_VERSION(11, 0, 7) ||
+		     amdgpu_ip_version(adev, MP0_HWIP, 0) ==
+			     IP_VERSION(11, 0, 11) ||
+		     amdgpu_ip_version(adev, MP0_HWIP, 0) ==
+			     IP_VERSION(11, 0, 12)) &&
 		    (ucode->ucode_id == AMDGPU_UCODE_ID_SDMA1 ||
 		     ucode->ucode_id == AMDGPU_UCODE_ID_SDMA2 ||
 		     ucode->ucode_id == AMDGPU_UCODE_ID_SDMA3))
 			/* PSP only receive one SDMA fw for sienna_cichlid,
-			 * as all four sdma fw are same */
+			 * as all four sdma fw are same
+			 */
 			continue;
 
 		psp_print_fw_hdr(psp, ucode);
 
-		ret = psp_execute_non_psp_fw_load(psp, ucode);
+		ret = psp_execute_ip_fw_load(psp, ucode);
 		if (ret)
 			return ret;
 
-		/* Start rlc autoload after psp recieved all the gfx firmware */
+		/* Start rlc autoload after psp received all the gfx firmware */
 		if (psp->autoload_supported && ucode->ucode_id == (amdgpu_sriov_vf(adev) ?
-		    AMDGPU_UCODE_ID_CP_MEC2 : AMDGPU_UCODE_ID_RLC_G)) {
+		    adev->virt.autoload_ucode_id : AMDGPU_UCODE_ID_RLC_G)) {
 			ret = psp_rlc_autoload_start(psp);
 			if (ret) {
-				DRM_ERROR("Failed to start rlc autoload\n");
+				dev_err(adev->dev, "Failed to start rlc autoload\n");
 				return ret;
 			}
 		}
@@ -2387,76 +3112,55 @@ static int psp_load_fw(struct amdgpu_device *adev)
 	struct psp_context *psp = &adev->psp;
 
 	if (amdgpu_sriov_vf(adev) && amdgpu_in_reset(adev)) {
-		psp_ring_stop(psp, PSP_RING_TYPE__KM); /* should not destroy ring, only stop */
-		goto skip_memalloc;
-	}
-
-	if (amdgpu_sriov_vf(adev)) {
-		ret = amdgpu_bo_create_kernel(adev, PSP_1_MEG, PSP_1_MEG,
-						AMDGPU_GEM_DOMAIN_VRAM,
-						&psp->fw_pri_bo,
-						&psp->fw_pri_mc_addr,
-						&psp->fw_pri_buf);
+		/* should not destroy ring, only stop */
+		psp_ring_stop(psp, PSP_RING_TYPE__KM);
 	} else {
-		ret = amdgpu_bo_create_kernel(adev, PSP_1_MEG, PSP_1_MEG,
-						AMDGPU_GEM_DOMAIN_GTT,
-						&psp->fw_pri_bo,
-						&psp->fw_pri_mc_addr,
-						&psp->fw_pri_buf);
-	}
-
-	if (ret)
-		goto failed;
-
-	ret = amdgpu_bo_create_kernel(adev, PSP_FENCE_BUFFER_SIZE, PAGE_SIZE,
-					AMDGPU_GEM_DOMAIN_VRAM,
-					&psp->fence_buf_bo,
-					&psp->fence_buf_mc_addr,
-					&psp->fence_buf);
-	if (ret)
-		goto failed;
-
-	ret = amdgpu_bo_create_kernel(adev, PSP_CMD_BUFFER_SIZE, PAGE_SIZE,
-				      AMDGPU_GEM_DOMAIN_VRAM,
-				      &psp->cmd_buf_bo, &psp->cmd_buf_mc_addr,
-				      (void **)&psp->cmd_buf_mem);
-	if (ret)
-		goto failed;
-
-	memset(psp->fence_buf, 0, PSP_FENCE_BUFFER_SIZE);
+		memset(psp->fence_buf, 0, PSP_FENCE_BUFFER_SIZE);
 
-	ret = psp_ring_init(psp, PSP_RING_TYPE__KM);
-	if (ret) {
-		DRM_ERROR("PSP ring init failed!\n");
-		goto failed;
+		ret = psp_ring_init(psp, PSP_RING_TYPE__KM);
+		if (ret) {
+			dev_err(adev->dev, "PSP ring init failed!\n");
+			goto failed;
+		}
 	}
 
-skip_memalloc:
 	ret = psp_hw_start(psp);
 	if (ret)
 		goto failed;
 
 	ret = psp_load_non_psp_fw(psp);
 	if (ret)
-		goto failed;
+		goto failed1;
 
 	ret = psp_asd_initialize(psp);
 	if (ret) {
-		DRM_ERROR("PSP load asd failed!\n");
-		return ret;
+		dev_err(adev->dev, "PSP load asd failed!\n");
+		goto failed1;
 	}
 
 	ret = psp_rl_load(adev);
 	if (ret) {
-		DRM_ERROR("PSP load RL failed!\n");
-		return ret;
+		dev_err(adev->dev, "PSP load RL failed!\n");
+		goto failed1;
+	}
+
+	if (amdgpu_sriov_vf(adev) && amdgpu_in_reset(adev)) {
+		if (adev->gmc.xgmi.num_physical_nodes > 1) {
+			ret = psp_xgmi_initialize(psp, false, true);
+			/* Warning the XGMI seesion initialize failure
+			 * Instead of stop driver initialization
+			 */
+			if (ret)
+				dev_err(psp->adev->dev,
+					"XGMI: Failed to initialize XGMI session\n");
+		}
 	}
 
 	if (psp->ta_fw) {
 		ret = psp_ras_initialize(psp);
 		if (ret)
 			dev_err(psp->adev->dev,
-					"RAS: Failed to initialize RAS\n");
+				"RAS: Failed to initialize RAS\n");
 
 		ret = psp_hdcp_initialize(psp);
 		if (ret)
@@ -2481,32 +3185,32 @@ skip_memalloc:
 
 	return 0;
 
+failed1:
+	psp_free_shared_bufs(psp);
 failed:
 	/*
 	 * all cleanup jobs (xgmi terminate, ras terminate,
 	 * ring destroy, cmd/fence/fw buffers destory,
 	 * psp->cmd destory) are delayed to psp_hw_fini
 	 */
+	psp_ring_destroy(psp, PSP_RING_TYPE__KM);
 	return ret;
 }
 
-static int psp_hw_init(void *handle)
+static int psp_hw_init(struct amdgpu_ip_block *ip_block)
 {
 	int ret;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	mutex_lock(&adev->firmware.mutex);
-	/*
-	 * This sequence is just used on hw_init only once, no need on
-	 * resume.
-	 */
+
 	ret = amdgpu_ucode_init_bo(adev);
 	if (ret)
 		goto failed;
 
 	ret = psp_load_fw(adev);
 	if (ret) {
-		DRM_ERROR("PSP firmware loading failed\n");
+		dev_err(adev->dev, "PSP firmware loading failed\n");
 		goto failed;
 	}
 
@@ -2519,9 +3223,9 @@ failed:
 	return -EINVAL;
 }
 
-static int psp_hw_fini(void *handle)
+static int psp_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	struct psp_context *psp = &adev->psp;
 
 	if (psp->ta_fw) {
@@ -2530,105 +3234,104 @@ static int psp_hw_fini(void *handle)
 		psp_rap_terminate(psp);
 		psp_dtm_terminate(psp);
 		psp_hdcp_terminate(psp);
+
+		if (adev->gmc.xgmi.num_physical_nodes > 1)
+			psp_xgmi_terminate(psp);
 	}
 
 	psp_asd_terminate(psp);
-
 	psp_tmr_terminate(psp);
-	psp_ring_destroy(psp, PSP_RING_TYPE__KM);
 
-	amdgpu_bo_free_kernel(&psp->fw_pri_bo,
-			      &psp->fw_pri_mc_addr, &psp->fw_pri_buf);
-	amdgpu_bo_free_kernel(&psp->fence_buf_bo,
-			      &psp->fence_buf_mc_addr, &psp->fence_buf);
-	amdgpu_bo_free_kernel(&psp->cmd_buf_bo, &psp->cmd_buf_mc_addr,
-			      (void **)&psp->cmd_buf_mem);
+	psp_ring_destroy(psp, PSP_RING_TYPE__KM);
 
 	return 0;
 }
 
-static int psp_suspend(void *handle)
+static int psp_suspend(struct amdgpu_ip_block *ip_block)
 {
-	int ret;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	int ret = 0;
+	struct amdgpu_device *adev = ip_block->adev;
 	struct psp_context *psp = &adev->psp;
 
 	if (adev->gmc.xgmi.num_physical_nodes > 1 &&
 	    psp->xgmi_context.context.initialized) {
 		ret = psp_xgmi_terminate(psp);
 		if (ret) {
-			DRM_ERROR("Failed to terminate xgmi ta\n");
-			return ret;
+			dev_err(adev->dev, "Failed to terminate xgmi ta\n");
+			goto out;
 		}
 	}
 
 	if (psp->ta_fw) {
 		ret = psp_ras_terminate(psp);
 		if (ret) {
-			DRM_ERROR("Failed to terminate ras ta\n");
-			return ret;
+			dev_err(adev->dev, "Failed to terminate ras ta\n");
+			goto out;
 		}
 		ret = psp_hdcp_terminate(psp);
 		if (ret) {
-			DRM_ERROR("Failed to terminate hdcp ta\n");
-			return ret;
+			dev_err(adev->dev, "Failed to terminate hdcp ta\n");
+			goto out;
 		}
 		ret = psp_dtm_terminate(psp);
 		if (ret) {
-			DRM_ERROR("Failed to terminate dtm ta\n");
-			return ret;
+			dev_err(adev->dev, "Failed to terminate dtm ta\n");
+			goto out;
 		}
 		ret = psp_rap_terminate(psp);
 		if (ret) {
-			DRM_ERROR("Failed to terminate rap ta\n");
-			return ret;
+			dev_err(adev->dev, "Failed to terminate rap ta\n");
+			goto out;
 		}
 		ret = psp_securedisplay_terminate(psp);
 		if (ret) {
-			DRM_ERROR("Failed to terminate securedisplay ta\n");
-			return ret;
+			dev_err(adev->dev, "Failed to terminate securedisplay ta\n");
+			goto out;
 		}
 	}
 
 	ret = psp_asd_terminate(psp);
 	if (ret) {
-		DRM_ERROR("Failed to terminate asd\n");
-		return ret;
+		dev_err(adev->dev, "Failed to terminate asd\n");
+		goto out;
 	}
 
 	ret = psp_tmr_terminate(psp);
 	if (ret) {
-		DRM_ERROR("Failed to terminate tmr\n");
-		return ret;
+		dev_err(adev->dev, "Failed to terminate tmr\n");
+		goto out;
 	}
 
 	ret = psp_ring_stop(psp, PSP_RING_TYPE__KM);
-	if (ret) {
-		DRM_ERROR("PSP ring stop failed\n");
-		return ret;
-	}
+	if (ret)
+		dev_err(adev->dev, "PSP ring stop failed\n");
 
-	return 0;
+out:
+	return ret;
 }
 
-static int psp_resume(void *handle)
+static int psp_resume(struct amdgpu_ip_block *ip_block)
 {
 	int ret;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	struct psp_context *psp = &adev->psp;
 
-	DRM_INFO("PSP is resuming...\n");
+	dev_info(adev->dev, "PSP is resuming...\n");
 
 	if (psp->mem_train_ctx.enable_mem_training) {
 		ret = psp_mem_training(psp, PSP_MEM_TRAIN_RESUME);
 		if (ret) {
-			DRM_ERROR("Failed to process memory training!\n");
+			dev_err(adev->dev, "Failed to process memory training!\n");
 			return ret;
 		}
 	}
 
 	mutex_lock(&adev->firmware.mutex);
 
+	ret = amdgpu_ucode_init_bo(adev);
+	if (ret)
+		goto failed;
+
 	ret = psp_hw_start(psp);
 	if (ret)
 		goto failed;
@@ -2639,7 +3342,7 @@ static int psp_resume(void *handle)
 
 	ret = psp_asd_initialize(psp);
 	if (ret) {
-		DRM_ERROR("PSP load asd failed!\n");
+		dev_err(adev->dev, "PSP load asd failed!\n");
 		goto failed;
 	}
 
@@ -2663,7 +3366,7 @@ static int psp_resume(void *handle)
 		ret = psp_ras_initialize(psp);
 		if (ret)
 			dev_err(psp->adev->dev,
-					"RAS: Failed to initialize RAS\n");
+				"RAS: Failed to initialize RAS\n");
 
 		ret = psp_hdcp_initialize(psp);
 		if (ret)
@@ -2691,7 +3394,7 @@ static int psp_resume(void *handle)
 	return 0;
 
 failed:
-	DRM_ERROR("PSP resume failed\n");
+	dev_err(adev->dev, "PSP resume failed\n");
 	mutex_unlock(&adev->firmware.mutex);
 	return ret;
 }
@@ -2725,19 +3428,6 @@ int psp_rlc_autoload_start(struct psp_context *psp)
 	return ret;
 }
 
-int psp_update_vcn_sram(struct amdgpu_device *adev, int inst_idx,
-			uint64_t cmd_gpu_addr, int cmd_size)
-{
-	struct amdgpu_firmware_info ucode = {0};
-
-	ucode.ucode_id = inst_idx ? AMDGPU_UCODE_ID_VCN1_RAM :
-		AMDGPU_UCODE_ID_VCN0_RAM;
-	ucode.mc_addr = cmd_gpu_addr;
-	ucode.ucode_size = cmd_size;
-
-	return psp_execute_non_psp_fw_load(&adev->psp, &ucode);
-}
-
 int psp_ring_cmd_submit(struct psp_context *psp,
 			uint64_t cmd_buf_mc_addr,
 			uint64_t fence_mc_addr,
@@ -2765,9 +3455,11 @@ int psp_ring_cmd_submit(struct psp_context *psp,
 		write_frame = ring_buffer_start + (psp_write_ptr_reg / rb_frame_size_dw);
 	/* Check invalid write_frame ptr address */
 	if ((write_frame < ring_buffer_start) || (ring_buffer_end < write_frame)) {
-		DRM_ERROR("ring_buffer_start = %p; ring_buffer_end = %p; write_frame = %p\n",
-			  ring_buffer_start, ring_buffer_end, write_frame);
-		DRM_ERROR("write_frame is pointing to address out of bounds\n");
+		dev_err(adev->dev,
+			"ring_buffer_start = %p; ring_buffer_end = %p; write_frame = %p\n",
+			ring_buffer_start, ring_buffer_end, write_frame);
+		dev_err(adev->dev,
+			"write_frame is pointing to address out of bounds\n");
 		return -EINVAL;
 	}
 
@@ -2788,25 +3480,14 @@ int psp_ring_cmd_submit(struct psp_context *psp,
 	return 0;
 }
 
-int psp_init_asd_microcode(struct psp_context *psp,
-			   const char *chip_name)
+int psp_init_asd_microcode(struct psp_context *psp, const char *chip_name)
 {
 	struct amdgpu_device *adev = psp->adev;
-	char fw_name[PSP_FW_NAME_LEN];
 	const struct psp_firmware_header_v1_0 *asd_hdr;
 	int err = 0;
 
-	if (!chip_name) {
-		dev_err(adev->dev, "invalid chip name for asd microcode\n");
-		return -EINVAL;
-	}
-
-	snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_asd.bin", chip_name);
-	err = request_firmware(&adev->psp.asd_fw, fw_name, adev->dev);
-	if (err)
-		goto out;
-
-	err = amdgpu_ucode_validate(adev->psp.asd_fw);
+	err = amdgpu_ucode_request(adev, &adev->psp.asd_fw, AMDGPU_UCODE_REQUIRED,
+				   "amdgpu/%s_asd.bin", chip_name);
 	if (err)
 		goto out;
 
@@ -2818,31 +3499,18 @@ int psp_init_asd_microcode(struct psp_context *psp,
 				le32_to_cpu(asd_hdr->header.ucode_array_offset_bytes);
 	return 0;
 out:
-	dev_err(adev->dev, "fail to initialize asd microcode\n");
-	release_firmware(adev->psp.asd_fw);
-	adev->psp.asd_fw = NULL;
+	amdgpu_ucode_release(&adev->psp.asd_fw);
 	return err;
 }
 
-int psp_init_toc_microcode(struct psp_context *psp,
-			   const char *chip_name)
+int psp_init_toc_microcode(struct psp_context *psp, const char *chip_name)
 {
 	struct amdgpu_device *adev = psp->adev;
-	char fw_name[PSP_FW_NAME_LEN];
 	const struct psp_firmware_header_v1_0 *toc_hdr;
 	int err = 0;
 
-	if (!chip_name) {
-		dev_err(adev->dev, "invalid chip name for toc microcode\n");
-		return -EINVAL;
-	}
-
-	snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_toc.bin", chip_name);
-	err = request_firmware(&adev->psp.toc_fw, fw_name, adev->dev);
-	if (err)
-		goto out;
-
-	err = amdgpu_ucode_validate(adev->psp.toc_fw);
+	err = amdgpu_ucode_request(adev, &adev->psp.toc_fw, AMDGPU_UCODE_REQUIRED,
+				   "amdgpu/%s_toc.bin", chip_name);
 	if (err)
 		goto out;
 
@@ -2854,9 +3522,7 @@ int psp_init_toc_microcode(struct psp_context *psp,
 				le32_to_cpu(toc_hdr->header.ucode_array_offset_bytes);
 	return 0;
 out:
-	dev_err(adev->dev, "fail to request/validate toc microcode\n");
-	release_firmware(adev->psp.toc_fw);
-	adev->psp.toc_fw = NULL;
+	amdgpu_ucode_release(&adev->psp.toc_fw);
 	return err;
 }
 
@@ -2878,7 +3544,7 @@ static int parse_sos_bin_descriptor(struct psp_context *psp,
 		psp->sos.fw_version        = le32_to_cpu(desc->fw_version);
 		psp->sos.feature_version   = le32_to_cpu(desc->fw_version);
 		psp->sos.size_bytes        = le32_to_cpu(desc->size_bytes);
-		psp->sos.start_addr 	   = ucode_start_addr;
+		psp->sos.start_addr	   = ucode_start_addr;
 		break;
 	case PSP_FW_TYPE_PSP_SYS_DRV:
 		psp->sys.fw_version        = le32_to_cpu(desc->fw_version);
@@ -2928,6 +3594,24 @@ static int parse_sos_bin_descriptor(struct psp_context *psp,
 		psp->dbg_drv.size_bytes         = le32_to_cpu(desc->size_bytes);
 		psp->dbg_drv.start_addr         = ucode_start_addr;
 		break;
+	case PSP_FW_TYPE_PSP_RAS_DRV:
+		psp->ras_drv.fw_version         = le32_to_cpu(desc->fw_version);
+		psp->ras_drv.feature_version    = le32_to_cpu(desc->fw_version);
+		psp->ras_drv.size_bytes         = le32_to_cpu(desc->size_bytes);
+		psp->ras_drv.start_addr         = ucode_start_addr;
+		break;
+	case PSP_FW_TYPE_PSP_IPKEYMGR_DRV:
+		psp->ipkeymgr_drv.fw_version         = le32_to_cpu(desc->fw_version);
+		psp->ipkeymgr_drv.feature_version    = le32_to_cpu(desc->fw_version);
+		psp->ipkeymgr_drv.size_bytes         = le32_to_cpu(desc->size_bytes);
+		psp->ipkeymgr_drv.start_addr         = ucode_start_addr;
+		break;
+	case PSP_FW_TYPE_PSP_SPDM_DRV:
+		psp->spdm_drv.fw_version	= le32_to_cpu(desc->fw_version);
+		psp->spdm_drv.feature_version	= le32_to_cpu(desc->fw_version);
+		psp->spdm_drv.size_bytes	= le32_to_cpu(desc->size_bytes);
+		psp->spdm_drv.start_addr	= ucode_start_addr;
+		break;
 	default:
 		dev_warn(psp->adev->dev, "Unsupported PSP FW type: %d\n", desc->fw_type);
 		break;
@@ -2947,7 +3631,7 @@ static int psp_init_sos_base_fw(struct amdgpu_device *adev)
 		le32_to_cpu(sos_hdr->header.ucode_array_offset_bytes);
 
 	if (adev->gmc.xgmi.connected_to_cpu ||
-	    (adev->ip_versions[MP0_HWIP][0] != IP_VERSION(13, 0, 2))) {
+	    (amdgpu_ip_version(adev, MP0_HWIP, 0) != IP_VERSION(13, 0, 2))) {
 		adev->psp.sos.fw_version = le32_to_cpu(sos_hdr->header.ucode_version);
 		adev->psp.sos.feature_version = le32_to_cpu(sos_hdr->sos.fw_version);
 
@@ -2957,7 +3641,6 @@ static int psp_init_sos_base_fw(struct amdgpu_device *adev)
 		adev->psp.sos.size_bytes = le32_to_cpu(sos_hdr->sos.size_bytes);
 		adev->psp.sos.start_addr = ucode_array_start_addr +
 				le32_to_cpu(sos_hdr->sos.offset_bytes);
-		adev->psp.xgmi_context.supports_extended_data = false;
 	} else {
 		/* Load alternate PSP SOS FW */
 		sos_hdr_v1_3 = (const struct psp_firmware_header_v1_3 *)adev->psp.sos_fw->data;
@@ -2972,7 +3655,6 @@ static int psp_init_sos_base_fw(struct amdgpu_device *adev)
 		adev->psp.sos.size_bytes = le32_to_cpu(sos_hdr_v1_3->sos_aux.size_bytes);
 		adev->psp.sos.start_addr = ucode_array_start_addr +
 			le32_to_cpu(sos_hdr_v1_3->sos_aux.offset_bytes);
-		adev->psp.xgmi_context.supports_extended_data = true;
 	}
 
 	if ((adev->psp.sys.size_bytes == 0) || (adev->psp.sos.size_bytes == 0)) {
@@ -2983,31 +3665,26 @@ static int psp_init_sos_base_fw(struct amdgpu_device *adev)
 	return 0;
 }
 
-int psp_init_sos_microcode(struct psp_context *psp,
-			   const char *chip_name)
+int psp_init_sos_microcode(struct psp_context *psp, const char *chip_name)
 {
 	struct amdgpu_device *adev = psp->adev;
-	char fw_name[PSP_FW_NAME_LEN];
 	const struct psp_firmware_header_v1_0 *sos_hdr;
 	const struct psp_firmware_header_v1_1 *sos_hdr_v1_1;
 	const struct psp_firmware_header_v1_2 *sos_hdr_v1_2;
 	const struct psp_firmware_header_v1_3 *sos_hdr_v1_3;
 	const struct psp_firmware_header_v2_0 *sos_hdr_v2_0;
-	int err = 0;
+	const struct psp_firmware_header_v2_1 *sos_hdr_v2_1;
+	int fw_index, fw_bin_count, start_index = 0;
+	const struct psp_fw_bin_desc *fw_bin;
 	uint8_t *ucode_array_start_addr;
-	int fw_index = 0;
-
-	if (!chip_name) {
-		dev_err(adev->dev, "invalid chip name for sos microcode\n");
-		return -EINVAL;
-	}
-
-	snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_sos.bin", chip_name);
-	err = request_firmware(&adev->psp.sos_fw, fw_name, adev->dev);
-	if (err)
-		goto out;
+	int err = 0;
 
-	err = amdgpu_ucode_validate(adev->psp.sos_fw);
+	if (amdgpu_is_kicker_fw(adev))
+		err = amdgpu_ucode_request(adev, &adev->psp.sos_fw, AMDGPU_UCODE_REQUIRED,
+					   "amdgpu/%s_sos_kicker.bin", chip_name);
+	else
+		err = amdgpu_ucode_request(adev, &adev->psp.sos_fw, AMDGPU_UCODE_REQUIRED,
+					   "amdgpu/%s_sos.bin", chip_name);
 	if (err)
 		goto out;
 
@@ -3056,15 +3733,30 @@ int psp_init_sos_microcode(struct psp_context *psp,
 	case 2:
 		sos_hdr_v2_0 = (const struct psp_firmware_header_v2_0 *)adev->psp.sos_fw->data;
 
-		if (le32_to_cpu(sos_hdr_v2_0->psp_fw_bin_count) >= UCODE_MAX_PSP_PACKAGING) {
+		fw_bin_count = le32_to_cpu(sos_hdr_v2_0->psp_fw_bin_count);
+
+		if (fw_bin_count >= UCODE_MAX_PSP_PACKAGING) {
 			dev_err(adev->dev, "packed SOS count exceeds maximum limit\n");
 			err = -EINVAL;
 			goto out;
 		}
 
-		for (fw_index = 0; fw_index < le32_to_cpu(sos_hdr_v2_0->psp_fw_bin_count); fw_index++) {
-			err = parse_sos_bin_descriptor(psp,
-						       &sos_hdr_v2_0->psp_fw_bin[fw_index],
+		if (sos_hdr_v2_0->header.header_version_minor == 1) {
+			sos_hdr_v2_1 = (const struct psp_firmware_header_v2_1 *)adev->psp.sos_fw->data;
+
+			fw_bin = sos_hdr_v2_1->psp_fw_bin;
+
+			if (psp_is_aux_sos_load_required(psp))
+				start_index = le32_to_cpu(sos_hdr_v2_1->psp_aux_fw_bin_index);
+			else
+				fw_bin_count -= le32_to_cpu(sos_hdr_v2_1->psp_aux_fw_bin_index);
+
+		} else {
+			fw_bin = sos_hdr_v2_0->psp_fw_bin;
+		}
+
+		for (fw_index = start_index; fw_index < fw_bin_count; fw_index++) {
+			err = parse_sos_bin_descriptor(psp, fw_bin + fw_index,
 						       sos_hdr_v2_0);
 			if (err)
 				goto out;
@@ -3079,14 +3771,41 @@ int psp_init_sos_microcode(struct psp_context *psp,
 
 	return 0;
 out:
-	dev_err(adev->dev,
-		"failed to init sos firmware\n");
-	release_firmware(adev->psp.sos_fw);
-	adev->psp.sos_fw = NULL;
+	amdgpu_ucode_release(&adev->psp.sos_fw);
 
 	return err;
 }
 
+static bool is_ta_fw_applicable(struct psp_context *psp,
+			     const struct psp_fw_bin_desc *desc)
+{
+	struct amdgpu_device *adev = psp->adev;
+	uint32_t fw_version;
+
+	switch (desc->fw_type) {
+	case TA_FW_TYPE_PSP_XGMI:
+	case TA_FW_TYPE_PSP_XGMI_AUX:
+		/* for now, AUX TA only exists on 13.0.6 ta bin,
+		 * from v20.00.0x.14
+		 */
+		if (amdgpu_ip_version(adev, MP0_HWIP, 0) ==
+		    IP_VERSION(13, 0, 6)) {
+			fw_version = le32_to_cpu(desc->fw_version);
+
+			if (adev->flags & AMD_IS_APU &&
+			    (fw_version & 0xff) >= 0x14)
+				return desc->fw_type == TA_FW_TYPE_PSP_XGMI_AUX;
+			else
+				return desc->fw_type == TA_FW_TYPE_PSP_XGMI;
+		}
+		break;
+	default:
+		break;
+	}
+
+	return true;
+}
+
 static int parse_ta_bin_descriptor(struct psp_context *psp,
 				   const struct psp_fw_bin_desc *desc,
 				   const struct ta_firmware_header_v2_0 *ta_hdr)
@@ -3096,6 +3815,9 @@ static int parse_ta_bin_descriptor(struct psp_context *psp,
 	if (!psp || !desc || !ta_hdr)
 		return -EINVAL;
 
+	if (!is_ta_fw_applicable(psp, desc))
+		return 0;
+
 	ucode_start_addr  = (uint8_t *)ta_hdr +
 			    le32_to_cpu(desc->offset_bytes) +
 			    le32_to_cpu(ta_hdr->header.ucode_array_offset_bytes);
@@ -3108,6 +3830,7 @@ static int parse_ta_bin_descriptor(struct psp_context *psp,
 		psp->asd_context.bin_desc.start_addr        = ucode_start_addr;
 		break;
 	case TA_FW_TYPE_PSP_XGMI:
+	case TA_FW_TYPE_PSP_XGMI_AUX:
 		psp->xgmi_context.context.bin_desc.fw_version       = le32_to_cpu(desc->fw_version);
 		psp->xgmi_context.context.bin_desc.size_bytes       = le32_to_cpu(desc->size_bytes);
 		psp->xgmi_context.context.bin_desc.start_addr       = ucode_start_addr;
@@ -3148,41 +3871,76 @@ static int parse_ta_bin_descriptor(struct psp_context *psp,
 	return 0;
 }
 
-int psp_init_ta_microcode(struct psp_context *psp,
-			  const char *chip_name)
+static int parse_ta_v1_microcode(struct psp_context *psp)
 {
+	const struct ta_firmware_header_v1_0 *ta_hdr;
 	struct amdgpu_device *adev = psp->adev;
-	char fw_name[PSP_FW_NAME_LEN];
-	const struct ta_firmware_header_v2_0 *ta_hdr;
-	int err = 0;
-	int ta_index = 0;
 
-	if (!chip_name) {
-		dev_err(adev->dev, "invalid chip name for ta microcode\n");
+	ta_hdr = (const struct ta_firmware_header_v1_0 *) adev->psp.ta_fw->data;
+
+	if (le16_to_cpu(ta_hdr->header.header_version_major) != 1)
 		return -EINVAL;
-	}
 
-	snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_ta.bin", chip_name);
-	err = request_firmware(&adev->psp.ta_fw, fw_name, adev->dev);
-	if (err)
-		goto out;
+	adev->psp.xgmi_context.context.bin_desc.fw_version =
+		le32_to_cpu(ta_hdr->xgmi.fw_version);
+	adev->psp.xgmi_context.context.bin_desc.size_bytes =
+		le32_to_cpu(ta_hdr->xgmi.size_bytes);
+	adev->psp.xgmi_context.context.bin_desc.start_addr =
+		(uint8_t *)ta_hdr +
+		le32_to_cpu(ta_hdr->header.ucode_array_offset_bytes);
+
+	adev->psp.ras_context.context.bin_desc.fw_version =
+		le32_to_cpu(ta_hdr->ras.fw_version);
+	adev->psp.ras_context.context.bin_desc.size_bytes =
+		le32_to_cpu(ta_hdr->ras.size_bytes);
+	adev->psp.ras_context.context.bin_desc.start_addr =
+		(uint8_t *)adev->psp.xgmi_context.context.bin_desc.start_addr +
+		le32_to_cpu(ta_hdr->ras.offset_bytes);
+
+	adev->psp.hdcp_context.context.bin_desc.fw_version =
+		le32_to_cpu(ta_hdr->hdcp.fw_version);
+	adev->psp.hdcp_context.context.bin_desc.size_bytes =
+		le32_to_cpu(ta_hdr->hdcp.size_bytes);
+	adev->psp.hdcp_context.context.bin_desc.start_addr =
+		(uint8_t *)ta_hdr +
+		le32_to_cpu(ta_hdr->header.ucode_array_offset_bytes);
+
+	adev->psp.dtm_context.context.bin_desc.fw_version =
+		le32_to_cpu(ta_hdr->dtm.fw_version);
+	adev->psp.dtm_context.context.bin_desc.size_bytes =
+		le32_to_cpu(ta_hdr->dtm.size_bytes);
+	adev->psp.dtm_context.context.bin_desc.start_addr =
+		(uint8_t *)adev->psp.hdcp_context.context.bin_desc.start_addr +
+		le32_to_cpu(ta_hdr->dtm.offset_bytes);
+
+	adev->psp.securedisplay_context.context.bin_desc.fw_version =
+		le32_to_cpu(ta_hdr->securedisplay.fw_version);
+	adev->psp.securedisplay_context.context.bin_desc.size_bytes =
+		le32_to_cpu(ta_hdr->securedisplay.size_bytes);
+	adev->psp.securedisplay_context.context.bin_desc.start_addr =
+		(uint8_t *)adev->psp.hdcp_context.context.bin_desc.start_addr +
+		le32_to_cpu(ta_hdr->securedisplay.offset_bytes);
+
+	adev->psp.ta_fw_version = le32_to_cpu(ta_hdr->header.ucode_version);
 
-	err = amdgpu_ucode_validate(adev->psp.ta_fw);
-	if (err)
-		goto out;
+	return 0;
+}
+
+static int parse_ta_v2_microcode(struct psp_context *psp)
+{
+	const struct ta_firmware_header_v2_0 *ta_hdr;
+	struct amdgpu_device *adev = psp->adev;
+	int err = 0;
+	int ta_index = 0;
 
 	ta_hdr = (const struct ta_firmware_header_v2_0 *)adev->psp.ta_fw->data;
 
-	if (le16_to_cpu(ta_hdr->header.header_version_major) != 2) {
-		dev_err(adev->dev, "unsupported TA header version\n");
-		err = -EINVAL;
-		goto out;
-	}
+	if (le16_to_cpu(ta_hdr->header.header_version_major) != 2)
+		return -EINVAL;
 
 	if (le32_to_cpu(ta_hdr->ta_fw_bin_count) >= UCODE_MAX_PSP_PACKAGING) {
 		dev_err(adev->dev, "packed TA count exceeds maximum limit\n");
-		err = -EINVAL;
-		goto out;
+		return -EINVAL;
 	}
 
 	for (ta_index = 0; ta_index < le32_to_cpu(ta_hdr->ta_fw_bin_count); ta_index++) {
@@ -3190,24 +3948,131 @@ int psp_init_ta_microcode(struct psp_context *psp,
 					      &ta_hdr->ta_fw_bin[ta_index],
 					      ta_hdr);
 		if (err)
-			goto out;
+			return err;
 	}
 
 	return 0;
+}
+
+int psp_init_ta_microcode(struct psp_context *psp, const char *chip_name)
+{
+	const struct common_firmware_header *hdr;
+	struct amdgpu_device *adev = psp->adev;
+	int err;
+
+	if (amdgpu_is_kicker_fw(adev))
+		err = amdgpu_ucode_request(adev, &adev->psp.ta_fw, AMDGPU_UCODE_REQUIRED,
+					   "amdgpu/%s_ta_kicker.bin", chip_name);
+	else
+		err = amdgpu_ucode_request(adev, &adev->psp.ta_fw, AMDGPU_UCODE_REQUIRED,
+					   "amdgpu/%s_ta.bin", chip_name);
+	if (err)
+		return err;
+
+	hdr = (const struct common_firmware_header *)adev->psp.ta_fw->data;
+	switch (le16_to_cpu(hdr->header_version_major)) {
+	case 1:
+		err = parse_ta_v1_microcode(psp);
+		break;
+	case 2:
+		err = parse_ta_v2_microcode(psp);
+		break;
+	default:
+		dev_err(adev->dev, "unsupported TA header version\n");
+		err = -EINVAL;
+	}
+
+	if (err)
+		amdgpu_ucode_release(&adev->psp.ta_fw);
+
+	return err;
+}
+
+int psp_init_cap_microcode(struct psp_context *psp, const char *chip_name)
+{
+	struct amdgpu_device *adev = psp->adev;
+	const struct psp_firmware_header_v1_0 *cap_hdr_v1_0;
+	struct amdgpu_firmware_info *info = NULL;
+	int err = 0;
+
+	if (!amdgpu_sriov_vf(adev)) {
+		dev_err(adev->dev, "cap microcode should only be loaded under SRIOV\n");
+		return -EINVAL;
+	}
+
+	err = amdgpu_ucode_request(adev, &adev->psp.cap_fw, AMDGPU_UCODE_OPTIONAL,
+				   "amdgpu/%s_cap.bin", chip_name);
+	if (err) {
+		if (err == -ENODEV) {
+			dev_warn(adev->dev, "cap microcode does not exist, skip\n");
+			err = 0;
+		} else {
+			dev_err(adev->dev, "fail to initialize cap microcode\n");
+		}
+		goto out;
+	}
+
+	info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CAP];
+	info->ucode_id = AMDGPU_UCODE_ID_CAP;
+	info->fw = adev->psp.cap_fw;
+	cap_hdr_v1_0 = (const struct psp_firmware_header_v1_0 *)
+		adev->psp.cap_fw->data;
+	adev->firmware.fw_size += ALIGN(
+			le32_to_cpu(cap_hdr_v1_0->header.ucode_size_bytes), PAGE_SIZE);
+	adev->psp.cap_fw_version = le32_to_cpu(cap_hdr_v1_0->header.ucode_version);
+	adev->psp.cap_feature_version = le32_to_cpu(cap_hdr_v1_0->sos.fw_version);
+	adev->psp.cap_ucode_size = le32_to_cpu(cap_hdr_v1_0->header.ucode_size_bytes);
+
+	return 0;
+
 out:
-	dev_err(adev->dev, "fail to initialize ta microcode\n");
-	release_firmware(adev->psp.ta_fw);
-	adev->psp.ta_fw = NULL;
+	amdgpu_ucode_release(&adev->psp.cap_fw);
 	return err;
 }
 
-static int psp_set_clockgating_state(void *handle,
-				     enum amd_clockgating_state state)
+int psp_config_sq_perfmon(struct psp_context *psp,
+		uint32_t xcp_id, bool core_override_enable,
+		bool reg_override_enable, bool perfmon_override_enable)
+{
+	int ret;
+
+	if (amdgpu_sriov_vf(psp->adev))
+		return 0;
+
+	if (xcp_id > MAX_XCP) {
+		dev_err(psp->adev->dev, "invalid xcp_id %d\n", xcp_id);
+		return -EINVAL;
+	}
+
+	if (amdgpu_ip_version(psp->adev, MP0_HWIP, 0) != IP_VERSION(13, 0, 6)) {
+		dev_err(psp->adev->dev, "Unsupported MP0 version 0x%x for CONFIG_SQ_PERFMON command\n",
+			amdgpu_ip_version(psp->adev, MP0_HWIP, 0));
+		return -EINVAL;
+	}
+	struct psp_gfx_cmd_resp *cmd = acquire_psp_cmd_buf(psp);
+
+	cmd->cmd_id	=	GFX_CMD_ID_CONFIG_SQ_PERFMON;
+	cmd->cmd.config_sq_perfmon.gfx_xcp_mask	=	BIT_MASK(xcp_id);
+	cmd->cmd.config_sq_perfmon.core_override	=	core_override_enable;
+	cmd->cmd.config_sq_perfmon.reg_override	=	reg_override_enable;
+	cmd->cmd.config_sq_perfmon.perfmon_override = perfmon_override_enable;
+
+	ret = psp_cmd_submit_buf(psp, NULL, cmd, psp->fence_buf_mc_addr);
+	if (ret)
+		dev_warn(psp->adev->dev, "PSP failed to config sq: xcp%d core%d reg%d perfmon%d\n",
+			xcp_id, core_override_enable, reg_override_enable, perfmon_override_enable);
+
+	release_psp_cmd_buf(psp);
+	return ret;
+}
+
+static int psp_set_clockgating_state(struct amdgpu_ip_block *ip_block,
+					enum amd_clockgating_state state)
 {
 	return 0;
 }
 
-static int psp_set_powergating_state(void *handle,
+static int psp_set_powergating_state(struct amdgpu_ip_block *ip_block,
 				     enum amd_powergating_state state)
 {
 	return 0;
@@ -3219,11 +4084,13 @@ static ssize_t psp_usbc_pd_fw_sysfs_read(struct device *dev,
 {
 	struct drm_device *ddev = dev_get_drvdata(dev);
 	struct amdgpu_device *adev = drm_to_adev(ddev);
+	struct amdgpu_ip_block *ip_block;
 	uint32_t fw_ver;
 	int ret;
 
-	if (!adev->ip_blocks[AMD_IP_BLOCK_TYPE_PSP].status.late_initialized) {
-		DRM_INFO("PSP block is not ready yet.");
+	ip_block = amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_PSP);
+	if (!ip_block || !ip_block->status.late_initialized) {
+		dev_info(adev->dev, "PSP block is not ready yet\n.");
 		return -EBUSY;
 	}
 
@@ -3232,7 +4099,7 @@ static ssize_t psp_usbc_pd_fw_sysfs_read(struct device *dev,
 	mutex_unlock(&adev->psp.mutex);
 
 	if (ret) {
-		DRM_ERROR("Failed to read USBC PD FW, err = %d", ret);
+		dev_err(adev->dev, "Failed to read USBC PD FW, err = %d\n", ret);
 		return ret;
 	}
 
@@ -3247,31 +4114,32 @@ static ssize_t psp_usbc_pd_fw_sysfs_write(struct device *dev,
 	struct drm_device *ddev = dev_get_drvdata(dev);
 	struct amdgpu_device *adev = drm_to_adev(ddev);
 	int ret, idx;
-	char fw_name[100];
 	const struct firmware *usbc_pd_fw;
 	struct amdgpu_bo *fw_buf_bo = NULL;
 	uint64_t fw_pri_mc_addr;
 	void *fw_pri_cpu_addr;
+	struct amdgpu_ip_block *ip_block;
 
-	if (!adev->ip_blocks[AMD_IP_BLOCK_TYPE_PSP].status.late_initialized) {
-		DRM_INFO("PSP block is not ready yet.");
+	ip_block = amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_PSP);
+	if (!ip_block || !ip_block->status.late_initialized) {
+		dev_err(adev->dev, "PSP block is not ready yet.");
 		return -EBUSY;
 	}
 
 	if (!drm_dev_enter(ddev, &idx))
 		return -ENODEV;
 
-	snprintf(fw_name, sizeof(fw_name), "amdgpu/%s", buf);
-	ret = request_firmware(&usbc_pd_fw, fw_name, adev->dev);
+	ret = amdgpu_ucode_request(adev, &usbc_pd_fw, AMDGPU_UCODE_REQUIRED,
+				   "amdgpu/%s", buf);
 	if (ret)
 		goto fail;
 
 	/* LFB address which is aligned to 1MB boundary per PSP request */
 	ret = amdgpu_bo_create_kernel(adev, usbc_pd_fw->size, 0x100000,
-						AMDGPU_GEM_DOMAIN_VRAM,
-						&fw_buf_bo,
-						&fw_pri_mc_addr,
-						&fw_pri_cpu_addr);
+				      AMDGPU_GEM_DOMAIN_VRAM |
+				      AMDGPU_GEM_DOMAIN_GTT,
+				      &fw_buf_bo, &fw_pri_mc_addr,
+				      &fw_pri_cpu_addr);
 	if (ret)
 		goto rel_buf;
 
@@ -3284,10 +4152,10 @@ static ssize_t psp_usbc_pd_fw_sysfs_write(struct device *dev,
 	amdgpu_bo_free_kernel(&fw_buf_bo, &fw_pri_mc_addr, &fw_pri_cpu_addr);
 
 rel_buf:
-	release_firmware(usbc_pd_fw);
+	amdgpu_ucode_release(&usbc_pd_fw);
 fail:
 	if (ret) {
-		DRM_ERROR("Failed to load USBC PD FW, err = %d", ret);
+		dev_err(adev->dev, "Failed to load USBC PD FW, err = %d", ret);
 		count = ret;
 	}
 
@@ -3308,7 +4176,12 @@ void psp_copy_fw(struct psp_context *psp, uint8_t *start_addr, uint32_t bin_size
 	drm_dev_exit(idx);
 }
 
-static DEVICE_ATTR(usbc_pd_fw, S_IRUGO | S_IWUSR,
+/**
+ * DOC: usbc_pd_fw
+ * Reading from this file will retrieve the USB-C PD firmware version. Writing to
+ * this file will trigger the update process.
+ */
+static DEVICE_ATTR(usbc_pd_fw, 0644,
 		   psp_usbc_pd_fw_sysfs_read,
 		   psp_usbc_pd_fw_sysfs_write);
 
@@ -3317,41 +4190,285 @@ int is_psp_fw_valid(struct psp_bin_desc bin)
 	return bin.size_bytes;
 }
 
-const struct amd_ip_funcs psp_ip_funcs = {
-	.name = "psp",
-	.early_init = psp_early_init,
-	.late_init = NULL,
-	.sw_init = psp_sw_init,
-	.sw_fini = psp_sw_fini,
-	.hw_init = psp_hw_init,
-	.hw_fini = psp_hw_fini,
-	.suspend = psp_suspend,
-	.resume = psp_resume,
-	.is_idle = NULL,
-	.check_soft_reset = NULL,
-	.wait_for_idle = NULL,
-	.soft_reset = NULL,
-	.set_clockgating_state = psp_set_clockgating_state,
-	.set_powergating_state = psp_set_powergating_state,
+static ssize_t amdgpu_psp_vbflash_write(struct file *filp, struct kobject *kobj,
+					const struct bin_attribute *bin_attr,
+					char *buffer, loff_t pos, size_t count)
+{
+	struct device *dev = kobj_to_dev(kobj);
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = drm_to_adev(ddev);
+
+	adev->psp.vbflash_done = false;
+
+	/* Safeguard against memory drain */
+	if (adev->psp.vbflash_image_size > AMD_VBIOS_FILE_MAX_SIZE_B) {
+		dev_err(adev->dev, "File size cannot exceed %u\n", AMD_VBIOS_FILE_MAX_SIZE_B);
+		kvfree(adev->psp.vbflash_tmp_buf);
+		adev->psp.vbflash_tmp_buf = NULL;
+		adev->psp.vbflash_image_size = 0;
+		return -ENOMEM;
+	}
+
+	/* TODO Just allocate max for now and optimize to realloc later if needed */
+	if (!adev->psp.vbflash_tmp_buf) {
+		adev->psp.vbflash_tmp_buf = kvmalloc(AMD_VBIOS_FILE_MAX_SIZE_B, GFP_KERNEL);
+		if (!adev->psp.vbflash_tmp_buf)
+			return -ENOMEM;
+	}
+
+	mutex_lock(&adev->psp.mutex);
+	memcpy(adev->psp.vbflash_tmp_buf + pos, buffer, count);
+	adev->psp.vbflash_image_size += count;
+	mutex_unlock(&adev->psp.mutex);
+
+	dev_dbg(adev->dev, "IFWI staged for update\n");
+
+	return count;
+}
+
+static ssize_t amdgpu_psp_vbflash_read(struct file *filp, struct kobject *kobj,
+				       const struct bin_attribute *bin_attr, char *buffer,
+				       loff_t pos, size_t count)
+{
+	struct device *dev = kobj_to_dev(kobj);
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = drm_to_adev(ddev);
+	struct amdgpu_bo *fw_buf_bo = NULL;
+	uint64_t fw_pri_mc_addr;
+	void *fw_pri_cpu_addr;
+	int ret;
+
+	if (adev->psp.vbflash_image_size == 0)
+		return -EINVAL;
+
+	dev_dbg(adev->dev, "PSP IFWI flash process initiated\n");
+
+	ret = amdgpu_bo_create_kernel(adev, adev->psp.vbflash_image_size,
+					AMDGPU_GPU_PAGE_SIZE,
+					AMDGPU_GEM_DOMAIN_VRAM,
+					&fw_buf_bo,
+					&fw_pri_mc_addr,
+					&fw_pri_cpu_addr);
+	if (ret)
+		goto rel_buf;
+
+	memcpy_toio(fw_pri_cpu_addr, adev->psp.vbflash_tmp_buf, adev->psp.vbflash_image_size);
+
+	mutex_lock(&adev->psp.mutex);
+	ret = psp_update_spirom(&adev->psp, fw_pri_mc_addr);
+	mutex_unlock(&adev->psp.mutex);
+
+	amdgpu_bo_free_kernel(&fw_buf_bo, &fw_pri_mc_addr, &fw_pri_cpu_addr);
+
+rel_buf:
+	kvfree(adev->psp.vbflash_tmp_buf);
+	adev->psp.vbflash_tmp_buf = NULL;
+	adev->psp.vbflash_image_size = 0;
+
+	if (ret) {
+		dev_err(adev->dev, "Failed to load IFWI, err = %d\n", ret);
+		return ret;
+	}
+
+	dev_dbg(adev->dev, "PSP IFWI flash process done\n");
+	return 0;
+}
+
+/**
+ * DOC: psp_vbflash
+ * Writing to this file will stage an IFWI for update. Reading from this file
+ * will trigger the update process.
+ */
+static const struct bin_attribute psp_vbflash_bin_attr = {
+	.attr = {.name = "psp_vbflash", .mode = 0660},
+	.size = 0,
+	.write = amdgpu_psp_vbflash_write,
+	.read = amdgpu_psp_vbflash_read,
+};
+
+/**
+ * DOC: psp_vbflash_status
+ * The status of the flash process.
+ * 0: IFWI flash not complete.
+ * 1: IFWI flash complete.
+ */
+static ssize_t amdgpu_psp_vbflash_status(struct device *dev,
+					 struct device_attribute *attr,
+					 char *buf)
+{
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = drm_to_adev(ddev);
+	uint32_t vbflash_status;
+
+	vbflash_status = psp_vbflash_status(&adev->psp);
+	if (!adev->psp.vbflash_done)
+		vbflash_status = 0;
+	else if (adev->psp.vbflash_done && !(vbflash_status & 0x80000000))
+		vbflash_status = 1;
+
+	return sysfs_emit(buf, "0x%x\n", vbflash_status);
+}
+static DEVICE_ATTR(psp_vbflash_status, 0440, amdgpu_psp_vbflash_status, NULL);
+
+static const struct bin_attribute *const bin_flash_attrs[] = {
+	&psp_vbflash_bin_attr,
+	NULL
+};
+
+static struct attribute *flash_attrs[] = {
+	&dev_attr_psp_vbflash_status.attr,
+	&dev_attr_usbc_pd_fw.attr,
+	NULL
+};
+
+static umode_t amdgpu_flash_attr_is_visible(struct kobject *kobj, struct attribute *attr, int idx)
+{
+	struct device *dev = kobj_to_dev(kobj);
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = drm_to_adev(ddev);
+
+	if (attr == &dev_attr_usbc_pd_fw.attr)
+		return adev->psp.sup_pd_fw_up ? 0660 : 0;
+
+	return adev->psp.sup_ifwi_up ? 0440 : 0;
+}
+
+static umode_t amdgpu_bin_flash_attr_is_visible(struct kobject *kobj,
+						const struct bin_attribute *attr,
+						int idx)
+{
+	struct device *dev = kobj_to_dev(kobj);
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = drm_to_adev(ddev);
+
+	return adev->psp.sup_ifwi_up ? 0660 : 0;
+}
+
+const struct attribute_group amdgpu_flash_attr_group = {
+	.attrs = flash_attrs,
+	.bin_attrs = bin_flash_attrs,
+	.is_bin_visible = amdgpu_bin_flash_attr_is_visible,
+	.is_visible = amdgpu_flash_attr_is_visible,
 };
 
-static int psp_sysfs_init(struct amdgpu_device *adev)
+#if defined(CONFIG_DEBUG_FS)
+static int psp_read_spirom_debugfs_open(struct inode *inode, struct file *filp)
 {
-	int ret = device_create_file(adev->dev, &dev_attr_usbc_pd_fw);
+	struct amdgpu_device *adev = filp->f_inode->i_private;
+	struct spirom_bo *bo_triplet;
+	int ret;
+
+	/* serialize the open() file calling */
+	if (!mutex_trylock(&adev->psp.mutex))
+		return -EBUSY;
+
+	/*
+	 * make sure only one userpace process is alive for dumping so that
+	 * only one memory buffer of AMD_VBIOS_FILE_MAX_SIZE * 2 is consumed.
+	 * let's say the case where one process try opening the file while
+	 * another one has proceeded to read or release. In this way, eliminate
+	 * the use of mutex for read() or release() callback as well.
+	 */
+	if (adev->psp.spirom_dump_trip) {
+		mutex_unlock(&adev->psp.mutex);
+		return -EBUSY;
+	}
+
+	bo_triplet = kzalloc(sizeof(struct spirom_bo), GFP_KERNEL);
+	if (!bo_triplet) {
+		mutex_unlock(&adev->psp.mutex);
+		return -ENOMEM;
+	}
+
+	ret = amdgpu_bo_create_kernel(adev, AMD_VBIOS_FILE_MAX_SIZE_B * 2,
+				      AMDGPU_GPU_PAGE_SIZE,
+				      AMDGPU_GEM_DOMAIN_GTT,
+				      &bo_triplet->bo,
+				      &bo_triplet->mc_addr,
+				      &bo_triplet->cpu_addr);
+	if (ret)
+		goto rel_trip;
 
+	ret = psp_dump_spirom(&adev->psp, bo_triplet->mc_addr);
 	if (ret)
-		DRM_ERROR("Failed to create USBC PD FW control file!");
+		goto rel_bo;
 
+	adev->psp.spirom_dump_trip = bo_triplet;
+	mutex_unlock(&adev->psp.mutex);
+	return 0;
+rel_bo:
+	amdgpu_bo_free_kernel(&bo_triplet->bo, &bo_triplet->mc_addr,
+			      &bo_triplet->cpu_addr);
+rel_trip:
+	kfree(bo_triplet);
+	mutex_unlock(&adev->psp.mutex);
+	dev_err(adev->dev, "Trying IFWI dump fails, err = %d\n", ret);
 	return ret;
 }
 
-static void psp_sysfs_fini(struct amdgpu_device *adev)
+static ssize_t psp_read_spirom_debugfs_read(struct file *filp, char __user *buf, size_t size,
+					    loff_t *pos)
 {
-	device_remove_file(adev->dev, &dev_attr_usbc_pd_fw);
+	struct amdgpu_device *adev = filp->f_inode->i_private;
+	struct spirom_bo *bo_triplet = adev->psp.spirom_dump_trip;
+
+	if (!bo_triplet)
+		return -EINVAL;
+
+	return simple_read_from_buffer(buf,
+				       size,
+				       pos, bo_triplet->cpu_addr,
+				       AMD_VBIOS_FILE_MAX_SIZE_B * 2);
 }
 
-const struct amdgpu_ip_block_version psp_v3_1_ip_block =
+static int psp_read_spirom_debugfs_release(struct inode *inode, struct file *filp)
+{
+	struct amdgpu_device *adev = filp->f_inode->i_private;
+	struct spirom_bo *bo_triplet = adev->psp.spirom_dump_trip;
+
+	if (bo_triplet) {
+		amdgpu_bo_free_kernel(&bo_triplet->bo, &bo_triplet->mc_addr,
+				      &bo_triplet->cpu_addr);
+		kfree(bo_triplet);
+	}
+
+	adev->psp.spirom_dump_trip = NULL;
+	return 0;
+}
+
+static const struct file_operations psp_dump_spirom_debugfs_ops = {
+	.owner = THIS_MODULE,
+	.open = psp_read_spirom_debugfs_open,
+	.read = psp_read_spirom_debugfs_read,
+	.release = psp_read_spirom_debugfs_release,
+	.llseek = default_llseek,
+};
+#endif
+
+void amdgpu_psp_debugfs_init(struct amdgpu_device *adev)
 {
+#if defined(CONFIG_DEBUG_FS)
+	struct drm_minor *minor = adev_to_drm(adev)->primary;
+
+	debugfs_create_file_size("psp_spirom_dump", 0444, minor->debugfs_root,
+				 adev, &psp_dump_spirom_debugfs_ops, AMD_VBIOS_FILE_MAX_SIZE_B * 2);
+#endif
+}
+
+const struct amd_ip_funcs psp_ip_funcs = {
+	.name = "psp",
+	.early_init = psp_early_init,
+	.sw_init = psp_sw_init,
+	.sw_fini = psp_sw_fini,
+	.hw_init = psp_hw_init,
+	.hw_fini = psp_hw_fini,
+	.suspend = psp_suspend,
+	.resume = psp_resume,
+	.set_clockgating_state = psp_set_clockgating_state,
+	.set_powergating_state = psp_set_powergating_state,
+};
+
+const struct amdgpu_ip_block_version psp_v3_1_ip_block = {
 	.type = AMD_IP_BLOCK_TYPE_PSP,
 	.major = 3,
 	.minor = 1,
@@ -3359,8 +4476,7 @@ const struct amdgpu_ip_block_version psp_v3_1_ip_block =
 	.funcs = &psp_ip_funcs,
 };
 
-const struct amdgpu_ip_block_version psp_v10_0_ip_block =
-{
+const struct amdgpu_ip_block_version psp_v10_0_ip_block = {
 	.type = AMD_IP_BLOCK_TYPE_PSP,
 	.major = 10,
 	.minor = 0,
@@ -3368,8 +4484,7 @@ const struct amdgpu_ip_block_version psp_v10_0_ip_block =
 	.funcs = &psp_ip_funcs,
 };
 
-const struct amdgpu_ip_block_version psp_v11_0_ip_block =
-{
+const struct amdgpu_ip_block_version psp_v11_0_ip_block = {
 	.type = AMD_IP_BLOCK_TYPE_PSP,
 	.major = 11,
 	.minor = 0,
@@ -3385,8 +4500,7 @@ const struct amdgpu_ip_block_version psp_v11_0_8_ip_block = {
 	.funcs = &psp_ip_funcs,
 };
 
-const struct amdgpu_ip_block_version psp_v12_0_ip_block =
-{
+const struct amdgpu_ip_block_version psp_v12_0_ip_block = {
 	.type = AMD_IP_BLOCK_TYPE_PSP,
 	.major = 12,
 	.minor = 0,
@@ -3401,3 +4515,19 @@ const struct amdgpu_ip_block_version psp_v13_0_ip_block = {
 	.rev = 0,
 	.funcs = &psp_ip_funcs,
 };
+
+const struct amdgpu_ip_block_version psp_v13_0_4_ip_block = {
+	.type = AMD_IP_BLOCK_TYPE_PSP,
+	.major = 13,
+	.minor = 0,
+	.rev = 4,
+	.funcs = &psp_ip_funcs,
+};
+
+const struct amdgpu_ip_block_version psp_v14_0_ip_block = {
+	.type = AMD_IP_BLOCK_TYPE_PSP,
+	.major = 14,
+	.minor = 0,
+	.rev = 0,
+	.funcs = &psp_ip_funcs,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h
index f29afabbff1f..237b624aa51c 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h
@@ -36,8 +36,34 @@
 #define PSP_CMD_BUFFER_SIZE	0x1000
 #define PSP_1_MEG		0x100000
 #define PSP_TMR_SIZE(adev)	((adev)->asic_type == CHIP_ALDEBARAN ? 0x800000 : 0x400000)
+#define PSP_TMR_ALIGNMENT	0x100000
 #define PSP_FW_NAME_LEN		0x24
 
+/* VBIOS gfl defines */
+#define MBOX_READY_MASK 0x80000000
+#define MBOX_STATUS_MASK 0x0000FFFF
+#define MBOX_COMMAND_MASK 0x00FF0000
+#define MBOX_READY_FLAG 0x80000000
+#define C2PMSG_CMD_SPI_UPDATE_ROM_IMAGE_ADDR_LO 0x2
+#define C2PMSG_CMD_SPI_UPDATE_ROM_IMAGE_ADDR_HI 0x3
+#define C2PMSG_CMD_SPI_UPDATE_FLASH_IMAGE 0x4
+#define C2PMSG_CMD_SPI_GET_ROM_IMAGE_ADDR_LO 0xf
+#define C2PMSG_CMD_SPI_GET_ROM_IMAGE_ADDR_HI 0x10
+#define C2PMSG_CMD_SPI_GET_FLASH_IMAGE 0x11
+
+/* Command register bit 31 set to indicate readiness */
+#define MBOX_TOS_READY_FLAG (GFX_FLAG_RESPONSE)
+#define MBOX_TOS_READY_MASK (GFX_CMD_RESPONSE_MASK | GFX_CMD_STATUS_MASK)
+
+/* Values to check for a successful GFX_CMD response wait. Check against
+ * both status bits and response state - helps to detect a command failure
+ * or other unexpected cases like a device drop reading all 0xFFs
+ */
+#define MBOX_TOS_RESP_FLAG (GFX_FLAG_RESPONSE)
+#define MBOX_TOS_RESP_MASK (GFX_CMD_RESPONSE_MASK | GFX_CMD_STATUS_MASK)
+
+extern const struct attribute_group amdgpu_flash_attr_group;
+
 enum psp_shared_mem_size {
 	PSP_ASD_SHARED_MEM_SIZE				= 0x0,
 	PSP_XGMI_SHARED_MEM_SIZE			= 0x4000,
@@ -48,6 +74,17 @@ enum psp_shared_mem_size {
 	PSP_SECUREDISPLAY_SHARED_MEM_SIZE	= 0x4000,
 };
 
+enum ta_type_id {
+	TA_TYPE_XGMI = 1,
+	TA_TYPE_RAS,
+	TA_TYPE_HDCP,
+	TA_TYPE_DTM,
+	TA_TYPE_RAP,
+	TA_TYPE_SECUREDISPLAY,
+
+	TA_TYPE_MAX_INDEX,
+};
+
 struct psp_context;
 struct psp_xgmi_node_info;
 struct psp_xgmi_topology_info;
@@ -58,15 +95,18 @@ enum psp_bootloader_cmd {
 	PSP_BL__LOAD_SOSDRV		= 0x20000,
 	PSP_BL__LOAD_KEY_DATABASE	= 0x80000,
 	PSP_BL__LOAD_SOCDRV             = 0xB0000,
-	PSP_BL__LOAD_INTFDRV            = 0xC0000,
-	PSP_BL__LOAD_DBGDRV             = 0xD0000,
+	PSP_BL__LOAD_DBGDRV             = 0xC0000,
+	PSP_BL__LOAD_HADDRV		= PSP_BL__LOAD_DBGDRV,
+	PSP_BL__LOAD_INTFDRV		= 0xD0000,
+	PSP_BL__LOAD_RASDRV		= 0xE0000,
+	PSP_BL__LOAD_IPKEYMGRDRV	= 0xF0000,
 	PSP_BL__DRAM_LONG_TRAIN		= 0x100000,
 	PSP_BL__DRAM_SHORT_TRAIN	= 0x200000,
 	PSP_BL__LOAD_TOS_SPL_TABLE	= 0x10000000,
+	PSP_BL__LOAD_SPDMDRV		= 0x20000000,
 };
 
-enum psp_ring_type
-{
+enum psp_ring_type {
 	PSP_RING_TYPE__INVALID = 0,
 	/*
 	 * These values map to the way the PSP kernel identifies the
@@ -76,8 +116,7 @@ enum psp_ring_type
 	PSP_RING_TYPE__KM = 2  /* Kernel mode ring (formerly called GPCOM) */
 };
 
-struct psp_ring
-{
+struct psp_ring {
 	enum psp_ring_type		ring_type;
 	struct psp_gfx_rb_frame		*ring_mem;
 	uint64_t			ring_mem_mc_addr;
@@ -91,20 +130,26 @@ enum psp_reg_prog_id {
 	PSP_REG_IH_RB_CNTL        = 0,  /* register IH_RB_CNTL */
 	PSP_REG_IH_RB_CNTL_RING1  = 1,  /* register IH_RB_CNTL_RING1 */
 	PSP_REG_IH_RB_CNTL_RING2  = 2,  /* register IH_RB_CNTL_RING2 */
+	PSP_REG_MMHUB_L1_TLB_CNTL = 25,
 	PSP_REG_LAST
 };
 
-struct psp_funcs
-{
+#define PSP_WAITREG_CHANGED BIT(0) /* check if the value has changed */
+#define PSP_WAITREG_NOVERBOSE BIT(1) /* No error verbose */
+
+struct psp_funcs {
 	int (*init_microcode)(struct psp_context *psp);
+	int (*wait_for_bootloader)(struct psp_context *psp);
 	int (*bootloader_load_kdb)(struct psp_context *psp);
 	int (*bootloader_load_spl)(struct psp_context *psp);
 	int (*bootloader_load_sysdrv)(struct psp_context *psp);
 	int (*bootloader_load_soc_drv)(struct psp_context *psp);
 	int (*bootloader_load_intf_drv)(struct psp_context *psp);
 	int (*bootloader_load_dbg_drv)(struct psp_context *psp);
+	int (*bootloader_load_ras_drv)(struct psp_context *psp);
+	int (*bootloader_load_ipkeymgr_drv)(struct psp_context *psp);
+	int (*bootloader_load_spdm_drv)(struct psp_context *psp);
 	int (*bootloader_load_sos)(struct psp_context *psp);
-	int (*ring_init)(struct psp_context *psp, enum psp_ring_type ring_type);
 	int (*ring_create)(struct psp_context *psp,
 			   enum psp_ring_type ring_type);
 	int (*ring_stop)(struct psp_context *psp,
@@ -118,6 +163,21 @@ struct psp_funcs
 	void (*ring_set_wptr)(struct psp_context *psp, uint32_t value);
 	int (*load_usbc_pd_fw)(struct psp_context *psp, uint64_t fw_pri_mc_addr);
 	int (*read_usbc_pd_fw)(struct psp_context *psp, uint32_t *fw_ver);
+	int (*update_spirom)(struct psp_context *psp, uint64_t fw_pri_mc_addr);
+	int (*dump_spirom)(struct psp_context *psp, uint64_t fw_pri_mc_addr);
+	int (*vbflash_stat)(struct psp_context *psp);
+	int (*fatal_error_recovery_quirk)(struct psp_context *psp);
+	bool (*get_ras_capability)(struct psp_context *psp);
+	bool (*is_aux_sos_load_required)(struct psp_context *psp);
+	bool (*is_reload_needed)(struct psp_context *psp);
+	int (*reg_program_no_ring)(struct psp_context *psp, uint32_t val,
+				   enum psp_reg_prog_id id);
+};
+
+struct ta_funcs {
+	int (*fn_ta_initialize)(struct psp_context *psp);
+	int (*fn_ta_invoke)(struct psp_context *psp, uint32_t ta_cmd_id);
+	int (*fn_ta_terminate)(struct psp_context *psp);
 };
 
 #define AMDGPU_XGMI_MAX_CONNECTED_NODES		64
@@ -127,6 +187,7 @@ struct psp_xgmi_node_info {
 	uint8_t					is_sharing_enabled;
 	enum ta_xgmi_assigned_sdma_engine	sdma_engine;
 	uint8_t					num_links;
+	struct xgmi_connected_port_num		port_num[TA_XGMI__MAX_PORT_NUM];
 };
 
 struct psp_xgmi_topology_info {
@@ -151,9 +212,11 @@ struct ta_mem_context {
 struct ta_context {
 	bool			initialized;
 	uint32_t		session_id;
+	uint32_t		resp_status;
 	struct ta_mem_context	mem_context;
 	struct psp_bin_desc		bin_desc;
 	enum psp_gfx_cmd_id		ta_load_type;
+	enum ta_type_id		ta_type;
 };
 
 struct ta_cp_context {
@@ -165,18 +228,20 @@ struct psp_xgmi_context {
 	struct ta_context		context;
 	struct psp_xgmi_topology_info	top_info;
 	bool				supports_extended_data;
+	uint8_t				xgmi_ta_caps;
 };
 
 struct psp_ras_context {
 	struct ta_context		context;
 	struct amdgpu_ras		*ras;
+	struct mutex			mutex;
 };
 
 #define MEM_TRAIN_SYSTEM_SIGNATURE		0x54534942
 #define GDDR6_MEM_TRAINING_DATA_SIZE_IN_BYTES	0x1000
 #define GDDR6_MEM_TRAINING_OFFSET		0x8000
 /*Define the VRAM size that will be encroached by BIST training.*/
-#define GDDR6_MEM_TRAINING_ENCROACHED_SIZE	0x2000000
+#define BIST_MEM_TRAINING_ENCROACHED_SIZE	0x2000000
 
 enum psp_memory_training_init_flag {
 	PSP_MEM_TRAIN_NOT_SUPPORT	= 0x0,
@@ -231,6 +296,7 @@ enum psp_runtime_entry_type {
 	PSP_RUNTIME_ENTRY_TYPE_MGPU_WAFL	= 0x3,  /* WAFL runtime data */
 	PSP_RUNTIME_ENTRY_TYPE_MGPU_XGMI	= 0x4,  /* XGMI runtime data */
 	PSP_RUNTIME_ENTRY_TYPE_BOOT_CONFIG	= 0x5,  /* Boot Config runtime data */
+	PSP_RUNTIME_ENTRY_TYPE_PPTABLE_ERR_STATUS = 0x6, /* SCPM validation data */
 };
 
 /* PSP runtime DB header */
@@ -265,19 +331,39 @@ enum psp_runtime_boot_cfg_feature {
 	BOOT_CFG_FEATURE_TWO_STAGE_DRAM_TRAINING    = 0x2,
 };
 
+/* PSP run time DB SCPM authentication defines */
+enum psp_runtime_scpm_authentication {
+	SCPM_DISABLE                     = 0x0,
+	SCPM_ENABLE                      = 0x1,
+	SCPM_ENABLE_WITH_SCPM_ERR        = 0x2,
+};
+
 /* PSP runtime DB boot config entry */
 struct psp_runtime_boot_cfg_entry {
 	uint32_t boot_cfg_bitmask;
 	uint32_t reserved;
 };
 
-struct psp_context
-{
-	struct amdgpu_device            *adev;
-	struct psp_ring                 km_ring;
+/* PSP runtime DB SCPM entry */
+struct psp_runtime_scpm_entry {
+	enum psp_runtime_scpm_authentication scpm_status;
+};
+
+#if defined(CONFIG_DEBUG_FS)
+struct spirom_bo {
+	struct amdgpu_bo *bo;
+	uint64_t mc_addr;
+	void *cpu_addr;
+};
+#endif
+
+struct psp_context {
+	struct amdgpu_device		*adev;
+	struct psp_ring			km_ring;
 	struct psp_gfx_cmd_resp		*cmd;
 
 	const struct psp_funcs		*funcs;
+	const struct ta_funcs		*ta_funcs;
 
 	/* firmware buffer */
 	struct amdgpu_bo		*fw_pri_bo;
@@ -295,17 +381,23 @@ struct psp_context
 	struct psp_bin_desc		soc_drv;
 	struct psp_bin_desc		intf_drv;
 	struct psp_bin_desc		dbg_drv;
+	struct psp_bin_desc		ras_drv;
+	struct psp_bin_desc		ipkeymgr_drv;
+	struct psp_bin_desc		spdm_drv;
 
 	/* tmr buffer */
 	struct amdgpu_bo		*tmr_bo;
 	uint64_t			tmr_mc_addr;
 
 	/* asd firmware */
-	const struct firmware	*asd_fw;
+	const struct firmware		*asd_fw;
 
 	/* toc firmware */
 	const struct firmware		*toc_fw;
 
+	/* cap firmware */
+	const struct firmware		*cap_fw;
+
 	/* fence buffer */
 	struct amdgpu_bo		*fence_buf_bo;
 	uint64_t			fence_buf_mc_addr;
@@ -320,6 +412,8 @@ struct psp_context
 	atomic_t			fence_value;
 	/* flag to mark whether gfx fw autoload is supported or not */
 	bool				autoload_supported;
+	/* flag to mark whether psp use runtime TMR or boottime TMR */
+	bool				boot_time_tmr;
 	/* flag to mark whether df cstate management centralized to PMFW */
 	bool				pmfw_centralized_cstate_management;
 
@@ -327,6 +421,10 @@ struct psp_context
 	const struct firmware		*ta_fw;
 	uint32_t			ta_fw_version;
 
+	uint32_t			cap_fw_version;
+	uint32_t			cap_feature_version;
+	uint32_t			cap_ucode_size;
+
 	struct ta_context		asd_context;
 	struct psp_xgmi_context		xgmi_context;
 	struct psp_ras_context		ras_context;
@@ -338,6 +436,17 @@ struct psp_context
 	struct psp_memory_training_context mem_train_ctx;
 
 	uint32_t			boot_cfg_bitmask;
+
+	/* firmware upgrades supported */
+	bool				sup_pd_fw_up;
+	bool				sup_ifwi_up;
+
+	char				*vbflash_tmp_buf;
+	size_t				vbflash_image_size;
+	bool				vbflash_done;
+#if defined(CONFIG_DEBUG_FS)
+	struct spirom_bo *spirom_dump_trip;
+#endif
 };
 
 struct amdgpu_psp_funcs {
@@ -346,7 +455,6 @@ struct amdgpu_psp_funcs {
 };
 
 
-#define psp_ring_init(psp, type) (psp)->funcs->ring_init((psp), (type))
 #define psp_ring_create(psp, type) (psp)->funcs->ring_create((psp), (type))
 #define psp_ring_stop(psp, type) (psp)->funcs->ring_stop((psp), (type))
 #define psp_ring_destroy(psp, type) ((psp)->funcs->ring_destroy((psp), (type)))
@@ -364,6 +472,15 @@ struct amdgpu_psp_funcs {
 		((psp)->funcs->bootloader_load_intf_drv ? (psp)->funcs->bootloader_load_intf_drv((psp)) : 0)
 #define psp_bootloader_load_dbg_drv(psp) \
 		((psp)->funcs->bootloader_load_dbg_drv ? (psp)->funcs->bootloader_load_dbg_drv((psp)) : 0)
+#define psp_bootloader_load_ras_drv(psp) \
+		((psp)->funcs->bootloader_load_ras_drv ? \
+		(psp)->funcs->bootloader_load_ras_drv((psp)) : 0)
+#define psp_bootloader_load_ipkeymgr_drv(psp) \
+		((psp)->funcs->bootloader_load_ipkeymgr_drv ? \
+		 (psp)->funcs->bootloader_load_ipkeymgr_drv((psp)) : 0)
+#define psp_bootloader_load_spdm_drv(psp) \
+		((psp)->funcs->bootloader_load_spdm_drv ? \
+		 (psp)->funcs->bootloader_load_spdm_drv((psp)) : 0)
 #define psp_bootloader_load_sos(psp) \
 		((psp)->funcs->bootloader_load_sos ? (psp)->funcs->bootloader_load_sos((psp)) : 0)
 #define psp_smu_reload_quirk(psp) \
@@ -384,6 +501,29 @@ struct amdgpu_psp_funcs {
 	((psp)->funcs->read_usbc_pd_fw ? \
 	(psp)->funcs->read_usbc_pd_fw((psp), fw_ver) : -EINVAL)
 
+#define psp_update_spirom(psp, fw_pri_mc_addr) \
+	((psp)->funcs->update_spirom ? \
+	(psp)->funcs->update_spirom((psp), fw_pri_mc_addr) : -EINVAL)
+
+#define psp_dump_spirom(psp, fw_pri_mc_addr) \
+	((psp)->funcs->dump_spirom ? \
+	(psp)->funcs->dump_spirom((psp), fw_pri_mc_addr) : -EINVAL)
+
+#define psp_vbflash_status(psp) \
+	((psp)->funcs->vbflash_stat ? \
+	(psp)->funcs->vbflash_stat((psp)) : -EINVAL)
+
+#define psp_fatal_error_recovery_quirk(psp) \
+	((psp)->funcs->fatal_error_recovery_quirk ? \
+	(psp)->funcs->fatal_error_recovery_quirk((psp)) : 0)
+
+#define psp_is_aux_sos_load_required(psp) \
+	((psp)->funcs->is_aux_sos_load_required ? (psp)->funcs->is_aux_sos_load_required((psp)) : 0)
+
+#define psp_reg_program_no_ring(psp, val, id) \
+	((psp)->funcs->reg_program_no_ring ? \
+	(psp)->funcs->reg_program_no_ring((psp), val, id) : -EINVAL)
+
 extern const struct amd_ip_funcs psp_ip_funcs;
 
 extern const struct amdgpu_ip_block_version psp_v3_1_ip_block;
@@ -392,13 +532,27 @@ extern const struct amdgpu_ip_block_version psp_v11_0_ip_block;
 extern const struct amdgpu_ip_block_version psp_v11_0_8_ip_block;
 extern const struct amdgpu_ip_block_version psp_v12_0_ip_block;
 extern const struct amdgpu_ip_block_version psp_v13_0_ip_block;
+extern const struct amdgpu_ip_block_version psp_v13_0_4_ip_block;
+extern const struct amdgpu_ip_block_version psp_v14_0_ip_block;
+
+int psp_wait_for(struct psp_context *psp, uint32_t reg_index,
+		 uint32_t field_val, uint32_t mask, uint32_t flags);
+extern int psp_wait_for_spirom_update(struct psp_context *psp, uint32_t reg_index,
+			uint32_t field_val, uint32_t mask, uint32_t msec_timeout);
 
-extern int psp_wait_for(struct psp_context *psp, uint32_t reg_index,
-			uint32_t field_val, uint32_t mask, bool check_changed);
+int psp_execute_ip_fw_load(struct psp_context *psp,
+			   struct amdgpu_firmware_info *ucode);
 
 int psp_gpu_reset(struct amdgpu_device *adev);
-int psp_update_vcn_sram(struct amdgpu_device *adev, int inst_idx,
-			uint64_t cmd_gpu_addr, int cmd_size);
+
+int psp_ta_init_shared_buf(struct psp_context *psp,
+				  struct ta_mem_context *mem_ctx);
+void psp_ta_free_shared_buf(struct ta_mem_context *mem_ctx);
+int psp_ta_unload(struct psp_context *psp, struct ta_context *context);
+int psp_ta_load(struct psp_context *psp, struct ta_context *context);
+int psp_ta_invoke(struct psp_context *psp,
+			uint32_t ta_cmd_id,
+			struct ta_context *context);
 
 int psp_xgmi_initialize(struct psp_context *psp, bool set_extended_data, bool load_ta);
 int psp_xgmi_terminate(struct psp_context *psp);
@@ -412,12 +566,16 @@ int psp_xgmi_get_topology_info(struct psp_context *psp,
 int psp_xgmi_set_topology_info(struct psp_context *psp,
 			       int number_devices,
 			       struct psp_xgmi_topology_info *topology);
-
+int psp_ras_initialize(struct psp_context *psp);
 int psp_ras_invoke(struct psp_context *psp, uint32_t ta_cmd_id);
 int psp_ras_enable_features(struct psp_context *psp,
 		union ta_ras_cmd_input *info, bool enable);
 int psp_ras_trigger_error(struct psp_context *psp,
-			  struct ta_ras_trigger_error_input *info);
+			  struct ta_ras_trigger_error_input *info, uint32_t instance_mask);
+int psp_ras_terminate(struct psp_context *psp);
+int psp_ras_query_address(struct psp_context *psp,
+			  struct ta_ras_query_address_input *addr_in,
+			  struct ta_ras_query_address_output *addr_out);
 
 int psp_hdcp_invoke(struct psp_context *psp, uint32_t ta_cmd_id);
 int psp_dtm_invoke(struct psp_context *psp, uint32_t ta_cmd_id);
@@ -440,12 +598,29 @@ int psp_init_sos_microcode(struct psp_context *psp,
 			   const char *chip_name);
 int psp_init_ta_microcode(struct psp_context *psp,
 			  const char *chip_name);
+int psp_init_cap_microcode(struct psp_context *psp,
+			  const char *chip_name);
 int psp_get_fw_attestation_records_addr(struct psp_context *psp,
 					uint64_t *output_ptr);
-
+int psp_update_fw_reservation(struct psp_context *psp);
 int psp_load_fw_list(struct psp_context *psp,
 		     struct amdgpu_firmware_info **ucode_list, int ucode_count);
 void psp_copy_fw(struct psp_context *psp, uint8_t *start_addr, uint32_t bin_size);
 
+int psp_spatial_partition(struct psp_context *psp, int mode);
+int psp_memory_partition(struct psp_context *psp, int mode);
+
 int is_psp_fw_valid(struct psp_bin_desc bin);
+
+int amdgpu_psp_wait_for_bootloader(struct amdgpu_device *adev);
+bool amdgpu_psp_get_ras_capability(struct psp_context *psp);
+
+int psp_config_sq_perfmon(struct psp_context *psp, uint32_t xcp_id,
+	bool core_override_enable, bool reg_override_enable, bool perfmon_override_enable);
+bool amdgpu_psp_tos_reload_needed(struct amdgpu_device *adev);
+int amdgpu_psp_reg_program_no_ring(struct psp_context *psp, uint32_t val,
+				   enum psp_reg_prog_id id);
+void amdgpu_psp_debugfs_init(struct amdgpu_device *adev);
+
+
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp_ta.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp_ta.c
new file mode 100644
index 000000000000..6e8aad91bcd3
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp_ta.c
@@ -0,0 +1,392 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "amdgpu.h"
+#include "amdgpu_psp_ta.h"
+
+#if defined(CONFIG_DEBUG_FS)
+
+static ssize_t ta_if_load_debugfs_write(struct file *fp, const char *buf,
+					    size_t len, loff_t *off);
+static ssize_t ta_if_unload_debugfs_write(struct file *fp, const char *buf,
+					    size_t len, loff_t *off);
+static ssize_t ta_if_invoke_debugfs_write(struct file *fp, const char *buf,
+					    size_t len, loff_t *off);
+
+static uint32_t get_bin_version(const uint8_t *bin)
+{
+	const struct common_firmware_header *hdr =
+			     (const struct common_firmware_header *)bin;
+
+	return hdr->ucode_version;
+}
+
+static int prep_ta_mem_context(struct ta_mem_context *mem_context,
+					     uint8_t *shared_buf,
+					     uint32_t shared_buf_len)
+{
+	if (mem_context->shared_mem_size < shared_buf_len)
+		return -EINVAL;
+	memset(mem_context->shared_buf, 0, mem_context->shared_mem_size);
+	memcpy((void *)mem_context->shared_buf, shared_buf, shared_buf_len);
+
+	return 0;
+}
+
+static bool is_ta_type_valid(enum ta_type_id ta_type)
+{
+	switch (ta_type) {
+	case TA_TYPE_RAS:
+		return true;
+	default:
+		return false;
+	}
+}
+
+static const struct ta_funcs ras_ta_funcs = {
+	.fn_ta_initialize = psp_ras_initialize,
+	.fn_ta_invoke    = psp_ras_invoke,
+	.fn_ta_terminate = psp_ras_terminate
+};
+
+static void set_ta_context_funcs(struct psp_context *psp,
+						      enum ta_type_id ta_type,
+						      struct ta_context **pcontext)
+{
+	switch (ta_type) {
+	case TA_TYPE_RAS:
+		*pcontext = &psp->ras_context.context;
+		psp->ta_funcs = &ras_ta_funcs;
+		break;
+	default:
+		break;
+	}
+}
+
+static const struct file_operations ta_load_debugfs_fops = {
+	.write  = ta_if_load_debugfs_write,
+	.llseek = default_llseek,
+	.owner  = THIS_MODULE
+};
+
+static const struct file_operations ta_unload_debugfs_fops = {
+	.write  = ta_if_unload_debugfs_write,
+	.llseek = default_llseek,
+	.owner  = THIS_MODULE
+};
+
+static const struct file_operations ta_invoke_debugfs_fops = {
+	.write  = ta_if_invoke_debugfs_write,
+	.llseek = default_llseek,
+	.owner  = THIS_MODULE
+};
+
+/*
+ * DOC: AMDGPU TA debugfs interfaces
+ *
+ * Three debugfs interfaces can be opened by a program to
+ * load/invoke/unload TA,
+ *
+ * - /sys/kernel/debug/dri/<N>/ta_if/ta_load
+ * - /sys/kernel/debug/dri/<N>/ta_if/ta_invoke
+ * - /sys/kernel/debug/dri/<N>/ta_if/ta_unload
+ *
+ * How to use the interfaces in a program?
+ *
+ * A program needs to provide transmit buffer to the interfaces
+ * and will receive buffer from the interfaces below,
+ *
+ * - For TA load debugfs interface:
+ *   Transmit buffer:
+ *    - TA type (4bytes)
+ *    - TA bin length (4bytes)
+ *    - TA bin
+ *   Receive buffer:
+ *    - TA ID (4bytes)
+ *
+ * - For TA invoke debugfs interface:
+ *   Transmit buffer:
+ *    - TA type (4bytes)
+ *    - TA ID (4bytes)
+ *    - TA CMD ID (4bytes)
+ *    - TA shard buf length
+ *      (4bytes, value not beyond TA shared memory size)
+ *    - TA shared buf
+ *   Receive buffer:
+ *    - TA shared buf
+ *
+ * - For TA unload debugfs interface:
+ *   Transmit buffer:
+ *    - TA type (4bytes)
+ *    - TA ID (4bytes)
+ */
+
+static ssize_t ta_if_load_debugfs_write(struct file *fp, const char *buf, size_t len, loff_t *off)
+{
+	uint32_t ta_type    = 0;
+	uint32_t ta_bin_len = 0;
+	uint8_t  *ta_bin    = NULL;
+	uint32_t copy_pos   = 0;
+	int      ret        = 0;
+
+	struct amdgpu_device *adev    = (struct amdgpu_device *)file_inode(fp)->i_private;
+	struct psp_context   *psp     = &adev->psp;
+	struct ta_context    *context = NULL;
+
+	if (!buf)
+		return -EINVAL;
+
+	ret = copy_from_user((void *)&ta_type, &buf[copy_pos], sizeof(uint32_t));
+	if (ret || (!is_ta_type_valid(ta_type)))
+		return -EFAULT;
+
+	copy_pos += sizeof(uint32_t);
+
+	ret = copy_from_user((void *)&ta_bin_len, &buf[copy_pos], sizeof(uint32_t));
+	if (ret)
+		return -EFAULT;
+
+	if (ta_bin_len > PSP_1_MEG)
+		return -EINVAL;
+
+	copy_pos += sizeof(uint32_t);
+
+	ta_bin = memdup_user(&buf[copy_pos], ta_bin_len);
+	if (IS_ERR(ta_bin))
+		return PTR_ERR(ta_bin);
+
+	/* Set TA context and functions */
+	set_ta_context_funcs(psp, ta_type, &context);
+
+	if (!psp->ta_funcs || !psp->ta_funcs->fn_ta_terminate) {
+		dev_err(adev->dev, "Unsupported function to terminate TA\n");
+		ret = -EOPNOTSUPP;
+		goto err_free_bin;
+	}
+
+	/*
+	 * Allocate TA shared buf in case shared buf was freed
+	 * due to loading TA failed before.
+	 */
+	if (!context->mem_context.shared_buf) {
+		ret = psp_ta_init_shared_buf(psp, &context->mem_context);
+		if (ret) {
+			ret = -ENOMEM;
+			goto err_free_bin;
+		}
+	}
+
+	ret = psp_fn_ta_terminate(psp);
+	if (ret || context->resp_status) {
+		dev_err(adev->dev,
+			"Failed to unload embedded TA (%d) and status (0x%X)\n",
+			ret, context->resp_status);
+		if (!ret)
+			ret = -EINVAL;
+		goto err_free_ta_shared_buf;
+	}
+
+	/* Prepare TA context for TA initialization */
+	context->ta_type                     = ta_type;
+	context->bin_desc.fw_version         = get_bin_version(ta_bin);
+	context->bin_desc.size_bytes         = ta_bin_len;
+	context->bin_desc.start_addr         = ta_bin;
+
+	if (!psp->ta_funcs->fn_ta_initialize) {
+		dev_err(adev->dev, "Unsupported function to initialize TA\n");
+		ret = -EOPNOTSUPP;
+		goto err_free_ta_shared_buf;
+	}
+
+	ret = psp_fn_ta_initialize(psp);
+	if (ret || context->resp_status) {
+		dev_err(adev->dev, "Failed to load TA via debugfs (%d) and status (0x%X)\n",
+			ret, context->resp_status);
+		if (!ret)
+			ret = -EINVAL;
+		goto err_free_ta_shared_buf;
+	}
+
+	if (copy_to_user((char *)buf, (void *)&context->session_id, sizeof(uint32_t)))
+		ret = -EFAULT;
+
+err_free_ta_shared_buf:
+	/* Only free TA shared buf when returns error code */
+	if (ret && context->mem_context.shared_buf)
+		psp_ta_free_shared_buf(&context->mem_context);
+err_free_bin:
+	kfree(ta_bin);
+
+	return ret;
+}
+
+static ssize_t ta_if_unload_debugfs_write(struct file *fp, const char *buf, size_t len, loff_t *off)
+{
+	uint32_t ta_type    = 0;
+	uint32_t ta_id      = 0;
+	uint32_t copy_pos   = 0;
+	int      ret        = 0;
+
+	struct amdgpu_device *adev    = (struct amdgpu_device *)file_inode(fp)->i_private;
+	struct psp_context   *psp     = &adev->psp;
+	struct ta_context    *context = NULL;
+
+	if (!buf)
+		return -EINVAL;
+
+	ret = copy_from_user((void *)&ta_type, &buf[copy_pos], sizeof(uint32_t));
+	if (ret || (!is_ta_type_valid(ta_type)))
+		return -EFAULT;
+
+	copy_pos += sizeof(uint32_t);
+
+	ret = copy_from_user((void *)&ta_id, &buf[copy_pos], sizeof(uint32_t));
+	if (ret)
+		return -EFAULT;
+
+	set_ta_context_funcs(psp, ta_type, &context);
+	context->session_id = ta_id;
+
+	if (!psp->ta_funcs || !psp->ta_funcs->fn_ta_terminate) {
+		dev_err(adev->dev, "Unsupported function to terminate TA\n");
+		return -EOPNOTSUPP;
+	}
+
+	ret = psp_fn_ta_terminate(psp);
+	if (ret || context->resp_status) {
+		dev_err(adev->dev, "Failed to unload TA via debugfs (%d) and status (0x%X)\n",
+			ret, context->resp_status);
+		if (!ret)
+			ret = -EINVAL;
+	}
+
+	if (context->mem_context.shared_buf)
+		psp_ta_free_shared_buf(&context->mem_context);
+
+	return ret;
+}
+
+static ssize_t ta_if_invoke_debugfs_write(struct file *fp, const char *buf, size_t len, loff_t *off)
+{
+	uint32_t ta_type        = 0;
+	uint32_t ta_id          = 0;
+	uint32_t cmd_id         = 0;
+	uint32_t shared_buf_len = 0;
+	uint8_t *shared_buf     = NULL;
+	uint32_t copy_pos       = 0;
+	int      ret            = 0;
+
+	struct amdgpu_device *adev    = (struct amdgpu_device *)file_inode(fp)->i_private;
+	struct psp_context   *psp     = &adev->psp;
+	struct ta_context    *context = NULL;
+
+	if (!buf)
+		return -EINVAL;
+
+	ret = copy_from_user((void *)&ta_type, &buf[copy_pos], sizeof(uint32_t));
+	if (ret)
+		return -EFAULT;
+	copy_pos += sizeof(uint32_t);
+
+	ret = copy_from_user((void *)&ta_id, &buf[copy_pos], sizeof(uint32_t));
+	if (ret)
+		return -EFAULT;
+	copy_pos += sizeof(uint32_t);
+
+	ret = copy_from_user((void *)&cmd_id, &buf[copy_pos], sizeof(uint32_t));
+	if (ret)
+		return -EFAULT;
+	copy_pos += sizeof(uint32_t);
+
+	ret = copy_from_user((void *)&shared_buf_len, &buf[copy_pos], sizeof(uint32_t));
+	if (ret)
+		return -EFAULT;
+	copy_pos += sizeof(uint32_t);
+
+	shared_buf = memdup_user(&buf[copy_pos], shared_buf_len);
+	if (IS_ERR(shared_buf))
+		return PTR_ERR(shared_buf);
+
+	set_ta_context_funcs(psp, ta_type, &context);
+
+	if (!context || !context->initialized) {
+		dev_err(adev->dev, "TA is not initialized\n");
+		ret = -EINVAL;
+		goto err_free_shared_buf;
+	}
+
+	if (!psp->ta_funcs || !psp->ta_funcs->fn_ta_invoke) {
+		dev_err(adev->dev, "Unsupported function to invoke TA\n");
+		ret = -EOPNOTSUPP;
+		goto err_free_shared_buf;
+	}
+
+	context->session_id = ta_id;
+
+	mutex_lock(&psp->ras_context.mutex);
+	ret = prep_ta_mem_context(&context->mem_context, shared_buf, shared_buf_len);
+	if (ret)
+		goto err_free_shared_buf;
+
+	ret = psp_fn_ta_invoke(psp, cmd_id);
+	if (ret || context->resp_status) {
+		dev_err(adev->dev, "Failed to invoke TA via debugfs (%d) and status (0x%X)\n",
+			ret, context->resp_status);
+		if (!ret) {
+			ret = -EINVAL;
+			goto err_free_shared_buf;
+		}
+	}
+
+	if (copy_to_user((char *)&buf[copy_pos], context->mem_context.shared_buf, shared_buf_len))
+		ret = -EFAULT;
+
+err_free_shared_buf:
+	mutex_unlock(&psp->ras_context.mutex);
+	kfree(shared_buf);
+
+	return ret;
+}
+
+void amdgpu_ta_if_debugfs_init(struct amdgpu_device *adev)
+{
+	struct drm_minor *minor = adev_to_drm(adev)->primary;
+
+	struct dentry *dir = debugfs_create_dir("ta_if", minor->debugfs_root);
+
+	debugfs_create_file("ta_load", 0200, dir, adev,
+				     &ta_load_debugfs_fops);
+
+	debugfs_create_file("ta_unload", 0200, dir,
+				     adev, &ta_unload_debugfs_fops);
+
+	debugfs_create_file("ta_invoke", 0200, dir,
+				     adev, &ta_invoke_debugfs_fops);
+}
+
+#else
+void amdgpu_ta_if_debugfs_init(struct amdgpu_device *adev)
+{
+
+}
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp_ta.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp_ta.h
new file mode 100644
index 000000000000..14cd1c81c3e6
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp_ta.h
@@ -0,0 +1,34 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __AMDGPU_PSP_TA_H__
+#define __AMDGPU_PSP_TA_H__
+
+/* Calling set_ta_context_funcs is required before using the following macros */
+#define psp_fn_ta_initialize(psp) ((psp)->ta_funcs->fn_ta_initialize((psp)))
+#define psp_fn_ta_invoke(psp, ta_cmd_id) ((psp)->ta_funcs->fn_ta_invoke((psp), (ta_cmd_id)))
+#define psp_fn_ta_terminate(psp) ((psp)->ta_funcs->fn_ta_terminate((psp)))
+
+void amdgpu_ta_if_debugfs_init(struct amdgpu_device *adev);
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_rap.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_rap.c
index 12010c988c8b..bacf888735db 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_rap.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_rap.c
@@ -101,7 +101,6 @@ static ssize_t amdgpu_rap_debugfs_write(struct file *f, const char __user *buf,
 	}
 
 	amdgpu_gfx_off_ctrl(adev, true);
-	pm_runtime_mark_last_busy(dev->dev);
 	pm_runtime_put_autosuspend(dev->dev);
 
 	return size;
@@ -116,7 +115,6 @@ static const struct file_operations amdgpu_rap_debugfs_ops = {
 
 void amdgpu_rap_debugfs_init(struct amdgpu_device *adev)
 {
-#if defined(CONFIG_DEBUG_FS)
 	struct drm_minor *minor = adev_to_drm(adev)->primary;
 
 	if (!adev->psp.rap_context.context.initialized)
@@ -124,5 +122,4 @@ void amdgpu_rap_debugfs_init(struct amdgpu_device *adev)
 
 	debugfs_create_file("rap_test", S_IWUSR, minor->debugfs_root,
 				adev, &amdgpu_rap_debugfs_ops);
-#endif
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
index 08133de21fdd..2a6cf7963dde 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
@@ -28,13 +28,21 @@
 #include <linux/reboot.h>
 #include <linux/syscalls.h>
 #include <linux/pm_runtime.h>
+#include <linux/list_sort.h>
 
 #include "amdgpu.h"
 #include "amdgpu_ras.h"
 #include "amdgpu_atomfirmware.h"
 #include "amdgpu_xgmi.h"
 #include "ivsrcid/nbio/irqsrcs_nbif_7_4.h"
+#include "nbio_v4_3.h"
+#include "nbif_v6_3_1.h"
+#include "nbio_v7_9.h"
 #include "atom.h"
+#include "amdgpu_reset.h"
+#include "amdgpu_psp.h"
+#include "amdgpu_ras_mgr.h"
+
 #ifdef CONFIG_X86_MCE_AMD
 #include <asm/mce.h>
 
@@ -66,6 +74,11 @@ const char *ras_block_string[] = {
 	"mp1",
 	"fuse",
 	"mca",
+	"vcn",
+	"jpeg",
+	"ih",
+	"mpio",
+	"mmsch",
 };
 
 const char *ras_mca_block_string[] = {
@@ -75,12 +88,20 @@ const char *ras_mca_block_string[] = {
 	"mca_iohc",
 };
 
+struct amdgpu_ras_block_list {
+	/* ras block link */
+	struct list_head node;
+
+	struct amdgpu_ras_block_object *ras_obj;
+};
+
 const char *get_ras_block_str(struct ras_common_if *ras_block)
 {
 	if (!ras_block)
 		return "NULL";
 
-	if (ras_block->block >= AMDGPU_RAS_BLOCK_COUNT)
+	if (ras_block->block >= AMDGPU_RAS_BLOCK_COUNT ||
+	    ras_block->block >= ARRAY_SIZE(ras_block_string))
 		return "OUT OF RANGE";
 
 	if (ras_block->block == AMDGPU_RAS_BLOCK__MCA)
@@ -89,6 +110,9 @@ const char *get_ras_block_str(struct ras_common_if *ras_block)
 	return ras_block_string[ras_block->block];
 }
 
+#define ras_block_str(_BLOCK_) \
+	(((_BLOCK_) < ARRAY_SIZE(ras_block_string)) ? ras_block_string[_BLOCK_] : "Out Of Range")
+
 #define ras_err_str(i) (ras_error_string[ffs(i)])
 
 #define RAS_DEFAULT_FLAGS (AMDGPU_RAS_FLAG_INIT_BY_VBIOS)
@@ -99,6 +123,15 @@ const char *get_ras_block_str(struct ras_common_if *ras_block)
 /* typical ECC bad page rate is 1 bad page per 100MB VRAM */
 #define RAS_BAD_PAGE_COVER              (100 * 1024 * 1024ULL)
 
+#define MAX_UMC_POISON_POLLING_TIME_ASYNC  10
+
+#define AMDGPU_RAS_RETIRE_PAGE_INTERVAL 100  //ms
+
+#define MAX_FLUSH_RETIRE_DWORK_TIMES  100
+
+#define BYPASS_ALLOCATED_ADDRESS        0x0
+#define BYPASS_INITIALIZATION_ADDRESS   0x1
+
 enum amdgpu_ras_retire_page_reservation {
 	AMDGPU_RAS_RETIRE_PAGE_RESERVED,
 	AMDGPU_RAS_RETIRE_PAGE_PENDING,
@@ -107,12 +140,18 @@ enum amdgpu_ras_retire_page_reservation {
 
 atomic_t amdgpu_ras_in_intr = ATOMIC_INIT(0);
 
-static bool amdgpu_ras_check_bad_page_unlock(struct amdgpu_ras *con,
+static int amdgpu_ras_check_bad_page_unlock(struct amdgpu_ras *con,
 				uint64_t addr);
-static bool amdgpu_ras_check_bad_page(struct amdgpu_device *adev,
+static int amdgpu_ras_check_bad_page(struct amdgpu_device *adev,
 				uint64_t addr);
+
+static void amdgpu_ras_critical_region_init(struct amdgpu_device *adev);
+static void amdgpu_ras_critical_region_fini(struct amdgpu_device *adev);
+
 #ifdef CONFIG_X86_MCE_AMD
 static void amdgpu_register_bad_pages_mca_notifier(struct amdgpu_device *adev);
+static void
+amdgpu_unregister_bad_pages_mca_notifier(struct amdgpu_device *adev);
 struct mce_notifier_adev_list {
 	struct amdgpu_device *devs[MAX_GPU_INSTANCE];
 	int num_gpu;
@@ -136,40 +175,39 @@ static bool amdgpu_ras_get_error_query_ready(struct amdgpu_device *adev)
 
 static int amdgpu_reserve_page_direct(struct amdgpu_device *adev, uint64_t address)
 {
-	struct ras_err_data err_data = {0, 0, 0, NULL};
+	struct ras_err_data err_data;
 	struct eeprom_table_record err_rec;
+	int ret;
 
-	if ((address >= adev->gmc.mc_vram_size) ||
-	    (address >= RAS_UMC_INJECT_ADDR_LIMIT)) {
+	ret = amdgpu_ras_check_bad_page(adev, address);
+	if (ret == -EINVAL) {
 		dev_warn(adev->dev,
-		         "RAS WARN: input address 0x%llx is invalid.\n",
-		         address);
+			"RAS WARN: input address 0x%llx is invalid.\n",
+			address);
 		return -EINVAL;
-	}
-
-	if (amdgpu_ras_check_bad_page(adev, address)) {
+	} else if (ret == 1) {
 		dev_warn(adev->dev,
-			 "RAS WARN: 0x%llx has already been marked as bad page!\n",
-			 address);
+			"RAS WARN: 0x%llx has already been marked as bad page!\n",
+			address);
 		return 0;
 	}
 
-	memset(&err_rec, 0x0, sizeof(struct eeprom_table_record));
-
-	err_rec.address = address;
-	err_rec.retired_page = address >> AMDGPU_GPU_PAGE_SHIFT;
-	err_rec.ts = (uint64_t)ktime_get_real_seconds();
-	err_rec.err_type = AMDGPU_RAS_EEPROM_ERR_NON_RECOVERABLE;
+	ret = amdgpu_ras_error_data_init(&err_data);
+	if (ret)
+		return ret;
 
+	memset(&err_rec, 0x0, sizeof(struct eeprom_table_record));
 	err_data.err_addr = &err_rec;
-	err_data.err_addr_cnt = 1;
+	amdgpu_umc_fill_error_record(&err_data, address, address, 0, 0);
 
 	if (amdgpu_bad_page_threshold != 0) {
 		amdgpu_ras_add_bad_pages(adev, err_data.err_addr,
-					 err_data.err_addr_cnt);
-		amdgpu_ras_save_bad_pages(adev);
+					 err_data.err_addr_cnt, false);
+		amdgpu_ras_save_bad_pages(adev, NULL);
 	}
 
+	amdgpu_ras_error_data_fini(&err_data);
+
 	dev_warn(adev->dev, "WARNING: THIS IS ONLY FOR TEST PURPOSES AND WILL CORRUPT RAS EEPROM\n");
 	dev_warn(adev->dev, "Clear EEPROM:\n");
 	dev_warn(adev->dev, "    echo 1 > /sys/kernel/debug/dri/0/ras/ras_eeprom_reset\n");
@@ -177,6 +215,56 @@ static int amdgpu_reserve_page_direct(struct amdgpu_device *adev, uint64_t addre
 	return 0;
 }
 
+static int amdgpu_check_address_validity(struct amdgpu_device *adev,
+			uint64_t address, uint64_t flags)
+{
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+	struct amdgpu_vram_block_info blk_info;
+	uint64_t page_pfns[32] = {0};
+	int i, ret, count;
+	bool hit = false;
+
+	if (amdgpu_ip_version(adev, UMC_HWIP, 0) < IP_VERSION(12, 0, 0))
+		return 0;
+
+	if (amdgpu_sriov_vf(adev)) {
+		if (amdgpu_virt_check_vf_critical_region(adev, address, &hit))
+			return -EPERM;
+		return hit ? -EACCES : 0;
+	}
+
+	if ((address >= adev->gmc.mc_vram_size) ||
+	    (address >= RAS_UMC_INJECT_ADDR_LIMIT))
+		return -EFAULT;
+
+	count = amdgpu_umc_lookup_bad_pages_in_a_row(adev,
+				address, page_pfns, ARRAY_SIZE(page_pfns));
+	if (count <= 0)
+		return -EPERM;
+
+	for (i = 0; i < count; i++) {
+		memset(&blk_info, 0, sizeof(blk_info));
+		ret = amdgpu_vram_mgr_query_address_block_info(&adev->mman.vram_mgr,
+					page_pfns[i] << AMDGPU_GPU_PAGE_SHIFT, &blk_info);
+		if (!ret) {
+			/* The input address that needs to be checked is allocated by
+			 * current calling process, so it is necessary to exclude
+			 * the calling process.
+			 */
+			if ((flags == BYPASS_ALLOCATED_ADDRESS) &&
+			    ((blk_info.task.pid != task_pid_nr(current)) ||
+				strncmp(blk_info.task.comm, current->comm, TASK_COMM_LEN)))
+				return -EACCES;
+			else if ((flags == BYPASS_INITIALIZATION_ADDRESS) &&
+				(blk_info.task.pid == con->init_task_pid) &&
+				!strncmp(blk_info.task.comm, con->init_task_comm, TASK_COMM_LEN))
+				return -EACCES;
+		}
+	}
+
+	return 0;
+}
+
 static ssize_t amdgpu_ras_debugfs_read(struct file *f, char __user *buf,
 					size_t size, loff_t *pos)
 {
@@ -190,6 +278,13 @@ static ssize_t amdgpu_ras_debugfs_read(struct file *f, char __user *buf,
 	if (amdgpu_ras_query_error_status(obj->adev, &info))
 		return -EINVAL;
 
+	/* Hardware counter will be reset automatically after the query on Vega20 and Arcturus */
+	if (amdgpu_ip_version(obj->adev, MP0_HWIP, 0) != IP_VERSION(11, 0, 2) &&
+	    amdgpu_ip_version(obj->adev, MP0_HWIP, 0) != IP_VERSION(11, 0, 4)) {
+		if (amdgpu_ras_reset_error_status(obj->adev, info.head.block))
+			dev_warn(obj->adev->dev, "Failed to reset error counter and error status");
+	}
+
 	s = snprintf(val, sizeof(val), "%s: %lu\n%s: %lu\n",
 			"ue", info.ue_count,
 			"ce", info.ce_count);
@@ -239,6 +334,8 @@ static int amdgpu_ras_debugfs_ctrl_parse_data(struct file *f,
 	int block_id;
 	uint32_t sub_block;
 	u64 address, value;
+	/* default value is 0 if the mask is not set by user */
+	u32 instance_mask = 0;
 
 	if (*pos)
 		return -EINVAL;
@@ -258,6 +355,8 @@ static int amdgpu_ras_debugfs_ctrl_parse_data(struct file *f,
 		op = 2;
 	else if (strstr(str, "retire_page") != NULL)
 		op = 3;
+	else if (strstr(str, "check_address") != NULL)
+		op = 4;
 	else if (str[0] && str[1] && str[2] && str[3])
 		/* ascii string, but commands are not matched. */
 		return -EINVAL;
@@ -272,24 +371,39 @@ static int amdgpu_ras_debugfs_ctrl_parse_data(struct file *f,
 			data->inject.address = address;
 
 			return 0;
+		} else if (op == 4) {
+			if (sscanf(str, "%*s 0x%llx 0x%llx", &address, &value) != 2 &&
+			    sscanf(str, "%*s %llu %llu", &address, &value) != 2)
+				return -EINVAL;
+
+			data->op = op;
+			data->inject.address = address;
+			data->inject.value = value;
+			return 0;
 		}
 
 		if (amdgpu_ras_find_block_id_by_name(block_name, &block_id))
 			return -EINVAL;
 
 		data->head.block = block_id;
-		/* only ue and ce errors are supported */
+		/* only ue, ce and poison errors are supported */
 		if (!memcmp("ue", err, 2))
 			data->head.type = AMDGPU_RAS_ERROR__MULTI_UNCORRECTABLE;
 		else if (!memcmp("ce", err, 2))
 			data->head.type = AMDGPU_RAS_ERROR__SINGLE_CORRECTABLE;
+		else if (!memcmp("poison", err, 6))
+			data->head.type = AMDGPU_RAS_ERROR__POISON;
 		else
 			return -EINVAL;
 
 		data->op = op;
 
 		if (op == 2) {
-			if (sscanf(str, "%*s %*s %*s 0x%x 0x%llx 0x%llx",
+			if (sscanf(str, "%*s %*s %*s 0x%x 0x%llx 0x%llx 0x%x",
+				   &sub_block, &address, &value, &instance_mask) != 4 &&
+			    sscanf(str, "%*s %*s %*s %u %llu %llu %u",
+				   &sub_block, &address, &value, &instance_mask) != 4 &&
+				sscanf(str, "%*s %*s %*s 0x%x 0x%llx 0x%llx",
 				   &sub_block, &address, &value) != 3 &&
 			    sscanf(str, "%*s %*s %*s %u %llu %llu",
 				   &sub_block, &address, &value) != 3)
@@ -297,6 +411,7 @@ static int amdgpu_ras_debugfs_ctrl_parse_data(struct file *f,
 			data->head.sub_block_index = sub_block;
 			data->inject.address = address;
 			data->inject.value = value;
+			data->inject.instance_mask = instance_mask;
 		}
 	} else {
 		if (size < sizeof(*data))
@@ -309,6 +424,46 @@ static int amdgpu_ras_debugfs_ctrl_parse_data(struct file *f,
 	return 0;
 }
 
+static void amdgpu_ras_instance_mask_check(struct amdgpu_device *adev,
+				struct ras_debug_if *data)
+{
+	int num_xcc = adev->gfx.xcc_mask ? NUM_XCC(adev->gfx.xcc_mask) : 1;
+	uint32_t mask, inst_mask = data->inject.instance_mask;
+
+	/* no need to set instance mask if there is only one instance */
+	if (num_xcc <= 1 && inst_mask) {
+		data->inject.instance_mask = 0;
+		dev_dbg(adev->dev,
+			"RAS inject mask(0x%x) isn't supported and force it to 0.\n",
+			inst_mask);
+
+		return;
+	}
+
+	switch (data->head.block) {
+	case AMDGPU_RAS_BLOCK__GFX:
+		mask = GENMASK(num_xcc - 1, 0);
+		break;
+	case AMDGPU_RAS_BLOCK__SDMA:
+		mask = GENMASK(adev->sdma.num_instances - 1, 0);
+		break;
+	case AMDGPU_RAS_BLOCK__VCN:
+	case AMDGPU_RAS_BLOCK__JPEG:
+		mask = GENMASK(adev->vcn.num_vcn_inst - 1, 0);
+		break;
+	default:
+		mask = inst_mask;
+		break;
+	}
+
+	/* remove invalid bits in instance mask */
+	data->inject.instance_mask &= mask;
+	if (inst_mask != data->inject.instance_mask)
+		dev_dbg(adev->dev,
+			"Adjust RAS inject mask 0x%x to 0x%x\n",
+			inst_mask, data->inject.instance_mask);
+}
+
 /**
  * DOC: AMDGPU RAS debugfs control interface
  *
@@ -324,7 +479,7 @@ static int amdgpu_ras_debugfs_ctrl_parse_data(struct file *f,
  * sub_block_index: some IPs have subcomponets. say, GFX, sDMA.
  * name: the name of IP.
  *
- * inject has two more members than head, they are address, value.
+ * inject has three more members than head, they are address, value and mask.
  * As their names indicate, inject operation will write the
  * value to the address.
  *
@@ -348,7 +503,7 @@ static int amdgpu_ras_debugfs_ctrl_parse_data(struct file *f,
  *
  *	echo "disable <block>" > /sys/kernel/debug/dri/<N>/ras/ras_ctrl
  *	echo "enable  <block> <error>" > /sys/kernel/debug/dri/<N>/ras/ras_ctrl
- *	echo "inject  <block> <error> <sub-block> <address> <value> > /sys/kernel/debug/dri/<N>/ras/ras_ctrl
+ *	echo "inject  <block> <error> <sub-block> <address> <value> <mask>" > /sys/kernel/debug/dri/<N>/ras/ras_ctrl
  *
  * Where N, is the card which you want to affect.
  *
@@ -359,19 +514,21 @@ static int amdgpu_ras_debugfs_ctrl_parse_data(struct file *f,
  * The block is one of: umc, sdma, gfx, etc.
  *	see ras_block_string[] for details
  *
- * The error type is one of: ue, ce, where,
+ * The error type is one of: ue, ce and poison where,
  *	ue is multi-uncorrectable
  *	ce is single-correctable
+ *	poison is poison
  *
  * The sub-block is a the sub-block index, pass 0 if there is no sub-block.
  * The address and value are hexadecimal numbers, leading 0x is optional.
+ * The mask means instance mask, is optional, default value is 0x1.
  *
  * For instance,
  *
  * .. code-block:: bash
  *
  *	echo inject umc ue 0x0 0x0 0x0 > /sys/kernel/debug/dri/0/ras/ras_ctrl
- *	echo inject umc ce 0 0 0 > /sys/kernel/debug/dri/0/ras/ras_ctrl
+ *	echo inject umc ce 0 0 0 3 > /sys/kernel/debug/dri/0/ras/ras_ctrl
  *	echo disable umc > /sys/kernel/debug/dri/0/ras/ras_ctrl
  *
  * How to check the result of the operation?
@@ -412,6 +569,9 @@ static ssize_t amdgpu_ras_debugfs_ctrl_write(struct file *f,
 			return size;
 		else
 			return ret;
+	} else if (data.op == 4) {
+		ret = amdgpu_check_address_validity(adev, data.inject.address, data.inject.value);
+		return ret ? ret : size;
 	}
 
 	if (!amdgpu_ras_is_supported(adev, data.head.block))
@@ -425,24 +585,21 @@ static ssize_t amdgpu_ras_debugfs_ctrl_write(struct file *f,
 		ret = amdgpu_ras_feature_enable(adev, &data.head, 1);
 		break;
 	case 2:
-		if ((data.inject.address >= adev->gmc.mc_vram_size) ||
-		    (data.inject.address >= RAS_UMC_INJECT_ADDR_LIMIT)) {
-			dev_warn(adev->dev, "RAS WARN: input address "
-					"0x%llx is invalid.",
+		/* umc ce/ue error injection for a bad page is not allowed */
+		if (data.head.block == AMDGPU_RAS_BLOCK__UMC)
+			ret = amdgpu_ras_check_bad_page(adev, data.inject.address);
+		if (ret == -EINVAL) {
+			dev_warn(adev->dev, "RAS WARN: input address 0x%llx is invalid.",
 					data.inject.address);
-			ret = -EINVAL;
 			break;
-		}
-
-		/* umc ce/ue error injection for a bad page is not allowed */
-		if ((data.head.block == AMDGPU_RAS_BLOCK__UMC) &&
-		    amdgpu_ras_check_bad_page(adev, data.inject.address)) {
-			dev_warn(adev->dev, "RAS WARN: inject: 0x%llx has "
-				 "already been marked as bad!\n",
-				 data.inject.address);
+		} else if (ret == 1) {
+			dev_warn(adev->dev, "RAS WARN: inject: 0x%llx has already been marked as bad!\n",
+					data.inject.address);
 			break;
 		}
 
+		amdgpu_ras_instance_mask_check(adev, &data);
+
 		/* data.inject.address is offset instead of absolute gpu address */
 		ret = amdgpu_ras_error_inject(adev, &data.inject);
 		break;
@@ -452,11 +609,13 @@ static ssize_t amdgpu_ras_debugfs_ctrl_write(struct file *f,
 	}
 
 	if (ret)
-		return -EINVAL;
+		return ret;
 
 	return size;
 }
 
+static int amdgpu_uniras_clear_badpages_info(struct amdgpu_device *adev);
+
 /**
  * DOC: AMDGPU RAS debugfs EEPROM table reset interface
  *
@@ -481,6 +640,11 @@ static ssize_t amdgpu_ras_debugfs_eeprom_write(struct file *f,
 		(struct amdgpu_device *)file_inode(f)->i_private;
 	int ret;
 
+	if (amdgpu_uniras_enabled(adev)) {
+		ret = amdgpu_uniras_clear_badpages_info(adev);
+		return ret ? ret : size;
+	}
+
 	ret = amdgpu_ras_eeprom_reset_table(
 		&(amdgpu_ras_get_context(adev)->eeprom_control));
 
@@ -543,13 +707,18 @@ static ssize_t amdgpu_ras_sysfs_read(struct device *dev,
 	if (amdgpu_ras_query_error_status(obj->adev, &info))
 		return -EINVAL;
 
-	if (obj->adev->asic_type == CHIP_ALDEBARAN) {
+	if (amdgpu_ip_version(obj->adev, MP0_HWIP, 0) != IP_VERSION(11, 0, 2) &&
+	    amdgpu_ip_version(obj->adev, MP0_HWIP, 0) != IP_VERSION(11, 0, 4)) {
 		if (amdgpu_ras_reset_error_status(obj->adev, info.head.block))
-			DRM_WARN("Failed to reset error counter and error status");
+			dev_warn(obj->adev->dev, "Failed to reset error counter and error status");
 	}
 
-	return sysfs_emit(buf, "%s: %lu\n%s: %lu\n", "ue", info.ue_count,
-			  "ce", info.ce_count);
+	if (info.head.block == AMDGPU_RAS_BLOCK__UMC)
+		return sysfs_emit(buf, "%s: %lu\n%s: %lu\n%s: %lu\n", "ue", info.ue_count,
+				"ce", info.ce_count, "de", info.de_count);
+	else
+		return sysfs_emit(buf, "%s: %lu\n%s: %lu\n", "ue", info.ue_count,
+				"ce", info.ce_count);
 }
 
 /* obj begin */
@@ -559,8 +728,11 @@ static ssize_t amdgpu_ras_sysfs_read(struct device *dev,
 
 static inline void put_obj(struct ras_manager *obj)
 {
-	if (obj && (--obj->use == 0))
+	if (obj && (--obj->use == 0)) {
 		list_del(&obj->node);
+		amdgpu_ras_error_data_fini(&obj->err_data);
+	}
+
 	if (obj && (obj->use < 0))
 		DRM_ERROR("RAS ERROR: Unbalance obj(%s) use\n", get_ras_block_str(&obj->head));
 }
@@ -590,6 +762,9 @@ static struct ras_manager *amdgpu_ras_create_obj(struct amdgpu_device *adev,
 	if (alive_obj(obj))
 		return NULL;
 
+	if (amdgpu_ras_error_data_init(&obj->err_data))
+		return NULL;
+
 	obj->head = *head;
 	obj->adev = adev;
 	list_add(&obj->node, &con->head);
@@ -700,42 +875,50 @@ int amdgpu_ras_feature_enable(struct amdgpu_device *adev,
 	if (!con)
 		return -EINVAL;
 
-	info = kzalloc(sizeof(union ta_ras_cmd_input), GFP_KERNEL);
-	if (!info)
-		return -ENOMEM;
-
-	if (!enable) {
-		info->disable_features = (struct ta_ras_disable_features_input) {
-			.block_id =  amdgpu_ras_block_to_ta(head->block),
-			.error_type = amdgpu_ras_error_to_ta(head->type),
-		};
-	} else {
-		info->enable_features = (struct ta_ras_enable_features_input) {
-			.block_id =  amdgpu_ras_block_to_ta(head->block),
-			.error_type = amdgpu_ras_error_to_ta(head->type),
-		};
-	}
+	/* For non-gfx ip, do not enable ras feature if it is not allowed */
+	/* For gfx ip, regardless of feature support status, */
+	/* Force issue enable or disable ras feature commands */
+	if (head->block != AMDGPU_RAS_BLOCK__GFX &&
+	    !amdgpu_ras_is_feature_allowed(adev, head))
+		return 0;
 
-	/* Do not enable if it is not allowed. */
-	WARN_ON(enable && !amdgpu_ras_is_feature_allowed(adev, head));
+	/* Only enable gfx ras feature from host side */
+	if (head->block == AMDGPU_RAS_BLOCK__GFX &&
+	    !amdgpu_sriov_vf(adev) &&
+	    !amdgpu_ras_intr_triggered()) {
+		info = kzalloc(sizeof(union ta_ras_cmd_input), GFP_KERNEL);
+		if (!info)
+			return -ENOMEM;
+
+		if (!enable) {
+			info->disable_features = (struct ta_ras_disable_features_input) {
+				.block_id =  amdgpu_ras_block_to_ta(head->block),
+				.error_type = amdgpu_ras_error_to_ta(head->type),
+			};
+		} else {
+			info->enable_features = (struct ta_ras_enable_features_input) {
+				.block_id =  amdgpu_ras_block_to_ta(head->block),
+				.error_type = amdgpu_ras_error_to_ta(head->type),
+			};
+		}
 
-	if (!amdgpu_ras_intr_triggered()) {
 		ret = psp_ras_enable_features(&adev->psp, info, enable);
 		if (ret) {
 			dev_err(adev->dev, "ras %s %s failed poison:%d ret:%d\n",
 				enable ? "enable":"disable",
 				get_ras_block_str(head),
 				amdgpu_ras_is_poison_mode_supported(adev), ret);
-			goto out;
+			kfree(info);
+			return ret;
 		}
+
+		kfree(info);
 	}
 
 	/* setup the obj */
 	__amdgpu_ras_feature_enable(adev, head, enable);
-	ret = 0;
-out:
-	kfree(info);
-	return ret;
+
+	return 0;
 }
 
 /* Only used in device probe stage and called only once. */
@@ -774,7 +957,7 @@ int amdgpu_ras_feature_enable_on_boot(struct amdgpu_device *adev,
 			if (ret)
 				return ret;
 
-			/* gfx block ras dsiable cmd must send to ras-ta */
+			/* gfx block ras disable cmd must send to ras-ta */
 			if (head->block == AMDGPU_RAS_BLOCK__GFX)
 				con->features |= BIT(head->block);
 
@@ -866,220 +1049,626 @@ static int amdgpu_ras_enable_all_features(struct amdgpu_device *adev,
 }
 /* feature ctl end */
 
+static int amdgpu_ras_block_match_default(struct amdgpu_ras_block_object *block_obj,
+		enum amdgpu_ras_block block)
+{
+	if (!block_obj)
+		return -EINVAL;
+
+	if (block_obj->ras_comm.block == block)
+		return 0;
+
+	return -EINVAL;
+}
 
-void amdgpu_ras_mca_query_error_status(struct amdgpu_device *adev,
-				       struct ras_common_if *ras_block,
-				       struct ras_err_data  *err_data)
+static struct amdgpu_ras_block_object *amdgpu_ras_get_ras_block(struct amdgpu_device *adev,
+					enum amdgpu_ras_block block, uint32_t sub_block_index)
 {
-	switch (ras_block->sub_block_index) {
-	case AMDGPU_RAS_MCA_BLOCK__MP0:
-		if (adev->mca.mp0.ras_funcs &&
-		    adev->mca.mp0.ras_funcs->query_ras_error_count)
-			adev->mca.mp0.ras_funcs->query_ras_error_count(adev, &err_data);
-		break;
-	case AMDGPU_RAS_MCA_BLOCK__MP1:
-		if (adev->mca.mp1.ras_funcs &&
-		    adev->mca.mp1.ras_funcs->query_ras_error_count)
-			adev->mca.mp1.ras_funcs->query_ras_error_count(adev, &err_data);
-		break;
-	case AMDGPU_RAS_MCA_BLOCK__MPIO:
-		if (adev->mca.mpio.ras_funcs &&
-		    adev->mca.mpio.ras_funcs->query_ras_error_count)
-			adev->mca.mpio.ras_funcs->query_ras_error_count(adev, &err_data);
-		break;
-	default:
-		break;
+	struct amdgpu_ras_block_list *node, *tmp;
+	struct amdgpu_ras_block_object *obj;
+
+	if (block >= AMDGPU_RAS_BLOCK__LAST)
+		return NULL;
+
+	list_for_each_entry_safe(node, tmp, &adev->ras_list, node) {
+		if (!node->ras_obj) {
+			dev_warn(adev->dev, "Warning: abnormal ras list node.\n");
+			continue;
+		}
+
+		obj = node->ras_obj;
+		if (obj->ras_block_match) {
+			if (obj->ras_block_match(obj, block, sub_block_index) == 0)
+				return obj;
+		} else {
+			if (amdgpu_ras_block_match_default(obj, block) == 0)
+				return obj;
+		}
 	}
+
+	return NULL;
 }
 
-/* query/inject/cure begin */
-int amdgpu_ras_query_error_status(struct amdgpu_device *adev,
-				  struct ras_query_if *info)
+static void amdgpu_ras_get_ecc_info(struct amdgpu_device *adev, struct ras_err_data *err_data)
 {
-	struct ras_manager *obj = amdgpu_ras_find_obj(adev, &info->head);
-	struct ras_err_data err_data = {0, 0, 0, NULL};
-	int i;
+	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
+	int ret = 0;
 
-	if (!obj)
-		return -EINVAL;
+	/*
+	 * choosing right query method according to
+	 * whether smu support query error information
+	 */
+	ret = amdgpu_dpm_get_ecc_info(adev, (void *)&(ras->umc_ecc));
+	if (ret == -EOPNOTSUPP) {
+		if (adev->umc.ras && adev->umc.ras->ras_block.hw_ops &&
+			adev->umc.ras->ras_block.hw_ops->query_ras_error_count)
+			adev->umc.ras->ras_block.hw_ops->query_ras_error_count(adev, err_data);
 
-	switch (info->head.block) {
-	case AMDGPU_RAS_BLOCK__UMC:
-		if (adev->umc.ras_funcs &&
-		    adev->umc.ras_funcs->query_ras_error_count)
-			adev->umc.ras_funcs->query_ras_error_count(adev, &err_data);
 		/* umc query_ras_error_address is also responsible for clearing
 		 * error status
 		 */
-		if (adev->umc.ras_funcs &&
-		    adev->umc.ras_funcs->query_ras_error_address)
-			adev->umc.ras_funcs->query_ras_error_address(adev, &err_data);
-		break;
-	case AMDGPU_RAS_BLOCK__SDMA:
-		if (adev->sdma.funcs->query_ras_error_count) {
-			for (i = 0; i < adev->sdma.num_instances; i++)
-				adev->sdma.funcs->query_ras_error_count(adev, i,
-									&err_data);
+		if (adev->umc.ras && adev->umc.ras->ras_block.hw_ops &&
+		    adev->umc.ras->ras_block.hw_ops->query_ras_error_address)
+			adev->umc.ras->ras_block.hw_ops->query_ras_error_address(adev, err_data);
+	} else if (!ret) {
+		if (adev->umc.ras &&
+			adev->umc.ras->ecc_info_query_ras_error_count)
+			adev->umc.ras->ecc_info_query_ras_error_count(adev, err_data);
+
+		if (adev->umc.ras &&
+			adev->umc.ras->ecc_info_query_ras_error_address)
+			adev->umc.ras->ecc_info_query_ras_error_address(adev, err_data);
+	}
+}
+
+static void amdgpu_ras_error_print_error_data(struct amdgpu_device *adev,
+					      struct ras_manager *ras_mgr,
+					      struct ras_err_data *err_data,
+					      struct ras_query_context *qctx,
+					      const char *blk_name,
+					      bool is_ue,
+					      bool is_de)
+{
+	struct amdgpu_smuio_mcm_config_info *mcm_info;
+	struct ras_err_node *err_node;
+	struct ras_err_info *err_info;
+	u64 event_id = qctx->evid.event_id;
+
+	if (is_ue) {
+		for_each_ras_error(err_node, err_data) {
+			err_info = &err_node->err_info;
+			mcm_info = &err_info->mcm_info;
+			if (err_info->ue_count) {
+				RAS_EVENT_LOG(adev, event_id, "socket: %d, die: %d, "
+					      "%lld new uncorrectable hardware errors detected in %s block\n",
+					      mcm_info->socket_id,
+					      mcm_info->die_id,
+					      err_info->ue_count,
+					      blk_name);
+			}
 		}
-		break;
-	case AMDGPU_RAS_BLOCK__GFX:
-		if (adev->gfx.ras_funcs &&
-		    adev->gfx.ras_funcs->query_ras_error_count)
-			adev->gfx.ras_funcs->query_ras_error_count(adev, &err_data);
 
-		if (adev->gfx.ras_funcs &&
-		    adev->gfx.ras_funcs->query_ras_error_status)
-			adev->gfx.ras_funcs->query_ras_error_status(adev);
-		break;
-	case AMDGPU_RAS_BLOCK__MMHUB:
-		if (adev->mmhub.ras_funcs &&
-		    adev->mmhub.ras_funcs->query_ras_error_count)
-			adev->mmhub.ras_funcs->query_ras_error_count(adev, &err_data);
-
-		if (adev->mmhub.ras_funcs &&
-		    adev->mmhub.ras_funcs->query_ras_error_status)
-			adev->mmhub.ras_funcs->query_ras_error_status(adev);
-		break;
-	case AMDGPU_RAS_BLOCK__PCIE_BIF:
-		if (adev->nbio.ras_funcs &&
-		    adev->nbio.ras_funcs->query_ras_error_count)
-			adev->nbio.ras_funcs->query_ras_error_count(adev, &err_data);
-		break;
-	case AMDGPU_RAS_BLOCK__XGMI_WAFL:
-		if (adev->gmc.xgmi.ras_funcs &&
-		    adev->gmc.xgmi.ras_funcs->query_ras_error_count)
-			adev->gmc.xgmi.ras_funcs->query_ras_error_count(adev, &err_data);
-		break;
-	case AMDGPU_RAS_BLOCK__HDP:
-		if (adev->hdp.ras_funcs &&
-		    adev->hdp.ras_funcs->query_ras_error_count)
-			adev->hdp.ras_funcs->query_ras_error_count(adev, &err_data);
-		break;
-	case AMDGPU_RAS_BLOCK__MCA:
-		amdgpu_ras_mca_query_error_status(adev, &info->head, &err_data);
-		break;
-	default:
-		break;
+		for_each_ras_error(err_node, &ras_mgr->err_data) {
+			err_info = &err_node->err_info;
+			mcm_info = &err_info->mcm_info;
+			RAS_EVENT_LOG(adev, event_id, "socket: %d, die: %d, "
+				      "%lld uncorrectable hardware errors detected in total in %s block\n",
+				      mcm_info->socket_id, mcm_info->die_id, err_info->ue_count, blk_name);
+		}
+
+	} else {
+		if (is_de) {
+			for_each_ras_error(err_node, err_data) {
+				err_info = &err_node->err_info;
+				mcm_info = &err_info->mcm_info;
+				if (err_info->de_count) {
+					RAS_EVENT_LOG(adev, event_id, "socket: %d, die: %d, "
+						      "%lld new deferred hardware errors detected in %s block\n",
+						      mcm_info->socket_id,
+						      mcm_info->die_id,
+						      err_info->de_count,
+						      blk_name);
+				}
+			}
+
+			for_each_ras_error(err_node, &ras_mgr->err_data) {
+				err_info = &err_node->err_info;
+				mcm_info = &err_info->mcm_info;
+				RAS_EVENT_LOG(adev, event_id, "socket: %d, die: %d, "
+					      "%lld deferred hardware errors detected in total in %s block\n",
+					      mcm_info->socket_id, mcm_info->die_id,
+					      err_info->de_count, blk_name);
+			}
+		} else {
+			if (adev->debug_disable_ce_logs)
+				return;
+
+			for_each_ras_error(err_node, err_data) {
+				err_info = &err_node->err_info;
+				mcm_info = &err_info->mcm_info;
+				if (err_info->ce_count) {
+					RAS_EVENT_LOG(adev, event_id, "socket: %d, die: %d, "
+						      "%lld new correctable hardware errors detected in %s block\n",
+						      mcm_info->socket_id,
+						      mcm_info->die_id,
+						      err_info->ce_count,
+						      blk_name);
+				}
+			}
+
+			for_each_ras_error(err_node, &ras_mgr->err_data) {
+				err_info = &err_node->err_info;
+				mcm_info = &err_info->mcm_info;
+				RAS_EVENT_LOG(adev, event_id, "socket: %d, die: %d, "
+					      "%lld correctable hardware errors detected in total in %s block\n",
+					      mcm_info->socket_id, mcm_info->die_id,
+					      err_info->ce_count, blk_name);
+			}
+		}
 	}
+}
 
-	obj->err_data.ue_count += err_data.ue_count;
-	obj->err_data.ce_count += err_data.ce_count;
+static inline bool err_data_has_source_info(struct ras_err_data *data)
+{
+	return !list_empty(&data->err_node_list);
+}
 
-	info->ue_count = obj->err_data.ue_count;
-	info->ce_count = obj->err_data.ce_count;
+static void amdgpu_ras_error_generate_report(struct amdgpu_device *adev,
+					     struct ras_query_if *query_if,
+					     struct ras_err_data *err_data,
+					     struct ras_query_context *qctx)
+{
+	struct ras_manager *ras_mgr = amdgpu_ras_find_obj(adev, &query_if->head);
+	const char *blk_name = get_ras_block_str(&query_if->head);
+	u64 event_id = qctx->evid.event_id;
+
+	if (err_data->ce_count) {
+		if (err_data_has_source_info(err_data)) {
+			amdgpu_ras_error_print_error_data(adev, ras_mgr, err_data, qctx,
+							  blk_name, false, false);
+		} else if (!adev->aid_mask &&
+			   adev->smuio.funcs &&
+			   adev->smuio.funcs->get_socket_id &&
+			   adev->smuio.funcs->get_die_id) {
+			RAS_EVENT_LOG(adev, event_id, "socket: %d, die: %d "
+				      "%ld correctable hardware errors "
+				      "detected in %s block\n",
+				      adev->smuio.funcs->get_socket_id(adev),
+				      adev->smuio.funcs->get_die_id(adev),
+				      ras_mgr->err_data.ce_count,
+				      blk_name);
+		} else {
+			RAS_EVENT_LOG(adev, event_id, "%ld correctable hardware errors "
+				      "detected in %s block\n",
+				      ras_mgr->err_data.ce_count,
+				      blk_name);
+		}
+	}
 
-	if (err_data.ce_count) {
-		if (adev->smuio.funcs &&
-		    adev->smuio.funcs->get_socket_id &&
-		    adev->smuio.funcs->get_die_id) {
-			dev_info(adev->dev, "socket: %d, die: %d "
-					"%ld correctable hardware errors "
-					"detected in %s block, no user "
-					"action is needed.\n",
-					adev->smuio.funcs->get_socket_id(adev),
-					adev->smuio.funcs->get_die_id(adev),
-					obj->err_data.ce_count,
-					get_ras_block_str(&info->head));
+	if (err_data->ue_count) {
+		if (err_data_has_source_info(err_data)) {
+			amdgpu_ras_error_print_error_data(adev, ras_mgr, err_data, qctx,
+							  blk_name, true, false);
+		} else if (!adev->aid_mask &&
+			   adev->smuio.funcs &&
+			   adev->smuio.funcs->get_socket_id &&
+			   adev->smuio.funcs->get_die_id) {
+			RAS_EVENT_LOG(adev, event_id, "socket: %d, die: %d "
+				      "%ld uncorrectable hardware errors "
+				      "detected in %s block\n",
+				      adev->smuio.funcs->get_socket_id(adev),
+				      adev->smuio.funcs->get_die_id(adev),
+				      ras_mgr->err_data.ue_count,
+				      blk_name);
 		} else {
-			dev_info(adev->dev, "%ld correctable hardware errors "
-					"detected in %s block, no user "
-					"action is needed.\n",
-					obj->err_data.ce_count,
-					get_ras_block_str(&info->head));
-		}
-	}
-	if (err_data.ue_count) {
-		if (adev->smuio.funcs &&
-		    adev->smuio.funcs->get_socket_id &&
-		    adev->smuio.funcs->get_die_id) {
-			dev_info(adev->dev, "socket: %d, die: %d "
-					"%ld uncorrectable hardware errors "
-					"detected in %s block\n",
-					adev->smuio.funcs->get_socket_id(adev),
-					adev->smuio.funcs->get_die_id(adev),
-					obj->err_data.ue_count,
-					get_ras_block_str(&info->head));
+			RAS_EVENT_LOG(adev, event_id, "%ld uncorrectable hardware errors "
+				      "detected in %s block\n",
+				      ras_mgr->err_data.ue_count,
+				      blk_name);
+		}
+	}
+
+	if (err_data->de_count) {
+		if (err_data_has_source_info(err_data)) {
+			amdgpu_ras_error_print_error_data(adev, ras_mgr, err_data, qctx,
+							  blk_name, false, true);
+		} else if (!adev->aid_mask &&
+			   adev->smuio.funcs &&
+			   adev->smuio.funcs->get_socket_id &&
+			   adev->smuio.funcs->get_die_id) {
+			RAS_EVENT_LOG(adev, event_id, "socket: %d, die: %d "
+				      "%ld deferred hardware errors "
+				      "detected in %s block\n",
+				      adev->smuio.funcs->get_socket_id(adev),
+				      adev->smuio.funcs->get_die_id(adev),
+				      ras_mgr->err_data.de_count,
+				      blk_name);
 		} else {
-			dev_info(adev->dev, "%ld uncorrectable hardware errors "
-					"detected in %s block\n",
-					obj->err_data.ue_count,
-					get_ras_block_str(&info->head));
+			RAS_EVENT_LOG(adev, event_id, "%ld deferred hardware errors "
+				      "detected in %s block\n",
+				      ras_mgr->err_data.de_count,
+				      blk_name);
 		}
 	}
+}
 
-	if (!amdgpu_persistent_edc_harvesting_supported(adev))
-		amdgpu_ras_reset_error_status(adev, info->head.block);
+static void amdgpu_ras_virt_error_generate_report(struct amdgpu_device *adev,
+						  struct ras_query_if *query_if,
+						  struct ras_err_data *err_data,
+						  struct ras_query_context *qctx)
+{
+	unsigned long new_ue, new_ce, new_de;
+	struct ras_manager *obj = amdgpu_ras_find_obj(adev, &query_if->head);
+	const char *blk_name = get_ras_block_str(&query_if->head);
+	u64 event_id = qctx->evid.event_id;
+
+	new_ce = err_data->ce_count - obj->err_data.ce_count;
+	new_ue = err_data->ue_count - obj->err_data.ue_count;
+	new_de = err_data->de_count - obj->err_data.de_count;
+
+	if (new_ce) {
+		RAS_EVENT_LOG(adev, event_id, "%lu correctable hardware errors "
+			      "detected in %s block\n",
+			      new_ce,
+			      blk_name);
+	}
+
+	if (new_ue) {
+		RAS_EVENT_LOG(adev, event_id, "%lu uncorrectable hardware errors "
+			      "detected in %s block\n",
+			      new_ue,
+			      blk_name);
+	}
+
+	if (new_de) {
+		RAS_EVENT_LOG(adev, event_id, "%lu deferred hardware errors "
+			      "detected in %s block\n",
+			      new_de,
+			      blk_name);
+	}
+}
+
+static void amdgpu_rasmgr_error_data_statistic_update(struct ras_manager *obj, struct ras_err_data *err_data)
+{
+	struct ras_err_node *err_node;
+	struct ras_err_info *err_info;
+
+	if (err_data_has_source_info(err_data)) {
+		for_each_ras_error(err_node, err_data) {
+			err_info = &err_node->err_info;
+			amdgpu_ras_error_statistic_de_count(&obj->err_data,
+					&err_info->mcm_info, err_info->de_count);
+			amdgpu_ras_error_statistic_ce_count(&obj->err_data,
+					&err_info->mcm_info, err_info->ce_count);
+			amdgpu_ras_error_statistic_ue_count(&obj->err_data,
+					&err_info->mcm_info, err_info->ue_count);
+		}
+	} else {
+		/* for legacy asic path which doesn't has error source info */
+		obj->err_data.ue_count += err_data->ue_count;
+		obj->err_data.ce_count += err_data->ce_count;
+		obj->err_data.de_count += err_data->de_count;
+	}
+}
+
+static void amdgpu_ras_mgr_virt_error_data_statistics_update(struct ras_manager *obj,
+							     struct ras_err_data *err_data)
+{
+	/* Host reports absolute counts */
+	obj->err_data.ue_count = err_data->ue_count;
+	obj->err_data.ce_count = err_data->ce_count;
+	obj->err_data.de_count = err_data->de_count;
+}
+
+static struct ras_manager *get_ras_manager(struct amdgpu_device *adev, enum amdgpu_ras_block blk)
+{
+	struct ras_common_if head;
+
+	memset(&head, 0, sizeof(head));
+	head.block = blk;
+
+	return amdgpu_ras_find_obj(adev, &head);
+}
+
+int amdgpu_ras_bind_aca(struct amdgpu_device *adev, enum amdgpu_ras_block blk,
+			const struct aca_info *aca_info, void *data)
+{
+	struct ras_manager *obj;
+
+	/* in resume phase, no need to create aca fs node */
+	if (adev->in_suspend || amdgpu_reset_in_recovery(adev))
+		return 0;
+
+	obj = get_ras_manager(adev, blk);
+	if (!obj)
+		return -EINVAL;
+
+	return amdgpu_aca_add_handle(adev, &obj->aca_handle, ras_block_str(blk), aca_info, data);
+}
+
+int amdgpu_ras_unbind_aca(struct amdgpu_device *adev, enum amdgpu_ras_block blk)
+{
+	struct ras_manager *obj;
+
+	obj = get_ras_manager(adev, blk);
+	if (!obj)
+		return -EINVAL;
+
+	amdgpu_aca_remove_handle(&obj->aca_handle);
 
 	return 0;
 }
 
-int amdgpu_ras_reset_error_status(struct amdgpu_device *adev,
-		enum amdgpu_ras_block block)
+static int amdgpu_aca_log_ras_error_data(struct amdgpu_device *adev, enum amdgpu_ras_block blk,
+					 enum aca_error_type type, struct ras_err_data *err_data,
+					 struct ras_query_context *qctx)
+{
+	struct ras_manager *obj;
+
+	obj = get_ras_manager(adev, blk);
+	if (!obj)
+		return -EINVAL;
+
+	return amdgpu_aca_get_error_data(adev, &obj->aca_handle, type, err_data, qctx);
+}
+
+ssize_t amdgpu_ras_aca_sysfs_read(struct device *dev, struct device_attribute *attr,
+				  struct aca_handle *handle, char *buf, void *data)
 {
-	if (!amdgpu_ras_is_supported(adev, block))
+	struct ras_manager *obj = container_of(handle, struct ras_manager, aca_handle);
+	struct ras_query_if info = {
+		.head = obj->head,
+	};
+
+	if (!amdgpu_ras_get_error_query_ready(obj->adev))
+		return sysfs_emit(buf, "Query currently inaccessible\n");
+
+	if (amdgpu_ras_query_error_status(obj->adev, &info))
 		return -EINVAL;
 
-	switch (block) {
-	case AMDGPU_RAS_BLOCK__GFX:
-		if (adev->gfx.ras_funcs &&
-		    adev->gfx.ras_funcs->reset_ras_error_count)
-			adev->gfx.ras_funcs->reset_ras_error_count(adev);
+	return sysfs_emit(buf, "%s: %lu\n%s: %lu\n%s: %lu\n", "ue", info.ue_count,
+			  "ce", info.ce_count, "de", info.de_count);
+}
 
-		if (adev->gfx.ras_funcs &&
-		    adev->gfx.ras_funcs->reset_ras_error_status)
-			adev->gfx.ras_funcs->reset_ras_error_status(adev);
-		break;
-	case AMDGPU_RAS_BLOCK__MMHUB:
-		if (adev->mmhub.ras_funcs &&
-		    adev->mmhub.ras_funcs->reset_ras_error_count)
-			adev->mmhub.ras_funcs->reset_ras_error_count(adev);
-
-		if (adev->mmhub.ras_funcs &&
-		    adev->mmhub.ras_funcs->reset_ras_error_status)
-			adev->mmhub.ras_funcs->reset_ras_error_status(adev);
-		break;
-	case AMDGPU_RAS_BLOCK__SDMA:
-		if (adev->sdma.funcs->reset_ras_error_count)
-			adev->sdma.funcs->reset_ras_error_count(adev);
-		break;
-	case AMDGPU_RAS_BLOCK__HDP:
-		if (adev->hdp.ras_funcs &&
-		    adev->hdp.ras_funcs->reset_ras_error_count)
-			adev->hdp.ras_funcs->reset_ras_error_count(adev);
-		break;
-	default:
-		break;
+static int amdgpu_ras_query_error_status_helper(struct amdgpu_device *adev,
+						struct ras_query_if *info,
+						struct ras_err_data *err_data,
+						struct ras_query_context *qctx,
+						unsigned int error_query_mode)
+{
+	enum amdgpu_ras_block blk = info ? info->head.block : AMDGPU_RAS_BLOCK_COUNT;
+	struct amdgpu_ras_block_object *block_obj = NULL;
+	int ret;
+
+	if (blk == AMDGPU_RAS_BLOCK_COUNT)
+		return -EINVAL;
+
+	if (error_query_mode == AMDGPU_RAS_INVALID_ERROR_QUERY)
+		return -EINVAL;
+
+	if (error_query_mode == AMDGPU_RAS_VIRT_ERROR_COUNT_QUERY) {
+		return amdgpu_virt_req_ras_err_count(adev, blk, err_data);
+	} else if (error_query_mode == AMDGPU_RAS_DIRECT_ERROR_QUERY) {
+		if (info->head.block == AMDGPU_RAS_BLOCK__UMC) {
+			amdgpu_ras_get_ecc_info(adev, err_data);
+		} else {
+			block_obj = amdgpu_ras_get_ras_block(adev, info->head.block, 0);
+			if (!block_obj || !block_obj->hw_ops) {
+				dev_dbg_once(adev->dev, "%s doesn't config RAS function\n",
+					     get_ras_block_str(&info->head));
+				return -EINVAL;
+			}
+
+			if (block_obj->hw_ops->query_ras_error_count)
+				block_obj->hw_ops->query_ras_error_count(adev, err_data);
+
+			if ((info->head.block == AMDGPU_RAS_BLOCK__SDMA) ||
+			    (info->head.block == AMDGPU_RAS_BLOCK__GFX) ||
+			    (info->head.block == AMDGPU_RAS_BLOCK__MMHUB)) {
+				if (block_obj->hw_ops->query_ras_error_status)
+					block_obj->hw_ops->query_ras_error_status(adev);
+			}
+		}
+	} else {
+		if (amdgpu_aca_is_enabled(adev)) {
+			ret = amdgpu_aca_log_ras_error_data(adev, blk, ACA_ERROR_TYPE_UE, err_data, qctx);
+			if (ret)
+				return ret;
+
+			ret = amdgpu_aca_log_ras_error_data(adev, blk, ACA_ERROR_TYPE_CE, err_data, qctx);
+			if (ret)
+				return ret;
+
+			ret = amdgpu_aca_log_ras_error_data(adev, blk, ACA_ERROR_TYPE_DEFERRED, err_data, qctx);
+			if (ret)
+				return ret;
+		} else {
+			/* FIXME: add code to check return value later */
+			amdgpu_mca_smu_log_ras_error(adev, blk, AMDGPU_MCA_ERROR_TYPE_UE, err_data, qctx);
+			amdgpu_mca_smu_log_ras_error(adev, blk, AMDGPU_MCA_ERROR_TYPE_CE, err_data, qctx);
+		}
 	}
 
 	return 0;
 }
 
-/* Trigger XGMI/WAFL error */
-static int amdgpu_ras_error_inject_xgmi(struct amdgpu_device *adev,
-				 struct ta_ras_trigger_error_input *block_info)
+/* query/inject/cure begin */
+static int amdgpu_ras_query_error_status_with_event(struct amdgpu_device *adev,
+						    struct ras_query_if *info,
+						    enum ras_event_type type)
 {
+	struct ras_manager *obj = amdgpu_ras_find_obj(adev, &info->head);
+	struct ras_err_data err_data;
+	struct ras_query_context qctx;
+	unsigned int error_query_mode;
 	int ret;
 
-	if (amdgpu_dpm_set_df_cstate(adev, DF_CSTATE_DISALLOW))
-		dev_warn(adev->dev, "Failed to disallow df cstate");
+	if (!obj)
+		return -EINVAL;
+
+	ret = amdgpu_ras_error_data_init(&err_data);
+	if (ret)
+		return ret;
+
+	if (!amdgpu_ras_get_error_query_mode(adev, &error_query_mode))
+		return -EINVAL;
+
+	memset(&qctx, 0, sizeof(qctx));
+	qctx.evid.type = type;
+	qctx.evid.event_id = amdgpu_ras_acquire_event_id(adev, type);
 
-	if (amdgpu_dpm_allow_xgmi_power_down(adev, false))
-		dev_warn(adev->dev, "Failed to disallow XGMI power down");
+	if (!down_read_trylock(&adev->reset_domain->sem)) {
+		ret = -EIO;
+		goto out_fini_err_data;
+	}
 
-	ret = psp_ras_trigger_error(&adev->psp, block_info);
+	ret = amdgpu_ras_query_error_status_helper(adev, info,
+						   &err_data,
+						   &qctx,
+						   error_query_mode);
+	up_read(&adev->reset_domain->sem);
+	if (ret)
+		goto out_fini_err_data;
 
-	if (amdgpu_ras_intr_triggered())
+	if (error_query_mode != AMDGPU_RAS_VIRT_ERROR_COUNT_QUERY) {
+		amdgpu_rasmgr_error_data_statistic_update(obj, &err_data);
+		amdgpu_ras_error_generate_report(adev, info, &err_data, &qctx);
+	} else {
+		/* Host provides absolute error counts. First generate the report
+		 * using the previous VF internal count against new host count.
+		 * Then Update VF internal count.
+		 */
+		amdgpu_ras_virt_error_generate_report(adev, info, &err_data, &qctx);
+		amdgpu_ras_mgr_virt_error_data_statistics_update(obj, &err_data);
+	}
+
+	info->ue_count = obj->err_data.ue_count;
+	info->ce_count = obj->err_data.ce_count;
+	info->de_count = obj->err_data.de_count;
+
+out_fini_err_data:
+	amdgpu_ras_error_data_fini(&err_data);
+
+	return ret;
+}
+
+static int amdgpu_uniras_clear_badpages_info(struct amdgpu_device *adev)
+{
+	struct ras_cmd_dev_handle req = {0};
+	int ret;
+
+	ret = amdgpu_ras_mgr_handle_ras_cmd(adev, RAS_CMD__CLEAR_BAD_PAGE_INFO,
+				&req, sizeof(req), NULL, 0);
+	if (ret) {
+		dev_err(adev->dev, "Failed to clear bad pages info, ret: %d\n", ret);
 		return ret;
+	}
+
+	return 0;
+}
+
+static int amdgpu_uniras_query_block_ecc(struct amdgpu_device *adev,
+			struct ras_query_if *info)
+{
+	struct ras_cmd_block_ecc_info_req req = {0};
+	struct ras_cmd_block_ecc_info_rsp rsp = {0};
+	int ret;
+
+	if (!info)
+		return -EINVAL;
 
-	if (amdgpu_dpm_allow_xgmi_power_down(adev, true))
-		dev_warn(adev->dev, "Failed to allow XGMI power down");
+	req.block_id = info->head.block;
+	req.subblock_id = info->head.sub_block_index;
 
-	if (amdgpu_dpm_set_df_cstate(adev, DF_CSTATE_ALLOW))
-		dev_warn(adev->dev, "Failed to allow df cstate");
+	ret = amdgpu_ras_mgr_handle_ras_cmd(adev, RAS_CMD__GET_BLOCK_ECC_STATUS,
+				&req, sizeof(req), &rsp, sizeof(rsp));
+	if (!ret) {
+		info->ce_count = rsp.ce_count;
+		info->ue_count = rsp.ue_count;
+		info->de_count = rsp.de_count;
+	}
 
 	return ret;
 }
 
+int amdgpu_ras_query_error_status(struct amdgpu_device *adev, struct ras_query_if *info)
+{
+	if (amdgpu_uniras_enabled(adev))
+		return amdgpu_uniras_query_block_ecc(adev, info);
+	else
+		return amdgpu_ras_query_error_status_with_event(adev, info, RAS_EVENT_TYPE_INVALID);
+}
+
+int amdgpu_ras_reset_error_count(struct amdgpu_device *adev,
+		enum amdgpu_ras_block block)
+{
+	struct amdgpu_ras_block_object *block_obj = amdgpu_ras_get_ras_block(adev, block, 0);
+	const struct amdgpu_mca_smu_funcs *mca_funcs = adev->mca.mca_funcs;
+	const struct aca_smu_funcs *smu_funcs = adev->aca.smu_funcs;
+
+	if (!block_obj || !block_obj->hw_ops) {
+		dev_dbg_once(adev->dev, "%s doesn't config RAS function\n",
+				ras_block_str(block));
+		return -EOPNOTSUPP;
+	}
+
+	if (!amdgpu_ras_is_supported(adev, block) ||
+	    !amdgpu_ras_get_aca_debug_mode(adev))
+		return -EOPNOTSUPP;
+
+	if (amdgpu_sriov_vf(adev))
+		return -EOPNOTSUPP;
+
+	/* skip ras error reset in gpu reset */
+	if ((amdgpu_in_reset(adev) || amdgpu_ras_in_recovery(adev)) &&
+	    ((smu_funcs && smu_funcs->set_debug_mode) ||
+	     (mca_funcs && mca_funcs->mca_set_debug_mode)))
+		return -EOPNOTSUPP;
+
+	if (block_obj->hw_ops->reset_ras_error_count)
+		block_obj->hw_ops->reset_ras_error_count(adev);
+
+	return 0;
+}
+
+int amdgpu_ras_reset_error_status(struct amdgpu_device *adev,
+		enum amdgpu_ras_block block)
+{
+	struct amdgpu_ras_block_object *block_obj = amdgpu_ras_get_ras_block(adev, block, 0);
+
+	if (amdgpu_ras_reset_error_count(adev, block) == -EOPNOTSUPP)
+		return 0;
+
+	if ((block == AMDGPU_RAS_BLOCK__GFX) ||
+	    (block == AMDGPU_RAS_BLOCK__MMHUB)) {
+		if (block_obj->hw_ops->reset_ras_error_status)
+			block_obj->hw_ops->reset_ras_error_status(adev);
+	}
+
+	return 0;
+}
+
+static int amdgpu_uniras_error_inject(struct amdgpu_device *adev,
+		struct ras_inject_if *info)
+{
+	struct ras_cmd_inject_error_req inject_req;
+	struct ras_cmd_inject_error_rsp rsp;
+
+	if (!info)
+		return -EINVAL;
+
+	memset(&inject_req, 0, sizeof(inject_req));
+	inject_req.block_id = info->head.block;
+	inject_req.subblock_id = info->head.sub_block_index;
+	inject_req.address = info->address;
+	inject_req.error_type = info->head.type;
+	inject_req.instance_mask = info->instance_mask;
+	inject_req.method = info->value;
+
+	return amdgpu_ras_mgr_handle_ras_cmd(adev, RAS_CMD__INJECT_ERROR,
+			&inject_req, sizeof(inject_req), &rsp, sizeof(rsp));
+}
+
 /* wrapper of psp_ras_trigger_error */
 int amdgpu_ras_error_inject(struct amdgpu_device *adev,
 		struct ras_inject_if *info)
@@ -1092,40 +1681,44 @@ int amdgpu_ras_error_inject(struct amdgpu_device *adev,
 		.address = info->address,
 		.value = info->value,
 	};
-	int ret = 0;
+	int ret = -EINVAL;
+	struct amdgpu_ras_block_object *block_obj = amdgpu_ras_get_ras_block(adev,
+							info->head.block,
+							info->head.sub_block_index);
+
+	if (amdgpu_uniras_enabled(adev))
+		return amdgpu_uniras_error_inject(adev, info);
+
+	/* inject on guest isn't allowed, return success directly */
+	if (amdgpu_sriov_vf(adev))
+		return 0;
 
 	if (!obj)
 		return -EINVAL;
 
+	if (!block_obj || !block_obj->hw_ops)	{
+		dev_dbg_once(adev->dev, "%s doesn't config RAS function\n",
+			     get_ras_block_str(&info->head));
+		return -EINVAL;
+	}
+
 	/* Calculate XGMI relative offset */
-	if (adev->gmc.xgmi.num_physical_nodes > 1) {
+	if (adev->gmc.xgmi.num_physical_nodes > 1 &&
+	    info->head.block != AMDGPU_RAS_BLOCK__GFX) {
 		block_info.address =
 			amdgpu_xgmi_get_relative_phy_addr(adev,
 							  block_info.address);
 	}
 
-	switch (info->head.block) {
-	case AMDGPU_RAS_BLOCK__GFX:
-		if (adev->gfx.ras_funcs &&
-		    adev->gfx.ras_funcs->ras_error_inject)
-			ret = adev->gfx.ras_funcs->ras_error_inject(adev, info);
-		else
-			ret = -EINVAL;
-		break;
-	case AMDGPU_RAS_BLOCK__UMC:
-	case AMDGPU_RAS_BLOCK__SDMA:
-	case AMDGPU_RAS_BLOCK__MMHUB:
-	case AMDGPU_RAS_BLOCK__PCIE_BIF:
-	case AMDGPU_RAS_BLOCK__MCA:
-		ret = psp_ras_trigger_error(&adev->psp, &block_info);
-		break;
-	case AMDGPU_RAS_BLOCK__XGMI_WAFL:
-		ret = amdgpu_ras_error_inject_xgmi(adev, &block_info);
-		break;
-	default:
-		dev_info(adev->dev, "%s error injection is not supported yet\n",
-			 get_ras_block_str(&info->head));
-		ret = -EINVAL;
+	if (block_obj->hw_ops->ras_error_inject) {
+		if (info->head.block == AMDGPU_RAS_BLOCK__GFX)
+			ret = block_obj->hw_ops->ras_error_inject(adev, info, info->instance_mask);
+		else /* Special ras_error_inject is defined (e.g: xgmi) */
+			ret = block_obj->hw_ops->ras_error_inject(adev, &block_info,
+						info->instance_mask);
+	} else {
+		/* default path */
+		ret = psp_ras_trigger_error(&adev->psp, &block_info, info->instance_mask);
 	}
 
 	if (ret)
@@ -1136,11 +1729,54 @@ int amdgpu_ras_error_inject(struct amdgpu_device *adev,
 }
 
 /**
- * amdgpu_ras_query_error_count -- Get error counts of all IPs
- * adev: pointer to AMD GPU device
- * ce_count: pointer to an integer to be set to the count of correctible errors.
- * ue_count: pointer to an integer to be set to the count of uncorrectible
+ * amdgpu_ras_query_error_count_helper -- Get error counter for specific IP
+ * @adev: pointer to AMD GPU device
+ * @ce_count: pointer to an integer to be set to the count of correctible errors.
+ * @ue_count: pointer to an integer to be set to the count of uncorrectible errors.
+ * @query_info: pointer to ras_query_if
+ *
+ * Return 0 for query success or do nothing, otherwise return an error
+ * on failures
+ */
+static int amdgpu_ras_query_error_count_helper(struct amdgpu_device *adev,
+					       unsigned long *ce_count,
+					       unsigned long *ue_count,
+					       struct ras_query_if *query_info)
+{
+	int ret;
+
+	if (!query_info)
+		/* do nothing if query_info is not specified */
+		return 0;
+
+	ret = amdgpu_ras_query_error_status(adev, query_info);
+	if (ret)
+		return ret;
+
+	*ce_count += query_info->ce_count;
+	*ue_count += query_info->ue_count;
+
+	/* some hardware/IP supports read to clear
+	 * no need to explictly reset the err status after the query call */
+	if (amdgpu_ip_version(adev, MP0_HWIP, 0) != IP_VERSION(11, 0, 2) &&
+	    amdgpu_ip_version(adev, MP0_HWIP, 0) != IP_VERSION(11, 0, 4)) {
+		if (amdgpu_ras_reset_error_status(adev, query_info->head.block))
+			dev_warn(adev->dev,
+				 "Failed to reset error counter and error status\n");
+	}
+
+	return 0;
+}
+
+/**
+ * amdgpu_ras_query_error_count -- Get error counts of all IPs or specific IP
+ * @adev: pointer to AMD GPU device
+ * @ce_count: pointer to an integer to be set to the count of correctible errors.
+ * @ue_count: pointer to an integer to be set to the count of uncorrectible
  * errors.
+ * @query_info: pointer to ras_query_if if the query request is only for
+ * specific ip block; if info is NULL, then the qurey request is for
+ * all the ip blocks that support query ras error counters/status
  *
  * If set, @ce_count or @ue_count, count and return the corresponding
  * error counts in those integer pointers. Return 0 if the device
@@ -1148,11 +1784,13 @@ int amdgpu_ras_error_inject(struct amdgpu_device *adev,
  */
 int amdgpu_ras_query_error_count(struct amdgpu_device *adev,
 				 unsigned long *ce_count,
-				 unsigned long *ue_count)
+				 unsigned long *ue_count,
+				 struct ras_query_if *query_info)
 {
 	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
 	struct ras_manager *obj;
 	unsigned long ce, ue;
+	int ret;
 
 	if (!adev->ras_enabled || !con)
 		return -EOPNOTSUPP;
@@ -1164,20 +1802,23 @@ int amdgpu_ras_query_error_count(struct amdgpu_device *adev,
 
 	ce = 0;
 	ue = 0;
-	list_for_each_entry(obj, &con->head, node) {
-		struct ras_query_if info = {
-			.head = obj->head,
-		};
-		int res;
-
-		res = amdgpu_ras_query_error_status(adev, &info);
-		if (res)
-			return res;
-
-		ce += info.ce_count;
-		ue += info.ue_count;
+	if (!query_info) {
+		/* query all the ip blocks that support ras query interface */
+		list_for_each_entry(obj, &con->head, node) {
+			struct ras_query_if info = {
+				.head = obj->head,
+			};
+
+			ret = amdgpu_ras_query_error_count_helper(adev, &ce, &ue, &info);
+		}
+	} else {
+		/* query specific ip block */
+		ret = amdgpu_ras_query_error_count_helper(adev, &ce, &ue, query_info);
 	}
 
+	if (ret)
+		return ret;
+
 	if (ce_count)
 		*ce_count = ce;
 
@@ -1192,7 +1833,9 @@ int amdgpu_ras_query_error_count(struct amdgpu_device *adev,
 /* sysfs begin */
 
 static int amdgpu_ras_badpages_read(struct amdgpu_device *adev,
-		struct ras_badpage **bps, unsigned int *count);
+		struct ras_badpage *bps, uint32_t count, uint32_t start);
+static int amdgpu_uniras_badpages_read(struct amdgpu_device *adev,
+		struct ras_badpage *bps, uint32_t count, uint32_t start);
 
 static char *amdgpu_ras_badpage_flags_str(unsigned int flags)
 {
@@ -1238,7 +1881,7 @@ static char *amdgpu_ras_badpage_flags_str(unsigned int flags)
  */
 
 static ssize_t amdgpu_ras_sysfs_badpages_read(struct file *f,
-		struct kobject *kobj, struct bin_attribute *attr,
+		struct kobject *kobj, const struct bin_attribute *attr,
 		char *buf, loff_t ppos, size_t count)
 {
 	struct amdgpu_ras *con =
@@ -1250,19 +1893,50 @@ static ssize_t amdgpu_ras_sysfs_badpages_read(struct file *f,
 	unsigned int end = div64_ul(ppos + count - 1, element_size);
 	ssize_t s = 0;
 	struct ras_badpage *bps = NULL;
-	unsigned int bps_count = 0;
+	int bps_count = 0, i, status;
+	uint64_t address;
 
 	memset(buf, 0, count);
 
-	if (amdgpu_ras_badpages_read(adev, &bps, &bps_count))
+	bps_count = end - start;
+	bps = kmalloc_array(bps_count, sizeof(*bps), GFP_KERNEL);
+	if (!bps)
+		return 0;
+
+	memset(bps, 0, sizeof(*bps) * bps_count);
+
+	if (amdgpu_uniras_enabled(adev))
+		bps_count = amdgpu_uniras_badpages_read(adev, bps, bps_count, start);
+	else
+		bps_count = amdgpu_ras_badpages_read(adev, bps, bps_count, start);
+
+	if (bps_count <= 0) {
+		kfree(bps);
 		return 0;
+	}
+
+	for (i = 0; i < bps_count; i++) {
+		address = ((uint64_t)bps[i].bp) << AMDGPU_GPU_PAGE_SHIFT;
+		if (amdgpu_ras_check_critical_address(adev, address))
+			continue;
+
+		bps[i].size = AMDGPU_GPU_PAGE_SIZE;
+
+		status = amdgpu_vram_mgr_query_page_status(&adev->mman.vram_mgr,
+					address);
+		if (status == -EBUSY)
+			bps[i].flags = AMDGPU_RAS_RETIRE_PAGE_PENDING;
+		else if (status == -ENOENT)
+			bps[i].flags = AMDGPU_RAS_RETIRE_PAGE_FAULT;
+		else
+			bps[i].flags = AMDGPU_RAS_RETIRE_PAGE_RESERVED;
 
-	for (; start < end && start < bps_count; start++)
 		s += scnprintf(&buf[s], element_size + 1,
 				"0x%08x : 0x%08x : %1s\n",
-				bps[start].bp,
-				bps[start].size,
-				amdgpu_ras_badpage_flags_str(bps[start].flags));
+				bps[i].bp,
+				bps[i].size,
+				amdgpu_ras_badpage_flags_str(bps[i].flags));
+	}
 
 	kfree(bps);
 
@@ -1275,23 +1949,106 @@ static ssize_t amdgpu_ras_sysfs_features_read(struct device *dev,
 	struct amdgpu_ras *con =
 		container_of(attr, struct amdgpu_ras, features_attr);
 
-	return scnprintf(buf, PAGE_SIZE, "feature mask: 0x%x\n", con->features);
+	return sysfs_emit(buf, "feature mask: 0x%x\n", con->features);
+}
+
+static bool amdgpu_ras_get_version_info(struct amdgpu_device *adev, u32 *major,
+			u32 *minor, u32 *rev)
+{
+	int i;
+
+	if (!adev || !major || !minor || !rev || !amdgpu_uniras_enabled(adev))
+		return false;
+
+	for (i = 0; i < adev->num_ip_blocks; i++) {
+		if (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_RAS) {
+			*major = adev->ip_blocks[i].version->major;
+			*minor = adev->ip_blocks[i].version->minor;
+			*rev = adev->ip_blocks[i].version->rev;
+			return true;
+		}
+	}
+
+	return false;
+}
+
+static ssize_t amdgpu_ras_sysfs_version_show(struct device *dev,
+		struct device_attribute *attr, char *buf)
+{
+	struct amdgpu_ras *con =
+		container_of(attr, struct amdgpu_ras, version_attr);
+	u32 major, minor, rev;
+	ssize_t size = 0;
+
+	size += sysfs_emit_at(buf, size, "table version: 0x%x\n",
+			con->eeprom_control.tbl_hdr.version);
+
+	if (amdgpu_ras_get_version_info(con->adev, &major, &minor, &rev))
+		size += sysfs_emit_at(buf, size, "ras version: %u.%u.%u\n",
+			major, minor, rev);
+
+	return size;
+}
+
+static ssize_t amdgpu_ras_sysfs_schema_show(struct device *dev,
+		struct device_attribute *attr, char *buf)
+{
+	struct amdgpu_ras *con =
+		container_of(attr, struct amdgpu_ras, schema_attr);
+	return sysfs_emit(buf, "schema: 0x%x\n", con->schema);
+}
+
+static struct {
+	enum ras_event_type type;
+	const char *name;
+} dump_event[] = {
+	{RAS_EVENT_TYPE_FATAL, "Fatal Error"},
+	{RAS_EVENT_TYPE_POISON_CREATION, "Poison Creation"},
+	{RAS_EVENT_TYPE_POISON_CONSUMPTION, "Poison Consumption"},
+};
+
+static ssize_t amdgpu_ras_sysfs_event_state_show(struct device *dev,
+						 struct device_attribute *attr, char *buf)
+{
+	struct amdgpu_ras *con =
+		container_of(attr, struct amdgpu_ras, event_state_attr);
+	struct ras_event_manager *event_mgr = con->event_mgr;
+	struct ras_event_state *event_state;
+	int i, size = 0;
+
+	if (!event_mgr)
+		return -EINVAL;
+
+	size += sysfs_emit_at(buf, size, "current seqno: %llu\n", atomic64_read(&event_mgr->seqno));
+	for (i = 0; i < ARRAY_SIZE(dump_event); i++) {
+		event_state = &event_mgr->event_state[dump_event[i].type];
+		size += sysfs_emit_at(buf, size, "%s: count:%llu, last_seqno:%llu\n",
+				      dump_event[i].name,
+				      atomic64_read(&event_state->count),
+				      event_state->last_seqno);
+	}
+
+	return (ssize_t)size;
 }
 
 static void amdgpu_ras_sysfs_remove_bad_page_node(struct amdgpu_device *adev)
 {
 	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
 
-	sysfs_remove_file_from_group(&adev->dev->kobj,
+	if (adev->dev->kobj.sd)
+		sysfs_remove_file_from_group(&adev->dev->kobj,
 				&con->badpages_attr.attr,
 				RAS_FS_NAME);
 }
 
-static int amdgpu_ras_sysfs_remove_feature_node(struct amdgpu_device *adev)
+static int amdgpu_ras_sysfs_remove_dev_attr_node(struct amdgpu_device *adev)
 {
 	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
 	struct attribute *attrs[] = {
 		&con->features_attr.attr,
+		&con->version_attr.attr,
+		&con->schema_attr.attr,
+		&con->event_state_attr.attr,
 		NULL
 	};
 	struct attribute_group group = {
@@ -1299,24 +2056,30 @@ static int amdgpu_ras_sysfs_remove_feature_node(struct amdgpu_device *adev)
 		.attrs = attrs,
 	};
 
-	sysfs_remove_group(&adev->dev->kobj, &group);
+	if (adev->dev->kobj.sd)
+		sysfs_remove_group(&adev->dev->kobj, &group);
 
 	return 0;
 }
 
 int amdgpu_ras_sysfs_create(struct amdgpu_device *adev,
-		struct ras_fs_if *head)
+		struct ras_common_if *head)
 {
-	struct ras_manager *obj = amdgpu_ras_find_obj(adev, &head->head);
+	struct ras_manager *obj = amdgpu_ras_find_obj(adev, head);
+
+	if (amdgpu_aca_is_enabled(adev))
+		return 0;
 
 	if (!obj || obj->attr_inuse)
 		return -EINVAL;
 
+	if (amdgpu_sriov_vf(adev) && !amdgpu_virt_ras_telemetry_block_en(adev, head->block))
+		return 0;
+
 	get_obj(obj);
 
-	memcpy(obj->fs_data.sysfs_name,
-			head->sysfs_name,
-			sizeof(obj->fs_data.sysfs_name));
+	snprintf(obj->fs_data.sysfs_name, sizeof(obj->fs_data.sysfs_name),
+		"%s_err_count", head->name);
 
 	obj->sysfs_attr = (struct device_attribute){
 		.attr = {
@@ -1344,10 +2107,14 @@ int amdgpu_ras_sysfs_remove(struct amdgpu_device *adev,
 {
 	struct ras_manager *obj = amdgpu_ras_find_obj(adev, head);
 
+	if (amdgpu_aca_is_enabled(adev))
+		return 0;
+
 	if (!obj || !obj->attr_inuse)
 		return -EINVAL;
 
-	sysfs_remove_file_from_group(&adev->dev->kobj,
+	if (adev->dev->kobj.sd)
+		sysfs_remove_file_from_group(&adev->dev->kobj,
 				&obj->sysfs_attr.attr,
 				RAS_FS_NAME);
 	obj->attr_inuse = 0;
@@ -1368,7 +2135,7 @@ static int amdgpu_ras_sysfs_remove_all(struct amdgpu_device *adev)
 	if (amdgpu_bad_page_threshold != 0)
 		amdgpu_ras_sysfs_remove_bad_page_node(adev);
 
-	amdgpu_ras_sysfs_remove_feature_node(adev);
+	amdgpu_ras_sysfs_remove_dev_attr_node(adev);
 
 	return 0;
 }
@@ -1396,6 +2163,7 @@ static int amdgpu_ras_sysfs_remove_all(struct amdgpu_device *adev)
 static struct dentry *amdgpu_ras_debugfs_create_ctrl_node(struct amdgpu_device *adev)
 {
 	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+	struct amdgpu_ras_eeprom_control *eeprom = &con->eeprom_control;
 	struct drm_minor  *minor = adev_to_drm(adev)->primary;
 	struct dentry     *dir;
 
@@ -1406,6 +2174,7 @@ static struct dentry *amdgpu_ras_debugfs_create_ctrl_node(struct amdgpu_device *
 			    &amdgpu_ras_debugfs_eeprom_ops);
 	debugfs_create_u32("bad_page_cnt_threshold", 0444, dir,
 			   &con->bad_page_cnt_threshold);
+	debugfs_create_u32("ras_num_recs", 0444, dir, &eeprom->ras_num_recs);
 	debugfs_create_x32("ras_hw_enabled", 0444, dir, &adev->ras_hw_enabled);
 	debugfs_create_x32("ras_enabled", 0444, dir, &adev->ras_enabled);
 	debugfs_create_file("ras_eeprom_size", S_IRUGO, dir, adev,
@@ -1453,6 +2222,24 @@ static void amdgpu_ras_debugfs_create(struct amdgpu_device *adev,
 			    obj, &amdgpu_ras_debugfs_ops);
 }
 
+static bool amdgpu_ras_aca_is_supported(struct amdgpu_device *adev)
+{
+	bool ret;
+
+	switch (amdgpu_ip_version(adev, MP0_HWIP, 0)) {
+	case IP_VERSION(13, 0, 6):
+	case IP_VERSION(13, 0, 12):
+	case IP_VERSION(13, 0, 14):
+		ret = true;
+		break;
+	default:
+		ret = false;
+		break;
+	}
+
+	return ret;
+}
+
 void amdgpu_ras_debugfs_create_all(struct amdgpu_device *adev)
 {
 	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
@@ -1478,15 +2265,28 @@ void amdgpu_ras_debugfs_create_all(struct amdgpu_device *adev)
 			amdgpu_ras_debugfs_create(adev, &fs_info, dir);
 		}
 	}
+
+	if (amdgpu_ras_aca_is_supported(adev)) {
+		if (amdgpu_aca_is_enabled(adev))
+			amdgpu_aca_smu_debugfs_init(adev, dir);
+		else
+			amdgpu_mca_smu_debugfs_init(adev, dir);
+	}
 }
 
 /* debugfs end */
 
 /* ras fs */
-static BIN_ATTR(gpu_vram_bad_pages, S_IRUGO,
-		amdgpu_ras_sysfs_badpages_read, NULL, 0);
+static const BIN_ATTR(gpu_vram_bad_pages, S_IRUGO,
+		      amdgpu_ras_sysfs_badpages_read, NULL, 0);
 static DEVICE_ATTR(features, S_IRUGO,
 		amdgpu_ras_sysfs_features_read, NULL);
+static DEVICE_ATTR(version, 0444,
+		amdgpu_ras_sysfs_version_show, NULL);
+static DEVICE_ATTR(schema, 0444,
+		amdgpu_ras_sysfs_schema_show, NULL);
+static DEVICE_ATTR(event_state, 0444,
+		   amdgpu_ras_sysfs_event_state_show, NULL);
 static int amdgpu_ras_fs_init(struct amdgpu_device *adev)
 {
 	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
@@ -1495,26 +2295,41 @@ static int amdgpu_ras_fs_init(struct amdgpu_device *adev)
 	};
 	struct attribute *attrs[] = {
 		&con->features_attr.attr,
+		&con->version_attr.attr,
+		&con->schema_attr.attr,
+		&con->event_state_attr.attr,
 		NULL
 	};
-	struct bin_attribute *bin_attrs[] = {
+	const struct bin_attribute *bin_attrs[] = {
 		NULL,
 		NULL,
 	};
 	int r;
 
+	group.attrs = attrs;
+
 	/* add features entry */
 	con->features_attr = dev_attr_features;
-	group.attrs = attrs;
 	sysfs_attr_init(attrs[0]);
 
+	/* add version entry */
+	con->version_attr = dev_attr_version;
+	sysfs_attr_init(attrs[1]);
+
+	/* add schema entry */
+	con->schema_attr = dev_attr_schema;
+	sysfs_attr_init(attrs[2]);
+
+	/* add event_state entry */
+	con->event_state_attr = dev_attr_event_state;
+	sysfs_attr_init(attrs[3]);
+
 	if (amdgpu_bad_page_threshold != 0) {
 		/* add bad_page_features entry */
-		bin_attr_gpu_vram_bad_pages.private = NULL;
 		con->badpages_attr = bin_attr_gpu_vram_bad_pages;
+		sysfs_bin_attr_init(&con->badpages_attr);
 		bin_attrs[0] = &con->badpages_attr;
 		group.bin_attrs = bin_attrs;
-		sysfs_bin_attr_init(bin_attrs[0]);
 	}
 
 	r = sysfs_create_group(&adev->dev->kobj, &group);
@@ -1543,12 +2358,162 @@ static int amdgpu_ras_fs_fini(struct amdgpu_device *adev)
 /* ras fs end */
 
 /* ih begin */
+
+/* For the hardware that cannot enable bif ring for both ras_controller_irq
+ * and ras_err_evnet_athub_irq ih cookies, the driver has to poll status
+ * register to check whether the interrupt is triggered or not, and properly
+ * ack the interrupt if it is there
+ */
+void amdgpu_ras_interrupt_fatal_error_handler(struct amdgpu_device *adev)
+{
+	/* Fatal error events are handled on host side */
+	if (amdgpu_sriov_vf(adev))
+		return;
+	/*
+	 * If the current interrupt is caused by a non-fatal RAS error, skip
+	 * check for fatal error. For fatal errors, FED status of all devices
+	 * in XGMI hive gets set when the first device gets fatal error
+	 * interrupt. The error gets propagated to other devices as well, so
+	 * make sure to ack the interrupt regardless of FED status.
+	 */
+	if (!amdgpu_ras_get_fed_status(adev) &&
+	    amdgpu_ras_is_err_state(adev, AMDGPU_RAS_BLOCK__ANY))
+		return;
+
+	if (amdgpu_uniras_enabled(adev)) {
+		amdgpu_ras_mgr_handle_fatal_interrupt(adev, NULL);
+		return;
+	}
+
+	if (adev->nbio.ras &&
+	    adev->nbio.ras->handle_ras_controller_intr_no_bifring)
+		adev->nbio.ras->handle_ras_controller_intr_no_bifring(adev);
+
+	if (adev->nbio.ras &&
+	    adev->nbio.ras->handle_ras_err_event_athub_intr_no_bifring)
+		adev->nbio.ras->handle_ras_err_event_athub_intr_no_bifring(adev);
+}
+
+static void amdgpu_ras_interrupt_poison_consumption_handler(struct ras_manager *obj,
+				struct amdgpu_iv_entry *entry)
+{
+	bool poison_stat = false;
+	struct amdgpu_device *adev = obj->adev;
+	struct amdgpu_ras_block_object *block_obj =
+		amdgpu_ras_get_ras_block(adev, obj->head.block, 0);
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+	enum ras_event_type type = RAS_EVENT_TYPE_POISON_CONSUMPTION;
+	u64 event_id;
+	int ret;
+
+	if (!block_obj || !con)
+		return;
+
+	ret = amdgpu_ras_mark_ras_event(adev, type);
+	if (ret)
+		return;
+
+	amdgpu_ras_set_err_poison(adev, block_obj->ras_comm.block);
+	/* both query_poison_status and handle_poison_consumption are optional,
+	 * but at least one of them should be implemented if we need poison
+	 * consumption handler
+	 */
+	if (block_obj->hw_ops && block_obj->hw_ops->query_poison_status) {
+		poison_stat = block_obj->hw_ops->query_poison_status(adev);
+		if (!poison_stat) {
+			/* Not poison consumption interrupt, no need to handle it */
+			dev_info(adev->dev, "No RAS poison status in %s poison IH.\n",
+					block_obj->ras_comm.name);
+
+			return;
+		}
+	}
+
+	amdgpu_umc_poison_handler(adev, obj->head.block, 0);
+
+	if (block_obj->hw_ops && block_obj->hw_ops->handle_poison_consumption)
+		poison_stat = block_obj->hw_ops->handle_poison_consumption(adev);
+
+	/* gpu reset is fallback for failed and default cases.
+	 * For RMA case, amdgpu_umc_poison_handler will handle gpu reset.
+	 */
+	if (poison_stat && !amdgpu_ras_is_rma(adev)) {
+		event_id = amdgpu_ras_acquire_event_id(adev, type);
+		RAS_EVENT_LOG(adev, event_id,
+			      "GPU reset for %s RAS poison consumption is issued!\n",
+			      block_obj->ras_comm.name);
+		amdgpu_ras_reset_gpu(adev);
+	}
+
+	if (!poison_stat)
+		amdgpu_gfx_poison_consumption_handler(adev, entry);
+}
+
+static void amdgpu_ras_interrupt_poison_creation_handler(struct ras_manager *obj,
+				struct amdgpu_iv_entry *entry)
+{
+	struct amdgpu_device *adev = obj->adev;
+	enum ras_event_type type = RAS_EVENT_TYPE_POISON_CREATION;
+	u64 event_id;
+	int ret;
+
+	ret = amdgpu_ras_mark_ras_event(adev, type);
+	if (ret)
+		return;
+
+	event_id = amdgpu_ras_acquire_event_id(adev, type);
+	RAS_EVENT_LOG(adev, event_id, "Poison is created\n");
+
+	if (amdgpu_ip_version(obj->adev, UMC_HWIP, 0) >= IP_VERSION(12, 0, 0)) {
+		struct amdgpu_ras *con = amdgpu_ras_get_context(obj->adev);
+
+		atomic_inc(&con->page_retirement_req_cnt);
+		atomic_inc(&con->poison_creation_count);
+
+		wake_up(&con->page_retirement_wq);
+	}
+}
+
+static void amdgpu_ras_interrupt_umc_handler(struct ras_manager *obj,
+				struct amdgpu_iv_entry *entry)
+{
+	struct ras_ih_data *data = &obj->ih_data;
+	struct ras_err_data err_data;
+	int ret;
+
+	if (!data->cb)
+		return;
+
+	ret = amdgpu_ras_error_data_init(&err_data);
+	if (ret)
+		return;
+
+	/* Let IP handle its data, maybe we need get the output
+	 * from the callback to update the error type/count, etc
+	 */
+	amdgpu_ras_set_fed(obj->adev, true);
+	ret = data->cb(obj->adev, &err_data, entry);
+	/* ue will trigger an interrupt, and in that case
+	 * we need do a reset to recovery the whole system.
+	 * But leave IP do that recovery, here we just dispatch
+	 * the error.
+	 */
+	if (ret == AMDGPU_RAS_SUCCESS) {
+		/* these counts could be left as 0 if
+		 * some blocks do not count error number
+		 */
+		obj->err_data.ue_count += err_data.ue_count;
+		obj->err_data.ce_count += err_data.ce_count;
+		obj->err_data.de_count += err_data.de_count;
+	}
+
+	amdgpu_ras_error_data_fini(&err_data);
+}
+
 static void amdgpu_ras_interrupt_handler(struct ras_manager *obj)
 {
 	struct ras_ih_data *data = &obj->ih_data;
 	struct amdgpu_iv_entry entry;
-	int ret;
-	struct ras_err_data err_data = {0, 0, 0, NULL};
 
 	while (data->rptr != data->wptr) {
 		rmb();
@@ -1559,29 +2524,17 @@ static void amdgpu_ras_interrupt_handler(struct ras_manager *obj)
 		data->rptr = (data->aligned_element_size +
 				data->rptr) % data->ring_size;
 
-		if (data->cb) {
-			if (amdgpu_ras_is_poison_mode_supported(obj->adev) &&
-			    obj->head.block == AMDGPU_RAS_BLOCK__UMC)
-				dev_info(obj->adev->dev,
-						"Poison is created, no user action is needed.\n");
-			else {
-				/* Let IP handle its data, maybe we need get the output
-				 * from the callback to udpate the error type/count, etc
-				 */
-				ret = data->cb(obj->adev, &err_data, &entry);
-				/* ue will trigger an interrupt, and in that case
-				 * we need do a reset to recovery the whole system.
-				 * But leave IP do that recovery, here we just dispatch
-				 * the error.
-				 */
-				if (ret == AMDGPU_RAS_SUCCESS) {
-					/* these counts could be left as 0 if
-					 * some blocks do not count error number
-					 */
-					obj->err_data.ue_count += err_data.ue_count;
-					obj->err_data.ce_count += err_data.ce_count;
-				}
-			}
+		if (amdgpu_ras_is_poison_mode_supported(obj->adev)) {
+			if (obj->head.block == AMDGPU_RAS_BLOCK__UMC)
+				amdgpu_ras_interrupt_poison_creation_handler(obj, &entry);
+			else
+				amdgpu_ras_interrupt_poison_consumption_handler(obj, &entry);
+		} else {
+			if (obj->head.block == AMDGPU_RAS_BLOCK__UMC)
+				amdgpu_ras_interrupt_umc_handler(obj, &entry);
+			else
+				dev_warn(obj->adev->dev,
+					"No RAS interrupt handler for non-UMC block with poison disabled.\n");
 		}
 	}
 }
@@ -1599,12 +2552,25 @@ static void amdgpu_ras_interrupt_process_handler(struct work_struct *work)
 int amdgpu_ras_interrupt_dispatch(struct amdgpu_device *adev,
 		struct ras_dispatch_if *info)
 {
-	struct ras_manager *obj = amdgpu_ras_find_obj(adev, &info->head);
-	struct ras_ih_data *data = &obj->ih_data;
+	struct ras_manager *obj;
+	struct ras_ih_data *data;
+
+	if (amdgpu_uniras_enabled(adev)) {
+		struct ras_ih_info ih_info;
 
+		memset(&ih_info, 0, sizeof(ih_info));
+		ih_info.block = info->head.block;
+		memcpy(&ih_info.iv_entry, info->entry, sizeof(struct amdgpu_iv_entry));
+
+		return amdgpu_ras_mgr_handle_controller_interrupt(adev, &ih_info);
+	}
+
+	obj = amdgpu_ras_find_obj(adev, &info->head);
 	if (!obj)
 		return -EINVAL;
 
+	data = &obj->ih_data;
+
 	if (data->inuse == 0)
 		return 0;
 
@@ -1622,9 +2588,9 @@ int amdgpu_ras_interrupt_dispatch(struct amdgpu_device *adev,
 }
 
 int amdgpu_ras_interrupt_remove_handler(struct amdgpu_device *adev,
-		struct ras_ih_if *info)
+		struct ras_common_if *head)
 {
-	struct ras_manager *obj = amdgpu_ras_find_obj(adev, &info->head);
+	struct ras_manager *obj = amdgpu_ras_find_obj(adev, head);
 	struct ras_ih_data *data;
 
 	if (!obj)
@@ -1644,24 +2610,27 @@ int amdgpu_ras_interrupt_remove_handler(struct amdgpu_device *adev,
 }
 
 int amdgpu_ras_interrupt_add_handler(struct amdgpu_device *adev,
-		struct ras_ih_if *info)
+		struct ras_common_if *head)
 {
-	struct ras_manager *obj = amdgpu_ras_find_obj(adev, &info->head);
+	struct ras_manager *obj = amdgpu_ras_find_obj(adev, head);
 	struct ras_ih_data *data;
+	struct amdgpu_ras_block_object *ras_obj;
 
 	if (!obj) {
 		/* in case we registe the IH before enable ras feature */
-		obj = amdgpu_ras_create_obj(adev, &info->head);
+		obj = amdgpu_ras_create_obj(adev, head);
 		if (!obj)
 			return -EINVAL;
 	} else
 		get_obj(obj);
 
+	ras_obj = container_of(head, struct amdgpu_ras_block_object, ras_comm);
+
 	data = &obj->ih_data;
 	/* add the callback.etc */
 	*data = (struct ras_ih_data) {
 		.inuse = 0,
-		.cb = info->cb,
+		.cb = ras_obj->ras_cb,
 		.element_size = sizeof(struct amdgpu_iv_entry),
 		.rptr = 0,
 		.wptr = 0,
@@ -1690,10 +2659,7 @@ static int amdgpu_ras_interrupt_remove_all(struct amdgpu_device *adev)
 	struct ras_manager *obj, *tmp;
 
 	list_for_each_entry_safe(obj, tmp, &con->head, node) {
-		struct ras_ih_if info = {
-			.head = obj->head,
-		};
-		amdgpu_ras_interrupt_remove_handler(adev, &info);
+		amdgpu_ras_interrupt_remove_handler(adev, &obj->head);
 	}
 
 	return 0;
@@ -1701,7 +2667,7 @@ static int amdgpu_ras_interrupt_remove_all(struct amdgpu_device *adev)
 /* ih end */
 
 /* traversal all IPs except NBIO to query error counter */
-static void amdgpu_ras_log_on_err_counter(struct amdgpu_device *adev)
+static void amdgpu_ras_log_on_err_counter(struct amdgpu_device *adev, enum ras_event_type type)
 {
 	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
 	struct ras_manager *obj;
@@ -1723,7 +2689,28 @@ static void amdgpu_ras_log_on_err_counter(struct amdgpu_device *adev)
 		if (info.head.block == AMDGPU_RAS_BLOCK__PCIE_BIF)
 			continue;
 
-		amdgpu_ras_query_error_status(adev, &info);
+		/*
+		 * this is a workaround for aldebaran, skip send msg to
+		 * smu to get ecc_info table due to smu handle get ecc
+		 * info table failed temporarily.
+		 * should be removed until smu fix handle ecc_info table.
+		 */
+		if ((info.head.block == AMDGPU_RAS_BLOCK__UMC) &&
+		    (amdgpu_ip_version(adev, MP1_HWIP, 0) ==
+		     IP_VERSION(13, 0, 2)))
+			continue;
+
+		amdgpu_ras_query_error_status_with_event(adev, &info, type);
+
+		if (amdgpu_ip_version(adev, MP0_HWIP, 0) !=
+			    IP_VERSION(11, 0, 2) &&
+		    amdgpu_ip_version(adev, MP0_HWIP, 0) !=
+			    IP_VERSION(11, 0, 4) &&
+		    amdgpu_ip_version(adev, MP0_HWIP, 0) !=
+			    IP_VERSION(13, 0, 0)) {
+			if (amdgpu_ras_reset_error_status(adev, info.head.block))
+				dev_warn(adev->dev, "Failed to reset error counter and error status");
+		}
 	}
 }
 
@@ -1731,24 +2718,28 @@ static void amdgpu_ras_log_on_err_counter(struct amdgpu_device *adev)
 static void amdgpu_ras_error_status_query(struct amdgpu_device *adev,
 					  struct ras_query_if *info)
 {
+	struct amdgpu_ras_block_object *block_obj;
 	/*
 	 * Only two block need to query read/write
 	 * RspStatus at current state
 	 */
-	switch (info->head.block) {
-	case AMDGPU_RAS_BLOCK__GFX:
-		if (adev->gfx.ras_funcs &&
-		    adev->gfx.ras_funcs->query_ras_error_status)
-			adev->gfx.ras_funcs->query_ras_error_status(adev);
-		break;
-	case AMDGPU_RAS_BLOCK__MMHUB:
-		if (adev->mmhub.ras_funcs &&
-		    adev->mmhub.ras_funcs->query_ras_error_status)
-			adev->mmhub.ras_funcs->query_ras_error_status(adev);
-		break;
-	default:
-		break;
+	if ((info->head.block != AMDGPU_RAS_BLOCK__GFX) &&
+		(info->head.block != AMDGPU_RAS_BLOCK__MMHUB))
+		return;
+
+	block_obj = amdgpu_ras_get_ras_block(adev,
+					info->head.block,
+					info->head.sub_block_index);
+
+	if (!block_obj || !block_obj->hw_ops) {
+		dev_dbg_once(adev->dev, "%s doesn't config RAS function\n",
+			     get_ras_block_str(&info->head));
+		return;
 	}
+
+	if (block_obj->hw_ops->query_ras_error_status)
+		block_obj->hw_ops->query_ras_error_status(adev);
+
 }
 
 static void amdgpu_ras_query_err_status(struct amdgpu_device *adev)
@@ -1768,55 +2759,121 @@ static void amdgpu_ras_query_err_status(struct amdgpu_device *adev)
 	}
 }
 
-/* recovery begin */
-
-/* return 0 on success.
- * caller need free bps.
- */
 static int amdgpu_ras_badpages_read(struct amdgpu_device *adev,
-		struct ras_badpage **bps, unsigned int *count)
+		struct ras_badpage *bps, uint32_t count, uint32_t start)
 {
 	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
 	struct ras_err_handler_data *data;
-	int i = 0;
-	int ret = 0, status;
+	int r = 0;
+	uint32_t i;
 
 	if (!con || !con->eh_data || !bps || !count)
 		return -EINVAL;
 
 	mutex_lock(&con->recovery_lock);
 	data = con->eh_data;
-	if (!data || data->count == 0) {
-		*bps = NULL;
-		ret = -EINVAL;
-		goto out;
+	if (start < data->count) {
+		for (i = start; i < data->count; i++) {
+			if (!data->bps[i].ts)
+				continue;
+
+			bps[r].bp = data->bps[i].retired_page;
+			r++;
+			if (r >= count)
+				break;
+		}
 	}
+	mutex_unlock(&con->recovery_lock);
 
-	*bps = kmalloc(sizeof(struct ras_badpage) * data->count, GFP_KERNEL);
-	if (!*bps) {
-		ret = -ENOMEM;
-		goto out;
-	}
+	return r;
+}
 
-	for (; i < data->count; i++) {
-		(*bps)[i] = (struct ras_badpage){
-			.bp = data->bps[i].retired_page,
-			.size = AMDGPU_GPU_PAGE_SIZE,
-			.flags = AMDGPU_RAS_RETIRE_PAGE_RESERVED,
-		};
-		status = amdgpu_vram_mgr_query_page_status(
-				ttm_manager_type(&adev->mman.bdev, TTM_PL_VRAM),
-				data->bps[i].retired_page);
-		if (status == -EBUSY)
-			(*bps)[i].flags = AMDGPU_RAS_RETIRE_PAGE_PENDING;
-		else if (status == -ENOENT)
-			(*bps)[i].flags = AMDGPU_RAS_RETIRE_PAGE_FAULT;
+static int amdgpu_uniras_badpages_read(struct amdgpu_device *adev,
+		struct ras_badpage *bps, uint32_t count, uint32_t start)
+{
+	struct ras_cmd_bad_pages_info_req cmd_input;
+	struct ras_cmd_bad_pages_info_rsp *output;
+	uint32_t group, start_group, end_group;
+	uint32_t pos, pos_in_group;
+	int r = 0, i;
+
+	if (!bps || !count)
+		return -EINVAL;
+
+	output = kmalloc(sizeof(*output), GFP_KERNEL);
+	if (!output)
+		return -ENOMEM;
+
+	memset(&cmd_input, 0, sizeof(cmd_input));
+
+	start_group = start / RAS_CMD_MAX_BAD_PAGES_PER_GROUP;
+	end_group = (start + count + RAS_CMD_MAX_BAD_PAGES_PER_GROUP - 1) /
+				RAS_CMD_MAX_BAD_PAGES_PER_GROUP;
+
+	pos = start;
+	for (group = start_group; group < end_group; group++) {
+		memset(output, 0, sizeof(*output));
+		cmd_input.group_index = group;
+		if (amdgpu_ras_mgr_handle_ras_cmd(adev, RAS_CMD__GET_BAD_PAGES,
+			&cmd_input, sizeof(cmd_input), output, sizeof(*output)))
+			goto out;
+
+		if (pos >= output->bp_total_cnt)
+			goto out;
+
+		pos_in_group = pos - group * RAS_CMD_MAX_BAD_PAGES_PER_GROUP;
+		for (i = pos_in_group; i < output->bp_in_group; i++, pos++) {
+			if (!output->records[i].ts)
+				continue;
+
+			bps[r].bp = output->records[i].retired_page;
+			r++;
+			if (r >= count)
+				goto out;
+		}
 	}
 
-	*count = data->count;
 out:
-	mutex_unlock(&con->recovery_lock);
-	return ret;
+	kfree(output);
+	return r;
+}
+
+static void amdgpu_ras_set_fed_all(struct amdgpu_device *adev,
+				   struct amdgpu_hive_info *hive, bool status)
+{
+	struct amdgpu_device *tmp_adev;
+
+	if (hive) {
+		list_for_each_entry(tmp_adev, &hive->device_list, gmc.xgmi.head)
+			amdgpu_ras_set_fed(tmp_adev, status);
+	} else {
+		amdgpu_ras_set_fed(adev, status);
+	}
+}
+
+bool amdgpu_ras_in_recovery(struct amdgpu_device *adev)
+{
+	struct amdgpu_hive_info *hive = amdgpu_get_xgmi_hive(adev);
+	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
+	int hive_ras_recovery = 0;
+
+	if (hive) {
+		hive_ras_recovery = atomic_read(&hive->ras_recovery);
+		amdgpu_put_xgmi_hive(hive);
+	}
+
+	if (ras && (atomic_read(&ras->in_recovery) || hive_ras_recovery))
+		return true;
+
+	return false;
+}
+
+static enum ras_event_type amdgpu_ras_get_fatal_error_event(struct amdgpu_device *adev)
+{
+	if (amdgpu_ras_intr_triggered())
+		return RAS_EVENT_TYPE_FATAL;
+	else
+		return RAS_EVENT_TYPE_POISON_CONSUMPTION;
 }
 
 static void amdgpu_ras_do_recovery(struct work_struct *work)
@@ -1826,9 +2883,26 @@ static void amdgpu_ras_do_recovery(struct work_struct *work)
 	struct amdgpu_device *remote_adev = NULL;
 	struct amdgpu_device *adev = ras->adev;
 	struct list_head device_list, *device_list_handle =  NULL;
+	struct amdgpu_hive_info *hive = amdgpu_get_xgmi_hive(adev);
+	unsigned int error_query_mode;
+	enum ras_event_type type;
+
+	if (hive) {
+		atomic_set(&hive->ras_recovery, 1);
 
+		/* If any device which is part of the hive received RAS fatal
+		 * error interrupt, set fatal error status on all. This
+		 * condition will need a recovery, and flag will be cleared
+		 * as part of recovery.
+		 */
+		list_for_each_entry(remote_adev, &hive->device_list,
+				    gmc.xgmi.head)
+			if (amdgpu_ras_get_fed_status(remote_adev)) {
+				amdgpu_ras_set_fed_all(adev, hive, true);
+				break;
+			}
+	}
 	if (!ras->disable_ras_err_cnt_harvest) {
-		struct amdgpu_hive_info *hive = amdgpu_get_xgmi_hive(adev);
 
 		/* Build list of devices to query RAS related errors */
 		if  (hive && adev->gmc.xgmi.num_physical_nodes > 1) {
@@ -1839,18 +2913,64 @@ static void amdgpu_ras_do_recovery(struct work_struct *work)
 			device_list_handle = &device_list;
 		}
 
+		if (amdgpu_ras_get_error_query_mode(adev, &error_query_mode)) {
+			if (error_query_mode == AMDGPU_RAS_FIRMWARE_ERROR_QUERY) {
+				/* wait 500ms to ensure pmfw polling mca bank info done */
+				msleep(500);
+			}
+		}
+
+		type = amdgpu_ras_get_fatal_error_event(adev);
 		list_for_each_entry(remote_adev,
 				device_list_handle, gmc.xgmi.head) {
-			amdgpu_ras_query_err_status(remote_adev);
-			amdgpu_ras_log_on_err_counter(remote_adev);
+			if (amdgpu_uniras_enabled(remote_adev)) {
+				amdgpu_ras_mgr_update_ras_ecc(remote_adev);
+			} else {
+				amdgpu_ras_query_err_status(remote_adev);
+				amdgpu_ras_log_on_err_counter(remote_adev, type);
+			}
 		}
 
-		amdgpu_put_xgmi_hive(hive);
 	}
 
-	if (amdgpu_device_should_recover_gpu(ras->adev))
-		amdgpu_device_gpu_recover(ras->adev, NULL);
+	if (amdgpu_device_should_recover_gpu(ras->adev)) {
+		struct amdgpu_reset_context reset_context;
+		memset(&reset_context, 0, sizeof(reset_context));
+
+		reset_context.method = AMD_RESET_METHOD_NONE;
+		reset_context.reset_req_dev = adev;
+		reset_context.src = AMDGPU_RESET_SRC_RAS;
+		set_bit(AMDGPU_SKIP_COREDUMP, &reset_context.flags);
+
+		/* Perform full reset in fatal error mode */
+		if (!amdgpu_ras_is_poison_mode_supported(ras->adev))
+			set_bit(AMDGPU_NEED_FULL_RESET, &reset_context.flags);
+		else {
+			clear_bit(AMDGPU_NEED_FULL_RESET, &reset_context.flags);
+
+			if (ras->gpu_reset_flags & AMDGPU_RAS_GPU_RESET_MODE2_RESET) {
+				ras->gpu_reset_flags &= ~AMDGPU_RAS_GPU_RESET_MODE2_RESET;
+				reset_context.method = AMD_RESET_METHOD_MODE2;
+			}
+
+			/* Fatal error occurs in poison mode, mode1 reset is used to
+			 * recover gpu.
+			 */
+			if (ras->gpu_reset_flags & AMDGPU_RAS_GPU_RESET_MODE1_RESET) {
+				ras->gpu_reset_flags &= ~AMDGPU_RAS_GPU_RESET_MODE1_RESET;
+				set_bit(AMDGPU_NEED_FULL_RESET, &reset_context.flags);
+
+				psp_fatal_error_recovery_quirk(&adev->psp);
+			}
+		}
+
+		amdgpu_device_gpu_recover(ras->adev, NULL, &reset_context);
+	}
 	atomic_set(&ras->in_recovery, 0);
+	if (hive) {
+		atomic_set(&hive->ras_recovery, 0);
+		amdgpu_put_xgmi_hive(hive);
+	}
 }
 
 /* alloc/realloc bps array */
@@ -1860,10 +2980,9 @@ static int amdgpu_ras_realloc_eh_data_space(struct amdgpu_device *adev,
 	unsigned int old_space = data->count + data->space_left;
 	unsigned int new_space = old_space + pages;
 	unsigned int align_space = ALIGN(new_space, 512);
-	void *bps = kmalloc(align_space * sizeof(*data->bps), GFP_KERNEL);
+	void *bps = kmalloc_array(align_space, sizeof(*data->bps), GFP_KERNEL);
 
 	if (!bps) {
-		kfree(bps);
 		return -ENOMEM;
 	}
 
@@ -1878,44 +2997,283 @@ static int amdgpu_ras_realloc_eh_data_space(struct amdgpu_device *adev,
 	return 0;
 }
 
-/* it deal with vram only. */
-int amdgpu_ras_add_bad_pages(struct amdgpu_device *adev,
-		struct eeprom_table_record *bps, int pages)
+static int amdgpu_ras_mca2pa_by_idx(struct amdgpu_device *adev,
+			struct eeprom_table_record *bps,
+			struct ras_err_data *err_data)
 {
-	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
-	struct ras_err_handler_data *data;
+	struct ta_ras_query_address_input addr_in;
+	uint32_t socket = 0;
 	int ret = 0;
-	uint32_t i;
 
-	if (!con || !con->eh_data || !bps || pages <= 0)
-		return 0;
+	if (adev->smuio.funcs && adev->smuio.funcs->get_socket_id)
+		socket = adev->smuio.funcs->get_socket_id(adev);
 
-	mutex_lock(&con->recovery_lock);
-	data = con->eh_data;
-	if (!data)
-		goto out;
+	/* reinit err_data */
+	err_data->err_addr_cnt = 0;
+	err_data->err_addr_len = adev->umc.retire_unit;
+
+	memset(&addr_in, 0, sizeof(addr_in));
+	addr_in.ma.err_addr = bps->address;
+	addr_in.ma.socket_id = socket;
+	addr_in.ma.ch_inst = bps->mem_channel;
+	if (!amdgpu_ras_smu_eeprom_supported(adev)) {
+		/* tell RAS TA the node instance is not used */
+		addr_in.ma.node_inst = TA_RAS_INV_NODE;
+	} else {
+		addr_in.ma.umc_inst = bps->mcumc_id;
+		addr_in.ma.node_inst = bps->cu;
+	}
+
+	if (adev->umc.ras && adev->umc.ras->convert_ras_err_addr)
+		ret = adev->umc.ras->convert_ras_err_addr(adev, err_data,
+				&addr_in, NULL, false);
+
+	return ret;
+}
+
+static int amdgpu_ras_mca2pa(struct amdgpu_device *adev,
+			struct eeprom_table_record *bps,
+			struct ras_err_data *err_data)
+{
+	struct ta_ras_query_address_input addr_in;
+	uint32_t die_id, socket = 0;
+
+	if (adev->smuio.funcs && adev->smuio.funcs->get_socket_id)
+		socket = adev->smuio.funcs->get_socket_id(adev);
 
-	for (i = 0; i < pages; i++) {
+	/* although die id is gotten from PA in nps1 mode, the id is
+	 * fitable for any nps mode
+	 */
+	if (adev->umc.ras && adev->umc.ras->get_die_id_from_pa)
+		die_id = adev->umc.ras->get_die_id_from_pa(adev, bps->address,
+					bps->retired_page << AMDGPU_GPU_PAGE_SHIFT);
+	else
+		return -EINVAL;
+
+	/* reinit err_data */
+	err_data->err_addr_cnt = 0;
+	err_data->err_addr_len = adev->umc.retire_unit;
+
+	memset(&addr_in, 0, sizeof(addr_in));
+	addr_in.ma.err_addr = bps->address;
+	addr_in.ma.ch_inst = bps->mem_channel;
+	addr_in.ma.umc_inst = bps->mcumc_id;
+	addr_in.ma.node_inst = die_id;
+	addr_in.ma.socket_id = socket;
+
+	if (adev->umc.ras && adev->umc.ras->convert_ras_err_addr)
+		return adev->umc.ras->convert_ras_err_addr(adev, err_data,
+					&addr_in, NULL, false);
+	else
+		return  -EINVAL;
+}
+
+static int __amdgpu_ras_restore_bad_pages(struct amdgpu_device *adev,
+					struct eeprom_table_record *bps, int count)
+{
+	int j;
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+	struct ras_err_handler_data *data = con->eh_data;
+
+	for (j = 0; j < count; j++) {
 		if (amdgpu_ras_check_bad_page_unlock(con,
-			bps[i].retired_page << AMDGPU_GPU_PAGE_SHIFT))
+			bps[j].retired_page << AMDGPU_GPU_PAGE_SHIFT)) {
+			data->count++;
+			data->space_left--;
 			continue;
+		}
 
 		if (!data->space_left &&
-			amdgpu_ras_realloc_eh_data_space(adev, data, 256)) {
-			ret = -ENOMEM;
-			goto out;
+		    amdgpu_ras_realloc_eh_data_space(adev, data, 256)) {
+			return -ENOMEM;
 		}
 
-		amdgpu_vram_mgr_reserve_range(
-			ttm_manager_type(&adev->mman.bdev, TTM_PL_VRAM),
-			bps[i].retired_page << AMDGPU_GPU_PAGE_SHIFT,
-			AMDGPU_GPU_PAGE_SIZE);
+		amdgpu_ras_reserve_page(adev, bps[j].retired_page);
 
-		memcpy(&data->bps[data->count], &bps[i], sizeof(*data->bps));
+		memcpy(&data->bps[data->count], &(bps[j]),
+				sizeof(struct eeprom_table_record));
 		data->count++;
 		data->space_left--;
+		con->bad_page_num++;
+	}
+
+	return 0;
+}
+
+static int __amdgpu_ras_convert_rec_array_from_rom(struct amdgpu_device *adev,
+				struct eeprom_table_record *bps, struct ras_err_data *err_data,
+				enum amdgpu_memory_partition nps)
+{
+	int i = 0;
+	enum amdgpu_memory_partition save_nps;
+
+	save_nps = (bps[0].retired_page >> UMC_NPS_SHIFT) & UMC_NPS_MASK;
+
+	/*old asics just have pa in eeprom*/
+	if (IP_VERSION_MAJ(amdgpu_ip_version(adev, UMC_HWIP, 0)) < 12) {
+		memcpy(err_data->err_addr, bps,
+			sizeof(struct eeprom_table_record) * adev->umc.retire_unit);
+		goto out;
+	}
+
+	for (i = 0; i < adev->umc.retire_unit; i++)
+		bps[i].retired_page &= ~(UMC_NPS_MASK << UMC_NPS_SHIFT);
+
+	if (save_nps) {
+		if (save_nps == nps) {
+			if (amdgpu_umc_pages_in_a_row(adev, err_data,
+					bps[0].retired_page << AMDGPU_GPU_PAGE_SHIFT))
+				return -EINVAL;
+			for (i = 0; i < adev->umc.retire_unit; i++) {
+				err_data->err_addr[i].address = bps[0].address;
+				err_data->err_addr[i].mem_channel = bps[0].mem_channel;
+				err_data->err_addr[i].bank = bps[0].bank;
+				err_data->err_addr[i].err_type = bps[0].err_type;
+				err_data->err_addr[i].mcumc_id = bps[0].mcumc_id;
+			}
+		} else {
+			if (amdgpu_ras_mca2pa_by_idx(adev, &bps[0], err_data))
+				return -EINVAL;
+		}
+	} else {
+		if (bps[0].address == 0) {
+			/* for specific old eeprom data, mca address is not stored,
+			 * calc it from pa
+			 */
+			if (amdgpu_umc_pa2mca(adev, bps[0].retired_page << AMDGPU_GPU_PAGE_SHIFT,
+				&(bps[0].address), AMDGPU_NPS1_PARTITION_MODE))
+				return -EINVAL;
+		}
+
+		if (amdgpu_ras_mca2pa(adev, &bps[0], err_data)) {
+			if (nps == AMDGPU_NPS1_PARTITION_MODE)
+				memcpy(err_data->err_addr, bps,
+					sizeof(struct eeprom_table_record) * adev->umc.retire_unit);
+			else
+				return -EOPNOTSUPP;
+		}
 	}
+
 out:
+	return __amdgpu_ras_restore_bad_pages(adev, err_data->err_addr, adev->umc.retire_unit);
+}
+
+static int __amdgpu_ras_convert_rec_from_rom(struct amdgpu_device *adev,
+				struct eeprom_table_record *bps, struct ras_err_data *err_data,
+				enum amdgpu_memory_partition nps)
+{
+	int i = 0;
+	enum amdgpu_memory_partition save_nps;
+
+	if (!amdgpu_ras_smu_eeprom_supported(adev)) {
+		save_nps = (bps->retired_page >> UMC_NPS_SHIFT) & UMC_NPS_MASK;
+		bps->retired_page &= ~(UMC_NPS_MASK << UMC_NPS_SHIFT);
+	} else {
+		/* if pmfw manages eeprom, save_nps is not stored on eeprom,
+		 * we should always convert mca address into physical address,
+		 * make save_nps different from nps
+		 */
+		save_nps = nps + 1;
+	}
+
+	if (save_nps == nps) {
+		if (amdgpu_umc_pages_in_a_row(adev, err_data,
+				bps->retired_page << AMDGPU_GPU_PAGE_SHIFT))
+			return -EINVAL;
+		for (i = 0; i < adev->umc.retire_unit; i++) {
+			err_data->err_addr[i].address = bps->address;
+			err_data->err_addr[i].mem_channel = bps->mem_channel;
+			err_data->err_addr[i].bank = bps->bank;
+			err_data->err_addr[i].err_type = bps->err_type;
+			err_data->err_addr[i].mcumc_id = bps->mcumc_id;
+		}
+	} else {
+		if (bps->address) {
+			if (amdgpu_ras_mca2pa_by_idx(adev, bps, err_data))
+				return -EINVAL;
+		} else {
+			/* for specific old eeprom data, mca address is not stored,
+			 * calc it from pa
+			 */
+			if (amdgpu_umc_pa2mca(adev, bps->retired_page << AMDGPU_GPU_PAGE_SHIFT,
+				&(bps->address), AMDGPU_NPS1_PARTITION_MODE))
+				return -EINVAL;
+
+			if (amdgpu_ras_mca2pa(adev, bps, err_data))
+				return -EOPNOTSUPP;
+		}
+	}
+
+	return __amdgpu_ras_restore_bad_pages(adev, err_data->err_addr,
+									adev->umc.retire_unit);
+}
+
+/* it deal with vram only. */
+int amdgpu_ras_add_bad_pages(struct amdgpu_device *adev,
+		struct eeprom_table_record *bps, int pages, bool from_rom)
+{
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+	struct ras_err_data err_data;
+	struct amdgpu_ras_eeprom_control *control =
+			&adev->psp.ras_context.ras->eeprom_control;
+	enum amdgpu_memory_partition nps = AMDGPU_NPS1_PARTITION_MODE;
+	int ret = 0;
+	uint32_t i = 0;
+
+	if (!con || !con->eh_data || !bps || pages <= 0)
+		return 0;
+
+	if (from_rom) {
+		err_data.err_addr =
+			kcalloc(adev->umc.retire_unit,
+				sizeof(struct eeprom_table_record), GFP_KERNEL);
+		if (!err_data.err_addr) {
+			dev_warn(adev->dev, "Failed to alloc UMC error address record in mca2pa conversion!\n");
+			return -ENOMEM;
+		}
+
+		if (adev->gmc.gmc_funcs->query_mem_partition_mode)
+			nps = adev->gmc.gmc_funcs->query_mem_partition_mode(adev);
+	}
+
+	mutex_lock(&con->recovery_lock);
+
+	if (from_rom) {
+		/* there is no pa recs in V3, so skip pa recs processing */
+		if ((control->tbl_hdr.version < RAS_TABLE_VER_V3) &&
+		    !amdgpu_ras_smu_eeprom_supported(adev)) {
+			for (i = 0; i < pages; i++) {
+				if (control->ras_num_recs - i >= adev->umc.retire_unit) {
+					if ((bps[i].address == bps[i + 1].address) &&
+						(bps[i].mem_channel == bps[i + 1].mem_channel)) {
+						/* deal with retire_unit records a time */
+						ret = __amdgpu_ras_convert_rec_array_from_rom(adev,
+										&bps[i], &err_data, nps);
+						if (ret)
+							con->bad_page_num -= adev->umc.retire_unit;
+						i += (adev->umc.retire_unit - 1);
+					} else {
+						break;
+					}
+				} else {
+					break;
+				}
+			}
+		}
+		for (; i < pages; i++) {
+			ret = __amdgpu_ras_convert_rec_from_rom(adev,
+				&bps[i], &err_data, nps);
+			if (ret)
+				con->bad_page_num -= adev->umc.retire_unit;
+		}
+
+		con->eh_data->count_saved = con->eh_data->count;
+	} else {
+		ret = __amdgpu_ras_restore_bad_pages(adev, bps, pages);
+	}
+
+	if (from_rom)
+		kfree(err_data.err_addr);
 	mutex_unlock(&con->recovery_lock);
 
 	return ret;
@@ -1924,30 +3282,70 @@ out:
 /*
  * write error record array to eeprom, the function should be
  * protected by recovery_lock
+ * new_cnt: new added UE count, excluding reserved bad pages, can be NULL
  */
-int amdgpu_ras_save_bad_pages(struct amdgpu_device *adev)
+int amdgpu_ras_save_bad_pages(struct amdgpu_device *adev,
+		unsigned long *new_cnt)
 {
 	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
 	struct ras_err_handler_data *data;
 	struct amdgpu_ras_eeprom_control *control;
-	int save_count;
+	int save_count, unit_num, i;
+
+	if (!con || !con->eh_data) {
+		if (new_cnt)
+			*new_cnt = 0;
+
+		return 0;
+	}
+
+	if (!con->eeprom_control.is_eeprom_valid) {
+		dev_warn(adev->dev,
+			"Failed to save EEPROM table data because of EEPROM data corruption!");
+		if (new_cnt)
+			*new_cnt = 0;
 
-	if (!con || !con->eh_data)
 		return 0;
+	}
 
+	mutex_lock(&con->recovery_lock);
 	control = &con->eeprom_control;
 	data = con->eh_data;
-	save_count = data->count - control->ras_num_recs;
+	if (amdgpu_ras_smu_eeprom_supported(adev))
+		unit_num = control->ras_num_recs -
+			control->ras_num_recs_old;
+	else
+		unit_num = data->count / adev->umc.retire_unit -
+			control->ras_num_recs;
+
+	save_count = con->bad_page_num - control->ras_num_bad_pages;
+	mutex_unlock(&con->recovery_lock);
+
+	if (new_cnt)
+		*new_cnt = unit_num;
+
 	/* only new entries are saved */
-	if (save_count > 0) {
-		if (amdgpu_ras_eeprom_append(control,
-					     &data->bps[control->ras_num_recs],
-					     save_count)) {
-			dev_err(adev->dev, "Failed to save EEPROM table data!");
-			return -EIO;
+	if (unit_num && save_count) {
+		/*old asics only save pa to eeprom like before*/
+		if (IP_VERSION_MAJ(amdgpu_ip_version(adev, UMC_HWIP, 0)) < 12) {
+			if (amdgpu_ras_eeprom_append(control,
+					&data->bps[data->count_saved], unit_num)) {
+				dev_err(adev->dev, "Failed to save EEPROM table data!");
+				return -EIO;
+			}
+		} else {
+			for (i = 0; i < unit_num; i++) {
+				if (amdgpu_ras_eeprom_append(control,
+						&data->bps[data->count_saved +
+						i * adev->umc.retire_unit], 1)) {
+					dev_err(adev->dev, "Failed to save EEPROM table data!");
+					return -EIO;
+				}
+			}
 		}
 
 		dev_info(adev->dev, "Saved %d pages to EEPROM table.\n", save_count);
+		data->count_saved = data->count;
 	}
 
 	return 0;
@@ -1962,7 +3360,7 @@ static int amdgpu_ras_load_bad_pages(struct amdgpu_device *adev)
 	struct amdgpu_ras_eeprom_control *control =
 		&adev->psp.ras_context.ras->eeprom_control;
 	struct eeprom_table_record *bps;
-	int ret;
+	int ret, i = 0;
 
 	/* no bad page record, skip eeprom access */
 	if (control->ras_num_recs == 0 || amdgpu_bad_page_threshold == 0)
@@ -1973,27 +3371,72 @@ static int amdgpu_ras_load_bad_pages(struct amdgpu_device *adev)
 		return -ENOMEM;
 
 	ret = amdgpu_ras_eeprom_read(control, bps, control->ras_num_recs);
-	if (ret)
+	if (ret) {
 		dev_err(adev->dev, "Failed to load EEPROM table records!");
-	else
-		ret = amdgpu_ras_add_bad_pages(adev, bps, control->ras_num_recs);
+	} else {
+		if (adev->umc.ras && adev->umc.ras->convert_ras_err_addr) {
+			/*In V3, there is no pa recs, and some cases(when address==0) may be parsed
+			as pa recs, so add verion check to avoid it.
+			*/
+			if ((control->tbl_hdr.version < RAS_TABLE_VER_V3) &&
+			    !amdgpu_ras_smu_eeprom_supported(adev)) {
+				for (i = 0; i < control->ras_num_recs; i++) {
+					if ((control->ras_num_recs - i) >= adev->umc.retire_unit) {
+						if ((bps[i].address == bps[i + 1].address) &&
+							(bps[i].mem_channel == bps[i + 1].mem_channel)) {
+							control->ras_num_pa_recs += adev->umc.retire_unit;
+							i += (adev->umc.retire_unit - 1);
+						} else {
+							control->ras_num_mca_recs +=
+										(control->ras_num_recs - i);
+							break;
+						}
+					} else {
+						control->ras_num_mca_recs += (control->ras_num_recs - i);
+						break;
+					}
+				}
+			} else {
+				control->ras_num_mca_recs = control->ras_num_recs;
+			}
+		}
+
+		ret = amdgpu_ras_add_bad_pages(adev, bps, control->ras_num_recs, true);
+		if (ret)
+			goto out;
+
+		ret = amdgpu_ras_eeprom_check(control);
+		if (ret)
+			goto out;
 
+		/* HW not usable */
+		if (amdgpu_ras_is_rma(adev))
+			ret = -EHWPOISON;
+	}
+
+out:
 	kfree(bps);
 	return ret;
 }
 
-static bool amdgpu_ras_check_bad_page_unlock(struct amdgpu_ras *con,
+static int amdgpu_ras_check_bad_page_unlock(struct amdgpu_ras *con,
 				uint64_t addr)
 {
 	struct ras_err_handler_data *data = con->eh_data;
+	struct amdgpu_device *adev = con->adev;
 	int i;
 
+	if ((addr >= adev->gmc.mc_vram_size &&
+	    adev->gmc.mc_vram_size) ||
+	    (addr >= RAS_UMC_INJECT_ADDR_LIMIT))
+		return -EINVAL;
+
 	addr >>= AMDGPU_GPU_PAGE_SHIFT;
 	for (i = 0; i < data->count; i++)
 		if (addr == data->bps[i].retired_page)
-			return true;
+			return 1;
 
-	return false;
+	return 0;
 }
 
 /*
@@ -2001,11 +3444,11 @@ static bool amdgpu_ras_check_bad_page_unlock(struct amdgpu_ras *con,
  *
  * Note: this check is only for umc block
  */
-static bool amdgpu_ras_check_bad_page(struct amdgpu_device *adev,
+static int amdgpu_ras_check_bad_page(struct amdgpu_device *adev,
 				uint64_t addr)
 {
 	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
-	bool ret = false;
+	int ret = 0;
 
 	if (!con || !con->eh_data)
 		return ret;
@@ -2022,45 +3465,382 @@ static void amdgpu_ras_validate_threshold(struct amdgpu_device *adev,
 	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
 
 	/*
-	 * Justification of value bad_page_cnt_threshold in ras structure
-	 *
-	 * Generally, -1 <= amdgpu_bad_page_threshold <= max record length
-	 * in eeprom, and introduce two scenarios accordingly.
-	 *
-	 * Bad page retirement enablement:
-	 *    - If amdgpu_bad_page_threshold = -1,
-	 *      bad_page_cnt_threshold = typical value by formula.
-	 *
-	 *    - When the value from user is 0 < amdgpu_bad_page_threshold <
-	 *      max record length in eeprom, use it directly.
-	 *
-	 * Bad page retirement disablement:
-	 *    - If amdgpu_bad_page_threshold = 0, bad page retirement
-	 *      functionality is disabled, and bad_page_cnt_threshold will
-	 *      take no effect.
+	 * amdgpu_bad_page_threshold is used to config
+	 * the threshold for the number of bad pages.
+	 * -1:  Threshold is set to default value
+	 *      Driver will issue a warning message when threshold is reached
+	 *      and continue runtime services.
+	 * 0:   Disable bad page retirement
+	 *      Driver will not retire bad pages
+	 *      which is intended for debugging purpose.
+	 * -2:  Threshold is determined by a formula
+	 *      that assumes 1 bad page per 100M of local memory.
+	 *      Driver will continue runtime services when threhold is reached.
+	 * 0 < threshold < max number of bad page records in EEPROM,
+	 *      A user-defined threshold is set
+	 *      Driver will halt runtime services when this custom threshold is reached.
 	 */
-
-	if (amdgpu_bad_page_threshold < 0) {
+	if (amdgpu_bad_page_threshold == -2) {
 		u64 val = adev->gmc.mc_vram_size;
 
 		do_div(val, RAS_BAD_PAGE_COVER);
 		con->bad_page_cnt_threshold = min(lower_32_bits(val),
 						  max_count);
+	} else if (amdgpu_bad_page_threshold == -1) {
+		con->bad_page_cnt_threshold = ((con->reserved_pages_in_bytes) >> 21) << 4;
 	} else {
 		con->bad_page_cnt_threshold = min_t(int, max_count,
 						    amdgpu_bad_page_threshold);
 	}
 }
 
-int amdgpu_ras_recovery_init(struct amdgpu_device *adev)
+int amdgpu_ras_put_poison_req(struct amdgpu_device *adev,
+		enum amdgpu_ras_block block, uint16_t pasid,
+		pasid_notify pasid_fn, void *data, uint32_t reset)
+{
+	int ret = 0;
+	struct ras_poison_msg poison_msg;
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+
+	memset(&poison_msg, 0, sizeof(poison_msg));
+	poison_msg.block = block;
+	poison_msg.pasid = pasid;
+	poison_msg.reset = reset;
+	poison_msg.pasid_fn = pasid_fn;
+	poison_msg.data = data;
+
+	ret = kfifo_put(&con->poison_fifo, poison_msg);
+	if (!ret) {
+		dev_err(adev->dev, "Poison message fifo is full!\n");
+		return -ENOSPC;
+	}
+
+	return 0;
+}
+
+static int amdgpu_ras_get_poison_req(struct amdgpu_device *adev,
+		struct ras_poison_msg *poison_msg)
+{
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+
+	return kfifo_get(&con->poison_fifo, poison_msg);
+}
+
+static void amdgpu_ras_ecc_log_init(struct ras_ecc_log_info *ecc_log)
+{
+	mutex_init(&ecc_log->lock);
+
+	INIT_RADIX_TREE(&ecc_log->de_page_tree, GFP_KERNEL);
+	ecc_log->de_queried_count = 0;
+	ecc_log->consumption_q_count = 0;
+}
+
+static void amdgpu_ras_ecc_log_fini(struct ras_ecc_log_info *ecc_log)
+{
+	struct radix_tree_iter iter;
+	void __rcu **slot;
+	struct ras_ecc_err *ecc_err;
+
+	mutex_lock(&ecc_log->lock);
+	radix_tree_for_each_slot(slot, &ecc_log->de_page_tree, &iter, 0) {
+		ecc_err = radix_tree_deref_slot(slot);
+		kfree(ecc_err->err_pages.pfn);
+		kfree(ecc_err);
+		radix_tree_iter_delete(&ecc_log->de_page_tree, &iter, slot);
+	}
+	mutex_unlock(&ecc_log->lock);
+
+	mutex_destroy(&ecc_log->lock);
+	ecc_log->de_queried_count = 0;
+	ecc_log->consumption_q_count = 0;
+}
+
+static bool amdgpu_ras_schedule_retirement_dwork(struct amdgpu_ras *con,
+				uint32_t delayed_ms)
+{
+	int ret;
+
+	mutex_lock(&con->umc_ecc_log.lock);
+	ret = radix_tree_tagged(&con->umc_ecc_log.de_page_tree,
+			UMC_ECC_NEW_DETECTED_TAG);
+	mutex_unlock(&con->umc_ecc_log.lock);
+
+	if (ret)
+		schedule_delayed_work(&con->page_retirement_dwork,
+			msecs_to_jiffies(delayed_ms));
+
+	return ret ? true : false;
+}
+
+static void amdgpu_ras_do_page_retirement(struct work_struct *work)
+{
+	struct amdgpu_ras *con = container_of(work, struct amdgpu_ras,
+					      page_retirement_dwork.work);
+	struct amdgpu_device *adev = con->adev;
+	struct ras_err_data err_data;
+
+	/* If gpu reset is ongoing, delay retiring the bad pages */
+	if (amdgpu_in_reset(adev) || amdgpu_ras_in_recovery(adev)) {
+		amdgpu_ras_schedule_retirement_dwork(con,
+				AMDGPU_RAS_RETIRE_PAGE_INTERVAL * 3);
+		return;
+	}
+
+	amdgpu_ras_error_data_init(&err_data);
+
+	amdgpu_umc_handle_bad_pages(adev, &err_data);
+
+	amdgpu_ras_error_data_fini(&err_data);
+
+	amdgpu_ras_schedule_retirement_dwork(con,
+			AMDGPU_RAS_RETIRE_PAGE_INTERVAL);
+}
+
+static int amdgpu_ras_poison_creation_handler(struct amdgpu_device *adev,
+				uint32_t poison_creation_count)
+{
+	int ret = 0;
+	struct ras_ecc_log_info *ecc_log;
+	struct ras_query_if info;
+	u32 timeout = MAX_UMC_POISON_POLLING_TIME_ASYNC;
+	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
+	u64 de_queried_count;
+	u64 consumption_q_count;
+	enum ras_event_type type = RAS_EVENT_TYPE_POISON_CREATION;
+
+	memset(&info, 0, sizeof(info));
+	info.head.block = AMDGPU_RAS_BLOCK__UMC;
+
+	ecc_log = &ras->umc_ecc_log;
+	ecc_log->de_queried_count = 0;
+	ecc_log->consumption_q_count = 0;
+
+	do {
+		ret = amdgpu_ras_query_error_status_with_event(adev, &info, type);
+		if (ret)
+			return ret;
+
+		de_queried_count = ecc_log->de_queried_count;
+		consumption_q_count = ecc_log->consumption_q_count;
+
+		if (de_queried_count && consumption_q_count)
+			break;
+
+		msleep(100);
+	} while (--timeout);
+
+	if (de_queried_count)
+		schedule_delayed_work(&ras->page_retirement_dwork, 0);
+
+	if (amdgpu_ras_is_rma(adev) && atomic_cmpxchg(&ras->rma_in_recovery, 0, 1) == 0)
+		amdgpu_ras_reset_gpu(adev);
+
+	return 0;
+}
+
+static void amdgpu_ras_clear_poison_fifo(struct amdgpu_device *adev)
+{
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+	struct ras_poison_msg msg;
+	int ret;
+
+	do {
+		ret = kfifo_get(&con->poison_fifo, &msg);
+	} while (ret);
+}
+
+static int amdgpu_ras_poison_consumption_handler(struct amdgpu_device *adev,
+			uint32_t msg_count, uint32_t *gpu_reset)
+{
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+	uint32_t reset_flags = 0, reset = 0;
+	struct ras_poison_msg msg;
+	int ret, i;
+
+	kgd2kfd_set_sram_ecc_flag(adev->kfd.dev);
+
+	for (i = 0; i < msg_count; i++) {
+		ret = amdgpu_ras_get_poison_req(adev, &msg);
+		if (!ret)
+			continue;
+
+		if (msg.pasid_fn)
+			msg.pasid_fn(adev, msg.pasid, msg.data);
+
+		reset_flags |= msg.reset;
+	}
+
+	/*
+	 * Try to ensure poison creation handler is completed first
+	 * to set rma if bad page exceed threshold.
+	 */
+	flush_delayed_work(&con->page_retirement_dwork);
+
+	/* for RMA, amdgpu_ras_poison_creation_handler will trigger gpu reset */
+	if (reset_flags && !amdgpu_ras_is_rma(adev)) {
+		if (reset_flags & AMDGPU_RAS_GPU_RESET_MODE1_RESET)
+			reset = AMDGPU_RAS_GPU_RESET_MODE1_RESET;
+		else if (reset_flags & AMDGPU_RAS_GPU_RESET_MODE2_RESET)
+			reset = AMDGPU_RAS_GPU_RESET_MODE2_RESET;
+		else
+			reset = reset_flags;
+
+		con->gpu_reset_flags |= reset;
+		amdgpu_ras_reset_gpu(adev);
+
+		*gpu_reset = reset;
+
+		/* Wait for gpu recovery to complete */
+		flush_work(&con->recovery_work);
+	}
+
+	return 0;
+}
+
+static int amdgpu_ras_page_retirement_thread(void *param)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)param;
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+	uint32_t poison_creation_count, msg_count;
+	uint32_t gpu_reset;
+	int ret;
+
+	while (!kthread_should_stop()) {
+
+		wait_event_interruptible(con->page_retirement_wq,
+				kthread_should_stop() ||
+				atomic_read(&con->page_retirement_req_cnt));
+
+		if (kthread_should_stop())
+			break;
+
+		mutex_lock(&con->poison_lock);
+		gpu_reset = 0;
+
+		do {
+			poison_creation_count = atomic_read(&con->poison_creation_count);
+			ret = amdgpu_ras_poison_creation_handler(adev, poison_creation_count);
+			if (ret == -EIO)
+				break;
+
+			if (poison_creation_count) {
+				atomic_sub(poison_creation_count, &con->poison_creation_count);
+				atomic_sub(poison_creation_count, &con->page_retirement_req_cnt);
+			}
+		} while (atomic_read(&con->poison_creation_count) &&
+			!atomic_read(&con->poison_consumption_count));
+
+		if (ret != -EIO) {
+			msg_count = kfifo_len(&con->poison_fifo);
+			if (msg_count) {
+				ret = amdgpu_ras_poison_consumption_handler(adev,
+						msg_count, &gpu_reset);
+				if ((ret != -EIO) &&
+				    (gpu_reset != AMDGPU_RAS_GPU_RESET_MODE1_RESET))
+					atomic_sub(msg_count, &con->page_retirement_req_cnt);
+			}
+		}
+
+		if ((ret == -EIO) || (gpu_reset == AMDGPU_RAS_GPU_RESET_MODE1_RESET)) {
+			/* gpu mode-1 reset is ongoing or just completed ras mode-1 reset */
+			/* Clear poison creation request */
+			atomic_set(&con->poison_creation_count, 0);
+			atomic_set(&con->poison_consumption_count, 0);
+
+			/* Clear poison fifo */
+			amdgpu_ras_clear_poison_fifo(adev);
+
+			/* Clear all poison requests */
+			atomic_set(&con->page_retirement_req_cnt, 0);
+
+			if (ret == -EIO) {
+				/* Wait for mode-1 reset to complete */
+				down_read(&adev->reset_domain->sem);
+				up_read(&adev->reset_domain->sem);
+			}
+
+			/* Wake up work to save bad pages to eeprom */
+			schedule_delayed_work(&con->page_retirement_dwork, 0);
+		} else if (gpu_reset) {
+			/* gpu just completed mode-2 reset or other reset */
+			/* Clear poison consumption messages cached in fifo */
+			msg_count = kfifo_len(&con->poison_fifo);
+			if (msg_count) {
+				amdgpu_ras_clear_poison_fifo(adev);
+				atomic_sub(msg_count, &con->page_retirement_req_cnt);
+			}
+
+			atomic_set(&con->poison_consumption_count, 0);
+
+			/* Wake up work to save bad pages to eeprom */
+			schedule_delayed_work(&con->page_retirement_dwork, 0);
+		}
+		mutex_unlock(&con->poison_lock);
+	}
+
+	return 0;
+}
+
+int amdgpu_ras_init_badpage_info(struct amdgpu_device *adev)
+{
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+	struct amdgpu_ras_eeprom_control *control;
+	int ret;
+
+	if (!con || amdgpu_sriov_vf(adev))
+		return 0;
+
+	if (amdgpu_uniras_enabled(adev))
+		return 0;
+
+	control = &con->eeprom_control;
+	con->ras_smu_drv = amdgpu_dpm_get_ras_smu_driver(adev);
+
+	ret = amdgpu_ras_eeprom_init(control);
+	control->is_eeprom_valid = !ret;
+
+	if (!adev->umc.ras || !adev->umc.ras->convert_ras_err_addr)
+		control->ras_num_pa_recs = control->ras_num_recs;
+
+	if (adev->umc.ras &&
+	    adev->umc.ras->get_retire_flip_bits)
+		adev->umc.ras->get_retire_flip_bits(adev);
+
+	if (control->ras_num_recs && control->is_eeprom_valid) {
+		ret = amdgpu_ras_load_bad_pages(adev);
+		if (ret) {
+			control->is_eeprom_valid = false;
+			return 0;
+		}
+
+		amdgpu_dpm_send_hbm_bad_pages_num(
+			adev, control->ras_num_bad_pages);
+
+		if (con->update_channel_flag == true) {
+			amdgpu_dpm_send_hbm_bad_channel_flag(
+				adev, control->bad_channel_bitmap);
+			con->update_channel_flag = false;
+		}
+
+		/* The format action is only applied to new ASICs */
+		if (IP_VERSION_MAJ(amdgpu_ip_version(adev, UMC_HWIP, 0)) >= 12 &&
+		    control->tbl_hdr.version < RAS_TABLE_VER_V3)
+			if (!amdgpu_ras_eeprom_reset_table(control))
+				if (amdgpu_ras_save_bad_pages(adev, NULL))
+					dev_warn(adev->dev, "Failed to format RAS EEPROM data in V3 version!\n");
+	}
+
+	return 0;
+}
+
+int amdgpu_ras_recovery_init(struct amdgpu_device *adev, bool init_bp_info)
 {
 	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
 	struct ras_err_handler_data **data;
 	u32  max_eeprom_records_count = 0;
-	bool exc_err_limit = false;
 	int ret;
 
-	if (!con)
+	if (!con || amdgpu_sriov_vf(adev))
 		return 0;
 
 	/* Allow access to RAS EEPROM via debugfs, when the ASIC
@@ -2074,42 +3854,44 @@ int amdgpu_ras_recovery_init(struct amdgpu_device *adev)
 		return 0;
 
 	data = &con->eh_data;
-	*data = kmalloc(sizeof(**data), GFP_KERNEL | __GFP_ZERO);
+	*data = kzalloc(sizeof(**data), GFP_KERNEL);
 	if (!*data) {
 		ret = -ENOMEM;
 		goto out;
 	}
 
 	mutex_init(&con->recovery_lock);
+	mutex_init(&con->poison_lock);
 	INIT_WORK(&con->recovery_work, amdgpu_ras_do_recovery);
 	atomic_set(&con->in_recovery, 0);
+	atomic_set(&con->rma_in_recovery, 0);
+	con->eeprom_control.bad_channel_bitmap = 0;
 
-	max_eeprom_records_count = amdgpu_ras_eeprom_max_record_count();
+	max_eeprom_records_count = amdgpu_ras_eeprom_max_record_count(&con->eeprom_control);
 	amdgpu_ras_validate_threshold(adev, max_eeprom_records_count);
 
-	/* Todo: During test the SMU might fail to read the eeprom through I2C
-	 * when the GPU is pending on XGMI reset during probe time
-	 * (Mostly after second bus reset), skip it now
-	 */
-	if (adev->gmc.xgmi.pending_reset)
-		return 0;
-	ret = amdgpu_ras_eeprom_init(&con->eeprom_control, &exc_err_limit);
-	/*
-	 * This calling fails when exc_err_limit is true or
-	 * ret != 0.
-	 */
-	if (exc_err_limit || ret)
-		goto free;
-
-	if (con->eeprom_control.ras_num_recs) {
-		ret = amdgpu_ras_load_bad_pages(adev);
+	if (init_bp_info) {
+		ret = amdgpu_ras_init_badpage_info(adev);
 		if (ret)
 			goto free;
+	}
 
-		if (adev->smu.ppt_funcs && adev->smu.ppt_funcs->send_hbm_bad_pages_num)
-			adev->smu.ppt_funcs->send_hbm_bad_pages_num(&adev->smu, con->eeprom_control.ras_num_recs);
+	mutex_init(&con->page_rsv_lock);
+	INIT_KFIFO(con->poison_fifo);
+	mutex_init(&con->page_retirement_lock);
+	init_waitqueue_head(&con->page_retirement_wq);
+	atomic_set(&con->page_retirement_req_cnt, 0);
+	atomic_set(&con->poison_creation_count, 0);
+	atomic_set(&con->poison_consumption_count, 0);
+	con->page_retirement_thread =
+		kthread_run(amdgpu_ras_page_retirement_thread, adev, "umc_page_retirement");
+	if (IS_ERR(con->page_retirement_thread)) {
+		con->page_retirement_thread = NULL;
+		dev_warn(adev->dev, "Failed to create umc_page_retirement thread!!!\n");
 	}
 
+	INIT_DELAYED_WORK(&con->page_retirement_dwork, amdgpu_ras_do_page_retirement);
+	amdgpu_ras_ecc_log_init(&con->umc_ecc_log);
 #ifdef CONFIG_X86_MCE_AMD
 	if ((adev->asic_type == CHIP_ALDEBARAN) &&
 	    (adev->gmc.xgmi.connected_to_cpu))
@@ -2128,7 +3910,7 @@ out:
 	 * Except error threshold exceeding case, other failure cases in this
 	 * function would not fail amdgpu driver init.
 	 */
-	if (!exc_err_limit)
+	if (!amdgpu_ras_is_rma(adev))
 		ret = 0;
 	else
 		ret = -EINVAL;
@@ -2140,25 +3922,75 @@ static int amdgpu_ras_recovery_fini(struct amdgpu_device *adev)
 {
 	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
 	struct ras_err_handler_data *data = con->eh_data;
+	int max_flush_timeout = MAX_FLUSH_RETIRE_DWORK_TIMES;
+	bool ret;
 
 	/* recovery_init failed to init it, fini is useless */
 	if (!data)
 		return 0;
 
+	/* Save all cached bad pages to eeprom */
+	do {
+		flush_delayed_work(&con->page_retirement_dwork);
+		ret = amdgpu_ras_schedule_retirement_dwork(con, 0);
+	} while (ret && max_flush_timeout--);
+
+	if (con->page_retirement_thread)
+		kthread_stop(con->page_retirement_thread);
+
+	atomic_set(&con->page_retirement_req_cnt, 0);
+	atomic_set(&con->poison_creation_count, 0);
+
+	mutex_destroy(&con->page_rsv_lock);
+
 	cancel_work_sync(&con->recovery_work);
 
+	cancel_delayed_work_sync(&con->page_retirement_dwork);
+
+	amdgpu_ras_ecc_log_fini(&con->umc_ecc_log);
+
 	mutex_lock(&con->recovery_lock);
 	con->eh_data = NULL;
 	kfree(data->bps);
 	kfree(data);
 	mutex_unlock(&con->recovery_lock);
 
+	amdgpu_ras_critical_region_init(adev);
+#ifdef CONFIG_X86_MCE_AMD
+	amdgpu_unregister_bad_pages_mca_notifier(adev);
+#endif
 	return 0;
 }
 /* recovery end */
 
 static bool amdgpu_ras_asic_supported(struct amdgpu_device *adev)
 {
+	if (amdgpu_sriov_vf(adev)) {
+		switch (amdgpu_ip_version(adev, MP0_HWIP, 0)) {
+		case IP_VERSION(13, 0, 2):
+		case IP_VERSION(13, 0, 6):
+		case IP_VERSION(13, 0, 12):
+		case IP_VERSION(13, 0, 14):
+			return true;
+		default:
+			return false;
+		}
+	}
+
+	if (adev->asic_type == CHIP_IP_DISCOVERY) {
+		switch (amdgpu_ip_version(adev, MP0_HWIP, 0)) {
+		case IP_VERSION(13, 0, 0):
+		case IP_VERSION(13, 0, 6):
+		case IP_VERSION(13, 0, 10):
+		case IP_VERSION(13, 0, 12):
+		case IP_VERSION(13, 0, 14):
+		case IP_VERSION(14, 0, 3):
+			return true;
+		default:
+			return false;
+		}
+	}
+
 	return adev->asic_type == CHIP_VEGA10 ||
 		adev->asic_type == CHIP_VEGA20 ||
 		adev->asic_type == CHIP_ARCTURUS ||
@@ -2178,13 +4010,95 @@ static void amdgpu_ras_get_quirks(struct amdgpu_device *adev)
 	if (!ctx)
 		return;
 
-	if (strnstr(ctx->vbios_version, "D16406",
-		    sizeof(ctx->vbios_version)) ||
-		strnstr(ctx->vbios_version, "D36002",
-			sizeof(ctx->vbios_version)))
+	if (strnstr(ctx->vbios_pn, "D16406",
+		    sizeof(ctx->vbios_pn)) ||
+		strnstr(ctx->vbios_pn, "D36002",
+			sizeof(ctx->vbios_pn)))
 		adev->ras_hw_enabled |= (1 << AMDGPU_RAS_BLOCK__GFX);
 }
 
+/* Query ras capablity via atomfirmware interface */
+static void amdgpu_ras_query_ras_capablity_from_vbios(struct amdgpu_device *adev)
+{
+	/* mem_ecc cap */
+	if (amdgpu_atomfirmware_mem_ecc_supported(adev)) {
+		dev_info(adev->dev, "MEM ECC is active.\n");
+		adev->ras_hw_enabled |= (1 << AMDGPU_RAS_BLOCK__UMC |
+					 1 << AMDGPU_RAS_BLOCK__DF);
+	} else {
+		dev_info(adev->dev, "MEM ECC is not presented.\n");
+	}
+
+	/* sram_ecc cap */
+	if (amdgpu_atomfirmware_sram_ecc_supported(adev)) {
+		dev_info(adev->dev, "SRAM ECC is active.\n");
+		if (!amdgpu_sriov_vf(adev))
+			adev->ras_hw_enabled |= ~(1 << AMDGPU_RAS_BLOCK__UMC |
+						  1 << AMDGPU_RAS_BLOCK__DF);
+		else
+			adev->ras_hw_enabled |= (1 << AMDGPU_RAS_BLOCK__PCIE_BIF |
+						 1 << AMDGPU_RAS_BLOCK__SDMA |
+						 1 << AMDGPU_RAS_BLOCK__GFX);
+
+		/*
+		 * VCN/JPEG RAS can be supported on both bare metal and
+		 * SRIOV environment
+		 */
+		if (amdgpu_ip_version(adev, VCN_HWIP, 0) == IP_VERSION(2, 6, 0) ||
+		    amdgpu_ip_version(adev, VCN_HWIP, 0) == IP_VERSION(4, 0, 0) ||
+		    amdgpu_ip_version(adev, VCN_HWIP, 0) == IP_VERSION(4, 0, 3) ||
+		    amdgpu_ip_version(adev, VCN_HWIP, 0) == IP_VERSION(5, 0, 1))
+			adev->ras_hw_enabled |= (1 << AMDGPU_RAS_BLOCK__VCN |
+						 1 << AMDGPU_RAS_BLOCK__JPEG);
+		else
+			adev->ras_hw_enabled &= ~(1 << AMDGPU_RAS_BLOCK__VCN |
+						  1 << AMDGPU_RAS_BLOCK__JPEG);
+
+		/*
+		 * XGMI RAS is not supported if xgmi num physical nodes
+		 * is zero
+		 */
+		if (!adev->gmc.xgmi.num_physical_nodes)
+			adev->ras_hw_enabled &= ~(1 << AMDGPU_RAS_BLOCK__XGMI_WAFL);
+	} else {
+		dev_info(adev->dev, "SRAM ECC is not presented.\n");
+	}
+}
+
+/* Query poison mode from umc/df IP callbacks */
+static void amdgpu_ras_query_poison_mode(struct amdgpu_device *adev)
+{
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+	bool df_poison, umc_poison;
+
+	/* poison setting is useless on SRIOV guest */
+	if (amdgpu_sriov_vf(adev) || !con)
+		return;
+
+	/* Init poison supported flag, the default value is false */
+	if (adev->gmc.xgmi.connected_to_cpu ||
+	    adev->gmc.is_app_apu) {
+		/* enabled by default when GPU is connected to CPU */
+		con->poison_supported = true;
+	} else if (adev->df.funcs &&
+	    adev->df.funcs->query_ras_poison_mode &&
+	    adev->umc.ras &&
+	    adev->umc.ras->query_ras_poison_mode) {
+		df_poison =
+			adev->df.funcs->query_ras_poison_mode(adev);
+		umc_poison =
+			adev->umc.ras->query_ras_poison_mode(adev);
+
+		/* Only poison is set in both DF and UMC, we can support it */
+		if (df_poison && umc_poison)
+			con->poison_supported = true;
+		else if (df_poison != umc_poison)
+			dev_warn(adev->dev,
+				"Poison setting is inconsistent in DF/UMC(%d:%d)!\n",
+				df_poison, umc_poison);
+	}
+}
+
 /*
  * check hardware's ras ability which will be saved in hw_supported.
  * if hardware does not support ras, we can skip some ras initializtion and
@@ -2198,26 +4112,21 @@ static void amdgpu_ras_check_supported(struct amdgpu_device *adev)
 {
 	adev->ras_hw_enabled = adev->ras_enabled = 0;
 
-	if (amdgpu_sriov_vf(adev) || !adev->is_atom_fw ||
-	    !amdgpu_ras_asic_supported(adev))
+	if (!amdgpu_ras_asic_supported(adev))
 		return;
 
-	if (!adev->gmc.xgmi.connected_to_cpu) {
-		if (amdgpu_atomfirmware_mem_ecc_supported(adev)) {
-			dev_info(adev->dev, "MEM ECC is active.\n");
-			adev->ras_hw_enabled |= (1 << AMDGPU_RAS_BLOCK__UMC |
-						   1 << AMDGPU_RAS_BLOCK__DF);
-		} else {
-			dev_info(adev->dev, "MEM ECC is not presented.\n");
-		}
+	if (amdgpu_sriov_vf(adev)) {
+		if (amdgpu_virt_get_ras_capability(adev))
+			goto init_ras_enabled_flag;
+	}
 
-		if (amdgpu_atomfirmware_sram_ecc_supported(adev)) {
-			dev_info(adev->dev, "SRAM ECC is active.\n");
-			adev->ras_hw_enabled |= ~(1 << AMDGPU_RAS_BLOCK__UMC |
-						    1 << AMDGPU_RAS_BLOCK__DF);
-		} else {
-			dev_info(adev->dev, "SRAM ECC is not presented.\n");
-		}
+	/* query ras capability from psp */
+	if (amdgpu_psp_get_ras_capability(&adev->psp))
+		goto init_ras_enabled_flag;
+
+	/* query ras capablity from bios */
+	if (!adev->gmc.xgmi.connected_to_cpu && !adev->gmc.is_app_apu) {
+		amdgpu_ras_query_ras_capablity_from_vbios(adev);
 	} else {
 		/* driver only manages a few IP blocks RAS feature
 		 * when GPU is connected cpu through XGMI */
@@ -2226,13 +4135,31 @@ static void amdgpu_ras_check_supported(struct amdgpu_device *adev)
 					   1 << AMDGPU_RAS_BLOCK__MMHUB);
 	}
 
+	/* apply asic specific settings (vega20 only for now) */
 	amdgpu_ras_get_quirks(adev);
 
+	/* query poison mode from umc/df ip callback */
+	amdgpu_ras_query_poison_mode(adev);
+
+init_ras_enabled_flag:
 	/* hw_supported needs to be aligned with RAS block mask. */
 	adev->ras_hw_enabled &= AMDGPU_RAS_BLOCK_MASK;
 
 	adev->ras_enabled = amdgpu_ras_enable == 0 ? 0 :
 		adev->ras_hw_enabled & amdgpu_ras_mask;
+
+	/* aca is disabled by default except for psp v13_0_6/v13_0_12/v13_0_14 */
+	if (!amdgpu_sriov_vf(adev)) {
+		adev->aca.is_enabled =
+			(amdgpu_ip_version(adev, MP0_HWIP, 0) == IP_VERSION(13, 0, 6) ||
+			amdgpu_ip_version(adev, MP0_HWIP, 0) == IP_VERSION(13, 0, 12) ||
+			amdgpu_ip_version(adev, MP0_HWIP, 0) == IP_VERSION(13, 0, 14));
+	}
+
+	/* bad page feature is not applicable to specific app platform */
+	if (adev->gmc.is_app_apu &&
+	    amdgpu_ip_version(adev, UMC_HWIP, 0) == IP_VERSION(12, 0, 0))
+		amdgpu_bad_page_threshold = 0;
 }
 
 static void amdgpu_ras_counte_dw(struct work_struct *work)
@@ -2250,29 +4177,92 @@ static void amdgpu_ras_counte_dw(struct work_struct *work)
 
 	/* Cache new values.
 	 */
-	if (amdgpu_ras_query_error_count(adev, &ce_count, &ue_count) == 0) {
+	if (amdgpu_ras_query_error_count(adev, &ce_count, &ue_count, NULL) == 0) {
 		atomic_set(&con->ras_ce_count, ce_count);
 		atomic_set(&con->ras_ue_count, ue_count);
 	}
 
-	pm_runtime_mark_last_busy(dev->dev);
 Out:
 	pm_runtime_put_autosuspend(dev->dev);
 }
 
+static int amdgpu_get_ras_schema(struct amdgpu_device *adev)
+{
+	return  amdgpu_ras_is_poison_mode_supported(adev) ? AMDGPU_RAS_ERROR__POISON : 0 |
+			AMDGPU_RAS_ERROR__SINGLE_CORRECTABLE |
+			AMDGPU_RAS_ERROR__MULTI_UNCORRECTABLE |
+			AMDGPU_RAS_ERROR__PARITY;
+}
+
+static void ras_event_mgr_init(struct ras_event_manager *mgr)
+{
+	struct ras_event_state *event_state;
+	int i;
+
+	memset(mgr, 0, sizeof(*mgr));
+	atomic64_set(&mgr->seqno, 0);
+
+	for (i = 0; i < ARRAY_SIZE(mgr->event_state); i++) {
+		event_state = &mgr->event_state[i];
+		event_state->last_seqno = RAS_EVENT_INVALID_ID;
+		atomic64_set(&event_state->count, 0);
+	}
+}
+
+static void amdgpu_ras_event_mgr_init(struct amdgpu_device *adev)
+{
+	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
+	struct amdgpu_hive_info *hive;
+
+	if (!ras)
+		return;
+
+	hive = amdgpu_get_xgmi_hive(adev);
+	ras->event_mgr = hive ? &hive->event_mgr : &ras->__event_mgr;
+
+	/* init event manager with node 0 on xgmi system */
+	if (!amdgpu_reset_in_recovery(adev)) {
+		if (!hive || adev->gmc.xgmi.node_id == 0)
+			ras_event_mgr_init(ras->event_mgr);
+	}
+
+	if (hive)
+		amdgpu_put_xgmi_hive(hive);
+}
+
+static void amdgpu_ras_init_reserved_vram_size(struct amdgpu_device *adev)
+{
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+
+	if (!con || (adev->flags & AMD_IS_APU))
+		return;
+
+	switch (amdgpu_ip_version(adev, MP0_HWIP, 0)) {
+	case IP_VERSION(13, 0, 2):
+	case IP_VERSION(13, 0, 6):
+	case IP_VERSION(13, 0, 12):
+		con->reserved_pages_in_bytes = AMDGPU_RAS_RESERVED_VRAM_SIZE_DEFAULT;
+		break;
+	case IP_VERSION(13, 0, 14):
+		con->reserved_pages_in_bytes = (AMDGPU_RAS_RESERVED_VRAM_SIZE_DEFAULT << 1);
+		break;
+	default:
+		break;
+	}
+}
+
 int amdgpu_ras_init(struct amdgpu_device *adev)
 {
 	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
 	int r;
-	bool df_poison, umc_poison;
 
 	if (con)
 		return 0;
 
-	con = kmalloc(sizeof(struct amdgpu_ras) +
+	con = kzalloc(sizeof(*con) +
 			sizeof(struct ras_manager) * AMDGPU_RAS_BLOCK_COUNT +
 			sizeof(struct ras_manager) * AMDGPU_RAS_MCA_BLOCK_COUNT,
-			GFP_KERNEL|__GFP_ZERO);
+			GFP_KERNEL);
 	if (!con)
 		return -ENOMEM;
 
@@ -2301,7 +4291,9 @@ int amdgpu_ras_init(struct amdgpu_device *adev)
 		goto release_con;
 	}
 
+	con->update_channel_flag = false;
 	con->features = 0;
+	con->schema = 0;
 	INIT_LIST_HEAD(&con->head);
 	/* Might need get this flag from vbios. */
 	con->flags = RAS_DEFAULT_FLAGS;
@@ -2309,54 +4301,95 @@ int amdgpu_ras_init(struct amdgpu_device *adev)
 	/* initialize nbio ras function ahead of any other
 	 * ras functions so hardware fatal error interrupt
 	 * can be enabled as early as possible */
-	switch (adev->asic_type) {
-	case CHIP_VEGA20:
-	case CHIP_ARCTURUS:
-	case CHIP_ALDEBARAN:
+	switch (amdgpu_ip_version(adev, NBIO_HWIP, 0)) {
+	case IP_VERSION(7, 4, 0):
+	case IP_VERSION(7, 4, 1):
+	case IP_VERSION(7, 4, 4):
 		if (!adev->gmc.xgmi.connected_to_cpu)
-			adev->nbio.ras_funcs = &nbio_v7_4_ras_funcs;
+			adev->nbio.ras = &nbio_v7_4_ras;
+		break;
+	case IP_VERSION(4, 3, 0):
+		if (adev->ras_hw_enabled & (1 << AMDGPU_RAS_BLOCK__DF))
+			/* unlike other generation of nbio ras,
+			 * nbio v4_3 only support fatal error interrupt
+			 * to inform software that DF is freezed due to
+			 * system fatal error event. driver should not
+			 * enable nbio ras in such case. Instead,
+			 * check DF RAS */
+			adev->nbio.ras = &nbio_v4_3_ras;
+		break;
+	case IP_VERSION(6, 3, 1):
+		if (adev->ras_hw_enabled & (1 << AMDGPU_RAS_BLOCK__DF))
+			/* unlike other generation of nbio ras,
+			 * nbif v6_3_1 only support fatal error interrupt
+			 * to inform software that DF is freezed due to
+			 * system fatal error event. driver should not
+			 * enable nbio ras in such case. Instead,
+			 * check DF RAS
+			 */
+			adev->nbio.ras = &nbif_v6_3_1_ras;
+		break;
+	case IP_VERSION(7, 9, 0):
+	case IP_VERSION(7, 9, 1):
+		if (!adev->gmc.is_app_apu)
+			adev->nbio.ras = &nbio_v7_9_ras;
 		break;
 	default:
 		/* nbio ras is not available */
 		break;
 	}
 
-	if (adev->nbio.ras_funcs &&
-	    adev->nbio.ras_funcs->init_ras_controller_interrupt) {
-		r = adev->nbio.ras_funcs->init_ras_controller_interrupt(adev);
+	/* nbio ras block needs to be enabled ahead of other ras blocks
+	 * to handle fatal error */
+	r = amdgpu_nbio_ras_sw_init(adev);
+	if (r)
+		return r;
+
+	if (adev->nbio.ras &&
+	    adev->nbio.ras->init_ras_controller_interrupt) {
+		r = adev->nbio.ras->init_ras_controller_interrupt(adev);
 		if (r)
 			goto release_con;
 	}
 
-	if (adev->nbio.ras_funcs &&
-	    adev->nbio.ras_funcs->init_ras_err_event_athub_interrupt) {
-		r = adev->nbio.ras_funcs->init_ras_err_event_athub_interrupt(adev);
+	if (adev->nbio.ras &&
+	    adev->nbio.ras->init_ras_err_event_athub_interrupt) {
+		r = adev->nbio.ras->init_ras_err_event_athub_interrupt(adev);
 		if (r)
 			goto release_con;
 	}
 
-	/* Init poison supported flag, the default value is false */
-	if (adev->df.funcs &&
-	    adev->df.funcs->query_ras_poison_mode &&
-	    adev->umc.ras_funcs &&
-	    adev->umc.ras_funcs->query_ras_poison_mode) {
-		df_poison =
-			adev->df.funcs->query_ras_poison_mode(adev);
-		umc_poison =
-			adev->umc.ras_funcs->query_ras_poison_mode(adev);
-		/* Only poison is set in both DF and UMC, we can support it */
-		if (df_poison && umc_poison)
-			con->poison_supported = true;
-		else if (df_poison != umc_poison)
-			dev_warn(adev->dev, "Poison setting is inconsistent in DF/UMC(%d:%d)!\n",
-					df_poison, umc_poison);
-	}
+	/* Packed socket_id to ras feature mask bits[31:29] */
+	if (adev->smuio.funcs &&
+	    adev->smuio.funcs->get_socket_id)
+		con->features |= ((adev->smuio.funcs->get_socket_id(adev)) <<
+					AMDGPU_RAS_FEATURES_SOCKETID_SHIFT);
+
+	/* Get RAS schema for particular SOC */
+	con->schema = amdgpu_get_ras_schema(adev);
+
+	amdgpu_ras_init_reserved_vram_size(adev);
 
 	if (amdgpu_ras_fs_init(adev)) {
 		r = -EINVAL;
 		goto release_con;
 	}
 
+	if (amdgpu_ras_aca_is_supported(adev)) {
+		if (amdgpu_aca_is_enabled(adev))
+			r = amdgpu_aca_init(adev);
+		else
+			r = amdgpu_mca_init(adev);
+		if (r)
+			goto release_con;
+	}
+
+	con->init_task_pid = task_pid_nr(current);
+	get_task_comm(con->init_task_comm, current);
+
+	mutex_init(&con->critical_region_lock);
+	INIT_LIST_HEAD(&con->critical_region_head);
+
 	dev_info(adev->dev, "RAS INFO: ras initialized successfully, "
 		 "hardware ability[%x] ras_mask[%x]\n",
 		 adev->ras_hw_enabled, adev->ras_enabled);
@@ -2371,7 +4404,8 @@ release_con:
 
 int amdgpu_persistent_edc_harvesting_supported(struct amdgpu_device *adev)
 {
-	if (adev->gmc.xgmi.connected_to_cpu)
+	if (adev->gmc.xgmi.connected_to_cpu ||
+	    adev->gmc.is_app_apu)
 		return 1;
 	return 0;
 }
@@ -2406,12 +4440,12 @@ bool amdgpu_ras_is_poison_mode_supported(struct amdgpu_device *adev)
 }
 
 /* helper function to handle common stuff in ip late init phase */
-int amdgpu_ras_late_init(struct amdgpu_device *adev,
-			 struct ras_common_if *ras_block,
-			 struct ras_fs_if *fs_info,
-			 struct ras_ih_if *ih_info)
+int amdgpu_ras_block_late_init(struct amdgpu_device *adev,
+			 struct ras_common_if *ras_block)
 {
+	struct amdgpu_ras_block_object *ras_obj = NULL;
 	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+	struct ras_query_if *query_info;
 	unsigned long ue_count, ce_count;
 	int r;
 
@@ -2423,7 +4457,7 @@ int amdgpu_ras_late_init(struct amdgpu_device *adev,
 
 	r = amdgpu_ras_feature_enable_on_boot(adev, ras_block, 1);
 	if (r) {
-		if (adev->in_suspend || amdgpu_in_reset(adev)) {
+		if (adev->in_suspend || amdgpu_reset_in_recovery(adev)) {
 			/* in resume phase, if fail to enable ras,
 			 * clean up all ras fs nodes, and disable ras */
 			goto cleanup;
@@ -2435,49 +4469,75 @@ int amdgpu_ras_late_init(struct amdgpu_device *adev,
 	amdgpu_persistent_edc_harvesting(adev, ras_block);
 
 	/* in resume phase, no need to create ras fs node */
-	if (adev->in_suspend || amdgpu_in_reset(adev))
+	if (adev->in_suspend || amdgpu_reset_in_recovery(adev))
 		return 0;
 
-	if (ih_info->cb) {
-		r = amdgpu_ras_interrupt_add_handler(adev, ih_info);
+	ras_obj = container_of(ras_block, struct amdgpu_ras_block_object, ras_comm);
+	if (ras_obj->ras_cb || (ras_obj->hw_ops &&
+	    (ras_obj->hw_ops->query_poison_status ||
+	    ras_obj->hw_ops->handle_poison_consumption))) {
+		r = amdgpu_ras_interrupt_add_handler(adev, ras_block);
 		if (r)
-			goto interrupt;
+			goto cleanup;
 	}
 
-	r = amdgpu_ras_sysfs_create(adev, fs_info);
-	if (r)
-		goto sysfs;
+	if (ras_obj->hw_ops &&
+	    (ras_obj->hw_ops->query_ras_error_count ||
+	     ras_obj->hw_ops->query_ras_error_status)) {
+		r = amdgpu_ras_sysfs_create(adev, ras_block);
+		if (r)
+			goto interrupt;
 
-	/* Those are the cached values at init.
-	 */
-	if (amdgpu_ras_query_error_count(adev, &ce_count, &ue_count) == 0) {
-		atomic_set(&con->ras_ce_count, ce_count);
-		atomic_set(&con->ras_ue_count, ue_count);
+		/* Those are the cached values at init.
+		 */
+		query_info = kzalloc(sizeof(*query_info), GFP_KERNEL);
+		if (!query_info)
+			return -ENOMEM;
+		memcpy(&query_info->head, ras_block, sizeof(struct ras_common_if));
+
+		if (amdgpu_ras_query_error_count(adev, &ce_count, &ue_count, query_info) == 0) {
+			atomic_set(&con->ras_ce_count, ce_count);
+			atomic_set(&con->ras_ue_count, ue_count);
+		}
+
+		kfree(query_info);
 	}
 
 	return 0;
-cleanup:
-	amdgpu_ras_sysfs_remove(adev, ras_block);
-sysfs:
-	if (ih_info->cb)
-		amdgpu_ras_interrupt_remove_handler(adev, ih_info);
+
 interrupt:
+	if (ras_obj->ras_cb)
+		amdgpu_ras_interrupt_remove_handler(adev, ras_block);
+cleanup:
 	amdgpu_ras_feature_enable(adev, ras_block, 0);
 	return r;
 }
 
+static int amdgpu_ras_block_late_init_default(struct amdgpu_device *adev,
+			 struct ras_common_if *ras_block)
+{
+	return amdgpu_ras_block_late_init(adev, ras_block);
+}
+
 /* helper function to remove ras fs node and interrupt handler */
-void amdgpu_ras_late_fini(struct amdgpu_device *adev,
-			  struct ras_common_if *ras_block,
-			  struct ras_ih_if *ih_info)
+void amdgpu_ras_block_late_fini(struct amdgpu_device *adev,
+			  struct ras_common_if *ras_block)
 {
-	if (!ras_block || !ih_info)
+	struct amdgpu_ras_block_object *ras_obj;
+	if (!ras_block)
 		return;
 
 	amdgpu_ras_sysfs_remove(adev, ras_block);
-	if (ih_info->cb)
-		amdgpu_ras_interrupt_remove_handler(adev, ih_info);
-	amdgpu_ras_feature_enable(adev, ras_block, 0);
+
+	ras_obj = container_of(ras_block, struct amdgpu_ras_block_object, ras_comm);
+	if (ras_obj->ras_cb)
+		amdgpu_ras_interrupt_remove_handler(adev, ras_block);
+}
+
+static void amdgpu_ras_block_late_fini_default(struct amdgpu_device *adev,
+			  struct ras_common_if *ras_block)
+{
+	return amdgpu_ras_block_late_fini(adev, ras_block);
 }
 
 /* do some init work after IP late init as dependence.
@@ -2526,10 +4586,64 @@ void amdgpu_ras_suspend(struct amdgpu_device *adev)
 
 	amdgpu_ras_disable_all_features(adev, 0);
 	/* Make sure all ras objects are disabled. */
-	if (con->features)
+	if (AMDGPU_RAS_GET_FEATURES(con->features))
 		amdgpu_ras_disable_all_features(adev, 1);
 }
 
+int amdgpu_ras_late_init(struct amdgpu_device *adev)
+{
+	struct amdgpu_ras_block_list *node, *tmp;
+	struct amdgpu_ras_block_object *obj;
+	int r;
+
+	amdgpu_ras_event_mgr_init(adev);
+
+	if (amdgpu_ras_aca_is_supported(adev)) {
+		if (amdgpu_reset_in_recovery(adev)) {
+			if (amdgpu_aca_is_enabled(adev))
+				r = amdgpu_aca_reset(adev);
+			else
+				r = amdgpu_mca_reset(adev);
+			if (r)
+				return r;
+		}
+
+		if (!amdgpu_sriov_vf(adev)) {
+			if (amdgpu_aca_is_enabled(adev))
+				amdgpu_ras_set_aca_debug_mode(adev, false);
+			else
+				amdgpu_ras_set_mca_debug_mode(adev, false);
+		}
+	}
+
+	/* Guest side doesn't need init ras feature */
+	if (amdgpu_sriov_vf(adev) && !amdgpu_sriov_ras_telemetry_en(adev))
+		return 0;
+
+	list_for_each_entry_safe(node, tmp, &adev->ras_list, node) {
+		obj = node->ras_obj;
+		if (!obj) {
+			dev_warn(adev->dev, "Warning: abnormal ras list node.\n");
+			continue;
+		}
+
+		if (!amdgpu_ras_is_supported(adev, obj->ras_comm.block))
+			continue;
+
+		if (obj->ras_late_init) {
+			r = obj->ras_late_init(adev, &obj->ras_comm);
+			if (r) {
+				dev_err(adev->dev, "%s failed to execute ras_late_init! ret:%d\n",
+					obj->ras_comm.name, r);
+				return r;
+			}
+		} else
+			amdgpu_ras_block_late_init_default(adev, &obj->ras_comm);
+	}
+
+	return 0;
+}
+
 /* do some fini work before IP fini as dependence */
 int amdgpu_ras_pre_fini(struct amdgpu_device *adev)
 {
@@ -2540,25 +4654,53 @@ int amdgpu_ras_pre_fini(struct amdgpu_device *adev)
 
 
 	/* Need disable ras on all IPs here before ip [hw/sw]fini */
-	amdgpu_ras_disable_all_features(adev, 0);
+	if (AMDGPU_RAS_GET_FEATURES(con->features))
+		amdgpu_ras_disable_all_features(adev, 0);
 	amdgpu_ras_recovery_fini(adev);
 	return 0;
 }
 
 int amdgpu_ras_fini(struct amdgpu_device *adev)
 {
+	struct amdgpu_ras_block_list *ras_node, *tmp;
+	struct amdgpu_ras_block_object *obj = NULL;
 	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
 
 	if (!adev->ras_enabled || !con)
 		return 0;
 
+	amdgpu_ras_critical_region_fini(adev);
+	mutex_destroy(&con->critical_region_lock);
+
+	list_for_each_entry_safe(ras_node, tmp, &adev->ras_list, node) {
+		if (ras_node->ras_obj) {
+			obj = ras_node->ras_obj;
+			if (amdgpu_ras_is_supported(adev, obj->ras_comm.block) &&
+			    obj->ras_fini)
+				obj->ras_fini(adev, &obj->ras_comm);
+			else
+				amdgpu_ras_block_late_fini_default(adev, &obj->ras_comm);
+		}
+
+		/* Clear ras blocks from ras_list and free ras block list node */
+		list_del(&ras_node->node);
+		kfree(ras_node);
+	}
+
 	amdgpu_ras_fs_fini(adev);
 	amdgpu_ras_interrupt_remove_all(adev);
 
-	WARN(con->features, "Feature mask is not cleared");
+	if (amdgpu_ras_aca_is_supported(adev)) {
+		if (amdgpu_aca_is_enabled(adev))
+			amdgpu_aca_fini(adev);
+		else
+			amdgpu_mca_fini(adev);
+	}
 
-	if (con->features)
-		amdgpu_ras_disable_all_features(adev, 1);
+	WARN(AMDGPU_RAS_GET_FEATURES(con->features), "Feature mask is not cleared");
+
+	if (AMDGPU_RAS_GET_FEATURES(con->features))
+		amdgpu_ras_disable_all_features(adev, 0);
 
 	cancel_delayed_work_sync(&con->ras_counte_delay_work);
 
@@ -2568,18 +4710,161 @@ int amdgpu_ras_fini(struct amdgpu_device *adev)
 	return 0;
 }
 
-void amdgpu_ras_global_ras_isr(struct amdgpu_device *adev)
+bool amdgpu_ras_get_fed_status(struct amdgpu_device *adev)
 {
-	amdgpu_ras_check_supported(adev);
-	if (!adev->ras_hw_enabled)
-		return;
+	struct amdgpu_ras *ras;
+
+	ras = amdgpu_ras_get_context(adev);
+	if (!ras)
+		return false;
+
+	return test_bit(AMDGPU_RAS_BLOCK__LAST, &ras->ras_err_state);
+}
+
+void amdgpu_ras_set_fed(struct amdgpu_device *adev, bool status)
+{
+	struct amdgpu_ras *ras;
+
+	ras = amdgpu_ras_get_context(adev);
+	if (ras) {
+		if (status)
+			set_bit(AMDGPU_RAS_BLOCK__LAST, &ras->ras_err_state);
+		else
+			clear_bit(AMDGPU_RAS_BLOCK__LAST, &ras->ras_err_state);
+	}
+}
+
+void amdgpu_ras_clear_err_state(struct amdgpu_device *adev)
+{
+	struct amdgpu_ras *ras;
+
+	ras = amdgpu_ras_get_context(adev);
+	if (ras) {
+		ras->ras_err_state = 0;
+		ras->gpu_reset_flags = 0;
+	}
+}
 
+void amdgpu_ras_set_err_poison(struct amdgpu_device *adev,
+			       enum amdgpu_ras_block block)
+{
+	struct amdgpu_ras *ras;
+
+	ras = amdgpu_ras_get_context(adev);
+	if (ras)
+		set_bit(block, &ras->ras_err_state);
+}
+
+bool amdgpu_ras_is_err_state(struct amdgpu_device *adev, int block)
+{
+	struct amdgpu_ras *ras;
+
+	ras = amdgpu_ras_get_context(adev);
+	if (ras) {
+		if (block == AMDGPU_RAS_BLOCK__ANY)
+			return (ras->ras_err_state != 0);
+		else
+			return test_bit(block, &ras->ras_err_state) ||
+			       test_bit(AMDGPU_RAS_BLOCK__LAST,
+					&ras->ras_err_state);
+	}
+
+	return false;
+}
+
+static struct ras_event_manager *__get_ras_event_mgr(struct amdgpu_device *adev)
+{
+	struct amdgpu_ras *ras;
+
+	ras = amdgpu_ras_get_context(adev);
+	if (!ras)
+		return NULL;
+
+	return ras->event_mgr;
+}
+
+int amdgpu_ras_mark_ras_event_caller(struct amdgpu_device *adev, enum ras_event_type type,
+				     const void *caller)
+{
+	struct ras_event_manager *event_mgr;
+	struct ras_event_state *event_state;
+	int ret = 0;
+
+	if (amdgpu_uniras_enabled(adev))
+		return 0;
+
+	if (type >= RAS_EVENT_TYPE_COUNT) {
+		ret = -EINVAL;
+		goto out;
+	}
+
+	event_mgr = __get_ras_event_mgr(adev);
+	if (!event_mgr) {
+		ret = -EINVAL;
+		goto out;
+	}
+
+	event_state = &event_mgr->event_state[type];
+	event_state->last_seqno = atomic64_inc_return(&event_mgr->seqno);
+	atomic64_inc(&event_state->count);
+
+out:
+	if (ret && caller)
+		dev_warn(adev->dev, "failed mark ras event (%d) in %ps, ret:%d\n",
+			 (int)type, caller, ret);
+
+	return ret;
+}
+
+u64 amdgpu_ras_acquire_event_id(struct amdgpu_device *adev, enum ras_event_type type)
+{
+	struct ras_event_manager *event_mgr;
+	u64 id;
+
+	if (type >= RAS_EVENT_TYPE_COUNT)
+		return RAS_EVENT_INVALID_ID;
+
+	switch (type) {
+	case RAS_EVENT_TYPE_FATAL:
+	case RAS_EVENT_TYPE_POISON_CREATION:
+	case RAS_EVENT_TYPE_POISON_CONSUMPTION:
+		event_mgr = __get_ras_event_mgr(adev);
+		if (!event_mgr)
+			return RAS_EVENT_INVALID_ID;
+
+		id = event_mgr->event_state[type].last_seqno;
+		break;
+	case RAS_EVENT_TYPE_INVALID:
+	default:
+		id = RAS_EVENT_INVALID_ID;
+		break;
+	}
+
+	return id;
+}
+
+int amdgpu_ras_global_ras_isr(struct amdgpu_device *adev)
+{
 	if (atomic_cmpxchg(&amdgpu_ras_in_intr, 0, 1) == 0) {
-		dev_info(adev->dev, "uncorrectable hardware error"
-			"(ERREVENT_ATHUB_INTERRUPT) detected!\n");
+		struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
+		enum ras_event_type type = RAS_EVENT_TYPE_FATAL;
+		u64 event_id = RAS_EVENT_INVALID_ID;
+
+		if (amdgpu_uniras_enabled(adev))
+			return 0;
 
+		if (!amdgpu_ras_mark_ras_event(adev, type))
+			event_id = amdgpu_ras_acquire_event_id(adev, type);
+
+		RAS_EVENT_LOG(adev, event_id, "uncorrectable hardware error"
+			      "(ERREVENT_ATHUB_INTERRUPT) detected!\n");
+
+		amdgpu_ras_set_fed(adev, true);
+		ras->gpu_reset_flags |= AMDGPU_RAS_GPU_RESET_MODE1_RESET;
 		amdgpu_ras_reset_gpu(adev);
 	}
+
+	return -EBUSY;
 }
 
 bool amdgpu_ras_need_emergency_restart(struct amdgpu_device *adev)
@@ -2636,18 +4921,14 @@ static int amdgpu_bad_page_notifier(struct notifier_block *nb,
 	struct mce *m = (struct mce *)data;
 	struct amdgpu_device *adev = NULL;
 	uint32_t gpu_id = 0;
-	uint32_t umc_inst = 0;
-	uint32_t ch_inst, channel_index = 0;
-	struct ras_err_data err_data = {0, 0, 0, NULL};
-	struct eeprom_table_record err_rec;
-	uint64_t retired_page;
+	uint32_t umc_inst = 0, ch_inst = 0;
 
 	/*
 	 * If the error was generated in UMC_V2, which belongs to GPU UMCs,
 	 * and error occurred in DramECC (Extended error code = 0) then only
 	 * process the error, else bail out.
 	 */
-	if (!m || !((smca_get_bank_type(m->bank) == SMCA_UMC_V2) &&
+	if (!m || !((smca_get_bank_type(m->extcpu, m->bank) == SMCA_UMC_V2) &&
 		    (XEC(m->status, 0x3f) == 0x0)))
 		return NOTIFY_DONE;
 
@@ -2679,37 +4960,10 @@ static int amdgpu_bad_page_notifier(struct notifier_block *nb,
 	dev_info(adev->dev, "Uncorrectable error detected in UMC inst: %d, chan_idx: %d",
 			     umc_inst, ch_inst);
 
-	memset(&err_rec, 0x0, sizeof(struct eeprom_table_record));
-
-	/*
-	 * Translate UMC channel address to Physical address
-	 */
-	channel_index =
-		adev->umc.channel_idx_tbl[umc_inst * adev->umc.channel_inst_num
-					  + ch_inst];
-
-	retired_page = ADDR_OF_8KB_BLOCK(m->addr) |
-			ADDR_OF_256B_BLOCK(channel_index) |
-			OFFSET_IN_256B_BLOCK(m->addr);
-
-	err_rec.address = m->addr;
-	err_rec.retired_page = retired_page >> AMDGPU_GPU_PAGE_SHIFT;
-	err_rec.ts = (uint64_t)ktime_get_real_seconds();
-	err_rec.err_type = AMDGPU_RAS_EEPROM_ERR_NON_RECOVERABLE;
-	err_rec.cu = 0;
-	err_rec.mem_channel = channel_index;
-	err_rec.mcumc_id = umc_inst;
-
-	err_data.err_addr = &err_rec;
-	err_data.err_addr_cnt = 1;
-
-	if (amdgpu_bad_page_threshold != 0) {
-		amdgpu_ras_add_bad_pages(adev, err_data.err_addr,
-						err_data.err_addr_cnt);
-		amdgpu_ras_save_bad_pages(adev);
-	}
-
-	return NOTIFY_OK;
+	if (!amdgpu_umc_page_retirement_mca(adev, m->addr, ch_inst, umc_inst))
+		return NOTIFY_OK;
+	else
+		return NOTIFY_DONE;
 }
 
 static struct notifier_block amdgpu_bad_page_nb = {
@@ -2738,4 +4992,757 @@ static void amdgpu_register_bad_pages_mca_notifier(struct amdgpu_device *adev)
 		notifier_registered = true;
 	}
 }
+static void amdgpu_unregister_bad_pages_mca_notifier(struct amdgpu_device *adev)
+{
+	int i, j;
+
+	if (!notifier_registered && !mce_adev_list.num_gpu)
+		return;
+	for (i = 0, j = 0; i < mce_adev_list.num_gpu; i++) {
+		if (mce_adev_list.devs[i] == adev)
+			mce_adev_list.devs[i] = NULL;
+		if (!mce_adev_list.devs[i])
+			++j;
+	}
+
+	if (j == mce_adev_list.num_gpu) {
+		mce_adev_list.num_gpu = 0;
+		/* Unregister x86 notifier with MCE subsystem. */
+		if (notifier_registered) {
+			mce_unregister_decode_chain(&amdgpu_bad_page_nb);
+			notifier_registered = false;
+		}
+	}
+}
 #endif
+
+struct amdgpu_ras *amdgpu_ras_get_context(struct amdgpu_device *adev)
+{
+	if (!adev)
+		return NULL;
+
+	return adev->psp.ras_context.ras;
+}
+
+int amdgpu_ras_set_context(struct amdgpu_device *adev, struct amdgpu_ras *ras_con)
+{
+	if (!adev)
+		return -EINVAL;
+
+	adev->psp.ras_context.ras = ras_con;
+	return 0;
+}
+
+/* check if ras is supported on block, say, sdma, gfx */
+int amdgpu_ras_is_supported(struct amdgpu_device *adev,
+		unsigned int block)
+{
+	int ret = 0;
+	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
+
+	if (block >= AMDGPU_RAS_BLOCK_COUNT)
+		return 0;
+
+	ret = ras && (adev->ras_enabled & (1 << block));
+
+	/* For the special asic with mem ecc enabled but sram ecc
+	 * not enabled, even if the ras block is not supported on
+	 * .ras_enabled, if the asic supports poison mode and the
+	 * ras block has ras configuration, it can be considered
+	 * that the ras block supports ras function.
+	 */
+	if (!ret &&
+	    (block == AMDGPU_RAS_BLOCK__GFX ||
+	     block == AMDGPU_RAS_BLOCK__SDMA ||
+	     block == AMDGPU_RAS_BLOCK__VCN ||
+	     block == AMDGPU_RAS_BLOCK__JPEG) &&
+		(amdgpu_ras_mask & (1 << block)) &&
+	    amdgpu_ras_is_poison_mode_supported(adev) &&
+	    amdgpu_ras_get_ras_block(adev, block, 0))
+		ret = 1;
+
+	return ret;
+}
+
+int amdgpu_ras_reset_gpu(struct amdgpu_device *adev)
+{
+	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
+
+	/* mode1 is the only selection for RMA status */
+	if (amdgpu_ras_is_rma(adev)) {
+		ras->gpu_reset_flags = 0;
+		ras->gpu_reset_flags |= AMDGPU_RAS_GPU_RESET_MODE1_RESET;
+	}
+
+	if (atomic_cmpxchg(&ras->in_recovery, 0, 1) == 0) {
+		struct amdgpu_hive_info *hive = amdgpu_get_xgmi_hive(adev);
+		int hive_ras_recovery = 0;
+
+		if (hive) {
+			hive_ras_recovery = atomic_read(&hive->ras_recovery);
+			amdgpu_put_xgmi_hive(hive);
+		}
+		/* In the case of multiple GPUs, after a GPU has started
+		 * resetting all GPUs on hive, other GPUs do not need to
+		 * trigger GPU reset again.
+		 */
+		if (!hive_ras_recovery)
+			amdgpu_reset_domain_schedule(ras->adev->reset_domain, &ras->recovery_work);
+		else
+			atomic_set(&ras->in_recovery, 0);
+	} else {
+		flush_work(&ras->recovery_work);
+		amdgpu_reset_domain_schedule(ras->adev->reset_domain, &ras->recovery_work);
+	}
+
+	return 0;
+}
+
+int amdgpu_ras_set_mca_debug_mode(struct amdgpu_device *adev, bool enable)
+{
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+	int ret = 0;
+
+	if (con) {
+		ret = amdgpu_mca_smu_set_debug_mode(adev, enable);
+		if (!ret)
+			con->is_aca_debug_mode = enable;
+	}
+
+	return ret;
+}
+
+int amdgpu_ras_set_aca_debug_mode(struct amdgpu_device *adev, bool enable)
+{
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+	int ret = 0;
+
+	if (con) {
+		if (amdgpu_aca_is_enabled(adev))
+			ret = amdgpu_aca_smu_set_debug_mode(adev, enable);
+		else
+			ret = amdgpu_mca_smu_set_debug_mode(adev, enable);
+		if (!ret)
+			con->is_aca_debug_mode = enable;
+	}
+
+	return ret;
+}
+
+bool amdgpu_ras_get_aca_debug_mode(struct amdgpu_device *adev)
+{
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+	const struct aca_smu_funcs *smu_funcs = adev->aca.smu_funcs;
+	const struct amdgpu_mca_smu_funcs *mca_funcs = adev->mca.mca_funcs;
+
+	if (!con)
+		return false;
+
+	if ((amdgpu_aca_is_enabled(adev) && smu_funcs && smu_funcs->set_debug_mode) ||
+	    (!amdgpu_aca_is_enabled(adev) && mca_funcs && mca_funcs->mca_set_debug_mode))
+		return con->is_aca_debug_mode;
+	else
+		return true;
+}
+
+bool amdgpu_ras_get_error_query_mode(struct amdgpu_device *adev,
+				     unsigned int *error_query_mode)
+{
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+	const struct amdgpu_mca_smu_funcs *mca_funcs = adev->mca.mca_funcs;
+	const struct aca_smu_funcs *smu_funcs = adev->aca.smu_funcs;
+
+	if (!con) {
+		*error_query_mode = AMDGPU_RAS_INVALID_ERROR_QUERY;
+		return false;
+	}
+
+	if (amdgpu_sriov_vf(adev)) {
+		*error_query_mode = AMDGPU_RAS_VIRT_ERROR_COUNT_QUERY;
+	} else if ((smu_funcs && smu_funcs->set_debug_mode) || (mca_funcs && mca_funcs->mca_set_debug_mode)) {
+		*error_query_mode =
+			(con->is_aca_debug_mode) ? AMDGPU_RAS_DIRECT_ERROR_QUERY : AMDGPU_RAS_FIRMWARE_ERROR_QUERY;
+	} else {
+		*error_query_mode = AMDGPU_RAS_DIRECT_ERROR_QUERY;
+	}
+
+	return true;
+}
+
+/* Register each ip ras block into amdgpu ras */
+int amdgpu_ras_register_ras_block(struct amdgpu_device *adev,
+		struct amdgpu_ras_block_object *ras_block_obj)
+{
+	struct amdgpu_ras_block_list *ras_node;
+	if (!adev || !ras_block_obj)
+		return -EINVAL;
+
+	ras_node = kzalloc(sizeof(*ras_node), GFP_KERNEL);
+	if (!ras_node)
+		return -ENOMEM;
+
+	INIT_LIST_HEAD(&ras_node->node);
+	ras_node->ras_obj = ras_block_obj;
+	list_add_tail(&ras_node->node, &adev->ras_list);
+
+	return 0;
+}
+
+void amdgpu_ras_get_error_type_name(uint32_t err_type, char *err_type_name)
+{
+	if (!err_type_name)
+		return;
+
+	switch (err_type) {
+	case AMDGPU_RAS_ERROR__SINGLE_CORRECTABLE:
+		sprintf(err_type_name, "correctable");
+		break;
+	case AMDGPU_RAS_ERROR__MULTI_UNCORRECTABLE:
+		sprintf(err_type_name, "uncorrectable");
+		break;
+	default:
+		sprintf(err_type_name, "unknown");
+		break;
+	}
+}
+
+bool amdgpu_ras_inst_get_memory_id_field(struct amdgpu_device *adev,
+					 const struct amdgpu_ras_err_status_reg_entry *reg_entry,
+					 uint32_t instance,
+					 uint32_t *memory_id)
+{
+	uint32_t err_status_lo_data, err_status_lo_offset;
+
+	if (!reg_entry)
+		return false;
+
+	err_status_lo_offset =
+		AMDGPU_RAS_REG_ENTRY_OFFSET(reg_entry->hwip, instance,
+					    reg_entry->seg_lo, reg_entry->reg_lo);
+	err_status_lo_data = RREG32(err_status_lo_offset);
+
+	if ((reg_entry->flags & AMDGPU_RAS_ERR_STATUS_VALID) &&
+	    !REG_GET_FIELD(err_status_lo_data, ERR_STATUS_LO, ERR_STATUS_VALID_FLAG))
+		return false;
+
+	*memory_id = REG_GET_FIELD(err_status_lo_data, ERR_STATUS_LO, MEMORY_ID);
+
+	return true;
+}
+
+bool amdgpu_ras_inst_get_err_cnt_field(struct amdgpu_device *adev,
+				       const struct amdgpu_ras_err_status_reg_entry *reg_entry,
+				       uint32_t instance,
+				       unsigned long *err_cnt)
+{
+	uint32_t err_status_hi_data, err_status_hi_offset;
+
+	if (!reg_entry)
+		return false;
+
+	err_status_hi_offset =
+		AMDGPU_RAS_REG_ENTRY_OFFSET(reg_entry->hwip, instance,
+					    reg_entry->seg_hi, reg_entry->reg_hi);
+	err_status_hi_data = RREG32(err_status_hi_offset);
+
+	if ((reg_entry->flags & AMDGPU_RAS_ERR_INFO_VALID) &&
+	    !REG_GET_FIELD(err_status_hi_data, ERR_STATUS_HI, ERR_INFO_VALID_FLAG))
+		/* keep the check here in case we need to refer to the result later */
+		dev_dbg(adev->dev, "Invalid err_info field\n");
+
+	/* read err count */
+	*err_cnt = REG_GET_FIELD(err_status_hi_data, ERR_STATUS, ERR_CNT);
+
+	return true;
+}
+
+void amdgpu_ras_inst_query_ras_error_count(struct amdgpu_device *adev,
+					   const struct amdgpu_ras_err_status_reg_entry *reg_list,
+					   uint32_t reg_list_size,
+					   const struct amdgpu_ras_memory_id_entry *mem_list,
+					   uint32_t mem_list_size,
+					   uint32_t instance,
+					   uint32_t err_type,
+					   unsigned long *err_count)
+{
+	uint32_t memory_id;
+	unsigned long err_cnt;
+	char err_type_name[16];
+	uint32_t i, j;
+
+	for (i = 0; i < reg_list_size; i++) {
+		/* query memory_id from err_status_lo */
+		if (!amdgpu_ras_inst_get_memory_id_field(adev, &reg_list[i],
+							 instance, &memory_id))
+			continue;
+
+		/* query err_cnt from err_status_hi */
+		if (!amdgpu_ras_inst_get_err_cnt_field(adev, &reg_list[i],
+						       instance, &err_cnt) ||
+		    !err_cnt)
+			continue;
+
+		*err_count += err_cnt;
+
+		/* log the errors */
+		amdgpu_ras_get_error_type_name(err_type, err_type_name);
+		if (!mem_list) {
+			/* memory_list is not supported */
+			dev_info(adev->dev,
+				 "%ld %s hardware errors detected in %s, instance: %d, memory_id: %d\n",
+				 err_cnt, err_type_name,
+				 reg_list[i].block_name,
+				 instance, memory_id);
+		} else {
+			for (j = 0; j < mem_list_size; j++) {
+				if (memory_id == mem_list[j].memory_id) {
+					dev_info(adev->dev,
+						 "%ld %s hardware errors detected in %s, instance: %d, memory block: %s\n",
+						 err_cnt, err_type_name,
+						 reg_list[i].block_name,
+						 instance, mem_list[j].name);
+					break;
+				}
+			}
+		}
+	}
+}
+
+void amdgpu_ras_inst_reset_ras_error_count(struct amdgpu_device *adev,
+					   const struct amdgpu_ras_err_status_reg_entry *reg_list,
+					   uint32_t reg_list_size,
+					   uint32_t instance)
+{
+	uint32_t err_status_lo_offset, err_status_hi_offset;
+	uint32_t i;
+
+	for (i = 0; i < reg_list_size; i++) {
+		err_status_lo_offset =
+			AMDGPU_RAS_REG_ENTRY_OFFSET(reg_list[i].hwip, instance,
+						    reg_list[i].seg_lo, reg_list[i].reg_lo);
+		err_status_hi_offset =
+			AMDGPU_RAS_REG_ENTRY_OFFSET(reg_list[i].hwip, instance,
+						    reg_list[i].seg_hi, reg_list[i].reg_hi);
+		WREG32(err_status_lo_offset, 0);
+		WREG32(err_status_hi_offset, 0);
+	}
+}
+
+int amdgpu_ras_error_data_init(struct ras_err_data *err_data)
+{
+	memset(err_data, 0, sizeof(*err_data));
+
+	INIT_LIST_HEAD(&err_data->err_node_list);
+
+	return 0;
+}
+
+static void amdgpu_ras_error_node_release(struct ras_err_node *err_node)
+{
+	if (!err_node)
+		return;
+
+	list_del(&err_node->node);
+	kvfree(err_node);
+}
+
+void amdgpu_ras_error_data_fini(struct ras_err_data *err_data)
+{
+	struct ras_err_node *err_node, *tmp;
+
+	list_for_each_entry_safe(err_node, tmp, &err_data->err_node_list, node)
+		amdgpu_ras_error_node_release(err_node);
+}
+
+static struct ras_err_node *amdgpu_ras_error_find_node_by_id(struct ras_err_data *err_data,
+							     struct amdgpu_smuio_mcm_config_info *mcm_info)
+{
+	struct ras_err_node *err_node;
+	struct amdgpu_smuio_mcm_config_info *ref_id;
+
+	if (!err_data || !mcm_info)
+		return NULL;
+
+	for_each_ras_error(err_node, err_data) {
+		ref_id = &err_node->err_info.mcm_info;
+
+		if (mcm_info->socket_id == ref_id->socket_id &&
+		    mcm_info->die_id == ref_id->die_id)
+			return err_node;
+	}
+
+	return NULL;
+}
+
+static struct ras_err_node *amdgpu_ras_error_node_new(void)
+{
+	struct ras_err_node *err_node;
+
+	err_node = kvzalloc(sizeof(*err_node), GFP_KERNEL);
+	if (!err_node)
+		return NULL;
+
+	INIT_LIST_HEAD(&err_node->node);
+
+	return err_node;
+}
+
+static int ras_err_info_cmp(void *priv, const struct list_head *a, const struct list_head *b)
+{
+	struct ras_err_node *nodea = container_of(a, struct ras_err_node, node);
+	struct ras_err_node *nodeb = container_of(b, struct ras_err_node, node);
+	struct amdgpu_smuio_mcm_config_info *infoa = &nodea->err_info.mcm_info;
+	struct amdgpu_smuio_mcm_config_info *infob = &nodeb->err_info.mcm_info;
+
+	if (unlikely(infoa->socket_id != infob->socket_id))
+		return infoa->socket_id - infob->socket_id;
+	else
+		return infoa->die_id - infob->die_id;
+
+	return 0;
+}
+
+static struct ras_err_info *amdgpu_ras_error_get_info(struct ras_err_data *err_data,
+				struct amdgpu_smuio_mcm_config_info *mcm_info)
+{
+	struct ras_err_node *err_node;
+
+	err_node = amdgpu_ras_error_find_node_by_id(err_data, mcm_info);
+	if (err_node)
+		return &err_node->err_info;
+
+	err_node = amdgpu_ras_error_node_new();
+	if (!err_node)
+		return NULL;
+
+	memcpy(&err_node->err_info.mcm_info, mcm_info, sizeof(*mcm_info));
+
+	err_data->err_list_count++;
+	list_add_tail(&err_node->node, &err_data->err_node_list);
+	list_sort(NULL, &err_data->err_node_list, ras_err_info_cmp);
+
+	return &err_node->err_info;
+}
+
+int amdgpu_ras_error_statistic_ue_count(struct ras_err_data *err_data,
+					struct amdgpu_smuio_mcm_config_info *mcm_info,
+					u64 count)
+{
+	struct ras_err_info *err_info;
+
+	if (!err_data || !mcm_info)
+		return -EINVAL;
+
+	if (!count)
+		return 0;
+
+	err_info = amdgpu_ras_error_get_info(err_data, mcm_info);
+	if (!err_info)
+		return -EINVAL;
+
+	err_info->ue_count += count;
+	err_data->ue_count += count;
+
+	return 0;
+}
+
+int amdgpu_ras_error_statistic_ce_count(struct ras_err_data *err_data,
+					struct amdgpu_smuio_mcm_config_info *mcm_info,
+					u64 count)
+{
+	struct ras_err_info *err_info;
+
+	if (!err_data || !mcm_info)
+		return -EINVAL;
+
+	if (!count)
+		return 0;
+
+	err_info = amdgpu_ras_error_get_info(err_data, mcm_info);
+	if (!err_info)
+		return -EINVAL;
+
+	err_info->ce_count += count;
+	err_data->ce_count += count;
+
+	return 0;
+}
+
+int amdgpu_ras_error_statistic_de_count(struct ras_err_data *err_data,
+					struct amdgpu_smuio_mcm_config_info *mcm_info,
+					u64 count)
+{
+	struct ras_err_info *err_info;
+
+	if (!err_data || !mcm_info)
+		return -EINVAL;
+
+	if (!count)
+		return 0;
+
+	err_info = amdgpu_ras_error_get_info(err_data, mcm_info);
+	if (!err_info)
+		return -EINVAL;
+
+	err_info->de_count += count;
+	err_data->de_count += count;
+
+	return 0;
+}
+
+#define mmMP0_SMN_C2PMSG_92	0x1609C
+#define mmMP0_SMN_C2PMSG_126	0x160BE
+static void amdgpu_ras_boot_time_error_reporting(struct amdgpu_device *adev,
+						 u32 instance)
+{
+	u32 socket_id, aid_id, hbm_id;
+	u32 fw_status;
+	u32 boot_error;
+	u64 reg_addr;
+
+	/* The pattern for smn addressing in other SOC could be different from
+	 * the one for aqua_vanjaram. We should revisit the code if the pattern
+	 * is changed. In such case, replace the aqua_vanjaram implementation
+	 * with more common helper */
+	reg_addr = (mmMP0_SMN_C2PMSG_92 << 2) +
+		   aqua_vanjaram_encode_ext_smn_addressing(instance);
+	fw_status = amdgpu_device_indirect_rreg_ext(adev, reg_addr);
+
+	reg_addr = (mmMP0_SMN_C2PMSG_126 << 2) +
+		   aqua_vanjaram_encode_ext_smn_addressing(instance);
+	boot_error = amdgpu_device_indirect_rreg_ext(adev, reg_addr);
+
+	socket_id = AMDGPU_RAS_GPU_ERR_SOCKET_ID(boot_error);
+	aid_id = AMDGPU_RAS_GPU_ERR_AID_ID(boot_error);
+	hbm_id = ((1 == AMDGPU_RAS_GPU_ERR_HBM_ID(boot_error)) ? 0 : 1);
+
+	if (AMDGPU_RAS_GPU_ERR_MEM_TRAINING(boot_error))
+		dev_info(adev->dev,
+			 "socket: %d, aid: %d, hbm: %d, fw_status: 0x%x, memory training failed\n",
+			 socket_id, aid_id, hbm_id, fw_status);
+
+	if (AMDGPU_RAS_GPU_ERR_FW_LOAD(boot_error))
+		dev_info(adev->dev,
+			 "socket: %d, aid: %d, fw_status: 0x%x, firmware load failed at boot time\n",
+			 socket_id, aid_id, fw_status);
+
+	if (AMDGPU_RAS_GPU_ERR_WAFL_LINK_TRAINING(boot_error))
+		dev_info(adev->dev,
+			 "socket: %d, aid: %d, fw_status: 0x%x, wafl link training failed\n",
+			 socket_id, aid_id, fw_status);
+
+	if (AMDGPU_RAS_GPU_ERR_XGMI_LINK_TRAINING(boot_error))
+		dev_info(adev->dev,
+			 "socket: %d, aid: %d, fw_status: 0x%x, xgmi link training failed\n",
+			 socket_id, aid_id, fw_status);
+
+	if (AMDGPU_RAS_GPU_ERR_USR_CP_LINK_TRAINING(boot_error))
+		dev_info(adev->dev,
+			 "socket: %d, aid: %d, fw_status: 0x%x, usr cp link training failed\n",
+			 socket_id, aid_id, fw_status);
+
+	if (AMDGPU_RAS_GPU_ERR_USR_DP_LINK_TRAINING(boot_error))
+		dev_info(adev->dev,
+			 "socket: %d, aid: %d, fw_status: 0x%x, usr dp link training failed\n",
+			 socket_id, aid_id, fw_status);
+
+	if (AMDGPU_RAS_GPU_ERR_HBM_MEM_TEST(boot_error))
+		dev_info(adev->dev,
+			 "socket: %d, aid: %d, hbm: %d, fw_status: 0x%x, hbm memory test failed\n",
+			 socket_id, aid_id, hbm_id, fw_status);
+
+	if (AMDGPU_RAS_GPU_ERR_HBM_BIST_TEST(boot_error))
+		dev_info(adev->dev,
+			 "socket: %d, aid: %d, hbm: %d, fw_status: 0x%x, hbm bist test failed\n",
+			 socket_id, aid_id, hbm_id, fw_status);
+
+	if (AMDGPU_RAS_GPU_ERR_DATA_ABORT(boot_error))
+		dev_info(adev->dev,
+			 "socket: %d, aid: %d, fw_status: 0x%x, data abort exception\n",
+			 socket_id, aid_id, fw_status);
+
+	if (AMDGPU_RAS_GPU_ERR_GENERIC(boot_error))
+		dev_info(adev->dev,
+			 "socket: %d, aid: %d, fw_status: 0x%x, Boot Controller Generic Error\n",
+			 socket_id, aid_id, fw_status);
+}
+
+static bool amdgpu_ras_boot_error_detected(struct amdgpu_device *adev,
+					   u32 instance)
+{
+	u64 reg_addr;
+	u32 reg_data;
+	int retry_loop;
+
+	reg_addr = (mmMP0_SMN_C2PMSG_92 << 2) +
+		   aqua_vanjaram_encode_ext_smn_addressing(instance);
+
+	for (retry_loop = 0; retry_loop < AMDGPU_RAS_BOOT_STATUS_POLLING_LIMIT; retry_loop++) {
+		reg_data = amdgpu_device_indirect_rreg_ext(adev, reg_addr);
+		if ((reg_data & AMDGPU_RAS_BOOT_STATUS_MASK) == AMDGPU_RAS_BOOT_STEADY_STATUS)
+			return false;
+		else
+			msleep(1);
+	}
+
+	return true;
+}
+
+void amdgpu_ras_query_boot_status(struct amdgpu_device *adev, u32 num_instances)
+{
+	u32 i;
+
+	for (i = 0; i < num_instances; i++) {
+		if (amdgpu_ras_boot_error_detected(adev, i))
+			amdgpu_ras_boot_time_error_reporting(adev, i);
+	}
+}
+
+int amdgpu_ras_reserve_page(struct amdgpu_device *adev, uint64_t pfn)
+{
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+	struct amdgpu_vram_mgr *mgr = &adev->mman.vram_mgr;
+	uint64_t start = pfn << AMDGPU_GPU_PAGE_SHIFT;
+	int ret = 0;
+
+	if (amdgpu_ras_check_critical_address(adev, start))
+		return 0;
+
+	mutex_lock(&con->page_rsv_lock);
+	ret = amdgpu_vram_mgr_query_page_status(mgr, start);
+	if (ret == -ENOENT)
+		ret = amdgpu_vram_mgr_reserve_range(mgr, start, AMDGPU_GPU_PAGE_SIZE);
+	mutex_unlock(&con->page_rsv_lock);
+
+	return ret;
+}
+
+void amdgpu_ras_event_log_print(struct amdgpu_device *adev, u64 event_id,
+				const char *fmt, ...)
+{
+	struct va_format vaf;
+	va_list args;
+
+	va_start(args, fmt);
+	vaf.fmt = fmt;
+	vaf.va = &args;
+
+	if (RAS_EVENT_ID_IS_VALID(event_id))
+		dev_printk(KERN_INFO, adev->dev, "{%llu}%pV", event_id, &vaf);
+	else
+		dev_printk(KERN_INFO, adev->dev, "%pV", &vaf);
+
+	va_end(args);
+}
+
+bool amdgpu_ras_is_rma(struct amdgpu_device *adev)
+{
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+
+	if (amdgpu_uniras_enabled(adev))
+		return amdgpu_ras_mgr_is_rma(adev);
+
+	if (!con)
+		return false;
+
+	return con->is_rma;
+}
+
+int amdgpu_ras_add_critical_region(struct amdgpu_device *adev,
+			struct amdgpu_bo *bo)
+{
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+	struct amdgpu_vram_mgr_resource *vres;
+	struct ras_critical_region *region;
+	struct drm_buddy_block *block;
+	int ret = 0;
+
+	if (!bo || !bo->tbo.resource)
+		return -EINVAL;
+
+	vres = to_amdgpu_vram_mgr_resource(bo->tbo.resource);
+
+	mutex_lock(&con->critical_region_lock);
+
+	/* Check if the bo had been recorded */
+	list_for_each_entry(region, &con->critical_region_head, node)
+		if (region->bo == bo)
+			goto out;
+
+	/* Record new critical amdgpu bo */
+	list_for_each_entry(block, &vres->blocks, link) {
+		region = kzalloc(sizeof(*region), GFP_KERNEL);
+		if (!region) {
+			ret = -ENOMEM;
+			goto out;
+		}
+		region->bo = bo;
+		region->start = amdgpu_vram_mgr_block_start(block);
+		region->size = amdgpu_vram_mgr_block_size(block);
+		list_add_tail(&region->node, &con->critical_region_head);
+	}
+
+out:
+	mutex_unlock(&con->critical_region_lock);
+
+	return ret;
+}
+
+static void amdgpu_ras_critical_region_init(struct amdgpu_device *adev)
+{
+	amdgpu_ras_add_critical_region(adev, adev->mman.fw_reserved_memory);
+}
+
+static void amdgpu_ras_critical_region_fini(struct amdgpu_device *adev)
+{
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+	struct ras_critical_region *region, *tmp;
+
+	mutex_lock(&con->critical_region_lock);
+	list_for_each_entry_safe(region, tmp, &con->critical_region_head, node) {
+		list_del(&region->node);
+		kfree(region);
+	}
+	mutex_unlock(&con->critical_region_lock);
+}
+
+bool amdgpu_ras_check_critical_address(struct amdgpu_device *adev, uint64_t addr)
+{
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+	struct ras_critical_region *region;
+	bool ret = false;
+
+	mutex_lock(&con->critical_region_lock);
+	list_for_each_entry(region, &con->critical_region_head, node) {
+		if ((region->start <= addr) &&
+		    (addr < (region->start + region->size))) {
+			ret = true;
+			break;
+		}
+	}
+	mutex_unlock(&con->critical_region_lock);
+
+	return ret;
+}
+
+void amdgpu_ras_pre_reset(struct amdgpu_device *adev,
+					  struct list_head *device_list)
+{
+	struct amdgpu_device *tmp_adev = NULL;
+
+	list_for_each_entry(tmp_adev, device_list, reset_list) {
+		if (amdgpu_uniras_enabled(tmp_adev))
+			amdgpu_ras_mgr_pre_reset(tmp_adev);
+	}
+}
+
+void amdgpu_ras_post_reset(struct amdgpu_device *adev,
+					  struct list_head *device_list)
+{
+	struct amdgpu_device *tmp_adev = NULL;
+
+	list_for_each_entry(tmp_adev, device_list, reset_list) {
+		if (amdgpu_uniras_enabled(tmp_adev))
+			amdgpu_ras_mgr_post_reset(tmp_adev);
+	}
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h
index e36f4de9fa55..ff44190d7d98 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h
@@ -26,12 +26,57 @@
 
 #include <linux/debugfs.h>
 #include <linux/list.h>
-#include "amdgpu.h"
-#include "amdgpu_psp.h"
+#include <linux/kfifo.h>
+#include <linux/radix-tree.h>
 #include "ta_ras_if.h"
 #include "amdgpu_ras_eeprom.h"
+#include "amdgpu_smuio.h"
+#include "amdgpu_aca.h"
+
+struct amdgpu_iv_entry;
+
+#define AMDGPU_RAS_GPU_ERR_MEM_TRAINING(x)		AMDGPU_GET_REG_FIELD(x, 0, 0)
+#define AMDGPU_RAS_GPU_ERR_FW_LOAD(x)			AMDGPU_GET_REG_FIELD(x, 1, 1)
+#define AMDGPU_RAS_GPU_ERR_WAFL_LINK_TRAINING(x)	AMDGPU_GET_REG_FIELD(x, 2, 2)
+#define AMDGPU_RAS_GPU_ERR_XGMI_LINK_TRAINING(x)	AMDGPU_GET_REG_FIELD(x, 3, 3)
+#define AMDGPU_RAS_GPU_ERR_USR_CP_LINK_TRAINING(x)	AMDGPU_GET_REG_FIELD(x, 4, 4)
+#define AMDGPU_RAS_GPU_ERR_USR_DP_LINK_TRAINING(x)	AMDGPU_GET_REG_FIELD(x, 5, 5)
+#define AMDGPU_RAS_GPU_ERR_HBM_MEM_TEST(x)		AMDGPU_GET_REG_FIELD(x, 6, 6)
+#define AMDGPU_RAS_GPU_ERR_HBM_BIST_TEST(x)		AMDGPU_GET_REG_FIELD(x, 7, 7)
+#define AMDGPU_RAS_GPU_ERR_SOCKET_ID(x)			AMDGPU_GET_REG_FIELD(x, 10, 8)
+#define AMDGPU_RAS_GPU_ERR_AID_ID(x)			AMDGPU_GET_REG_FIELD(x, 12, 11)
+#define AMDGPU_RAS_GPU_ERR_HBM_ID(x)			AMDGPU_GET_REG_FIELD(x, 14, 13)
+#define AMDGPU_RAS_GPU_ERR_DATA_ABORT(x)		AMDGPU_GET_REG_FIELD(x, 29, 29)
+#define AMDGPU_RAS_GPU_ERR_GENERIC(x)			AMDGPU_GET_REG_FIELD(x, 30, 30)
+
+#define AMDGPU_RAS_BOOT_STATUS_POLLING_LIMIT	100
+#define AMDGPU_RAS_BOOT_STEADY_STATUS		0xBA
+#define AMDGPU_RAS_BOOT_STATUS_MASK		0xFF
 
 #define AMDGPU_RAS_FLAG_INIT_BY_VBIOS		(0x1 << 0)
+/* position of instance value in sub_block_index of
+ * ta_ras_trigger_error_input, the sub block uses lower 12 bits
+ */
+#define AMDGPU_RAS_INST_MASK 0xfffff000
+#define AMDGPU_RAS_INST_SHIFT 0xc
+
+#define AMDGPU_RAS_FEATURES_SOCKETID_SHIFT 29
+#define AMDGPU_RAS_FEATURES_SOCKETID_MASK 0xe0000000
+
+/* Reserve 8 physical dram row for possible retirement.
+ * In worst cases, it will lose 8 * 2MB memory in vram domain */
+#define AMDGPU_RAS_RESERVED_VRAM_SIZE_DEFAULT	(16ULL << 20)
+/* The high three bits indicates socketid */
+#define AMDGPU_RAS_GET_FEATURES(val)  ((val) & ~AMDGPU_RAS_FEATURES_SOCKETID_MASK)
+
+#define RAS_EVENT_INVALID_ID		(BIT_ULL(63))
+#define RAS_EVENT_ID_IS_VALID(x)	(!((x) & BIT_ULL(63)))
+
+#define RAS_EVENT_LOG(adev, id, fmt, ...)	\
+	amdgpu_ras_event_log_print((adev), (id), (fmt), ##__VA_ARGS__)
+
+#define amdgpu_ras_mark_ras_event(adev, type)	\
+	(amdgpu_ras_mark_ras_event_caller((adev), (type), __builtin_return_address(0)))
 
 enum amdgpu_ras_block {
 	AMDGPU_RAS_BLOCK__UMC = 0,
@@ -49,8 +94,14 @@ enum amdgpu_ras_block {
 	AMDGPU_RAS_BLOCK__MP1,
 	AMDGPU_RAS_BLOCK__FUSE,
 	AMDGPU_RAS_BLOCK__MCA,
-
-	AMDGPU_RAS_BLOCK__LAST
+	AMDGPU_RAS_BLOCK__VCN,
+	AMDGPU_RAS_BLOCK__JPEG,
+	AMDGPU_RAS_BLOCK__IH,
+	AMDGPU_RAS_BLOCK__MPIO,
+	AMDGPU_RAS_BLOCK__MMSCH,
+
+	AMDGPU_RAS_BLOCK__LAST,
+	AMDGPU_RAS_BLOCK__ANY = -1
 };
 
 enum amdgpu_ras_mca_block {
@@ -312,6 +363,53 @@ enum amdgpu_ras_ret {
 	AMDGPU_RAS_PT,
 };
 
+enum amdgpu_ras_error_query_mode {
+	AMDGPU_RAS_INVALID_ERROR_QUERY		= 0,
+	AMDGPU_RAS_DIRECT_ERROR_QUERY		= 1,
+	AMDGPU_RAS_FIRMWARE_ERROR_QUERY		= 2,
+	AMDGPU_RAS_VIRT_ERROR_COUNT_QUERY	= 3,
+};
+
+/* ras error status reisger fields */
+#define ERR_STATUS_LO__ERR_STATUS_VALID_FLAG__SHIFT	0x0
+#define ERR_STATUS_LO__ERR_STATUS_VALID_FLAG_MASK	0x00000001L
+#define ERR_STATUS_LO__MEMORY_ID__SHIFT			0x18
+#define ERR_STATUS_LO__MEMORY_ID_MASK			0xFF000000L
+#define ERR_STATUS_HI__ERR_INFO_VALID_FLAG__SHIFT	0x2
+#define ERR_STATUS_HI__ERR_INFO_VALID_FLAG_MASK		0x00000004L
+#define ERR_STATUS__ERR_CNT__SHIFT			0x17
+#define ERR_STATUS__ERR_CNT_MASK			0x03800000L
+
+#define AMDGPU_RAS_REG_ENTRY(ip, inst, reg_lo, reg_hi) \
+	ip##_HWIP, inst, reg_lo##_BASE_IDX, reg_lo, reg_hi##_BASE_IDX, reg_hi
+
+#define AMDGPU_RAS_REG_ENTRY_OFFSET(hwip, ip_inst, segment, reg) \
+	(adev->reg_offset[hwip][ip_inst][segment] + (reg))
+
+#define AMDGPU_RAS_ERR_INFO_VALID	(1 << 0)
+#define AMDGPU_RAS_ERR_STATUS_VALID	(1 << 1)
+#define AMDGPU_RAS_ERR_ADDRESS_VALID	(1 << 2)
+
+#define AMDGPU_RAS_GPU_RESET_MODE2_RESET  (0x1 << 0)
+#define AMDGPU_RAS_GPU_RESET_MODE1_RESET  (0x1 << 1)
+
+struct amdgpu_ras_err_status_reg_entry {
+	uint32_t hwip;
+	uint32_t ip_inst;
+	uint32_t seg_lo;
+	uint32_t reg_lo;
+	uint32_t seg_hi;
+	uint32_t reg_hi;
+	uint32_t reg_inst;
+	uint32_t flags;
+	const char *block_name;
+};
+
+struct amdgpu_ras_memory_id_entry {
+	uint32_t memory_id;
+	const char *name;
+};
+
 struct ras_common_if {
 	enum amdgpu_ras_block block;
 	enum amdgpu_ras_error_type type;
@@ -319,13 +417,130 @@ struct ras_common_if {
 	char name[32];
 };
 
+#define MAX_UMC_CHANNEL_NUM 32
+
+struct ecc_info_per_ch {
+	uint16_t ce_count_lo_chip;
+	uint16_t ce_count_hi_chip;
+	uint64_t mca_umc_status;
+	uint64_t mca_umc_addr;
+	uint64_t mca_ceumc_addr;
+};
+
+struct umc_ecc_info {
+	struct ecc_info_per_ch ecc[MAX_UMC_CHANNEL_NUM];
+
+	/* Determine smu ecctable whether support
+	 * record correctable error address
+	 */
+	int record_ce_addr_supported;
+};
+
+enum ras_event_type {
+	RAS_EVENT_TYPE_INVALID = 0,
+	RAS_EVENT_TYPE_FATAL,
+	RAS_EVENT_TYPE_POISON_CREATION,
+	RAS_EVENT_TYPE_POISON_CONSUMPTION,
+	RAS_EVENT_TYPE_COUNT,
+};
+
+struct ras_event_state {
+	u64 last_seqno;
+	atomic64_t count;
+};
+
+struct ras_event_manager {
+	atomic64_t seqno;
+	struct ras_event_state event_state[RAS_EVENT_TYPE_COUNT];
+};
+
+struct ras_event_id {
+	enum ras_event_type type;
+	u64 event_id;
+};
+
+struct ras_query_context {
+	struct ras_event_id evid;
+};
+
+typedef int (*pasid_notify)(struct amdgpu_device *adev,
+		uint16_t pasid, void *data);
+
+struct ras_poison_msg {
+	enum amdgpu_ras_block block;
+	uint16_t pasid;
+	uint32_t reset;
+	pasid_notify pasid_fn;
+	void *data;
+};
+
+struct ras_err_pages {
+	uint32_t count;
+	uint64_t *pfn;
+};
+
+struct ras_ecc_err {
+	uint64_t status;
+	uint64_t ipid;
+	uint64_t addr;
+	uint64_t pa_pfn;
+	/* save global channel index across all UMC instances */
+	uint32_t channel_idx;
+	struct ras_err_pages err_pages;
+};
+
+struct ras_ecc_log_info {
+	struct mutex lock;
+	struct radix_tree_root de_page_tree;
+	uint64_t de_queried_count;
+	uint64_t consumption_q_count;
+};
+
+struct ras_critical_region {
+	struct list_head node;
+	struct amdgpu_bo *bo;
+	uint64_t start;
+	uint64_t size;
+};
+
+struct ras_eeprom_table_version {
+	uint32_t minor    : 16;
+	uint32_t major    : 16;
+};
+
+struct ras_eeprom_smu_funcs {
+	int (*get_ras_table_version)(struct amdgpu_device *adev,
+							uint32_t *table_version);
+	int (*get_badpage_count)(struct amdgpu_device *adev, uint32_t *count, uint32_t timeout);
+	int (*get_badpage_mca_addr)(struct amdgpu_device *adev, uint16_t index, uint64_t *mca_addr);
+	int (*set_timestamp)(struct amdgpu_device *adev, uint64_t timestamp);
+	int (*get_timestamp)(struct amdgpu_device *adev,
+							uint16_t index, uint64_t *timestamp);
+	int (*get_badpage_ipid)(struct amdgpu_device *adev, uint16_t index, uint64_t *ipid);
+	int (*erase_ras_table)(struct amdgpu_device *adev, uint32_t *result);
+};
+
+enum ras_smu_feature_flags {
+	RAS_SMU_FEATURE_BIT__RAS_EEPROM = BIT_ULL(0),
+};
+
+struct ras_smu_drv {
+	const struct ras_eeprom_smu_funcs *smu_eeprom_funcs;
+	void (*ras_smu_feature_flags)(struct amdgpu_device *adev, uint64_t *flags);
+};
+
 struct amdgpu_ras {
+	void *ras_mgr;
 	/* ras infrastructure */
 	/* for ras itself. */
 	uint32_t features;
+	uint32_t schema;
 	struct list_head head;
 	/* sysfs */
 	struct device_attribute features_attr;
+	struct device_attribute version_attr;
+	struct device_attribute schema_attr;
+	struct device_attribute event_state_attr;
 	struct bin_attribute badpages_attr;
 	struct dentry *de_ras_eeprom_table;
 	/* block array */
@@ -334,6 +549,7 @@ struct amdgpu_ras {
 	/* gpu recovery */
 	struct work_struct recovery_work;
 	atomic_t in_recovery;
+	atomic_t rma_in_recovery;
 	struct amdgpu_device *adev;
 	/* error handler data */
 	struct ras_err_handler_data *eh_data;
@@ -358,25 +574,92 @@ struct amdgpu_ras {
 	struct delayed_work ras_counte_delay_work;
 	atomic_t ras_ue_count;
 	atomic_t ras_ce_count;
+
+	/* record umc error info queried from smu */
+	struct umc_ecc_info umc_ecc;
+
+	/* Indicates smu whether need update bad channel info */
+	bool update_channel_flag;
+	/* Record status of smu mca debug mode */
+	bool is_aca_debug_mode;
+	bool is_rma;
+
+	/* Record special requirements of gpu reset caller */
+	uint32_t  gpu_reset_flags;
+
+	struct task_struct *page_retirement_thread;
+	wait_queue_head_t page_retirement_wq;
+	struct mutex page_retirement_lock;
+	atomic_t page_retirement_req_cnt;
+	atomic_t poison_creation_count;
+	atomic_t poison_consumption_count;
+	struct mutex page_rsv_lock;
+	DECLARE_KFIFO(poison_fifo, struct ras_poison_msg, 128);
+	struct ras_ecc_log_info  umc_ecc_log;
+	struct delayed_work page_retirement_dwork;
+
+	/* ras errors detected */
+	unsigned long ras_err_state;
+
+	/* RAS event manager */
+	struct ras_event_manager __event_mgr;
+	struct ras_event_manager *event_mgr;
+
+	uint64_t reserved_pages_in_bytes;
+
+	pid_t init_task_pid;
+	char init_task_comm[TASK_COMM_LEN];
+
+	int bad_page_num;
+
+	struct list_head critical_region_head;
+	struct mutex critical_region_lock;
+
+	/* Protect poison injection */
+	struct mutex poison_lock;
+
+	/* Disable/Enable uniras switch */
+	bool uniras_enabled;
+	const struct ras_smu_drv *ras_smu_drv;
 };
 
 struct ras_fs_data {
-	char sysfs_name[32];
+	char sysfs_name[48];
 	char debugfs_name[32];
 };
 
+struct ras_err_info {
+	struct amdgpu_smuio_mcm_config_info mcm_info;
+	u64 ce_count;
+	u64 ue_count;
+	u64 de_count;
+};
+
+struct ras_err_node {
+	struct list_head node;
+	struct ras_err_info err_info;
+};
+
 struct ras_err_data {
 	unsigned long ue_count;
 	unsigned long ce_count;
+	unsigned long de_count;
 	unsigned long err_addr_cnt;
 	struct eeprom_table_record *err_addr;
+	unsigned long err_addr_len;
+	u32 err_list_count;
+	struct list_head err_node_list;
 };
 
+#define for_each_ras_error(err_node, err_data) \
+	list_for_each_entry(err_node, &(err_data)->err_node_list, node)
+
 struct ras_err_handler_data {
 	/* point to bad page records array */
 	struct eeprom_table_record *bps;
 	/* the count of entries */
 	int count;
+	int count_saved;
 	/* the space can place new entries */
 	int space_left;
 };
@@ -419,6 +702,8 @@ struct ras_manager {
 	struct ras_ih_data ih_data;
 
 	struct ras_err_data err_data;
+
+	struct aca_handle aca_handle;
 };
 
 struct ras_badpage {
@@ -438,12 +723,14 @@ struct ras_query_if {
 	struct ras_common_if head;
 	unsigned long ue_count;
 	unsigned long ce_count;
+	unsigned long de_count;
 };
 
 struct ras_inject_if {
 	struct ras_common_if head;
 	uint64_t address;
 	uint64_t value;
+	uint32_t instance_mask;
 };
 
 struct ras_cure_if {
@@ -468,6 +755,30 @@ struct ras_debug_if {
 	};
 	int op;
 };
+
+struct amdgpu_ras_block_object {
+	struct ras_common_if  ras_comm;
+
+	int (*ras_block_match)(struct amdgpu_ras_block_object *block_obj,
+				enum amdgpu_ras_block block, uint32_t sub_block_index);
+	int (*ras_late_init)(struct amdgpu_device *adev, struct ras_common_if *ras_block);
+	void (*ras_fini)(struct amdgpu_device *adev, struct ras_common_if *ras_block);
+	ras_ih_cb ras_cb;
+	const struct amdgpu_ras_block_hw_ops *hw_ops;
+};
+
+struct amdgpu_ras_block_hw_ops {
+	int  (*ras_error_inject)(struct amdgpu_device *adev,
+			void *inject_if, uint32_t instance_mask);
+	void (*query_ras_error_count)(struct amdgpu_device *adev, void *ras_error_status);
+	void (*query_ras_error_status)(struct amdgpu_device *adev);
+	void (*query_ras_error_address)(struct amdgpu_device *adev, void *ras_error_status);
+	void (*reset_ras_error_count)(struct amdgpu_device *adev);
+	void (*reset_ras_error_status)(struct amdgpu_device *adev);
+	bool (*query_poison_status)(struct amdgpu_device *adev);
+	bool (*handle_poison_consumption)(struct amdgpu_device *adev);
+};
+
 /* work flow
  * vbios
  * 1: ras feature enable (enabled by default)
@@ -482,43 +793,23 @@ struct ras_debug_if {
  * 8: feature disable
  */
 
-#define amdgpu_ras_get_context(adev)		((adev)->psp.ras_context.ras)
-#define amdgpu_ras_set_context(adev, ras_con)	((adev)->psp.ras_context.ras = (ras_con))
-
-/* check if ras is supported on block, say, sdma, gfx */
-static inline int amdgpu_ras_is_supported(struct amdgpu_device *adev,
-		unsigned int block)
-{
-	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
-
-	if (block >= AMDGPU_RAS_BLOCK_COUNT)
-		return 0;
-	return ras && (adev->ras_enabled & (1 << block));
-}
-
-int amdgpu_ras_recovery_init(struct amdgpu_device *adev);
+int amdgpu_ras_init_badpage_info(struct amdgpu_device *adev);
+int amdgpu_ras_recovery_init(struct amdgpu_device *adev, bool init_bp_info);
 
 void amdgpu_ras_resume(struct amdgpu_device *adev);
 void amdgpu_ras_suspend(struct amdgpu_device *adev);
 
 int amdgpu_ras_query_error_count(struct amdgpu_device *adev,
 				 unsigned long *ce_count,
-				 unsigned long *ue_count);
+				 unsigned long *ue_count,
+				 struct ras_query_if *query_info);
 
 /* error handling functions */
 int amdgpu_ras_add_bad_pages(struct amdgpu_device *adev,
-		struct eeprom_table_record *bps, int pages);
-
-int amdgpu_ras_save_bad_pages(struct amdgpu_device *adev);
-
-static inline int amdgpu_ras_reset_gpu(struct amdgpu_device *adev)
-{
-	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
+		struct eeprom_table_record *bps, int pages, bool from_rom);
 
-	if (atomic_cmpxchg(&ras->in_recovery, 0, 1) == 0)
-		schedule_work(&ras->recovery_work);
-	return 0;
-}
+int amdgpu_ras_save_bad_pages(struct amdgpu_device *adev,
+		unsigned long *new_cnt);
 
 static inline enum ta_ras_block
 amdgpu_ras_block_to_ta(enum amdgpu_ras_block block) {
@@ -553,6 +844,16 @@ amdgpu_ras_block_to_ta(enum amdgpu_ras_block block) {
 		return TA_RAS_BLOCK__FUSE;
 	case AMDGPU_RAS_BLOCK__MCA:
 		return TA_RAS_BLOCK__MCA;
+	case AMDGPU_RAS_BLOCK__VCN:
+		return TA_RAS_BLOCK__VCN;
+	case AMDGPU_RAS_BLOCK__JPEG:
+		return TA_RAS_BLOCK__JPEG;
+	case AMDGPU_RAS_BLOCK__IH:
+		return TA_RAS_BLOCK__IH;
+	case AMDGPU_RAS_BLOCK__MPIO:
+		return TA_RAS_BLOCK__MPIO;
+	case AMDGPU_RAS_BLOCK__MMSCH:
+		return TA_RAS_BLOCK__MMSCH;
 	default:
 		WARN_ONCE(1, "RAS ERROR: unexpected block id %d\n", block);
 		return TA_RAS_BLOCK__UMC;
@@ -580,15 +881,15 @@ amdgpu_ras_error_to_ta(enum amdgpu_ras_error_type error) {
 
 /* called in ip_init and ip_fini */
 int amdgpu_ras_init(struct amdgpu_device *adev);
+int amdgpu_ras_late_init(struct amdgpu_device *adev);
 int amdgpu_ras_fini(struct amdgpu_device *adev);
 int amdgpu_ras_pre_fini(struct amdgpu_device *adev);
-int amdgpu_ras_late_init(struct amdgpu_device *adev,
-			 struct ras_common_if *ras_block,
-			 struct ras_fs_if *fs_info,
-			 struct ras_ih_if *ih_info);
-void amdgpu_ras_late_fini(struct amdgpu_device *adev,
-			  struct ras_common_if *ras_block,
-			  struct ras_ih_if *ih_info);
+
+int amdgpu_ras_block_late_init(struct amdgpu_device *adev,
+			struct ras_common_if *ras_block);
+
+void amdgpu_ras_block_late_fini(struct amdgpu_device *adev,
+			  struct ras_common_if *ras_block);
 
 int amdgpu_ras_feature_enable(struct amdgpu_device *adev,
 		struct ras_common_if *head, bool enable);
@@ -597,7 +898,7 @@ int amdgpu_ras_feature_enable_on_boot(struct amdgpu_device *adev,
 		struct ras_common_if *head, bool enable);
 
 int amdgpu_ras_sysfs_create(struct amdgpu_device *adev,
-		struct ras_fs_if *head);
+		struct ras_common_if *head);
 
 int amdgpu_ras_sysfs_remove(struct amdgpu_device *adev,
 		struct ras_common_if *head);
@@ -607,6 +908,8 @@ void amdgpu_ras_debugfs_create_all(struct amdgpu_device *adev);
 int amdgpu_ras_query_error_status(struct amdgpu_device *adev,
 		struct ras_query_if *info);
 
+int amdgpu_ras_reset_error_count(struct amdgpu_device *adev,
+		enum amdgpu_ras_block block);
 int amdgpu_ras_reset_error_status(struct amdgpu_device *adev,
 		enum amdgpu_ras_block block);
 
@@ -614,10 +917,10 @@ int amdgpu_ras_error_inject(struct amdgpu_device *adev,
 		struct ras_inject_if *info);
 
 int amdgpu_ras_interrupt_add_handler(struct amdgpu_device *adev,
-		struct ras_ih_if *info);
+		struct ras_common_if *head);
 
 int amdgpu_ras_interrupt_remove_handler(struct amdgpu_device *adev,
-		struct ras_ih_if *info);
+		struct ras_common_if *head);
 
 int amdgpu_ras_interrupt_dispatch(struct amdgpu_device *adev,
 		struct ras_dispatch_if *info);
@@ -637,7 +940,7 @@ static inline void amdgpu_ras_intr_cleared(void)
 	atomic_set(&amdgpu_ras_in_intr, 0);
 }
 
-void amdgpu_ras_global_ras_isr(struct amdgpu_device *adev);
+int amdgpu_ras_global_ras_isr(struct amdgpu_device *adev);
 
 void amdgpu_ras_set_error_query_ready(struct amdgpu_device *adev, bool ready);
 
@@ -651,4 +954,94 @@ const char *get_ras_block_str(struct ras_common_if *ras_block);
 
 bool amdgpu_ras_is_poison_mode_supported(struct amdgpu_device *adev);
 
+int amdgpu_ras_is_supported(struct amdgpu_device *adev, unsigned int block);
+
+int amdgpu_ras_reset_gpu(struct amdgpu_device *adev);
+
+struct amdgpu_ras* amdgpu_ras_get_context(struct amdgpu_device *adev);
+
+int amdgpu_ras_set_context(struct amdgpu_device *adev, struct amdgpu_ras *ras_con);
+
+int amdgpu_ras_set_mca_debug_mode(struct amdgpu_device *adev, bool enable);
+int amdgpu_ras_set_aca_debug_mode(struct amdgpu_device *adev, bool enable);
+bool amdgpu_ras_get_aca_debug_mode(struct amdgpu_device *adev);
+bool amdgpu_ras_get_error_query_mode(struct amdgpu_device *adev,
+				     unsigned int *mode);
+
+int amdgpu_ras_register_ras_block(struct amdgpu_device *adev,
+				struct amdgpu_ras_block_object *ras_block_obj);
+void amdgpu_ras_interrupt_fatal_error_handler(struct amdgpu_device *adev);
+void amdgpu_ras_get_error_type_name(uint32_t err_type, char *err_type_name);
+bool amdgpu_ras_inst_get_memory_id_field(struct amdgpu_device *adev,
+					 const struct amdgpu_ras_err_status_reg_entry *reg_entry,
+					 uint32_t instance,
+					 uint32_t *memory_id);
+bool amdgpu_ras_inst_get_err_cnt_field(struct amdgpu_device *adev,
+				       const struct amdgpu_ras_err_status_reg_entry *reg_entry,
+				       uint32_t instance,
+				       unsigned long *err_cnt);
+void amdgpu_ras_inst_query_ras_error_count(struct amdgpu_device *adev,
+					   const struct amdgpu_ras_err_status_reg_entry *reg_list,
+					   uint32_t reg_list_size,
+					   const struct amdgpu_ras_memory_id_entry *mem_list,
+					   uint32_t mem_list_size,
+					   uint32_t instance,
+					   uint32_t err_type,
+					   unsigned long *err_count);
+void amdgpu_ras_inst_reset_ras_error_count(struct amdgpu_device *adev,
+					   const struct amdgpu_ras_err_status_reg_entry *reg_list,
+					   uint32_t reg_list_size,
+					   uint32_t instance);
+
+int amdgpu_ras_error_data_init(struct ras_err_data *err_data);
+void amdgpu_ras_error_data_fini(struct ras_err_data *err_data);
+int amdgpu_ras_error_statistic_ce_count(struct ras_err_data *err_data,
+					struct amdgpu_smuio_mcm_config_info *mcm_info,
+					u64 count);
+int amdgpu_ras_error_statistic_ue_count(struct ras_err_data *err_data,
+					struct amdgpu_smuio_mcm_config_info *mcm_info,
+					u64 count);
+int amdgpu_ras_error_statistic_de_count(struct ras_err_data *err_data,
+					struct amdgpu_smuio_mcm_config_info *mcm_info,
+					u64 count);
+void amdgpu_ras_query_boot_status(struct amdgpu_device *adev, u32 num_instances);
+int amdgpu_ras_bind_aca(struct amdgpu_device *adev, enum amdgpu_ras_block blk,
+			       const struct aca_info *aca_info, void *data);
+int amdgpu_ras_unbind_aca(struct amdgpu_device *adev, enum amdgpu_ras_block blk);
+
+ssize_t amdgpu_ras_aca_sysfs_read(struct device *dev, struct device_attribute *attr,
+				  struct aca_handle *handle, char *buf, void *data);
+
+void amdgpu_ras_set_fed(struct amdgpu_device *adev, bool status);
+bool amdgpu_ras_get_fed_status(struct amdgpu_device *adev);
+void amdgpu_ras_set_err_poison(struct amdgpu_device *adev,
+			       enum amdgpu_ras_block block);
+void amdgpu_ras_clear_err_state(struct amdgpu_device *adev);
+bool amdgpu_ras_is_err_state(struct amdgpu_device *adev, int block);
+
+u64 amdgpu_ras_acquire_event_id(struct amdgpu_device *adev, enum ras_event_type type);
+int amdgpu_ras_mark_ras_event_caller(struct amdgpu_device *adev, enum ras_event_type type,
+				     const void *caller);
+
+int amdgpu_ras_reserve_page(struct amdgpu_device *adev, uint64_t pfn);
+
+int amdgpu_ras_add_critical_region(struct amdgpu_device *adev, struct amdgpu_bo *bo);
+bool amdgpu_ras_check_critical_address(struct amdgpu_device *adev, uint64_t addr);
+
+int amdgpu_ras_put_poison_req(struct amdgpu_device *adev,
+		enum amdgpu_ras_block block, uint16_t pasid,
+		pasid_notify pasid_fn, void *data, uint32_t reset);
+
+bool amdgpu_ras_in_recovery(struct amdgpu_device *adev);
+
+__printf(3, 4)
+void amdgpu_ras_event_log_print(struct amdgpu_device *adev, u64 event_id,
+				const char *fmt, ...);
+
+bool amdgpu_ras_is_rma(struct amdgpu_device *adev);
+
+void amdgpu_ras_pre_reset(struct amdgpu_device *adev,
+					  struct list_head *device_list);
+void amdgpu_ras_post_reset(struct amdgpu_device *adev,
+					  struct list_head *device_list);
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c
index 05117eda105b..64dd7a81bff5 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c
@@ -31,14 +31,35 @@
 #include <linux/debugfs.h>
 #include <linux/uaccess.h>
 
-#define EEPROM_I2C_MADDR_VEGA20         0x0
-#define EEPROM_I2C_MADDR_ARCTURUS       0x40000
-#define EEPROM_I2C_MADDR_ARCTURUS_D342  0x0
-#define EEPROM_I2C_MADDR_SIENNA_CICHLID 0x0
-#define EEPROM_I2C_MADDR_ALDEBARAN      0x0
+#include "amdgpu_reset.h"
+#include "amdgpu_ras_mgr.h"
+
+/* These are memory addresses as would be seen by one or more EEPROM
+ * chips strung on the I2C bus, usually by manipulating pins 1-3 of a
+ * set of EEPROM devices. They form a continuous memory space.
+ *
+ * The I2C device address includes the device type identifier, 1010b,
+ * which is a reserved value and indicates that this is an I2C EEPROM
+ * device. It also includes the top 3 bits of the 19 bit EEPROM memory
+ * address, namely bits 18, 17, and 16. This makes up the 7 bit
+ * address sent on the I2C bus with bit 0 being the direction bit,
+ * which is not represented here, and sent by the hardware directly.
+ *
+ * For instance,
+ *   50h = 1010000b => device type identifier 1010b, bits 18:16 = 000b, address 0.
+ *   54h = 1010100b => --"--, bits 18:16 = 100b, address 40000h.
+ *   56h = 1010110b => --"--, bits 18:16 = 110b, address 60000h.
+ * Depending on the size of the I2C EEPROM device(s), bits 18:16 may
+ * address memory in a device or a device on the I2C bus, depending on
+ * the status of pins 1-3. See top of amdgpu_eeprom.c.
+ *
+ * The RAS table lives either at address 0 or address 40000h of EEPROM.
+ */
+#define EEPROM_I2C_MADDR_0      0x0
+#define EEPROM_I2C_MADDR_4      0x40000
 
 /*
- * The 2 macros bellow represent the actual size in bytes that
+ * The 2 macros below represent the actual size in bytes that
  * those entities occupy in the EEPROM memory.
  * RAS_TABLE_RECORD_SIZE is different than sizeof(eeprom_table_record) which
  * uses uint64 to store 6b fields such as retired_page.
@@ -48,11 +69,24 @@
 
 /* Table hdr is 'AMDR' */
 #define RAS_TABLE_HDR_VAL       0x414d4452
-#define RAS_TABLE_VER           0x00010000
 
 /* Bad GPU tag ‘BADG’ */
 #define RAS_TABLE_HDR_BAD       0x42414447
 
+/*
+ * EEPROM Table structure v1
+ * ---------------------------------
+ * |                               |
+ * |     EEPROM TABLE HEADER       |
+ * |      ( size 20 Bytes )        |
+ * |                               |
+ * ---------------------------------
+ * |                               |
+ * |    BAD PAGE RECORD AREA       |
+ * |                               |
+ * ---------------------------------
+ */
+
 /* Assume 2-Mbit size EEPROM and take up the whole space. */
 #define RAS_TBL_SIZE_BYTES      (256 * 1024)
 #define RAS_TABLE_START         0
@@ -61,6 +95,37 @@
 #define RAS_MAX_RECORD_COUNT    ((RAS_TBL_SIZE_BYTES - RAS_TABLE_HEADER_SIZE) \
 				 / RAS_TABLE_RECORD_SIZE)
 
+/*
+ * EEPROM Table structrue v2.1
+ * ---------------------------------
+ * |                               |
+ * |     EEPROM TABLE HEADER       |
+ * |      ( size 20 Bytes )        |
+ * |                               |
+ * ---------------------------------
+ * |                               |
+ * |     EEPROM TABLE RAS INFO     |
+ * | (available info size 4 Bytes) |
+ * |  ( reserved size 252 Bytes )  |
+ * |                               |
+ * ---------------------------------
+ * |                               |
+ * |     BAD PAGE RECORD AREA      |
+ * |                               |
+ * ---------------------------------
+ */
+
+/* EEPROM Table V2_1 */
+#define RAS_TABLE_V2_1_INFO_SIZE       256
+#define RAS_TABLE_V2_1_INFO_START      RAS_TABLE_HEADER_SIZE
+#define RAS_RECORD_START_V2_1          (RAS_HDR_START + RAS_TABLE_HEADER_SIZE + \
+					RAS_TABLE_V2_1_INFO_SIZE)
+#define RAS_MAX_RECORD_COUNT_V2_1      ((RAS_TBL_SIZE_BYTES - RAS_TABLE_HEADER_SIZE - \
+					RAS_TABLE_V2_1_INFO_SIZE) \
+					/ RAS_TABLE_RECORD_SIZE)
+
+#define RAS_SMU_MESSAGE_TIMEOUT_MS 1000 /* 1s */
+
 /* Given a zero-based index of an EEPROM RAS record, yields the EEPROM
  * offset off of RAS_TABLE_START.  That is, this is something you can
  * add to control->i2c_address, and then tell I2C layer to read
@@ -83,43 +148,40 @@
 #define RAS_NUM_RECS(_tbl_hdr)  (((_tbl_hdr)->tbl_size - \
 				  RAS_TABLE_HEADER_SIZE) / RAS_TABLE_RECORD_SIZE)
 
-#define to_amdgpu_device(x) (container_of(x, struct amdgpu_ras, eeprom_control))->adev
+#define RAS_NUM_RECS_V2_1(_tbl_hdr)  (((_tbl_hdr)->tbl_size - \
+				       RAS_TABLE_HEADER_SIZE - \
+				       RAS_TABLE_V2_1_INFO_SIZE) / RAS_TABLE_RECORD_SIZE)
 
-static bool __is_ras_eeprom_supported(struct amdgpu_device *adev)
-{
-	return  adev->asic_type == CHIP_VEGA20 ||
-		adev->asic_type == CHIP_ARCTURUS ||
-		adev->asic_type == CHIP_SIENNA_CICHLID ||
-		adev->asic_type == CHIP_ALDEBARAN;
-}
+#define to_amdgpu_device(x) ((container_of(x, struct amdgpu_ras, eeprom_control))->adev)
 
-static bool __get_eeprom_i2c_addr_arct(struct amdgpu_device *adev,
-				       struct amdgpu_ras_eeprom_control *control)
+static bool __is_ras_eeprom_supported(struct amdgpu_device *adev)
 {
-	struct atom_context *atom_ctx = adev->mode_info.atom_context;
-
-	if (!control || !atom_ctx)
+	switch (amdgpu_ip_version(adev, MP1_HWIP, 0)) {
+	case IP_VERSION(11, 0, 2): /* VEGA20 and ARCTURUS */
+	case IP_VERSION(11, 0, 7): /* Sienna cichlid */
+	case IP_VERSION(13, 0, 0):
+	case IP_VERSION(13, 0, 2): /* Aldebaran */
+	case IP_VERSION(13, 0, 10):
+		return true;
+	case IP_VERSION(13, 0, 6):
+	case IP_VERSION(13, 0, 12):
+	case IP_VERSION(13, 0, 14):
+		return (adev->gmc.is_app_apu) ? false : true;
+	default:
 		return false;
-
-	if (strnstr(atom_ctx->vbios_version,
-	            "D342",
-		    sizeof(atom_ctx->vbios_version)))
-		control->i2c_address = EEPROM_I2C_MADDR_ARCTURUS_D342;
-	else
-		control->i2c_address = EEPROM_I2C_MADDR_ARCTURUS;
-
-	return true;
+	}
 }
 
 static bool __get_eeprom_i2c_addr(struct amdgpu_device *adev,
 				  struct amdgpu_ras_eeprom_control *control)
 {
+	struct atom_context *atom_ctx = adev->mode_info.atom_context;
 	u8 i2c_addr;
 
 	if (!control)
 		return false;
 
-	if (amdgpu_atomfirmware_ras_rom_addr(adev, &i2c_addr)) {
+	if (adev->bios && amdgpu_atomfirmware_ras_rom_addr(adev, &i2c_addr)) {
 		/* The address given by VBIOS is an 8-bit, wire-format
 		 * address, i.e. the most significant byte.
 		 *
@@ -134,27 +196,44 @@ static bool __get_eeprom_i2c_addr(struct amdgpu_device *adev,
 		return true;
 	}
 
-	switch (adev->asic_type) {
-	case CHIP_VEGA20:
-		control->i2c_address = EEPROM_I2C_MADDR_VEGA20;
-		break;
-
-	case CHIP_ARCTURUS:
-		return __get_eeprom_i2c_addr_arct(adev, control);
-
-	case CHIP_SIENNA_CICHLID:
-		control->i2c_address = EEPROM_I2C_MADDR_SIENNA_CICHLID;
-		break;
-
-	case CHIP_ALDEBARAN:
-		control->i2c_address = EEPROM_I2C_MADDR_ALDEBARAN;
-		break;
-
+	switch (amdgpu_ip_version(adev, MP1_HWIP, 0)) {
+	case IP_VERSION(11, 0, 2):
+		/* VEGA20 and ARCTURUS */
+		if (adev->asic_type == CHIP_VEGA20)
+			control->i2c_address = EEPROM_I2C_MADDR_0;
+		else if (strnstr(atom_ctx->vbios_pn,
+				 "D342",
+				 sizeof(atom_ctx->vbios_pn)))
+			control->i2c_address = EEPROM_I2C_MADDR_0;
+		else
+			control->i2c_address = EEPROM_I2C_MADDR_4;
+		return true;
+	case IP_VERSION(11, 0, 7):
+		control->i2c_address = EEPROM_I2C_MADDR_0;
+		return true;
+	case IP_VERSION(13, 0, 2):
+		if (strnstr(atom_ctx->vbios_pn, "D673",
+			    sizeof(atom_ctx->vbios_pn)))
+			control->i2c_address = EEPROM_I2C_MADDR_4;
+		else
+			control->i2c_address = EEPROM_I2C_MADDR_0;
+		return true;
+	case IP_VERSION(13, 0, 0):
+		if (strnstr(atom_ctx->vbios_pn, "D707",
+			    sizeof(atom_ctx->vbios_pn)))
+			control->i2c_address = EEPROM_I2C_MADDR_0;
+		else
+			control->i2c_address = EEPROM_I2C_MADDR_4;
+		return true;
+	case IP_VERSION(13, 0, 6):
+	case IP_VERSION(13, 0, 10):
+	case IP_VERSION(13, 0, 12):
+	case IP_VERSION(13, 0, 14):
+		control->i2c_address = EEPROM_I2C_MADDR_4;
+		return true;
 	default:
 		return false;
 	}
-
-	return true;
 }
 
 static void
@@ -193,23 +272,89 @@ static int __write_table_header(struct amdgpu_ras_eeprom_control *control)
 	__encode_table_header_to_buf(&control->tbl_hdr, buf);
 
 	/* i2c may be unstable in gpu reset */
-	down_read(&adev->reset_sem);
-	res = amdgpu_eeprom_write(&adev->pm.smu_i2c,
+	down_read(&adev->reset_domain->sem);
+	res = amdgpu_eeprom_write(adev->pm.ras_eeprom_i2c_bus,
 				  control->i2c_address +
 				  control->ras_header_offset,
 				  buf, RAS_TABLE_HEADER_SIZE);
-	up_read(&adev->reset_sem);
+	up_read(&adev->reset_domain->sem);
 
 	if (res < 0) {
-		DRM_ERROR("Failed to write EEPROM table header:%d", res);
+		dev_err(adev->dev, "Failed to write EEPROM table header:%d",
+			res);
 	} else if (res < RAS_TABLE_HEADER_SIZE) {
-		DRM_ERROR("Short write:%d out of %d\n",
-			  res, RAS_TABLE_HEADER_SIZE);
+		dev_err(adev->dev, "Short write:%d out of %d\n", res,
+			RAS_TABLE_HEADER_SIZE);
+		res = -EIO;
+	} else {
+		res = 0;
+	}
+
+	return res;
+}
+
+static void
+__encode_table_ras_info_to_buf(struct amdgpu_ras_eeprom_table_ras_info *rai,
+			       unsigned char *buf)
+{
+	u32 *pp = (uint32_t *)buf;
+	u32 tmp;
+
+	tmp = ((uint32_t)(rai->rma_status) & 0xFF) |
+	      (((uint32_t)(rai->health_percent) << 8) & 0xFF00) |
+	      (((uint32_t)(rai->ecc_page_threshold) << 16) & 0xFFFF0000);
+	pp[0] = cpu_to_le32(tmp);
+}
+
+static void
+__decode_table_ras_info_from_buf(struct amdgpu_ras_eeprom_table_ras_info *rai,
+				 unsigned char *buf)
+{
+	u32 *pp = (uint32_t *)buf;
+	u32 tmp;
+
+	tmp = le32_to_cpu(pp[0]);
+	rai->rma_status = tmp & 0xFF;
+	rai->health_percent = (tmp >> 8) & 0xFF;
+	rai->ecc_page_threshold = (tmp >> 16) & 0xFFFF;
+}
+
+static int __write_table_ras_info(struct amdgpu_ras_eeprom_control *control)
+{
+	struct amdgpu_device *adev = to_amdgpu_device(control);
+	u8 *buf;
+	int res;
+
+	buf = kzalloc(RAS_TABLE_V2_1_INFO_SIZE, GFP_KERNEL);
+	if (!buf) {
+		dev_err(adev->dev,
+			"Failed to alloc buf to write table ras info\n");
+		return -ENOMEM;
+	}
+
+	__encode_table_ras_info_to_buf(&control->tbl_rai, buf);
+
+	/* i2c may be unstable in gpu reset */
+	down_read(&adev->reset_domain->sem);
+	res = amdgpu_eeprom_write(adev->pm.ras_eeprom_i2c_bus,
+				  control->i2c_address +
+				  control->ras_info_offset,
+				  buf, RAS_TABLE_V2_1_INFO_SIZE);
+	up_read(&adev->reset_domain->sem);
+
+	if (res < 0) {
+		dev_err(adev->dev, "Failed to write EEPROM table ras info:%d",
+			res);
+	} else if (res < RAS_TABLE_V2_1_INFO_SIZE) {
+		dev_err(adev->dev, "Short write:%d out of %d\n", res,
+			RAS_TABLE_V2_1_INFO_SIZE);
 		res = -EIO;
 	} else {
 		res = 0;
 	}
 
+	kfree(buf);
+
 	return res;
 }
 
@@ -229,6 +374,21 @@ static u8 __calc_hdr_byte_sum(const struct amdgpu_ras_eeprom_control *control)
 	return csum;
 }
 
+static u8 __calc_ras_info_byte_sum(const struct amdgpu_ras_eeprom_control *control)
+{
+	int ii;
+	u8  *pp, csum;
+	size_t sz;
+
+	sz = sizeof(control->tbl_rai);
+	pp = (u8 *) &control->tbl_rai;
+	csum = 0;
+	for (ii = 0; ii < sz; ii++, pp++)
+		csum += *pp;
+
+	return csum;
+}
+
 static int amdgpu_ras_eeprom_correct_header_tag(
 	struct amdgpu_ras_eeprom_control *control,
 	uint32_t header)
@@ -254,6 +414,25 @@ static int amdgpu_ras_eeprom_correct_header_tag(
 	return res;
 }
 
+static void amdgpu_ras_set_eeprom_table_version(struct amdgpu_ras_eeprom_control *control)
+{
+	struct amdgpu_device *adev = to_amdgpu_device(control);
+	struct amdgpu_ras_eeprom_table_header *hdr = &control->tbl_hdr;
+
+	switch (amdgpu_ip_version(adev, UMC_HWIP, 0)) {
+	case IP_VERSION(8, 10, 0):
+		hdr->version = RAS_TABLE_VER_V2_1;
+		return;
+	case IP_VERSION(12, 0, 0):
+	case IP_VERSION(12, 5, 0):
+		hdr->version = RAS_TABLE_VER_V3;
+		return;
+	default:
+		hdr->version = RAS_TABLE_VER_V1;
+		return;
+	}
+}
+
 /**
  * amdgpu_ras_eeprom_reset_table -- Reset the RAS EEPROM table
  * @control: pointer to control structure
@@ -263,25 +442,73 @@ static int amdgpu_ras_eeprom_correct_header_tag(
  */
 int amdgpu_ras_eeprom_reset_table(struct amdgpu_ras_eeprom_control *control)
 {
+	struct amdgpu_device *adev = to_amdgpu_device(control);
 	struct amdgpu_ras_eeprom_table_header *hdr = &control->tbl_hdr;
+	struct amdgpu_ras_eeprom_table_ras_info *rai = &control->tbl_rai;
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+	u32 erase_res = 0;
 	u8 csum;
 	int res;
 
 	mutex_lock(&control->ras_tbl_mutex);
 
-	hdr->header = RAS_TABLE_HDR_VAL;
-	hdr->version = RAS_TABLE_VER;
-	hdr->first_rec_offset = RAS_RECORD_START;
-	hdr->tbl_size = RAS_TABLE_HEADER_SIZE;
+	if (!amdgpu_ras_smu_eeprom_supported(adev)) {
+		hdr->header = RAS_TABLE_HDR_VAL;
+		amdgpu_ras_set_eeprom_table_version(control);
+
+		if (hdr->version >= RAS_TABLE_VER_V2_1) {
+			hdr->first_rec_offset = RAS_RECORD_START_V2_1;
+			hdr->tbl_size = RAS_TABLE_HEADER_SIZE +
+					RAS_TABLE_V2_1_INFO_SIZE;
+			rai->rma_status = GPU_HEALTH_USABLE;
+
+			control->ras_record_offset = RAS_RECORD_START_V2_1;
+			control->ras_max_record_count = RAS_MAX_RECORD_COUNT_V2_1;
+			/**
+			 * GPU health represented as a percentage.
+			 * 0 means worst health, 100 means fully health.
+			 */
+			rai->health_percent = 100;
+			/* ecc_page_threshold = 0 means disable bad page retirement */
+			rai->ecc_page_threshold = con->bad_page_cnt_threshold;
+		} else {
+			hdr->first_rec_offset = RAS_RECORD_START;
+			hdr->tbl_size = RAS_TABLE_HEADER_SIZE;
+
+			control->ras_record_offset = RAS_RECORD_START;
+			control->ras_max_record_count = RAS_MAX_RECORD_COUNT;
+		}
 
-	csum = __calc_hdr_byte_sum(control);
-	csum = -csum;
-	hdr->checksum = csum;
-	res = __write_table_header(control);
+		csum = __calc_hdr_byte_sum(control);
+		if (hdr->version >= RAS_TABLE_VER_V2_1)
+			csum += __calc_ras_info_byte_sum(control);
+		csum = -csum;
+		hdr->checksum = csum;
+		res = __write_table_header(control);
+		if (!res && hdr->version > RAS_TABLE_VER_V1)
+			res = __write_table_ras_info(control);
+	} else {
+		res = amdgpu_ras_smu_erase_ras_table(adev, &erase_res);
+		if (res || erase_res) {
+			dev_warn(adev->dev, "RAS EEPROM reset failed, res:%d result:%d",
+										res, erase_res);
+			if (!res)
+				res = -EIO;
+		}
+	}
 
 	control->ras_num_recs = 0;
+	control->ras_num_bad_pages = 0;
+	control->ras_num_mca_recs = 0;
+	control->ras_num_pa_recs = 0;
 	control->ras_fri = 0;
 
+	amdgpu_dpm_send_hbm_bad_pages_num(adev, control->ras_num_bad_pages);
+
+	control->bad_channel_bitmap = 0;
+	amdgpu_dpm_send_hbm_bad_channel_flag(adev, control->bad_channel_bitmap);
+	con->update_channel_flag = false;
+
 	amdgpu_ras_debugfs_set_ret_size(control);
 
 	mutex_unlock(&control->ras_tbl_mutex);
@@ -349,7 +576,11 @@ bool amdgpu_ras_eeprom_check_err_threshold(struct amdgpu_device *adev)
 {
 	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
 
-	if (!__is_ras_eeprom_supported(adev))
+	if (amdgpu_uniras_enabled(adev))
+		return amdgpu_ras_mgr_check_eeprom_safety_watermark(adev);
+
+	if (!__is_ras_eeprom_supported(adev) ||
+	    !amdgpu_bad_page_threshold)
 		return false;
 
 	/* skip check eeprom table for VEGA20 Gaming */
@@ -360,10 +591,19 @@ bool amdgpu_ras_eeprom_check_err_threshold(struct amdgpu_device *adev)
 			return false;
 
 	if (con->eeprom_control.tbl_hdr.header == RAS_TABLE_HDR_BAD) {
-		dev_warn(adev->dev, "This GPU is in BAD status.");
-		dev_warn(adev->dev, "Please retire it or set a larger "
-			 "threshold value when reloading driver.\n");
-		return true;
+		if (con->eeprom_control.ras_num_bad_pages > con->bad_page_cnt_threshold)
+			dev_warn(adev->dev, "RAS records:%d exceed threshold:%d",
+				 con->eeprom_control.ras_num_bad_pages, con->bad_page_cnt_threshold);
+		if ((amdgpu_bad_page_threshold == -1) ||
+		    (amdgpu_bad_page_threshold == -2)) {
+			dev_warn(adev->dev,
+				 "Please consult AMD Service Action Guide (SAG) for appropriate service procedures.\n");
+			return false;
+		} else {
+			dev_warn(adev->dev,
+				 "Please consider adjusting the customized threshold.\n");
+			return true;
+		}
 	}
 
 	return false;
@@ -387,21 +627,21 @@ static int __amdgpu_ras_eeprom_write(struct amdgpu_ras_eeprom_control *control,
 	int res;
 
 	/* i2c may be unstable in gpu reset */
-	down_read(&adev->reset_sem);
+	down_read(&adev->reset_domain->sem);
 	buf_size = num * RAS_TABLE_RECORD_SIZE;
-	res = amdgpu_eeprom_write(&adev->pm.smu_i2c,
+	res = amdgpu_eeprom_write(adev->pm.ras_eeprom_i2c_bus,
 				  control->i2c_address +
 				  RAS_INDEX_TO_OFFSET(control, fri),
 				  buf, buf_size);
-	up_read(&adev->reset_sem);
+	up_read(&adev->reset_domain->sem);
 	if (res < 0) {
-		DRM_ERROR("Writing %d EEPROM table records error:%d",
-			  num, res);
+		dev_err(adev->dev, "Writing %d EEPROM table records error:%d",
+			num, res);
 	} else if (res < buf_size) {
 		/* Short write, return error.
 		 */
-		DRM_ERROR("Wrote %d records out of %d",
-			  res / RAS_TABLE_RECORD_SIZE, num);
+		dev_err(adev->dev, "Wrote %d records out of %d",
+			res / RAS_TABLE_RECORD_SIZE, num);
 		res = -EIO;
 	} else {
 		res = 0;
@@ -415,6 +655,8 @@ amdgpu_ras_eeprom_append_table(struct amdgpu_ras_eeprom_control *control,
 			       struct eeprom_table_record *record,
 			       const u32 num)
 {
+	struct amdgpu_ras *con = amdgpu_ras_get_context(to_amdgpu_device(control));
+	struct amdgpu_device *adev = to_amdgpu_device(control);
 	u32 a, b, i;
 	u8 *buf, *pp;
 	int res;
@@ -426,9 +668,17 @@ amdgpu_ras_eeprom_append_table(struct amdgpu_ras_eeprom_control *control,
 	/* Encode all of them in one go.
 	 */
 	pp = buf;
-	for (i = 0; i < num; i++, pp += RAS_TABLE_RECORD_SIZE)
+	for (i = 0; i < num; i++, pp += RAS_TABLE_RECORD_SIZE) {
 		__encode_table_record_to_buf(control, &record[i], pp);
 
+		/* update bad channel bitmap */
+		if ((record[i].mem_channel < BITS_PER_TYPE(control->bad_channel_bitmap)) &&
+		    !(control->bad_channel_bitmap & (1 << record[i].mem_channel))) {
+			control->bad_channel_bitmap |= 1 << record[i].mem_channel;
+			con->update_channel_flag = true;
+		}
+	}
+
 	/* a, first record index to write into.
 	 * b, last record index to write into.
 	 * a = first index to read (fri) + number of records in the table,
@@ -509,6 +759,14 @@ amdgpu_ras_eeprom_append_table(struct amdgpu_ras_eeprom_control *control,
 	control->ras_num_recs = 1 + (control->ras_max_record_count + b
 				     - control->ras_fri)
 		% control->ras_max_record_count;
+
+	/*old asics only save pa to eeprom like before*/
+	if (IP_VERSION_MAJ(amdgpu_ip_version(adev, UMC_HWIP, 0)) < 12)
+		control->ras_num_pa_recs += num;
+	else
+		control->ras_num_mca_recs += num;
+
+	control->ras_num_bad_pages = con->bad_page_num;
 Out:
 	kfree(buf);
 	return res;
@@ -526,44 +784,75 @@ amdgpu_ras_eeprom_update_header(struct amdgpu_ras_eeprom_control *control)
 	/* Modify the header if it exceeds.
 	 */
 	if (amdgpu_bad_page_threshold != 0 &&
-	    control->ras_num_recs >= ras->bad_page_cnt_threshold) {
+	    control->ras_num_bad_pages > ras->bad_page_cnt_threshold) {
 		dev_warn(adev->dev,
 			"Saved bad pages %d reaches threshold value %d\n",
-			control->ras_num_recs, ras->bad_page_cnt_threshold);
-		control->tbl_hdr.header = RAS_TABLE_HDR_BAD;
+			control->ras_num_bad_pages, ras->bad_page_cnt_threshold);
+
+		if (adev->cper.enabled && !amdgpu_uniras_enabled(adev) &&
+		    amdgpu_cper_generate_bp_threshold_record(adev))
+			dev_warn(adev->dev, "fail to generate bad page threshold cper records\n");
+
+		if ((amdgpu_bad_page_threshold != -1) &&
+		    (amdgpu_bad_page_threshold != -2)) {
+			control->tbl_hdr.header = RAS_TABLE_HDR_BAD;
+			if (control->tbl_hdr.version >= RAS_TABLE_VER_V2_1) {
+				control->tbl_rai.rma_status = GPU_RETIRED__ECC_REACH_THRESHOLD;
+				control->tbl_rai.health_percent = 0;
+			}
+			ras->is_rma = true;
+		}
+
+		/* ignore the -ENOTSUPP return value */
+		amdgpu_dpm_send_rma_reason(adev);
 	}
 
-	control->tbl_hdr.version = RAS_TABLE_VER;
-	control->tbl_hdr.first_rec_offset = RAS_INDEX_TO_OFFSET(control, control->ras_fri);
-	control->tbl_hdr.tbl_size = RAS_TABLE_HEADER_SIZE + control->ras_num_recs * RAS_TABLE_RECORD_SIZE;
+	if (control->tbl_hdr.version >= RAS_TABLE_VER_V2_1)
+		control->tbl_hdr.tbl_size = RAS_TABLE_HEADER_SIZE +
+					    RAS_TABLE_V2_1_INFO_SIZE +
+					    control->ras_num_recs * RAS_TABLE_RECORD_SIZE;
+	else
+		control->tbl_hdr.tbl_size = RAS_TABLE_HEADER_SIZE +
+					    control->ras_num_recs * RAS_TABLE_RECORD_SIZE;
 	control->tbl_hdr.checksum = 0;
 
 	buf_size = control->ras_num_recs * RAS_TABLE_RECORD_SIZE;
 	buf = kcalloc(control->ras_num_recs, RAS_TABLE_RECORD_SIZE, GFP_KERNEL);
 	if (!buf) {
-		DRM_ERROR("allocating memory for table of size %d bytes failed\n",
-			  control->tbl_hdr.tbl_size);
+		dev_err(adev->dev,
+			"allocating memory for table of size %d bytes failed\n",
+			control->tbl_hdr.tbl_size);
 		res = -ENOMEM;
 		goto Out;
 	}
 
-	down_read(&adev->reset_sem);
-	res = amdgpu_eeprom_read(&adev->pm.smu_i2c,
+	down_read(&adev->reset_domain->sem);
+	res = amdgpu_eeprom_read(adev->pm.ras_eeprom_i2c_bus,
 				 control->i2c_address +
 				 control->ras_record_offset,
 				 buf, buf_size);
-	up_read(&adev->reset_sem);
+	up_read(&adev->reset_domain->sem);
 	if (res < 0) {
-		DRM_ERROR("EEPROM failed reading records:%d\n",
-			  res);
+		dev_err(adev->dev, "EEPROM failed reading records:%d\n", res);
 		goto Out;
 	} else if (res < buf_size) {
-		DRM_ERROR("EEPROM read %d out of %d bytes\n",
-			  res, buf_size);
+		dev_err(adev->dev, "EEPROM read %d out of %d bytes\n", res,
+			buf_size);
 		res = -EIO;
 		goto Out;
 	}
 
+	/**
+	 * bad page records have been stored in eeprom,
+	 * now calculate gpu health percent
+	 */
+	if (amdgpu_bad_page_threshold != 0 &&
+	    control->tbl_hdr.version >= RAS_TABLE_VER_V2_1 &&
+	    control->ras_num_bad_pages <= ras->bad_page_cnt_threshold)
+		control->tbl_rai.health_percent = ((ras->bad_page_cnt_threshold -
+						   control->ras_num_bad_pages) * 100) /
+						   ras->bad_page_cnt_threshold;
+
 	/* Recalc the checksum.
 	 */
 	csum = 0;
@@ -571,15 +860,84 @@ amdgpu_ras_eeprom_update_header(struct amdgpu_ras_eeprom_control *control)
 		csum += *pp;
 
 	csum += __calc_hdr_byte_sum(control);
+	if (control->tbl_hdr.version >= RAS_TABLE_VER_V2_1)
+		csum += __calc_ras_info_byte_sum(control);
 	/* avoid sign extension when assigning to "checksum" */
 	csum = -csum;
 	control->tbl_hdr.checksum = csum;
 	res = __write_table_header(control);
+	if (!res && control->tbl_hdr.version > RAS_TABLE_VER_V1)
+		res = __write_table_ras_info(control);
 Out:
 	kfree(buf);
 	return res;
 }
 
+int amdgpu_ras_eeprom_update_record_num(struct amdgpu_ras_eeprom_control *control)
+{
+	struct amdgpu_device *adev = to_amdgpu_device(control);
+	int ret, retry = 20;
+
+	if (!amdgpu_ras_smu_eeprom_supported(adev))
+		return 0;
+
+	control->ras_num_recs_old = control->ras_num_recs;
+
+	do {
+		/* 1000ms timeout is long enough, smu_get_badpage_count won't
+		 * return -EBUSY before timeout.
+		 */
+		ret = amdgpu_ras_smu_get_badpage_count(adev,
+			&(control->ras_num_recs), RAS_SMU_MESSAGE_TIMEOUT_MS);
+		if (!ret &&
+		    (control->ras_num_recs_old == control->ras_num_recs)) {
+			/* record number update in PMFW needs some time,
+			 * smu_get_badpage_count may return immediately without
+			 * count update, sleep for a while and retry again.
+			 */
+			msleep(50);
+			retry--;
+		} else {
+			break;
+		}
+	} while (retry);
+
+	/* no update of record number is not a real failure,
+	 * don't print warning here
+	 */
+	if (!ret && (control->ras_num_recs_old == control->ras_num_recs))
+		ret = -EINVAL;
+
+	return ret;
+}
+
+static int amdgpu_ras_smu_eeprom_append(struct amdgpu_ras_eeprom_control *control)
+{
+	struct amdgpu_device *adev = to_amdgpu_device(control);
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+
+	if (!amdgpu_ras_smu_eeprom_supported(adev) || !con)
+		return 0;
+
+	control->ras_num_bad_pages = con->bad_page_num;
+
+	if (amdgpu_bad_page_threshold != 0 &&
+	    control->ras_num_bad_pages > con->bad_page_cnt_threshold) {
+		dev_warn(adev->dev,
+			"Saved bad pages %d reaches threshold value %d\n",
+			control->ras_num_bad_pages, con->bad_page_cnt_threshold);
+
+		if (adev->cper.enabled && amdgpu_cper_generate_bp_threshold_record(adev))
+			dev_warn(adev->dev, "fail to generate bad page threshold cper records\n");
+
+		if ((amdgpu_bad_page_threshold != -1) &&
+		    (amdgpu_bad_page_threshold != -2))
+			con->is_rma = true;
+	}
+
+	return 0;
+}
+
 /**
  * amdgpu_ras_eeprom_append -- append records to the EEPROM RAS table
  * @control: pointer to control structure
@@ -598,20 +956,32 @@ int amdgpu_ras_eeprom_append(struct amdgpu_ras_eeprom_control *control,
 			     const u32 num)
 {
 	struct amdgpu_device *adev = to_amdgpu_device(control);
-	int res;
+	int res, i;
+	uint64_t nps = AMDGPU_NPS1_PARTITION_MODE;
 
 	if (!__is_ras_eeprom_supported(adev))
 		return 0;
 
+	if (amdgpu_ras_smu_eeprom_supported(adev))
+		return amdgpu_ras_smu_eeprom_append(control);
+
 	if (num == 0) {
-		DRM_ERROR("will not append 0 records\n");
+		dev_err(adev->dev, "will not append 0 records\n");
 		return -EINVAL;
 	} else if (num > control->ras_max_record_count) {
-		DRM_ERROR("cannot append %d records than the size of table %d\n",
-			  num, control->ras_max_record_count);
+		dev_err(adev->dev,
+			"cannot append %d records than the size of table %d\n",
+			num, control->ras_max_record_count);
 		return -EINVAL;
 	}
 
+	if (adev->gmc.gmc_funcs->query_mem_partition_mode)
+		nps = adev->gmc.gmc_funcs->query_mem_partition_mode(adev);
+
+	/* set the new channel index flag */
+	for (i = 0; i < num; i++)
+		record[i].retired_page |= (nps << UMC_NPS_SHIFT);
+
 	mutex_lock(&control->ras_tbl_mutex);
 
 	res = amdgpu_ras_eeprom_append_table(control, record, num);
@@ -621,6 +991,11 @@ int amdgpu_ras_eeprom_append(struct amdgpu_ras_eeprom_control *control,
 		amdgpu_ras_debugfs_set_ret_size(control);
 
 	mutex_unlock(&control->ras_tbl_mutex);
+
+	/* clear channel index flag, the flag is only saved on eeprom */
+	for (i = 0; i < num; i++)
+		record[i].retired_page &= ~(nps << UMC_NPS_SHIFT);
+
 	return res;
 }
 
@@ -642,21 +1017,21 @@ static int __amdgpu_ras_eeprom_read(struct amdgpu_ras_eeprom_control *control,
 	int res;
 
 	/* i2c may be unstable in gpu reset */
-	down_read(&adev->reset_sem);
+	down_read(&adev->reset_domain->sem);
 	buf_size = num * RAS_TABLE_RECORD_SIZE;
-	res = amdgpu_eeprom_read(&adev->pm.smu_i2c,
+	res = amdgpu_eeprom_read(adev->pm.ras_eeprom_i2c_bus,
 				 control->i2c_address +
 				 RAS_INDEX_TO_OFFSET(control, fri),
 				 buf, buf_size);
-	up_read(&adev->reset_sem);
+	up_read(&adev->reset_domain->sem);
 	if (res < 0) {
-		DRM_ERROR("Reading %d EEPROM table records error:%d",
-			  num, res);
+		dev_err(adev->dev, "Reading %d EEPROM table records error:%d",
+			num, res);
 	} else if (res < buf_size) {
 		/* Short read, return error.
 		 */
-		DRM_ERROR("Read %d records out of %d",
-			  res / RAS_TABLE_RECORD_SIZE, num);
+		dev_err(adev->dev, "Read %d records out of %d",
+			res / RAS_TABLE_RECORD_SIZE, num);
 		res = -EIO;
 	} else {
 		res = 0;
@@ -665,6 +1040,50 @@ static int __amdgpu_ras_eeprom_read(struct amdgpu_ras_eeprom_control *control,
 	return res;
 }
 
+int amdgpu_ras_eeprom_read_idx(struct amdgpu_ras_eeprom_control *control,
+			struct eeprom_table_record *record, u32 rec_idx,
+			const u32 num)
+{
+	struct amdgpu_device *adev = to_amdgpu_device(control);
+	uint64_t ts, end_idx;
+	int i, ret;
+	u64 mca, ipid;
+
+	if (!amdgpu_ras_smu_eeprom_supported(adev))
+		return 0;
+
+	if (!adev->umc.ras || !adev->umc.ras->mca_ipid_parse)
+		return -EOPNOTSUPP;
+
+	end_idx = rec_idx + num;
+	for (i = rec_idx; i < end_idx; i++) {
+		ret = amdgpu_ras_smu_get_badpage_mca_addr(adev, i, &mca);
+		if (ret)
+			return ret;
+
+		ret = amdgpu_ras_smu_get_badpage_ipid(adev, i, &ipid);
+		if (ret)
+			return ret;
+
+		ret = amdgpu_ras_smu_get_timestamp(adev, i, &ts);
+		if (ret)
+			return ret;
+
+		record[i - rec_idx].address = mca;
+		/* retired_page (pa) is unused now */
+		record[i - rec_idx].retired_page = 0x1ULL;
+		record[i - rec_idx].ts = ts;
+		record[i - rec_idx].err_type = AMDGPU_RAS_EEPROM_ERR_NON_RECOVERABLE;
+
+		adev->umc.ras->mca_ipid_parse(adev, ipid,
+			(uint32_t *)&(record[i - rec_idx].cu),
+			(uint32_t *)&(record[i - rec_idx].mem_channel),
+			(uint32_t *)&(record[i - rec_idx].mcumc_id), NULL);
+	}
+
+	return 0;
+}
+
 /**
  * amdgpu_ras_eeprom_read -- read EEPROM
  * @control: pointer to control structure
@@ -681,19 +1100,23 @@ int amdgpu_ras_eeprom_read(struct amdgpu_ras_eeprom_control *control,
 			   const u32 num)
 {
 	struct amdgpu_device *adev = to_amdgpu_device(control);
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
 	int i, res;
 	u8 *buf, *pp;
 	u32 g0, g1;
 
+	if (amdgpu_ras_smu_eeprom_supported(adev))
+		return amdgpu_ras_eeprom_read_idx(control, record, 0, num);
+
 	if (!__is_ras_eeprom_supported(adev))
 		return 0;
 
 	if (num == 0) {
-		DRM_ERROR("will not read 0 records\n");
+		dev_err(adev->dev, "will not read 0 records\n");
 		return -EINVAL;
 	} else if (num > control->ras_num_recs) {
-		DRM_ERROR("too many records to read:%d available:%d\n",
-			  num, control->ras_num_recs);
+		dev_err(adev->dev, "too many records to read:%d available:%d\n",
+			num, control->ras_num_recs);
 		return -EINVAL;
 	}
 
@@ -748,8 +1171,16 @@ int amdgpu_ras_eeprom_read(struct amdgpu_ras_eeprom_control *control,
 	/* Read up everything? Then transform.
 	 */
 	pp = buf;
-	for (i = 0; i < num; i++, pp += RAS_TABLE_RECORD_SIZE)
+	for (i = 0; i < num; i++, pp += RAS_TABLE_RECORD_SIZE) {
 		__decode_table_record_from_buf(control, &record[i], pp);
+
+		/* update bad channel bitmap */
+		if ((record[i].mem_channel < BITS_PER_TYPE(control->bad_channel_bitmap)) &&
+		    !(control->bad_channel_bitmap & (1 << record[i].mem_channel))) {
+			control->bad_channel_bitmap |= 1 << record[i].mem_channel;
+			con->update_channel_flag = true;
+		}
+	}
 Out:
 	kfree(buf);
 	mutex_unlock(&control->ras_tbl_mutex);
@@ -757,9 +1188,15 @@ Out:
 	return res;
 }
 
-uint32_t amdgpu_ras_eeprom_max_record_count(void)
+uint32_t amdgpu_ras_eeprom_max_record_count(struct amdgpu_ras_eeprom_control *control)
 {
-	return RAS_MAX_RECORD_COUNT;
+	/* get available eeprom table version first before eeprom table init */
+	amdgpu_ras_set_eeprom_table_version(control);
+
+	if (control->tbl_hdr.version >= RAS_TABLE_VER_V2_1)
+		return RAS_MAX_RECORD_COUNT_V2_1;
+	else
+		return RAS_MAX_RECORD_COUNT;
 }
 
 static ssize_t
@@ -842,6 +1279,10 @@ static ssize_t amdgpu_ras_debugfs_table_read(struct file *f, char __user *buf,
 	int res = -EFAULT;
 	size_t data_len;
 
+	/* pmfw manages eeprom data by itself */
+	if (amdgpu_ras_smu_eeprom_supported(adev))
+		return 0;
+
 	mutex_lock(&control->ras_tbl_mutex);
 
 	/* We want *pos - data_len > 0, which means there's
@@ -1001,20 +1442,27 @@ static int __verify_ras_table_checksum(struct amdgpu_ras_eeprom_control *control
 	int buf_size, res;
 	u8  csum, *buf, *pp;
 
-	buf_size = RAS_TABLE_HEADER_SIZE +
-		control->ras_num_recs * RAS_TABLE_RECORD_SIZE;
+	if (control->tbl_hdr.version >= RAS_TABLE_VER_V2_1)
+		buf_size = RAS_TABLE_HEADER_SIZE +
+			   RAS_TABLE_V2_1_INFO_SIZE +
+			   control->ras_num_recs * RAS_TABLE_RECORD_SIZE;
+	else
+		buf_size = RAS_TABLE_HEADER_SIZE +
+			   control->ras_num_recs * RAS_TABLE_RECORD_SIZE;
+
 	buf = kzalloc(buf_size, GFP_KERNEL);
 	if (!buf) {
-		DRM_ERROR("Out of memory checking RAS table checksum.\n");
+		dev_err(adev->dev,
+			"Out of memory checking RAS table checksum.\n");
 		return -ENOMEM;
 	}
 
-	res = amdgpu_eeprom_read(&adev->pm.smu_i2c,
+	res = amdgpu_eeprom_read(adev->pm.ras_eeprom_i2c_bus,
 				 control->i2c_address +
 				 control->ras_header_offset,
 				 buf, buf_size);
 	if (res < buf_size) {
-		DRM_ERROR("Partial read for checksum, res:%d\n", res);
+		dev_err(adev->dev, "Partial read for checksum, res:%d\n", res);
 		/* On partial reads, return -EIO.
 		 */
 		if (res >= 0)
@@ -1030,8 +1478,78 @@ Out:
 	return res < 0 ? res : csum;
 }
 
-int amdgpu_ras_eeprom_init(struct amdgpu_ras_eeprom_control *control,
-			   bool *exceed_err_limit)
+static int __read_table_ras_info(struct amdgpu_ras_eeprom_control *control)
+{
+	struct amdgpu_ras_eeprom_table_ras_info *rai = &control->tbl_rai;
+	struct amdgpu_device *adev = to_amdgpu_device(control);
+	unsigned char *buf;
+	int res;
+
+	buf = kzalloc(RAS_TABLE_V2_1_INFO_SIZE, GFP_KERNEL);
+	if (!buf) {
+		dev_err(adev->dev,
+			"Failed to alloc buf to read EEPROM table ras info\n");
+		return -ENOMEM;
+	}
+
+	/**
+	 * EEPROM table V2_1 supports ras info,
+	 * read EEPROM table ras info
+	 */
+	res = amdgpu_eeprom_read(adev->pm.ras_eeprom_i2c_bus,
+				 control->i2c_address + control->ras_info_offset,
+				 buf, RAS_TABLE_V2_1_INFO_SIZE);
+	if (res < RAS_TABLE_V2_1_INFO_SIZE) {
+		dev_err(adev->dev,
+			"Failed to read EEPROM table ras info, res:%d", res);
+		res = res >= 0 ? -EIO : res;
+		goto Out;
+	}
+
+	__decode_table_ras_info_from_buf(rai, buf);
+
+Out:
+	kfree(buf);
+	return res == RAS_TABLE_V2_1_INFO_SIZE ? 0 : res;
+}
+
+static int amdgpu_ras_smu_eeprom_init(struct amdgpu_ras_eeprom_control *control)
+{
+	struct amdgpu_device *adev = to_amdgpu_device(control);
+	struct amdgpu_ras_eeprom_table_header *hdr = &control->tbl_hdr;
+	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
+	uint64_t local_time;
+	int res;
+
+	ras->is_rma = false;
+
+	if (!__is_ras_eeprom_supported(adev))
+		return 0;
+	mutex_init(&control->ras_tbl_mutex);
+
+	res = amdgpu_ras_smu_get_table_version(adev, &(hdr->version));
+	if (res)
+		return res;
+
+	res = amdgpu_ras_smu_get_badpage_count(adev,
+								&(control->ras_num_recs), 100);
+	if (res)
+		return res;
+
+	local_time = (uint64_t)ktime_get_real_seconds();
+	res = amdgpu_ras_smu_set_timestamp(adev, local_time);
+	if (res)
+		return res;
+
+	control->ras_max_record_count = 4000;
+
+	control->ras_num_mca_recs = 0;
+	control->ras_num_pa_recs = 0;
+
+	return 0;
+}
+
+int amdgpu_ras_eeprom_init(struct amdgpu_ras_eeprom_control *control)
 {
 	struct amdgpu_device *adev = to_amdgpu_device(control);
 	unsigned char buf[RAS_TABLE_HEADER_SIZE] = { 0 };
@@ -1039,58 +1557,177 @@ int amdgpu_ras_eeprom_init(struct amdgpu_ras_eeprom_control *control,
 	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
 	int res;
 
-	*exceed_err_limit = false;
+	if (amdgpu_ras_smu_eeprom_supported(adev))
+		return amdgpu_ras_smu_eeprom_init(control);
+
+	ras->is_rma = false;
 
 	if (!__is_ras_eeprom_supported(adev))
 		return 0;
 
 	/* Verify i2c adapter is initialized */
-	if (!adev->pm.smu_i2c.algo)
+	if (!adev->pm.ras_eeprom_i2c_bus || !adev->pm.ras_eeprom_i2c_bus->algo)
 		return -ENOENT;
 
 	if (!__get_eeprom_i2c_addr(adev, control))
 		return -EINVAL;
 
 	control->ras_header_offset = RAS_HDR_START;
-	control->ras_record_offset = RAS_RECORD_START;
-	control->ras_max_record_count  = RAS_MAX_RECORD_COUNT;
+	control->ras_info_offset = RAS_TABLE_V2_1_INFO_START;
 	mutex_init(&control->ras_tbl_mutex);
 
 	/* Read the table header from EEPROM address */
-	res = amdgpu_eeprom_read(&adev->pm.smu_i2c,
+	res = amdgpu_eeprom_read(adev->pm.ras_eeprom_i2c_bus,
 				 control->i2c_address + control->ras_header_offset,
 				 buf, RAS_TABLE_HEADER_SIZE);
 	if (res < RAS_TABLE_HEADER_SIZE) {
-		DRM_ERROR("Failed to read EEPROM table header, res:%d", res);
+		dev_err(adev->dev, "Failed to read EEPROM table header, res:%d",
+			res);
 		return res >= 0 ? -EIO : res;
 	}
 
 	__decode_table_header_from_buf(hdr, buf);
 
-	control->ras_num_recs = RAS_NUM_RECS(hdr);
+	if (hdr->header != RAS_TABLE_HDR_VAL &&
+	    hdr->header != RAS_TABLE_HDR_BAD) {
+		dev_info(adev->dev, "Creating a new EEPROM table");
+		return amdgpu_ras_eeprom_reset_table(control);
+	}
+
+	switch (hdr->version) {
+	case RAS_TABLE_VER_V2_1:
+	case RAS_TABLE_VER_V3:
+		control->ras_num_recs = RAS_NUM_RECS_V2_1(hdr);
+		control->ras_record_offset = RAS_RECORD_START_V2_1;
+		control->ras_max_record_count = RAS_MAX_RECORD_COUNT_V2_1;
+		break;
+	case RAS_TABLE_VER_V1:
+		control->ras_num_recs = RAS_NUM_RECS(hdr);
+		control->ras_record_offset = RAS_RECORD_START;
+		control->ras_max_record_count = RAS_MAX_RECORD_COUNT;
+		break;
+	default:
+		dev_err(adev->dev,
+			"RAS header invalid, unsupported version: %u",
+			hdr->version);
+		return -EINVAL;
+	}
+
+	if (control->ras_num_recs > control->ras_max_record_count) {
+		dev_err(adev->dev,
+			"RAS header invalid, records in header: %u max allowed :%u",
+			control->ras_num_recs, control->ras_max_record_count);
+		return -EINVAL;
+	}
+
 	control->ras_fri = RAS_OFFSET_TO_INDEX(control, hdr->first_rec_offset);
+	control->ras_num_mca_recs = 0;
+	control->ras_num_pa_recs = 0;
+	return 0;
+}
+
+static int amdgpu_ras_smu_eeprom_check(struct amdgpu_ras_eeprom_control *control)
+{
+	struct amdgpu_device *adev = to_amdgpu_device(control);
+	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
+
+	if (!__is_ras_eeprom_supported(adev))
+		return 0;
+
+	control->ras_num_bad_pages = ras->bad_page_num;
+
+	if ((ras->bad_page_cnt_threshold < control->ras_num_bad_pages) &&
+	    amdgpu_bad_page_threshold != 0) {
+		dev_warn(adev->dev,
+			"RAS records:%d exceed threshold:%d\n",
+			control->ras_num_bad_pages, ras->bad_page_cnt_threshold);
+		if ((amdgpu_bad_page_threshold == -1) ||
+			(amdgpu_bad_page_threshold == -2)) {
+			dev_warn(adev->dev,
+				 "Please consult AMD Service Action Guide (SAG) for appropriate service procedures\n");
+		} else {
+			ras->is_rma = true;
+			dev_warn(adev->dev,
+				 "User defined threshold is set, runtime service will be halt when threshold is reached\n");
+		}
+
+		return 0;
+	}
+
+	dev_dbg(adev->dev,
+		"Found existing EEPROM table with %d records",
+		control->ras_num_bad_pages);
+
+	/* Warn if we are at 90% of the threshold or above
+	 */
+	if (10 * control->ras_num_bad_pages >= 9 * ras->bad_page_cnt_threshold)
+		dev_warn(adev->dev, "RAS records:%u exceeds 90%% of threshold:%d",
+				control->ras_num_bad_pages,
+				ras->bad_page_cnt_threshold);
+	return 0;
+}
+
+int amdgpu_ras_eeprom_check(struct amdgpu_ras_eeprom_control *control)
+{
+	struct amdgpu_device *adev = to_amdgpu_device(control);
+	struct amdgpu_ras_eeprom_table_header *hdr = &control->tbl_hdr;
+	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
+	int res = 0;
+
+	if (amdgpu_ras_smu_eeprom_supported(adev))
+		return amdgpu_ras_smu_eeprom_check(control);
+
+	if (!__is_ras_eeprom_supported(adev))
+		return 0;
+
+	/* Verify i2c adapter is initialized */
+	if (!adev->pm.ras_eeprom_i2c_bus || !adev->pm.ras_eeprom_i2c_bus->algo)
+		return -ENOENT;
+
+	if (!__get_eeprom_i2c_addr(adev, control))
+		return -EINVAL;
+
+	control->ras_num_bad_pages = ras->bad_page_num;
 
 	if (hdr->header == RAS_TABLE_HDR_VAL) {
-		DRM_DEBUG_DRIVER("Found existing EEPROM table with %d records",
-				 control->ras_num_recs);
+		dev_dbg(adev->dev,
+			"Found existing EEPROM table with %d records",
+			control->ras_num_bad_pages);
+
+		if (hdr->version >= RAS_TABLE_VER_V2_1) {
+			res = __read_table_ras_info(control);
+			if (res)
+				return res;
+		}
+
 		res = __verify_ras_table_checksum(control);
 		if (res)
-			DRM_ERROR("RAS table incorrect checksum or error:%d\n",
-				  res);
+			dev_err(adev->dev,
+				"RAS table incorrect checksum or error:%d\n",
+				res);
 
 		/* Warn if we are at 90% of the threshold or above
 		 */
-		if (10 * control->ras_num_recs >= 9 * ras->bad_page_cnt_threshold)
+		if (10 * control->ras_num_bad_pages >= 9 * ras->bad_page_cnt_threshold)
 			dev_warn(adev->dev, "RAS records:%u exceeds 90%% of threshold:%d",
-					control->ras_num_recs,
+					control->ras_num_bad_pages,
 					ras->bad_page_cnt_threshold);
 	} else if (hdr->header == RAS_TABLE_HDR_BAD &&
 		   amdgpu_bad_page_threshold != 0) {
+		if (hdr->version >= RAS_TABLE_VER_V2_1) {
+			res = __read_table_ras_info(control);
+			if (res)
+				return res;
+		}
+
 		res = __verify_ras_table_checksum(control);
-		if (res)
-			DRM_ERROR("RAS Table incorrect checksum or error:%d\n",
-				  res);
-		if (ras->bad_page_cnt_threshold > control->ras_num_recs) {
+		if (res) {
+			dev_err(adev->dev,
+				"RAS Table incorrect checksum or error:%d\n",
+				res);
+			return -EINVAL;
+		}
+		if (ras->bad_page_cnt_threshold >= control->ras_num_bad_pages) {
 			/* This means that, the threshold was increased since
 			 * the last time the system was booted, and now,
 			 * ras->bad_page_cnt_threshold - control->num_recs > 0,
@@ -1100,29 +1737,194 @@ int amdgpu_ras_eeprom_init(struct amdgpu_ras_eeprom_control *control,
 			dev_info(adev->dev,
 				 "records:%d threshold:%d, resetting "
 				 "RAS table header signature",
-				 control->ras_num_recs,
+				 control->ras_num_bad_pages,
 				 ras->bad_page_cnt_threshold);
 			res = amdgpu_ras_eeprom_correct_header_tag(control,
 								   RAS_TABLE_HDR_VAL);
 		} else {
-			dev_err(adev->dev, "RAS records:%d exceed threshold:%d",
-				control->ras_num_recs, ras->bad_page_cnt_threshold);
-			if (amdgpu_bad_page_threshold == -2) {
-				dev_warn(adev->dev, "GPU will be initialized due to bad_page_threshold = -2.");
+			dev_warn(adev->dev,
+				"RAS records:%d exceed threshold:%d\n",
+				control->ras_num_bad_pages, ras->bad_page_cnt_threshold);
+			if ((amdgpu_bad_page_threshold == -1) ||
+			    (amdgpu_bad_page_threshold == -2)) {
 				res = 0;
+				dev_warn(adev->dev,
+					 "Please consult AMD Service Action Guide (SAG) for appropriate service procedures\n");
 			} else {
-				*exceed_err_limit = true;
-				dev_err(adev->dev,
-					"RAS records:%d exceed threshold:%d, "
-					"GPU will not be initialized. Replace this GPU or increase the threshold",
-					control->ras_num_recs, ras->bad_page_cnt_threshold);
+				ras->is_rma = true;
+				dev_warn(adev->dev,
+					 "User defined threshold is set, runtime service will be halt when threshold is reached\n");
 			}
 		}
-	} else {
-		DRM_INFO("Creating a new EEPROM table");
-
-		res = amdgpu_ras_eeprom_reset_table(control);
 	}
 
 	return res < 0 ? res : 0;
 }
+
+void amdgpu_ras_eeprom_check_and_recover(struct amdgpu_device *adev)
+{
+	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
+	struct amdgpu_ras_eeprom_control *control;
+	int res;
+
+	if (!__is_ras_eeprom_supported(adev) || !ras ||
+	    amdgpu_ras_smu_eeprom_supported(adev))
+		return;
+	control = &ras->eeprom_control;
+	if (!control->is_eeprom_valid)
+		return;
+	res = __verify_ras_table_checksum(control);
+	if (res) {
+		dev_warn(adev->dev,
+			"RAS table incorrect checksum or error:%d, try to recover\n",
+			res);
+		if (!amdgpu_ras_eeprom_reset_table(control))
+			if (!amdgpu_ras_save_bad_pages(adev, NULL))
+				if (!__verify_ras_table_checksum(control)) {
+					dev_info(adev->dev, "RAS table recovery succeed\n");
+					return;
+				}
+		dev_err(adev->dev, "RAS table recovery failed\n");
+		control->is_eeprom_valid = false;
+	}
+	return;
+}
+
+static const struct ras_smu_drv *amdgpu_ras_get_smu_ras_drv(struct amdgpu_device *adev)
+{
+	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
+
+	if (!ras)
+		return NULL;
+
+	return ras->ras_smu_drv;
+}
+
+static uint64_t amdgpu_ras_smu_get_feature_flags(struct amdgpu_device *adev)
+{
+	const struct ras_smu_drv *ras_smu_drv = amdgpu_ras_get_smu_ras_drv(adev);
+	uint64_t flags = 0ULL;
+
+	if (!ras_smu_drv)
+		goto out;
+
+	if (ras_smu_drv->ras_smu_feature_flags)
+		ras_smu_drv->ras_smu_feature_flags(adev, &flags);
+
+out:
+	return flags;
+}
+
+bool amdgpu_ras_smu_eeprom_supported(struct amdgpu_device *adev)
+{
+	const struct ras_smu_drv *smu_ras_drv = amdgpu_ras_get_smu_ras_drv(adev);
+	uint64_t flags = 0ULL;
+
+	if (!__is_ras_eeprom_supported(adev) || !smu_ras_drv)
+		return false;
+
+	if (!smu_ras_drv->smu_eeprom_funcs)
+		return false;
+
+	flags = amdgpu_ras_smu_get_feature_flags(adev);
+
+	return !!(flags & RAS_SMU_FEATURE_BIT__RAS_EEPROM);
+}
+
+int amdgpu_ras_smu_get_table_version(struct amdgpu_device *adev,
+				     uint32_t *table_version)
+{
+	const struct ras_smu_drv *smu_ras_drv = amdgpu_ras_get_smu_ras_drv(adev);
+
+	if (!amdgpu_ras_smu_eeprom_supported(adev))
+		return -EOPNOTSUPP;
+
+	if (smu_ras_drv->smu_eeprom_funcs->get_ras_table_version)
+		return smu_ras_drv->smu_eeprom_funcs->get_ras_table_version(adev,
+										 table_version);
+	return -EOPNOTSUPP;
+}
+
+int amdgpu_ras_smu_get_badpage_count(struct amdgpu_device *adev,
+				     uint32_t *count, uint32_t timeout)
+{
+	const struct ras_smu_drv *smu_ras_drv = amdgpu_ras_get_smu_ras_drv(adev);
+
+	if (!amdgpu_ras_smu_eeprom_supported(adev))
+		return -EOPNOTSUPP;
+
+	if (smu_ras_drv->smu_eeprom_funcs->get_badpage_count)
+		return smu_ras_drv->smu_eeprom_funcs->get_badpage_count(adev,
+									     count, timeout);
+	return -EOPNOTSUPP;
+}
+
+int amdgpu_ras_smu_get_badpage_mca_addr(struct amdgpu_device *adev,
+					uint16_t index, uint64_t *mca_addr)
+{
+	const struct ras_smu_drv *smu_ras_drv = amdgpu_ras_get_smu_ras_drv(adev);
+
+	if (!amdgpu_ras_smu_eeprom_supported(adev))
+		return -EOPNOTSUPP;
+
+	if (smu_ras_drv->smu_eeprom_funcs->get_badpage_mca_addr)
+		return smu_ras_drv->smu_eeprom_funcs->get_badpage_mca_addr(adev,
+										index, mca_addr);
+	return -EOPNOTSUPP;
+}
+
+int amdgpu_ras_smu_set_timestamp(struct amdgpu_device *adev,
+				 uint64_t timestamp)
+{
+	const struct ras_smu_drv *smu_ras_drv = amdgpu_ras_get_smu_ras_drv(adev);
+
+	if (!amdgpu_ras_smu_eeprom_supported(adev))
+		return -EOPNOTSUPP;
+
+	if (smu_ras_drv->smu_eeprom_funcs->set_timestamp)
+		return smu_ras_drv->smu_eeprom_funcs->set_timestamp(adev,
+									 timestamp);
+	return -EOPNOTSUPP;
+}
+
+int amdgpu_ras_smu_get_timestamp(struct amdgpu_device *adev,
+				 uint16_t index, uint64_t *timestamp)
+{
+	const struct ras_smu_drv *smu_ras_drv = amdgpu_ras_get_smu_ras_drv(adev);
+
+	if (!amdgpu_ras_smu_eeprom_supported(adev))
+		return -EOPNOTSUPP;
+
+	if (smu_ras_drv->smu_eeprom_funcs->get_timestamp)
+		return smu_ras_drv->smu_eeprom_funcs->get_timestamp(adev,
+									 index, timestamp);
+	return -EOPNOTSUPP;
+}
+
+int amdgpu_ras_smu_get_badpage_ipid(struct amdgpu_device *adev,
+				    uint16_t index, uint64_t *ipid)
+{
+	const struct ras_smu_drv *smu_ras_drv = amdgpu_ras_get_smu_ras_drv(adev);
+
+	if (!amdgpu_ras_smu_eeprom_supported(adev))
+		return -EOPNOTSUPP;
+
+	if (smu_ras_drv->smu_eeprom_funcs->get_badpage_ipid)
+		return smu_ras_drv->smu_eeprom_funcs->get_badpage_ipid(adev,
+									    index, ipid);
+	return -EOPNOTSUPP;
+}
+
+int amdgpu_ras_smu_erase_ras_table(struct amdgpu_device *adev,
+				   uint32_t *result)
+{
+	const struct ras_smu_drv *smu_ras_drv = amdgpu_ras_get_smu_ras_drv(adev);
+
+	if (!amdgpu_ras_smu_eeprom_supported(adev))
+		return -EOPNOTSUPP;
+
+	if (smu_ras_drv->smu_eeprom_funcs->erase_ras_table)
+		return smu_ras_drv->smu_eeprom_funcs->erase_ras_table(adev,
+									   result);
+	return -EOPNOTSUPP;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.h
index 6bb00578bfbb..2e5d63957e71 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.h
@@ -26,8 +26,17 @@
 
 #include <linux/i2c.h>
 
+#define RAS_TABLE_VER_V1           0x00010000
+#define RAS_TABLE_VER_V2_1         0x00021000
+#define RAS_TABLE_VER_V3           0x00030000
+
 struct amdgpu_device;
 
+enum amdgpu_ras_gpu_health_status {
+	GPU_HEALTH_USABLE = 0,
+	GPU_RETIRED__ECC_REACH_THRESHOLD = 2,
+};
+
 enum amdgpu_ras_eeprom_err_type {
 	AMDGPU_RAS_EEPROM_ERR_NA,
 	AMDGPU_RAS_EEPROM_ERR_RECOVERABLE,
@@ -43,9 +52,18 @@ struct amdgpu_ras_eeprom_table_header {
 	uint32_t checksum;
 } __packed;
 
+struct amdgpu_ras_eeprom_table_ras_info {
+	u8  rma_status;
+	u8  health_percent;
+	u16 ecc_page_threshold;
+	u32 padding[64 - 1];
+} __packed;
+
 struct amdgpu_ras_eeprom_control {
 	struct amdgpu_ras_eeprom_table_header tbl_hdr;
 
+	struct amdgpu_ras_eeprom_table_ras_info tbl_rai;
+
 	/* Base I2C EEPPROM 19-bit memory address,
 	 * where the table is located. For more information,
 	 * see top of amdgpu_eeprom.c.
@@ -58,11 +76,24 @@ struct amdgpu_ras_eeprom_control {
 	 * right after the header.
 	 */
 	u32 ras_header_offset;
+	u32 ras_info_offset;
 	u32 ras_record_offset;
 
 	/* Number of records in the table.
 	 */
 	u32 ras_num_recs;
+	u32 ras_num_recs_old;
+
+	/* the bad page number is ras_num_recs or
+	 * ras_num_recs * umc.retire_unit
+	 */
+	u32 ras_num_bad_pages;
+
+	/* Number of records store mca address */
+	u32 ras_num_mca_recs;
+
+	/* Number of records store physical address */
+	u32 ras_num_pa_recs;
 
 	/* First record index to read, 0-based.
 	 * Range is [0, num_recs-1]. This is
@@ -80,6 +111,12 @@ struct amdgpu_ras_eeprom_control {
 	/* Protect table access via this mutex.
 	 */
 	struct mutex ras_tbl_mutex;
+
+	/* Record channel info which occurred bad pages
+	 */
+	u32 bad_channel_bitmap;
+
+	bool is_eeprom_valid;
 };
 
 /*
@@ -107,8 +144,7 @@ struct eeprom_table_record {
 	unsigned char mcumc_id;
 } __packed;
 
-int amdgpu_ras_eeprom_init(struct amdgpu_ras_eeprom_control *control,
-			   bool *exceed_err_limit);
+int amdgpu_ras_eeprom_init(struct amdgpu_ras_eeprom_control *control);
 
 int amdgpu_ras_eeprom_reset_table(struct amdgpu_ras_eeprom_control *control);
 
@@ -120,10 +156,43 @@ int amdgpu_ras_eeprom_read(struct amdgpu_ras_eeprom_control *control,
 int amdgpu_ras_eeprom_append(struct amdgpu_ras_eeprom_control *control,
 			     struct eeprom_table_record *records, const u32 num);
 
-uint32_t amdgpu_ras_eeprom_max_record_count(void);
+uint32_t amdgpu_ras_eeprom_max_record_count(struct amdgpu_ras_eeprom_control *control);
 
 void amdgpu_ras_debugfs_set_ret_size(struct amdgpu_ras_eeprom_control *control);
 
+int amdgpu_ras_eeprom_check(struct amdgpu_ras_eeprom_control *control);
+
+void amdgpu_ras_eeprom_check_and_recover(struct amdgpu_device *adev);
+
+bool amdgpu_ras_smu_eeprom_supported(struct amdgpu_device *adev);
+
+int amdgpu_ras_smu_get_table_version(struct amdgpu_device *adev,
+							uint32_t *table_version);
+
+int amdgpu_ras_smu_get_badpage_count(struct amdgpu_device *adev,
+								uint32_t *count, uint32_t timeout);
+
+int amdgpu_ras_smu_get_badpage_mca_addr(struct amdgpu_device *adev,
+								uint16_t index, uint64_t *mca_addr);
+
+int amdgpu_ras_smu_set_timestamp(struct amdgpu_device *adev,
+										uint64_t timestamp);
+
+int amdgpu_ras_smu_get_timestamp(struct amdgpu_device *adev,
+							uint16_t index, uint64_t *timestamp);
+
+int amdgpu_ras_smu_get_badpage_ipid(struct amdgpu_device *adev,
+								uint16_t index, uint64_t *ipid);
+
+int amdgpu_ras_smu_erase_ras_table(struct amdgpu_device *adev,
+									uint32_t *result);
+
+int amdgpu_ras_eeprom_read_idx(struct amdgpu_ras_eeprom_control *control,
+			struct eeprom_table_record *record, u32 rec_idx,
+			const u32 num);
+
+int amdgpu_ras_eeprom_update_record_num(struct amdgpu_ras_eeprom_control *control);
+
 extern const struct file_operations amdgpu_ras_debugfs_eeprom_size_ops;
 extern const struct file_operations amdgpu_ras_debugfs_eeprom_table_ops;
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_res_cursor.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_res_cursor.h
index acfa207cf970..be2e56ce1355 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_res_cursor.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_res_cursor.h
@@ -30,12 +30,15 @@
 #include <drm/ttm/ttm_resource.h>
 #include <drm/ttm/ttm_range_manager.h>
 
+#include "amdgpu_vram_mgr.h"
+
 /* state back for walking over vram_mgr and gtt_mgr allocations */
 struct amdgpu_res_cursor {
 	uint64_t		start;
 	uint64_t		size;
 	uint64_t		remaining;
-	struct drm_mm_node	*node;
+	void			*node;
+	uint32_t		mem_type;
 };
 
 /**
@@ -52,27 +55,64 @@ static inline void amdgpu_res_first(struct ttm_resource *res,
 				    uint64_t start, uint64_t size,
 				    struct amdgpu_res_cursor *cur)
 {
+	struct drm_buddy_block *block;
+	struct list_head *head, *next;
 	struct drm_mm_node *node;
 
-	if (!res || res->mem_type == TTM_PL_SYSTEM) {
-		cur->start = start;
-		cur->size = size;
+	if (!res)
+		goto fallback;
+
+	BUG_ON(start + size > res->size);
+
+	cur->mem_type = res->mem_type;
+
+	switch (cur->mem_type) {
+	case TTM_PL_VRAM:
+		head = &to_amdgpu_vram_mgr_resource(res)->blocks;
+
+		block = list_first_entry_or_null(head,
+						 struct drm_buddy_block,
+						 link);
+		if (!block)
+			goto fallback;
+
+		while (start >= amdgpu_vram_mgr_block_size(block)) {
+			start -= amdgpu_vram_mgr_block_size(block);
+
+			next = block->link.next;
+			if (next != head)
+				block = list_entry(next, struct drm_buddy_block, link);
+		}
+
+		cur->start = amdgpu_vram_mgr_block_start(block) + start;
+		cur->size = min(amdgpu_vram_mgr_block_size(block) - start, size);
 		cur->remaining = size;
-		cur->node = NULL;
-		WARN_ON(res && start + size > res->num_pages << PAGE_SHIFT);
-		return;
-	}
+		cur->node = block;
+		break;
+	case TTM_PL_TT:
+	case AMDGPU_PL_DOORBELL:
+	case AMDGPU_PL_MMIO_REMAP:
+		node = to_ttm_range_mgr_node(res)->mm_nodes;
+		while (start >= node->size << PAGE_SHIFT)
+			start -= node++->size << PAGE_SHIFT;
 
-	BUG_ON(start + size > res->num_pages << PAGE_SHIFT);
+		cur->start = (node->start << PAGE_SHIFT) + start;
+		cur->size = min((node->size << PAGE_SHIFT) - start, size);
+		cur->remaining = size;
+		cur->node = node;
+		break;
+	default:
+		goto fallback;
+	}
 
-	node = to_ttm_range_mgr_node(res)->mm_nodes;
-	while (start >= node->size << PAGE_SHIFT)
-		start -= node++->size << PAGE_SHIFT;
+	return;
 
-	cur->start = (node->start << PAGE_SHIFT) + start;
-	cur->size = min((node->size << PAGE_SHIFT) - start, size);
+fallback:
+	cur->start = start;
+	cur->size = size;
 	cur->remaining = size;
-	cur->node = node;
+	cur->node = NULL;
+	WARN_ON(res && start + size > res->size);
 }
 
 /**
@@ -85,7 +125,9 @@ static inline void amdgpu_res_first(struct ttm_resource *res,
  */
 static inline void amdgpu_res_next(struct amdgpu_res_cursor *cur, uint64_t size)
 {
-	struct drm_mm_node *node = cur->node;
+	struct drm_buddy_block *block;
+	struct drm_mm_node *node;
+	struct list_head *next;
 
 	BUG_ON(size > cur->remaining);
 
@@ -99,9 +141,54 @@ static inline void amdgpu_res_next(struct amdgpu_res_cursor *cur, uint64_t size)
 		return;
 	}
 
-	cur->node = ++node;
-	cur->start = node->start << PAGE_SHIFT;
-	cur->size = min(node->size << PAGE_SHIFT, cur->remaining);
+	switch (cur->mem_type) {
+	case TTM_PL_VRAM:
+		block = cur->node;
+
+		next = block->link.next;
+		block = list_entry(next, struct drm_buddy_block, link);
+
+		cur->node = block;
+		cur->start = amdgpu_vram_mgr_block_start(block);
+		cur->size = min(amdgpu_vram_mgr_block_size(block), cur->remaining);
+		break;
+	case TTM_PL_TT:
+	case AMDGPU_PL_DOORBELL:
+	case AMDGPU_PL_MMIO_REMAP:
+		node = cur->node;
+
+		cur->node = ++node;
+		cur->start = node->start << PAGE_SHIFT;
+		cur->size = min(node->size << PAGE_SHIFT, cur->remaining);
+		break;
+	default:
+		return;
+	}
+}
+
+/**
+ * amdgpu_res_cleared - check if blocks are cleared
+ *
+ * @cur: the cursor to extract the block
+ *
+ * Check if the @cur block is cleared
+ */
+static inline bool amdgpu_res_cleared(struct amdgpu_res_cursor *cur)
+{
+	struct drm_buddy_block *block;
+
+	switch (cur->mem_type) {
+	case TTM_PL_VRAM:
+		block = cur->node;
+
+		if (!amdgpu_vram_mgr_is_cleared(block))
+			return false;
+		break;
+	default:
+		return false;
+	}
+
+	return true;
 }
 
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_reset.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_reset.c
index 02afd4115675..28c4ad62f50e 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_reset.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_reset.c
@@ -23,23 +23,176 @@
 
 #include "amdgpu_reset.h"
 #include "aldebaran.h"
+#include "sienna_cichlid.h"
+#include "smu_v13_0_10.h"
 
-int amdgpu_reset_add_handler(struct amdgpu_reset_control *reset_ctl,
-			     struct amdgpu_reset_handler *handler)
+static int amdgpu_reset_xgmi_reset_on_init_suspend(struct amdgpu_device *adev)
 {
-	/* TODO: Check if handler exists? */
-	list_add_tail(&handler->handler_list, &reset_ctl->reset_handlers);
+	int i;
+
+	for (i = adev->num_ip_blocks - 1; i >= 0; i--) {
+		if (!adev->ip_blocks[i].status.valid)
+			continue;
+		if (!adev->ip_blocks[i].status.hw)
+			continue;
+		/* displays are handled in phase1 */
+		if (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_DCE)
+			continue;
+
+		/* XXX handle errors */
+		amdgpu_ip_block_suspend(&adev->ip_blocks[i]);
+		adev->ip_blocks[i].status.hw = false;
+	}
+
+	/* VCN FW shared region is in frambuffer, there are some flags
+	 * initialized in that region during sw_init. Make sure the region is
+	 * backed up.
+	 */
+	amdgpu_vcn_save_vcpu_bo(adev);
+
 	return 0;
 }
 
+static int amdgpu_reset_xgmi_reset_on_init_prep_hwctxt(
+	struct amdgpu_reset_control *reset_ctl,
+	struct amdgpu_reset_context *reset_context)
+{
+	struct list_head *reset_device_list = reset_context->reset_device_list;
+	struct amdgpu_device *tmp_adev;
+	int r;
+
+	list_for_each_entry(tmp_adev, reset_device_list, reset_list) {
+		amdgpu_unregister_gpu_instance(tmp_adev);
+		r = amdgpu_reset_xgmi_reset_on_init_suspend(tmp_adev);
+		if (r) {
+			dev_err(tmp_adev->dev,
+				"xgmi reset on init: prepare for reset failed");
+			return r;
+		}
+	}
+
+	return r;
+}
+
+static int amdgpu_reset_xgmi_reset_on_init_restore_hwctxt(
+	struct amdgpu_reset_control *reset_ctl,
+	struct amdgpu_reset_context *reset_context)
+{
+	struct list_head *reset_device_list = reset_context->reset_device_list;
+	struct amdgpu_device *tmp_adev = NULL;
+	int r;
+
+	r = amdgpu_device_reinit_after_reset(reset_context);
+	if (r)
+		return r;
+	list_for_each_entry(tmp_adev, reset_device_list, reset_list) {
+		if (!tmp_adev->kfd.init_complete) {
+			kgd2kfd_init_zone_device(tmp_adev);
+			amdgpu_amdkfd_device_init(tmp_adev);
+			amdgpu_amdkfd_drm_client_create(tmp_adev);
+		}
+	}
+
+	return r;
+}
+
+static int amdgpu_reset_xgmi_reset_on_init_perform_reset(
+	struct amdgpu_reset_control *reset_ctl,
+	struct amdgpu_reset_context *reset_context)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)reset_ctl->handle;
+	struct list_head *reset_device_list = reset_context->reset_device_list;
+	struct amdgpu_device *tmp_adev = NULL;
+	int r;
+
+	dev_dbg(adev->dev, "xgmi roi - hw reset\n");
+
+	list_for_each_entry(tmp_adev, reset_device_list, reset_list) {
+		mutex_lock(&tmp_adev->reset_cntl->reset_lock);
+		tmp_adev->reset_cntl->active_reset =
+			amdgpu_asic_reset_method(adev);
+	}
+	r = 0;
+	/* Mode1 reset needs to be triggered on all devices together */
+	list_for_each_entry(tmp_adev, reset_device_list, reset_list) {
+		/* For XGMI run all resets in parallel to speed up the process */
+		if (!queue_work(system_unbound_wq, &tmp_adev->xgmi_reset_work))
+			r = -EALREADY;
+		if (r) {
+			dev_err(tmp_adev->dev,
+				"xgmi reset on init: reset failed with error, %d",
+				r);
+			break;
+		}
+	}
+
+	/* For XGMI wait for all resets to complete before proceed */
+	if (!r) {
+		list_for_each_entry(tmp_adev, reset_device_list, reset_list) {
+			flush_work(&tmp_adev->xgmi_reset_work);
+			r = tmp_adev->asic_reset_res;
+			if (r)
+				break;
+		}
+	}
+
+	list_for_each_entry(tmp_adev, reset_device_list, reset_list) {
+		mutex_unlock(&tmp_adev->reset_cntl->reset_lock);
+		tmp_adev->reset_cntl->active_reset = AMD_RESET_METHOD_NONE;
+	}
+
+	return r;
+}
+
+int amdgpu_reset_do_xgmi_reset_on_init(
+	struct amdgpu_reset_context *reset_context)
+{
+	struct list_head *reset_device_list = reset_context->reset_device_list;
+	struct amdgpu_device *adev;
+	int r;
+
+	if (!reset_device_list || list_empty(reset_device_list) ||
+	    list_is_singular(reset_device_list))
+		return -EINVAL;
+
+	adev = list_first_entry(reset_device_list, struct amdgpu_device,
+				reset_list);
+	r = amdgpu_reset_prepare_hwcontext(adev, reset_context);
+	if (r)
+		return r;
+
+	r = amdgpu_reset_perform_reset(adev, reset_context);
+
+	return r;
+}
+
+struct amdgpu_reset_handler xgmi_reset_on_init_handler = {
+	.reset_method = AMD_RESET_METHOD_ON_INIT,
+	.prepare_env = NULL,
+	.prepare_hwcontext = amdgpu_reset_xgmi_reset_on_init_prep_hwctxt,
+	.perform_reset = amdgpu_reset_xgmi_reset_on_init_perform_reset,
+	.restore_hwcontext = amdgpu_reset_xgmi_reset_on_init_restore_hwctxt,
+	.restore_env = NULL,
+	.do_reset = NULL,
+};
+
 int amdgpu_reset_init(struct amdgpu_device *adev)
 {
 	int ret = 0;
 
-	switch (adev->asic_type) {
-	case CHIP_ALDEBARAN:
+	switch (amdgpu_ip_version(adev, MP1_HWIP, 0)) {
+	case IP_VERSION(13, 0, 2):
+	case IP_VERSION(13, 0, 6):
+	case IP_VERSION(13, 0, 12):
+	case IP_VERSION(13, 0, 14):
 		ret = aldebaran_reset_init(adev);
 		break;
+	case IP_VERSION(11, 0, 7):
+		ret = sienna_cichlid_reset_init(adev);
+		break;
+	case IP_VERSION(13, 0, 10):
+		ret = smu_v13_0_10_reset_init(adev);
+		break;
 	default:
 		break;
 	}
@@ -51,10 +204,19 @@ int amdgpu_reset_fini(struct amdgpu_device *adev)
 {
 	int ret = 0;
 
-	switch (adev->asic_type) {
-	case CHIP_ALDEBARAN:
+	switch (amdgpu_ip_version(adev, MP1_HWIP, 0)) {
+	case IP_VERSION(13, 0, 2):
+	case IP_VERSION(13, 0, 6):
+	case IP_VERSION(13, 0, 12):
+	case IP_VERSION(13, 0, 14):
 		ret = aldebaran_reset_fini(adev);
 		break;
+	case IP_VERSION(11, 0, 7):
+		ret = sienna_cichlid_reset_fini(adev);
+		break;
+	case IP_VERSION(13, 0, 10):
+		ret = smu_v13_0_10_reset_fini(adev);
+		break;
 	default:
 		break;
 	}
@@ -71,7 +233,7 @@ int amdgpu_reset_prepare_hwcontext(struct amdgpu_device *adev,
 		reset_handler = adev->reset_cntl->get_reset_handler(
 			adev->reset_cntl, reset_context);
 	if (!reset_handler)
-		return -ENOSYS;
+		return -EOPNOTSUPP;
 
 	return reset_handler->prepare_hwcontext(adev->reset_cntl,
 						reset_context);
@@ -87,7 +249,7 @@ int amdgpu_reset_perform_reset(struct amdgpu_device *adev,
 		reset_handler = adev->reset_cntl->get_reset_handler(
 			adev->reset_cntl, reset_context);
 	if (!reset_handler)
-		return -ENOSYS;
+		return -EOPNOTSUPP;
 
 	ret = reset_handler->perform_reset(adev->reset_cntl, reset_context);
 	if (ret)
@@ -96,3 +258,97 @@ int amdgpu_reset_perform_reset(struct amdgpu_device *adev,
 	return reset_handler->restore_hwcontext(adev->reset_cntl,
 						reset_context);
 }
+
+
+void amdgpu_reset_destroy_reset_domain(struct kref *ref)
+{
+	struct amdgpu_reset_domain *reset_domain = container_of(ref,
+								struct amdgpu_reset_domain,
+								refcount);
+	if (reset_domain->wq)
+		destroy_workqueue(reset_domain->wq);
+
+	kvfree(reset_domain);
+}
+
+struct amdgpu_reset_domain *amdgpu_reset_create_reset_domain(enum amdgpu_reset_domain_type type,
+							     char *wq_name)
+{
+	struct amdgpu_reset_domain *reset_domain;
+
+	reset_domain = kvzalloc(sizeof(struct amdgpu_reset_domain), GFP_KERNEL);
+	if (!reset_domain) {
+		DRM_ERROR("Failed to allocate amdgpu_reset_domain!");
+		return NULL;
+	}
+
+	reset_domain->type = type;
+	kref_init(&reset_domain->refcount);
+
+	reset_domain->wq = create_singlethread_workqueue(wq_name);
+	if (!reset_domain->wq) {
+		DRM_ERROR("Failed to allocate wq for amdgpu_reset_domain!");
+		amdgpu_reset_put_reset_domain(reset_domain);
+		return NULL;
+
+	}
+
+	atomic_set(&reset_domain->in_gpu_reset, 0);
+	atomic_set(&reset_domain->reset_res, 0);
+	init_rwsem(&reset_domain->sem);
+
+	return reset_domain;
+}
+
+void amdgpu_device_lock_reset_domain(struct amdgpu_reset_domain *reset_domain)
+{
+	atomic_set(&reset_domain->in_gpu_reset, 1);
+	down_write(&reset_domain->sem);
+}
+
+
+void amdgpu_device_unlock_reset_domain(struct amdgpu_reset_domain *reset_domain)
+{
+	atomic_set(&reset_domain->in_gpu_reset, 0);
+	up_write(&reset_domain->sem);
+}
+
+void amdgpu_reset_get_desc(struct amdgpu_reset_context *rst_ctxt, char *buf,
+			   size_t len)
+{
+	if (!buf || !len)
+		return;
+
+	switch (rst_ctxt->src) {
+	case AMDGPU_RESET_SRC_JOB:
+		if (rst_ctxt->job) {
+			snprintf(buf, len, "job hang on ring:%s",
+				 rst_ctxt->job->base.sched->name);
+		} else {
+			strscpy(buf, "job hang", len);
+		}
+		break;
+	case AMDGPU_RESET_SRC_RAS:
+		strscpy(buf, "RAS error", len);
+		break;
+	case AMDGPU_RESET_SRC_MES:
+		strscpy(buf, "MES hang", len);
+		break;
+	case AMDGPU_RESET_SRC_HWS:
+		strscpy(buf, "HWS hang", len);
+		break;
+	case AMDGPU_RESET_SRC_USER:
+		strscpy(buf, "user trigger", len);
+		break;
+	case AMDGPU_RESET_SRC_USERQ:
+		strscpy(buf, "user queue trigger", len);
+		break;
+	default:
+		strscpy(buf, "unknown", len);
+	}
+}
+
+bool amdgpu_reset_in_recovery(struct amdgpu_device *adev)
+{
+	return (adev->init_lvl->level == AMDGPU_INIT_LEVEL_RESET_RECOVERY);
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_reset.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_reset.h
index e00d38d9160a..07b4d37f1db6 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_reset.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_reset.h
@@ -26,10 +26,24 @@
 
 #include "amdgpu.h"
 
+#define AMDGPU_RESET_MAX_HANDLERS 5
+
 enum AMDGPU_RESET_FLAGS {
 
 	AMDGPU_NEED_FULL_RESET = 0,
 	AMDGPU_SKIP_HW_RESET = 1,
+	AMDGPU_SKIP_COREDUMP = 2,
+	AMDGPU_HOST_FLR = 3,
+};
+
+enum AMDGPU_RESET_SRCS {
+	AMDGPU_RESET_SRC_UNKNOWN,
+	AMDGPU_RESET_SRC_JOB,
+	AMDGPU_RESET_SRC_RAS,
+	AMDGPU_RESET_SRC_MES,
+	AMDGPU_RESET_SRC_HWS,
+	AMDGPU_RESET_SRC_USER,
+	AMDGPU_RESET_SRC_USERQ,
 };
 
 struct amdgpu_reset_context {
@@ -37,12 +51,13 @@ struct amdgpu_reset_context {
 	struct amdgpu_device *reset_req_dev;
 	struct amdgpu_job *job;
 	struct amdgpu_hive_info *hive;
+	struct list_head *reset_device_list;
 	unsigned long flags;
+	enum AMDGPU_RESET_SRCS src;
 };
 
 struct amdgpu_reset_handler {
 	enum amd_reset_method reset_method;
-	struct list_head handler_list;
 	int (*prepare_env)(struct amdgpu_reset_control *reset_ctl,
 			   struct amdgpu_reset_context *context);
 	int (*prepare_hwcontext)(struct amdgpu_reset_control *reset_ctl,
@@ -61,7 +76,8 @@ struct amdgpu_reset_control {
 	void *handle;
 	struct work_struct reset_work;
 	struct mutex reset_lock;
-	struct list_head reset_handlers;
+	struct amdgpu_reset_handler *(
+		*reset_handlers)[AMDGPU_RESET_MAX_HANDLERS];
 	atomic_t in_reset;
 	enum amd_reset_method active_reset;
 	struct amdgpu_reset_handler *(*get_reset_handler)(
@@ -70,6 +86,21 @@ struct amdgpu_reset_control {
 	void (*async_reset)(struct work_struct *work);
 };
 
+
+enum amdgpu_reset_domain_type {
+	SINGLE_DEVICE,
+	XGMI_HIVE
+};
+
+struct amdgpu_reset_domain {
+	struct kref refcount;
+	struct workqueue_struct *wq;
+	enum amdgpu_reset_domain_type type;
+	struct rw_semaphore sem;
+	atomic_t in_gpu_reset;
+	atomic_t reset_res;
+};
+
 int amdgpu_reset_init(struct amdgpu_device *adev);
 int amdgpu_reset_fini(struct amdgpu_device *adev);
 
@@ -79,7 +110,67 @@ int amdgpu_reset_prepare_hwcontext(struct amdgpu_device *adev,
 int amdgpu_reset_perform_reset(struct amdgpu_device *adev,
 			       struct amdgpu_reset_context *reset_context);
 
-int amdgpu_reset_add_handler(struct amdgpu_reset_control *reset_ctl,
-			     struct amdgpu_reset_handler *handler);
+int amdgpu_reset_prepare_env(struct amdgpu_device *adev,
+			     struct amdgpu_reset_context *reset_context);
+int amdgpu_reset_restore_env(struct amdgpu_device *adev,
+			     struct amdgpu_reset_context *reset_context);
+
+struct amdgpu_reset_domain *amdgpu_reset_create_reset_domain(enum amdgpu_reset_domain_type type,
+							     char *wq_name);
+
+void amdgpu_reset_destroy_reset_domain(struct kref *ref);
+
+static inline bool amdgpu_reset_get_reset_domain(struct amdgpu_reset_domain *domain)
+{
+	return kref_get_unless_zero(&domain->refcount) != 0;
+}
+
+static inline void amdgpu_reset_put_reset_domain(struct amdgpu_reset_domain *domain)
+{
+	if (domain)
+		kref_put(&domain->refcount, amdgpu_reset_destroy_reset_domain);
+}
+
+static inline bool amdgpu_reset_domain_schedule(struct amdgpu_reset_domain *domain,
+						struct work_struct *work)
+{
+	return queue_work(domain->wq, work);
+}
+
+static inline bool amdgpu_reset_pending(struct amdgpu_reset_domain *domain)
+{
+	lockdep_assert_held(&domain->sem);
+	return rwsem_is_contended(&domain->sem);
+}
+
+void amdgpu_device_lock_reset_domain(struct amdgpu_reset_domain *reset_domain);
+
+void amdgpu_device_unlock_reset_domain(struct amdgpu_reset_domain *reset_domain);
+
+void amdgpu_reset_get_desc(struct amdgpu_reset_context *rst_ctxt, char *buf,
+			   size_t len);
+
+#define for_each_handler(i, handler, reset_ctl)                  \
+	for (i = 0; (i < AMDGPU_RESET_MAX_HANDLERS) &&           \
+		    (handler = (*reset_ctl->reset_handlers)[i]); \
+	     ++i)
+
+extern struct amdgpu_reset_handler xgmi_reset_on_init_handler;
+int amdgpu_reset_do_xgmi_reset_on_init(
+	struct amdgpu_reset_context *reset_context);
+
+bool amdgpu_reset_in_recovery(struct amdgpu_device *adev);
+
+static inline void amdgpu_reset_set_dpc_status(struct amdgpu_device *adev,
+					       bool status)
+{
+	adev->pcie_reset_ctx.occurs_dpc = status;
+	adev->no_hw_access = status;
+}
+
+static inline bool amdgpu_reset_in_dpc(struct amdgpu_device *adev)
+{
+	return adev->pcie_reset_ctx.occurs_dpc;
+}
 
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c
index ab2351ba9574..c596b6df2e2d 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c
@@ -33,6 +33,7 @@
 
 #include <drm/amdgpu_drm.h>
 #include "amdgpu.h"
+#include "amdgpu_ras_mgr.h"
 #include "atom.h"
 
 /*
@@ -50,6 +51,26 @@
  */
 
 /**
+ * amdgpu_ring_max_ibs - Return max IBs that fit in a single submission.
+ *
+ * @type: ring type for which to return the limit.
+ */
+unsigned int amdgpu_ring_max_ibs(enum amdgpu_ring_type type)
+{
+	switch (type) {
+	case AMDGPU_RING_TYPE_GFX:
+		/* Need to keep at least 192 on GFX7+ for old radv. */
+		return 192;
+	case AMDGPU_RING_TYPE_COMPUTE:
+		return 125;
+	case AMDGPU_RING_TYPE_VCN_JPEG:
+		return 16;
+	default:
+		return 49;
+	}
+}
+
+/**
  * amdgpu_ring_alloc - allocate space on the ring buffer
  *
  * @ring: amdgpu_ring structure holding ring information
@@ -58,7 +79,7 @@
  * Allocate @ndw dwords in the ring buffer (all asics).
  * Returns 0 on success, error on failure.
  */
-int amdgpu_ring_alloc(struct amdgpu_ring *ring, unsigned ndw)
+int amdgpu_ring_alloc(struct amdgpu_ring *ring, unsigned int ndw)
 {
 	/* Align requested size with padding so unlock_commit can
 	 * pad safely */
@@ -79,6 +100,29 @@ int amdgpu_ring_alloc(struct amdgpu_ring *ring, unsigned ndw)
 	return 0;
 }
 
+/**
+ * amdgpu_ring_alloc_reemit - allocate space on the ring buffer for reemit
+ *
+ * @ring: amdgpu_ring structure holding ring information
+ * @ndw: number of dwords to allocate in the ring buffer
+ *
+ * Allocate @ndw dwords in the ring buffer (all asics).
+ * doesn't check the max_dw limit as we may be reemitting
+ * several submissions.
+ */
+static void amdgpu_ring_alloc_reemit(struct amdgpu_ring *ring, unsigned int ndw)
+{
+	/* Align requested size with padding so unlock_commit can
+	 * pad safely */
+	ndw = (ndw + ring->funcs->align_mask) & ~ring->funcs->align_mask;
+
+	ring->count_dw = ndw;
+	ring->wptr_old = ring->wptr;
+
+	if (ring->funcs->begin_use)
+		ring->funcs->begin_use(ring);
+}
+
 /** amdgpu_ring_insert_nop - insert NOP packets
  *
  * @ring: amdgpu_ring structure holding ring information
@@ -88,10 +132,22 @@ int amdgpu_ring_alloc(struct amdgpu_ring *ring, unsigned ndw)
  */
 void amdgpu_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count)
 {
-	int i;
+	uint32_t occupied, chunk1, chunk2;
+
+	occupied = ring->wptr & ring->buf_mask;
+	chunk1 = ring->buf_mask + 1 - occupied;
+	chunk1 = (chunk1 >= count) ? count : chunk1;
+	chunk2 = count - chunk1;
+
+	if (chunk1)
+		memset32(&ring->ring[occupied], ring->funcs->nop, chunk1);
 
-	for (i = 0; i < count; i++)
-		amdgpu_ring_write(ring, ring->funcs->nop);
+	if (chunk2)
+		memset32(ring->ring, ring->funcs->nop, chunk2);
+
+	ring->wptr += count;
+	ring->wptr &= ring->ptr_mask;
+	ring->count_dw -= count;
 }
 
 /**
@@ -104,8 +160,16 @@ void amdgpu_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count)
  */
 void amdgpu_ring_generic_pad_ib(struct amdgpu_ring *ring, struct amdgpu_ib *ib)
 {
-	while (ib->length_dw & ring->funcs->align_mask)
-		ib->ptr[ib->length_dw++] = ring->funcs->nop;
+	u32 align_mask = ring->funcs->align_mask;
+	u32 count = ib->length_dw & align_mask;
+
+	if (count) {
+		count = align_mask + 1 - count;
+
+		memset32(&ib->ptr[ib->length_dw], ring->funcs->nop, count);
+
+		ib->length_dw += count;
+	}
 }
 
 /**
@@ -121,11 +185,16 @@ void amdgpu_ring_commit(struct amdgpu_ring *ring)
 {
 	uint32_t count;
 
+	if (ring->count_dw < 0)
+		DRM_ERROR("amdgpu: writing more dwords to the ring than expected!\n");
+
 	/* We pad to match fetch size */
 	count = ring->funcs->align_mask + 1 -
 		(ring->wptr & ring->funcs->align_mask);
-	count %= ring->funcs->align_mask + 1;
-	ring->funcs->insert_nop(ring, count);
+	count &= ring->funcs->align_mask;
+
+	if (count != 0)
+		ring->funcs->insert_nop(ring, count);
 
 	mb();
 	amdgpu_ring_set_wptr(ring);
@@ -149,6 +218,12 @@ void amdgpu_ring_undo(struct amdgpu_ring *ring)
 		ring->funcs->end_use(ring);
 }
 
+#define amdgpu_ring_get_gpu_addr(ring, offset)				\
+	 (ring->adev->wb.gpu_addr + offset * 4)
+
+#define amdgpu_ring_get_cpu_addr(ring, offset)				\
+	 (&ring->adev->wb.wb[offset])
+
 /**
  * amdgpu_ring_init - init driver ring struct.
  *
@@ -172,6 +247,7 @@ int amdgpu_ring_init(struct amdgpu_device *adev, struct amdgpu_ring *ring,
 	int sched_hw_submission = amdgpu_sched_hw_submission;
 	u32 *num_sched;
 	u32 hw_ip;
+	unsigned int max_ibs_dw;
 
 	/* Set the hw submission limit higher for KIQ because
 	 * it's used for a number of gfx/compute tasks by both
@@ -181,6 +257,8 @@ int amdgpu_ring_init(struct amdgpu_device *adev, struct amdgpu_ring *ring,
 	 */
 	if (ring->funcs->type == AMDGPU_RING_TYPE_KIQ)
 		sched_hw_submission = max(sched_hw_submission, 256);
+	if (ring->funcs->type == AMDGPU_RING_TYPE_MES)
+		sched_hw_submission = 8;
 	else if (ring == &adev->sdma.instance[0].page)
 		sched_hw_submission = 256;
 
@@ -189,10 +267,14 @@ int amdgpu_ring_init(struct amdgpu_device *adev, struct amdgpu_ring *ring,
 			return -EINVAL;
 
 		ring->adev = adev;
+		ring->num_hw_submission = sched_hw_submission;
+		ring->sched_score = sched_score;
+		ring->vmid_wait = dma_fence_get_stub();
+
 		ring->idx = adev->num_rings++;
 		adev->rings[ring->idx] = ring;
-		r = amdgpu_fence_driver_init_ring(ring, sched_hw_submission,
-						  sched_score);
+
+		r = amdgpu_fence_driver_init_ring(ring);
 		if (r)
 			return r;
 	}
@@ -217,44 +299,89 @@ int amdgpu_ring_init(struct amdgpu_device *adev, struct amdgpu_ring *ring,
 
 	r = amdgpu_device_wb_get(adev, &ring->trail_fence_offs);
 	if (r) {
-		dev_err(adev->dev,
-			"(%d) ring trail_fence_offs wb alloc failed\n", r);
+		dev_err(adev->dev, "(%d) ring trail_fence_offs wb alloc failed\n", r);
 		return r;
 	}
-	ring->trail_fence_gpu_addr =
-		adev->wb.gpu_addr + (ring->trail_fence_offs * 4);
-	ring->trail_fence_cpu_addr = &adev->wb.wb[ring->trail_fence_offs];
 
 	r = amdgpu_device_wb_get(adev, &ring->cond_exe_offs);
 	if (r) {
 		dev_err(adev->dev, "(%d) ring cond_exec_polling wb alloc failed\n", r);
 		return r;
 	}
-	ring->cond_exe_gpu_addr = adev->wb.gpu_addr + (ring->cond_exe_offs * 4);
-	ring->cond_exe_cpu_addr = &adev->wb.wb[ring->cond_exe_offs];
+
+	ring->fence_gpu_addr =
+		amdgpu_ring_get_gpu_addr(ring, ring->fence_offs);
+	ring->fence_cpu_addr =
+		amdgpu_ring_get_cpu_addr(ring, ring->fence_offs);
+
+	ring->rptr_gpu_addr =
+		amdgpu_ring_get_gpu_addr(ring, ring->rptr_offs);
+	ring->rptr_cpu_addr =
+		amdgpu_ring_get_cpu_addr(ring, ring->rptr_offs);
+
+	ring->wptr_gpu_addr =
+		amdgpu_ring_get_gpu_addr(ring, ring->wptr_offs);
+	ring->wptr_cpu_addr =
+		amdgpu_ring_get_cpu_addr(ring, ring->wptr_offs);
+
+	ring->trail_fence_gpu_addr =
+		amdgpu_ring_get_gpu_addr(ring, ring->trail_fence_offs);
+	ring->trail_fence_cpu_addr =
+		amdgpu_ring_get_cpu_addr(ring, ring->trail_fence_offs);
+
+	ring->cond_exe_gpu_addr =
+		amdgpu_ring_get_gpu_addr(ring, ring->cond_exe_offs);
+	ring->cond_exe_cpu_addr =
+		amdgpu_ring_get_cpu_addr(ring, ring->cond_exe_offs);
+
 	/* always set cond_exec_polling to CONTINUE */
 	*ring->cond_exe_cpu_addr = 1;
 
-	r = amdgpu_fence_driver_start_ring(ring, irq_src, irq_type);
-	if (r) {
-		dev_err(adev->dev, "failed initializing fences (%d).\n", r);
-		return r;
-	}
+	if (ring->funcs->type != AMDGPU_RING_TYPE_CPER) {
+		r = amdgpu_fence_driver_start_ring(ring, irq_src, irq_type);
+		if (r) {
+			dev_err(adev->dev, "failed initializing fences (%d).\n", r);
+			return r;
+		}
+
+		max_ibs_dw = ring->funcs->emit_frame_size +
+			     amdgpu_ring_max_ibs(ring->funcs->type) * ring->funcs->emit_ib_size;
+		max_ibs_dw = (max_ibs_dw + ring->funcs->align_mask) & ~ring->funcs->align_mask;
+
+		if (WARN_ON(max_ibs_dw > max_dw))
+			max_dw = max_ibs_dw;
 
-	ring->ring_size = roundup_pow_of_two(max_dw * 4 * sched_hw_submission);
+		ring->ring_size = roundup_pow_of_two(max_dw * 4 * sched_hw_submission);
+	} else {
+		ring->ring_size = roundup_pow_of_two(max_dw * 4);
+		ring->count_dw = (ring->ring_size - 4) >> 2;
+		/* ring buffer is empty now */
+		ring->wptr = *ring->rptr_cpu_addr = 0;
+	}
 
 	ring->buf_mask = (ring->ring_size / 4) - 1;
 	ring->ptr_mask = ring->funcs->support_64bit_ptrs ?
 		0xffffffffffffffff : ring->buf_mask;
+	/*  Initialize cached_rptr to 0 */
+	ring->cached_rptr = 0;
+
+	if (!ring->ring_backup) {
+		ring->ring_backup = kvzalloc(ring->ring_size, GFP_KERNEL);
+		if (!ring->ring_backup)
+			return -ENOMEM;
+	}
+
 	/* Allocate ring buffer */
 	if (ring->ring_obj == NULL) {
-		r = amdgpu_bo_create_kernel(adev, ring->ring_size + ring->funcs->extra_dw, PAGE_SIZE,
+		r = amdgpu_bo_create_kernel(adev, ring->ring_size + ring->funcs->extra_bytes,
+					    PAGE_SIZE,
 					    AMDGPU_GEM_DOMAIN_GTT,
 					    &ring->ring_obj,
 					    &ring->gpu_addr,
 					    (void **)&ring->ring);
 		if (r) {
 			dev_err(adev->dev, "(%d) ring create failed\n", r);
+			kvfree(ring->ring_backup);
 			return r;
 		}
 		amdgpu_ring_clear_ring(ring);
@@ -263,7 +390,7 @@ int amdgpu_ring_init(struct amdgpu_device *adev, struct amdgpu_ring *ring,
 	ring->max_dw = max_dw;
 	ring->hw_prio = hw_prio;
 
-	if (!ring->no_scheduler) {
+	if (!ring->no_scheduler && ring->funcs->type < AMDGPU_HW_IP_NUM) {
 		hw_ip = ring->funcs->type;
 		num_sched = &adev->gpu_sched[hw_ip][hw_prio].num_scheds;
 		adev->gpu_sched[hw_ip][hw_prio].sched[(*num_sched)++] =
@@ -298,12 +425,12 @@ void amdgpu_ring_fini(struct amdgpu_ring *ring)
 	amdgpu_bo_free_kernel(&ring->ring_obj,
 			      &ring->gpu_addr,
 			      (void **)&ring->ring);
+	kvfree(ring->ring_backup);
+	ring->ring_backup = NULL;
 
 	dma_fence_put(ring->vmid_wait);
 	ring->vmid_wait = NULL;
 	ring->me = 0;
-
-	ring->adev->rings[ring->idx] = NULL;
 }
 
 /**
@@ -338,17 +465,30 @@ void amdgpu_ring_emit_reg_write_reg_wait_helper(struct amdgpu_ring *ring,
 bool amdgpu_ring_soft_recovery(struct amdgpu_ring *ring, unsigned int vmid,
 			       struct dma_fence *fence)
 {
-	ktime_t deadline = ktime_add_us(ktime_get(), 10000);
+	unsigned long flags;
+	ktime_t deadline;
+	bool ret;
+
+	deadline = ktime_add_us(ktime_get(), 10000);
 
 	if (amdgpu_sriov_vf(ring->adev) || !ring->funcs->soft_recovery || !fence)
 		return false;
 
-	atomic_inc(&ring->adev->gpu_reset_counter);
+	spin_lock_irqsave(fence->lock, flags);
+	if (!dma_fence_is_signaled_locked(fence))
+		dma_fence_set_error(fence, -ENODATA);
+	spin_unlock_irqrestore(fence->lock, flags);
+
 	while (!dma_fence_is_signaled(fence) &&
 	       ktime_to_ns(ktime_sub(deadline, ktime_get())) > 0)
 		ring->funcs->soft_recovery(ring, vmid);
 
-	return dma_fence_is_signaled(fence);
+	ret = dma_fence_is_signaled(fence);
+	/* increment the counter only if soft reset worked */
+	if (ret)
+		atomic_inc(&ring->adev->gpu_reset_counter);
+
+	return ret;
 }
 
 /*
@@ -356,6 +496,66 @@ bool amdgpu_ring_soft_recovery(struct amdgpu_ring *ring, unsigned int vmid,
  */
 #if defined(CONFIG_DEBUG_FS)
 
+static ssize_t amdgpu_ras_cper_debugfs_read(struct file *f, char __user *buf,
+					    size_t size, loff_t *offset)
+{
+	const uint8_t ring_header_size = 12;
+	struct amdgpu_ring *ring = file_inode(f)->i_private;
+	struct ras_cmd_cper_snapshot_req *snapshot_req __free(kfree) =
+		kzalloc(sizeof(struct ras_cmd_cper_snapshot_req), GFP_KERNEL);
+	struct ras_cmd_cper_snapshot_rsp *snapshot_rsp __free(kfree) =
+		kzalloc(sizeof(struct ras_cmd_cper_snapshot_rsp), GFP_KERNEL);
+	struct ras_cmd_cper_record_req *record_req __free(kfree) =
+		kzalloc(sizeof(struct ras_cmd_cper_record_req), GFP_KERNEL);
+	struct ras_cmd_cper_record_rsp *record_rsp __free(kfree) =
+		kzalloc(sizeof(struct ras_cmd_cper_record_rsp), GFP_KERNEL);
+	uint8_t *ring_header __free(kfree) =
+		kzalloc(ring_header_size, GFP_KERNEL);
+	uint32_t total_cper_num;
+	uint64_t start_cper_id;
+	int r;
+
+	if (!snapshot_req || !snapshot_rsp || !record_req || !record_rsp ||
+	    !ring_header)
+		return -ENOMEM;
+
+	if (!(*offset)) {
+		/* Need at least 12 bytes for the header on the first read */
+		if (size < ring_header_size)
+			return -EINVAL;
+
+		if (copy_to_user(buf, ring_header, ring_header_size))
+			return -EFAULT;
+		buf += ring_header_size;
+		size -= ring_header_size;
+	}
+
+	r = amdgpu_ras_mgr_handle_ras_cmd(ring->adev,
+					  RAS_CMD__GET_CPER_SNAPSHOT,
+					  snapshot_req, sizeof(struct ras_cmd_cper_snapshot_req),
+					  snapshot_rsp, sizeof(struct ras_cmd_cper_snapshot_rsp));
+	if (r || !snapshot_rsp->total_cper_num)
+		return r;
+
+	start_cper_id = snapshot_rsp->start_cper_id;
+	total_cper_num = snapshot_rsp->total_cper_num;
+
+	record_req->buf_ptr = (uint64_t)(uintptr_t)buf;
+	record_req->buf_size = size;
+	record_req->cper_start_id = start_cper_id + *offset;
+	record_req->cper_num = total_cper_num;
+	r = amdgpu_ras_mgr_handle_ras_cmd(ring->adev, RAS_CMD__GET_CPER_RECORD,
+					  record_req, sizeof(struct ras_cmd_cper_record_req),
+					  record_rsp, sizeof(struct ras_cmd_cper_record_rsp));
+	if (r)
+		return r;
+
+	r = *offset ? record_rsp->real_data_size : record_rsp->real_data_size + ring_header_size;
+	(*offset) += record_rsp->real_cper_num;
+
+	return r;
+}
+
 /* Layout of file is 12 bytes consisting of
  * - rptr
  * - wptr
@@ -367,8 +567,13 @@ static ssize_t amdgpu_debugfs_ring_read(struct file *f, char __user *buf,
 					size_t size, loff_t *pos)
 {
 	struct amdgpu_ring *ring = file_inode(f)->i_private;
-	int r, i;
 	uint32_t value, result, early[3];
+	uint64_t p;
+	loff_t i;
+	int r;
+
+	if (ring->funcs->type == AMDGPU_RING_TYPE_CPER && amdgpu_uniras_enabled(ring->adev))
+		return amdgpu_ras_cper_debugfs_read(f, buf, size, pos);
 
 	if (*pos & 3 || size & 3)
 		return -EINVAL;
@@ -376,13 +581,18 @@ static ssize_t amdgpu_debugfs_ring_read(struct file *f, char __user *buf,
 	result = 0;
 
 	if (*pos < 12) {
+		if (ring->funcs->type == AMDGPU_RING_TYPE_CPER)
+			mutex_lock(&ring->adev->cper.ring_lock);
+
 		early[0] = amdgpu_ring_get_rptr(ring) & ring->buf_mask;
 		early[1] = amdgpu_ring_get_wptr(ring) & ring->buf_mask;
 		early[2] = ring->wptr & ring->buf_mask;
 		for (i = *pos / 4; i < 3 && size; i++) {
 			r = put_user(early[i], (uint32_t *)buf);
-			if (r)
-				return r;
+			if (r) {
+				result = r;
+				goto out;
+			}
 			buf += 4;
 			result += 4;
 			size -= 4;
@@ -390,29 +600,113 @@ static ssize_t amdgpu_debugfs_ring_read(struct file *f, char __user *buf,
 		}
 	}
 
-	while (size) {
-		if (*pos >= (ring->ring_size + 12))
-			return result;
+	if (ring->funcs->type != AMDGPU_RING_TYPE_CPER) {
+		while (size) {
+			if (*pos >= (ring->ring_size + 12))
+				return result;
 
-		value = ring->ring[(*pos - 12)/4];
-		r = put_user(value, (uint32_t *)buf);
-		if (r)
-			return r;
-		buf += 4;
-		result += 4;
-		size -= 4;
-		*pos += 4;
+			value = ring->ring[(*pos - 12)/4];
+			r = put_user(value, (uint32_t *)buf);
+			if (r)
+				return r;
+			buf += 4;
+			result += 4;
+			size -= 4;
+			*pos += 4;
+		}
+	} else {
+		p = early[0];
+		if (early[0] <= early[1])
+			size = (early[1] - early[0]);
+		else
+			size = ring->ring_size - (early[0] - early[1]);
+
+		while (size) {
+			if (p == early[1])
+				goto out;
+
+			value = ring->ring[p];
+			r = put_user(value, (uint32_t *)buf);
+			if (r) {
+				result = r;
+				goto out;
+			}
+
+			buf += 4;
+			result += 4;
+			size--;
+			p++;
+			p &= ring->ptr_mask;
+		}
 	}
 
+out:
+	if (ring->funcs->type == AMDGPU_RING_TYPE_CPER)
+		mutex_unlock(&ring->adev->cper.ring_lock);
+
 	return result;
 }
 
+static ssize_t amdgpu_debugfs_virt_ring_read(struct file *f, char __user *buf,
+	size_t size, loff_t *pos)
+{
+	struct amdgpu_ring *ring = file_inode(f)->i_private;
+
+	if (*pos & 3 || size & 3)
+		return -EINVAL;
+
+	if (ring->funcs->type == AMDGPU_RING_TYPE_CPER)
+		amdgpu_virt_req_ras_cper_dump(ring->adev, false);
+
+	return amdgpu_debugfs_ring_read(f, buf, size, pos);
+}
+
 static const struct file_operations amdgpu_debugfs_ring_fops = {
 	.owner = THIS_MODULE,
 	.read = amdgpu_debugfs_ring_read,
 	.llseek = default_llseek
 };
 
+static const struct file_operations amdgpu_debugfs_virt_ring_fops = {
+	.owner = THIS_MODULE,
+	.read = amdgpu_debugfs_virt_ring_read,
+	.llseek = default_llseek
+};
+
+static ssize_t amdgpu_debugfs_mqd_read(struct file *f, char __user *buf,
+				       size_t size, loff_t *pos)
+{
+	struct amdgpu_ring *ring = file_inode(f)->i_private;
+	ssize_t bytes = min_t(ssize_t, ring->mqd_size - *pos, size);
+	void *from = ((u8 *)ring->mqd_ptr) + *pos;
+
+	if (*pos > ring->mqd_size)
+		return 0;
+
+	if (copy_to_user(buf, from, bytes))
+		return -EFAULT;
+
+	*pos += bytes;
+	return bytes;
+}
+
+static const struct file_operations amdgpu_debugfs_mqd_fops = {
+	.owner = THIS_MODULE,
+	.read = amdgpu_debugfs_mqd_read,
+	.llseek = default_llseek
+};
+
+static int amdgpu_debugfs_ring_error(void *data, u64 val)
+{
+	struct amdgpu_ring *ring = data;
+
+	amdgpu_fence_driver_set_error(ring, val);
+	return 0;
+}
+
+DEFINE_DEBUGFS_ATTRIBUTE_SIGNED(amdgpu_debugfs_error_fops, NULL,
+				amdgpu_debugfs_ring_error, "%lld\n");
+
 #endif
 
 void amdgpu_debugfs_ring_init(struct amdgpu_device *adev,
@@ -424,9 +718,25 @@ void amdgpu_debugfs_ring_init(struct amdgpu_device *adev,
 	char name[32];
 
 	sprintf(name, "amdgpu_ring_%s", ring->name);
-	debugfs_create_file_size(name, S_IFREG | S_IRUGO, root, ring,
-				 &amdgpu_debugfs_ring_fops,
-				 ring->ring_size + 12);
+	if (amdgpu_sriov_vf(adev))
+		debugfs_create_file_size(name, S_IFREG | 0444, root, ring,
+					 &amdgpu_debugfs_virt_ring_fops,
+					 ring->ring_size + 12);
+	else
+		debugfs_create_file_size(name, S_IFREG | 0444, root, ring,
+					 &amdgpu_debugfs_ring_fops,
+					 ring->ring_size + 12);
+
+	if (ring->mqd_obj) {
+		sprintf(name, "amdgpu_mqd_%s", ring->name);
+		debugfs_create_file_size(name, S_IFREG | 0444, root, ring,
+					 &amdgpu_debugfs_mqd_fops,
+					 ring->mqd_size);
+	}
+
+	sprintf(name, "amdgpu_error_%s", ring->name);
+	debugfs_create_file(name, 0200, root, ring,
+			    &amdgpu_debugfs_error_fops);
 
 #endif
 }
@@ -454,5 +764,164 @@ int amdgpu_ring_test_helper(struct amdgpu_ring *ring)
 			      ring->name);
 
 	ring->sched.ready = !r;
+
 	return r;
 }
+
+static void amdgpu_ring_to_mqd_prop(struct amdgpu_ring *ring,
+				    struct amdgpu_mqd_prop *prop)
+{
+	struct amdgpu_device *adev = ring->adev;
+	bool is_high_prio_compute = ring->funcs->type == AMDGPU_RING_TYPE_COMPUTE &&
+				    amdgpu_gfx_is_high_priority_compute_queue(adev, ring);
+	bool is_high_prio_gfx = ring->funcs->type == AMDGPU_RING_TYPE_GFX &&
+				amdgpu_gfx_is_high_priority_graphics_queue(adev, ring);
+
+	memset(prop, 0, sizeof(*prop));
+
+	prop->mqd_gpu_addr = ring->mqd_gpu_addr;
+	prop->hqd_base_gpu_addr = ring->gpu_addr;
+	prop->rptr_gpu_addr = ring->rptr_gpu_addr;
+	prop->wptr_gpu_addr = ring->wptr_gpu_addr;
+	prop->queue_size = ring->ring_size;
+	prop->eop_gpu_addr = ring->eop_gpu_addr;
+	prop->use_doorbell = ring->use_doorbell;
+	prop->doorbell_index = ring->doorbell_index;
+	prop->kernel_queue = true;
+
+	/* map_queues packet doesn't need activate the queue,
+	 * so only kiq need set this field.
+	 */
+	prop->hqd_active = ring->funcs->type == AMDGPU_RING_TYPE_KIQ;
+
+	prop->allow_tunneling = is_high_prio_compute;
+	if (is_high_prio_compute || is_high_prio_gfx) {
+		prop->hqd_pipe_priority = AMDGPU_GFX_PIPE_PRIO_HIGH;
+		prop->hqd_queue_priority = AMDGPU_GFX_QUEUE_PRIORITY_MAXIMUM;
+	}
+}
+
+int amdgpu_ring_init_mqd(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct amdgpu_mqd *mqd_mgr;
+	struct amdgpu_mqd_prop prop;
+
+	amdgpu_ring_to_mqd_prop(ring, &prop);
+
+	ring->wptr = 0;
+
+	if (ring->funcs->type == AMDGPU_RING_TYPE_KIQ)
+		mqd_mgr = &adev->mqds[AMDGPU_HW_IP_COMPUTE];
+	else
+		mqd_mgr = &adev->mqds[ring->funcs->type];
+
+	return mqd_mgr->init_mqd(adev, ring->mqd_ptr, &prop);
+}
+
+void amdgpu_ring_ib_begin(struct amdgpu_ring *ring)
+{
+	if (ring->is_sw_ring)
+		amdgpu_sw_ring_ib_begin(ring);
+}
+
+void amdgpu_ring_ib_end(struct amdgpu_ring *ring)
+{
+	if (ring->is_sw_ring)
+		amdgpu_sw_ring_ib_end(ring);
+}
+
+void amdgpu_ring_ib_on_emit_cntl(struct amdgpu_ring *ring)
+{
+	if (ring->is_sw_ring)
+		amdgpu_sw_ring_ib_mark_offset(ring, AMDGPU_MUX_OFFSET_TYPE_CONTROL);
+}
+
+void amdgpu_ring_ib_on_emit_ce(struct amdgpu_ring *ring)
+{
+	if (ring->is_sw_ring)
+		amdgpu_sw_ring_ib_mark_offset(ring, AMDGPU_MUX_OFFSET_TYPE_CE);
+}
+
+void amdgpu_ring_ib_on_emit_de(struct amdgpu_ring *ring)
+{
+	if (ring->is_sw_ring)
+		amdgpu_sw_ring_ib_mark_offset(ring, AMDGPU_MUX_OFFSET_TYPE_DE);
+}
+
+bool amdgpu_ring_sched_ready(struct amdgpu_ring *ring)
+{
+	if (!ring)
+		return false;
+
+	if (ring->no_scheduler || !drm_sched_wqueue_ready(&ring->sched))
+		return false;
+
+	return true;
+}
+
+void amdgpu_ring_reset_helper_begin(struct amdgpu_ring *ring,
+				    struct amdgpu_fence *guilty_fence)
+{
+	/* Stop the scheduler to prevent anybody else from touching the ring buffer. */
+	drm_sched_wqueue_stop(&ring->sched);
+	/* back up the non-guilty commands */
+	amdgpu_ring_backup_unprocessed_commands(ring, guilty_fence);
+}
+
+int amdgpu_ring_reset_helper_end(struct amdgpu_ring *ring,
+				 struct amdgpu_fence *guilty_fence)
+{
+	unsigned int i;
+	int r;
+
+	/* verify that the ring is functional */
+	r = amdgpu_ring_test_ring(ring);
+	if (r)
+		return r;
+
+	/* signal the guilty fence and set an error on all fences from the context */
+	if (guilty_fence)
+		amdgpu_fence_driver_guilty_force_completion(guilty_fence);
+	/* Re-emit the non-guilty commands */
+	if (ring->ring_backup_entries_to_copy) {
+		amdgpu_ring_alloc_reemit(ring, ring->ring_backup_entries_to_copy);
+		for (i = 0; i < ring->ring_backup_entries_to_copy; i++)
+			amdgpu_ring_write(ring, ring->ring_backup[i]);
+		amdgpu_ring_commit(ring);
+	}
+	/* Start the scheduler again */
+	drm_sched_wqueue_start(&ring->sched);
+	return 0;
+}
+
+bool amdgpu_ring_is_reset_type_supported(struct amdgpu_ring *ring,
+					 u32 reset_type)
+{
+	switch (ring->funcs->type) {
+	case AMDGPU_RING_TYPE_GFX:
+		if (ring->adev->gfx.gfx_supported_reset & reset_type)
+			return true;
+		break;
+	case AMDGPU_RING_TYPE_COMPUTE:
+		if (ring->adev->gfx.compute_supported_reset & reset_type)
+			return true;
+		break;
+	case AMDGPU_RING_TYPE_SDMA:
+		if (ring->adev->sdma.supported_reset & reset_type)
+			return true;
+		break;
+	case AMDGPU_RING_TYPE_VCN_DEC:
+	case AMDGPU_RING_TYPE_VCN_ENC:
+		if (ring->adev->vcn.supported_reset & reset_type)
+			return true;
+		break;
+	case AMDGPU_RING_TYPE_VCN_JPEG:
+		if (ring->adev->jpeg.supported_reset & reset_type)
+			return true;
+		break;
+	default:
+		break;
+	}
+	return false;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h
index 4d380e79752c..7a27c6c4bb44 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h
@@ -27,14 +27,24 @@
 #include <drm/amdgpu_drm.h>
 #include <drm/gpu_scheduler.h>
 #include <drm/drm_print.h>
+#include <drm/drm_suballoc.h>
+
+struct amdgpu_device;
+struct amdgpu_ring;
+struct amdgpu_ib;
+struct amdgpu_cs_parser;
+struct amdgpu_job;
+struct amdgpu_vm;
 
 /* max number of rings */
-#define AMDGPU_MAX_RINGS		28
-#define AMDGPU_MAX_HWIP_RINGS		8
+#define AMDGPU_MAX_RINGS		149
+#define AMDGPU_MAX_HWIP_RINGS		64
 #define AMDGPU_MAX_GFX_RINGS		2
+#define AMDGPU_MAX_SW_GFX_RINGS         2
 #define AMDGPU_MAX_COMPUTE_RINGS	8
 #define AMDGPU_MAX_VCE_RINGS		3
 #define AMDGPU_MAX_UVD_ENC_RINGS	2
+#define AMDGPU_MAX_VPE_RINGS		2
 
 enum amdgpu_ring_priority_level {
 	AMDGPU_RING_PRIO_0,
@@ -52,9 +62,7 @@ enum amdgpu_ring_priority_level {
 #define AMDGPU_FENCE_FLAG_64BIT         (1 << 0)
 #define AMDGPU_FENCE_FLAG_INT           (1 << 1)
 #define AMDGPU_FENCE_FLAG_TC_WB_ONLY    (1 << 2)
-
-/* fence flag bit to indicate the face is embedded in job*/
-#define AMDGPU_FENCE_FLAG_EMBED_IN_JOB_BIT		(DMA_FENCE_FLAG_USER_BITS + 1)
+#define AMDGPU_FENCE_FLAG_EXEC          (1 << 3)
 
 #define to_amdgpu_ring(s) container_of((s), struct amdgpu_ring, sched)
 
@@ -70,8 +78,12 @@ enum amdgpu_ring_type {
 	AMDGPU_RING_TYPE_VCN_DEC	= AMDGPU_HW_IP_VCN_DEC,
 	AMDGPU_RING_TYPE_VCN_ENC	= AMDGPU_HW_IP_VCN_ENC,
 	AMDGPU_RING_TYPE_VCN_JPEG	= AMDGPU_HW_IP_VCN_JPEG,
+	AMDGPU_RING_TYPE_VPE		= AMDGPU_HW_IP_VPE,
 	AMDGPU_RING_TYPE_KIQ,
-	AMDGPU_RING_TYPE_MES
+	AMDGPU_RING_TYPE_MES,
+	AMDGPU_RING_TYPE_UMSCH_MM,
+	AMDGPU_RING_TYPE_CPER,
+	AMDGPU_RING_TYPE_MAX,
 };
 
 enum amdgpu_ib_pool_type {
@@ -85,11 +97,13 @@ enum amdgpu_ib_pool_type {
 	AMDGPU_IB_POOL_MAX
 };
 
-struct amdgpu_device;
-struct amdgpu_ring;
-struct amdgpu_ib;
-struct amdgpu_cs_parser;
-struct amdgpu_job;
+struct amdgpu_ib {
+	struct drm_suballoc		*sa_bo;
+	uint32_t			length_dw;
+	uint64_t			gpu_addr;
+	uint32_t			*ptr;
+	uint32_t			flags;
+};
 
 struct amdgpu_sched {
 	u32				num_scheds;
@@ -101,10 +115,11 @@ struct amdgpu_sched {
  */
 struct amdgpu_fence_driver {
 	uint64_t			gpu_addr;
-	volatile uint32_t		*cpu_addr;
+	uint32_t			*cpu_addr;
 	/* sync_seq is protected by ring emission lock */
 	uint32_t			sync_seq;
 	atomic_t			last_seq;
+	u64				signalled_wptr;
 	bool				initialized;
 	struct amdgpu_irq_src		*irq_src;
 	unsigned			irq_type;
@@ -114,11 +129,35 @@ struct amdgpu_fence_driver {
 	struct dma_fence		**fences;
 };
 
+/*
+ * Fences mark an event in the GPUs pipeline and are used
+ * for GPU/CPU synchronization.  When the fence is written,
+ * it is expected that all buffers associated with that fence
+ * are no longer in use by the associated ring on the GPU and
+ * that the relevant GPU caches have been flushed.
+ */
+
+struct amdgpu_fence {
+	struct dma_fence base;
+
+	/* RB, DMA, etc. */
+	struct amdgpu_ring		*ring;
+	ktime_t				start_timestamp;
+
+	/* wptr for the fence for resets */
+	u64				wptr;
+	/* fence context for resets */
+	u64				context;
+};
+
+extern const struct drm_sched_backend_ops amdgpu_sched_ops;
+
+void amdgpu_fence_driver_set_error(struct amdgpu_ring *ring, int error);
 void amdgpu_fence_driver_force_completion(struct amdgpu_ring *ring);
+void amdgpu_fence_driver_guilty_force_completion(struct amdgpu_fence *af);
+void amdgpu_fence_save_wptr(struct amdgpu_fence *af);
 
-int amdgpu_fence_driver_init_ring(struct amdgpu_ring *ring,
-				  unsigned num_hw_submission,
-				  atomic_t *sched_score);
+int amdgpu_fence_driver_init_ring(struct amdgpu_ring *ring);
 int amdgpu_fence_driver_start_ring(struct amdgpu_ring *ring,
 				   struct amdgpu_irq_src *irq_src,
 				   unsigned irq_type);
@@ -126,8 +165,8 @@ void amdgpu_fence_driver_hw_init(struct amdgpu_device *adev);
 void amdgpu_fence_driver_hw_fini(struct amdgpu_device *adev);
 int amdgpu_fence_driver_sw_init(struct amdgpu_device *adev);
 void amdgpu_fence_driver_sw_fini(struct amdgpu_device *adev);
-int amdgpu_fence_emit(struct amdgpu_ring *ring, struct dma_fence **fence, struct amdgpu_job *job,
-		      unsigned flags);
+int amdgpu_fence_emit(struct amdgpu_ring *ring, struct amdgpu_fence *af,
+		      unsigned int flags);
 int amdgpu_fence_emit_polling(struct amdgpu_ring *ring, uint32_t *s,
 			      uint32_t timeout);
 bool amdgpu_fence_process(struct amdgpu_ring *ring);
@@ -137,27 +176,64 @@ signed long amdgpu_fence_wait_polling(struct amdgpu_ring *ring,
 				      signed long timeout);
 unsigned amdgpu_fence_count_emitted(struct amdgpu_ring *ring);
 
+void amdgpu_fence_driver_isr_toggle(struct amdgpu_device *adev, bool stop);
+
+u64 amdgpu_fence_last_unsignaled_time_us(struct amdgpu_ring *ring);
+void amdgpu_fence_update_start_timestamp(struct amdgpu_ring *ring, uint32_t seq,
+					 ktime_t timestamp);
+
 /*
  * Rings.
  */
 
 /* provided by hw blocks that expose a ring buffer for commands */
 struct amdgpu_ring_funcs {
+	/**
+	 * @type:
+	 *
+	 * GFX, Compute, SDMA, UVD, VCE, VCN, VPE, KIQ, MES, UMSCH, and CPER
+	 * use ring buffers. The type field just identifies which component the
+	 * ring buffer is associated with.
+	 */
 	enum amdgpu_ring_type	type;
 	uint32_t		align_mask;
+
+	/**
+	 * @nop:
+	 *
+	 * Every block in the amdgpu has no-op instructions (e.g., GFX 10
+	 * uses PACKET3(PACKET3_NOP, 0x3FFF), VCN 5 uses VCN_ENC_CMD_NO_OP,
+	 * etc). This field receives the specific no-op for the component
+	 * that initializes the ring.
+	 */
 	u32			nop;
 	bool			support_64bit_ptrs;
 	bool			no_user_fence;
-	unsigned		vmhub;
-	unsigned		extra_dw;
+	bool			secure_submission_supported;
+
+	/**
+	 * @extra_bytes:
+	 *
+	 * Optional extra space in bytes that is added to the ring size
+	 * when allocating the BO that holds the contents of the ring.
+	 * This space isn't used for command submission to the ring,
+	 * but is just there to satisfy some hardware requirements or
+	 * implement workarounds. It's up to the implementation of each
+	 * specific ring to initialize this space.
+	 */
+	unsigned		extra_bytes;
 
 	/* ring read/write ptr handling */
 	u64 (*get_rptr)(struct amdgpu_ring *ring);
 	u64 (*get_wptr)(struct amdgpu_ring *ring);
 	void (*set_wptr)(struct amdgpu_ring *ring);
 	/* validating and patching of IBs */
-	int (*parse_cs)(struct amdgpu_cs_parser *p, uint32_t ib_idx);
-	int (*patch_cs_in_place)(struct amdgpu_cs_parser *p, uint32_t ib_idx);
+	int (*parse_cs)(struct amdgpu_cs_parser *p,
+			struct amdgpu_job *job,
+			struct amdgpu_ib *ib);
+	int (*patch_cs_in_place)(struct amdgpu_cs_parser *p,
+				 struct amdgpu_job *job,
+				 struct amdgpu_ib *ib);
 	/* constants to calculate how many DW are needed for an emit */
 	unsigned emit_frame_size;
 	unsigned emit_ib_size;
@@ -185,13 +261,14 @@ struct amdgpu_ring_funcs {
 	void (*insert_end)(struct amdgpu_ring *ring);
 	/* pad the indirect buffer to the necessary number of dw */
 	void (*pad_ib)(struct amdgpu_ring *ring, struct amdgpu_ib *ib);
-	unsigned (*init_cond_exec)(struct amdgpu_ring *ring);
-	void (*patch_cond_exec)(struct amdgpu_ring *ring, unsigned offset);
+	unsigned (*init_cond_exec)(struct amdgpu_ring *ring, uint64_t addr);
 	/* note usage for clock and power gating */
 	void (*begin_use)(struct amdgpu_ring *ring);
 	void (*end_use)(struct amdgpu_ring *ring);
 	void (*emit_switch_buffer) (struct amdgpu_ring *ring);
 	void (*emit_cntxcntl) (struct amdgpu_ring *ring, uint32_t flags);
+	void (*emit_gfx_shadow)(struct amdgpu_ring *ring, u64 shadow_va, u64 csa_va,
+				u64 gds_va, bool init_shadow, int vmid);
 	void (*emit_rreg)(struct amdgpu_ring *ring, uint32_t reg,
 			  uint32_t reg_val_offs);
 	void (*emit_wreg)(struct amdgpu_ring *ring, uint32_t reg, uint32_t val);
@@ -207,8 +284,17 @@ struct amdgpu_ring_funcs {
 	int (*preempt_ib)(struct amdgpu_ring *ring);
 	void (*emit_mem_sync)(struct amdgpu_ring *ring);
 	void (*emit_wave_limit)(struct amdgpu_ring *ring, bool enable);
+	void (*patch_cntl)(struct amdgpu_ring *ring, unsigned offset);
+	void (*patch_ce)(struct amdgpu_ring *ring, unsigned offset);
+	void (*patch_de)(struct amdgpu_ring *ring, unsigned offset);
+	int (*reset)(struct amdgpu_ring *ring, unsigned int vmid,
+		     struct amdgpu_fence *timedout_fence);
+	void (*emit_cleaner_shader)(struct amdgpu_ring *ring);
 };
 
+/**
+ * amdgpu_ring - Holds ring information
+ */
 struct amdgpu_ring {
 	struct amdgpu_device		*adev;
 	const struct amdgpu_ring_funcs	*funcs;
@@ -216,49 +302,128 @@ struct amdgpu_ring {
 	struct drm_gpu_scheduler	sched;
 
 	struct amdgpu_bo	*ring_obj;
-	volatile uint32_t	*ring;
+	uint32_t		*ring;
+	/* backups for resets */
+	uint32_t		*ring_backup;
+	unsigned int		ring_backup_entries_to_copy;
 	unsigned		rptr_offs;
+	u64			rptr_gpu_addr;
+	u32			*rptr_cpu_addr;
+
+	/**
+	 * @wptr:
+	 *
+	 * This is part of the Ring buffer implementation and represents the
+	 * write pointer. The wptr determines where the host has written.
+	 */
 	u64			wptr;
+
+	/**
+	 * @wptr_old:
+	 *
+	 * Before update wptr with the new value, usually the old value is
+	 * stored in the wptr_old.
+	 */
 	u64			wptr_old;
 	unsigned		ring_size;
+
+	/**
+	 * @max_dw:
+	 *
+	 * Maximum number of DWords for ring allocation. This information is
+	 * provided at the ring initialization time, and each IP block can
+	 * specify a specific value. Check places that invoke
+	 * amdgpu_ring_init() to see the maximum size per block.
+	 */
 	unsigned		max_dw;
+
+	/**
+	 * @count_dw:
+	 *
+	 * This value starts with the maximum amount of DWords supported by the
+	 * ring. This value is updated based on the ring manipulation.
+	 */
 	int			count_dw;
 	uint64_t		gpu_addr;
+
+	/**
+	 * @ptr_mask:
+	 *
+	 * Some IPs provide support for 64-bit pointers and others for 32-bit
+	 * only; this behavior is component-specific and defined by the field
+	 * support_64bit_ptr. If the IP block supports 64-bits, the mask
+	 * 0xffffffffffffffff is set; otherwise, this value assumes buf_mask.
+	 * Notice that this field is used to keep wptr under a valid range.
+	 */
 	uint64_t		ptr_mask;
+
+	/**
+	 * @buf_mask:
+	 *
+	 * Buffer mask is a value used to keep wptr count under its
+	 * thresholding. Buffer mask initialized during the ring buffer
+	 * initialization time, and it is defined as (ring_size / 4) -1.
+	 */
 	uint32_t		buf_mask;
 	u32			idx;
+	u32			xcc_id;
+	u32			xcp_id;
 	u32			me;
 	u32			pipe;
 	u32			queue;
 	struct amdgpu_bo	*mqd_obj;
 	uint64_t                mqd_gpu_addr;
 	void                    *mqd_ptr;
+	unsigned                mqd_size;
 	uint64_t                eop_gpu_addr;
 	u32			doorbell_index;
 	bool			use_doorbell;
 	bool			use_pollmem;
 	unsigned		wptr_offs;
+	u64			wptr_gpu_addr;
+
+	/**
+	 * @wptr_cpu_addr:
+	 *
+	 * This is the CPU address pointer in the writeback slot. This is used
+	 * to commit changes to the GPU.
+	 */
+	u32			*wptr_cpu_addr;
 	unsigned		fence_offs;
+	u64			fence_gpu_addr;
+	u32			*fence_cpu_addr;
 	uint64_t		current_ctx;
 	char			name[16];
 	u32                     trail_seq;
 	unsigned		trail_fence_offs;
 	u64			trail_fence_gpu_addr;
-	volatile u32		*trail_fence_cpu_addr;
+	u32			*trail_fence_cpu_addr;
 	unsigned		cond_exe_offs;
 	u64			cond_exe_gpu_addr;
-	volatile u32		*cond_exe_cpu_addr;
+	u32			*cond_exe_cpu_addr;
+	unsigned int		set_q_mode_offs;
+	u32			*set_q_mode_ptr;
+	u64			set_q_mode_token;
+	unsigned		vm_hub;
 	unsigned		vm_inv_eng;
 	struct dma_fence	*vmid_wait;
 	bool			has_compute_vm_bug;
 	bool			no_scheduler;
+	bool			no_user_submission;
 	int			hw_prio;
+	unsigned 		num_hw_submission;
+	atomic_t		*sched_score;
+
+	bool            is_sw_ring;
+	unsigned int    entry_index;
+	/* store the cached rptr to restore after reset */
+	uint64_t cached_rptr;
 };
 
-#define amdgpu_ring_parse_cs(r, p, ib) ((r)->funcs->parse_cs((p), (ib)))
-#define amdgpu_ring_patch_cs_in_place(r, p, ib) ((r)->funcs->patch_cs_in_place((p), (ib)))
+#define amdgpu_ring_parse_cs(r, p, job, ib) ((r)->funcs->parse_cs((p), (job), (ib)))
+#define amdgpu_ring_patch_cs_in_place(r, p, job, ib) ((r)->funcs->patch_cs_in_place((p), (job), (ib)))
 #define amdgpu_ring_test_ring(r) (r)->funcs->test_ring((r))
-#define amdgpu_ring_test_ib(r, t) (r)->funcs->test_ib((r), (t))
+#define amdgpu_ring_test_ib(r, t) ((r)->funcs->test_ib ? (r)->funcs->test_ib((r), (t)) : 0)
 #define amdgpu_ring_get_rptr(r) (r)->funcs->get_rptr((r))
 #define amdgpu_ring_get_wptr(r) (r)->funcs->get_wptr((r))
 #define amdgpu_ring_set_wptr(r) (r)->funcs->set_wptr((r))
@@ -270,24 +435,35 @@ struct amdgpu_ring {
 #define amdgpu_ring_emit_hdp_flush(r) (r)->funcs->emit_hdp_flush((r))
 #define amdgpu_ring_emit_switch_buffer(r) (r)->funcs->emit_switch_buffer((r))
 #define amdgpu_ring_emit_cntxcntl(r, d) (r)->funcs->emit_cntxcntl((r), (d))
+#define amdgpu_ring_emit_gfx_shadow(r, s, c, g, i, v) ((r)->funcs->emit_gfx_shadow((r), (s), (c), (g), (i), (v)))
 #define amdgpu_ring_emit_rreg(r, d, o) (r)->funcs->emit_rreg((r), (d), (o))
 #define amdgpu_ring_emit_wreg(r, d, v) (r)->funcs->emit_wreg((r), (d), (v))
 #define amdgpu_ring_emit_reg_wait(r, d, v, m) (r)->funcs->emit_reg_wait((r), (d), (v), (m))
 #define amdgpu_ring_emit_reg_write_reg_wait(r, d0, d1, v, m) (r)->funcs->emit_reg_write_reg_wait((r), (d0), (d1), (v), (m))
 #define amdgpu_ring_emit_frame_cntl(r, b, s) (r)->funcs->emit_frame_cntl((r), (b), (s))
 #define amdgpu_ring_pad_ib(r, ib) ((r)->funcs->pad_ib((r), (ib)))
-#define amdgpu_ring_init_cond_exec(r) (r)->funcs->init_cond_exec((r))
-#define amdgpu_ring_patch_cond_exec(r,o) (r)->funcs->patch_cond_exec((r),(o))
+#define amdgpu_ring_init_cond_exec(r, a) (r)->funcs->init_cond_exec((r), (a))
 #define amdgpu_ring_preempt_ib(r) (r)->funcs->preempt_ib(r)
+#define amdgpu_ring_patch_cntl(r, o) ((r)->funcs->patch_cntl((r), (o)))
+#define amdgpu_ring_patch_ce(r, o) ((r)->funcs->patch_ce((r), (o)))
+#define amdgpu_ring_patch_de(r, o) ((r)->funcs->patch_de((r), (o)))
+#define amdgpu_ring_reset(r, v, f) (r)->funcs->reset((r), (v), (f))
 
+unsigned int amdgpu_ring_max_ibs(enum amdgpu_ring_type type);
 int amdgpu_ring_alloc(struct amdgpu_ring *ring, unsigned ndw);
+void amdgpu_ring_ib_begin(struct amdgpu_ring *ring);
+void amdgpu_ring_ib_end(struct amdgpu_ring *ring);
+void amdgpu_ring_ib_on_emit_cntl(struct amdgpu_ring *ring);
+void amdgpu_ring_ib_on_emit_ce(struct amdgpu_ring *ring);
+void amdgpu_ring_ib_on_emit_de(struct amdgpu_ring *ring);
+
 void amdgpu_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count);
 void amdgpu_ring_generic_pad_ib(struct amdgpu_ring *ring, struct amdgpu_ib *ib);
 void amdgpu_ring_commit(struct amdgpu_ring *ring);
 void amdgpu_ring_undo(struct amdgpu_ring *ring);
 int amdgpu_ring_init(struct amdgpu_device *adev, struct amdgpu_ring *ring,
-		     unsigned int ring_size, struct amdgpu_irq_src *irq_src,
-		     unsigned int irq_type, unsigned int prio,
+		     unsigned int max_dw, struct amdgpu_irq_src *irq_src,
+		     unsigned int irq_type, unsigned int hw_prio,
 		     atomic_t *sched_score);
 void amdgpu_ring_fini(struct amdgpu_ring *ring);
 void amdgpu_ring_emit_reg_write_reg_wait_helper(struct amdgpu_ring *ring,
@@ -304,16 +480,11 @@ static inline void amdgpu_ring_set_preempt_cond_exec(struct amdgpu_ring *ring,
 
 static inline void amdgpu_ring_clear_ring(struct amdgpu_ring *ring)
 {
-	int i = 0;
-	while (i <= ring->buf_mask)
-		ring->ring[i++] = ring->funcs->nop;
-
+	memset32(ring->ring, ring->funcs->nop, ring->buf_mask + 1);
 }
 
 static inline void amdgpu_ring_write(struct amdgpu_ring *ring, uint32_t v)
 {
-	if (ring->count_dw <= 0)
-		DRM_ERROR("amdgpu: writing more dwords to the ring than expected!\n");
 	ring->ring[ring->wptr++ & ring->buf_mask] = v;
 	ring->wptr &= ring->ptr_mask;
 	ring->count_dw--;
@@ -323,26 +494,20 @@ static inline void amdgpu_ring_write_multiple(struct amdgpu_ring *ring,
 					      void *src, int count_dw)
 {
 	unsigned occupied, chunk1, chunk2;
-	void *dst;
-
-	if (unlikely(ring->count_dw < count_dw))
-		DRM_ERROR("amdgpu: writing more dwords to the ring than expected!\n");
 
 	occupied = ring->wptr & ring->buf_mask;
-	dst = (void *)&ring->ring[occupied];
 	chunk1 = ring->buf_mask + 1 - occupied;
-	chunk1 = (chunk1 >= count_dw) ? count_dw: chunk1;
+	chunk1 = (chunk1 >= count_dw) ? count_dw : chunk1;
 	chunk2 = count_dw - chunk1;
 	chunk1 <<= 2;
 	chunk2 <<= 2;
 
 	if (chunk1)
-		memcpy(dst, src, chunk1);
+		memcpy(&ring->ring[occupied], src, chunk1);
 
 	if (chunk2) {
 		src += chunk1;
-		dst = (void *)ring->ring;
-		memcpy(dst, src, chunk2);
+		memcpy(ring->ring, src, chunk2);
 	}
 
 	ring->wptr += count_dw;
@@ -350,8 +515,66 @@ static inline void amdgpu_ring_write_multiple(struct amdgpu_ring *ring,
 	ring->count_dw -= count_dw;
 }
 
+/**
+ * amdgpu_ring_patch_cond_exec - patch dw count of conditional execute
+ * @ring: amdgpu_ring structure
+ * @offset: offset returned by amdgpu_ring_init_cond_exec
+ *
+ * Calculate the dw count and patch it into a cond_exec command.
+ */
+static inline void amdgpu_ring_patch_cond_exec(struct amdgpu_ring *ring,
+					       unsigned int offset)
+{
+	unsigned cur;
+
+	if (!ring->funcs->init_cond_exec)
+		return;
+
+	WARN_ON(offset > ring->buf_mask);
+	WARN_ON(ring->ring[offset] != 0);
+
+	cur = (ring->wptr - 1) & ring->buf_mask;
+	if (cur < offset)
+		cur += ring->ring_size >> 2;
+	ring->ring[offset] = cur - offset;
+}
+
 int amdgpu_ring_test_helper(struct amdgpu_ring *ring);
 
 void amdgpu_debugfs_ring_init(struct amdgpu_device *adev,
 			      struct amdgpu_ring *ring);
+
+int amdgpu_ring_init_mqd(struct amdgpu_ring *ring);
+
+static inline u32 amdgpu_ib_get_value(struct amdgpu_ib *ib, int idx)
+{
+	return ib->ptr[idx];
+}
+
+static inline void amdgpu_ib_set_value(struct amdgpu_ib *ib, int idx,
+				       uint32_t value)
+{
+	ib->ptr[idx] = value;
+}
+
+int amdgpu_ib_get(struct amdgpu_device *adev, struct amdgpu_vm *vm,
+		  unsigned size,
+		  enum amdgpu_ib_pool_type pool,
+		  struct amdgpu_ib *ib);
+void amdgpu_ib_free(struct amdgpu_ib *ib, struct dma_fence *f);
+int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs,
+		       struct amdgpu_ib *ibs, struct amdgpu_job *job,
+		       struct dma_fence **f);
+int amdgpu_ib_pool_init(struct amdgpu_device *adev);
+void amdgpu_ib_pool_fini(struct amdgpu_device *adev);
+int amdgpu_ib_ring_tests(struct amdgpu_device *adev);
+bool amdgpu_ring_sched_ready(struct amdgpu_ring *ring);
+void amdgpu_ring_backup_unprocessed_commands(struct amdgpu_ring *ring,
+					     struct amdgpu_fence *guilty_fence);
+void amdgpu_ring_reset_helper_begin(struct amdgpu_ring *ring,
+				    struct amdgpu_fence *guilty_fence);
+int amdgpu_ring_reset_helper_end(struct amdgpu_ring *ring,
+				 struct amdgpu_fence *guilty_fence);
+bool amdgpu_ring_is_reset_type_supported(struct amdgpu_ring *ring,
+					 u32 reset_type);
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring_mux.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring_mux.c
new file mode 100644
index 000000000000..7e7d6c3865bc
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring_mux.c
@@ -0,0 +1,576 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include <linux/slab.h>
+#include <drm/drm_print.h>
+
+#include "amdgpu_ring_mux.h"
+#include "amdgpu_ring.h"
+#include "amdgpu.h"
+
+#define AMDGPU_MUX_RESUBMIT_JIFFIES_TIMEOUT (HZ / 2)
+#define AMDGPU_MAX_LAST_UNSIGNALED_THRESHOLD_US 10000
+
+static const struct ring_info {
+	unsigned int hw_pio;
+	const char *ring_name;
+} sw_ring_info[] = {
+	{ AMDGPU_RING_PRIO_DEFAULT, "gfx_low"},
+	{ AMDGPU_RING_PRIO_2, "gfx_high"},
+};
+
+static struct kmem_cache *amdgpu_mux_chunk_slab;
+
+static inline struct amdgpu_mux_entry *amdgpu_ring_mux_sw_entry(struct amdgpu_ring_mux *mux,
+								struct amdgpu_ring *ring)
+{
+	return ring->entry_index < mux->ring_entry_size ?
+			&mux->ring_entry[ring->entry_index] : NULL;
+}
+
+/* copy packages on sw ring range[begin, end) */
+static void amdgpu_ring_mux_copy_pkt_from_sw_ring(struct amdgpu_ring_mux *mux,
+						  struct amdgpu_ring *ring,
+						  u64 s_start, u64 s_end)
+{
+	u64 start, end;
+	struct amdgpu_ring *real_ring = mux->real_ring;
+
+	start = s_start & ring->buf_mask;
+	end = s_end & ring->buf_mask;
+
+	if (start == end) {
+		DRM_ERROR("no more data copied from sw ring\n");
+		return;
+	}
+	if (start > end) {
+		amdgpu_ring_alloc(real_ring, (ring->ring_size >> 2) + end - start);
+		amdgpu_ring_write_multiple(real_ring, (void *)&ring->ring[start],
+					   (ring->ring_size >> 2) - start);
+		amdgpu_ring_write_multiple(real_ring, (void *)&ring->ring[0], end);
+	} else {
+		amdgpu_ring_alloc(real_ring, end - start);
+		amdgpu_ring_write_multiple(real_ring, (void *)&ring->ring[start], end - start);
+	}
+}
+
+static void amdgpu_mux_resubmit_chunks(struct amdgpu_ring_mux *mux)
+{
+	struct amdgpu_mux_entry *e = NULL;
+	struct amdgpu_mux_chunk *chunk;
+	uint32_t seq, last_seq;
+	int i;
+
+	/*find low priority entries:*/
+	if (!mux->s_resubmit)
+		return;
+
+	for (i = 0; i < mux->num_ring_entries; i++) {
+		if (mux->ring_entry[i].ring->hw_prio <= AMDGPU_RING_PRIO_DEFAULT) {
+			e = &mux->ring_entry[i];
+			break;
+		}
+	}
+
+	if (!e) {
+		DRM_ERROR("%s no low priority ring found\n", __func__);
+		return;
+	}
+
+	last_seq = atomic_read(&e->ring->fence_drv.last_seq);
+	seq = mux->seqno_to_resubmit;
+	if (last_seq < seq) {
+		/*resubmit all the fences between (last_seq, seq]*/
+		list_for_each_entry(chunk, &e->list, entry) {
+			if (chunk->sync_seq > last_seq && chunk->sync_seq <= seq) {
+				amdgpu_fence_update_start_timestamp(e->ring,
+								    chunk->sync_seq,
+								    ktime_get());
+				if (chunk->sync_seq ==
+					le32_to_cpu(*(e->ring->fence_drv.cpu_addr + 2))) {
+					if (chunk->cntl_offset <= e->ring->buf_mask)
+						amdgpu_ring_patch_cntl(e->ring,
+								       chunk->cntl_offset);
+					if (chunk->ce_offset <= e->ring->buf_mask)
+						amdgpu_ring_patch_ce(e->ring, chunk->ce_offset);
+					if (chunk->de_offset <= e->ring->buf_mask)
+						amdgpu_ring_patch_de(e->ring, chunk->de_offset);
+				}
+				amdgpu_ring_mux_copy_pkt_from_sw_ring(mux, e->ring,
+								      chunk->start,
+								      chunk->end);
+				mux->wptr_resubmit = chunk->end;
+				amdgpu_ring_commit(mux->real_ring);
+			}
+		}
+	}
+
+	timer_delete(&mux->resubmit_timer);
+	mux->s_resubmit = false;
+}
+
+static void amdgpu_ring_mux_schedule_resubmit(struct amdgpu_ring_mux *mux)
+{
+	mod_timer(&mux->resubmit_timer, jiffies + AMDGPU_MUX_RESUBMIT_JIFFIES_TIMEOUT);
+}
+
+static void amdgpu_mux_resubmit_fallback(struct timer_list *t)
+{
+	struct amdgpu_ring_mux *mux = timer_container_of(mux, t,
+							 resubmit_timer);
+
+	if (!spin_trylock(&mux->lock)) {
+		amdgpu_ring_mux_schedule_resubmit(mux);
+		DRM_ERROR("reschedule resubmit\n");
+		return;
+	}
+	amdgpu_mux_resubmit_chunks(mux);
+	spin_unlock(&mux->lock);
+}
+
+int amdgpu_ring_mux_init(struct amdgpu_ring_mux *mux, struct amdgpu_ring *ring,
+			 unsigned int entry_size)
+{
+	mux->real_ring = ring;
+	mux->num_ring_entries = 0;
+
+	mux->ring_entry = kcalloc(entry_size, sizeof(struct amdgpu_mux_entry), GFP_KERNEL);
+	if (!mux->ring_entry)
+		return -ENOMEM;
+
+	mux->ring_entry_size = entry_size;
+	mux->s_resubmit = false;
+
+	amdgpu_mux_chunk_slab = KMEM_CACHE(amdgpu_mux_chunk, SLAB_HWCACHE_ALIGN);
+	if (!amdgpu_mux_chunk_slab) {
+		DRM_ERROR("create amdgpu_mux_chunk cache failed\n");
+		return -ENOMEM;
+	}
+
+	spin_lock_init(&mux->lock);
+	timer_setup(&mux->resubmit_timer, amdgpu_mux_resubmit_fallback, 0);
+
+	return 0;
+}
+
+void amdgpu_ring_mux_fini(struct amdgpu_ring_mux *mux)
+{
+	struct amdgpu_mux_entry *e;
+	struct amdgpu_mux_chunk *chunk, *chunk2;
+	int i;
+
+	for (i = 0; i < mux->num_ring_entries; i++) {
+		e = &mux->ring_entry[i];
+		list_for_each_entry_safe(chunk, chunk2, &e->list, entry) {
+			list_del(&chunk->entry);
+			kmem_cache_free(amdgpu_mux_chunk_slab, chunk);
+		}
+	}
+	kmem_cache_destroy(amdgpu_mux_chunk_slab);
+	kfree(mux->ring_entry);
+	mux->ring_entry = NULL;
+	mux->num_ring_entries = 0;
+	mux->ring_entry_size = 0;
+}
+
+int amdgpu_ring_mux_add_sw_ring(struct amdgpu_ring_mux *mux, struct amdgpu_ring *ring)
+{
+	struct amdgpu_mux_entry *e;
+
+	if (mux->num_ring_entries >= mux->ring_entry_size) {
+		DRM_ERROR("add sw ring exceeding max entry size\n");
+		return -ENOENT;
+	}
+
+	e = &mux->ring_entry[mux->num_ring_entries];
+	ring->entry_index = mux->num_ring_entries;
+	e->ring = ring;
+
+	INIT_LIST_HEAD(&e->list);
+	mux->num_ring_entries += 1;
+	return 0;
+}
+
+void amdgpu_ring_mux_set_wptr(struct amdgpu_ring_mux *mux, struct amdgpu_ring *ring, u64 wptr)
+{
+	struct amdgpu_mux_entry *e;
+
+	spin_lock(&mux->lock);
+
+	if (ring->hw_prio <= AMDGPU_RING_PRIO_DEFAULT)
+		amdgpu_mux_resubmit_chunks(mux);
+
+	e = amdgpu_ring_mux_sw_entry(mux, ring);
+	if (!e) {
+		DRM_ERROR("cannot find entry for sw ring\n");
+		spin_unlock(&mux->lock);
+		return;
+	}
+
+	/* We could skip this set wptr as preemption in process. */
+	if (ring->hw_prio <= AMDGPU_RING_PRIO_DEFAULT && mux->pending_trailing_fence_signaled) {
+		spin_unlock(&mux->lock);
+		return;
+	}
+
+	e->sw_cptr = e->sw_wptr;
+	/* Update cptr if the package already copied in resubmit functions */
+	if (ring->hw_prio <= AMDGPU_RING_PRIO_DEFAULT && e->sw_cptr < mux->wptr_resubmit)
+		e->sw_cptr = mux->wptr_resubmit;
+	e->sw_wptr = wptr;
+	e->start_ptr_in_hw_ring = mux->real_ring->wptr;
+
+	/* Skip copying for the packages already resubmitted.*/
+	if (ring->hw_prio > AMDGPU_RING_PRIO_DEFAULT || mux->wptr_resubmit < wptr) {
+		amdgpu_ring_mux_copy_pkt_from_sw_ring(mux, ring, e->sw_cptr, wptr);
+		e->end_ptr_in_hw_ring = mux->real_ring->wptr;
+		amdgpu_ring_commit(mux->real_ring);
+	} else {
+		e->end_ptr_in_hw_ring = mux->real_ring->wptr;
+	}
+	spin_unlock(&mux->lock);
+}
+
+u64 amdgpu_ring_mux_get_wptr(struct amdgpu_ring_mux *mux, struct amdgpu_ring *ring)
+{
+	struct amdgpu_mux_entry *e;
+
+	e = amdgpu_ring_mux_sw_entry(mux, ring);
+	if (!e) {
+		DRM_ERROR("cannot find entry for sw ring\n");
+		return 0;
+	}
+
+	return e->sw_wptr;
+}
+
+/**
+ * amdgpu_ring_mux_get_rptr - get the readptr of the software ring
+ * @mux: the multiplexer the software rings attach to
+ * @ring: the software ring of which we calculate the readptr
+ *
+ * The return value of the readptr is not precise while the other rings could
+ * write data onto the real ring buffer.After overwriting on the real ring, we
+ * can not decide if our packages have been excuted or not read yet. However,
+ * this function is only called by the tools such as umr to collect the latest
+ * packages for the hang analysis. We assume the hang happens near our latest
+ * submit. Thus we could use the following logic to give the clue:
+ * If the readptr is between start and end, then we return the copy pointer
+ * plus the distance from start to readptr. If the readptr is before start, we
+ * return the copy pointer. Lastly, if the readptr is past end, we return the
+ * write pointer.
+ */
+u64 amdgpu_ring_mux_get_rptr(struct amdgpu_ring_mux *mux, struct amdgpu_ring *ring)
+{
+	struct amdgpu_mux_entry *e;
+	u64 readp, offset, start, end;
+
+	e = amdgpu_ring_mux_sw_entry(mux, ring);
+	if (!e) {
+		DRM_ERROR("no sw entry found!\n");
+		return 0;
+	}
+
+	readp = amdgpu_ring_get_rptr(mux->real_ring);
+
+	start = e->start_ptr_in_hw_ring & mux->real_ring->buf_mask;
+	end = e->end_ptr_in_hw_ring & mux->real_ring->buf_mask;
+	if (start > end) {
+		if (readp <= end)
+			readp += mux->real_ring->ring_size >> 2;
+		end += mux->real_ring->ring_size >> 2;
+	}
+
+	if (start <= readp && readp <= end) {
+		offset = readp - start;
+		e->sw_rptr = (e->sw_cptr + offset) & ring->buf_mask;
+	} else if (readp < start) {
+		e->sw_rptr = e->sw_cptr;
+	} else {
+		/* end < readptr */
+		e->sw_rptr = e->sw_wptr;
+	}
+
+	return e->sw_rptr;
+}
+
+u64 amdgpu_sw_ring_get_rptr_gfx(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct amdgpu_ring_mux *mux = &adev->gfx.muxer;
+
+	WARN_ON(!ring->is_sw_ring);
+	return amdgpu_ring_mux_get_rptr(mux, ring);
+}
+
+u64 amdgpu_sw_ring_get_wptr_gfx(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct amdgpu_ring_mux *mux = &adev->gfx.muxer;
+
+	WARN_ON(!ring->is_sw_ring);
+	return amdgpu_ring_mux_get_wptr(mux, ring);
+}
+
+void amdgpu_sw_ring_set_wptr_gfx(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct amdgpu_ring_mux *mux = &adev->gfx.muxer;
+
+	WARN_ON(!ring->is_sw_ring);
+	amdgpu_ring_mux_set_wptr(mux, ring, ring->wptr);
+}
+
+/* Override insert_nop to prevent emitting nops to the software rings */
+void amdgpu_sw_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count)
+{
+	WARN_ON(!ring->is_sw_ring);
+}
+
+const char *amdgpu_sw_ring_name(int idx)
+{
+	return idx < ARRAY_SIZE(sw_ring_info) ?
+		sw_ring_info[idx].ring_name : NULL;
+}
+
+unsigned int amdgpu_sw_ring_priority(int idx)
+{
+	return idx < ARRAY_SIZE(sw_ring_info) ?
+		sw_ring_info[idx].hw_pio : AMDGPU_RING_PRIO_DEFAULT;
+}
+
+/*Scan on low prio rings to have unsignaled fence and high ring has no fence.*/
+static int amdgpu_mcbp_scan(struct amdgpu_ring_mux *mux)
+{
+	struct amdgpu_ring *ring;
+	int i, need_preempt;
+
+	need_preempt = 0;
+	for (i = 0; i < mux->num_ring_entries; i++) {
+		ring = mux->ring_entry[i].ring;
+		if (ring->hw_prio > AMDGPU_RING_PRIO_DEFAULT &&
+		    amdgpu_fence_count_emitted(ring) > 0)
+			return 0;
+		if (ring->hw_prio <= AMDGPU_RING_PRIO_DEFAULT &&
+		    amdgpu_fence_last_unsignaled_time_us(ring) >
+		    AMDGPU_MAX_LAST_UNSIGNALED_THRESHOLD_US)
+			need_preempt = 1;
+	}
+	return need_preempt && !mux->s_resubmit;
+}
+
+/* Trigger Mid-Command Buffer Preemption (MCBP) and find if we need to resubmit. */
+static int amdgpu_mcbp_trigger_preempt(struct amdgpu_ring_mux *mux)
+{
+	int r;
+
+	spin_lock(&mux->lock);
+	mux->pending_trailing_fence_signaled = true;
+	r = amdgpu_ring_preempt_ib(mux->real_ring);
+	spin_unlock(&mux->lock);
+	return r;
+}
+
+void amdgpu_sw_ring_ib_begin(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct amdgpu_ring_mux *mux = &adev->gfx.muxer;
+
+	WARN_ON(!ring->is_sw_ring);
+	if (adev->gfx.mcbp && ring->hw_prio > AMDGPU_RING_PRIO_DEFAULT) {
+		if (amdgpu_mcbp_scan(mux) > 0)
+			amdgpu_mcbp_trigger_preempt(mux);
+		return;
+	}
+
+	amdgpu_ring_mux_start_ib(mux, ring);
+}
+
+void amdgpu_sw_ring_ib_end(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct amdgpu_ring_mux *mux = &adev->gfx.muxer;
+
+	WARN_ON(!ring->is_sw_ring);
+	if (adev->gfx.mcbp && ring->hw_prio > AMDGPU_RING_PRIO_DEFAULT)
+		return;
+	amdgpu_ring_mux_end_ib(mux, ring);
+}
+
+void amdgpu_sw_ring_ib_mark_offset(struct amdgpu_ring *ring, enum amdgpu_ring_mux_offset_type type)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct amdgpu_ring_mux *mux = &adev->gfx.muxer;
+	unsigned offset;
+
+	if (ring->hw_prio > AMDGPU_RING_PRIO_DEFAULT)
+		return;
+
+	offset = ring->wptr & ring->buf_mask;
+
+	amdgpu_ring_mux_ib_mark_offset(mux, ring, offset, type);
+}
+
+void amdgpu_ring_mux_start_ib(struct amdgpu_ring_mux *mux, struct amdgpu_ring *ring)
+{
+	struct amdgpu_mux_entry *e;
+	struct amdgpu_mux_chunk *chunk;
+
+	spin_lock(&mux->lock);
+	amdgpu_mux_resubmit_chunks(mux);
+	spin_unlock(&mux->lock);
+
+	e = amdgpu_ring_mux_sw_entry(mux, ring);
+	if (!e) {
+		DRM_ERROR("cannot find entry!\n");
+		return;
+	}
+
+	chunk = kmem_cache_alloc(amdgpu_mux_chunk_slab, GFP_KERNEL);
+	if (!chunk) {
+		DRM_ERROR("alloc amdgpu_mux_chunk_slab failed\n");
+		return;
+	}
+
+	chunk->start = ring->wptr;
+	/* the initialized value used to check if they are set by the ib submission*/
+	chunk->cntl_offset = ring->buf_mask + 1;
+	chunk->de_offset = ring->buf_mask + 1;
+	chunk->ce_offset = ring->buf_mask + 1;
+	list_add_tail(&chunk->entry, &e->list);
+}
+
+static void scan_and_remove_signaled_chunk(struct amdgpu_ring_mux *mux, struct amdgpu_ring *ring)
+{
+	uint32_t last_seq = 0;
+	struct amdgpu_mux_entry *e;
+	struct amdgpu_mux_chunk *chunk, *tmp;
+
+	e = amdgpu_ring_mux_sw_entry(mux, ring);
+	if (!e) {
+		DRM_ERROR("cannot find entry!\n");
+		return;
+	}
+
+	last_seq = atomic_read(&ring->fence_drv.last_seq);
+
+	list_for_each_entry_safe(chunk, tmp, &e->list, entry) {
+		if (chunk->sync_seq <= last_seq) {
+			list_del(&chunk->entry);
+			kmem_cache_free(amdgpu_mux_chunk_slab, chunk);
+		}
+	}
+}
+
+void amdgpu_ring_mux_ib_mark_offset(struct amdgpu_ring_mux *mux,
+				    struct amdgpu_ring *ring, u64 offset,
+				    enum amdgpu_ring_mux_offset_type type)
+{
+	struct amdgpu_mux_entry *e;
+	struct amdgpu_mux_chunk *chunk;
+
+	e = amdgpu_ring_mux_sw_entry(mux, ring);
+	if (!e) {
+		DRM_ERROR("cannot find entry!\n");
+		return;
+	}
+
+	chunk = list_last_entry(&e->list, struct amdgpu_mux_chunk, entry);
+	if (!chunk) {
+		DRM_ERROR("cannot find chunk!\n");
+		return;
+	}
+
+	switch (type) {
+	case AMDGPU_MUX_OFFSET_TYPE_CONTROL:
+		chunk->cntl_offset = offset;
+		break;
+	case AMDGPU_MUX_OFFSET_TYPE_DE:
+		chunk->de_offset = offset;
+		break;
+	case AMDGPU_MUX_OFFSET_TYPE_CE:
+		chunk->ce_offset = offset;
+		break;
+	default:
+		DRM_ERROR("invalid type (%d)\n", type);
+		break;
+	}
+}
+
+void amdgpu_ring_mux_end_ib(struct amdgpu_ring_mux *mux, struct amdgpu_ring *ring)
+{
+	struct amdgpu_mux_entry *e;
+	struct amdgpu_mux_chunk *chunk;
+
+	e = amdgpu_ring_mux_sw_entry(mux, ring);
+	if (!e) {
+		DRM_ERROR("cannot find entry!\n");
+		return;
+	}
+
+	chunk = list_last_entry(&e->list, struct amdgpu_mux_chunk, entry);
+	if (!chunk) {
+		DRM_ERROR("cannot find chunk!\n");
+		return;
+	}
+
+	chunk->end = ring->wptr;
+	chunk->sync_seq = READ_ONCE(ring->fence_drv.sync_seq);
+
+	scan_and_remove_signaled_chunk(mux, ring);
+}
+
+bool amdgpu_mcbp_handle_trailing_fence_irq(struct amdgpu_ring_mux *mux)
+{
+	struct amdgpu_mux_entry *e;
+	struct amdgpu_ring *ring = NULL;
+	int i;
+
+	if (!mux->pending_trailing_fence_signaled)
+		return false;
+
+	if (mux->real_ring->trail_seq != le32_to_cpu(*mux->real_ring->trail_fence_cpu_addr))
+		return false;
+
+	for (i = 0; i < mux->num_ring_entries; i++) {
+		e = &mux->ring_entry[i];
+		if (e->ring->hw_prio <= AMDGPU_RING_PRIO_DEFAULT) {
+			ring = e->ring;
+			break;
+		}
+	}
+
+	if (!ring) {
+		DRM_ERROR("cannot find low priority ring\n");
+		return false;
+	}
+
+	amdgpu_fence_process(ring);
+	if (amdgpu_fence_count_emitted(ring) > 0) {
+		mux->s_resubmit = true;
+		mux->seqno_to_resubmit = ring->fence_drv.sync_seq;
+		amdgpu_ring_mux_schedule_resubmit(mux);
+	}
+
+	mux->pending_trailing_fence_signaled = false;
+	return true;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring_mux.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring_mux.h
new file mode 100644
index 000000000000..d3186b570b82
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring_mux.h
@@ -0,0 +1,127 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __AMDGPU_RING_MUX__
+#define __AMDGPU_RING_MUX__
+
+#include <linux/timer.h>
+#include <linux/spinlock.h>
+#include "amdgpu_ring.h"
+
+struct amdgpu_ring;
+
+/**
+ * struct amdgpu_mux_entry - the entry recording software rings copying information.
+ * @ring: the pointer to the software ring.
+ * @start_ptr_in_hw_ring: last start location copied to in the hardware ring.
+ * @end_ptr_in_hw_ring: last end location copied to in the hardware ring.
+ * @sw_cptr: the position of the copy pointer in the sw ring.
+ * @sw_rptr: the read pointer in software ring.
+ * @sw_wptr: the write pointer in software ring.
+ * @list: list head for amdgpu_mux_chunk
+ */
+struct amdgpu_mux_entry {
+	struct amdgpu_ring      *ring;
+	u64                     start_ptr_in_hw_ring;
+	u64                     end_ptr_in_hw_ring;
+	u64                     sw_cptr;
+	u64                     sw_rptr;
+	u64                     sw_wptr;
+	struct list_head        list;
+};
+
+enum amdgpu_ring_mux_offset_type {
+	AMDGPU_MUX_OFFSET_TYPE_CONTROL,
+	AMDGPU_MUX_OFFSET_TYPE_DE,
+	AMDGPU_MUX_OFFSET_TYPE_CE,
+};
+
+enum ib_complete_status {
+	/* IB not started/reset value, default value. */
+	IB_COMPLETION_STATUS_DEFAULT = 0,
+	/* IB preempted, started but not completed. */
+	IB_COMPLETION_STATUS_PREEMPTED = 1,
+	/* IB completed. */
+	IB_COMPLETION_STATUS_COMPLETED = 2,
+};
+
+struct amdgpu_ring_mux {
+	struct amdgpu_ring      *real_ring;
+
+	struct amdgpu_mux_entry *ring_entry;
+	unsigned int            num_ring_entries;
+	unsigned int            ring_entry_size;
+	/*the lock for copy data from different software rings*/
+	spinlock_t              lock;
+	bool                    s_resubmit;
+	uint32_t                seqno_to_resubmit;
+	u64                     wptr_resubmit;
+	struct timer_list       resubmit_timer;
+
+	bool                    pending_trailing_fence_signaled;
+};
+
+/**
+ * struct amdgpu_mux_chunk - save the location of indirect buffer's package on softare rings.
+ * @entry: the list entry.
+ * @sync_seq: the fence seqno related with the saved IB.
+ * @start:- start location on the software ring.
+ * @end:- end location on the software ring.
+ * @control_offset:- the PRE_RESUME bit position used for resubmission.
+ * @de_offset:- the anchor in write_data for de meta of resubmission.
+ * @ce_offset:- the anchor in write_data for ce meta of resubmission.
+ */
+struct amdgpu_mux_chunk {
+	struct list_head        entry;
+	uint32_t                sync_seq;
+	u64                     start;
+	u64                     end;
+	u64                     cntl_offset;
+	u64                     de_offset;
+	u64                     ce_offset;
+};
+
+int amdgpu_ring_mux_init(struct amdgpu_ring_mux *mux, struct amdgpu_ring *ring,
+			 unsigned int entry_size);
+void amdgpu_ring_mux_fini(struct amdgpu_ring_mux *mux);
+int amdgpu_ring_mux_add_sw_ring(struct amdgpu_ring_mux *mux, struct amdgpu_ring *ring);
+void amdgpu_ring_mux_set_wptr(struct amdgpu_ring_mux *mux, struct amdgpu_ring *ring, u64 wptr);
+u64 amdgpu_ring_mux_get_wptr(struct amdgpu_ring_mux *mux, struct amdgpu_ring *ring);
+u64 amdgpu_ring_mux_get_rptr(struct amdgpu_ring_mux *mux, struct amdgpu_ring *ring);
+void amdgpu_ring_mux_start_ib(struct amdgpu_ring_mux *mux, struct amdgpu_ring *ring);
+void amdgpu_ring_mux_end_ib(struct amdgpu_ring_mux *mux, struct amdgpu_ring *ring);
+void amdgpu_ring_mux_ib_mark_offset(struct amdgpu_ring_mux *mux, struct amdgpu_ring *ring,
+				    u64 offset, enum amdgpu_ring_mux_offset_type type);
+bool amdgpu_mcbp_handle_trailing_fence_irq(struct amdgpu_ring_mux *mux);
+
+u64 amdgpu_sw_ring_get_rptr_gfx(struct amdgpu_ring *ring);
+u64 amdgpu_sw_ring_get_wptr_gfx(struct amdgpu_ring *ring);
+void amdgpu_sw_ring_set_wptr_gfx(struct amdgpu_ring *ring);
+void amdgpu_sw_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count);
+void amdgpu_sw_ring_ib_begin(struct amdgpu_ring *ring);
+void amdgpu_sw_ring_ib_end(struct amdgpu_ring *ring);
+void amdgpu_sw_ring_ib_mark_offset(struct amdgpu_ring *ring, enum amdgpu_ring_mux_offset_type type);
+const char *amdgpu_sw_ring_name(int idx);
+unsigned int amdgpu_sw_ring_priority(int idx);
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.c
index 6373bfb47d55..5aa830a02d80 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.c
@@ -31,12 +31,13 @@
  * amdgpu_gfx_rlc_enter_safe_mode - Set RLC into safe mode
  *
  * @adev: amdgpu_device pointer
+ * @xcc_id: xcc accelerated compute core id
  *
  * Set RLC enter into safe mode if RLC is enabled and haven't in safe mode.
  */
-void amdgpu_gfx_rlc_enter_safe_mode(struct amdgpu_device *adev)
+void amdgpu_gfx_rlc_enter_safe_mode(struct amdgpu_device *adev, int xcc_id)
 {
-	if (adev->gfx.rlc.in_safe_mode)
+	if (adev->gfx.rlc.in_safe_mode[xcc_id])
 		return;
 
 	/* if RLC is not enabled, do nothing */
@@ -46,8 +47,8 @@ void amdgpu_gfx_rlc_enter_safe_mode(struct amdgpu_device *adev)
 	if (adev->cg_flags &
 	    (AMD_CG_SUPPORT_GFX_CGCG | AMD_CG_SUPPORT_GFX_MGCG |
 	     AMD_CG_SUPPORT_GFX_3D_CGCG)) {
-		adev->gfx.rlc.funcs->set_safe_mode(adev);
-		adev->gfx.rlc.in_safe_mode = true;
+		adev->gfx.rlc.funcs->set_safe_mode(adev, xcc_id);
+		adev->gfx.rlc.in_safe_mode[xcc_id] = true;
 	}
 }
 
@@ -55,12 +56,13 @@ void amdgpu_gfx_rlc_enter_safe_mode(struct amdgpu_device *adev)
  * amdgpu_gfx_rlc_exit_safe_mode - Set RLC out of safe mode
  *
  * @adev: amdgpu_device pointer
+ * @xcc_id: xcc accelerated compute core id
  *
  * Set RLC exit safe mode if RLC is enabled and have entered into safe mode.
  */
-void amdgpu_gfx_rlc_exit_safe_mode(struct amdgpu_device *adev)
+void amdgpu_gfx_rlc_exit_safe_mode(struct amdgpu_device *adev, int xcc_id)
 {
-	if (!(adev->gfx.rlc.in_safe_mode))
+	if (!(adev->gfx.rlc.in_safe_mode[xcc_id]))
 		return;
 
 	/* if RLC is not enabled, do nothing */
@@ -70,8 +72,8 @@ void amdgpu_gfx_rlc_exit_safe_mode(struct amdgpu_device *adev)
 	if (adev->cg_flags &
 	    (AMD_CG_SUPPORT_GFX_CGCG | AMD_CG_SUPPORT_GFX_MGCG |
 	     AMD_CG_SUPPORT_GFX_3D_CGCG)) {
-		adev->gfx.rlc.funcs->unset_safe_mode(adev);
-		adev->gfx.rlc.in_safe_mode = false;
+		adev->gfx.rlc.funcs->unset_safe_mode(adev, xcc_id);
+		adev->gfx.rlc.in_safe_mode[xcc_id] = false;
 	}
 }
 
@@ -87,13 +89,14 @@ void amdgpu_gfx_rlc_exit_safe_mode(struct amdgpu_device *adev)
 int amdgpu_gfx_rlc_init_sr(struct amdgpu_device *adev, u32 dws)
 {
 	const u32 *src_ptr;
-	volatile u32 *dst_ptr;
+	u32 *dst_ptr;
 	u32 i;
 	int r;
 
 	/* allocate save restore block */
 	r = amdgpu_bo_create_reserved(adev, dws * 4, PAGE_SIZE,
-				      AMDGPU_GEM_DOMAIN_VRAM,
+				      AMDGPU_GEM_DOMAIN_VRAM |
+				      AMDGPU_GEM_DOMAIN_GTT,
 				      &adev->gfx.rlc.save_restore_obj,
 				      &adev->gfx.rlc.save_restore_gpu_addr,
 				      (void **)&adev->gfx.rlc.sr_ptr);
@@ -130,7 +133,8 @@ int amdgpu_gfx_rlc_init_csb(struct amdgpu_device *adev)
 	/* allocate clear state block */
 	adev->gfx.rlc.clear_state_size = dws = adev->gfx.rlc.funcs->get_csb_size(adev);
 	r = amdgpu_bo_create_kernel(adev, dws * 4, PAGE_SIZE,
-				      AMDGPU_GEM_DOMAIN_VRAM,
+				      AMDGPU_GEM_DOMAIN_VRAM |
+				      AMDGPU_GEM_DOMAIN_GTT,
 				      &adev->gfx.rlc.clear_state_obj,
 				      &adev->gfx.rlc.clear_state_gpu_addr,
 				      (void **)&adev->gfx.rlc.cs_ptr);
@@ -156,7 +160,8 @@ int amdgpu_gfx_rlc_init_cpt(struct amdgpu_device *adev)
 	int r;
 
 	r = amdgpu_bo_create_reserved(adev, adev->gfx.rlc.cp_table_size,
-				      PAGE_SIZE, AMDGPU_GEM_DOMAIN_VRAM,
+				      PAGE_SIZE, AMDGPU_GEM_DOMAIN_VRAM |
+				      AMDGPU_GEM_DOMAIN_GTT,
 				      &adev->gfx.rlc.cp_table_obj,
 				      &adev->gfx.rlc.cp_table_gpu_addr,
 				      (void **)&adev->gfx.rlc.cp_table_ptr);
@@ -184,7 +189,7 @@ int amdgpu_gfx_rlc_init_cpt(struct amdgpu_device *adev)
 void amdgpu_gfx_rlc_setup_cp_table(struct amdgpu_device *adev)
 {
 	const __le32 *fw_data;
-	volatile u32 *dst_ptr;
+	u32 *dst_ptr;
 	int me, i, max_me;
 	u32 bo_offset = 0;
 	u32 table_offset, table_size;
@@ -236,7 +241,7 @@ void amdgpu_gfx_rlc_setup_cp_table(struct amdgpu_device *adev)
 			table_size = le32_to_cpu(hdr->jt_size);
 		}
 
-		for (i = 0; i < table_size; i ++) {
+		for (i = 0; i < table_size; i++) {
 			dst_ptr[bo_offset + i] =
 				cpu_to_le32(le32_to_cpu(fw_data[table_offset + i]));
 		}
@@ -272,3 +277,274 @@ void amdgpu_gfx_rlc_fini(struct amdgpu_device *adev)
 			      &adev->gfx.rlc.cp_table_gpu_addr,
 			      (void **)&adev->gfx.rlc.cp_table_ptr);
 }
+
+static int amdgpu_gfx_rlc_init_microcode_v2_0(struct amdgpu_device *adev)
+{
+	const struct common_firmware_header *common_hdr;
+	const struct rlc_firmware_header_v2_0 *rlc_hdr;
+	struct amdgpu_firmware_info *info;
+	unsigned int *tmp;
+	unsigned int i;
+
+	rlc_hdr = (const struct rlc_firmware_header_v2_0 *)adev->gfx.rlc_fw->data;
+
+	adev->gfx.rlc_fw_version = le32_to_cpu(rlc_hdr->header.ucode_version);
+	adev->gfx.rlc_feature_version = le32_to_cpu(rlc_hdr->ucode_feature_version);
+	adev->gfx.rlc.save_and_restore_offset =
+		le32_to_cpu(rlc_hdr->save_and_restore_offset);
+	adev->gfx.rlc.clear_state_descriptor_offset =
+		le32_to_cpu(rlc_hdr->clear_state_descriptor_offset);
+	adev->gfx.rlc.avail_scratch_ram_locations =
+		le32_to_cpu(rlc_hdr->avail_scratch_ram_locations);
+	adev->gfx.rlc.reg_restore_list_size =
+		le32_to_cpu(rlc_hdr->reg_restore_list_size);
+	adev->gfx.rlc.reg_list_format_start =
+		le32_to_cpu(rlc_hdr->reg_list_format_start);
+	adev->gfx.rlc.reg_list_format_separate_start =
+		le32_to_cpu(rlc_hdr->reg_list_format_separate_start);
+	adev->gfx.rlc.starting_offsets_start =
+		le32_to_cpu(rlc_hdr->starting_offsets_start);
+	adev->gfx.rlc.reg_list_format_size_bytes =
+		le32_to_cpu(rlc_hdr->reg_list_format_size_bytes);
+	adev->gfx.rlc.reg_list_size_bytes =
+		le32_to_cpu(rlc_hdr->reg_list_size_bytes);
+	adev->gfx.rlc.register_list_format =
+		kmalloc(adev->gfx.rlc.reg_list_format_size_bytes +
+			adev->gfx.rlc.reg_list_size_bytes, GFP_KERNEL);
+	if (!adev->gfx.rlc.register_list_format) {
+		dev_err(adev->dev, "failed to allocate memory for rlc register_list_format\n");
+		return -ENOMEM;
+	}
+
+	tmp = (unsigned int *)((uintptr_t)rlc_hdr +
+			le32_to_cpu(rlc_hdr->reg_list_format_array_offset_bytes));
+	for (i = 0 ; i < (rlc_hdr->reg_list_format_size_bytes >> 2); i++)
+		adev->gfx.rlc.register_list_format[i] = le32_to_cpu(tmp[i]);
+
+	adev->gfx.rlc.register_restore = adev->gfx.rlc.register_list_format + i;
+
+	tmp = (unsigned int *)((uintptr_t)rlc_hdr +
+			le32_to_cpu(rlc_hdr->reg_list_array_offset_bytes));
+	for (i = 0 ; i < (rlc_hdr->reg_list_size_bytes >> 2); i++)
+		adev->gfx.rlc.register_restore[i] = le32_to_cpu(tmp[i]);
+
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
+		info = &adev->firmware.ucode[AMDGPU_UCODE_ID_RLC_G];
+		info->ucode_id = AMDGPU_UCODE_ID_RLC_G;
+		info->fw = adev->gfx.rlc_fw;
+		if (info->fw) {
+			common_hdr = (const struct common_firmware_header *)info->fw->data;
+			adev->firmware.fw_size +=
+				ALIGN(le32_to_cpu(common_hdr->ucode_size_bytes), PAGE_SIZE);
+		}
+	}
+
+	return 0;
+}
+
+static void amdgpu_gfx_rlc_init_microcode_v2_1(struct amdgpu_device *adev)
+{
+	const struct rlc_firmware_header_v2_1 *rlc_hdr;
+	struct amdgpu_firmware_info *info;
+
+	rlc_hdr = (const struct rlc_firmware_header_v2_1 *)adev->gfx.rlc_fw->data;
+	adev->gfx.rlc_srlc_fw_version = le32_to_cpu(rlc_hdr->save_restore_list_cntl_ucode_ver);
+	adev->gfx.rlc_srlc_feature_version = le32_to_cpu(rlc_hdr->save_restore_list_cntl_feature_ver);
+	adev->gfx.rlc.save_restore_list_cntl_size_bytes = le32_to_cpu(rlc_hdr->save_restore_list_cntl_size_bytes);
+	adev->gfx.rlc.save_restore_list_cntl = (u8 *)rlc_hdr + le32_to_cpu(rlc_hdr->save_restore_list_cntl_offset_bytes);
+	adev->gfx.rlc_srlg_fw_version = le32_to_cpu(rlc_hdr->save_restore_list_gpm_ucode_ver);
+	adev->gfx.rlc_srlg_feature_version = le32_to_cpu(rlc_hdr->save_restore_list_gpm_feature_ver);
+	adev->gfx.rlc.save_restore_list_gpm_size_bytes = le32_to_cpu(rlc_hdr->save_restore_list_gpm_size_bytes);
+	adev->gfx.rlc.save_restore_list_gpm = (u8 *)rlc_hdr + le32_to_cpu(rlc_hdr->save_restore_list_gpm_offset_bytes);
+	adev->gfx.rlc_srls_fw_version = le32_to_cpu(rlc_hdr->save_restore_list_srm_ucode_ver);
+	adev->gfx.rlc_srls_feature_version = le32_to_cpu(rlc_hdr->save_restore_list_srm_feature_ver);
+	adev->gfx.rlc.save_restore_list_srm_size_bytes = le32_to_cpu(rlc_hdr->save_restore_list_srm_size_bytes);
+	adev->gfx.rlc.save_restore_list_srm = (u8 *)rlc_hdr + le32_to_cpu(rlc_hdr->save_restore_list_srm_offset_bytes);
+	adev->gfx.rlc.reg_list_format_direct_reg_list_length =
+		le32_to_cpu(rlc_hdr->reg_list_format_direct_reg_list_length);
+
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
+		if (adev->gfx.rlc.save_restore_list_cntl_size_bytes) {
+			info = &adev->firmware.ucode[AMDGPU_UCODE_ID_RLC_RESTORE_LIST_CNTL];
+			info->ucode_id = AMDGPU_UCODE_ID_RLC_RESTORE_LIST_CNTL;
+			info->fw = adev->gfx.rlc_fw;
+			adev->firmware.fw_size +=
+				ALIGN(adev->gfx.rlc.save_restore_list_cntl_size_bytes, PAGE_SIZE);
+		}
+
+		if (adev->gfx.rlc.save_restore_list_gpm_size_bytes) {
+			info = &adev->firmware.ucode[AMDGPU_UCODE_ID_RLC_RESTORE_LIST_GPM_MEM];
+			info->ucode_id = AMDGPU_UCODE_ID_RLC_RESTORE_LIST_GPM_MEM;
+			info->fw = adev->gfx.rlc_fw;
+			adev->firmware.fw_size +=
+				ALIGN(adev->gfx.rlc.save_restore_list_gpm_size_bytes, PAGE_SIZE);
+		}
+
+		if (adev->gfx.rlc.save_restore_list_srm_size_bytes) {
+			info = &adev->firmware.ucode[AMDGPU_UCODE_ID_RLC_RESTORE_LIST_SRM_MEM];
+			info->ucode_id = AMDGPU_UCODE_ID_RLC_RESTORE_LIST_SRM_MEM;
+			info->fw = adev->gfx.rlc_fw;
+			adev->firmware.fw_size +=
+				ALIGN(adev->gfx.rlc.save_restore_list_srm_size_bytes, PAGE_SIZE);
+		}
+	}
+}
+
+static void amdgpu_gfx_rlc_init_microcode_v2_2(struct amdgpu_device *adev)
+{
+	const struct rlc_firmware_header_v2_2 *rlc_hdr;
+	struct amdgpu_firmware_info *info;
+
+	rlc_hdr = (const struct rlc_firmware_header_v2_2 *)adev->gfx.rlc_fw->data;
+	adev->gfx.rlc.rlc_iram_ucode_size_bytes = le32_to_cpu(rlc_hdr->rlc_iram_ucode_size_bytes);
+	adev->gfx.rlc.rlc_iram_ucode = (u8 *)rlc_hdr + le32_to_cpu(rlc_hdr->rlc_iram_ucode_offset_bytes);
+	adev->gfx.rlc.rlc_dram_ucode_size_bytes = le32_to_cpu(rlc_hdr->rlc_dram_ucode_size_bytes);
+	adev->gfx.rlc.rlc_dram_ucode = (u8 *)rlc_hdr + le32_to_cpu(rlc_hdr->rlc_dram_ucode_offset_bytes);
+
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
+		if (adev->gfx.rlc.rlc_iram_ucode_size_bytes) {
+			info = &adev->firmware.ucode[AMDGPU_UCODE_ID_RLC_IRAM];
+			info->ucode_id = AMDGPU_UCODE_ID_RLC_IRAM;
+			info->fw = adev->gfx.rlc_fw;
+			adev->firmware.fw_size +=
+				ALIGN(adev->gfx.rlc.rlc_iram_ucode_size_bytes, PAGE_SIZE);
+		}
+
+		if (adev->gfx.rlc.rlc_dram_ucode_size_bytes) {
+			info = &adev->firmware.ucode[AMDGPU_UCODE_ID_RLC_DRAM];
+			info->ucode_id = AMDGPU_UCODE_ID_RLC_DRAM;
+			info->fw = adev->gfx.rlc_fw;
+			adev->firmware.fw_size +=
+				ALIGN(adev->gfx.rlc.rlc_dram_ucode_size_bytes, PAGE_SIZE);
+		}
+	}
+}
+
+static void amdgpu_gfx_rlc_init_microcode_v2_3(struct amdgpu_device *adev)
+{
+	const struct rlc_firmware_header_v2_3 *rlc_hdr;
+	struct amdgpu_firmware_info *info;
+
+	rlc_hdr = (const struct rlc_firmware_header_v2_3 *)adev->gfx.rlc_fw->data;
+	adev->gfx.rlcp_ucode_version = le32_to_cpu(rlc_hdr->rlcp_ucode_version);
+	adev->gfx.rlcp_ucode_feature_version = le32_to_cpu(rlc_hdr->rlcp_ucode_feature_version);
+	adev->gfx.rlc.rlcp_ucode_size_bytes = le32_to_cpu(rlc_hdr->rlcp_ucode_size_bytes);
+	adev->gfx.rlc.rlcp_ucode = (u8 *)rlc_hdr + le32_to_cpu(rlc_hdr->rlcp_ucode_offset_bytes);
+
+	adev->gfx.rlcv_ucode_version = le32_to_cpu(rlc_hdr->rlcv_ucode_version);
+	adev->gfx.rlcv_ucode_feature_version = le32_to_cpu(rlc_hdr->rlcv_ucode_feature_version);
+	adev->gfx.rlc.rlcv_ucode_size_bytes = le32_to_cpu(rlc_hdr->rlcv_ucode_size_bytes);
+	adev->gfx.rlc.rlcv_ucode = (u8 *)rlc_hdr + le32_to_cpu(rlc_hdr->rlcv_ucode_offset_bytes);
+
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
+		if (adev->gfx.rlc.rlcp_ucode_size_bytes) {
+			info = &adev->firmware.ucode[AMDGPU_UCODE_ID_RLC_P];
+			info->ucode_id = AMDGPU_UCODE_ID_RLC_P;
+			info->fw = adev->gfx.rlc_fw;
+			adev->firmware.fw_size +=
+				ALIGN(adev->gfx.rlc.rlcp_ucode_size_bytes, PAGE_SIZE);
+		}
+
+		if (adev->gfx.rlc.rlcv_ucode_size_bytes) {
+			info = &adev->firmware.ucode[AMDGPU_UCODE_ID_RLC_V];
+			info->ucode_id = AMDGPU_UCODE_ID_RLC_V;
+			info->fw = adev->gfx.rlc_fw;
+			adev->firmware.fw_size +=
+				ALIGN(adev->gfx.rlc.rlcv_ucode_size_bytes, PAGE_SIZE);
+		}
+	}
+}
+
+static void amdgpu_gfx_rlc_init_microcode_v2_4(struct amdgpu_device *adev)
+{
+	const struct rlc_firmware_header_v2_4 *rlc_hdr;
+	struct amdgpu_firmware_info *info;
+
+	rlc_hdr = (const struct rlc_firmware_header_v2_4 *)adev->gfx.rlc_fw->data;
+	adev->gfx.rlc.global_tap_delays_ucode_size_bytes = le32_to_cpu(rlc_hdr->global_tap_delays_ucode_size_bytes);
+	adev->gfx.rlc.global_tap_delays_ucode = (u8 *)rlc_hdr + le32_to_cpu(rlc_hdr->global_tap_delays_ucode_offset_bytes);
+	adev->gfx.rlc.se0_tap_delays_ucode_size_bytes = le32_to_cpu(rlc_hdr->se0_tap_delays_ucode_size_bytes);
+	adev->gfx.rlc.se0_tap_delays_ucode = (u8 *)rlc_hdr + le32_to_cpu(rlc_hdr->se0_tap_delays_ucode_offset_bytes);
+	adev->gfx.rlc.se1_tap_delays_ucode_size_bytes = le32_to_cpu(rlc_hdr->se1_tap_delays_ucode_size_bytes);
+	adev->gfx.rlc.se1_tap_delays_ucode = (u8 *)rlc_hdr + le32_to_cpu(rlc_hdr->se1_tap_delays_ucode_offset_bytes);
+	adev->gfx.rlc.se2_tap_delays_ucode_size_bytes = le32_to_cpu(rlc_hdr->se2_tap_delays_ucode_size_bytes);
+	adev->gfx.rlc.se2_tap_delays_ucode = (u8 *)rlc_hdr + le32_to_cpu(rlc_hdr->se2_tap_delays_ucode_offset_bytes);
+	adev->gfx.rlc.se3_tap_delays_ucode_size_bytes = le32_to_cpu(rlc_hdr->se3_tap_delays_ucode_size_bytes);
+	adev->gfx.rlc.se3_tap_delays_ucode = (u8 *)rlc_hdr + le32_to_cpu(rlc_hdr->se3_tap_delays_ucode_offset_bytes);
+
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
+		if (adev->gfx.rlc.global_tap_delays_ucode_size_bytes) {
+			info = &adev->firmware.ucode[AMDGPU_UCODE_ID_GLOBAL_TAP_DELAYS];
+			info->ucode_id = AMDGPU_UCODE_ID_GLOBAL_TAP_DELAYS;
+			info->fw = adev->gfx.rlc_fw;
+			adev->firmware.fw_size +=
+				ALIGN(adev->gfx.rlc.global_tap_delays_ucode_size_bytes, PAGE_SIZE);
+		}
+
+		if (adev->gfx.rlc.se0_tap_delays_ucode_size_bytes) {
+			info = &adev->firmware.ucode[AMDGPU_UCODE_ID_SE0_TAP_DELAYS];
+			info->ucode_id = AMDGPU_UCODE_ID_SE0_TAP_DELAYS;
+			info->fw = adev->gfx.rlc_fw;
+			adev->firmware.fw_size +=
+				ALIGN(adev->gfx.rlc.se0_tap_delays_ucode_size_bytes, PAGE_SIZE);
+		}
+
+		if (adev->gfx.rlc.se1_tap_delays_ucode_size_bytes) {
+			info = &adev->firmware.ucode[AMDGPU_UCODE_ID_SE1_TAP_DELAYS];
+			info->ucode_id = AMDGPU_UCODE_ID_SE1_TAP_DELAYS;
+			info->fw = adev->gfx.rlc_fw;
+			adev->firmware.fw_size +=
+				ALIGN(adev->gfx.rlc.se1_tap_delays_ucode_size_bytes, PAGE_SIZE);
+		}
+
+		if (adev->gfx.rlc.se2_tap_delays_ucode_size_bytes) {
+			info = &adev->firmware.ucode[AMDGPU_UCODE_ID_SE2_TAP_DELAYS];
+			info->ucode_id = AMDGPU_UCODE_ID_SE2_TAP_DELAYS;
+			info->fw = adev->gfx.rlc_fw;
+			adev->firmware.fw_size +=
+				ALIGN(adev->gfx.rlc.se2_tap_delays_ucode_size_bytes, PAGE_SIZE);
+		}
+
+		if (adev->gfx.rlc.se3_tap_delays_ucode_size_bytes) {
+			info = &adev->firmware.ucode[AMDGPU_UCODE_ID_SE3_TAP_DELAYS];
+			info->ucode_id = AMDGPU_UCODE_ID_SE3_TAP_DELAYS;
+			info->fw = adev->gfx.rlc_fw;
+			adev->firmware.fw_size +=
+				ALIGN(adev->gfx.rlc.se3_tap_delays_ucode_size_bytes, PAGE_SIZE);
+		}
+	}
+}
+
+int amdgpu_gfx_rlc_init_microcode(struct amdgpu_device *adev,
+				  uint16_t version_major,
+				  uint16_t version_minor)
+{
+	int err;
+
+	if (version_major < 2) {
+		/* only support rlc_hdr v2.x and onwards */
+		dev_err(adev->dev, "unsupported rlc fw hdr\n");
+		return -EINVAL;
+	}
+
+	/* is_rlc_v2_1 is still used in APU code path */
+	if (version_major == 2 && version_minor == 1)
+		adev->gfx.rlc.is_rlc_v2_1 = true;
+
+	err = amdgpu_gfx_rlc_init_microcode_v2_0(adev);
+	if (err) {
+		dev_err(adev->dev, "fail to init rlc v2_0 microcode\n");
+		return err;
+	}
+
+	if (version_minor >= 1)
+		amdgpu_gfx_rlc_init_microcode_v2_1(adev);
+	if (version_minor >= 2)
+		amdgpu_gfx_rlc_init_microcode_v2_2(adev);
+	if (version_minor == 3)
+		amdgpu_gfx_rlc_init_microcode_v2_3(adev);
+	if (version_minor == 4)
+		amdgpu_gfx_rlc_init_microcode_v2_4(adev);
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.h
index 00afd0dcae86..2ce310b31942 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.h
@@ -26,6 +26,8 @@
 
 #include "clearstate_defs.h"
 
+#define AMDGPU_MAX_RLC_INSTANCES	8
+
 /* firmware ID used in rlc toc */
 typedef enum _FIRMWARE_ID_ {
 	FIRMWARE_ID_INVALID					= 0,
@@ -69,6 +71,94 @@ typedef enum _FIRMWARE_ID_ {
 	FIRMWARE_ID_MAX                                         = 38,
 } FIRMWARE_ID;
 
+typedef enum _SOC21_FIRMWARE_ID_ {
+    SOC21_FIRMWARE_ID_INVALID                     = 0,
+    SOC21_FIRMWARE_ID_RLC_G_UCODE                 = 1,
+    SOC21_FIRMWARE_ID_RLC_TOC                     = 2,
+    SOC21_FIRMWARE_ID_RLCG_SCRATCH                = 3,
+    SOC21_FIRMWARE_ID_RLC_SRM_ARAM                = 4,
+    SOC21_FIRMWARE_ID_RLC_P_UCODE                 = 5,
+    SOC21_FIRMWARE_ID_RLC_V_UCODE                 = 6,
+    SOC21_FIRMWARE_ID_RLX6_UCODE                  = 7,
+    SOC21_FIRMWARE_ID_RLX6_UCODE_CORE1            = 8,
+    SOC21_FIRMWARE_ID_RLX6_DRAM_BOOT              = 9,
+    SOC21_FIRMWARE_ID_RLX6_DRAM_BOOT_CORE1        = 10,
+    SOC21_FIRMWARE_ID_SDMA_UCODE_TH0              = 11,
+    SOC21_FIRMWARE_ID_SDMA_UCODE_TH1              = 12,
+    SOC21_FIRMWARE_ID_CP_PFP                      = 13,
+    SOC21_FIRMWARE_ID_CP_ME                       = 14,
+    SOC21_FIRMWARE_ID_CP_MEC                      = 15,
+    SOC21_FIRMWARE_ID_RS64_MES_P0                 = 16,
+    SOC21_FIRMWARE_ID_RS64_MES_P1                 = 17,
+    SOC21_FIRMWARE_ID_RS64_PFP                    = 18,
+    SOC21_FIRMWARE_ID_RS64_ME                     = 19,
+    SOC21_FIRMWARE_ID_RS64_MEC                    = 20,
+    SOC21_FIRMWARE_ID_RS64_MES_P0_STACK           = 21,
+    SOC21_FIRMWARE_ID_RS64_MES_P1_STACK           = 22,
+    SOC21_FIRMWARE_ID_RS64_PFP_P0_STACK           = 23,
+    SOC21_FIRMWARE_ID_RS64_PFP_P1_STACK           = 24,
+    SOC21_FIRMWARE_ID_RS64_ME_P0_STACK            = 25,
+    SOC21_FIRMWARE_ID_RS64_ME_P1_STACK            = 26,
+    SOC21_FIRMWARE_ID_RS64_MEC_P0_STACK           = 27,
+    SOC21_FIRMWARE_ID_RS64_MEC_P1_STACK           = 28,
+    SOC21_FIRMWARE_ID_RS64_MEC_P2_STACK           = 29,
+    SOC21_FIRMWARE_ID_RS64_MEC_P3_STACK           = 30,
+    SOC21_FIRMWARE_ID_RLC_SRM_DRAM_SR             = 31,
+    SOC21_FIRMWARE_ID_RLCG_SCRATCH_SR             = 32,
+    SOC21_FIRMWARE_ID_RLCP_SCRATCH_SR             = 33,
+    SOC21_FIRMWARE_ID_RLCV_SCRATCH_SR             = 34,
+    SOC21_FIRMWARE_ID_RLX6_DRAM_SR                = 35,
+    SOC21_FIRMWARE_ID_RLX6_DRAM_SR_CORE1          = 36,
+    SOC21_FIRMWARE_ID_MAX                         = 37
+} SOC21_FIRMWARE_ID;
+
+typedef enum _SOC24_FIRMWARE_ID_ {
+    SOC24_FIRMWARE_ID_INVALID                     = 0,
+    SOC24_FIRMWARE_ID_RLC_G_UCODE                 = 1,
+    SOC24_FIRMWARE_ID_RLC_TOC                     = 2,
+    SOC24_FIRMWARE_ID_RLCG_SCRATCH                = 3,
+    SOC24_FIRMWARE_ID_RLC_SRM_ARAM                = 4,
+    SOC24_FIRMWARE_ID_RLC_P_UCODE                 = 5,
+    SOC24_FIRMWARE_ID_RLC_V_UCODE                 = 6,
+    SOC24_FIRMWARE_ID_RLX6_UCODE                  = 7,
+    SOC24_FIRMWARE_ID_RLX6_UCODE_CORE1            = 8,
+    SOC24_FIRMWARE_ID_RLX6_DRAM_BOOT              = 9,
+    SOC24_FIRMWARE_ID_RLX6_DRAM_BOOT_CORE1        = 10,
+    SOC24_FIRMWARE_ID_SDMA_UCODE_TH0              = 11,
+    SOC24_FIRMWARE_ID_SDMA_UCODE_TH1              = 12,
+    SOC24_FIRMWARE_ID_CP_PFP                      = 13,
+    SOC24_FIRMWARE_ID_CP_ME                       = 14,
+    SOC24_FIRMWARE_ID_CP_MEC                      = 15,
+    SOC24_FIRMWARE_ID_RS64_MES_P0                 = 16,
+    SOC24_FIRMWARE_ID_RS64_MES_P1                 = 17,
+    SOC24_FIRMWARE_ID_RS64_PFP                    = 18,
+    SOC24_FIRMWARE_ID_RS64_ME                     = 19,
+    SOC24_FIRMWARE_ID_RS64_MEC                    = 20,
+    SOC24_FIRMWARE_ID_RS64_MES_P0_STACK           = 21,
+    SOC24_FIRMWARE_ID_RS64_MES_P1_STACK           = 22,
+    SOC24_FIRMWARE_ID_RS64_PFP_P0_STACK           = 23,
+    SOC24_FIRMWARE_ID_RS64_PFP_P1_STACK           = 24,
+    SOC24_FIRMWARE_ID_RS64_ME_P0_STACK            = 25,
+    SOC24_FIRMWARE_ID_RS64_ME_P1_STACK            = 26,
+    SOC24_FIRMWARE_ID_RS64_MEC_P0_STACK           = 27,
+    SOC24_FIRMWARE_ID_RS64_MEC_P1_STACK           = 28,
+    SOC24_FIRMWARE_ID_RS64_MEC_P2_STACK           = 29,
+    SOC24_FIRMWARE_ID_RS64_MEC_P3_STACK           = 30,
+    SOC24_FIRMWARE_ID_RLC_SRM_DRAM_SR             = 31,
+    SOC24_FIRMWARE_ID_RLCG_SCRATCH_SR             = 32,
+    SOC24_FIRMWARE_ID_RLCP_SCRATCH_SR             = 33,
+    SOC24_FIRMWARE_ID_RLCV_SCRATCH_SR             = 34,
+    SOC24_FIRMWARE_ID_RLX6_DRAM_SR                = 35,
+    SOC24_FIRMWARE_ID_RLX6_DRAM_SR_CORE1          = 36,
+    SOC24_FIRMWARE_ID_RLCDEBUGLOG                 = 37,
+    SOC24_FIRMWARE_ID_SRIOV_DEBUG                 = 38,
+    SOC24_FIRMWARE_ID_SRIOV_CSA_RLC               = 39,
+    SOC24_FIRMWARE_ID_SRIOV_CSA_SDMA              = 40,
+    SOC24_FIRMWARE_ID_SRIOV_CSA_CP                = 41,
+    SOC24_FIRMWARE_ID_UMF_ZONE_PAD                = 42,
+    SOC24_FIRMWARE_ID_MAX                         = 43
+} SOC24_FIRMWARE_ID;
+
 typedef struct _RLC_TABLE_OF_CONTENT {
 	union {
 		unsigned int	DW0;
@@ -112,47 +202,96 @@ typedef struct _RLC_TABLE_OF_CONTENT {
 	};
 } RLC_TABLE_OF_CONTENT;
 
+typedef struct _RLC_TABLE_OF_CONTENT_V2 {
+	union {
+		unsigned int    DW0;
+		struct {
+			uint32_t offset         : 25;
+			uint32_t id             : 7;
+		};
+	};
+
+	union {
+		unsigned int    DW1;
+		struct {
+			uint32_t reserved0              : 1;
+			uint32_t reserved1              : 1;
+			uint32_t reserved2              : 1;
+			uint32_t memory_destination     : 2;
+			uint32_t vfflr_image_code       : 4;
+			uint32_t reserved9              : 1;
+			uint32_t reserved10             : 1;
+			uint32_t reserved11             : 1;
+			uint32_t size_x16               : 1;
+			uint32_t reserved13             : 1;
+			uint32_t size                   : 18;
+		};
+	};
+} RLC_TABLE_OF_CONTENT_V2;
+
 #define RLC_TOC_MAX_SIZE		64
 
 struct amdgpu_rlc_funcs {
 	bool (*is_rlc_enabled)(struct amdgpu_device *adev);
-	void (*set_safe_mode)(struct amdgpu_device *adev);
-	void (*unset_safe_mode)(struct amdgpu_device *adev);
+	void (*set_safe_mode)(struct amdgpu_device *adev, int xcc_id);
+	void (*unset_safe_mode)(struct amdgpu_device *adev, int xcc_id);
 	int  (*init)(struct amdgpu_device *adev);
 	u32  (*get_csb_size)(struct amdgpu_device *adev);
-	void (*get_csb_buffer)(struct amdgpu_device *adev, volatile u32 *buffer);
+
+	/**
+	 * @get_csb_buffer: Get the clear state to be put into the hardware.
+	 *
+	 * The parameter adev is used to get the CS data and other gfx info,
+	 * and buffer is the RLC CS pointer
+	 *
+	 * Sometimes, the user space puts a request to clear the state in the
+	 * command buffer; this function provides the clear state that gets put
+	 * into the hardware. Note that the driver programs Clear State
+	 * Indirect Buffer (CSB) explicitly when it sets up the kernel rings,
+	 * and it also provides a pointer to it which is used by the firmware
+	 * to load the clear state in some cases.
+	 */
+	void (*get_csb_buffer)(struct amdgpu_device *adev, u32 *buffer);
 	int  (*get_cp_table_num)(struct amdgpu_device *adev);
 	int  (*resume)(struct amdgpu_device *adev);
 	void (*stop)(struct amdgpu_device *adev);
 	void (*reset)(struct amdgpu_device *adev);
 	void (*start)(struct amdgpu_device *adev);
-	void (*update_spm_vmid)(struct amdgpu_device *adev, unsigned vmid);
-	void (*sriov_wreg)(struct amdgpu_device *adev, u32 offset, u32 v, u32 acc_flags, u32 hwip);
-	u32 (*sriov_rreg)(struct amdgpu_device *adev, u32 offset, u32 acc_flags, u32 hwip);
+	void (*update_spm_vmid)(struct amdgpu_device *adev, struct amdgpu_ring *ring, unsigned vmid);
 	bool (*is_rlcg_access_range)(struct amdgpu_device *adev, uint32_t reg);
 };
 
+struct amdgpu_rlcg_reg_access_ctrl {
+	uint32_t scratch_reg0;
+	uint32_t scratch_reg1;
+	uint32_t scratch_reg2;
+	uint32_t scratch_reg3;
+	uint32_t grbm_cntl;
+	uint32_t grbm_idx;
+	uint32_t spare_int;
+};
+
 struct amdgpu_rlc {
 	/* for power gating */
 	struct amdgpu_bo        *save_restore_obj;
 	uint64_t                save_restore_gpu_addr;
-	volatile uint32_t       *sr_ptr;
+	uint32_t		*sr_ptr;
 	const u32               *reg_list;
 	u32                     reg_list_size;
 	/* for clear state */
 	struct amdgpu_bo        *clear_state_obj;
 	uint64_t                clear_state_gpu_addr;
-	volatile uint32_t       *cs_ptr;
+	uint32_t		*cs_ptr;
 	const struct cs_section_def   *cs_data;
 	u32                     clear_state_size;
 	/* for cp tables */
 	struct amdgpu_bo        *cp_table_obj;
 	uint64_t                cp_table_gpu_addr;
-	volatile uint32_t       *cp_table_ptr;
+	uint32_t		*cp_table_ptr;
 	u32                     cp_table_size;
 
 	/* safe mode for updating CG/PG state */
-	bool in_safe_mode;
+	bool in_safe_mode[AMDGPU_MAX_RLC_INSTANCES];
 	const struct amdgpu_rlc_funcs *funcs;
 
 	/* for firmware data */
@@ -171,6 +310,13 @@ struct amdgpu_rlc {
 	u32 save_restore_list_srm_size_bytes;
 	u32 rlc_iram_ucode_size_bytes;
 	u32 rlc_dram_ucode_size_bytes;
+	u32 rlcp_ucode_size_bytes;
+	u32 rlcv_ucode_size_bytes;
+	u32 global_tap_delays_ucode_size_bytes;
+	u32 se0_tap_delays_ucode_size_bytes;
+	u32 se1_tap_delays_ucode_size_bytes;
+	u32 se2_tap_delays_ucode_size_bytes;
+	u32 se3_tap_delays_ucode_size_bytes;
 
 	u32 *register_list_format;
 	u32 *register_restore;
@@ -179,6 +325,13 @@ struct amdgpu_rlc {
 	u8 *save_restore_list_srm;
 	u8 *rlc_iram_ucode;
 	u8 *rlc_dram_ucode;
+	u8 *rlcp_ucode;
+	u8 *rlcv_ucode;
+	u8 *global_tap_delays_ucode;
+	u8 *se0_tap_delays_ucode;
+	u8 *se1_tap_delays_ucode;
+	u8 *se2_tap_delays_ucode;
+	u8 *se3_tap_delays_ucode;
 
 	bool is_rlc_v2_1;
 
@@ -191,14 +344,20 @@ struct amdgpu_rlc {
 	struct amdgpu_bo	*rlc_toc_bo;
 	uint64_t		rlc_toc_gpu_addr;
 	void			*rlc_toc_buf;
+
+	bool rlcg_reg_access_supported;
+	/* registers for rlcg indirect reg access */
+	struct amdgpu_rlcg_reg_access_ctrl reg_access_ctrl[AMDGPU_MAX_RLC_INSTANCES];
 };
 
-void amdgpu_gfx_rlc_enter_safe_mode(struct amdgpu_device *adev);
-void amdgpu_gfx_rlc_exit_safe_mode(struct amdgpu_device *adev);
+void amdgpu_gfx_rlc_enter_safe_mode(struct amdgpu_device *adev, int xcc_id);
+void amdgpu_gfx_rlc_exit_safe_mode(struct amdgpu_device *adev, int xcc_id);
 int amdgpu_gfx_rlc_init_sr(struct amdgpu_device *adev, u32 dws);
 int amdgpu_gfx_rlc_init_csb(struct amdgpu_device *adev);
 int amdgpu_gfx_rlc_init_cpt(struct amdgpu_device *adev);
 void amdgpu_gfx_rlc_setup_cp_table(struct amdgpu_device *adev);
 void amdgpu_gfx_rlc_fini(struct amdgpu_device *adev);
-
+int amdgpu_gfx_rlc_init_microcode(struct amdgpu_device *adev,
+				  uint16_t version_major,
+				  uint16_t version_minor);
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sa.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_sa.c
index 524d10b21041..39070b2a4c04 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sa.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sa.c
@@ -44,327 +44,62 @@
 
 #include "amdgpu.h"
 
-static void amdgpu_sa_bo_remove_locked(struct amdgpu_sa_bo *sa_bo);
-static void amdgpu_sa_bo_try_free(struct amdgpu_sa_manager *sa_manager);
-
 int amdgpu_sa_bo_manager_init(struct amdgpu_device *adev,
 			      struct amdgpu_sa_manager *sa_manager,
-			      unsigned size, u32 align, u32 domain)
+			      unsigned int size, u32 suballoc_align, u32 domain)
 {
-	int i, r;
-
-	init_waitqueue_head(&sa_manager->wq);
-	sa_manager->bo = NULL;
-	sa_manager->size = size;
-	sa_manager->domain = domain;
-	sa_manager->align = align;
-	sa_manager->hole = &sa_manager->olist;
-	INIT_LIST_HEAD(&sa_manager->olist);
-	for (i = 0; i < AMDGPU_SA_NUM_FENCE_LISTS; ++i)
-		INIT_LIST_HEAD(&sa_manager->flist[i]);
+	int r;
 
-	r = amdgpu_bo_create_kernel(adev, size, align, domain, &sa_manager->bo,
-				&sa_manager->gpu_addr, &sa_manager->cpu_ptr);
+	r = amdgpu_bo_create_kernel(adev, size, AMDGPU_GPU_PAGE_SIZE, domain,
+				    &sa_manager->bo, &sa_manager->gpu_addr,
+				    &sa_manager->cpu_ptr);
 	if (r) {
 		dev_err(adev->dev, "(%d) failed to allocate bo for manager\n", r);
 		return r;
 	}
 
-	memset(sa_manager->cpu_ptr, 0, sa_manager->size);
+	memset(sa_manager->cpu_ptr, 0, size);
+	drm_suballoc_manager_init(&sa_manager->base, size, suballoc_align);
 	return r;
 }
 
 void amdgpu_sa_bo_manager_fini(struct amdgpu_device *adev,
 			       struct amdgpu_sa_manager *sa_manager)
 {
-	struct amdgpu_sa_bo *sa_bo, *tmp;
-
 	if (sa_manager->bo == NULL) {
 		dev_err(adev->dev, "no bo for sa manager\n");
 		return;
 	}
 
-	if (!list_empty(&sa_manager->olist)) {
-		sa_manager->hole = &sa_manager->olist,
-		amdgpu_sa_bo_try_free(sa_manager);
-		if (!list_empty(&sa_manager->olist)) {
-			dev_err(adev->dev, "sa_manager is not empty, clearing anyway\n");
-		}
-	}
-	list_for_each_entry_safe(sa_bo, tmp, &sa_manager->olist, olist) {
-		amdgpu_sa_bo_remove_locked(sa_bo);
-	}
+	drm_suballoc_manager_fini(&sa_manager->base);
 
 	amdgpu_bo_free_kernel(&sa_manager->bo, &sa_manager->gpu_addr, &sa_manager->cpu_ptr);
-	sa_manager->size = 0;
 }
 
-static void amdgpu_sa_bo_remove_locked(struct amdgpu_sa_bo *sa_bo)
-{
-	struct amdgpu_sa_manager *sa_manager = sa_bo->manager;
-	if (sa_manager->hole == &sa_bo->olist) {
-		sa_manager->hole = sa_bo->olist.prev;
-	}
-	list_del_init(&sa_bo->olist);
-	list_del_init(&sa_bo->flist);
-	dma_fence_put(sa_bo->fence);
-	kfree(sa_bo);
-}
-
-static void amdgpu_sa_bo_try_free(struct amdgpu_sa_manager *sa_manager)
+int amdgpu_sa_bo_new(struct amdgpu_sa_manager *sa_manager,
+		     struct drm_suballoc **sa_bo,
+		     unsigned int size)
 {
-	struct amdgpu_sa_bo *sa_bo, *tmp;
+	struct drm_suballoc *sa = drm_suballoc_new(&sa_manager->base, size,
+						   GFP_KERNEL, false, 0);
 
-	if (sa_manager->hole->next == &sa_manager->olist)
-		return;
+	if (IS_ERR(sa)) {
+		*sa_bo = NULL;
 
-	sa_bo = list_entry(sa_manager->hole->next, struct amdgpu_sa_bo, olist);
-	list_for_each_entry_safe_from(sa_bo, tmp, &sa_manager->olist, olist) {
-		if (sa_bo->fence == NULL ||
-		    !dma_fence_is_signaled(sa_bo->fence)) {
-			return;
-		}
-		amdgpu_sa_bo_remove_locked(sa_bo);
+		return PTR_ERR(sa);
 	}
-}
 
-static inline unsigned amdgpu_sa_bo_hole_soffset(struct amdgpu_sa_manager *sa_manager)
-{
-	struct list_head *hole = sa_manager->hole;
-
-	if (hole != &sa_manager->olist) {
-		return list_entry(hole, struct amdgpu_sa_bo, olist)->eoffset;
-	}
+	*sa_bo = sa;
 	return 0;
 }
 
-static inline unsigned amdgpu_sa_bo_hole_eoffset(struct amdgpu_sa_manager *sa_manager)
-{
-	struct list_head *hole = sa_manager->hole;
-
-	if (hole->next != &sa_manager->olist) {
-		return list_entry(hole->next, struct amdgpu_sa_bo, olist)->soffset;
-	}
-	return sa_manager->size;
-}
-
-static bool amdgpu_sa_bo_try_alloc(struct amdgpu_sa_manager *sa_manager,
-				   struct amdgpu_sa_bo *sa_bo,
-				   unsigned size, unsigned align)
+void amdgpu_sa_bo_free(struct drm_suballoc **sa_bo, struct dma_fence *fence)
 {
-	unsigned soffset, eoffset, wasted;
-
-	soffset = amdgpu_sa_bo_hole_soffset(sa_manager);
-	eoffset = amdgpu_sa_bo_hole_eoffset(sa_manager);
-	wasted = (align - (soffset % align)) % align;
-
-	if ((eoffset - soffset) >= (size + wasted)) {
-		soffset += wasted;
-
-		sa_bo->manager = sa_manager;
-		sa_bo->soffset = soffset;
-		sa_bo->eoffset = soffset + size;
-		list_add(&sa_bo->olist, sa_manager->hole);
-		INIT_LIST_HEAD(&sa_bo->flist);
-		sa_manager->hole = &sa_bo->olist;
-		return true;
-	}
-	return false;
-}
-
-/**
- * amdgpu_sa_event - Check if we can stop waiting
- *
- * @sa_manager: pointer to the sa_manager
- * @size: number of bytes we want to allocate
- * @align: alignment we need to match
- *
- * Check if either there is a fence we can wait for or
- * enough free memory to satisfy the allocation directly
- */
-static bool amdgpu_sa_event(struct amdgpu_sa_manager *sa_manager,
-			    unsigned size, unsigned align)
-{
-	unsigned soffset, eoffset, wasted;
-	int i;
-
-	for (i = 0; i < AMDGPU_SA_NUM_FENCE_LISTS; ++i)
-		if (!list_empty(&sa_manager->flist[i]))
-			return true;
-
-	soffset = amdgpu_sa_bo_hole_soffset(sa_manager);
-	eoffset = amdgpu_sa_bo_hole_eoffset(sa_manager);
-	wasted = (align - (soffset % align)) % align;
-
-	if ((eoffset - soffset) >= (size + wasted)) {
-		return true;
-	}
-
-	return false;
-}
-
-static bool amdgpu_sa_bo_next_hole(struct amdgpu_sa_manager *sa_manager,
-				   struct dma_fence **fences,
-				   unsigned *tries)
-{
-	struct amdgpu_sa_bo *best_bo = NULL;
-	unsigned i, soffset, best, tmp;
-
-	/* if hole points to the end of the buffer */
-	if (sa_manager->hole->next == &sa_manager->olist) {
-		/* try again with its beginning */
-		sa_manager->hole = &sa_manager->olist;
-		return true;
-	}
-
-	soffset = amdgpu_sa_bo_hole_soffset(sa_manager);
-	/* to handle wrap around we add sa_manager->size */
-	best = sa_manager->size * 2;
-	/* go over all fence list and try to find the closest sa_bo
-	 * of the current last
-	 */
-	for (i = 0; i < AMDGPU_SA_NUM_FENCE_LISTS; ++i) {
-		struct amdgpu_sa_bo *sa_bo;
-
-		fences[i] = NULL;
-
-		if (list_empty(&sa_manager->flist[i]))
-			continue;
-
-		sa_bo = list_first_entry(&sa_manager->flist[i],
-					 struct amdgpu_sa_bo, flist);
-
-		if (!dma_fence_is_signaled(sa_bo->fence)) {
-			fences[i] = sa_bo->fence;
-			continue;
-		}
-
-		/* limit the number of tries each ring gets */
-		if (tries[i] > 2) {
-			continue;
-		}
-
-		tmp = sa_bo->soffset;
-		if (tmp < soffset) {
-			/* wrap around, pretend it's after */
-			tmp += sa_manager->size;
-		}
-		tmp -= soffset;
-		if (tmp < best) {
-			/* this sa bo is the closest one */
-			best = tmp;
-			best_bo = sa_bo;
-		}
-	}
-
-	if (best_bo) {
-		uint32_t idx = best_bo->fence->context;
-
-		idx %= AMDGPU_SA_NUM_FENCE_LISTS;
-		++tries[idx];
-		sa_manager->hole = best_bo->olist.prev;
-
-		/* we knew that this one is signaled,
-		   so it's save to remote it */
-		amdgpu_sa_bo_remove_locked(best_bo);
-		return true;
-	}
-	return false;
-}
-
-int amdgpu_sa_bo_new(struct amdgpu_sa_manager *sa_manager,
-		     struct amdgpu_sa_bo **sa_bo,
-		     unsigned size, unsigned align)
-{
-	struct dma_fence *fences[AMDGPU_SA_NUM_FENCE_LISTS];
-	unsigned tries[AMDGPU_SA_NUM_FENCE_LISTS];
-	unsigned count;
-	int i, r;
-	signed long t;
-
-	if (WARN_ON_ONCE(align > sa_manager->align))
-		return -EINVAL;
-
-	if (WARN_ON_ONCE(size > sa_manager->size))
-		return -EINVAL;
-
-	*sa_bo = kmalloc(sizeof(struct amdgpu_sa_bo), GFP_KERNEL);
-	if (!(*sa_bo))
-		return -ENOMEM;
-	(*sa_bo)->manager = sa_manager;
-	(*sa_bo)->fence = NULL;
-	INIT_LIST_HEAD(&(*sa_bo)->olist);
-	INIT_LIST_HEAD(&(*sa_bo)->flist);
-
-	spin_lock(&sa_manager->wq.lock);
-	do {
-		for (i = 0; i < AMDGPU_SA_NUM_FENCE_LISTS; ++i)
-			tries[i] = 0;
-
-		do {
-			amdgpu_sa_bo_try_free(sa_manager);
-
-			if (amdgpu_sa_bo_try_alloc(sa_manager, *sa_bo,
-						   size, align)) {
-				spin_unlock(&sa_manager->wq.lock);
-				return 0;
-			}
-
-			/* see if we can skip over some allocations */
-		} while (amdgpu_sa_bo_next_hole(sa_manager, fences, tries));
-
-		for (i = 0, count = 0; i < AMDGPU_SA_NUM_FENCE_LISTS; ++i)
-			if (fences[i])
-				fences[count++] = dma_fence_get(fences[i]);
-
-		if (count) {
-			spin_unlock(&sa_manager->wq.lock);
-			t = dma_fence_wait_any_timeout(fences, count, false,
-						       MAX_SCHEDULE_TIMEOUT,
-						       NULL);
-			for (i = 0; i < count; ++i)
-				dma_fence_put(fences[i]);
-
-			r = (t > 0) ? 0 : t;
-			spin_lock(&sa_manager->wq.lock);
-		} else {
-			/* if we have nothing to wait for block */
-			r = wait_event_interruptible_locked(
-				sa_manager->wq,
-				amdgpu_sa_event(sa_manager, size, align)
-			);
-		}
-
-	} while (!r);
-
-	spin_unlock(&sa_manager->wq.lock);
-	kfree(*sa_bo);
-	*sa_bo = NULL;
-	return r;
-}
-
-void amdgpu_sa_bo_free(struct amdgpu_device *adev, struct amdgpu_sa_bo **sa_bo,
-		       struct dma_fence *fence)
-{
-	struct amdgpu_sa_manager *sa_manager;
-
 	if (sa_bo == NULL || *sa_bo == NULL) {
 		return;
 	}
 
-	sa_manager = (*sa_bo)->manager;
-	spin_lock(&sa_manager->wq.lock);
-	if (fence && !dma_fence_is_signaled(fence)) {
-		uint32_t idx;
-
-		(*sa_bo)->fence = dma_fence_get(fence);
-		idx = fence->context % AMDGPU_SA_NUM_FENCE_LISTS;
-		list_add_tail(&(*sa_bo)->flist, &sa_manager->flist[idx]);
-	} else {
-		amdgpu_sa_bo_remove_locked(*sa_bo);
-	}
-	wake_up_all_locked(&sa_manager->wq);
-	spin_unlock(&sa_manager->wq.lock);
+	drm_suballoc_free(*sa_bo, fence);
 	*sa_bo = NULL;
 }
 
@@ -373,26 +108,8 @@ void amdgpu_sa_bo_free(struct amdgpu_device *adev, struct amdgpu_sa_bo **sa_bo,
 void amdgpu_sa_bo_dump_debug_info(struct amdgpu_sa_manager *sa_manager,
 				  struct seq_file *m)
 {
-	struct amdgpu_sa_bo *i;
-
-	spin_lock(&sa_manager->wq.lock);
-	list_for_each_entry(i, &sa_manager->olist, olist) {
-		uint64_t soffset = i->soffset + sa_manager->gpu_addr;
-		uint64_t eoffset = i->eoffset + sa_manager->gpu_addr;
-		if (&i->olist == sa_manager->hole) {
-			seq_printf(m, ">");
-		} else {
-			seq_printf(m, " ");
-		}
-		seq_printf(m, "[0x%010llx 0x%010llx] size %8lld",
-			   soffset, eoffset, eoffset - soffset);
+	struct drm_printer p = drm_seq_file_printer(m);
 
-		if (i->fence)
-			seq_printf(m, " protected by 0x%016llx on context %llu",
-				   i->fence->seqno, i->fence->context);
-
-		seq_printf(m, "\n");
-	}
-	spin_unlock(&sa_manager->wq.lock);
+	drm_suballoc_dump_debug_info(&sa_manager->base, &p, sa_manager->gpu_addr);
 }
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sched.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_sched.c
index e9b45089a28a..341beec59537 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sched.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sched.c
@@ -22,7 +22,6 @@
  * Authors: Andres Rodriguez <andresx7@gmail.com>
  */
 
-#include <linux/fdtable.h>
 #include <linux/file.h>
 #include <linux/pid.h>
 
@@ -36,25 +35,26 @@ static int amdgpu_sched_process_priority_override(struct amdgpu_device *adev,
 						  int fd,
 						  int32_t priority)
 {
-	struct fd f = fdget(fd);
+	CLASS(fd, f)(fd);
 	struct amdgpu_fpriv *fpriv;
+	struct amdgpu_ctx_mgr *mgr;
 	struct amdgpu_ctx *ctx;
 	uint32_t id;
 	int r;
 
-	if (!f.file)
+	if (fd_empty(f))
 		return -EINVAL;
 
-	r = amdgpu_file_to_fpriv(f.file, &fpriv);
-	if (r) {
-		fdput(f);
+	r = amdgpu_file_to_fpriv(fd_file(f), &fpriv);
+	if (r)
 		return r;
-	}
 
-	idr_for_each_entry(&fpriv->ctx_mgr.ctx_handles, ctx, id)
+	mgr = &fpriv->ctx_mgr;
+	mutex_lock(&mgr->lock);
+	idr_for_each_entry(&mgr->ctx_handles, ctx, id)
 		amdgpu_ctx_priority_override(ctx, priority);
+	mutex_unlock(&mgr->lock);
 
-	fdput(f);
 	return 0;
 }
 
@@ -63,31 +63,25 @@ static int amdgpu_sched_context_priority_override(struct amdgpu_device *adev,
 						  unsigned ctx_id,
 						  int32_t priority)
 {
-	struct fd f = fdget(fd);
+	CLASS(fd, f)(fd);
 	struct amdgpu_fpriv *fpriv;
 	struct amdgpu_ctx *ctx;
 	int r;
 
-	if (!f.file)
+	if (fd_empty(f))
 		return -EINVAL;
 
-	r = amdgpu_file_to_fpriv(f.file, &fpriv);
-	if (r) {
-		fdput(f);
+	r = amdgpu_file_to_fpriv(fd_file(f), &fpriv);
+	if (r)
 		return r;
-	}
 
 	ctx = amdgpu_ctx_get(fpriv, ctx_id);
 
-	if (!ctx) {
-		fdput(f);
+	if (!ctx)
 		return -EINVAL;
-	}
 
 	amdgpu_ctx_priority_override(ctx, priority);
 	amdgpu_ctx_put(ctx);
-	fdput(f);
-
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.c
index 65debb65a5df..8b8a04138711 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.c
@@ -21,9 +21,13 @@
  *
  */
 
+#include <linux/firmware.h>
 #include "amdgpu.h"
 #include "amdgpu_sdma.h"
 #include "amdgpu_ras.h"
+#include "amdgpu_reset.h"
+#include "gc/gc_10_1_0_offset.h"
+#include "gc/gc_10_3_0_sh_mask.h"
 
 #define AMDGPU_CSA_SDMA_SIZE 64
 /* SDMA CSA reside in the 3rd page of CSA */
@@ -63,7 +67,7 @@ int amdgpu_sdma_get_index_from_ring(struct amdgpu_ring *ring, uint32_t *index)
 }
 
 uint64_t amdgpu_sdma_get_csa_mc_addr(struct amdgpu_ring *ring,
-				     unsigned vmid)
+				     unsigned int vmid)
 {
 	struct amdgpu_device *adev = ring->adev;
 	uint64_t csa_mc_addr;
@@ -71,7 +75,7 @@ uint64_t amdgpu_sdma_get_csa_mc_addr(struct amdgpu_ring *ring,
 	int r;
 
 	/* don't enable OS preemption on SDMA under SRIOV */
-	if (amdgpu_sriov_vf(adev) || vmid == 0 || !amdgpu_mcbp)
+	if (amdgpu_sriov_vf(adev) || vmid == 0 || !adev->gfx.mcbp)
 		return 0;
 
 	r = amdgpu_sdma_get_index_from_ring(ring, &index);
@@ -87,78 +91,39 @@ uint64_t amdgpu_sdma_get_csa_mc_addr(struct amdgpu_ring *ring,
 }
 
 int amdgpu_sdma_ras_late_init(struct amdgpu_device *adev,
-			      void *ras_ih_info)
+			      struct ras_common_if *ras_block)
 {
 	int r, i;
-	struct ras_ih_if *ih_info = (struct ras_ih_if *)ras_ih_info;
-	struct ras_fs_if fs_info = {
-		.sysfs_name = "sdma_err_count",
-	};
-
-	if (!ih_info)
-		return -EINVAL;
-
-	if (!adev->sdma.ras_if) {
-		adev->sdma.ras_if = kmalloc(sizeof(struct ras_common_if), GFP_KERNEL);
-		if (!adev->sdma.ras_if)
-			return -ENOMEM;
-		adev->sdma.ras_if->block = AMDGPU_RAS_BLOCK__SDMA;
-		adev->sdma.ras_if->type = AMDGPU_RAS_ERROR__MULTI_UNCORRECTABLE;
-		adev->sdma.ras_if->sub_block_index = 0;
-	}
-	fs_info.head = ih_info->head = *adev->sdma.ras_if;
 
-	r = amdgpu_ras_late_init(adev, adev->sdma.ras_if,
-				 &fs_info, ih_info);
+	r = amdgpu_ras_block_late_init(adev, ras_block);
 	if (r)
-		goto free;
+		return r;
 
-	if (amdgpu_ras_is_supported(adev, adev->sdma.ras_if->block)) {
+	if (amdgpu_ras_is_supported(adev, ras_block->block)) {
 		for (i = 0; i < adev->sdma.num_instances; i++) {
 			r = amdgpu_irq_get(adev, &adev->sdma.ecc_irq,
 				AMDGPU_SDMA_IRQ_INSTANCE0 + i);
 			if (r)
 				goto late_fini;
 		}
-	} else {
-		r = 0;
-		goto free;
 	}
 
 	return 0;
 
 late_fini:
-	amdgpu_ras_late_fini(adev, adev->sdma.ras_if, ih_info);
-free:
-	kfree(adev->sdma.ras_if);
-	adev->sdma.ras_if = NULL;
+	amdgpu_ras_block_late_fini(adev, ras_block);
 	return r;
 }
 
-void amdgpu_sdma_ras_fini(struct amdgpu_device *adev)
-{
-	if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__SDMA) &&
-			adev->sdma.ras_if) {
-		struct ras_common_if *ras_if = adev->sdma.ras_if;
-		struct ras_ih_if ih_info = {
-			.head = *ras_if,
-			/* the cb member will not be used by
-			 * amdgpu_ras_interrupt_remove_handler, init it only
-			 * to cheat the check in ras_late_fini
-			 */
-			.cb = amdgpu_sdma_process_ras_data_cb,
-		};
-
-		amdgpu_ras_late_fini(adev, ras_if, &ih_info);
-		kfree(ras_if);
-	}
-}
-
 int amdgpu_sdma_process_ras_data_cb(struct amdgpu_device *adev,
 		void *err_data,
 		struct amdgpu_iv_entry *entry)
 {
 	kgd2kfd_set_sram_ecc_flag(adev->kfd.dev);
+
+	if (amdgpu_sriov_vf(adev))
+		return AMDGPU_RAS_SUCCESS;
+
 	amdgpu_ras_reset_gpu(adev);
 
 	return AMDGPU_RAS_SUCCESS;
@@ -181,3 +146,466 @@ int amdgpu_sdma_process_ecc_irq(struct amdgpu_device *adev,
 	amdgpu_ras_interrupt_dispatch(adev, &ih_data);
 	return 0;
 }
+
+static int amdgpu_sdma_init_inst_ctx(struct amdgpu_sdma_instance *sdma_inst)
+{
+	uint16_t version_major;
+	const struct common_firmware_header *header = NULL;
+	const struct sdma_firmware_header_v1_0 *hdr;
+	const struct sdma_firmware_header_v2_0 *hdr_v2;
+	const struct sdma_firmware_header_v3_0 *hdr_v3;
+
+	header = (const struct common_firmware_header *)
+		sdma_inst->fw->data;
+	version_major = le16_to_cpu(header->header_version_major);
+
+	switch (version_major) {
+	case 1:
+		hdr = (const struct sdma_firmware_header_v1_0 *)sdma_inst->fw->data;
+		sdma_inst->fw_version = le32_to_cpu(hdr->header.ucode_version);
+		sdma_inst->feature_version = le32_to_cpu(hdr->ucode_feature_version);
+		break;
+	case 2:
+		hdr_v2 = (const struct sdma_firmware_header_v2_0 *)sdma_inst->fw->data;
+		sdma_inst->fw_version = le32_to_cpu(hdr_v2->header.ucode_version);
+		sdma_inst->feature_version = le32_to_cpu(hdr_v2->ucode_feature_version);
+		break;
+	case 3:
+		hdr_v3 = (const struct sdma_firmware_header_v3_0 *)sdma_inst->fw->data;
+		sdma_inst->fw_version = le32_to_cpu(hdr_v3->header.ucode_version);
+		sdma_inst->feature_version = le32_to_cpu(hdr_v3->ucode_feature_version);
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	if (sdma_inst->feature_version >= 20)
+		sdma_inst->burst_nop = true;
+
+	return 0;
+}
+
+void amdgpu_sdma_destroy_inst_ctx(struct amdgpu_device *adev,
+				  bool duplicate)
+{
+	int i;
+
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		amdgpu_ucode_release(&adev->sdma.instance[i].fw);
+		if (duplicate)
+			break;
+	}
+
+	memset((void *)adev->sdma.instance, 0,
+	       sizeof(struct amdgpu_sdma_instance) * AMDGPU_MAX_SDMA_INSTANCES);
+}
+
+int amdgpu_sdma_init_microcode(struct amdgpu_device *adev,
+			       u32 instance, bool duplicate)
+{
+	struct amdgpu_firmware_info *info = NULL;
+	const struct common_firmware_header *header = NULL;
+	int err, i;
+	const struct sdma_firmware_header_v2_0 *sdma_hdr;
+	const struct sdma_firmware_header_v3_0 *sdma_hv3;
+	uint16_t version_major;
+	char ucode_prefix[30];
+
+	amdgpu_ucode_ip_version_decode(adev, SDMA0_HWIP, ucode_prefix, sizeof(ucode_prefix));
+	if (instance == 0)
+		err = amdgpu_ucode_request(adev, &adev->sdma.instance[instance].fw,
+					   AMDGPU_UCODE_REQUIRED,
+					   "amdgpu/%s.bin", ucode_prefix);
+	else
+		err = amdgpu_ucode_request(adev, &adev->sdma.instance[instance].fw,
+					   AMDGPU_UCODE_REQUIRED,
+					   "amdgpu/%s%d.bin", ucode_prefix, instance);
+	if (err)
+		goto out;
+
+	header = (const struct common_firmware_header *)
+		adev->sdma.instance[instance].fw->data;
+	version_major = le16_to_cpu(header->header_version_major);
+
+	if ((duplicate && instance) || (!duplicate && version_major > 1)) {
+		err = -EINVAL;
+		goto out;
+	}
+
+	err = amdgpu_sdma_init_inst_ctx(&adev->sdma.instance[instance]);
+	if (err)
+		goto out;
+
+	if (duplicate) {
+		for (i = 1; i < adev->sdma.num_instances; i++)
+			memcpy((void *)&adev->sdma.instance[i],
+			       (void *)&adev->sdma.instance[0],
+			       sizeof(struct amdgpu_sdma_instance));
+	}
+
+	DRM_DEBUG("psp_load == '%s'\n",
+		  adev->firmware.load_type == AMDGPU_FW_LOAD_PSP ? "true" : "false");
+
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
+		switch (version_major) {
+		case 1:
+			for (i = 0; i < adev->sdma.num_instances; i++) {
+				if (!duplicate && (instance != i))
+					continue;
+				else {
+					/* Use a single copy per SDMA firmware type. PSP uses the same instance for all
+					 * groups of SDMAs */
+					if ((amdgpu_ip_version(adev, SDMA0_HWIP, 0) ==
+						IP_VERSION(4, 4, 2) ||
+					     amdgpu_ip_version(adev, SDMA0_HWIP, 0) ==
+						IP_VERSION(4, 4, 4) ||
+					     amdgpu_ip_version(adev, SDMA0_HWIP, 0) ==
+						IP_VERSION(4, 4, 5)) &&
+					    adev->firmware.load_type ==
+						AMDGPU_FW_LOAD_PSP &&
+					    adev->sdma.num_inst_per_aid == i) {
+						break;
+					}
+					info = &adev->firmware.ucode[AMDGPU_UCODE_ID_SDMA0 + i];
+					info->ucode_id = AMDGPU_UCODE_ID_SDMA0 + i;
+					info->fw = adev->sdma.instance[i].fw;
+					adev->firmware.fw_size +=
+						ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE);
+				}
+			}
+			break;
+		case 2:
+			sdma_hdr = (const struct sdma_firmware_header_v2_0 *)
+				adev->sdma.instance[0].fw->data;
+			info = &adev->firmware.ucode[AMDGPU_UCODE_ID_SDMA_UCODE_TH0];
+			info->ucode_id = AMDGPU_UCODE_ID_SDMA_UCODE_TH0;
+			info->fw = adev->sdma.instance[0].fw;
+			adev->firmware.fw_size +=
+				ALIGN(le32_to_cpu(sdma_hdr->ctx_ucode_size_bytes), PAGE_SIZE);
+			info = &adev->firmware.ucode[AMDGPU_UCODE_ID_SDMA_UCODE_TH1];
+			info->ucode_id = AMDGPU_UCODE_ID_SDMA_UCODE_TH1;
+			info->fw = adev->sdma.instance[0].fw;
+			adev->firmware.fw_size +=
+				ALIGN(le32_to_cpu(sdma_hdr->ctl_ucode_size_bytes), PAGE_SIZE);
+			break;
+		case 3:
+			sdma_hv3 = (const struct sdma_firmware_header_v3_0 *)
+				adev->sdma.instance[0].fw->data;
+			info = &adev->firmware.ucode[AMDGPU_UCODE_ID_SDMA_RS64];
+			info->ucode_id = AMDGPU_UCODE_ID_SDMA_RS64;
+			info->fw = adev->sdma.instance[0].fw;
+			adev->firmware.fw_size +=
+				ALIGN(le32_to_cpu(sdma_hv3->ucode_size_bytes), PAGE_SIZE);
+			break;
+		default:
+			err = -EINVAL;
+		}
+	}
+
+out:
+	if (err)
+		amdgpu_sdma_destroy_inst_ctx(adev, duplicate);
+	return err;
+}
+
+int amdgpu_sdma_ras_sw_init(struct amdgpu_device *adev)
+{
+	int err = 0;
+	struct amdgpu_sdma_ras *ras = NULL;
+
+	/* adev->sdma.ras is NULL, which means sdma does not
+	 * support ras function, then do nothing here.
+	 */
+	if (!adev->sdma.ras)
+		return 0;
+
+	ras = adev->sdma.ras;
+
+	err = amdgpu_ras_register_ras_block(adev, &ras->ras_block);
+	if (err) {
+		dev_err(adev->dev, "Failed to register sdma ras block!\n");
+		return err;
+	}
+
+	strcpy(ras->ras_block.ras_comm.name, "sdma");
+	ras->ras_block.ras_comm.block = AMDGPU_RAS_BLOCK__SDMA;
+	ras->ras_block.ras_comm.type = AMDGPU_RAS_ERROR__MULTI_UNCORRECTABLE;
+	adev->sdma.ras_if = &ras->ras_block.ras_comm;
+
+	/* If not define special ras_late_init function, use default ras_late_init */
+	if (!ras->ras_block.ras_late_init)
+		ras->ras_block.ras_late_init = amdgpu_sdma_ras_late_init;
+
+	/* If not defined special ras_cb function, use default ras_cb */
+	if (!ras->ras_block.ras_cb)
+		ras->ras_block.ras_cb = amdgpu_sdma_process_ras_data_cb;
+
+	return 0;
+}
+
+/*
+ * debugfs for to enable/disable sdma job submission to specific core.
+ */
+#if defined(CONFIG_DEBUG_FS)
+static int amdgpu_debugfs_sdma_sched_mask_set(void *data, u64 val)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)data;
+	u64 i, num_ring;
+	u64 mask = 0;
+	struct amdgpu_ring *ring, *page = NULL;
+
+	if (!adev)
+		return -ENODEV;
+
+	/* Determine the number of rings per SDMA instance
+	 * (1 for sdma gfx ring, 2 if page queue exists)
+	 */
+	if (adev->sdma.has_page_queue)
+		num_ring = 2;
+	else
+		num_ring = 1;
+
+	/* Calculate the maximum possible mask value
+	 * based on the number of SDMA instances and rings
+	*/
+	mask = BIT_ULL(adev->sdma.num_instances * num_ring) - 1;
+
+	if ((val & mask) == 0)
+		return -EINVAL;
+
+	for (i = 0; i < adev->sdma.num_instances; ++i) {
+		ring = &adev->sdma.instance[i].ring;
+		if (adev->sdma.has_page_queue)
+			page = &adev->sdma.instance[i].page;
+		if (val & BIT_ULL(i * num_ring))
+			ring->sched.ready = true;
+		else
+			ring->sched.ready = false;
+
+		if (page) {
+			if (val & BIT_ULL(i * num_ring + 1))
+				page->sched.ready = true;
+			else
+				page->sched.ready = false;
+		}
+	}
+	/* publish sched.ready flag update effective immediately across smp */
+	smp_rmb();
+	return 0;
+}
+
+static int amdgpu_debugfs_sdma_sched_mask_get(void *data, u64 *val)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)data;
+	u64 i, num_ring;
+	u64 mask = 0;
+	struct amdgpu_ring *ring, *page = NULL;
+
+	if (!adev)
+		return -ENODEV;
+
+	/* Determine the number of rings per SDMA instance
+	 * (1 for sdma gfx ring, 2 if page queue exists)
+	 */
+	if (adev->sdma.has_page_queue)
+		num_ring = 2;
+	else
+		num_ring = 1;
+
+	for (i = 0; i < adev->sdma.num_instances; ++i) {
+		ring = &adev->sdma.instance[i].ring;
+		if (adev->sdma.has_page_queue)
+			page = &adev->sdma.instance[i].page;
+
+		if (ring->sched.ready)
+			mask |= BIT_ULL(i * num_ring);
+		else
+			mask &= ~BIT_ULL(i * num_ring);
+
+		if (page) {
+			if (page->sched.ready)
+				mask |= BIT_ULL(i * num_ring + 1);
+			else
+				mask &= ~BIT_ULL(i * num_ring + 1);
+		}
+	}
+
+	*val = mask;
+	return 0;
+}
+
+DEFINE_DEBUGFS_ATTRIBUTE(amdgpu_debugfs_sdma_sched_mask_fops,
+			 amdgpu_debugfs_sdma_sched_mask_get,
+			 amdgpu_debugfs_sdma_sched_mask_set, "%llx\n");
+
+#endif
+
+void amdgpu_debugfs_sdma_sched_mask_init(struct amdgpu_device *adev)
+{
+#if defined(CONFIG_DEBUG_FS)
+	struct drm_minor *minor = adev_to_drm(adev)->primary;
+	struct dentry *root = minor->debugfs_root;
+	char name[32];
+
+	if (!(adev->sdma.num_instances > 1))
+		return;
+	sprintf(name, "amdgpu_sdma_sched_mask");
+	debugfs_create_file(name, 0600, root, adev,
+			    &amdgpu_debugfs_sdma_sched_mask_fops);
+#endif
+}
+
+static ssize_t amdgpu_get_sdma_reset_mask(struct device *dev,
+						struct device_attribute *attr,
+						char *buf)
+{
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = drm_to_adev(ddev);
+
+	if (!adev)
+		return -ENODEV;
+
+	return amdgpu_show_reset_mask(buf, adev->sdma.supported_reset);
+}
+
+static DEVICE_ATTR(sdma_reset_mask, 0444,
+		   amdgpu_get_sdma_reset_mask, NULL);
+
+int amdgpu_sdma_sysfs_reset_mask_init(struct amdgpu_device *adev)
+{
+	int r = 0;
+
+	if (!amdgpu_gpu_recovery)
+		return r;
+
+	if (adev->sdma.num_instances) {
+		r = device_create_file(adev->dev, &dev_attr_sdma_reset_mask);
+		if (r)
+			return r;
+	}
+
+	return r;
+}
+
+void amdgpu_sdma_sysfs_reset_mask_fini(struct amdgpu_device *adev)
+{
+	if (!amdgpu_gpu_recovery)
+		return;
+
+	if (adev->dev->kobj.sd) {
+		if (adev->sdma.num_instances)
+			device_remove_file(adev->dev, &dev_attr_sdma_reset_mask);
+	}
+}
+
+struct amdgpu_ring *amdgpu_sdma_get_shared_ring(struct amdgpu_device *adev, struct amdgpu_ring *ring)
+{
+	if (adev->sdma.has_page_queue &&
+	    (ring->me < adev->sdma.num_instances) &&
+	    (ring == &adev->sdma.instance[ring->me].ring))
+		return &adev->sdma.instance[ring->me].page;
+	else
+		return NULL;
+}
+
+/**
+* amdgpu_sdma_is_shared_inv_eng - Check if a ring is an SDMA ring that shares a VM invalidation engine
+* @adev: Pointer to the AMDGPU device structure
+* @ring: Pointer to the ring structure to check
+*
+* This function checks if the given ring is an SDMA ring that shares a VM invalidation engine.
+* It returns true if the ring is such an SDMA ring, false otherwise.
+*/
+bool amdgpu_sdma_is_shared_inv_eng(struct amdgpu_device *adev, struct amdgpu_ring *ring)
+{
+	int i = ring->me;
+
+	if (!adev->sdma.has_page_queue || i >= adev->sdma.num_instances)
+		return false;
+
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3) ||
+	    amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4) ||
+	    amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 5, 0))
+		return (ring == &adev->sdma.instance[i].page);
+	else
+		return false;
+}
+
+static int amdgpu_sdma_soft_reset(struct amdgpu_device *adev, u32 instance_id)
+{
+	struct amdgpu_sdma_instance *sdma_instance = &adev->sdma.instance[instance_id];
+
+	if (sdma_instance->funcs->soft_reset_kernel_queue)
+		return sdma_instance->funcs->soft_reset_kernel_queue(adev, instance_id);
+
+	return -EOPNOTSUPP;
+}
+
+/**
+ * amdgpu_sdma_reset_engine - Reset a specific SDMA engine
+ * @adev: Pointer to the AMDGPU device
+ * @instance_id: Logical ID of the SDMA engine instance to reset
+ * @caller_handles_kernel_queues: Skip kernel queue processing. Caller
+ * will handle it.
+ *
+ * Returns: 0 on success, or a negative error code on failure.
+ */
+int amdgpu_sdma_reset_engine(struct amdgpu_device *adev, uint32_t instance_id,
+			     bool caller_handles_kernel_queues)
+{
+	int ret = 0;
+	struct amdgpu_sdma_instance *sdma_instance = &adev->sdma.instance[instance_id];
+	struct amdgpu_ring *gfx_ring = &sdma_instance->ring;
+	struct amdgpu_ring *page_ring = &sdma_instance->page;
+
+	mutex_lock(&sdma_instance->engine_reset_mutex);
+
+	if (!caller_handles_kernel_queues) {
+		/* Stop the scheduler's work queue for the GFX and page rings if they are running.
+		 * This ensures that no new tasks are submitted to the queues while
+		 * the reset is in progress.
+		 */
+		drm_sched_wqueue_stop(&gfx_ring->sched);
+
+		if (adev->sdma.has_page_queue)
+			drm_sched_wqueue_stop(&page_ring->sched);
+	}
+
+	if (sdma_instance->funcs->stop_kernel_queue) {
+		sdma_instance->funcs->stop_kernel_queue(gfx_ring);
+		if (adev->sdma.has_page_queue)
+			sdma_instance->funcs->stop_kernel_queue(page_ring);
+	}
+
+	/* Perform the SDMA reset for the specified instance */
+	ret = amdgpu_sdma_soft_reset(adev, instance_id);
+	if (ret) {
+		dev_err(adev->dev, "Failed to reset SDMA logical instance %u\n", instance_id);
+		goto exit;
+	}
+
+	if (sdma_instance->funcs->start_kernel_queue) {
+		sdma_instance->funcs->start_kernel_queue(gfx_ring);
+		if (adev->sdma.has_page_queue)
+			sdma_instance->funcs->start_kernel_queue(page_ring);
+	}
+
+exit:
+	if (!caller_handles_kernel_queues) {
+		/* Restart the scheduler's work queue for the GFX and page rings
+		 * if they were stopped by this function. This allows new tasks
+		 * to be submitted to the queues after the reset is complete.
+		 */
+		if (!ret) {
+			amdgpu_fence_driver_force_completion(gfx_ring);
+			drm_sched_wqueue_start(&gfx_ring->sched);
+			if (adev->sdma.has_page_queue) {
+				amdgpu_fence_driver_force_completion(page_ring);
+				drm_sched_wqueue_start(&page_ring->sched);
+			}
+		}
+	}
+	mutex_unlock(&sdma_instance->engine_reset_mutex);
+
+	return ret;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.h
index f8fb755e3aa6..34311f32be4c 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.h
@@ -23,9 +23,10 @@
 
 #ifndef __AMDGPU_SDMA_H__
 #define __AMDGPU_SDMA_H__
+#include "amdgpu_ras.h"
 
 /* max number of IP instances */
-#define AMDGPU_MAX_SDMA_INSTANCES		8
+#define AMDGPU_MAX_SDMA_INSTANCES		16
 
 enum amdgpu_sdma_irq {
 	AMDGPU_SDMA_IRQ_INSTANCE0  = 0,
@@ -36,9 +37,25 @@ enum amdgpu_sdma_irq {
 	AMDGPU_SDMA_IRQ_INSTANCE5,
 	AMDGPU_SDMA_IRQ_INSTANCE6,
 	AMDGPU_SDMA_IRQ_INSTANCE7,
+	AMDGPU_SDMA_IRQ_INSTANCE8,
+	AMDGPU_SDMA_IRQ_INSTANCE9,
+	AMDGPU_SDMA_IRQ_INSTANCE10,
+	AMDGPU_SDMA_IRQ_INSTANCE11,
+	AMDGPU_SDMA_IRQ_INSTANCE12,
+	AMDGPU_SDMA_IRQ_INSTANCE13,
+	AMDGPU_SDMA_IRQ_INSTANCE14,
+	AMDGPU_SDMA_IRQ_INSTANCE15,
 	AMDGPU_SDMA_IRQ_LAST
 };
 
+#define NUM_SDMA(x) hweight32(x)
+
+struct amdgpu_sdma_funcs {
+	int (*stop_kernel_queue)(struct amdgpu_ring *ring);
+	int (*start_kernel_queue)(struct amdgpu_ring *ring);
+	int (*soft_reset_kernel_queue)(struct amdgpu_device *adev, u32 instance_id);
+};
+
 struct amdgpu_sdma_instance {
 	/* SDMA firmware */
 	const struct firmware	*fw;
@@ -48,32 +65,74 @@ struct amdgpu_sdma_instance {
 	struct amdgpu_ring	ring;
 	struct amdgpu_ring	page;
 	bool			burst_nop;
+	uint32_t		aid_id;
+
+	struct amdgpu_bo	*sdma_fw_obj;
+	uint64_t		sdma_fw_gpu_addr;
+	uint32_t		*sdma_fw_ptr;
+	struct mutex		engine_reset_mutex;
+	/* track guilty state of GFX and PAGE queues */
+	bool			gfx_guilty;
+	bool			page_guilty;
+	const struct amdgpu_sdma_funcs   *funcs;
 };
 
-struct amdgpu_sdma_ras_funcs {
-	int (*ras_late_init)(struct amdgpu_device *adev,
-			void *ras_ih_info);
-	void (*ras_fini)(struct amdgpu_device *adev);
-	int (*query_ras_error_count)(struct amdgpu_device *adev,
-			uint32_t instance, void *ras_error_status);
-	void (*reset_ras_error_count)(struct amdgpu_device *adev);
+enum amdgpu_sdma_ras_memory_id {
+	AMDGPU_SDMA_MBANK_DATA_BUF0 = 1,
+	AMDGPU_SDMA_MBANK_DATA_BUF1 = 2,
+	AMDGPU_SDMA_MBANK_DATA_BUF2 = 3,
+	AMDGPU_SDMA_MBANK_DATA_BUF3 = 4,
+	AMDGPU_SDMA_MBANK_DATA_BUF4 = 5,
+	AMDGPU_SDMA_MBANK_DATA_BUF5 = 6,
+	AMDGPU_SDMA_MBANK_DATA_BUF6 = 7,
+	AMDGPU_SDMA_MBANK_DATA_BUF7 = 8,
+	AMDGPU_SDMA_MBANK_DATA_BUF8 = 9,
+	AMDGPU_SDMA_MBANK_DATA_BUF9 = 10,
+	AMDGPU_SDMA_MBANK_DATA_BUF10 = 11,
+	AMDGPU_SDMA_MBANK_DATA_BUF11 = 12,
+	AMDGPU_SDMA_MBANK_DATA_BUF12 = 13,
+	AMDGPU_SDMA_MBANK_DATA_BUF13 = 14,
+	AMDGPU_SDMA_MBANK_DATA_BUF14 = 15,
+	AMDGPU_SDMA_MBANK_DATA_BUF15 = 16,
+	AMDGPU_SDMA_UCODE_BUF = 17,
+	AMDGPU_SDMA_RB_CMD_BUF = 18,
+	AMDGPU_SDMA_IB_CMD_BUF = 19,
+	AMDGPU_SDMA_UTCL1_RD_FIFO = 20,
+	AMDGPU_SDMA_UTCL1_RDBST_FIFO = 21,
+	AMDGPU_SDMA_UTCL1_WR_FIFO = 22,
+	AMDGPU_SDMA_DATA_LUT_FIFO = 23,
+	AMDGPU_SDMA_SPLIT_DAT_BUF = 24,
+	AMDGPU_SDMA_MEMORY_BLOCK_LAST,
+};
+
+struct amdgpu_sdma_ras {
+	struct amdgpu_ras_block_object ras_block;
 };
 
 struct amdgpu_sdma {
 	struct amdgpu_sdma_instance instance[AMDGPU_MAX_SDMA_INSTANCES];
 	struct amdgpu_irq_src	trap_irq;
 	struct amdgpu_irq_src	illegal_inst_irq;
+	struct amdgpu_irq_src	fence_irq;
 	struct amdgpu_irq_src	ecc_irq;
 	struct amdgpu_irq_src	vm_hole_irq;
 	struct amdgpu_irq_src	doorbell_invalid_irq;
 	struct amdgpu_irq_src	pool_timeout_irq;
 	struct amdgpu_irq_src	srbm_write_irq;
+	struct amdgpu_irq_src	ctxt_empty_irq;
 
 	int			num_instances;
+	uint32_t 		sdma_mask;
+	int			num_inst_per_aid;
 	uint32_t                    srbm_soft_reset;
 	bool			has_page_queue;
 	struct ras_common_if	*ras_if;
-	const struct amdgpu_sdma_ras_funcs	*funcs;
+	struct amdgpu_sdma_ras	*ras;
+	uint32_t		*ip_dump;
+	uint32_t 		supported_reset;
+	struct list_head	reset_callback_list;
+	bool			no_user_submission;
+	bool			disable_uq;
 };
 
 /*
@@ -95,7 +154,7 @@ struct amdgpu_buffer_funcs {
 				 uint64_t dst_offset,
 				 /* number of byte to transfer */
 				 uint32_t byte_count,
-				 bool tmz);
+				 uint32_t copy_flags);
 
 	/* maximum bytes in a single operation */
 	uint32_t	fill_max_bytes;
@@ -113,6 +172,9 @@ struct amdgpu_buffer_funcs {
 				 uint32_t byte_count);
 };
 
+int amdgpu_sdma_reset_engine(struct amdgpu_device *adev, uint32_t instance_id,
+			     bool caller_handles_kernel_queues);
+
 #define amdgpu_emit_copy_buffer(adev, ib, s, d, b, t) (adev)->mman.buffer_funcs->emit_copy_buffer((ib),  (s), (d), (b), (t))
 #define amdgpu_emit_fill_buffer(adev, ib, s, d, b) (adev)->mman.buffer_funcs->emit_fill_buffer((ib), (s), (d), (b))
 
@@ -121,12 +183,22 @@ amdgpu_sdma_get_instance_from_ring(struct amdgpu_ring *ring);
 int amdgpu_sdma_get_index_from_ring(struct amdgpu_ring *ring, uint32_t *index);
 uint64_t amdgpu_sdma_get_csa_mc_addr(struct amdgpu_ring *ring, unsigned vmid);
 int amdgpu_sdma_ras_late_init(struct amdgpu_device *adev,
-			      void *ras_ih_info);
-void amdgpu_sdma_ras_fini(struct amdgpu_device *adev);
+			      struct ras_common_if *ras_block);
 int amdgpu_sdma_process_ras_data_cb(struct amdgpu_device *adev,
 		void *err_data,
 		struct amdgpu_iv_entry *entry);
 int amdgpu_sdma_process_ecc_irq(struct amdgpu_device *adev,
 				      struct amdgpu_irq_src *source,
 				      struct amdgpu_iv_entry *entry);
+int amdgpu_sdma_init_microcode(struct amdgpu_device *adev, u32 instance,
+			       bool duplicate);
+void amdgpu_sdma_destroy_inst_ctx(struct amdgpu_device *adev,
+        bool duplicate);
+int amdgpu_sdma_ras_sw_init(struct amdgpu_device *adev);
+void amdgpu_debugfs_sdma_sched_mask_init(struct amdgpu_device *adev);
+int amdgpu_sdma_sysfs_reset_mask_init(struct amdgpu_device *adev);
+void amdgpu_sdma_sysfs_reset_mask_fini(struct amdgpu_device *adev);
+bool amdgpu_sdma_is_shared_inv_eng(struct amdgpu_device *adev, struct amdgpu_ring *ring);
+struct amdgpu_ring *amdgpu_sdma_get_shared_ring(struct amdgpu_device *adev,
+	struct amdgpu_ring *ring);
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_securedisplay.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_securedisplay.c
index cc7597a15fe9..3739be1b71e0 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_securedisplay.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_securedisplay.c
@@ -77,11 +77,11 @@ void psp_securedisplay_parse_resp_status(struct psp_context *psp,
 	}
 }
 
-void psp_prep_securedisplay_cmd_buf(struct psp_context *psp, struct securedisplay_cmd **cmd,
+void psp_prep_securedisplay_cmd_buf(struct psp_context *psp, struct ta_securedisplay_cmd **cmd,
 	enum ta_securedisplay_command command_id)
 {
-	*cmd = (struct securedisplay_cmd *)psp->securedisplay_context.context.mem_context.shared_buf;
-	memset(*cmd, 0, sizeof(struct securedisplay_cmd));
+	*cmd = (struct ta_securedisplay_cmd *)psp->securedisplay_context.context.mem_context.shared_buf;
+	memset(*cmd, 0, sizeof(struct ta_securedisplay_cmd));
 	(*cmd)->status = TA_SECUREDISPLAY_STATUS__GENERIC_FAILURE;
 	(*cmd)->cmd_id = command_id;
 }
@@ -93,7 +93,7 @@ static ssize_t amdgpu_securedisplay_debugfs_write(struct file *f, const char __u
 {
 	struct amdgpu_device *adev = (struct amdgpu_device *)file_inode(f)->i_private;
 	struct psp_context *psp = &adev->psp;
-	struct securedisplay_cmd *securedisplay_cmd;
+	struct ta_securedisplay_cmd *securedisplay_cmd;
 	struct drm_device *dev = adev_to_drm(adev);
 	uint32_t phy_id;
 	uint32_t op;
@@ -121,6 +121,7 @@ static ssize_t amdgpu_securedisplay_debugfs_write(struct file *f, const char __u
 
 	switch (op) {
 	case 1:
+		mutex_lock(&psp->securedisplay_context.mutex);
 		psp_prep_securedisplay_cmd_buf(psp, &securedisplay_cmd,
 			TA_SECUREDISPLAY_COMMAND__QUERY_TA);
 		ret = psp_securedisplay_invoke(psp, TA_SECUREDISPLAY_COMMAND__QUERY_TA);
@@ -131,8 +132,14 @@ static ssize_t amdgpu_securedisplay_debugfs_write(struct file *f, const char __u
 			else
 				psp_securedisplay_parse_resp_status(psp, securedisplay_cmd->status);
 		}
+		mutex_unlock(&psp->securedisplay_context.mutex);
 		break;
 	case 2:
+		if (size < 3 || phy_id >= TA_SECUREDISPLAY_MAX_PHY) {
+			dev_err(adev->dev, "Invalid input: %s\n", str);
+			return -EINVAL;
+		}
+		mutex_lock(&psp->securedisplay_context.mutex);
 		psp_prep_securedisplay_cmd_buf(psp, &securedisplay_cmd,
 			TA_SECUREDISPLAY_COMMAND__SEND_ROI_CRC);
 		securedisplay_cmd->securedisplay_in_message.send_roi_crc.phy_id = phy_id;
@@ -146,12 +153,12 @@ static ssize_t amdgpu_securedisplay_debugfs_write(struct file *f, const char __u
 				psp_securedisplay_parse_resp_status(psp, securedisplay_cmd->status);
 			}
 		}
+		mutex_unlock(&psp->securedisplay_context.mutex);
 		break;
 	default:
 		dev_err(adev->dev, "Invalid input: %s\n", str);
 	}
 
-	pm_runtime_mark_last_busy(dev->dev);
 	pm_runtime_put_autosuspend(dev->dev);
 
 	return size;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_securedisplay.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_securedisplay.h
index fe98574748f4..456ad68ed4b2 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_securedisplay.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_securedisplay.h
@@ -30,7 +30,7 @@
 void amdgpu_securedisplay_debugfs_init(struct amdgpu_device *adev);
 void psp_securedisplay_parse_resp_status(struct psp_context *psp,
 		enum ta_securedisplay_status status);
-void psp_prep_securedisplay_cmd_buf(struct psp_context *psp, struct securedisplay_cmd **cmd,
+void psp_prep_securedisplay_cmd_buf(struct psp_context *psp, struct ta_securedisplay_cmd **cmd,
 		enum ta_securedisplay_command command_id);
 
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_seq64.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_seq64.c
new file mode 100644
index 000000000000..a0b479d5fff1
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_seq64.c
@@ -0,0 +1,262 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "amdgpu.h"
+#include "amdgpu_seq64.h"
+
+#include <drm/drm_exec.h>
+
+/**
+ * DOC: amdgpu_seq64
+ *
+ * amdgpu_seq64 allocates a 64bit memory on each request in sequence order.
+ * seq64 driver is required for user queue fence memory allocation, TLB
+ * counters and VM updates. It has maximum count of 32768 64 bit slots.
+ */
+
+/**
+ * amdgpu_seq64_get_va_base - Get the seq64 va base address
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Returns:
+ * va base address on success
+ */
+static inline u64 amdgpu_seq64_get_va_base(struct amdgpu_device *adev)
+{
+	u64 addr = AMDGPU_VA_RESERVED_SEQ64_START(adev);
+
+	addr = amdgpu_gmc_sign_extend(addr);
+
+	return addr;
+}
+
+/**
+ * amdgpu_seq64_map - Map the seq64 memory to VM
+ *
+ * @adev: amdgpu_device pointer
+ * @vm: vm pointer
+ * @bo_va: bo_va pointer
+ *
+ * Map the seq64 memory to the given VM.
+ *
+ * Returns:
+ * 0 on success or a negative error code on failure
+ */
+int amdgpu_seq64_map(struct amdgpu_device *adev, struct amdgpu_vm *vm,
+		     struct amdgpu_bo_va **bo_va)
+{
+	struct amdgpu_bo *bo;
+	struct drm_exec exec;
+	u64 seq64_addr;
+	int r;
+
+	bo = adev->seq64.sbo;
+	if (!bo)
+		return -EINVAL;
+
+	drm_exec_init(&exec, DRM_EXEC_INTERRUPTIBLE_WAIT, 0);
+	drm_exec_until_all_locked(&exec) {
+		r = amdgpu_vm_lock_pd(vm, &exec, 0);
+		if (likely(!r))
+			r = drm_exec_lock_obj(&exec, &bo->tbo.base);
+		drm_exec_retry_on_contention(&exec);
+		if (unlikely(r))
+			goto error;
+	}
+
+	*bo_va = amdgpu_vm_bo_add(adev, vm, bo);
+	if (!*bo_va) {
+		r = -ENOMEM;
+		goto error;
+	}
+
+	seq64_addr = amdgpu_seq64_get_va_base(adev) & AMDGPU_GMC_HOLE_MASK;
+
+	r = amdgpu_vm_bo_map(adev, *bo_va, seq64_addr, 0,
+			     AMDGPU_VA_RESERVED_SEQ64_SIZE,
+			     AMDGPU_VM_PAGE_READABLE | AMDGPU_VM_MTYPE_UC);
+	if (r) {
+		DRM_ERROR("failed to do bo_map on userq sem, err=%d\n", r);
+		amdgpu_vm_bo_del(adev, *bo_va);
+		goto error;
+	}
+
+	r = amdgpu_vm_bo_update(adev, *bo_va, false);
+	if (r) {
+		DRM_ERROR("failed to do vm_bo_update on userq sem\n");
+		amdgpu_vm_bo_del(adev, *bo_va);
+		goto error;
+	}
+
+error:
+	drm_exec_fini(&exec);
+	return r;
+}
+
+/**
+ * amdgpu_seq64_unmap - Unmap the seq64 memory
+ *
+ * @adev: amdgpu_device pointer
+ * @fpriv: DRM file private
+ *
+ * Unmap the seq64 memory from the given VM.
+ */
+void amdgpu_seq64_unmap(struct amdgpu_device *adev, struct amdgpu_fpriv *fpriv)
+{
+	struct amdgpu_vm *vm;
+	struct amdgpu_bo *bo;
+	struct drm_exec exec;
+	int r;
+
+	if (!fpriv->seq64_va)
+		return;
+
+	bo = adev->seq64.sbo;
+	if (!bo)
+		return;
+
+	vm = &fpriv->vm;
+
+	drm_exec_init(&exec, 0, 0);
+	drm_exec_until_all_locked(&exec) {
+		r = amdgpu_vm_lock_pd(vm, &exec, 0);
+		if (likely(!r))
+			r = drm_exec_lock_obj(&exec, &bo->tbo.base);
+		drm_exec_retry_on_contention(&exec);
+		if (unlikely(r))
+			goto error;
+	}
+
+	amdgpu_vm_bo_del(adev, fpriv->seq64_va);
+
+	fpriv->seq64_va = NULL;
+
+error:
+	drm_exec_fini(&exec);
+}
+
+/**
+ * amdgpu_seq64_alloc - Allocate a 64 bit memory
+ *
+ * @adev: amdgpu_device pointer
+ * @va: VA to access the seq in process address space
+ * @gpu_addr: GPU address to access the seq
+ * @cpu_addr: CPU address to access the seq
+ *
+ * Alloc a 64 bit memory from seq64 pool.
+ *
+ * Returns:
+ * 0 on success or a negative error code on failure
+ */
+int amdgpu_seq64_alloc(struct amdgpu_device *adev, u64 *va,
+		       u64 *gpu_addr, u64 **cpu_addr)
+{
+	unsigned long bit_pos;
+
+	bit_pos = find_first_zero_bit(adev->seq64.used, adev->seq64.num_sem);
+	if (bit_pos >= adev->seq64.num_sem)
+		return -ENOSPC;
+
+	__set_bit(bit_pos, adev->seq64.used);
+
+	*va = bit_pos * sizeof(u64) + amdgpu_seq64_get_va_base(adev);
+
+	if (gpu_addr)
+		*gpu_addr = bit_pos * sizeof(u64) + adev->seq64.gpu_addr;
+
+	*cpu_addr = bit_pos + adev->seq64.cpu_base_addr;
+
+	return 0;
+}
+
+/**
+ * amdgpu_seq64_free - Free the given 64 bit memory
+ *
+ * @adev: amdgpu_device pointer
+ * @va: gpu start address to be freed
+ *
+ * Free the given 64 bit memory from seq64 pool.
+ */
+void amdgpu_seq64_free(struct amdgpu_device *adev, u64 va)
+{
+	unsigned long bit_pos;
+
+	bit_pos = (va - amdgpu_seq64_get_va_base(adev)) / sizeof(u64);
+	if (bit_pos < adev->seq64.num_sem)
+		__clear_bit(bit_pos, adev->seq64.used);
+}
+
+/**
+ * amdgpu_seq64_fini - Cleanup seq64 driver
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Free the memory space allocated for seq64.
+ *
+ */
+void amdgpu_seq64_fini(struct amdgpu_device *adev)
+{
+	amdgpu_bo_free_kernel(&adev->seq64.sbo,
+			      NULL,
+			      (void **)&adev->seq64.cpu_base_addr);
+}
+
+/**
+ * amdgpu_seq64_init - Initialize seq64 driver
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Allocate the required memory space for seq64.
+ *
+ * Returns:
+ * 0 on success or a negative error code on failure
+ */
+int amdgpu_seq64_init(struct amdgpu_device *adev)
+{
+	int r;
+
+	if (adev->seq64.sbo)
+		return 0;
+
+	/*
+	 * AMDGPU_MAX_SEQ64_SLOTS * sizeof(u64) * 8 = AMDGPU_MAX_SEQ64_SLOTS
+	 * 64bit slots
+	 */
+	r = amdgpu_bo_create_kernel(adev, AMDGPU_VA_RESERVED_SEQ64_SIZE,
+				    PAGE_SIZE, AMDGPU_GEM_DOMAIN_GTT,
+				    &adev->seq64.sbo, &adev->seq64.gpu_addr,
+				    (void **)&adev->seq64.cpu_base_addr);
+	if (r) {
+		dev_warn(adev->dev, "(%d) create seq64 failed\n", r);
+		return r;
+	}
+
+	memset(adev->seq64.cpu_base_addr, 0, AMDGPU_VA_RESERVED_SEQ64_SIZE);
+
+	adev->seq64.num_sem = AMDGPU_MAX_SEQ64_SLOTS;
+	memset(&adev->seq64.used, 0, sizeof(adev->seq64.used));
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_seq64.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_seq64.h
new file mode 100644
index 000000000000..26a249aaaee1
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_seq64.h
@@ -0,0 +1,49 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __AMDGPU_SEQ64_H__
+#define __AMDGPU_SEQ64_H__
+
+#include "amdgpu_vm.h"
+
+#define AMDGPU_MAX_SEQ64_SLOTS         (AMDGPU_VA_RESERVED_SEQ64_SIZE / sizeof(u64))
+
+struct amdgpu_seq64 {
+	struct amdgpu_bo *sbo;
+	u32 num_sem;
+	u64 gpu_addr;
+	u64 *cpu_base_addr;
+	DECLARE_BITMAP(used, AMDGPU_MAX_SEQ64_SLOTS);
+};
+
+void amdgpu_seq64_fini(struct amdgpu_device *adev);
+int amdgpu_seq64_init(struct amdgpu_device *adev);
+int amdgpu_seq64_alloc(struct amdgpu_device *adev, u64 *va, u64 *gpu_addr, u64 **cpu_addr);
+void amdgpu_seq64_free(struct amdgpu_device *adev, u64 gpu_addr);
+int amdgpu_seq64_map(struct amdgpu_device *adev, struct amdgpu_vm *vm,
+		     struct amdgpu_bo_va **bo_va);
+void amdgpu_seq64_unmap(struct amdgpu_device *adev, struct amdgpu_fpriv *fpriv);
+
+#endif
+
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_smuio.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_smuio.h
index 484bb3dcec47..ec9d12f85f39 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_smuio.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_smuio.h
@@ -23,14 +23,28 @@
 #ifndef __AMDGPU_SMUIO_H__
 #define __AMDGPU_SMUIO_H__
 
+enum amdgpu_pkg_type {
+	AMDGPU_PKG_TYPE_APU = 2,
+	AMDGPU_PKG_TYPE_CEM = 3,
+	AMDGPU_PKG_TYPE_OAM = 4,
+	AMDGPU_PKG_TYPE_UNKNOWN,
+};
+
+struct amdgpu_smuio_mcm_config_info {
+	int socket_id;
+	int die_id;
+};
+
 struct amdgpu_smuio_funcs {
 	u32 (*get_rom_index_offset)(struct amdgpu_device *adev);
 	u32 (*get_rom_data_offset)(struct amdgpu_device *adev);
 	void (*update_rom_clock_gating)(struct amdgpu_device *adev, bool enable);
-	void (*get_clock_gating_state)(struct amdgpu_device *adev, u32 *flags);
+	void (*get_clock_gating_state)(struct amdgpu_device *adev, u64 *flags);
 	u32 (*get_die_id)(struct amdgpu_device *adev);
 	u32 (*get_socket_id)(struct amdgpu_device *adev);
+	enum amdgpu_pkg_type (*get_pkg_type)(struct amdgpu_device *adev);
 	bool (*is_host_gpu_xgmi_supported)(struct amdgpu_device *adev);
+	u64 (*get_gpu_clock_counter)(struct amdgpu_device *adev);
 };
 
 struct amdgpu_smuio {
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c
index 862eb3c1c4c5..d6ae9974c952 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c
@@ -1,3 +1,4 @@
+// SPDX-License-Identifier: MIT
 /*
  * Copyright 2014 Advanced Micro Devices, Inc.
  * All Rights Reserved.
@@ -51,7 +52,6 @@ static struct kmem_cache *amdgpu_sync_slab;
 void amdgpu_sync_create(struct amdgpu_sync *sync)
 {
 	hash_init(sync->fences);
-	sync->last_vm_update = NULL;
 }
 
 /**
@@ -135,11 +135,16 @@ static bool amdgpu_sync_add_later(struct amdgpu_sync *sync, struct dma_fence *f)
 	struct amdgpu_sync_entry *e;
 
 	hash_for_each_possible(sync->fences, e, node, f->context) {
-		if (unlikely(e->fence->context != f->context))
-			continue;
+		if (dma_fence_is_signaled(e->fence)) {
+			dma_fence_put(e->fence);
+			e->fence = dma_fence_get(f);
+			return true;
+		}
 
-		amdgpu_sync_keep_later(&e->fence, f);
-		return true;
+		if (likely(e->fence->context == f->context)) {
+			amdgpu_sync_keep_later(&e->fence, f);
+			return true;
+		}
 	}
 	return false;
 }
@@ -149,10 +154,12 @@ static bool amdgpu_sync_add_later(struct amdgpu_sync *sync, struct dma_fence *f)
  *
  * @sync: sync object to add fence to
  * @f: fence to sync to
+ * @flags: memory allocation flags to use when allocating sync entry
  *
  * Add the fence to the sync object.
  */
-int amdgpu_sync_fence(struct amdgpu_sync *sync, struct dma_fence *f)
+int amdgpu_sync_fence(struct amdgpu_sync *sync, struct dma_fence *f,
+		      gfp_t flags)
 {
 	struct amdgpu_sync_entry *e;
 
@@ -162,7 +169,7 @@ int amdgpu_sync_fence(struct amdgpu_sync *sync, struct dma_fence *f)
 	if (amdgpu_sync_add_later(sync, f))
 		return 0;
 
-	e = kmem_cache_alloc(amdgpu_sync_slab, GFP_KERNEL);
+	e = kmem_cache_alloc(amdgpu_sync_slab, flags);
 	if (!e)
 		return -ENOMEM;
 
@@ -171,23 +178,6 @@ int amdgpu_sync_fence(struct amdgpu_sync *sync, struct dma_fence *f)
 	return 0;
 }
 
-/**
- * amdgpu_sync_vm_fence - remember to sync to this VM fence
- *
- * @sync: sync object to add fence to
- * @fence: the VM fence to add
- *
- * Add the fence to the sync object and remember it as VM update.
- */
-int amdgpu_sync_vm_fence(struct amdgpu_sync *sync, struct dma_fence *fence)
-{
-	if (!fence)
-		return 0;
-
-	amdgpu_sync_keep_later(&sync->last_vm_update, fence);
-	return amdgpu_sync_fence(sync, fence);
-}
-
 /* Determine based on the owner and mode if we should sync to a fence or not */
 static bool amdgpu_sync_test_fence(struct amdgpu_device *adev,
 				   enum amdgpu_sync_mode mode,
@@ -208,7 +198,8 @@ static bool amdgpu_sync_test_fence(struct amdgpu_device *adev,
 
 	/* Never sync to VM updates either. */
 	if (fence_owner == AMDGPU_FENCE_OWNER_VM &&
-	    owner != AMDGPU_FENCE_OWNER_UNDEFINED)
+	    owner != AMDGPU_FENCE_OWNER_UNDEFINED &&
+	    owner != AMDGPU_FENCE_OWNER_KFD)
 		return false;
 
 	/* Ignore fences depending on the sync mode */
@@ -252,44 +243,65 @@ int amdgpu_sync_resv(struct amdgpu_device *adev, struct amdgpu_sync *sync,
 		     struct dma_resv *resv, enum amdgpu_sync_mode mode,
 		     void *owner)
 {
-	struct dma_resv_list *flist;
+	struct dma_resv_iter cursor;
 	struct dma_fence *f;
-	unsigned i;
-	int r = 0;
+	int r;
 
 	if (resv == NULL)
 		return -EINVAL;
-
-	/* always sync to the exclusive fence */
-	f = dma_resv_excl_fence(resv);
-	dma_fence_chain_for_each(f, f) {
-		struct dma_fence_chain *chain = to_dma_fence_chain(f);
-
-		if (amdgpu_sync_test_fence(adev, mode, owner, chain ?
-					   chain->fence : f)) {
-			r = amdgpu_sync_fence(sync, f);
-			dma_fence_put(f);
-			if (r)
-				return r;
-			break;
+	/* Implicitly sync only to KERNEL, WRITE and READ */
+	dma_resv_for_each_fence(&cursor, resv, DMA_RESV_USAGE_READ, f) {
+		dma_fence_chain_for_each(f, f) {
+			struct dma_fence *tmp = dma_fence_chain_contained(f);
+
+			if (amdgpu_sync_test_fence(adev, mode, owner, tmp)) {
+				r = amdgpu_sync_fence(sync, f, GFP_KERNEL);
+				dma_fence_put(f);
+				if (r)
+					return r;
+				break;
+			}
 		}
 	}
+	return 0;
+}
 
-	flist = dma_resv_shared_list(resv);
-	if (!flist)
-		return 0;
+/**
+ * amdgpu_sync_kfd - sync to KFD fences
+ *
+ * @sync: sync object to add KFD fences to
+ * @resv: reservation object with KFD fences
+ *
+ * Extract all KFD fences and add them to the sync object.
+ */
+int amdgpu_sync_kfd(struct amdgpu_sync *sync, struct dma_resv *resv)
+{
+	struct dma_resv_iter cursor;
+	struct dma_fence *f;
+	int r = 0;
 
-	for (i = 0; i < flist->shared_count; ++i) {
-		f = rcu_dereference_protected(flist->shared[i],
-					      dma_resv_held(resv));
+	dma_resv_iter_begin(&cursor, resv, DMA_RESV_USAGE_BOOKKEEP);
+	dma_resv_for_each_fence_unlocked(&cursor, f) {
+		void *fence_owner = amdgpu_sync_get_owner(f);
 
-		if (amdgpu_sync_test_fence(adev, mode, owner, f)) {
-			r = amdgpu_sync_fence(sync, f);
-			if (r)
-				return r;
-		}
+		if (fence_owner != AMDGPU_FENCE_OWNER_KFD)
+			continue;
+
+		r = amdgpu_sync_fence(sync, f, GFP_KERNEL);
+		if (r)
+			break;
 	}
-	return 0;
+	dma_resv_iter_end(&cursor);
+
+	return r;
+}
+
+/* Free the entry back to the slab */
+static void amdgpu_sync_entry_free(struct amdgpu_sync_entry *e)
+{
+	hash_del(&e->node);
+	dma_fence_put(e->fence);
+	kmem_cache_free(amdgpu_sync_slab, e);
 }
 
 /**
@@ -313,9 +325,7 @@ struct dma_fence *amdgpu_sync_peek_fence(struct amdgpu_sync *sync,
 		struct drm_sched_fence *s_fence = to_drm_sched_fence(f);
 
 		if (dma_fence_is_signaled(f)) {
-			hash_del(&e->node);
-			dma_fence_put(f);
-			kmem_cache_free(amdgpu_sync_slab, e);
+			amdgpu_sync_entry_free(e);
 			continue;
 		}
 		if (ring && s_fence) {
@@ -349,6 +359,7 @@ struct dma_fence *amdgpu_sync_get_fence(struct amdgpu_sync *sync)
 	struct hlist_node *tmp;
 	struct dma_fence *f;
 	int i;
+
 	hash_for_each_safe(sync->fences, i, tmp, e, node) {
 
 		f = e->fence;
@@ -383,19 +394,64 @@ int amdgpu_sync_clone(struct amdgpu_sync *source, struct amdgpu_sync *clone)
 	hash_for_each_safe(source->fences, i, tmp, e, node) {
 		f = e->fence;
 		if (!dma_fence_is_signaled(f)) {
-			r = amdgpu_sync_fence(clone, f);
+			r = amdgpu_sync_fence(clone, f, GFP_KERNEL);
 			if (r)
 				return r;
 		} else {
-			hash_del(&e->node);
-			dma_fence_put(f);
-			kmem_cache_free(amdgpu_sync_slab, e);
+			amdgpu_sync_entry_free(e);
 		}
 	}
 
-	dma_fence_put(clone->last_vm_update);
-	clone->last_vm_update = dma_fence_get(source->last_vm_update);
+	return 0;
+}
+
+/**
+ * amdgpu_sync_move - move all fences from src to dst
+ *
+ * @src: source of the fences, empty after function
+ * @dst: destination for the fences
+ *
+ * Moves all fences from source to destination. All fences in destination are
+ * freed and source is empty after the function call.
+ */
+void amdgpu_sync_move(struct amdgpu_sync *src, struct amdgpu_sync *dst)
+{
+	unsigned int i;
+
+	amdgpu_sync_free(dst);
+
+	for (i = 0; i < HASH_SIZE(src->fences); ++i)
+		hlist_move_list(&src->fences[i], &dst->fences[i]);
+}
+
+/**
+ * amdgpu_sync_push_to_job - push fences into job
+ * @sync: sync object to get the fences from
+ * @job: job to push the fences into
+ *
+ * Add all unsignaled fences from sync to job.
+ */
+int amdgpu_sync_push_to_job(struct amdgpu_sync *sync, struct amdgpu_job *job)
+{
+	struct amdgpu_sync_entry *e;
+	struct hlist_node *tmp;
+	struct dma_fence *f;
+	int i, r;
+
+	hash_for_each_safe(sync->fences, i, tmp, e, node) {
+		f = e->fence;
+		if (dma_fence_is_signaled(f)) {
+			amdgpu_sync_entry_free(e);
+			continue;
+		}
 
+		dma_fence_get(f);
+		r = drm_sched_job_add_dependency(&job->base, f);
+		if (r) {
+			dma_fence_put(f);
+			return r;
+		}
+	}
 	return 0;
 }
 
@@ -410,9 +466,7 @@ int amdgpu_sync_wait(struct amdgpu_sync *sync, bool intr)
 		if (r)
 			return r;
 
-		hash_del(&e->node);
-		dma_fence_put(e->fence);
-		kmem_cache_free(amdgpu_sync_slab, e);
+		amdgpu_sync_entry_free(e);
 	}
 
 	return 0;
@@ -429,15 +483,10 @@ void amdgpu_sync_free(struct amdgpu_sync *sync)
 {
 	struct amdgpu_sync_entry *e;
 	struct hlist_node *tmp;
-	unsigned i;
-
-	hash_for_each_safe(sync->fences, i, tmp, e, node) {
-		hash_del(&e->node);
-		dma_fence_put(e->fence);
-		kmem_cache_free(amdgpu_sync_slab, e);
-	}
+	unsigned int i;
 
-	dma_fence_put(sync->last_vm_update);
+	hash_for_each_safe(sync->fences, i, tmp, e, node)
+		amdgpu_sync_entry_free(e);
 }
 
 /**
@@ -447,9 +496,7 @@ void amdgpu_sync_free(struct amdgpu_sync *sync)
  */
 int amdgpu_sync_init(void)
 {
-	amdgpu_sync_slab = kmem_cache_create(
-		"amdgpu_sync", sizeof(struct amdgpu_sync_entry), 0,
-		SLAB_HWCACHE_ALIGN, NULL);
+	amdgpu_sync_slab = KMEM_CACHE(amdgpu_sync_entry, SLAB_HWCACHE_ALIGN);
 	if (!amdgpu_sync_slab)
 		return -ENOMEM;
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.h
index 7c0fe20c470d..51eb4382c91e 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.h
@@ -30,6 +30,7 @@ struct dma_fence;
 struct dma_resv;
 struct amdgpu_device;
 struct amdgpu_ring;
+struct amdgpu_job;
 
 enum amdgpu_sync_mode {
 	AMDGPU_SYNC_ALWAYS,
@@ -43,19 +44,21 @@ enum amdgpu_sync_mode {
  */
 struct amdgpu_sync {
 	DECLARE_HASHTABLE(fences, 4);
-	struct dma_fence	*last_vm_update;
 };
 
 void amdgpu_sync_create(struct amdgpu_sync *sync);
-int amdgpu_sync_fence(struct amdgpu_sync *sync, struct dma_fence *f);
-int amdgpu_sync_vm_fence(struct amdgpu_sync *sync, struct dma_fence *fence);
+int amdgpu_sync_fence(struct amdgpu_sync *sync, struct dma_fence *f,
+		      gfp_t flags);
 int amdgpu_sync_resv(struct amdgpu_device *adev, struct amdgpu_sync *sync,
 		     struct dma_resv *resv, enum amdgpu_sync_mode mode,
 		     void *owner);
+int amdgpu_sync_kfd(struct amdgpu_sync *sync, struct dma_resv *resv);
 struct dma_fence *amdgpu_sync_peek_fence(struct amdgpu_sync *sync,
 				     struct amdgpu_ring *ring);
 struct dma_fence *amdgpu_sync_get_fence(struct amdgpu_sync *sync);
 int amdgpu_sync_clone(struct amdgpu_sync *source, struct amdgpu_sync *clone);
+void amdgpu_sync_move(struct amdgpu_sync *src, struct amdgpu_sync *dst);
+int amdgpu_sync_push_to_job(struct amdgpu_sync *sync, struct amdgpu_job *job);
 int amdgpu_sync_wait(struct amdgpu_sync *sync, bool intr);
 void amdgpu_sync_free(struct amdgpu_sync *sync);
 int amdgpu_sync_init(void);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_test.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_test.c
deleted file mode 100644
index 909d830b513e..000000000000
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_test.c
+++ /dev/null
@@ -1,250 +0,0 @@
-// SPDX-License-Identifier: GPL-2.0 OR MIT
-/*
- * Copyright 2009 VMware, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: Michel Dänzer
- */
-
-#include <drm/amdgpu_drm.h>
-#include "amdgpu.h"
-#include "amdgpu_uvd.h"
-#include "amdgpu_vce.h"
-
-/* Test BO GTT->VRAM and VRAM->GTT GPU copies across the whole GTT aperture */
-static void amdgpu_do_test_moves(struct amdgpu_device *adev)
-{
-	struct amdgpu_ring *ring = adev->mman.buffer_funcs_ring;
-	struct amdgpu_bo *vram_obj = NULL;
-	struct amdgpu_bo **gtt_obj = NULL;
-	struct amdgpu_bo_param bp;
-	uint64_t gart_addr, vram_addr;
-	unsigned n, size;
-	int i, r;
-
-	size = 1024 * 1024;
-
-	/* Number of tests =
-	 * (Total GTT - gart_pin_size - (2 transfer windows for buffer moves)) / test size
-	 */
-	n = adev->gmc.gart_size - atomic64_read(&adev->gart_pin_size);
-	n -= AMDGPU_GTT_MAX_TRANSFER_SIZE * AMDGPU_GTT_NUM_TRANSFER_WINDOWS *
-		AMDGPU_GPU_PAGE_SIZE;
-	n /= size;
-
-	gtt_obj = kcalloc(n, sizeof(*gtt_obj), GFP_KERNEL);
-	if (!gtt_obj) {
-		DRM_ERROR("Failed to allocate %d pointers\n", n);
-		r = 1;
-		goto out_cleanup;
-	}
-	memset(&bp, 0, sizeof(bp));
-	bp.size = size;
-	bp.byte_align = PAGE_SIZE;
-	bp.domain = AMDGPU_GEM_DOMAIN_VRAM;
-	bp.flags = 0;
-	bp.type = ttm_bo_type_kernel;
-	bp.resv = NULL;
-	bp.bo_ptr_size = sizeof(struct amdgpu_bo);
-
-	r = amdgpu_bo_create(adev, &bp, &vram_obj);
-	if (r) {
-		DRM_ERROR("Failed to create VRAM object\n");
-		goto out_cleanup;
-	}
-	r = amdgpu_bo_reserve(vram_obj, false);
-	if (unlikely(r != 0))
-		goto out_unref;
-	r = amdgpu_bo_pin(vram_obj, AMDGPU_GEM_DOMAIN_VRAM);
-	if (r) {
-		DRM_ERROR("Failed to pin VRAM object\n");
-		goto out_unres;
-	}
-	vram_addr = amdgpu_bo_gpu_offset(vram_obj);
-	for (i = 0; i < n; i++) {
-		void *gtt_map, *vram_map;
-		void **gart_start, **gart_end;
-		void **vram_start, **vram_end;
-		struct dma_fence *fence = NULL;
-
-		bp.domain = AMDGPU_GEM_DOMAIN_GTT;
-		r = amdgpu_bo_create(adev, &bp, gtt_obj + i);
-		if (r) {
-			DRM_ERROR("Failed to create GTT object %d\n", i);
-			goto out_lclean;
-		}
-
-		r = amdgpu_bo_reserve(gtt_obj[i], false);
-		if (unlikely(r != 0))
-			goto out_lclean_unref;
-		r = amdgpu_bo_pin(gtt_obj[i], AMDGPU_GEM_DOMAIN_GTT);
-		if (r) {
-			DRM_ERROR("Failed to pin GTT object %d\n", i);
-			goto out_lclean_unres;
-		}
-		r = amdgpu_ttm_alloc_gart(&gtt_obj[i]->tbo);
-		if (r) {
-			DRM_ERROR("%p bind failed\n", gtt_obj[i]);
-			goto out_lclean_unpin;
-		}
-		gart_addr = amdgpu_bo_gpu_offset(gtt_obj[i]);
-
-		r = amdgpu_bo_kmap(gtt_obj[i], &gtt_map);
-		if (r) {
-			DRM_ERROR("Failed to map GTT object %d\n", i);
-			goto out_lclean_unpin;
-		}
-
-		for (gart_start = gtt_map, gart_end = gtt_map + size;
-		     gart_start < gart_end;
-		     gart_start++)
-			*gart_start = gart_start;
-
-		amdgpu_bo_kunmap(gtt_obj[i]);
-
-		r = amdgpu_copy_buffer(ring, gart_addr, vram_addr,
-				       size, NULL, &fence, false, false, false);
-
-		if (r) {
-			DRM_ERROR("Failed GTT->VRAM copy %d\n", i);
-			goto out_lclean_unpin;
-		}
-
-		r = dma_fence_wait(fence, false);
-		if (r) {
-			DRM_ERROR("Failed to wait for GTT->VRAM fence %d\n", i);
-			goto out_lclean_unpin;
-		}
-
-		dma_fence_put(fence);
-		fence = NULL;
-
-		r = amdgpu_bo_kmap(vram_obj, &vram_map);
-		if (r) {
-			DRM_ERROR("Failed to map VRAM object after copy %d\n", i);
-			goto out_lclean_unpin;
-		}
-
-		for (gart_start = gtt_map, gart_end = gtt_map + size,
-		     vram_start = vram_map, vram_end = vram_map + size;
-		     vram_start < vram_end;
-		     gart_start++, vram_start++) {
-			if (*vram_start != gart_start) {
-				DRM_ERROR("Incorrect GTT->VRAM copy %d: Got 0x%p, "
-					  "expected 0x%p (GTT/VRAM offset "
-					  "0x%16llx/0x%16llx)\n",
-					  i, *vram_start, gart_start,
-					  (unsigned long long)
-					  (gart_addr - adev->gmc.gart_start +
-					   (void *)gart_start - gtt_map),
-					  (unsigned long long)
-					  (vram_addr - adev->gmc.vram_start +
-					   (void *)gart_start - gtt_map));
-				amdgpu_bo_kunmap(vram_obj);
-				goto out_lclean_unpin;
-			}
-			*vram_start = vram_start;
-		}
-
-		amdgpu_bo_kunmap(vram_obj);
-
-		r = amdgpu_copy_buffer(ring, vram_addr, gart_addr,
-				       size, NULL, &fence, false, false, false);
-
-		if (r) {
-			DRM_ERROR("Failed VRAM->GTT copy %d\n", i);
-			goto out_lclean_unpin;
-		}
-
-		r = dma_fence_wait(fence, false);
-		if (r) {
-			DRM_ERROR("Failed to wait for VRAM->GTT fence %d\n", i);
-			goto out_lclean_unpin;
-		}
-
-		dma_fence_put(fence);
-		fence = NULL;
-
-		r = amdgpu_bo_kmap(gtt_obj[i], &gtt_map);
-		if (r) {
-			DRM_ERROR("Failed to map GTT object after copy %d\n", i);
-			goto out_lclean_unpin;
-		}
-
-		for (gart_start = gtt_map, gart_end = gtt_map + size,
-		     vram_start = vram_map, vram_end = vram_map + size;
-		     gart_start < gart_end;
-		     gart_start++, vram_start++) {
-			if (*gart_start != vram_start) {
-				DRM_ERROR("Incorrect VRAM->GTT copy %d: Got 0x%p, "
-					  "expected 0x%p (VRAM/GTT offset "
-					  "0x%16llx/0x%16llx)\n",
-					  i, *gart_start, vram_start,
-					  (unsigned long long)
-					  (vram_addr - adev->gmc.vram_start +
-					   (void *)vram_start - vram_map),
-					  (unsigned long long)
-					  (gart_addr - adev->gmc.gart_start +
-					   (void *)vram_start - vram_map));
-				amdgpu_bo_kunmap(gtt_obj[i]);
-				goto out_lclean_unpin;
-			}
-		}
-
-		amdgpu_bo_kunmap(gtt_obj[i]);
-
-		DRM_INFO("Tested GTT->VRAM and VRAM->GTT copy for GTT offset 0x%llx\n",
-			 gart_addr - adev->gmc.gart_start);
-		continue;
-
-out_lclean_unpin:
-		amdgpu_bo_unpin(gtt_obj[i]);
-out_lclean_unres:
-		amdgpu_bo_unreserve(gtt_obj[i]);
-out_lclean_unref:
-		amdgpu_bo_unref(&gtt_obj[i]);
-out_lclean:
-		for (--i; i >= 0; --i) {
-			amdgpu_bo_unpin(gtt_obj[i]);
-			amdgpu_bo_unreserve(gtt_obj[i]);
-			amdgpu_bo_unref(&gtt_obj[i]);
-		}
-		if (fence)
-			dma_fence_put(fence);
-		break;
-	}
-
-	amdgpu_bo_unpin(vram_obj);
-out_unres:
-	amdgpu_bo_unreserve(vram_obj);
-out_unref:
-	amdgpu_bo_unref(&vram_obj);
-out_cleanup:
-	kfree(gtt_obj);
-	if (r) {
-		pr_warn("Error while testing BO move\n");
-	}
-}
-
-void amdgpu_test_moves(struct amdgpu_device *adev)
-{
-	if (adev->mman.buffer_funcs)
-		amdgpu_do_test_moves(adev);
-}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h
index d855cb53c7e0..d13e64a69e25 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h
@@ -127,7 +127,7 @@ TRACE_EVENT(amdgpu_bo_create,
 
 	    TP_fast_assign(
 			   __entry->bo = bo;
-			   __entry->pages = bo->tbo.resource->num_pages;
+			   __entry->pages = PFN_UP(bo->tbo.resource->size);
 			   __entry->type = bo->tbo.resource->mem_type;
 			   __entry->prefer = bo->preferred_domains;
 			   __entry->allow = bo->allowed_domains;
@@ -140,8 +140,10 @@ TRACE_EVENT(amdgpu_bo_create,
 );
 
 TRACE_EVENT(amdgpu_cs,
-	    TP_PROTO(struct amdgpu_cs_parser *p, int i),
-	    TP_ARGS(p, i),
+	    TP_PROTO(struct amdgpu_cs_parser *p,
+		     struct amdgpu_job *job,
+		     struct amdgpu_ib *ib),
+	    TP_ARGS(p, job, ib),
 	    TP_STRUCT__entry(
 			     __field(struct amdgpu_bo_list *, bo_list)
 			     __field(u32, ring)
@@ -151,10 +153,10 @@ TRACE_EVENT(amdgpu_cs,
 
 	    TP_fast_assign(
 			   __entry->bo_list = p->bo_list;
-			   __entry->ring = to_amdgpu_ring(p->entity->rq->sched)->idx;
-			   __entry->dw = p->job->ibs[i].length_dw;
+			   __entry->ring = to_amdgpu_ring(job->base.entity->rq->sched)->idx;
+			   __entry->dw = ib->length_dw;
 			   __entry->fences = amdgpu_fence_count_emitted(
-				to_amdgpu_ring(p->entity->rq->sched));
+				to_amdgpu_ring(job->base.entity->rq->sched));
 			   ),
 	    TP_printk("bo_list=%p, ring=%u, dw=%u, fences=%u",
 		      __entry->bo_list, __entry->ring, __entry->dw,
@@ -165,25 +167,23 @@ TRACE_EVENT(amdgpu_cs_ioctl,
 	    TP_PROTO(struct amdgpu_job *job),
 	    TP_ARGS(job),
 	    TP_STRUCT__entry(
-			     __field(uint64_t, sched_job_id)
 			     __string(timeline, AMDGPU_JOB_GET_TIMELINE_NAME(job))
-			     __field(unsigned int, context)
-			     __field(unsigned int, seqno)
+			     __field(u64, context)
+			     __field(u64, seqno)
 			     __field(struct dma_fence *, fence)
 			     __string(ring, to_amdgpu_ring(job->base.sched)->name)
 			     __field(u32, num_ibs)
 			     ),
 
 	    TP_fast_assign(
-			   __entry->sched_job_id = job->base.id;
-			   __assign_str(timeline, AMDGPU_JOB_GET_TIMELINE_NAME(job));
+			   __assign_str(timeline);
 			   __entry->context = job->base.s_fence->finished.context;
 			   __entry->seqno = job->base.s_fence->finished.seqno;
-			   __assign_str(ring, to_amdgpu_ring(job->base.sched)->name);
+			   __assign_str(ring);
 			   __entry->num_ibs = job->num_ibs;
 			   ),
-	    TP_printk("sched_job=%llu, timeline=%s, context=%u, seqno=%u, ring_name=%s, num_ibs=%u",
-		      __entry->sched_job_id, __get_str(timeline), __entry->context,
+	    TP_printk("timeline=%s, fence=%llu:%llu, ring_name=%s, num_ibs=%u",
+		      __get_str(timeline), __entry->context,
 		      __entry->seqno, __get_str(ring), __entry->num_ibs)
 );
 
@@ -191,24 +191,22 @@ TRACE_EVENT(amdgpu_sched_run_job,
 	    TP_PROTO(struct amdgpu_job *job),
 	    TP_ARGS(job),
 	    TP_STRUCT__entry(
-			     __field(uint64_t, sched_job_id)
 			     __string(timeline, AMDGPU_JOB_GET_TIMELINE_NAME(job))
-			     __field(unsigned int, context)
-			     __field(unsigned int, seqno)
+			     __field(u64, context)
+			     __field(u64, seqno)
 			     __string(ring, to_amdgpu_ring(job->base.sched)->name)
 			     __field(u32, num_ibs)
 			     ),
 
 	    TP_fast_assign(
-			   __entry->sched_job_id = job->base.id;
-			   __assign_str(timeline, AMDGPU_JOB_GET_TIMELINE_NAME(job));
+			   __assign_str(timeline);
 			   __entry->context = job->base.s_fence->finished.context;
 			   __entry->seqno = job->base.s_fence->finished.seqno;
-			   __assign_str(ring, to_amdgpu_ring(job->base.sched)->name);
+			   __assign_str(ring);
 			   __entry->num_ibs = job->num_ibs;
 			   ),
-	    TP_printk("sched_job=%llu, timeline=%s, context=%u, seqno=%u, ring_name=%s, num_ibs=%u",
-		      __entry->sched_job_id, __get_str(timeline), __entry->context,
+	    TP_printk("timeline=%s, fence=%llu:%llu, ring_name=%s, num_ibs=%u",
+		      __get_str(timeline), __entry->context,
 		      __entry->seqno, __get_str(ring), __entry->num_ibs)
 );
 
@@ -229,9 +227,9 @@ TRACE_EVENT(amdgpu_vm_grab_id,
 
 	    TP_fast_assign(
 			   __entry->pasid = vm->pasid;
-			   __assign_str(ring, ring->name);
+			   __assign_str(ring);
 			   __entry->vmid = job->vmid;
-			   __entry->vm_hub = ring->funcs->vmhub,
+			   __entry->vm_hub = ring->vm_hub,
 			   __entry->pd_addr = job->vm_pd_addr;
 			   __entry->needs_flush = job->vm_needs_flush;
 			   ),
@@ -358,11 +356,10 @@ TRACE_EVENT(amdgpu_vm_update_ptes,
 			}
 	),
 	TP_printk("pid:%u vm_ctx:0x%llx start:0x%010llx end:0x%010llx,"
-		  " flags:0x%llx, incr:%llu, dst:\n%s%s", __entry->pid,
+		  " flags:0x%llx, incr:%llu, dst:\n%s", __entry->pid,
 		  __entry->vm_ctx, __entry->start, __entry->end,
 		  __entry->flags, __entry->incr,  __print_array(
-		  __get_dynamic_array(dst), min(__entry->nptes, 32u), 8),
-		  __entry->nptes > 32 ? "..." : "")
+		  __get_dynamic_array(dst), __entry->nptes, 8))
 );
 
 TRACE_EVENT(amdgpu_vm_set_ptes,
@@ -424,14 +421,14 @@ TRACE_EVENT(amdgpu_vm_flush,
 			     ),
 
 	    TP_fast_assign(
-			   __assign_str(ring, ring->name);
+			   __assign_str(ring);
 			   __entry->vmid = vmid;
-			   __entry->vm_hub = ring->funcs->vmhub;
+			   __entry->vm_hub = ring->vm_hub;
 			   __entry->pd_addr = pd_addr;
 			   ),
 	    TP_printk("ring=%s, id=%u, hub=%u, pd_addr=%010Lx",
 		      __get_str(ring), __entry->vmid,
-		      __entry->vm_hub,__entry->pd_addr)
+		      __entry->vm_hub, __entry->pd_addr)
 );
 
 DECLARE_EVENT_CLASS(amdgpu_pasid,
@@ -456,6 +453,38 @@ DEFINE_EVENT(amdgpu_pasid, amdgpu_pasid_freed,
 	    TP_ARGS(pasid)
 );
 
+TRACE_EVENT(amdgpu_isolation,
+	    TP_PROTO(void *prev, void *next),
+	    TP_ARGS(prev, next),
+	    TP_STRUCT__entry(
+			     __field(void *, prev)
+			     __field(void *, next)
+			     ),
+
+	    TP_fast_assign(
+			   __entry->prev = prev;
+			   __entry->next = next;
+			   ),
+	    TP_printk("prev=%p, next=%p",
+		      __entry->prev,
+		      __entry->next)
+);
+
+TRACE_EVENT(amdgpu_cleaner_shader,
+	    TP_PROTO(struct amdgpu_ring *ring, struct dma_fence *fence),
+	    TP_ARGS(ring, fence),
+	    TP_STRUCT__entry(
+			     __string(ring, ring->name)
+			     __field(u64, seqno)
+			     ),
+
+	    TP_fast_assign(
+			   __assign_str(ring);
+			   __entry->seqno = fence->seqno;
+			   ),
+	    TP_printk("ring=%s, seqno=%Lu", __get_str(ring), __entry->seqno)
+);
+
 TRACE_EVENT(amdgpu_bo_list_set,
 	    TP_PROTO(struct amdgpu_bo_list *list, struct amdgpu_bo *bo),
 	    TP_ARGS(list, bo),
@@ -493,7 +522,7 @@ TRACE_EVENT(amdgpu_cs_bo_status,
 );
 
 TRACE_EVENT(amdgpu_bo_move,
-	    TP_PROTO(struct amdgpu_bo* bo, uint32_t new_placement, uint32_t old_placement),
+	    TP_PROTO(struct amdgpu_bo *bo, uint32_t new_placement, uint32_t old_placement),
 	    TP_ARGS(bo, new_placement, old_placement),
 	    TP_STRUCT__entry(
 			__field(struct amdgpu_bo *, bo)
@@ -518,23 +547,35 @@ TRACE_EVENT(amdgpu_ib_pipe_sync,
 	    TP_ARGS(sched_job, fence),
 	    TP_STRUCT__entry(
 			     __string(ring, sched_job->base.sched->name)
-			     __field(uint64_t, id)
 			     __field(struct dma_fence *, fence)
-			     __field(uint64_t, ctx)
-			     __field(unsigned, seqno)
+			     __field(u64, ctx)
+			     __field(u64, seqno)
 			     ),
 
 	    TP_fast_assign(
-			   __assign_str(ring, sched_job->base.sched->name);
-			   __entry->id = sched_job->base.id;
+			   __assign_str(ring);
 			   __entry->fence = fence;
 			   __entry->ctx = fence->context;
 			   __entry->seqno = fence->seqno;
 			   ),
-	    TP_printk("job ring=%s, id=%llu, need pipe sync to fence=%p, context=%llu, seq=%u",
-		      __get_str(ring), __entry->id,
-		      __entry->fence, __entry->ctx,
-		      __entry->seqno)
+	    TP_printk("job ring=%s need pipe sync to fence=%llu:%llu",
+		      __get_str(ring), __entry->ctx, __entry->seqno)
+);
+
+TRACE_EVENT(amdgpu_reset_reg_dumps,
+	    TP_PROTO(uint32_t address, uint32_t value),
+	    TP_ARGS(address, value),
+	    TP_STRUCT__entry(
+			     __field(uint32_t, address)
+			     __field(uint32_t, value)
+			     ),
+	    TP_fast_assign(
+			   __entry->address = address;
+			   __entry->value = value;
+			   ),
+	    TP_printk("amdgpu register dump 0x%x: 0x%x",
+		      __entry->address,
+		      __entry->value)
 );
 
 #undef AMDGPU_JOB_GET_TIMELINE_NAME
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_trace_points.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_trace_points.c
index 57c6c39ba064..b96d885f6e33 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_trace_points.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_trace_points.c
@@ -23,6 +23,7 @@
  */
 
 #include <drm/amdgpu_drm.h>
+#include "amdgpu_cs.h"
 #include "amdgpu.h"
 
 #define CREATE_TRACE_POINTS
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index c875f1cdd2af..2b931e855abd 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -38,15 +38,15 @@
 #include <linux/seq_file.h>
 #include <linux/slab.h>
 #include <linux/swap.h>
-#include <linux/swiotlb.h>
 #include <linux/dma-buf.h>
 #include <linux/sizes.h>
 #include <linux/module.h>
 
-#include <drm/ttm/ttm_bo_api.h>
-#include <drm/ttm/ttm_bo_driver.h>
+#include <drm/drm_drv.h>
+#include <drm/ttm/ttm_bo.h>
 #include <drm/ttm/ttm_placement.h>
 #include <drm/ttm/ttm_range_manager.h>
+#include <drm/ttm/ttm_tt.h>
 
 #include <drm/amdgpu_drm.h>
 
@@ -56,13 +56,14 @@
 #include "amdgpu_amdkfd.h"
 #include "amdgpu_sdma.h"
 #include "amdgpu_ras.h"
+#include "amdgpu_hmm.h"
 #include "amdgpu_atomfirmware.h"
 #include "amdgpu_res_cursor.h"
 #include "bif/bif_4_1_d.h"
 
-MODULE_IMPORT_NS(DMA_BUF);
+MODULE_IMPORT_NS("DMA_BUF");
 
-#define AMDGPU_TTM_VRAM_MAX_DW_READ	(size_t)128
+#define AMDGPU_TTM_VRAM_MAX_DW_READ	((size_t)128)
 
 static int amdgpu_ttm_backend_bind(struct ttm_device *bdev,
 				   struct ttm_tt *ttm,
@@ -101,32 +102,19 @@ static void amdgpu_evict_flags(struct ttm_buffer_object *bo,
 	/* Don't handle scatter gather BOs */
 	if (bo->type == ttm_bo_type_sg) {
 		placement->num_placement = 0;
-		placement->num_busy_placement = 0;
 		return;
 	}
 
 	/* Object isn't an AMDGPU object so ignore */
 	if (!amdgpu_bo_is_amdgpu_bo(bo)) {
 		placement->placement = &placements;
-		placement->busy_placement = &placements;
 		placement->num_placement = 1;
-		placement->num_busy_placement = 1;
 		return;
 	}
 
 	abo = ttm_to_amdgpu_bo(bo);
-	if (abo->flags & AMDGPU_AMDKFD_CREATE_SVM_BO) {
-		struct dma_fence *fence;
-		struct dma_resv *resv = &bo->base._resv;
-
-		rcu_read_lock();
-		fence = rcu_dereference(resv->fence_excl);
-		if (fence && !fence->ops->signaled)
-			dma_fence_enable_sw_signaling(fence);
-
+	if (abo->flags & AMDGPU_GEM_CREATE_DISCARDABLE) {
 		placement->num_placement = 0;
-		placement->num_busy_placement = 0;
-		rcu_read_unlock();
 		return;
 	}
 
@@ -134,17 +122,19 @@ static void amdgpu_evict_flags(struct ttm_buffer_object *bo,
 	case AMDGPU_PL_GDS:
 	case AMDGPU_PL_GWS:
 	case AMDGPU_PL_OA:
+	case AMDGPU_PL_DOORBELL:
+	case AMDGPU_PL_MMIO_REMAP:
 		placement->num_placement = 0;
-		placement->num_busy_placement = 0;
 		return;
 
 	case TTM_PL_VRAM:
 		if (!adev->mman.buffer_funcs_enabled) {
 			/* Move to system memory */
 			amdgpu_bo_placement_from_domain(abo, AMDGPU_GEM_DOMAIN_CPU);
+
 		} else if (!amdgpu_gmc_vram_full_visible(&adev->gmc) &&
 			   !(abo->flags & AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED) &&
-			   amdgpu_bo_in_cpu_visible_vram(abo)) {
+			   amdgpu_res_cpu_visible(adev, bo->resource)) {
 
 			/* Try evicting to the CPU inaccessible part of VRAM
 			 * first, but only set GTT as busy placement, so this
@@ -156,8 +146,7 @@ static void amdgpu_evict_flags(struct ttm_buffer_object *bo,
 							AMDGPU_GEM_DOMAIN_CPU);
 			abo->placements[0].fpfn = adev->gmc.visible_vram_size >> PAGE_SHIFT;
 			abo->placements[0].lpfn = 0;
-			abo->placement.busy_placement = &abo->placements[1];
-			abo->placement.num_busy_placement = 1;
+			abo->placements[0].flags |= TTM_PL_FLAG_DESIRED;
 		} else {
 			/* Move to GTT memory */
 			amdgpu_bo_placement_from_domain(abo, AMDGPU_GEM_DOMAIN_GTT |
@@ -178,10 +167,10 @@ static void amdgpu_evict_flags(struct ttm_buffer_object *bo,
  * @bo: buffer object to map
  * @mem: memory object to map
  * @mm_cur: range to map
- * @num_pages: number of pages to map
  * @window: which GART window to use
  * @ring: DMA ring to use for the copy
  * @tmz: if we should setup a TMZ enabled mapping
+ * @size: in number of bytes to map, out number of bytes mapped
  * @addr: resulting address inside the MC address space
  *
  * Setup one of the GART windows to access a specific piece of memory or return
@@ -190,23 +179,22 @@ static void amdgpu_evict_flags(struct ttm_buffer_object *bo,
 static int amdgpu_ttm_map_buffer(struct ttm_buffer_object *bo,
 				 struct ttm_resource *mem,
 				 struct amdgpu_res_cursor *mm_cur,
-				 unsigned num_pages, unsigned window,
-				 struct amdgpu_ring *ring, bool tmz,
-				 uint64_t *addr)
+				 unsigned int window, struct amdgpu_ring *ring,
+				 bool tmz, uint64_t *size, uint64_t *addr)
 {
 	struct amdgpu_device *adev = ring->adev;
-	struct amdgpu_job *job;
-	unsigned num_dw, num_bytes;
-	struct dma_fence *fence;
+	unsigned int offset, num_pages, num_dw, num_bytes;
 	uint64_t src_addr, dst_addr;
+	struct amdgpu_job *job;
 	void *cpu_addr;
 	uint64_t flags;
-	unsigned int i;
 	int r;
 
 	BUG_ON(adev->mman.buffer_funcs->copy_max_bytes <
 	       AMDGPU_GTT_MAX_TRANSFER_SIZE * 8);
-	BUG_ON(mem->mem_type == AMDGPU_PL_PREEMPT);
+
+	if (WARN_ON(mem->mem_type == AMDGPU_PL_PREEMPT))
+		return -EINVAL;
 
 	/* Map only what can't be accessed directly */
 	if (!tmz && mem->start != AMDGPU_BO_INVALID_OFFSET) {
@@ -215,16 +203,31 @@ static int amdgpu_ttm_map_buffer(struct ttm_buffer_object *bo,
 		return 0;
 	}
 
+
+	/*
+	 * If start begins at an offset inside the page, then adjust the size
+	 * and addr accordingly
+	 */
+	offset = mm_cur->start & ~PAGE_MASK;
+
+	num_pages = PFN_UP(*size + offset);
+	num_pages = min_t(uint32_t, num_pages, AMDGPU_GTT_MAX_TRANSFER_SIZE);
+
+	*size = min(*size, (uint64_t)num_pages * PAGE_SIZE - offset);
+
 	*addr = adev->gmc.gart_start;
 	*addr += (u64)window * AMDGPU_GTT_MAX_TRANSFER_SIZE *
 		AMDGPU_GPU_PAGE_SIZE;
-	*addr += mm_cur->start & ~PAGE_MASK;
+	*addr += offset;
 
 	num_dw = ALIGN(adev->mman.buffer_funcs->copy_num_dw, 8);
 	num_bytes = num_pages * 8 * AMDGPU_GPU_PAGES_IN_CPU_PAGE;
 
-	r = amdgpu_job_alloc_with_ib(adev, num_dw * 4 + num_bytes,
-				     AMDGPU_IB_POOL_DELAYED, &job);
+	r = amdgpu_job_alloc_with_ib(adev, &adev->mman.high_pr,
+				     AMDGPU_FENCE_OWNER_UNDEFINED,
+				     num_dw * 4 + num_bytes,
+				     AMDGPU_IB_POOL_DELAYED, &job,
+				     AMDGPU_KERNEL_JOB_ID_TTM_MAP_BUFFER);
 	if (r)
 		return r;
 
@@ -234,7 +237,7 @@ static int amdgpu_ttm_map_buffer(struct ttm_buffer_object *bo,
 	dst_addr = amdgpu_bo_gpu_offset(adev->gart.bo);
 	dst_addr += window * AMDGPU_GTT_MAX_TRANSFER_SIZE * 8;
 	amdgpu_emit_copy_buffer(adev, &job->ibs[0], src_addr,
-				dst_addr, num_bytes, false);
+				dst_addr, num_bytes, 0);
 
 	amdgpu_ring_pad_ib(ring, &job->ibs[0]);
 	WARN_ON(job->ibs[0].length_dw > num_dw);
@@ -249,38 +252,15 @@ static int amdgpu_ttm_map_buffer(struct ttm_buffer_object *bo,
 		dma_addr_t *dma_addr;
 
 		dma_addr = &bo->ttm->dma_address[mm_cur->start >> PAGE_SHIFT];
-		r = amdgpu_gart_map(adev, 0, num_pages, dma_addr, flags,
-				    cpu_addr);
-		if (r)
-			goto error_free;
+		amdgpu_gart_map(adev, 0, num_pages, dma_addr, flags, cpu_addr);
 	} else {
-		dma_addr_t dma_address;
+		u64 pa = mm_cur->start + adev->vm_manager.vram_base_offset;
 
-		dma_address = mm_cur->start;
-		dma_address += adev->vm_manager.vram_base_offset;
-
-		for (i = 0; i < num_pages; ++i) {
-			r = amdgpu_gart_map(adev, i << PAGE_SHIFT, 1,
-					    &dma_address, flags, cpu_addr);
-			if (r)
-				goto error_free;
-
-			dma_address += PAGE_SIZE;
-		}
+		amdgpu_gart_map_vram_range(adev, pa, 0, num_pages, flags, cpu_addr);
 	}
 
-	r = amdgpu_job_submit(job, &adev->mman.entity,
-			      AMDGPU_FENCE_OWNER_UNDEFINED, &fence);
-	if (r)
-		goto error_free;
-
-	dma_fence_put(fence);
-
-	return r;
-
-error_free:
-	amdgpu_job_free(job);
-	return r;
+	dma_fence_put(amdgpu_job_submit(job));
+	return 0;
 }
 
 /**
@@ -298,23 +278,24 @@ error_free:
  * move and different for a BO to BO copy.
  *
  */
-int amdgpu_ttm_copy_mem_to_mem(struct amdgpu_device *adev,
-			       const struct amdgpu_copy_mem *src,
-			       const struct amdgpu_copy_mem *dst,
-			       uint64_t size, bool tmz,
-			       struct dma_resv *resv,
-			       struct dma_fence **f)
+__attribute__((nonnull))
+static int amdgpu_ttm_copy_mem_to_mem(struct amdgpu_device *adev,
+				      const struct amdgpu_copy_mem *src,
+				      const struct amdgpu_copy_mem *dst,
+				      uint64_t size, bool tmz,
+				      struct dma_resv *resv,
+				      struct dma_fence **f)
 {
-	const uint32_t GTT_MAX_BYTES = (AMDGPU_GTT_MAX_TRANSFER_SIZE *
-					AMDGPU_GPU_PAGE_SIZE);
-
 	struct amdgpu_ring *ring = adev->mman.buffer_funcs_ring;
 	struct amdgpu_res_cursor src_mm, dst_mm;
 	struct dma_fence *fence = NULL;
 	int r = 0;
+	uint32_t copy_flags = 0;
+	struct amdgpu_bo *abo_src, *abo_dst;
 
 	if (!adev->mman.buffer_funcs_enabled) {
-		DRM_ERROR("Trying to move memory with ring turned off.\n");
+		dev_err(adev->dev,
+			"Trying to move memory with ring turned off.\n");
 		return -EINVAL;
 	}
 
@@ -323,34 +304,49 @@ int amdgpu_ttm_copy_mem_to_mem(struct amdgpu_device *adev,
 
 	mutex_lock(&adev->mman.gtt_window_lock);
 	while (src_mm.remaining) {
-		uint32_t src_page_offset = src_mm.start & ~PAGE_MASK;
-		uint32_t dst_page_offset = dst_mm.start & ~PAGE_MASK;
+		uint64_t from, to, cur_size, tiling_flags;
+		uint32_t num_type, data_format, max_com, write_compress_disable;
 		struct dma_fence *next;
-		uint32_t cur_size;
-		uint64_t from, to;
 
-		/* Copy size cannot exceed GTT_MAX_BYTES. So if src or dst
-		 * begins at an offset, then adjust the size accordingly
-		 */
-		cur_size = max(src_page_offset, dst_page_offset);
-		cur_size = min(min3(src_mm.size, dst_mm.size, size),
-			       (uint64_t)(GTT_MAX_BYTES - cur_size));
+		/* Never copy more than 256MiB at once to avoid a timeout */
+		cur_size = min3(src_mm.size, dst_mm.size, 256ULL << 20);
 
 		/* Map src to window 0 and dst to window 1. */
 		r = amdgpu_ttm_map_buffer(src->bo, src->mem, &src_mm,
-					  PFN_UP(cur_size + src_page_offset),
-					  0, ring, tmz, &from);
+					  0, ring, tmz, &cur_size, &from);
 		if (r)
 			goto error;
 
 		r = amdgpu_ttm_map_buffer(dst->bo, dst->mem, &dst_mm,
-					  PFN_UP(cur_size + dst_page_offset),
-					  1, ring, tmz, &to);
+					  1, ring, tmz, &cur_size, &to);
 		if (r)
 			goto error;
 
-		r = amdgpu_copy_buffer(ring, from, to, cur_size,
-				       resv, &next, false, true, tmz);
+		abo_src = ttm_to_amdgpu_bo(src->bo);
+		abo_dst = ttm_to_amdgpu_bo(dst->bo);
+		if (tmz)
+			copy_flags |= AMDGPU_COPY_FLAGS_TMZ;
+		if ((abo_src->flags & AMDGPU_GEM_CREATE_GFX12_DCC) &&
+		    (abo_src->tbo.resource->mem_type == TTM_PL_VRAM))
+			copy_flags |= AMDGPU_COPY_FLAGS_READ_DECOMPRESSED;
+		if ((abo_dst->flags & AMDGPU_GEM_CREATE_GFX12_DCC) &&
+		    (dst->mem->mem_type == TTM_PL_VRAM)) {
+			copy_flags |= AMDGPU_COPY_FLAGS_WRITE_COMPRESSED;
+			amdgpu_bo_get_tiling_flags(abo_dst, &tiling_flags);
+			max_com = AMDGPU_TILING_GET(tiling_flags, GFX12_DCC_MAX_COMPRESSED_BLOCK);
+			num_type = AMDGPU_TILING_GET(tiling_flags, GFX12_DCC_NUMBER_TYPE);
+			data_format = AMDGPU_TILING_GET(tiling_flags, GFX12_DCC_DATA_FORMAT);
+			write_compress_disable =
+				AMDGPU_TILING_GET(tiling_flags, GFX12_DCC_WRITE_COMPRESS_DISABLE);
+			copy_flags |= (AMDGPU_COPY_FLAGS_SET(MAX_COMPRESSED, max_com) |
+				       AMDGPU_COPY_FLAGS_SET(NUMBER_TYPE, num_type) |
+				       AMDGPU_COPY_FLAGS_SET(DATA_FORMAT, data_format) |
+				       AMDGPU_COPY_FLAGS_SET(WRITE_COMPRESS_DISABLE,
+							     write_compress_disable));
+		}
+
+		r = amdgpu_copy_buffer(ring, from, to, cur_size, resv,
+				       &next, false, true, copy_flags);
 		if (r)
 			goto error;
 
@@ -362,9 +358,7 @@ int amdgpu_ttm_copy_mem_to_mem(struct amdgpu_device *adev,
 	}
 error:
 	mutex_unlock(&adev->mman.gtt_window_lock);
-	if (f)
-		*f = dma_fence_get(fence);
-	dma_fence_put(fence);
+	*f = fence;
 	return r;
 }
 
@@ -393,7 +387,7 @@ static int amdgpu_move_blit(struct ttm_buffer_object *bo,
 	dst.offset = 0;
 
 	r = amdgpu_ttm_copy_mem_to_mem(adev, &src, &dst,
-				       new_mem->num_pages << PAGE_SHIFT,
+				       new_mem->size,
 				       amdgpu_bo_encrypted(abo),
 				       bo->base.resv, &fence);
 	if (r)
@@ -404,11 +398,12 @@ static int amdgpu_move_blit(struct ttm_buffer_object *bo,
 	    (abo->flags & AMDGPU_GEM_CREATE_VRAM_WIPE_ON_RELEASE)) {
 		struct dma_fence *wipe_fence = NULL;
 
-		r = amdgpu_fill_buffer(ttm_to_amdgpu_bo(bo), AMDGPU_POISON,
-				       NULL, &wipe_fence);
+		r = amdgpu_fill_buffer(abo, 0, NULL, &wipe_fence,
+				       false, AMDGPU_KERNEL_JOB_ID_MOVE_BLIT);
 		if (r) {
 			goto error;
 		} else if (wipe_fence) {
+			amdgpu_vram_mgr_set_cleared(bo->resource);
 			dma_fence_put(fence);
 			fence = wipe_fence;
 		}
@@ -429,30 +424,56 @@ error:
 	return r;
 }
 
-/*
- * amdgpu_mem_visible - Check that memory can be accessed by ttm_bo_move_memcpy
+/**
+ * amdgpu_res_cpu_visible - Check that resource can be accessed by CPU
+ * @adev: amdgpu device
+ * @res: the resource to check
  *
- * Called by amdgpu_bo_move()
+ * Returns: true if the full resource is CPU visible, false otherwise.
  */
-static bool amdgpu_mem_visible(struct amdgpu_device *adev,
-			       struct ttm_resource *mem)
+bool amdgpu_res_cpu_visible(struct amdgpu_device *adev,
+			    struct ttm_resource *res)
 {
-	uint64_t mem_size = (u64)mem->num_pages << PAGE_SHIFT;
 	struct amdgpu_res_cursor cursor;
 
-	if (mem->mem_type == TTM_PL_SYSTEM ||
-	    mem->mem_type == TTM_PL_TT)
+	if (!res)
+		return false;
+
+	if (res->mem_type == TTM_PL_SYSTEM || res->mem_type == TTM_PL_TT ||
+	    res->mem_type == AMDGPU_PL_PREEMPT || res->mem_type == AMDGPU_PL_DOORBELL ||
+	    res->mem_type == AMDGPU_PL_MMIO_REMAP)
 		return true;
-	if (mem->mem_type != TTM_PL_VRAM)
+
+	if (res->mem_type != TTM_PL_VRAM)
 		return false;
 
-	amdgpu_res_first(mem, 0, mem_size, &cursor);
+	amdgpu_res_first(res, 0, res->size, &cursor);
+	while (cursor.remaining) {
+		if ((cursor.start + cursor.size) > adev->gmc.visible_vram_size)
+			return false;
+		amdgpu_res_next(&cursor, cursor.size);
+	}
+
+	return true;
+}
+
+/*
+ * amdgpu_res_copyable - Check that memory can be accessed by ttm_bo_move_memcpy
+ *
+ * Called by amdgpu_bo_move()
+ */
+static bool amdgpu_res_copyable(struct amdgpu_device *adev,
+				struct ttm_resource *mem)
+{
+	if (!amdgpu_res_cpu_visible(adev, mem))
+		return false;
 
 	/* ttm_resource_ioremap only supports contiguous memory */
-	if (cursor.size != mem_size)
+	if (mem->mem_type == TTM_PL_VRAM &&
+	    !(mem->placement & TTM_PL_FLAG_CONTIGUOUS))
 		return false;
 
-	return cursor.start + cursor.size <= adev->gmc.visible_vram_size;
+	return true;
 }
 
 /*
@@ -477,22 +498,21 @@ static int amdgpu_bo_move(struct ttm_buffer_object *bo, bool evict,
 			return r;
 	}
 
-	/* Can't move a pinned BO */
 	abo = ttm_to_amdgpu_bo(bo);
-	if (WARN_ON_ONCE(abo->tbo.pin_count > 0))
-		return -EINVAL;
-
 	adev = amdgpu_ttm_adev(bo->bdev);
 
-	if (old_mem->mem_type == TTM_PL_SYSTEM && bo->ttm == NULL) {
+	if (!old_mem || (old_mem->mem_type == TTM_PL_SYSTEM &&
+			 bo->ttm == NULL)) {
+		amdgpu_bo_move_notify(bo, evict, new_mem);
 		ttm_bo_move_null(bo, new_mem);
-		goto out;
+		return 0;
 	}
 	if (old_mem->mem_type == TTM_PL_SYSTEM &&
 	    (new_mem->mem_type == TTM_PL_TT ||
 	     new_mem->mem_type == AMDGPU_PL_PREEMPT)) {
+		amdgpu_bo_move_notify(bo, evict, new_mem);
 		ttm_bo_move_null(bo, new_mem);
-		goto out;
+		return 0;
 	}
 	if ((old_mem->mem_type == TTM_PL_TT ||
 	     old_mem->mem_type == AMDGPU_PL_PREEMPT) &&
@@ -502,20 +522,26 @@ static int amdgpu_bo_move(struct ttm_buffer_object *bo, bool evict,
 			return r;
 
 		amdgpu_ttm_backend_unbind(bo->bdev, bo->ttm);
+		amdgpu_bo_move_notify(bo, evict, new_mem);
 		ttm_resource_free(bo, &bo->resource);
 		ttm_bo_assign_mem(bo, new_mem);
-		goto out;
+		return 0;
 	}
 
 	if (old_mem->mem_type == AMDGPU_PL_GDS ||
 	    old_mem->mem_type == AMDGPU_PL_GWS ||
 	    old_mem->mem_type == AMDGPU_PL_OA ||
+	    old_mem->mem_type == AMDGPU_PL_DOORBELL ||
+	    old_mem->mem_type == AMDGPU_PL_MMIO_REMAP ||
 	    new_mem->mem_type == AMDGPU_PL_GDS ||
 	    new_mem->mem_type == AMDGPU_PL_GWS ||
-	    new_mem->mem_type == AMDGPU_PL_OA) {
+	    new_mem->mem_type == AMDGPU_PL_OA ||
+	    new_mem->mem_type == AMDGPU_PL_DOORBELL ||
+	    new_mem->mem_type == AMDGPU_PL_MMIO_REMAP) {
 		/* Nothing to save here */
+		amdgpu_bo_move_notify(bo, evict, new_mem);
 		ttm_bo_move_null(bo, new_mem);
-		goto out;
+		return 0;
 	}
 
 	if (bo->type == ttm_bo_type_device &&
@@ -527,27 +553,28 @@ static int amdgpu_bo_move(struct ttm_buffer_object *bo, bool evict,
 		abo->flags &= ~AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED;
 	}
 
-	if (adev->mman.buffer_funcs_enabled) {
-		if (((old_mem->mem_type == TTM_PL_SYSTEM &&
-		      new_mem->mem_type == TTM_PL_VRAM) ||
-		     (old_mem->mem_type == TTM_PL_VRAM &&
-		      new_mem->mem_type == TTM_PL_SYSTEM))) {
-			hop->fpfn = 0;
-			hop->lpfn = 0;
-			hop->mem_type = TTM_PL_TT;
-			hop->flags = TTM_PL_FLAG_TEMPORARY;
-			return -EMULTIHOP;
-		}
+	if (adev->mman.buffer_funcs_enabled &&
+	    ((old_mem->mem_type == TTM_PL_SYSTEM &&
+	      new_mem->mem_type == TTM_PL_VRAM) ||
+	     (old_mem->mem_type == TTM_PL_VRAM &&
+	      new_mem->mem_type == TTM_PL_SYSTEM))) {
+		hop->fpfn = 0;
+		hop->lpfn = 0;
+		hop->mem_type = TTM_PL_TT;
+		hop->flags = TTM_PL_FLAG_TEMPORARY;
+		return -EMULTIHOP;
+	}
 
+	amdgpu_bo_move_notify(bo, evict, new_mem);
+	if (adev->mman.buffer_funcs_enabled)
 		r = amdgpu_move_blit(bo, evict, new_mem, old_mem);
-	} else {
+	else
 		r = -ENODEV;
-	}
 
 	if (r) {
 		/* Check that all memory is CPU accessible */
-		if (!amdgpu_mem_visible(adev, old_mem) ||
-		    !amdgpu_mem_visible(adev, new_mem)) {
+		if (!amdgpu_res_copyable(adev, old_mem) ||
+		    !amdgpu_res_copyable(adev, new_mem)) {
 			pr_err("Move buffer fallback to memcpy unavailable\n");
 			return r;
 		}
@@ -557,10 +584,10 @@ static int amdgpu_bo_move(struct ttm_buffer_object *bo, bool evict,
 			return r;
 	}
 
-out:
-	/* update statistics */
+	/* update statistics after the move */
+	if (evict)
+		atomic64_inc(&adev->num_evictions);
 	atomic64_add(bo->base.size, &adev->num_bytes_moved);
-	amdgpu_bo_move_notify(bo, evict, new_mem);
 	return 0;
 }
 
@@ -573,7 +600,6 @@ static int amdgpu_ttm_io_mem_reserve(struct ttm_device *bdev,
 				     struct ttm_resource *mem)
 {
 	struct amdgpu_device *adev = amdgpu_ttm_adev(bdev);
-	size_t bus_size = (size_t)mem->num_pages << PAGE_SHIFT;
 
 	switch (mem->mem_type) {
 	case TTM_PL_SYSTEM:
@@ -584,9 +610,6 @@ static int amdgpu_ttm_io_mem_reserve(struct ttm_device *bdev,
 		break;
 	case TTM_PL_VRAM:
 		mem->bus.offset = mem->start << PAGE_SHIFT;
-		/* check if it's visible */
-		if ((mem->bus.offset + bus_size) > adev->gmc.visible_vram_size)
-			return -EINVAL;
 
 		if (adev->mman.aper_base_kaddr &&
 		    mem->placement & TTM_PL_FLAG_CONTIGUOUS)
@@ -596,6 +619,18 @@ static int amdgpu_ttm_io_mem_reserve(struct ttm_device *bdev,
 		mem->bus.offset += adev->gmc.aper_base;
 		mem->bus.is_iomem = true;
 		break;
+	case AMDGPU_PL_DOORBELL:
+		mem->bus.offset = mem->start << PAGE_SHIFT;
+		mem->bus.offset += adev->doorbell.base;
+		mem->bus.is_iomem = true;
+		mem->bus.caching = ttm_uncached;
+		break;
+	case AMDGPU_PL_MMIO_REMAP:
+		mem->bus.offset = mem->start << PAGE_SHIFT;
+		mem->bus.offset += adev->rmmio_remap.bus_addr;
+		mem->bus.is_iomem = true;
+		mem->bus.caching = ttm_uncached;
+		break;
 	default:
 		return -EINVAL;
 	}
@@ -610,6 +645,12 @@ static unsigned long amdgpu_ttm_io_mem_pfn(struct ttm_buffer_object *bo,
 
 	amdgpu_res_first(bo->resource, (u64)page_offset << PAGE_SHIFT, 0,
 			 &cursor);
+
+	if (bo->resource->mem_type == AMDGPU_PL_DOORBELL)
+		return ((uint64_t)(adev->doorbell.base + cursor.start)) >> PAGE_SHIFT;
+	else if (bo->resource->mem_type == AMDGPU_PL_MMIO_REMAP)
+		return ((uint64_t)(adev->rmmio_remap.bus_addr + cursor.start)) >> PAGE_SHIFT;
+
 	return (adev->gmc.aper_base + cursor.start) >> PAGE_SHIFT;
 }
 
@@ -645,23 +686,25 @@ struct amdgpu_ttm_tt {
 	struct task_struct	*usertask;
 	uint32_t		userflags;
 	bool			bound;
-#if IS_ENABLED(CONFIG_DRM_AMDGPU_USERPTR)
-	struct hmm_range	*range;
-#endif
+	int32_t			pool_id;
 };
 
+#define ttm_to_amdgpu_ttm_tt(ptr)	container_of(ptr, struct amdgpu_ttm_tt, ttm)
+
 #ifdef CONFIG_DRM_AMDGPU_USERPTR
 /*
  * amdgpu_ttm_tt_get_user_pages - get device accessible pages that back user
  * memory and start HMM tracking CPU page table update
  *
  * Calling function must call amdgpu_ttm_tt_userptr_range_done() once and only
- * once afterwards to stop HMM tracking
+ * once afterwards to stop HMM tracking. Its the caller responsibility to ensure
+ * that range is a valid memory and it is freed too.
  */
-int amdgpu_ttm_tt_get_user_pages(struct amdgpu_bo *bo, struct page **pages)
+int amdgpu_ttm_tt_get_user_pages(struct amdgpu_bo *bo,
+				 struct amdgpu_hmm_range *range)
 {
 	struct ttm_tt *ttm = bo->tbo.ttm;
-	struct amdgpu_ttm_tt *gtt = (void *)ttm;
+	struct amdgpu_ttm_tt *gtt = ttm_to_amdgpu_ttm_tt(ttm);
 	unsigned long start = gtt->userptr;
 	struct vm_area_struct *vma;
 	struct mm_struct *mm;
@@ -674,10 +717,6 @@ int amdgpu_ttm_tt_get_user_pages(struct amdgpu_bo *bo, struct page **pages)
 		return -EFAULT;
 	}
 
-	/* Another get_user_pages is running at the same time?? */
-	if (WARN_ON(gtt->range))
-		return -EFAULT;
-
 	if (!mmget_not_zero(mm)) /* Happens during process shutdown */
 		return -ESRCH;
 
@@ -694,9 +733,8 @@ int amdgpu_ttm_tt_get_user_pages(struct amdgpu_bo *bo, struct page **pages)
 	}
 
 	readonly = amdgpu_ttm_tt_is_readonly(ttm);
-	r = amdgpu_hmm_range_get_pages(&bo->notifier, mm, pages, start,
-				       ttm->num_pages, &gtt->range, readonly,
-				       true, NULL);
+	r = amdgpu_hmm_range_get_pages(&bo->notifier, start, ttm->num_pages,
+				       readonly, NULL, range);
 out_unlock:
 	mmap_read_unlock(mm);
 	if (r)
@@ -707,37 +745,6 @@ out_unlock:
 	return r;
 }
 
-/*
- * amdgpu_ttm_tt_userptr_range_done - stop HMM track the CPU page table change
- * Check if the pages backing this ttm range have been invalidated
- *
- * Returns: true if pages are still valid
- */
-bool amdgpu_ttm_tt_get_user_pages_done(struct ttm_tt *ttm)
-{
-	struct amdgpu_ttm_tt *gtt = (void *)ttm;
-	bool r = false;
-
-	if (!gtt || !gtt->userptr)
-		return false;
-
-	DRM_DEBUG_DRIVER("user_pages_done 0x%llx pages 0x%x\n",
-		gtt->userptr, ttm->num_pages);
-
-	WARN_ONCE(!gtt->range || !gtt->range->hmm_pfns,
-		"No user pages to check\n");
-
-	if (gtt->range) {
-		/*
-		 * FIXME: Must always hold notifier_lock for this, and must
-		 * not ignore the return code.
-		 */
-		r = amdgpu_hmm_range_get_pages_done(gtt->range);
-		gtt->range = NULL;
-	}
-
-	return !r;
-}
 #endif
 
 /*
@@ -747,12 +754,12 @@ bool amdgpu_ttm_tt_get_user_pages_done(struct ttm_tt *ttm)
  * that backs user memory and will ultimately be mapped into the device
  * address space.
  */
-void amdgpu_ttm_tt_set_user_pages(struct ttm_tt *ttm, struct page **pages)
+void amdgpu_ttm_tt_set_user_pages(struct ttm_tt *ttm, struct amdgpu_hmm_range *range)
 {
 	unsigned long i;
 
 	for (i = 0; i < ttm->num_pages; ++i)
-		ttm->pages[i] = pages ? pages[i] : NULL;
+		ttm->pages[i] = range ? hmm_pfn_to_page(range->hmm_range.hmm_pfns[i]) : NULL;
 }
 
 /*
@@ -764,7 +771,7 @@ static int amdgpu_ttm_tt_pin_userptr(struct ttm_device *bdev,
 				     struct ttm_tt *ttm)
 {
 	struct amdgpu_device *adev = amdgpu_ttm_adev(bdev);
-	struct amdgpu_ttm_tt *gtt = (void *)ttm;
+	struct amdgpu_ttm_tt *gtt = ttm_to_amdgpu_ttm_tt(ttm);
 	int write = !(gtt->userflags & AMDGPU_GEM_USERPTR_READONLY);
 	enum dma_data_direction direction = write ?
 		DMA_BIDIRECTIONAL : DMA_TO_DEVICE;
@@ -780,7 +787,7 @@ static int amdgpu_ttm_tt_pin_userptr(struct ttm_device *bdev,
 	/* Map SG to device */
 	r = dma_map_sgtable(adev->dev, ttm->sg, direction, 0);
 	if (r)
-		goto release_sg;
+		goto release_sg_table;
 
 	/* convert SG to linear array of pages and dma addresses */
 	drm_prime_sg_to_dma_addr_array(ttm->sg, gtt->ttm.dma_address,
@@ -788,6 +795,8 @@ static int amdgpu_ttm_tt_pin_userptr(struct ttm_device *bdev,
 
 	return 0;
 
+release_sg_table:
+	sg_free_table(ttm->sg);
 release_sg:
 	kfree(ttm->sg);
 	ttm->sg = NULL;
@@ -801,7 +810,7 @@ static void amdgpu_ttm_tt_unpin_userptr(struct ttm_device *bdev,
 					struct ttm_tt *ttm)
 {
 	struct amdgpu_device *adev = amdgpu_ttm_adev(bdev);
-	struct amdgpu_ttm_tt *gtt = (void *)ttm;
+	struct amdgpu_ttm_tt *gtt = ttm_to_amdgpu_ttm_tt(ttm);
 	int write = !(gtt->userflags & AMDGPU_GEM_USERPTR_READONLY);
 	enum dma_data_direction direction = write ?
 		DMA_BIDIRECTIONAL : DMA_TO_DEVICE;
@@ -813,64 +822,63 @@ static void amdgpu_ttm_tt_unpin_userptr(struct ttm_device *bdev,
 	/* unmap the pages mapped to the device */
 	dma_unmap_sgtable(adev->dev, ttm->sg, direction, 0);
 	sg_free_table(ttm->sg);
+}
 
-#if IS_ENABLED(CONFIG_DRM_AMDGPU_USERPTR)
-	if (gtt->range) {
-		unsigned long i;
+/*
+ * total_pages is constructed as MQD0+CtrlStack0 + MQD1+CtrlStack1 + ...
+ * MQDn+CtrlStackn where n is the number of XCCs per partition.
+ * pages_per_xcc is the size of one MQD+CtrlStack. The first page is MQD
+ * and uses memory type default, UC. The rest of pages_per_xcc are
+ * Ctrl stack and modify their memory type to NC.
+ */
+static void amdgpu_ttm_gart_bind_gfx9_mqd(struct amdgpu_device *adev,
+				struct ttm_tt *ttm, uint64_t flags)
+{
+	struct amdgpu_ttm_tt *gtt = (void *)ttm;
+	uint64_t total_pages = ttm->num_pages;
+	int num_xcc = max(1U, adev->gfx.num_xcc_per_xcp);
+	uint64_t page_idx, pages_per_xcc;
+	int i;
+	uint64_t ctrl_flags = AMDGPU_PTE_MTYPE_VG10(flags, AMDGPU_MTYPE_NC);
 
-		for (i = 0; i < ttm->num_pages; i++) {
-			if (ttm->pages[i] !=
-			    hmm_pfn_to_page(gtt->range->hmm_pfns[i]))
-				break;
-		}
+	pages_per_xcc = total_pages;
+	do_div(pages_per_xcc, num_xcc);
 
-		WARN((i == ttm->num_pages), "Missing get_user_page_done\n");
+	for (i = 0, page_idx = 0; i < num_xcc; i++, page_idx += pages_per_xcc) {
+		/* MQD page: use default flags */
+		amdgpu_gart_bind(adev,
+				gtt->offset + (page_idx << PAGE_SHIFT),
+				1, &gtt->ttm.dma_address[page_idx], flags);
+		/*
+		 * Ctrl pages - modify the memory type to NC (ctrl_flags) from
+		 * the second page of the BO onward.
+		 */
+		amdgpu_gart_bind(adev,
+				gtt->offset + ((page_idx + 1) << PAGE_SHIFT),
+				pages_per_xcc - 1,
+				&gtt->ttm.dma_address[page_idx + 1],
+				ctrl_flags);
 	}
-#endif
 }
 
-static int amdgpu_ttm_gart_bind(struct amdgpu_device *adev,
-				struct ttm_buffer_object *tbo,
-				uint64_t flags)
+static void amdgpu_ttm_gart_bind(struct amdgpu_device *adev,
+				 struct ttm_buffer_object *tbo,
+				 uint64_t flags)
 {
 	struct amdgpu_bo *abo = ttm_to_amdgpu_bo(tbo);
 	struct ttm_tt *ttm = tbo->ttm;
-	struct amdgpu_ttm_tt *gtt = (void *)ttm;
-	int r;
+	struct amdgpu_ttm_tt *gtt = ttm_to_amdgpu_ttm_tt(ttm);
 
 	if (amdgpu_bo_encrypted(abo))
 		flags |= AMDGPU_PTE_TMZ;
 
 	if (abo->flags & AMDGPU_GEM_CREATE_CP_MQD_GFX9) {
-		uint64_t page_idx = 1;
-
-		r = amdgpu_gart_bind(adev, gtt->offset, page_idx,
-				gtt->ttm.dma_address, flags);
-		if (r)
-			goto gart_bind_fail;
-
-		/* The memory type of the first page defaults to UC. Now
-		 * modify the memory type to NC from the second page of
-		 * the BO onward.
-		 */
-		flags &= ~AMDGPU_PTE_MTYPE_VG10_MASK;
-		flags |= AMDGPU_PTE_MTYPE_VG10(AMDGPU_MTYPE_NC);
-
-		r = amdgpu_gart_bind(adev,
-				gtt->offset + (page_idx << PAGE_SHIFT),
-				ttm->num_pages - page_idx,
-				&(gtt->ttm.dma_address[page_idx]), flags);
+		amdgpu_ttm_gart_bind_gfx9_mqd(adev, ttm, flags);
 	} else {
-		r = amdgpu_gart_bind(adev, gtt->offset, ttm->num_pages,
-				     gtt->ttm.dma_address, flags);
+		amdgpu_gart_bind(adev, gtt->offset, ttm->num_pages,
+				 gtt->ttm.dma_address, flags);
 	}
-
-gart_bind_fail:
-	if (r)
-		DRM_ERROR("failed to bind %u pages at 0x%08llX\n",
-			  ttm->num_pages, gtt->offset);
-
-	return r;
+	gtt->bound = true;
 }
 
 /*
@@ -884,9 +892,9 @@ static int amdgpu_ttm_backend_bind(struct ttm_device *bdev,
 				   struct ttm_resource *bo_mem)
 {
 	struct amdgpu_device *adev = amdgpu_ttm_adev(bdev);
-	struct amdgpu_ttm_tt *gtt = (void*)ttm;
+	struct amdgpu_ttm_tt *gtt = ttm_to_amdgpu_ttm_tt(ttm);
 	uint64_t flags;
-	int r = 0;
+	int r;
 
 	if (!bo_mem)
 		return -EINVAL;
@@ -897,7 +905,7 @@ static int amdgpu_ttm_backend_bind(struct ttm_device *bdev,
 	if (gtt->userptr) {
 		r = amdgpu_ttm_tt_pin_userptr(bdev, ttm);
 		if (r) {
-			DRM_ERROR("failed to pin userptr\n");
+			dev_err(adev->dev, "failed to pin userptr\n");
 			return r;
 		}
 	} else if (ttm->page_flags & TTM_TT_FLAG_EXTERNAL) {
@@ -922,11 +930,6 @@ static int amdgpu_ttm_backend_bind(struct ttm_device *bdev,
 		     ttm->num_pages, bo_mem, ttm);
 	}
 
-	if (bo_mem->mem_type == AMDGPU_PL_GDS ||
-	    bo_mem->mem_type == AMDGPU_PL_GWS ||
-	    bo_mem->mem_type == AMDGPU_PL_OA)
-		return -EINVAL;
-
 	if (bo_mem->mem_type != TTM_PL_TT ||
 	    !amdgpu_gtt_mgr_has_gart_addr(bo_mem)) {
 		gtt->offset = AMDGPU_BO_INVALID_OFFSET;
@@ -938,14 +941,10 @@ static int amdgpu_ttm_backend_bind(struct ttm_device *bdev,
 
 	/* bind pages into GART page tables */
 	gtt->offset = (u64)bo_mem->start << PAGE_SHIFT;
-	r = amdgpu_gart_bind(adev, gtt->offset, ttm->num_pages,
-		gtt->ttm.dma_address, flags);
-
-	if (r)
-		DRM_ERROR("failed to bind %u pages at 0x%08llX\n",
-			  ttm->num_pages, gtt->offset);
+	amdgpu_gart_bind(adev, gtt->offset, ttm->num_pages,
+			 gtt->ttm.dma_address, flags);
 	gtt->bound = true;
-	return r;
+	return 0;
 }
 
 /*
@@ -960,7 +959,7 @@ int amdgpu_ttm_alloc_gart(struct ttm_buffer_object *bo)
 {
 	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->bdev);
 	struct ttm_operation_ctx ctx = { false, false };
-	struct amdgpu_ttm_tt *gtt = (void *)bo->ttm;
+	struct amdgpu_ttm_tt *gtt = ttm_to_amdgpu_ttm_tt(bo->ttm);
 	struct ttm_placement placement;
 	struct ttm_place placements;
 	struct ttm_resource *tmp;
@@ -971,16 +970,12 @@ int amdgpu_ttm_alloc_gart(struct ttm_buffer_object *bo)
 		return 0;
 
 	addr = amdgpu_gmc_agp_addr(bo);
-	if (addr != AMDGPU_BO_INVALID_OFFSET) {
-		bo->resource->start = addr >> PAGE_SHIFT;
+	if (addr != AMDGPU_BO_INVALID_OFFSET)
 		return 0;
-	}
 
 	/* allocate GART space */
 	placement.num_placement = 1;
 	placement.placement = &placements;
-	placement.num_busy_placement = 1;
-	placement.busy_placement = &placements;
 	placements.fpfn = 0;
 	placements.lpfn = adev->gmc.gart_size >> PAGE_SHIFT;
 	placements.mem_type = TTM_PL_TT;
@@ -995,12 +990,7 @@ int amdgpu_ttm_alloc_gart(struct ttm_buffer_object *bo)
 
 	/* Bind pages */
 	gtt->offset = (u64)tmp->start << PAGE_SHIFT;
-	r = amdgpu_ttm_gart_bind(adev, bo, flags);
-	if (unlikely(r)) {
-		ttm_resource_free(bo, &tmp);
-		return r;
-	}
-
+	amdgpu_ttm_gart_bind(adev, bo, flags);
 	amdgpu_gart_invalidate_tlb(adev);
 	ttm_resource_free(bo, &bo->resource);
 	ttm_bo_assign_mem(bo, tmp);
@@ -1014,19 +1004,16 @@ int amdgpu_ttm_alloc_gart(struct ttm_buffer_object *bo)
  * Called by amdgpu_gtt_mgr_recover() from amdgpu_device_reset() to
  * rebind GTT pages during a GPU reset.
  */
-int amdgpu_ttm_recover_gart(struct ttm_buffer_object *tbo)
+void amdgpu_ttm_recover_gart(struct ttm_buffer_object *tbo)
 {
 	struct amdgpu_device *adev = amdgpu_ttm_adev(tbo->bdev);
 	uint64_t flags;
-	int r;
 
 	if (!tbo->ttm)
-		return 0;
+		return;
 
 	flags = amdgpu_ttm_tt_pte_flags(adev, tbo->ttm, tbo->resource);
-	r = amdgpu_ttm_gart_bind(adev, tbo, flags);
-
-	return r;
+	amdgpu_ttm_gart_bind(adev, tbo, flags);
 }
 
 /*
@@ -1039,13 +1026,12 @@ static void amdgpu_ttm_backend_unbind(struct ttm_device *bdev,
 				      struct ttm_tt *ttm)
 {
 	struct amdgpu_device *adev = amdgpu_ttm_adev(bdev);
-	struct amdgpu_ttm_tt *gtt = (void *)ttm;
-	int r;
+	struct amdgpu_ttm_tt *gtt = ttm_to_amdgpu_ttm_tt(ttm);
 
 	/* if the pages have userptr pinning then clear that first */
 	if (gtt->userptr) {
 		amdgpu_ttm_tt_unpin_userptr(bdev, ttm);
-	} else if (ttm->sg && gtt->gobj->import_attach) {
+	} else if (ttm->sg && drm_gem_is_imported(gtt->gobj)) {
 		struct dma_buf_attachment *attach;
 
 		attach = gtt->gobj->import_attach;
@@ -1060,17 +1046,14 @@ static void amdgpu_ttm_backend_unbind(struct ttm_device *bdev,
 		return;
 
 	/* unbind shouldn't be done for GDS/GWS/OA in ttm_bo_clean_mm */
-	r = amdgpu_gart_unbind(adev, gtt->offset, ttm->num_pages);
-	if (r)
-		DRM_ERROR("failed to unbind %u pages at 0x%08llX\n",
-			  gtt->ttm.num_pages, gtt->offset);
+	amdgpu_gart_unbind(adev, gtt->offset, ttm->num_pages);
 	gtt->bound = false;
 }
 
 static void amdgpu_ttm_backend_destroy(struct ttm_device *bdev,
 				       struct ttm_tt *ttm)
 {
-	struct amdgpu_ttm_tt *gtt = (void *)ttm;
+	struct amdgpu_ttm_tt *gtt = ttm_to_amdgpu_ttm_tt(ttm);
 
 	if (gtt->usertask)
 		put_task_struct(gtt->usertask);
@@ -1090,15 +1073,20 @@ static void amdgpu_ttm_backend_destroy(struct ttm_device *bdev,
 static struct ttm_tt *amdgpu_ttm_tt_create(struct ttm_buffer_object *bo,
 					   uint32_t page_flags)
 {
+	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->bdev);
 	struct amdgpu_bo *abo = ttm_to_amdgpu_bo(bo);
 	struct amdgpu_ttm_tt *gtt;
 	enum ttm_caching caching;
 
 	gtt = kzalloc(sizeof(struct amdgpu_ttm_tt), GFP_KERNEL);
-	if (gtt == NULL) {
+	if (!gtt)
 		return NULL;
-	}
+
 	gtt->gobj = &bo->base;
+	if (adev->gmc.mem_partitions && abo->xcp_id >= 0)
+		gtt->pool_id = KFD_XCP_MEM_ID(adev, abo->xcp_id);
+	else
+		gtt->pool_id = abo->xcp_id;
 
 	if (abo->flags & AMDGPU_GEM_CREATE_CPU_GTT_USWC)
 		caching = ttm_write_combined;
@@ -1124,7 +1112,8 @@ static int amdgpu_ttm_tt_populate(struct ttm_device *bdev,
 				  struct ttm_operation_ctx *ctx)
 {
 	struct amdgpu_device *adev = amdgpu_ttm_adev(bdev);
-	struct amdgpu_ttm_tt *gtt = (void *)ttm;
+	struct amdgpu_ttm_tt *gtt = ttm_to_amdgpu_ttm_tt(ttm);
+	struct ttm_pool *pool;
 	pgoff_t i;
 	int ret;
 
@@ -1139,7 +1128,11 @@ static int amdgpu_ttm_tt_populate(struct ttm_device *bdev,
 	if (ttm->page_flags & TTM_TT_FLAG_EXTERNAL)
 		return 0;
 
-	ret = ttm_pool_alloc(&adev->mman.bdev.pool, ttm, ctx);
+	if (adev->mman.ttm_pools && gtt->pool_id >= 0)
+		pool = &adev->mman.ttm_pools[gtt->pool_id];
+	else
+		pool = &adev->mman.bdev.pool;
+	ret = ttm_pool_alloc(pool, ttm, ctx);
 	if (ret)
 		return ret;
 
@@ -1158,8 +1151,9 @@ static int amdgpu_ttm_tt_populate(struct ttm_device *bdev,
 static void amdgpu_ttm_tt_unpopulate(struct ttm_device *bdev,
 				     struct ttm_tt *ttm)
 {
-	struct amdgpu_ttm_tt *gtt = (void *)ttm;
+	struct amdgpu_ttm_tt *gtt = ttm_to_amdgpu_ttm_tt(ttm);
 	struct amdgpu_device *adev;
+	struct ttm_pool *pool;
 	pgoff_t i;
 
 	amdgpu_ttm_backend_unbind(bdev, ttm);
@@ -1178,7 +1172,33 @@ static void amdgpu_ttm_tt_unpopulate(struct ttm_device *bdev,
 		ttm->pages[i]->mapping = NULL;
 
 	adev = amdgpu_ttm_adev(bdev);
-	return ttm_pool_free(&adev->mman.bdev.pool, ttm);
+
+	if (adev->mman.ttm_pools && gtt->pool_id >= 0)
+		pool = &adev->mman.ttm_pools[gtt->pool_id];
+	else
+		pool = &adev->mman.bdev.pool;
+
+	return ttm_pool_free(pool, ttm);
+}
+
+/**
+ * amdgpu_ttm_tt_get_userptr - Return the userptr GTT ttm_tt for the current
+ * task
+ *
+ * @tbo: The ttm_buffer_object that contains the userptr
+ * @user_addr:  The returned value
+ */
+int amdgpu_ttm_tt_get_userptr(const struct ttm_buffer_object *tbo,
+			      uint64_t *user_addr)
+{
+	struct amdgpu_ttm_tt *gtt;
+
+	if (!tbo->ttm)
+		return -EINVAL;
+
+	gtt = (void *)tbo->ttm;
+	*user_addr = gtt->userptr;
+	return 0;
 }
 
 /**
@@ -1189,8 +1209,9 @@ static void amdgpu_ttm_tt_unpopulate(struct ttm_device *bdev,
  * @addr:  The address in the current tasks VM space to use
  * @flags: Requirements of userptr object.
  *
- * Called by amdgpu_gem_userptr_ioctl() to bind userptr pages
- * to current task
+ * Called by amdgpu_gem_userptr_ioctl() and kfd_ioctl_alloc_memory_of_gpu() to
+ * bind userptr pages to current task and by kfd_ioctl_acquire_vm() to
+ * initialize GPU VM for a KFD process.
  */
 int amdgpu_ttm_tt_set_userptr(struct ttm_buffer_object *bo,
 			      uint64_t addr, uint32_t flags)
@@ -1207,7 +1228,7 @@ int amdgpu_ttm_tt_set_userptr(struct ttm_buffer_object *bo,
 	/* Set TTM_TT_FLAG_EXTERNAL before populate but after create. */
 	bo->ttm->page_flags |= TTM_TT_FLAG_EXTERNAL;
 
-	gtt = (void *)bo->ttm;
+	gtt = ttm_to_amdgpu_ttm_tt(bo->ttm);
 	gtt->userptr = addr;
 	gtt->userflags = flags;
 
@@ -1224,7 +1245,7 @@ int amdgpu_ttm_tt_set_userptr(struct ttm_buffer_object *bo,
  */
 struct mm_struct *amdgpu_ttm_tt_get_usermm(struct ttm_tt *ttm)
 {
-	struct amdgpu_ttm_tt *gtt = (void *)ttm;
+	struct amdgpu_ttm_tt *gtt = ttm_to_amdgpu_ttm_tt(ttm);
 
 	if (gtt == NULL)
 		return NULL;
@@ -1243,7 +1264,7 @@ struct mm_struct *amdgpu_ttm_tt_get_usermm(struct ttm_tt *ttm)
 bool amdgpu_ttm_tt_affect_userptr(struct ttm_tt *ttm, unsigned long start,
 				  unsigned long end, unsigned long *userptr)
 {
-	struct amdgpu_ttm_tt *gtt = (void *)ttm;
+	struct amdgpu_ttm_tt *gtt = ttm_to_amdgpu_ttm_tt(ttm);
 	unsigned long size;
 
 	if (gtt == NULL || !gtt->userptr)
@@ -1266,7 +1287,7 @@ bool amdgpu_ttm_tt_affect_userptr(struct ttm_tt *ttm, unsigned long start,
  */
 bool amdgpu_ttm_tt_is_userptr(struct ttm_tt *ttm)
 {
-	struct amdgpu_ttm_tt *gtt = (void *)ttm;
+	struct amdgpu_ttm_tt *gtt = ttm_to_amdgpu_ttm_tt(ttm);
 
 	if (gtt == NULL || !gtt->userptr)
 		return false;
@@ -1279,7 +1300,7 @@ bool amdgpu_ttm_tt_is_userptr(struct ttm_tt *ttm)
  */
 bool amdgpu_ttm_tt_is_readonly(struct ttm_tt *ttm)
 {
-	struct amdgpu_ttm_tt *gtt = (void *)ttm;
+	struct amdgpu_ttm_tt *gtt = ttm_to_amdgpu_ttm_tt(ttm);
 
 	if (gtt == NULL)
 		return false;
@@ -1303,10 +1324,12 @@ uint64_t amdgpu_ttm_tt_pde_flags(struct ttm_tt *ttm, struct ttm_resource *mem)
 		flags |= AMDGPU_PTE_VALID;
 
 	if (mem && (mem->mem_type == TTM_PL_TT ||
-		    mem->mem_type == AMDGPU_PL_PREEMPT)) {
+		    mem->mem_type == AMDGPU_PL_DOORBELL ||
+		    mem->mem_type == AMDGPU_PL_PREEMPT ||
+		    mem->mem_type == AMDGPU_PL_MMIO_REMAP)) {
 		flags |= AMDGPU_PTE_SYSTEM;
 
-		if (ttm->caching == ttm_cached)
+		if (ttm && ttm->caching == ttm_cached)
 			flags |= AMDGPU_PTE_SNOOPED;
 	}
 
@@ -1352,11 +1375,11 @@ uint64_t amdgpu_ttm_tt_pte_flags(struct amdgpu_device *adev, struct ttm_tt *ttm,
 static bool amdgpu_ttm_bo_eviction_valuable(struct ttm_buffer_object *bo,
 					    const struct ttm_place *place)
 {
-	unsigned long num_pages = bo->resource->num_pages;
-	struct amdgpu_res_cursor cursor;
-	struct dma_resv_list *flist;
+	struct dma_resv_iter resv_cursor;
 	struct dma_fence *f;
-	int i;
+
+	if (!amdgpu_bo_is_amdgpu_bo(bo))
+		return ttm_bo_eviction_valuable(bo, place);
 
 	/* Swapout? */
 	if (bo->resource->mem_type == TTM_PL_SYSTEM)
@@ -1370,50 +1393,27 @@ static bool amdgpu_ttm_bo_eviction_valuable(struct ttm_buffer_object *bo,
 	 * If true, then return false as any KFD process needs all its BOs to
 	 * be resident to run successfully
 	 */
-	flist = dma_resv_shared_list(bo->base.resv);
-	if (flist) {
-		for (i = 0; i < flist->shared_count; ++i) {
-			f = rcu_dereference_protected(flist->shared[i],
-				dma_resv_held(bo->base.resv));
-			if (amdkfd_fence_check_mm(f, current->mm))
-				return false;
-		}
+	dma_resv_for_each_fence(&resv_cursor, bo->base.resv,
+				DMA_RESV_USAGE_BOOKKEEP, f) {
+		if (amdkfd_fence_check_mm(f, current->mm) &&
+		    !(place->flags & TTM_PL_FLAG_CONTIGUOUS))
+			return false;
 	}
 
-	switch (bo->resource->mem_type) {
-	case AMDGPU_PL_PREEMPT:
-		/* Preemptible BOs don't own system resources managed by the
-		 * driver (pages, VRAM, GART space). They point to resources
-		 * owned by someone else (e.g. pageable memory in user mode
-		 * or a DMABuf). They are used in a preemptible context so we
-		 * can guarantee no deadlocks and good QoS in case of MMU
-		 * notifiers or DMABuf move notifiers from the resource owner.
-		 */
+	/* Preemptible BOs don't own system resources managed by the
+	 * driver (pages, VRAM, GART space). They point to resources
+	 * owned by someone else (e.g. pageable memory in user mode
+	 * or a DMABuf). They are used in a preemptible context so we
+	 * can guarantee no deadlocks and good QoS in case of MMU
+	 * notifiers or DMABuf move notifiers from the resource owner.
+	 */
+	if (bo->resource->mem_type == AMDGPU_PL_PREEMPT)
 		return false;
-	case TTM_PL_TT:
-		if (amdgpu_bo_is_amdgpu_bo(bo) &&
-		    amdgpu_bo_encrypted(ttm_to_amdgpu_bo(bo)))
-			return false;
-		return true;
-
-	case TTM_PL_VRAM:
-		/* Check each drm MM node individually */
-		amdgpu_res_first(bo->resource, 0, (u64)num_pages << PAGE_SHIFT,
-				 &cursor);
-		while (cursor.remaining) {
-			if (place->fpfn < PFN_DOWN(cursor.start + cursor.size)
-			    && !(place->lpfn &&
-				 place->lpfn <= PFN_DOWN(cursor.start)))
-				return true;
 
-			amdgpu_res_next(&cursor, cursor.size);
-		}
+	if (bo->resource->mem_type == TTM_PL_TT &&
+	    amdgpu_bo_encrypted(ttm_to_amdgpu_bo(bo)))
 		return false;
 
-	default:
-		break;
-	}
-
 	return ttm_bo_eviction_valuable(bo, place);
 }
 
@@ -1452,6 +1452,68 @@ static void amdgpu_ttm_vram_mm_access(struct amdgpu_device *adev, loff_t pos,
 	}
 }
 
+static int amdgpu_ttm_access_memory_sdma(struct ttm_buffer_object *bo,
+					unsigned long offset, void *buf,
+					int len, int write)
+{
+	struct amdgpu_bo *abo = ttm_to_amdgpu_bo(bo);
+	struct amdgpu_device *adev = amdgpu_ttm_adev(abo->tbo.bdev);
+	struct amdgpu_res_cursor src_mm;
+	struct amdgpu_job *job;
+	struct dma_fence *fence;
+	uint64_t src_addr, dst_addr;
+	unsigned int num_dw;
+	int r, idx;
+
+	if (len != PAGE_SIZE)
+		return -EINVAL;
+
+	if (!adev->mman.sdma_access_ptr)
+		return -EACCES;
+
+	if (!drm_dev_enter(adev_to_drm(adev), &idx))
+		return -ENODEV;
+
+	if (write)
+		memcpy(adev->mman.sdma_access_ptr, buf, len);
+
+	num_dw = ALIGN(adev->mman.buffer_funcs->copy_num_dw, 8);
+	r = amdgpu_job_alloc_with_ib(adev, &adev->mman.high_pr,
+				     AMDGPU_FENCE_OWNER_UNDEFINED,
+				     num_dw * 4, AMDGPU_IB_POOL_DELAYED,
+				     &job,
+				     AMDGPU_KERNEL_JOB_ID_TTM_ACCESS_MEMORY_SDMA);
+	if (r)
+		goto out;
+
+	mutex_lock(&adev->mman.gtt_window_lock);
+	amdgpu_res_first(abo->tbo.resource, offset, len, &src_mm);
+	src_addr = amdgpu_ttm_domain_start(adev, bo->resource->mem_type) +
+		src_mm.start;
+	dst_addr = amdgpu_bo_gpu_offset(adev->mman.sdma_access_bo);
+	if (write)
+		swap(src_addr, dst_addr);
+
+	amdgpu_emit_copy_buffer(adev, &job->ibs[0], src_addr, dst_addr,
+				PAGE_SIZE, 0);
+
+	amdgpu_ring_pad_ib(adev->mman.buffer_funcs_ring, &job->ibs[0]);
+	WARN_ON(job->ibs[0].length_dw > num_dw);
+
+	fence = amdgpu_job_submit(job);
+	mutex_unlock(&adev->mman.gtt_window_lock);
+
+	if (!dma_fence_wait_timeout(fence, false, adev->sdma_timeout))
+		r = -ETIMEDOUT;
+	dma_fence_put(fence);
+
+	if (!(r || write))
+		memcpy(buf, adev->mman.sdma_access_ptr, len);
+out:
+	drm_dev_exit(idx);
+	return r;
+}
+
 /**
  * amdgpu_ttm_access_memory - Read or Write memory that backs a buffer object.
  *
@@ -1476,6 +1538,10 @@ static int amdgpu_ttm_access_memory(struct ttm_buffer_object *bo,
 	if (bo->resource->mem_type != TTM_PL_VRAM)
 		return -EIO;
 
+	if (amdgpu_device_has_timeouts_enabled(adev) &&
+			!amdgpu_ttm_access_memory_sdma(bo, offset, buf, len, write))
+		return len;
+
 	amdgpu_res_first(bo->resource, offset, len, &cursor);
 	while (cursor.remaining) {
 		size_t count, size = cursor.size;
@@ -1517,7 +1583,6 @@ static struct ttm_device_funcs amdgpu_bo_driver = {
 	.io_mem_reserve = &amdgpu_ttm_io_mem_reserve,
 	.io_mem_pfn = amdgpu_ttm_io_mem_pfn,
 	.access_memory = &amdgpu_ttm_access_memory,
-	.del_from_lru_notify = &amdgpu_vm_del_from_lru_notify
 };
 
 /*
@@ -1536,6 +1601,23 @@ static void amdgpu_ttm_fw_reserve_vram_fini(struct amdgpu_device *adev)
 		NULL, &adev->mman.fw_vram_usage_va);
 }
 
+/*
+ * Driver Reservation functions
+ */
+/**
+ * amdgpu_ttm_drv_reserve_vram_fini - free drv reserved vram
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * free drv reserved vram if it has been reserved.
+ */
+static void amdgpu_ttm_drv_reserve_vram_fini(struct amdgpu_device *adev)
+{
+	amdgpu_bo_free_kernel(&adev->mman.drv_vram_usage_reserved_bo,
+						  NULL,
+						  &adev->mman.drv_vram_usage_va);
+}
+
 /**
  * amdgpu_ttm_fw_reserve_vram_init - create bo vram reservation from fw
  *
@@ -1557,11 +1639,35 @@ static int amdgpu_ttm_fw_reserve_vram_init(struct amdgpu_device *adev)
 	return amdgpu_bo_create_kernel_at(adev,
 					  adev->mman.fw_vram_usage_start_offset,
 					  adev->mman.fw_vram_usage_size,
-					  AMDGPU_GEM_DOMAIN_VRAM,
 					  &adev->mman.fw_vram_usage_reserved_bo,
 					  &adev->mman.fw_vram_usage_va);
 }
 
+/**
+ * amdgpu_ttm_drv_reserve_vram_init - create bo vram reservation from driver
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * create bo vram reservation from drv.
+ */
+static int amdgpu_ttm_drv_reserve_vram_init(struct amdgpu_device *adev)
+{
+	u64 vram_size = adev->gmc.visible_vram_size;
+
+	adev->mman.drv_vram_usage_va = NULL;
+	adev->mman.drv_vram_usage_reserved_bo = NULL;
+
+	if (adev->mman.drv_vram_usage_size == 0 ||
+	    adev->mman.drv_vram_usage_size > vram_size)
+		return 0;
+
+	return amdgpu_bo_create_kernel_at(adev,
+					  adev->mman.drv_vram_usage_start_offset,
+					  adev->mman.drv_vram_usage_size,
+					  &adev->mman.drv_vram_usage_reserved_bo,
+					  &adev->mman.drv_vram_usage_va);
+}
+
 /*
  * Memoy training reservation functions
  */
@@ -1584,14 +1690,15 @@ static int amdgpu_ttm_training_reserve_vram_fini(struct amdgpu_device *adev)
 	return 0;
 }
 
-static void amdgpu_ttm_training_data_block_init(struct amdgpu_device *adev)
+static void amdgpu_ttm_training_data_block_init(struct amdgpu_device *adev,
+						uint32_t reserve_size)
 {
 	struct psp_memory_training_context *ctx = &adev->psp.mem_train_ctx;
 
 	memset(ctx, 0, sizeof(*ctx));
 
 	ctx->c2p_train_data_offset =
-		ALIGN((adev->gmc.mc_vram_size - adev->mman.discovery_tmr_size - SZ_1M), SZ_1M);
+		ALIGN((adev->gmc.mc_vram_size - reserve_size - SZ_1M), SZ_1M);
 	ctx->p2c_train_data_offset =
 		(adev->gmc.mc_vram_size - GDDR6_MEM_TRAINING_OFFSET);
 	ctx->train_data_size =
@@ -1609,11 +1716,12 @@ static void amdgpu_ttm_training_data_block_init(struct amdgpu_device *adev)
  */
 static int amdgpu_ttm_reserve_tmr(struct amdgpu_device *adev)
 {
-	int ret;
 	struct psp_memory_training_context *ctx = &adev->psp.mem_train_ctx;
 	bool mem_train_support = false;
+	uint32_t reserve_size = 0;
+	int ret;
 
-	if (!amdgpu_sriov_vf(adev)) {
+	if (adev->bios && !amdgpu_sriov_vf(adev)) {
 		if (amdgpu_atomfirmware_mem_training_supported(adev))
 			mem_train_support = true;
 		else
@@ -1627,43 +1735,135 @@ static int amdgpu_ttm_reserve_tmr(struct amdgpu_device *adev)
 	 * Otherwise, fallback to legacy approach to check and reserve tmr block for ip
 	 * discovery data and G6 memory training data respectively
 	 */
-	adev->mman.discovery_tmr_size =
-		amdgpu_atomfirmware_get_fw_reserved_fb_size(adev);
-	if (!adev->mman.discovery_tmr_size)
-		adev->mman.discovery_tmr_size = DISCOVERY_TMR_OFFSET;
+	if (adev->bios)
+		reserve_size =
+			amdgpu_atomfirmware_get_fw_reserved_fb_size(adev);
+
+	if (!adev->bios &&
+	    (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3) ||
+	     amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4) ||
+	     amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 5, 0)))
+		reserve_size = max(reserve_size, (uint32_t)280 << 20);
+	else if (!reserve_size)
+		reserve_size = DISCOVERY_TMR_OFFSET;
 
 	if (mem_train_support) {
 		/* reserve vram for mem train according to TMR location */
-		amdgpu_ttm_training_data_block_init(adev);
+		amdgpu_ttm_training_data_block_init(adev, reserve_size);
 		ret = amdgpu_bo_create_kernel_at(adev,
-					 ctx->c2p_train_data_offset,
-					 ctx->train_data_size,
-					 AMDGPU_GEM_DOMAIN_VRAM,
-					 &ctx->c2p_bo,
-					 NULL);
+						 ctx->c2p_train_data_offset,
+						 ctx->train_data_size,
+						 &ctx->c2p_bo,
+						 NULL);
 		if (ret) {
-			DRM_ERROR("alloc c2p_bo failed(%d)!\n", ret);
+			dev_err(adev->dev, "alloc c2p_bo failed(%d)!\n", ret);
 			amdgpu_ttm_training_reserve_vram_fini(adev);
 			return ret;
 		}
 		ctx->init = PSP_MEM_TRAIN_RESERVE_SUCCESS;
 	}
 
-	ret = amdgpu_bo_create_kernel_at(adev,
-				adev->gmc.real_vram_size - adev->mman.discovery_tmr_size,
-				adev->mman.discovery_tmr_size,
-				AMDGPU_GEM_DOMAIN_VRAM,
-				&adev->mman.discovery_memory,
-				NULL);
+	ret = amdgpu_bo_create_kernel_at(
+		adev, adev->gmc.real_vram_size - reserve_size, reserve_size,
+		&adev->mman.fw_reserved_memory, NULL);
 	if (ret) {
-		DRM_ERROR("alloc tmr failed(%d)!\n", ret);
-		amdgpu_bo_free_kernel(&adev->mman.discovery_memory, NULL, NULL);
+		dev_err(adev->dev, "alloc tmr failed(%d)!\n", ret);
+		amdgpu_bo_free_kernel(&adev->mman.fw_reserved_memory, NULL,
+				      NULL);
 		return ret;
 	}
 
 	return 0;
 }
 
+static int amdgpu_ttm_pools_init(struct amdgpu_device *adev)
+{
+	int i;
+
+	if (!adev->gmc.is_app_apu || !adev->gmc.num_mem_partitions)
+		return 0;
+
+	adev->mman.ttm_pools = kcalloc(adev->gmc.num_mem_partitions,
+				       sizeof(*adev->mman.ttm_pools),
+				       GFP_KERNEL);
+	if (!adev->mman.ttm_pools)
+		return -ENOMEM;
+
+	for (i = 0; i < adev->gmc.num_mem_partitions; i++) {
+		ttm_pool_init(&adev->mman.ttm_pools[i], adev->dev,
+			      adev->gmc.mem_partitions[i].numa.node,
+			      TTM_ALLOCATION_POOL_BENEFICIAL_ORDER(get_order(SZ_2M)));
+	}
+	return 0;
+}
+
+static void amdgpu_ttm_pools_fini(struct amdgpu_device *adev)
+{
+	int i;
+
+	if (!adev->gmc.is_app_apu || !adev->mman.ttm_pools)
+		return;
+
+	for (i = 0; i < adev->gmc.num_mem_partitions; i++)
+		ttm_pool_fini(&adev->mman.ttm_pools[i]);
+
+	kfree(adev->mman.ttm_pools);
+	adev->mman.ttm_pools = NULL;
+}
+
+/**
+ * amdgpu_ttm_mmio_remap_bo_init - Allocate the singleton 4K MMIO_REMAP BO
+ * @adev: amdgpu device
+ *
+ * Allocates a one-page (4K) GEM BO in AMDGPU_GEM_DOMAIN_MMIO_REMAP when the
+ * hardware exposes a remap base (adev->rmmio_remap.bus_addr) and the host
+ * PAGE_SIZE is <= AMDGPU_GPU_PAGE_SIZE (4K). The BO is created as a regular
+ * GEM object (amdgpu_bo_create).
+ *
+ * Return:
+ *  * 0 on success or intentional skip (feature not present/unsupported)
+ *  * negative errno on allocation failure
+ */
+static int amdgpu_ttm_mmio_remap_bo_init(struct amdgpu_device *adev)
+{
+	struct amdgpu_bo_param bp;
+	int r;
+
+	/* Skip if HW doesn't expose remap, or if PAGE_SIZE > AMDGPU_GPU_PAGE_SIZE (4K). */
+	if (!adev->rmmio_remap.bus_addr || PAGE_SIZE > AMDGPU_GPU_PAGE_SIZE)
+		return 0;
+
+	memset(&bp, 0, sizeof(bp));
+
+	/* Create exactly one GEM BO in the MMIO_REMAP domain. */
+	bp.type        = ttm_bo_type_device;          /* userspace-mappable GEM */
+	bp.size        = AMDGPU_GPU_PAGE_SIZE;        /* 4K */
+	bp.byte_align  = AMDGPU_GPU_PAGE_SIZE;
+	bp.domain      = AMDGPU_GEM_DOMAIN_MMIO_REMAP;
+	bp.flags       = 0;
+	bp.resv        = NULL;
+	bp.bo_ptr_size = sizeof(struct amdgpu_bo);
+
+	r = amdgpu_bo_create(adev, &bp, &adev->rmmio_remap.bo);
+	if (r)
+		return r;
+
+	return 0;
+}
+
+/**
+ * amdgpu_ttm_mmio_remap_bo_fini - Free the singleton MMIO_REMAP BO
+ * @adev: amdgpu device
+ *
+ * Frees the kernel-owned MMIO_REMAP BO if it was allocated by
+ * amdgpu_ttm_mmio_remap_bo_init().
+ */
+static void amdgpu_ttm_mmio_remap_bo_fini(struct amdgpu_device *adev)
+{
+	amdgpu_bo_unref(&adev->rmmio_remap.bo);
+	adev->rmmio_remap.bo = NULL;
+}
+
 /*
  * amdgpu_ttm_init - Init the memory management (ttm) as well as various
  * gtt/vram related fields.
@@ -1677,34 +1877,40 @@ int amdgpu_ttm_init(struct amdgpu_device *adev)
 {
 	uint64_t gtt_size;
 	int r;
-	u64 vis_vram_limit;
 
 	mutex_init(&adev->mman.gtt_window_lock);
 
+	dma_set_max_seg_size(adev->dev, UINT_MAX);
 	/* No others user of address space so set it to 0 */
 	r = ttm_device_init(&adev->mman.bdev, &amdgpu_bo_driver, adev->dev,
 			       adev_to_drm(adev)->anon_inode->i_mapping,
 			       adev_to_drm(adev)->vma_offset_manager,
-			       adev->need_swiotlb,
-			       dma_addressing_limited(adev->dev));
+			       (adev->need_swiotlb ?
+				TTM_ALLOCATION_POOL_USE_DMA_ALLOC : 0) |
+			       (dma_addressing_limited(adev->dev) ?
+				TTM_ALLOCATION_POOL_USE_DMA32 : 0) |
+			       TTM_ALLOCATION_POOL_BENEFICIAL_ORDER(get_order(SZ_2M)));
 	if (r) {
-		DRM_ERROR("failed initializing buffer object driver(%d).\n", r);
+		dev_err(adev->dev,
+			"failed initializing buffer object driver(%d).\n", r);
 		return r;
 	}
-	adev->mman.initialized = true;
 
-	/* Initialize VRAM pool with all of VRAM divided into pages */
-	r = amdgpu_vram_mgr_init(adev);
+	r = amdgpu_ttm_pools_init(adev);
 	if (r) {
-		DRM_ERROR("Failed initializing VRAM heap.\n");
+		dev_err(adev->dev, "failed to init ttm pools(%d).\n", r);
 		return r;
 	}
+	adev->mman.initialized = true;
 
-	/* Reduce size of CPU-visible VRAM if requested */
-	vis_vram_limit = (u64)amdgpu_vis_vram_limit * 1024 * 1024;
-	if (amdgpu_vis_vram_limit > 0 &&
-	    vis_vram_limit <= adev->gmc.visible_vram_size)
-		adev->gmc.visible_vram_size = vis_vram_limit;
+	if (!adev->gmc.is_app_apu) {
+		/* Initialize VRAM pool with all of VRAM divided into pages */
+		r = amdgpu_vram_mgr_init(adev);
+		if (r) {
+			dev_err(adev->dev, "Failed initializing VRAM heap.\n");
+			return r;
+		}
+	}
 
 	/* Change the size here instead of the init above so only lpfn is affected */
 	amdgpu_ttm_set_buffer_funcs_status(adev, false);
@@ -1714,6 +1920,9 @@ int amdgpu_ttm_init(struct amdgpu_device *adev)
 		adev->mman.aper_base_kaddr = ioremap_cache(adev->gmc.aper_base,
 				adev->gmc.visible_vram_size);
 
+	else if (adev->gmc.is_app_apu)
+		DRM_DEBUG_DRIVER(
+			"No need to ioremap when real vram size is 0\n");
 	else
 #endif
 		adev->mman.aper_base_kaddr = ioremap_wc(adev->gmc.aper_base,
@@ -1725,16 +1934,23 @@ int amdgpu_ttm_init(struct amdgpu_device *adev)
 	 *place on the VRAM, so reserve it early.
 	 */
 	r = amdgpu_ttm_fw_reserve_vram_init(adev);
-	if (r) {
+	if (r)
+		return r;
+
+	/*
+	 * The reserved VRAM for the driver must be pinned to a specific
+	 * location in VRAM, so reserve it early.
+	 */
+	r = amdgpu_ttm_drv_reserve_vram_init(adev);
+	if (r)
 		return r;
-	}
 
 	/*
-	 * only NAVI10 and onwards ASIC support for IP discovery.
-	 * If IP discovery enabled, a block of memory should be
-	 * reserved for IP discovey.
+	 * only NAVI10 and later ASICs support IP discovery.
+	 * If IP discovery is enabled, a block of memory should be
+	 * reserved for it.
 	 */
-	if (adev->mman.discovery_bin) {
+	if (adev->discovery.reserve_tmr) {
 		r = amdgpu_ttm_reserve_tmr(adev);
 		if (r)
 			return r;
@@ -1743,78 +1959,125 @@ int amdgpu_ttm_init(struct amdgpu_device *adev)
 	/* allocate memory as required for VGA
 	 * This is used for VGA emulation and pre-OS scanout buffers to
 	 * avoid display artifacts while transitioning between pre-OS
-	 * and driver.  */
-	r = amdgpu_bo_create_kernel_at(adev, 0, adev->mman.stolen_vga_size,
-				       AMDGPU_GEM_DOMAIN_VRAM,
-				       &adev->mman.stolen_vga_memory,
-				       NULL);
-	if (r)
-		return r;
-	r = amdgpu_bo_create_kernel_at(adev, adev->mman.stolen_vga_size,
-				       adev->mman.stolen_extended_size,
-				       AMDGPU_GEM_DOMAIN_VRAM,
-				       &adev->mman.stolen_extended_memory,
-				       NULL);
-	if (r)
-		return r;
-	r = amdgpu_bo_create_kernel_at(adev, adev->mman.stolen_reserved_offset,
-				       adev->mman.stolen_reserved_size,
-				       AMDGPU_GEM_DOMAIN_VRAM,
-				       &adev->mman.stolen_reserved_memory,
-				       NULL);
-	if (r)
-		return r;
+	 * and driver.
+	 */
+	if (!adev->gmc.is_app_apu) {
+		r = amdgpu_bo_create_kernel_at(adev, 0,
+					       adev->mman.stolen_vga_size,
+					       &adev->mman.stolen_vga_memory,
+					       NULL);
+		if (r)
+			return r;
+
+		r = amdgpu_bo_create_kernel_at(adev, adev->mman.stolen_vga_size,
+					       adev->mman.stolen_extended_size,
+					       &adev->mman.stolen_extended_memory,
+					       NULL);
+
+		if (r)
+			return r;
+
+		r = amdgpu_bo_create_kernel_at(adev,
+					       adev->mman.stolen_reserved_offset,
+					       adev->mman.stolen_reserved_size,
+					       &adev->mman.stolen_reserved_memory,
+					       NULL);
+		if (r)
+			return r;
+	} else {
+		DRM_DEBUG_DRIVER("Skipped stolen memory reservation\n");
+	}
 
-	DRM_INFO("amdgpu: %uM of VRAM memory ready\n",
-		 (unsigned) (adev->gmc.real_vram_size / (1024 * 1024)));
+	dev_info(adev->dev, "amdgpu: %uM of VRAM memory ready\n",
+		 (unsigned int)(adev->gmc.real_vram_size / (1024 * 1024)));
 
-	/* Compute GTT size, either bsaed on 3/4th the size of RAM size
-	 * or whatever the user passed on module init */
-	if (amdgpu_gtt_size == -1) {
-		struct sysinfo si;
+	/* Compute GTT size, either based on TTM limit
+	 * or whatever the user passed on module init.
+	 */
+	gtt_size = ttm_tt_pages_limit() << PAGE_SHIFT;
+	if (amdgpu_gtt_size != -1) {
+		uint64_t configured_size = (uint64_t)amdgpu_gtt_size << 20;
+
+		drm_warn(&adev->ddev,
+			"Configuring gttsize via module parameter is deprecated, please use ttm.pages_limit\n");
+		if (gtt_size != configured_size)
+			drm_warn(&adev->ddev,
+				"GTT size has been set as %llu but TTM size has been set as %llu, this is unusual\n",
+				configured_size, gtt_size);
 
-		si_meminfo(&si);
-		gtt_size = min(max((AMDGPU_DEFAULT_GTT_SIZE_MB << 20),
-			       adev->gmc.mc_vram_size),
-			       ((uint64_t)si.totalram * si.mem_unit * 3/4));
+		gtt_size = configured_size;
 	}
-	else
-		gtt_size = (uint64_t)amdgpu_gtt_size << 20;
 
 	/* Initialize GTT memory pool */
 	r = amdgpu_gtt_mgr_init(adev, gtt_size);
 	if (r) {
-		DRM_ERROR("Failed initializing GTT heap.\n");
+		dev_err(adev->dev, "Failed initializing GTT heap.\n");
+		return r;
+	}
+	dev_info(adev->dev, "amdgpu: %uM of GTT memory ready.\n",
+		 (unsigned int)(gtt_size / (1024 * 1024)));
+
+	if (adev->flags & AMD_IS_APU) {
+		if (adev->gmc.real_vram_size < gtt_size)
+			adev->apu_prefer_gtt = true;
+	}
+
+	/* Initialize doorbell pool on PCI BAR */
+	r = amdgpu_ttm_init_on_chip(adev, AMDGPU_PL_DOORBELL, adev->doorbell.size / PAGE_SIZE);
+	if (r) {
+		dev_err(adev->dev, "Failed initializing doorbell heap.\n");
+		return r;
+	}
+
+	/* Create a boorbell page for kernel usages */
+	r = amdgpu_doorbell_create_kernel_doorbells(adev);
+	if (r) {
+		dev_err(adev->dev, "Failed to initialize kernel doorbells.\n");
+		return r;
+	}
+
+	/* Initialize MMIO-remap pool (single page 4K) */
+	r = amdgpu_ttm_init_on_chip(adev, AMDGPU_PL_MMIO_REMAP, 1);
+	if (r) {
+		dev_err(adev->dev, "Failed initializing MMIO-remap heap.\n");
 		return r;
 	}
-	DRM_INFO("amdgpu: %uM of GTT memory ready.\n",
-		 (unsigned)(gtt_size / (1024 * 1024)));
+
+	/* Allocate the singleton MMIO_REMAP BO (4K) if supported */
+	r = amdgpu_ttm_mmio_remap_bo_init(adev);
+	if (r)
+		return r;
 
 	/* Initialize preemptible memory pool */
 	r = amdgpu_preempt_mgr_init(adev);
 	if (r) {
-		DRM_ERROR("Failed initializing PREEMPT heap.\n");
+		dev_err(adev->dev, "Failed initializing PREEMPT heap.\n");
 		return r;
 	}
 
 	/* Initialize various on-chip memory pools */
 	r = amdgpu_ttm_init_on_chip(adev, AMDGPU_PL_GDS, adev->gds.gds_size);
 	if (r) {
-		DRM_ERROR("Failed initializing GDS heap.\n");
+		dev_err(adev->dev, "Failed initializing GDS heap.\n");
 		return r;
 	}
 
 	r = amdgpu_ttm_init_on_chip(adev, AMDGPU_PL_GWS, adev->gds.gws_size);
 	if (r) {
-		DRM_ERROR("Failed initializing gws heap.\n");
+		dev_err(adev->dev, "Failed initializing gws heap.\n");
 		return r;
 	}
 
 	r = amdgpu_ttm_init_on_chip(adev, AMDGPU_PL_OA, adev->gds.oa_size);
 	if (r) {
-		DRM_ERROR("Failed initializing oa heap.\n");
+		dev_err(adev->dev, "Failed initializing oa heap.\n");
 		return r;
 	}
+	if (amdgpu_bo_create_kernel(adev, PAGE_SIZE, PAGE_SIZE,
+				AMDGPU_GEM_DOMAIN_GTT,
+				&adev->mman.sdma_access_bo, NULL,
+				&adev->mman.sdma_access_ptr))
+		DRM_WARN("Debug VRAM access will use slowpath MM access\n");
 
 	return 0;
 }
@@ -1824,29 +2087,57 @@ int amdgpu_ttm_init(struct amdgpu_device *adev)
  */
 void amdgpu_ttm_fini(struct amdgpu_device *adev)
 {
+	int idx;
+
 	if (!adev->mman.initialized)
 		return;
 
+	amdgpu_ttm_pools_fini(adev);
+
 	amdgpu_ttm_training_reserve_vram_fini(adev);
 	/* return the stolen vga memory back to VRAM */
-	amdgpu_bo_free_kernel(&adev->mman.stolen_vga_memory, NULL, NULL);
-	amdgpu_bo_free_kernel(&adev->mman.stolen_extended_memory, NULL, NULL);
-	/* return the IP Discovery TMR memory back to VRAM */
-	amdgpu_bo_free_kernel(&adev->mman.discovery_memory, NULL, NULL);
-	if (adev->mman.stolen_reserved_size)
-		amdgpu_bo_free_kernel(&adev->mman.stolen_reserved_memory,
-				      NULL, NULL);
+	if (!adev->gmc.is_app_apu) {
+		amdgpu_bo_free_kernel(&adev->mman.stolen_vga_memory, NULL, NULL);
+		amdgpu_bo_free_kernel(&adev->mman.stolen_extended_memory, NULL, NULL);
+		/* return the FW reserved memory back to VRAM */
+		amdgpu_bo_free_kernel(&adev->mman.fw_reserved_memory, NULL,
+				      NULL);
+		amdgpu_bo_free_kernel(&adev->mman.fw_reserved_memory_extend, NULL,
+				      NULL);
+		if (adev->mman.stolen_reserved_size)
+			amdgpu_bo_free_kernel(&adev->mman.stolen_reserved_memory,
+					      NULL, NULL);
+	}
+	amdgpu_bo_free_kernel(&adev->mman.sdma_access_bo, NULL,
+					&adev->mman.sdma_access_ptr);
+
+	amdgpu_ttm_mmio_remap_bo_fini(adev);
 	amdgpu_ttm_fw_reserve_vram_fini(adev);
+	amdgpu_ttm_drv_reserve_vram_fini(adev);
+
+	if (drm_dev_enter(adev_to_drm(adev), &idx)) {
+
+		if (adev->mman.aper_base_kaddr)
+			iounmap(adev->mman.aper_base_kaddr);
+		adev->mman.aper_base_kaddr = NULL;
 
-	amdgpu_vram_mgr_fini(adev);
+		drm_dev_exit(idx);
+	}
+
+	if (!adev->gmc.is_app_apu)
+		amdgpu_vram_mgr_fini(adev);
 	amdgpu_gtt_mgr_fini(adev);
 	amdgpu_preempt_mgr_fini(adev);
+	amdgpu_doorbell_fini(adev);
+
 	ttm_range_man_fini(&adev->mman.bdev, AMDGPU_PL_GDS);
 	ttm_range_man_fini(&adev->mman.bdev, AMDGPU_PL_GWS);
 	ttm_range_man_fini(&adev->mman.bdev, AMDGPU_PL_OA);
+	ttm_range_man_fini(&adev->mman.bdev, AMDGPU_PL_DOORBELL);
+	ttm_range_man_fini(&adev->mman.bdev, AMDGPU_PL_MMIO_REMAP);
 	ttm_device_fini(&adev->mman.bdev);
 	adev->mman.initialized = false;
-	DRM_INFO("amdgpu: ttm finalized\n");
+	dev_info(adev->dev, "amdgpu: ttm finalized\n");
 }
 
 /**
@@ -1865,7 +2156,7 @@ void amdgpu_ttm_set_buffer_funcs_status(struct amdgpu_device *adev, bool enable)
 	int r;
 
 	if (!adev->mman.initialized || amdgpu_in_reset(adev) ||
-	    adev->mman.buffer_funcs_enabled == enable)
+	    adev->mman.buffer_funcs_enabled == enable || adev->gmc.is_app_apu)
 		return;
 
 	if (enable) {
@@ -1874,18 +2165,32 @@ void amdgpu_ttm_set_buffer_funcs_status(struct amdgpu_device *adev, bool enable)
 
 		ring = adev->mman.buffer_funcs_ring;
 		sched = &ring->sched;
-		r = drm_sched_entity_init(&adev->mman.entity,
+		r = drm_sched_entity_init(&adev->mman.high_pr,
 					  DRM_SCHED_PRIORITY_KERNEL, &sched,
 					  1, NULL);
 		if (r) {
-			DRM_ERROR("Failed setting up TTM BO move entity (%d)\n",
-				  r);
+			dev_err(adev->dev,
+				"Failed setting up TTM BO move entity (%d)\n",
+				r);
 			return;
 		}
+
+		r = drm_sched_entity_init(&adev->mman.low_pr,
+					  DRM_SCHED_PRIORITY_NORMAL, &sched,
+					  1, NULL);
+		if (r) {
+			dev_err(adev->dev,
+				"Failed setting up TTM BO move entity (%d)\n",
+				r);
+			goto error_free_entity;
+		}
 	} else {
-		drm_sched_entity_destroy(&adev->mman.entity);
-		dma_fence_put(man->move);
-		man->move = NULL;
+		drm_sched_entity_destroy(&adev->mman.high_pr);
+		drm_sched_entity_destroy(&adev->mman.low_pr);
+		/* Drop all the old fences since re-creating the scheduler entities
+		 * will allocate new contexts.
+		 */
+		ttm_resource_manager_cleanup(man);
 	}
 
 	/* this just adjusts TTM size idea, which sets lpfn to the correct value */
@@ -1893,60 +2198,81 @@ void amdgpu_ttm_set_buffer_funcs_status(struct amdgpu_device *adev, bool enable)
 		size = adev->gmc.real_vram_size;
 	else
 		size = adev->gmc.visible_vram_size;
-	man->size = size >> PAGE_SHIFT;
+	man->size = size;
 	adev->mman.buffer_funcs_enabled = enable;
+
+	return;
+
+error_free_entity:
+	drm_sched_entity_destroy(&adev->mman.high_pr);
+}
+
+static int amdgpu_ttm_prepare_job(struct amdgpu_device *adev,
+				  bool direct_submit,
+				  unsigned int num_dw,
+				  struct dma_resv *resv,
+				  bool vm_needs_flush,
+				  struct amdgpu_job **job,
+				  bool delayed, u64 k_job_id)
+{
+	enum amdgpu_ib_pool_type pool = direct_submit ?
+		AMDGPU_IB_POOL_DIRECT :
+		AMDGPU_IB_POOL_DELAYED;
+	int r;
+	struct drm_sched_entity *entity = delayed ? &adev->mman.low_pr :
+						    &adev->mman.high_pr;
+	r = amdgpu_job_alloc_with_ib(adev, entity,
+				     AMDGPU_FENCE_OWNER_UNDEFINED,
+				     num_dw * 4, pool, job, k_job_id);
+	if (r)
+		return r;
+
+	if (vm_needs_flush) {
+		(*job)->vm_pd_addr = amdgpu_gmc_pd_addr(adev->gmc.pdb0_bo ?
+							adev->gmc.pdb0_bo :
+							adev->gart.bo);
+		(*job)->vm_needs_flush = true;
+	}
+	if (!resv)
+		return 0;
+
+	return drm_sched_job_add_resv_dependencies(&(*job)->base, resv,
+						   DMA_RESV_USAGE_BOOKKEEP);
 }
 
 int amdgpu_copy_buffer(struct amdgpu_ring *ring, uint64_t src_offset,
 		       uint64_t dst_offset, uint32_t byte_count,
 		       struct dma_resv *resv,
 		       struct dma_fence **fence, bool direct_submit,
-		       bool vm_needs_flush, bool tmz)
+		       bool vm_needs_flush, uint32_t copy_flags)
 {
-	enum amdgpu_ib_pool_type pool = direct_submit ? AMDGPU_IB_POOL_DIRECT :
-		AMDGPU_IB_POOL_DELAYED;
 	struct amdgpu_device *adev = ring->adev;
+	unsigned int num_loops, num_dw;
 	struct amdgpu_job *job;
-
 	uint32_t max_bytes;
-	unsigned num_loops, num_dw;
-	unsigned i;
+	unsigned int i;
 	int r;
 
-	if (direct_submit && !ring->sched.ready) {
-		DRM_ERROR("Trying to move memory with ring turned off.\n");
+	if (!direct_submit && !ring->sched.ready) {
+		dev_err(adev->dev,
+			"Trying to move memory with ring turned off.\n");
 		return -EINVAL;
 	}
 
 	max_bytes = adev->mman.buffer_funcs->copy_max_bytes;
 	num_loops = DIV_ROUND_UP(byte_count, max_bytes);
 	num_dw = ALIGN(num_loops * adev->mman.buffer_funcs->copy_num_dw, 8);
-
-	r = amdgpu_job_alloc_with_ib(adev, num_dw * 4, pool, &job);
+	r = amdgpu_ttm_prepare_job(adev, direct_submit, num_dw,
+				   resv, vm_needs_flush, &job, false,
+				   AMDGPU_KERNEL_JOB_ID_TTM_COPY_BUFFER);
 	if (r)
 		return r;
 
-	if (vm_needs_flush) {
-		job->vm_pd_addr = amdgpu_gmc_pd_addr(adev->gmc.pdb0_bo ?
-					adev->gmc.pdb0_bo : adev->gart.bo);
-		job->vm_needs_flush = true;
-	}
-	if (resv) {
-		r = amdgpu_sync_resv(adev, &job->sync, resv,
-				     AMDGPU_SYNC_ALWAYS,
-				     AMDGPU_FENCE_OWNER_UNDEFINED);
-		if (r) {
-			DRM_ERROR("sync failed (%d).\n", r);
-			goto error_free;
-		}
-	}
-
 	for (i = 0; i < num_loops; i++) {
 		uint32_t cur_size_in_bytes = min(byte_count, max_bytes);
 
 		amdgpu_emit_copy_buffer(adev, &job->ibs[0], src_offset,
-					dst_offset, cur_size_in_bytes, tmz);
-
+					dst_offset, cur_size_in_bytes, copy_flags);
 		src_offset += cur_size_in_bytes;
 		dst_offset += cur_size_in_bytes;
 		byte_count -= cur_size_in_bytes;
@@ -1957,8 +2283,7 @@ int amdgpu_copy_buffer(struct amdgpu_ring *ring, uint64_t src_offset,
 	if (direct_submit)
 		r = amdgpu_job_submit_direct(job, ring, fence);
 	else
-		r = amdgpu_job_submit(job, &adev->mman.entity,
-				      AMDGPU_FENCE_OWNER_UNDEFINED, fence);
+		*fence = amdgpu_job_submit(job);
 	if (r)
 		goto error_free;
 
@@ -1966,94 +2291,163 @@ int amdgpu_copy_buffer(struct amdgpu_ring *ring, uint64_t src_offset,
 
 error_free:
 	amdgpu_job_free(job);
-	DRM_ERROR("Error scheduling IBs (%d)\n", r);
+	dev_err(adev->dev, "Error scheduling IBs (%d)\n", r);
 	return r;
 }
 
-int amdgpu_fill_buffer(struct amdgpu_bo *bo,
-		       uint32_t src_data,
-		       struct dma_resv *resv,
-		       struct dma_fence **fence)
+static int amdgpu_ttm_fill_mem(struct amdgpu_ring *ring, uint32_t src_data,
+			       uint64_t dst_addr, uint32_t byte_count,
+			       struct dma_resv *resv,
+			       struct dma_fence **fence,
+			       bool vm_needs_flush, bool delayed,
+			       u64 k_job_id)
 {
-	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
-	uint32_t max_bytes = adev->mman.buffer_funcs->fill_max_bytes;
-	struct amdgpu_ring *ring = adev->mman.buffer_funcs_ring;
-
-	struct amdgpu_res_cursor cursor;
+	struct amdgpu_device *adev = ring->adev;
 	unsigned int num_loops, num_dw;
-	uint64_t num_bytes;
-
 	struct amdgpu_job *job;
+	uint32_t max_bytes;
+	unsigned int i;
 	int r;
 
-	if (!adev->mman.buffer_funcs_enabled) {
-		DRM_ERROR("Trying to clear memory with ring turned off.\n");
-		return -EINVAL;
+	max_bytes = adev->mman.buffer_funcs->fill_max_bytes;
+	num_loops = DIV_ROUND_UP_ULL(byte_count, max_bytes);
+	num_dw = ALIGN(num_loops * adev->mman.buffer_funcs->fill_num_dw, 8);
+	r = amdgpu_ttm_prepare_job(adev, false, num_dw, resv, vm_needs_flush,
+				   &job, delayed, k_job_id);
+	if (r)
+		return r;
+
+	for (i = 0; i < num_loops; i++) {
+		uint32_t cur_size = min(byte_count, max_bytes);
+
+		amdgpu_emit_fill_buffer(adev, &job->ibs[0], src_data, dst_addr,
+					cur_size);
+
+		dst_addr += cur_size;
+		byte_count -= cur_size;
 	}
 
-	if (bo->tbo.resource->mem_type == AMDGPU_PL_PREEMPT) {
-		DRM_ERROR("Trying to clear preemptible memory.\n");
+	amdgpu_ring_pad_ib(ring, &job->ibs[0]);
+	WARN_ON(job->ibs[0].length_dw > num_dw);
+	*fence = amdgpu_job_submit(job);
+	return 0;
+}
+
+/**
+ * amdgpu_ttm_clear_buffer - clear memory buffers
+ * @bo: amdgpu buffer object
+ * @resv: reservation object
+ * @fence: dma_fence associated with the operation
+ *
+ * Clear the memory buffer resource.
+ *
+ * Returns:
+ * 0 for success or a negative error code on failure.
+ */
+int amdgpu_ttm_clear_buffer(struct amdgpu_bo *bo,
+			    struct dma_resv *resv,
+			    struct dma_fence **fence)
+{
+	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
+	struct amdgpu_ring *ring = adev->mman.buffer_funcs_ring;
+	struct amdgpu_res_cursor cursor;
+	u64 addr;
+	int r = 0;
+
+	if (!adev->mman.buffer_funcs_enabled)
 		return -EINVAL;
-	}
 
-	if (bo->tbo.resource->mem_type == TTM_PL_TT) {
-		r = amdgpu_ttm_alloc_gart(&bo->tbo);
-		if (r)
-			return r;
-	}
+	if (!fence)
+		return -EINVAL;
 
-	num_bytes = bo->tbo.resource->num_pages << PAGE_SHIFT;
-	num_loops = 0;
+	*fence = dma_fence_get_stub();
 
-	amdgpu_res_first(bo->tbo.resource, 0, num_bytes, &cursor);
+	amdgpu_res_first(bo->tbo.resource, 0, amdgpu_bo_size(bo), &cursor);
+
+	mutex_lock(&adev->mman.gtt_window_lock);
 	while (cursor.remaining) {
-		num_loops += DIV_ROUND_UP_ULL(cursor.size, max_bytes);
-		amdgpu_res_next(&cursor, cursor.size);
-	}
-	num_dw = num_loops * adev->mman.buffer_funcs->fill_num_dw;
+		struct dma_fence *next = NULL;
+		u64 size;
 
-	/* for IB padding */
-	num_dw += 64;
+		if (amdgpu_res_cleared(&cursor)) {
+			amdgpu_res_next(&cursor, cursor.size);
+			continue;
+		}
 
-	r = amdgpu_job_alloc_with_ib(adev, num_dw * 4, AMDGPU_IB_POOL_DELAYED,
-				     &job);
-	if (r)
-		return r;
+		/* Never clear more than 256MiB at once to avoid timeouts */
+		size = min(cursor.size, 256ULL << 20);
 
-	if (resv) {
-		r = amdgpu_sync_resv(adev, &job->sync, resv,
-				     AMDGPU_SYNC_ALWAYS,
-				     AMDGPU_FENCE_OWNER_UNDEFINED);
-		if (r) {
-			DRM_ERROR("sync failed (%d).\n", r);
-			goto error_free;
-		}
+		r = amdgpu_ttm_map_buffer(&bo->tbo, bo->tbo.resource, &cursor,
+					  1, ring, false, &size, &addr);
+		if (r)
+			goto err;
+
+		r = amdgpu_ttm_fill_mem(ring, 0, addr, size, resv,
+					&next, true, true,
+					AMDGPU_KERNEL_JOB_ID_TTM_CLEAR_BUFFER);
+		if (r)
+			goto err;
+
+		dma_fence_put(*fence);
+		*fence = next;
+
+		amdgpu_res_next(&cursor, size);
 	}
+err:
+	mutex_unlock(&adev->mman.gtt_window_lock);
 
-	amdgpu_res_first(bo->tbo.resource, 0, num_bytes, &cursor);
-	while (cursor.remaining) {
-		uint32_t cur_size = min_t(uint64_t, cursor.size, max_bytes);
-		uint64_t dst_addr = cursor.start;
+	return r;
+}
 
-		dst_addr += amdgpu_ttm_domain_start(adev,
-						    bo->tbo.resource->mem_type);
-		amdgpu_emit_fill_buffer(adev, &job->ibs[0], src_data, dst_addr,
-					cur_size);
+int amdgpu_fill_buffer(struct amdgpu_bo *bo,
+			uint32_t src_data,
+			struct dma_resv *resv,
+			struct dma_fence **f,
+			bool delayed,
+			u64 k_job_id)
+{
+	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
+	struct amdgpu_ring *ring = adev->mman.buffer_funcs_ring;
+	struct dma_fence *fence = NULL;
+	struct amdgpu_res_cursor dst;
+	int r;
 
-		amdgpu_res_next(&cursor, cur_size);
+	if (!adev->mman.buffer_funcs_enabled) {
+		dev_err(adev->dev,
+			"Trying to clear memory with ring turned off.\n");
+		return -EINVAL;
 	}
 
-	amdgpu_ring_pad_ib(ring, &job->ibs[0]);
-	WARN_ON(job->ibs[0].length_dw > num_dw);
-	r = amdgpu_job_submit(job, &adev->mman.entity,
-			      AMDGPU_FENCE_OWNER_UNDEFINED, fence);
-	if (r)
-		goto error_free;
+	amdgpu_res_first(bo->tbo.resource, 0, amdgpu_bo_size(bo), &dst);
 
-	return 0;
+	mutex_lock(&adev->mman.gtt_window_lock);
+	while (dst.remaining) {
+		struct dma_fence *next;
+		uint64_t cur_size, to;
 
-error_free:
-	amdgpu_job_free(job);
+		/* Never fill more than 256MiB at once to avoid timeouts */
+		cur_size = min(dst.size, 256ULL << 20);
+
+		r = amdgpu_ttm_map_buffer(&bo->tbo, bo->tbo.resource, &dst,
+					  1, ring, false, &cur_size, &to);
+		if (r)
+			goto error;
+
+		r = amdgpu_ttm_fill_mem(ring, src_data, to, cur_size, resv,
+					&next, true, delayed, k_job_id);
+		if (r)
+			goto error;
+
+		dma_fence_put(fence);
+		fence = next;
+
+		amdgpu_res_next(&dst, cur_size);
+	}
+error:
+	mutex_unlock(&adev->mman.gtt_window_lock);
+	if (f)
+		*f = dma_fence_get(fence);
+	dma_fence_put(fence);
 	return r;
 }
 
@@ -2080,7 +2474,7 @@ int amdgpu_ttm_evict_resources(struct amdgpu_device *adev, int mem_type)
 		man = ttm_manager_type(&adev->mman.bdev, mem_type);
 		break;
 	default:
-		DRM_ERROR("Trying to evict invalid memory type\n");
+		dev_err(adev->dev, "Trying to evict invalid memory type\n");
 		return -EINVAL;
 	}
 
@@ -2089,73 +2483,13 @@ int amdgpu_ttm_evict_resources(struct amdgpu_device *adev, int mem_type)
 
 #if defined(CONFIG_DEBUG_FS)
 
-static int amdgpu_mm_vram_table_show(struct seq_file *m, void *unused)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)m->private;
-	struct ttm_resource_manager *man = ttm_manager_type(&adev->mman.bdev,
-							    TTM_PL_VRAM);
-	struct drm_printer p = drm_seq_file_printer(m);
-
-	man->func->debug(man, &p);
-	return 0;
-}
-
 static int amdgpu_ttm_page_pool_show(struct seq_file *m, void *unused)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)m->private;
+	struct amdgpu_device *adev = m->private;
 
 	return ttm_pool_debugfs(&adev->mman.bdev.pool, m);
 }
 
-static int amdgpu_mm_tt_table_show(struct seq_file *m, void *unused)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)m->private;
-	struct ttm_resource_manager *man = ttm_manager_type(&adev->mman.bdev,
-							    TTM_PL_TT);
-	struct drm_printer p = drm_seq_file_printer(m);
-
-	man->func->debug(man, &p);
-	return 0;
-}
-
-static int amdgpu_mm_gds_table_show(struct seq_file *m, void *unused)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)m->private;
-	struct ttm_resource_manager *man = ttm_manager_type(&adev->mman.bdev,
-							    AMDGPU_PL_GDS);
-	struct drm_printer p = drm_seq_file_printer(m);
-
-	man->func->debug(man, &p);
-	return 0;
-}
-
-static int amdgpu_mm_gws_table_show(struct seq_file *m, void *unused)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)m->private;
-	struct ttm_resource_manager *man = ttm_manager_type(&adev->mman.bdev,
-							    AMDGPU_PL_GWS);
-	struct drm_printer p = drm_seq_file_printer(m);
-
-	man->func->debug(man, &p);
-	return 0;
-}
-
-static int amdgpu_mm_oa_table_show(struct seq_file *m, void *unused)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)m->private;
-	struct ttm_resource_manager *man = ttm_manager_type(&adev->mman.bdev,
-							    AMDGPU_PL_OA);
-	struct drm_printer p = drm_seq_file_printer(m);
-
-	man->func->debug(man, &p);
-	return 0;
-}
-
-DEFINE_SHOW_ATTRIBUTE(amdgpu_mm_vram_table);
-DEFINE_SHOW_ATTRIBUTE(amdgpu_mm_tt_table);
-DEFINE_SHOW_ATTRIBUTE(amdgpu_mm_gds_table);
-DEFINE_SHOW_ATTRIBUTE(amdgpu_mm_gws_table);
-DEFINE_SHOW_ATTRIBUTE(amdgpu_mm_oa_table);
 DEFINE_SHOW_ATTRIBUTE(amdgpu_ttm_page_pool);
 
 /*
@@ -2265,7 +2599,7 @@ static ssize_t amdgpu_iomem_read(struct file *f, char __user *buf,
 		struct page *p;
 		void *ptr;
 
-		bytes = bytes < size ? bytes : size;
+		bytes = min(bytes, size);
 
 		/* Translate the bus address to a physical address.  If
 		 * the domain is NULL it means there is no IOMMU active
@@ -2281,9 +2615,9 @@ static ssize_t amdgpu_iomem_read(struct file *f, char __user *buf,
 		if (p->mapping != adev->mman.bdev.dev_mapping)
 			return -EPERM;
 
-		ptr = kmap(p);
+		ptr = kmap_local_page(p);
 		r = copy_to_user(buf, ptr + off, bytes);
-		kunmap(p);
+		kunmap_local(ptr);
 		if (r)
 			return -EFAULT;
 
@@ -2320,7 +2654,7 @@ static ssize_t amdgpu_iomem_write(struct file *f, const char __user *buf,
 		struct page *p;
 		void *ptr;
 
-		bytes = bytes < size ? bytes : size;
+		bytes = min(bytes, size);
 
 		addr = dom ? iommu_iova_to_phys(dom, addr) : addr;
 
@@ -2332,9 +2666,9 @@ static ssize_t amdgpu_iomem_write(struct file *f, const char __user *buf,
 		if (p->mapping != adev->mman.bdev.dev_mapping)
 			return -EPERM;
 
-		ptr = kmap(p);
+		ptr = kmap_local_page(p);
 		r = copy_from_user(ptr + off, buf, bytes);
-		kunmap(p);
+		kunmap_local(ptr);
 		if (r)
 			return -EFAULT;
 
@@ -2365,17 +2699,23 @@ void amdgpu_ttm_debugfs_init(struct amdgpu_device *adev)
 				 &amdgpu_ttm_vram_fops, adev->gmc.mc_vram_size);
 	debugfs_create_file("amdgpu_iomem", 0444, root, adev,
 			    &amdgpu_ttm_iomem_fops);
-	debugfs_create_file("amdgpu_vram_mm", 0444, root, adev,
-			    &amdgpu_mm_vram_table_fops);
-	debugfs_create_file("amdgpu_gtt_mm", 0444, root, adev,
-			    &amdgpu_mm_tt_table_fops);
-	debugfs_create_file("amdgpu_gds_mm", 0444, root, adev,
-			    &amdgpu_mm_gds_table_fops);
-	debugfs_create_file("amdgpu_gws_mm", 0444, root, adev,
-			    &amdgpu_mm_gws_table_fops);
-	debugfs_create_file("amdgpu_oa_mm", 0444, root, adev,
-			    &amdgpu_mm_oa_table_fops);
 	debugfs_create_file("ttm_page_pool", 0444, root, adev,
 			    &amdgpu_ttm_page_pool_fops);
+	ttm_resource_manager_create_debugfs(ttm_manager_type(&adev->mman.bdev,
+							     TTM_PL_VRAM),
+					    root, "amdgpu_vram_mm");
+	ttm_resource_manager_create_debugfs(ttm_manager_type(&adev->mman.bdev,
+							     TTM_PL_TT),
+					    root, "amdgpu_gtt_mm");
+	ttm_resource_manager_create_debugfs(ttm_manager_type(&adev->mman.bdev,
+							     AMDGPU_PL_GDS),
+					    root, "amdgpu_gds_mm");
+	ttm_resource_manager_create_debugfs(ttm_manager_type(&adev->mman.bdev,
+							     AMDGPU_PL_GWS),
+					    root, "amdgpu_gws_mm");
+	ttm_resource_manager_create_debugfs(ttm_manager_type(&adev->mman.bdev,
+							     AMDGPU_PL_OA),
+					    root, "amdgpu_oa_mm");
+
 #endif
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h
index 7346ecff4438..577ee04ce0bf 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h
@@ -26,42 +26,35 @@
 
 #include <linux/dma-direction.h>
 #include <drm/gpu_scheduler.h>
-#include "amdgpu.h"
+#include <drm/ttm/ttm_placement.h>
+#include "amdgpu_vram_mgr.h"
+#include "amdgpu_hmm.h"
 
 #define AMDGPU_PL_GDS		(TTM_PL_PRIV + 0)
 #define AMDGPU_PL_GWS		(TTM_PL_PRIV + 1)
 #define AMDGPU_PL_OA		(TTM_PL_PRIV + 2)
 #define AMDGPU_PL_PREEMPT	(TTM_PL_PRIV + 3)
+#define AMDGPU_PL_DOORBELL	(TTM_PL_PRIV + 4)
+#define AMDGPU_PL_MMIO_REMAP	(TTM_PL_PRIV + 5)
+#define __AMDGPU_PL_NUM	(TTM_PL_PRIV + 6)
 
 #define AMDGPU_GTT_MAX_TRANSFER_SIZE	512
 #define AMDGPU_GTT_NUM_TRANSFER_WINDOWS	2
 
-#define AMDGPU_POISON	0xd0bed0be
+extern const struct attribute_group amdgpu_vram_mgr_attr_group;
+extern const struct attribute_group amdgpu_gtt_mgr_attr_group;
 
-struct amdgpu_vram_mgr {
-	struct ttm_resource_manager manager;
-	struct drm_mm mm;
-	spinlock_t lock;
-	struct list_head reservations_pending;
-	struct list_head reserved_pages;
-	atomic64_t usage;
-	atomic64_t vis_usage;
-};
+struct hmm_range;
 
 struct amdgpu_gtt_mgr {
 	struct ttm_resource_manager manager;
 	struct drm_mm mm;
 	spinlock_t lock;
-	atomic64_t used;
-};
-
-struct amdgpu_preempt_mgr {
-	struct ttm_resource_manager manager;
-	atomic64_t used;
 };
 
 struct amdgpu_mman {
 	struct ttm_device		bdev;
+	struct ttm_pool			*ttm_pools;
 	bool				initialized;
 	void __iomem			*aper_base_kaddr;
 
@@ -71,12 +64,14 @@ struct amdgpu_mman {
 	bool					buffer_funcs_enabled;
 
 	struct mutex				gtt_window_lock;
-	/* Scheduler entity for buffer moves */
-	struct drm_sched_entity			entity;
+	/* High priority scheduler entity for buffer moves */
+	struct drm_sched_entity			high_pr;
+	/* Low priority scheduler entity for VRAM clearing */
+	struct drm_sched_entity			low_pr;
 
 	struct amdgpu_vram_mgr vram_mgr;
 	struct amdgpu_gtt_mgr gtt_mgr;
-	struct amdgpu_preempt_mgr preempt_mgr;
+	struct ttm_resource_manager preempt_mgr;
 
 	uint64_t		stolen_vga_size;
 	struct amdgpu_bo	*stolen_vga_memory;
@@ -88,16 +83,25 @@ struct amdgpu_mman {
 	uint64_t		stolen_reserved_offset;
 	uint64_t		stolen_reserved_size;
 
-	/* discovery */
-	uint8_t				*discovery_bin;
-	uint32_t			discovery_tmr_size;
-	struct amdgpu_bo		*discovery_memory;
+	/* fw reserved memory */
+	struct amdgpu_bo		*fw_reserved_memory;
+	struct amdgpu_bo		*fw_reserved_memory_extend;
 
 	/* firmware VRAM reservation */
 	u64		fw_vram_usage_start_offset;
 	u64		fw_vram_usage_size;
 	struct amdgpu_bo	*fw_vram_usage_reserved_bo;
 	void		*fw_vram_usage_va;
+
+	/* driver VRAM reservation */
+	u64		drv_vram_usage_start_offset;
+	u64		drv_vram_usage_size;
+	struct amdgpu_bo	*drv_vram_usage_reserved_bo;
+	void		*drv_vram_usage_va;
+
+	/* PAGE_SIZE'd BO for process memory r/w over SDMA. */
+	struct amdgpu_bo	*sdma_access_bo;
+	void			*sdma_access_ptr;
 };
 
 struct amdgpu_copy_mem {
@@ -106,6 +110,23 @@ struct amdgpu_copy_mem {
 	unsigned long			offset;
 };
 
+#define AMDGPU_COPY_FLAGS_TMZ		(1 << 0)
+#define AMDGPU_COPY_FLAGS_READ_DECOMPRESSED	(1 << 1)
+#define AMDGPU_COPY_FLAGS_WRITE_COMPRESSED	(1 << 2)
+#define AMDGPU_COPY_FLAGS_MAX_COMPRESSED_SHIFT		3
+#define AMDGPU_COPY_FLAGS_MAX_COMPRESSED_MASK		0x03
+#define AMDGPU_COPY_FLAGS_NUMBER_TYPE_SHIFT		5
+#define AMDGPU_COPY_FLAGS_NUMBER_TYPE_MASK		0x07
+#define AMDGPU_COPY_FLAGS_DATA_FORMAT_SHIFT		8
+#define AMDGPU_COPY_FLAGS_DATA_FORMAT_MASK		0x3f
+#define AMDGPU_COPY_FLAGS_WRITE_COMPRESS_DISABLE_SHIFT	14
+#define AMDGPU_COPY_FLAGS_WRITE_COMPRESS_DISABLE_MASK	0x1
+
+#define AMDGPU_COPY_FLAGS_SET(field, value) \
+	(((__u32)(value) & AMDGPU_COPY_FLAGS_##field##_MASK) << AMDGPU_COPY_FLAGS_##field##_SHIFT)
+#define AMDGPU_COPY_FLAGS_GET(value, field) \
+	(((__u32)(value) >> AMDGPU_COPY_FLAGS_##field##_SHIFT) & AMDGPU_COPY_FLAGS_##field##_MASK)
+
 int amdgpu_gtt_mgr_init(struct amdgpu_device *adev, uint64_t gtt_size);
 void amdgpu_gtt_mgr_fini(struct amdgpu_device *adev);
 int amdgpu_preempt_mgr_init(struct amdgpu_device *adev);
@@ -114,8 +135,7 @@ int amdgpu_vram_mgr_init(struct amdgpu_device *adev);
 void amdgpu_vram_mgr_fini(struct amdgpu_device *adev);
 
 bool amdgpu_gtt_mgr_has_gart_addr(struct ttm_resource *mem);
-uint64_t amdgpu_gtt_mgr_usage(struct ttm_resource_manager *man);
-int amdgpu_gtt_mgr_recover(struct ttm_resource_manager *man);
+void amdgpu_gtt_mgr_recover(struct amdgpu_gtt_mgr *mgr);
 
 uint64_t amdgpu_preempt_mgr_usage(struct ttm_resource_manager *man);
 
@@ -129,54 +149,53 @@ int amdgpu_vram_mgr_alloc_sgt(struct amdgpu_device *adev,
 void amdgpu_vram_mgr_free_sgt(struct device *dev,
 			      enum dma_data_direction dir,
 			      struct sg_table *sgt);
-uint64_t amdgpu_vram_mgr_usage(struct ttm_resource_manager *man);
-uint64_t amdgpu_vram_mgr_vis_usage(struct ttm_resource_manager *man);
-int amdgpu_vram_mgr_reserve_range(struct ttm_resource_manager *man,
+uint64_t amdgpu_vram_mgr_vis_usage(struct amdgpu_vram_mgr *mgr);
+int amdgpu_vram_mgr_reserve_range(struct amdgpu_vram_mgr *mgr,
 				  uint64_t start, uint64_t size);
-int amdgpu_vram_mgr_query_page_status(struct ttm_resource_manager *man,
+int amdgpu_vram_mgr_query_page_status(struct amdgpu_vram_mgr *mgr,
 				      uint64_t start);
+void amdgpu_vram_mgr_clear_reset_blocks(struct amdgpu_device *adev);
+
+bool amdgpu_res_cpu_visible(struct amdgpu_device *adev,
+			    struct ttm_resource *res);
 
 int amdgpu_ttm_init(struct amdgpu_device *adev);
 void amdgpu_ttm_fini(struct amdgpu_device *adev);
 void amdgpu_ttm_set_buffer_funcs_status(struct amdgpu_device *adev,
 					bool enable);
-
 int amdgpu_copy_buffer(struct amdgpu_ring *ring, uint64_t src_offset,
 		       uint64_t dst_offset, uint32_t byte_count,
 		       struct dma_resv *resv,
 		       struct dma_fence **fence, bool direct_submit,
-		       bool vm_needs_flush, bool tmz);
-int amdgpu_ttm_copy_mem_to_mem(struct amdgpu_device *adev,
-			       const struct amdgpu_copy_mem *src,
-			       const struct amdgpu_copy_mem *dst,
-			       uint64_t size, bool tmz,
-			       struct dma_resv *resv,
-			       struct dma_fence **f);
+		       bool vm_needs_flush, uint32_t copy_flags);
+int amdgpu_ttm_clear_buffer(struct amdgpu_bo *bo,
+			    struct dma_resv *resv,
+			    struct dma_fence **fence);
 int amdgpu_fill_buffer(struct amdgpu_bo *bo,
 			uint32_t src_data,
 			struct dma_resv *resv,
-			struct dma_fence **fence);
+			struct dma_fence **fence,
+			bool delayed,
+			u64 k_job_id);
 
 int amdgpu_ttm_alloc_gart(struct ttm_buffer_object *bo);
-int amdgpu_ttm_recover_gart(struct ttm_buffer_object *tbo);
+void amdgpu_ttm_recover_gart(struct ttm_buffer_object *tbo);
 uint64_t amdgpu_ttm_domain_start(struct amdgpu_device *adev, uint32_t type);
 
 #if IS_ENABLED(CONFIG_DRM_AMDGPU_USERPTR)
-int amdgpu_ttm_tt_get_user_pages(struct amdgpu_bo *bo, struct page **pages);
-bool amdgpu_ttm_tt_get_user_pages_done(struct ttm_tt *ttm);
+int amdgpu_ttm_tt_get_user_pages(struct amdgpu_bo *bo,
+				 struct amdgpu_hmm_range *range);
 #else
 static inline int amdgpu_ttm_tt_get_user_pages(struct amdgpu_bo *bo,
-					       struct page **pages)
+					       struct amdgpu_hmm_range *range)
 {
 	return -EPERM;
 }
-static inline bool amdgpu_ttm_tt_get_user_pages_done(struct ttm_tt *ttm)
-{
-	return false;
-}
 #endif
 
-void amdgpu_ttm_tt_set_user_pages(struct ttm_tt *ttm, struct page **pages);
+void amdgpu_ttm_tt_set_user_pages(struct ttm_tt *ttm, struct amdgpu_hmm_range *range);
+int amdgpu_ttm_tt_get_userptr(const struct ttm_buffer_object *tbo,
+			      uint64_t *user_addr);
 int amdgpu_ttm_tt_set_userptr(struct ttm_buffer_object *bo,
 			      uint64_t addr, uint32_t flags);
 bool amdgpu_ttm_tt_has_userptr(struct ttm_tt *ttm);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c
index ca3350502618..e96f24e9ad57 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c
@@ -28,6 +28,13 @@
 #include "amdgpu.h"
 #include "amdgpu_ucode.h"
 
+#define AMDGPU_UCODE_NAME_MAX		(128)
+
+static const struct kicker_device kicker_device_list[] = {
+	{0x744B, 0x00},
+	{0x7551, 0xC8}
+};
+
 static void amdgpu_ucode_print_common_hdr(const struct common_firmware_header *hdr)
 {
 	DRM_DEBUG("size_bytes: %u\n", le32_to_cpu(hdr->size_bytes));
@@ -115,6 +122,12 @@ void amdgpu_ucode_print_gfx_hdr(const struct common_firmware_header *hdr)
 			  le32_to_cpu(gfx_hdr->ucode_feature_version));
 		DRM_DEBUG("jt_offset: %u\n", le32_to_cpu(gfx_hdr->jt_offset));
 		DRM_DEBUG("jt_size: %u\n", le32_to_cpu(gfx_hdr->jt_size));
+	} else if (version_major == 2) {
+		const struct gfx_firmware_header_v2_0 *gfx_hdr =
+			container_of(hdr, struct gfx_firmware_header_v2_0, header);
+
+		DRM_DEBUG("ucode_feature_version: %u\n",
+			  le32_to_cpu(gfx_hdr->ucode_feature_version));
 	} else {
 		DRM_ERROR("Unknown GFX ucode version: %u.%u\n", version_major, version_minor);
 	}
@@ -145,70 +158,138 @@ void amdgpu_ucode_print_rlc_hdr(const struct common_firmware_header *hdr)
 	} else if (version_major == 2) {
 		const struct rlc_firmware_header_v2_0 *rlc_hdr =
 			container_of(hdr, struct rlc_firmware_header_v2_0, header);
+		const struct rlc_firmware_header_v2_1 *rlc_hdr_v2_1 =
+			container_of(rlc_hdr, struct rlc_firmware_header_v2_1, v2_0);
+		const struct rlc_firmware_header_v2_2 *rlc_hdr_v2_2 =
+			container_of(rlc_hdr_v2_1, struct rlc_firmware_header_v2_2, v2_1);
+		const struct rlc_firmware_header_v2_3 *rlc_hdr_v2_3 =
+			container_of(rlc_hdr_v2_2, struct rlc_firmware_header_v2_3, v2_2);
+		const struct rlc_firmware_header_v2_4 *rlc_hdr_v2_4 =
+			container_of(rlc_hdr_v2_3, struct rlc_firmware_header_v2_4, v2_3);
 
-		DRM_DEBUG("ucode_feature_version: %u\n",
-			  le32_to_cpu(rlc_hdr->ucode_feature_version));
-		DRM_DEBUG("jt_offset: %u\n", le32_to_cpu(rlc_hdr->jt_offset));
-		DRM_DEBUG("jt_size: %u\n", le32_to_cpu(rlc_hdr->jt_size));
-		DRM_DEBUG("save_and_restore_offset: %u\n",
-			  le32_to_cpu(rlc_hdr->save_and_restore_offset));
-		DRM_DEBUG("clear_state_descriptor_offset: %u\n",
-			  le32_to_cpu(rlc_hdr->clear_state_descriptor_offset));
-		DRM_DEBUG("avail_scratch_ram_locations: %u\n",
-			  le32_to_cpu(rlc_hdr->avail_scratch_ram_locations));
-		DRM_DEBUG("reg_restore_list_size: %u\n",
-			  le32_to_cpu(rlc_hdr->reg_restore_list_size));
-		DRM_DEBUG("reg_list_format_start: %u\n",
-			  le32_to_cpu(rlc_hdr->reg_list_format_start));
-		DRM_DEBUG("reg_list_format_separate_start: %u\n",
-			  le32_to_cpu(rlc_hdr->reg_list_format_separate_start));
-		DRM_DEBUG("starting_offsets_start: %u\n",
-			  le32_to_cpu(rlc_hdr->starting_offsets_start));
-		DRM_DEBUG("reg_list_format_size_bytes: %u\n",
-			  le32_to_cpu(rlc_hdr->reg_list_format_size_bytes));
-		DRM_DEBUG("reg_list_format_array_offset_bytes: %u\n",
-			  le32_to_cpu(rlc_hdr->reg_list_format_array_offset_bytes));
-		DRM_DEBUG("reg_list_size_bytes: %u\n",
-			  le32_to_cpu(rlc_hdr->reg_list_size_bytes));
-		DRM_DEBUG("reg_list_array_offset_bytes: %u\n",
-			  le32_to_cpu(rlc_hdr->reg_list_array_offset_bytes));
-		DRM_DEBUG("reg_list_format_separate_size_bytes: %u\n",
-			  le32_to_cpu(rlc_hdr->reg_list_format_separate_size_bytes));
-		DRM_DEBUG("reg_list_format_separate_array_offset_bytes: %u\n",
-			  le32_to_cpu(rlc_hdr->reg_list_format_separate_array_offset_bytes));
-		DRM_DEBUG("reg_list_separate_size_bytes: %u\n",
-			  le32_to_cpu(rlc_hdr->reg_list_separate_size_bytes));
-		DRM_DEBUG("reg_list_separate_array_offset_bytes: %u\n",
-			  le32_to_cpu(rlc_hdr->reg_list_separate_array_offset_bytes));
-		if (version_minor == 1) {
-			const struct rlc_firmware_header_v2_1 *v2_1 =
-				container_of(rlc_hdr, struct rlc_firmware_header_v2_1, v2_0);
+		switch (version_minor) {
+		case 0:
+			/* rlc_hdr v2_0 */
+			DRM_DEBUG("ucode_feature_version: %u\n",
+				  le32_to_cpu(rlc_hdr->ucode_feature_version));
+			DRM_DEBUG("jt_offset: %u\n", le32_to_cpu(rlc_hdr->jt_offset));
+			DRM_DEBUG("jt_size: %u\n", le32_to_cpu(rlc_hdr->jt_size));
+			DRM_DEBUG("save_and_restore_offset: %u\n",
+				  le32_to_cpu(rlc_hdr->save_and_restore_offset));
+			DRM_DEBUG("clear_state_descriptor_offset: %u\n",
+				  le32_to_cpu(rlc_hdr->clear_state_descriptor_offset));
+			DRM_DEBUG("avail_scratch_ram_locations: %u\n",
+				  le32_to_cpu(rlc_hdr->avail_scratch_ram_locations));
+			DRM_DEBUG("reg_restore_list_size: %u\n",
+				  le32_to_cpu(rlc_hdr->reg_restore_list_size));
+			DRM_DEBUG("reg_list_format_start: %u\n",
+				  le32_to_cpu(rlc_hdr->reg_list_format_start));
+			DRM_DEBUG("reg_list_format_separate_start: %u\n",
+				  le32_to_cpu(rlc_hdr->reg_list_format_separate_start));
+			DRM_DEBUG("starting_offsets_start: %u\n",
+				  le32_to_cpu(rlc_hdr->starting_offsets_start));
+			DRM_DEBUG("reg_list_format_size_bytes: %u\n",
+				  le32_to_cpu(rlc_hdr->reg_list_format_size_bytes));
+			DRM_DEBUG("reg_list_format_array_offset_bytes: %u\n",
+				  le32_to_cpu(rlc_hdr->reg_list_format_array_offset_bytes));
+			DRM_DEBUG("reg_list_size_bytes: %u\n",
+				  le32_to_cpu(rlc_hdr->reg_list_size_bytes));
+			DRM_DEBUG("reg_list_array_offset_bytes: %u\n",
+				  le32_to_cpu(rlc_hdr->reg_list_array_offset_bytes));
+			DRM_DEBUG("reg_list_format_separate_size_bytes: %u\n",
+				  le32_to_cpu(rlc_hdr->reg_list_format_separate_size_bytes));
+			DRM_DEBUG("reg_list_format_separate_array_offset_bytes: %u\n",
+				  le32_to_cpu(rlc_hdr->reg_list_format_separate_array_offset_bytes));
+			DRM_DEBUG("reg_list_separate_size_bytes: %u\n",
+				  le32_to_cpu(rlc_hdr->reg_list_separate_size_bytes));
+			DRM_DEBUG("reg_list_separate_array_offset_bytes: %u\n",
+				  le32_to_cpu(rlc_hdr->reg_list_separate_array_offset_bytes));
+			break;
+		case 1:
+			/* rlc_hdr v2_1 */
 			DRM_DEBUG("reg_list_format_direct_reg_list_length: %u\n",
-				  le32_to_cpu(v2_1->reg_list_format_direct_reg_list_length));
+				  le32_to_cpu(rlc_hdr_v2_1->reg_list_format_direct_reg_list_length));
 			DRM_DEBUG("save_restore_list_cntl_ucode_ver: %u\n",
-				  le32_to_cpu(v2_1->save_restore_list_cntl_ucode_ver));
+				  le32_to_cpu(rlc_hdr_v2_1->save_restore_list_cntl_ucode_ver));
 			DRM_DEBUG("save_restore_list_cntl_feature_ver: %u\n",
-				  le32_to_cpu(v2_1->save_restore_list_cntl_feature_ver));
+				  le32_to_cpu(rlc_hdr_v2_1->save_restore_list_cntl_feature_ver));
 			DRM_DEBUG("save_restore_list_cntl_size_bytes %u\n",
-				  le32_to_cpu(v2_1->save_restore_list_cntl_size_bytes));
+				  le32_to_cpu(rlc_hdr_v2_1->save_restore_list_cntl_size_bytes));
 			DRM_DEBUG("save_restore_list_cntl_offset_bytes: %u\n",
-				  le32_to_cpu(v2_1->save_restore_list_cntl_offset_bytes));
+				  le32_to_cpu(rlc_hdr_v2_1->save_restore_list_cntl_offset_bytes));
 			DRM_DEBUG("save_restore_list_gpm_ucode_ver: %u\n",
-				  le32_to_cpu(v2_1->save_restore_list_gpm_ucode_ver));
+				  le32_to_cpu(rlc_hdr_v2_1->save_restore_list_gpm_ucode_ver));
 			DRM_DEBUG("save_restore_list_gpm_feature_ver: %u\n",
-				  le32_to_cpu(v2_1->save_restore_list_gpm_feature_ver));
+				  le32_to_cpu(rlc_hdr_v2_1->save_restore_list_gpm_feature_ver));
 			DRM_DEBUG("save_restore_list_gpm_size_bytes %u\n",
-				  le32_to_cpu(v2_1->save_restore_list_gpm_size_bytes));
+				  le32_to_cpu(rlc_hdr_v2_1->save_restore_list_gpm_size_bytes));
 			DRM_DEBUG("save_restore_list_gpm_offset_bytes: %u\n",
-				  le32_to_cpu(v2_1->save_restore_list_gpm_offset_bytes));
+				  le32_to_cpu(rlc_hdr_v2_1->save_restore_list_gpm_offset_bytes));
 			DRM_DEBUG("save_restore_list_srm_ucode_ver: %u\n",
-				  le32_to_cpu(v2_1->save_restore_list_srm_ucode_ver));
+				  le32_to_cpu(rlc_hdr_v2_1->save_restore_list_srm_ucode_ver));
 			DRM_DEBUG("save_restore_list_srm_feature_ver: %u\n",
-				  le32_to_cpu(v2_1->save_restore_list_srm_feature_ver));
+				  le32_to_cpu(rlc_hdr_v2_1->save_restore_list_srm_feature_ver));
 			DRM_DEBUG("save_restore_list_srm_size_bytes %u\n",
-				  le32_to_cpu(v2_1->save_restore_list_srm_size_bytes));
+				  le32_to_cpu(rlc_hdr_v2_1->save_restore_list_srm_size_bytes));
 			DRM_DEBUG("save_restore_list_srm_offset_bytes: %u\n",
-				  le32_to_cpu(v2_1->save_restore_list_srm_offset_bytes));
+				  le32_to_cpu(rlc_hdr_v2_1->save_restore_list_srm_offset_bytes));
+			break;
+		case 2:
+			/* rlc_hdr v2_2 */
+			DRM_DEBUG("rlc_iram_ucode_size_bytes: %u\n",
+				  le32_to_cpu(rlc_hdr_v2_2->rlc_iram_ucode_size_bytes));
+			DRM_DEBUG("rlc_iram_ucode_offset_bytes: %u\n",
+				  le32_to_cpu(rlc_hdr_v2_2->rlc_iram_ucode_offset_bytes));
+			DRM_DEBUG("rlc_dram_ucode_size_bytes: %u\n",
+				  le32_to_cpu(rlc_hdr_v2_2->rlc_dram_ucode_size_bytes));
+			DRM_DEBUG("rlc_dram_ucode_offset_bytes: %u\n",
+				  le32_to_cpu(rlc_hdr_v2_2->rlc_dram_ucode_offset_bytes));
+			break;
+		case 3:
+			/* rlc_hdr v2_3 */
+			DRM_DEBUG("rlcp_ucode_version: %u\n",
+				  le32_to_cpu(rlc_hdr_v2_3->rlcp_ucode_version));
+			DRM_DEBUG("rlcp_ucode_feature_version: %u\n",
+				  le32_to_cpu(rlc_hdr_v2_3->rlcp_ucode_feature_version));
+			DRM_DEBUG("rlcp_ucode_size_bytes: %u\n",
+				  le32_to_cpu(rlc_hdr_v2_3->rlcp_ucode_size_bytes));
+			DRM_DEBUG("rlcp_ucode_offset_bytes: %u\n",
+				  le32_to_cpu(rlc_hdr_v2_3->rlcp_ucode_offset_bytes));
+			DRM_DEBUG("rlcv_ucode_version: %u\n",
+				  le32_to_cpu(rlc_hdr_v2_3->rlcv_ucode_version));
+			DRM_DEBUG("rlcv_ucode_feature_version: %u\n",
+				  le32_to_cpu(rlc_hdr_v2_3->rlcv_ucode_feature_version));
+			DRM_DEBUG("rlcv_ucode_size_bytes: %u\n",
+				  le32_to_cpu(rlc_hdr_v2_3->rlcv_ucode_size_bytes));
+			DRM_DEBUG("rlcv_ucode_offset_bytes: %u\n",
+				  le32_to_cpu(rlc_hdr_v2_3->rlcv_ucode_offset_bytes));
+			break;
+		case 4:
+			/* rlc_hdr v2_4 */
+			DRM_DEBUG("global_tap_delays_ucode_size_bytes :%u\n",
+				  le32_to_cpu(rlc_hdr_v2_4->global_tap_delays_ucode_size_bytes));
+			DRM_DEBUG("global_tap_delays_ucode_offset_bytes: %u\n",
+				  le32_to_cpu(rlc_hdr_v2_4->global_tap_delays_ucode_offset_bytes));
+			DRM_DEBUG("se0_tap_delays_ucode_size_bytes :%u\n",
+				  le32_to_cpu(rlc_hdr_v2_4->se0_tap_delays_ucode_size_bytes));
+			DRM_DEBUG("se0_tap_delays_ucode_offset_bytes: %u\n",
+				  le32_to_cpu(rlc_hdr_v2_4->se0_tap_delays_ucode_offset_bytes));
+			DRM_DEBUG("se1_tap_delays_ucode_size_bytes :%u\n",
+				  le32_to_cpu(rlc_hdr_v2_4->se1_tap_delays_ucode_size_bytes));
+			DRM_DEBUG("se1_tap_delays_ucode_offset_bytes: %u\n",
+				  le32_to_cpu(rlc_hdr_v2_4->se1_tap_delays_ucode_offset_bytes));
+			DRM_DEBUG("se2_tap_delays_ucode_size_bytes :%u\n",
+				  le32_to_cpu(rlc_hdr_v2_4->se2_tap_delays_ucode_size_bytes));
+			DRM_DEBUG("se2_tap_delays_ucode_offset_bytes: %u\n",
+				  le32_to_cpu(rlc_hdr_v2_4->se2_tap_delays_ucode_offset_bytes));
+			DRM_DEBUG("se3_tap_delays_ucode_size_bytes :%u\n",
+				  le32_to_cpu(rlc_hdr_v2_4->se3_tap_delays_ucode_size_bytes));
+			DRM_DEBUG("se3_tap_delays_ucode_offset_bytes: %u\n",
+				  le32_to_cpu(rlc_hdr_v2_4->se3_tap_delays_ucode_offset_bytes));
+			break;
+		default:
+			DRM_ERROR("Unknown RLC v2 ucode: v2.%u\n", version_minor);
+			break;
 		}
 	} else {
 		DRM_ERROR("Unknown RLC ucode version: %u.%u\n", version_major, version_minor);
@@ -238,6 +319,23 @@ void amdgpu_ucode_print_sdma_hdr(const struct common_firmware_header *hdr)
 				container_of(sdma_hdr, struct sdma_firmware_header_v1_1, v1_0);
 			DRM_DEBUG("digest_size: %u\n", le32_to_cpu(sdma_v1_1_hdr->digest_size));
 		}
+	} else if (version_major == 2) {
+		const struct sdma_firmware_header_v2_0 *sdma_hdr =
+			container_of(hdr, struct sdma_firmware_header_v2_0, header);
+
+		DRM_DEBUG("ucode_feature_version: %u\n",
+			  le32_to_cpu(sdma_hdr->ucode_feature_version));
+		DRM_DEBUG("ctx_jt_offset: %u\n", le32_to_cpu(sdma_hdr->ctx_jt_offset));
+		DRM_DEBUG("ctx_jt_size: %u\n", le32_to_cpu(sdma_hdr->ctx_jt_size));
+		DRM_DEBUG("ctl_ucode_offset: %u\n", le32_to_cpu(sdma_hdr->ctl_ucode_offset));
+		DRM_DEBUG("ctl_jt_offset: %u\n", le32_to_cpu(sdma_hdr->ctl_jt_offset));
+		DRM_DEBUG("ctl_jt_size: %u\n", le32_to_cpu(sdma_hdr->ctl_jt_size));
+	} else if (version_major == 3) {
+		const struct sdma_firmware_header_v3_0 *sdma_hdr =
+			container_of(hdr, struct sdma_firmware_header_v3_0, header);
+
+		DRM_DEBUG("ucode_reversion: %u\n",
+			  le32_to_cpu(sdma_hdr->ucode_feature_version));
 	} else {
 		DRM_ERROR("Unknown SDMA ucode version: %u.%u\n",
 			  version_major, version_minor);
@@ -248,6 +346,8 @@ void amdgpu_ucode_print_psp_hdr(const struct common_firmware_header *hdr)
 {
 	uint16_t version_major = le16_to_cpu(hdr->header_version_major);
 	uint16_t version_minor = le16_to_cpu(hdr->header_version_minor);
+	uint32_t fw_index;
+	const struct psp_fw_bin_desc *desc;
 
 	DRM_DEBUG("PSP\n");
 	amdgpu_ucode_print_common_hdr(hdr);
@@ -312,6 +412,77 @@ void amdgpu_ucode_print_psp_hdr(const struct common_firmware_header *hdr)
 			DRM_DEBUG("spl_size_bytes: %u\n",
 				  le32_to_cpu(psp_hdr_v1_3->spl.size_bytes));
 		}
+	} else if (version_major == 2) {
+		const struct psp_firmware_header_v2_0 *psp_hdr_v2_0 =
+			 container_of(hdr, struct psp_firmware_header_v2_0, header);
+		for (fw_index = 0; fw_index < le32_to_cpu(psp_hdr_v2_0->psp_fw_bin_count); fw_index++) {
+			desc = &(psp_hdr_v2_0->psp_fw_bin[fw_index]);
+			switch (desc->fw_type) {
+			case PSP_FW_TYPE_PSP_SOS:
+				DRM_DEBUG("psp_sos_version: %u\n",
+					  le32_to_cpu(desc->fw_version));
+				DRM_DEBUG("psp_sos_size_bytes: %u\n",
+					  le32_to_cpu(desc->size_bytes));
+				break;
+			case PSP_FW_TYPE_PSP_SYS_DRV:
+				DRM_DEBUG("psp_sys_drv_version: %u\n",
+					  le32_to_cpu(desc->fw_version));
+				DRM_DEBUG("psp_sys_drv_size_bytes: %u\n",
+					  le32_to_cpu(desc->size_bytes));
+				break;
+			case PSP_FW_TYPE_PSP_KDB:
+				DRM_DEBUG("psp_kdb_version: %u\n",
+					  le32_to_cpu(desc->fw_version));
+				DRM_DEBUG("psp_kdb_size_bytes: %u\n",
+					  le32_to_cpu(desc->size_bytes));
+				break;
+			case PSP_FW_TYPE_PSP_TOC:
+				DRM_DEBUG("psp_toc_version: %u\n",
+					  le32_to_cpu(desc->fw_version));
+				DRM_DEBUG("psp_toc_size_bytes: %u\n",
+					  le32_to_cpu(desc->size_bytes));
+				break;
+			case PSP_FW_TYPE_PSP_SPL:
+				DRM_DEBUG("psp_spl_version: %u\n",
+					  le32_to_cpu(desc->fw_version));
+				DRM_DEBUG("psp_spl_size_bytes: %u\n",
+					  le32_to_cpu(desc->size_bytes));
+				break;
+			case PSP_FW_TYPE_PSP_RL:
+				DRM_DEBUG("psp_rl_version: %u\n",
+					  le32_to_cpu(desc->fw_version));
+				DRM_DEBUG("psp_rl_size_bytes: %u\n",
+					  le32_to_cpu(desc->size_bytes));
+				break;
+			case PSP_FW_TYPE_PSP_SOC_DRV:
+				DRM_DEBUG("psp_soc_drv_version: %u\n",
+					  le32_to_cpu(desc->fw_version));
+				DRM_DEBUG("psp_soc_drv_size_bytes: %u\n",
+					  le32_to_cpu(desc->size_bytes));
+				break;
+			case PSP_FW_TYPE_PSP_INTF_DRV:
+				DRM_DEBUG("psp_intf_drv_version: %u\n",
+					  le32_to_cpu(desc->fw_version));
+				DRM_DEBUG("psp_intf_drv_size_bytes: %u\n",
+					  le32_to_cpu(desc->size_bytes));
+				break;
+			case PSP_FW_TYPE_PSP_DBG_DRV:
+				DRM_DEBUG("psp_dbg_drv_version: %u\n",
+					  le32_to_cpu(desc->fw_version));
+				DRM_DEBUG("psp_dbg_drv_size_bytes: %u\n",
+					  le32_to_cpu(desc->size_bytes));
+				break;
+			case PSP_FW_TYPE_PSP_RAS_DRV:
+				DRM_DEBUG("psp_ras_drv_version: %u\n",
+					  le32_to_cpu(desc->fw_version));
+				DRM_DEBUG("psp_ras_drv_size_bytes: %u\n",
+					  le32_to_cpu(desc->size_bytes));
+				break;
+			default:
+				DRM_DEBUG("Unsupported PSP fw type: %d\n", desc->fw_type);
+				break;
+			}
+		}
 	} else {
 		DRM_ERROR("Unknown PSP ucode version: %u.%u\n",
 			  version_major, version_minor);
@@ -339,7 +510,7 @@ void amdgpu_ucode_print_gpu_info_hdr(const struct common_firmware_header *hdr)
 	}
 }
 
-int amdgpu_ucode_validate(const struct firmware *fw)
+static int amdgpu_ucode_validate(const struct firmware *fw)
 {
 	const struct common_firmware_header *hdr =
 		(const struct common_firmware_header *)fw->data;
@@ -355,8 +526,8 @@ bool amdgpu_ucode_hdr_version(union amdgpu_firmware_header *hdr,
 {
 	if ((hdr->common.header_version_major == hdr_major) &&
 		(hdr->common.header_version_minor == hdr_minor))
-		return false;
-	return true;
+		return true;
+	return false;
 }
 
 enum amdgpu_firmware_load_type
@@ -389,26 +560,6 @@ amdgpu_ucode_get_load_type(struct amdgpu_device *adev, int load_type)
 	case CHIP_POLARIS12:
 	case CHIP_VEGAM:
 		return AMDGPU_FW_LOAD_SMU;
-	case CHIP_VEGA10:
-	case CHIP_RAVEN:
-	case CHIP_VEGA12:
-	case CHIP_VEGA20:
-	case CHIP_ARCTURUS:
-	case CHIP_RENOIR:
-	case CHIP_NAVI10:
-	case CHIP_NAVI14:
-	case CHIP_NAVI12:
-	case CHIP_SIENNA_CICHLID:
-	case CHIP_NAVY_FLOUNDER:
-	case CHIP_VANGOGH:
-	case CHIP_DIMGREY_CAVEFISH:
-	case CHIP_ALDEBARAN:
-	case CHIP_BEIGE_GOBY:
-	case CHIP_YELLOW_CARP:
-		if (!load_type)
-			return AMDGPU_FW_LOAD_DIRECT;
-		else
-			return AMDGPU_FW_LOAD_PSP;
 	case CHIP_CYAN_SKILLFISH:
 		if (!(load_type &&
 		      adev->apu_flags & AMD_APU_IS_CYAN_SKILLFISH2))
@@ -418,6 +569,8 @@ amdgpu_ucode_get_load_type(struct amdgpu_device *adev, int load_type)
 	default:
 		if (!load_type)
 			return AMDGPU_FW_LOAD_DIRECT;
+		else if (load_type == 3)
+			return AMDGPU_FW_LOAD_RLC_BACKDOOR_AUTO;
 		else
 			return AMDGPU_FW_LOAD_PSP;
 	}
@@ -442,6 +595,10 @@ const char *amdgpu_ucode_name(enum AMDGPU_UCODE_ID ucode_id)
 		return "SDMA6";
 	case AMDGPU_UCODE_ID_SDMA7:
 		return "SDMA7";
+	case AMDGPU_UCODE_ID_SDMA_UCODE_TH0:
+		return "SDMA_CTX";
+	case AMDGPU_UCODE_ID_SDMA_UCODE_TH1:
+		return "SDMA_CTL";
 	case AMDGPU_UCODE_ID_CP_CE:
 		return "CP_CE";
 	case AMDGPU_UCODE_ID_CP_PFP:
@@ -460,6 +617,10 @@ const char *amdgpu_ucode_name(enum AMDGPU_UCODE_ID ucode_id)
 		return "CP_MES";
 	case AMDGPU_UCODE_ID_CP_MES_DATA:
 		return "CP_MES_DATA";
+	case AMDGPU_UCODE_ID_CP_MES1:
+		return "CP_MES_KIQ";
+	case AMDGPU_UCODE_ID_CP_MES1_DATA:
+		return "CP_MES_KIQ_DATA";
 	case AMDGPU_UCODE_ID_RLC_RESTORE_LIST_CNTL:
 		return "RLC_RESTORE_LIST_CNTL";
 	case AMDGPU_UCODE_ID_RLC_RESTORE_LIST_GPM_MEM:
@@ -472,10 +633,32 @@ const char *amdgpu_ucode_name(enum AMDGPU_UCODE_ID ucode_id)
 		return "RLC_DRAM";
 	case AMDGPU_UCODE_ID_RLC_G:
 		return "RLC_G";
+	case AMDGPU_UCODE_ID_RLC_P:
+		return "RLC_P";
+	case AMDGPU_UCODE_ID_RLC_V:
+		return "RLC_V";
+	case AMDGPU_UCODE_ID_GLOBAL_TAP_DELAYS:
+		return "GLOBAL_TAP_DELAYS";
+	case AMDGPU_UCODE_ID_SE0_TAP_DELAYS:
+		return "SE0_TAP_DELAYS";
+	case AMDGPU_UCODE_ID_SE1_TAP_DELAYS:
+		return "SE1_TAP_DELAYS";
+	case AMDGPU_UCODE_ID_SE2_TAP_DELAYS:
+		return "SE2_TAP_DELAYS";
+	case AMDGPU_UCODE_ID_SE3_TAP_DELAYS:
+		return "SE3_TAP_DELAYS";
+	case AMDGPU_UCODE_ID_IMU_I:
+		return "IMU_I";
+	case AMDGPU_UCODE_ID_IMU_D:
+		return "IMU_D";
 	case AMDGPU_UCODE_ID_STORAGE:
 		return "STORAGE";
 	case AMDGPU_UCODE_ID_SMC:
 		return "SMC";
+	case AMDGPU_UCODE_ID_PPTABLE:
+		return "PPTABLE";
+	case AMDGPU_UCODE_ID_P2S_TABLE:
+		return "P2STABLE";
 	case AMDGPU_UCODE_ID_UVD:
 		return "UVD";
 	case AMDGPU_UCODE_ID_UVD1:
@@ -496,20 +679,72 @@ const char *amdgpu_ucode_name(enum AMDGPU_UCODE_ID ucode_id)
 		return "VCN1_RAM";
 	case AMDGPU_UCODE_ID_DMCUB:
 		return "DMCUB";
+	case AMDGPU_UCODE_ID_CAP:
+		return "CAP";
+	case AMDGPU_UCODE_ID_VPE_CTX:
+		return "VPE_CTX";
+	case AMDGPU_UCODE_ID_VPE_CTL:
+		return "VPE_CTL";
+	case AMDGPU_UCODE_ID_VPE:
+		return "VPE";
+	case AMDGPU_UCODE_ID_UMSCH_MM_UCODE:
+		return "UMSCH_MM_UCODE";
+	case AMDGPU_UCODE_ID_UMSCH_MM_DATA:
+		return "UMSCH_MM_DATA";
+	case AMDGPU_UCODE_ID_UMSCH_MM_CMD_BUFFER:
+		return "UMSCH_MM_CMD_BUFFER";
+	case AMDGPU_UCODE_ID_JPEG_RAM:
+		return "JPEG";
+	case AMDGPU_UCODE_ID_SDMA_RS64:
+		return "RS64_SDMA";
+	case AMDGPU_UCODE_ID_CP_RS64_PFP:
+		return "RS64_PFP";
+	case AMDGPU_UCODE_ID_CP_RS64_ME:
+		return "RS64_ME";
+	case AMDGPU_UCODE_ID_CP_RS64_MEC:
+		return "RS64_MEC";
+	case AMDGPU_UCODE_ID_CP_RS64_PFP_P0_STACK:
+		return "RS64_PFP_P0_STACK";
+	case AMDGPU_UCODE_ID_CP_RS64_PFP_P1_STACK:
+		return "RS64_PFP_P1_STACK";
+	case AMDGPU_UCODE_ID_CP_RS64_ME_P0_STACK:
+		return "RS64_ME_P0_STACK";
+	case AMDGPU_UCODE_ID_CP_RS64_ME_P1_STACK:
+		return "RS64_ME_P1_STACK";
+	case AMDGPU_UCODE_ID_CP_RS64_MEC_P0_STACK:
+		return "RS64_MEC_P0_STACK";
+	case AMDGPU_UCODE_ID_CP_RS64_MEC_P1_STACK:
+		return "RS64_MEC_P1_STACK";
+	case AMDGPU_UCODE_ID_CP_RS64_MEC_P2_STACK:
+		return "RS64_MEC_P2_STACK";
+	case AMDGPU_UCODE_ID_CP_RS64_MEC_P3_STACK:
+		return "RS64_MEC_P3_STACK";
+	case AMDGPU_UCODE_ID_ISP:
+		return "ISP";
 	default:
 		return "UNKNOWN UCODE";
 	}
 }
 
+static inline int amdgpu_ucode_is_valid(uint32_t fw_version)
+{
+	if (!fw_version)
+		return -EINVAL;
+
+	return 0;
+}
+
 #define FW_VERSION_ATTR(name, mode, field)				\
 static ssize_t show_##name(struct device *dev,				\
-			  struct device_attribute *attr,		\
-			  char *buf)					\
+			   struct device_attribute *attr, char *buf)	\
 {									\
 	struct drm_device *ddev = dev_get_drvdata(dev);			\
 	struct amdgpu_device *adev = drm_to_adev(ddev);			\
 									\
-	return sysfs_emit(buf, "0x%08x\n", adev->field);	\
+	if (!buf)							\
+		return amdgpu_ucode_is_valid(adev->field);		\
+									\
+	return sysfs_emit(buf, "0x%08x\n", adev->field);		\
 }									\
 static DEVICE_ATTR(name, mode, show_##name, NULL)
 
@@ -525,6 +760,7 @@ FW_VERSION_ATTR(rlc_srlg_fw_version, 0444, gfx.rlc_srlg_fw_version);
 FW_VERSION_ATTR(rlc_srls_fw_version, 0444, gfx.rlc_srls_fw_version);
 FW_VERSION_ATTR(mec_fw_version, 0444, gfx.mec_fw_version);
 FW_VERSION_ATTR(mec2_fw_version, 0444, gfx.mec2_fw_version);
+FW_VERSION_ATTR(imu_fw_version, 0444, gfx.imu_fw_version);
 FW_VERSION_ATTR(sos_fw_version, 0444, psp.sos.fw_version);
 FW_VERSION_ATTR(asd_fw_version, 0444, psp.asd_context.bin_desc.fw_version);
 FW_VERSION_ATTR(ta_ras_fw_version, 0444, psp.ras_context.context.bin_desc.fw_version);
@@ -534,6 +770,10 @@ FW_VERSION_ATTR(sdma_fw_version, 0444, sdma.instance[0].fw_version);
 FW_VERSION_ATTR(sdma2_fw_version, 0444, sdma.instance[1].fw_version);
 FW_VERSION_ATTR(vcn_fw_version, 0444, vcn.fw_version);
 FW_VERSION_ATTR(dmcu_fw_version, 0444, dm.dmcu_fw_version);
+FW_VERSION_ATTR(dmcub_fw_version, 0444, dm.dmcub_fw_version);
+FW_VERSION_ATTR(mes_fw_version, 0444, mes.sched_version & AMDGPU_MES_VERSION_MASK);
+FW_VERSION_ATTR(mes_kiq_fw_version, 0444, mes.kiq_version & AMDGPU_MES_VERSION_MASK);
+FW_VERSION_ATTR(pldm_fw_version, 0444, firmware.pldm_version);
 
 static struct attribute *fw_attrs[] = {
 	&dev_attr_vce_fw_version.attr, &dev_attr_uvd_fw_version.attr,
@@ -546,12 +786,30 @@ static struct attribute *fw_attrs[] = {
 	&dev_attr_ta_ras_fw_version.attr, &dev_attr_ta_xgmi_fw_version.attr,
 	&dev_attr_smc_fw_version.attr, &dev_attr_sdma_fw_version.attr,
 	&dev_attr_sdma2_fw_version.attr, &dev_attr_vcn_fw_version.attr,
-	&dev_attr_dmcu_fw_version.attr, NULL
+	&dev_attr_dmcu_fw_version.attr, &dev_attr_dmcub_fw_version.attr,
+	&dev_attr_imu_fw_version.attr, &dev_attr_mes_fw_version.attr,
+	&dev_attr_mes_kiq_fw_version.attr, &dev_attr_pldm_fw_version.attr,
+	NULL
 };
 
+#define to_dev_attr(x) container_of(x, struct device_attribute, attr)
+
+static umode_t amdgpu_ucode_sys_visible(struct kobject *kobj,
+					struct attribute *attr, int idx)
+{
+	struct device_attribute *dev_attr = to_dev_attr(attr);
+	struct device *dev = kobj_to_dev(kobj);
+
+	if (dev_attr->show(dev, dev_attr, NULL) == -EINVAL)
+		return 0;
+
+	return attr->mode;
+}
+
 static const struct attribute_group fw_attr_group = {
 	.name = "fw_version",
-	.attrs = fw_attrs
+	.attrs = fw_attrs,
+	.is_visible = amdgpu_ucode_sys_visible
 };
 
 int amdgpu_ucode_sysfs_init(struct amdgpu_device *adev)
@@ -570,12 +828,18 @@ static int amdgpu_ucode_init_single_fw(struct amdgpu_device *adev,
 {
 	const struct common_firmware_header *header = NULL;
 	const struct gfx_firmware_header_v1_0 *cp_hdr = NULL;
+	const struct gfx_firmware_header_v2_0 *cpv2_hdr = NULL;
 	const struct dmcu_firmware_header_v1_0 *dmcu_hdr = NULL;
 	const struct dmcub_firmware_header_v1_0 *dmcub_hdr = NULL;
 	const struct mes_firmware_header_v1_0 *mes_hdr = NULL;
+	const struct sdma_firmware_header_v2_0 *sdma_hdr = NULL;
+	const struct sdma_firmware_header_v3_0 *sdmav3_hdr = NULL;
+	const struct imu_firmware_header_v1_0 *imu_hdr = NULL;
+	const struct vpe_firmware_header_v1_0 *vpe_hdr = NULL;
+	const struct umsch_mm_firmware_header_v1_0 *umsch_mm_hdr = NULL;
 	u8 *ucode_addr;
 
-	if (NULL == ucode->fw)
+	if (!ucode->fw)
 		return 0;
 
 	ucode->mc_addr = mc_addr;
@@ -586,12 +850,33 @@ static int amdgpu_ucode_init_single_fw(struct amdgpu_device *adev,
 
 	header = (const struct common_firmware_header *)ucode->fw->data;
 	cp_hdr = (const struct gfx_firmware_header_v1_0 *)ucode->fw->data;
+	cpv2_hdr = (const struct gfx_firmware_header_v2_0 *)ucode->fw->data;
 	dmcu_hdr = (const struct dmcu_firmware_header_v1_0 *)ucode->fw->data;
 	dmcub_hdr = (const struct dmcub_firmware_header_v1_0 *)ucode->fw->data;
 	mes_hdr = (const struct mes_firmware_header_v1_0 *)ucode->fw->data;
+	sdma_hdr = (const struct sdma_firmware_header_v2_0 *)ucode->fw->data;
+	sdmav3_hdr = (const struct sdma_firmware_header_v3_0 *)ucode->fw->data;
+	imu_hdr = (const struct imu_firmware_header_v1_0 *)ucode->fw->data;
+	vpe_hdr = (const struct vpe_firmware_header_v1_0 *)ucode->fw->data;
+	umsch_mm_hdr = (const struct umsch_mm_firmware_header_v1_0 *)ucode->fw->data;
 
 	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
 		switch (ucode->ucode_id) {
+		case AMDGPU_UCODE_ID_SDMA_UCODE_TH0:
+			ucode->ucode_size = le32_to_cpu(sdma_hdr->ctx_ucode_size_bytes);
+			ucode_addr = (u8 *)ucode->fw->data +
+				le32_to_cpu(sdma_hdr->header.ucode_array_offset_bytes);
+			break;
+		case AMDGPU_UCODE_ID_SDMA_UCODE_TH1:
+			ucode->ucode_size = le32_to_cpu(sdma_hdr->ctl_ucode_size_bytes);
+			ucode_addr = (u8 *)ucode->fw->data +
+				le32_to_cpu(sdma_hdr->ctl_ucode_offset);
+			break;
+		case AMDGPU_UCODE_ID_SDMA_RS64:
+			ucode->ucode_size = le32_to_cpu(sdmav3_hdr->ucode_size_bytes);
+			ucode_addr = (u8 *)ucode->fw->data +
+				le32_to_cpu(sdmav3_hdr->header.ucode_array_offset_bytes);
+			break;
 		case AMDGPU_UCODE_ID_CP_MEC1:
 		case AMDGPU_UCODE_ID_CP_MEC2:
 			ucode->ucode_size = le32_to_cpu(header->ucode_size_bytes) -
@@ -626,6 +911,34 @@ static int amdgpu_ucode_init_single_fw(struct amdgpu_device *adev,
 			ucode->ucode_size = adev->gfx.rlc.rlc_dram_ucode_size_bytes;
 			ucode_addr = adev->gfx.rlc.rlc_dram_ucode;
 			break;
+		case AMDGPU_UCODE_ID_RLC_P:
+			ucode->ucode_size = adev->gfx.rlc.rlcp_ucode_size_bytes;
+			ucode_addr = adev->gfx.rlc.rlcp_ucode;
+			break;
+		case AMDGPU_UCODE_ID_RLC_V:
+			ucode->ucode_size = adev->gfx.rlc.rlcv_ucode_size_bytes;
+			ucode_addr = adev->gfx.rlc.rlcv_ucode;
+			break;
+		case AMDGPU_UCODE_ID_GLOBAL_TAP_DELAYS:
+			ucode->ucode_size = adev->gfx.rlc.global_tap_delays_ucode_size_bytes;
+			ucode_addr = adev->gfx.rlc.global_tap_delays_ucode;
+			break;
+		case AMDGPU_UCODE_ID_SE0_TAP_DELAYS:
+			ucode->ucode_size = adev->gfx.rlc.se0_tap_delays_ucode_size_bytes;
+			ucode_addr = adev->gfx.rlc.se0_tap_delays_ucode;
+			break;
+		case AMDGPU_UCODE_ID_SE1_TAP_DELAYS:
+			ucode->ucode_size = adev->gfx.rlc.se1_tap_delays_ucode_size_bytes;
+			ucode_addr = adev->gfx.rlc.se1_tap_delays_ucode;
+			break;
+		case AMDGPU_UCODE_ID_SE2_TAP_DELAYS:
+			ucode->ucode_size = adev->gfx.rlc.se2_tap_delays_ucode_size_bytes;
+			ucode_addr = adev->gfx.rlc.se2_tap_delays_ucode;
+			break;
+		case AMDGPU_UCODE_ID_SE3_TAP_DELAYS:
+			ucode->ucode_size = adev->gfx.rlc.se3_tap_delays_ucode_size_bytes;
+			ucode_addr = adev->gfx.rlc.se3_tap_delays_ucode;
+			break;
 		case AMDGPU_UCODE_ID_CP_MES:
 			ucode->ucode_size = le32_to_cpu(mes_hdr->mes_ucode_size_bytes);
 			ucode_addr = (u8 *)ucode->fw->data +
@@ -636,6 +949,16 @@ static int amdgpu_ucode_init_single_fw(struct amdgpu_device *adev,
 			ucode_addr = (u8 *)ucode->fw->data +
 				le32_to_cpu(mes_hdr->mes_ucode_data_offset_bytes);
 			break;
+		case AMDGPU_UCODE_ID_CP_MES1:
+			ucode->ucode_size = le32_to_cpu(mes_hdr->mes_ucode_size_bytes);
+			ucode_addr = (u8 *)ucode->fw->data +
+				le32_to_cpu(mes_hdr->mes_ucode_offset_bytes);
+			break;
+		case AMDGPU_UCODE_ID_CP_MES1_DATA:
+			ucode->ucode_size = le32_to_cpu(mes_hdr->mes_ucode_data_size_bytes);
+			ucode_addr = (u8 *)ucode->fw->data +
+				le32_to_cpu(mes_hdr->mes_ucode_data_offset_bytes);
+			break;
 		case AMDGPU_UCODE_ID_DMCU_ERAM:
 			ucode->ucode_size = le32_to_cpu(header->ucode_size_bytes) -
 				le32_to_cpu(dmcu_hdr->intv_size_bytes);
@@ -653,6 +976,100 @@ static int amdgpu_ucode_init_single_fw(struct amdgpu_device *adev,
 			ucode_addr = (u8 *)ucode->fw->data +
 				le32_to_cpu(header->ucode_array_offset_bytes);
 			break;
+		case AMDGPU_UCODE_ID_PPTABLE:
+			ucode->ucode_size = ucode->fw->size;
+			ucode_addr = (u8 *)ucode->fw->data;
+			break;
+		case AMDGPU_UCODE_ID_P2S_TABLE:
+			ucode->ucode_size = ucode->fw->size;
+			ucode_addr = (u8 *)ucode->fw->data;
+			break;
+		case AMDGPU_UCODE_ID_IMU_I:
+			ucode->ucode_size = le32_to_cpu(imu_hdr->imu_iram_ucode_size_bytes);
+			ucode_addr = (u8 *)ucode->fw->data +
+				le32_to_cpu(imu_hdr->header.ucode_array_offset_bytes);
+			break;
+		case AMDGPU_UCODE_ID_IMU_D:
+			ucode->ucode_size = le32_to_cpu(imu_hdr->imu_dram_ucode_size_bytes);
+			ucode_addr = (u8 *)ucode->fw->data +
+				le32_to_cpu(imu_hdr->header.ucode_array_offset_bytes) +
+				le32_to_cpu(imu_hdr->imu_iram_ucode_size_bytes);
+			break;
+		case AMDGPU_UCODE_ID_CP_RS64_PFP:
+			ucode->ucode_size = le32_to_cpu(cpv2_hdr->ucode_size_bytes);
+			ucode_addr = (u8 *)ucode->fw->data +
+				le32_to_cpu(header->ucode_array_offset_bytes);
+			break;
+		case AMDGPU_UCODE_ID_CP_RS64_PFP_P0_STACK:
+			ucode->ucode_size = le32_to_cpu(cpv2_hdr->data_size_bytes);
+			ucode_addr = (u8 *)ucode->fw->data +
+				le32_to_cpu(cpv2_hdr->data_offset_bytes);
+			break;
+		case AMDGPU_UCODE_ID_CP_RS64_PFP_P1_STACK:
+			ucode->ucode_size = le32_to_cpu(cpv2_hdr->data_size_bytes);
+			ucode_addr = (u8 *)ucode->fw->data +
+				le32_to_cpu(cpv2_hdr->data_offset_bytes);
+			break;
+		case AMDGPU_UCODE_ID_CP_RS64_ME:
+			ucode->ucode_size = le32_to_cpu(cpv2_hdr->ucode_size_bytes);
+			ucode_addr = (u8 *)ucode->fw->data +
+				le32_to_cpu(header->ucode_array_offset_bytes);
+			break;
+		case AMDGPU_UCODE_ID_CP_RS64_ME_P0_STACK:
+			ucode->ucode_size = le32_to_cpu(cpv2_hdr->data_size_bytes);
+			ucode_addr = (u8 *)ucode->fw->data +
+				le32_to_cpu(cpv2_hdr->data_offset_bytes);
+			break;
+		case AMDGPU_UCODE_ID_CP_RS64_ME_P1_STACK:
+			ucode->ucode_size = le32_to_cpu(cpv2_hdr->data_size_bytes);
+			ucode_addr = (u8 *)ucode->fw->data +
+				le32_to_cpu(cpv2_hdr->data_offset_bytes);
+			break;
+		case AMDGPU_UCODE_ID_CP_RS64_MEC:
+			ucode->ucode_size = le32_to_cpu(cpv2_hdr->ucode_size_bytes);
+			ucode_addr = (u8 *)ucode->fw->data +
+				le32_to_cpu(header->ucode_array_offset_bytes);
+			break;
+		case AMDGPU_UCODE_ID_CP_RS64_MEC_P0_STACK:
+			ucode->ucode_size = le32_to_cpu(cpv2_hdr->data_size_bytes);
+			ucode_addr = (u8 *)ucode->fw->data +
+				le32_to_cpu(cpv2_hdr->data_offset_bytes);
+			break;
+		case AMDGPU_UCODE_ID_CP_RS64_MEC_P1_STACK:
+			ucode->ucode_size = le32_to_cpu(cpv2_hdr->data_size_bytes);
+			ucode_addr = (u8 *)ucode->fw->data +
+				le32_to_cpu(cpv2_hdr->data_offset_bytes);
+			break;
+		case AMDGPU_UCODE_ID_CP_RS64_MEC_P2_STACK:
+			ucode->ucode_size = le32_to_cpu(cpv2_hdr->data_size_bytes);
+			ucode_addr = (u8 *)ucode->fw->data +
+				le32_to_cpu(cpv2_hdr->data_offset_bytes);
+			break;
+		case AMDGPU_UCODE_ID_CP_RS64_MEC_P3_STACK:
+			ucode->ucode_size = le32_to_cpu(cpv2_hdr->data_size_bytes);
+			ucode_addr = (u8 *)ucode->fw->data +
+				le32_to_cpu(cpv2_hdr->data_offset_bytes);
+			break;
+		case AMDGPU_UCODE_ID_VPE_CTX:
+			ucode->ucode_size = le32_to_cpu(vpe_hdr->ctx_ucode_size_bytes);
+			ucode_addr = (u8 *)ucode->fw->data +
+				le32_to_cpu(vpe_hdr->header.ucode_array_offset_bytes);
+			break;
+		case AMDGPU_UCODE_ID_VPE_CTL:
+			ucode->ucode_size = le32_to_cpu(vpe_hdr->ctl_ucode_size_bytes);
+			ucode_addr = (u8 *)ucode->fw->data +
+				le32_to_cpu(vpe_hdr->ctl_ucode_offset);
+			break;
+		case AMDGPU_UCODE_ID_UMSCH_MM_UCODE:
+			ucode->ucode_size = le32_to_cpu(umsch_mm_hdr->umsch_mm_ucode_size_bytes);
+			ucode_addr = (u8 *)ucode->fw->data +
+				le32_to_cpu(umsch_mm_hdr->header.ucode_array_offset_bytes);
+			break;
+		case AMDGPU_UCODE_ID_UMSCH_MM_DATA:
+			ucode->ucode_size = le32_to_cpu(umsch_mm_hdr->umsch_mm_ucode_data_size_bytes);
+			ucode_addr = (u8 *)ucode->fw->data +
+				le32_to_cpu(umsch_mm_hdr->umsch_mm_ucode_data_offset_bytes);
+			break;
 		default:
 			ucode->ucode_size = le32_to_cpu(header->ucode_size_bytes);
 			ucode_addr = (u8 *)ucode->fw->data +
@@ -678,7 +1095,7 @@ static int amdgpu_ucode_patch_jt(struct amdgpu_firmware_info *ucode,
 	uint8_t *src_addr = NULL;
 	uint8_t *dst_addr = NULL;
 
-	if (NULL == ucode->fw)
+	if (!ucode->fw)
 		return 0;
 
 	comm_hdr = (const struct common_firmware_header *)ucode->fw->data;
@@ -696,9 +1113,11 @@ static int amdgpu_ucode_patch_jt(struct amdgpu_firmware_info *ucode,
 
 int amdgpu_ucode_create_bo(struct amdgpu_device *adev)
 {
-	if (adev->firmware.load_type != AMDGPU_FW_LOAD_DIRECT) {
+	if ((adev->firmware.load_type != AMDGPU_FW_LOAD_DIRECT) &&
+	    (adev->firmware.load_type != AMDGPU_FW_LOAD_RLC_BACKDOOR_AUTO)) {
 		amdgpu_bo_create_kernel(adev, adev->firmware.fw_size, PAGE_SIZE,
-			amdgpu_sriov_vf(adev) ? AMDGPU_GEM_DOMAIN_VRAM : AMDGPU_GEM_DOMAIN_GTT,
+			(amdgpu_sriov_vf(adev) || adev->debug_use_vram_fw_buf) ?
+			AMDGPU_GEM_DOMAIN_VRAM : AMDGPU_GEM_DOMAIN_GTT,
 			&adev->firmware.fw_buf,
 			&adev->firmware.fw_buf_mc,
 			&adev->firmware.fw_buf_ptr);
@@ -714,8 +1133,7 @@ int amdgpu_ucode_create_bo(struct amdgpu_device *adev)
 
 void amdgpu_ucode_free_bo(struct amdgpu_device *adev)
 {
-	if (adev->firmware.load_type != AMDGPU_FW_LOAD_DIRECT)
-		amdgpu_bo_free_kernel(&adev->firmware.fw_buf,
+	amdgpu_bo_free_kernel(&adev->firmware.fw_buf,
 		&adev->firmware.fw_buf_mc,
 		&adev->firmware.fw_buf_ptr);
 }
@@ -742,6 +1160,9 @@ int amdgpu_ucode_init_bo(struct amdgpu_device *adev)
 		adev->firmware.max_ucodes = AMDGPU_UCODE_ID_MAXIMUM;
 	}
 
+	if (amdgpu_virt_xgmi_migrate_enabled(adev) && adev->firmware.fw_buf)
+		adev->firmware.fw_buf_mc = amdgpu_bo_fb_aper_addr(adev->firmware.fw_buf);
+
 	for (i = 0; i < adev->firmware.max_ucodes; i++) {
 		ucode = &adev->firmware.ucode[i];
 		if (ucode->fw) {
@@ -750,6 +1171,7 @@ int amdgpu_ucode_init_bo(struct amdgpu_device *adev)
 			if (i == AMDGPU_UCODE_ID_CP_MEC1 &&
 			    adev->firmware.load_type != AMDGPU_FW_LOAD_PSP) {
 				const struct gfx_firmware_header_v1_0 *cp_hdr;
+
 				cp_hdr = (const struct gfx_firmware_header_v1_0 *)ucode->fw->data;
 				amdgpu_ucode_patch_jt(ucode,  adev->firmware.fw_buf_mc + fw_offset,
 						    adev->firmware.fw_buf_ptr + fw_offset);
@@ -760,3 +1182,335 @@ int amdgpu_ucode_init_bo(struct amdgpu_device *adev)
 	}
 	return 0;
 }
+
+static const char *amdgpu_ucode_legacy_naming(struct amdgpu_device *adev, int block_type)
+{
+	if (block_type == MP0_HWIP) {
+		switch (amdgpu_ip_version(adev, MP0_HWIP, 0)) {
+		case IP_VERSION(9, 0, 0):
+			switch (adev->asic_type) {
+			case CHIP_VEGA10:
+				return "vega10";
+			case CHIP_VEGA12:
+				return "vega12";
+			default:
+				return NULL;
+			}
+		case IP_VERSION(10, 0, 0):
+		case IP_VERSION(10, 0, 1):
+			if (adev->asic_type == CHIP_RAVEN) {
+				if (adev->apu_flags & AMD_APU_IS_RAVEN2)
+					return "raven2";
+				else if (adev->apu_flags & AMD_APU_IS_PICASSO)
+					return "picasso";
+				return "raven";
+			}
+			break;
+		case IP_VERSION(11, 0, 0):
+			return "navi10";
+		case IP_VERSION(11, 0, 2):
+			return "vega20";
+		case IP_VERSION(11, 0, 3):
+			return "renoir";
+		case IP_VERSION(11, 0, 4):
+			return "arcturus";
+		case IP_VERSION(11, 0, 5):
+			return "navi14";
+		case IP_VERSION(11, 0, 7):
+			return "sienna_cichlid";
+		case IP_VERSION(11, 0, 9):
+			return "navi12";
+		case IP_VERSION(11, 0, 11):
+			return "navy_flounder";
+		case IP_VERSION(11, 0, 12):
+			return "dimgrey_cavefish";
+		case IP_VERSION(11, 0, 13):
+			return "beige_goby";
+		case IP_VERSION(11, 5, 0):
+		case IP_VERSION(11, 5, 2):
+			return "vangogh";
+		case IP_VERSION(12, 0, 1):
+			return "green_sardine";
+		case IP_VERSION(13, 0, 2):
+			return "aldebaran";
+		case IP_VERSION(13, 0, 1):
+		case IP_VERSION(13, 0, 3):
+			return "yellow_carp";
+		}
+	} else if (block_type == MP1_HWIP) {
+		switch (amdgpu_ip_version(adev, MP1_HWIP, 0)) {
+		case IP_VERSION(9, 0, 0):
+		case IP_VERSION(10, 0, 0):
+		case IP_VERSION(10, 0, 1):
+		case IP_VERSION(11, 0, 2):
+			if (adev->asic_type == CHIP_ARCTURUS)
+				return "arcturus_smc";
+			return NULL;
+		case IP_VERSION(11, 0, 0):
+			return "navi10_smc";
+		case IP_VERSION(11, 0, 5):
+			return "navi14_smc";
+		case IP_VERSION(11, 0, 9):
+			return "navi12_smc";
+		case IP_VERSION(11, 0, 7):
+			return "sienna_cichlid_smc";
+		case IP_VERSION(11, 0, 11):
+			return "navy_flounder_smc";
+		case IP_VERSION(11, 0, 12):
+			return "dimgrey_cavefish_smc";
+		case IP_VERSION(11, 0, 13):
+			return "beige_goby_smc";
+		case IP_VERSION(13, 0, 2):
+			return "aldebaran_smc";
+		}
+	} else if (block_type == SDMA0_HWIP) {
+		switch (amdgpu_ip_version(adev, SDMA0_HWIP, 0)) {
+		case IP_VERSION(4, 0, 0):
+			return "vega10_sdma";
+		case IP_VERSION(4, 0, 1):
+			return "vega12_sdma";
+		case IP_VERSION(4, 1, 0):
+		case IP_VERSION(4, 1, 1):
+			if (adev->apu_flags & AMD_APU_IS_RAVEN2)
+				return "raven2_sdma";
+			else if (adev->apu_flags & AMD_APU_IS_PICASSO)
+				return "picasso_sdma";
+			return "raven_sdma";
+		case IP_VERSION(4, 1, 2):
+			if (adev->apu_flags & AMD_APU_IS_RENOIR)
+				return "renoir_sdma";
+			return "green_sardine_sdma";
+		case IP_VERSION(4, 2, 0):
+			return "vega20_sdma";
+		case IP_VERSION(4, 2, 2):
+			return "arcturus_sdma";
+		case IP_VERSION(4, 4, 0):
+			return "aldebaran_sdma";
+		case IP_VERSION(5, 0, 0):
+			return "navi10_sdma";
+		case IP_VERSION(5, 0, 1):
+			return "cyan_skillfish2_sdma";
+		case IP_VERSION(5, 0, 2):
+			return "navi14_sdma";
+		case IP_VERSION(5, 0, 5):
+			return "navi12_sdma";
+		case IP_VERSION(5, 2, 0):
+			return "sienna_cichlid_sdma";
+		case IP_VERSION(5, 2, 2):
+			return "navy_flounder_sdma";
+		case IP_VERSION(5, 2, 4):
+			return "dimgrey_cavefish_sdma";
+		case IP_VERSION(5, 2, 5):
+			return "beige_goby_sdma";
+		case IP_VERSION(5, 2, 3):
+			return "yellow_carp_sdma";
+		case IP_VERSION(5, 2, 1):
+			return "vangogh_sdma";
+		}
+	} else if (block_type == UVD_HWIP) {
+		switch (amdgpu_ip_version(adev, UVD_HWIP, 0)) {
+		case IP_VERSION(1, 0, 0):
+		case IP_VERSION(1, 0, 1):
+			if (adev->apu_flags & AMD_APU_IS_RAVEN2)
+				return "raven2_vcn";
+			else if (adev->apu_flags & AMD_APU_IS_PICASSO)
+				return "picasso_vcn";
+			return "raven_vcn";
+		case IP_VERSION(2, 5, 0):
+			return "arcturus_vcn";
+		case IP_VERSION(2, 2, 0):
+			if (adev->apu_flags & AMD_APU_IS_RENOIR)
+				return "renoir_vcn";
+			return "green_sardine_vcn";
+		case IP_VERSION(2, 6, 0):
+			return "aldebaran_vcn";
+		case IP_VERSION(2, 0, 0):
+			return "navi10_vcn";
+		case IP_VERSION(2, 0, 2):
+			if (adev->asic_type == CHIP_NAVI12)
+				return "navi12_vcn";
+			return "navi14_vcn";
+		case IP_VERSION(3, 0, 0):
+		case IP_VERSION(3, 0, 64):
+		case IP_VERSION(3, 0, 192):
+			if (amdgpu_ip_version(adev, GC_HWIP, 0) ==
+			    IP_VERSION(10, 3, 0))
+				return "sienna_cichlid_vcn";
+			return "navy_flounder_vcn";
+		case IP_VERSION(3, 0, 2):
+			return "vangogh_vcn";
+		case IP_VERSION(3, 0, 16):
+			return "dimgrey_cavefish_vcn";
+		case IP_VERSION(3, 0, 33):
+			return "beige_goby_vcn";
+		case IP_VERSION(3, 1, 1):
+			return "yellow_carp_vcn";
+		}
+	} else if (block_type == GC_HWIP) {
+		switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+		case IP_VERSION(9, 0, 1):
+			return "vega10";
+		case IP_VERSION(9, 2, 1):
+			return "vega12";
+		case IP_VERSION(9, 4, 0):
+			return "vega20";
+		case IP_VERSION(9, 2, 2):
+		case IP_VERSION(9, 1, 0):
+			if (adev->apu_flags & AMD_APU_IS_RAVEN2)
+				return "raven2";
+			else if (adev->apu_flags & AMD_APU_IS_PICASSO)
+				return "picasso";
+			return "raven";
+		case IP_VERSION(9, 4, 1):
+			return "arcturus";
+		case IP_VERSION(9, 3, 0):
+			if (adev->apu_flags & AMD_APU_IS_RENOIR)
+				return "renoir";
+			return "green_sardine";
+		case IP_VERSION(9, 4, 2):
+			return "aldebaran";
+		case IP_VERSION(10, 1, 10):
+			return "navi10";
+		case IP_VERSION(10, 1, 1):
+			return "navi14";
+		case IP_VERSION(10, 1, 2):
+			return "navi12";
+		case IP_VERSION(10, 3, 0):
+			return "sienna_cichlid";
+		case IP_VERSION(10, 3, 2):
+			return "navy_flounder";
+		case IP_VERSION(10, 3, 1):
+			return "vangogh";
+		case IP_VERSION(10, 3, 4):
+			return "dimgrey_cavefish";
+		case IP_VERSION(10, 3, 5):
+			return "beige_goby";
+		case IP_VERSION(10, 3, 3):
+			return "yellow_carp";
+		case IP_VERSION(10, 1, 3):
+		case IP_VERSION(10, 1, 4):
+			return "cyan_skillfish2";
+		}
+	}
+	return NULL;
+}
+
+bool amdgpu_is_kicker_fw(struct amdgpu_device *adev)
+{
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(kicker_device_list); i++) {
+		if (adev->pdev->device == kicker_device_list[i].device &&
+		    adev->pdev->revision == kicker_device_list[i].revision)
+			return true;
+	}
+
+	return false;
+}
+
+void amdgpu_ucode_ip_version_decode(struct amdgpu_device *adev, int block_type, char *ucode_prefix, int len)
+{
+	int maj, min, rev;
+	char *ip_name;
+	const char *legacy;
+	uint32_t version = amdgpu_ip_version(adev, block_type, 0);
+
+	legacy = amdgpu_ucode_legacy_naming(adev, block_type);
+	if (legacy) {
+		snprintf(ucode_prefix, len, "%s", legacy);
+		return;
+	}
+
+	switch (block_type) {
+	case GC_HWIP:
+		ip_name = "gc";
+		break;
+	case SDMA0_HWIP:
+		ip_name = "sdma";
+		break;
+	case MP0_HWIP:
+		ip_name = "psp";
+		break;
+	case MP1_HWIP:
+		ip_name = "smu";
+		break;
+	case UVD_HWIP:
+		ip_name = "vcn";
+		break;
+	case VPE_HWIP:
+		ip_name = "vpe";
+		break;
+	case ISP_HWIP:
+		ip_name = "isp";
+		break;
+	default:
+		BUG();
+	}
+
+	maj = IP_VERSION_MAJ(version);
+	min = IP_VERSION_MIN(version);
+	rev = IP_VERSION_REV(version);
+
+	snprintf(ucode_prefix, len, "%s_%d_%d_%d", ip_name, maj, min, rev);
+}
+
+/*
+ * amdgpu_ucode_request - Fetch and validate amdgpu microcode
+ *
+ * @adev: amdgpu device
+ * @fw: pointer to load firmware to
+ * @required: whether the firmware is required
+ * @fmt: firmware name format string
+ * @...: variable arguments
+ *
+ * This is a helper that will use request_firmware and amdgpu_ucode_validate
+ * to load and run basic validation on firmware. If the load fails, remap
+ * the error code to -ENODEV, so that early_init functions will fail to load.
+ */
+int amdgpu_ucode_request(struct amdgpu_device *adev, const struct firmware **fw,
+			 enum amdgpu_ucode_required required, const char *fmt, ...)
+{
+	char fname[AMDGPU_UCODE_NAME_MAX];
+	va_list ap;
+	int r;
+
+	va_start(ap, fmt);
+	r = vsnprintf(fname, sizeof(fname), fmt, ap);
+	va_end(ap);
+	if (r == sizeof(fname)) {
+		dev_warn(adev->dev, "amdgpu firmware name buffer overflow\n");
+		return -EOVERFLOW;
+	}
+
+	if (required == AMDGPU_UCODE_REQUIRED)
+		r = request_firmware(fw, fname, adev->dev);
+	else {
+		r = firmware_request_nowarn(fw, fname, adev->dev);
+		if (r)
+			drm_info(&adev->ddev, "Optional firmware \"%s\" was not found\n", fname);
+	}
+	if (r)
+		return -ENODEV;
+
+	r = amdgpu_ucode_validate(*fw);
+	if (r)
+		/*
+		 * The amdgpu_ucode_request() should be paired with amdgpu_ucode_release()
+		 * regardless of success/failure, and the amdgpu_ucode_release() takes care of
+		 * firmware release and need to avoid redundant release FW operation here.
+		 */
+		dev_dbg(adev->dev, "\"%s\" failed to validate\n", fname);
+
+	return r;
+}
+
+/*
+ * amdgpu_ucode_release - Release firmware microcode
+ *
+ * @fw: pointer to firmware to release
+ */
+void amdgpu_ucode_release(const struct firmware **fw)
+{
+	release_firmware(*fw);
+	*fw = NULL;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.h
index 7c2538db3cd5..6349aad6da35 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.h
@@ -25,6 +25,8 @@
 
 #include "amdgpu_socbb.h"
 
+#define RS64_FW_UC_START_ADDR_LO 0x3000
+
 struct common_firmware_header {
 	uint32_t size_bytes; /* size of the entire header+image(s) in bytes */
 	uint32_t header_size_bytes; /* size of just the header in bytes */
@@ -124,6 +126,10 @@ enum psp_fw_type {
 	PSP_FW_TYPE_PSP_SOC_DRV,
 	PSP_FW_TYPE_PSP_INTF_DRV,
 	PSP_FW_TYPE_PSP_DBG_DRV,
+	PSP_FW_TYPE_PSP_RAS_DRV,
+	PSP_FW_TYPE_PSP_IPKEYMGR_DRV,
+	PSP_FW_TYPE_PSP_SPDM_DRV,
+	PSP_FW_TYPE_MAX_INDEX,
 };
 
 /* version_major=2, version_minor=0 */
@@ -133,6 +139,14 @@ struct psp_firmware_header_v2_0 {
 	struct psp_fw_bin_desc psp_fw_bin[];
 };
 
+/* version_major=2, version_minor=1 */
+struct psp_firmware_header_v2_1 {
+	struct common_firmware_header header;
+	uint32_t psp_fw_bin_count;
+	uint32_t psp_aux_fw_bin_index;
+	struct psp_fw_bin_desc psp_fw_bin[];
+};
+
 /* version_major=1, version_minor=0 */
 struct ta_firmware_header_v1_0 {
 	struct common_firmware_header header;
@@ -152,6 +166,7 @@ enum ta_fw_type {
 	TA_FW_TYPE_PSP_DTM,
 	TA_FW_TYPE_PSP_RAP,
 	TA_FW_TYPE_PSP_SECUREDISPLAY,
+	TA_FW_TYPE_PSP_XGMI_AUX,
 	TA_FW_TYPE_MAX_INDEX,
 };
 
@@ -170,6 +185,18 @@ struct gfx_firmware_header_v1_0 {
 	uint32_t jt_size;  /* size of jt */
 };
 
+/* version_major=2, version_minor=0 */
+struct gfx_firmware_header_v2_0 {
+	struct common_firmware_header header;
+	uint32_t ucode_feature_version;
+	uint32_t ucode_size_bytes;
+	uint32_t ucode_offset_bytes;
+	uint32_t data_size_bytes;
+	uint32_t data_offset_bytes;
+	uint32_t ucode_start_addr_lo;
+	uint32_t ucode_start_addr_hi;
+};
+
 /* version_major=1, version_minor=0 */
 struct mes_firmware_header_v1_0 {
 	struct common_firmware_header header;
@@ -236,7 +263,7 @@ struct rlc_firmware_header_v2_1 {
 	uint32_t save_restore_list_srm_offset_bytes;
 };
 
-/* version_major=2, version_minor=1 */
+/* version_major=2, version_minor=2 */
 struct rlc_firmware_header_v2_2 {
 	struct rlc_firmware_header_v2_1 v2_1;
 	uint32_t rlc_iram_ucode_size_bytes;
@@ -245,6 +272,34 @@ struct rlc_firmware_header_v2_2 {
 	uint32_t rlc_dram_ucode_offset_bytes;
 };
 
+/* version_major=2, version_minor=3 */
+struct rlc_firmware_header_v2_3 {
+    struct rlc_firmware_header_v2_2 v2_2;
+    uint32_t rlcp_ucode_version;
+    uint32_t rlcp_ucode_feature_version;
+    uint32_t rlcp_ucode_size_bytes;
+    uint32_t rlcp_ucode_offset_bytes;
+    uint32_t rlcv_ucode_version;
+    uint32_t rlcv_ucode_feature_version;
+    uint32_t rlcv_ucode_size_bytes;
+    uint32_t rlcv_ucode_offset_bytes;
+};
+
+/* version_major=2, version_minor=4 */
+struct rlc_firmware_header_v2_4 {
+    struct rlc_firmware_header_v2_3 v2_3;
+    uint32_t global_tap_delays_ucode_size_bytes;
+    uint32_t global_tap_delays_ucode_offset_bytes;
+    uint32_t se0_tap_delays_ucode_size_bytes;
+    uint32_t se0_tap_delays_ucode_offset_bytes;
+    uint32_t se1_tap_delays_ucode_size_bytes;
+    uint32_t se1_tap_delays_ucode_offset_bytes;
+    uint32_t se2_tap_delays_ucode_size_bytes;
+    uint32_t se2_tap_delays_ucode_offset_bytes;
+    uint32_t se3_tap_delays_ucode_size_bytes;
+    uint32_t se3_tap_delays_ucode_offset_bytes;
+};
+
 /* version_major=1, version_minor=0 */
 struct sdma_firmware_header_v1_0 {
 	struct common_firmware_header header;
@@ -260,6 +315,57 @@ struct sdma_firmware_header_v1_1 {
 	uint32_t digest_size;
 };
 
+/* version_major=2, version_minor=0 */
+struct sdma_firmware_header_v2_0 {
+	struct common_firmware_header header;
+	uint32_t ucode_feature_version;
+	uint32_t ctx_ucode_size_bytes; /* context thread ucode size */
+	uint32_t ctx_jt_offset; /* context thread jt location */
+	uint32_t ctx_jt_size; /* context thread size of jt */
+	uint32_t ctl_ucode_offset;
+	uint32_t ctl_ucode_size_bytes; /* control thread ucode size */
+	uint32_t ctl_jt_offset; /* control thread jt location */
+	uint32_t ctl_jt_size; /* control thread size of jt */
+};
+
+/* version_major=1, version_minor=0 */
+struct vpe_firmware_header_v1_0 {
+	struct common_firmware_header header;
+	uint32_t ucode_feature_version;
+	uint32_t ctx_ucode_size_bytes; /* context thread ucode size */
+	uint32_t ctx_jt_offset; /* context thread jt location */
+	uint32_t ctx_jt_size; /* context thread size of jt */
+	uint32_t ctl_ucode_offset;
+	uint32_t ctl_ucode_size_bytes; /* control thread ucode size */
+	uint32_t ctl_jt_offset; /* control thread jt location */
+	uint32_t ctl_jt_size; /* control thread size of jt */
+};
+
+/* version_major=1, version_minor=0 */
+struct umsch_mm_firmware_header_v1_0 {
+	struct common_firmware_header header;
+	uint32_t umsch_mm_ucode_version;
+	uint32_t umsch_mm_ucode_size_bytes;
+	uint32_t umsch_mm_ucode_offset_bytes;
+	uint32_t umsch_mm_ucode_data_version;
+	uint32_t umsch_mm_ucode_data_size_bytes;
+	uint32_t umsch_mm_ucode_data_offset_bytes;
+	uint32_t umsch_mm_irq_start_addr_lo;
+	uint32_t umsch_mm_irq_start_addr_hi;
+	uint32_t umsch_mm_uc_start_addr_lo;
+	uint32_t umsch_mm_uc_start_addr_hi;
+	uint32_t umsch_mm_data_start_addr_lo;
+	uint32_t umsch_mm_data_start_addr_hi;
+};
+
+/* version_major=3, version_minor=0 */
+struct sdma_firmware_header_v3_0 {
+    struct common_firmware_header header;
+    uint32_t ucode_feature_version;
+    uint32_t ucode_offset_bytes;
+    uint32_t ucode_size_bytes;
+};
+
 /* gpu info payload */
 struct gpu_info_firmware_v1_0 {
 	uint32_t gc_num_se;
@@ -313,6 +419,15 @@ struct dmcub_firmware_header_v1_0 {
 	uint32_t bss_data_bytes; /* size of bss/data region, in bytes */
 };
 
+/* version_major=1, version_minor=0 */
+struct imu_firmware_header_v1_0 {
+    struct common_firmware_header header;
+    uint32_t imu_iram_ucode_size_bytes;
+    uint32_t imu_iram_ucode_offset_bytes;
+    uint32_t imu_dram_ucode_size_bytes;
+    uint32_t imu_dram_ucode_offset_bytes;
+};
+
 /* header is fixed size */
 union amdgpu_firmware_header {
 	struct common_firmware_header common;
@@ -323,27 +438,36 @@ union amdgpu_firmware_header {
 	struct psp_firmware_header_v1_1 psp_v1_1;
 	struct psp_firmware_header_v1_3 psp_v1_3;
 	struct psp_firmware_header_v2_0 psp_v2_0;
+	struct psp_firmware_header_v2_0 psp_v2_1;
 	struct ta_firmware_header_v1_0 ta;
 	struct ta_firmware_header_v2_0 ta_v2_0;
 	struct gfx_firmware_header_v1_0 gfx;
+	struct gfx_firmware_header_v2_0 gfx_v2_0;
 	struct rlc_firmware_header_v1_0 rlc;
 	struct rlc_firmware_header_v2_0 rlc_v2_0;
 	struct rlc_firmware_header_v2_1 rlc_v2_1;
+	struct rlc_firmware_header_v2_2 rlc_v2_2;
+	struct rlc_firmware_header_v2_3 rlc_v2_3;
+	struct rlc_firmware_header_v2_4 rlc_v2_4;
 	struct sdma_firmware_header_v1_0 sdma;
 	struct sdma_firmware_header_v1_1 sdma_v1_1;
+	struct sdma_firmware_header_v2_0 sdma_v2_0;
+	struct sdma_firmware_header_v3_0 sdma_v3_0;
 	struct gpu_info_firmware_header_v1_0 gpu_info;
 	struct dmcu_firmware_header_v1_0 dmcu;
 	struct dmcub_firmware_header_v1_0 dmcub;
+	struct imu_firmware_header_v1_0 imu;
 	uint8_t raw[0x100];
 };
 
-#define UCODE_MAX_PSP_PACKAGING ((sizeof(union amdgpu_firmware_header) - sizeof(struct common_firmware_header) - 4) / sizeof(struct psp_fw_bin_desc))
+#define UCODE_MAX_PSP_PACKAGING (((sizeof(union amdgpu_firmware_header) - sizeof(struct common_firmware_header) - 4) / sizeof(struct psp_fw_bin_desc)) * 2)
 
 /*
  * fw loading support
  */
 enum AMDGPU_UCODE_ID {
-	AMDGPU_UCODE_ID_SDMA0 = 0,
+	AMDGPU_UCODE_ID_CAP = 0,
+	AMDGPU_UCODE_ID_SDMA0,
 	AMDGPU_UCODE_ID_SDMA1,
 	AMDGPU_UCODE_ID_SDMA2,
 	AMDGPU_UCODE_ID_SDMA3,
@@ -351,23 +475,49 @@ enum AMDGPU_UCODE_ID {
 	AMDGPU_UCODE_ID_SDMA5,
 	AMDGPU_UCODE_ID_SDMA6,
 	AMDGPU_UCODE_ID_SDMA7,
+	AMDGPU_UCODE_ID_SDMA_UCODE_TH0,
+	AMDGPU_UCODE_ID_SDMA_UCODE_TH1,
+	AMDGPU_UCODE_ID_SDMA_RS64,
 	AMDGPU_UCODE_ID_CP_CE,
 	AMDGPU_UCODE_ID_CP_PFP,
 	AMDGPU_UCODE_ID_CP_ME,
+	AMDGPU_UCODE_ID_CP_RS64_PFP,
+	AMDGPU_UCODE_ID_CP_RS64_ME,
+	AMDGPU_UCODE_ID_CP_RS64_MEC,
+	AMDGPU_UCODE_ID_CP_RS64_PFP_P0_STACK,
+	AMDGPU_UCODE_ID_CP_RS64_PFP_P1_STACK,
+	AMDGPU_UCODE_ID_CP_RS64_ME_P0_STACK,
+	AMDGPU_UCODE_ID_CP_RS64_ME_P1_STACK,
+	AMDGPU_UCODE_ID_CP_RS64_MEC_P0_STACK,
+	AMDGPU_UCODE_ID_CP_RS64_MEC_P1_STACK,
+	AMDGPU_UCODE_ID_CP_RS64_MEC_P2_STACK,
+	AMDGPU_UCODE_ID_CP_RS64_MEC_P3_STACK,
 	AMDGPU_UCODE_ID_CP_MEC1,
 	AMDGPU_UCODE_ID_CP_MEC1_JT,
 	AMDGPU_UCODE_ID_CP_MEC2,
 	AMDGPU_UCODE_ID_CP_MEC2_JT,
 	AMDGPU_UCODE_ID_CP_MES,
 	AMDGPU_UCODE_ID_CP_MES_DATA,
+	AMDGPU_UCODE_ID_CP_MES1,
+	AMDGPU_UCODE_ID_CP_MES1_DATA,
+	AMDGPU_UCODE_ID_IMU_I,
+	AMDGPU_UCODE_ID_IMU_D,
+	AMDGPU_UCODE_ID_GLOBAL_TAP_DELAYS,
+	AMDGPU_UCODE_ID_SE0_TAP_DELAYS,
+	AMDGPU_UCODE_ID_SE1_TAP_DELAYS,
+	AMDGPU_UCODE_ID_SE2_TAP_DELAYS,
+	AMDGPU_UCODE_ID_SE3_TAP_DELAYS,
 	AMDGPU_UCODE_ID_RLC_RESTORE_LIST_CNTL,
 	AMDGPU_UCODE_ID_RLC_RESTORE_LIST_GPM_MEM,
 	AMDGPU_UCODE_ID_RLC_RESTORE_LIST_SRM_MEM,
 	AMDGPU_UCODE_ID_RLC_IRAM,
 	AMDGPU_UCODE_ID_RLC_DRAM,
+	AMDGPU_UCODE_ID_RLC_P,
+	AMDGPU_UCODE_ID_RLC_V,
 	AMDGPU_UCODE_ID_RLC_G,
 	AMDGPU_UCODE_ID_STORAGE,
 	AMDGPU_UCODE_ID_SMC,
+	AMDGPU_UCODE_ID_PPTABLE,
 	AMDGPU_UCODE_ID_UVD,
 	AMDGPU_UCODE_ID_UVD1,
 	AMDGPU_UCODE_ID_VCE,
@@ -378,6 +528,15 @@ enum AMDGPU_UCODE_ID {
 	AMDGPU_UCODE_ID_VCN0_RAM,
 	AMDGPU_UCODE_ID_VCN1_RAM,
 	AMDGPU_UCODE_ID_DMCUB,
+	AMDGPU_UCODE_ID_VPE_CTX,
+	AMDGPU_UCODE_ID_VPE_CTL,
+	AMDGPU_UCODE_ID_VPE,
+	AMDGPU_UCODE_ID_UMSCH_MM_UCODE,
+	AMDGPU_UCODE_ID_UMSCH_MM_DATA,
+	AMDGPU_UCODE_ID_UMSCH_MM_CMD_BUFFER,
+	AMDGPU_UCODE_ID_P2S_TABLE,
+	AMDGPU_UCODE_ID_JPEG_RAM,
+	AMDGPU_UCODE_ID_ISP,
 	AMDGPU_UCODE_ID_MAXIMUM,
 };
 
@@ -390,11 +549,16 @@ enum AMDGPU_UCODE_STATUS {
 
 enum amdgpu_firmware_load_type {
 	AMDGPU_FW_LOAD_DIRECT = 0,
-	AMDGPU_FW_LOAD_SMU,
 	AMDGPU_FW_LOAD_PSP,
+	AMDGPU_FW_LOAD_SMU,
 	AMDGPU_FW_LOAD_RLC_BACKDOOR_AUTO,
 };
 
+enum amdgpu_ucode_required {
+	AMDGPU_UCODE_OPTIONAL,
+	AMDGPU_UCODE_REQUIRED,
+};
+
 /* conform to smu_ucode_xfer_cz.h */
 #define AMDGPU_SDMA0_UCODE_LOADED	0x00000001
 #define AMDGPU_SDMA1_UCODE_LOADED	0x00000002
@@ -438,16 +602,26 @@ struct amdgpu_firmware {
 
 	void *fw_buf_ptr;
 	uint64_t fw_buf_mc;
+	uint32_t pldm_version;
+};
+
+struct kicker_device{
+	unsigned short device;
+	u8 revision;
 };
 
 void amdgpu_ucode_print_mc_hdr(const struct common_firmware_header *hdr);
 void amdgpu_ucode_print_smc_hdr(const struct common_firmware_header *hdr);
+void amdgpu_ucode_print_imu_hdr(const struct common_firmware_header *hdr);
 void amdgpu_ucode_print_gfx_hdr(const struct common_firmware_header *hdr);
 void amdgpu_ucode_print_rlc_hdr(const struct common_firmware_header *hdr);
 void amdgpu_ucode_print_sdma_hdr(const struct common_firmware_header *hdr);
 void amdgpu_ucode_print_psp_hdr(const struct common_firmware_header *hdr);
 void amdgpu_ucode_print_gpu_info_hdr(const struct common_firmware_header *hdr);
-int amdgpu_ucode_validate(const struct firmware *fw);
+__printf(4, 5)
+int amdgpu_ucode_request(struct amdgpu_device *adev, const struct firmware **fw,
+			 enum amdgpu_ucode_required required, const char *fmt, ...);
+void amdgpu_ucode_release(const struct firmware **fw);
 bool amdgpu_ucode_hdr_version(union amdgpu_firmware_header *hdr,
 				uint16_t hdr_major, uint16_t hdr_minor);
 
@@ -462,4 +636,7 @@ amdgpu_ucode_get_load_type(struct amdgpu_device *adev, int load_type);
 
 const char *amdgpu_ucode_name(enum AMDGPU_UCODE_ID ucode_id);
 
+void amdgpu_ucode_ip_version_decode(struct amdgpu_device *adev, int block_type, char *ucode_prefix, int len);
+bool amdgpu_is_kicker_fw(struct amdgpu_device *adev);
+
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c
index a90029ee9733..3f0b0e9af4f3 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c
@@ -21,126 +21,363 @@
  *
  */
 
-#include "amdgpu_ras.h"
+#include <linux/sort.h>
+#include "amdgpu.h"
+#include "umc_v6_7.h"
+#include "amdgpu_ras_mgr.h"
+#define MAX_UMC_POISON_POLLING_TIME_SYNC   20  //ms
 
-int amdgpu_umc_ras_late_init(struct amdgpu_device *adev)
+#define MAX_UMC_HASH_STRING_SIZE  256
+
+static int amdgpu_umc_convert_error_address(struct amdgpu_device *adev,
+				    struct ras_err_data *err_data, uint64_t err_addr,
+				    uint32_t ch_inst, uint32_t umc_inst)
 {
-	int r;
-	struct ras_fs_if fs_info = {
-		.sysfs_name = "umc_err_count",
-	};
-	struct ras_ih_if ih_info = {
-		.cb = amdgpu_umc_process_ras_data_cb,
-	};
+	switch (amdgpu_ip_version(adev, UMC_HWIP, 0)) {
+	case IP_VERSION(6, 7, 0):
+		umc_v6_7_convert_error_address(adev,
+				err_data, err_addr, ch_inst, umc_inst);
+		break;
+	default:
+		dev_warn(adev->dev,
+			 "UMC address to Physical address translation is not supported\n");
+		return AMDGPU_RAS_FAIL;
+	}
+
+	return AMDGPU_RAS_SUCCESS;
+}
+
+int amdgpu_umc_page_retirement_mca(struct amdgpu_device *adev,
+			uint64_t err_addr, uint32_t ch_inst, uint32_t umc_inst)
+{
+	struct ras_err_data err_data;
+	int ret;
+
+	ret = amdgpu_ras_error_data_init(&err_data);
+	if (ret)
+		return ret;
 
-	if (!adev->umc.ras_if) {
-		adev->umc.ras_if =
-			kmalloc(sizeof(struct ras_common_if), GFP_KERNEL);
-		if (!adev->umc.ras_if)
-			return -ENOMEM;
-		adev->umc.ras_if->block = AMDGPU_RAS_BLOCK__UMC;
-		adev->umc.ras_if->type = AMDGPU_RAS_ERROR__MULTI_UNCORRECTABLE;
-		adev->umc.ras_if->sub_block_index = 0;
+	err_data.err_addr =
+		kcalloc(adev->umc.max_ras_err_cnt_per_query,
+			sizeof(struct eeprom_table_record), GFP_KERNEL);
+	if (!err_data.err_addr) {
+		dev_warn(adev->dev,
+			"Failed to alloc memory for umc error record in MCA notifier!\n");
+		ret = AMDGPU_RAS_FAIL;
+		goto out_fini_err_data;
 	}
-	ih_info.head = fs_info.head = *adev->umc.ras_if;
 
-	r = amdgpu_ras_late_init(adev, adev->umc.ras_if,
-				 &fs_info, &ih_info);
-	if (r)
-		goto free;
+	err_data.err_addr_len = adev->umc.max_ras_err_cnt_per_query;
 
-	if (amdgpu_ras_is_supported(adev, adev->umc.ras_if->block)) {
-		r = amdgpu_irq_get(adev, &adev->gmc.ecc_irq, 0);
-		if (r)
-			goto late_fini;
-	} else {
-		r = 0;
-		goto free;
+	/*
+	 * Translate UMC channel address to Physical address
+	 */
+	ret = amdgpu_umc_convert_error_address(adev, &err_data, err_addr,
+					ch_inst, umc_inst);
+	if (ret)
+		goto out_free_err_addr;
+
+	if (amdgpu_bad_page_threshold != 0) {
+		amdgpu_ras_add_bad_pages(adev, err_data.err_addr,
+						err_data.err_addr_cnt, false);
+		amdgpu_ras_save_bad_pages(adev, NULL);
 	}
 
-	/* ras init of specific umc version */
-	if (adev->umc.ras_funcs &&
-	    adev->umc.ras_funcs->err_cnt_init)
-		adev->umc.ras_funcs->err_cnt_init(adev);
+out_free_err_addr:
+	kfree(err_data.err_addr);
 
-	return 0;
+out_fini_err_data:
+	amdgpu_ras_error_data_fini(&err_data);
 
-late_fini:
-	amdgpu_ras_late_fini(adev, adev->umc.ras_if, &ih_info);
-free:
-	kfree(adev->umc.ras_if);
-	adev->umc.ras_if = NULL;
-	return r;
+	return ret;
 }
 
-void amdgpu_umc_ras_fini(struct amdgpu_device *adev)
+void amdgpu_umc_handle_bad_pages(struct amdgpu_device *adev,
+			void *ras_error_status)
 {
-	if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__UMC) &&
-			adev->umc.ras_if) {
-		struct ras_common_if *ras_if = adev->umc.ras_if;
-		struct ras_ih_if ih_info = {
-			.head = *ras_if,
-			.cb = amdgpu_umc_process_ras_data_cb,
-		};
+	struct ras_err_data *err_data = (struct ras_err_data *)ras_error_status;
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+	struct amdgpu_ras_eeprom_control *control = &con->eeprom_control;
+	unsigned int error_query_mode;
+	int ret = 0;
+	unsigned long err_count;
+
+	amdgpu_ras_get_error_query_mode(adev, &error_query_mode);
 
-		amdgpu_ras_late_fini(adev, ras_if, &ih_info);
-		kfree(ras_if);
+	err_data->err_addr =
+		kcalloc(adev->umc.max_ras_err_cnt_per_query,
+			sizeof(struct eeprom_table_record), GFP_KERNEL);
+
+	/* still call query_ras_error_address to clear error status
+	 * even NOMEM error is encountered
+	 */
+	if (!err_data->err_addr)
+		dev_warn(adev->dev,
+			"Failed to alloc memory for umc error address record!\n");
+	else
+		err_data->err_addr_len = adev->umc.max_ras_err_cnt_per_query;
+
+	mutex_lock(&con->page_retirement_lock);
+	if (!amdgpu_ras_smu_eeprom_supported(adev)) {
+		ret = amdgpu_dpm_get_ecc_info(adev, (void *)&(con->umc_ecc));
+		if (ret == -EOPNOTSUPP &&
+		    error_query_mode == AMDGPU_RAS_DIRECT_ERROR_QUERY) {
+			if (adev->umc.ras && adev->umc.ras->ras_block.hw_ops &&
+			    adev->umc.ras->ras_block.hw_ops->query_ras_error_count)
+				adev->umc.ras->ras_block.hw_ops->query_ras_error_count(adev,
+								ras_error_status);
+
+			if (adev->umc.ras && adev->umc.ras->ras_block.hw_ops &&
+			    adev->umc.ras->ras_block.hw_ops->query_ras_error_address &&
+			    adev->umc.max_ras_err_cnt_per_query) {
+				err_data->err_addr =
+					kcalloc(adev->umc.max_ras_err_cnt_per_query,
+						sizeof(struct eeprom_table_record), GFP_KERNEL);
+
+				/* still call query_ras_error_address to clear error status
+				 * even NOMEM error is encountered
+				 */
+				if (!err_data->err_addr)
+					dev_warn(adev->dev,
+						"Failed to alloc memory for umc error address record!\n");
+				else
+					err_data->err_addr_len =
+						adev->umc.max_ras_err_cnt_per_query;
+
+				/* umc query_ras_error_address is also responsible for clearing
+				 * error status
+				 */
+				adev->umc.ras->ras_block.hw_ops->query_ras_error_address(adev,
+								ras_error_status);
+			}
+		} else if (error_query_mode == AMDGPU_RAS_FIRMWARE_ERROR_QUERY ||
+		    (!ret && error_query_mode == AMDGPU_RAS_DIRECT_ERROR_QUERY)) {
+			if (adev->umc.ras &&
+			    adev->umc.ras->ecc_info_query_ras_error_count)
+				adev->umc.ras->ecc_info_query_ras_error_count(adev,
+								ras_error_status);
+
+			if (adev->umc.ras &&
+			    adev->umc.ras->ecc_info_query_ras_error_address &&
+			    adev->umc.max_ras_err_cnt_per_query) {
+				err_data->err_addr =
+					kcalloc(adev->umc.max_ras_err_cnt_per_query,
+						sizeof(struct eeprom_table_record), GFP_KERNEL);
+
+				/* still call query_ras_error_address to clear error status
+				 * even NOMEM error is encountered
+				 */
+				if (!err_data->err_addr)
+					dev_warn(adev->dev,
+						"Failed to alloc memory for umc error address record!\n");
+				else
+					err_data->err_addr_len =
+						adev->umc.max_ras_err_cnt_per_query;
+
+				/* umc query_ras_error_address is also responsible for clearing
+				 * error status
+				 */
+				adev->umc.ras->ecc_info_query_ras_error_address(adev,
+								ras_error_status);
+			}
+		}
+	} else {
+		if (!amdgpu_ras_eeprom_update_record_num(control)) {
+			err_data->err_addr_cnt = err_data->de_count =
+				control->ras_num_recs -	control->ras_num_recs_old;
+			amdgpu_ras_eeprom_read_idx(control, err_data->err_addr,
+				control->ras_num_recs_old, err_data->de_count);
+		}
 	}
+
+	/* only uncorrectable error needs gpu reset */
+	if (err_data->ue_count || err_data->de_count) {
+		err_count = err_data->ue_count + err_data->de_count;
+		if ((amdgpu_bad_page_threshold != 0) &&
+			err_data->err_addr_cnt) {
+			amdgpu_ras_add_bad_pages(adev, err_data->err_addr,
+				err_data->err_addr_cnt, amdgpu_ras_smu_eeprom_supported(adev));
+			amdgpu_ras_save_bad_pages(adev, &err_count);
+
+			amdgpu_dpm_send_hbm_bad_pages_num(adev,
+					con->eeprom_control.ras_num_bad_pages);
+
+			if (con->update_channel_flag == true) {
+				amdgpu_dpm_send_hbm_bad_channel_flag(adev, con->eeprom_control.bad_channel_bitmap);
+				con->update_channel_flag = false;
+			}
+		}
+	}
+
+	kfree(err_data->err_addr);
+	err_data->err_addr = NULL;
+
+	mutex_unlock(&con->page_retirement_lock);
 }
 
-int amdgpu_umc_process_ras_data_cb(struct amdgpu_device *adev,
+static int amdgpu_umc_do_page_retirement(struct amdgpu_device *adev,
 		void *ras_error_status,
-		struct amdgpu_iv_entry *entry)
+		struct amdgpu_iv_entry *entry,
+		uint32_t reset)
 {
 	struct ras_err_data *err_data = (struct ras_err_data *)ras_error_status;
 	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
 
 	kgd2kfd_set_sram_ecc_flag(adev->kfd.dev);
-	if (adev->umc.ras_funcs &&
-	    adev->umc.ras_funcs->query_ras_error_count)
-	    adev->umc.ras_funcs->query_ras_error_count(adev, ras_error_status);
-
-	if (adev->umc.ras_funcs &&
-	    adev->umc.ras_funcs->query_ras_error_address &&
-	    adev->umc.max_ras_err_cnt_per_query) {
-		err_data->err_addr =
-			kcalloc(adev->umc.max_ras_err_cnt_per_query,
-				sizeof(struct eeprom_table_record), GFP_KERNEL);
+	amdgpu_umc_handle_bad_pages(adev, ras_error_status);
+
+	if ((err_data->ue_count || err_data->de_count) &&
+	    (reset || amdgpu_ras_is_rma(adev))) {
+		con->gpu_reset_flags |= reset;
+		amdgpu_ras_reset_gpu(adev);
+	}
 
-		/* still call query_ras_error_address to clear error status
-		 * even NOMEM error is encountered
-		 */
-		if(!err_data->err_addr)
-			dev_warn(adev->dev, "Failed to alloc memory for "
-					"umc error address record!\n");
+	return AMDGPU_RAS_SUCCESS;
+}
+
+int amdgpu_umc_pasid_poison_handler(struct amdgpu_device *adev,
+			enum amdgpu_ras_block block, uint16_t pasid,
+			pasid_notify pasid_fn, void *data, uint32_t reset)
+{
+	int ret = AMDGPU_RAS_SUCCESS;
 
-		/* umc query_ras_error_address is also responsible for clearing
-		 * error status
-		 */
-		adev->umc.ras_funcs->query_ras_error_address(adev, ras_error_status);
+	if (adev->gmc.xgmi.connected_to_cpu ||
+		adev->gmc.is_app_apu) {
+		if (reset) {
+			/* MCA poison handler is only responsible for GPU reset,
+			 * let MCA notifier do page retirement.
+			 */
+			kgd2kfd_set_sram_ecc_flag(adev->kfd.dev);
+			amdgpu_ras_reset_gpu(adev);
+		}
+		return ret;
 	}
 
-	/* only uncorrectable error needs gpu reset */
-	if (err_data->ue_count) {
-		dev_info(adev->dev, "%ld uncorrectable hardware errors "
-				"detected in UMC block\n",
-				err_data->ue_count);
+	if (!amdgpu_sriov_vf(adev)) {
+		if (amdgpu_ip_version(adev, UMC_HWIP, 0) < IP_VERSION(12, 0, 0)) {
+			struct ras_err_data err_data;
+			struct ras_common_if head = {
+				.block = AMDGPU_RAS_BLOCK__UMC,
+			};
+			struct ras_manager *obj = amdgpu_ras_find_obj(adev, &head);
 
-		if ((amdgpu_bad_page_threshold != 0) &&
-			err_data->err_addr_cnt) {
-			amdgpu_ras_add_bad_pages(adev, err_data->err_addr,
-						err_data->err_addr_cnt);
-			amdgpu_ras_save_bad_pages(adev);
+			ret = amdgpu_ras_error_data_init(&err_data);
+			if (ret)
+				return ret;
+
+			ret = amdgpu_umc_do_page_retirement(adev, &err_data, NULL, reset);
+
+			if (ret == AMDGPU_RAS_SUCCESS && obj) {
+				obj->err_data.ue_count += err_data.ue_count;
+				obj->err_data.ce_count += err_data.ce_count;
+				obj->err_data.de_count += err_data.de_count;
+			}
+
+			amdgpu_ras_error_data_fini(&err_data);
+		} else if (amdgpu_uniras_enabled(adev)) {
+			struct ras_ih_info ih_info = {0};
 
-			if (adev->smu.ppt_funcs && adev->smu.ppt_funcs->send_hbm_bad_pages_num)
-				adev->smu.ppt_funcs->send_hbm_bad_pages_num(&adev->smu, con->eeprom_control.ras_num_recs);
+			ih_info.block = block;
+			ih_info.pasid = pasid;
+			ih_info.reset = reset;
+			ih_info.pasid_fn = pasid_fn;
+			ih_info.data = data;
+			amdgpu_ras_mgr_handle_consumer_interrupt(adev, &ih_info);
+		} else {
+			struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+			int ret;
+
+			ret = amdgpu_ras_put_poison_req(adev,
+				block, pasid, pasid_fn, data, reset);
+			if (!ret) {
+				atomic_inc(&con->page_retirement_req_cnt);
+				atomic_inc(&con->poison_consumption_count);
+				wake_up(&con->page_retirement_wq);
+			}
 		}
+	} else {
+		if (adev->virt.ops && adev->virt.ops->ras_poison_handler)
+			adev->virt.ops->ras_poison_handler(adev, block);
+		else
+			dev_warn(adev->dev,
+				"No ras_poison_handler interface in SRIOV!\n");
+	}
 
-		amdgpu_ras_reset_gpu(adev);
+	return ret;
+}
+
+int amdgpu_umc_poison_handler(struct amdgpu_device *adev,
+			enum amdgpu_ras_block block, uint32_t reset)
+{
+	return amdgpu_umc_pasid_poison_handler(adev,
+				block, 0, NULL, NULL, reset);
+}
+
+int amdgpu_umc_process_ras_data_cb(struct amdgpu_device *adev,
+		void *ras_error_status,
+		struct amdgpu_iv_entry *entry)
+{
+	return amdgpu_umc_do_page_retirement(adev, ras_error_status, entry,
+				AMDGPU_RAS_GPU_RESET_MODE1_RESET);
+}
+
+int amdgpu_umc_ras_sw_init(struct amdgpu_device *adev)
+{
+	int err;
+	struct amdgpu_umc_ras *ras;
+
+	if (!adev->umc.ras)
+		return 0;
+
+	ras = adev->umc.ras;
+
+	err = amdgpu_ras_register_ras_block(adev, &ras->ras_block);
+	if (err) {
+		dev_err(adev->dev, "Failed to register umc ras block!\n");
+		return err;
 	}
 
-	kfree(err_data->err_addr);
-	return AMDGPU_RAS_SUCCESS;
+	strcpy(adev->umc.ras->ras_block.ras_comm.name, "umc");
+	ras->ras_block.ras_comm.block = AMDGPU_RAS_BLOCK__UMC;
+	ras->ras_block.ras_comm.type = AMDGPU_RAS_ERROR__MULTI_UNCORRECTABLE;
+	adev->umc.ras_if = &ras->ras_block.ras_comm;
+
+	if (!ras->ras_block.ras_late_init)
+		ras->ras_block.ras_late_init = amdgpu_umc_ras_late_init;
+
+	if (!ras->ras_block.ras_cb)
+		ras->ras_block.ras_cb = amdgpu_umc_process_ras_data_cb;
+
+	return 0;
+}
+
+int amdgpu_umc_ras_late_init(struct amdgpu_device *adev, struct ras_common_if *ras_block)
+{
+	int r;
+
+	r = amdgpu_ras_block_late_init(adev, ras_block);
+	if (r)
+		return r;
+
+	if (amdgpu_sriov_vf(adev))
+		return r;
+
+	if (amdgpu_ras_is_supported(adev, ras_block->block)) {
+		r = amdgpu_irq_get(adev, &adev->gmc.ecc_irq, 0);
+		if (r)
+			goto late_fini;
+	}
+
+	/* ras init of specific umc version */
+	if (adev->umc.ras &&
+	    adev->umc.ras->err_cnt_init)
+		adev->umc.ras->err_cnt_init(adev);
+
+	return 0;
+
+late_fini:
+	amdgpu_ras_block_late_fini(adev, ras_block);
+	return r;
 }
 
 int amdgpu_umc_process_ecc_irq(struct amdgpu_device *adev,
@@ -160,3 +397,231 @@ int amdgpu_umc_process_ecc_irq(struct amdgpu_device *adev,
 	amdgpu_ras_interrupt_dispatch(adev, &ih_data);
 	return 0;
 }
+
+int amdgpu_umc_fill_error_record(struct ras_err_data *err_data,
+		uint64_t err_addr,
+		uint64_t retired_page,
+		uint32_t channel_index,
+		uint32_t umc_inst)
+{
+	struct eeprom_table_record *err_rec;
+
+	if (!err_data ||
+	    !err_data->err_addr ||
+	    (err_data->err_addr_cnt >= err_data->err_addr_len))
+		return -EINVAL;
+
+	err_rec = &err_data->err_addr[err_data->err_addr_cnt];
+
+	err_rec->address = err_addr;
+	/* page frame address is saved */
+	err_rec->retired_page = retired_page >> AMDGPU_GPU_PAGE_SHIFT;
+	err_rec->ts = (uint64_t)ktime_get_real_seconds();
+	err_rec->err_type = AMDGPU_RAS_EEPROM_ERR_NON_RECOVERABLE;
+	err_rec->cu = 0;
+	err_rec->mem_channel = channel_index;
+	err_rec->mcumc_id = umc_inst;
+
+	err_data->err_addr_cnt++;
+
+	return 0;
+}
+
+static int amdgpu_umc_loop_all_aid(struct amdgpu_device *adev, umc_func func,
+				   void *data)
+{
+	uint32_t umc_node_inst;
+	uint32_t node_inst;
+	uint32_t umc_inst;
+	uint32_t ch_inst;
+	int ret;
+
+	/*
+	 * This loop is done based on the following -
+	 * umc.active mask = mask of active umc instances across all nodes
+	 * umc.umc_inst_num = maximum number of umc instancess per node
+	 * umc.node_inst_num = maximum number of node instances
+	 * Channel instances are not assumed to be harvested.
+	 */
+	dev_dbg(adev->dev, "active umcs :%lx umc_inst per node: %d",
+		adev->umc.active_mask, adev->umc.umc_inst_num);
+	for_each_set_bit(umc_node_inst, &(adev->umc.active_mask),
+			 adev->umc.node_inst_num * adev->umc.umc_inst_num) {
+		node_inst = umc_node_inst / adev->umc.umc_inst_num;
+		umc_inst = umc_node_inst % adev->umc.umc_inst_num;
+		LOOP_UMC_CH_INST(ch_inst) {
+			dev_dbg(adev->dev,
+				"node_inst :%d umc_inst: %d ch_inst: %d",
+				node_inst, umc_inst, ch_inst);
+			ret = func(adev, node_inst, umc_inst, ch_inst, data);
+			if (ret) {
+				dev_err(adev->dev,
+					"Node %d umc %d ch %d func returns %d\n",
+					node_inst, umc_inst, ch_inst, ret);
+				return ret;
+			}
+		}
+	}
+
+	return 0;
+}
+
+int amdgpu_umc_loop_channels(struct amdgpu_device *adev,
+			umc_func func, void *data)
+{
+	uint32_t node_inst       = 0;
+	uint32_t umc_inst        = 0;
+	uint32_t ch_inst         = 0;
+	int ret = 0;
+
+	if (adev->aid_mask)
+		return amdgpu_umc_loop_all_aid(adev, func, data);
+
+	if (adev->umc.node_inst_num) {
+		LOOP_UMC_EACH_NODE_INST_AND_CH(node_inst, umc_inst, ch_inst) {
+			ret = func(adev, node_inst, umc_inst, ch_inst, data);
+			if (ret) {
+				dev_err(adev->dev, "Node %d umc %d ch %d func returns %d\n",
+					node_inst, umc_inst, ch_inst, ret);
+				return ret;
+			}
+		}
+	} else {
+		LOOP_UMC_INST_AND_CH(umc_inst, ch_inst) {
+			ret = func(adev, 0, umc_inst, ch_inst, data);
+			if (ret) {
+				dev_err(adev->dev, "Umc %d ch %d func returns %d\n",
+					umc_inst, ch_inst, ret);
+				return ret;
+			}
+		}
+	}
+
+	return 0;
+}
+
+int amdgpu_umc_update_ecc_status(struct amdgpu_device *adev,
+				uint64_t status, uint64_t ipid, uint64_t addr)
+{
+	if (adev->umc.ras->update_ecc_status)
+		return adev->umc.ras->update_ecc_status(adev,
+					status, ipid, addr);
+	return 0;
+}
+
+int amdgpu_umc_logs_ecc_err(struct amdgpu_device *adev,
+		struct radix_tree_root *ecc_tree, struct ras_ecc_err *ecc_err)
+{
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+	struct ras_ecc_log_info *ecc_log;
+	int ret;
+
+	ecc_log = &con->umc_ecc_log;
+
+	mutex_lock(&ecc_log->lock);
+	ret = radix_tree_insert(ecc_tree, ecc_err->pa_pfn, ecc_err);
+	if (!ret)
+		radix_tree_tag_set(ecc_tree,
+			ecc_err->pa_pfn, UMC_ECC_NEW_DETECTED_TAG);
+	mutex_unlock(&ecc_log->lock);
+
+	return ret;
+}
+
+int amdgpu_umc_pages_in_a_row(struct amdgpu_device *adev,
+			struct ras_err_data *err_data, uint64_t pa_addr)
+{
+	struct ta_ras_query_address_output addr_out;
+
+	/* reinit err_data */
+	err_data->err_addr_cnt = 0;
+	err_data->err_addr_len = adev->umc.retire_unit;
+
+	addr_out.pa.pa = pa_addr;
+	if (adev->umc.ras && adev->umc.ras->convert_ras_err_addr)
+		return adev->umc.ras->convert_ras_err_addr(adev, err_data, NULL,
+				&addr_out, false);
+	else
+		return -EINVAL;
+}
+
+int amdgpu_umc_lookup_bad_pages_in_a_row(struct amdgpu_device *adev,
+			uint64_t pa_addr, uint64_t *pfns, int len)
+{
+	int i, ret;
+	struct ras_err_data err_data;
+
+	err_data.err_addr = kcalloc(adev->umc.retire_unit,
+				sizeof(struct eeprom_table_record), GFP_KERNEL);
+	if (!err_data.err_addr) {
+		dev_warn(adev->dev, "Failed to alloc memory in bad page lookup!\n");
+		return 0;
+	}
+
+	ret = amdgpu_umc_pages_in_a_row(adev, &err_data, pa_addr);
+	if (ret)
+		goto out;
+
+	for (i = 0; i < adev->umc.retire_unit; i++) {
+		if (i >= len)
+			goto out;
+
+		pfns[i] = err_data.err_addr[i].retired_page;
+	}
+	ret = i;
+	adev->umc.err_addr_cnt = err_data.err_addr_cnt;
+
+out:
+	kfree(err_data.err_addr);
+	return ret;
+}
+
+int amdgpu_umc_mca_to_addr(struct amdgpu_device *adev,
+			uint64_t err_addr, uint32_t ch, uint32_t umc,
+			uint32_t node, uint32_t socket,
+			struct ta_ras_query_address_output *addr_out, bool dump_addr)
+{
+	struct ta_ras_query_address_input addr_in;
+	int ret;
+
+	memset(&addr_in, 0, sizeof(addr_in));
+	addr_in.ma.err_addr = err_addr;
+	addr_in.ma.ch_inst = ch;
+	addr_in.ma.umc_inst = umc;
+	addr_in.ma.node_inst = node;
+	addr_in.ma.socket_id = socket;
+
+	if (adev->umc.ras && adev->umc.ras->convert_ras_err_addr) {
+		ret = adev->umc.ras->convert_ras_err_addr(adev, NULL, &addr_in,
+				addr_out, dump_addr);
+		if (ret)
+			return ret;
+	} else {
+		return 0;
+	}
+
+	return 0;
+}
+
+int amdgpu_umc_pa2mca(struct amdgpu_device *adev,
+		uint64_t pa, uint64_t *mca, enum amdgpu_memory_partition nps)
+{
+	struct ta_ras_query_address_input addr_in;
+	struct ta_ras_query_address_output addr_out;
+	int ret;
+
+	/* nps: the pa belongs to */
+	addr_in.pa.pa = pa | ((uint64_t)nps << 58);
+	addr_in.addr_type = TA_RAS_PA_TO_MCA;
+	ret = psp_ras_query_address(&adev->psp, &addr_in, &addr_out);
+	if (ret) {
+		dev_warn(adev->dev, "Failed to query RAS MCA address for 0x%llx",
+			pa);
+
+		return ret;
+	}
+
+	*mca = addr_out.ma.err_addr;
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h
index 1f5fe2315236..28dff750c47e 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h
@@ -20,7 +20,8 @@
  */
 #ifndef __AMDGPU_UMC_H__
 #define __AMDGPU_UMC_H__
-
+#include "amdgpu_ras.h"
+#include "amdgpu_mca.h"
 /*
  * (addr / 256) * 4096, the higher 26 bits in ErrorAddr
  * is the index of 4KB block
@@ -31,6 +32,11 @@
  * is the index of 8KB block
  */
 #define ADDR_OF_8KB_BLOCK(addr)			(((addr) & ~0xffULL) << 5)
+/*
+ * (addr / 256) * 32768, the higher 26 bits in ErrorAddr
+ * is the index of 8KB block
+ */
+#define ADDR_OF_32KB_BLOCK(addr)			(((addr) & ~0xffULL) << 7)
 /* channel index is the index of 256B block */
 #define ADDR_OF_256B_BLOCK(channel_index)	((channel_index) << 8)
 /* offset in 256B block */
@@ -40,15 +46,75 @@
 #define LOOP_UMC_CH_INST(ch_inst) for ((ch_inst) = 0; (ch_inst) < adev->umc.channel_inst_num; (ch_inst)++)
 #define LOOP_UMC_INST_AND_CH(umc_inst, ch_inst) LOOP_UMC_INST((umc_inst)) LOOP_UMC_CH_INST((ch_inst))
 
-struct amdgpu_umc_ras_funcs {
+#define LOOP_UMC_NODE_INST(node_inst) \
+		for_each_set_bit((node_inst), &(adev->umc.active_mask), adev->umc.node_inst_num)
+
+#define LOOP_UMC_EACH_NODE_INST_AND_CH(node_inst, umc_inst, ch_inst) \
+		LOOP_UMC_NODE_INST((node_inst)) LOOP_UMC_INST_AND_CH((umc_inst), (ch_inst))
+
+/* Page retirement tag */
+#define UMC_ECC_NEW_DETECTED_TAG       0x1
+/*
+ * a flag to indicate v2 of channel index stored in eeprom
+ *
+ * v1 (legacy way): store channel index within a umc instance in eeprom
+ *    range in UMC v12: 0 ~ 7
+ * v2: store global channel index in eeprom
+ *    range in UMC v12: 0 ~ 127
+ *
+ * NOTE: it's better to store it in eeprom_table_record.mem_channel,
+ * but there is only 8 bits in mem_channel, and the channel number may
+ * increase in the future, we decide to save it in
+ * eeprom_table_record.retired_page. retired_page is useless in v2,
+ * we depend on eeprom_table_record.address instead of retired_page in v2.
+ * Only 48 bits are saved on eeprom, use bit 47 here.
+ */
+#define UMC_CHANNEL_IDX_V2	BIT_ULL(47)
+
+/*
+ * save nps value to eeprom_table_record.retired_page[47:40],
+ * the channel index flag above will be retired.
+ */
+#define UMC_NPS_SHIFT 40
+#define UMC_NPS_MASK 0xffULL
+
+/* three column bits and one row bit in MCA address flip
+ * in bad page retirement
+ */
+#define RETIRE_FLIP_BITS_NUM 4
+
+struct amdgpu_umc_flip_bits {
+	uint32_t flip_bits_in_pa[RETIRE_FLIP_BITS_NUM];
+	uint32_t flip_row_bit;
+	uint32_t r13_in_pa;
+	uint32_t bit_num;
+};
+
+typedef int (*umc_func)(struct amdgpu_device *adev, uint32_t node_inst,
+			uint32_t umc_inst, uint32_t ch_inst, void *data);
+
+struct amdgpu_umc_ras {
+	struct amdgpu_ras_block_object ras_block;
 	void (*err_cnt_init)(struct amdgpu_device *adev);
-	int (*ras_late_init)(struct amdgpu_device *adev);
-	void (*ras_fini)(struct amdgpu_device *adev);
-	void (*query_ras_error_count)(struct amdgpu_device *adev,
+	bool (*query_ras_poison_mode)(struct amdgpu_device *adev);
+	void (*ecc_info_query_ras_error_count)(struct amdgpu_device *adev,
 				      void *ras_error_status);
-	void (*query_ras_error_address)(struct amdgpu_device *adev,
+	void (*ecc_info_query_ras_error_address)(struct amdgpu_device *adev,
 					void *ras_error_status);
-	bool (*query_ras_poison_mode)(struct amdgpu_device *adev);
+	bool (*check_ecc_err_status)(struct amdgpu_device *adev,
+			enum amdgpu_mca_error_type type, void *ras_error_status);
+	int (*update_ecc_status)(struct amdgpu_device *adev,
+			uint64_t status, uint64_t ipid, uint64_t addr);
+	int (*convert_ras_err_addr)(struct amdgpu_device *adev,
+			struct ras_err_data *err_data,
+			struct ta_ras_query_address_input *addr_in,
+			struct ta_ras_query_address_output *addr_out,
+			bool dump_addr);
+	uint32_t (*get_die_id_from_pa)(struct amdgpu_device *adev,
+			uint64_t mca_addr, uint64_t retired_page);
+	void (*get_retire_flip_bits)(struct amdgpu_device *adev);
+	void (*mca_ipid_parse)(struct amdgpu_device *adev, uint64_t ipid,
+			uint32_t *did, uint32_t *ch, uint32_t *umc_inst, uint32_t *sid);
 };
 
 struct amdgpu_umc_funcs {
@@ -62,22 +128,69 @@ struct amdgpu_umc {
 	uint32_t channel_inst_num;
 	/* number of umc instance with memory map register access */
 	uint32_t umc_inst_num;
+
+	/* Total number of umc node instance including harvest one */
+	uint32_t node_inst_num;
+
 	/* UMC regiser per channel offset */
 	uint32_t channel_offs;
+	/* how many pages are retired in one UE */
+	uint32_t retire_unit;
 	/* channel index table of interleaved memory */
 	const uint32_t *channel_idx_tbl;
 	struct ras_common_if *ras_if;
 
 	const struct amdgpu_umc_funcs *funcs;
-	const struct amdgpu_umc_ras_funcs *ras_funcs;
+	struct amdgpu_umc_ras *ras;
+
+	/* active mask for umc node instance */
+	unsigned long active_mask;
+
+	struct amdgpu_umc_flip_bits flip_bits;
+
+	unsigned long err_addr_cnt;
 };
 
-int amdgpu_umc_ras_late_init(struct amdgpu_device *adev);
-void amdgpu_umc_ras_fini(struct amdgpu_device *adev);
-int amdgpu_umc_process_ras_data_cb(struct amdgpu_device *adev,
-		void *ras_error_status,
-		struct amdgpu_iv_entry *entry);
+int amdgpu_umc_ras_sw_init(struct amdgpu_device *adev);
+int amdgpu_umc_ras_late_init(struct amdgpu_device *adev, struct ras_common_if *ras_block);
+int amdgpu_umc_poison_handler(struct amdgpu_device *adev,
+			enum amdgpu_ras_block block, uint32_t reset);
+int amdgpu_umc_pasid_poison_handler(struct amdgpu_device *adev,
+			enum amdgpu_ras_block block, uint16_t pasid,
+			pasid_notify pasid_fn, void *data, uint32_t reset);
 int amdgpu_umc_process_ecc_irq(struct amdgpu_device *adev,
 		struct amdgpu_irq_src *source,
 		struct amdgpu_iv_entry *entry);
+int amdgpu_umc_fill_error_record(struct ras_err_data *err_data,
+		uint64_t err_addr,
+		uint64_t retired_page,
+		uint32_t channel_index,
+		uint32_t umc_inst);
+
+int amdgpu_umc_process_ras_data_cb(struct amdgpu_device *adev,
+		void *ras_error_status,
+		struct amdgpu_iv_entry *entry);
+int amdgpu_umc_page_retirement_mca(struct amdgpu_device *adev,
+			uint64_t err_addr, uint32_t ch_inst, uint32_t umc_inst);
+
+int amdgpu_umc_loop_channels(struct amdgpu_device *adev,
+			umc_func func, void *data);
+
+int amdgpu_umc_update_ecc_status(struct amdgpu_device *adev,
+				uint64_t status, uint64_t ipid, uint64_t addr);
+int amdgpu_umc_logs_ecc_err(struct amdgpu_device *adev,
+		struct radix_tree_root *ecc_tree, struct ras_ecc_err *ecc_err);
+
+void amdgpu_umc_handle_bad_pages(struct amdgpu_device *adev,
+			void *ras_error_status);
+int amdgpu_umc_pages_in_a_row(struct amdgpu_device *adev,
+			struct ras_err_data *err_data, uint64_t pa_addr);
+int amdgpu_umc_lookup_bad_pages_in_a_row(struct amdgpu_device *adev,
+			uint64_t pa_addr, uint64_t *pfns, int len);
+int amdgpu_umc_mca_to_addr(struct amdgpu_device *adev,
+			uint64_t err_addr, uint32_t ch, uint32_t umc,
+			uint32_t node, uint32_t socket,
+			struct ta_ras_query_address_output *addr_out, bool dump_addr);
+int amdgpu_umc_pa2mca(struct amdgpu_device *adev,
+		uint64_t pa, uint64_t *mca, enum amdgpu_memory_partition nps);
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_umr.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_umr.h
index 919d9d401750..5b27fc41ffbf 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_umr.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_umr.h
@@ -35,17 +35,51 @@ struct amdgpu_debugfs_regs2_iocdata {
 	} srbm;
 };
 
+struct amdgpu_debugfs_regs2_iocdata_v2 {
+	__u32 use_srbm, use_grbm, pg_lock;
+	struct {
+		__u32 se, sh, instance;
+	} grbm;
+	struct {
+		__u32 me, pipe, queue, vmid;
+	} srbm;
+	u32 xcc_id;
+};
+
+struct amdgpu_debugfs_gprwave_iocdata {
+	u32 gpr_or_wave, se, sh, cu, wave, simd, xcc_id;
+	struct {
+		u32 thread, vpgr_or_sgpr;
+	} gpr;
+};
+
 /*
  * MMIO debugfs state data (per file* handle)
  */
 struct amdgpu_debugfs_regs2_data {
 	struct amdgpu_device *adev;
 	struct mutex lock;
-	struct amdgpu_debugfs_regs2_iocdata id;
+	struct amdgpu_debugfs_regs2_iocdata_v2 id;
+};
+
+struct amdgpu_debugfs_gprwave_data {
+	struct amdgpu_device *adev;
+	struct mutex lock;
+	struct amdgpu_debugfs_gprwave_iocdata id;
 };
 
 enum AMDGPU_DEBUGFS_REGS2_CMDS {
-	AMDGPU_DEBUGFS_REGS2_CMD_SET_STATE=0,
+	AMDGPU_DEBUGFS_REGS2_CMD_SET_STATE = 0,
+	AMDGPU_DEBUGFS_REGS2_CMD_SET_STATE_V2,
+};
+
+enum AMDGPU_DEBUGFS_GPRWAVE_CMDS {
+	AMDGPU_DEBUGFS_GPRWAVE_CMD_SET_STATE = 0,
 };
 
+//reg2 interface
 #define AMDGPU_DEBUGFS_REGS2_IOC_SET_STATE _IOWR(0x20, AMDGPU_DEBUGFS_REGS2_CMD_SET_STATE, struct amdgpu_debugfs_regs2_iocdata)
+#define AMDGPU_DEBUGFS_REGS2_IOC_SET_STATE_V2 _IOWR(0x20, AMDGPU_DEBUGFS_REGS2_CMD_SET_STATE_V2, struct amdgpu_debugfs_regs2_iocdata_v2)
+
+//gprwave interface
+#define AMDGPU_DEBUGFS_GPRWAVE_IOC_SET_STATE _IOWR(0x20, AMDGPU_DEBUGFS_GPRWAVE_CMD_SET_STATE, struct amdgpu_debugfs_gprwave_iocdata)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_umsch_mm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_umsch_mm.c
new file mode 100644
index 000000000000..cd707d70a0bf
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_umsch_mm.c
@@ -0,0 +1,550 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/firmware.h>
+#include <linux/module.h>
+#include <linux/debugfs.h>
+#include <drm/drm_exec.h>
+#include <drm/drm_drv.h>
+
+#include "amdgpu.h"
+#include "amdgpu_umsch_mm.h"
+#include "umsch_mm_v4_0.h"
+
+MODULE_FIRMWARE("amdgpu/umsch_mm_4_0_0.bin");
+
+int amdgpu_umsch_mm_submit_pkt(struct amdgpu_umsch_mm *umsch, void *pkt, int ndws)
+{
+	struct amdgpu_ring *ring = &umsch->ring;
+
+	if (amdgpu_ring_alloc(ring, ndws))
+		return -ENOMEM;
+
+	amdgpu_ring_write_multiple(ring, pkt, ndws);
+	amdgpu_ring_commit(ring);
+
+	return 0;
+}
+
+int amdgpu_umsch_mm_query_fence(struct amdgpu_umsch_mm *umsch)
+{
+	struct amdgpu_ring *ring = &umsch->ring;
+	struct amdgpu_device *adev = ring->adev;
+	int r;
+
+	r = amdgpu_fence_wait_polling(ring, ring->fence_drv.sync_seq, adev->usec_timeout);
+	if (r < 1) {
+		dev_err(adev->dev, "ring umsch timeout, emitted fence %u\n",
+			ring->fence_drv.sync_seq);
+		return -ETIMEDOUT;
+	}
+
+	return 0;
+}
+
+static void umsch_mm_ring_set_wptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_umsch_mm *umsch = (struct amdgpu_umsch_mm *)ring;
+	struct amdgpu_device *adev = ring->adev;
+
+	if (ring->use_doorbell)
+		WDOORBELL32(ring->doorbell_index, ring->wptr << 2);
+	else
+		WREG32(umsch->rb_wptr, ring->wptr << 2);
+}
+
+static u64 umsch_mm_ring_get_rptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_umsch_mm *umsch = (struct amdgpu_umsch_mm *)ring;
+	struct amdgpu_device *adev = ring->adev;
+
+	return RREG32(umsch->rb_rptr);
+}
+
+static u64 umsch_mm_ring_get_wptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_umsch_mm *umsch = (struct amdgpu_umsch_mm *)ring;
+	struct amdgpu_device *adev = ring->adev;
+
+	return RREG32(umsch->rb_wptr);
+}
+
+static const struct amdgpu_ring_funcs umsch_v4_0_ring_funcs = {
+	.type = AMDGPU_RING_TYPE_UMSCH_MM,
+	.align_mask = 0,
+	.nop = 0,
+	.support_64bit_ptrs = false,
+	.get_rptr = umsch_mm_ring_get_rptr,
+	.get_wptr = umsch_mm_ring_get_wptr,
+	.set_wptr = umsch_mm_ring_set_wptr,
+	.insert_nop = amdgpu_ring_insert_nop,
+};
+
+int amdgpu_umsch_mm_ring_init(struct amdgpu_umsch_mm *umsch)
+{
+	struct amdgpu_device *adev = container_of(umsch, struct amdgpu_device, umsch_mm);
+	struct amdgpu_ring *ring = &umsch->ring;
+
+	ring->vm_hub = AMDGPU_MMHUB0(0);
+	ring->use_doorbell = true;
+	ring->no_scheduler = true;
+	ring->doorbell_index = (AMDGPU_NAVI10_DOORBELL64_VCN0_1 << 1) + 6;
+
+	snprintf(ring->name, sizeof(ring->name), "umsch");
+
+	return amdgpu_ring_init(adev, ring, 1024, NULL, 0, AMDGPU_RING_PRIO_DEFAULT, NULL);
+}
+
+int amdgpu_umsch_mm_init_microcode(struct amdgpu_umsch_mm *umsch)
+{
+	const struct umsch_mm_firmware_header_v1_0 *umsch_mm_hdr;
+	struct amdgpu_device *adev = umsch->ring.adev;
+	const char *fw_name = NULL;
+	int r;
+
+	switch (amdgpu_ip_version(adev, VCN_HWIP, 0)) {
+	case IP_VERSION(4, 0, 5):
+	case IP_VERSION(4, 0, 6):
+		fw_name = "4_0_0";
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	r = amdgpu_ucode_request(adev, &adev->umsch_mm.fw, AMDGPU_UCODE_REQUIRED,
+				 "amdgpu/umsch_mm_%s.bin", fw_name);
+	if (r) {
+		release_firmware(adev->umsch_mm.fw);
+		adev->umsch_mm.fw = NULL;
+		return r;
+	}
+
+	umsch_mm_hdr = (const struct umsch_mm_firmware_header_v1_0 *)adev->umsch_mm.fw->data;
+
+	adev->umsch_mm.ucode_size = le32_to_cpu(umsch_mm_hdr->umsch_mm_ucode_size_bytes);
+	adev->umsch_mm.data_size = le32_to_cpu(umsch_mm_hdr->umsch_mm_ucode_data_size_bytes);
+
+	adev->umsch_mm.irq_start_addr =
+		le32_to_cpu(umsch_mm_hdr->umsch_mm_irq_start_addr_lo) |
+		((uint64_t)(le32_to_cpu(umsch_mm_hdr->umsch_mm_irq_start_addr_hi)) << 32);
+	adev->umsch_mm.uc_start_addr =
+		le32_to_cpu(umsch_mm_hdr->umsch_mm_uc_start_addr_lo) |
+		((uint64_t)(le32_to_cpu(umsch_mm_hdr->umsch_mm_uc_start_addr_hi)) << 32);
+	adev->umsch_mm.data_start_addr =
+		le32_to_cpu(umsch_mm_hdr->umsch_mm_data_start_addr_lo) |
+		((uint64_t)(le32_to_cpu(umsch_mm_hdr->umsch_mm_data_start_addr_hi)) << 32);
+
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
+		struct amdgpu_firmware_info *info;
+
+		info = &adev->firmware.ucode[AMDGPU_UCODE_ID_UMSCH_MM_UCODE];
+		info->ucode_id = AMDGPU_UCODE_ID_UMSCH_MM_UCODE;
+		info->fw = adev->umsch_mm.fw;
+		adev->firmware.fw_size +=
+			ALIGN(le32_to_cpu(umsch_mm_hdr->umsch_mm_ucode_size_bytes), PAGE_SIZE);
+
+		info = &adev->firmware.ucode[AMDGPU_UCODE_ID_UMSCH_MM_DATA];
+		info->ucode_id = AMDGPU_UCODE_ID_UMSCH_MM_DATA;
+		info->fw = adev->umsch_mm.fw;
+		adev->firmware.fw_size +=
+			ALIGN(le32_to_cpu(umsch_mm_hdr->umsch_mm_ucode_data_size_bytes), PAGE_SIZE);
+	}
+
+	return 0;
+}
+
+int amdgpu_umsch_mm_allocate_ucode_buffer(struct amdgpu_umsch_mm *umsch)
+{
+	const struct umsch_mm_firmware_header_v1_0 *umsch_mm_hdr;
+	struct amdgpu_device *adev = umsch->ring.adev;
+	const __le32 *fw_data;
+	uint32_t fw_size;
+	int r;
+
+	umsch_mm_hdr = (const struct umsch_mm_firmware_header_v1_0 *)
+		       adev->umsch_mm.fw->data;
+
+	fw_data = (const __le32 *)(adev->umsch_mm.fw->data +
+		  le32_to_cpu(umsch_mm_hdr->umsch_mm_ucode_offset_bytes));
+	fw_size = le32_to_cpu(umsch_mm_hdr->umsch_mm_ucode_size_bytes);
+
+	r = amdgpu_bo_create_reserved(adev, fw_size,
+				      4 * 1024, AMDGPU_GEM_DOMAIN_VRAM,
+				      &adev->umsch_mm.ucode_fw_obj,
+				      &adev->umsch_mm.ucode_fw_gpu_addr,
+				      (void **)&adev->umsch_mm.ucode_fw_ptr);
+	if (r) {
+		dev_err(adev->dev, "(%d) failed to create umsch_mm fw ucode bo\n", r);
+		return r;
+	}
+
+	memcpy(adev->umsch_mm.ucode_fw_ptr, fw_data, fw_size);
+
+	amdgpu_bo_kunmap(adev->umsch_mm.ucode_fw_obj);
+	amdgpu_bo_unreserve(adev->umsch_mm.ucode_fw_obj);
+	return 0;
+}
+
+int amdgpu_umsch_mm_allocate_ucode_data_buffer(struct amdgpu_umsch_mm *umsch)
+{
+	const struct umsch_mm_firmware_header_v1_0 *umsch_mm_hdr;
+	struct amdgpu_device *adev = umsch->ring.adev;
+	const __le32 *fw_data;
+	uint32_t fw_size;
+	int r;
+
+	umsch_mm_hdr = (const struct umsch_mm_firmware_header_v1_0 *)
+		       adev->umsch_mm.fw->data;
+
+	fw_data = (const __le32 *)(adev->umsch_mm.fw->data +
+		  le32_to_cpu(umsch_mm_hdr->umsch_mm_ucode_data_offset_bytes));
+	fw_size = le32_to_cpu(umsch_mm_hdr->umsch_mm_ucode_data_size_bytes);
+
+	r = amdgpu_bo_create_reserved(adev, fw_size,
+				      64 * 1024, AMDGPU_GEM_DOMAIN_VRAM,
+				      &adev->umsch_mm.data_fw_obj,
+				      &adev->umsch_mm.data_fw_gpu_addr,
+				      (void **)&adev->umsch_mm.data_fw_ptr);
+	if (r) {
+		dev_err(adev->dev, "(%d) failed to create umsch_mm fw data bo\n", r);
+		return r;
+	}
+
+	memcpy(adev->umsch_mm.data_fw_ptr, fw_data, fw_size);
+
+	amdgpu_bo_kunmap(adev->umsch_mm.data_fw_obj);
+	amdgpu_bo_unreserve(adev->umsch_mm.data_fw_obj);
+	return 0;
+}
+
+int amdgpu_umsch_mm_psp_execute_cmd_buf(struct amdgpu_umsch_mm *umsch)
+{
+	struct amdgpu_device *adev = umsch->ring.adev;
+	struct amdgpu_firmware_info ucode = {
+		.ucode_id = AMDGPU_UCODE_ID_UMSCH_MM_CMD_BUFFER,
+		.mc_addr = adev->umsch_mm.cmd_buf_gpu_addr,
+		.ucode_size = ((uintptr_t)adev->umsch_mm.cmd_buf_curr_ptr -
+			      (uintptr_t)adev->umsch_mm.cmd_buf_ptr),
+	};
+
+	return psp_execute_ip_fw_load(&adev->psp, &ucode);
+}
+
+static void umsch_mm_agdb_index_init(struct amdgpu_device *adev)
+{
+	uint32_t umsch_mm_agdb_start;
+	int i;
+
+	umsch_mm_agdb_start = adev->doorbell_index.max_assignment + 1;
+	umsch_mm_agdb_start = roundup(umsch_mm_agdb_start, 1024);
+	umsch_mm_agdb_start += (AMDGPU_NAVI10_DOORBELL64_VCN0_1 << 1);
+
+	for (i = 0; i < CONTEXT_PRIORITY_NUM_LEVELS; i++)
+		adev->umsch_mm.agdb_index[i] = umsch_mm_agdb_start + i;
+}
+
+static int umsch_mm_init(struct amdgpu_device *adev)
+{
+	int r;
+
+	adev->umsch_mm.vmid_mask_mm_vpe = 0xf00;
+	adev->umsch_mm.engine_mask = (1 << UMSCH_SWIP_ENGINE_TYPE_VPE);
+	adev->umsch_mm.vpe_hqd_mask = 0xfe;
+
+	r = amdgpu_device_wb_get(adev, &adev->umsch_mm.wb_index);
+	if (r) {
+		dev_err(adev->dev, "failed to alloc wb for umsch: %d\n", r);
+		return r;
+	}
+
+	adev->umsch_mm.sch_ctx_gpu_addr = adev->wb.gpu_addr +
+					  (adev->umsch_mm.wb_index * 4);
+
+	r = amdgpu_bo_create_kernel(adev, PAGE_SIZE, PAGE_SIZE,
+				    AMDGPU_GEM_DOMAIN_GTT,
+				    &adev->umsch_mm.cmd_buf_obj,
+				    &adev->umsch_mm.cmd_buf_gpu_addr,
+				    (void **)&adev->umsch_mm.cmd_buf_ptr);
+	if (r) {
+		dev_err(adev->dev, "failed to allocate cmdbuf bo %d\n", r);
+		amdgpu_device_wb_free(adev, adev->umsch_mm.wb_index);
+		return r;
+	}
+
+	r = amdgpu_bo_create_kernel(adev, AMDGPU_UMSCHFW_LOG_SIZE, PAGE_SIZE,
+				    AMDGPU_GEM_DOMAIN_VRAM |
+				    AMDGPU_GEM_DOMAIN_GTT,
+				    &adev->umsch_mm.dbglog_bo,
+				    &adev->umsch_mm.log_gpu_addr,
+				    &adev->umsch_mm.log_cpu_addr);
+	if (r) {
+		dev_err(adev->dev, "(%d) failed to allocate umsch debug bo\n", r);
+		return r;
+	}
+
+	mutex_init(&adev->umsch_mm.mutex_hidden);
+
+	umsch_mm_agdb_index_init(adev);
+
+	return 0;
+}
+
+
+static int umsch_mm_early_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	switch (amdgpu_ip_version(adev, VCN_HWIP, 0)) {
+	case IP_VERSION(4, 0, 5):
+	case IP_VERSION(4, 0, 6):
+		umsch_mm_v4_0_set_funcs(&adev->umsch_mm);
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	adev->umsch_mm.ring.funcs = &umsch_v4_0_ring_funcs;
+	umsch_mm_set_regs(&adev->umsch_mm);
+
+	return 0;
+}
+
+static int umsch_mm_late_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	if (amdgpu_in_reset(adev) || adev->in_s0ix || adev->in_suspend)
+		return 0;
+
+	return 0;
+}
+
+static int umsch_mm_sw_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int r;
+
+	r = umsch_mm_init(adev);
+	if (r)
+		return r;
+
+	amdgpu_umsch_fwlog_init(&adev->umsch_mm);
+	r = umsch_mm_ring_init(&adev->umsch_mm);
+	if (r)
+		return r;
+
+	r = umsch_mm_init_microcode(&adev->umsch_mm);
+	if (r)
+		return r;
+
+	return 0;
+}
+
+static int umsch_mm_sw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	release_firmware(adev->umsch_mm.fw);
+	adev->umsch_mm.fw = NULL;
+
+	amdgpu_ring_fini(&adev->umsch_mm.ring);
+
+	mutex_destroy(&adev->umsch_mm.mutex_hidden);
+
+	amdgpu_bo_free_kernel(&adev->umsch_mm.cmd_buf_obj,
+			      &adev->umsch_mm.cmd_buf_gpu_addr,
+			      (void **)&adev->umsch_mm.cmd_buf_ptr);
+
+	amdgpu_bo_free_kernel(&adev->umsch_mm.dbglog_bo,
+				    &adev->umsch_mm.log_gpu_addr,
+				    (void **)&adev->umsch_mm.log_cpu_addr);
+
+	amdgpu_device_wb_free(adev, adev->umsch_mm.wb_index);
+
+	return 0;
+}
+
+static int umsch_mm_hw_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int r;
+
+	r = umsch_mm_load_microcode(&adev->umsch_mm);
+	if (r)
+		return r;
+
+	umsch_mm_ring_start(&adev->umsch_mm);
+
+	r = umsch_mm_set_hw_resources(&adev->umsch_mm);
+	if (r)
+		return r;
+
+	return 0;
+}
+
+static int umsch_mm_hw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	umsch_mm_ring_stop(&adev->umsch_mm);
+
+	amdgpu_bo_free_kernel(&adev->umsch_mm.data_fw_obj,
+			      &adev->umsch_mm.data_fw_gpu_addr,
+			      (void **)&adev->umsch_mm.data_fw_ptr);
+
+	amdgpu_bo_free_kernel(&adev->umsch_mm.ucode_fw_obj,
+			      &adev->umsch_mm.ucode_fw_gpu_addr,
+			      (void **)&adev->umsch_mm.ucode_fw_ptr);
+	return 0;
+}
+
+static int umsch_mm_suspend(struct amdgpu_ip_block *ip_block)
+{
+	return umsch_mm_hw_fini(ip_block);
+}
+
+static int umsch_mm_resume(struct amdgpu_ip_block *ip_block)
+{
+	return umsch_mm_hw_init(ip_block);
+}
+
+void amdgpu_umsch_fwlog_init(struct amdgpu_umsch_mm *umsch_mm)
+{
+#if defined(CONFIG_DEBUG_FS)
+	void *fw_log_cpu_addr = umsch_mm->log_cpu_addr;
+	volatile struct amdgpu_umsch_fwlog *log_buf = fw_log_cpu_addr;
+
+	log_buf->header_size = sizeof(struct amdgpu_umsch_fwlog);
+	log_buf->buffer_size = AMDGPU_UMSCHFW_LOG_SIZE;
+	log_buf->rptr = log_buf->header_size;
+	log_buf->wptr = log_buf->header_size;
+	log_buf->wrapped = 0;
+#endif
+}
+
+/*
+ * debugfs for mapping umsch firmware log buffer.
+ */
+#if defined(CONFIG_DEBUG_FS)
+static ssize_t amdgpu_debugfs_umsch_fwlog_read(struct file *f, char __user *buf,
+					     size_t size, loff_t *pos)
+{
+	struct amdgpu_umsch_mm *umsch_mm;
+	void *log_buf;
+	volatile struct amdgpu_umsch_fwlog *plog;
+	unsigned int read_pos, write_pos, available, i, read_bytes = 0;
+	unsigned int read_num[2] = {0};
+
+	umsch_mm = file_inode(f)->i_private;
+	if (!umsch_mm)
+		return -ENODEV;
+
+	if (!umsch_mm->log_cpu_addr)
+		return -EFAULT;
+
+	log_buf = umsch_mm->log_cpu_addr;
+
+	plog = (volatile struct amdgpu_umsch_fwlog *)log_buf;
+	read_pos = plog->rptr;
+	write_pos = plog->wptr;
+
+	if (read_pos > AMDGPU_UMSCHFW_LOG_SIZE || write_pos > AMDGPU_UMSCHFW_LOG_SIZE)
+		return -EFAULT;
+
+	if (!size || (read_pos == write_pos))
+		return 0;
+
+	if (write_pos > read_pos) {
+		available = write_pos - read_pos;
+		read_num[0] = min_t(size_t, size, available);
+	} else {
+		read_num[0] = AMDGPU_UMSCHFW_LOG_SIZE - read_pos;
+		available = read_num[0] + write_pos - plog->header_size;
+		if (size > available)
+			read_num[1] = write_pos - plog->header_size;
+		else if (size > read_num[0])
+			read_num[1] = size - read_num[0];
+		else
+			read_num[0] = size;
+	}
+
+	for (i = 0; i < 2; i++) {
+		if (read_num[i]) {
+			if (read_pos == AMDGPU_UMSCHFW_LOG_SIZE)
+				read_pos = plog->header_size;
+			if (read_num[i] == copy_to_user((buf + read_bytes),
+							(log_buf + read_pos), read_num[i]))
+				return -EFAULT;
+
+			read_bytes += read_num[i];
+			read_pos += read_num[i];
+		}
+	}
+
+	plog->rptr = read_pos;
+	*pos += read_bytes;
+	return read_bytes;
+}
+
+static const struct file_operations amdgpu_debugfs_umschfwlog_fops = {
+	.owner = THIS_MODULE,
+	.read = amdgpu_debugfs_umsch_fwlog_read,
+	.llseek = default_llseek
+};
+#endif
+
+void amdgpu_debugfs_umsch_fwlog_init(struct amdgpu_device *adev,
+			struct amdgpu_umsch_mm *umsch_mm)
+{
+#if defined(CONFIG_DEBUG_FS)
+	struct drm_minor *minor = adev_to_drm(adev)->primary;
+	struct dentry *root = minor->debugfs_root;
+	char name[32];
+
+	sprintf(name, "amdgpu_umsch_fwlog");
+	debugfs_create_file_size(name, S_IFREG | 0444, root, umsch_mm,
+				 &amdgpu_debugfs_umschfwlog_fops,
+				 AMDGPU_UMSCHFW_LOG_SIZE);
+#endif
+}
+
+static const struct amd_ip_funcs umsch_mm_v4_0_ip_funcs = {
+	.name = "umsch_mm_v4_0",
+	.early_init = umsch_mm_early_init,
+	.late_init = umsch_mm_late_init,
+	.sw_init = umsch_mm_sw_init,
+	.sw_fini = umsch_mm_sw_fini,
+	.hw_init = umsch_mm_hw_init,
+	.hw_fini = umsch_mm_hw_fini,
+	.suspend = umsch_mm_suspend,
+	.resume = umsch_mm_resume,
+};
+
+const struct amdgpu_ip_block_version umsch_mm_v4_0_ip_block = {
+	.type = AMD_IP_BLOCK_TYPE_UMSCH_MM,
+	.major = 4,
+	.minor = 0,
+	.rev = 0,
+	.funcs = &umsch_mm_v4_0_ip_funcs,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_umsch_mm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_umsch_mm.h
new file mode 100644
index 000000000000..2c771a753778
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_umsch_mm.h
@@ -0,0 +1,246 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __AMDGPU_UMSCH_MM_H__
+#define __AMDGPU_UMSCH_MM_H__
+
+enum UMSCH_SWIP_ENGINE_TYPE {
+	UMSCH_SWIP_ENGINE_TYPE_VCN0 = 0,
+	UMSCH_SWIP_ENGINE_TYPE_VCN1 = 1,
+	UMSCH_SWIP_ENGINE_TYPE_VCN = 2,
+	UMSCH_SWIP_ENGINE_TYPE_VPE = 3,
+	UMSCH_SWIP_ENGINE_TYPE_MAX
+};
+
+enum UMSCH_CONTEXT_PRIORITY_LEVEL {
+	CONTEXT_PRIORITY_LEVEL_IDLE = 0,
+	CONTEXT_PRIORITY_LEVEL_NORMAL = 1,
+	CONTEXT_PRIORITY_LEVEL_FOCUS = 2,
+	CONTEXT_PRIORITY_LEVEL_REALTIME = 3,
+	CONTEXT_PRIORITY_NUM_LEVELS
+};
+
+struct umsch_mm_set_resource_input {
+	uint32_t vmid_mask_mm_vcn;
+	uint32_t vmid_mask_mm_vpe;
+	uint32_t collaboration_mask_vpe;
+	uint32_t logging_vmid;
+	uint32_t engine_mask;
+	union {
+		struct {
+			uint32_t disable_reset : 1;
+			uint32_t disable_umsch_mm_log : 1;
+			uint32_t use_rs64mem_for_proc_ctx_csa : 1;
+			uint32_t reserved : 29;
+		};
+		uint32_t uint32_all;
+	};
+};
+
+struct amdgpu_umsch_fwlog {
+	uint32_t rptr;
+	uint32_t wptr;
+	uint32_t buffer_size;
+	uint32_t header_size;
+	uint32_t wrapped;
+};
+
+struct umsch_mm_add_queue_input {
+	uint32_t process_id;
+	uint64_t page_table_base_addr;
+	uint64_t process_va_start;
+	uint64_t process_va_end;
+	uint64_t process_quantum;
+	uint64_t process_csa_addr;
+	uint64_t context_quantum;
+	uint64_t context_csa_addr;
+	uint32_t inprocess_context_priority;
+	enum UMSCH_CONTEXT_PRIORITY_LEVEL context_global_priority_level;
+	uint32_t doorbell_offset_0;
+	uint32_t doorbell_offset_1;
+	enum UMSCH_SWIP_ENGINE_TYPE engine_type;
+	uint32_t affinity;
+	uint64_t mqd_addr;
+	uint64_t h_context;
+	uint64_t h_queue;
+	uint32_t vm_context_cntl;
+
+	uint32_t process_csa_array_index;
+	uint32_t context_csa_array_index;
+
+	struct {
+		uint32_t is_context_suspended : 1;
+		uint32_t collaboration_mode : 1;
+		uint32_t reserved : 30;
+	};
+};
+
+struct umsch_mm_remove_queue_input {
+	uint32_t doorbell_offset_0;
+	uint32_t doorbell_offset_1;
+	uint64_t context_csa_addr;
+	uint32_t context_csa_array_index;
+};
+
+struct MQD_INFO {
+	uint32_t rb_base_hi;
+	uint32_t rb_base_lo;
+	uint32_t rb_size;
+	uint32_t wptr_val;
+	uint32_t rptr_val;
+	uint32_t unmapped;
+	uint32_t vmid;
+};
+
+struct amdgpu_umsch_mm;
+
+struct umsch_mm_funcs {
+	int (*set_hw_resources)(struct amdgpu_umsch_mm *umsch);
+	int (*add_queue)(struct amdgpu_umsch_mm *umsch,
+			 struct umsch_mm_add_queue_input *input);
+	int (*remove_queue)(struct amdgpu_umsch_mm *umsch,
+			    struct umsch_mm_remove_queue_input *input);
+	int (*set_regs)(struct amdgpu_umsch_mm *umsch);
+	int (*init_microcode)(struct amdgpu_umsch_mm *umsch);
+	int (*load_microcode)(struct amdgpu_umsch_mm *umsch);
+	int (*ring_init)(struct amdgpu_umsch_mm *umsch);
+	int (*ring_start)(struct amdgpu_umsch_mm *umsch);
+	int (*ring_stop)(struct amdgpu_umsch_mm *umsch);
+	int (*ring_fini)(struct amdgpu_umsch_mm *umsch);
+};
+
+struct amdgpu_umsch_mm {
+	struct amdgpu_ring		ring;
+
+	uint32_t			rb_wptr;
+	uint32_t			rb_rptr;
+
+	const struct umsch_mm_funcs	*funcs;
+
+	const struct firmware		*fw;
+	uint32_t			fw_version;
+	uint32_t			feature_version;
+
+	struct amdgpu_bo		*ucode_fw_obj;
+	uint64_t			ucode_fw_gpu_addr;
+	uint32_t			*ucode_fw_ptr;
+	uint64_t			irq_start_addr;
+	uint64_t			uc_start_addr;
+	uint32_t			ucode_size;
+
+	struct amdgpu_bo		*data_fw_obj;
+	uint64_t			data_fw_gpu_addr;
+	uint32_t			*data_fw_ptr;
+	uint64_t			data_start_addr;
+	uint32_t			data_size;
+
+	struct amdgpu_bo		*cmd_buf_obj;
+	uint64_t			cmd_buf_gpu_addr;
+	uint32_t			*cmd_buf_ptr;
+	uint32_t			*cmd_buf_curr_ptr;
+
+	uint32_t			wb_index;
+	uint64_t			sch_ctx_gpu_addr;
+	uint32_t			*sch_ctx_cpu_addr;
+
+	uint32_t			vmid_mask_mm_vcn;
+	uint32_t			vmid_mask_mm_vpe;
+	uint32_t			engine_mask;
+	uint32_t			vcn0_hqd_mask;
+	uint32_t			vcn1_hqd_mask;
+	uint32_t			vcn_hqd_mask[2];
+	uint32_t			vpe_hqd_mask;
+	uint32_t			agdb_index[CONTEXT_PRIORITY_NUM_LEVELS];
+
+	struct mutex			mutex_hidden;
+	struct amdgpu_bo		*dbglog_bo;
+	void				*log_cpu_addr;
+	uint64_t			log_gpu_addr;
+	uint32_t			mem_size;
+	uint32_t			log_offset;
+};
+
+int amdgpu_umsch_mm_submit_pkt(struct amdgpu_umsch_mm *umsch, void *pkt, int ndws);
+int amdgpu_umsch_mm_query_fence(struct amdgpu_umsch_mm *umsch);
+
+int amdgpu_umsch_mm_init_microcode(struct amdgpu_umsch_mm *umsch);
+int amdgpu_umsch_mm_allocate_ucode_buffer(struct amdgpu_umsch_mm *umsch);
+int amdgpu_umsch_mm_allocate_ucode_data_buffer(struct amdgpu_umsch_mm *umsch);
+
+int amdgpu_umsch_mm_psp_execute_cmd_buf(struct amdgpu_umsch_mm *umsch);
+
+int amdgpu_umsch_mm_ring_init(struct amdgpu_umsch_mm *umsch);
+
+void amdgpu_debugfs_umsch_fwlog_init(struct amdgpu_device *adev,
+			struct amdgpu_umsch_mm *umsch);
+
+void amdgpu_umsch_fwlog_init(struct amdgpu_umsch_mm *umsch_mm);
+
+#define WREG32_SOC15_UMSCH(reg, value)								\
+	do {											\
+		uint32_t reg_offset = adev->reg_offset[VCN_HWIP][0][reg##_BASE_IDX] + reg;	\
+		if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {				\
+			*adev->umsch_mm.cmd_buf_curr_ptr++ = (reg_offset << 2);			\
+			*adev->umsch_mm.cmd_buf_curr_ptr++ = value;				\
+		} else	{									\
+			WREG32(reg_offset, value);						\
+		}										\
+	} while (0)
+
+#define umsch_mm_set_hw_resources(umsch) \
+	((umsch)->funcs->set_hw_resources ? (umsch)->funcs->set_hw_resources((umsch)) : 0)
+#define umsch_mm_add_queue(umsch, input) \
+	((umsch)->funcs->add_queue ? (umsch)->funcs->add_queue((umsch), (input)) : 0)
+#define umsch_mm_remove_queue(umsch, input) \
+	((umsch)->funcs->remove_queue ? (umsch)->funcs->remove_queue((umsch), (input)) : 0)
+
+#define umsch_mm_set_regs(umsch) \
+	((umsch)->funcs->set_regs ? (umsch)->funcs->set_regs((umsch)) : 0)
+#define umsch_mm_init_microcode(umsch) \
+	((umsch)->funcs->init_microcode ? (umsch)->funcs->init_microcode((umsch)) : 0)
+#define umsch_mm_load_microcode(umsch) \
+	((umsch)->funcs->load_microcode ? (umsch)->funcs->load_microcode((umsch)) : 0)
+
+#define umsch_mm_ring_init(umsch) \
+	((umsch)->funcs->ring_init ? (umsch)->funcs->ring_init((umsch)) : 0)
+#define umsch_mm_ring_start(umsch) \
+	((umsch)->funcs->ring_start ? (umsch)->funcs->ring_start((umsch)) : 0)
+#define umsch_mm_ring_stop(umsch) \
+	((umsch)->funcs->ring_stop ? (umsch)->funcs->ring_stop((umsch)) : 0)
+#define umsch_mm_ring_fini(umsch) \
+	((umsch)->funcs->ring_fini ? (umsch)->funcs->ring_fini((umsch)) : 0)
+
+static inline void amdgpu_umsch_mm_lock(struct amdgpu_umsch_mm *umsch)
+{
+	mutex_lock(&umsch->mutex_hidden);
+}
+
+static inline void amdgpu_umsch_mm_unlock(struct amdgpu_umsch_mm *umsch)
+{
+	mutex_unlock(&umsch->mutex_hidden);
+}
+
+extern const struct amdgpu_ip_block_version umsch_mm_v4_0_ip_block;
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_userq.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_userq.c
new file mode 100644
index 000000000000..9a969175900e
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_userq.c
@@ -0,0 +1,1482 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include <drm/drm_auth.h>
+#include <drm/drm_exec.h>
+#include <linux/pm_runtime.h>
+#include <drm/drm_drv.h>
+
+#include "amdgpu.h"
+#include "amdgpu_reset.h"
+#include "amdgpu_vm.h"
+#include "amdgpu_userq.h"
+#include "amdgpu_hmm.h"
+#include "amdgpu_userq_fence.h"
+
+u32 amdgpu_userq_get_supported_ip_mask(struct amdgpu_device *adev)
+{
+	int i;
+	u32 userq_ip_mask = 0;
+
+	for (i = 0; i < AMDGPU_HW_IP_NUM; i++) {
+		if (adev->userq_funcs[i])
+			userq_ip_mask |= (1 << i);
+	}
+
+	return userq_ip_mask;
+}
+
+static bool amdgpu_userq_is_reset_type_supported(struct amdgpu_device *adev,
+				enum amdgpu_ring_type ring_type, int reset_type)
+{
+
+	if (ring_type < 0 || ring_type >= AMDGPU_RING_TYPE_MAX)
+		return false;
+
+	switch (ring_type) {
+	case AMDGPU_RING_TYPE_GFX:
+		if (adev->gfx.gfx_supported_reset & reset_type)
+			return true;
+		break;
+	case AMDGPU_RING_TYPE_COMPUTE:
+		if (adev->gfx.compute_supported_reset & reset_type)
+			return true;
+		break;
+	case AMDGPU_RING_TYPE_SDMA:
+		if (adev->sdma.supported_reset & reset_type)
+			return true;
+		break;
+	case AMDGPU_RING_TYPE_VCN_DEC:
+	case AMDGPU_RING_TYPE_VCN_ENC:
+		if (adev->vcn.supported_reset & reset_type)
+			return true;
+		break;
+	case AMDGPU_RING_TYPE_VCN_JPEG:
+		if (adev->jpeg.supported_reset & reset_type)
+			return true;
+		break;
+	default:
+		break;
+	}
+	return false;
+}
+
+static void amdgpu_userq_gpu_reset(struct amdgpu_device *adev)
+{
+	if (amdgpu_device_should_recover_gpu(adev)) {
+		amdgpu_reset_domain_schedule(adev->reset_domain,
+					     &adev->userq_reset_work);
+		/* Wait for the reset job to complete */
+		flush_work(&adev->userq_reset_work);
+	}
+}
+
+static int
+amdgpu_userq_detect_and_reset_queues(struct amdgpu_userq_mgr *uq_mgr)
+{
+	struct amdgpu_device *adev = uq_mgr->adev;
+	const int queue_types[] = {
+		AMDGPU_RING_TYPE_COMPUTE,
+		AMDGPU_RING_TYPE_GFX,
+		AMDGPU_RING_TYPE_SDMA
+	};
+	const int num_queue_types = ARRAY_SIZE(queue_types);
+	bool gpu_reset = false;
+	int r = 0;
+	int i;
+
+	/* Warning if current process mutex is not held */
+	WARN_ON(!mutex_is_locked(&uq_mgr->userq_mutex));
+
+	if (unlikely(adev->debug_disable_gpu_ring_reset)) {
+		dev_err(adev->dev, "userq reset disabled by debug mask\n");
+		return 0;
+	}
+
+	/*
+	 * If GPU recovery feature is disabled system-wide,
+	 * skip all reset detection logic
+	 */
+	if (!amdgpu_gpu_recovery)
+		return 0;
+
+	/*
+	 * Iterate through all queue types to detect and reset problematic queues
+	 * Process each queue type in the defined order
+	 */
+	for (i = 0; i < num_queue_types; i++) {
+		int ring_type = queue_types[i];
+		const struct amdgpu_userq_funcs *funcs = adev->userq_funcs[ring_type];
+
+		if (!amdgpu_userq_is_reset_type_supported(adev, ring_type, AMDGPU_RESET_TYPE_PER_QUEUE))
+				continue;
+
+		if (atomic_read(&uq_mgr->userq_count[ring_type]) > 0 &&
+		    funcs && funcs->detect_and_reset) {
+			r = funcs->detect_and_reset(adev, ring_type);
+			if (r) {
+				gpu_reset = true;
+				break;
+			}
+		}
+	}
+
+	if (gpu_reset)
+		amdgpu_userq_gpu_reset(adev);
+
+	return r;
+}
+
+static int amdgpu_userq_buffer_va_list_add(struct amdgpu_usermode_queue *queue,
+					   struct amdgpu_bo_va_mapping *va_map, u64 addr)
+{
+	struct amdgpu_userq_va_cursor *va_cursor;
+	struct userq_va_list;
+
+	va_cursor = kzalloc(sizeof(*va_cursor), GFP_KERNEL);
+	if (!va_cursor)
+		return -ENOMEM;
+
+	INIT_LIST_HEAD(&va_cursor->list);
+	va_cursor->gpu_addr = addr;
+	atomic_set(&va_map->bo_va->userq_va_mapped, 1);
+	list_add(&va_cursor->list, &queue->userq_va_list);
+
+	return 0;
+}
+
+int amdgpu_userq_input_va_validate(struct amdgpu_usermode_queue *queue,
+				   u64 addr, u64 expected_size)
+{
+	struct amdgpu_bo_va_mapping *va_map;
+	struct amdgpu_vm *vm = queue->vm;
+	u64 user_addr;
+	u64 size;
+	int r = 0;
+
+	user_addr = (addr & AMDGPU_GMC_HOLE_MASK) >> AMDGPU_GPU_PAGE_SHIFT;
+	size = expected_size >> AMDGPU_GPU_PAGE_SHIFT;
+
+	r = amdgpu_bo_reserve(vm->root.bo, false);
+	if (r)
+		return r;
+
+	va_map = amdgpu_vm_bo_lookup_mapping(vm, user_addr);
+	if (!va_map) {
+		r = -EINVAL;
+		goto out_err;
+	}
+	/* Only validate the userq whether resident in the VM mapping range */
+	if (user_addr >= va_map->start  &&
+	    va_map->last - user_addr + 1 >= size) {
+		amdgpu_userq_buffer_va_list_add(queue, va_map, user_addr);
+		amdgpu_bo_unreserve(vm->root.bo);
+		return 0;
+	}
+
+	r = -EINVAL;
+out_err:
+	amdgpu_bo_unreserve(vm->root.bo);
+	return r;
+}
+
+static bool amdgpu_userq_buffer_va_mapped(struct amdgpu_vm *vm, u64 addr)
+{
+	struct amdgpu_bo_va_mapping *mapping;
+	bool r;
+
+	if (amdgpu_bo_reserve(vm->root.bo, false))
+		return false;
+
+	mapping = amdgpu_vm_bo_lookup_mapping(vm, addr);
+	if (!IS_ERR_OR_NULL(mapping) && atomic_read(&mapping->bo_va->userq_va_mapped))
+		r = true;
+	else
+		r = false;
+	amdgpu_bo_unreserve(vm->root.bo);
+
+	return r;
+}
+
+static bool amdgpu_userq_buffer_vas_mapped(struct amdgpu_usermode_queue *queue)
+{
+	struct amdgpu_userq_va_cursor *va_cursor, *tmp;
+	int r = 0;
+
+	list_for_each_entry_safe(va_cursor, tmp, &queue->userq_va_list, list) {
+		r += amdgpu_userq_buffer_va_mapped(queue->vm, va_cursor->gpu_addr);
+		dev_dbg(queue->userq_mgr->adev->dev,
+			"validate the userq mapping:%p va:%llx r:%d\n",
+			queue, va_cursor->gpu_addr, r);
+	}
+
+	if (r != 0)
+		return true;
+
+	return false;
+}
+
+static void amdgpu_userq_buffer_va_list_del(struct amdgpu_bo_va_mapping *mapping,
+					    struct amdgpu_userq_va_cursor *va_cursor)
+{
+	atomic_set(&mapping->bo_va->userq_va_mapped, 0);
+	list_del(&va_cursor->list);
+	kfree(va_cursor);
+}
+
+static int amdgpu_userq_buffer_vas_list_cleanup(struct amdgpu_device *adev,
+						struct amdgpu_usermode_queue *queue)
+{
+	struct amdgpu_userq_va_cursor *va_cursor, *tmp;
+	struct amdgpu_bo_va_mapping *mapping;
+	int r;
+
+	r = amdgpu_bo_reserve(queue->vm->root.bo, false);
+	if (r)
+		return r;
+
+	list_for_each_entry_safe(va_cursor, tmp, &queue->userq_va_list, list) {
+		mapping = amdgpu_vm_bo_lookup_mapping(queue->vm, va_cursor->gpu_addr);
+		if (!mapping) {
+			r = -EINVAL;
+			goto err;
+		}
+		dev_dbg(adev->dev, "delete the userq:%p va:%llx\n",
+			queue, va_cursor->gpu_addr);
+		amdgpu_userq_buffer_va_list_del(mapping, va_cursor);
+	}
+err:
+	amdgpu_bo_unreserve(queue->vm->root.bo);
+	return r;
+}
+
+static int
+amdgpu_userq_preempt_helper(struct amdgpu_userq_mgr *uq_mgr,
+			  struct amdgpu_usermode_queue *queue)
+{
+	struct amdgpu_device *adev = uq_mgr->adev;
+	const struct amdgpu_userq_funcs *userq_funcs =
+		adev->userq_funcs[queue->queue_type];
+	bool found_hung_queue = false;
+	int r = 0;
+
+	if (queue->state == AMDGPU_USERQ_STATE_MAPPED) {
+		r = userq_funcs->preempt(uq_mgr, queue);
+		if (r) {
+			queue->state = AMDGPU_USERQ_STATE_HUNG;
+			found_hung_queue = true;
+		} else {
+			queue->state = AMDGPU_USERQ_STATE_PREEMPTED;
+		}
+	}
+
+	if (found_hung_queue)
+		amdgpu_userq_detect_and_reset_queues(uq_mgr);
+
+	return r;
+}
+
+static int
+amdgpu_userq_restore_helper(struct amdgpu_userq_mgr *uq_mgr,
+			struct amdgpu_usermode_queue *queue)
+{
+	struct amdgpu_device *adev = uq_mgr->adev;
+	const struct amdgpu_userq_funcs *userq_funcs =
+		adev->userq_funcs[queue->queue_type];
+	int r = 0;
+
+	if (queue->state == AMDGPU_USERQ_STATE_PREEMPTED) {
+		r = userq_funcs->restore(uq_mgr, queue);
+		if (r) {
+			queue->state = AMDGPU_USERQ_STATE_HUNG;
+		} else {
+			queue->state = AMDGPU_USERQ_STATE_MAPPED;
+		}
+	}
+
+	return r;
+}
+
+static int
+amdgpu_userq_unmap_helper(struct amdgpu_userq_mgr *uq_mgr,
+			  struct amdgpu_usermode_queue *queue)
+{
+	struct amdgpu_device *adev = uq_mgr->adev;
+	const struct amdgpu_userq_funcs *userq_funcs =
+		adev->userq_funcs[queue->queue_type];
+	bool found_hung_queue = false;
+	int r = 0;
+
+	if ((queue->state == AMDGPU_USERQ_STATE_MAPPED) ||
+		(queue->state == AMDGPU_USERQ_STATE_PREEMPTED)) {
+		r = userq_funcs->unmap(uq_mgr, queue);
+		if (r) {
+			queue->state = AMDGPU_USERQ_STATE_HUNG;
+			found_hung_queue = true;
+		} else {
+			queue->state = AMDGPU_USERQ_STATE_UNMAPPED;
+		}
+	}
+
+	if (found_hung_queue)
+		amdgpu_userq_detect_and_reset_queues(uq_mgr);
+
+	return r;
+}
+
+static int
+amdgpu_userq_map_helper(struct amdgpu_userq_mgr *uq_mgr,
+			struct amdgpu_usermode_queue *queue)
+{
+	struct amdgpu_device *adev = uq_mgr->adev;
+	const struct amdgpu_userq_funcs *userq_funcs =
+		adev->userq_funcs[queue->queue_type];
+	int r = 0;
+
+	if (queue->state == AMDGPU_USERQ_STATE_UNMAPPED) {
+		r = userq_funcs->map(uq_mgr, queue);
+		if (r) {
+			queue->state = AMDGPU_USERQ_STATE_HUNG;
+			amdgpu_userq_detect_and_reset_queues(uq_mgr);
+		} else {
+			queue->state = AMDGPU_USERQ_STATE_MAPPED;
+		}
+	}
+
+	return r;
+}
+
+static int
+amdgpu_userq_wait_for_last_fence(struct amdgpu_userq_mgr *uq_mgr,
+				 struct amdgpu_usermode_queue *queue)
+{
+	struct dma_fence *f = queue->last_fence;
+	int ret = 0;
+
+	if (f && !dma_fence_is_signaled(f)) {
+		ret = dma_fence_wait_timeout(f, true, MAX_SCHEDULE_TIMEOUT);
+		if (ret <= 0) {
+			drm_file_err(uq_mgr->file, "Timed out waiting for fence=%llu:%llu\n",
+				     f->context, f->seqno);
+			queue->state = AMDGPU_USERQ_STATE_HUNG;
+			return -ETIME;
+		}
+	}
+
+	return ret;
+}
+
+static void
+amdgpu_userq_cleanup(struct amdgpu_userq_mgr *uq_mgr,
+		     struct amdgpu_usermode_queue *queue,
+		     int queue_id)
+{
+	struct amdgpu_device *adev = uq_mgr->adev;
+	const struct amdgpu_userq_funcs *uq_funcs = adev->userq_funcs[queue->queue_type];
+
+	/* Wait for mode-1 reset to complete */
+	down_read(&adev->reset_domain->sem);
+
+	/* Drop the userq reference. */
+	amdgpu_userq_buffer_vas_list_cleanup(adev, queue);
+	uq_funcs->mqd_destroy(uq_mgr, queue);
+	amdgpu_userq_fence_driver_free(queue);
+	/* Use interrupt-safe locking since IRQ handlers may access these XArrays */
+	xa_erase_irq(&uq_mgr->userq_mgr_xa, (unsigned long)queue_id);
+	xa_erase_irq(&adev->userq_doorbell_xa, queue->doorbell_index);
+	queue->userq_mgr = NULL;
+	list_del(&queue->userq_va_list);
+	kfree(queue);
+
+	up_read(&adev->reset_domain->sem);
+}
+
+static struct amdgpu_usermode_queue *
+amdgpu_userq_find(struct amdgpu_userq_mgr *uq_mgr, int qid)
+{
+	return xa_load(&uq_mgr->userq_mgr_xa, qid);
+}
+
+void
+amdgpu_userq_ensure_ev_fence(struct amdgpu_userq_mgr *uq_mgr,
+			     struct amdgpu_eviction_fence_mgr *evf_mgr)
+{
+	struct amdgpu_eviction_fence *ev_fence;
+
+retry:
+	/* Flush any pending resume work to create ev_fence */
+	flush_delayed_work(&uq_mgr->resume_work);
+
+	mutex_lock(&uq_mgr->userq_mutex);
+	spin_lock(&evf_mgr->ev_fence_lock);
+	ev_fence = evf_mgr->ev_fence;
+	spin_unlock(&evf_mgr->ev_fence_lock);
+	if (!ev_fence || dma_fence_is_signaled(&ev_fence->base)) {
+		mutex_unlock(&uq_mgr->userq_mutex);
+		/*
+		 * Looks like there was no pending resume work,
+		 * add one now to create a valid eviction fence
+		 */
+		schedule_delayed_work(&uq_mgr->resume_work, 0);
+		goto retry;
+	}
+}
+
+int amdgpu_userq_create_object(struct amdgpu_userq_mgr *uq_mgr,
+			       struct amdgpu_userq_obj *userq_obj,
+			       int size)
+{
+	struct amdgpu_device *adev = uq_mgr->adev;
+	struct amdgpu_bo_param bp;
+	int r;
+
+	memset(&bp, 0, sizeof(bp));
+	bp.byte_align = PAGE_SIZE;
+	bp.domain = AMDGPU_GEM_DOMAIN_GTT;
+	bp.flags = AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS |
+		   AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED;
+	bp.type = ttm_bo_type_kernel;
+	bp.size = size;
+	bp.resv = NULL;
+	bp.bo_ptr_size = sizeof(struct amdgpu_bo);
+
+	r = amdgpu_bo_create(adev, &bp, &userq_obj->obj);
+	if (r) {
+		drm_file_err(uq_mgr->file, "Failed to allocate BO for userqueue (%d)", r);
+		return r;
+	}
+
+	r = amdgpu_bo_reserve(userq_obj->obj, true);
+	if (r) {
+		drm_file_err(uq_mgr->file, "Failed to reserve BO to map (%d)", r);
+		goto free_obj;
+	}
+
+	r = amdgpu_ttm_alloc_gart(&(userq_obj->obj)->tbo);
+	if (r) {
+		drm_file_err(uq_mgr->file, "Failed to alloc GART for userqueue object (%d)", r);
+		goto unresv;
+	}
+
+	r = amdgpu_bo_kmap(userq_obj->obj, &userq_obj->cpu_ptr);
+	if (r) {
+		drm_file_err(uq_mgr->file, "Failed to map BO for userqueue (%d)", r);
+		goto unresv;
+	}
+
+	userq_obj->gpu_addr = amdgpu_bo_gpu_offset(userq_obj->obj);
+	amdgpu_bo_unreserve(userq_obj->obj);
+	memset(userq_obj->cpu_ptr, 0, size);
+	return 0;
+
+unresv:
+	amdgpu_bo_unreserve(userq_obj->obj);
+
+free_obj:
+	amdgpu_bo_unref(&userq_obj->obj);
+	return r;
+}
+
+void amdgpu_userq_destroy_object(struct amdgpu_userq_mgr *uq_mgr,
+				 struct amdgpu_userq_obj *userq_obj)
+{
+	amdgpu_bo_kunmap(userq_obj->obj);
+	amdgpu_bo_unref(&userq_obj->obj);
+}
+
+uint64_t
+amdgpu_userq_get_doorbell_index(struct amdgpu_userq_mgr *uq_mgr,
+				struct amdgpu_db_info *db_info,
+				struct drm_file *filp)
+{
+	uint64_t index;
+	struct drm_gem_object *gobj;
+	struct amdgpu_userq_obj *db_obj = db_info->db_obj;
+	int r, db_size;
+
+	gobj = drm_gem_object_lookup(filp, db_info->doorbell_handle);
+	if (gobj == NULL) {
+		drm_file_err(uq_mgr->file, "Can't find GEM object for doorbell\n");
+		return -EINVAL;
+	}
+
+	db_obj->obj = amdgpu_bo_ref(gem_to_amdgpu_bo(gobj));
+	drm_gem_object_put(gobj);
+
+	r = amdgpu_bo_reserve(db_obj->obj, true);
+	if (r) {
+		drm_file_err(uq_mgr->file, "[Usermode queues] Failed to pin doorbell object\n");
+		goto unref_bo;
+	}
+
+	/* Pin the BO before generating the index, unpin in queue destroy */
+	r = amdgpu_bo_pin(db_obj->obj, AMDGPU_GEM_DOMAIN_DOORBELL);
+	if (r) {
+		drm_file_err(uq_mgr->file, "[Usermode queues] Failed to pin doorbell object\n");
+		goto unresv_bo;
+	}
+
+	switch (db_info->queue_type) {
+	case AMDGPU_HW_IP_GFX:
+	case AMDGPU_HW_IP_COMPUTE:
+	case AMDGPU_HW_IP_DMA:
+		db_size = sizeof(u64);
+		break;
+	default:
+		drm_file_err(uq_mgr->file, "[Usermode queues] IP %d not support\n",
+			     db_info->queue_type);
+		r = -EINVAL;
+		goto unpin_bo;
+	}
+
+	index = amdgpu_doorbell_index_on_bar(uq_mgr->adev, db_obj->obj,
+					     db_info->doorbell_offset, db_size);
+	drm_dbg_driver(adev_to_drm(uq_mgr->adev),
+		       "[Usermode queues] doorbell index=%lld\n", index);
+	amdgpu_bo_unreserve(db_obj->obj);
+	return index;
+
+unpin_bo:
+	amdgpu_bo_unpin(db_obj->obj);
+unresv_bo:
+	amdgpu_bo_unreserve(db_obj->obj);
+unref_bo:
+	amdgpu_bo_unref(&db_obj->obj);
+	return r;
+}
+
+static int
+amdgpu_userq_destroy(struct drm_file *filp, int queue_id)
+{
+	struct amdgpu_fpriv *fpriv = filp->driver_priv;
+	struct amdgpu_userq_mgr *uq_mgr = &fpriv->userq_mgr;
+	struct amdgpu_device *adev = uq_mgr->adev;
+	struct amdgpu_usermode_queue *queue;
+	int r = 0;
+
+	cancel_delayed_work_sync(&uq_mgr->resume_work);
+	mutex_lock(&uq_mgr->userq_mutex);
+
+	queue = amdgpu_userq_find(uq_mgr, queue_id);
+	if (!queue) {
+		drm_dbg_driver(adev_to_drm(uq_mgr->adev), "Invalid queue id to destroy\n");
+		mutex_unlock(&uq_mgr->userq_mutex);
+		return -EINVAL;
+	}
+	amdgpu_userq_wait_for_last_fence(uq_mgr, queue);
+	r = amdgpu_bo_reserve(queue->db_obj.obj, true);
+	if (!r) {
+		amdgpu_bo_unpin(queue->db_obj.obj);
+		amdgpu_bo_unreserve(queue->db_obj.obj);
+	}
+	amdgpu_bo_unref(&queue->db_obj.obj);
+	atomic_dec(&uq_mgr->userq_count[queue->queue_type]);
+#if defined(CONFIG_DEBUG_FS)
+	debugfs_remove_recursive(queue->debugfs_queue);
+#endif
+	amdgpu_userq_detect_and_reset_queues(uq_mgr);
+	r = amdgpu_userq_unmap_helper(uq_mgr, queue);
+	/*TODO: It requires a reset for userq hw unmap error*/
+	if (unlikely(r != AMDGPU_USERQ_STATE_UNMAPPED)) {
+		drm_warn(adev_to_drm(uq_mgr->adev), "trying to destroy a HW mapping userq\n");
+		queue->state = AMDGPU_USERQ_STATE_HUNG;
+	}
+	amdgpu_userq_cleanup(uq_mgr, queue, queue_id);
+	mutex_unlock(&uq_mgr->userq_mutex);
+
+	pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
+
+	return r;
+}
+
+static int amdgpu_userq_priority_permit(struct drm_file *filp,
+					int priority)
+{
+	if (priority < AMDGPU_USERQ_CREATE_FLAGS_QUEUE_PRIORITY_HIGH)
+		return 0;
+
+	if (capable(CAP_SYS_NICE))
+		return 0;
+
+	if (drm_is_current_master(filp))
+		return 0;
+
+	return -EACCES;
+}
+
+#if defined(CONFIG_DEBUG_FS)
+static int amdgpu_mqd_info_read(struct seq_file *m, void *unused)
+{
+	struct amdgpu_usermode_queue *queue = m->private;
+	struct amdgpu_bo *bo;
+	int r;
+
+	if (!queue || !queue->mqd.obj)
+		return -EINVAL;
+
+	bo = amdgpu_bo_ref(queue->mqd.obj);
+	r = amdgpu_bo_reserve(bo, true);
+	if (r) {
+		amdgpu_bo_unref(&bo);
+		return -EINVAL;
+	}
+
+	seq_printf(m, "queue_type: %d\n", queue->queue_type);
+	seq_printf(m, "mqd_gpu_address: 0x%llx\n", amdgpu_bo_gpu_offset(queue->mqd.obj));
+
+	amdgpu_bo_unreserve(bo);
+	amdgpu_bo_unref(&bo);
+
+	return 0;
+}
+
+static int amdgpu_mqd_info_open(struct inode *inode, struct file *file)
+{
+	return single_open(file, amdgpu_mqd_info_read, inode->i_private);
+}
+
+static const struct file_operations amdgpu_mqd_info_fops = {
+	.owner = THIS_MODULE,
+	.open = amdgpu_mqd_info_open,
+	.read = seq_read,
+	.llseek = seq_lseek,
+	.release = single_release,
+};
+#endif
+
+static int
+amdgpu_userq_create(struct drm_file *filp, union drm_amdgpu_userq *args)
+{
+	struct amdgpu_fpriv *fpriv = filp->driver_priv;
+	struct amdgpu_userq_mgr *uq_mgr = &fpriv->userq_mgr;
+	struct amdgpu_device *adev = uq_mgr->adev;
+	const struct amdgpu_userq_funcs *uq_funcs;
+	struct amdgpu_usermode_queue *queue;
+	struct amdgpu_db_info db_info;
+	char *queue_name;
+	bool skip_map_queue;
+	u32 qid;
+	uint64_t index;
+	int r = 0;
+	int priority =
+		(args->in.flags & AMDGPU_USERQ_CREATE_FLAGS_QUEUE_PRIORITY_MASK) >>
+		AMDGPU_USERQ_CREATE_FLAGS_QUEUE_PRIORITY_SHIFT;
+
+	r = amdgpu_userq_priority_permit(filp, priority);
+	if (r)
+		return r;
+
+	r = pm_runtime_get_sync(adev_to_drm(adev)->dev);
+	if (r < 0) {
+		drm_file_err(uq_mgr->file, "pm_runtime_get_sync() failed for userqueue create\n");
+		pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
+		return r;
+	}
+
+	/*
+	 * There could be a situation that we are creating a new queue while
+	 * the other queues under this UQ_mgr are suspended. So if there is any
+	 * resume work pending, wait for it to get done.
+	 *
+	 * This will also make sure we have a valid eviction fence ready to be used.
+	 */
+	amdgpu_userq_ensure_ev_fence(&fpriv->userq_mgr, &fpriv->evf_mgr);
+
+	uq_funcs = adev->userq_funcs[args->in.ip_type];
+	if (!uq_funcs) {
+		drm_file_err(uq_mgr->file, "Usermode queue is not supported for this IP (%u)\n",
+			     args->in.ip_type);
+		r = -EINVAL;
+		goto unlock;
+	}
+
+	queue = kzalloc(sizeof(struct amdgpu_usermode_queue), GFP_KERNEL);
+	if (!queue) {
+		drm_file_err(uq_mgr->file, "Failed to allocate memory for queue\n");
+		r = -ENOMEM;
+		goto unlock;
+	}
+
+	INIT_LIST_HEAD(&queue->userq_va_list);
+	queue->doorbell_handle = args->in.doorbell_handle;
+	queue->queue_type = args->in.ip_type;
+	queue->vm = &fpriv->vm;
+	queue->priority = priority;
+
+	db_info.queue_type = queue->queue_type;
+	db_info.doorbell_handle = queue->doorbell_handle;
+	db_info.db_obj = &queue->db_obj;
+	db_info.doorbell_offset = args->in.doorbell_offset;
+
+	/* Validate the userq virtual address.*/
+	if (amdgpu_userq_input_va_validate(queue, args->in.queue_va, args->in.queue_size) ||
+	    amdgpu_userq_input_va_validate(queue, args->in.rptr_va, AMDGPU_GPU_PAGE_SIZE) ||
+	    amdgpu_userq_input_va_validate(queue, args->in.wptr_va, AMDGPU_GPU_PAGE_SIZE)) {
+		r = -EINVAL;
+		kfree(queue);
+		goto unlock;
+	}
+
+	/* Convert relative doorbell offset into absolute doorbell index */
+	index = amdgpu_userq_get_doorbell_index(uq_mgr, &db_info, filp);
+	if (index == (uint64_t)-EINVAL) {
+		drm_file_err(uq_mgr->file, "Failed to get doorbell for queue\n");
+		kfree(queue);
+		r = -EINVAL;
+		goto unlock;
+	}
+
+	queue->doorbell_index = index;
+	xa_init_flags(&queue->fence_drv_xa, XA_FLAGS_ALLOC);
+	r = amdgpu_userq_fence_driver_alloc(adev, queue);
+	if (r) {
+		drm_file_err(uq_mgr->file, "Failed to alloc fence driver\n");
+		goto unlock;
+	}
+
+	r = uq_funcs->mqd_create(uq_mgr, &args->in, queue);
+	if (r) {
+		drm_file_err(uq_mgr->file, "Failed to create Queue\n");
+		amdgpu_userq_fence_driver_free(queue);
+		kfree(queue);
+		goto unlock;
+	}
+
+	/* Wait for mode-1 reset to complete */
+	down_read(&adev->reset_domain->sem);
+	r = xa_err(xa_store_irq(&adev->userq_doorbell_xa, index, queue, GFP_KERNEL));
+	if (r) {
+		kfree(queue);
+		up_read(&adev->reset_domain->sem);
+		goto unlock;
+	}
+
+	r = xa_alloc(&uq_mgr->userq_mgr_xa, &qid, queue, XA_LIMIT(1, AMDGPU_MAX_USERQ_COUNT), GFP_KERNEL);
+	if (r) {
+		drm_file_err(uq_mgr->file, "Failed to allocate a queue id\n");
+		amdgpu_userq_fence_driver_free(queue);
+		uq_funcs->mqd_destroy(uq_mgr, queue);
+		kfree(queue);
+		r = -ENOMEM;
+		up_read(&adev->reset_domain->sem);
+		goto unlock;
+	}
+	up_read(&adev->reset_domain->sem);
+	queue->userq_mgr = uq_mgr;
+
+	/* don't map the queue if scheduling is halted */
+	if (adev->userq_halt_for_enforce_isolation &&
+	    ((queue->queue_type == AMDGPU_HW_IP_GFX) ||
+	     (queue->queue_type == AMDGPU_HW_IP_COMPUTE)))
+		skip_map_queue = true;
+	else
+		skip_map_queue = false;
+	if (!skip_map_queue) {
+		r = amdgpu_userq_map_helper(uq_mgr, queue);
+		if (r) {
+			drm_file_err(uq_mgr->file, "Failed to map Queue\n");
+			xa_erase(&uq_mgr->userq_mgr_xa, qid);
+			amdgpu_userq_fence_driver_free(queue);
+			uq_funcs->mqd_destroy(uq_mgr, queue);
+			kfree(queue);
+			goto unlock;
+		}
+	}
+
+	queue_name = kasprintf(GFP_KERNEL, "queue-%d", qid);
+	if (!queue_name) {
+		r = -ENOMEM;
+		goto unlock;
+	}
+
+#if defined(CONFIG_DEBUG_FS)
+	/* Queue dentry per client to hold MQD information   */
+	queue->debugfs_queue = debugfs_create_dir(queue_name, filp->debugfs_client);
+	debugfs_create_file("mqd_info", 0444, queue->debugfs_queue, queue, &amdgpu_mqd_info_fops);
+#endif
+	kfree(queue_name);
+
+	args->out.queue_id = qid;
+	atomic_inc(&uq_mgr->userq_count[queue->queue_type]);
+
+unlock:
+	mutex_unlock(&uq_mgr->userq_mutex);
+
+	return r;
+}
+
+static int amdgpu_userq_input_args_validate(struct drm_device *dev,
+					union drm_amdgpu_userq *args,
+					struct drm_file *filp)
+{
+	struct amdgpu_device *adev = drm_to_adev(dev);
+
+	switch (args->in.op) {
+	case AMDGPU_USERQ_OP_CREATE:
+		if (args->in.flags & ~(AMDGPU_USERQ_CREATE_FLAGS_QUEUE_PRIORITY_MASK |
+				       AMDGPU_USERQ_CREATE_FLAGS_QUEUE_SECURE))
+			return -EINVAL;
+		/* Usermode queues are only supported for GFX IP as of now */
+		if (args->in.ip_type != AMDGPU_HW_IP_GFX &&
+		    args->in.ip_type != AMDGPU_HW_IP_DMA &&
+		    args->in.ip_type != AMDGPU_HW_IP_COMPUTE) {
+			drm_file_err(filp, "Usermode queue doesn't support IP type %u\n",
+				     args->in.ip_type);
+			return -EINVAL;
+		}
+
+		if ((args->in.flags & AMDGPU_USERQ_CREATE_FLAGS_QUEUE_SECURE) &&
+		    (args->in.ip_type != AMDGPU_HW_IP_GFX) &&
+		    (args->in.ip_type != AMDGPU_HW_IP_COMPUTE) &&
+		    !amdgpu_is_tmz(adev)) {
+			drm_file_err(filp, "Secure only supported on GFX/Compute queues\n");
+			return -EINVAL;
+		}
+
+		if (args->in.queue_va == AMDGPU_BO_INVALID_OFFSET ||
+		    args->in.queue_va == 0 ||
+		    args->in.queue_size == 0) {
+			drm_file_err(filp, "invalidate userq queue va or size\n");
+			return -EINVAL;
+		}
+		if (!args->in.wptr_va || !args->in.rptr_va) {
+			drm_file_err(filp, "invalidate userq queue rptr or wptr\n");
+			return -EINVAL;
+		}
+		break;
+	case AMDGPU_USERQ_OP_FREE:
+		if (args->in.ip_type ||
+		    args->in.doorbell_handle ||
+		    args->in.doorbell_offset ||
+		    args->in.flags ||
+		    args->in.queue_va ||
+		    args->in.queue_size ||
+		    args->in.rptr_va ||
+		    args->in.wptr_va ||
+		    args->in.mqd ||
+		    args->in.mqd_size)
+			return -EINVAL;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+int amdgpu_userq_ioctl(struct drm_device *dev, void *data,
+		       struct drm_file *filp)
+{
+	union drm_amdgpu_userq *args = data;
+	int r;
+
+	if (amdgpu_userq_input_args_validate(dev, args, filp) < 0)
+		return -EINVAL;
+
+	switch (args->in.op) {
+	case AMDGPU_USERQ_OP_CREATE:
+		r = amdgpu_userq_create(filp, args);
+		if (r)
+			drm_file_err(filp, "Failed to create usermode queue\n");
+		break;
+
+	case AMDGPU_USERQ_OP_FREE:
+		r = amdgpu_userq_destroy(filp, args->in.queue_id);
+		if (r)
+			drm_file_err(filp, "Failed to destroy usermode queue\n");
+		break;
+
+	default:
+		drm_dbg_driver(dev, "Invalid user queue op specified: %d\n", args->in.op);
+		return -EINVAL;
+	}
+
+	return r;
+}
+
+static int
+amdgpu_userq_restore_all(struct amdgpu_userq_mgr *uq_mgr)
+{
+	struct amdgpu_usermode_queue *queue;
+	unsigned long queue_id;
+	int ret = 0, r;
+
+	/* Resume all the queues for this process */
+	xa_for_each(&uq_mgr->userq_mgr_xa, queue_id, queue) {
+
+		if (!amdgpu_userq_buffer_vas_mapped(queue)) {
+			drm_file_err(uq_mgr->file,
+				     "trying restore queue without va mapping\n");
+			queue->state = AMDGPU_USERQ_STATE_INVALID_VA;
+			continue;
+		}
+
+		r = amdgpu_userq_restore_helper(uq_mgr, queue);
+		if (r)
+			ret = r;
+	}
+
+	if (ret)
+		drm_file_err(uq_mgr->file, "Failed to map all the queues\n");
+	return ret;
+}
+
+static int amdgpu_userq_validate_vm(void *param, struct amdgpu_bo *bo)
+{
+	struct ttm_operation_ctx ctx = { false, false };
+
+	amdgpu_bo_placement_from_domain(bo, bo->allowed_domains);
+	return ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
+}
+
+/* Handle all BOs on the invalidated list, validate them and update the PTs */
+static int
+amdgpu_userq_bo_validate(struct amdgpu_device *adev, struct drm_exec *exec,
+			 struct amdgpu_vm *vm)
+{
+	struct ttm_operation_ctx ctx = { false, false };
+	struct amdgpu_bo_va *bo_va;
+	struct amdgpu_bo *bo;
+	int ret;
+
+	spin_lock(&vm->status_lock);
+	while (!list_empty(&vm->invalidated)) {
+		bo_va = list_first_entry(&vm->invalidated,
+					 struct amdgpu_bo_va,
+					 base.vm_status);
+		spin_unlock(&vm->status_lock);
+
+		bo = bo_va->base.bo;
+		ret = drm_exec_prepare_obj(exec, &bo->tbo.base, 2);
+		if (unlikely(ret))
+			return ret;
+
+		amdgpu_bo_placement_from_domain(bo, bo->allowed_domains);
+		ret = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
+		if (ret)
+			return ret;
+
+		/* This moves the bo_va to the done list */
+		ret = amdgpu_vm_bo_update(adev, bo_va, false);
+		if (ret)
+			return ret;
+
+		spin_lock(&vm->status_lock);
+	}
+	spin_unlock(&vm->status_lock);
+
+	return 0;
+}
+
+/* Make sure the whole VM is ready to be used */
+static int
+amdgpu_userq_vm_validate(struct amdgpu_userq_mgr *uq_mgr)
+{
+	struct amdgpu_fpriv *fpriv = uq_mgr_to_fpriv(uq_mgr);
+	bool invalidated = false, new_addition = false;
+	struct ttm_operation_ctx ctx = { true, false };
+	struct amdgpu_device *adev = uq_mgr->adev;
+	struct amdgpu_hmm_range *range;
+	struct amdgpu_vm *vm = &fpriv->vm;
+	unsigned long key, tmp_key;
+	struct amdgpu_bo_va *bo_va;
+	struct amdgpu_bo *bo;
+	struct drm_exec exec;
+	struct xarray xa;
+	int ret;
+
+	xa_init(&xa);
+
+retry_lock:
+	drm_exec_init(&exec, DRM_EXEC_IGNORE_DUPLICATES, 0);
+	drm_exec_until_all_locked(&exec) {
+		ret = amdgpu_vm_lock_pd(vm, &exec, 1);
+		drm_exec_retry_on_contention(&exec);
+		if (unlikely(ret))
+			goto unlock_all;
+
+		ret = amdgpu_vm_lock_done_list(vm, &exec, 1);
+		drm_exec_retry_on_contention(&exec);
+		if (unlikely(ret))
+			goto unlock_all;
+
+		/* This validates PDs, PTs and per VM BOs */
+		ret = amdgpu_vm_validate(adev, vm, NULL,
+					 amdgpu_userq_validate_vm,
+					 NULL);
+		if (unlikely(ret))
+			goto unlock_all;
+
+		/* This locks and validates the remaining evicted BOs */
+		ret = amdgpu_userq_bo_validate(adev, &exec, vm);
+		drm_exec_retry_on_contention(&exec);
+		if (unlikely(ret))
+			goto unlock_all;
+	}
+
+	if (invalidated) {
+		xa_for_each(&xa, tmp_key, range) {
+			bo = range->bo;
+			amdgpu_bo_placement_from_domain(bo, AMDGPU_GEM_DOMAIN_CPU);
+			ret = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
+			if (ret)
+				goto unlock_all;
+
+			amdgpu_ttm_tt_set_user_pages(bo->tbo.ttm, range);
+
+			amdgpu_bo_placement_from_domain(bo, AMDGPU_GEM_DOMAIN_GTT);
+			ret = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
+			if (ret)
+				goto unlock_all;
+		}
+		invalidated = false;
+	}
+
+	ret = amdgpu_vm_handle_moved(adev, vm, NULL);
+	if (ret)
+		goto unlock_all;
+
+	key = 0;
+	/* Validate User Ptr BOs */
+	list_for_each_entry(bo_va, &vm->done, base.vm_status) {
+		bo = bo_va->base.bo;
+		if (!bo)
+			continue;
+
+		if (!amdgpu_ttm_tt_is_userptr(bo->tbo.ttm))
+			continue;
+
+		range = xa_load(&xa, key);
+		if (range && range->bo != bo) {
+			xa_erase(&xa, key);
+			amdgpu_hmm_range_free(range);
+			range = NULL;
+		}
+
+		if (!range) {
+			range = amdgpu_hmm_range_alloc(bo);
+			if (!range) {
+				ret = -ENOMEM;
+				goto unlock_all;
+			}
+
+			xa_store(&xa, key, range, GFP_KERNEL);
+			new_addition = true;
+		}
+		key++;
+	}
+
+	if (new_addition) {
+		drm_exec_fini(&exec);
+		xa_for_each(&xa, tmp_key, range) {
+			if (!range)
+				continue;
+			bo = range->bo;
+			ret = amdgpu_ttm_tt_get_user_pages(bo, range);
+			if (ret)
+				goto unlock_all;
+		}
+
+		invalidated = true;
+		new_addition = false;
+		goto retry_lock;
+	}
+
+	ret = amdgpu_vm_update_pdes(adev, vm, false);
+	if (ret)
+		goto unlock_all;
+
+	/*
+	 * We need to wait for all VM updates to finish before restarting the
+	 * queues. Using the done list like that is now ok since everything is
+	 * locked in place.
+	 */
+	list_for_each_entry(bo_va, &vm->done, base.vm_status)
+		dma_fence_wait(bo_va->last_pt_update, false);
+	dma_fence_wait(vm->last_update, false);
+
+	ret = amdgpu_eviction_fence_replace_fence(&fpriv->evf_mgr, &exec);
+	if (ret)
+		drm_file_err(uq_mgr->file, "Failed to replace eviction fence\n");
+
+unlock_all:
+	drm_exec_fini(&exec);
+	xa_for_each(&xa, tmp_key, range) {
+		if (!range)
+			continue;
+		bo = range->bo;
+		amdgpu_hmm_range_free(range);
+	}
+	xa_destroy(&xa);
+	return ret;
+}
+
+static void amdgpu_userq_restore_worker(struct work_struct *work)
+{
+	struct amdgpu_userq_mgr *uq_mgr = work_to_uq_mgr(work, resume_work.work);
+	struct amdgpu_fpriv *fpriv = uq_mgr_to_fpriv(uq_mgr);
+	int ret;
+
+	flush_delayed_work(&fpriv->evf_mgr.suspend_work);
+
+	mutex_lock(&uq_mgr->userq_mutex);
+
+	ret = amdgpu_userq_vm_validate(uq_mgr);
+	if (ret) {
+		drm_file_err(uq_mgr->file, "Failed to validate BOs to restore\n");
+		goto unlock;
+	}
+
+	ret = amdgpu_userq_restore_all(uq_mgr);
+	if (ret) {
+		drm_file_err(uq_mgr->file, "Failed to restore all queues\n");
+		goto unlock;
+	}
+
+unlock:
+	mutex_unlock(&uq_mgr->userq_mutex);
+}
+
+static int
+amdgpu_userq_evict_all(struct amdgpu_userq_mgr *uq_mgr)
+{
+	struct amdgpu_usermode_queue *queue;
+	unsigned long queue_id;
+	int ret = 0, r;
+
+	amdgpu_userq_detect_and_reset_queues(uq_mgr);
+	/* Try to unmap all the queues in this process ctx */
+	xa_for_each(&uq_mgr->userq_mgr_xa, queue_id, queue) {
+		r = amdgpu_userq_preempt_helper(uq_mgr, queue);
+		if (r)
+			ret = r;
+	}
+
+	if (ret)
+		drm_file_err(uq_mgr->file, "Couldn't unmap all the queues\n");
+	return ret;
+}
+
+void amdgpu_userq_reset_work(struct work_struct *work)
+{
+	struct amdgpu_device *adev = container_of(work, struct amdgpu_device,
+						  userq_reset_work);
+	struct amdgpu_reset_context reset_context;
+
+	memset(&reset_context, 0, sizeof(reset_context));
+
+	reset_context.method = AMD_RESET_METHOD_NONE;
+	reset_context.reset_req_dev = adev;
+	reset_context.src = AMDGPU_RESET_SRC_USERQ;
+	set_bit(AMDGPU_NEED_FULL_RESET, &reset_context.flags);
+	/*set_bit(AMDGPU_SKIP_COREDUMP, &reset_context.flags);*/
+
+	amdgpu_device_gpu_recover(adev, NULL, &reset_context);
+}
+
+static int
+amdgpu_userq_wait_for_signal(struct amdgpu_userq_mgr *uq_mgr)
+{
+	struct amdgpu_usermode_queue *queue;
+	unsigned long queue_id;
+	int ret;
+
+	xa_for_each(&uq_mgr->userq_mgr_xa, queue_id, queue) {
+		struct dma_fence *f = queue->last_fence;
+
+		if (!f || dma_fence_is_signaled(f))
+			continue;
+		ret = dma_fence_wait_timeout(f, true, msecs_to_jiffies(100));
+		if (ret <= 0) {
+			drm_file_err(uq_mgr->file, "Timed out waiting for fence=%llu:%llu\n",
+				     f->context, f->seqno);
+			return -ETIMEDOUT;
+		}
+	}
+
+	return 0;
+}
+
+void
+amdgpu_userq_evict(struct amdgpu_userq_mgr *uq_mgr,
+		   struct amdgpu_eviction_fence *ev_fence)
+{
+	struct amdgpu_fpriv *fpriv = uq_mgr_to_fpriv(uq_mgr);
+	struct amdgpu_eviction_fence_mgr *evf_mgr = &fpriv->evf_mgr;
+	struct amdgpu_device *adev = uq_mgr->adev;
+	int ret;
+
+	/* Wait for any pending userqueue fence work to finish */
+	ret = amdgpu_userq_wait_for_signal(uq_mgr);
+	if (ret)
+		dev_err(adev->dev, "Not evicting userqueue, timeout waiting for work\n");
+
+	ret = amdgpu_userq_evict_all(uq_mgr);
+	if (ret)
+		dev_err(adev->dev, "Failed to evict userqueue\n");
+
+	/* Signal current eviction fence */
+	amdgpu_eviction_fence_signal(evf_mgr, ev_fence);
+
+	if (evf_mgr->fd_closing) {
+		cancel_delayed_work_sync(&uq_mgr->resume_work);
+		return;
+	}
+
+	/* Schedule a resume work */
+	schedule_delayed_work(&uq_mgr->resume_work, 0);
+}
+
+int amdgpu_userq_mgr_init(struct amdgpu_userq_mgr *userq_mgr, struct drm_file *file_priv,
+			  struct amdgpu_device *adev)
+{
+	mutex_init(&userq_mgr->userq_mutex);
+	xa_init_flags(&userq_mgr->userq_mgr_xa, XA_FLAGS_ALLOC);
+	userq_mgr->adev = adev;
+	userq_mgr->file = file_priv;
+
+	INIT_DELAYED_WORK(&userq_mgr->resume_work, amdgpu_userq_restore_worker);
+	return 0;
+}
+
+void amdgpu_userq_mgr_fini(struct amdgpu_userq_mgr *userq_mgr)
+{
+	struct amdgpu_usermode_queue *queue;
+	unsigned long queue_id;
+
+	cancel_delayed_work_sync(&userq_mgr->resume_work);
+
+	mutex_lock(&userq_mgr->userq_mutex);
+	amdgpu_userq_detect_and_reset_queues(userq_mgr);
+	xa_for_each(&userq_mgr->userq_mgr_xa, queue_id, queue) {
+		amdgpu_userq_wait_for_last_fence(userq_mgr, queue);
+		amdgpu_userq_unmap_helper(userq_mgr, queue);
+		amdgpu_userq_cleanup(userq_mgr, queue, queue_id);
+	}
+
+	xa_destroy(&userq_mgr->userq_mgr_xa);
+	mutex_unlock(&userq_mgr->userq_mutex);
+	mutex_destroy(&userq_mgr->userq_mutex);
+}
+
+int amdgpu_userq_suspend(struct amdgpu_device *adev)
+{
+	u32 ip_mask = amdgpu_userq_get_supported_ip_mask(adev);
+	struct amdgpu_usermode_queue *queue;
+	struct amdgpu_userq_mgr *uqm;
+	unsigned long queue_id;
+	int r;
+
+	if (!ip_mask)
+		return 0;
+
+	xa_for_each(&adev->userq_doorbell_xa, queue_id, queue) {
+		uqm = queue->userq_mgr;
+		cancel_delayed_work_sync(&uqm->resume_work);
+		guard(mutex)(&uqm->userq_mutex);
+		amdgpu_userq_detect_and_reset_queues(uqm);
+		if (adev->in_s0ix)
+			r = amdgpu_userq_preempt_helper(uqm, queue);
+		else
+			r = amdgpu_userq_unmap_helper(uqm, queue);
+		if (r)
+			return r;
+	}
+	return 0;
+}
+
+int amdgpu_userq_resume(struct amdgpu_device *adev)
+{
+	u32 ip_mask = amdgpu_userq_get_supported_ip_mask(adev);
+	struct amdgpu_usermode_queue *queue;
+	struct amdgpu_userq_mgr *uqm;
+	unsigned long queue_id;
+	int r;
+
+	if (!ip_mask)
+		return 0;
+
+	xa_for_each(&adev->userq_doorbell_xa, queue_id, queue) {
+		uqm = queue->userq_mgr;
+		guard(mutex)(&uqm->userq_mutex);
+		if (adev->in_s0ix)
+			r = amdgpu_userq_restore_helper(uqm, queue);
+		else
+			r = amdgpu_userq_map_helper(uqm, queue);
+		if (r)
+			return r;
+	}
+
+	return 0;
+}
+
+int amdgpu_userq_stop_sched_for_enforce_isolation(struct amdgpu_device *adev,
+						  u32 idx)
+{
+	u32 ip_mask = amdgpu_userq_get_supported_ip_mask(adev);
+	struct amdgpu_usermode_queue *queue;
+	struct amdgpu_userq_mgr *uqm;
+	unsigned long queue_id;
+	int ret = 0, r;
+
+	/* only need to stop gfx/compute */
+	if (!(ip_mask & ((1 << AMDGPU_HW_IP_GFX) | (1 << AMDGPU_HW_IP_COMPUTE))))
+		return 0;
+
+	if (adev->userq_halt_for_enforce_isolation)
+		dev_warn(adev->dev, "userq scheduling already stopped!\n");
+	adev->userq_halt_for_enforce_isolation = true;
+	xa_for_each(&adev->userq_doorbell_xa, queue_id, queue) {
+		uqm = queue->userq_mgr;
+		cancel_delayed_work_sync(&uqm->resume_work);
+		mutex_lock(&uqm->userq_mutex);
+		if (((queue->queue_type == AMDGPU_HW_IP_GFX) ||
+		     (queue->queue_type == AMDGPU_HW_IP_COMPUTE)) &&
+		    (queue->xcp_id == idx)) {
+			amdgpu_userq_detect_and_reset_queues(uqm);
+			r = amdgpu_userq_preempt_helper(uqm, queue);
+			if (r)
+				ret = r;
+		}
+		mutex_unlock(&uqm->userq_mutex);
+	}
+
+	return ret;
+}
+
+int amdgpu_userq_start_sched_for_enforce_isolation(struct amdgpu_device *adev,
+						   u32 idx)
+{
+	u32 ip_mask = amdgpu_userq_get_supported_ip_mask(adev);
+	struct amdgpu_usermode_queue *queue;
+	struct amdgpu_userq_mgr *uqm;
+	unsigned long queue_id;
+	int ret = 0, r;
+
+	/* only need to stop gfx/compute */
+	if (!(ip_mask & ((1 << AMDGPU_HW_IP_GFX) | (1 << AMDGPU_HW_IP_COMPUTE))))
+		return 0;
+
+	if (!adev->userq_halt_for_enforce_isolation)
+		dev_warn(adev->dev, "userq scheduling already started!\n");
+	adev->userq_halt_for_enforce_isolation = false;
+	xa_for_each(&adev->userq_doorbell_xa, queue_id, queue) {
+		uqm = queue->userq_mgr;
+		mutex_lock(&uqm->userq_mutex);
+			if (((queue->queue_type == AMDGPU_HW_IP_GFX) ||
+			     (queue->queue_type == AMDGPU_HW_IP_COMPUTE)) &&
+			    (queue->xcp_id == idx)) {
+				r = amdgpu_userq_restore_helper(uqm, queue);
+				if (r)
+					ret = r;
+			}
+		mutex_unlock(&uqm->userq_mutex);
+	}
+
+	return ret;
+}
+
+int amdgpu_userq_gem_va_unmap_validate(struct amdgpu_device *adev,
+				       struct amdgpu_bo_va_mapping *mapping,
+				       uint64_t saddr)
+{
+	u32 ip_mask = amdgpu_userq_get_supported_ip_mask(adev);
+	struct amdgpu_bo_va *bo_va = mapping->bo_va;
+	struct dma_resv *resv = bo_va->base.bo->tbo.base.resv;
+	int ret = 0;
+
+	if (!ip_mask)
+		return 0;
+
+	dev_warn_once(adev->dev, "now unmapping a vital queue va:%llx\n", saddr);
+	/**
+	 * The userq VA mapping reservation should include the eviction fence,
+	 * if the eviction fence can't signal successfully during unmapping,
+	 * then driver will warn to flag this improper unmap of the userq VA.
+	 * Note: The eviction fence may be attached to different BOs, and this
+	 * unmap is only for one kind of userq VAs, so at this point suppose
+	 * the eviction fence is always unsignaled.
+	 */
+	if (!dma_resv_test_signaled(resv, DMA_RESV_USAGE_BOOKKEEP)) {
+		ret = dma_resv_wait_timeout(resv, DMA_RESV_USAGE_BOOKKEEP, true,
+					    MAX_SCHEDULE_TIMEOUT);
+		if (ret <= 0)
+			return -EBUSY;
+	}
+
+	return 0;
+}
+
+void amdgpu_userq_pre_reset(struct amdgpu_device *adev)
+{
+	const struct amdgpu_userq_funcs *userq_funcs;
+	struct amdgpu_usermode_queue *queue;
+	struct amdgpu_userq_mgr *uqm;
+	unsigned long queue_id;
+
+	xa_for_each(&adev->userq_doorbell_xa, queue_id, queue) {
+		uqm = queue->userq_mgr;
+		cancel_delayed_work_sync(&uqm->resume_work);
+		if (queue->state == AMDGPU_USERQ_STATE_MAPPED) {
+			amdgpu_userq_wait_for_last_fence(uqm, queue);
+			userq_funcs = adev->userq_funcs[queue->queue_type];
+			userq_funcs->unmap(uqm, queue);
+			/* just mark all queues as hung at this point.
+			 * if unmap succeeds, we could map again
+			 * in amdgpu_userq_post_reset() if vram is not lost
+			 */
+			queue->state = AMDGPU_USERQ_STATE_HUNG;
+			amdgpu_userq_fence_driver_force_completion(queue);
+		}
+	}
+}
+
+int amdgpu_userq_post_reset(struct amdgpu_device *adev, bool vram_lost)
+{
+	/* if any queue state is AMDGPU_USERQ_STATE_UNMAPPED
+	 * at this point, we should be able to map it again
+	 * and continue if vram is not lost.
+	 */
+	struct amdgpu_userq_mgr *uqm;
+	struct amdgpu_usermode_queue *queue;
+	const struct amdgpu_userq_funcs *userq_funcs;
+	unsigned long queue_id;
+	int r = 0;
+
+	xa_for_each(&adev->userq_doorbell_xa, queue_id, queue) {
+		uqm = queue->userq_mgr;
+		if (queue->state == AMDGPU_USERQ_STATE_HUNG && !vram_lost) {
+			userq_funcs = adev->userq_funcs[queue->queue_type];
+			/* Re-map queue */
+			r = userq_funcs->map(uqm, queue);
+			if (r) {
+				dev_err(adev->dev, "Failed to remap queue %ld\n", queue_id);
+				continue;
+			}
+			queue->state = AMDGPU_USERQ_STATE_MAPPED;
+		}
+	}
+
+	return r;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_userq.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_userq.h
new file mode 100644
index 000000000000..c37444427a14
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_userq.h
@@ -0,0 +1,161 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef AMDGPU_USERQ_H_
+#define AMDGPU_USERQ_H_
+#include "amdgpu_eviction_fence.h"
+
+#define AMDGPU_MAX_USERQ_COUNT 512
+
+#define to_ev_fence(f) container_of(f, struct amdgpu_eviction_fence, base)
+#define uq_mgr_to_fpriv(u) container_of(u, struct amdgpu_fpriv, userq_mgr)
+#define work_to_uq_mgr(w, name) container_of(w, struct amdgpu_userq_mgr, name)
+
+enum amdgpu_userq_state {
+	AMDGPU_USERQ_STATE_UNMAPPED = 0,
+	AMDGPU_USERQ_STATE_MAPPED,
+	AMDGPU_USERQ_STATE_PREEMPTED,
+	AMDGPU_USERQ_STATE_HUNG,
+	AMDGPU_USERQ_STATE_INVALID_VA,
+};
+
+struct amdgpu_mqd_prop;
+
+struct amdgpu_userq_obj {
+	void		 *cpu_ptr;
+	uint64_t	 gpu_addr;
+	struct amdgpu_bo *obj;
+};
+
+struct amdgpu_userq_va_cursor {
+	u64			gpu_addr;
+	struct list_head	list;
+};
+
+struct amdgpu_usermode_queue {
+	int			queue_type;
+	enum amdgpu_userq_state state;
+	uint64_t		doorbell_handle;
+	uint64_t		doorbell_index;
+	uint64_t		flags;
+	struct amdgpu_mqd_prop	*userq_prop;
+	struct amdgpu_userq_mgr *userq_mgr;
+	struct amdgpu_vm	*vm;
+	struct amdgpu_userq_obj mqd;
+	struct amdgpu_userq_obj	db_obj;
+	struct amdgpu_userq_obj fw_obj;
+	struct amdgpu_userq_obj wptr_obj;
+	struct xarray		fence_drv_xa;
+	struct amdgpu_userq_fence_driver *fence_drv;
+	struct dma_fence	*last_fence;
+	u32			xcp_id;
+	int			priority;
+	struct dentry		*debugfs_queue;
+
+	struct list_head	userq_va_list;
+};
+
+struct amdgpu_userq_funcs {
+	int (*mqd_create)(struct amdgpu_userq_mgr *uq_mgr,
+			  struct drm_amdgpu_userq_in *args,
+			  struct amdgpu_usermode_queue *queue);
+	void (*mqd_destroy)(struct amdgpu_userq_mgr *uq_mgr,
+			    struct amdgpu_usermode_queue *uq);
+	int (*unmap)(struct amdgpu_userq_mgr *uq_mgr,
+		     struct amdgpu_usermode_queue *queue);
+	int (*map)(struct amdgpu_userq_mgr *uq_mgr,
+		   struct amdgpu_usermode_queue *queue);
+	int (*preempt)(struct amdgpu_userq_mgr *uq_mgr,
+		   struct amdgpu_usermode_queue *queue);
+	int (*restore)(struct amdgpu_userq_mgr *uq_mgr,
+		   struct amdgpu_usermode_queue *queue);
+	int (*detect_and_reset)(struct amdgpu_device *adev,
+		  int queue_type);
+};
+
+/* Usermode queues for gfx */
+struct amdgpu_userq_mgr {
+	/**
+	 * @userq_mgr_xa: Per-process user queue map (queue ID → queue)
+	 * Key: queue_id (unique ID within the process's userq manager)
+	 * Value: struct amdgpu_usermode_queue
+	 */
+	struct xarray			userq_mgr_xa;
+	struct mutex			userq_mutex;
+	struct amdgpu_device		*adev;
+	struct delayed_work		resume_work;
+	struct drm_file			*file;
+	atomic_t                        userq_count[AMDGPU_RING_TYPE_MAX];
+};
+
+struct amdgpu_db_info {
+	uint64_t doorbell_handle;
+	uint32_t queue_type;
+	uint32_t doorbell_offset;
+	struct amdgpu_userq_obj	*db_obj;
+};
+
+int amdgpu_userq_ioctl(struct drm_device *dev, void *data, struct drm_file *filp);
+
+int amdgpu_userq_mgr_init(struct amdgpu_userq_mgr *userq_mgr, struct drm_file *file_priv,
+			  struct amdgpu_device *adev);
+
+void amdgpu_userq_mgr_fini(struct amdgpu_userq_mgr *userq_mgr);
+
+int amdgpu_userq_create_object(struct amdgpu_userq_mgr *uq_mgr,
+			       struct amdgpu_userq_obj *userq_obj,
+			       int size);
+
+void amdgpu_userq_destroy_object(struct amdgpu_userq_mgr *uq_mgr,
+				 struct amdgpu_userq_obj *userq_obj);
+
+void amdgpu_userq_evict(struct amdgpu_userq_mgr *uq_mgr,
+			struct amdgpu_eviction_fence *ev_fence);
+
+void amdgpu_userq_ensure_ev_fence(struct amdgpu_userq_mgr *userq_mgr,
+				  struct amdgpu_eviction_fence_mgr *evf_mgr);
+
+uint64_t amdgpu_userq_get_doorbell_index(struct amdgpu_userq_mgr *uq_mgr,
+					 struct amdgpu_db_info *db_info,
+					     struct drm_file *filp);
+
+u32 amdgpu_userq_get_supported_ip_mask(struct amdgpu_device *adev);
+
+int amdgpu_userq_suspend(struct amdgpu_device *adev);
+int amdgpu_userq_resume(struct amdgpu_device *adev);
+
+int amdgpu_userq_stop_sched_for_enforce_isolation(struct amdgpu_device *adev,
+						  u32 idx);
+int amdgpu_userq_start_sched_for_enforce_isolation(struct amdgpu_device *adev,
+						   u32 idx);
+void amdgpu_userq_reset_work(struct work_struct *work);
+void amdgpu_userq_pre_reset(struct amdgpu_device *adev);
+int amdgpu_userq_post_reset(struct amdgpu_device *adev, bool vram_lost);
+
+int amdgpu_userq_input_va_validate(struct amdgpu_usermode_queue *queue,
+				   u64 addr, u64 expected_size);
+int amdgpu_userq_gem_va_unmap_validate(struct amdgpu_device *adev,
+				       struct amdgpu_bo_va_mapping *mapping,
+				       uint64_t saddr);
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_userq_fence.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_userq_fence.c
new file mode 100644
index 000000000000..eba9fb359047
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_userq_fence.c
@@ -0,0 +1,1011 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/kref.h>
+#include <linux/slab.h>
+#include <linux/dma-fence-unwrap.h>
+
+#include <drm/drm_exec.h>
+#include <drm/drm_syncobj.h>
+
+#include "amdgpu.h"
+#include "amdgpu_userq_fence.h"
+
+static const struct dma_fence_ops amdgpu_userq_fence_ops;
+static struct kmem_cache *amdgpu_userq_fence_slab;
+
+int amdgpu_userq_fence_slab_init(void)
+{
+	amdgpu_userq_fence_slab = kmem_cache_create("amdgpu_userq_fence",
+						    sizeof(struct amdgpu_userq_fence),
+						    0,
+						    SLAB_HWCACHE_ALIGN,
+						    NULL);
+	if (!amdgpu_userq_fence_slab)
+		return -ENOMEM;
+
+	return 0;
+}
+
+void amdgpu_userq_fence_slab_fini(void)
+{
+	rcu_barrier();
+	kmem_cache_destroy(amdgpu_userq_fence_slab);
+}
+
+static inline struct amdgpu_userq_fence *to_amdgpu_userq_fence(struct dma_fence *f)
+{
+	if (!f || f->ops != &amdgpu_userq_fence_ops)
+		return NULL;
+
+	return container_of(f, struct amdgpu_userq_fence, base);
+}
+
+static u64 amdgpu_userq_fence_read(struct amdgpu_userq_fence_driver *fence_drv)
+{
+	return le64_to_cpu(*fence_drv->cpu_addr);
+}
+
+static void
+amdgpu_userq_fence_write(struct amdgpu_userq_fence_driver *fence_drv,
+			 u64 seq)
+{
+	if (fence_drv->cpu_addr)
+		*fence_drv->cpu_addr = cpu_to_le64(seq);
+}
+
+int amdgpu_userq_fence_driver_alloc(struct amdgpu_device *adev,
+				    struct amdgpu_usermode_queue *userq)
+{
+	struct amdgpu_userq_fence_driver *fence_drv;
+	unsigned long flags;
+	int r;
+
+	fence_drv = kzalloc(sizeof(*fence_drv), GFP_KERNEL);
+	if (!fence_drv)
+		return -ENOMEM;
+
+	/* Acquire seq64 memory */
+	r = amdgpu_seq64_alloc(adev, &fence_drv->va, &fence_drv->gpu_addr,
+			       &fence_drv->cpu_addr);
+	if (r)
+		goto free_fence_drv;
+
+	memset(fence_drv->cpu_addr, 0, sizeof(u64));
+
+	kref_init(&fence_drv->refcount);
+	INIT_LIST_HEAD(&fence_drv->fences);
+	spin_lock_init(&fence_drv->fence_list_lock);
+
+	fence_drv->adev = adev;
+	fence_drv->context = dma_fence_context_alloc(1);
+	get_task_comm(fence_drv->timeline_name, current);
+
+	xa_lock_irqsave(&adev->userq_xa, flags);
+	r = xa_err(__xa_store(&adev->userq_xa, userq->doorbell_index,
+			      fence_drv, GFP_KERNEL));
+	xa_unlock_irqrestore(&adev->userq_xa, flags);
+	if (r)
+		goto free_seq64;
+
+	userq->fence_drv = fence_drv;
+
+	return 0;
+
+free_seq64:
+	amdgpu_seq64_free(adev, fence_drv->va);
+free_fence_drv:
+	kfree(fence_drv);
+
+	return r;
+}
+
+static void amdgpu_userq_walk_and_drop_fence_drv(struct xarray *xa)
+{
+	struct amdgpu_userq_fence_driver *fence_drv;
+	unsigned long index;
+
+	if (xa_empty(xa))
+		return;
+
+	xa_lock(xa);
+	xa_for_each(xa, index, fence_drv) {
+		__xa_erase(xa, index);
+		amdgpu_userq_fence_driver_put(fence_drv);
+	}
+
+	xa_unlock(xa);
+}
+
+void
+amdgpu_userq_fence_driver_free(struct amdgpu_usermode_queue *userq)
+{
+	amdgpu_userq_walk_and_drop_fence_drv(&userq->fence_drv_xa);
+	xa_destroy(&userq->fence_drv_xa);
+	/* Drop the fence_drv reference held by user queue */
+	amdgpu_userq_fence_driver_put(userq->fence_drv);
+}
+
+void amdgpu_userq_fence_driver_process(struct amdgpu_userq_fence_driver *fence_drv)
+{
+	struct amdgpu_userq_fence *userq_fence, *tmp;
+	struct dma_fence *fence;
+	unsigned long flags;
+	u64 rptr;
+	int i;
+
+	if (!fence_drv)
+		return;
+
+	spin_lock_irqsave(&fence_drv->fence_list_lock, flags);
+	rptr = amdgpu_userq_fence_read(fence_drv);
+
+	list_for_each_entry_safe(userq_fence, tmp, &fence_drv->fences, link) {
+		fence = &userq_fence->base;
+
+		if (rptr < fence->seqno)
+			break;
+
+		dma_fence_signal(fence);
+
+		for (i = 0; i < userq_fence->fence_drv_array_count; i++)
+			amdgpu_userq_fence_driver_put(userq_fence->fence_drv_array[i]);
+
+		list_del(&userq_fence->link);
+		dma_fence_put(fence);
+	}
+	spin_unlock_irqrestore(&fence_drv->fence_list_lock, flags);
+}
+
+void amdgpu_userq_fence_driver_destroy(struct kref *ref)
+{
+	struct amdgpu_userq_fence_driver *fence_drv = container_of(ref,
+					 struct amdgpu_userq_fence_driver,
+					 refcount);
+	struct amdgpu_userq_fence_driver *xa_fence_drv;
+	struct amdgpu_device *adev = fence_drv->adev;
+	struct amdgpu_userq_fence *fence, *tmp;
+	struct xarray *xa = &adev->userq_xa;
+	unsigned long index, flags;
+	struct dma_fence *f;
+
+	spin_lock_irqsave(&fence_drv->fence_list_lock, flags);
+	list_for_each_entry_safe(fence, tmp, &fence_drv->fences, link) {
+		f = &fence->base;
+
+		if (!dma_fence_is_signaled(f)) {
+			dma_fence_set_error(f, -ECANCELED);
+			dma_fence_signal(f);
+		}
+
+		list_del(&fence->link);
+		dma_fence_put(f);
+	}
+	spin_unlock_irqrestore(&fence_drv->fence_list_lock, flags);
+
+	xa_lock_irqsave(xa, flags);
+	xa_for_each(xa, index, xa_fence_drv)
+		if (xa_fence_drv == fence_drv)
+			__xa_erase(xa, index);
+	xa_unlock_irqrestore(xa, flags);
+
+	/* Free seq64 memory */
+	amdgpu_seq64_free(adev, fence_drv->va);
+	kfree(fence_drv);
+}
+
+void amdgpu_userq_fence_driver_get(struct amdgpu_userq_fence_driver *fence_drv)
+{
+	kref_get(&fence_drv->refcount);
+}
+
+void amdgpu_userq_fence_driver_put(struct amdgpu_userq_fence_driver *fence_drv)
+{
+	kref_put(&fence_drv->refcount, amdgpu_userq_fence_driver_destroy);
+}
+
+static int amdgpu_userq_fence_alloc(struct amdgpu_userq_fence **userq_fence)
+{
+	*userq_fence = kmem_cache_alloc(amdgpu_userq_fence_slab, GFP_ATOMIC);
+	return *userq_fence ? 0 : -ENOMEM;
+}
+
+static int amdgpu_userq_fence_create(struct amdgpu_usermode_queue *userq,
+				     struct amdgpu_userq_fence *userq_fence,
+				     u64 seq, struct dma_fence **f)
+{
+	struct amdgpu_userq_fence_driver *fence_drv;
+	struct dma_fence *fence;
+	unsigned long flags;
+
+	fence_drv = userq->fence_drv;
+	if (!fence_drv)
+		return -EINVAL;
+
+	spin_lock_init(&userq_fence->lock);
+	INIT_LIST_HEAD(&userq_fence->link);
+	fence = &userq_fence->base;
+	userq_fence->fence_drv = fence_drv;
+
+	dma_fence_init64(fence, &amdgpu_userq_fence_ops, &userq_fence->lock,
+			 fence_drv->context, seq);
+
+	amdgpu_userq_fence_driver_get(fence_drv);
+	dma_fence_get(fence);
+
+	if (!xa_empty(&userq->fence_drv_xa)) {
+		struct amdgpu_userq_fence_driver *stored_fence_drv;
+		unsigned long index, count = 0;
+		int i = 0;
+
+		xa_lock(&userq->fence_drv_xa);
+		xa_for_each(&userq->fence_drv_xa, index, stored_fence_drv)
+			count++;
+
+		userq_fence->fence_drv_array =
+			kvmalloc_array(count,
+				       sizeof(struct amdgpu_userq_fence_driver *),
+				       GFP_ATOMIC);
+
+		if (userq_fence->fence_drv_array) {
+			xa_for_each(&userq->fence_drv_xa, index, stored_fence_drv) {
+				userq_fence->fence_drv_array[i] = stored_fence_drv;
+				__xa_erase(&userq->fence_drv_xa, index);
+				i++;
+			}
+		}
+
+		userq_fence->fence_drv_array_count = i;
+		xa_unlock(&userq->fence_drv_xa);
+	} else {
+		userq_fence->fence_drv_array = NULL;
+		userq_fence->fence_drv_array_count = 0;
+	}
+
+	/* Check if hardware has already processed the job */
+	spin_lock_irqsave(&fence_drv->fence_list_lock, flags);
+	if (!dma_fence_is_signaled(fence))
+		list_add_tail(&userq_fence->link, &fence_drv->fences);
+	else
+		dma_fence_put(fence);
+
+	spin_unlock_irqrestore(&fence_drv->fence_list_lock, flags);
+
+	*f = fence;
+
+	return 0;
+}
+
+static const char *amdgpu_userq_fence_get_driver_name(struct dma_fence *f)
+{
+	return "amdgpu_userq_fence";
+}
+
+static const char *amdgpu_userq_fence_get_timeline_name(struct dma_fence *f)
+{
+	struct amdgpu_userq_fence *fence = to_amdgpu_userq_fence(f);
+
+	return fence->fence_drv->timeline_name;
+}
+
+static bool amdgpu_userq_fence_signaled(struct dma_fence *f)
+{
+	struct amdgpu_userq_fence *fence = to_amdgpu_userq_fence(f);
+	struct amdgpu_userq_fence_driver *fence_drv = fence->fence_drv;
+	u64 rptr, wptr;
+
+	rptr = amdgpu_userq_fence_read(fence_drv);
+	wptr = fence->base.seqno;
+
+	if (rptr >= wptr)
+		return true;
+
+	return false;
+}
+
+static void amdgpu_userq_fence_free(struct rcu_head *rcu)
+{
+	struct dma_fence *fence = container_of(rcu, struct dma_fence, rcu);
+	struct amdgpu_userq_fence *userq_fence = to_amdgpu_userq_fence(fence);
+	struct amdgpu_userq_fence_driver *fence_drv = userq_fence->fence_drv;
+
+	/* Release the fence driver reference */
+	amdgpu_userq_fence_driver_put(fence_drv);
+
+	kvfree(userq_fence->fence_drv_array);
+	kmem_cache_free(amdgpu_userq_fence_slab, userq_fence);
+}
+
+static void amdgpu_userq_fence_release(struct dma_fence *f)
+{
+	call_rcu(&f->rcu, amdgpu_userq_fence_free);
+}
+
+static const struct dma_fence_ops amdgpu_userq_fence_ops = {
+	.get_driver_name = amdgpu_userq_fence_get_driver_name,
+	.get_timeline_name = amdgpu_userq_fence_get_timeline_name,
+	.signaled = amdgpu_userq_fence_signaled,
+	.release = amdgpu_userq_fence_release,
+};
+
+/**
+ * amdgpu_userq_fence_read_wptr - Read the userq wptr value
+ *
+ * @queue: user mode queue structure pointer
+ * @wptr: write pointer value
+ *
+ * Read the wptr value from userq's MQD. The userq signal IOCTL
+ * creates a dma_fence for the shared buffers that expects the
+ * RPTR value written to seq64 memory >= WPTR.
+ *
+ * Returns wptr value on success, error on failure.
+ */
+static int amdgpu_userq_fence_read_wptr(struct amdgpu_usermode_queue *queue,
+					u64 *wptr)
+{
+	struct amdgpu_bo_va_mapping *mapping;
+	struct amdgpu_bo *bo;
+	u64 addr, *ptr;
+	int r;
+
+	r = amdgpu_bo_reserve(queue->vm->root.bo, false);
+	if (r)
+		return r;
+
+	addr = queue->userq_prop->wptr_gpu_addr;
+	addr &= AMDGPU_GMC_HOLE_MASK;
+
+	mapping = amdgpu_vm_bo_lookup_mapping(queue->vm, addr >> PAGE_SHIFT);
+	if (!mapping) {
+		amdgpu_bo_unreserve(queue->vm->root.bo);
+		DRM_ERROR("Failed to lookup amdgpu_bo_va_mapping\n");
+		return -EINVAL;
+	}
+
+	bo = amdgpu_bo_ref(mapping->bo_va->base.bo);
+	amdgpu_bo_unreserve(queue->vm->root.bo);
+	r = amdgpu_bo_reserve(bo, true);
+	if (r) {
+		amdgpu_bo_unref(&bo);
+		DRM_ERROR("Failed to reserve userqueue wptr bo");
+		return r;
+	}
+
+	r = amdgpu_bo_kmap(bo, (void **)&ptr);
+	if (r) {
+		DRM_ERROR("Failed mapping the userqueue wptr bo");
+		goto map_error;
+	}
+
+	*wptr = le64_to_cpu(*ptr);
+
+	amdgpu_bo_kunmap(bo);
+	amdgpu_bo_unreserve(bo);
+	amdgpu_bo_unref(&bo);
+
+	return 0;
+
+map_error:
+	amdgpu_bo_unreserve(bo);
+	amdgpu_bo_unref(&bo);
+
+	return r;
+}
+
+static void amdgpu_userq_fence_cleanup(struct dma_fence *fence)
+{
+	dma_fence_put(fence);
+}
+
+static void
+amdgpu_userq_fence_driver_set_error(struct amdgpu_userq_fence *fence,
+				    int error)
+{
+	struct amdgpu_userq_fence_driver *fence_drv = fence->fence_drv;
+	unsigned long flags;
+	struct dma_fence *f;
+
+	spin_lock_irqsave(&fence_drv->fence_list_lock, flags);
+
+	f = rcu_dereference_protected(&fence->base,
+				      lockdep_is_held(&fence_drv->fence_list_lock));
+	if (f && !dma_fence_is_signaled_locked(f))
+		dma_fence_set_error(f, error);
+	spin_unlock_irqrestore(&fence_drv->fence_list_lock, flags);
+}
+
+void
+amdgpu_userq_fence_driver_force_completion(struct amdgpu_usermode_queue *userq)
+{
+	struct dma_fence *f = userq->last_fence;
+
+	if (f) {
+		struct amdgpu_userq_fence *fence = to_amdgpu_userq_fence(f);
+		struct amdgpu_userq_fence_driver *fence_drv = fence->fence_drv;
+		u64 wptr = fence->base.seqno;
+
+		amdgpu_userq_fence_driver_set_error(fence, -ECANCELED);
+		amdgpu_userq_fence_write(fence_drv, wptr);
+		amdgpu_userq_fence_driver_process(fence_drv);
+
+	}
+}
+
+int amdgpu_userq_signal_ioctl(struct drm_device *dev, void *data,
+			      struct drm_file *filp)
+{
+	struct amdgpu_fpriv *fpriv = filp->driver_priv;
+	struct amdgpu_userq_mgr *userq_mgr = &fpriv->userq_mgr;
+	struct drm_amdgpu_userq_signal *args = data;
+	struct drm_gem_object **gobj_write = NULL;
+	struct drm_gem_object **gobj_read = NULL;
+	struct amdgpu_usermode_queue *queue;
+	struct amdgpu_userq_fence *userq_fence;
+	struct drm_syncobj **syncobj = NULL;
+	u32 *bo_handles_write, num_write_bo_handles;
+	u32 *syncobj_handles, num_syncobj_handles;
+	u32 *bo_handles_read, num_read_bo_handles;
+	int r, i, entry, rentry, wentry;
+	struct dma_fence *fence;
+	struct drm_exec exec;
+	u64 wptr;
+
+	num_syncobj_handles = args->num_syncobj_handles;
+	syncobj_handles = memdup_user(u64_to_user_ptr(args->syncobj_handles),
+				      size_mul(sizeof(u32), num_syncobj_handles));
+	if (IS_ERR(syncobj_handles))
+		return PTR_ERR(syncobj_handles);
+
+	/* Array of pointers to the looked up syncobjs */
+	syncobj = kmalloc_array(num_syncobj_handles, sizeof(*syncobj), GFP_KERNEL);
+	if (!syncobj) {
+		r = -ENOMEM;
+		goto free_syncobj_handles;
+	}
+
+	for (entry = 0; entry < num_syncobj_handles; entry++) {
+		syncobj[entry] = drm_syncobj_find(filp, syncobj_handles[entry]);
+		if (!syncobj[entry]) {
+			r = -ENOENT;
+			goto free_syncobj;
+		}
+	}
+
+	num_read_bo_handles = args->num_bo_read_handles;
+	bo_handles_read = memdup_user(u64_to_user_ptr(args->bo_read_handles),
+				      sizeof(u32) * num_read_bo_handles);
+	if (IS_ERR(bo_handles_read)) {
+		r = PTR_ERR(bo_handles_read);
+		goto free_syncobj;
+	}
+
+	/* Array of pointers to the GEM read objects */
+	gobj_read = kmalloc_array(num_read_bo_handles, sizeof(*gobj_read), GFP_KERNEL);
+	if (!gobj_read) {
+		r = -ENOMEM;
+		goto free_bo_handles_read;
+	}
+
+	for (rentry = 0; rentry < num_read_bo_handles; rentry++) {
+		gobj_read[rentry] = drm_gem_object_lookup(filp, bo_handles_read[rentry]);
+		if (!gobj_read[rentry]) {
+			r = -ENOENT;
+			goto put_gobj_read;
+		}
+	}
+
+	num_write_bo_handles = args->num_bo_write_handles;
+	bo_handles_write = memdup_user(u64_to_user_ptr(args->bo_write_handles),
+				       sizeof(u32) * num_write_bo_handles);
+	if (IS_ERR(bo_handles_write)) {
+		r = PTR_ERR(bo_handles_write);
+		goto put_gobj_read;
+	}
+
+	/* Array of pointers to the GEM write objects */
+	gobj_write = kmalloc_array(num_write_bo_handles, sizeof(*gobj_write), GFP_KERNEL);
+	if (!gobj_write) {
+		r = -ENOMEM;
+		goto free_bo_handles_write;
+	}
+
+	for (wentry = 0; wentry < num_write_bo_handles; wentry++) {
+		gobj_write[wentry] = drm_gem_object_lookup(filp, bo_handles_write[wentry]);
+		if (!gobj_write[wentry]) {
+			r = -ENOENT;
+			goto put_gobj_write;
+		}
+	}
+
+	/* Retrieve the user queue */
+	queue = xa_load(&userq_mgr->userq_mgr_xa, args->queue_id);
+	if (!queue) {
+		r = -ENOENT;
+		goto put_gobj_write;
+	}
+
+	r = amdgpu_userq_fence_read_wptr(queue, &wptr);
+	if (r)
+		goto put_gobj_write;
+
+	r = amdgpu_userq_fence_alloc(&userq_fence);
+	if (r)
+		goto put_gobj_write;
+
+	/* We are here means UQ is active, make sure the eviction fence is valid */
+	amdgpu_userq_ensure_ev_fence(&fpriv->userq_mgr, &fpriv->evf_mgr);
+
+	/* Create a new fence */
+	r = amdgpu_userq_fence_create(queue, userq_fence, wptr, &fence);
+	if (r) {
+		mutex_unlock(&userq_mgr->userq_mutex);
+		kmem_cache_free(amdgpu_userq_fence_slab, userq_fence);
+		goto put_gobj_write;
+	}
+
+	dma_fence_put(queue->last_fence);
+	queue->last_fence = dma_fence_get(fence);
+	mutex_unlock(&userq_mgr->userq_mutex);
+
+	drm_exec_init(&exec, DRM_EXEC_INTERRUPTIBLE_WAIT,
+		      (num_read_bo_handles + num_write_bo_handles));
+
+	/* Lock all BOs with retry handling */
+	drm_exec_until_all_locked(&exec) {
+		r = drm_exec_prepare_array(&exec, gobj_read, num_read_bo_handles, 1);
+		drm_exec_retry_on_contention(&exec);
+		if (r) {
+			amdgpu_userq_fence_cleanup(fence);
+			goto exec_fini;
+		}
+
+		r = drm_exec_prepare_array(&exec, gobj_write, num_write_bo_handles, 1);
+		drm_exec_retry_on_contention(&exec);
+		if (r) {
+			amdgpu_userq_fence_cleanup(fence);
+			goto exec_fini;
+		}
+	}
+
+	for (i = 0; i < num_read_bo_handles; i++) {
+		if (!gobj_read || !gobj_read[i]->resv)
+			continue;
+
+		dma_resv_add_fence(gobj_read[i]->resv, fence,
+				   DMA_RESV_USAGE_READ);
+	}
+
+	for (i = 0; i < num_write_bo_handles; i++) {
+		if (!gobj_write || !gobj_write[i]->resv)
+			continue;
+
+		dma_resv_add_fence(gobj_write[i]->resv, fence,
+				   DMA_RESV_USAGE_WRITE);
+	}
+
+	/* Add the created fence to syncobj/BO's */
+	for (i = 0; i < num_syncobj_handles; i++)
+		drm_syncobj_replace_fence(syncobj[i], fence);
+
+	/* drop the reference acquired in fence creation function */
+	dma_fence_put(fence);
+
+exec_fini:
+	drm_exec_fini(&exec);
+put_gobj_write:
+	while (wentry-- > 0)
+		drm_gem_object_put(gobj_write[wentry]);
+	kfree(gobj_write);
+free_bo_handles_write:
+	kfree(bo_handles_write);
+put_gobj_read:
+	while (rentry-- > 0)
+		drm_gem_object_put(gobj_read[rentry]);
+	kfree(gobj_read);
+free_bo_handles_read:
+	kfree(bo_handles_read);
+free_syncobj:
+	while (entry-- > 0)
+		if (syncobj[entry])
+			drm_syncobj_put(syncobj[entry]);
+	kfree(syncobj);
+free_syncobj_handles:
+	kfree(syncobj_handles);
+
+	return r;
+}
+
+int amdgpu_userq_wait_ioctl(struct drm_device *dev, void *data,
+			    struct drm_file *filp)
+{
+	u32 *syncobj_handles, *timeline_points, *timeline_handles, *bo_handles_read, *bo_handles_write;
+	u32 num_syncobj, num_read_bo_handles, num_write_bo_handles;
+	struct drm_amdgpu_userq_fence_info *fence_info = NULL;
+	struct drm_amdgpu_userq_wait *wait_info = data;
+	struct amdgpu_fpriv *fpriv = filp->driver_priv;
+	struct amdgpu_userq_mgr *userq_mgr = &fpriv->userq_mgr;
+	struct amdgpu_usermode_queue *waitq;
+	struct drm_gem_object **gobj_write;
+	struct drm_gem_object **gobj_read;
+	struct dma_fence **fences = NULL;
+	u16 num_points, num_fences = 0;
+	int r, i, rentry, wentry, cnt;
+	struct drm_exec exec;
+
+	num_read_bo_handles = wait_info->num_bo_read_handles;
+	bo_handles_read = memdup_user(u64_to_user_ptr(wait_info->bo_read_handles),
+				      size_mul(sizeof(u32), num_read_bo_handles));
+	if (IS_ERR(bo_handles_read))
+		return PTR_ERR(bo_handles_read);
+
+	num_write_bo_handles = wait_info->num_bo_write_handles;
+	bo_handles_write = memdup_user(u64_to_user_ptr(wait_info->bo_write_handles),
+				       size_mul(sizeof(u32), num_write_bo_handles));
+	if (IS_ERR(bo_handles_write)) {
+		r = PTR_ERR(bo_handles_write);
+		goto free_bo_handles_read;
+	}
+
+	num_syncobj = wait_info->num_syncobj_handles;
+	syncobj_handles = memdup_user(u64_to_user_ptr(wait_info->syncobj_handles),
+				      size_mul(sizeof(u32), num_syncobj));
+	if (IS_ERR(syncobj_handles)) {
+		r = PTR_ERR(syncobj_handles);
+		goto free_bo_handles_write;
+	}
+
+	num_points = wait_info->num_syncobj_timeline_handles;
+	timeline_handles = memdup_user(u64_to_user_ptr(wait_info->syncobj_timeline_handles),
+				       sizeof(u32) * num_points);
+	if (IS_ERR(timeline_handles)) {
+		r = PTR_ERR(timeline_handles);
+		goto free_syncobj_handles;
+	}
+
+	timeline_points = memdup_user(u64_to_user_ptr(wait_info->syncobj_timeline_points),
+				      sizeof(u32) * num_points);
+	if (IS_ERR(timeline_points)) {
+		r = PTR_ERR(timeline_points);
+		goto free_timeline_handles;
+	}
+
+	gobj_read = kmalloc_array(num_read_bo_handles, sizeof(*gobj_read), GFP_KERNEL);
+	if (!gobj_read) {
+		r = -ENOMEM;
+		goto free_timeline_points;
+	}
+
+	for (rentry = 0; rentry < num_read_bo_handles; rentry++) {
+		gobj_read[rentry] = drm_gem_object_lookup(filp, bo_handles_read[rentry]);
+		if (!gobj_read[rentry]) {
+			r = -ENOENT;
+			goto put_gobj_read;
+		}
+	}
+
+	gobj_write = kmalloc_array(num_write_bo_handles, sizeof(*gobj_write), GFP_KERNEL);
+	if (!gobj_write) {
+		r = -ENOMEM;
+		goto put_gobj_read;
+	}
+
+	for (wentry = 0; wentry < num_write_bo_handles; wentry++) {
+		gobj_write[wentry] = drm_gem_object_lookup(filp, bo_handles_write[wentry]);
+		if (!gobj_write[wentry]) {
+			r = -ENOENT;
+			goto put_gobj_write;
+		}
+	}
+
+	drm_exec_init(&exec, DRM_EXEC_INTERRUPTIBLE_WAIT,
+		      (num_read_bo_handles + num_write_bo_handles));
+
+	/* Lock all BOs with retry handling */
+	drm_exec_until_all_locked(&exec) {
+		r = drm_exec_prepare_array(&exec, gobj_read, num_read_bo_handles, 1);
+		drm_exec_retry_on_contention(&exec);
+		if (r) {
+			drm_exec_fini(&exec);
+			goto put_gobj_write;
+		}
+
+		r = drm_exec_prepare_array(&exec, gobj_write, num_write_bo_handles, 1);
+		drm_exec_retry_on_contention(&exec);
+		if (r) {
+			drm_exec_fini(&exec);
+			goto put_gobj_write;
+		}
+	}
+
+	if (!wait_info->num_fences) {
+		if (num_points) {
+			struct dma_fence_unwrap iter;
+			struct dma_fence *fence;
+			struct dma_fence *f;
+
+			for (i = 0; i < num_points; i++) {
+				r = drm_syncobj_find_fence(filp, timeline_handles[i],
+							   timeline_points[i],
+							   DRM_SYNCOBJ_WAIT_FLAGS_WAIT_FOR_SUBMIT,
+							   &fence);
+				if (r)
+					goto exec_fini;
+
+				dma_fence_unwrap_for_each(f, &iter, fence)
+					num_fences++;
+
+				dma_fence_put(fence);
+			}
+		}
+
+		/* Count syncobj's fence */
+		for (i = 0; i < num_syncobj; i++) {
+			struct dma_fence *fence;
+
+			r = drm_syncobj_find_fence(filp, syncobj_handles[i],
+						   0,
+						   DRM_SYNCOBJ_WAIT_FLAGS_WAIT_FOR_SUBMIT,
+						   &fence);
+			if (r)
+				goto exec_fini;
+
+			num_fences++;
+			dma_fence_put(fence);
+		}
+
+		/* Count GEM objects fence */
+		for (i = 0; i < num_read_bo_handles; i++) {
+			struct dma_resv_iter resv_cursor;
+			struct dma_fence *fence;
+
+			dma_resv_for_each_fence(&resv_cursor, gobj_read[i]->resv,
+						DMA_RESV_USAGE_READ, fence)
+				num_fences++;
+		}
+
+		for (i = 0; i < num_write_bo_handles; i++) {
+			struct dma_resv_iter resv_cursor;
+			struct dma_fence *fence;
+
+			dma_resv_for_each_fence(&resv_cursor, gobj_write[i]->resv,
+						DMA_RESV_USAGE_WRITE, fence)
+				num_fences++;
+		}
+
+		/*
+		 * Passing num_fences = 0 means that userspace doesn't want to
+		 * retrieve userq_fence_info. If num_fences = 0 we skip filling
+		 * userq_fence_info and return the actual number of fences on
+		 * args->num_fences.
+		 */
+		wait_info->num_fences = num_fences;
+	} else {
+		/* Array of fence info */
+		fence_info = kmalloc_array(wait_info->num_fences, sizeof(*fence_info), GFP_KERNEL);
+		if (!fence_info) {
+			r = -ENOMEM;
+			goto exec_fini;
+		}
+
+		/* Array of fences */
+		fences = kmalloc_array(wait_info->num_fences, sizeof(*fences), GFP_KERNEL);
+		if (!fences) {
+			r = -ENOMEM;
+			goto free_fence_info;
+		}
+
+		/* Retrieve GEM read objects fence */
+		for (i = 0; i < num_read_bo_handles; i++) {
+			struct dma_resv_iter resv_cursor;
+			struct dma_fence *fence;
+
+			dma_resv_for_each_fence(&resv_cursor, gobj_read[i]->resv,
+						DMA_RESV_USAGE_READ, fence) {
+				if (WARN_ON_ONCE(num_fences >= wait_info->num_fences)) {
+					r = -EINVAL;
+					goto free_fences;
+				}
+
+				fences[num_fences++] = fence;
+				dma_fence_get(fence);
+			}
+		}
+
+		/* Retrieve GEM write objects fence */
+		for (i = 0; i < num_write_bo_handles; i++) {
+			struct dma_resv_iter resv_cursor;
+			struct dma_fence *fence;
+
+			dma_resv_for_each_fence(&resv_cursor, gobj_write[i]->resv,
+						DMA_RESV_USAGE_WRITE, fence) {
+				if (WARN_ON_ONCE(num_fences >= wait_info->num_fences)) {
+					r = -EINVAL;
+					goto free_fences;
+				}
+
+				fences[num_fences++] = fence;
+				dma_fence_get(fence);
+			}
+		}
+
+		if (num_points) {
+			struct dma_fence_unwrap iter;
+			struct dma_fence *fence;
+			struct dma_fence *f;
+
+			for (i = 0; i < num_points; i++) {
+				r = drm_syncobj_find_fence(filp, timeline_handles[i],
+							   timeline_points[i],
+							   DRM_SYNCOBJ_WAIT_FLAGS_WAIT_FOR_SUBMIT,
+							   &fence);
+				if (r)
+					goto free_fences;
+
+				dma_fence_unwrap_for_each(f, &iter, fence) {
+					if (WARN_ON_ONCE(num_fences >= wait_info->num_fences)) {
+						r = -EINVAL;
+						goto free_fences;
+					}
+
+					dma_fence_get(f);
+					fences[num_fences++] = f;
+				}
+
+				dma_fence_put(fence);
+			}
+		}
+
+		/* Retrieve syncobj's fence */
+		for (i = 0; i < num_syncobj; i++) {
+			struct dma_fence *fence;
+
+			r = drm_syncobj_find_fence(filp, syncobj_handles[i],
+						   0,
+						   DRM_SYNCOBJ_WAIT_FLAGS_WAIT_FOR_SUBMIT,
+						   &fence);
+			if (r)
+				goto free_fences;
+
+			if (WARN_ON_ONCE(num_fences >= wait_info->num_fences)) {
+				r = -EINVAL;
+				goto free_fences;
+			}
+
+			fences[num_fences++] = fence;
+		}
+
+		/*
+		 * Keep only the latest fences to reduce the number of values
+		 * given back to userspace.
+		 */
+		num_fences = dma_fence_dedup_array(fences, num_fences);
+
+		waitq = xa_load(&userq_mgr->userq_mgr_xa, wait_info->waitq_id);
+		if (!waitq) {
+			r = -EINVAL;
+			goto free_fences;
+		}
+
+		for (i = 0, cnt = 0; i < num_fences; i++) {
+			struct amdgpu_userq_fence_driver *fence_drv;
+			struct amdgpu_userq_fence *userq_fence;
+			u32 index;
+
+			userq_fence = to_amdgpu_userq_fence(fences[i]);
+			if (!userq_fence) {
+				/*
+				 * Just waiting on other driver fences should
+				 * be good for now
+				 */
+				r = dma_fence_wait(fences[i], true);
+				if (r) {
+					dma_fence_put(fences[i]);
+					goto free_fences;
+				}
+
+				dma_fence_put(fences[i]);
+				continue;
+			}
+
+			fence_drv = userq_fence->fence_drv;
+			/*
+			 * We need to make sure the user queue release their reference
+			 * to the fence drivers at some point before queue destruction.
+			 * Otherwise, we would gather those references until we don't
+			 * have any more space left and crash.
+			 */
+			r = xa_alloc(&waitq->fence_drv_xa, &index, fence_drv,
+				     xa_limit_32b, GFP_KERNEL);
+			if (r)
+				goto free_fences;
+
+			amdgpu_userq_fence_driver_get(fence_drv);
+
+			/* Store drm syncobj's gpu va address and value */
+			fence_info[cnt].va = fence_drv->va;
+			fence_info[cnt].value = fences[i]->seqno;
+
+			dma_fence_put(fences[i]);
+			/* Increment the actual userq fence count */
+			cnt++;
+		}
+
+		wait_info->num_fences = cnt;
+		/* Copy userq fence info to user space */
+		if (copy_to_user(u64_to_user_ptr(wait_info->out_fences),
+				 fence_info, wait_info->num_fences * sizeof(*fence_info))) {
+			r = -EFAULT;
+			goto free_fences;
+		}
+
+		kfree(fences);
+		kfree(fence_info);
+	}
+
+	drm_exec_fini(&exec);
+	for (i = 0; i < num_read_bo_handles; i++)
+		drm_gem_object_put(gobj_read[i]);
+	kfree(gobj_read);
+
+	for (i = 0; i < num_write_bo_handles; i++)
+		drm_gem_object_put(gobj_write[i]);
+	kfree(gobj_write);
+
+	kfree(timeline_points);
+	kfree(timeline_handles);
+	kfree(syncobj_handles);
+	kfree(bo_handles_write);
+	kfree(bo_handles_read);
+
+	return 0;
+
+free_fences:
+	while (num_fences-- > 0)
+		dma_fence_put(fences[num_fences]);
+	kfree(fences);
+free_fence_info:
+	kfree(fence_info);
+exec_fini:
+	drm_exec_fini(&exec);
+put_gobj_write:
+	while (wentry-- > 0)
+		drm_gem_object_put(gobj_write[wentry]);
+	kfree(gobj_write);
+put_gobj_read:
+	while (rentry-- > 0)
+		drm_gem_object_put(gobj_read[rentry]);
+	kfree(gobj_read);
+free_timeline_points:
+	kfree(timeline_points);
+free_timeline_handles:
+	kfree(timeline_handles);
+free_syncobj_handles:
+	kfree(syncobj_handles);
+free_bo_handles_write:
+	kfree(bo_handles_write);
+free_bo_handles_read:
+	kfree(bo_handles_read);
+
+	return r;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_userq_fence.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_userq_fence.h
new file mode 100644
index 000000000000..d76add2afc77
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_userq_fence.h
@@ -0,0 +1,77 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __AMDGPU_USERQ_FENCE_H__
+#define __AMDGPU_USERQ_FENCE_H__
+
+#include <linux/types.h>
+
+#include "amdgpu_userq.h"
+
+struct amdgpu_userq_fence {
+	struct dma_fence base;
+	/*
+	 * This lock is necessary to synchronize the
+	 * userqueue dma fence operations.
+	 */
+	spinlock_t lock;
+	struct list_head link;
+	unsigned long fence_drv_array_count;
+	struct amdgpu_userq_fence_driver *fence_drv;
+	struct amdgpu_userq_fence_driver **fence_drv_array;
+};
+
+struct amdgpu_userq_fence_driver {
+	struct kref refcount;
+	u64 va;
+	u64 gpu_addr;
+	u64 *cpu_addr;
+	u64 context;
+	/*
+	 * This lock is necesaary to synchronize the access
+	 * to the fences list by the fence driver.
+	 */
+	spinlock_t fence_list_lock;
+	struct list_head fences;
+	struct amdgpu_device *adev;
+	char timeline_name[TASK_COMM_LEN];
+};
+
+int amdgpu_userq_fence_slab_init(void);
+void amdgpu_userq_fence_slab_fini(void);
+
+void amdgpu_userq_fence_driver_get(struct amdgpu_userq_fence_driver *fence_drv);
+void amdgpu_userq_fence_driver_put(struct amdgpu_userq_fence_driver *fence_drv);
+int amdgpu_userq_fence_driver_alloc(struct amdgpu_device *adev,
+				    struct amdgpu_usermode_queue *userq);
+void amdgpu_userq_fence_driver_free(struct amdgpu_usermode_queue *userq);
+void amdgpu_userq_fence_driver_process(struct amdgpu_userq_fence_driver *fence_drv);
+void amdgpu_userq_fence_driver_force_completion(struct amdgpu_usermode_queue *userq);
+void amdgpu_userq_fence_driver_destroy(struct kref *ref);
+int amdgpu_userq_signal_ioctl(struct drm_device *dev, void *data,
+			      struct drm_file *filp);
+int amdgpu_userq_wait_ioctl(struct drm_device *dev, void *data,
+			    struct drm_file *filp);
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_utils.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_utils.h
new file mode 100644
index 000000000000..1e40ca3b1584
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_utils.h
@@ -0,0 +1,91 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2025 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef AMDGPU_UTILS_H_
+#define AMDGPU_UTILS_H_
+
+/* ---------- Generic 2‑bit capability attribute encoding ----------
+ * 00 INVALID, 01 RO, 10 WO, 11 RW
+ */
+enum amdgpu_cap_attr {
+	AMDGPU_CAP_ATTR_INVALID = 0,
+	AMDGPU_CAP_ATTR_RO      = 1 << 0,
+	AMDGPU_CAP_ATTR_WO      = 1 << 1,
+	AMDGPU_CAP_ATTR_RW      = (AMDGPU_CAP_ATTR_RO | AMDGPU_CAP_ATTR_WO),
+};
+
+#define AMDGPU_CAP_ATTR_BITS 2
+#define AMDGPU_CAP_ATTR_MAX  ((1U << AMDGPU_CAP_ATTR_BITS) - 1)
+
+/* Internal helper to build helpers for a given enum NAME */
+#define DECLARE_ATTR_CAP_CLASS_HELPERS(NAME)							\
+enum { NAME##_BITMAP_BITS = NAME##_COUNT * AMDGPU_CAP_ATTR_BITS };				\
+struct NAME##_caps {										\
+	DECLARE_BITMAP(bmap, NAME##_BITMAP_BITS);						\
+};												\
+static inline unsigned int NAME##_ATTR_START(enum NAME##_cap_id cap)				\
+{ return (unsigned int)cap * AMDGPU_CAP_ATTR_BITS; }						\
+static inline void NAME##_attr_init(struct NAME##_caps *c)					\
+{ if (c) bitmap_zero(c->bmap, NAME##_BITMAP_BITS); }						\
+static inline int NAME##_attr_set(struct NAME##_caps *c,					\
+				  enum NAME##_cap_id cap, enum amdgpu_cap_attr attr)		\
+{												\
+	if (!c)											\
+		return -EINVAL;									\
+	if (cap >= NAME##_COUNT)								\
+		return -EINVAL;									\
+	if ((unsigned int)attr > AMDGPU_CAP_ATTR_MAX)						\
+		return -EINVAL;									\
+	bitmap_write(c->bmap, (unsigned long)attr,						\
+			NAME##_ATTR_START(cap), AMDGPU_CAP_ATTR_BITS);				\
+	return 0;										\
+}												\
+static inline int NAME##_attr_get(const struct NAME##_caps *c,					\
+				  enum NAME##_cap_id cap, enum amdgpu_cap_attr *out)		\
+{												\
+	unsigned long v;									\
+	if (!c || !out)										\
+		return -EINVAL;									\
+	if (cap >= NAME##_COUNT)								\
+		return -EINVAL;									\
+	v = bitmap_read(c->bmap, NAME##_ATTR_START(cap), AMDGPU_CAP_ATTR_BITS);			\
+	*out = (enum amdgpu_cap_attr)v;								\
+	return 0;										\
+}												\
+static inline bool NAME##_cap_is_ro(const struct NAME##_caps *c, enum NAME##_cap_id id)		\
+{ enum amdgpu_cap_attr a; return !NAME##_attr_get(c, id, &a) && a == AMDGPU_CAP_ATTR_RO; }	\
+static inline bool NAME##_cap_is_wo(const struct NAME##_caps *c, enum NAME##_cap_id id)		\
+{ enum amdgpu_cap_attr a; return !NAME##_attr_get(c, id, &a) && a == AMDGPU_CAP_ATTR_WO; }	\
+static inline bool NAME##_cap_is_rw(const struct NAME##_caps *c, enum NAME##_cap_id id)		\
+{ enum amdgpu_cap_attr a; return !NAME##_attr_get(c, id, &a) && a == AMDGPU_CAP_ATTR_RW; }
+
+/* Element expander for enum creation */
+#define _CAP_ENUM_ELEM(x) x,
+
+/* Public macro: declare enum + helpers from an X‑macro list */
+#define DECLARE_ATTR_CAP_CLASS(NAME, LIST_MACRO)						\
+	enum NAME##_cap_id { LIST_MACRO(_CAP_ENUM_ELEM) NAME##_COUNT };				\
+	DECLARE_ATTR_CAP_CLASS_HELPERS(NAME)
+
+#endif /* AMDGPU_UTILS_H_ */
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c
index 6f8de11a17f1..5c38f0d30c87 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c
@@ -37,6 +37,7 @@
 #include "amdgpu.h"
 #include "amdgpu_pm.h"
 #include "amdgpu_uvd.h"
+#include "amdgpu_cs.h"
 #include "cikd.h"
 #include "uvd/uvd_4_2_d.h"
 
@@ -95,16 +96,16 @@
  */
 struct amdgpu_uvd_cs_ctx {
 	struct amdgpu_cs_parser *parser;
-	unsigned reg, count;
-	unsigned data0, data1;
-	unsigned idx;
-	unsigned ib_idx;
+	unsigned int reg, count;
+	unsigned int data0, data1;
+	unsigned int idx;
+	struct amdgpu_ib *ib;
 
 	/* does the IB has a msg command */
 	bool has_msg_cmd;
 
 	/* minimum buffer sizes */
-	unsigned *buf_sizes;
+	unsigned int *buf_sizes;
 };
 
 #ifdef CONFIG_DRM_AMDGPU_SI
@@ -185,7 +186,7 @@ int amdgpu_uvd_sw_init(struct amdgpu_device *adev)
 	unsigned long bo_size;
 	const char *fw_name;
 	const struct common_firmware_header *hdr;
-	unsigned family_id;
+	unsigned int family_id;
 	int i, j, r;
 
 	INIT_DELAYED_WORK(&adev->uvd.idle_work, amdgpu_uvd_idle_work_handler);
@@ -259,19 +260,11 @@ int amdgpu_uvd_sw_init(struct amdgpu_device *adev)
 		return -EINVAL;
 	}
 
-	r = request_firmware(&adev->uvd.fw, fw_name, adev->dev);
-	if (r) {
-		dev_err(adev->dev, "amdgpu_uvd: Can't load firmware \"%s\"\n",
-			fw_name);
-		return r;
-	}
-
-	r = amdgpu_ucode_validate(adev->uvd.fw);
+	r = amdgpu_ucode_request(adev, &adev->uvd.fw, AMDGPU_UCODE_REQUIRED, "%s", fw_name);
 	if (r) {
 		dev_err(adev->dev, "amdgpu_uvd: Can't validate firmware \"%s\"\n",
 			fw_name);
-		release_firmware(adev->uvd.fw);
-		adev->uvd.fw = NULL;
+		amdgpu_ucode_release(&adev->uvd.fw);
 		return r;
 	}
 
@@ -282,7 +275,7 @@ int amdgpu_uvd_sw_init(struct amdgpu_device *adev)
 	family_id = le32_to_cpu(hdr->ucode_version) & 0xff;
 
 	if (adev->asic_type < CHIP_VEGA20) {
-		unsigned version_major, version_minor;
+		unsigned int version_major, version_minor;
 
 		version_major = (le32_to_cpu(hdr->ucode_version) >> 24) & 0xff;
 		version_minor = (le32_to_cpu(hdr->ucode_version) >> 8) & 0xff;
@@ -330,8 +323,11 @@ int amdgpu_uvd_sw_init(struct amdgpu_device *adev)
 		if (adev->uvd.harvest_config & (1 << j))
 			continue;
 		r = amdgpu_bo_create_kernel(adev, bo_size, PAGE_SIZE,
-					    AMDGPU_GEM_DOMAIN_VRAM, &adev->uvd.inst[j].vcpu_bo,
-					    &adev->uvd.inst[j].gpu_addr, &adev->uvd.inst[j].cpu_addr);
+					    AMDGPU_GEM_DOMAIN_VRAM |
+					    AMDGPU_GEM_DOMAIN_GTT,
+					    &adev->uvd.inst[j].vcpu_bo,
+					    &adev->uvd.inst[j].gpu_addr,
+					    &adev->uvd.inst[j].cpu_addr);
 		if (r) {
 			dev_err(adev->dev, "(%d) failed to allocate UVD bo\n", r);
 			return r;
@@ -393,7 +389,7 @@ int amdgpu_uvd_sw_fini(struct amdgpu_device *adev)
 			amdgpu_ring_fini(&adev->uvd.inst[j].ring_enc[i]);
 	}
 	amdgpu_bo_free_kernel(&adev->uvd.ib_bo, NULL, &addr);
-	release_firmware(adev->uvd.fw);
+	amdgpu_ucode_release(&adev->uvd.fw);
 
 	return 0;
 }
@@ -402,32 +398,32 @@ int amdgpu_uvd_sw_fini(struct amdgpu_device *adev)
  * amdgpu_uvd_entity_init - init entity
  *
  * @adev: amdgpu_device pointer
+ * @ring: amdgpu_ring pointer to check
  *
+ * Initialize the entity used for handle management in the kernel driver.
  */
-int amdgpu_uvd_entity_init(struct amdgpu_device *adev)
+int amdgpu_uvd_entity_init(struct amdgpu_device *adev, struct amdgpu_ring *ring)
 {
-	struct amdgpu_ring *ring;
-	struct drm_gpu_scheduler *sched;
-	int r;
+	if (ring == &adev->uvd.inst[0].ring) {
+		struct drm_gpu_scheduler *sched = &ring->sched;
+		int r;
 
-	ring = &adev->uvd.inst[0].ring;
-	sched = &ring->sched;
-	r = drm_sched_entity_init(&adev->uvd.entity, DRM_SCHED_PRIORITY_NORMAL,
-				  &sched, 1, NULL);
-	if (r) {
-		DRM_ERROR("Failed setting up UVD kernel entity.\n");
-		return r;
+		r = drm_sched_entity_init(&adev->uvd.entity, DRM_SCHED_PRIORITY_NORMAL,
+					  &sched, 1, NULL);
+		if (r) {
+			DRM_ERROR("Failed setting up UVD kernel entity.\n");
+			return r;
+		}
 	}
 
 	return 0;
 }
 
-int amdgpu_uvd_suspend(struct amdgpu_device *adev)
+int amdgpu_uvd_prepare_suspend(struct amdgpu_device *adev)
 {
-	unsigned size;
+	unsigned int size;
 	void *ptr;
 	int i, j, idx;
-	bool in_ras_intr = amdgpu_ras_intr_triggered();
 
 	cancel_delayed_work_sync(&adev->uvd.idle_work);
 
@@ -456,7 +452,7 @@ int amdgpu_uvd_suspend(struct amdgpu_device *adev)
 
 		if (drm_dev_enter(adev_to_drm(adev), &idx)) {
 			/* re-write 0 since err_event_athub will corrupt VCPU buffer */
-			if (in_ras_intr)
+			if (amdgpu_ras_intr_triggered())
 				memset(adev->uvd.inst[j].saved_bo, 0, size);
 			else
 				memcpy_fromio(adev->uvd.inst[j].saved_bo, ptr, size);
@@ -465,7 +461,12 @@ int amdgpu_uvd_suspend(struct amdgpu_device *adev)
 		}
 	}
 
-	if (in_ras_intr)
+	return 0;
+}
+
+int amdgpu_uvd_suspend(struct amdgpu_device *adev)
+{
+	if (amdgpu_ras_intr_triggered())
 		DRM_WARN("UVD VCPU state may lost due to RAS ERREVENT_ATHUB_INTERRUPT\n");
 
 	return 0;
@@ -473,7 +474,7 @@ int amdgpu_uvd_suspend(struct amdgpu_device *adev)
 
 int amdgpu_uvd_resume(struct amdgpu_device *adev)
 {
-	unsigned size;
+	unsigned int size;
 	void *ptr;
 	int i, idx;
 
@@ -495,7 +496,7 @@ int amdgpu_uvd_resume(struct amdgpu_device *adev)
 			adev->uvd.inst[i].saved_bo = NULL;
 		} else {
 			const struct common_firmware_header *hdr;
-			unsigned offset;
+			unsigned int offset;
 
 			hdr = (const struct common_firmware_header *)adev->uvd.fw->data;
 			if (adev->firmware.load_type != AMDGPU_FW_LOAD_PSP) {
@@ -546,9 +547,12 @@ void amdgpu_uvd_free_handles(struct amdgpu_device *adev, struct drm_file *filp)
 static void amdgpu_uvd_force_into_uvd_segment(struct amdgpu_bo *abo)
 {
 	int i;
+
 	for (i = 0; i < abo->placement.num_placement; ++i) {
 		abo->placements[i].fpfn = 0 >> PAGE_SHIFT;
 		abo->placements[i].lpfn = (256 * 1024 * 1024) >> PAGE_SHIFT;
+		if (abo->placements[i].mem_type == TTM_PL_VRAM)
+			abo->placements[i].flags |= TTM_PL_FLAG_CONTIGUOUS;
 	}
 }
 
@@ -557,8 +561,8 @@ static u64 amdgpu_uvd_get_addr_from_ctx(struct amdgpu_uvd_cs_ctx *ctx)
 	uint32_t lo, hi;
 	uint64_t addr;
 
-	lo = amdgpu_get_ib_value(ctx->parser, ctx->ib_idx, ctx->data0);
-	hi = amdgpu_get_ib_value(ctx->parser, ctx->ib_idx, ctx->data1);
+	lo = amdgpu_ib_get_value(ctx->ib, ctx->data0);
+	hi = amdgpu_ib_get_value(ctx->ib, ctx->data1);
 	addr = ((uint64_t)lo) | (((uint64_t)hi) << 32);
 
 	return addr;
@@ -583,16 +587,17 @@ static int amdgpu_uvd_cs_pass1(struct amdgpu_uvd_cs_ctx *ctx)
 
 	r = amdgpu_cs_find_mapping(ctx->parser, addr, &bo, &mapping);
 	if (r) {
-		DRM_ERROR("Can't find BO for addr 0x%08Lx\n", addr);
+		DRM_ERROR("Can't find BO for addr 0x%08llx\n", addr);
 		return r;
 	}
 
 	if (!ctx->parser->adev->uvd.address_64_bit) {
 		/* check if it's a message or feedback command */
-		cmd = amdgpu_get_ib_value(ctx->parser, ctx->ib_idx, ctx->idx) >> 1;
+		cmd = amdgpu_ib_get_value(ctx->ib, ctx->idx) >> 1;
 		if (cmd == 0x0 || cmd == 0x3) {
 			/* yes, force it into VRAM */
 			uint32_t domain = AMDGPU_GEM_DOMAIN_VRAM;
+
 			amdgpu_bo_placement_from_domain(bo, domain);
 		}
 		amdgpu_uvd_force_into_uvd_segment(bo);
@@ -613,21 +618,21 @@ static int amdgpu_uvd_cs_pass1(struct amdgpu_uvd_cs_ctx *ctx)
  * Peek into the decode message and calculate the necessary buffer sizes.
  */
 static int amdgpu_uvd_cs_msg_decode(struct amdgpu_device *adev, uint32_t *msg,
-	unsigned buf_sizes[])
+	unsigned int buf_sizes[])
 {
-	unsigned stream_type = msg[4];
-	unsigned width = msg[6];
-	unsigned height = msg[7];
-	unsigned dpb_size = msg[9];
-	unsigned pitch = msg[28];
-	unsigned level = msg[57];
+	unsigned int stream_type = msg[4];
+	unsigned int width = msg[6];
+	unsigned int height = msg[7];
+	unsigned int dpb_size = msg[9];
+	unsigned int pitch = msg[28];
+	unsigned int level = msg[57];
 
-	unsigned width_in_mb = width / 16;
-	unsigned height_in_mb = ALIGN(height / 16, 2);
-	unsigned fs_in_mb = width_in_mb * height_in_mb;
+	unsigned int width_in_mb = width / 16;
+	unsigned int height_in_mb = ALIGN(height / 16, 2);
+	unsigned int fs_in_mb = width_in_mb * height_in_mb;
 
-	unsigned image_size, tmp, min_dpb_size, num_dpb_buffer;
-	unsigned min_ctx_size = ~0;
+	unsigned int image_size, tmp, min_dpb_size, num_dpb_buffer;
+	unsigned int min_ctx_size = ~0;
 
 	image_size = width * height;
 	image_size += image_size / 2;
@@ -635,7 +640,7 @@ static int amdgpu_uvd_cs_msg_decode(struct amdgpu_device *adev, uint32_t *msg,
 
 	switch (stream_type) {
 	case 0: /* H264 */
-		switch(level) {
+		switch (level) {
 		case 30:
 			num_dpb_buffer = 8100 / fs_in_mb;
 			break;
@@ -713,7 +718,7 @@ static int amdgpu_uvd_cs_msg_decode(struct amdgpu_device *adev, uint32_t *msg,
 		break;
 
 	case 7: /* H264 Perf */
-		switch(level) {
+		switch (level) {
 		case 30:
 			num_dpb_buffer = 8100 / fs_in_mb;
 			break;
@@ -746,7 +751,7 @@ static int amdgpu_uvd_cs_msg_decode(struct amdgpu_device *adev, uint32_t *msg,
 		/* reference picture buffer */
 		min_dpb_size = image_size * num_dpb_buffer;
 
-		if (!adev->uvd.use_ctx_buf){
+		if (!adev->uvd.use_ctx_buf) {
 			/* macroblock context buffer */
 			min_dpb_size +=
 				width_in_mb * height_in_mb * num_dpb_buffer * 192;
@@ -809,7 +814,7 @@ static int amdgpu_uvd_cs_msg_decode(struct amdgpu_device *adev, uint32_t *msg,
  * Make sure that we don't open up to many sessions.
  */
 static int amdgpu_uvd_cs_msg(struct amdgpu_uvd_cs_ctx *ctx,
-			     struct amdgpu_bo *bo, unsigned offset)
+			     struct amdgpu_bo *bo, unsigned int offset)
 {
 	struct amdgpu_device *adev = ctx->parser->adev;
 	int32_t *msg, msg_type, handle;
@@ -834,6 +839,7 @@ static int amdgpu_uvd_cs_msg(struct amdgpu_uvd_cs_ctx *ctx,
 	handle = msg[2];
 
 	if (handle == 0) {
+		amdgpu_bo_kunmap(bo);
 		DRM_ERROR("Invalid UVD handle!\n");
 		return -EINVAL;
 	}
@@ -892,6 +898,7 @@ static int amdgpu_uvd_cs_msg(struct amdgpu_uvd_cs_ctx *ctx,
 		DRM_ERROR("Illegal UVD message type (%d)!\n", msg_type);
 	}
 
+	amdgpu_bo_kunmap(bo);
 	return -EINVAL;
 }
 
@@ -913,7 +920,7 @@ static int amdgpu_uvd_cs_pass2(struct amdgpu_uvd_cs_ctx *ctx)
 
 	r = amdgpu_cs_find_mapping(ctx->parser, addr, &bo, &mapping);
 	if (r) {
-		DRM_ERROR("Can't find BO for addr 0x%08Lx\n", addr);
+		DRM_ERROR("Can't find BO for addr 0x%08llx\n", addr);
 		return r;
 	}
 
@@ -925,16 +932,14 @@ static int amdgpu_uvd_cs_pass2(struct amdgpu_uvd_cs_ctx *ctx)
 	addr -= mapping->start * AMDGPU_GPU_PAGE_SIZE;
 	start += addr;
 
-	amdgpu_set_ib_value(ctx->parser, ctx->ib_idx, ctx->data0,
-			    lower_32_bits(start));
-	amdgpu_set_ib_value(ctx->parser, ctx->ib_idx, ctx->data1,
-			    upper_32_bits(start));
+	amdgpu_ib_set_value(ctx->ib, ctx->data0, lower_32_bits(start));
+	amdgpu_ib_set_value(ctx->ib, ctx->data1, upper_32_bits(start));
 
-	cmd = amdgpu_get_ib_value(ctx->parser, ctx->ib_idx, ctx->idx) >> 1;
+	cmd = amdgpu_ib_get_value(ctx->ib, ctx->idx) >> 1;
 	if (cmd < 0x4) {
 		if ((end - start) < ctx->buf_sizes[cmd]) {
 			DRM_ERROR("buffer (%d) to small (%d / %d)!\n", cmd,
-				  (unsigned)(end - start),
+				  (unsigned int)(end - start),
 				  ctx->buf_sizes[cmd]);
 			return -EINVAL;
 		}
@@ -942,7 +947,7 @@ static int amdgpu_uvd_cs_pass2(struct amdgpu_uvd_cs_ctx *ctx)
 	} else if (cmd == 0x206) {
 		if ((end - start) < ctx->buf_sizes[4]) {
 			DRM_ERROR("buffer (%d) to small (%d / %d)!\n", cmd,
-					  (unsigned)(end - start),
+					  (unsigned int)(end - start),
 					  ctx->buf_sizes[4]);
 			return -EINVAL;
 		}
@@ -953,14 +958,14 @@ static int amdgpu_uvd_cs_pass2(struct amdgpu_uvd_cs_ctx *ctx)
 
 	if (!ctx->parser->adev->uvd.address_64_bit) {
 		if ((start >> 28) != ((end - 1) >> 28)) {
-			DRM_ERROR("reloc %LX-%LX crossing 256MB boundary!\n",
+			DRM_ERROR("reloc %llx-%llx crossing 256MB boundary!\n",
 				  start, end);
 			return -EINVAL;
 		}
 
 		if ((cmd == 0 || cmd == 0x3) &&
 		    (start >> 28) != (ctx->parser->adev->uvd.inst->gpu_addr >> 28)) {
-			DRM_ERROR("msg/fb buffer %LX-%LX out of 256MB segment!\n",
+			DRM_ERROR("msg/fb buffer %llx-%llx out of 256MB segment!\n",
 				  start, end);
 			return -EINVAL;
 		}
@@ -990,14 +995,13 @@ static int amdgpu_uvd_cs_pass2(struct amdgpu_uvd_cs_ctx *ctx)
 static int amdgpu_uvd_cs_reg(struct amdgpu_uvd_cs_ctx *ctx,
 			     int (*cb)(struct amdgpu_uvd_cs_ctx *ctx))
 {
-	struct amdgpu_ib *ib = &ctx->parser->job->ibs[ctx->ib_idx];
 	int i, r;
 
 	ctx->idx++;
 	for (i = 0; i <= ctx->count; ++i) {
-		unsigned reg = ctx->reg + i;
+		unsigned int reg = ctx->reg + i;
 
-		if (ctx->idx >= ib->length_dw) {
+		if (ctx->idx >= ctx->ib->length_dw) {
 			DRM_ERROR("Register command after end of CS!\n");
 			return -EINVAL;
 		}
@@ -1037,12 +1041,12 @@ static int amdgpu_uvd_cs_reg(struct amdgpu_uvd_cs_ctx *ctx,
 static int amdgpu_uvd_cs_packets(struct amdgpu_uvd_cs_ctx *ctx,
 				 int (*cb)(struct amdgpu_uvd_cs_ctx *ctx))
 {
-	struct amdgpu_ib *ib = &ctx->parser->job->ibs[ctx->ib_idx];
 	int r;
 
-	for (ctx->idx = 0 ; ctx->idx < ib->length_dw; ) {
-		uint32_t cmd = amdgpu_get_ib_value(ctx->parser, ctx->ib_idx, ctx->idx);
-		unsigned type = CP_PACKET_GET_TYPE(cmd);
+	for (ctx->idx = 0 ; ctx->idx < ctx->ib->length_dw; ) {
+		uint32_t cmd = amdgpu_ib_get_value(ctx->ib, ctx->idx);
+		unsigned int type = CP_PACKET_GET_TYPE(cmd);
+
 		switch (type) {
 		case PACKET_TYPE0:
 			ctx->reg = CP_PACKET0_GET_REG(cmd);
@@ -1066,25 +1070,26 @@ static int amdgpu_uvd_cs_packets(struct amdgpu_uvd_cs_ctx *ctx,
  * amdgpu_uvd_ring_parse_cs - UVD command submission parser
  *
  * @parser: Command submission parser context
- * @ib_idx: Which indirect buffer to use
+ * @job: the job to parse
+ * @ib: the IB to patch
  *
  * Parse the command stream, patch in addresses as necessary.
  */
-int amdgpu_uvd_ring_parse_cs(struct amdgpu_cs_parser *parser, uint32_t ib_idx)
+int amdgpu_uvd_ring_parse_cs(struct amdgpu_cs_parser *parser,
+			     struct amdgpu_job *job,
+			     struct amdgpu_ib *ib)
 {
 	struct amdgpu_uvd_cs_ctx ctx = {};
-	unsigned buf_sizes[] = {
+	unsigned int buf_sizes[] = {
 		[0x00000000]	=	2048,
 		[0x00000001]	=	0xFFFFFFFF,
 		[0x00000002]	=	0xFFFFFFFF,
 		[0x00000003]	=	2048,
 		[0x00000004]	=	0xFFFFFFFF,
 	};
-	struct amdgpu_ib *ib = &parser->job->ibs[ib_idx];
 	int r;
 
-	parser->job->vm = NULL;
-	ib->gpu_addr = amdgpu_sa_bo_gpu_addr(ib->sa_bo);
+	job->vm = NULL;
 
 	if (ib->length_dw % 16) {
 		DRM_ERROR("UVD IB length (%d) not 16 dwords aligned!\n",
@@ -1094,7 +1099,7 @@ int amdgpu_uvd_ring_parse_cs(struct amdgpu_cs_parser *parser, uint32_t ib_idx)
 
 	ctx.parser = parser;
 	ctx.buf_sizes = buf_sizes;
-	ctx.ib_idx = ib_idx;
+	ctx.ib = ib;
 
 	/* first round only required on chips without UVD 64 bit address support */
 	if (!parser->adev->uvd.address_64_bit) {
@@ -1122,30 +1127,29 @@ static int amdgpu_uvd_send_msg(struct amdgpu_ring *ring, struct amdgpu_bo *bo,
 {
 	struct amdgpu_device *adev = ring->adev;
 	struct dma_fence *f = NULL;
+	uint32_t offset, data[4];
 	struct amdgpu_job *job;
 	struct amdgpu_ib *ib;
-	uint32_t data[4];
 	uint64_t addr;
-	long r;
-	int i;
-	unsigned offset_idx = 0;
-	unsigned offset[3] = { UVD_BASE_SI, 0, 0 };
+	int i, r;
 
-	r = amdgpu_job_alloc_with_ib(adev, 64, direct ? AMDGPU_IB_POOL_DIRECT :
-				     AMDGPU_IB_POOL_DELAYED, &job);
+	r = amdgpu_job_alloc_with_ib(ring->adev, &adev->uvd.entity,
+				     AMDGPU_FENCE_OWNER_UNDEFINED,
+				     64, direct ? AMDGPU_IB_POOL_DIRECT :
+				     AMDGPU_IB_POOL_DELAYED, &job,
+				     AMDGPU_KERNEL_JOB_ID_VCN_RING_TEST);
 	if (r)
 		return r;
 
-	if (adev->asic_type >= CHIP_VEGA10) {
-		offset_idx = 1 + ring->me;
-		offset[1] = adev->reg_offset[UVD_HWIP][0][1];
-		offset[2] = adev->reg_offset[UVD_HWIP][1][1];
-	}
+	if (adev->asic_type >= CHIP_VEGA10)
+		offset = adev->reg_offset[UVD_HWIP][ring->me][1];
+	else
+		offset = UVD_BASE_SI;
 
-	data[0] = PACKET0(offset[offset_idx] + UVD_GPCOM_VCPU_DATA0, 0);
-	data[1] = PACKET0(offset[offset_idx] + UVD_GPCOM_VCPU_DATA1, 0);
-	data[2] = PACKET0(offset[offset_idx] + UVD_GPCOM_VCPU_CMD, 0);
-	data[3] = PACKET0(offset[offset_idx] + UVD_NO_OP, 0);
+	data[0] = PACKET0(offset + UVD_GPCOM_VCPU_DATA0, 0);
+	data[1] = PACKET0(offset + UVD_GPCOM_VCPU_DATA1, 0);
+	data[2] = PACKET0(offset + UVD_GPCOM_VCPU_CMD, 0);
+	data[3] = PACKET0(offset + UVD_NO_OP, 0);
 
 	ib = &job->ibs[0];
 	addr = amdgpu_bo_gpu_offset(bo);
@@ -1162,27 +1166,17 @@ static int amdgpu_uvd_send_msg(struct amdgpu_ring *ring, struct amdgpu_bo *bo,
 	ib->length_dw = 16;
 
 	if (direct) {
-		r = dma_resv_wait_timeout(bo->tbo.base.resv, true, false,
-					  msecs_to_jiffies(10));
-		if (r == 0)
-			r = -ETIMEDOUT;
-		if (r < 0)
-			goto err_free;
-
 		r = amdgpu_job_submit_direct(job, ring, &f);
 		if (r)
 			goto err_free;
 	} else {
-		r = amdgpu_sync_resv(adev, &job->sync, bo->tbo.base.resv,
-				     AMDGPU_SYNC_ALWAYS,
-				     AMDGPU_FENCE_OWNER_UNDEFINED);
+		r = drm_sched_job_add_resv_dependencies(&job->base,
+							bo->tbo.base.resv,
+							DMA_RESV_USAGE_KERNEL);
 		if (r)
 			goto err_free;
 
-		r = amdgpu_job_submit(job, &adev->uvd.entity,
-				      AMDGPU_FENCE_OWNER_UNDEFINED, &f);
-		if (r)
-			goto err_free;
+		f = amdgpu_job_submit(job);
 	}
 
 	amdgpu_bo_reserve(bo, true);
@@ -1201,8 +1195,9 @@ err_free:
 }
 
 /* multiple fence commands without any stream commands in between can
-   crash the vcpu so just try to emmit a dummy create/destroy msg to
-   avoid this */
+ * crash the vcpu so just try to emmit a dummy create/destroy msg to
+ * avoid this
+ */
 int amdgpu_uvd_get_create_msg(struct amdgpu_ring *ring, uint32_t handle,
 			      struct dma_fence **fence)
 {
@@ -1268,15 +1263,14 @@ static void amdgpu_uvd_idle_work_handler(struct work_struct *work)
 {
 	struct amdgpu_device *adev =
 		container_of(work, struct amdgpu_device, uvd.idle_work.work);
-	unsigned fences = 0, i, j;
+	unsigned int fences = 0, i, j;
 
 	for (i = 0; i < adev->uvd.num_uvd_inst; ++i) {
 		if (adev->uvd.harvest_config & (1 << i))
 			continue;
 		fences += amdgpu_fence_count_emitted(&adev->uvd.inst[i].ring);
-		for (j = 0; j < adev->uvd.num_enc_rings; ++j) {
+		for (j = 0; j < adev->uvd.num_enc_rings; ++j)
 			fences += amdgpu_fence_count_emitted(&adev->uvd.inst[i].ring_enc[j]);
-		}
 	}
 
 	if (fences == 0) {
@@ -1372,7 +1366,7 @@ error:
  */
 uint32_t amdgpu_uvd_used_handles(struct amdgpu_device *adev)
 {
-	unsigned i;
+	unsigned int i;
 	uint32_t used_handles = 0;
 
 	for (i = 0; i < adev->uvd.max_handles; ++i) {
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.h
index 76ac9699885d..9dfad2f48ef4 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.h
@@ -73,7 +73,8 @@ struct amdgpu_uvd {
 
 int amdgpu_uvd_sw_init(struct amdgpu_device *adev);
 int amdgpu_uvd_sw_fini(struct amdgpu_device *adev);
-int amdgpu_uvd_entity_init(struct amdgpu_device *adev);
+int amdgpu_uvd_entity_init(struct amdgpu_device *adev, struct amdgpu_ring *ring);
+int amdgpu_uvd_prepare_suspend(struct amdgpu_device *adev);
 int amdgpu_uvd_suspend(struct amdgpu_device *adev);
 int amdgpu_uvd_resume(struct amdgpu_device *adev);
 int amdgpu_uvd_get_create_msg(struct amdgpu_ring *ring, uint32_t handle,
@@ -82,7 +83,9 @@ int amdgpu_uvd_get_destroy_msg(struct amdgpu_ring *ring, uint32_t handle,
 			       bool direct, struct dma_fence **fence);
 void amdgpu_uvd_free_handles(struct amdgpu_device *adev,
 			     struct drm_file *filp);
-int amdgpu_uvd_ring_parse_cs(struct amdgpu_cs_parser *parser, uint32_t ib_idx);
+int amdgpu_uvd_ring_parse_cs(struct amdgpu_cs_parser *parser,
+			     struct amdgpu_job *job,
+			     struct amdgpu_ib *ib);
 void amdgpu_uvd_ring_begin_use(struct amdgpu_ring *ring);
 void amdgpu_uvd_ring_end_use(struct amdgpu_ring *ring);
 int amdgpu_uvd_ring_test_ib(struct amdgpu_ring *ring, long timeout);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c
index 688bef1649b5..a7d8f1ce6ac2 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c
@@ -34,12 +34,16 @@
 #include "amdgpu.h"
 #include "amdgpu_pm.h"
 #include "amdgpu_vce.h"
+#include "amdgpu_cs.h"
 #include "cikd.h"
 
 /* 1 second timeout */
 #define VCE_IDLE_TIMEOUT	msecs_to_jiffies(1000)
 
 /* Firmware Names */
+#ifdef CONFIG_DRM_AMDGPU_SI
+#define FIRMWARE_VCE_V1_0	"amdgpu/vce_1_0_0.bin"
+#endif
 #ifdef CONFIG_DRM_AMDGPU_CIK
 #define FIRMWARE_BONAIRE	"amdgpu/bonaire_vce.bin"
 #define FIRMWARE_KABINI	"amdgpu/kabini_vce.bin"
@@ -60,6 +64,9 @@
 #define FIRMWARE_VEGA12		"amdgpu/vega12_vce.bin"
 #define FIRMWARE_VEGA20		"amdgpu/vega20_vce.bin"
 
+#ifdef CONFIG_DRM_AMDGPU_SI
+MODULE_FIRMWARE(FIRMWARE_VCE_V1_0);
+#endif
 #ifdef CONFIG_DRM_AMDGPU_CIK
 MODULE_FIRMWARE(FIRMWARE_BONAIRE);
 MODULE_FIRMWARE(FIRMWARE_KABINI);
@@ -87,90 +94,93 @@ static int amdgpu_vce_get_destroy_msg(struct amdgpu_ring *ring, uint32_t handle,
 				      bool direct, struct dma_fence **fence);
 
 /**
- * amdgpu_vce_sw_init - allocate memory, load vce firmware
+ * amdgpu_vce_firmware_name() - determine the firmware file name for VCE
  *
  * @adev: amdgpu_device pointer
- * @size: size for the new BO
  *
- * First step to get VCE online, allocate memory and load the firmware
+ * Each chip that has VCE IP may need a different firmware.
+ * This function returns the name of the VCE firmware file
+ * appropriate for the current chip.
  */
-int amdgpu_vce_sw_init(struct amdgpu_device *adev, unsigned long size)
+static const char *amdgpu_vce_firmware_name(struct amdgpu_device *adev)
 {
-	const char *fw_name;
-	const struct common_firmware_header *hdr;
-	unsigned ucode_version, version_major, version_minor, binary_id;
-	int i, r;
-
 	switch (adev->asic_type) {
+#ifdef CONFIG_DRM_AMDGPU_SI
+	case CHIP_PITCAIRN:
+	case CHIP_TAHITI:
+	case CHIP_VERDE:
+		return FIRMWARE_VCE_V1_0;
+#endif
 #ifdef CONFIG_DRM_AMDGPU_CIK
 	case CHIP_BONAIRE:
-		fw_name = FIRMWARE_BONAIRE;
-		break;
+		return FIRMWARE_BONAIRE;
 	case CHIP_KAVERI:
-		fw_name = FIRMWARE_KAVERI;
-		break;
+		return FIRMWARE_KAVERI;
 	case CHIP_KABINI:
-		fw_name = FIRMWARE_KABINI;
-		break;
+		return FIRMWARE_KABINI;
 	case CHIP_HAWAII:
-		fw_name = FIRMWARE_HAWAII;
-		break;
+		return FIRMWARE_HAWAII;
 	case CHIP_MULLINS:
-		fw_name = FIRMWARE_MULLINS;
-		break;
+		return FIRMWARE_MULLINS;
 #endif
 	case CHIP_TONGA:
-		fw_name = FIRMWARE_TONGA;
-		break;
+		return  FIRMWARE_TONGA;
 	case CHIP_CARRIZO:
-		fw_name = FIRMWARE_CARRIZO;
-		break;
+		return  FIRMWARE_CARRIZO;
 	case CHIP_FIJI:
-		fw_name = FIRMWARE_FIJI;
-		break;
+		return  FIRMWARE_FIJI;
 	case CHIP_STONEY:
-		fw_name = FIRMWARE_STONEY;
-		break;
+		return  FIRMWARE_STONEY;
 	case CHIP_POLARIS10:
-		fw_name = FIRMWARE_POLARIS10;
-		break;
+		return  FIRMWARE_POLARIS10;
 	case CHIP_POLARIS11:
-		fw_name = FIRMWARE_POLARIS11;
-		break;
+		return  FIRMWARE_POLARIS11;
 	case CHIP_POLARIS12:
-		fw_name = FIRMWARE_POLARIS12;
-		break;
+		return  FIRMWARE_POLARIS12;
 	case CHIP_VEGAM:
-		fw_name = FIRMWARE_VEGAM;
-		break;
+		return  FIRMWARE_VEGAM;
 	case CHIP_VEGA10:
-		fw_name = FIRMWARE_VEGA10;
-		break;
+		return  FIRMWARE_VEGA10;
 	case CHIP_VEGA12:
-		fw_name = FIRMWARE_VEGA12;
-		break;
+		return  FIRMWARE_VEGA12;
 	case CHIP_VEGA20:
-		fw_name = FIRMWARE_VEGA20;
-		break;
+		return  FIRMWARE_VEGA20;
 
 	default:
-		return -EINVAL;
+		return NULL;
 	}
+}
 
-	r = request_firmware(&adev->vce.fw, fw_name, adev->dev);
-	if (r) {
-		dev_err(adev->dev, "amdgpu_vce: Can't load firmware \"%s\"\n",
-			fw_name);
-		return r;
-	}
+/**
+ * amdgpu_vce_early_init() - try to load VCE firmware
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Tries to load the VCE firmware.
+ *
+ * When not found, returns ENOENT so that the driver can
+ * still load and initialize the rest of the IP blocks.
+ * The GPU can function just fine without VCE, they will just
+ * not support video encoding.
+ */
+int amdgpu_vce_early_init(struct amdgpu_device *adev)
+{
+	const char *fw_name = amdgpu_vce_firmware_name(adev);
+	const struct common_firmware_header *hdr;
+	unsigned int ucode_version, version_major, version_minor, binary_id;
+	int r;
 
-	r = amdgpu_ucode_validate(adev->vce.fw);
+	if (!fw_name)
+		return -ENOENT;
+
+	r = amdgpu_ucode_request(adev, &adev->vce.fw, AMDGPU_UCODE_REQUIRED, "%s", fw_name);
 	if (r) {
-		dev_err(adev->dev, "amdgpu_vce: Can't validate firmware \"%s\"\n",
-			fw_name);
-		release_firmware(adev->vce.fw);
-		adev->vce.fw = NULL;
-		return r;
+		dev_err(adev->dev,
+			"amdgpu_vce: Firmware \"%s\" not found or failed to validate (%d)\n",
+			fw_name, r);
+
+		amdgpu_ucode_release(&adev->vce.fw);
+		return -ENOENT;
 	}
 
 	hdr = (const struct common_firmware_header *)adev->vce.fw->data;
@@ -179,13 +189,39 @@ int amdgpu_vce_sw_init(struct amdgpu_device *adev, unsigned long size)
 	version_major = (ucode_version >> 20) & 0xfff;
 	version_minor = (ucode_version >> 8) & 0xfff;
 	binary_id = ucode_version & 0xff;
-	DRM_INFO("Found VCE firmware Version: %d.%d Binary ID: %d\n",
+	dev_info(adev->dev, "Found VCE firmware Version: %d.%d Binary ID: %d\n",
 		version_major, version_minor, binary_id);
 	adev->vce.fw_version = ((version_major << 24) | (version_minor << 16) |
 				(binary_id << 8));
 
+	return 0;
+}
+
+/**
+ * amdgpu_vce_sw_init() - allocate memory for VCE BO
+ *
+ * @adev: amdgpu_device pointer
+ * @size: size for the new BO
+ *
+ * First step to get VCE online: allocate memory for VCE BO.
+ * The VCE firmware binary is copied into the VCE BO later,
+ * in amdgpu_vce_resume. The VCE executes its code from the
+ * VCE BO and also uses the space in this BO for its stack and data.
+ *
+ * Ideally this BO should be placed in VRAM for optimal performance,
+ * although technically it also runs from system RAM (albeit slowly).
+ */
+int amdgpu_vce_sw_init(struct amdgpu_device *adev, unsigned long size)
+{
+	int i, r;
+
+	if (!adev->vce.fw)
+		return -ENOENT;
+
 	r = amdgpu_bo_create_kernel(adev, size, PAGE_SIZE,
-				    AMDGPU_GEM_DOMAIN_VRAM, &adev->vce.vcpu_bo,
+				    AMDGPU_GEM_DOMAIN_VRAM |
+				    AMDGPU_GEM_DOMAIN_GTT,
+				    &adev->vce.vcpu_bo,
 				    &adev->vce.gpu_addr, &adev->vce.cpu_addr);
 	if (r) {
 		dev_err(adev->dev, "(%d) failed to allocate VCE bo\n", r);
@@ -212,22 +248,22 @@ int amdgpu_vce_sw_init(struct amdgpu_device *adev, unsigned long size)
  */
 int amdgpu_vce_sw_fini(struct amdgpu_device *adev)
 {
-	unsigned i;
+	unsigned int i;
 
 	if (adev->vce.vcpu_bo == NULL)
 		return 0;
 
 	drm_sched_entity_destroy(&adev->vce.entity);
 
-	amdgpu_bo_free_kernel(&adev->vce.vcpu_bo, &adev->vce.gpu_addr,
-		(void **)&adev->vce.cpu_addr);
-
 	for (i = 0; i < adev->vce.num_rings; i++)
 		amdgpu_ring_fini(&adev->vce.ring[i]);
 
-	release_firmware(adev->vce.fw);
+	amdgpu_ucode_release(&adev->vce.fw);
 	mutex_destroy(&adev->vce.idle_mutex);
 
+	amdgpu_bo_free_kernel(&adev->vce.vcpu_bo, &adev->vce.gpu_addr,
+		(void **)&adev->vce.cpu_addr);
+
 	return 0;
 }
 
@@ -235,21 +271,22 @@ int amdgpu_vce_sw_fini(struct amdgpu_device *adev)
  * amdgpu_vce_entity_init - init entity
  *
  * @adev: amdgpu_device pointer
+ * @ring: amdgpu_ring pointer to check
  *
+ * Initialize the entity used for handle management in the kernel driver.
  */
-int amdgpu_vce_entity_init(struct amdgpu_device *adev)
+int amdgpu_vce_entity_init(struct amdgpu_device *adev, struct amdgpu_ring *ring)
 {
-	struct amdgpu_ring *ring;
-	struct drm_gpu_scheduler *sched;
-	int r;
-
-	ring = &adev->vce.ring[0];
-	sched = &ring->sched;
-	r = drm_sched_entity_init(&adev->vce.entity, DRM_SCHED_PRIORITY_NORMAL,
-				  &sched, 1, NULL);
-	if (r != 0) {
-		DRM_ERROR("Failed setting up VCE run queue.\n");
-		return r;
+	if (ring == &adev->vce.ring[0]) {
+		struct drm_gpu_scheduler *sched = &ring->sched;
+		int r;
+
+		r = drm_sched_entity_init(&adev->vce.entity, DRM_SCHED_PRIORITY_NORMAL,
+					  &sched, 1, NULL);
+		if (r != 0) {
+			DRM_ERROR("Failed setting up VCE run queue.\n");
+			return r;
+		}
 	}
 
 	return 0;
@@ -289,40 +326,23 @@ int amdgpu_vce_suspend(struct amdgpu_device *adev)
  */
 int amdgpu_vce_resume(struct amdgpu_device *adev)
 {
-	void *cpu_addr;
 	const struct common_firmware_header *hdr;
-	unsigned offset;
-	int r, idx;
+	unsigned int offset;
+	int idx;
 
 	if (adev->vce.vcpu_bo == NULL)
 		return -EINVAL;
 
-	r = amdgpu_bo_reserve(adev->vce.vcpu_bo, false);
-	if (r) {
-		dev_err(adev->dev, "(%d) failed to reserve VCE bo\n", r);
-		return r;
-	}
-
-	r = amdgpu_bo_kmap(adev->vce.vcpu_bo, &cpu_addr);
-	if (r) {
-		amdgpu_bo_unreserve(adev->vce.vcpu_bo);
-		dev_err(adev->dev, "(%d) VCE map failed\n", r);
-		return r;
-	}
-
 	hdr = (const struct common_firmware_header *)adev->vce.fw->data;
 	offset = le32_to_cpu(hdr->ucode_array_offset_bytes);
 
 	if (drm_dev_enter(adev_to_drm(adev), &idx)) {
-		memcpy_toio(cpu_addr, adev->vce.fw->data + offset,
+		memset_io(adev->vce.cpu_addr, 0, amdgpu_bo_size(adev->vce.vcpu_bo));
+		memcpy_toio(adev->vce.cpu_addr, adev->vce.fw->data + offset,
 			    adev->vce.fw->size - offset);
 		drm_dev_exit(idx);
 	}
 
-	amdgpu_bo_kunmap(adev->vce.vcpu_bo);
-
-	amdgpu_bo_unreserve(adev->vce.vcpu_bo);
-
 	return 0;
 }
 
@@ -337,7 +357,7 @@ static void amdgpu_vce_idle_work_handler(struct work_struct *work)
 {
 	struct amdgpu_device *adev =
 		container_of(work, struct amdgpu_device, vce.idle_work.work);
-	unsigned i, count = 0;
+	unsigned int i, count = 0;
 
 	for (i = 0; i < adev->vce.num_rings; i++)
 		count += amdgpu_fence_count_emitted(&adev->vce.ring[i]);
@@ -414,6 +434,7 @@ void amdgpu_vce_free_handles(struct amdgpu_device *adev, struct drm_file *filp)
 {
 	struct amdgpu_ring *ring = &adev->vce.ring[0];
 	int i, r;
+
 	for (i = 0; i < AMDGPU_MAX_VCE_HANDLES; ++i) {
 		uint32_t handle = atomic_read(&adev->vce.handles[i]);
 
@@ -430,11 +451,28 @@ void amdgpu_vce_free_handles(struct amdgpu_device *adev, struct drm_file *filp)
 }
 
 /**
+ * amdgpu_vce_required_gart_pages() - gets number of GART pages required by VCE
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Returns how many GART pages we need before GTT for the VCE IP block.
+ * For VCE1, see vce_v1_0_ensure_vcpu_bo_32bit_addr for details.
+ * For VCE2+, this is not needed so return zero.
+ */
+u32 amdgpu_vce_required_gart_pages(struct amdgpu_device *adev)
+{
+	/* VCE IP block not added yet, so can't use amdgpu_ip_version */
+	if (adev->family == AMDGPU_FAMILY_SI)
+		return 512;
+
+	return 0;
+}
+
+/**
  * amdgpu_vce_get_create_msg - generate a VCE create msg
  *
  * @ring: ring we should submit the msg to
  * @handle: VCE session handle to use
- * @bo: amdgpu object for which we query the offset
  * @fence: optional fence to return
  *
  * Open up a stream for HW test
@@ -442,7 +480,7 @@ void amdgpu_vce_free_handles(struct amdgpu_device *adev, struct drm_file *filp)
 static int amdgpu_vce_get_create_msg(struct amdgpu_ring *ring, uint32_t handle,
 				     struct dma_fence **fence)
 {
-	const unsigned ib_size_dw = 1024;
+	const unsigned int ib_size_dw = 1024;
 	struct amdgpu_job *job;
 	struct amdgpu_ib *ib;
 	struct amdgpu_ib ib_msg;
@@ -450,8 +488,10 @@ static int amdgpu_vce_get_create_msg(struct amdgpu_ring *ring, uint32_t handle,
 	uint64_t addr;
 	int i, r;
 
-	r = amdgpu_job_alloc_with_ib(ring->adev, ib_size_dw * 4,
-				     AMDGPU_IB_POOL_DIRECT, &job);
+	r = amdgpu_job_alloc_with_ib(ring->adev, &ring->adev->vce.entity,
+				     AMDGPU_FENCE_OWNER_UNDEFINED,
+				     ib_size_dw * 4, AMDGPU_IB_POOL_DIRECT,
+				     &job, AMDGPU_KERNEL_JOB_ID_VCN_RING_TEST);
 	if (r)
 		return r;
 
@@ -505,7 +545,7 @@ static int amdgpu_vce_get_create_msg(struct amdgpu_ring *ring, uint32_t handle,
 		ib->ptr[i] = 0x0;
 
 	r = amdgpu_job_submit_direct(job, ring, &f);
-	amdgpu_ib_free(ring->adev, &ib_msg, f);
+	amdgpu_ib_free(&ib_msg, f);
 	if (r)
 		goto err;
 
@@ -532,15 +572,18 @@ err:
 static int amdgpu_vce_get_destroy_msg(struct amdgpu_ring *ring, uint32_t handle,
 				      bool direct, struct dma_fence **fence)
 {
-	const unsigned ib_size_dw = 1024;
+	const unsigned int ib_size_dw = 1024;
 	struct amdgpu_job *job;
 	struct amdgpu_ib *ib;
 	struct dma_fence *f = NULL;
 	int i, r;
 
-	r = amdgpu_job_alloc_with_ib(ring->adev, ib_size_dw * 4,
+	r = amdgpu_job_alloc_with_ib(ring->adev, &ring->adev->vce.entity,
+				     AMDGPU_FENCE_OWNER_UNDEFINED,
+				     ib_size_dw * 4,
 				     direct ? AMDGPU_IB_POOL_DIRECT :
-				     AMDGPU_IB_POOL_DELAYED, &job);
+				     AMDGPU_IB_POOL_DELAYED, &job,
+				     AMDGPU_KERNEL_JOB_ID_VCN_RING_TEST);
 	if (r)
 		return r;
 
@@ -570,8 +613,7 @@ static int amdgpu_vce_get_destroy_msg(struct amdgpu_ring *ring, uint32_t handle,
 	if (direct)
 		r = amdgpu_job_submit_direct(job, ring, &f);
 	else
-		r = amdgpu_job_submit(job, &ring->adev->vce.entity,
-				      AMDGPU_FENCE_OWNER_UNDEFINED, &f);
+		f = amdgpu_job_submit(job);
 	if (r)
 		goto err;
 
@@ -588,8 +630,8 @@ err:
 /**
  * amdgpu_vce_validate_bo - make sure not to cross 4GB boundary
  *
- * @p: parser context
- * @ib_idx: indirect buffer to use
+ * @p: cs parser
+ * @ib: indirect buffer to use
  * @lo: address of lower dword
  * @hi: address of higher dword
  * @size: minimum size
@@ -597,19 +639,20 @@ err:
  *
  * Make sure that no BO cross a 4GB boundary.
  */
-static int amdgpu_vce_validate_bo(struct amdgpu_cs_parser *p, uint32_t ib_idx,
-				  int lo, int hi, unsigned size, int32_t index)
+static int amdgpu_vce_validate_bo(struct amdgpu_cs_parser *p,
+				  struct amdgpu_ib *ib, int lo, int hi,
+				  unsigned int size, int32_t index)
 {
 	int64_t offset = ((uint64_t)size) * ((int64_t)index);
 	struct ttm_operation_ctx ctx = { false, false };
 	struct amdgpu_bo_va_mapping *mapping;
-	unsigned i, fpfn, lpfn;
+	unsigned int i, fpfn, lpfn;
 	struct amdgpu_bo *bo;
 	uint64_t addr;
 	int r;
 
-	addr = ((uint64_t)amdgpu_get_ib_value(p, ib_idx, lo)) |
-	       ((uint64_t)amdgpu_get_ib_value(p, ib_idx, hi)) << 32;
+	addr = ((uint64_t)amdgpu_ib_get_value(ib, lo)) |
+	       ((uint64_t)amdgpu_ib_get_value(ib, hi)) << 32;
 	if (index >= 0) {
 		addr += offset;
 		fpfn = PAGE_ALIGN(offset) >> PAGE_SHIFT;
@@ -621,7 +664,7 @@ static int amdgpu_vce_validate_bo(struct amdgpu_cs_parser *p, uint32_t ib_idx,
 
 	r = amdgpu_cs_find_mapping(p, addr, &bo, &mapping);
 	if (r) {
-		DRM_ERROR("Can't find BO for addr 0x%010Lx %d %d %d %d\n",
+		DRM_ERROR("Can't find BO for addr 0x%010llx %d %d %d %d\n",
 			  addr, lo, hi, size, index);
 		return r;
 	}
@@ -639,7 +682,7 @@ static int amdgpu_vce_validate_bo(struct amdgpu_cs_parser *p, uint32_t ib_idx,
  * amdgpu_vce_cs_reloc - command submission relocation
  *
  * @p: parser context
- * @ib_idx: indirect buffer to use
+ * @ib: indirect buffer to use
  * @lo: address of lower dword
  * @hi: address of higher dword
  * @size: minimum size
@@ -647,8 +690,8 @@ static int amdgpu_vce_validate_bo(struct amdgpu_cs_parser *p, uint32_t ib_idx,
  *
  * Patch relocation inside command stream with real buffer address
  */
-static int amdgpu_vce_cs_reloc(struct amdgpu_cs_parser *p, uint32_t ib_idx,
-			       int lo, int hi, unsigned size, uint32_t index)
+static int amdgpu_vce_cs_reloc(struct amdgpu_cs_parser *p, struct amdgpu_ib *ib,
+			       int lo, int hi, unsigned int size, uint32_t index)
 {
 	struct amdgpu_bo_va_mapping *mapping;
 	struct amdgpu_bo *bo;
@@ -658,20 +701,20 @@ static int amdgpu_vce_cs_reloc(struct amdgpu_cs_parser *p, uint32_t ib_idx,
 	if (index == 0xffffffff)
 		index = 0;
 
-	addr = ((uint64_t)amdgpu_get_ib_value(p, ib_idx, lo)) |
-	       ((uint64_t)amdgpu_get_ib_value(p, ib_idx, hi)) << 32;
+	addr = ((uint64_t)amdgpu_ib_get_value(ib, lo)) |
+	       ((uint64_t)amdgpu_ib_get_value(ib, hi)) << 32;
 	addr += ((uint64_t)size) * ((uint64_t)index);
 
 	r = amdgpu_cs_find_mapping(p, addr, &bo, &mapping);
 	if (r) {
-		DRM_ERROR("Can't find BO for addr 0x%010Lx %d %d %d %d\n",
+		DRM_ERROR("Can't find BO for addr 0x%010llx %d %d %d %d\n",
 			  addr, lo, hi, size, index);
 		return r;
 	}
 
 	if ((addr + (uint64_t)size) >
 	    (mapping->last + 1) * AMDGPU_GPU_PAGE_SIZE) {
-		DRM_ERROR("BO too small for addr 0x%010Lx %d %d\n",
+		DRM_ERROR("BO too small for addr 0x%010llx %d %d\n",
 			  addr, lo, hi);
 		return -EINVAL;
 	}
@@ -680,8 +723,8 @@ static int amdgpu_vce_cs_reloc(struct amdgpu_cs_parser *p, uint32_t ib_idx,
 	addr += amdgpu_bo_gpu_offset(bo);
 	addr -= ((uint64_t)size) * ((uint64_t)index);
 
-	amdgpu_set_ib_value(p, ib_idx, lo, lower_32_bits(addr));
-	amdgpu_set_ib_value(p, ib_idx, hi, upper_32_bits(addr));
+	amdgpu_ib_set_value(ib, lo, lower_32_bits(addr));
+	amdgpu_ib_set_value(ib, hi, upper_32_bits(addr));
 
 	return 0;
 }
@@ -694,12 +737,12 @@ static int amdgpu_vce_cs_reloc(struct amdgpu_cs_parser *p, uint32_t ib_idx,
  * @allocated: allocated a new handle?
  *
  * Validates the handle and return the found session index or -EINVAL
- * we we don't have another free session index.
+ * we don't have another free session index.
  */
 static int amdgpu_vce_validate_handle(struct amdgpu_cs_parser *p,
 				      uint32_t handle, uint32_t *allocated)
 {
-	unsigned i;
+	unsigned int i;
 
 	/* validate the handle */
 	for (i = 0; i < AMDGPU_MAX_VCE_HANDLES; ++i) {
@@ -730,27 +773,29 @@ static int amdgpu_vce_validate_handle(struct amdgpu_cs_parser *p,
  * amdgpu_vce_ring_parse_cs - parse and validate the command stream
  *
  * @p: parser context
- * @ib_idx: indirect buffer to use
+ * @job: the job to parse
+ * @ib: the IB to patch
  */
-int amdgpu_vce_ring_parse_cs(struct amdgpu_cs_parser *p, uint32_t ib_idx)
+int amdgpu_vce_ring_parse_cs(struct amdgpu_cs_parser *p,
+			     struct amdgpu_job *job,
+			     struct amdgpu_ib *ib)
 {
-	struct amdgpu_ib *ib = &p->job->ibs[ib_idx];
-	unsigned fb_idx = 0, bs_idx = 0;
+	unsigned int fb_idx = 0, bs_idx = 0;
 	int session_idx = -1;
 	uint32_t destroyed = 0;
 	uint32_t created = 0;
 	uint32_t allocated = 0;
 	uint32_t tmp, handle = 0;
-	uint32_t *size = &tmp;
-	unsigned idx;
+	uint32_t dummy = 0xffffffff;
+	uint32_t *size = &dummy;
+	unsigned int idx;
 	int i, r = 0;
 
-	p->job->vm = NULL;
-	ib->gpu_addr = amdgpu_sa_bo_gpu_addr(ib->sa_bo);
+	job->vm = NULL;
 
 	for (idx = 0; idx < ib->length_dw;) {
-		uint32_t len = amdgpu_get_ib_value(p, ib_idx, idx);
-		uint32_t cmd = amdgpu_get_ib_value(p, ib_idx, idx + 1);
+		uint32_t len = amdgpu_ib_get_value(ib, idx);
+		uint32_t cmd = amdgpu_ib_get_value(ib, idx + 1);
 
 		if ((len < 8) || (len & 3)) {
 			DRM_ERROR("invalid VCE command length (%d)!\n", len);
@@ -760,52 +805,52 @@ int amdgpu_vce_ring_parse_cs(struct amdgpu_cs_parser *p, uint32_t ib_idx)
 
 		switch (cmd) {
 		case 0x00000002: /* task info */
-			fb_idx = amdgpu_get_ib_value(p, ib_idx, idx + 6);
-			bs_idx = amdgpu_get_ib_value(p, ib_idx, idx + 7);
+			fb_idx = amdgpu_ib_get_value(ib, idx + 6);
+			bs_idx = amdgpu_ib_get_value(ib, idx + 7);
 			break;
 
 		case 0x03000001: /* encode */
-			r = amdgpu_vce_validate_bo(p, ib_idx, idx + 10,
-						   idx + 9, 0, 0);
+			r = amdgpu_vce_validate_bo(p, ib, idx + 10, idx + 9,
+						   0, 0);
 			if (r)
 				goto out;
 
-			r = amdgpu_vce_validate_bo(p, ib_idx, idx + 12,
-						   idx + 11, 0, 0);
+			r = amdgpu_vce_validate_bo(p, ib, idx + 12, idx + 11,
+						   0, 0);
 			if (r)
 				goto out;
 			break;
 
 		case 0x05000001: /* context buffer */
-			r = amdgpu_vce_validate_bo(p, ib_idx, idx + 3,
-						   idx + 2, 0, 0);
+			r = amdgpu_vce_validate_bo(p, ib, idx + 3, idx + 2,
+						   0, 0);
 			if (r)
 				goto out;
 			break;
 
 		case 0x05000004: /* video bitstream buffer */
-			tmp = amdgpu_get_ib_value(p, ib_idx, idx + 4);
-			r = amdgpu_vce_validate_bo(p, ib_idx, idx + 3, idx + 2,
+			tmp = amdgpu_ib_get_value(ib, idx + 4);
+			r = amdgpu_vce_validate_bo(p, ib, idx + 3, idx + 2,
 						   tmp, bs_idx);
 			if (r)
 				goto out;
 			break;
 
 		case 0x05000005: /* feedback buffer */
-			r = amdgpu_vce_validate_bo(p, ib_idx, idx + 3, idx + 2,
+			r = amdgpu_vce_validate_bo(p, ib, idx + 3, idx + 2,
 						   4096, fb_idx);
 			if (r)
 				goto out;
 			break;
 
 		case 0x0500000d: /* MV buffer */
-			r = amdgpu_vce_validate_bo(p, ib_idx, idx + 3,
-							idx + 2, 0, 0);
+			r = amdgpu_vce_validate_bo(p, ib, idx + 3, idx + 2,
+						   0, 0);
 			if (r)
 				goto out;
 
-			r = amdgpu_vce_validate_bo(p, ib_idx, idx + 8,
-							idx + 7, 0, 0);
+			r = amdgpu_vce_validate_bo(p, ib, idx + 8, idx + 7,
+						   0, 0);
 			if (r)
 				goto out;
 			break;
@@ -815,12 +860,12 @@ int amdgpu_vce_ring_parse_cs(struct amdgpu_cs_parser *p, uint32_t ib_idx)
 	}
 
 	for (idx = 0; idx < ib->length_dw;) {
-		uint32_t len = amdgpu_get_ib_value(p, ib_idx, idx);
-		uint32_t cmd = amdgpu_get_ib_value(p, ib_idx, idx + 1);
+		uint32_t len = amdgpu_ib_get_value(ib, idx);
+		uint32_t cmd = amdgpu_ib_get_value(ib, idx + 1);
 
 		switch (cmd) {
 		case 0x00000001: /* session */
-			handle = amdgpu_get_ib_value(p, ib_idx, idx + 2);
+			handle = amdgpu_ib_get_value(ib, idx + 2);
 			session_idx = amdgpu_vce_validate_handle(p, handle,
 								 &allocated);
 			if (session_idx < 0) {
@@ -831,8 +876,8 @@ int amdgpu_vce_ring_parse_cs(struct amdgpu_cs_parser *p, uint32_t ib_idx)
 			break;
 
 		case 0x00000002: /* task info */
-			fb_idx = amdgpu_get_ib_value(p, ib_idx, idx + 6);
-			bs_idx = amdgpu_get_ib_value(p, ib_idx, idx + 7);
+			fb_idx = amdgpu_ib_get_value(ib, idx + 6);
+			bs_idx = amdgpu_ib_get_value(ib, idx + 7);
 			break;
 
 		case 0x01000001: /* create */
@@ -847,8 +892,8 @@ int amdgpu_vce_ring_parse_cs(struct amdgpu_cs_parser *p, uint32_t ib_idx)
 				goto out;
 			}
 
-			*size = amdgpu_get_ib_value(p, ib_idx, idx + 8) *
-				amdgpu_get_ib_value(p, ib_idx, idx + 10) *
+			*size = amdgpu_ib_get_value(ib, idx + 8) *
+				amdgpu_ib_get_value(ib, idx + 10) *
 				8 * 3 / 2;
 			break;
 
@@ -877,12 +922,12 @@ int amdgpu_vce_ring_parse_cs(struct amdgpu_cs_parser *p, uint32_t ib_idx)
 			break;
 
 		case 0x03000001: /* encode */
-			r = amdgpu_vce_cs_reloc(p, ib_idx, idx + 10, idx + 9,
+			r = amdgpu_vce_cs_reloc(p, ib, idx + 10, idx + 9,
 						*size, 0);
 			if (r)
 				goto out;
 
-			r = amdgpu_vce_cs_reloc(p, ib_idx, idx + 12, idx + 11,
+			r = amdgpu_vce_cs_reloc(p, ib, idx + 12, idx + 11,
 						*size / 3, 0);
 			if (r)
 				goto out;
@@ -893,35 +938,35 @@ int amdgpu_vce_ring_parse_cs(struct amdgpu_cs_parser *p, uint32_t ib_idx)
 			break;
 
 		case 0x05000001: /* context buffer */
-			r = amdgpu_vce_cs_reloc(p, ib_idx, idx + 3, idx + 2,
+			r = amdgpu_vce_cs_reloc(p, ib, idx + 3, idx + 2,
 						*size * 2, 0);
 			if (r)
 				goto out;
 			break;
 
 		case 0x05000004: /* video bitstream buffer */
-			tmp = amdgpu_get_ib_value(p, ib_idx, idx + 4);
-			r = amdgpu_vce_cs_reloc(p, ib_idx, idx + 3, idx + 2,
+			tmp = amdgpu_ib_get_value(ib, idx + 4);
+			r = amdgpu_vce_cs_reloc(p, ib, idx + 3, idx + 2,
 						tmp, bs_idx);
 			if (r)
 				goto out;
 			break;
 
 		case 0x05000005: /* feedback buffer */
-			r = amdgpu_vce_cs_reloc(p, ib_idx, idx + 3, idx + 2,
+			r = amdgpu_vce_cs_reloc(p, ib, idx + 3, idx + 2,
 						4096, fb_idx);
 			if (r)
 				goto out;
 			break;
 
 		case 0x0500000d: /* MV buffer */
-			r = amdgpu_vce_cs_reloc(p, ib_idx, idx + 3,
-							idx + 2, *size, 0);
+			r = amdgpu_vce_cs_reloc(p, ib, idx + 3,
+						idx + 2, *size, 0);
 			if (r)
 				goto out;
 
-			r = amdgpu_vce_cs_reloc(p, ib_idx, idx + 8,
-							idx + 7, *size / 12, 0);
+			r = amdgpu_vce_cs_reloc(p, ib, idx + 8,
+						idx + 7, *size / 12, 0);
 			if (r)
 				goto out;
 			break;
@@ -966,11 +1011,13 @@ out:
  * amdgpu_vce_ring_parse_cs_vm - parse the command stream in VM mode
  *
  * @p: parser context
- * @ib_idx: indirect buffer to use
+ * @job: the job to parse
+ * @ib: the IB to patch
  */
-int amdgpu_vce_ring_parse_cs_vm(struct amdgpu_cs_parser *p, uint32_t ib_idx)
+int amdgpu_vce_ring_parse_cs_vm(struct amdgpu_cs_parser *p,
+				struct amdgpu_job *job,
+				struct amdgpu_ib *ib)
 {
-	struct amdgpu_ib *ib = &p->job->ibs[ib_idx];
 	int session_idx = -1;
 	uint32_t destroyed = 0;
 	uint32_t created = 0;
@@ -979,8 +1026,8 @@ int amdgpu_vce_ring_parse_cs_vm(struct amdgpu_cs_parser *p, uint32_t ib_idx)
 	int i, r = 0, idx = 0;
 
 	while (idx < ib->length_dw) {
-		uint32_t len = amdgpu_get_ib_value(p, ib_idx, idx);
-		uint32_t cmd = amdgpu_get_ib_value(p, ib_idx, idx + 1);
+		uint32_t len = amdgpu_ib_get_value(ib, idx);
+		uint32_t cmd = amdgpu_ib_get_value(ib, idx + 1);
 
 		if ((len < 8) || (len & 3)) {
 			DRM_ERROR("invalid VCE command length (%d)!\n", len);
@@ -990,7 +1037,7 @@ int amdgpu_vce_ring_parse_cs_vm(struct amdgpu_cs_parser *p, uint32_t ib_idx)
 
 		switch (cmd) {
 		case 0x00000001: /* session */
-			handle = amdgpu_get_ib_value(p, ib_idx, idx + 2);
+			handle = amdgpu_ib_get_value(ib, idx + 2);
 			session_idx = amdgpu_vce_validate_handle(p, handle,
 								 &allocated);
 			if (session_idx < 0) {
@@ -1039,7 +1086,6 @@ out:
 	if (!r) {
 		/* No error, free all destroyed handle slots */
 		tmp = destroyed;
-		amdgpu_ib_free(p->adev, ib, NULL);
 	} else {
 		/* Error during parsing, free all allocated handle slots */
 		tmp = allocated;
@@ -1082,7 +1128,7 @@ void amdgpu_vce_ring_emit_ib(struct amdgpu_ring *ring,
  *
  */
 void amdgpu_vce_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 seq,
-				unsigned flags)
+				unsigned int flags)
 {
 	WARN_ON(flags & AMDGPU_FENCE_FLAG_64BIT);
 
@@ -1104,7 +1150,7 @@ int amdgpu_vce_ring_test_ring(struct amdgpu_ring *ring)
 {
 	struct amdgpu_device *adev = ring->adev;
 	uint32_t rptr;
-	unsigned i;
+	unsigned int i;
 	int r, timeout = adev->usec_timeout;
 
 	/* skip ring test for sriov*/
@@ -1169,7 +1215,7 @@ error:
 
 enum amdgpu_ring_priority_level amdgpu_vce_get_ring_prio(int ring)
 {
-	switch(ring) {
+	switch (ring) {
 	case 0:
 		return AMDGPU_RING_PRIO_0;
 	case 1:
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.h
index be4a6e773c5b..1c3464ce5037 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.h
@@ -51,16 +51,22 @@ struct amdgpu_vce {
 	struct drm_sched_entity	entity;
 	uint32_t                srbm_soft_reset;
 	unsigned		num_rings;
+	uint32_t		keyselect;
 };
 
+int amdgpu_vce_early_init(struct amdgpu_device *adev);
 int amdgpu_vce_sw_init(struct amdgpu_device *adev, unsigned long size);
 int amdgpu_vce_sw_fini(struct amdgpu_device *adev);
-int amdgpu_vce_entity_init(struct amdgpu_device *adev);
+int amdgpu_vce_entity_init(struct amdgpu_device *adev, struct amdgpu_ring *ring);
 int amdgpu_vce_suspend(struct amdgpu_device *adev);
 int amdgpu_vce_resume(struct amdgpu_device *adev);
 void amdgpu_vce_free_handles(struct amdgpu_device *adev, struct drm_file *filp);
-int amdgpu_vce_ring_parse_cs(struct amdgpu_cs_parser *p, uint32_t ib_idx);
-int amdgpu_vce_ring_parse_cs_vm(struct amdgpu_cs_parser *p, uint32_t ib_idx);
+u32 amdgpu_vce_required_gart_pages(struct amdgpu_device *adev);
+int amdgpu_vce_ring_parse_cs(struct amdgpu_cs_parser *p, struct amdgpu_job *job,
+			     struct amdgpu_ib *ib);
+int amdgpu_vce_ring_parse_cs_vm(struct amdgpu_cs_parser *p,
+				struct amdgpu_job *job,
+				struct amdgpu_ib *ib);
 void amdgpu_vce_ring_emit_ib(struct amdgpu_ring *ring, struct amdgpu_job *job,
 				struct amdgpu_ib *ib, uint32_t flags);
 void amdgpu_vce_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 seq,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
index 585961c2f5f2..5e0786ea911b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
@@ -1,5 +1,5 @@
 /*
- * Copyright 2016 Advanced Micro Devices, Inc.
+ * Copyright 2016-2024 Advanced Micro Devices, Inc.
  * All Rights Reserved.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
@@ -26,7 +26,9 @@
 
 #include <linux/firmware.h>
 #include <linux/module.h>
+#include <linux/dmi.h>
 #include <linux/pci.h>
+#include <linux/debugfs.h>
 #include <drm/drm_drv.h>
 
 #include "amdgpu.h"
@@ -35,22 +37,32 @@
 #include "soc15d.h"
 
 /* Firmware Names */
-#define FIRMWARE_RAVEN		"amdgpu/raven_vcn.bin"
-#define FIRMWARE_PICASSO	"amdgpu/picasso_vcn.bin"
-#define FIRMWARE_RAVEN2		"amdgpu/raven2_vcn.bin"
-#define FIRMWARE_ARCTURUS	"amdgpu/arcturus_vcn.bin"
-#define FIRMWARE_RENOIR		"amdgpu/renoir_vcn.bin"
-#define FIRMWARE_GREEN_SARDINE	"amdgpu/green_sardine_vcn.bin"
-#define FIRMWARE_NAVI10		"amdgpu/navi10_vcn.bin"
-#define FIRMWARE_NAVI14		"amdgpu/navi14_vcn.bin"
-#define FIRMWARE_NAVI12		"amdgpu/navi12_vcn.bin"
-#define FIRMWARE_SIENNA_CICHLID	"amdgpu/sienna_cichlid_vcn.bin"
-#define FIRMWARE_NAVY_FLOUNDER	"amdgpu/navy_flounder_vcn.bin"
-#define FIRMWARE_VANGOGH	"amdgpu/vangogh_vcn.bin"
+#define FIRMWARE_RAVEN			"amdgpu/raven_vcn.bin"
+#define FIRMWARE_PICASSO		"amdgpu/picasso_vcn.bin"
+#define FIRMWARE_RAVEN2			"amdgpu/raven2_vcn.bin"
+#define FIRMWARE_ARCTURUS		"amdgpu/arcturus_vcn.bin"
+#define FIRMWARE_RENOIR			"amdgpu/renoir_vcn.bin"
+#define FIRMWARE_GREEN_SARDINE		"amdgpu/green_sardine_vcn.bin"
+#define FIRMWARE_NAVI10			"amdgpu/navi10_vcn.bin"
+#define FIRMWARE_NAVI14			"amdgpu/navi14_vcn.bin"
+#define FIRMWARE_NAVI12			"amdgpu/navi12_vcn.bin"
+#define FIRMWARE_SIENNA_CICHLID		"amdgpu/sienna_cichlid_vcn.bin"
+#define FIRMWARE_NAVY_FLOUNDER		"amdgpu/navy_flounder_vcn.bin"
+#define FIRMWARE_VANGOGH		"amdgpu/vangogh_vcn.bin"
 #define FIRMWARE_DIMGREY_CAVEFISH	"amdgpu/dimgrey_cavefish_vcn.bin"
-#define FIRMWARE_ALDEBARAN	"amdgpu/aldebaran_vcn.bin"
-#define FIRMWARE_BEIGE_GOBY	"amdgpu/beige_goby_vcn.bin"
-#define FIRMWARE_YELLOW_CARP	"amdgpu/yellow_carp_vcn.bin"
+#define FIRMWARE_ALDEBARAN		"amdgpu/aldebaran_vcn.bin"
+#define FIRMWARE_BEIGE_GOBY		"amdgpu/beige_goby_vcn.bin"
+#define FIRMWARE_YELLOW_CARP		"amdgpu/yellow_carp_vcn.bin"
+#define FIRMWARE_VCN_3_1_2		"amdgpu/vcn_3_1_2.bin"
+#define FIRMWARE_VCN4_0_0		"amdgpu/vcn_4_0_0.bin"
+#define FIRMWARE_VCN4_0_2		"amdgpu/vcn_4_0_2.bin"
+#define FIRMWARE_VCN4_0_3		"amdgpu/vcn_4_0_3.bin"
+#define FIRMWARE_VCN4_0_4		"amdgpu/vcn_4_0_4.bin"
+#define FIRMWARE_VCN4_0_5		"amdgpu/vcn_4_0_5.bin"
+#define FIRMWARE_VCN4_0_6		"amdgpu/vcn_4_0_6.bin"
+#define FIRMWARE_VCN4_0_6_1		"amdgpu/vcn_4_0_6_1.bin"
+#define FIRMWARE_VCN5_0_0		"amdgpu/vcn_5_0_0.bin"
+#define FIRMWARE_VCN5_0_1		"amdgpu/vcn_5_0_1.bin"
 
 MODULE_FIRMWARE(FIRMWARE_RAVEN);
 MODULE_FIRMWARE(FIRMWARE_PICASSO);
@@ -68,124 +80,93 @@ MODULE_FIRMWARE(FIRMWARE_VANGOGH);
 MODULE_FIRMWARE(FIRMWARE_DIMGREY_CAVEFISH);
 MODULE_FIRMWARE(FIRMWARE_BEIGE_GOBY);
 MODULE_FIRMWARE(FIRMWARE_YELLOW_CARP);
+MODULE_FIRMWARE(FIRMWARE_VCN_3_1_2);
+MODULE_FIRMWARE(FIRMWARE_VCN4_0_0);
+MODULE_FIRMWARE(FIRMWARE_VCN4_0_2);
+MODULE_FIRMWARE(FIRMWARE_VCN4_0_3);
+MODULE_FIRMWARE(FIRMWARE_VCN4_0_4);
+MODULE_FIRMWARE(FIRMWARE_VCN4_0_5);
+MODULE_FIRMWARE(FIRMWARE_VCN4_0_6);
+MODULE_FIRMWARE(FIRMWARE_VCN4_0_6_1);
+MODULE_FIRMWARE(FIRMWARE_VCN5_0_0);
+MODULE_FIRMWARE(FIRMWARE_VCN5_0_1);
 
 static void amdgpu_vcn_idle_work_handler(struct work_struct *work);
+static void amdgpu_vcn_reg_dump_fini(struct amdgpu_device *adev);
 
-int amdgpu_vcn_sw_init(struct amdgpu_device *adev)
+int amdgpu_vcn_early_init(struct amdgpu_device *adev, int i)
+{
+	char ucode_prefix[25];
+	int r;
+
+	adev->vcn.inst[i].adev = adev;
+	adev->vcn.inst[i].inst = i;
+	amdgpu_ucode_ip_version_decode(adev, UVD_HWIP, ucode_prefix, sizeof(ucode_prefix));
+
+	if (i != 0 && adev->vcn.per_inst_fw) {
+		r = amdgpu_ucode_request(adev, &adev->vcn.inst[i].fw,
+					 AMDGPU_UCODE_REQUIRED,
+					 "amdgpu/%s_%d.bin", ucode_prefix, i);
+		if (r)
+			amdgpu_ucode_release(&adev->vcn.inst[i].fw);
+	} else {
+		if (!adev->vcn.inst[0].fw) {
+			r = amdgpu_ucode_request(adev, &adev->vcn.inst[0].fw,
+						 AMDGPU_UCODE_REQUIRED,
+						 "amdgpu/%s.bin", ucode_prefix);
+			if (r)
+				amdgpu_ucode_release(&adev->vcn.inst[0].fw);
+		} else {
+			r = 0;
+		}
+		adev->vcn.inst[i].fw = adev->vcn.inst[0].fw;
+	}
+
+	return r;
+}
+
+int amdgpu_vcn_sw_init(struct amdgpu_device *adev, int i)
 {
 	unsigned long bo_size;
-	const char *fw_name;
 	const struct common_firmware_header *hdr;
 	unsigned char fw_check;
-	int i, r;
-
-	INIT_DELAYED_WORK(&adev->vcn.idle_work, amdgpu_vcn_idle_work_handler);
-	mutex_init(&adev->vcn.vcn_pg_lock);
-	mutex_init(&adev->vcn.vcn1_jpeg1_workaround);
-	atomic_set(&adev->vcn.total_submission_cnt, 0);
-	for (i = 0; i < adev->vcn.num_vcn_inst; i++)
-		atomic_set(&adev->vcn.inst[i].dpg_enc_submission_cnt, 0);
-
-	switch (adev->ip_versions[UVD_HWIP][0]) {
-	case IP_VERSION(1, 0, 0):
-	case IP_VERSION(1, 0, 1):
-		if (adev->apu_flags & AMD_APU_IS_RAVEN2)
-			fw_name = FIRMWARE_RAVEN2;
-		else if (adev->apu_flags & AMD_APU_IS_PICASSO)
-			fw_name = FIRMWARE_PICASSO;
-		else
-			fw_name = FIRMWARE_RAVEN;
-		break;
-	case IP_VERSION(2, 5, 0):
-		fw_name = FIRMWARE_ARCTURUS;
-		if ((adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) &&
-		    (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG))
-			adev->vcn.indirect_sram = true;
-		break;
-	case IP_VERSION(2, 2, 0):
-		if (adev->apu_flags & AMD_APU_IS_RENOIR)
-			fw_name = FIRMWARE_RENOIR;
-		else
-			fw_name = FIRMWARE_GREEN_SARDINE;
-
-		if ((adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) &&
-		    (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG))
-			adev->vcn.indirect_sram = true;
-		break;
-	case IP_VERSION(2, 6, 0):
-		fw_name = FIRMWARE_ALDEBARAN;
-		if ((adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) &&
-		    (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG))
-			adev->vcn.indirect_sram = true;
-		break;
-	case IP_VERSION(2, 0, 0):
-		fw_name = FIRMWARE_NAVI10;
-		if ((adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) &&
-		    (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG))
-			adev->vcn.indirect_sram = true;
-		break;
-	case IP_VERSION(2, 0, 2):
-		if (adev->asic_type == CHIP_NAVI12)
-			fw_name = FIRMWARE_NAVI12;
-		else
-			fw_name = FIRMWARE_NAVI14;
-		if ((adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) &&
-		    (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG))
-			adev->vcn.indirect_sram = true;
-		break;
-	case IP_VERSION(3, 0, 0):
-	case IP_VERSION(3, 0, 64):
-	case IP_VERSION(3, 0, 192):
-		if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 3, 0))
-			fw_name = FIRMWARE_SIENNA_CICHLID;
-		else
-			fw_name = FIRMWARE_NAVY_FLOUNDER;
-		if ((adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) &&
-		    (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG))
-			adev->vcn.indirect_sram = true;
-		break;
-	case IP_VERSION(3, 0, 2):
-		fw_name = FIRMWARE_VANGOGH;
-		break;
-	case IP_VERSION(3, 0, 16):
-		fw_name = FIRMWARE_DIMGREY_CAVEFISH;
-		if ((adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) &&
-		    (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG))
-			adev->vcn.indirect_sram = true;
-		break;
-	case IP_VERSION(3, 0, 33):
-		fw_name = FIRMWARE_BEIGE_GOBY;
-		if ((adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) &&
-		    (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG))
-			adev->vcn.indirect_sram = true;
-		break;
-	case IP_VERSION(3, 1, 1):
-		fw_name = FIRMWARE_YELLOW_CARP;
-		if ((adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) &&
-		    (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG))
-			adev->vcn.indirect_sram = true;
-		break;
-	default:
-		return -EINVAL;
-	}
+	unsigned int fw_shared_size, log_offset;
+	int r;
 
-	r = request_firmware(&adev->vcn.fw, fw_name, adev->dev);
-	if (r) {
-		dev_err(adev->dev, "amdgpu_vcn: Can't load firmware \"%s\"\n",
-			fw_name);
-		return r;
+	mutex_init(&adev->vcn.inst[i].vcn1_jpeg1_workaround);
+	mutex_init(&adev->vcn.inst[i].vcn_pg_lock);
+	mutex_init(&adev->vcn.inst[i].engine_reset_mutex);
+	atomic_set(&adev->vcn.inst[i].total_submission_cnt, 0);
+	INIT_DELAYED_WORK(&adev->vcn.inst[i].idle_work, amdgpu_vcn_idle_work_handler);
+	atomic_set(&adev->vcn.inst[i].dpg_enc_submission_cnt, 0);
+	if ((adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) &&
+	    (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG))
+		adev->vcn.inst[i].indirect_sram = true;
+
+	/*
+	 * Some Steam Deck's BIOS versions are incompatible with the
+	 * indirect SRAM mode, leading to amdgpu being unable to get
+	 * properly probed (and even potentially crashing the kernel).
+	 * Hence, check for these versions here - notice this is
+	 * restricted to Vangogh (Deck's APU).
+	 */
+	if (amdgpu_ip_version(adev, UVD_HWIP, 0) == IP_VERSION(3, 0, 2)) {
+		const char *bios_ver = dmi_get_system_info(DMI_BIOS_VERSION);
+
+		if (bios_ver && (!strncmp("F7A0113", bios_ver, 7) ||
+				 !strncmp("F7A0114", bios_ver, 7))) {
+			adev->vcn.inst[i].indirect_sram = false;
+			dev_info(adev->dev,
+				 "Steam Deck quirk: indirect SRAM disabled on BIOS %s\n", bios_ver);
+		}
 	}
 
-	r = amdgpu_ucode_validate(adev->vcn.fw);
-	if (r) {
-		dev_err(adev->dev, "amdgpu_vcn: Can't validate firmware \"%s\"\n",
-			fw_name);
-		release_firmware(adev->vcn.fw);
-		adev->vcn.fw = NULL;
-		return r;
-	}
+	/* from vcn4 and above, only unified queue is used */
+	adev->vcn.inst[i].using_unified_queue =
+		amdgpu_ip_version(adev, UVD_HWIP, 0) >= IP_VERSION(4, 0, 0);
 
-	hdr = (const struct common_firmware_header *)adev->vcn.fw->data;
+	hdr = (const struct common_firmware_header *)adev->vcn.inst[i].fw->data;
+	adev->vcn.inst[i].fw_version = le32_to_cpu(hdr->ucode_version);
 	adev->vcn.fw_version = le32_to_cpu(hdr->ucode_version);
 
 	/* Bit 20-23, it is encode major and non-zero for new naming convention.
@@ -203,283 +184,378 @@ int amdgpu_vcn_sw_init(struct amdgpu_device *adev)
 		enc_major = fw_check;
 		dec_ver = (le32_to_cpu(hdr->ucode_version) >> 24) & 0xf;
 		vep = (le32_to_cpu(hdr->ucode_version) >> 28) & 0xf;
-		DRM_INFO("Found VCN firmware Version ENC: %u.%u DEC: %u VEP: %u Revision: %u\n",
-			enc_major, enc_minor, dec_ver, vep, fw_rev);
+		dev_info(adev->dev,
+			 "[VCN instance %d] Found VCN firmware Version ENC: %u.%u DEC: %u VEP: %u Revision: %u\n",
+			 i, enc_major, enc_minor, dec_ver, vep, fw_rev);
 	} else {
 		unsigned int version_major, version_minor, family_id;
 
 		family_id = le32_to_cpu(hdr->ucode_version) & 0xff;
 		version_major = (le32_to_cpu(hdr->ucode_version) >> 24) & 0xff;
 		version_minor = (le32_to_cpu(hdr->ucode_version) >> 8) & 0xff;
-		DRM_INFO("Found VCN firmware Version: %u.%u Family ID: %u\n",
-			version_major, version_minor, family_id);
+		dev_info(adev->dev, "[VCN instance %d] Found VCN firmware Version: %u.%u Family ID: %u\n",
+			 i, version_major, version_minor, family_id);
 	}
 
 	bo_size = AMDGPU_VCN_STACK_SIZE + AMDGPU_VCN_CONTEXT_SIZE;
 	if (adev->firmware.load_type != AMDGPU_FW_LOAD_PSP)
 		bo_size += AMDGPU_GPU_PAGE_ALIGN(le32_to_cpu(hdr->ucode_size_bytes) + 8);
-	bo_size += AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_fw_shared));
 
-	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
-		if (adev->vcn.harvest_config & (1 << i))
-			continue;
+	if (amdgpu_ip_version(adev, UVD_HWIP, 0) >= IP_VERSION(5, 0, 0)) {
+		fw_shared_size = AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_vcn5_fw_shared));
+		log_offset = offsetof(struct amdgpu_vcn5_fw_shared, fw_log);
+	} else if (amdgpu_ip_version(adev, UVD_HWIP, 0) >= IP_VERSION(4, 0, 0)) {
+		fw_shared_size = AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_vcn4_fw_shared));
+		log_offset = offsetof(struct amdgpu_vcn4_fw_shared, fw_log);
+	} else {
+		fw_shared_size = AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_fw_shared));
+		log_offset = offsetof(struct amdgpu_fw_shared, fw_log);
+	}
 
-		r = amdgpu_bo_create_kernel(adev, bo_size, PAGE_SIZE,
-						AMDGPU_GEM_DOMAIN_VRAM, &adev->vcn.inst[i].vcpu_bo,
-						&adev->vcn.inst[i].gpu_addr, &adev->vcn.inst[i].cpu_addr);
+	bo_size += fw_shared_size;
+
+	if (amdgpu_vcnfw_log)
+		bo_size += AMDGPU_VCNFW_LOG_SIZE;
+
+	r = amdgpu_bo_create_kernel(adev, bo_size, PAGE_SIZE,
+				    AMDGPU_GEM_DOMAIN_VRAM |
+				    AMDGPU_GEM_DOMAIN_GTT,
+				    &adev->vcn.inst[i].vcpu_bo,
+				    &adev->vcn.inst[i].gpu_addr,
+				    &adev->vcn.inst[i].cpu_addr);
+	if (r) {
+		dev_err(adev->dev, "(%d) failed to allocate vcn bo\n", r);
+		return r;
+	}
+
+	adev->vcn.inst[i].fw_shared.cpu_addr = adev->vcn.inst[i].cpu_addr +
+		bo_size - fw_shared_size;
+	adev->vcn.inst[i].fw_shared.gpu_addr = adev->vcn.inst[i].gpu_addr +
+		bo_size - fw_shared_size;
+
+	adev->vcn.inst[i].fw_shared.mem_size = fw_shared_size;
+
+	if (amdgpu_vcnfw_log) {
+		adev->vcn.inst[i].fw_shared.cpu_addr -= AMDGPU_VCNFW_LOG_SIZE;
+		adev->vcn.inst[i].fw_shared.gpu_addr -= AMDGPU_VCNFW_LOG_SIZE;
+		adev->vcn.inst[i].fw_shared.log_offset = log_offset;
+	}
+
+	if (adev->vcn.inst[i].indirect_sram) {
+		r = amdgpu_bo_create_kernel(adev, 64 * 2 * 4, PAGE_SIZE,
+					    AMDGPU_GEM_DOMAIN_VRAM |
+					    AMDGPU_GEM_DOMAIN_GTT,
+					    &adev->vcn.inst[i].dpg_sram_bo,
+					    &adev->vcn.inst[i].dpg_sram_gpu_addr,
+					    &adev->vcn.inst[i].dpg_sram_cpu_addr);
 		if (r) {
-			dev_err(adev->dev, "(%d) failed to allocate vcn bo\n", r);
+			dev_err(adev->dev, "VCN %d (%d) failed to allocate DPG bo\n", i, r);
 			return r;
 		}
-
-		adev->vcn.inst[i].fw_shared_cpu_addr = adev->vcn.inst[i].cpu_addr +
-				bo_size - AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_fw_shared));
-		adev->vcn.inst[i].fw_shared_gpu_addr = adev->vcn.inst[i].gpu_addr +
-				bo_size - AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_fw_shared));
-
-		if (adev->vcn.indirect_sram) {
-			r = amdgpu_bo_create_kernel(adev, 64 * 2 * 4, PAGE_SIZE,
-					AMDGPU_GEM_DOMAIN_VRAM, &adev->vcn.inst[i].dpg_sram_bo,
-					&adev->vcn.inst[i].dpg_sram_gpu_addr, &adev->vcn.inst[i].dpg_sram_cpu_addr);
-			if (r) {
-				dev_err(adev->dev, "VCN %d (%d) failed to allocate DPG bo\n", i, r);
-				return r;
-			}
-		}
 	}
 
 	return 0;
 }
 
-int amdgpu_vcn_sw_fini(struct amdgpu_device *adev)
+void amdgpu_vcn_sw_fini(struct amdgpu_device *adev, int i)
 {
-	int i, j;
+	int j;
 
-	for (j = 0; j < adev->vcn.num_vcn_inst; ++j) {
-		if (adev->vcn.harvest_config & (1 << j))
-			continue;
+	if (adev->vcn.harvest_config & (1 << i))
+		return;
 
-		if (adev->vcn.indirect_sram) {
-			amdgpu_bo_free_kernel(&adev->vcn.inst[j].dpg_sram_bo,
-						  &adev->vcn.inst[j].dpg_sram_gpu_addr,
-						  (void **)&adev->vcn.inst[j].dpg_sram_cpu_addr);
-		}
-		kvfree(adev->vcn.inst[j].saved_bo);
+	amdgpu_bo_free_kernel(
+		&adev->vcn.inst[i].dpg_sram_bo,
+		&adev->vcn.inst[i].dpg_sram_gpu_addr,
+		(void **)&adev->vcn.inst[i].dpg_sram_cpu_addr);
+
+	kvfree(adev->vcn.inst[i].saved_bo);
 
-		amdgpu_bo_free_kernel(&adev->vcn.inst[j].vcpu_bo,
-					  &adev->vcn.inst[j].gpu_addr,
-					  (void **)&adev->vcn.inst[j].cpu_addr);
+	amdgpu_bo_free_kernel(&adev->vcn.inst[i].vcpu_bo,
+			      &adev->vcn.inst[i].gpu_addr,
+			      (void **)&adev->vcn.inst[i].cpu_addr);
 
-		amdgpu_ring_fini(&adev->vcn.inst[j].ring_dec);
+	amdgpu_ring_fini(&adev->vcn.inst[i].ring_dec);
 
-		for (i = 0; i < adev->vcn.num_enc_rings; ++i)
-			amdgpu_ring_fini(&adev->vcn.inst[j].ring_enc[i]);
+	for (j = 0; j < adev->vcn.inst[i].num_enc_rings; ++j)
+		amdgpu_ring_fini(&adev->vcn.inst[i].ring_enc[j]);
+
+	if (adev->vcn.per_inst_fw) {
+		amdgpu_ucode_release(&adev->vcn.inst[i].fw);
+	} else {
+		amdgpu_ucode_release(&adev->vcn.inst[0].fw);
+		adev->vcn.inst[i].fw = NULL;
 	}
 
-	release_firmware(adev->vcn.fw);
-	mutex_destroy(&adev->vcn.vcn1_jpeg1_workaround);
-	mutex_destroy(&adev->vcn.vcn_pg_lock);
+	if (adev->vcn.reg_list)
+		amdgpu_vcn_reg_dump_fini(adev);
 
-	return 0;
+	mutex_destroy(&adev->vcn.inst[i].vcn_pg_lock);
+	mutex_destroy(&adev->vcn.inst[i].vcn1_jpeg1_workaround);
 }
 
 bool amdgpu_vcn_is_disabled_vcn(struct amdgpu_device *adev, enum vcn_ring_type type, uint32_t vcn_instance)
 {
 	bool ret = false;
+	int vcn_config = adev->vcn.inst[vcn_instance].vcn_config;
 
-	int major;
-	int minor;
-	int revision;
-
-	/* if cannot find IP data, then this VCN does not exist */
-	if (amdgpu_discovery_get_vcn_version(adev, vcn_instance, &major, &minor, &revision) != 0)
-		return true;
-
-	if ((type == VCN_ENCODE_RING) && (revision & VCN_BLOCK_ENCODE_DISABLE_MASK)) {
+	if ((type == VCN_ENCODE_RING) && (vcn_config & VCN_BLOCK_ENCODE_DISABLE_MASK))
 		ret = true;
-	} else if ((type == VCN_DECODE_RING) && (revision & VCN_BLOCK_DECODE_DISABLE_MASK)) {
+	else if ((type == VCN_DECODE_RING) && (vcn_config & VCN_BLOCK_DECODE_DISABLE_MASK))
 		ret = true;
-	} else if ((type == VCN_UNIFIED_RING) && (revision & VCN_BLOCK_QUEUE_DISABLE_MASK)) {
+	else if ((type == VCN_UNIFIED_RING) && (vcn_config & VCN_BLOCK_QUEUE_DISABLE_MASK))
 		ret = true;
-	}
 
 	return ret;
 }
 
-int amdgpu_vcn_suspend(struct amdgpu_device *adev)
+static int amdgpu_vcn_save_vcpu_bo_inst(struct amdgpu_device *adev, int i)
 {
-	unsigned size;
+	unsigned int size;
 	void *ptr;
-	int i, idx;
+	int idx;
 
-	cancel_delayed_work_sync(&adev->vcn.idle_work);
+	if (adev->vcn.harvest_config & (1 << i))
+		return 0;
+	if (adev->vcn.inst[i].vcpu_bo == NULL)
+		return 0;
 
-	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
-		if (adev->vcn.harvest_config & (1 << i))
-			continue;
-		if (adev->vcn.inst[i].vcpu_bo == NULL)
-			return 0;
+	size = amdgpu_bo_size(adev->vcn.inst[i].vcpu_bo);
+	ptr = adev->vcn.inst[i].cpu_addr;
 
-		size = amdgpu_bo_size(adev->vcn.inst[i].vcpu_bo);
-		ptr = adev->vcn.inst[i].cpu_addr;
+	adev->vcn.inst[i].saved_bo = kvmalloc(size, GFP_KERNEL);
+	if (!adev->vcn.inst[i].saved_bo)
+		return -ENOMEM;
 
-		adev->vcn.inst[i].saved_bo = kvmalloc(size, GFP_KERNEL);
-		if (!adev->vcn.inst[i].saved_bo)
-			return -ENOMEM;
+	if (drm_dev_enter(adev_to_drm(adev), &idx)) {
+		memcpy_fromio(adev->vcn.inst[i].saved_bo, ptr, size);
+		drm_dev_exit(idx);
+	}
 
-		if (drm_dev_enter(adev_to_drm(adev), &idx)) {
-			memcpy_fromio(adev->vcn.inst[i].saved_bo, ptr, size);
-			drm_dev_exit(idx);
-		}
+	return 0;
+}
+
+int amdgpu_vcn_save_vcpu_bo(struct amdgpu_device *adev)
+{
+	int ret, i;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		ret = amdgpu_vcn_save_vcpu_bo_inst(adev, i);
+		if (ret)
+			return ret;
 	}
+
 	return 0;
 }
 
-int amdgpu_vcn_resume(struct amdgpu_device *adev)
+int amdgpu_vcn_suspend(struct amdgpu_device *adev, int i)
+{
+	bool in_ras_intr = amdgpu_ras_intr_triggered();
+
+	if (adev->vcn.harvest_config & (1 << i))
+		return 0;
+
+	/* err_event_athub and dpc recovery will corrupt VCPU buffer, so we need to
+	 * restore fw data and clear buffer in amdgpu_vcn_resume() */
+	if (in_ras_intr || adev->pcie_reset_ctx.in_link_reset)
+		return 0;
+
+	return amdgpu_vcn_save_vcpu_bo_inst(adev, i);
+}
+
+int amdgpu_vcn_resume(struct amdgpu_device *adev, int i)
 {
-	unsigned size;
+	unsigned int size;
 	void *ptr;
-	int i, idx;
+	int idx;
 
-	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
-		if (adev->vcn.harvest_config & (1 << i))
-			continue;
-		if (adev->vcn.inst[i].vcpu_bo == NULL)
-			return -EINVAL;
+	if (adev->vcn.harvest_config & (1 << i))
+		return 0;
+	if (adev->vcn.inst[i].vcpu_bo == NULL)
+		return -EINVAL;
+
+	size = amdgpu_bo_size(adev->vcn.inst[i].vcpu_bo);
+	ptr = adev->vcn.inst[i].cpu_addr;
 
-		size = amdgpu_bo_size(adev->vcn.inst[i].vcpu_bo);
-		ptr = adev->vcn.inst[i].cpu_addr;
+	if (adev->vcn.inst[i].saved_bo != NULL) {
+		if (drm_dev_enter(adev_to_drm(adev), &idx)) {
+			memcpy_toio(ptr, adev->vcn.inst[i].saved_bo, size);
+			drm_dev_exit(idx);
+		}
+		kvfree(adev->vcn.inst[i].saved_bo);
+		adev->vcn.inst[i].saved_bo = NULL;
+	} else {
+		const struct common_firmware_header *hdr;
+		unsigned int offset;
 
-		if (adev->vcn.inst[i].saved_bo != NULL) {
+		hdr = (const struct common_firmware_header *)adev->vcn.inst[i].fw->data;
+		if (adev->firmware.load_type != AMDGPU_FW_LOAD_PSP) {
+			offset = le32_to_cpu(hdr->ucode_array_offset_bytes);
 			if (drm_dev_enter(adev_to_drm(adev), &idx)) {
-				memcpy_toio(ptr, adev->vcn.inst[i].saved_bo, size);
+				memcpy_toio(adev->vcn.inst[i].cpu_addr,
+					    adev->vcn.inst[i].fw->data + offset,
+					    le32_to_cpu(hdr->ucode_size_bytes));
 				drm_dev_exit(idx);
 			}
-			kvfree(adev->vcn.inst[i].saved_bo);
-			adev->vcn.inst[i].saved_bo = NULL;
-		} else {
-			const struct common_firmware_header *hdr;
-			unsigned offset;
-
-			hdr = (const struct common_firmware_header *)adev->vcn.fw->data;
-			if (adev->firmware.load_type != AMDGPU_FW_LOAD_PSP) {
-				offset = le32_to_cpu(hdr->ucode_array_offset_bytes);
-				if (drm_dev_enter(adev_to_drm(adev), &idx)) {
-					memcpy_toio(adev->vcn.inst[i].cpu_addr, adev->vcn.fw->data + offset,
-						    le32_to_cpu(hdr->ucode_size_bytes));
-					drm_dev_exit(idx);
-				}
-				size -= le32_to_cpu(hdr->ucode_size_bytes);
-				ptr += le32_to_cpu(hdr->ucode_size_bytes);
-			}
-			memset_io(ptr, 0, size);
+			size -= le32_to_cpu(hdr->ucode_size_bytes);
+			ptr += le32_to_cpu(hdr->ucode_size_bytes);
 		}
+		memset_io(ptr, 0, size);
 	}
+
 	return 0;
 }
 
-static void amdgpu_vcn_idle_work_handler(struct work_struct *work)
+void amdgpu_vcn_get_profile(struct amdgpu_device *adev)
 {
-	struct amdgpu_device *adev =
-		container_of(work, struct amdgpu_device, vcn.idle_work.work);
-	unsigned int fences = 0, fence[AMDGPU_MAX_VCN_INSTANCES] = {0};
-	unsigned int i, j;
-	int r = 0;
+	int r;
 
-	for (j = 0; j < adev->vcn.num_vcn_inst; ++j) {
-		if (adev->vcn.harvest_config & (1 << j))
-			continue;
+	mutex_lock(&adev->vcn.workload_profile_mutex);
 
-		for (i = 0; i < adev->vcn.num_enc_rings; ++i) {
-			fence[j] += amdgpu_fence_count_emitted(&adev->vcn.inst[j].ring_enc[i]);
+	if (adev->vcn.workload_profile_active) {
+		mutex_unlock(&adev->vcn.workload_profile_mutex);
+		return;
+	}
+	r = amdgpu_dpm_switch_power_profile(adev, PP_SMC_POWER_PROFILE_VIDEO,
+					    true);
+	if (r)
+		dev_warn(adev->dev,
+			 "(%d) failed to enable video power profile mode\n", r);
+	else
+		adev->vcn.workload_profile_active = true;
+	mutex_unlock(&adev->vcn.workload_profile_mutex);
+}
+
+void amdgpu_vcn_put_profile(struct amdgpu_device *adev)
+{
+	bool pg = true;
+	int r, i;
+
+	mutex_lock(&adev->vcn.workload_profile_mutex);
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+		if (adev->vcn.inst[i].cur_state != AMD_PG_STATE_GATE) {
+			pg = false;
+			break;
 		}
+	}
+
+	if (pg) {
+		r = amdgpu_dpm_switch_power_profile(
+			adev, PP_SMC_POWER_PROFILE_VIDEO, false);
+		if (r)
+			dev_warn(
+				adev->dev,
+				"(%d) failed to disable video power profile mode\n",
+				r);
+		else
+			adev->vcn.workload_profile_active = false;
+	}
 
-		if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)	{
-			struct dpg_pause_state new_state;
+	mutex_unlock(&adev->vcn.workload_profile_mutex);
+}
 
-			if (fence[j] ||
-				unlikely(atomic_read(&adev->vcn.inst[j].dpg_enc_submission_cnt)))
-				new_state.fw_based = VCN_DPG_STATE__PAUSE;
-			else
-				new_state.fw_based = VCN_DPG_STATE__UNPAUSE;
+static void amdgpu_vcn_idle_work_handler(struct work_struct *work)
+{
+	struct amdgpu_vcn_inst *vcn_inst =
+		container_of(work, struct amdgpu_vcn_inst, idle_work.work);
+	struct amdgpu_device *adev = vcn_inst->adev;
+	unsigned int fences = 0, fence[AMDGPU_MAX_VCN_INSTANCES] = {0};
+	unsigned int i = vcn_inst->inst, j;
 
-			adev->vcn.pause_dpg_mode(adev, j, &new_state);
-		}
+	if (adev->vcn.harvest_config & (1 << i))
+		return;
+
+	for (j = 0; j < adev->vcn.inst[i].num_enc_rings; ++j)
+		fence[i] += amdgpu_fence_count_emitted(&vcn_inst->ring_enc[j]);
+
+	/* Only set DPG pause for VCN3 or below, VCN4 and above will be handled by FW */
+	if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG &&
+	    !adev->vcn.inst[i].using_unified_queue) {
+		struct dpg_pause_state new_state;
 
-		fence[j] += amdgpu_fence_count_emitted(&adev->vcn.inst[j].ring_dec);
-		fences += fence[j];
+		if (fence[i] ||
+		    unlikely(atomic_read(&vcn_inst->dpg_enc_submission_cnt)))
+			new_state.fw_based = VCN_DPG_STATE__PAUSE;
+		else
+			new_state.fw_based = VCN_DPG_STATE__UNPAUSE;
+
+		adev->vcn.inst[i].pause_dpg_mode(vcn_inst, &new_state);
 	}
 
-	if (!fences && !atomic_read(&adev->vcn.total_submission_cnt)) {
-		amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_VCN,
-		       AMD_PG_STATE_GATE);
-		r = amdgpu_dpm_switch_power_profile(adev, PP_SMC_POWER_PROFILE_VIDEO,
-				false);
-		if (r)
-			dev_warn(adev->dev, "(%d) failed to disable video power profile mode\n", r);
+	fence[i] += amdgpu_fence_count_emitted(&vcn_inst->ring_dec);
+	fences += fence[i];
+
+	if (!fences && !atomic_read(&vcn_inst->total_submission_cnt)) {
+		mutex_lock(&vcn_inst->vcn_pg_lock);
+		vcn_inst->set_pg_state(vcn_inst, AMD_PG_STATE_GATE);
+		mutex_unlock(&vcn_inst->vcn_pg_lock);
+		amdgpu_vcn_put_profile(adev);
+
 	} else {
-		schedule_delayed_work(&adev->vcn.idle_work, VCN_IDLE_TIMEOUT);
+		schedule_delayed_work(&vcn_inst->idle_work, VCN_IDLE_TIMEOUT);
 	}
 }
 
 void amdgpu_vcn_ring_begin_use(struct amdgpu_ring *ring)
 {
 	struct amdgpu_device *adev = ring->adev;
-	int r = 0;
+	struct amdgpu_vcn_inst *vcn_inst = &adev->vcn.inst[ring->me];
 
-	atomic_inc(&adev->vcn.total_submission_cnt);
+	atomic_inc(&vcn_inst->total_submission_cnt);
 
-	if (!cancel_delayed_work_sync(&adev->vcn.idle_work)) {
-		r = amdgpu_dpm_switch_power_profile(adev, PP_SMC_POWER_PROFILE_VIDEO,
-				true);
-		if (r)
-			dev_warn(adev->dev, "(%d) failed to switch to video power profile mode\n", r);
-	}
+	cancel_delayed_work_sync(&vcn_inst->idle_work);
 
-	mutex_lock(&adev->vcn.vcn_pg_lock);
-	amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_VCN,
-	       AMD_PG_STATE_UNGATE);
+	mutex_lock(&vcn_inst->vcn_pg_lock);
+	vcn_inst->set_pg_state(vcn_inst, AMD_PG_STATE_UNGATE);
 
-	if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)	{
+	/* Only set DPG pause for VCN3 or below, VCN4 and above will be handled by FW */
+	if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG &&
+	    !vcn_inst->using_unified_queue) {
 		struct dpg_pause_state new_state;
 
 		if (ring->funcs->type == AMDGPU_RING_TYPE_VCN_ENC) {
-			atomic_inc(&adev->vcn.inst[ring->me].dpg_enc_submission_cnt);
+			atomic_inc(&vcn_inst->dpg_enc_submission_cnt);
 			new_state.fw_based = VCN_DPG_STATE__PAUSE;
 		} else {
 			unsigned int fences = 0;
 			unsigned int i;
 
-			for (i = 0; i < adev->vcn.num_enc_rings; ++i)
-				fences += amdgpu_fence_count_emitted(&adev->vcn.inst[ring->me].ring_enc[i]);
+			for (i = 0; i < vcn_inst->num_enc_rings; ++i)
+				fences += amdgpu_fence_count_emitted(&vcn_inst->ring_enc[i]);
 
-			if (fences || atomic_read(&adev->vcn.inst[ring->me].dpg_enc_submission_cnt))
+			if (fences || atomic_read(&vcn_inst->dpg_enc_submission_cnt))
 				new_state.fw_based = VCN_DPG_STATE__PAUSE;
 			else
 				new_state.fw_based = VCN_DPG_STATE__UNPAUSE;
 		}
 
-		adev->vcn.pause_dpg_mode(adev, ring->me, &new_state);
+		vcn_inst->pause_dpg_mode(vcn_inst, &new_state);
 	}
-	mutex_unlock(&adev->vcn.vcn_pg_lock);
+	mutex_unlock(&vcn_inst->vcn_pg_lock);
+	amdgpu_vcn_get_profile(adev);
 }
 
 void amdgpu_vcn_ring_end_use(struct amdgpu_ring *ring)
 {
+	struct amdgpu_device *adev = ring->adev;
+
+	/* Only set DPG pause for VCN3 or below, VCN4 and above will be handled by FW */
 	if (ring->adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG &&
-		ring->funcs->type == AMDGPU_RING_TYPE_VCN_ENC)
+	    ring->funcs->type == AMDGPU_RING_TYPE_VCN_ENC &&
+	    !adev->vcn.inst[ring->me].using_unified_queue)
 		atomic_dec(&ring->adev->vcn.inst[ring->me].dpg_enc_submission_cnt);
 
-	atomic_dec(&ring->adev->vcn.total_submission_cnt);
+	atomic_dec(&ring->adev->vcn.inst[ring->me].total_submission_cnt);
 
-	schedule_delayed_work(&ring->adev->vcn.idle_work, VCN_IDLE_TIMEOUT);
+	schedule_delayed_work(&ring->adev->vcn.inst[ring->me].idle_work,
+			      VCN_IDLE_TIMEOUT);
 }
 
 int amdgpu_vcn_dec_ring_test_ring(struct amdgpu_ring *ring)
 {
 	struct amdgpu_device *adev = ring->adev;
 	uint32_t tmp = 0;
-	unsigned i;
+	unsigned int i;
 	int r;
 
 	/* VCN in SRIOV does not support direct register read/write */
@@ -490,7 +566,7 @@ int amdgpu_vcn_dec_ring_test_ring(struct amdgpu_ring *ring)
 	r = amdgpu_ring_alloc(ring, 3);
 	if (r)
 		return r;
-	amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.scratch9, 0));
+	amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.scratch9, 0));
 	amdgpu_ring_write(ring, 0xDEADBEEF);
 	amdgpu_ring_commit(ring);
 	for (i = 0; i < adev->usec_timeout; i++) {
@@ -541,27 +617,28 @@ static int amdgpu_vcn_dec_send_msg(struct amdgpu_ring *ring,
 				   struct amdgpu_ib *ib_msg,
 				   struct dma_fence **fence)
 {
+	u64 addr = AMDGPU_GPU_PAGE_ALIGN(ib_msg->gpu_addr);
 	struct amdgpu_device *adev = ring->adev;
 	struct dma_fence *f = NULL;
 	struct amdgpu_job *job;
 	struct amdgpu_ib *ib;
-	uint64_t addr = AMDGPU_GPU_PAGE_ALIGN(ib_msg->gpu_addr);
 	int i, r;
 
-	r = amdgpu_job_alloc_with_ib(adev, 64,
-					AMDGPU_IB_POOL_DIRECT, &job);
+	r = amdgpu_job_alloc_with_ib(ring->adev, NULL, NULL,
+				     64, AMDGPU_IB_POOL_DIRECT,
+				     &job, AMDGPU_KERNEL_JOB_ID_VCN_RING_TEST);
 	if (r)
 		goto err;
 
 	ib = &job->ibs[0];
-	ib->ptr[0] = PACKET0(adev->vcn.internal.data0, 0);
+	ib->ptr[0] = PACKET0(adev->vcn.inst[ring->me].internal.data0, 0);
 	ib->ptr[1] = addr;
-	ib->ptr[2] = PACKET0(adev->vcn.internal.data1, 0);
+	ib->ptr[2] = PACKET0(adev->vcn.inst[ring->me].internal.data1, 0);
 	ib->ptr[3] = addr >> 32;
-	ib->ptr[4] = PACKET0(adev->vcn.internal.cmd, 0);
+	ib->ptr[4] = PACKET0(adev->vcn.inst[ring->me].internal.cmd, 0);
 	ib->ptr[5] = 0;
 	for (i = 6; i < 16; i += 2) {
-		ib->ptr[i] = PACKET0(adev->vcn.internal.nop, 0);
+		ib->ptr[i] = PACKET0(adev->vcn.inst[ring->me].internal.nop, 0);
 		ib->ptr[i+1] = 0;
 	}
 	ib->length_dw = 16;
@@ -570,7 +647,7 @@ static int amdgpu_vcn_dec_send_msg(struct amdgpu_ring *ring,
 	if (r)
 		goto err_free;
 
-	amdgpu_ib_free(adev, ib_msg, f);
+	amdgpu_ib_free(ib_msg, f);
 
 	if (fence)
 		*fence = dma_fence_get(f);
@@ -581,7 +658,7 @@ static int amdgpu_vcn_dec_send_msg(struct amdgpu_ring *ring,
 err_free:
 	amdgpu_job_free(job);
 err:
-	amdgpu_ib_free(adev, ib_msg, f);
+	amdgpu_ib_free(ib_msg, f);
 	return r;
 }
 
@@ -679,27 +756,70 @@ error:
 	return r;
 }
 
+static uint32_t *amdgpu_vcn_unified_ring_ib_header(struct amdgpu_ib *ib,
+						uint32_t ib_pack_in_dw, bool enc)
+{
+	uint32_t *ib_checksum;
+
+	ib->ptr[ib->length_dw++] = 0x00000010; /* single queue checksum */
+	ib->ptr[ib->length_dw++] = 0x30000002;
+	ib_checksum = &ib->ptr[ib->length_dw++];
+	ib->ptr[ib->length_dw++] = ib_pack_in_dw;
+
+	ib->ptr[ib->length_dw++] = 0x00000010; /* engine info */
+	ib->ptr[ib->length_dw++] = 0x30000001;
+	ib->ptr[ib->length_dw++] = enc ? 0x2 : 0x3;
+	ib->ptr[ib->length_dw++] = ib_pack_in_dw * sizeof(uint32_t);
+
+	return ib_checksum;
+}
+
+static void amdgpu_vcn_unified_ring_ib_checksum(uint32_t **ib_checksum,
+						uint32_t ib_pack_in_dw)
+{
+	uint32_t i;
+	uint32_t checksum = 0;
+
+	for (i = 0; i < ib_pack_in_dw; i++)
+		checksum += *(*ib_checksum + 2 + i);
+
+	**ib_checksum = checksum;
+}
+
 static int amdgpu_vcn_dec_sw_send_msg(struct amdgpu_ring *ring,
 				      struct amdgpu_ib *ib_msg,
 				      struct dma_fence **fence)
 {
 	struct amdgpu_vcn_decode_buffer *decode_buffer = NULL;
-	const unsigned int ib_size_dw = 64;
+	unsigned int ib_size_dw = 64;
 	struct amdgpu_device *adev = ring->adev;
 	struct dma_fence *f = NULL;
 	struct amdgpu_job *job;
 	struct amdgpu_ib *ib;
 	uint64_t addr = AMDGPU_GPU_PAGE_ALIGN(ib_msg->gpu_addr);
+	uint32_t *ib_checksum;
+	uint32_t ib_pack_in_dw;
 	int i, r;
 
-	r = amdgpu_job_alloc_with_ib(adev, ib_size_dw * 4,
-				AMDGPU_IB_POOL_DIRECT, &job);
+	if (adev->vcn.inst[ring->me].using_unified_queue)
+		ib_size_dw += 8;
+
+	r = amdgpu_job_alloc_with_ib(ring->adev, NULL, NULL,
+				     ib_size_dw * 4, AMDGPU_IB_POOL_DIRECT,
+				     &job, AMDGPU_KERNEL_JOB_ID_VCN_RING_TEST);
 	if (r)
 		goto err;
 
 	ib = &job->ibs[0];
 	ib->length_dw = 0;
 
+	/* single queue headers */
+	if (adev->vcn.inst[ring->me].using_unified_queue) {
+		ib_pack_in_dw = sizeof(struct amdgpu_vcn_decode_buffer) / sizeof(uint32_t)
+						+ 4 + 2; /* engine info + decoding ib in dw */
+		ib_checksum = amdgpu_vcn_unified_ring_ib_header(ib, ib_pack_in_dw, false);
+	}
+
 	ib->ptr[ib->length_dw++] = sizeof(struct amdgpu_vcn_decode_buffer) + 8;
 	ib->ptr[ib->length_dw++] = cpu_to_le32(AMDGPU_VCN_IB_FLAG_DECODE_BUFFER);
 	decode_buffer = (struct amdgpu_vcn_decode_buffer *)&(ib->ptr[ib->length_dw]);
@@ -713,11 +833,14 @@ static int amdgpu_vcn_dec_sw_send_msg(struct amdgpu_ring *ring,
 	for (i = ib->length_dw; i < ib_size_dw; ++i)
 		ib->ptr[i] = 0x0;
 
+	if (adev->vcn.inst[ring->me].using_unified_queue)
+		amdgpu_vcn_unified_ring_ib_checksum(&ib_checksum, ib_pack_in_dw);
+
 	r = amdgpu_job_submit_direct(job, ring, &f);
 	if (r)
 		goto err_free;
 
-	amdgpu_ib_free(adev, ib_msg, f);
+	amdgpu_ib_free(ib_msg, f);
 
 	if (fence)
 		*fence = dma_fence_get(f);
@@ -728,7 +851,7 @@ static int amdgpu_vcn_dec_sw_send_msg(struct amdgpu_ring *ring,
 err_free:
 	amdgpu_job_free(job);
 err:
-	amdgpu_ib_free(adev, ib_msg, f);
+	amdgpu_ib_free(ib_msg, f);
 	return r;
 }
 
@@ -768,7 +891,7 @@ int amdgpu_vcn_enc_ring_test_ring(struct amdgpu_ring *ring)
 {
 	struct amdgpu_device *adev = ring->adev;
 	uint32_t rptr;
-	unsigned i;
+	unsigned int i;
 	int r;
 
 	if (amdgpu_sriov_vf(adev))
@@ -799,15 +922,21 @@ static int amdgpu_vcn_enc_get_create_msg(struct amdgpu_ring *ring, uint32_t hand
 					 struct amdgpu_ib *ib_msg,
 					 struct dma_fence **fence)
 {
-	const unsigned ib_size_dw = 16;
+	unsigned int ib_size_dw = 16;
+	struct amdgpu_device *adev = ring->adev;
 	struct amdgpu_job *job;
 	struct amdgpu_ib *ib;
 	struct dma_fence *f = NULL;
+	uint32_t *ib_checksum = NULL;
 	uint64_t addr;
 	int i, r;
 
-	r = amdgpu_job_alloc_with_ib(ring->adev, ib_size_dw * 4,
-					AMDGPU_IB_POOL_DIRECT, &job);
+	if (adev->vcn.inst[ring->me].using_unified_queue)
+		ib_size_dw += 8;
+
+	r = amdgpu_job_alloc_with_ib(ring->adev, NULL, NULL,
+				     ib_size_dw * 4, AMDGPU_IB_POOL_DIRECT,
+				     &job, AMDGPU_KERNEL_JOB_ID_VCN_RING_TEST);
 	if (r)
 		return r;
 
@@ -815,12 +944,16 @@ static int amdgpu_vcn_enc_get_create_msg(struct amdgpu_ring *ring, uint32_t hand
 	addr = AMDGPU_GPU_PAGE_ALIGN(ib_msg->gpu_addr);
 
 	ib->length_dw = 0;
+
+	if (adev->vcn.inst[ring->me].using_unified_queue)
+		ib_checksum = amdgpu_vcn_unified_ring_ib_header(ib, 0x11, true);
+
 	ib->ptr[ib->length_dw++] = 0x00000018;
 	ib->ptr[ib->length_dw++] = 0x00000001; /* session info */
 	ib->ptr[ib->length_dw++] = handle;
 	ib->ptr[ib->length_dw++] = upper_32_bits(addr);
 	ib->ptr[ib->length_dw++] = addr;
-	ib->ptr[ib->length_dw++] = 0x0000000b;
+	ib->ptr[ib->length_dw++] = 0x00000000;
 
 	ib->ptr[ib->length_dw++] = 0x00000014;
 	ib->ptr[ib->length_dw++] = 0x00000002; /* task info */
@@ -834,6 +967,9 @@ static int amdgpu_vcn_enc_get_create_msg(struct amdgpu_ring *ring, uint32_t hand
 	for (i = ib->length_dw; i < ib_size_dw; ++i)
 		ib->ptr[i] = 0x0;
 
+	if (adev->vcn.inst[ring->me].using_unified_queue)
+		amdgpu_vcn_unified_ring_ib_checksum(&ib_checksum, 0x11);
+
 	r = amdgpu_job_submit_direct(job, ring, &f);
 	if (r)
 		goto err;
@@ -853,15 +989,21 @@ static int amdgpu_vcn_enc_get_destroy_msg(struct amdgpu_ring *ring, uint32_t han
 					  struct amdgpu_ib *ib_msg,
 					  struct dma_fence **fence)
 {
-	const unsigned ib_size_dw = 16;
+	unsigned int ib_size_dw = 16;
+	struct amdgpu_device *adev = ring->adev;
 	struct amdgpu_job *job;
 	struct amdgpu_ib *ib;
 	struct dma_fence *f = NULL;
+	uint32_t *ib_checksum = NULL;
 	uint64_t addr;
 	int i, r;
 
-	r = amdgpu_job_alloc_with_ib(ring->adev, ib_size_dw * 4,
-					AMDGPU_IB_POOL_DIRECT, &job);
+	if (adev->vcn.inst[ring->me].using_unified_queue)
+		ib_size_dw += 8;
+
+	r = amdgpu_job_alloc_with_ib(ring->adev, NULL, NULL,
+				     ib_size_dw * 4, AMDGPU_IB_POOL_DIRECT,
+				     &job, AMDGPU_KERNEL_JOB_ID_VCN_RING_TEST);
 	if (r)
 		return r;
 
@@ -869,12 +1011,16 @@ static int amdgpu_vcn_enc_get_destroy_msg(struct amdgpu_ring *ring, uint32_t han
 	addr = AMDGPU_GPU_PAGE_ALIGN(ib_msg->gpu_addr);
 
 	ib->length_dw = 0;
+
+	if (adev->vcn.inst[ring->me].using_unified_queue)
+		ib_checksum = amdgpu_vcn_unified_ring_ib_header(ib, 0x11, true);
+
 	ib->ptr[ib->length_dw++] = 0x00000018;
 	ib->ptr[ib->length_dw++] = 0x00000001;
 	ib->ptr[ib->length_dw++] = handle;
 	ib->ptr[ib->length_dw++] = upper_32_bits(addr);
 	ib->ptr[ib->length_dw++] = addr;
-	ib->ptr[ib->length_dw++] = 0x0000000b;
+	ib->ptr[ib->length_dw++] = 0x00000000;
 
 	ib->ptr[ib->length_dw++] = 0x00000014;
 	ib->ptr[ib->length_dw++] = 0x00000002;
@@ -888,6 +1034,9 @@ static int amdgpu_vcn_enc_get_destroy_msg(struct amdgpu_ring *ring, uint32_t han
 	for (i = ib->length_dw; i < ib_size_dw; ++i)
 		ib->ptr[i] = 0x0;
 
+	if (adev->vcn.inst[ring->me].using_unified_queue)
+		amdgpu_vcn_unified_ring_ib_checksum(&ib_checksum, 0x11);
+
 	r = amdgpu_job_submit_direct(job, ring, &f);
 	if (r)
 		goto err;
@@ -932,15 +1081,33 @@ int amdgpu_vcn_enc_ring_test_ib(struct amdgpu_ring *ring, long timeout)
 		r = 0;
 
 error:
-	amdgpu_ib_free(adev, &ib, fence);
+	amdgpu_ib_free(&ib, fence);
 	dma_fence_put(fence);
 
 	return r;
 }
 
+int amdgpu_vcn_unified_ring_test_ib(struct amdgpu_ring *ring, long timeout)
+{
+	struct amdgpu_device *adev = ring->adev;
+	long r;
+
+	if ((amdgpu_ip_version(adev, UVD_HWIP, 0) != IP_VERSION(4, 0, 3)) &&
+	    (amdgpu_ip_version(adev, UVD_HWIP, 0) != IP_VERSION(5, 0, 1))) {
+		r = amdgpu_vcn_enc_ring_test_ib(ring, timeout);
+		if (r)
+			goto error;
+	}
+
+	r =  amdgpu_vcn_dec_sw_ring_test_ib(ring, timeout);
+
+error:
+	return r;
+}
+
 enum amdgpu_ring_priority_level amdgpu_vcn_get_enc_ring_prio(int ring)
 {
-	switch(ring) {
+	switch (ring) {
 	case 0:
 		return AMDGPU_RING_PRIO_0;
 	case 1:
@@ -952,29 +1119,519 @@ enum amdgpu_ring_priority_level amdgpu_vcn_get_enc_ring_prio(int ring)
 	}
 }
 
-void amdgpu_vcn_setup_ucode(struct amdgpu_device *adev)
+void amdgpu_vcn_setup_ucode(struct amdgpu_device *adev, int i)
 {
-	int i;
 	unsigned int idx;
 
 	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
 		const struct common_firmware_header *hdr;
-		hdr = (const struct common_firmware_header *)adev->vcn.fw->data;
 
+		if (adev->vcn.harvest_config & (1 << i))
+			return;
+
+		if ((amdgpu_ip_version(adev, UVD_HWIP, 0) == IP_VERSION(4, 0, 3) ||
+		     amdgpu_ip_version(adev, UVD_HWIP, 0) == IP_VERSION(5, 0, 1))
+		    && (i > 0))
+			return;
+
+		hdr = (const struct common_firmware_header *)adev->vcn.inst[i].fw->data;
+		/* currently only support 2 FW instances */
+		if (i >= 2) {
+			dev_info(adev->dev, "More then 2 VCN FW instances!\n");
+			return;
+		}
+		idx = AMDGPU_UCODE_ID_VCN + i;
+		adev->firmware.ucode[idx].ucode_id = idx;
+		adev->firmware.ucode[idx].fw = adev->vcn.inst[i].fw;
+		adev->firmware.fw_size +=
+			ALIGN(le32_to_cpu(hdr->ucode_size_bytes), PAGE_SIZE);
+	}
+}
+
+/*
+ * debugfs for mapping vcn firmware log buffer.
+ */
+#if defined(CONFIG_DEBUG_FS)
+static ssize_t amdgpu_debugfs_vcn_fwlog_read(struct file *f, char __user *buf,
+					     size_t size, loff_t *pos)
+{
+	struct amdgpu_vcn_inst *vcn;
+	void *log_buf;
+	struct amdgpu_vcn_fwlog *plog;
+	unsigned int read_pos, write_pos, available, i, read_bytes = 0;
+	unsigned int read_num[2] = {0};
+
+	vcn = file_inode(f)->i_private;
+	if (!vcn)
+		return -ENODEV;
+
+	if (!vcn->fw_shared.cpu_addr || !amdgpu_vcnfw_log)
+		return -EFAULT;
+
+	log_buf = vcn->fw_shared.cpu_addr + vcn->fw_shared.mem_size;
+
+	plog = (struct amdgpu_vcn_fwlog *)log_buf;
+	read_pos = plog->rptr;
+	write_pos = plog->wptr;
+
+	if (read_pos > AMDGPU_VCNFW_LOG_SIZE || write_pos > AMDGPU_VCNFW_LOG_SIZE)
+		return -EFAULT;
+
+	if (!size || (read_pos == write_pos))
+		return 0;
+
+	if (write_pos > read_pos) {
+		available = write_pos - read_pos;
+		read_num[0] = min_t(size_t, size, available);
+	} else {
+		read_num[0] = AMDGPU_VCNFW_LOG_SIZE - read_pos;
+		available = read_num[0] + write_pos - plog->header_size;
+		if (size > available)
+			read_num[1] = write_pos - plog->header_size;
+		else if (size > read_num[0])
+			read_num[1] = size - read_num[0];
+		else
+			read_num[0] = size;
+	}
+
+	for (i = 0; i < 2; i++) {
+		if (read_num[i]) {
+			if (read_pos == AMDGPU_VCNFW_LOG_SIZE)
+				read_pos = plog->header_size;
+			if (read_num[i] == copy_to_user((buf + read_bytes),
+							(log_buf + read_pos), read_num[i]))
+				return -EFAULT;
+
+			read_bytes += read_num[i];
+			read_pos += read_num[i];
+		}
+	}
+
+	plog->rptr = read_pos;
+	*pos += read_bytes;
+	return read_bytes;
+}
+
+static const struct file_operations amdgpu_debugfs_vcnfwlog_fops = {
+	.owner = THIS_MODULE,
+	.read = amdgpu_debugfs_vcn_fwlog_read,
+	.llseek = default_llseek
+};
+#endif
+
+void amdgpu_debugfs_vcn_fwlog_init(struct amdgpu_device *adev, uint8_t i,
+				   struct amdgpu_vcn_inst *vcn)
+{
+#if defined(CONFIG_DEBUG_FS)
+	struct drm_minor *minor = adev_to_drm(adev)->primary;
+	struct dentry *root = minor->debugfs_root;
+	char name[32];
+
+	sprintf(name, "amdgpu_vcn_%d_fwlog", i);
+	debugfs_create_file_size(name, S_IFREG | 0444, root, vcn,
+				 &amdgpu_debugfs_vcnfwlog_fops,
+				 AMDGPU_VCNFW_LOG_SIZE);
+#endif
+}
+
+void amdgpu_vcn_fwlog_init(struct amdgpu_vcn_inst *vcn)
+{
+#if defined(CONFIG_DEBUG_FS)
+	uint32_t *flag = vcn->fw_shared.cpu_addr;
+	void *fw_log_cpu_addr = vcn->fw_shared.cpu_addr + vcn->fw_shared.mem_size;
+	uint64_t fw_log_gpu_addr = vcn->fw_shared.gpu_addr + vcn->fw_shared.mem_size;
+	struct amdgpu_vcn_fwlog *log_buf = fw_log_cpu_addr;
+	struct amdgpu_fw_shared_fw_logging *fw_log = vcn->fw_shared.cpu_addr
+							 + vcn->fw_shared.log_offset;
+	*flag |= cpu_to_le32(AMDGPU_VCN_FW_LOGGING_FLAG);
+	fw_log->is_enabled = 1;
+	fw_log->addr_lo = cpu_to_le32(fw_log_gpu_addr & 0xFFFFFFFF);
+	fw_log->addr_hi = cpu_to_le32(fw_log_gpu_addr >> 32);
+	fw_log->size = cpu_to_le32(AMDGPU_VCNFW_LOG_SIZE);
+
+	log_buf->header_size = sizeof(struct amdgpu_vcn_fwlog);
+	log_buf->buffer_size = AMDGPU_VCNFW_LOG_SIZE;
+	log_buf->rptr = log_buf->header_size;
+	log_buf->wptr = log_buf->header_size;
+	log_buf->wrapped = 0;
+#endif
+}
+
+int amdgpu_vcn_process_poison_irq(struct amdgpu_device *adev,
+				struct amdgpu_irq_src *source,
+				struct amdgpu_iv_entry *entry)
+{
+	struct ras_common_if *ras_if = adev->vcn.ras_if;
+	struct ras_dispatch_if ih_data = {
+		.entry = entry,
+	};
+
+	if (!ras_if)
+		return 0;
+
+	if (!amdgpu_sriov_vf(adev)) {
+		ih_data.head = *ras_if;
+		amdgpu_ras_interrupt_dispatch(adev, &ih_data);
+	} else {
+		if (adev->virt.ops && adev->virt.ops->ras_poison_handler)
+			adev->virt.ops->ras_poison_handler(adev, ras_if->block);
+		else
+			dev_warn(adev->dev,
+				"No ras_poison_handler interface in SRIOV for VCN!\n");
+	}
+
+	return 0;
+}
+
+int amdgpu_vcn_ras_late_init(struct amdgpu_device *adev, struct ras_common_if *ras_block)
+{
+	int r, i;
+
+	r = amdgpu_ras_block_late_init(adev, ras_block);
+	if (r)
+		return r;
+
+	if (amdgpu_ras_is_supported(adev, ras_block->block)) {
 		for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
-			if (adev->vcn.harvest_config & (1 << i))
+			if (adev->vcn.harvest_config & (1 << i) ||
+			    !adev->vcn.inst[i].ras_poison_irq.funcs)
 				continue;
-			/* currently only support 2 FW instances */
-			if (i >= 2) {
-				dev_info(adev->dev, "More then 2 VCN FW instances!\n");
-				break;
-			}
-			idx = AMDGPU_UCODE_ID_VCN + i;
-			adev->firmware.ucode[idx].ucode_id = idx;
-			adev->firmware.ucode[idx].fw = adev->vcn.fw;
-			adev->firmware.fw_size +=
-				ALIGN(le32_to_cpu(hdr->ucode_size_bytes), PAGE_SIZE);
+
+			r = amdgpu_irq_get(adev, &adev->vcn.inst[i].ras_poison_irq, 0);
+			if (r)
+				goto late_fini;
+		}
+	}
+	return 0;
+
+late_fini:
+	amdgpu_ras_block_late_fini(adev, ras_block);
+	return r;
+}
+
+int amdgpu_vcn_ras_sw_init(struct amdgpu_device *adev)
+{
+	int err;
+	struct amdgpu_vcn_ras *ras;
+
+	if (!adev->vcn.ras)
+		return 0;
+
+	ras = adev->vcn.ras;
+	err = amdgpu_ras_register_ras_block(adev, &ras->ras_block);
+	if (err) {
+		dev_err(adev->dev, "Failed to register vcn ras block!\n");
+		return err;
+	}
+
+	strcpy(ras->ras_block.ras_comm.name, "vcn");
+	ras->ras_block.ras_comm.block = AMDGPU_RAS_BLOCK__VCN;
+	ras->ras_block.ras_comm.type = AMDGPU_RAS_ERROR__POISON;
+	adev->vcn.ras_if = &ras->ras_block.ras_comm;
+
+	if (!ras->ras_block.ras_late_init)
+		ras->ras_block.ras_late_init = amdgpu_vcn_ras_late_init;
+
+	return 0;
+}
+
+int amdgpu_vcn_psp_update_sram(struct amdgpu_device *adev, int inst_idx,
+			       enum AMDGPU_UCODE_ID ucode_id)
+{
+	struct amdgpu_firmware_info ucode = {
+		.ucode_id = (ucode_id ? ucode_id :
+			    (inst_idx ? AMDGPU_UCODE_ID_VCN1_RAM :
+					AMDGPU_UCODE_ID_VCN0_RAM)),
+		.mc_addr = adev->vcn.inst[inst_idx].dpg_sram_gpu_addr,
+		.ucode_size = ((uintptr_t)adev->vcn.inst[inst_idx].dpg_sram_curr_addr -
+			      (uintptr_t)adev->vcn.inst[inst_idx].dpg_sram_cpu_addr),
+	};
+
+	return psp_execute_ip_fw_load(&adev->psp, &ucode);
+}
+
+static ssize_t amdgpu_get_vcn_reset_mask(struct device *dev,
+						struct device_attribute *attr,
+						char *buf)
+{
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = drm_to_adev(ddev);
+
+	if (!adev)
+		return -ENODEV;
+
+	return amdgpu_show_reset_mask(buf, adev->vcn.supported_reset);
+}
+
+static DEVICE_ATTR(vcn_reset_mask, 0444,
+		   amdgpu_get_vcn_reset_mask, NULL);
+
+int amdgpu_vcn_sysfs_reset_mask_init(struct amdgpu_device *adev)
+{
+	int r = 0;
+
+	if (adev->vcn.num_vcn_inst) {
+		r = device_create_file(adev->dev, &dev_attr_vcn_reset_mask);
+		if (r)
+			return r;
+	}
+
+	return r;
+}
+
+void amdgpu_vcn_sysfs_reset_mask_fini(struct amdgpu_device *adev)
+{
+	if (adev->dev->kobj.sd) {
+		if (adev->vcn.num_vcn_inst)
+			device_remove_file(adev->dev, &dev_attr_vcn_reset_mask);
+	}
+}
+
+/*
+ * debugfs to enable/disable vcn job submission to specific core or
+ * instance. It is created only if the queue type is unified.
+ */
+#if defined(CONFIG_DEBUG_FS)
+static int amdgpu_debugfs_vcn_sched_mask_set(void *data, u64 val)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)data;
+	u32 i;
+	u64 mask;
+	struct amdgpu_ring *ring;
+
+	if (!adev)
+		return -ENODEV;
+
+	mask = (1ULL << adev->vcn.num_vcn_inst) - 1;
+	if ((val & mask) == 0)
+		return -EINVAL;
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		ring = &adev->vcn.inst[i].ring_enc[0];
+		if (val & (1ULL << i))
+			ring->sched.ready = true;
+		else
+			ring->sched.ready = false;
+	}
+	/* publish sched.ready flag update effective immediately across smp */
+	smp_rmb();
+	return 0;
+}
+
+static int amdgpu_debugfs_vcn_sched_mask_get(void *data, u64 *val)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)data;
+	u32 i;
+	u64 mask = 0;
+	struct amdgpu_ring *ring;
+
+	if (!adev)
+		return -ENODEV;
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		ring = &adev->vcn.inst[i].ring_enc[0];
+		if (ring->sched.ready)
+			mask |= 1ULL << i;
+		}
+	*val = mask;
+	return 0;
+}
+
+DEFINE_DEBUGFS_ATTRIBUTE(amdgpu_debugfs_vcn_sched_mask_fops,
+			 amdgpu_debugfs_vcn_sched_mask_get,
+			 amdgpu_debugfs_vcn_sched_mask_set, "%llx\n");
+#endif
+
+void amdgpu_debugfs_vcn_sched_mask_init(struct amdgpu_device *adev)
+{
+#if defined(CONFIG_DEBUG_FS)
+	struct drm_minor *minor = adev_to_drm(adev)->primary;
+	struct dentry *root = minor->debugfs_root;
+	char name[32];
+
+	if (adev->vcn.num_vcn_inst <= 1 || !adev->vcn.inst[0].using_unified_queue)
+		return;
+	sprintf(name, "amdgpu_vcn_sched_mask");
+	debugfs_create_file(name, 0600, root, adev,
+			    &amdgpu_debugfs_vcn_sched_mask_fops);
+#endif
+}
+
+/**
+ * vcn_set_powergating_state - set VCN block powergating state
+ *
+ * @ip_block: amdgpu_ip_block pointer
+ * @state: power gating state
+ *
+ * Set VCN block powergating state
+ */
+int vcn_set_powergating_state(struct amdgpu_ip_block *ip_block,
+			      enum amd_powergating_state state)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int ret = 0, i;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		struct amdgpu_vcn_inst *vinst = &adev->vcn.inst[i];
+
+		ret |= vinst->set_pg_state(vinst, state);
+	}
+
+	return ret;
+}
+
+/**
+ * amdgpu_vcn_reset_engine - Reset a specific VCN engine
+ * @adev: Pointer to the AMDGPU device
+ * @instance_id: VCN engine instance to reset
+ *
+ * Returns: 0 on success, or a negative error code on failure.
+ */
+static int amdgpu_vcn_reset_engine(struct amdgpu_device *adev,
+				   uint32_t instance_id)
+{
+	struct amdgpu_vcn_inst *vinst = &adev->vcn.inst[instance_id];
+	int r, i;
+
+	mutex_lock(&vinst->engine_reset_mutex);
+	/* Stop the scheduler's work queue for the dec and enc rings if they are running.
+	 * This ensures that no new tasks are submitted to the queues while
+	 * the reset is in progress.
+	 */
+	drm_sched_wqueue_stop(&vinst->ring_dec.sched);
+	for (i = 0; i < vinst->num_enc_rings; i++)
+		drm_sched_wqueue_stop(&vinst->ring_enc[i].sched);
+
+	/* Perform the VCN reset for the specified instance */
+	r = vinst->reset(vinst);
+	if (r)
+		goto unlock;
+	r = amdgpu_ring_test_ring(&vinst->ring_dec);
+	if (r)
+		goto unlock;
+	for (i = 0; i < vinst->num_enc_rings; i++) {
+		r = amdgpu_ring_test_ring(&vinst->ring_enc[i]);
+		if (r)
+			goto unlock;
+	}
+	amdgpu_fence_driver_force_completion(&vinst->ring_dec);
+	for (i = 0; i < vinst->num_enc_rings; i++)
+		amdgpu_fence_driver_force_completion(&vinst->ring_enc[i]);
+
+	/* Restart the scheduler's work queue for the dec and enc rings
+	 * if they were stopped by this function. This allows new tasks
+	 * to be submitted to the queues after the reset is complete.
+	 */
+	drm_sched_wqueue_start(&vinst->ring_dec.sched);
+	for (i = 0; i < vinst->num_enc_rings; i++)
+		drm_sched_wqueue_start(&vinst->ring_enc[i].sched);
+
+unlock:
+	mutex_unlock(&vinst->engine_reset_mutex);
+
+	return r;
+}
+
+/**
+ * amdgpu_vcn_ring_reset - Reset a VCN ring
+ * @ring: ring to reset
+ * @vmid: vmid of guilty job
+ * @timedout_fence: fence of timed out job
+ *
+ * This helper is for VCN blocks without unified queues because
+ * resetting the engine resets all queues in that case.  With
+ * unified queues we have one queue per engine.
+ * Returns: 0 on success, or a negative error code on failure.
+ */
+int amdgpu_vcn_ring_reset(struct amdgpu_ring *ring,
+			  unsigned int vmid,
+			  struct amdgpu_fence *timedout_fence)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	if (adev->vcn.inst[ring->me].using_unified_queue)
+		return -EINVAL;
+
+	return amdgpu_vcn_reset_engine(adev, ring->me);
+}
+
+int amdgpu_vcn_reg_dump_init(struct amdgpu_device *adev,
+			     const struct amdgpu_hwip_reg_entry *reg, u32 count)
+{
+	adev->vcn.ip_dump = kcalloc(adev->vcn.num_vcn_inst * count,
+				     sizeof(uint32_t), GFP_KERNEL);
+	if (!adev->vcn.ip_dump)
+		return -ENOMEM;
+	adev->vcn.reg_list = reg;
+	adev->vcn.reg_count = count;
+
+	return 0;
+}
+
+static void amdgpu_vcn_reg_dump_fini(struct amdgpu_device *adev)
+{
+	kfree(adev->vcn.ip_dump);
+	adev->vcn.ip_dump = NULL;
+	adev->vcn.reg_list = NULL;
+	adev->vcn.reg_count = 0;
+}
+
+void amdgpu_vcn_dump_ip_state(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, j;
+	bool is_powered;
+	u32 inst_off;
+
+	if (!adev->vcn.ip_dump)
+		return;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+		if (adev->vcn.harvest_config & (1 << i))
+			continue;
+
+		inst_off = i * adev->vcn.reg_count;
+		/* mmUVD_POWER_STATUS is always readable and is the first in reg_list */
+		adev->vcn.ip_dump[inst_off] =
+			RREG32(SOC15_REG_ENTRY_OFFSET_INST(adev->vcn.reg_list[0], i));
+		is_powered = (adev->vcn.ip_dump[inst_off] &
+			      UVD_POWER_STATUS__UVD_POWER_STATUS_TILES_OFF) !=
+			      UVD_POWER_STATUS__UVD_POWER_STATUS_TILES_OFF;
+
+		if (is_powered)
+			for (j = 1; j < adev->vcn.reg_count; j++)
+				adev->vcn.ip_dump[inst_off + j] =
+				RREG32(SOC15_REG_ENTRY_OFFSET_INST(adev->vcn.reg_list[j], i));
+	}
+}
+
+void amdgpu_vcn_print_ip_state(struct amdgpu_ip_block *ip_block, struct drm_printer *p)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, j;
+	bool is_powered;
+	u32 inst_off;
+
+	if (!adev->vcn.ip_dump)
+		return;
+
+	drm_printf(p, "num_instances:%d\n", adev->vcn.num_vcn_inst);
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+		if (adev->vcn.harvest_config & (1 << i)) {
+			drm_printf(p, "\nHarvested Instance:VCN%d Skipping dump\n", i);
+			continue;
+		}
+
+		inst_off = i * adev->vcn.reg_count;
+		is_powered = (adev->vcn.ip_dump[inst_off] &
+			      UVD_POWER_STATUS__UVD_POWER_STATUS_TILES_OFF) !=
+			      UVD_POWER_STATUS__UVD_POWER_STATUS_TILES_OFF;
+
+		if (is_powered) {
+			drm_printf(p, "\nActive Instance:VCN%d\n", i);
+			for (j = 0; j < adev->vcn.reg_count; j++)
+				drm_printf(p, "%-50s \t 0x%08x\n", adev->vcn.reg_list[j].reg_name,
+					   adev->vcn.ip_dump[inst_off + j]);
+		} else {
+			drm_printf(p, "\nInactive Instance:VCN%d\n", i);
 		}
-		dev_info(adev->dev, "Will use PSP to load VCN firmware\n");
 	}
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h
index bfa27ea94804..82624b44e661 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h
@@ -1,5 +1,5 @@
 /*
- * Copyright 2016 Advanced Micro Devices, Inc.
+ * Copyright 2016-2024 Advanced Micro Devices, Inc. All rights reserved.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -24,14 +24,16 @@
 #ifndef __AMDGPU_VCN_H__
 #define __AMDGPU_VCN_H__
 
+#include "amdgpu_ras.h"
+
 #define AMDGPU_VCN_STACK_SIZE		(128*1024)
 #define AMDGPU_VCN_CONTEXT_SIZE 	(512*1024)
 
 #define AMDGPU_VCN_FIRMWARE_OFFSET	256
 #define AMDGPU_VCN_MAX_ENC_RINGS	3
 
-#define AMDGPU_MAX_VCN_INSTANCES	2
-#define AMDGPU_MAX_VCN_ENC_RINGS  AMDGPU_VCN_MAX_ENC_RINGS * AMDGPU_MAX_VCN_INSTANCES
+#define AMDGPU_MAX_VCN_INSTANCES	4
+#define AMDGPU_MAX_VCN_ENC_RINGS  (AMDGPU_VCN_MAX_ENC_RINGS * AMDGPU_MAX_VCN_INSTANCES)
 
 #define AMDGPU_VCN_HARVEST_VCN0 (1 << 0)
 #define AMDGPU_VCN_HARVEST_VCN1 (1 << 1)
@@ -63,10 +65,7 @@
 #define VCN_ENC_CMD_REG_WRITE		0x0000000b
 #define VCN_ENC_CMD_REG_WAIT		0x0000000c
 
-#define VCN_VID_SOC_ADDRESS_2_0 	0x1fa00
-#define VCN1_VID_SOC_ADDRESS_3_0 	0x48200
 #define VCN_AON_SOC_ADDRESS_2_0 	0x1f800
-#define VCN1_AON_SOC_ADDRESS_3_0 	0x48000
 #define VCN_VID_IP_ADDRESS_2_0		0x0
 #define VCN_AON_IP_ADDRESS_2_0		0x30000
 
@@ -101,7 +100,8 @@
 
 #define SOC15_DPG_MODE_OFFSET(ip, inst_idx, reg) 						\
 	({											\
-		uint32_t internal_reg_offset, addr;						\
+		/* To avoid a -Wunused-but-set-variable warning. */				\
+		uint32_t internal_reg_offset __maybe_unused, addr;				\
 		bool video_range, video1_range, aon_range, aon1_range;				\
 												\
 		addr = (adev->reg_offset[ip##_HWIP][inst_idx][reg##_BASE_IDX] + reg);		\
@@ -141,27 +141,112 @@
 		RREG32_SOC15(VCN, inst_idx, mmUVD_DPG_LMA_DATA);				\
 	})
 
-#define WREG32_SOC15_DPG_MODE(inst_idx, offset, value, mask_en, indirect)			\
-	do {											\
-		if (!indirect) {								\
-			WREG32_SOC15(VCN, inst_idx, mmUVD_DPG_LMA_DATA, value);			\
-			WREG32_SOC15(VCN, inst_idx, mmUVD_DPG_LMA_CTL, 				\
-				(0x1 << UVD_DPG_LMA_CTL__READ_WRITE__SHIFT |			\
-				 mask_en << UVD_DPG_LMA_CTL__MASK_EN__SHIFT |			\
-				 offset << UVD_DPG_LMA_CTL__READ_WRITE_ADDR__SHIFT));		\
-		} else {									\
-			*adev->vcn.inst[inst_idx].dpg_sram_curr_addr++ = offset;		\
-			*adev->vcn.inst[inst_idx].dpg_sram_curr_addr++ = value;			\
-		}										\
+#define WREG32_SOC15_DPG_MODE(inst_idx, offset, value, mask_en, indirect)             \
+	do {                                                                          \
+		if (!indirect) {                                                      \
+			WREG32_SOC15(VCN, GET_INST(VCN, inst_idx),                    \
+				     mmUVD_DPG_LMA_DATA, value);                      \
+			WREG32_SOC15(                                                 \
+				VCN, GET_INST(VCN, inst_idx),                         \
+				mmUVD_DPG_LMA_CTL,                                    \
+				(0x1 << UVD_DPG_LMA_CTL__READ_WRITE__SHIFT |          \
+				 mask_en << UVD_DPG_LMA_CTL__MASK_EN__SHIFT |         \
+				 offset << UVD_DPG_LMA_CTL__READ_WRITE_ADDR__SHIFT)); \
+		} else {                                                              \
+			*adev->vcn.inst[inst_idx].dpg_sram_curr_addr++ =              \
+				offset;                                               \
+			*adev->vcn.inst[inst_idx].dpg_sram_curr_addr++ =              \
+				value;                                                \
+		}                                                                     \
 	} while (0)
 
+#define SOC24_DPG_MODE_OFFSET(ip, inst_idx, reg)						\
+	({											\
+		/* To avoid a -Wunused-but-set-variable warning. */				\
+		uint32_t internal_reg_offset __maybe_unused, addr;				\
+		bool video_range, video1_range, aon_range, aon1_range;				\
+												\
+		addr = (adev->reg_offset[ip##_HWIP][inst_idx][reg##_BASE_IDX] + reg);		\
+		addr <<= 2;									\
+		video_range = ((((0xFFFFF & addr) >= (VCN_VID_SOC_ADDRESS)) &&			\
+				((0xFFFFF & addr) < ((VCN_VID_SOC_ADDRESS + 0x2600)))));	\
+		video1_range = ((((0xFFFFF & addr) >= (VCN1_VID_SOC_ADDRESS)) &&		\
+				((0xFFFFF & addr) < ((VCN1_VID_SOC_ADDRESS + 0x2600)))));	\
+		aon_range   = ((((0xFFFFF & addr) >= (VCN_AON_SOC_ADDRESS)) &&			\
+				((0xFFFFF & addr) < ((VCN_AON_SOC_ADDRESS + 0x600)))));		\
+		aon1_range   = ((((0xFFFFF & addr) >= (VCN1_AON_SOC_ADDRESS)) &&		\
+				((0xFFFFF & addr) < ((VCN1_AON_SOC_ADDRESS + 0x600)))));	\
+		if (video_range)								\
+			internal_reg_offset = ((0xFFFFF & addr) - (VCN_VID_SOC_ADDRESS) +	\
+				(VCN_VID_IP_ADDRESS));						\
+		else if (aon_range)								\
+			internal_reg_offset = ((0xFFFFF & addr) - (VCN_AON_SOC_ADDRESS) +	\
+				(VCN_AON_IP_ADDRESS));						\
+		else if (video1_range)								\
+			internal_reg_offset = ((0xFFFFF & addr) - (VCN1_VID_SOC_ADDRESS) +	\
+				(VCN_VID_IP_ADDRESS));						\
+		else if (aon1_range)								\
+			internal_reg_offset = ((0xFFFFF & addr) - (VCN1_AON_SOC_ADDRESS) +	\
+				(VCN_AON_IP_ADDRESS));						\
+		else										\
+			internal_reg_offset = (0xFFFFF & addr);					\
+												\
+		internal_reg_offset >>= 2;							\
+	})
+
+#define WREG32_SOC24_DPG_MODE(inst_idx, offset, value, mask_en, indirect)		\
+	do {										\
+		if (!indirect) {							\
+			WREG32_SOC15(VCN, GET_INST(VCN, inst_idx),			\
+				     regUVD_DPG_LMA_DATA, value);			\
+			WREG32_SOC15(							\
+				VCN, GET_INST(VCN, inst_idx),				\
+				regUVD_DPG_LMA_CTL,					\
+				(0x1 << UVD_DPG_LMA_CTL__READ_WRITE__SHIFT |		\
+				 mask_en << UVD_DPG_LMA_CTL__MASK_EN__SHIFT |		\
+				 offset << UVD_DPG_LMA_CTL__READ_WRITE_ADDR__SHIFT));	\
+		} else {								\
+			*adev->vcn.inst[inst_idx].dpg_sram_curr_addr++ =		\
+				offset;							\
+			*adev->vcn.inst[inst_idx].dpg_sram_curr_addr++ =		\
+				value;							\
+		}									\
+	} while (0)
+
+#define AMDGPU_FW_SHARED_FLAG_0_UNIFIED_QUEUE (1 << 2)
+#define AMDGPU_FW_SHARED_FLAG_0_DRM_KEY_INJECT (1 << 4)
 #define AMDGPU_VCN_FW_SHARED_FLAG_0_RB	(1 << 6)
 #define AMDGPU_VCN_MULTI_QUEUE_FLAG	(1 << 8)
 #define AMDGPU_VCN_SW_RING_FLAG		(1 << 9)
+#define AMDGPU_VCN_FW_LOGGING_FLAG	(1 << 10)
+#define AMDGPU_VCN_SMU_VERSION_INFO_FLAG (1 << 11)
+#define AMDGPU_VCN_SMU_DPM_INTERFACE_FLAG (1 << 11)
+#define AMDGPU_VCN_VF_RB_SETUP_FLAG (1 << 14)
+#define AMDGPU_VCN_VF_RB_DECOUPLE_FLAG (1 << 15)
+
+#define MAX_NUM_VCN_RB_SETUP 4
 
 #define AMDGPU_VCN_IB_FLAG_DECODE_BUFFER	0x00000001
 #define AMDGPU_VCN_CMD_FLAG_MSG_BUFFER		0x00000001
 
+#define VCN_CODEC_DISABLE_MASK_AV1  (1 << 0)
+#define VCN_CODEC_DISABLE_MASK_VP9  (1 << 1)
+#define VCN_CODEC_DISABLE_MASK_HEVC (1 << 2)
+#define VCN_CODEC_DISABLE_MASK_H264 (1 << 3)
+
+#define AMDGPU_VCN_SMU_DPM_INTERFACE_DGPU (0)
+#define AMDGPU_VCN_SMU_DPM_INTERFACE_APU (1)
+
+#define AMDGPU_DRM_KEY_INJECT_WORKAROUND_VCNFW_ASD_HANDSHAKING 2
+
+struct amdgpu_hwip_reg_entry;
+
+enum amdgpu_vcn_caps {
+	AMDGPU_VCN_RRMT_ENABLED,
+};
+
+#define AMDGPU_VCN_CAPS(caps) BIT(AMDGPU_VCN_##caps)
+
 enum fw_queue_mode {
 	FW_QUEUE_RING_RESET = 1,
 	FW_QUEUE_DPG_HOLD_OFF = 2,
@@ -205,7 +290,16 @@ struct amdgpu_vcn_reg{
 	unsigned	scratch9;
 };
 
+struct amdgpu_vcn_fw_shared {
+	void        *cpu_addr;
+	uint64_t    gpu_addr;
+	uint32_t    mem_size;
+	uint32_t    log_offset;
+};
+
 struct amdgpu_vcn_inst {
+	struct amdgpu_device	*adev;
+	int			inst;
 	struct amdgpu_bo	*vcpu_bo;
 	void			*cpu_addr;
 	uint64_t		gpu_addr;
@@ -214,6 +308,7 @@ struct amdgpu_vcn_inst {
 	struct amdgpu_ring	ring_enc[AMDGPU_VCN_MAX_ENC_RINGS];
 	atomic_t		sched_score;
 	struct amdgpu_irq_src	irq;
+	struct amdgpu_irq_src	ras_poison_irq;
 	struct amdgpu_vcn_reg	external;
 	struct amdgpu_bo	*dpg_sram_bo;
 	struct dpg_pause_state	pause_state;
@@ -221,28 +316,58 @@ struct amdgpu_vcn_inst {
 	uint64_t		dpg_sram_gpu_addr;
 	uint32_t		*dpg_sram_curr_addr;
 	atomic_t		dpg_enc_submission_cnt;
-	void			*fw_shared_cpu_addr;
-	uint64_t		fw_shared_gpu_addr;
-};
-
-struct amdgpu_vcn {
-	unsigned		fw_version;
+	struct amdgpu_vcn_fw_shared fw_shared;
+	uint8_t			aid_id;
+	const struct firmware	*fw; /* VCN firmware */
+	uint8_t			vcn_config;
+	uint32_t		vcn_codec_disable_mask;
+	atomic_t		total_submission_cnt;
+	struct mutex		vcn_pg_lock;
+	enum amd_powergating_state cur_state;
 	struct delayed_work	idle_work;
-	const struct firmware	*fw;	/* VCN firmware */
+	unsigned		fw_version;
 	unsigned		num_enc_rings;
-	enum amd_powergating_state cur_state;
 	bool			indirect_sram;
+	struct amdgpu_vcn_reg	 internal;
+	struct mutex		vcn1_jpeg1_workaround;
+	int (*pause_dpg_mode)(struct amdgpu_vcn_inst *vinst,
+			      struct dpg_pause_state *new_state);
+	int (*set_pg_state)(struct amdgpu_vcn_inst *vinst,
+			    enum amd_powergating_state state);
+	int (*reset)(struct amdgpu_vcn_inst *vinst);
+	bool using_unified_queue;
+	struct mutex		engine_reset_mutex;
+};
+
+struct amdgpu_vcn_ras {
+	struct amdgpu_ras_block_object ras_block;
+};
 
+struct amdgpu_vcn {
 	uint8_t	num_vcn_inst;
 	struct amdgpu_vcn_inst	 inst[AMDGPU_MAX_VCN_INSTANCES];
-	struct amdgpu_vcn_reg	 internal;
-	struct mutex		 vcn_pg_lock;
-	struct mutex		vcn1_jpeg1_workaround;
-	atomic_t		 total_submission_cnt;
 
 	unsigned	harvest_config;
-	int (*pause_dpg_mode)(struct amdgpu_device *adev,
-		int inst_idx, struct dpg_pause_state *new_state);
+
+	struct ras_common_if    *ras_if;
+	struct amdgpu_vcn_ras   *ras;
+
+	uint16_t inst_mask;
+	uint8_t	num_inst_per_aid;
+
+	/* IP reg dump */
+	uint32_t		*ip_dump;
+
+	uint32_t		supported_reset;
+	uint32_t		caps;
+
+	bool			per_inst_fw;
+	unsigned		fw_version;
+
+	bool			workload_profile_active;
+	struct mutex            workload_profile_mutex;
+	u32 reg_count;
+	const struct amdgpu_hwip_reg_entry *reg_list;
 };
 
 struct amdgpu_fw_shared_rb_ptrs_struct {
@@ -264,6 +389,25 @@ struct amdgpu_fw_shared_sw_ring {
 	uint8_t padding[3];
 };
 
+struct amdgpu_fw_shared_unified_queue_struct {
+	uint8_t is_enabled;
+	uint8_t queue_mode;
+	uint8_t queue_status;
+	uint8_t padding[5];
+};
+
+struct amdgpu_fw_shared_fw_logging {
+	uint8_t is_enabled;
+	uint32_t addr_lo;
+	uint32_t addr_hi;
+	uint32_t size;
+};
+
+struct amdgpu_fw_shared_smu_interface_info {
+	uint8_t smu_interface_type;
+	uint8_t padding[3];
+};
+
 struct amdgpu_fw_shared {
 	uint32_t present_flag_0;
 	uint8_t pad[44];
@@ -271,6 +415,66 @@ struct amdgpu_fw_shared {
 	uint8_t pad1[1];
 	struct amdgpu_fw_shared_multi_queue multi_queue;
 	struct amdgpu_fw_shared_sw_ring sw_ring;
+	struct amdgpu_fw_shared_fw_logging fw_log;
+	struct amdgpu_fw_shared_smu_interface_info smu_interface_info;
+};
+
+struct amdgpu_vcn_rb_setup_info {
+	uint32_t  rb_addr_lo;
+	uint32_t  rb_addr_hi;
+	uint32_t  rb_size;
+};
+
+struct amdgpu_fw_shared_rb_setup {
+	uint32_t is_rb_enabled_flags;
+
+	union {
+		struct {
+			uint32_t rb_addr_lo;
+			uint32_t rb_addr_hi;
+			uint32_t  rb_size;
+			uint32_t  rb4_addr_lo;
+			uint32_t  rb4_addr_hi;
+			uint32_t  rb4_size;
+			uint32_t  reserved[6];
+		};
+
+		struct {
+			struct amdgpu_vcn_rb_setup_info rb_info[MAX_NUM_VCN_RB_SETUP];
+		};
+	};
+};
+
+struct amdgpu_fw_shared_drm_key_wa {
+	uint8_t  method;
+	uint8_t  reserved[3];
+};
+
+struct amdgpu_fw_shared_queue_decouple {
+	uint8_t  is_enabled;
+	uint8_t  reserved[7];
+};
+
+struct amdgpu_vcn4_fw_shared {
+	uint32_t present_flag_0;
+	uint8_t pad[12];
+	struct amdgpu_fw_shared_unified_queue_struct sq;
+	uint8_t pad1[8];
+	struct amdgpu_fw_shared_fw_logging fw_log;
+	uint8_t pad2[20];
+	struct amdgpu_fw_shared_rb_setup rb_setup;
+	struct amdgpu_fw_shared_smu_interface_info smu_dpm_interface;
+	struct amdgpu_fw_shared_drm_key_wa drm_key_wa;
+	uint8_t pad3[9];
+	struct amdgpu_fw_shared_queue_decouple decouple;
+};
+
+struct amdgpu_vcn_fwlog {
+	uint32_t rptr;
+	uint32_t wptr;
+	uint32_t buffer_size;
+	uint32_t header_size;
+	uint8_t wrapped;
 };
 
 struct amdgpu_vcn_decode_buffer {
@@ -280,6 +484,28 @@ struct amdgpu_vcn_decode_buffer {
 	uint32_t pad[30];
 };
 
+struct amdgpu_vcn_rb_metadata {
+	uint32_t size;
+	uint32_t present_flag_0;
+
+	uint8_t version;
+	uint8_t ring_id;
+	uint8_t pad[26];
+};
+
+struct amdgpu_vcn5_fw_shared {
+	uint32_t present_flag_0;
+	uint8_t pad[12];
+	struct amdgpu_fw_shared_unified_queue_struct sq;
+	uint8_t pad1[8];
+	struct amdgpu_fw_shared_fw_logging fw_log;
+	uint8_t pad2[20];
+	struct amdgpu_fw_shared_rb_setup rb_setup;
+	struct amdgpu_fw_shared_smu_interface_info smu_dpm_interface;
+	struct amdgpu_fw_shared_drm_key_wa drm_key_wa;
+	uint8_t pad3[404];
+};
+
 #define VCN_BLOCK_ENCODE_DISABLE_MASK 0x80
 #define VCN_BLOCK_DECODE_DISABLE_MASK 0x40
 #define VCN_BLOCK_QUEUE_DISABLE_MASK 0xC0
@@ -290,10 +516,11 @@ enum vcn_ring_type {
 	VCN_UNIFIED_RING,
 };
 
-int amdgpu_vcn_sw_init(struct amdgpu_device *adev);
-int amdgpu_vcn_sw_fini(struct amdgpu_device *adev);
-int amdgpu_vcn_suspend(struct amdgpu_device *adev);
-int amdgpu_vcn_resume(struct amdgpu_device *adev);
+int amdgpu_vcn_early_init(struct amdgpu_device *adev, int i);
+int amdgpu_vcn_sw_init(struct amdgpu_device *adev, int i);
+void amdgpu_vcn_sw_fini(struct amdgpu_device *adev, int i);
+int amdgpu_vcn_suspend(struct amdgpu_device *adev, int i);
+int amdgpu_vcn_resume(struct amdgpu_device *adev, int i);
 void amdgpu_vcn_ring_begin_use(struct amdgpu_ring *ring);
 void amdgpu_vcn_ring_end_use(struct amdgpu_ring *ring);
 
@@ -304,12 +531,43 @@ int amdgpu_vcn_dec_ring_test_ring(struct amdgpu_ring *ring);
 int amdgpu_vcn_dec_ring_test_ib(struct amdgpu_ring *ring, long timeout);
 int amdgpu_vcn_dec_sw_ring_test_ring(struct amdgpu_ring *ring);
 int amdgpu_vcn_dec_sw_ring_test_ib(struct amdgpu_ring *ring, long timeout);
+int amdgpu_vcn_unified_ring_test_ib(struct amdgpu_ring *ring, long timeout);
 
 int amdgpu_vcn_enc_ring_test_ring(struct amdgpu_ring *ring);
 int amdgpu_vcn_enc_ring_test_ib(struct amdgpu_ring *ring, long timeout);
 
 enum amdgpu_ring_priority_level amdgpu_vcn_get_enc_ring_prio(int ring);
 
-void amdgpu_vcn_setup_ucode(struct amdgpu_device *adev);
+void amdgpu_vcn_setup_ucode(struct amdgpu_device *adev, int i);
+
+void amdgpu_vcn_fwlog_init(struct amdgpu_vcn_inst *vcn);
+void amdgpu_debugfs_vcn_fwlog_init(struct amdgpu_device *adev,
+                                   uint8_t i, struct amdgpu_vcn_inst *vcn);
+
+int amdgpu_vcn_process_poison_irq(struct amdgpu_device *adev,
+			struct amdgpu_irq_src *source,
+			struct amdgpu_iv_entry *entry);
+int amdgpu_vcn_ras_late_init(struct amdgpu_device *adev,
+			struct ras_common_if *ras_block);
+int amdgpu_vcn_ras_sw_init(struct amdgpu_device *adev);
+
+int amdgpu_vcn_psp_update_sram(struct amdgpu_device *adev, int inst_idx,
+			       enum AMDGPU_UCODE_ID ucode_id);
+int amdgpu_vcn_save_vcpu_bo(struct amdgpu_device *adev);
+int amdgpu_vcn_sysfs_reset_mask_init(struct amdgpu_device *adev);
+void amdgpu_vcn_sysfs_reset_mask_fini(struct amdgpu_device *adev);
+void amdgpu_debugfs_vcn_sched_mask_init(struct amdgpu_device *adev);
+
+int vcn_set_powergating_state(struct amdgpu_ip_block *ip_block,
+			      enum amd_powergating_state state);
+int amdgpu_vcn_ring_reset(struct amdgpu_ring *ring,
+			  unsigned int vmid,
+			  struct amdgpu_fence *guilty_fence);
+int amdgpu_vcn_reg_dump_init(struct amdgpu_device *adev,
+			     const struct amdgpu_hwip_reg_entry *reg, u32 count);
+void amdgpu_vcn_dump_ip_state(struct amdgpu_ip_block *ip_block);
+void amdgpu_vcn_print_ip_state(struct amdgpu_ip_block *ip_block, struct drm_printer *p);
+void amdgpu_vcn_get_profile(struct amdgpu_device *adev);
+void amdgpu_vcn_put_profile(struct amdgpu_device *adev);
 
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
index 04cf9b207e62..47a6ce4fdc74 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
@@ -23,10 +23,17 @@
 
 #include <linux/module.h>
 
+#ifdef CONFIG_X86
+#include <asm/hypervisor.h>
+#endif
+
 #include <drm/drm_drv.h>
+#include <xen/xen.h>
 
 #include "amdgpu.h"
 #include "amdgpu_ras.h"
+#include "amdgpu_reset.h"
+#include "amdgpu_dpm.h"
 #include "vi.h"
 #include "soc15.h"
 #include "nv.h"
@@ -37,6 +44,18 @@
 		vf2pf_info->ucode_info[ucode].version = ver; \
 	} while (0)
 
+#define mmRCC_CONFIG_MEMSIZE    0xde3
+
+const char *amdgpu_virt_dynamic_crit_table_name[] = {
+	"IP DISCOVERY",
+	"VBIOS IMG",
+	"RAS TELEMETRY",
+	"DATA EXCHANGE",
+	"BAD PAGE INFO",
+	"INIT HEADER",
+	"LAST",
+};
+
 bool amdgpu_virt_mmio_blocked(struct amdgpu_device *adev)
 {
 	/* By now all MMIO pages except mailbox are blocked */
@@ -51,7 +70,8 @@ void amdgpu_virt_init_setting(struct amdgpu_device *adev)
 
 	/* enable virtual display */
 	if (adev->asic_type != CHIP_ALDEBARAN &&
-	    adev->asic_type != CHIP_ARCTURUS) {
+	    adev->asic_type != CHIP_ARCTURUS &&
+	    ((adev->pdev->class >> 8) != PCI_CLASS_ACCELERATOR_PROCESSING)) {
 		if (adev->mode_info.num_crtc == 0)
 			adev->mode_info.num_crtc = 1;
 		adev->enable_virtual_display = true;
@@ -59,52 +79,10 @@ void amdgpu_virt_init_setting(struct amdgpu_device *adev)
 	ddev->driver_features &= ~DRIVER_ATOMIC;
 	adev->cg_flags = 0;
 	adev->pg_flags = 0;
-}
-
-void amdgpu_virt_kiq_reg_write_reg_wait(struct amdgpu_device *adev,
-					uint32_t reg0, uint32_t reg1,
-					uint32_t ref, uint32_t mask)
-{
-	struct amdgpu_kiq *kiq = &adev->gfx.kiq;
-	struct amdgpu_ring *ring = &kiq->ring;
-	signed long r, cnt = 0;
-	unsigned long flags;
-	uint32_t seq;
-
-	spin_lock_irqsave(&kiq->ring_lock, flags);
-	amdgpu_ring_alloc(ring, 32);
-	amdgpu_ring_emit_reg_write_reg_wait(ring, reg0, reg1,
-					    ref, mask);
-	r = amdgpu_fence_emit_polling(ring, &seq, MAX_KIQ_REG_WAIT);
-	if (r)
-		goto failed_undo;
-
-	amdgpu_ring_commit(ring);
-	spin_unlock_irqrestore(&kiq->ring_lock, flags);
-
-	r = amdgpu_fence_wait_polling(ring, seq, MAX_KIQ_REG_WAIT);
-
-	/* don't wait anymore for IRQ context */
-	if (r < 1 && in_interrupt())
-		goto failed_kiq;
-
-	might_sleep();
-	while (r < 1 && cnt++ < MAX_KIQ_REG_TRY) {
 
-		msleep(MAX_KIQ_REG_BAILOUT_INTERVAL);
-		r = amdgpu_fence_wait_polling(ring, seq, MAX_KIQ_REG_WAIT);
-	}
-
-	if (cnt > MAX_KIQ_REG_TRY)
-		goto failed_kiq;
-
-	return;
-
-failed_undo:
-	amdgpu_ring_undo(ring);
-	spin_unlock_irqrestore(&kiq->ring_lock, flags);
-failed_kiq:
-	dev_err(adev->dev, "failed to write reg %x wait reg %x\n", reg0, reg1);
+	/* Reduce kcq number to 2 to reduce latency */
+	if (amdgpu_num_kcq == -1)
+		amdgpu_num_kcq = 2;
 }
 
 /**
@@ -121,8 +99,10 @@ int amdgpu_virt_request_full_gpu(struct amdgpu_device *adev, bool init)
 
 	if (virt->ops && virt->ops->req_full_gpu) {
 		r = virt->ops->req_full_gpu(adev, init);
-		if (r)
+		if (r) {
+			adev->no_hw_access = true;
 			return r;
+		}
 
 		adev->virt.caps &= ~AMDGPU_SRIOV_CAPS_RUNTIME;
 	}
@@ -182,9 +162,24 @@ void amdgpu_virt_request_init_data(struct amdgpu_device *adev)
 		virt->ops->req_init_data(adev);
 
 	if (adev->virt.req_init_data_ver > 0)
-		DRM_INFO("host supports REQ_INIT_DATA handshake\n");
+		dev_info(adev->dev, "host supports REQ_INIT_DATA handshake of critical_region_version %d\n",
+				 adev->virt.req_init_data_ver);
 	else
-		DRM_WARN("host doesn't support REQ_INIT_DATA handshake\n");
+		dev_warn(adev->dev, "host doesn't support REQ_INIT_DATA handshake\n");
+}
+
+/**
+ * amdgpu_virt_ready_to_reset() - send ready to reset to host
+ * @adev:	amdgpu device.
+ * Send ready to reset message to GPU hypervisor to signal we have stopped GPU
+ * activity and is ready for host FLR
+ */
+void amdgpu_virt_ready_to_reset(struct amdgpu_device *adev)
+{
+	struct amdgpu_virt *virt = &adev->virt;
+
+	if (virt->ops && virt->ops->reset_gpu)
+		virt->ops->ready_to_reset(adev);
 }
 
 /**
@@ -217,17 +212,18 @@ int amdgpu_virt_alloc_mm_table(struct amdgpu_device *adev)
 		return 0;
 
 	r = amdgpu_bo_create_kernel(adev, PAGE_SIZE, PAGE_SIZE,
-				    AMDGPU_GEM_DOMAIN_VRAM,
+				    AMDGPU_GEM_DOMAIN_VRAM |
+				    AMDGPU_GEM_DOMAIN_GTT,
 				    &adev->virt.mm_table.bo,
 				    &adev->virt.mm_table.gpu_addr,
 				    (void *)&adev->virt.mm_table.cpu_addr);
 	if (r) {
-		DRM_ERROR("failed to alloc mm table and error = %d.\n", r);
+		dev_err(adev->dev, "failed to alloc mm table and error = %d.\n", r);
 		return r;
 	}
 
 	memset((void *)adev->virt.mm_table.cpu_addr, 0, PAGE_SIZE);
-	DRM_INFO("MM table gpu addr = 0x%llx, cpu addr = %p.\n",
+	dev_info(adev->dev, "MM table gpu addr = 0x%llx, cpu addr = %p.\n",
 		 adev->virt.mm_table.gpu_addr,
 		 adev->virt.mm_table.cpu_addr);
 	return 0;
@@ -249,6 +245,22 @@ void amdgpu_virt_free_mm_table(struct amdgpu_device *adev)
 	adev->virt.mm_table.gpu_addr = 0;
 }
 
+/**
+ * amdgpu_virt_rcvd_ras_interrupt() - receive ras interrupt
+ * @adev:	amdgpu device.
+ * Check whether host sent RAS error message
+ * Return: true if found, otherwise false
+ */
+bool amdgpu_virt_rcvd_ras_interrupt(struct amdgpu_device *adev)
+{
+	struct amdgpu_virt *virt = &adev->virt;
+
+	if (!virt->ops || !virt->ops->rcvd_ras_intr)
+		return false;
+
+	return virt->ops->rcvd_ras_intr(adev);
+}
+
 
 unsigned int amd_sriov_msg_checksum(void *obj,
 				unsigned long obj_size,
@@ -283,17 +295,15 @@ static int amdgpu_virt_init_ras_err_handler_data(struct amdgpu_device *adev)
 
 	*data = kmalloc(sizeof(struct amdgpu_virt_ras_err_handler_data), GFP_KERNEL);
 	if (!*data)
-		return -ENOMEM;
+		goto data_failure;
 
-	bps = kmalloc_array(align_space, sizeof((*data)->bps), GFP_KERNEL);
-	bps_bo = kmalloc_array(align_space, sizeof((*data)->bps_bo), GFP_KERNEL);
+	bps = kmalloc_array(align_space, sizeof(*(*data)->bps), GFP_KERNEL);
+	if (!bps)
+		goto bps_failure;
 
-	if (!bps || !bps_bo) {
-		kfree(bps);
-		kfree(bps_bo);
-		kfree(*data);
-		return -ENOMEM;
-	}
+	bps_bo = kmalloc_array(align_space, sizeof(*(*data)->bps_bo), GFP_KERNEL);
+	if (!bps_bo)
+		goto bps_bo_failure;
 
 	(*data)->bps = bps;
 	(*data)->bps_bo = bps_bo;
@@ -303,6 +313,13 @@ static int amdgpu_virt_init_ras_err_handler_data(struct amdgpu_device *adev)
 	virt->ras_init_done = true;
 
 	return 0;
+
+bps_bo_failure:
+	kfree(bps);
+bps_failure:
+	kfree(*data);
+data_failure:
+	return -ENOMEM;
 }
 
 static void amdgpu_virt_ras_release_bp(struct amdgpu_device *adev)
@@ -317,8 +334,10 @@ static void amdgpu_virt_ras_release_bp(struct amdgpu_device *adev)
 
 	for (i = data->last_reserved - 1; i >= 0; i--) {
 		bo = data->bps_bo[i];
-		amdgpu_bo_free_kernel(&bo, NULL, NULL);
-		data->bps_bo[i] = bo;
+		if (bo) {
+			amdgpu_bo_free_kernel(&bo, NULL, NULL);
+			data->bps_bo[i] = bo;
+		}
 		data->last_reserved = i;
 	}
 }
@@ -358,6 +377,8 @@ static void amdgpu_virt_ras_reserve_bps(struct amdgpu_device *adev)
 {
 	struct amdgpu_virt *virt = &adev->virt;
 	struct amdgpu_virt_ras_err_handler_data *data = virt->virt_eh_data;
+	struct amdgpu_vram_mgr *mgr = &adev->mman.vram_mgr;
+	struct ttm_resource_manager *man = &mgr->manager;
 	struct amdgpu_bo *bo = NULL;
 	uint64_t bp;
 	int i;
@@ -373,13 +394,20 @@ static void amdgpu_virt_ras_reserve_bps(struct amdgpu_device *adev)
 		 * 2) a ras bad page has been reserved (duplicate error injection
 		 *    for one page);
 		 */
-		if (amdgpu_bo_create_kernel_at(adev, bp << AMDGPU_GPU_PAGE_SHIFT,
-					       AMDGPU_GPU_PAGE_SIZE,
-					       AMDGPU_GEM_DOMAIN_VRAM,
-					       &bo, NULL))
-			DRM_DEBUG("RAS WARN: reserve vram for retired page %llx fail\n", bp);
-
-		data->bps_bo[i] = bo;
+		if  (ttm_resource_manager_used(man)) {
+			amdgpu_vram_mgr_reserve_range(&adev->mman.vram_mgr,
+				bp << AMDGPU_GPU_PAGE_SHIFT,
+				AMDGPU_GPU_PAGE_SIZE);
+			data->bps_bo[i] = NULL;
+		} else {
+			if (amdgpu_bo_create_kernel_at(adev, bp << AMDGPU_GPU_PAGE_SHIFT,
+							AMDGPU_GPU_PAGE_SIZE,
+							&bo, NULL))
+				dev_dbg(adev->dev,
+						"RAS WARN: reserve vram for retired page %llx fail\n",
+						bp);
+			data->bps_bo[i] = bo;
+		}
 		data->last_reserved = i + 1;
 		bo = NULL;
 	}
@@ -408,11 +436,19 @@ static void amdgpu_virt_add_bad_page(struct amdgpu_device *adev,
 	struct eeprom_table_record bp;
 	uint64_t retired_page;
 	uint32_t bp_idx, bp_cnt;
+	void *vram_usage_va = NULL;
+
+	if (adev->mman.fw_vram_usage_va)
+		vram_usage_va = adev->mman.fw_vram_usage_va;
+	else
+		vram_usage_va = adev->mman.drv_vram_usage_va;
+
+	memset(&bp, 0, sizeof(bp));
 
 	if (bp_block_size) {
 		bp_cnt = bp_block_size / sizeof(uint64_t);
 		for (bp_idx = 0; bp_idx < bp_cnt; bp_idx++) {
-			retired_page = *(uint64_t *)(adev->mman.fw_vram_usage_va +
+			retired_page = *(uint64_t *)(vram_usage_va +
 					bp_block_offset + bp_idx * sizeof(uint64_t));
 			bp.retired_page = retired_page;
 
@@ -439,7 +475,7 @@ static int amdgpu_virt_read_pf2vf_data(struct amdgpu_device *adev)
 		return -EINVAL;
 
 	if (pf2vf_info->size > 1024) {
-		DRM_ERROR("invalid pf2vf message size\n");
+		dev_err(adev->dev, "invalid pf2vf message size: 0x%x\n", pf2vf_info->size);
 		return -EINVAL;
 	}
 
@@ -450,7 +486,9 @@ static int amdgpu_virt_read_pf2vf_data(struct amdgpu_device *adev)
 			adev->virt.fw_reserve.p_pf2vf, pf2vf_info->size,
 			adev->virt.fw_reserve.checksum_key, checksum);
 		if (checksum != checkval) {
-			DRM_ERROR("invalid pf2vf message\n");
+			dev_err(adev->dev,
+				"invalid pf2vf message: header checksum=0x%x calculated checksum=0x%x\n",
+				checksum, checkval);
 			return -EINVAL;
 		}
 
@@ -464,7 +502,9 @@ static int amdgpu_virt_read_pf2vf_data(struct amdgpu_device *adev)
 			adev->virt.fw_reserve.p_pf2vf, pf2vf_info->size,
 			0, checksum);
 		if (checksum != checkval) {
-			DRM_ERROR("invalid pf2vf message\n");
+			dev_err(adev->dev,
+				"invalid pf2vf message: header checksum=0x%x calculated checksum=0x%x\n",
+				checksum, checkval);
 			return -EINVAL;
 		}
 
@@ -493,14 +533,17 @@ static int amdgpu_virt_read_pf2vf_data(struct amdgpu_device *adev)
 			tmp = ((struct amd_sriov_msg_pf2vf_info *)pf2vf_info)->mm_bw_management[i].encode_max_frame_pixels;
 			adev->virt.encode_max_frame_pixels = max(tmp, adev->virt.encode_max_frame_pixels);
 		}
-		if((adev->virt.decode_max_dimension_pixels > 0) || (adev->virt.encode_max_dimension_pixels > 0))
+		if ((adev->virt.decode_max_dimension_pixels > 0) || (adev->virt.encode_max_dimension_pixels > 0))
 			adev->virt.is_mm_bw_enabled = true;
 
 		adev->unique_id =
 			((struct amd_sriov_msg_pf2vf_info *)pf2vf_info)->uuid;
+		adev->virt.ras_en_caps.all = ((struct amd_sriov_msg_pf2vf_info *)pf2vf_info)->ras_en_caps.all;
+		adev->virt.ras_telemetry_en_caps.all =
+			((struct amd_sriov_msg_pf2vf_info *)pf2vf_info)->ras_telemetry_en_caps.all;
 		break;
 	default:
-		DRM_ERROR("invalid pf2vf version\n");
+		dev_err(adev->dev, "invalid pf2vf version: 0x%x\n", pf2vf_info->version);
 		return -EINVAL;
 	}
 
@@ -548,7 +591,6 @@ static void amdgpu_virt_populate_vf2pf_ucode_info(struct amdgpu_device *adev)
 static int amdgpu_virt_write_vf2pf_data(struct amdgpu_device *adev)
 {
 	struct amd_sriov_msg_vf2pf_info *vf2pf_info;
-	struct ttm_resource_manager *vram_man = ttm_manager_type(&adev->mman.bdev, TTM_PL_VRAM);
 
 	vf2pf_info = (struct amd_sriov_msg_vf2pf_info *) adev->virt.fw_reserve.p_vf2pf;
 
@@ -571,8 +613,10 @@ static int amdgpu_virt_write_vf2pf_data(struct amdgpu_device *adev)
 	vf2pf_info->driver_cert = 0;
 	vf2pf_info->os_info.all = 0;
 
-	vf2pf_info->fb_usage = amdgpu_vram_mgr_usage(vram_man) >> 20;
-	vf2pf_info->fb_vis_usage = amdgpu_vram_mgr_vis_usage(vram_man) >> 20;
+	vf2pf_info->fb_usage = ttm_resource_manager_used(&adev->mman.vram_mgr.manager) ?
+		 ttm_resource_manager_usage(&adev->mman.vram_mgr.manager) >> 20 : 0;
+	vf2pf_info->fb_vis_usage =
+		amdgpu_vram_mgr_vis_usage(&adev->mman.vram_mgr) >> 20;
 	vf2pf_info->fb_size = adev->gmc.real_vram_size >> 20;
 	vf2pf_info->fb_vis_size = adev->gmc.visible_vram_size >> 20;
 
@@ -585,9 +629,15 @@ static int amdgpu_virt_write_vf2pf_data(struct amdgpu_device *adev)
 	vf2pf_info->decode_usage = 0;
 
 	vf2pf_info->dummy_page_addr = (uint64_t)adev->dummy_page_addr;
+	if (amdgpu_sriov_is_mes_info_enable(adev)) {
+		vf2pf_info->mes_info_addr =
+			(uint64_t)(adev->mes.resource_1_gpu_addr[0] + AMDGPU_GPU_PAGE_SIZE);
+		vf2pf_info->mes_info_size =
+			adev->mes.resource_1[0]->tbo.base.size - AMDGPU_GPU_PAGE_SIZE;
+	}
 	vf2pf_info->checksum =
 		amd_sriov_msg_checksum(
-		vf2pf_info, vf2pf_info->header.size, 0, 0);
+		vf2pf_info, sizeof(*vf2pf_info), 0, 0);
 
 	return 0;
 }
@@ -598,18 +648,59 @@ static void amdgpu_virt_update_vf2pf_work_item(struct work_struct *work)
 	int ret;
 
 	ret = amdgpu_virt_read_pf2vf_data(adev);
-	if (ret)
+	if (ret) {
+		adev->virt.vf2pf_update_retry_cnt++;
+
+		if ((amdgpu_virt_rcvd_ras_interrupt(adev) ||
+			adev->virt.vf2pf_update_retry_cnt >= AMDGPU_VF2PF_UPDATE_MAX_RETRY_LIMIT) &&
+			amdgpu_sriov_runtime(adev)) {
+
+			amdgpu_ras_set_fed(adev, true);
+			if (amdgpu_reset_domain_schedule(adev->reset_domain,
+							&adev->kfd.reset_work))
+				return;
+			else
+				dev_err(adev->dev, "Failed to queue work! at %s", __func__);
+		}
+
 		goto out;
+	}
+
+	adev->virt.vf2pf_update_retry_cnt = 0;
 	amdgpu_virt_write_vf2pf_data(adev);
 
 out:
 	schedule_delayed_work(&(adev->virt.vf2pf_work), adev->virt.vf2pf_update_interval_ms);
 }
 
+static int amdgpu_virt_read_exchange_data_from_mem(struct amdgpu_device *adev, uint32_t *pfvf_data)
+{
+	uint32_t dataexchange_offset =
+		adev->virt.crit_regn_tbl[AMD_SRIOV_MSG_DATAEXCHANGE_TABLE_ID].offset;
+	uint32_t dataexchange_size =
+		adev->virt.crit_regn_tbl[AMD_SRIOV_MSG_DATAEXCHANGE_TABLE_ID].size_kb << 10;
+	uint64_t pos = 0;
+
+	dev_info(adev->dev,
+			"Got data exchange info from dynamic crit_region_table at offset 0x%x with size of 0x%x bytes.\n",
+			dataexchange_offset, dataexchange_size);
+
+	if (!IS_ALIGNED(dataexchange_offset, 4) || !IS_ALIGNED(dataexchange_size, 4)) {
+		dev_err(adev->dev, "Data exchange data not aligned to 4 bytes\n");
+		return -EINVAL;
+	}
+
+	pos = (uint64_t)dataexchange_offset;
+	amdgpu_device_vram_access(adev, pos, pfvf_data,
+			dataexchange_size, false);
+
+	return 0;
+}
+
 void amdgpu_virt_fini_data_exchange(struct amdgpu_device *adev)
 {
 	if (adev->virt.vf2pf_update_interval_ms != 0) {
-		DRM_INFO("clean up the vf2pf work item\n");
+		dev_info(adev->dev, "clean up the vf2pf work item\n");
 		cancel_delayed_work_sync(&adev->virt.vf2pf_work);
 		adev->virt.vf2pf_update_interval_ms = 0;
 	}
@@ -617,58 +708,118 @@ void amdgpu_virt_fini_data_exchange(struct amdgpu_device *adev)
 
 void amdgpu_virt_init_data_exchange(struct amdgpu_device *adev)
 {
-	uint64_t bp_block_offset = 0;
-	uint32_t bp_block_size = 0;
-	struct amd_sriov_msg_pf2vf_info *pf2vf_v2 = NULL;
+	uint32_t *pfvf_data = NULL;
 
 	adev->virt.fw_reserve.p_pf2vf = NULL;
 	adev->virt.fw_reserve.p_vf2pf = NULL;
 	adev->virt.vf2pf_update_interval_ms = 0;
+	adev->virt.vf2pf_update_retry_cnt = 0;
 
-	if (adev->mman.fw_vram_usage_va != NULL) {
-		adev->virt.vf2pf_update_interval_ms = 2000;
+	if (adev->mman.fw_vram_usage_va && adev->mman.drv_vram_usage_va) {
+		dev_warn(adev->dev, "Currently fw_vram and drv_vram should not have values at the same time!");
+	} else if (adev->mman.fw_vram_usage_va || adev->mman.drv_vram_usage_va) {
+		/* go through this logic in ip_init and reset to init workqueue*/
+		amdgpu_virt_exchange_data(adev);
 
-		adev->virt.fw_reserve.p_pf2vf =
-			(struct amd_sriov_msg_pf2vf_info_header *)
-			(adev->mman.fw_vram_usage_va + (AMD_SRIOV_MSG_PF2VF_OFFSET_KB << 10));
-		adev->virt.fw_reserve.p_vf2pf =
-			(struct amd_sriov_msg_vf2pf_info_header *)
-			(adev->mman.fw_vram_usage_va + (AMD_SRIOV_MSG_VF2PF_OFFSET_KB << 10));
+		INIT_DELAYED_WORK(&adev->virt.vf2pf_work, amdgpu_virt_update_vf2pf_work_item);
+		schedule_delayed_work(&(adev->virt.vf2pf_work), msecs_to_jiffies(adev->virt.vf2pf_update_interval_ms));
+	} else if (adev->bios != NULL) {
+		/* got through this logic in early init stage to get necessary flags, e.g. rlcg_acc related*/
+		if (adev->virt.req_init_data_ver == GPU_CRIT_REGION_V2) {
+			pfvf_data =
+				kzalloc(adev->virt.crit_regn_tbl[AMD_SRIOV_MSG_DATAEXCHANGE_TABLE_ID].size_kb << 10,
+					GFP_KERNEL);
+			if (!pfvf_data) {
+				dev_err(adev->dev, "Failed to allocate memory for pfvf_data\n");
+				return;
+			}
 
-		amdgpu_virt_read_pf2vf_data(adev);
-		amdgpu_virt_write_vf2pf_data(adev);
+			if (amdgpu_virt_read_exchange_data_from_mem(adev, pfvf_data))
+				goto free_pfvf_data;
 
-		/* bad page handling for version 2 */
-		if (adev->virt.fw_reserve.p_pf2vf->version == 2) {
-				pf2vf_v2 = (struct amd_sriov_msg_pf2vf_info *)adev->virt.fw_reserve.p_pf2vf;
+			adev->virt.fw_reserve.p_pf2vf =
+				(struct amd_sriov_msg_pf2vf_info_header *)pfvf_data;
 
-				bp_block_offset = ((uint64_t)pf2vf_v2->bp_block_offset_low & 0xFFFFFFFF) |
-						((((uint64_t)pf2vf_v2->bp_block_offset_high) << 32) & 0xFFFFFFFF00000000);
-				bp_block_size = pf2vf_v2->bp_block_size;
+			amdgpu_virt_read_pf2vf_data(adev);
 
-				if (bp_block_size && !adev->virt.ras_init_done)
-					amdgpu_virt_init_ras_err_handler_data(adev);
+free_pfvf_data:
+			kfree(pfvf_data);
+			pfvf_data = NULL;
+			adev->virt.fw_reserve.p_pf2vf = NULL;
+		} else {
+			adev->virt.fw_reserve.p_pf2vf =
+				(struct amd_sriov_msg_pf2vf_info_header *)
+				(adev->bios + (AMD_SRIOV_MSG_PF2VF_OFFSET_KB_V1 << 10));
 
-				if (adev->virt.ras_init_done)
-					amdgpu_virt_add_bad_page(adev, bp_block_offset, bp_block_size);
+			amdgpu_virt_read_pf2vf_data(adev);
+		}
+	}
+}
+
+
+void amdgpu_virt_exchange_data(struct amdgpu_device *adev)
+{
+	uint64_t bp_block_offset = 0;
+	uint32_t bp_block_size = 0;
+	struct amd_sriov_msg_pf2vf_info *pf2vf_v2 = NULL;
+
+	if (adev->mman.fw_vram_usage_va || adev->mman.drv_vram_usage_va) {
+		if (adev->mman.fw_vram_usage_va) {
+			if (adev->virt.req_init_data_ver == GPU_CRIT_REGION_V2) {
+				adev->virt.fw_reserve.p_pf2vf =
+					(struct amd_sriov_msg_pf2vf_info_header *)
+					(adev->mman.fw_vram_usage_va +
+					adev->virt.crit_regn_tbl[AMD_SRIOV_MSG_DATAEXCHANGE_TABLE_ID].offset);
+				adev->virt.fw_reserve.p_vf2pf =
+					(struct amd_sriov_msg_vf2pf_info_header *)
+					(adev->mman.fw_vram_usage_va +
+					adev->virt.crit_regn_tbl[AMD_SRIOV_MSG_DATAEXCHANGE_TABLE_ID].offset +
+					(AMD_SRIOV_MSG_SIZE_KB << 10));
+				adev->virt.fw_reserve.ras_telemetry =
+					(adev->mman.fw_vram_usage_va +
+					adev->virt.crit_regn_tbl[AMD_SRIOV_MSG_RAS_TELEMETRY_TABLE_ID].offset);
+			} else {
+				adev->virt.fw_reserve.p_pf2vf =
+					(struct amd_sriov_msg_pf2vf_info_header *)
+					(adev->mman.fw_vram_usage_va + (AMD_SRIOV_MSG_PF2VF_OFFSET_KB_V1 << 10));
+				adev->virt.fw_reserve.p_vf2pf =
+					(struct amd_sriov_msg_vf2pf_info_header *)
+					(adev->mman.fw_vram_usage_va + (AMD_SRIOV_MSG_VF2PF_OFFSET_KB_V1 << 10));
+				adev->virt.fw_reserve.ras_telemetry =
+					(adev->mman.fw_vram_usage_va + (AMD_SRIOV_MSG_RAS_TELEMETRY_OFFSET_KB_V1 << 10));
 			}
-	} else if (adev->bios != NULL) {
-		adev->virt.fw_reserve.p_pf2vf =
-			(struct amd_sriov_msg_pf2vf_info_header *)
-			(adev->bios + (AMD_SRIOV_MSG_PF2VF_OFFSET_KB << 10));
+		} else if (adev->mman.drv_vram_usage_va) {
+			adev->virt.fw_reserve.p_pf2vf =
+				(struct amd_sriov_msg_pf2vf_info_header *)
+				(adev->mman.drv_vram_usage_va + (AMD_SRIOV_MSG_PF2VF_OFFSET_KB_V1 << 10));
+			adev->virt.fw_reserve.p_vf2pf =
+				(struct amd_sriov_msg_vf2pf_info_header *)
+				(adev->mman.drv_vram_usage_va + (AMD_SRIOV_MSG_VF2PF_OFFSET_KB_V1 << 10));
+			adev->virt.fw_reserve.ras_telemetry =
+				(adev->mman.drv_vram_usage_va + (AMD_SRIOV_MSG_RAS_TELEMETRY_OFFSET_KB_V1 << 10));
+		}
 
 		amdgpu_virt_read_pf2vf_data(adev);
+		amdgpu_virt_write_vf2pf_data(adev);
 
-		return;
-	}
+		/* bad page handling for version 2 */
+		if (adev->virt.fw_reserve.p_pf2vf->version == 2) {
+			pf2vf_v2 = (struct amd_sriov_msg_pf2vf_info *)adev->virt.fw_reserve.p_pf2vf;
 
-	if (adev->virt.vf2pf_update_interval_ms != 0) {
-		INIT_DELAYED_WORK(&adev->virt.vf2pf_work, amdgpu_virt_update_vf2pf_work_item);
-		schedule_delayed_work(&(adev->virt.vf2pf_work), adev->virt.vf2pf_update_interval_ms);
+			bp_block_offset = ((uint64_t)pf2vf_v2->bp_block_offset_low & 0xFFFFFFFF) |
+				((((uint64_t)pf2vf_v2->bp_block_offset_high) << 32) & 0xFFFFFFFF00000000);
+			bp_block_size = pf2vf_v2->bp_block_size;
+
+			if (bp_block_size && !adev->virt.ras_init_done)
+				amdgpu_virt_init_ras_err_handler_data(adev);
+
+			if (adev->virt.ras_init_done)
+				amdgpu_virt_add_bad_page(adev, bp_block_offset, bp_block_size);
+		}
 	}
 }
 
-void amdgpu_detect_virtualization(struct amdgpu_device *adev)
+static u32 amdgpu_virt_init_detect_asic(struct amdgpu_device *adev)
 {
 	uint32_t reg;
 
@@ -684,6 +835,7 @@ void amdgpu_detect_virtualization(struct amdgpu_device *adev)
 	case CHIP_SIENNA_CICHLID:
 	case CHIP_ARCTURUS:
 	case CHIP_ALDEBARAN:
+	case CHIP_IP_DISCOVERY:
 		reg = RREG32(mmRCC_IOV_FUNC_IDENTIFIER);
 		break;
 	default: /* other chip doesn't support SRIOV */
@@ -698,18 +850,36 @@ void amdgpu_detect_virtualization(struct amdgpu_device *adev)
 		adev->virt.caps |= AMDGPU_SRIOV_CAPS_ENABLE_IOV;
 
 	if (!reg) {
-		if (is_virtual_machine())	/* passthrough mode exclus sriov mod */
+		/* passthrough mode exclus sriov mod */
+		if (is_virtual_machine() && !xen_initial_domain())
 			adev->virt.caps |= AMDGPU_PASSTHROUGH_MODE;
 	}
 
+	return reg;
+}
+
+static bool amdgpu_virt_init_req_data(struct amdgpu_device *adev, u32 reg)
+{
+	bool is_sriov = false;
+
 	/* we have the ability to check now */
 	if (amdgpu_sriov_vf(adev)) {
+		is_sriov = true;
+
 		switch (adev->asic_type) {
 		case CHIP_TONGA:
 		case CHIP_FIJI:
 			vi_set_virt_ops(adev);
 			break;
 		case CHIP_VEGA10:
+			soc15_set_virt_ops(adev);
+#ifdef CONFIG_X86
+			/* not send GPU_INIT_DATA with MS_HYPERV*/
+			if (!hypervisor_is_type(X86_HYPER_MS_HYPERV))
+#endif
+				/* send a dummy GPU_INIT_DATA request to host on vega10 */
+				amdgpu_virt_request_init_data(adev);
+			break;
 		case CHIP_VEGA20:
 		case CHIP_ARCTURUS:
 		case CHIP_ALDEBARAN:
@@ -718,15 +888,258 @@ void amdgpu_detect_virtualization(struct amdgpu_device *adev)
 		case CHIP_NAVI10:
 		case CHIP_NAVI12:
 		case CHIP_SIENNA_CICHLID:
+		case CHIP_IP_DISCOVERY:
 			nv_set_virt_ops(adev);
 			/* try send GPU_INIT_DATA request to host */
 			amdgpu_virt_request_init_data(adev);
 			break;
 		default: /* other chip doesn't support SRIOV */
-			DRM_ERROR("Unknown asic type: %d!\n", adev->asic_type);
+			is_sriov = false;
+			dev_err(adev->dev, "Unknown asic type: %d!\n", adev->asic_type);
 			break;
 		}
 	}
+
+	return is_sriov;
+}
+
+static void amdgpu_virt_init_ras(struct amdgpu_device *adev)
+{
+	ratelimit_state_init(&adev->virt.ras.ras_error_cnt_rs, 5 * HZ, 1);
+	ratelimit_state_init(&adev->virt.ras.ras_cper_dump_rs, 5 * HZ, 1);
+	ratelimit_state_init(&adev->virt.ras.ras_chk_criti_rs, 5 * HZ, 1);
+
+	ratelimit_set_flags(&adev->virt.ras.ras_error_cnt_rs,
+			    RATELIMIT_MSG_ON_RELEASE);
+	ratelimit_set_flags(&adev->virt.ras.ras_cper_dump_rs,
+			    RATELIMIT_MSG_ON_RELEASE);
+	ratelimit_set_flags(&adev->virt.ras.ras_chk_criti_rs,
+			    RATELIMIT_MSG_ON_RELEASE);
+
+	mutex_init(&adev->virt.ras.ras_telemetry_mutex);
+	mutex_init(&adev->virt.access_req_mutex);
+
+	adev->virt.ras.cper_rptr = 0;
+}
+
+static uint8_t amdgpu_virt_crit_region_calc_checksum(uint8_t *buf_start, uint8_t *buf_end)
+{
+	uint32_t sum = 0;
+
+	if (buf_start >= buf_end)
+		return 0;
+
+	for (; buf_start < buf_end; buf_start++)
+		sum += buf_start[0];
+
+	return 0xffffffff - sum;
+}
+
+int amdgpu_virt_init_critical_region(struct amdgpu_device *adev)
+{
+	struct amd_sriov_msg_init_data_header *init_data_hdr = NULL;
+	u64 init_hdr_offset = adev->virt.init_data_header.offset;
+	u64 init_hdr_size = (u64)adev->virt.init_data_header.size_kb << 10; /* KB → bytes */
+	u64 vram_size;
+	u64 end;
+	int r = 0;
+	uint8_t checksum = 0;
+
+	/* Skip below init if critical region version != v2 */
+	if (adev->virt.req_init_data_ver != GPU_CRIT_REGION_V2)
+		return 0;
+
+	if (init_hdr_offset < 0) {
+		dev_err(adev->dev, "Invalid init header offset\n");
+		return -EINVAL;
+	}
+
+	vram_size = RREG32(mmRCC_CONFIG_MEMSIZE);
+	if (!vram_size || vram_size == U32_MAX)
+		return -EINVAL;
+	vram_size <<= 20;
+
+	if (check_add_overflow(init_hdr_offset, init_hdr_size, &end) || end > vram_size) {
+		dev_err(adev->dev, "init_data_header exceeds VRAM size, exiting\n");
+		return -EINVAL;
+	}
+
+	/* Allocate for init_data_hdr */
+	init_data_hdr = kzalloc(sizeof(struct amd_sriov_msg_init_data_header), GFP_KERNEL);
+	if (!init_data_hdr)
+		return -ENOMEM;
+
+	amdgpu_device_vram_access(adev, (uint64_t)init_hdr_offset, (uint32_t *)init_data_hdr,
+					sizeof(struct amd_sriov_msg_init_data_header), false);
+
+	/* Table validation */
+	if (strncmp(init_data_hdr->signature,
+				AMDGPU_SRIOV_CRIT_DATA_SIGNATURE,
+				AMDGPU_SRIOV_CRIT_DATA_SIG_LEN) != 0) {
+		dev_err(adev->dev, "Invalid init data signature: %.4s\n",
+			init_data_hdr->signature);
+		r = -EINVAL;
+		goto out;
+	}
+
+	checksum = amdgpu_virt_crit_region_calc_checksum(
+			(uint8_t *)&init_data_hdr->initdata_offset,
+			(uint8_t *)init_data_hdr +
+			sizeof(struct amd_sriov_msg_init_data_header));
+	if (checksum != init_data_hdr->checksum) {
+		dev_err(adev->dev, "Found unmatching checksum from calculation 0x%x and init_data 0x%x\n",
+				checksum, init_data_hdr->checksum);
+		r = -EINVAL;
+		goto out;
+	}
+
+	memset(&adev->virt.crit_regn, 0, sizeof(adev->virt.crit_regn));
+	memset(adev->virt.crit_regn_tbl, 0, sizeof(adev->virt.crit_regn_tbl));
+
+	adev->virt.crit_regn.offset = init_data_hdr->initdata_offset;
+	adev->virt.crit_regn.size_kb = init_data_hdr->initdata_size_in_kb;
+
+	/* Validation and initialization for each table entry */
+	if (IS_SRIOV_CRIT_REGN_ENTRY_VALID(init_data_hdr, AMD_SRIOV_MSG_IPD_TABLE_ID)) {
+		if (!init_data_hdr->ip_discovery_size_in_kb ||
+				init_data_hdr->ip_discovery_size_in_kb > DISCOVERY_TMR_SIZE) {
+			dev_err(adev->dev, "Invalid %s size: 0x%x\n",
+				amdgpu_virt_dynamic_crit_table_name[AMD_SRIOV_MSG_IPD_TABLE_ID],
+				init_data_hdr->ip_discovery_size_in_kb);
+			r = -EINVAL;
+			goto out;
+		}
+
+		adev->virt.crit_regn_tbl[AMD_SRIOV_MSG_IPD_TABLE_ID].offset =
+			init_data_hdr->ip_discovery_offset;
+		adev->virt.crit_regn_tbl[AMD_SRIOV_MSG_IPD_TABLE_ID].size_kb =
+			init_data_hdr->ip_discovery_size_in_kb;
+	}
+
+	if (IS_SRIOV_CRIT_REGN_ENTRY_VALID(init_data_hdr, AMD_SRIOV_MSG_VBIOS_IMG_TABLE_ID)) {
+		if (!init_data_hdr->vbios_img_size_in_kb) {
+			dev_err(adev->dev, "Invalid %s size: 0x%x\n",
+				amdgpu_virt_dynamic_crit_table_name[AMD_SRIOV_MSG_VBIOS_IMG_TABLE_ID],
+				init_data_hdr->vbios_img_size_in_kb);
+			r = -EINVAL;
+			goto out;
+		}
+
+		adev->virt.crit_regn_tbl[AMD_SRIOV_MSG_VBIOS_IMG_TABLE_ID].offset =
+			init_data_hdr->vbios_img_offset;
+		adev->virt.crit_regn_tbl[AMD_SRIOV_MSG_VBIOS_IMG_TABLE_ID].size_kb =
+			init_data_hdr->vbios_img_size_in_kb;
+	}
+
+	if (IS_SRIOV_CRIT_REGN_ENTRY_VALID(init_data_hdr, AMD_SRIOV_MSG_RAS_TELEMETRY_TABLE_ID)) {
+		if (!init_data_hdr->ras_tele_info_size_in_kb) {
+			dev_err(adev->dev, "Invalid %s size: 0x%x\n",
+				amdgpu_virt_dynamic_crit_table_name[AMD_SRIOV_MSG_RAS_TELEMETRY_TABLE_ID],
+				init_data_hdr->ras_tele_info_size_in_kb);
+			r = -EINVAL;
+			goto out;
+		}
+
+		adev->virt.crit_regn_tbl[AMD_SRIOV_MSG_RAS_TELEMETRY_TABLE_ID].offset =
+			init_data_hdr->ras_tele_info_offset;
+		adev->virt.crit_regn_tbl[AMD_SRIOV_MSG_RAS_TELEMETRY_TABLE_ID].size_kb =
+			init_data_hdr->ras_tele_info_size_in_kb;
+	}
+
+	if (IS_SRIOV_CRIT_REGN_ENTRY_VALID(init_data_hdr, AMD_SRIOV_MSG_DATAEXCHANGE_TABLE_ID)) {
+		if (!init_data_hdr->dataexchange_size_in_kb) {
+			dev_err(adev->dev, "Invalid %s size: 0x%x\n",
+				amdgpu_virt_dynamic_crit_table_name[AMD_SRIOV_MSG_DATAEXCHANGE_TABLE_ID],
+				init_data_hdr->dataexchange_size_in_kb);
+			r = -EINVAL;
+			goto out;
+		}
+
+		adev->virt.crit_regn_tbl[AMD_SRIOV_MSG_DATAEXCHANGE_TABLE_ID].offset =
+			init_data_hdr->dataexchange_offset;
+		adev->virt.crit_regn_tbl[AMD_SRIOV_MSG_DATAEXCHANGE_TABLE_ID].size_kb =
+			init_data_hdr->dataexchange_size_in_kb;
+	}
+
+	if (IS_SRIOV_CRIT_REGN_ENTRY_VALID(init_data_hdr, AMD_SRIOV_MSG_BAD_PAGE_INFO_TABLE_ID)) {
+		if (!init_data_hdr->bad_page_size_in_kb) {
+			dev_err(adev->dev, "Invalid %s size: 0x%x\n",
+				amdgpu_virt_dynamic_crit_table_name[AMD_SRIOV_MSG_BAD_PAGE_INFO_TABLE_ID],
+				init_data_hdr->bad_page_size_in_kb);
+			r = -EINVAL;
+			goto out;
+		}
+
+		adev->virt.crit_regn_tbl[AMD_SRIOV_MSG_BAD_PAGE_INFO_TABLE_ID].offset =
+			init_data_hdr->bad_page_info_offset;
+		adev->virt.crit_regn_tbl[AMD_SRIOV_MSG_BAD_PAGE_INFO_TABLE_ID].size_kb =
+			init_data_hdr->bad_page_size_in_kb;
+	}
+
+	/* Validation for critical region info */
+	if (adev->virt.crit_regn_tbl[AMD_SRIOV_MSG_IPD_TABLE_ID].size_kb > DISCOVERY_TMR_SIZE) {
+		dev_err(adev->dev, "Invalid IP discovery size: 0x%x\n",
+				adev->virt.crit_regn_tbl[AMD_SRIOV_MSG_IPD_TABLE_ID].size_kb);
+		r = -EINVAL;
+		goto out;
+	}
+
+	/* reserved memory starts from crit region base offset with the size of 5MB */
+	adev->mman.fw_vram_usage_start_offset = adev->virt.crit_regn.offset;
+	adev->mman.fw_vram_usage_size = adev->virt.crit_regn.size_kb << 10;
+	dev_info(adev->dev,
+		"critical region v%d requested to reserve memory start at %08llx with %llu KB.\n",
+			init_data_hdr->version,
+			adev->mman.fw_vram_usage_start_offset,
+			adev->mman.fw_vram_usage_size >> 10);
+
+	adev->virt.is_dynamic_crit_regn_enabled = true;
+
+out:
+	kfree(init_data_hdr);
+	init_data_hdr = NULL;
+
+	return r;
+}
+
+int amdgpu_virt_get_dynamic_data_info(struct amdgpu_device *adev,
+	int data_id, uint8_t *binary, u32 *size)
+{
+	uint32_t data_offset = 0;
+	uint32_t data_size = 0;
+	enum amd_sriov_msg_table_id_enum data_table_id = data_id;
+
+	if (data_table_id >= AMD_SRIOV_MSG_MAX_TABLE_ID)
+		return -EINVAL;
+
+	data_offset = adev->virt.crit_regn_tbl[data_table_id].offset;
+	data_size = adev->virt.crit_regn_tbl[data_table_id].size_kb << 10;
+
+	/* Validate on input params */
+	if (!binary || !size || *size < (uint64_t)data_size)
+		return -EINVAL;
+
+	/* Proceed to copy the dynamic content */
+	amdgpu_device_vram_access(adev,
+			(uint64_t)data_offset, (uint32_t *)binary, data_size, false);
+	*size = (uint64_t)data_size;
+
+	dev_dbg(adev->dev,
+		"Got %s info from dynamic crit_region_table at offset 0x%x with size of 0x%x bytes.\n",
+		amdgpu_virt_dynamic_crit_table_name[data_id], data_offset, data_size);
+
+	return 0;
+}
+
+void amdgpu_virt_init(struct amdgpu_device *adev)
+{
+	bool is_sriov = false;
+	uint32_t reg = amdgpu_virt_init_detect_asic(adev);
+
+	is_sriov = amdgpu_virt_init_req_data(adev, reg);
+
+	if (is_sriov)
+		amdgpu_virt_init_ras(adev);
 }
 
 static bool amdgpu_virt_access_debugfs_is_mmio(struct amdgpu_device *adev)
@@ -775,6 +1188,90 @@ enum amdgpu_sriov_vf_mode amdgpu_virt_get_sriov_vf_mode(struct amdgpu_device *ad
 	return mode;
 }
 
+void amdgpu_virt_pre_reset(struct amdgpu_device *adev)
+{
+	/* stop the data exchange thread */
+	amdgpu_virt_fini_data_exchange(adev);
+	amdgpu_dpm_set_mp1_state(adev, PP_MP1_STATE_FLR);
+}
+
+void amdgpu_virt_post_reset(struct amdgpu_device *adev)
+{
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(11, 0, 3)) {
+		/* force set to GFXOFF state after reset,
+		 * to avoid some invalid operation before GC enable
+		 */
+		adev->gfx.is_poweron = false;
+	}
+
+	adev->mes.ring[0].sched.ready = false;
+}
+
+bool amdgpu_virt_fw_load_skip_check(struct amdgpu_device *adev, uint32_t ucode_id)
+{
+	switch (amdgpu_ip_version(adev, MP0_HWIP, 0)) {
+	case IP_VERSION(13, 0, 0):
+		/* no vf autoload, white list */
+		if (ucode_id == AMDGPU_UCODE_ID_VCN1 ||
+		    ucode_id == AMDGPU_UCODE_ID_VCN)
+			return false;
+		else
+			return true;
+	case IP_VERSION(11, 0, 9):
+	case IP_VERSION(11, 0, 7):
+		/* black list for CHIP_NAVI12 and CHIP_SIENNA_CICHLID */
+		if (ucode_id == AMDGPU_UCODE_ID_RLC_G
+		    || ucode_id == AMDGPU_UCODE_ID_RLC_RESTORE_LIST_CNTL
+		    || ucode_id == AMDGPU_UCODE_ID_RLC_RESTORE_LIST_GPM_MEM
+		    || ucode_id == AMDGPU_UCODE_ID_RLC_RESTORE_LIST_SRM_MEM
+		    || ucode_id == AMDGPU_UCODE_ID_SMC)
+			return true;
+		else
+			return false;
+	case IP_VERSION(13, 0, 10):
+		/* white list */
+		if (ucode_id == AMDGPU_UCODE_ID_CAP
+		|| ucode_id == AMDGPU_UCODE_ID_CP_RS64_PFP
+		|| ucode_id == AMDGPU_UCODE_ID_CP_RS64_ME
+		|| ucode_id == AMDGPU_UCODE_ID_CP_RS64_MEC
+		|| ucode_id == AMDGPU_UCODE_ID_CP_RS64_PFP_P0_STACK
+		|| ucode_id == AMDGPU_UCODE_ID_CP_RS64_PFP_P1_STACK
+		|| ucode_id == AMDGPU_UCODE_ID_CP_RS64_ME_P0_STACK
+		|| ucode_id == AMDGPU_UCODE_ID_CP_RS64_ME_P1_STACK
+		|| ucode_id == AMDGPU_UCODE_ID_CP_RS64_MEC_P0_STACK
+		|| ucode_id == AMDGPU_UCODE_ID_CP_RS64_MEC_P1_STACK
+		|| ucode_id == AMDGPU_UCODE_ID_CP_RS64_MEC_P2_STACK
+		|| ucode_id == AMDGPU_UCODE_ID_CP_RS64_MEC_P3_STACK
+		|| ucode_id == AMDGPU_UCODE_ID_CP_MES
+		|| ucode_id == AMDGPU_UCODE_ID_CP_MES_DATA
+		|| ucode_id == AMDGPU_UCODE_ID_CP_MES1
+		|| ucode_id == AMDGPU_UCODE_ID_CP_MES1_DATA
+		|| ucode_id == AMDGPU_UCODE_ID_VCN1
+		|| ucode_id == AMDGPU_UCODE_ID_VCN)
+			return false;
+		else
+			return true;
+	default:
+		/* lagacy black list */
+		if (ucode_id == AMDGPU_UCODE_ID_SDMA0
+		    || ucode_id == AMDGPU_UCODE_ID_SDMA1
+		    || ucode_id == AMDGPU_UCODE_ID_SDMA2
+		    || ucode_id == AMDGPU_UCODE_ID_SDMA3
+		    || ucode_id == AMDGPU_UCODE_ID_SDMA4
+		    || ucode_id == AMDGPU_UCODE_ID_SDMA5
+		    || ucode_id == AMDGPU_UCODE_ID_SDMA6
+		    || ucode_id == AMDGPU_UCODE_ID_SDMA7
+		    || ucode_id == AMDGPU_UCODE_ID_RLC_G
+		    || ucode_id == AMDGPU_UCODE_ID_RLC_RESTORE_LIST_CNTL
+		    || ucode_id == AMDGPU_UCODE_ID_RLC_RESTORE_LIST_GPM_MEM
+		    || ucode_id == AMDGPU_UCODE_ID_RLC_RESTORE_LIST_SRM_MEM
+		    || ucode_id == AMDGPU_UCODE_ID_SMC)
+			return true;
+		else
+			return false;
+	}
+}
+
 void amdgpu_virt_update_sriov_video_codec(struct amdgpu_device *adev,
 			struct amdgpu_video_codec_info *encode, uint32_t encode_array_size,
 			struct amdgpu_video_codec_info *decode, uint32_t decode_array_size)
@@ -806,3 +1303,545 @@ void amdgpu_virt_update_sriov_video_codec(struct amdgpu_device *adev,
 		}
 	}
 }
+
+bool amdgpu_virt_get_rlcg_reg_access_flag(struct amdgpu_device *adev,
+						 u32 acc_flags, u32 hwip,
+						 bool write, u32 *rlcg_flag)
+{
+	bool ret = false;
+
+	switch (hwip) {
+	case GC_HWIP:
+		if (amdgpu_sriov_reg_indirect_gc(adev)) {
+			*rlcg_flag =
+				write ? AMDGPU_RLCG_GC_WRITE : AMDGPU_RLCG_GC_READ;
+			ret = true;
+		/* only in new version, AMDGPU_REGS_NO_KIQ and
+		 * AMDGPU_REGS_RLC are enabled simultaneously */
+		} else if ((acc_flags & AMDGPU_REGS_RLC) &&
+				!(acc_flags & AMDGPU_REGS_NO_KIQ) && write) {
+			*rlcg_flag = AMDGPU_RLCG_GC_WRITE_LEGACY;
+			ret = true;
+		}
+		break;
+	case MMHUB_HWIP:
+		if (amdgpu_sriov_reg_indirect_mmhub(adev) &&
+		    (acc_flags & AMDGPU_REGS_RLC) && write) {
+			*rlcg_flag = AMDGPU_RLCG_MMHUB_WRITE;
+			ret = true;
+		}
+		break;
+	default:
+		break;
+	}
+	return ret;
+}
+
+u32 amdgpu_virt_rlcg_reg_rw(struct amdgpu_device *adev, u32 offset, u32 v, u32 flag, u32 xcc_id)
+{
+	struct amdgpu_rlcg_reg_access_ctrl *reg_access_ctrl;
+	uint32_t timeout = 50000;
+	uint32_t i, tmp;
+	uint32_t ret = 0;
+	void *scratch_reg0;
+	void *scratch_reg1;
+	void *scratch_reg2;
+	void *scratch_reg3;
+	void *spare_int;
+	unsigned long flags;
+
+	if (!adev->gfx.rlc.rlcg_reg_access_supported) {
+		dev_err(adev->dev,
+			"indirect registers access through rlcg is not available\n");
+		return 0;
+	}
+
+	if (adev->gfx.xcc_mask && (((1 << xcc_id) & adev->gfx.xcc_mask) == 0)) {
+		dev_err(adev->dev, "invalid xcc\n");
+		return 0;
+	}
+
+	if (amdgpu_device_skip_hw_access(adev))
+		return 0;
+
+	reg_access_ctrl = &adev->gfx.rlc.reg_access_ctrl[xcc_id];
+	scratch_reg0 = (void __iomem *)adev->rmmio + 4 * reg_access_ctrl->scratch_reg0;
+	scratch_reg1 = (void __iomem *)adev->rmmio + 4 * reg_access_ctrl->scratch_reg1;
+	scratch_reg2 = (void __iomem *)adev->rmmio + 4 * reg_access_ctrl->scratch_reg2;
+	scratch_reg3 = (void __iomem *)adev->rmmio + 4 * reg_access_ctrl->scratch_reg3;
+
+	spin_lock_irqsave(&adev->virt.rlcg_reg_lock, flags);
+
+	if (reg_access_ctrl->spare_int)
+		spare_int = (void __iomem *)adev->rmmio + 4 * reg_access_ctrl->spare_int;
+
+	if (offset == reg_access_ctrl->grbm_cntl) {
+		/* if the target reg offset is grbm_cntl, write to scratch_reg2 */
+		writel(v, scratch_reg2);
+		if (flag == AMDGPU_RLCG_GC_WRITE_LEGACY)
+			writel(v, ((void __iomem *)adev->rmmio) + (offset * 4));
+	} else if (offset == reg_access_ctrl->grbm_idx) {
+		/* if the target reg offset is grbm_idx, write to scratch_reg3 */
+		writel(v, scratch_reg3);
+		if (flag == AMDGPU_RLCG_GC_WRITE_LEGACY)
+			writel(v, ((void __iomem *)adev->rmmio) + (offset * 4));
+	} else {
+		/*
+		 * SCRATCH_REG0 	= read/write value
+		 * SCRATCH_REG1[30:28]	= command
+		 * SCRATCH_REG1[19:0]	= address in dword
+		 * SCRATCH_REG1[27:24]	= Error reporting
+		 */
+		writel(v, scratch_reg0);
+		writel((offset | flag), scratch_reg1);
+		if (reg_access_ctrl->spare_int)
+			writel(1, spare_int);
+
+		for (i = 0; i < timeout; i++) {
+			tmp = readl(scratch_reg1);
+			if (!(tmp & AMDGPU_RLCG_SCRATCH1_ADDRESS_MASK))
+				break;
+			udelay(10);
+		}
+
+		tmp = readl(scratch_reg1);
+		if (i >= timeout || (tmp & AMDGPU_RLCG_SCRATCH1_ERROR_MASK) != 0) {
+			if (amdgpu_sriov_rlcg_error_report_enabled(adev)) {
+				if (tmp & AMDGPU_RLCG_VFGATE_DISABLED) {
+					dev_err(adev->dev,
+						"vfgate is disabled, rlcg failed to program reg: 0x%05x\n", offset);
+				} else if (tmp & AMDGPU_RLCG_WRONG_OPERATION_TYPE) {
+					dev_err(adev->dev,
+						"wrong operation type, rlcg failed to program reg: 0x%05x\n", offset);
+				} else if (tmp & AMDGPU_RLCG_REG_NOT_IN_RANGE) {
+					dev_err(adev->dev,
+						"register is not in range, rlcg failed to program reg: 0x%05x\n", offset);
+				} else {
+					dev_err(adev->dev,
+						"unknown error type, rlcg failed to program reg: 0x%05x\n", offset);
+				}
+			} else {
+				dev_err(adev->dev,
+					"timeout: rlcg faled to program reg: 0x%05x\n", offset);
+			}
+		}
+	}
+
+	ret = readl(scratch_reg0);
+
+	spin_unlock_irqrestore(&adev->virt.rlcg_reg_lock, flags);
+
+	return ret;
+}
+
+void amdgpu_sriov_wreg(struct amdgpu_device *adev,
+		       u32 offset, u32 value,
+		       u32 acc_flags, u32 hwip, u32 xcc_id)
+{
+	u32 rlcg_flag;
+
+	if (amdgpu_device_skip_hw_access(adev))
+		return;
+
+	if (!amdgpu_sriov_runtime(adev) &&
+		amdgpu_virt_get_rlcg_reg_access_flag(adev, acc_flags, hwip, true, &rlcg_flag)) {
+		amdgpu_virt_rlcg_reg_rw(adev, offset, value, rlcg_flag, xcc_id);
+		return;
+	}
+
+	if (acc_flags & AMDGPU_REGS_NO_KIQ)
+		WREG32_NO_KIQ(offset, value);
+	else
+		WREG32(offset, value);
+}
+
+u32 amdgpu_sriov_rreg(struct amdgpu_device *adev,
+		      u32 offset, u32 acc_flags, u32 hwip, u32 xcc_id)
+{
+	u32 rlcg_flag;
+
+	if (amdgpu_device_skip_hw_access(adev))
+		return 0;
+
+	if (!amdgpu_sriov_runtime(adev) &&
+		amdgpu_virt_get_rlcg_reg_access_flag(adev, acc_flags, hwip, false, &rlcg_flag))
+		return amdgpu_virt_rlcg_reg_rw(adev, offset, 0, rlcg_flag, xcc_id);
+
+	if (acc_flags & AMDGPU_REGS_NO_KIQ)
+		return RREG32_NO_KIQ(offset);
+	else
+		return RREG32(offset);
+}
+
+bool amdgpu_sriov_xnack_support(struct amdgpu_device *adev)
+{
+	bool xnack_mode = true;
+
+	if (amdgpu_sriov_vf(adev) &&
+	    amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 2))
+		xnack_mode = false;
+
+	return xnack_mode;
+}
+
+bool amdgpu_virt_get_ras_capability(struct amdgpu_device *adev)
+{
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+
+	if (!amdgpu_sriov_ras_caps_en(adev))
+		return false;
+
+	if (adev->virt.ras_en_caps.bits.block_umc)
+		adev->ras_hw_enabled |= BIT(AMDGPU_RAS_BLOCK__UMC);
+	if (adev->virt.ras_en_caps.bits.block_sdma)
+		adev->ras_hw_enabled |= BIT(AMDGPU_RAS_BLOCK__SDMA);
+	if (adev->virt.ras_en_caps.bits.block_gfx)
+		adev->ras_hw_enabled |= BIT(AMDGPU_RAS_BLOCK__GFX);
+	if (adev->virt.ras_en_caps.bits.block_mmhub)
+		adev->ras_hw_enabled |= BIT(AMDGPU_RAS_BLOCK__MMHUB);
+	if (adev->virt.ras_en_caps.bits.block_athub)
+		adev->ras_hw_enabled |= BIT(AMDGPU_RAS_BLOCK__ATHUB);
+	if (adev->virt.ras_en_caps.bits.block_pcie_bif)
+		adev->ras_hw_enabled |= BIT(AMDGPU_RAS_BLOCK__PCIE_BIF);
+	if (adev->virt.ras_en_caps.bits.block_hdp)
+		adev->ras_hw_enabled |= BIT(AMDGPU_RAS_BLOCK__HDP);
+	if (adev->virt.ras_en_caps.bits.block_xgmi_wafl)
+		adev->ras_hw_enabled |= BIT(AMDGPU_RAS_BLOCK__XGMI_WAFL);
+	if (adev->virt.ras_en_caps.bits.block_df)
+		adev->ras_hw_enabled |= BIT(AMDGPU_RAS_BLOCK__DF);
+	if (adev->virt.ras_en_caps.bits.block_smn)
+		adev->ras_hw_enabled |= BIT(AMDGPU_RAS_BLOCK__SMN);
+	if (adev->virt.ras_en_caps.bits.block_sem)
+		adev->ras_hw_enabled |= BIT(AMDGPU_RAS_BLOCK__SEM);
+	if (adev->virt.ras_en_caps.bits.block_mp0)
+		adev->ras_hw_enabled |= BIT(AMDGPU_RAS_BLOCK__MP0);
+	if (adev->virt.ras_en_caps.bits.block_mp1)
+		adev->ras_hw_enabled |= BIT(AMDGPU_RAS_BLOCK__MP1);
+	if (adev->virt.ras_en_caps.bits.block_fuse)
+		adev->ras_hw_enabled |= BIT(AMDGPU_RAS_BLOCK__FUSE);
+	if (adev->virt.ras_en_caps.bits.block_mca)
+		adev->ras_hw_enabled |= BIT(AMDGPU_RAS_BLOCK__MCA);
+	if (adev->virt.ras_en_caps.bits.block_vcn)
+		adev->ras_hw_enabled |= BIT(AMDGPU_RAS_BLOCK__VCN);
+	if (adev->virt.ras_en_caps.bits.block_jpeg)
+		adev->ras_hw_enabled |= BIT(AMDGPU_RAS_BLOCK__JPEG);
+	if (adev->virt.ras_en_caps.bits.block_ih)
+		adev->ras_hw_enabled |= BIT(AMDGPU_RAS_BLOCK__IH);
+	if (adev->virt.ras_en_caps.bits.block_mpio)
+		adev->ras_hw_enabled |= BIT(AMDGPU_RAS_BLOCK__MPIO);
+
+	if (adev->virt.ras_en_caps.bits.poison_propogation_mode)
+		con->poison_supported = true; /* Poison is handled by host */
+
+	return true;
+}
+
+static inline enum amd_sriov_ras_telemetry_gpu_block
+amdgpu_ras_block_to_sriov(struct amdgpu_device *adev, enum amdgpu_ras_block block) {
+	switch (block) {
+	case AMDGPU_RAS_BLOCK__UMC:
+		return RAS_TELEMETRY_GPU_BLOCK_UMC;
+	case AMDGPU_RAS_BLOCK__SDMA:
+		return RAS_TELEMETRY_GPU_BLOCK_SDMA;
+	case AMDGPU_RAS_BLOCK__GFX:
+		return RAS_TELEMETRY_GPU_BLOCK_GFX;
+	case AMDGPU_RAS_BLOCK__MMHUB:
+		return RAS_TELEMETRY_GPU_BLOCK_MMHUB;
+	case AMDGPU_RAS_BLOCK__ATHUB:
+		return RAS_TELEMETRY_GPU_BLOCK_ATHUB;
+	case AMDGPU_RAS_BLOCK__PCIE_BIF:
+		return RAS_TELEMETRY_GPU_BLOCK_PCIE_BIF;
+	case AMDGPU_RAS_BLOCK__HDP:
+		return RAS_TELEMETRY_GPU_BLOCK_HDP;
+	case AMDGPU_RAS_BLOCK__XGMI_WAFL:
+		return RAS_TELEMETRY_GPU_BLOCK_XGMI_WAFL;
+	case AMDGPU_RAS_BLOCK__DF:
+		return RAS_TELEMETRY_GPU_BLOCK_DF;
+	case AMDGPU_RAS_BLOCK__SMN:
+		return RAS_TELEMETRY_GPU_BLOCK_SMN;
+	case AMDGPU_RAS_BLOCK__SEM:
+		return RAS_TELEMETRY_GPU_BLOCK_SEM;
+	case AMDGPU_RAS_BLOCK__MP0:
+		return RAS_TELEMETRY_GPU_BLOCK_MP0;
+	case AMDGPU_RAS_BLOCK__MP1:
+		return RAS_TELEMETRY_GPU_BLOCK_MP1;
+	case AMDGPU_RAS_BLOCK__FUSE:
+		return RAS_TELEMETRY_GPU_BLOCK_FUSE;
+	case AMDGPU_RAS_BLOCK__MCA:
+		return RAS_TELEMETRY_GPU_BLOCK_MCA;
+	case AMDGPU_RAS_BLOCK__VCN:
+		return RAS_TELEMETRY_GPU_BLOCK_VCN;
+	case AMDGPU_RAS_BLOCK__JPEG:
+		return RAS_TELEMETRY_GPU_BLOCK_JPEG;
+	case AMDGPU_RAS_BLOCK__IH:
+		return RAS_TELEMETRY_GPU_BLOCK_IH;
+	case AMDGPU_RAS_BLOCK__MPIO:
+		return RAS_TELEMETRY_GPU_BLOCK_MPIO;
+	default:
+		dev_warn(adev->dev, "Unsupported SRIOV RAS telemetry block 0x%x\n",
+			      block);
+		return RAS_TELEMETRY_GPU_BLOCK_COUNT;
+	}
+}
+
+static int amdgpu_virt_cache_host_error_counts(struct amdgpu_device *adev,
+					       struct amdsriov_ras_telemetry *host_telemetry)
+{
+	struct amd_sriov_ras_telemetry_error_count *tmp = NULL;
+	uint32_t checksum, used_size;
+
+	checksum = host_telemetry->header.checksum;
+	used_size = host_telemetry->header.used_size;
+
+	if (used_size > (AMD_SRIOV_MSG_RAS_TELEMETRY_SIZE_KB_V1 << 10))
+		return 0;
+
+	tmp = kmemdup(&host_telemetry->body.error_count, used_size, GFP_KERNEL);
+	if (!tmp)
+		return -ENOMEM;
+
+	if (checksum != amd_sriov_msg_checksum(tmp, used_size, 0, 0))
+		goto out;
+
+	memcpy(&adev->virt.count_cache, tmp,
+	       min(used_size, sizeof(adev->virt.count_cache)));
+out:
+	kfree(tmp);
+
+	return 0;
+}
+
+static int amdgpu_virt_req_ras_err_count_internal(struct amdgpu_device *adev, bool force_update)
+{
+	struct amdgpu_virt *virt = &adev->virt;
+
+	if (!virt->ops || !virt->ops->req_ras_err_count)
+		return -EOPNOTSUPP;
+
+	/* Host allows 15 ras telemetry requests per 60 seconds. Afterwhich, the Host
+	 * will ignore incoming guest messages. Ratelimit the guest messages to
+	 * prevent guest self DOS.
+	 */
+	if (__ratelimit(&virt->ras.ras_error_cnt_rs) || force_update) {
+		mutex_lock(&virt->ras.ras_telemetry_mutex);
+		if (!virt->ops->req_ras_err_count(adev))
+			amdgpu_virt_cache_host_error_counts(adev,
+				virt->fw_reserve.ras_telemetry);
+		mutex_unlock(&virt->ras.ras_telemetry_mutex);
+	}
+
+	return 0;
+}
+
+/* Bypass ACA interface and query ECC counts directly from host */
+int amdgpu_virt_req_ras_err_count(struct amdgpu_device *adev, enum amdgpu_ras_block block,
+				  struct ras_err_data *err_data)
+{
+	enum amd_sriov_ras_telemetry_gpu_block sriov_block;
+
+	sriov_block = amdgpu_ras_block_to_sriov(adev, block);
+
+	if (sriov_block >= RAS_TELEMETRY_GPU_BLOCK_COUNT ||
+	    !amdgpu_sriov_ras_telemetry_block_en(adev, sriov_block))
+		return -EOPNOTSUPP;
+
+	/* Host Access may be lost during reset, just return last cached data. */
+	if (down_read_trylock(&adev->reset_domain->sem)) {
+		amdgpu_virt_req_ras_err_count_internal(adev, false);
+		up_read(&adev->reset_domain->sem);
+	}
+
+	err_data->ue_count = adev->virt.count_cache.block[sriov_block].ue_count;
+	err_data->ce_count = adev->virt.count_cache.block[sriov_block].ce_count;
+	err_data->de_count = adev->virt.count_cache.block[sriov_block].de_count;
+
+	return 0;
+}
+
+static int
+amdgpu_virt_write_cpers_to_ring(struct amdgpu_device *adev,
+				struct amdsriov_ras_telemetry *host_telemetry,
+				u32 *more)
+{
+	struct amd_sriov_ras_cper_dump *cper_dump = NULL;
+	struct cper_hdr *entry = NULL;
+	struct amdgpu_ring *ring = &adev->cper.ring_buf;
+	uint32_t checksum, used_size, i;
+	int ret = 0;
+
+	checksum = host_telemetry->header.checksum;
+	used_size = host_telemetry->header.used_size;
+
+	if (used_size > (AMD_SRIOV_MSG_RAS_TELEMETRY_SIZE_KB_V1 << 10))
+		return -EINVAL;
+
+	cper_dump = kmemdup(&host_telemetry->body.cper_dump, used_size, GFP_KERNEL);
+	if (!cper_dump)
+		return -ENOMEM;
+
+	if (checksum != amd_sriov_msg_checksum(cper_dump, used_size, 0, 0)) {
+		ret = -EINVAL;
+		goto out;
+	}
+
+	*more = cper_dump->more;
+
+	if (cper_dump->wptr < adev->virt.ras.cper_rptr) {
+		dev_warn(
+			adev->dev,
+			"guest specified rptr that was too high! guest rptr: 0x%llx, host rptr: 0x%llx\n",
+			adev->virt.ras.cper_rptr, cper_dump->wptr);
+
+		adev->virt.ras.cper_rptr = cper_dump->wptr;
+		goto out;
+	}
+
+	entry = (struct cper_hdr *)&cper_dump->buf[0];
+
+	for (i = 0; i < cper_dump->count; i++) {
+		amdgpu_cper_ring_write(ring, entry, entry->record_length);
+		entry = (struct cper_hdr *)((char *)entry +
+					    entry->record_length);
+	}
+
+	if (cper_dump->overflow_count)
+		dev_warn(adev->dev,
+			 "host reported CPER overflow of 0x%llx entries!\n",
+			 cper_dump->overflow_count);
+
+	adev->virt.ras.cper_rptr = cper_dump->wptr;
+out:
+	kfree(cper_dump);
+
+	return ret;
+}
+
+static int amdgpu_virt_req_ras_cper_dump_internal(struct amdgpu_device *adev)
+{
+	struct amdgpu_virt *virt = &adev->virt;
+	int ret = 0;
+	uint32_t more = 0;
+
+	if (!virt->ops || !virt->ops->req_ras_cper_dump)
+		return -EOPNOTSUPP;
+
+	do {
+		if (!virt->ops->req_ras_cper_dump(adev, virt->ras.cper_rptr))
+			ret = amdgpu_virt_write_cpers_to_ring(
+				adev, virt->fw_reserve.ras_telemetry, &more);
+		else
+			ret = 0;
+	} while (more && !ret);
+
+	return ret;
+}
+
+int amdgpu_virt_req_ras_cper_dump(struct amdgpu_device *adev, bool force_update)
+{
+	struct amdgpu_virt *virt = &adev->virt;
+	int ret = 0;
+
+	if (!amdgpu_sriov_ras_cper_en(adev))
+		return -EOPNOTSUPP;
+
+	if ((__ratelimit(&virt->ras.ras_cper_dump_rs) || force_update) &&
+	    down_read_trylock(&adev->reset_domain->sem)) {
+		mutex_lock(&virt->ras.ras_telemetry_mutex);
+		ret = amdgpu_virt_req_ras_cper_dump_internal(adev);
+		mutex_unlock(&virt->ras.ras_telemetry_mutex);
+		up_read(&adev->reset_domain->sem);
+	}
+
+	return ret;
+}
+
+int amdgpu_virt_ras_telemetry_post_reset(struct amdgpu_device *adev)
+{
+	unsigned long ue_count, ce_count;
+
+	if (amdgpu_sriov_ras_telemetry_en(adev)) {
+		amdgpu_virt_req_ras_err_count_internal(adev, true);
+		amdgpu_ras_query_error_count(adev, &ce_count, &ue_count, NULL);
+	}
+
+	return 0;
+}
+
+bool amdgpu_virt_ras_telemetry_block_en(struct amdgpu_device *adev,
+					enum amdgpu_ras_block block)
+{
+	enum amd_sriov_ras_telemetry_gpu_block sriov_block;
+
+	sriov_block = amdgpu_ras_block_to_sriov(adev, block);
+
+	if (sriov_block >= RAS_TELEMETRY_GPU_BLOCK_COUNT ||
+	    !amdgpu_sriov_ras_telemetry_block_en(adev, sriov_block))
+		return false;
+
+	return true;
+}
+
+/*
+ * amdgpu_virt_request_bad_pages() - request bad pages
+ * @adev: amdgpu device.
+ * Send command to GPU hypervisor to write new bad pages into the shared PF2VF region
+ */
+void amdgpu_virt_request_bad_pages(struct amdgpu_device *adev)
+{
+	struct amdgpu_virt *virt = &adev->virt;
+
+	if (virt->ops && virt->ops->req_bad_pages)
+		virt->ops->req_bad_pages(adev);
+}
+
+static int amdgpu_virt_cache_chk_criti_hit(struct amdgpu_device *adev,
+					   struct amdsriov_ras_telemetry *host_telemetry,
+					   bool *hit)
+{
+	struct amd_sriov_ras_chk_criti *tmp = NULL;
+	uint32_t checksum, used_size;
+
+	checksum = host_telemetry->header.checksum;
+	used_size = host_telemetry->header.used_size;
+
+	if (used_size > (AMD_SRIOV_MSG_RAS_TELEMETRY_SIZE_KB_V1 << 10))
+		return 0;
+
+	tmp = kmemdup(&host_telemetry->body.chk_criti, used_size, GFP_KERNEL);
+	if (!tmp)
+		return -ENOMEM;
+
+	if (checksum != amd_sriov_msg_checksum(tmp, used_size, 0, 0))
+		goto out;
+
+	if (hit)
+		*hit = tmp->hit ? true : false;
+
+out:
+	kfree(tmp);
+
+	return 0;
+}
+
+int amdgpu_virt_check_vf_critical_region(struct amdgpu_device *adev, u64 addr, bool *hit)
+{
+	struct amdgpu_virt *virt = &adev->virt;
+	int r = -EPERM;
+
+	if (!virt->ops || !virt->ops->req_ras_chk_criti)
+		return -EOPNOTSUPP;
+
+	/* Host allows 15 ras telemetry requests per 60 seconds. Afterwhich, the Host
+	 * will ignore incoming guest messages. Ratelimit the guest messages to
+	 * prevent guest self DOS.
+	 */
+	if (__ratelimit(&virt->ras.ras_chk_criti_rs)) {
+		mutex_lock(&virt->ras.ras_telemetry_mutex);
+		if (!virt->ops->req_ras_chk_criti(adev, addr))
+			r = amdgpu_virt_cache_chk_criti_hit(
+				adev, virt->fw_reserve.ras_telemetry, hit);
+		mutex_unlock(&virt->ras.ras_telemetry_mutex);
+	}
+
+	return r;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h
index 8d4c20bb71c5..01d5bca2dee1 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h
@@ -31,12 +31,35 @@
 #define AMDGPU_SRIOV_CAPS_IS_VF        (1 << 2) /* this GPU is a virtual function */
 #define AMDGPU_PASSTHROUGH_MODE        (1 << 3) /* thw whole GPU is pass through for VM */
 #define AMDGPU_SRIOV_CAPS_RUNTIME      (1 << 4) /* is out of full access mode */
+#define AMDGPU_VF_MMIO_ACCESS_PROTECT  (1 << 5) /* MMIO write access is not allowed in sriov runtime */
+
+/* flags for indirect register access path supported by rlcg for sriov */
+#define AMDGPU_RLCG_GC_WRITE_LEGACY    (0x8 << 28)
+#define AMDGPU_RLCG_GC_WRITE           (0x0 << 28)
+#define AMDGPU_RLCG_GC_READ            (0x1 << 28)
+#define AMDGPU_RLCG_MMHUB_WRITE        (0x2 << 28)
+
+/* error code for indirect register access path supported by rlcg for sriov */
+#define AMDGPU_RLCG_VFGATE_DISABLED		0x4000000
+#define AMDGPU_RLCG_WRONG_OPERATION_TYPE	0x2000000
+#define AMDGPU_RLCG_REG_NOT_IN_RANGE		0x1000000
+
+#define AMDGPU_RLCG_SCRATCH1_ADDRESS_MASK	0xFFFFF
+#define AMDGPU_RLCG_SCRATCH1_ERROR_MASK	0xF000000
 
 /* all asic after AI use this offset */
 #define mmRCC_IOV_FUNC_IDENTIFIER 0xDE5
 /* tonga/fiji use this offset */
 #define mmBIF_IOV_FUNC_IDENTIFIER 0x1503
 
+#define AMDGPU_VF2PF_UPDATE_MAX_RETRY_LIMIT 2
+
+/* Signature used to validate the SR-IOV dynamic critical region init data header ("INDA") */
+#define AMDGPU_SRIOV_CRIT_DATA_SIGNATURE "INDA"
+#define AMDGPU_SRIOV_CRIT_DATA_SIG_LEN   4
+
+#define IS_SRIOV_CRIT_REGN_ENTRY_VALID(hdr, id) ((hdr)->valid_tables & (1 << (id)))
+
 enum amdgpu_sriov_vf_mode {
 	SRIOV_VF_MODE_BARE_METAL = 0,
 	SRIOV_VF_MODE_ONE_VF,
@@ -61,6 +84,8 @@ struct amdgpu_vf_error_buffer {
 	uint64_t data[AMDGPU_VF_ERROR_ENTRY_SIZE];
 };
 
+enum idh_request;
+
 /**
  * struct amdgpu_virt_ops - amdgpu device virt operations
  */
@@ -69,8 +94,17 @@ struct amdgpu_virt_ops {
 	int (*rel_full_gpu)(struct amdgpu_device *adev, bool init);
 	int (*req_init_data)(struct amdgpu_device *adev);
 	int (*reset_gpu)(struct amdgpu_device *adev);
+	void (*ready_to_reset)(struct amdgpu_device *adev);
 	int (*wait_reset)(struct amdgpu_device *adev);
-	void (*trans_msg)(struct amdgpu_device *adev, u32 req, u32 data1, u32 data2, u32 data3);
+	void (*trans_msg)(struct amdgpu_device *adev, enum idh_request req,
+			  u32 data1, u32 data2, u32 data3);
+	void (*ras_poison_handler)(struct amdgpu_device *adev,
+					enum amdgpu_ras_block block);
+	bool (*rcvd_ras_intr)(struct amdgpu_device *adev);
+	int (*req_ras_err_count)(struct amdgpu_device *adev);
+	int (*req_ras_cper_dump)(struct amdgpu_device *adev, u64 vf_rptr);
+	int (*req_bad_pages)(struct amdgpu_device *adev);
+	int (*req_ras_chk_criti)(struct amdgpu_device *adev, u64 addr);
 };
 
 /*
@@ -79,6 +113,7 @@ struct amdgpu_virt_ops {
 struct amdgpu_virt_fw_reserve {
 	struct amd_sriov_msg_pf2vf_info_header *p_pf2vf;
 	struct amd_sriov_msg_vf2pf_info_header *p_vf2pf;
+	void *ras_telemetry;
 	unsigned int checksum_key;
 };
 
@@ -106,15 +141,29 @@ enum AMDGIM_FEATURE_FLAG {
 	AMDGIM_FEATURE_PP_ONE_VF = (1 << 4),
 	/* Indirect Reg Access enabled */
 	AMDGIM_FEATURE_INDIRECT_REG_ACCESS = (1 << 5),
+	/* AV1 Support MODE*/
+	AMDGIM_FEATURE_AV1_SUPPORT = (1 << 6),
+	/* VCN RB decouple */
+	AMDGIM_FEATURE_VCN_RB_DECOUPLE = (1 << 7),
+	/* MES info */
+	AMDGIM_FEATURE_MES_INFO_ENABLE = (1 << 8),
+	AMDGIM_FEATURE_RAS_CAPS = (1 << 9),
+	AMDGIM_FEATURE_RAS_TELEMETRY = (1 << 10),
+	AMDGIM_FEATURE_RAS_CPER = (1 << 11),
+	AMDGIM_FEATURE_XGMI_TA_EXT_PEER_LINK = (1 << 12),
 };
 
 enum AMDGIM_REG_ACCESS_FLAG {
 	/* Use PSP to program IH_RB_CNTL */
-	AMDGIM_FEATURE_IH_REG_PSP_EN     = (1 << 0),
+	AMDGIM_FEATURE_IH_REG_PSP_EN      = (1 << 0),
 	/* Use RLC to program MMHUB regs */
-	AMDGIM_FEATURE_MMHUB_REG_RLC_EN  = (1 << 1),
+	AMDGIM_FEATURE_MMHUB_REG_RLC_EN   = (1 << 1),
 	/* Use RLC to program GC regs */
-	AMDGIM_FEATURE_GC_REG_RLC_EN     = (1 << 2),
+	AMDGIM_FEATURE_GC_REG_RLC_EN      = (1 << 2),
+	/* Use PSP to program L1_TLB_CNTL */
+	AMDGIM_FEATURE_L1_TLB_CNTL_PSP_EN = (1 << 3),
+	/* Use RLCG to program SQ_CONFIG1 */
+	AMDGIM_FEATURE_REG_ACCESS_SQ_CONFIG = (1 << 4),
 };
 
 struct amdgim_pf2vf_info_v1 {
@@ -208,6 +257,23 @@ struct amdgpu_virt_ras_err_handler_data {
 	int last_reserved;
 };
 
+struct amdgpu_virt_ras {
+	struct ratelimit_state ras_error_cnt_rs;
+	struct ratelimit_state ras_cper_dump_rs;
+	struct ratelimit_state ras_chk_criti_rs;
+	struct mutex ras_telemetry_mutex;
+	uint64_t cper_rptr;
+};
+
+#define AMDGPU_VIRT_CAPS_LIST(X) X(AMDGPU_VIRT_CAP_POWER_LIMIT)
+
+DECLARE_ATTR_CAP_CLASS(amdgpu_virt, AMDGPU_VIRT_CAPS_LIST);
+
+struct amdgpu_virt_region {
+	uint32_t offset;
+	uint32_t size_kb;
+};
+
 /* GPU virtualization */
 struct amdgpu_virt {
 	uint32_t			caps;
@@ -217,11 +283,16 @@ struct amdgpu_virt {
 	uint32_t			reg_val_offs;
 	struct amdgpu_irq_src		ack_irq;
 	struct amdgpu_irq_src		rcv_irq;
+
 	struct work_struct		flr_work;
+	struct work_struct		req_bad_pages_work;
+	struct work_struct		handle_bad_pages_work;
+
 	struct amdgpu_mm_table		mm_table;
 	const struct amdgpu_virt_ops	*ops;
 	struct amdgpu_vf_error_buffer	vf_errors;
 	struct amdgpu_virt_fw_reserve	fw_reserve;
+	struct amdgpu_virt_caps virt_caps;
 	uint32_t gim_feature;
 	uint32_t reg_access_mode;
 	int req_init_data_ver;
@@ -230,9 +301,16 @@ struct amdgpu_virt {
 	bool ras_init_done;
 	uint32_t reg_access;
 
+	/* dynamic(v2) critical regions */
+	struct amdgpu_virt_region init_data_header;
+	struct amdgpu_virt_region crit_regn;
+	struct amdgpu_virt_region crit_regn_tbl[AMD_SRIOV_MSG_MAX_TABLE_ID];
+	bool is_dynamic_crit_regn_enabled;
+
 	/* vf2pf message */
 	struct delayed_work vf2pf_work;
 	uint32_t vf2pf_update_interval_ms;
+	int vf2pf_update_retry_cnt;
 
 	/* multimedia bandwidth config */
 	bool     is_mm_bw_enabled;
@@ -240,6 +318,22 @@ struct amdgpu_virt {
 	uint32_t decode_max_frame_pixels;
 	uint32_t encode_max_dimension_pixels;
 	uint32_t encode_max_frame_pixels;
+
+	/* the ucode id to signal the autoload */
+	uint32_t autoload_ucode_id;
+
+	/* Spinlock to protect access to the RLCG register interface */
+	spinlock_t rlcg_reg_lock;
+
+	struct mutex access_req_mutex;
+
+	union amd_sriov_ras_caps ras_en_caps;
+	union amd_sriov_ras_caps ras_telemetry_en_caps;
+	struct amdgpu_virt_ras ras;
+	struct amd_sriov_ras_telemetry_error_count count_cache;
+
+	/* hibernate and resume with different VF feature for xgmi enabled system */
+	bool is_xgmi_node_migrate_enabled;
 };
 
 struct amdgpu_video_codec_info;
@@ -275,13 +369,44 @@ struct amdgpu_video_codec_info;
 (amdgpu_sriov_vf((adev)) && \
 	((adev)->virt.reg_access & (AMDGIM_FEATURE_GC_REG_RLC_EN)))
 
+#define amdgpu_sriov_reg_indirect_l1_tlb_cntl(adev) \
+(amdgpu_sriov_vf((adev)) && \
+	((adev)->virt.reg_access & (AMDGIM_FEATURE_L1_TLB_CNTL_PSP_EN)))
+
+#define amdgpu_sriov_rlcg_error_report_enabled(adev) \
+        (amdgpu_sriov_reg_indirect_mmhub(adev) || amdgpu_sriov_reg_indirect_gc(adev))
+
+#define amdgpu_sriov_reg_access_sq_config(adev) \
+(amdgpu_sriov_vf((adev)) && \
+	((adev)->virt.reg_access & (AMDGIM_FEATURE_REG_ACCESS_SQ_CONFIG)))
+
 #define amdgpu_passthrough(adev) \
 ((adev)->virt.caps & AMDGPU_PASSTHROUGH_MODE)
 
+#define amdgpu_sriov_vf_mmio_access_protection(adev) \
+((adev)->virt.caps & AMDGPU_VF_MMIO_ACCESS_PROTECT)
+
+#define amdgpu_sriov_ras_caps_en(adev) \
+((adev)->virt.gim_feature & AMDGIM_FEATURE_RAS_CAPS)
+
+#define amdgpu_sriov_ras_telemetry_en(adev) \
+(((adev)->virt.gim_feature & AMDGIM_FEATURE_RAS_TELEMETRY) && (adev)->virt.fw_reserve.ras_telemetry)
+
+#define amdgpu_sriov_ras_telemetry_block_en(adev, sriov_blk) \
+(amdgpu_sriov_ras_telemetry_en((adev)) && (adev)->virt.ras_telemetry_en_caps.all & BIT(sriov_blk))
+
+#define amdgpu_sriov_ras_cper_en(adev) \
+((adev)->virt.gim_feature & AMDGIM_FEATURE_RAS_CPER)
+
+#define amdgpu_sriov_xgmi_ta_ext_peer_link_en(adev) \
+((adev)->virt.gim_feature & AMDGIM_FEATURE_XGMI_TA_EXT_PEER_LINK)
+
 static inline bool is_virtual_machine(void)
 {
-#ifdef CONFIG_X86
+#if defined(CONFIG_X86)
 	return boot_cpu_has(X86_FEATURE_HYPERVISOR);
+#elif defined(CONFIG_ARM64)
+	return !is_kernel_in_hyp_mode();
 #else
 	return false;
 #endif
@@ -289,27 +414,42 @@ static inline bool is_virtual_machine(void)
 
 #define amdgpu_sriov_is_pp_one_vf(adev) \
 	((adev)->virt.gim_feature & AMDGIM_FEATURE_PP_ONE_VF)
+#define amdgpu_sriov_multi_vf_mode(adev) \
+	(amdgpu_sriov_vf(adev) && !amdgpu_sriov_is_pp_one_vf(adev))
 #define amdgpu_sriov_is_debug(adev) \
 	((!amdgpu_in_reset(adev)) && adev->virt.tdr_debug)
 #define amdgpu_sriov_is_normal(adev) \
 	((!amdgpu_in_reset(adev)) && (!adev->virt.tdr_debug))
+#define amdgpu_sriov_is_av1_support(adev) \
+	((adev)->virt.gim_feature & AMDGIM_FEATURE_AV1_SUPPORT)
+#define amdgpu_sriov_is_vcn_rb_decouple(adev) \
+	((adev)->virt.gim_feature & AMDGIM_FEATURE_VCN_RB_DECOUPLE)
+#define amdgpu_sriov_is_mes_info_enable(adev) \
+	((adev)->virt.gim_feature & AMDGIM_FEATURE_MES_INFO_ENABLE)
+
+#define amdgpu_virt_xgmi_migrate_enabled(adev) \
+	((adev)->virt.is_xgmi_node_migrate_enabled && (adev)->gmc.xgmi.node_segment_size != 0)
 
 bool amdgpu_virt_mmio_blocked(struct amdgpu_device *adev);
 void amdgpu_virt_init_setting(struct amdgpu_device *adev);
-void amdgpu_virt_kiq_reg_write_reg_wait(struct amdgpu_device *adev,
-					uint32_t reg0, uint32_t rreg1,
-					uint32_t ref, uint32_t mask);
 int amdgpu_virt_request_full_gpu(struct amdgpu_device *adev, bool init);
 int amdgpu_virt_release_full_gpu(struct amdgpu_device *adev, bool init);
 int amdgpu_virt_reset_gpu(struct amdgpu_device *adev);
 void amdgpu_virt_request_init_data(struct amdgpu_device *adev);
+void amdgpu_virt_ready_to_reset(struct amdgpu_device *adev);
 int amdgpu_virt_wait_reset(struct amdgpu_device *adev);
 int amdgpu_virt_alloc_mm_table(struct amdgpu_device *adev);
 void amdgpu_virt_free_mm_table(struct amdgpu_device *adev);
+bool amdgpu_virt_rcvd_ras_interrupt(struct amdgpu_device *adev);
 void amdgpu_virt_release_ras_err_handler_data(struct amdgpu_device *adev);
 void amdgpu_virt_init_data_exchange(struct amdgpu_device *adev);
+void amdgpu_virt_exchange_data(struct amdgpu_device *adev);
 void amdgpu_virt_fini_data_exchange(struct amdgpu_device *adev);
-void amdgpu_detect_virtualization(struct amdgpu_device *adev);
+void amdgpu_virt_init(struct amdgpu_device *adev);
+
+int amdgpu_virt_init_critical_region(struct amdgpu_device *adev);
+int amdgpu_virt_get_dynamic_data_info(struct amdgpu_device *adev,
+	int data_id, uint8_t *binary, u32 *size);
 
 bool amdgpu_virt_can_access_debugfs(struct amdgpu_device *adev);
 int amdgpu_virt_enable_access_debugfs(struct amdgpu_device *adev);
@@ -320,4 +460,27 @@ enum amdgpu_sriov_vf_mode amdgpu_virt_get_sriov_vf_mode(struct amdgpu_device *ad
 void amdgpu_virt_update_sriov_video_codec(struct amdgpu_device *adev,
 			struct amdgpu_video_codec_info *encode, uint32_t encode_array_size,
 			struct amdgpu_video_codec_info *decode, uint32_t decode_array_size);
+void amdgpu_sriov_wreg(struct amdgpu_device *adev,
+		       u32 offset, u32 value,
+		       u32 acc_flags, u32 hwip, u32 xcc_id);
+u32 amdgpu_sriov_rreg(struct amdgpu_device *adev,
+		      u32 offset, u32 acc_flags, u32 hwip, u32 xcc_id);
+bool amdgpu_virt_fw_load_skip_check(struct amdgpu_device *adev,
+			uint32_t ucode_id);
+void amdgpu_virt_pre_reset(struct amdgpu_device *adev);
+void amdgpu_virt_post_reset(struct amdgpu_device *adev);
+bool amdgpu_sriov_xnack_support(struct amdgpu_device *adev);
+bool amdgpu_virt_get_rlcg_reg_access_flag(struct amdgpu_device *adev,
+					  u32 acc_flags, u32 hwip,
+					  bool write, u32 *rlcg_flag);
+u32 amdgpu_virt_rlcg_reg_rw(struct amdgpu_device *adev, u32 offset, u32 v, u32 flag, u32 xcc_id);
+bool amdgpu_virt_get_ras_capability(struct amdgpu_device *adev);
+int amdgpu_virt_req_ras_err_count(struct amdgpu_device *adev, enum amdgpu_ras_block block,
+				  struct ras_err_data *err_data);
+int amdgpu_virt_req_ras_cper_dump(struct amdgpu_device *adev, bool force_update);
+int amdgpu_virt_ras_telemetry_post_reset(struct amdgpu_device *adev);
+bool amdgpu_virt_ras_telemetry_block_en(struct amdgpu_device *adev,
+					enum amdgpu_ras_block block);
+void amdgpu_virt_request_bad_pages(struct amdgpu_device *adev);
+int amdgpu_virt_check_vf_critical_region(struct amdgpu_device *adev, u64 addr, bool *hit);
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vkms.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vkms.c
index ac9a8cd21c4b..79bad9cbe2ab 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vkms.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vkms.c
@@ -1,7 +1,9 @@
 // SPDX-License-Identifier: GPL-2.0+
 
 #include <drm/drm_atomic_helper.h>
+#include <drm/drm_edid.h>
 #include <drm/drm_simple_kms_helper.h>
+#include <drm/drm_gem_framebuffer_helper.h>
 #include <drm/drm_vblank.h>
 
 #include "amdgpu.h"
@@ -12,10 +14,11 @@
 #include "dce_v8_0.h"
 #endif
 #include "dce_v10_0.h"
-#include "dce_v11_0.h"
 #include "ivsrcid/ivsrcid_vislands30.h"
 #include "amdgpu_vkms.h"
 #include "amdgpu_display.h"
+#include "atom.h"
+#include "amdgpu_irq.h"
 
 /**
  * DOC: amdgpu_vkms
@@ -41,46 +44,44 @@ static const u32 amdgpu_vkms_formats[] = {
 
 static enum hrtimer_restart amdgpu_vkms_vblank_simulate(struct hrtimer *timer)
 {
-	struct amdgpu_vkms_output *output = container_of(timer,
-							 struct amdgpu_vkms_output,
-							 vblank_hrtimer);
-	struct drm_crtc *crtc = &output->crtc;
+	struct amdgpu_crtc *amdgpu_crtc = container_of(timer, struct amdgpu_crtc, vblank_timer);
+	struct drm_crtc *crtc = &amdgpu_crtc->base;
+	struct amdgpu_vkms_output *output = drm_crtc_to_amdgpu_vkms_output(crtc);
 	u64 ret_overrun;
 	bool ret;
 
-	ret_overrun = hrtimer_forward_now(&output->vblank_hrtimer,
+	ret_overrun = hrtimer_forward_now(&amdgpu_crtc->vblank_timer,
 					  output->period_ns);
-	WARN_ON(ret_overrun != 1);
+	if (ret_overrun != 1)
+		DRM_WARN("%s: vblank timer overrun\n", __func__);
 
 	ret = drm_crtc_handle_vblank(crtc);
+	/* Don't queue timer again when vblank is disabled. */
 	if (!ret)
-		DRM_ERROR("amdgpu_vkms failure on handling vblank");
+		return HRTIMER_NORESTART;
 
 	return HRTIMER_RESTART;
 }
 
 static int amdgpu_vkms_enable_vblank(struct drm_crtc *crtc)
 {
-	struct drm_device *dev = crtc->dev;
-	unsigned int pipe = drm_crtc_index(crtc);
-	struct drm_vblank_crtc *vblank = &dev->vblank[pipe];
+	struct drm_vblank_crtc *vblank = drm_crtc_vblank_crtc(crtc);
 	struct amdgpu_vkms_output *out = drm_crtc_to_amdgpu_vkms_output(crtc);
+	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
 
 	drm_calc_timestamping_constants(crtc, &crtc->mode);
 
-	hrtimer_init(&out->vblank_hrtimer, CLOCK_MONOTONIC, HRTIMER_MODE_REL);
-	out->vblank_hrtimer.function = &amdgpu_vkms_vblank_simulate;
 	out->period_ns = ktime_set(0, vblank->framedur_ns);
-	hrtimer_start(&out->vblank_hrtimer, out->period_ns, HRTIMER_MODE_REL);
+	hrtimer_start(&amdgpu_crtc->vblank_timer, out->period_ns, HRTIMER_MODE_REL);
 
 	return 0;
 }
 
 static void amdgpu_vkms_disable_vblank(struct drm_crtc *crtc)
 {
-	struct amdgpu_vkms_output *out = drm_crtc_to_amdgpu_vkms_output(crtc);
+	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
 
-	hrtimer_cancel(&out->vblank_hrtimer);
+	hrtimer_try_to_cancel(&amdgpu_crtc->vblank_timer);
 }
 
 static bool amdgpu_vkms_get_vblank_timestamp(struct drm_crtc *crtc,
@@ -88,17 +89,16 @@ static bool amdgpu_vkms_get_vblank_timestamp(struct drm_crtc *crtc,
 					     ktime_t *vblank_time,
 					     bool in_vblank_irq)
 {
-	struct drm_device *dev = crtc->dev;
-	unsigned int pipe = crtc->index;
 	struct amdgpu_vkms_output *output = drm_crtc_to_amdgpu_vkms_output(crtc);
-	struct drm_vblank_crtc *vblank = &dev->vblank[pipe];
+	struct drm_vblank_crtc *vblank = drm_crtc_vblank_crtc(crtc);
+	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
 
 	if (!READ_ONCE(vblank->enabled)) {
 		*vblank_time = ktime_get();
 		return true;
 	}
 
-	*vblank_time = READ_ONCE(output->vblank_hrtimer.node.expires);
+	*vblank_time = READ_ONCE(amdgpu_crtc->vblank_timer.node.expires);
 
 	if (WARN_ON(*vblank_time == vblank->time))
 		return true;
@@ -142,15 +142,16 @@ static void amdgpu_vkms_crtc_atomic_disable(struct drm_crtc *crtc,
 static void amdgpu_vkms_crtc_atomic_flush(struct drm_crtc *crtc,
 					  struct drm_atomic_state *state)
 {
+	unsigned long flags;
 	if (crtc->state->event) {
-		spin_lock(&crtc->dev->event_lock);
+		spin_lock_irqsave(&crtc->dev->event_lock, flags);
 
 		if (drm_crtc_vblank_get(crtc) != 0)
 			drm_crtc_send_vblank_event(crtc, crtc->state->event);
 		else
 			drm_crtc_arm_vblank_event(crtc, crtc->state->event);
 
-		spin_unlock(&crtc->dev->event_lock);
+		spin_unlock_irqrestore(&crtc->dev->event_lock, flags);
 
 		crtc->state->event = NULL;
 	}
@@ -165,6 +166,8 @@ static const struct drm_crtc_helper_funcs amdgpu_vkms_crtc_helper_funcs = {
 static int amdgpu_vkms_crtc_init(struct drm_device *dev, struct drm_crtc *crtc,
 			  struct drm_plane *primary, struct drm_plane *cursor)
 {
+	struct amdgpu_device *adev = drm_to_adev(dev);
+	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
 	int ret;
 
 	ret = drm_crtc_init_with_planes(dev, crtc, primary, cursor,
@@ -176,6 +179,17 @@ static int amdgpu_vkms_crtc_init(struct drm_device *dev, struct drm_crtc *crtc,
 
 	drm_crtc_helper_add(crtc, &amdgpu_vkms_crtc_helper_funcs);
 
+	amdgpu_crtc->crtc_id = drm_crtc_index(crtc);
+	adev->mode_info.crtcs[drm_crtc_index(crtc)] = amdgpu_crtc;
+
+	amdgpu_crtc->pll_id = ATOM_PPLL_INVALID;
+	amdgpu_crtc->encoder = NULL;
+	amdgpu_crtc->connector = NULL;
+	amdgpu_crtc->vsync_timer_enabled = AMDGPU_IRQ_STATE_DISABLE;
+
+	hrtimer_setup(&amdgpu_crtc->vblank_timer, &amdgpu_vkms_vblank_simulate, CLOCK_MONOTONIC,
+		      HRTIMER_MODE_REL);
+
 	return ret;
 }
 
@@ -222,6 +236,8 @@ static int amdgpu_vkms_conn_get_modes(struct drm_connector *connector)
 
 	for (i = 0; i < ARRAY_SIZE(common_modes); i++) {
 		mode = drm_cvt_mode(dev, common_modes[i].w, common_modes[i].h, 60, false, false, false);
+		if (!mode)
+			continue;
 		drm_mode_probed_add(connector, mode);
 	}
 
@@ -266,8 +282,8 @@ static int amdgpu_vkms_plane_atomic_check(struct drm_plane *plane,
 		return PTR_ERR(crtc_state);
 
 	ret = drm_atomic_helper_check_plane_state(new_plane_state, crtc_state,
-						  DRM_PLANE_HELPER_NO_SCALING,
-						  DRM_PLANE_HELPER_NO_SCALING,
+						  DRM_PLANE_NO_SCALING,
+						  DRM_PLANE_NO_SCALING,
 						  false, true);
 	if (ret != 0)
 		return ret;
@@ -286,9 +302,6 @@ static int amdgpu_vkms_prepare_fb(struct drm_plane *plane,
 	struct drm_gem_object *obj;
 	struct amdgpu_device *adev;
 	struct amdgpu_bo *rbo;
-	struct list_head list;
-	struct ttm_validate_buffer tv;
-	struct ww_acquire_ctx ticket;
 	uint32_t domain;
 	int r;
 
@@ -297,61 +310,80 @@ static int amdgpu_vkms_prepare_fb(struct drm_plane *plane,
 		return 0;
 	}
 	afb = to_amdgpu_framebuffer(new_state->fb);
-	obj = new_state->fb->obj[0];
+
+	obj = drm_gem_fb_get_obj(new_state->fb, 0);
+	if (!obj) {
+		DRM_ERROR("Failed to get obj from framebuffer\n");
+		return -EINVAL;
+	}
+
 	rbo = gem_to_amdgpu_bo(obj);
 	adev = amdgpu_ttm_adev(rbo->tbo.bdev);
-	INIT_LIST_HEAD(&list);
-
-	tv.bo = &rbo->tbo;
-	tv.num_shared = 1;
-	list_add(&tv.head, &list);
 
-	r = ttm_eu_reserve_buffers(&ticket, &list, false, NULL);
+	r = amdgpu_bo_reserve(rbo, true);
 	if (r) {
 		dev_err(adev->dev, "fail to reserve bo (%d)\n", r);
 		return r;
 	}
 
+	r = dma_resv_reserve_fences(rbo->tbo.base.resv, 1);
+	if (r) {
+		dev_err(adev->dev, "allocating fence slot failed (%d)\n", r);
+		goto error_unlock;
+	}
+
 	if (plane->type != DRM_PLANE_TYPE_CURSOR)
 		domain = amdgpu_display_supported_domains(adev, rbo->flags);
 	else
 		domain = AMDGPU_GEM_DOMAIN_VRAM;
 
+	rbo->flags |= AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS;
 	r = amdgpu_bo_pin(rbo, domain);
 	if (unlikely(r != 0)) {
 		if (r != -ERESTARTSYS)
 			DRM_ERROR("Failed to pin framebuffer with error %d\n", r);
-		ttm_eu_backoff_reservation(&ticket, &list);
-		return r;
+		goto error_unlock;
 	}
 
 	r = amdgpu_ttm_alloc_gart(&rbo->tbo);
 	if (unlikely(r != 0)) {
-		amdgpu_bo_unpin(rbo);
-		ttm_eu_backoff_reservation(&ticket, &list);
 		DRM_ERROR("%p bind failed\n", rbo);
-		return r;
+		goto error_unpin;
 	}
 
-	ttm_eu_backoff_reservation(&ticket, &list);
+	amdgpu_bo_unreserve(rbo);
 
 	afb->address = amdgpu_bo_gpu_offset(rbo);
 
 	amdgpu_bo_ref(rbo);
 
 	return 0;
+
+error_unpin:
+	amdgpu_bo_unpin(rbo);
+
+error_unlock:
+	amdgpu_bo_unreserve(rbo);
+	return r;
 }
 
 static void amdgpu_vkms_cleanup_fb(struct drm_plane *plane,
 				   struct drm_plane_state *old_state)
 {
 	struct amdgpu_bo *rbo;
+	struct drm_gem_object *obj;
 	int r;
 
 	if (!old_state->fb)
 		return;
 
-	rbo = gem_to_amdgpu_bo(old_state->fb->obj[0]);
+	obj = drm_gem_fb_get_obj(old_state->fb, 0);
+	if (!obj) {
+		DRM_ERROR("Failed to get obj from framebuffer\n");
+		return;
+	}
+
+	rbo = gem_to_amdgpu_bo(obj);
 	r = amdgpu_bo_reserve(rbo, false);
 	if (unlikely(r)) {
 		DRM_ERROR("failed to reserve rbo before unpin\n");
@@ -396,12 +428,12 @@ static struct drm_plane *amdgpu_vkms_plane_init(struct drm_device *dev,
 	return plane;
 }
 
-int amdgpu_vkms_output_init(struct drm_device *dev,
-			    struct amdgpu_vkms_output *output, int index)
+static int amdgpu_vkms_output_init(struct drm_device *dev, struct
+				   amdgpu_vkms_output *output, int index)
 {
 	struct drm_connector *connector = &output->connector;
 	struct drm_encoder *encoder = &output->encoder;
-	struct drm_crtc *crtc = &output->crtc;
+	struct drm_crtc *crtc = &output->crtc.base;
 	struct drm_plane *primary, *cursor = NULL;
 	int ret;
 
@@ -460,10 +492,15 @@ const struct drm_mode_config_funcs amdgpu_vkms_mode_funcs = {
 	.atomic_commit = drm_atomic_helper_commit,
 };
 
-static int amdgpu_vkms_sw_init(void *handle)
+static int amdgpu_vkms_sw_init(struct amdgpu_ip_block *ip_block)
 {
 	int r, i;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
+
+	adev->amdgpu_vkms_output = kcalloc(adev->mode_info.num_crtc,
+		sizeof(struct amdgpu_vkms_output), GFP_KERNEL);
+	if (!adev->amdgpu_vkms_output)
+		return -ENOMEM;
 
 	adev_to_drm(adev)->max_vblank_count = 0;
 
@@ -475,16 +512,12 @@ static int amdgpu_vkms_sw_init(void *handle)
 	adev_to_drm(adev)->mode_config.preferred_depth = 24;
 	adev_to_drm(adev)->mode_config.prefer_shadow = 1;
 
-	adev_to_drm(adev)->mode_config.fb_base = adev->gmc.aper_base;
+	adev_to_drm(adev)->mode_config.fb_modifiers_not_supported = true;
 
 	r = amdgpu_display_modeset_create_props(adev);
 	if (r)
 		return r;
 
-	adev->amdgpu_vkms_output = kcalloc(adev->mode_info.num_crtc, sizeof(struct amdgpu_vkms_output), GFP_KERNEL);
-	if (!adev->amdgpu_vkms_output)
-		return -ENOMEM;
-
 	/* allocate crtcs, encoders, connectors */
 	for (i = 0; i < adev->mode_info.num_crtc; i++) {
 		r = amdgpu_vkms_output_init(adev_to_drm(adev), &adev->amdgpu_vkms_output[i], i);
@@ -492,33 +525,38 @@ static int amdgpu_vkms_sw_init(void *handle)
 			return r;
 	}
 
+	r = drm_vblank_init(adev_to_drm(adev), adev->mode_info.num_crtc);
+	if (r)
+		return r;
+
 	drm_kms_helper_poll_init(adev_to_drm(adev));
 
 	adev->mode_info.mode_config_initialized = true;
 	return 0;
 }
 
-static int amdgpu_vkms_sw_fini(void *handle)
+static int amdgpu_vkms_sw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int i = 0;
 
 	for (i = 0; i < adev->mode_info.num_crtc; i++)
-		if (adev->amdgpu_vkms_output[i].vblank_hrtimer.function)
-			hrtimer_cancel(&adev->amdgpu_vkms_output[i].vblank_hrtimer);
-
-	kfree(adev->mode_info.bios_hardcoded_edid);
-	kfree(adev->amdgpu_vkms_output);
+		if (adev->mode_info.crtcs[i])
+			hrtimer_cancel(&adev->mode_info.crtcs[i]->vblank_timer);
 
 	drm_kms_helper_poll_fini(adev_to_drm(adev));
+	drm_mode_config_cleanup(adev_to_drm(adev));
 
 	adev->mode_info.mode_config_initialized = false;
+
+	drm_edid_free(adev->mode_info.bios_hardcoded_edid);
+	kfree(adev->amdgpu_vkms_output);
 	return 0;
 }
 
-static int amdgpu_vkms_hw_init(void *handle)
+static int amdgpu_vkms_hw_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	switch (adev->asic_type) {
 #ifdef CONFIG_DRM_AMDGPU_SI
@@ -542,13 +580,6 @@ static int amdgpu_vkms_hw_init(void *handle)
 	case CHIP_TONGA:
 		dce_v10_0_disable_dce(adev);
 		break;
-	case CHIP_CARRIZO:
-	case CHIP_STONEY:
-	case CHIP_POLARIS10:
-	case CHIP_POLARIS11:
-	case CHIP_VEGAM:
-		dce_v11_0_disable_dce(adev);
-		break;
 	case CHIP_TOPAZ:
 #ifdef CONFIG_DRM_AMDGPU_SI
 	case CHIP_HAINAN:
@@ -561,55 +592,45 @@ static int amdgpu_vkms_hw_init(void *handle)
 	return 0;
 }
 
-static int amdgpu_vkms_hw_fini(void *handle)
+static int amdgpu_vkms_hw_fini(struct amdgpu_ip_block *ip_block)
 {
 	return 0;
 }
 
-static int amdgpu_vkms_suspend(void *handle)
+static int amdgpu_vkms_suspend(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int r;
 
 	r = drm_mode_config_helper_suspend(adev_to_drm(adev));
 	if (r)
 		return r;
-	return amdgpu_vkms_hw_fini(handle);
+
+	return 0;
 }
 
-static int amdgpu_vkms_resume(void *handle)
+static int amdgpu_vkms_resume(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 	int r;
 
-	r = amdgpu_vkms_hw_init(handle);
+	r = amdgpu_vkms_hw_init(ip_block);
 	if (r)
 		return r;
-	return drm_mode_config_helper_resume(adev_to_drm(adev));
+	return drm_mode_config_helper_resume(adev_to_drm(ip_block->adev));
 }
 
-static bool amdgpu_vkms_is_idle(void *handle)
+static bool amdgpu_vkms_is_idle(struct amdgpu_ip_block *ip_block)
 {
 	return true;
 }
 
-static int amdgpu_vkms_wait_for_idle(void *handle)
-{
-	return 0;
-}
-
-static int amdgpu_vkms_soft_reset(void *handle)
-{
-	return 0;
-}
-
-static int amdgpu_vkms_set_clockgating_state(void *handle,
+static int amdgpu_vkms_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_clockgating_state state)
 {
 	return 0;
 }
 
-static int amdgpu_vkms_set_powergating_state(void *handle,
+static int amdgpu_vkms_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_powergating_state state)
 {
 	return 0;
@@ -617,8 +638,6 @@ static int amdgpu_vkms_set_powergating_state(void *handle,
 
 static const struct amd_ip_funcs amdgpu_vkms_ip_funcs = {
 	.name = "amdgpu_vkms",
-	.early_init = NULL,
-	.late_init = NULL,
 	.sw_init = amdgpu_vkms_sw_init,
 	.sw_fini = amdgpu_vkms_sw_fini,
 	.hw_init = amdgpu_vkms_hw_init,
@@ -626,14 +645,11 @@ static const struct amd_ip_funcs amdgpu_vkms_ip_funcs = {
 	.suspend = amdgpu_vkms_suspend,
 	.resume = amdgpu_vkms_resume,
 	.is_idle = amdgpu_vkms_is_idle,
-	.wait_for_idle = amdgpu_vkms_wait_for_idle,
-	.soft_reset = amdgpu_vkms_soft_reset,
 	.set_clockgating_state = amdgpu_vkms_set_clockgating_state,
 	.set_powergating_state = amdgpu_vkms_set_powergating_state,
 };
 
-const struct amdgpu_ip_block_version amdgpu_vkms_ip_block =
-{
+const struct amdgpu_ip_block_version amdgpu_vkms_ip_block = {
 	.type = AMD_IP_BLOCK_TYPE_DCE,
 	.major = 1,
 	.minor = 0,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vkms.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vkms.h
index 97f1b79c0724..4f8722ff37c2 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vkms.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vkms.h
@@ -10,15 +10,14 @@
 #define YRES_MAX  16384
 
 #define drm_crtc_to_amdgpu_vkms_output(target) \
-	container_of(target, struct amdgpu_vkms_output, crtc)
+	container_of(target, struct amdgpu_vkms_output, crtc.base)
 
 extern const struct amdgpu_ip_block_version amdgpu_vkms_ip_block;
 
 struct amdgpu_vkms_output {
-	struct drm_crtc crtc;
+	struct amdgpu_crtc crtc;
 	struct drm_encoder encoder;
 	struct drm_connector connector;
-	struct hrtimer vblank_hrtimer;
 	ktime_t period_ns;
 	struct drm_pending_vblank_event *event;
 };
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
index 0e7dc23f78e7..a67285118c37 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
@@ -33,7 +33,10 @@
 
 #include <drm/amdgpu_drm.h>
 #include <drm/drm_drv.h>
+#include <drm/ttm/ttm_tt.h>
+#include <drm/drm_exec.h>
 #include "amdgpu.h"
+#include "amdgpu_vm.h"
 #include "amdgpu_trace.h"
 #include "amdgpu_amdkfd.h"
 #include "amdgpu_gmc.h"
@@ -45,22 +48,43 @@
 /**
  * DOC: GPUVM
  *
- * GPUVM is similar to the legacy gart on older asics, however
- * rather than there being a single global gart table
- * for the entire GPU, there are multiple VM page tables active
- * at any given time.  The VM page tables can contain a mix
- * vram pages and system memory pages and system memory pages
+ * GPUVM is the MMU functionality provided on the GPU.
+ * GPUVM is similar to the legacy GART on older asics, however
+ * rather than there being a single global GART table
+ * for the entire GPU, there can be multiple GPUVM page tables active
+ * at any given time.  The GPUVM page tables can contain a mix
+ * VRAM pages and system pages (both memory and MMIO) and system pages
  * can be mapped as snooped (cached system pages) or unsnooped
  * (uncached system pages).
- * Each VM has an ID associated with it and there is a page table
- * associated with each VMID.  When execting a command buffer,
- * the kernel tells the the ring what VMID to use for that command
+ *
+ * Each active GPUVM has an ID associated with it and there is a page table
+ * linked with each VMID.  When executing a command buffer,
+ * the kernel tells the engine what VMID to use for that command
  * buffer.  VMIDs are allocated dynamically as commands are submitted.
  * The userspace drivers maintain their own address space and the kernel
  * sets up their pages tables accordingly when they submit their
  * command buffers and a VMID is assigned.
- * Cayman/Trinity support up to 8 active VMs at any given time;
- * SI supports 16.
+ * The hardware supports up to 16 active GPUVMs at any given time.
+ *
+ * Each GPUVM is represented by a 1-2 or 1-5 level page table, depending
+ * on the ASIC family.  GPUVM supports RWX attributes on each page as well
+ * as other features such as encryption and caching attributes.
+ *
+ * VMID 0 is special.  It is the GPUVM used for the kernel driver.  In
+ * addition to an aperture managed by a page table, VMID 0 also has
+ * several other apertures.  There is an aperture for direct access to VRAM
+ * and there is a legacy AGP aperture which just forwards accesses directly
+ * to the matching system physical addresses (or IOVAs when an IOMMU is
+ * present).  These apertures provide direct access to these memories without
+ * incurring the overhead of a page table.  VMID 0 is used by the kernel
+ * driver for tasks like memory management.
+ *
+ * GPU clients (i.e., engines on the GPU) use GPUVM VMIDs to access memory.
+ * For user applications, each application can have their own unique GPUVM
+ * address space.  The application manages the address space and the kernel
+ * driver manages the GPUVM page tables for each process.  If an GPU client
+ * accesses an invalid page, it will generate a GPU page fault, similar to
+ * accessing an invalid page on a CPU.
  */
 
 #define START(node) ((node)->start)
@@ -89,171 +113,29 @@ struct amdgpu_prt_cb {
 };
 
 /**
- * amdgpu_vm_set_pasid - manage pasid and vm ptr mapping
- *
- * @adev: amdgpu_device pointer
- * @vm: amdgpu_vm pointer
- * @pasid: the pasid the VM is using on this GPU
- *
- * Set the pasid this VM is using on this GPU, can also be used to remove the
- * pasid by passing in zero.
- *
- */
-int amdgpu_vm_set_pasid(struct amdgpu_device *adev, struct amdgpu_vm *vm,
-			u32 pasid)
-{
-	int r;
-
-	if (vm->pasid == pasid)
-		return 0;
-
-	if (vm->pasid) {
-		r = xa_err(xa_erase_irq(&adev->vm_manager.pasids, vm->pasid));
-		if (r < 0)
-			return r;
-
-		vm->pasid = 0;
-	}
-
-	if (pasid) {
-		r = xa_err(xa_store_irq(&adev->vm_manager.pasids, pasid, vm,
-					GFP_KERNEL));
-		if (r < 0)
-			return r;
-
-		vm->pasid = pasid;
-	}
-
-
-	return 0;
-}
-
-/*
- * vm eviction_lock can be taken in MMU notifiers. Make sure no reclaim-FS
- * happens while holding this lock anywhere to prevent deadlocks when
- * an MMU notifier runs in reclaim-FS context.
- */
-static inline void amdgpu_vm_eviction_lock(struct amdgpu_vm *vm)
-{
-	mutex_lock(&vm->eviction_lock);
-	vm->saved_flags = memalloc_noreclaim_save();
-}
-
-static inline int amdgpu_vm_eviction_trylock(struct amdgpu_vm *vm)
-{
-	if (mutex_trylock(&vm->eviction_lock)) {
-		vm->saved_flags = memalloc_noreclaim_save();
-		return 1;
-	}
-	return 0;
-}
-
-static inline void amdgpu_vm_eviction_unlock(struct amdgpu_vm *vm)
-{
-	memalloc_noreclaim_restore(vm->saved_flags);
-	mutex_unlock(&vm->eviction_lock);
-}
-
-/**
- * amdgpu_vm_level_shift - return the addr shift for each level
- *
- * @adev: amdgpu_device pointer
- * @level: VMPT level
- *
- * Returns:
- * The number of bits the pfn needs to be right shifted for a level.
- */
-static unsigned amdgpu_vm_level_shift(struct amdgpu_device *adev,
-				      unsigned level)
-{
-	switch (level) {
-	case AMDGPU_VM_PDB2:
-	case AMDGPU_VM_PDB1:
-	case AMDGPU_VM_PDB0:
-		return 9 * (AMDGPU_VM_PDB0 - level) +
-			adev->vm_manager.block_size;
-	case AMDGPU_VM_PTB:
-		return 0;
-	default:
-		return ~0;
-	}
-}
-
-/**
- * amdgpu_vm_num_entries - return the number of entries in a PD/PT
- *
- * @adev: amdgpu_device pointer
- * @level: VMPT level
- *
- * Returns:
- * The number of entries in a page directory or page table.
- */
-static unsigned amdgpu_vm_num_entries(struct amdgpu_device *adev,
-				      unsigned level)
-{
-	unsigned shift = amdgpu_vm_level_shift(adev,
-					       adev->vm_manager.root_level);
-
-	if (level == adev->vm_manager.root_level)
-		/* For the root directory */
-		return round_up(adev->vm_manager.max_pfn, 1ULL << shift)
-			>> shift;
-	else if (level != AMDGPU_VM_PTB)
-		/* Everything in between */
-		return 512;
-	else
-		/* For the page tables on the leaves */
-		return AMDGPU_VM_PTE_COUNT(adev);
-}
-
-/**
- * amdgpu_vm_num_ats_entries - return the number of ATS entries in the root PD
- *
- * @adev: amdgpu_device pointer
- *
- * Returns:
- * The number of entries in the root page directory which needs the ATS setting.
+ * struct amdgpu_vm_tlb_seq_struct - Helper to increment the TLB flush sequence
  */
-static unsigned amdgpu_vm_num_ats_entries(struct amdgpu_device *adev)
-{
-	unsigned shift;
-
-	shift = amdgpu_vm_level_shift(adev, adev->vm_manager.root_level);
-	return AMDGPU_GMC_HOLE_START >> (shift + AMDGPU_GPU_PAGE_SHIFT);
-}
+struct amdgpu_vm_tlb_seq_struct {
+	/**
+	 * @vm: pointer to the amdgpu_vm structure to set the fence sequence on
+	 */
+	struct amdgpu_vm *vm;
 
-/**
- * amdgpu_vm_entries_mask - the mask to get the entry number of a PD/PT
- *
- * @adev: amdgpu_device pointer
- * @level: VMPT level
- *
- * Returns:
- * The mask to extract the entry number of a PD/PT from an address.
- */
-static uint32_t amdgpu_vm_entries_mask(struct amdgpu_device *adev,
-				       unsigned int level)
-{
-	if (level <= adev->vm_manager.root_level)
-		return 0xffffffff;
-	else if (level != AMDGPU_VM_PTB)
-		return 0x1ff;
-	else
-		return AMDGPU_VM_PTE_COUNT(adev) - 1;
-}
+	/**
+	 * @cb: callback
+	 */
+	struct dma_fence_cb cb;
+};
 
 /**
- * amdgpu_vm_bo_size - returns the size of the BOs in bytes
- *
- * @adev: amdgpu_device pointer
- * @level: VMPT level
+ * amdgpu_vm_assert_locked - check if VM is correctly locked
+ * @vm: the VM which schould be tested
  *
- * Returns:
- * The size of the BO for a page directory or page table in bytes.
+ * Asserts that the VM root PD is locked.
  */
-static unsigned amdgpu_vm_bo_size(struct amdgpu_device *adev, unsigned level)
+static void amdgpu_vm_assert_locked(struct amdgpu_vm *vm)
 {
-	return AMDGPU_GPU_PAGE_ALIGN(amdgpu_vm_num_entries(adev, level) * 8);
+	dma_resv_assert_held(vm->root.bo->tbo.base.resv);
 }
 
 /**
@@ -270,10 +152,13 @@ static void amdgpu_vm_bo_evicted(struct amdgpu_vm_bo_base *vm_bo)
 	struct amdgpu_bo *bo = vm_bo->bo;
 
 	vm_bo->moved = true;
+	amdgpu_vm_assert_locked(vm);
+	spin_lock(&vm_bo->vm->status_lock);
 	if (bo->tbo.type == ttm_bo_type_kernel)
 		list_move(&vm_bo->vm_status, &vm->evicted);
 	else
 		list_move_tail(&vm_bo->vm_status, &vm->evicted);
+	spin_unlock(&vm_bo->vm->status_lock);
 }
 /**
  * amdgpu_vm_bo_moved - vm_bo is moved
@@ -285,7 +170,10 @@ static void amdgpu_vm_bo_evicted(struct amdgpu_vm_bo_base *vm_bo)
  */
 static void amdgpu_vm_bo_moved(struct amdgpu_vm_bo_base *vm_bo)
 {
+	amdgpu_vm_assert_locked(vm_bo->vm);
+	spin_lock(&vm_bo->vm->status_lock);
 	list_move(&vm_bo->vm_status, &vm_bo->vm->moved);
+	spin_unlock(&vm_bo->vm->status_lock);
 }
 
 /**
@@ -298,7 +186,10 @@ static void amdgpu_vm_bo_moved(struct amdgpu_vm_bo_base *vm_bo)
  */
 static void amdgpu_vm_bo_idle(struct amdgpu_vm_bo_base *vm_bo)
 {
+	amdgpu_vm_assert_locked(vm_bo->vm);
+	spin_lock(&vm_bo->vm->status_lock);
 	list_move(&vm_bo->vm_status, &vm_bo->vm->idle);
+	spin_unlock(&vm_bo->vm->status_lock);
 	vm_bo->moved = false;
 }
 
@@ -312,9 +203,25 @@ static void amdgpu_vm_bo_idle(struct amdgpu_vm_bo_base *vm_bo)
  */
 static void amdgpu_vm_bo_invalidated(struct amdgpu_vm_bo_base *vm_bo)
 {
-	spin_lock(&vm_bo->vm->invalidated_lock);
+	spin_lock(&vm_bo->vm->status_lock);
 	list_move(&vm_bo->vm_status, &vm_bo->vm->invalidated);
-	spin_unlock(&vm_bo->vm->invalidated_lock);
+	spin_unlock(&vm_bo->vm->status_lock);
+}
+
+/**
+ * amdgpu_vm_bo_evicted_user - vm_bo is evicted
+ *
+ * @vm_bo: vm_bo which is evicted
+ *
+ * State for BOs used by user mode queues which are not at the location they
+ * should be.
+ */
+static void amdgpu_vm_bo_evicted_user(struct amdgpu_vm_bo_base *vm_bo)
+{
+	vm_bo->moved = true;
+	spin_lock(&vm_bo->vm->status_lock);
+	list_move(&vm_bo->vm_status, &vm_bo->vm->evicted_user);
+	spin_unlock(&vm_bo->vm->status_lock);
 }
 
 /**
@@ -327,10 +234,14 @@ static void amdgpu_vm_bo_invalidated(struct amdgpu_vm_bo_base *vm_bo)
  */
 static void amdgpu_vm_bo_relocated(struct amdgpu_vm_bo_base *vm_bo)
 {
-	if (vm_bo->bo->parent)
+	amdgpu_vm_assert_locked(vm_bo->vm);
+	if (vm_bo->bo->parent) {
+		spin_lock(&vm_bo->vm->status_lock);
 		list_move(&vm_bo->vm_status, &vm_bo->vm->relocated);
-	else
+		spin_unlock(&vm_bo->vm->status_lock);
+	} else {
 		amdgpu_vm_bo_idle(vm_bo);
+	}
 }
 
 /**
@@ -343,330 +254,256 @@ static void amdgpu_vm_bo_relocated(struct amdgpu_vm_bo_base *vm_bo)
  */
 static void amdgpu_vm_bo_done(struct amdgpu_vm_bo_base *vm_bo)
 {
-	spin_lock(&vm_bo->vm->invalidated_lock);
+	amdgpu_vm_assert_locked(vm_bo->vm);
+	spin_lock(&vm_bo->vm->status_lock);
 	list_move(&vm_bo->vm_status, &vm_bo->vm->done);
-	spin_unlock(&vm_bo->vm->invalidated_lock);
+	spin_unlock(&vm_bo->vm->status_lock);
 }
 
 /**
- * amdgpu_vm_bo_base_init - Adds bo to the list of bos associated with the vm
- *
- * @base: base structure for tracking BO usage in a VM
- * @vm: vm to which bo is to be added
- * @bo: amdgpu buffer object
- *
- * Initialize a bo_va_base structure and add it to the appropriate lists
+ * amdgpu_vm_bo_reset_state_machine - reset the vm_bo state machine
+ * @vm: the VM which state machine to reset
  *
+ * Move all vm_bo object in the VM into a state where they will be updated
+ * again during validation.
  */
-static void amdgpu_vm_bo_base_init(struct amdgpu_vm_bo_base *base,
-				   struct amdgpu_vm *vm,
-				   struct amdgpu_bo *bo)
+static void amdgpu_vm_bo_reset_state_machine(struct amdgpu_vm *vm)
 {
-	base->vm = vm;
-	base->bo = bo;
-	base->next = NULL;
-	INIT_LIST_HEAD(&base->vm_status);
-
-	if (!bo)
-		return;
-	base->next = bo->vm_bo;
-	bo->vm_bo = base;
+	struct amdgpu_vm_bo_base *vm_bo, *tmp;
 
-	if (bo->tbo.base.resv != vm->root.bo->tbo.base.resv)
-		return;
+	amdgpu_vm_assert_locked(vm);
 
-	vm->bulk_moveable = false;
-	if (bo->tbo.type == ttm_bo_type_kernel && bo->parent)
-		amdgpu_vm_bo_relocated(base);
-	else
-		amdgpu_vm_bo_idle(base);
+	spin_lock(&vm->status_lock);
+	list_splice_init(&vm->done, &vm->invalidated);
+	list_for_each_entry(vm_bo, &vm->invalidated, vm_status)
+		vm_bo->moved = true;
 
-	if (bo->preferred_domains &
-	    amdgpu_mem_type_to_domain(bo->tbo.resource->mem_type))
-		return;
+	list_for_each_entry_safe(vm_bo, tmp, &vm->idle, vm_status) {
+		struct amdgpu_bo *bo = vm_bo->bo;
 
-	/*
-	 * we checked all the prerequisites, but it looks like this per vm bo
-	 * is currently evicted. add the bo to the evicted list to make sure it
-	 * is validated on next vm use to avoid fault.
-	 * */
-	amdgpu_vm_bo_evicted(base);
+		vm_bo->moved = true;
+		if (!bo || bo->tbo.type != ttm_bo_type_kernel)
+			list_move(&vm_bo->vm_status, &vm_bo->vm->moved);
+		else if (bo->parent)
+			list_move(&vm_bo->vm_status, &vm_bo->vm->relocated);
+	}
+	spin_unlock(&vm->status_lock);
 }
 
 /**
- * amdgpu_vm_pt_parent - get the parent page directory
- *
- * @pt: child page table
+ * amdgpu_vm_update_shared - helper to update shared memory stat
+ * @base: base structure for tracking BO usage in a VM
  *
- * Helper to get the parent entry for the child page table. NULL if we are at
- * the root page directory.
- */
-static struct amdgpu_vm_bo_base *amdgpu_vm_pt_parent(struct amdgpu_vm_bo_base *pt)
-{
-	struct amdgpu_bo *parent = pt->bo->parent;
-
-	if (!parent)
-		return NULL;
-
-	return parent->vm_bo;
+ * Takes the vm status_lock and updates the shared memory stat. If the basic
+ * stat changed (e.g. buffer was moved) amdgpu_vm_update_stats need to be called
+ * as well.
+ */
+static void amdgpu_vm_update_shared(struct amdgpu_vm_bo_base *base)
+{
+	struct amdgpu_vm *vm = base->vm;
+	struct amdgpu_bo *bo = base->bo;
+	uint64_t size = amdgpu_bo_size(bo);
+	uint32_t bo_memtype = amdgpu_bo_mem_stats_placement(bo);
+	bool shared;
+
+	dma_resv_assert_held(bo->tbo.base.resv);
+	spin_lock(&vm->status_lock);
+	shared = drm_gem_object_is_shared_for_memory_stats(&bo->tbo.base);
+	if (base->shared != shared) {
+		base->shared = shared;
+		if (shared) {
+			vm->stats[bo_memtype].drm.shared += size;
+			vm->stats[bo_memtype].drm.private -= size;
+		} else {
+			vm->stats[bo_memtype].drm.shared -= size;
+			vm->stats[bo_memtype].drm.private += size;
+		}
+	}
+	spin_unlock(&vm->status_lock);
 }
 
-/*
- * amdgpu_vm_pt_cursor - state for for_each_amdgpu_vm_pt
- */
-struct amdgpu_vm_pt_cursor {
-	uint64_t pfn;
-	struct amdgpu_vm_bo_base *parent;
-	struct amdgpu_vm_bo_base *entry;
-	unsigned level;
-};
-
 /**
- * amdgpu_vm_pt_start - start PD/PT walk
- *
- * @adev: amdgpu_device pointer
- * @vm: amdgpu_vm structure
- * @start: start address of the walk
- * @cursor: state to initialize
+ * amdgpu_vm_bo_update_shared - callback when bo gets shared/unshared
+ * @bo: amdgpu buffer object
  *
- * Initialize a amdgpu_vm_pt_cursor to start a walk.
+ * Update the per VM stats for all the vm if needed from private to shared or
+ * vice versa.
  */
-static void amdgpu_vm_pt_start(struct amdgpu_device *adev,
-			       struct amdgpu_vm *vm, uint64_t start,
-			       struct amdgpu_vm_pt_cursor *cursor)
+void amdgpu_vm_bo_update_shared(struct amdgpu_bo *bo)
 {
-	cursor->pfn = start;
-	cursor->parent = NULL;
-	cursor->entry = &vm->root;
-	cursor->level = adev->vm_manager.root_level;
+	struct amdgpu_vm_bo_base *base;
+
+	for (base = bo->vm_bo; base; base = base->next)
+		amdgpu_vm_update_shared(base);
 }
 
 /**
- * amdgpu_vm_pt_descendant - go to child node
- *
- * @adev: amdgpu_device pointer
- * @cursor: current state
+ * amdgpu_vm_update_stats_locked - helper to update normal memory stat
+ * @base: base structure for tracking BO usage in a VM
+ * @res:  the ttm_resource to use for the purpose of accounting, may or may not
+ *        be bo->tbo.resource
+ * @sign: if we should add (+1) or subtract (-1) from the stat
  *
- * Walk to the child node of the current node.
- * Returns:
- * True if the walk was possible, false otherwise.
+ * Caller need to have the vm status_lock held. Useful for when multiple update
+ * need to happen at the same time.
  */
-static bool amdgpu_vm_pt_descendant(struct amdgpu_device *adev,
-				    struct amdgpu_vm_pt_cursor *cursor)
+static void amdgpu_vm_update_stats_locked(struct amdgpu_vm_bo_base *base,
+			    struct ttm_resource *res, int sign)
 {
-	unsigned mask, shift, idx;
+	struct amdgpu_vm *vm = base->vm;
+	struct amdgpu_bo *bo = base->bo;
+	int64_t size = sign * amdgpu_bo_size(bo);
+	uint32_t bo_memtype = amdgpu_bo_mem_stats_placement(bo);
 
-	if ((cursor->level == AMDGPU_VM_PTB) || !cursor->entry ||
-	    !cursor->entry->bo)
-		return false;
+	/* For drm-total- and drm-shared-, BO are accounted by their preferred
+	 * placement, see also amdgpu_bo_mem_stats_placement.
+	 */
+	if (base->shared)
+		vm->stats[bo_memtype].drm.shared += size;
+	else
+		vm->stats[bo_memtype].drm.private += size;
 
-	mask = amdgpu_vm_entries_mask(adev, cursor->level);
-	shift = amdgpu_vm_level_shift(adev, cursor->level);
+	if (res && res->mem_type < __AMDGPU_PL_NUM) {
+		uint32_t res_memtype = res->mem_type;
 
-	++cursor->level;
-	idx = (cursor->pfn >> shift) & mask;
-	cursor->parent = cursor->entry;
-	cursor->entry = &to_amdgpu_bo_vm(cursor->entry->bo)->entries[idx];
-	return true;
+		vm->stats[res_memtype].drm.resident += size;
+		/* BO only count as purgeable if it is resident,
+		 * since otherwise there's nothing to purge.
+		 */
+		if (bo->flags & AMDGPU_GEM_CREATE_DISCARDABLE)
+			vm->stats[res_memtype].drm.purgeable += size;
+		if (!(bo->preferred_domains & amdgpu_mem_type_to_domain(res_memtype)))
+			vm->stats[bo_memtype].evicted += size;
+	}
 }
 
 /**
- * amdgpu_vm_pt_sibling - go to sibling node
- *
- * @adev: amdgpu_device pointer
- * @cursor: current state
+ * amdgpu_vm_update_stats - helper to update normal memory stat
+ * @base: base structure for tracking BO usage in a VM
+ * @res:  the ttm_resource to use for the purpose of accounting, may or may not
+ *        be bo->tbo.resource
+ * @sign: if we should add (+1) or subtract (-1) from the stat
  *
- * Walk to the sibling node of the current node.
- * Returns:
- * True if the walk was possible, false otherwise.
+ * Updates the basic memory stat when bo is added/deleted/moved.
  */
-static bool amdgpu_vm_pt_sibling(struct amdgpu_device *adev,
-				 struct amdgpu_vm_pt_cursor *cursor)
+void amdgpu_vm_update_stats(struct amdgpu_vm_bo_base *base,
+			    struct ttm_resource *res, int sign)
 {
-	unsigned shift, num_entries;
-
-	/* Root doesn't have a sibling */
-	if (!cursor->parent)
-		return false;
-
-	/* Go to our parents and see if we got a sibling */
-	shift = amdgpu_vm_level_shift(adev, cursor->level - 1);
-	num_entries = amdgpu_vm_num_entries(adev, cursor->level - 1);
+	struct amdgpu_vm *vm = base->vm;
 
-	if (cursor->entry == &to_amdgpu_bo_vm(cursor->parent->bo)->entries[num_entries - 1])
-		return false;
-
-	cursor->pfn += 1ULL << shift;
-	cursor->pfn &= ~((1ULL << shift) - 1);
-	++cursor->entry;
-	return true;
+	spin_lock(&vm->status_lock);
+	amdgpu_vm_update_stats_locked(base, res, sign);
+	spin_unlock(&vm->status_lock);
 }
 
 /**
- * amdgpu_vm_pt_ancestor - go to parent node
+ * amdgpu_vm_bo_base_init - Adds bo to the list of bos associated with the vm
+ *
+ * @base: base structure for tracking BO usage in a VM
+ * @vm: vm to which bo is to be added
+ * @bo: amdgpu buffer object
  *
- * @cursor: current state
+ * Initialize a bo_va_base structure and add it to the appropriate lists
  *
- * Walk to the parent node of the current node.
- * Returns:
- * True if the walk was possible, false otherwise.
  */
-static bool amdgpu_vm_pt_ancestor(struct amdgpu_vm_pt_cursor *cursor)
+void amdgpu_vm_bo_base_init(struct amdgpu_vm_bo_base *base,
+			    struct amdgpu_vm *vm, struct amdgpu_bo *bo)
 {
-	if (!cursor->parent)
-		return false;
+	base->vm = vm;
+	base->bo = bo;
+	base->next = NULL;
+	INIT_LIST_HEAD(&base->vm_status);
 
-	--cursor->level;
-	cursor->entry = cursor->parent;
-	cursor->parent = amdgpu_vm_pt_parent(cursor->parent);
-	return true;
-}
+	if (!bo)
+		return;
+	base->next = bo->vm_bo;
+	bo->vm_bo = base;
 
-/**
- * amdgpu_vm_pt_next - get next PD/PT in hieratchy
- *
- * @adev: amdgpu_device pointer
- * @cursor: current state
- *
- * Walk the PD/PT tree to the next node.
- */
-static void amdgpu_vm_pt_next(struct amdgpu_device *adev,
-			      struct amdgpu_vm_pt_cursor *cursor)
-{
-	/* First try a newborn child */
-	if (amdgpu_vm_pt_descendant(adev, cursor))
+	spin_lock(&vm->status_lock);
+	base->shared = drm_gem_object_is_shared_for_memory_stats(&bo->tbo.base);
+	amdgpu_vm_update_stats_locked(base, bo->tbo.resource, +1);
+	spin_unlock(&vm->status_lock);
+
+	if (!amdgpu_vm_is_bo_always_valid(vm, bo))
 		return;
 
-	/* If that didn't worked try to find a sibling */
-	while (!amdgpu_vm_pt_sibling(adev, cursor)) {
-		/* No sibling, go to our parents and grandparents */
-		if (!amdgpu_vm_pt_ancestor(cursor)) {
-			cursor->pfn = ~0ll;
-			return;
-		}
-	}
-}
+	dma_resv_assert_held(vm->root.bo->tbo.base.resv);
 
-/**
- * amdgpu_vm_pt_first_dfs - start a deep first search
- *
- * @adev: amdgpu_device structure
- * @vm: amdgpu_vm structure
- * @start: optional cursor to start with
- * @cursor: state to initialize
- *
- * Starts a deep first traversal of the PD/PT tree.
- */
-static void amdgpu_vm_pt_first_dfs(struct amdgpu_device *adev,
-				   struct amdgpu_vm *vm,
-				   struct amdgpu_vm_pt_cursor *start,
-				   struct amdgpu_vm_pt_cursor *cursor)
-{
-	if (start)
-		*cursor = *start;
+	ttm_bo_set_bulk_move(&bo->tbo, &vm->lru_bulk_move);
+	if (bo->tbo.type == ttm_bo_type_kernel && bo->parent)
+		amdgpu_vm_bo_relocated(base);
 	else
-		amdgpu_vm_pt_start(adev, vm, 0, cursor);
-	while (amdgpu_vm_pt_descendant(adev, cursor));
-}
-
-/**
- * amdgpu_vm_pt_continue_dfs - check if the deep first search should continue
- *
- * @start: starting point for the search
- * @entry: current entry
- *
- * Returns:
- * True when the search should continue, false otherwise.
- */
-static bool amdgpu_vm_pt_continue_dfs(struct amdgpu_vm_pt_cursor *start,
-				      struct amdgpu_vm_bo_base *entry)
-{
-	return entry && (!start || entry != start->entry);
-}
+		amdgpu_vm_bo_idle(base);
 
-/**
- * amdgpu_vm_pt_next_dfs - get the next node for a deep first search
- *
- * @adev: amdgpu_device structure
- * @cursor: current state
- *
- * Move the cursor to the next node in a deep first search.
- */
-static void amdgpu_vm_pt_next_dfs(struct amdgpu_device *adev,
-				  struct amdgpu_vm_pt_cursor *cursor)
-{
-	if (!cursor->entry)
+	if (bo->preferred_domains &
+	    amdgpu_mem_type_to_domain(bo->tbo.resource->mem_type))
 		return;
 
-	if (!cursor->parent)
-		cursor->entry = NULL;
-	else if (amdgpu_vm_pt_sibling(adev, cursor))
-		while (amdgpu_vm_pt_descendant(adev, cursor));
-	else
-		amdgpu_vm_pt_ancestor(cursor);
+	/*
+	 * we checked all the prerequisites, but it looks like this per vm bo
+	 * is currently evicted. add the bo to the evicted list to make sure it
+	 * is validated on next vm use to avoid fault.
+	 * */
+	amdgpu_vm_bo_evicted(base);
 }
 
-/*
- * for_each_amdgpu_vm_pt_dfs_safe - safe deep first search of all PDs/PTs
- */
-#define for_each_amdgpu_vm_pt_dfs_safe(adev, vm, start, cursor, entry)		\
-	for (amdgpu_vm_pt_first_dfs((adev), (vm), (start), &(cursor)),		\
-	     (entry) = (cursor).entry, amdgpu_vm_pt_next_dfs((adev), &(cursor));\
-	     amdgpu_vm_pt_continue_dfs((start), (entry));			\
-	     (entry) = (cursor).entry, amdgpu_vm_pt_next_dfs((adev), &(cursor)))
-
 /**
- * amdgpu_vm_get_pd_bo - add the VM PD to a validation list
+ * amdgpu_vm_lock_pd - lock PD in drm_exec
  *
  * @vm: vm providing the BOs
- * @validated: head of validation list
- * @entry: entry to add
+ * @exec: drm execution context
+ * @num_fences: number of extra fences to reserve
  *
- * Add the page directory to the list of BOs to
- * validate for command submission.
+ * Lock the VM root PD in the DRM execution context.
  */
-void amdgpu_vm_get_pd_bo(struct amdgpu_vm *vm,
-			 struct list_head *validated,
-			 struct amdgpu_bo_list_entry *entry)
+int amdgpu_vm_lock_pd(struct amdgpu_vm *vm, struct drm_exec *exec,
+		      unsigned int num_fences)
 {
-	entry->priority = 0;
-	entry->tv.bo = &vm->root.bo->tbo;
-	/* Two for VM updates, one for TTM and one for the CS job */
-	entry->tv.num_shared = 4;
-	entry->user_pages = NULL;
-	list_add(&entry->tv.head, validated);
+	/* We need at least two fences for the VM PD/PT updates */
+	return drm_exec_prepare_obj(exec, &vm->root.bo->tbo.base,
+				    2 + num_fences);
 }
 
 /**
- * amdgpu_vm_del_from_lru_notify - update bulk_moveable flag
- *
- * @bo: BO which was removed from the LRU
+ * amdgpu_vm_lock_done_list - lock all BOs on the done list
+ * @vm: vm providing the BOs
+ * @exec: drm execution context
+ * @num_fences: number of extra fences to reserve
  *
- * Make sure the bulk_moveable flag is updated when a BO is removed from the
- * LRU.
+ * Lock the BOs on the done list in the DRM execution context.
  */
-void amdgpu_vm_del_from_lru_notify(struct ttm_buffer_object *bo)
+int amdgpu_vm_lock_done_list(struct amdgpu_vm *vm, struct drm_exec *exec,
+			     unsigned int num_fences)
 {
-	struct amdgpu_bo *abo;
-	struct amdgpu_vm_bo_base *bo_base;
+	struct list_head *prev = &vm->done;
+	struct amdgpu_bo_va *bo_va;
+	struct amdgpu_bo *bo;
+	int ret;
 
-	if (!amdgpu_bo_is_amdgpu_bo(bo))
-		return;
+	/* We can only trust prev->next while holding the lock */
+	spin_lock(&vm->status_lock);
+	while (!list_is_head(prev->next, &vm->done)) {
+		bo_va = list_entry(prev->next, typeof(*bo_va), base.vm_status);
 
-	if (bo->pin_count)
-		return;
+		bo = bo_va->base.bo;
+		if (bo) {
+			amdgpu_bo_ref(bo);
+			spin_unlock(&vm->status_lock);
 
-	abo = ttm_to_amdgpu_bo(bo);
-	if (!abo->parent)
-		return;
-	for (bo_base = abo->vm_bo; bo_base; bo_base = bo_base->next) {
-		struct amdgpu_vm *vm = bo_base->vm;
+			ret = drm_exec_prepare_obj(exec, &bo->tbo.base, 1);
+			amdgpu_bo_unref(&bo);
+			if (unlikely(ret))
+				return ret;
 
-		if (abo->tbo.base.resv == vm->root.bo->tbo.base.resv)
-			vm->bulk_moveable = false;
+			spin_lock(&vm->status_lock);
+		}
+		prev = prev->next;
 	}
+	spin_unlock(&vm->status_lock);
 
+	return 0;
 }
+
 /**
  * amdgpu_vm_move_to_lru_tail - move all BOs to the end of LRU
  *
@@ -679,71 +516,110 @@ void amdgpu_vm_del_from_lru_notify(struct ttm_buffer_object *bo)
 void amdgpu_vm_move_to_lru_tail(struct amdgpu_device *adev,
 				struct amdgpu_vm *vm)
 {
-	struct amdgpu_vm_bo_base *bo_base;
+	spin_lock(&adev->mman.bdev.lru_lock);
+	ttm_lru_bulk_move_tail(&vm->lru_bulk_move);
+	spin_unlock(&adev->mman.bdev.lru_lock);
+}
 
-	if (vm->bulk_moveable) {
-		spin_lock(&adev->mman.bdev.lru_lock);
-		ttm_bo_bulk_move_lru_tail(&vm->lru_bulk_move);
-		spin_unlock(&adev->mman.bdev.lru_lock);
-		return;
-	}
+/* Create scheduler entities for page table updates */
+static int amdgpu_vm_init_entities(struct amdgpu_device *adev,
+				   struct amdgpu_vm *vm)
+{
+	int r;
 
-	memset(&vm->lru_bulk_move, 0, sizeof(vm->lru_bulk_move));
+	r = drm_sched_entity_init(&vm->immediate, DRM_SCHED_PRIORITY_NORMAL,
+				  adev->vm_manager.vm_pte_scheds,
+				  adev->vm_manager.vm_pte_num_scheds, NULL);
+	if (r)
+		goto error;
 
-	spin_lock(&adev->mman.bdev.lru_lock);
-	list_for_each_entry(bo_base, &vm->idle, vm_status) {
-		struct amdgpu_bo *bo = bo_base->bo;
-		struct amdgpu_bo *shadow = amdgpu_bo_shadowed(bo);
+	return drm_sched_entity_init(&vm->delayed, DRM_SCHED_PRIORITY_NORMAL,
+				     adev->vm_manager.vm_pte_scheds,
+				     adev->vm_manager.vm_pte_num_scheds, NULL);
 
-		if (!bo->parent)
-			continue;
+error:
+	drm_sched_entity_destroy(&vm->immediate);
+	return r;
+}
 
-		ttm_bo_move_to_lru_tail(&bo->tbo, bo->tbo.resource,
-					&vm->lru_bulk_move);
-		if (shadow)
-			ttm_bo_move_to_lru_tail(&shadow->tbo,
-						shadow->tbo.resource,
-						&vm->lru_bulk_move);
-	}
-	spin_unlock(&adev->mman.bdev.lru_lock);
+/* Destroy the entities for page table updates again */
+static void amdgpu_vm_fini_entities(struct amdgpu_vm *vm)
+{
+	drm_sched_entity_destroy(&vm->immediate);
+	drm_sched_entity_destroy(&vm->delayed);
+}
 
-	vm->bulk_moveable = true;
+/**
+ * amdgpu_vm_generation - return the page table re-generation counter
+ * @adev: the amdgpu_device
+ * @vm: optional VM to check, might be NULL
+ *
+ * Returns a page table re-generation token to allow checking if submissions
+ * are still valid to use this VM. The VM parameter might be NULL in which case
+ * just the VRAM lost counter will be used.
+ */
+uint64_t amdgpu_vm_generation(struct amdgpu_device *adev, struct amdgpu_vm *vm)
+{
+	uint64_t result = (u64)atomic_read(&adev->vram_lost_counter) << 32;
+
+	if (!vm)
+		return result;
+
+	result += lower_32_bits(vm->generation);
+	/* Add one if the page tables will be re-generated on next CS */
+	if (drm_sched_entity_error(&vm->delayed))
+		++result;
+
+	return result;
 }
 
 /**
- * amdgpu_vm_validate_pt_bos - validate the page table BOs
+ * amdgpu_vm_validate - validate evicted BOs tracked in the VM
  *
  * @adev: amdgpu device pointer
  * @vm: vm providing the BOs
+ * @ticket: optional reservation ticket used to reserve the VM
  * @validate: callback to do the validation
  * @param: parameter for the validation callback
  *
- * Validate the page table BOs on command submission if neccessary.
+ * Validate the page table BOs and per-VM BOs on command submission if
+ * necessary. If a ticket is given, also try to validate evicted user queue
+ * BOs. They must already be reserved with the given ticket.
  *
  * Returns:
  * Validation result.
  */
-int amdgpu_vm_validate_pt_bos(struct amdgpu_device *adev, struct amdgpu_vm *vm,
-			      int (*validate)(void *p, struct amdgpu_bo *bo),
-			      void *param)
+int amdgpu_vm_validate(struct amdgpu_device *adev, struct amdgpu_vm *vm,
+		       struct ww_acquire_ctx *ticket,
+		       int (*validate)(void *p, struct amdgpu_bo *bo),
+		       void *param)
 {
-	struct amdgpu_vm_bo_base *bo_base, *tmp;
+	uint64_t new_vm_generation = amdgpu_vm_generation(adev, vm);
+	struct amdgpu_vm_bo_base *bo_base;
+	struct amdgpu_bo *bo;
 	int r;
 
-	vm->bulk_moveable &= list_empty(&vm->evicted);
+	if (vm->generation != new_vm_generation) {
+		vm->generation = new_vm_generation;
+		amdgpu_vm_bo_reset_state_machine(vm);
+		amdgpu_vm_fini_entities(vm);
+		r = amdgpu_vm_init_entities(adev, vm);
+		if (r)
+			return r;
+	}
+
+	spin_lock(&vm->status_lock);
+	while (!list_empty(&vm->evicted)) {
+		bo_base = list_first_entry(&vm->evicted,
+					   struct amdgpu_vm_bo_base,
+					   vm_status);
+		spin_unlock(&vm->status_lock);
 
-	list_for_each_entry_safe(bo_base, tmp, &vm->evicted, vm_status) {
-		struct amdgpu_bo *bo = bo_base->bo;
-		struct amdgpu_bo *shadow = amdgpu_bo_shadowed(bo);
+		bo = bo_base->bo;
 
 		r = validate(param, bo);
 		if (r)
 			return r;
-		if (shadow) {
-			r = validate(param, shadow);
-			if (r)
-				return r;
-		}
 
 		if (bo->tbo.type != ttm_bo_type_kernel) {
 			amdgpu_vm_bo_moved(bo_base);
@@ -751,334 +627,67 @@ int amdgpu_vm_validate_pt_bos(struct amdgpu_device *adev, struct amdgpu_vm *vm,
 			vm->update_funcs->map_table(to_amdgpu_bo_vm(bo));
 			amdgpu_vm_bo_relocated(bo_base);
 		}
+		spin_lock(&vm->status_lock);
 	}
+	while (ticket && !list_empty(&vm->evicted_user)) {
+		bo_base = list_first_entry(&vm->evicted_user,
+					   struct amdgpu_vm_bo_base,
+					   vm_status);
+		spin_unlock(&vm->status_lock);
 
-	amdgpu_vm_eviction_lock(vm);
-	vm->evicting = false;
-	amdgpu_vm_eviction_unlock(vm);
-
-	return 0;
-}
-
-/**
- * amdgpu_vm_ready - check VM is ready for updates
- *
- * @vm: VM to check
- *
- * Check if all VM PDs/PTs are ready for updates
- *
- * Returns:
- * True if eviction list is empty.
- */
-bool amdgpu_vm_ready(struct amdgpu_vm *vm)
-{
-	return list_empty(&vm->evicted);
-}
-
-/**
- * amdgpu_vm_clear_bo - initially clear the PDs/PTs
- *
- * @adev: amdgpu_device pointer
- * @vm: VM to clear BO from
- * @vmbo: BO to clear
- * @immediate: use an immediate update
- *
- * Root PD needs to be reserved when calling this.
- *
- * Returns:
- * 0 on success, errno otherwise.
- */
-static int amdgpu_vm_clear_bo(struct amdgpu_device *adev,
-			      struct amdgpu_vm *vm,
-			      struct amdgpu_bo_vm *vmbo,
-			      bool immediate)
-{
-	struct ttm_operation_ctx ctx = { true, false };
-	unsigned level = adev->vm_manager.root_level;
-	struct amdgpu_vm_update_params params;
-	struct amdgpu_bo *ancestor = &vmbo->bo;
-	struct amdgpu_bo *bo = &vmbo->bo;
-	unsigned entries, ats_entries;
-	uint64_t addr;
-	int r, idx;
-
-	/* Figure out our place in the hierarchy */
-	if (ancestor->parent) {
-		++level;
-		while (ancestor->parent->parent) {
-			++level;
-			ancestor = ancestor->parent;
-		}
-	}
-
-	entries = amdgpu_bo_size(bo) / 8;
-	if (!vm->pte_support_ats) {
-		ats_entries = 0;
-
-	} else if (!bo->parent) {
-		ats_entries = amdgpu_vm_num_ats_entries(adev);
-		ats_entries = min(ats_entries, entries);
-		entries -= ats_entries;
-
-	} else {
-		struct amdgpu_vm_bo_base *pt;
-
-		pt = ancestor->vm_bo;
-		ats_entries = amdgpu_vm_num_ats_entries(adev);
-		if ((pt - to_amdgpu_bo_vm(vm->root.bo)->entries) >= ats_entries) {
-			ats_entries = 0;
-		} else {
-			ats_entries = entries;
-			entries = 0;
-		}
-	}
-
-	r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
-	if (r)
-		return r;
-
-	if (vmbo->shadow) {
-		struct amdgpu_bo *shadow = vmbo->shadow;
+		bo = bo_base->bo;
+		dma_resv_assert_held(bo->tbo.base.resv);
 
-		r = ttm_bo_validate(&shadow->tbo, &shadow->placement, &ctx);
+		r = validate(param, bo);
 		if (r)
 			return r;
-	}
-
-	if (!drm_dev_enter(adev_to_drm(adev), &idx))
-		return -ENODEV;
-
-	r = vm->update_funcs->map_table(vmbo);
-	if (r)
-		goto exit;
-
-	memset(&params, 0, sizeof(params));
-	params.adev = adev;
-	params.vm = vm;
-	params.immediate = immediate;
-
-	r = vm->update_funcs->prepare(&params, NULL, AMDGPU_SYNC_EXPLICIT);
-	if (r)
-		goto exit;
 
-	addr = 0;
-	if (ats_entries) {
-		uint64_t value = 0, flags;
+		amdgpu_vm_bo_invalidated(bo_base);
 
-		flags = AMDGPU_PTE_DEFAULT_ATC;
-		if (level != AMDGPU_VM_PTB) {
-			/* Handle leaf PDEs as PTEs */
-			flags |= AMDGPU_PDE_PTE;
-			amdgpu_gmc_get_vm_pde(adev, level, &value, &flags);
-		}
-
-		r = vm->update_funcs->update(&params, vmbo, addr, 0, ats_entries,
-					     value, flags);
-		if (r)
-			goto exit;
-
-		addr += ats_entries * 8;
+		spin_lock(&vm->status_lock);
 	}
+	spin_unlock(&vm->status_lock);
 
-	if (entries) {
-		uint64_t value = 0, flags = 0;
-
-		if (adev->asic_type >= CHIP_VEGA10) {
-			if (level != AMDGPU_VM_PTB) {
-				/* Handle leaf PDEs as PTEs */
-				flags |= AMDGPU_PDE_PTE;
-				amdgpu_gmc_get_vm_pde(adev, level,
-						      &value, &flags);
-			} else {
-				/* Workaround for fault priority problem on GMC9 */
-				flags = AMDGPU_PTE_EXECUTABLE;
-			}
-		}
-
-		r = vm->update_funcs->update(&params, vmbo, addr, 0, entries,
-					     value, flags);
-		if (r)
-			goto exit;
-	}
-
-	r = vm->update_funcs->commit(&params, NULL);
-exit:
-	drm_dev_exit(idx);
-	return r;
-}
-
-/**
- * amdgpu_vm_pt_create - create bo for PD/PT
- *
- * @adev: amdgpu_device pointer
- * @vm: requesting vm
- * @level: the page table level
- * @immediate: use a immediate update
- * @vmbo: pointer to the buffer object pointer
- */
-static int amdgpu_vm_pt_create(struct amdgpu_device *adev,
-			       struct amdgpu_vm *vm,
-			       int level, bool immediate,
-			       struct amdgpu_bo_vm **vmbo)
-{
-	struct amdgpu_bo_param bp;
-	struct amdgpu_bo *bo;
-	struct dma_resv *resv;
-	unsigned int num_entries;
-	int r;
-
-	memset(&bp, 0, sizeof(bp));
-
-	bp.size = amdgpu_vm_bo_size(adev, level);
-	bp.byte_align = AMDGPU_GPU_PAGE_SIZE;
-	bp.domain = AMDGPU_GEM_DOMAIN_VRAM;
-	bp.domain = amdgpu_bo_get_preferred_domain(adev, bp.domain);
-	bp.flags = AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS |
-		AMDGPU_GEM_CREATE_CPU_GTT_USWC;
-
-	if (level < AMDGPU_VM_PTB)
-		num_entries = amdgpu_vm_num_entries(adev, level);
-	else
-		num_entries = 0;
-
-	bp.bo_ptr_size = struct_size((*vmbo), entries, num_entries);
-
-	if (vm->use_cpu_for_update)
-		bp.flags |= AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED;
-
-	bp.type = ttm_bo_type_kernel;
-	bp.no_wait_gpu = immediate;
-	if (vm->root.bo)
-		bp.resv = vm->root.bo->tbo.base.resv;
-
-	r = amdgpu_bo_create_vm(adev, &bp, vmbo);
-	if (r)
-		return r;
-
-	bo = &(*vmbo)->bo;
-	if (vm->is_compute_context || (adev->flags & AMD_IS_APU)) {
-		(*vmbo)->shadow = NULL;
-		return 0;
-	}
-
-	if (!bp.resv)
-		WARN_ON(dma_resv_lock(bo->tbo.base.resv,
-				      NULL));
-	resv = bp.resv;
-	memset(&bp, 0, sizeof(bp));
-	bp.size = amdgpu_vm_bo_size(adev, level);
-	bp.domain = AMDGPU_GEM_DOMAIN_GTT;
-	bp.flags = AMDGPU_GEM_CREATE_CPU_GTT_USWC;
-	bp.type = ttm_bo_type_kernel;
-	bp.resv = bo->tbo.base.resv;
-	bp.bo_ptr_size = sizeof(struct amdgpu_bo);
-
-	r = amdgpu_bo_create(adev, &bp, &(*vmbo)->shadow);
-
-	if (!resv)
-		dma_resv_unlock(bo->tbo.base.resv);
-
-	if (r) {
-		amdgpu_bo_unref(&bo);
-		return r;
-	}
-
-	(*vmbo)->shadow->parent = amdgpu_bo_ref(bo);
-	amdgpu_bo_add_to_shadow_list(*vmbo);
+	amdgpu_vm_eviction_lock(vm);
+	vm->evicting = false;
+	amdgpu_vm_eviction_unlock(vm);
 
 	return 0;
 }
 
 /**
- * amdgpu_vm_alloc_pts - Allocate a specific page table
+ * amdgpu_vm_ready - check VM is ready for updates
  *
- * @adev: amdgpu_device pointer
- * @vm: VM to allocate page tables for
- * @cursor: Which page table to allocate
- * @immediate: use an immediate update
+ * @vm: VM to check
  *
- * Make sure a specific page table or directory is allocated.
+ * Check if all VM PDs/PTs are ready for updates
  *
  * Returns:
- * 1 if page table needed to be allocated, 0 if page table was already
- * allocated, negative errno if an error occurred.
+ * True if VM is not evicting and all VM entities are not stopped
  */
-static int amdgpu_vm_alloc_pts(struct amdgpu_device *adev,
-			       struct amdgpu_vm *vm,
-			       struct amdgpu_vm_pt_cursor *cursor,
-			       bool immediate)
+bool amdgpu_vm_ready(struct amdgpu_vm *vm)
 {
-	struct amdgpu_vm_bo_base *entry = cursor->entry;
-	struct amdgpu_bo *pt_bo;
-	struct amdgpu_bo_vm *pt;
-	int r;
-
-	if (entry->bo)
-		return 0;
-
-	r = amdgpu_vm_pt_create(adev, vm, cursor->level, immediate, &pt);
-	if (r)
-		return r;
-
-	/* Keep a reference to the root directory to avoid
-	 * freeing them up in the wrong order.
-	 */
-	pt_bo = &pt->bo;
-	pt_bo->parent = amdgpu_bo_ref(cursor->parent->bo);
-	amdgpu_vm_bo_base_init(entry, vm, pt_bo);
-	r = amdgpu_vm_clear_bo(adev, vm, pt, immediate);
-	if (r)
-		goto error_free_pt;
+	bool ret;
 
-	return 0;
-
-error_free_pt:
-	amdgpu_bo_unref(&pt->shadow);
-	amdgpu_bo_unref(&pt_bo);
-	return r;
-}
+	amdgpu_vm_assert_locked(vm);
 
-/**
- * amdgpu_vm_free_table - fre one PD/PT
- *
- * @entry: PDE to free
- */
-static void amdgpu_vm_free_table(struct amdgpu_vm_bo_base *entry)
-{
-	struct amdgpu_bo *shadow;
-
-	if (!entry->bo)
-		return;
-	shadow = amdgpu_bo_shadowed(entry->bo);
-	entry->bo->vm_bo = NULL;
-	list_del(&entry->vm_status);
-	amdgpu_bo_unref(&shadow);
-	amdgpu_bo_unref(&entry->bo);
-}
+	amdgpu_vm_eviction_lock(vm);
+	ret = !vm->evicting;
+	amdgpu_vm_eviction_unlock(vm);
 
-/**
- * amdgpu_vm_free_pts - free PD/PT levels
- *
- * @adev: amdgpu device structure
- * @vm: amdgpu vm structure
- * @start: optional cursor where to start freeing PDs/PTs
- *
- * Free the page directory or page table level and all sub levels.
- */
-static void amdgpu_vm_free_pts(struct amdgpu_device *adev,
-			       struct amdgpu_vm *vm,
-			       struct amdgpu_vm_pt_cursor *start)
-{
-	struct amdgpu_vm_pt_cursor cursor;
-	struct amdgpu_vm_bo_base *entry;
+	spin_lock(&vm->status_lock);
+	ret &= list_empty(&vm->evicted);
+	spin_unlock(&vm->status_lock);
 
-	vm->bulk_moveable = false;
+	spin_lock(&vm->immediate.lock);
+	ret &= !vm->immediate.stopped;
+	spin_unlock(&vm->immediate.lock);
 
-	for_each_amdgpu_vm_pt_dfs_safe(adev, vm, start, cursor, entry)
-		amdgpu_vm_free_table(entry);
+	spin_lock(&vm->delayed.lock);
+	ret &= !vm->delayed.stopped;
+	spin_unlock(&vm->delayed.lock);
 
-	if (start)
-		amdgpu_vm_free_table(start->entry);
+	return ret;
 }
 
 /**
@@ -1129,27 +738,22 @@ bool amdgpu_vm_need_pipeline_sync(struct amdgpu_ring *ring,
 				  struct amdgpu_job *job)
 {
 	struct amdgpu_device *adev = ring->adev;
-	unsigned vmhub = ring->funcs->vmhub;
+	unsigned vmhub = ring->vm_hub;
 	struct amdgpu_vmid_mgr *id_mgr = &adev->vm_manager.id_mgr[vmhub];
-	struct amdgpu_vmid *id;
-	bool gds_switch_needed;
-	bool vm_flush_needed = job->vm_needs_flush || ring->has_compute_vm_bug;
 
 	if (job->vmid == 0)
 		return false;
-	id = &id_mgr->ids[job->vmid];
-	gds_switch_needed = ring->funcs->emit_gds_switch && (
-		id->gds_base != job->gds_base ||
-		id->gds_size != job->gds_size ||
-		id->gws_base != job->gws_base ||
-		id->gws_size != job->gws_size ||
-		id->oa_base != job->oa_base ||
-		id->oa_size != job->oa_size);
-
-	if (amdgpu_vmid_had_gpu_reset(adev, id))
+
+	if (job->vm_needs_flush || ring->has_compute_vm_bug)
+		return true;
+
+	if (ring->funcs->emit_gds_switch && job->gds_switch_needed)
 		return true;
 
-	return vm_flush_needed || gds_switch_needed;
+	if (amdgpu_vmid_had_gpu_reset(adev, &id_mgr->ids[job->vmid]))
+		return true;
+
+	return false;
 }
 
 /**
@@ -1168,30 +772,25 @@ int amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_job *job,
 		    bool need_pipe_sync)
 {
 	struct amdgpu_device *adev = ring->adev;
-	unsigned vmhub = ring->funcs->vmhub;
+	struct amdgpu_isolation *isolation = &adev->isolation[ring->xcp_id];
+	unsigned vmhub = ring->vm_hub;
 	struct amdgpu_vmid_mgr *id_mgr = &adev->vm_manager.id_mgr[vmhub];
 	struct amdgpu_vmid *id = &id_mgr->ids[job->vmid];
-	bool gds_switch_needed = ring->funcs->emit_gds_switch && (
-		id->gds_base != job->gds_base ||
-		id->gds_size != job->gds_size ||
-		id->gws_base != job->gws_base ||
-		id->gws_size != job->gws_size ||
-		id->oa_base != job->oa_base ||
-		id->oa_size != job->oa_size);
+	bool spm_update_needed = job->spm_update_needed;
+	bool gds_switch_needed = ring->funcs->emit_gds_switch &&
+		job->gds_switch_needed;
 	bool vm_flush_needed = job->vm_needs_flush;
-	struct dma_fence *fence = NULL;
+	bool cleaner_shader_needed = false;
 	bool pasid_mapping_needed = false;
-	unsigned patch_offset = 0;
-	bool update_spm_vmid_needed = (job->vm && (job->vm->reserved_vmid[vmhub] != NULL));
+	struct dma_fence *fence = NULL;
+	unsigned int patch;
 	int r;
 
-	if (update_spm_vmid_needed && adev->gfx.rlc.funcs->update_spm_vmid)
-		adev->gfx.rlc.funcs->update_spm_vmid(adev, job->vmid);
-
 	if (amdgpu_vmid_had_gpu_reset(adev, id)) {
 		gds_switch_needed = true;
 		vm_flush_needed = true;
 		pasid_mapping_needed = true;
+		spm_update_needed = true;
 	}
 
 	mutex_lock(&id_mgr->lock);
@@ -1206,15 +805,26 @@ int amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_job *job,
 	pasid_mapping_needed &= adev->gmc.gmc_funcs->emit_pasid_mapping &&
 		ring->funcs->emit_wreg;
 
-	if (!vm_flush_needed && !gds_switch_needed && !need_pipe_sync)
+	cleaner_shader_needed = job->run_cleaner_shader &&
+		adev->gfx.enable_cleaner_shader &&
+		ring->funcs->emit_cleaner_shader && job->base.s_fence &&
+		&job->base.s_fence->scheduled == isolation->spearhead;
+
+	if (!vm_flush_needed && !gds_switch_needed && !need_pipe_sync &&
+	    !cleaner_shader_needed)
 		return 0;
 
+	amdgpu_ring_ib_begin(ring);
 	if (ring->funcs->init_cond_exec)
-		patch_offset = amdgpu_ring_init_cond_exec(ring);
+		patch = amdgpu_ring_init_cond_exec(ring,
+						   ring->cond_exe_gpu_addr);
 
 	if (need_pipe_sync)
 		amdgpu_ring_emit_pipeline_sync(ring);
 
+	if (cleaner_shader_needed)
+		ring->funcs->emit_cleaner_shader(ring);
+
 	if (vm_flush_needed) {
 		trace_amdgpu_vm_flush(ring, job->vmid, job->vm_pd_addr);
 		amdgpu_ring_emit_vm_flush(ring, job->vmid, job->vm_pd_addr);
@@ -1223,10 +833,24 @@ int amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_job *job,
 	if (pasid_mapping_needed)
 		amdgpu_gmc_emit_pasid_mapping(ring, job->vmid, job->pasid);
 
-	if (vm_flush_needed || pasid_mapping_needed) {
-		r = amdgpu_fence_emit(ring, &fence, NULL, 0);
+	if (spm_update_needed && adev->gfx.rlc.funcs->update_spm_vmid)
+		adev->gfx.rlc.funcs->update_spm_vmid(adev, ring, job->vmid);
+
+	if (ring->funcs->emit_gds_switch &&
+	    gds_switch_needed) {
+		amdgpu_ring_emit_gds_switch(ring, job->vmid, job->gds_base,
+					    job->gds_size, job->gws_base,
+					    job->gws_size, job->oa_base,
+					    job->oa_size);
+	}
+
+	if (vm_flush_needed || pasid_mapping_needed || cleaner_shader_needed) {
+		r = amdgpu_fence_emit(ring, job->hw_vm_fence, 0);
 		if (r)
 			return r;
+		fence = &job->hw_vm_fence->base;
+		/* get a ref for the job */
+		dma_fence_get(fence);
 	}
 
 	if (vm_flush_needed) {
@@ -1245,29 +869,29 @@ int amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_job *job,
 		id->pasid_mapping = dma_fence_get(fence);
 		mutex_unlock(&id_mgr->lock);
 	}
-	dma_fence_put(fence);
 
-	if (ring->funcs->emit_gds_switch && gds_switch_needed) {
-		id->gds_base = job->gds_base;
-		id->gds_size = job->gds_size;
-		id->gws_base = job->gws_base;
-		id->gws_size = job->gws_size;
-		id->oa_base = job->oa_base;
-		id->oa_size = job->oa_size;
-		amdgpu_ring_emit_gds_switch(ring, job->vmid, job->gds_base,
-					    job->gds_size, job->gws_base,
-					    job->gws_size, job->oa_base,
-					    job->oa_size);
+	/*
+	 * Make sure that all other submissions wait for the cleaner shader to
+	 * finish before we push them to the HW.
+	 */
+	if (cleaner_shader_needed) {
+		trace_amdgpu_cleaner_shader(ring, fence);
+		mutex_lock(&adev->enforce_isolation_mutex);
+		dma_fence_put(isolation->spearhead);
+		isolation->spearhead = dma_fence_get(fence);
+		mutex_unlock(&adev->enforce_isolation_mutex);
 	}
+	dma_fence_put(fence);
 
-	if (ring->funcs->patch_cond_exec)
-		amdgpu_ring_patch_cond_exec(ring, patch_offset);
+	amdgpu_ring_patch_cond_exec(ring, patch);
 
 	/* the double SWITCH_BUFFER here *cannot* be skipped by COND_EXEC */
 	if (ring->funcs->emit_switch_buffer) {
 		amdgpu_ring_emit_switch_buffer(ring);
 		amdgpu_ring_emit_switch_buffer(ring);
 	}
+
+	amdgpu_ring_ib_end(ring);
 	return 0;
 }
 
@@ -1328,53 +952,6 @@ uint64_t amdgpu_vm_map_gart(const dma_addr_t *pages_addr, uint64_t addr)
 }
 
 /**
- * amdgpu_vm_update_pde - update a single level in the hierarchy
- *
- * @params: parameters for the update
- * @vm: requested vm
- * @entry: entry to update
- *
- * Makes sure the requested entry in parent is up to date.
- */
-static int amdgpu_vm_update_pde(struct amdgpu_vm_update_params *params,
-				struct amdgpu_vm *vm,
-				struct amdgpu_vm_bo_base *entry)
-{
-	struct amdgpu_vm_bo_base *parent = amdgpu_vm_pt_parent(entry);
-	struct amdgpu_bo *bo = parent->bo, *pbo;
-	uint64_t pde, pt, flags;
-	unsigned level;
-
-	for (level = 0, pbo = bo->parent; pbo; ++level)
-		pbo = pbo->parent;
-
-	level += params->adev->vm_manager.root_level;
-	amdgpu_gmc_get_pde_for_bo(entry->bo, level, &pt, &flags);
-	pde = (entry - to_amdgpu_bo_vm(parent->bo)->entries) * 8;
-	return vm->update_funcs->update(params, to_amdgpu_bo_vm(bo), pde, pt,
-					1, 0, flags);
-}
-
-/**
- * amdgpu_vm_invalidate_pds - mark all PDs as invalid
- *
- * @adev: amdgpu_device pointer
- * @vm: related vm
- *
- * Mark all PD level as invalid after an error.
- */
-static void amdgpu_vm_invalidate_pds(struct amdgpu_device *adev,
-				     struct amdgpu_vm *vm)
-{
-	struct amdgpu_vm_pt_cursor cursor;
-	struct amdgpu_vm_bo_base *entry;
-
-	for_each_amdgpu_vm_pt_dfs_safe(adev, vm, NULL, cursor, entry)
-		if (entry->bo && !entry->moved)
-			amdgpu_vm_bo_relocated(entry);
-}
-
-/**
  * amdgpu_vm_update_pdes - make sure that all directories are valid
  *
  * @adev: amdgpu_device pointer
@@ -1390,9 +967,18 @@ int amdgpu_vm_update_pdes(struct amdgpu_device *adev,
 			  struct amdgpu_vm *vm, bool immediate)
 {
 	struct amdgpu_vm_update_params params;
+	struct amdgpu_vm_bo_base *entry;
+	bool flush_tlb_needed = false;
+	LIST_HEAD(relocated);
 	int r, idx;
 
-	if (list_empty(&vm->relocated))
+	amdgpu_vm_assert_locked(vm);
+
+	spin_lock(&vm->status_lock);
+	list_splice_init(&vm->relocated, &relocated);
+	spin_unlock(&vm->status_lock);
+
+	if (list_empty(&relocated))
 		return 0;
 
 	if (!drm_dev_enter(adev_to_drm(adev), &idx))
@@ -1403,19 +989,16 @@ int amdgpu_vm_update_pdes(struct amdgpu_device *adev,
 	params.vm = vm;
 	params.immediate = immediate;
 
-	r = vm->update_funcs->prepare(&params, NULL, AMDGPU_SYNC_EXPLICIT);
+	r = vm->update_funcs->prepare(&params, NULL,
+				      AMDGPU_KERNEL_JOB_ID_VM_UPDATE_PDES);
 	if (r)
-		goto exit;
-
-	while (!list_empty(&vm->relocated)) {
-		struct amdgpu_vm_bo_base *entry;
+		goto error;
 
-		entry = list_first_entry(&vm->relocated,
-					 struct amdgpu_vm_bo_base,
-					 vm_status);
-		amdgpu_vm_bo_idle(entry);
+	list_for_each_entry(entry, &relocated, vm_status) {
+		/* vm_flush_needed after updating moved PDEs */
+		flush_tlb_needed |= entry->moved;
 
-		r = amdgpu_vm_update_pde(&params, vm, entry);
+		r = amdgpu_vm_pde_update(&params, entry);
 		if (r)
 			goto error;
 	}
@@ -1423,323 +1006,149 @@ int amdgpu_vm_update_pdes(struct amdgpu_device *adev,
 	r = vm->update_funcs->commit(&params, &vm->last_update);
 	if (r)
 		goto error;
-	drm_dev_exit(idx);
-	return 0;
+
+	if (flush_tlb_needed)
+		atomic64_inc(&vm->tlb_seq);
+
+	while (!list_empty(&relocated)) {
+		entry = list_first_entry(&relocated, struct amdgpu_vm_bo_base,
+					 vm_status);
+		amdgpu_vm_bo_idle(entry);
+	}
 
 error:
-	amdgpu_vm_invalidate_pds(adev, vm);
-exit:
 	drm_dev_exit(idx);
 	return r;
 }
 
-/*
- * amdgpu_vm_update_flags - figure out flags for PTE updates
+/**
+ * amdgpu_vm_tlb_seq_cb - make sure to increment tlb sequence
+ * @fence: unused
+ * @cb: the callback structure
  *
- * Make sure to set the right flags for the PTEs at the desired level.
+ * Increments the tlb sequence to make sure that future CS execute a VM flush.
  */
-static void amdgpu_vm_update_flags(struct amdgpu_vm_update_params *params,
-				   struct amdgpu_bo_vm *pt, unsigned int level,
-				   uint64_t pe, uint64_t addr,
-				   unsigned int count, uint32_t incr,
-				   uint64_t flags)
-
+static void amdgpu_vm_tlb_seq_cb(struct dma_fence *fence,
+				 struct dma_fence_cb *cb)
 {
-	if (level != AMDGPU_VM_PTB) {
-		flags |= AMDGPU_PDE_PTE;
-		amdgpu_gmc_get_vm_pde(params->adev, level, &addr, &flags);
-
-	} else if (params->adev->asic_type >= CHIP_VEGA10 &&
-		   !(flags & AMDGPU_PTE_VALID) &&
-		   !(flags & AMDGPU_PTE_PRT)) {
+	struct amdgpu_vm_tlb_seq_struct *tlb_cb;
 
-		/* Workaround for fault priority problem on GMC9 */
-		flags |= AMDGPU_PTE_EXECUTABLE;
-	}
-
-	params->vm->update_funcs->update(params, pt, pe, addr, count, incr,
-					 flags);
+	tlb_cb = container_of(cb, typeof(*tlb_cb), cb);
+	atomic64_inc(&tlb_cb->vm->tlb_seq);
+	kfree(tlb_cb);
 }
 
 /**
- * amdgpu_vm_fragment - get fragment for PTEs
+ * amdgpu_vm_tlb_flush - prepare TLB flush
  *
- * @params: see amdgpu_vm_update_params definition
- * @start: first PTE to handle
- * @end: last PTE to handle
- * @flags: hw mapping flags
- * @frag: resulting fragment size
- * @frag_end: end of this fragment
+ * @params: parameters for update
+ * @fence: input fence to sync TLB flush with
+ * @tlb_cb: the callback structure
  *
- * Returns the first possible fragment for the start and end address.
+ * Increments the tlb sequence to make sure that future CS execute a VM flush.
  */
-static void amdgpu_vm_fragment(struct amdgpu_vm_update_params *params,
-			       uint64_t start, uint64_t end, uint64_t flags,
-			       unsigned int *frag, uint64_t *frag_end)
+static void
+amdgpu_vm_tlb_flush(struct amdgpu_vm_update_params *params,
+		    struct dma_fence **fence,
+		    struct amdgpu_vm_tlb_seq_struct *tlb_cb)
 {
-	/**
-	 * The MC L1 TLB supports variable sized pages, based on a fragment
-	 * field in the PTE. When this field is set to a non-zero value, page
-	 * granularity is increased from 4KB to (1 << (12 + frag)). The PTE
-	 * flags are considered valid for all PTEs within the fragment range
-	 * and corresponding mappings are assumed to be physically contiguous.
-	 *
-	 * The L1 TLB can store a single PTE for the whole fragment,
-	 * significantly increasing the space available for translation
-	 * caching. This leads to large improvements in throughput when the
-	 * TLB is under pressure.
-	 *
-	 * The L2 TLB distributes small and large fragments into two
-	 * asymmetric partitions. The large fragment cache is significantly
-	 * larger. Thus, we try to use large fragments wherever possible.
-	 * Userspace can support this by aligning virtual base address and
-	 * allocation size to the fragment size.
-	 *
-	 * Starting with Vega10 the fragment size only controls the L1. The L2
-	 * is now directly feed with small/huge/giant pages from the walker.
-	 */
-	unsigned max_frag;
+	struct amdgpu_vm *vm = params->vm;
 
-	if (params->adev->asic_type < CHIP_VEGA10)
-		max_frag = params->adev->vm_manager.fragment_size;
-	else
-		max_frag = 31;
-
-	/* system pages are non continuously */
-	if (params->pages_addr) {
-		*frag = 0;
-		*frag_end = end;
+	tlb_cb->vm = vm;
+	if (!fence || !*fence) {
+		amdgpu_vm_tlb_seq_cb(NULL, &tlb_cb->cb);
 		return;
 	}
 
-	/* This intentionally wraps around if no bit is set */
-	*frag = min((unsigned)ffs(start) - 1, (unsigned)fls64(end - start) - 1);
-	if (*frag >= max_frag) {
-		*frag = max_frag;
-		*frag_end = end & ~((1ULL << max_frag) - 1);
+	if (!dma_fence_add_callback(*fence, &tlb_cb->cb,
+				    amdgpu_vm_tlb_seq_cb)) {
+		dma_fence_put(vm->last_tlb_flush);
+		vm->last_tlb_flush = dma_fence_get(*fence);
 	} else {
-		*frag_end = start + (1 << *frag);
+		amdgpu_vm_tlb_seq_cb(NULL, &tlb_cb->cb);
 	}
-}
-
-/**
- * amdgpu_vm_update_ptes - make sure that page tables are valid
- *
- * @params: see amdgpu_vm_update_params definition
- * @start: start of GPU address range
- * @end: end of GPU address range
- * @dst: destination address to map to, the next dst inside the function
- * @flags: mapping flags
- *
- * Update the page tables in the range @start - @end.
- *
- * Returns:
- * 0 for success, -EINVAL for failure.
- */
-static int amdgpu_vm_update_ptes(struct amdgpu_vm_update_params *params,
-				 uint64_t start, uint64_t end,
-				 uint64_t dst, uint64_t flags)
-{
-	struct amdgpu_device *adev = params->adev;
-	struct amdgpu_vm_pt_cursor cursor;
-	uint64_t frag_start = start, frag_end;
-	unsigned int frag;
-	int r;
-
-	/* figure out the initial fragment */
-	amdgpu_vm_fragment(params, frag_start, end, flags, &frag, &frag_end);
-
-	/* walk over the address space and update the PTs */
-	amdgpu_vm_pt_start(adev, params->vm, start, &cursor);
-	while (cursor.pfn < end) {
-		unsigned shift, parent_shift, mask;
-		uint64_t incr, entry_end, pe_start;
-		struct amdgpu_bo *pt;
-
-		if (!params->unlocked) {
-			/* make sure that the page tables covering the
-			 * address range are actually allocated
-			 */
-			r = amdgpu_vm_alloc_pts(params->adev, params->vm,
-						&cursor, params->immediate);
-			if (r)
-				return r;
-		}
-
-		shift = amdgpu_vm_level_shift(adev, cursor.level);
-		parent_shift = amdgpu_vm_level_shift(adev, cursor.level - 1);
-		if (params->unlocked) {
-			/* Unlocked updates are only allowed on the leaves */
-			if (amdgpu_vm_pt_descendant(adev, &cursor))
-				continue;
-		} else if (adev->asic_type < CHIP_VEGA10 &&
-			   (flags & AMDGPU_PTE_VALID)) {
-			/* No huge page support before GMC v9 */
-			if (cursor.level != AMDGPU_VM_PTB) {
-				if (!amdgpu_vm_pt_descendant(adev, &cursor))
-					return -ENOENT;
-				continue;
-			}
-		} else if (frag < shift) {
-			/* We can't use this level when the fragment size is
-			 * smaller than the address shift. Go to the next
-			 * child entry and try again.
-			 */
-			if (amdgpu_vm_pt_descendant(adev, &cursor))
-				continue;
-		} else if (frag >= parent_shift) {
-			/* If the fragment size is even larger than the parent
-			 * shift we should go up one level and check it again.
-			 */
-			if (!amdgpu_vm_pt_ancestor(&cursor))
-				return -EINVAL;
-			continue;
-		}
-
-		pt = cursor.entry->bo;
-		if (!pt) {
-			/* We need all PDs and PTs for mapping something, */
-			if (flags & AMDGPU_PTE_VALID)
-				return -ENOENT;
-
-			/* but unmapping something can happen at a higher
-			 * level.
-			 */
-			if (!amdgpu_vm_pt_ancestor(&cursor))
-				return -EINVAL;
-
-			pt = cursor.entry->bo;
-			shift = parent_shift;
-			frag_end = max(frag_end, ALIGN(frag_start + 1,
-				   1ULL << shift));
-		}
 
-		/* Looks good so far, calculate parameters for the update */
-		incr = (uint64_t)AMDGPU_GPU_PAGE_SIZE << shift;
-		mask = amdgpu_vm_entries_mask(adev, cursor.level);
-		pe_start = ((cursor.pfn >> shift) & mask) * 8;
-		entry_end = ((uint64_t)mask + 1) << shift;
-		entry_end += cursor.pfn & ~(entry_end - 1);
-		entry_end = min(entry_end, end);
-
-		do {
-			struct amdgpu_vm *vm = params->vm;
-			uint64_t upd_end = min(entry_end, frag_end);
-			unsigned nptes = (upd_end - frag_start) >> shift;
-			uint64_t upd_flags = flags | AMDGPU_PTE_FRAG(frag);
-
-			/* This can happen when we set higher level PDs to
-			 * silent to stop fault floods.
-			 */
-			nptes = max(nptes, 1u);
-
-			trace_amdgpu_vm_update_ptes(params, frag_start, upd_end,
-						    nptes, dst, incr, upd_flags,
-						    vm->task_info.pid,
-						    vm->immediate.fence_context);
-			amdgpu_vm_update_flags(params, to_amdgpu_bo_vm(pt),
-					       cursor.level, pe_start, dst,
-					       nptes, incr, upd_flags);
-
-			pe_start += nptes * 8;
-			dst += nptes * incr;
-
-			frag_start = upd_end;
-			if (frag_start >= frag_end) {
-				/* figure out the next fragment */
-				amdgpu_vm_fragment(params, frag_start, end,
-						   flags, &frag, &frag_end);
-				if (frag < shift)
-					break;
-			}
-		} while (frag_start < entry_end);
-
-		if (amdgpu_vm_pt_descendant(adev, &cursor)) {
-			/* Free all child entries.
-			 * Update the tables with the flags and addresses and free up subsequent
-			 * tables in the case of huge pages or freed up areas.
-			 * This is the maximum you can free, because all other page tables are not
-			 * completely covered by the range and so potentially still in use.
-			 */
-			while (cursor.pfn < frag_start) {
-				/* Make sure previous mapping is freed */
-				if (cursor.entry->bo) {
-					params->table_freed = true;
-					amdgpu_vm_free_pts(adev, params->vm, &cursor);
-				}
-				amdgpu_vm_pt_next(adev, &cursor);
-			}
+	/* Prepare a TLB flush fence to be attached to PTs */
+	if (!params->unlocked) {
+		amdgpu_vm_tlb_fence_create(params->adev, vm, fence);
 
-		} else if (frag >= shift) {
-			/* or just move on to the next on the same level. */
-			amdgpu_vm_pt_next(adev, &cursor);
-		}
+		/* Makes sure no PD/PT is freed before the flush */
+		dma_resv_add_fence(vm->root.bo->tbo.base.resv, *fence,
+				   DMA_RESV_USAGE_BOOKKEEP);
 	}
-
-	return 0;
 }
 
 /**
- * amdgpu_vm_bo_update_mapping - update a mapping in the vm page table
+ * amdgpu_vm_update_range - update a range in the vm page table
  *
- * @adev: amdgpu_device pointer of the VM
- * @bo_adev: amdgpu_device pointer of the mapped BO
- * @vm: requested vm
+ * @adev: amdgpu_device pointer to use for commands
+ * @vm: the VM to update the range
  * @immediate: immediate submission in a page fault
  * @unlocked: unlocked invalidation during MM callback
- * @resv: fences we need to sync to
+ * @flush_tlb: trigger tlb invalidation after update completed
+ * @allow_override: change MTYPE for local NUMA nodes
+ * @sync: fences we need to sync to
  * @start: start of mapped range
  * @last: last mapped entry
  * @flags: flags for the entries
  * @offset: offset into nodes and pages_addr
+ * @vram_base: base for vram mappings
  * @res: ttm_resource to map
  * @pages_addr: DMA addresses to use for mapping
  * @fence: optional resulting fence
- * @table_freed: return true if page table is freed
  *
  * Fill in the page table entries between @start and @last.
  *
  * Returns:
- * 0 for success, -EINVAL for failure.
+ * 0 for success, negative erro code for failure.
  */
-int amdgpu_vm_bo_update_mapping(struct amdgpu_device *adev,
-				struct amdgpu_device *bo_adev,
-				struct amdgpu_vm *vm, bool immediate,
-				bool unlocked, struct dma_resv *resv,
-				uint64_t start, uint64_t last,
-				uint64_t flags, uint64_t offset,
-				struct ttm_resource *res,
-				dma_addr_t *pages_addr,
-				struct dma_fence **fence,
-				bool *table_freed)
+int amdgpu_vm_update_range(struct amdgpu_device *adev, struct amdgpu_vm *vm,
+			   bool immediate, bool unlocked, bool flush_tlb,
+			   bool allow_override, struct amdgpu_sync *sync,
+			   uint64_t start, uint64_t last, uint64_t flags,
+			   uint64_t offset, uint64_t vram_base,
+			   struct ttm_resource *res, dma_addr_t *pages_addr,
+			   struct dma_fence **fence)
 {
+	struct amdgpu_vm_tlb_seq_struct *tlb_cb;
 	struct amdgpu_vm_update_params params;
 	struct amdgpu_res_cursor cursor;
-	enum amdgpu_sync_mode sync_mode;
 	int r, idx;
 
 	if (!drm_dev_enter(adev_to_drm(adev), &idx))
 		return -ENODEV;
 
+	tlb_cb = kmalloc(sizeof(*tlb_cb), GFP_KERNEL);
+	if (!tlb_cb) {
+		drm_dev_exit(idx);
+		return -ENOMEM;
+	}
+
+	/* Vega20+XGMI where PTEs get inadvertently cached in L2 texture cache,
+	 * heavy-weight flush TLB unconditionally.
+	 */
+	flush_tlb |= adev->gmc.xgmi.num_physical_nodes &&
+		     amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 0);
+
+	/*
+	 * On GFX8 and older any 8 PTE block with a valid bit set enters the TLB
+	 */
+	flush_tlb |= amdgpu_ip_version(adev, GC_HWIP, 0) < IP_VERSION(9, 0, 0);
+
 	memset(&params, 0, sizeof(params));
 	params.adev = adev;
 	params.vm = vm;
 	params.immediate = immediate;
 	params.pages_addr = pages_addr;
 	params.unlocked = unlocked;
-
-	/* Implicitly sync to command submissions in the same VM before
-	 * unmapping. Sync to moving fences before mapping.
-	 */
-	if (!(flags & AMDGPU_PTE_VALID))
-		sync_mode = AMDGPU_SYNC_EQ_OWNER;
-	else
-		sync_mode = AMDGPU_SYNC_EXPLICIT;
+	params.needs_flush = flush_tlb;
+	params.allow_override = allow_override;
+	INIT_LIST_HEAD(&params.tlb_flush_waitlist);
 
 	amdgpu_vm_eviction_lock(vm);
 	if (vm->evicting) {
 		r = -EBUSY;
-		goto error_unlock;
+		goto error_free;
 	}
 
 	if (!unlocked && !dma_fence_is_signaled(vm->last_unlocked)) {
@@ -1750,9 +1159,10 @@ int amdgpu_vm_bo_update_mapping(struct amdgpu_device *adev,
 		dma_fence_put(tmp);
 	}
 
-	r = vm->update_funcs->prepare(&params, resv, sync_mode);
+	r = vm->update_funcs->prepare(&params, sync,
+				      AMDGPU_KERNEL_JOB_ID_VM_UPDATE_RANGE);
 	if (r)
-		goto error_unlock;
+		goto error_free;
 
 	amdgpu_res_first(pages_addr ? NULL : res, offset,
 			 (last - start + 1) * AMDGPU_GPU_PAGE_SIZE, &cursor);
@@ -1779,6 +1189,8 @@ int amdgpu_vm_bo_update_mapping(struct amdgpu_device *adev,
 					    pages_addr[idx - 1] + PAGE_SIZE))
 						break;
 				}
+				if (!contiguous)
+					count--;
 				num_entries = count *
 					AMDGPU_GPU_PAGES_IN_CPU_PAGE;
 			}
@@ -1791,84 +1203,53 @@ int amdgpu_vm_bo_update_mapping(struct amdgpu_device *adev,
 				params.pages_addr = NULL;
 			}
 
-		} else if (flags & (AMDGPU_PTE_VALID | AMDGPU_PTE_PRT)) {
-			addr = bo_adev->vm_manager.vram_base_offset +
-				cursor.start;
+		} else if (flags & (AMDGPU_PTE_VALID | AMDGPU_PTE_PRT_FLAG(adev))) {
+			addr = vram_base + cursor.start;
 		} else {
 			addr = 0;
 		}
 
 		tmp = start + num_entries;
-		r = amdgpu_vm_update_ptes(&params, start, tmp, addr, flags);
+		r = amdgpu_vm_ptes_update(&params, start, tmp, addr, flags);
 		if (r)
-			goto error_unlock;
+			goto error_free;
 
 		amdgpu_res_next(&cursor, num_entries * AMDGPU_GPU_PAGE_SIZE);
 		start = tmp;
 	}
 
 	r = vm->update_funcs->commit(&params, fence);
+	if (r)
+		goto error_free;
 
-	if (table_freed)
-		*table_freed = *table_freed || params.table_freed;
+	if (params.needs_flush) {
+		amdgpu_vm_tlb_flush(&params, fence, tlb_cb);
+		tlb_cb = NULL;
+	}
 
-error_unlock:
+	amdgpu_vm_pt_free_list(adev, &params);
+
+error_free:
+	kfree(tlb_cb);
 	amdgpu_vm_eviction_unlock(vm);
 	drm_dev_exit(idx);
 	return r;
 }
 
-void amdgpu_vm_get_memory(struct amdgpu_vm *vm, uint64_t *vram_mem,
-				uint64_t *gtt_mem, uint64_t *cpu_mem)
+void amdgpu_vm_get_memory(struct amdgpu_vm *vm,
+			  struct amdgpu_mem_stats stats[__AMDGPU_PL_NUM])
 {
-	struct amdgpu_bo_va *bo_va, *tmp;
-
-	list_for_each_entry_safe(bo_va, tmp, &vm->idle, base.vm_status) {
-		if (!bo_va->base.bo)
-			continue;
-		amdgpu_bo_get_memory(bo_va->base.bo, vram_mem,
-				gtt_mem, cpu_mem);
-	}
-	list_for_each_entry_safe(bo_va, tmp, &vm->evicted, base.vm_status) {
-		if (!bo_va->base.bo)
-			continue;
-		amdgpu_bo_get_memory(bo_va->base.bo, vram_mem,
-				gtt_mem, cpu_mem);
-	}
-	list_for_each_entry_safe(bo_va, tmp, &vm->relocated, base.vm_status) {
-		if (!bo_va->base.bo)
-			continue;
-		amdgpu_bo_get_memory(bo_va->base.bo, vram_mem,
-				gtt_mem, cpu_mem);
-	}
-	list_for_each_entry_safe(bo_va, tmp, &vm->moved, base.vm_status) {
-		if (!bo_va->base.bo)
-			continue;
-		amdgpu_bo_get_memory(bo_va->base.bo, vram_mem,
-				gtt_mem, cpu_mem);
-	}
-	spin_lock(&vm->invalidated_lock);
-	list_for_each_entry_safe(bo_va, tmp, &vm->invalidated, base.vm_status) {
-		if (!bo_va->base.bo)
-			continue;
-		amdgpu_bo_get_memory(bo_va->base.bo, vram_mem,
-				gtt_mem, cpu_mem);
-	}
-	list_for_each_entry_safe(bo_va, tmp, &vm->done, base.vm_status) {
-		if (!bo_va->base.bo)
-			continue;
-		amdgpu_bo_get_memory(bo_va->base.bo, vram_mem,
-				gtt_mem, cpu_mem);
-	}
-	spin_unlock(&vm->invalidated_lock);
+	spin_lock(&vm->status_lock);
+	memcpy(stats, vm->stats, sizeof(*stats) * __AMDGPU_PL_NUM);
+	spin_unlock(&vm->status_lock);
 }
+
 /**
  * amdgpu_vm_bo_update - update all BO mappings in the vm page table
  *
  * @adev: amdgpu_device pointer
  * @bo_va: requested BO and VM object
  * @clear: if true clear the entries
- * @table_freed: return true if page table is freed
  *
  * Fill in the page table entries for @bo_va.
  *
@@ -1876,59 +1257,90 @@ void amdgpu_vm_get_memory(struct amdgpu_vm *vm, uint64_t *vram_mem,
  * 0 for success, -EINVAL for failure.
  */
 int amdgpu_vm_bo_update(struct amdgpu_device *adev, struct amdgpu_bo_va *bo_va,
-			bool clear, bool *table_freed)
+			bool clear)
 {
 	struct amdgpu_bo *bo = bo_va->base.bo;
 	struct amdgpu_vm *vm = bo_va->base.vm;
 	struct amdgpu_bo_va_mapping *mapping;
+	struct dma_fence **last_update;
 	dma_addr_t *pages_addr = NULL;
 	struct ttm_resource *mem;
-	struct dma_fence **last_update;
-	struct dma_resv *resv;
+	struct amdgpu_sync sync;
+	bool flush_tlb = clear;
+	uint64_t vram_base;
 	uint64_t flags;
-	struct amdgpu_device *bo_adev = adev;
+	bool uncached;
 	int r;
 
-	if (clear || !bo) {
+	amdgpu_sync_create(&sync);
+	if (clear) {
+		mem = NULL;
+
+		/* Implicitly sync to command submissions in the same VM before
+		 * unmapping.
+		 */
+		r = amdgpu_sync_resv(adev, &sync, vm->root.bo->tbo.base.resv,
+				     AMDGPU_SYNC_EQ_OWNER, vm);
+		if (r)
+			goto error_free;
+		if (bo) {
+			r = amdgpu_sync_kfd(&sync, bo->tbo.base.resv);
+			if (r)
+				goto error_free;
+		}
+	} else if (!bo) {
 		mem = NULL;
-		resv = vm->root.bo->tbo.base.resv;
+
+		/* PRT map operations don't need to sync to anything. */
+
 	} else {
 		struct drm_gem_object *obj = &bo->tbo.base;
 
-		resv = bo->tbo.base.resv;
-		if (obj->import_attach && bo_va->is_xgmi) {
+		if (drm_gem_is_imported(obj) && bo_va->is_xgmi) {
 			struct dma_buf *dma_buf = obj->import_attach->dmabuf;
 			struct drm_gem_object *gobj = dma_buf->priv;
 			struct amdgpu_bo *abo = gem_to_amdgpu_bo(gobj);
 
-			if (abo->tbo.resource->mem_type == TTM_PL_VRAM)
+			if (abo->tbo.resource &&
+			    abo->tbo.resource->mem_type == TTM_PL_VRAM)
 				bo = gem_to_amdgpu_bo(gobj);
 		}
 		mem = bo->tbo.resource;
-		if (mem->mem_type == TTM_PL_TT ||
-		    mem->mem_type == AMDGPU_PL_PREEMPT)
+		if (mem && (mem->mem_type == TTM_PL_TT ||
+			    mem->mem_type == AMDGPU_PL_PREEMPT))
 			pages_addr = bo->tbo.ttm->dma_address;
+
+		/* Implicitly sync to moving fences before mapping anything */
+		r = amdgpu_sync_resv(adev, &sync, bo->tbo.base.resv,
+				     AMDGPU_SYNC_EXPLICIT, vm);
+		if (r)
+			goto error_free;
 	}
 
 	if (bo) {
+		struct amdgpu_device *bo_adev;
+
 		flags = amdgpu_ttm_tt_pte_flags(adev, bo->tbo.ttm, mem);
 
 		if (amdgpu_bo_encrypted(bo))
 			flags |= AMDGPU_PTE_TMZ;
 
 		bo_adev = amdgpu_ttm_adev(bo->tbo.bdev);
+		vram_base = bo_adev->vm_manager.vram_base_offset;
+		uncached = (bo->flags & AMDGPU_GEM_CREATE_UNCACHED) != 0;
 	} else {
 		flags = 0x0;
+		vram_base = 0;
+		uncached = false;
 	}
 
-	if (clear || (bo && bo->tbo.base.resv ==
-		      vm->root.bo->tbo.base.resv))
+	if (clear || amdgpu_vm_is_bo_always_valid(vm, bo))
 		last_update = &vm->last_update;
 	else
 		last_update = &bo_va->last_pt_update;
 
 	if (!clear && bo_va->base.moved) {
-		bo_va->base.moved = false;
+		flush_tlb = true;
 		list_splice_init(&bo_va->valids, &bo_va->invalids);
 
 	} else if (bo_va->cleared != clear) {
@@ -1941,34 +1353,34 @@ int amdgpu_vm_bo_update(struct amdgpu_device *adev, struct amdgpu_bo_va *bo_va,
 		/* normally,bo_va->flags only contians READABLE and WIRTEABLE bit go here
 		 * but in case of something, we filter the flags in first place
 		 */
-		if (!(mapping->flags & AMDGPU_PTE_READABLE))
+		if (!(mapping->flags & AMDGPU_VM_PAGE_READABLE))
 			update_flags &= ~AMDGPU_PTE_READABLE;
-		if (!(mapping->flags & AMDGPU_PTE_WRITEABLE))
+		if (!(mapping->flags & AMDGPU_VM_PAGE_WRITEABLE))
 			update_flags &= ~AMDGPU_PTE_WRITEABLE;
 
 		/* Apply ASIC specific mapping flags */
-		amdgpu_gmc_get_vm_pte(adev, mapping, &update_flags);
+		amdgpu_gmc_get_vm_pte(adev, vm, bo, mapping->flags,
+				      &update_flags);
 
 		trace_amdgpu_vm_bo_update(mapping);
 
-		r = amdgpu_vm_bo_update_mapping(adev, bo_adev, vm, false, false,
-						resv, mapping->start,
-						mapping->last, update_flags,
-						mapping->offset, mem,
-						pages_addr, last_update, table_freed);
+		r = amdgpu_vm_update_range(adev, vm, false, false, flush_tlb,
+					   !uncached, &sync, mapping->start,
+					   mapping->last, update_flags,
+					   mapping->offset, vram_base, mem,
+					   pages_addr, last_update);
 		if (r)
-			return r;
+			goto error_free;
 	}
 
 	/* If the BO is not in its preferred location add it back to
 	 * the evicted list so that it gets validated again on the
 	 * next command submission.
 	 */
-	if (bo && bo->tbo.base.resv == vm->root.bo->tbo.base.resv) {
-		uint32_t mem_type = bo->tbo.resource->mem_type;
-
-		if (!(bo->preferred_domains &
-		      amdgpu_mem_type_to_domain(mem_type)))
+	if (amdgpu_vm_is_bo_always_valid(vm, bo)) {
+		if (bo->tbo.resource &&
+		    !(bo->preferred_domains &
+		      amdgpu_mem_type_to_domain(bo->tbo.resource->mem_type)))
 			amdgpu_vm_bo_evicted(&bo_va->base);
 		else
 			amdgpu_vm_bo_idle(&bo_va->base);
@@ -1978,13 +1390,16 @@ int amdgpu_vm_bo_update(struct amdgpu_device *adev, struct amdgpu_bo_va *bo_va,
 
 	list_splice_init(&bo_va->invalids, &bo_va->valids);
 	bo_va->cleared = clear;
+	bo_va->base.moved = false;
 
 	if (trace_amdgpu_vm_bo_mapping_enabled()) {
 		list_for_each_entry(mapping, &bo_va->valids, list)
 			trace_amdgpu_vm_bo_mapping(mapping);
 	}
 
-	return 0;
+error_free:
+	amdgpu_sync_free(&sync);
+	return r;
 }
 
 /**
@@ -2086,7 +1501,7 @@ static void amdgpu_vm_free_mapping(struct amdgpu_device *adev,
 				   struct amdgpu_bo_va_mapping *mapping,
 				   struct dma_fence *fence)
 {
-	if (mapping->flags & AMDGPU_PTE_PRT)
+	if (mapping->flags & AMDGPU_VM_PAGE_PRT)
 		amdgpu_vm_add_prt_cb(adev, fence);
 	kfree(mapping);
 }
@@ -2102,30 +1517,14 @@ static void amdgpu_vm_free_mapping(struct amdgpu_device *adev,
 static void amdgpu_vm_prt_fini(struct amdgpu_device *adev, struct amdgpu_vm *vm)
 {
 	struct dma_resv *resv = vm->root.bo->tbo.base.resv;
-	struct dma_fence *excl, **shared;
-	unsigned i, shared_count;
-	int r;
-
-	r = dma_resv_get_fences(resv, &excl, &shared_count, &shared);
-	if (r) {
-		/* Not enough memory to grab the fence list, as last resort
-		 * block for all the fences to complete.
-		 */
-		dma_resv_wait_timeout(resv, true, false,
-						    MAX_SCHEDULE_TIMEOUT);
-		return;
-	}
+	struct dma_resv_iter cursor;
+	struct dma_fence *fence;
 
-	/* Add a callback for each fence in the reservation object */
-	amdgpu_vm_prt_get(adev);
-	amdgpu_vm_add_prt_cb(adev, excl);
-
-	for (i = 0; i < shared_count; ++i) {
+	dma_resv_for_each_fence(&cursor, resv, DMA_RESV_USAGE_BOOKKEEP, fence) {
+		/* Add a callback for each fence in the reservation object */
 		amdgpu_vm_prt_get(adev);
-		amdgpu_vm_add_prt_cb(adev, shared[i]);
+		amdgpu_vm_add_prt_cb(adev, fence);
 	}
-
-	kfree(shared);
 }
 
 /**
@@ -2147,29 +1546,34 @@ int amdgpu_vm_clear_freed(struct amdgpu_device *adev,
 			  struct amdgpu_vm *vm,
 			  struct dma_fence **fence)
 {
-	struct dma_resv *resv = vm->root.bo->tbo.base.resv;
 	struct amdgpu_bo_va_mapping *mapping;
-	uint64_t init_pte_value = 0;
 	struct dma_fence *f = NULL;
+	struct amdgpu_sync sync;
 	int r;
 
+
+	/*
+	 * Implicitly sync to command submissions in the same VM before
+	 * unmapping.
+	 */
+	amdgpu_sync_create(&sync);
+	r = amdgpu_sync_resv(adev, &sync, vm->root.bo->tbo.base.resv,
+			     AMDGPU_SYNC_EQ_OWNER, vm);
+	if (r)
+		goto error_free;
+
 	while (!list_empty(&vm->freed)) {
 		mapping = list_first_entry(&vm->freed,
 			struct amdgpu_bo_va_mapping, list);
 		list_del(&mapping->list);
 
-		if (vm->pte_support_ats &&
-		    mapping->start < AMDGPU_GMC_HOLE_START)
-			init_pte_value = AMDGPU_PTE_DEFAULT_ATC;
-
-		r = amdgpu_vm_bo_update_mapping(adev, adev, vm, false, false,
-						resv, mapping->start,
-						mapping->last, init_pte_value,
-						0, NULL, NULL, &f, NULL);
+		r = amdgpu_vm_update_range(adev, vm, false, false, true, false,
+					   &sync, mapping->start, mapping->last,
+					   0, 0, 0, NULL, NULL, &f);
 		amdgpu_vm_free_mapping(adev, vm, mapping, f);
 		if (r) {
 			dma_fence_put(f);
-			return r;
+			goto error_free;
 		}
 	}
 
@@ -2180,7 +1584,9 @@ int amdgpu_vm_clear_freed(struct amdgpu_device *adev,
 		dma_fence_put(f);
 	}
 
-	return 0;
+error_free:
+	amdgpu_sync_free(&sync);
+	return r;
 
 }
 
@@ -2189,6 +1595,7 @@ int amdgpu_vm_clear_freed(struct amdgpu_device *adev,
  *
  * @adev: amdgpu_device pointer
  * @vm: requested vm
+ * @ticket: optional reservation ticket used to reserve the VM
  *
  * Make sure all BOs which are moved are updated in the PTs.
  *
@@ -2198,48 +1605,116 @@ int amdgpu_vm_clear_freed(struct amdgpu_device *adev,
  * PTs have to be reserved!
  */
 int amdgpu_vm_handle_moved(struct amdgpu_device *adev,
-			   struct amdgpu_vm *vm)
+			   struct amdgpu_vm *vm,
+			   struct ww_acquire_ctx *ticket)
 {
-	struct amdgpu_bo_va *bo_va, *tmp;
+	struct amdgpu_bo_va *bo_va;
 	struct dma_resv *resv;
-	bool clear;
+	bool clear, unlock;
 	int r;
 
-	list_for_each_entry_safe(bo_va, tmp, &vm->moved, base.vm_status) {
+	spin_lock(&vm->status_lock);
+	while (!list_empty(&vm->moved)) {
+		bo_va = list_first_entry(&vm->moved, struct amdgpu_bo_va,
+					 base.vm_status);
+		spin_unlock(&vm->status_lock);
+
 		/* Per VM BOs never need to bo cleared in the page tables */
-		r = amdgpu_vm_bo_update(adev, bo_va, false, NULL);
+		r = amdgpu_vm_bo_update(adev, bo_va, false);
 		if (r)
 			return r;
+		spin_lock(&vm->status_lock);
 	}
 
-	spin_lock(&vm->invalidated_lock);
 	while (!list_empty(&vm->invalidated)) {
 		bo_va = list_first_entry(&vm->invalidated, struct amdgpu_bo_va,
 					 base.vm_status);
 		resv = bo_va->base.bo->tbo.base.resv;
-		spin_unlock(&vm->invalidated_lock);
+		spin_unlock(&vm->status_lock);
 
 		/* Try to reserve the BO to avoid clearing its ptes */
-		if (!amdgpu_vm_debug && dma_resv_trylock(resv))
+		if (!adev->debug_vm && dma_resv_trylock(resv)) {
+			clear = false;
+			unlock = true;
+		/* The caller is already holding the reservation lock */
+		} else if (ticket && dma_resv_locking_ctx(resv) == ticket) {
 			clear = false;
+			unlock = false;
 		/* Somebody else is using the BO right now */
-		else
+		} else {
 			clear = true;
+			unlock = false;
+		}
+
+		r = amdgpu_vm_bo_update(adev, bo_va, clear);
 
-		r = amdgpu_vm_bo_update(adev, bo_va, clear, NULL);
+		if (unlock)
+			dma_resv_unlock(resv);
 		if (r)
 			return r;
 
-		if (!clear)
-			dma_resv_unlock(resv);
-		spin_lock(&vm->invalidated_lock);
+		/* Remember evicted DMABuf imports in compute VMs for later
+		 * validation
+		 */
+		if (vm->is_compute_context &&
+		    drm_gem_is_imported(&bo_va->base.bo->tbo.base) &&
+		    (!bo_va->base.bo->tbo.resource ||
+		     bo_va->base.bo->tbo.resource->mem_type == TTM_PL_SYSTEM))
+			amdgpu_vm_bo_evicted_user(&bo_va->base);
+
+		spin_lock(&vm->status_lock);
 	}
-	spin_unlock(&vm->invalidated_lock);
+	spin_unlock(&vm->status_lock);
 
 	return 0;
 }
 
 /**
+ * amdgpu_vm_flush_compute_tlb - Flush TLB on compute VM
+ *
+ * @adev: amdgpu_device pointer
+ * @vm: requested vm
+ * @flush_type: flush type
+ * @xcc_mask: mask of XCCs that belong to the compute partition in need of a TLB flush.
+ *
+ * Flush TLB if needed for a compute VM.
+ *
+ * Returns:
+ * 0 for success.
+ */
+int amdgpu_vm_flush_compute_tlb(struct amdgpu_device *adev,
+				struct amdgpu_vm *vm,
+				uint32_t flush_type,
+				uint32_t xcc_mask)
+{
+	uint64_t tlb_seq = amdgpu_vm_tlb_seq(vm);
+	bool all_hub = false;
+	int xcc = 0, r = 0;
+
+	WARN_ON_ONCE(!vm->is_compute_context);
+
+	/*
+	 * It can be that we race and lose here, but that is extremely unlikely
+	 * and the worst thing which could happen is that we flush the changes
+	 * into the TLB once more which is harmless.
+	 */
+	if (atomic64_xchg(&vm->kfd_last_flushed_seq, tlb_seq) == tlb_seq)
+		return 0;
+
+	if (adev->family == AMDGPU_FAMILY_AI ||
+	    adev->family == AMDGPU_FAMILY_RV)
+		all_hub = true;
+
+	for_each_inst(xcc, xcc_mask) {
+		r = amdgpu_gmc_flush_gpu_tlb_pasid(adev, vm->pasid, flush_type,
+						   all_hub, xcc);
+		if (r)
+			break;
+	}
+	return r;
+}
+
+/**
  * amdgpu_vm_bo_add - add a bo to a specific vm
  *
  * @adev: amdgpu_device pointer
@@ -2267,12 +1742,14 @@ struct amdgpu_bo_va *amdgpu_vm_bo_add(struct amdgpu_device *adev,
 	amdgpu_vm_bo_base_init(&bo_va->base, vm, bo);
 
 	bo_va->ref_count = 1;
+	bo_va->last_pt_update = dma_fence_get_stub();
 	INIT_LIST_HEAD(&bo_va->valids);
 	INIT_LIST_HEAD(&bo_va->invalids);
 
 	if (!bo)
 		return bo_va;
 
+	dma_resv_assert_held(bo->tbo.base.resv);
 	if (amdgpu_dmabuf_is_xgmi_accessible(adev, bo)) {
 		bo_va->is_xgmi = true;
 		/* Power up XGMI if it can be potentially used */
@@ -2303,16 +1780,46 @@ static void amdgpu_vm_bo_insert_map(struct amdgpu_device *adev,
 	list_add(&mapping->list, &bo_va->invalids);
 	amdgpu_vm_it_insert(mapping, &vm->va);
 
-	if (mapping->flags & AMDGPU_PTE_PRT)
+	if (mapping->flags & AMDGPU_VM_PAGE_PRT)
 		amdgpu_vm_prt_get(adev);
 
-	if (bo && bo->tbo.base.resv == vm->root.bo->tbo.base.resv &&
-	    !bo_va->base.moved) {
-		list_move(&bo_va->base.vm_status, &vm->moved);
-	}
+	if (amdgpu_vm_is_bo_always_valid(vm, bo) && !bo_va->base.moved)
+		amdgpu_vm_bo_moved(&bo_va->base);
+
 	trace_amdgpu_vm_bo_map(bo_va, mapping);
 }
 
+/* Validate operation parameters to prevent potential abuse */
+static int amdgpu_vm_verify_parameters(struct amdgpu_device *adev,
+					  struct amdgpu_bo *bo,
+					  uint64_t saddr,
+					  uint64_t offset,
+					  uint64_t size)
+{
+	uint64_t tmp, lpfn;
+
+	if (saddr & AMDGPU_GPU_PAGE_MASK
+	    || offset & AMDGPU_GPU_PAGE_MASK
+	    || size & AMDGPU_GPU_PAGE_MASK)
+		return -EINVAL;
+
+	if (check_add_overflow(saddr, size, &tmp)
+	    || check_add_overflow(offset, size, &tmp)
+	    || size == 0 /* which also leads to end < begin */)
+		return -EINVAL;
+
+	/* make sure object fit at this offset */
+	if (bo && offset + size > amdgpu_bo_size(bo))
+		return -EINVAL;
+
+	/* Ensure last pfn not exceed max_pfn */
+	lpfn = (saddr + size - 1) >> AMDGPU_GPU_PAGE_SHIFT;
+	if (lpfn >= adev->vm_manager.max_pfn)
+		return -EINVAL;
+
+	return 0;
+}
+
 /**
  * amdgpu_vm_bo_map - map bo inside a vm
  *
@@ -2333,27 +1840,20 @@ static void amdgpu_vm_bo_insert_map(struct amdgpu_device *adev,
 int amdgpu_vm_bo_map(struct amdgpu_device *adev,
 		     struct amdgpu_bo_va *bo_va,
 		     uint64_t saddr, uint64_t offset,
-		     uint64_t size, uint64_t flags)
+		     uint64_t size, uint32_t flags)
 {
 	struct amdgpu_bo_va_mapping *mapping, *tmp;
 	struct amdgpu_bo *bo = bo_va->base.bo;
 	struct amdgpu_vm *vm = bo_va->base.vm;
 	uint64_t eaddr;
+	int r;
 
-	/* validate the parameters */
-	if (saddr & ~PAGE_MASK || offset & ~PAGE_MASK ||
-	    size == 0 || size & ~PAGE_MASK)
-		return -EINVAL;
-
-	/* make sure object fit at this offset */
-	eaddr = saddr + size - 1;
-	if (saddr >= eaddr ||
-	    (bo && offset + size > amdgpu_bo_size(bo)) ||
-	    (eaddr >= adev->vm_manager.max_pfn << AMDGPU_GPU_PAGE_SHIFT))
-		return -EINVAL;
+	r = amdgpu_vm_verify_parameters(adev, bo, saddr, offset, size);
+	if (r)
+		return r;
 
 	saddr /= AMDGPU_GPU_PAGE_SIZE;
-	eaddr /= AMDGPU_GPU_PAGE_SIZE;
+	eaddr = saddr + (size - 1) / AMDGPU_GPU_PAGE_SIZE;
 
 	tmp = amdgpu_vm_it_iter_first(&vm->va, saddr, eaddr);
 	if (tmp) {
@@ -2399,24 +1899,16 @@ int amdgpu_vm_bo_map(struct amdgpu_device *adev,
 int amdgpu_vm_bo_replace_map(struct amdgpu_device *adev,
 			     struct amdgpu_bo_va *bo_va,
 			     uint64_t saddr, uint64_t offset,
-			     uint64_t size, uint64_t flags)
+			     uint64_t size, uint32_t flags)
 {
 	struct amdgpu_bo_va_mapping *mapping;
 	struct amdgpu_bo *bo = bo_va->base.bo;
 	uint64_t eaddr;
 	int r;
 
-	/* validate the parameters */
-	if (saddr & ~PAGE_MASK || offset & ~PAGE_MASK ||
-	    size == 0 || size & ~PAGE_MASK)
-		return -EINVAL;
-
-	/* make sure object fit at this offset */
-	eaddr = saddr + size - 1;
-	if (saddr >= eaddr ||
-	    (bo && offset + size > amdgpu_bo_size(bo)) ||
-	    (eaddr >= adev->vm_manager.max_pfn << AMDGPU_GPU_PAGE_SHIFT))
-		return -EINVAL;
+	r = amdgpu_vm_verify_parameters(adev, bo, saddr, offset, size);
+	if (r)
+		return r;
 
 	/* Allocate all the needed memory */
 	mapping = kmalloc(sizeof(*mapping), GFP_KERNEL);
@@ -2430,7 +1922,7 @@ int amdgpu_vm_bo_replace_map(struct amdgpu_device *adev,
 	}
 
 	saddr /= AMDGPU_GPU_PAGE_SIZE;
-	eaddr /= AMDGPU_GPU_PAGE_SIZE;
+	eaddr = saddr + (size - 1) / AMDGPU_GPU_PAGE_SIZE;
 
 	mapping->start = saddr;
 	mapping->last = eaddr;
@@ -2463,6 +1955,7 @@ int amdgpu_vm_bo_unmap(struct amdgpu_device *adev,
 	struct amdgpu_bo_va_mapping *mapping;
 	struct amdgpu_vm *vm = bo_va->base.vm;
 	bool valid = true;
+	int r;
 
 	saddr /= AMDGPU_GPU_PAGE_SIZE;
 
@@ -2483,6 +1976,17 @@ int amdgpu_vm_bo_unmap(struct amdgpu_device *adev,
 			return -ENOENT;
 	}
 
+	/* It's unlikely to happen that the mapping userq hasn't been idled
+	 * during user requests GEM unmap IOCTL except for forcing the unmap
+	 * from user space.
+	 */
+	if (unlikely(atomic_read(&bo_va->userq_va_mapped) > 0)) {
+		r = amdgpu_userq_gem_va_unmap_validate(adev, mapping, saddr);
+		if (unlikely(r == -EBUSY))
+			dev_warn_once(adev->dev,
+				      "Attempt to unmap an active userq buffer\n");
+	}
+
 	list_del(&mapping->list);
 	amdgpu_vm_it_remove(mapping, &vm->va);
 	mapping->bo_va = NULL;
@@ -2517,10 +2021,14 @@ int amdgpu_vm_bo_clear_mappings(struct amdgpu_device *adev,
 	struct amdgpu_bo_va_mapping *before, *after, *tmp, *next;
 	LIST_HEAD(removed);
 	uint64_t eaddr;
+	int r;
+
+	r = amdgpu_vm_verify_parameters(adev, NULL, saddr, 0, size);
+	if (r)
+		return r;
 
-	eaddr = saddr + size - 1;
 	saddr /= AMDGPU_GPU_PAGE_SIZE;
-	eaddr /= AMDGPU_GPU_PAGE_SIZE;
+	eaddr = saddr + (size - 1) / AMDGPU_GPU_PAGE_SIZE;
 
 	/* Allocate all the needed memory */
 	before = kzalloc(sizeof(*before), GFP_KERNEL);
@@ -2582,18 +2090,30 @@ int amdgpu_vm_bo_clear_mappings(struct amdgpu_device *adev,
 
 	/* Insert partial mapping before the range */
 	if (!list_empty(&before->list)) {
+		struct amdgpu_bo *bo = before->bo_va->base.bo;
+
 		amdgpu_vm_it_insert(before, &vm->va);
-		if (before->flags & AMDGPU_PTE_PRT)
+		if (before->flags & AMDGPU_VM_PAGE_PRT)
 			amdgpu_vm_prt_get(adev);
+
+		if (amdgpu_vm_is_bo_always_valid(vm, bo) &&
+		    !before->bo_va->base.moved)
+			amdgpu_vm_bo_moved(&before->bo_va->base);
 	} else {
 		kfree(before);
 	}
 
 	/* Insert partial mapping after the range */
 	if (!list_empty(&after->list)) {
+		struct amdgpu_bo *bo = after->bo_va->base.bo;
+
 		amdgpu_vm_it_insert(after, &vm->va);
-		if (after->flags & AMDGPU_PTE_PRT)
+		if (after->flags & AMDGPU_VM_PAGE_PRT)
 			amdgpu_vm_prt_get(adev);
+
+		if (amdgpu_vm_is_bo_always_valid(vm, bo) &&
+		    !after->bo_va->base.moved)
+			amdgpu_vm_bo_moved(&after->bo_va->base);
 	} else {
 		kfree(after);
 	}
@@ -2650,7 +2170,7 @@ void amdgpu_vm_bo_trace_cs(struct amdgpu_vm *vm, struct ww_acquire_ctx *ticket)
 }
 
 /**
- * amdgpu_vm_bo_rmv - remove a bo to a specific vm
+ * amdgpu_vm_bo_del - remove a bo from a specific vm
  *
  * @adev: amdgpu_device pointer
  * @bo_va: requested bo_va
@@ -2659,7 +2179,7 @@ void amdgpu_vm_bo_trace_cs(struct amdgpu_vm *vm, struct ww_acquire_ctx *ticket)
  *
  * Object have to be reserved!
  */
-void amdgpu_vm_bo_rmv(struct amdgpu_device *adev,
+void amdgpu_vm_bo_del(struct amdgpu_device *adev,
 		      struct amdgpu_bo_va *bo_va)
 {
 	struct amdgpu_bo_va_mapping *mapping, *next;
@@ -2667,23 +2187,27 @@ void amdgpu_vm_bo_rmv(struct amdgpu_device *adev,
 	struct amdgpu_vm *vm = bo_va->base.vm;
 	struct amdgpu_vm_bo_base **base;
 
+	dma_resv_assert_held(vm->root.bo->tbo.base.resv);
+
 	if (bo) {
-		if (bo->tbo.base.resv == vm->root.bo->tbo.base.resv)
-			vm->bulk_moveable = false;
+		dma_resv_assert_held(bo->tbo.base.resv);
+		if (amdgpu_vm_is_bo_always_valid(vm, bo))
+			ttm_bo_set_bulk_move(&bo->tbo, NULL);
 
 		for (base = &bo_va->base.bo->vm_bo; *base;
 		     base = &(*base)->next) {
 			if (*base != &bo_va->base)
 				continue;
 
+			amdgpu_vm_update_stats(*base, bo->tbo.resource, -1);
 			*base = bo_va->base.next;
 			break;
 		}
 	}
 
-	spin_lock(&vm->invalidated_lock);
+	spin_lock(&vm->status_lock);
 	list_del(&bo_va->base.vm_status);
-	spin_unlock(&vm->invalidated_lock);
+	spin_unlock(&vm->status_lock);
 
 	list_for_each_entry_safe(mapping, next, &bo_va->valids, list) {
 		list_del(&mapping->list);
@@ -2723,7 +2247,7 @@ bool amdgpu_vm_evictable(struct amdgpu_bo *bo)
 		return true;
 
 	/* Don't evict VM page tables while they are busy */
-	if (!dma_resv_test_signaled(bo->tbo.base.resv, true))
+	if (!dma_resv_test_signaled(bo->tbo.base.resv, DMA_RESV_USAGE_BOOKKEEP))
 		return false;
 
 	/* Try to block ongoing updates */
@@ -2744,25 +2268,19 @@ bool amdgpu_vm_evictable(struct amdgpu_bo *bo)
 /**
  * amdgpu_vm_bo_invalidate - mark the bo as invalid
  *
- * @adev: amdgpu_device pointer
  * @bo: amdgpu buffer object
  * @evicted: is the BO evicted
  *
  * Mark @bo as invalid.
  */
-void amdgpu_vm_bo_invalidate(struct amdgpu_device *adev,
-			     struct amdgpu_bo *bo, bool evicted)
+void amdgpu_vm_bo_invalidate(struct amdgpu_bo *bo, bool evicted)
 {
 	struct amdgpu_vm_bo_base *bo_base;
 
-	/* shadow bo doesn't have bo base, its validation needs its parent */
-	if (bo->parent && (amdgpu_bo_shadowed(bo->parent) == bo))
-		bo = bo->parent;
-
 	for (bo_base = bo->vm_bo; bo_base; bo_base = bo_base->next) {
 		struct amdgpu_vm *vm = bo_base->vm;
 
-		if (evicted && bo->tbo.base.resv == vm->root.bo->tbo.base.resv) {
+		if (evicted && amdgpu_vm_is_bo_always_valid(vm, bo)) {
 			amdgpu_vm_bo_evicted(bo_base);
 			continue;
 		}
@@ -2773,7 +2291,7 @@ void amdgpu_vm_bo_invalidate(struct amdgpu_device *adev,
 
 		if (bo->tbo.type == ttm_bo_type_kernel)
 			amdgpu_vm_bo_relocated(bo_base);
-		else if (bo->tbo.base.resv == vm->root.bo->tbo.base.resv)
+		else if (amdgpu_vm_is_bo_always_valid(vm, bo))
 			amdgpu_vm_bo_moved(bo_base);
 		else
 			amdgpu_vm_bo_invalidated(bo_base);
@@ -2781,6 +2299,32 @@ void amdgpu_vm_bo_invalidate(struct amdgpu_device *adev,
 }
 
 /**
+ * amdgpu_vm_bo_move - handle BO move
+ *
+ * @bo: amdgpu buffer object
+ * @new_mem: the new placement of the BO move
+ * @evicted: is the BO evicted
+ *
+ * Update the memory stats for the new placement and mark @bo as invalid.
+ */
+void amdgpu_vm_bo_move(struct amdgpu_bo *bo, struct ttm_resource *new_mem,
+		       bool evicted)
+{
+	struct amdgpu_vm_bo_base *bo_base;
+
+	for (bo_base = bo->vm_bo; bo_base; bo_base = bo_base->next) {
+		struct amdgpu_vm *vm = bo_base->vm;
+
+		spin_lock(&vm->status_lock);
+		amdgpu_vm_update_stats_locked(bo_base, bo->tbo.resource, -1);
+		amdgpu_vm_update_stats_locked(bo_base, new_mem, +1);
+		spin_unlock(&vm->status_lock);
+	}
+
+	amdgpu_vm_bo_invalidate(bo, evicted);
+}
+
+/**
  * amdgpu_vm_get_block_size - calculate VM page table size as power of two
  *
  * @vm_size: VM size
@@ -2850,7 +2394,7 @@ void amdgpu_vm_adjust_size(struct amdgpu_device *adev, uint32_t min_vm_size,
 		phys_ram_gb = ((uint64_t)si.totalram * si.mem_unit +
 			       (1 << 30) - 1) >> 30;
 		vm_size = roundup_pow_of_two(
-			min(max(phys_ram_gb * 3, min_vm_size), max_size));
+			clamp(phys_ram_gb * 3, min_vm_size, max_size));
 	}
 
 	adev->vm_manager.max_pfn = (uint64_t)vm_size << 18;
@@ -2859,7 +2403,7 @@ void amdgpu_vm_adjust_size(struct amdgpu_device *adev, uint32_t min_vm_size,
 	if (amdgpu_vm_block_size != -1)
 		tmp >>= amdgpu_vm_block_size - 9;
 	tmp = DIV_ROUND_UP(fls64(tmp) - 1, 9) - 1;
-	adev->vm_manager.num_level = min(max_level, (unsigned)tmp);
+	adev->vm_manager.num_level = min_t(unsigned int, max_level, tmp);
 	switch (adev->vm_manager.num_level) {
 	case 3:
 		adev->vm_manager.root_level = AMDGPU_VM_PDB2;
@@ -2889,10 +2433,11 @@ void amdgpu_vm_adjust_size(struct amdgpu_device *adev, uint32_t min_vm_size,
 	else
 		adev->vm_manager.fragment_size = amdgpu_vm_fragment_size;
 
-	DRM_INFO("vm size is %u GB, %u levels, block size is %u-bit, fragment size is %u-bit\n",
-		 vm_size, adev->vm_manager.num_level + 1,
-		 adev->vm_manager.block_size,
-		 adev->vm_manager.fragment_size);
+	dev_info(
+		adev->dev,
+		"vm size is %u GB, %u levels, block size is %u-bit, fragment size is %u-bit\n",
+		vm_size, adev->vm_manager.num_level + 1,
+		adev->vm_manager.block_size, adev->vm_manager.fragment_size);
 }
 
 /**
@@ -2903,12 +2448,114 @@ void amdgpu_vm_adjust_size(struct amdgpu_device *adev, uint32_t min_vm_size,
  */
 long amdgpu_vm_wait_idle(struct amdgpu_vm *vm, long timeout)
 {
-	timeout = dma_resv_wait_timeout(vm->root.bo->tbo.base.resv, true,
-					true, timeout);
+	timeout = drm_sched_entity_flush(&vm->immediate, timeout);
 	if (timeout <= 0)
 		return timeout;
 
-	return dma_fence_wait_timeout(vm->last_unlocked, true, timeout);
+	return drm_sched_entity_flush(&vm->delayed, timeout);
+}
+
+static void amdgpu_vm_destroy_task_info(struct kref *kref)
+{
+	struct amdgpu_task_info *ti = container_of(kref, struct amdgpu_task_info, refcount);
+
+	kfree(ti);
+}
+
+static inline struct amdgpu_vm *
+amdgpu_vm_get_vm_from_pasid(struct amdgpu_device *adev, u32 pasid)
+{
+	struct amdgpu_vm *vm;
+	unsigned long flags;
+
+	xa_lock_irqsave(&adev->vm_manager.pasids, flags);
+	vm = xa_load(&adev->vm_manager.pasids, pasid);
+	xa_unlock_irqrestore(&adev->vm_manager.pasids, flags);
+
+	return vm;
+}
+
+/**
+ * amdgpu_vm_put_task_info - reference down the vm task_info ptr
+ *
+ * @task_info: task_info struct under discussion.
+ *
+ * frees the vm task_info ptr at the last put
+ */
+void amdgpu_vm_put_task_info(struct amdgpu_task_info *task_info)
+{
+	if (task_info)
+		kref_put(&task_info->refcount, amdgpu_vm_destroy_task_info);
+}
+
+/**
+ * amdgpu_vm_get_task_info_vm - Extracts task info for a vm.
+ *
+ * @vm: VM to get info from
+ *
+ * Returns the reference counted task_info structure, which must be
+ * referenced down with amdgpu_vm_put_task_info.
+ */
+struct amdgpu_task_info *
+amdgpu_vm_get_task_info_vm(struct amdgpu_vm *vm)
+{
+	struct amdgpu_task_info *ti = NULL;
+
+	if (vm) {
+		ti = vm->task_info;
+		kref_get(&vm->task_info->refcount);
+	}
+
+	return ti;
+}
+
+/**
+ * amdgpu_vm_get_task_info_pasid - Extracts task info for a PASID.
+ *
+ * @adev: drm device pointer
+ * @pasid: PASID identifier for VM
+ *
+ * Returns the reference counted task_info structure, which must be
+ * referenced down with amdgpu_vm_put_task_info.
+ */
+struct amdgpu_task_info *
+amdgpu_vm_get_task_info_pasid(struct amdgpu_device *adev, u32 pasid)
+{
+	return amdgpu_vm_get_task_info_vm(
+			amdgpu_vm_get_vm_from_pasid(adev, pasid));
+}
+
+static int amdgpu_vm_create_task_info(struct amdgpu_vm *vm)
+{
+	vm->task_info = kzalloc(sizeof(struct amdgpu_task_info), GFP_KERNEL);
+	if (!vm->task_info)
+		return -ENOMEM;
+
+	kref_init(&vm->task_info->refcount);
+	return 0;
+}
+
+/**
+ * amdgpu_vm_set_task_info - Sets VMs task info.
+ *
+ * @vm: vm for which to set the info
+ */
+void amdgpu_vm_set_task_info(struct amdgpu_vm *vm)
+{
+	if (!vm->task_info)
+		return;
+
+	if (vm->task_info->task.pid == current->pid)
+		return;
+
+	vm->task_info->task.pid = current->pid;
+	get_task_comm(vm->task_info->task.comm, current);
+
+	if (current->group_leader->mm != current->mm)
+		return;
+
+	vm->task_info->tgid = current->group_leader->pid;
+	get_task_comm(vm->task_info->process_name, current->group_leader);
 }
 
 /**
@@ -2916,13 +2563,16 @@ long amdgpu_vm_wait_idle(struct amdgpu_vm *vm, long timeout)
  *
  * @adev: amdgpu_device pointer
  * @vm: requested vm
+ * @xcp_id: GPU partition selection id
+ * @pasid: the pasid the VM is using on this GPU
  *
  * Init @vm fields.
  *
  * Returns:
  * 0 for success, error for failure.
  */
-int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm)
+int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm,
+		   int32_t xcp_id, uint32_t pasid)
 {
 	struct amdgpu_bo *root_bo;
 	struct amdgpu_bo_vm *root;
@@ -2932,35 +2582,29 @@ int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm)
 	for (i = 0; i < AMDGPU_MAX_VMHUBS; i++)
 		vm->reserved_vmid[i] = NULL;
 	INIT_LIST_HEAD(&vm->evicted);
+	INIT_LIST_HEAD(&vm->evicted_user);
 	INIT_LIST_HEAD(&vm->relocated);
 	INIT_LIST_HEAD(&vm->moved);
 	INIT_LIST_HEAD(&vm->idle);
 	INIT_LIST_HEAD(&vm->invalidated);
-	spin_lock_init(&vm->invalidated_lock);
+	spin_lock_init(&vm->status_lock);
 	INIT_LIST_HEAD(&vm->freed);
 	INIT_LIST_HEAD(&vm->done);
+	INIT_KFIFO(vm->faults);
 
-	/* create scheduler entities for page table updates */
-	r = drm_sched_entity_init(&vm->immediate, DRM_SCHED_PRIORITY_NORMAL,
-				  adev->vm_manager.vm_pte_scheds,
-				  adev->vm_manager.vm_pte_num_scheds, NULL);
+	r = amdgpu_vm_init_entities(adev, vm);
 	if (r)
 		return r;
 
-	r = drm_sched_entity_init(&vm->delayed, DRM_SCHED_PRIORITY_NORMAL,
-				  adev->vm_manager.vm_pte_scheds,
-				  adev->vm_manager.vm_pte_num_scheds, NULL);
-	if (r)
-		goto error_free_immediate;
+	ttm_lru_bulk_move_init(&vm->lru_bulk_move);
 
-	vm->pte_support_ats = false;
 	vm->is_compute_context = false;
 
 	vm->use_cpu_for_update = !!(adev->vm_manager.vm_update_mode &
 				    AMDGPU_VM_USE_CPU_FOR_GFX);
 
-	DRM_DEBUG_DRIVER("VM update mode is %s\n",
-			 vm->use_cpu_for_update ? "CPU" : "SDMA");
+	dev_dbg(adev->dev, "VM update mode is %s\n",
+		vm->use_cpu_for_update ? "CPU" : "SDMA");
 	WARN_ONCE((vm->use_cpu_for_update &&
 		   !amdgpu_gmc_vram_full_visible(&adev->gmc)),
 		  "CPU update of VM recommended only for large BAR system\n");
@@ -2969,84 +2613,75 @@ int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm)
 		vm->update_funcs = &amdgpu_vm_cpu_funcs;
 	else
 		vm->update_funcs = &amdgpu_vm_sdma_funcs;
-	vm->last_update = NULL;
+
+	vm->last_update = dma_fence_get_stub();
 	vm->last_unlocked = dma_fence_get_stub();
+	vm->last_tlb_flush = dma_fence_get_stub();
+	vm->generation = amdgpu_vm_generation(adev, NULL);
 
 	mutex_init(&vm->eviction_lock);
 	vm->evicting = false;
+	vm->tlb_fence_context = dma_fence_context_alloc(1);
 
 	r = amdgpu_vm_pt_create(adev, vm, adev->vm_manager.root_level,
-				false, &root);
+				false, &root, xcp_id);
 	if (r)
 		goto error_free_delayed;
-	root_bo = &root->bo;
+
+	root_bo = amdgpu_bo_ref(&root->bo);
 	r = amdgpu_bo_reserve(root_bo, true);
+	if (r) {
+		amdgpu_bo_unref(&root_bo);
+		goto error_free_delayed;
+	}
+
+	amdgpu_vm_bo_base_init(&vm->root, vm, root_bo);
+	r = dma_resv_reserve_fences(root_bo->tbo.base.resv, 1);
 	if (r)
 		goto error_free_root;
 
-	r = dma_resv_reserve_shared(root_bo->tbo.base.resv, 1);
+	r = amdgpu_vm_pt_clear(adev, vm, root, false);
 	if (r)
-		goto error_unreserve;
-
-	amdgpu_vm_bo_base_init(&vm->root, vm, root_bo);
+		goto error_free_root;
 
-	r = amdgpu_vm_clear_bo(adev, vm, root, false);
+	r = amdgpu_vm_create_task_info(vm);
 	if (r)
-		goto error_unreserve;
-
-	amdgpu_bo_unreserve(vm->root.bo);
+		dev_dbg(adev->dev, "Failed to create task info for VM\n");
 
-	INIT_KFIFO(vm->faults);
+	/* Store new PASID in XArray (if non-zero) */
+	if (pasid != 0) {
+		r = xa_err(xa_store_irq(&adev->vm_manager.pasids, pasid, vm, GFP_KERNEL));
+		if (r < 0)
+			goto error_free_root;
 
-	return 0;
+		vm->pasid = pasid;
+	}
 
-error_unreserve:
 	amdgpu_bo_unreserve(vm->root.bo);
+	amdgpu_bo_unref(&root_bo);
+
+	return 0;
 
 error_free_root:
-	amdgpu_bo_unref(&root->shadow);
+	/* If PASID was partially set, erase it from XArray before failing */
+	if (vm->pasid != 0) {
+		xa_erase_irq(&adev->vm_manager.pasids, vm->pasid);
+		vm->pasid = 0;
+	}
+	amdgpu_vm_pt_free_root(adev, vm);
+	amdgpu_bo_unreserve(vm->root.bo);
 	amdgpu_bo_unref(&root_bo);
-	vm->root.bo = NULL;
 
 error_free_delayed:
+	dma_fence_put(vm->last_tlb_flush);
 	dma_fence_put(vm->last_unlocked);
-	drm_sched_entity_destroy(&vm->delayed);
-
-error_free_immediate:
-	drm_sched_entity_destroy(&vm->immediate);
+	ttm_lru_bulk_move_fini(&adev->mman.bdev, &vm->lru_bulk_move);
+	amdgpu_vm_fini_entities(vm);
 
 	return r;
 }
 
 /**
- * amdgpu_vm_check_clean_reserved - check if a VM is clean
- *
- * @adev: amdgpu_device pointer
- * @vm: the VM to check
- *
- * check all entries of the root PD, if any subsequent PDs are allocated,
- * it means there are page table creating and filling, and is no a clean
- * VM
- *
- * Returns:
- *	0 if this VM is clean
- */
-static int amdgpu_vm_check_clean_reserved(struct amdgpu_device *adev,
-					  struct amdgpu_vm *vm)
-{
-	enum amdgpu_vm_level root = adev->vm_manager.root_level;
-	unsigned int entries = amdgpu_vm_num_entries(adev, root);
-	unsigned int i = 0;
-
-	for (i = 0; i < entries; i++) {
-		if (to_amdgpu_bo_vm(vm->root.bo)->entries[i].bo)
-			return -EINVAL;
-	}
-
-	return 0;
-}
-
-/**
  * amdgpu_vm_make_compute - Turn a GFX VM into a compute VM
  *
  * @adev: amdgpu_device pointer
@@ -3067,35 +2702,17 @@ static int amdgpu_vm_check_clean_reserved(struct amdgpu_device *adev,
  */
 int amdgpu_vm_make_compute(struct amdgpu_device *adev, struct amdgpu_vm *vm)
 {
-	bool pte_support_ats = (adev->asic_type == CHIP_RAVEN);
 	int r;
 
 	r = amdgpu_bo_reserve(vm->root.bo, true);
 	if (r)
 		return r;
 
-	/* Sanity checks */
-	r = amdgpu_vm_check_clean_reserved(adev, vm);
-	if (r)
-		goto unreserve_bo;
-
-	/* Check if PD needs to be reinitialized and do it before
-	 * changing any other state, in case it fails.
-	 */
-	if (pte_support_ats != vm->pte_support_ats) {
-		vm->pte_support_ats = pte_support_ats;
-		r = amdgpu_vm_clear_bo(adev, vm,
-				       to_amdgpu_bo_vm(vm->root.bo),
-				       false);
-		if (r)
-			goto unreserve_bo;
-	}
-
 	/* Update VM state */
 	vm->use_cpu_for_update = !!(adev->vm_manager.vm_update_mode &
 				    AMDGPU_VM_USE_CPU_FOR_COMPUTE);
-	DRM_DEBUG_DRIVER("VM update mode is %s\n",
-			 vm->use_cpu_for_update ? "CPU" : "SDMA");
+	dev_dbg(adev->dev, "VM update mode is %s\n",
+		vm->use_cpu_for_update ? "CPU" : "SDMA");
 	WARN_ONCE((vm->use_cpu_for_update &&
 		   !amdgpu_gmc_vram_full_visible(&adev->gmc)),
 		  "CPU update of VM recommended only for large BAR system\n");
@@ -3108,35 +2725,31 @@ int amdgpu_vm_make_compute(struct amdgpu_device *adev, struct amdgpu_vm *vm)
 			goto unreserve_bo;
 
 		vm->update_funcs = &amdgpu_vm_cpu_funcs;
+		r = amdgpu_vm_pt_map_tables(adev, vm);
+		if (r)
+			goto unreserve_bo;
+
 	} else {
 		vm->update_funcs = &amdgpu_vm_sdma_funcs;
 	}
+
 	dma_fence_put(vm->last_update);
-	vm->last_update = NULL;
+	vm->last_update = dma_fence_get_stub();
 	vm->is_compute_context = true;
 
-	/* Free the shadow bo for compute VM */
-	amdgpu_bo_unref(&to_amdgpu_bo_vm(vm->root.bo)->shadow);
-
-	goto unreserve_bo;
-
 unreserve_bo:
 	amdgpu_bo_unreserve(vm->root.bo);
 	return r;
 }
 
-/**
- * amdgpu_vm_release_compute - release a compute vm
- * @adev: amdgpu_device pointer
- * @vm: a vm turned into compute vm by calling amdgpu_vm_make_compute
- *
- * This is a correspondant of amdgpu_vm_make_compute. It decouples compute
- * pasid from vm. Compute should stop use of vm after this call.
- */
-void amdgpu_vm_release_compute(struct amdgpu_device *adev, struct amdgpu_vm *vm)
+static int amdgpu_vm_stats_is_zero(struct amdgpu_vm *vm)
 {
-	amdgpu_vm_set_pasid(adev, vm, 0);
-	vm->is_compute_context = false;
+	for (int i = 0; i < __AMDGPU_PL_NUM; ++i) {
+		if (!(drm_memory_stats_is_zero(&vm->stats[i].drm) &&
+		      vm->stats[i].evicted == 0))
+			return false;
+	}
+	return true;
 }
 
 /**
@@ -3153,18 +2766,28 @@ void amdgpu_vm_fini(struct amdgpu_device *adev, struct amdgpu_vm *vm)
 	struct amdgpu_bo_va_mapping *mapping, *tmp;
 	bool prt_fini_needed = !!adev->gmc.gmc_funcs->set_prt;
 	struct amdgpu_bo *root;
+	unsigned long flags;
 	int i;
 
 	amdgpu_amdkfd_gpuvm_destroy_cb(adev, vm);
 
 	root = amdgpu_bo_ref(vm->root.bo);
 	amdgpu_bo_reserve(root, true);
-	amdgpu_vm_set_pasid(adev, vm, 0);
+	/* Remove PASID mapping before destroying VM */
+	if (vm->pasid != 0) {
+		xa_erase_irq(&adev->vm_manager.pasids, vm->pasid);
+		vm->pasid = 0;
+	}
 	dma_fence_wait(vm->last_unlocked, false);
 	dma_fence_put(vm->last_unlocked);
+	dma_fence_wait(vm->last_tlb_flush, false);
+	/* Make sure that all fence callbacks have completed */
+	spin_lock_irqsave(vm->last_tlb_flush->lock, flags);
+	spin_unlock_irqrestore(vm->last_tlb_flush->lock, flags);
+	dma_fence_put(vm->last_tlb_flush);
 
 	list_for_each_entry_safe(mapping, tmp, &vm->freed, list) {
-		if (mapping->flags & AMDGPU_PTE_PRT && prt_fini_needed) {
+		if (mapping->flags & AMDGPU_VM_PAGE_PRT && prt_fini_needed) {
 			amdgpu_vm_prt_fini(adev, vm);
 			prt_fini_needed = false;
 		}
@@ -3173,13 +2796,12 @@ void amdgpu_vm_fini(struct amdgpu_device *adev, struct amdgpu_vm *vm)
 		amdgpu_vm_free_mapping(adev, vm, mapping, NULL);
 	}
 
-	amdgpu_vm_free_pts(adev, vm, NULL);
+	amdgpu_vm_pt_free_root(adev, vm);
 	amdgpu_bo_unreserve(root);
 	amdgpu_bo_unref(&root);
 	WARN_ON(vm->root.bo);
 
-	drm_sched_entity_destroy(&vm->immediate);
-	drm_sched_entity_destroy(&vm->delayed);
+	amdgpu_vm_fini_entities(vm);
 
 	if (!RB_EMPTY_ROOT(&vm->va.rb_root)) {
 		dev_err(adev->dev, "still active bo inside vm\n");
@@ -3194,8 +2816,22 @@ void amdgpu_vm_fini(struct amdgpu_device *adev, struct amdgpu_vm *vm)
 	}
 
 	dma_fence_put(vm->last_update);
-	for (i = 0; i < AMDGPU_MAX_VMHUBS; i++)
+
+	for (i = 0; i < AMDGPU_MAX_VMHUBS; i++) {
 		amdgpu_vmid_free_reserved(adev, vm, i);
+	}
+
+	ttm_lru_bulk_move_fini(&adev->mman.bdev, &vm->lru_bulk_move);
+
+	if (!amdgpu_vm_stats_is_zero(vm)) {
+		struct amdgpu_task_info *ti = vm->task_info;
+
+		dev_warn(adev->dev,
+			 "VM memory stats for proc %s(%d) task %s(%d) is non-zero when fini\n",
+			 ti->process_name, ti->task.pid, ti->task.comm, ti->tgid);
+	}
+
+	amdgpu_vm_put_task_info(vm->task_info);
 }
 
 /**
@@ -3207,8 +2843,6 @@ void amdgpu_vm_fini(struct amdgpu_device *adev, struct amdgpu_vm *vm)
  */
 void amdgpu_vm_manager_init(struct amdgpu_device *adev)
 {
-	unsigned i;
-
 	/* Concurrent flushes are only possible starting with Vega10 and
 	 * are broken on Navi10 and Navi14.
 	 */
@@ -3217,11 +2851,6 @@ void amdgpu_vm_manager_init(struct amdgpu_device *adev)
 					      adev->asic_type == CHIP_NAVI14);
 	amdgpu_vmid_mgr_init(adev);
 
-	adev->vm_manager.fence_context =
-		dma_fence_context_alloc(AMDGPU_MAX_RINGS);
-	for (i = 0; i < AMDGPU_MAX_RINGS; ++i)
-		adev->vm_manager.seqno[i] = 0;
-
 	spin_lock_init(&adev->vm_manager.prt_lock);
 	atomic_set(&adev->vm_manager.num_prt_users, 0);
 
@@ -3230,7 +2859,11 @@ void amdgpu_vm_manager_init(struct amdgpu_device *adev)
 	 */
 #ifdef CONFIG_X86_64
 	if (amdgpu_vm_update_mode == -1) {
-		if (amdgpu_gmc_vram_full_visible(&adev->gmc))
+		/* For asic with VF MMIO access protection
+		 * avoid using CPU for VM table updates
+		 */
+		if (amdgpu_gmc_vram_full_visible(&adev->gmc) &&
+		    !amdgpu_sriov_vf_mmio_access_protection(adev))
 			adev->vm_manager.vm_update_mode =
 				AMDGPU_VM_USE_CPU_FOR_COMPUTE;
 		else
@@ -3274,34 +2907,18 @@ int amdgpu_vm_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 	union drm_amdgpu_vm *args = data;
 	struct amdgpu_device *adev = drm_to_adev(dev);
 	struct amdgpu_fpriv *fpriv = filp->driver_priv;
-	long timeout = msecs_to_jiffies(2000);
-	int r;
+	struct amdgpu_vm *vm = &fpriv->vm;
+
+	/* No valid flags defined yet */
+	if (args->in.flags)
+		return -EINVAL;
 
 	switch (args->in.op) {
 	case AMDGPU_VM_OP_RESERVE_VMID:
 		/* We only have requirement to reserve vmid from gfxhub */
-		r = amdgpu_vmid_alloc_reserved(adev, &fpriv->vm,
-					       AMDGPU_GFXHUB_0);
-		if (r)
-			return r;
-		break;
+		return amdgpu_vmid_alloc_reserved(adev, vm, AMDGPU_GFXHUB(0));
 	case AMDGPU_VM_OP_UNRESERVE_VMID:
-		if (amdgpu_sriov_runtime(adev))
-			timeout = 8 * timeout;
-
-		/* Wait vm idle to make sure the vmid set in SPM_VMID is
-		 * not referenced anymore.
-		 */
-		r = amdgpu_bo_reserve(fpriv->vm.root.bo, true);
-		if (r)
-			return r;
-
-		r = amdgpu_vm_wait_idle(&fpriv->vm, timeout);
-		if (r < 0)
-			return r;
-
-		amdgpu_bo_unreserve(fpriv->vm.root.bo);
-		amdgpu_vmid_free_reserved(adev, &fpriv->vm, AMDGPU_GFXHUB_0);
+		amdgpu_vmid_free_reserved(adev, vm, AMDGPU_GFXHUB(0));
 		break;
 	default:
 		return -EINVAL;
@@ -3311,51 +2928,13 @@ int amdgpu_vm_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 }
 
 /**
- * amdgpu_vm_get_task_info - Extracts task info for a PASID.
- *
- * @adev: drm device pointer
- * @pasid: PASID identifier for VM
- * @task_info: task_info to fill.
- */
-void amdgpu_vm_get_task_info(struct amdgpu_device *adev, u32 pasid,
-			 struct amdgpu_task_info *task_info)
-{
-	struct amdgpu_vm *vm;
-	unsigned long flags;
-
-	xa_lock_irqsave(&adev->vm_manager.pasids, flags);
-
-	vm = xa_load(&adev->vm_manager.pasids, pasid);
-	if (vm)
-		*task_info = vm->task_info;
-
-	xa_unlock_irqrestore(&adev->vm_manager.pasids, flags);
-}
-
-/**
- * amdgpu_vm_set_task_info - Sets VMs task info.
- *
- * @vm: vm for which to set the info
- */
-void amdgpu_vm_set_task_info(struct amdgpu_vm *vm)
-{
-	if (vm->task_info.pid)
-		return;
-
-	vm->task_info.pid = current->pid;
-	get_task_comm(vm->task_info.task_name, current);
-
-	if (current->group_leader->mm != current->mm)
-		return;
-
-	vm->task_info.tgid = current->group_leader->pid;
-	get_task_comm(vm->task_info.process_name, current->group_leader);
-}
-
-/**
  * amdgpu_vm_handle_fault - graceful handling of VM faults.
  * @adev: amdgpu device pointer
  * @pasid: PASID of the VM
+ * @ts: Timestamp of the fault
+ * @vmid: VMID, only used for GFX 9.4.3.
+ * @node_id: Node_id received in IH cookie. Only applicable for
+ *           GFX 9.4.3.
  * @addr: Address of the fault
  * @write_fault: true is write fault, false is read fault
  *
@@ -3363,7 +2942,8 @@ void amdgpu_vm_set_task_info(struct amdgpu_vm *vm)
  * shouldn't be reported any more.
  */
 bool amdgpu_vm_handle_fault(struct amdgpu_device *adev, u32 pasid,
-			    uint64_t addr, bool write_fault)
+			    u32 vmid, u32 node_id, uint64_t addr, uint64_t ts,
+			    bool write_fault)
 {
 	bool is_compute_context = false;
 	struct amdgpu_bo *root;
@@ -3387,8 +2967,8 @@ bool amdgpu_vm_handle_fault(struct amdgpu_device *adev, u32 pasid,
 
 	addr /= AMDGPU_GPU_PAGE_SIZE;
 
-	if (is_compute_context &&
-	    !svm_range_restore_pages(adev, pasid, addr, write_fault)) {
+	if (is_compute_context && !svm_range_restore_pages(adev, pasid, vmid,
+	    node_id, addr, ts, write_fault)) {
 		amdgpu_bo_unref(&root);
 		return true;
 	}
@@ -3413,8 +2993,7 @@ bool amdgpu_vm_handle_fault(struct amdgpu_device *adev, u32 pasid,
 		/* Intentionally setting invalid PTE flag
 		 * combination to force a no-retry-fault
 		 */
-		flags = AMDGPU_PTE_EXECUTABLE | AMDGPU_PDE_PTE |
-			AMDGPU_PTE_TF;
+		flags = AMDGPU_VM_NORETRY_FLAGS;
 		value = 0;
 	} else if (amdgpu_vm_fault_stop == AMDGPU_VM_FAULT_STOP_NEVER) {
 		/* Redirect the access to the dummy page */
@@ -3427,15 +3006,14 @@ bool amdgpu_vm_handle_fault(struct amdgpu_device *adev, u32 pasid,
 		value = 0;
 	}
 
-	r = dma_resv_reserve_shared(root->tbo.base.resv, 1);
+	r = dma_resv_reserve_fences(root->tbo.base.resv, 1);
 	if (r) {
 		pr_debug("failed %d to reserve fence slot\n", r);
 		goto error_unlock;
 	}
 
-	r = amdgpu_vm_bo_update_mapping(adev, adev, vm, true, false, NULL, addr,
-					addr, flags, value, NULL, NULL, NULL,
-					NULL);
+	r = amdgpu_vm_update_range(adev, vm, true, false, false, false,
+				   NULL, addr, addr, flags, value, 0, NULL, NULL, NULL);
 	if (r)
 		goto error_unlock;
 
@@ -3444,7 +3022,7 @@ bool amdgpu_vm_handle_fault(struct amdgpu_device *adev, u32 pasid,
 error_unlock:
 	amdgpu_bo_unreserve(root);
 	if (r < 0)
-		DRM_ERROR("Can't handle page fault (%d)\n", r);
+		dev_err(adev->dev, "Can't handle page fault (%d)\n", r);
 
 error_unref:
 	amdgpu_bo_unref(&root);
@@ -3478,6 +3056,9 @@ void amdgpu_debugfs_vm_bo_info(struct amdgpu_vm *vm, struct seq_file *m)
 	unsigned int total_done_objs = 0;
 	unsigned int id = 0;
 
+	amdgpu_vm_assert_locked(vm);
+
+	spin_lock(&vm->status_lock);
 	seq_puts(m, "\tIdle BOs:\n");
 	list_for_each_entry_safe(bo_va, tmp, &vm->idle, base.vm_status) {
 		if (!bo_va->base.bo)
@@ -3515,7 +3096,6 @@ void amdgpu_debugfs_vm_bo_info(struct amdgpu_vm *vm, struct seq_file *m)
 	id = 0;
 
 	seq_puts(m, "\tInvalidated BOs:\n");
-	spin_lock(&vm->invalidated_lock);
 	list_for_each_entry_safe(bo_va, tmp, &vm->invalidated, base.vm_status) {
 		if (!bo_va->base.bo)
 			continue;
@@ -3530,7 +3110,7 @@ void amdgpu_debugfs_vm_bo_info(struct amdgpu_vm *vm, struct seq_file *m)
 			continue;
 		total_done += amdgpu_bo_print_info(id++, bo_va->base.bo, m);
 	}
-	spin_unlock(&vm->invalidated_lock);
+	spin_unlock(&vm->status_lock);
 	total_done_objs = id;
 
 	seq_printf(m, "\tTotal idle size:        %12lld\tobjs:\t%d\n", total_idle,
@@ -3547,3 +3127,83 @@ void amdgpu_debugfs_vm_bo_info(struct amdgpu_vm *vm, struct seq_file *m)
 		   total_done_objs);
 }
 #endif
+
+/**
+ * amdgpu_vm_update_fault_cache - update cached fault into.
+ * @adev: amdgpu device pointer
+ * @pasid: PASID of the VM
+ * @addr: Address of the fault
+ * @status: GPUVM fault status register
+ * @vmhub: which vmhub got the fault
+ *
+ * Cache the fault info for later use by userspace in debugging.
+ */
+void amdgpu_vm_update_fault_cache(struct amdgpu_device *adev,
+				  unsigned int pasid,
+				  uint64_t addr,
+				  uint32_t status,
+				  unsigned int vmhub)
+{
+	struct amdgpu_vm *vm;
+	unsigned long flags;
+
+	xa_lock_irqsave(&adev->vm_manager.pasids, flags);
+
+	vm = xa_load(&adev->vm_manager.pasids, pasid);
+	/* Don't update the fault cache if status is 0.  In the multiple
+	 * fault case, subsequent faults will return a 0 status which is
+	 * useless for userspace and replaces the useful fault status, so
+	 * only update if status is non-0.
+	 */
+	if (vm && status) {
+		vm->fault_info.addr = addr;
+		vm->fault_info.status = status;
+		/*
+		 * Update the fault information globally for later usage
+		 * when vm could be stale or freed.
+		 */
+		adev->vm_manager.fault_info.addr = addr;
+		adev->vm_manager.fault_info.vmhub = vmhub;
+		adev->vm_manager.fault_info.status = status;
+
+		if (AMDGPU_IS_GFXHUB(vmhub)) {
+			vm->fault_info.vmhub = AMDGPU_VMHUB_TYPE_GFX;
+			vm->fault_info.vmhub |=
+				(vmhub - AMDGPU_GFXHUB_START) << AMDGPU_VMHUB_IDX_SHIFT;
+		} else if (AMDGPU_IS_MMHUB0(vmhub)) {
+			vm->fault_info.vmhub = AMDGPU_VMHUB_TYPE_MM0;
+			vm->fault_info.vmhub |=
+				(vmhub - AMDGPU_MMHUB0_START) << AMDGPU_VMHUB_IDX_SHIFT;
+		} else if (AMDGPU_IS_MMHUB1(vmhub)) {
+			vm->fault_info.vmhub = AMDGPU_VMHUB_TYPE_MM1;
+			vm->fault_info.vmhub |=
+				(vmhub - AMDGPU_MMHUB1_START) << AMDGPU_VMHUB_IDX_SHIFT;
+		} else {
+			WARN_ONCE(1, "Invalid vmhub %u\n", vmhub);
+		}
+	}
+	xa_unlock_irqrestore(&adev->vm_manager.pasids, flags);
+}
+
+/**
+ * amdgpu_vm_is_bo_always_valid - check if the BO is VM always valid
+ *
+ * @vm: VM to test against.
+ * @bo: BO to be tested.
+ *
+ * Returns true if the BO shares the dma_resv object with the root PD and is
+ * always guaranteed to be valid inside the VM.
+ */
+bool amdgpu_vm_is_bo_always_valid(struct amdgpu_vm *vm, struct amdgpu_bo *bo)
+{
+	return bo && bo->tbo.base.resv == vm->root.bo->tbo.base.resv;
+}
+
+void amdgpu_vm_print_task_info(struct amdgpu_device *adev,
+			       struct amdgpu_task_info *task_info)
+{
+	dev_err(adev->dev,
+		" Process %s pid %d thread %s pid %d\n",
+		task_info->process_name, task_info->tgid,
+		task_info->task.comm, task_info->task.pid);
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
index 85fcfb8c5efd..15d757c016cb 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
@@ -29,12 +29,15 @@
 #include <linux/rbtree.h>
 #include <drm/gpu_scheduler.h>
 #include <drm/drm_file.h>
-#include <drm/ttm/ttm_bo_driver.h>
+#include <drm/ttm/ttm_bo.h>
 #include <linux/sched/mm.h>
 
 #include "amdgpu_sync.h"
 #include "amdgpu_ring.h"
 #include "amdgpu_ids.h"
+#include "amdgpu_ttm.h"
+
+struct drm_exec;
 
 struct amdgpu_bo_va;
 struct amdgpu_job;
@@ -83,10 +86,19 @@ struct amdgpu_bo_vm;
 /* PDE Block Fragment Size for VEGA10 */
 #define AMDGPU_PDE_BFS(a)	((uint64_t)a << 59)
 
+/* Flag combination to set no-retry with TF disabled */
+#define AMDGPU_VM_NORETRY_FLAGS	(AMDGPU_PTE_EXECUTABLE | AMDGPU_PDE_PTE | \
+				AMDGPU_PTE_TF)
 
+/* Flag combination to set no-retry with TF enabled */
+#define AMDGPU_VM_NORETRY_FLAGS_TF (AMDGPU_PTE_VALID | AMDGPU_PTE_SYSTEM | \
+				   AMDGPU_PTE_PRT)
 /* For GFX9 */
-#define AMDGPU_PTE_MTYPE_VG10(a)	((uint64_t)(a) << 57)
-#define AMDGPU_PTE_MTYPE_VG10_MASK	AMDGPU_PTE_MTYPE_VG10(3ULL)
+#define AMDGPU_PTE_MTYPE_VG10_SHIFT(mtype)	((uint64_t)(mtype) << 57)
+#define AMDGPU_PTE_MTYPE_VG10_MASK	AMDGPU_PTE_MTYPE_VG10_SHIFT(3ULL)
+#define AMDGPU_PTE_MTYPE_VG10(flags, mtype)			\
+	(((uint64_t)(flags) & (~AMDGPU_PTE_MTYPE_VG10_MASK)) |	\
+	  AMDGPU_PTE_MTYPE_VG10_SHIFT(mtype))
 
 #define AMDGPU_MTYPE_NC 0
 #define AMDGPU_MTYPE_CC 2
@@ -99,28 +111,74 @@ struct amdgpu_bo_vm;
                                 | AMDGPU_PTE_MTYPE_VG10(AMDGPU_MTYPE_CC))
 
 /* gfx10 */
-#define AMDGPU_PTE_MTYPE_NV10(a)       ((uint64_t)(a) << 48)
-#define AMDGPU_PTE_MTYPE_NV10_MASK     AMDGPU_PTE_MTYPE_NV10(7ULL)
+#define AMDGPU_PTE_MTYPE_NV10_SHIFT(mtype)	((uint64_t)(mtype) << 48)
+#define AMDGPU_PTE_MTYPE_NV10_MASK     AMDGPU_PTE_MTYPE_NV10_SHIFT(7ULL)
+#define AMDGPU_PTE_MTYPE_NV10(flags, mtype)			\
+	(((uint64_t)(flags) & (~AMDGPU_PTE_MTYPE_NV10_MASK)) |	\
+	  AMDGPU_PTE_MTYPE_NV10_SHIFT(mtype))
+
+/* gfx12 */
+#define AMDGPU_PTE_PRT_GFX12		(1ULL << 56)
+#define AMDGPU_PTE_PRT_FLAG(adev)	\
+	((amdgpu_ip_version((adev), GC_HWIP, 0) >= IP_VERSION(12, 0, 0)) ? AMDGPU_PTE_PRT_GFX12 : AMDGPU_PTE_PRT)
+
+#define AMDGPU_PTE_MTYPE_GFX12_SHIFT(mtype)	((uint64_t)(mtype) << 54)
+#define AMDGPU_PTE_MTYPE_GFX12_MASK	AMDGPU_PTE_MTYPE_GFX12_SHIFT(3ULL)
+#define AMDGPU_PTE_MTYPE_GFX12(flags, mtype)				\
+	(((uint64_t)(flags) & (~AMDGPU_PTE_MTYPE_GFX12_MASK)) |	\
+	  AMDGPU_PTE_MTYPE_GFX12_SHIFT(mtype))
+
+#define AMDGPU_PTE_DCC			(1ULL << 58)
+#define AMDGPU_PTE_IS_PTE		(1ULL << 63)
+
+/* PDE Block Fragment Size for gfx v12 */
+#define AMDGPU_PDE_BFS_GFX12(a)		((uint64_t)((a) & 0x1fULL) << 58)
+#define AMDGPU_PDE_BFS_FLAG(adev, a)	\
+	((amdgpu_ip_version((adev), GC_HWIP, 0) >= IP_VERSION(12, 0, 0)) ? AMDGPU_PDE_BFS_GFX12(a) : AMDGPU_PDE_BFS(a))
+/* PDE is handled as PTE for gfx v12 */
+#define AMDGPU_PDE_PTE_GFX12		(1ULL << 63)
+#define AMDGPU_PDE_PTE_FLAG(adev)	\
+	((amdgpu_ip_version((adev), GC_HWIP, 0) >= IP_VERSION(12, 0, 0)) ? AMDGPU_PDE_PTE_GFX12 : AMDGPU_PDE_PTE)
 
 /* How to program VM fault handling */
 #define AMDGPU_VM_FAULT_STOP_NEVER	0
 #define AMDGPU_VM_FAULT_STOP_FIRST	1
 #define AMDGPU_VM_FAULT_STOP_ALWAYS	2
 
-/* Reserve 4MB VRAM for page tables */
+/* How much VRAM be reserved for page tables */
 #define AMDGPU_VM_RESERVED_VRAM		(8ULL << 20)
 
-/* max number of VMHUB */
-#define AMDGPU_MAX_VMHUBS			3
-#define AMDGPU_GFXHUB_0				0
-#define AMDGPU_MMHUB_0				1
-#define AMDGPU_MMHUB_1				2
-
-/* Reserve 2MB at top/bottom of address space for kernel use */
-#define AMDGPU_VA_RESERVED_SIZE			(2ULL << 20)
-
-/* max vmids dedicated for process */
-#define AMDGPU_VM_MAX_RESERVED_VMID	1
+/*
+ * max number of VMHUB
+ * layout: max 8 GFXHUB + 4 MMHUB0 + 1 MMHUB1
+ */
+#define AMDGPU_MAX_VMHUBS			13
+#define AMDGPU_GFXHUB_START			0
+#define AMDGPU_MMHUB0_START			8
+#define AMDGPU_MMHUB1_START			12
+#define AMDGPU_GFXHUB(x)			(AMDGPU_GFXHUB_START + (x))
+#define AMDGPU_MMHUB0(x)			(AMDGPU_MMHUB0_START + (x))
+#define AMDGPU_MMHUB1(x)			(AMDGPU_MMHUB1_START + (x))
+
+#define AMDGPU_IS_GFXHUB(x) ((x) >= AMDGPU_GFXHUB_START && (x) < AMDGPU_MMHUB0_START)
+#define AMDGPU_IS_MMHUB0(x) ((x) >= AMDGPU_MMHUB0_START && (x) < AMDGPU_MMHUB1_START)
+#define AMDGPU_IS_MMHUB1(x) ((x) >= AMDGPU_MMHUB1_START && (x) < AMDGPU_MAX_VMHUBS)
+
+/* Reserve space at top/bottom of address space for kernel use */
+#define AMDGPU_VA_RESERVED_CSA_SIZE		(2ULL << 20)
+#define AMDGPU_VA_RESERVED_CSA_START(adev)	(((adev)->vm_manager.max_pfn \
+						  << AMDGPU_GPU_PAGE_SHIFT)  \
+						 - AMDGPU_VA_RESERVED_CSA_SIZE)
+#define AMDGPU_VA_RESERVED_SEQ64_SIZE		(2ULL << 20)
+#define AMDGPU_VA_RESERVED_SEQ64_START(adev)	(AMDGPU_VA_RESERVED_CSA_START(adev) \
+						 - AMDGPU_VA_RESERVED_SEQ64_SIZE)
+#define AMDGPU_VA_RESERVED_TRAP_SIZE		(2ULL << 12)
+#define AMDGPU_VA_RESERVED_TRAP_START(adev)	(AMDGPU_VA_RESERVED_SEQ64_START(adev) \
+						 - AMDGPU_VA_RESERVED_TRAP_SIZE)
+#define AMDGPU_VA_RESERVED_BOTTOM		(1ULL << 16)
+#define AMDGPU_VA_RESERVED_TOP			(AMDGPU_VA_RESERVED_TRAP_SIZE + \
+						 AMDGPU_VA_RESERVED_SEQ64_SIZE + \
+						 AMDGPU_VA_RESERVED_CSA_SIZE)
 
 /* See vm_update_mode */
 #define AMDGPU_VM_USE_CPU_FOR_GFX (1 << 0)
@@ -145,9 +203,13 @@ struct amdgpu_vm_bo_base {
 	/* protected by bo being reserved */
 	struct amdgpu_vm_bo_base	*next;
 
-	/* protected by spinlock */
+	/* protected by vm status_lock */
 	struct list_head		vm_status;
 
+	/* if the bo is counted as shared in mem stats
+	 * protected by vm status_lock */
+	bool				shared;
+
 	/* protected by the BO being reserved */
 	bool				moved;
 };
@@ -174,10 +236,10 @@ struct amdgpu_vm_pte_funcs {
 };
 
 struct amdgpu_task_info {
-	char	process_name[TASK_COMM_LEN];
-	char	task_name[TASK_COMM_LEN];
-	pid_t	pid;
-	pid_t	tgid;
+	struct drm_wedge_task_info task;
+	char		process_name[TASK_COMM_LEN];
+	pid_t		tgid;
+	struct kref	refcount;
 };
 
 /**
@@ -227,15 +289,26 @@ struct amdgpu_vm_update_params {
 	unsigned int num_dw_left;
 
 	/**
-	 * @table_freed: return true if page table is freed when updating
+	 * @needs_flush: true whenever we need to invalidate the TLB
+	 */
+	bool needs_flush;
+
+	/**
+	 * @allow_override: true for memory that is not uncached: allows MTYPE
+	 * to be overridden for NUMA local memory.
 	 */
-	bool table_freed;
+	bool allow_override;
+
+	/**
+	 * @tlb_flush_waitlist: temporary storage for BOs until tlb_flush
+	 */
+	struct list_head tlb_flush_waitlist;
 };
 
 struct amdgpu_vm_update_funcs {
 	int (*map_table)(struct amdgpu_bo_vm *bo);
-	int (*prepare)(struct amdgpu_vm_update_params *p, struct dma_resv *resv,
-		       enum amdgpu_sync_mode sync_mode);
+	int (*prepare)(struct amdgpu_vm_update_params *p,
+		       struct amdgpu_sync *sync, u64 k_job_id);
 	int (*update)(struct amdgpu_vm_update_params *p,
 		      struct amdgpu_bo_vm *bo, uint64_t pe, uint64_t addr,
 		      unsigned count, uint32_t incr, uint64_t flags);
@@ -243,6 +316,22 @@ struct amdgpu_vm_update_funcs {
 		      struct dma_fence **fence);
 };
 
+struct amdgpu_vm_fault_info {
+	/* fault address */
+	uint64_t	addr;
+	/* fault status register */
+	uint32_t	status;
+	/* which vmhub? gfxhub, mmhub, etc. */
+	unsigned int	vmhub;
+};
+
+struct amdgpu_mem_stats {
+	struct drm_memory_stats drm;
+
+	/* buffers that requested this placement but are currently evicted */
+	uint64_t evicted;
+};
+
 struct amdgpu_vm {
 	/* tree of virtual addresses mapped */
 	struct rb_root_cached	va;
@@ -254,7 +343,20 @@ struct amdgpu_vm {
 	bool			evicting;
 	unsigned int		saved_flags;
 
-	/* BOs who needs a validation */
+	/* Lock to protect vm_bo add/del/move on all lists of vm */
+	spinlock_t		status_lock;
+
+	/* Memory statistics for this vm, protected by status_lock */
+	struct amdgpu_mem_stats stats[__AMDGPU_PL_NUM];
+
+	/*
+	 * The following lists contain amdgpu_vm_bo_base objects for either
+	 * PDs, PTs or per VM BOs. The state transits are:
+	 *
+	 * evicted -> relocated (PDs, PTs) or moved (per VM BOs) -> idle
+	 */
+
+	/* Per-VM and PT BOs who needs a validation */
 	struct list_head	evicted;
 
 	/* PT BOs which relocated and their parent need an update */
@@ -266,16 +368,29 @@ struct amdgpu_vm {
 	/* All BOs of this VM not currently in the state machine */
 	struct list_head	idle;
 
+	/*
+	 * The following lists contain amdgpu_vm_bo_base objects for BOs which
+	 * have their own dma_resv object and not depend on the root PD. Their
+	 * state transits are:
+	 *
+	 * evicted_user or invalidated -> done
+	 */
+
+	/* BOs for user mode queues that need a validation */
+	struct list_head	evicted_user;
+
 	/* regular invalidated BOs, but not yet updated in the PT */
 	struct list_head	invalidated;
-	spinlock_t		invalidated_lock;
-
-	/* BO mappings freed, but not yet updated in the PT */
-	struct list_head	freed;
 
 	/* BOs which are invalidated, has been updated in the PTs */
 	struct list_head        done;
 
+	/*
+	 * This list contains amdgpu_bo_va_mapping objects which have been freed
+	 * but not updated in the PTs
+	 */
+	struct list_head	freed;
+
 	/* contains the page directory */
 	struct amdgpu_vm_bo_base     root;
 	struct dma_fence	*last_update;
@@ -284,11 +399,19 @@ struct amdgpu_vm {
 	struct drm_sched_entity	immediate;
 	struct drm_sched_entity	delayed;
 
+	/* Last finished delayed update */
+	atomic64_t		tlb_seq;
+	struct dma_fence	*last_tlb_flush;
+	atomic64_t		kfd_last_flushed_seq;
+	uint64_t		tlb_fence_context;
+
+	/* How many times we had to re-generate the page tables */
+	uint64_t		generation;
+
 	/* Last unlocked submission to the scheduler entities */
 	struct dma_fence	*last_unlocked;
 
 	unsigned int		pasid;
-	/* dedicated to vm */
 	struct amdgpu_vmid	*reserved_vmid[AMDGPU_MAX_VMHUBS];
 
 	/* Flag to indicate if VM tables are updated by CPU or GPU (SDMA) */
@@ -297,9 +420,6 @@ struct amdgpu_vm {
 	/* Functions to use for VM table updates */
 	const struct amdgpu_vm_update_funcs	*update_funcs;
 
-	/* Flag to indicate ATS support from PTE for GFX9 */
-	bool			pte_support_ats;
-
 	/* Up to 128 pending retry page faults */
 	DECLARE_KFIFO(faults, u64, 128);
 
@@ -313,14 +433,18 @@ struct amdgpu_vm {
 	uint64_t		pd_phys_addr;
 
 	/* Some basic info about the task */
-	struct amdgpu_task_info task_info;
+	struct amdgpu_task_info *task_info;
 
 	/* Store positions of group of BOs */
 	struct ttm_lru_bulk_move lru_bulk_move;
-	/* mark whether can do the bulk move */
-	bool			bulk_moveable;
 	/* Flag to indicate if VM is used for compute */
 	bool			is_compute_context;
+
+	/* Memory partition number, -1 means any partition */
+	int8_t			mem_id;
+
+	/* cached fault info */
+	struct amdgpu_vm_fault_info fault_info;
 };
 
 struct amdgpu_vm_manager {
@@ -329,10 +453,6 @@ struct amdgpu_vm_manager {
 	unsigned int				first_kfd_vmid;
 	bool					concurrent_flush;
 
-	/* Handling of VM fences */
-	u64					fence_context;
-	unsigned				seqno[AMDGPU_MAX_RINGS];
-
 	uint64_t				max_pfn;
 	uint32_t				num_level;
 	uint32_t				block_size;
@@ -360,6 +480,8 @@ struct amdgpu_vm_manager {
 	 * look up VM of a page fault
 	 */
 	struct xarray				pasids;
+	/* Global registration of recent page fault information */
+	struct amdgpu_vm_fault_info	fault_info;
 };
 
 struct amdgpu_bo_va_mapping;
@@ -374,21 +496,20 @@ extern const struct amdgpu_vm_update_funcs amdgpu_vm_sdma_funcs;
 void amdgpu_vm_manager_init(struct amdgpu_device *adev);
 void amdgpu_vm_manager_fini(struct amdgpu_device *adev);
 
-int amdgpu_vm_set_pasid(struct amdgpu_device *adev, struct amdgpu_vm *vm,
-			u32 pasid);
-
 long amdgpu_vm_wait_idle(struct amdgpu_vm *vm, long timeout);
-int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm);
+int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm, int32_t xcp_id, uint32_t pasid);
 int amdgpu_vm_make_compute(struct amdgpu_device *adev, struct amdgpu_vm *vm);
-void amdgpu_vm_release_compute(struct amdgpu_device *adev, struct amdgpu_vm *vm);
 void amdgpu_vm_fini(struct amdgpu_device *adev, struct amdgpu_vm *vm);
-void amdgpu_vm_get_pd_bo(struct amdgpu_vm *vm,
-			 struct list_head *validated,
-			 struct amdgpu_bo_list_entry *entry);
+int amdgpu_vm_lock_pd(struct amdgpu_vm *vm, struct drm_exec *exec,
+		      unsigned int num_fences);
+int amdgpu_vm_lock_done_list(struct amdgpu_vm *vm, struct drm_exec *exec,
+			     unsigned int num_fences);
 bool amdgpu_vm_ready(struct amdgpu_vm *vm);
-int amdgpu_vm_validate_pt_bos(struct amdgpu_device *adev, struct amdgpu_vm *vm,
-			      int (*callback)(void *p, struct amdgpu_bo *bo),
-			      void *param);
+uint64_t amdgpu_vm_generation(struct amdgpu_device *adev, struct amdgpu_vm *vm);
+int amdgpu_vm_validate(struct amdgpu_device *adev, struct amdgpu_vm *vm,
+		       struct ww_acquire_ctx *ticket,
+		       int (*callback)(void *p, struct amdgpu_bo *bo),
+		       void *param);
 int amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_job *job, bool need_pipe_sync);
 int amdgpu_vm_update_pdes(struct amdgpu_device *adev,
 			  struct amdgpu_vm *vm, bool immediate);
@@ -396,22 +517,31 @@ int amdgpu_vm_clear_freed(struct amdgpu_device *adev,
 			  struct amdgpu_vm *vm,
 			  struct dma_fence **fence);
 int amdgpu_vm_handle_moved(struct amdgpu_device *adev,
-			   struct amdgpu_vm *vm);
-int amdgpu_vm_bo_update_mapping(struct amdgpu_device *adev,
-				struct amdgpu_device *bo_adev,
-				struct amdgpu_vm *vm, bool immediate,
-				bool unlocked, struct dma_resv *resv,
-				uint64_t start, uint64_t last,
-				uint64_t flags, uint64_t offset,
-				struct ttm_resource *res,
-				dma_addr_t *pages_addr,
-				struct dma_fence **fence, bool *free_table);
+			   struct amdgpu_vm *vm,
+			   struct ww_acquire_ctx *ticket);
+int amdgpu_vm_flush_compute_tlb(struct amdgpu_device *adev,
+				struct amdgpu_vm *vm,
+				uint32_t flush_type,
+				uint32_t xcc_mask);
+void amdgpu_vm_bo_base_init(struct amdgpu_vm_bo_base *base,
+			    struct amdgpu_vm *vm, struct amdgpu_bo *bo);
+int amdgpu_vm_update_range(struct amdgpu_device *adev, struct amdgpu_vm *vm,
+			   bool immediate, bool unlocked, bool flush_tlb,
+			   bool allow_override, struct amdgpu_sync *sync,
+			   uint64_t start, uint64_t last, uint64_t flags,
+			   uint64_t offset, uint64_t vram_base,
+			   struct ttm_resource *res, dma_addr_t *pages_addr,
+			   struct dma_fence **fence);
 int amdgpu_vm_bo_update(struct amdgpu_device *adev,
 			struct amdgpu_bo_va *bo_va,
-			bool clear, bool *table_freed);
+			bool clear);
 bool amdgpu_vm_evictable(struct amdgpu_bo *bo);
-void amdgpu_vm_bo_invalidate(struct amdgpu_device *adev,
-			     struct amdgpu_bo *bo, bool evicted);
+void amdgpu_vm_bo_invalidate(struct amdgpu_bo *bo, bool evicted);
+void amdgpu_vm_update_stats(struct amdgpu_vm_bo_base *base,
+			    struct ttm_resource *new_res, int sign);
+void amdgpu_vm_bo_update_shared(struct amdgpu_bo *bo);
+void amdgpu_vm_bo_move(struct amdgpu_bo *bo, struct ttm_resource *new_mem,
+		       bool evicted);
 uint64_t amdgpu_vm_map_gart(const dma_addr_t *pages_addr, uint64_t addr);
 struct amdgpu_bo_va *amdgpu_vm_bo_find(struct amdgpu_vm *vm,
 				       struct amdgpu_bo *bo);
@@ -421,11 +551,11 @@ struct amdgpu_bo_va *amdgpu_vm_bo_add(struct amdgpu_device *adev,
 int amdgpu_vm_bo_map(struct amdgpu_device *adev,
 		     struct amdgpu_bo_va *bo_va,
 		     uint64_t addr, uint64_t offset,
-		     uint64_t size, uint64_t flags);
+		     uint64_t size, uint32_t flags);
 int amdgpu_vm_bo_replace_map(struct amdgpu_device *adev,
 			     struct amdgpu_bo_va *bo_va,
 			     uint64_t addr, uint64_t offset,
-			     uint64_t size, uint64_t flags);
+			     uint64_t size, uint32_t flags);
 int amdgpu_vm_bo_unmap(struct amdgpu_device *adev,
 		       struct amdgpu_bo_va *bo_va,
 		       uint64_t addr);
@@ -435,7 +565,7 @@ int amdgpu_vm_bo_clear_mappings(struct amdgpu_device *adev,
 struct amdgpu_bo_va_mapping *amdgpu_vm_bo_lookup_mapping(struct amdgpu_vm *vm,
 							 uint64_t addr);
 void amdgpu_vm_bo_trace_cs(struct amdgpu_vm *vm, struct ww_acquire_ctx *ticket);
-void amdgpu_vm_bo_rmv(struct amdgpu_device *adev,
+void amdgpu_vm_bo_del(struct amdgpu_device *adev,
 		      struct amdgpu_bo_va *bo_va);
 void amdgpu_vm_adjust_size(struct amdgpu_device *adev, uint32_t min_vm_size,
 			   uint32_t fragment_size_default, unsigned max_level,
@@ -445,21 +575,117 @@ bool amdgpu_vm_need_pipeline_sync(struct amdgpu_ring *ring,
 				  struct amdgpu_job *job);
 void amdgpu_vm_check_compute_bug(struct amdgpu_device *adev);
 
-void amdgpu_vm_get_task_info(struct amdgpu_device *adev, u32 pasid,
-			     struct amdgpu_task_info *task_info);
+struct amdgpu_task_info *
+amdgpu_vm_get_task_info_pasid(struct amdgpu_device *adev, u32 pasid);
+
+struct amdgpu_task_info *
+amdgpu_vm_get_task_info_vm(struct amdgpu_vm *vm);
+
+void amdgpu_vm_put_task_info(struct amdgpu_task_info *task_info);
+
 bool amdgpu_vm_handle_fault(struct amdgpu_device *adev, u32 pasid,
-			    uint64_t addr, bool write_fault);
+			    u32 vmid, u32 node_id, uint64_t addr, uint64_t ts,
+			    bool write_fault);
 
 void amdgpu_vm_set_task_info(struct amdgpu_vm *vm);
 
 void amdgpu_vm_move_to_lru_tail(struct amdgpu_device *adev,
 				struct amdgpu_vm *vm);
-void amdgpu_vm_del_from_lru_notify(struct ttm_buffer_object *bo);
-void amdgpu_vm_get_memory(struct amdgpu_vm *vm, uint64_t *vram_mem,
-				uint64_t *gtt_mem, uint64_t *cpu_mem);
+void amdgpu_vm_get_memory(struct amdgpu_vm *vm,
+			  struct amdgpu_mem_stats stats[__AMDGPU_PL_NUM]);
+
+int amdgpu_vm_pt_clear(struct amdgpu_device *adev, struct amdgpu_vm *vm,
+		       struct amdgpu_bo_vm *vmbo, bool immediate);
+int amdgpu_vm_pt_create(struct amdgpu_device *adev, struct amdgpu_vm *vm,
+			int level, bool immediate, struct amdgpu_bo_vm **vmbo,
+			int32_t xcp_id);
+void amdgpu_vm_pt_free_root(struct amdgpu_device *adev, struct amdgpu_vm *vm);
+
+int amdgpu_vm_pde_update(struct amdgpu_vm_update_params *params,
+			 struct amdgpu_vm_bo_base *entry);
+int amdgpu_vm_ptes_update(struct amdgpu_vm_update_params *params,
+			  uint64_t start, uint64_t end,
+			  uint64_t dst, uint64_t flags);
+void amdgpu_vm_pt_free_work(struct work_struct *work);
+void amdgpu_vm_pt_free_list(struct amdgpu_device *adev,
+			    struct amdgpu_vm_update_params *params);
 
 #if defined(CONFIG_DEBUG_FS)
 void amdgpu_debugfs_vm_bo_info(struct amdgpu_vm *vm, struct seq_file *m);
 #endif
 
+int amdgpu_vm_pt_map_tables(struct amdgpu_device *adev, struct amdgpu_vm *vm);
+
+bool amdgpu_vm_is_bo_always_valid(struct amdgpu_vm *vm, struct amdgpu_bo *bo);
+
+/**
+ * amdgpu_vm_tlb_seq - return tlb flush sequence number
+ * @vm: the amdgpu_vm structure to query
+ *
+ * Returns the tlb flush sequence number which indicates that the VM TLBs needs
+ * to be invalidated whenever the sequence number change.
+ */
+static inline uint64_t amdgpu_vm_tlb_seq(struct amdgpu_vm *vm)
+{
+	unsigned long flags;
+	spinlock_t *lock;
+
+	/*
+	 * Workaround to stop racing between the fence signaling and handling
+	 * the cb. The lock is static after initially setting it up, just make
+	 * sure that the dma_fence structure isn't freed up.
+	 */
+	rcu_read_lock();
+	lock = vm->last_tlb_flush->lock;
+	rcu_read_unlock();
+
+	spin_lock_irqsave(lock, flags);
+	spin_unlock_irqrestore(lock, flags);
+
+	return atomic64_read(&vm->tlb_seq);
+}
+
+/*
+ * vm eviction_lock can be taken in MMU notifiers. Make sure no reclaim-FS
+ * happens while holding this lock anywhere to prevent deadlocks when
+ * an MMU notifier runs in reclaim-FS context.
+ */
+static inline void amdgpu_vm_eviction_lock(struct amdgpu_vm *vm)
+{
+	mutex_lock(&vm->eviction_lock);
+	vm->saved_flags = memalloc_noreclaim_save();
+}
+
+static inline bool amdgpu_vm_eviction_trylock(struct amdgpu_vm *vm)
+{
+	if (mutex_trylock(&vm->eviction_lock)) {
+		vm->saved_flags = memalloc_noreclaim_save();
+		return true;
+	}
+	return false;
+}
+
+static inline void amdgpu_vm_eviction_unlock(struct amdgpu_vm *vm)
+{
+	memalloc_noreclaim_restore(vm->saved_flags);
+	mutex_unlock(&vm->eviction_lock);
+}
+
+void amdgpu_vm_update_fault_cache(struct amdgpu_device *adev,
+				  unsigned int pasid,
+				  uint64_t addr,
+				  uint32_t status,
+				  unsigned int vmhub);
+void amdgpu_vm_tlb_fence_create(struct amdgpu_device *adev,
+				 struct amdgpu_vm *vm,
+				 struct dma_fence **fence);
+
+void amdgpu_vm_print_task_info(struct amdgpu_device *adev,
+			       struct amdgpu_task_info *task_info);
+
+#define amdgpu_vm_bo_va_for_each_valid_mapping(bo_va, mapping) \
+		list_for_each_entry(mapping, &(bo_va)->valids, list)
+#define amdgpu_vm_bo_va_for_each_invalid_mapping(bo_va, mapping) \
+		list_for_each_entry(mapping, &(bo_va)->invalids, list)
+
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_cpu.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_cpu.c
index e3fbf0f10add..22e2e5b47341 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_cpu.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_cpu.c
@@ -31,6 +31,7 @@
  */
 static int amdgpu_vm_cpu_map_table(struct amdgpu_bo_vm *table)
 {
+	table->bo.flags |= AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED;
 	return amdgpu_bo_kmap(&table->bo, NULL);
 }
 
@@ -38,20 +39,20 @@ static int amdgpu_vm_cpu_map_table(struct amdgpu_bo_vm *table)
  * amdgpu_vm_cpu_prepare - prepare page table update with the CPU
  *
  * @p: see amdgpu_vm_update_params definition
- * @resv: reservation object with embedded fence
- * @sync_mode: synchronization mode
+ * @sync: sync obj with fences to wait on
+ * @k_job_id: the id for tracing/debug purposes
  *
  * Returns:
  * Negativ errno, 0 for success.
  */
 static int amdgpu_vm_cpu_prepare(struct amdgpu_vm_update_params *p,
-				 struct dma_resv *resv,
-				 enum amdgpu_sync_mode sync_mode)
+				 struct amdgpu_sync *sync,
+				 u64 k_job_id)
 {
-	if (!resv)
+	if (!sync)
 		return 0;
 
-	return amdgpu_bo_sync_wait_resv(p->adev, resv, sync_mode, p->vm, true);
+	return amdgpu_sync_wait(sync, true);
 }
 
 /**
@@ -74,13 +75,12 @@ static int amdgpu_vm_cpu_update(struct amdgpu_vm_update_params *p,
 {
 	unsigned int i;
 	uint64_t value;
-	int r;
+	long r;
 
-	if (vmbo->bo.tbo.moving) {
-		r = dma_fence_wait(vmbo->bo.tbo.moving, true);
-		if (r)
-			return r;
-	}
+	r = dma_resv_wait_timeout(vmbo->bo.tbo.base.resv, DMA_RESV_USAGE_KERNEL,
+				  true, MAX_SCHEDULE_TIMEOUT);
+	if (r < 0)
+		return r;
 
 	pe += (unsigned long)amdgpu_bo_kptr(&vmbo->bo);
 
@@ -108,7 +108,9 @@ static int amdgpu_vm_cpu_update(struct amdgpu_vm_update_params *p,
 static int amdgpu_vm_cpu_commit(struct amdgpu_vm_update_params *p,
 				struct dma_fence **fence)
 {
-	/* Flush HDP */
+	if (p->needs_flush)
+		atomic64_inc(&p->vm->tlb_seq);
+
 	mb();
 	amdgpu_device_flush_hdp(p->adev, NULL);
 	return 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_pt.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_pt.c
new file mode 100644
index 000000000000..f794fb1cc06e
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_pt.c
@@ -0,0 +1,976 @@
+// SPDX-License-Identifier: GPL-2.0 OR MIT
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#include <drm/drm_drv.h>
+
+#include "amdgpu.h"
+#include "amdgpu_trace.h"
+#include "amdgpu_vm.h"
+#include "amdgpu_job.h"
+
+/*
+ * amdgpu_vm_pt_cursor - state for for_each_amdgpu_vm_pt
+ */
+struct amdgpu_vm_pt_cursor {
+	uint64_t pfn;
+	struct amdgpu_vm_bo_base *parent;
+	struct amdgpu_vm_bo_base *entry;
+	unsigned int level;
+};
+
+/**
+ * amdgpu_vm_pt_level_shift - return the addr shift for each level
+ *
+ * @adev: amdgpu_device pointer
+ * @level: VMPT level
+ *
+ * Returns:
+ * The number of bits the pfn needs to be right shifted for a level.
+ */
+static unsigned int amdgpu_vm_pt_level_shift(struct amdgpu_device *adev,
+					     unsigned int level)
+{
+	switch (level) {
+	case AMDGPU_VM_PDB2:
+	case AMDGPU_VM_PDB1:
+	case AMDGPU_VM_PDB0:
+		return 9 * (AMDGPU_VM_PDB0 - level) +
+			adev->vm_manager.block_size;
+	case AMDGPU_VM_PTB:
+		return 0;
+	default:
+		return ~0;
+	}
+}
+
+/**
+ * amdgpu_vm_pt_num_entries - return the number of entries in a PD/PT
+ *
+ * @adev: amdgpu_device pointer
+ * @level: VMPT level
+ *
+ * Returns:
+ * The number of entries in a page directory or page table.
+ */
+static unsigned int amdgpu_vm_pt_num_entries(struct amdgpu_device *adev,
+					     unsigned int level)
+{
+	unsigned int shift;
+
+	shift = amdgpu_vm_pt_level_shift(adev, adev->vm_manager.root_level);
+	if (level == adev->vm_manager.root_level)
+		/* For the root directory */
+		return round_up(adev->vm_manager.max_pfn, 1ULL << shift)
+			>> shift;
+	else if (level != AMDGPU_VM_PTB)
+		/* Everything in between */
+		return 512;
+
+	/* For the page tables on the leaves */
+	return AMDGPU_VM_PTE_COUNT(adev);
+}
+
+/**
+ * amdgpu_vm_pt_entries_mask - the mask to get the entry number of a PD/PT
+ *
+ * @adev: amdgpu_device pointer
+ * @level: VMPT level
+ *
+ * Returns:
+ * The mask to extract the entry number of a PD/PT from an address.
+ */
+static uint32_t amdgpu_vm_pt_entries_mask(struct amdgpu_device *adev,
+					  unsigned int level)
+{
+	if (level <= adev->vm_manager.root_level)
+		return 0xffffffff;
+	else if (level != AMDGPU_VM_PTB)
+		return 0x1ff;
+	else
+		return AMDGPU_VM_PTE_COUNT(adev) - 1;
+}
+
+/**
+ * amdgpu_vm_pt_size - returns the size of the page table in bytes
+ *
+ * @adev: amdgpu_device pointer
+ * @level: VMPT level
+ *
+ * Returns:
+ * The size of the BO for a page directory or page table in bytes.
+ */
+static unsigned int amdgpu_vm_pt_size(struct amdgpu_device *adev,
+				      unsigned int level)
+{
+	return AMDGPU_GPU_PAGE_ALIGN(amdgpu_vm_pt_num_entries(adev, level) * 8);
+}
+
+/**
+ * amdgpu_vm_pt_parent - get the parent page directory
+ *
+ * @pt: child page table
+ *
+ * Helper to get the parent entry for the child page table. NULL if we are at
+ * the root page directory.
+ */
+static struct amdgpu_vm_bo_base *
+amdgpu_vm_pt_parent(struct amdgpu_vm_bo_base *pt)
+{
+	struct amdgpu_bo *parent = pt->bo->parent;
+
+	if (!parent)
+		return NULL;
+
+	return parent->vm_bo;
+}
+
+/**
+ * amdgpu_vm_pt_start - start PD/PT walk
+ *
+ * @adev: amdgpu_device pointer
+ * @vm: amdgpu_vm structure
+ * @start: start address of the walk
+ * @cursor: state to initialize
+ *
+ * Initialize a amdgpu_vm_pt_cursor to start a walk.
+ */
+static void amdgpu_vm_pt_start(struct amdgpu_device *adev,
+			       struct amdgpu_vm *vm, uint64_t start,
+			       struct amdgpu_vm_pt_cursor *cursor)
+{
+	cursor->pfn = start;
+	cursor->parent = NULL;
+	cursor->entry = &vm->root;
+	cursor->level = adev->vm_manager.root_level;
+}
+
+/**
+ * amdgpu_vm_pt_descendant - go to child node
+ *
+ * @adev: amdgpu_device pointer
+ * @cursor: current state
+ *
+ * Walk to the child node of the current node.
+ * Returns:
+ * True if the walk was possible, false otherwise.
+ */
+static bool amdgpu_vm_pt_descendant(struct amdgpu_device *adev,
+				    struct amdgpu_vm_pt_cursor *cursor)
+{
+	unsigned int mask, shift, idx;
+
+	if ((cursor->level == AMDGPU_VM_PTB) || !cursor->entry ||
+	    !cursor->entry->bo)
+		return false;
+
+	mask = amdgpu_vm_pt_entries_mask(adev, cursor->level);
+	shift = amdgpu_vm_pt_level_shift(adev, cursor->level);
+
+	++cursor->level;
+	idx = (cursor->pfn >> shift) & mask;
+	cursor->parent = cursor->entry;
+	cursor->entry = &to_amdgpu_bo_vm(cursor->entry->bo)->entries[idx];
+	return true;
+}
+
+/**
+ * amdgpu_vm_pt_sibling - go to sibling node
+ *
+ * @adev: amdgpu_device pointer
+ * @cursor: current state
+ *
+ * Walk to the sibling node of the current node.
+ * Returns:
+ * True if the walk was possible, false otherwise.
+ */
+static bool amdgpu_vm_pt_sibling(struct amdgpu_device *adev,
+				 struct amdgpu_vm_pt_cursor *cursor)
+{
+
+	unsigned int shift, num_entries;
+	struct amdgpu_bo_vm *parent;
+
+	/* Root doesn't have a sibling */
+	if (!cursor->parent)
+		return false;
+
+	/* Go to our parents and see if we got a sibling */
+	shift = amdgpu_vm_pt_level_shift(adev, cursor->level - 1);
+	num_entries = amdgpu_vm_pt_num_entries(adev, cursor->level - 1);
+	parent = to_amdgpu_bo_vm(cursor->parent->bo);
+
+	if (cursor->entry == &parent->entries[num_entries - 1])
+		return false;
+
+	cursor->pfn += 1ULL << shift;
+	cursor->pfn &= ~((1ULL << shift) - 1);
+	++cursor->entry;
+	return true;
+}
+
+/**
+ * amdgpu_vm_pt_ancestor - go to parent node
+ *
+ * @cursor: current state
+ *
+ * Walk to the parent node of the current node.
+ * Returns:
+ * True if the walk was possible, false otherwise.
+ */
+static bool amdgpu_vm_pt_ancestor(struct amdgpu_vm_pt_cursor *cursor)
+{
+	if (!cursor->parent)
+		return false;
+
+	--cursor->level;
+	cursor->entry = cursor->parent;
+	cursor->parent = amdgpu_vm_pt_parent(cursor->parent);
+	return true;
+}
+
+/**
+ * amdgpu_vm_pt_next - get next PD/PT in hieratchy
+ *
+ * @adev: amdgpu_device pointer
+ * @cursor: current state
+ *
+ * Walk the PD/PT tree to the next node.
+ */
+static void amdgpu_vm_pt_next(struct amdgpu_device *adev,
+			      struct amdgpu_vm_pt_cursor *cursor)
+{
+	/* First try a newborn child */
+	if (amdgpu_vm_pt_descendant(adev, cursor))
+		return;
+
+	/* If that didn't worked try to find a sibling */
+	while (!amdgpu_vm_pt_sibling(adev, cursor)) {
+		/* No sibling, go to our parents and grandparents */
+		if (!amdgpu_vm_pt_ancestor(cursor)) {
+			cursor->pfn = ~0ll;
+			return;
+		}
+	}
+}
+
+/**
+ * amdgpu_vm_pt_first_dfs - start a deep first search
+ *
+ * @adev: amdgpu_device structure
+ * @vm: amdgpu_vm structure
+ * @start: optional cursor to start with
+ * @cursor: state to initialize
+ *
+ * Starts a deep first traversal of the PD/PT tree.
+ */
+static void amdgpu_vm_pt_first_dfs(struct amdgpu_device *adev,
+				   struct amdgpu_vm *vm,
+				   struct amdgpu_vm_pt_cursor *start,
+				   struct amdgpu_vm_pt_cursor *cursor)
+{
+	if (start)
+		*cursor = *start;
+	else
+		amdgpu_vm_pt_start(adev, vm, 0, cursor);
+
+	while (amdgpu_vm_pt_descendant(adev, cursor))
+		;
+}
+
+/**
+ * amdgpu_vm_pt_continue_dfs - check if the deep first search should continue
+ *
+ * @start: starting point for the search
+ * @entry: current entry
+ *
+ * Returns:
+ * True when the search should continue, false otherwise.
+ */
+static bool amdgpu_vm_pt_continue_dfs(struct amdgpu_vm_pt_cursor *start,
+				      struct amdgpu_vm_bo_base *entry)
+{
+	return entry && (!start || entry != start->entry);
+}
+
+/**
+ * amdgpu_vm_pt_next_dfs - get the next node for a deep first search
+ *
+ * @adev: amdgpu_device structure
+ * @cursor: current state
+ *
+ * Move the cursor to the next node in a deep first search.
+ */
+static void amdgpu_vm_pt_next_dfs(struct amdgpu_device *adev,
+				  struct amdgpu_vm_pt_cursor *cursor)
+{
+	if (!cursor->entry)
+		return;
+
+	if (!cursor->parent)
+		cursor->entry = NULL;
+	else if (amdgpu_vm_pt_sibling(adev, cursor))
+		while (amdgpu_vm_pt_descendant(adev, cursor))
+			;
+	else
+		amdgpu_vm_pt_ancestor(cursor);
+}
+
+/*
+ * for_each_amdgpu_vm_pt_dfs_safe - safe deep first search of all PDs/PTs
+ */
+#define for_each_amdgpu_vm_pt_dfs_safe(adev, vm, start, cursor, entry)		\
+	for (amdgpu_vm_pt_first_dfs((adev), (vm), (start), &(cursor)),		\
+	     (entry) = (cursor).entry, amdgpu_vm_pt_next_dfs((adev), &(cursor));\
+	     amdgpu_vm_pt_continue_dfs((start), (entry));			\
+	     (entry) = (cursor).entry, amdgpu_vm_pt_next_dfs((adev), &(cursor)))
+
+/**
+ * amdgpu_vm_pt_clear - initially clear the PDs/PTs
+ *
+ * @adev: amdgpu_device pointer
+ * @vm: VM to clear BO from
+ * @vmbo: BO to clear
+ * @immediate: use an immediate update
+ *
+ * Root PD needs to be reserved when calling this.
+ *
+ * Returns:
+ * 0 on success, errno otherwise.
+ */
+int amdgpu_vm_pt_clear(struct amdgpu_device *adev, struct amdgpu_vm *vm,
+		       struct amdgpu_bo_vm *vmbo, bool immediate)
+{
+	unsigned int level = adev->vm_manager.root_level;
+	struct ttm_operation_ctx ctx = { true, false };
+	struct amdgpu_vm_update_params params;
+	struct amdgpu_bo *ancestor = &vmbo->bo;
+	unsigned int entries;
+	struct amdgpu_bo *bo = &vmbo->bo;
+	uint64_t addr;
+	int r, idx;
+
+	/* Figure out our place in the hierarchy */
+	if (ancestor->parent) {
+		++level;
+		while (ancestor->parent->parent) {
+			++level;
+			ancestor = ancestor->parent;
+		}
+	}
+
+	entries = amdgpu_bo_size(bo) / 8;
+
+	r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
+	if (r)
+		return r;
+
+	if (!drm_dev_enter(adev_to_drm(adev), &idx))
+		return -ENODEV;
+
+	r = vm->update_funcs->map_table(vmbo);
+	if (r)
+		goto exit;
+
+	memset(&params, 0, sizeof(params));
+	params.adev = adev;
+	params.vm = vm;
+	params.immediate = immediate;
+
+	r = vm->update_funcs->prepare(&params, NULL,
+				      AMDGPU_KERNEL_JOB_ID_VM_PT_CLEAR);
+	if (r)
+		goto exit;
+
+	addr = 0;
+
+	uint64_t value = 0, flags = 0;
+	if (adev->asic_type >= CHIP_VEGA10) {
+		if (level != AMDGPU_VM_PTB) {
+			/* Handle leaf PDEs as PTEs */
+			flags |= AMDGPU_PDE_PTE_FLAG(adev);
+			amdgpu_gmc_get_vm_pde(adev, level,
+					      &value, &flags);
+		} else {
+			/* Workaround for fault priority problem on GMC9 */
+			flags = AMDGPU_PTE_EXECUTABLE;
+		}
+	}
+
+	r = vm->update_funcs->update(&params, vmbo, addr, 0, entries,
+				     value, flags);
+	if (r)
+		goto exit;
+
+	r = vm->update_funcs->commit(&params, NULL);
+exit:
+	drm_dev_exit(idx);
+	return r;
+}
+
+/**
+ * amdgpu_vm_pt_create - create bo for PD/PT
+ *
+ * @adev: amdgpu_device pointer
+ * @vm: requesting vm
+ * @level: the page table level
+ * @immediate: use a immediate update
+ * @vmbo: pointer to the buffer object pointer
+ * @xcp_id: GPU partition id
+ */
+int amdgpu_vm_pt_create(struct amdgpu_device *adev, struct amdgpu_vm *vm,
+			int level, bool immediate, struct amdgpu_bo_vm **vmbo,
+			int32_t xcp_id)
+{
+	struct amdgpu_bo_param bp;
+	unsigned int num_entries;
+
+	memset(&bp, 0, sizeof(bp));
+
+	bp.size = amdgpu_vm_pt_size(adev, level);
+	bp.byte_align = AMDGPU_GPU_PAGE_SIZE;
+
+	if (!adev->gmc.is_app_apu)
+		bp.domain = AMDGPU_GEM_DOMAIN_VRAM;
+	else
+		bp.domain = AMDGPU_GEM_DOMAIN_GTT;
+
+	bp.domain = amdgpu_bo_get_preferred_domain(adev, bp.domain);
+	bp.flags = AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS |
+		AMDGPU_GEM_CREATE_CPU_GTT_USWC;
+
+	if (level < AMDGPU_VM_PTB)
+		num_entries = amdgpu_vm_pt_num_entries(adev, level);
+	else
+		num_entries = 0;
+
+	bp.bo_ptr_size = struct_size((*vmbo), entries, num_entries);
+
+	if (vm->use_cpu_for_update)
+		bp.flags |= AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED;
+
+	bp.type = ttm_bo_type_kernel;
+	bp.no_wait_gpu = immediate;
+	bp.xcp_id_plus1 = xcp_id + 1;
+
+	if (vm->root.bo)
+		bp.resv = vm->root.bo->tbo.base.resv;
+
+	return amdgpu_bo_create_vm(adev, &bp, vmbo);
+}
+
+/**
+ * amdgpu_vm_pt_alloc - Allocate a specific page table
+ *
+ * @adev: amdgpu_device pointer
+ * @vm: VM to allocate page tables for
+ * @cursor: Which page table to allocate
+ * @immediate: use an immediate update
+ *
+ * Make sure a specific page table or directory is allocated.
+ *
+ * Returns:
+ * 1 if page table needed to be allocated, 0 if page table was already
+ * allocated, negative errno if an error occurred.
+ */
+static int amdgpu_vm_pt_alloc(struct amdgpu_device *adev,
+			      struct amdgpu_vm *vm,
+			      struct amdgpu_vm_pt_cursor *cursor,
+			      bool immediate)
+{
+	struct amdgpu_vm_bo_base *entry = cursor->entry;
+	struct amdgpu_bo *pt_bo;
+	struct amdgpu_bo_vm *pt;
+	int r;
+
+	if (entry->bo)
+		return 0;
+
+	amdgpu_vm_eviction_unlock(vm);
+	r = amdgpu_vm_pt_create(adev, vm, cursor->level, immediate, &pt,
+				vm->root.bo->xcp_id);
+	amdgpu_vm_eviction_lock(vm);
+	if (r)
+		return r;
+
+	/* Keep a reference to the root directory to avoid
+	 * freeing them up in the wrong order.
+	 */
+	pt_bo = &pt->bo;
+	pt_bo->parent = amdgpu_bo_ref(cursor->parent->bo);
+	amdgpu_vm_bo_base_init(entry, vm, pt_bo);
+	r = amdgpu_vm_pt_clear(adev, vm, pt, immediate);
+	if (r)
+		goto error_free_pt;
+
+	return 0;
+
+error_free_pt:
+	amdgpu_bo_unref(&pt_bo);
+	return r;
+}
+
+/**
+ * amdgpu_vm_pt_free - free one PD/PT
+ *
+ * @entry: PDE to free
+ */
+static void amdgpu_vm_pt_free(struct amdgpu_vm_bo_base *entry)
+{
+	if (!entry->bo)
+		return;
+
+	amdgpu_vm_update_stats(entry, entry->bo->tbo.resource, -1);
+	entry->bo->vm_bo = NULL;
+	ttm_bo_set_bulk_move(&entry->bo->tbo, NULL);
+
+	spin_lock(&entry->vm->status_lock);
+	list_del(&entry->vm_status);
+	spin_unlock(&entry->vm->status_lock);
+	amdgpu_bo_unref(&entry->bo);
+}
+
+/**
+ * amdgpu_vm_pt_free_list - free PD/PT levels
+ *
+ * @adev: amdgpu device structure
+ * @params: see amdgpu_vm_update_params definition
+ *
+ * Free the page directory objects saved in the flush list
+ */
+void amdgpu_vm_pt_free_list(struct amdgpu_device *adev,
+			    struct amdgpu_vm_update_params *params)
+{
+	struct amdgpu_vm_bo_base *entry, *next;
+	bool unlocked = params->unlocked;
+
+	if (list_empty(&params->tlb_flush_waitlist))
+		return;
+
+	/*
+	 * unlocked unmap clear page table leaves, warning to free the page entry.
+	 */
+	WARN_ON(unlocked);
+
+	list_for_each_entry_safe(entry, next, &params->tlb_flush_waitlist, vm_status)
+		amdgpu_vm_pt_free(entry);
+}
+
+/**
+ * amdgpu_vm_pt_add_list - add PD/PT level to the flush list
+ *
+ * @params: parameters for the update
+ * @cursor: first PT entry to start DF search from, non NULL
+ *
+ * This list will be freed after TLB flush.
+ */
+static void amdgpu_vm_pt_add_list(struct amdgpu_vm_update_params *params,
+				  struct amdgpu_vm_pt_cursor *cursor)
+{
+	struct amdgpu_vm_pt_cursor seek;
+	struct amdgpu_vm_bo_base *entry;
+
+	spin_lock(&params->vm->status_lock);
+	for_each_amdgpu_vm_pt_dfs_safe(params->adev, params->vm, cursor, seek, entry) {
+		if (entry && entry->bo)
+			list_move(&entry->vm_status, &params->tlb_flush_waitlist);
+	}
+
+	/* enter start node now */
+	list_move(&cursor->entry->vm_status, &params->tlb_flush_waitlist);
+	spin_unlock(&params->vm->status_lock);
+}
+
+/**
+ * amdgpu_vm_pt_free_root - free root PD
+ * @adev: amdgpu device structure
+ * @vm: amdgpu vm structure
+ *
+ * Free the root page directory and everything below it.
+ */
+void amdgpu_vm_pt_free_root(struct amdgpu_device *adev, struct amdgpu_vm *vm)
+{
+	struct amdgpu_vm_pt_cursor cursor;
+	struct amdgpu_vm_bo_base *entry;
+
+	for_each_amdgpu_vm_pt_dfs_safe(adev, vm, NULL, cursor, entry) {
+		if (entry)
+			amdgpu_vm_pt_free(entry);
+	}
+}
+
+/**
+ * amdgpu_vm_pde_update - update a single level in the hierarchy
+ *
+ * @params: parameters for the update
+ * @entry: entry to update
+ *
+ * Makes sure the requested entry in parent is up to date.
+ */
+int amdgpu_vm_pde_update(struct amdgpu_vm_update_params *params,
+			 struct amdgpu_vm_bo_base *entry)
+{
+	struct amdgpu_vm_bo_base *parent = amdgpu_vm_pt_parent(entry);
+	struct amdgpu_bo *bo, *pbo;
+	struct amdgpu_vm *vm = params->vm;
+	uint64_t pde, pt, flags;
+	unsigned int level;
+
+	if (WARN_ON(!parent))
+		return -EINVAL;
+
+	bo = parent->bo;
+	for (level = 0, pbo = bo->parent; pbo; ++level)
+		pbo = pbo->parent;
+
+	level += params->adev->vm_manager.root_level;
+	amdgpu_gmc_get_pde_for_bo(entry->bo, level, &pt, &flags);
+	pde = (entry - to_amdgpu_bo_vm(parent->bo)->entries) * 8;
+	return vm->update_funcs->update(params, to_amdgpu_bo_vm(bo), pde, pt,
+					1, 0, flags);
+}
+
+/**
+ * amdgpu_vm_pte_update_noretry_flags - Update PTE no-retry flags
+ *
+ * @adev: amdgpu_device pointer
+ * @flags: pointer to PTE flags
+ *
+ * Update PTE no-retry flags when TF is enabled.
+ */
+static void amdgpu_vm_pte_update_noretry_flags(struct amdgpu_device *adev,
+						uint64_t *flags)
+{
+	/*
+	 * Update no-retry flags with the corresponding TF
+	 * no-retry combination.
+	 */
+	if ((*flags & AMDGPU_VM_NORETRY_FLAGS) == AMDGPU_VM_NORETRY_FLAGS) {
+		*flags &= ~AMDGPU_VM_NORETRY_FLAGS;
+		*flags |= adev->gmc.noretry_flags;
+	}
+}
+
+/*
+ * amdgpu_vm_pte_update_flags - figure out flags for PTE updates
+ *
+ * Make sure to set the right flags for the PTEs at the desired level.
+ */
+static void amdgpu_vm_pte_update_flags(struct amdgpu_vm_update_params *params,
+				       struct amdgpu_bo_vm *pt,
+				       unsigned int level,
+				       uint64_t pe, uint64_t addr,
+				       unsigned int count, uint32_t incr,
+				       uint64_t flags)
+{
+	struct amdgpu_device *adev = params->adev;
+
+	if (level != AMDGPU_VM_PTB) {
+		flags |= AMDGPU_PDE_PTE_FLAG(params->adev);
+		amdgpu_gmc_get_vm_pde(adev, level, &addr, &flags);
+
+	} else if (adev->asic_type >= CHIP_VEGA10 &&
+		   !(flags & AMDGPU_PTE_VALID) &&
+		   !(flags & AMDGPU_PTE_PRT_FLAG(params->adev))) {
+
+		/* Workaround for fault priority problem on GMC9 */
+		flags |= AMDGPU_PTE_EXECUTABLE;
+	}
+
+	/*
+	 * Update no-retry flags to use the no-retry flag combination
+	 * with TF enabled. The AMDGPU_VM_NORETRY_FLAGS flag combination
+	 * does not work when TF is enabled. So, replace them with
+	 * AMDGPU_VM_NORETRY_FLAGS_TF flag combination which works for
+	 * all cases.
+	 */
+	if (level == AMDGPU_VM_PTB)
+		amdgpu_vm_pte_update_noretry_flags(adev, &flags);
+
+	/* APUs mapping system memory may need different MTYPEs on different
+	 * NUMA nodes. Only do this for contiguous ranges that can be assumed
+	 * to be on the same NUMA node.
+	 */
+	if ((flags & AMDGPU_PTE_SYSTEM) && (adev->flags & AMD_IS_APU) &&
+	    adev->gmc.gmc_funcs->override_vm_pte_flags &&
+	    num_possible_nodes() > 1 && !params->pages_addr && params->allow_override)
+		amdgpu_gmc_override_vm_pte_flags(adev, params->vm, addr, &flags);
+
+	params->vm->update_funcs->update(params, pt, pe, addr, count, incr,
+					 flags);
+}
+
+/**
+ * amdgpu_vm_pte_fragment - get fragment for PTEs
+ *
+ * @params: see amdgpu_vm_update_params definition
+ * @start: first PTE to handle
+ * @end: last PTE to handle
+ * @flags: hw mapping flags
+ * @frag: resulting fragment size
+ * @frag_end: end of this fragment
+ *
+ * Returns the first possible fragment for the start and end address.
+ */
+static void amdgpu_vm_pte_fragment(struct amdgpu_vm_update_params *params,
+				   uint64_t start, uint64_t end, uint64_t flags,
+				   unsigned int *frag, uint64_t *frag_end)
+{
+	/**
+	 * The MC L1 TLB supports variable sized pages, based on a fragment
+	 * field in the PTE. When this field is set to a non-zero value, page
+	 * granularity is increased from 4KB to (1 << (12 + frag)). The PTE
+	 * flags are considered valid for all PTEs within the fragment range
+	 * and corresponding mappings are assumed to be physically contiguous.
+	 *
+	 * The L1 TLB can store a single PTE for the whole fragment,
+	 * significantly increasing the space available for translation
+	 * caching. This leads to large improvements in throughput when the
+	 * TLB is under pressure.
+	 *
+	 * The L2 TLB distributes small and large fragments into two
+	 * asymmetric partitions. The large fragment cache is significantly
+	 * larger. Thus, we try to use large fragments wherever possible.
+	 * Userspace can support this by aligning virtual base address and
+	 * allocation size to the fragment size.
+	 *
+	 * Starting with Vega10 the fragment size only controls the L1. The L2
+	 * is now directly feed with small/huge/giant pages from the walker.
+	 */
+	unsigned int max_frag;
+
+	if (params->adev->asic_type < CHIP_VEGA10)
+		max_frag = params->adev->vm_manager.fragment_size;
+	else
+		max_frag = 31;
+
+	/* system pages are non continuously */
+	if (params->pages_addr) {
+		*frag = 0;
+		*frag_end = end;
+		return;
+	}
+
+	/* This intentionally wraps around if no bit is set */
+	*frag = min_t(unsigned int, ffs(start) - 1, fls64(end - start) - 1);
+	if (*frag >= max_frag) {
+		*frag = max_frag;
+		*frag_end = end & ~((1ULL << max_frag) - 1);
+	} else {
+		*frag_end = start + (1 << *frag);
+	}
+}
+
+/**
+ * amdgpu_vm_ptes_update - make sure that page tables are valid
+ *
+ * @params: see amdgpu_vm_update_params definition
+ * @start: start of GPU address range
+ * @end: end of GPU address range
+ * @dst: destination address to map to, the next dst inside the function
+ * @flags: mapping flags
+ *
+ * Update the page tables in the range @start - @end.
+ *
+ * Returns:
+ * 0 for success, -EINVAL for failure.
+ */
+int amdgpu_vm_ptes_update(struct amdgpu_vm_update_params *params,
+			  uint64_t start, uint64_t end,
+			  uint64_t dst, uint64_t flags)
+{
+	struct amdgpu_device *adev = params->adev;
+	struct amdgpu_vm_pt_cursor cursor;
+	uint64_t frag_start = start, frag_end;
+	unsigned int frag;
+	int r;
+
+	/* figure out the initial fragment */
+	amdgpu_vm_pte_fragment(params, frag_start, end, flags, &frag,
+			       &frag_end);
+
+	/* walk over the address space and update the PTs */
+	amdgpu_vm_pt_start(adev, params->vm, start, &cursor);
+	while (cursor.pfn < end) {
+		unsigned int shift, parent_shift, mask;
+		uint64_t incr, entry_end, pe_start;
+		struct amdgpu_bo *pt;
+
+		if (!params->unlocked) {
+			/* make sure that the page tables covering the
+			 * address range are actually allocated
+			 */
+			r = amdgpu_vm_pt_alloc(params->adev, params->vm,
+					       &cursor, params->immediate);
+			if (r)
+				return r;
+		}
+
+		shift = amdgpu_vm_pt_level_shift(adev, cursor.level);
+		parent_shift = amdgpu_vm_pt_level_shift(adev, cursor.level - 1);
+		if (params->unlocked) {
+			/* Unlocked updates are only allowed on the leaves */
+			if (amdgpu_vm_pt_descendant(adev, &cursor))
+				continue;
+		} else if (adev->asic_type < CHIP_VEGA10 &&
+			   (flags & AMDGPU_PTE_VALID)) {
+			/* No huge page support before GMC v9 */
+			if (cursor.level != AMDGPU_VM_PTB) {
+				if (!amdgpu_vm_pt_descendant(adev, &cursor))
+					return -ENOENT;
+				continue;
+			}
+		} else if (frag < shift) {
+			/* We can't use this level when the fragment size is
+			 * smaller than the address shift. Go to the next
+			 * child entry and try again.
+			 */
+			if (amdgpu_vm_pt_descendant(adev, &cursor))
+				continue;
+		} else if (frag >= parent_shift) {
+			/* If the fragment size is even larger than the parent
+			 * shift we should go up one level and check it again.
+			 */
+			if (!amdgpu_vm_pt_ancestor(&cursor))
+				return -EINVAL;
+			continue;
+		}
+
+		pt = cursor.entry->bo;
+		if (!pt) {
+			/* We need all PDs and PTs for mapping something, */
+			if (flags & AMDGPU_PTE_VALID)
+				return -ENOENT;
+
+			/* but unmapping something can happen at a higher
+			 * level.
+			 */
+			if (!amdgpu_vm_pt_ancestor(&cursor))
+				return -EINVAL;
+
+			pt = cursor.entry->bo;
+			shift = parent_shift;
+			frag_end = max(frag_end, ALIGN(frag_start + 1,
+				   1ULL << shift));
+		}
+
+		/* Looks good so far, calculate parameters for the update */
+		incr = (uint64_t)AMDGPU_GPU_PAGE_SIZE << shift;
+		mask = amdgpu_vm_pt_entries_mask(adev, cursor.level);
+		pe_start = ((cursor.pfn >> shift) & mask) * 8;
+
+		if (cursor.level < AMDGPU_VM_PTB && params->unlocked)
+			/*
+			 * MMU notifier callback unlocked unmap huge page, leave is PDE entry,
+			 * only clear one entry. Next entry search again for PDE or PTE leave.
+			 */
+			entry_end = 1ULL << shift;
+		else
+			entry_end = ((uint64_t)mask + 1) << shift;
+		entry_end += cursor.pfn & ~(entry_end - 1);
+		entry_end = min(entry_end, end);
+
+		do {
+			struct amdgpu_vm *vm = params->vm;
+			uint64_t upd_end = min(entry_end, frag_end);
+			unsigned int nptes = (upd_end - frag_start) >> shift;
+			uint64_t upd_flags = flags | AMDGPU_PTE_FRAG(frag);
+
+			/* This can happen when we set higher level PDs to
+			 * silent to stop fault floods.
+			 */
+			nptes = max(nptes, 1u);
+
+			trace_amdgpu_vm_update_ptes(params, frag_start, upd_end,
+						    min(nptes, 32u), dst, incr,
+						    upd_flags,
+						    vm->task_info ? vm->task_info->tgid : 0,
+						    vm->immediate.fence_context);
+			amdgpu_vm_pte_update_flags(params, to_amdgpu_bo_vm(pt),
+						   cursor.level, pe_start, dst,
+						   nptes, incr, upd_flags);
+
+			pe_start += nptes * 8;
+			dst += nptes * incr;
+
+			frag_start = upd_end;
+			if (frag_start >= frag_end) {
+				/* figure out the next fragment */
+				amdgpu_vm_pte_fragment(params, frag_start, end,
+						       flags, &frag, &frag_end);
+				if (frag < shift)
+					break;
+			}
+		} while (frag_start < entry_end);
+
+		if (amdgpu_vm_pt_descendant(adev, &cursor)) {
+			/* Free all child entries.
+			 * Update the tables with the flags and addresses and free up subsequent
+			 * tables in the case of huge pages or freed up areas.
+			 * This is the maximum you can free, because all other page tables are not
+			 * completely covered by the range and so potentially still in use.
+			 */
+			while (cursor.pfn < frag_start) {
+				/* Make sure previous mapping is freed */
+				if (cursor.entry->bo) {
+					params->needs_flush = true;
+					amdgpu_vm_pt_add_list(params, &cursor);
+				}
+				amdgpu_vm_pt_next(adev, &cursor);
+			}
+
+		} else if (frag >= shift) {
+			/* or just move on to the next on the same level. */
+			amdgpu_vm_pt_next(adev, &cursor);
+		}
+	}
+
+	return 0;
+}
+
+/**
+ * amdgpu_vm_pt_map_tables - have bo of root PD cpu accessible
+ * @adev: amdgpu device structure
+ * @vm: amdgpu vm structure
+ *
+ * make root page directory and everything below it cpu accessible.
+ */
+int amdgpu_vm_pt_map_tables(struct amdgpu_device *adev, struct amdgpu_vm *vm)
+{
+	struct amdgpu_vm_pt_cursor cursor;
+	struct amdgpu_vm_bo_base *entry;
+
+	for_each_amdgpu_vm_pt_dfs_safe(adev, vm, NULL, cursor, entry) {
+
+		struct amdgpu_bo_vm *bo;
+		int r;
+
+		if (entry->bo) {
+			bo = to_amdgpu_bo_vm(entry->bo);
+			r = vm->update_funcs->map_table(bo);
+			if (r)
+				return r;
+		}
+	}
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_sdma.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_sdma.c
index dbb551762805..36805dcfa159 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_sdma.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_sdma.c
@@ -35,47 +35,63 @@
  */
 static int amdgpu_vm_sdma_map_table(struct amdgpu_bo_vm *table)
 {
+	return amdgpu_ttm_alloc_gart(&table->bo.tbo);
+}
+
+/* Allocate a new job for @count PTE updates */
+static int amdgpu_vm_sdma_alloc_job(struct amdgpu_vm_update_params *p,
+				    unsigned int count, u64 k_job_id)
+{
+	enum amdgpu_ib_pool_type pool = p->immediate ? AMDGPU_IB_POOL_IMMEDIATE
+		: AMDGPU_IB_POOL_DELAYED;
+	struct drm_sched_entity *entity = p->immediate ? &p->vm->immediate
+		: &p->vm->delayed;
+	unsigned int ndw;
 	int r;
 
-	r = amdgpu_ttm_alloc_gart(&table->bo.tbo);
+	/* estimate how many dw we need */
+	ndw = AMDGPU_VM_SDMA_MIN_NUM_DW;
+	if (p->pages_addr)
+		ndw += count * 2;
+	ndw = min(ndw, AMDGPU_VM_SDMA_MAX_NUM_DW);
+
+	r = amdgpu_job_alloc_with_ib(p->adev, entity, AMDGPU_FENCE_OWNER_VM,
+				     ndw * 4, pool, &p->job, k_job_id);
 	if (r)
 		return r;
 
-	if (table->shadow)
-		r = amdgpu_ttm_alloc_gart(&table->shadow->tbo);
-
-	return r;
+	p->num_dw_left = ndw;
+	return 0;
 }
 
 /**
  * amdgpu_vm_sdma_prepare - prepare SDMA command submission
  *
  * @p: see amdgpu_vm_update_params definition
- * @resv: reservation object with embedded fence
- * @sync_mode: synchronization mode
+ * @sync: amdgpu_sync object with fences to wait for
+ * @k_job_id: identifier of the job, for tracing purpose
  *
  * Returns:
  * Negativ errno, 0 for success.
  */
 static int amdgpu_vm_sdma_prepare(struct amdgpu_vm_update_params *p,
-				  struct dma_resv *resv,
-				  enum amdgpu_sync_mode sync_mode)
+				  struct amdgpu_sync *sync, u64 k_job_id)
 {
-	enum amdgpu_ib_pool_type pool = p->immediate ? AMDGPU_IB_POOL_IMMEDIATE
-		: AMDGPU_IB_POOL_DELAYED;
-	unsigned int ndw = AMDGPU_VM_SDMA_MIN_NUM_DW;
 	int r;
 
-	r = amdgpu_job_alloc_with_ib(p->adev, ndw * 4, pool, &p->job);
+	r = amdgpu_vm_sdma_alloc_job(p, 0, k_job_id);
 	if (r)
 		return r;
 
-	p->num_dw_left = ndw;
-
-	if (!resv)
+	if (!sync)
 		return 0;
 
-	return amdgpu_sync_resv(p->adev, &p->job->sync, resv, sync_mode, p->vm);
+	r = amdgpu_sync_push_to_job(sync, p->job);
+	if (r) {
+		p->num_dw_left = 0;
+		amdgpu_job_free(p->job);
+	}
+	return r;
 }
 
 /**
@@ -91,38 +107,42 @@ static int amdgpu_vm_sdma_commit(struct amdgpu_vm_update_params *p,
 				 struct dma_fence **fence)
 {
 	struct amdgpu_ib *ib = p->job->ibs;
-	struct drm_sched_entity *entity;
 	struct amdgpu_ring *ring;
 	struct dma_fence *f;
-	int r;
 
-	entity = p->immediate ? &p->vm->immediate : &p->vm->delayed;
-	ring = container_of(entity->rq->sched, struct amdgpu_ring, sched);
+	ring = container_of(p->vm->delayed.rq->sched, struct amdgpu_ring,
+			    sched);
 
 	WARN_ON(ib->length_dw == 0);
 	amdgpu_ring_pad_ib(ring, ib);
+
+	if (p->needs_flush)
+		atomic64_inc(&p->vm->tlb_seq);
+
 	WARN_ON(ib->length_dw > p->num_dw_left);
-	r = amdgpu_job_submit(p->job, entity, AMDGPU_FENCE_OWNER_VM, &f);
-	if (r)
-		goto error;
+	f = amdgpu_job_submit(p->job);
 
 	if (p->unlocked) {
 		struct dma_fence *tmp = dma_fence_get(f);
 
-		swap(p->vm->last_unlocked, f);
+		swap(p->vm->last_unlocked, tmp);
 		dma_fence_put(tmp);
 	} else {
-		amdgpu_bo_fence(p->vm->root.bo, f, true);
+		dma_resv_add_fence(p->vm->root.bo->tbo.base.resv, f,
+				   DMA_RESV_USAGE_BOOKKEEP);
 	}
 
-	if (fence && !p->immediate)
+	if (fence && !p->immediate) {
+		/*
+		 * Most hw generations now have a separate queue for page table
+		 * updates, but when the queue is shared with userspace we need
+		 * the extra CPU round trip to correctly flush the TLB.
+		 */
+		set_bit(DRM_SCHED_FENCE_DONT_PIPELINE, &f->flags);
 		swap(*fence, f);
+	}
 	dma_fence_put(f);
 	return 0;
-
-error:
-	amdgpu_job_free(p->job);
-	return r;
 }
 
 /**
@@ -144,7 +164,7 @@ static void amdgpu_vm_sdma_copy_ptes(struct amdgpu_vm_update_params *p,
 
 	src += p->num_dw_left * 4;
 
-	pe += amdgpu_gmc_sign_extend(amdgpu_bo_gpu_offset_no_check(bo));
+	pe += amdgpu_bo_gpu_offset_no_check(bo);
 	trace_amdgpu_vm_copy_ptes(pe, src, count, p->immediate);
 
 	amdgpu_vm_copy_pte(p->adev, ib, pe, src, count);
@@ -171,7 +191,7 @@ static void amdgpu_vm_sdma_set_ptes(struct amdgpu_vm_update_params *p,
 {
 	struct amdgpu_ib *ib = p->job->ibs;
 
-	pe += amdgpu_gmc_sign_extend(amdgpu_bo_gpu_offset_no_check(bo));
+	pe += amdgpu_bo_gpu_offset_no_check(bo);
 	trace_amdgpu_vm_set_ptes(pe, addr, count, incr, flags, p->immediate);
 	if (count < 3) {
 		amdgpu_vm_write_pte(p->adev, ib, pe, addr | flags,
@@ -202,16 +222,24 @@ static int amdgpu_vm_sdma_update(struct amdgpu_vm_update_params *p,
 				 uint64_t flags)
 {
 	struct amdgpu_bo *bo = &vmbo->bo;
-	enum amdgpu_ib_pool_type pool = p->immediate ? AMDGPU_IB_POOL_IMMEDIATE
-		: AMDGPU_IB_POOL_DELAYED;
+	struct dma_resv_iter cursor;
 	unsigned int i, ndw, nptes;
+	struct dma_fence *fence;
 	uint64_t *pte;
 	int r;
 
 	/* Wait for PD/PT moves to be completed */
-	r = amdgpu_sync_fence(&p->job->sync, bo->tbo.moving);
-	if (r)
-		return r;
+	dma_resv_iter_begin(&cursor, bo->tbo.base.resv, DMA_RESV_USAGE_KERNEL);
+	dma_resv_for_each_fence_unlocked(&cursor, fence) {
+		dma_fence_get(fence);
+		r = drm_sched_job_add_dependency(&p->job->base, fence);
+		if (r) {
+			dma_fence_put(fence);
+			dma_resv_iter_end(&cursor);
+			return r;
+		}
+	}
+	dma_resv_iter_end(&cursor);
 
 	do {
 		ndw = p->num_dw_left;
@@ -222,34 +250,21 @@ static int amdgpu_vm_sdma_update(struct amdgpu_vm_update_params *p,
 			if (r)
 				return r;
 
-			/* estimate how many dw we need */
-			ndw = 32;
-			if (p->pages_addr)
-				ndw += count * 2;
-			ndw = max(ndw, AMDGPU_VM_SDMA_MIN_NUM_DW);
-			ndw = min(ndw, AMDGPU_VM_SDMA_MAX_NUM_DW);
-
-			r = amdgpu_job_alloc_with_ib(p->adev, ndw * 4, pool,
-						     &p->job);
+			r = amdgpu_vm_sdma_alloc_job(p, count,
+						     AMDGPU_KERNEL_JOB_ID_VM_UPDATE);
 			if (r)
 				return r;
-
-			p->num_dw_left = ndw;
 		}
 
 		if (!p->pages_addr) {
 			/* set page commands needed */
-			if (vmbo->shadow)
-				amdgpu_vm_sdma_set_ptes(p, vmbo->shadow, pe, addr,
-							count, incr, flags);
 			amdgpu_vm_sdma_set_ptes(p, bo, pe, addr, count,
 						incr, flags);
 			return 0;
 		}
 
 		/* copy commands needed */
-		ndw -= p->adev->vm_manager.vm_pte_funcs->copy_pte_num_dw *
-			(vmbo->shadow ? 2 : 1);
+		ndw -= p->adev->vm_manager.vm_pte_funcs->copy_pte_num_dw;
 
 		/* for padding */
 		ndw -= 7;
@@ -264,8 +279,6 @@ static int amdgpu_vm_sdma_update(struct amdgpu_vm_update_params *p,
 			pte[i] |= flags;
 		}
 
-		if (vmbo->shadow)
-			amdgpu_vm_sdma_copy_ptes(p, vmbo->shadow, pe, nptes);
 		amdgpu_vm_sdma_copy_ptes(p, bo, pe, nptes);
 
 		pe += nptes * 8;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_tlb_fence.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_tlb_fence.c
new file mode 100644
index 000000000000..5d26797356a3
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_tlb_fence.c
@@ -0,0 +1,111 @@
+// SPDX-License-Identifier: GPL-2.0 OR MIT
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#include <linux/dma-fence.h>
+#include <linux/workqueue.h>
+
+#include "amdgpu.h"
+#include "amdgpu_vm.h"
+#include "amdgpu_gmc.h"
+
+struct amdgpu_tlb_fence {
+	struct dma_fence	base;
+	struct amdgpu_device	*adev;
+	struct dma_fence	*dependency;
+	struct work_struct	work;
+	spinlock_t		lock;
+	uint16_t		pasid;
+
+};
+
+static const char *amdgpu_tlb_fence_get_driver_name(struct dma_fence *fence)
+{
+	return "amdgpu tlb fence";
+}
+
+static const char *amdgpu_tlb_fence_get_timeline_name(struct dma_fence *f)
+{
+	return "amdgpu tlb timeline";
+}
+
+static void amdgpu_tlb_fence_work(struct work_struct *work)
+{
+	struct amdgpu_tlb_fence *f = container_of(work, typeof(*f), work);
+	int r;
+
+	if (f->dependency) {
+		dma_fence_wait(f->dependency, false);
+		dma_fence_put(f->dependency);
+		f->dependency = NULL;
+	}
+
+	r = amdgpu_gmc_flush_gpu_tlb_pasid(f->adev, f->pasid, 2, true, 0);
+	if (r) {
+		dev_err(f->adev->dev, "TLB flush failed for PASID %d.\n",
+			f->pasid);
+		dma_fence_set_error(&f->base, r);
+	}
+
+	dma_fence_signal(&f->base);
+	dma_fence_put(&f->base);
+}
+
+static const struct dma_fence_ops amdgpu_tlb_fence_ops = {
+	.get_driver_name = amdgpu_tlb_fence_get_driver_name,
+	.get_timeline_name = amdgpu_tlb_fence_get_timeline_name
+};
+
+void amdgpu_vm_tlb_fence_create(struct amdgpu_device *adev, struct amdgpu_vm *vm,
+				struct dma_fence **fence)
+{
+	struct amdgpu_tlb_fence *f;
+
+	f = kmalloc(sizeof(*f), GFP_KERNEL);
+	if (!f) {
+		/*
+		 * We can't fail since the PDEs and PTEs are already updated, so
+		 * just block for the dependency and execute the TLB flush
+		 */
+		if (*fence)
+			dma_fence_wait(*fence, false);
+
+		amdgpu_gmc_flush_gpu_tlb_pasid(adev, vm->pasid, 2, true, 0);
+		*fence = dma_fence_get_stub();
+		return;
+	}
+
+	f->adev = adev;
+	f->dependency = *fence;
+	f->pasid = vm->pasid;
+	INIT_WORK(&f->work, amdgpu_tlb_fence_work);
+	spin_lock_init(&f->lock);
+
+	dma_fence_init64(&f->base, &amdgpu_tlb_fence_ops, &f->lock,
+			 vm->tlb_fence_context, atomic64_read(&vm->tlb_seq));
+
+	/* TODO: We probably need a separate wq here */
+	dma_fence_get(&f->base);
+	schedule_work(&f->work);
+
+	*fence = &f->base;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vpe.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vpe.c
new file mode 100644
index 000000000000..aa78c2ee9e21
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vpe.c
@@ -0,0 +1,1018 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#include <linux/firmware.h>
+#include <drm/drm_drv.h>
+
+#include "amdgpu.h"
+#include "amdgpu_ucode.h"
+#include "amdgpu_vpe.h"
+#include "amdgpu_smu.h"
+#include "soc15_common.h"
+#include "vpe_v6_1.h"
+
+#define AMDGPU_CSA_VPE_SIZE 	64
+/* VPE CSA resides in the 4th page of CSA */
+#define AMDGPU_CSA_VPE_OFFSET 	(4096 * 3)
+
+/* 1 second timeout */
+#define VPE_IDLE_TIMEOUT	msecs_to_jiffies(1000)
+
+#define VPE_MAX_DPM_LEVEL			4
+#define FIXED1_8_BITS_PER_FRACTIONAL_PART	8
+#define GET_PRATIO_INTEGER_PART(x)		((x) >> FIXED1_8_BITS_PER_FRACTIONAL_PART)
+
+static void vpe_set_ring_funcs(struct amdgpu_device *adev);
+
+static inline uint16_t div16_u16_rem(uint16_t dividend, uint16_t divisor, uint16_t *remainder)
+{
+	*remainder = dividend % divisor;
+	return dividend / divisor;
+}
+
+static inline uint16_t complete_integer_division_u16(
+	uint16_t dividend,
+	uint16_t divisor,
+	uint16_t *remainder)
+{
+	return div16_u16_rem(dividend, divisor, (uint16_t *)remainder);
+}
+
+static uint16_t vpe_u1_8_from_fraction(uint16_t numerator, uint16_t denominator)
+{
+	u16 arg1_value = numerator;
+	u16 arg2_value = denominator;
+
+	uint16_t remainder;
+
+	/* determine integer part */
+	uint16_t res_value = complete_integer_division_u16(
+		arg1_value, arg2_value, &remainder);
+
+	if (res_value > 127 /* CHAR_MAX */)
+		return 0;
+
+	/* determine fractional part */
+	{
+		unsigned int i = FIXED1_8_BITS_PER_FRACTIONAL_PART;
+
+		do {
+			remainder <<= 1;
+
+			res_value <<= 1;
+
+			if (remainder >= arg2_value) {
+				res_value |= 1;
+				remainder -= arg2_value;
+			}
+		} while (--i != 0);
+	}
+
+	/* round up LSB */
+	{
+		uint16_t summand = (remainder << 1) >= arg2_value;
+
+		if ((res_value + summand) > 32767 /* SHRT_MAX */)
+			return 0;
+
+		res_value += summand;
+	}
+
+	return res_value;
+}
+
+static uint16_t vpe_internal_get_pratio(uint16_t from_frequency, uint16_t to_frequency)
+{
+	uint16_t pratio = vpe_u1_8_from_fraction(from_frequency, to_frequency);
+
+	if (GET_PRATIO_INTEGER_PART(pratio) > 1)
+		pratio = 0;
+
+	return pratio;
+}
+
+/*
+ * VPE has 4 DPM levels from level 0 (lowerest) to 3 (highest),
+ * VPE FW will dynamically decide which level should be used according to current loading.
+ *
+ * Get VPE and SOC clocks from PM, and select the appropriate four clock values,
+ * calculate the ratios of adjusting from one clock to another.
+ * The VPE FW can then request the appropriate frequency from the PMFW.
+ */
+int amdgpu_vpe_configure_dpm(struct amdgpu_vpe *vpe)
+{
+	struct amdgpu_device *adev = vpe->ring.adev;
+	uint32_t dpm_ctl;
+
+	if (adev->pm.dpm_enabled) {
+		struct dpm_clocks clock_table = { 0 };
+		struct dpm_clock *VPEClks;
+		struct dpm_clock *SOCClks;
+		uint32_t idx;
+		uint32_t vpeclk_enalbled_num = 0;
+		uint32_t pratio_vmax_vnorm = 0, pratio_vnorm_vmid = 0, pratio_vmid_vmin = 0;
+		uint16_t pratio_vmin_freq = 0, pratio_vmid_freq = 0, pratio_vnorm_freq = 0, pratio_vmax_freq = 0;
+
+		dpm_ctl = RREG32(vpe_get_reg_offset(vpe, 0, vpe->regs.dpm_enable));
+		dpm_ctl |= 1; /* DPM enablement */
+		WREG32(vpe_get_reg_offset(vpe, 0, vpe->regs.dpm_enable), dpm_ctl);
+
+		/* Get VPECLK and SOCCLK */
+		if (amdgpu_dpm_get_dpm_clock_table(adev, &clock_table)) {
+			dev_dbg(adev->dev, "%s: get clock failed!\n", __func__);
+			goto disable_dpm;
+		}
+
+		SOCClks = clock_table.SocClocks;
+		VPEClks = clock_table.VPEClocks;
+
+		/* Comfirm enabled vpe clk num
+		 * Enabled VPE clocks are ordered from low to high in VPEClks
+		 * The highest valid clock index+1 is the number of VPEClks
+		 */
+		for (idx = PP_SMU_NUM_VPECLK_DPM_LEVELS; idx && !vpeclk_enalbled_num; idx--)
+			if (VPEClks[idx-1].Freq)
+				vpeclk_enalbled_num = idx;
+
+		/* vpe dpm only cares 4 levels. */
+		for (idx = 0; idx < VPE_MAX_DPM_LEVEL; idx++) {
+			uint32_t soc_dpm_level;
+			uint32_t min_freq;
+
+			if (idx == 0)
+				soc_dpm_level = 0;
+			else
+				soc_dpm_level = (idx * 2) + 1;
+
+			/* clamp the max level */
+			if (soc_dpm_level > vpeclk_enalbled_num - 1)
+				soc_dpm_level = vpeclk_enalbled_num - 1;
+
+			min_freq = (SOCClks[soc_dpm_level].Freq < VPEClks[soc_dpm_level].Freq) ?
+				   SOCClks[soc_dpm_level].Freq : VPEClks[soc_dpm_level].Freq;
+
+			switch (idx) {
+			case 0:
+				pratio_vmin_freq = min_freq;
+				break;
+			case 1:
+				pratio_vmid_freq = min_freq;
+				break;
+			case 2:
+				pratio_vnorm_freq = min_freq;
+				break;
+			case 3:
+				pratio_vmax_freq = min_freq;
+				break;
+			default:
+				break;
+			}
+		}
+
+		if (pratio_vmin_freq && pratio_vmid_freq && pratio_vnorm_freq && pratio_vmax_freq) {
+			uint32_t pratio_ctl;
+
+			pratio_vmax_vnorm = (uint32_t)vpe_internal_get_pratio(pratio_vmax_freq, pratio_vnorm_freq);
+			pratio_vnorm_vmid = (uint32_t)vpe_internal_get_pratio(pratio_vnorm_freq, pratio_vmid_freq);
+			pratio_vmid_vmin = (uint32_t)vpe_internal_get_pratio(pratio_vmid_freq, pratio_vmin_freq);
+
+			pratio_ctl = pratio_vmax_vnorm | (pratio_vnorm_vmid << 9) | (pratio_vmid_vmin << 18);
+			WREG32(vpe_get_reg_offset(vpe, 0, vpe->regs.dpm_pratio), pratio_ctl);		/* PRatio */
+			WREG32(vpe_get_reg_offset(vpe, 0, vpe->regs.dpm_request_interval), 24000);	/* 1ms, unit=1/24MHz */
+			WREG32(vpe_get_reg_offset(vpe, 0, vpe->regs.dpm_decision_threshold), 1200000);	/* 50ms */
+			WREG32(vpe_get_reg_offset(vpe, 0, vpe->regs.dpm_busy_clamp_threshold), 1200000);/* 50ms */
+			WREG32(vpe_get_reg_offset(vpe, 0, vpe->regs.dpm_idle_clamp_threshold), 1200000);/* 50ms */
+			dev_dbg(adev->dev, "%s: configure vpe dpm pratio done!\n", __func__);
+		} else {
+			dev_dbg(adev->dev, "%s: invalid pratio parameters!\n", __func__);
+			goto disable_dpm;
+		}
+	}
+	return 0;
+
+disable_dpm:
+	dpm_ctl = RREG32(vpe_get_reg_offset(vpe, 0, vpe->regs.dpm_enable));
+	dpm_ctl &= 0xfffffffe; /* Disable DPM */
+	WREG32(vpe_get_reg_offset(vpe, 0, vpe->regs.dpm_enable), dpm_ctl);
+	dev_dbg(adev->dev, "%s: disable vpe dpm\n", __func__);
+	return -EINVAL;
+}
+
+int amdgpu_vpe_psp_update_sram(struct amdgpu_device *adev)
+{
+	struct amdgpu_firmware_info ucode = {
+		.ucode_id = AMDGPU_UCODE_ID_VPE,
+		.mc_addr = adev->vpe.cmdbuf_gpu_addr,
+		.ucode_size = 8,
+	};
+
+	return psp_execute_ip_fw_load(&adev->psp, &ucode);
+}
+
+int amdgpu_vpe_init_microcode(struct amdgpu_vpe *vpe)
+{
+	struct amdgpu_device *adev = vpe->ring.adev;
+	const struct vpe_firmware_header_v1_0 *vpe_hdr;
+	char fw_prefix[32];
+	int ret;
+
+	amdgpu_ucode_ip_version_decode(adev, VPE_HWIP, fw_prefix, sizeof(fw_prefix));
+	ret = amdgpu_ucode_request(adev, &adev->vpe.fw, AMDGPU_UCODE_REQUIRED,
+				   "amdgpu/%s.bin", fw_prefix);
+	if (ret)
+		goto out;
+
+	vpe_hdr = (const struct vpe_firmware_header_v1_0 *)adev->vpe.fw->data;
+	adev->vpe.fw_version = le32_to_cpu(vpe_hdr->header.ucode_version);
+	adev->vpe.feature_version = le32_to_cpu(vpe_hdr->ucode_feature_version);
+
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
+		struct amdgpu_firmware_info *info;
+
+		info = &adev->firmware.ucode[AMDGPU_UCODE_ID_VPE_CTX];
+		info->ucode_id = AMDGPU_UCODE_ID_VPE_CTX;
+		info->fw = adev->vpe.fw;
+		adev->firmware.fw_size +=
+			ALIGN(le32_to_cpu(vpe_hdr->ctx_ucode_size_bytes), PAGE_SIZE);
+
+		info = &adev->firmware.ucode[AMDGPU_UCODE_ID_VPE_CTL];
+		info->ucode_id = AMDGPU_UCODE_ID_VPE_CTL;
+		info->fw = adev->vpe.fw;
+		adev->firmware.fw_size +=
+			ALIGN(le32_to_cpu(vpe_hdr->ctl_ucode_size_bytes), PAGE_SIZE);
+	}
+
+	return 0;
+out:
+	dev_err(adev->dev, "fail to initialize vpe microcode\n");
+	release_firmware(adev->vpe.fw);
+	adev->vpe.fw = NULL;
+	return ret;
+}
+
+int amdgpu_vpe_ring_init(struct amdgpu_vpe *vpe)
+{
+	struct amdgpu_device *adev = container_of(vpe, struct amdgpu_device, vpe);
+	struct amdgpu_ring *ring = &vpe->ring;
+	int ret;
+
+	ring->ring_obj = NULL;
+	ring->use_doorbell = true;
+	ring->vm_hub = AMDGPU_MMHUB0(0);
+	ring->doorbell_index = (adev->doorbell_index.vpe_ring << 1);
+	snprintf(ring->name, 4, "vpe");
+
+	ret = amdgpu_ring_init(adev, ring, 1024, &vpe->trap_irq, 0,
+			     AMDGPU_RING_PRIO_DEFAULT, NULL);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+int amdgpu_vpe_ring_fini(struct amdgpu_vpe *vpe)
+{
+	amdgpu_ring_fini(&vpe->ring);
+
+	return 0;
+}
+
+static int vpe_early_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	struct amdgpu_vpe *vpe = &adev->vpe;
+
+	switch (amdgpu_ip_version(adev, VPE_HWIP, 0)) {
+	case IP_VERSION(6, 1, 0):
+	case IP_VERSION(6, 1, 3):
+		vpe_v6_1_set_funcs(vpe);
+		break;
+	case IP_VERSION(6, 1, 1):
+		vpe_v6_1_set_funcs(vpe);
+		vpe->collaborate_mode = true;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	vpe_set_ring_funcs(adev);
+	vpe_set_regs(vpe);
+
+	dev_info(adev->dev, "VPE: collaborate mode %s", vpe->collaborate_mode ? "true" : "false");
+
+	return 0;
+}
+
+static bool vpe_need_dpm0_at_power_down(struct amdgpu_device *adev)
+{
+	switch (amdgpu_ip_version(adev, VPE_HWIP, 0)) {
+	case IP_VERSION(6, 1, 1):
+		return adev->pm.fw_version < 0x0a640500;
+	default:
+		return false;
+	}
+}
+
+static int vpe_get_dpm_level(struct amdgpu_device *adev)
+{
+	struct amdgpu_vpe *vpe = &adev->vpe;
+
+	if (!adev->pm.dpm_enabled)
+		return 0;
+
+	return RREG32(vpe_get_reg_offset(vpe, 0, vpe->regs.dpm_request_lv));
+}
+
+static void vpe_idle_work_handler(struct work_struct *work)
+{
+	struct amdgpu_device *adev =
+		container_of(work, struct amdgpu_device, vpe.idle_work.work);
+	unsigned int fences = 0;
+
+	fences += amdgpu_fence_count_emitted(&adev->vpe.ring);
+	if (fences)
+		goto reschedule;
+
+	if (vpe_need_dpm0_at_power_down(adev) && vpe_get_dpm_level(adev) != 0)
+		goto reschedule;
+
+	amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_VPE, AMD_PG_STATE_GATE);
+	return;
+
+reschedule:
+	schedule_delayed_work(&adev->vpe.idle_work, VPE_IDLE_TIMEOUT);
+}
+
+static int vpe_common_init(struct amdgpu_vpe *vpe)
+{
+	struct amdgpu_device *adev = container_of(vpe, struct amdgpu_device, vpe);
+	int r;
+
+	r = amdgpu_bo_create_kernel(adev, PAGE_SIZE, PAGE_SIZE,
+				    AMDGPU_GEM_DOMAIN_GTT,
+				    &adev->vpe.cmdbuf_obj,
+				    &adev->vpe.cmdbuf_gpu_addr,
+				    (void **)&adev->vpe.cmdbuf_cpu_addr);
+	if (r) {
+		dev_err(adev->dev, "VPE: failed to allocate cmdbuf bo %d\n", r);
+		return r;
+	}
+
+	vpe->context_started = false;
+	INIT_DELAYED_WORK(&adev->vpe.idle_work, vpe_idle_work_handler);
+
+	return 0;
+}
+
+static int vpe_sw_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	struct amdgpu_vpe *vpe = &adev->vpe;
+	int ret;
+
+	ret = vpe_common_init(vpe);
+	if (ret)
+		goto out;
+
+	ret = vpe_irq_init(vpe);
+	if (ret)
+		goto out;
+
+	ret = vpe_ring_init(vpe);
+	if (ret)
+		goto out;
+
+	ret = vpe_init_microcode(vpe);
+	if (ret)
+		goto out;
+
+	adev->vpe.supported_reset =
+		 amdgpu_get_soft_full_reset_mask(&adev->vpe.ring);
+	if (!amdgpu_sriov_vf(adev))
+		adev->vpe.supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE;
+	ret = amdgpu_vpe_sysfs_reset_mask_init(adev);
+	if (ret)
+		goto out;
+out:
+	return ret;
+}
+
+static int vpe_sw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	struct amdgpu_vpe *vpe = &adev->vpe;
+
+	release_firmware(vpe->fw);
+	vpe->fw = NULL;
+
+	amdgpu_vpe_sysfs_reset_mask_fini(adev);
+	vpe_ring_fini(vpe);
+
+	amdgpu_bo_free_kernel(&adev->vpe.cmdbuf_obj,
+			      &adev->vpe.cmdbuf_gpu_addr,
+			      (void **)&adev->vpe.cmdbuf_cpu_addr);
+
+	return 0;
+}
+
+static int vpe_hw_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	struct amdgpu_vpe *vpe = &adev->vpe;
+	int ret;
+
+	/* Power on VPE */
+	ret = amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_VPE,
+						     AMD_PG_STATE_UNGATE);
+	if (ret)
+		return ret;
+
+	ret = vpe_load_microcode(vpe);
+	if (ret)
+		return ret;
+
+	ret = vpe_ring_start(vpe);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+static int vpe_hw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	struct amdgpu_vpe *vpe = &adev->vpe;
+
+	cancel_delayed_work_sync(&adev->vpe.idle_work);
+
+	vpe_ring_stop(vpe);
+
+	/* Power off VPE */
+	amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_VPE, AMD_PG_STATE_GATE);
+
+	return 0;
+}
+
+static int vpe_suspend(struct amdgpu_ip_block *ip_block)
+{
+	return vpe_hw_fini(ip_block);
+}
+
+static int vpe_resume(struct amdgpu_ip_block *ip_block)
+{
+	return vpe_hw_init(ip_block);
+}
+
+static void vpe_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count)
+{
+	int i;
+
+	for (i = 0; i < count; i++)
+		if (i == 0)
+			amdgpu_ring_write(ring, ring->funcs->nop |
+				VPE_CMD_NOP_HEADER_COUNT(count - 1));
+		else
+			amdgpu_ring_write(ring, ring->funcs->nop);
+}
+
+static uint64_t vpe_get_csa_mc_addr(struct amdgpu_ring *ring, uint32_t vmid)
+{
+	struct amdgpu_device *adev = ring->adev;
+	uint32_t index = 0;
+	uint64_t csa_mc_addr;
+
+	if (amdgpu_sriov_vf(adev) || vmid == 0 || !adev->gfx.mcbp)
+		return 0;
+
+	csa_mc_addr = amdgpu_csa_vaddr(adev) + AMDGPU_CSA_VPE_OFFSET +
+		      index * AMDGPU_CSA_VPE_SIZE;
+
+	return csa_mc_addr;
+}
+
+static void vpe_ring_emit_pred_exec(struct amdgpu_ring *ring,
+				    uint32_t device_select,
+				    uint32_t exec_count)
+{
+	if (!ring->adev->vpe.collaborate_mode)
+		return;
+
+	amdgpu_ring_write(ring, VPE_CMD_HEADER(VPE_CMD_OPCODE_PRED_EXE, 0) |
+				(device_select << 16));
+	amdgpu_ring_write(ring, exec_count & 0x1fff);
+}
+
+static void vpe_ring_emit_ib(struct amdgpu_ring *ring,
+			     struct amdgpu_job *job,
+			     struct amdgpu_ib *ib,
+			     uint32_t flags)
+{
+	uint32_t vmid = AMDGPU_JOB_GET_VMID(job);
+	uint64_t csa_mc_addr = vpe_get_csa_mc_addr(ring, vmid);
+
+	amdgpu_ring_write(ring, VPE_CMD_HEADER(VPE_CMD_OPCODE_INDIRECT, 0) |
+				VPE_CMD_INDIRECT_HEADER_VMID(vmid & 0xf));
+
+	/* base must be 32 byte aligned */
+	amdgpu_ring_write(ring, ib->gpu_addr & 0xffffffe0);
+	amdgpu_ring_write(ring, upper_32_bits(ib->gpu_addr));
+	amdgpu_ring_write(ring, ib->length_dw);
+	amdgpu_ring_write(ring, lower_32_bits(csa_mc_addr));
+	amdgpu_ring_write(ring, upper_32_bits(csa_mc_addr));
+}
+
+static void vpe_ring_emit_fence(struct amdgpu_ring *ring, uint64_t addr,
+				uint64_t seq, unsigned int flags)
+{
+	int i = 0;
+
+	do {
+		/* write the fence */
+		amdgpu_ring_write(ring, VPE_CMD_HEADER(VPE_CMD_OPCODE_FENCE, 0));
+		/* zero in first two bits */
+		WARN_ON_ONCE(addr & 0x3);
+		amdgpu_ring_write(ring, lower_32_bits(addr));
+		amdgpu_ring_write(ring, upper_32_bits(addr));
+		amdgpu_ring_write(ring, i == 0 ? lower_32_bits(seq) : upper_32_bits(seq));
+		addr += 4;
+	} while ((flags & AMDGPU_FENCE_FLAG_64BIT) && (i++ < 1));
+
+	if (flags & AMDGPU_FENCE_FLAG_INT) {
+		/* generate an interrupt */
+		amdgpu_ring_write(ring, VPE_CMD_HEADER(VPE_CMD_OPCODE_TRAP, 0));
+		amdgpu_ring_write(ring, 0);
+	}
+
+}
+
+static void vpe_ring_emit_pipeline_sync(struct amdgpu_ring *ring)
+{
+	uint32_t seq = ring->fence_drv.sync_seq;
+	uint64_t addr = ring->fence_drv.gpu_addr;
+
+	vpe_ring_emit_pred_exec(ring, 0, 6);
+
+	/* wait for idle */
+	amdgpu_ring_write(ring, VPE_CMD_HEADER(VPE_CMD_OPCODE_POLL_REGMEM,
+				VPE_POLL_REGMEM_SUBOP_REGMEM) |
+				VPE_CMD_POLL_REGMEM_HEADER_FUNC(3) | /* equal */
+				VPE_CMD_POLL_REGMEM_HEADER_MEM(1));
+	amdgpu_ring_write(ring, addr & 0xfffffffc);
+	amdgpu_ring_write(ring, upper_32_bits(addr));
+	amdgpu_ring_write(ring, seq); /* reference */
+	amdgpu_ring_write(ring, 0xffffffff); /* mask */
+	amdgpu_ring_write(ring, VPE_CMD_POLL_REGMEM_DW5_RETRY_COUNT(0xfff) |
+				VPE_CMD_POLL_REGMEM_DW5_INTERVAL(4));
+}
+
+static void vpe_ring_emit_wreg(struct amdgpu_ring *ring, uint32_t reg, uint32_t val)
+{
+	vpe_ring_emit_pred_exec(ring, 0, 3);
+
+	amdgpu_ring_write(ring, VPE_CMD_HEADER(VPE_CMD_OPCODE_REG_WRITE, 0));
+	amdgpu_ring_write(ring,	reg << 2);
+	amdgpu_ring_write(ring, val);
+}
+
+static void vpe_ring_emit_reg_wait(struct amdgpu_ring *ring, uint32_t reg,
+				   uint32_t val, uint32_t mask)
+{
+	vpe_ring_emit_pred_exec(ring, 0, 6);
+
+	amdgpu_ring_write(ring, VPE_CMD_HEADER(VPE_CMD_OPCODE_POLL_REGMEM,
+				VPE_POLL_REGMEM_SUBOP_REGMEM) |
+				VPE_CMD_POLL_REGMEM_HEADER_FUNC(3) | /* equal */
+				VPE_CMD_POLL_REGMEM_HEADER_MEM(0));
+	amdgpu_ring_write(ring, reg << 2);
+	amdgpu_ring_write(ring, 0);
+	amdgpu_ring_write(ring, val); /* reference */
+	amdgpu_ring_write(ring, mask); /* mask */
+	amdgpu_ring_write(ring, VPE_CMD_POLL_REGMEM_DW5_RETRY_COUNT(0xfff) |
+				VPE_CMD_POLL_REGMEM_DW5_INTERVAL(10));
+}
+
+static void vpe_ring_emit_vm_flush(struct amdgpu_ring *ring, unsigned int vmid,
+				   uint64_t pd_addr)
+{
+	amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pd_addr);
+}
+
+static unsigned int vpe_ring_init_cond_exec(struct amdgpu_ring *ring,
+					    uint64_t addr)
+{
+	unsigned int ret;
+
+	amdgpu_ring_write(ring, VPE_CMD_HEADER(VPE_CMD_OPCODE_COND_EXE, 0));
+	amdgpu_ring_write(ring, lower_32_bits(addr));
+	amdgpu_ring_write(ring, upper_32_bits(addr));
+	amdgpu_ring_write(ring, 1);
+	ret = ring->wptr & ring->buf_mask;
+	amdgpu_ring_write(ring, 0);
+
+	return ret;
+}
+
+static int vpe_ring_preempt_ib(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct amdgpu_vpe *vpe = &adev->vpe;
+	uint32_t preempt_reg = vpe->regs.queue0_preempt;
+	int i, r = 0;
+
+	/* assert preemption condition */
+	amdgpu_ring_set_preempt_cond_exec(ring, false);
+
+	/* emit the trailing fence */
+	ring->trail_seq += 1;
+	amdgpu_ring_alloc(ring, 10);
+	vpe_ring_emit_fence(ring, ring->trail_fence_gpu_addr, ring->trail_seq, 0);
+	amdgpu_ring_commit(ring);
+
+	/* assert IB preemption */
+	WREG32(vpe_get_reg_offset(vpe, ring->me, preempt_reg), 1);
+
+	/* poll the trailing fence */
+	for (i = 0; i < adev->usec_timeout; i++) {
+		if (ring->trail_seq ==
+		    le32_to_cpu(*(ring->trail_fence_cpu_addr)))
+			break;
+		udelay(1);
+	}
+
+	if (i >= adev->usec_timeout) {
+		r = -EINVAL;
+		dev_err(adev->dev, "ring %d failed to be preempted\n", ring->idx);
+	}
+
+	/* deassert IB preemption */
+	WREG32(vpe_get_reg_offset(vpe, ring->me, preempt_reg), 0);
+
+	/* deassert the preemption condition */
+	amdgpu_ring_set_preempt_cond_exec(ring, true);
+
+	return r;
+}
+
+static int vpe_set_clockgating_state(struct amdgpu_ip_block *ip_block,
+				     enum amd_clockgating_state state)
+{
+	return 0;
+}
+
+static int vpe_set_powergating_state(struct amdgpu_ip_block *ip_block,
+				     enum amd_powergating_state state)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	struct amdgpu_vpe *vpe = &adev->vpe;
+
+	if (!adev->pm.dpm_enabled)
+		dev_err(adev->dev, "Without PM, cannot support powergating\n");
+
+	dev_dbg(adev->dev, "%s: %s!\n", __func__, (state == AMD_PG_STATE_GATE) ? "GATE":"UNGATE");
+
+	if (state == AMD_PG_STATE_GATE) {
+		amdgpu_dpm_enable_vpe(adev, false);
+		vpe->context_started = false;
+	} else {
+		amdgpu_dpm_enable_vpe(adev, true);
+	}
+
+	return 0;
+}
+
+static uint64_t vpe_ring_get_rptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct amdgpu_vpe *vpe = &adev->vpe;
+	uint64_t rptr;
+
+	if (ring->use_doorbell) {
+		rptr = atomic64_read((atomic64_t *)ring->rptr_cpu_addr);
+		dev_dbg(adev->dev, "rptr/doorbell before shift == 0x%016llx\n", rptr);
+	} else {
+		rptr = RREG32(vpe_get_reg_offset(vpe, ring->me, vpe->regs.queue0_rb_rptr_hi));
+		rptr = rptr << 32;
+		rptr |= RREG32(vpe_get_reg_offset(vpe, ring->me, vpe->regs.queue0_rb_rptr_lo));
+		dev_dbg(adev->dev, "rptr before shift [%i] == 0x%016llx\n", ring->me, rptr);
+	}
+
+	return (rptr >> 2);
+}
+
+static uint64_t vpe_ring_get_wptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct amdgpu_vpe *vpe = &adev->vpe;
+	uint64_t wptr;
+
+	if (ring->use_doorbell) {
+		wptr = atomic64_read((atomic64_t *)ring->wptr_cpu_addr);
+		dev_dbg(adev->dev, "wptr/doorbell before shift == 0x%016llx\n", wptr);
+	} else {
+		wptr = RREG32(vpe_get_reg_offset(vpe, ring->me, vpe->regs.queue0_rb_wptr_hi));
+		wptr = wptr << 32;
+		wptr |= RREG32(vpe_get_reg_offset(vpe, ring->me, vpe->regs.queue0_rb_wptr_lo));
+		dev_dbg(adev->dev, "wptr before shift [%i] == 0x%016llx\n", ring->me, wptr);
+	}
+
+	return (wptr >> 2);
+}
+
+static void vpe_ring_set_wptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct amdgpu_vpe *vpe = &adev->vpe;
+
+	if (ring->use_doorbell) {
+		dev_dbg(adev->dev, "Using doorbell, \
+			wptr_offs == 0x%08x, \
+			lower_32_bits(ring->wptr) << 2 == 0x%08x, \
+			upper_32_bits(ring->wptr) << 2 == 0x%08x\n",
+			ring->wptr_offs,
+			lower_32_bits(ring->wptr << 2),
+			upper_32_bits(ring->wptr << 2));
+		atomic64_set((atomic64_t *)ring->wptr_cpu_addr, ring->wptr << 2);
+		WDOORBELL64(ring->doorbell_index, ring->wptr << 2);
+		if (vpe->collaborate_mode)
+			WDOORBELL64(ring->doorbell_index + 4, ring->wptr << 2);
+	} else {
+		int i;
+
+		for (i = 0; i < vpe->num_instances; i++) {
+			dev_dbg(adev->dev, "Not using doorbell, \
+				regVPEC_QUEUE0_RB_WPTR == 0x%08x, \
+				regVPEC_QUEUE0_RB_WPTR_HI == 0x%08x\n",
+				lower_32_bits(ring->wptr << 2),
+				upper_32_bits(ring->wptr << 2));
+			WREG32(vpe_get_reg_offset(vpe, i, vpe->regs.queue0_rb_wptr_lo),
+			       lower_32_bits(ring->wptr << 2));
+			WREG32(vpe_get_reg_offset(vpe, i, vpe->regs.queue0_rb_wptr_hi),
+			       upper_32_bits(ring->wptr << 2));
+		}
+	}
+}
+
+static int vpe_ring_test_ring(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	const uint32_t test_pattern = 0xdeadbeef;
+	uint32_t index, i;
+	uint64_t wb_addr;
+	int ret;
+
+	ret = amdgpu_device_wb_get(adev, &index);
+	if (ret) {
+		dev_err(adev->dev, "(%d) failed to allocate wb slot\n", ret);
+		return ret;
+	}
+
+	adev->wb.wb[index] = 0;
+	wb_addr = adev->wb.gpu_addr + (index * 4);
+
+	ret = amdgpu_ring_alloc(ring, 4);
+	if (ret) {
+		dev_err(adev->dev, "amdgpu: dma failed to lock ring %d (%d).\n", ring->idx, ret);
+		goto out;
+	}
+
+	amdgpu_ring_write(ring, VPE_CMD_HEADER(VPE_CMD_OPCODE_FENCE, 0));
+	amdgpu_ring_write(ring, lower_32_bits(wb_addr));
+	amdgpu_ring_write(ring, upper_32_bits(wb_addr));
+	amdgpu_ring_write(ring, test_pattern);
+	amdgpu_ring_commit(ring);
+
+	for (i = 0; i < adev->usec_timeout; i++) {
+		if (le32_to_cpu(adev->wb.wb[index]) == test_pattern)
+			goto out;
+		udelay(1);
+	}
+
+	ret = -ETIMEDOUT;
+out:
+	amdgpu_device_wb_free(adev, index);
+
+	return ret;
+}
+
+static int vpe_ring_test_ib(struct amdgpu_ring *ring, long timeout)
+{
+	struct amdgpu_device *adev = ring->adev;
+	const uint32_t test_pattern = 0xdeadbeef;
+	struct amdgpu_ib ib = {};
+	struct dma_fence *f = NULL;
+	uint32_t index;
+	uint64_t wb_addr;
+	int ret;
+
+	ret = amdgpu_device_wb_get(adev, &index);
+	if (ret) {
+		dev_err(adev->dev, "(%d) failed to allocate wb slot\n", ret);
+		return ret;
+	}
+
+	adev->wb.wb[index] = 0;
+	wb_addr = adev->wb.gpu_addr + (index * 4);
+
+	ret = amdgpu_ib_get(adev, NULL, 256, AMDGPU_IB_POOL_DIRECT, &ib);
+	if (ret)
+		goto err0;
+
+	ib.ptr[0] = VPE_CMD_HEADER(VPE_CMD_OPCODE_FENCE, 0);
+	ib.ptr[1] = lower_32_bits(wb_addr);
+	ib.ptr[2] = upper_32_bits(wb_addr);
+	ib.ptr[3] = test_pattern;
+	ib.ptr[4] = VPE_CMD_HEADER(VPE_CMD_OPCODE_NOP, 0);
+	ib.ptr[5] = VPE_CMD_HEADER(VPE_CMD_OPCODE_NOP, 0);
+	ib.ptr[6] = VPE_CMD_HEADER(VPE_CMD_OPCODE_NOP, 0);
+	ib.ptr[7] = VPE_CMD_HEADER(VPE_CMD_OPCODE_NOP, 0);
+	ib.length_dw = 8;
+
+	ret = amdgpu_ib_schedule(ring, 1, &ib, NULL, &f);
+	if (ret)
+		goto err1;
+
+	ret = dma_fence_wait_timeout(f, false, timeout);
+	if (ret <= 0) {
+		ret = ret ? : -ETIMEDOUT;
+		goto err1;
+	}
+
+	ret = (le32_to_cpu(adev->wb.wb[index]) == test_pattern) ? 0 : -EINVAL;
+
+err1:
+	amdgpu_ib_free(&ib, NULL);
+	dma_fence_put(f);
+err0:
+	amdgpu_device_wb_free(adev, index);
+
+	return ret;
+}
+
+static void vpe_ring_begin_use(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct amdgpu_vpe *vpe = &adev->vpe;
+
+	cancel_delayed_work_sync(&adev->vpe.idle_work);
+
+	/* Power on VPE and notify VPE of new context  */
+	if (!vpe->context_started) {
+		uint32_t context_notify;
+
+		/* Power on VPE */
+		amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_VPE, AMD_PG_STATE_UNGATE);
+
+		/* Indicates that a job from a new context has been submitted. */
+		context_notify = RREG32(vpe_get_reg_offset(vpe, 0, vpe->regs.context_indicator));
+		if ((context_notify & 0x1) == 0)
+			context_notify |= 0x1;
+		else
+			context_notify &= ~(0x1);
+		WREG32(vpe_get_reg_offset(vpe, 0, vpe->regs.context_indicator), context_notify);
+		vpe->context_started = true;
+	}
+}
+
+static void vpe_ring_end_use(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	schedule_delayed_work(&adev->vpe.idle_work, VPE_IDLE_TIMEOUT);
+}
+
+static int vpe_ring_reset(struct amdgpu_ring *ring,
+			  unsigned int vmid,
+			  struct amdgpu_fence *timedout_fence)
+{
+	struct amdgpu_device *adev = ring->adev;
+	int r;
+
+	amdgpu_ring_reset_helper_begin(ring, timedout_fence);
+
+	r = amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_VPE,
+						   AMD_PG_STATE_GATE);
+	if (r)
+		return r;
+	r = amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_VPE,
+						   AMD_PG_STATE_UNGATE);
+	if (r)
+		return r;
+
+	return amdgpu_ring_reset_helper_end(ring, timedout_fence);
+}
+
+static ssize_t amdgpu_get_vpe_reset_mask(struct device *dev,
+						struct device_attribute *attr,
+						char *buf)
+{
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = drm_to_adev(ddev);
+
+	if (!adev)
+		return -ENODEV;
+
+	return amdgpu_show_reset_mask(buf, adev->vpe.supported_reset);
+}
+
+static DEVICE_ATTR(vpe_reset_mask, 0444,
+		   amdgpu_get_vpe_reset_mask, NULL);
+
+int amdgpu_vpe_sysfs_reset_mask_init(struct amdgpu_device *adev)
+{
+	int r = 0;
+
+	if (adev->vpe.num_instances) {
+		r = device_create_file(adev->dev, &dev_attr_vpe_reset_mask);
+		if (r)
+			return r;
+	}
+
+	return r;
+}
+
+void amdgpu_vpe_sysfs_reset_mask_fini(struct amdgpu_device *adev)
+{
+	if (adev->dev->kobj.sd) {
+		if (adev->vpe.num_instances)
+			device_remove_file(adev->dev, &dev_attr_vpe_reset_mask);
+	}
+}
+
+static const struct amdgpu_ring_funcs vpe_ring_funcs = {
+	.type = AMDGPU_RING_TYPE_VPE,
+	.align_mask = 0xf,
+	.nop = VPE_CMD_HEADER(VPE_CMD_OPCODE_NOP, 0),
+	.support_64bit_ptrs = true,
+	.get_rptr = vpe_ring_get_rptr,
+	.get_wptr = vpe_ring_get_wptr,
+	.set_wptr = vpe_ring_set_wptr,
+	.emit_frame_size =
+		5 + /* vpe_ring_init_cond_exec */
+		6 + /* vpe_ring_emit_pipeline_sync */
+		10 + 10 + 10 + /* vpe_ring_emit_fence */
+		/* vpe_ring_emit_vm_flush */
+		SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 +
+		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 6,
+	.emit_ib_size = 7 + 6,
+	.emit_ib = vpe_ring_emit_ib,
+	.emit_pipeline_sync = vpe_ring_emit_pipeline_sync,
+	.emit_fence = vpe_ring_emit_fence,
+	.emit_vm_flush = vpe_ring_emit_vm_flush,
+	.emit_wreg = vpe_ring_emit_wreg,
+	.emit_reg_wait = vpe_ring_emit_reg_wait,
+	.emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper,
+	.insert_nop = vpe_ring_insert_nop,
+	.pad_ib = amdgpu_ring_generic_pad_ib,
+	.test_ring = vpe_ring_test_ring,
+	.test_ib = vpe_ring_test_ib,
+	.init_cond_exec = vpe_ring_init_cond_exec,
+	.preempt_ib = vpe_ring_preempt_ib,
+	.begin_use = vpe_ring_begin_use,
+	.end_use = vpe_ring_end_use,
+	.reset = vpe_ring_reset,
+};
+
+static void vpe_set_ring_funcs(struct amdgpu_device *adev)
+{
+	adev->vpe.ring.funcs = &vpe_ring_funcs;
+}
+
+const struct amd_ip_funcs vpe_ip_funcs = {
+	.name = "vpe_v6_1",
+	.early_init = vpe_early_init,
+	.sw_init = vpe_sw_init,
+	.sw_fini = vpe_sw_fini,
+	.hw_init = vpe_hw_init,
+	.hw_fini = vpe_hw_fini,
+	.suspend = vpe_suspend,
+	.resume = vpe_resume,
+	.set_clockgating_state = vpe_set_clockgating_state,
+	.set_powergating_state = vpe_set_powergating_state,
+};
+
+const struct amdgpu_ip_block_version vpe_v6_1_ip_block = {
+	.type = AMD_IP_BLOCK_TYPE_VPE,
+	.major = 6,
+	.minor = 1,
+	.rev = 0,
+	.funcs = &vpe_ip_funcs,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vpe.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vpe.h
new file mode 100644
index 000000000000..695da740a97e
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vpe.h
@@ -0,0 +1,111 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#ifndef __AMDGPU_VPE_H__
+#define __AMDGPU_VPE_H__
+
+#include "amdgpu_ring.h"
+#include "amdgpu_irq.h"
+#include "vpe_6_1_fw_if.h"
+
+#define AMDGPU_MAX_VPE_INSTANCES 2
+
+struct amdgpu_vpe;
+
+struct vpe_funcs {
+	uint32_t (*get_reg_offset)(struct amdgpu_vpe *vpe, uint32_t inst, uint32_t offset);
+	int (*set_regs)(struct amdgpu_vpe *vpe);
+	int (*irq_init)(struct amdgpu_vpe *vpe);
+	int (*init_microcode)(struct amdgpu_vpe *vpe);
+	int (*load_microcode)(struct amdgpu_vpe *vpe);
+	int (*ring_init)(struct amdgpu_vpe *vpe);
+	int (*ring_start)(struct amdgpu_vpe *vpe);
+	int (*ring_stop)(struct amdgpu_vpe *vpe);
+	int (*ring_fini)(struct amdgpu_vpe *vpe);
+};
+
+struct vpe_regs {
+	uint32_t queue0_rb_rptr_lo;
+	uint32_t queue0_rb_rptr_hi;
+	uint32_t queue0_rb_wptr_lo;
+	uint32_t queue0_rb_wptr_hi;
+	uint32_t queue0_preempt;
+
+	uint32_t dpm_enable;
+	uint32_t dpm_pratio;
+	uint32_t dpm_request_interval;
+	uint32_t dpm_decision_threshold;
+	uint32_t dpm_busy_clamp_threshold;
+	uint32_t dpm_idle_clamp_threshold;
+	uint32_t dpm_request_lv;
+	uint32_t context_indicator;
+};
+
+struct amdgpu_vpe {
+	struct amdgpu_ring		ring;
+	struct amdgpu_irq_src		trap_irq;
+
+	const struct vpe_funcs		*funcs;
+	struct vpe_regs			regs;
+
+	const struct firmware		*fw;
+	uint32_t			fw_version;
+	uint32_t			feature_version;
+
+	struct amdgpu_bo		*cmdbuf_obj;
+	uint64_t			cmdbuf_gpu_addr;
+	uint32_t			*cmdbuf_cpu_addr;
+	struct delayed_work		idle_work;
+	bool				context_started;
+
+	uint32_t			num_instances;
+	bool				collaborate_mode;
+	uint32_t			supported_reset;
+};
+
+int amdgpu_vpe_psp_update_sram(struct amdgpu_device *adev);
+int amdgpu_vpe_init_microcode(struct amdgpu_vpe *vpe);
+int amdgpu_vpe_ring_init(struct amdgpu_vpe *vpe);
+int amdgpu_vpe_ring_fini(struct amdgpu_vpe *vpe);
+int amdgpu_vpe_configure_dpm(struct amdgpu_vpe *vpe);
+void amdgpu_vpe_sysfs_reset_mask_fini(struct amdgpu_device *adev);
+int amdgpu_vpe_sysfs_reset_mask_init(struct amdgpu_device *adev);
+
+#define vpe_ring_init(vpe) ((vpe)->funcs->ring_init ? (vpe)->funcs->ring_init((vpe)) : 0)
+#define vpe_ring_start(vpe) ((vpe)->funcs->ring_start ? (vpe)->funcs->ring_start((vpe)) : 0)
+#define vpe_ring_stop(vpe) ((vpe)->funcs->ring_stop ? (vpe)->funcs->ring_stop((vpe)) : 0)
+#define vpe_ring_fini(vpe) ((vpe)->funcs->ring_fini ? (vpe)->funcs->ring_fini((vpe)) : 0)
+
+#define vpe_get_reg_offset(vpe, inst, offset) \
+		((vpe)->funcs->get_reg_offset ? (vpe)->funcs->get_reg_offset((vpe), (inst), (offset)) : 0)
+#define vpe_set_regs(vpe) \
+		((vpe)->funcs->set_regs ? (vpe)->funcs->set_regs((vpe)) : 0)
+#define vpe_irq_init(vpe) \
+		((vpe)->funcs->irq_init ? (vpe)->funcs->irq_init((vpe)) : 0)
+#define vpe_init_microcode(vpe) \
+		((vpe)->funcs->init_microcode ? (vpe)->funcs->init_microcode((vpe)) : 0)
+#define vpe_load_microcode(vpe) \
+		((vpe)->funcs->load_microcode ? (vpe)->funcs->load_microcode((vpe)) : 0)
+
+extern const struct amdgpu_ip_block_version vpe_v6_1_ip_block;
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c
index 7b2b0980ec41..9d934c07fa6b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c
@@ -24,16 +24,20 @@
 
 #include <linux/dma-mapping.h>
 #include <drm/ttm/ttm_range_manager.h>
+#include <drm/drm_drv.h>
 
 #include "amdgpu.h"
 #include "amdgpu_vm.h"
 #include "amdgpu_res_cursor.h"
-#include "amdgpu_atomfirmware.h"
 #include "atom.h"
 
+#define AMDGPU_MAX_SG_SEGMENT_SIZE	(2UL << 30)
+
 struct amdgpu_vram_reservation {
-	struct list_head node;
-	struct drm_mm_node mm_node;
+	u64 start;
+	u64 size;
+	struct list_head allocated;
+	struct list_head blocks;
 };
 
 static inline struct amdgpu_vram_mgr *
@@ -48,6 +52,44 @@ to_amdgpu_device(struct amdgpu_vram_mgr *mgr)
 	return container_of(mgr, struct amdgpu_device, mman.vram_mgr);
 }
 
+static inline struct drm_buddy_block *
+amdgpu_vram_mgr_first_block(struct list_head *list)
+{
+	return list_first_entry_or_null(list, struct drm_buddy_block, link);
+}
+
+static inline bool amdgpu_is_vram_mgr_blocks_contiguous(struct list_head *head)
+{
+	struct drm_buddy_block *block;
+	u64 start, size;
+
+	block = amdgpu_vram_mgr_first_block(head);
+	if (!block)
+		return false;
+
+	while (head != block->link.next) {
+		start = amdgpu_vram_mgr_block_start(block);
+		size = amdgpu_vram_mgr_block_size(block);
+
+		block = list_entry(block->link.next, struct drm_buddy_block, link);
+		if (start + size != amdgpu_vram_mgr_block_start(block))
+			return false;
+	}
+
+	return true;
+}
+
+static inline u64 amdgpu_vram_mgr_blocks_size(struct list_head *head)
+{
+	struct drm_buddy_block *block;
+	u64 size = 0;
+
+	list_for_each_entry(block, head, link)
+		size += amdgpu_vram_mgr_block_size(block);
+
+	return size;
+}
+
 /**
  * DOC: mem_info_vram_total
  *
@@ -96,10 +138,9 @@ static ssize_t amdgpu_mem_info_vram_used_show(struct device *dev,
 {
 	struct drm_device *ddev = dev_get_drvdata(dev);
 	struct amdgpu_device *adev = drm_to_adev(ddev);
-	struct ttm_resource_manager *man;
+	struct ttm_resource_manager *man = &adev->mman.vram_mgr.manager;
 
-	man = ttm_manager_type(&adev->mman.bdev, TTM_PL_VRAM);
-	return sysfs_emit(buf, "%llu\n", amdgpu_vram_mgr_usage(man));
+	return sysfs_emit(buf, "%llu\n", ttm_resource_manager_usage(man));
 }
 
 /**
@@ -116,10 +157,9 @@ static ssize_t amdgpu_mem_info_vis_vram_used_show(struct device *dev,
 {
 	struct drm_device *ddev = dev_get_drvdata(dev);
 	struct amdgpu_device *adev = drm_to_adev(ddev);
-	struct ttm_resource_manager *man;
 
-	man = ttm_manager_type(&adev->mman.bdev, TTM_PL_VRAM);
-	return sysfs_emit(buf, "%llu\n", amdgpu_vram_mgr_vis_usage(man));
+	return sysfs_emit(buf, "%llu\n",
+			  amdgpu_vram_mgr_vis_usage(&adev->mman.vram_mgr));
 }
 
 /**
@@ -183,23 +223,41 @@ static struct attribute *amdgpu_vram_mgr_attributes[] = {
 	NULL
 };
 
+static umode_t amdgpu_vram_attrs_is_visible(struct kobject *kobj,
+					    struct attribute *attr, int i)
+{
+	struct device *dev = kobj_to_dev(kobj);
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = drm_to_adev(ddev);
+
+	if (attr == &dev_attr_mem_info_vram_vendor.attr &&
+	    !adev->gmc.vram_vendor)
+		return 0;
+
+	if (!ttm_resource_manager_used(&adev->mman.vram_mgr.manager))
+		return 0;
+
+	return attr->mode;
+}
+
 const struct attribute_group amdgpu_vram_mgr_attr_group = {
-	.attrs = amdgpu_vram_mgr_attributes
+	.attrs = amdgpu_vram_mgr_attributes,
+	.is_visible = amdgpu_vram_attrs_is_visible
 };
 
 /**
- * amdgpu_vram_mgr_vis_size - Calculate visible node size
+ * amdgpu_vram_mgr_vis_size - Calculate visible block size
  *
  * @adev: amdgpu_device pointer
- * @node: MM node structure
+ * @block: DRM BUDDY block structure
  *
- * Calculate how many bytes of the MM node are inside visible VRAM
+ * Calculate how many bytes of the DRM BUDDY block are inside visible VRAM
  */
 static u64 amdgpu_vram_mgr_vis_size(struct amdgpu_device *adev,
-				    struct drm_mm_node *node)
+				    struct drm_buddy_block *block)
 {
-	uint64_t start = node->start << PAGE_SHIFT;
-	uint64_t end = (node->size + node->start) << PAGE_SHIFT;
+	u64 start = amdgpu_vram_mgr_block_start(block);
+	u64 end = start + amdgpu_vram_mgr_block_size(block);
 
 	if (start >= adev->gmc.visible_vram_size)
 		return 0;
@@ -220,9 +278,9 @@ u64 amdgpu_vram_mgr_bo_visible_size(struct amdgpu_bo *bo)
 {
 	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
 	struct ttm_resource *res = bo->tbo.resource;
-	unsigned pages = res->num_pages;
-	struct drm_mm_node *mm;
-	u64 usage;
+	struct amdgpu_vram_mgr_resource *vres = to_amdgpu_vram_mgr_resource(res);
+	struct drm_buddy_block *block;
+	u64 usage = 0;
 
 	if (amdgpu_gmc_vram_full_visible(&adev->gmc))
 		return amdgpu_bo_size(bo);
@@ -230,9 +288,8 @@ u64 amdgpu_vram_mgr_bo_visible_size(struct amdgpu_bo *bo)
 	if (res->start >= adev->gmc.visible_vram_size >> PAGE_SHIFT)
 		return 0;
 
-	mm = &container_of(res, struct ttm_range_mgr_node, base)->mm_nodes[0];
-	for (usage = 0; pages; pages -= mm->size, mm++)
-		usage += amdgpu_vram_mgr_vis_size(adev, mm);
+	list_for_each_entry(block, &vres->blocks, link)
+		usage += amdgpu_vram_mgr_vis_size(adev, block);
 
 	return usage;
 }
@@ -242,51 +299,61 @@ static void amdgpu_vram_mgr_do_reserve(struct ttm_resource_manager *man)
 {
 	struct amdgpu_vram_mgr *mgr = to_vram_mgr(man);
 	struct amdgpu_device *adev = to_amdgpu_device(mgr);
-	struct drm_mm *mm = &mgr->mm;
+	struct drm_buddy *mm = &mgr->mm;
 	struct amdgpu_vram_reservation *rsv, *temp;
+	struct drm_buddy_block *block;
 	uint64_t vis_usage;
 
-	list_for_each_entry_safe(rsv, temp, &mgr->reservations_pending, node) {
-		if (drm_mm_reserve_node(mm, &rsv->mm_node))
+	list_for_each_entry_safe(rsv, temp, &mgr->reservations_pending, blocks) {
+		if (drm_buddy_alloc_blocks(mm, rsv->start, rsv->start + rsv->size,
+					   rsv->size, mm->chunk_size, &rsv->allocated,
+					   DRM_BUDDY_RANGE_ALLOCATION))
+			continue;
+
+		block = amdgpu_vram_mgr_first_block(&rsv->allocated);
+		if (!block)
 			continue;
 
 		dev_dbg(adev->dev, "Reservation 0x%llx - %lld, Succeeded\n",
-			rsv->mm_node.start, rsv->mm_node.size);
+			rsv->start, rsv->size);
 
-		vis_usage = amdgpu_vram_mgr_vis_size(adev, &rsv->mm_node);
+		vis_usage = amdgpu_vram_mgr_vis_size(adev, block);
 		atomic64_add(vis_usage, &mgr->vis_usage);
-		atomic64_add(rsv->mm_node.size << PAGE_SHIFT, &mgr->usage);
-		list_move(&rsv->node, &mgr->reserved_pages);
+		spin_lock(&man->bdev->lru_lock);
+		man->usage += rsv->size;
+		spin_unlock(&man->bdev->lru_lock);
+		list_move(&rsv->blocks, &mgr->reserved_pages);
 	}
 }
 
 /**
  * amdgpu_vram_mgr_reserve_range - Reserve a range from VRAM
  *
- * @man: TTM memory type manager
+ * @mgr: amdgpu_vram_mgr pointer
  * @start: start address of the range in VRAM
  * @size: size of the range
  *
- * Reserve memory from start addess with the specified size in VRAM
+ * Reserve memory from start address with the specified size in VRAM
  */
-int amdgpu_vram_mgr_reserve_range(struct ttm_resource_manager *man,
+int amdgpu_vram_mgr_reserve_range(struct amdgpu_vram_mgr *mgr,
 				  uint64_t start, uint64_t size)
 {
-	struct amdgpu_vram_mgr *mgr = to_vram_mgr(man);
 	struct amdgpu_vram_reservation *rsv;
 
 	rsv = kzalloc(sizeof(*rsv), GFP_KERNEL);
 	if (!rsv)
 		return -ENOMEM;
 
-	INIT_LIST_HEAD(&rsv->node);
-	rsv->mm_node.start = start >> PAGE_SHIFT;
-	rsv->mm_node.size = size >> PAGE_SHIFT;
+	INIT_LIST_HEAD(&rsv->allocated);
+	INIT_LIST_HEAD(&rsv->blocks);
 
-	spin_lock(&mgr->lock);
-	list_add_tail(&mgr->reservations_pending, &rsv->node);
-	amdgpu_vram_mgr_do_reserve(man);
-	spin_unlock(&mgr->lock);
+	rsv->start = start;
+	rsv->size = size;
+
+	mutex_lock(&mgr->lock);
+	list_add_tail(&rsv->blocks, &mgr->reservations_pending);
+	amdgpu_vram_mgr_do_reserve(&mgr->manager);
+	mutex_unlock(&mgr->lock);
 
 	return 0;
 }
@@ -294,7 +361,7 @@ int amdgpu_vram_mgr_reserve_range(struct ttm_resource_manager *man,
 /**
  * amdgpu_vram_mgr_query_page_status - query the reservation status
  *
- * @man: TTM memory type manager
+ * @mgr: amdgpu_vram_mgr pointer
  * @start: start address of a page in VRAM
  *
  * Returns:
@@ -302,26 +369,25 @@ int amdgpu_vram_mgr_reserve_range(struct ttm_resource_manager *man,
  *	0: the page has been reserved
  *	-ENOENT: the input page is not a reservation
  */
-int amdgpu_vram_mgr_query_page_status(struct ttm_resource_manager *man,
+int amdgpu_vram_mgr_query_page_status(struct amdgpu_vram_mgr *mgr,
 				      uint64_t start)
 {
-	struct amdgpu_vram_mgr *mgr = to_vram_mgr(man);
 	struct amdgpu_vram_reservation *rsv;
 	int ret;
 
-	spin_lock(&mgr->lock);
+	mutex_lock(&mgr->lock);
 
-	list_for_each_entry(rsv, &mgr->reservations_pending, node) {
-		if ((rsv->mm_node.start <= start) &&
-		    (start < (rsv->mm_node.start + rsv->mm_node.size))) {
+	list_for_each_entry(rsv, &mgr->reservations_pending, blocks) {
+		if (rsv->start <= start &&
+		    (start < (rsv->start + rsv->size))) {
 			ret = -EBUSY;
 			goto out;
 		}
 	}
 
-	list_for_each_entry(rsv, &mgr->reserved_pages, node) {
-		if ((rsv->mm_node.start <= start) &&
-		    (start < (rsv->mm_node.start + rsv->mm_node.size))) {
+	list_for_each_entry(rsv, &mgr->reserved_pages, blocks) {
+		if (rsv->start <= start &&
+		    (start < (rsv->start + rsv->size))) {
 			ret = 0;
 			goto out;
 		}
@@ -329,30 +395,37 @@ int amdgpu_vram_mgr_query_page_status(struct ttm_resource_manager *man,
 
 	ret = -ENOENT;
 out:
-	spin_unlock(&mgr->lock);
+	mutex_unlock(&mgr->lock);
 	return ret;
 }
 
-/**
- * amdgpu_vram_mgr_virt_start - update virtual start address
- *
- * @mem: ttm_resource to update
- * @node: just allocated node
- *
- * Calculate a virtual BO start address to easily check if everything is CPU
- * accessible.
- */
-static void amdgpu_vram_mgr_virt_start(struct ttm_resource *mem,
-				       struct drm_mm_node *node)
+int amdgpu_vram_mgr_query_address_block_info(struct amdgpu_vram_mgr *mgr,
+			uint64_t address, struct amdgpu_vram_block_info *info)
 {
-	unsigned long start;
+	struct amdgpu_vram_mgr_resource *vres;
+	struct drm_buddy_block *block;
+	u64 start, size;
+	int ret = -ENOENT;
+
+	mutex_lock(&mgr->lock);
+	list_for_each_entry(vres, &mgr->allocated_vres_list, vres_node) {
+		list_for_each_entry(block, &vres->blocks, link) {
+			start = amdgpu_vram_mgr_block_start(block);
+			size = amdgpu_vram_mgr_block_size(block);
+			if ((start <= address) && (address < (start + size))) {
+				info->start = start;
+				info->size = size;
+				memcpy(&info->task, &vres->task, sizeof(vres->task));
+				ret = 0;
+				goto out;
+			}
+		}
+	}
 
-	start = node->start + node->size;
-	if (start > mem->num_pages)
-		start -= mem->num_pages;
-	else
-		start = 0;
-	mem->start = max(mem->start, start);
+out:
+	mutex_unlock(&mgr->lock);
+
+	return ret;
 }
 
 /**
@@ -370,116 +443,183 @@ static int amdgpu_vram_mgr_new(struct ttm_resource_manager *man,
 			       const struct ttm_place *place,
 			       struct ttm_resource **res)
 {
-	unsigned long lpfn, num_nodes, pages_per_node, pages_left, pages;
 	struct amdgpu_vram_mgr *mgr = to_vram_mgr(man);
 	struct amdgpu_device *adev = to_amdgpu_device(mgr);
-	uint64_t vis_usage = 0, mem_bytes, max_bytes;
-	struct ttm_range_mgr_node *node;
-	struct drm_mm *mm = &mgr->mm;
-	enum drm_mm_insert_mode mode;
-	unsigned i;
+	struct amdgpu_bo *bo = ttm_to_amdgpu_bo(tbo);
+	u64 vis_usage = 0, max_bytes, min_block_size;
+	struct amdgpu_vram_mgr_resource *vres;
+	u64 size, remaining_size, lpfn, fpfn;
+	unsigned int adjust_dcc_size = 0;
+	struct drm_buddy *mm = &mgr->mm;
+	struct drm_buddy_block *block;
+	unsigned long pages_per_block;
 	int r;
 
-	lpfn = place->lpfn;
-	if (!lpfn)
+	lpfn = (u64)place->lpfn << PAGE_SHIFT;
+	if (!lpfn || lpfn > man->size)
 		lpfn = man->size;
 
+	fpfn = (u64)place->fpfn << PAGE_SHIFT;
+
 	max_bytes = adev->gmc.mc_vram_size;
 	if (tbo->type != ttm_bo_type_kernel)
 		max_bytes -= AMDGPU_VM_RESERVED_VRAM;
 
-	/* bail out quickly if there's likely not enough VRAM for this BO */
-	mem_bytes = tbo->base.size;
-	if (atomic64_add_return(mem_bytes, &mgr->usage) > max_bytes) {
-		r = -ENOSPC;
-		goto error_sub;
-	}
-
-	if (place->flags & TTM_PL_FLAG_CONTIGUOUS) {
-		pages_per_node = ~0ul;
-		num_nodes = 1;
+	if (bo->flags & AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS) {
+		pages_per_block = ~0ul;
 	} else {
 #ifdef CONFIG_TRANSPARENT_HUGEPAGE
-		pages_per_node = HPAGE_PMD_NR;
+		pages_per_block = HPAGE_PMD_NR;
 #else
 		/* default to 2MB */
-		pages_per_node = 2UL << (20UL - PAGE_SHIFT);
+		pages_per_block = 2UL << (20UL - PAGE_SHIFT);
 #endif
-		pages_per_node = max_t(uint32_t, pages_per_node,
-				       tbo->page_alignment);
-		num_nodes = DIV_ROUND_UP_ULL(PFN_UP(mem_bytes), pages_per_node);
+		pages_per_block = max_t(u32, pages_per_block,
+					tbo->page_alignment);
 	}
 
-	node = kvmalloc(struct_size(node, mm_nodes, num_nodes),
-			GFP_KERNEL | __GFP_ZERO);
-	if (!node) {
-		r = -ENOMEM;
-		goto error_sub;
+	vres = kzalloc(sizeof(*vres), GFP_KERNEL);
+	if (!vres)
+		return -ENOMEM;
+
+	ttm_resource_init(tbo, place, &vres->base);
+
+	/* bail out quickly if there's likely not enough VRAM for this BO */
+	if (ttm_resource_manager_usage(man) > max_bytes) {
+		r = -ENOSPC;
+		goto error_fini;
 	}
 
-	ttm_resource_init(tbo, place, &node->base);
+	INIT_LIST_HEAD(&vres->blocks);
 
-	mode = DRM_MM_INSERT_BEST;
 	if (place->flags & TTM_PL_FLAG_TOPDOWN)
-		mode = DRM_MM_INSERT_HIGH;
-
-	pages_left = node->base.num_pages;
-
-	/* Limit maximum size to 2GB due to SG table limitations */
-	pages = min(pages_left, 2UL << (30 - PAGE_SHIFT));
-
-	i = 0;
-	spin_lock(&mgr->lock);
-	while (pages_left) {
-		uint32_t alignment = tbo->page_alignment;
-
-		if (pages >= pages_per_node)
-			alignment = pages_per_node;
-
-		r = drm_mm_insert_node_in_range(mm, &node->mm_nodes[i], pages,
-						alignment, 0, place->fpfn,
-						lpfn, mode);
-		if (unlikely(r)) {
-			if (pages > pages_per_node) {
-				if (is_power_of_2(pages))
-					pages = pages / 2;
-				else
-					pages = rounddown_pow_of_two(pages);
-				continue;
-			}
-			goto error_free;
+		vres->flags |= DRM_BUDDY_TOPDOWN_ALLOCATION;
+
+	if (bo->flags & AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS)
+		vres->flags |= DRM_BUDDY_CONTIGUOUS_ALLOCATION;
+
+	if (bo->flags & AMDGPU_GEM_CREATE_VRAM_CLEARED)
+		vres->flags |= DRM_BUDDY_CLEAR_ALLOCATION;
+
+	if (fpfn || lpfn != mgr->mm.size)
+		/* Allocate blocks in desired range */
+		vres->flags |= DRM_BUDDY_RANGE_ALLOCATION;
+
+	if (bo->flags & AMDGPU_GEM_CREATE_GFX12_DCC &&
+	    adev->gmc.gmc_funcs->get_dcc_alignment)
+		adjust_dcc_size = amdgpu_gmc_get_dcc_alignment(adev);
+
+	remaining_size = (u64)vres->base.size;
+	if (bo->flags & AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS && adjust_dcc_size) {
+		unsigned int dcc_size;
+
+		dcc_size = roundup_pow_of_two(vres->base.size + adjust_dcc_size);
+		remaining_size = (u64)dcc_size;
+
+		vres->flags |= DRM_BUDDY_TRIM_DISABLE;
+	}
+
+	mutex_lock(&mgr->lock);
+	while (remaining_size) {
+		if (tbo->page_alignment)
+			min_block_size = (u64)tbo->page_alignment << PAGE_SHIFT;
+		else
+			min_block_size = mgr->default_page_size;
+
+		size = remaining_size;
+
+		if (bo->flags & AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS && adjust_dcc_size)
+			min_block_size = size;
+		else if ((size >= (u64)pages_per_block << PAGE_SHIFT) &&
+			 !(size & (((u64)pages_per_block << PAGE_SHIFT) - 1)))
+			min_block_size = (u64)pages_per_block << PAGE_SHIFT;
+
+		BUG_ON(min_block_size < mm->chunk_size);
+
+		r = drm_buddy_alloc_blocks(mm, fpfn,
+					   lpfn,
+					   size,
+					   min_block_size,
+					   &vres->blocks,
+					   vres->flags);
+
+		if (unlikely(r == -ENOSPC) && pages_per_block == ~0ul &&
+		    !(place->flags & TTM_PL_FLAG_CONTIGUOUS)) {
+			vres->flags &= ~DRM_BUDDY_CONTIGUOUS_ALLOCATION;
+			pages_per_block = max_t(u32, 2UL << (20UL - PAGE_SHIFT),
+						tbo->page_alignment);
+
+			continue;
 		}
 
-		vis_usage += amdgpu_vram_mgr_vis_size(adev, &node->mm_nodes[i]);
-		amdgpu_vram_mgr_virt_start(&node->base, &node->mm_nodes[i]);
-		pages_left -= pages;
-		++i;
+		if (unlikely(r))
+			goto error_free_blocks;
+
+		if (size > remaining_size)
+			remaining_size = 0;
+		else
+			remaining_size -= size;
+	}
+
+	vres->task.pid = task_pid_nr(current);
+	get_task_comm(vres->task.comm, current);
+	list_add_tail(&vres->vres_node, &mgr->allocated_vres_list);
+
+	if (bo->flags & AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS && adjust_dcc_size) {
+		struct drm_buddy_block *dcc_block;
+		unsigned long dcc_start;
+		u64 trim_start;
+
+		dcc_block = amdgpu_vram_mgr_first_block(&vres->blocks);
+		/* Adjust the start address for DCC buffers only */
+		dcc_start =
+			roundup((unsigned long)amdgpu_vram_mgr_block_start(dcc_block),
+				adjust_dcc_size);
+		trim_start = (u64)dcc_start;
+		drm_buddy_block_trim(mm, &trim_start,
+				     (u64)vres->base.size,
+				     &vres->blocks);
+	}
+	mutex_unlock(&mgr->lock);
+
+	vres->base.start = 0;
+	size = max_t(u64, amdgpu_vram_mgr_blocks_size(&vres->blocks),
+		     vres->base.size);
+	list_for_each_entry(block, &vres->blocks, link) {
+		unsigned long start;
 
-		if (pages > pages_left)
-			pages = pages_left;
+		start = amdgpu_vram_mgr_block_start(block) +
+			amdgpu_vram_mgr_block_size(block);
+		start >>= PAGE_SHIFT;
+
+		if (start > PFN_UP(size))
+			start -= PFN_UP(size);
+		else
+			start = 0;
+		vres->base.start = max(vres->base.start, start);
+
+		vis_usage += amdgpu_vram_mgr_vis_size(adev, block);
 	}
-	spin_unlock(&mgr->lock);
 
-	if (i == 1)
-		node->base.placement |= TTM_PL_FLAG_CONTIGUOUS;
+	if (amdgpu_is_vram_mgr_blocks_contiguous(&vres->blocks))
+		vres->base.placement |= TTM_PL_FLAG_CONTIGUOUS;
 
 	if (adev->gmc.xgmi.connected_to_cpu)
-		node->base.bus.caching = ttm_cached;
+		vres->base.bus.caching = ttm_cached;
 	else
-		node->base.bus.caching = ttm_write_combined;
+		vres->base.bus.caching = ttm_write_combined;
 
 	atomic64_add(vis_usage, &mgr->vis_usage);
-	*res = &node->base;
+	*res = &vres->base;
 	return 0;
 
-error_free:
-	while (i--)
-		drm_mm_remove_node(&node->mm_nodes[i]);
-	spin_unlock(&mgr->lock);
-	kvfree(node);
+error_free_blocks:
+	drm_buddy_free_list(mm, &vres->blocks, 0);
+	mutex_unlock(&mgr->lock);
+error_fini:
+	ttm_resource_fini(man, &vres->base);
+	kfree(vres);
 
-error_sub:
-	atomic64_sub(mem_bytes, &mgr->usage);
 	return r;
 }
 
@@ -494,28 +634,29 @@ error_sub:
 static void amdgpu_vram_mgr_del(struct ttm_resource_manager *man,
 				struct ttm_resource *res)
 {
-	struct ttm_range_mgr_node *node = to_ttm_range_mgr_node(res);
+	struct amdgpu_vram_mgr_resource *vres = to_amdgpu_vram_mgr_resource(res);
 	struct amdgpu_vram_mgr *mgr = to_vram_mgr(man);
 	struct amdgpu_device *adev = to_amdgpu_device(mgr);
-	uint64_t usage = 0, vis_usage = 0;
-	unsigned i, pages;
+	struct drm_buddy *mm = &mgr->mm;
+	struct drm_buddy_block *block;
+	uint64_t vis_usage = 0;
 
-	spin_lock(&mgr->lock);
-	for (i = 0, pages = res->num_pages; pages;
-	     pages -= node->mm_nodes[i].size, ++i) {
-		struct drm_mm_node *mm = &node->mm_nodes[i];
+	mutex_lock(&mgr->lock);
 
-		drm_mm_remove_node(mm);
-		usage += mm->size << PAGE_SHIFT;
-		vis_usage += amdgpu_vram_mgr_vis_size(adev, mm);
-	}
+	list_del(&vres->vres_node);
+	memset(&vres->task, 0, sizeof(vres->task));
+
+	list_for_each_entry(block, &vres->blocks, link)
+		vis_usage += amdgpu_vram_mgr_vis_size(adev, block);
+
+	drm_buddy_free_list(mm, &vres->blocks, vres->flags);
 	amdgpu_vram_mgr_do_reserve(man);
-	spin_unlock(&mgr->lock);
+	mutex_unlock(&mgr->lock);
 
-	atomic64_sub(usage, &mgr->usage);
 	atomic64_sub(vis_usage, &mgr->vis_usage);
 
-	kvfree(node);
+	ttm_resource_fini(man, res);
+	kfree(vres);
 }
 
 /**
@@ -547,11 +688,11 @@ int amdgpu_vram_mgr_alloc_sgt(struct amdgpu_device *adev,
 	if (!*sgt)
 		return -ENOMEM;
 
-	/* Determine the number of DRM_MM nodes to export */
+	/* Determine the number of DRM_BUDDY blocks to export */
 	amdgpu_res_first(res, offset, length, &cursor);
 	while (cursor.remaining) {
 		num_entries++;
-		amdgpu_res_next(&cursor, cursor.size);
+		amdgpu_res_next(&cursor, min(cursor.size, AMDGPU_MAX_SG_SEGMENT_SIZE));
 	}
 
 	r = sg_alloc_table(*sgt, num_entries, GFP_KERNEL);
@@ -563,15 +704,15 @@ int amdgpu_vram_mgr_alloc_sgt(struct amdgpu_device *adev,
 		sg->length = 0;
 
 	/*
-	 * Walk down DRM_MM nodes to populate scatterlist nodes
-	 * @note: Use iterator api to get first the DRM_MM node
+	 * Walk down DRM_BUDDY blocks to populate scatterlist nodes
+	 * @note: Use iterator api to get first the DRM_BUDDY block
 	 * and the number of bytes from it. Access the following
-	 * DRM_MM node(s) if more buffer needs to exported
+	 * DRM_BUDDY block(s) if more buffer needs to exported
 	 */
 	amdgpu_res_first(res, offset, length, &cursor);
 	for_each_sgtable_sg((*sgt), sg, i) {
 		phys_addr_t phys = cursor.start + adev->gmc.aper_base;
-		size_t size = cursor.size;
+		unsigned long size = min(cursor.size, AMDGPU_MAX_SG_SEGMENT_SIZE);
 		dma_addr_t addr;
 
 		addr = dma_map_resource(dev, phys, size, dir,
@@ -584,7 +725,7 @@ int amdgpu_vram_mgr_alloc_sgt(struct amdgpu_device *adev,
 		sg_dma_address(sg) = addr;
 		sg_dma_len(sg) = size;
 
-		amdgpu_res_next(&cursor, cursor.size);
+		amdgpu_res_next(&cursor, size);
 	}
 
 	return 0;
@@ -630,31 +771,98 @@ void amdgpu_vram_mgr_free_sgt(struct device *dev,
 }
 
 /**
- * amdgpu_vram_mgr_usage - how many bytes are used in this domain
+ * amdgpu_vram_mgr_vis_usage - how many bytes are used in the visible part
+ *
+ * @mgr: amdgpu_vram_mgr pointer
+ *
+ * Returns how many bytes are used in the visible part of VRAM
+ */
+uint64_t amdgpu_vram_mgr_vis_usage(struct amdgpu_vram_mgr *mgr)
+{
+	return atomic64_read(&mgr->vis_usage);
+}
+
+/**
+ * amdgpu_vram_mgr_clear_reset_blocks - reset clear blocks
+ *
+ * @adev: amdgpu device pointer
+ *
+ * Reset the cleared drm buddy blocks.
+ */
+void amdgpu_vram_mgr_clear_reset_blocks(struct amdgpu_device *adev)
+{
+	struct amdgpu_vram_mgr *mgr = &adev->mman.vram_mgr;
+	struct drm_buddy *mm = &mgr->mm;
+
+	mutex_lock(&mgr->lock);
+	drm_buddy_reset_clear(mm, false);
+	mutex_unlock(&mgr->lock);
+}
+
+/**
+ * amdgpu_vram_mgr_intersects - test each drm buddy block for intersection
  *
  * @man: TTM memory type manager
+ * @res: The resource to test
+ * @place: The place to test against
+ * @size: Size of the new allocation
  *
- * Returns how many bytes are used in this domain.
+ * Test each drm buddy block for intersection for eviction decision.
  */
-uint64_t amdgpu_vram_mgr_usage(struct ttm_resource_manager *man)
+static bool amdgpu_vram_mgr_intersects(struct ttm_resource_manager *man,
+				       struct ttm_resource *res,
+				       const struct ttm_place *place,
+				       size_t size)
 {
-	struct amdgpu_vram_mgr *mgr = to_vram_mgr(man);
+	struct amdgpu_vram_mgr_resource *mgr = to_amdgpu_vram_mgr_resource(res);
+	struct drm_buddy_block *block;
+
+	/* Check each drm buddy block individually */
+	list_for_each_entry(block, &mgr->blocks, link) {
+		unsigned long fpfn =
+			amdgpu_vram_mgr_block_start(block) >> PAGE_SHIFT;
+		unsigned long lpfn = fpfn +
+			(amdgpu_vram_mgr_block_size(block) >> PAGE_SHIFT);
+
+		if (place->fpfn < lpfn &&
+		    (!place->lpfn || place->lpfn > fpfn))
+			return true;
+	}
 
-	return atomic64_read(&mgr->usage);
+	return false;
 }
 
 /**
- * amdgpu_vram_mgr_vis_usage - how many bytes are used in the visible part
+ * amdgpu_vram_mgr_compatible - test each drm buddy block for compatibility
  *
  * @man: TTM memory type manager
+ * @res: The resource to test
+ * @place: The place to test against
+ * @size: Size of the new allocation
  *
- * Returns how many bytes are used in the visible part of VRAM
+ * Test each drm buddy block for placement compatibility.
  */
-uint64_t amdgpu_vram_mgr_vis_usage(struct ttm_resource_manager *man)
+static bool amdgpu_vram_mgr_compatible(struct ttm_resource_manager *man,
+				       struct ttm_resource *res,
+				       const struct ttm_place *place,
+				       size_t size)
 {
-	struct amdgpu_vram_mgr *mgr = to_vram_mgr(man);
+	struct amdgpu_vram_mgr_resource *mgr = to_amdgpu_vram_mgr_resource(res);
+	struct drm_buddy_block *block;
+
+	/* Check each drm buddy block individually */
+	list_for_each_entry(block, &mgr->blocks, link) {
+		unsigned long fpfn =
+			amdgpu_vram_mgr_block_start(block) >> PAGE_SHIFT;
+		unsigned long lpfn = fpfn +
+			(amdgpu_vram_mgr_block_size(block) >> PAGE_SHIFT);
+
+		if (fpfn < place->fpfn ||
+		    (place->lpfn && lpfn > place->lpfn))
+			return false;
+	}
 
-	return atomic64_read(&mgr->vis_usage);
+	return true;
 }
 
 /**
@@ -669,19 +877,30 @@ static void amdgpu_vram_mgr_debug(struct ttm_resource_manager *man,
 				  struct drm_printer *printer)
 {
 	struct amdgpu_vram_mgr *mgr = to_vram_mgr(man);
+	struct drm_buddy *mm = &mgr->mm;
+	struct amdgpu_vram_reservation *rsv;
+
+	drm_printf(printer, "  vis usage:%llu\n",
+		   amdgpu_vram_mgr_vis_usage(mgr));
 
-	spin_lock(&mgr->lock);
-	drm_mm_print(&mgr->mm, printer);
-	spin_unlock(&mgr->lock);
+	mutex_lock(&mgr->lock);
+	drm_printf(printer, "default_page_size: %lluKiB\n",
+		   mgr->default_page_size >> 10);
 
-	drm_printf(printer, "man size:%llu pages, ram usage:%lluMB, vis usage:%lluMB\n",
-		   man->size, amdgpu_vram_mgr_usage(man) >> 20,
-		   amdgpu_vram_mgr_vis_usage(man) >> 20);
+	drm_buddy_print(mm, printer);
+
+	drm_printf(printer, "reserved:\n");
+	list_for_each_entry(rsv, &mgr->reserved_pages, blocks)
+		drm_printf(printer, "%#018llx-%#018llx: %llu\n",
+			rsv->start, rsv->start + rsv->size, rsv->size);
+	mutex_unlock(&mgr->lock);
 }
 
 static const struct ttm_resource_manager_func amdgpu_vram_mgr_func = {
 	.alloc	= amdgpu_vram_mgr_new,
 	.free	= amdgpu_vram_mgr_del,
+	.intersects = amdgpu_vram_mgr_intersects,
+	.compatible = amdgpu_vram_mgr_compatible,
 	.debug	= amdgpu_vram_mgr_debug
 };
 
@@ -696,15 +915,24 @@ int amdgpu_vram_mgr_init(struct amdgpu_device *adev)
 {
 	struct amdgpu_vram_mgr *mgr = &adev->mman.vram_mgr;
 	struct ttm_resource_manager *man = &mgr->manager;
+	int err;
 
-	ttm_resource_manager_init(man, adev->gmc.real_vram_size >> PAGE_SHIFT);
+	man->cg = drmm_cgroup_register_region(adev_to_drm(adev), "vram", adev->gmc.real_vram_size);
+	if (IS_ERR(man->cg))
+		return PTR_ERR(man->cg);
+	ttm_resource_manager_init(man, &adev->mman.bdev,
+				  adev->gmc.real_vram_size);
 
-	man->func = &amdgpu_vram_mgr_func;
-
-	drm_mm_init(&mgr->mm, 0, man->size);
-	spin_lock_init(&mgr->lock);
+	mutex_init(&mgr->lock);
 	INIT_LIST_HEAD(&mgr->reservations_pending);
 	INIT_LIST_HEAD(&mgr->reserved_pages);
+	INIT_LIST_HEAD(&mgr->allocated_vres_list);
+	mgr->default_page_size = PAGE_SIZE;
+
+	man->func = &amdgpu_vram_mgr_func;
+	err = drm_buddy_init(&mgr->mm, man->size, PAGE_SIZE);
+	if (err)
+		return err;
 
 	ttm_set_driver_manager(&adev->mman.bdev, TTM_PL_VRAM, &mgr->manager);
 	ttm_resource_manager_set_used(man, true);
@@ -732,16 +960,17 @@ void amdgpu_vram_mgr_fini(struct amdgpu_device *adev)
 	if (ret)
 		return;
 
-	spin_lock(&mgr->lock);
-	list_for_each_entry_safe(rsv, temp, &mgr->reservations_pending, node)
+	mutex_lock(&mgr->lock);
+	list_for_each_entry_safe(rsv, temp, &mgr->reservations_pending, blocks)
 		kfree(rsv);
 
-	list_for_each_entry_safe(rsv, temp, &mgr->reserved_pages, node) {
-		drm_mm_remove_node(&rsv->mm_node);
+	list_for_each_entry_safe(rsv, temp, &mgr->reserved_pages, blocks) {
+		drm_buddy_free_list(&mgr->mm, &rsv->allocated, 0);
 		kfree(rsv);
 	}
-	drm_mm_takedown(&mgr->mm);
-	spin_unlock(&mgr->lock);
+	if (!adev->gmc.is_app_apu)
+		drm_buddy_fini(&mgr->mm);
+	mutex_unlock(&mgr->lock);
 
 	ttm_resource_manager_cleanup(man);
 	ttm_set_driver_manager(&adev->mman.bdev, TTM_PL_VRAM, NULL);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.h
new file mode 100644
index 000000000000..5f5fd9a911c2
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.h
@@ -0,0 +1,92 @@
+/* SPDX-License-Identifier: MIT
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __AMDGPU_VRAM_MGR_H__
+#define __AMDGPU_VRAM_MGR_H__
+
+#include <drm/drm_buddy.h>
+
+struct amdgpu_vram_mgr {
+	struct ttm_resource_manager manager;
+	struct drm_buddy mm;
+	/* protects access to buffer objects */
+	struct mutex lock;
+	struct list_head reservations_pending;
+	struct list_head reserved_pages;
+	atomic64_t vis_usage;
+	u64 default_page_size;
+	struct list_head allocated_vres_list;
+};
+
+struct amdgpu_vres_task {
+	pid_t pid;
+	char  comm[TASK_COMM_LEN];
+};
+
+struct amdgpu_vram_block_info {
+	u64 start;
+	u64 size;
+	struct amdgpu_vres_task task;
+};
+
+struct amdgpu_vram_mgr_resource {
+	struct ttm_resource base;
+	struct list_head blocks;
+	unsigned long flags;
+	struct list_head vres_node;
+	struct amdgpu_vres_task task;
+};
+
+static inline u64 amdgpu_vram_mgr_block_start(struct drm_buddy_block *block)
+{
+	return drm_buddy_block_offset(block);
+}
+
+static inline u64 amdgpu_vram_mgr_block_size(struct drm_buddy_block *block)
+{
+	return (u64)PAGE_SIZE << drm_buddy_block_order(block);
+}
+
+static inline bool amdgpu_vram_mgr_is_cleared(struct drm_buddy_block *block)
+{
+	return drm_buddy_block_is_clear(block);
+}
+
+static inline struct amdgpu_vram_mgr_resource *
+to_amdgpu_vram_mgr_resource(struct ttm_resource *res)
+{
+	return container_of(res, struct amdgpu_vram_mgr_resource, base);
+}
+
+static inline void amdgpu_vram_mgr_set_cleared(struct ttm_resource *res)
+{
+	struct amdgpu_vram_mgr_resource *ares = to_amdgpu_vram_mgr_resource(res);
+
+	WARN_ON(ares->flags & DRM_BUDDY_CLEARED);
+	ares->flags |= DRM_BUDDY_CLEARED;
+}
+
+int amdgpu_vram_mgr_query_address_block_info(struct amdgpu_vram_mgr *mgr,
+		uint64_t address, struct amdgpu_vram_block_info *info);
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_xcp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_xcp.c
new file mode 100644
index 000000000000..1083db8cea2e
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_xcp.c
@@ -0,0 +1,1107 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include "amdgpu.h"
+#include "amdgpu_xcp.h"
+#include "amdgpu_drv.h"
+
+#include <drm/drm_drv.h>
+#include "../amdxcp/amdgpu_xcp_drv.h"
+
+static void amdgpu_xcp_sysfs_entries_init(struct amdgpu_xcp_mgr *xcp_mgr);
+static void amdgpu_xcp_sysfs_entries_update(struct amdgpu_xcp_mgr *xcp_mgr);
+
+static int __amdgpu_xcp_run(struct amdgpu_xcp_mgr *xcp_mgr,
+			    struct amdgpu_xcp_ip *xcp_ip, int xcp_state)
+{
+	int (*run_func)(void *handle, uint32_t inst_mask);
+	int ret = 0;
+
+	if (!xcp_ip || !xcp_ip->valid || !xcp_ip->ip_funcs)
+		return 0;
+
+	run_func = NULL;
+
+	switch (xcp_state) {
+	case AMDGPU_XCP_PREPARE_SUSPEND:
+		run_func = xcp_ip->ip_funcs->prepare_suspend;
+		break;
+	case AMDGPU_XCP_SUSPEND:
+		run_func = xcp_ip->ip_funcs->suspend;
+		break;
+	case AMDGPU_XCP_PREPARE_RESUME:
+		run_func = xcp_ip->ip_funcs->prepare_resume;
+		break;
+	case AMDGPU_XCP_RESUME:
+		run_func = xcp_ip->ip_funcs->resume;
+		break;
+	}
+
+	if (run_func)
+		ret = run_func(xcp_mgr->adev, xcp_ip->inst_mask);
+
+	return ret;
+}
+
+static int amdgpu_xcp_run_transition(struct amdgpu_xcp_mgr *xcp_mgr, int xcp_id,
+				     int state)
+{
+	struct amdgpu_xcp_ip *xcp_ip;
+	struct amdgpu_xcp *xcp;
+	int i, ret;
+
+	if (xcp_id >= MAX_XCP || !xcp_mgr->xcp[xcp_id].valid)
+		return -EINVAL;
+
+	xcp = &xcp_mgr->xcp[xcp_id];
+	for (i = 0; i < AMDGPU_XCP_MAX_BLOCKS; ++i) {
+		xcp_ip = &xcp->ip[i];
+		ret = __amdgpu_xcp_run(xcp_mgr, xcp_ip, state);
+		if (ret)
+			break;
+	}
+
+	return ret;
+}
+
+int amdgpu_xcp_prepare_suspend(struct amdgpu_xcp_mgr *xcp_mgr, int xcp_id)
+{
+	return amdgpu_xcp_run_transition(xcp_mgr, xcp_id,
+					 AMDGPU_XCP_PREPARE_SUSPEND);
+}
+
+int amdgpu_xcp_suspend(struct amdgpu_xcp_mgr *xcp_mgr, int xcp_id)
+{
+	return amdgpu_xcp_run_transition(xcp_mgr, xcp_id, AMDGPU_XCP_SUSPEND);
+}
+
+int amdgpu_xcp_prepare_resume(struct amdgpu_xcp_mgr *xcp_mgr, int xcp_id)
+{
+	return amdgpu_xcp_run_transition(xcp_mgr, xcp_id,
+					 AMDGPU_XCP_PREPARE_RESUME);
+}
+
+int amdgpu_xcp_resume(struct amdgpu_xcp_mgr *xcp_mgr, int xcp_id)
+{
+	return amdgpu_xcp_run_transition(xcp_mgr, xcp_id, AMDGPU_XCP_RESUME);
+}
+
+static void __amdgpu_xcp_add_block(struct amdgpu_xcp_mgr *xcp_mgr, int xcp_id,
+				   struct amdgpu_xcp_ip *ip)
+{
+	struct amdgpu_xcp *xcp;
+
+	if (!ip)
+		return;
+
+	xcp = &xcp_mgr->xcp[xcp_id];
+	xcp->ip[ip->ip_id] = *ip;
+	xcp->ip[ip->ip_id].valid = true;
+
+	xcp->valid = true;
+}
+
+static void __amdgpu_xcp_set_unique_id(struct amdgpu_xcp_mgr *xcp_mgr,
+				       int xcp_id)
+{
+	struct amdgpu_xcp *xcp = &xcp_mgr->xcp[xcp_id];
+	struct amdgpu_device *adev = xcp_mgr->adev;
+	uint32_t inst_mask;
+	uint64_t uid;
+	int i;
+
+	if (!amdgpu_xcp_get_inst_details(xcp, AMDGPU_XCP_GFX, &inst_mask) &&
+	    inst_mask) {
+		i = GET_INST(GC, (ffs(inst_mask) - 1));
+		uid = amdgpu_device_get_uid(xcp_mgr->adev->uid_info,
+					    AMDGPU_UID_TYPE_XCD, i);
+		if (uid)
+			xcp->unique_id = uid;
+	}
+}
+
+int amdgpu_xcp_init(struct amdgpu_xcp_mgr *xcp_mgr, int num_xcps, int mode)
+{
+	struct amdgpu_device *adev = xcp_mgr->adev;
+	struct amdgpu_xcp_ip ip;
+	uint8_t mem_id;
+	int i, j, ret;
+
+	if (!num_xcps || num_xcps > MAX_XCP)
+		return -EINVAL;
+
+	xcp_mgr->mode = mode;
+
+	for (i = 0; i < MAX_XCP; ++i)
+		xcp_mgr->xcp[i].valid = false;
+
+	/* This is needed for figuring out memory id of xcp */
+	xcp_mgr->num_xcp_per_mem_partition = num_xcps / xcp_mgr->adev->gmc.num_mem_partitions;
+
+	for (i = 0; i < num_xcps; ++i) {
+		for (j = AMDGPU_XCP_GFXHUB; j < AMDGPU_XCP_MAX_BLOCKS; ++j) {
+			ret = xcp_mgr->funcs->get_ip_details(xcp_mgr, i, j,
+							     &ip);
+			if (ret)
+				continue;
+
+			__amdgpu_xcp_add_block(xcp_mgr, i, &ip);
+		}
+
+		xcp_mgr->xcp[i].id = i;
+
+		if (xcp_mgr->funcs->get_xcp_mem_id) {
+			ret = xcp_mgr->funcs->get_xcp_mem_id(
+				xcp_mgr, &xcp_mgr->xcp[i], &mem_id);
+			if (ret)
+				continue;
+			else
+				xcp_mgr->xcp[i].mem_id = mem_id;
+		}
+		__amdgpu_xcp_set_unique_id(xcp_mgr, i);
+	}
+
+	xcp_mgr->num_xcps = num_xcps;
+	amdgpu_xcp_update_partition_sched_list(adev);
+
+	return 0;
+}
+
+static int __amdgpu_xcp_switch_partition_mode(struct amdgpu_xcp_mgr *xcp_mgr,
+					      int mode)
+{
+	int ret, curr_mode, num_xcps = 0;
+
+	if (!xcp_mgr->funcs || !xcp_mgr->funcs->switch_partition_mode)
+		return 0;
+
+	mutex_lock(&xcp_mgr->xcp_lock);
+
+	curr_mode = xcp_mgr->mode;
+	/* State set to transient mode */
+	xcp_mgr->mode = AMDGPU_XCP_MODE_TRANS;
+
+	ret = xcp_mgr->funcs->switch_partition_mode(xcp_mgr, mode, &num_xcps);
+
+	if (ret) {
+		/* Failed, get whatever mode it's at now */
+		if (xcp_mgr->funcs->query_partition_mode)
+			xcp_mgr->mode = amdgpu_xcp_query_partition_mode(
+				xcp_mgr, AMDGPU_XCP_FL_LOCKED);
+		else
+			xcp_mgr->mode = curr_mode;
+
+		goto out;
+	}
+	amdgpu_xcp_sysfs_entries_update(xcp_mgr);
+out:
+	mutex_unlock(&xcp_mgr->xcp_lock);
+
+	return ret;
+}
+
+int amdgpu_xcp_switch_partition_mode(struct amdgpu_xcp_mgr *xcp_mgr, int mode)
+{
+	if (!xcp_mgr || mode == AMDGPU_XCP_MODE_NONE)
+		return -EINVAL;
+
+	if (xcp_mgr->mode == mode)
+		return 0;
+
+	return __amdgpu_xcp_switch_partition_mode(xcp_mgr, mode);
+}
+
+int amdgpu_xcp_restore_partition_mode(struct amdgpu_xcp_mgr *xcp_mgr)
+{
+	if (!xcp_mgr || xcp_mgr->mode == AMDGPU_XCP_MODE_NONE)
+		return 0;
+
+	return __amdgpu_xcp_switch_partition_mode(xcp_mgr, xcp_mgr->mode);
+}
+
+static bool __amdgpu_xcp_is_cached_mode_valid(struct amdgpu_xcp_mgr *xcp_mgr)
+{
+	if (!xcp_mgr->funcs || !xcp_mgr->funcs->query_partition_mode)
+		return true;
+
+	if (!amdgpu_sriov_vf(xcp_mgr->adev) &&
+	    xcp_mgr->mode == AMDGPU_XCP_MODE_NONE)
+		return true;
+
+	if (xcp_mgr->mode != AMDGPU_XCP_MODE_NONE &&
+	    xcp_mgr->mode != AMDGPU_XCP_MODE_TRANS)
+		return true;
+
+	return false;
+}
+
+int amdgpu_xcp_query_partition_mode(struct amdgpu_xcp_mgr *xcp_mgr, u32 flags)
+{
+	int mode;
+
+	if (__amdgpu_xcp_is_cached_mode_valid(xcp_mgr))
+		return xcp_mgr->mode;
+
+	if (!(flags & AMDGPU_XCP_FL_LOCKED))
+		mutex_lock(&xcp_mgr->xcp_lock);
+	mode = xcp_mgr->funcs->query_partition_mode(xcp_mgr);
+
+	/* First time query for VF, set the mode here */
+	if (amdgpu_sriov_vf(xcp_mgr->adev) &&
+	    xcp_mgr->mode == AMDGPU_XCP_MODE_NONE)
+		xcp_mgr->mode = mode;
+
+	if (xcp_mgr->mode != AMDGPU_XCP_MODE_TRANS && mode != xcp_mgr->mode)
+		dev_WARN(
+			xcp_mgr->adev->dev,
+			"Cached partition mode %d not matching with device mode %d",
+			xcp_mgr->mode, mode);
+
+	if (!(flags & AMDGPU_XCP_FL_LOCKED))
+		mutex_unlock(&xcp_mgr->xcp_lock);
+
+	return mode;
+}
+
+static int amdgpu_xcp_dev_alloc(struct amdgpu_device *adev)
+{
+	struct drm_device *p_ddev;
+	struct drm_device *ddev;
+	int i, ret;
+
+	ddev = adev_to_drm(adev);
+
+	/* xcp #0 shares drm device setting with adev */
+	adev->xcp_mgr->xcp->ddev = ddev;
+
+	for (i = 1; i < MAX_XCP; i++) {
+		ret = amdgpu_xcp_drm_dev_alloc(&p_ddev);
+		if (ret == -ENOSPC) {
+			dev_warn(adev->dev,
+			"Skip xcp node #%d when out of drm node resource.", i);
+			ret = 0;
+			goto out;
+		} else if (ret) {
+			goto out;
+		}
+
+		/* Redirect all IOCTLs to the primary device */
+		adev->xcp_mgr->xcp[i].rdev = p_ddev->render->dev;
+		adev->xcp_mgr->xcp[i].pdev = p_ddev->primary->dev;
+		adev->xcp_mgr->xcp[i].driver = (struct drm_driver *)p_ddev->driver;
+		adev->xcp_mgr->xcp[i].vma_offset_manager = p_ddev->vma_offset_manager;
+		p_ddev->render->dev = ddev;
+		p_ddev->primary->dev = ddev;
+		p_ddev->vma_offset_manager = ddev->vma_offset_manager;
+		p_ddev->driver = &amdgpu_partition_driver;
+		adev->xcp_mgr->xcp[i].ddev = p_ddev;
+
+		dev_set_drvdata(p_ddev->dev, &adev->xcp_mgr->xcp[i]);
+	}
+	ret = 0;
+out:
+	amdgpu_xcp_sysfs_entries_init(adev->xcp_mgr);
+
+	return ret;
+}
+
+int amdgpu_xcp_mgr_init(struct amdgpu_device *adev, int init_mode,
+			int init_num_xcps,
+			struct amdgpu_xcp_mgr_funcs *xcp_funcs)
+{
+	struct amdgpu_xcp_mgr *xcp_mgr;
+	int i;
+
+	if (!xcp_funcs || !xcp_funcs->get_ip_details)
+		return -EINVAL;
+
+	xcp_mgr = kzalloc(sizeof(*xcp_mgr), GFP_KERNEL);
+
+	if (!xcp_mgr)
+		return -ENOMEM;
+
+	xcp_mgr->adev = adev;
+	xcp_mgr->funcs = xcp_funcs;
+	xcp_mgr->mode = init_mode;
+	mutex_init(&xcp_mgr->xcp_lock);
+
+	if (init_mode != AMDGPU_XCP_MODE_NONE)
+		amdgpu_xcp_init(xcp_mgr, init_num_xcps, init_mode);
+
+	adev->xcp_mgr = xcp_mgr;
+	for (i = 0; i < MAX_XCP; ++i)
+		xcp_mgr->xcp[i].xcp_mgr = xcp_mgr;
+
+	return amdgpu_xcp_dev_alloc(adev);
+}
+
+int amdgpu_xcp_get_partition(struct amdgpu_xcp_mgr *xcp_mgr,
+			     enum AMDGPU_XCP_IP_BLOCK ip, int instance)
+{
+	struct amdgpu_xcp *xcp;
+	int i, id_mask = 0;
+
+	if (ip >= AMDGPU_XCP_MAX_BLOCKS)
+		return -EINVAL;
+
+	for (i = 0; i < xcp_mgr->num_xcps; ++i) {
+		xcp = &xcp_mgr->xcp[i];
+		if ((xcp->valid) && (xcp->ip[ip].valid) &&
+		    (xcp->ip[ip].inst_mask & BIT(instance)))
+			id_mask |= BIT(i);
+	}
+
+	if (!id_mask)
+		id_mask = -ENXIO;
+
+	return id_mask;
+}
+
+int amdgpu_xcp_get_inst_details(struct amdgpu_xcp *xcp,
+				enum AMDGPU_XCP_IP_BLOCK ip,
+				uint32_t *inst_mask)
+{
+	if (!xcp->valid || !inst_mask || !(xcp->ip[ip].valid))
+		return -EINVAL;
+
+	*inst_mask = xcp->ip[ip].inst_mask;
+
+	return 0;
+}
+
+int amdgpu_xcp_dev_register(struct amdgpu_device *adev,
+			const struct pci_device_id *ent)
+{
+	int i, ret;
+
+	if (!adev->xcp_mgr)
+		return 0;
+
+	for (i = 1; i < MAX_XCP; i++) {
+		if (!adev->xcp_mgr->xcp[i].ddev)
+			break;
+
+		ret = drm_dev_register(adev->xcp_mgr->xcp[i].ddev, ent->driver_data);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+void amdgpu_xcp_dev_unplug(struct amdgpu_device *adev)
+{
+	struct drm_device *p_ddev;
+	int i;
+
+	if (!adev->xcp_mgr)
+		return;
+
+	for (i = 1; i < MAX_XCP; i++) {
+		if (!adev->xcp_mgr->xcp[i].ddev)
+			break;
+
+		p_ddev = adev->xcp_mgr->xcp[i].ddev;
+		drm_dev_unplug(p_ddev);
+		p_ddev->render->dev = adev->xcp_mgr->xcp[i].rdev;
+		p_ddev->primary->dev = adev->xcp_mgr->xcp[i].pdev;
+		p_ddev->driver =  adev->xcp_mgr->xcp[i].driver;
+		p_ddev->vma_offset_manager = adev->xcp_mgr->xcp[i].vma_offset_manager;
+		amdgpu_xcp_drm_dev_free(p_ddev);
+	}
+}
+
+int amdgpu_xcp_open_device(struct amdgpu_device *adev,
+			   struct amdgpu_fpriv *fpriv,
+			   struct drm_file *file_priv)
+{
+	int i;
+
+	if (!adev->xcp_mgr)
+		return 0;
+
+	fpriv->xcp_id = AMDGPU_XCP_NO_PARTITION;
+	for (i = 0; i < MAX_XCP; ++i) {
+		if (!adev->xcp_mgr->xcp[i].ddev)
+			break;
+
+		if (file_priv->minor == adev->xcp_mgr->xcp[i].ddev->render) {
+			if (adev->xcp_mgr->xcp[i].valid == FALSE) {
+				dev_err(adev->dev, "renderD%d partition %d not valid!",
+						file_priv->minor->index, i);
+				return -ENOENT;
+			}
+			dev_dbg(adev->dev, "renderD%d partition %d opened!",
+					file_priv->minor->index, i);
+			fpriv->xcp_id = i;
+			break;
+		}
+	}
+
+	fpriv->vm.mem_id = fpriv->xcp_id == AMDGPU_XCP_NO_PARTITION ? -1 :
+				adev->xcp_mgr->xcp[fpriv->xcp_id].mem_id;
+	return 0;
+}
+
+void amdgpu_xcp_release_sched(struct amdgpu_device *adev,
+				  struct amdgpu_ctx_entity *entity)
+{
+	struct drm_gpu_scheduler *sched;
+	struct amdgpu_ring *ring;
+
+	if (!adev->xcp_mgr)
+		return;
+
+	sched = entity->entity.rq->sched;
+	if (drm_sched_wqueue_ready(sched)) {
+		ring = to_amdgpu_ring(entity->entity.rq->sched);
+		atomic_dec(&adev->xcp_mgr->xcp[ring->xcp_id].ref_cnt);
+	}
+}
+
+int amdgpu_xcp_select_scheds(struct amdgpu_device *adev,
+			     u32 hw_ip, u32 hw_prio,
+			     struct amdgpu_fpriv *fpriv,
+			     unsigned int *num_scheds,
+			     struct drm_gpu_scheduler ***scheds)
+{
+	u32 sel_xcp_id;
+	int i;
+	struct amdgpu_xcp_mgr *xcp_mgr = adev->xcp_mgr;
+
+	if (fpriv->xcp_id == AMDGPU_XCP_NO_PARTITION) {
+		u32 least_ref_cnt = ~0;
+
+		fpriv->xcp_id = 0;
+		for (i = 0; i < xcp_mgr->num_xcps; i++) {
+			u32 total_ref_cnt;
+
+			total_ref_cnt = atomic_read(&xcp_mgr->xcp[i].ref_cnt);
+			if (total_ref_cnt < least_ref_cnt) {
+				fpriv->xcp_id = i;
+				least_ref_cnt = total_ref_cnt;
+			}
+		}
+	}
+	sel_xcp_id = fpriv->xcp_id;
+
+	if (xcp_mgr->xcp[sel_xcp_id].gpu_sched[hw_ip][hw_prio].num_scheds) {
+		*num_scheds =
+			xcp_mgr->xcp[fpriv->xcp_id].gpu_sched[hw_ip][hw_prio].num_scheds;
+		*scheds =
+			xcp_mgr->xcp[fpriv->xcp_id].gpu_sched[hw_ip][hw_prio].sched;
+		atomic_inc(&adev->xcp_mgr->xcp[sel_xcp_id].ref_cnt);
+		dev_dbg(adev->dev, "Selected partition #%d", sel_xcp_id);
+	} else {
+		dev_err(adev->dev, "Failed to schedule partition #%d.", sel_xcp_id);
+		return -ENOENT;
+	}
+
+	return 0;
+}
+
+static void amdgpu_set_xcp_id(struct amdgpu_device *adev,
+			      uint32_t inst_idx,
+			      struct amdgpu_ring *ring)
+{
+	int xcp_id;
+	enum AMDGPU_XCP_IP_BLOCK ip_blk;
+	uint32_t inst_mask;
+
+	ring->xcp_id = AMDGPU_XCP_NO_PARTITION;
+	if (ring->funcs->type == AMDGPU_RING_TYPE_COMPUTE)
+		adev->gfx.enforce_isolation[0].xcp_id = ring->xcp_id;
+	if ((adev->xcp_mgr->mode == AMDGPU_XCP_MODE_NONE) ||
+	    (ring->funcs->type == AMDGPU_RING_TYPE_CPER))
+		return;
+
+	inst_mask = 1 << inst_idx;
+
+	switch (ring->funcs->type) {
+	case AMDGPU_HW_IP_GFX:
+	case AMDGPU_RING_TYPE_COMPUTE:
+	case AMDGPU_RING_TYPE_KIQ:
+		ip_blk = AMDGPU_XCP_GFX;
+		break;
+	case AMDGPU_RING_TYPE_SDMA:
+		ip_blk = AMDGPU_XCP_SDMA;
+		break;
+	case AMDGPU_RING_TYPE_VCN_ENC:
+	case AMDGPU_RING_TYPE_VCN_JPEG:
+		ip_blk = AMDGPU_XCP_VCN;
+		break;
+	default:
+		dev_err(adev->dev, "Not support ring type %d!", ring->funcs->type);
+		return;
+	}
+
+	for (xcp_id = 0; xcp_id < adev->xcp_mgr->num_xcps; xcp_id++) {
+		if (adev->xcp_mgr->xcp[xcp_id].ip[ip_blk].inst_mask & inst_mask) {
+			ring->xcp_id = xcp_id;
+			dev_dbg(adev->dev, "ring:%s xcp_id :%u", ring->name,
+				ring->xcp_id);
+			if (ring->funcs->type == AMDGPU_RING_TYPE_COMPUTE)
+				adev->gfx.enforce_isolation[xcp_id].xcp_id = xcp_id;
+			break;
+		}
+	}
+}
+
+static void amdgpu_xcp_gpu_sched_update(struct amdgpu_device *adev,
+					struct amdgpu_ring *ring,
+					unsigned int sel_xcp_id)
+{
+	unsigned int *num_gpu_sched;
+
+	num_gpu_sched = &adev->xcp_mgr->xcp[sel_xcp_id]
+			.gpu_sched[ring->funcs->type][ring->hw_prio].num_scheds;
+	adev->xcp_mgr->xcp[sel_xcp_id].gpu_sched[ring->funcs->type][ring->hw_prio]
+			.sched[(*num_gpu_sched)++] = &ring->sched;
+	dev_dbg(adev->dev, "%s :[%d] gpu_sched[%d][%d] = %d",
+		ring->name, sel_xcp_id, ring->funcs->type,
+		ring->hw_prio, *num_gpu_sched);
+}
+
+static int amdgpu_xcp_sched_list_update(struct amdgpu_device *adev)
+{
+	struct amdgpu_ring *ring;
+	int i;
+
+	for (i = 0; i < MAX_XCP; i++) {
+		atomic_set(&adev->xcp_mgr->xcp[i].ref_cnt, 0);
+		memset(adev->xcp_mgr->xcp[i].gpu_sched, 0, sizeof(adev->xcp_mgr->xcp->gpu_sched));
+	}
+
+	if (adev->xcp_mgr->mode == AMDGPU_XCP_MODE_NONE)
+		return 0;
+
+	for (i = 0; i < AMDGPU_MAX_RINGS; i++) {
+		ring = adev->rings[i];
+		if (!ring || !ring->sched.ready || ring->no_scheduler)
+			continue;
+
+		amdgpu_xcp_gpu_sched_update(adev, ring, ring->xcp_id);
+
+		/* VCN may be shared by two partitions under CPX MODE in certain
+		 * configs.
+		 */
+		if ((ring->funcs->type == AMDGPU_RING_TYPE_VCN_ENC ||
+		     ring->funcs->type == AMDGPU_RING_TYPE_VCN_JPEG) &&
+		    (adev->xcp_mgr->num_xcps > adev->vcn.num_vcn_inst))
+			amdgpu_xcp_gpu_sched_update(adev, ring, ring->xcp_id + 1);
+	}
+
+	return 0;
+}
+
+int amdgpu_xcp_update_partition_sched_list(struct amdgpu_device *adev)
+{
+	int i;
+
+	for (i = 0; i < adev->num_rings; i++) {
+		struct amdgpu_ring *ring = adev->rings[i];
+
+		if (ring->funcs->type == AMDGPU_RING_TYPE_COMPUTE ||
+			ring->funcs->type == AMDGPU_RING_TYPE_KIQ)
+			amdgpu_set_xcp_id(adev, ring->xcc_id, ring);
+		else
+			amdgpu_set_xcp_id(adev, ring->me, ring);
+	}
+
+	return amdgpu_xcp_sched_list_update(adev);
+}
+
+void amdgpu_xcp_update_supported_modes(struct amdgpu_xcp_mgr *xcp_mgr)
+{
+	struct amdgpu_device *adev = xcp_mgr->adev;
+
+	xcp_mgr->supp_xcp_modes = 0;
+
+	switch (NUM_XCC(adev->gfx.xcc_mask)) {
+	case 8:
+		xcp_mgr->supp_xcp_modes = BIT(AMDGPU_SPX_PARTITION_MODE) |
+					  BIT(AMDGPU_DPX_PARTITION_MODE) |
+					  BIT(AMDGPU_QPX_PARTITION_MODE) |
+					  BIT(AMDGPU_CPX_PARTITION_MODE);
+		break;
+	case 6:
+		xcp_mgr->supp_xcp_modes = BIT(AMDGPU_SPX_PARTITION_MODE) |
+					  BIT(AMDGPU_TPX_PARTITION_MODE) |
+					  BIT(AMDGPU_CPX_PARTITION_MODE);
+		break;
+	case 4:
+		xcp_mgr->supp_xcp_modes = BIT(AMDGPU_SPX_PARTITION_MODE) |
+					  BIT(AMDGPU_DPX_PARTITION_MODE) |
+					  BIT(AMDGPU_CPX_PARTITION_MODE);
+		break;
+	case 2:
+		xcp_mgr->supp_xcp_modes = BIT(AMDGPU_SPX_PARTITION_MODE) |
+					  BIT(AMDGPU_CPX_PARTITION_MODE);
+		break;
+	case 1:
+		xcp_mgr->supp_xcp_modes = BIT(AMDGPU_SPX_PARTITION_MODE) |
+					  BIT(AMDGPU_CPX_PARTITION_MODE);
+		break;
+
+	default:
+		break;
+	}
+}
+
+int amdgpu_xcp_pre_partition_switch(struct amdgpu_xcp_mgr *xcp_mgr, u32 flags)
+{
+	/* TODO:
+	 * Stop user queues and threads, and make sure GPU is empty of work.
+	 */
+
+	if (flags & AMDGPU_XCP_OPS_KFD)
+		amdgpu_amdkfd_device_fini_sw(xcp_mgr->adev);
+
+	return 0;
+}
+
+int amdgpu_xcp_post_partition_switch(struct amdgpu_xcp_mgr *xcp_mgr, u32 flags)
+{
+	int ret = 0;
+
+	if (flags & AMDGPU_XCP_OPS_KFD) {
+		amdgpu_amdkfd_device_probe(xcp_mgr->adev);
+		amdgpu_amdkfd_device_init(xcp_mgr->adev);
+		/* If KFD init failed, return failure */
+		if (!xcp_mgr->adev->kfd.init_complete)
+			ret = -EIO;
+	}
+
+	return ret;
+}
+
+/*====================== xcp sysfs - configuration ======================*/
+#define XCP_CFG_SYSFS_RES_ATTR_SHOW(_name)                         \
+	static ssize_t amdgpu_xcp_res_sysfs_##_name##_show(        \
+		struct amdgpu_xcp_res_details *xcp_res, char *buf) \
+	{                                                          \
+		return sysfs_emit(buf, "%d\n", xcp_res->_name);    \
+	}
+
+struct amdgpu_xcp_res_sysfs_attribute {
+	struct attribute attr;
+	ssize_t (*show)(struct amdgpu_xcp_res_details *xcp_res, char *buf);
+};
+
+#define XCP_CFG_SYSFS_RES_ATTR(_name)                                        \
+	struct amdgpu_xcp_res_sysfs_attribute xcp_res_sysfs_attr_##_name = { \
+		.attr = { .name = __stringify(_name), .mode = 0400 },        \
+		.show = amdgpu_xcp_res_sysfs_##_name##_show,                 \
+	}
+
+XCP_CFG_SYSFS_RES_ATTR_SHOW(num_inst)
+XCP_CFG_SYSFS_RES_ATTR(num_inst);
+XCP_CFG_SYSFS_RES_ATTR_SHOW(num_shared)
+XCP_CFG_SYSFS_RES_ATTR(num_shared);
+
+#define XCP_CFG_SYSFS_RES_ATTR_PTR(_name) xcp_res_sysfs_attr_##_name.attr
+
+static struct attribute *xcp_cfg_res_sysfs_attrs[] = {
+	&XCP_CFG_SYSFS_RES_ATTR_PTR(num_inst),
+	&XCP_CFG_SYSFS_RES_ATTR_PTR(num_shared), NULL
+};
+
+static const char *xcp_desc[] = {
+	[AMDGPU_SPX_PARTITION_MODE] = "SPX",
+	[AMDGPU_DPX_PARTITION_MODE] = "DPX",
+	[AMDGPU_TPX_PARTITION_MODE] = "TPX",
+	[AMDGPU_QPX_PARTITION_MODE] = "QPX",
+	[AMDGPU_CPX_PARTITION_MODE] = "CPX",
+};
+
+static const char *nps_desc[] = {
+	[UNKNOWN_MEMORY_PARTITION_MODE] = "UNKNOWN",
+	[AMDGPU_NPS1_PARTITION_MODE] = "NPS1",
+	[AMDGPU_NPS2_PARTITION_MODE] = "NPS2",
+	[AMDGPU_NPS3_PARTITION_MODE] = "NPS3",
+	[AMDGPU_NPS4_PARTITION_MODE] = "NPS4",
+	[AMDGPU_NPS6_PARTITION_MODE] = "NPS6",
+	[AMDGPU_NPS8_PARTITION_MODE] = "NPS8",
+};
+
+ATTRIBUTE_GROUPS(xcp_cfg_res_sysfs);
+
+#define to_xcp_attr(x) \
+	container_of(x, struct amdgpu_xcp_res_sysfs_attribute, attr)
+#define to_xcp_res(x) container_of(x, struct amdgpu_xcp_res_details, kobj)
+
+static ssize_t xcp_cfg_res_sysfs_attr_show(struct kobject *kobj,
+					   struct attribute *attr, char *buf)
+{
+	struct amdgpu_xcp_res_sysfs_attribute *attribute;
+	struct amdgpu_xcp_res_details *xcp_res;
+
+	attribute = to_xcp_attr(attr);
+	xcp_res = to_xcp_res(kobj);
+
+	if (!attribute->show)
+		return -EIO;
+
+	return attribute->show(xcp_res, buf);
+}
+
+static const struct sysfs_ops xcp_cfg_res_sysfs_ops = {
+	.show = xcp_cfg_res_sysfs_attr_show,
+};
+
+static const struct kobj_type xcp_cfg_res_sysfs_ktype = {
+	.sysfs_ops = &xcp_cfg_res_sysfs_ops,
+	.default_groups = xcp_cfg_res_sysfs_groups,
+};
+
+const char *xcp_res_names[] = {
+	[AMDGPU_XCP_RES_XCC] = "xcc",
+	[AMDGPU_XCP_RES_DMA] = "dma",
+	[AMDGPU_XCP_RES_DEC] = "dec",
+	[AMDGPU_XCP_RES_JPEG] = "jpeg",
+};
+
+static int amdgpu_xcp_get_res_info(struct amdgpu_xcp_mgr *xcp_mgr,
+				   int mode,
+				   struct amdgpu_xcp_cfg *xcp_cfg)
+{
+	if (xcp_mgr->funcs && xcp_mgr->funcs->get_xcp_res_info)
+		return xcp_mgr->funcs->get_xcp_res_info(xcp_mgr, mode, xcp_cfg);
+
+	return -EOPNOTSUPP;
+}
+
+#define to_xcp_cfg(x) container_of(x, struct amdgpu_xcp_cfg, kobj)
+static ssize_t supported_xcp_configs_show(struct kobject *kobj,
+					  struct kobj_attribute *attr, char *buf)
+{
+	struct amdgpu_xcp_cfg *xcp_cfg = to_xcp_cfg(kobj);
+	struct amdgpu_xcp_mgr *xcp_mgr = xcp_cfg->xcp_mgr;
+	int size = 0, mode;
+	char *sep = "";
+
+	if (!xcp_mgr || !xcp_mgr->supp_xcp_modes)
+		return sysfs_emit(buf, "Not supported\n");
+
+	for_each_inst(mode, xcp_mgr->supp_xcp_modes) {
+		size += sysfs_emit_at(buf, size, "%s%s", sep, xcp_desc[mode]);
+		sep = ", ";
+	}
+
+	size += sysfs_emit_at(buf, size, "\n");
+
+	return size;
+}
+
+static ssize_t supported_nps_configs_show(struct kobject *kobj,
+					  struct kobj_attribute *attr, char *buf)
+{
+	struct amdgpu_xcp_cfg *xcp_cfg = to_xcp_cfg(kobj);
+	int size = 0, mode;
+	char *sep = "";
+
+	if (!xcp_cfg || !xcp_cfg->compatible_nps_modes)
+		return sysfs_emit(buf, "Not supported\n");
+
+	for_each_inst(mode, xcp_cfg->compatible_nps_modes) {
+		size += sysfs_emit_at(buf, size, "%s%s", sep, nps_desc[mode]);
+		sep = ", ";
+	}
+
+	size += sysfs_emit_at(buf, size, "\n");
+
+	return size;
+}
+
+static ssize_t xcp_config_show(struct kobject *kobj,
+			       struct kobj_attribute *attr, char *buf)
+{
+	struct amdgpu_xcp_cfg *xcp_cfg = to_xcp_cfg(kobj);
+
+	return sysfs_emit(buf, "%s\n",
+			  amdgpu_gfx_compute_mode_desc(xcp_cfg->mode));
+}
+
+static ssize_t xcp_config_store(struct kobject *kobj,
+				struct kobj_attribute *attr,
+				const char *buf, size_t size)
+{
+	struct amdgpu_xcp_cfg *xcp_cfg = to_xcp_cfg(kobj);
+	int mode, r;
+
+	if (!strncasecmp("SPX", buf, strlen("SPX")))
+		mode = AMDGPU_SPX_PARTITION_MODE;
+	else if (!strncasecmp("DPX", buf, strlen("DPX")))
+		mode = AMDGPU_DPX_PARTITION_MODE;
+	else if (!strncasecmp("TPX", buf, strlen("TPX")))
+		mode = AMDGPU_TPX_PARTITION_MODE;
+	else if (!strncasecmp("QPX", buf, strlen("QPX")))
+		mode = AMDGPU_QPX_PARTITION_MODE;
+	else if (!strncasecmp("CPX", buf, strlen("CPX")))
+		mode = AMDGPU_CPX_PARTITION_MODE;
+	else
+		return -EINVAL;
+
+	r = amdgpu_xcp_get_res_info(xcp_cfg->xcp_mgr, mode, xcp_cfg);
+
+	if (r)
+		return r;
+
+	xcp_cfg->mode = mode;
+	return size;
+}
+
+static struct kobj_attribute xcp_cfg_sysfs_mode =
+	__ATTR_RW_MODE(xcp_config, 0644);
+
+static void xcp_cfg_sysfs_release(struct kobject *kobj)
+{
+	struct amdgpu_xcp_cfg *xcp_cfg = to_xcp_cfg(kobj);
+
+	kfree(xcp_cfg);
+}
+
+static const struct kobj_type xcp_cfg_sysfs_ktype = {
+	.release = xcp_cfg_sysfs_release,
+	.sysfs_ops = &kobj_sysfs_ops,
+};
+
+static struct kobj_attribute supp_part_sysfs_mode =
+	__ATTR_RO(supported_xcp_configs);
+
+static struct kobj_attribute supp_nps_sysfs_mode =
+	__ATTR_RO(supported_nps_configs);
+
+static const struct attribute *xcp_attrs[] = {
+	&supp_part_sysfs_mode.attr,
+	&xcp_cfg_sysfs_mode.attr,
+	NULL,
+};
+
+static void amdgpu_xcp_cfg_sysfs_init(struct amdgpu_device *adev)
+{
+	struct amdgpu_xcp_res_details *xcp_res;
+	struct amdgpu_xcp_cfg *xcp_cfg;
+	int i, r, j, rid, mode;
+
+	if (!adev->xcp_mgr)
+		return;
+
+	xcp_cfg = kzalloc(sizeof(*xcp_cfg), GFP_KERNEL);
+	if (!xcp_cfg)
+		return;
+	xcp_cfg->xcp_mgr = adev->xcp_mgr;
+
+	r = kobject_init_and_add(&xcp_cfg->kobj, &xcp_cfg_sysfs_ktype,
+				 &adev->dev->kobj, "compute_partition_config");
+	if (r)
+		goto err1;
+
+	r = sysfs_create_files(&xcp_cfg->kobj, xcp_attrs);
+	if (r)
+		goto err1;
+
+	if (adev->gmc.supported_nps_modes != 0) {
+		r = sysfs_create_file(&xcp_cfg->kobj, &supp_nps_sysfs_mode.attr);
+		if (r) {
+			sysfs_remove_files(&xcp_cfg->kobj, xcp_attrs);
+			goto err1;
+		}
+	}
+
+	mode = (xcp_cfg->xcp_mgr->mode ==
+		AMDGPU_UNKNOWN_COMPUTE_PARTITION_MODE) ?
+		       AMDGPU_SPX_PARTITION_MODE :
+		       xcp_cfg->xcp_mgr->mode;
+	r = amdgpu_xcp_get_res_info(xcp_cfg->xcp_mgr, mode, xcp_cfg);
+	if (r) {
+		sysfs_remove_file(&xcp_cfg->kobj, &supp_nps_sysfs_mode.attr);
+		sysfs_remove_files(&xcp_cfg->kobj, xcp_attrs);
+		goto err1;
+	}
+
+	xcp_cfg->mode = mode;
+	for (i = 0; i < xcp_cfg->num_res; i++) {
+		xcp_res = &xcp_cfg->xcp_res[i];
+		rid = xcp_res->id;
+		r = kobject_init_and_add(&xcp_res->kobj,
+					 &xcp_cfg_res_sysfs_ktype,
+					 &xcp_cfg->kobj, "%s",
+					 xcp_res_names[rid]);
+		if (r)
+			goto err;
+	}
+
+	adev->xcp_mgr->xcp_cfg = xcp_cfg;
+	return;
+err:
+	for (j = 0; j < i; j++) {
+		xcp_res = &xcp_cfg->xcp_res[i];
+		kobject_put(&xcp_res->kobj);
+	}
+
+	sysfs_remove_file(&xcp_cfg->kobj, &supp_nps_sysfs_mode.attr);
+	sysfs_remove_files(&xcp_cfg->kobj, xcp_attrs);
+err1:
+	kobject_put(&xcp_cfg->kobj);
+}
+
+static void amdgpu_xcp_cfg_sysfs_fini(struct amdgpu_device *adev)
+{
+	struct amdgpu_xcp_res_details *xcp_res;
+	struct amdgpu_xcp_cfg *xcp_cfg;
+	int i;
+
+	if (!adev->xcp_mgr || !adev->xcp_mgr->xcp_cfg)
+		return;
+
+	xcp_cfg = adev->xcp_mgr->xcp_cfg;
+	for (i = 0; i < xcp_cfg->num_res; i++) {
+		xcp_res = &xcp_cfg->xcp_res[i];
+		kobject_put(&xcp_res->kobj);
+	}
+
+	sysfs_remove_file(&xcp_cfg->kobj, &supp_nps_sysfs_mode.attr);
+	sysfs_remove_files(&xcp_cfg->kobj, xcp_attrs);
+	kobject_put(&xcp_cfg->kobj);
+}
+
+/*====================== xcp sysfs - data entries ======================*/
+
+#define to_xcp(x) container_of(x, struct amdgpu_xcp, kobj)
+
+static ssize_t xcp_metrics_show(struct kobject *kobj,
+				struct kobj_attribute *attr, char *buf)
+{
+	struct amdgpu_xcp *xcp = to_xcp(kobj);
+	struct amdgpu_xcp_mgr *xcp_mgr;
+	ssize_t size;
+
+	xcp_mgr = xcp->xcp_mgr;
+	size = amdgpu_dpm_get_xcp_metrics(xcp_mgr->adev, xcp->id, NULL);
+	if (size <= 0)
+		return size;
+
+	if (size > PAGE_SIZE)
+		return -ENOSPC;
+
+	return amdgpu_dpm_get_xcp_metrics(xcp_mgr->adev, xcp->id, buf);
+}
+
+static umode_t amdgpu_xcp_attrs_is_visible(struct kobject *kobj,
+					   struct attribute *attr, int n)
+{
+	struct amdgpu_xcp *xcp = to_xcp(kobj);
+
+	if (!xcp || !xcp->valid)
+		return 0;
+
+	return attr->mode;
+}
+
+static struct kobj_attribute xcp_sysfs_metrics = __ATTR_RO(xcp_metrics);
+
+static struct attribute *amdgpu_xcp_attrs[] = {
+	&xcp_sysfs_metrics.attr,
+	NULL,
+};
+
+static const struct attribute_group amdgpu_xcp_attrs_group = {
+	.attrs = amdgpu_xcp_attrs,
+	.is_visible = amdgpu_xcp_attrs_is_visible
+};
+
+static const struct kobj_type xcp_sysfs_ktype = {
+	.sysfs_ops = &kobj_sysfs_ops,
+};
+
+static void amdgpu_xcp_sysfs_entries_fini(struct amdgpu_xcp_mgr *xcp_mgr, int n)
+{
+	struct amdgpu_xcp *xcp;
+
+	for (n--; n >= 0; n--) {
+		xcp = &xcp_mgr->xcp[n];
+		if (!xcp->ddev || !xcp->valid)
+			continue;
+		sysfs_remove_group(&xcp->kobj, &amdgpu_xcp_attrs_group);
+		kobject_put(&xcp->kobj);
+	}
+}
+
+static void amdgpu_xcp_sysfs_entries_init(struct amdgpu_xcp_mgr *xcp_mgr)
+{
+	struct amdgpu_xcp *xcp;
+	int i, r;
+
+	for (i = 0; i < MAX_XCP; i++) {
+		/* Redirect all IOCTLs to the primary device */
+		xcp = &xcp_mgr->xcp[i];
+		if (!xcp->ddev)
+			break;
+		r = kobject_init_and_add(&xcp->kobj, &xcp_sysfs_ktype,
+					 &xcp->ddev->dev->kobj, "xcp");
+		if (r)
+			goto out;
+
+		r = sysfs_create_group(&xcp->kobj, &amdgpu_xcp_attrs_group);
+		if (r)
+			goto out;
+	}
+
+	return;
+out:
+	kobject_put(&xcp->kobj);
+}
+
+static void amdgpu_xcp_sysfs_entries_update(struct amdgpu_xcp_mgr *xcp_mgr)
+{
+	struct amdgpu_xcp *xcp;
+	int i;
+
+	for (i = 0; i < MAX_XCP; i++) {
+		/* Redirect all IOCTLs to the primary device */
+		xcp = &xcp_mgr->xcp[i];
+		if (!xcp->ddev)
+			continue;
+		sysfs_update_group(&xcp->kobj, &amdgpu_xcp_attrs_group);
+	}
+
+	return;
+}
+
+void amdgpu_xcp_sysfs_init(struct amdgpu_device *adev)
+{
+	if (!adev->xcp_mgr)
+		return;
+
+	amdgpu_xcp_cfg_sysfs_init(adev);
+
+	return;
+}
+
+void amdgpu_xcp_sysfs_fini(struct amdgpu_device *adev)
+{
+	if (!adev->xcp_mgr)
+		return;
+	amdgpu_xcp_sysfs_entries_fini(adev->xcp_mgr, MAX_XCP);
+	amdgpu_xcp_cfg_sysfs_fini(adev);
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_xcp.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_xcp.h
new file mode 100644
index 000000000000..1928d9e224fc
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_xcp.h
@@ -0,0 +1,217 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef AMDGPU_XCP_H
+#define AMDGPU_XCP_H
+
+#include <linux/pci.h>
+#include <linux/xarray.h>
+
+#include "amdgpu_ctx.h"
+
+#define MAX_XCP 8
+
+#define AMDGPU_XCP_MODE_NONE -1
+#define AMDGPU_XCP_MODE_TRANS -2
+
+#define AMDGPU_XCP_FL_NONE 0
+#define AMDGPU_XCP_FL_LOCKED (1 << 0)
+
+#define AMDGPU_XCP_NO_PARTITION (~0)
+
+#define AMDGPU_XCP_OPS_KFD	(1 << 0)
+
+struct amdgpu_fpriv;
+
+enum AMDGPU_XCP_IP_BLOCK {
+	AMDGPU_XCP_GFXHUB,
+	AMDGPU_XCP_GFX,
+	AMDGPU_XCP_SDMA,
+	AMDGPU_XCP_VCN,
+	AMDGPU_XCP_MAX_BLOCKS
+};
+
+enum AMDGPU_XCP_STATE {
+	AMDGPU_XCP_PREPARE_SUSPEND,
+	AMDGPU_XCP_SUSPEND,
+	AMDGPU_XCP_PREPARE_RESUME,
+	AMDGPU_XCP_RESUME,
+};
+
+enum amdgpu_xcp_res_id {
+	AMDGPU_XCP_RES_XCC,
+	AMDGPU_XCP_RES_DMA,
+	AMDGPU_XCP_RES_DEC,
+	AMDGPU_XCP_RES_JPEG,
+	AMDGPU_XCP_RES_MAX,
+};
+
+struct amdgpu_xcp_res_details {
+	enum amdgpu_xcp_res_id id;
+	u8 num_inst;
+	u8 num_shared;
+	struct kobject kobj;
+};
+
+struct amdgpu_xcp_cfg {
+	u8 mode;
+	struct amdgpu_xcp_res_details xcp_res[AMDGPU_XCP_RES_MAX];
+	u8 num_res;
+	struct amdgpu_xcp_mgr *xcp_mgr;
+	struct kobject kobj;
+	u16 compatible_nps_modes;
+};
+
+struct amdgpu_xcp_ip_funcs {
+	int (*prepare_suspend)(void *handle, uint32_t inst_mask);
+	int (*suspend)(void *handle, uint32_t inst_mask);
+	int (*prepare_resume)(void *handle, uint32_t inst_mask);
+	int (*resume)(void *handle, uint32_t inst_mask);
+};
+
+struct amdgpu_xcp_ip {
+	struct amdgpu_xcp_ip_funcs *ip_funcs;
+	uint32_t inst_mask;
+
+	enum AMDGPU_XCP_IP_BLOCK ip_id;
+	bool valid;
+};
+
+struct amdgpu_xcp {
+	struct amdgpu_xcp_ip ip[AMDGPU_XCP_MAX_BLOCKS];
+
+	uint8_t id;
+	uint8_t mem_id;
+	bool valid;
+	atomic_t	ref_cnt;
+	struct drm_device *ddev;
+	struct drm_device *rdev;
+	struct drm_device *pdev;
+	struct drm_driver *driver;
+	struct drm_vma_offset_manager *vma_offset_manager;
+	struct amdgpu_sched	gpu_sched[AMDGPU_HW_IP_NUM][AMDGPU_RING_PRIO_MAX];
+	struct amdgpu_xcp_mgr *xcp_mgr;
+	struct kobject kobj;
+	uint64_t unique_id;
+};
+
+struct amdgpu_xcp_mgr {
+	struct amdgpu_device *adev;
+	struct mutex xcp_lock;
+	struct amdgpu_xcp_mgr_funcs *funcs;
+
+	struct amdgpu_xcp xcp[MAX_XCP];
+	uint8_t num_xcps;
+	int8_t mode;
+
+	 /* Used to determine KFD memory size limits per XCP */
+	unsigned int num_xcp_per_mem_partition;
+	struct amdgpu_xcp_cfg *xcp_cfg;
+	uint32_t supp_xcp_modes;
+	uint32_t avail_xcp_modes;
+};
+
+struct amdgpu_xcp_mgr_funcs {
+	int (*switch_partition_mode)(struct amdgpu_xcp_mgr *xcp_mgr, int mode,
+				     int *num_xcps);
+	int (*query_partition_mode)(struct amdgpu_xcp_mgr *xcp_mgr);
+	int (*get_ip_details)(struct amdgpu_xcp_mgr *xcp_mgr, int xcp_id,
+			      enum AMDGPU_XCP_IP_BLOCK ip_id,
+			      struct amdgpu_xcp_ip *ip);
+	int (*get_xcp_mem_id)(struct amdgpu_xcp_mgr *xcp_mgr,
+			      struct amdgpu_xcp *xcp, uint8_t *mem_id);
+	int (*get_xcp_res_info)(struct amdgpu_xcp_mgr *xcp_mgr,
+				int mode,
+				struct amdgpu_xcp_cfg *xcp_cfg);
+	int (*prepare_suspend)(struct amdgpu_xcp_mgr *xcp_mgr, int xcp_id);
+	int (*suspend)(struct amdgpu_xcp_mgr *xcp_mgr, int xcp_id);
+	int (*prepare_resume)(struct amdgpu_xcp_mgr *xcp_mgr, int xcp_id);
+	int (*resume)(struct amdgpu_xcp_mgr *xcp_mgr, int xcp_id);
+};
+
+int amdgpu_xcp_prepare_suspend(struct amdgpu_xcp_mgr *xcp_mgr, int xcp_id);
+int amdgpu_xcp_suspend(struct amdgpu_xcp_mgr *xcp_mgr, int xcp_id);
+int amdgpu_xcp_prepare_resume(struct amdgpu_xcp_mgr *xcp_mgr, int xcp_id);
+int amdgpu_xcp_resume(struct amdgpu_xcp_mgr *xcp_mgr, int xcp_id);
+
+int amdgpu_xcp_mgr_init(struct amdgpu_device *adev, int init_mode,
+			int init_xcps, struct amdgpu_xcp_mgr_funcs *xcp_funcs);
+int amdgpu_xcp_init(struct amdgpu_xcp_mgr *xcp_mgr, int num_xcps, int mode);
+int amdgpu_xcp_query_partition_mode(struct amdgpu_xcp_mgr *xcp_mgr, u32 flags);
+int amdgpu_xcp_switch_partition_mode(struct amdgpu_xcp_mgr *xcp_mgr, int mode);
+int amdgpu_xcp_restore_partition_mode(struct amdgpu_xcp_mgr *xcp_mgr);
+int amdgpu_xcp_get_partition(struct amdgpu_xcp_mgr *xcp_mgr,
+			     enum AMDGPU_XCP_IP_BLOCK ip, int instance);
+
+int amdgpu_xcp_get_inst_details(struct amdgpu_xcp *xcp,
+				enum AMDGPU_XCP_IP_BLOCK ip,
+				uint32_t *inst_mask);
+
+int amdgpu_xcp_dev_register(struct amdgpu_device *adev,
+				const struct pci_device_id *ent);
+void amdgpu_xcp_dev_unplug(struct amdgpu_device *adev);
+int amdgpu_xcp_open_device(struct amdgpu_device *adev,
+			   struct amdgpu_fpriv *fpriv,
+			   struct drm_file *file_priv);
+void amdgpu_xcp_release_sched(struct amdgpu_device *adev,
+			      struct amdgpu_ctx_entity *entity);
+int amdgpu_xcp_select_scheds(struct amdgpu_device *adev,
+			     u32 hw_ip, u32 hw_prio,
+			     struct amdgpu_fpriv *fpriv,
+			     unsigned int *num_scheds,
+			     struct drm_gpu_scheduler ***scheds);
+void amdgpu_xcp_update_supported_modes(struct amdgpu_xcp_mgr *xcp_mgr);
+int amdgpu_xcp_update_partition_sched_list(struct amdgpu_device *adev);
+int amdgpu_xcp_pre_partition_switch(struct amdgpu_xcp_mgr *xcp_mgr, u32 flags);
+int amdgpu_xcp_post_partition_switch(struct amdgpu_xcp_mgr *xcp_mgr, u32 flags);
+void amdgpu_xcp_sysfs_init(struct amdgpu_device *adev);
+void amdgpu_xcp_sysfs_fini(struct amdgpu_device *adev);
+
+static inline int amdgpu_xcp_get_num_xcp(struct amdgpu_xcp_mgr *xcp_mgr)
+{
+	if (!xcp_mgr)
+		return 1;
+	else
+		return xcp_mgr->num_xcps;
+}
+
+static inline struct amdgpu_xcp *
+amdgpu_get_next_xcp(struct amdgpu_xcp_mgr *xcp_mgr, int *from)
+{
+	if (!xcp_mgr)
+		return NULL;
+
+	while (*from < MAX_XCP) {
+		if (xcp_mgr->xcp[*from].valid)
+			return &xcp_mgr->xcp[*from];
+		++(*from);
+	}
+
+	return NULL;
+}
+
+#define for_each_xcp(xcp_mgr, xcp, i)                            \
+	for (i = 0, xcp = amdgpu_get_next_xcp(xcp_mgr, &i); xcp; \
+	     ++i, xcp = amdgpu_get_next_xcp(xcp_mgr, &i))
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c
index 567df2db23ac..aad530c46a9f 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c
@@ -29,12 +29,21 @@
 #include "df/df_3_6_offset.h"
 #include "xgmi/xgmi_4_0_0_smn.h"
 #include "xgmi/xgmi_4_0_0_sh_mask.h"
+#include "xgmi/xgmi_6_1_0_sh_mask.h"
 #include "wafl/wafl2_4_0_0_smn.h"
 #include "wafl/wafl2_4_0_0_sh_mask.h"
 
-#define smnPCS_XGMI23_PCS_ERROR_STATUS   0x11a01210
+#include "amdgpu_reset.h"
+
 #define smnPCS_XGMI3X16_PCS_ERROR_STATUS 0x11a0020c
+#define smnPCS_XGMI3X16_PCS_ERROR_NONCORRECTABLE_MASK   0x11a00218
 #define smnPCS_GOPX1_PCS_ERROR_STATUS    0x12200210
+#define smnPCS_GOPX1_PCS_ERROR_NONCORRECTABLE_MASK      0x12200218
+
+#define XGMI_STATE_DISABLE                      0xD1
+#define XGMI_STATE_LS0                          0x81
+#define XGMI_LINK_ACTIVE			1
+#define XGMI_LINK_INACTIVE			0
 
 static DEFINE_MUTEX(xgmi_mutex);
 
@@ -67,17 +76,6 @@ static const int wafl_pcs_err_status_reg_arct[] = {
 	smnPCS_GOPX1_0_PCS_GOPX1_PCS_ERROR_STATUS + 0x100000,
 };
 
-static const int xgmi23_pcs_err_status_reg_aldebaran[] = {
-	smnPCS_XGMI23_PCS_ERROR_STATUS,
-	smnPCS_XGMI23_PCS_ERROR_STATUS + 0x100000,
-	smnPCS_XGMI23_PCS_ERROR_STATUS + 0x200000,
-	smnPCS_XGMI23_PCS_ERROR_STATUS + 0x300000,
-	smnPCS_XGMI23_PCS_ERROR_STATUS + 0x400000,
-	smnPCS_XGMI23_PCS_ERROR_STATUS + 0x500000,
-	smnPCS_XGMI23_PCS_ERROR_STATUS + 0x600000,
-	smnPCS_XGMI23_PCS_ERROR_STATUS + 0x700000
-};
-
 static const int xgmi3x16_pcs_err_status_reg_aldebaran[] = {
 	smnPCS_XGMI3X16_PCS_ERROR_STATUS,
 	smnPCS_XGMI3X16_PCS_ERROR_STATUS + 0x100000,
@@ -89,11 +87,74 @@ static const int xgmi3x16_pcs_err_status_reg_aldebaran[] = {
 	smnPCS_XGMI3X16_PCS_ERROR_STATUS + 0x700000
 };
 
+static const int xgmi3x16_pcs_err_noncorrectable_mask_reg_aldebaran[] = {
+	smnPCS_XGMI3X16_PCS_ERROR_NONCORRECTABLE_MASK,
+	smnPCS_XGMI3X16_PCS_ERROR_NONCORRECTABLE_MASK + 0x100000,
+	smnPCS_XGMI3X16_PCS_ERROR_NONCORRECTABLE_MASK + 0x200000,
+	smnPCS_XGMI3X16_PCS_ERROR_NONCORRECTABLE_MASK + 0x300000,
+	smnPCS_XGMI3X16_PCS_ERROR_NONCORRECTABLE_MASK + 0x400000,
+	smnPCS_XGMI3X16_PCS_ERROR_NONCORRECTABLE_MASK + 0x500000,
+	smnPCS_XGMI3X16_PCS_ERROR_NONCORRECTABLE_MASK + 0x600000,
+	smnPCS_XGMI3X16_PCS_ERROR_NONCORRECTABLE_MASK + 0x700000
+};
+
 static const int walf_pcs_err_status_reg_aldebaran[] = {
 	smnPCS_GOPX1_PCS_ERROR_STATUS,
 	smnPCS_GOPX1_PCS_ERROR_STATUS + 0x100000
 };
 
+static const int walf_pcs_err_noncorrectable_mask_reg_aldebaran[] = {
+	smnPCS_GOPX1_PCS_ERROR_NONCORRECTABLE_MASK,
+	smnPCS_GOPX1_PCS_ERROR_NONCORRECTABLE_MASK + 0x100000
+};
+
+static const int xgmi3x16_pcs_err_status_reg_v6_4[] = {
+	smnPCS_XGMI3X16_PCS_ERROR_STATUS,
+	smnPCS_XGMI3X16_PCS_ERROR_STATUS + 0x100000
+};
+
+static const int xgmi3x16_pcs_err_noncorrectable_mask_reg_v6_4[] = {
+	smnPCS_XGMI3X16_PCS_ERROR_NONCORRECTABLE_MASK,
+	smnPCS_XGMI3X16_PCS_ERROR_NONCORRECTABLE_MASK + 0x100000
+};
+
+static const u64 xgmi_v6_4_0_mca_base_array[] = {
+	0x11a09200,
+	0x11b09200,
+};
+
+static const char *xgmi_v6_4_0_ras_error_code_ext[32] = {
+	[0x00] = "XGMI PCS DataLossErr",
+	[0x01] = "XGMI PCS TrainingErr",
+	[0x02] = "XGMI PCS FlowCtrlAckErr",
+	[0x03] = "XGMI PCS RxFifoUnderflowErr",
+	[0x04] = "XGMI PCS RxFifoOverflowErr",
+	[0x05] = "XGMI PCS CRCErr",
+	[0x06] = "XGMI PCS BERExceededErr",
+	[0x07] = "XGMI PCS TxMetaDataErr",
+	[0x08] = "XGMI PCS ReplayBufParityErr",
+	[0x09] = "XGMI PCS DataParityErr",
+	[0x0a] = "XGMI PCS ReplayFifoOverflowErr",
+	[0x0b] = "XGMI PCS ReplayFifoUnderflowErr",
+	[0x0c] = "XGMI PCS ElasticFifoOverflowErr",
+	[0x0d] = "XGMI PCS DeskewErr",
+	[0x0e] = "XGMI PCS FlowCtrlCRCErr",
+	[0x0f] = "XGMI PCS DataStartupLimitErr",
+	[0x10] = "XGMI PCS FCInitTimeoutErr",
+	[0x11] = "XGMI PCS RecoveryTimeoutErr",
+	[0x12] = "XGMI PCS ReadySerialTimeoutErr",
+	[0x13] = "XGMI PCS ReadySerialAttemptErr",
+	[0x14] = "XGMI PCS RecoveryAttemptErr",
+	[0x15] = "XGMI PCS RecoveryRelockAttemptErr",
+	[0x16] = "XGMI PCS ReplayAttemptErr",
+	[0x17] = "XGMI PCS SyncHdrErr",
+	[0x18] = "XGMI PCS TxReplayTimeoutErr",
+	[0x19] = "XGMI PCS RxReplayTimeoutErr",
+	[0x1a] = "XGMI PCS LinkSubTxTimeoutErr",
+	[0x1b] = "XGMI PCS LinkSubRxTimeoutErr",
+	[0x1c] = "XGMI PCS RxCMDPktErr",
+};
+
 static const struct amdgpu_pcs_ras_field xgmi_pcs_ras_fields[] = {
 	{"XGMI PCS DataLossErr",
 	 SOC15_REG_FIELD(XGMI0_PCS_GOPX16_PCS_ERROR_STATUS, DataLossErr)},
@@ -172,6 +233,143 @@ static const struct amdgpu_pcs_ras_field wafl_pcs_ras_fields[] = {
 	 SOC15_REG_FIELD(PCS_GOPX1_0_PCS_GOPX1_PCS_ERROR_STATUS, RecoveryRelockAttemptErr)},
 };
 
+static const struct amdgpu_pcs_ras_field xgmi3x16_pcs_ras_fields[] = {
+	{"XGMI3X16 PCS DataLossErr",
+	 SOC15_REG_FIELD(PCS_XGMI3X16_PCS_ERROR_STATUS, DataLossErr)},
+	{"XGMI3X16 PCS TrainingErr",
+	 SOC15_REG_FIELD(PCS_XGMI3X16_PCS_ERROR_STATUS, TrainingErr)},
+	{"XGMI3X16 PCS FlowCtrlAckErr",
+	 SOC15_REG_FIELD(PCS_XGMI3X16_PCS_ERROR_STATUS, FlowCtrlAckErr)},
+	{"XGMI3X16 PCS RxFifoUnderflowErr",
+	 SOC15_REG_FIELD(PCS_XGMI3X16_PCS_ERROR_STATUS, RxFifoUnderflowErr)},
+	{"XGMI3X16 PCS RxFifoOverflowErr",
+	 SOC15_REG_FIELD(PCS_XGMI3X16_PCS_ERROR_STATUS, RxFifoOverflowErr)},
+	{"XGMI3X16 PCS CRCErr",
+	 SOC15_REG_FIELD(PCS_XGMI3X16_PCS_ERROR_STATUS, CRCErr)},
+	{"XGMI3X16 PCS BERExceededErr",
+	 SOC15_REG_FIELD(PCS_XGMI3X16_PCS_ERROR_STATUS, BERExceededErr)},
+	{"XGMI3X16 PCS TxVcidDataErr",
+	 SOC15_REG_FIELD(PCS_XGMI3X16_PCS_ERROR_STATUS, TxVcidDataErr)},
+	{"XGMI3X16 PCS ReplayBufParityErr",
+	 SOC15_REG_FIELD(PCS_XGMI3X16_PCS_ERROR_STATUS, ReplayBufParityErr)},
+	{"XGMI3X16 PCS DataParityErr",
+	 SOC15_REG_FIELD(PCS_XGMI3X16_PCS_ERROR_STATUS, DataParityErr)},
+	{"XGMI3X16 PCS ReplayFifoOverflowErr",
+	 SOC15_REG_FIELD(PCS_XGMI3X16_PCS_ERROR_STATUS, ReplayFifoOverflowErr)},
+	{"XGMI3X16 PCS ReplayFifoUnderflowErr",
+	 SOC15_REG_FIELD(PCS_XGMI3X16_PCS_ERROR_STATUS, ReplayFifoUnderflowErr)},
+	{"XGMI3X16 PCS ElasticFifoOverflowErr",
+	 SOC15_REG_FIELD(PCS_XGMI3X16_PCS_ERROR_STATUS, ElasticFifoOverflowErr)},
+	{"XGMI3X16 PCS DeskewErr",
+	 SOC15_REG_FIELD(PCS_XGMI3X16_PCS_ERROR_STATUS, DeskewErr)},
+	{"XGMI3X16 PCS FlowCtrlCRCErr",
+	 SOC15_REG_FIELD(PCS_XGMI3X16_PCS_ERROR_STATUS, FlowCtrlCRCErr)},
+	{"XGMI3X16 PCS DataStartupLimitErr",
+	 SOC15_REG_FIELD(PCS_XGMI3X16_PCS_ERROR_STATUS, DataStartupLimitErr)},
+	{"XGMI3X16 PCS FCInitTimeoutErr",
+	 SOC15_REG_FIELD(PCS_XGMI3X16_PCS_ERROR_STATUS, FCInitTimeoutErr)},
+	{"XGMI3X16 PCS RecoveryTimeoutErr",
+	 SOC15_REG_FIELD(PCS_XGMI3X16_PCS_ERROR_STATUS, RecoveryTimeoutErr)},
+	{"XGMI3X16 PCS ReadySerialTimeoutErr",
+	 SOC15_REG_FIELD(PCS_XGMI3X16_PCS_ERROR_STATUS, ReadySerialTimeoutErr)},
+	{"XGMI3X16 PCS ReadySerialAttemptErr",
+	 SOC15_REG_FIELD(PCS_XGMI3X16_PCS_ERROR_STATUS, ReadySerialAttemptErr)},
+	{"XGMI3X16 PCS RecoveryAttemptErr",
+	 SOC15_REG_FIELD(PCS_XGMI3X16_PCS_ERROR_STATUS, RecoveryAttemptErr)},
+	{"XGMI3X16 PCS RecoveryRelockAttemptErr",
+	 SOC15_REG_FIELD(PCS_XGMI3X16_PCS_ERROR_STATUS, RecoveryRelockAttemptErr)},
+	{"XGMI3X16 PCS ReplayAttemptErr",
+	 SOC15_REG_FIELD(PCS_XGMI3X16_PCS_ERROR_STATUS, ReplayAttemptErr)},
+	{"XGMI3X16 PCS SyncHdrErr",
+	 SOC15_REG_FIELD(PCS_XGMI3X16_PCS_ERROR_STATUS, SyncHdrErr)},
+	{"XGMI3X16 PCS TxReplayTimeoutErr",
+	 SOC15_REG_FIELD(PCS_XGMI3X16_PCS_ERROR_STATUS, TxReplayTimeoutErr)},
+	{"XGMI3X16 PCS RxReplayTimeoutErr",
+	 SOC15_REG_FIELD(PCS_XGMI3X16_PCS_ERROR_STATUS, RxReplayTimeoutErr)},
+	{"XGMI3X16 PCS LinkSubTxTimeoutErr",
+	 SOC15_REG_FIELD(PCS_XGMI3X16_PCS_ERROR_STATUS, LinkSubTxTimeoutErr)},
+	{"XGMI3X16 PCS LinkSubRxTimeoutErr",
+	 SOC15_REG_FIELD(PCS_XGMI3X16_PCS_ERROR_STATUS, LinkSubRxTimeoutErr)},
+	{"XGMI3X16 PCS RxCMDPktErr",
+	 SOC15_REG_FIELD(PCS_XGMI3X16_PCS_ERROR_STATUS, RxCMDPktErr)},
+};
+
+int amdgpu_xgmi_get_ext_link(struct amdgpu_device *adev, int link_num)
+{
+	int link_map_6_4_x[8] = { 0, 3, 1, 2, 7, 6, 4, 5 };
+
+	if (adev->gmc.xgmi.num_physical_nodes <= 1)
+		return -EINVAL;
+
+	switch (amdgpu_ip_version(adev, XGMI_HWIP, 0)) {
+	case IP_VERSION(6, 4, 0):
+	case IP_VERSION(6, 4, 1):
+		if (link_num < ARRAY_SIZE(link_map_6_4_x))
+			return link_map_6_4_x[link_num];
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return -EINVAL;
+}
+
+static u32 xgmi_v6_4_get_link_status(struct amdgpu_device *adev, int global_link_num)
+{
+	const u32 smn_xgmi_6_4_pcs_state_hist1[2] = { 0x11a00070, 0x11b00070 };
+	const u32 smn_xgmi_6_4_1_pcs_state_hist1[2] = { 0x12100070,
+							0x11b00070 };
+	u32 i, n;
+	u64 addr;
+
+	switch (amdgpu_ip_version(adev, XGMI_HWIP, 0)) {
+	case IP_VERSION(6, 4, 0):
+		n = ARRAY_SIZE(smn_xgmi_6_4_pcs_state_hist1);
+		addr = smn_xgmi_6_4_pcs_state_hist1[global_link_num % n];
+		break;
+	case IP_VERSION(6, 4, 1):
+		n = ARRAY_SIZE(smn_xgmi_6_4_1_pcs_state_hist1);
+		addr = smn_xgmi_6_4_1_pcs_state_hist1[global_link_num % n];
+		break;
+	default:
+		return U32_MAX;
+	}
+
+	i = global_link_num / n;
+
+	if (!(adev->aid_mask & BIT(i)))
+		return U32_MAX;
+
+	addr += adev->asic_funcs->encode_ext_smn_addressing(i);
+
+	return RREG32_PCIE_EXT(addr);
+}
+
+int amdgpu_get_xgmi_link_status(struct amdgpu_device *adev, int global_link_num)
+{
+	u32 xgmi_state_reg_val;
+
+	if (adev->gmc.xgmi.num_physical_nodes <= 1)
+		return -EINVAL;
+
+	switch (amdgpu_ip_version(adev, XGMI_HWIP, 0)) {
+	case IP_VERSION(6, 4, 0):
+	case IP_VERSION(6, 4, 1):
+		xgmi_state_reg_val = xgmi_v6_4_get_link_status(adev, global_link_num);
+		break;
+	default:
+		return -EOPNOTSUPP;
+	}
+
+	if ((xgmi_state_reg_val & 0xFF) == XGMI_STATE_DISABLE)
+		return -ENOLINK;
+
+	if ((xgmi_state_reg_val & 0xFF) == XGMI_STATE_LS0)
+		return XGMI_LINK_ACTIVE;
+
+	return XGMI_LINK_INACTIVE;
+}
+
 /**
  * DOC: AMDGPU XGMI Support
  *
@@ -208,6 +406,7 @@ static struct attribute *amdgpu_xgmi_hive_attrs[] = {
 	&amdgpu_xgmi_hive_id,
 	NULL
 };
+ATTRIBUTE_GROUPS(amdgpu_xgmi_hive);
 
 static ssize_t amdgpu_xgmi_show_attrs(struct kobject *kobj,
 	struct attribute *attr, char *buf)
@@ -226,6 +425,9 @@ static void amdgpu_xgmi_hive_release(struct kobject *kobj)
 	struct amdgpu_hive_info *hive = container_of(
 		kobj, struct amdgpu_hive_info, kobj);
 
+	amdgpu_reset_put_reset_domain(hive->reset_domain);
+	hive->reset_domain = NULL;
+
 	mutex_destroy(&hive->hive_lock);
 	kfree(hive);
 }
@@ -234,10 +436,10 @@ static const struct sysfs_ops amdgpu_xgmi_hive_ops = {
 	.show = amdgpu_xgmi_show_attrs,
 };
 
-struct kobj_type amdgpu_xgmi_hive_type = {
+static const struct kobj_type amdgpu_xgmi_hive_type = {
 	.release = amdgpu_xgmi_hive_release,
 	.sysfs_ops = &amdgpu_xgmi_hive_ops,
-	.default_attrs = amdgpu_xgmi_hive_attrs,
+	.default_groups = amdgpu_xgmi_hive_groups,
 };
 
 static ssize_t amdgpu_xgmi_show_device_id(struct device *dev,
@@ -251,6 +453,82 @@ static ssize_t amdgpu_xgmi_show_device_id(struct device *dev,
 
 }
 
+static ssize_t amdgpu_xgmi_show_physical_id(struct device *dev,
+				     struct device_attribute *attr,
+				     char *buf)
+{
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = drm_to_adev(ddev);
+
+	return sysfs_emit(buf, "%u\n", adev->gmc.xgmi.physical_node_id);
+
+}
+
+static ssize_t amdgpu_xgmi_show_num_hops(struct device *dev,
+					struct device_attribute *attr,
+					char *buf)
+{
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = drm_to_adev(ddev);
+	struct psp_xgmi_topology_info *top = &adev->psp.xgmi_context.top_info;
+	int i;
+
+	for (i = 0; i < top->num_nodes; i++)
+		sprintf(buf + 3 * i, "%02x ", top->nodes[i].num_hops);
+
+	return sysfs_emit(buf, "%s\n", buf);
+}
+
+static ssize_t amdgpu_xgmi_show_num_links(struct device *dev,
+					struct device_attribute *attr,
+					char *buf)
+{
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = drm_to_adev(ddev);
+	struct psp_xgmi_topology_info *top = &adev->psp.xgmi_context.top_info;
+	int i;
+
+	for (i = 0; i < top->num_nodes; i++)
+		sprintf(buf + 3 * i, "%02x ", top->nodes[i].num_links);
+
+	return sysfs_emit(buf, "%s\n", buf);
+}
+
+static ssize_t amdgpu_xgmi_show_connected_port_num(struct device *dev,
+					struct device_attribute *attr,
+					char *buf)
+{
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = drm_to_adev(ddev);
+	struct psp_xgmi_topology_info *top = &adev->psp.xgmi_context.top_info;
+	int i, j, size = 0;
+	int current_node;
+	/*
+	 * get the node id in the sysfs for the current socket and show
+	 * it in the port num info output in the sysfs for easy reading.
+	 * it is NOT the one retrieved from xgmi ta.
+	 */
+	for (i = 0; i < top->num_nodes; i++) {
+		if (top->nodes[i].node_id == adev->gmc.xgmi.node_id) {
+			current_node = i;
+			break;
+		}
+	}
+
+	if (i == top->num_nodes)
+		return -EINVAL;
+
+	for (i = 0; i < top->num_nodes; i++) {
+		for (j = 0; j < top->nodes[i].num_links; j++)
+			/* node id in sysfs starts from 1 rather than 0 so +1 here */
+			size += sysfs_emit_at(buf, size, "%02x:%02x ->  %02x:%02x\n", current_node + 1,
+					      top->nodes[i].port_num[j].src_xgmi_port_num, i + 1,
+					      top->nodes[i].port_num[j].dst_xgmi_port_num);
+	}
+
+	return size;
+}
+
 #define AMDGPU_XGMI_SET_FICAA(o)	((o) | 0x456801)
 static ssize_t amdgpu_xgmi_show_error(struct device *dev,
 				      struct device_attribute *attr,
@@ -265,6 +543,11 @@ static ssize_t amdgpu_xgmi_show_error(struct device *dev,
 	ficaa_pie_ctl_in = AMDGPU_XGMI_SET_FICAA(0x200);
 	ficaa_pie_status_in = AMDGPU_XGMI_SET_FICAA(0x208);
 
+	if ((!adev->df.funcs) ||
+	    (!adev->df.funcs->get_fica) ||
+	    (!adev->df.funcs->set_fica))
+		return -EINVAL;
+
 	fica_out = adev->df.funcs->get_fica(adev, ficaa_pie_ctl_in);
 	if (fica_out != 0x1f)
 		pr_err("xGMI error counters not enabled!\n");
@@ -281,7 +564,11 @@ static ssize_t amdgpu_xgmi_show_error(struct device *dev,
 
 
 static DEVICE_ATTR(xgmi_device_id, S_IRUGO, amdgpu_xgmi_show_device_id, NULL);
+static DEVICE_ATTR(xgmi_physical_id, 0444, amdgpu_xgmi_show_physical_id, NULL);
 static DEVICE_ATTR(xgmi_error, S_IRUGO, amdgpu_xgmi_show_error, NULL);
+static DEVICE_ATTR(xgmi_num_hops, S_IRUGO, amdgpu_xgmi_show_num_hops, NULL);
+static DEVICE_ATTR(xgmi_num_links, S_IRUGO, amdgpu_xgmi_show_num_links, NULL);
+static DEVICE_ATTR(xgmi_port_num, S_IRUGO, amdgpu_xgmi_show_connected_port_num, NULL);
 
 static int amdgpu_xgmi_sysfs_add_dev_info(struct amdgpu_device *adev,
 					 struct amdgpu_hive_info *hive)
@@ -296,11 +583,33 @@ static int amdgpu_xgmi_sysfs_add_dev_info(struct amdgpu_device *adev,
 		return ret;
 	}
 
+	ret = device_create_file(adev->dev, &dev_attr_xgmi_physical_id);
+	if (ret) {
+		dev_err(adev->dev, "XGMI: Failed to create device file xgmi_physical_id\n");
+		return ret;
+	}
+
 	/* Create xgmi error file */
 	ret = device_create_file(adev->dev, &dev_attr_xgmi_error);
 	if (ret)
 		pr_err("failed to create xgmi_error\n");
 
+	/* Create xgmi num hops file */
+	ret = device_create_file(adev->dev, &dev_attr_xgmi_num_hops);
+	if (ret)
+		pr_err("failed to create xgmi_num_hops\n");
+
+	/* Create xgmi num links file */
+	ret = device_create_file(adev->dev, &dev_attr_xgmi_num_links);
+	if (ret)
+		pr_err("failed to create xgmi_num_links\n");
+
+	/* Create xgmi port num file if supported */
+	if (adev->psp.xgmi_context.xgmi_ta_caps & EXTEND_PEER_LINK_INFO_CMD_FLAG) {
+		ret = device_create_file(adev->dev, &dev_attr_xgmi_port_num);
+		if (ret)
+			dev_err(adev->dev, "failed to create xgmi_port_num\n");
+	}
 
 	/* Create sysfs link to hive info folder on the first device */
 	if (hive->kobj.parent != (&adev->dev->kobj)) {
@@ -328,6 +637,12 @@ remove_link:
 
 remove_file:
 	device_remove_file(adev->dev, &dev_attr_xgmi_device_id);
+	device_remove_file(adev->dev, &dev_attr_xgmi_physical_id);
+	device_remove_file(adev->dev, &dev_attr_xgmi_error);
+	device_remove_file(adev->dev, &dev_attr_xgmi_num_hops);
+	device_remove_file(adev->dev, &dev_attr_xgmi_num_links);
+	if (adev->psp.xgmi_context.xgmi_ta_caps & EXTEND_PEER_LINK_INFO_CMD_FLAG)
+		device_remove_file(adev->dev, &dev_attr_xgmi_port_num);
 
 success:
 	return ret;
@@ -340,7 +655,12 @@ static void amdgpu_xgmi_sysfs_rem_dev_info(struct amdgpu_device *adev,
 	memset(node, 0, sizeof(node));
 
 	device_remove_file(adev->dev, &dev_attr_xgmi_device_id);
+	device_remove_file(adev->dev, &dev_attr_xgmi_physical_id);
 	device_remove_file(adev->dev, &dev_attr_xgmi_error);
+	device_remove_file(adev->dev, &dev_attr_xgmi_num_hops);
+	device_remove_file(adev->dev, &dev_attr_xgmi_num_links);
+	if (adev->psp.xgmi_context.xgmi_ta_caps & EXTEND_PEER_LINK_INFO_CMD_FLAG)
+		device_remove_file(adev->dev, &dev_attr_xgmi_port_num);
 
 	if (hive->kobj.parent != (&adev->dev->kobj))
 		sysfs_remove_link(&adev->dev->kobj,"xgmi_hive_info");
@@ -375,6 +695,7 @@ struct amdgpu_hive_info *amdgpu_get_xgmi_hive(struct amdgpu_device *adev)
 	hive = kzalloc(sizeof(*hive), GFP_KERNEL);
 	if (!hive) {
 		dev_err(adev->dev, "XGMI: allocation failed\n");
+		ret = -ENOMEM;
 		hive = NULL;
 		goto pro_end;
 	}
@@ -387,20 +708,48 @@ struct amdgpu_hive_info *amdgpu_get_xgmi_hive(struct amdgpu_device *adev)
 	if (ret) {
 		dev_err(adev->dev, "XGMI: failed initializing kobject for xgmi hive\n");
 		kobject_put(&hive->kobj);
-		kfree(hive);
 		hive = NULL;
 		goto pro_end;
 	}
 
+	/**
+	 * Only init hive->reset_domain for none SRIOV configuration. For SRIOV,
+	 * Host driver decide how to reset the GPU either through FLR or chain reset.
+	 * Guest side will get individual notifications from the host for the FLR
+	 * if necessary.
+	 */
+	if (!amdgpu_sriov_vf(adev)) {
+	/**
+	 * Avoid recreating reset domain when hive is reconstructed for the case
+	 * of reset the devices in the XGMI hive during probe for passthrough GPU
+	 * See https://www.spinics.net/lists/amd-gfx/msg58836.html
+	 */
+		if (adev->reset_domain->type != XGMI_HIVE) {
+			hive->reset_domain =
+				amdgpu_reset_create_reset_domain(XGMI_HIVE, "amdgpu-reset-hive");
+			if (!hive->reset_domain) {
+				dev_err(adev->dev, "XGMI: failed initializing reset domain for xgmi hive\n");
+				ret = -ENOMEM;
+				kobject_put(&hive->kobj);
+				hive = NULL;
+				goto pro_end;
+			}
+		} else {
+			amdgpu_reset_get_reset_domain(adev->reset_domain);
+			hive->reset_domain = adev->reset_domain;
+		}
+	}
+
 	hive->hive_id = adev->gmc.xgmi.hive_id;
 	INIT_LIST_HEAD(&hive->device_list);
 	INIT_LIST_HEAD(&hive->node);
 	mutex_init(&hive->hive_lock);
-	atomic_set(&hive->in_reset, 0);
 	atomic_set(&hive->number_devices, 0);
 	task_barrier_init(&hive->tb);
 	hive->pstate = AMDGPU_XGMI_PSTATE_UNKNOWN;
 	hive->hi_req_gpu = NULL;
+	atomic_set(&hive->requested_nps_mode, UNKNOWN_MEMORY_PARTITION_MODE);
+
 	/*
 	 * hive pstate on boot is high in vega20 so we have to go to low
 	 * pstate on after boot.
@@ -485,6 +834,9 @@ int amdgpu_xgmi_update_topology(struct amdgpu_hive_info *hive, struct amdgpu_dev
 {
 	int ret;
 
+	if (amdgpu_sriov_vf(adev))
+		return 0;
+
 	/* Each psp need to set the latest topology */
 	ret = psp_xgmi_set_topology_info(&adev->psp,
 					 atomic_read(&hive->number_devices),
@@ -506,28 +858,88 @@ int amdgpu_xgmi_update_topology(struct amdgpu_hive_info *hive, struct amdgpu_dev
  * num_hops[2:0] = number of hops
  */
 int amdgpu_xgmi_get_hops_count(struct amdgpu_device *adev,
-		struct amdgpu_device *peer_adev)
+			       struct amdgpu_device *peer_adev)
 {
 	struct psp_xgmi_topology_info *top = &adev->psp.xgmi_context.top_info;
 	uint8_t num_hops_mask = 0x7;
 	int i;
 
+	if (!adev->gmc.xgmi.supported)
+		return 0;
+
 	for (i = 0 ; i < top->num_nodes; ++i)
 		if (top->nodes[i].node_id == peer_adev->gmc.xgmi.node_id)
 			return top->nodes[i].num_hops & num_hops_mask;
-	return	-EINVAL;
+
+	dev_err(adev->dev, "Failed to get xgmi hops count for peer %d.\n",
+		peer_adev->gmc.xgmi.physical_node_id);
+
+	return 0;
+}
+
+int amdgpu_xgmi_get_bandwidth(struct amdgpu_device *adev, struct amdgpu_device *peer_adev,
+			      enum amdgpu_xgmi_bw_mode bw_mode, enum amdgpu_xgmi_bw_unit bw_unit,
+			      uint32_t *min_bw, uint32_t *max_bw)
+{
+	bool peer_mode = bw_mode == AMDGPU_XGMI_BW_MODE_PER_PEER;
+	int unit_scale = bw_unit == AMDGPU_XGMI_BW_UNIT_MBYTES ? 1000 : 1;
+	int num_lanes = adev->gmc.xgmi.max_width;
+	int speed = adev->gmc.xgmi.max_speed;
+	int num_links = !peer_mode ? 1 : -1;
+
+	if (!(min_bw && max_bw))
+		return -EINVAL;
+
+	*min_bw = 0;
+	*max_bw = 0;
+
+	if (!adev->gmc.xgmi.supported)
+		return -ENODATA;
+
+	if (peer_mode && !peer_adev)
+		return -EINVAL;
+
+	if (peer_mode) {
+		struct psp_xgmi_topology_info *top = &adev->psp.xgmi_context.top_info;
+		int i;
+
+		for (i = 0 ; i < top->num_nodes; ++i) {
+			if (top->nodes[i].node_id != peer_adev->gmc.xgmi.node_id)
+				continue;
+
+			num_links =  top->nodes[i].num_links;
+			break;
+		}
+	}
+
+	if (num_links == -1) {
+		dev_err(adev->dev, "Failed to get number of xgmi links for peer %d.\n",
+			peer_adev->gmc.xgmi.physical_node_id);
+	} else if (num_links) {
+		int per_link_bw = (speed * num_lanes * unit_scale)/BITS_PER_BYTE;
+
+		*min_bw = per_link_bw;
+		*max_bw = num_links * per_link_bw;
+	}
+
+	return 0;
 }
 
-int amdgpu_xgmi_get_num_links(struct amdgpu_device *adev,
-		struct amdgpu_device *peer_adev)
+bool amdgpu_xgmi_get_is_sharing_enabled(struct amdgpu_device *adev,
+					struct amdgpu_device *peer_adev)
 {
 	struct psp_xgmi_topology_info *top = &adev->psp.xgmi_context.top_info;
 	int i;
 
+	/* Sharing should always be enabled for non-SRIOV. */
+	if (!amdgpu_sriov_vf(adev))
+		return true;
+
 	for (i = 0 ; i < top->num_nodes; ++i)
 		if (top->nodes[i].node_id == peer_adev->gmc.xgmi.node_id)
-			return top->nodes[i].num_links;
-	return	-EINVAL;
+			return !!top->nodes[i].is_sharing_enabled;
+
+	return false;
 }
 
 /*
@@ -568,8 +980,7 @@ int amdgpu_xgmi_add_device(struct amdgpu_device *adev)
 	if (!adev->gmc.xgmi.supported)
 		return 0;
 
-	if (!adev->gmc.xgmi.pending_reset &&
-	    amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_PSP)) {
+	if (amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_PSP)) {
 		ret = psp_xgmi_initialize(&adev->psp, false, true);
 		if (ret) {
 			dev_err(adev->dev,
@@ -615,8 +1026,7 @@ int amdgpu_xgmi_add_device(struct amdgpu_device *adev)
 
 	task_barrier_add_task(&hive->tb);
 
-	if (!adev->gmc.xgmi.pending_reset &&
-	    amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_PSP)) {
+	if (amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_PSP)) {
 		list_for_each_entry(tmp_adev, &hive->device_list, gmc.xgmi.head) {
 			/* update node list for other device in the hive */
 			if (tmp_adev != adev) {
@@ -630,18 +1040,33 @@ int amdgpu_xgmi_add_device(struct amdgpu_device *adev)
 				goto exit_unlock;
 		}
 
-		/* get latest topology info for each device from psp */
-		list_for_each_entry(tmp_adev, &hive->device_list, gmc.xgmi.head) {
-			ret = psp_xgmi_get_topology_info(&tmp_adev->psp, count,
-					&tmp_adev->psp.xgmi_context.top_info, false);
+		if (amdgpu_sriov_vf(adev) &&
+			adev->psp.xgmi_context.xgmi_ta_caps & EXTEND_PEER_LINK_INFO_CMD_FLAG) {
+			/* only get topology for VF being init if it can support full duplex */
+			ret = psp_xgmi_get_topology_info(&adev->psp, count,
+						&adev->psp.xgmi_context.top_info, false);
 			if (ret) {
-				dev_err(tmp_adev->dev,
+				dev_err(adev->dev,
 					"XGMI: Get topology failure on device %llx, hive %llx, ret %d",
-					tmp_adev->gmc.xgmi.node_id,
-					tmp_adev->gmc.xgmi.hive_id, ret);
-				/* To do : continue with some node failed or disable the whole hive */
+					adev->gmc.xgmi.node_id,
+					adev->gmc.xgmi.hive_id, ret);
+				/* To do: continue with some node failed or disable the whole hive*/
 				goto exit_unlock;
 			}
+		} else {
+			/* get latest topology info for each device from psp */
+			list_for_each_entry(tmp_adev, &hive->device_list, gmc.xgmi.head) {
+				ret = psp_xgmi_get_topology_info(&tmp_adev->psp, count,
+					&tmp_adev->psp.xgmi_context.top_info, false);
+				if (ret) {
+					dev_err(tmp_adev->dev,
+						"XGMI: Get topology failure on device %llx, hive %llx, ret %d",
+						tmp_adev->gmc.xgmi.node_id,
+						tmp_adev->gmc.xgmi.hive_id, ret);
+					/* To do : continue with some node failed or disable the whole hive */
+					goto exit_unlock;
+				}
+			}
 		}
 
 		/* get topology again for hives that support extended data */
@@ -673,7 +1098,7 @@ int amdgpu_xgmi_add_device(struct amdgpu_device *adev)
 		}
 	}
 
-	if (!ret && !adev->gmc.xgmi.pending_reset)
+	if (!ret)
 		ret = amdgpu_xgmi_sysfs_add_dev_info(adev, hive);
 
 exit_unlock:
@@ -723,56 +1148,94 @@ int amdgpu_xgmi_remove_device(struct amdgpu_device *adev)
 		amdgpu_put_xgmi_hive(hive);
 	}
 
-	return psp_xgmi_terminate(&adev->psp);
+	return 0;
+}
+
+static int xgmi_v6_4_0_aca_bank_parser(struct aca_handle *handle, struct aca_bank *bank,
+				       enum aca_smu_type type, void *data)
+{
+	struct amdgpu_device *adev = handle->adev;
+	struct aca_bank_info info;
+	const char *error_str;
+	u64 status, count;
+	int ret, ext_error_code;
+
+	ret = aca_bank_info_decode(bank, &info);
+	if (ret)
+		return ret;
+
+	status = bank->regs[ACA_REG_IDX_STATUS];
+	ext_error_code = ACA_REG__STATUS__ERRORCODEEXT(status);
+
+	error_str = ext_error_code < ARRAY_SIZE(xgmi_v6_4_0_ras_error_code_ext) ?
+		xgmi_v6_4_0_ras_error_code_ext[ext_error_code] : NULL;
+	if (error_str)
+		dev_info(adev->dev, "%s detected\n", error_str);
+
+	count = ACA_REG__MISC0__ERRCNT(bank->regs[ACA_REG_IDX_MISC0]);
+
+	switch (type) {
+	case ACA_SMU_TYPE_UE:
+		if (ext_error_code != 0 && ext_error_code != 9)
+			count = 0ULL;
+
+		bank->aca_err_type = ACA_ERROR_TYPE_UE;
+		ret = aca_error_cache_log_bank_error(handle, &info, ACA_ERROR_TYPE_UE, count);
+		break;
+	case ACA_SMU_TYPE_CE:
+		count = ext_error_code == 6 ? count : 0ULL;
+		bank->aca_err_type = ACA_ERROR_TYPE_CE;
+		ret = aca_error_cache_log_bank_error(handle, &info, bank->aca_err_type, count);
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return ret;
 }
 
-static int amdgpu_xgmi_ras_late_init(struct amdgpu_device *adev)
+static const struct aca_bank_ops xgmi_v6_4_0_aca_bank_ops = {
+	.aca_bank_parser = xgmi_v6_4_0_aca_bank_parser,
+};
+
+static const struct aca_info xgmi_v6_4_0_aca_info = {
+	.hwip = ACA_HWIP_TYPE_PCS_XGMI,
+	.mask = ACA_ERROR_UE_MASK | ACA_ERROR_CE_MASK,
+	.bank_ops = &xgmi_v6_4_0_aca_bank_ops,
+};
+
+static int amdgpu_xgmi_ras_late_init(struct amdgpu_device *adev, struct ras_common_if *ras_block)
 {
 	int r;
-	struct ras_ih_if ih_info = {
-		.cb = NULL,
-	};
-	struct ras_fs_if fs_info = {
-		.sysfs_name = "xgmi_wafl_err_count",
-	};
 
 	if (!adev->gmc.xgmi.supported ||
 	    adev->gmc.xgmi.num_physical_nodes == 0)
 		return 0;
 
-	adev->gmc.xgmi.ras_funcs->reset_ras_error_count(adev);
+	amdgpu_ras_reset_error_count(adev, AMDGPU_RAS_BLOCK__XGMI_WAFL);
 
-	if (!adev->gmc.xgmi.ras_if) {
-		adev->gmc.xgmi.ras_if = kmalloc(sizeof(struct ras_common_if), GFP_KERNEL);
-		if (!adev->gmc.xgmi.ras_if)
-			return -ENOMEM;
-		adev->gmc.xgmi.ras_if->block = AMDGPU_RAS_BLOCK__XGMI_WAFL;
-		adev->gmc.xgmi.ras_if->type = AMDGPU_RAS_ERROR__MULTI_UNCORRECTABLE;
-		adev->gmc.xgmi.ras_if->sub_block_index = 0;
-	}
-	ih_info.head = fs_info.head = *adev->gmc.xgmi.ras_if;
-	r = amdgpu_ras_late_init(adev, adev->gmc.xgmi.ras_if,
-				 &fs_info, &ih_info);
-	if (r || !amdgpu_ras_is_supported(adev, adev->gmc.xgmi.ras_if->block)) {
-		kfree(adev->gmc.xgmi.ras_if);
-		adev->gmc.xgmi.ras_if = NULL;
+	r = amdgpu_ras_block_late_init(adev, ras_block);
+	if (r)
+		return r;
+
+	switch (amdgpu_ip_version(adev, XGMI_HWIP, 0)) {
+	case IP_VERSION(6, 4, 0):
+	case IP_VERSION(6, 4, 1):
+		r = amdgpu_ras_bind_aca(adev, AMDGPU_RAS_BLOCK__XGMI_WAFL,
+					&xgmi_v6_4_0_aca_info, NULL);
+		if (r)
+			goto late_fini;
+		break;
+	default:
+		break;
 	}
 
-	return r;
-}
+	return 0;
 
-static void amdgpu_xgmi_ras_fini(struct amdgpu_device *adev)
-{
-	if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__XGMI_WAFL) &&
-			adev->gmc.xgmi.ras_if) {
-		struct ras_common_if *ras_if = adev->gmc.xgmi.ras_if;
-		struct ras_ih_if ih_info = {
-			.cb = NULL,
-		};
+late_fini:
+	amdgpu_ras_block_late_fini(adev, ras_block);
 
-		amdgpu_ras_late_fini(adev, ras_if, &ih_info);
-		kfree(ras_if);
-	}
+	return r;
 }
 
 uint64_t amdgpu_xgmi_get_relative_phy_addr(struct amdgpu_device *adev,
@@ -788,7 +1251,7 @@ static void pcs_clear_status(struct amdgpu_device *adev, uint32_t pcs_status_reg
 	WREG32_PCIE(pcs_status_reg, 0);
 }
 
-static void amdgpu_xgmi_reset_ras_error_count(struct amdgpu_device *adev)
+static void amdgpu_xgmi_legacy_reset_ras_error_count(struct amdgpu_device *adev)
 {
 	uint32_t i;
 
@@ -804,9 +1267,6 @@ static void amdgpu_xgmi_reset_ras_error_count(struct amdgpu_device *adev)
 					 xgmi_pcs_err_status_reg_vg20[i]);
 		break;
 	case CHIP_ALDEBARAN:
-		for (i = 0; i < ARRAY_SIZE(xgmi23_pcs_err_status_reg_aldebaran); i++)
-			pcs_clear_status(adev,
-					 xgmi23_pcs_err_status_reg_aldebaran[i]);
 		for (i = 0; i < ARRAY_SIZE(xgmi3x16_pcs_err_status_reg_aldebaran); i++)
 			pcs_clear_status(adev,
 					 xgmi3x16_pcs_err_status_reg_aldebaran[i]);
@@ -817,58 +1277,116 @@ static void amdgpu_xgmi_reset_ras_error_count(struct amdgpu_device *adev)
 	default:
 		break;
 	}
+
+	switch (amdgpu_ip_version(adev, XGMI_HWIP, 0)) {
+	case IP_VERSION(6, 4, 0):
+	case IP_VERSION(6, 4, 1):
+		for (i = 0; i < ARRAY_SIZE(xgmi3x16_pcs_err_status_reg_v6_4); i++)
+			pcs_clear_status(adev,
+					xgmi3x16_pcs_err_status_reg_v6_4[i]);
+		break;
+	default:
+		break;
+	}
+}
+
+static void __xgmi_v6_4_0_reset_error_count(struct amdgpu_device *adev, int xgmi_inst, u64 mca_base)
+{
+	WREG64_MCA(xgmi_inst, mca_base, ACA_REG_IDX_STATUS, 0ULL);
+}
+
+static void xgmi_v6_4_0_reset_error_count(struct amdgpu_device *adev, int xgmi_inst)
+{
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(xgmi_v6_4_0_mca_base_array); i++)
+		__xgmi_v6_4_0_reset_error_count(adev, xgmi_inst, xgmi_v6_4_0_mca_base_array[i]);
+}
+
+static void xgmi_v6_4_0_reset_ras_error_count(struct amdgpu_device *adev)
+{
+	int i;
+
+	for_each_inst(i, adev->aid_mask)
+		xgmi_v6_4_0_reset_error_count(adev, i);
+}
+
+static void amdgpu_xgmi_reset_ras_error_count(struct amdgpu_device *adev)
+{
+	switch (amdgpu_ip_version(adev, XGMI_HWIP, 0)) {
+	case IP_VERSION(6, 4, 0):
+	case IP_VERSION(6, 4, 1):
+		xgmi_v6_4_0_reset_ras_error_count(adev);
+		break;
+	default:
+		amdgpu_xgmi_legacy_reset_ras_error_count(adev);
+		break;
+	}
 }
 
 static int amdgpu_xgmi_query_pcs_error_status(struct amdgpu_device *adev,
 					      uint32_t value,
+						  uint32_t mask_value,
 					      uint32_t *ue_count,
 					      uint32_t *ce_count,
-					      bool is_xgmi_pcs)
+					      bool is_xgmi_pcs,
+						  bool check_mask)
 {
 	int i;
-	int ue_cnt;
+	int ue_cnt = 0;
+	const struct amdgpu_pcs_ras_field *pcs_ras_fields = NULL;
+	uint32_t field_array_size = 0;
 
 	if (is_xgmi_pcs) {
-		/* query xgmi pcs error status,
-		 * only ue is supported */
-		for (i = 0; i < ARRAY_SIZE(xgmi_pcs_ras_fields); i ++) {
-			ue_cnt = (value &
-				  xgmi_pcs_ras_fields[i].pcs_err_mask) >>
-				  xgmi_pcs_ras_fields[i].pcs_err_shift;
-			if (ue_cnt) {
-				dev_info(adev->dev, "%s detected\n",
-					 xgmi_pcs_ras_fields[i].err_name);
-				*ue_count += ue_cnt;
-			}
+		if (amdgpu_ip_version(adev, XGMI_HWIP, 0) ==
+		    IP_VERSION(6, 1, 0) ||
+		    amdgpu_ip_version(adev, XGMI_HWIP, 0) ==
+		    IP_VERSION(6, 4, 0) ||
+		    amdgpu_ip_version(adev, XGMI_HWIP, 0) ==
+		    IP_VERSION(6, 4, 1)) {
+			pcs_ras_fields = &xgmi3x16_pcs_ras_fields[0];
+			field_array_size = ARRAY_SIZE(xgmi3x16_pcs_ras_fields);
+		} else {
+			pcs_ras_fields = &xgmi_pcs_ras_fields[0];
+			field_array_size = ARRAY_SIZE(xgmi_pcs_ras_fields);
 		}
 	} else {
-		/* query wafl pcs error status,
-		 * only ue is supported */
-		for (i = 0; i < ARRAY_SIZE(wafl_pcs_ras_fields); i++) {
-			ue_cnt = (value &
-				  wafl_pcs_ras_fields[i].pcs_err_mask) >>
-				  wafl_pcs_ras_fields[i].pcs_err_shift;
-			if (ue_cnt) {
-				dev_info(adev->dev, "%s detected\n",
-					 wafl_pcs_ras_fields[i].err_name);
-				*ue_count += ue_cnt;
-			}
+		pcs_ras_fields = &wafl_pcs_ras_fields[0];
+		field_array_size = ARRAY_SIZE(wafl_pcs_ras_fields);
+	}
+
+	if (check_mask)
+		value = value & ~mask_value;
+
+	/* query xgmi/walf pcs error status,
+	 * only ue is supported */
+	for (i = 0; value && i < field_array_size; i++) {
+		ue_cnt = (value &
+				pcs_ras_fields[i].pcs_err_mask) >>
+				pcs_ras_fields[i].pcs_err_shift;
+		if (ue_cnt) {
+			dev_info(adev->dev, "%s detected\n",
+				 pcs_ras_fields[i].err_name);
+			*ue_count += ue_cnt;
 		}
+
+		/* reset bit value if the bit is checked */
+		value &= ~(pcs_ras_fields[i].pcs_err_mask);
 	}
 
 	return 0;
 }
 
-static int amdgpu_xgmi_query_ras_error_count(struct amdgpu_device *adev,
-					     void *ras_error_status)
+static void amdgpu_xgmi_legacy_query_ras_error_count(struct amdgpu_device *adev,
+						     void *ras_error_status)
 {
 	struct ras_err_data *err_data = (struct ras_err_data *)ras_error_status;
-	int i;
-	uint32_t data;
+	int i, supported = 1;
+	uint32_t data, mask_data = 0;
 	uint32_t ue_cnt = 0, ce_cnt = 0;
 
 	if (!amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__XGMI_WAFL))
-		return -EINVAL;
+		return ;
 
 	err_data->ue_count = 0;
 	err_data->ce_count = 0;
@@ -879,15 +1397,15 @@ static int amdgpu_xgmi_query_ras_error_count(struct amdgpu_device *adev,
 		for (i = 0; i < ARRAY_SIZE(xgmi_pcs_err_status_reg_arct); i++) {
 			data = RREG32_PCIE(xgmi_pcs_err_status_reg_arct[i]);
 			if (data)
-				amdgpu_xgmi_query_pcs_error_status(adev,
-						data, &ue_cnt, &ce_cnt, true);
+				amdgpu_xgmi_query_pcs_error_status(adev, data,
+						mask_data, &ue_cnt, &ce_cnt, true, false);
 		}
 		/* check wafl pcs error */
 		for (i = 0; i < ARRAY_SIZE(wafl_pcs_err_status_reg_arct); i++) {
 			data = RREG32_PCIE(wafl_pcs_err_status_reg_arct[i]);
 			if (data)
-				amdgpu_xgmi_query_pcs_error_status(adev,
-						data, &ue_cnt, &ce_cnt, false);
+				amdgpu_xgmi_query_pcs_error_status(adev, data,
+						mask_data, &ue_cnt, &ce_cnt, false, false);
 		}
 		break;
 	case CHIP_VEGA20:
@@ -895,56 +1413,366 @@ static int amdgpu_xgmi_query_ras_error_count(struct amdgpu_device *adev,
 		for (i = 0; i < ARRAY_SIZE(xgmi_pcs_err_status_reg_vg20); i++) {
 			data = RREG32_PCIE(xgmi_pcs_err_status_reg_vg20[i]);
 			if (data)
-				amdgpu_xgmi_query_pcs_error_status(adev,
-						data, &ue_cnt, &ce_cnt, true);
+				amdgpu_xgmi_query_pcs_error_status(adev, data,
+						mask_data, &ue_cnt, &ce_cnt, true, false);
 		}
 		/* check wafl pcs error */
 		for (i = 0; i < ARRAY_SIZE(wafl_pcs_err_status_reg_vg20); i++) {
 			data = RREG32_PCIE(wafl_pcs_err_status_reg_vg20[i]);
 			if (data)
-				amdgpu_xgmi_query_pcs_error_status(adev,
-						data, &ue_cnt, &ce_cnt, false);
+				amdgpu_xgmi_query_pcs_error_status(adev, data,
+						mask_data, &ue_cnt, &ce_cnt, false, false);
 		}
 		break;
 	case CHIP_ALDEBARAN:
-		/* check xgmi23 pcs error */
-		for (i = 0; i < ARRAY_SIZE(xgmi23_pcs_err_status_reg_aldebaran); i++) {
-			data = RREG32_PCIE(xgmi23_pcs_err_status_reg_aldebaran[i]);
-			if (data)
-				amdgpu_xgmi_query_pcs_error_status(adev,
-						data, &ue_cnt, &ce_cnt, true);
-		}
 		/* check xgmi3x16 pcs error */
 		for (i = 0; i < ARRAY_SIZE(xgmi3x16_pcs_err_status_reg_aldebaran); i++) {
 			data = RREG32_PCIE(xgmi3x16_pcs_err_status_reg_aldebaran[i]);
+			mask_data =
+				RREG32_PCIE(xgmi3x16_pcs_err_noncorrectable_mask_reg_aldebaran[i]);
 			if (data)
-				amdgpu_xgmi_query_pcs_error_status(adev,
-						data, &ue_cnt, &ce_cnt, true);
+				amdgpu_xgmi_query_pcs_error_status(adev, data,
+						mask_data, &ue_cnt, &ce_cnt, true, true);
 		}
 		/* check wafl pcs error */
 		for (i = 0; i < ARRAY_SIZE(walf_pcs_err_status_reg_aldebaran); i++) {
 			data = RREG32_PCIE(walf_pcs_err_status_reg_aldebaran[i]);
+			mask_data =
+				RREG32_PCIE(walf_pcs_err_noncorrectable_mask_reg_aldebaran[i]);
+			if (data)
+				amdgpu_xgmi_query_pcs_error_status(adev, data,
+						mask_data, &ue_cnt, &ce_cnt, false, true);
+		}
+		break;
+	default:
+		supported = 0;
+		break;
+	}
+
+	switch (amdgpu_ip_version(adev, XGMI_HWIP, 0)) {
+	case IP_VERSION(6, 4, 0):
+	case IP_VERSION(6, 4, 1):
+		/* check xgmi3x16 pcs error */
+		for (i = 0; i < ARRAY_SIZE(xgmi3x16_pcs_err_status_reg_v6_4); i++) {
+			data = RREG32_PCIE(xgmi3x16_pcs_err_status_reg_v6_4[i]);
+			mask_data =
+				RREG32_PCIE(xgmi3x16_pcs_err_noncorrectable_mask_reg_v6_4[i]);
 			if (data)
-				amdgpu_xgmi_query_pcs_error_status(adev,
-						data, &ue_cnt, &ce_cnt, false);
+				amdgpu_xgmi_query_pcs_error_status(adev, data,
+						mask_data, &ue_cnt, &ce_cnt, true, true);
 		}
 		break;
 	default:
-		dev_warn(adev->dev, "XGMI RAS error query not supported");
+		if (!supported)
+			dev_warn(adev->dev, "XGMI RAS error query not supported");
 		break;
 	}
 
-	adev->gmc.xgmi.ras_funcs->reset_ras_error_count(adev);
+	amdgpu_ras_reset_error_count(adev, AMDGPU_RAS_BLOCK__XGMI_WAFL);
 
 	err_data->ue_count += ue_cnt;
 	err_data->ce_count += ce_cnt;
+}
 
-	return 0;
+static enum aca_error_type xgmi_v6_4_0_pcs_mca_get_error_type(struct amdgpu_device *adev, u64 status)
+{
+	const char *error_str;
+	int ext_error_code;
+
+	ext_error_code = ACA_REG__STATUS__ERRORCODEEXT(status);
+
+	error_str = ext_error_code < ARRAY_SIZE(xgmi_v6_4_0_ras_error_code_ext) ?
+		xgmi_v6_4_0_ras_error_code_ext[ext_error_code] : NULL;
+	if (error_str)
+		dev_info(adev->dev, "%s detected\n", error_str);
+
+	switch (ext_error_code) {
+	case 0:
+		return ACA_ERROR_TYPE_UE;
+	case 6:
+		return ACA_ERROR_TYPE_CE;
+	default:
+		return -EINVAL;
+	}
+
+	return -EINVAL;
+}
+
+static void __xgmi_v6_4_0_query_error_count(struct amdgpu_device *adev, struct amdgpu_smuio_mcm_config_info *mcm_info,
+					    u64 mca_base, struct ras_err_data *err_data)
+{
+	int xgmi_inst = mcm_info->die_id;
+	u64 status = 0;
+
+	status = RREG64_MCA(xgmi_inst, mca_base, ACA_REG_IDX_STATUS);
+	if (!ACA_REG__STATUS__VAL(status))
+		return;
+
+	switch (xgmi_v6_4_0_pcs_mca_get_error_type(adev, status)) {
+	case ACA_ERROR_TYPE_UE:
+		amdgpu_ras_error_statistic_ue_count(err_data, mcm_info, 1ULL);
+		break;
+	case ACA_ERROR_TYPE_CE:
+		amdgpu_ras_error_statistic_ce_count(err_data, mcm_info, 1ULL);
+		break;
+	default:
+		break;
+	}
+
+	WREG64_MCA(xgmi_inst, mca_base, ACA_REG_IDX_STATUS, 0ULL);
 }
 
-const struct amdgpu_xgmi_ras_funcs xgmi_ras_funcs = {
-	.ras_late_init = amdgpu_xgmi_ras_late_init,
-	.ras_fini = amdgpu_xgmi_ras_fini,
+static void xgmi_v6_4_0_query_error_count(struct amdgpu_device *adev, int xgmi_inst, struct ras_err_data *err_data)
+{
+	struct amdgpu_smuio_mcm_config_info mcm_info = {
+		.socket_id = adev->smuio.funcs->get_socket_id(adev),
+		.die_id = xgmi_inst,
+	};
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(xgmi_v6_4_0_mca_base_array); i++)
+		__xgmi_v6_4_0_query_error_count(adev, &mcm_info, xgmi_v6_4_0_mca_base_array[i], err_data);
+}
+
+static void xgmi_v6_4_0_query_ras_error_count(struct amdgpu_device *adev, void *ras_error_status)
+{
+	struct ras_err_data *err_data = (struct ras_err_data *)ras_error_status;
+	int i;
+
+	for_each_inst(i, adev->aid_mask)
+		xgmi_v6_4_0_query_error_count(adev, i, err_data);
+}
+
+static void amdgpu_xgmi_query_ras_error_count(struct amdgpu_device *adev,
+					      void *ras_error_status)
+{
+	switch (amdgpu_ip_version(adev, XGMI_HWIP, 0)) {
+	case IP_VERSION(6, 4, 0):
+	case IP_VERSION(6, 4, 1):
+		xgmi_v6_4_0_query_ras_error_count(adev, ras_error_status);
+		break;
+	default:
+		amdgpu_xgmi_legacy_query_ras_error_count(adev, ras_error_status);
+		break;
+	}
+}
+
+/* Trigger XGMI/WAFL error */
+static int amdgpu_ras_error_inject_xgmi(struct amdgpu_device *adev,
+			void *inject_if, uint32_t instance_mask)
+{
+	int ret1, ret2;
+	struct ta_ras_trigger_error_input *block_info =
+				(struct ta_ras_trigger_error_input *)inject_if;
+
+	if (amdgpu_dpm_set_df_cstate(adev, DF_CSTATE_DISALLOW))
+		dev_warn(adev->dev, "Failed to disallow df cstate");
+
+	ret1 = amdgpu_dpm_set_pm_policy(adev, PP_PM_POLICY_XGMI_PLPD, XGMI_PLPD_DISALLOW);
+	if (ret1 && ret1 != -EOPNOTSUPP)
+		dev_warn(adev->dev, "Failed to disallow XGMI power down");
+
+	ret2 = psp_ras_trigger_error(&adev->psp, block_info, instance_mask);
+
+	if (amdgpu_ras_intr_triggered())
+		return ret2;
+
+	ret1 = amdgpu_dpm_set_pm_policy(adev, PP_PM_POLICY_XGMI_PLPD, XGMI_PLPD_DEFAULT);
+	if (ret1 && ret1 != -EOPNOTSUPP)
+		dev_warn(adev->dev, "Failed to allow XGMI power down");
+
+	if (amdgpu_dpm_set_df_cstate(adev, DF_CSTATE_ALLOW))
+		dev_warn(adev->dev, "Failed to allow df cstate");
+
+	return ret2;
+}
+
+struct amdgpu_ras_block_hw_ops  xgmi_ras_hw_ops = {
 	.query_ras_error_count = amdgpu_xgmi_query_ras_error_count,
 	.reset_ras_error_count = amdgpu_xgmi_reset_ras_error_count,
+	.ras_error_inject = amdgpu_ras_error_inject_xgmi,
+};
+
+struct amdgpu_xgmi_ras xgmi_ras = {
+	.ras_block = {
+		.hw_ops = &xgmi_ras_hw_ops,
+		.ras_late_init = amdgpu_xgmi_ras_late_init,
+	},
 };
+
+int amdgpu_xgmi_ras_sw_init(struct amdgpu_device *adev)
+{
+	int err;
+	struct amdgpu_xgmi_ras *ras;
+
+	if (!adev->gmc.xgmi.ras)
+		return 0;
+
+	ras = adev->gmc.xgmi.ras;
+	err = amdgpu_ras_register_ras_block(adev, &ras->ras_block);
+	if (err) {
+		dev_err(adev->dev, "Failed to register xgmi_wafl_pcs ras block!\n");
+		return err;
+	}
+
+	strcpy(ras->ras_block.ras_comm.name, "xgmi_wafl");
+	ras->ras_block.ras_comm.block = AMDGPU_RAS_BLOCK__XGMI_WAFL;
+	ras->ras_block.ras_comm.type = AMDGPU_RAS_ERROR__MULTI_UNCORRECTABLE;
+	adev->gmc.xgmi.ras_if = &ras->ras_block.ras_comm;
+
+	return 0;
+}
+
+static void amdgpu_xgmi_reset_on_init_work(struct work_struct *work)
+{
+	struct amdgpu_hive_info *hive =
+		container_of(work, struct amdgpu_hive_info, reset_on_init_work);
+	struct amdgpu_reset_context reset_context;
+	struct amdgpu_device *tmp_adev;
+	struct list_head device_list;
+	int r;
+
+	mutex_lock(&hive->hive_lock);
+
+	INIT_LIST_HEAD(&device_list);
+	list_for_each_entry(tmp_adev, &hive->device_list, gmc.xgmi.head)
+		list_add_tail(&tmp_adev->reset_list, &device_list);
+
+	tmp_adev = list_first_entry(&device_list, struct amdgpu_device,
+				    reset_list);
+	amdgpu_device_lock_reset_domain(tmp_adev->reset_domain);
+
+	reset_context.method = AMD_RESET_METHOD_ON_INIT;
+	reset_context.reset_req_dev = tmp_adev;
+	reset_context.hive = hive;
+	reset_context.reset_device_list = &device_list;
+	set_bit(AMDGPU_NEED_FULL_RESET, &reset_context.flags);
+	set_bit(AMDGPU_SKIP_COREDUMP, &reset_context.flags);
+
+	amdgpu_reset_do_xgmi_reset_on_init(&reset_context);
+	mutex_unlock(&hive->hive_lock);
+	amdgpu_device_unlock_reset_domain(tmp_adev->reset_domain);
+
+	list_for_each_entry(tmp_adev, &hive->device_list, gmc.xgmi.head) {
+		r = amdgpu_ras_init_badpage_info(tmp_adev);
+		if (r && r != -EHWPOISON)
+			dev_err(tmp_adev->dev,
+				"error during bad page data initialization");
+	}
+}
+
+static void amdgpu_xgmi_schedule_reset_on_init(struct amdgpu_hive_info *hive)
+{
+	INIT_WORK(&hive->reset_on_init_work, amdgpu_xgmi_reset_on_init_work);
+	amdgpu_reset_domain_schedule(hive->reset_domain,
+				     &hive->reset_on_init_work);
+}
+
+int amdgpu_xgmi_reset_on_init(struct amdgpu_device *adev)
+{
+	struct amdgpu_hive_info *hive;
+	bool reset_scheduled;
+	int num_devs;
+
+	hive = amdgpu_get_xgmi_hive(adev);
+	if (!hive)
+		return -EINVAL;
+
+	mutex_lock(&hive->hive_lock);
+	num_devs = atomic_read(&hive->number_devices);
+	reset_scheduled = false;
+	if (num_devs == adev->gmc.xgmi.num_physical_nodes) {
+		amdgpu_xgmi_schedule_reset_on_init(hive);
+		reset_scheduled = true;
+	}
+
+	mutex_unlock(&hive->hive_lock);
+	amdgpu_put_xgmi_hive(hive);
+
+	if (reset_scheduled)
+		flush_work(&hive->reset_on_init_work);
+
+	return 0;
+}
+
+int amdgpu_xgmi_request_nps_change(struct amdgpu_device *adev,
+				   struct amdgpu_hive_info *hive,
+				   int req_nps_mode)
+{
+	struct amdgpu_device *tmp_adev;
+	int cur_nps_mode, r;
+
+	/* This is expected to be called only during unload of driver. The
+	 * request needs to be placed only once for all devices in the hive. If
+	 * one of them fail, revert the request for previous successful devices.
+	 * After placing the request, make hive mode as UNKNOWN so that other
+	 * devices don't request anymore.
+	 */
+	mutex_lock(&hive->hive_lock);
+	if (atomic_read(&hive->requested_nps_mode) ==
+	    UNKNOWN_MEMORY_PARTITION_MODE) {
+		dev_dbg(adev->dev, "Unexpected entry for hive NPS change");
+		mutex_unlock(&hive->hive_lock);
+		return 0;
+	}
+	list_for_each_entry(tmp_adev, &hive->device_list, gmc.xgmi.head) {
+		r = adev->gmc.gmc_funcs->request_mem_partition_mode(
+			tmp_adev, req_nps_mode);
+		if (r)
+			break;
+	}
+	if (r) {
+		/* Request back current mode if one of the requests failed */
+		cur_nps_mode =
+			adev->gmc.gmc_funcs->query_mem_partition_mode(tmp_adev);
+		list_for_each_entry_continue_reverse(
+			tmp_adev, &hive->device_list, gmc.xgmi.head)
+			adev->gmc.gmc_funcs->request_mem_partition_mode(
+				tmp_adev, cur_nps_mode);
+	}
+	/* Set to UNKNOWN so that other devices don't request anymore */
+	atomic_set(&hive->requested_nps_mode, UNKNOWN_MEMORY_PARTITION_MODE);
+	mutex_unlock(&hive->hive_lock);
+
+	return r;
+}
+
+bool amdgpu_xgmi_same_hive(struct amdgpu_device *adev,
+			   struct amdgpu_device *bo_adev)
+{
+	return (amdgpu_use_xgmi_p2p && adev != bo_adev &&
+		adev->gmc.xgmi.hive_id &&
+		adev->gmc.xgmi.hive_id == bo_adev->gmc.xgmi.hive_id);
+}
+
+void amdgpu_xgmi_early_init(struct amdgpu_device *adev)
+{
+	if (!adev->gmc.xgmi.supported)
+		return;
+
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	case IP_VERSION(9, 4, 0):
+	case IP_VERSION(9, 4, 1):
+	case IP_VERSION(9, 4, 2):
+		/* 25 GT/s */
+		adev->gmc.xgmi.max_speed = 25;
+		adev->gmc.xgmi.max_width = 16;
+		break;
+	case IP_VERSION(9, 4, 3):
+	case IP_VERSION(9, 4, 4):
+	case IP_VERSION(9, 5, 0):
+		/* 32 GT/s */
+		adev->gmc.xgmi.max_speed = 32;
+		adev->gmc.xgmi.max_width = 16;
+		break;
+	default:
+		break;
+	}
+}
+
+void amgpu_xgmi_set_max_speed_width(struct amdgpu_device *adev,
+				    uint16_t max_speed, uint8_t max_width)
+{
+	adev->gmc.xgmi.max_speed = max_speed;
+	adev->gmc.xgmi.max_width = max_width;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.h
index d2189bf7d428..5f36aff17e79 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.h
@@ -23,8 +23,7 @@
 #define __AMDGPU_XGMI_H__
 
 #include <drm/task_barrier.h>
-#include "amdgpu_psp.h"
-
+#include "amdgpu_ras.h"
 
 struct amdgpu_hive_info {
 	struct kobject kobj;
@@ -33,7 +32,6 @@ struct amdgpu_hive_info {
 	struct list_head node;
 	atomic_t number_devices;
 	struct mutex hive_lock;
-	atomic_t in_reset;
 	int hi_req_count;
 	struct amdgpu_device *hi_req_gpu;
 	struct task_barrier tb;
@@ -42,6 +40,12 @@ struct amdgpu_hive_info {
 		AMDGPU_XGMI_PSTATE_MAX_VEGA20,
 		AMDGPU_XGMI_PSTATE_UNKNOWN
 	} pstate;
+
+	struct amdgpu_reset_domain *reset_domain;
+	atomic_t ras_recovery;
+	struct ras_event_manager event_mgr;
+	struct work_struct reset_on_init_work;
+	atomic_t requested_nps_mode;
 };
 
 struct amdgpu_pcs_ras_field {
@@ -50,25 +54,80 @@ struct amdgpu_pcs_ras_field {
 	uint32_t pcs_err_shift;
 };
 
-extern const struct amdgpu_xgmi_ras_funcs xgmi_ras_funcs;
+/**
+ * Bandwidth range reporting comes in two modes.
+ *
+ * PER_LINK - range for any xgmi link
+ * PER_PEER - range of max of single xgmi link to max of multiple links based on source peer
+ */
+enum amdgpu_xgmi_bw_mode {
+	AMDGPU_XGMI_BW_MODE_PER_LINK = 0,
+	AMDGPU_XGMI_BW_MODE_PER_PEER
+};
+
+enum amdgpu_xgmi_bw_unit {
+	AMDGPU_XGMI_BW_UNIT_GBYTES = 0,
+	AMDGPU_XGMI_BW_UNIT_MBYTES
+};
+
+struct amdgpu_xgmi_ras {
+	struct amdgpu_ras_block_object ras_block;
+};
+extern struct amdgpu_xgmi_ras xgmi_ras;
+
+struct amdgpu_xgmi {
+	/* from psp */
+	u64 node_id;
+	u64 hive_id;
+	/* fixed per family */
+	u64 node_segment_size;
+	/* physical node (0-3) */
+	unsigned physical_node_id;
+	/* number of nodes (0-4) */
+	unsigned num_physical_nodes;
+	/* gpu list in the same hive */
+	struct list_head head;
+	bool supported;
+	struct ras_common_if *ras_if;
+	bool connected_to_cpu;
+	struct amdgpu_xgmi_ras *ras;
+	uint16_t max_speed;
+	uint8_t max_width;
+};
+
 struct amdgpu_hive_info *amdgpu_get_xgmi_hive(struct amdgpu_device *adev);
 void amdgpu_put_xgmi_hive(struct amdgpu_hive_info *hive);
 int amdgpu_xgmi_update_topology(struct amdgpu_hive_info *hive, struct amdgpu_device *adev);
 int amdgpu_xgmi_add_device(struct amdgpu_device *adev);
 int amdgpu_xgmi_remove_device(struct amdgpu_device *adev);
 int amdgpu_xgmi_set_pstate(struct amdgpu_device *adev, int pstate);
-int amdgpu_xgmi_get_hops_count(struct amdgpu_device *adev,
-		struct amdgpu_device *peer_adev);
-int amdgpu_xgmi_get_num_links(struct amdgpu_device *adev,
-		struct amdgpu_device *peer_adev);
+int amdgpu_xgmi_get_hops_count(struct amdgpu_device *adev, struct amdgpu_device *peer_adev);
+int amdgpu_xgmi_get_bandwidth(struct amdgpu_device *adev, struct amdgpu_device *peer_adev,
+			      enum amdgpu_xgmi_bw_mode bw_mode, enum amdgpu_xgmi_bw_unit bw_unit,
+			      uint32_t *min_bw, uint32_t *max_bw);
+bool amdgpu_xgmi_get_is_sharing_enabled(struct amdgpu_device *adev,
+					struct amdgpu_device *peer_adev);
 uint64_t amdgpu_xgmi_get_relative_phy_addr(struct amdgpu_device *adev,
 					   uint64_t addr);
-static inline bool amdgpu_xgmi_same_hive(struct amdgpu_device *adev,
-		struct amdgpu_device *bo_adev)
-{
-	return (adev != bo_adev &&
-		adev->gmc.xgmi.hive_id &&
-		adev->gmc.xgmi.hive_id == bo_adev->gmc.xgmi.hive_id);
-}
+bool amdgpu_xgmi_same_hive(struct amdgpu_device *adev,
+			   struct amdgpu_device *bo_adev);
+int amdgpu_xgmi_ras_sw_init(struct amdgpu_device *adev);
+int amdgpu_xgmi_reset_on_init(struct amdgpu_device *adev);
+
+int amdgpu_xgmi_request_nps_change(struct amdgpu_device *adev,
+				   struct amdgpu_hive_info *hive,
+				   int req_nps_mode);
+int amdgpu_get_xgmi_link_status(struct amdgpu_device *adev,
+				int global_link_num);
+int amdgpu_xgmi_get_ext_link(struct amdgpu_device *adev, int link_num);
+
+void amdgpu_xgmi_early_init(struct amdgpu_device *adev);
+uint32_t amdgpu_xgmi_get_max_bandwidth(struct amdgpu_device *adev);
+
+void amgpu_xgmi_set_max_speed_width(struct amdgpu_device *adev,
+				    uint16_t max_speed, uint8_t max_width);
+
+/* Cleanup macro for use with __free(xgmi_put_hive) */
+DEFINE_FREE(xgmi_put_hive, struct amdgpu_hive_info *, if (_T) amdgpu_put_xgmi_hive(_T))
 
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgv_sriovmsg.h b/drivers/gpu/drm/amd/amdgpu/amdgv_sriovmsg.h
index 7326b6c1b71c..3cdb1e0eca37 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgv_sriovmsg.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgv_sriovmsg.h
@@ -1,45 +1,106 @@
 /*
- * Copyright 2018-2019 Advanced Micro Devices, Inc.
+ * Copyright (c) 2018-2021 Advanced Micro Devices, Inc. All rights reserved.
  *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
  *
  * The above copyright notice and this permission notice shall be included in
  * all copies or substantial portions of the Software.
  *
  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
  * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
  */
 
 #ifndef AMDGV_SRIOV_MSG__H_
 #define AMDGV_SRIOV_MSG__H_
 
-/* unit in kilobytes */
-#define AMD_SRIOV_MSG_VBIOS_OFFSET              0
-#define AMD_SRIOV_MSG_VBIOS_SIZE_KB             64
-#define AMD_SRIOV_MSG_DATAEXCHANGE_OFFSET_KB    AMD_SRIOV_MSG_VBIOS_SIZE_KB
-#define AMD_SRIOV_MSG_DATAEXCHANGE_SIZE_KB      4
+#define AMD_SRIOV_MSG_SIZE_KB                           1
 
 /*
- * layout
- * 0           64KB        65KB        66KB
- * |   VBIOS   |   PF2VF   |   VF2PF   |   Bad Page   | ...
- * |   64KB    |   1KB     |   1KB     |
+ * layout v1
+ * 0           64KB        65KB        66KB           68KB                   132KB
+ * |   VBIOS   |   PF2VF   |   VF2PF   |   Bad Page   | RAS Telemetry Region | ...
+ * |   64KB    |   1KB     |   1KB     |   2KB        | 64KB                 | ...
  */
-#define AMD_SRIOV_MSG_SIZE_KB                   1
-#define AMD_SRIOV_MSG_PF2VF_OFFSET_KB           AMD_SRIOV_MSG_DATAEXCHANGE_OFFSET_KB
-#define AMD_SRIOV_MSG_VF2PF_OFFSET_KB           (AMD_SRIOV_MSG_PF2VF_OFFSET_KB + AMD_SRIOV_MSG_SIZE_KB)
-#define AMD_SRIOV_MSG_BAD_PAGE_OFFSET_KB        (AMD_SRIOV_MSG_VF2PF_OFFSET_KB + AMD_SRIOV_MSG_SIZE_KB)
+
+/*
+ * layout v2 (offsets are dynamically allocated and the offsets below are examples)
+ * 0           1KB         64KB        65KB        66KB           68KB                   132KB
+ * |  INITD_H  |   VBIOS   |   PF2VF   |   VF2PF   |   Bad Page   | RAS Telemetry Region | ...
+ * |   1KB     |   64KB    |   1KB     |   1KB     |   2KB        | 64KB                 | ...
+ *
+ * Note: PF2VF + VF2PF + Bad Page = DataExchange region (allocated contiguously)
+ */
+
+/* v1 layout sizes */
+#define AMD_SRIOV_MSG_VBIOS_SIZE_KB_V1			64
+#define AMD_SRIOV_MSG_PF2VF_SIZE_KB_V1			1
+#define AMD_SRIOV_MSG_VF2PF_SIZE_KB_V1			1
+#define AMD_SRIOV_MSG_BAD_PAGE_SIZE_KB_V1		2
+#define AMD_SRIOV_MSG_RAS_TELEMETRY_SIZE_KB_V1		64
+#define AMD_SRIOV_MSG_DATAEXCHANGE_SIZE_KB_V1		\
+	(AMD_SRIOV_MSG_PF2VF_SIZE_KB_V1 + AMD_SRIOV_MSG_VF2PF_SIZE_KB_V1 + \
+	 AMD_SRIOV_MSG_BAD_PAGE_SIZE_KB_V1)
+
+/* v1 offsets */
+#define AMD_SRIOV_MSG_VBIOS_OFFSET_V1			0
+#define AMD_SRIOV_MSG_DATAEXCHANGE_OFFSET_KB_V1		AMD_SRIOV_MSG_VBIOS_SIZE_KB_V1
+#define AMD_SRIOV_MSG_TMR_OFFSET_KB			2048
+#define AMD_SRIOV_MSG_PF2VF_OFFSET_KB_V1		AMD_SRIOV_MSG_DATAEXCHANGE_OFFSET_KB_V1
+#define AMD_SRIOV_MSG_VF2PF_OFFSET_KB_V1		\
+	(AMD_SRIOV_MSG_PF2VF_OFFSET_KB_V1 + AMD_SRIOV_MSG_SIZE_KB)
+#define AMD_SRIOV_MSG_BAD_PAGE_OFFSET_KB_V1		\
+	(AMD_SRIOV_MSG_VF2PF_OFFSET_KB_V1 + AMD_SRIOV_MSG_SIZE_KB)
+#define AMD_SRIOV_MSG_RAS_TELEMETRY_OFFSET_KB_V1	\
+	(AMD_SRIOV_MSG_BAD_PAGE_OFFSET_KB_V1 + AMD_SRIOV_MSG_BAD_PAGE_SIZE_KB_V1)
+#define AMD_SRIOV_MSG_INIT_DATA_TOT_SIZE_KB_V1		\
+	(AMD_SRIOV_MSG_VBIOS_SIZE_KB_V1 + AMD_SRIOV_MSG_DATAEXCHANGE_SIZE_KB_V1 + \
+	 AMD_SRIOV_MSG_RAS_TELEMETRY_SIZE_KB_V1)
+
+enum amd_sriov_crit_region_version {
+	GPU_CRIT_REGION_V1 = 1,
+	GPU_CRIT_REGION_V2 = 2,
+};
+
+/* v2 layout offset enum (in order of allocation) */
+enum amd_sriov_msg_table_id_enum {
+	AMD_SRIOV_MSG_IPD_TABLE_ID = 0,
+	AMD_SRIOV_MSG_VBIOS_IMG_TABLE_ID,
+	AMD_SRIOV_MSG_RAS_TELEMETRY_TABLE_ID,
+	AMD_SRIOV_MSG_DATAEXCHANGE_TABLE_ID,
+	AMD_SRIOV_MSG_BAD_PAGE_INFO_TABLE_ID,
+	AMD_SRIOV_MSG_INITD_H_TABLE_ID,
+	AMD_SRIOV_MSG_MAX_TABLE_ID,
+};
+
+struct amd_sriov_msg_init_data_header {
+	char     signature[4];  /* "INDA"  */
+	uint32_t version;
+	uint32_t checksum;
+	uint32_t initdata_offset; /* 0 */
+	uint32_t initdata_size_in_kb; /* 5MB */
+	uint32_t valid_tables;
+	uint32_t vbios_img_offset;
+	uint32_t vbios_img_size_in_kb;
+	uint32_t dataexchange_offset;
+	uint32_t dataexchange_size_in_kb;
+	uint32_t ras_tele_info_offset;
+	uint32_t ras_tele_info_size_in_kb;
+	uint32_t ip_discovery_offset;
+	uint32_t ip_discovery_size_in_kb;
+	uint32_t bad_page_info_offset;
+	uint32_t bad_page_size_in_kb;
+	uint32_t reserved[8];
+};
 
 /*
  * PF2VF history log:
@@ -51,10 +112,10 @@
  * v2 defined in amdgim
  * v3 current
  */
-#define AMD_SRIOV_MSG_FW_VRAM_PF2VF_VER			2
-#define AMD_SRIOV_MSG_FW_VRAM_VF2PF_VER			3
+#define AMD_SRIOV_MSG_FW_VRAM_PF2VF_VER 2
+#define AMD_SRIOV_MSG_FW_VRAM_VF2PF_VER 3
 
-#define AMD_SRIOV_MSG_RESERVE_UCODE		24
+#define AMD_SRIOV_MSG_RESERVE_UCODE 24
 
 #define AMD_SRIOV_MSG_RESERVE_VCN_INST 4
 
@@ -83,37 +144,74 @@ enum amd_sriov_ucode_engine_id {
 	AMD_SRIOV_UCODE_ID__MAX
 };
 
-#pragma pack(push, 1) 	// PF2VF / VF2PF data areas are byte packed
+#pragma pack(push, 1) // PF2VF / VF2PF data areas are byte packed
 
 union amd_sriov_msg_feature_flags {
 	struct {
-		uint32_t  error_log_collect  : 1;
-		uint32_t  host_load_ucodes   : 1;
-		uint32_t  host_flr_vramlost  : 1;
-		uint32_t  mm_bw_management   : 1;
-		uint32_t  pp_one_vf_mode     : 1;
-		uint32_t  reg_indirect_acc   : 1;
-		uint32_t  reserved           : 26;
+		uint32_t error_log_collect	: 1;
+		uint32_t host_load_ucodes	: 1;
+		uint32_t host_flr_vramlost	: 1;
+		uint32_t mm_bw_management	: 1;
+		uint32_t pp_one_vf_mode		: 1;
+		uint32_t reg_indirect_acc	: 1;
+		uint32_t av1_support		: 1;
+		uint32_t vcn_rb_decouple	: 1;
+		uint32_t mes_info_dump_enable	: 1;
+		uint32_t ras_caps		: 1;
+		uint32_t ras_telemetry		: 1;
+		uint32_t ras_cper		: 1;
+		uint32_t xgmi_ta_ext_peer_link	: 1;
+		uint32_t reserved		: 19;
 	} flags;
-	uint32_t      all;
+	uint32_t all;
 };
 
 union amd_sriov_reg_access_flags {
 	struct {
-		uint32_t vf_reg_access_ih 	 : 1;
-		uint32_t vf_reg_access_mmhub : 1;
-		uint32_t vf_reg_access_gc 	 : 1;
-		uint32_t reserved	         : 29;
+		uint32_t vf_reg_access_ih		: 1;
+		uint32_t vf_reg_access_mmhub		: 1;
+		uint32_t vf_reg_access_gc		: 1;
+		uint32_t vf_reg_access_l1_tlb_cntl	: 1;
+		uint32_t vf_reg_access_sq_config	: 1;
+		uint32_t reserved			: 27;
 	} flags;
 	uint32_t all;
 };
 
+union amd_sriov_ras_caps {
+	struct {
+		uint64_t block_umc			: 1;
+		uint64_t block_sdma			: 1;
+		uint64_t block_gfx			: 1;
+		uint64_t block_mmhub			: 1;
+		uint64_t block_athub			: 1;
+		uint64_t block_pcie_bif			: 1;
+		uint64_t block_hdp			: 1;
+		uint64_t block_xgmi_wafl		: 1;
+		uint64_t block_df			: 1;
+		uint64_t block_smn			: 1;
+		uint64_t block_sem			: 1;
+		uint64_t block_mp0			: 1;
+		uint64_t block_mp1			: 1;
+		uint64_t block_fuse			: 1;
+		uint64_t block_mca			: 1;
+		uint64_t block_vcn			: 1;
+		uint64_t block_jpeg			: 1;
+		uint64_t block_ih			: 1;
+		uint64_t block_mpio			: 1;
+		uint64_t block_mmsch			: 1;
+		uint64_t poison_propogation_mode	: 1;
+		uint64_t reserved			: 43;
+	} bits;
+	uint64_t all;
+};
+
 union amd_sriov_msg_os_info {
 	struct {
-		uint32_t  windows            : 1;
-		uint32_t  reserved           : 31;
+		uint32_t windows  : 1;
+		uint32_t reserved : 31;
 	} info;
-	uint32_t      all;
+	uint32_t all;
 };
 
 struct amd_sriov_msg_uuid_info {
@@ -156,6 +254,7 @@ struct amd_sriov_msg_pf2vf_info_header {
 	uint32_t reserved[2];
 };
 
+#define AMD_SRIOV_MSG_PF2VF_INFO_FILLED_SIZE (55)
 struct amd_sriov_msg_pf2vf_info {
 	/* header contains size and version */
 	struct amd_sriov_msg_pf2vf_info_header header;
@@ -204,11 +303,19 @@ struct amd_sriov_msg_pf2vf_info {
 	} mm_bw_management[AMD_SRIOV_MSG_RESERVE_VCN_INST];
 	/* UUID info */
 	struct amd_sriov_msg_uuid_info uuid_info;
-	/* pcie atomic Ops info */
-	uint32_t pcie_atomic_ops_enabled_flags;
+	/* PCIE atomic ops support flag */
+	uint32_t pcie_atomic_ops_support_flags;
+	/* Portion of GPU memory occupied by VF.  MAX value is 65535, but set to uint32_t to maintain alignment with reserved size */
+	uint32_t gpu_capacity;
+	/* vf bdf on host pci tree for debug only */
+	uint32_t bdf_on_host;
+	uint32_t more_bp;	//Reserved for future use.
+	union amd_sriov_ras_caps ras_en_caps;
+	union amd_sriov_ras_caps ras_telemetry_en_caps;
+
 	/* reserved */
-	uint32_t reserved[256 - 48];
-};
+	uint32_t reserved[256 - AMD_SRIOV_MSG_PF2VF_INFO_FILLED_SIZE];
+} __packed;
 
 struct amd_sriov_msg_vf2pf_info_header {
 	/* the total structure size in byte */
@@ -219,12 +326,13 @@ struct amd_sriov_msg_vf2pf_info_header {
 	uint32_t reserved[2];
 };
 
+#define AMD_SRIOV_MSG_VF2PF_INFO_FILLED_SIZE (73)
 struct amd_sriov_msg_vf2pf_info {
 	/* header contains size and version */
 	struct amd_sriov_msg_vf2pf_info_header header;
 	uint32_t checksum;
 	/* driver version */
-	uint8_t  driver_version[64];
+	uint8_t driver_version[64];
 	/* driver certification, 1=WHQL, 0=None */
 	uint32_t driver_cert;
 	/* guest OS type and version */
@@ -258,14 +366,16 @@ struct amd_sriov_msg_vf2pf_info {
 	uint32_t fb_size;
 	/* guest ucode data, each one is 1.25 Dword */
 	struct {
-		uint8_t  id;
+		uint8_t id;
 		uint32_t version;
 	} ucode_info[AMD_SRIOV_MSG_RESERVE_UCODE];
 	uint64_t dummy_page_addr;
-
+	/* FB allocated for guest MES to record UQ info */
+	uint64_t mes_info_addr;
+	uint32_t mes_info_size;
 	/* reserved */
-	uint32_t reserved[256-70];
-};
+	uint32_t reserved[256 - AMD_SRIOV_MSG_VF2PF_INFO_FILLED_SIZE];
+} __packed;
 
 /* mailbox message send from guest to host  */
 enum amd_sriov_mailbox_request_message {
@@ -275,22 +385,98 @@ enum amd_sriov_mailbox_request_message {
 	MB_REQ_MSG_REL_GPU_FINI_ACCESS,
 	MB_REQ_MSG_REQ_GPU_RESET_ACCESS,
 	MB_REQ_MSG_REQ_GPU_INIT_DATA,
-
-	MB_REQ_MSG_LOG_VF_ERROR       = 200,
+	MB_REQ_MSG_PSP_VF_CMD_RELAY,
+
+	MB_REQ_MSG_LOG_VF_ERROR = 200,
+	MB_REQ_MSG_READY_TO_RESET = 201,
+	MB_REQ_MSG_RAS_POISON = 202,
+	MB_REQ_RAS_ERROR_COUNT = 203,
+	MB_REQ_RAS_CPER_DUMP = 204,
+	MB_REQ_RAS_BAD_PAGES = 205,
 };
 
 /* mailbox message send from host to guest  */
 enum amd_sriov_mailbox_response_message {
-	MB_RES_MSG_CLR_MSG_BUF = 0,
-	MB_RES_MSG_READY_TO_ACCESS_GPU = 1,
-	MB_RES_MSG_FLR_NOTIFICATION,
-	MB_RES_MSG_FLR_NOTIFICATION_COMPLETION,
-	MB_RES_MSG_SUCCESS,
-	MB_RES_MSG_FAIL,
-	MB_RES_MSG_QUERY_ALIVE,
-	MB_RES_MSG_GPU_INIT_DATA_READY,
-
-	MB_RES_MSG_TEXT_MESSAGE = 255
+	MB_RES_MSG_CLR_MSG_BUF			= 0,
+	MB_RES_MSG_READY_TO_ACCESS_GPU		= 1,
+	MB_RES_MSG_FLR_NOTIFICATION		= 2,
+	MB_RES_MSG_FLR_NOTIFICATION_COMPLETION  = 3,
+	MB_RES_MSG_SUCCESS			= 4,
+	MB_RES_MSG_FAIL				= 5,
+	MB_RES_MSG_QUERY_ALIVE			= 6,
+	MB_RES_MSG_GPU_INIT_DATA_READY		= 7,
+	MB_RES_MSG_RAS_POISON_READY		= 8,
+	MB_RES_MSG_PF_SOFT_FLR_NOTIFICATION	= 9,
+	MB_RES_MSG_GPU_RMA			= 10,
+	MB_RES_MSG_RAS_ERROR_COUNT_READY	= 11,
+	MB_REQ_RAS_CPER_DUMP_READY		= 14,
+	MB_RES_MSG_RAS_BAD_PAGES_READY		= 15,
+	MB_RES_MSG_RAS_BAD_PAGES_NOTIFICATION	= 16,
+	MB_RES_MSG_UNRECOV_ERR_NOTIFICATION	= 17,
+	MB_RES_MSG_TEXT_MESSAGE			= 255
+};
+
+enum amd_sriov_ras_telemetry_gpu_block {
+	RAS_TELEMETRY_GPU_BLOCK_UMC		= 0,
+	RAS_TELEMETRY_GPU_BLOCK_SDMA		= 1,
+	RAS_TELEMETRY_GPU_BLOCK_GFX		= 2,
+	RAS_TELEMETRY_GPU_BLOCK_MMHUB		= 3,
+	RAS_TELEMETRY_GPU_BLOCK_ATHUB		= 4,
+	RAS_TELEMETRY_GPU_BLOCK_PCIE_BIF	= 5,
+	RAS_TELEMETRY_GPU_BLOCK_HDP		= 6,
+	RAS_TELEMETRY_GPU_BLOCK_XGMI_WAFL	= 7,
+	RAS_TELEMETRY_GPU_BLOCK_DF		= 8,
+	RAS_TELEMETRY_GPU_BLOCK_SMN		= 9,
+	RAS_TELEMETRY_GPU_BLOCK_SEM		= 10,
+	RAS_TELEMETRY_GPU_BLOCK_MP0		= 11,
+	RAS_TELEMETRY_GPU_BLOCK_MP1		= 12,
+	RAS_TELEMETRY_GPU_BLOCK_FUSE		= 13,
+	RAS_TELEMETRY_GPU_BLOCK_MCA		= 14,
+	RAS_TELEMETRY_GPU_BLOCK_VCN		= 15,
+	RAS_TELEMETRY_GPU_BLOCK_JPEG		= 16,
+	RAS_TELEMETRY_GPU_BLOCK_IH		= 17,
+	RAS_TELEMETRY_GPU_BLOCK_MPIO		= 18,
+	RAS_TELEMETRY_GPU_BLOCK_COUNT		= 19,
+};
+
+struct amd_sriov_ras_telemetry_header {
+	uint32_t checksum;
+	uint32_t used_size;
+	uint32_t reserved[2];
+};
+
+struct amd_sriov_ras_telemetry_error_count {
+	struct {
+		uint32_t ce_count;
+		uint32_t ue_count;
+		uint32_t de_count;
+		uint32_t ce_overflow_count;
+		uint32_t ue_overflow_count;
+		uint32_t de_overflow_count;
+		uint32_t reserved[6];
+	} block[RAS_TELEMETRY_GPU_BLOCK_COUNT];
+};
+
+struct amd_sriov_ras_cper_dump {
+	uint32_t more;
+	uint64_t overflow_count;
+	uint64_t count;
+	uint64_t wptr;
+	uint32_t buf[];
+};
+
+struct amd_sriov_ras_chk_criti {
+	uint32_t hit;
+};
+
+struct amdsriov_ras_telemetry {
+	struct amd_sriov_ras_telemetry_header header;
+
+	union {
+		struct amd_sriov_ras_telemetry_error_count error_count;
+		struct amd_sriov_ras_cper_dump cper_dump;
+		struct amd_sriov_ras_chk_criti chk_criti;
+	} body;
 };
 
 /* version data stored in MAILBOX_MSGBUF_RCV_DW1 for future expansion */
@@ -298,17 +484,15 @@ enum amd_sriov_gpu_init_data_version {
 	GPU_INIT_DATA_READY_V1 = 1,
 };
 
-#pragma pack(pop)	// Restore previous packing option
+#pragma pack(pop) // Restore previous packing option
 
 /* checksum function between host and guest */
-unsigned int amd_sriov_msg_checksum(void *obj,
-				unsigned long obj_size,
-				unsigned int key,
-				unsigned int checksum);
+unsigned int amd_sriov_msg_checksum(void *obj, unsigned long obj_size, unsigned int key,
+				    unsigned int checksum);
 
 /* assertion at compile time */
 #ifdef __linux__
-#define stringification(s) _stringification(s)
+#define stringification(s)  _stringification(s)
 #define _stringification(s) #s
 
 _Static_assert(
@@ -319,13 +503,11 @@ _Static_assert(
 	sizeof(struct amd_sriov_msg_pf2vf_info) == AMD_SRIOV_MSG_SIZE_KB << 10,
 	"amd_sriov_msg_pf2vf_info must be " stringification(AMD_SRIOV_MSG_SIZE_KB) " KB");
 
-_Static_assert(
-	AMD_SRIOV_MSG_RESERVE_UCODE % 4 == 0,
-	"AMD_SRIOV_MSG_RESERVE_UCODE must be multiple of 4");
+_Static_assert(AMD_SRIOV_MSG_RESERVE_UCODE % 4 == 0,
+	       "AMD_SRIOV_MSG_RESERVE_UCODE must be multiple of 4");
 
-_Static_assert(
-	AMD_SRIOV_MSG_RESERVE_UCODE > AMD_SRIOV_UCODE_ID__MAX,
-	"AMD_SRIOV_MSG_RESERVE_UCODE must be bigger than AMD_SRIOV_UCODE_ID__MAX");
+_Static_assert(AMD_SRIOV_MSG_RESERVE_UCODE > AMD_SRIOV_UCODE_ID__MAX,
+	       "AMD_SRIOV_MSG_RESERVE_UCODE must be bigger than AMD_SRIOV_UCODE_ID__MAX");
 
 #undef _stringification
 #undef stringification
diff --git a/drivers/gpu/drm/amd/amdgpu/aqua_vanjaram.c b/drivers/gpu/drm/amd/amdgpu/aqua_vanjaram.c
new file mode 100644
index 000000000000..f9e2edf5260b
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/aqua_vanjaram.c
@@ -0,0 +1,986 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include "amdgpu.h"
+#include "soc15.h"
+
+#include "soc15_common.h"
+#include "amdgpu_reg_state.h"
+#include "amdgpu_xcp.h"
+#include "gfx_v9_4_3.h"
+#include "gfxhub_v1_2.h"
+#include "sdma_v4_4_2.h"
+#include "amdgpu_ip.h"
+
+#define XCP_INST_MASK(num_inst, xcp_id)                                        \
+	(num_inst ? GENMASK(num_inst - 1, 0) << (xcp_id * num_inst) : 0)
+
+void aqua_vanjaram_doorbell_index_init(struct amdgpu_device *adev)
+{
+	int i;
+
+	adev->doorbell_index.kiq = AMDGPU_DOORBELL_LAYOUT1_KIQ_START;
+
+	adev->doorbell_index.mec_ring0 = AMDGPU_DOORBELL_LAYOUT1_MEC_RING_START;
+
+	adev->doorbell_index.userqueue_start = AMDGPU_DOORBELL_LAYOUT1_USERQUEUE_START;
+	adev->doorbell_index.userqueue_end = AMDGPU_DOORBELL_LAYOUT1_USERQUEUE_END;
+	adev->doorbell_index.xcc_doorbell_range = AMDGPU_DOORBELL_LAYOUT1_XCC_RANGE;
+
+	adev->doorbell_index.sdma_doorbell_range = 20;
+	for (i = 0; i < adev->sdma.num_instances; i++)
+		adev->doorbell_index.sdma_engine[i] =
+			AMDGPU_DOORBELL_LAYOUT1_sDMA_ENGINE_START +
+			i * (adev->doorbell_index.sdma_doorbell_range >> 1);
+
+	adev->doorbell_index.ih = AMDGPU_DOORBELL_LAYOUT1_IH;
+	adev->doorbell_index.vcn.vcn_ring0_1 = AMDGPU_DOORBELL_LAYOUT1_VCN_START;
+
+	adev->doorbell_index.first_non_cp = AMDGPU_DOORBELL_LAYOUT1_FIRST_NON_CP;
+	adev->doorbell_index.last_non_cp = AMDGPU_DOORBELL_LAYOUT1_LAST_NON_CP;
+
+	adev->doorbell_index.max_assignment = AMDGPU_DOORBELL_LAYOUT1_MAX_ASSIGNMENT << 1;
+}
+
+/* Fixed pattern for smn addressing on different AIDs:
+ *   bit[34]: indicate cross AID access
+ *   bit[33:32]: indicate target AID id
+ * AID id range is 0 ~ 3 as maximum AID number is 4.
+ */
+u64 aqua_vanjaram_encode_ext_smn_addressing(int ext_id)
+{
+	u64 ext_offset;
+
+	/* local routing and bit[34:32] will be zeros */
+	if (ext_id == 0)
+		return 0;
+
+	/* Initiated from host, accessing to all non-zero aids are cross traffic */
+	ext_offset = ((u64)(ext_id & 0x3) << 32) | (1ULL << 34);
+
+	return ext_offset;
+}
+
+static enum amdgpu_gfx_partition
+__aqua_vanjaram_calc_xcp_mode(struct amdgpu_xcp_mgr *xcp_mgr)
+{
+	struct amdgpu_device *adev = xcp_mgr->adev;
+	int num_xcc, num_xcc_per_xcp = 0, mode = 0;
+
+	num_xcc = NUM_XCC(xcp_mgr->adev->gfx.xcc_mask);
+	if (adev->gfx.funcs->get_xccs_per_xcp)
+		num_xcc_per_xcp = adev->gfx.funcs->get_xccs_per_xcp(adev);
+	if ((num_xcc_per_xcp) && (num_xcc % num_xcc_per_xcp == 0))
+		mode = num_xcc / num_xcc_per_xcp;
+
+	if (num_xcc_per_xcp == 1)
+		return AMDGPU_CPX_PARTITION_MODE;
+
+	switch (mode) {
+	case 1:
+		return AMDGPU_SPX_PARTITION_MODE;
+	case 2:
+		return AMDGPU_DPX_PARTITION_MODE;
+	case 3:
+		return AMDGPU_TPX_PARTITION_MODE;
+	case 4:
+		return AMDGPU_QPX_PARTITION_MODE;
+	default:
+		return AMDGPU_UNKNOWN_COMPUTE_PARTITION_MODE;
+	}
+
+	return AMDGPU_UNKNOWN_COMPUTE_PARTITION_MODE;
+}
+
+static int aqua_vanjaram_query_partition_mode(struct amdgpu_xcp_mgr *xcp_mgr)
+{
+	enum amdgpu_gfx_partition derv_mode,
+		mode = AMDGPU_UNKNOWN_COMPUTE_PARTITION_MODE;
+	struct amdgpu_device *adev = xcp_mgr->adev;
+
+	derv_mode = __aqua_vanjaram_calc_xcp_mode(xcp_mgr);
+
+	if (amdgpu_sriov_vf(adev))
+		return derv_mode;
+
+	if (adev->nbio.funcs->get_compute_partition_mode) {
+		mode = adev->nbio.funcs->get_compute_partition_mode(adev);
+		if (mode != derv_mode) {
+			dev_warn(
+				adev->dev,
+				"Mismatch in compute partition mode - reported : %d derived : %d",
+				mode, derv_mode);
+			if (derv_mode == AMDGPU_UNKNOWN_COMPUTE_PARTITION_MODE)
+				amdgpu_device_bus_status_check(adev);
+		}
+	}
+
+	return mode;
+}
+
+static int __aqua_vanjaram_get_xcc_per_xcp(struct amdgpu_xcp_mgr *xcp_mgr, int mode)
+{
+	int num_xcc, num_xcc_per_xcp = 0;
+
+	num_xcc = NUM_XCC(xcp_mgr->adev->gfx.xcc_mask);
+
+	switch (mode) {
+	case AMDGPU_SPX_PARTITION_MODE:
+		num_xcc_per_xcp = num_xcc;
+		break;
+	case AMDGPU_DPX_PARTITION_MODE:
+		num_xcc_per_xcp = num_xcc / 2;
+		break;
+	case AMDGPU_TPX_PARTITION_MODE:
+		num_xcc_per_xcp = num_xcc / 3;
+		break;
+	case AMDGPU_QPX_PARTITION_MODE:
+		num_xcc_per_xcp = num_xcc / 4;
+		break;
+	case AMDGPU_CPX_PARTITION_MODE:
+		num_xcc_per_xcp = 1;
+		break;
+	}
+
+	return num_xcc_per_xcp;
+}
+
+static int __aqua_vanjaram_get_xcp_ip_info(struct amdgpu_xcp_mgr *xcp_mgr, int xcp_id,
+				    enum AMDGPU_XCP_IP_BLOCK ip_id,
+				    struct amdgpu_xcp_ip *ip)
+{
+	struct amdgpu_device *adev = xcp_mgr->adev;
+	int num_sdma, num_vcn, num_shared_vcn, num_xcp;
+	int num_xcc_xcp, num_sdma_xcp, num_vcn_xcp;
+
+	num_sdma = adev->sdma.num_instances;
+	num_vcn = adev->vcn.num_vcn_inst;
+	num_shared_vcn = 1;
+
+	num_xcc_xcp = adev->gfx.num_xcc_per_xcp;
+	num_xcp = NUM_XCC(adev->gfx.xcc_mask) / num_xcc_xcp;
+
+	switch (xcp_mgr->mode) {
+	case AMDGPU_SPX_PARTITION_MODE:
+	case AMDGPU_DPX_PARTITION_MODE:
+	case AMDGPU_TPX_PARTITION_MODE:
+	case AMDGPU_QPX_PARTITION_MODE:
+	case AMDGPU_CPX_PARTITION_MODE:
+		num_sdma_xcp = DIV_ROUND_UP(num_sdma, num_xcp);
+		num_vcn_xcp = DIV_ROUND_UP(num_vcn, num_xcp);
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	if (num_vcn && num_xcp > num_vcn)
+		num_shared_vcn = num_xcp / num_vcn;
+
+	switch (ip_id) {
+	case AMDGPU_XCP_GFXHUB:
+		ip->inst_mask = XCP_INST_MASK(num_xcc_xcp, xcp_id);
+		ip->ip_funcs = &gfxhub_v1_2_xcp_funcs;
+		break;
+	case AMDGPU_XCP_GFX:
+		ip->inst_mask = XCP_INST_MASK(num_xcc_xcp, xcp_id);
+		ip->ip_funcs = &gfx_v9_4_3_xcp_funcs;
+		break;
+	case AMDGPU_XCP_SDMA:
+		ip->inst_mask = XCP_INST_MASK(num_sdma_xcp, xcp_id);
+		ip->ip_funcs = &sdma_v4_4_2_xcp_funcs;
+		break;
+	case AMDGPU_XCP_VCN:
+		ip->inst_mask =
+			XCP_INST_MASK(num_vcn_xcp, xcp_id / num_shared_vcn);
+		/* TODO : Assign IP funcs */
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	ip->ip_id = ip_id;
+
+	return 0;
+}
+
+static int __aqua_vanjaram_get_px_mode_info(struct amdgpu_xcp_mgr *xcp_mgr,
+					    int px_mode, int *num_xcp,
+					    uint16_t *nps_modes)
+{
+	struct amdgpu_device *adev = xcp_mgr->adev;
+	uint32_t gc_ver = amdgpu_ip_version(adev, GC_HWIP, 0);
+
+	if (!num_xcp || !nps_modes || !(xcp_mgr->supp_xcp_modes & BIT(px_mode)))
+		return -EINVAL;
+
+	switch (px_mode) {
+	case AMDGPU_SPX_PARTITION_MODE:
+		*num_xcp = 1;
+		*nps_modes = BIT(AMDGPU_NPS1_PARTITION_MODE);
+		break;
+	case AMDGPU_DPX_PARTITION_MODE:
+		*num_xcp = 2;
+		*nps_modes = BIT(AMDGPU_NPS1_PARTITION_MODE) |
+			     BIT(AMDGPU_NPS2_PARTITION_MODE);
+		break;
+	case AMDGPU_TPX_PARTITION_MODE:
+		*num_xcp = 3;
+		*nps_modes = BIT(AMDGPU_NPS1_PARTITION_MODE) |
+			     BIT(AMDGPU_NPS4_PARTITION_MODE);
+		break;
+	case AMDGPU_QPX_PARTITION_MODE:
+		*num_xcp = 4;
+		*nps_modes = BIT(AMDGPU_NPS1_PARTITION_MODE) |
+			     BIT(AMDGPU_NPS4_PARTITION_MODE);
+		if (gc_ver == IP_VERSION(9, 5, 0))
+			*nps_modes |= BIT(AMDGPU_NPS2_PARTITION_MODE);
+		break;
+	case AMDGPU_CPX_PARTITION_MODE:
+		*num_xcp = NUM_XCC(adev->gfx.xcc_mask);
+		*nps_modes = BIT(AMDGPU_NPS1_PARTITION_MODE) |
+			     BIT(AMDGPU_NPS4_PARTITION_MODE);
+		if (gc_ver == IP_VERSION(9, 5, 0))
+			*nps_modes |= BIT(AMDGPU_NPS2_PARTITION_MODE);
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int aqua_vanjaram_get_xcp_res_info(struct amdgpu_xcp_mgr *xcp_mgr,
+					  int mode,
+					  struct amdgpu_xcp_cfg *xcp_cfg)
+{
+	struct amdgpu_device *adev = xcp_mgr->adev;
+	int max_res[AMDGPU_XCP_RES_MAX] = {};
+	bool res_lt_xcp;
+	int num_xcp, i, r;
+	u16 nps_modes;
+
+	if (!(xcp_mgr->supp_xcp_modes & BIT(mode)))
+		return -EINVAL;
+
+	max_res[AMDGPU_XCP_RES_XCC] = NUM_XCC(adev->gfx.xcc_mask);
+	max_res[AMDGPU_XCP_RES_DMA] = adev->sdma.num_instances;
+	max_res[AMDGPU_XCP_RES_DEC] = adev->vcn.num_vcn_inst;
+	max_res[AMDGPU_XCP_RES_JPEG] = adev->jpeg.num_jpeg_inst;
+
+	r = __aqua_vanjaram_get_px_mode_info(xcp_mgr, mode, &num_xcp, &nps_modes);
+	if (r)
+		return r;
+
+	xcp_cfg->compatible_nps_modes =
+		(adev->gmc.supported_nps_modes & nps_modes);
+	xcp_cfg->num_res = ARRAY_SIZE(max_res);
+
+	for (i = 0; i < xcp_cfg->num_res; i++) {
+		res_lt_xcp = max_res[i] < num_xcp;
+		xcp_cfg->xcp_res[i].id = i;
+		xcp_cfg->xcp_res[i].num_inst =
+			res_lt_xcp ? 1 : max_res[i] / num_xcp;
+		xcp_cfg->xcp_res[i].num_inst =
+			i == AMDGPU_XCP_RES_JPEG ?
+			xcp_cfg->xcp_res[i].num_inst *
+			adev->jpeg.num_jpeg_rings : xcp_cfg->xcp_res[i].num_inst;
+		xcp_cfg->xcp_res[i].num_shared =
+			res_lt_xcp ? num_xcp / max_res[i] : 1;
+	}
+
+	return 0;
+}
+
+static enum amdgpu_gfx_partition
+__aqua_vanjaram_get_auto_mode(struct amdgpu_xcp_mgr *xcp_mgr)
+{
+	struct amdgpu_device *adev = xcp_mgr->adev;
+	int num_xcc;
+
+	num_xcc = NUM_XCC(xcp_mgr->adev->gfx.xcc_mask);
+
+	if (adev->gmc.num_mem_partitions == 1)
+		return AMDGPU_SPX_PARTITION_MODE;
+
+	if (adev->gmc.num_mem_partitions == num_xcc)
+		return AMDGPU_CPX_PARTITION_MODE;
+
+	if (adev->gmc.num_mem_partitions == num_xcc / 2)
+		return (adev->flags & AMD_IS_APU) ? AMDGPU_TPX_PARTITION_MODE :
+						    AMDGPU_CPX_PARTITION_MODE;
+
+	if (adev->gmc.num_mem_partitions == 2 && !(adev->flags & AMD_IS_APU))
+		return AMDGPU_DPX_PARTITION_MODE;
+
+	return AMDGPU_UNKNOWN_COMPUTE_PARTITION_MODE;
+}
+
+static bool __aqua_vanjaram_is_valid_mode(struct amdgpu_xcp_mgr *xcp_mgr,
+					  enum amdgpu_gfx_partition mode)
+{
+	struct amdgpu_device *adev = xcp_mgr->adev;
+	int num_xcc, num_xccs_per_xcp, r;
+	int num_xcp, nps_mode;
+	u16 supp_nps_modes;
+	bool comp_mode;
+
+	nps_mode = adev->gmc.gmc_funcs->query_mem_partition_mode(adev);
+	r = __aqua_vanjaram_get_px_mode_info(xcp_mgr, mode, &num_xcp,
+					       &supp_nps_modes);
+	if (r)
+		return false;
+
+	comp_mode = !!(BIT(nps_mode) & supp_nps_modes);
+	num_xcc = NUM_XCC(adev->gfx.xcc_mask);
+	switch (mode) {
+	case AMDGPU_SPX_PARTITION_MODE:
+		return comp_mode && num_xcc > 0;
+	case AMDGPU_DPX_PARTITION_MODE:
+		return comp_mode && (num_xcc % 4) == 0;
+	case AMDGPU_TPX_PARTITION_MODE:
+		return comp_mode && ((num_xcc % 3) == 0);
+	case AMDGPU_QPX_PARTITION_MODE:
+		num_xccs_per_xcp = num_xcc / 4;
+		return comp_mode && (num_xccs_per_xcp >= 2);
+	case AMDGPU_CPX_PARTITION_MODE:
+		return comp_mode && (num_xcc > 1);
+	default:
+		return false;
+	}
+
+	return false;
+}
+
+static void __aqua_vanjaram_update_available_partition_mode(struct amdgpu_xcp_mgr *xcp_mgr)
+{
+	int mode;
+
+	xcp_mgr->avail_xcp_modes = 0;
+
+	for_each_inst(mode, xcp_mgr->supp_xcp_modes) {
+		if (__aqua_vanjaram_is_valid_mode(xcp_mgr, mode))
+			xcp_mgr->avail_xcp_modes |= BIT(mode);
+	}
+}
+
+static int aqua_vanjaram_switch_partition_mode(struct amdgpu_xcp_mgr *xcp_mgr,
+					       int mode, int *num_xcps)
+{
+	int num_xcc_per_xcp, num_xcc, ret;
+	struct amdgpu_device *adev;
+	u32 flags = 0;
+
+	adev = xcp_mgr->adev;
+	num_xcc = NUM_XCC(adev->gfx.xcc_mask);
+
+	if (mode == AMDGPU_AUTO_COMPUTE_PARTITION_MODE) {
+		mode = __aqua_vanjaram_get_auto_mode(xcp_mgr);
+		if (mode == AMDGPU_UNKNOWN_COMPUTE_PARTITION_MODE) {
+			dev_err(adev->dev,
+				"Invalid config, no compatible compute partition mode found, available memory partitions: %d",
+				adev->gmc.num_mem_partitions);
+			return -EINVAL;
+		}
+	} else if (!__aqua_vanjaram_is_valid_mode(xcp_mgr, mode)) {
+		dev_err(adev->dev,
+			"Invalid compute partition mode requested, requested: %s, available memory partitions: %d",
+			amdgpu_gfx_compute_mode_desc(mode), adev->gmc.num_mem_partitions);
+		return -EINVAL;
+	}
+
+	if (adev->kfd.init_complete && !amdgpu_in_reset(adev) &&
+		!adev->in_suspend)
+		flags |= AMDGPU_XCP_OPS_KFD;
+
+	if (flags & AMDGPU_XCP_OPS_KFD) {
+		ret = amdgpu_amdkfd_check_and_lock_kfd(adev);
+		if (ret)
+			goto out;
+	}
+
+	ret = amdgpu_xcp_pre_partition_switch(xcp_mgr, flags);
+	if (ret)
+		goto unlock;
+
+	num_xcc_per_xcp = __aqua_vanjaram_get_xcc_per_xcp(xcp_mgr, mode);
+	if (adev->gfx.funcs->switch_partition_mode)
+		adev->gfx.funcs->switch_partition_mode(xcp_mgr->adev,
+						       num_xcc_per_xcp);
+
+	/* Init info about new xcps */
+	*num_xcps = num_xcc / num_xcc_per_xcp;
+	amdgpu_xcp_init(xcp_mgr, *num_xcps, mode);
+
+	ret = amdgpu_xcp_post_partition_switch(xcp_mgr, flags);
+	if (!ret)
+		__aqua_vanjaram_update_available_partition_mode(xcp_mgr);
+unlock:
+	if (flags & AMDGPU_XCP_OPS_KFD)
+		amdgpu_amdkfd_unlock_kfd(adev);
+out:
+	return ret;
+}
+
+static int __aqua_vanjaram_get_xcp_mem_id(struct amdgpu_device *adev,
+					  int xcc_id, uint8_t *mem_id)
+{
+	/* memory/spatial modes validation check is already done */
+	*mem_id = xcc_id / adev->gfx.num_xcc_per_xcp;
+	*mem_id /= adev->xcp_mgr->num_xcp_per_mem_partition;
+
+	return 0;
+}
+
+static int aqua_vanjaram_get_xcp_mem_id(struct amdgpu_xcp_mgr *xcp_mgr,
+					struct amdgpu_xcp *xcp, uint8_t *mem_id)
+{
+	struct amdgpu_numa_info numa_info;
+	struct amdgpu_device *adev;
+	uint32_t xcc_mask;
+	int r, i, xcc_id;
+
+	adev = xcp_mgr->adev;
+	/* TODO: BIOS is not returning the right info now
+	 * Check on this later
+	 */
+	/*
+	if (adev->gmc.gmc_funcs->query_mem_partition_mode)
+		mode = adev->gmc.gmc_funcs->query_mem_partition_mode(adev);
+	*/
+	if (adev->gmc.num_mem_partitions == 1) {
+		/* Only one range */
+		*mem_id = 0;
+		return 0;
+	}
+
+	r = amdgpu_xcp_get_inst_details(xcp, AMDGPU_XCP_GFX, &xcc_mask);
+	if (r || !xcc_mask)
+		return -EINVAL;
+
+	xcc_id = ffs(xcc_mask) - 1;
+	if (!adev->gmc.is_app_apu)
+		return __aqua_vanjaram_get_xcp_mem_id(adev, xcc_id, mem_id);
+
+	r = amdgpu_acpi_get_mem_info(adev, xcc_id, &numa_info);
+
+	if (r)
+		return r;
+
+	r = -EINVAL;
+	for (i = 0; i < adev->gmc.num_mem_partitions; ++i) {
+		if (adev->gmc.mem_partitions[i].numa.node == numa_info.nid) {
+			*mem_id = i;
+			r = 0;
+			break;
+		}
+	}
+
+	return r;
+}
+
+static int aqua_vanjaram_get_xcp_ip_details(struct amdgpu_xcp_mgr *xcp_mgr, int xcp_id,
+				     enum AMDGPU_XCP_IP_BLOCK ip_id,
+				     struct amdgpu_xcp_ip *ip)
+{
+	if (!ip)
+		return -EINVAL;
+
+	return __aqua_vanjaram_get_xcp_ip_info(xcp_mgr, xcp_id, ip_id, ip);
+}
+
+struct amdgpu_xcp_mgr_funcs aqua_vanjaram_xcp_funcs = {
+	.switch_partition_mode = &aqua_vanjaram_switch_partition_mode,
+	.query_partition_mode = &aqua_vanjaram_query_partition_mode,
+	.get_ip_details = &aqua_vanjaram_get_xcp_ip_details,
+	.get_xcp_res_info = &aqua_vanjaram_get_xcp_res_info,
+	.get_xcp_mem_id = &aqua_vanjaram_get_xcp_mem_id,
+};
+
+static int aqua_vanjaram_xcp_mgr_init(struct amdgpu_device *adev)
+{
+	int ret;
+
+	if (amdgpu_sriov_vf(adev))
+		aqua_vanjaram_xcp_funcs.switch_partition_mode = NULL;
+
+	ret = amdgpu_xcp_mgr_init(adev, AMDGPU_UNKNOWN_COMPUTE_PARTITION_MODE, 1,
+				  &aqua_vanjaram_xcp_funcs);
+	if (ret)
+		return ret;
+
+	amdgpu_xcp_update_supported_modes(adev->xcp_mgr);
+	/* TODO: Default memory node affinity init */
+
+	return ret;
+}
+
+int aqua_vanjaram_init_soc_config(struct amdgpu_device *adev)
+{
+	u32 mask, avail_inst, inst_mask = adev->sdma.sdma_mask;
+	int ret, i;
+
+	/* generally 1 AID supports 4 instances */
+	adev->sdma.num_inst_per_aid = 4;
+	adev->sdma.num_instances = NUM_SDMA(adev->sdma.sdma_mask);
+
+	adev->aid_mask = i = 1;
+	inst_mask >>= adev->sdma.num_inst_per_aid;
+
+	for (mask = (1 << adev->sdma.num_inst_per_aid) - 1; inst_mask;
+	     inst_mask >>= adev->sdma.num_inst_per_aid, ++i) {
+		avail_inst = inst_mask & mask;
+		if (avail_inst == mask || avail_inst == 0x3 ||
+		    avail_inst == 0xc)
+			adev->aid_mask |= (1 << i);
+	}
+
+	/* Harvest config is not used for aqua vanjaram. VCN and JPEGs will be
+	 * addressed based on logical instance ids.
+	 */
+	adev->vcn.harvest_config = 0;
+	adev->vcn.num_inst_per_aid = 1;
+	adev->vcn.num_vcn_inst = hweight32(adev->vcn.inst_mask);
+	adev->jpeg.harvest_config = 0;
+	adev->jpeg.num_inst_per_aid = 1;
+	adev->jpeg.num_jpeg_inst = hweight32(adev->jpeg.inst_mask);
+
+	ret = aqua_vanjaram_xcp_mgr_init(adev);
+	if (ret)
+		return ret;
+
+	amdgpu_ip_map_init(adev);
+
+	return 0;
+}
+
+static void aqua_read_smn(struct amdgpu_device *adev,
+			  struct amdgpu_smn_reg_data *regdata,
+			  uint64_t smn_addr)
+{
+	regdata->addr = smn_addr;
+	regdata->value = RREG32_PCIE(smn_addr);
+}
+
+struct aqua_reg_list {
+	uint64_t start_addr;
+	uint32_t num_regs;
+	uint32_t incrx;
+};
+
+#define DW_ADDR_INCR	4
+
+static void aqua_read_smn_ext(struct amdgpu_device *adev,
+			      struct amdgpu_smn_reg_data *regdata,
+			      uint64_t smn_addr, int i)
+{
+	regdata->addr =
+		smn_addr + adev->asic_funcs->encode_ext_smn_addressing(i);
+	regdata->value = RREG32_PCIE_EXT(regdata->addr);
+}
+
+#define smnreg_0x1A340218	0x1A340218
+#define smnreg_0x1A3402E4	0x1A3402E4
+#define smnreg_0x1A340294	0x1A340294
+#define smreg_0x1A380088	0x1A380088
+
+#define NUM_PCIE_SMN_REGS	14
+
+static struct aqua_reg_list pcie_reg_addrs[] = {
+	{ smnreg_0x1A340218, 1, 0 },
+	{ smnreg_0x1A3402E4, 1, 0 },
+	{ smnreg_0x1A340294, 6, DW_ADDR_INCR },
+	{ smreg_0x1A380088, 6, DW_ADDR_INCR },
+};
+
+static ssize_t aqua_vanjaram_read_pcie_state(struct amdgpu_device *adev,
+					     void *buf, size_t max_size)
+{
+	struct amdgpu_reg_state_pcie_v1_0 *pcie_reg_state;
+	uint32_t start_addr, incrx, num_regs, szbuf;
+	struct amdgpu_regs_pcie_v1_0 *pcie_regs;
+	struct amdgpu_smn_reg_data *reg_data;
+	struct pci_dev *us_pdev, *ds_pdev;
+	int aer_cap, r, n;
+
+	if (!buf || !max_size)
+		return -EINVAL;
+
+	pcie_reg_state = (struct amdgpu_reg_state_pcie_v1_0 *)buf;
+
+	szbuf = sizeof(*pcie_reg_state) +
+		amdgpu_reginst_size(1, sizeof(*pcie_regs), NUM_PCIE_SMN_REGS);
+	/* Only one instance of pcie regs */
+	if (max_size < szbuf)
+		return -EOVERFLOW;
+
+	pcie_regs = (struct amdgpu_regs_pcie_v1_0 *)((uint8_t *)buf +
+						     sizeof(*pcie_reg_state));
+	pcie_regs->inst_header.instance = 0;
+	pcie_regs->inst_header.state = AMDGPU_INST_S_OK;
+	pcie_regs->inst_header.num_smn_regs = NUM_PCIE_SMN_REGS;
+
+	reg_data = pcie_regs->smn_reg_values;
+
+	for (r = 0; r < ARRAY_SIZE(pcie_reg_addrs); r++) {
+		start_addr = pcie_reg_addrs[r].start_addr;
+		incrx = pcie_reg_addrs[r].incrx;
+		num_regs = pcie_reg_addrs[r].num_regs;
+		for (n = 0; n < num_regs; n++) {
+			aqua_read_smn(adev, reg_data, start_addr + n * incrx);
+			++reg_data;
+		}
+	}
+
+	ds_pdev = pci_upstream_bridge(adev->pdev);
+	us_pdev = pci_upstream_bridge(ds_pdev);
+
+	pcie_capability_read_word(us_pdev, PCI_EXP_DEVSTA,
+				  &pcie_regs->device_status);
+	pcie_capability_read_word(us_pdev, PCI_EXP_LNKSTA,
+				  &pcie_regs->link_status);
+
+	aer_cap = pci_find_ext_capability(us_pdev, PCI_EXT_CAP_ID_ERR);
+	if (aer_cap) {
+		pci_read_config_dword(us_pdev, aer_cap + PCI_ERR_COR_STATUS,
+				      &pcie_regs->pcie_corr_err_status);
+		pci_read_config_dword(us_pdev, aer_cap + PCI_ERR_UNCOR_STATUS,
+				      &pcie_regs->pcie_uncorr_err_status);
+	}
+
+	pci_read_config_dword(us_pdev, PCI_PRIMARY_BUS,
+			      &pcie_regs->sub_bus_number_latency);
+
+	pcie_reg_state->common_header.structure_size = szbuf;
+	pcie_reg_state->common_header.format_revision = 1;
+	pcie_reg_state->common_header.content_revision = 0;
+	pcie_reg_state->common_header.state_type = AMDGPU_REG_STATE_TYPE_PCIE;
+	pcie_reg_state->common_header.num_instances = 1;
+
+	return pcie_reg_state->common_header.structure_size;
+}
+
+#define smnreg_0x11A00050	0x11A00050
+#define smnreg_0x11A00180	0x11A00180
+#define smnreg_0x11A00070	0x11A00070
+#define smnreg_0x11A00200	0x11A00200
+#define smnreg_0x11A0020C	0x11A0020C
+#define smnreg_0x11A00210	0x11A00210
+#define smnreg_0x11A00108	0x11A00108
+
+#define XGMI_LINK_REG(smnreg, l) ((smnreg) | (l << 20))
+
+#define NUM_XGMI_SMN_REGS 25
+
+static struct aqua_reg_list xgmi_reg_addrs[] = {
+	{ smnreg_0x11A00050, 1, 0 },
+	{ smnreg_0x11A00180, 16, DW_ADDR_INCR },
+	{ smnreg_0x11A00070, 4, DW_ADDR_INCR },
+	{ smnreg_0x11A00200, 1, 0 },
+	{ smnreg_0x11A0020C, 1, 0 },
+	{ smnreg_0x11A00210, 1, 0 },
+	{ smnreg_0x11A00108, 1, 0 },
+};
+
+static ssize_t aqua_vanjaram_read_xgmi_state(struct amdgpu_device *adev,
+					     void *buf, size_t max_size)
+{
+	struct amdgpu_reg_state_xgmi_v1_0 *xgmi_reg_state;
+	uint32_t start_addr, incrx, num_regs, szbuf;
+	struct amdgpu_regs_xgmi_v1_0 *xgmi_regs;
+	struct amdgpu_smn_reg_data *reg_data;
+	const int max_xgmi_instances = 8;
+	int inst = 0, i, j, r, n;
+	const int xgmi_inst = 2;
+	void *p;
+
+	if (!buf || !max_size)
+		return -EINVAL;
+
+	xgmi_reg_state = (struct amdgpu_reg_state_xgmi_v1_0 *)buf;
+
+	szbuf = sizeof(*xgmi_reg_state) +
+		amdgpu_reginst_size(max_xgmi_instances, sizeof(*xgmi_regs),
+				    NUM_XGMI_SMN_REGS);
+	/* Only one instance of pcie regs */
+	if (max_size < szbuf)
+		return -EOVERFLOW;
+
+	p = &xgmi_reg_state->xgmi_state_regs[0];
+	for_each_inst(i, adev->aid_mask) {
+		for (j = 0; j < xgmi_inst; ++j) {
+			xgmi_regs = (struct amdgpu_regs_xgmi_v1_0 *)p;
+			xgmi_regs->inst_header.instance = inst++;
+
+			xgmi_regs->inst_header.state = AMDGPU_INST_S_OK;
+			xgmi_regs->inst_header.num_smn_regs = NUM_XGMI_SMN_REGS;
+
+			reg_data = xgmi_regs->smn_reg_values;
+
+			for (r = 0; r < ARRAY_SIZE(xgmi_reg_addrs); r++) {
+				start_addr = xgmi_reg_addrs[r].start_addr;
+				incrx = xgmi_reg_addrs[r].incrx;
+				num_regs = xgmi_reg_addrs[r].num_regs;
+
+				for (n = 0; n < num_regs; n++) {
+					aqua_read_smn_ext(
+						adev, reg_data,
+						XGMI_LINK_REG(start_addr, j) +
+							n * incrx,
+						i);
+					++reg_data;
+				}
+			}
+			p = reg_data;
+		}
+	}
+
+	xgmi_reg_state->common_header.structure_size = szbuf;
+	xgmi_reg_state->common_header.format_revision = 1;
+	xgmi_reg_state->common_header.content_revision = 0;
+	xgmi_reg_state->common_header.state_type = AMDGPU_REG_STATE_TYPE_XGMI;
+	xgmi_reg_state->common_header.num_instances = max_xgmi_instances;
+
+	return xgmi_reg_state->common_header.structure_size;
+}
+
+#define smnreg_0x11C00070	0x11C00070
+#define smnreg_0x11C00210	0x11C00210
+
+static struct aqua_reg_list wafl_reg_addrs[] = {
+	{ smnreg_0x11C00070, 4, DW_ADDR_INCR },
+	{ smnreg_0x11C00210, 1, 0 },
+};
+
+#define WAFL_LINK_REG(smnreg, l) ((smnreg) | (l << 20))
+
+#define NUM_WAFL_SMN_REGS 5
+
+static ssize_t aqua_vanjaram_read_wafl_state(struct amdgpu_device *adev,
+					     void *buf, size_t max_size)
+{
+	struct amdgpu_reg_state_wafl_v1_0 *wafl_reg_state;
+	uint32_t start_addr, incrx, num_regs, szbuf;
+	struct amdgpu_regs_wafl_v1_0 *wafl_regs;
+	struct amdgpu_smn_reg_data *reg_data;
+	const int max_wafl_instances = 8;
+	int inst = 0, i, j, r, n;
+	const int wafl_inst = 2;
+	void *p;
+
+	if (!buf || !max_size)
+		return -EINVAL;
+
+	wafl_reg_state = (struct amdgpu_reg_state_wafl_v1_0 *)buf;
+
+	szbuf = sizeof(*wafl_reg_state) +
+		amdgpu_reginst_size(max_wafl_instances, sizeof(*wafl_regs),
+				    NUM_WAFL_SMN_REGS);
+
+	if (max_size < szbuf)
+		return -EOVERFLOW;
+
+	p = &wafl_reg_state->wafl_state_regs[0];
+	for_each_inst(i, adev->aid_mask) {
+		for (j = 0; j < wafl_inst; ++j) {
+			wafl_regs = (struct amdgpu_regs_wafl_v1_0 *)p;
+			wafl_regs->inst_header.instance = inst++;
+
+			wafl_regs->inst_header.state = AMDGPU_INST_S_OK;
+			wafl_regs->inst_header.num_smn_regs = NUM_WAFL_SMN_REGS;
+
+			reg_data = wafl_regs->smn_reg_values;
+
+			for (r = 0; r < ARRAY_SIZE(wafl_reg_addrs); r++) {
+				start_addr = wafl_reg_addrs[r].start_addr;
+				incrx = wafl_reg_addrs[r].incrx;
+				num_regs = wafl_reg_addrs[r].num_regs;
+				for (n = 0; n < num_regs; n++) {
+					aqua_read_smn_ext(
+						adev, reg_data,
+						WAFL_LINK_REG(start_addr, j) +
+							n * incrx,
+						i);
+					++reg_data;
+				}
+			}
+			p = reg_data;
+		}
+	}
+
+	wafl_reg_state->common_header.structure_size = szbuf;
+	wafl_reg_state->common_header.format_revision = 1;
+	wafl_reg_state->common_header.content_revision = 0;
+	wafl_reg_state->common_header.state_type = AMDGPU_REG_STATE_TYPE_WAFL;
+	wafl_reg_state->common_header.num_instances = max_wafl_instances;
+
+	return wafl_reg_state->common_header.structure_size;
+}
+
+#define smnreg_0x1B311060 0x1B311060
+#define smnreg_0x1B411060 0x1B411060
+#define smnreg_0x1B511060 0x1B511060
+#define smnreg_0x1B611060 0x1B611060
+
+#define smnreg_0x1C307120 0x1C307120
+#define smnreg_0x1C317120 0x1C317120
+
+#define smnreg_0x1C320830 0x1C320830
+#define smnreg_0x1C380830 0x1C380830
+#define smnreg_0x1C3D0830 0x1C3D0830
+#define smnreg_0x1C420830 0x1C420830
+
+#define smnreg_0x1C320100 0x1C320100
+#define smnreg_0x1C380100 0x1C380100
+#define smnreg_0x1C3D0100 0x1C3D0100
+#define smnreg_0x1C420100 0x1C420100
+
+#define smnreg_0x1B310500 0x1B310500
+#define smnreg_0x1C300400 0x1C300400
+
+#define USR_CAKE_INCR 0x11000
+#define USR_LINK_INCR 0x100000
+#define USR_CP_INCR 0x10000
+
+#define NUM_USR_SMN_REGS	20
+
+struct aqua_reg_list usr_reg_addrs[] = {
+	{ smnreg_0x1B311060, 4, DW_ADDR_INCR },
+	{ smnreg_0x1B411060, 4, DW_ADDR_INCR },
+	{ smnreg_0x1B511060, 4, DW_ADDR_INCR },
+	{ smnreg_0x1B611060, 4, DW_ADDR_INCR },
+	{ smnreg_0x1C307120, 2, DW_ADDR_INCR },
+	{ smnreg_0x1C317120, 2, DW_ADDR_INCR },
+};
+
+#define NUM_USR1_SMN_REGS	46
+struct aqua_reg_list usr1_reg_addrs[] = {
+	{ smnreg_0x1C320830, 6, USR_CAKE_INCR },
+	{ smnreg_0x1C380830, 5, USR_CAKE_INCR },
+	{ smnreg_0x1C3D0830, 5, USR_CAKE_INCR },
+	{ smnreg_0x1C420830, 4, USR_CAKE_INCR },
+	{ smnreg_0x1C320100, 6, USR_CAKE_INCR },
+	{ smnreg_0x1C380100, 5, USR_CAKE_INCR },
+	{ smnreg_0x1C3D0100, 5, USR_CAKE_INCR },
+	{ smnreg_0x1C420100, 4, USR_CAKE_INCR },
+	{ smnreg_0x1B310500, 4, USR_LINK_INCR },
+	{ smnreg_0x1C300400, 2, USR_CP_INCR },
+};
+
+static ssize_t aqua_vanjaram_read_usr_state(struct amdgpu_device *adev,
+					    void *buf, size_t max_size,
+					    int reg_state)
+{
+	uint32_t start_addr, incrx, num_regs, szbuf, num_smn;
+	struct amdgpu_reg_state_usr_v1_0 *usr_reg_state;
+	struct amdgpu_regs_usr_v1_0 *usr_regs;
+	struct amdgpu_smn_reg_data *reg_data;
+	const int max_usr_instances = 4;
+	struct aqua_reg_list *reg_addrs;
+	int inst = 0, i, n, r, arr_size;
+	void *p;
+
+	if (!buf || !max_size)
+		return -EINVAL;
+
+	switch (reg_state) {
+	case AMDGPU_REG_STATE_TYPE_USR:
+		arr_size = ARRAY_SIZE(usr_reg_addrs);
+		reg_addrs = usr_reg_addrs;
+		num_smn = NUM_USR_SMN_REGS;
+		break;
+	case AMDGPU_REG_STATE_TYPE_USR_1:
+		arr_size = ARRAY_SIZE(usr1_reg_addrs);
+		reg_addrs = usr1_reg_addrs;
+		num_smn = NUM_USR1_SMN_REGS;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	usr_reg_state = (struct amdgpu_reg_state_usr_v1_0 *)buf;
+
+	szbuf = sizeof(*usr_reg_state) + amdgpu_reginst_size(max_usr_instances,
+							     sizeof(*usr_regs),
+							     num_smn);
+	if (max_size < szbuf)
+		return -EOVERFLOW;
+
+	p = &usr_reg_state->usr_state_regs[0];
+	for_each_inst(i, adev->aid_mask) {
+		usr_regs = (struct amdgpu_regs_usr_v1_0 *)p;
+		usr_regs->inst_header.instance = inst++;
+		usr_regs->inst_header.state = AMDGPU_INST_S_OK;
+		usr_regs->inst_header.num_smn_regs = num_smn;
+		reg_data = usr_regs->smn_reg_values;
+
+		for (r = 0; r < arr_size; r++) {
+			start_addr = reg_addrs[r].start_addr;
+			incrx = reg_addrs[r].incrx;
+			num_regs = reg_addrs[r].num_regs;
+			for (n = 0; n < num_regs; n++) {
+				aqua_read_smn_ext(adev, reg_data,
+						  start_addr + n * incrx, i);
+				reg_data++;
+			}
+		}
+		p = reg_data;
+	}
+
+	usr_reg_state->common_header.structure_size = szbuf;
+	usr_reg_state->common_header.format_revision = 1;
+	usr_reg_state->common_header.content_revision = 0;
+	usr_reg_state->common_header.state_type = AMDGPU_REG_STATE_TYPE_USR;
+	usr_reg_state->common_header.num_instances = max_usr_instances;
+
+	return usr_reg_state->common_header.structure_size;
+}
+
+ssize_t aqua_vanjaram_get_reg_state(struct amdgpu_device *adev,
+				    enum amdgpu_reg_state reg_state, void *buf,
+				    size_t max_size)
+{
+	ssize_t size;
+
+	switch (reg_state) {
+	case AMDGPU_REG_STATE_TYPE_PCIE:
+		size = aqua_vanjaram_read_pcie_state(adev, buf, max_size);
+		break;
+	case AMDGPU_REG_STATE_TYPE_XGMI:
+		size = aqua_vanjaram_read_xgmi_state(adev, buf, max_size);
+		break;
+	case AMDGPU_REG_STATE_TYPE_WAFL:
+		size = aqua_vanjaram_read_wafl_state(adev, buf, max_size);
+		break;
+	case AMDGPU_REG_STATE_TYPE_USR:
+		size = aqua_vanjaram_read_usr_state(adev, buf, max_size,
+						    AMDGPU_REG_STATE_TYPE_USR);
+		break;
+	case AMDGPU_REG_STATE_TYPE_USR_1:
+		size = aqua_vanjaram_read_usr_state(
+			adev, buf, max_size, AMDGPU_REG_STATE_TYPE_USR_1);
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return size;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/athub_v1_0.c b/drivers/gpu/drm/amd/amdgpu/athub_v1_0.c
index 3ea557864320..42f4e163e251 100644
--- a/drivers/gpu/drm/amd/amdgpu/athub_v1_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/athub_v1_0.c
@@ -68,12 +68,13 @@ int athub_v1_0_set_clockgating(struct amdgpu_device *adev,
 	if (amdgpu_sriov_vf(adev))
 		return 0;
 
-	switch (adev->asic_type) {
-	case CHIP_VEGA10:
-	case CHIP_VEGA12:
-	case CHIP_VEGA20:
-	case CHIP_RAVEN:
-	case CHIP_RENOIR:
+	switch (amdgpu_ip_version(adev, ATHUB_HWIP, 0)) {
+	case IP_VERSION(9, 0, 0):
+	case IP_VERSION(9, 1, 0):
+	case IP_VERSION(9, 2, 0):
+	case IP_VERSION(9, 3, 0):
+	case IP_VERSION(9, 4, 0):
+	case IP_VERSION(1, 5, 0):
 		athub_update_medium_grain_clock_gating(adev,
 				state == AMD_CG_STATE_GATE);
 		athub_update_medium_grain_light_sleep(adev,
@@ -86,7 +87,7 @@ int athub_v1_0_set_clockgating(struct amdgpu_device *adev,
 	return 0;
 }
 
-void athub_v1_0_get_clockgating(struct amdgpu_device *adev, u32 *flags)
+void athub_v1_0_get_clockgating(struct amdgpu_device *adev, u64 *flags)
 {
 	int data;
 
diff --git a/drivers/gpu/drm/amd/amdgpu/athub_v1_0.h b/drivers/gpu/drm/amd/amdgpu/athub_v1_0.h
index b279af59e34f..6be0a6704ea7 100644
--- a/drivers/gpu/drm/amd/amdgpu/athub_v1_0.h
+++ b/drivers/gpu/drm/amd/amdgpu/athub_v1_0.h
@@ -25,6 +25,6 @@
 
 int athub_v1_0_set_clockgating(struct amdgpu_device *adev,
 			       enum amd_clockgating_state state);
-void athub_v1_0_get_clockgating(struct amdgpu_device *adev, u32 *flags);
+void athub_v1_0_get_clockgating(struct amdgpu_device *adev, u64 *flags);
 
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/athub_v2_0.c b/drivers/gpu/drm/amd/amdgpu/athub_v2_0.c
index ab6a07e5e8c4..5a122f50a6e7 100644
--- a/drivers/gpu/drm/amd/amdgpu/athub_v2_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/athub_v2_0.c
@@ -77,7 +77,8 @@ int athub_v2_0_set_clockgating(struct amdgpu_device *adev,
 	if (amdgpu_sriov_vf(adev))
 		return 0;
 
-	switch (adev->ip_versions[ATHUB_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, ATHUB_HWIP, 0)) {
+	case IP_VERSION(1, 3, 1):
 	case IP_VERSION(2, 0, 0):
 	case IP_VERSION(2, 0, 2):
 		athub_v2_0_update_medium_grain_clock_gating(adev,
@@ -92,7 +93,7 @@ int athub_v2_0_set_clockgating(struct amdgpu_device *adev,
 	return 0;
 }
 
-void athub_v2_0_get_clockgating(struct amdgpu_device *adev, u32 *flags)
+void athub_v2_0_get_clockgating(struct amdgpu_device *adev, u64 *flags)
 {
 	int data;
 
diff --git a/drivers/gpu/drm/amd/amdgpu/athub_v2_0.h b/drivers/gpu/drm/amd/amdgpu/athub_v2_0.h
index 02932c1c8bab..8b763f6dfd81 100644
--- a/drivers/gpu/drm/amd/amdgpu/athub_v2_0.h
+++ b/drivers/gpu/drm/amd/amdgpu/athub_v2_0.h
@@ -25,6 +25,6 @@
 
 int athub_v2_0_set_clockgating(struct amdgpu_device *adev,
 			       enum amd_clockgating_state state);
-void athub_v2_0_get_clockgating(struct amdgpu_device *adev, u32 *flags);
+void athub_v2_0_get_clockgating(struct amdgpu_device *adev, u64 *flags);
 
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/athub_v2_1.c b/drivers/gpu/drm/amd/amdgpu/athub_v2_1.c
index 2edefd10e56c..e143fcc46148 100644
--- a/drivers/gpu/drm/amd/amdgpu/athub_v2_1.c
+++ b/drivers/gpu/drm/amd/amdgpu/athub_v2_1.c
@@ -70,10 +70,11 @@ int athub_v2_1_set_clockgating(struct amdgpu_device *adev,
 	if (amdgpu_sriov_vf(adev))
 		return 0;
 
-	switch (adev->ip_versions[ATHUB_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, ATHUB_HWIP, 0)) {
 	case IP_VERSION(2, 1, 0):
 	case IP_VERSION(2, 1, 1):
 	case IP_VERSION(2, 1, 2):
+	case IP_VERSION(2, 4, 0):
 		athub_v2_1_update_medium_grain_clock_gating(adev, state == AMD_CG_STATE_GATE);
 		athub_v2_1_update_medium_grain_light_sleep(adev, state == AMD_CG_STATE_GATE);
 		break;
@@ -84,7 +85,7 @@ int athub_v2_1_set_clockgating(struct amdgpu_device *adev,
 	return 0;
 }
 
-void athub_v2_1_get_clockgating(struct amdgpu_device *adev, u32 *flags)
+void athub_v2_1_get_clockgating(struct amdgpu_device *adev, u64 *flags)
 {
 	int data;
 
diff --git a/drivers/gpu/drm/amd/amdgpu/athub_v2_1.h b/drivers/gpu/drm/amd/amdgpu/athub_v2_1.h
index 5e6824c0f591..b799f14bce03 100644
--- a/drivers/gpu/drm/amd/amdgpu/athub_v2_1.h
+++ b/drivers/gpu/drm/amd/amdgpu/athub_v2_1.h
@@ -25,6 +25,6 @@
 
 int athub_v2_1_set_clockgating(struct amdgpu_device *adev,
 			       enum amd_clockgating_state state);
-void athub_v2_1_get_clockgating(struct amdgpu_device *adev, u32 *flags);
+void athub_v2_1_get_clockgating(struct amdgpu_device *adev, u64 *flags);
 
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/athub_v3_0.c b/drivers/gpu/drm/amd/amdgpu/athub_v3_0.c
new file mode 100644
index 000000000000..d1bba9c64e16
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/athub_v3_0.c
@@ -0,0 +1,139 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "amdgpu.h"
+#include "athub_v3_0.h"
+#include "athub/athub_3_0_0_offset.h"
+#include "athub/athub_3_0_0_sh_mask.h"
+#include "navi10_enum.h"
+#include "soc15_common.h"
+
+#define regATHUB_MISC_CNTL_V3_0_1			0x00d7
+#define regATHUB_MISC_CNTL_V3_0_1_BASE_IDX		0
+#define regATHUB_MISC_CNTL_V3_3_0			0x00d8
+#define regATHUB_MISC_CNTL_V3_3_0_BASE_IDX		0
+
+
+static uint32_t athub_v3_0_get_cg_cntl(struct amdgpu_device *adev)
+{
+	uint32_t data;
+
+	switch (amdgpu_ip_version(adev, ATHUB_HWIP, 0)) {
+	case IP_VERSION(3, 0, 1):
+		data = RREG32_SOC15(ATHUB, 0, regATHUB_MISC_CNTL_V3_0_1);
+		break;
+	case IP_VERSION(3, 3, 0):
+		data = RREG32_SOC15(ATHUB, 0, regATHUB_MISC_CNTL_V3_3_0);
+		break;
+	default:
+		data = RREG32_SOC15(ATHUB, 0, regATHUB_MISC_CNTL);
+		break;
+	}
+	return data;
+}
+
+static void athub_v3_0_set_cg_cntl(struct amdgpu_device *adev, uint32_t data)
+{
+	switch (amdgpu_ip_version(adev, ATHUB_HWIP, 0)) {
+	case IP_VERSION(3, 0, 1):
+		WREG32_SOC15(ATHUB, 0, regATHUB_MISC_CNTL_V3_0_1, data);
+		break;
+	case IP_VERSION(3, 3, 0):
+		WREG32_SOC15(ATHUB, 0, regATHUB_MISC_CNTL_V3_3_0, data);
+		break;
+	default:
+		WREG32_SOC15(ATHUB, 0, regATHUB_MISC_CNTL, data);
+		break;
+	}
+}
+
+static void
+athub_v3_0_update_medium_grain_clock_gating(struct amdgpu_device *adev,
+					    bool enable)
+{
+	uint32_t def, data;
+
+	def = data = athub_v3_0_get_cg_cntl(adev);
+
+	if (enable && (adev->cg_flags & AMD_CG_SUPPORT_ATHUB_MGCG))
+		data |= ATHUB_MISC_CNTL__CG_ENABLE_MASK;
+	else
+		data &= ~ATHUB_MISC_CNTL__CG_ENABLE_MASK;
+
+	if (def != data)
+		athub_v3_0_set_cg_cntl(adev, data);
+}
+
+static void
+athub_v3_0_update_medium_grain_light_sleep(struct amdgpu_device *adev,
+					   bool enable)
+{
+	uint32_t def, data;
+
+	def = data = athub_v3_0_get_cg_cntl(adev);
+
+	if (enable && (adev->cg_flags & AMD_CG_SUPPORT_ATHUB_LS))
+		data |= ATHUB_MISC_CNTL__CG_MEM_LS_ENABLE_MASK;
+	else
+		data &= ~ATHUB_MISC_CNTL__CG_MEM_LS_ENABLE_MASK;
+
+	if (def != data)
+		athub_v3_0_set_cg_cntl(adev, data);
+}
+
+int athub_v3_0_set_clockgating(struct amdgpu_device *adev,
+			       enum amd_clockgating_state state)
+{
+	if (amdgpu_sriov_vf(adev))
+		return 0;
+
+	switch (amdgpu_ip_version(adev, ATHUB_HWIP, 0)) {
+	case IP_VERSION(3, 0, 0):
+	case IP_VERSION(3, 0, 1):
+	case IP_VERSION(3, 0, 2):
+	case IP_VERSION(3, 3, 0):
+		athub_v3_0_update_medium_grain_clock_gating(adev,
+				state == AMD_CG_STATE_GATE);
+		athub_v3_0_update_medium_grain_light_sleep(adev,
+				state == AMD_CG_STATE_GATE);
+		break;
+	default:
+		break;
+	}
+
+	return 0;
+}
+
+void athub_v3_0_get_clockgating(struct amdgpu_device *adev, u64 *flags)
+{
+	int data;
+
+	/* AMD_CG_SUPPORT_ATHUB_MGCG */
+	data = athub_v3_0_get_cg_cntl(adev);
+	if (data & ATHUB_MISC_CNTL__CG_ENABLE_MASK)
+		*flags |= AMD_CG_SUPPORT_ATHUB_MGCG;
+
+	/* AMD_CG_SUPPORT_ATHUB_LS */
+	if (data & ATHUB_MISC_CNTL__CG_MEM_LS_ENABLE_MASK)
+		*flags |= AMD_CG_SUPPORT_ATHUB_LS;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/athub_v3_0.h b/drivers/gpu/drm/amd/amdgpu/athub_v3_0.h
new file mode 100644
index 000000000000..e08a7d564365
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/athub_v3_0.h
@@ -0,0 +1,30 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __ATHUB_V3_0_H__
+#define __ATHUB_V3_0_H__
+
+int athub_v3_0_set_clockgating(struct amdgpu_device *adev,
+			       enum amd_clockgating_state state);
+void athub_v3_0_get_clockgating(struct amdgpu_device *adev, u64 *flags);
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/athub_v4_1_0.c b/drivers/gpu/drm/amd/amdgpu/athub_v4_1_0.c
new file mode 100644
index 000000000000..8a0773b80864
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/athub_v4_1_0.c
@@ -0,0 +1,122 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "amdgpu.h"
+#include "athub_v4_1_0.h"
+#include "athub/athub_4_1_0_offset.h"
+#include "athub/athub_4_1_0_sh_mask.h"
+#include "soc15_common.h"
+
+static uint32_t athub_v4_1_0_get_cg_cntl(struct amdgpu_device *adev)
+{
+	uint32_t data;
+
+	switch (amdgpu_ip_version(adev, ATHUB_HWIP, 0)) {
+	case IP_VERSION(4, 1, 0):
+		data = RREG32_SOC15(ATHUB, 0, regATHUB_MISC_CNTL);
+		break;
+	default:
+		data = 0;
+		break;
+	}
+	return data;
+}
+
+static void athub_v4_1_0_set_cg_cntl(struct amdgpu_device *adev, uint32_t data)
+{
+	switch (amdgpu_ip_version(adev, ATHUB_HWIP, 0)) {
+	case IP_VERSION(4, 1, 0):
+		WREG32_SOC15(ATHUB, 0, regATHUB_MISC_CNTL, data);
+		break;
+	default:
+		break;
+	}
+}
+
+static void
+athub_v4_1_0_update_medium_grain_clock_gating(struct amdgpu_device *adev,
+					      bool enable)
+{
+	uint32_t def, data;
+
+	def = data = athub_v4_1_0_get_cg_cntl(adev);
+
+	if (enable && (adev->cg_flags & AMD_CG_SUPPORT_ATHUB_MGCG))
+		data |= ATHUB_MISC_CNTL__CG_ENABLE_MASK;
+	else
+		data &= ~ATHUB_MISC_CNTL__CG_ENABLE_MASK;
+
+	if (def != data)
+		athub_v4_1_0_set_cg_cntl(adev, data);
+}
+
+static void
+athub_v4_1_0_update_medium_grain_light_sleep(struct amdgpu_device *adev,
+					     bool enable)
+{
+	uint32_t def, data;
+
+	def = data = athub_v4_1_0_get_cg_cntl(adev);
+
+	if (enable && (adev->cg_flags & AMD_CG_SUPPORT_ATHUB_LS))
+		data |= ATHUB_MISC_CNTL__CG_MEM_LS_ENABLE_MASK;
+	else
+		data &= ~ATHUB_MISC_CNTL__CG_MEM_LS_ENABLE_MASK;
+
+	if (def != data)
+		athub_v4_1_0_set_cg_cntl(adev, data);
+}
+
+int athub_v4_1_0_set_clockgating(struct amdgpu_device *adev,
+				 enum amd_clockgating_state state)
+{
+	if (amdgpu_sriov_vf(adev))
+		return 0;
+
+	switch (amdgpu_ip_version(adev, ATHUB_HWIP, 0)) {
+	case IP_VERSION(4, 1, 0):
+		athub_v4_1_0_update_medium_grain_clock_gating(adev,
+				state == AMD_CG_STATE_GATE);
+		athub_v4_1_0_update_medium_grain_light_sleep(adev,
+				state == AMD_CG_STATE_GATE);
+		break;
+	default:
+		break;
+	}
+
+	return 0;
+}
+
+void athub_v4_1_0_get_clockgating(struct amdgpu_device *adev, u64 *flags)
+{
+	int data;
+
+	/* AMD_CG_SUPPORT_ATHUB_MGCG */
+	data = athub_v4_1_0_get_cg_cntl(adev);
+	if (data & ATHUB_MISC_CNTL__CG_ENABLE_MASK)
+		*flags |= AMD_CG_SUPPORT_ATHUB_MGCG;
+
+	/* AMD_CG_SUPPORT_ATHUB_LS */
+	if (data & ATHUB_MISC_CNTL__CG_MEM_LS_ENABLE_MASK)
+		*flags |= AMD_CG_SUPPORT_ATHUB_LS;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/athub_v4_1_0.h b/drivers/gpu/drm/amd/amdgpu/athub_v4_1_0.h
new file mode 100644
index 000000000000..4d18d0998fa8
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/athub_v4_1_0.h
@@ -0,0 +1,30 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __ATHUB_V4_1_0_H__
+#define __ATHUB_V4_1_0_H__
+
+int athub_v4_1_0_set_clockgating(struct amdgpu_device *adev,
+				 enum amd_clockgating_state state);
+void athub_v4_1_0_get_clockgating(struct amdgpu_device *adev, u64 *flags);
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/atom.c b/drivers/gpu/drm/amd/amdgpu/atom.c
index 6fa2229b7229..7a063e44d429 100644
--- a/drivers/gpu/drm/amd/amdgpu/atom.c
+++ b/drivers/gpu/drm/amd/amdgpu/atom.c
@@ -25,7 +25,9 @@
 #include <linux/module.h>
 #include <linux/sched.h>
 #include <linux/slab.h>
-#include <asm/unaligned.h>
+#include <linux/string_helpers.h>
+
+#include <linux/unaligned.h>
 
 #include <drm/drm_util.h>
 
@@ -60,6 +62,7 @@
 typedef struct {
 	struct atom_context *ctx;
 	uint32_t *ps, *ws;
+	int ps_size, ws_size;
 	int ps_shift;
 	uint16_t start;
 	unsigned last_jump;
@@ -68,8 +71,8 @@ typedef struct {
 } atom_exec_context;
 
 int amdgpu_atom_debug;
-static int amdgpu_atom_execute_table_locked(struct atom_context *ctx, int index, uint32_t *params);
-int amdgpu_atom_execute_table(struct atom_context *ctx, int index, uint32_t *params);
+static int amdgpu_atom_execute_table_locked(struct atom_context *ctx, int index, uint32_t *params, int params_size);
+int amdgpu_atom_execute_table(struct atom_context *ctx, int index, uint32_t *params, int params_size);
 
 static uint32_t atom_arg_mask[8] =
 	{ 0xFFFFFFFF, 0xFFFF, 0xFFFF00, 0xFFFF0000, 0xFF, 0xFF00, 0xFF0000,
@@ -221,7 +224,10 @@ static uint32_t atom_get_src_int(atom_exec_context *ctx, uint8_t attr,
 		(*ptr)++;
 		/* get_unaligned_le32 avoids unaligned accesses from atombios
 		 * tables, noticed on a DEC Alpha. */
-		val = get_unaligned_le32((u32 *)&ctx->ps[idx]);
+		if (idx < ctx->ps_size)
+			val = get_unaligned_le32((u32 *)&ctx->ps[idx]);
+		else
+			pr_info("PS index out of range: %i > %i\n", idx, ctx->ps_size);
 		if (print)
 			DEBUG("PS[0x%02X,0x%04X]", idx, val);
 		break;
@@ -259,7 +265,10 @@ static uint32_t atom_get_src_int(atom_exec_context *ctx, uint8_t attr,
 			val = gctx->reg_block;
 			break;
 		default:
-			val = ctx->ws[idx];
+			if (idx < ctx->ws_size)
+				val = ctx->ws[idx];
+			else
+				pr_info("WS index out of range: %i > %i\n", idx, ctx->ws_size);
 		}
 		break;
 	case ATOM_ARG_ID:
@@ -292,7 +301,7 @@ static uint32_t atom_get_src_int(atom_exec_context *ctx, uint8_t attr,
 			(*ptr) += 4;
 			if (print)
 				DEBUG("IMM 0x%08X\n", val);
-			return val;
+			break;
 		case ATOM_SRC_WORD0:
 		case ATOM_SRC_WORD8:
 		case ATOM_SRC_WORD16:
@@ -300,7 +309,7 @@ static uint32_t atom_get_src_int(atom_exec_context *ctx, uint8_t attr,
 			(*ptr) += 2;
 			if (print)
 				DEBUG("IMM 0x%04X\n", val);
-			return val;
+			break;
 		case ATOM_SRC_BYTE0:
 		case ATOM_SRC_BYTE8:
 		case ATOM_SRC_BYTE16:
@@ -309,9 +318,9 @@ static uint32_t atom_get_src_int(atom_exec_context *ctx, uint8_t attr,
 			(*ptr)++;
 			if (print)
 				DEBUG("IMM 0x%02X\n", val);
-			return val;
+			break;
 		}
-		return 0;
+		return val;
 	case ATOM_ARG_PLL:
 		idx = U8(*ptr);
 		(*ptr)++;
@@ -393,7 +402,6 @@ static void atom_skip_src_int(atom_exec_context *ctx, uint8_t attr, int *ptr)
 			(*ptr)++;
 			return;
 		}
-		return;
 	}
 }
 
@@ -494,6 +502,10 @@ static void atom_put_dst(atom_exec_context *ctx, int arg, uint8_t attr,
 		idx = U8(*ptr);
 		(*ptr)++;
 		DEBUG("PS[0x%02X]", idx);
+		if (idx >= ctx->ps_size) {
+			pr_info("PS index out of range: %i > %i\n", idx, ctx->ps_size);
+			return;
+		}
 		ctx->ps[idx] = cpu_to_le32(val);
 		break;
 	case ATOM_ARG_WS:
@@ -526,6 +538,10 @@ static void atom_put_dst(atom_exec_context *ctx, int arg, uint8_t attr,
 			gctx->reg_block = val;
 			break;
 		default:
+			if (idx >= ctx->ws_size) {
+				pr_info("WS index out of range: %i > %i\n", idx, ctx->ws_size);
+				return;
+			}
 			ctx->ws[idx] = val;
 		}
 		break;
@@ -623,7 +639,7 @@ static void atom_op_calltable(atom_exec_context *ctx, int *ptr, int arg)
 	else
 		SDEBUG("   table: %d\n", idx);
 	if (U16(ctx->ctx->cmd_table + 4 + 2 * idx))
-		r = amdgpu_atom_execute_table_locked(ctx->ctx, idx, ctx->ps + ctx->ps_shift);
+		r = amdgpu_atom_execute_table_locked(ctx->ctx, idx, ctx->ps + ctx->ps_shift, ctx->ps_size - ctx->ps_shift);
 	if (r) {
 		ctx->abort = true;
 	}
@@ -740,7 +756,7 @@ static void atom_op_jump(atom_exec_context *ctx, int *ptr, int arg)
 		break;
 	}
 	if (arg != ATOM_COND_ALWAYS)
-		SDEBUG("   taken: %s\n", execute ? "yes" : "no");
+		SDEBUG("   taken: %s\n", str_yes_no(execute));
 	SDEBUG("   target: 0x%04X\n", target);
 	if (execute) {
 		if (ctx->last_jump == (ctx->start + target)) {
@@ -1202,7 +1218,7 @@ static struct {
 	atom_op_div32, ATOM_ARG_WS},
 };
 
-static int amdgpu_atom_execute_table_locked(struct atom_context *ctx, int index, uint32_t *params)
+static int amdgpu_atom_execute_table_locked(struct atom_context *ctx, int index, uint32_t *params, int params_size)
 {
 	int base = CU16(ctx->cmd_table + 4 + 2 * index);
 	int len, ws, ps, ptr;
@@ -1224,12 +1240,21 @@ static int amdgpu_atom_execute_table_locked(struct atom_context *ctx, int index,
 	ectx.ps_shift = ps / 4;
 	ectx.start = base;
 	ectx.ps = params;
+	ectx.ps_size = params_size;
 	ectx.abort = false;
 	ectx.last_jump = 0;
-	if (ws)
+	ectx.last_jump_jiffies = 0;
+	if (ws) {
 		ectx.ws = kcalloc(4, ws, GFP_KERNEL);
-	else
+		if (!ectx.ws) {
+			ret = -ENOMEM;
+			goto free;
+		}
+		ectx.ws_size = ws;
+	} else {
 		ectx.ws = NULL;
+		ectx.ws_size = 0;
+	}
 
 	debug_depth++;
 	while (1) {
@@ -1263,7 +1288,7 @@ free:
 	return ret;
 }
 
-int amdgpu_atom_execute_table(struct atom_context *ctx, int index, uint32_t *params)
+int amdgpu_atom_execute_table(struct atom_context *ctx, int index, uint32_t *params, int params_size)
 {
 	int r;
 
@@ -1279,7 +1304,7 @@ int amdgpu_atom_execute_table(struct atom_context *ctx, int index, uint32_t *par
 	/* reset divmul */
 	ctx->divmul[0] = 0;
 	ctx->divmul[1] = 0;
-	r = amdgpu_atom_execute_table_locked(ctx, index, params);
+	r = amdgpu_atom_execute_table_locked(ctx, index, params, params_size);
 	mutex_unlock(&ctx->mutex);
 	return r;
 }
@@ -1423,6 +1448,7 @@ static void atom_get_vbios_pn(struct atom_context *ctx)
 		if (vbios_str == NULL)
 			vbios_str += sizeof(BIOS_ATOM_PREFIX) - 1;
 	}
+	OPTIMIZER_HIDE_VAR(vbios_str);
 	if (vbios_str != NULL && *vbios_str == 0)
 		vbios_str++;
 
@@ -1436,14 +1462,33 @@ static void atom_get_vbios_pn(struct atom_context *ctx)
 
 		ctx->vbios_pn[count] = 0;
 	}
+
+	pr_info("ATOM BIOS: %s\n", ctx->vbios_pn);
 }
 
 static void atom_get_vbios_version(struct atom_context *ctx)
 {
+	unsigned short start = 3, end;
 	unsigned char *vbios_ver;
+	unsigned char *p_rom;
+
+	p_rom = ctx->bios;
+	/* Search from strings offset if it's present */
+	start = *(unsigned short *)(p_rom +
+				    OFFSET_TO_GET_ATOMBIOS_STRING_START);
+
+	/* Search till atom rom header start point */
+	end = *(unsigned short *)(p_rom + OFFSET_TO_ATOM_ROM_HEADER_POINTER);
+
+	/* Use hardcoded offsets, if the offsets are not populated */
+	if (end <= start) {
+		start = 3;
+		end = 1024;
+	}
 
 	/* find anchor ATOMBIOSBK-AMD */
-	vbios_ver = atom_find_str_in_rom(ctx, BIOS_VERSION_PREFIX, 3, 1024, 64);
+	vbios_ver =
+		atom_find_str_in_rom(ctx, BIOS_VERSION_PREFIX, start, end, 64);
 	if (vbios_ver != NULL) {
 		/* skip ATOMBIOSBK-AMD VER */
 		vbios_ver += 18;
@@ -1453,16 +1498,36 @@ static void atom_get_vbios_version(struct atom_context *ctx)
 	}
 }
 
+static void atom_get_vbios_build(struct atom_context *ctx)
+{
+	unsigned char *atom_rom_hdr;
+	unsigned char *str;
+	uint16_t base, len;
+
+	base = CU16(ATOM_ROM_TABLE_PTR);
+	atom_rom_hdr = CSTR(base);
+
+	str = CSTR(CU16(base + ATOM_ROM_CFG_PTR));
+	/* Skip config string */
+	while (str < atom_rom_hdr && *str++)
+		;
+	/* Skip change list string */
+	while (str < atom_rom_hdr && *str++)
+		;
+
+	len = min(atom_rom_hdr - str, STRLEN_NORMAL);
+	if (len)
+		strscpy(ctx->build_num, str, len);
+}
+
 struct atom_context *amdgpu_atom_parse(struct card_info *card, void *bios)
 {
 	int base;
 	struct atom_context *ctx =
 	    kzalloc(sizeof(struct atom_context), GFP_KERNEL);
-	char *str;
 	struct _ATOM_ROM_HEADER *atom_rom_header;
 	struct _ATOM_MASTER_DATA_TABLE *master_table;
 	struct _ATOM_FIRMWARE_INFO *atom_fw_info;
-	u16 idx;
 
 	if (!ctx)
 		return NULL;
@@ -1500,16 +1565,6 @@ struct atom_context *amdgpu_atom_parse(struct card_info *card, void *bios)
 		return NULL;
 	}
 
-	idx = CU16(ATOM_ROM_PART_NUMBER_PTR);
-	if (idx == 0)
-		idx = 0x80;
-
-	str = CSTR(idx);
-	if (*str != '\0') {
-		pr_info("ATOM BIOS: %s\n", str);
-		strlcpy(ctx->vbios_version, str, sizeof(ctx->vbios_version));
-	}
-
 	atom_rom_header = (struct _ATOM_ROM_HEADER *)CSTR(base);
 	if (atom_rom_header->usMasterDataTableOffset != 0) {
 		master_table = (struct _ATOM_MASTER_DATA_TABLE *)
@@ -1525,6 +1580,7 @@ struct atom_context *amdgpu_atom_parse(struct card_info *card, void *bios)
 	atom_get_vbios_pn(ctx);
 	atom_get_vbios_date(ctx);
 	atom_get_vbios_version(ctx);
+	atom_get_vbios_build(ctx);
 
 	return ctx;
 }
@@ -1544,7 +1600,7 @@ int amdgpu_atom_asic_init(struct atom_context *ctx)
 
 	if (!CU16(ctx->cmd_table + 4 + 2 * ATOM_CMD_INIT))
 		return 1;
-	ret = amdgpu_atom_execute_table(ctx, ATOM_CMD_INIT, ps);
+	ret = amdgpu_atom_execute_table(ctx, ATOM_CMD_INIT, ps, 16);
 	if (ret)
 		return ret;
 
diff --git a/drivers/gpu/drm/amd/amdgpu/atom.h b/drivers/gpu/drm/amd/amdgpu/atom.h
index 0c1839824520..825ff28731f5 100644
--- a/drivers/gpu/drm/amd/amdgpu/atom.h
+++ b/drivers/gpu/drm/amd/amdgpu/atom.h
@@ -33,11 +33,11 @@ struct drm_device;
 #define ATOM_ATI_MAGIC_PTR	0x30
 #define ATOM_ATI_MAGIC		" 761295520"
 #define ATOM_ROM_TABLE_PTR	0x48
-#define ATOM_ROM_PART_NUMBER_PTR	0x6E
 
 #define ATOM_ROM_MAGIC		"ATOM"
 #define ATOM_ROM_MAGIC_PTR	4
 
+#define ATOM_ROM_CFG_PTR 0xC
 #define ATOM_ROM_MSG_PTR	0x10
 #define ATOM_ROM_CMD_PTR	0x1E
 #define ATOM_ROM_DATA_PTR	0x20
@@ -118,12 +118,15 @@ struct drm_device;
 
 struct card_info {
 	struct drm_device *dev;
-	void (* reg_write)(struct card_info *, uint32_t, uint32_t);   /*  filled by driver */
-	uint32_t (* reg_read)(struct card_info *, uint32_t);          /*  filled by driver */
-	void (* mc_write)(struct card_info *, uint32_t, uint32_t);   /*  filled by driver */
-	uint32_t (* mc_read)(struct card_info *, uint32_t);          /*  filled by driver */
-	void (* pll_write)(struct card_info *, uint32_t, uint32_t);   /*  filled by driver */
-	uint32_t (* pll_read)(struct card_info *, uint32_t);          /*  filled by driver */
+	void (*reg_write)(struct card_info *info,
+			  u32 reg, uint32_t val);   /*  filled by driver */
+	uint32_t (*reg_read)(struct card_info *info, uint32_t reg);          /*  filled by driver */
+	void (*mc_write)(struct card_info *info,
+			 u32 reg, uint32_t val);   /*  filled by driver */
+	uint32_t (*mc_read)(struct card_info *info, uint32_t reg);          /*  filled by driver */
+	void (*pll_write)(struct card_info *info,
+			  u32 reg, uint32_t val);   /*  filled by driver */
+	uint32_t (*pll_read)(struct card_info *info, uint32_t reg);          /*  filled by driver */
 };
 
 struct atom_context {
@@ -143,21 +146,21 @@ struct atom_context {
 	int io_mode;
 	uint32_t *scratch;
 	int scratch_size_bytes;
-	char vbios_version[20];
 
 	uint8_t name[STRLEN_LONG];
 	uint8_t vbios_pn[STRLEN_LONG];
 	uint32_t version;
 	uint8_t vbios_ver_str[STRLEN_NORMAL];
 	uint8_t date[STRLEN_NORMAL];
+	uint8_t build_num[STRLEN_NORMAL];
 };
 
 extern int amdgpu_atom_debug;
 
-struct atom_context *amdgpu_atom_parse(struct card_info *, void *);
-int amdgpu_atom_execute_table(struct atom_context *, int, uint32_t *);
-int amdgpu_atom_asic_init(struct atom_context *);
-void amdgpu_atom_destroy(struct atom_context *);
+struct atom_context *amdgpu_atom_parse(struct card_info *card, void *bios);
+int amdgpu_atom_execute_table(struct atom_context *ctx, int index, uint32_t *params, int params_size);
+int amdgpu_atom_asic_init(struct atom_context *ctx);
+void amdgpu_atom_destroy(struct atom_context *ctx);
 bool amdgpu_atom_parse_data_header(struct atom_context *ctx, int index, uint16_t *size,
 			    uint8_t *frev, uint8_t *crev, uint16_t *data_start);
 bool amdgpu_atom_parse_cmd_header(struct atom_context *ctx, int index,
diff --git a/drivers/gpu/drm/amd/amdgpu/atombios_crtc.c b/drivers/gpu/drm/amd/amdgpu/atombios_crtc.c
index afad094f84c2..3dfc28840a7d 100644
--- a/drivers/gpu/drm/amd/amdgpu/atombios_crtc.c
+++ b/drivers/gpu/drm/amd/amdgpu/atombios_crtc.c
@@ -24,7 +24,6 @@
  *          Alex Deucher
  */
 
-#include <drm/drm_crtc_helper.h>
 #include <drm/amdgpu_drm.h>
 #include <drm/drm_fixed.h>
 #include "amdgpu.h"
@@ -78,7 +77,7 @@ void amdgpu_atombios_crtc_overscan_setup(struct drm_crtc *crtc,
 		args.usOverscanTop = cpu_to_le16(amdgpu_crtc->v_border);
 		break;
 	}
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 }
 
 void amdgpu_atombios_crtc_scaler_setup(struct drm_crtc *crtc)
@@ -107,7 +106,7 @@ void amdgpu_atombios_crtc_scaler_setup(struct drm_crtc *crtc)
 		args.ucEnable = ATOM_SCALER_DISABLE;
 		break;
 	}
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 }
 
 void amdgpu_atombios_crtc_lock(struct drm_crtc *crtc, int lock)
@@ -124,7 +123,7 @@ void amdgpu_atombios_crtc_lock(struct drm_crtc *crtc, int lock)
 	args.ucCRTC = amdgpu_crtc->crtc_id;
 	args.ucEnable = lock;
 
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 }
 
 void amdgpu_atombios_crtc_enable(struct drm_crtc *crtc, int state)
@@ -140,7 +139,7 @@ void amdgpu_atombios_crtc_enable(struct drm_crtc *crtc, int state)
 	args.ucCRTC = amdgpu_crtc->crtc_id;
 	args.ucEnable = state;
 
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 }
 
 void amdgpu_atombios_crtc_blank(struct drm_crtc *crtc, int state)
@@ -156,7 +155,7 @@ void amdgpu_atombios_crtc_blank(struct drm_crtc *crtc, int state)
 	args.ucCRTC = amdgpu_crtc->crtc_id;
 	args.ucBlanking = state;
 
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 }
 
 void amdgpu_atombios_crtc_powergate(struct drm_crtc *crtc, int state)
@@ -172,7 +171,7 @@ void amdgpu_atombios_crtc_powergate(struct drm_crtc *crtc, int state)
 	args.ucDispPipeId = amdgpu_crtc->crtc_id;
 	args.ucEnable = state;
 
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 }
 
 void amdgpu_atombios_crtc_powergate_init(struct amdgpu_device *adev)
@@ -184,7 +183,7 @@ void amdgpu_atombios_crtc_powergate_init(struct amdgpu_device *adev)
 
 	args.ucEnable = ATOM_INIT;
 
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 }
 
 void amdgpu_atombios_crtc_set_dtd_timing(struct drm_crtc *crtc,
@@ -229,7 +228,7 @@ void amdgpu_atombios_crtc_set_dtd_timing(struct drm_crtc *crtc,
 	args.susModeMiscInfo.usAccess = cpu_to_le16(misc);
 	args.ucCRTC = amdgpu_crtc->crtc_id;
 
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 }
 
 union atom_enable_ss {
@@ -294,7 +293,7 @@ static void amdgpu_atombios_crtc_program_ss(struct amdgpu_device *adev,
 	args.v3.usSpreadSpectrumStep = cpu_to_le16(ss->step);
 	args.v3.ucEnable = enable;
 
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 }
 
 union adjust_pixel_clock {
@@ -396,7 +395,7 @@ static u32 amdgpu_atombios_crtc_adjust_pll(struct drm_crtc *crtc,
 					ADJUST_DISPLAY_CONFIG_SS_ENABLE;
 
 			amdgpu_atom_execute_table(adev->mode_info.atom_context,
-					   index, (uint32_t *)&args);
+					   index, (uint32_t *)&args, sizeof(args));
 			adjusted_clock = le16_to_cpu(args.v1.usPixelClock) * 10;
 			break;
 		case 3:
@@ -429,7 +428,7 @@ static u32 amdgpu_atombios_crtc_adjust_pll(struct drm_crtc *crtc,
 				args.v3.sInput.ucExtTransmitterID = 0;
 
 			amdgpu_atom_execute_table(adev->mode_info.atom_context,
-					   index, (uint32_t *)&args);
+					   index, (uint32_t *)&args, sizeof(args));
 			adjusted_clock = le32_to_cpu(args.v3.sOutput.ulDispPllFreq) * 10;
 			if (args.v3.sOutput.ucRefDiv) {
 				amdgpu_crtc->pll_flags |= AMDGPU_PLL_USE_FRAC_FB_DIV;
@@ -515,7 +514,7 @@ void amdgpu_atombios_crtc_set_disp_eng_pll(struct amdgpu_device *adev,
 		DRM_ERROR("Unknown table version %d %d\n", frev, crev);
 		return;
 	}
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 }
 
 union set_dce_clock {
@@ -545,7 +544,7 @@ u32 amdgpu_atombios_crtc_set_dce_clock(struct amdgpu_device *adev,
 			args.v2_1.asParam.ulDCEClkFreq = cpu_to_le32(freq); /* 10kHz units */
 			args.v2_1.asParam.ucDCEClkType = clk_type;
 			args.v2_1.asParam.ucDCEClkSrc = clk_src;
-			amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+			amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 			ret_freq = le32_to_cpu(args.v2_1.asParam.ulDCEClkFreq) * 10;
 			break;
 		default:
@@ -741,7 +740,7 @@ void amdgpu_atombios_crtc_program_pll(struct drm_crtc *crtc,
 		return;
 	}
 
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 }
 
 int amdgpu_atombios_crtc_prepare_pll(struct drm_crtc *crtc,
diff --git a/drivers/gpu/drm/amd/amdgpu/atombios_dp.c b/drivers/gpu/drm/amd/amdgpu/atombios_dp.c
index f327becb022f..492813ab1b54 100644
--- a/drivers/gpu/drm/amd/amdgpu/atombios_dp.c
+++ b/drivers/gpu/drm/amd/amdgpu/atombios_dp.c
@@ -26,6 +26,8 @@
  */
 
 #include <drm/amdgpu_drm.h>
+#include <drm/display/drm_dp_helper.h>
+
 #include "amdgpu.h"
 
 #include "atom.h"
@@ -34,7 +36,6 @@
 #include "atombios_dp.h"
 #include "amdgpu_connectors.h"
 #include "amdgpu_atombios.h"
-#include <drm/drm_dp_helper.h>
 
 /* move these to drm_dp_helper.c/h */
 #define DP_LINK_CONFIGURATION_SIZE 9
@@ -82,7 +83,7 @@ static int amdgpu_atombios_dp_process_aux_ch(struct amdgpu_i2c_chan *chan,
 	args.v2.ucDelay = delay / 10;
 	args.v2.ucHPD_ID = chan->rec.hpd;
 
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 
 	*ack = args.v2.ucReplyStatus;
 
@@ -300,7 +301,7 @@ static u8 amdgpu_atombios_dp_encoder_service(struct amdgpu_device *adev,
 	args.ucLaneNum = lane_num;
 	args.ucStatus = 0;
 
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 	return args.ucStatus;
 }
 
@@ -429,7 +430,7 @@ void amdgpu_atombios_dp_set_link_config(struct drm_connector *connector,
 }
 
 int amdgpu_atombios_dp_mode_valid_helper(struct drm_connector *connector,
-				  struct drm_display_mode *mode)
+				  const struct drm_display_mode *mode)
 {
 	struct amdgpu_connector *amdgpu_connector = to_amdgpu_connector(connector);
 	struct amdgpu_connector_atom_dig *dig_connector;
@@ -457,8 +458,8 @@ bool amdgpu_atombios_dp_needs_link_train(struct amdgpu_connector *amdgpu_connect
 	u8 link_status[DP_LINK_STATUS_SIZE];
 	struct amdgpu_connector_atom_dig *dig = amdgpu_connector->con_priv;
 
-	if (drm_dp_dpcd_read_link_status(&amdgpu_connector->ddc_bus->aux, link_status)
-	    <= 0)
+	if (drm_dp_dpcd_read_link_status(&amdgpu_connector->ddc_bus->aux,
+					 link_status) < 0)
 		return false;
 	if (drm_dp_channel_eq_ok(link_status, dig->dp_lane_count))
 		return false;
@@ -615,7 +616,7 @@ amdgpu_atombios_dp_link_train_cr(struct amdgpu_atombios_dp_link_train_info *dp_i
 		drm_dp_link_train_clock_recovery_delay(dp_info->aux, dp_info->dpcd);
 
 		if (drm_dp_dpcd_read_link_status(dp_info->aux,
-						 dp_info->link_status) <= 0) {
+						 dp_info->link_status) < 0) {
 			DRM_ERROR("displayport link status failed\n");
 			break;
 		}
@@ -680,7 +681,7 @@ amdgpu_atombios_dp_link_train_ce(struct amdgpu_atombios_dp_link_train_info *dp_i
 		drm_dp_link_train_channel_eq_delay(dp_info->aux, dp_info->dpcd);
 
 		if (drm_dp_dpcd_read_link_status(dp_info->aux,
-						 dp_info->link_status) <= 0) {
+						 dp_info->link_status) < 0) {
 			DRM_ERROR("displayport link status failed\n");
 			break;
 		}
diff --git a/drivers/gpu/drm/amd/amdgpu/atombios_dp.h b/drivers/gpu/drm/amd/amdgpu/atombios_dp.h
index f59d85eaddf0..3e24acf8133f 100644
--- a/drivers/gpu/drm/amd/amdgpu/atombios_dp.h
+++ b/drivers/gpu/drm/amd/amdgpu/atombios_dp.h
@@ -32,7 +32,7 @@ int amdgpu_atombios_dp_get_panel_mode(struct drm_encoder *encoder,
 void amdgpu_atombios_dp_set_link_config(struct drm_connector *connector,
 				 const struct drm_display_mode *mode);
 int amdgpu_atombios_dp_mode_valid_helper(struct drm_connector *connector,
-				  struct drm_display_mode *mode);
+				  const struct drm_display_mode *mode);
 bool amdgpu_atombios_dp_needs_link_train(struct amdgpu_connector *amdgpu_connector);
 void amdgpu_atombios_dp_set_rx_power_state(struct drm_connector *connector,
 				    u8 power_state);
diff --git a/drivers/gpu/drm/amd/amdgpu/atombios_encoders.c b/drivers/gpu/drm/amd/amdgpu/atombios_encoders.c
index 6134ed964027..a51f3414b65d 100644
--- a/drivers/gpu/drm/amd/amdgpu/atombios_encoders.c
+++ b/drivers/gpu/drm/amd/amdgpu/atombios_encoders.c
@@ -26,7 +26,9 @@
 
 #include <linux/pci.h>
 
-#include <drm/drm_crtc_helper.h>
+#include <acpi/video.h>
+
+#include <drm/drm_edid.h>
 #include <drm/amdgpu_drm.h>
 #include "amdgpu.h"
 #include "amdgpu_connectors.h"
@@ -118,8 +120,6 @@ amdgpu_atombios_encoder_set_backlight_level(struct amdgpu_encoder *amdgpu_encode
 	}
 }
 
-#if defined(CONFIG_BACKLIGHT_CLASS_DEVICE) || defined(CONFIG_BACKLIGHT_CLASS_DEVICE_MODULE)
-
 static u8 amdgpu_atombios_encoder_backlight_level(struct backlight_device *bd)
 {
 	u8 level;
@@ -184,7 +184,12 @@ void amdgpu_atombios_encoder_init_backlight(struct amdgpu_encoder *amdgpu_encode
 		return;
 
 	if (!(adev->mode_info.firmware_flags & ATOM_BIOS_INFO_BL_CONTROLLED_BY_GPU))
-		return;
+		goto register_acpi_backlight;
+
+	if (!acpi_video_backlight_use_native()) {
+		drm_info(dev, "Skipping amdgpu atom DIG backlight registration\n");
+		goto register_acpi_backlight;
+	}
 
 	pdata = kmalloc(sizeof(struct amdgpu_backlight_privdata), GFP_KERNEL);
 	if (!pdata) {
@@ -210,7 +215,7 @@ void amdgpu_atombios_encoder_init_backlight(struct amdgpu_encoder *amdgpu_encode
 	dig->bl_dev = bd;
 
 	bd->props.brightness = amdgpu_atombios_encoder_get_backlight_brightness(bd);
-	bd->props.power = FB_BLANK_UNBLANK;
+	bd->props.power = BACKLIGHT_POWER_ON;
 	backlight_update_status(bd);
 
 	DRM_INFO("amdgpu atom DIG backlight initialized\n");
@@ -220,6 +225,10 @@ void amdgpu_atombios_encoder_init_backlight(struct amdgpu_encoder *amdgpu_encode
 error:
 	kfree(pdata);
 	return;
+
+register_acpi_backlight:
+	/* Try registering an ACPI video backlight device instead. */
+	acpi_video_register_backlight();
 }
 
 void
@@ -251,18 +260,6 @@ amdgpu_atombios_encoder_fini_backlight(struct amdgpu_encoder *amdgpu_encoder)
 	}
 }
 
-#else /* !CONFIG_BACKLIGHT_CLASS_DEVICE */
-
-void amdgpu_atombios_encoder_init_backlight(struct amdgpu_encoder *encoder)
-{
-}
-
-void amdgpu_atombios_encoder_fini_backlight(struct amdgpu_encoder *encoder)
-{
-}
-
-#endif
-
 bool amdgpu_atombios_encoder_is_digital(struct drm_encoder *encoder)
 {
 	struct amdgpu_encoder *amdgpu_encoder = to_amdgpu_encoder(encoder);
@@ -338,7 +335,7 @@ amdgpu_atombios_encoder_setup_dac(struct drm_encoder *encoder, int action)
 	args.ucDacStandard = ATOM_DAC1_PS2;
 	args.usPixelClock = cpu_to_le16(amdgpu_encoder->pixel_clock / 10);
 
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 
 }
 
@@ -435,7 +432,7 @@ amdgpu_atombios_encoder_setup_dvo(struct drm_encoder *encoder, int action)
 		break;
 	}
 
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 }
 
 int amdgpu_atombios_encoder_get_encoder_mode(struct drm_encoder *encoder)
@@ -469,7 +466,7 @@ int amdgpu_atombios_encoder_get_encoder_mode(struct drm_encoder *encoder)
 			if (amdgpu_connector->use_digital &&
 			    (amdgpu_connector->audio == AMDGPU_AUDIO_ENABLE))
 				return ATOM_ENCODER_MODE_HDMI;
-			else if (drm_detect_hdmi_monitor(amdgpu_connector_edid(connector)) &&
+			else if (connector->display_info.is_hdmi &&
 				 (amdgpu_connector->audio == AMDGPU_AUDIO_AUTO))
 				return ATOM_ENCODER_MODE_HDMI;
 			else if (amdgpu_connector->use_digital)
@@ -488,7 +485,7 @@ int amdgpu_atombios_encoder_get_encoder_mode(struct drm_encoder *encoder)
 		if (amdgpu_audio != 0) {
 			if (amdgpu_connector->audio == AMDGPU_AUDIO_ENABLE)
 				return ATOM_ENCODER_MODE_HDMI;
-			else if (drm_detect_hdmi_monitor(amdgpu_connector_edid(connector)) &&
+			else if (connector->display_info.is_hdmi &&
 				 (amdgpu_connector->audio == AMDGPU_AUDIO_AUTO))
 				return ATOM_ENCODER_MODE_HDMI;
 			else
@@ -506,7 +503,7 @@ int amdgpu_atombios_encoder_get_encoder_mode(struct drm_encoder *encoder)
 		} else if (amdgpu_audio != 0) {
 			if (amdgpu_connector->audio == AMDGPU_AUDIO_ENABLE)
 				return ATOM_ENCODER_MODE_HDMI;
-			else if (drm_detect_hdmi_monitor(amdgpu_connector_edid(connector)) &&
+			else if (connector->display_info.is_hdmi &&
 				 (amdgpu_connector->audio == AMDGPU_AUDIO_AUTO))
 				return ATOM_ENCODER_MODE_HDMI;
 			else
@@ -735,7 +732,7 @@ amdgpu_atombios_encoder_setup_dig_encoder(struct drm_encoder *encoder,
 		break;
 	}
 
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 
 }
 
@@ -765,7 +762,6 @@ amdgpu_atombios_encoder_setup_dig_transmitter(struct drm_encoder *encoder, int a
 	int dp_clock = 0;
 	int dp_lane_count = 0;
 	int connector_object_id = 0;
-	int igp_lane_info = 0;
 	int dig_encoder = dig->dig_encoder;
 	int hpd_id = AMDGPU_HPD_NONE;
 
@@ -848,26 +844,6 @@ amdgpu_atombios_encoder_setup_dig_transmitter(struct drm_encoder *encoder, int a
 			else
 				args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_DIG1_ENCODER;
 
-			if ((adev->flags & AMD_IS_APU) &&
-			    (amdgpu_encoder->encoder_id == ENCODER_OBJECT_ID_INTERNAL_UNIPHY)) {
-				if (is_dp ||
-				    !amdgpu_dig_monitor_is_duallink(encoder, amdgpu_encoder->pixel_clock)) {
-					if (igp_lane_info & 0x1)
-						args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_LANE_0_3;
-					else if (igp_lane_info & 0x2)
-						args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_LANE_4_7;
-					else if (igp_lane_info & 0x4)
-						args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_LANE_8_11;
-					else if (igp_lane_info & 0x8)
-						args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_LANE_12_15;
-				} else {
-					if (igp_lane_info & 0x3)
-						args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_LANE_0_7;
-					else if (igp_lane_info & 0xc)
-						args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_LANE_8_15;
-				}
-			}
-
 			if (dig->linkb)
 				args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_LINKB;
 			else
@@ -1160,7 +1136,7 @@ amdgpu_atombios_encoder_setup_dig_transmitter(struct drm_encoder *encoder, int a
 		break;
 	}
 
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 }
 
 bool
@@ -1188,7 +1164,7 @@ amdgpu_atombios_encoder_set_edp_panel_power(struct drm_connector *connector,
 
 	args.v1.ucAction = action;
 
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 
 	/* wait for the panel to power up */
 	if (action == ATOM_TRANSMITTER_ACTION_POWER_ON) {
@@ -1312,7 +1288,7 @@ amdgpu_atombios_encoder_setup_external_encoder(struct drm_encoder *encoder,
 		DRM_ERROR("Unknown table version: %d, %d\n", frev, crev);
 		return;
 	}
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 }
 
 static void
@@ -1657,7 +1633,7 @@ amdgpu_atombios_encoder_set_crtc_source(struct drm_encoder *encoder)
 		return;
 	}
 
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 }
 
 /* This only needs to be called once at startup */
@@ -1730,7 +1706,7 @@ amdgpu_atombios_encoder_dac_load_detect(struct drm_encoder *encoder,
 				args.sDacload.ucMisc = DAC_LOAD_MISC_YPrPb;
 		}
 
-		amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+		amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 
 		return true;
 	} else
@@ -2088,24 +2064,25 @@ amdgpu_atombios_encoder_get_lcd_info(struct amdgpu_encoder *encoder)
 				case LCD_FAKE_EDID_PATCH_RECORD_TYPE:
 					fake_edid_record = (ATOM_FAKE_EDID_PATCH_RECORD *)record;
 					if (fake_edid_record->ucFakeEDIDLength) {
-						struct edid *edid;
-						int edid_size =
-							max((int)EDID_LENGTH, (int)fake_edid_record->ucFakeEDIDLength);
-						edid = kmalloc(edid_size, GFP_KERNEL);
-						if (edid) {
-							memcpy((u8 *)edid, (u8 *)&fake_edid_record->ucFakeEDIDString[0],
-							       fake_edid_record->ucFakeEDIDLength);
-
-							if (drm_edid_is_valid(edid)) {
-								adev->mode_info.bios_hardcoded_edid = edid;
-								adev->mode_info.bios_hardcoded_edid_size = edid_size;
-							} else
-								kfree(edid);
-						}
+						const struct drm_edid *edid;
+						int edid_size;
+
+						if (fake_edid_record->ucFakeEDIDLength == 128)
+							edid_size = fake_edid_record->ucFakeEDIDLength;
+						else
+							edid_size = fake_edid_record->ucFakeEDIDLength * 128;
+						edid = drm_edid_alloc(fake_edid_record->ucFakeEDIDString, edid_size);
+						if (drm_edid_valid(edid))
+							adev->mode_info.bios_hardcoded_edid = edid;
+						else
+							drm_edid_free(edid);
+						record += struct_size(fake_edid_record,
+								      ucFakeEDIDString,
+								      edid_size);
+					} else {
+						/* empty fake edid record must be 3 bytes long */
+						record += sizeof(ATOM_FAKE_EDID_PATCH_RECORD) + 1;
 					}
-					record += fake_edid_record->ucFakeEDIDLength ?
-						fake_edid_record->ucFakeEDIDLength + 2 :
-						sizeof(ATOM_FAKE_EDID_PATCH_RECORD);
 					break;
 				case LCD_PANEL_RESOLUTION_RECORD_TYPE:
 					panel_res_record = (ATOM_PANEL_RESOLUTION_PATCH_RECORD *)record;
diff --git a/drivers/gpu/drm/amd/amdgpu/atombios_i2c.c b/drivers/gpu/drm/amd/amdgpu/atombios_i2c.c
index af0335535f82..a6501114322f 100644
--- a/drivers/gpu/drm/amd/amdgpu/atombios_i2c.c
+++ b/drivers/gpu/drm/amd/amdgpu/atombios_i2c.c
@@ -86,7 +86,7 @@ static int amdgpu_atombios_i2c_process_i2c_ch(struct amdgpu_i2c_chan *chan,
 	args.ucSlaveAddr = slave_addr << 1;
 	args.ucLineNumber = chan->rec.i2c_id;
 
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 
 	/* error */
 	if (args.ucStatus != HW_ASSISTED_I2C_STATUS_SUCCESS) {
@@ -172,5 +172,5 @@ void amdgpu_atombios_i2c_channel_trans(struct amdgpu_device *adev, u8 slave_addr
 	args.ucSlaveAddr = slave_addr;
 	args.ucLineNumber = line_number;
 
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/cik.c b/drivers/gpu/drm/amd/amdgpu/cik.c
index 54f28c075f21..9cd63b4177bf 100644
--- a/drivers/gpu/drm/amd/amdgpu/cik.c
+++ b/drivers/gpu/drm/amd/amdgpu/cik.c
@@ -1141,12 +1141,12 @@ static uint32_t cik_get_register_value(struct amdgpu_device *adev,
 
 		mutex_lock(&adev->grbm_idx_mutex);
 		if (se_num != 0xffffffff || sh_num != 0xffffffff)
-			amdgpu_gfx_select_se_sh(adev, se_num, sh_num, 0xffffffff);
+			amdgpu_gfx_select_se_sh(adev, se_num, sh_num, 0xffffffff, 0);
 
 		val = RREG32(reg_offset);
 
 		if (se_num != 0xffffffff || sh_num != 0xffffffff)
-			amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+			amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
 		mutex_unlock(&adev->grbm_idx_mutex);
 		return val;
 	} else {
@@ -1375,14 +1375,14 @@ static int cik_asic_pci_config_reset(struct amdgpu_device *adev)
 	return r;
 }
 
-static bool cik_asic_supports_baco(struct amdgpu_device *adev)
+static int cik_asic_supports_baco(struct amdgpu_device *adev)
 {
 	switch (adev->asic_type) {
 	case CHIP_BONAIRE:
 	case CHIP_HAWAII:
 		return amdgpu_dpm_is_baco_supported(adev);
 	default:
-		return false;
+		return 0;
 	}
 }
 
@@ -1428,6 +1428,10 @@ static int cik_asic_reset(struct amdgpu_device *adev)
 {
 	int r;
 
+	/* APUs don't have full asic reset */
+	if (adev->flags & AMD_IS_APU)
+		return 0;
+
 	if (cik_asic_reset_method(adev) == AMD_RESET_METHOD_BACO) {
 		dev_info(adev->dev, "BACO reset\n");
 		r = amdgpu_dpm_baco_reset(adev);
@@ -1570,17 +1574,8 @@ static void cik_pcie_gen3_enable(struct amdgpu_device *adev)
 			u16 bridge_cfg2, gpu_cfg2;
 			u32 max_lw, current_lw, tmp;
 
-			pcie_capability_read_word(root, PCI_EXP_LNKCTL,
-						  &bridge_cfg);
-			pcie_capability_read_word(adev->pdev, PCI_EXP_LNKCTL,
-						  &gpu_cfg);
-
-			tmp16 = bridge_cfg | PCI_EXP_LNKCTL_HAWD;
-			pcie_capability_write_word(root, PCI_EXP_LNKCTL, tmp16);
-
-			tmp16 = gpu_cfg | PCI_EXP_LNKCTL_HAWD;
-			pcie_capability_write_word(adev->pdev, PCI_EXP_LNKCTL,
-						   tmp16);
+			pcie_capability_set_word(root, PCI_EXP_LNKCTL, PCI_EXP_LNKCTL_HAWD);
+			pcie_capability_set_word(adev->pdev, PCI_EXP_LNKCTL, PCI_EXP_LNKCTL_HAWD);
 
 			tmp = RREG32_PCIE(ixPCIE_LC_STATUS1);
 			max_lw = (tmp & PCIE_LC_STATUS1__LC_DETECTED_LINK_WIDTH_MASK) >>
@@ -1633,45 +1628,28 @@ static void cik_pcie_gen3_enable(struct amdgpu_device *adev)
 				msleep(100);
 
 				/* linkctl */
-				pcie_capability_read_word(root, PCI_EXP_LNKCTL,
-							  &tmp16);
-				tmp16 &= ~PCI_EXP_LNKCTL_HAWD;
-				tmp16 |= (bridge_cfg & PCI_EXP_LNKCTL_HAWD);
-				pcie_capability_write_word(root, PCI_EXP_LNKCTL,
-							   tmp16);
-
-				pcie_capability_read_word(adev->pdev,
-							  PCI_EXP_LNKCTL,
-							  &tmp16);
-				tmp16 &= ~PCI_EXP_LNKCTL_HAWD;
-				tmp16 |= (gpu_cfg & PCI_EXP_LNKCTL_HAWD);
-				pcie_capability_write_word(adev->pdev,
-							   PCI_EXP_LNKCTL,
-							   tmp16);
+				pcie_capability_clear_and_set_word(root, PCI_EXP_LNKCTL,
+								   PCI_EXP_LNKCTL_HAWD,
+								   bridge_cfg &
+								   PCI_EXP_LNKCTL_HAWD);
+				pcie_capability_clear_and_set_word(adev->pdev, PCI_EXP_LNKCTL,
+								   PCI_EXP_LNKCTL_HAWD,
+								   gpu_cfg &
+								   PCI_EXP_LNKCTL_HAWD);
 
 				/* linkctl2 */
-				pcie_capability_read_word(root, PCI_EXP_LNKCTL2,
-							  &tmp16);
-				tmp16 &= ~(PCI_EXP_LNKCTL2_ENTER_COMP |
-					   PCI_EXP_LNKCTL2_TX_MARGIN);
-				tmp16 |= (bridge_cfg2 &
-					  (PCI_EXP_LNKCTL2_ENTER_COMP |
-					   PCI_EXP_LNKCTL2_TX_MARGIN));
-				pcie_capability_write_word(root,
-							   PCI_EXP_LNKCTL2,
-							   tmp16);
-
-				pcie_capability_read_word(adev->pdev,
-							  PCI_EXP_LNKCTL2,
-							  &tmp16);
-				tmp16 &= ~(PCI_EXP_LNKCTL2_ENTER_COMP |
-					   PCI_EXP_LNKCTL2_TX_MARGIN);
-				tmp16 |= (gpu_cfg2 &
-					  (PCI_EXP_LNKCTL2_ENTER_COMP |
-					   PCI_EXP_LNKCTL2_TX_MARGIN));
-				pcie_capability_write_word(adev->pdev,
-							   PCI_EXP_LNKCTL2,
-							   tmp16);
+				pcie_capability_clear_and_set_word(root, PCI_EXP_LNKCTL2,
+								   PCI_EXP_LNKCTL2_ENTER_COMP |
+								   PCI_EXP_LNKCTL2_TX_MARGIN,
+								   bridge_cfg2 &
+								   (PCI_EXP_LNKCTL2_ENTER_COMP |
+								    PCI_EXP_LNKCTL2_TX_MARGIN));
+				pcie_capability_clear_and_set_word(adev->pdev, PCI_EXP_LNKCTL2,
+								   PCI_EXP_LNKCTL2_ENTER_COMP |
+								   PCI_EXP_LNKCTL2_TX_MARGIN,
+								   gpu_cfg2 &
+								   (PCI_EXP_LNKCTL2_ENTER_COMP |
+								    PCI_EXP_LNKCTL2_TX_MARGIN));
 
 				tmp = RREG32_PCIE(ixPCIE_LC_CNTL4);
 				tmp &= ~PCIE_LC_CNTL4__LC_SET_QUIESCE_MASK;
@@ -1686,16 +1664,15 @@ static void cik_pcie_gen3_enable(struct amdgpu_device *adev)
 	speed_cntl &= ~PCIE_LC_SPEED_CNTL__LC_FORCE_DIS_SW_SPEED_CHANGE_MASK;
 	WREG32_PCIE(ixPCIE_LC_SPEED_CNTL, speed_cntl);
 
-	pcie_capability_read_word(adev->pdev, PCI_EXP_LNKCTL2, &tmp16);
-	tmp16 &= ~PCI_EXP_LNKCTL2_TLS;
-
+	tmp16 = 0;
 	if (adev->pm.pcie_gen_mask & CAIL_PCIE_LINK_SPEED_SUPPORT_GEN3)
 		tmp16 |= PCI_EXP_LNKCTL2_TLS_8_0GT; /* gen3 */
 	else if (adev->pm.pcie_gen_mask & CAIL_PCIE_LINK_SPEED_SUPPORT_GEN2)
 		tmp16 |= PCI_EXP_LNKCTL2_TLS_5_0GT; /* gen2 */
 	else
 		tmp16 |= PCI_EXP_LNKCTL2_TLS_2_5GT; /* gen1 */
-	pcie_capability_write_word(adev->pdev, PCI_EXP_LNKCTL2, tmp16);
+	pcie_capability_clear_and_set_word(adev->pdev, PCI_EXP_LNKCTL2,
+					   PCI_EXP_LNKCTL2_TLS, tmp16);
 
 	speed_cntl = RREG32_PCIE(ixPCIE_LC_SPEED_CNTL);
 	speed_cntl |= PCIE_LC_SPEED_CNTL__LC_INITIATE_LINK_SPEED_CHANGE_MASK;
@@ -1715,16 +1692,12 @@ static void cik_program_aspm(struct amdgpu_device *adev)
 	bool disable_l0s = false, disable_l1 = false, disable_plloff_in_l1 = false;
 	bool disable_clkreq = false;
 
-	if (amdgpu_aspm == 0)
+	if (!amdgpu_device_should_use_aspm(adev))
 		return;
 
 	if (pci_is_root_bus(adev->pdev->bus))
 		return;
 
-	/* XXX double check APUs */
-	if (adev->flags & AMD_IS_APU)
-		return;
-
 	orig = data = RREG32_PCIE(ixPCIE_LC_N_FTS_CNTL);
 	data &= ~PCIE_LC_N_FTS_CNTL__LC_XMIT_N_FTS_MASK;
 	data |= (0x24 << PCIE_LC_N_FTS_CNTL__LC_XMIT_N_FTS__SHIFT) |
@@ -2012,9 +1985,9 @@ static const struct amdgpu_asic_funcs cik_asic_funcs =
 	.query_video_codecs = &cik_query_video_codecs,
 };
 
-static int cik_common_early_init(void *handle)
+static int cik_common_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	adev->smc_rreg = &cik_smc_rreg;
 	adev->smc_wreg = &cik_smc_wreg;
@@ -2151,19 +2124,9 @@ static int cik_common_early_init(void *handle)
 	return 0;
 }
 
-static int cik_common_sw_init(void *handle)
-{
-	return 0;
-}
-
-static int cik_common_sw_fini(void *handle)
-{
-	return 0;
-}
-
-static int cik_common_hw_init(void *handle)
+static int cik_common_hw_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	/* move the golden regs per IP block */
 	cik_init_golden_registers(adev);
@@ -2175,48 +2138,36 @@ static int cik_common_hw_init(void *handle)
 	return 0;
 }
 
-static int cik_common_hw_fini(void *handle)
+static int cik_common_hw_fini(struct amdgpu_ip_block *ip_block)
 {
 	return 0;
 }
 
-static int cik_common_suspend(void *handle)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	return cik_common_hw_fini(adev);
-}
-
-static int cik_common_resume(void *handle)
+static int cik_common_resume(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	return cik_common_hw_init(adev);
+	return cik_common_hw_init(ip_block);
 }
 
-static bool cik_common_is_idle(void *handle)
+static bool cik_common_is_idle(struct amdgpu_ip_block *ip_block)
 {
 	return true;
 }
 
-static int cik_common_wait_for_idle(void *handle)
-{
-	return 0;
-}
 
-static int cik_common_soft_reset(void *handle)
+
+static int cik_common_soft_reset(struct amdgpu_ip_block *ip_block)
 {
 	/* XXX hard reset?? */
 	return 0;
 }
 
-static int cik_common_set_clockgating_state(void *handle,
+static int cik_common_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					    enum amd_clockgating_state state)
 {
 	return 0;
 }
 
-static int cik_common_set_powergating_state(void *handle,
+static int cik_common_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					    enum amd_powergating_state state)
 {
 	return 0;
@@ -2225,15 +2176,10 @@ static int cik_common_set_powergating_state(void *handle,
 static const struct amd_ip_funcs cik_common_ip_funcs = {
 	.name = "cik_common",
 	.early_init = cik_common_early_init,
-	.late_init = NULL,
-	.sw_init = cik_common_sw_init,
-	.sw_fini = cik_common_sw_fini,
 	.hw_init = cik_common_hw_init,
 	.hw_fini = cik_common_hw_fini,
-	.suspend = cik_common_suspend,
 	.resume = cik_common_resume,
 	.is_idle = cik_common_is_idle,
-	.wait_for_idle = cik_common_wait_for_idle,
 	.soft_reset = cik_common_soft_reset,
 	.set_clockgating_state = cik_common_set_clockgating_state,
 	.set_powergating_state = cik_common_set_powergating_state,
diff --git a/drivers/gpu/drm/amd/amdgpu/cik_ih.c b/drivers/gpu/drm/amd/amdgpu/cik_ih.c
index df385ffc9768..876a3256dba4 100644
--- a/drivers/gpu/drm/amd/amdgpu/cik_ih.c
+++ b/drivers/gpu/drm/amd/amdgpu/cik_ih.c
@@ -156,6 +156,9 @@ static int cik_ih_irq_init(struct amdgpu_device *adev)
 	/* enable irqs */
 	cik_ih_enable_interrupts(adev);
 
+	if (adev->irq.ih_soft.ring_size)
+		adev->irq.ih_soft.enabled = true;
+
 	return 0;
 }
 
@@ -192,6 +195,9 @@ static u32 cik_ih_get_wptr(struct amdgpu_device *adev,
 
 	wptr = le32_to_cpu(*ih->wptr_cpu);
 
+	if (ih == &adev->irq.ih_soft)
+		goto out;
+
 	if (wptr & IH_RB_WPTR__RB_OVERFLOW_MASK) {
 		wptr &= ~IH_RB_WPTR__RB_OVERFLOW_MASK;
 		/* When a ring buffer overflow happen start parsing interrupt
@@ -204,7 +210,15 @@ static u32 cik_ih_get_wptr(struct amdgpu_device *adev,
 		tmp = RREG32(mmIH_RB_CNTL);
 		tmp |= IH_RB_CNTL__WPTR_OVERFLOW_CLEAR_MASK;
 		WREG32(mmIH_RB_CNTL, tmp);
+
+		/* Unset the CLEAR_OVERFLOW bit immediately so new overflows
+		 * can be detected.
+		 */
+		tmp &= ~IH_RB_CNTL__WPTR_OVERFLOW_CLEAR_MASK;
+		WREG32(mmIH_RB_CNTL, tmp);
 	}
+
+out:
 	return (wptr & ih->ptr_mask);
 }
 
@@ -277,9 +291,9 @@ static void cik_ih_set_rptr(struct amdgpu_device *adev,
 	WREG32(mmIH_RB_RPTR, ih->rptr);
 }
 
-static int cik_ih_early_init(void *handle)
+static int cik_ih_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int ret;
 
 	ret = amdgpu_irq_add_domain(adev);
@@ -291,23 +305,27 @@ static int cik_ih_early_init(void *handle)
 	return 0;
 }
 
-static int cik_ih_sw_init(void *handle)
+static int cik_ih_sw_init(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	r = amdgpu_ih_ring_init(adev, &adev->irq.ih, 64 * 1024, false);
 	if (r)
 		return r;
 
+	r = amdgpu_ih_ring_init(adev, &adev->irq.ih_soft, IH_SW_RING_SIZE, true);
+	if (r)
+		return r;
+
 	r = amdgpu_irq_init(adev);
 
 	return r;
 }
 
-static int cik_ih_sw_fini(void *handle)
+static int cik_ih_sw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	amdgpu_irq_fini_sw(adev);
 	amdgpu_irq_remove_domain(adev);
@@ -315,39 +333,33 @@ static int cik_ih_sw_fini(void *handle)
 	return 0;
 }
 
-static int cik_ih_hw_init(void *handle)
+static int cik_ih_hw_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	return cik_ih_irq_init(adev);
 }
 
-static int cik_ih_hw_fini(void *handle)
+static int cik_ih_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	cik_ih_irq_disable(adev);
+	cik_ih_irq_disable(ip_block->adev);
 
 	return 0;
 }
 
-static int cik_ih_suspend(void *handle)
+static int cik_ih_suspend(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	return cik_ih_hw_fini(adev);
+	return cik_ih_hw_fini(ip_block);
 }
 
-static int cik_ih_resume(void *handle)
+static int cik_ih_resume(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	return cik_ih_hw_init(adev);
+	return cik_ih_hw_init(ip_block);
 }
 
-static bool cik_ih_is_idle(void *handle)
+static bool cik_ih_is_idle(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	u32 tmp = RREG32(mmSRBM_STATUS);
 
 	if (tmp & SRBM_STATUS__IH_BUSY_MASK)
@@ -356,11 +368,11 @@ static bool cik_ih_is_idle(void *handle)
 	return true;
 }
 
-static int cik_ih_wait_for_idle(void *handle)
+static int cik_ih_wait_for_idle(struct amdgpu_ip_block *ip_block)
 {
 	unsigned i;
 	u32 tmp;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	for (i = 0; i < adev->usec_timeout; i++) {
 		/* read MC_STATUS */
@@ -372,9 +384,9 @@ static int cik_ih_wait_for_idle(void *handle)
 	return -ETIMEDOUT;
 }
 
-static int cik_ih_soft_reset(void *handle)
+static int cik_ih_soft_reset(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	u32 srbm_soft_reset = 0;
 	u32 tmp = RREG32(mmSRBM_STATUS);
@@ -402,13 +414,13 @@ static int cik_ih_soft_reset(void *handle)
 	return 0;
 }
 
-static int cik_ih_set_clockgating_state(void *handle,
+static int cik_ih_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_clockgating_state state)
 {
 	return 0;
 }
 
-static int cik_ih_set_powergating_state(void *handle,
+static int cik_ih_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_powergating_state state)
 {
 	return 0;
@@ -417,7 +429,6 @@ static int cik_ih_set_powergating_state(void *handle,
 static const struct amd_ip_funcs cik_ih_ip_funcs = {
 	.name = "cik_ih",
 	.early_init = cik_ih_early_init,
-	.late_init = NULL,
 	.sw_init = cik_ih_sw_init,
 	.sw_fini = cik_ih_sw_fini,
 	.hw_init = cik_ih_hw_init,
@@ -442,8 +453,7 @@ static void cik_ih_set_interrupt_funcs(struct amdgpu_device *adev)
 	adev->irq.ih_funcs = &cik_ih_funcs;
 }
 
-const struct amdgpu_ip_block_version cik_ih_ip_block =
-{
+const struct amdgpu_ip_block_version cik_ih_ip_block = {
 	.type = AMD_IP_BLOCK_TYPE_IH,
 	.major = 2,
 	.minor = 0,
diff --git a/drivers/gpu/drm/amd/amdgpu/cik_sdma.c b/drivers/gpu/drm/amd/amdgpu/cik_sdma.c
index c8ebd108548d..9e8715b4739d 100644
--- a/drivers/gpu/drm/amd/amdgpu/cik_sdma.c
+++ b/drivers/gpu/drm/amd/amdgpu/cik_sdma.c
@@ -54,7 +54,9 @@ static void cik_sdma_set_ring_funcs(struct amdgpu_device *adev);
 static void cik_sdma_set_irq_funcs(struct amdgpu_device *adev);
 static void cik_sdma_set_buffer_funcs(struct amdgpu_device *adev);
 static void cik_sdma_set_vm_pte_funcs(struct amdgpu_device *adev);
-static int cik_sdma_soft_reset(void *handle);
+static int cik_sdma_soft_reset(struct amdgpu_ip_block *ip_block);
+
+u32 amdgpu_cik_gpu_check_soft_reset(struct amdgpu_device *adev);
 
 MODULE_FIRMWARE("amdgpu/bonaire_sdma.bin");
 MODULE_FIRMWARE("amdgpu/bonaire_sdma1.bin");
@@ -67,16 +69,12 @@ MODULE_FIRMWARE("amdgpu/kabini_sdma1.bin");
 MODULE_FIRMWARE("amdgpu/mullins_sdma.bin");
 MODULE_FIRMWARE("amdgpu/mullins_sdma1.bin");
 
-u32 amdgpu_cik_gpu_check_soft_reset(struct amdgpu_device *adev);
-
-
 static void cik_sdma_free_microcode(struct amdgpu_device *adev)
 {
 	int i;
-	for (i = 0; i < adev->sdma.num_instances; i++) {
-			release_firmware(adev->sdma.instance[i].fw);
-			adev->sdma.instance[i].fw = NULL;
-	}
+
+	for (i = 0; i < adev->sdma.num_instances; i++)
+		amdgpu_ucode_release(&adev->sdma.instance[i].fw);
 }
 
 /*
@@ -108,7 +106,6 @@ static void cik_sdma_free_microcode(struct amdgpu_device *adev)
 static int cik_sdma_init_microcode(struct amdgpu_device *adev)
 {
 	const char *chip_name;
-	char fw_name[30];
 	int err = 0, i;
 
 	DRM_DEBUG("\n");
@@ -134,21 +131,22 @@ static int cik_sdma_init_microcode(struct amdgpu_device *adev)
 
 	for (i = 0; i < adev->sdma.num_instances; i++) {
 		if (i == 0)
-			snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_sdma.bin", chip_name);
+			err = amdgpu_ucode_request(adev, &adev->sdma.instance[i].fw,
+						   AMDGPU_UCODE_REQUIRED,
+						   "amdgpu/%s_sdma.bin", chip_name);
 		else
-			snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_sdma1.bin", chip_name);
-		err = request_firmware(&adev->sdma.instance[i].fw, fw_name, adev->dev);
+			err = amdgpu_ucode_request(adev, &adev->sdma.instance[i].fw,
+						   AMDGPU_UCODE_REQUIRED,
+						   "amdgpu/%s_sdma1.bin", chip_name);
 		if (err)
 			goto out;
-		err = amdgpu_ucode_validate(adev->sdma.instance[i].fw);
 	}
 out:
 	if (err) {
-		pr_err("cik_sdma: Failed to load firmware \"%s\"\n", fw_name);
-		for (i = 0; i < adev->sdma.num_instances; i++) {
-			release_firmware(adev->sdma.instance[i].fw);
-			adev->sdma.instance[i].fw = NULL;
-		}
+		pr_err("cik_sdma: Failed to load firmware \"%s_sdma%s.bin\"\n",
+		       chip_name, i == 0 ? "" : "1");
+		for (i = 0; i < adev->sdma.num_instances; i++)
+			amdgpu_ucode_release(&adev->sdma.instance[i].fw);
 	}
 	return err;
 }
@@ -164,7 +162,7 @@ static uint64_t cik_sdma_ring_get_rptr(struct amdgpu_ring *ring)
 {
 	u32 rptr;
 
-	rptr = ring->adev->wb.wb[ring->rptr_offs];
+	rptr = *ring->rptr_cpu_addr;
 
 	return (rptr & 0x3fffc) >> 2;
 }
@@ -195,7 +193,7 @@ static void cik_sdma_ring_set_wptr(struct amdgpu_ring *ring)
 	struct amdgpu_device *adev = ring->adev;
 
 	WREG32(mmSDMA0_GFX_RB_WPTR + sdma_offsets[ring->me],
-	       (lower_32_bits(ring->wptr) << 2) & 0x3fffc);
+	       (ring->wptr << 2) & 0x3fffc);
 }
 
 static void cik_sdma_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count)
@@ -309,15 +307,9 @@ static void cik_sdma_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 seq
  */
 static void cik_sdma_gfx_stop(struct amdgpu_device *adev)
 {
-	struct amdgpu_ring *sdma0 = &adev->sdma.instance[0].ring;
-	struct amdgpu_ring *sdma1 = &adev->sdma.instance[1].ring;
 	u32 rb_cntl;
 	int i;
 
-	if ((adev->mman.buffer_funcs_ring == sdma0) ||
-	    (adev->mman.buffer_funcs_ring == sdma1))
-			amdgpu_ttm_set_buffer_funcs_status(adev, false);
-
 	for (i = 0; i < adev->sdma.num_instances; i++) {
 		rb_cntl = RREG32(mmSDMA0_GFX_RB_CNTL + sdma_offsets[i]);
 		rb_cntl &= ~SDMA0_GFX_RB_CNTL__RB_ENABLE_MASK;
@@ -436,12 +428,10 @@ static int cik_sdma_gfx_resume(struct amdgpu_device *adev)
 	struct amdgpu_ring *ring;
 	u32 rb_cntl, ib_cntl;
 	u32 rb_bufsz;
-	u32 wb_offset;
 	int i, j, r;
 
 	for (i = 0; i < adev->sdma.num_instances; i++) {
 		ring = &adev->sdma.instance[i].ring;
-		wb_offset = (ring->rptr_offs * 4);
 
 		mutex_lock(&adev->srbm_mutex);
 		for (j = 0; j < 16; j++) {
@@ -477,9 +467,9 @@ static int cik_sdma_gfx_resume(struct amdgpu_device *adev)
 
 		/* set the wb address whether it's enabled or not */
 		WREG32(mmSDMA0_GFX_RB_RPTR_ADDR_HI + sdma_offsets[i],
-		       upper_32_bits(adev->wb.gpu_addr + wb_offset) & 0xFFFFFFFF);
+		       upper_32_bits(ring->rptr_gpu_addr) & 0xFFFFFFFF);
 		WREG32(mmSDMA0_GFX_RB_RPTR_ADDR_LO + sdma_offsets[i],
-		       ((adev->wb.gpu_addr + wb_offset) & 0xFFFFFFFC));
+		       ((ring->rptr_gpu_addr) & 0xFFFFFFFC));
 
 		rb_cntl |= SDMA0_GFX_RB_CNTL__RPTR_WRITEBACK_ENABLE_MASK;
 
@@ -487,7 +477,7 @@ static int cik_sdma_gfx_resume(struct amdgpu_device *adev)
 		WREG32(mmSDMA0_GFX_RB_BASE_HI + sdma_offsets[i], ring->gpu_addr >> 40);
 
 		ring->wptr = 0;
-		WREG32(mmSDMA0_GFX_RB_WPTR + sdma_offsets[i], lower_32_bits(ring->wptr) << 2);
+		WREG32(mmSDMA0_GFX_RB_WPTR + sdma_offsets[i], ring->wptr << 2);
 
 		/* enable DMA RB */
 		WREG32(mmSDMA0_GFX_RB_CNTL + sdma_offsets[i],
@@ -499,8 +489,6 @@ static int cik_sdma_gfx_resume(struct amdgpu_device *adev)
 #endif
 		/* enable DMA IBs */
 		WREG32(mmSDMA0_GFX_IB_CNTL + sdma_offsets[i], ib_cntl);
-
-		ring->sched.ready = true;
 	}
 
 	cik_sdma_enable(adev, true);
@@ -510,9 +498,6 @@ static int cik_sdma_gfx_resume(struct amdgpu_device *adev)
 		r = amdgpu_ring_test_helper(ring);
 		if (r)
 			return r;
-
-		if (adev->mman.buffer_funcs_ring == ring)
-			amdgpu_ttm_set_buffer_funcs_status(adev, true);
 	}
 
 	return 0;
@@ -712,7 +697,7 @@ static int cik_sdma_ring_test_ib(struct amdgpu_ring *ring, long timeout)
 		r = -EINVAL;
 
 err1:
-	amdgpu_ib_free(adev, &ib, NULL);
+	amdgpu_ib_free(&ib, NULL);
 	dma_fence_put(f);
 err0:
 	amdgpu_device_wb_free(adev, index);
@@ -934,12 +919,17 @@ static void cik_enable_sdma_mgls(struct amdgpu_device *adev,
 	}
 }
 
-static int cik_sdma_early_init(void *handle)
+static int cik_sdma_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
+	int r;
 
 	adev->sdma.num_instances = SDMA_MAX_INSTANCE;
 
+	r = cik_sdma_init_microcode(adev);
+	if (r)
+		return r;
+
 	cik_sdma_set_ring_funcs(adev);
 	cik_sdma_set_irq_funcs(adev);
 	cik_sdma_set_buffer_funcs(adev);
@@ -948,18 +938,12 @@ static int cik_sdma_early_init(void *handle)
 	return 0;
 }
 
-static int cik_sdma_sw_init(void *handle)
+static int cik_sdma_sw_init(struct amdgpu_ip_block *ip_block)
 {
 	struct amdgpu_ring *ring;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int r, i;
 
-	r = cik_sdma_init_microcode(adev);
-	if (r) {
-		DRM_ERROR("Failed to load sdma firmware!\n");
-		return r;
-	}
-
 	/* SDMA trap event */
 	r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 224,
 			      &adev->sdma.trap_irq);
@@ -994,9 +978,9 @@ static int cik_sdma_sw_init(void *handle)
 	return r;
 }
 
-static int cik_sdma_sw_fini(void *handle)
+static int cik_sdma_sw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int i;
 
 	for (i = 0; i < adev->sdma.num_instances; i++)
@@ -1006,21 +990,16 @@ static int cik_sdma_sw_fini(void *handle)
 	return 0;
 }
 
-static int cik_sdma_hw_init(void *handle)
+static int cik_sdma_hw_init(struct amdgpu_ip_block *ip_block)
 {
-	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
-	r = cik_sdma_start(adev);
-	if (r)
-		return r;
-
-	return r;
+	return cik_sdma_start(adev);
 }
 
-static int cik_sdma_hw_fini(void *handle)
+static int cik_sdma_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	cik_ctx_switch_enable(adev, false);
 	cik_sdma_enable(adev, false);
@@ -1028,25 +1007,21 @@ static int cik_sdma_hw_fini(void *handle)
 	return 0;
 }
 
-static int cik_sdma_suspend(void *handle)
+static int cik_sdma_suspend(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	return cik_sdma_hw_fini(adev);
+	return cik_sdma_hw_fini(ip_block);
 }
 
-static int cik_sdma_resume(void *handle)
+static int cik_sdma_resume(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	cik_sdma_soft_reset(ip_block);
 
-	cik_sdma_soft_reset(handle);
-
-	return cik_sdma_hw_init(adev);
+	return cik_sdma_hw_init(ip_block);
 }
 
-static bool cik_sdma_is_idle(void *handle)
+static bool cik_sdma_is_idle(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	u32 tmp = RREG32(mmSRBM_STATUS2);
 
 	if (tmp & (SRBM_STATUS2__SDMA_BUSY_MASK |
@@ -1056,27 +1031,23 @@ static bool cik_sdma_is_idle(void *handle)
 	return true;
 }
 
-static int cik_sdma_wait_for_idle(void *handle)
+static int cik_sdma_wait_for_idle(struct amdgpu_ip_block *ip_block)
 {
 	unsigned i;
-	u32 tmp;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	for (i = 0; i < adev->usec_timeout; i++) {
-		tmp = RREG32(mmSRBM_STATUS2) & (SRBM_STATUS2__SDMA_BUSY_MASK |
-				SRBM_STATUS2__SDMA1_BUSY_MASK);
-
-		if (!tmp)
+		if (cik_sdma_is_idle(ip_block))
 			return 0;
 		udelay(1);
 	}
 	return -ETIMEDOUT;
 }
 
-static int cik_sdma_soft_reset(void *handle)
+static int cik_sdma_soft_reset(struct amdgpu_ip_block *ip_block)
 {
 	u32 srbm_soft_reset = 0;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	u32 tmp;
 
 	/* sdma0 */
@@ -1210,11 +1181,11 @@ static int cik_sdma_process_illegal_inst_irq(struct amdgpu_device *adev,
 	return 0;
 }
 
-static int cik_sdma_set_clockgating_state(void *handle,
+static int cik_sdma_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_clockgating_state state)
 {
 	bool gate = false;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (state == AMD_CG_STATE_GATE)
 		gate = true;
@@ -1225,7 +1196,7 @@ static int cik_sdma_set_clockgating_state(void *handle,
 	return 0;
 }
 
-static int cik_sdma_set_powergating_state(void *handle,
+static int cik_sdma_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_powergating_state state)
 {
 	return 0;
@@ -1234,7 +1205,6 @@ static int cik_sdma_set_powergating_state(void *handle,
 static const struct amd_ip_funcs cik_sdma_ip_funcs = {
 	.name = "cik_sdma",
 	.early_init = cik_sdma_early_init,
-	.late_init = NULL,
 	.sw_init = cik_sdma_sw_init,
 	.sw_fini = cik_sdma_sw_fini,
 	.hw_init = cik_sdma_hw_init,
@@ -1308,7 +1278,7 @@ static void cik_sdma_set_irq_funcs(struct amdgpu_device *adev)
  * @src_offset: src GPU address
  * @dst_offset: dst GPU address
  * @byte_count: number of bytes to xfer
- * @tmz: is this a secure operation
+ * @copy_flags: unused
  *
  * Copy GPU buffers using the DMA engine (CIK).
  * Used by the amdgpu ttm implementation to move pages if
@@ -1318,7 +1288,7 @@ static void cik_sdma_emit_copy_buffer(struct amdgpu_ib *ib,
 				      uint64_t src_offset,
 				      uint64_t dst_offset,
 				      uint32_t byte_count,
-				      bool tmz)
+				      uint32_t copy_flags)
 {
 	ib->ptr[ib->length_dw++] = SDMA_PACKET(SDMA_OPCODE_COPY, SDMA_COPY_SUB_OPCODE_LINEAR, 0);
 	ib->ptr[ib->length_dw++] = byte_count;
diff --git a/drivers/gpu/drm/amd/amdgpu/cikd.h b/drivers/gpu/drm/amd/amdgpu/cikd.h
index 55982c0064b5..8aca4f2734f2 100644
--- a/drivers/gpu/drm/amd/amdgpu/cikd.h
+++ b/drivers/gpu/drm/amd/amdgpu/cikd.h
@@ -51,8 +51,14 @@
 #define HPD4_REGISTER_OFFSET                 (0x1813 - 0x1807)
 #define HPD5_REGISTER_OFFSET                 (0x1816 - 0x1807)
 
-#define BONAIRE_GB_ADDR_CONFIG_GOLDEN        0x12010001
-#define HAWAII_GB_ADDR_CONFIG_GOLDEN         0x12011003
+/* audio endpt instance offsets */
+#define AUD0_REGISTER_OFFSET                 (0x1780 - 0x1780)
+#define AUD1_REGISTER_OFFSET                 (0x1786 - 0x1780)
+#define AUD2_REGISTER_OFFSET                 (0x178c - 0x1780)
+#define AUD3_REGISTER_OFFSET                 (0x1792 - 0x1780)
+#define AUD4_REGISTER_OFFSET                 (0x1798 - 0x1780)
+#define AUD5_REGISTER_OFFSET                 (0x179d - 0x1780)
+#define AUD6_REGISTER_OFFSET                 (0x17a4 - 0x1780)
 
 #define		PIPEID(x)					((x) << 0)
 #define		MEID(x)						((x) << 2)
@@ -364,6 +370,7 @@
 		 * 1 - Stream
 		 * 2 - Bypass
 		 */
+#define		EOP_EXEC				(1 << 28) /* For Trailing Fence */
 #define		DATA_SEL(x)                             ((x) << 29)
 		/* 0 - discard
 		 * 1 - send low 32bit data
diff --git a/drivers/gpu/drm/amd/amdgpu/clearstate_gfx11.h b/drivers/gpu/drm/amd/amdgpu/clearstate_gfx11.h
new file mode 100644
index 000000000000..a8b29d33c464
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/clearstate_gfx11.h
@@ -0,0 +1,997 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __CLEARSTATE_GFX11_H_
+#define __CLEARSTATE_GFX11_H_
+
+static const unsigned int gfx11_SECT_CONTEXT_def_1[] =
+{
+    0x00000000, // DB_RENDER_CONTROL
+    0x00000000, // DB_COUNT_CONTROL
+    0x00000000, // DB_DEPTH_VIEW
+    0x00000000, // DB_RENDER_OVERRIDE
+    0x00000000, // DB_RENDER_OVERRIDE2
+    0x00000000, // DB_HTILE_DATA_BASE
+    0, // HOLE
+    0x00000000, // DB_DEPTH_SIZE_XY
+    0x00000000, // DB_DEPTH_BOUNDS_MIN
+    0x00000000, // DB_DEPTH_BOUNDS_MAX
+    0x00000000, // DB_STENCIL_CLEAR
+    0x00000000, // DB_DEPTH_CLEAR
+    0x00000000, // PA_SC_SCREEN_SCISSOR_TL
+    0x40004000, // PA_SC_SCREEN_SCISSOR_BR
+    0, // HOLE
+    0x00000000, // DB_RESERVED_REG_2
+    0x00000000, // DB_Z_INFO
+    0x00000000, // DB_STENCIL_INFO
+    0x00000000, // DB_Z_READ_BASE
+    0x00000000, // DB_STENCIL_READ_BASE
+    0x00000000, // DB_Z_WRITE_BASE
+    0x00000000, // DB_STENCIL_WRITE_BASE
+    0x00000000, // DB_RESERVED_REG_1
+    0x00000000, // DB_RESERVED_REG_3
+    0x00000000, // DB_SPI_VRS_CENTER_LOCATION
+    0, // HOLE
+    0x00000000, // DB_Z_READ_BASE_HI
+    0x00000000, // DB_STENCIL_READ_BASE_HI
+    0x00000000, // DB_Z_WRITE_BASE_HI
+    0x00000000, // DB_STENCIL_WRITE_BASE_HI
+    0x00000000, // DB_HTILE_DATA_BASE_HI
+    0x00150055, // DB_RMI_L2_CACHE_CONTROL
+    0x00000000, // TA_BC_BASE_ADDR
+    0x00000000, // TA_BC_BASE_ADDR_HI
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // COHER_DEST_BASE_HI_0
+    0x00000000, // COHER_DEST_BASE_HI_1
+    0x00000000, // COHER_DEST_BASE_HI_2
+    0x00000000, // COHER_DEST_BASE_HI_3
+    0x00000000, // COHER_DEST_BASE_2
+    0x00000000, // COHER_DEST_BASE_3
+    0x00000000, // PA_SC_WINDOW_OFFSET
+    0x80000000, // PA_SC_WINDOW_SCISSOR_TL
+    0x40004000, // PA_SC_WINDOW_SCISSOR_BR
+    0x0000ffff, // PA_SC_CLIPRECT_RULE
+    0x00000000, // PA_SC_CLIPRECT_0_TL
+    0x40004000, // PA_SC_CLIPRECT_0_BR
+    0x00000000, // PA_SC_CLIPRECT_1_TL
+    0x40004000, // PA_SC_CLIPRECT_1_BR
+    0x00000000, // PA_SC_CLIPRECT_2_TL
+    0x40004000, // PA_SC_CLIPRECT_2_BR
+    0x00000000, // PA_SC_CLIPRECT_3_TL
+    0x40004000, // PA_SC_CLIPRECT_3_BR
+    0xaa99aaaa, // PA_SC_EDGERULE
+    0x00000000, // PA_SU_HARDWARE_SCREEN_OFFSET
+    0xffffffff, // CB_TARGET_MASK
+    0xffffffff, // CB_SHADER_MASK
+    0x80000000, // PA_SC_GENERIC_SCISSOR_TL
+    0x40004000, // PA_SC_GENERIC_SCISSOR_BR
+    0x00000000, // COHER_DEST_BASE_0
+    0x00000000, // COHER_DEST_BASE_1
+    0x80000000, // PA_SC_VPORT_SCISSOR_0_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_0_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_1_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_1_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_2_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_2_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_3_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_3_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_4_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_4_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_5_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_5_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_6_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_6_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_7_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_7_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_8_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_8_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_9_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_9_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_10_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_10_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_11_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_11_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_12_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_12_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_13_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_13_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_14_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_14_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_15_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_15_BR
+    0x00000000, // PA_SC_VPORT_ZMIN_0
+    0x3f800000, // PA_SC_VPORT_ZMAX_0
+    0x00000000, // PA_SC_VPORT_ZMIN_1
+    0x3f800000, // PA_SC_VPORT_ZMAX_1
+    0x00000000, // PA_SC_VPORT_ZMIN_2
+    0x3f800000, // PA_SC_VPORT_ZMAX_2
+    0x00000000, // PA_SC_VPORT_ZMIN_3
+    0x3f800000, // PA_SC_VPORT_ZMAX_3
+    0x00000000, // PA_SC_VPORT_ZMIN_4
+    0x3f800000, // PA_SC_VPORT_ZMAX_4
+    0x00000000, // PA_SC_VPORT_ZMIN_5
+    0x3f800000, // PA_SC_VPORT_ZMAX_5
+    0x00000000, // PA_SC_VPORT_ZMIN_6
+    0x3f800000, // PA_SC_VPORT_ZMAX_6
+    0x00000000, // PA_SC_VPORT_ZMIN_7
+    0x3f800000, // PA_SC_VPORT_ZMAX_7
+    0x00000000, // PA_SC_VPORT_ZMIN_8
+    0x3f800000, // PA_SC_VPORT_ZMAX_8
+    0x00000000, // PA_SC_VPORT_ZMIN_9
+    0x3f800000, // PA_SC_VPORT_ZMAX_9
+    0x00000000, // PA_SC_VPORT_ZMIN_10
+    0x3f800000, // PA_SC_VPORT_ZMAX_10
+    0x00000000, // PA_SC_VPORT_ZMIN_11
+    0x3f800000, // PA_SC_VPORT_ZMAX_11
+    0x00000000, // PA_SC_VPORT_ZMIN_12
+    0x3f800000, // PA_SC_VPORT_ZMAX_12
+    0x00000000, // PA_SC_VPORT_ZMIN_13
+    0x3f800000, // PA_SC_VPORT_ZMAX_13
+    0x00000000, // PA_SC_VPORT_ZMIN_14
+    0x3f800000, // PA_SC_VPORT_ZMAX_14
+    0x00000000, // PA_SC_VPORT_ZMIN_15
+    0x3f800000, // PA_SC_VPORT_ZMAX_15
+    0x00000000, // PA_SC_RASTER_CONFIG
+    0x00000000, // PA_SC_RASTER_CONFIG_1
+    0x00000000, // PA_SC_SCREEN_EXTENT_CONTROL
+};
+static const unsigned int gfx11_SECT_CONTEXT_def_2[] =
+{
+    0x00000000, // CP_PERFMON_CNTX_CNTL
+    0x00000000, // CP_PIPEID
+    0x00000000, // CP_VMID
+    0x00000000, // CONTEXT_RESERVED_REG0
+    0x00000000, // CONTEXT_RESERVED_REG1
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // PA_SC_FSR_EN
+    0x00000000, // PA_SC_FSR_FBW_RECURSIONS_X
+    0x00000000, // PA_SC_FSR_FBW_RECURSIONS_Y
+    0, // HOLE
+    0x00000000, // PA_SC_VRS_OVERRIDE_CNTL
+    0x00000000, // PA_SC_VRS_RATE_FEEDBACK_BASE
+    0x00000000, // PA_SC_VRS_RATE_FEEDBACK_BASE_EXT
+    0x00000000, // PA_SC_VRS_RATE_FEEDBACK_SIZE_XY
+    0x00000000, // PA_SC_BINNER_OUTPUT_TIMEOUT_CNTL
+    0x00000000, // PA_SC_VRS_RATE_CACHE_CNTL
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // PA_SC_VRS_RATE_BASE
+    0x00000000, // PA_SC_VRS_RATE_BASE_EXT
+    0x00000000, // PA_SC_VRS_RATE_SIZE_XY
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // VGT_MULTI_PRIM_IB_RESET_INDX
+    0x00550055, // CB_RMI_GL2_CACHE_CONTROL
+    0x00000000, // CB_BLEND_RED
+    0x00000000, // CB_BLEND_GREEN
+    0x00000000, // CB_BLEND_BLUE
+    0x00000000, // CB_BLEND_ALPHA
+    0x00000000, // CB_FDCC_CONTROL
+    0x00000000, // CB_COVERAGE_OUT_CONTROL
+    0x00000000, // DB_STENCIL_CONTROL
+    0x01000000, // DB_STENCILREFMASK
+    0x01000000, // DB_STENCILREFMASK_BF
+    0, // HOLE
+    0x00000000, // PA_CL_VPORT_XSCALE
+    0x00000000, // PA_CL_VPORT_XOFFSET
+    0x00000000, // PA_CL_VPORT_YSCALE
+    0x00000000, // PA_CL_VPORT_YOFFSET
+    0x00000000, // PA_CL_VPORT_ZSCALE
+    0x00000000, // PA_CL_VPORT_ZOFFSET
+    0x00000000, // PA_CL_VPORT_XSCALE_1
+    0x00000000, // PA_CL_VPORT_XOFFSET_1
+    0x00000000, // PA_CL_VPORT_YSCALE_1
+    0x00000000, // PA_CL_VPORT_YOFFSET_1
+    0x00000000, // PA_CL_VPORT_ZSCALE_1
+    0x00000000, // PA_CL_VPORT_ZOFFSET_1
+    0x00000000, // PA_CL_VPORT_XSCALE_2
+    0x00000000, // PA_CL_VPORT_XOFFSET_2
+    0x00000000, // PA_CL_VPORT_YSCALE_2
+    0x00000000, // PA_CL_VPORT_YOFFSET_2
+    0x00000000, // PA_CL_VPORT_ZSCALE_2
+    0x00000000, // PA_CL_VPORT_ZOFFSET_2
+    0x00000000, // PA_CL_VPORT_XSCALE_3
+    0x00000000, // PA_CL_VPORT_XOFFSET_3
+    0x00000000, // PA_CL_VPORT_YSCALE_3
+    0x00000000, // PA_CL_VPORT_YOFFSET_3
+    0x00000000, // PA_CL_VPORT_ZSCALE_3
+    0x00000000, // PA_CL_VPORT_ZOFFSET_3
+    0x00000000, // PA_CL_VPORT_XSCALE_4
+    0x00000000, // PA_CL_VPORT_XOFFSET_4
+    0x00000000, // PA_CL_VPORT_YSCALE_4
+    0x00000000, // PA_CL_VPORT_YOFFSET_4
+    0x00000000, // PA_CL_VPORT_ZSCALE_4
+    0x00000000, // PA_CL_VPORT_ZOFFSET_4
+    0x00000000, // PA_CL_VPORT_XSCALE_5
+    0x00000000, // PA_CL_VPORT_XOFFSET_5
+    0x00000000, // PA_CL_VPORT_YSCALE_5
+    0x00000000, // PA_CL_VPORT_YOFFSET_5
+    0x00000000, // PA_CL_VPORT_ZSCALE_5
+    0x00000000, // PA_CL_VPORT_ZOFFSET_5
+    0x00000000, // PA_CL_VPORT_XSCALE_6
+    0x00000000, // PA_CL_VPORT_XOFFSET_6
+    0x00000000, // PA_CL_VPORT_YSCALE_6
+    0x00000000, // PA_CL_VPORT_YOFFSET_6
+    0x00000000, // PA_CL_VPORT_ZSCALE_6
+    0x00000000, // PA_CL_VPORT_ZOFFSET_6
+    0x00000000, // PA_CL_VPORT_XSCALE_7
+    0x00000000, // PA_CL_VPORT_XOFFSET_7
+    0x00000000, // PA_CL_VPORT_YSCALE_7
+    0x00000000, // PA_CL_VPORT_YOFFSET_7
+    0x00000000, // PA_CL_VPORT_ZSCALE_7
+    0x00000000, // PA_CL_VPORT_ZOFFSET_7
+    0x00000000, // PA_CL_VPORT_XSCALE_8
+    0x00000000, // PA_CL_VPORT_XOFFSET_8
+    0x00000000, // PA_CL_VPORT_YSCALE_8
+    0x00000000, // PA_CL_VPORT_YOFFSET_8
+    0x00000000, // PA_CL_VPORT_ZSCALE_8
+    0x00000000, // PA_CL_VPORT_ZOFFSET_8
+    0x00000000, // PA_CL_VPORT_XSCALE_9
+    0x00000000, // PA_CL_VPORT_XOFFSET_9
+    0x00000000, // PA_CL_VPORT_YSCALE_9
+    0x00000000, // PA_CL_VPORT_YOFFSET_9
+    0x00000000, // PA_CL_VPORT_ZSCALE_9
+    0x00000000, // PA_CL_VPORT_ZOFFSET_9
+    0x00000000, // PA_CL_VPORT_XSCALE_10
+    0x00000000, // PA_CL_VPORT_XOFFSET_10
+    0x00000000, // PA_CL_VPORT_YSCALE_10
+    0x00000000, // PA_CL_VPORT_YOFFSET_10
+    0x00000000, // PA_CL_VPORT_ZSCALE_10
+    0x00000000, // PA_CL_VPORT_ZOFFSET_10
+    0x00000000, // PA_CL_VPORT_XSCALE_11
+    0x00000000, // PA_CL_VPORT_XOFFSET_11
+    0x00000000, // PA_CL_VPORT_YSCALE_11
+    0x00000000, // PA_CL_VPORT_YOFFSET_11
+    0x00000000, // PA_CL_VPORT_ZSCALE_11
+    0x00000000, // PA_CL_VPORT_ZOFFSET_11
+    0x00000000, // PA_CL_VPORT_XSCALE_12
+    0x00000000, // PA_CL_VPORT_XOFFSET_12
+    0x00000000, // PA_CL_VPORT_YSCALE_12
+    0x00000000, // PA_CL_VPORT_YOFFSET_12
+    0x00000000, // PA_CL_VPORT_ZSCALE_12
+    0x00000000, // PA_CL_VPORT_ZOFFSET_12
+    0x00000000, // PA_CL_VPORT_XSCALE_13
+    0x00000000, // PA_CL_VPORT_XOFFSET_13
+    0x00000000, // PA_CL_VPORT_YSCALE_13
+    0x00000000, // PA_CL_VPORT_YOFFSET_13
+    0x00000000, // PA_CL_VPORT_ZSCALE_13
+    0x00000000, // PA_CL_VPORT_ZOFFSET_13
+    0x00000000, // PA_CL_VPORT_XSCALE_14
+    0x00000000, // PA_CL_VPORT_XOFFSET_14
+    0x00000000, // PA_CL_VPORT_YSCALE_14
+    0x00000000, // PA_CL_VPORT_YOFFSET_14
+    0x00000000, // PA_CL_VPORT_ZSCALE_14
+    0x00000000, // PA_CL_VPORT_ZOFFSET_14
+    0x00000000, // PA_CL_VPORT_XSCALE_15
+    0x00000000, // PA_CL_VPORT_XOFFSET_15
+    0x00000000, // PA_CL_VPORT_YSCALE_15
+    0x00000000, // PA_CL_VPORT_YOFFSET_15
+    0x00000000, // PA_CL_VPORT_ZSCALE_15
+    0x00000000, // PA_CL_VPORT_ZOFFSET_15
+    0x00000000, // PA_CL_UCP_0_X
+    0x00000000, // PA_CL_UCP_0_Y
+    0x00000000, // PA_CL_UCP_0_Z
+    0x00000000, // PA_CL_UCP_0_W
+    0x00000000, // PA_CL_UCP_1_X
+    0x00000000, // PA_CL_UCP_1_Y
+    0x00000000, // PA_CL_UCP_1_Z
+    0x00000000, // PA_CL_UCP_1_W
+    0x00000000, // PA_CL_UCP_2_X
+    0x00000000, // PA_CL_UCP_2_Y
+    0x00000000, // PA_CL_UCP_2_Z
+    0x00000000, // PA_CL_UCP_2_W
+    0x00000000, // PA_CL_UCP_3_X
+    0x00000000, // PA_CL_UCP_3_Y
+    0x00000000, // PA_CL_UCP_3_Z
+    0x00000000, // PA_CL_UCP_3_W
+    0x00000000, // PA_CL_UCP_4_X
+    0x00000000, // PA_CL_UCP_4_Y
+    0x00000000, // PA_CL_UCP_4_Z
+    0x00000000, // PA_CL_UCP_4_W
+    0x00000000, // PA_CL_UCP_5_X
+    0x00000000, // PA_CL_UCP_5_Y
+    0x00000000, // PA_CL_UCP_5_Z
+    0x00000000, // PA_CL_UCP_5_W
+    0x00000000, // PA_CL_PROG_NEAR_CLIP_Z
+    0x00000000, // PA_RATE_CNTL
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SPI_PS_INPUT_CNTL_0
+    0x00000000, // SPI_PS_INPUT_CNTL_1
+    0x00000000, // SPI_PS_INPUT_CNTL_2
+    0x00000000, // SPI_PS_INPUT_CNTL_3
+    0x00000000, // SPI_PS_INPUT_CNTL_4
+    0x00000000, // SPI_PS_INPUT_CNTL_5
+    0x00000000, // SPI_PS_INPUT_CNTL_6
+    0x00000000, // SPI_PS_INPUT_CNTL_7
+    0x00000000, // SPI_PS_INPUT_CNTL_8
+    0x00000000, // SPI_PS_INPUT_CNTL_9
+    0x00000000, // SPI_PS_INPUT_CNTL_10
+    0x00000000, // SPI_PS_INPUT_CNTL_11
+    0x00000000, // SPI_PS_INPUT_CNTL_12
+    0x00000000, // SPI_PS_INPUT_CNTL_13
+    0x00000000, // SPI_PS_INPUT_CNTL_14
+    0x00000000, // SPI_PS_INPUT_CNTL_15
+    0x00000000, // SPI_PS_INPUT_CNTL_16
+    0x00000000, // SPI_PS_INPUT_CNTL_17
+    0x00000000, // SPI_PS_INPUT_CNTL_18
+    0x00000000, // SPI_PS_INPUT_CNTL_19
+    0x00000000, // SPI_PS_INPUT_CNTL_20
+    0x00000000, // SPI_PS_INPUT_CNTL_21
+    0x00000000, // SPI_PS_INPUT_CNTL_22
+    0x00000000, // SPI_PS_INPUT_CNTL_23
+    0x00000000, // SPI_PS_INPUT_CNTL_24
+    0x00000000, // SPI_PS_INPUT_CNTL_25
+    0x00000000, // SPI_PS_INPUT_CNTL_26
+    0x00000000, // SPI_PS_INPUT_CNTL_27
+    0x00000000, // SPI_PS_INPUT_CNTL_28
+    0x00000000, // SPI_PS_INPUT_CNTL_29
+    0x00000000, // SPI_PS_INPUT_CNTL_30
+    0x00000000, // SPI_PS_INPUT_CNTL_31
+    0x00000000, // SPI_VS_OUT_CONFIG
+    0, // HOLE
+    0x00000000, // SPI_PS_INPUT_ENA
+    0x00000000, // SPI_PS_INPUT_ADDR
+    0x00000000, // SPI_INTERP_CONTROL_0
+    0x00000002, // SPI_PS_IN_CONTROL
+    0x00000000, // SPI_BARYC_SSAA_CNTL
+    0x00000000, // SPI_BARYC_CNTL
+    0, // HOLE
+    0x00000000, // SPI_TMPRING_SIZE
+    0x00000000, // SPI_GFX_SCRATCH_BASE_LO
+    0x00000000, // SPI_GFX_SCRATCH_BASE_HI
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SPI_SHADER_IDX_FORMAT
+    0x00000000, // SPI_SHADER_POS_FORMAT
+    0x00000000, // SPI_SHADER_Z_FORMAT
+    0x00000000, // SPI_SHADER_COL_FORMAT
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SX_PS_DOWNCONVERT_CONTROL
+    0x00000000, // SX_PS_DOWNCONVERT
+    0x00000000, // SX_BLEND_OPT_EPSILON
+    0x00000000, // SX_BLEND_OPT_CONTROL
+    0x00000000, // SX_MRT0_BLEND_OPT
+    0x00000000, // SX_MRT1_BLEND_OPT
+    0x00000000, // SX_MRT2_BLEND_OPT
+    0x00000000, // SX_MRT3_BLEND_OPT
+    0x00000000, // SX_MRT4_BLEND_OPT
+    0x00000000, // SX_MRT5_BLEND_OPT
+    0x00000000, // SX_MRT6_BLEND_OPT
+    0x00000000, // SX_MRT7_BLEND_OPT
+    0x00000000, // CB_BLEND0_CONTROL
+    0x00000000, // CB_BLEND1_CONTROL
+    0x00000000, // CB_BLEND2_CONTROL
+    0x00000000, // CB_BLEND3_CONTROL
+    0x00000000, // CB_BLEND4_CONTROL
+    0x00000000, // CB_BLEND5_CONTROL
+    0x00000000, // CB_BLEND6_CONTROL
+    0x00000000, // CB_BLEND7_CONTROL
+};
+static const unsigned int gfx11_SECT_CONTEXT_def_3[] =
+{
+    0x00000000, // PA_CL_POINT_X_RAD
+    0x00000000, // PA_CL_POINT_Y_RAD
+    0x00000000, // PA_CL_POINT_SIZE
+    0x00000000, // PA_CL_POINT_CULL_RAD
+};
+static const unsigned int gfx11_SECT_CONTEXT_def_4[] =
+{
+    0x00000000, // GE_MAX_OUTPUT_PER_SUBGROUP
+    0x00000000, // DB_DEPTH_CONTROL
+    0x00000000, // DB_EQAA
+    0x00000000, // CB_COLOR_CONTROL
+    0x00000000, // DB_SHADER_CONTROL
+    0x00090000, // PA_CL_CLIP_CNTL
+    0x00000004, // PA_SU_SC_MODE_CNTL
+    0x00000000, // PA_CL_VTE_CNTL
+    0x00000000, // PA_CL_VS_OUT_CNTL
+    0x00000000, // PA_CL_NANINF_CNTL
+    0x00000000, // PA_SU_LINE_STIPPLE_CNTL
+    0x00000000, // PA_SU_LINE_STIPPLE_SCALE
+    0x00000000, // PA_SU_PRIM_FILTER_CNTL
+    0x00000000, // PA_SU_SMALL_PRIM_FILTER_CNTL
+    0, // HOLE
+    0x00000000, // PA_CL_NGG_CNTL
+    0x00000000, // PA_SU_OVER_RASTERIZATION_CNTL
+    0x00000000, // PA_STEREO_CNTL
+    0x00000000, // PA_STATE_STEREO_X
+    0x00000000, // PA_CL_VRS_CNTL
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // PA_SU_POINT_SIZE
+    0x00000000, // PA_SU_POINT_MINMAX
+    0x00000000, // PA_SU_LINE_CNTL
+    0x00000000, // PA_SC_LINE_STIPPLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // VGT_HOS_MAX_TESS_LEVEL
+    0x00000000, // VGT_HOS_MIN_TESS_LEVEL
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // VGT_GS_ONCHIP_CNTL
+    0x00000000, // PA_SC_MODE_CNTL_0
+    0x00000000, // PA_SC_MODE_CNTL_1
+    0x00000000, // VGT_ENHANCE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // IA_ENHANCE
+};
+static const unsigned int gfx11_SECT_CONTEXT_def_5[] =
+{
+    0x00000000, // WD_ENHANCE
+    0x00000000, // VGT_PRIMITIVEID_EN
+};
+static const unsigned int gfx11_SECT_CONTEXT_def_6[] =
+{
+    0x00000000, // VGT_PRIMITIVEID_RESET
+};
+static const unsigned int gfx11_SECT_CONTEXT_def_7[] =
+{
+    0x00000000, // VGT_DRAW_PAYLOAD_CNTL
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // VGT_ESGS_RING_ITEMSIZE
+    0, // HOLE
+    0x00000000, // VGT_REUSE_OFF
+    0, // HOLE
+    0x00000000, // DB_HTILE_SURFACE
+    0x00000000, // DB_SRESULTS_COMPARE_STATE0
+    0x00000000, // DB_SRESULTS_COMPARE_STATE1
+    0x00000000, // DB_PRELOAD_CONTROL
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // VGT_STRMOUT_DRAW_OPAQUE_OFFSET
+    0x00000000, // VGT_STRMOUT_DRAW_OPAQUE_BUFFER_FILLED_SIZE
+    0x00000000, // VGT_STRMOUT_DRAW_OPAQUE_VERTEX_STRIDE
+    0, // HOLE
+    0x00000000, // VGT_GS_MAX_VERT_OUT
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // GE_NGG_SUBGRP_CNTL
+    0x00000000, // VGT_TESS_DISTRIBUTION
+    0x00000000, // VGT_SHADER_STAGES_EN
+    0x00000000, // VGT_LS_HS_CONFIG
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // VGT_TF_PARAM
+    0x00000000, // DB_ALPHA_TO_MASK
+    0, // HOLE
+    0x00000000, // PA_SU_POLY_OFFSET_DB_FMT_CNTL
+    0x00000000, // PA_SU_POLY_OFFSET_CLAMP
+    0x00000000, // PA_SU_POLY_OFFSET_FRONT_SCALE
+    0x00000000, // PA_SU_POLY_OFFSET_FRONT_OFFSET
+    0x00000000, // PA_SU_POLY_OFFSET_BACK_SCALE
+    0x00000000, // PA_SU_POLY_OFFSET_BACK_OFFSET
+    0x00000000, // VGT_GS_INSTANCE_CNT
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // PA_SC_CENTROID_PRIORITY_0
+    0x00000000, // PA_SC_CENTROID_PRIORITY_1
+    0x00001000, // PA_SC_LINE_CNTL
+    0x00000000, // PA_SC_AA_CONFIG
+    0x00000005, // PA_SU_VTX_CNTL
+    0x3f800000, // PA_CL_GB_VERT_CLIP_ADJ
+    0x3f800000, // PA_CL_GB_VERT_DISC_ADJ
+    0x3f800000, // PA_CL_GB_HORZ_CLIP_ADJ
+    0x3f800000, // PA_CL_GB_HORZ_DISC_ADJ
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_0
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_1
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_2
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_3
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y0_0
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y0_1
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y0_2
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y0_3
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y1_0
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y1_1
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y1_2
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y1_3
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y1_0
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y1_1
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y1_2
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y1_3
+    0xffffffff, // PA_SC_AA_MASK_X0Y0_X1Y0
+    0xffffffff, // PA_SC_AA_MASK_X0Y1_X1Y1
+    0x00000000, // PA_SC_SHADER_CONTROL
+    0x00000003, // PA_SC_BINNER_CNTL_0
+    0x00000000, // PA_SC_BINNER_CNTL_1
+    0x00100000, // PA_SC_CONSERVATIVE_RASTERIZATION_CNTL
+    0x00000000, // PA_SC_NGG_MODE_CNTL
+    0x00000000, // PA_SC_BINNER_CNTL_2
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // CB_COLOR0_BASE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // CB_COLOR0_VIEW
+    0x00000000, // CB_COLOR0_INFO
+    0x00000000, // CB_COLOR0_ATTRIB
+    0x00000000, // CB_COLOR0_FDCC_CONTROL
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // CB_COLOR0_DCC_BASE
+    0, // HOLE
+    0x00000000, // CB_COLOR1_BASE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // CB_COLOR1_VIEW
+    0x00000000, // CB_COLOR1_INFO
+    0x00000000, // CB_COLOR1_ATTRIB
+    0x00000000, // CB_COLOR1_FDCC_CONTROL
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // CB_COLOR1_DCC_BASE
+    0, // HOLE
+    0x00000000, // CB_COLOR2_BASE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // CB_COLOR2_VIEW
+    0x00000000, // CB_COLOR2_INFO
+    0x00000000, // CB_COLOR2_ATTRIB
+    0x00000000, // CB_COLOR2_FDCC_CONTROL
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // CB_COLOR2_DCC_BASE
+    0, // HOLE
+    0x00000000, // CB_COLOR3_BASE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // CB_COLOR3_VIEW
+    0x00000000, // CB_COLOR3_INFO
+    0x00000000, // CB_COLOR3_ATTRIB
+    0x00000000, // CB_COLOR3_FDCC_CONTROL
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // CB_COLOR3_DCC_BASE
+    0, // HOLE
+    0x00000000, // CB_COLOR4_BASE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // CB_COLOR4_VIEW
+    0x00000000, // CB_COLOR4_INFO
+    0x00000000, // CB_COLOR4_ATTRIB
+    0x00000000, // CB_COLOR4_FDCC_CONTROL
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // CB_COLOR4_DCC_BASE
+    0, // HOLE
+    0x00000000, // CB_COLOR5_BASE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // CB_COLOR5_VIEW
+    0x00000000, // CB_COLOR5_INFO
+    0x00000000, // CB_COLOR5_ATTRIB
+    0x00000000, // CB_COLOR5_FDCC_CONTROL
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // CB_COLOR5_DCC_BASE
+    0, // HOLE
+    0x00000000, // CB_COLOR6_BASE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // CB_COLOR6_VIEW
+    0x00000000, // CB_COLOR6_INFO
+    0x00000000, // CB_COLOR6_ATTRIB
+    0x00000000, // CB_COLOR6_FDCC_CONTROL
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // CB_COLOR6_DCC_BASE
+    0, // HOLE
+    0x00000000, // CB_COLOR7_BASE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // CB_COLOR7_VIEW
+    0x00000000, // CB_COLOR7_INFO
+    0x00000000, // CB_COLOR7_ATTRIB
+    0x00000000, // CB_COLOR7_FDCC_CONTROL
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // CB_COLOR7_DCC_BASE
+    0, // HOLE
+    0x00000000, // CB_COLOR0_BASE_EXT
+    0x00000000, // CB_COLOR1_BASE_EXT
+    0x00000000, // CB_COLOR2_BASE_EXT
+    0x00000000, // CB_COLOR3_BASE_EXT
+    0x00000000, // CB_COLOR4_BASE_EXT
+    0x00000000, // CB_COLOR5_BASE_EXT
+    0x00000000, // CB_COLOR6_BASE_EXT
+    0x00000000, // CB_COLOR7_BASE_EXT
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // CB_COLOR0_DCC_BASE_EXT
+    0x00000000, // CB_COLOR1_DCC_BASE_EXT
+    0x00000000, // CB_COLOR2_DCC_BASE_EXT
+    0x00000000, // CB_COLOR3_DCC_BASE_EXT
+    0x00000000, // CB_COLOR4_DCC_BASE_EXT
+    0x00000000, // CB_COLOR5_DCC_BASE_EXT
+    0x00000000, // CB_COLOR6_DCC_BASE_EXT
+    0x00000000, // CB_COLOR7_DCC_BASE_EXT
+    0x00000000, // CB_COLOR0_ATTRIB2
+    0x00000000, // CB_COLOR1_ATTRIB2
+    0x00000000, // CB_COLOR2_ATTRIB2
+    0x00000000, // CB_COLOR3_ATTRIB2
+    0x00000000, // CB_COLOR4_ATTRIB2
+    0x00000000, // CB_COLOR5_ATTRIB2
+    0x00000000, // CB_COLOR6_ATTRIB2
+    0x00000000, // CB_COLOR7_ATTRIB2
+    0x00000000, // CB_COLOR0_ATTRIB3
+    0x00000000, // CB_COLOR1_ATTRIB3
+    0x00000000, // CB_COLOR2_ATTRIB3
+    0x00000000, // CB_COLOR3_ATTRIB3
+    0x00000000, // CB_COLOR4_ATTRIB3
+    0x00000000, // CB_COLOR5_ATTRIB3
+    0x00000000, // CB_COLOR6_ATTRIB3
+    0x00000000, // CB_COLOR7_ATTRIB3
+};
+static const struct cs_extent_def gfx11_SECT_CONTEXT_defs[] =
+{
+    {gfx11_SECT_CONTEXT_def_1, 0x0000a000, 215 },
+    {gfx11_SECT_CONTEXT_def_2, 0x0000a0d8, 272 },
+    {gfx11_SECT_CONTEXT_def_3, 0x0000a1f5, 4 },
+    {gfx11_SECT_CONTEXT_def_4, 0x0000a1ff, 158 },
+    {gfx11_SECT_CONTEXT_def_5, 0x0000a2a0, 2 },
+    {gfx11_SECT_CONTEXT_def_6, 0x0000a2a3, 1 },
+    {gfx11_SECT_CONTEXT_def_7, 0x0000a2a6, 282 },
+    { 0, 0, 0 }
+};
+static const struct cs_section_def gfx11_cs_data[] = {
+    { gfx11_SECT_CONTEXT_defs, SECT_CONTEXT },
+    { 0, SECT_NONE }
+};
+
+#endif /* __CLEARSTATE_GFX11_H_ */
diff --git a/drivers/gpu/drm/amd/amdgpu/clearstate_gfx12.h b/drivers/gpu/drm/amd/amdgpu/clearstate_gfx12.h
new file mode 100644
index 000000000000..2f6c9d11d5ae
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/clearstate_gfx12.h
@@ -0,0 +1,121 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __CLEARSTATE_GFX12_H_
+#define __CLEARSTATE_GFX12_H_
+
+static const unsigned int gfx12_SECT_CONTEXT_def_1[] = {
+0x00000000, //mmSC_MEM_TEMPORAL
+0x00000000, //mmSC_MEM_SPEC_READ
+0x00000000, //mmPA_SC_VPORT_0_TL
+0x00000000, //mmPA_SC_VPORT_0_BR
+0x00000000, //mmPA_SC_VPORT_1_TL
+0x00000000, //mmPA_SC_VPORT_1_BR
+0x00000000, //mmPA_SC_VPORT_2_TL
+0x00000000, //mmPA_SC_VPORT_2_BR
+0x00000000, //mmPA_SC_VPORT_3_TL
+0x00000000, //mmPA_SC_VPORT_3_BR
+0x00000000, //mmPA_SC_VPORT_4_TL
+0x00000000, //mmPA_SC_VPORT_4_BR
+0x00000000, //mmPA_SC_VPORT_5_TL
+0x00000000, //mmPA_SC_VPORT_5_BR
+0x00000000, //mmPA_SC_VPORT_6_TL
+0x00000000, //mmPA_SC_VPORT_6_BR
+0x00000000, //mmPA_SC_VPORT_7_TL
+0x00000000, //mmPA_SC_VPORT_7_BR
+0x00000000, //mmPA_SC_VPORT_8_TL
+0x00000000, //mmPA_SC_VPORT_8_BR
+0x00000000, //mmPA_SC_VPORT_9_TL
+0x00000000, //mmPA_SC_VPORT_9_BR
+0x00000000, //mmPA_SC_VPORT_10_TL
+0x00000000, //mmPA_SC_VPORT_10_BR
+0x00000000, //mmPA_SC_VPORT_11_TL
+0x00000000, //mmPA_SC_VPORT_11_BR
+0x00000000, //mmPA_SC_VPORT_12_TL
+0x00000000, //mmPA_SC_VPORT_12_BR
+0x00000000, //mmPA_SC_VPORT_13_TL
+0x00000000, //mmPA_SC_VPORT_13_BR
+0x00000000, //mmPA_SC_VPORT_14_TL
+0x00000000, //mmPA_SC_VPORT_14_BR
+0x00000000, //mmPA_SC_VPORT_15_TL
+0x00000000, //mmPA_SC_VPORT_15_BR
+};
+
+static const unsigned int gfx12_SECT_CONTEXT_def_2[] = {
+0x00000000, //mmPA_CL_PROG_NEAR_CLIP_Z
+0x00000000, //mmPA_RATE_CNTL
+};
+
+static const unsigned int gfx12_SECT_CONTEXT_def_3[] = {
+0x00000000, //mmCP_PERFMON_CNTX_CNTL
+};
+
+static const unsigned int gfx12_SECT_CONTEXT_def_4[] = {
+0x00000000, //mmCONTEXT_RESERVED_REG0
+0x00000000, //mmCONTEXT_RESERVED_REG1
+0x00000000, //mmPA_SC_CLIPRECT_0_EXT
+0x00000000, //mmPA_SC_CLIPRECT_1_EXT
+0x00000000, //mmPA_SC_CLIPRECT_2_EXT
+0x00000000, //mmPA_SC_CLIPRECT_3_EXT
+};
+
+static const unsigned int gfx12_SECT_CONTEXT_def_5[] = {
+0x00000000, //mmPA_SC_HIZ_INFO
+0x00000000, //mmPA_SC_HIS_INFO
+0x00000000, //mmPA_SC_HIZ_BASE
+0x00000000, //mmPA_SC_HIZ_BASE_EXT
+0x00000000, //mmPA_SC_HIZ_SIZE_XY
+0x00000000, //mmPA_SC_HIS_BASE
+0x00000000, //mmPA_SC_HIS_BASE_EXT
+0x00000000, //mmPA_SC_HIS_SIZE_XY
+0x00000000, //mmPA_SC_BINNER_OUTPUT_TIMEOUT_CNTL
+0x00000000, //mmPA_SC_BINNER_DYNAMIC_BATCH_LIMIT
+0x00000000, //mmPA_SC_HISZ_CONTROL
+};
+
+static const unsigned int gfx12_SECT_CONTEXT_def_6[] = {
+0x00000000, //mmCB_MEM0_INFO
+0x00000000, //mmCB_MEM1_INFO
+0x00000000, //mmCB_MEM2_INFO
+0x00000000, //mmCB_MEM3_INFO
+0x00000000, //mmCB_MEM4_INFO
+0x00000000, //mmCB_MEM5_INFO
+0x00000000, //mmCB_MEM6_INFO
+0x00000000, //mmCB_MEM7_INFO
+};
+
+static const struct cs_extent_def gfx12_SECT_CONTEXT_defs[] = {
+    {gfx12_SECT_CONTEXT_def_1, 0x0000a03e, 34 },
+    {gfx12_SECT_CONTEXT_def_2, 0x0000a0cc, 2 },
+    {gfx12_SECT_CONTEXT_def_3, 0x0000a0d8, 1 },
+    {gfx12_SECT_CONTEXT_def_4, 0x0000a0db, 6 },
+    {gfx12_SECT_CONTEXT_def_5, 0x0000a2e5, 11 },
+    {gfx12_SECT_CONTEXT_def_6, 0x0000a3c0, 8 },
+    { 0, 0, 0 }
+};
+
+static const struct cs_section_def gfx12_cs_data[] = {
+    { gfx12_SECT_CONTEXT_defs, SECT_CONTEXT },
+    { 0, SECT_NONE }
+};
+
+#endif /* __CLEARSTATE_GFX12_H_ */
diff --git a/drivers/gpu/drm/amd/amdgpu/clearstate_gfx9.h b/drivers/gpu/drm/amd/amdgpu/clearstate_gfx9.h
index 567a904804bc..9c85ca6358c1 100644
--- a/drivers/gpu/drm/amd/amdgpu/clearstate_gfx9.h
+++ b/drivers/gpu/drm/amd/amdgpu/clearstate_gfx9.h
@@ -21,8 +21,7 @@
  *
  */
 
-static const unsigned int gfx9_SECT_CONTEXT_def_1[] =
-{
+static const unsigned int gfx9_SECT_CONTEXT_def_1[] = {
     0x00000000, // DB_RENDER_CONTROL
     0x00000000, // DB_COUNT_CONTROL
     0x00000000, // DB_DEPTH_VIEW
@@ -236,8 +235,7 @@ static const unsigned int gfx9_SECT_CONTEXT_def_1[] =
     0x00000000, // PA_SC_VPORT_ZMIN_15
     0x3f800000, // PA_SC_VPORT_ZMAX_15
 };
-static const unsigned int gfx9_SECT_CONTEXT_def_2[] =
-{
+static const unsigned int gfx9_SECT_CONTEXT_def_2[] = {
     0x00000000, // PA_SC_SCREEN_EXTENT_CONTROL
     0x00000000, // PA_SC_TILE_STEERING_OVERRIDE
     0x00000000, // CP_PERFMON_CNTX_CNTL
@@ -521,15 +519,13 @@ static const unsigned int gfx9_SECT_CONTEXT_def_2[] =
     0x00000000, // CB_MRT6_EPITCH
     0x00000000, // CB_MRT7_EPITCH
 };
-static const unsigned int gfx9_SECT_CONTEXT_def_3[] =
-{
+static const unsigned int gfx9_SECT_CONTEXT_def_3[] = {
     0x00000000, // PA_CL_POINT_X_RAD
     0x00000000, // PA_CL_POINT_Y_RAD
     0x00000000, // PA_CL_POINT_SIZE
     0x00000000, // PA_CL_POINT_CULL_RAD
 };
-static const unsigned int gfx9_SECT_CONTEXT_def_4[] =
-{
+static const unsigned int gfx9_SECT_CONTEXT_def_4[] = {
     0x00000000, // DB_DEPTH_CONTROL
     0x00000000, // DB_EQAA
     0x00000000, // CB_COLOR_CONTROL
@@ -688,17 +684,14 @@ static const unsigned int gfx9_SECT_CONTEXT_def_4[] =
     0x00000000, // VGT_GS_OUT_PRIM_TYPE
     0x00000000, // IA_ENHANCE
 };
-static const unsigned int gfx9_SECT_CONTEXT_def_5[] =
-{
+static const unsigned int gfx9_SECT_CONTEXT_def_5[] = {
     0x00000000, // WD_ENHANCE
     0x00000000, // VGT_PRIMITIVEID_EN
 };
-static const unsigned int gfx9_SECT_CONTEXT_def_6[] =
-{
+static const unsigned int gfx9_SECT_CONTEXT_def_6[] = {
     0x00000000, // VGT_PRIMITIVEID_RESET
 };
-static const unsigned int gfx9_SECT_CONTEXT_def_7[] =
-{
+static const unsigned int gfx9_SECT_CONTEXT_def_7[] = {
     0x00000000, // VGT_GS_MAX_PRIMS_PER_SUBGROUP
     0x00000000, // VGT_DRAW_PAYLOAD_CNTL
     0, // HOLE
@@ -766,8 +759,7 @@ static const unsigned int gfx9_SECT_CONTEXT_def_7[] =
     0x00000000, // VGT_STRMOUT_CONFIG
     0x00000000, // VGT_STRMOUT_BUFFER_CONFIG
 };
-static const unsigned int gfx9_SECT_CONTEXT_def_8[] =
-{
+static const unsigned int gfx9_SECT_CONTEXT_def_8[] = {
     0x00000000, // PA_SC_CENTROID_PRIORITY_0
     0x00000000, // PA_SC_CENTROID_PRIORITY_1
     0x00001000, // PA_SC_LINE_CNTL
@@ -924,8 +916,7 @@ static const unsigned int gfx9_SECT_CONTEXT_def_8[] =
     0x00000000, // CB_COLOR7_DCC_BASE
     0x00000000, // CB_COLOR7_DCC_BASE_EXT
 };
-static const struct cs_extent_def gfx9_SECT_CONTEXT_defs[] =
-{
+static const struct cs_extent_def gfx9_SECT_CONTEXT_defs[] = {
     {gfx9_SECT_CONTEXT_def_1, 0x0000a000, 212 },
     {gfx9_SECT_CONTEXT_def_2, 0x0000a0d6, 282 },
     {gfx9_SECT_CONTEXT_def_3, 0x0000a1f5, 4 },
diff --git a/drivers/gpu/drm/amd/amdgpu/clearstate_si.h b/drivers/gpu/drm/amd/amdgpu/clearstate_si.h
index 66e39cdb5cb0..5fd96ddd7f0f 100644
--- a/drivers/gpu/drm/amd/amdgpu/clearstate_si.h
+++ b/drivers/gpu/drm/amd/amdgpu/clearstate_si.h
@@ -21,8 +21,7 @@
  *
  */
 
-static const u32 si_SECT_CONTEXT_def_1[] =
-{
+static const u32 si_SECT_CONTEXT_def_1[] = {
     0x00000000, // DB_RENDER_CONTROL
     0x00000000, // DB_COUNT_CONTROL
     0x00000000, // DB_DEPTH_VIEW
@@ -236,8 +235,7 @@ static const u32 si_SECT_CONTEXT_def_1[] =
     0x00000000, // PA_SC_VPORT_ZMIN_15
     0x3f800000, // PA_SC_VPORT_ZMAX_15
 };
-static const u32 si_SECT_CONTEXT_def_2[] =
-{
+static const u32 si_SECT_CONTEXT_def_2[] = {
     0x00000000, // CP_PERFMON_CNTX_CNTL
     0x00000000, // CP_RINGID
     0x00000000, // CP_VMID
@@ -511,8 +509,7 @@ static const u32 si_SECT_CONTEXT_def_2[] =
     0x00000000, // CB_BLEND6_CONTROL
     0x00000000, // CB_BLEND7_CONTROL
 };
-static const u32 si_SECT_CONTEXT_def_3[] =
-{
+static const u32 si_SECT_CONTEXT_def_3[] = {
     0x00000000, // PA_CL_POINT_X_RAD
     0x00000000, // PA_CL_POINT_Y_RAD
     0x00000000, // PA_CL_POINT_SIZE
@@ -520,8 +517,7 @@ static const u32 si_SECT_CONTEXT_def_3[] =
     0x00000000, // VGT_DMA_BASE_HI
     0x00000000, // VGT_DMA_BASE
 };
-static const u32 si_SECT_CONTEXT_def_4[] =
-{
+static const u32 si_SECT_CONTEXT_def_4[] = {
     0x00000000, // DB_DEPTH_CONTROL
     0x00000000, // DB_EQAA
     0x00000000, // CB_COLOR_CONTROL
@@ -680,16 +676,13 @@ static const u32 si_SECT_CONTEXT_def_4[] =
     0x00000000, // VGT_GS_OUT_PRIM_TYPE
     0x00000000, // IA_ENHANCE
 };
-static const u32 si_SECT_CONTEXT_def_5[] =
-{
+static const u32 si_SECT_CONTEXT_def_5[] = {
     0x00000000, // VGT_PRIMITIVEID_EN
 };
-static const u32 si_SECT_CONTEXT_def_6[] =
-{
+static const u32 si_SECT_CONTEXT_def_6[] = {
     0x00000000, // VGT_PRIMITIVEID_RESET
 };
-static const u32 si_SECT_CONTEXT_def_7[] =
-{
+static const u32 si_SECT_CONTEXT_def_7[] = {
     0x00000000, // VGT_MULTI_PRIM_IB_RESET_EN
     0, // HOLE
     0, // HOLE
@@ -924,8 +917,7 @@ static const u32 si_SECT_CONTEXT_def_7[] =
     0x00000000, // CB_COLOR7_CLEAR_WORD0
     0x00000000, // CB_COLOR7_CLEAR_WORD1
 };
-static const struct cs_extent_def si_SECT_CONTEXT_defs[] =
-{
+static const struct cs_extent_def si_SECT_CONTEXT_defs[] = {
     {si_SECT_CONTEXT_def_1, 0x0000a000, 212 },
     {si_SECT_CONTEXT_def_2, 0x0000a0d8, 272 },
     {si_SECT_CONTEXT_def_3, 0x0000a1f5, 6 },
diff --git a/drivers/gpu/drm/amd/amdgpu/cyan_skillfish_reg_init.c b/drivers/gpu/drm/amd/amdgpu/cyan_skillfish_reg_init.c
new file mode 100644
index 000000000000..ed1e25661706
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/cyan_skillfish_reg_init.c
@@ -0,0 +1,56 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright 2018 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include "amdgpu.h"
+#include "nv.h"
+
+#include "soc15_common.h"
+#include "soc15_hw_ip.h"
+#include "cyan_skillfish_ip_offset.h"
+
+int cyan_skillfish_reg_base_init(struct amdgpu_device *adev)
+{
+	/* HW has more IP blocks,  only initialized the blocke needed by driver */
+	uint32_t i;
+
+	adev->gfx.xcc_mask = 1;
+	for (i = 0 ; i < MAX_INSTANCE ; ++i) {
+		adev->reg_offset[GC_HWIP][i] = (uint32_t *)(&(GC_BASE.instance[i]));
+		adev->reg_offset[HDP_HWIP][i] = (uint32_t *)(&(HDP_BASE.instance[i]));
+		adev->reg_offset[MMHUB_HWIP][i] = (uint32_t *)(&(MMHUB_BASE.instance[i]));
+		adev->reg_offset[ATHUB_HWIP][i] = (uint32_t *)(&(ATHUB_BASE.instance[i]));
+		adev->reg_offset[NBIO_HWIP][i] = (uint32_t *)(&(NBIO_BASE.instance[i]));
+		adev->reg_offset[MP0_HWIP][i] = (uint32_t *)(&(MP0_BASE.instance[i]));
+		adev->reg_offset[MP1_HWIP][i] = (uint32_t *)(&(MP1_BASE.instance[i]));
+		adev->reg_offset[VCN_HWIP][i] = (uint32_t *)(&(UVD0_BASE.instance[i]));
+		adev->reg_offset[DF_HWIP][i] = (uint32_t *)(&(DF_BASE.instance[i]));
+		adev->reg_offset[DCE_HWIP][i] = (uint32_t *)(&(DMU_BASE.instance[i]));
+		adev->reg_offset[OSSSYS_HWIP][i] = (uint32_t *)(&(OSSSYS_BASE.instance[i]));
+		adev->reg_offset[SDMA0_HWIP][i] = (uint32_t *)(&(GC_BASE.instance[i]));
+		adev->reg_offset[SDMA1_HWIP][i] = (uint32_t *)(&(GC_BASE.instance[i]));
+		adev->reg_offset[SMUIO_HWIP][i] = (uint32_t *)(&(SMUIO_BASE.instance[i]));
+		adev->reg_offset[THM_HWIP][i] = (uint32_t *)(&(THM_BASE.instance[i]));
+		adev->reg_offset[CLK_HWIP][i] = (uint32_t *)(&(CLK_BASE.instance[i]));
+	}
+	return 0;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/cz_ih.c b/drivers/gpu/drm/amd/amdgpu/cz_ih.c
index b8c47e0cf37a..bc7a2e06ab5f 100644
--- a/drivers/gpu/drm/amd/amdgpu/cz_ih.c
+++ b/drivers/gpu/drm/amd/amdgpu/cz_ih.c
@@ -157,6 +157,9 @@ static int cz_ih_irq_init(struct amdgpu_device *adev)
 	/* enable interrupts */
 	cz_ih_enable_interrupts(adev);
 
+	if (adev->irq.ih_soft.ring_size)
+		adev->irq.ih_soft.enabled = true;
+
 	return 0;
 }
 
@@ -194,6 +197,9 @@ static u32 cz_ih_get_wptr(struct amdgpu_device *adev,
 
 	wptr = le32_to_cpu(*ih->wptr_cpu);
 
+	if (ih == &adev->irq.ih_soft)
+		goto out;
+
 	if (!REG_GET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW))
 		goto out;
 
@@ -216,6 +222,11 @@ static u32 cz_ih_get_wptr(struct amdgpu_device *adev,
 	tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 1);
 	WREG32(mmIH_RB_CNTL, tmp);
 
+	/* Unset the CLEAR_OVERFLOW bit immediately so new overflows
+	 * can be detected.
+	 */
+	tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 0);
+	WREG32(mmIH_RB_CNTL, tmp);
 
 out:
 	return (wptr & ih->ptr_mask);
@@ -269,9 +280,9 @@ static void cz_ih_set_rptr(struct amdgpu_device *adev,
 	WREG32(mmIH_RB_RPTR, ih->rptr);
 }
 
-static int cz_ih_early_init(void *handle)
+static int cz_ih_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int ret;
 
 	ret = amdgpu_irq_add_domain(adev);
@@ -283,23 +294,27 @@ static int cz_ih_early_init(void *handle)
 	return 0;
 }
 
-static int cz_ih_sw_init(void *handle)
+static int cz_ih_sw_init(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	r = amdgpu_ih_ring_init(adev, &adev->irq.ih, 64 * 1024, false);
 	if (r)
 		return r;
 
+	r = amdgpu_ih_ring_init(adev, &adev->irq.ih_soft, IH_SW_RING_SIZE, true);
+	if (r)
+		return r;
+
 	r = amdgpu_irq_init(adev);
 
 	return r;
 }
 
-static int cz_ih_sw_fini(void *handle)
+static int cz_ih_sw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	amdgpu_irq_fini_sw(adev);
 	amdgpu_irq_remove_domain(adev);
@@ -307,10 +322,10 @@ static int cz_ih_sw_fini(void *handle)
 	return 0;
 }
 
-static int cz_ih_hw_init(void *handle)
+static int cz_ih_hw_init(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	r = cz_ih_irq_init(adev);
 	if (r)
@@ -319,32 +334,26 @@ static int cz_ih_hw_init(void *handle)
 	return 0;
 }
 
-static int cz_ih_hw_fini(void *handle)
+static int cz_ih_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	cz_ih_irq_disable(adev);
+	cz_ih_irq_disable(ip_block->adev);
 
 	return 0;
 }
 
-static int cz_ih_suspend(void *handle)
+static int cz_ih_suspend(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	return cz_ih_hw_fini(adev);
+	return cz_ih_hw_fini(ip_block);
 }
 
-static int cz_ih_resume(void *handle)
+static int cz_ih_resume(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	return cz_ih_hw_init(adev);
+	return cz_ih_hw_init(ip_block);
 }
 
-static bool cz_ih_is_idle(void *handle)
+static bool cz_ih_is_idle(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	u32 tmp = RREG32(mmSRBM_STATUS);
 
 	if (REG_GET_FIELD(tmp, SRBM_STATUS, IH_BUSY))
@@ -353,11 +362,11 @@ static bool cz_ih_is_idle(void *handle)
 	return true;
 }
 
-static int cz_ih_wait_for_idle(void *handle)
+static int cz_ih_wait_for_idle(struct amdgpu_ip_block *ip_block)
 {
 	unsigned i;
 	u32 tmp;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	for (i = 0; i < adev->usec_timeout; i++) {
 		/* read MC_STATUS */
@@ -369,10 +378,10 @@ static int cz_ih_wait_for_idle(void *handle)
 	return -ETIMEDOUT;
 }
 
-static int cz_ih_soft_reset(void *handle)
+static int cz_ih_soft_reset(struct amdgpu_ip_block *ip_block)
 {
 	u32 srbm_soft_reset = 0;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	u32 tmp = RREG32(mmSRBM_STATUS);
 
 	if (tmp & SRBM_STATUS__IH_BUSY_MASK)
@@ -399,14 +408,14 @@ static int cz_ih_soft_reset(void *handle)
 	return 0;
 }
 
-static int cz_ih_set_clockgating_state(void *handle,
+static int cz_ih_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_clockgating_state state)
 {
 	// TODO
 	return 0;
 }
 
-static int cz_ih_set_powergating_state(void *handle,
+static int cz_ih_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_powergating_state state)
 {
 	// TODO
@@ -416,7 +425,6 @@ static int cz_ih_set_powergating_state(void *handle,
 static const struct amd_ip_funcs cz_ih_ip_funcs = {
 	.name = "cz_ih",
 	.early_init = cz_ih_early_init,
-	.late_init = NULL,
 	.sw_init = cz_ih_sw_init,
 	.sw_fini = cz_ih_sw_fini,
 	.hw_init = cz_ih_hw_init,
diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c
index d1570a462a51..72ca6538b2e4 100644
--- a/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c
@@ -21,7 +21,10 @@
  *
  */
 
+#include <drm/drm_edid.h>
 #include <drm/drm_fourcc.h>
+#include <drm/drm_modeset_helper.h>
+#include <drm/drm_modeset_helper_vtables.h>
 #include <drm/drm_vblank.h>
 
 #include "amdgpu.h"
@@ -49,9 +52,9 @@
 
 static void dce_v10_0_set_display_funcs(struct amdgpu_device *adev);
 static void dce_v10_0_set_irq_funcs(struct amdgpu_device *adev);
+static void dce_v10_0_hpd_int_ack(struct amdgpu_device *adev, int hpd);
 
-static const u32 crtc_offsets[] =
-{
+static const u32 crtc_offsets[] = {
 	CRTC0_REGISTER_OFFSET,
 	CRTC1_REGISTER_OFFSET,
 	CRTC2_REGISTER_OFFSET,
@@ -61,8 +64,7 @@ static const u32 crtc_offsets[] =
 	CRTC6_REGISTER_OFFSET
 };
 
-static const u32 hpd_offsets[] =
-{
+static const u32 hpd_offsets[] = {
 	HPD0_REGISTER_OFFSET,
 	HPD1_REGISTER_OFFSET,
 	HPD2_REGISTER_OFFSET,
@@ -119,30 +121,26 @@ static const struct {
 	.hpd = DISP_INTERRUPT_STATUS_CONTINUE5__DC_HPD6_INTERRUPT_MASK
 } };
 
-static const u32 golden_settings_tonga_a11[] =
-{
+static const u32 golden_settings_tonga_a11[] = {
 	mmDCI_CLK_CNTL, 0x00000080, 0x00000000,
 	mmFBC_DEBUG_COMP, 0x000000f0, 0x00000070,
 	mmFBC_MISC, 0x1f311fff, 0x12300000,
 	mmHDMI_CONTROL, 0x31000111, 0x00000011,
 };
 
-static const u32 tonga_mgcg_cgcg_init[] =
-{
+static const u32 tonga_mgcg_cgcg_init[] = {
 	mmXDMA_CLOCK_GATING_CNTL, 0xffffffff, 0x00000100,
 	mmXDMA_MEM_POWER_CNTL, 0x00000101, 0x00000000,
 };
 
-static const u32 golden_settings_fiji_a10[] =
-{
+static const u32 golden_settings_fiji_a10[] = {
 	mmDCI_CLK_CNTL, 0x00000080, 0x00000000,
 	mmFBC_DEBUG_COMP, 0x000000f0, 0x00000070,
 	mmFBC_MISC, 0x1f311fff, 0x12300000,
 	mmHDMI_CONTROL, 0x31000111, 0x00000011,
 };
 
-static const u32 fiji_mgcg_cgcg_init[] =
-{
+static const u32 fiji_mgcg_cgcg_init[] = {
 	mmXDMA_CLOCK_GATING_CNTL, 0xffffffff, 0x00000100,
 	mmXDMA_MEM_POWER_CNTL, 0x00000101, 0x00000000,
 };
@@ -367,6 +365,7 @@ static void dce_v10_0_hpd_init(struct amdgpu_device *adev)
 				    AMDGPU_HPD_DISCONNECT_INT_DELAY_IN_MS);
 		WREG32(mmDC_HPD_TOGGLE_FILT_CNTL + hpd_offsets[amdgpu_connector->hpd.hpd], tmp);
 
+		dce_v10_0_hpd_int_ack(adev, amdgpu_connector->hpd.hpd);
 		dce_v10_0_hpd_set_polarity(adev, amdgpu_connector->hpd.hpd);
 		amdgpu_irq_get(adev, &adev->hpd_irq,
 			       amdgpu_connector->hpd.hpd);
@@ -1040,7 +1039,7 @@ static void dce_v10_0_program_watermarks(struct amdgpu_device *adev,
 					    (u32)mode->clock);
 		line_time = (u32) div_u64((u64)mode->crtc_htotal * 1000000,
 					  (u32)mode->clock);
-		line_time = min(line_time, (u32)65535);
+		line_time = min_t(u32, line_time, 65535);
 
 		/* watermark for high clocks */
 		if (adev->pm.dpm_enabled) {
@@ -1070,7 +1069,7 @@ static void dce_v10_0_program_watermarks(struct amdgpu_device *adev,
 		wm_high.num_heads = num_heads;
 
 		/* set for high clocks */
-		latency_watermark_a = min(dce_v10_0_latency_watermark(&wm_high), (u32)65535);
+		latency_watermark_a = min_t(u32, dce_v10_0_latency_watermark(&wm_high), 65535);
 
 		/* possibly force display priority to high */
 		/* should really do this at mode validation time... */
@@ -1109,7 +1108,7 @@ static void dce_v10_0_program_watermarks(struct amdgpu_device *adev,
 		wm_low.num_heads = num_heads;
 
 		/* set for low clocks */
-		latency_watermark_b = min(dce_v10_0_latency_watermark(&wm_low), (u32)65535);
+		latency_watermark_b = min_t(u32, dce_v10_0_latency_watermark(&wm_low), 65535);
 
 		/* possibly force display priority to high */
 		/* should really do this at mode validation time... */
@@ -1142,8 +1141,7 @@ static void dce_v10_0_program_watermarks(struct amdgpu_device *adev,
 
 	/* save values for DPM */
 	amdgpu_crtc->line_time = line_time;
-	amdgpu_crtc->wm_high = latency_watermark_a;
-	amdgpu_crtc->wm_low = latency_watermark_b;
+
 	/* Save number of lines the linebuffer leads before the scanout */
 	amdgpu_crtc->lb_vblank_lead_lines = lb_vblank_lead_lines;
 }
@@ -1300,7 +1298,7 @@ static void dce_v10_0_audio_write_speaker_allocation(struct drm_encoder *encoder
 		return;
 	}
 
-	sad_count = drm_edid_to_speaker_allocation(amdgpu_connector_edid(connector), &sadb);
+	sad_count = drm_edid_to_speaker_allocation(amdgpu_connector->edid, &sadb);
 	if (sad_count < 0) {
 		DRM_ERROR("Couldn't read Speaker Allocation Data Block: %d\n", sad_count);
 		sad_count = 0;
@@ -1370,7 +1368,7 @@ static void dce_v10_0_audio_write_sad_regs(struct drm_encoder *encoder)
 		return;
 	}
 
-	sad_count = drm_edid_to_sad(amdgpu_connector_edid(connector), &sads);
+	sad_count = drm_edid_to_sad(amdgpu_connector->edid, &sads);
 	if (sad_count < 0)
 		DRM_ERROR("Couldn't read SADs: %d\n", sad_count);
 	if (sad_count <= 0)
@@ -1423,8 +1421,7 @@ static void dce_v10_0_audio_enable(struct amdgpu_device *adev,
 			   enable ? AZALIA_F0_CODEC_PIN_CONTROL_HOT_PLUG_CONTROL__AUDIO_ENABLED_MASK : 0);
 }
 
-static const u32 pin_offsets[] =
-{
+static const u32 pin_offsets[] = {
 	AUD0_REGISTER_OFFSET,
 	AUD1_REGISTER_OFFSET,
 	AUD2_REGISTER_OFFSET,
@@ -1464,17 +1461,12 @@ static int dce_v10_0_audio_init(struct amdgpu_device *adev)
 
 static void dce_v10_0_audio_fini(struct amdgpu_device *adev)
 {
-	int i;
-
 	if (!amdgpu_audio)
 		return;
 
 	if (!adev->mode_info.audio.enabled)
 		return;
 
-	for (i = 0; i < adev->mode_info.audio.num_pins; i++)
-		dce_v10_0_audio_enable(adev, &adev->mode_info.audio.pin[i], false);
-
 	adev->mode_info.audio.enabled = false;
 }
 
@@ -1809,8 +1801,7 @@ static void dce_v10_0_afmt_fini(struct amdgpu_device *adev)
 	}
 }
 
-static const u32 vga_control_regs[6] =
-{
+static const u32 vga_control_regs[6] = {
 	mmD1VGA_CONTROL,
 	mmD2VGA_CONTROL,
 	mmD3VGA_CONTROL,
@@ -1884,6 +1875,7 @@ static int dce_v10_0_crtc_do_set_base(struct drm_crtc *crtc,
 		return r;
 
 	if (!atomic) {
+		abo->flags |= AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS;
 		r = amdgpu_bo_pin(abo, AMDGPU_GEM_DOMAIN_VRAM);
 		if (unlikely(r != 0)) {
 			amdgpu_bo_unreserve(abo);
@@ -2404,6 +2396,7 @@ static int dce_v10_0_crtc_cursor_set2(struct drm_crtc *crtc,
 		return ret;
 	}
 
+	aobj->flags |= AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS;
 	ret = amdgpu_bo_pin(aobj, AMDGPU_GEM_DOMAIN_VRAM);
 	amdgpu_bo_unreserve(aobj);
 	if (ret) {
@@ -2532,7 +2525,7 @@ static void dce_v10_0_crtc_dpms(struct drm_crtc *crtc, int mode)
 		break;
 	}
 	/* adjust pm to dpms */
-	amdgpu_pm_compute_clocks(adev);
+	amdgpu_dpm_compute_clocks(adev);
 }
 
 static void dce_v10_0_crtc_prepare(struct drm_crtc *crtc)
@@ -2688,6 +2681,32 @@ static const struct drm_crtc_helper_funcs dce_v10_0_crtc_helper_funcs = {
 	.get_scanout_position = amdgpu_crtc_get_scanout_position,
 };
 
+static void dce_v10_0_panic_flush(struct drm_plane *plane)
+{
+	struct drm_framebuffer *fb;
+	struct amdgpu_crtc *amdgpu_crtc;
+	struct amdgpu_device *adev;
+	uint32_t fb_format;
+
+	if (!plane->fb)
+		return;
+
+	fb = plane->fb;
+	amdgpu_crtc = to_amdgpu_crtc(plane->crtc);
+	adev = drm_to_adev(fb->dev);
+
+	/* Disable DC tiling */
+	fb_format = RREG32(mmGRPH_CONTROL + amdgpu_crtc->crtc_offset);
+	fb_format &= ~GRPH_CONTROL__GRPH_ARRAY_MODE_MASK;
+	WREG32(mmGRPH_CONTROL + amdgpu_crtc->crtc_offset, fb_format);
+
+}
+
+static const struct drm_plane_helper_funcs dce_v10_0_drm_primary_plane_helper_funcs = {
+	.get_scanout_buffer = amdgpu_display_get_scanout_buffer,
+	.panic_flush = dce_v10_0_panic_flush,
+};
+
 static int dce_v10_0_crtc_init(struct amdgpu_device *adev, int index)
 {
 	struct amdgpu_crtc *amdgpu_crtc;
@@ -2735,13 +2754,14 @@ static int dce_v10_0_crtc_init(struct amdgpu_device *adev, int index)
 	amdgpu_crtc->encoder = NULL;
 	amdgpu_crtc->connector = NULL;
 	drm_crtc_helper_add(&amdgpu_crtc->base, &dce_v10_0_crtc_helper_funcs);
+	drm_plane_helper_add(amdgpu_crtc->base.primary, &dce_v10_0_drm_primary_plane_helper_funcs);
 
 	return 0;
 }
 
-static int dce_v10_0_early_init(void *handle)
+static int dce_v10_0_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	adev->audio_endpt_rreg = &dce_v10_0_audio_endpt_rreg;
 	adev->audio_endpt_wreg = &dce_v10_0_audio_endpt_wreg;
@@ -2766,10 +2786,10 @@ static int dce_v10_0_early_init(void *handle)
 	return 0;
 }
 
-static int dce_v10_0_sw_init(void *handle)
+static int dce_v10_0_sw_init(struct amdgpu_ip_block *ip_block)
 {
 	int r, i;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	for (i = 0; i < adev->mode_info.num_crtc; i++) {
 		r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, i + 1, &adev->crtc_irq);
@@ -2798,7 +2818,7 @@ static int dce_v10_0_sw_init(void *handle)
 	adev_to_drm(adev)->mode_config.preferred_depth = 24;
 	adev_to_drm(adev)->mode_config.prefer_shadow = 1;
 
-	adev_to_drm(adev)->mode_config.fb_base = adev->gmc.aper_base;
+	adev_to_drm(adev)->mode_config.fb_modifiers_not_supported = true;
 
 	r = amdgpu_display_modeset_create_props(adev);
 	if (r)
@@ -2828,17 +2848,28 @@ static int dce_v10_0_sw_init(void *handle)
 	if (r)
 		return r;
 
+	/* Disable vblank IRQs aggressively for power-saving */
+	/* XXX: can this be enabled for DC? */
+	adev_to_drm(adev)->vblank_disable_immediate = true;
+
+	r = drm_vblank_init(adev_to_drm(adev), adev->mode_info.num_crtc);
+	if (r)
+		return r;
+
+	INIT_DELAYED_WORK(&adev->hotplug_work,
+		  amdgpu_display_hotplug_work_func);
+
 	drm_kms_helper_poll_init(adev_to_drm(adev));
 
 	adev->mode_info.mode_config_initialized = true;
 	return 0;
 }
 
-static int dce_v10_0_sw_fini(void *handle)
+static int dce_v10_0_sw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
-	kfree(adev->mode_info.bios_hardcoded_edid);
+	drm_edid_free(adev->mode_info.bios_hardcoded_edid);
 
 	drm_kms_helper_poll_fini(adev_to_drm(adev));
 
@@ -2852,10 +2883,10 @@ static int dce_v10_0_sw_fini(void *handle)
 	return 0;
 }
 
-static int dce_v10_0_hw_init(void *handle)
+static int dce_v10_0_hw_init(struct amdgpu_ip_block *ip_block)
 {
 	int i;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	dce_v10_0_init_golden_registers(adev);
 
@@ -2877,10 +2908,10 @@ static int dce_v10_0_hw_init(void *handle)
 	return 0;
 }
 
-static int dce_v10_0_hw_fini(void *handle)
+static int dce_v10_0_hw_fini(struct amdgpu_ip_block *ip_block)
 {
 	int i;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	dce_v10_0_hpd_fini(adev);
 
@@ -2890,12 +2921,14 @@ static int dce_v10_0_hw_fini(void *handle)
 
 	dce_v10_0_pageflip_interrupt_fini(adev);
 
+	flush_delayed_work(&adev->hotplug_work);
+
 	return 0;
 }
 
-static int dce_v10_0_suspend(void *handle)
+static int dce_v10_0_suspend(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int r;
 
 	r = amdgpu_display_suspend_helper(adev);
@@ -2905,18 +2938,18 @@ static int dce_v10_0_suspend(void *handle)
 	adev->mode_info.bl_level =
 		amdgpu_atombios_encoder_get_backlight_level_from_reg(adev);
 
-	return dce_v10_0_hw_fini(handle);
+	return dce_v10_0_hw_fini(ip_block);
 }
 
-static int dce_v10_0_resume(void *handle)
+static int dce_v10_0_resume(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int ret;
 
 	amdgpu_atombios_encoder_set_backlight_level_to_reg(adev,
 							   adev->mode_info.bl_level);
 
-	ret = dce_v10_0_hw_init(handle);
+	ret = dce_v10_0_hw_init(ip_block);
 
 	/* turn on the BL */
 	if (adev->mode_info.bl_encoder) {
@@ -2931,27 +2964,22 @@ static int dce_v10_0_resume(void *handle)
 	return amdgpu_display_resume_helper(adev);
 }
 
-static bool dce_v10_0_is_idle(void *handle)
+static bool dce_v10_0_is_idle(struct amdgpu_ip_block *ip_block)
 {
 	return true;
 }
 
-static int dce_v10_0_wait_for_idle(void *handle)
+static bool dce_v10_0_check_soft_reset(struct amdgpu_ip_block *ip_block)
 {
-	return 0;
-}
-
-static bool dce_v10_0_check_soft_reset(void *handle)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	return dce_v10_0_is_display_hung(adev);
 }
 
-static int dce_v10_0_soft_reset(void *handle)
+static int dce_v10_0_soft_reset(struct amdgpu_ip_block *ip_block)
 {
 	u32 srbm_soft_reset = 0, tmp;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (dce_v10_0_is_display_hung(adev))
 		srbm_soft_reset |= SRBM_SOFT_RESET__SOFT_RESET_DC_MASK;
@@ -3041,7 +3069,7 @@ static int dce_v10_0_set_hpd_irq_state(struct amdgpu_device *adev,
 	u32 tmp;
 
 	if (hpd >= adev->mode_info.num_hpd) {
-		DRM_DEBUG("invalid hdp %d\n", hpd);
+		DRM_DEBUG("invalid hpd %d\n", hpd);
 		return 0;
 	}
 
@@ -3193,7 +3221,7 @@ static void dce_v10_0_hpd_int_ack(struct amdgpu_device *adev,
 	u32 tmp;
 
 	if (hpd >= adev->mode_info.num_hpd) {
-		DRM_DEBUG("invalid hdp %d\n", hpd);
+		DRM_DEBUG("invalid hpd %d\n", hpd);
 		return;
 	}
 
@@ -3288,20 +3316,20 @@ static int dce_v10_0_hpd_irq(struct amdgpu_device *adev,
 
 	if (disp_int & mask) {
 		dce_v10_0_hpd_int_ack(adev, hpd);
-		schedule_work(&adev->hotplug_work);
+		schedule_delayed_work(&adev->hotplug_work, 0);
 		DRM_DEBUG("IH: HPD%d\n", hpd + 1);
 	}
 
 	return 0;
 }
 
-static int dce_v10_0_set_clockgating_state(void *handle,
+static int dce_v10_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_clockgating_state state)
 {
 	return 0;
 }
 
-static int dce_v10_0_set_powergating_state(void *handle,
+static int dce_v10_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_powergating_state state)
 {
 	return 0;
@@ -3310,7 +3338,6 @@ static int dce_v10_0_set_powergating_state(void *handle,
 static const struct amd_ip_funcs dce_v10_0_ip_funcs = {
 	.name = "dce_v10_0",
 	.early_init = dce_v10_0_early_init,
-	.late_init = NULL,
 	.sw_init = dce_v10_0_sw_init,
 	.sw_fini = dce_v10_0_sw_fini,
 	.hw_init = dce_v10_0_hw_init,
@@ -3318,7 +3345,6 @@ static const struct amd_ip_funcs dce_v10_0_ip_funcs = {
 	.suspend = dce_v10_0_suspend,
 	.resume = dce_v10_0_resume,
 	.is_idle = dce_v10_0_is_idle,
-	.wait_for_idle = dce_v10_0_wait_for_idle,
 	.check_soft_reset = dce_v10_0_check_soft_reset,
 	.soft_reset = dce_v10_0_soft_reset,
 	.set_clockgating_state = dce_v10_0_set_clockgating_state,
@@ -3636,8 +3662,7 @@ static void dce_v10_0_set_irq_funcs(struct amdgpu_device *adev)
 	adev->hpd_irq.funcs = &dce_v10_0_hpd_irq_funcs;
 }
 
-const struct amdgpu_ip_block_version dce_v10_0_ip_block =
-{
+const struct amdgpu_ip_block_version dce_v10_0_ip_block = {
 	.type = AMD_IP_BLOCK_TYPE_DCE,
 	.major = 10,
 	.minor = 0,
@@ -3645,8 +3670,7 @@ const struct amdgpu_ip_block_version dce_v10_0_ip_block =
 	.funcs = &dce_v10_0_ip_funcs,
 };
 
-const struct amdgpu_ip_block_version dce_v10_1_ip_block =
-{
+const struct amdgpu_ip_block_version dce_v10_1_ip_block = {
 	.type = AMD_IP_BLOCK_TYPE_DCE,
 	.major = 10,
 	.minor = 1,
diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c
deleted file mode 100644
index 18a7b3bd633b..000000000000
--- a/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c
+++ /dev/null
@@ -1,3784 +0,0 @@
-/*
- * Copyright 2014 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- */
-
-#include <drm/drm_fourcc.h>
-#include <drm/drm_vblank.h>
-
-#include "amdgpu.h"
-#include "amdgpu_pm.h"
-#include "amdgpu_i2c.h"
-#include "vid.h"
-#include "atom.h"
-#include "amdgpu_atombios.h"
-#include "atombios_crtc.h"
-#include "atombios_encoders.h"
-#include "amdgpu_pll.h"
-#include "amdgpu_connectors.h"
-#include "amdgpu_display.h"
-#include "dce_v11_0.h"
-
-#include "dce/dce_11_0_d.h"
-#include "dce/dce_11_0_sh_mask.h"
-#include "dce/dce_11_0_enum.h"
-#include "oss/oss_3_0_d.h"
-#include "oss/oss_3_0_sh_mask.h"
-#include "gmc/gmc_8_1_d.h"
-#include "gmc/gmc_8_1_sh_mask.h"
-
-#include "ivsrcid/ivsrcid_vislands30.h"
-
-static void dce_v11_0_set_display_funcs(struct amdgpu_device *adev);
-static void dce_v11_0_set_irq_funcs(struct amdgpu_device *adev);
-
-static const u32 crtc_offsets[] =
-{
-	CRTC0_REGISTER_OFFSET,
-	CRTC1_REGISTER_OFFSET,
-	CRTC2_REGISTER_OFFSET,
-	CRTC3_REGISTER_OFFSET,
-	CRTC4_REGISTER_OFFSET,
-	CRTC5_REGISTER_OFFSET,
-	CRTC6_REGISTER_OFFSET
-};
-
-static const u32 hpd_offsets[] =
-{
-	HPD0_REGISTER_OFFSET,
-	HPD1_REGISTER_OFFSET,
-	HPD2_REGISTER_OFFSET,
-	HPD3_REGISTER_OFFSET,
-	HPD4_REGISTER_OFFSET,
-	HPD5_REGISTER_OFFSET
-};
-
-static const uint32_t dig_offsets[] = {
-	DIG0_REGISTER_OFFSET,
-	DIG1_REGISTER_OFFSET,
-	DIG2_REGISTER_OFFSET,
-	DIG3_REGISTER_OFFSET,
-	DIG4_REGISTER_OFFSET,
-	DIG5_REGISTER_OFFSET,
-	DIG6_REGISTER_OFFSET,
-	DIG7_REGISTER_OFFSET,
-	DIG8_REGISTER_OFFSET
-};
-
-static const struct {
-	uint32_t        reg;
-	uint32_t        vblank;
-	uint32_t        vline;
-	uint32_t        hpd;
-
-} interrupt_status_offsets[] = { {
-	.reg = mmDISP_INTERRUPT_STATUS,
-	.vblank = DISP_INTERRUPT_STATUS__LB_D1_VBLANK_INTERRUPT_MASK,
-	.vline = DISP_INTERRUPT_STATUS__LB_D1_VLINE_INTERRUPT_MASK,
-	.hpd = DISP_INTERRUPT_STATUS__DC_HPD1_INTERRUPT_MASK
-}, {
-	.reg = mmDISP_INTERRUPT_STATUS_CONTINUE,
-	.vblank = DISP_INTERRUPT_STATUS_CONTINUE__LB_D2_VBLANK_INTERRUPT_MASK,
-	.vline = DISP_INTERRUPT_STATUS_CONTINUE__LB_D2_VLINE_INTERRUPT_MASK,
-	.hpd = DISP_INTERRUPT_STATUS_CONTINUE__DC_HPD2_INTERRUPT_MASK
-}, {
-	.reg = mmDISP_INTERRUPT_STATUS_CONTINUE2,
-	.vblank = DISP_INTERRUPT_STATUS_CONTINUE2__LB_D3_VBLANK_INTERRUPT_MASK,
-	.vline = DISP_INTERRUPT_STATUS_CONTINUE2__LB_D3_VLINE_INTERRUPT_MASK,
-	.hpd = DISP_INTERRUPT_STATUS_CONTINUE2__DC_HPD3_INTERRUPT_MASK
-}, {
-	.reg = mmDISP_INTERRUPT_STATUS_CONTINUE3,
-	.vblank = DISP_INTERRUPT_STATUS_CONTINUE3__LB_D4_VBLANK_INTERRUPT_MASK,
-	.vline = DISP_INTERRUPT_STATUS_CONTINUE3__LB_D4_VLINE_INTERRUPT_MASK,
-	.hpd = DISP_INTERRUPT_STATUS_CONTINUE3__DC_HPD4_INTERRUPT_MASK
-}, {
-	.reg = mmDISP_INTERRUPT_STATUS_CONTINUE4,
-	.vblank = DISP_INTERRUPT_STATUS_CONTINUE4__LB_D5_VBLANK_INTERRUPT_MASK,
-	.vline = DISP_INTERRUPT_STATUS_CONTINUE4__LB_D5_VLINE_INTERRUPT_MASK,
-	.hpd = DISP_INTERRUPT_STATUS_CONTINUE4__DC_HPD5_INTERRUPT_MASK
-}, {
-	.reg = mmDISP_INTERRUPT_STATUS_CONTINUE5,
-	.vblank = DISP_INTERRUPT_STATUS_CONTINUE5__LB_D6_VBLANK_INTERRUPT_MASK,
-	.vline = DISP_INTERRUPT_STATUS_CONTINUE5__LB_D6_VLINE_INTERRUPT_MASK,
-	.hpd = DISP_INTERRUPT_STATUS_CONTINUE5__DC_HPD6_INTERRUPT_MASK
-} };
-
-static const u32 cz_golden_settings_a11[] =
-{
-	mmCRTC_DOUBLE_BUFFER_CONTROL, 0x00010101, 0x00010000,
-	mmFBC_MISC, 0x1f311fff, 0x14300000,
-};
-
-static const u32 cz_mgcg_cgcg_init[] =
-{
-	mmXDMA_CLOCK_GATING_CNTL, 0xffffffff, 0x00000100,
-	mmXDMA_MEM_POWER_CNTL, 0x00000101, 0x00000000,
-};
-
-static const u32 stoney_golden_settings_a11[] =
-{
-	mmCRTC_DOUBLE_BUFFER_CONTROL, 0x00010101, 0x00010000,
-	mmFBC_MISC, 0x1f311fff, 0x14302000,
-};
-
-static const u32 polaris11_golden_settings_a11[] =
-{
-	mmDCI_CLK_CNTL, 0x00000080, 0x00000000,
-	mmFBC_DEBUG_COMP, 0x000000f0, 0x00000070,
-	mmFBC_DEBUG1, 0xffffffff, 0x00000008,
-	mmFBC_MISC, 0x9f313fff, 0x14302008,
-	mmHDMI_CONTROL, 0x313f031f, 0x00000011,
-};
-
-static const u32 polaris10_golden_settings_a11[] =
-{
-	mmDCI_CLK_CNTL, 0x00000080, 0x00000000,
-	mmFBC_DEBUG_COMP, 0x000000f0, 0x00000070,
-	mmFBC_MISC, 0x9f313fff, 0x14302008,
-	mmHDMI_CONTROL, 0x313f031f, 0x00000011,
-};
-
-static void dce_v11_0_init_golden_registers(struct amdgpu_device *adev)
-{
-	switch (adev->asic_type) {
-	case CHIP_CARRIZO:
-		amdgpu_device_program_register_sequence(adev,
-							cz_mgcg_cgcg_init,
-							ARRAY_SIZE(cz_mgcg_cgcg_init));
-		amdgpu_device_program_register_sequence(adev,
-							cz_golden_settings_a11,
-							ARRAY_SIZE(cz_golden_settings_a11));
-		break;
-	case CHIP_STONEY:
-		amdgpu_device_program_register_sequence(adev,
-							stoney_golden_settings_a11,
-							ARRAY_SIZE(stoney_golden_settings_a11));
-		break;
-	case CHIP_POLARIS11:
-	case CHIP_POLARIS12:
-		amdgpu_device_program_register_sequence(adev,
-							polaris11_golden_settings_a11,
-							ARRAY_SIZE(polaris11_golden_settings_a11));
-		break;
-	case CHIP_POLARIS10:
-	case CHIP_VEGAM:
-		amdgpu_device_program_register_sequence(adev,
-							polaris10_golden_settings_a11,
-							ARRAY_SIZE(polaris10_golden_settings_a11));
-		break;
-	default:
-		break;
-	}
-}
-
-static u32 dce_v11_0_audio_endpt_rreg(struct amdgpu_device *adev,
-				     u32 block_offset, u32 reg)
-{
-	unsigned long flags;
-	u32 r;
-
-	spin_lock_irqsave(&adev->audio_endpt_idx_lock, flags);
-	WREG32(mmAZALIA_F0_CODEC_ENDPOINT_INDEX + block_offset, reg);
-	r = RREG32(mmAZALIA_F0_CODEC_ENDPOINT_DATA + block_offset);
-	spin_unlock_irqrestore(&adev->audio_endpt_idx_lock, flags);
-
-	return r;
-}
-
-static void dce_v11_0_audio_endpt_wreg(struct amdgpu_device *adev,
-				      u32 block_offset, u32 reg, u32 v)
-{
-	unsigned long flags;
-
-	spin_lock_irqsave(&adev->audio_endpt_idx_lock, flags);
-	WREG32(mmAZALIA_F0_CODEC_ENDPOINT_INDEX + block_offset, reg);
-	WREG32(mmAZALIA_F0_CODEC_ENDPOINT_DATA + block_offset, v);
-	spin_unlock_irqrestore(&adev->audio_endpt_idx_lock, flags);
-}
-
-static u32 dce_v11_0_vblank_get_counter(struct amdgpu_device *adev, int crtc)
-{
-	if (crtc < 0 || crtc >= adev->mode_info.num_crtc)
-		return 0;
-	else
-		return RREG32(mmCRTC_STATUS_FRAME_COUNT + crtc_offsets[crtc]);
-}
-
-static void dce_v11_0_pageflip_interrupt_init(struct amdgpu_device *adev)
-{
-	unsigned i;
-
-	/* Enable pflip interrupts */
-	for (i = 0; i < adev->mode_info.num_crtc; i++)
-		amdgpu_irq_get(adev, &adev->pageflip_irq, i);
-}
-
-static void dce_v11_0_pageflip_interrupt_fini(struct amdgpu_device *adev)
-{
-	unsigned i;
-
-	/* Disable pflip interrupts */
-	for (i = 0; i < adev->mode_info.num_crtc; i++)
-		amdgpu_irq_put(adev, &adev->pageflip_irq, i);
-}
-
-/**
- * dce_v11_0_page_flip - pageflip callback.
- *
- * @adev: amdgpu_device pointer
- * @crtc_id: crtc to cleanup pageflip on
- * @crtc_base: new address of the crtc (GPU MC address)
- * @async: asynchronous flip
- *
- * Triggers the actual pageflip by updating the primary
- * surface base address.
- */
-static void dce_v11_0_page_flip(struct amdgpu_device *adev,
-				int crtc_id, u64 crtc_base, bool async)
-{
-	struct amdgpu_crtc *amdgpu_crtc = adev->mode_info.crtcs[crtc_id];
-	struct drm_framebuffer *fb = amdgpu_crtc->base.primary->fb;
-	u32 tmp;
-
-	/* flip immediate for async, default is vsync */
-	tmp = RREG32(mmGRPH_FLIP_CONTROL + amdgpu_crtc->crtc_offset);
-	tmp = REG_SET_FIELD(tmp, GRPH_FLIP_CONTROL,
-			    GRPH_SURFACE_UPDATE_IMMEDIATE_EN, async ? 1 : 0);
-	WREG32(mmGRPH_FLIP_CONTROL + amdgpu_crtc->crtc_offset, tmp);
-	/* update pitch */
-	WREG32(mmGRPH_PITCH + amdgpu_crtc->crtc_offset,
-	       fb->pitches[0] / fb->format->cpp[0]);
-	/* update the scanout addresses */
-	WREG32(mmGRPH_PRIMARY_SURFACE_ADDRESS_HIGH + amdgpu_crtc->crtc_offset,
-	       upper_32_bits(crtc_base));
-	/* writing to the low address triggers the update */
-	WREG32(mmGRPH_PRIMARY_SURFACE_ADDRESS + amdgpu_crtc->crtc_offset,
-	       lower_32_bits(crtc_base));
-	/* post the write */
-	RREG32(mmGRPH_PRIMARY_SURFACE_ADDRESS + amdgpu_crtc->crtc_offset);
-}
-
-static int dce_v11_0_crtc_get_scanoutpos(struct amdgpu_device *adev, int crtc,
-					u32 *vbl, u32 *position)
-{
-	if ((crtc < 0) || (crtc >= adev->mode_info.num_crtc))
-		return -EINVAL;
-
-	*vbl = RREG32(mmCRTC_V_BLANK_START_END + crtc_offsets[crtc]);
-	*position = RREG32(mmCRTC_STATUS_POSITION + crtc_offsets[crtc]);
-
-	return 0;
-}
-
-/**
- * dce_v11_0_hpd_sense - hpd sense callback.
- *
- * @adev: amdgpu_device pointer
- * @hpd: hpd (hotplug detect) pin
- *
- * Checks if a digital monitor is connected (evergreen+).
- * Returns true if connected, false if not connected.
- */
-static bool dce_v11_0_hpd_sense(struct amdgpu_device *adev,
-			       enum amdgpu_hpd_id hpd)
-{
-	bool connected = false;
-
-	if (hpd >= adev->mode_info.num_hpd)
-		return connected;
-
-	if (RREG32(mmDC_HPD_INT_STATUS + hpd_offsets[hpd]) &
-	    DC_HPD_INT_STATUS__DC_HPD_SENSE_MASK)
-		connected = true;
-
-	return connected;
-}
-
-/**
- * dce_v11_0_hpd_set_polarity - hpd set polarity callback.
- *
- * @adev: amdgpu_device pointer
- * @hpd: hpd (hotplug detect) pin
- *
- * Set the polarity of the hpd pin (evergreen+).
- */
-static void dce_v11_0_hpd_set_polarity(struct amdgpu_device *adev,
-				      enum amdgpu_hpd_id hpd)
-{
-	u32 tmp;
-	bool connected = dce_v11_0_hpd_sense(adev, hpd);
-
-	if (hpd >= adev->mode_info.num_hpd)
-		return;
-
-	tmp = RREG32(mmDC_HPD_INT_CONTROL + hpd_offsets[hpd]);
-	if (connected)
-		tmp = REG_SET_FIELD(tmp, DC_HPD_INT_CONTROL, DC_HPD_INT_POLARITY, 0);
-	else
-		tmp = REG_SET_FIELD(tmp, DC_HPD_INT_CONTROL, DC_HPD_INT_POLARITY, 1);
-	WREG32(mmDC_HPD_INT_CONTROL + hpd_offsets[hpd], tmp);
-}
-
-/**
- * dce_v11_0_hpd_init - hpd setup callback.
- *
- * @adev: amdgpu_device pointer
- *
- * Setup the hpd pins used by the card (evergreen+).
- * Enable the pin, set the polarity, and enable the hpd interrupts.
- */
-static void dce_v11_0_hpd_init(struct amdgpu_device *adev)
-{
-	struct drm_device *dev = adev_to_drm(adev);
-	struct drm_connector *connector;
-	struct drm_connector_list_iter iter;
-	u32 tmp;
-
-	drm_connector_list_iter_begin(dev, &iter);
-	drm_for_each_connector_iter(connector, &iter) {
-		struct amdgpu_connector *amdgpu_connector = to_amdgpu_connector(connector);
-
-		if (amdgpu_connector->hpd.hpd >= adev->mode_info.num_hpd)
-			continue;
-
-		if (connector->connector_type == DRM_MODE_CONNECTOR_eDP ||
-		    connector->connector_type == DRM_MODE_CONNECTOR_LVDS) {
-			/* don't try to enable hpd on eDP or LVDS avoid breaking the
-			 * aux dp channel on imac and help (but not completely fix)
-			 * https://bugzilla.redhat.com/show_bug.cgi?id=726143
-			 * also avoid interrupt storms during dpms.
-			 */
-			tmp = RREG32(mmDC_HPD_INT_CONTROL + hpd_offsets[amdgpu_connector->hpd.hpd]);
-			tmp = REG_SET_FIELD(tmp, DC_HPD_INT_CONTROL, DC_HPD_INT_EN, 0);
-			WREG32(mmDC_HPD_INT_CONTROL + hpd_offsets[amdgpu_connector->hpd.hpd], tmp);
-			continue;
-		}
-
-		tmp = RREG32(mmDC_HPD_CONTROL + hpd_offsets[amdgpu_connector->hpd.hpd]);
-		tmp = REG_SET_FIELD(tmp, DC_HPD_CONTROL, DC_HPD_EN, 1);
-		WREG32(mmDC_HPD_CONTROL + hpd_offsets[amdgpu_connector->hpd.hpd], tmp);
-
-		tmp = RREG32(mmDC_HPD_TOGGLE_FILT_CNTL + hpd_offsets[amdgpu_connector->hpd.hpd]);
-		tmp = REG_SET_FIELD(tmp, DC_HPD_TOGGLE_FILT_CNTL,
-				    DC_HPD_CONNECT_INT_DELAY,
-				    AMDGPU_HPD_CONNECT_INT_DELAY_IN_MS);
-		tmp = REG_SET_FIELD(tmp, DC_HPD_TOGGLE_FILT_CNTL,
-				    DC_HPD_DISCONNECT_INT_DELAY,
-				    AMDGPU_HPD_DISCONNECT_INT_DELAY_IN_MS);
-		WREG32(mmDC_HPD_TOGGLE_FILT_CNTL + hpd_offsets[amdgpu_connector->hpd.hpd], tmp);
-
-		dce_v11_0_hpd_set_polarity(adev, amdgpu_connector->hpd.hpd);
-		amdgpu_irq_get(adev, &adev->hpd_irq, amdgpu_connector->hpd.hpd);
-	}
-	drm_connector_list_iter_end(&iter);
-}
-
-/**
- * dce_v11_0_hpd_fini - hpd tear down callback.
- *
- * @adev: amdgpu_device pointer
- *
- * Tear down the hpd pins used by the card (evergreen+).
- * Disable the hpd interrupts.
- */
-static void dce_v11_0_hpd_fini(struct amdgpu_device *adev)
-{
-	struct drm_device *dev = adev_to_drm(adev);
-	struct drm_connector *connector;
-	struct drm_connector_list_iter iter;
-	u32 tmp;
-
-	drm_connector_list_iter_begin(dev, &iter);
-	drm_for_each_connector_iter(connector, &iter) {
-		struct amdgpu_connector *amdgpu_connector = to_amdgpu_connector(connector);
-
-		if (amdgpu_connector->hpd.hpd >= adev->mode_info.num_hpd)
-			continue;
-
-		tmp = RREG32(mmDC_HPD_CONTROL + hpd_offsets[amdgpu_connector->hpd.hpd]);
-		tmp = REG_SET_FIELD(tmp, DC_HPD_CONTROL, DC_HPD_EN, 0);
-		WREG32(mmDC_HPD_CONTROL + hpd_offsets[amdgpu_connector->hpd.hpd], tmp);
-
-		amdgpu_irq_put(adev, &adev->hpd_irq, amdgpu_connector->hpd.hpd);
-	}
-	drm_connector_list_iter_end(&iter);
-}
-
-static u32 dce_v11_0_hpd_get_gpio_reg(struct amdgpu_device *adev)
-{
-	return mmDC_GPIO_HPD_A;
-}
-
-static bool dce_v11_0_is_display_hung(struct amdgpu_device *adev)
-{
-	u32 crtc_hung = 0;
-	u32 crtc_status[6];
-	u32 i, j, tmp;
-
-	for (i = 0; i < adev->mode_info.num_crtc; i++) {
-		tmp = RREG32(mmCRTC_CONTROL + crtc_offsets[i]);
-		if (REG_GET_FIELD(tmp, CRTC_CONTROL, CRTC_MASTER_EN)) {
-			crtc_status[i] = RREG32(mmCRTC_STATUS_HV_COUNT + crtc_offsets[i]);
-			crtc_hung |= (1 << i);
-		}
-	}
-
-	for (j = 0; j < 10; j++) {
-		for (i = 0; i < adev->mode_info.num_crtc; i++) {
-			if (crtc_hung & (1 << i)) {
-				tmp = RREG32(mmCRTC_STATUS_HV_COUNT + crtc_offsets[i]);
-				if (tmp != crtc_status[i])
-					crtc_hung &= ~(1 << i);
-			}
-		}
-		if (crtc_hung == 0)
-			return false;
-		udelay(100);
-	}
-
-	return true;
-}
-
-static void dce_v11_0_set_vga_render_state(struct amdgpu_device *adev,
-					   bool render)
-{
-	u32 tmp;
-
-	/* Lockout access through VGA aperture*/
-	tmp = RREG32(mmVGA_HDP_CONTROL);
-	if (render)
-		tmp = REG_SET_FIELD(tmp, VGA_HDP_CONTROL, VGA_MEMORY_DISABLE, 0);
-	else
-		tmp = REG_SET_FIELD(tmp, VGA_HDP_CONTROL, VGA_MEMORY_DISABLE, 1);
-	WREG32(mmVGA_HDP_CONTROL, tmp);
-
-	/* disable VGA render */
-	tmp = RREG32(mmVGA_RENDER_CONTROL);
-	if (render)
-		tmp = REG_SET_FIELD(tmp, VGA_RENDER_CONTROL, VGA_VSTATUS_CNTL, 1);
-	else
-		tmp = REG_SET_FIELD(tmp, VGA_RENDER_CONTROL, VGA_VSTATUS_CNTL, 0);
-	WREG32(mmVGA_RENDER_CONTROL, tmp);
-}
-
-static int dce_v11_0_get_num_crtc (struct amdgpu_device *adev)
-{
-	int num_crtc = 0;
-
-	switch (adev->asic_type) {
-	case CHIP_CARRIZO:
-		num_crtc = 3;
-		break;
-	case CHIP_STONEY:
-		num_crtc = 2;
-		break;
-	case CHIP_POLARIS10:
-	case CHIP_VEGAM:
-		num_crtc = 6;
-		break;
-	case CHIP_POLARIS11:
-	case CHIP_POLARIS12:
-		num_crtc = 5;
-		break;
-	default:
-		num_crtc = 0;
-	}
-	return num_crtc;
-}
-
-void dce_v11_0_disable_dce(struct amdgpu_device *adev)
-{
-	/*Disable VGA render and enabled crtc, if has DCE engine*/
-	if (amdgpu_atombios_has_dce_engine_info(adev)) {
-		u32 tmp;
-		int crtc_enabled, i;
-
-		dce_v11_0_set_vga_render_state(adev, false);
-
-		/*Disable crtc*/
-		for (i = 0; i < dce_v11_0_get_num_crtc(adev); i++) {
-			crtc_enabled = REG_GET_FIELD(RREG32(mmCRTC_CONTROL + crtc_offsets[i]),
-									 CRTC_CONTROL, CRTC_MASTER_EN);
-			if (crtc_enabled) {
-				WREG32(mmCRTC_UPDATE_LOCK + crtc_offsets[i], 1);
-				tmp = RREG32(mmCRTC_CONTROL + crtc_offsets[i]);
-				tmp = REG_SET_FIELD(tmp, CRTC_CONTROL, CRTC_MASTER_EN, 0);
-				WREG32(mmCRTC_CONTROL + crtc_offsets[i], tmp);
-				WREG32(mmCRTC_UPDATE_LOCK + crtc_offsets[i], 0);
-			}
-		}
-	}
-}
-
-static void dce_v11_0_program_fmt(struct drm_encoder *encoder)
-{
-	struct drm_device *dev = encoder->dev;
-	struct amdgpu_device *adev = drm_to_adev(dev);
-	struct amdgpu_encoder *amdgpu_encoder = to_amdgpu_encoder(encoder);
-	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(encoder->crtc);
-	struct drm_connector *connector = amdgpu_get_connector_for_encoder(encoder);
-	int bpc = 0;
-	u32 tmp = 0;
-	enum amdgpu_connector_dither dither = AMDGPU_FMT_DITHER_DISABLE;
-
-	if (connector) {
-		struct amdgpu_connector *amdgpu_connector = to_amdgpu_connector(connector);
-		bpc = amdgpu_connector_get_monitor_bpc(connector);
-		dither = amdgpu_connector->dither;
-	}
-
-	/* LVDS/eDP FMT is set up by atom */
-	if (amdgpu_encoder->devices & ATOM_DEVICE_LCD_SUPPORT)
-		return;
-
-	/* not needed for analog */
-	if ((amdgpu_encoder->encoder_id == ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC1) ||
-	    (amdgpu_encoder->encoder_id == ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC2))
-		return;
-
-	if (bpc == 0)
-		return;
-
-	switch (bpc) {
-	case 6:
-		if (dither == AMDGPU_FMT_DITHER_ENABLE) {
-			/* XXX sort out optimal dither settings */
-			tmp = REG_SET_FIELD(tmp, FMT_BIT_DEPTH_CONTROL, FMT_FRAME_RANDOM_ENABLE, 1);
-			tmp = REG_SET_FIELD(tmp, FMT_BIT_DEPTH_CONTROL, FMT_HIGHPASS_RANDOM_ENABLE, 1);
-			tmp = REG_SET_FIELD(tmp, FMT_BIT_DEPTH_CONTROL, FMT_SPATIAL_DITHER_EN, 1);
-			tmp = REG_SET_FIELD(tmp, FMT_BIT_DEPTH_CONTROL, FMT_SPATIAL_DITHER_DEPTH, 0);
-		} else {
-			tmp = REG_SET_FIELD(tmp, FMT_BIT_DEPTH_CONTROL, FMT_TRUNCATE_EN, 1);
-			tmp = REG_SET_FIELD(tmp, FMT_BIT_DEPTH_CONTROL, FMT_TRUNCATE_DEPTH, 0);
-		}
-		break;
-	case 8:
-		if (dither == AMDGPU_FMT_DITHER_ENABLE) {
-			/* XXX sort out optimal dither settings */
-			tmp = REG_SET_FIELD(tmp, FMT_BIT_DEPTH_CONTROL, FMT_FRAME_RANDOM_ENABLE, 1);
-			tmp = REG_SET_FIELD(tmp, FMT_BIT_DEPTH_CONTROL, FMT_HIGHPASS_RANDOM_ENABLE, 1);
-			tmp = REG_SET_FIELD(tmp, FMT_BIT_DEPTH_CONTROL, FMT_RGB_RANDOM_ENABLE, 1);
-			tmp = REG_SET_FIELD(tmp, FMT_BIT_DEPTH_CONTROL, FMT_SPATIAL_DITHER_EN, 1);
-			tmp = REG_SET_FIELD(tmp, FMT_BIT_DEPTH_CONTROL, FMT_SPATIAL_DITHER_DEPTH, 1);
-		} else {
-			tmp = REG_SET_FIELD(tmp, FMT_BIT_DEPTH_CONTROL, FMT_TRUNCATE_EN, 1);
-			tmp = REG_SET_FIELD(tmp, FMT_BIT_DEPTH_CONTROL, FMT_TRUNCATE_DEPTH, 1);
-		}
-		break;
-	case 10:
-		if (dither == AMDGPU_FMT_DITHER_ENABLE) {
-			/* XXX sort out optimal dither settings */
-			tmp = REG_SET_FIELD(tmp, FMT_BIT_DEPTH_CONTROL, FMT_FRAME_RANDOM_ENABLE, 1);
-			tmp = REG_SET_FIELD(tmp, FMT_BIT_DEPTH_CONTROL, FMT_HIGHPASS_RANDOM_ENABLE, 1);
-			tmp = REG_SET_FIELD(tmp, FMT_BIT_DEPTH_CONTROL, FMT_RGB_RANDOM_ENABLE, 1);
-			tmp = REG_SET_FIELD(tmp, FMT_BIT_DEPTH_CONTROL, FMT_SPATIAL_DITHER_EN, 1);
-			tmp = REG_SET_FIELD(tmp, FMT_BIT_DEPTH_CONTROL, FMT_SPATIAL_DITHER_DEPTH, 2);
-		} else {
-			tmp = REG_SET_FIELD(tmp, FMT_BIT_DEPTH_CONTROL, FMT_TRUNCATE_EN, 1);
-			tmp = REG_SET_FIELD(tmp, FMT_BIT_DEPTH_CONTROL, FMT_TRUNCATE_DEPTH, 2);
-		}
-		break;
-	default:
-		/* not needed */
-		break;
-	}
-
-	WREG32(mmFMT_BIT_DEPTH_CONTROL + amdgpu_crtc->crtc_offset, tmp);
-}
-
-
-/* display watermark setup */
-/**
- * dce_v11_0_line_buffer_adjust - Set up the line buffer
- *
- * @adev: amdgpu_device pointer
- * @amdgpu_crtc: the selected display controller
- * @mode: the current display mode on the selected display
- * controller
- *
- * Setup up the line buffer allocation for
- * the selected display controller (CIK).
- * Returns the line buffer size in pixels.
- */
-static u32 dce_v11_0_line_buffer_adjust(struct amdgpu_device *adev,
-				       struct amdgpu_crtc *amdgpu_crtc,
-				       struct drm_display_mode *mode)
-{
-	u32 tmp, buffer_alloc, i, mem_cfg;
-	u32 pipe_offset = amdgpu_crtc->crtc_id;
-	/*
-	 * Line Buffer Setup
-	 * There are 6 line buffers, one for each display controllers.
-	 * There are 3 partitions per LB. Select the number of partitions
-	 * to enable based on the display width.  For display widths larger
-	 * than 4096, you need use to use 2 display controllers and combine
-	 * them using the stereo blender.
-	 */
-	if (amdgpu_crtc->base.enabled && mode) {
-		if (mode->crtc_hdisplay < 1920) {
-			mem_cfg = 1;
-			buffer_alloc = 2;
-		} else if (mode->crtc_hdisplay < 2560) {
-			mem_cfg = 2;
-			buffer_alloc = 2;
-		} else if (mode->crtc_hdisplay < 4096) {
-			mem_cfg = 0;
-			buffer_alloc = (adev->flags & AMD_IS_APU) ? 2 : 4;
-		} else {
-			DRM_DEBUG_KMS("Mode too big for LB!\n");
-			mem_cfg = 0;
-			buffer_alloc = (adev->flags & AMD_IS_APU) ? 2 : 4;
-		}
-	} else {
-		mem_cfg = 1;
-		buffer_alloc = 0;
-	}
-
-	tmp = RREG32(mmLB_MEMORY_CTRL + amdgpu_crtc->crtc_offset);
-	tmp = REG_SET_FIELD(tmp, LB_MEMORY_CTRL, LB_MEMORY_CONFIG, mem_cfg);
-	WREG32(mmLB_MEMORY_CTRL + amdgpu_crtc->crtc_offset, tmp);
-
-	tmp = RREG32(mmPIPE0_DMIF_BUFFER_CONTROL + pipe_offset);
-	tmp = REG_SET_FIELD(tmp, PIPE0_DMIF_BUFFER_CONTROL, DMIF_BUFFERS_ALLOCATED, buffer_alloc);
-	WREG32(mmPIPE0_DMIF_BUFFER_CONTROL + pipe_offset, tmp);
-
-	for (i = 0; i < adev->usec_timeout; i++) {
-		tmp = RREG32(mmPIPE0_DMIF_BUFFER_CONTROL + pipe_offset);
-		if (REG_GET_FIELD(tmp, PIPE0_DMIF_BUFFER_CONTROL, DMIF_BUFFERS_ALLOCATION_COMPLETED))
-			break;
-		udelay(1);
-	}
-
-	if (amdgpu_crtc->base.enabled && mode) {
-		switch (mem_cfg) {
-		case 0:
-		default:
-			return 4096 * 2;
-		case 1:
-			return 1920 * 2;
-		case 2:
-			return 2560 * 2;
-		}
-	}
-
-	/* controller not enabled, so no lb used */
-	return 0;
-}
-
-/**
- * cik_get_number_of_dram_channels - get the number of dram channels
- *
- * @adev: amdgpu_device pointer
- *
- * Look up the number of video ram channels (CIK).
- * Used for display watermark bandwidth calculations
- * Returns the number of dram channels
- */
-static u32 cik_get_number_of_dram_channels(struct amdgpu_device *adev)
-{
-	u32 tmp = RREG32(mmMC_SHARED_CHMAP);
-
-	switch (REG_GET_FIELD(tmp, MC_SHARED_CHMAP, NOOFCHAN)) {
-	case 0:
-	default:
-		return 1;
-	case 1:
-		return 2;
-	case 2:
-		return 4;
-	case 3:
-		return 8;
-	case 4:
-		return 3;
-	case 5:
-		return 6;
-	case 6:
-		return 10;
-	case 7:
-		return 12;
-	case 8:
-		return 16;
-	}
-}
-
-struct dce10_wm_params {
-	u32 dram_channels; /* number of dram channels */
-	u32 yclk;          /* bandwidth per dram data pin in kHz */
-	u32 sclk;          /* engine clock in kHz */
-	u32 disp_clk;      /* display clock in kHz */
-	u32 src_width;     /* viewport width */
-	u32 active_time;   /* active display time in ns */
-	u32 blank_time;    /* blank time in ns */
-	bool interlaced;    /* mode is interlaced */
-	fixed20_12 vsc;    /* vertical scale ratio */
-	u32 num_heads;     /* number of active crtcs */
-	u32 bytes_per_pixel; /* bytes per pixel display + overlay */
-	u32 lb_size;       /* line buffer allocated to pipe */
-	u32 vtaps;         /* vertical scaler taps */
-};
-
-/**
- * dce_v11_0_dram_bandwidth - get the dram bandwidth
- *
- * @wm: watermark calculation data
- *
- * Calculate the raw dram bandwidth (CIK).
- * Used for display watermark bandwidth calculations
- * Returns the dram bandwidth in MBytes/s
- */
-static u32 dce_v11_0_dram_bandwidth(struct dce10_wm_params *wm)
-{
-	/* Calculate raw DRAM Bandwidth */
-	fixed20_12 dram_efficiency; /* 0.7 */
-	fixed20_12 yclk, dram_channels, bandwidth;
-	fixed20_12 a;
-
-	a.full = dfixed_const(1000);
-	yclk.full = dfixed_const(wm->yclk);
-	yclk.full = dfixed_div(yclk, a);
-	dram_channels.full = dfixed_const(wm->dram_channels * 4);
-	a.full = dfixed_const(10);
-	dram_efficiency.full = dfixed_const(7);
-	dram_efficiency.full = dfixed_div(dram_efficiency, a);
-	bandwidth.full = dfixed_mul(dram_channels, yclk);
-	bandwidth.full = dfixed_mul(bandwidth, dram_efficiency);
-
-	return dfixed_trunc(bandwidth);
-}
-
-/**
- * dce_v11_0_dram_bandwidth_for_display - get the dram bandwidth for display
- *
- * @wm: watermark calculation data
- *
- * Calculate the dram bandwidth used for display (CIK).
- * Used for display watermark bandwidth calculations
- * Returns the dram bandwidth for display in MBytes/s
- */
-static u32 dce_v11_0_dram_bandwidth_for_display(struct dce10_wm_params *wm)
-{
-	/* Calculate DRAM Bandwidth and the part allocated to display. */
-	fixed20_12 disp_dram_allocation; /* 0.3 to 0.7 */
-	fixed20_12 yclk, dram_channels, bandwidth;
-	fixed20_12 a;
-
-	a.full = dfixed_const(1000);
-	yclk.full = dfixed_const(wm->yclk);
-	yclk.full = dfixed_div(yclk, a);
-	dram_channels.full = dfixed_const(wm->dram_channels * 4);
-	a.full = dfixed_const(10);
-	disp_dram_allocation.full = dfixed_const(3); /* XXX worse case value 0.3 */
-	disp_dram_allocation.full = dfixed_div(disp_dram_allocation, a);
-	bandwidth.full = dfixed_mul(dram_channels, yclk);
-	bandwidth.full = dfixed_mul(bandwidth, disp_dram_allocation);
-
-	return dfixed_trunc(bandwidth);
-}
-
-/**
- * dce_v11_0_data_return_bandwidth - get the data return bandwidth
- *
- * @wm: watermark calculation data
- *
- * Calculate the data return bandwidth used for display (CIK).
- * Used for display watermark bandwidth calculations
- * Returns the data return bandwidth in MBytes/s
- */
-static u32 dce_v11_0_data_return_bandwidth(struct dce10_wm_params *wm)
-{
-	/* Calculate the display Data return Bandwidth */
-	fixed20_12 return_efficiency; /* 0.8 */
-	fixed20_12 sclk, bandwidth;
-	fixed20_12 a;
-
-	a.full = dfixed_const(1000);
-	sclk.full = dfixed_const(wm->sclk);
-	sclk.full = dfixed_div(sclk, a);
-	a.full = dfixed_const(10);
-	return_efficiency.full = dfixed_const(8);
-	return_efficiency.full = dfixed_div(return_efficiency, a);
-	a.full = dfixed_const(32);
-	bandwidth.full = dfixed_mul(a, sclk);
-	bandwidth.full = dfixed_mul(bandwidth, return_efficiency);
-
-	return dfixed_trunc(bandwidth);
-}
-
-/**
- * dce_v11_0_dmif_request_bandwidth - get the dmif bandwidth
- *
- * @wm: watermark calculation data
- *
- * Calculate the dmif bandwidth used for display (CIK).
- * Used for display watermark bandwidth calculations
- * Returns the dmif bandwidth in MBytes/s
- */
-static u32 dce_v11_0_dmif_request_bandwidth(struct dce10_wm_params *wm)
-{
-	/* Calculate the DMIF Request Bandwidth */
-	fixed20_12 disp_clk_request_efficiency; /* 0.8 */
-	fixed20_12 disp_clk, bandwidth;
-	fixed20_12 a, b;
-
-	a.full = dfixed_const(1000);
-	disp_clk.full = dfixed_const(wm->disp_clk);
-	disp_clk.full = dfixed_div(disp_clk, a);
-	a.full = dfixed_const(32);
-	b.full = dfixed_mul(a, disp_clk);
-
-	a.full = dfixed_const(10);
-	disp_clk_request_efficiency.full = dfixed_const(8);
-	disp_clk_request_efficiency.full = dfixed_div(disp_clk_request_efficiency, a);
-
-	bandwidth.full = dfixed_mul(b, disp_clk_request_efficiency);
-
-	return dfixed_trunc(bandwidth);
-}
-
-/**
- * dce_v11_0_available_bandwidth - get the min available bandwidth
- *
- * @wm: watermark calculation data
- *
- * Calculate the min available bandwidth used for display (CIK).
- * Used for display watermark bandwidth calculations
- * Returns the min available bandwidth in MBytes/s
- */
-static u32 dce_v11_0_available_bandwidth(struct dce10_wm_params *wm)
-{
-	/* Calculate the Available bandwidth. Display can use this temporarily but not in average. */
-	u32 dram_bandwidth = dce_v11_0_dram_bandwidth(wm);
-	u32 data_return_bandwidth = dce_v11_0_data_return_bandwidth(wm);
-	u32 dmif_req_bandwidth = dce_v11_0_dmif_request_bandwidth(wm);
-
-	return min(dram_bandwidth, min(data_return_bandwidth, dmif_req_bandwidth));
-}
-
-/**
- * dce_v11_0_average_bandwidth - get the average available bandwidth
- *
- * @wm: watermark calculation data
- *
- * Calculate the average available bandwidth used for display (CIK).
- * Used for display watermark bandwidth calculations
- * Returns the average available bandwidth in MBytes/s
- */
-static u32 dce_v11_0_average_bandwidth(struct dce10_wm_params *wm)
-{
-	/* Calculate the display mode Average Bandwidth
-	 * DisplayMode should contain the source and destination dimensions,
-	 * timing, etc.
-	 */
-	fixed20_12 bpp;
-	fixed20_12 line_time;
-	fixed20_12 src_width;
-	fixed20_12 bandwidth;
-	fixed20_12 a;
-
-	a.full = dfixed_const(1000);
-	line_time.full = dfixed_const(wm->active_time + wm->blank_time);
-	line_time.full = dfixed_div(line_time, a);
-	bpp.full = dfixed_const(wm->bytes_per_pixel);
-	src_width.full = dfixed_const(wm->src_width);
-	bandwidth.full = dfixed_mul(src_width, bpp);
-	bandwidth.full = dfixed_mul(bandwidth, wm->vsc);
-	bandwidth.full = dfixed_div(bandwidth, line_time);
-
-	return dfixed_trunc(bandwidth);
-}
-
-/**
- * dce_v11_0_latency_watermark - get the latency watermark
- *
- * @wm: watermark calculation data
- *
- * Calculate the latency watermark (CIK).
- * Used for display watermark bandwidth calculations
- * Returns the latency watermark in ns
- */
-static u32 dce_v11_0_latency_watermark(struct dce10_wm_params *wm)
-{
-	/* First calculate the latency in ns */
-	u32 mc_latency = 2000; /* 2000 ns. */
-	u32 available_bandwidth = dce_v11_0_available_bandwidth(wm);
-	u32 worst_chunk_return_time = (512 * 8 * 1000) / available_bandwidth;
-	u32 cursor_line_pair_return_time = (128 * 4 * 1000) / available_bandwidth;
-	u32 dc_latency = 40000000 / wm->disp_clk; /* dc pipe latency */
-	u32 other_heads_data_return_time = ((wm->num_heads + 1) * worst_chunk_return_time) +
-		(wm->num_heads * cursor_line_pair_return_time);
-	u32 latency = mc_latency + other_heads_data_return_time + dc_latency;
-	u32 max_src_lines_per_dst_line, lb_fill_bw, line_fill_time;
-	u32 tmp, dmif_size = 12288;
-	fixed20_12 a, b, c;
-
-	if (wm->num_heads == 0)
-		return 0;
-
-	a.full = dfixed_const(2);
-	b.full = dfixed_const(1);
-	if ((wm->vsc.full > a.full) ||
-	    ((wm->vsc.full > b.full) && (wm->vtaps >= 3)) ||
-	    (wm->vtaps >= 5) ||
-	    ((wm->vsc.full >= a.full) && wm->interlaced))
-		max_src_lines_per_dst_line = 4;
-	else
-		max_src_lines_per_dst_line = 2;
-
-	a.full = dfixed_const(available_bandwidth);
-	b.full = dfixed_const(wm->num_heads);
-	a.full = dfixed_div(a, b);
-	tmp = div_u64((u64) dmif_size * (u64) wm->disp_clk, mc_latency + 512);
-	tmp = min(dfixed_trunc(a), tmp);
-
-	lb_fill_bw = min(tmp, wm->disp_clk * wm->bytes_per_pixel / 1000);
-
-	a.full = dfixed_const(max_src_lines_per_dst_line * wm->src_width * wm->bytes_per_pixel);
-	b.full = dfixed_const(1000);
-	c.full = dfixed_const(lb_fill_bw);
-	b.full = dfixed_div(c, b);
-	a.full = dfixed_div(a, b);
-	line_fill_time = dfixed_trunc(a);
-
-	if (line_fill_time < wm->active_time)
-		return latency;
-	else
-		return latency + (line_fill_time - wm->active_time);
-
-}
-
-/**
- * dce_v11_0_average_bandwidth_vs_dram_bandwidth_for_display - check
- * average and available dram bandwidth
- *
- * @wm: watermark calculation data
- *
- * Check if the display average bandwidth fits in the display
- * dram bandwidth (CIK).
- * Used for display watermark bandwidth calculations
- * Returns true if the display fits, false if not.
- */
-static bool dce_v11_0_average_bandwidth_vs_dram_bandwidth_for_display(struct dce10_wm_params *wm)
-{
-	if (dce_v11_0_average_bandwidth(wm) <=
-	    (dce_v11_0_dram_bandwidth_for_display(wm) / wm->num_heads))
-		return true;
-	else
-		return false;
-}
-
-/**
- * dce_v11_0_average_bandwidth_vs_available_bandwidth - check
- * average and available bandwidth
- *
- * @wm: watermark calculation data
- *
- * Check if the display average bandwidth fits in the display
- * available bandwidth (CIK).
- * Used for display watermark bandwidth calculations
- * Returns true if the display fits, false if not.
- */
-static bool dce_v11_0_average_bandwidth_vs_available_bandwidth(struct dce10_wm_params *wm)
-{
-	if (dce_v11_0_average_bandwidth(wm) <=
-	    (dce_v11_0_available_bandwidth(wm) / wm->num_heads))
-		return true;
-	else
-		return false;
-}
-
-/**
- * dce_v11_0_check_latency_hiding - check latency hiding
- *
- * @wm: watermark calculation data
- *
- * Check latency hiding (CIK).
- * Used for display watermark bandwidth calculations
- * Returns true if the display fits, false if not.
- */
-static bool dce_v11_0_check_latency_hiding(struct dce10_wm_params *wm)
-{
-	u32 lb_partitions = wm->lb_size / wm->src_width;
-	u32 line_time = wm->active_time + wm->blank_time;
-	u32 latency_tolerant_lines;
-	u32 latency_hiding;
-	fixed20_12 a;
-
-	a.full = dfixed_const(1);
-	if (wm->vsc.full > a.full)
-		latency_tolerant_lines = 1;
-	else {
-		if (lb_partitions <= (wm->vtaps + 1))
-			latency_tolerant_lines = 1;
-		else
-			latency_tolerant_lines = 2;
-	}
-
-	latency_hiding = (latency_tolerant_lines * line_time + wm->blank_time);
-
-	if (dce_v11_0_latency_watermark(wm) <= latency_hiding)
-		return true;
-	else
-		return false;
-}
-
-/**
- * dce_v11_0_program_watermarks - program display watermarks
- *
- * @adev: amdgpu_device pointer
- * @amdgpu_crtc: the selected display controller
- * @lb_size: line buffer size
- * @num_heads: number of display controllers in use
- *
- * Calculate and program the display watermarks for the
- * selected display controller (CIK).
- */
-static void dce_v11_0_program_watermarks(struct amdgpu_device *adev,
-					struct amdgpu_crtc *amdgpu_crtc,
-					u32 lb_size, u32 num_heads)
-{
-	struct drm_display_mode *mode = &amdgpu_crtc->base.mode;
-	struct dce10_wm_params wm_low, wm_high;
-	u32 active_time;
-	u32 line_time = 0;
-	u32 latency_watermark_a = 0, latency_watermark_b = 0;
-	u32 tmp, wm_mask, lb_vblank_lead_lines = 0;
-
-	if (amdgpu_crtc->base.enabled && num_heads && mode) {
-		active_time = (u32) div_u64((u64)mode->crtc_hdisplay * 1000000,
-					    (u32)mode->clock);
-		line_time = (u32) div_u64((u64)mode->crtc_htotal * 1000000,
-					  (u32)mode->clock);
-		line_time = min(line_time, (u32)65535);
-
-		/* watermark for high clocks */
-		if (adev->pm.dpm_enabled) {
-			wm_high.yclk =
-				amdgpu_dpm_get_mclk(adev, false) * 10;
-			wm_high.sclk =
-				amdgpu_dpm_get_sclk(adev, false) * 10;
-		} else {
-			wm_high.yclk = adev->pm.current_mclk * 10;
-			wm_high.sclk = adev->pm.current_sclk * 10;
-		}
-
-		wm_high.disp_clk = mode->clock;
-		wm_high.src_width = mode->crtc_hdisplay;
-		wm_high.active_time = active_time;
-		wm_high.blank_time = line_time - wm_high.active_time;
-		wm_high.interlaced = false;
-		if (mode->flags & DRM_MODE_FLAG_INTERLACE)
-			wm_high.interlaced = true;
-		wm_high.vsc = amdgpu_crtc->vsc;
-		wm_high.vtaps = 1;
-		if (amdgpu_crtc->rmx_type != RMX_OFF)
-			wm_high.vtaps = 2;
-		wm_high.bytes_per_pixel = 4; /* XXX: get this from fb config */
-		wm_high.lb_size = lb_size;
-		wm_high.dram_channels = cik_get_number_of_dram_channels(adev);
-		wm_high.num_heads = num_heads;
-
-		/* set for high clocks */
-		latency_watermark_a = min(dce_v11_0_latency_watermark(&wm_high), (u32)65535);
-
-		/* possibly force display priority to high */
-		/* should really do this at mode validation time... */
-		if (!dce_v11_0_average_bandwidth_vs_dram_bandwidth_for_display(&wm_high) ||
-		    !dce_v11_0_average_bandwidth_vs_available_bandwidth(&wm_high) ||
-		    !dce_v11_0_check_latency_hiding(&wm_high) ||
-		    (adev->mode_info.disp_priority == 2)) {
-			DRM_DEBUG_KMS("force priority to high\n");
-		}
-
-		/* watermark for low clocks */
-		if (adev->pm.dpm_enabled) {
-			wm_low.yclk =
-				amdgpu_dpm_get_mclk(adev, true) * 10;
-			wm_low.sclk =
-				amdgpu_dpm_get_sclk(adev, true) * 10;
-		} else {
-			wm_low.yclk = adev->pm.current_mclk * 10;
-			wm_low.sclk = adev->pm.current_sclk * 10;
-		}
-
-		wm_low.disp_clk = mode->clock;
-		wm_low.src_width = mode->crtc_hdisplay;
-		wm_low.active_time = active_time;
-		wm_low.blank_time = line_time - wm_low.active_time;
-		wm_low.interlaced = false;
-		if (mode->flags & DRM_MODE_FLAG_INTERLACE)
-			wm_low.interlaced = true;
-		wm_low.vsc = amdgpu_crtc->vsc;
-		wm_low.vtaps = 1;
-		if (amdgpu_crtc->rmx_type != RMX_OFF)
-			wm_low.vtaps = 2;
-		wm_low.bytes_per_pixel = 4; /* XXX: get this from fb config */
-		wm_low.lb_size = lb_size;
-		wm_low.dram_channels = cik_get_number_of_dram_channels(adev);
-		wm_low.num_heads = num_heads;
-
-		/* set for low clocks */
-		latency_watermark_b = min(dce_v11_0_latency_watermark(&wm_low), (u32)65535);
-
-		/* possibly force display priority to high */
-		/* should really do this at mode validation time... */
-		if (!dce_v11_0_average_bandwidth_vs_dram_bandwidth_for_display(&wm_low) ||
-		    !dce_v11_0_average_bandwidth_vs_available_bandwidth(&wm_low) ||
-		    !dce_v11_0_check_latency_hiding(&wm_low) ||
-		    (adev->mode_info.disp_priority == 2)) {
-			DRM_DEBUG_KMS("force priority to high\n");
-		}
-		lb_vblank_lead_lines = DIV_ROUND_UP(lb_size, mode->crtc_hdisplay);
-	}
-
-	/* select wm A */
-	wm_mask = RREG32(mmDPG_WATERMARK_MASK_CONTROL + amdgpu_crtc->crtc_offset);
-	tmp = REG_SET_FIELD(wm_mask, DPG_WATERMARK_MASK_CONTROL, URGENCY_WATERMARK_MASK, 1);
-	WREG32(mmDPG_WATERMARK_MASK_CONTROL + amdgpu_crtc->crtc_offset, tmp);
-	tmp = RREG32(mmDPG_PIPE_URGENCY_CONTROL + amdgpu_crtc->crtc_offset);
-	tmp = REG_SET_FIELD(tmp, DPG_PIPE_URGENCY_CONTROL, URGENCY_LOW_WATERMARK, latency_watermark_a);
-	tmp = REG_SET_FIELD(tmp, DPG_PIPE_URGENCY_CONTROL, URGENCY_HIGH_WATERMARK, line_time);
-	WREG32(mmDPG_PIPE_URGENCY_CONTROL + amdgpu_crtc->crtc_offset, tmp);
-	/* select wm B */
-	tmp = REG_SET_FIELD(wm_mask, DPG_WATERMARK_MASK_CONTROL, URGENCY_WATERMARK_MASK, 2);
-	WREG32(mmDPG_WATERMARK_MASK_CONTROL + amdgpu_crtc->crtc_offset, tmp);
-	tmp = RREG32(mmDPG_PIPE_URGENCY_CONTROL + amdgpu_crtc->crtc_offset);
-	tmp = REG_SET_FIELD(tmp, DPG_PIPE_URGENCY_CONTROL, URGENCY_LOW_WATERMARK, latency_watermark_b);
-	tmp = REG_SET_FIELD(tmp, DPG_PIPE_URGENCY_CONTROL, URGENCY_HIGH_WATERMARK, line_time);
-	WREG32(mmDPG_PIPE_URGENCY_CONTROL + amdgpu_crtc->crtc_offset, tmp);
-	/* restore original selection */
-	WREG32(mmDPG_WATERMARK_MASK_CONTROL + amdgpu_crtc->crtc_offset, wm_mask);
-
-	/* save values for DPM */
-	amdgpu_crtc->line_time = line_time;
-	amdgpu_crtc->wm_high = latency_watermark_a;
-	amdgpu_crtc->wm_low = latency_watermark_b;
-	/* Save number of lines the linebuffer leads before the scanout */
-	amdgpu_crtc->lb_vblank_lead_lines = lb_vblank_lead_lines;
-}
-
-/**
- * dce_v11_0_bandwidth_update - program display watermarks
- *
- * @adev: amdgpu_device pointer
- *
- * Calculate and program the display watermarks and line
- * buffer allocation (CIK).
- */
-static void dce_v11_0_bandwidth_update(struct amdgpu_device *adev)
-{
-	struct drm_display_mode *mode = NULL;
-	u32 num_heads = 0, lb_size;
-	int i;
-
-	amdgpu_display_update_priority(adev);
-
-	for (i = 0; i < adev->mode_info.num_crtc; i++) {
-		if (adev->mode_info.crtcs[i]->base.enabled)
-			num_heads++;
-	}
-	for (i = 0; i < adev->mode_info.num_crtc; i++) {
-		mode = &adev->mode_info.crtcs[i]->base.mode;
-		lb_size = dce_v11_0_line_buffer_adjust(adev, adev->mode_info.crtcs[i], mode);
-		dce_v11_0_program_watermarks(adev, adev->mode_info.crtcs[i],
-					    lb_size, num_heads);
-	}
-}
-
-static void dce_v11_0_audio_get_connected_pins(struct amdgpu_device *adev)
-{
-	int i;
-	u32 offset, tmp;
-
-	for (i = 0; i < adev->mode_info.audio.num_pins; i++) {
-		offset = adev->mode_info.audio.pin[i].offset;
-		tmp = RREG32_AUDIO_ENDPT(offset,
-					 ixAZALIA_F0_CODEC_PIN_CONTROL_RESPONSE_CONFIGURATION_DEFAULT);
-		if (((tmp &
-		AZALIA_F0_CODEC_PIN_CONTROL_RESPONSE_CONFIGURATION_DEFAULT__PORT_CONNECTIVITY_MASK) >>
-		AZALIA_F0_CODEC_PIN_CONTROL_RESPONSE_CONFIGURATION_DEFAULT__PORT_CONNECTIVITY__SHIFT) == 1)
-			adev->mode_info.audio.pin[i].connected = false;
-		else
-			adev->mode_info.audio.pin[i].connected = true;
-	}
-}
-
-static struct amdgpu_audio_pin *dce_v11_0_audio_get_pin(struct amdgpu_device *adev)
-{
-	int i;
-
-	dce_v11_0_audio_get_connected_pins(adev);
-
-	for (i = 0; i < adev->mode_info.audio.num_pins; i++) {
-		if (adev->mode_info.audio.pin[i].connected)
-			return &adev->mode_info.audio.pin[i];
-	}
-	DRM_ERROR("No connected audio pins found!\n");
-	return NULL;
-}
-
-static void dce_v11_0_afmt_audio_select_pin(struct drm_encoder *encoder)
-{
-	struct amdgpu_device *adev = drm_to_adev(encoder->dev);
-	struct amdgpu_encoder *amdgpu_encoder = to_amdgpu_encoder(encoder);
-	struct amdgpu_encoder_atom_dig *dig = amdgpu_encoder->enc_priv;
-	u32 tmp;
-
-	if (!dig || !dig->afmt || !dig->afmt->pin)
-		return;
-
-	tmp = RREG32(mmAFMT_AUDIO_SRC_CONTROL + dig->afmt->offset);
-	tmp = REG_SET_FIELD(tmp, AFMT_AUDIO_SRC_CONTROL, AFMT_AUDIO_SRC_SELECT, dig->afmt->pin->id);
-	WREG32(mmAFMT_AUDIO_SRC_CONTROL + dig->afmt->offset, tmp);
-}
-
-static void dce_v11_0_audio_write_latency_fields(struct drm_encoder *encoder,
-						struct drm_display_mode *mode)
-{
-	struct drm_device *dev = encoder->dev;
-	struct amdgpu_device *adev = drm_to_adev(dev);
-	struct amdgpu_encoder *amdgpu_encoder = to_amdgpu_encoder(encoder);
-	struct amdgpu_encoder_atom_dig *dig = amdgpu_encoder->enc_priv;
-	struct drm_connector *connector;
-	struct drm_connector_list_iter iter;
-	struct amdgpu_connector *amdgpu_connector = NULL;
-	u32 tmp;
-	int interlace = 0;
-
-	if (!dig || !dig->afmt || !dig->afmt->pin)
-		return;
-
-	drm_connector_list_iter_begin(dev, &iter);
-	drm_for_each_connector_iter(connector, &iter) {
-		if (connector->encoder == encoder) {
-			amdgpu_connector = to_amdgpu_connector(connector);
-			break;
-		}
-	}
-	drm_connector_list_iter_end(&iter);
-
-	if (!amdgpu_connector) {
-		DRM_ERROR("Couldn't find encoder's connector\n");
-		return;
-	}
-
-	if (mode->flags & DRM_MODE_FLAG_INTERLACE)
-		interlace = 1;
-	if (connector->latency_present[interlace]) {
-		tmp = REG_SET_FIELD(0, AZALIA_F0_CODEC_PIN_CONTROL_RESPONSE_LIPSYNC,
-				    VIDEO_LIPSYNC, connector->video_latency[interlace]);
-		tmp = REG_SET_FIELD(0, AZALIA_F0_CODEC_PIN_CONTROL_RESPONSE_LIPSYNC,
-				    AUDIO_LIPSYNC, connector->audio_latency[interlace]);
-	} else {
-		tmp = REG_SET_FIELD(0, AZALIA_F0_CODEC_PIN_CONTROL_RESPONSE_LIPSYNC,
-				    VIDEO_LIPSYNC, 0);
-		tmp = REG_SET_FIELD(0, AZALIA_F0_CODEC_PIN_CONTROL_RESPONSE_LIPSYNC,
-				    AUDIO_LIPSYNC, 0);
-	}
-	WREG32_AUDIO_ENDPT(dig->afmt->pin->offset,
-			   ixAZALIA_F0_CODEC_PIN_CONTROL_RESPONSE_LIPSYNC, tmp);
-}
-
-static void dce_v11_0_audio_write_speaker_allocation(struct drm_encoder *encoder)
-{
-	struct drm_device *dev = encoder->dev;
-	struct amdgpu_device *adev = drm_to_adev(dev);
-	struct amdgpu_encoder *amdgpu_encoder = to_amdgpu_encoder(encoder);
-	struct amdgpu_encoder_atom_dig *dig = amdgpu_encoder->enc_priv;
-	struct drm_connector *connector;
-	struct drm_connector_list_iter iter;
-	struct amdgpu_connector *amdgpu_connector = NULL;
-	u32 tmp;
-	u8 *sadb = NULL;
-	int sad_count;
-
-	if (!dig || !dig->afmt || !dig->afmt->pin)
-		return;
-
-	drm_connector_list_iter_begin(dev, &iter);
-	drm_for_each_connector_iter(connector, &iter) {
-		if (connector->encoder == encoder) {
-			amdgpu_connector = to_amdgpu_connector(connector);
-			break;
-		}
-	}
-	drm_connector_list_iter_end(&iter);
-
-	if (!amdgpu_connector) {
-		DRM_ERROR("Couldn't find encoder's connector\n");
-		return;
-	}
-
-	sad_count = drm_edid_to_speaker_allocation(amdgpu_connector_edid(connector), &sadb);
-	if (sad_count < 0) {
-		DRM_ERROR("Couldn't read Speaker Allocation Data Block: %d\n", sad_count);
-		sad_count = 0;
-	}
-
-	/* program the speaker allocation */
-	tmp = RREG32_AUDIO_ENDPT(dig->afmt->pin->offset,
-				 ixAZALIA_F0_CODEC_PIN_CONTROL_CHANNEL_SPEAKER);
-	tmp = REG_SET_FIELD(tmp, AZALIA_F0_CODEC_PIN_CONTROL_CHANNEL_SPEAKER,
-			    DP_CONNECTION, 0);
-	/* set HDMI mode */
-	tmp = REG_SET_FIELD(tmp, AZALIA_F0_CODEC_PIN_CONTROL_CHANNEL_SPEAKER,
-			    HDMI_CONNECTION, 1);
-	if (sad_count)
-		tmp = REG_SET_FIELD(tmp, AZALIA_F0_CODEC_PIN_CONTROL_CHANNEL_SPEAKER,
-				    SPEAKER_ALLOCATION, sadb[0]);
-	else
-		tmp = REG_SET_FIELD(tmp, AZALIA_F0_CODEC_PIN_CONTROL_CHANNEL_SPEAKER,
-				    SPEAKER_ALLOCATION, 5); /* stereo */
-	WREG32_AUDIO_ENDPT(dig->afmt->pin->offset,
-			   ixAZALIA_F0_CODEC_PIN_CONTROL_CHANNEL_SPEAKER, tmp);
-
-	kfree(sadb);
-}
-
-static void dce_v11_0_audio_write_sad_regs(struct drm_encoder *encoder)
-{
-	struct drm_device *dev = encoder->dev;
-	struct amdgpu_device *adev = drm_to_adev(dev);
-	struct amdgpu_encoder *amdgpu_encoder = to_amdgpu_encoder(encoder);
-	struct amdgpu_encoder_atom_dig *dig = amdgpu_encoder->enc_priv;
-	struct drm_connector *connector;
-	struct drm_connector_list_iter iter;
-	struct amdgpu_connector *amdgpu_connector = NULL;
-	struct cea_sad *sads;
-	int i, sad_count;
-
-	static const u16 eld_reg_to_type[][2] = {
-		{ ixAZALIA_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR0, HDMI_AUDIO_CODING_TYPE_PCM },
-		{ ixAZALIA_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR1, HDMI_AUDIO_CODING_TYPE_AC3 },
-		{ ixAZALIA_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR2, HDMI_AUDIO_CODING_TYPE_MPEG1 },
-		{ ixAZALIA_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR3, HDMI_AUDIO_CODING_TYPE_MP3 },
-		{ ixAZALIA_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR4, HDMI_AUDIO_CODING_TYPE_MPEG2 },
-		{ ixAZALIA_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR5, HDMI_AUDIO_CODING_TYPE_AAC_LC },
-		{ ixAZALIA_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR6, HDMI_AUDIO_CODING_TYPE_DTS },
-		{ ixAZALIA_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR7, HDMI_AUDIO_CODING_TYPE_ATRAC },
-		{ ixAZALIA_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR9, HDMI_AUDIO_CODING_TYPE_EAC3 },
-		{ ixAZALIA_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR10, HDMI_AUDIO_CODING_TYPE_DTS_HD },
-		{ ixAZALIA_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR11, HDMI_AUDIO_CODING_TYPE_MLP },
-		{ ixAZALIA_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR13, HDMI_AUDIO_CODING_TYPE_WMA_PRO },
-	};
-
-	if (!dig || !dig->afmt || !dig->afmt->pin)
-		return;
-
-	drm_connector_list_iter_begin(dev, &iter);
-	drm_for_each_connector_iter(connector, &iter) {
-		if (connector->encoder == encoder) {
-			amdgpu_connector = to_amdgpu_connector(connector);
-			break;
-		}
-	}
-	drm_connector_list_iter_end(&iter);
-
-	if (!amdgpu_connector) {
-		DRM_ERROR("Couldn't find encoder's connector\n");
-		return;
-	}
-
-	sad_count = drm_edid_to_sad(amdgpu_connector_edid(connector), &sads);
-	if (sad_count < 0)
-		DRM_ERROR("Couldn't read SADs: %d\n", sad_count);
-	if (sad_count <= 0)
-		return;
-	BUG_ON(!sads);
-
-	for (i = 0; i < ARRAY_SIZE(eld_reg_to_type); i++) {
-		u32 tmp = 0;
-		u8 stereo_freqs = 0;
-		int max_channels = -1;
-		int j;
-
-		for (j = 0; j < sad_count; j++) {
-			struct cea_sad *sad = &sads[j];
-
-			if (sad->format == eld_reg_to_type[i][1]) {
-				if (sad->channels > max_channels) {
-					tmp = REG_SET_FIELD(tmp, AZALIA_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR0,
-							    MAX_CHANNELS, sad->channels);
-					tmp = REG_SET_FIELD(tmp, AZALIA_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR0,
-							    DESCRIPTOR_BYTE_2, sad->byte2);
-					tmp = REG_SET_FIELD(tmp, AZALIA_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR0,
-							    SUPPORTED_FREQUENCIES, sad->freq);
-					max_channels = sad->channels;
-				}
-
-				if (sad->format == HDMI_AUDIO_CODING_TYPE_PCM)
-					stereo_freqs |= sad->freq;
-				else
-					break;
-			}
-		}
-
-		tmp = REG_SET_FIELD(tmp, AZALIA_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR0,
-				    SUPPORTED_FREQUENCIES_STEREO, stereo_freqs);
-		WREG32_AUDIO_ENDPT(dig->afmt->pin->offset, eld_reg_to_type[i][0], tmp);
-	}
-
-	kfree(sads);
-}
-
-static void dce_v11_0_audio_enable(struct amdgpu_device *adev,
-				  struct amdgpu_audio_pin *pin,
-				  bool enable)
-{
-	if (!pin)
-		return;
-
-	WREG32_AUDIO_ENDPT(pin->offset, ixAZALIA_F0_CODEC_PIN_CONTROL_HOT_PLUG_CONTROL,
-			   enable ? AZALIA_F0_CODEC_PIN_CONTROL_HOT_PLUG_CONTROL__AUDIO_ENABLED_MASK : 0);
-}
-
-static const u32 pin_offsets[] =
-{
-	AUD0_REGISTER_OFFSET,
-	AUD1_REGISTER_OFFSET,
-	AUD2_REGISTER_OFFSET,
-	AUD3_REGISTER_OFFSET,
-	AUD4_REGISTER_OFFSET,
-	AUD5_REGISTER_OFFSET,
-	AUD6_REGISTER_OFFSET,
-	AUD7_REGISTER_OFFSET,
-};
-
-static int dce_v11_0_audio_init(struct amdgpu_device *adev)
-{
-	int i;
-
-	if (!amdgpu_audio)
-		return 0;
-
-	adev->mode_info.audio.enabled = true;
-
-	switch (adev->asic_type) {
-	case CHIP_CARRIZO:
-	case CHIP_STONEY:
-		adev->mode_info.audio.num_pins = 7;
-		break;
-	case CHIP_POLARIS10:
-	case CHIP_VEGAM:
-		adev->mode_info.audio.num_pins = 8;
-		break;
-	case CHIP_POLARIS11:
-	case CHIP_POLARIS12:
-		adev->mode_info.audio.num_pins = 6;
-		break;
-	default:
-		return -EINVAL;
-	}
-
-	for (i = 0; i < adev->mode_info.audio.num_pins; i++) {
-		adev->mode_info.audio.pin[i].channels = -1;
-		adev->mode_info.audio.pin[i].rate = -1;
-		adev->mode_info.audio.pin[i].bits_per_sample = -1;
-		adev->mode_info.audio.pin[i].status_bits = 0;
-		adev->mode_info.audio.pin[i].category_code = 0;
-		adev->mode_info.audio.pin[i].connected = false;
-		adev->mode_info.audio.pin[i].offset = pin_offsets[i];
-		adev->mode_info.audio.pin[i].id = i;
-		/* disable audio.  it will be set up later */
-		/* XXX remove once we switch to ip funcs */
-		dce_v11_0_audio_enable(adev, &adev->mode_info.audio.pin[i], false);
-	}
-
-	return 0;
-}
-
-static void dce_v11_0_audio_fini(struct amdgpu_device *adev)
-{
-	int i;
-
-	if (!amdgpu_audio)
-		return;
-
-	if (!adev->mode_info.audio.enabled)
-		return;
-
-	for (i = 0; i < adev->mode_info.audio.num_pins; i++)
-		dce_v11_0_audio_enable(adev, &adev->mode_info.audio.pin[i], false);
-
-	adev->mode_info.audio.enabled = false;
-}
-
-/*
- * update the N and CTS parameters for a given pixel clock rate
- */
-static void dce_v11_0_afmt_update_ACR(struct drm_encoder *encoder, uint32_t clock)
-{
-	struct drm_device *dev = encoder->dev;
-	struct amdgpu_device *adev = drm_to_adev(dev);
-	struct amdgpu_afmt_acr acr = amdgpu_afmt_acr(clock);
-	struct amdgpu_encoder *amdgpu_encoder = to_amdgpu_encoder(encoder);
-	struct amdgpu_encoder_atom_dig *dig = amdgpu_encoder->enc_priv;
-	u32 tmp;
-
-	tmp = RREG32(mmHDMI_ACR_32_0 + dig->afmt->offset);
-	tmp = REG_SET_FIELD(tmp, HDMI_ACR_32_0, HDMI_ACR_CTS_32, acr.cts_32khz);
-	WREG32(mmHDMI_ACR_32_0 + dig->afmt->offset, tmp);
-	tmp = RREG32(mmHDMI_ACR_32_1 + dig->afmt->offset);
-	tmp = REG_SET_FIELD(tmp, HDMI_ACR_32_1, HDMI_ACR_N_32, acr.n_32khz);
-	WREG32(mmHDMI_ACR_32_1 + dig->afmt->offset, tmp);
-
-	tmp = RREG32(mmHDMI_ACR_44_0 + dig->afmt->offset);
-	tmp = REG_SET_FIELD(tmp, HDMI_ACR_44_0, HDMI_ACR_CTS_44, acr.cts_44_1khz);
-	WREG32(mmHDMI_ACR_44_0 + dig->afmt->offset, tmp);
-	tmp = RREG32(mmHDMI_ACR_44_1 + dig->afmt->offset);
-	tmp = REG_SET_FIELD(tmp, HDMI_ACR_44_1, HDMI_ACR_N_44, acr.n_44_1khz);
-	WREG32(mmHDMI_ACR_44_1 + dig->afmt->offset, tmp);
-
-	tmp = RREG32(mmHDMI_ACR_48_0 + dig->afmt->offset);
-	tmp = REG_SET_FIELD(tmp, HDMI_ACR_48_0, HDMI_ACR_CTS_48, acr.cts_48khz);
-	WREG32(mmHDMI_ACR_48_0 + dig->afmt->offset, tmp);
-	tmp = RREG32(mmHDMI_ACR_48_1 + dig->afmt->offset);
-	tmp = REG_SET_FIELD(tmp, HDMI_ACR_48_1, HDMI_ACR_N_48, acr.n_48khz);
-	WREG32(mmHDMI_ACR_48_1 + dig->afmt->offset, tmp);
-
-}
-
-/*
- * build a HDMI Video Info Frame
- */
-static void dce_v11_0_afmt_update_avi_infoframe(struct drm_encoder *encoder,
-					       void *buffer, size_t size)
-{
-	struct drm_device *dev = encoder->dev;
-	struct amdgpu_device *adev = drm_to_adev(dev);
-	struct amdgpu_encoder *amdgpu_encoder = to_amdgpu_encoder(encoder);
-	struct amdgpu_encoder_atom_dig *dig = amdgpu_encoder->enc_priv;
-	uint8_t *frame = buffer + 3;
-	uint8_t *header = buffer;
-
-	WREG32(mmAFMT_AVI_INFO0 + dig->afmt->offset,
-		frame[0x0] | (frame[0x1] << 8) | (frame[0x2] << 16) | (frame[0x3] << 24));
-	WREG32(mmAFMT_AVI_INFO1 + dig->afmt->offset,
-		frame[0x4] | (frame[0x5] << 8) | (frame[0x6] << 16) | (frame[0x7] << 24));
-	WREG32(mmAFMT_AVI_INFO2 + dig->afmt->offset,
-		frame[0x8] | (frame[0x9] << 8) | (frame[0xA] << 16) | (frame[0xB] << 24));
-	WREG32(mmAFMT_AVI_INFO3 + dig->afmt->offset,
-		frame[0xC] | (frame[0xD] << 8) | (header[1] << 24));
-}
-
-static void dce_v11_0_audio_set_dto(struct drm_encoder *encoder, u32 clock)
-{
-	struct drm_device *dev = encoder->dev;
-	struct amdgpu_device *adev = drm_to_adev(dev);
-	struct amdgpu_encoder *amdgpu_encoder = to_amdgpu_encoder(encoder);
-	struct amdgpu_encoder_atom_dig *dig = amdgpu_encoder->enc_priv;
-	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(encoder->crtc);
-	u32 dto_phase = 24 * 1000;
-	u32 dto_modulo = clock;
-	u32 tmp;
-
-	if (!dig || !dig->afmt)
-		return;
-
-	/* XXX two dtos; generally use dto0 for hdmi */
-	/* Express [24MHz / target pixel clock] as an exact rational
-	 * number (coefficient of two integer numbers.  DCCG_AUDIO_DTOx_PHASE
-	 * is the numerator, DCCG_AUDIO_DTOx_MODULE is the denominator
-	 */
-	tmp = RREG32(mmDCCG_AUDIO_DTO_SOURCE);
-	tmp = REG_SET_FIELD(tmp, DCCG_AUDIO_DTO_SOURCE, DCCG_AUDIO_DTO0_SOURCE_SEL,
-			    amdgpu_crtc->crtc_id);
-	WREG32(mmDCCG_AUDIO_DTO_SOURCE, tmp);
-	WREG32(mmDCCG_AUDIO_DTO0_PHASE, dto_phase);
-	WREG32(mmDCCG_AUDIO_DTO0_MODULE, dto_modulo);
-}
-
-/*
- * update the info frames with the data from the current display mode
- */
-static void dce_v11_0_afmt_setmode(struct drm_encoder *encoder,
-				  struct drm_display_mode *mode)
-{
-	struct drm_device *dev = encoder->dev;
-	struct amdgpu_device *adev = drm_to_adev(dev);
-	struct amdgpu_encoder *amdgpu_encoder = to_amdgpu_encoder(encoder);
-	struct amdgpu_encoder_atom_dig *dig = amdgpu_encoder->enc_priv;
-	struct drm_connector *connector = amdgpu_get_connector_for_encoder(encoder);
-	u8 buffer[HDMI_INFOFRAME_HEADER_SIZE + HDMI_AVI_INFOFRAME_SIZE];
-	struct hdmi_avi_infoframe frame;
-	ssize_t err;
-	u32 tmp;
-	int bpc = 8;
-
-	if (!dig || !dig->afmt)
-		return;
-
-	/* Silent, r600_hdmi_enable will raise WARN for us */
-	if (!dig->afmt->enabled)
-		return;
-
-	/* hdmi deep color mode general control packets setup, if bpc > 8 */
-	if (encoder->crtc) {
-		struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(encoder->crtc);
-		bpc = amdgpu_crtc->bpc;
-	}
-
-	/* disable audio prior to setting up hw */
-	dig->afmt->pin = dce_v11_0_audio_get_pin(adev);
-	dce_v11_0_audio_enable(adev, dig->afmt->pin, false);
-
-	dce_v11_0_audio_set_dto(encoder, mode->clock);
-
-	tmp = RREG32(mmHDMI_VBI_PACKET_CONTROL + dig->afmt->offset);
-	tmp = REG_SET_FIELD(tmp, HDMI_VBI_PACKET_CONTROL, HDMI_NULL_SEND, 1);
-	WREG32(mmHDMI_VBI_PACKET_CONTROL + dig->afmt->offset, tmp); /* send null packets when required */
-
-	WREG32(mmAFMT_AUDIO_CRC_CONTROL + dig->afmt->offset, 0x1000);
-
-	tmp = RREG32(mmHDMI_CONTROL + dig->afmt->offset);
-	switch (bpc) {
-	case 0:
-	case 6:
-	case 8:
-	case 16:
-	default:
-		tmp = REG_SET_FIELD(tmp, HDMI_CONTROL, HDMI_DEEP_COLOR_ENABLE, 0);
-		tmp = REG_SET_FIELD(tmp, HDMI_CONTROL, HDMI_DEEP_COLOR_DEPTH, 0);
-		DRM_DEBUG("%s: Disabling hdmi deep color for %d bpc.\n",
-			  connector->name, bpc);
-		break;
-	case 10:
-		tmp = REG_SET_FIELD(tmp, HDMI_CONTROL, HDMI_DEEP_COLOR_ENABLE, 1);
-		tmp = REG_SET_FIELD(tmp, HDMI_CONTROL, HDMI_DEEP_COLOR_DEPTH, 1);
-		DRM_DEBUG("%s: Enabling hdmi deep color 30 for 10 bpc.\n",
-			  connector->name);
-		break;
-	case 12:
-		tmp = REG_SET_FIELD(tmp, HDMI_CONTROL, HDMI_DEEP_COLOR_ENABLE, 1);
-		tmp = REG_SET_FIELD(tmp, HDMI_CONTROL, HDMI_DEEP_COLOR_DEPTH, 2);
-		DRM_DEBUG("%s: Enabling hdmi deep color 36 for 12 bpc.\n",
-			  connector->name);
-		break;
-	}
-	WREG32(mmHDMI_CONTROL + dig->afmt->offset, tmp);
-
-	tmp = RREG32(mmHDMI_VBI_PACKET_CONTROL + dig->afmt->offset);
-	tmp = REG_SET_FIELD(tmp, HDMI_VBI_PACKET_CONTROL, HDMI_NULL_SEND, 1); /* send null packets when required */
-	tmp = REG_SET_FIELD(tmp, HDMI_VBI_PACKET_CONTROL, HDMI_GC_SEND, 1); /* send general control packets */
-	tmp = REG_SET_FIELD(tmp, HDMI_VBI_PACKET_CONTROL, HDMI_GC_CONT, 1); /* send general control packets every frame */
-	WREG32(mmHDMI_VBI_PACKET_CONTROL + dig->afmt->offset, tmp);
-
-	tmp = RREG32(mmHDMI_INFOFRAME_CONTROL0 + dig->afmt->offset);
-	/* enable audio info frames (frames won't be set until audio is enabled) */
-	tmp = REG_SET_FIELD(tmp, HDMI_INFOFRAME_CONTROL0, HDMI_AUDIO_INFO_SEND, 1);
-	/* required for audio info values to be updated */
-	tmp = REG_SET_FIELD(tmp, HDMI_INFOFRAME_CONTROL0, HDMI_AUDIO_INFO_CONT, 1);
-	WREG32(mmHDMI_INFOFRAME_CONTROL0 + dig->afmt->offset, tmp);
-
-	tmp = RREG32(mmAFMT_INFOFRAME_CONTROL0 + dig->afmt->offset);
-	/* required for audio info values to be updated */
-	tmp = REG_SET_FIELD(tmp, AFMT_INFOFRAME_CONTROL0, AFMT_AUDIO_INFO_UPDATE, 1);
-	WREG32(mmAFMT_INFOFRAME_CONTROL0 + dig->afmt->offset, tmp);
-
-	tmp = RREG32(mmHDMI_INFOFRAME_CONTROL1 + dig->afmt->offset);
-	/* anything other than 0 */
-	tmp = REG_SET_FIELD(tmp, HDMI_INFOFRAME_CONTROL1, HDMI_AUDIO_INFO_LINE, 2);
-	WREG32(mmHDMI_INFOFRAME_CONTROL1 + dig->afmt->offset, tmp);
-
-	WREG32(mmHDMI_GC + dig->afmt->offset, 0); /* unset HDMI_GC_AVMUTE */
-
-	tmp = RREG32(mmHDMI_AUDIO_PACKET_CONTROL + dig->afmt->offset);
-	/* set the default audio delay */
-	tmp = REG_SET_FIELD(tmp, HDMI_AUDIO_PACKET_CONTROL, HDMI_AUDIO_DELAY_EN, 1);
-	/* should be suffient for all audio modes and small enough for all hblanks */
-	tmp = REG_SET_FIELD(tmp, HDMI_AUDIO_PACKET_CONTROL, HDMI_AUDIO_PACKETS_PER_LINE, 3);
-	WREG32(mmHDMI_AUDIO_PACKET_CONTROL + dig->afmt->offset, tmp);
-
-	tmp = RREG32(mmAFMT_AUDIO_PACKET_CONTROL + dig->afmt->offset);
-	/* allow 60958 channel status fields to be updated */
-	tmp = REG_SET_FIELD(tmp, AFMT_AUDIO_PACKET_CONTROL, AFMT_60958_CS_UPDATE, 1);
-	WREG32(mmAFMT_AUDIO_PACKET_CONTROL + dig->afmt->offset, tmp);
-
-	tmp = RREG32(mmHDMI_ACR_PACKET_CONTROL + dig->afmt->offset);
-	if (bpc > 8)
-		/* clear SW CTS value */
-		tmp = REG_SET_FIELD(tmp, HDMI_ACR_PACKET_CONTROL, HDMI_ACR_SOURCE, 0);
-	else
-		/* select SW CTS value */
-		tmp = REG_SET_FIELD(tmp, HDMI_ACR_PACKET_CONTROL, HDMI_ACR_SOURCE, 1);
-	/* allow hw to sent ACR packets when required */
-	tmp = REG_SET_FIELD(tmp, HDMI_ACR_PACKET_CONTROL, HDMI_ACR_AUTO_SEND, 1);
-	WREG32(mmHDMI_ACR_PACKET_CONTROL + dig->afmt->offset, tmp);
-
-	dce_v11_0_afmt_update_ACR(encoder, mode->clock);
-
-	tmp = RREG32(mmAFMT_60958_0 + dig->afmt->offset);
-	tmp = REG_SET_FIELD(tmp, AFMT_60958_0, AFMT_60958_CS_CHANNEL_NUMBER_L, 1);
-	WREG32(mmAFMT_60958_0 + dig->afmt->offset, tmp);
-
-	tmp = RREG32(mmAFMT_60958_1 + dig->afmt->offset);
-	tmp = REG_SET_FIELD(tmp, AFMT_60958_1, AFMT_60958_CS_CHANNEL_NUMBER_R, 2);
-	WREG32(mmAFMT_60958_1 + dig->afmt->offset, tmp);
-
-	tmp = RREG32(mmAFMT_60958_2 + dig->afmt->offset);
-	tmp = REG_SET_FIELD(tmp, AFMT_60958_2, AFMT_60958_CS_CHANNEL_NUMBER_2, 3);
-	tmp = REG_SET_FIELD(tmp, AFMT_60958_2, AFMT_60958_CS_CHANNEL_NUMBER_3, 4);
-	tmp = REG_SET_FIELD(tmp, AFMT_60958_2, AFMT_60958_CS_CHANNEL_NUMBER_4, 5);
-	tmp = REG_SET_FIELD(tmp, AFMT_60958_2, AFMT_60958_CS_CHANNEL_NUMBER_5, 6);
-	tmp = REG_SET_FIELD(tmp, AFMT_60958_2, AFMT_60958_CS_CHANNEL_NUMBER_6, 7);
-	tmp = REG_SET_FIELD(tmp, AFMT_60958_2, AFMT_60958_CS_CHANNEL_NUMBER_7, 8);
-	WREG32(mmAFMT_60958_2 + dig->afmt->offset, tmp);
-
-	dce_v11_0_audio_write_speaker_allocation(encoder);
-
-	WREG32(mmAFMT_AUDIO_PACKET_CONTROL2 + dig->afmt->offset,
-	       (0xff << AFMT_AUDIO_PACKET_CONTROL2__AFMT_AUDIO_CHANNEL_ENABLE__SHIFT));
-
-	dce_v11_0_afmt_audio_select_pin(encoder);
-	dce_v11_0_audio_write_sad_regs(encoder);
-	dce_v11_0_audio_write_latency_fields(encoder, mode);
-
-	err = drm_hdmi_avi_infoframe_from_display_mode(&frame, connector, mode);
-	if (err < 0) {
-		DRM_ERROR("failed to setup AVI infoframe: %zd\n", err);
-		return;
-	}
-
-	err = hdmi_avi_infoframe_pack(&frame, buffer, sizeof(buffer));
-	if (err < 0) {
-		DRM_ERROR("failed to pack AVI infoframe: %zd\n", err);
-		return;
-	}
-
-	dce_v11_0_afmt_update_avi_infoframe(encoder, buffer, sizeof(buffer));
-
-	tmp = RREG32(mmHDMI_INFOFRAME_CONTROL0 + dig->afmt->offset);
-	/* enable AVI info frames */
-	tmp = REG_SET_FIELD(tmp, HDMI_INFOFRAME_CONTROL0, HDMI_AVI_INFO_SEND, 1);
-	/* required for audio info values to be updated */
-	tmp = REG_SET_FIELD(tmp, HDMI_INFOFRAME_CONTROL0, HDMI_AVI_INFO_CONT, 1);
-	WREG32(mmHDMI_INFOFRAME_CONTROL0 + dig->afmt->offset, tmp);
-
-	tmp = RREG32(mmHDMI_INFOFRAME_CONTROL1 + dig->afmt->offset);
-	tmp = REG_SET_FIELD(tmp, HDMI_INFOFRAME_CONTROL1, HDMI_AVI_INFO_LINE, 2);
-	WREG32(mmHDMI_INFOFRAME_CONTROL1 + dig->afmt->offset, tmp);
-
-	tmp = RREG32(mmAFMT_AUDIO_PACKET_CONTROL + dig->afmt->offset);
-	/* send audio packets */
-	tmp = REG_SET_FIELD(tmp, AFMT_AUDIO_PACKET_CONTROL, AFMT_AUDIO_SAMPLE_SEND, 1);
-	WREG32(mmAFMT_AUDIO_PACKET_CONTROL + dig->afmt->offset, tmp);
-
-	WREG32(mmAFMT_RAMP_CONTROL0 + dig->afmt->offset, 0x00FFFFFF);
-	WREG32(mmAFMT_RAMP_CONTROL1 + dig->afmt->offset, 0x007FFFFF);
-	WREG32(mmAFMT_RAMP_CONTROL2 + dig->afmt->offset, 0x00000001);
-	WREG32(mmAFMT_RAMP_CONTROL3 + dig->afmt->offset, 0x00000001);
-
-	/* enable audio after to setting up hw */
-	dce_v11_0_audio_enable(adev, dig->afmt->pin, true);
-}
-
-static void dce_v11_0_afmt_enable(struct drm_encoder *encoder, bool enable)
-{
-	struct drm_device *dev = encoder->dev;
-	struct amdgpu_device *adev = drm_to_adev(dev);
-	struct amdgpu_encoder *amdgpu_encoder = to_amdgpu_encoder(encoder);
-	struct amdgpu_encoder_atom_dig *dig = amdgpu_encoder->enc_priv;
-
-	if (!dig || !dig->afmt)
-		return;
-
-	/* Silent, r600_hdmi_enable will raise WARN for us */
-	if (enable && dig->afmt->enabled)
-		return;
-	if (!enable && !dig->afmt->enabled)
-		return;
-
-	if (!enable && dig->afmt->pin) {
-		dce_v11_0_audio_enable(adev, dig->afmt->pin, false);
-		dig->afmt->pin = NULL;
-	}
-
-	dig->afmt->enabled = enable;
-
-	DRM_DEBUG("%sabling AFMT interface @ 0x%04X for encoder 0x%x\n",
-		  enable ? "En" : "Dis", dig->afmt->offset, amdgpu_encoder->encoder_id);
-}
-
-static int dce_v11_0_afmt_init(struct amdgpu_device *adev)
-{
-	int i;
-
-	for (i = 0; i < adev->mode_info.num_dig; i++)
-		adev->mode_info.afmt[i] = NULL;
-
-	/* DCE11 has audio blocks tied to DIG encoders */
-	for (i = 0; i < adev->mode_info.num_dig; i++) {
-		adev->mode_info.afmt[i] = kzalloc(sizeof(struct amdgpu_afmt), GFP_KERNEL);
-		if (adev->mode_info.afmt[i]) {
-			adev->mode_info.afmt[i]->offset = dig_offsets[i];
-			adev->mode_info.afmt[i]->id = i;
-		} else {
-			int j;
-			for (j = 0; j < i; j++) {
-				kfree(adev->mode_info.afmt[j]);
-				adev->mode_info.afmt[j] = NULL;
-			}
-			return -ENOMEM;
-		}
-	}
-	return 0;
-}
-
-static void dce_v11_0_afmt_fini(struct amdgpu_device *adev)
-{
-	int i;
-
-	for (i = 0; i < adev->mode_info.num_dig; i++) {
-		kfree(adev->mode_info.afmt[i]);
-		adev->mode_info.afmt[i] = NULL;
-	}
-}
-
-static const u32 vga_control_regs[6] =
-{
-	mmD1VGA_CONTROL,
-	mmD2VGA_CONTROL,
-	mmD3VGA_CONTROL,
-	mmD4VGA_CONTROL,
-	mmD5VGA_CONTROL,
-	mmD6VGA_CONTROL,
-};
-
-static void dce_v11_0_vga_enable(struct drm_crtc *crtc, bool enable)
-{
-	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
-	struct drm_device *dev = crtc->dev;
-	struct amdgpu_device *adev = drm_to_adev(dev);
-	u32 vga_control;
-
-	vga_control = RREG32(vga_control_regs[amdgpu_crtc->crtc_id]) & ~1;
-	if (enable)
-		WREG32(vga_control_regs[amdgpu_crtc->crtc_id], vga_control | 1);
-	else
-		WREG32(vga_control_regs[amdgpu_crtc->crtc_id], vga_control);
-}
-
-static void dce_v11_0_grph_enable(struct drm_crtc *crtc, bool enable)
-{
-	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
-	struct drm_device *dev = crtc->dev;
-	struct amdgpu_device *adev = drm_to_adev(dev);
-
-	if (enable)
-		WREG32(mmGRPH_ENABLE + amdgpu_crtc->crtc_offset, 1);
-	else
-		WREG32(mmGRPH_ENABLE + amdgpu_crtc->crtc_offset, 0);
-}
-
-static int dce_v11_0_crtc_do_set_base(struct drm_crtc *crtc,
-				     struct drm_framebuffer *fb,
-				     int x, int y, int atomic)
-{
-	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
-	struct drm_device *dev = crtc->dev;
-	struct amdgpu_device *adev = drm_to_adev(dev);
-	struct drm_framebuffer *target_fb;
-	struct drm_gem_object *obj;
-	struct amdgpu_bo *abo;
-	uint64_t fb_location, tiling_flags;
-	uint32_t fb_format, fb_pitch_pixels;
-	u32 fb_swap = REG_SET_FIELD(0, GRPH_SWAP_CNTL, GRPH_ENDIAN_SWAP, ENDIAN_NONE);
-	u32 pipe_config;
-	u32 tmp, viewport_w, viewport_h;
-	int r;
-	bool bypass_lut = false;
-
-	/* no fb bound */
-	if (!atomic && !crtc->primary->fb) {
-		DRM_DEBUG_KMS("No FB bound\n");
-		return 0;
-	}
-
-	if (atomic)
-		target_fb = fb;
-	else
-		target_fb = crtc->primary->fb;
-
-	/* If atomic, assume fb object is pinned & idle & fenced and
-	 * just update base pointers
-	 */
-	obj = target_fb->obj[0];
-	abo = gem_to_amdgpu_bo(obj);
-	r = amdgpu_bo_reserve(abo, false);
-	if (unlikely(r != 0))
-		return r;
-
-	if (!atomic) {
-		r = amdgpu_bo_pin(abo, AMDGPU_GEM_DOMAIN_VRAM);
-		if (unlikely(r != 0)) {
-			amdgpu_bo_unreserve(abo);
-			return -EINVAL;
-		}
-	}
-	fb_location = amdgpu_bo_gpu_offset(abo);
-
-	amdgpu_bo_get_tiling_flags(abo, &tiling_flags);
-	amdgpu_bo_unreserve(abo);
-
-	pipe_config = AMDGPU_TILING_GET(tiling_flags, PIPE_CONFIG);
-
-	switch (target_fb->format->format) {
-	case DRM_FORMAT_C8:
-		fb_format = REG_SET_FIELD(0, GRPH_CONTROL, GRPH_DEPTH, 0);
-		fb_format = REG_SET_FIELD(fb_format, GRPH_CONTROL, GRPH_FORMAT, 0);
-		break;
-	case DRM_FORMAT_XRGB4444:
-	case DRM_FORMAT_ARGB4444:
-		fb_format = REG_SET_FIELD(0, GRPH_CONTROL, GRPH_DEPTH, 1);
-		fb_format = REG_SET_FIELD(fb_format, GRPH_CONTROL, GRPH_FORMAT, 2);
-#ifdef __BIG_ENDIAN
-		fb_swap = REG_SET_FIELD(fb_swap, GRPH_SWAP_CNTL, GRPH_ENDIAN_SWAP,
-					ENDIAN_8IN16);
-#endif
-		break;
-	case DRM_FORMAT_XRGB1555:
-	case DRM_FORMAT_ARGB1555:
-		fb_format = REG_SET_FIELD(0, GRPH_CONTROL, GRPH_DEPTH, 1);
-		fb_format = REG_SET_FIELD(fb_format, GRPH_CONTROL, GRPH_FORMAT, 0);
-#ifdef __BIG_ENDIAN
-		fb_swap = REG_SET_FIELD(fb_swap, GRPH_SWAP_CNTL, GRPH_ENDIAN_SWAP,
-					ENDIAN_8IN16);
-#endif
-		break;
-	case DRM_FORMAT_BGRX5551:
-	case DRM_FORMAT_BGRA5551:
-		fb_format = REG_SET_FIELD(0, GRPH_CONTROL, GRPH_DEPTH, 1);
-		fb_format = REG_SET_FIELD(fb_format, GRPH_CONTROL, GRPH_FORMAT, 5);
-#ifdef __BIG_ENDIAN
-		fb_swap = REG_SET_FIELD(fb_swap, GRPH_SWAP_CNTL, GRPH_ENDIAN_SWAP,
-					ENDIAN_8IN16);
-#endif
-		break;
-	case DRM_FORMAT_RGB565:
-		fb_format = REG_SET_FIELD(0, GRPH_CONTROL, GRPH_DEPTH, 1);
-		fb_format = REG_SET_FIELD(fb_format, GRPH_CONTROL, GRPH_FORMAT, 1);
-#ifdef __BIG_ENDIAN
-		fb_swap = REG_SET_FIELD(fb_swap, GRPH_SWAP_CNTL, GRPH_ENDIAN_SWAP,
-					ENDIAN_8IN16);
-#endif
-		break;
-	case DRM_FORMAT_XRGB8888:
-	case DRM_FORMAT_ARGB8888:
-		fb_format = REG_SET_FIELD(0, GRPH_CONTROL, GRPH_DEPTH, 2);
-		fb_format = REG_SET_FIELD(fb_format, GRPH_CONTROL, GRPH_FORMAT, 0);
-#ifdef __BIG_ENDIAN
-		fb_swap = REG_SET_FIELD(fb_swap, GRPH_SWAP_CNTL, GRPH_ENDIAN_SWAP,
-					ENDIAN_8IN32);
-#endif
-		break;
-	case DRM_FORMAT_XRGB2101010:
-	case DRM_FORMAT_ARGB2101010:
-		fb_format = REG_SET_FIELD(0, GRPH_CONTROL, GRPH_DEPTH, 2);
-		fb_format = REG_SET_FIELD(fb_format, GRPH_CONTROL, GRPH_FORMAT, 1);
-#ifdef __BIG_ENDIAN
-		fb_swap = REG_SET_FIELD(fb_swap, GRPH_SWAP_CNTL, GRPH_ENDIAN_SWAP,
-					ENDIAN_8IN32);
-#endif
-		/* Greater 8 bpc fb needs to bypass hw-lut to retain precision */
-		bypass_lut = true;
-		break;
-	case DRM_FORMAT_BGRX1010102:
-	case DRM_FORMAT_BGRA1010102:
-		fb_format = REG_SET_FIELD(0, GRPH_CONTROL, GRPH_DEPTH, 2);
-		fb_format = REG_SET_FIELD(fb_format, GRPH_CONTROL, GRPH_FORMAT, 4);
-#ifdef __BIG_ENDIAN
-		fb_swap = REG_SET_FIELD(fb_swap, GRPH_SWAP_CNTL, GRPH_ENDIAN_SWAP,
-					ENDIAN_8IN32);
-#endif
-		/* Greater 8 bpc fb needs to bypass hw-lut to retain precision */
-		bypass_lut = true;
-		break;
-	case DRM_FORMAT_XBGR8888:
-	case DRM_FORMAT_ABGR8888:
-		fb_format = REG_SET_FIELD(0, GRPH_CONTROL, GRPH_DEPTH, 2);
-		fb_format = REG_SET_FIELD(fb_format, GRPH_CONTROL, GRPH_FORMAT, 0);
-		fb_swap = REG_SET_FIELD(fb_swap, GRPH_SWAP_CNTL, GRPH_RED_CROSSBAR, 2);
-		fb_swap = REG_SET_FIELD(fb_swap, GRPH_SWAP_CNTL, GRPH_BLUE_CROSSBAR, 2);
-#ifdef __BIG_ENDIAN
-		fb_swap = REG_SET_FIELD(fb_swap, GRPH_SWAP_CNTL, GRPH_ENDIAN_SWAP,
-					ENDIAN_8IN32);
-#endif
-		break;
-	default:
-		DRM_ERROR("Unsupported screen format %p4cc\n",
-			  &target_fb->format->format);
-		return -EINVAL;
-	}
-
-	if (AMDGPU_TILING_GET(tiling_flags, ARRAY_MODE) == ARRAY_2D_TILED_THIN1) {
-		unsigned bankw, bankh, mtaspect, tile_split, num_banks;
-
-		bankw = AMDGPU_TILING_GET(tiling_flags, BANK_WIDTH);
-		bankh = AMDGPU_TILING_GET(tiling_flags, BANK_HEIGHT);
-		mtaspect = AMDGPU_TILING_GET(tiling_flags, MACRO_TILE_ASPECT);
-		tile_split = AMDGPU_TILING_GET(tiling_flags, TILE_SPLIT);
-		num_banks = AMDGPU_TILING_GET(tiling_flags, NUM_BANKS);
-
-		fb_format = REG_SET_FIELD(fb_format, GRPH_CONTROL, GRPH_NUM_BANKS, num_banks);
-		fb_format = REG_SET_FIELD(fb_format, GRPH_CONTROL, GRPH_ARRAY_MODE,
-					  ARRAY_2D_TILED_THIN1);
-		fb_format = REG_SET_FIELD(fb_format, GRPH_CONTROL, GRPH_TILE_SPLIT,
-					  tile_split);
-		fb_format = REG_SET_FIELD(fb_format, GRPH_CONTROL, GRPH_BANK_WIDTH, bankw);
-		fb_format = REG_SET_FIELD(fb_format, GRPH_CONTROL, GRPH_BANK_HEIGHT, bankh);
-		fb_format = REG_SET_FIELD(fb_format, GRPH_CONTROL, GRPH_MACRO_TILE_ASPECT,
-					  mtaspect);
-		fb_format = REG_SET_FIELD(fb_format, GRPH_CONTROL, GRPH_MICRO_TILE_MODE,
-					  ADDR_SURF_MICRO_TILING_DISPLAY);
-	} else if (AMDGPU_TILING_GET(tiling_flags, ARRAY_MODE) == ARRAY_1D_TILED_THIN1) {
-		fb_format = REG_SET_FIELD(fb_format, GRPH_CONTROL, GRPH_ARRAY_MODE,
-					  ARRAY_1D_TILED_THIN1);
-	}
-
-	fb_format = REG_SET_FIELD(fb_format, GRPH_CONTROL, GRPH_PIPE_CONFIG,
-				  pipe_config);
-
-	dce_v11_0_vga_enable(crtc, false);
-
-	/* Make sure surface address is updated at vertical blank rather than
-	 * horizontal blank
-	 */
-	tmp = RREG32(mmGRPH_FLIP_CONTROL + amdgpu_crtc->crtc_offset);
-	tmp = REG_SET_FIELD(tmp, GRPH_FLIP_CONTROL,
-			    GRPH_SURFACE_UPDATE_H_RETRACE_EN, 0);
-	WREG32(mmGRPH_FLIP_CONTROL + amdgpu_crtc->crtc_offset, tmp);
-
-	WREG32(mmGRPH_PRIMARY_SURFACE_ADDRESS_HIGH + amdgpu_crtc->crtc_offset,
-	       upper_32_bits(fb_location));
-	WREG32(mmGRPH_SECONDARY_SURFACE_ADDRESS_HIGH + amdgpu_crtc->crtc_offset,
-	       upper_32_bits(fb_location));
-	WREG32(mmGRPH_PRIMARY_SURFACE_ADDRESS + amdgpu_crtc->crtc_offset,
-	       (u32)fb_location & GRPH_PRIMARY_SURFACE_ADDRESS__GRPH_PRIMARY_SURFACE_ADDRESS_MASK);
-	WREG32(mmGRPH_SECONDARY_SURFACE_ADDRESS + amdgpu_crtc->crtc_offset,
-	       (u32) fb_location & GRPH_SECONDARY_SURFACE_ADDRESS__GRPH_SECONDARY_SURFACE_ADDRESS_MASK);
-	WREG32(mmGRPH_CONTROL + amdgpu_crtc->crtc_offset, fb_format);
-	WREG32(mmGRPH_SWAP_CNTL + amdgpu_crtc->crtc_offset, fb_swap);
-
-	/*
-	 * The LUT only has 256 slots for indexing by a 8 bpc fb. Bypass the LUT
-	 * for > 8 bpc scanout to avoid truncation of fb indices to 8 msb's, to
-	 * retain the full precision throughout the pipeline.
-	 */
-	tmp = RREG32(mmGRPH_LUT_10BIT_BYPASS + amdgpu_crtc->crtc_offset);
-	if (bypass_lut)
-		tmp = REG_SET_FIELD(tmp, GRPH_LUT_10BIT_BYPASS, GRPH_LUT_10BIT_BYPASS_EN, 1);
-	else
-		tmp = REG_SET_FIELD(tmp, GRPH_LUT_10BIT_BYPASS, GRPH_LUT_10BIT_BYPASS_EN, 0);
-	WREG32(mmGRPH_LUT_10BIT_BYPASS + amdgpu_crtc->crtc_offset, tmp);
-
-	if (bypass_lut)
-		DRM_DEBUG_KMS("Bypassing hardware LUT due to 10 bit fb scanout.\n");
-
-	WREG32(mmGRPH_SURFACE_OFFSET_X + amdgpu_crtc->crtc_offset, 0);
-	WREG32(mmGRPH_SURFACE_OFFSET_Y + amdgpu_crtc->crtc_offset, 0);
-	WREG32(mmGRPH_X_START + amdgpu_crtc->crtc_offset, 0);
-	WREG32(mmGRPH_Y_START + amdgpu_crtc->crtc_offset, 0);
-	WREG32(mmGRPH_X_END + amdgpu_crtc->crtc_offset, target_fb->width);
-	WREG32(mmGRPH_Y_END + amdgpu_crtc->crtc_offset, target_fb->height);
-
-	fb_pitch_pixels = target_fb->pitches[0] / target_fb->format->cpp[0];
-	WREG32(mmGRPH_PITCH + amdgpu_crtc->crtc_offset, fb_pitch_pixels);
-
-	dce_v11_0_grph_enable(crtc, true);
-
-	WREG32(mmLB_DESKTOP_HEIGHT + amdgpu_crtc->crtc_offset,
-	       target_fb->height);
-
-	x &= ~3;
-	y &= ~1;
-	WREG32(mmVIEWPORT_START + amdgpu_crtc->crtc_offset,
-	       (x << 16) | y);
-	viewport_w = crtc->mode.hdisplay;
-	viewport_h = (crtc->mode.vdisplay + 1) & ~1;
-	WREG32(mmVIEWPORT_SIZE + amdgpu_crtc->crtc_offset,
-	       (viewport_w << 16) | viewport_h);
-
-	/* set pageflip to happen anywhere in vblank interval */
-	WREG32(mmCRTC_MASTER_UPDATE_MODE + amdgpu_crtc->crtc_offset, 0);
-
-	if (!atomic && fb && fb != crtc->primary->fb) {
-		abo = gem_to_amdgpu_bo(fb->obj[0]);
-		r = amdgpu_bo_reserve(abo, true);
-		if (unlikely(r != 0))
-			return r;
-		amdgpu_bo_unpin(abo);
-		amdgpu_bo_unreserve(abo);
-	}
-
-	/* Bytes per pixel may have changed */
-	dce_v11_0_bandwidth_update(adev);
-
-	return 0;
-}
-
-static void dce_v11_0_set_interleave(struct drm_crtc *crtc,
-				     struct drm_display_mode *mode)
-{
-	struct drm_device *dev = crtc->dev;
-	struct amdgpu_device *adev = drm_to_adev(dev);
-	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
-	u32 tmp;
-
-	tmp = RREG32(mmLB_DATA_FORMAT + amdgpu_crtc->crtc_offset);
-	if (mode->flags & DRM_MODE_FLAG_INTERLACE)
-		tmp = REG_SET_FIELD(tmp, LB_DATA_FORMAT, INTERLEAVE_EN, 1);
-	else
-		tmp = REG_SET_FIELD(tmp, LB_DATA_FORMAT, INTERLEAVE_EN, 0);
-	WREG32(mmLB_DATA_FORMAT + amdgpu_crtc->crtc_offset, tmp);
-}
-
-static void dce_v11_0_crtc_load_lut(struct drm_crtc *crtc)
-{
-	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
-	struct drm_device *dev = crtc->dev;
-	struct amdgpu_device *adev = drm_to_adev(dev);
-	u16 *r, *g, *b;
-	int i;
-	u32 tmp;
-
-	DRM_DEBUG_KMS("%d\n", amdgpu_crtc->crtc_id);
-
-	tmp = RREG32(mmINPUT_CSC_CONTROL + amdgpu_crtc->crtc_offset);
-	tmp = REG_SET_FIELD(tmp, INPUT_CSC_CONTROL, INPUT_CSC_GRPH_MODE, 0);
-	WREG32(mmINPUT_CSC_CONTROL + amdgpu_crtc->crtc_offset, tmp);
-
-	tmp = RREG32(mmPRESCALE_GRPH_CONTROL + amdgpu_crtc->crtc_offset);
-	tmp = REG_SET_FIELD(tmp, PRESCALE_GRPH_CONTROL, GRPH_PRESCALE_BYPASS, 1);
-	WREG32(mmPRESCALE_GRPH_CONTROL + amdgpu_crtc->crtc_offset, tmp);
-
-	tmp = RREG32(mmINPUT_GAMMA_CONTROL + amdgpu_crtc->crtc_offset);
-	tmp = REG_SET_FIELD(tmp, INPUT_GAMMA_CONTROL, GRPH_INPUT_GAMMA_MODE, 0);
-	WREG32(mmINPUT_GAMMA_CONTROL + amdgpu_crtc->crtc_offset, tmp);
-
-	WREG32(mmDC_LUT_CONTROL + amdgpu_crtc->crtc_offset, 0);
-
-	WREG32(mmDC_LUT_BLACK_OFFSET_BLUE + amdgpu_crtc->crtc_offset, 0);
-	WREG32(mmDC_LUT_BLACK_OFFSET_GREEN + amdgpu_crtc->crtc_offset, 0);
-	WREG32(mmDC_LUT_BLACK_OFFSET_RED + amdgpu_crtc->crtc_offset, 0);
-
-	WREG32(mmDC_LUT_WHITE_OFFSET_BLUE + amdgpu_crtc->crtc_offset, 0xffff);
-	WREG32(mmDC_LUT_WHITE_OFFSET_GREEN + amdgpu_crtc->crtc_offset, 0xffff);
-	WREG32(mmDC_LUT_WHITE_OFFSET_RED + amdgpu_crtc->crtc_offset, 0xffff);
-
-	WREG32(mmDC_LUT_RW_MODE + amdgpu_crtc->crtc_offset, 0);
-	WREG32(mmDC_LUT_WRITE_EN_MASK + amdgpu_crtc->crtc_offset, 0x00000007);
-
-	WREG32(mmDC_LUT_RW_INDEX + amdgpu_crtc->crtc_offset, 0);
-	r = crtc->gamma_store;
-	g = r + crtc->gamma_size;
-	b = g + crtc->gamma_size;
-	for (i = 0; i < 256; i++) {
-		WREG32(mmDC_LUT_30_COLOR + amdgpu_crtc->crtc_offset,
-		       ((*r++ & 0xffc0) << 14) |
-		       ((*g++ & 0xffc0) << 4) |
-		       (*b++ >> 6));
-	}
-
-	tmp = RREG32(mmDEGAMMA_CONTROL + amdgpu_crtc->crtc_offset);
-	tmp = REG_SET_FIELD(tmp, DEGAMMA_CONTROL, GRPH_DEGAMMA_MODE, 0);
-	tmp = REG_SET_FIELD(tmp, DEGAMMA_CONTROL, CURSOR_DEGAMMA_MODE, 0);
-	tmp = REG_SET_FIELD(tmp, DEGAMMA_CONTROL, CURSOR2_DEGAMMA_MODE, 0);
-	WREG32(mmDEGAMMA_CONTROL + amdgpu_crtc->crtc_offset, tmp);
-
-	tmp = RREG32(mmGAMUT_REMAP_CONTROL + amdgpu_crtc->crtc_offset);
-	tmp = REG_SET_FIELD(tmp, GAMUT_REMAP_CONTROL, GRPH_GAMUT_REMAP_MODE, 0);
-	WREG32(mmGAMUT_REMAP_CONTROL + amdgpu_crtc->crtc_offset, tmp);
-
-	tmp = RREG32(mmREGAMMA_CONTROL + amdgpu_crtc->crtc_offset);
-	tmp = REG_SET_FIELD(tmp, REGAMMA_CONTROL, GRPH_REGAMMA_MODE, 0);
-	WREG32(mmREGAMMA_CONTROL + amdgpu_crtc->crtc_offset, tmp);
-
-	tmp = RREG32(mmOUTPUT_CSC_CONTROL + amdgpu_crtc->crtc_offset);
-	tmp = REG_SET_FIELD(tmp, OUTPUT_CSC_CONTROL, OUTPUT_CSC_GRPH_MODE, 0);
-	WREG32(mmOUTPUT_CSC_CONTROL + amdgpu_crtc->crtc_offset, tmp);
-
-	/* XXX match this to the depth of the crtc fmt block, move to modeset? */
-	WREG32(mmDENORM_CONTROL + amdgpu_crtc->crtc_offset, 0);
-	/* XXX this only needs to be programmed once per crtc at startup,
-	 * not sure where the best place for it is
-	 */
-	tmp = RREG32(mmALPHA_CONTROL + amdgpu_crtc->crtc_offset);
-	tmp = REG_SET_FIELD(tmp, ALPHA_CONTROL, CURSOR_ALPHA_BLND_ENA, 1);
-	WREG32(mmALPHA_CONTROL + amdgpu_crtc->crtc_offset, tmp);
-}
-
-static int dce_v11_0_pick_dig_encoder(struct drm_encoder *encoder)
-{
-	struct amdgpu_encoder *amdgpu_encoder = to_amdgpu_encoder(encoder);
-	struct amdgpu_encoder_atom_dig *dig = amdgpu_encoder->enc_priv;
-
-	switch (amdgpu_encoder->encoder_id) {
-	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
-		if (dig->linkb)
-			return 1;
-		else
-			return 0;
-	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
-		if (dig->linkb)
-			return 3;
-		else
-			return 2;
-	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
-		if (dig->linkb)
-			return 5;
-		else
-			return 4;
-	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY3:
-		return 6;
-	default:
-		DRM_ERROR("invalid encoder_id: 0x%x\n", amdgpu_encoder->encoder_id);
-		return 0;
-	}
-}
-
-/**
- * dce_v11_0_pick_pll - Allocate a PPLL for use by the crtc.
- *
- * @crtc: drm crtc
- *
- * Returns the PPLL (Pixel PLL) to be used by the crtc.  For DP monitors
- * a single PPLL can be used for all DP crtcs/encoders.  For non-DP
- * monitors a dedicated PPLL must be used.  If a particular board has
- * an external DP PLL, return ATOM_PPLL_INVALID to skip PLL programming
- * as there is no need to program the PLL itself.  If we are not able to
- * allocate a PLL, return ATOM_PPLL_INVALID to skip PLL programming to
- * avoid messing up an existing monitor.
- *
- * Asic specific PLL information
- *
- * DCE 10.x
- * Tonga
- * - PPLL1, PPLL2 are available for all UNIPHY (both DP and non-DP)
- * CI
- * - PPLL0, PPLL1, PPLL2 are available for all UNIPHY (both DP and non-DP) and DAC
- *
- */
-static u32 dce_v11_0_pick_pll(struct drm_crtc *crtc)
-{
-	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
-	struct drm_device *dev = crtc->dev;
-	struct amdgpu_device *adev = drm_to_adev(dev);
-	u32 pll_in_use;
-	int pll;
-
-	if ((adev->asic_type == CHIP_POLARIS10) ||
-	    (adev->asic_type == CHIP_POLARIS11) ||
-	    (adev->asic_type == CHIP_POLARIS12) ||
-	    (adev->asic_type == CHIP_VEGAM)) {
-		struct amdgpu_encoder *amdgpu_encoder =
-			to_amdgpu_encoder(amdgpu_crtc->encoder);
-		struct amdgpu_encoder_atom_dig *dig = amdgpu_encoder->enc_priv;
-
-		if (ENCODER_MODE_IS_DP(amdgpu_atombios_encoder_get_encoder_mode(amdgpu_crtc->encoder)))
-			return ATOM_DP_DTO;
-
-		switch (amdgpu_encoder->encoder_id) {
-		case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
-			if (dig->linkb)
-				return ATOM_COMBOPHY_PLL1;
-			else
-				return ATOM_COMBOPHY_PLL0;
-		case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
-			if (dig->linkb)
-				return ATOM_COMBOPHY_PLL3;
-			else
-				return ATOM_COMBOPHY_PLL2;
-		case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
-			if (dig->linkb)
-				return ATOM_COMBOPHY_PLL5;
-			else
-				return ATOM_COMBOPHY_PLL4;
-		default:
-			DRM_ERROR("invalid encoder_id: 0x%x\n", amdgpu_encoder->encoder_id);
-			return ATOM_PPLL_INVALID;
-		}
-	}
-
-	if (ENCODER_MODE_IS_DP(amdgpu_atombios_encoder_get_encoder_mode(amdgpu_crtc->encoder))) {
-		if (adev->clock.dp_extclk)
-			/* skip PPLL programming if using ext clock */
-			return ATOM_PPLL_INVALID;
-		else {
-			/* use the same PPLL for all DP monitors */
-			pll = amdgpu_pll_get_shared_dp_ppll(crtc);
-			if (pll != ATOM_PPLL_INVALID)
-				return pll;
-		}
-	} else {
-		/* use the same PPLL for all monitors with the same clock */
-		pll = amdgpu_pll_get_shared_nondp_ppll(crtc);
-		if (pll != ATOM_PPLL_INVALID)
-			return pll;
-	}
-
-	/* XXX need to determine what plls are available on each DCE11 part */
-	pll_in_use = amdgpu_pll_get_use_mask(crtc);
-	if (adev->flags & AMD_IS_APU) {
-		if (!(pll_in_use & (1 << ATOM_PPLL1)))
-			return ATOM_PPLL1;
-		if (!(pll_in_use & (1 << ATOM_PPLL0)))
-			return ATOM_PPLL0;
-		DRM_ERROR("unable to allocate a PPLL\n");
-		return ATOM_PPLL_INVALID;
-	} else {
-		if (!(pll_in_use & (1 << ATOM_PPLL2)))
-			return ATOM_PPLL2;
-		if (!(pll_in_use & (1 << ATOM_PPLL1)))
-			return ATOM_PPLL1;
-		if (!(pll_in_use & (1 << ATOM_PPLL0)))
-			return ATOM_PPLL0;
-		DRM_ERROR("unable to allocate a PPLL\n");
-		return ATOM_PPLL_INVALID;
-	}
-	return ATOM_PPLL_INVALID;
-}
-
-static void dce_v11_0_lock_cursor(struct drm_crtc *crtc, bool lock)
-{
-	struct amdgpu_device *adev = drm_to_adev(crtc->dev);
-	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
-	uint32_t cur_lock;
-
-	cur_lock = RREG32(mmCUR_UPDATE + amdgpu_crtc->crtc_offset);
-	if (lock)
-		cur_lock = REG_SET_FIELD(cur_lock, CUR_UPDATE, CURSOR_UPDATE_LOCK, 1);
-	else
-		cur_lock = REG_SET_FIELD(cur_lock, CUR_UPDATE, CURSOR_UPDATE_LOCK, 0);
-	WREG32(mmCUR_UPDATE + amdgpu_crtc->crtc_offset, cur_lock);
-}
-
-static void dce_v11_0_hide_cursor(struct drm_crtc *crtc)
-{
-	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
-	struct amdgpu_device *adev = drm_to_adev(crtc->dev);
-	u32 tmp;
-
-	tmp = RREG32(mmCUR_CONTROL + amdgpu_crtc->crtc_offset);
-	tmp = REG_SET_FIELD(tmp, CUR_CONTROL, CURSOR_EN, 0);
-	WREG32(mmCUR_CONTROL + amdgpu_crtc->crtc_offset, tmp);
-}
-
-static void dce_v11_0_show_cursor(struct drm_crtc *crtc)
-{
-	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
-	struct amdgpu_device *adev = drm_to_adev(crtc->dev);
-	u32 tmp;
-
-	WREG32(mmCUR_SURFACE_ADDRESS_HIGH + amdgpu_crtc->crtc_offset,
-	       upper_32_bits(amdgpu_crtc->cursor_addr));
-	WREG32(mmCUR_SURFACE_ADDRESS + amdgpu_crtc->crtc_offset,
-	       lower_32_bits(amdgpu_crtc->cursor_addr));
-
-	tmp = RREG32(mmCUR_CONTROL + amdgpu_crtc->crtc_offset);
-	tmp = REG_SET_FIELD(tmp, CUR_CONTROL, CURSOR_EN, 1);
-	tmp = REG_SET_FIELD(tmp, CUR_CONTROL, CURSOR_MODE, 2);
-	WREG32(mmCUR_CONTROL + amdgpu_crtc->crtc_offset, tmp);
-}
-
-static int dce_v11_0_cursor_move_locked(struct drm_crtc *crtc,
-					int x, int y)
-{
-	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
-	struct amdgpu_device *adev = drm_to_adev(crtc->dev);
-	int xorigin = 0, yorigin = 0;
-
-	amdgpu_crtc->cursor_x = x;
-	amdgpu_crtc->cursor_y = y;
-
-	/* avivo cursor are offset into the total surface */
-	x += crtc->x;
-	y += crtc->y;
-	DRM_DEBUG("x %d y %d c->x %d c->y %d\n", x, y, crtc->x, crtc->y);
-
-	if (x < 0) {
-		xorigin = min(-x, amdgpu_crtc->max_cursor_width - 1);
-		x = 0;
-	}
-	if (y < 0) {
-		yorigin = min(-y, amdgpu_crtc->max_cursor_height - 1);
-		y = 0;
-	}
-
-	WREG32(mmCUR_POSITION + amdgpu_crtc->crtc_offset, (x << 16) | y);
-	WREG32(mmCUR_HOT_SPOT + amdgpu_crtc->crtc_offset, (xorigin << 16) | yorigin);
-	WREG32(mmCUR_SIZE + amdgpu_crtc->crtc_offset,
-	       ((amdgpu_crtc->cursor_width - 1) << 16) | (amdgpu_crtc->cursor_height - 1));
-
-	return 0;
-}
-
-static int dce_v11_0_crtc_cursor_move(struct drm_crtc *crtc,
-				      int x, int y)
-{
-	int ret;
-
-	dce_v11_0_lock_cursor(crtc, true);
-	ret = dce_v11_0_cursor_move_locked(crtc, x, y);
-	dce_v11_0_lock_cursor(crtc, false);
-
-	return ret;
-}
-
-static int dce_v11_0_crtc_cursor_set2(struct drm_crtc *crtc,
-				      struct drm_file *file_priv,
-				      uint32_t handle,
-				      uint32_t width,
-				      uint32_t height,
-				      int32_t hot_x,
-				      int32_t hot_y)
-{
-	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
-	struct drm_gem_object *obj;
-	struct amdgpu_bo *aobj;
-	int ret;
-
-	if (!handle) {
-		/* turn off cursor */
-		dce_v11_0_hide_cursor(crtc);
-		obj = NULL;
-		goto unpin;
-	}
-
-	if ((width > amdgpu_crtc->max_cursor_width) ||
-	    (height > amdgpu_crtc->max_cursor_height)) {
-		DRM_ERROR("bad cursor width or height %d x %d\n", width, height);
-		return -EINVAL;
-	}
-
-	obj = drm_gem_object_lookup(file_priv, handle);
-	if (!obj) {
-		DRM_ERROR("Cannot find cursor object %x for crtc %d\n", handle, amdgpu_crtc->crtc_id);
-		return -ENOENT;
-	}
-
-	aobj = gem_to_amdgpu_bo(obj);
-	ret = amdgpu_bo_reserve(aobj, false);
-	if (ret != 0) {
-		drm_gem_object_put(obj);
-		return ret;
-	}
-
-	ret = amdgpu_bo_pin(aobj, AMDGPU_GEM_DOMAIN_VRAM);
-	amdgpu_bo_unreserve(aobj);
-	if (ret) {
-		DRM_ERROR("Failed to pin new cursor BO (%d)\n", ret);
-		drm_gem_object_put(obj);
-		return ret;
-	}
-	amdgpu_crtc->cursor_addr = amdgpu_bo_gpu_offset(aobj);
-
-	dce_v11_0_lock_cursor(crtc, true);
-
-	if (width != amdgpu_crtc->cursor_width ||
-	    height != amdgpu_crtc->cursor_height ||
-	    hot_x != amdgpu_crtc->cursor_hot_x ||
-	    hot_y != amdgpu_crtc->cursor_hot_y) {
-		int x, y;
-
-		x = amdgpu_crtc->cursor_x + amdgpu_crtc->cursor_hot_x - hot_x;
-		y = amdgpu_crtc->cursor_y + amdgpu_crtc->cursor_hot_y - hot_y;
-
-		dce_v11_0_cursor_move_locked(crtc, x, y);
-
-		amdgpu_crtc->cursor_width = width;
-		amdgpu_crtc->cursor_height = height;
-		amdgpu_crtc->cursor_hot_x = hot_x;
-		amdgpu_crtc->cursor_hot_y = hot_y;
-	}
-
-	dce_v11_0_show_cursor(crtc);
-	dce_v11_0_lock_cursor(crtc, false);
-
-unpin:
-	if (amdgpu_crtc->cursor_bo) {
-		struct amdgpu_bo *aobj = gem_to_amdgpu_bo(amdgpu_crtc->cursor_bo);
-		ret = amdgpu_bo_reserve(aobj, true);
-		if (likely(ret == 0)) {
-			amdgpu_bo_unpin(aobj);
-			amdgpu_bo_unreserve(aobj);
-		}
-		drm_gem_object_put(amdgpu_crtc->cursor_bo);
-	}
-
-	amdgpu_crtc->cursor_bo = obj;
-	return 0;
-}
-
-static void dce_v11_0_cursor_reset(struct drm_crtc *crtc)
-{
-	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
-
-	if (amdgpu_crtc->cursor_bo) {
-		dce_v11_0_lock_cursor(crtc, true);
-
-		dce_v11_0_cursor_move_locked(crtc, amdgpu_crtc->cursor_x,
-					     amdgpu_crtc->cursor_y);
-
-		dce_v11_0_show_cursor(crtc);
-
-		dce_v11_0_lock_cursor(crtc, false);
-	}
-}
-
-static int dce_v11_0_crtc_gamma_set(struct drm_crtc *crtc, u16 *red, u16 *green,
-				    u16 *blue, uint32_t size,
-				    struct drm_modeset_acquire_ctx *ctx)
-{
-	dce_v11_0_crtc_load_lut(crtc);
-
-	return 0;
-}
-
-static void dce_v11_0_crtc_destroy(struct drm_crtc *crtc)
-{
-	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
-
-	drm_crtc_cleanup(crtc);
-	kfree(amdgpu_crtc);
-}
-
-static const struct drm_crtc_funcs dce_v11_0_crtc_funcs = {
-	.cursor_set2 = dce_v11_0_crtc_cursor_set2,
-	.cursor_move = dce_v11_0_crtc_cursor_move,
-	.gamma_set = dce_v11_0_crtc_gamma_set,
-	.set_config = amdgpu_display_crtc_set_config,
-	.destroy = dce_v11_0_crtc_destroy,
-	.page_flip_target = amdgpu_display_crtc_page_flip_target,
-	.get_vblank_counter = amdgpu_get_vblank_counter_kms,
-	.enable_vblank = amdgpu_enable_vblank_kms,
-	.disable_vblank = amdgpu_disable_vblank_kms,
-	.get_vblank_timestamp = drm_crtc_vblank_helper_get_vblank_timestamp,
-};
-
-static void dce_v11_0_crtc_dpms(struct drm_crtc *crtc, int mode)
-{
-	struct drm_device *dev = crtc->dev;
-	struct amdgpu_device *adev = drm_to_adev(dev);
-	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
-	unsigned type;
-
-	switch (mode) {
-	case DRM_MODE_DPMS_ON:
-		amdgpu_crtc->enabled = true;
-		amdgpu_atombios_crtc_enable(crtc, ATOM_ENABLE);
-		dce_v11_0_vga_enable(crtc, true);
-		amdgpu_atombios_crtc_blank(crtc, ATOM_DISABLE);
-		dce_v11_0_vga_enable(crtc, false);
-		/* Make sure VBLANK and PFLIP interrupts are still enabled */
-		type = amdgpu_display_crtc_idx_to_irq_type(adev,
-						amdgpu_crtc->crtc_id);
-		amdgpu_irq_update(adev, &adev->crtc_irq, type);
-		amdgpu_irq_update(adev, &adev->pageflip_irq, type);
-		drm_crtc_vblank_on(crtc);
-		dce_v11_0_crtc_load_lut(crtc);
-		break;
-	case DRM_MODE_DPMS_STANDBY:
-	case DRM_MODE_DPMS_SUSPEND:
-	case DRM_MODE_DPMS_OFF:
-		drm_crtc_vblank_off(crtc);
-		if (amdgpu_crtc->enabled) {
-			dce_v11_0_vga_enable(crtc, true);
-			amdgpu_atombios_crtc_blank(crtc, ATOM_ENABLE);
-			dce_v11_0_vga_enable(crtc, false);
-		}
-		amdgpu_atombios_crtc_enable(crtc, ATOM_DISABLE);
-		amdgpu_crtc->enabled = false;
-		break;
-	}
-	/* adjust pm to dpms */
-	amdgpu_pm_compute_clocks(adev);
-}
-
-static void dce_v11_0_crtc_prepare(struct drm_crtc *crtc)
-{
-	/* disable crtc pair power gating before programming */
-	amdgpu_atombios_crtc_powergate(crtc, ATOM_DISABLE);
-	amdgpu_atombios_crtc_lock(crtc, ATOM_ENABLE);
-	dce_v11_0_crtc_dpms(crtc, DRM_MODE_DPMS_OFF);
-}
-
-static void dce_v11_0_crtc_commit(struct drm_crtc *crtc)
-{
-	dce_v11_0_crtc_dpms(crtc, DRM_MODE_DPMS_ON);
-	amdgpu_atombios_crtc_lock(crtc, ATOM_DISABLE);
-}
-
-static void dce_v11_0_crtc_disable(struct drm_crtc *crtc)
-{
-	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
-	struct drm_device *dev = crtc->dev;
-	struct amdgpu_device *adev = drm_to_adev(dev);
-	struct amdgpu_atom_ss ss;
-	int i;
-
-	dce_v11_0_crtc_dpms(crtc, DRM_MODE_DPMS_OFF);
-	if (crtc->primary->fb) {
-		int r;
-		struct amdgpu_bo *abo;
-
-		abo = gem_to_amdgpu_bo(crtc->primary->fb->obj[0]);
-		r = amdgpu_bo_reserve(abo, true);
-		if (unlikely(r))
-			DRM_ERROR("failed to reserve abo before unpin\n");
-		else {
-			amdgpu_bo_unpin(abo);
-			amdgpu_bo_unreserve(abo);
-		}
-	}
-	/* disable the GRPH */
-	dce_v11_0_grph_enable(crtc, false);
-
-	amdgpu_atombios_crtc_powergate(crtc, ATOM_ENABLE);
-
-	for (i = 0; i < adev->mode_info.num_crtc; i++) {
-		if (adev->mode_info.crtcs[i] &&
-		    adev->mode_info.crtcs[i]->enabled &&
-		    i != amdgpu_crtc->crtc_id &&
-		    amdgpu_crtc->pll_id == adev->mode_info.crtcs[i]->pll_id) {
-			/* one other crtc is using this pll don't turn
-			 * off the pll
-			 */
-			goto done;
-		}
-	}
-
-	switch (amdgpu_crtc->pll_id) {
-	case ATOM_PPLL0:
-	case ATOM_PPLL1:
-	case ATOM_PPLL2:
-		/* disable the ppll */
-		amdgpu_atombios_crtc_program_pll(crtc, amdgpu_crtc->crtc_id, amdgpu_crtc->pll_id,
-						 0, 0, ATOM_DISABLE, 0, 0, 0, 0, 0, false, &ss);
-		break;
-	case ATOM_COMBOPHY_PLL0:
-	case ATOM_COMBOPHY_PLL1:
-	case ATOM_COMBOPHY_PLL2:
-	case ATOM_COMBOPHY_PLL3:
-	case ATOM_COMBOPHY_PLL4:
-	case ATOM_COMBOPHY_PLL5:
-		/* disable the ppll */
-		amdgpu_atombios_crtc_program_pll(crtc, ATOM_CRTC_INVALID, amdgpu_crtc->pll_id,
-						 0, 0, ATOM_DISABLE, 0, 0, 0, 0, 0, false, &ss);
-		break;
-	default:
-		break;
-	}
-done:
-	amdgpu_crtc->pll_id = ATOM_PPLL_INVALID;
-	amdgpu_crtc->adjusted_clock = 0;
-	amdgpu_crtc->encoder = NULL;
-	amdgpu_crtc->connector = NULL;
-}
-
-static int dce_v11_0_crtc_mode_set(struct drm_crtc *crtc,
-				  struct drm_display_mode *mode,
-				  struct drm_display_mode *adjusted_mode,
-				  int x, int y, struct drm_framebuffer *old_fb)
-{
-	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
-	struct drm_device *dev = crtc->dev;
-	struct amdgpu_device *adev = drm_to_adev(dev);
-
-	if (!amdgpu_crtc->adjusted_clock)
-		return -EINVAL;
-
-	if ((adev->asic_type == CHIP_POLARIS10) ||
-	    (adev->asic_type == CHIP_POLARIS11) ||
-	    (adev->asic_type == CHIP_POLARIS12) ||
-	    (adev->asic_type == CHIP_VEGAM)) {
-		struct amdgpu_encoder *amdgpu_encoder =
-			to_amdgpu_encoder(amdgpu_crtc->encoder);
-		int encoder_mode =
-			amdgpu_atombios_encoder_get_encoder_mode(amdgpu_crtc->encoder);
-
-		/* SetPixelClock calculates the plls and ss values now */
-		amdgpu_atombios_crtc_program_pll(crtc, amdgpu_crtc->crtc_id,
-						 amdgpu_crtc->pll_id,
-						 encoder_mode, amdgpu_encoder->encoder_id,
-						 adjusted_mode->clock, 0, 0, 0, 0,
-						 amdgpu_crtc->bpc, amdgpu_crtc->ss_enabled, &amdgpu_crtc->ss);
-	} else {
-		amdgpu_atombios_crtc_set_pll(crtc, adjusted_mode);
-	}
-	amdgpu_atombios_crtc_set_dtd_timing(crtc, adjusted_mode);
-	dce_v11_0_crtc_do_set_base(crtc, old_fb, x, y, 0);
-	amdgpu_atombios_crtc_overscan_setup(crtc, mode, adjusted_mode);
-	amdgpu_atombios_crtc_scaler_setup(crtc);
-	dce_v11_0_cursor_reset(crtc);
-	/* update the hw version fpr dpm */
-	amdgpu_crtc->hw_mode = *adjusted_mode;
-
-	return 0;
-}
-
-static bool dce_v11_0_crtc_mode_fixup(struct drm_crtc *crtc,
-				     const struct drm_display_mode *mode,
-				     struct drm_display_mode *adjusted_mode)
-{
-	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
-	struct drm_device *dev = crtc->dev;
-	struct drm_encoder *encoder;
-
-	/* assign the encoder to the amdgpu crtc to avoid repeated lookups later */
-	list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) {
-		if (encoder->crtc == crtc) {
-			amdgpu_crtc->encoder = encoder;
-			amdgpu_crtc->connector = amdgpu_get_connector_for_encoder(encoder);
-			break;
-		}
-	}
-	if ((amdgpu_crtc->encoder == NULL) || (amdgpu_crtc->connector == NULL)) {
-		amdgpu_crtc->encoder = NULL;
-		amdgpu_crtc->connector = NULL;
-		return false;
-	}
-	if (!amdgpu_display_crtc_scaling_mode_fixup(crtc, mode, adjusted_mode))
-		return false;
-	if (amdgpu_atombios_crtc_prepare_pll(crtc, adjusted_mode))
-		return false;
-	/* pick pll */
-	amdgpu_crtc->pll_id = dce_v11_0_pick_pll(crtc);
-	/* if we can't get a PPLL for a non-DP encoder, fail */
-	if ((amdgpu_crtc->pll_id == ATOM_PPLL_INVALID) &&
-	    !ENCODER_MODE_IS_DP(amdgpu_atombios_encoder_get_encoder_mode(amdgpu_crtc->encoder)))
-		return false;
-
-	return true;
-}
-
-static int dce_v11_0_crtc_set_base(struct drm_crtc *crtc, int x, int y,
-				  struct drm_framebuffer *old_fb)
-{
-	return dce_v11_0_crtc_do_set_base(crtc, old_fb, x, y, 0);
-}
-
-static int dce_v11_0_crtc_set_base_atomic(struct drm_crtc *crtc,
-					 struct drm_framebuffer *fb,
-					 int x, int y, enum mode_set_atomic state)
-{
-	return dce_v11_0_crtc_do_set_base(crtc, fb, x, y, 1);
-}
-
-static const struct drm_crtc_helper_funcs dce_v11_0_crtc_helper_funcs = {
-	.dpms = dce_v11_0_crtc_dpms,
-	.mode_fixup = dce_v11_0_crtc_mode_fixup,
-	.mode_set = dce_v11_0_crtc_mode_set,
-	.mode_set_base = dce_v11_0_crtc_set_base,
-	.mode_set_base_atomic = dce_v11_0_crtc_set_base_atomic,
-	.prepare = dce_v11_0_crtc_prepare,
-	.commit = dce_v11_0_crtc_commit,
-	.disable = dce_v11_0_crtc_disable,
-	.get_scanout_position = amdgpu_crtc_get_scanout_position,
-};
-
-static int dce_v11_0_crtc_init(struct amdgpu_device *adev, int index)
-{
-	struct amdgpu_crtc *amdgpu_crtc;
-
-	amdgpu_crtc = kzalloc(sizeof(struct amdgpu_crtc) +
-			      (AMDGPUFB_CONN_LIMIT * sizeof(struct drm_connector *)), GFP_KERNEL);
-	if (amdgpu_crtc == NULL)
-		return -ENOMEM;
-
-	drm_crtc_init(adev_to_drm(adev), &amdgpu_crtc->base, &dce_v11_0_crtc_funcs);
-
-	drm_mode_crtc_set_gamma_size(&amdgpu_crtc->base, 256);
-	amdgpu_crtc->crtc_id = index;
-	adev->mode_info.crtcs[index] = amdgpu_crtc;
-
-	amdgpu_crtc->max_cursor_width = 128;
-	amdgpu_crtc->max_cursor_height = 128;
-	adev_to_drm(adev)->mode_config.cursor_width = amdgpu_crtc->max_cursor_width;
-	adev_to_drm(adev)->mode_config.cursor_height = amdgpu_crtc->max_cursor_height;
-
-	switch (amdgpu_crtc->crtc_id) {
-	case 0:
-	default:
-		amdgpu_crtc->crtc_offset = CRTC0_REGISTER_OFFSET;
-		break;
-	case 1:
-		amdgpu_crtc->crtc_offset = CRTC1_REGISTER_OFFSET;
-		break;
-	case 2:
-		amdgpu_crtc->crtc_offset = CRTC2_REGISTER_OFFSET;
-		break;
-	case 3:
-		amdgpu_crtc->crtc_offset = CRTC3_REGISTER_OFFSET;
-		break;
-	case 4:
-		amdgpu_crtc->crtc_offset = CRTC4_REGISTER_OFFSET;
-		break;
-	case 5:
-		amdgpu_crtc->crtc_offset = CRTC5_REGISTER_OFFSET;
-		break;
-	}
-
-	amdgpu_crtc->pll_id = ATOM_PPLL_INVALID;
-	amdgpu_crtc->adjusted_clock = 0;
-	amdgpu_crtc->encoder = NULL;
-	amdgpu_crtc->connector = NULL;
-	drm_crtc_helper_add(&amdgpu_crtc->base, &dce_v11_0_crtc_helper_funcs);
-
-	return 0;
-}
-
-static int dce_v11_0_early_init(void *handle)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	adev->audio_endpt_rreg = &dce_v11_0_audio_endpt_rreg;
-	adev->audio_endpt_wreg = &dce_v11_0_audio_endpt_wreg;
-
-	dce_v11_0_set_display_funcs(adev);
-
-	adev->mode_info.num_crtc = dce_v11_0_get_num_crtc(adev);
-
-	switch (adev->asic_type) {
-	case CHIP_CARRIZO:
-		adev->mode_info.num_hpd = 6;
-		adev->mode_info.num_dig = 9;
-		break;
-	case CHIP_STONEY:
-		adev->mode_info.num_hpd = 6;
-		adev->mode_info.num_dig = 9;
-		break;
-	case CHIP_POLARIS10:
-	case CHIP_VEGAM:
-		adev->mode_info.num_hpd = 6;
-		adev->mode_info.num_dig = 6;
-		break;
-	case CHIP_POLARIS11:
-	case CHIP_POLARIS12:
-		adev->mode_info.num_hpd = 5;
-		adev->mode_info.num_dig = 5;
-		break;
-	default:
-		/* FIXME: not supported yet */
-		return -EINVAL;
-	}
-
-	dce_v11_0_set_irq_funcs(adev);
-
-	return 0;
-}
-
-static int dce_v11_0_sw_init(void *handle)
-{
-	int r, i;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	for (i = 0; i < adev->mode_info.num_crtc; i++) {
-		r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, i + 1, &adev->crtc_irq);
-		if (r)
-			return r;
-	}
-
-	for (i = VISLANDS30_IV_SRCID_D1_GRPH_PFLIP; i < 20; i += 2) {
-		r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, i, &adev->pageflip_irq);
-		if (r)
-			return r;
-	}
-
-	/* HPD hotplug */
-	r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_HOTPLUG_DETECT_A, &adev->hpd_irq);
-	if (r)
-		return r;
-
-	adev_to_drm(adev)->mode_config.funcs = &amdgpu_mode_funcs;
-
-	adev_to_drm(adev)->mode_config.async_page_flip = true;
-
-	adev_to_drm(adev)->mode_config.max_width = 16384;
-	adev_to_drm(adev)->mode_config.max_height = 16384;
-
-	adev_to_drm(adev)->mode_config.preferred_depth = 24;
-	adev_to_drm(adev)->mode_config.prefer_shadow = 1;
-
-	adev_to_drm(adev)->mode_config.fb_base = adev->gmc.aper_base;
-
-	r = amdgpu_display_modeset_create_props(adev);
-	if (r)
-		return r;
-
-	adev_to_drm(adev)->mode_config.max_width = 16384;
-	adev_to_drm(adev)->mode_config.max_height = 16384;
-
-
-	/* allocate crtcs */
-	for (i = 0; i < adev->mode_info.num_crtc; i++) {
-		r = dce_v11_0_crtc_init(adev, i);
-		if (r)
-			return r;
-	}
-
-	if (amdgpu_atombios_get_connector_info_from_object_table(adev))
-		amdgpu_display_print_display_setup(adev_to_drm(adev));
-	else
-		return -EINVAL;
-
-	/* setup afmt */
-	r = dce_v11_0_afmt_init(adev);
-	if (r)
-		return r;
-
-	r = dce_v11_0_audio_init(adev);
-	if (r)
-		return r;
-
-	drm_kms_helper_poll_init(adev_to_drm(adev));
-
-	adev->mode_info.mode_config_initialized = true;
-	return 0;
-}
-
-static int dce_v11_0_sw_fini(void *handle)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	kfree(adev->mode_info.bios_hardcoded_edid);
-
-	drm_kms_helper_poll_fini(adev_to_drm(adev));
-
-	dce_v11_0_audio_fini(adev);
-
-	dce_v11_0_afmt_fini(adev);
-
-	drm_mode_config_cleanup(adev_to_drm(adev));
-	adev->mode_info.mode_config_initialized = false;
-
-	return 0;
-}
-
-static int dce_v11_0_hw_init(void *handle)
-{
-	int i;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	dce_v11_0_init_golden_registers(adev);
-
-	/* disable vga render */
-	dce_v11_0_set_vga_render_state(adev, false);
-	/* init dig PHYs, disp eng pll */
-	amdgpu_atombios_crtc_powergate_init(adev);
-	amdgpu_atombios_encoder_init_dig(adev);
-	if ((adev->asic_type == CHIP_POLARIS10) ||
-	    (adev->asic_type == CHIP_POLARIS11) ||
-	    (adev->asic_type == CHIP_POLARIS12) ||
-	    (adev->asic_type == CHIP_VEGAM)) {
-		amdgpu_atombios_crtc_set_dce_clock(adev, adev->clock.default_dispclk,
-						   DCE_CLOCK_TYPE_DISPCLK, ATOM_GCK_DFS);
-		amdgpu_atombios_crtc_set_dce_clock(adev, 0,
-						   DCE_CLOCK_TYPE_DPREFCLK, ATOM_GCK_DFS);
-	} else {
-		amdgpu_atombios_crtc_set_disp_eng_pll(adev, adev->clock.default_dispclk);
-	}
-
-	/* initialize hpd */
-	dce_v11_0_hpd_init(adev);
-
-	for (i = 0; i < adev->mode_info.audio.num_pins; i++) {
-		dce_v11_0_audio_enable(adev, &adev->mode_info.audio.pin[i], false);
-	}
-
-	dce_v11_0_pageflip_interrupt_init(adev);
-
-	return 0;
-}
-
-static int dce_v11_0_hw_fini(void *handle)
-{
-	int i;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	dce_v11_0_hpd_fini(adev);
-
-	for (i = 0; i < adev->mode_info.audio.num_pins; i++) {
-		dce_v11_0_audio_enable(adev, &adev->mode_info.audio.pin[i], false);
-	}
-
-	dce_v11_0_pageflip_interrupt_fini(adev);
-
-	return 0;
-}
-
-static int dce_v11_0_suspend(void *handle)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-	int r;
-
-	r = amdgpu_display_suspend_helper(adev);
-	if (r)
-		return r;
-
-	adev->mode_info.bl_level =
-		amdgpu_atombios_encoder_get_backlight_level_from_reg(adev);
-
-	return dce_v11_0_hw_fini(handle);
-}
-
-static int dce_v11_0_resume(void *handle)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-	int ret;
-
-	amdgpu_atombios_encoder_set_backlight_level_to_reg(adev,
-							   adev->mode_info.bl_level);
-
-	ret = dce_v11_0_hw_init(handle);
-
-	/* turn on the BL */
-	if (adev->mode_info.bl_encoder) {
-		u8 bl_level = amdgpu_display_backlight_get_level(adev,
-								  adev->mode_info.bl_encoder);
-		amdgpu_display_backlight_set_level(adev, adev->mode_info.bl_encoder,
-						    bl_level);
-	}
-	if (ret)
-		return ret;
-
-	return amdgpu_display_resume_helper(adev);
-}
-
-static bool dce_v11_0_is_idle(void *handle)
-{
-	return true;
-}
-
-static int dce_v11_0_wait_for_idle(void *handle)
-{
-	return 0;
-}
-
-static int dce_v11_0_soft_reset(void *handle)
-{
-	u32 srbm_soft_reset = 0, tmp;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	if (dce_v11_0_is_display_hung(adev))
-		srbm_soft_reset |= SRBM_SOFT_RESET__SOFT_RESET_DC_MASK;
-
-	if (srbm_soft_reset) {
-		tmp = RREG32(mmSRBM_SOFT_RESET);
-		tmp |= srbm_soft_reset;
-		dev_info(adev->dev, "SRBM_SOFT_RESET=0x%08X\n", tmp);
-		WREG32(mmSRBM_SOFT_RESET, tmp);
-		tmp = RREG32(mmSRBM_SOFT_RESET);
-
-		udelay(50);
-
-		tmp &= ~srbm_soft_reset;
-		WREG32(mmSRBM_SOFT_RESET, tmp);
-		tmp = RREG32(mmSRBM_SOFT_RESET);
-
-		/* Wait a little for things to settle down */
-		udelay(50);
-	}
-	return 0;
-}
-
-static void dce_v11_0_set_crtc_vblank_interrupt_state(struct amdgpu_device *adev,
-						     int crtc,
-						     enum amdgpu_interrupt_state state)
-{
-	u32 lb_interrupt_mask;
-
-	if (crtc >= adev->mode_info.num_crtc) {
-		DRM_DEBUG("invalid crtc %d\n", crtc);
-		return;
-	}
-
-	switch (state) {
-	case AMDGPU_IRQ_STATE_DISABLE:
-		lb_interrupt_mask = RREG32(mmLB_INTERRUPT_MASK + crtc_offsets[crtc]);
-		lb_interrupt_mask = REG_SET_FIELD(lb_interrupt_mask, LB_INTERRUPT_MASK,
-						  VBLANK_INTERRUPT_MASK, 0);
-		WREG32(mmLB_INTERRUPT_MASK + crtc_offsets[crtc], lb_interrupt_mask);
-		break;
-	case AMDGPU_IRQ_STATE_ENABLE:
-		lb_interrupt_mask = RREG32(mmLB_INTERRUPT_MASK + crtc_offsets[crtc]);
-		lb_interrupt_mask = REG_SET_FIELD(lb_interrupt_mask, LB_INTERRUPT_MASK,
-						  VBLANK_INTERRUPT_MASK, 1);
-		WREG32(mmLB_INTERRUPT_MASK + crtc_offsets[crtc], lb_interrupt_mask);
-		break;
-	default:
-		break;
-	}
-}
-
-static void dce_v11_0_set_crtc_vline_interrupt_state(struct amdgpu_device *adev,
-						    int crtc,
-						    enum amdgpu_interrupt_state state)
-{
-	u32 lb_interrupt_mask;
-
-	if (crtc >= adev->mode_info.num_crtc) {
-		DRM_DEBUG("invalid crtc %d\n", crtc);
-		return;
-	}
-
-	switch (state) {
-	case AMDGPU_IRQ_STATE_DISABLE:
-		lb_interrupt_mask = RREG32(mmLB_INTERRUPT_MASK + crtc_offsets[crtc]);
-		lb_interrupt_mask = REG_SET_FIELD(lb_interrupt_mask, LB_INTERRUPT_MASK,
-						  VLINE_INTERRUPT_MASK, 0);
-		WREG32(mmLB_INTERRUPT_MASK + crtc_offsets[crtc], lb_interrupt_mask);
-		break;
-	case AMDGPU_IRQ_STATE_ENABLE:
-		lb_interrupt_mask = RREG32(mmLB_INTERRUPT_MASK + crtc_offsets[crtc]);
-		lb_interrupt_mask = REG_SET_FIELD(lb_interrupt_mask, LB_INTERRUPT_MASK,
-						  VLINE_INTERRUPT_MASK, 1);
-		WREG32(mmLB_INTERRUPT_MASK + crtc_offsets[crtc], lb_interrupt_mask);
-		break;
-	default:
-		break;
-	}
-}
-
-static int dce_v11_0_set_hpd_irq_state(struct amdgpu_device *adev,
-					struct amdgpu_irq_src *source,
-					unsigned hpd,
-					enum amdgpu_interrupt_state state)
-{
-	u32 tmp;
-
-	if (hpd >= adev->mode_info.num_hpd) {
-		DRM_DEBUG("invalid hdp %d\n", hpd);
-		return 0;
-	}
-
-	switch (state) {
-	case AMDGPU_IRQ_STATE_DISABLE:
-		tmp = RREG32(mmDC_HPD_INT_CONTROL + hpd_offsets[hpd]);
-		tmp = REG_SET_FIELD(tmp, DC_HPD_INT_CONTROL, DC_HPD_INT_EN, 0);
-		WREG32(mmDC_HPD_INT_CONTROL + hpd_offsets[hpd], tmp);
-		break;
-	case AMDGPU_IRQ_STATE_ENABLE:
-		tmp = RREG32(mmDC_HPD_INT_CONTROL + hpd_offsets[hpd]);
-		tmp = REG_SET_FIELD(tmp, DC_HPD_INT_CONTROL, DC_HPD_INT_EN, 1);
-		WREG32(mmDC_HPD_INT_CONTROL + hpd_offsets[hpd], tmp);
-		break;
-	default:
-		break;
-	}
-
-	return 0;
-}
-
-static int dce_v11_0_set_crtc_irq_state(struct amdgpu_device *adev,
-					struct amdgpu_irq_src *source,
-					unsigned type,
-					enum amdgpu_interrupt_state state)
-{
-	switch (type) {
-	case AMDGPU_CRTC_IRQ_VBLANK1:
-		dce_v11_0_set_crtc_vblank_interrupt_state(adev, 0, state);
-		break;
-	case AMDGPU_CRTC_IRQ_VBLANK2:
-		dce_v11_0_set_crtc_vblank_interrupt_state(adev, 1, state);
-		break;
-	case AMDGPU_CRTC_IRQ_VBLANK3:
-		dce_v11_0_set_crtc_vblank_interrupt_state(adev, 2, state);
-		break;
-	case AMDGPU_CRTC_IRQ_VBLANK4:
-		dce_v11_0_set_crtc_vblank_interrupt_state(adev, 3, state);
-		break;
-	case AMDGPU_CRTC_IRQ_VBLANK5:
-		dce_v11_0_set_crtc_vblank_interrupt_state(adev, 4, state);
-		break;
-	case AMDGPU_CRTC_IRQ_VBLANK6:
-		dce_v11_0_set_crtc_vblank_interrupt_state(adev, 5, state);
-		break;
-	case AMDGPU_CRTC_IRQ_VLINE1:
-		dce_v11_0_set_crtc_vline_interrupt_state(adev, 0, state);
-		break;
-	case AMDGPU_CRTC_IRQ_VLINE2:
-		dce_v11_0_set_crtc_vline_interrupt_state(adev, 1, state);
-		break;
-	case AMDGPU_CRTC_IRQ_VLINE3:
-		dce_v11_0_set_crtc_vline_interrupt_state(adev, 2, state);
-		break;
-	case AMDGPU_CRTC_IRQ_VLINE4:
-		dce_v11_0_set_crtc_vline_interrupt_state(adev, 3, state);
-		break;
-	case AMDGPU_CRTC_IRQ_VLINE5:
-		dce_v11_0_set_crtc_vline_interrupt_state(adev, 4, state);
-		break;
-	 case AMDGPU_CRTC_IRQ_VLINE6:
-		dce_v11_0_set_crtc_vline_interrupt_state(adev, 5, state);
-		break;
-	default:
-		break;
-	}
-	return 0;
-}
-
-static int dce_v11_0_set_pageflip_irq_state(struct amdgpu_device *adev,
-					    struct amdgpu_irq_src *src,
-					    unsigned type,
-					    enum amdgpu_interrupt_state state)
-{
-	u32 reg;
-
-	if (type >= adev->mode_info.num_crtc) {
-		DRM_ERROR("invalid pageflip crtc %d\n", type);
-		return -EINVAL;
-	}
-
-	reg = RREG32(mmGRPH_INTERRUPT_CONTROL + crtc_offsets[type]);
-	if (state == AMDGPU_IRQ_STATE_DISABLE)
-		WREG32(mmGRPH_INTERRUPT_CONTROL + crtc_offsets[type],
-		       reg & ~GRPH_INTERRUPT_CONTROL__GRPH_PFLIP_INT_MASK_MASK);
-	else
-		WREG32(mmGRPH_INTERRUPT_CONTROL + crtc_offsets[type],
-		       reg | GRPH_INTERRUPT_CONTROL__GRPH_PFLIP_INT_MASK_MASK);
-
-	return 0;
-}
-
-static int dce_v11_0_pageflip_irq(struct amdgpu_device *adev,
-				  struct amdgpu_irq_src *source,
-				  struct amdgpu_iv_entry *entry)
-{
-	unsigned long flags;
-	unsigned crtc_id;
-	struct amdgpu_crtc *amdgpu_crtc;
-	struct amdgpu_flip_work *works;
-
-	crtc_id = (entry->src_id - 8) >> 1;
-	amdgpu_crtc = adev->mode_info.crtcs[crtc_id];
-
-	if (crtc_id >= adev->mode_info.num_crtc) {
-		DRM_ERROR("invalid pageflip crtc %d\n", crtc_id);
-		return -EINVAL;
-	}
-
-	if (RREG32(mmGRPH_INTERRUPT_STATUS + crtc_offsets[crtc_id]) &
-	    GRPH_INTERRUPT_STATUS__GRPH_PFLIP_INT_OCCURRED_MASK)
-		WREG32(mmGRPH_INTERRUPT_STATUS + crtc_offsets[crtc_id],
-		       GRPH_INTERRUPT_STATUS__GRPH_PFLIP_INT_CLEAR_MASK);
-
-	/* IRQ could occur when in initial stage */
-	if(amdgpu_crtc == NULL)
-		return 0;
-
-	spin_lock_irqsave(&adev_to_drm(adev)->event_lock, flags);
-	works = amdgpu_crtc->pflip_works;
-	if (amdgpu_crtc->pflip_status != AMDGPU_FLIP_SUBMITTED){
-		DRM_DEBUG_DRIVER("amdgpu_crtc->pflip_status = %d != "
-						 "AMDGPU_FLIP_SUBMITTED(%d)\n",
-						 amdgpu_crtc->pflip_status,
-						 AMDGPU_FLIP_SUBMITTED);
-		spin_unlock_irqrestore(&adev_to_drm(adev)->event_lock, flags);
-		return 0;
-	}
-
-	/* page flip completed. clean up */
-	amdgpu_crtc->pflip_status = AMDGPU_FLIP_NONE;
-	amdgpu_crtc->pflip_works = NULL;
-
-	/* wakeup usersapce */
-	if(works->event)
-		drm_crtc_send_vblank_event(&amdgpu_crtc->base, works->event);
-
-	spin_unlock_irqrestore(&adev_to_drm(adev)->event_lock, flags);
-
-	drm_crtc_vblank_put(&amdgpu_crtc->base);
-	schedule_work(&works->unpin_work);
-
-	return 0;
-}
-
-static void dce_v11_0_hpd_int_ack(struct amdgpu_device *adev,
-				  int hpd)
-{
-	u32 tmp;
-
-	if (hpd >= adev->mode_info.num_hpd) {
-		DRM_DEBUG("invalid hdp %d\n", hpd);
-		return;
-	}
-
-	tmp = RREG32(mmDC_HPD_INT_CONTROL + hpd_offsets[hpd]);
-	tmp = REG_SET_FIELD(tmp, DC_HPD_INT_CONTROL, DC_HPD_INT_ACK, 1);
-	WREG32(mmDC_HPD_INT_CONTROL + hpd_offsets[hpd], tmp);
-}
-
-static void dce_v11_0_crtc_vblank_int_ack(struct amdgpu_device *adev,
-					  int crtc)
-{
-	u32 tmp;
-
-	if (crtc < 0 || crtc >= adev->mode_info.num_crtc) {
-		DRM_DEBUG("invalid crtc %d\n", crtc);
-		return;
-	}
-
-	tmp = RREG32(mmLB_VBLANK_STATUS + crtc_offsets[crtc]);
-	tmp = REG_SET_FIELD(tmp, LB_VBLANK_STATUS, VBLANK_ACK, 1);
-	WREG32(mmLB_VBLANK_STATUS + crtc_offsets[crtc], tmp);
-}
-
-static void dce_v11_0_crtc_vline_int_ack(struct amdgpu_device *adev,
-					 int crtc)
-{
-	u32 tmp;
-
-	if (crtc < 0 || crtc >= adev->mode_info.num_crtc) {
-		DRM_DEBUG("invalid crtc %d\n", crtc);
-		return;
-	}
-
-	tmp = RREG32(mmLB_VLINE_STATUS + crtc_offsets[crtc]);
-	tmp = REG_SET_FIELD(tmp, LB_VLINE_STATUS, VLINE_ACK, 1);
-	WREG32(mmLB_VLINE_STATUS + crtc_offsets[crtc], tmp);
-}
-
-static int dce_v11_0_crtc_irq(struct amdgpu_device *adev,
-				struct amdgpu_irq_src *source,
-				struct amdgpu_iv_entry *entry)
-{
-	unsigned crtc = entry->src_id - 1;
-	uint32_t disp_int = RREG32(interrupt_status_offsets[crtc].reg);
-	unsigned int irq_type = amdgpu_display_crtc_idx_to_irq_type(adev,
-								    crtc);
-
-	switch (entry->src_data[0]) {
-	case 0: /* vblank */
-		if (disp_int & interrupt_status_offsets[crtc].vblank)
-			dce_v11_0_crtc_vblank_int_ack(adev, crtc);
-		else
-			DRM_DEBUG("IH: IH event w/o asserted irq bit?\n");
-
-		if (amdgpu_irq_enabled(adev, source, irq_type)) {
-			drm_handle_vblank(adev_to_drm(adev), crtc);
-		}
-		DRM_DEBUG("IH: D%d vblank\n", crtc + 1);
-
-		break;
-	case 1: /* vline */
-		if (disp_int & interrupt_status_offsets[crtc].vline)
-			dce_v11_0_crtc_vline_int_ack(adev, crtc);
-		else
-			DRM_DEBUG("IH: IH event w/o asserted irq bit?\n");
-
-		DRM_DEBUG("IH: D%d vline\n", crtc + 1);
-
-		break;
-	default:
-		DRM_DEBUG("Unhandled interrupt: %d %d\n", entry->src_id, entry->src_data[0]);
-		break;
-	}
-
-	return 0;
-}
-
-static int dce_v11_0_hpd_irq(struct amdgpu_device *adev,
-			     struct amdgpu_irq_src *source,
-			     struct amdgpu_iv_entry *entry)
-{
-	uint32_t disp_int, mask;
-	unsigned hpd;
-
-	if (entry->src_data[0] >= adev->mode_info.num_hpd) {
-		DRM_DEBUG("Unhandled interrupt: %d %d\n", entry->src_id, entry->src_data[0]);
-		return 0;
-	}
-
-	hpd = entry->src_data[0];
-	disp_int = RREG32(interrupt_status_offsets[hpd].reg);
-	mask = interrupt_status_offsets[hpd].hpd;
-
-	if (disp_int & mask) {
-		dce_v11_0_hpd_int_ack(adev, hpd);
-		schedule_work(&adev->hotplug_work);
-		DRM_DEBUG("IH: HPD%d\n", hpd + 1);
-	}
-
-	return 0;
-}
-
-static int dce_v11_0_set_clockgating_state(void *handle,
-					  enum amd_clockgating_state state)
-{
-	return 0;
-}
-
-static int dce_v11_0_set_powergating_state(void *handle,
-					  enum amd_powergating_state state)
-{
-	return 0;
-}
-
-static const struct amd_ip_funcs dce_v11_0_ip_funcs = {
-	.name = "dce_v11_0",
-	.early_init = dce_v11_0_early_init,
-	.late_init = NULL,
-	.sw_init = dce_v11_0_sw_init,
-	.sw_fini = dce_v11_0_sw_fini,
-	.hw_init = dce_v11_0_hw_init,
-	.hw_fini = dce_v11_0_hw_fini,
-	.suspend = dce_v11_0_suspend,
-	.resume = dce_v11_0_resume,
-	.is_idle = dce_v11_0_is_idle,
-	.wait_for_idle = dce_v11_0_wait_for_idle,
-	.soft_reset = dce_v11_0_soft_reset,
-	.set_clockgating_state = dce_v11_0_set_clockgating_state,
-	.set_powergating_state = dce_v11_0_set_powergating_state,
-};
-
-static void
-dce_v11_0_encoder_mode_set(struct drm_encoder *encoder,
-			  struct drm_display_mode *mode,
-			  struct drm_display_mode *adjusted_mode)
-{
-	struct amdgpu_encoder *amdgpu_encoder = to_amdgpu_encoder(encoder);
-
-	amdgpu_encoder->pixel_clock = adjusted_mode->clock;
-
-	/* need to call this here rather than in prepare() since we need some crtc info */
-	amdgpu_atombios_encoder_dpms(encoder, DRM_MODE_DPMS_OFF);
-
-	/* set scaler clears this on some chips */
-	dce_v11_0_set_interleave(encoder->crtc, mode);
-
-	if (amdgpu_atombios_encoder_get_encoder_mode(encoder) == ATOM_ENCODER_MODE_HDMI) {
-		dce_v11_0_afmt_enable(encoder, true);
-		dce_v11_0_afmt_setmode(encoder, adjusted_mode);
-	}
-}
-
-static void dce_v11_0_encoder_prepare(struct drm_encoder *encoder)
-{
-	struct amdgpu_device *adev = drm_to_adev(encoder->dev);
-	struct amdgpu_encoder *amdgpu_encoder = to_amdgpu_encoder(encoder);
-	struct drm_connector *connector = amdgpu_get_connector_for_encoder(encoder);
-
-	if ((amdgpu_encoder->active_device &
-	     (ATOM_DEVICE_DFP_SUPPORT | ATOM_DEVICE_LCD_SUPPORT)) ||
-	    (amdgpu_encoder_get_dp_bridge_encoder_id(encoder) !=
-	     ENCODER_OBJECT_ID_NONE)) {
-		struct amdgpu_encoder_atom_dig *dig = amdgpu_encoder->enc_priv;
-		if (dig) {
-			dig->dig_encoder = dce_v11_0_pick_dig_encoder(encoder);
-			if (amdgpu_encoder->active_device & ATOM_DEVICE_DFP_SUPPORT)
-				dig->afmt = adev->mode_info.afmt[dig->dig_encoder];
-		}
-	}
-
-	amdgpu_atombios_scratch_regs_lock(adev, true);
-
-	if (connector) {
-		struct amdgpu_connector *amdgpu_connector = to_amdgpu_connector(connector);
-
-		/* select the clock/data port if it uses a router */
-		if (amdgpu_connector->router.cd_valid)
-			amdgpu_i2c_router_select_cd_port(amdgpu_connector);
-
-		/* turn eDP panel on for mode set */
-		if (connector->connector_type == DRM_MODE_CONNECTOR_eDP)
-			amdgpu_atombios_encoder_set_edp_panel_power(connector,
-							     ATOM_TRANSMITTER_ACTION_POWER_ON);
-	}
-
-	/* this is needed for the pll/ss setup to work correctly in some cases */
-	amdgpu_atombios_encoder_set_crtc_source(encoder);
-	/* set up the FMT blocks */
-	dce_v11_0_program_fmt(encoder);
-}
-
-static void dce_v11_0_encoder_commit(struct drm_encoder *encoder)
-{
-	struct drm_device *dev = encoder->dev;
-	struct amdgpu_device *adev = drm_to_adev(dev);
-
-	/* need to call this here as we need the crtc set up */
-	amdgpu_atombios_encoder_dpms(encoder, DRM_MODE_DPMS_ON);
-	amdgpu_atombios_scratch_regs_lock(adev, false);
-}
-
-static void dce_v11_0_encoder_disable(struct drm_encoder *encoder)
-{
-	struct amdgpu_encoder *amdgpu_encoder = to_amdgpu_encoder(encoder);
-	struct amdgpu_encoder_atom_dig *dig;
-
-	amdgpu_atombios_encoder_dpms(encoder, DRM_MODE_DPMS_OFF);
-
-	if (amdgpu_atombios_encoder_is_digital(encoder)) {
-		if (amdgpu_atombios_encoder_get_encoder_mode(encoder) == ATOM_ENCODER_MODE_HDMI)
-			dce_v11_0_afmt_enable(encoder, false);
-		dig = amdgpu_encoder->enc_priv;
-		dig->dig_encoder = -1;
-	}
-	amdgpu_encoder->active_device = 0;
-}
-
-/* these are handled by the primary encoders */
-static void dce_v11_0_ext_prepare(struct drm_encoder *encoder)
-{
-
-}
-
-static void dce_v11_0_ext_commit(struct drm_encoder *encoder)
-{
-
-}
-
-static void
-dce_v11_0_ext_mode_set(struct drm_encoder *encoder,
-		      struct drm_display_mode *mode,
-		      struct drm_display_mode *adjusted_mode)
-{
-
-}
-
-static void dce_v11_0_ext_disable(struct drm_encoder *encoder)
-{
-
-}
-
-static void
-dce_v11_0_ext_dpms(struct drm_encoder *encoder, int mode)
-{
-
-}
-
-static const struct drm_encoder_helper_funcs dce_v11_0_ext_helper_funcs = {
-	.dpms = dce_v11_0_ext_dpms,
-	.prepare = dce_v11_0_ext_prepare,
-	.mode_set = dce_v11_0_ext_mode_set,
-	.commit = dce_v11_0_ext_commit,
-	.disable = dce_v11_0_ext_disable,
-	/* no detect for TMDS/LVDS yet */
-};
-
-static const struct drm_encoder_helper_funcs dce_v11_0_dig_helper_funcs = {
-	.dpms = amdgpu_atombios_encoder_dpms,
-	.mode_fixup = amdgpu_atombios_encoder_mode_fixup,
-	.prepare = dce_v11_0_encoder_prepare,
-	.mode_set = dce_v11_0_encoder_mode_set,
-	.commit = dce_v11_0_encoder_commit,
-	.disable = dce_v11_0_encoder_disable,
-	.detect = amdgpu_atombios_encoder_dig_detect,
-};
-
-static const struct drm_encoder_helper_funcs dce_v11_0_dac_helper_funcs = {
-	.dpms = amdgpu_atombios_encoder_dpms,
-	.mode_fixup = amdgpu_atombios_encoder_mode_fixup,
-	.prepare = dce_v11_0_encoder_prepare,
-	.mode_set = dce_v11_0_encoder_mode_set,
-	.commit = dce_v11_0_encoder_commit,
-	.detect = amdgpu_atombios_encoder_dac_detect,
-};
-
-static void dce_v11_0_encoder_destroy(struct drm_encoder *encoder)
-{
-	struct amdgpu_encoder *amdgpu_encoder = to_amdgpu_encoder(encoder);
-	if (amdgpu_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT))
-		amdgpu_atombios_encoder_fini_backlight(amdgpu_encoder);
-	kfree(amdgpu_encoder->enc_priv);
-	drm_encoder_cleanup(encoder);
-	kfree(amdgpu_encoder);
-}
-
-static const struct drm_encoder_funcs dce_v11_0_encoder_funcs = {
-	.destroy = dce_v11_0_encoder_destroy,
-};
-
-static void dce_v11_0_encoder_add(struct amdgpu_device *adev,
-				 uint32_t encoder_enum,
-				 uint32_t supported_device,
-				 u16 caps)
-{
-	struct drm_device *dev = adev_to_drm(adev);
-	struct drm_encoder *encoder;
-	struct amdgpu_encoder *amdgpu_encoder;
-
-	/* see if we already added it */
-	list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) {
-		amdgpu_encoder = to_amdgpu_encoder(encoder);
-		if (amdgpu_encoder->encoder_enum == encoder_enum) {
-			amdgpu_encoder->devices |= supported_device;
-			return;
-		}
-
-	}
-
-	/* add a new one */
-	amdgpu_encoder = kzalloc(sizeof(struct amdgpu_encoder), GFP_KERNEL);
-	if (!amdgpu_encoder)
-		return;
-
-	encoder = &amdgpu_encoder->base;
-	switch (adev->mode_info.num_crtc) {
-	case 1:
-		encoder->possible_crtcs = 0x1;
-		break;
-	case 2:
-	default:
-		encoder->possible_crtcs = 0x3;
-		break;
-	case 3:
-		encoder->possible_crtcs = 0x7;
-		break;
-	case 4:
-		encoder->possible_crtcs = 0xf;
-		break;
-	case 5:
-		encoder->possible_crtcs = 0x1f;
-		break;
-	case 6:
-		encoder->possible_crtcs = 0x3f;
-		break;
-	}
-
-	amdgpu_encoder->enc_priv = NULL;
-
-	amdgpu_encoder->encoder_enum = encoder_enum;
-	amdgpu_encoder->encoder_id = (encoder_enum & OBJECT_ID_MASK) >> OBJECT_ID_SHIFT;
-	amdgpu_encoder->devices = supported_device;
-	amdgpu_encoder->rmx_type = RMX_OFF;
-	amdgpu_encoder->underscan_type = UNDERSCAN_OFF;
-	amdgpu_encoder->is_ext_encoder = false;
-	amdgpu_encoder->caps = caps;
-
-	switch (amdgpu_encoder->encoder_id) {
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC1:
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC2:
-		drm_encoder_init(dev, encoder, &dce_v11_0_encoder_funcs,
-				 DRM_MODE_ENCODER_DAC, NULL);
-		drm_encoder_helper_add(encoder, &dce_v11_0_dac_helper_funcs);
-		break;
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DVO1:
-	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
-	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
-	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
-	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY3:
-		if (amdgpu_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT)) {
-			amdgpu_encoder->rmx_type = RMX_FULL;
-			drm_encoder_init(dev, encoder, &dce_v11_0_encoder_funcs,
-					 DRM_MODE_ENCODER_LVDS, NULL);
-			amdgpu_encoder->enc_priv = amdgpu_atombios_encoder_get_lcd_info(amdgpu_encoder);
-		} else if (amdgpu_encoder->devices & (ATOM_DEVICE_CRT_SUPPORT)) {
-			drm_encoder_init(dev, encoder, &dce_v11_0_encoder_funcs,
-					 DRM_MODE_ENCODER_DAC, NULL);
-			amdgpu_encoder->enc_priv = amdgpu_atombios_encoder_get_dig_info(amdgpu_encoder);
-		} else {
-			drm_encoder_init(dev, encoder, &dce_v11_0_encoder_funcs,
-					 DRM_MODE_ENCODER_TMDS, NULL);
-			amdgpu_encoder->enc_priv = amdgpu_atombios_encoder_get_dig_info(amdgpu_encoder);
-		}
-		drm_encoder_helper_add(encoder, &dce_v11_0_dig_helper_funcs);
-		break;
-	case ENCODER_OBJECT_ID_SI170B:
-	case ENCODER_OBJECT_ID_CH7303:
-	case ENCODER_OBJECT_ID_EXTERNAL_SDVOA:
-	case ENCODER_OBJECT_ID_EXTERNAL_SDVOB:
-	case ENCODER_OBJECT_ID_TITFP513:
-	case ENCODER_OBJECT_ID_VT1623:
-	case ENCODER_OBJECT_ID_HDMI_SI1930:
-	case ENCODER_OBJECT_ID_TRAVIS:
-	case ENCODER_OBJECT_ID_NUTMEG:
-		/* these are handled by the primary encoders */
-		amdgpu_encoder->is_ext_encoder = true;
-		if (amdgpu_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT))
-			drm_encoder_init(dev, encoder, &dce_v11_0_encoder_funcs,
-					 DRM_MODE_ENCODER_LVDS, NULL);
-		else if (amdgpu_encoder->devices & (ATOM_DEVICE_CRT_SUPPORT))
-			drm_encoder_init(dev, encoder, &dce_v11_0_encoder_funcs,
-					 DRM_MODE_ENCODER_DAC, NULL);
-		else
-			drm_encoder_init(dev, encoder, &dce_v11_0_encoder_funcs,
-					 DRM_MODE_ENCODER_TMDS, NULL);
-		drm_encoder_helper_add(encoder, &dce_v11_0_ext_helper_funcs);
-		break;
-	}
-}
-
-static const struct amdgpu_display_funcs dce_v11_0_display_funcs = {
-	.bandwidth_update = &dce_v11_0_bandwidth_update,
-	.vblank_get_counter = &dce_v11_0_vblank_get_counter,
-	.backlight_set_level = &amdgpu_atombios_encoder_set_backlight_level,
-	.backlight_get_level = &amdgpu_atombios_encoder_get_backlight_level,
-	.hpd_sense = &dce_v11_0_hpd_sense,
-	.hpd_set_polarity = &dce_v11_0_hpd_set_polarity,
-	.hpd_get_gpio_reg = &dce_v11_0_hpd_get_gpio_reg,
-	.page_flip = &dce_v11_0_page_flip,
-	.page_flip_get_scanoutpos = &dce_v11_0_crtc_get_scanoutpos,
-	.add_encoder = &dce_v11_0_encoder_add,
-	.add_connector = &amdgpu_connector_add,
-};
-
-static void dce_v11_0_set_display_funcs(struct amdgpu_device *adev)
-{
-	adev->mode_info.funcs = &dce_v11_0_display_funcs;
-}
-
-static const struct amdgpu_irq_src_funcs dce_v11_0_crtc_irq_funcs = {
-	.set = dce_v11_0_set_crtc_irq_state,
-	.process = dce_v11_0_crtc_irq,
-};
-
-static const struct amdgpu_irq_src_funcs dce_v11_0_pageflip_irq_funcs = {
-	.set = dce_v11_0_set_pageflip_irq_state,
-	.process = dce_v11_0_pageflip_irq,
-};
-
-static const struct amdgpu_irq_src_funcs dce_v11_0_hpd_irq_funcs = {
-	.set = dce_v11_0_set_hpd_irq_state,
-	.process = dce_v11_0_hpd_irq,
-};
-
-static void dce_v11_0_set_irq_funcs(struct amdgpu_device *adev)
-{
-	if (adev->mode_info.num_crtc > 0)
-		adev->crtc_irq.num_types = AMDGPU_CRTC_IRQ_VLINE1 + adev->mode_info.num_crtc;
-	else
-		adev->crtc_irq.num_types = 0;
-	adev->crtc_irq.funcs = &dce_v11_0_crtc_irq_funcs;
-
-	adev->pageflip_irq.num_types = adev->mode_info.num_crtc;
-	adev->pageflip_irq.funcs = &dce_v11_0_pageflip_irq_funcs;
-
-	adev->hpd_irq.num_types = adev->mode_info.num_hpd;
-	adev->hpd_irq.funcs = &dce_v11_0_hpd_irq_funcs;
-}
-
-const struct amdgpu_ip_block_version dce_v11_0_ip_block =
-{
-	.type = AMD_IP_BLOCK_TYPE_DCE,
-	.major = 11,
-	.minor = 0,
-	.rev = 0,
-	.funcs = &dce_v11_0_ip_funcs,
-};
-
-const struct amdgpu_ip_block_version dce_v11_2_ip_block =
-{
-	.type = AMD_IP_BLOCK_TYPE_DCE,
-	.major = 11,
-	.minor = 2,
-	.rev = 0,
-	.funcs = &dce_v11_0_ip_funcs,
-};
diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v11_0.h b/drivers/gpu/drm/amd/amdgpu/dce_v11_0.h
deleted file mode 100644
index 0d878ca3acba..000000000000
--- a/drivers/gpu/drm/amd/amdgpu/dce_v11_0.h
+++ /dev/null
@@ -1,32 +0,0 @@
-/*
- * Copyright 2014 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- */
-
-#ifndef __DCE_V11_0_H__
-#define __DCE_V11_0_H__
-
-extern const struct amdgpu_ip_block_version dce_v11_0_ip_block;
-extern const struct amdgpu_ip_block_version dce_v11_2_ip_block;
-
-void dce_v11_0_disable_dce(struct amdgpu_device *adev);
-
-#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c
index c7803dc2b2d5..acc887a58518 100644
--- a/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c
@@ -23,7 +23,10 @@
 
 #include <linux/pci.h>
 
+#include <drm/drm_edid.h>
 #include <drm/drm_fourcc.h>
+#include <drm/drm_modeset_helper.h>
+#include <drm/drm_modeset_helper_vtables.h>
 #include <drm/drm_vblank.h>
 
 #include "amdgpu.h"
@@ -37,18 +40,25 @@
 #include "amdgpu_connectors.h"
 #include "amdgpu_display.h"
 
+#include "dce_v6_0.h"
+#include "sid.h"
+
 #include "bif/bif_3_0_d.h"
 #include "bif/bif_3_0_sh_mask.h"
+
 #include "oss/oss_1_0_d.h"
 #include "oss/oss_1_0_sh_mask.h"
+
 #include "gca/gfx_6_0_d.h"
 #include "gca/gfx_6_0_sh_mask.h"
+#include "gca/gfx_7_2_enum.h"
+
 #include "gmc/gmc_6_0_d.h"
 #include "gmc/gmc_6_0_sh_mask.h"
+
 #include "dce/dce_6_0_d.h"
 #include "dce/dce_6_0_sh_mask.h"
-#include "gca/gfx_7_2_enum.h"
-#include "dce_v6_0.h"
+
 #include "si_enums.h"
 
 static void dce_v6_0_set_display_funcs(struct amdgpu_device *adev);
@@ -56,31 +66,31 @@ static void dce_v6_0_set_irq_funcs(struct amdgpu_device *adev);
 
 static const u32 crtc_offsets[6] =
 {
-	SI_CRTC0_REGISTER_OFFSET,
-	SI_CRTC1_REGISTER_OFFSET,
-	SI_CRTC2_REGISTER_OFFSET,
-	SI_CRTC3_REGISTER_OFFSET,
-	SI_CRTC4_REGISTER_OFFSET,
-	SI_CRTC5_REGISTER_OFFSET
+	CRTC0_REGISTER_OFFSET,
+	CRTC1_REGISTER_OFFSET,
+	CRTC2_REGISTER_OFFSET,
+	CRTC3_REGISTER_OFFSET,
+	CRTC4_REGISTER_OFFSET,
+	CRTC5_REGISTER_OFFSET
 };
 
 static const u32 hpd_offsets[] =
 {
-	mmDC_HPD1_INT_STATUS - mmDC_HPD1_INT_STATUS,
-	mmDC_HPD2_INT_STATUS - mmDC_HPD1_INT_STATUS,
-	mmDC_HPD3_INT_STATUS - mmDC_HPD1_INT_STATUS,
-	mmDC_HPD4_INT_STATUS - mmDC_HPD1_INT_STATUS,
-	mmDC_HPD5_INT_STATUS - mmDC_HPD1_INT_STATUS,
-	mmDC_HPD6_INT_STATUS - mmDC_HPD1_INT_STATUS,
+	HPD0_REGISTER_OFFSET,
+	HPD1_REGISTER_OFFSET,
+	HPD2_REGISTER_OFFSET,
+	HPD3_REGISTER_OFFSET,
+	HPD4_REGISTER_OFFSET,
+	HPD5_REGISTER_OFFSET
 };
 
 static const uint32_t dig_offsets[] = {
-	SI_CRTC0_REGISTER_OFFSET,
-	SI_CRTC1_REGISTER_OFFSET,
-	SI_CRTC2_REGISTER_OFFSET,
-	SI_CRTC3_REGISTER_OFFSET,
-	SI_CRTC4_REGISTER_OFFSET,
-	SI_CRTC5_REGISTER_OFFSET,
+	CRTC0_REGISTER_OFFSET,
+	CRTC1_REGISTER_OFFSET,
+	CRTC2_REGISTER_OFFSET,
+	CRTC3_REGISTER_OFFSET,
+	CRTC4_REGISTER_OFFSET,
+	CRTC5_REGISTER_OFFSET,
 	(0x13830 - 0x7030) >> 2,
 };
 
@@ -203,9 +213,9 @@ static void dce_v6_0_page_flip(struct amdgpu_device *adev,
 	/* update the scanout addresses */
 	WREG32(mmGRPH_PRIMARY_SURFACE_ADDRESS_HIGH + amdgpu_crtc->crtc_offset,
 	       upper_32_bits(crtc_base));
+	/* writing to the low address triggers the update */
 	WREG32(mmGRPH_PRIMARY_SURFACE_ADDRESS + amdgpu_crtc->crtc_offset,
 	       (u32)crtc_base);
-
 	/* post the write */
 	RREG32(mmGRPH_PRIMARY_SURFACE_ADDRESS + amdgpu_crtc->crtc_offset);
 }
@@ -215,11 +225,11 @@ static int dce_v6_0_crtc_get_scanoutpos(struct amdgpu_device *adev, int crtc,
 {
 	if ((crtc < 0) || (crtc >= adev->mode_info.num_crtc))
 		return -EINVAL;
+
 	*vbl = RREG32(mmCRTC_V_BLANK_START_END + crtc_offsets[crtc]);
 	*position = RREG32(mmCRTC_STATUS_POSITION + crtc_offsets[crtc]);
 
 	return 0;
-
 }
 
 /**
@@ -239,7 +249,8 @@ static bool dce_v6_0_hpd_sense(struct amdgpu_device *adev,
 	if (hpd >= adev->mode_info.num_hpd)
 		return connected;
 
-	if (RREG32(mmDC_HPD1_INT_STATUS + hpd_offsets[hpd]) & DC_HPD1_INT_STATUS__DC_HPD1_SENSE_MASK)
+	if (RREG32(mmDC_HPD1_INT_STATUS + hpd_offsets[hpd]) &
+	    DC_HPD1_INT_STATUS__DC_HPD1_SENSE_MASK)
 		connected = true;
 
 	return connected;
@@ -270,6 +281,21 @@ static void dce_v6_0_hpd_set_polarity(struct amdgpu_device *adev,
 	WREG32(mmDC_HPD1_INT_CONTROL + hpd_offsets[hpd], tmp);
 }
 
+static void dce_v6_0_hpd_int_ack(struct amdgpu_device *adev,
+				 int hpd)
+{
+	u32 tmp;
+
+	if (hpd >= adev->mode_info.num_hpd) {
+		DRM_DEBUG("invalid hpd %d\n", hpd);
+		return;
+	}
+
+	tmp = RREG32(mmDC_HPD1_INT_CONTROL + hpd_offsets[hpd]);
+	tmp |= DC_HPD1_INT_CONTROL__DC_HPD1_INT_ACK_MASK;
+	WREG32(mmDC_HPD1_INT_CONTROL + hpd_offsets[hpd], tmp);
+}
+
 /**
  * dce_v6_0_hpd_init - hpd setup callback.
  *
@@ -309,6 +335,7 @@ static void dce_v6_0_hpd_init(struct amdgpu_device *adev)
 			continue;
 		}
 
+		dce_v6_0_hpd_int_ack(adev, amdgpu_connector->hpd.hpd);
 		dce_v6_0_hpd_set_polarity(adev, amdgpu_connector->hpd.hpd);
 		amdgpu_irq_get(adev, &adev->hpd_irq, amdgpu_connector->hpd.hpd);
 	}
@@ -339,7 +366,7 @@ static void dce_v6_0_hpd_fini(struct amdgpu_device *adev)
 
 		tmp = RREG32(mmDC_HPD1_CONTROL + hpd_offsets[amdgpu_connector->hpd.hpd]);
 		tmp &= ~DC_HPD1_CONTROL__DC_HPD1_EN_MASK;
-		WREG32(mmDC_HPD1_CONTROL + hpd_offsets[amdgpu_connector->hpd.hpd], 0);
+		WREG32(mmDC_HPD1_CONTROL + hpd_offsets[amdgpu_connector->hpd.hpd], tmp);
 
 		amdgpu_irq_put(adev, &adev->hpd_irq, amdgpu_connector->hpd.hpd);
 	}
@@ -351,13 +378,41 @@ static u32 dce_v6_0_hpd_get_gpio_reg(struct amdgpu_device *adev)
 	return mmDC_GPIO_HPD_A;
 }
 
+static bool dce_v6_0_is_display_hung(struct amdgpu_device *adev)
+{
+	u32 crtc_hung = 0;
+	u32 crtc_status[6];
+	u32 i, j, tmp;
+
+	for (i = 0; i < adev->mode_info.num_crtc; i++) {
+		if (RREG32(mmCRTC_CONTROL + crtc_offsets[i]) & CRTC_CONTROL__CRTC_MASTER_EN_MASK) {
+			crtc_status[i] = RREG32(mmCRTC_STATUS_HV_COUNT + crtc_offsets[i]);
+			crtc_hung |= (1 << i);
+		}
+	}
+
+	for (j = 0; j < 10; j++) {
+		for (i = 0; i < adev->mode_info.num_crtc; i++) {
+			if (crtc_hung & (1 << i)) {
+				tmp = RREG32(mmCRTC_STATUS_HV_COUNT + crtc_offsets[i]);
+				if (tmp != crtc_status[i])
+					crtc_hung &= ~(1 << i);
+			}
+		}
+		if (crtc_hung == 0)
+			return false;
+		udelay(100);
+	}
+
+	return true;
+}
+
 static void dce_v6_0_set_vga_render_state(struct amdgpu_device *adev,
 					  bool render)
 {
 	if (!render)
 		WREG32(mmVGA_RENDER_CONTROL,
-			RREG32(mmVGA_RENDER_CONTROL) & VGA_VSTATUS_CNTL);
-
+		       RREG32(mmVGA_RENDER_CONTROL) & ~VGA_RENDER_CONTROL__VGA_VSTATUS_CNTL_MASK);
 }
 
 static int dce_v6_0_get_num_crtc(struct amdgpu_device *adev)
@@ -400,7 +455,6 @@ void dce_v6_0_disable_dce(struct amdgpu_device *adev)
 
 static void dce_v6_0_program_fmt(struct drm_encoder *encoder)
 {
-
 	struct drm_device *dev = encoder->dev;
 	struct amdgpu_device *adev = drm_to_adev(dev);
 	struct amdgpu_encoder *amdgpu_encoder = to_amdgpu_encoder(encoder);
@@ -843,7 +897,7 @@ static void dce_v6_0_program_watermarks(struct amdgpu_device *adev,
 					    (u32)mode->clock);
 		line_time = (u32) div_u64((u64)mode->crtc_htotal * 1000000,
 					  (u32)mode->clock);
-		line_time = min(line_time, (u32)65535);
+		line_time = min_t(u32, line_time, 65535);
 		priority_a_cnt = 0;
 		priority_b_cnt = 0;
 
@@ -876,8 +930,8 @@ static void dce_v6_0_program_watermarks(struct amdgpu_device *adev,
 		wm_high.dram_channels = dram_channels;
 		wm_high.num_heads = num_heads;
 
-		if (adev->pm.dpm_enabled) {
 		/* watermark for low clocks */
+		if (adev->pm.dpm_enabled) {
 			wm_low.yclk =
 				amdgpu_dpm_get_mclk(adev, true) * 10;
 			wm_low.sclk =
@@ -904,9 +958,9 @@ static void dce_v6_0_program_watermarks(struct amdgpu_device *adev,
 		wm_low.num_heads = num_heads;
 
 		/* set for high clocks */
-		latency_watermark_a = min(dce_v6_0_latency_watermark(&wm_high), (u32)65535);
+		latency_watermark_a = min_t(u32, dce_v6_0_latency_watermark(&wm_high), 65535);
 		/* set for low clocks */
-		latency_watermark_b = min(dce_v6_0_latency_watermark(&wm_low), (u32)65535);
+		latency_watermark_b = min_t(u32, dce_v6_0_latency_watermark(&wm_low), 65535);
 
 		/* possibly force display priority to high */
 		/* should really do this at mode validation time... */
@@ -957,16 +1011,16 @@ static void dce_v6_0_program_watermarks(struct amdgpu_device *adev,
 	/* select wm A */
 	arb_control3 = RREG32(mmDPG_PIPE_ARBITRATION_CONTROL3 + amdgpu_crtc->crtc_offset);
 	tmp = arb_control3;
-	tmp &= ~LATENCY_WATERMARK_MASK(3);
-	tmp |= LATENCY_WATERMARK_MASK(1);
+	tmp &= ~(3 << DPG_PIPE_ARBITRATION_CONTROL3__URGENCY_WATERMARK_MASK__SHIFT);
+	tmp |= (1 << DPG_PIPE_ARBITRATION_CONTROL3__URGENCY_WATERMARK_MASK__SHIFT);
 	WREG32(mmDPG_PIPE_ARBITRATION_CONTROL3 + amdgpu_crtc->crtc_offset, tmp);
 	WREG32(mmDPG_PIPE_URGENCY_CONTROL + amdgpu_crtc->crtc_offset,
 	       ((latency_watermark_a << DPG_PIPE_URGENCY_CONTROL__URGENCY_LOW_WATERMARK__SHIFT)  |
 		(line_time << DPG_PIPE_URGENCY_CONTROL__URGENCY_HIGH_WATERMARK__SHIFT)));
 	/* select wm B */
 	tmp = RREG32(mmDPG_PIPE_ARBITRATION_CONTROL3 + amdgpu_crtc->crtc_offset);
-	tmp &= ~LATENCY_WATERMARK_MASK(3);
-	tmp |= LATENCY_WATERMARK_MASK(2);
+	tmp &= ~(3 << DPG_PIPE_ARBITRATION_CONTROL3__URGENCY_WATERMARK_MASK__SHIFT);
+	tmp |= (2 << DPG_PIPE_ARBITRATION_CONTROL3__URGENCY_WATERMARK_MASK__SHIFT);
 	WREG32(mmDPG_PIPE_ARBITRATION_CONTROL3 + amdgpu_crtc->crtc_offset, tmp);
 	WREG32(mmDPG_PIPE_URGENCY_CONTROL + amdgpu_crtc->crtc_offset,
 	       ((latency_watermark_b << DPG_PIPE_URGENCY_CONTROL__URGENCY_LOW_WATERMARK__SHIFT) |
@@ -980,13 +1034,26 @@ static void dce_v6_0_program_watermarks(struct amdgpu_device *adev,
 
 	/* save values for DPM */
 	amdgpu_crtc->line_time = line_time;
-	amdgpu_crtc->wm_high = latency_watermark_a;
 
 	/* Save number of lines the linebuffer leads before the scanout */
 	amdgpu_crtc->lb_vblank_lead_lines = lb_vblank_lead_lines;
 }
 
 /* watermark setup */
+/**
+ * dce_v6_0_line_buffer_adjust - Set up the line buffer
+ *
+ * @adev: amdgpu_device pointer
+ * @amdgpu_crtc: the selected display controller
+ * @mode: the current display mode on the selected display
+ * controller
+ * @other_mode: the display mode of another display controller
+ *              that may be sharing the line buffer
+ *
+ * Setup up the line buffer allocation for
+ * the selected display controller (CIK).
+ * Returns the line buffer size in pixels.
+ */
 static u32 dce_v6_0_line_buffer_adjust(struct amdgpu_device *adev,
 				   struct amdgpu_crtc *amdgpu_crtc,
 				   struct drm_display_mode *mode,
@@ -1021,7 +1088,7 @@ static u32 dce_v6_0_line_buffer_adjust(struct amdgpu_device *adev,
 	}
 
 	WREG32(mmDC_LB_MEMORY_SPLIT + amdgpu_crtc->crtc_offset,
-	       DC_LB_MEMORY_CONFIG(tmp));
+	       (tmp << DC_LB_MEMORY_SPLIT__DC_LB_MEMORY_CONFIG__SHIFT));
 
 	WREG32(mmPIPE0_DMIF_BUFFER_CONTROL + pipe_offset,
 	       (buffer_alloc << PIPE0_DMIF_BUFFER_CONTROL__DMIF_BUFFERS_ALLOCATED__SHIFT));
@@ -1198,7 +1265,7 @@ static void dce_v6_0_audio_write_speaker_allocation(struct drm_encoder *encoder)
 		return;
 	}
 
-	sad_count = drm_edid_to_speaker_allocation(amdgpu_connector_edid(connector), &sadb);
+	sad_count = drm_edid_to_speaker_allocation(amdgpu_connector->edid, &sadb);
 	if (sad_count < 0) {
 		DRM_ERROR("Couldn't read Speaker Allocation Data Block: %d\n", sad_count);
 		sad_count = 0;
@@ -1238,6 +1305,7 @@ static void dce_v6_0_audio_write_sad_regs(struct drm_encoder *encoder)
 	struct amdgpu_device *adev = drm_to_adev(dev);
 	struct amdgpu_encoder *amdgpu_encoder = to_amdgpu_encoder(encoder);
 	struct amdgpu_encoder_atom_dig *dig = amdgpu_encoder->enc_priv;
+	u32 offset;
 	struct drm_connector *connector;
 	struct drm_connector_list_iter iter;
 	struct amdgpu_connector *amdgpu_connector = NULL;
@@ -1259,6 +1327,11 @@ static void dce_v6_0_audio_write_sad_regs(struct drm_encoder *encoder)
 		{ ixAZALIA_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR13, HDMI_AUDIO_CODING_TYPE_WMA_PRO },
 	};
 
+	if (!dig || !dig->afmt || !dig->afmt->pin)
+		return;
+
+	offset = dig->afmt->pin->offset;
+
 	drm_connector_list_iter_begin(dev, &iter);
 	drm_for_each_connector_iter(connector, &iter) {
 		if (connector->encoder == encoder) {
@@ -1273,14 +1346,14 @@ static void dce_v6_0_audio_write_sad_regs(struct drm_encoder *encoder)
 		return;
 	}
 
-	sad_count = drm_edid_to_sad(amdgpu_connector_edid(connector), &sads);
+	sad_count = drm_edid_to_sad(amdgpu_connector->edid, &sads);
 	if (sad_count < 0)
 		DRM_ERROR("Couldn't read SADs: %d\n", sad_count);
 	if (sad_count <= 0)
 		return;
 
 	for (i = 0; i < ARRAY_SIZE(eld_reg_to_type); i++) {
-		u32 tmp = 0;
+		u32 value = 0;
 		u8 stereo_freqs = 0;
 		int max_channels = -1;
 		int j;
@@ -1290,12 +1363,12 @@ static void dce_v6_0_audio_write_sad_regs(struct drm_encoder *encoder)
 
 			if (sad->format == eld_reg_to_type[i][1]) {
 				if (sad->channels > max_channels) {
-					tmp = REG_SET_FIELD(tmp, AZALIA_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR0,
-							MAX_CHANNELS, sad->channels);
-					tmp = REG_SET_FIELD(tmp, AZALIA_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR0,
-							DESCRIPTOR_BYTE_2, sad->byte2);
-					tmp = REG_SET_FIELD(tmp, AZALIA_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR0,
-							SUPPORTED_FREQUENCIES, sad->freq);
+					value = (sad->channels <<
+						AZALIA_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR0__MAX_CHANNELS__SHIFT) |
+					       (sad->byte2 <<
+						AZALIA_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR0__DESCRIPTOR_BYTE_2__SHIFT) |
+					       (sad->freq <<
+						AZALIA_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR0__SUPPORTED_FREQUENCIES__SHIFT);
 					max_channels = sad->channels;
 				}
 
@@ -1306,13 +1379,13 @@ static void dce_v6_0_audio_write_sad_regs(struct drm_encoder *encoder)
 			}
 		}
 
-		tmp = REG_SET_FIELD(tmp, AZALIA_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR0,
-				SUPPORTED_FREQUENCIES_STEREO, stereo_freqs);
-		WREG32_AUDIO_ENDPT(dig->afmt->pin->offset, eld_reg_to_type[i][0], tmp);
+		value |= (stereo_freqs <<
+			AZALIA_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR0__SUPPORTED_FREQUENCIES_STEREO__SHIFT);
+
+		WREG32_AUDIO_ENDPT(offset, eld_reg_to_type[i][0], value);
 	}
 
 	kfree(sads);
-
 }
 
 static void dce_v6_0_audio_enable(struct amdgpu_device *adev,
@@ -1328,13 +1401,13 @@ static void dce_v6_0_audio_enable(struct amdgpu_device *adev,
 
 static const u32 pin_offsets[7] =
 {
-	(0x1780 - 0x1780),
-	(0x1786 - 0x1780),
-	(0x178c - 0x1780),
-	(0x1792 - 0x1780),
-	(0x1798 - 0x1780),
-	(0x179d - 0x1780),
-	(0x17a4 - 0x1780),
+	AUD0_REGISTER_OFFSET,
+	AUD1_REGISTER_OFFSET,
+	AUD2_REGISTER_OFFSET,
+	AUD3_REGISTER_OFFSET,
+	AUD4_REGISTER_OFFSET,
+	AUD5_REGISTER_OFFSET,
+	AUD6_REGISTER_OFFSET,
 };
 
 static int dce_v6_0_audio_init(struct amdgpu_device *adev)
@@ -1367,6 +1440,8 @@ static int dce_v6_0_audio_init(struct amdgpu_device *adev)
 		adev->mode_info.audio.pin[i].connected = false;
 		adev->mode_info.audio.pin[i].offset = pin_offsets[i];
 		adev->mode_info.audio.pin[i].id = i;
+		/* disable audio.  it will be set up later */
+		/* XXX remove once we switch to ip funcs */
 		dce_v6_0_audio_enable(adev, &adev->mode_info.audio.pin[i], false);
 	}
 
@@ -1375,17 +1450,12 @@ static int dce_v6_0_audio_init(struct amdgpu_device *adev)
 
 static void dce_v6_0_audio_fini(struct amdgpu_device *adev)
 {
-	int i;
-
 	if (!amdgpu_audio)
 		return;
 
 	if (!adev->mode_info.audio.enabled)
 		return;
 
-	for (i = 0; i < adev->mode_info.audio.num_pins; i++)
-		dce_v6_0_audio_enable(adev, &adev->mode_info.audio.pin[i], false);
-
 	adev->mode_info.audio.enabled = false;
 }
 
@@ -1816,7 +1886,7 @@ static int dce_v6_0_crtc_do_set_base(struct drm_crtc *crtc,
 	struct amdgpu_bo *abo;
 	uint64_t fb_location, tiling_flags;
 	uint32_t fb_format, fb_pitch_pixels, pipe_config;
-	u32 fb_swap = GRPH_ENDIAN_SWAP(GRPH_ENDIAN_NONE);
+	u32 fb_swap = (GRPH_ENDIAN_NONE << GRPH_SWAP_CNTL__GRPH_ENDIAN_SWAP__SHIFT);
 	u32 viewport_w, viewport_h;
 	int r;
 	bool bypass_lut = false;
@@ -1842,6 +1912,7 @@ static int dce_v6_0_crtc_do_set_base(struct drm_crtc *crtc,
 		return r;
 
 	if (!atomic) {
+		abo->flags |= AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS;
 		r = amdgpu_bo_pin(abo, AMDGPU_GEM_DOMAIN_VRAM);
 		if (unlikely(r != 0)) {
 			amdgpu_bo_unreserve(abo);
@@ -1855,76 +1926,76 @@ static int dce_v6_0_crtc_do_set_base(struct drm_crtc *crtc,
 
 	switch (target_fb->format->format) {
 	case DRM_FORMAT_C8:
-		fb_format = (GRPH_DEPTH(GRPH_DEPTH_8BPP) |
-			     GRPH_FORMAT(GRPH_FORMAT_INDEXED));
+		fb_format = ((GRPH_DEPTH_8BPP << GRPH_CONTROL__GRPH_DEPTH__SHIFT) |
+			     (GRPH_FORMAT_INDEXED << GRPH_CONTROL__GRPH_FORMAT__SHIFT));
 		break;
 	case DRM_FORMAT_XRGB4444:
 	case DRM_FORMAT_ARGB4444:
-		fb_format = (GRPH_DEPTH(GRPH_DEPTH_16BPP) |
-			     GRPH_FORMAT(GRPH_FORMAT_ARGB4444));
+		fb_format = ((GRPH_DEPTH_16BPP << GRPH_CONTROL__GRPH_DEPTH__SHIFT) |
+			     (GRPH_FORMAT_ARGB4444 << GRPH_CONTROL__GRPH_FORMAT__SHIFT));
 #ifdef __BIG_ENDIAN
-		fb_swap = GRPH_ENDIAN_SWAP(GRPH_ENDIAN_8IN16);
+		fb_swap = (GRPH_ENDIAN_8IN16 << GRPH_SWAP_CNTL__GRPH_ENDIAN_SWAP__SHIFT);
 #endif
 		break;
 	case DRM_FORMAT_XRGB1555:
 	case DRM_FORMAT_ARGB1555:
-		fb_format = (GRPH_DEPTH(GRPH_DEPTH_16BPP) |
-			     GRPH_FORMAT(GRPH_FORMAT_ARGB1555));
+		fb_format = ((GRPH_DEPTH_16BPP << GRPH_CONTROL__GRPH_DEPTH__SHIFT) |
+			     (GRPH_FORMAT_ARGB1555 << GRPH_CONTROL__GRPH_FORMAT__SHIFT));
 #ifdef __BIG_ENDIAN
-		fb_swap = GRPH_ENDIAN_SWAP(GRPH_ENDIAN_8IN16);
+		fb_swap = (GRPH_ENDIAN_8IN16 << GRPH_SWAP_CNTL__GRPH_ENDIAN_SWAP__SHIFT);
 #endif
 		break;
 	case DRM_FORMAT_BGRX5551:
 	case DRM_FORMAT_BGRA5551:
-		fb_format = (GRPH_DEPTH(GRPH_DEPTH_16BPP) |
-			     GRPH_FORMAT(GRPH_FORMAT_BGRA5551));
+		fb_format = ((GRPH_DEPTH_16BPP << GRPH_CONTROL__GRPH_DEPTH__SHIFT) |
+			     (GRPH_FORMAT_BGRA5551 << GRPH_CONTROL__GRPH_FORMAT__SHIFT));
 #ifdef __BIG_ENDIAN
-		fb_swap = GRPH_ENDIAN_SWAP(GRPH_ENDIAN_8IN16);
+		fb_swap = (GRPH_ENDIAN_8IN16 << GRPH_SWAP_CNTL__GRPH_ENDIAN_SWAP__SHIFT);
 #endif
 		break;
 	case DRM_FORMAT_RGB565:
-		fb_format = (GRPH_DEPTH(GRPH_DEPTH_16BPP) |
-			     GRPH_FORMAT(GRPH_FORMAT_ARGB565));
+		fb_format = ((GRPH_DEPTH_16BPP << GRPH_CONTROL__GRPH_DEPTH__SHIFT) |
+			     (GRPH_FORMAT_ARGB565 << GRPH_CONTROL__GRPH_FORMAT__SHIFT));
 #ifdef __BIG_ENDIAN
-		fb_swap = GRPH_ENDIAN_SWAP(GRPH_ENDIAN_8IN16);
+		fb_swap = (GRPH_ENDIAN_8IN16 << GRPH_SWAP_CNTL__GRPH_ENDIAN_SWAP__SHIFT);
 #endif
 		break;
 	case DRM_FORMAT_XRGB8888:
 	case DRM_FORMAT_ARGB8888:
-		fb_format = (GRPH_DEPTH(GRPH_DEPTH_32BPP) |
-			     GRPH_FORMAT(GRPH_FORMAT_ARGB8888));
+		fb_format = ((GRPH_DEPTH_32BPP << GRPH_CONTROL__GRPH_DEPTH__SHIFT) |
+			     (GRPH_FORMAT_ARGB8888 << GRPH_CONTROL__GRPH_FORMAT__SHIFT));
 #ifdef __BIG_ENDIAN
-		fb_swap = GRPH_ENDIAN_SWAP(GRPH_ENDIAN_8IN32);
+		fb_swap = (GRPH_ENDIAN_8IN32 << GRPH_SWAP_CNTL__GRPH_ENDIAN_SWAP__SHIFT);
 #endif
 		break;
 	case DRM_FORMAT_XRGB2101010:
 	case DRM_FORMAT_ARGB2101010:
-		fb_format = (GRPH_DEPTH(GRPH_DEPTH_32BPP) |
-			     GRPH_FORMAT(GRPH_FORMAT_ARGB2101010));
+		fb_format = ((GRPH_DEPTH_32BPP << GRPH_CONTROL__GRPH_DEPTH__SHIFT) |
+			     (GRPH_FORMAT_ARGB2101010 << GRPH_CONTROL__GRPH_FORMAT__SHIFT));
 #ifdef __BIG_ENDIAN
-		fb_swap = GRPH_ENDIAN_SWAP(GRPH_ENDIAN_8IN32);
+		fb_swap = (GRPH_ENDIAN_8IN32 << GRPH_SWAP_CNTL__GRPH_ENDIAN_SWAP__SHIFT);
 #endif
 		/* Greater 8 bpc fb needs to bypass hw-lut to retain precision */
 		bypass_lut = true;
 		break;
 	case DRM_FORMAT_BGRX1010102:
 	case DRM_FORMAT_BGRA1010102:
-		fb_format = (GRPH_DEPTH(GRPH_DEPTH_32BPP) |
-			     GRPH_FORMAT(GRPH_FORMAT_BGRA1010102));
+		fb_format = ((GRPH_DEPTH_32BPP << GRPH_CONTROL__GRPH_DEPTH__SHIFT) |
+			     (GRPH_FORMAT_BGRA1010102 << GRPH_CONTROL__GRPH_FORMAT__SHIFT));
 #ifdef __BIG_ENDIAN
-		fb_swap = GRPH_ENDIAN_SWAP(GRPH_ENDIAN_8IN32);
+		fb_swap = (GRPH_ENDIAN_8IN32 << GRPH_SWAP_CNTL__GRPH_ENDIAN_SWAP__SHIFT);
 #endif
 		/* Greater 8 bpc fb needs to bypass hw-lut to retain precision */
 		bypass_lut = true;
 		break;
 	case DRM_FORMAT_XBGR8888:
 	case DRM_FORMAT_ABGR8888:
-		fb_format = (GRPH_DEPTH(GRPH_DEPTH_32BPP) |
-			     GRPH_FORMAT(GRPH_FORMAT_ARGB8888));
-		fb_swap = (GRPH_RED_CROSSBAR(GRPH_RED_SEL_B) |
-			   GRPH_BLUE_CROSSBAR(GRPH_BLUE_SEL_R));
+		fb_format = ((GRPH_DEPTH_32BPP << GRPH_CONTROL__GRPH_DEPTH__SHIFT) |
+			     (GRPH_FORMAT_ARGB8888 << GRPH_CONTROL__GRPH_FORMAT__SHIFT));
+		fb_swap = ((GRPH_RED_SEL_B << GRPH_SWAP_CNTL__GRPH_RED_CROSSBAR__SHIFT) |
+			   (GRPH_BLUE_SEL_R << GRPH_SWAP_CNTL__GRPH_BLUE_CROSSBAR__SHIFT));
 #ifdef __BIG_ENDIAN
-		fb_swap |= GRPH_ENDIAN_SWAP(GRPH_ENDIAN_8IN32);
+		fb_swap |= (GRPH_ENDIAN_8IN32 << GRPH_SWAP_CNTL__GRPH_ENDIAN_SWAP__SHIFT);
 #endif
 		break;
 	default:
@@ -1942,18 +2013,18 @@ static int dce_v6_0_crtc_do_set_base(struct drm_crtc *crtc,
 		tile_split = AMDGPU_TILING_GET(tiling_flags, TILE_SPLIT);
 		num_banks = AMDGPU_TILING_GET(tiling_flags, NUM_BANKS);
 
-		fb_format |= GRPH_NUM_BANKS(num_banks);
-		fb_format |= GRPH_ARRAY_MODE(GRPH_ARRAY_2D_TILED_THIN1);
-		fb_format |= GRPH_TILE_SPLIT(tile_split);
-		fb_format |= GRPH_BANK_WIDTH(bankw);
-		fb_format |= GRPH_BANK_HEIGHT(bankh);
-		fb_format |= GRPH_MACRO_TILE_ASPECT(mtaspect);
+		fb_format |= (num_banks << GRPH_CONTROL__GRPH_NUM_BANKS__SHIFT);
+		fb_format |= (GRPH_ARRAY_2D_TILED_THIN1 << GRPH_CONTROL__GRPH_ARRAY_MODE__SHIFT);
+		fb_format |= (tile_split << GRPH_CONTROL__GRPH_TILE_SPLIT__SHIFT);
+		fb_format |= (bankw << GRPH_CONTROL__GRPH_BANK_WIDTH__SHIFT);
+		fb_format |= (bankh << GRPH_CONTROL__GRPH_BANK_HEIGHT__SHIFT);
+		fb_format |= (mtaspect << GRPH_CONTROL__GRPH_MACRO_TILE_ASPECT__SHIFT);
 	} else if (AMDGPU_TILING_GET(tiling_flags, ARRAY_MODE) == ARRAY_1D_TILED_THIN1) {
-		fb_format |= GRPH_ARRAY_MODE(GRPH_ARRAY_1D_TILED_THIN1);
+		fb_format |= (GRPH_ARRAY_1D_TILED_THIN1 << GRPH_CONTROL__GRPH_ARRAY_MODE__SHIFT);
 	}
 
 	pipe_config = AMDGPU_TILING_GET(tiling_flags, PIPE_CONFIG);
-	fb_format |= GRPH_PIPE_CONFIG(pipe_config);
+	fb_format |= (pipe_config << GRPH_CONTROL__GRPH_PIPE_CONFIG__SHIFT);
 
 	dce_v6_0_vga_enable(crtc, false);
 
@@ -1969,7 +2040,7 @@ static int dce_v6_0_crtc_do_set_base(struct drm_crtc *crtc,
 	WREG32(mmGRPH_PRIMARY_SURFACE_ADDRESS + amdgpu_crtc->crtc_offset,
 	       (u32)fb_location & GRPH_PRIMARY_SURFACE_ADDRESS__GRPH_PRIMARY_SURFACE_ADDRESS_MASK);
 	WREG32(mmGRPH_SECONDARY_SURFACE_ADDRESS + amdgpu_crtc->crtc_offset,
-	       (u32) fb_location & GRPH_PRIMARY_SURFACE_ADDRESS__GRPH_PRIMARY_SURFACE_ADDRESS_MASK);
+	       (u32) fb_location & GRPH_SECONDARY_SURFACE_ADDRESS__GRPH_SECONDARY_SURFACE_ADDRESS_MASK);
 	WREG32(mmGRPH_CONTROL + amdgpu_crtc->crtc_offset, fb_format);
 	WREG32(mmGRPH_SWAP_CNTL + amdgpu_crtc->crtc_offset, fb_swap);
 
@@ -2037,14 +2108,13 @@ static void dce_v6_0_set_interleave(struct drm_crtc *crtc,
 
 	if (mode->flags & DRM_MODE_FLAG_INTERLACE)
 		WREG32(mmDATA_FORMAT + amdgpu_crtc->crtc_offset,
-		       INTERLEAVE_EN);
+			DATA_FORMAT__INTERLEAVE_EN_MASK);
 	else
 		WREG32(mmDATA_FORMAT + amdgpu_crtc->crtc_offset, 0);
 }
 
 static void dce_v6_0_crtc_load_lut(struct drm_crtc *crtc)
 {
-
 	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
 	struct drm_device *dev = crtc->dev;
 	struct amdgpu_device *adev = drm_to_adev(dev);
@@ -2054,15 +2124,15 @@ static void dce_v6_0_crtc_load_lut(struct drm_crtc *crtc)
 	DRM_DEBUG_KMS("%d\n", amdgpu_crtc->crtc_id);
 
 	WREG32(mmINPUT_CSC_CONTROL + amdgpu_crtc->crtc_offset,
-	       ((0 << INPUT_CSC_CONTROL__INPUT_CSC_GRPH_MODE__SHIFT) |
-		(0 << INPUT_CSC_CONTROL__INPUT_CSC_OVL_MODE__SHIFT)));
+	       ((INPUT_CSC_BYPASS << INPUT_CSC_CONTROL__INPUT_CSC_GRPH_MODE__SHIFT) |
+		(INPUT_CSC_BYPASS << INPUT_CSC_CONTROL__INPUT_CSC_OVL_MODE__SHIFT)));
 	WREG32(mmPRESCALE_GRPH_CONTROL + amdgpu_crtc->crtc_offset,
 	       PRESCALE_GRPH_CONTROL__GRPH_PRESCALE_BYPASS_MASK);
 	WREG32(mmPRESCALE_OVL_CONTROL + amdgpu_crtc->crtc_offset,
 	       PRESCALE_OVL_CONTROL__OVL_PRESCALE_BYPASS_MASK);
 	WREG32(mmINPUT_GAMMA_CONTROL + amdgpu_crtc->crtc_offset,
-	       ((0 << INPUT_GAMMA_CONTROL__GRPH_INPUT_GAMMA_MODE__SHIFT) |
-		(0 << INPUT_GAMMA_CONTROL__OVL_INPUT_GAMMA_MODE__SHIFT)));
+	       ((INPUT_GAMMA_USE_LUT << INPUT_GAMMA_CONTROL__GRPH_INPUT_GAMMA_MODE__SHIFT) |
+		(INPUT_GAMMA_USE_LUT << INPUT_GAMMA_CONTROL__OVL_INPUT_GAMMA_MODE__SHIFT)));
 
 	WREG32(mmDC_LUT_CONTROL + amdgpu_crtc->crtc_offset, 0);
 
@@ -2089,19 +2159,19 @@ static void dce_v6_0_crtc_load_lut(struct drm_crtc *crtc)
 	}
 
 	WREG32(mmDEGAMMA_CONTROL + amdgpu_crtc->crtc_offset,
-	       ((0 << DEGAMMA_CONTROL__GRPH_DEGAMMA_MODE__SHIFT) |
-		(0 << DEGAMMA_CONTROL__OVL_DEGAMMA_MODE__SHIFT) |
-		ICON_DEGAMMA_MODE(0) |
-		(0 << DEGAMMA_CONTROL__CURSOR_DEGAMMA_MODE__SHIFT)));
+	       ((DEGAMMA_BYPASS << DEGAMMA_CONTROL__GRPH_DEGAMMA_MODE__SHIFT) |
+		(DEGAMMA_BYPASS << DEGAMMA_CONTROL__OVL_DEGAMMA_MODE__SHIFT) |
+		(DEGAMMA_BYPASS << DEGAMMA_CONTROL__ICON_DEGAMMA_MODE__SHIFT) |
+		(DEGAMMA_BYPASS << DEGAMMA_CONTROL__CURSOR_DEGAMMA_MODE__SHIFT)));
 	WREG32(mmGAMUT_REMAP_CONTROL + amdgpu_crtc->crtc_offset,
-	       ((0 << GAMUT_REMAP_CONTROL__GRPH_GAMUT_REMAP_MODE__SHIFT) |
-		(0 << GAMUT_REMAP_CONTROL__OVL_GAMUT_REMAP_MODE__SHIFT)));
+	       ((GAMUT_REMAP_BYPASS << GAMUT_REMAP_CONTROL__GRPH_GAMUT_REMAP_MODE__SHIFT) |
+		(GAMUT_REMAP_BYPASS << GAMUT_REMAP_CONTROL__OVL_GAMUT_REMAP_MODE__SHIFT)));
 	WREG32(mmREGAMMA_CONTROL + amdgpu_crtc->crtc_offset,
-	       ((0 << REGAMMA_CONTROL__GRPH_REGAMMA_MODE__SHIFT) |
-		(0 << REGAMMA_CONTROL__OVL_REGAMMA_MODE__SHIFT)));
+	       ((REGAMMA_BYPASS << REGAMMA_CONTROL__GRPH_REGAMMA_MODE__SHIFT) |
+		(REGAMMA_BYPASS << REGAMMA_CONTROL__OVL_REGAMMA_MODE__SHIFT)));
 	WREG32(mmOUTPUT_CSC_CONTROL + amdgpu_crtc->crtc_offset,
-	       ((0 << OUTPUT_CSC_CONTROL__OUTPUT_CSC_GRPH_MODE__SHIFT) |
-		(0 << OUTPUT_CSC_CONTROL__OUTPUT_CSC_OVL_MODE__SHIFT)));
+	       ((OUTPUT_CSC_BYPASS << OUTPUT_CSC_CONTROL__OUTPUT_CSC_GRPH_MODE__SHIFT) |
+		(OUTPUT_CSC_BYPASS << OUTPUT_CSC_CONTROL__OUTPUT_CSC_OVL_MODE__SHIFT)));
 	/* XXX match this to the depth of the crtc fmt block, move to modeset? */
 	WREG32(0x1a50 + amdgpu_crtc->crtc_offset, 0);
 
@@ -2196,8 +2266,6 @@ static void dce_v6_0_hide_cursor(struct drm_crtc *crtc)
 	WREG32(mmCUR_CONTROL + amdgpu_crtc->crtc_offset,
 	       (CURSOR_24_8_PRE_MULT << CUR_CONTROL__CURSOR_MODE__SHIFT) |
 	       (CURSOR_URGENT_1_2 << CUR_CONTROL__CURSOR_URGENT_CONTROL__SHIFT));
-
-
 }
 
 static void dce_v6_0_show_cursor(struct drm_crtc *crtc)
@@ -2214,7 +2282,6 @@ static void dce_v6_0_show_cursor(struct drm_crtc *crtc)
 	       CUR_CONTROL__CURSOR_EN_MASK |
 	       (CURSOR_24_8_PRE_MULT << CUR_CONTROL__CURSOR_MODE__SHIFT) |
 	       (CURSOR_URGENT_1_2 << CUR_CONTROL__CURSOR_URGENT_CONTROL__SHIFT));
-
 }
 
 static int dce_v6_0_cursor_move_locked(struct drm_crtc *crtc,
@@ -2302,6 +2369,7 @@ static int dce_v6_0_crtc_cursor_set2(struct drm_crtc *crtc,
 		return ret;
 	}
 
+	aobj->flags |= AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS;
 	ret = amdgpu_bo_pin(aobj, AMDGPU_GEM_DOMAIN_VRAM);
 	amdgpu_bo_unreserve(aobj);
 	if (ret) {
@@ -2424,7 +2492,7 @@ static void dce_v6_0_crtc_dpms(struct drm_crtc *crtc, int mode)
 		break;
 	}
 	/* adjust pm to dpms */
-	amdgpu_pm_compute_clocks(adev);
+	amdgpu_dpm_compute_clocks(adev);
 }
 
 static void dce_v6_0_crtc_prepare(struct drm_crtc *crtc)
@@ -2524,7 +2592,6 @@ static bool dce_v6_0_crtc_mode_fixup(struct drm_crtc *crtc,
 				     const struct drm_display_mode *mode,
 				     struct drm_display_mode *adjusted_mode)
 {
-
 	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
 	struct drm_device *dev = crtc->dev;
 	struct drm_encoder *encoder;
@@ -2581,6 +2648,32 @@ static const struct drm_crtc_helper_funcs dce_v6_0_crtc_helper_funcs = {
 	.get_scanout_position = amdgpu_crtc_get_scanout_position,
 };
 
+static void dce_v6_0_panic_flush(struct drm_plane *plane)
+{
+	struct drm_framebuffer *fb;
+	struct amdgpu_crtc *amdgpu_crtc;
+	struct amdgpu_device *adev;
+	uint32_t fb_format;
+
+	if (!plane->fb)
+		return;
+
+	fb = plane->fb;
+	amdgpu_crtc = to_amdgpu_crtc(plane->crtc);
+	adev = drm_to_adev(fb->dev);
+
+	/* Disable DC tiling */
+	fb_format = RREG32(mmGRPH_CONTROL + amdgpu_crtc->crtc_offset);
+	fb_format &= ~GRPH_CONTROL__GRPH_ARRAY_MODE_MASK;
+	WREG32(mmGRPH_CONTROL + amdgpu_crtc->crtc_offset, fb_format);
+
+}
+
+static const struct drm_plane_helper_funcs dce_v6_0_drm_primary_plane_helper_funcs = {
+	.get_scanout_buffer = amdgpu_display_get_scanout_buffer,
+	.panic_flush = dce_v6_0_panic_flush,
+};
+
 static int dce_v6_0_crtc_init(struct amdgpu_device *adev, int index)
 {
 	struct amdgpu_crtc *amdgpu_crtc;
@@ -2608,13 +2701,14 @@ static int dce_v6_0_crtc_init(struct amdgpu_device *adev, int index)
 	amdgpu_crtc->encoder = NULL;
 	amdgpu_crtc->connector = NULL;
 	drm_crtc_helper_add(&amdgpu_crtc->base, &dce_v6_0_crtc_helper_funcs);
+	drm_plane_helper_add(amdgpu_crtc->base.primary, &dce_v6_0_drm_primary_plane_helper_funcs);
 
 	return 0;
 }
 
-static int dce_v6_0_early_init(void *handle)
+static int dce_v6_0_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	adev->audio_endpt_rreg = &dce_v6_0_audio_endpt_rreg;
 	adev->audio_endpt_wreg = &dce_v6_0_audio_endpt_wreg;
@@ -2643,11 +2737,10 @@ static int dce_v6_0_early_init(void *handle)
 	return 0;
 }
 
-static int dce_v6_0_sw_init(void *handle)
+static int dce_v6_0_sw_init(struct amdgpu_ip_block *ip_block)
 {
 	int r, i;
-	bool ret;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	for (i = 0; i < adev->mode_info.num_crtc; i++) {
 		r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, i + 1, &adev->crtc_irq);
@@ -2674,7 +2767,7 @@ static int dce_v6_0_sw_init(void *handle)
 	adev_to_drm(adev)->mode_config.max_height = 16384;
 	adev_to_drm(adev)->mode_config.preferred_depth = 24;
 	adev_to_drm(adev)->mode_config.prefer_shadow = 1;
-	adev_to_drm(adev)->mode_config.fb_base = adev->gmc.aper_base;
+	adev_to_drm(adev)->mode_config.fb_modifiers_not_supported = true;
 
 	r = amdgpu_display_modeset_create_props(adev);
 	if (r)
@@ -2690,8 +2783,7 @@ static int dce_v6_0_sw_init(void *handle)
 			return r;
 	}
 
-	ret = amdgpu_atombios_get_connector_info_from_object_table(adev);
-	if (ret)
+	if (amdgpu_atombios_get_connector_info_from_object_table(adev))
 		amdgpu_display_print_display_setup(adev_to_drm(adev));
 	else
 		return -EINVAL;
@@ -2705,16 +2797,28 @@ static int dce_v6_0_sw_init(void *handle)
 	if (r)
 		return r;
 
+	/* Disable vblank IRQs aggressively for power-saving */
+	/* XXX: can this be enabled for DC? */
+	adev_to_drm(adev)->vblank_disable_immediate = true;
+
+	r = drm_vblank_init(adev_to_drm(adev), adev->mode_info.num_crtc);
+	if (r)
+		return r;
+
+	/* Pre-DCE11 */
+	INIT_DELAYED_WORK(&adev->hotplug_work,
+		  amdgpu_display_hotplug_work_func);
+
 	drm_kms_helper_poll_init(adev_to_drm(adev));
 
 	return r;
 }
 
-static int dce_v6_0_sw_fini(void *handle)
+static int dce_v6_0_sw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
-	kfree(adev->mode_info.bios_hardcoded_edid);
+	drm_edid_free(adev->mode_info.bios_hardcoded_edid);
 
 	drm_kms_helper_poll_fini(adev_to_drm(adev));
 
@@ -2727,10 +2831,10 @@ static int dce_v6_0_sw_fini(void *handle)
 	return 0;
 }
 
-static int dce_v6_0_hw_init(void *handle)
+static int dce_v6_0_hw_init(struct amdgpu_ip_block *ip_block)
 {
 	int i;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	/* disable vga render */
 	dce_v6_0_set_vga_render_state(adev, false);
@@ -2750,10 +2854,10 @@ static int dce_v6_0_hw_init(void *handle)
 	return 0;
 }
 
-static int dce_v6_0_hw_fini(void *handle)
+static int dce_v6_0_hw_fini(struct amdgpu_ip_block *ip_block)
 {
 	int i;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	dce_v6_0_hpd_fini(adev);
 
@@ -2763,12 +2867,14 @@ static int dce_v6_0_hw_fini(void *handle)
 
 	dce_v6_0_pageflip_interrupt_fini(adev);
 
+	flush_delayed_work(&adev->hotplug_work);
+
 	return 0;
 }
 
-static int dce_v6_0_suspend(void *handle)
+static int dce_v6_0_suspend(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int r;
 
 	r = amdgpu_display_suspend_helper(adev);
@@ -2777,18 +2883,18 @@ static int dce_v6_0_suspend(void *handle)
 	adev->mode_info.bl_level =
 		amdgpu_atombios_encoder_get_backlight_level_from_reg(adev);
 
-	return dce_v6_0_hw_fini(handle);
+	return dce_v6_0_hw_fini(ip_block);
 }
 
-static int dce_v6_0_resume(void *handle)
+static int dce_v6_0_resume(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int ret;
 
 	amdgpu_atombios_encoder_set_backlight_level_to_reg(adev,
 							   adev->mode_info.bl_level);
 
-	ret = dce_v6_0_hw_init(handle);
+	ret = dce_v6_0_hw_init(ip_block);
 
 	/* turn on the BL */
 	if (adev->mode_info.bl_encoder) {
@@ -2803,19 +2909,35 @@ static int dce_v6_0_resume(void *handle)
 	return amdgpu_display_resume_helper(adev);
 }
 
-static bool dce_v6_0_is_idle(void *handle)
+static bool dce_v6_0_is_idle(struct amdgpu_ip_block *ip_block)
 {
 	return true;
 }
 
-static int dce_v6_0_wait_for_idle(void *handle)
+static int dce_v6_0_soft_reset(struct amdgpu_ip_block *ip_block)
 {
-	return 0;
-}
+	u32 srbm_soft_reset = 0, tmp;
+	struct amdgpu_device *adev = ip_block->adev;
 
-static int dce_v6_0_soft_reset(void *handle)
-{
-	DRM_INFO("xxxx: dce_v6_0_soft_reset --- no impl!!\n");
+	if (dce_v6_0_is_display_hung(adev))
+		srbm_soft_reset |= SRBM_SOFT_RESET__SOFT_RESET_DC_MASK;
+
+	if (srbm_soft_reset) {
+		tmp = RREG32(mmSRBM_SOFT_RESET);
+		tmp |= srbm_soft_reset;
+		dev_info(adev->dev, "SRBM_SOFT_RESET=0x%08X\n", tmp);
+		WREG32(mmSRBM_SOFT_RESET, tmp);
+		tmp = RREG32(mmSRBM_SOFT_RESET);
+
+		udelay(50);
+
+		tmp &= ~srbm_soft_reset;
+		WREG32(mmSRBM_SOFT_RESET, tmp);
+		tmp = RREG32(mmSRBM_SOFT_RESET);
+
+		/* Wait a little for things to settle down */
+		udelay(50);
+	}
 	return 0;
 }
 
@@ -2832,22 +2954,22 @@ static void dce_v6_0_set_crtc_vblank_interrupt_state(struct amdgpu_device *adev,
 
 	switch (crtc) {
 	case 0:
-		reg_block = SI_CRTC0_REGISTER_OFFSET;
+		reg_block = CRTC0_REGISTER_OFFSET;
 		break;
 	case 1:
-		reg_block = SI_CRTC1_REGISTER_OFFSET;
+		reg_block = CRTC1_REGISTER_OFFSET;
 		break;
 	case 2:
-		reg_block = SI_CRTC2_REGISTER_OFFSET;
+		reg_block = CRTC2_REGISTER_OFFSET;
 		break;
 	case 3:
-		reg_block = SI_CRTC3_REGISTER_OFFSET;
+		reg_block = CRTC3_REGISTER_OFFSET;
 		break;
 	case 4:
-		reg_block = SI_CRTC4_REGISTER_OFFSET;
+		reg_block = CRTC4_REGISTER_OFFSET;
 		break;
 	case 5:
-		reg_block = SI_CRTC5_REGISTER_OFFSET;
+		reg_block = CRTC5_REGISTER_OFFSET;
 		break;
 	default:
 		DRM_DEBUG("invalid crtc %d\n", crtc);
@@ -2857,12 +2979,12 @@ static void dce_v6_0_set_crtc_vblank_interrupt_state(struct amdgpu_device *adev,
 	switch (state) {
 	case AMDGPU_IRQ_STATE_DISABLE:
 		interrupt_mask = RREG32(mmINT_MASK + reg_block);
-		interrupt_mask &= ~VBLANK_INT_MASK;
+		interrupt_mask &= ~INT_MASK__VBLANK_INT_MASK;
 		WREG32(mmINT_MASK + reg_block, interrupt_mask);
 		break;
 	case AMDGPU_IRQ_STATE_ENABLE:
 		interrupt_mask = RREG32(mmINT_MASK + reg_block);
-		interrupt_mask |= VBLANK_INT_MASK;
+		interrupt_mask |= INT_MASK__VBLANK_INT_MASK;
 		WREG32(mmINT_MASK + reg_block, interrupt_mask);
 		break;
 	default:
@@ -2877,28 +2999,28 @@ static void dce_v6_0_set_crtc_vline_interrupt_state(struct amdgpu_device *adev,
 
 }
 
-static int dce_v6_0_set_hpd_interrupt_state(struct amdgpu_device *adev,
+static int dce_v6_0_set_hpd_irq_state(struct amdgpu_device *adev,
 					    struct amdgpu_irq_src *src,
-					    unsigned type,
+					    unsigned hpd,
 					    enum amdgpu_interrupt_state state)
 {
 	u32 dc_hpd_int_cntl;
 
-	if (type >= adev->mode_info.num_hpd) {
-		DRM_DEBUG("invalid hdp %d\n", type);
+	if (hpd >= adev->mode_info.num_hpd) {
+		DRM_DEBUG("invalid hpd %d\n", hpd);
 		return 0;
 	}
 
 	switch (state) {
 	case AMDGPU_IRQ_STATE_DISABLE:
-		dc_hpd_int_cntl = RREG32(mmDC_HPD1_INT_CONTROL + hpd_offsets[type]);
-		dc_hpd_int_cntl &= ~DC_HPDx_INT_EN;
-		WREG32(mmDC_HPD1_INT_CONTROL + hpd_offsets[type], dc_hpd_int_cntl);
+		dc_hpd_int_cntl = RREG32(mmDC_HPD1_INT_CONTROL + hpd_offsets[hpd]);
+		dc_hpd_int_cntl &= ~DC_HPD1_INT_CONTROL__DC_HPD1_INT_EN_MASK;
+		WREG32(mmDC_HPD1_INT_CONTROL + hpd_offsets[hpd], dc_hpd_int_cntl);
 		break;
 	case AMDGPU_IRQ_STATE_ENABLE:
-		dc_hpd_int_cntl = RREG32(mmDC_HPD1_INT_CONTROL + hpd_offsets[type]);
-		dc_hpd_int_cntl |= DC_HPDx_INT_EN;
-		WREG32(mmDC_HPD1_INT_CONTROL + hpd_offsets[type], dc_hpd_int_cntl);
+		dc_hpd_int_cntl = RREG32(mmDC_HPD1_INT_CONTROL + hpd_offsets[hpd]);
+		dc_hpd_int_cntl |= DC_HPD1_INT_CONTROL__DC_HPD1_INT_EN_MASK;
+		WREG32(mmDC_HPD1_INT_CONTROL + hpd_offsets[hpd], dc_hpd_int_cntl);
 		break;
 	default:
 		break;
@@ -2907,7 +3029,7 @@ static int dce_v6_0_set_hpd_interrupt_state(struct amdgpu_device *adev,
 	return 0;
 }
 
-static int dce_v6_0_set_crtc_interrupt_state(struct amdgpu_device *adev,
+static int dce_v6_0_set_crtc_irq_state(struct amdgpu_device *adev,
 					     struct amdgpu_irq_src *src,
 					     unsigned type,
 					     enum amdgpu_interrupt_state state)
@@ -2967,7 +3089,7 @@ static int dce_v6_0_crtc_irq(struct amdgpu_device *adev,
 	switch (entry->src_data[0]) {
 	case 0: /* vblank */
 		if (disp_int & interrupt_status_offsets[crtc].vblank)
-			WREG32(mmVBLANK_STATUS + crtc_offsets[crtc], VBLANK_ACK);
+			WREG32(mmVBLANK_STATUS + crtc_offsets[crtc], VBLANK_STATUS__VBLANK_ACK_MASK);
 		else
 			DRM_DEBUG("IH: IH event w/o asserted irq bit?\n");
 
@@ -2978,7 +3100,7 @@ static int dce_v6_0_crtc_irq(struct amdgpu_device *adev,
 		break;
 	case 1: /* vline */
 		if (disp_int & interrupt_status_offsets[crtc].vline)
-			WREG32(mmVLINE_STATUS + crtc_offsets[crtc], VLINE_ACK);
+			WREG32(mmVLINE_STATUS + crtc_offsets[crtc], VLINE_STATUS__VLINE_ACK_MASK);
 		else
 			DRM_DEBUG("IH: IH event w/o asserted irq bit?\n");
 
@@ -2992,7 +3114,7 @@ static int dce_v6_0_crtc_irq(struct amdgpu_device *adev,
 	return 0;
 }
 
-static int dce_v6_0_set_pageflip_interrupt_state(struct amdgpu_device *adev,
+static int dce_v6_0_set_pageflip_irq_state(struct amdgpu_device *adev,
 						 struct amdgpu_irq_src *src,
 						 unsigned type,
 						 enum amdgpu_interrupt_state state)
@@ -3043,7 +3165,7 @@ static int dce_v6_0_pageflip_irq(struct amdgpu_device *adev,
 
 	spin_lock_irqsave(&adev_to_drm(adev)->event_lock, flags);
 	works = amdgpu_crtc->pflip_works;
-	if (amdgpu_crtc->pflip_status != AMDGPU_FLIP_SUBMITTED){
+	if (amdgpu_crtc->pflip_status != AMDGPU_FLIP_SUBMITTED) {
 		DRM_DEBUG_DRIVER("amdgpu_crtc->pflip_status = %d != "
 						"AMDGPU_FLIP_SUBMITTED(%d)\n",
 						amdgpu_crtc->pflip_status,
@@ -3072,7 +3194,7 @@ static int dce_v6_0_hpd_irq(struct amdgpu_device *adev,
 			    struct amdgpu_irq_src *source,
 			    struct amdgpu_iv_entry *entry)
 {
-	uint32_t disp_int, mask, tmp;
+	uint32_t disp_int, mask;
 	unsigned hpd;
 
 	if (entry->src_data[0] >= adev->mode_info.num_hpd) {
@@ -3085,24 +3207,21 @@ static int dce_v6_0_hpd_irq(struct amdgpu_device *adev,
 	mask = interrupt_status_offsets[hpd].hpd;
 
 	if (disp_int & mask) {
-		tmp = RREG32(mmDC_HPD1_INT_CONTROL + hpd_offsets[hpd]);
-		tmp |= DC_HPD1_INT_CONTROL__DC_HPD1_INT_ACK_MASK;
-		WREG32(mmDC_HPD1_INT_CONTROL + hpd_offsets[hpd], tmp);
-		schedule_work(&adev->hotplug_work);
+		dce_v6_0_hpd_int_ack(adev, hpd);
+		schedule_delayed_work(&adev->hotplug_work, 0);
 		DRM_DEBUG("IH: HPD%d\n", hpd + 1);
 	}
 
 	return 0;
-
 }
 
-static int dce_v6_0_set_clockgating_state(void *handle,
+static int dce_v6_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_clockgating_state state)
 {
 	return 0;
 }
 
-static int dce_v6_0_set_powergating_state(void *handle,
+static int dce_v6_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_powergating_state state)
 {
 	return 0;
@@ -3111,7 +3230,6 @@ static int dce_v6_0_set_powergating_state(void *handle,
 static const struct amd_ip_funcs dce_v6_0_ip_funcs = {
 	.name = "dce_v6_0",
 	.early_init = dce_v6_0_early_init,
-	.late_init = NULL,
 	.sw_init = dce_v6_0_sw_init,
 	.sw_fini = dce_v6_0_sw_fini,
 	.hw_init = dce_v6_0_hw_init,
@@ -3119,18 +3237,15 @@ static const struct amd_ip_funcs dce_v6_0_ip_funcs = {
 	.suspend = dce_v6_0_suspend,
 	.resume = dce_v6_0_resume,
 	.is_idle = dce_v6_0_is_idle,
-	.wait_for_idle = dce_v6_0_wait_for_idle,
 	.soft_reset = dce_v6_0_soft_reset,
 	.set_clockgating_state = dce_v6_0_set_clockgating_state,
 	.set_powergating_state = dce_v6_0_set_powergating_state,
 };
 
-static void
-dce_v6_0_encoder_mode_set(struct drm_encoder *encoder,
+static void dce_v6_0_encoder_mode_set(struct drm_encoder *encoder,
 			  struct drm_display_mode *mode,
 			  struct drm_display_mode *adjusted_mode)
 {
-
 	struct amdgpu_encoder *amdgpu_encoder = to_amdgpu_encoder(encoder);
 	int em = amdgpu_atombios_encoder_get_encoder_mode(encoder);
 
@@ -3150,7 +3265,6 @@ dce_v6_0_encoder_mode_set(struct drm_encoder *encoder,
 
 static void dce_v6_0_encoder_prepare(struct drm_encoder *encoder)
 {
-
 	struct amdgpu_device *adev = drm_to_adev(encoder->dev);
 	struct amdgpu_encoder *amdgpu_encoder = to_amdgpu_encoder(encoder);
 	struct drm_connector *connector = amdgpu_get_connector_for_encoder(encoder);
@@ -3190,7 +3304,6 @@ static void dce_v6_0_encoder_prepare(struct drm_encoder *encoder)
 
 static void dce_v6_0_encoder_commit(struct drm_encoder *encoder)
 {
-
 	struct drm_device *dev = encoder->dev;
 	struct amdgpu_device *adev = drm_to_adev(dev);
 
@@ -3201,7 +3314,6 @@ static void dce_v6_0_encoder_commit(struct drm_encoder *encoder)
 
 static void dce_v6_0_encoder_disable(struct drm_encoder *encoder)
 {
-
 	struct amdgpu_encoder *amdgpu_encoder = to_amdgpu_encoder(encoder);
 	struct amdgpu_encoder_atom_dig *dig;
 	int em = amdgpu_atombios_encoder_get_encoder_mode(encoder);
@@ -3228,8 +3340,7 @@ static void dce_v6_0_ext_commit(struct drm_encoder *encoder)
 
 }
 
-static void
-dce_v6_0_ext_mode_set(struct drm_encoder *encoder,
+static void dce_v6_0_ext_mode_set(struct drm_encoder *encoder,
 		      struct drm_display_mode *mode,
 		      struct drm_display_mode *adjusted_mode)
 {
@@ -3241,8 +3352,7 @@ static void dce_v6_0_ext_disable(struct drm_encoder *encoder)
 
 }
 
-static void
-dce_v6_0_ext_dpms(struct drm_encoder *encoder, int mode)
+static void dce_v6_0_ext_dpms(struct drm_encoder *encoder, int mode)
 {
 
 }
@@ -3313,7 +3423,6 @@ static void dce_v6_0_encoder_add(struct amdgpu_device *adev,
 			amdgpu_encoder->devices |= supported_device;
 			return;
 		}
-
 	}
 
 	/* add a new one */
@@ -3420,17 +3529,17 @@ static void dce_v6_0_set_display_funcs(struct amdgpu_device *adev)
 }
 
 static const struct amdgpu_irq_src_funcs dce_v6_0_crtc_irq_funcs = {
-	.set = dce_v6_0_set_crtc_interrupt_state,
+	.set = dce_v6_0_set_crtc_irq_state,
 	.process = dce_v6_0_crtc_irq,
 };
 
 static const struct amdgpu_irq_src_funcs dce_v6_0_pageflip_irq_funcs = {
-	.set = dce_v6_0_set_pageflip_interrupt_state,
+	.set = dce_v6_0_set_pageflip_irq_state,
 	.process = dce_v6_0_pageflip_irq,
 };
 
 static const struct amdgpu_irq_src_funcs dce_v6_0_hpd_irq_funcs = {
-	.set = dce_v6_0_set_hpd_interrupt_state,
+	.set = dce_v6_0_set_hpd_irq_state,
 	.process = dce_v6_0_hpd_irq,
 };
 
diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c
index b200b9e722d9..2ccd6aad8dd6 100644
--- a/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c
@@ -21,7 +21,10 @@
  *
  */
 
+#include <drm/drm_edid.h>
 #include <drm/drm_fourcc.h>
+#include <drm/drm_modeset_helper.h>
+#include <drm/drm_modeset_helper_vtables.h>
 #include <drm/drm_vblank.h>
 
 #include "amdgpu.h"
@@ -51,8 +54,7 @@
 static void dce_v8_0_set_display_funcs(struct amdgpu_device *adev);
 static void dce_v8_0_set_irq_funcs(struct amdgpu_device *adev);
 
-static const u32 crtc_offsets[6] =
-{
+static const u32 crtc_offsets[6] = {
 	CRTC0_REGISTER_OFFSET,
 	CRTC1_REGISTER_OFFSET,
 	CRTC2_REGISTER_OFFSET,
@@ -61,8 +63,7 @@ static const u32 crtc_offsets[6] =
 	CRTC5_REGISTER_OFFSET
 };
 
-static const u32 hpd_offsets[] =
-{
+static const u32 hpd_offsets[] = {
 	HPD0_REGISTER_OFFSET,
 	HPD1_REGISTER_OFFSET,
 	HPD2_REGISTER_OFFSET,
@@ -264,6 +265,21 @@ static void dce_v8_0_hpd_set_polarity(struct amdgpu_device *adev,
 	WREG32(mmDC_HPD1_INT_CONTROL + hpd_offsets[hpd], tmp);
 }
 
+static void dce_v8_0_hpd_int_ack(struct amdgpu_device *adev,
+				 int hpd)
+{
+	u32 tmp;
+
+	if (hpd >= adev->mode_info.num_hpd) {
+		DRM_DEBUG("invalid hpd %d\n", hpd);
+		return;
+	}
+
+	tmp = RREG32(mmDC_HPD1_INT_CONTROL + hpd_offsets[hpd]);
+	tmp |= DC_HPD1_INT_CONTROL__DC_HPD1_INT_ACK_MASK;
+	WREG32(mmDC_HPD1_INT_CONTROL + hpd_offsets[hpd], tmp);
+}
+
 /**
  * dce_v8_0_hpd_init - hpd setup callback.
  *
@@ -303,6 +319,7 @@ static void dce_v8_0_hpd_init(struct amdgpu_device *adev)
 			continue;
 		}
 
+		dce_v8_0_hpd_int_ack(adev, amdgpu_connector->hpd.hpd);
 		dce_v8_0_hpd_set_polarity(adev, amdgpu_connector->hpd.hpd);
 		amdgpu_irq_get(adev, &adev->hpd_irq, amdgpu_connector->hpd.hpd);
 	}
@@ -333,7 +350,7 @@ static void dce_v8_0_hpd_fini(struct amdgpu_device *adev)
 
 		tmp = RREG32(mmDC_HPD1_CONTROL + hpd_offsets[amdgpu_connector->hpd.hpd]);
 		tmp &= ~DC_HPD1_CONTROL__DC_HPD1_EN_MASK;
-		WREG32(mmDC_HPD1_CONTROL + hpd_offsets[amdgpu_connector->hpd.hpd], 0);
+		WREG32(mmDC_HPD1_CONTROL + hpd_offsets[amdgpu_connector->hpd.hpd], tmp);
 
 		amdgpu_irq_put(adev, &adev->hpd_irq, amdgpu_connector->hpd.hpd);
 	}
@@ -975,7 +992,7 @@ static void dce_v8_0_program_watermarks(struct amdgpu_device *adev,
 					    (u32)mode->clock);
 		line_time = (u32) div_u64((u64)mode->crtc_htotal * 1000000,
 					  (u32)mode->clock);
-		line_time = min(line_time, (u32)65535);
+		line_time = min_t(u32, line_time, 65535);
 
 		/* watermark for high clocks */
 		if (adev->pm.dpm_enabled) {
@@ -1005,7 +1022,7 @@ static void dce_v8_0_program_watermarks(struct amdgpu_device *adev,
 		wm_high.num_heads = num_heads;
 
 		/* set for high clocks */
-		latency_watermark_a = min(dce_v8_0_latency_watermark(&wm_high), (u32)65535);
+		latency_watermark_a = min_t(u32, dce_v8_0_latency_watermark(&wm_high), 65535);
 
 		/* possibly force display priority to high */
 		/* should really do this at mode validation time... */
@@ -1044,7 +1061,7 @@ static void dce_v8_0_program_watermarks(struct amdgpu_device *adev,
 		wm_low.num_heads = num_heads;
 
 		/* set for low clocks */
-		latency_watermark_b = min(dce_v8_0_latency_watermark(&wm_low), (u32)65535);
+		latency_watermark_b = min_t(u32, dce_v8_0_latency_watermark(&wm_low), 65535);
 
 		/* possibly force display priority to high */
 		/* should really do this at mode validation time... */
@@ -1079,8 +1096,7 @@ static void dce_v8_0_program_watermarks(struct amdgpu_device *adev,
 
 	/* save values for DPM */
 	amdgpu_crtc->line_time = line_time;
-	amdgpu_crtc->wm_high = latency_watermark_a;
-	amdgpu_crtc->wm_low = latency_watermark_b;
+
 	/* Save number of lines the linebuffer leads before the scanout */
 	amdgpu_crtc->lb_vblank_lead_lines = lb_vblank_lead_lines;
 }
@@ -1255,7 +1271,7 @@ static void dce_v8_0_audio_write_speaker_allocation(struct drm_encoder *encoder)
 		return;
 	}
 
-	sad_count = drm_edid_to_speaker_allocation(amdgpu_connector_edid(connector), &sadb);
+	sad_count = drm_edid_to_speaker_allocation(amdgpu_connector->edid, &sadb);
 	if (sad_count < 0) {
 		DRM_ERROR("Couldn't read Speaker Allocation Data Block: %d\n", sad_count);
 		sad_count = 0;
@@ -1323,7 +1339,7 @@ static void dce_v8_0_audio_write_sad_regs(struct drm_encoder *encoder)
 		return;
 	}
 
-	sad_count = drm_edid_to_sad(amdgpu_connector_edid(connector), &sads);
+	sad_count = drm_edid_to_sad(amdgpu_connector->edid, &sads);
 	if (sad_count < 0)
 		DRM_ERROR("Couldn't read SADs: %d\n", sad_count);
 	if (sad_count <= 0)
@@ -1343,9 +1359,9 @@ static void dce_v8_0_audio_write_sad_regs(struct drm_encoder *encoder)
 				if (sad->channels > max_channels) {
 					value = (sad->channels <<
 						 AZALIA_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR0__MAX_CHANNELS__SHIFT) |
-					        (sad->byte2 <<
+						(sad->byte2 <<
 						 AZALIA_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR0__DESCRIPTOR_BYTE_2__SHIFT) |
-					        (sad->freq <<
+						(sad->freq <<
 						 AZALIA_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR0__SUPPORTED_FREQUENCIES__SHIFT);
 					max_channels = sad->channels;
 				}
@@ -1377,15 +1393,14 @@ static void dce_v8_0_audio_enable(struct amdgpu_device *adev,
 		enable ? AZALIA_F0_CODEC_PIN_CONTROL_HOT_PLUG_CONTROL__AUDIO_ENABLED_MASK : 0);
 }
 
-static const u32 pin_offsets[7] =
-{
-	(0x1780 - 0x1780),
-	(0x1786 - 0x1780),
-	(0x178c - 0x1780),
-	(0x1792 - 0x1780),
-	(0x1798 - 0x1780),
-	(0x179d - 0x1780),
-	(0x17a4 - 0x1780),
+static const u32 pin_offsets[7] = {
+	AUD0_REGISTER_OFFSET,
+	AUD1_REGISTER_OFFSET,
+	AUD2_REGISTER_OFFSET,
+	AUD3_REGISTER_OFFSET,
+	AUD4_REGISTER_OFFSET,
+	AUD5_REGISTER_OFFSET,
+	AUD6_REGISTER_OFFSET,
 };
 
 static int dce_v8_0_audio_init(struct amdgpu_device *adev)
@@ -1427,17 +1442,12 @@ static int dce_v8_0_audio_init(struct amdgpu_device *adev)
 
 static void dce_v8_0_audio_fini(struct amdgpu_device *adev)
 {
-	int i;
-
 	if (!amdgpu_audio)
 		return;
 
 	if (!adev->mode_info.audio.enabled)
 		return;
 
-	for (i = 0; i < adev->mode_info.audio.num_pins; i++)
-		dce_v8_0_audio_enable(adev, &adev->mode_info.audio.pin[i], false);
-
 	adev->mode_info.audio.enabled = false;
 }
 
@@ -1738,8 +1748,7 @@ static void dce_v8_0_afmt_fini(struct amdgpu_device *adev)
 	}
 }
 
-static const u32 vga_control_regs[6] =
-{
+static const u32 vga_control_regs[6] = {
 	mmD1VGA_CONTROL,
 	mmD2VGA_CONTROL,
 	mmD3VGA_CONTROL,
@@ -1813,6 +1822,7 @@ static int dce_v8_0_crtc_do_set_base(struct drm_crtc *crtc,
 		return r;
 
 	if (!atomic) {
+		abo->flags |= AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS;
 		r = amdgpu_bo_pin(abo, AMDGPU_GEM_DOMAIN_VRAM);
 		if (unlikely(r != 0)) {
 			amdgpu_bo_unreserve(abo);
@@ -1893,9 +1903,9 @@ static int dce_v8_0_crtc_do_set_base(struct drm_crtc *crtc,
 	case DRM_FORMAT_XBGR8888:
 	case DRM_FORMAT_ABGR8888:
 		fb_format = ((GRPH_DEPTH_32BPP << GRPH_CONTROL__GRPH_DEPTH__SHIFT) |
-		             (GRPH_FORMAT_ARGB8888 << GRPH_CONTROL__GRPH_FORMAT__SHIFT));
+				(GRPH_FORMAT_ARGB8888 << GRPH_CONTROL__GRPH_FORMAT__SHIFT));
 		fb_swap = ((GRPH_RED_SEL_B << GRPH_SWAP_CNTL__GRPH_RED_CROSSBAR__SHIFT) |
-		           (GRPH_BLUE_SEL_R << GRPH_SWAP_CNTL__GRPH_BLUE_CROSSBAR__SHIFT));
+			(GRPH_BLUE_SEL_R << GRPH_SWAP_CNTL__GRPH_BLUE_CROSSBAR__SHIFT));
 #ifdef __BIG_ENDIAN
 		fb_swap |= (GRPH_ENDIAN_8IN32 << GRPH_SWAP_CNTL__GRPH_ENDIAN_SWAP__SHIFT);
 #endif
@@ -2092,22 +2102,18 @@ static int dce_v8_0_pick_dig_encoder(struct drm_encoder *encoder)
 			return 1;
 		else
 			return 0;
-		break;
 	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
 		if (dig->linkb)
 			return 3;
 		else
 			return 2;
-		break;
 	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
 		if (dig->linkb)
 			return 5;
 		else
 			return 4;
-		break;
 	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY3:
 		return 6;
-		break;
 	default:
 		DRM_ERROR("invalid encoder_id: 0x%x\n", amdgpu_encoder->encoder_id);
 		return 0;
@@ -2309,6 +2315,7 @@ static int dce_v8_0_crtc_cursor_set2(struct drm_crtc *crtc,
 		return ret;
 	}
 
+	aobj->flags |= AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS;
 	ret = amdgpu_bo_pin(aobj, AMDGPU_GEM_DOMAIN_VRAM);
 	amdgpu_bo_unreserve(aobj);
 	if (ret) {
@@ -2437,7 +2444,7 @@ static void dce_v8_0_crtc_dpms(struct drm_crtc *crtc, int mode)
 		break;
 	}
 	/* adjust pm to dpms */
-	amdgpu_pm_compute_clocks(adev);
+	amdgpu_dpm_compute_clocks(adev);
 }
 
 static void dce_v8_0_crtc_prepare(struct drm_crtc *crtc)
@@ -2600,6 +2607,31 @@ static const struct drm_crtc_helper_funcs dce_v8_0_crtc_helper_funcs = {
 	.get_scanout_position = amdgpu_crtc_get_scanout_position,
 };
 
+static void dce_v8_0_panic_flush(struct drm_plane *plane)
+{
+	struct drm_framebuffer *fb;
+	struct amdgpu_crtc *amdgpu_crtc;
+	struct amdgpu_device *adev;
+	uint32_t fb_format;
+
+	if (!plane->fb)
+		return;
+
+	fb = plane->fb;
+	amdgpu_crtc = to_amdgpu_crtc(plane->crtc);
+	adev = drm_to_adev(fb->dev);
+
+	/* Disable DC tiling */
+	fb_format = RREG32(mmGRPH_CONTROL + amdgpu_crtc->crtc_offset);
+	fb_format &= ~GRPH_CONTROL__GRPH_ARRAY_MODE_MASK;
+	WREG32(mmGRPH_CONTROL + amdgpu_crtc->crtc_offset, fb_format);
+}
+
+static const struct drm_plane_helper_funcs dce_v8_0_drm_primary_plane_helper_funcs = {
+	.get_scanout_buffer = amdgpu_display_get_scanout_buffer,
+	.panic_flush = dce_v8_0_panic_flush,
+};
+
 static int dce_v8_0_crtc_init(struct amdgpu_device *adev, int index)
 {
 	struct amdgpu_crtc *amdgpu_crtc;
@@ -2627,13 +2659,14 @@ static int dce_v8_0_crtc_init(struct amdgpu_device *adev, int index)
 	amdgpu_crtc->encoder = NULL;
 	amdgpu_crtc->connector = NULL;
 	drm_crtc_helper_add(&amdgpu_crtc->base, &dce_v8_0_crtc_helper_funcs);
+	drm_plane_helper_add(amdgpu_crtc->base.primary, &dce_v8_0_drm_primary_plane_helper_funcs);
 
 	return 0;
 }
 
-static int dce_v8_0_early_init(void *handle)
+static int dce_v8_0_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	adev->audio_endpt_rreg = &dce_v8_0_audio_endpt_rreg;
 	adev->audio_endpt_wreg = &dce_v8_0_audio_endpt_wreg;
@@ -2667,10 +2700,10 @@ static int dce_v8_0_early_init(void *handle)
 	return 0;
 }
 
-static int dce_v8_0_sw_init(void *handle)
+static int dce_v8_0_sw_init(struct amdgpu_ip_block *ip_block)
 {
 	int r, i;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	for (i = 0; i < adev->mode_info.num_crtc; i++) {
 		r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, i + 1, &adev->crtc_irq);
@@ -2697,9 +2730,13 @@ static int dce_v8_0_sw_init(void *handle)
 	adev_to_drm(adev)->mode_config.max_height = 16384;
 
 	adev_to_drm(adev)->mode_config.preferred_depth = 24;
-	adev_to_drm(adev)->mode_config.prefer_shadow = 1;
+	if (adev->asic_type == CHIP_HAWAII)
+		/* disable prefer shadow for now due to hibernation issues */
+		adev_to_drm(adev)->mode_config.prefer_shadow = 0;
+	else
+		adev_to_drm(adev)->mode_config.prefer_shadow = 1;
 
-	adev_to_drm(adev)->mode_config.fb_base = adev->gmc.aper_base;
+	adev_to_drm(adev)->mode_config.fb_modifiers_not_supported = true;
 
 	r = amdgpu_display_modeset_create_props(adev);
 	if (r)
@@ -2729,17 +2766,29 @@ static int dce_v8_0_sw_init(void *handle)
 	if (r)
 		return r;
 
+	/* Disable vblank IRQs aggressively for power-saving */
+	/* XXX: can this be enabled for DC? */
+	adev_to_drm(adev)->vblank_disable_immediate = true;
+
+	r = drm_vblank_init(adev_to_drm(adev), adev->mode_info.num_crtc);
+	if (r)
+		return r;
+
+	/* Pre-DCE11 */
+	INIT_DELAYED_WORK(&adev->hotplug_work,
+		  amdgpu_display_hotplug_work_func);
+
 	drm_kms_helper_poll_init(adev_to_drm(adev));
 
 	adev->mode_info.mode_config_initialized = true;
 	return 0;
 }
 
-static int dce_v8_0_sw_fini(void *handle)
+static int dce_v8_0_sw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
-	kfree(adev->mode_info.bios_hardcoded_edid);
+	drm_edid_free(adev->mode_info.bios_hardcoded_edid);
 
 	drm_kms_helper_poll_fini(adev_to_drm(adev));
 
@@ -2753,10 +2802,10 @@ static int dce_v8_0_sw_fini(void *handle)
 	return 0;
 }
 
-static int dce_v8_0_hw_init(void *handle)
+static int dce_v8_0_hw_init(struct amdgpu_ip_block *ip_block)
 {
 	int i;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	/* disable vga render */
 	dce_v8_0_set_vga_render_state(adev, false);
@@ -2776,10 +2825,10 @@ static int dce_v8_0_hw_init(void *handle)
 	return 0;
 }
 
-static int dce_v8_0_hw_fini(void *handle)
+static int dce_v8_0_hw_fini(struct amdgpu_ip_block *ip_block)
 {
 	int i;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	dce_v8_0_hpd_fini(adev);
 
@@ -2789,12 +2838,14 @@ static int dce_v8_0_hw_fini(void *handle)
 
 	dce_v8_0_pageflip_interrupt_fini(adev);
 
+	flush_delayed_work(&adev->hotplug_work);
+
 	return 0;
 }
 
-static int dce_v8_0_suspend(void *handle)
+static int dce_v8_0_suspend(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int r;
 
 	r = amdgpu_display_suspend_helper(adev);
@@ -2804,18 +2855,18 @@ static int dce_v8_0_suspend(void *handle)
 	adev->mode_info.bl_level =
 		amdgpu_atombios_encoder_get_backlight_level_from_reg(adev);
 
-	return dce_v8_0_hw_fini(handle);
+	return dce_v8_0_hw_fini(ip_block);
 }
 
-static int dce_v8_0_resume(void *handle)
+static int dce_v8_0_resume(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int ret;
 
 	amdgpu_atombios_encoder_set_backlight_level_to_reg(adev,
 							   adev->mode_info.bl_level);
 
-	ret = dce_v8_0_hw_init(handle);
+	ret = dce_v8_0_hw_init(ip_block);
 
 	/* turn on the BL */
 	if (adev->mode_info.bl_encoder) {
@@ -2830,20 +2881,15 @@ static int dce_v8_0_resume(void *handle)
 	return amdgpu_display_resume_helper(adev);
 }
 
-static bool dce_v8_0_is_idle(void *handle)
+static bool dce_v8_0_is_idle(struct amdgpu_ip_block *ip_block)
 {
 	return true;
 }
 
-static int dce_v8_0_wait_for_idle(void *handle)
-{
-	return 0;
-}
-
-static int dce_v8_0_soft_reset(void *handle)
+static int dce_v8_0_soft_reset(struct amdgpu_ip_block *ip_block)
 {
 	u32 srbm_soft_reset = 0, tmp;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (dce_v8_0_is_display_hung(adev))
 		srbm_soft_reset |= SRBM_SOFT_RESET__SOFT_RESET_DC_MASK;
@@ -2969,7 +3015,7 @@ static void dce_v8_0_set_crtc_vline_interrupt_state(struct amdgpu_device *adev,
 	}
 }
 
-static int dce_v8_0_set_hpd_interrupt_state(struct amdgpu_device *adev,
+static int dce_v8_0_set_hpd_irq_state(struct amdgpu_device *adev,
 					    struct amdgpu_irq_src *src,
 					    unsigned type,
 					    enum amdgpu_interrupt_state state)
@@ -2977,7 +3023,7 @@ static int dce_v8_0_set_hpd_interrupt_state(struct amdgpu_device *adev,
 	u32 dc_hpd_int_cntl;
 
 	if (type >= adev->mode_info.num_hpd) {
-		DRM_DEBUG("invalid hdp %d\n", type);
+		DRM_DEBUG("invalid hpd %d\n", type);
 		return 0;
 	}
 
@@ -2999,7 +3045,7 @@ static int dce_v8_0_set_hpd_interrupt_state(struct amdgpu_device *adev,
 	return 0;
 }
 
-static int dce_v8_0_set_crtc_interrupt_state(struct amdgpu_device *adev,
+static int dce_v8_0_set_crtc_irq_state(struct amdgpu_device *adev,
 					     struct amdgpu_irq_src *src,
 					     unsigned type,
 					     enum amdgpu_interrupt_state state)
@@ -3084,7 +3130,7 @@ static int dce_v8_0_crtc_irq(struct amdgpu_device *adev,
 	return 0;
 }
 
-static int dce_v8_0_set_pageflip_interrupt_state(struct amdgpu_device *adev,
+static int dce_v8_0_set_pageflip_irq_state(struct amdgpu_device *adev,
 						 struct amdgpu_irq_src *src,
 						 unsigned type,
 						 enum amdgpu_interrupt_state state)
@@ -3135,7 +3181,7 @@ static int dce_v8_0_pageflip_irq(struct amdgpu_device *adev,
 
 	spin_lock_irqsave(&adev_to_drm(adev)->event_lock, flags);
 	works = amdgpu_crtc->pflip_works;
-	if (amdgpu_crtc->pflip_status != AMDGPU_FLIP_SUBMITTED){
+	if (amdgpu_crtc->pflip_status != AMDGPU_FLIP_SUBMITTED) {
 		DRM_DEBUG_DRIVER("amdgpu_crtc->pflip_status = %d != "
 						"AMDGPU_FLIP_SUBMITTED(%d)\n",
 						amdgpu_crtc->pflip_status,
@@ -3164,7 +3210,7 @@ static int dce_v8_0_hpd_irq(struct amdgpu_device *adev,
 			    struct amdgpu_irq_src *source,
 			    struct amdgpu_iv_entry *entry)
 {
-	uint32_t disp_int, mask, tmp;
+	uint32_t disp_int, mask;
 	unsigned hpd;
 
 	if (entry->src_data[0] >= adev->mode_info.num_hpd) {
@@ -3177,10 +3223,8 @@ static int dce_v8_0_hpd_irq(struct amdgpu_device *adev,
 	mask = interrupt_status_offsets[hpd].hpd;
 
 	if (disp_int & mask) {
-		tmp = RREG32(mmDC_HPD1_INT_CONTROL + hpd_offsets[hpd]);
-		tmp |= DC_HPD1_INT_CONTROL__DC_HPD1_INT_ACK_MASK;
-		WREG32(mmDC_HPD1_INT_CONTROL + hpd_offsets[hpd], tmp);
-		schedule_work(&adev->hotplug_work);
+		dce_v8_0_hpd_int_ack(adev, hpd);
+		schedule_delayed_work(&adev->hotplug_work, 0);
 		DRM_DEBUG("IH: HPD%d\n", hpd + 1);
 	}
 
@@ -3188,13 +3232,13 @@ static int dce_v8_0_hpd_irq(struct amdgpu_device *adev,
 
 }
 
-static int dce_v8_0_set_clockgating_state(void *handle,
+static int dce_v8_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_clockgating_state state)
 {
 	return 0;
 }
 
-static int dce_v8_0_set_powergating_state(void *handle,
+static int dce_v8_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_powergating_state state)
 {
 	return 0;
@@ -3203,7 +3247,6 @@ static int dce_v8_0_set_powergating_state(void *handle,
 static const struct amd_ip_funcs dce_v8_0_ip_funcs = {
 	.name = "dce_v8_0",
 	.early_init = dce_v8_0_early_init,
-	.late_init = NULL,
 	.sw_init = dce_v8_0_sw_init,
 	.sw_fini = dce_v8_0_sw_fini,
 	.hw_init = dce_v8_0_hw_init,
@@ -3211,7 +3254,6 @@ static const struct amd_ip_funcs dce_v8_0_ip_funcs = {
 	.suspend = dce_v8_0_suspend,
 	.resume = dce_v8_0_resume,
 	.is_idle = dce_v8_0_is_idle,
-	.wait_for_idle = dce_v8_0_wait_for_idle,
 	.soft_reset = dce_v8_0_soft_reset,
 	.set_clockgating_state = dce_v8_0_set_clockgating_state,
 	.set_powergating_state = dce_v8_0_set_powergating_state,
@@ -3499,17 +3541,17 @@ static void dce_v8_0_set_display_funcs(struct amdgpu_device *adev)
 }
 
 static const struct amdgpu_irq_src_funcs dce_v8_0_crtc_irq_funcs = {
-	.set = dce_v8_0_set_crtc_interrupt_state,
+	.set = dce_v8_0_set_crtc_irq_state,
 	.process = dce_v8_0_crtc_irq,
 };
 
 static const struct amdgpu_irq_src_funcs dce_v8_0_pageflip_irq_funcs = {
-	.set = dce_v8_0_set_pageflip_interrupt_state,
+	.set = dce_v8_0_set_pageflip_irq_state,
 	.process = dce_v8_0_pageflip_irq,
 };
 
 static const struct amdgpu_irq_src_funcs dce_v8_0_hpd_irq_funcs = {
-	.set = dce_v8_0_set_hpd_interrupt_state,
+	.set = dce_v8_0_set_hpd_irq_state,
 	.process = dce_v8_0_hpd_irq,
 };
 
@@ -3528,8 +3570,7 @@ static void dce_v8_0_set_irq_funcs(struct amdgpu_device *adev)
 	adev->hpd_irq.funcs = &dce_v8_0_hpd_irq_funcs;
 }
 
-const struct amdgpu_ip_block_version dce_v8_0_ip_block =
-{
+const struct amdgpu_ip_block_version dce_v8_0_ip_block = {
 	.type = AMD_IP_BLOCK_TYPE_DCE,
 	.major = 8,
 	.minor = 0,
@@ -3537,8 +3578,7 @@ const struct amdgpu_ip_block_version dce_v8_0_ip_block =
 	.funcs = &dce_v8_0_ip_funcs,
 };
 
-const struct amdgpu_ip_block_version dce_v8_1_ip_block =
-{
+const struct amdgpu_ip_block_version dce_v8_1_ip_block = {
 	.type = AMD_IP_BLOCK_TYPE_DCE,
 	.major = 8,
 	.minor = 1,
@@ -3546,8 +3586,7 @@ const struct amdgpu_ip_block_version dce_v8_1_ip_block =
 	.funcs = &dce_v8_0_ip_funcs,
 };
 
-const struct amdgpu_ip_block_version dce_v8_2_ip_block =
-{
+const struct amdgpu_ip_block_version dce_v8_2_ip_block = {
 	.type = AMD_IP_BLOCK_TYPE_DCE,
 	.major = 8,
 	.minor = 2,
@@ -3555,8 +3594,7 @@ const struct amdgpu_ip_block_version dce_v8_2_ip_block =
 	.funcs = &dce_v8_0_ip_funcs,
 };
 
-const struct amdgpu_ip_block_version dce_v8_3_ip_block =
-{
+const struct amdgpu_ip_block_version dce_v8_3_ip_block = {
 	.type = AMD_IP_BLOCK_TYPE_DCE,
 	.major = 8,
 	.minor = 3,
@@ -3564,8 +3602,7 @@ const struct amdgpu_ip_block_version dce_v8_3_ip_block =
 	.funcs = &dce_v8_0_ip_funcs,
 };
 
-const struct amdgpu_ip_block_version dce_v8_5_ip_block =
-{
+const struct amdgpu_ip_block_version dce_v8_5_ip_block = {
 	.type = AMD_IP_BLOCK_TYPE_DCE,
 	.major = 8,
 	.minor = 5,
diff --git a/drivers/gpu/drm/amd/amdgpu/df_v1_7.c b/drivers/gpu/drm/amd/amdgpu/df_v1_7.c
index 2d01ac0d4c11..cd298556f7a6 100644
--- a/drivers/gpu/drm/amd/amdgpu/df_v1_7.c
+++ b/drivers/gpu/drm/amd/amdgpu/df_v1_7.c
@@ -70,6 +70,8 @@ static u32 df_v1_7_get_hbm_channel_number(struct amdgpu_device *adev)
 	int fb_channel_number;
 
 	fb_channel_number = adev->df.funcs->get_fb_channel_number(adev);
+	if (fb_channel_number >= ARRAY_SIZE(df_v1_7_channel_number))
+		fb_channel_number = 0;
 
 	return df_v1_7_channel_number[fb_channel_number];
 }
@@ -94,12 +96,12 @@ static void df_v1_7_update_medium_grain_clock_gating(struct amdgpu_device *adev,
 		WREG32_SOC15(DF, 0, mmDF_PIE_AON0_DfGlobalClkGater, tmp);
 	}
 
-	/* Exit boradcast mode */
+	/* Exit broadcast mode */
 	adev->df.funcs->enable_broadcast_mode(adev, false);
 }
 
 static void df_v1_7_get_clockgating_state(struct amdgpu_device *adev,
-					  u32 *flags)
+					  u64 *flags)
 {
 	u32 tmp;
 
diff --git a/drivers/gpu/drm/amd/amdgpu/df_v3_6.c b/drivers/gpu/drm/amd/amdgpu/df_v3_6.c
index 43c5e3ec9a39..621aeca53880 100644
--- a/drivers/gpu/drm/amd/amdgpu/df_v3_6.c
+++ b/drivers/gpu/drm/amd/amdgpu/df_v3_6.c
@@ -254,8 +254,8 @@ static void df_v3_6_sw_init(struct amdgpu_device *adev)
 
 static void df_v3_6_sw_fini(struct amdgpu_device *adev)
 {
-
-	device_remove_file(adev->dev, &dev_attr_df_cntr_avail);
+	if (adev->dev->kobj.sd)
+		device_remove_file(adev->dev, &dev_attr_df_cntr_avail);
 
 }
 
@@ -332,7 +332,7 @@ static void df_v3_6_update_medium_grain_clock_gating(struct amdgpu_device *adev,
 }
 
 static void df_v3_6_get_clockgating_state(struct amdgpu_device *adev,
-					  u32 *flags)
+					  u64 *flags)
 {
 	u32 tmp;
 
@@ -458,7 +458,7 @@ static int df_v3_6_pmc_add_cntr(struct amdgpu_device *adev,
 
 #define DEFERRED_ARM_MASK	(1 << 31)
 static int df_v3_6_pmc_set_deferred(struct amdgpu_device *adev,
-				    int counter_idx, uint64_t config,
+				    uint64_t config, int counter_idx,
 				    bool is_deferred)
 {
 
@@ -476,8 +476,8 @@ static int df_v3_6_pmc_set_deferred(struct amdgpu_device *adev,
 }
 
 static bool df_v3_6_pmc_is_deferred(struct amdgpu_device *adev,
-				    int counter_idx,
-				    uint64_t config)
+				    uint64_t config,
+				    int counter_idx)
 {
 	return	(df_v3_6_pmc_has_counter(adev, config, counter_idx) &&
 			(adev->df_perfmon_config_assign_mask[counter_idx]
diff --git a/drivers/gpu/drm/amd/amdgpu/df_v4_15.c b/drivers/gpu/drm/amd/amdgpu/df_v4_15.c
new file mode 100644
index 000000000000..2a573e33908b
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/df_v4_15.c
@@ -0,0 +1,45 @@
+/*
+ * Copyright 2024 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include "amdgpu.h"
+#include "df_v4_15.h"
+
+#include "df/df_4_15_offset.h"
+#include "df/df_4_15_sh_mask.h"
+
+static void df_v4_15_hw_init(struct amdgpu_device *adev)
+{
+	if (adev->have_atomics_support) {
+		uint32_t tmp;
+		uint32_t dis_lcl_proc = (1 <<  1 |
+					1 <<  2 |
+					1 << 13);
+
+		tmp = RREG32_SOC15(DF, 0, regNCSConfigurationRegister1);
+		tmp |= (dis_lcl_proc << NCSConfigurationRegister1__DisIntAtomicsLclProcessing__SHIFT);
+		WREG32_SOC15(DF, 0, regNCSConfigurationRegister1, tmp);
+	}
+}
+
+const struct amdgpu_df_funcs df_v4_15_funcs = {
+	.hw_init = df_v4_15_hw_init
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/df_v4_15.h b/drivers/gpu/drm/amd/amdgpu/df_v4_15.h
new file mode 100644
index 000000000000..dddf2422112a
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/df_v4_15.h
@@ -0,0 +1,30 @@
+/*
+ * Copyright 2024 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __DF_V4_15_H__
+#define __DF_V4_15_H__
+
+extern const struct amdgpu_df_funcs df_v4_15_funcs;
+
+#endif /* __DF_V4_15_H__ */
+
diff --git a/drivers/gpu/drm/amd/amdgpu/df_v4_3.c b/drivers/gpu/drm/amd/amdgpu/df_v4_3.c
new file mode 100644
index 000000000000..e8b9e19ede2e
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/df_v4_3.c
@@ -0,0 +1,61 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include "amdgpu.h"
+#include "df_v4_3.h"
+
+#include "df/df_4_3_offset.h"
+#include "df/df_4_3_sh_mask.h"
+
+static bool df_v4_3_query_ras_poison_mode(struct amdgpu_device *adev)
+{
+	uint32_t hw_assert_msklo, hw_assert_mskhi;
+	uint32_t v0, v1, v28, v31;
+
+	hw_assert_msklo = RREG32_SOC15(DF, 0,
+				regDF_CS_UMC_AON0_HardwareAssertMaskLow);
+	hw_assert_mskhi = RREG32_SOC15(DF, 0,
+				regDF_NCS_PG0_HardwareAssertMaskHigh);
+
+	v0 = REG_GET_FIELD(hw_assert_msklo,
+		DF_CS_UMC_AON0_HardwareAssertMaskLow, HWAssertMsk0);
+	v1 = REG_GET_FIELD(hw_assert_msklo,
+		DF_CS_UMC_AON0_HardwareAssertMaskLow, HWAssertMsk1);
+	v28 = REG_GET_FIELD(hw_assert_mskhi,
+		DF_NCS_PG0_HardwareAssertMaskHigh, HWAssertMsk28);
+	v31 = REG_GET_FIELD(hw_assert_mskhi,
+		DF_NCS_PG0_HardwareAssertMaskHigh, HWAssertMsk31);
+
+	if (v0 && v1 && v28 && v31)
+		return true;
+	else if (!v0 && !v1 && !v28 && !v31)
+		return false;
+	else {
+		dev_warn(adev->dev, "DF poison setting is inconsistent(%d:%d:%d:%d)!\n",
+				v0, v1, v28, v31);
+		return false;
+	}
+}
+
+const struct amdgpu_df_funcs df_v4_3_funcs = {
+	.query_ras_poison_mode = df_v4_3_query_ras_poison_mode,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/df_v4_3.h b/drivers/gpu/drm/amd/amdgpu/df_v4_3.h
new file mode 100644
index 000000000000..06ef0724edd3
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/df_v4_3.h
@@ -0,0 +1,31 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __DF_V4_3_H__
+#define __DF_V4_3_H__
+
+#include "soc15_common.h"
+
+extern const struct amdgpu_df_funcs df_v4_3_funcs;
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/df_v4_6_2.c b/drivers/gpu/drm/amd/amdgpu/df_v4_6_2.c
new file mode 100644
index 000000000000..a47960a0babd
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/df_v4_6_2.c
@@ -0,0 +1,34 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include "amdgpu.h"
+#include "df_v4_6_2.h"
+
+static bool df_v4_6_2_query_ras_poison_mode(struct amdgpu_device *adev)
+{
+	/* return true since related regs are inaccessible */
+	return true;
+}
+
+const struct amdgpu_df_funcs df_v4_6_2_funcs = {
+	.query_ras_poison_mode = df_v4_6_2_query_ras_poison_mode,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/df_v4_6_2.h b/drivers/gpu/drm/amd/amdgpu/df_v4_6_2.h
new file mode 100644
index 000000000000..3bc3e6d216e2
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/df_v4_6_2.h
@@ -0,0 +1,31 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __DF_V4_6_2_H__
+#define __DF_V4_6_2_H__
+
+#include "soc15_common.h"
+
+extern const struct amdgpu_df_funcs df_v4_6_2_funcs;
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
index dbe7442fb25c..d75b9940f248 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
@@ -40,11 +40,11 @@
 #include "ivsrcid/gfx/irqsrcs_gfx_10_1.h"
 
 #include "soc15.h"
-#include "soc15d.h"
 #include "soc15_common.h"
 #include "clearstate_gfx10.h"
 #include "v10_structs.h"
 #include "gfx_v10_0.h"
+#include "gfx_v10_0_cleaner_shader.h"
 #include "nbio_v2_3.h"
 
 /*
@@ -53,13 +53,9 @@
  * 2. Async ring
  */
 #define GFX10_NUM_GFX_RINGS_NV1X	1
-#define GFX10_NUM_GFX_RINGS_Sienna_Cichlid	1
+#define GFX10_NUM_GFX_RINGS_Sienna_Cichlid	2
 #define GFX10_MEC_HPD_SIZE	2048
 
-#define RLCG_VFGATE_DISABLED	0x4000000
-#define RLCG_WRONG_OPERATION_TYPE	0x2000000
-#define RLCG_NOT_IN_RANGE	0x1000000
-
 #define F32_CE_PROGRAM_RAM_SIZE		65536
 #define RLCG_UCODE_LOADING_START_ADDRESS	0x00002000L
 
@@ -106,10 +102,21 @@
 #define mmGCR_GENERAL_CNTL_Sienna_Cichlid			0x1580
 #define mmGCR_GENERAL_CNTL_Sienna_Cichlid_BASE_IDX	0
 
+#define mmGOLDEN_TSC_COUNT_UPPER_Cyan_Skillfish                0x0105
+#define mmGOLDEN_TSC_COUNT_UPPER_Cyan_Skillfish_BASE_IDX       1
+#define mmGOLDEN_TSC_COUNT_LOWER_Cyan_Skillfish                0x0106
+#define mmGOLDEN_TSC_COUNT_LOWER_Cyan_Skillfish_BASE_IDX       1
+
 #define mmGOLDEN_TSC_COUNT_UPPER_Vangogh                0x0025
 #define mmGOLDEN_TSC_COUNT_UPPER_Vangogh_BASE_IDX       1
 #define mmGOLDEN_TSC_COUNT_LOWER_Vangogh                0x0026
 #define mmGOLDEN_TSC_COUNT_LOWER_Vangogh_BASE_IDX       1
+
+#define mmGOLDEN_TSC_COUNT_UPPER_GC_10_3_6                0x002d
+#define mmGOLDEN_TSC_COUNT_UPPER_GC_10_3_6_BASE_IDX       1
+#define mmGOLDEN_TSC_COUNT_LOWER_GC_10_3_6                0x002e
+#define mmGOLDEN_TSC_COUNT_LOWER_GC_10_3_6_BASE_IDX       1
+
 #define mmSPI_CONFIG_CNTL_1_Vangogh		 0x2441
 #define mmSPI_CONFIG_CNTL_1_Vangogh_BASE_IDX	 1
 #define mmVGT_TF_MEMORY_BASE_HI_Vangogh          0x2261
@@ -180,14 +187,6 @@
 #define mmRLC_SPARE_INT_0_Sienna_Cichlid               0x4ca5
 #define mmRLC_SPARE_INT_0_Sienna_Cichlid_BASE_IDX      1
 
-#define GFX_RLCG_GC_WRITE_OLD	(0x8 << 28)
-#define GFX_RLCG_GC_WRITE	(0x0 << 28)
-#define GFX_RLCG_GC_READ	(0x1 << 28)
-#define GFX_RLCG_MMHUB_WRITE	(0x2 << 28)
-
-#define RLCG_ERROR_REPORT_ENABLED(adev) \
-	(amdgpu_sriov_reg_indirect_mmhub(adev) || amdgpu_sriov_reg_indirect_gc(adev))
-
 MODULE_FIRMWARE("amdgpu/navi10_ce.bin");
 MODULE_FIRMWARE("amdgpu/navi10_pfp.bin");
 MODULE_FIRMWARE("amdgpu/navi10_me.bin");
@@ -256,13 +255,6 @@ MODULE_FIRMWARE("amdgpu/yellow_carp_mec.bin");
 MODULE_FIRMWARE("amdgpu/yellow_carp_mec2.bin");
 MODULE_FIRMWARE("amdgpu/yellow_carp_rlc.bin");
 
-MODULE_FIRMWARE("amdgpu/cyan_skillfish_ce.bin");
-MODULE_FIRMWARE("amdgpu/cyan_skillfish_pfp.bin");
-MODULE_FIRMWARE("amdgpu/cyan_skillfish_me.bin");
-MODULE_FIRMWARE("amdgpu/cyan_skillfish_mec.bin");
-MODULE_FIRMWARE("amdgpu/cyan_skillfish_mec2.bin");
-MODULE_FIRMWARE("amdgpu/cyan_skillfish_rlc.bin");
-
 MODULE_FIRMWARE("amdgpu/cyan_skillfish2_ce.bin");
 MODULE_FIRMWARE("amdgpu/cyan_skillfish2_pfp.bin");
 MODULE_FIRMWARE("amdgpu/cyan_skillfish2_me.bin");
@@ -270,8 +262,225 @@ MODULE_FIRMWARE("amdgpu/cyan_skillfish2_mec.bin");
 MODULE_FIRMWARE("amdgpu/cyan_skillfish2_mec2.bin");
 MODULE_FIRMWARE("amdgpu/cyan_skillfish2_rlc.bin");
 
-static const struct soc15_reg_golden golden_settings_gc_10_1[] =
-{
+MODULE_FIRMWARE("amdgpu/gc_10_3_6_ce.bin");
+MODULE_FIRMWARE("amdgpu/gc_10_3_6_pfp.bin");
+MODULE_FIRMWARE("amdgpu/gc_10_3_6_me.bin");
+MODULE_FIRMWARE("amdgpu/gc_10_3_6_mec.bin");
+MODULE_FIRMWARE("amdgpu/gc_10_3_6_mec2.bin");
+MODULE_FIRMWARE("amdgpu/gc_10_3_6_rlc.bin");
+
+MODULE_FIRMWARE("amdgpu/gc_10_3_7_ce.bin");
+MODULE_FIRMWARE("amdgpu/gc_10_3_7_pfp.bin");
+MODULE_FIRMWARE("amdgpu/gc_10_3_7_me.bin");
+MODULE_FIRMWARE("amdgpu/gc_10_3_7_mec.bin");
+MODULE_FIRMWARE("amdgpu/gc_10_3_7_mec2.bin");
+MODULE_FIRMWARE("amdgpu/gc_10_3_7_rlc.bin");
+
+static const struct amdgpu_hwip_reg_entry gc_reg_list_10_1[] = {
+	SOC15_REG_ENTRY_STR(GC, 0, mmGRBM_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmGRBM_STATUS2),
+	SOC15_REG_ENTRY_STR(GC, 0, mmGRBM_STATUS3),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_STALLED_STAT1),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_STALLED_STAT2),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CPC_STALLED_STAT1),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CPF_STALLED_STAT1),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_BUSY_STAT),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CPC_BUSY_STAT),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CPF_BUSY_STAT),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CPC_BUSY_STAT2),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CPF_BUSY_STAT2),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CPF_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_GFX_ERROR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_GFX_HPD_STATUS0),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_RB_BASE),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_RB_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_RB_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_RB0_BASE),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_RB0_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_RB0_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_RB1_BASE),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_RB1_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_RB1_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_RB2_BASE),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_RB2_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_RB2_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_IB1_CMD_BUFSZ),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_IB2_CMD_BUFSZ),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_IB1_CMD_BUFSZ),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_IB2_CMD_BUFSZ),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_IB1_BASE_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_IB1_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_IB1_BUFSZ),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_IB2_BASE_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_IB2_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_IB2_BUFSZ),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_IB1_BASE_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_IB1_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_IB1_BUFSZ),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_IB2_BASE_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_IB2_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_IB2_BUFSZ),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCPF_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCPC_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCPG_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmGDS_PROTECTION_FAULT),
+	SOC15_REG_ENTRY_STR(GC, 0, mmGDS_VM_PROTECTION_FAULT),
+	SOC15_REG_ENTRY_STR(GC, 0, mmIA_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmIA_UTCL1_STATUS_2),
+	SOC15_REG_ENTRY_STR(GC, 0, mmPA_CL_CNTL_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmRLC_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmRMI_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSQC_DCACHE_UTCL0_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSQC_ICACHE_UTCL0_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSQG_UTCL0_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmTCP_UTCL0_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmWD_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmGCVM_L2_PROTECTION_FAULT_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, mmGCVM_L2_PROTECTION_FAULT_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_DEBUG),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_MEC_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_MES_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_INSTR_PNTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_MEC1_INSTR_PNTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_MEC2_INSTR_PNTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_MES_DEBUG_INTERRUPT_INSTR_PNTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_MES_INSTR_PNTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_ME_INSTR_PNTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_PFP_INSTR_PNTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CPC_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmRLC_STAT),
+	SOC15_REG_ENTRY_STR(GC, 0, mmRLC_SMU_COMMAND),
+	SOC15_REG_ENTRY_STR(GC, 0, mmRLC_SMU_MESSAGE),
+	SOC15_REG_ENTRY_STR(GC, 0, mmRLC_SMU_ARGUMENT_1),
+	SOC15_REG_ENTRY_STR(GC, 0, mmRLC_SMU_ARGUMENT_2),
+	SOC15_REG_ENTRY_STR(GC, 0, mmRLC_SMU_ARGUMENT_3),
+	SOC15_REG_ENTRY_STR(GC, 0, mmRLC_SMU_ARGUMENT_4),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSMU_RLC_RESPONSE),
+	SOC15_REG_ENTRY_STR(GC, 0, mmRLC_SAFE_MODE),
+	SOC15_REG_ENTRY_STR(GC, 0, mmRLC_SMU_SAFE_MODE),
+	SOC15_REG_ENTRY_STR(GC, 0, mmRLC_RLCS_GPM_STAT_2),
+	SOC15_REG_ENTRY_STR(GC, 0, mmRLC_SPP_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmRLC_RLCS_BOOTLOAD_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmRLC_INT_STAT),
+	SOC15_REG_ENTRY_STR(GC, 0, mmRLC_GPM_GENERAL_6),
+	SOC15_REG_ENTRY_STR(GC, 0, mmRLC_GPM_DEBUG_INST_A),
+	SOC15_REG_ENTRY_STR(GC, 0, mmRLC_GPM_DEBUG_INST_B),
+	SOC15_REG_ENTRY_STR(GC, 0, mmRLC_GPM_DEBUG_INST_ADDR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmRLC_LX6_CORE_PDEBUG_INST),
+	/* cp header registers */
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_MES_HEADER_DUMP),
+	/* SE status registers */
+	SOC15_REG_ENTRY_STR(GC, 0, mmGRBM_STATUS_SE0),
+	SOC15_REG_ENTRY_STR(GC, 0, mmGRBM_STATUS_SE1),
+	SOC15_REG_ENTRY_STR(GC, 0, mmGRBM_STATUS_SE2),
+	SOC15_REG_ENTRY_STR(GC, 0, mmGRBM_STATUS_SE3)
+};
+
+static const struct amdgpu_hwip_reg_entry gc_cp_reg_list_10[] = {
+	/* compute registers */
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_VMID),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_PERSISTENT_STATE),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_PIPE_PRIORITY),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_QUEUE_PRIORITY),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_QUANTUM),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_PQ_BASE),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_PQ_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_PQ_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_PQ_WPTR_POLL_ADDR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_PQ_WPTR_POLL_ADDR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_PQ_DOORBELL_CONTROL),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_PQ_CONTROL),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_IB_BASE_ADDR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_IB_BASE_ADDR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_IB_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_IB_CONTROL),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_DEQUEUE_REQUEST),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_EOP_BASE_ADDR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_EOP_BASE_ADDR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_EOP_CONTROL),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_EOP_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_EOP_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_EOP_EVENTS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_CTX_SAVE_BASE_ADDR_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_CTX_SAVE_BASE_ADDR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_CTX_SAVE_CONTROL),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_CNTL_STACK_OFFSET),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_CNTL_STACK_SIZE),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_WG_STATE_OFFSET),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_CTX_SAVE_SIZE),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_GDS_RESOURCE_STATE),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_ERROR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_EOP_WPTR_MEM),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_PQ_WPTR_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_PQ_WPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_SUSPEND_CNTL_STACK_OFFSET),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_SUSPEND_CNTL_STACK_DW_CNT),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_SUSPEND_WG_STATE_OFFSET),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_DEQUEUE_STATUS),
+	/* cp header registers */
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_MEC_ME1_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_MEC_ME1_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_MEC_ME1_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_MEC_ME1_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_MEC_ME1_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_MEC_ME1_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_MEC_ME1_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_MEC_ME1_HEADER_DUMP),
+};
+
+static const struct amdgpu_hwip_reg_entry gc_gfx_queue_reg_list_10[] = {
+	/* gfx queue registers */
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_GFX_HQD_ACTIVE),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_GFX_HQD_QUEUE_PRIORITY),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_GFX_HQD_BASE),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_GFX_HQD_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_GFX_HQD_OFFSET),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_GFX_HQD_CSMD_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_GFX_HQD_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_GFX_HQD_WPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_GFX_HQD_DEQUEUE_REQUEST),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_GFX_HQD_MAPPED),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_GFX_HQD_QUE_MGR_CONTROL),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_GFX_HQD_HQ_CONTROL0),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_GFX_HQD_HQ_STATUS0),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_GFX_HQD_CE_WPTR_POLL_ADDR_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_GFX_HQD_CE_WPTR_POLL_ADDR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_GFX_HQD_CE_OFFSET),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_GFX_HQD_CE_CSMD_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_GFX_HQD_CE_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_GFX_HQD_CE_WPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_GFX_MQD_BASE_ADDR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_GFX_MQD_BASE_ADDR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_RB_WPTR_POLL_ADDR_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_RB_WPTR_POLL_ADDR_HI),
+	/* gfx header registers */
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_PFP_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_PFP_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_PFP_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_PFP_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_PFP_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_PFP_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_PFP_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_PFP_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_ME_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_ME_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_ME_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_ME_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_ME_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_ME_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_ME_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_ME_HEADER_DUMP),
+};
+
+static const struct soc15_reg_golden golden_settings_gc_10_1[] = {
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmCB_HW_CONTROL_4, 0xffffffff, 0x00400014),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmCGTT_CPF_CLK_CTRL, 0xfcff8fff, 0xf8000100),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmCGTT_SPI_CLK_CTRL, 0xcd000000, 0x0d000100),
@@ -314,13 +523,11 @@ static const struct soc15_reg_golden golden_settings_gc_10_1[] =
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmUTCL1_CTRL, 0x00c00000, 0x00c00000)
 };
 
-static const struct soc15_reg_golden golden_settings_gc_10_0_nv10[] =
-{
+static const struct soc15_reg_golden golden_settings_gc_10_0_nv10[] = {
 	/* Pending on emulation bring up */
 };
 
-static const struct soc15_reg_golden golden_settings_gc_rlc_spm_10_0_nv10[] =
-{
+static const struct soc15_reg_golden golden_settings_gc_rlc_spm_10_0_nv10[] = {
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGRBM_GFX_INDEX, 0xe0000000, 0x0),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGRBM_GFX_INDEX, 0xffffff, 0x0),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmRLC_SPM_GLB_SAMPLEDELAY_IND_ADDR, 0xFFFFFFFF, 0x28),
@@ -1375,8 +1582,7 @@ static const struct soc15_reg_golden golden_settings_gc_rlc_spm_10_0_nv10[] =
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGRBM_GFX_INDEX, 0xFFFFFFFF, 0xe0000000)
 };
 
-static const struct soc15_reg_golden golden_settings_gc_10_1_1[] =
-{
+static const struct soc15_reg_golden golden_settings_gc_10_1_1[] = {
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmCB_HW_CONTROL_4, 0xffffffff, 0x003c0014),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmCGTT_GS_NGG_CLK_CTRL, 0xffff8fff, 0xffff8100),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmCGTT_IA_CLK_CTRL, 0xffff0fff, 0xffff0100),
@@ -1417,8 +1623,7 @@ static const struct soc15_reg_golden golden_settings_gc_10_1_1[] =
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmUTCL1_CTRL, 0x00c00000, 0x00c00000),
 };
 
-static const struct soc15_reg_golden golden_settings_gc_10_1_2[] =
-{
+static const struct soc15_reg_golden golden_settings_gc_10_1_2[] = {
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmCB_HW_CONTROL_4, 0x003e001f, 0x003c0014),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmCGTT_GS_NGG_CLK_CTRL, 0xffff8fff, 0xffff8100),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmCGTT_IA_CLK_CTRL, 0xffff0fff, 0xffff0100),
@@ -1463,150 +1668,11 @@ static const struct soc15_reg_golden golden_settings_gc_10_1_2[] =
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmUTCL1_CTRL, 0xffffffff, 0x00c00000)
 };
 
-static bool gfx_v10_get_rlcg_flag(struct amdgpu_device *adev, u32 acc_flags, u32 hwip,
-				 int write, u32 *rlcg_flag)
-{
-	switch (hwip) {
-	case GC_HWIP:
-		if (amdgpu_sriov_reg_indirect_gc(adev)) {
-			*rlcg_flag = write ? GFX_RLCG_GC_WRITE : GFX_RLCG_GC_READ;
-
-			return true;
-		/* only in new version, AMDGPU_REGS_NO_KIQ and AMDGPU_REGS_RLC enabled simultaneously */
-		} else if ((acc_flags & AMDGPU_REGS_RLC) && !(acc_flags & AMDGPU_REGS_NO_KIQ)) {
-			*rlcg_flag = GFX_RLCG_GC_WRITE_OLD;
-
-			return true;
-		}
-
-		break;
-	case MMHUB_HWIP:
-		if (amdgpu_sriov_reg_indirect_mmhub(adev) &&
-		    (acc_flags & AMDGPU_REGS_RLC) && write) {
-			*rlcg_flag = GFX_RLCG_MMHUB_WRITE;
-			return true;
-		}
-
-		break;
-	default:
-		DRM_DEBUG("Not program register by RLCG\n");
-	}
-
-	return false;
-}
-
-static u32 gfx_v10_rlcg_rw(struct amdgpu_device *adev, u32 offset, u32 v, uint32_t flag)
-{
-	static void *scratch_reg0;
-	static void *scratch_reg1;
-	static void *scratch_reg2;
-	static void *scratch_reg3;
-	static void *spare_int;
-	static uint32_t grbm_cntl;
-	static uint32_t grbm_idx;
-	uint32_t i = 0;
-	uint32_t retries = 50000;
-	u32 ret = 0;
-	u32 tmp;
-
-	scratch_reg0 = adev->rmmio +
-		       (adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG0_BASE_IDX] + mmSCRATCH_REG0) * 4;
-	scratch_reg1 = adev->rmmio +
-		       (adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG1_BASE_IDX] + mmSCRATCH_REG1) * 4;
-	scratch_reg2 = adev->rmmio +
-		       (adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG0_BASE_IDX] + mmSCRATCH_REG2) * 4;
-	scratch_reg3 = adev->rmmio +
-		       (adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG1_BASE_IDX] + mmSCRATCH_REG3) * 4;
-
-	if (adev->ip_versions[GC_HWIP][0] >= IP_VERSION(10, 3, 0)) {
-		spare_int = adev->rmmio +
-			    (adev->reg_offset[GC_HWIP][0][mmRLC_SPARE_INT_0_Sienna_Cichlid_BASE_IDX]
-			     + mmRLC_SPARE_INT_0_Sienna_Cichlid) * 4;
-	} else {
-		spare_int = adev->rmmio +
-			    (adev->reg_offset[GC_HWIP][0][mmRLC_SPARE_INT_BASE_IDX] + mmRLC_SPARE_INT) * 4;
-	}
-
-	grbm_cntl = adev->reg_offset[GC_HWIP][0][mmGRBM_GFX_CNTL_BASE_IDX] + mmGRBM_GFX_CNTL;
-	grbm_idx = adev->reg_offset[GC_HWIP][0][mmGRBM_GFX_INDEX_BASE_IDX] + mmGRBM_GFX_INDEX;
-
-	if (offset == grbm_cntl || offset == grbm_idx) {
-		if (offset  == grbm_cntl)
-			writel(v, scratch_reg2);
-		else if (offset == grbm_idx)
-			writel(v, scratch_reg3);
-
-		writel(v, ((void __iomem *)adev->rmmio) + (offset * 4));
-	} else {
-		writel(v, scratch_reg0);
-		writel(offset | flag, scratch_reg1);
-		writel(1, spare_int);
-
-		for (i = 0; i < retries; i++) {
-			tmp = readl(scratch_reg1);
-			if (!(tmp & flag))
-				break;
-
-			udelay(10);
-		}
-
-		if (i >= retries) {
-			if (RLCG_ERROR_REPORT_ENABLED(adev)) {
-				if (tmp & RLCG_VFGATE_DISABLED)
-					pr_err("The vfgate is disabled, program reg:0x%05x failed!\n", offset);
-				else if (tmp & RLCG_WRONG_OPERATION_TYPE)
-					pr_err("Wrong operation type, program reg:0x%05x failed!\n", offset);
-				else if (tmp & RLCG_NOT_IN_RANGE)
-					pr_err("The register is not in range, program reg:0x%05x failed!\n", offset);
-				else
-					pr_err("Unknown error type, program reg:0x%05x failed!\n", offset);
-			} else
-				pr_err("timeout: rlcg program reg:0x%05x failed!\n", offset);
-		}
-	}
-
-	ret = readl(scratch_reg0);
-
-	return ret;
-}
-
-static void gfx_v10_sriov_wreg(struct amdgpu_device *adev, u32 offset, u32 value, u32 acc_flags, u32 hwip)
-{
-	u32 rlcg_flag;
-
-	if (!amdgpu_sriov_runtime(adev) &&
-	    gfx_v10_get_rlcg_flag(adev, acc_flags, hwip, 1, &rlcg_flag)) {
-		gfx_v10_rlcg_rw(adev, offset, value, rlcg_flag);
-		return;
-	}
-
-	if (acc_flags & AMDGPU_REGS_NO_KIQ)
-		WREG32_NO_KIQ(offset, value);
-	else
-		WREG32(offset, value);
-}
-
-static u32 gfx_v10_sriov_rreg(struct amdgpu_device *adev, u32 offset, u32 acc_flags, u32 hwip)
-{
-	u32 rlcg_flag;
-
-	if (!amdgpu_sriov_runtime(adev) &&
-	    gfx_v10_get_rlcg_flag(adev, acc_flags, hwip, 0, &rlcg_flag))
-		return gfx_v10_rlcg_rw(adev, offset, 0, rlcg_flag);
-
-	if (acc_flags & AMDGPU_REGS_NO_KIQ)
-		return RREG32_NO_KIQ(offset);
-	else
-		return RREG32(offset);
-}
-
-static const struct soc15_reg_golden golden_settings_gc_10_1_nv14[] =
-{
+static const struct soc15_reg_golden golden_settings_gc_10_1_nv14[] = {
 	/* Pending on emulation bring up */
 };
 
-static const struct soc15_reg_golden golden_settings_gc_rlc_spm_10_1_nv14[] =
-{
+static const struct soc15_reg_golden golden_settings_gc_rlc_spm_10_1_nv14[] = {
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGRBM_GFX_INDEX, 0xE0000000L, 0x0),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGRBM_GFX_INDEX, 0xffffff, 0x0),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmRLC_SPM_GLB_SAMPLEDELAY_IND_ADDR, 0xFFFFFFFF, 0x28),
@@ -2229,13 +2295,11 @@ static const struct soc15_reg_golden golden_settings_gc_rlc_spm_10_1_nv14[] =
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGRBM_GFX_INDEX, 0xFFFFFFFF, 0xe0000000)
 };
 
-static const struct soc15_reg_golden golden_settings_gc_10_1_2_nv12[] =
-{
+static const struct soc15_reg_golden golden_settings_gc_10_1_2_nv12[] = {
 	/* Pending on emulation bring up */
 };
 
-static const struct soc15_reg_golden golden_settings_gc_rlc_spm_10_1_2_nv12[] =
-{
+static const struct soc15_reg_golden golden_settings_gc_rlc_spm_10_1_2_nv12[] = {
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGRBM_GFX_INDEX, 0xe0000000L, 0x0),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGRBM_GFX_INDEX, 0xffffff, 0x0),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmRLC_SPM_GLB_SAMPLEDELAY_IND_ADDR, 0xFFFFFFFF, 0x28),
@@ -3290,8 +3354,7 @@ static const struct soc15_reg_golden golden_settings_gc_rlc_spm_10_1_2_nv12[] =
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGRBM_GFX_INDEX, 0xFFFFFFFF, 0xe0000000)
 };
 
-static const struct soc15_reg_golden golden_settings_gc_10_3[] =
-{
+static const struct soc15_reg_golden golden_settings_gc_10_3[] = {
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmCGTT_SPI_CS_CLK_CTRL, 0x78000000, 0x78000100),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmCGTT_SPI_PS_CLK_CTRL, 0xff7f0fff, 0x78000100),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmCGTT_SPI_RA0_CLK_CTRL, 0xff7f0fff, 0x30000100),
@@ -3300,7 +3363,7 @@ static const struct soc15_reg_golden golden_settings_gc_10_3[] =
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmDB_DEBUG3, 0xffffffff, 0x00000280),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmDB_DEBUG4, 0xffffffff, 0x00800000),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmDB_EXCEPTION_CONTROL, 0x7fff0f1f, 0x00b80000),
-	SOC15_REG_GOLDEN_VALUE(GC, 0 ,mmGCEA_SDP_TAG_RESERVE0, 0xffffffff, 0x10100100),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGCEA_SDP_TAG_RESERVE0, 0xffffffff, 0x10100100),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGCEA_SDP_TAG_RESERVE1, 0xffffffff, 0x17000088),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGCR_GENERAL_CNTL_Sienna_Cichlid, 0x1ff1ffff, 0x00000500),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGCUTCL2_CGTT_CLK_CTRL_Sienna_Cichlid, 0xff000000, 0xff008080),
@@ -3337,13 +3400,11 @@ static const struct soc15_reg_golden golden_settings_gc_10_3[] =
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmUTCL1_CTRL, 0xffbfffff, 0x00a00000)
 };
 
-static const struct soc15_reg_golden golden_settings_gc_10_3_sienna_cichlid[] =
-{
+static const struct soc15_reg_golden golden_settings_gc_10_3_sienna_cichlid[] = {
 	/* Pending on emulation bring up */
 };
 
-static const struct soc15_reg_golden golden_settings_gc_10_3_2[] =
-{
+static const struct soc15_reg_golden golden_settings_gc_10_3_2[] = {
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmCGTT_SPI_CS_CLK_CTRL, 0xff7f0fff, 0x78000100),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmCGTT_SPI_PS_CLK_CTRL, 0xff7f0fff, 0x78000100),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmCGTT_SPI_RA0_CLK_CTRL, 0xff7f0fff, 0x30000100),
@@ -3390,8 +3451,7 @@ static const struct soc15_reg_golden golden_settings_gc_10_3_2[] =
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmLDS_CONFIG,  0x00000020, 0x00000020),
 };
 
-static const struct soc15_reg_golden golden_settings_gc_10_3_vangogh[] =
-{
+static const struct soc15_reg_golden golden_settings_gc_10_3_vangogh[] = {
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmCGTT_SPI_RA0_CLK_CTRL, 0xff7f0fff, 0x30000100),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmCGTT_SPI_RA1_CLK_CTRL, 0xff7f0fff, 0x7e000100),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmCH_PIPE_STEER, 0x000000ff, 0x000000e4),
@@ -3421,15 +3481,14 @@ static const struct soc15_reg_golden golden_settings_gc_10_3_vangogh[] =
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmLDS_CONFIG,  0x00000020, 0x00000020),
 };
 
-static const struct soc15_reg_golden golden_settings_gc_10_3_3[] =
-{
+static const struct soc15_reg_golden golden_settings_gc_10_3_3[] = {
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmCGTT_SPI_CS_CLK_CTRL, 0xff7f0fff, 0x78000100),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmCH_PIPE_STEER, 0x000000ff, 0x000000e4),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmCPF_GCR_CNTL, 0x0007ffff, 0x0000c200),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmDB_DEBUG3, 0xffffffff, 0x00000280),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmDB_DEBUG4, 0xffffffff, 0x00800000),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGB_ADDR_CONFIG, 0x0c1807ff, 0x00000242),
-	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGCR_GENERAL_CNTL, 0x1ff1ffff, 0x00000500),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGCR_GENERAL_CNTL_Vangogh, 0x1ff1ffff, 0x00000500),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGL1_PIPE_STEER, 0x000000ff, 0x000000e4),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGL2_PIPE_STEER_0, 0x77777777, 0x32103210),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGL2_PIPE_STEER_1, 0x77777777, 0x32103210),
@@ -3445,8 +3504,7 @@ static const struct soc15_reg_golden golden_settings_gc_10_3_3[] =
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmUTCL1_CTRL, 0xffffffff, 0x00100000)
 };
 
-static const struct soc15_reg_golden golden_settings_gc_10_3_4[] =
-{
+static const struct soc15_reg_golden golden_settings_gc_10_3_4[] = {
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmCGTT_SPI_CS_CLK_CTRL, 0x78000000, 0x78000100),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmCGTT_SPI_RA0_CLK_CTRL, 0x30000000, 0x30000100),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmCGTT_SPI_RA1_CLK_CTRL, 0x7e000000, 0x7e000100),
@@ -3516,7 +3574,7 @@ static const struct soc15_reg_golden golden_settings_gc_10_3_5[] = {
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_PERFCOUNTER7_SELECT, 0xf0f001ff, 0x00000000),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_PERFCOUNTER8_SELECT, 0xf0f001ff, 0x00000000),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQ_PERFCOUNTER9_SELECT, 0xf0f001ff, 0x00000000),
-	SOC15_REG_GOLDEN_VALUE(GC, 0, mmTA_CNTL_AUX,0xfff7ffff, 0x01030000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmTA_CNTL_AUX, 0xfff7ffff, 0x01030000),
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmUTCL1_CTRL, 0xffbfffff, 0x00a00000)
 };
 
@@ -3557,6 +3615,56 @@ static const struct soc15_reg_golden golden_settings_gc_10_0_cyan_skillfish[] =
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmUTCL1_CTRL, 0x00800000, 0x00800000)
 };
 
+static const struct soc15_reg_golden golden_settings_gc_10_3_6[] = {
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmCGTT_SPI_CS_CLK_CTRL, 0xff7f0fff, 0x78000100),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmCH_PIPE_STEER, 0x000000ff, 0x00000044),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmCPF_GCR_CNTL, 0x0007ffff, 0x0000c200),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmDB_DEBUG3, 0xffffffff, 0x00000280),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmDB_DEBUG4, 0xffffffff, 0x00800000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGB_ADDR_CONFIG, 0x0c1807ff, 0x00000042),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGCR_GENERAL_CNTL_Vangogh, 0x1ff1ffff, 0x00000500),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGL1_PIPE_STEER, 0x000000ff, 0x00000044),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGL2_PIPE_STEER_0, 0x77777777, 0x32103210),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGL2_PIPE_STEER_1, 0x77777777, 0x32103210),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGL2A_ADDR_MATCH_MASK, 0xffffffff, 0xfffffff3),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGL2C_ADDR_MATCH_MASK, 0xffffffff, 0xfffffff3),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGL2C_CM_CTRL1, 0xff8fff0f, 0x580f1008),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGL2C_CTRL3, 0xf7ffffff, 0x00f80988),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmLDS_CONFIG, 0x000001ff, 0x00000020),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmPA_CL_ENHANCE, 0xf17fffff, 0x01200007),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmPA_SC_BINNER_TIMEOUT_COUNTER, 0xffffffff, 0x00000800),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmPA_SC_ENHANCE_2, 0xffffffbf, 0x00000820),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQG_CONFIG, 0x000017ff, 0x00001000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSX_DEBUG_1, 0xffffff7f, 0x00010020),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmTA_CNTL_AUX, 0xfff7ffff, 0x01030000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmUTCL1_CTRL, 0xffffffff, 0x00100000)
+};
+
+static const struct soc15_reg_golden golden_settings_gc_10_3_7[] = {
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmCGTT_SPI_CS_CLK_CTRL, 0xff7f0fff, 0x78000100),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmCH_PIPE_STEER, 0x000000ff, 0x000000e4),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmCPF_GCR_CNTL, 0x0007ffff, 0x0000c200),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmDB_DEBUG3, 0xffffffff, 0x00000280),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmDB_DEBUG4, 0xffffffff, 0x00800000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGB_ADDR_CONFIG, 0x0c1807ff, 0x00000041),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGCR_GENERAL_CNTL_Vangogh, 0x1ff1ffff, 0x00000500),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGL1_PIPE_STEER, 0x000000ff, 0x000000e4),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGL2_PIPE_STEER_0, 0x77777777, 0x32103210),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGL2_PIPE_STEER_1, 0x77777777, 0x32103210),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGL2A_ADDR_MATCH_MASK, 0xffffffff, 0xffffffff),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGL2C_ADDR_MATCH_MASK, 0xffffffff, 0xffffffff),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGL2C_CM_CTRL1, 0xff8fff0f, 0x580f1008),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGL2C_CTRL3, 0xf7ffffff, 0x00f80988),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmLDS_CONFIG, 0x000001ff, 0x00000020),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmPA_CL_ENHANCE, 0xf000003f, 0x01200007),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmPA_SC_BINNER_TIMEOUT_COUNTER, 0xffffffff, 0x00000800),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmPA_SC_ENHANCE_2, 0xffffffbf, 0x00000820),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSQG_CONFIG, 0x000017ff, 0x00001000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSX_DEBUG_1, 0xffffff7f, 0x00010020),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmTA_CNTL_AUX, 0xfff7ffff, 0x01030000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmUTCL1_CTRL, 0xffffffff, 0x00100000)
+};
+
 #define DEFAULT_SH_MEM_CONFIG \
 	((SH_MEM_ADDRESS_MODE_64 << SH_MEM_CONFIG__ADDRESS_MODE__SHIFT) | \
 	 (SH_MEM_ALIGNMENT_MODE_UNALIGNED << SH_MEM_CONFIG__ALIGNMENT_MODE__SHIFT) | \
@@ -3570,11 +3678,12 @@ static void gfx_v10_0_set_ring_funcs(struct amdgpu_device *adev);
 static void gfx_v10_0_set_irq_funcs(struct amdgpu_device *adev);
 static void gfx_v10_0_set_gds_init(struct amdgpu_device *adev);
 static void gfx_v10_0_set_rlc_funcs(struct amdgpu_device *adev);
+static void gfx_v10_0_set_mqd_funcs(struct amdgpu_device *adev);
 static int gfx_v10_0_get_cu_info(struct amdgpu_device *adev,
 				 struct amdgpu_cu_info *cu_info);
 static uint64_t gfx_v10_0_get_gpu_clock_counter(struct amdgpu_device *adev);
 static void gfx_v10_0_select_se_sh(struct amdgpu_device *adev, u32 se_num,
-				   u32 sh_num, u32 instance);
+				   u32 sh_num, u32 instance, int xcc_id);
 static u32 gfx_v10_0_get_wgp_active_bitmap_per_sh(struct amdgpu_device *adev);
 
 static int gfx_v10_0_rlc_backdoor_autoload_buffer_init(struct amdgpu_device *adev);
@@ -3587,16 +3696,29 @@ static void gfx_v10_0_ring_emit_frame_cntl(struct amdgpu_ring *ring, bool start,
 static u32 gfx_v10_3_get_disabled_sa(struct amdgpu_device *adev);
 static void gfx_v10_3_program_pbb_mode(struct amdgpu_device *adev);
 static void gfx_v10_3_set_power_brake_sequence(struct amdgpu_device *adev);
-
+static void gfx_v10_0_ring_invalidate_tlbs(struct amdgpu_ring *ring,
+					   uint16_t pasid, uint32_t flush_type,
+					   bool all_hub, uint8_t dst_sel);
+static void gfx_v10_0_update_spm_vmid_internal(struct amdgpu_device *adev,
+					       unsigned int vmid);
+
+static int gfx_v10_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
+					  enum amd_powergating_state state);
 static void gfx10_kiq_set_resources(struct amdgpu_ring *kiq_ring, uint64_t queue_mask)
 {
+	struct amdgpu_device *adev = kiq_ring->adev;
+	u64 shader_mc_addr;
+
+	/* Cleaner shader MC address */
+	shader_mc_addr = adev->gfx.cleaner_shader_gpu_addr >> 8;
+
 	amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_SET_RESOURCES, 6));
 	amdgpu_ring_write(kiq_ring, PACKET3_SET_RESOURCES_VMID_MASK(0) |
 			  PACKET3_SET_RESOURCES_QUEUE_TYPE(0));	/* vmid_mask:0 queue_type:0 (KIQ) */
 	amdgpu_ring_write(kiq_ring, lower_32_bits(queue_mask));	/* queue mask lo */
 	amdgpu_ring_write(kiq_ring, upper_32_bits(queue_mask));	/* queue mask hi */
-	amdgpu_ring_write(kiq_ring, 0);	/* gws mask lo */
-	amdgpu_ring_write(kiq_ring, 0);	/* gws mask hi */
+	amdgpu_ring_write(kiq_ring, lower_32_bits(shader_mc_addr)); /* cleaner shader addr lo */
+	amdgpu_ring_write(kiq_ring, upper_32_bits(shader_mc_addr)); /* cleaner shader addr hi */
 	amdgpu_ring_write(kiq_ring, 0);	/* oac mask */
 	amdgpu_ring_write(kiq_ring, 0);	/* gds heap base:0, gds heap size:0 */
 }
@@ -3604,10 +3726,23 @@ static void gfx10_kiq_set_resources(struct amdgpu_ring *kiq_ring, uint64_t queue
 static void gfx10_kiq_map_queues(struct amdgpu_ring *kiq_ring,
 				 struct amdgpu_ring *ring)
 {
-	struct amdgpu_device *adev = kiq_ring->adev;
 	uint64_t mqd_addr = amdgpu_bo_gpu_offset(ring->mqd_obj);
-	uint64_t wptr_addr = adev->wb.gpu_addr + (ring->wptr_offs * 4);
-	uint32_t eng_sel = ring->funcs->type == AMDGPU_RING_TYPE_GFX ? 4 : 0;
+	uint64_t wptr_addr = ring->wptr_gpu_addr;
+	uint32_t eng_sel = 0;
+
+	switch (ring->funcs->type) {
+	case AMDGPU_RING_TYPE_COMPUTE:
+		eng_sel = 0;
+		break;
+	case AMDGPU_RING_TYPE_GFX:
+		eng_sel = 4;
+		break;
+	case AMDGPU_RING_TYPE_MES:
+		eng_sel = 5;
+		break;
+	default:
+		WARN_ON(1);
+	}
 
 	amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_MAP_QUEUES, 5));
 	/* Q_sel:0, vmid:0, vidmem: 1, engine:0, num_Q:1*/
@@ -3680,12 +3815,59 @@ static void gfx10_kiq_invalidate_tlbs(struct amdgpu_ring *kiq_ring,
 				uint16_t pasid, uint32_t flush_type,
 				bool all_hub)
 {
-	amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_INVALIDATE_TLBS, 0));
-	amdgpu_ring_write(kiq_ring,
-			PACKET3_INVALIDATE_TLBS_DST_SEL(1) |
-			PACKET3_INVALIDATE_TLBS_ALL_HUB(all_hub) |
-			PACKET3_INVALIDATE_TLBS_PASID(pasid) |
-			PACKET3_INVALIDATE_TLBS_FLUSH_TYPE(flush_type));
+	gfx_v10_0_ring_invalidate_tlbs(kiq_ring, pasid, flush_type, all_hub, 1);
+}
+
+static void gfx_v10_0_kiq_reset_hw_queue(struct amdgpu_ring *kiq_ring, uint32_t queue_type,
+					uint32_t me_id, uint32_t pipe_id, uint32_t queue_id,
+					uint32_t xcc_id, uint32_t vmid)
+{
+	struct amdgpu_device *adev = kiq_ring->adev;
+	unsigned i;
+	uint32_t tmp;
+
+	/* enter save mode */
+	amdgpu_gfx_rlc_enter_safe_mode(adev, xcc_id);
+	mutex_lock(&adev->srbm_mutex);
+	nv_grbm_select(adev, me_id, pipe_id, queue_id, 0);
+
+	if (queue_type == AMDGPU_RING_TYPE_COMPUTE) {
+		WREG32_SOC15(GC, 0, mmCP_HQD_DEQUEUE_REQUEST, 0x2);
+		WREG32_SOC15(GC, 0, mmSPI_COMPUTE_QUEUE_RESET, 0x1);
+		/* wait till dequeue take effects */
+		for (i = 0; i < adev->usec_timeout; i++) {
+			if (!(RREG32_SOC15(GC, 0, mmCP_HQD_ACTIVE) & 1))
+				break;
+			udelay(1);
+		}
+		if (i >= adev->usec_timeout)
+			dev_err(adev->dev, "fail to wait on hqd deactive\n");
+	} else if (queue_type == AMDGPU_RING_TYPE_GFX) {
+		WREG32_SOC15(GC, 0, mmGRBM_GFX_INDEX,
+			     (uint32_t)(0x1 << GRBM_GFX_INDEX__SE_BROADCAST_WRITES__SHIFT));
+		tmp = REG_SET_FIELD(0, CP_VMID_RESET, RESET_REQUEST, 1 << vmid);
+		if (pipe_id == 0)
+			tmp = REG_SET_FIELD(tmp, CP_VMID_RESET, PIPE0_QUEUES, 1 << queue_id);
+		else
+			tmp = REG_SET_FIELD(tmp, CP_VMID_RESET, PIPE1_QUEUES, 1 << queue_id);
+		WREG32_SOC15(GC, 0, mmCP_VMID_RESET, tmp);
+
+		/* wait till dequeue take effects */
+		for (i = 0; i < adev->usec_timeout; i++) {
+			if (!(RREG32_SOC15(GC, 0, mmCP_GFX_HQD_ACTIVE) & 1))
+				break;
+			udelay(1);
+		}
+		if (i >= adev->usec_timeout)
+			dev_err(adev->dev, "failed to wait on gfx hqd deactivate\n");
+	} else {
+		dev_err(adev->dev, "reset queue_type(%d) not supported\n", queue_type);
+	}
+
+	nv_grbm_select(adev, 0, 0, 0, 0);
+	mutex_unlock(&adev->srbm_mutex);
+	/* exit safe mode */
+	amdgpu_gfx_rlc_exit_safe_mode(adev, 0);
 }
 
 static const struct kiq_pm4_funcs gfx_v10_0_kiq_pm4_funcs = {
@@ -3694,6 +3876,7 @@ static const struct kiq_pm4_funcs gfx_v10_0_kiq_pm4_funcs = {
 	.kiq_unmap_queues = gfx10_kiq_unmap_queues,
 	.kiq_query_status = gfx10_kiq_query_status,
 	.kiq_invalidate_tlbs = gfx10_kiq_invalidate_tlbs,
+	.kiq_reset_hw_queue = gfx_v10_0_kiq_reset_hw_queue,
 	.set_resources_size = 8,
 	.map_queues_size = 7,
 	.unmap_queues_size = 6,
@@ -3703,12 +3886,12 @@ static const struct kiq_pm4_funcs gfx_v10_0_kiq_pm4_funcs = {
 
 static void gfx_v10_0_set_kiq_pm4_funcs(struct amdgpu_device *adev)
 {
-	adev->gfx.kiq.pmf = &gfx_v10_0_kiq_pm4_funcs;
+	adev->gfx.kiq[0].pmf = &gfx_v10_0_kiq_pm4_funcs;
 }
 
 static void gfx_v10_0_init_spm_golden_registers(struct amdgpu_device *adev)
 {
-	switch (adev->ip_versions[GC_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 	case IP_VERSION(10, 1, 10):
 		soc15_program_register_sequence(adev,
 						golden_settings_gc_rlc_spm_10_0_nv10,
@@ -3731,7 +3914,10 @@ static void gfx_v10_0_init_spm_golden_registers(struct amdgpu_device *adev)
 
 static void gfx_v10_0_init_golden_registers(struct amdgpu_device *adev)
 {
-	switch (adev->ip_versions[GC_HWIP][0]) {
+	if (amdgpu_sriov_vf(adev))
+		return;
+
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 	case IP_VERSION(10, 1, 10):
 		soc15_program_register_sequence(adev,
 						golden_settings_gc_10_1,
@@ -3781,8 +3967,8 @@ static void gfx_v10_0_init_golden_registers(struct amdgpu_device *adev)
 		break;
 	case IP_VERSION(10, 3, 4):
 		soc15_program_register_sequence(adev,
-                                                golden_settings_gc_10_3_4,
-                                                (const u32)ARRAY_SIZE(golden_settings_gc_10_3_4));
+						golden_settings_gc_10_3_4,
+						(const u32)ARRAY_SIZE(golden_settings_gc_10_3_4));
 		break;
 	case IP_VERSION(10, 3, 5):
 		soc15_program_register_sequence(adev,
@@ -3790,23 +3976,27 @@ static void gfx_v10_0_init_golden_registers(struct amdgpu_device *adev)
 						(const u32)ARRAY_SIZE(golden_settings_gc_10_3_5));
 		break;
 	case IP_VERSION(10, 1, 3):
+	case IP_VERSION(10, 1, 4):
 		soc15_program_register_sequence(adev,
 						golden_settings_gc_10_0_cyan_skillfish,
 						(const u32)ARRAY_SIZE(golden_settings_gc_10_0_cyan_skillfish));
 		break;
+	case IP_VERSION(10, 3, 6):
+		soc15_program_register_sequence(adev,
+						golden_settings_gc_10_3_6,
+						(const u32)ARRAY_SIZE(golden_settings_gc_10_3_6));
+		break;
+	case IP_VERSION(10, 3, 7):
+		soc15_program_register_sequence(adev,
+						golden_settings_gc_10_3_7,
+						(const u32)ARRAY_SIZE(golden_settings_gc_10_3_7));
+		break;
 	default:
 		break;
 	}
 	gfx_v10_0_init_spm_golden_registers(adev);
 }
 
-static void gfx_v10_0_scratch_init(struct amdgpu_device *adev)
-{
-	adev->gfx.scratch.num_reg = 8;
-	adev->gfx.scratch.reg_base = SOC15_REG_OFFSET(GC, 0, mmSCRATCH_REG0);
-	adev->gfx.scratch.free_mask = (1u << adev->gfx.scratch.num_reg) - 1;
-}
-
 static void gfx_v10_0_write_data_to_reg(struct amdgpu_ring *ring, int eng_sel,
 				       bool wc, uint32_t reg, uint32_t val)
 {
@@ -3843,29 +4033,22 @@ static void gfx_v10_0_wait_reg_mem(struct amdgpu_ring *ring, int eng_sel,
 static int gfx_v10_0_ring_test_ring(struct amdgpu_ring *ring)
 {
 	struct amdgpu_device *adev = ring->adev;
-	uint32_t scratch;
+	uint32_t scratch = SOC15_REG_OFFSET(GC, 0, mmSCRATCH_REG0);
 	uint32_t tmp = 0;
-	unsigned i;
+	unsigned int i;
 	int r;
 
-	r = amdgpu_gfx_scratch_get(adev, &scratch);
-	if (r) {
-		DRM_ERROR("amdgpu: cp failed to get scratch reg (%d).\n", r);
-		return r;
-	}
-
 	WREG32(scratch, 0xCAFEDEAD);
-
 	r = amdgpu_ring_alloc(ring, 3);
 	if (r) {
 		DRM_ERROR("amdgpu: cp failed to lock ring %d (%d).\n",
 			  ring->idx, r);
-		amdgpu_gfx_scratch_free(adev, scratch);
 		return r;
 	}
 
 	amdgpu_ring_write(ring, PACKET3(PACKET3_SET_UCONFIG_REG, 1));
-	amdgpu_ring_write(ring, (scratch - PACKET3_SET_UCONFIG_REG_START));
+	amdgpu_ring_write(ring, scratch -
+			  PACKET3_SET_UCONFIG_REG_START);
 	amdgpu_ring_write(ring, 0xDEADBEEF);
 	amdgpu_ring_commit(ring);
 
@@ -3882,8 +4065,6 @@ static int gfx_v10_0_ring_test_ring(struct amdgpu_ring *ring)
 	if (i >= adev->usec_timeout)
 		r = -ETIMEDOUT;
 
-	amdgpu_gfx_scratch_free(adev, scratch);
-
 	return r;
 }
 
@@ -3892,22 +4073,26 @@ static int gfx_v10_0_ring_test_ib(struct amdgpu_ring *ring, long timeout)
 	struct amdgpu_device *adev = ring->adev;
 	struct amdgpu_ib ib;
 	struct dma_fence *f = NULL;
-	unsigned index;
+	unsigned int index;
 	uint64_t gpu_addr;
-	uint32_t tmp;
+	uint32_t *cpu_ptr;
 	long r;
 
+	memset(&ib, 0, sizeof(ib));
+
 	r = amdgpu_device_wb_get(adev, &index);
 	if (r)
 		return r;
 
 	gpu_addr = adev->wb.gpu_addr + (index * 4);
 	adev->wb.wb[index] = cpu_to_le32(0xCAFEDEAD);
-	memset(&ib, 0, sizeof(ib));
-	r = amdgpu_ib_get(adev, NULL, 16,
-					AMDGPU_IB_POOL_DIRECT, &ib);
-	if (r)
+	cpu_ptr = &adev->wb.wb[index];
+
+	r = amdgpu_ib_get(adev, NULL, 20, AMDGPU_IB_POOL_DIRECT, &ib);
+	if (r) {
+		DRM_ERROR("amdgpu: failed to get ib (%ld).\n", r);
 		goto err1;
+	}
 
 	ib.ptr[0] = PACKET3(PACKET3_WRITE_DATA, 3);
 	ib.ptr[1] = WRITE_DATA_DST_SEL(5) | WR_CONFIRM;
@@ -3928,13 +4113,12 @@ static int gfx_v10_0_ring_test_ib(struct amdgpu_ring *ring, long timeout)
 		goto err2;
 	}
 
-	tmp = adev->wb.wb[index];
-	if (tmp == 0xDEADBEEF)
+	if (le32_to_cpu(*cpu_ptr) == 0xDEADBEEF)
 		r = 0;
 	else
 		r = -EINVAL;
 err2:
-	amdgpu_ib_free(adev, &ib, NULL);
+	amdgpu_ib_free(&ib, NULL);
 	dma_fence_put(f);
 err1:
 	amdgpu_device_wb_free(adev, index);
@@ -3943,18 +4127,12 @@ err1:
 
 static void gfx_v10_0_free_microcode(struct amdgpu_device *adev)
 {
-	release_firmware(adev->gfx.pfp_fw);
-	adev->gfx.pfp_fw = NULL;
-	release_firmware(adev->gfx.me_fw);
-	adev->gfx.me_fw = NULL;
-	release_firmware(adev->gfx.ce_fw);
-	adev->gfx.ce_fw = NULL;
-	release_firmware(adev->gfx.rlc_fw);
-	adev->gfx.rlc_fw = NULL;
-	release_firmware(adev->gfx.mec_fw);
-	adev->gfx.mec_fw = NULL;
-	release_firmware(adev->gfx.mec2_fw);
-	adev->gfx.mec2_fw = NULL;
+	amdgpu_ucode_release(&adev->gfx.pfp_fw);
+	amdgpu_ucode_release(&adev->gfx.me_fw);
+	amdgpu_ucode_release(&adev->gfx.ce_fw);
+	amdgpu_ucode_release(&adev->gfx.rlc_fw);
+	amdgpu_ucode_release(&adev->gfx.mec_fw);
+	amdgpu_ucode_release(&adev->gfx.mec2_fw);
 
 	kfree(adev->gfx.rlc.register_list_format);
 }
@@ -3963,11 +4141,12 @@ static void gfx_v10_0_check_fw_write_wait(struct amdgpu_device *adev)
 {
 	adev->gfx.cp_fw_write_wait = false;
 
-	switch (adev->ip_versions[GC_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 	case IP_VERSION(10, 1, 10):
 	case IP_VERSION(10, 1, 2):
 	case IP_VERSION(10, 1, 1):
 	case IP_VERSION(10, 1, 3):
+	case IP_VERSION(10, 1, 4):
 		if ((adev->gfx.me_fw_version >= 0x00000046) &&
 		    (adev->gfx.me_feature_version >= 27) &&
 		    (adev->gfx.pfp_fw_version >= 0x00000068) &&
@@ -3981,7 +4160,9 @@ static void gfx_v10_0_check_fw_write_wait(struct amdgpu_device *adev)
 	case IP_VERSION(10, 3, 1):
 	case IP_VERSION(10, 3, 4):
 	case IP_VERSION(10, 3, 5):
+	case IP_VERSION(10, 3, 6):
 	case IP_VERSION(10, 3, 3):
+	case IP_VERSION(10, 3, 7):
 		adev->gfx.cp_fw_write_wait = true;
 		break;
 	default:
@@ -3992,39 +4173,6 @@ static void gfx_v10_0_check_fw_write_wait(struct amdgpu_device *adev)
 		DRM_WARN_ONCE("CP firmware version too old, please update!");
 }
 
-
-static void gfx_v10_0_init_rlc_ext_microcode(struct amdgpu_device *adev)
-{
-	const struct rlc_firmware_header_v2_1 *rlc_hdr;
-
-	rlc_hdr = (const struct rlc_firmware_header_v2_1 *)adev->gfx.rlc_fw->data;
-	adev->gfx.rlc_srlc_fw_version = le32_to_cpu(rlc_hdr->save_restore_list_cntl_ucode_ver);
-	adev->gfx.rlc_srlc_feature_version = le32_to_cpu(rlc_hdr->save_restore_list_cntl_feature_ver);
-	adev->gfx.rlc.save_restore_list_cntl_size_bytes = le32_to_cpu(rlc_hdr->save_restore_list_cntl_size_bytes);
-	adev->gfx.rlc.save_restore_list_cntl = (u8 *)rlc_hdr + le32_to_cpu(rlc_hdr->save_restore_list_cntl_offset_bytes);
-	adev->gfx.rlc_srlg_fw_version = le32_to_cpu(rlc_hdr->save_restore_list_gpm_ucode_ver);
-	adev->gfx.rlc_srlg_feature_version = le32_to_cpu(rlc_hdr->save_restore_list_gpm_feature_ver);
-	adev->gfx.rlc.save_restore_list_gpm_size_bytes = le32_to_cpu(rlc_hdr->save_restore_list_gpm_size_bytes);
-	adev->gfx.rlc.save_restore_list_gpm = (u8 *)rlc_hdr + le32_to_cpu(rlc_hdr->save_restore_list_gpm_offset_bytes);
-	adev->gfx.rlc_srls_fw_version = le32_to_cpu(rlc_hdr->save_restore_list_srm_ucode_ver);
-	adev->gfx.rlc_srls_feature_version = le32_to_cpu(rlc_hdr->save_restore_list_srm_feature_ver);
-	adev->gfx.rlc.save_restore_list_srm_size_bytes = le32_to_cpu(rlc_hdr->save_restore_list_srm_size_bytes);
-	adev->gfx.rlc.save_restore_list_srm = (u8 *)rlc_hdr + le32_to_cpu(rlc_hdr->save_restore_list_srm_offset_bytes);
-	adev->gfx.rlc.reg_list_format_direct_reg_list_length =
-			le32_to_cpu(rlc_hdr->reg_list_format_direct_reg_list_length);
-}
-
-static void gfx_v10_0_init_rlc_iram_dram_microcode(struct amdgpu_device *adev)
-{
-	const struct rlc_firmware_header_v2_2 *rlc_hdr;
-
-	rlc_hdr = (const struct rlc_firmware_header_v2_2 *)adev->gfx.rlc_fw->data;
-	adev->gfx.rlc.rlc_iram_ucode_size_bytes = le32_to_cpu(rlc_hdr->rlc_iram_ucode_size_bytes);
-	adev->gfx.rlc.rlc_iram_ucode = (u8 *)rlc_hdr + le32_to_cpu(rlc_hdr->rlc_iram_ucode_offset_bytes);
-	adev->gfx.rlc.rlc_dram_ucode_size_bytes = le32_to_cpu(rlc_hdr->rlc_dram_ucode_size_bytes);
-	adev->gfx.rlc.rlc_dram_ucode = (u8 *)rlc_hdr + le32_to_cpu(rlc_hdr->rlc_dram_ucode_offset_bytes);
-}
-
 static bool gfx_v10_0_navi10_gfxoff_should_enable(struct amdgpu_device *adev)
 {
 	bool ret = false;
@@ -4039,12 +4187,12 @@ static bool gfx_v10_0_navi10_gfxoff_should_enable(struct amdgpu_device *adev)
 		break;
 	}
 
-	return ret ;
+	return ret;
 }
 
 static void gfx_v10_0_check_gfxoff_flag(struct amdgpu_device *adev)
 {
-	switch (adev->ip_versions[GC_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 	case IP_VERSION(10, 1, 10):
 		if (!gfx_v10_0_navi10_gfxoff_should_enable(adev))
 			adev->pm.pp_feature &= ~PP_GFXOFF_MASK;
@@ -4056,300 +4204,87 @@ static void gfx_v10_0_check_gfxoff_flag(struct amdgpu_device *adev)
 
 static int gfx_v10_0_init_microcode(struct amdgpu_device *adev)
 {
-	const char *chip_name;
-	char fw_name[40];
-	char *wks = "";
+	char fw_name[53];
+	char ucode_prefix[30];
+	const char *wks = "";
 	int err;
-	struct amdgpu_firmware_info *info = NULL;
-	const struct common_firmware_header *header = NULL;
-	const struct gfx_firmware_header_v1_0 *cp_hdr;
 	const struct rlc_firmware_header_v2_0 *rlc_hdr;
-	unsigned int *tmp = NULL;
-	unsigned int i = 0;
 	uint16_t version_major;
 	uint16_t version_minor;
 
 	DRM_DEBUG("\n");
 
-	switch (adev->ip_versions[GC_HWIP][0]) {
-	case IP_VERSION(10, 1, 10):
-		chip_name = "navi10";
-		break;
-	case IP_VERSION(10, 1, 1):
-		chip_name = "navi14";
-		if (!(adev->pdev->device == 0x7340 &&
-		      adev->pdev->revision != 0x00))
-			wks = "_wks";
-		break;
-	case IP_VERSION(10, 1, 2):
-		chip_name = "navi12";
-		break;
-	case IP_VERSION(10, 3, 0):
-		chip_name = "sienna_cichlid";
-		break;
-	case IP_VERSION(10, 3, 2):
-		chip_name = "navy_flounder";
-		break;
-	case IP_VERSION(10, 3, 1):
-		chip_name = "vangogh";
-		break;
-	case IP_VERSION(10, 3, 4):
-		chip_name = "dimgrey_cavefish";
-		break;
-	case IP_VERSION(10, 3, 5):
-		chip_name = "beige_goby";
-		break;
-	case IP_VERSION(10, 3, 3):
-		chip_name = "yellow_carp";
-		break;
-	case IP_VERSION(10, 1, 3):
-		if (adev->apu_flags & AMD_APU_IS_CYAN_SKILLFISH2)
-			chip_name = "cyan_skillfish2";
-		else
-			chip_name = "cyan_skillfish";
-		break;
-	default:
-		BUG();
-	}
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(10, 1, 1) &&
+	    (!(adev->pdev->device == 0x7340 && adev->pdev->revision != 0x00)))
+		wks = "_wks";
+	amdgpu_ucode_ip_version_decode(adev, GC_HWIP, ucode_prefix, sizeof(ucode_prefix));
 
-	snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_pfp%s.bin", chip_name, wks);
-	err = request_firmware(&adev->gfx.pfp_fw, fw_name, adev->dev);
+	err = amdgpu_ucode_request(adev, &adev->gfx.pfp_fw,
+				   AMDGPU_UCODE_REQUIRED,
+				   "amdgpu/%s_pfp%s.bin", ucode_prefix, wks);
 	if (err)
 		goto out;
-	err = amdgpu_ucode_validate(adev->gfx.pfp_fw);
-	if (err)
-		goto out;
-	cp_hdr = (const struct gfx_firmware_header_v1_0 *)adev->gfx.pfp_fw->data;
-	adev->gfx.pfp_fw_version = le32_to_cpu(cp_hdr->header.ucode_version);
-	adev->gfx.pfp_feature_version = le32_to_cpu(cp_hdr->ucode_feature_version);
+	amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_PFP);
 
-	snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_me%s.bin", chip_name, wks);
-	err = request_firmware(&adev->gfx.me_fw, fw_name, adev->dev);
-	if (err)
-		goto out;
-	err = amdgpu_ucode_validate(adev->gfx.me_fw);
+	err = amdgpu_ucode_request(adev, &adev->gfx.me_fw,
+				   AMDGPU_UCODE_REQUIRED,
+				   "amdgpu/%s_me%s.bin", ucode_prefix, wks);
 	if (err)
 		goto out;
-	cp_hdr = (const struct gfx_firmware_header_v1_0 *)adev->gfx.me_fw->data;
-	adev->gfx.me_fw_version = le32_to_cpu(cp_hdr->header.ucode_version);
-	adev->gfx.me_feature_version = le32_to_cpu(cp_hdr->ucode_feature_version);
+	amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_ME);
 
-	snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_ce%s.bin", chip_name, wks);
-	err = request_firmware(&adev->gfx.ce_fw, fw_name, adev->dev);
-	if (err)
-		goto out;
-	err = amdgpu_ucode_validate(adev->gfx.ce_fw);
+	err = amdgpu_ucode_request(adev, &adev->gfx.ce_fw,
+				   AMDGPU_UCODE_REQUIRED,
+				   "amdgpu/%s_ce%s.bin", ucode_prefix, wks);
 	if (err)
 		goto out;
-	cp_hdr = (const struct gfx_firmware_header_v1_0 *)adev->gfx.ce_fw->data;
-	adev->gfx.ce_fw_version = le32_to_cpu(cp_hdr->header.ucode_version);
-	adev->gfx.ce_feature_version = le32_to_cpu(cp_hdr->ucode_feature_version);
+	amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_CE);
 
 	if (!amdgpu_sriov_vf(adev)) {
-		snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_rlc.bin", chip_name);
+		snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_rlc.bin", ucode_prefix);
 		err = request_firmware(&adev->gfx.rlc_fw, fw_name, adev->dev);
 		if (err)
 			goto out;
-		err = amdgpu_ucode_validate(adev->gfx.rlc_fw);
+
+		/* don't validate this firmware. There are apparently firmwares
+		 * in the wild with incorrect size in the header
+		 */
 		rlc_hdr = (const struct rlc_firmware_header_v2_0 *)adev->gfx.rlc_fw->data;
 		version_major = le16_to_cpu(rlc_hdr->header.header_version_major);
 		version_minor = le16_to_cpu(rlc_hdr->header.header_version_minor);
-
-		adev->gfx.rlc_fw_version = le32_to_cpu(rlc_hdr->header.ucode_version);
-		adev->gfx.rlc_feature_version = le32_to_cpu(rlc_hdr->ucode_feature_version);
-		adev->gfx.rlc.save_and_restore_offset =
-			le32_to_cpu(rlc_hdr->save_and_restore_offset);
-		adev->gfx.rlc.clear_state_descriptor_offset =
-			le32_to_cpu(rlc_hdr->clear_state_descriptor_offset);
-		adev->gfx.rlc.avail_scratch_ram_locations =
-			le32_to_cpu(rlc_hdr->avail_scratch_ram_locations);
-		adev->gfx.rlc.reg_restore_list_size =
-			le32_to_cpu(rlc_hdr->reg_restore_list_size);
-		adev->gfx.rlc.reg_list_format_start =
-			le32_to_cpu(rlc_hdr->reg_list_format_start);
-		adev->gfx.rlc.reg_list_format_separate_start =
-			le32_to_cpu(rlc_hdr->reg_list_format_separate_start);
-		adev->gfx.rlc.starting_offsets_start =
-			le32_to_cpu(rlc_hdr->starting_offsets_start);
-		adev->gfx.rlc.reg_list_format_size_bytes =
-			le32_to_cpu(rlc_hdr->reg_list_format_size_bytes);
-		adev->gfx.rlc.reg_list_size_bytes =
-			le32_to_cpu(rlc_hdr->reg_list_size_bytes);
-		adev->gfx.rlc.register_list_format =
-			kmalloc(adev->gfx.rlc.reg_list_format_size_bytes +
-					adev->gfx.rlc.reg_list_size_bytes, GFP_KERNEL);
-		if (!adev->gfx.rlc.register_list_format) {
-			err = -ENOMEM;
+		err = amdgpu_gfx_rlc_init_microcode(adev, version_major, version_minor);
+		if (err)
 			goto out;
-		}
-
-		tmp = (unsigned int *)((uintptr_t)rlc_hdr +
-							   le32_to_cpu(rlc_hdr->reg_list_format_array_offset_bytes));
-		for (i = 0 ; i < (rlc_hdr->reg_list_format_size_bytes >> 2); i++)
-			adev->gfx.rlc.register_list_format[i] =	le32_to_cpu(tmp[i]);
-
-		adev->gfx.rlc.register_restore = adev->gfx.rlc.register_list_format + i;
-
-		tmp = (unsigned int *)((uintptr_t)rlc_hdr +
-							   le32_to_cpu(rlc_hdr->reg_list_array_offset_bytes));
-		for (i = 0 ; i < (rlc_hdr->reg_list_size_bytes >> 2); i++)
-			adev->gfx.rlc.register_restore[i] = le32_to_cpu(tmp[i]);
-
-		if (version_major == 2) {
-			if (version_minor >= 1)
-				gfx_v10_0_init_rlc_ext_microcode(adev);
-			if (version_minor == 2)
-				gfx_v10_0_init_rlc_iram_dram_microcode(adev);
-		}
 	}
 
-	snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_mec%s.bin", chip_name, wks);
-	err = request_firmware(&adev->gfx.mec_fw, fw_name, adev->dev);
-	if (err)
-		goto out;
-	err = amdgpu_ucode_validate(adev->gfx.mec_fw);
+	err = amdgpu_ucode_request(adev, &adev->gfx.mec_fw,
+				   AMDGPU_UCODE_REQUIRED,
+				   "amdgpu/%s_mec%s.bin", ucode_prefix, wks);
 	if (err)
 		goto out;
-	cp_hdr = (const struct gfx_firmware_header_v1_0 *)adev->gfx.mec_fw->data;
-	adev->gfx.mec_fw_version = le32_to_cpu(cp_hdr->header.ucode_version);
-	adev->gfx.mec_feature_version = le32_to_cpu(cp_hdr->ucode_feature_version);
+	amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_MEC1);
+	amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_MEC1_JT);
 
-	snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_mec2%s.bin", chip_name, wks);
-	err = request_firmware(&adev->gfx.mec2_fw, fw_name, adev->dev);
+	err = amdgpu_ucode_request(adev, &adev->gfx.mec2_fw,
+				   AMDGPU_UCODE_REQUIRED,
+				   "amdgpu/%s_mec2%s.bin", ucode_prefix, wks);
 	if (!err) {
-		err = amdgpu_ucode_validate(adev->gfx.mec2_fw);
-		if (err)
-			goto out;
-		cp_hdr = (const struct gfx_firmware_header_v1_0 *)
-		adev->gfx.mec2_fw->data;
-		adev->gfx.mec2_fw_version =
-		le32_to_cpu(cp_hdr->header.ucode_version);
-		adev->gfx.mec2_feature_version =
-		le32_to_cpu(cp_hdr->ucode_feature_version);
+		amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_MEC2);
+		amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_MEC2_JT);
 	} else {
 		err = 0;
 		adev->gfx.mec2_fw = NULL;
 	}
 
-	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
-		info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_PFP];
-		info->ucode_id = AMDGPU_UCODE_ID_CP_PFP;
-		info->fw = adev->gfx.pfp_fw;
-		header = (const struct common_firmware_header *)info->fw->data;
-		adev->firmware.fw_size +=
-			ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE);
-
-		info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_ME];
-		info->ucode_id = AMDGPU_UCODE_ID_CP_ME;
-		info->fw = adev->gfx.me_fw;
-		header = (const struct common_firmware_header *)info->fw->data;
-		adev->firmware.fw_size +=
-			ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE);
-
-		info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_CE];
-		info->ucode_id = AMDGPU_UCODE_ID_CP_CE;
-		info->fw = adev->gfx.ce_fw;
-		header = (const struct common_firmware_header *)info->fw->data;
-		adev->firmware.fw_size +=
-			ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE);
-
-		info = &adev->firmware.ucode[AMDGPU_UCODE_ID_RLC_G];
-		info->ucode_id = AMDGPU_UCODE_ID_RLC_G;
-		info->fw = adev->gfx.rlc_fw;
-		if (info->fw) {
-			header = (const struct common_firmware_header *)info->fw->data;
-			adev->firmware.fw_size +=
-				ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE);
-		}
-		if (adev->gfx.rlc.save_restore_list_cntl_size_bytes &&
-		    adev->gfx.rlc.save_restore_list_gpm_size_bytes &&
-		    adev->gfx.rlc.save_restore_list_srm_size_bytes) {
-			info = &adev->firmware.ucode[AMDGPU_UCODE_ID_RLC_RESTORE_LIST_CNTL];
-			info->ucode_id = AMDGPU_UCODE_ID_RLC_RESTORE_LIST_CNTL;
-			info->fw = adev->gfx.rlc_fw;
-			adev->firmware.fw_size +=
-				ALIGN(adev->gfx.rlc.save_restore_list_cntl_size_bytes, PAGE_SIZE);
-
-			info = &adev->firmware.ucode[AMDGPU_UCODE_ID_RLC_RESTORE_LIST_GPM_MEM];
-			info->ucode_id = AMDGPU_UCODE_ID_RLC_RESTORE_LIST_GPM_MEM;
-			info->fw = adev->gfx.rlc_fw;
-			adev->firmware.fw_size +=
-				ALIGN(adev->gfx.rlc.save_restore_list_gpm_size_bytes, PAGE_SIZE);
-
-			info = &adev->firmware.ucode[AMDGPU_UCODE_ID_RLC_RESTORE_LIST_SRM_MEM];
-			info->ucode_id = AMDGPU_UCODE_ID_RLC_RESTORE_LIST_SRM_MEM;
-			info->fw = adev->gfx.rlc_fw;
-			adev->firmware.fw_size +=
-				ALIGN(adev->gfx.rlc.save_restore_list_srm_size_bytes, PAGE_SIZE);
-
-			if (adev->gfx.rlc.rlc_iram_ucode_size_bytes &&
-			    adev->gfx.rlc.rlc_dram_ucode_size_bytes) {
-				info = &adev->firmware.ucode[AMDGPU_UCODE_ID_RLC_IRAM];
-				info->ucode_id = AMDGPU_UCODE_ID_RLC_IRAM;
-				info->fw = adev->gfx.rlc_fw;
-				adev->firmware.fw_size +=
-					ALIGN(adev->gfx.rlc.rlc_iram_ucode_size_bytes, PAGE_SIZE);
-
-				info = &adev->firmware.ucode[AMDGPU_UCODE_ID_RLC_DRAM];
-				info->ucode_id = AMDGPU_UCODE_ID_RLC_DRAM;
-				info->fw = adev->gfx.rlc_fw;
-				adev->firmware.fw_size +=
-					ALIGN(adev->gfx.rlc.rlc_dram_ucode_size_bytes, PAGE_SIZE);
-			}
-		}
-
-		info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_MEC1];
-		info->ucode_id = AMDGPU_UCODE_ID_CP_MEC1;
-		info->fw = adev->gfx.mec_fw;
-		header = (const struct common_firmware_header *)info->fw->data;
-		cp_hdr = (const struct gfx_firmware_header_v1_0 *)info->fw->data;
-		adev->firmware.fw_size +=
-			ALIGN(le32_to_cpu(header->ucode_size_bytes) -
-			      le32_to_cpu(cp_hdr->jt_size) * 4, PAGE_SIZE);
-
-		info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_MEC1_JT];
-		info->ucode_id = AMDGPU_UCODE_ID_CP_MEC1_JT;
-		info->fw = adev->gfx.mec_fw;
-		adev->firmware.fw_size +=
-			ALIGN(le32_to_cpu(cp_hdr->jt_size) * 4, PAGE_SIZE);
-
-		if (adev->gfx.mec2_fw) {
-			info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_MEC2];
-			info->ucode_id = AMDGPU_UCODE_ID_CP_MEC2;
-			info->fw = adev->gfx.mec2_fw;
-			header = (const struct common_firmware_header *)info->fw->data;
-			cp_hdr = (const struct gfx_firmware_header_v1_0 *)info->fw->data;
-			adev->firmware.fw_size +=
-				ALIGN(le32_to_cpu(header->ucode_size_bytes) -
-				      le32_to_cpu(cp_hdr->jt_size) * 4,
-				      PAGE_SIZE);
-			info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_MEC2_JT];
-			info->ucode_id = AMDGPU_UCODE_ID_CP_MEC2_JT;
-			info->fw = adev->gfx.mec2_fw;
-			adev->firmware.fw_size +=
-				ALIGN(le32_to_cpu(cp_hdr->jt_size) * 4,
-				      PAGE_SIZE);
-		}
-	}
-
 	gfx_v10_0_check_fw_write_wait(adev);
 out:
 	if (err) {
-		dev_err(adev->dev,
-			"gfx10: Failed to load firmware \"%s\"\n",
-			fw_name);
-		release_firmware(adev->gfx.pfp_fw);
-		adev->gfx.pfp_fw = NULL;
-		release_firmware(adev->gfx.me_fw);
-		adev->gfx.me_fw = NULL;
-		release_firmware(adev->gfx.ce_fw);
-		adev->gfx.ce_fw = NULL;
-		release_firmware(adev->gfx.rlc_fw);
-		adev->gfx.rlc_fw = NULL;
-		release_firmware(adev->gfx.mec_fw);
-		adev->gfx.mec_fw = NULL;
-		release_firmware(adev->gfx.mec2_fw);
-		adev->gfx.mec2_fw = NULL;
+		amdgpu_ucode_release(&adev->gfx.pfp_fw);
+		amdgpu_ucode_release(&adev->gfx.me_fw);
+		amdgpu_ucode_release(&adev->gfx.ce_fw);
+		amdgpu_ucode_release(&adev->gfx.rlc_fw);
+		amdgpu_ucode_release(&adev->gfx.mec_fw);
+		amdgpu_ucode_release(&adev->gfx.mec2_fw);
 	}
 
 	gfx_v10_0_check_gfxoff_flag(adev);
@@ -4387,12 +4322,9 @@ static u32 gfx_v10_0_get_csb_size(struct amdgpu_device *adev)
 	return count;
 }
 
-static void gfx_v10_0_get_csb_buffer(struct amdgpu_device *adev,
-				    volatile u32 *buffer)
+static void gfx_v10_0_get_csb_buffer(struct amdgpu_device *adev, u32 *buffer)
 {
-	u32 count = 0, i;
-	const struct cs_section_def *sect = NULL;
-	const struct cs_extent_def *ext = NULL;
+	u32 count = 0;
 	int ctx_reg_offset;
 
 	if (adev->gfx.rlc.cs_data == NULL)
@@ -4400,39 +4332,15 @@ static void gfx_v10_0_get_csb_buffer(struct amdgpu_device *adev,
 	if (buffer == NULL)
 		return;
 
-	buffer[count++] = cpu_to_le32(PACKET3(PACKET3_PREAMBLE_CNTL, 0));
-	buffer[count++] = cpu_to_le32(PACKET3_PREAMBLE_BEGIN_CLEAR_STATE);
-
-	buffer[count++] = cpu_to_le32(PACKET3(PACKET3_CONTEXT_CONTROL, 1));
-	buffer[count++] = cpu_to_le32(0x80000000);
-	buffer[count++] = cpu_to_le32(0x80000000);
+	count = amdgpu_gfx_csb_preamble_start(buffer);
+	count = amdgpu_gfx_csb_data_parser(adev, buffer, count);
 
-	for (sect = adev->gfx.rlc.cs_data; sect->section != NULL; ++sect) {
-		for (ext = sect->section; ext->extent != NULL; ++ext) {
-			if (sect->id == SECT_CONTEXT) {
-				buffer[count++] =
-					cpu_to_le32(PACKET3(PACKET3_SET_CONTEXT_REG, ext->reg_count));
-				buffer[count++] = cpu_to_le32(ext->reg_index -
-						PACKET3_SET_CONTEXT_REG_START);
-				for (i = 0; i < ext->reg_count; i++)
-					buffer[count++] = cpu_to_le32(ext->extent[i]);
-			} else {
-				return;
-			}
-		}
-	}
-
-	ctx_reg_offset =
-		SOC15_REG_OFFSET(GC, 0, mmPA_SC_TILE_STEERING_OVERRIDE) - PACKET3_SET_CONTEXT_REG_START;
+	ctx_reg_offset = SOC15_REG_OFFSET(GC, 0, mmPA_SC_TILE_STEERING_OVERRIDE) - PACKET3_SET_CONTEXT_REG_START;
 	buffer[count++] = cpu_to_le32(PACKET3(PACKET3_SET_CONTEXT_REG, 1));
 	buffer[count++] = cpu_to_le32(ctx_reg_offset);
 	buffer[count++] = cpu_to_le32(adev->gfx.config.pa_sc_tile_steering_override);
 
-	buffer[count++] = cpu_to_le32(PACKET3(PACKET3_PREAMBLE_CNTL, 0));
-	buffer[count++] = cpu_to_le32(PACKET3_PREAMBLE_END_CLEAR_STATE);
-
-	buffer[count++] = cpu_to_le32(PACKET3(PACKET3_CLEAR_STATE, 0));
-	buffer[count++] = cpu_to_le32(0);
+	amdgpu_gfx_csb_preamble_end(buffer, count);
 }
 
 static void gfx_v10_0_rlc_fini(struct amdgpu_device *adev)
@@ -4448,6 +4356,30 @@ static void gfx_v10_0_rlc_fini(struct amdgpu_device *adev)
 			(void **)&adev->gfx.rlc.cp_table_ptr);
 }
 
+static void gfx_v10_0_init_rlcg_reg_access_ctrl(struct amdgpu_device *adev)
+{
+	struct amdgpu_rlcg_reg_access_ctrl *reg_access_ctrl;
+
+	reg_access_ctrl = &adev->gfx.rlc.reg_access_ctrl[0];
+	reg_access_ctrl->scratch_reg0 = SOC15_REG_OFFSET(GC, 0, mmSCRATCH_REG0);
+	reg_access_ctrl->scratch_reg1 = SOC15_REG_OFFSET(GC, 0, mmSCRATCH_REG1);
+	reg_access_ctrl->scratch_reg2 = SOC15_REG_OFFSET(GC, 0, mmSCRATCH_REG2);
+	reg_access_ctrl->scratch_reg3 = SOC15_REG_OFFSET(GC, 0, mmSCRATCH_REG3);
+	reg_access_ctrl->grbm_cntl = SOC15_REG_OFFSET(GC, 0, mmGRBM_GFX_CNTL);
+	reg_access_ctrl->grbm_idx = SOC15_REG_OFFSET(GC, 0, mmGRBM_GFX_INDEX);
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	case IP_VERSION(10, 3, 0):
+		reg_access_ctrl->spare_int =
+			SOC15_REG_OFFSET(GC, 0, mmRLC_SPARE_INT_0_Sienna_Cichlid);
+		break;
+	default:
+		reg_access_ctrl->spare_int =
+			SOC15_REG_OFFSET(GC, 0, mmRLC_SPARE_INT);
+		break;
+	}
+	adev->gfx.rlc.rlcg_reg_access_supported = true;
+}
+
 static int gfx_v10_0_rlc_init(struct amdgpu_device *adev)
 {
 	const struct cs_section_def *cs_data;
@@ -4464,10 +4396,6 @@ static int gfx_v10_0_rlc_init(struct amdgpu_device *adev)
 			return r;
 	}
 
-	/* init spm vmid with 0xf */
-	if (adev->gfx.rlc.funcs->update_spm_vmid)
-		adev->gfx.rlc.funcs->update_spm_vmid(adev, 0xf);
-
 	return 0;
 }
 
@@ -4477,19 +4405,11 @@ static void gfx_v10_0_mec_fini(struct amdgpu_device *adev)
 	amdgpu_bo_free_kernel(&adev->gfx.mec.mec_fw_obj, NULL, NULL);
 }
 
-static int gfx_v10_0_me_init(struct amdgpu_device *adev)
+static void gfx_v10_0_me_init(struct amdgpu_device *adev)
 {
-	int r;
-
 	bitmap_zero(adev->gfx.me.queue_bitmap, AMDGPU_MAX_GFX_QUEUES);
 
 	amdgpu_gfx_graphics_queue_acquire(adev);
-
-	r = gfx_v10_0_init_microcode(adev);
-	if (r)
-		DRM_ERROR("Failed to load gfx firmware!\n");
-
-	return r;
 }
 
 static int gfx_v10_0_mec_init(struct amdgpu_device *adev)
@@ -4497,13 +4417,13 @@ static int gfx_v10_0_mec_init(struct amdgpu_device *adev)
 	int r;
 	u32 *hpd;
 	const __le32 *fw_data = NULL;
-	unsigned fw_size;
+	unsigned int fw_size;
 	u32 *fw = NULL;
 	size_t mec_hpd_size;
 
 	const struct gfx_firmware_header_v1_0 *mec_hdr = NULL;
 
-	bitmap_zero(adev->gfx.mec.queue_bitmap, AMDGPU_MAX_COMPUTE_QUEUES);
+	bitmap_zero(adev->gfx.mec_bitmap[0].queue_bitmap, AMDGPU_MAX_COMPUTE_QUEUES);
 
 	/* take ownership of the relevant compute queues */
 	amdgpu_gfx_compute_queue_acquire(adev);
@@ -4575,11 +4495,12 @@ static void wave_read_regs(struct amdgpu_device *adev, uint32_t wave,
 		*(out++) = RREG32_SOC15(GC, 0, mmSQ_IND_DATA);
 }
 
-static void gfx_v10_0_read_wave_data(struct amdgpu_device *adev, uint32_t simd, uint32_t wave, uint32_t *dst, int *no_fields)
+static void gfx_v10_0_read_wave_data(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd, uint32_t wave, uint32_t *dst, int *no_fields)
 {
 	/* in gfx10 the SIMD_ID is specified as part of the INSTANCE
 	 * field when performing a select_se_sh so it should be
-	 * zero here */
+	 * zero here
+	 */
 	WARN_ON(simd != 0);
 
 	/* type 2 wave data */
@@ -4602,7 +4523,7 @@ static void gfx_v10_0_read_wave_data(struct amdgpu_device *adev, uint32_t simd,
 	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_MODE);
 }
 
-static void gfx_v10_0_read_wave_sgprs(struct amdgpu_device *adev, uint32_t simd,
+static void gfx_v10_0_read_wave_sgprs(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd,
 				     uint32_t wave, uint32_t start,
 				     uint32_t size, uint32_t *dst)
 {
@@ -4613,7 +4534,7 @@ static void gfx_v10_0_read_wave_sgprs(struct amdgpu_device *adev, uint32_t simd,
 		dst);
 }
 
-static void gfx_v10_0_read_wave_vgprs(struct amdgpu_device *adev, uint32_t simd,
+static void gfx_v10_0_read_wave_vgprs(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd,
 				      uint32_t wave, uint32_t thread,
 				      uint32_t start, uint32_t size,
 				      uint32_t *dst)
@@ -4624,7 +4545,7 @@ static void gfx_v10_0_read_wave_vgprs(struct amdgpu_device *adev, uint32_t simd,
 }
 
 static void gfx_v10_0_select_me_pipe_q(struct amdgpu_device *adev,
-				       u32 me, u32 pipe, u32 q, u32 vm)
+				       u32 me, u32 pipe, u32 q, u32 vm, u32 xcc_id)
 {
 	nv_grbm_select(adev, me, pipe, q, vm);
 }
@@ -4660,9 +4581,7 @@ static void gfx_v10_0_gpu_early_init(struct amdgpu_device *adev)
 {
 	u32 gb_addr_config;
 
-	adev->gfx.funcs = &gfx_v10_0_gfx_funcs;
-
-	switch (adev->ip_versions[GC_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 	case IP_VERSION(10, 1, 10):
 	case IP_VERSION(10, 1, 1):
 	case IP_VERSION(10, 1, 2):
@@ -4678,7 +4597,9 @@ static void gfx_v10_0_gpu_early_init(struct amdgpu_device *adev)
 	case IP_VERSION(10, 3, 1):
 	case IP_VERSION(10, 3, 4):
 	case IP_VERSION(10, 3, 5):
+	case IP_VERSION(10, 3, 6):
 	case IP_VERSION(10, 3, 3):
+	case IP_VERSION(10, 3, 7):
 		adev->gfx.config.max_hw_contexts = 8;
 		adev->gfx.config.sc_prim_fifo_size_frontend = 0x20;
 		adev->gfx.config.sc_prim_fifo_size_backend = 0x100;
@@ -4689,6 +4610,7 @@ static void gfx_v10_0_gpu_early_init(struct amdgpu_device *adev)
 			1 << REG_GET_FIELD(gb_addr_config, GB_ADDR_CONFIG, NUM_PKRS);
 		break;
 	case IP_VERSION(10, 1, 3):
+	case IP_VERSION(10, 1, 4):
 		adev->gfx.config.max_hw_contexts = 8;
 		adev->gfx.config.sc_prim_fifo_size_frontend = 0x20;
 		adev->gfx.config.sc_prim_fifo_size_backend = 0x100;
@@ -4727,9 +4649,9 @@ static void gfx_v10_0_gpu_early_init(struct amdgpu_device *adev)
 static int gfx_v10_0_gfx_ring_init(struct amdgpu_device *adev, int ring_id,
 				   int me, int pipe, int queue)
 {
-	int r;
 	struct amdgpu_ring *ring;
 	unsigned int irq_type;
+	unsigned int hw_prio;
 
 	ring = &adev->gfx.gfx_ring[ring_id];
 
@@ -4744,21 +4666,20 @@ static int gfx_v10_0_gfx_ring_init(struct amdgpu_device *adev, int ring_id,
 		ring->doorbell_index = adev->doorbell_index.gfx_ring0 << 1;
 	else
 		ring->doorbell_index = adev->doorbell_index.gfx_ring1 << 1;
+	ring->vm_hub = AMDGPU_GFXHUB(0);
 	sprintf(ring->name, "gfx_%d.%d.%d", ring->me, ring->pipe, ring->queue);
 
 	irq_type = AMDGPU_CP_IRQ_GFX_ME0_PIPE0_EOP + ring->pipe;
-	r = amdgpu_ring_init(adev, ring, 1024, &adev->gfx.eop_irq, irq_type,
-			     AMDGPU_RING_PRIO_DEFAULT, NULL);
-	if (r)
-		return r;
-	return 0;
+	hw_prio = amdgpu_gfx_is_high_priority_graphics_queue(adev, ring) ?
+			AMDGPU_GFX_PIPE_PRIO_HIGH : AMDGPU_GFX_PIPE_PRIO_NORMAL;
+	return amdgpu_ring_init(adev, ring, 1024, &adev->gfx.eop_irq, irq_type,
+				hw_prio, NULL);
 }
 
 static int gfx_v10_0_compute_ring_init(struct amdgpu_device *adev, int ring_id,
 				       int mec, int pipe, int queue)
 {
-	int r;
-	unsigned irq_type;
+	unsigned int irq_type;
 	struct amdgpu_ring *ring;
 	unsigned int hw_prio;
 
@@ -4774,36 +4695,78 @@ static int gfx_v10_0_compute_ring_init(struct amdgpu_device *adev, int ring_id,
 	ring->doorbell_index = (adev->doorbell_index.mec_ring0 + ring_id) << 1;
 	ring->eop_gpu_addr = adev->gfx.mec.hpd_eop_gpu_addr
 				+ (ring_id * GFX10_MEC_HPD_SIZE);
+	ring->vm_hub = AMDGPU_GFXHUB(0);
 	sprintf(ring->name, "comp_%d.%d.%d", ring->me, ring->pipe, ring->queue);
 
 	irq_type = AMDGPU_CP_IRQ_COMPUTE_MEC1_PIPE0_EOP
 		+ ((ring->me - 1) * adev->gfx.mec.num_pipe_per_mec)
 		+ ring->pipe;
 	hw_prio = amdgpu_gfx_is_high_priority_compute_queue(adev, ring) ?
-			AMDGPU_GFX_PIPE_PRIO_HIGH : AMDGPU_GFX_PIPE_PRIO_NORMAL;
+			AMDGPU_RING_PRIO_2 : AMDGPU_RING_PRIO_DEFAULT;
 	/* type-2 packets are deprecated on MEC, use type-3 instead */
-	r = amdgpu_ring_init(adev, ring, 1024, &adev->gfx.eop_irq, irq_type,
+	return amdgpu_ring_init(adev, ring, 1024, &adev->gfx.eop_irq, irq_type,
 			     hw_prio, NULL);
-	if (r)
-		return r;
+}
 
-	return 0;
+static void gfx_v10_0_alloc_ip_dump(struct amdgpu_device *adev)
+{
+	uint32_t reg_count = ARRAY_SIZE(gc_reg_list_10_1);
+	uint32_t *ptr;
+	uint32_t inst;
+
+	ptr = kcalloc(reg_count, sizeof(uint32_t), GFP_KERNEL);
+	if (!ptr) {
+		DRM_ERROR("Failed to allocate memory for GFX IP Dump\n");
+		adev->gfx.ip_dump_core = NULL;
+	} else {
+		adev->gfx.ip_dump_core = ptr;
+	}
+
+	/* Allocate memory for compute queue registers for all the instances */
+	reg_count = ARRAY_SIZE(gc_cp_reg_list_10);
+	inst = adev->gfx.mec.num_mec * adev->gfx.mec.num_pipe_per_mec *
+		adev->gfx.mec.num_queue_per_pipe;
+
+	ptr = kcalloc(reg_count * inst, sizeof(uint32_t), GFP_KERNEL);
+	if (!ptr) {
+		DRM_ERROR("Failed to allocate memory for Compute Queues IP Dump\n");
+		adev->gfx.ip_dump_compute_queues = NULL;
+	} else {
+		adev->gfx.ip_dump_compute_queues = ptr;
+	}
+
+	/* Allocate memory for gfx queue registers for all the instances */
+	reg_count = ARRAY_SIZE(gc_gfx_queue_reg_list_10);
+	inst = adev->gfx.me.num_me * adev->gfx.me.num_pipe_per_me *
+		adev->gfx.me.num_queue_per_pipe;
+
+	ptr = kcalloc(reg_count * inst, sizeof(uint32_t), GFP_KERNEL);
+	if (!ptr) {
+		DRM_ERROR("Failed to allocate memory for GFX Queues IP Dump\n");
+		adev->gfx.ip_dump_gfx_queues = NULL;
+	} else {
+		adev->gfx.ip_dump_gfx_queues = ptr;
+	}
 }
 
-static int gfx_v10_0_sw_init(void *handle)
+static int gfx_v10_0_sw_init(struct amdgpu_ip_block *ip_block)
 {
 	int i, j, k, r, ring_id = 0;
-	struct amdgpu_kiq *kiq;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	int xcc_id = 0;
+	struct amdgpu_device *adev = ip_block->adev;
+	int num_queue_per_pipe = 1; /* we only enable 1 KGQ per pipe */
+
+	INIT_DELAYED_WORK(&adev->gfx.idle_work, amdgpu_gfx_profile_idle_work_handler);
 
-	switch (adev->ip_versions[GC_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 	case IP_VERSION(10, 1, 10):
 	case IP_VERSION(10, 1, 1):
 	case IP_VERSION(10, 1, 2):
 	case IP_VERSION(10, 1, 3):
+	case IP_VERSION(10, 1, 4):
 		adev->gfx.me.num_me = 1;
 		adev->gfx.me.num_pipe_per_me = 1;
-		adev->gfx.me.num_queue_per_pipe = 1;
+		adev->gfx.me.num_queue_per_pipe = 8;
 		adev->gfx.mec.num_mec = 2;
 		adev->gfx.mec.num_pipe_per_mec = 4;
 		adev->gfx.mec.num_queue_per_pipe = 8;
@@ -4813,10 +4776,12 @@ static int gfx_v10_0_sw_init(void *handle)
 	case IP_VERSION(10, 3, 1):
 	case IP_VERSION(10, 3, 4):
 	case IP_VERSION(10, 3, 5):
+	case IP_VERSION(10, 3, 6):
 	case IP_VERSION(10, 3, 3):
+	case IP_VERSION(10, 3, 7):
 		adev->gfx.me.num_me = 1;
-		adev->gfx.me.num_pipe_per_me = 1;
-		adev->gfx.me.num_queue_per_pipe = 1;
+		adev->gfx.me.num_pipe_per_me = 2;
+		adev->gfx.me.num_queue_per_pipe = 2;
 		adev->gfx.mec.num_mec = 2;
 		adev->gfx.mec.num_pipe_per_mec = 4;
 		adev->gfx.mec.num_queue_per_pipe = 4;
@@ -4830,11 +4795,79 @@ static int gfx_v10_0_sw_init(void *handle)
 		adev->gfx.mec.num_queue_per_pipe = 8;
 		break;
 	}
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	case IP_VERSION(10, 1, 10):
+	case IP_VERSION(10, 1, 1):
+	case IP_VERSION(10, 1, 2):
+		adev->gfx.cleaner_shader_ptr = gfx_10_1_10_cleaner_shader_hex;
+		adev->gfx.cleaner_shader_size = sizeof(gfx_10_1_10_cleaner_shader_hex);
+		if (adev->gfx.me_fw_version >= 101 &&
+		    adev->gfx.pfp_fw_version  >= 158 &&
+		    adev->gfx.mec_fw_version >= 151) {
+			adev->gfx.enable_cleaner_shader = true;
+			r = amdgpu_gfx_cleaner_shader_sw_init(adev, adev->gfx.cleaner_shader_size);
+			if (r) {
+				adev->gfx.enable_cleaner_shader = false;
+				dev_err(adev->dev, "Failed to initialize cleaner shader\n");
+			}
+		}
+		break;
+	case IP_VERSION(10, 3, 0):
+	case IP_VERSION(10, 3, 1):
+	case IP_VERSION(10, 3, 2):
+	case IP_VERSION(10, 3, 3):
+	case IP_VERSION(10, 3, 4):
+	case IP_VERSION(10, 3, 5):
+		adev->gfx.cleaner_shader_ptr = gfx_10_3_0_cleaner_shader_hex;
+		adev->gfx.cleaner_shader_size = sizeof(gfx_10_3_0_cleaner_shader_hex);
+		if (adev->gfx.me_fw_version >= 64 &&
+		    adev->gfx.pfp_fw_version >= 100 &&
+		    adev->gfx.mec_fw_version >= 122) {
+			adev->gfx.enable_cleaner_shader = true;
+			r = amdgpu_gfx_cleaner_shader_sw_init(adev, adev->gfx.cleaner_shader_size);
+			if (r) {
+				adev->gfx.enable_cleaner_shader = false;
+				dev_err(adev->dev, "Failed to initialize cleaner shader\n");
+			}
+		}
+		break;
+	case IP_VERSION(10, 3, 6):
+		adev->gfx.cleaner_shader_ptr = gfx_10_3_0_cleaner_shader_hex;
+		adev->gfx.cleaner_shader_size = sizeof(gfx_10_3_0_cleaner_shader_hex);
+		if (adev->gfx.me_fw_version >= 14 &&
+		    adev->gfx.pfp_fw_version >= 17 &&
+		    adev->gfx.mec_fw_version >= 24) {
+			adev->gfx.enable_cleaner_shader = true;
+			r = amdgpu_gfx_cleaner_shader_sw_init(adev, adev->gfx.cleaner_shader_size);
+			if (r) {
+				adev->gfx.enable_cleaner_shader = false;
+				dev_err(adev->dev, "Failed to initialize cleaner shader\n");
+			}
+		}
+		break;
+	case IP_VERSION(10, 3, 7):
+		adev->gfx.cleaner_shader_ptr = gfx_10_3_0_cleaner_shader_hex;
+		adev->gfx.cleaner_shader_size = sizeof(gfx_10_3_0_cleaner_shader_hex);
+		if (adev->gfx.me_fw_version >= 4 &&
+		    adev->gfx.pfp_fw_version >= 9 &&
+		    adev->gfx.mec_fw_version >= 12) {
+			adev->gfx.enable_cleaner_shader = true;
+			r = amdgpu_gfx_cleaner_shader_sw_init(adev, adev->gfx.cleaner_shader_size);
+			if (r) {
+				adev->gfx.enable_cleaner_shader = false;
+				dev_err(adev->dev, "Failed to initialize cleaner shader\n");
+			}
+		}
+		break;
+	default:
+		adev->gfx.enable_cleaner_shader = false;
+		break;
+	}
 
 	/* KIQ event */
 	r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_GRBM_CP,
 			      GFX_10_1__SRCID__CP_IB2_INTERRUPT_PKT,
-			      &adev->gfx.kiq.irq);
+			      &adev->gfx.kiq[0].irq);
 	if (r)
 		return r;
 
@@ -4845,6 +4878,13 @@ static int gfx_v10_0_sw_init(void *handle)
 	if (r)
 		return r;
 
+	/* Bad opcode Event */
+	r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_GRBM_CP,
+			      GFX_10_1__SRCID__CP_BAD_OPCODE_ERROR,
+			      &adev->gfx.bad_op_irq);
+	if (r)
+		return r;
+
 	/* Privileged reg */
 	r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_GRBM_CP, GFX_10_1__SRCID__CP_PRIV_REG_FAULT,
 			      &adev->gfx.priv_reg_irq);
@@ -4859,16 +4899,16 @@ static int gfx_v10_0_sw_init(void *handle)
 
 	adev->gfx.gfx_current_status = AMDGPU_GFX_NORMAL_MODE;
 
-	gfx_v10_0_scratch_init(adev);
-
-	r = gfx_v10_0_me_init(adev);
-	if (r)
-		return r;
+	gfx_v10_0_me_init(adev);
 
-	r = gfx_v10_0_rlc_init(adev);
-	if (r) {
-		DRM_ERROR("Failed to init rlc BOs!\n");
-		return r;
+	if (adev->gfx.rlc.funcs) {
+		if (adev->gfx.rlc.funcs->init) {
+			r = adev->gfx.rlc.funcs->init(adev);
+			if (r) {
+				dev_err(adev->dev, "Failed to init rlc BOs!\n");
+				return r;
+			}
+		}
 	}
 
 	r = gfx_v10_0_mec_init(adev);
@@ -4879,7 +4919,7 @@ static int gfx_v10_0_sw_init(void *handle)
 
 	/* set up the gfx ring */
 	for (i = 0; i < adev->gfx.me.num_me; i++) {
-		for (j = 0; j < adev->gfx.me.num_queue_per_pipe; j++) {
+		for (j = 0; j < num_queue_per_pipe; j++) {
 			for (k = 0; k < adev->gfx.me.num_pipe_per_me; k++) {
 				if (!amdgpu_gfx_is_me_queue_enabled(adev, i, k, j))
 					continue;
@@ -4898,8 +4938,8 @@ static int gfx_v10_0_sw_init(void *handle)
 	for (i = 0; i < adev->gfx.mec.num_mec; ++i) {
 		for (j = 0; j < adev->gfx.mec.num_queue_per_pipe; j++) {
 			for (k = 0; k < adev->gfx.mec.num_pipe_per_mec; k++) {
-				if (!amdgpu_gfx_is_mec_queue_enabled(adev, i, k,
-								     j))
+				if (!amdgpu_gfx_is_mec_queue_enabled(adev, 0, i,
+								     k, j))
 					continue;
 
 				r = gfx_v10_0_compute_ring_init(adev, ring_id,
@@ -4912,18 +4952,27 @@ static int gfx_v10_0_sw_init(void *handle)
 		}
 	}
 
-	r = amdgpu_gfx_kiq_init(adev, GFX10_MEC_HPD_SIZE);
+	adev->gfx.gfx_supported_reset =
+		amdgpu_get_soft_full_reset_mask(&adev->gfx.gfx_ring[0]);
+	adev->gfx.compute_supported_reset =
+		amdgpu_get_soft_full_reset_mask(&adev->gfx.compute_ring[0]);
+	if (!amdgpu_sriov_vf(adev) &&
+	    !adev->debug_disable_gpu_ring_reset) {
+		adev->gfx.compute_supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE;
+		adev->gfx.gfx_supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE;
+	}
+
+	r = amdgpu_gfx_kiq_init(adev, GFX10_MEC_HPD_SIZE, 0);
 	if (r) {
 		DRM_ERROR("Failed to init KIQ BOs!\n");
 		return r;
 	}
 
-	kiq = &adev->gfx.kiq;
-	r = amdgpu_gfx_kiq_init_ring(adev, &kiq->ring, &kiq->irq);
+	r = amdgpu_gfx_kiq_init_ring(adev, xcc_id);
 	if (r)
 		return r;
 
-	r = amdgpu_gfx_mqd_sw_init(adev, sizeof(struct v10_compute_mqd));
+	r = amdgpu_gfx_mqd_sw_init(adev, sizeof(struct v10_compute_mqd), 0);
 	if (r)
 		return r;
 
@@ -4938,6 +4987,12 @@ static int gfx_v10_0_sw_init(void *handle)
 
 	gfx_v10_0_gpu_early_init(adev);
 
+	gfx_v10_0_alloc_ip_dump(adev);
+
+	r = amdgpu_gfx_sysfs_init(adev);
+	if (r)
+		return r;
+
 	return 0;
 }
 
@@ -4962,19 +5017,22 @@ static void gfx_v10_0_me_fini(struct amdgpu_device *adev)
 			      (void **)&adev->gfx.me.me_fw_ptr);
 }
 
-static int gfx_v10_0_sw_fini(void *handle)
+static int gfx_v10_0_sw_fini(struct amdgpu_ip_block *ip_block)
 {
 	int i;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	for (i = 0; i < adev->gfx.num_gfx_rings; i++)
 		amdgpu_ring_fini(&adev->gfx.gfx_ring[i]);
 	for (i = 0; i < adev->gfx.num_compute_rings; i++)
 		amdgpu_ring_fini(&adev->gfx.compute_ring[i]);
 
-	amdgpu_gfx_mqd_sw_fini(adev);
-	amdgpu_gfx_kiq_free_ring(&adev->gfx.kiq.ring);
-	amdgpu_gfx_kiq_fini(adev);
+	amdgpu_gfx_mqd_sw_fini(adev, 0);
+
+	amdgpu_gfx_kiq_free_ring(&adev->gfx.kiq[0].ring);
+	amdgpu_gfx_kiq_fini(adev, 0);
+
+	amdgpu_gfx_cleaner_shader_sw_fini(adev);
 
 	gfx_v10_0_pfp_fini(adev);
 	gfx_v10_0_ce_fini(adev);
@@ -4986,12 +5044,17 @@ static int gfx_v10_0_sw_fini(void *handle)
 		gfx_v10_0_rlc_backdoor_autoload_buffer_fini(adev);
 
 	gfx_v10_0_free_microcode(adev);
+	amdgpu_gfx_sysfs_fini(adev);
+
+	kfree(adev->gfx.ip_dump_core);
+	kfree(adev->gfx.ip_dump_compute_queues);
+	kfree(adev->gfx.ip_dump_gfx_queues);
 
 	return 0;
 }
 
 static void gfx_v10_0_select_se_sh(struct amdgpu_device *adev, u32 se_num,
-				   u32 sh_num, u32 instance)
+				   u32 sh_num, u32 instance, int xcc_id)
 {
 	u32 data;
 
@@ -5046,17 +5109,21 @@ static void gfx_v10_0_setup_rb(struct amdgpu_device *adev)
 	for (i = 0; i < adev->gfx.config.max_shader_engines; i++) {
 		for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) {
 			bitmap = i * adev->gfx.config.max_sh_per_se + j;
-			if (((adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 3, 0)) ||
-				(adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 3, 3))) &&
+			if (((amdgpu_ip_version(adev, GC_HWIP, 0) ==
+			      IP_VERSION(10, 3, 0)) ||
+			     (amdgpu_ip_version(adev, GC_HWIP, 0) ==
+			      IP_VERSION(10, 3, 3)) ||
+			     (amdgpu_ip_version(adev, GC_HWIP, 0) ==
+			      IP_VERSION(10, 3, 6))) &&
 			    ((gfx_v10_3_get_disabled_sa(adev) >> bitmap) & 1))
 				continue;
-			gfx_v10_0_select_se_sh(adev, i, j, 0xffffffff);
+			gfx_v10_0_select_se_sh(adev, i, j, 0xffffffff, 0);
 			data = gfx_v10_0_get_rb_active_bitmap(adev);
 			active_rbs |= data << ((i * adev->gfx.config.max_sh_per_se + j) *
 					       rb_bitmap_width_per_sh);
 		}
 	}
-	gfx_v10_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+	gfx_v10_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
 	mutex_unlock(&adev->grbm_idx_mutex);
 
 	adev->gfx.config.backend_enable_mask = active_rbs;
@@ -5073,8 +5140,9 @@ static u32 gfx_v10_0_init_pa_sc_tile_steering_override(struct amdgpu_device *ade
 	uint32_t pa_sc_tile_steering_override;
 
 	/* for ASICs that integrates GFX v10.3
-	 * pa_sc_tile_steering_override should be set to 0 */
-	if (adev->ip_versions[GC_HWIP][0] >= IP_VERSION(10, 3, 0))
+	 * pa_sc_tile_steering_override should be set to 0
+	 */
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) >= IP_VERSION(10, 3, 0))
 		return 0;
 
 	/* init num_sc */
@@ -5103,6 +5171,29 @@ static u32 gfx_v10_0_init_pa_sc_tile_steering_override(struct amdgpu_device *ade
 
 #define DEFAULT_SH_MEM_BASES	(0x6000)
 
+static void gfx_v10_0_debug_trap_config_init(struct amdgpu_device *adev,
+				uint32_t first_vmid,
+				uint32_t last_vmid)
+{
+	uint32_t data;
+	uint32_t trap_config_vmid_mask = 0;
+	int i;
+
+	/* Calculate trap config vmid mask */
+	for (i = first_vmid; i < last_vmid; i++)
+		trap_config_vmid_mask |= (1 << i);
+
+	data = REG_SET_FIELD(0, SPI_GDBG_TRAP_CONFIG,
+			VMID_SEL, trap_config_vmid_mask);
+	data = REG_SET_FIELD(data, SPI_GDBG_TRAP_CONFIG,
+			TRAP_EN, 1);
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_CONFIG), data);
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_MASK), 0);
+
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_DATA0), 0);
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_DATA1), 0);
+}
+
 static void gfx_v10_0_init_compute_vmid(struct amdgpu_device *adev)
 {
 	int i;
@@ -5126,14 +5217,19 @@ static void gfx_v10_0_init_compute_vmid(struct amdgpu_device *adev)
 	nv_grbm_select(adev, 0, 0, 0, 0);
 	mutex_unlock(&adev->srbm_mutex);
 
-	/* Initialize all compute VMIDs to have no GDS, GWS, or OA
-	   acccess. These should be enabled by FW for target VMIDs. */
+	/*
+	 * Initialize all compute VMIDs to have no GDS, GWS, or OA
+	 * access. These should be enabled by FW for target VMIDs.
+	 */
 	for (i = adev->vm_manager.first_kfd_vmid; i < AMDGPU_NUM_VMID; i++) {
 		WREG32_SOC15_OFFSET(GC, 0, mmGDS_VMID0_BASE, 2 * i, 0);
 		WREG32_SOC15_OFFSET(GC, 0, mmGDS_VMID0_SIZE, 2 * i, 0);
 		WREG32_SOC15_OFFSET(GC, 0, mmGDS_GWS_VMID0, i, 0);
 		WREG32_SOC15_OFFSET(GC, 0, mmGDS_OA_VMID0, i, 0);
 	}
+
+	gfx_v10_0_debug_trap_config_init(adev, adev->vm_manager.first_kfd_vmid,
+					AMDGPU_NUM_VMID);
 }
 
 static void gfx_v10_0_init_gds_vmid(struct amdgpu_device *adev)
@@ -5185,7 +5281,7 @@ static void gfx_v10_0_tcp_harvest(struct amdgpu_device *adev)
 	mutex_lock(&adev->grbm_idx_mutex);
 	for (i = 0; i < adev->gfx.config.max_shader_engines; i++) {
 		for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) {
-			gfx_v10_0_select_se_sh(adev, i, j, 0xffffffff);
+			gfx_v10_0_select_se_sh(adev, i, j, 0xffffffff, 0);
 			wgp_active_bitmap = gfx_v10_0_get_wgp_active_bitmap_per_sh(adev);
 			/*
 			 * Set corresponding TCP bits for the inactive WGPs in
@@ -5218,7 +5314,7 @@ static void gfx_v10_0_tcp_harvest(struct amdgpu_device *adev)
 		}
 	}
 
-	gfx_v10_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+	gfx_v10_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
 	mutex_unlock(&adev->grbm_idx_mutex);
 }
 
@@ -5227,7 +5323,7 @@ static void gfx_v10_0_get_tcc_info(struct amdgpu_device *adev)
 	/* TCCs are global (not instanced). */
 	uint32_t tcc_disable;
 
-	if (adev->ip_versions[GC_HWIP][0] >= IP_VERSION(10, 3, 0)) {
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) >= IP_VERSION(10, 3, 0)) {
 		tcc_disable = RREG32_SOC15(GC, 0, mmCGTS_TCC_DISABLE_gc_10_3) |
 			      RREG32_SOC15(GC, 0, mmCGTS_USER_TCC_DISABLE_gc_10_3);
 	} else {
@@ -5245,7 +5341,8 @@ static void gfx_v10_0_constants_init(struct amdgpu_device *adev)
 	u32 tmp;
 	int i;
 
-	WREG32_FIELD15(GC, 0, GRBM_CNTL, READ_TIMEOUT, 0xff);
+	if (!amdgpu_sriov_vf(adev))
+		WREG32_FIELD15(GC, 0, GRBM_CNTL, READ_TIMEOUT, 0xff);
 
 	gfx_v10_0_setup_rb(adev);
 	gfx_v10_0_get_cu_info(adev, &adev->gfx.cu_info);
@@ -5256,7 +5353,7 @@ static void gfx_v10_0_constants_init(struct amdgpu_device *adev)
 	/* XXX SH_MEM regs */
 	/* where to put LDS, scratch, GPUVM in FSA64 space */
 	mutex_lock(&adev->srbm_mutex);
-	for (i = 0; i < adev->vm_manager.id_mgr[AMDGPU_GFXHUB_0].num_ids; i++) {
+	for (i = 0; i < adev->vm_manager.id_mgr[AMDGPU_GFXHUB(0)].num_ids; i++) {
 		nv_grbm_select(adev, 0, 0, 0, i);
 		/* CP and shaders */
 		WREG32_SOC15(GC, 0, mmSH_MEM_CONFIG, DEFAULT_SH_MEM_CONFIG);
@@ -5277,26 +5374,74 @@ static void gfx_v10_0_constants_init(struct amdgpu_device *adev)
 
 }
 
+static u32 gfx_v10_0_get_cpg_int_cntl(struct amdgpu_device *adev,
+				      int me, int pipe)
+{
+	if (me != 0)
+		return 0;
+
+	switch (pipe) {
+	case 0:
+		return SOC15_REG_OFFSET(GC, 0, mmCP_INT_CNTL_RING0);
+	case 1:
+		return SOC15_REG_OFFSET(GC, 0, mmCP_INT_CNTL_RING1);
+	default:
+		return 0;
+	}
+}
+
+static u32 gfx_v10_0_get_cpc_int_cntl(struct amdgpu_device *adev,
+				      int me, int pipe)
+{
+	/*
+	 * amdgpu controls only the first MEC. That's why this function only
+	 * handles the setting of interrupts for this specific MEC. All other
+	 * pipes' interrupts are set by amdkfd.
+	 */
+	if (me != 1)
+		return 0;
+
+	switch (pipe) {
+	case 0:
+		return SOC15_REG_OFFSET(GC, 0, mmCP_ME1_PIPE0_INT_CNTL);
+	case 1:
+		return SOC15_REG_OFFSET(GC, 0, mmCP_ME1_PIPE1_INT_CNTL);
+	case 2:
+		return SOC15_REG_OFFSET(GC, 0, mmCP_ME1_PIPE2_INT_CNTL);
+	case 3:
+		return SOC15_REG_OFFSET(GC, 0, mmCP_ME1_PIPE3_INT_CNTL);
+	default:
+		return 0;
+	}
+}
+
 static void gfx_v10_0_enable_gui_idle_interrupt(struct amdgpu_device *adev,
 					       bool enable)
 {
-	u32 tmp;
+	u32 tmp, cp_int_cntl_reg;
+	int i, j;
 
 	if (amdgpu_sriov_vf(adev))
 		return;
 
-	tmp = RREG32_SOC15(GC, 0, mmCP_INT_CNTL_RING0);
-
-	tmp = REG_SET_FIELD(tmp, CP_INT_CNTL_RING0, CNTX_BUSY_INT_ENABLE,
-			    enable ? 1 : 0);
-	tmp = REG_SET_FIELD(tmp, CP_INT_CNTL_RING0, CNTX_EMPTY_INT_ENABLE,
-			    enable ? 1 : 0);
-	tmp = REG_SET_FIELD(tmp, CP_INT_CNTL_RING0, CMP_BUSY_INT_ENABLE,
-			    enable ? 1 : 0);
-	tmp = REG_SET_FIELD(tmp, CP_INT_CNTL_RING0, GFX_IDLE_INT_ENABLE,
-			    enable ? 1 : 0);
-
-	WREG32_SOC15(GC, 0, mmCP_INT_CNTL_RING0, tmp);
+	for (i = 0; i < adev->gfx.me.num_me; i++) {
+		for (j = 0; j < adev->gfx.me.num_pipe_per_me; j++) {
+			cp_int_cntl_reg = gfx_v10_0_get_cpg_int_cntl(adev, i, j);
+
+			if (cp_int_cntl_reg) {
+				tmp = RREG32_SOC15_IP(GC, cp_int_cntl_reg);
+				tmp = REG_SET_FIELD(tmp, CP_INT_CNTL_RING0, CNTX_BUSY_INT_ENABLE,
+						    enable ? 1 : 0);
+				tmp = REG_SET_FIELD(tmp, CP_INT_CNTL_RING0, CNTX_EMPTY_INT_ENABLE,
+						    enable ? 1 : 0);
+				tmp = REG_SET_FIELD(tmp, CP_INT_CNTL_RING0, CMP_BUSY_INT_ENABLE,
+						    enable ? 1 : 0);
+				tmp = REG_SET_FIELD(tmp, CP_INT_CNTL_RING0, GFX_IDLE_INT_ENABLE,
+						    enable ? 1 : 0);
+				WREG32_SOC15_IP(GC, cp_int_cntl_reg, tmp);
+			}
+		}
+	}
 }
 
 static int gfx_v10_0_init_csb(struct amdgpu_device *adev)
@@ -5304,7 +5449,7 @@ static int gfx_v10_0_init_csb(struct amdgpu_device *adev)
 	adev->gfx.rlc.funcs->get_csb_buffer(adev, adev->gfx.rlc.cs_ptr);
 
 	/* csib */
-	if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 1, 2)) {
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(10, 1, 2)) {
 		WREG32_SOC15_RLC(GC, 0, mmRLC_CSIB_ADDR_HI,
 				adev->gfx.rlc.clear_state_gpu_addr >> 32);
 		WREG32_SOC15_RLC(GC, 0, mmRLC_CSIB_ADDR_LO,
@@ -5360,8 +5505,10 @@ static void gfx_v10_0_rlc_smu_handshake_cntl(struct amdgpu_device *adev,
 
 static void gfx_v10_0_rlc_start(struct amdgpu_device *adev)
 {
-	/* TODO: enable rlc & smu handshake until smu
-	 * and gfxoff feature works as expected */
+	/*
+	 * TODO: enable rlc & smu handshake until smu
+	 * and gfxoff feature works as expected
+	 */
 	if (!(amdgpu_pp_feature_mask & PP_GFXOFF_MASK))
 		gfx_v10_0_rlc_smu_handshake_cntl(adev, false);
 
@@ -5384,7 +5531,7 @@ static int gfx_v10_0_rlc_load_microcode(struct amdgpu_device *adev)
 {
 	const struct rlc_firmware_header_v2_0 *hdr;
 	const __le32 *fw_data;
-	unsigned i, fw_size;
+	unsigned int i, fw_size;
 
 	if (!adev->gfx.rlc_fw)
 		return -EINVAL;
@@ -5421,6 +5568,8 @@ static int gfx_v10_0_rlc_resume(struct amdgpu_device *adev)
 
 		gfx_v10_0_init_csb(adev);
 
+		gfx_v10_0_update_spm_vmid_internal(adev, 0xf);
+
 		if (!amdgpu_sriov_vf(adev)) /* enable RLC SRM */
 			gfx_v10_0_rlc_enable_srm(adev);
 	} else {
@@ -5451,6 +5600,8 @@ static int gfx_v10_0_rlc_resume(struct amdgpu_device *adev)
 
 		gfx_v10_0_init_csb(adev);
 
+		gfx_v10_0_update_spm_vmid_internal(adev, 0xf);
+
 		adev->gfx.rlc.funcs->start(adev);
 
 		if (adev->firmware.load_type == AMDGPU_FW_LOAD_RLC_BACKDOOR_AUTO) {
@@ -5459,6 +5610,7 @@ static int gfx_v10_0_rlc_resume(struct amdgpu_device *adev)
 				return r;
 		}
 	}
+
 	return 0;
 }
 
@@ -5926,11 +6078,13 @@ static int gfx_v10_0_cp_gfx_enable(struct amdgpu_device *adev, bool enable)
 	tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, PFP_HALT, enable ? 0 : 1);
 	tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, CE_HALT, enable ? 0 : 1);
 
-	if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 1, 2)) {
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(10, 1, 2))
 		WREG32_SOC15_RLC(GC, 0, mmCP_ME_CNTL, tmp);
-	} else {
+	else
 		WREG32_SOC15(GC, 0, mmCP_ME_CNTL, tmp);
-	}
+
+	if (amdgpu_in_reset(adev) && !enable)
+		return 0;
 
 	for (i = 0; i < adev->usec_timeout; i++) {
 		if (RREG32_SOC15(GC, 0, mmCP_STAT) == 0)
@@ -5949,7 +6103,7 @@ static int gfx_v10_0_cp_gfx_load_pfp_microcode(struct amdgpu_device *adev)
 	int r;
 	const struct gfx_firmware_header_v1_0 *pfp_hdr;
 	const __le32 *fw_data;
-	unsigned i, fw_size;
+	unsigned int i, fw_size;
 	uint32_t tmp;
 	uint32_t usec_timeout = 50000;  /* wait for 50ms */
 
@@ -5998,7 +6152,7 @@ static int gfx_v10_0_cp_gfx_load_pfp_microcode(struct amdgpu_device *adev)
 	}
 
 	if (amdgpu_emu_mode == 1)
-		adev->hdp.funcs->flush_hdp(adev, NULL);
+		amdgpu_device_flush_hdp(adev, NULL);
 
 	tmp = RREG32_SOC15(GC, 0, mmCP_PFP_IC_BASE_CNTL);
 	tmp = REG_SET_FIELD(tmp, CP_PFP_IC_BASE_CNTL, VMID, 0);
@@ -6027,7 +6181,7 @@ static int gfx_v10_0_cp_gfx_load_ce_microcode(struct amdgpu_device *adev)
 	int r;
 	const struct gfx_firmware_header_v1_0 *ce_hdr;
 	const __le32 *fw_data;
-	unsigned i, fw_size;
+	unsigned int i, fw_size;
 	uint32_t tmp;
 	uint32_t usec_timeout = 50000;  /* wait for 50ms */
 
@@ -6076,7 +6230,7 @@ static int gfx_v10_0_cp_gfx_load_ce_microcode(struct amdgpu_device *adev)
 	}
 
 	if (amdgpu_emu_mode == 1)
-		adev->hdp.funcs->flush_hdp(adev, NULL);
+		amdgpu_device_flush_hdp(adev, NULL);
 
 	tmp = RREG32_SOC15(GC, 0, mmCP_CE_IC_BASE_CNTL);
 	tmp = REG_SET_FIELD(tmp, CP_CE_IC_BASE_CNTL, VMID, 0);
@@ -6104,7 +6258,7 @@ static int gfx_v10_0_cp_gfx_load_me_microcode(struct amdgpu_device *adev)
 	int r;
 	const struct gfx_firmware_header_v1_0 *me_hdr;
 	const __le32 *fw_data;
-	unsigned i, fw_size;
+	unsigned int i, fw_size;
 	uint32_t tmp;
 	uint32_t usec_timeout = 50000;  /* wait for 50ms */
 
@@ -6153,7 +6307,7 @@ static int gfx_v10_0_cp_gfx_load_me_microcode(struct amdgpu_device *adev)
 	}
 
 	if (amdgpu_emu_mode == 1)
-		adev->hdp.funcs->flush_hdp(adev, NULL);
+		amdgpu_device_flush_hdp(adev, NULL);
 
 	tmp = RREG32_SOC15(GC, 0, mmCP_ME_IC_BASE_CNTL);
 	tmp = REG_SET_FIELD(tmp, CP_ME_IC_BASE_CNTL, VMID, 0);
@@ -6315,13 +6469,15 @@ static void gfx_v10_0_cp_gfx_set_doorbell(struct amdgpu_device *adev,
 		}
 		WREG32_SOC15(GC, 0, mmCP_RB_DOORBELL_CONTROL, tmp);
 	}
-	switch (adev->ip_versions[GC_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 	case IP_VERSION(10, 3, 0):
 	case IP_VERSION(10, 3, 2):
 	case IP_VERSION(10, 3, 1):
 	case IP_VERSION(10, 3, 4):
 	case IP_VERSION(10, 3, 5):
+	case IP_VERSION(10, 3, 6):
 	case IP_VERSION(10, 3, 3):
+	case IP_VERSION(10, 3, 7):
 		tmp = REG_SET_FIELD(0, CP_RB_DOORBELL_RANGE_LOWER,
 				    DOORBELL_RANGE_LOWER_Sienna_Cichlid, ring->doorbell_index);
 		WREG32_SOC15(GC, 0, mmCP_RB_DOORBELL_RANGE_LOWER, tmp);
@@ -6346,7 +6502,6 @@ static int gfx_v10_0_cp_gfx_resume(struct amdgpu_device *adev)
 	u32 tmp;
 	u32 rb_bufsz;
 	u64 rb_addr, rptr_addr, wptr_gpu_addr;
-	u32 i;
 
 	/* Set the write pointer delay */
 	WREG32_SOC15(GC, 0, mmCP_RB_WPTR_DELAY, 0);
@@ -6373,13 +6528,13 @@ static int gfx_v10_0_cp_gfx_resume(struct amdgpu_device *adev)
 	WREG32_SOC15(GC, 0, mmCP_RB0_WPTR, lower_32_bits(ring->wptr));
 	WREG32_SOC15(GC, 0, mmCP_RB0_WPTR_HI, upper_32_bits(ring->wptr));
 
-	/* set the wb address wether it's enabled or not */
-	rptr_addr = adev->wb.gpu_addr + (ring->rptr_offs * 4);
+	/* set the wb address whether it's enabled or not */
+	rptr_addr = ring->rptr_gpu_addr;
 	WREG32_SOC15(GC, 0, mmCP_RB0_RPTR_ADDR, lower_32_bits(rptr_addr));
 	WREG32_SOC15(GC, 0, mmCP_RB0_RPTR_ADDR_HI, upper_32_bits(rptr_addr) &
 		     CP_RB_RPTR_ADDR_HI__RB_RPTR_ADDR_HI_MASK);
 
-	wptr_gpu_addr = adev->wb.gpu_addr + (ring->wptr_offs * 4);
+	wptr_gpu_addr = ring->wptr_gpu_addr;
 	WREG32_SOC15(GC, 0, mmCP_RB_WPTR_POLL_ADDR_LO,
 		     lower_32_bits(wptr_gpu_addr));
 	WREG32_SOC15(GC, 0, mmCP_RB_WPTR_POLL_ADDR_HI,
@@ -6411,12 +6566,12 @@ static int gfx_v10_0_cp_gfx_resume(struct amdgpu_device *adev)
 		ring->wptr = 0;
 		WREG32_SOC15(GC, 0, mmCP_RB1_WPTR, lower_32_bits(ring->wptr));
 		WREG32_SOC15(GC, 0, mmCP_RB1_WPTR_HI, upper_32_bits(ring->wptr));
-		/* Set the wb address wether it's enabled or not */
-		rptr_addr = adev->wb.gpu_addr + (ring->rptr_offs * 4);
+		/* Set the wb address whether it's enabled or not */
+		rptr_addr = ring->rptr_gpu_addr;
 		WREG32_SOC15(GC, 0, mmCP_RB1_RPTR_ADDR, lower_32_bits(rptr_addr));
 		WREG32_SOC15(GC, 0, mmCP_RB1_RPTR_ADDR_HI, upper_32_bits(rptr_addr) &
 			     CP_RB1_RPTR_ADDR_HI__RB_RPTR_ADDR_HI_MASK);
-		wptr_gpu_addr = adev->wb.gpu_addr + (ring->wptr_offs * 4);
+		wptr_gpu_addr = ring->wptr_gpu_addr;
 		WREG32_SOC15(GC, 0, mmCP_RB_WPTR_POLL_ADDR_LO,
 			     lower_32_bits(wptr_gpu_addr));
 		WREG32_SOC15(GC, 0, mmCP_RB_WPTR_POLL_ADDR_HI,
@@ -6441,24 +6596,21 @@ static int gfx_v10_0_cp_gfx_resume(struct amdgpu_device *adev)
 	/* start the ring */
 	gfx_v10_0_cp_gfx_start(adev);
 
-	for (i = 0; i < adev->gfx.num_gfx_rings; i++) {
-		ring = &adev->gfx.gfx_ring[i];
-		ring->sched.ready = true;
-	}
-
 	return 0;
 }
 
 static void gfx_v10_0_cp_compute_enable(struct amdgpu_device *adev, bool enable)
 {
 	if (enable) {
-		switch (adev->ip_versions[GC_HWIP][0]) {
+		switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 		case IP_VERSION(10, 3, 0):
 		case IP_VERSION(10, 3, 2):
 		case IP_VERSION(10, 3, 1):
 		case IP_VERSION(10, 3, 4):
 		case IP_VERSION(10, 3, 5):
+		case IP_VERSION(10, 3, 6):
 		case IP_VERSION(10, 3, 3):
+		case IP_VERSION(10, 3, 7):
 			WREG32_SOC15(GC, 0, mmCP_MEC_CNTL_Sienna_Cichlid, 0);
 			break;
 		default:
@@ -6466,13 +6618,15 @@ static void gfx_v10_0_cp_compute_enable(struct amdgpu_device *adev, bool enable)
 			break;
 		}
 	} else {
-		switch (adev->ip_versions[GC_HWIP][0]) {
+		switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 		case IP_VERSION(10, 3, 0):
 		case IP_VERSION(10, 3, 2):
 		case IP_VERSION(10, 3, 1):
 		case IP_VERSION(10, 3, 4):
 		case IP_VERSION(10, 3, 5):
+		case IP_VERSION(10, 3, 6):
 		case IP_VERSION(10, 3, 3):
+		case IP_VERSION(10, 3, 7):
 			WREG32_SOC15(GC, 0, mmCP_MEC_CNTL_Sienna_Cichlid,
 				     (CP_MEC_CNTL__MEC_ME1_HALT_MASK |
 				      CP_MEC_CNTL__MEC_ME2_HALT_MASK));
@@ -6483,7 +6637,7 @@ static void gfx_v10_0_cp_compute_enable(struct amdgpu_device *adev, bool enable)
 				      CP_MEC_CNTL__MEC_ME2_HALT_MASK));
 			break;
 		}
-		adev->gfx.kiq.ring.sched.ready = false;
+		adev->gfx.kiq[0].ring.sched.ready = false;
 	}
 	udelay(50);
 }
@@ -6492,7 +6646,7 @@ static int gfx_v10_0_cp_compute_load_microcode(struct amdgpu_device *adev)
 {
 	const struct gfx_firmware_header_v1_0 *mec_hdr;
 	const __le32 *fw_data;
-	unsigned i;
+	unsigned int i;
 	u32 tmp;
 	u32 usec_timeout = 50000; /* Wait for 50 ms */
 
@@ -6528,7 +6682,7 @@ static int gfx_v10_0_cp_compute_load_microcode(struct amdgpu_device *adev)
 	}
 
 	if (amdgpu_emu_mode == 1)
-		adev->hdp.funcs->flush_hdp(adev, NULL);
+		amdgpu_device_flush_hdp(adev, NULL);
 
 	tmp = RREG32_SOC15(GC, 0, mmCP_CPC_IC_BASE_CNTL);
 	tmp = REG_SET_FIELD(tmp, CP_CPC_IC_BASE_CNTL, CACHE_POLICY, 0);
@@ -6564,35 +6718,51 @@ static void gfx_v10_0_kiq_setting(struct amdgpu_ring *ring)
 	struct amdgpu_device *adev = ring->adev;
 
 	/* tell RLC which is KIQ queue */
-	switch (adev->ip_versions[GC_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 	case IP_VERSION(10, 3, 0):
 	case IP_VERSION(10, 3, 2):
 	case IP_VERSION(10, 3, 1):
 	case IP_VERSION(10, 3, 4):
 	case IP_VERSION(10, 3, 5):
+	case IP_VERSION(10, 3, 6):
 	case IP_VERSION(10, 3, 3):
+	case IP_VERSION(10, 3, 7):
 		tmp = RREG32_SOC15(GC, 0, mmRLC_CP_SCHEDULERS_Sienna_Cichlid);
 		tmp &= 0xffffff00;
 		tmp |= (ring->me << 5) | (ring->pipe << 3) | (ring->queue);
-		WREG32_SOC15(GC, 0, mmRLC_CP_SCHEDULERS_Sienna_Cichlid, tmp);
-		tmp |= 0x80;
-		WREG32_SOC15(GC, 0, mmRLC_CP_SCHEDULERS_Sienna_Cichlid, tmp);
+		WREG32_SOC15(GC, 0, mmRLC_CP_SCHEDULERS_Sienna_Cichlid, tmp | 0x80);
 		break;
 	default:
 		tmp = RREG32_SOC15(GC, 0, mmRLC_CP_SCHEDULERS);
 		tmp &= 0xffffff00;
 		tmp |= (ring->me << 5) | (ring->pipe << 3) | (ring->queue);
-		WREG32_SOC15(GC, 0, mmRLC_CP_SCHEDULERS, tmp);
-		tmp |= 0x80;
-		WREG32_SOC15(GC, 0, mmRLC_CP_SCHEDULERS, tmp);
+		WREG32_SOC15(GC, 0, mmRLC_CP_SCHEDULERS, tmp | 0x80);
 		break;
 	}
 }
 
-static int gfx_v10_0_gfx_mqd_init(struct amdgpu_ring *ring)
+static void gfx_v10_0_gfx_mqd_set_priority(struct amdgpu_device *adev,
+					   struct v10_gfx_mqd *mqd,
+					   struct amdgpu_mqd_prop *prop)
 {
-	struct amdgpu_device *adev = ring->adev;
-	struct v10_gfx_mqd *mqd = ring->mqd_ptr;
+	bool priority = 0;
+	u32 tmp;
+
+	/* set up default queue priority level
+	 * 0x0 = low priority, 0x1 = high priority
+	 */
+	if (prop->hqd_pipe_priority == AMDGPU_GFX_PIPE_PRIO_HIGH)
+		priority = 1;
+
+	tmp = RREG32_SOC15(GC, 0, mmCP_GFX_HQD_QUEUE_PRIORITY);
+	tmp = REG_SET_FIELD(tmp, CP_GFX_HQD_QUEUE_PRIORITY, PRIORITY_LEVEL, priority);
+	mqd->cp_gfx_hqd_queue_priority = tmp;
+}
+
+static int gfx_v10_0_gfx_mqd_init(struct amdgpu_device *adev, void *m,
+				  struct amdgpu_mqd_prop *prop)
+{
+	struct v10_gfx_mqd *mqd = m;
 	uint64_t hqd_gpu_addr, wb_gpu_addr;
 	uint32_t tmp;
 	uint32_t rb_bufsz;
@@ -6602,8 +6772,8 @@ static int gfx_v10_0_gfx_mqd_init(struct amdgpu_ring *ring)
 	mqd->cp_gfx_hqd_wptr_hi = 0;
 
 	/* set the pointer to the MQD */
-	mqd->cp_mqd_base_addr = ring->mqd_gpu_addr & 0xfffffffc;
-	mqd->cp_mqd_base_addr_hi = upper_32_bits(ring->mqd_gpu_addr);
+	mqd->cp_mqd_base_addr = prop->mqd_gpu_addr & 0xfffffffc;
+	mqd->cp_mqd_base_addr_hi = upper_32_bits(prop->mqd_gpu_addr);
 
 	/* set up mqd control */
 	tmp = RREG32_SOC15(GC, 0, mmCP_GFX_MQD_CONTROL);
@@ -6617,11 +6787,8 @@ static int gfx_v10_0_gfx_mqd_init(struct amdgpu_ring *ring)
 	tmp = REG_SET_FIELD(tmp, CP_GFX_HQD_VMID, VMID, 0);
 	mqd->cp_gfx_hqd_vmid = 0;
 
-	/* set up default queue priority level
-	 * 0x0 = low priority, 0x1 = high priority */
-	tmp = RREG32_SOC15(GC, 0, mmCP_GFX_HQD_QUEUE_PRIORITY);
-	tmp = REG_SET_FIELD(tmp, CP_GFX_HQD_QUEUE_PRIORITY, PRIORITY_LEVEL, 0);
-	mqd->cp_gfx_hqd_queue_priority = tmp;
+	/* set up gfx queue priority */
+	gfx_v10_0_gfx_mqd_set_priority(adev, mqd, prop);
 
 	/* set up time quantum */
 	tmp = RREG32_SOC15(GC, 0, mmCP_GFX_HQD_QUANTUM);
@@ -6629,23 +6796,23 @@ static int gfx_v10_0_gfx_mqd_init(struct amdgpu_ring *ring)
 	mqd->cp_gfx_hqd_quantum = tmp;
 
 	/* set up gfx hqd base. this is similar as CP_RB_BASE */
-	hqd_gpu_addr = ring->gpu_addr >> 8;
+	hqd_gpu_addr = prop->hqd_base_gpu_addr >> 8;
 	mqd->cp_gfx_hqd_base = hqd_gpu_addr;
 	mqd->cp_gfx_hqd_base_hi = upper_32_bits(hqd_gpu_addr);
 
 	/* set up hqd_rptr_addr/_hi, similar as CP_RB_RPTR */
-	wb_gpu_addr = adev->wb.gpu_addr + (ring->rptr_offs * 4);
+	wb_gpu_addr = prop->rptr_gpu_addr;
 	mqd->cp_gfx_hqd_rptr_addr = wb_gpu_addr & 0xfffffffc;
 	mqd->cp_gfx_hqd_rptr_addr_hi =
 		upper_32_bits(wb_gpu_addr) & 0xffff;
 
 	/* set up rb_wptr_poll addr */
-	wb_gpu_addr = adev->wb.gpu_addr + (ring->wptr_offs * 4);
+	wb_gpu_addr = prop->wptr_gpu_addr;
 	mqd->cp_rb_wptr_poll_addr_lo = wb_gpu_addr & 0xfffffffc;
 	mqd->cp_rb_wptr_poll_addr_hi = upper_32_bits(wb_gpu_addr) & 0xffff;
 
 	/* set up the gfx_hqd_control, similar as CP_RB0_CNTL */
-	rb_bufsz = order_base_2(ring->ring_size / 4) - 1;
+	rb_bufsz = order_base_2(prop->queue_size / 4) - 1;
 	tmp = RREG32_SOC15(GC, 0, mmCP_GFX_HQD_CNTL);
 	tmp = REG_SET_FIELD(tmp, CP_GFX_HQD_CNTL, RB_BUFSZ, rb_bufsz);
 	tmp = REG_SET_FIELD(tmp, CP_GFX_HQD_CNTL, RB_BLKSZ, rb_bufsz - 2);
@@ -6656,9 +6823,9 @@ static int gfx_v10_0_gfx_mqd_init(struct amdgpu_ring *ring)
 
 	/* set up cp_doorbell_control */
 	tmp = RREG32_SOC15(GC, 0, mmCP_RB_DOORBELL_CONTROL);
-	if (ring->use_doorbell) {
+	if (prop->use_doorbell) {
 		tmp = REG_SET_FIELD(tmp, CP_RB_DOORBELL_CONTROL,
-				    DOORBELL_OFFSET, ring->doorbell_index);
+				    DOORBELL_OFFSET, prop->doorbell_index);
 		tmp = REG_SET_FIELD(tmp, CP_RB_DOORBELL_CONTROL,
 				    DOORBELL_EN, 1);
 	} else
@@ -6666,13 +6833,7 @@ static int gfx_v10_0_gfx_mqd_init(struct amdgpu_ring *ring)
 				    DOORBELL_EN, 0);
 	mqd->cp_rb_doorbell_control = tmp;
 
-	/*if there are 2 gfx rings, set the lower doorbell range of the first ring,
-	 *otherwise the range of the second ring will override the first ring */
-	if (ring->doorbell_index == adev->doorbell_index.gfx_ring0 << 1)
-		gfx_v10_0_cp_gfx_set_doorbell(adev, ring);
-
 	/* reset read and write pointers, similar to CP_RB0_WPTR/_RPTR */
-	ring->wptr = 0;
 	mqd->cp_gfx_hqd_rptr = RREG32_SOC15(GC, 0, mmCP_GFX_HQD_RPTR);
 
 	/* active the queue */
@@ -6681,175 +6842,71 @@ static int gfx_v10_0_gfx_mqd_init(struct amdgpu_ring *ring)
 	return 0;
 }
 
-#ifdef BRING_UP_DEBUG
-static int gfx_v10_0_gfx_queue_init_register(struct amdgpu_ring *ring)
-{
-	struct amdgpu_device *adev = ring->adev;
-	struct v10_gfx_mqd *mqd = ring->mqd_ptr;
-
-	/* set mmCP_GFX_HQD_WPTR/_HI to 0 */
-	WREG32_SOC15(GC, 0, mmCP_GFX_HQD_WPTR, mqd->cp_gfx_hqd_wptr);
-	WREG32_SOC15(GC, 0, mmCP_GFX_HQD_WPTR_HI, mqd->cp_gfx_hqd_wptr_hi);
-
-	/* set GFX_MQD_BASE */
-	WREG32_SOC15(GC, 0, mmCP_MQD_BASE_ADDR, mqd->cp_mqd_base_addr);
-	WREG32_SOC15(GC, 0, mmCP_MQD_BASE_ADDR_HI, mqd->cp_mqd_base_addr_hi);
-
-	/* set GFX_MQD_CONTROL */
-	WREG32_SOC15(GC, 0, mmCP_GFX_MQD_CONTROL, mqd->cp_gfx_mqd_control);
-
-	/* set GFX_HQD_VMID to 0 */
-	WREG32_SOC15(GC, 0, mmCP_GFX_HQD_VMID, mqd->cp_gfx_hqd_vmid);
-
-	WREG32_SOC15(GC, 0, mmCP_GFX_HQD_QUEUE_PRIORITY,
-			mqd->cp_gfx_hqd_queue_priority);
-	WREG32_SOC15(GC, 0, mmCP_GFX_HQD_QUANTUM, mqd->cp_gfx_hqd_quantum);
-
-	/* set GFX_HQD_BASE, similar as CP_RB_BASE */
-	WREG32_SOC15(GC, 0, mmCP_GFX_HQD_BASE, mqd->cp_gfx_hqd_base);
-	WREG32_SOC15(GC, 0, mmCP_GFX_HQD_BASE_HI, mqd->cp_gfx_hqd_base_hi);
-
-	/* set GFX_HQD_RPTR_ADDR, similar as CP_RB_RPTR */
-	WREG32_SOC15(GC, 0, mmCP_GFX_HQD_RPTR_ADDR, mqd->cp_gfx_hqd_rptr_addr);
-	WREG32_SOC15(GC, 0, mmCP_GFX_HQD_RPTR_ADDR_HI, mqd->cp_gfx_hqd_rptr_addr_hi);
-
-	/* set GFX_HQD_CNTL, similar as CP_RB_CNTL */
-	WREG32_SOC15(GC, 0, mmCP_GFX_HQD_CNTL, mqd->cp_gfx_hqd_cntl);
-
-	/* set RB_WPTR_POLL_ADDR */
-	WREG32_SOC15(GC, 0, mmCP_RB_WPTR_POLL_ADDR_LO, mqd->cp_rb_wptr_poll_addr_lo);
-	WREG32_SOC15(GC, 0, mmCP_RB_WPTR_POLL_ADDR_HI, mqd->cp_rb_wptr_poll_addr_hi);
-
-	/* set RB_DOORBELL_CONTROL */
-	WREG32_SOC15(GC, 0, mmCP_RB_DOORBELL_CONTROL, mqd->cp_rb_doorbell_control);
-
-	/* active the queue */
-	WREG32_SOC15(GC, 0, mmCP_GFX_HQD_ACTIVE, mqd->cp_gfx_hqd_active);
-
-	return 0;
-}
-#endif
-
-static int gfx_v10_0_gfx_init_queue(struct amdgpu_ring *ring)
+static int gfx_v10_0_kgq_init_queue(struct amdgpu_ring *ring, bool reset)
 {
 	struct amdgpu_device *adev = ring->adev;
 	struct v10_gfx_mqd *mqd = ring->mqd_ptr;
 	int mqd_idx = ring - &adev->gfx.gfx_ring[0];
 
-	if (!amdgpu_in_reset(adev) && !adev->in_suspend) {
+	if (!reset && !amdgpu_in_reset(adev) && !adev->in_suspend) {
 		memset((void *)mqd, 0, sizeof(*mqd));
 		mutex_lock(&adev->srbm_mutex);
 		nv_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0);
-		gfx_v10_0_gfx_mqd_init(ring);
-#ifdef BRING_UP_DEBUG
-		gfx_v10_0_gfx_queue_init_register(ring);
-#endif
+		amdgpu_ring_init_mqd(ring);
+
+		/*
+		 * if there are 2 gfx rings, set the lower doorbell
+		 * range of the first ring, otherwise the range of
+		 * the second ring will override the first ring
+		 */
+		if (ring->doorbell_index == adev->doorbell_index.gfx_ring0 << 1)
+			gfx_v10_0_cp_gfx_set_doorbell(adev, ring);
+
 		nv_grbm_select(adev, 0, 0, 0, 0);
 		mutex_unlock(&adev->srbm_mutex);
 		if (adev->gfx.me.mqd_backup[mqd_idx])
-			memcpy(adev->gfx.me.mqd_backup[mqd_idx], mqd, sizeof(*mqd));
-	} else if (amdgpu_in_reset(adev)) {
-		/* reset mqd with the backup copy */
-		if (adev->gfx.me.mqd_backup[mqd_idx])
-			memcpy(mqd, adev->gfx.me.mqd_backup[mqd_idx], sizeof(*mqd));
-		/* reset the ring */
-		ring->wptr = 0;
-		adev->wb.wb[ring->wptr_offs] = 0;
-		amdgpu_ring_clear_ring(ring);
-#ifdef BRING_UP_DEBUG
+			memcpy_fromio(adev->gfx.me.mqd_backup[mqd_idx], mqd, sizeof(*mqd));
+	} else {
 		mutex_lock(&adev->srbm_mutex);
 		nv_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0);
-		gfx_v10_0_gfx_queue_init_register(ring);
+		if (ring->doorbell_index == adev->doorbell_index.gfx_ring0 << 1)
+			gfx_v10_0_cp_gfx_set_doorbell(adev, ring);
+
 		nv_grbm_select(adev, 0, 0, 0, 0);
 		mutex_unlock(&adev->srbm_mutex);
-#endif
-	} else {
+		/* restore mqd with the backup copy */
+		if (adev->gfx.me.mqd_backup[mqd_idx])
+			memcpy_toio(mqd, adev->gfx.me.mqd_backup[mqd_idx], sizeof(*mqd));
+		/* reset the ring */
+		ring->wptr = 0;
+		*ring->wptr_cpu_addr = 0;
 		amdgpu_ring_clear_ring(ring);
 	}
 
 	return 0;
 }
 
-#ifndef BRING_UP_DEBUG
-static int gfx_v10_0_kiq_enable_kgq(struct amdgpu_device *adev)
-{
-	struct amdgpu_kiq *kiq = &adev->gfx.kiq;
-	struct amdgpu_ring *kiq_ring = &adev->gfx.kiq.ring;
-	int r, i;
-
-	if (!kiq->pmf || !kiq->pmf->kiq_map_queues)
-		return -EINVAL;
-
-	r = amdgpu_ring_alloc(kiq_ring, kiq->pmf->map_queues_size *
-					adev->gfx.num_gfx_rings);
-	if (r) {
-		DRM_ERROR("Failed to lock KIQ (%d).\n", r);
-		return r;
-	}
-
-	for (i = 0; i < adev->gfx.num_gfx_rings; i++)
-		kiq->pmf->kiq_map_queues(kiq_ring, &adev->gfx.gfx_ring[i]);
-
-	return amdgpu_ring_test_helper(kiq_ring);
-}
-#endif
-
 static int gfx_v10_0_cp_async_gfx_ring_resume(struct amdgpu_device *adev)
 {
 	int r, i;
-	struct amdgpu_ring *ring;
 
 	for (i = 0; i < adev->gfx.num_gfx_rings; i++) {
-		ring = &adev->gfx.gfx_ring[i];
-
-		r = amdgpu_bo_reserve(ring->mqd_obj, false);
-		if (unlikely(r != 0))
-			goto done;
-
-		r = amdgpu_bo_kmap(ring->mqd_obj, (void **)&ring->mqd_ptr);
-		if (!r) {
-			r = gfx_v10_0_gfx_init_queue(ring);
-			amdgpu_bo_kunmap(ring->mqd_obj);
-			ring->mqd_ptr = NULL;
-		}
-		amdgpu_bo_unreserve(ring->mqd_obj);
+		r = gfx_v10_0_kgq_init_queue(&adev->gfx.gfx_ring[i], false);
 		if (r)
-			goto done;
+			return r;
 	}
-#ifndef BRING_UP_DEBUG
-	r = gfx_v10_0_kiq_enable_kgq(adev);
-	if (r)
-		goto done;
-#endif
-	r = gfx_v10_0_cp_gfx_start(adev);
+
+	r = amdgpu_gfx_enable_kgq(adev, 0);
 	if (r)
-		goto done;
+		return r;
 
-	for (i = 0; i < adev->gfx.num_gfx_rings; i++) {
-		ring = &adev->gfx.gfx_ring[i];
-		ring->sched.ready = true;
-	}
-done:
-	return r;
+	return gfx_v10_0_cp_gfx_start(adev);
 }
 
-static void gfx_v10_0_compute_mqd_set_priority(struct amdgpu_ring *ring, struct v10_compute_mqd *mqd)
+static int gfx_v10_0_compute_mqd_init(struct amdgpu_device *adev, void *m,
+				      struct amdgpu_mqd_prop *prop)
 {
-	struct amdgpu_device *adev = ring->adev;
-
-	if (ring->funcs->type == AMDGPU_RING_TYPE_COMPUTE) {
-		if (amdgpu_gfx_is_high_priority_compute_queue(adev, ring)) {
-			mqd->cp_hqd_pipe_priority = AMDGPU_GFX_PIPE_PRIO_HIGH;
-			mqd->cp_hqd_queue_priority =
-				AMDGPU_GFX_QUEUE_PRIORITY_MAXIMUM;
-		}
-	}
-}
-
-static int gfx_v10_0_compute_mqd_init(struct amdgpu_ring *ring)
-{
-	struct amdgpu_device *adev = ring->adev;
-	struct v10_compute_mqd *mqd = ring->mqd_ptr;
+	struct v10_compute_mqd *mqd = m;
 	uint64_t hqd_gpu_addr, wb_gpu_addr, eop_base_addr;
 	uint32_t tmp;
 
@@ -6861,7 +6918,7 @@ static int gfx_v10_0_compute_mqd_init(struct amdgpu_ring *ring)
 	mqd->compute_static_thread_mgmt_se3 = 0xffffffff;
 	mqd->compute_misc_reserved = 0x00000003;
 
-	eop_base_addr = ring->eop_gpu_addr >> 8;
+	eop_base_addr = prop->eop_gpu_addr >> 8;
 	mqd->cp_hqd_eop_base_addr_lo = eop_base_addr;
 	mqd->cp_hqd_eop_base_addr_hi = upper_32_bits(eop_base_addr);
 
@@ -6875,9 +6932,9 @@ static int gfx_v10_0_compute_mqd_init(struct amdgpu_ring *ring)
 	/* enable doorbell? */
 	tmp = RREG32_SOC15(GC, 0, mmCP_HQD_PQ_DOORBELL_CONTROL);
 
-	if (ring->use_doorbell) {
+	if (prop->use_doorbell) {
 		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
-				    DOORBELL_OFFSET, ring->doorbell_index);
+				    DOORBELL_OFFSET, prop->doorbell_index);
 		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
 				    DOORBELL_EN, 1);
 		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
@@ -6892,15 +6949,14 @@ static int gfx_v10_0_compute_mqd_init(struct amdgpu_ring *ring)
 	mqd->cp_hqd_pq_doorbell_control = tmp;
 
 	/* disable the queue if it's active */
-	ring->wptr = 0;
 	mqd->cp_hqd_dequeue_request = 0;
 	mqd->cp_hqd_pq_rptr = 0;
 	mqd->cp_hqd_pq_wptr_lo = 0;
 	mqd->cp_hqd_pq_wptr_hi = 0;
 
 	/* set the pointer to the MQD */
-	mqd->cp_mqd_base_addr_lo = ring->mqd_gpu_addr & 0xfffffffc;
-	mqd->cp_mqd_base_addr_hi = upper_32_bits(ring->mqd_gpu_addr);
+	mqd->cp_mqd_base_addr_lo = prop->mqd_gpu_addr & 0xfffffffc;
+	mqd->cp_mqd_base_addr_hi = upper_32_bits(prop->mqd_gpu_addr);
 
 	/* set MQD vmid to 0 */
 	tmp = RREG32_SOC15(GC, 0, mmCP_MQD_CONTROL);
@@ -6908,55 +6964,38 @@ static int gfx_v10_0_compute_mqd_init(struct amdgpu_ring *ring)
 	mqd->cp_mqd_control = tmp;
 
 	/* set the pointer to the HQD, this is similar CP_RB0_BASE/_HI */
-	hqd_gpu_addr = ring->gpu_addr >> 8;
+	hqd_gpu_addr = prop->hqd_base_gpu_addr >> 8;
 	mqd->cp_hqd_pq_base_lo = hqd_gpu_addr;
 	mqd->cp_hqd_pq_base_hi = upper_32_bits(hqd_gpu_addr);
 
 	/* set up the HQD, this is similar to CP_RB0_CNTL */
 	tmp = RREG32_SOC15(GC, 0, mmCP_HQD_PQ_CONTROL);
 	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, QUEUE_SIZE,
-			    (order_base_2(ring->ring_size / 4) - 1));
+			    (order_base_2(prop->queue_size / 4) - 1));
 	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, RPTR_BLOCK_SIZE,
-			    ((order_base_2(AMDGPU_GPU_PAGE_SIZE / 4) - 1) << 8));
+			    (order_base_2(AMDGPU_GPU_PAGE_SIZE / 4) - 1));
 #ifdef __BIG_ENDIAN
 	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, ENDIAN_SWAP, 1);
 #endif
-	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, UNORD_DISPATCH, 0);
-	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, TUNNEL_DISPATCH, 0);
+	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, UNORD_DISPATCH, 1);
+	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, TUNNEL_DISPATCH,
+			    prop->allow_tunneling);
 	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, PRIV_STATE, 1);
 	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, KMD_QUEUE, 1);
 	mqd->cp_hqd_pq_control = tmp;
 
 	/* set the wb address whether it's enabled or not */
-	wb_gpu_addr = adev->wb.gpu_addr + (ring->rptr_offs * 4);
+	wb_gpu_addr = prop->rptr_gpu_addr;
 	mqd->cp_hqd_pq_rptr_report_addr_lo = wb_gpu_addr & 0xfffffffc;
 	mqd->cp_hqd_pq_rptr_report_addr_hi =
 		upper_32_bits(wb_gpu_addr) & 0xffff;
 
 	/* only used if CP_PQ_WPTR_POLL_CNTL.CP_PQ_WPTR_POLL_CNTL__EN_MASK=1 */
-	wb_gpu_addr = adev->wb.gpu_addr + (ring->wptr_offs * 4);
+	wb_gpu_addr = prop->wptr_gpu_addr;
 	mqd->cp_hqd_pq_wptr_poll_addr_lo = wb_gpu_addr & 0xfffffffc;
 	mqd->cp_hqd_pq_wptr_poll_addr_hi = upper_32_bits(wb_gpu_addr) & 0xffff;
 
-	tmp = 0;
-	/* enable the doorbell if requested */
-	if (ring->use_doorbell) {
-		tmp = RREG32_SOC15(GC, 0, mmCP_HQD_PQ_DOORBELL_CONTROL);
-		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
-				DOORBELL_OFFSET, ring->doorbell_index);
-
-		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
-				    DOORBELL_EN, 1);
-		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
-				    DOORBELL_SOURCE, 0);
-		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
-				    DOORBELL_HIT, 0);
-	}
-
-	mqd->cp_hqd_pq_doorbell_control = tmp;
-
 	/* reset read and write pointers, similar to CP_RB0_WPTR/_RPTR */
-	ring->wptr = 0;
 	mqd->cp_hqd_pq_rptr = RREG32_SOC15(GC, 0, mmCP_HQD_PQ_RPTR);
 
 	/* set the vmid for the queue */
@@ -6972,13 +7011,10 @@ static int gfx_v10_0_compute_mqd_init(struct amdgpu_ring *ring)
 	mqd->cp_hqd_ib_control = tmp;
 
 	/* set static priority for a compute queue/ring */
-	gfx_v10_0_compute_mqd_set_priority(ring, mqd);
+	mqd->cp_hqd_pipe_priority = prop->hqd_pipe_priority;
+	mqd->cp_hqd_queue_priority = prop->hqd_queue_priority;
 
-	/* map_queues packet doesn't need activate the queue,
-	 * so only kiq need set this field.
-	 */
-	if (ring->funcs->type == AMDGPU_RING_TYPE_KIQ)
-		mqd->cp_hqd_active = 1;
+	mqd->cp_hqd_active = prop->hqd_active;
 
 	return 0;
 }
@@ -6996,20 +7032,6 @@ static int gfx_v10_0_kiq_init_register(struct amdgpu_ring *ring)
 	/* disable wptr polling */
 	WREG32_FIELD15(GC, 0, CP_PQ_WPTR_POLL_CNTL, EN, 0);
 
-	/* write the EOP addr */
-	WREG32_SOC15(GC, 0, mmCP_HQD_EOP_BASE_ADDR,
-	       mqd->cp_hqd_eop_base_addr_lo);
-	WREG32_SOC15(GC, 0, mmCP_HQD_EOP_BASE_ADDR_HI,
-	       mqd->cp_hqd_eop_base_addr_hi);
-
-	/* set the EOP size, register value is 2^(EOP_SIZE+1) dwords */
-	WREG32_SOC15(GC, 0, mmCP_HQD_EOP_CONTROL,
-	       mqd->cp_hqd_eop_control);
-
-	/* enable doorbell? */
-	WREG32_SOC15(GC, 0, mmCP_HQD_PQ_DOORBELL_CONTROL,
-	       mqd->cp_hqd_pq_doorbell_control);
-
 	/* disable the queue if it's active */
 	if (RREG32_SOC15(GC, 0, mmCP_HQD_ACTIVE) & 1) {
 		WREG32_SOC15(GC, 0, mmCP_HQD_DEQUEUE_REQUEST, 1);
@@ -7028,6 +7050,19 @@ static int gfx_v10_0_kiq_init_register(struct amdgpu_ring *ring)
 		       mqd->cp_hqd_pq_wptr_hi);
 	}
 
+	/* disable doorbells */
+	WREG32_SOC15(GC, 0, mmCP_HQD_PQ_DOORBELL_CONTROL, 0);
+
+	/* write the EOP addr */
+	WREG32_SOC15(GC, 0, mmCP_HQD_EOP_BASE_ADDR,
+	       mqd->cp_hqd_eop_base_addr_lo);
+	WREG32_SOC15(GC, 0, mmCP_HQD_EOP_BASE_ADDR_HI,
+	       mqd->cp_hqd_eop_base_addr_hi);
+
+	/* set the EOP size, register value is 2^(EOP_SIZE+1) dwords */
+	WREG32_SOC15(GC, 0, mmCP_HQD_EOP_CONTROL,
+	       mqd->cp_hqd_eop_control);
+
 	/* set the pointer to the MQD */
 	WREG32_SOC15(GC, 0, mmCP_MQD_BASE_ADDR,
 	       mqd->cp_mqd_base_addr_lo);
@@ -7097,14 +7132,13 @@ static int gfx_v10_0_kiq_init_queue(struct amdgpu_ring *ring)
 {
 	struct amdgpu_device *adev = ring->adev;
 	struct v10_compute_mqd *mqd = ring->mqd_ptr;
-	int mqd_idx = AMDGPU_MAX_COMPUTE_RINGS;
 
 	gfx_v10_0_kiq_setting(ring);
 
 	if (amdgpu_in_reset(adev)) { /* for GPU_RESET case */
 		/* reset MQD to a clean status */
-		if (adev->gfx.mec.mqd_backup[mqd_idx])
-			memcpy(mqd, adev->gfx.mec.mqd_backup[mqd_idx], sizeof(*mqd));
+		if (adev->gfx.kiq[0].mqd_backup)
+			memcpy_toio(mqd, adev->gfx.kiq[0].mqd_backup, sizeof(*mqd));
 
 		/* reset ring buffer */
 		ring->wptr = 0;
@@ -7117,46 +7151,45 @@ static int gfx_v10_0_kiq_init_queue(struct amdgpu_ring *ring)
 		mutex_unlock(&adev->srbm_mutex);
 	} else {
 		memset((void *)mqd, 0, sizeof(*mqd));
+		if (amdgpu_sriov_vf(adev) && adev->in_suspend)
+			amdgpu_ring_clear_ring(ring);
 		mutex_lock(&adev->srbm_mutex);
 		nv_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0);
-		gfx_v10_0_compute_mqd_init(ring);
+		amdgpu_ring_init_mqd(ring);
 		gfx_v10_0_kiq_init_register(ring);
 		nv_grbm_select(adev, 0, 0, 0, 0);
 		mutex_unlock(&adev->srbm_mutex);
 
-		if (adev->gfx.mec.mqd_backup[mqd_idx])
-			memcpy(adev->gfx.mec.mqd_backup[mqd_idx], mqd, sizeof(*mqd));
+		if (adev->gfx.kiq[0].mqd_backup)
+			memcpy_fromio(adev->gfx.kiq[0].mqd_backup, mqd, sizeof(*mqd));
 	}
 
 	return 0;
 }
 
-static int gfx_v10_0_kcq_init_queue(struct amdgpu_ring *ring)
+static int gfx_v10_0_kcq_init_queue(struct amdgpu_ring *ring, bool restore)
 {
 	struct amdgpu_device *adev = ring->adev;
 	struct v10_compute_mqd *mqd = ring->mqd_ptr;
 	int mqd_idx = ring - &adev->gfx.compute_ring[0];
 
-	if (!amdgpu_in_reset(adev) && !adev->in_suspend) {
+	if (!restore && !amdgpu_in_reset(adev) && !adev->in_suspend) {
 		memset((void *)mqd, 0, sizeof(*mqd));
 		mutex_lock(&adev->srbm_mutex);
 		nv_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0);
-		gfx_v10_0_compute_mqd_init(ring);
+		amdgpu_ring_init_mqd(ring);
 		nv_grbm_select(adev, 0, 0, 0, 0);
 		mutex_unlock(&adev->srbm_mutex);
 
 		if (adev->gfx.mec.mqd_backup[mqd_idx])
-			memcpy(adev->gfx.mec.mqd_backup[mqd_idx], mqd, sizeof(*mqd));
-	} else if (amdgpu_in_reset(adev)) { /* for GPU_RESET case */
-		/* reset MQD to a clean status */
+			memcpy_fromio(adev->gfx.mec.mqd_backup[mqd_idx], mqd, sizeof(*mqd));
+	} else {
+		/* restore MQD to a clean status */
 		if (adev->gfx.mec.mqd_backup[mqd_idx])
-			memcpy(mqd, adev->gfx.mec.mqd_backup[mqd_idx], sizeof(*mqd));
-
+			memcpy_toio(mqd, adev->gfx.mec.mqd_backup[mqd_idx], sizeof(*mqd));
 		/* reset ring buffer */
 		ring->wptr = 0;
-		atomic64_set((atomic64_t *)&adev->wb.wb[ring->wptr_offs], 0);
-		amdgpu_ring_clear_ring(ring);
-	} else {
+		atomic64_set((atomic64_t *)ring->wptr_cpu_addr, 0);
 		amdgpu_ring_clear_ring(ring);
 	}
 
@@ -7165,54 +7198,24 @@ static int gfx_v10_0_kcq_init_queue(struct amdgpu_ring *ring)
 
 static int gfx_v10_0_kiq_resume(struct amdgpu_device *adev)
 {
-	struct amdgpu_ring *ring;
-	int r;
-
-	ring = &adev->gfx.kiq.ring;
-
-	r = amdgpu_bo_reserve(ring->mqd_obj, false);
-	if (unlikely(r != 0))
-		return r;
-
-	r = amdgpu_bo_kmap(ring->mqd_obj, (void **)&ring->mqd_ptr);
-	if (unlikely(r != 0))
-		return r;
-
-	gfx_v10_0_kiq_init_queue(ring);
-	amdgpu_bo_kunmap(ring->mqd_obj);
-	ring->mqd_ptr = NULL;
-	amdgpu_bo_unreserve(ring->mqd_obj);
-	ring->sched.ready = true;
+	gfx_v10_0_kiq_init_queue(&adev->gfx.kiq[0].ring);
 	return 0;
 }
 
 static int gfx_v10_0_kcq_resume(struct amdgpu_device *adev)
 {
-	struct amdgpu_ring *ring = NULL;
-	int r = 0, i;
+	int i, r;
 
 	gfx_v10_0_cp_compute_enable(adev, true);
 
 	for (i = 0; i < adev->gfx.num_compute_rings; i++) {
-		ring = &adev->gfx.compute_ring[i];
-
-		r = amdgpu_bo_reserve(ring->mqd_obj, false);
-		if (unlikely(r != 0))
-			goto done;
-		r = amdgpu_bo_kmap(ring->mqd_obj, (void **)&ring->mqd_ptr);
-		if (!r) {
-			r = gfx_v10_0_kcq_init_queue(ring);
-			amdgpu_bo_kunmap(ring->mqd_obj);
-			ring->mqd_ptr = NULL;
-		}
-		amdgpu_bo_unreserve(ring->mqd_obj);
+		r = gfx_v10_0_kcq_init_queue(&adev->gfx.compute_ring[i],
+					     false);
 		if (r)
-			goto done;
+			return r;
 	}
 
-	r = amdgpu_gfx_enable_kcq(adev);
-done:
-	return r;
+	return amdgpu_gfx_enable_kcq(adev, 0);
 }
 
 static int gfx_v10_0_cp_resume(struct amdgpu_device *adev)
@@ -7279,9 +7282,11 @@ static bool gfx_v10_0_check_grbm_cam_remapping(struct amdgpu_device *adev)
 {
 	uint32_t data, pattern = 0xDEADBEEF;
 
-	/* check if mmVGT_ESGS_RING_SIZE_UMD
-	 * has been remapped to mmVGT_ESGS_RING_SIZE */
-	switch (adev->ip_versions[GC_HWIP][0]) {
+	/*
+	 * check if mmVGT_ESGS_RING_SIZE_UMD
+	 * has been remapped to mmVGT_ESGS_RING_SIZE
+	 */
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 	case IP_VERSION(10, 3, 0):
 	case IP_VERSION(10, 3, 2):
 	case IP_VERSION(10, 3, 4):
@@ -7291,15 +7296,15 @@ static bool gfx_v10_0_check_grbm_cam_remapping(struct amdgpu_device *adev)
 		WREG32_SOC15(GC, 0, mmVGT_ESGS_RING_SIZE_UMD, pattern);
 
 		if (RREG32_SOC15(GC, 0, mmVGT_ESGS_RING_SIZE_Sienna_Cichlid) == pattern) {
-			WREG32_SOC15(GC, 0, mmVGT_ESGS_RING_SIZE_UMD , data);
+			WREG32_SOC15(GC, 0, mmVGT_ESGS_RING_SIZE_UMD, data);
 			return true;
-		} else {
-			WREG32_SOC15(GC, 0, mmVGT_ESGS_RING_SIZE_Sienna_Cichlid, data);
-			return false;
 		}
+		WREG32_SOC15(GC, 0, mmVGT_ESGS_RING_SIZE_Sienna_Cichlid, data);
 		break;
 	case IP_VERSION(10, 3, 1):
 	case IP_VERSION(10, 3, 3):
+	case IP_VERSION(10, 3, 6):
+	case IP_VERSION(10, 3, 7):
 		return true;
 	default:
 		data = RREG32_SOC15(GC, 0, mmVGT_ESGS_RING_SIZE);
@@ -7309,12 +7314,12 @@ static bool gfx_v10_0_check_grbm_cam_remapping(struct amdgpu_device *adev)
 		if (RREG32_SOC15(GC, 0, mmVGT_ESGS_RING_SIZE) == pattern) {
 			WREG32_SOC15(GC, 0, mmVGT_ESGS_RING_SIZE_UMD, data);
 			return true;
-		} else {
-			WREG32_SOC15(GC, 0, mmVGT_ESGS_RING_SIZE, data);
-			return false;
 		}
+		WREG32_SOC15(GC, 0, mmVGT_ESGS_RING_SIZE, data);
 		break;
 	}
+
+	return false;
 }
 
 static void gfx_v10_0_setup_grbm_cam_remapping(struct amdgpu_device *adev)
@@ -7324,17 +7329,21 @@ static void gfx_v10_0_setup_grbm_cam_remapping(struct amdgpu_device *adev)
 	if (amdgpu_sriov_vf(adev))
 		return;
 
-	/* initialize cam_index to 0
-	 * index will auto-inc after each data writting */
+	/*
+	 * Initialize cam_index to 0
+	 * index will auto-inc after each data writing
+	 */
 	WREG32_SOC15(GC, 0, mmGRBM_CAM_INDEX, 0);
 
-	switch (adev->ip_versions[GC_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 	case IP_VERSION(10, 3, 0):
 	case IP_VERSION(10, 3, 2):
 	case IP_VERSION(10, 3, 1):
 	case IP_VERSION(10, 3, 4):
 	case IP_VERSION(10, 3, 5):
+	case IP_VERSION(10, 3, 6):
 	case IP_VERSION(10, 3, 3):
+	case IP_VERSION(10, 3, 7):
 		/* mmVGT_TF_RING_SIZE_UMD -> mmVGT_TF_RING_SIZE */
 		data = (SOC15_REG_OFFSET(GC, 0, mmVGT_TF_RING_SIZE_UMD) <<
 			GRBM_CAM_DATA__CAM_ADDR__SHIFT) |
@@ -7453,6 +7462,7 @@ static void gfx_v10_0_setup_grbm_cam_remapping(struct amdgpu_device *adev)
 static void gfx_v10_0_disable_gpa_mode(struct amdgpu_device *adev)
 {
 	uint32_t data;
+
 	data = RREG32_SOC15(GC, 0, mmCPC_PSP_DEBUG);
 	data |= CPC_PSP_DEBUG__GPA_OVERRIDE_MASK;
 	WREG32_SOC15(GC, 0, mmCPC_PSP_DEBUG, data);
@@ -7462,25 +7472,26 @@ static void gfx_v10_0_disable_gpa_mode(struct amdgpu_device *adev)
 	WREG32_SOC15(GC, 0, mmCPG_PSP_DEBUG, data);
 }
 
-static int gfx_v10_0_hw_init(void *handle)
+static int gfx_v10_0_hw_init(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (!amdgpu_emu_mode)
 		gfx_v10_0_init_golden_registers(adev);
 
+	amdgpu_gfx_cleaner_shader_init(adev, adev->gfx.cleaner_shader_size,
+				       adev->gfx.cleaner_shader_ptr);
+
 	if (adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT) {
 		/**
 		 * For gfx 10, rlc firmware loading relies on smu firmware is
 		 * loaded firstly, so in direct type, it has to load smc ucode
 		 * here before rlc.
 		 */
-		if (!(adev->flags & AMD_IS_APU)) {
-			r = amdgpu_pm_load_smu_firmware(adev, NULL);
-			if (r)
-				return r;
-		}
+		r = amdgpu_pm_load_smu_firmware(adev, NULL);
+		if (r)
+			return r;
 		gfx_v10_0_disable_gpa_mode(adev);
 	}
 
@@ -7498,80 +7509,56 @@ static int gfx_v10_0_hw_init(void *handle)
 	 * init golden registers and rlc resume may override some registers,
 	 * reconfig them here
 	 */
-	if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 1, 10) ||
-	    adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 1, 1) ||
-	    adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 1, 2))
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(10, 1, 10) ||
+	    amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(10, 1, 1) ||
+	    amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(10, 1, 2))
 		gfx_v10_0_tcp_harvest(adev);
 
 	r = gfx_v10_0_cp_resume(adev);
 	if (r)
 		return r;
 
-	if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 3, 0))
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(10, 3, 0))
 		gfx_v10_3_program_pbb_mode(adev);
 
-	if (adev->ip_versions[GC_HWIP][0] >= IP_VERSION(10, 3, 0))
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) >= IP_VERSION(10, 3, 0) && !amdgpu_sriov_vf(adev))
 		gfx_v10_3_set_power_brake_sequence(adev);
 
 	return r;
 }
 
-#ifndef BRING_UP_DEBUG
-static int gfx_v10_0_kiq_disable_kgq(struct amdgpu_device *adev)
+static int gfx_v10_0_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_kiq *kiq = &adev->gfx.kiq;
-	struct amdgpu_ring *kiq_ring = &kiq->ring;
-	int i;
+	struct amdgpu_device *adev = ip_block->adev;
 
-	if (!kiq->pmf || !kiq->pmf->kiq_unmap_queues)
-		return -EINVAL;
-
-	if (amdgpu_ring_alloc(kiq_ring, kiq->pmf->unmap_queues_size *
-					adev->gfx.num_gfx_rings))
-		return -ENOMEM;
-
-	for (i = 0; i < adev->gfx.num_gfx_rings; i++)
-		kiq->pmf->kiq_unmap_queues(kiq_ring, &adev->gfx.gfx_ring[i],
-					   PREEMPT_QUEUES, 0, 0);
-
-	return amdgpu_ring_test_helper(kiq_ring);
-}
-#endif
-
-static int gfx_v10_0_hw_fini(void *handle)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-	int r;
-	uint32_t tmp;
+	cancel_delayed_work_sync(&adev->gfx.idle_work);
 
 	amdgpu_irq_put(adev, &adev->gfx.priv_reg_irq, 0);
 	amdgpu_irq_put(adev, &adev->gfx.priv_inst_irq, 0);
+	amdgpu_irq_put(adev, &adev->gfx.bad_op_irq, 0);
+
+	/* WA added for Vangogh asic fixing the SMU suspend failure
+	 * It needs to set power gating again during gfxoff control
+	 * otherwise the gfxoff disallowing will be failed to set.
+	 */
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(10, 3, 1))
+		gfx_v10_0_set_powergating_state(ip_block, AMD_PG_STATE_UNGATE);
 
 	if (!adev->no_hw_access) {
-#ifndef BRING_UP_DEBUG
 		if (amdgpu_async_gfx_ring) {
-			r = gfx_v10_0_kiq_disable_kgq(adev);
-			if (r)
+			if (amdgpu_gfx_disable_kgq(adev, 0))
 				DRM_ERROR("KGQ disable failed\n");
 		}
-#endif
-		if (amdgpu_gfx_disable_kcq(adev))
+
+		if (amdgpu_gfx_disable_kcq(adev, 0))
 			DRM_ERROR("KCQ disable failed\n");
 	}
 
 	if (amdgpu_sriov_vf(adev)) {
 		gfx_v10_0_cp_gfx_enable(adev, false);
-		/* Program KIQ position of RLC_CP_SCHEDULERS during destroy */
-		if (adev->ip_versions[GC_HWIP][0] >= IP_VERSION(10, 3, 0)) {
-			tmp = RREG32_SOC15(GC, 0, mmRLC_CP_SCHEDULERS_Sienna_Cichlid);
-			tmp &= 0xffffff00;
-			WREG32_SOC15(GC, 0, mmRLC_CP_SCHEDULERS_Sienna_Cichlid, tmp);
-		} else {
-			tmp = RREG32_SOC15(GC, 0, mmRLC_CP_SCHEDULERS);
-			tmp &= 0xffffff00;
-			WREG32_SOC15(GC, 0, mmRLC_CP_SCHEDULERS, tmp);
-		}
-
+		/* Remove the steps of clearing KIQ position.
+		 * It causes GFX hang when another Win guest is rendering.
+		 */
 		return 0;
 	}
 	gfx_v10_0_cp_enable(adev, false);
@@ -7580,19 +7567,19 @@ static int gfx_v10_0_hw_fini(void *handle)
 	return 0;
 }
 
-static int gfx_v10_0_suspend(void *handle)
+static int gfx_v10_0_suspend(struct amdgpu_ip_block *ip_block)
 {
-	return gfx_v10_0_hw_fini(handle);
+	return gfx_v10_0_hw_fini(ip_block);
 }
 
-static int gfx_v10_0_resume(void *handle)
+static int gfx_v10_0_resume(struct amdgpu_ip_block *ip_block)
 {
-	return gfx_v10_0_hw_init(handle);
+	return gfx_v10_0_hw_init(ip_block);
 }
 
-static bool gfx_v10_0_is_idle(void *handle)
+static bool gfx_v10_0_is_idle(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (REG_GET_FIELD(RREG32_SOC15(GC, 0, mmGRBM_STATUS),
 				GRBM_STATUS, GUI_ACTIVE))
@@ -7601,11 +7588,11 @@ static bool gfx_v10_0_is_idle(void *handle)
 		return true;
 }
 
-static int gfx_v10_0_wait_for_idle(void *handle)
+static int gfx_v10_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
 {
-	unsigned i;
+	unsigned int i;
 	u32 tmp;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	for (i = 0; i < adev->usec_timeout; i++) {
 		/* read MC_STATUS */
@@ -7619,11 +7606,11 @@ static int gfx_v10_0_wait_for_idle(void *handle)
 	return -ETIMEDOUT;
 }
 
-static int gfx_v10_0_soft_reset(void *handle)
+static int gfx_v10_0_soft_reset(struct amdgpu_ip_block *ip_block)
 {
 	u32 grbm_soft_reset = 0;
 	u32 tmp;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	/* GRBM_STATUS */
 	tmp = RREG32_SOC15(GC, 0, mmGRBM_STATUS);
@@ -7648,12 +7635,13 @@ static int gfx_v10_0_soft_reset(void *handle)
 
 	/* GRBM_STATUS2 */
 	tmp = RREG32_SOC15(GC, 0, mmGRBM_STATUS2);
-	switch (adev->ip_versions[GC_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 	case IP_VERSION(10, 3, 0):
 	case IP_VERSION(10, 3, 2):
 	case IP_VERSION(10, 3, 1):
 	case IP_VERSION(10, 3, 4):
 	case IP_VERSION(10, 3, 5):
+	case IP_VERSION(10, 3, 6):
 	case IP_VERSION(10, 3, 3):
 		if (REG_GET_FIELD(tmp, GRBM_STATUS2, RLC_BUSY_Sienna_Cichlid))
 			grbm_soft_reset = REG_SET_FIELD(grbm_soft_reset,
@@ -7680,19 +7668,17 @@ static int gfx_v10_0_soft_reset(void *handle)
 		/* Disable MEC parsing/prefetching */
 		gfx_v10_0_cp_compute_enable(adev, false);
 
-		if (grbm_soft_reset) {
-			tmp = RREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET);
-			tmp |= grbm_soft_reset;
-			dev_info(adev->dev, "GRBM_SOFT_RESET=0x%08X\n", tmp);
-			WREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET, tmp);
-			tmp = RREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET);
+		tmp = RREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET);
+		tmp |= grbm_soft_reset;
+		dev_info(adev->dev, "GRBM_SOFT_RESET=0x%08X\n", tmp);
+		WREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET, tmp);
+		tmp = RREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET);
 
-			udelay(50);
+		udelay(50);
 
-			tmp &= ~grbm_soft_reset;
-			WREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET, tmp);
-			tmp = RREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET);
-		}
+		tmp &= ~grbm_soft_reset;
+		WREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET, tmp);
+		tmp = RREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET);
 
 		/* Wait a little for things to settle down */
 		udelay(50);
@@ -7704,9 +7690,26 @@ static uint64_t gfx_v10_0_get_gpu_clock_counter(struct amdgpu_device *adev)
 {
 	uint64_t clock, clock_lo, clock_hi, hi_check;
 
-	switch (adev->ip_versions[GC_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	case IP_VERSION(10, 1, 3):
+	case IP_VERSION(10, 1, 4):
+		preempt_disable();
+		clock_hi = RREG32_SOC15_NO_KIQ(SMUIO, 0, mmGOLDEN_TSC_COUNT_UPPER_Cyan_Skillfish);
+		clock_lo = RREG32_SOC15_NO_KIQ(SMUIO, 0, mmGOLDEN_TSC_COUNT_LOWER_Cyan_Skillfish);
+		hi_check = RREG32_SOC15_NO_KIQ(SMUIO, 0, mmGOLDEN_TSC_COUNT_UPPER_Cyan_Skillfish);
+		/* The SMUIO TSC clock frequency is 100MHz, which sets 32-bit carry over
+		 * roughly every 42 seconds.
+		 */
+		if (hi_check != clock_hi) {
+			clock_lo = RREG32_SOC15_NO_KIQ(SMUIO, 0, mmGOLDEN_TSC_COUNT_LOWER_Cyan_Skillfish);
+			clock_hi = hi_check;
+		}
+		preempt_enable();
+		clock = clock_lo | (clock_hi << 32ULL);
+		break;
 	case IP_VERSION(10, 3, 1):
 	case IP_VERSION(10, 3, 3):
+	case IP_VERSION(10, 3, 7):
 		preempt_disable();
 		clock_hi = RREG32_SOC15_NO_KIQ(SMUIO, 0, mmGOLDEN_TSC_COUNT_UPPER_Vangogh);
 		clock_lo = RREG32_SOC15_NO_KIQ(SMUIO, 0, mmGOLDEN_TSC_COUNT_LOWER_Vangogh);
@@ -7721,6 +7724,21 @@ static uint64_t gfx_v10_0_get_gpu_clock_counter(struct amdgpu_device *adev)
 		preempt_enable();
 		clock = clock_lo | (clock_hi << 32ULL);
 		break;
+	case IP_VERSION(10, 3, 6):
+		preempt_disable();
+		clock_hi = RREG32_SOC15_NO_KIQ(SMUIO, 0, mmGOLDEN_TSC_COUNT_UPPER_GC_10_3_6);
+		clock_lo = RREG32_SOC15_NO_KIQ(SMUIO, 0, mmGOLDEN_TSC_COUNT_LOWER_GC_10_3_6);
+		hi_check = RREG32_SOC15_NO_KIQ(SMUIO, 0, mmGOLDEN_TSC_COUNT_UPPER_GC_10_3_6);
+		/* The SMUIO TSC clock frequency is 100MHz, which sets 32-bit carry over
+		 * roughly every 42 seconds.
+		 */
+		if (hi_check != clock_hi) {
+			clock_lo = RREG32_SOC15_NO_KIQ(SMUIO, 0, mmGOLDEN_TSC_COUNT_LOWER_GC_10_3_6);
+			clock_hi = hi_check;
+		}
+		preempt_enable();
+		clock = clock_lo | (clock_hi << 32ULL);
+		break;
 	default:
 		preempt_disable();
 		clock_hi = RREG32_SOC15_NO_KIQ(SMUIO, 0, mmGOLDEN_TSC_COUNT_UPPER);
@@ -7769,15 +7787,18 @@ static void gfx_v10_0_ring_emit_gds_switch(struct amdgpu_ring *ring,
 				    (1 << (oa_size + oa_base)) - (1 << oa_base));
 }
 
-static int gfx_v10_0_early_init(void *handle)
+static int gfx_v10_0_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
-	switch (adev->ip_versions[GC_HWIP][0]) {
+	adev->gfx.funcs = &gfx_v10_0_gfx_funcs;
+
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 	case IP_VERSION(10, 1, 10):
 	case IP_VERSION(10, 1, 1):
 	case IP_VERSION(10, 1, 2):
 	case IP_VERSION(10, 1, 3):
+	case IP_VERSION(10, 1, 4):
 		adev->gfx.num_gfx_rings = GFX10_NUM_GFX_RINGS_NV1X;
 		break;
 	case IP_VERSION(10, 3, 0):
@@ -7785,7 +7806,9 @@ static int gfx_v10_0_early_init(void *handle)
 	case IP_VERSION(10, 3, 1):
 	case IP_VERSION(10, 3, 4):
 	case IP_VERSION(10, 3, 5):
+	case IP_VERSION(10, 3, 6):
 	case IP_VERSION(10, 3, 3):
+	case IP_VERSION(10, 3, 7):
 		adev->gfx.num_gfx_rings = GFX10_NUM_GFX_RINGS_Sienna_Cichlid;
 		break;
 	default:
@@ -7800,13 +7823,17 @@ static int gfx_v10_0_early_init(void *handle)
 	gfx_v10_0_set_irq_funcs(adev);
 	gfx_v10_0_set_gds_init(adev);
 	gfx_v10_0_set_rlc_funcs(adev);
+	gfx_v10_0_set_mqd_funcs(adev);
 
-	return 0;
+	/* init rlcg reg access ctrl */
+	gfx_v10_0_init_rlcg_reg_access_ctrl(adev);
+
+	return gfx_v10_0_init_microcode(adev);
 }
 
-static int gfx_v10_0_late_init(void *handle)
+static int gfx_v10_0_late_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int r;
 
 	r = amdgpu_irq_get(adev, &adev->gfx.priv_reg_irq, 0);
@@ -7817,6 +7844,10 @@ static int gfx_v10_0_late_init(void *handle)
 	if (r)
 		return r;
 
+	r = amdgpu_irq_get(adev, &adev->gfx.bad_op_irq, 0);
+	if (r)
+		return r;
+
 	return 0;
 }
 
@@ -7829,21 +7860,23 @@ static bool gfx_v10_0_is_rlc_enabled(struct amdgpu_device *adev)
 	return (REG_GET_FIELD(rlc_cntl, RLC_CNTL, RLC_ENABLE_F32)) ? true : false;
 }
 
-static void gfx_v10_0_set_safe_mode(struct amdgpu_device *adev)
+static void gfx_v10_0_set_safe_mode(struct amdgpu_device *adev, int xcc_id)
 {
 	uint32_t data;
-	unsigned i;
+	unsigned int i;
 
 	data = RLC_SAFE_MODE__CMD_MASK;
 	data |= (1 << RLC_SAFE_MODE__MESSAGE__SHIFT);
 
-	switch (adev->ip_versions[GC_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 	case IP_VERSION(10, 3, 0):
 	case IP_VERSION(10, 3, 2):
 	case IP_VERSION(10, 3, 1):
 	case IP_VERSION(10, 3, 4):
 	case IP_VERSION(10, 3, 5):
+	case IP_VERSION(10, 3, 6):
 	case IP_VERSION(10, 3, 3):
+	case IP_VERSION(10, 3, 7):
 		WREG32_SOC15(GC, 0, mmRLC_SAFE_MODE_Sienna_Cichlid, data);
 
 		/* wait for RLC_SAFE_MODE */
@@ -7868,18 +7901,20 @@ static void gfx_v10_0_set_safe_mode(struct amdgpu_device *adev)
 	}
 }
 
-static void gfx_v10_0_unset_safe_mode(struct amdgpu_device *adev)
+static void gfx_v10_0_unset_safe_mode(struct amdgpu_device *adev, int xcc_id)
 {
 	uint32_t data;
 
 	data = RLC_SAFE_MODE__CMD_MASK;
-	switch (adev->ip_versions[GC_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 	case IP_VERSION(10, 3, 0):
 	case IP_VERSION(10, 3, 2):
 	case IP_VERSION(10, 3, 1):
 	case IP_VERSION(10, 3, 4):
 	case IP_VERSION(10, 3, 5):
+	case IP_VERSION(10, 3, 6):
 	case IP_VERSION(10, 3, 3):
+	case IP_VERSION(10, 3, 7):
 		WREG32_SOC15(GC, 0, mmRLC_SAFE_MODE_Sienna_Cichlid, data);
 		break;
 	default:
@@ -8182,7 +8217,7 @@ static void gfx_v10_0_apply_medium_grain_clock_gating_workaround(struct amdgpu_d
 		mmCGTS_SA1_QUAD1_SM_CTRL_REG
 	};
 
-	if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 1, 2)) {
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(10, 1, 2)) {
 		for (i = 0; i < ARRAY_SIZE(tcp_ctrl_regs_nv12); i++) {
 			reg_idx = adev->reg_offset[GC_HWIP][0][mmCGTS_SA0_WGP00_CU0_TCP_CTRL_REG_BASE_IDX] +
 				  tcp_ctrl_regs_nv12[i];
@@ -8213,7 +8248,7 @@ static void gfx_v10_0_apply_medium_grain_clock_gating_workaround(struct amdgpu_d
 static int gfx_v10_0_update_gfx_clock_gating(struct amdgpu_device *adev,
 					    bool enable)
 {
-	amdgpu_gfx_rlc_enter_safe_mode(adev);
+	amdgpu_gfx_rlc_enter_safe_mode(adev, 0);
 
 	if (enable) {
 		/* enable FGCG firstly*/
@@ -8227,9 +8262,12 @@ static int gfx_v10_0_update_gfx_clock_gating(struct amdgpu_device *adev,
 		/* ===  CGCG + CGLS === */
 		gfx_v10_0_update_coarse_grain_clock_gating(adev, enable);
 
-		if ((adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 1, 10)) ||
-		    (adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 1, 1)) ||
-		    (adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 1, 2)))
+		if ((amdgpu_ip_version(adev, GC_HWIP, 0) ==
+		     IP_VERSION(10, 1, 10)) ||
+		    (amdgpu_ip_version(adev, GC_HWIP, 0) ==
+		     IP_VERSION(10, 1, 1)) ||
+		    (amdgpu_ip_version(adev, GC_HWIP, 0) ==
+		     IP_VERSION(10, 1, 2)))
 			gfx_v10_0_apply_medium_grain_clock_gating_workaround(adev);
 	} else {
 		/* CGCG/CGLS should be disabled before MGCG/MGLS
@@ -8252,31 +8290,39 @@ static int gfx_v10_0_update_gfx_clock_gating(struct amdgpu_device *adev,
 	     AMD_CG_SUPPORT_GFX_3D_CGLS))
 		gfx_v10_0_enable_gui_idle_interrupt(adev, enable);
 
-	amdgpu_gfx_rlc_exit_safe_mode(adev);
+	amdgpu_gfx_rlc_exit_safe_mode(adev, 0);
 
 	return 0;
 }
 
-static void gfx_v10_0_update_spm_vmid(struct amdgpu_device *adev, unsigned vmid)
+static void gfx_v10_0_update_spm_vmid_internal(struct amdgpu_device *adev,
+					       unsigned int vmid)
 {
-	u32 reg, data;
-
-	amdgpu_gfx_off_ctrl(adev, false);
+	u32 reg, pre_data, data;
 
-	/* not for *_SOC15 */
 	reg = SOC15_REG_OFFSET(GC, 0, mmRLC_SPM_MC_CNTL);
-	if (amdgpu_sriov_is_pp_one_vf(adev))
-		data = RREG32_NO_KIQ(reg);
+	/* not for *_SOC15 */
+	if (amdgpu_sriov_is_pp_one_vf(adev) && !amdgpu_sriov_runtime(adev))
+		pre_data = RREG32_NO_KIQ(reg);
 	else
-		data = RREG32_SOC15(GC, 0, mmRLC_SPM_MC_CNTL);
+		pre_data = RREG32(reg);
 
-	data &= ~RLC_SPM_MC_CNTL__RLC_SPM_VMID_MASK;
+	data = pre_data & (~RLC_SPM_MC_CNTL__RLC_SPM_VMID_MASK);
 	data |= (vmid & RLC_SPM_MC_CNTL__RLC_SPM_VMID_MASK) << RLC_SPM_MC_CNTL__RLC_SPM_VMID__SHIFT;
 
-	if (amdgpu_sriov_is_pp_one_vf(adev))
-		WREG32_SOC15_NO_KIQ(GC, 0, mmRLC_SPM_MC_CNTL, data);
-	else
-		WREG32_SOC15(GC, 0, mmRLC_SPM_MC_CNTL, data);
+	if (pre_data != data) {
+		if (amdgpu_sriov_is_pp_one_vf(adev) && !amdgpu_sriov_runtime(adev)) {
+			WREG32_SOC15_NO_KIQ(GC, 0, mmRLC_SPM_MC_CNTL, data);
+		} else
+			WREG32_SOC15(GC, 0, mmRLC_SPM_MC_CNTL, data);
+	}
+}
+
+static void gfx_v10_0_update_spm_vmid(struct amdgpu_device *adev, struct amdgpu_ring *ring, unsigned int vmid)
+{
+	amdgpu_gfx_off_ctrl(adev, false);
+
+	gfx_v10_0_update_spm_vmid_internal(adev, vmid);
 
 	amdgpu_gfx_off_ctrl(adev, true);
 }
@@ -8330,9 +8376,11 @@ static void gfx_v10_cntl_power_gating(struct amdgpu_device *adev, bool enable)
 	 * Power/performance team will optimize it and might give a new value later.
 	 */
 	if (enable && (adev->pg_flags & AMD_PG_SUPPORT_GFX_PG)) {
-		switch (adev->ip_versions[GC_HWIP][0]) {
+		switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 		case IP_VERSION(10, 3, 1):
 		case IP_VERSION(10, 3, 3):
+		case IP_VERSION(10, 3, 6):
+		case IP_VERSION(10, 3, 7):
 			data = 0x4E20 & RLC_PG_DELAY_3__CGCG_ACTIVE_BEFORE_CGPG_MASK_Vangogh;
 			WREG32_SOC15(GC, 0, mmRLC_PG_DELAY_3, data);
 			break;
@@ -8344,11 +8392,11 @@ static void gfx_v10_cntl_power_gating(struct amdgpu_device *adev, bool enable)
 
 static void gfx_v10_cntl_pg(struct amdgpu_device *adev, bool enable)
 {
-	amdgpu_gfx_rlc_enter_safe_mode(adev);
+	amdgpu_gfx_rlc_enter_safe_mode(adev, 0);
 
 	gfx_v10_cntl_power_gating(adev, enable);
 
-	amdgpu_gfx_rlc_exit_safe_mode(adev);
+	amdgpu_gfx_rlc_exit_safe_mode(adev, 0);
 }
 
 static const struct amdgpu_rlc_funcs gfx_v10_0_rlc_funcs = {
@@ -8377,21 +8425,19 @@ static const struct amdgpu_rlc_funcs gfx_v10_0_rlc_funcs_sriov = {
 	.reset = gfx_v10_0_rlc_reset,
 	.start = gfx_v10_0_rlc_start,
 	.update_spm_vmid = gfx_v10_0_update_spm_vmid,
-	.sriov_wreg = gfx_v10_sriov_wreg,
-	.sriov_rreg = gfx_v10_sriov_rreg,
 	.is_rlcg_access_range = gfx_v10_0_is_rlcg_access_range,
 };
 
-static int gfx_v10_0_set_powergating_state(void *handle,
+static int gfx_v10_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_powergating_state state)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	bool enable = (state == AMD_PG_STATE_GATE);
 
 	if (amdgpu_sriov_vf(adev))
 		return 0;
 
-	switch (adev->ip_versions[GC_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 	case IP_VERSION(10, 1, 10):
 	case IP_VERSION(10, 1, 1):
 	case IP_VERSION(10, 1, 2):
@@ -8403,8 +8449,16 @@ static int gfx_v10_0_set_powergating_state(void *handle,
 		break;
 	case IP_VERSION(10, 3, 1):
 	case IP_VERSION(10, 3, 3):
+	case IP_VERSION(10, 3, 6):
+	case IP_VERSION(10, 3, 7):
+		if (!enable)
+			amdgpu_gfx_off_ctrl(adev, false);
+
 		gfx_v10_cntl_pg(adev, enable);
-		amdgpu_gfx_off_ctrl(adev, enable);
+
+		if (enable)
+			amdgpu_gfx_off_ctrl(adev, true);
+
 		break;
 	default:
 		break;
@@ -8412,15 +8466,15 @@ static int gfx_v10_0_set_powergating_state(void *handle,
 	return 0;
 }
 
-static int gfx_v10_0_set_clockgating_state(void *handle,
+static int gfx_v10_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_clockgating_state state)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (amdgpu_sriov_vf(adev))
 		return 0;
 
-	switch (adev->ip_versions[GC_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 	case IP_VERSION(10, 1, 10):
 	case IP_VERSION(10, 1, 1):
 	case IP_VERSION(10, 1, 2):
@@ -8429,7 +8483,9 @@ static int gfx_v10_0_set_clockgating_state(void *handle,
 	case IP_VERSION(10, 3, 1):
 	case IP_VERSION(10, 3, 4):
 	case IP_VERSION(10, 3, 5):
+	case IP_VERSION(10, 3, 6):
 	case IP_VERSION(10, 3, 3):
+	case IP_VERSION(10, 3, 7):
 		gfx_v10_0_update_gfx_clock_gating(adev,
 						 state == AMD_CG_STATE_GATE);
 		break;
@@ -8439,9 +8495,9 @@ static int gfx_v10_0_set_clockgating_state(void *handle,
 	return 0;
 }
 
-static void gfx_v10_0_get_clockgating_state(void *handle, u32 *flags)
+static void gfx_v10_0_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int data;
 
 	/* AMD_CG_SUPPORT_GFX_FGCG */
@@ -8485,7 +8541,8 @@ static void gfx_v10_0_get_clockgating_state(void *handle, u32 *flags)
 
 static u64 gfx_v10_0_ring_get_rptr_gfx(struct amdgpu_ring *ring)
 {
-	return ring->adev->wb.wb[ring->rptr_offs]; /* gfx10 is 32bit rptr*/
+	/* gfx10 is 32bit rptr*/
+	return *(uint32_t *)ring->rptr_cpu_addr;
 }
 
 static u64 gfx_v10_0_ring_get_wptr_gfx(struct amdgpu_ring *ring)
@@ -8495,7 +8552,7 @@ static u64 gfx_v10_0_ring_get_wptr_gfx(struct amdgpu_ring *ring)
 
 	/* XXX check if swapping is necessary on BE */
 	if (ring->use_doorbell) {
-		wptr = atomic64_read((atomic64_t *)&adev->wb.wb[ring->wptr_offs]);
+		wptr = atomic64_read((atomic64_t *)ring->wptr_cpu_addr);
 	} else {
 		wptr = RREG32_SOC15(GC, 0, mmCP_RB0_WPTR);
 		wptr += (u64)RREG32_SOC15(GC, 0, mmCP_RB0_WPTR_HI) << 32;
@@ -8510,17 +8567,21 @@ static void gfx_v10_0_ring_set_wptr_gfx(struct amdgpu_ring *ring)
 
 	if (ring->use_doorbell) {
 		/* XXX check if swapping is necessary on BE */
-		atomic64_set((atomic64_t *)&adev->wb.wb[ring->wptr_offs], ring->wptr);
+		atomic64_set((atomic64_t *)ring->wptr_cpu_addr,
+			     ring->wptr);
 		WDOORBELL64(ring->doorbell_index, ring->wptr);
 	} else {
-		WREG32_SOC15(GC, 0, mmCP_RB0_WPTR, lower_32_bits(ring->wptr));
-		WREG32_SOC15(GC, 0, mmCP_RB0_WPTR_HI, upper_32_bits(ring->wptr));
+		WREG32_SOC15(GC, 0, mmCP_RB0_WPTR,
+			     lower_32_bits(ring->wptr));
+		WREG32_SOC15(GC, 0, mmCP_RB0_WPTR_HI,
+			     upper_32_bits(ring->wptr));
 	}
 }
 
 static u64 gfx_v10_0_ring_get_rptr_compute(struct amdgpu_ring *ring)
 {
-	return ring->adev->wb.wb[ring->rptr_offs]; /* gfx10 hardware is 32bit rptr */
+	/* gfx10 hardware is 32bit rptr */
+	return *(uint32_t *)ring->rptr_cpu_addr;
 }
 
 static u64 gfx_v10_0_ring_get_wptr_compute(struct amdgpu_ring *ring)
@@ -8529,7 +8590,7 @@ static u64 gfx_v10_0_ring_get_wptr_compute(struct amdgpu_ring *ring)
 
 	/* XXX check if swapping is necessary on BE */
 	if (ring->use_doorbell)
-		wptr = atomic64_read((atomic64_t *)&ring->adev->wb.wb[ring->wptr_offs]);
+		wptr = atomic64_read((atomic64_t *)ring->wptr_cpu_addr);
 	else
 		BUG();
 	return wptr;
@@ -8539,9 +8600,9 @@ static void gfx_v10_0_ring_set_wptr_compute(struct amdgpu_ring *ring)
 {
 	struct amdgpu_device *adev = ring->adev;
 
-	/* XXX check if swapping is necessary on BE */
 	if (ring->use_doorbell) {
-		atomic64_set((atomic64_t *)&adev->wb.wb[ring->wptr_offs], ring->wptr);
+		atomic64_set((atomic64_t *)ring->wptr_cpu_addr,
+			     ring->wptr);
 		WDOORBELL64(ring->doorbell_index, ring->wptr);
 	} else {
 		BUG(); /* only DOORBELL method supported on gfx10 now */
@@ -8567,7 +8628,7 @@ static void gfx_v10_0_ring_emit_hdp_flush(struct amdgpu_ring *ring)
 		}
 		reg_mem_engine = 0;
 	} else {
-		ref_and_mask = nbio_hf_reg->ref_and_mask_cp0;
+		ref_and_mask = nbio_hf_reg->ref_and_mask_cp0 << ring->pipe;
 		reg_mem_engine = 1; /* pfp */
 	}
 
@@ -8582,7 +8643,7 @@ static void gfx_v10_0_ring_emit_ib_gfx(struct amdgpu_ring *ring,
 				       struct amdgpu_ib *ib,
 				       uint32_t flags)
 {
-	unsigned vmid = AMDGPU_JOB_GET_VMID(job);
+	unsigned int vmid = AMDGPU_JOB_GET_VMID(job);
 	u32 header, control = 0;
 
 	if (ib->flags & AMDGPU_IB_FLAG_CE)
@@ -8592,7 +8653,7 @@ static void gfx_v10_0_ring_emit_ib_gfx(struct amdgpu_ring *ring,
 
 	control |= ib->length_dw | (vmid << 24);
 
-	if ((amdgpu_sriov_vf(ring->adev) || amdgpu_mcbp) && (ib->flags & AMDGPU_IB_FLAG_PREEMPT)) {
+	if (ring->adev->gfx.mcbp && (ib->flags & AMDGPU_IB_FLAG_PREEMPT)) {
 		control |= INDIRECT_BUFFER_PRE_ENB(1);
 
 		if (flags & AMDGPU_IB_PREEMPTED)
@@ -8619,7 +8680,7 @@ static void gfx_v10_0_ring_emit_ib_compute(struct amdgpu_ring *ring,
 					   struct amdgpu_ib *ib,
 					   uint32_t flags)
 {
-	unsigned vmid = AMDGPU_JOB_GET_VMID(job);
+	unsigned int vmid = AMDGPU_JOB_GET_VMID(job);
 	u32 control = INDIRECT_BUFFER_VALID | ib->length_dw | (vmid << 24);
 
 	/* Currently, there is a high possibility to get wave ID mismatch
@@ -8650,7 +8711,7 @@ static void gfx_v10_0_ring_emit_ib_compute(struct amdgpu_ring *ring,
 }
 
 static void gfx_v10_0_ring_emit_fence(struct amdgpu_ring *ring, u64 addr,
-				     u64 seq, unsigned flags)
+				     u64 seq, unsigned int flags)
 {
 	bool write64bit = flags & AMDGPU_FENCE_FLAG_64BIT;
 	bool int_sel = flags & AMDGPU_FENCE_FLAG_INT;
@@ -8692,8 +8753,20 @@ static void gfx_v10_0_ring_emit_pipeline_sync(struct amdgpu_ring *ring)
 			       upper_32_bits(addr), seq, 0xffffffff, 4);
 }
 
+static void gfx_v10_0_ring_invalidate_tlbs(struct amdgpu_ring *ring,
+				   uint16_t pasid, uint32_t flush_type,
+				   bool all_hub, uint8_t dst_sel)
+{
+	amdgpu_ring_write(ring, PACKET3(PACKET3_INVALIDATE_TLBS, 0));
+	amdgpu_ring_write(ring,
+			  PACKET3_INVALIDATE_TLBS_DST_SEL(dst_sel) |
+			  PACKET3_INVALIDATE_TLBS_ALL_HUB(all_hub) |
+			  PACKET3_INVALIDATE_TLBS_PASID(pasid) |
+			  PACKET3_INVALIDATE_TLBS_FLUSH_TYPE(flush_type));
+}
+
 static void gfx_v10_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
-					 unsigned vmid, uint64_t pd_addr)
+					 unsigned int vmid, uint64_t pd_addr)
 {
 	amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pd_addr);
 
@@ -8743,7 +8816,7 @@ static void gfx_v10_0_ring_emit_cntxcntl(struct amdgpu_ring *ring,
 {
 	uint32_t dw2 = 0;
 
-	if (amdgpu_mcbp || amdgpu_sriov_vf(ring->adev))
+	if (ring->adev->gfx.mcbp)
 		gfx_v10_0_ring_emit_ce_meta(ring,
 				    (!amdgpu_sriov_vf(ring->adev) && flags & AMDGPU_IB_PREEMPTED) ? true : false);
 
@@ -8772,38 +8845,28 @@ static void gfx_v10_0_ring_emit_cntxcntl(struct amdgpu_ring *ring,
 	amdgpu_ring_write(ring, 0);
 }
 
-static unsigned gfx_v10_0_ring_emit_init_cond_exec(struct amdgpu_ring *ring)
+static unsigned int gfx_v10_0_ring_emit_init_cond_exec(struct amdgpu_ring *ring,
+						       uint64_t addr)
 {
-	unsigned ret;
+	unsigned int ret;
 
 	amdgpu_ring_write(ring, PACKET3(PACKET3_COND_EXEC, 3));
-	amdgpu_ring_write(ring, lower_32_bits(ring->cond_exe_gpu_addr));
-	amdgpu_ring_write(ring, upper_32_bits(ring->cond_exe_gpu_addr));
-	amdgpu_ring_write(ring, 0); /* discard following DWs if *cond_exec_gpu_addr==0 */
+	amdgpu_ring_write(ring, lower_32_bits(addr));
+	amdgpu_ring_write(ring, upper_32_bits(addr));
+	/* discard following DWs if *cond_exec_gpu_addr==0 */
+	amdgpu_ring_write(ring, 0);
 	ret = ring->wptr & ring->buf_mask;
-	amdgpu_ring_write(ring, 0x55aa55aa); /* patch dummy value later */
+	/* patch dummy value later */
+	amdgpu_ring_write(ring, 0);
 
 	return ret;
 }
 
-static void gfx_v10_0_ring_emit_patch_cond_exec(struct amdgpu_ring *ring, unsigned offset)
-{
-	unsigned cur;
-	BUG_ON(offset > ring->buf_mask);
-	BUG_ON(ring->ring[offset] != 0x55aa55aa);
-
-	cur = (ring->wptr - 1) & ring->buf_mask;
-	if (likely(cur > offset))
-		ring->ring[offset] = cur - offset;
-	else
-		ring->ring[offset] = (ring->buf_mask + 1) - offset + cur;
-}
-
 static int gfx_v10_0_ring_preempt_ib(struct amdgpu_ring *ring)
 {
 	int i, r = 0;
 	struct amdgpu_device *adev = ring->adev;
-	struct amdgpu_kiq *kiq = &adev->gfx.kiq;
+	struct amdgpu_kiq *kiq = &adev->gfx.kiq[0];
 	struct amdgpu_ring *kiq_ring = &kiq->ring;
 	unsigned long flags;
 
@@ -8850,26 +8913,26 @@ static void gfx_v10_0_ring_emit_ce_meta(struct amdgpu_ring *ring, bool resume)
 {
 	struct amdgpu_device *adev = ring->adev;
 	struct v10_ce_ib_state ce_payload = {0};
-	uint64_t csa_addr;
+	uint64_t offset, ce_payload_gpu_addr;
+	void *ce_payload_cpu_addr;
 	int cnt;
 
 	cnt = (sizeof(ce_payload) >> 2) + 4 - 2;
-	csa_addr = amdgpu_csa_vaddr(ring->adev);
+
+	offset = offsetof(struct v10_gfx_meta_data, ce_payload);
+	ce_payload_gpu_addr = amdgpu_csa_vaddr(ring->adev) + offset;
+	ce_payload_cpu_addr = adev->virt.csa_cpu_addr + offset;
 
 	amdgpu_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, cnt));
 	amdgpu_ring_write(ring, (WRITE_DATA_ENGINE_SEL(2) |
 				 WRITE_DATA_DST_SEL(8) |
 				 WR_CONFIRM) |
 				 WRITE_DATA_CACHE_POLICY(0));
-	amdgpu_ring_write(ring, lower_32_bits(csa_addr +
-			      offsetof(struct v10_gfx_meta_data, ce_payload)));
-	amdgpu_ring_write(ring, upper_32_bits(csa_addr +
-			      offsetof(struct v10_gfx_meta_data, ce_payload)));
+	amdgpu_ring_write(ring, lower_32_bits(ce_payload_gpu_addr));
+	amdgpu_ring_write(ring, upper_32_bits(ce_payload_gpu_addr));
 
 	if (resume)
-		amdgpu_ring_write_multiple(ring, adev->virt.csa_cpu_addr +
-					   offsetof(struct v10_gfx_meta_data,
-						    ce_payload),
+		amdgpu_ring_write_multiple(ring, ce_payload_cpu_addr,
 					   sizeof(ce_payload) >> 2);
 	else
 		amdgpu_ring_write_multiple(ring, (void *)&ce_payload,
@@ -8880,12 +8943,18 @@ static void gfx_v10_0_ring_emit_de_meta(struct amdgpu_ring *ring, bool resume)
 {
 	struct amdgpu_device *adev = ring->adev;
 	struct v10_de_ib_state de_payload = {0};
-	uint64_t csa_addr, gds_addr;
+	uint64_t offset, gds_addr, de_payload_gpu_addr;
+	void *de_payload_cpu_addr;
 	int cnt;
 
-	csa_addr = amdgpu_csa_vaddr(ring->adev);
-	gds_addr = ALIGN(csa_addr + AMDGPU_CSA_SIZE - adev->gds.gds_size,
+	offset = offsetof(struct v10_gfx_meta_data, de_payload);
+	de_payload_gpu_addr = amdgpu_csa_vaddr(ring->adev) + offset;
+	de_payload_cpu_addr = adev->virt.csa_cpu_addr + offset;
+
+	gds_addr = ALIGN(amdgpu_csa_vaddr(ring->adev) +
+			 AMDGPU_CSA_SIZE - adev->gds.gds_size,
 			 PAGE_SIZE);
+
 	de_payload.gds_backup_addrlo = lower_32_bits(gds_addr);
 	de_payload.gds_backup_addrhi = upper_32_bits(gds_addr);
 
@@ -8895,15 +8964,11 @@ static void gfx_v10_0_ring_emit_de_meta(struct amdgpu_ring *ring, bool resume)
 				 WRITE_DATA_DST_SEL(8) |
 				 WR_CONFIRM) |
 				 WRITE_DATA_CACHE_POLICY(0));
-	amdgpu_ring_write(ring, lower_32_bits(csa_addr +
-			      offsetof(struct v10_gfx_meta_data, de_payload)));
-	amdgpu_ring_write(ring, upper_32_bits(csa_addr +
-			      offsetof(struct v10_gfx_meta_data, de_payload)));
+	amdgpu_ring_write(ring, lower_32_bits(de_payload_gpu_addr));
+	amdgpu_ring_write(ring, upper_32_bits(de_payload_gpu_addr));
 
 	if (resume)
-		amdgpu_ring_write_multiple(ring, adev->virt.csa_cpu_addr +
-					   offsetof(struct v10_gfx_meta_data,
-						    de_payload),
+		amdgpu_ring_write_multiple(ring, de_payload_cpu_addr,
 					   sizeof(de_payload) >> 2);
 	else
 		amdgpu_ring_write_multiple(ring, (void *)&de_payload,
@@ -8983,19 +9048,6 @@ static void gfx_v10_0_ring_emit_reg_write_reg_wait(struct amdgpu_ring *ring,
 							   ref, mask);
 }
 
-static void gfx_v10_0_ring_soft_recovery(struct amdgpu_ring *ring,
-					 unsigned vmid)
-{
-	struct amdgpu_device *adev = ring->adev;
-	uint32_t value = 0;
-
-	value = REG_SET_FIELD(value, SQ_CMD, CMD, 0x03);
-	value = REG_SET_FIELD(value, SQ_CMD, MODE, 0x01);
-	value = REG_SET_FIELD(value, SQ_CMD, CHECK_VMID, 1);
-	value = REG_SET_FIELD(value, SQ_CMD, VM_ID, vmid);
-	WREG32_SOC15(GC, 0, mmSQ_CMD, value);
-}
-
 static void
 gfx_v10_0_set_gfx_eop_interrupt_state(struct amdgpu_device *adev,
 				      uint32_t me, uint32_t pipe,
@@ -9093,7 +9145,7 @@ static void gfx_v10_0_set_compute_eop_interrupt_state(struct amdgpu_device *adev
 
 static int gfx_v10_0_set_eop_interrupt_state(struct amdgpu_device *adev,
 					    struct amdgpu_irq_src *src,
-					    unsigned type,
+					    unsigned int type,
 					    enum amdgpu_interrupt_state state)
 {
 	switch (type) {
@@ -9142,6 +9194,7 @@ static int gfx_v10_0_eop_irq(struct amdgpu_device *adev,
 	struct amdgpu_ring *ring;
 
 	DRM_DEBUG("IH: CP EOP\n");
+
 	me_id = (entry->ring_id & 0x0c) >> 2;
 	pipe_id = (entry->ring_id & 0x03) >> 0;
 	queue_id = (entry->ring_id & 0x70) >> 4;
@@ -9158,27 +9211,58 @@ static int gfx_v10_0_eop_irq(struct amdgpu_device *adev,
 		for (i = 0; i < adev->gfx.num_compute_rings; i++) {
 			ring = &adev->gfx.compute_ring[i];
 			/* Per-queue interrupt is supported for MEC starting from VI.
-			  * The interrupt can only be enabled/disabled per pipe instead of per queue.
-			  */
-			if ((ring->me == me_id) && (ring->pipe == pipe_id) && (ring->queue == queue_id))
+			 * The interrupt can only be enabled/disabled per pipe instead
+			 * of per queue.
+			 */
+			if ((ring->me == me_id) &&
+			    (ring->pipe == pipe_id) &&
+			    (ring->queue == queue_id))
 				amdgpu_fence_process(ring);
 		}
 		break;
 	}
+
 	return 0;
 }
 
 static int gfx_v10_0_set_priv_reg_fault_state(struct amdgpu_device *adev,
 					      struct amdgpu_irq_src *source,
-					      unsigned type,
+					      unsigned int type,
 					      enum amdgpu_interrupt_state state)
 {
+	u32 cp_int_cntl_reg, cp_int_cntl;
+	int i, j;
+
 	switch (state) {
 	case AMDGPU_IRQ_STATE_DISABLE:
 	case AMDGPU_IRQ_STATE_ENABLE:
-		WREG32_FIELD15(GC, 0, CP_INT_CNTL_RING0,
-			       PRIV_REG_INT_ENABLE,
-			       state == AMDGPU_IRQ_STATE_ENABLE ? 1 : 0);
+		for (i = 0; i < adev->gfx.me.num_me; i++) {
+			for (j = 0; j < adev->gfx.me.num_pipe_per_me; j++) {
+				cp_int_cntl_reg = gfx_v10_0_get_cpg_int_cntl(adev, i, j);
+
+				if (cp_int_cntl_reg) {
+					cp_int_cntl = RREG32_SOC15_IP(GC, cp_int_cntl_reg);
+					cp_int_cntl = REG_SET_FIELD(cp_int_cntl, CP_INT_CNTL_RING0,
+								    PRIV_REG_INT_ENABLE,
+								    state == AMDGPU_IRQ_STATE_ENABLE ? 1 : 0);
+					WREG32_SOC15_IP(GC, cp_int_cntl_reg, cp_int_cntl);
+				}
+			}
+		}
+		for (i = 0; i < adev->gfx.mec.num_mec; i++) {
+			for (j = 0; j < adev->gfx.mec.num_pipe_per_mec; j++) {
+				/* MECs start at 1 */
+				cp_int_cntl_reg = gfx_v10_0_get_cpc_int_cntl(adev, i + 1, j);
+
+				if (cp_int_cntl_reg) {
+					cp_int_cntl = RREG32_SOC15_IP(GC, cp_int_cntl_reg);
+					cp_int_cntl = REG_SET_FIELD(cp_int_cntl, CP_ME1_PIPE0_INT_CNTL,
+								    PRIV_REG_INT_ENABLE,
+								    state == AMDGPU_IRQ_STATE_ENABLE ? 1 : 0);
+					WREG32_SOC15_IP(GC, cp_int_cntl_reg, cp_int_cntl);
+				}
+			}
+		}
 		break;
 	default:
 		break;
@@ -9187,17 +9271,75 @@ static int gfx_v10_0_set_priv_reg_fault_state(struct amdgpu_device *adev,
 	return 0;
 }
 
+static int gfx_v10_0_set_bad_op_fault_state(struct amdgpu_device *adev,
+					    struct amdgpu_irq_src *source,
+					    unsigned type,
+					    enum amdgpu_interrupt_state state)
+{
+	u32 cp_int_cntl_reg, cp_int_cntl;
+	int i, j;
+
+	switch (state) {
+	case AMDGPU_IRQ_STATE_DISABLE:
+	case AMDGPU_IRQ_STATE_ENABLE:
+		for (i = 0; i < adev->gfx.me.num_me; i++) {
+			for (j = 0; j < adev->gfx.me.num_pipe_per_me; j++) {
+				cp_int_cntl_reg = gfx_v10_0_get_cpg_int_cntl(adev, i, j);
+
+				if (cp_int_cntl_reg) {
+					cp_int_cntl = RREG32_SOC15_IP(GC, cp_int_cntl_reg);
+					cp_int_cntl = REG_SET_FIELD(cp_int_cntl, CP_INT_CNTL_RING0,
+								    OPCODE_ERROR_INT_ENABLE,
+								    state == AMDGPU_IRQ_STATE_ENABLE ? 1 : 0);
+					WREG32_SOC15_IP(GC, cp_int_cntl_reg, cp_int_cntl);
+				}
+			}
+		}
+		for (i = 0; i < adev->gfx.mec.num_mec; i++) {
+			for (j = 0; j < adev->gfx.mec.num_pipe_per_mec; j++) {
+				/* MECs start at 1 */
+				cp_int_cntl_reg = gfx_v10_0_get_cpc_int_cntl(adev, i + 1, j);
+
+				if (cp_int_cntl_reg) {
+					cp_int_cntl = RREG32_SOC15_IP(GC, cp_int_cntl_reg);
+					cp_int_cntl = REG_SET_FIELD(cp_int_cntl, CP_ME1_PIPE0_INT_CNTL,
+								    OPCODE_ERROR_INT_ENABLE,
+								    state == AMDGPU_IRQ_STATE_ENABLE ? 1 : 0);
+					WREG32_SOC15_IP(GC, cp_int_cntl_reg, cp_int_cntl);
+				}
+			}
+		}
+		break;
+	default:
+		break;
+	}
+	return 0;
+}
+
 static int gfx_v10_0_set_priv_inst_fault_state(struct amdgpu_device *adev,
 					       struct amdgpu_irq_src *source,
-					       unsigned type,
+					       unsigned int type,
 					       enum amdgpu_interrupt_state state)
 {
+	u32 cp_int_cntl_reg, cp_int_cntl;
+	int i, j;
+
 	switch (state) {
 	case AMDGPU_IRQ_STATE_DISABLE:
 	case AMDGPU_IRQ_STATE_ENABLE:
-		WREG32_FIELD15(GC, 0, CP_INT_CNTL_RING0,
-			       PRIV_INSTR_INT_ENABLE,
-			       state == AMDGPU_IRQ_STATE_ENABLE ? 1 : 0);
+		for (i = 0; i < adev->gfx.me.num_me; i++) {
+			for (j = 0; j < adev->gfx.me.num_pipe_per_me; j++) {
+				cp_int_cntl_reg = gfx_v10_0_get_cpg_int_cntl(adev, i, j);
+
+				if (cp_int_cntl_reg) {
+					cp_int_cntl = RREG32_SOC15_IP(GC, cp_int_cntl_reg);
+					cp_int_cntl = REG_SET_FIELD(cp_int_cntl, CP_INT_CNTL_RING0,
+								    PRIV_INSTR_INT_ENABLE,
+								    state == AMDGPU_IRQ_STATE_ENABLE ? 1 : 0);
+					WREG32_SOC15_IP(GC, cp_int_cntl_reg, cp_int_cntl);
+				}
+			}
+		}
 		break;
 	default:
 		break;
@@ -9221,8 +9363,8 @@ static void gfx_v10_0_handle_priv_fault(struct amdgpu_device *adev,
 	case 0:
 		for (i = 0; i < adev->gfx.num_gfx_rings; i++) {
 			ring = &adev->gfx.gfx_ring[i];
-			/* we only enabled 1 gfx queue per pipe for now */
-			if (ring->me == me_id && ring->pipe == pipe_id)
+			if (ring->me == me_id && ring->pipe == pipe_id &&
+			    ring->queue == queue_id)
 				drm_sched_fault(&ring->sched);
 		}
 		break;
@@ -9249,6 +9391,15 @@ static int gfx_v10_0_priv_reg_irq(struct amdgpu_device *adev,
 	return 0;
 }
 
+static int gfx_v10_0_bad_op_irq(struct amdgpu_device *adev,
+				struct amdgpu_irq_src *source,
+				struct amdgpu_iv_entry *entry)
+{
+	DRM_ERROR("Illegal opcode in command stream \n");
+	gfx_v10_0_handle_priv_fault(adev, entry);
+	return 0;
+}
+
 static int gfx_v10_0_priv_inst_irq(struct amdgpu_device *adev,
 				   struct amdgpu_irq_src *source,
 				   struct amdgpu_iv_entry *entry)
@@ -9264,7 +9415,7 @@ static int gfx_v10_0_kiq_set_interrupt_state(struct amdgpu_device *adev,
 					     enum amdgpu_interrupt_state state)
 {
 	uint32_t tmp, target;
-	struct amdgpu_ring *ring = &(adev->gfx.kiq.ring);
+	struct amdgpu_ring *ring = &(adev->gfx.kiq[0].ring);
 
 	if (ring->me == 1)
 		target = SOC15_REG_OFFSET(GC, 0, mmCP_ME1_PIPE0_INT_CNTL);
@@ -9308,7 +9459,7 @@ static int gfx_v10_0_kiq_irq(struct amdgpu_device *adev,
 			     struct amdgpu_iv_entry *entry)
 {
 	u8 me_id, pipe_id, queue_id;
-	struct amdgpu_ring *ring = &(adev->gfx.kiq.ring);
+	struct amdgpu_ring *ring = &(adev->gfx.kiq[0].ring);
 
 	me_id = (entry->ring_id & 0x0c) >> 2;
 	pipe_id = (entry->ring_id & 0x03) >> 0;
@@ -9343,6 +9494,319 @@ static void gfx_v10_0_emit_mem_sync(struct amdgpu_ring *ring)
 	amdgpu_ring_write(ring, gcr_cntl); /* GCR_CNTL */
 }
 
+static void gfx_v10_ring_insert_nop(struct amdgpu_ring *ring, uint32_t num_nop)
+{
+	/* Header itself is a NOP packet */
+	if (num_nop == 1) {
+		amdgpu_ring_write(ring, ring->funcs->nop);
+		return;
+	}
+
+	/* Max HW optimization till 0x3ffe, followed by remaining one NOP at a time*/
+	amdgpu_ring_write(ring, PACKET3(PACKET3_NOP, min(num_nop - 2, 0x3ffe)));
+
+	/* Header is at index 0, followed by num_nops - 1 NOP packet's */
+	amdgpu_ring_insert_nop(ring, num_nop - 1);
+}
+
+static int gfx_v10_0_reset_kgq(struct amdgpu_ring *ring,
+			       unsigned int vmid,
+			       struct amdgpu_fence *timedout_fence)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct amdgpu_kiq *kiq = &adev->gfx.kiq[0];
+	struct amdgpu_ring *kiq_ring = &kiq->ring;
+	unsigned long flags;
+	u32 tmp;
+	u64 addr;
+	int r;
+
+	if (!kiq->pmf || !kiq->pmf->kiq_unmap_queues)
+		return -EINVAL;
+
+	amdgpu_ring_reset_helper_begin(ring, timedout_fence);
+
+	spin_lock_irqsave(&kiq->ring_lock, flags);
+
+	if (amdgpu_ring_alloc(kiq_ring, 5 + 7 + 7)) {
+		spin_unlock_irqrestore(&kiq->ring_lock, flags);
+		return -ENOMEM;
+	}
+
+	addr = amdgpu_bo_gpu_offset(ring->mqd_obj) +
+		offsetof(struct v10_gfx_mqd, cp_gfx_hqd_active);
+	tmp = REG_SET_FIELD(0, CP_VMID_RESET, RESET_REQUEST, 1 << vmid);
+	if (ring->pipe == 0)
+		tmp = REG_SET_FIELD(tmp, CP_VMID_RESET, PIPE0_QUEUES, 1 << ring->queue);
+	else
+		tmp = REG_SET_FIELD(tmp, CP_VMID_RESET, PIPE1_QUEUES, 1 << ring->queue);
+
+	gfx_v10_0_ring_emit_wreg(kiq_ring,
+				 SOC15_REG_OFFSET(GC, 0, mmCP_VMID_RESET), tmp);
+	gfx_v10_0_wait_reg_mem(kiq_ring, 0, 1, 0,
+			       lower_32_bits(addr), upper_32_bits(addr),
+			       0, 1, 0x20);
+	gfx_v10_0_ring_emit_reg_wait(kiq_ring,
+				     SOC15_REG_OFFSET(GC, 0, mmCP_VMID_RESET), 0, 0xffffffff);
+	amdgpu_ring_commit(kiq_ring);
+	r = amdgpu_ring_test_ring(kiq_ring);
+	spin_unlock_irqrestore(&kiq->ring_lock, flags);
+	if (r)
+		return r;
+
+	r = gfx_v10_0_kgq_init_queue(ring, true);
+	if (r) {
+		DRM_ERROR("fail to init kgq\n");
+		return r;
+	}
+
+	spin_lock_irqsave(&kiq->ring_lock, flags);
+
+	if (amdgpu_ring_alloc(kiq_ring, kiq->pmf->map_queues_size)) {
+		spin_unlock_irqrestore(&kiq->ring_lock, flags);
+		return -ENOMEM;
+	}
+	kiq->pmf->kiq_map_queues(kiq_ring, ring);
+	amdgpu_ring_commit(kiq_ring);
+	r = amdgpu_ring_test_ring(kiq_ring);
+	spin_unlock_irqrestore(&kiq->ring_lock, flags);
+	if (r)
+		return r;
+
+	return amdgpu_ring_reset_helper_end(ring, timedout_fence);
+}
+
+static int gfx_v10_0_reset_kcq(struct amdgpu_ring *ring,
+			       unsigned int vmid,
+			       struct amdgpu_fence *timedout_fence)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct amdgpu_kiq *kiq = &adev->gfx.kiq[0];
+	struct amdgpu_ring *kiq_ring = &kiq->ring;
+	unsigned long flags;
+	int i, r;
+
+	if (!kiq->pmf || !kiq->pmf->kiq_unmap_queues)
+		return -EINVAL;
+
+	amdgpu_ring_reset_helper_begin(ring, timedout_fence);
+
+	spin_lock_irqsave(&kiq->ring_lock, flags);
+
+	if (amdgpu_ring_alloc(kiq_ring, kiq->pmf->unmap_queues_size)) {
+		spin_unlock_irqrestore(&kiq->ring_lock, flags);
+		return -ENOMEM;
+	}
+
+	kiq->pmf->kiq_unmap_queues(kiq_ring, ring, RESET_QUEUES,
+				   0, 0);
+	amdgpu_ring_commit(kiq_ring);
+	r = amdgpu_ring_test_ring(kiq_ring);
+	spin_unlock_irqrestore(&kiq->ring_lock, flags);
+	if (r)
+		return r;
+
+	/* make sure dequeue is complete*/
+	amdgpu_gfx_rlc_enter_safe_mode(adev, 0);
+	mutex_lock(&adev->srbm_mutex);
+	nv_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0);
+	for (i = 0; i < adev->usec_timeout; i++) {
+		if (!(RREG32_SOC15(GC, 0, mmCP_HQD_ACTIVE) & 1))
+			break;
+		udelay(1);
+	}
+	if (i >= adev->usec_timeout)
+		r = -ETIMEDOUT;
+	nv_grbm_select(adev, 0, 0, 0, 0);
+	mutex_unlock(&adev->srbm_mutex);
+	amdgpu_gfx_rlc_exit_safe_mode(adev, 0);
+	if (r) {
+		dev_err(adev->dev, "fail to wait on hqd deactivate\n");
+		return r;
+	}
+
+	r = gfx_v10_0_kcq_init_queue(ring, true);
+	if (r) {
+		dev_err(adev->dev, "fail to init kcq\n");
+		return r;
+	}
+
+	spin_lock_irqsave(&kiq->ring_lock, flags);
+	if (amdgpu_ring_alloc(kiq_ring, kiq->pmf->map_queues_size)) {
+		spin_unlock_irqrestore(&kiq->ring_lock, flags);
+		return -ENOMEM;
+	}
+	kiq->pmf->kiq_map_queues(kiq_ring, ring);
+	amdgpu_ring_commit(kiq_ring);
+	r = amdgpu_ring_test_ring(kiq_ring);
+	spin_unlock_irqrestore(&kiq->ring_lock, flags);
+	if (r)
+		return r;
+
+	return amdgpu_ring_reset_helper_end(ring, timedout_fence);
+}
+
+static void gfx_v10_ip_print(struct amdgpu_ip_block *ip_block, struct drm_printer *p)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	uint32_t i, j, k, reg, index = 0;
+	uint32_t reg_count = ARRAY_SIZE(gc_reg_list_10_1);
+
+	if (!adev->gfx.ip_dump_core)
+		return;
+
+	for (i = 0; i < reg_count; i++)
+		drm_printf(p, "%-50s \t 0x%08x\n",
+			   gc_reg_list_10_1[i].reg_name,
+			   adev->gfx.ip_dump_core[i]);
+
+	/* print compute queue registers for all instances */
+	if (!adev->gfx.ip_dump_compute_queues)
+		return;
+
+	reg_count = ARRAY_SIZE(gc_cp_reg_list_10);
+	drm_printf(p, "\nnum_mec: %d num_pipe: %d num_queue: %d\n",
+		   adev->gfx.mec.num_mec,
+		   adev->gfx.mec.num_pipe_per_mec,
+		   adev->gfx.mec.num_queue_per_pipe);
+
+	for (i = 0; i < adev->gfx.mec.num_mec; i++) {
+		for (j = 0; j < adev->gfx.mec.num_pipe_per_mec; j++) {
+			for (k = 0; k < adev->gfx.mec.num_queue_per_pipe; k++) {
+				drm_printf(p, "\nmec %d, pipe %d, queue %d\n", i, j, k);
+				for (reg = 0; reg < reg_count; reg++) {
+					if (i && gc_cp_reg_list_10[reg].reg_offset == mmCP_MEC_ME1_HEADER_DUMP)
+						drm_printf(p, "%-50s \t 0x%08x\n",
+							   "mmCP_MEC_ME2_HEADER_DUMP",
+							   adev->gfx.ip_dump_compute_queues[index + reg]);
+					else
+						drm_printf(p, "%-50s \t 0x%08x\n",
+							   gc_cp_reg_list_10[reg].reg_name,
+							   adev->gfx.ip_dump_compute_queues[index + reg]);
+				}
+				index += reg_count;
+			}
+		}
+	}
+
+	/* print gfx queue registers for all instances */
+	if (!adev->gfx.ip_dump_gfx_queues)
+		return;
+
+	index = 0;
+	reg_count = ARRAY_SIZE(gc_gfx_queue_reg_list_10);
+	drm_printf(p, "\nnum_me: %d num_pipe: %d num_queue: %d\n",
+		   adev->gfx.me.num_me,
+		   adev->gfx.me.num_pipe_per_me,
+		   adev->gfx.me.num_queue_per_pipe);
+
+	for (i = 0; i < adev->gfx.me.num_me; i++) {
+		for (j = 0; j < adev->gfx.me.num_pipe_per_me; j++) {
+			for (k = 0; k < adev->gfx.me.num_queue_per_pipe; k++) {
+				drm_printf(p, "\nme %d, pipe %d, queue %d\n", i, j, k);
+				for (reg = 0; reg < reg_count; reg++) {
+					drm_printf(p, "%-50s \t 0x%08x\n",
+						   gc_gfx_queue_reg_list_10[reg].reg_name,
+						   adev->gfx.ip_dump_gfx_queues[index + reg]);
+				}
+				index += reg_count;
+			}
+		}
+	}
+}
+
+static void gfx_v10_ip_dump(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	uint32_t i, j, k, reg, index = 0;
+	uint32_t reg_count = ARRAY_SIZE(gc_reg_list_10_1);
+
+	if (!adev->gfx.ip_dump_core)
+		return;
+
+	amdgpu_gfx_off_ctrl(adev, false);
+	for (i = 0; i < reg_count; i++)
+		adev->gfx.ip_dump_core[i] = RREG32(SOC15_REG_ENTRY_OFFSET(gc_reg_list_10_1[i]));
+	amdgpu_gfx_off_ctrl(adev, true);
+
+	/* dump compute queue registers for all instances */
+	if (!adev->gfx.ip_dump_compute_queues)
+		return;
+
+	reg_count = ARRAY_SIZE(gc_cp_reg_list_10);
+	amdgpu_gfx_off_ctrl(adev, false);
+	mutex_lock(&adev->srbm_mutex);
+	for (i = 0; i < adev->gfx.mec.num_mec; i++) {
+		for (j = 0; j < adev->gfx.mec.num_pipe_per_mec; j++) {
+			for (k = 0; k < adev->gfx.mec.num_queue_per_pipe; k++) {
+				/* ME0 is for GFX so start from 1 for CP */
+				nv_grbm_select(adev, adev->gfx.me.num_me + i, j, k, 0);
+
+				for (reg = 0; reg < reg_count; reg++) {
+					if (i && gc_cp_reg_list_10[reg].reg_offset == mmCP_MEC_ME1_HEADER_DUMP)
+						adev->gfx.ip_dump_compute_queues[index + reg] =
+							RREG32(SOC15_REG_OFFSET(GC, 0, mmCP_MEC_ME2_HEADER_DUMP));
+					else
+						adev->gfx.ip_dump_compute_queues[index + reg] =
+							RREG32(SOC15_REG_ENTRY_OFFSET(
+								       gc_cp_reg_list_10[reg]));
+				}
+				index += reg_count;
+			}
+		}
+	}
+	nv_grbm_select(adev, 0, 0, 0, 0);
+	mutex_unlock(&adev->srbm_mutex);
+	amdgpu_gfx_off_ctrl(adev, true);
+
+	/* dump gfx queue registers for all instances */
+	if (!adev->gfx.ip_dump_gfx_queues)
+		return;
+
+	index = 0;
+	reg_count = ARRAY_SIZE(gc_gfx_queue_reg_list_10);
+	amdgpu_gfx_off_ctrl(adev, false);
+	mutex_lock(&adev->srbm_mutex);
+	for (i = 0; i < adev->gfx.me.num_me; i++) {
+		for (j = 0; j < adev->gfx.me.num_pipe_per_me; j++) {
+			for (k = 0; k < adev->gfx.me.num_queue_per_pipe; k++) {
+				nv_grbm_select(adev, i, j, k, 0);
+
+				for (reg = 0; reg < reg_count; reg++) {
+					adev->gfx.ip_dump_gfx_queues[index + reg] =
+						RREG32(SOC15_REG_ENTRY_OFFSET(
+							gc_gfx_queue_reg_list_10[reg]));
+				}
+				index += reg_count;
+			}
+		}
+	}
+	nv_grbm_select(adev, 0, 0, 0, 0);
+	mutex_unlock(&adev->srbm_mutex);
+	amdgpu_gfx_off_ctrl(adev, true);
+}
+
+static void gfx_v10_0_ring_emit_cleaner_shader(struct amdgpu_ring *ring)
+{
+	/* Emit the cleaner shader */
+	amdgpu_ring_write(ring, PACKET3(PACKET3_RUN_CLEANER_SHADER, 0));
+	amdgpu_ring_write(ring, 0);  /* RESERVED field, programmed to zero */
+}
+
+static void gfx_v10_0_ring_begin_use(struct amdgpu_ring *ring)
+{
+	amdgpu_gfx_profile_ring_begin_use(ring);
+
+	amdgpu_gfx_enforce_isolation_ring_begin_use(ring);
+}
+
+static void gfx_v10_0_ring_end_use(struct amdgpu_ring *ring)
+{
+	amdgpu_gfx_profile_ring_end_use(ring);
+
+	amdgpu_gfx_enforce_isolation_ring_end_use(ring);
+}
+
 static const struct amd_ip_funcs gfx_v10_0_ip_funcs = {
 	.name = "gfx_v10_0",
 	.early_init = gfx_v10_0_early_init,
@@ -9359,6 +9823,8 @@ static const struct amd_ip_funcs gfx_v10_0_ip_funcs = {
 	.set_clockgating_state = gfx_v10_0_set_clockgating_state,
 	.set_powergating_state = gfx_v10_0_set_powergating_state,
 	.get_clockgating_state = gfx_v10_0_get_clockgating_state,
+	.dump_ip_state = gfx_v10_ip_dump,
+	.print_ip_state = gfx_v10_ip_print,
 };
 
 static const struct amdgpu_ring_funcs gfx_v10_0_ring_funcs_gfx = {
@@ -9366,7 +9832,7 @@ static const struct amdgpu_ring_funcs gfx_v10_0_ring_funcs_gfx = {
 	.align_mask = 0xff,
 	.nop = PACKET3(PACKET3_NOP, 0x3FFF),
 	.support_64bit_ptrs = true,
-	.vmhub = AMDGPU_GFXHUB_0,
+	.secure_submission_supported = true,
 	.get_rptr = gfx_v10_0_ring_get_rptr_gfx,
 	.get_wptr = gfx_v10_0_ring_get_wptr_gfx,
 	.set_wptr = gfx_v10_0_ring_set_wptr_gfx,
@@ -9375,7 +9841,7 @@ static const struct amdgpu_ring_funcs gfx_v10_0_ring_funcs_gfx = {
 		7 + /* PIPELINE_SYNC */
 		SOC15_FLUSH_GPU_TLB_NUM_WREG * 5 +
 		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 7 +
-		2 + /* VM_FLUSH */
+		4 + /* VM_FLUSH */
 		8 + /* FENCE for VM_FLUSH */
 		20 + /* GDS switch */
 		4 + /* double SWITCH_BUFFER,
@@ -9391,7 +9857,8 @@ static const struct amdgpu_ring_funcs gfx_v10_0_ring_funcs_gfx = {
 		5 + /* HDP_INVL */
 		8 + 8 + /* FENCE x2 */
 		2 + /* SWITCH_BUFFER */
-		8, /* gfx_v10_0_emit_mem_sync */
+		8 + /* gfx_v10_0_emit_mem_sync */
+		2, /* gfx_v10_0_ring_emit_cleaner_shader */
 	.emit_ib_size =	4, /* gfx_v10_0_ring_emit_ib_gfx */
 	.emit_ib = gfx_v10_0_ring_emit_ib_gfx,
 	.emit_fence = gfx_v10_0_ring_emit_fence,
@@ -9401,19 +9868,21 @@ static const struct amdgpu_ring_funcs gfx_v10_0_ring_funcs_gfx = {
 	.emit_hdp_flush = gfx_v10_0_ring_emit_hdp_flush,
 	.test_ring = gfx_v10_0_ring_test_ring,
 	.test_ib = gfx_v10_0_ring_test_ib,
-	.insert_nop = amdgpu_ring_insert_nop,
+	.insert_nop = gfx_v10_ring_insert_nop,
 	.pad_ib = amdgpu_ring_generic_pad_ib,
 	.emit_switch_buffer = gfx_v10_0_ring_emit_sb,
 	.emit_cntxcntl = gfx_v10_0_ring_emit_cntxcntl,
 	.init_cond_exec = gfx_v10_0_ring_emit_init_cond_exec,
-	.patch_cond_exec = gfx_v10_0_ring_emit_patch_cond_exec,
 	.preempt_ib = gfx_v10_0_ring_preempt_ib,
 	.emit_frame_cntl = gfx_v10_0_ring_emit_frame_cntl,
 	.emit_wreg = gfx_v10_0_ring_emit_wreg,
 	.emit_reg_wait = gfx_v10_0_ring_emit_reg_wait,
 	.emit_reg_write_reg_wait = gfx_v10_0_ring_emit_reg_write_reg_wait,
-	.soft_recovery = gfx_v10_0_ring_soft_recovery,
 	.emit_mem_sync = gfx_v10_0_emit_mem_sync,
+	.reset = gfx_v10_0_reset_kgq,
+	.emit_cleaner_shader = gfx_v10_0_ring_emit_cleaner_shader,
+	.begin_use = gfx_v10_0_ring_begin_use,
+	.end_use = gfx_v10_0_ring_end_use,
 };
 
 static const struct amdgpu_ring_funcs gfx_v10_0_ring_funcs_compute = {
@@ -9421,7 +9890,6 @@ static const struct amdgpu_ring_funcs gfx_v10_0_ring_funcs_compute = {
 	.align_mask = 0xff,
 	.nop = PACKET3(PACKET3_NOP, 0x3FFF),
 	.support_64bit_ptrs = true,
-	.vmhub = AMDGPU_GFXHUB_0,
 	.get_rptr = gfx_v10_0_ring_get_rptr_compute,
 	.get_wptr = gfx_v10_0_ring_get_wptr_compute,
 	.set_wptr = gfx_v10_0_ring_set_wptr_compute,
@@ -9434,7 +9902,8 @@ static const struct amdgpu_ring_funcs gfx_v10_0_ring_funcs_compute = {
 		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 7 +
 		2 + /* gfx_v10_0_ring_emit_vm_flush */
 		8 + 8 + 8 + /* gfx_v10_0_ring_emit_fence x3 for user fence, vm fence */
-		8, /* gfx_v10_0_emit_mem_sync */
+		8 + /* gfx_v10_0_emit_mem_sync */
+		2, /* gfx_v10_0_ring_emit_cleaner_shader */
 	.emit_ib_size =	7, /* gfx_v10_0_ring_emit_ib_compute */
 	.emit_ib = gfx_v10_0_ring_emit_ib_compute,
 	.emit_fence = gfx_v10_0_ring_emit_fence,
@@ -9444,12 +9913,16 @@ static const struct amdgpu_ring_funcs gfx_v10_0_ring_funcs_compute = {
 	.emit_hdp_flush = gfx_v10_0_ring_emit_hdp_flush,
 	.test_ring = gfx_v10_0_ring_test_ring,
 	.test_ib = gfx_v10_0_ring_test_ib,
-	.insert_nop = amdgpu_ring_insert_nop,
+	.insert_nop = gfx_v10_ring_insert_nop,
 	.pad_ib = amdgpu_ring_generic_pad_ib,
 	.emit_wreg = gfx_v10_0_ring_emit_wreg,
 	.emit_reg_wait = gfx_v10_0_ring_emit_reg_wait,
 	.emit_reg_write_reg_wait = gfx_v10_0_ring_emit_reg_write_reg_wait,
 	.emit_mem_sync = gfx_v10_0_emit_mem_sync,
+	.reset = gfx_v10_0_reset_kcq,
+	.emit_cleaner_shader = gfx_v10_0_ring_emit_cleaner_shader,
+	.begin_use = gfx_v10_0_ring_begin_use,
+	.end_use = gfx_v10_0_ring_end_use,
 };
 
 static const struct amdgpu_ring_funcs gfx_v10_0_ring_funcs_kiq = {
@@ -9457,7 +9930,6 @@ static const struct amdgpu_ring_funcs gfx_v10_0_ring_funcs_kiq = {
 	.align_mask = 0xff,
 	.nop = PACKET3(PACKET3_NOP, 0x3FFF),
 	.support_64bit_ptrs = true,
-	.vmhub = AMDGPU_GFXHUB_0,
 	.get_rptr = gfx_v10_0_ring_get_rptr_compute,
 	.get_wptr = gfx_v10_0_ring_get_wptr_compute,
 	.set_wptr = gfx_v10_0_ring_set_wptr_compute,
@@ -9468,7 +9940,6 @@ static const struct amdgpu_ring_funcs gfx_v10_0_ring_funcs_kiq = {
 		7 + /* gfx_v10_0_ring_emit_pipeline_sync */
 		SOC15_FLUSH_GPU_TLB_NUM_WREG * 5 +
 		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 7 +
-		2 + /* gfx_v10_0_ring_emit_vm_flush */
 		8 + 8 + 8, /* gfx_v10_0_ring_emit_fence_kiq x3 for user fence, vm fence */
 	.emit_ib_size =	7, /* gfx_v10_0_ring_emit_ib_compute */
 	.emit_ib = gfx_v10_0_ring_emit_ib_compute,
@@ -9481,13 +9952,14 @@ static const struct amdgpu_ring_funcs gfx_v10_0_ring_funcs_kiq = {
 	.emit_wreg = gfx_v10_0_ring_emit_wreg,
 	.emit_reg_wait = gfx_v10_0_ring_emit_reg_wait,
 	.emit_reg_write_reg_wait = gfx_v10_0_ring_emit_reg_write_reg_wait,
+	.emit_hdp_flush = gfx_v10_0_ring_emit_hdp_flush,
 };
 
 static void gfx_v10_0_set_ring_funcs(struct amdgpu_device *adev)
 {
 	int i;
 
-	adev->gfx.kiq.ring.funcs = &gfx_v10_0_ring_funcs_kiq;
+	adev->gfx.kiq[0].ring.funcs = &gfx_v10_0_ring_funcs_kiq;
 
 	for (i = 0; i < adev->gfx.num_gfx_rings; i++)
 		adev->gfx.gfx_ring[i].funcs = &gfx_v10_0_ring_funcs_gfx;
@@ -9506,6 +9978,11 @@ static const struct amdgpu_irq_src_funcs gfx_v10_0_priv_reg_irq_funcs = {
 	.process = gfx_v10_0_priv_reg_irq,
 };
 
+static const struct amdgpu_irq_src_funcs gfx_v10_0_bad_op_irq_funcs = {
+	.set = gfx_v10_0_set_bad_op_fault_state,
+	.process = gfx_v10_0_bad_op_irq,
+};
+
 static const struct amdgpu_irq_src_funcs gfx_v10_0_priv_inst_irq_funcs = {
 	.set = gfx_v10_0_set_priv_inst_fault_state,
 	.process = gfx_v10_0_priv_inst_irq,
@@ -9521,27 +9998,33 @@ static void gfx_v10_0_set_irq_funcs(struct amdgpu_device *adev)
 	adev->gfx.eop_irq.num_types = AMDGPU_CP_IRQ_LAST;
 	adev->gfx.eop_irq.funcs = &gfx_v10_0_eop_irq_funcs;
 
-	adev->gfx.kiq.irq.num_types = AMDGPU_CP_KIQ_IRQ_LAST;
-	adev->gfx.kiq.irq.funcs = &gfx_v10_0_kiq_irq_funcs;
+	adev->gfx.kiq[0].irq.num_types = AMDGPU_CP_KIQ_IRQ_LAST;
+	adev->gfx.kiq[0].irq.funcs = &gfx_v10_0_kiq_irq_funcs;
 
 	adev->gfx.priv_reg_irq.num_types = 1;
 	adev->gfx.priv_reg_irq.funcs = &gfx_v10_0_priv_reg_irq_funcs;
 
+	adev->gfx.bad_op_irq.num_types = 1;
+	adev->gfx.bad_op_irq.funcs = &gfx_v10_0_bad_op_irq_funcs;
+
 	adev->gfx.priv_inst_irq.num_types = 1;
 	adev->gfx.priv_inst_irq.funcs = &gfx_v10_0_priv_inst_irq_funcs;
 }
 
 static void gfx_v10_0_set_rlc_funcs(struct amdgpu_device *adev)
 {
-	switch (adev->ip_versions[GC_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 	case IP_VERSION(10, 1, 10):
 	case IP_VERSION(10, 1, 1):
 	case IP_VERSION(10, 1, 3):
+	case IP_VERSION(10, 1, 4):
 	case IP_VERSION(10, 3, 2):
 	case IP_VERSION(10, 3, 1):
 	case IP_VERSION(10, 3, 4):
 	case IP_VERSION(10, 3, 5):
+	case IP_VERSION(10, 3, 6):
 	case IP_VERSION(10, 3, 3):
+	case IP_VERSION(10, 3, 7):
 		adev->gfx.rlc.funcs = &gfx_v10_0_rlc_funcs;
 		break;
 	case IP_VERSION(10, 1, 2):
@@ -9555,7 +10038,7 @@ static void gfx_v10_0_set_rlc_funcs(struct amdgpu_device *adev)
 
 static void gfx_v10_0_set_gds_init(struct amdgpu_device *adev)
 {
-	unsigned total_cu = adev->gfx.config.max_cu_per_sh *
+	unsigned int total_cu = adev->gfx.config.max_cu_per_sh *
 			    adev->gfx.config.max_sh_per_se *
 			    adev->gfx.config.max_shader_engines;
 
@@ -9565,6 +10048,20 @@ static void gfx_v10_0_set_gds_init(struct amdgpu_device *adev)
 	adev->gds.oa_size = 16;
 }
 
+static void gfx_v10_0_set_mqd_funcs(struct amdgpu_device *adev)
+{
+	/* set gfx eng mqd */
+	adev->mqds[AMDGPU_HW_IP_GFX].mqd_size =
+		sizeof(struct v10_gfx_mqd);
+	adev->mqds[AMDGPU_HW_IP_GFX].init_mqd =
+		gfx_v10_0_gfx_mqd_init;
+	/* set compute eng mqd */
+	adev->mqds[AMDGPU_HW_IP_COMPUTE].mqd_size =
+		sizeof(struct v10_compute_mqd);
+	adev->mqds[AMDGPU_HW_IP_COMPUTE].init_mqd =
+		gfx_v10_0_compute_mqd_init;
+}
+
 static void gfx_v10_0_set_user_wgp_inactive_bitmap_per_sh(struct amdgpu_device *adev,
 							  u32 bitmap)
 {
@@ -9622,7 +10119,7 @@ static int gfx_v10_0_get_cu_info(struct amdgpu_device *adev,
 {
 	int i, j, k, counter, active_cu_number = 0;
 	u32 mask, bitmap, ao_bitmap, ao_cu_mask = 0;
-	unsigned disable_masks[4 * 2];
+	unsigned int disable_masks[4 * 2];
 
 	if (!adev || !cu_info)
 		return -EINVAL;
@@ -9633,19 +10130,25 @@ static int gfx_v10_0_get_cu_info(struct amdgpu_device *adev,
 	for (i = 0; i < adev->gfx.config.max_shader_engines; i++) {
 		for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) {
 			bitmap = i * adev->gfx.config.max_sh_per_se + j;
-			if (((adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 3, 0)) ||
-				(adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 3, 3))) &&
+			if (((amdgpu_ip_version(adev, GC_HWIP, 0) ==
+			      IP_VERSION(10, 3, 0)) ||
+			     (amdgpu_ip_version(adev, GC_HWIP, 0) ==
+			      IP_VERSION(10, 3, 3)) ||
+			     (amdgpu_ip_version(adev, GC_HWIP, 0) ==
+			      IP_VERSION(10, 3, 6)) ||
+			     (amdgpu_ip_version(adev, GC_HWIP, 0) ==
+			      IP_VERSION(10, 3, 7))) &&
 			    ((gfx_v10_3_get_disabled_sa(adev) >> bitmap) & 1))
 				continue;
 			mask = 1;
 			ao_bitmap = 0;
 			counter = 0;
-			gfx_v10_0_select_se_sh(adev, i, j, 0xffffffff);
+			gfx_v10_0_select_se_sh(adev, i, j, 0xffffffff, 0);
 			if (i < 4 && j < 2)
 				gfx_v10_0_set_user_wgp_inactive_bitmap_per_sh(
 					adev, disable_masks[i * 2 + j]);
 			bitmap = gfx_v10_0_get_cu_active_bitmap_per_sh(adev);
-			cu_info->bitmap[i][j] = bitmap;
+			cu_info->bitmap[0][i][j] = bitmap;
 
 			for (k = 0; k < adev->gfx.config.max_cu_per_sh; k++) {
 				if (bitmap & mask) {
@@ -9661,7 +10164,7 @@ static int gfx_v10_0_get_cu_info(struct amdgpu_device *adev,
 			cu_info->ao_cu_bitmap[i][j] = ao_bitmap;
 		}
 	}
-	gfx_v10_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+	gfx_v10_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
 	mutex_unlock(&adev->grbm_idx_mutex);
 
 	cu_info->number = active_cu_number;
@@ -9737,8 +10240,7 @@ static void gfx_v10_3_set_power_brake_sequence(struct amdgpu_device *adev)
 		     (0x1 << DIDT_SQ_THROTTLE_CTRL__PWRBRK_STALL_EN__SHIFT));
 }
 
-const struct amdgpu_ip_block_version gfx_v10_0_ip_block =
-{
+const struct amdgpu_ip_block_version gfx_v10_0_ip_block = {
 	.type = AMD_IP_BLOCK_TYPE_GFX,
 	.major = 10,
 	.minor = 0,
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0_cleaner_shader.h b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0_cleaner_shader.h
new file mode 100644
index 000000000000..f67569ccf9f6
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0_cleaner_shader.h
@@ -0,0 +1,91 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2025 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+/* Define the cleaner shader gfx_10_1_10 */
+static const u32 gfx_10_1_10_cleaner_shader_hex[] = {
+	0xb0804004, 0xbf8a0000,
+	0xbf068100, 0xbf840023,
+	0xbe8203b8, 0xbefc0380,
+	0x7e008480, 0x7e028480,
+	0x7e048480, 0x7e068480,
+	0x7e088480, 0x7e0a8480,
+	0x7e0c8480, 0x7e0e8480,
+	0xbefc0302, 0x80828802,
+	0xbf84fff5, 0xbe8203ff,
+	0x80000000, 0x87020102,
+	0xbf840012, 0xbefe03c1,
+	0xbeff03c1, 0xd7650001,
+	0x0001007f, 0xd7660001,
+	0x0002027e, 0x16020288,
+	0xbe8203bf, 0xbefc03c1,
+	0xd9382000, 0x00020201,
+	0xd9386040, 0x00040401,
+	0xd70f6a01, 0x000202ff,
+	0x00000400, 0x80828102,
+	0xbf84fff7, 0xbefc03ff,
+	0x00000068, 0xbe803000,
+	0xbe813000, 0xbe823000,
+	0xbe833000, 0x80fc847c,
+	0xbf84fffa, 0xbeea0480,
+	0xbeec0480, 0xbeee0480,
+	0xbef00480, 0xbef20480,
+	0xbef40480, 0xbef60480,
+	0xbef80480, 0xbefa0480,
+	0xbf810000, 0xbf9f0000,
+	0xbf9f0000, 0xbf9f0000,
+	0xbf9f0000, 0xbf9f0000,
+};
+
+/* Define the cleaner shader gfx_10_3_0 */
+static const u32 gfx_10_3_0_cleaner_shader_hex[] = {
+	0xb0804004, 0xbf8a0000,
+	0xbe8203b8, 0xbefc0380,
+	0x7e008480, 0x7e028480,
+	0x7e048480, 0x7e068480,
+	0x7e088480, 0x7e0a8480,
+	0x7e0c8480, 0x7e0e8480,
+	0xbefc0302, 0x80828802,
+	0xbf84fff5, 0xbe8203ff,
+	0x80000000, 0x87020002,
+	0xbf840012, 0xbefe03c1,
+	0xbeff03c1, 0xd7650001,
+	0x0001007f, 0xd7660001,
+	0x0002027e, 0x16020288,
+	0xbe8203bf, 0xbefc03c1,
+	0xd9382000, 0x00020201,
+	0xd9386040, 0x00040401,
+	0xd70f6a01, 0x000202ff,
+	0x00000400, 0x80828102,
+	0xbf84fff7, 0xbefc03ff,
+	0x00000068, 0xbe803080,
+	0xbe813080, 0xbe823080,
+	0xbe833080, 0x80fc847c,
+	0xbf84fffa, 0xbeea0480,
+	0xbeec0480, 0xbeee0480,
+	0xbef00480, 0xbef20480,
+	0xbef40480, 0xbef60480,
+	0xbef80480, 0xbefa0480,
+	0xbf810000, 0xbf9f0000,
+	0xbf9f0000, 0xbf9f0000,
+	0xbf9f0000, 0xbf9f0000,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_1_10_cleaner_shader.asm b/drivers/gpu/drm/amd/amdgpu/gfx_v10_1_10_cleaner_shader.asm
new file mode 100644
index 000000000000..54f7ed9e2801
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_1_10_cleaner_shader.asm
@@ -0,0 +1,125 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2025 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+// This shader is to clean LDS, SGPRs and VGPRs. It is  first 64 Dwords or 256 bytes of 256 Dwords cleaner shader.
+
+// GFX10.1 : Clear SGPRs, VGPRs and LDS
+//   Launch 32 waves per CU (16 per SIMD) as a workgroup (threadgroup) to fill every wave slot
+//   Waves are "wave32" and have 64 VGPRs each, which uses all 1024 VGPRs per SIMD
+//   Waves are launched in "CU" mode, and the workgroup shares 64KB of LDS (half of the WGP's LDS)
+//      It takes 2 workgroups to use all of LDS: one on each CU of the WGP
+//   Each wave clears SGPRs 0 - 107
+//   Each wave clears VGPRs 0 - 63
+//   The first wave of the workgroup clears its 64KB of LDS
+//   The shader starts with "S_BARRIER" to ensure SPI has launched all waves of the workgroup
+//       before any wave in the workgroup could end.  Without this, it is possible not all SGPRs get cleared.
+
+
+shader main
+  asic(GFX10.1)
+  type(CS)
+  wave_size(32)
+// Note: original source code from SQ team
+//
+// Create 32 waves in a threadgroup (CS waves)
+// Each allocates 64 VGPRs
+// The workgroup allocates all of LDS (64kbytes)
+//
+// Takes about 2500 clocks to run.
+//   (theorhetical fastest = 1024clks vgpr + 640lds = 1660 clks)
+//
+  S_BARRIER
+  s_cmp_eq_u32 s0, 1                                // Bit0 is set, sgpr0 is set then clear VGPRS and LDS as FW set COMPUTE_USER_DATA_0
+  s_cbranch_scc0  label_0023                        // Clean VGPRs and LDS if sgpr0 of wave is set, scc = (s0 == 1)
+
+  s_mov_b32     s2, 0x00000038  // Loop 64/8=8 times  (loop unrolled for performance)
+  s_mov_b32     m0, 0
+  //
+  // CLEAR VGPRs
+  //
+label_0005:
+  v_movreld_b32     v0, 0
+  v_movreld_b32     v1, 0
+  v_movreld_b32     v2, 0
+  v_movreld_b32     v3, 0
+  v_movreld_b32     v4, 0
+  v_movreld_b32     v5, 0
+  v_movreld_b32     v6, 0
+  v_movreld_b32     v7, 0
+  s_mov_b32         m0, s2
+  s_sub_u32     s2, s2, 8
+  s_cbranch_scc0  label_0005
+  //
+  s_mov_b32     s2, 0x80000000                       // Bit31 is first_wave
+  s_and_b32     s2, s2, s1                           // sgpr0 has tg_size (first_wave) term as in ucode only COMPUTE_PGM_RSRC2.tg_size_en is set
+  s_cbranch_scc0  label_0023                         // Clean LDS if its first wave of ThreadGroup/WorkGroup
+  // CLEAR LDS
+  //
+  s_mov_b32 exec_lo, 0xffffffff
+  s_mov_b32 exec_hi, 0xffffffff
+  v_mbcnt_lo_u32_b32  v1, exec_hi, 0          // Set V1 to thread-ID (0..63)
+  v_mbcnt_hi_u32_b32  v1, exec_lo, v1        // Set V1 to thread-ID (0..63)
+  v_mul_u32_u24  v1, 0x00000008, v1          // * 8, so each thread is a double-dword address (8byte)
+  s_mov_b32     s2, 0x00000003f                    // 64 loop iterations
+  s_mov_b32     m0, 0xffffffff
+  // Clear all of LDS space
+  // Each FirstWave of WorkGroup clears 64kbyte block
+
+label_001F:
+  ds_write2_b64  v1, v[2:3], v[2:3] offset1:32
+  ds_write2_b64  v1, v[4:5], v[4:5] offset0:64 offset1:96
+  v_add_co_u32     v1, vcc, 0x00000400, v1
+  s_sub_u32     s2, s2, 1
+  s_cbranch_scc0  label_001F
+
+  //
+  // CLEAR SGPRs
+  //
+label_0023:
+  s_mov_b32     m0, 0x00000068  // Loop 108/4=27 times  (loop unrolled for performance)
+label_sgpr_loop:
+  s_movreld_b32     s0, s0
+  s_movreld_b32     s1, s0
+  s_movreld_b32     s2, s0
+  s_movreld_b32     s3, s0
+  s_sub_u32         m0, m0, 4
+  s_cbranch_scc0  label_sgpr_loop
+
+  //clear vcc
+  s_mov_b64 vcc, 0          //clear vcc
+  //s_setreg_imm32_b32 hw_reg_shader_flat_scratch_lo, 0   //clear  flat scratch lo SGPR
+  //s_setreg_imm32_b32 hw_reg_shader_flat_scratch_hi, 0    //clear  flat scratch hi SGPR
+  s_mov_b64 ttmp0, 0        //Clear ttmp0 and ttmp1
+  s_mov_b64 ttmp2, 0        //Clear ttmp2 and ttmp3
+  s_mov_b64 ttmp4, 0        //Clear ttmp4 and ttmp5
+  s_mov_b64 ttmp6, 0        //Clear ttmp6 and ttmp7
+  s_mov_b64 ttmp8, 0        //Clear ttmp8 and ttmp9
+  s_mov_b64 ttmp10, 0       //Clear ttmp10 and ttmp11
+  s_mov_b64 ttmp12, 0       //Clear ttmp12 and ttmp13
+  s_mov_b64 ttmp14, 0       //Clear ttmp14 and ttmp15
+
+ s_endpgm
+
+end
+
+
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_3_0_cleaner_shader.asm b/drivers/gpu/drm/amd/amdgpu/gfx_v10_3_0_cleaner_shader.asm
new file mode 100644
index 000000000000..0e1c246166c0
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_3_0_cleaner_shader.asm
@@ -0,0 +1,124 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2025 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+// This shader is to clean LDS, SGPRs and VGPRs. It is  first 64 Dwords or 256 bytes of 192 Dwords cleaner shader.
+//To turn this shader program on for complitaion change this to main and lower shader main to main_1
+
+// GFX10.3 : Clear SGPRs, VGPRs and LDS
+//   Launch 32 waves per CU (16 per SIMD) as a workgroup (threadgroup) to fill every wave slot
+//   Waves are "wave32" and have 64 VGPRs each, which uses all 1024 VGPRs per SIMD
+//   Waves are launched in "CU" mode, and the workgroup shares 64KB of LDS (half of the WGP's LDS)
+//      It takes 2 workgroups to use all of LDS: one on each CU of the WGP
+//   Each wave clears SGPRs 0 - 107
+//   Each wave clears VGPRs 0 - 63
+//   The first wave of the workgroup clears its 64KB of LDS
+//   The shader starts with "S_BARRIER" to ensure SPI has launched all waves of the workgroup
+//       before any wave in the workgroup could end.  Without this, it is possible not all SGPRs get cleared.
+
+
+shader main
+  asic(GFX10)
+  type(CS)
+  wave_size(32)
+// Note: original source code from SQ team
+
+//
+// Create 32 waves in a threadgroup (CS waves)
+// Each allocates 64 VGPRs
+// The workgroup allocates all of LDS (64kbytes)
+//
+// Takes about 2500 clocks to run.
+//   (theorhetical fastest = 1024clks vgpr + 640lds = 1660 clks)
+//
+  S_BARRIER
+  s_mov_b32     s2, 0x00000038  // Loop 64/8=8 times  (loop unrolled for performance)
+  s_mov_b32     m0, 0
+  //
+  // CLEAR VGPRs
+  //
+label_0005:
+  v_movreld_b32     v0, 0
+  v_movreld_b32     v1, 0
+  v_movreld_b32     v2, 0
+  v_movreld_b32     v3, 0
+  v_movreld_b32     v4, 0
+  v_movreld_b32     v5, 0
+  v_movreld_b32     v6, 0
+  v_movreld_b32     v7, 0
+  s_mov_b32         m0, s2
+  s_sub_u32     s2, s2, 8
+  s_cbranch_scc0  label_0005
+  //
+  s_mov_b32     s2, 0x80000000                     // Bit31 is first_wave
+  s_and_b32     s2, s2, s0                                  // sgpr0 has tg_size (first_wave) term as in ucode only COMPUTE_PGM_RSRC2.tg_size_en is set
+  s_cbranch_scc0  label_0023                         // Clean LDS if its first wave of ThreadGroup/WorkGroup
+  // CLEAR LDS
+  //
+  s_mov_b32 exec_lo, 0xffffffff
+  s_mov_b32 exec_hi, 0xffffffff
+  v_mbcnt_lo_u32_b32  v1, exec_hi, 0          // Set V1 to thread-ID (0..63)
+  v_mbcnt_hi_u32_b32  v1, exec_lo, v1        // Set V1 to thread-ID (0..63)
+  v_mul_u32_u24  v1, 0x00000008, v1          // * 8, so each thread is a double-dword address (8byte)
+  s_mov_b32     s2, 0x00000003f                    // 64 loop iterations
+  s_mov_b32     m0, 0xffffffff
+  // Clear all of LDS space
+  // Each FirstWave of WorkGroup clears 64kbyte block
+
+label_001F:
+  ds_write2_b64  v1, v[2:3], v[2:3] offset1:32
+  ds_write2_b64  v1, v[4:5], v[4:5] offset0:64 offset1:96
+  v_add_co_u32     v1, vcc, 0x00000400, v1
+  s_sub_u32     s2, s2, 1
+  s_cbranch_scc0  label_001F
+
+  //
+  // CLEAR SGPRs
+  //
+label_0023:
+  s_mov_b32     m0, 0x00000068  // Loop 108/4=27 times  (loop unrolled for performance)
+label_sgpr_loop:
+  s_movreld_b32     s0, 0
+  s_movreld_b32     s1, 0
+  s_movreld_b32     s2, 0
+  s_movreld_b32     s3, 0
+  s_sub_u32         m0, m0, 4
+  s_cbranch_scc0  label_sgpr_loop
+
+  //clear vcc
+  s_mov_b32 flat_scratch_lo, 0   //clear  flat scratch lo SGPR
+  s_mov_b32 flat_scratch_hi, 0   //clear  flat scratch hi SGPR
+  s_mov_b64 vcc, 0          //clear vcc
+  s_mov_b64 ttmp0, 0        //Clear ttmp0 and ttmp1
+  s_mov_b64 ttmp2, 0        //Clear ttmp2 and ttmp3
+  s_mov_b64 ttmp4, 0        //Clear ttmp4 and ttmp5
+  s_mov_b64 ttmp6, 0        //Clear ttmp6 and ttmp7
+  s_mov_b64 ttmp8, 0        //Clear ttmp8 and ttmp9
+  s_mov_b64 ttmp10, 0       //Clear ttmp10 and ttmp11
+  s_mov_b64 ttmp12, 0       //Clear ttmp12 and ttmp13
+  s_mov_b64 ttmp14, 0       //Clear ttmp14 and ttmp15
+
+ s_endpgm
+
+end
+
+
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c
new file mode 100644
index 000000000000..8a2ee2de390f
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c
@@ -0,0 +1,7538 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include <linux/delay.h>
+#include <linux/kernel.h>
+#include <linux/firmware.h>
+#include <linux/module.h>
+#include <linux/pci.h>
+#include "amdgpu.h"
+#include "amdgpu_gfx.h"
+#include "amdgpu_psp.h"
+#include "amdgpu_smu.h"
+#include "imu_v11_0.h"
+#include "soc21.h"
+#include "nvd.h"
+
+#include "gc/gc_11_0_0_offset.h"
+#include "gc/gc_11_0_0_sh_mask.h"
+#include "smuio/smuio_13_0_6_offset.h"
+#include "smuio/smuio_13_0_6_sh_mask.h"
+#include "navi10_enum.h"
+#include "ivsrcid/gfx/irqsrcs_gfx_11_0_0.h"
+
+#include "soc15.h"
+#include "clearstate_gfx11.h"
+#include "v11_structs.h"
+#include "gfx_v11_0.h"
+#include "gfx_v11_0_cleaner_shader.h"
+#include "gfx_v11_0_3.h"
+#include "nbio_v4_3.h"
+#include "mes_v11_0.h"
+#include "mes_userqueue.h"
+#include "amdgpu_userq_fence.h"
+
+#define GFX11_NUM_GFX_RINGS		1
+#define GFX11_MEC_HPD_SIZE	2048
+
+#define RLCG_UCODE_LOADING_START_ADDRESS	0x00002000L
+#define RLC_PG_DELAY_3_DEFAULT_GC_11_0_1	0x1388
+
+#define regCGTT_WD_CLK_CTRL		0x5086
+#define regCGTT_WD_CLK_CTRL_BASE_IDX	1
+#define regRLC_RLCS_BOOTLOAD_STATUS_gc_11_0_1	0x4e7e
+#define regRLC_RLCS_BOOTLOAD_STATUS_gc_11_0_1_BASE_IDX	1
+#define regPC_CONFIG_CNTL_1		0x194d
+#define regPC_CONFIG_CNTL_1_BASE_IDX	1
+
+#define regCP_GFX_MQD_CONTROL_DEFAULT                                             0x00000100
+#define regCP_GFX_HQD_VMID_DEFAULT                                                0x00000000
+#define regCP_GFX_HQD_QUEUE_PRIORITY_DEFAULT                                      0x00000000
+#define regCP_GFX_HQD_QUANTUM_DEFAULT                                             0x00000a01
+#define regCP_GFX_HQD_CNTL_DEFAULT                                                0x00a00000
+#define regCP_RB_DOORBELL_CONTROL_DEFAULT                                         0x00000000
+#define regCP_GFX_HQD_RPTR_DEFAULT                                                0x00000000
+
+#define regCP_HQD_EOP_CONTROL_DEFAULT                                             0x00000006
+#define regCP_HQD_PQ_DOORBELL_CONTROL_DEFAULT                                     0x00000000
+#define regCP_MQD_CONTROL_DEFAULT                                                 0x00000100
+#define regCP_HQD_PQ_CONTROL_DEFAULT                                              0x00308509
+#define regCP_HQD_PQ_DOORBELL_CONTROL_DEFAULT                                     0x00000000
+#define regCP_HQD_PQ_RPTR_DEFAULT                                                 0x00000000
+#define regCP_HQD_PERSISTENT_STATE_DEFAULT                                        0x0be05501
+#define regCP_HQD_IB_CONTROL_DEFAULT                                              0x00300000
+
+MODULE_FIRMWARE("amdgpu/gc_11_0_0_pfp.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_0_0_me.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_0_0_mec.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_0_0_rlc.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_0_0_rlc_kicker.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_0_0_rlc_1.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_0_0_toc.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_0_1_pfp.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_0_1_me.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_0_1_mec.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_0_1_rlc.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_0_2_pfp.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_0_2_me.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_0_2_mec.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_0_2_rlc.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_0_3_pfp.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_0_3_me.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_0_3_mec.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_0_3_rlc.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_0_4_pfp.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_0_4_me.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_0_4_mec.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_0_4_rlc.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_5_0_pfp.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_5_0_me.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_5_0_mec.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_5_0_rlc.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_5_1_pfp.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_5_1_me.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_5_1_mec.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_5_1_rlc.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_5_2_pfp.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_5_2_me.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_5_2_mec.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_5_2_rlc.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_5_3_pfp.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_5_3_me.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_5_3_mec.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_5_3_rlc.bin");
+
+static const struct amdgpu_hwip_reg_entry gc_reg_list_11_0[] = {
+	SOC15_REG_ENTRY_STR(GC, 0, regGRBM_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regGRBM_STATUS2),
+	SOC15_REG_ENTRY_STR(GC, 0, regGRBM_STATUS3),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_STALLED_STAT1),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_STALLED_STAT2),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_STALLED_STAT3),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_CPC_STALLED_STAT1),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_CPF_STALLED_STAT1),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_BUSY_STAT),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_CPC_BUSY_STAT),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_CPF_BUSY_STAT),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_CPC_BUSY_STAT2),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_CPF_BUSY_STAT2),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_CPF_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_ERROR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HPD_STATUS0),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_RB_BASE),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_RB_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_RB_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_RB0_BASE),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_RB0_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_RB0_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_RB1_BASE),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_RB1_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_RB1_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_IB1_CMD_BUFSZ),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_IB2_CMD_BUFSZ),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_IB1_BASE_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_IB1_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_IB1_BUFSZ),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_IB2_BASE_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_IB2_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_IB2_BUFSZ),
+	SOC15_REG_ENTRY_STR(GC, 0, regCPF_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regCPC_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regCPG_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regGDS_PROTECTION_FAULT),
+	SOC15_REG_ENTRY_STR(GC, 0, regGDS_VM_PROTECTION_FAULT),
+	SOC15_REG_ENTRY_STR(GC, 0, regIA_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regIA_UTCL1_STATUS_2),
+	SOC15_REG_ENTRY_STR(GC, 0, regPA_CL_CNTL_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regRLC_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regRMI_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regSQC_CACHES),
+	SOC15_REG_ENTRY_STR(GC, 0, regSQG_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regWD_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regGCVM_L2_PROTECTION_FAULT_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, regGCVM_L2_PROTECTION_FAULT_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_DEBUG),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MES_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC1_INSTR_PNTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MES_DEBUG_INTERRUPT_INSTR_PNTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MES_INSTR_PNTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_INSTR_PNTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_PFP_INSTR_PNTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_CPC_STATUS),
+	/* cp header registers */
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MES_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MES_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MES_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MES_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MES_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MES_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MES_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MES_HEADER_DUMP),
+	/* SE status registers */
+	SOC15_REG_ENTRY_STR(GC, 0, regGRBM_STATUS_SE0),
+	SOC15_REG_ENTRY_STR(GC, 0, regGRBM_STATUS_SE1),
+	SOC15_REG_ENTRY_STR(GC, 0, regGRBM_STATUS_SE2),
+	SOC15_REG_ENTRY_STR(GC, 0, regGRBM_STATUS_SE3),
+	SOC15_REG_ENTRY_STR(GC, 0, regGRBM_STATUS_SE4),
+	SOC15_REG_ENTRY_STR(GC, 0, regGRBM_STATUS_SE5)
+};
+
+static const struct amdgpu_hwip_reg_entry gc_cp_reg_list_11[] = {
+	/* compute registers */
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_VMID),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PERSISTENT_STATE),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PIPE_PRIORITY),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_QUEUE_PRIORITY),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_QUANTUM),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PQ_BASE),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PQ_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PQ_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PQ_WPTR_POLL_ADDR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PQ_WPTR_POLL_ADDR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PQ_DOORBELL_CONTROL),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PQ_CONTROL),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_IB_BASE_ADDR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_IB_BASE_ADDR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_IB_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_IB_CONTROL),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_DEQUEUE_REQUEST),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_EOP_BASE_ADDR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_EOP_BASE_ADDR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_EOP_CONTROL),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_EOP_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_EOP_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_EOP_EVENTS),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_CTX_SAVE_BASE_ADDR_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_CTX_SAVE_BASE_ADDR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_CTX_SAVE_CONTROL),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_CNTL_STACK_OFFSET),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_CNTL_STACK_SIZE),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_WG_STATE_OFFSET),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_CTX_SAVE_SIZE),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_GDS_RESOURCE_STATE),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_ERROR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_EOP_WPTR_MEM),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PQ_WPTR_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PQ_WPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_SUSPEND_CNTL_STACK_OFFSET),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_SUSPEND_CNTL_STACK_DW_CNT),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_SUSPEND_WG_STATE_OFFSET),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_DEQUEUE_STATUS),
+	/* cp header registers */
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP),
+};
+
+static const struct amdgpu_hwip_reg_entry gc_gfx_queue_reg_list_11[] = {
+	/* gfx queue registers */
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_ACTIVE),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_VMID),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_QUEUE_PRIORITY),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_QUANTUM),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_BASE),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_OFFSET),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_CSMD_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_WPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_DEQUEUE_REQUEST),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_MAPPED),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_QUE_MGR_CONTROL),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_HQ_CONTROL0),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_HQ_STATUS0),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_MQD_BASE_ADDR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_MQD_BASE_ADDR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_RB_WPTR_POLL_ADDR_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_RB_WPTR_POLL_ADDR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_RB_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_IB1_BASE_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_IB1_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_IB1_CMD_BUFSZ),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_IB1_BUFSZ),
+	/* cp header registers */
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_PFP_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_PFP_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_PFP_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_PFP_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_PFP_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_PFP_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_PFP_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_PFP_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP),
+};
+
+static const struct soc15_reg_golden golden_settings_gc_11_0[] = {
+	SOC15_REG_GOLDEN_VALUE(GC, 0, regTCP_CNTL, 0x20000000, 0x20000000)
+};
+
+static const struct soc15_reg_golden golden_settings_gc_11_0_1[] =
+{
+	SOC15_REG_GOLDEN_VALUE(GC, 0, regCGTT_GS_NGG_CLK_CTRL, 0x9fff8fff, 0x00000010),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, regCGTT_WD_CLK_CTRL, 0xffff8fff, 0x00000010),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, regCPF_GCR_CNTL, 0x0007ffff, 0x0000c200),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, regGL2C_CTRL3, 0xffff001b, 0x00f01988),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, regPA_CL_ENHANCE, 0xf0ffffff, 0x00880007),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, regPA_SC_ENHANCE_3, 0xfffffffd, 0x00000008),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, regPA_SC_VRS_SURFACE_CNTL_1, 0xfff891ff, 0x55480100),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, regTA_CNTL_AUX, 0xf7f7ffff, 0x01030000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, regTCP_CNTL2, 0xfcffffff, 0x0000000a)
+};
+
+#define DEFAULT_SH_MEM_CONFIG \
+	((SH_MEM_ADDRESS_MODE_64 << SH_MEM_CONFIG__ADDRESS_MODE__SHIFT) | \
+	 (SH_MEM_ALIGNMENT_MODE_UNALIGNED << SH_MEM_CONFIG__ALIGNMENT_MODE__SHIFT) | \
+	 (3 << SH_MEM_CONFIG__INITIAL_INST_PREFETCH__SHIFT))
+
+static void gfx_v11_0_disable_gpa_mode(struct amdgpu_device *adev);
+static void gfx_v11_0_set_ring_funcs(struct amdgpu_device *adev);
+static void gfx_v11_0_set_irq_funcs(struct amdgpu_device *adev);
+static void gfx_v11_0_set_gds_init(struct amdgpu_device *adev);
+static void gfx_v11_0_set_rlc_funcs(struct amdgpu_device *adev);
+static void gfx_v11_0_set_mqd_funcs(struct amdgpu_device *adev);
+static void gfx_v11_0_set_imu_funcs(struct amdgpu_device *adev);
+static int gfx_v11_0_get_cu_info(struct amdgpu_device *adev,
+                                 struct amdgpu_cu_info *cu_info);
+static uint64_t gfx_v11_0_get_gpu_clock_counter(struct amdgpu_device *adev);
+static void gfx_v11_0_select_se_sh(struct amdgpu_device *adev, u32 se_num,
+				   u32 sh_num, u32 instance, int xcc_id);
+static u32 gfx_v11_0_get_wgp_active_bitmap_per_sh(struct amdgpu_device *adev);
+
+static void gfx_v11_0_ring_emit_de_meta(struct amdgpu_ring *ring, bool resume);
+static void gfx_v11_0_ring_emit_frame_cntl(struct amdgpu_ring *ring, bool start, bool secure);
+static void gfx_v11_0_ring_emit_wreg(struct amdgpu_ring *ring, uint32_t reg,
+				     uint32_t val);
+static int gfx_v11_0_wait_for_rlc_autoload_complete(struct amdgpu_device *adev);
+static void gfx_v11_0_ring_invalidate_tlbs(struct amdgpu_ring *ring,
+					   uint16_t pasid, uint32_t flush_type,
+					   bool all_hub, uint8_t dst_sel);
+static void gfx_v11_0_set_safe_mode(struct amdgpu_device *adev, int xcc_id);
+static void gfx_v11_0_unset_safe_mode(struct amdgpu_device *adev, int xcc_id);
+static void gfx_v11_0_update_perf_clk(struct amdgpu_device *adev,
+				      bool enable);
+
+static void gfx11_kiq_set_resources(struct amdgpu_ring *kiq_ring, uint64_t queue_mask)
+{
+	struct amdgpu_device *adev = kiq_ring->adev;
+	u64 shader_mc_addr;
+
+	/* Cleaner shader MC address */
+	shader_mc_addr = adev->gfx.cleaner_shader_gpu_addr >> 8;
+
+	amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_SET_RESOURCES, 6));
+	amdgpu_ring_write(kiq_ring, PACKET3_SET_RESOURCES_VMID_MASK(0) |
+			  PACKET3_SET_RESOURCES_UNMAP_LATENTY(0xa) | /* unmap_latency: 0xa (~ 1s) */
+			  PACKET3_SET_RESOURCES_QUEUE_TYPE(0));	/* vmid_mask:0 queue_type:0 (KIQ) */
+	amdgpu_ring_write(kiq_ring, lower_32_bits(queue_mask));	/* queue mask lo */
+	amdgpu_ring_write(kiq_ring, upper_32_bits(queue_mask));	/* queue mask hi */
+	amdgpu_ring_write(kiq_ring, lower_32_bits(shader_mc_addr)); /* cleaner shader addr lo */
+	amdgpu_ring_write(kiq_ring, upper_32_bits(shader_mc_addr)); /* cleaner shader addr hi */
+	amdgpu_ring_write(kiq_ring, 0);	/* oac mask */
+	amdgpu_ring_write(kiq_ring, 0);	/* gds heap base:0, gds heap size:0 */
+}
+
+static void gfx11_kiq_map_queues(struct amdgpu_ring *kiq_ring,
+				 struct amdgpu_ring *ring)
+{
+	uint64_t mqd_addr = amdgpu_bo_gpu_offset(ring->mqd_obj);
+	uint64_t wptr_addr = ring->wptr_gpu_addr;
+	uint32_t me = 0, eng_sel = 0;
+
+	switch (ring->funcs->type) {
+	case AMDGPU_RING_TYPE_COMPUTE:
+		me = 1;
+		eng_sel = 0;
+		break;
+	case AMDGPU_RING_TYPE_GFX:
+		me = 0;
+		eng_sel = 4;
+		break;
+	case AMDGPU_RING_TYPE_MES:
+		me = 2;
+		eng_sel = 5;
+		break;
+	default:
+		WARN_ON(1);
+	}
+
+	amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_MAP_QUEUES, 5));
+	/* Q_sel:0, vmid:0, vidmem: 1, engine:0, num_Q:1*/
+	amdgpu_ring_write(kiq_ring, /* Q_sel: 0, vmid: 0, engine: 0, num_Q: 1 */
+			  PACKET3_MAP_QUEUES_QUEUE_SEL(0) | /* Queue_Sel */
+			  PACKET3_MAP_QUEUES_VMID(0) | /* VMID */
+			  PACKET3_MAP_QUEUES_QUEUE(ring->queue) |
+			  PACKET3_MAP_QUEUES_PIPE(ring->pipe) |
+			  PACKET3_MAP_QUEUES_ME((me)) |
+			  PACKET3_MAP_QUEUES_QUEUE_TYPE(0) | /*queue_type: normal compute queue */
+			  PACKET3_MAP_QUEUES_ALLOC_FORMAT(0) | /* alloc format: all_on_one_pipe */
+			  PACKET3_MAP_QUEUES_ENGINE_SEL(eng_sel) |
+			  PACKET3_MAP_QUEUES_NUM_QUEUES(1)); /* num_queues: must be 1 */
+	amdgpu_ring_write(kiq_ring, PACKET3_MAP_QUEUES_DOORBELL_OFFSET(ring->doorbell_index));
+	amdgpu_ring_write(kiq_ring, lower_32_bits(mqd_addr));
+	amdgpu_ring_write(kiq_ring, upper_32_bits(mqd_addr));
+	amdgpu_ring_write(kiq_ring, lower_32_bits(wptr_addr));
+	amdgpu_ring_write(kiq_ring, upper_32_bits(wptr_addr));
+}
+
+static void gfx11_kiq_unmap_queues(struct amdgpu_ring *kiq_ring,
+				   struct amdgpu_ring *ring,
+				   enum amdgpu_unmap_queues_action action,
+				   u64 gpu_addr, u64 seq)
+{
+	struct amdgpu_device *adev = kiq_ring->adev;
+	uint32_t eng_sel = ring->funcs->type == AMDGPU_RING_TYPE_GFX ? 4 : 0;
+
+	if (adev->enable_mes && !adev->gfx.kiq[0].ring.sched.ready) {
+		amdgpu_mes_unmap_legacy_queue(adev, ring, action, gpu_addr, seq);
+		return;
+	}
+
+	amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_UNMAP_QUEUES, 4));
+	amdgpu_ring_write(kiq_ring, /* Q_sel: 0, vmid: 0, engine: 0, num_Q: 1 */
+			  PACKET3_UNMAP_QUEUES_ACTION(action) |
+			  PACKET3_UNMAP_QUEUES_QUEUE_SEL(0) |
+			  PACKET3_UNMAP_QUEUES_ENGINE_SEL(eng_sel) |
+			  PACKET3_UNMAP_QUEUES_NUM_QUEUES(1));
+	amdgpu_ring_write(kiq_ring,
+		  PACKET3_UNMAP_QUEUES_DOORBELL_OFFSET0(ring->doorbell_index));
+
+	if (action == PREEMPT_QUEUES_NO_UNMAP) {
+		amdgpu_ring_write(kiq_ring, lower_32_bits(gpu_addr));
+		amdgpu_ring_write(kiq_ring, upper_32_bits(gpu_addr));
+		amdgpu_ring_write(kiq_ring, seq);
+	} else {
+		amdgpu_ring_write(kiq_ring, 0);
+		amdgpu_ring_write(kiq_ring, 0);
+		amdgpu_ring_write(kiq_ring, 0);
+	}
+}
+
+static void gfx11_kiq_query_status(struct amdgpu_ring *kiq_ring,
+				   struct amdgpu_ring *ring,
+				   u64 addr,
+				   u64 seq)
+{
+	uint32_t eng_sel = ring->funcs->type == AMDGPU_RING_TYPE_GFX ? 4 : 0;
+
+	amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_QUERY_STATUS, 5));
+	amdgpu_ring_write(kiq_ring,
+			  PACKET3_QUERY_STATUS_CONTEXT_ID(0) |
+			  PACKET3_QUERY_STATUS_INTERRUPT_SEL(0) |
+			  PACKET3_QUERY_STATUS_COMMAND(2));
+	amdgpu_ring_write(kiq_ring, /* Q_sel: 0, vmid: 0, engine: 0, num_Q: 1 */
+			  PACKET3_QUERY_STATUS_DOORBELL_OFFSET(ring->doorbell_index) |
+			  PACKET3_QUERY_STATUS_ENG_SEL(eng_sel));
+	amdgpu_ring_write(kiq_ring, lower_32_bits(addr));
+	amdgpu_ring_write(kiq_ring, upper_32_bits(addr));
+	amdgpu_ring_write(kiq_ring, lower_32_bits(seq));
+	amdgpu_ring_write(kiq_ring, upper_32_bits(seq));
+}
+
+static void gfx11_kiq_invalidate_tlbs(struct amdgpu_ring *kiq_ring,
+				uint16_t pasid, uint32_t flush_type,
+				bool all_hub)
+{
+	gfx_v11_0_ring_invalidate_tlbs(kiq_ring, pasid, flush_type, all_hub, 1);
+}
+
+static const struct kiq_pm4_funcs gfx_v11_0_kiq_pm4_funcs = {
+	.kiq_set_resources = gfx11_kiq_set_resources,
+	.kiq_map_queues = gfx11_kiq_map_queues,
+	.kiq_unmap_queues = gfx11_kiq_unmap_queues,
+	.kiq_query_status = gfx11_kiq_query_status,
+	.kiq_invalidate_tlbs = gfx11_kiq_invalidate_tlbs,
+	.set_resources_size = 8,
+	.map_queues_size = 7,
+	.unmap_queues_size = 6,
+	.query_status_size = 7,
+	.invalidate_tlbs_size = 2,
+};
+
+static void gfx_v11_0_set_kiq_pm4_funcs(struct amdgpu_device *adev)
+{
+	adev->gfx.kiq[0].pmf = &gfx_v11_0_kiq_pm4_funcs;
+}
+
+static void gfx_v11_0_init_golden_registers(struct amdgpu_device *adev)
+{
+	if (amdgpu_sriov_vf(adev))
+		return;
+
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	case IP_VERSION(11, 0, 1):
+	case IP_VERSION(11, 0, 4):
+		soc15_program_register_sequence(adev,
+						golden_settings_gc_11_0_1,
+						(const u32)ARRAY_SIZE(golden_settings_gc_11_0_1));
+		break;
+	default:
+		break;
+	}
+	soc15_program_register_sequence(adev,
+					golden_settings_gc_11_0,
+					(const u32)ARRAY_SIZE(golden_settings_gc_11_0));
+
+}
+
+static void gfx_v11_0_write_data_to_reg(struct amdgpu_ring *ring, int eng_sel,
+				       bool wc, uint32_t reg, uint32_t val)
+{
+	amdgpu_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
+	amdgpu_ring_write(ring, WRITE_DATA_ENGINE_SEL(eng_sel) |
+			  WRITE_DATA_DST_SEL(0) | (wc ? WR_CONFIRM : 0));
+	amdgpu_ring_write(ring, reg);
+	amdgpu_ring_write(ring, 0);
+	amdgpu_ring_write(ring, val);
+}
+
+static void gfx_v11_0_wait_reg_mem(struct amdgpu_ring *ring, int eng_sel,
+				  int mem_space, int opt, uint32_t addr0,
+				  uint32_t addr1, uint32_t ref, uint32_t mask,
+				  uint32_t inv)
+{
+	amdgpu_ring_write(ring, PACKET3(PACKET3_WAIT_REG_MEM, 5));
+	amdgpu_ring_write(ring,
+			  /* memory (1) or register (0) */
+			  (WAIT_REG_MEM_MEM_SPACE(mem_space) |
+			   WAIT_REG_MEM_OPERATION(opt) | /* wait */
+			   WAIT_REG_MEM_FUNCTION(3) |  /* equal */
+			   WAIT_REG_MEM_ENGINE(eng_sel)));
+
+	if (mem_space)
+		BUG_ON(addr0 & 0x3); /* Dword align */
+	amdgpu_ring_write(ring, addr0);
+	amdgpu_ring_write(ring, addr1);
+	amdgpu_ring_write(ring, ref);
+	amdgpu_ring_write(ring, mask);
+	amdgpu_ring_write(ring, inv); /* poll interval */
+}
+
+static void gfx_v11_ring_insert_nop(struct amdgpu_ring *ring, uint32_t num_nop)
+{
+	/* Header itself is a NOP packet */
+	if (num_nop == 1) {
+		amdgpu_ring_write(ring, ring->funcs->nop);
+		return;
+	}
+
+	/* Max HW optimization till 0x3ffe, followed by remaining one NOP at a time*/
+	amdgpu_ring_write(ring, PACKET3(PACKET3_NOP, min(num_nop - 2, 0x3ffe)));
+
+	/* Header is at index 0, followed by num_nops - 1 NOP packet's */
+	amdgpu_ring_insert_nop(ring, num_nop - 1);
+}
+
+static int gfx_v11_0_ring_test_ring(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	uint32_t scratch = SOC15_REG_OFFSET(GC, 0, regSCRATCH_REG0);
+	uint32_t tmp = 0;
+	unsigned i;
+	int r;
+
+	WREG32(scratch, 0xCAFEDEAD);
+	r = amdgpu_ring_alloc(ring, 5);
+	if (r) {
+		DRM_ERROR("amdgpu: cp failed to lock ring %d (%d).\n",
+			  ring->idx, r);
+		return r;
+	}
+
+	if (ring->funcs->type == AMDGPU_RING_TYPE_KIQ) {
+		gfx_v11_0_ring_emit_wreg(ring, scratch, 0xDEADBEEF);
+	} else {
+		amdgpu_ring_write(ring, PACKET3(PACKET3_SET_UCONFIG_REG, 1));
+		amdgpu_ring_write(ring, scratch -
+				  PACKET3_SET_UCONFIG_REG_START);
+		amdgpu_ring_write(ring, 0xDEADBEEF);
+	}
+	amdgpu_ring_commit(ring);
+
+	for (i = 0; i < adev->usec_timeout; i++) {
+		tmp = RREG32(scratch);
+		if (tmp == 0xDEADBEEF)
+			break;
+		if (amdgpu_emu_mode == 1)
+			msleep(1);
+		else
+			udelay(1);
+	}
+
+	if (i >= adev->usec_timeout)
+		r = -ETIMEDOUT;
+	return r;
+}
+
+static int gfx_v11_0_ring_test_ib(struct amdgpu_ring *ring, long timeout)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct amdgpu_ib ib;
+	struct dma_fence *f = NULL;
+	unsigned index;
+	uint64_t gpu_addr;
+	uint32_t *cpu_ptr;
+	long r;
+
+	/* MES KIQ fw hasn't indirect buffer support for now */
+	if (adev->enable_mes_kiq &&
+	    ring->funcs->type == AMDGPU_RING_TYPE_KIQ)
+		return 0;
+
+	memset(&ib, 0, sizeof(ib));
+
+	r = amdgpu_device_wb_get(adev, &index);
+	if (r)
+		return r;
+
+	gpu_addr = adev->wb.gpu_addr + (index * 4);
+	adev->wb.wb[index] = cpu_to_le32(0xCAFEDEAD);
+	cpu_ptr = &adev->wb.wb[index];
+
+	r = amdgpu_ib_get(adev, NULL, 20, AMDGPU_IB_POOL_DIRECT, &ib);
+	if (r) {
+		DRM_ERROR("amdgpu: failed to get ib (%ld).\n", r);
+		goto err1;
+	}
+
+	ib.ptr[0] = PACKET3(PACKET3_WRITE_DATA, 3);
+	ib.ptr[1] = WRITE_DATA_DST_SEL(5) | WR_CONFIRM;
+	ib.ptr[2] = lower_32_bits(gpu_addr);
+	ib.ptr[3] = upper_32_bits(gpu_addr);
+	ib.ptr[4] = 0xDEADBEEF;
+	ib.length_dw = 5;
+
+	r = amdgpu_ib_schedule(ring, 1, &ib, NULL, &f);
+	if (r)
+		goto err2;
+
+	r = dma_fence_wait_timeout(f, false, timeout);
+	if (r == 0) {
+		r = -ETIMEDOUT;
+		goto err2;
+	} else if (r < 0) {
+		goto err2;
+	}
+
+	if (le32_to_cpu(*cpu_ptr) == 0xDEADBEEF)
+		r = 0;
+	else
+		r = -EINVAL;
+err2:
+	amdgpu_ib_free(&ib, NULL);
+	dma_fence_put(f);
+err1:
+	amdgpu_device_wb_free(adev, index);
+	return r;
+}
+
+static void gfx_v11_0_free_microcode(struct amdgpu_device *adev)
+{
+	amdgpu_ucode_release(&adev->gfx.pfp_fw);
+	amdgpu_ucode_release(&adev->gfx.me_fw);
+	amdgpu_ucode_release(&adev->gfx.rlc_fw);
+	amdgpu_ucode_release(&adev->gfx.mec_fw);
+
+	kfree(adev->gfx.rlc.register_list_format);
+}
+
+static int gfx_v11_0_init_toc_microcode(struct amdgpu_device *adev, const char *ucode_prefix)
+{
+	const struct psp_firmware_header_v1_0 *toc_hdr;
+	int err = 0;
+
+	err = amdgpu_ucode_request(adev, &adev->psp.toc_fw,
+				   AMDGPU_UCODE_REQUIRED,
+				   "amdgpu/%s_toc.bin", ucode_prefix);
+	if (err)
+		goto out;
+
+	toc_hdr = (const struct psp_firmware_header_v1_0 *)adev->psp.toc_fw->data;
+	adev->psp.toc.fw_version = le32_to_cpu(toc_hdr->header.ucode_version);
+	adev->psp.toc.feature_version = le32_to_cpu(toc_hdr->sos.fw_version);
+	adev->psp.toc.size_bytes = le32_to_cpu(toc_hdr->header.ucode_size_bytes);
+	adev->psp.toc.start_addr = (uint8_t *)toc_hdr +
+				le32_to_cpu(toc_hdr->header.ucode_array_offset_bytes);
+	return 0;
+out:
+	amdgpu_ucode_release(&adev->psp.toc_fw);
+	return err;
+}
+
+static void gfx_v11_0_check_fw_cp_gfx_shadow(struct amdgpu_device *adev)
+{
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	case IP_VERSION(11, 0, 0):
+	case IP_VERSION(11, 0, 2):
+	case IP_VERSION(11, 0, 3):
+		if ((adev->gfx.me_fw_version >= 1505) &&
+		    (adev->gfx.pfp_fw_version >= 1600) &&
+		    (adev->gfx.mec_fw_version >= 512)) {
+			if (amdgpu_sriov_vf(adev))
+				adev->gfx.cp_gfx_shadow = true;
+			else
+				adev->gfx.cp_gfx_shadow = false;
+		}
+		break;
+	default:
+		adev->gfx.cp_gfx_shadow = false;
+		break;
+	}
+}
+
+static int gfx_v11_0_init_microcode(struct amdgpu_device *adev)
+{
+	char ucode_prefix[25];
+	int err;
+	const struct rlc_firmware_header_v2_0 *rlc_hdr;
+	uint16_t version_major;
+	uint16_t version_minor;
+
+	DRM_DEBUG("\n");
+
+	amdgpu_ucode_ip_version_decode(adev, GC_HWIP, ucode_prefix, sizeof(ucode_prefix));
+	err = amdgpu_ucode_request(adev, &adev->gfx.pfp_fw,
+				   AMDGPU_UCODE_REQUIRED,
+				   "amdgpu/%s_pfp.bin", ucode_prefix);
+	if (err)
+		goto out;
+	/* check pfp fw hdr version to decide if enable rs64 for gfx11.*/
+	adev->gfx.rs64_enable = amdgpu_ucode_hdr_version(
+				(union amdgpu_firmware_header *)
+				adev->gfx.pfp_fw->data, 2, 0);
+	if (adev->gfx.rs64_enable) {
+		dev_info(adev->dev, "CP RS64 enable\n");
+		amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_RS64_PFP);
+		amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_RS64_PFP_P0_STACK);
+		amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_RS64_PFP_P1_STACK);
+	} else {
+		amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_PFP);
+	}
+
+	err = amdgpu_ucode_request(adev, &adev->gfx.me_fw,
+				   AMDGPU_UCODE_REQUIRED,
+				   "amdgpu/%s_me.bin", ucode_prefix);
+	if (err)
+		goto out;
+	if (adev->gfx.rs64_enable) {
+		amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_RS64_ME);
+		amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_RS64_ME_P0_STACK);
+		amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_RS64_ME_P1_STACK);
+	} else {
+		amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_ME);
+	}
+
+	if (!amdgpu_sriov_vf(adev)) {
+		if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(11, 0, 0) &&
+		    adev->pdev->revision == 0xCE)
+			err = amdgpu_ucode_request(adev, &adev->gfx.rlc_fw,
+						   AMDGPU_UCODE_REQUIRED,
+						   "amdgpu/gc_11_0_0_rlc_1.bin");
+		else if (amdgpu_is_kicker_fw(adev))
+			err = amdgpu_ucode_request(adev, &adev->gfx.rlc_fw,
+						   AMDGPU_UCODE_REQUIRED,
+						   "amdgpu/%s_rlc_kicker.bin", ucode_prefix);
+		else
+			err = amdgpu_ucode_request(adev, &adev->gfx.rlc_fw,
+						   AMDGPU_UCODE_REQUIRED,
+						   "amdgpu/%s_rlc.bin", ucode_prefix);
+		if (err)
+			goto out;
+		rlc_hdr = (const struct rlc_firmware_header_v2_0 *)adev->gfx.rlc_fw->data;
+		version_major = le16_to_cpu(rlc_hdr->header.header_version_major);
+		version_minor = le16_to_cpu(rlc_hdr->header.header_version_minor);
+		err = amdgpu_gfx_rlc_init_microcode(adev, version_major, version_minor);
+		if (err)
+			goto out;
+	}
+
+	err = amdgpu_ucode_request(adev, &adev->gfx.mec_fw,
+				   AMDGPU_UCODE_REQUIRED,
+				   "amdgpu/%s_mec.bin", ucode_prefix);
+	if (err)
+		goto out;
+	if (adev->gfx.rs64_enable) {
+		amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_RS64_MEC);
+		amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_RS64_MEC_P0_STACK);
+		amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_RS64_MEC_P1_STACK);
+		amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_RS64_MEC_P2_STACK);
+		amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_RS64_MEC_P3_STACK);
+	} else {
+		amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_MEC1);
+		amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_MEC1_JT);
+	}
+
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_RLC_BACKDOOR_AUTO)
+		err = gfx_v11_0_init_toc_microcode(adev, ucode_prefix);
+
+	/* only one MEC for gfx 11.0.0. */
+	adev->gfx.mec2_fw = NULL;
+
+	gfx_v11_0_check_fw_cp_gfx_shadow(adev);
+
+	if (adev->gfx.imu.funcs && adev->gfx.imu.funcs->init_microcode) {
+		err = adev->gfx.imu.funcs->init_microcode(adev);
+		if (err)
+			DRM_ERROR("Failed to init imu firmware!\n");
+		return err;
+	}
+
+out:
+	if (err) {
+		amdgpu_ucode_release(&adev->gfx.pfp_fw);
+		amdgpu_ucode_release(&adev->gfx.me_fw);
+		amdgpu_ucode_release(&adev->gfx.rlc_fw);
+		amdgpu_ucode_release(&adev->gfx.mec_fw);
+	}
+
+	return err;
+}
+
+static u32 gfx_v11_0_get_csb_size(struct amdgpu_device *adev)
+{
+	u32 count = 0;
+	const struct cs_section_def *sect = NULL;
+	const struct cs_extent_def *ext = NULL;
+
+	/* begin clear state */
+	count += 2;
+	/* context control state */
+	count += 3;
+
+	for (sect = gfx11_cs_data; sect->section != NULL; ++sect) {
+		for (ext = sect->section; ext->extent != NULL; ++ext) {
+			if (sect->id == SECT_CONTEXT)
+				count += 2 + ext->reg_count;
+			else
+				return 0;
+		}
+	}
+
+	/* set PA_SC_TILE_STEERING_OVERRIDE */
+	count += 3;
+	/* end clear state */
+	count += 2;
+	/* clear state */
+	count += 2;
+
+	return count;
+}
+
+static void gfx_v11_0_get_csb_buffer(struct amdgpu_device *adev, u32 *buffer)
+{
+	u32 count = 0;
+	int ctx_reg_offset;
+
+	if (adev->gfx.rlc.cs_data == NULL)
+		return;
+	if (buffer == NULL)
+		return;
+
+	count = amdgpu_gfx_csb_preamble_start(buffer);
+	count = amdgpu_gfx_csb_data_parser(adev, buffer, count);
+
+	ctx_reg_offset = SOC15_REG_OFFSET(GC, 0, regPA_SC_TILE_STEERING_OVERRIDE) - PACKET3_SET_CONTEXT_REG_START;
+	buffer[count++] = cpu_to_le32(PACKET3(PACKET3_SET_CONTEXT_REG, 1));
+	buffer[count++] = cpu_to_le32(ctx_reg_offset);
+	buffer[count++] = cpu_to_le32(adev->gfx.config.pa_sc_tile_steering_override);
+
+	amdgpu_gfx_csb_preamble_end(buffer, count);
+}
+
+static void gfx_v11_0_rlc_fini(struct amdgpu_device *adev)
+{
+	/* clear state block */
+	amdgpu_bo_free_kernel(&adev->gfx.rlc.clear_state_obj,
+			&adev->gfx.rlc.clear_state_gpu_addr,
+			(void **)&adev->gfx.rlc.cs_ptr);
+
+	/* jump table block */
+	amdgpu_bo_free_kernel(&adev->gfx.rlc.cp_table_obj,
+			&adev->gfx.rlc.cp_table_gpu_addr,
+			(void **)&adev->gfx.rlc.cp_table_ptr);
+}
+
+static void gfx_v11_0_init_rlcg_reg_access_ctrl(struct amdgpu_device *adev)
+{
+	struct amdgpu_rlcg_reg_access_ctrl *reg_access_ctrl;
+
+	reg_access_ctrl = &adev->gfx.rlc.reg_access_ctrl[0];
+	reg_access_ctrl->scratch_reg0 = SOC15_REG_OFFSET(GC, 0, regSCRATCH_REG0);
+	reg_access_ctrl->scratch_reg1 = SOC15_REG_OFFSET(GC, 0, regSCRATCH_REG1);
+	reg_access_ctrl->scratch_reg2 = SOC15_REG_OFFSET(GC, 0, regSCRATCH_REG2);
+	reg_access_ctrl->scratch_reg3 = SOC15_REG_OFFSET(GC, 0, regSCRATCH_REG3);
+	reg_access_ctrl->grbm_cntl = SOC15_REG_OFFSET(GC, 0, regGRBM_GFX_CNTL);
+	reg_access_ctrl->grbm_idx = SOC15_REG_OFFSET(GC, 0, regGRBM_GFX_INDEX);
+	reg_access_ctrl->spare_int = SOC15_REG_OFFSET(GC, 0, regRLC_SPARE_INT_0);
+	adev->gfx.rlc.rlcg_reg_access_supported = true;
+}
+
+static int gfx_v11_0_rlc_init(struct amdgpu_device *adev)
+{
+	const struct cs_section_def *cs_data;
+	int r;
+
+	adev->gfx.rlc.cs_data = gfx11_cs_data;
+
+	cs_data = adev->gfx.rlc.cs_data;
+
+	if (cs_data) {
+		/* init clear state block */
+		r = amdgpu_gfx_rlc_init_csb(adev);
+		if (r)
+			return r;
+	}
+
+	/* init spm vmid with 0xf */
+	if (adev->gfx.rlc.funcs->update_spm_vmid)
+		adev->gfx.rlc.funcs->update_spm_vmid(adev, NULL, 0xf);
+
+	return 0;
+}
+
+static void gfx_v11_0_mec_fini(struct amdgpu_device *adev)
+{
+	amdgpu_bo_free_kernel(&adev->gfx.mec.hpd_eop_obj, NULL, NULL);
+	amdgpu_bo_free_kernel(&adev->gfx.mec.mec_fw_obj, NULL, NULL);
+	amdgpu_bo_free_kernel(&adev->gfx.mec.mec_fw_data_obj, NULL, NULL);
+}
+
+static void gfx_v11_0_me_init(struct amdgpu_device *adev)
+{
+	bitmap_zero(adev->gfx.me.queue_bitmap, AMDGPU_MAX_GFX_QUEUES);
+
+	amdgpu_gfx_graphics_queue_acquire(adev);
+}
+
+static int gfx_v11_0_mec_init(struct amdgpu_device *adev)
+{
+	int r;
+	u32 *hpd;
+	size_t mec_hpd_size;
+
+	bitmap_zero(adev->gfx.mec_bitmap[0].queue_bitmap, AMDGPU_MAX_COMPUTE_QUEUES);
+
+	/* take ownership of the relevant compute queues */
+	amdgpu_gfx_compute_queue_acquire(adev);
+	mec_hpd_size = adev->gfx.num_compute_rings * GFX11_MEC_HPD_SIZE;
+
+	if (mec_hpd_size) {
+		r = amdgpu_bo_create_reserved(adev, mec_hpd_size, PAGE_SIZE,
+					      AMDGPU_GEM_DOMAIN_GTT,
+					      &adev->gfx.mec.hpd_eop_obj,
+					      &adev->gfx.mec.hpd_eop_gpu_addr,
+					      (void **)&hpd);
+		if (r) {
+			dev_warn(adev->dev, "(%d) create HDP EOP bo failed\n", r);
+			gfx_v11_0_mec_fini(adev);
+			return r;
+		}
+
+		memset(hpd, 0, mec_hpd_size);
+
+		amdgpu_bo_kunmap(adev->gfx.mec.hpd_eop_obj);
+		amdgpu_bo_unreserve(adev->gfx.mec.hpd_eop_obj);
+	}
+
+	return 0;
+}
+
+static uint32_t wave_read_ind(struct amdgpu_device *adev, uint32_t wave, uint32_t address)
+{
+	WREG32_SOC15(GC, 0, regSQ_IND_INDEX,
+		(wave << SQ_IND_INDEX__WAVE_ID__SHIFT) |
+		(address << SQ_IND_INDEX__INDEX__SHIFT));
+	return RREG32_SOC15(GC, 0, regSQ_IND_DATA);
+}
+
+static void wave_read_regs(struct amdgpu_device *adev, uint32_t wave,
+			   uint32_t thread, uint32_t regno,
+			   uint32_t num, uint32_t *out)
+{
+	WREG32_SOC15(GC, 0, regSQ_IND_INDEX,
+		(wave << SQ_IND_INDEX__WAVE_ID__SHIFT) |
+		(regno << SQ_IND_INDEX__INDEX__SHIFT) |
+		(thread << SQ_IND_INDEX__WORKITEM_ID__SHIFT) |
+		(SQ_IND_INDEX__AUTO_INCR_MASK));
+	while (num--)
+		*(out++) = RREG32_SOC15(GC, 0, regSQ_IND_DATA);
+}
+
+static void gfx_v11_0_read_wave_data(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd, uint32_t wave, uint32_t *dst, int *no_fields)
+{
+	/* in gfx11 the SIMD_ID is specified as part of the INSTANCE
+	 * field when performing a select_se_sh so it should be
+	 * zero here */
+	WARN_ON(simd != 0);
+
+	/* type 3 wave data */
+	dst[(*no_fields)++] = 3;
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_STATUS);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_PC_LO);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_PC_HI);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_EXEC_LO);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_EXEC_HI);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_HW_ID1);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_HW_ID2);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_GPR_ALLOC);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_LDS_ALLOC);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_TRAPSTS);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_IB_STS);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_IB_STS2);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_IB_DBG1);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_M0);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_MODE);
+}
+
+static void gfx_v11_0_read_wave_sgprs(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd,
+				     uint32_t wave, uint32_t start,
+				     uint32_t size, uint32_t *dst)
+{
+	WARN_ON(simd != 0);
+
+	wave_read_regs(
+		adev, wave, 0, start + SQIND_WAVE_SGPRS_OFFSET, size,
+		dst);
+}
+
+static void gfx_v11_0_read_wave_vgprs(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd,
+				      uint32_t wave, uint32_t thread,
+				      uint32_t start, uint32_t size,
+				      uint32_t *dst)
+{
+	wave_read_regs(
+		adev, wave, thread,
+		start + SQIND_WAVE_VGPRS_OFFSET, size, dst);
+}
+
+static void gfx_v11_0_select_me_pipe_q(struct amdgpu_device *adev,
+					u32 me, u32 pipe, u32 q, u32 vm, u32 xcc_id)
+{
+	soc21_grbm_select(adev, me, pipe, q, vm);
+}
+
+/* all sizes are in bytes */
+#define MQD_SHADOW_BASE_SIZE      73728
+#define MQD_SHADOW_BASE_ALIGNMENT 256
+#define MQD_FWWORKAREA_SIZE       484
+#define MQD_FWWORKAREA_ALIGNMENT  256
+
+static void gfx_v11_0_get_gfx_shadow_info_nocheck(struct amdgpu_device *adev,
+					 struct amdgpu_gfx_shadow_info *shadow_info)
+{
+	shadow_info->shadow_size = MQD_SHADOW_BASE_SIZE;
+	shadow_info->shadow_alignment = MQD_SHADOW_BASE_ALIGNMENT;
+	shadow_info->csa_size = MQD_FWWORKAREA_SIZE;
+	shadow_info->csa_alignment = MQD_FWWORKAREA_ALIGNMENT;
+}
+
+static int gfx_v11_0_get_gfx_shadow_info(struct amdgpu_device *adev,
+					 struct amdgpu_gfx_shadow_info *shadow_info,
+					 bool skip_check)
+{
+	if (adev->gfx.cp_gfx_shadow || skip_check) {
+		gfx_v11_0_get_gfx_shadow_info_nocheck(adev, shadow_info);
+		return 0;
+	} else {
+		memset(shadow_info, 0, sizeof(struct amdgpu_gfx_shadow_info));
+		return -ENOTSUPP;
+	}
+}
+
+static const struct amdgpu_gfx_funcs gfx_v11_0_gfx_funcs = {
+	.get_gpu_clock_counter = &gfx_v11_0_get_gpu_clock_counter,
+	.select_se_sh = &gfx_v11_0_select_se_sh,
+	.read_wave_data = &gfx_v11_0_read_wave_data,
+	.read_wave_sgprs = &gfx_v11_0_read_wave_sgprs,
+	.read_wave_vgprs = &gfx_v11_0_read_wave_vgprs,
+	.select_me_pipe_q = &gfx_v11_0_select_me_pipe_q,
+	.update_perfmon_mgcg = &gfx_v11_0_update_perf_clk,
+	.get_gfx_shadow_info = &gfx_v11_0_get_gfx_shadow_info,
+};
+
+static int gfx_v11_0_gpu_early_init(struct amdgpu_device *adev)
+{
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	case IP_VERSION(11, 0, 0):
+	case IP_VERSION(11, 0, 2):
+		adev->gfx.config.max_hw_contexts = 8;
+		adev->gfx.config.sc_prim_fifo_size_frontend = 0x20;
+		adev->gfx.config.sc_prim_fifo_size_backend = 0x100;
+		adev->gfx.config.sc_hiz_tile_fifo_size = 0;
+		adev->gfx.config.sc_earlyz_tile_fifo_size = 0x4C0;
+		break;
+	case IP_VERSION(11, 0, 3):
+		adev->gfx.ras = &gfx_v11_0_3_ras;
+		adev->gfx.config.max_hw_contexts = 8;
+		adev->gfx.config.sc_prim_fifo_size_frontend = 0x20;
+		adev->gfx.config.sc_prim_fifo_size_backend = 0x100;
+		adev->gfx.config.sc_hiz_tile_fifo_size = 0;
+		adev->gfx.config.sc_earlyz_tile_fifo_size = 0x4C0;
+		break;
+	case IP_VERSION(11, 0, 1):
+	case IP_VERSION(11, 0, 4):
+	case IP_VERSION(11, 5, 0):
+	case IP_VERSION(11, 5, 1):
+	case IP_VERSION(11, 5, 2):
+	case IP_VERSION(11, 5, 3):
+		adev->gfx.config.max_hw_contexts = 8;
+		adev->gfx.config.sc_prim_fifo_size_frontend = 0x20;
+		adev->gfx.config.sc_prim_fifo_size_backend = 0x100;
+		adev->gfx.config.sc_hiz_tile_fifo_size = 0x80;
+		adev->gfx.config.sc_earlyz_tile_fifo_size = 0x300;
+		break;
+	default:
+		BUG();
+		break;
+	}
+
+	return 0;
+}
+
+static int gfx_v11_0_gfx_ring_init(struct amdgpu_device *adev, int ring_id,
+				   int me, int pipe, int queue)
+{
+	struct amdgpu_ring *ring;
+	unsigned int irq_type;
+	unsigned int hw_prio;
+
+	ring = &adev->gfx.gfx_ring[ring_id];
+
+	ring->me = me;
+	ring->pipe = pipe;
+	ring->queue = queue;
+
+	ring->ring_obj = NULL;
+	ring->use_doorbell = true;
+	if (adev->gfx.disable_kq) {
+		ring->no_scheduler = true;
+		ring->no_user_submission = true;
+	}
+
+	if (!ring_id)
+		ring->doorbell_index = adev->doorbell_index.gfx_ring0 << 1;
+	else
+		ring->doorbell_index = adev->doorbell_index.gfx_ring1 << 1;
+	ring->vm_hub = AMDGPU_GFXHUB(0);
+	sprintf(ring->name, "gfx_%d.%d.%d", ring->me, ring->pipe, ring->queue);
+
+	irq_type = AMDGPU_CP_IRQ_GFX_ME0_PIPE0_EOP + ring->pipe;
+	hw_prio = amdgpu_gfx_is_high_priority_graphics_queue(adev, ring) ?
+		AMDGPU_GFX_PIPE_PRIO_HIGH : AMDGPU_GFX_PIPE_PRIO_NORMAL;
+	return amdgpu_ring_init(adev, ring, 1024, &adev->gfx.eop_irq, irq_type,
+				hw_prio, NULL);
+}
+
+static int gfx_v11_0_compute_ring_init(struct amdgpu_device *adev, int ring_id,
+				       int mec, int pipe, int queue)
+{
+	int r;
+	unsigned irq_type;
+	struct amdgpu_ring *ring;
+	unsigned int hw_prio;
+
+	ring = &adev->gfx.compute_ring[ring_id];
+
+	/* mec0 is me1 */
+	ring->me = mec + 1;
+	ring->pipe = pipe;
+	ring->queue = queue;
+
+	ring->ring_obj = NULL;
+	ring->use_doorbell = true;
+	ring->doorbell_index = (adev->doorbell_index.mec_ring0 + ring_id) << 1;
+	ring->eop_gpu_addr = adev->gfx.mec.hpd_eop_gpu_addr
+				+ (ring_id * GFX11_MEC_HPD_SIZE);
+	ring->vm_hub = AMDGPU_GFXHUB(0);
+	sprintf(ring->name, "comp_%d.%d.%d", ring->me, ring->pipe, ring->queue);
+
+	irq_type = AMDGPU_CP_IRQ_COMPUTE_MEC1_PIPE0_EOP
+		+ ((ring->me - 1) * adev->gfx.mec.num_pipe_per_mec)
+		+ ring->pipe;
+	hw_prio = amdgpu_gfx_is_high_priority_compute_queue(adev, ring) ?
+			AMDGPU_GFX_PIPE_PRIO_HIGH : AMDGPU_GFX_PIPE_PRIO_NORMAL;
+	/* type-2 packets are deprecated on MEC, use type-3 instead */
+	r = amdgpu_ring_init(adev, ring, 1024, &adev->gfx.eop_irq, irq_type,
+			     hw_prio, NULL);
+	if (r)
+		return r;
+
+	return 0;
+}
+
+static struct {
+	SOC21_FIRMWARE_ID	id;
+	unsigned int		offset;
+	unsigned int		size;
+} rlc_autoload_info[SOC21_FIRMWARE_ID_MAX];
+
+static void gfx_v11_0_parse_rlc_toc(struct amdgpu_device *adev, void *rlc_toc)
+{
+	RLC_TABLE_OF_CONTENT *ucode = rlc_toc;
+
+	while (ucode && (ucode->id > SOC21_FIRMWARE_ID_INVALID) &&
+			(ucode->id < SOC21_FIRMWARE_ID_MAX)) {
+		rlc_autoload_info[ucode->id].id = ucode->id;
+		rlc_autoload_info[ucode->id].offset = ucode->offset * 4;
+		rlc_autoload_info[ucode->id].size = ucode->size * 4;
+
+		ucode++;
+	}
+}
+
+static uint32_t gfx_v11_0_calc_toc_total_size(struct amdgpu_device *adev)
+{
+	uint32_t total_size = 0;
+	SOC21_FIRMWARE_ID id;
+
+	gfx_v11_0_parse_rlc_toc(adev, adev->psp.toc.start_addr);
+
+	for (id = SOC21_FIRMWARE_ID_RLC_G_UCODE; id < SOC21_FIRMWARE_ID_MAX; id++)
+		total_size += rlc_autoload_info[id].size;
+
+	/* In case the offset in rlc toc ucode is aligned */
+	if (total_size < rlc_autoload_info[SOC21_FIRMWARE_ID_MAX-1].offset)
+		total_size = rlc_autoload_info[SOC21_FIRMWARE_ID_MAX-1].offset +
+			rlc_autoload_info[SOC21_FIRMWARE_ID_MAX-1].size;
+
+	return total_size;
+}
+
+static int gfx_v11_0_rlc_autoload_buffer_init(struct amdgpu_device *adev)
+{
+	int r;
+	uint32_t total_size;
+
+	total_size = gfx_v11_0_calc_toc_total_size(adev);
+
+	r = amdgpu_bo_create_reserved(adev, total_size, 64 * 1024,
+				      AMDGPU_GEM_DOMAIN_VRAM |
+				      AMDGPU_GEM_DOMAIN_GTT,
+				      &adev->gfx.rlc.rlc_autoload_bo,
+				      &adev->gfx.rlc.rlc_autoload_gpu_addr,
+				      (void **)&adev->gfx.rlc.rlc_autoload_ptr);
+
+	if (r) {
+		dev_err(adev->dev, "(%d) failed to create fw autoload bo\n", r);
+		return r;
+	}
+
+	return 0;
+}
+
+static void gfx_v11_0_rlc_backdoor_autoload_copy_ucode(struct amdgpu_device *adev,
+					      SOC21_FIRMWARE_ID id,
+			    		      const void *fw_data,
+					      uint32_t fw_size,
+					      uint32_t *fw_autoload_mask)
+{
+	uint32_t toc_offset;
+	uint32_t toc_fw_size;
+	char *ptr = adev->gfx.rlc.rlc_autoload_ptr;
+
+	if (id <= SOC21_FIRMWARE_ID_INVALID || id >= SOC21_FIRMWARE_ID_MAX)
+		return;
+
+	toc_offset = rlc_autoload_info[id].offset;
+	toc_fw_size = rlc_autoload_info[id].size;
+
+	if (fw_size == 0)
+		fw_size = toc_fw_size;
+
+	if (fw_size > toc_fw_size)
+		fw_size = toc_fw_size;
+
+	memcpy(ptr + toc_offset, fw_data, fw_size);
+
+	if (fw_size < toc_fw_size)
+		memset(ptr + toc_offset + fw_size, 0, toc_fw_size - fw_size);
+
+	if ((id != SOC21_FIRMWARE_ID_RS64_PFP) && (id != SOC21_FIRMWARE_ID_RS64_ME))
+		*(uint64_t *)fw_autoload_mask |= 1ULL << id;
+}
+
+static void gfx_v11_0_rlc_backdoor_autoload_copy_toc_ucode(struct amdgpu_device *adev,
+							uint32_t *fw_autoload_mask)
+{
+	void *data;
+	uint32_t size;
+	uint64_t *toc_ptr;
+
+	*(uint64_t *)fw_autoload_mask |= 0x1;
+
+	DRM_DEBUG("rlc autoload enabled fw: 0x%llx\n", *(uint64_t *)fw_autoload_mask);
+
+	data = adev->psp.toc.start_addr;
+	size = rlc_autoload_info[SOC21_FIRMWARE_ID_RLC_TOC].size;
+
+	toc_ptr = (uint64_t *)data + size / 8 - 1;
+	*toc_ptr = *(uint64_t *)fw_autoload_mask;
+
+	gfx_v11_0_rlc_backdoor_autoload_copy_ucode(adev, SOC21_FIRMWARE_ID_RLC_TOC,
+					data, size, fw_autoload_mask);
+}
+
+static void gfx_v11_0_rlc_backdoor_autoload_copy_gfx_ucode(struct amdgpu_device *adev,
+							uint32_t *fw_autoload_mask)
+{
+	const __le32 *fw_data;
+	uint32_t fw_size;
+	const struct gfx_firmware_header_v1_0 *cp_hdr;
+	const struct gfx_firmware_header_v2_0 *cpv2_hdr;
+	const struct rlc_firmware_header_v2_0 *rlc_hdr;
+	const struct rlc_firmware_header_v2_2 *rlcv22_hdr;
+	uint16_t version_major, version_minor;
+
+	if (adev->gfx.rs64_enable) {
+		/* pfp ucode */
+		cpv2_hdr = (const struct gfx_firmware_header_v2_0 *)
+			adev->gfx.pfp_fw->data;
+		/* instruction */
+		fw_data = (const __le32 *)(adev->gfx.pfp_fw->data +
+			le32_to_cpu(cpv2_hdr->ucode_offset_bytes));
+		fw_size = le32_to_cpu(cpv2_hdr->ucode_size_bytes);
+		gfx_v11_0_rlc_backdoor_autoload_copy_ucode(adev, SOC21_FIRMWARE_ID_RS64_PFP,
+						fw_data, fw_size, fw_autoload_mask);
+		/* data */
+		fw_data = (const __le32 *)(adev->gfx.pfp_fw->data +
+			le32_to_cpu(cpv2_hdr->data_offset_bytes));
+		fw_size = le32_to_cpu(cpv2_hdr->data_size_bytes);
+		gfx_v11_0_rlc_backdoor_autoload_copy_ucode(adev, SOC21_FIRMWARE_ID_RS64_PFP_P0_STACK,
+						fw_data, fw_size, fw_autoload_mask);
+		gfx_v11_0_rlc_backdoor_autoload_copy_ucode(adev, SOC21_FIRMWARE_ID_RS64_PFP_P1_STACK,
+						fw_data, fw_size, fw_autoload_mask);
+		/* me ucode */
+		cpv2_hdr = (const struct gfx_firmware_header_v2_0 *)
+			adev->gfx.me_fw->data;
+		/* instruction */
+		fw_data = (const __le32 *)(adev->gfx.me_fw->data +
+			le32_to_cpu(cpv2_hdr->ucode_offset_bytes));
+		fw_size = le32_to_cpu(cpv2_hdr->ucode_size_bytes);
+		gfx_v11_0_rlc_backdoor_autoload_copy_ucode(adev, SOC21_FIRMWARE_ID_RS64_ME,
+						fw_data, fw_size, fw_autoload_mask);
+		/* data */
+		fw_data = (const __le32 *)(adev->gfx.me_fw->data +
+			le32_to_cpu(cpv2_hdr->data_offset_bytes));
+		fw_size = le32_to_cpu(cpv2_hdr->data_size_bytes);
+		gfx_v11_0_rlc_backdoor_autoload_copy_ucode(adev, SOC21_FIRMWARE_ID_RS64_ME_P0_STACK,
+						fw_data, fw_size, fw_autoload_mask);
+		gfx_v11_0_rlc_backdoor_autoload_copy_ucode(adev, SOC21_FIRMWARE_ID_RS64_ME_P1_STACK,
+						fw_data, fw_size, fw_autoload_mask);
+		/* mec ucode */
+		cpv2_hdr = (const struct gfx_firmware_header_v2_0 *)
+			adev->gfx.mec_fw->data;
+		/* instruction */
+		fw_data = (const __le32 *) (adev->gfx.mec_fw->data +
+			le32_to_cpu(cpv2_hdr->ucode_offset_bytes));
+		fw_size = le32_to_cpu(cpv2_hdr->ucode_size_bytes);
+		gfx_v11_0_rlc_backdoor_autoload_copy_ucode(adev, SOC21_FIRMWARE_ID_RS64_MEC,
+						fw_data, fw_size, fw_autoload_mask);
+		/* data */
+		fw_data = (const __le32 *) (adev->gfx.mec_fw->data +
+			le32_to_cpu(cpv2_hdr->data_offset_bytes));
+		fw_size = le32_to_cpu(cpv2_hdr->data_size_bytes);
+		gfx_v11_0_rlc_backdoor_autoload_copy_ucode(adev, SOC21_FIRMWARE_ID_RS64_MEC_P0_STACK,
+						fw_data, fw_size, fw_autoload_mask);
+		gfx_v11_0_rlc_backdoor_autoload_copy_ucode(adev, SOC21_FIRMWARE_ID_RS64_MEC_P1_STACK,
+						fw_data, fw_size, fw_autoload_mask);
+		gfx_v11_0_rlc_backdoor_autoload_copy_ucode(adev, SOC21_FIRMWARE_ID_RS64_MEC_P2_STACK,
+						fw_data, fw_size, fw_autoload_mask);
+		gfx_v11_0_rlc_backdoor_autoload_copy_ucode(adev, SOC21_FIRMWARE_ID_RS64_MEC_P3_STACK,
+						fw_data, fw_size, fw_autoload_mask);
+	} else {
+		/* pfp ucode */
+		cp_hdr = (const struct gfx_firmware_header_v1_0 *)
+			adev->gfx.pfp_fw->data;
+		fw_data = (const __le32 *)(adev->gfx.pfp_fw->data +
+				le32_to_cpu(cp_hdr->header.ucode_array_offset_bytes));
+		fw_size = le32_to_cpu(cp_hdr->header.ucode_size_bytes);
+		gfx_v11_0_rlc_backdoor_autoload_copy_ucode(adev, SOC21_FIRMWARE_ID_CP_PFP,
+						fw_data, fw_size, fw_autoload_mask);
+
+		/* me ucode */
+		cp_hdr = (const struct gfx_firmware_header_v1_0 *)
+			adev->gfx.me_fw->data;
+		fw_data = (const __le32 *)(adev->gfx.me_fw->data +
+				le32_to_cpu(cp_hdr->header.ucode_array_offset_bytes));
+		fw_size = le32_to_cpu(cp_hdr->header.ucode_size_bytes);
+		gfx_v11_0_rlc_backdoor_autoload_copy_ucode(adev, SOC21_FIRMWARE_ID_CP_ME,
+						fw_data, fw_size, fw_autoload_mask);
+
+		/* mec ucode */
+		cp_hdr = (const struct gfx_firmware_header_v1_0 *)
+			adev->gfx.mec_fw->data;
+		fw_data = (const __le32 *) (adev->gfx.mec_fw->data +
+				le32_to_cpu(cp_hdr->header.ucode_array_offset_bytes));
+		fw_size = le32_to_cpu(cp_hdr->header.ucode_size_bytes) -
+			cp_hdr->jt_size * 4;
+		gfx_v11_0_rlc_backdoor_autoload_copy_ucode(adev, SOC21_FIRMWARE_ID_CP_MEC,
+						fw_data, fw_size, fw_autoload_mask);
+	}
+
+	/* rlc ucode */
+	rlc_hdr = (const struct rlc_firmware_header_v2_0 *)
+		adev->gfx.rlc_fw->data;
+	fw_data = (const __le32 *)(adev->gfx.rlc_fw->data +
+			le32_to_cpu(rlc_hdr->header.ucode_array_offset_bytes));
+	fw_size = le32_to_cpu(rlc_hdr->header.ucode_size_bytes);
+	gfx_v11_0_rlc_backdoor_autoload_copy_ucode(adev, SOC21_FIRMWARE_ID_RLC_G_UCODE,
+					fw_data, fw_size, fw_autoload_mask);
+
+	version_major = le16_to_cpu(rlc_hdr->header.header_version_major);
+	version_minor = le16_to_cpu(rlc_hdr->header.header_version_minor);
+	if (version_major == 2) {
+		if (version_minor >= 2) {
+			rlcv22_hdr = (const struct rlc_firmware_header_v2_2 *)adev->gfx.rlc_fw->data;
+
+			fw_data = (const __le32 *)(adev->gfx.rlc_fw->data +
+					le32_to_cpu(rlcv22_hdr->rlc_iram_ucode_offset_bytes));
+			fw_size = le32_to_cpu(rlcv22_hdr->rlc_iram_ucode_size_bytes);
+			gfx_v11_0_rlc_backdoor_autoload_copy_ucode(adev, SOC21_FIRMWARE_ID_RLX6_UCODE,
+					fw_data, fw_size, fw_autoload_mask);
+
+			fw_data = (const __le32 *)(adev->gfx.rlc_fw->data +
+					le32_to_cpu(rlcv22_hdr->rlc_dram_ucode_offset_bytes));
+			fw_size = le32_to_cpu(rlcv22_hdr->rlc_dram_ucode_size_bytes);
+			gfx_v11_0_rlc_backdoor_autoload_copy_ucode(adev, SOC21_FIRMWARE_ID_RLX6_DRAM_BOOT,
+					fw_data, fw_size, fw_autoload_mask);
+		}
+	}
+}
+
+static void gfx_v11_0_rlc_backdoor_autoload_copy_sdma_ucode(struct amdgpu_device *adev,
+							uint32_t *fw_autoload_mask)
+{
+	const __le32 *fw_data;
+	uint32_t fw_size;
+	const struct sdma_firmware_header_v2_0 *sdma_hdr;
+
+	sdma_hdr = (const struct sdma_firmware_header_v2_0 *)
+		adev->sdma.instance[0].fw->data;
+	fw_data = (const __le32 *) (adev->sdma.instance[0].fw->data +
+			le32_to_cpu(sdma_hdr->header.ucode_array_offset_bytes));
+	fw_size = le32_to_cpu(sdma_hdr->ctx_ucode_size_bytes);
+
+	gfx_v11_0_rlc_backdoor_autoload_copy_ucode(adev,
+			SOC21_FIRMWARE_ID_SDMA_UCODE_TH0, fw_data, fw_size, fw_autoload_mask);
+
+	fw_data = (const __le32 *) (adev->sdma.instance[0].fw->data +
+			le32_to_cpu(sdma_hdr->ctl_ucode_offset));
+	fw_size = le32_to_cpu(sdma_hdr->ctl_ucode_size_bytes);
+
+	gfx_v11_0_rlc_backdoor_autoload_copy_ucode(adev,
+			SOC21_FIRMWARE_ID_SDMA_UCODE_TH1, fw_data, fw_size, fw_autoload_mask);
+}
+
+static void gfx_v11_0_rlc_backdoor_autoload_copy_mes_ucode(struct amdgpu_device *adev,
+							uint32_t *fw_autoload_mask)
+{
+	const __le32 *fw_data;
+	unsigned fw_size;
+	const struct mes_firmware_header_v1_0 *mes_hdr;
+	int pipe, ucode_id, data_id;
+
+	for (pipe = 0; pipe < 2; pipe++) {
+		if (pipe==0) {
+			ucode_id = SOC21_FIRMWARE_ID_RS64_MES_P0;
+			data_id  = SOC21_FIRMWARE_ID_RS64_MES_P0_STACK;
+		} else {
+			ucode_id = SOC21_FIRMWARE_ID_RS64_MES_P1;
+			data_id  = SOC21_FIRMWARE_ID_RS64_MES_P1_STACK;
+		}
+
+		mes_hdr = (const struct mes_firmware_header_v1_0 *)
+			adev->mes.fw[pipe]->data;
+
+		fw_data = (const __le32 *)(adev->mes.fw[pipe]->data +
+				le32_to_cpu(mes_hdr->mes_ucode_offset_bytes));
+		fw_size = le32_to_cpu(mes_hdr->mes_ucode_size_bytes);
+
+		gfx_v11_0_rlc_backdoor_autoload_copy_ucode(adev,
+				ucode_id, fw_data, fw_size, fw_autoload_mask);
+
+		fw_data = (const __le32 *)(adev->mes.fw[pipe]->data +
+				le32_to_cpu(mes_hdr->mes_ucode_data_offset_bytes));
+		fw_size = le32_to_cpu(mes_hdr->mes_ucode_data_size_bytes);
+
+		gfx_v11_0_rlc_backdoor_autoload_copy_ucode(adev,
+				data_id, fw_data, fw_size, fw_autoload_mask);
+	}
+}
+
+static int gfx_v11_0_rlc_backdoor_autoload_enable(struct amdgpu_device *adev)
+{
+	uint32_t rlc_g_offset, rlc_g_size;
+	uint64_t gpu_addr;
+	uint32_t autoload_fw_id[2];
+
+	memset(autoload_fw_id, 0, sizeof(uint32_t) * 2);
+
+	/* RLC autoload sequence 2: copy ucode */
+	gfx_v11_0_rlc_backdoor_autoload_copy_sdma_ucode(adev, autoload_fw_id);
+	gfx_v11_0_rlc_backdoor_autoload_copy_gfx_ucode(adev, autoload_fw_id);
+	gfx_v11_0_rlc_backdoor_autoload_copy_mes_ucode(adev, autoload_fw_id);
+	gfx_v11_0_rlc_backdoor_autoload_copy_toc_ucode(adev, autoload_fw_id);
+
+	rlc_g_offset = rlc_autoload_info[SOC21_FIRMWARE_ID_RLC_G_UCODE].offset;
+	rlc_g_size = rlc_autoload_info[SOC21_FIRMWARE_ID_RLC_G_UCODE].size;
+	gpu_addr = adev->gfx.rlc.rlc_autoload_gpu_addr + rlc_g_offset;
+
+	WREG32_SOC15(GC, 0, regGFX_IMU_RLC_BOOTLOADER_ADDR_HI, upper_32_bits(gpu_addr));
+	WREG32_SOC15(GC, 0, regGFX_IMU_RLC_BOOTLOADER_ADDR_LO, lower_32_bits(gpu_addr));
+
+	WREG32_SOC15(GC, 0, regGFX_IMU_RLC_BOOTLOADER_SIZE, rlc_g_size);
+
+	/* RLC autoload sequence 3: load IMU fw */
+	if (adev->gfx.imu.funcs->load_microcode)
+		adev->gfx.imu.funcs->load_microcode(adev);
+	/* RLC autoload sequence 4 init IMU fw */
+	if (adev->gfx.imu.funcs->setup_imu)
+		adev->gfx.imu.funcs->setup_imu(adev);
+	if (adev->gfx.imu.funcs->start_imu)
+		adev->gfx.imu.funcs->start_imu(adev);
+
+	/* RLC autoload sequence 5 disable gpa mode */
+	gfx_v11_0_disable_gpa_mode(adev);
+
+	return 0;
+}
+
+static void gfx_v11_0_alloc_ip_dump(struct amdgpu_device *adev)
+{
+	uint32_t reg_count = ARRAY_SIZE(gc_reg_list_11_0);
+	uint32_t *ptr;
+	uint32_t inst;
+
+	ptr = kcalloc(reg_count, sizeof(uint32_t), GFP_KERNEL);
+	if (!ptr) {
+		DRM_ERROR("Failed to allocate memory for GFX IP Dump\n");
+		adev->gfx.ip_dump_core = NULL;
+	} else {
+		adev->gfx.ip_dump_core = ptr;
+	}
+
+	/* Allocate memory for compute queue registers for all the instances */
+	reg_count = ARRAY_SIZE(gc_cp_reg_list_11);
+	inst = adev->gfx.mec.num_mec * adev->gfx.mec.num_pipe_per_mec *
+		adev->gfx.mec.num_queue_per_pipe;
+
+	ptr = kcalloc(reg_count * inst, sizeof(uint32_t), GFP_KERNEL);
+	if (!ptr) {
+		DRM_ERROR("Failed to allocate memory for Compute Queues IP Dump\n");
+		adev->gfx.ip_dump_compute_queues = NULL;
+	} else {
+		adev->gfx.ip_dump_compute_queues = ptr;
+	}
+
+	/* Allocate memory for gfx queue registers for all the instances */
+	reg_count = ARRAY_SIZE(gc_gfx_queue_reg_list_11);
+	inst = adev->gfx.me.num_me * adev->gfx.me.num_pipe_per_me *
+		adev->gfx.me.num_queue_per_pipe;
+
+	ptr = kcalloc(reg_count * inst, sizeof(uint32_t), GFP_KERNEL);
+	if (!ptr) {
+		DRM_ERROR("Failed to allocate memory for GFX Queues IP Dump\n");
+		adev->gfx.ip_dump_gfx_queues = NULL;
+	} else {
+		adev->gfx.ip_dump_gfx_queues = ptr;
+	}
+}
+
+static int gfx_v11_0_sw_init(struct amdgpu_ip_block *ip_block)
+{
+	int i, j, k, r, ring_id;
+	int xcc_id = 0;
+	struct amdgpu_device *adev = ip_block->adev;
+	int num_queue_per_pipe = 1; /* we only enable 1 KGQ per pipe */
+
+	INIT_DELAYED_WORK(&adev->gfx.idle_work, amdgpu_gfx_profile_idle_work_handler);
+
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	case IP_VERSION(11, 0, 0):
+	case IP_VERSION(11, 0, 1):
+	case IP_VERSION(11, 0, 2):
+	case IP_VERSION(11, 0, 3):
+	case IP_VERSION(11, 0, 4):
+	case IP_VERSION(11, 5, 0):
+	case IP_VERSION(11, 5, 1):
+	case IP_VERSION(11, 5, 2):
+	case IP_VERSION(11, 5, 3):
+		adev->gfx.me.num_me = 1;
+		adev->gfx.me.num_pipe_per_me = 1;
+		adev->gfx.me.num_queue_per_pipe = 2;
+		adev->gfx.mec.num_mec = 1;
+		adev->gfx.mec.num_pipe_per_mec = 4;
+		adev->gfx.mec.num_queue_per_pipe = 4;
+		break;
+	default:
+		adev->gfx.me.num_me = 1;
+		adev->gfx.me.num_pipe_per_me = 1;
+		adev->gfx.me.num_queue_per_pipe = 1;
+		adev->gfx.mec.num_mec = 1;
+		adev->gfx.mec.num_pipe_per_mec = 4;
+		adev->gfx.mec.num_queue_per_pipe = 8;
+		break;
+	}
+
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	case IP_VERSION(11, 0, 0):
+	case IP_VERSION(11, 0, 2):
+	case IP_VERSION(11, 0, 3):
+		if (!adev->gfx.disable_uq &&
+		    adev->gfx.me_fw_version  >= 2420 &&
+		    adev->gfx.pfp_fw_version >= 2580 &&
+		    adev->gfx.mec_fw_version >= 2650 &&
+		    adev->mes.fw_version[0] >= 120) {
+			adev->userq_funcs[AMDGPU_HW_IP_GFX] = &userq_mes_funcs;
+			adev->userq_funcs[AMDGPU_HW_IP_COMPUTE] = &userq_mes_funcs;
+		}
+		break;
+	case IP_VERSION(11, 0, 1):
+	case IP_VERSION(11, 0, 4):
+	case IP_VERSION(11, 5, 0):
+	case IP_VERSION(11, 5, 1):
+	case IP_VERSION(11, 5, 2):
+	case IP_VERSION(11, 5, 3):
+		/* add firmware version checks here */
+		if (0 && !adev->gfx.disable_uq) {
+			adev->userq_funcs[AMDGPU_HW_IP_GFX] = &userq_mes_funcs;
+			adev->userq_funcs[AMDGPU_HW_IP_COMPUTE] = &userq_mes_funcs;
+		}
+		break;
+	default:
+		break;
+	}
+
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	case IP_VERSION(11, 0, 0):
+	case IP_VERSION(11, 0, 2):
+	case IP_VERSION(11, 0, 3):
+		adev->gfx.cleaner_shader_ptr = gfx_11_0_3_cleaner_shader_hex;
+		adev->gfx.cleaner_shader_size = sizeof(gfx_11_0_3_cleaner_shader_hex);
+		if (adev->gfx.me_fw_version  >= 2280 &&
+		    adev->gfx.pfp_fw_version >= 2370 &&
+		    adev->gfx.mec_fw_version >= 2450  &&
+		    adev->mes.fw_version[0] >= 99) {
+			adev->gfx.enable_cleaner_shader = true;
+			r = amdgpu_gfx_cleaner_shader_sw_init(adev, adev->gfx.cleaner_shader_size);
+			if (r) {
+				adev->gfx.enable_cleaner_shader = false;
+				dev_err(adev->dev, "Failed to initialize cleaner shader\n");
+			}
+		}
+		break;
+	case IP_VERSION(11, 0, 1):
+	case IP_VERSION(11, 0, 4):
+		adev->gfx.cleaner_shader_ptr = gfx_11_0_3_cleaner_shader_hex;
+		adev->gfx.cleaner_shader_size = sizeof(gfx_11_0_3_cleaner_shader_hex);
+		if (adev->gfx.pfp_fw_version >= 102 &&
+		    adev->gfx.mec_fw_version >= 66 &&
+		    adev->mes.fw_version[0] >= 128) {
+			adev->gfx.enable_cleaner_shader = true;
+			r = amdgpu_gfx_cleaner_shader_sw_init(adev, adev->gfx.cleaner_shader_size);
+			if (r) {
+				adev->gfx.enable_cleaner_shader = false;
+				dev_err(adev->dev, "Failed to initialize cleaner shader\n");
+			}
+		}
+		break;
+	case IP_VERSION(11, 5, 0):
+	case IP_VERSION(11, 5, 1):
+		adev->gfx.cleaner_shader_ptr = gfx_11_0_3_cleaner_shader_hex;
+		adev->gfx.cleaner_shader_size = sizeof(gfx_11_0_3_cleaner_shader_hex);
+		if (adev->gfx.mec_fw_version >= 26 &&
+		    adev->mes.fw_version[0] >= 114) {
+			adev->gfx.enable_cleaner_shader = true;
+			r = amdgpu_gfx_cleaner_shader_sw_init(adev, adev->gfx.cleaner_shader_size);
+			if (r) {
+				adev->gfx.enable_cleaner_shader = false;
+				dev_err(adev->dev, "Failed to initialize cleaner shader\n");
+			}
+		}
+		break;
+	case IP_VERSION(11, 5, 2):
+		adev->gfx.cleaner_shader_ptr = gfx_11_0_3_cleaner_shader_hex;
+		adev->gfx.cleaner_shader_size = sizeof(gfx_11_0_3_cleaner_shader_hex);
+		if (adev->gfx.me_fw_version  >= 12 &&
+		    adev->gfx.pfp_fw_version >= 15 &&
+		    adev->gfx.mec_fw_version >= 15) {
+			adev->gfx.enable_cleaner_shader = true;
+			r = amdgpu_gfx_cleaner_shader_sw_init(adev, adev->gfx.cleaner_shader_size);
+			if (r) {
+				adev->gfx.enable_cleaner_shader = false;
+				dev_err(adev->dev, "Failed to initialize cleaner shader\n");
+			}
+		}
+		break;
+	case IP_VERSION(11, 5, 3):
+		adev->gfx.cleaner_shader_ptr = gfx_11_0_3_cleaner_shader_hex;
+		adev->gfx.cleaner_shader_size = sizeof(gfx_11_0_3_cleaner_shader_hex);
+		if (adev->gfx.me_fw_version  >= 7 &&
+		    adev->gfx.pfp_fw_version >= 8 &&
+		    adev->gfx.mec_fw_version >= 8) {
+			adev->gfx.enable_cleaner_shader = true;
+			r = amdgpu_gfx_cleaner_shader_sw_init(adev, adev->gfx.cleaner_shader_size);
+			if (r) {
+				adev->gfx.enable_cleaner_shader = false;
+				dev_err(adev->dev, "Failed to initialize cleaner shader\n");
+			}
+		}
+		break;
+	default:
+		adev->gfx.enable_cleaner_shader = false;
+		break;
+	}
+
+	/* Enable CG flag in one VF mode for enabling RLC safe mode enter/exit */
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(11, 0, 3) &&
+	    amdgpu_sriov_is_pp_one_vf(adev))
+		adev->cg_flags = AMD_CG_SUPPORT_GFX_CGCG;
+
+	/* EOP Event */
+	r = amdgpu_irq_add_id(adev, SOC21_IH_CLIENTID_GRBM_CP,
+			      GFX_11_0_0__SRCID__CP_EOP_INTERRUPT,
+			      &adev->gfx.eop_irq);
+	if (r)
+		return r;
+
+	/* Bad opcode Event */
+	r = amdgpu_irq_add_id(adev, SOC21_IH_CLIENTID_GRBM_CP,
+			      GFX_11_0_0__SRCID__CP_BAD_OPCODE_ERROR,
+			      &adev->gfx.bad_op_irq);
+	if (r)
+		return r;
+
+	/* Privileged reg */
+	r = amdgpu_irq_add_id(adev, SOC21_IH_CLIENTID_GRBM_CP,
+			      GFX_11_0_0__SRCID__CP_PRIV_REG_FAULT,
+			      &adev->gfx.priv_reg_irq);
+	if (r)
+		return r;
+
+	/* Privileged inst */
+	r = amdgpu_irq_add_id(adev, SOC21_IH_CLIENTID_GRBM_CP,
+			      GFX_11_0_0__SRCID__CP_PRIV_INSTR_FAULT,
+			      &adev->gfx.priv_inst_irq);
+	if (r)
+		return r;
+
+	/* FED error */
+	r = amdgpu_irq_add_id(adev, SOC21_IH_CLIENTID_GFX,
+				  GFX_11_0_0__SRCID__RLC_GC_FED_INTERRUPT,
+				  &adev->gfx.rlc_gc_fed_irq);
+	if (r)
+		return r;
+
+	adev->gfx.gfx_current_status = AMDGPU_GFX_NORMAL_MODE;
+
+	gfx_v11_0_me_init(adev);
+
+	r = gfx_v11_0_rlc_init(adev);
+	if (r) {
+		DRM_ERROR("Failed to init rlc BOs!\n");
+		return r;
+	}
+
+	r = gfx_v11_0_mec_init(adev);
+	if (r) {
+		DRM_ERROR("Failed to init MEC BOs!\n");
+		return r;
+	}
+
+	if (adev->gfx.num_gfx_rings) {
+		ring_id = 0;
+		/* set up the gfx ring */
+		for (i = 0; i < adev->gfx.me.num_me; i++) {
+			for (j = 0; j < num_queue_per_pipe; j++) {
+				for (k = 0; k < adev->gfx.me.num_pipe_per_me; k++) {
+					if (!amdgpu_gfx_is_me_queue_enabled(adev, i, k, j))
+						continue;
+
+					r = gfx_v11_0_gfx_ring_init(adev, ring_id,
+								    i, k, j);
+					if (r)
+						return r;
+					ring_id++;
+				}
+			}
+		}
+	}
+
+	if (adev->gfx.num_compute_rings) {
+		ring_id = 0;
+		/* set up the compute queues - allocate horizontally across pipes */
+		for (i = 0; i < adev->gfx.mec.num_mec; ++i) {
+			for (j = 0; j < adev->gfx.mec.num_queue_per_pipe; j++) {
+				for (k = 0; k < adev->gfx.mec.num_pipe_per_mec; k++) {
+					if (!amdgpu_gfx_is_mec_queue_enabled(adev, 0, i,
+									     k, j))
+						continue;
+
+					r = gfx_v11_0_compute_ring_init(adev, ring_id,
+									i, k, j);
+					if (r)
+						return r;
+
+					ring_id++;
+				}
+			}
+		}
+	}
+
+	adev->gfx.gfx_supported_reset =
+		amdgpu_get_soft_full_reset_mask(&adev->gfx.gfx_ring[0]);
+	adev->gfx.compute_supported_reset =
+		amdgpu_get_soft_full_reset_mask(&adev->gfx.compute_ring[0]);
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	case IP_VERSION(11, 0, 0):
+	case IP_VERSION(11, 0, 2):
+	case IP_VERSION(11, 0, 3):
+		if ((adev->gfx.me_fw_version >= 2280) &&
+		    (adev->gfx.mec_fw_version >= 2410) &&
+		    !amdgpu_sriov_vf(adev) &&
+		    !adev->debug_disable_gpu_ring_reset) {
+			adev->gfx.compute_supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE;
+			adev->gfx.gfx_supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE;
+		}
+		break;
+	default:
+		if (!amdgpu_sriov_vf(adev) &&
+		    !adev->debug_disable_gpu_ring_reset) {
+			adev->gfx.compute_supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE;
+			adev->gfx.gfx_supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE;
+		}
+		break;
+	}
+
+	if (!adev->enable_mes_kiq) {
+		r = amdgpu_gfx_kiq_init(adev, GFX11_MEC_HPD_SIZE, 0);
+		if (r) {
+			DRM_ERROR("Failed to init KIQ BOs!\n");
+			return r;
+		}
+
+		r = amdgpu_gfx_kiq_init_ring(adev, xcc_id);
+		if (r)
+			return r;
+	}
+
+	r = amdgpu_gfx_mqd_sw_init(adev, sizeof(struct v11_compute_mqd), 0);
+	if (r)
+		return r;
+
+	/* allocate visible FB for rlc auto-loading fw */
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_RLC_BACKDOOR_AUTO) {
+		r = gfx_v11_0_rlc_autoload_buffer_init(adev);
+		if (r)
+			return r;
+	}
+
+	r = gfx_v11_0_gpu_early_init(adev);
+	if (r)
+		return r;
+
+	if (amdgpu_gfx_ras_sw_init(adev)) {
+		dev_err(adev->dev, "Failed to initialize gfx ras block!\n");
+		return -EINVAL;
+	}
+
+	gfx_v11_0_alloc_ip_dump(adev);
+
+	r = amdgpu_gfx_sysfs_init(adev);
+	if (r)
+		return r;
+
+	return 0;
+}
+
+static void gfx_v11_0_pfp_fini(struct amdgpu_device *adev)
+{
+	amdgpu_bo_free_kernel(&adev->gfx.pfp.pfp_fw_obj,
+			      &adev->gfx.pfp.pfp_fw_gpu_addr,
+			      (void **)&adev->gfx.pfp.pfp_fw_ptr);
+
+	amdgpu_bo_free_kernel(&adev->gfx.pfp.pfp_fw_data_obj,
+			      &adev->gfx.pfp.pfp_fw_data_gpu_addr,
+			      (void **)&adev->gfx.pfp.pfp_fw_data_ptr);
+}
+
+static void gfx_v11_0_me_fini(struct amdgpu_device *adev)
+{
+	amdgpu_bo_free_kernel(&adev->gfx.me.me_fw_obj,
+			      &adev->gfx.me.me_fw_gpu_addr,
+			      (void **)&adev->gfx.me.me_fw_ptr);
+
+	amdgpu_bo_free_kernel(&adev->gfx.me.me_fw_data_obj,
+			       &adev->gfx.me.me_fw_data_gpu_addr,
+			       (void **)&adev->gfx.me.me_fw_data_ptr);
+}
+
+static void gfx_v11_0_rlc_autoload_buffer_fini(struct amdgpu_device *adev)
+{
+	amdgpu_bo_free_kernel(&adev->gfx.rlc.rlc_autoload_bo,
+			&adev->gfx.rlc.rlc_autoload_gpu_addr,
+			(void **)&adev->gfx.rlc.rlc_autoload_ptr);
+}
+
+static int gfx_v11_0_sw_fini(struct amdgpu_ip_block *ip_block)
+{
+	int i;
+	struct amdgpu_device *adev = ip_block->adev;
+
+	for (i = 0; i < adev->gfx.num_gfx_rings; i++)
+		amdgpu_ring_fini(&adev->gfx.gfx_ring[i]);
+	for (i = 0; i < adev->gfx.num_compute_rings; i++)
+		amdgpu_ring_fini(&adev->gfx.compute_ring[i]);
+
+	amdgpu_gfx_mqd_sw_fini(adev, 0);
+
+	if (!adev->enable_mes_kiq) {
+		amdgpu_gfx_kiq_free_ring(&adev->gfx.kiq[0].ring);
+		amdgpu_gfx_kiq_fini(adev, 0);
+	}
+
+	amdgpu_gfx_cleaner_shader_sw_fini(adev);
+
+	gfx_v11_0_pfp_fini(adev);
+	gfx_v11_0_me_fini(adev);
+	gfx_v11_0_rlc_fini(adev);
+	gfx_v11_0_mec_fini(adev);
+
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_RLC_BACKDOOR_AUTO)
+		gfx_v11_0_rlc_autoload_buffer_fini(adev);
+
+	gfx_v11_0_free_microcode(adev);
+
+	amdgpu_gfx_sysfs_fini(adev);
+
+	kfree(adev->gfx.ip_dump_core);
+	kfree(adev->gfx.ip_dump_compute_queues);
+	kfree(adev->gfx.ip_dump_gfx_queues);
+
+	return 0;
+}
+
+static void gfx_v11_0_select_se_sh(struct amdgpu_device *adev, u32 se_num,
+				   u32 sh_num, u32 instance, int xcc_id)
+{
+	u32 data;
+
+	if (instance == 0xffffffff)
+		data = REG_SET_FIELD(0, GRBM_GFX_INDEX,
+				     INSTANCE_BROADCAST_WRITES, 1);
+	else
+		data = REG_SET_FIELD(0, GRBM_GFX_INDEX, INSTANCE_INDEX,
+				     instance);
+
+	if (se_num == 0xffffffff)
+		data = REG_SET_FIELD(data, GRBM_GFX_INDEX, SE_BROADCAST_WRITES,
+				     1);
+	else
+		data = REG_SET_FIELD(data, GRBM_GFX_INDEX, SE_INDEX, se_num);
+
+	if (sh_num == 0xffffffff)
+		data = REG_SET_FIELD(data, GRBM_GFX_INDEX, SA_BROADCAST_WRITES,
+				     1);
+	else
+		data = REG_SET_FIELD(data, GRBM_GFX_INDEX, SA_INDEX, sh_num);
+
+	WREG32_SOC15(GC, 0, regGRBM_GFX_INDEX, data);
+}
+
+static u32 gfx_v11_0_get_sa_active_bitmap(struct amdgpu_device *adev)
+{
+	u32 gc_disabled_sa_mask, gc_user_disabled_sa_mask, sa_mask;
+
+	gc_disabled_sa_mask = RREG32_SOC15(GC, 0, regCC_GC_SA_UNIT_DISABLE);
+	gc_disabled_sa_mask = REG_GET_FIELD(gc_disabled_sa_mask,
+					   CC_GC_SA_UNIT_DISABLE,
+					   SA_DISABLE);
+	gc_user_disabled_sa_mask = RREG32_SOC15(GC, 0, regGC_USER_SA_UNIT_DISABLE);
+	gc_user_disabled_sa_mask = REG_GET_FIELD(gc_user_disabled_sa_mask,
+						 GC_USER_SA_UNIT_DISABLE,
+						 SA_DISABLE);
+	sa_mask = amdgpu_gfx_create_bitmask(adev->gfx.config.max_sh_per_se *
+					    adev->gfx.config.max_shader_engines);
+
+	return sa_mask & (~(gc_disabled_sa_mask | gc_user_disabled_sa_mask));
+}
+
+static u32 gfx_v11_0_get_rb_active_bitmap(struct amdgpu_device *adev)
+{
+	u32 gc_disabled_rb_mask, gc_user_disabled_rb_mask;
+	u32 rb_mask;
+
+	gc_disabled_rb_mask = RREG32_SOC15(GC, 0, regCC_RB_BACKEND_DISABLE);
+	gc_disabled_rb_mask = REG_GET_FIELD(gc_disabled_rb_mask,
+					    CC_RB_BACKEND_DISABLE,
+					    BACKEND_DISABLE);
+	gc_user_disabled_rb_mask = RREG32_SOC15(GC, 0, regGC_USER_RB_BACKEND_DISABLE);
+	gc_user_disabled_rb_mask = REG_GET_FIELD(gc_user_disabled_rb_mask,
+						 GC_USER_RB_BACKEND_DISABLE,
+						 BACKEND_DISABLE);
+	rb_mask = amdgpu_gfx_create_bitmask(adev->gfx.config.max_backends_per_se *
+					    adev->gfx.config.max_shader_engines);
+
+	return rb_mask & (~(gc_disabled_rb_mask | gc_user_disabled_rb_mask));
+}
+
+static void gfx_v11_0_setup_rb(struct amdgpu_device *adev)
+{
+	u32 rb_bitmap_per_sa;
+	u32 rb_bitmap_width_per_sa;
+	u32 max_sa;
+	u32 active_sa_bitmap;
+	u32 global_active_rb_bitmap;
+	u32 active_rb_bitmap = 0;
+	u32 i;
+
+	/* query sa bitmap from SA_UNIT_DISABLE registers */
+	active_sa_bitmap = gfx_v11_0_get_sa_active_bitmap(adev);
+	/* query rb bitmap from RB_BACKEND_DISABLE registers */
+	global_active_rb_bitmap = gfx_v11_0_get_rb_active_bitmap(adev);
+
+	/* generate active rb bitmap according to active sa bitmap */
+	max_sa = adev->gfx.config.max_shader_engines *
+		 adev->gfx.config.max_sh_per_se;
+	rb_bitmap_width_per_sa = adev->gfx.config.max_backends_per_se /
+				 adev->gfx.config.max_sh_per_se;
+	rb_bitmap_per_sa = amdgpu_gfx_create_bitmask(rb_bitmap_width_per_sa);
+
+	for (i = 0; i < max_sa; i++) {
+		if (active_sa_bitmap & (1 << i))
+			active_rb_bitmap |= (rb_bitmap_per_sa << (i * rb_bitmap_width_per_sa));
+	}
+
+	active_rb_bitmap &= global_active_rb_bitmap;
+	adev->gfx.config.backend_enable_mask = active_rb_bitmap;
+	adev->gfx.config.num_rbs = hweight32(active_rb_bitmap);
+}
+
+#define DEFAULT_SH_MEM_BASES	(0x6000)
+#define LDS_APP_BASE           0x1
+#define SCRATCH_APP_BASE       0x2
+
+static void gfx_v11_0_init_compute_vmid(struct amdgpu_device *adev)
+{
+	int i;
+	uint32_t sh_mem_bases;
+	uint32_t data;
+
+	/*
+	 * Configure apertures:
+	 * LDS:         0x60000000'00000000 - 0x60000001'00000000 (4GB)
+	 * Scratch:     0x60000001'00000000 - 0x60000002'00000000 (4GB)
+	 * GPUVM:       0x60010000'00000000 - 0x60020000'00000000 (1TB)
+	 */
+	sh_mem_bases = (LDS_APP_BASE << SH_MEM_BASES__SHARED_BASE__SHIFT) |
+			SCRATCH_APP_BASE;
+
+	mutex_lock(&adev->srbm_mutex);
+	for (i = adev->vm_manager.first_kfd_vmid; i < AMDGPU_NUM_VMID; i++) {
+		soc21_grbm_select(adev, 0, 0, 0, i);
+		/* CP and shaders */
+		WREG32_SOC15(GC, 0, regSH_MEM_CONFIG, DEFAULT_SH_MEM_CONFIG);
+		WREG32_SOC15(GC, 0, regSH_MEM_BASES, sh_mem_bases);
+
+		/* Enable trap for each kfd vmid. */
+		data = RREG32_SOC15(GC, 0, regSPI_GDBG_PER_VMID_CNTL);
+		data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, TRAP_EN, 1);
+		WREG32_SOC15(GC, 0, regSPI_GDBG_PER_VMID_CNTL, data);
+	}
+	soc21_grbm_select(adev, 0, 0, 0, 0);
+	mutex_unlock(&adev->srbm_mutex);
+
+	/*
+	 * Initialize all compute VMIDs to have no GDS, GWS, or OA
+	 * access. These should be enabled by FW for target VMIDs.
+	 */
+	for (i = adev->vm_manager.first_kfd_vmid; i < AMDGPU_NUM_VMID; i++) {
+		WREG32_SOC15_OFFSET(GC, 0, regGDS_VMID0_BASE, 2 * i, 0);
+		WREG32_SOC15_OFFSET(GC, 0, regGDS_VMID0_SIZE, 2 * i, 0);
+		WREG32_SOC15_OFFSET(GC, 0, regGDS_GWS_VMID0, i, 0);
+		WREG32_SOC15_OFFSET(GC, 0, regGDS_OA_VMID0, i, 0);
+	}
+}
+
+static void gfx_v11_0_init_gds_vmid(struct amdgpu_device *adev)
+{
+	int vmid;
+
+	/*
+	 * Initialize all compute and user-gfx VMIDs to have no GDS, GWS, or OA
+	 * access. Compute VMIDs should be enabled by FW for target VMIDs,
+	 * the driver can enable them for graphics. VMID0 should maintain
+	 * access so that HWS firmware can save/restore entries.
+	 */
+	for (vmid = 1; vmid < 16; vmid++) {
+		WREG32_SOC15_OFFSET(GC, 0, regGDS_VMID0_BASE, 2 * vmid, 0);
+		WREG32_SOC15_OFFSET(GC, 0, regGDS_VMID0_SIZE, 2 * vmid, 0);
+		WREG32_SOC15_OFFSET(GC, 0, regGDS_GWS_VMID0, vmid, 0);
+		WREG32_SOC15_OFFSET(GC, 0, regGDS_OA_VMID0, vmid, 0);
+	}
+}
+
+static void gfx_v11_0_tcp_harvest(struct amdgpu_device *adev)
+{
+	/* TODO: harvest feature to be added later. */
+}
+
+static void gfx_v11_0_get_tcc_info(struct amdgpu_device *adev)
+{
+	/* TCCs are global (not instanced). */
+	uint32_t tcc_disable = RREG32_SOC15(GC, 0, regCGTS_TCC_DISABLE) |
+			       RREG32_SOC15(GC, 0, regCGTS_USER_TCC_DISABLE);
+
+	adev->gfx.config.tcc_disabled_mask =
+		REG_GET_FIELD(tcc_disable, CGTS_TCC_DISABLE, TCC_DISABLE) |
+		(REG_GET_FIELD(tcc_disable, CGTS_TCC_DISABLE, HI_TCC_DISABLE) << 16);
+}
+
+static void gfx_v11_0_constants_init(struct amdgpu_device *adev)
+{
+	u32 tmp;
+	int i;
+
+	if (!amdgpu_sriov_vf(adev))
+		WREG32_FIELD15_PREREG(GC, 0, GRBM_CNTL, READ_TIMEOUT, 0xff);
+
+	gfx_v11_0_setup_rb(adev);
+	gfx_v11_0_get_cu_info(adev, &adev->gfx.cu_info);
+	gfx_v11_0_get_tcc_info(adev);
+	adev->gfx.config.pa_sc_tile_steering_override = 0;
+
+	/* Set whether texture coordinate truncation is conformant. */
+	tmp = RREG32_SOC15(GC, 0, regTA_CNTL2);
+	adev->gfx.config.ta_cntl2_truncate_coord_mode =
+		REG_GET_FIELD(tmp, TA_CNTL2, TRUNCATE_COORD_MODE);
+
+	/* XXX SH_MEM regs */
+	/* where to put LDS, scratch, GPUVM in FSA64 space */
+	mutex_lock(&adev->srbm_mutex);
+	for (i = 0; i < adev->vm_manager.id_mgr[AMDGPU_GFXHUB(0)].num_ids; i++) {
+		soc21_grbm_select(adev, 0, 0, 0, i);
+		/* CP and shaders */
+		WREG32_SOC15(GC, 0, regSH_MEM_CONFIG, DEFAULT_SH_MEM_CONFIG);
+		if (i != 0) {
+			tmp = REG_SET_FIELD(0, SH_MEM_BASES, PRIVATE_BASE,
+				(adev->gmc.private_aperture_start >> 48));
+			tmp = REG_SET_FIELD(tmp, SH_MEM_BASES, SHARED_BASE,
+				(adev->gmc.shared_aperture_start >> 48));
+			WREG32_SOC15(GC, 0, regSH_MEM_BASES, tmp);
+		}
+	}
+	soc21_grbm_select(adev, 0, 0, 0, 0);
+
+	mutex_unlock(&adev->srbm_mutex);
+
+	gfx_v11_0_init_compute_vmid(adev);
+	gfx_v11_0_init_gds_vmid(adev);
+}
+
+static u32 gfx_v11_0_get_cpg_int_cntl(struct amdgpu_device *adev,
+				      int me, int pipe)
+{
+	if (me != 0)
+		return 0;
+
+	switch (pipe) {
+	case 0:
+		return SOC15_REG_OFFSET(GC, 0, regCP_INT_CNTL_RING0);
+	case 1:
+		return SOC15_REG_OFFSET(GC, 0, regCP_INT_CNTL_RING1);
+	default:
+		return 0;
+	}
+}
+
+static u32 gfx_v11_0_get_cpc_int_cntl(struct amdgpu_device *adev,
+				      int me, int pipe)
+{
+	/*
+	 * amdgpu controls only the first MEC. That's why this function only
+	 * handles the setting of interrupts for this specific MEC. All other
+	 * pipes' interrupts are set by amdkfd.
+	 */
+	if (me != 1)
+		return 0;
+
+	switch (pipe) {
+	case 0:
+		return SOC15_REG_OFFSET(GC, 0, regCP_ME1_PIPE0_INT_CNTL);
+	case 1:
+		return SOC15_REG_OFFSET(GC, 0, regCP_ME1_PIPE1_INT_CNTL);
+	case 2:
+		return SOC15_REG_OFFSET(GC, 0, regCP_ME1_PIPE2_INT_CNTL);
+	case 3:
+		return SOC15_REG_OFFSET(GC, 0, regCP_ME1_PIPE3_INT_CNTL);
+	default:
+		return 0;
+	}
+}
+
+static void gfx_v11_0_enable_gui_idle_interrupt(struct amdgpu_device *adev,
+					       bool enable)
+{
+	u32 tmp, cp_int_cntl_reg;
+	int i, j;
+
+	if (amdgpu_sriov_vf(adev))
+		return;
+
+	for (i = 0; i < adev->gfx.me.num_me; i++) {
+		for (j = 0; j < adev->gfx.me.num_pipe_per_me; j++) {
+			cp_int_cntl_reg = gfx_v11_0_get_cpg_int_cntl(adev, i, j);
+
+			if (cp_int_cntl_reg) {
+				tmp = RREG32_SOC15_IP(GC, cp_int_cntl_reg);
+				tmp = REG_SET_FIELD(tmp, CP_INT_CNTL_RING0, CNTX_BUSY_INT_ENABLE,
+						    enable ? 1 : 0);
+				tmp = REG_SET_FIELD(tmp, CP_INT_CNTL_RING0, CNTX_EMPTY_INT_ENABLE,
+						    enable ? 1 : 0);
+				tmp = REG_SET_FIELD(tmp, CP_INT_CNTL_RING0, CMP_BUSY_INT_ENABLE,
+						    enable ? 1 : 0);
+				tmp = REG_SET_FIELD(tmp, CP_INT_CNTL_RING0, GFX_IDLE_INT_ENABLE,
+						    enable ? 1 : 0);
+				WREG32_SOC15_IP(GC, cp_int_cntl_reg, tmp);
+			}
+		}
+	}
+}
+
+static int gfx_v11_0_init_csb(struct amdgpu_device *adev)
+{
+	adev->gfx.rlc.funcs->get_csb_buffer(adev, adev->gfx.rlc.cs_ptr);
+
+	WREG32_SOC15(GC, 0, regRLC_CSIB_ADDR_HI,
+			adev->gfx.rlc.clear_state_gpu_addr >> 32);
+	WREG32_SOC15(GC, 0, regRLC_CSIB_ADDR_LO,
+			adev->gfx.rlc.clear_state_gpu_addr & 0xfffffffc);
+	WREG32_SOC15(GC, 0, regRLC_CSIB_LENGTH, adev->gfx.rlc.clear_state_size);
+
+	return 0;
+}
+
+static void gfx_v11_0_rlc_stop(struct amdgpu_device *adev)
+{
+	u32 tmp = RREG32_SOC15(GC, 0, regRLC_CNTL);
+
+	tmp = REG_SET_FIELD(tmp, RLC_CNTL, RLC_ENABLE_F32, 0);
+	WREG32_SOC15(GC, 0, regRLC_CNTL, tmp);
+}
+
+static void gfx_v11_0_rlc_reset(struct amdgpu_device *adev)
+{
+	WREG32_FIELD15_PREREG(GC, 0, GRBM_SOFT_RESET, SOFT_RESET_RLC, 1);
+	udelay(50);
+	WREG32_FIELD15_PREREG(GC, 0, GRBM_SOFT_RESET, SOFT_RESET_RLC, 0);
+	udelay(50);
+}
+
+static void gfx_v11_0_rlc_smu_handshake_cntl(struct amdgpu_device *adev,
+					     bool enable)
+{
+	uint32_t rlc_pg_cntl;
+
+	rlc_pg_cntl = RREG32_SOC15(GC, 0, regRLC_PG_CNTL);
+
+	if (!enable) {
+		/* RLC_PG_CNTL[23] = 0 (default)
+		 * RLC will wait for handshake acks with SMU
+		 * GFXOFF will be enabled
+		 * RLC_PG_CNTL[23] = 1
+		 * RLC will not issue any message to SMU
+		 * hence no handshake between SMU & RLC
+		 * GFXOFF will be disabled
+		 */
+		rlc_pg_cntl |= RLC_PG_CNTL__SMU_HANDSHAKE_DISABLE_MASK;
+	} else
+		rlc_pg_cntl &= ~RLC_PG_CNTL__SMU_HANDSHAKE_DISABLE_MASK;
+	WREG32_SOC15(GC, 0, regRLC_PG_CNTL, rlc_pg_cntl);
+}
+
+static void gfx_v11_0_rlc_start(struct amdgpu_device *adev)
+{
+	/* TODO: enable rlc & smu handshake until smu
+	 * and gfxoff feature works as expected */
+	if (!(amdgpu_pp_feature_mask & PP_GFXOFF_MASK))
+		gfx_v11_0_rlc_smu_handshake_cntl(adev, false);
+
+	WREG32_FIELD15_PREREG(GC, 0, RLC_CNTL, RLC_ENABLE_F32, 1);
+	udelay(50);
+}
+
+static void gfx_v11_0_rlc_enable_srm(struct amdgpu_device *adev)
+{
+	uint32_t tmp;
+
+	/* enable Save Restore Machine */
+	tmp = RREG32(SOC15_REG_OFFSET(GC, 0, regRLC_SRM_CNTL));
+	tmp |= RLC_SRM_CNTL__AUTO_INCR_ADDR_MASK;
+	tmp |= RLC_SRM_CNTL__SRM_ENABLE_MASK;
+	WREG32(SOC15_REG_OFFSET(GC, 0, regRLC_SRM_CNTL), tmp);
+}
+
+static void gfx_v11_0_load_rlcg_microcode(struct amdgpu_device *adev)
+{
+	const struct rlc_firmware_header_v2_0 *hdr;
+	const __le32 *fw_data;
+	unsigned i, fw_size;
+
+	hdr = (const struct rlc_firmware_header_v2_0 *)adev->gfx.rlc_fw->data;
+	fw_data = (const __le32 *)(adev->gfx.rlc_fw->data +
+			   le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+	fw_size = le32_to_cpu(hdr->header.ucode_size_bytes) / 4;
+
+	WREG32_SOC15(GC, 0, regRLC_GPM_UCODE_ADDR,
+		     RLCG_UCODE_LOADING_START_ADDRESS);
+
+	for (i = 0; i < fw_size; i++)
+		WREG32_SOC15(GC, 0, regRLC_GPM_UCODE_DATA,
+			     le32_to_cpup(fw_data++));
+
+	WREG32_SOC15(GC, 0, regRLC_GPM_UCODE_ADDR, adev->gfx.rlc_fw_version);
+}
+
+static void gfx_v11_0_load_rlc_iram_dram_microcode(struct amdgpu_device *adev)
+{
+	const struct rlc_firmware_header_v2_2 *hdr;
+	const __le32 *fw_data;
+	unsigned i, fw_size;
+	u32 tmp;
+
+	hdr = (const struct rlc_firmware_header_v2_2 *)adev->gfx.rlc_fw->data;
+
+	fw_data = (const __le32 *)(adev->gfx.rlc_fw->data +
+			le32_to_cpu(hdr->rlc_iram_ucode_offset_bytes));
+	fw_size = le32_to_cpu(hdr->rlc_iram_ucode_size_bytes) / 4;
+
+	WREG32_SOC15(GC, 0, regRLC_LX6_IRAM_ADDR, 0);
+
+	for (i = 0; i < fw_size; i++) {
+		if ((amdgpu_emu_mode == 1) && (i % 100 == 99))
+			msleep(1);
+		WREG32_SOC15(GC, 0, regRLC_LX6_IRAM_DATA,
+				le32_to_cpup(fw_data++));
+	}
+
+	WREG32_SOC15(GC, 0, regRLC_LX6_IRAM_ADDR, adev->gfx.rlc_fw_version);
+
+	fw_data = (const __le32 *)(adev->gfx.rlc_fw->data +
+			le32_to_cpu(hdr->rlc_dram_ucode_offset_bytes));
+	fw_size = le32_to_cpu(hdr->rlc_dram_ucode_size_bytes) / 4;
+
+	WREG32_SOC15(GC, 0, regRLC_LX6_DRAM_ADDR, 0);
+	for (i = 0; i < fw_size; i++) {
+		if ((amdgpu_emu_mode == 1) && (i % 100 == 99))
+			msleep(1);
+		WREG32_SOC15(GC, 0, regRLC_LX6_DRAM_DATA,
+				le32_to_cpup(fw_data++));
+	}
+
+	WREG32_SOC15(GC, 0, regRLC_LX6_IRAM_ADDR, adev->gfx.rlc_fw_version);
+
+	tmp = RREG32_SOC15(GC, 0, regRLC_LX6_CNTL);
+	tmp = REG_SET_FIELD(tmp, RLC_LX6_CNTL, PDEBUG_ENABLE, 1);
+	tmp = REG_SET_FIELD(tmp, RLC_LX6_CNTL, BRESET, 0);
+	WREG32_SOC15(GC, 0, regRLC_LX6_CNTL, tmp);
+}
+
+static void gfx_v11_0_load_rlcp_rlcv_microcode(struct amdgpu_device *adev)
+{
+	const struct rlc_firmware_header_v2_3 *hdr;
+	const __le32 *fw_data;
+	unsigned i, fw_size;
+	u32 tmp;
+
+	hdr = (const struct rlc_firmware_header_v2_3 *)adev->gfx.rlc_fw->data;
+
+	fw_data = (const __le32 *)(adev->gfx.rlc_fw->data +
+			le32_to_cpu(hdr->rlcp_ucode_offset_bytes));
+	fw_size = le32_to_cpu(hdr->rlcp_ucode_size_bytes) / 4;
+
+	WREG32_SOC15(GC, 0, regRLC_PACE_UCODE_ADDR, 0);
+
+	for (i = 0; i < fw_size; i++) {
+		if ((amdgpu_emu_mode == 1) && (i % 100 == 99))
+			msleep(1);
+		WREG32_SOC15(GC, 0, regRLC_PACE_UCODE_DATA,
+				le32_to_cpup(fw_data++));
+	}
+
+	WREG32_SOC15(GC, 0, regRLC_PACE_UCODE_ADDR, adev->gfx.rlc_fw_version);
+
+	tmp = RREG32_SOC15(GC, 0, regRLC_GPM_THREAD_ENABLE);
+	tmp = REG_SET_FIELD(tmp, RLC_GPM_THREAD_ENABLE, THREAD1_ENABLE, 1);
+	WREG32_SOC15(GC, 0, regRLC_GPM_THREAD_ENABLE, tmp);
+
+	fw_data = (const __le32 *)(adev->gfx.rlc_fw->data +
+			le32_to_cpu(hdr->rlcv_ucode_offset_bytes));
+	fw_size = le32_to_cpu(hdr->rlcv_ucode_size_bytes) / 4;
+
+	WREG32_SOC15(GC, 0, regRLC_GPU_IOV_UCODE_ADDR, 0);
+
+	for (i = 0; i < fw_size; i++) {
+		if ((amdgpu_emu_mode == 1) && (i % 100 == 99))
+			msleep(1);
+		WREG32_SOC15(GC, 0, regRLC_GPU_IOV_UCODE_DATA,
+				le32_to_cpup(fw_data++));
+	}
+
+	WREG32_SOC15(GC, 0, regRLC_GPU_IOV_UCODE_ADDR, adev->gfx.rlc_fw_version);
+
+	tmp = RREG32_SOC15(GC, 0, regRLC_GPU_IOV_F32_CNTL);
+	tmp = REG_SET_FIELD(tmp, RLC_GPU_IOV_F32_CNTL, ENABLE, 1);
+	WREG32_SOC15(GC, 0, regRLC_GPU_IOV_F32_CNTL, tmp);
+}
+
+static int gfx_v11_0_rlc_load_microcode(struct amdgpu_device *adev)
+{
+	const struct rlc_firmware_header_v2_0 *hdr;
+	uint16_t version_major;
+	uint16_t version_minor;
+
+	if (!adev->gfx.rlc_fw)
+		return -EINVAL;
+
+	hdr = (const struct rlc_firmware_header_v2_0 *)adev->gfx.rlc_fw->data;
+	amdgpu_ucode_print_rlc_hdr(&hdr->header);
+
+	version_major = le16_to_cpu(hdr->header.header_version_major);
+	version_minor = le16_to_cpu(hdr->header.header_version_minor);
+
+	if (version_major == 2) {
+		gfx_v11_0_load_rlcg_microcode(adev);
+		if (amdgpu_dpm == 1) {
+			if (version_minor >= 2)
+				gfx_v11_0_load_rlc_iram_dram_microcode(adev);
+			if (version_minor == 3)
+				gfx_v11_0_load_rlcp_rlcv_microcode(adev);
+		}
+
+		return 0;
+	}
+
+	return -EINVAL;
+}
+
+static int gfx_v11_0_rlc_resume(struct amdgpu_device *adev)
+{
+	int r;
+
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
+		gfx_v11_0_init_csb(adev);
+
+		if (!amdgpu_sriov_vf(adev)) /* enable RLC SRM */
+			gfx_v11_0_rlc_enable_srm(adev);
+	} else {
+		if (amdgpu_sriov_vf(adev)) {
+			gfx_v11_0_init_csb(adev);
+			return 0;
+		}
+
+		adev->gfx.rlc.funcs->stop(adev);
+
+		/* disable CG */
+		WREG32_SOC15(GC, 0, regRLC_CGCG_CGLS_CTRL, 0);
+
+		/* disable PG */
+		WREG32_SOC15(GC, 0, regRLC_PG_CNTL, 0);
+
+		if (adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT) {
+			/* legacy rlc firmware loading */
+			r = gfx_v11_0_rlc_load_microcode(adev);
+			if (r)
+				return r;
+		}
+
+		gfx_v11_0_init_csb(adev);
+
+		adev->gfx.rlc.funcs->start(adev);
+	}
+	return 0;
+}
+
+static int gfx_v11_0_config_me_cache(struct amdgpu_device *adev, uint64_t addr)
+{
+	uint32_t usec_timeout = 50000;  /* wait for 50ms */
+	uint32_t tmp;
+	int i;
+
+	/* Trigger an invalidation of the L1 instruction caches */
+	tmp = RREG32_SOC15(GC, 0, regCP_ME_IC_OP_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_ME_IC_OP_CNTL, INVALIDATE_CACHE, 1);
+	WREG32_SOC15(GC, 0, regCP_ME_IC_OP_CNTL, tmp);
+
+	/* Wait for invalidation complete */
+	for (i = 0; i < usec_timeout; i++) {
+		tmp = RREG32_SOC15(GC, 0, regCP_ME_IC_OP_CNTL);
+		if (1 == REG_GET_FIELD(tmp, CP_ME_IC_OP_CNTL,
+					INVALIDATE_CACHE_COMPLETE))
+			break;
+		udelay(1);
+	}
+
+	if (i >= usec_timeout) {
+		dev_err(adev->dev, "failed to invalidate instruction cache\n");
+		return -EINVAL;
+	}
+
+	if (amdgpu_emu_mode == 1)
+		amdgpu_device_flush_hdp(adev, NULL);
+
+	tmp = RREG32_SOC15(GC, 0, regCP_ME_IC_BASE_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_ME_IC_BASE_CNTL, VMID, 0);
+	tmp = REG_SET_FIELD(tmp, CP_ME_IC_BASE_CNTL, CACHE_POLICY, 0);
+	tmp = REG_SET_FIELD(tmp, CP_ME_IC_BASE_CNTL, EXE_DISABLE, 0);
+	tmp = REG_SET_FIELD(tmp, CP_ME_IC_BASE_CNTL, ADDRESS_CLAMP, 1);
+	WREG32_SOC15(GC, 0, regCP_ME_IC_BASE_CNTL, tmp);
+
+	/* Program me ucode address into intruction cache address register */
+	WREG32_SOC15(GC, 0, regCP_ME_IC_BASE_LO,
+			lower_32_bits(addr) & 0xFFFFF000);
+	WREG32_SOC15(GC, 0, regCP_ME_IC_BASE_HI,
+			upper_32_bits(addr));
+
+	return 0;
+}
+
+static int gfx_v11_0_config_pfp_cache(struct amdgpu_device *adev, uint64_t addr)
+{
+	uint32_t usec_timeout = 50000;  /* wait for 50ms */
+	uint32_t tmp;
+	int i;
+
+	/* Trigger an invalidation of the L1 instruction caches */
+	tmp = RREG32_SOC15(GC, 0, regCP_PFP_IC_OP_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_PFP_IC_OP_CNTL, INVALIDATE_CACHE, 1);
+	WREG32_SOC15(GC, 0, regCP_PFP_IC_OP_CNTL, tmp);
+
+	/* Wait for invalidation complete */
+	for (i = 0; i < usec_timeout; i++) {
+		tmp = RREG32_SOC15(GC, 0, regCP_PFP_IC_OP_CNTL);
+		if (1 == REG_GET_FIELD(tmp, CP_PFP_IC_OP_CNTL,
+					INVALIDATE_CACHE_COMPLETE))
+			break;
+		udelay(1);
+	}
+
+	if (i >= usec_timeout) {
+		dev_err(adev->dev, "failed to invalidate instruction cache\n");
+		return -EINVAL;
+	}
+
+	if (amdgpu_emu_mode == 1)
+		amdgpu_device_flush_hdp(adev, NULL);
+
+	tmp = RREG32_SOC15(GC, 0, regCP_PFP_IC_BASE_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_PFP_IC_BASE_CNTL, VMID, 0);
+	tmp = REG_SET_FIELD(tmp, CP_PFP_IC_BASE_CNTL, CACHE_POLICY, 0);
+	tmp = REG_SET_FIELD(tmp, CP_PFP_IC_BASE_CNTL, EXE_DISABLE, 0);
+	tmp = REG_SET_FIELD(tmp, CP_PFP_IC_BASE_CNTL, ADDRESS_CLAMP, 1);
+	WREG32_SOC15(GC, 0, regCP_PFP_IC_BASE_CNTL, tmp);
+
+	/* Program pfp ucode address into intruction cache address register */
+	WREG32_SOC15(GC, 0, regCP_PFP_IC_BASE_LO,
+			lower_32_bits(addr) & 0xFFFFF000);
+	WREG32_SOC15(GC, 0, regCP_PFP_IC_BASE_HI,
+			upper_32_bits(addr));
+
+	return 0;
+}
+
+static int gfx_v11_0_config_mec_cache(struct amdgpu_device *adev, uint64_t addr)
+{
+	uint32_t usec_timeout = 50000;  /* wait for 50ms */
+	uint32_t tmp;
+	int i;
+
+	/* Trigger an invalidation of the L1 instruction caches */
+	tmp = RREG32_SOC15(GC, 0, regCP_CPC_IC_OP_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_CPC_IC_OP_CNTL, INVALIDATE_CACHE, 1);
+
+	WREG32_SOC15(GC, 0, regCP_CPC_IC_OP_CNTL, tmp);
+
+	/* Wait for invalidation complete */
+	for (i = 0; i < usec_timeout; i++) {
+		tmp = RREG32_SOC15(GC, 0, regCP_CPC_IC_OP_CNTL);
+		if (1 == REG_GET_FIELD(tmp, CP_CPC_IC_OP_CNTL,
+					INVALIDATE_CACHE_COMPLETE))
+			break;
+		udelay(1);
+	}
+
+	if (i >= usec_timeout) {
+		dev_err(adev->dev, "failed to invalidate instruction cache\n");
+		return -EINVAL;
+	}
+
+	if (amdgpu_emu_mode == 1)
+		amdgpu_device_flush_hdp(adev, NULL);
+
+	tmp = RREG32_SOC15(GC, 0, regCP_CPC_IC_BASE_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_CPC_IC_BASE_CNTL, CACHE_POLICY, 0);
+	tmp = REG_SET_FIELD(tmp, CP_CPC_IC_BASE_CNTL, EXE_DISABLE, 0);
+	tmp = REG_SET_FIELD(tmp, CP_CPC_IC_BASE_CNTL, ADDRESS_CLAMP, 1);
+	WREG32_SOC15(GC, 0, regCP_CPC_IC_BASE_CNTL, tmp);
+
+	/* Program mec1 ucode address into intruction cache address register */
+	WREG32_SOC15(GC, 0, regCP_CPC_IC_BASE_LO,
+			lower_32_bits(addr) & 0xFFFFF000);
+	WREG32_SOC15(GC, 0, regCP_CPC_IC_BASE_HI,
+			upper_32_bits(addr));
+
+	return 0;
+}
+
+static int gfx_v11_0_config_pfp_cache_rs64(struct amdgpu_device *adev, uint64_t addr, uint64_t addr2)
+{
+	uint32_t usec_timeout = 50000;  /* wait for 50ms */
+	uint32_t tmp;
+	unsigned i, pipe_id;
+	const struct gfx_firmware_header_v2_0 *pfp_hdr;
+
+	pfp_hdr = (const struct gfx_firmware_header_v2_0 *)
+		adev->gfx.pfp_fw->data;
+
+	WREG32_SOC15(GC, 0, regCP_PFP_IC_BASE_LO,
+		lower_32_bits(addr));
+	WREG32_SOC15(GC, 0, regCP_PFP_IC_BASE_HI,
+		upper_32_bits(addr));
+
+	tmp = RREG32_SOC15(GC, 0, regCP_PFP_IC_BASE_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_PFP_IC_BASE_CNTL, VMID, 0);
+	tmp = REG_SET_FIELD(tmp, CP_PFP_IC_BASE_CNTL, CACHE_POLICY, 0);
+	tmp = REG_SET_FIELD(tmp, CP_PFP_IC_BASE_CNTL, EXE_DISABLE, 0);
+	WREG32_SOC15(GC, 0, regCP_PFP_IC_BASE_CNTL, tmp);
+
+	/*
+	 * Programming any of the CP_PFP_IC_BASE registers
+	 * forces invalidation of the ME L1 I$. Wait for the
+	 * invalidation complete
+	 */
+	for (i = 0; i < usec_timeout; i++) {
+		tmp = RREG32_SOC15(GC, 0, regCP_PFP_IC_OP_CNTL);
+		if (1 == REG_GET_FIELD(tmp, CP_PFP_IC_OP_CNTL,
+			INVALIDATE_CACHE_COMPLETE))
+			break;
+		udelay(1);
+	}
+
+	if (i >= usec_timeout) {
+		dev_err(adev->dev, "failed to invalidate instruction cache\n");
+		return -EINVAL;
+	}
+
+	/* Prime the L1 instruction caches */
+	tmp = RREG32_SOC15(GC, 0, regCP_PFP_IC_OP_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_PFP_IC_OP_CNTL, PRIME_ICACHE, 1);
+	WREG32_SOC15(GC, 0, regCP_PFP_IC_OP_CNTL, tmp);
+	/* Waiting for cache primed*/
+	for (i = 0; i < usec_timeout; i++) {
+		tmp = RREG32_SOC15(GC, 0, regCP_PFP_IC_OP_CNTL);
+		if (1 == REG_GET_FIELD(tmp, CP_PFP_IC_OP_CNTL,
+			ICACHE_PRIMED))
+			break;
+		udelay(1);
+	}
+
+	if (i >= usec_timeout) {
+		dev_err(adev->dev, "failed to prime instruction cache\n");
+		return -EINVAL;
+	}
+
+	mutex_lock(&adev->srbm_mutex);
+	for (pipe_id = 0; pipe_id < adev->gfx.me.num_pipe_per_me; pipe_id++) {
+		soc21_grbm_select(adev, 0, pipe_id, 0, 0);
+		WREG32_SOC15(GC, 0, regCP_PFP_PRGRM_CNTR_START,
+			(pfp_hdr->ucode_start_addr_hi << 30) |
+			(pfp_hdr->ucode_start_addr_lo >> 2));
+		WREG32_SOC15(GC, 0, regCP_PFP_PRGRM_CNTR_START_HI,
+			pfp_hdr->ucode_start_addr_hi >> 2);
+
+		/*
+		 * Program CP_ME_CNTL to reset given PIPE to take
+		 * effect of CP_PFP_PRGRM_CNTR_START.
+		 */
+		tmp = RREG32_SOC15(GC, 0, regCP_ME_CNTL);
+		if (pipe_id == 0)
+			tmp = REG_SET_FIELD(tmp, CP_ME_CNTL,
+					PFP_PIPE0_RESET, 1);
+		else
+			tmp = REG_SET_FIELD(tmp, CP_ME_CNTL,
+					PFP_PIPE1_RESET, 1);
+		WREG32_SOC15(GC, 0, regCP_ME_CNTL, tmp);
+
+		/* Clear pfp pipe0 reset bit. */
+		if (pipe_id == 0)
+			tmp = REG_SET_FIELD(tmp, CP_ME_CNTL,
+					PFP_PIPE0_RESET, 0);
+		else
+			tmp = REG_SET_FIELD(tmp, CP_ME_CNTL,
+					PFP_PIPE1_RESET, 0);
+		WREG32_SOC15(GC, 0, regCP_ME_CNTL, tmp);
+
+		WREG32_SOC15(GC, 0, regCP_GFX_RS64_DC_BASE0_LO,
+			lower_32_bits(addr2));
+		WREG32_SOC15(GC, 0, regCP_GFX_RS64_DC_BASE0_HI,
+			upper_32_bits(addr2));
+	}
+	soc21_grbm_select(adev, 0, 0, 0, 0);
+	mutex_unlock(&adev->srbm_mutex);
+
+	tmp = RREG32_SOC15(GC, 0, regCP_GFX_RS64_DC_BASE_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_GFX_RS64_DC_BASE_CNTL, VMID, 0);
+	tmp = REG_SET_FIELD(tmp, CP_GFX_RS64_DC_BASE_CNTL, CACHE_POLICY, 0);
+	WREG32_SOC15(GC, 0, regCP_GFX_RS64_DC_BASE_CNTL, tmp);
+
+	/* Invalidate the data caches */
+	tmp = RREG32_SOC15(GC, 0, regCP_GFX_RS64_DC_OP_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_GFX_RS64_DC_OP_CNTL, INVALIDATE_DCACHE, 1);
+	WREG32_SOC15(GC, 0, regCP_GFX_RS64_DC_OP_CNTL, tmp);
+
+	for (i = 0; i < usec_timeout; i++) {
+		tmp = RREG32_SOC15(GC, 0, regCP_GFX_RS64_DC_OP_CNTL);
+		if (1 == REG_GET_FIELD(tmp, CP_GFX_RS64_DC_OP_CNTL,
+			INVALIDATE_DCACHE_COMPLETE))
+			break;
+		udelay(1);
+	}
+
+	if (i >= usec_timeout) {
+		dev_err(adev->dev, "failed to invalidate RS64 data cache\n");
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int gfx_v11_0_config_me_cache_rs64(struct amdgpu_device *adev, uint64_t addr, uint64_t addr2)
+{
+	uint32_t usec_timeout = 50000;  /* wait for 50ms */
+	uint32_t tmp;
+	unsigned i, pipe_id;
+	const struct gfx_firmware_header_v2_0 *me_hdr;
+
+	me_hdr = (const struct gfx_firmware_header_v2_0 *)
+		adev->gfx.me_fw->data;
+
+	WREG32_SOC15(GC, 0, regCP_ME_IC_BASE_LO,
+		lower_32_bits(addr));
+	WREG32_SOC15(GC, 0, regCP_ME_IC_BASE_HI,
+		upper_32_bits(addr));
+
+	tmp = RREG32_SOC15(GC, 0, regCP_ME_IC_BASE_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_ME_IC_BASE_CNTL, VMID, 0);
+	tmp = REG_SET_FIELD(tmp, CP_ME_IC_BASE_CNTL, CACHE_POLICY, 0);
+	tmp = REG_SET_FIELD(tmp, CP_ME_IC_BASE_CNTL, EXE_DISABLE, 0);
+	WREG32_SOC15(GC, 0, regCP_ME_IC_BASE_CNTL, tmp);
+
+	/*
+	 * Programming any of the CP_ME_IC_BASE registers
+	 * forces invalidation of the ME L1 I$. Wait for the
+	 * invalidation complete
+	 */
+	for (i = 0; i < usec_timeout; i++) {
+		tmp = RREG32_SOC15(GC, 0, regCP_ME_IC_OP_CNTL);
+		if (1 == REG_GET_FIELD(tmp, CP_ME_IC_OP_CNTL,
+			INVALIDATE_CACHE_COMPLETE))
+			break;
+		udelay(1);
+	}
+
+	if (i >= usec_timeout) {
+		dev_err(adev->dev, "failed to invalidate instruction cache\n");
+		return -EINVAL;
+	}
+
+	/* Prime the instruction caches */
+	tmp = RREG32_SOC15(GC, 0, regCP_ME_IC_OP_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_ME_IC_OP_CNTL, PRIME_ICACHE, 1);
+	WREG32_SOC15(GC, 0, regCP_ME_IC_OP_CNTL, tmp);
+
+	/* Waiting for instruction cache primed*/
+	for (i = 0; i < usec_timeout; i++) {
+		tmp = RREG32_SOC15(GC, 0, regCP_ME_IC_OP_CNTL);
+		if (1 == REG_GET_FIELD(tmp, CP_ME_IC_OP_CNTL,
+			ICACHE_PRIMED))
+			break;
+		udelay(1);
+	}
+
+	if (i >= usec_timeout) {
+		dev_err(adev->dev, "failed to prime instruction cache\n");
+		return -EINVAL;
+	}
+
+	mutex_lock(&adev->srbm_mutex);
+	for (pipe_id = 0; pipe_id < adev->gfx.me.num_pipe_per_me; pipe_id++) {
+		soc21_grbm_select(adev, 0, pipe_id, 0, 0);
+		WREG32_SOC15(GC, 0, regCP_ME_PRGRM_CNTR_START,
+			(me_hdr->ucode_start_addr_hi << 30) |
+			(me_hdr->ucode_start_addr_lo >> 2) );
+		WREG32_SOC15(GC, 0, regCP_ME_PRGRM_CNTR_START_HI,
+			me_hdr->ucode_start_addr_hi>>2);
+
+		/*
+		 * Program CP_ME_CNTL to reset given PIPE to take
+		 * effect of CP_PFP_PRGRM_CNTR_START.
+		 */
+		tmp = RREG32_SOC15(GC, 0, regCP_ME_CNTL);
+		if (pipe_id == 0)
+			tmp = REG_SET_FIELD(tmp, CP_ME_CNTL,
+					ME_PIPE0_RESET, 1);
+		else
+			tmp = REG_SET_FIELD(tmp, CP_ME_CNTL,
+					ME_PIPE1_RESET, 1);
+		WREG32_SOC15(GC, 0, regCP_ME_CNTL, tmp);
+
+		/* Clear pfp pipe0 reset bit. */
+		if (pipe_id == 0)
+			tmp = REG_SET_FIELD(tmp, CP_ME_CNTL,
+					ME_PIPE0_RESET, 0);
+		else
+			tmp = REG_SET_FIELD(tmp, CP_ME_CNTL,
+					ME_PIPE1_RESET, 0);
+		WREG32_SOC15(GC, 0, regCP_ME_CNTL, tmp);
+
+		WREG32_SOC15(GC, 0, regCP_GFX_RS64_DC_BASE1_LO,
+			lower_32_bits(addr2));
+		WREG32_SOC15(GC, 0, regCP_GFX_RS64_DC_BASE1_HI,
+			upper_32_bits(addr2));
+	}
+	soc21_grbm_select(adev, 0, 0, 0, 0);
+	mutex_unlock(&adev->srbm_mutex);
+
+	tmp = RREG32_SOC15(GC, 0, regCP_GFX_RS64_DC_BASE_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_GFX_RS64_DC_BASE_CNTL, VMID, 0);
+	tmp = REG_SET_FIELD(tmp, CP_GFX_RS64_DC_BASE_CNTL, CACHE_POLICY, 0);
+	WREG32_SOC15(GC, 0, regCP_GFX_RS64_DC_BASE_CNTL, tmp);
+
+	/* Invalidate the data caches */
+	tmp = RREG32_SOC15(GC, 0, regCP_GFX_RS64_DC_OP_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_GFX_RS64_DC_OP_CNTL, INVALIDATE_DCACHE, 1);
+	WREG32_SOC15(GC, 0, regCP_GFX_RS64_DC_OP_CNTL, tmp);
+
+	for (i = 0; i < usec_timeout; i++) {
+		tmp = RREG32_SOC15(GC, 0, regCP_GFX_RS64_DC_OP_CNTL);
+		if (1 == REG_GET_FIELD(tmp, CP_GFX_RS64_DC_OP_CNTL,
+			INVALIDATE_DCACHE_COMPLETE))
+			break;
+		udelay(1);
+	}
+
+	if (i >= usec_timeout) {
+		dev_err(adev->dev, "failed to invalidate RS64 data cache\n");
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int gfx_v11_0_config_mec_cache_rs64(struct amdgpu_device *adev, uint64_t addr, uint64_t addr2)
+{
+	uint32_t usec_timeout = 50000;  /* wait for 50ms */
+	uint32_t tmp;
+	unsigned i;
+	const struct gfx_firmware_header_v2_0 *mec_hdr;
+
+	mec_hdr = (const struct gfx_firmware_header_v2_0 *)
+		adev->gfx.mec_fw->data;
+
+	tmp = RREG32_SOC15(GC, 0, regCP_CPC_IC_BASE_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_CPC_IC_BASE_CNTL, VMID, 0);
+	tmp = REG_SET_FIELD(tmp, CP_CPC_IC_BASE_CNTL, EXE_DISABLE, 0);
+	tmp = REG_SET_FIELD(tmp, CP_CPC_IC_BASE_CNTL, CACHE_POLICY, 0);
+	WREG32_SOC15(GC, 0, regCP_CPC_IC_BASE_CNTL, tmp);
+
+	tmp = RREG32_SOC15(GC, 0, regCP_MEC_DC_BASE_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_MEC_DC_BASE_CNTL, VMID, 0);
+	tmp = REG_SET_FIELD(tmp, CP_MEC_DC_BASE_CNTL, CACHE_POLICY, 0);
+	WREG32_SOC15(GC, 0, regCP_MEC_DC_BASE_CNTL, tmp);
+
+	mutex_lock(&adev->srbm_mutex);
+	for (i = 0; i < adev->gfx.mec.num_pipe_per_mec; i++) {
+		soc21_grbm_select(adev, 1, i, 0, 0);
+
+		WREG32_SOC15(GC, 0, regCP_MEC_MDBASE_LO, addr2);
+		WREG32_SOC15(GC, 0, regCP_MEC_MDBASE_HI,
+		     upper_32_bits(addr2));
+
+		WREG32_SOC15(GC, 0, regCP_MEC_RS64_PRGRM_CNTR_START,
+					mec_hdr->ucode_start_addr_lo >> 2 |
+					mec_hdr->ucode_start_addr_hi << 30);
+		WREG32_SOC15(GC, 0, regCP_MEC_RS64_PRGRM_CNTR_START_HI,
+					mec_hdr->ucode_start_addr_hi >> 2);
+
+		WREG32_SOC15(GC, 0, regCP_CPC_IC_BASE_LO, addr);
+		WREG32_SOC15(GC, 0, regCP_CPC_IC_BASE_HI,
+		     upper_32_bits(addr));
+	}
+	mutex_unlock(&adev->srbm_mutex);
+	soc21_grbm_select(adev, 0, 0, 0, 0);
+
+	/* Trigger an invalidation of the L1 instruction caches */
+	tmp = RREG32_SOC15(GC, 0, regCP_MEC_DC_OP_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_MEC_DC_OP_CNTL, INVALIDATE_DCACHE, 1);
+	WREG32_SOC15(GC, 0, regCP_MEC_DC_OP_CNTL, tmp);
+
+	/* Wait for invalidation complete */
+	for (i = 0; i < usec_timeout; i++) {
+		tmp = RREG32_SOC15(GC, 0, regCP_MEC_DC_OP_CNTL);
+		if (1 == REG_GET_FIELD(tmp, CP_MEC_DC_OP_CNTL,
+				       INVALIDATE_DCACHE_COMPLETE))
+			break;
+		udelay(1);
+	}
+
+	if (i >= usec_timeout) {
+		dev_err(adev->dev, "failed to invalidate instruction cache\n");
+		return -EINVAL;
+	}
+
+	/* Trigger an invalidation of the L1 instruction caches */
+	tmp = RREG32_SOC15(GC, 0, regCP_CPC_IC_OP_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_CPC_IC_OP_CNTL, INVALIDATE_CACHE, 1);
+	WREG32_SOC15(GC, 0, regCP_CPC_IC_OP_CNTL, tmp);
+
+	/* Wait for invalidation complete */
+	for (i = 0; i < usec_timeout; i++) {
+		tmp = RREG32_SOC15(GC, 0, regCP_CPC_IC_OP_CNTL);
+		if (1 == REG_GET_FIELD(tmp, CP_CPC_IC_OP_CNTL,
+				       INVALIDATE_CACHE_COMPLETE))
+			break;
+		udelay(1);
+	}
+
+	if (i >= usec_timeout) {
+		dev_err(adev->dev, "failed to invalidate instruction cache\n");
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static void gfx_v11_0_config_gfx_rs64(struct amdgpu_device *adev)
+{
+	const struct gfx_firmware_header_v2_0 *pfp_hdr;
+	const struct gfx_firmware_header_v2_0 *me_hdr;
+	const struct gfx_firmware_header_v2_0 *mec_hdr;
+	uint32_t pipe_id, tmp;
+
+	mec_hdr = (const struct gfx_firmware_header_v2_0 *)
+		adev->gfx.mec_fw->data;
+	me_hdr = (const struct gfx_firmware_header_v2_0 *)
+		adev->gfx.me_fw->data;
+	pfp_hdr = (const struct gfx_firmware_header_v2_0 *)
+		adev->gfx.pfp_fw->data;
+
+	/* config pfp program start addr */
+	for (pipe_id = 0; pipe_id < 2; pipe_id++) {
+		soc21_grbm_select(adev, 0, pipe_id, 0, 0);
+		WREG32_SOC15(GC, 0, regCP_PFP_PRGRM_CNTR_START,
+			(pfp_hdr->ucode_start_addr_hi << 30) |
+			(pfp_hdr->ucode_start_addr_lo >> 2));
+		WREG32_SOC15(GC, 0, regCP_PFP_PRGRM_CNTR_START_HI,
+			pfp_hdr->ucode_start_addr_hi >> 2);
+	}
+	soc21_grbm_select(adev, 0, 0, 0, 0);
+
+	/* reset pfp pipe */
+	tmp = RREG32_SOC15(GC, 0, regCP_ME_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, PFP_PIPE0_RESET, 1);
+	tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, PFP_PIPE1_RESET, 1);
+	WREG32_SOC15(GC, 0, regCP_ME_CNTL, tmp);
+
+	/* clear pfp pipe reset */
+	tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, PFP_PIPE0_RESET, 0);
+	tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, PFP_PIPE1_RESET, 0);
+	WREG32_SOC15(GC, 0, regCP_ME_CNTL, tmp);
+
+	/* config me program start addr */
+	for (pipe_id = 0; pipe_id < 2; pipe_id++) {
+		soc21_grbm_select(adev, 0, pipe_id, 0, 0);
+		WREG32_SOC15(GC, 0, regCP_ME_PRGRM_CNTR_START,
+			(me_hdr->ucode_start_addr_hi << 30) |
+			(me_hdr->ucode_start_addr_lo >> 2) );
+		WREG32_SOC15(GC, 0, regCP_ME_PRGRM_CNTR_START_HI,
+			me_hdr->ucode_start_addr_hi>>2);
+	}
+	soc21_grbm_select(adev, 0, 0, 0, 0);
+
+	/* reset me pipe */
+	tmp = RREG32_SOC15(GC, 0, regCP_ME_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, ME_PIPE0_RESET, 1);
+	tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, ME_PIPE1_RESET, 1);
+	WREG32_SOC15(GC, 0, regCP_ME_CNTL, tmp);
+
+	/* clear me pipe reset */
+	tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, ME_PIPE0_RESET, 0);
+	tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, ME_PIPE1_RESET, 0);
+	WREG32_SOC15(GC, 0, regCP_ME_CNTL, tmp);
+
+	/* config mec program start addr */
+	for (pipe_id = 0; pipe_id < 4; pipe_id++) {
+		soc21_grbm_select(adev, 1, pipe_id, 0, 0);
+		WREG32_SOC15(GC, 0, regCP_MEC_RS64_PRGRM_CNTR_START,
+					mec_hdr->ucode_start_addr_lo >> 2 |
+					mec_hdr->ucode_start_addr_hi << 30);
+		WREG32_SOC15(GC, 0, regCP_MEC_RS64_PRGRM_CNTR_START_HI,
+					mec_hdr->ucode_start_addr_hi >> 2);
+	}
+	soc21_grbm_select(adev, 0, 0, 0, 0);
+
+	/* reset mec pipe */
+	tmp = RREG32_SOC15(GC, 0, regCP_MEC_RS64_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_MEC_RS64_CNTL, MEC_PIPE0_RESET, 1);
+	tmp = REG_SET_FIELD(tmp, CP_MEC_RS64_CNTL, MEC_PIPE1_RESET, 1);
+	tmp = REG_SET_FIELD(tmp, CP_MEC_RS64_CNTL, MEC_PIPE2_RESET, 1);
+	tmp = REG_SET_FIELD(tmp, CP_MEC_RS64_CNTL, MEC_PIPE3_RESET, 1);
+	WREG32_SOC15(GC, 0, regCP_MEC_RS64_CNTL, tmp);
+
+	/* clear mec pipe reset */
+	tmp = REG_SET_FIELD(tmp, CP_MEC_RS64_CNTL, MEC_PIPE0_RESET, 0);
+	tmp = REG_SET_FIELD(tmp, CP_MEC_RS64_CNTL, MEC_PIPE1_RESET, 0);
+	tmp = REG_SET_FIELD(tmp, CP_MEC_RS64_CNTL, MEC_PIPE2_RESET, 0);
+	tmp = REG_SET_FIELD(tmp, CP_MEC_RS64_CNTL, MEC_PIPE3_RESET, 0);
+	WREG32_SOC15(GC, 0, regCP_MEC_RS64_CNTL, tmp);
+}
+
+static int gfx_v11_0_wait_for_rlc_autoload_complete(struct amdgpu_device *adev)
+{
+	uint32_t cp_status;
+	uint32_t bootload_status;
+	int i, r;
+	uint64_t addr, addr2;
+
+	for (i = 0; i < adev->usec_timeout; i++) {
+		cp_status = RREG32_SOC15(GC, 0, regCP_STAT);
+
+		if (amdgpu_ip_version(adev, GC_HWIP, 0) ==
+			    IP_VERSION(11, 0, 1) ||
+		    amdgpu_ip_version(adev, GC_HWIP, 0) ==
+			    IP_VERSION(11, 0, 4) ||
+		    amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(11, 5, 0) ||
+		    amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(11, 5, 1) ||
+		    amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(11, 5, 2) ||
+		    amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(11, 5, 3))
+			bootload_status = RREG32_SOC15(GC, 0,
+					regRLC_RLCS_BOOTLOAD_STATUS_gc_11_0_1);
+		else
+			bootload_status = RREG32_SOC15(GC, 0, regRLC_RLCS_BOOTLOAD_STATUS);
+
+		if ((cp_status == 0) &&
+		    (REG_GET_FIELD(bootload_status,
+			RLC_RLCS_BOOTLOAD_STATUS, BOOTLOAD_COMPLETE) == 1)) {
+			break;
+		}
+		udelay(1);
+	}
+
+	if (i >= adev->usec_timeout) {
+		dev_err(adev->dev, "rlc autoload: gc ucode autoload timeout\n");
+		return -ETIMEDOUT;
+	}
+
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_RLC_BACKDOOR_AUTO) {
+		if (adev->gfx.rs64_enable) {
+			addr = adev->gfx.rlc.rlc_autoload_gpu_addr +
+				rlc_autoload_info[SOC21_FIRMWARE_ID_RS64_ME].offset;
+			addr2 = adev->gfx.rlc.rlc_autoload_gpu_addr +
+				rlc_autoload_info[SOC21_FIRMWARE_ID_RS64_ME_P0_STACK].offset;
+			r = gfx_v11_0_config_me_cache_rs64(adev, addr, addr2);
+			if (r)
+				return r;
+			addr = adev->gfx.rlc.rlc_autoload_gpu_addr +
+				rlc_autoload_info[SOC21_FIRMWARE_ID_RS64_PFP].offset;
+			addr2 = adev->gfx.rlc.rlc_autoload_gpu_addr +
+				rlc_autoload_info[SOC21_FIRMWARE_ID_RS64_PFP_P0_STACK].offset;
+			r = gfx_v11_0_config_pfp_cache_rs64(adev, addr, addr2);
+			if (r)
+				return r;
+			addr = adev->gfx.rlc.rlc_autoload_gpu_addr +
+				rlc_autoload_info[SOC21_FIRMWARE_ID_RS64_MEC].offset;
+			addr2 = adev->gfx.rlc.rlc_autoload_gpu_addr +
+				rlc_autoload_info[SOC21_FIRMWARE_ID_RS64_MEC_P0_STACK].offset;
+			r = gfx_v11_0_config_mec_cache_rs64(adev, addr, addr2);
+			if (r)
+				return r;
+		} else {
+			addr = adev->gfx.rlc.rlc_autoload_gpu_addr +
+				rlc_autoload_info[SOC21_FIRMWARE_ID_CP_ME].offset;
+			r = gfx_v11_0_config_me_cache(adev, addr);
+			if (r)
+				return r;
+			addr = adev->gfx.rlc.rlc_autoload_gpu_addr +
+				rlc_autoload_info[SOC21_FIRMWARE_ID_CP_PFP].offset;
+			r = gfx_v11_0_config_pfp_cache(adev, addr);
+			if (r)
+				return r;
+			addr = adev->gfx.rlc.rlc_autoload_gpu_addr +
+				rlc_autoload_info[SOC21_FIRMWARE_ID_CP_MEC].offset;
+			r = gfx_v11_0_config_mec_cache(adev, addr);
+			if (r)
+				return r;
+		}
+	}
+
+	return 0;
+}
+
+static int gfx_v11_0_cp_gfx_enable(struct amdgpu_device *adev, bool enable)
+{
+	int i;
+	u32 tmp = RREG32_SOC15(GC, 0, regCP_ME_CNTL);
+
+	tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, ME_HALT, enable ? 0 : 1);
+	tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, PFP_HALT, enable ? 0 : 1);
+	WREG32_SOC15(GC, 0, regCP_ME_CNTL, tmp);
+
+	for (i = 0; i < adev->usec_timeout; i++) {
+		if (RREG32_SOC15(GC, 0, regCP_STAT) == 0)
+			break;
+		udelay(1);
+	}
+
+	if (i >= adev->usec_timeout)
+		DRM_ERROR("failed to %s cp gfx\n", enable ? "unhalt" : "halt");
+
+	return 0;
+}
+
+static int gfx_v11_0_cp_gfx_load_pfp_microcode(struct amdgpu_device *adev)
+{
+	int r;
+	const struct gfx_firmware_header_v1_0 *pfp_hdr;
+	const __le32 *fw_data;
+	unsigned i, fw_size;
+
+	pfp_hdr = (const struct gfx_firmware_header_v1_0 *)
+		adev->gfx.pfp_fw->data;
+
+	amdgpu_ucode_print_gfx_hdr(&pfp_hdr->header);
+
+	fw_data = (const __le32 *)(adev->gfx.pfp_fw->data +
+		le32_to_cpu(pfp_hdr->header.ucode_array_offset_bytes));
+	fw_size = le32_to_cpu(pfp_hdr->header.ucode_size_bytes);
+
+	r = amdgpu_bo_create_reserved(adev, pfp_hdr->header.ucode_size_bytes,
+				      PAGE_SIZE, AMDGPU_GEM_DOMAIN_GTT,
+				      &adev->gfx.pfp.pfp_fw_obj,
+				      &adev->gfx.pfp.pfp_fw_gpu_addr,
+				      (void **)&adev->gfx.pfp.pfp_fw_ptr);
+	if (r) {
+		dev_err(adev->dev, "(%d) failed to create pfp fw bo\n", r);
+		gfx_v11_0_pfp_fini(adev);
+		return r;
+	}
+
+	memcpy(adev->gfx.pfp.pfp_fw_ptr, fw_data, fw_size);
+
+	amdgpu_bo_kunmap(adev->gfx.pfp.pfp_fw_obj);
+	amdgpu_bo_unreserve(adev->gfx.pfp.pfp_fw_obj);
+
+	gfx_v11_0_config_pfp_cache(adev, adev->gfx.pfp.pfp_fw_gpu_addr);
+
+	WREG32_SOC15(GC, 0, regCP_HYP_PFP_UCODE_ADDR, 0);
+
+	for (i = 0; i < pfp_hdr->jt_size; i++)
+		WREG32_SOC15(GC, 0, regCP_HYP_PFP_UCODE_DATA,
+			     le32_to_cpup(fw_data + pfp_hdr->jt_offset + i));
+
+	WREG32_SOC15(GC, 0, regCP_HYP_PFP_UCODE_ADDR, adev->gfx.pfp_fw_version);
+
+	return 0;
+}
+
+static int gfx_v11_0_cp_gfx_load_pfp_microcode_rs64(struct amdgpu_device *adev)
+{
+	int r;
+	const struct gfx_firmware_header_v2_0 *pfp_hdr;
+	const __le32 *fw_ucode, *fw_data;
+	unsigned i, pipe_id, fw_ucode_size, fw_data_size;
+	uint32_t tmp;
+	uint32_t usec_timeout = 50000;  /* wait for 50ms */
+
+	pfp_hdr = (const struct gfx_firmware_header_v2_0 *)
+		adev->gfx.pfp_fw->data;
+
+	amdgpu_ucode_print_gfx_hdr(&pfp_hdr->header);
+
+	/* instruction */
+	fw_ucode = (const __le32 *)(adev->gfx.pfp_fw->data +
+		le32_to_cpu(pfp_hdr->ucode_offset_bytes));
+	fw_ucode_size = le32_to_cpu(pfp_hdr->ucode_size_bytes);
+	/* data */
+	fw_data = (const __le32 *)(adev->gfx.pfp_fw->data +
+		le32_to_cpu(pfp_hdr->data_offset_bytes));
+	fw_data_size = le32_to_cpu(pfp_hdr->data_size_bytes);
+
+	/* 64kb align */
+	r = amdgpu_bo_create_reserved(adev, fw_ucode_size,
+				      64 * 1024,
+				      AMDGPU_GEM_DOMAIN_VRAM |
+				      AMDGPU_GEM_DOMAIN_GTT,
+				      &adev->gfx.pfp.pfp_fw_obj,
+				      &adev->gfx.pfp.pfp_fw_gpu_addr,
+				      (void **)&adev->gfx.pfp.pfp_fw_ptr);
+	if (r) {
+		dev_err(adev->dev, "(%d) failed to create pfp ucode fw bo\n", r);
+		gfx_v11_0_pfp_fini(adev);
+		return r;
+	}
+
+	r = amdgpu_bo_create_reserved(adev, fw_data_size,
+				      64 * 1024,
+				      AMDGPU_GEM_DOMAIN_VRAM |
+				      AMDGPU_GEM_DOMAIN_GTT,
+				      &adev->gfx.pfp.pfp_fw_data_obj,
+				      &adev->gfx.pfp.pfp_fw_data_gpu_addr,
+				      (void **)&adev->gfx.pfp.pfp_fw_data_ptr);
+	if (r) {
+		dev_err(adev->dev, "(%d) failed to create pfp data fw bo\n", r);
+		gfx_v11_0_pfp_fini(adev);
+		return r;
+	}
+
+	memcpy(adev->gfx.pfp.pfp_fw_ptr, fw_ucode, fw_ucode_size);
+	memcpy(adev->gfx.pfp.pfp_fw_data_ptr, fw_data, fw_data_size);
+
+	amdgpu_bo_kunmap(adev->gfx.pfp.pfp_fw_obj);
+	amdgpu_bo_kunmap(adev->gfx.pfp.pfp_fw_data_obj);
+	amdgpu_bo_unreserve(adev->gfx.pfp.pfp_fw_obj);
+	amdgpu_bo_unreserve(adev->gfx.pfp.pfp_fw_data_obj);
+
+	if (amdgpu_emu_mode == 1)
+		amdgpu_device_flush_hdp(adev, NULL);
+
+	WREG32_SOC15(GC, 0, regCP_PFP_IC_BASE_LO,
+		lower_32_bits(adev->gfx.pfp.pfp_fw_gpu_addr));
+	WREG32_SOC15(GC, 0, regCP_PFP_IC_BASE_HI,
+		upper_32_bits(adev->gfx.pfp.pfp_fw_gpu_addr));
+
+	tmp = RREG32_SOC15(GC, 0, regCP_PFP_IC_BASE_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_PFP_IC_BASE_CNTL, VMID, 0);
+	tmp = REG_SET_FIELD(tmp, CP_PFP_IC_BASE_CNTL, CACHE_POLICY, 0);
+	tmp = REG_SET_FIELD(tmp, CP_PFP_IC_BASE_CNTL, EXE_DISABLE, 0);
+	WREG32_SOC15(GC, 0, regCP_PFP_IC_BASE_CNTL, tmp);
+
+	/*
+	 * Programming any of the CP_PFP_IC_BASE registers
+	 * forces invalidation of the ME L1 I$. Wait for the
+	 * invalidation complete
+	 */
+	for (i = 0; i < usec_timeout; i++) {
+		tmp = RREG32_SOC15(GC, 0, regCP_PFP_IC_OP_CNTL);
+		if (1 == REG_GET_FIELD(tmp, CP_PFP_IC_OP_CNTL,
+			INVALIDATE_CACHE_COMPLETE))
+			break;
+		udelay(1);
+	}
+
+	if (i >= usec_timeout) {
+		dev_err(adev->dev, "failed to invalidate instruction cache\n");
+		return -EINVAL;
+	}
+
+	/* Prime the L1 instruction caches */
+	tmp = RREG32_SOC15(GC, 0, regCP_PFP_IC_OP_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_PFP_IC_OP_CNTL, PRIME_ICACHE, 1);
+	WREG32_SOC15(GC, 0, regCP_PFP_IC_OP_CNTL, tmp);
+	/* Waiting for cache primed*/
+	for (i = 0; i < usec_timeout; i++) {
+		tmp = RREG32_SOC15(GC, 0, regCP_PFP_IC_OP_CNTL);
+		if (1 == REG_GET_FIELD(tmp, CP_PFP_IC_OP_CNTL,
+			ICACHE_PRIMED))
+			break;
+		udelay(1);
+	}
+
+	if (i >= usec_timeout) {
+		dev_err(adev->dev, "failed to prime instruction cache\n");
+		return -EINVAL;
+	}
+
+	mutex_lock(&adev->srbm_mutex);
+	for (pipe_id = 0; pipe_id < adev->gfx.me.num_pipe_per_me; pipe_id++) {
+		soc21_grbm_select(adev, 0, pipe_id, 0, 0);
+		WREG32_SOC15(GC, 0, regCP_PFP_PRGRM_CNTR_START,
+			(pfp_hdr->ucode_start_addr_hi << 30) |
+			(pfp_hdr->ucode_start_addr_lo >> 2) );
+		WREG32_SOC15(GC, 0, regCP_PFP_PRGRM_CNTR_START_HI,
+			pfp_hdr->ucode_start_addr_hi>>2);
+
+		/*
+		 * Program CP_ME_CNTL to reset given PIPE to take
+		 * effect of CP_PFP_PRGRM_CNTR_START.
+		 */
+		tmp = RREG32_SOC15(GC, 0, regCP_ME_CNTL);
+		if (pipe_id == 0)
+			tmp = REG_SET_FIELD(tmp, CP_ME_CNTL,
+					PFP_PIPE0_RESET, 1);
+		else
+			tmp = REG_SET_FIELD(tmp, CP_ME_CNTL,
+					PFP_PIPE1_RESET, 1);
+		WREG32_SOC15(GC, 0, regCP_ME_CNTL, tmp);
+
+		/* Clear pfp pipe0 reset bit. */
+		if (pipe_id == 0)
+			tmp = REG_SET_FIELD(tmp, CP_ME_CNTL,
+					PFP_PIPE0_RESET, 0);
+		else
+			tmp = REG_SET_FIELD(tmp, CP_ME_CNTL,
+					PFP_PIPE1_RESET, 0);
+		WREG32_SOC15(GC, 0, regCP_ME_CNTL, tmp);
+
+		WREG32_SOC15(GC, 0, regCP_GFX_RS64_DC_BASE0_LO,
+			lower_32_bits(adev->gfx.pfp.pfp_fw_data_gpu_addr));
+		WREG32_SOC15(GC, 0, regCP_GFX_RS64_DC_BASE0_HI,
+			upper_32_bits(adev->gfx.pfp.pfp_fw_data_gpu_addr));
+	}
+	soc21_grbm_select(adev, 0, 0, 0, 0);
+	mutex_unlock(&adev->srbm_mutex);
+
+	tmp = RREG32_SOC15(GC, 0, regCP_GFX_RS64_DC_BASE_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_GFX_RS64_DC_BASE_CNTL, VMID, 0);
+	tmp = REG_SET_FIELD(tmp, CP_GFX_RS64_DC_BASE_CNTL, CACHE_POLICY, 0);
+	WREG32_SOC15(GC, 0, regCP_GFX_RS64_DC_BASE_CNTL, tmp);
+
+	/* Invalidate the data caches */
+	tmp = RREG32_SOC15(GC, 0, regCP_GFX_RS64_DC_OP_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_GFX_RS64_DC_OP_CNTL, INVALIDATE_DCACHE, 1);
+	WREG32_SOC15(GC, 0, regCP_GFX_RS64_DC_OP_CNTL, tmp);
+
+	for (i = 0; i < usec_timeout; i++) {
+		tmp = RREG32_SOC15(GC, 0, regCP_GFX_RS64_DC_OP_CNTL);
+		if (1 == REG_GET_FIELD(tmp, CP_GFX_RS64_DC_OP_CNTL,
+			INVALIDATE_DCACHE_COMPLETE))
+			break;
+		udelay(1);
+	}
+
+	if (i >= usec_timeout) {
+		dev_err(adev->dev, "failed to invalidate RS64 data cache\n");
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int gfx_v11_0_cp_gfx_load_me_microcode(struct amdgpu_device *adev)
+{
+	int r;
+	const struct gfx_firmware_header_v1_0 *me_hdr;
+	const __le32 *fw_data;
+	unsigned i, fw_size;
+
+	me_hdr = (const struct gfx_firmware_header_v1_0 *)
+		adev->gfx.me_fw->data;
+
+	amdgpu_ucode_print_gfx_hdr(&me_hdr->header);
+
+	fw_data = (const __le32 *)(adev->gfx.me_fw->data +
+		le32_to_cpu(me_hdr->header.ucode_array_offset_bytes));
+	fw_size = le32_to_cpu(me_hdr->header.ucode_size_bytes);
+
+	r = amdgpu_bo_create_reserved(adev, me_hdr->header.ucode_size_bytes,
+				      PAGE_SIZE, AMDGPU_GEM_DOMAIN_GTT,
+				      &adev->gfx.me.me_fw_obj,
+				      &adev->gfx.me.me_fw_gpu_addr,
+				      (void **)&adev->gfx.me.me_fw_ptr);
+	if (r) {
+		dev_err(adev->dev, "(%d) failed to create me fw bo\n", r);
+		gfx_v11_0_me_fini(adev);
+		return r;
+	}
+
+	memcpy(adev->gfx.me.me_fw_ptr, fw_data, fw_size);
+
+	amdgpu_bo_kunmap(adev->gfx.me.me_fw_obj);
+	amdgpu_bo_unreserve(adev->gfx.me.me_fw_obj);
+
+	gfx_v11_0_config_me_cache(adev, adev->gfx.me.me_fw_gpu_addr);
+
+	WREG32_SOC15(GC, 0, regCP_HYP_ME_UCODE_ADDR, 0);
+
+	for (i = 0; i < me_hdr->jt_size; i++)
+		WREG32_SOC15(GC, 0, regCP_HYP_ME_UCODE_DATA,
+			     le32_to_cpup(fw_data + me_hdr->jt_offset + i));
+
+	WREG32_SOC15(GC, 0, regCP_HYP_ME_UCODE_ADDR, adev->gfx.me_fw_version);
+
+	return 0;
+}
+
+static int gfx_v11_0_cp_gfx_load_me_microcode_rs64(struct amdgpu_device *adev)
+{
+	int r;
+	const struct gfx_firmware_header_v2_0 *me_hdr;
+	const __le32 *fw_ucode, *fw_data;
+	unsigned i, pipe_id, fw_ucode_size, fw_data_size;
+	uint32_t tmp;
+	uint32_t usec_timeout = 50000;  /* wait for 50ms */
+
+	me_hdr = (const struct gfx_firmware_header_v2_0 *)
+		adev->gfx.me_fw->data;
+
+	amdgpu_ucode_print_gfx_hdr(&me_hdr->header);
+
+	/* instruction */
+	fw_ucode = (const __le32 *)(adev->gfx.me_fw->data +
+		le32_to_cpu(me_hdr->ucode_offset_bytes));
+	fw_ucode_size = le32_to_cpu(me_hdr->ucode_size_bytes);
+	/* data */
+	fw_data = (const __le32 *)(adev->gfx.me_fw->data +
+		le32_to_cpu(me_hdr->data_offset_bytes));
+	fw_data_size = le32_to_cpu(me_hdr->data_size_bytes);
+
+	/* 64kb align*/
+	r = amdgpu_bo_create_reserved(adev, fw_ucode_size,
+				      64 * 1024,
+				      AMDGPU_GEM_DOMAIN_VRAM |
+				      AMDGPU_GEM_DOMAIN_GTT,
+				      &adev->gfx.me.me_fw_obj,
+				      &adev->gfx.me.me_fw_gpu_addr,
+				      (void **)&adev->gfx.me.me_fw_ptr);
+	if (r) {
+		dev_err(adev->dev, "(%d) failed to create me ucode bo\n", r);
+		gfx_v11_0_me_fini(adev);
+		return r;
+	}
+
+	r = amdgpu_bo_create_reserved(adev, fw_data_size,
+				      64 * 1024,
+				      AMDGPU_GEM_DOMAIN_VRAM |
+				      AMDGPU_GEM_DOMAIN_GTT,
+				      &adev->gfx.me.me_fw_data_obj,
+				      &adev->gfx.me.me_fw_data_gpu_addr,
+				      (void **)&adev->gfx.me.me_fw_data_ptr);
+	if (r) {
+		dev_err(adev->dev, "(%d) failed to create me data bo\n", r);
+		gfx_v11_0_pfp_fini(adev);
+		return r;
+	}
+
+	memcpy(adev->gfx.me.me_fw_ptr, fw_ucode, fw_ucode_size);
+	memcpy(adev->gfx.me.me_fw_data_ptr, fw_data, fw_data_size);
+
+	amdgpu_bo_kunmap(adev->gfx.me.me_fw_obj);
+	amdgpu_bo_kunmap(adev->gfx.me.me_fw_data_obj);
+	amdgpu_bo_unreserve(adev->gfx.me.me_fw_obj);
+	amdgpu_bo_unreserve(adev->gfx.me.me_fw_data_obj);
+
+	if (amdgpu_emu_mode == 1)
+		amdgpu_device_flush_hdp(adev, NULL);
+
+	WREG32_SOC15(GC, 0, regCP_ME_IC_BASE_LO,
+		lower_32_bits(adev->gfx.me.me_fw_gpu_addr));
+	WREG32_SOC15(GC, 0, regCP_ME_IC_BASE_HI,
+		upper_32_bits(adev->gfx.me.me_fw_gpu_addr));
+
+	tmp = RREG32_SOC15(GC, 0, regCP_ME_IC_BASE_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_ME_IC_BASE_CNTL, VMID, 0);
+	tmp = REG_SET_FIELD(tmp, CP_ME_IC_BASE_CNTL, CACHE_POLICY, 0);
+	tmp = REG_SET_FIELD(tmp, CP_ME_IC_BASE_CNTL, EXE_DISABLE, 0);
+	WREG32_SOC15(GC, 0, regCP_ME_IC_BASE_CNTL, tmp);
+
+	/*
+	 * Programming any of the CP_ME_IC_BASE registers
+	 * forces invalidation of the ME L1 I$. Wait for the
+	 * invalidation complete
+	 */
+	for (i = 0; i < usec_timeout; i++) {
+		tmp = RREG32_SOC15(GC, 0, regCP_ME_IC_OP_CNTL);
+		if (1 == REG_GET_FIELD(tmp, CP_ME_IC_OP_CNTL,
+			INVALIDATE_CACHE_COMPLETE))
+			break;
+		udelay(1);
+	}
+
+	if (i >= usec_timeout) {
+		dev_err(adev->dev, "failed to invalidate instruction cache\n");
+		return -EINVAL;
+	}
+
+	/* Prime the instruction caches */
+	tmp = RREG32_SOC15(GC, 0, regCP_ME_IC_OP_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_ME_IC_OP_CNTL, PRIME_ICACHE, 1);
+	WREG32_SOC15(GC, 0, regCP_ME_IC_OP_CNTL, tmp);
+
+	/* Waiting for instruction cache primed*/
+	for (i = 0; i < usec_timeout; i++) {
+		tmp = RREG32_SOC15(GC, 0, regCP_ME_IC_OP_CNTL);
+		if (1 == REG_GET_FIELD(tmp, CP_ME_IC_OP_CNTL,
+			ICACHE_PRIMED))
+			break;
+		udelay(1);
+	}
+
+	if (i >= usec_timeout) {
+		dev_err(adev->dev, "failed to prime instruction cache\n");
+		return -EINVAL;
+	}
+
+	mutex_lock(&adev->srbm_mutex);
+	for (pipe_id = 0; pipe_id < adev->gfx.me.num_pipe_per_me; pipe_id++) {
+		soc21_grbm_select(adev, 0, pipe_id, 0, 0);
+		WREG32_SOC15(GC, 0, regCP_ME_PRGRM_CNTR_START,
+			(me_hdr->ucode_start_addr_hi << 30) |
+			(me_hdr->ucode_start_addr_lo >> 2) );
+		WREG32_SOC15(GC, 0, regCP_ME_PRGRM_CNTR_START_HI,
+			me_hdr->ucode_start_addr_hi>>2);
+
+		/*
+		 * Program CP_ME_CNTL to reset given PIPE to take
+		 * effect of CP_PFP_PRGRM_CNTR_START.
+		 */
+		tmp = RREG32_SOC15(GC, 0, regCP_ME_CNTL);
+		if (pipe_id == 0)
+			tmp = REG_SET_FIELD(tmp, CP_ME_CNTL,
+					ME_PIPE0_RESET, 1);
+		else
+			tmp = REG_SET_FIELD(tmp, CP_ME_CNTL,
+					ME_PIPE1_RESET, 1);
+		WREG32_SOC15(GC, 0, regCP_ME_CNTL, tmp);
+
+		/* Clear pfp pipe0 reset bit. */
+		if (pipe_id == 0)
+			tmp = REG_SET_FIELD(tmp, CP_ME_CNTL,
+					ME_PIPE0_RESET, 0);
+		else
+			tmp = REG_SET_FIELD(tmp, CP_ME_CNTL,
+					ME_PIPE1_RESET, 0);
+		WREG32_SOC15(GC, 0, regCP_ME_CNTL, tmp);
+
+		WREG32_SOC15(GC, 0, regCP_GFX_RS64_DC_BASE1_LO,
+			lower_32_bits(adev->gfx.me.me_fw_data_gpu_addr));
+		WREG32_SOC15(GC, 0, regCP_GFX_RS64_DC_BASE1_HI,
+			upper_32_bits(adev->gfx.me.me_fw_data_gpu_addr));
+	}
+	soc21_grbm_select(adev, 0, 0, 0, 0);
+	mutex_unlock(&adev->srbm_mutex);
+
+	tmp = RREG32_SOC15(GC, 0, regCP_GFX_RS64_DC_BASE_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_GFX_RS64_DC_BASE_CNTL, VMID, 0);
+	tmp = REG_SET_FIELD(tmp, CP_GFX_RS64_DC_BASE_CNTL, CACHE_POLICY, 0);
+	WREG32_SOC15(GC, 0, regCP_GFX_RS64_DC_BASE_CNTL, tmp);
+
+	/* Invalidate the data caches */
+	tmp = RREG32_SOC15(GC, 0, regCP_GFX_RS64_DC_OP_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_GFX_RS64_DC_OP_CNTL, INVALIDATE_DCACHE, 1);
+	WREG32_SOC15(GC, 0, regCP_GFX_RS64_DC_OP_CNTL, tmp);
+
+	for (i = 0; i < usec_timeout; i++) {
+		tmp = RREG32_SOC15(GC, 0, regCP_GFX_RS64_DC_OP_CNTL);
+		if (1 == REG_GET_FIELD(tmp, CP_GFX_RS64_DC_OP_CNTL,
+			INVALIDATE_DCACHE_COMPLETE))
+			break;
+		udelay(1);
+	}
+
+	if (i >= usec_timeout) {
+		dev_err(adev->dev, "failed to invalidate RS64 data cache\n");
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int gfx_v11_0_cp_gfx_load_microcode(struct amdgpu_device *adev)
+{
+	int r;
+
+	if (!adev->gfx.me_fw || !adev->gfx.pfp_fw)
+		return -EINVAL;
+
+	gfx_v11_0_cp_gfx_enable(adev, false);
+
+	if (adev->gfx.rs64_enable)
+		r = gfx_v11_0_cp_gfx_load_pfp_microcode_rs64(adev);
+	else
+		r = gfx_v11_0_cp_gfx_load_pfp_microcode(adev);
+	if (r) {
+		dev_err(adev->dev, "(%d) failed to load pfp fw\n", r);
+		return r;
+	}
+
+	if (adev->gfx.rs64_enable)
+		r = gfx_v11_0_cp_gfx_load_me_microcode_rs64(adev);
+	else
+		r = gfx_v11_0_cp_gfx_load_me_microcode(adev);
+	if (r) {
+		dev_err(adev->dev, "(%d) failed to load me fw\n", r);
+		return r;
+	}
+
+	return 0;
+}
+
+static int gfx_v11_0_cp_gfx_start(struct amdgpu_device *adev)
+{
+	struct amdgpu_ring *ring;
+	const struct cs_section_def *sect = NULL;
+	const struct cs_extent_def *ext = NULL;
+	int r, i;
+	int ctx_reg_offset;
+
+	/* init the CP */
+	WREG32_SOC15(GC, 0, regCP_MAX_CONTEXT,
+		     adev->gfx.config.max_hw_contexts - 1);
+	WREG32_SOC15(GC, 0, regCP_DEVICE_ID, 1);
+
+	if (!amdgpu_async_gfx_ring)
+		gfx_v11_0_cp_gfx_enable(adev, true);
+
+	ring = &adev->gfx.gfx_ring[0];
+	r = amdgpu_ring_alloc(ring, gfx_v11_0_get_csb_size(adev));
+	if (r) {
+		DRM_ERROR("amdgpu: cp failed to lock ring (%d).\n", r);
+		return r;
+	}
+
+	amdgpu_ring_write(ring, PACKET3(PACKET3_PREAMBLE_CNTL, 0));
+	amdgpu_ring_write(ring, PACKET3_PREAMBLE_BEGIN_CLEAR_STATE);
+
+	amdgpu_ring_write(ring, PACKET3(PACKET3_CONTEXT_CONTROL, 1));
+	amdgpu_ring_write(ring, 0x80000000);
+	amdgpu_ring_write(ring, 0x80000000);
+
+	for (sect = gfx11_cs_data; sect->section != NULL; ++sect) {
+		for (ext = sect->section; ext->extent != NULL; ++ext) {
+			if (sect->id == SECT_CONTEXT) {
+				amdgpu_ring_write(ring,
+						  PACKET3(PACKET3_SET_CONTEXT_REG,
+							  ext->reg_count));
+				amdgpu_ring_write(ring, ext->reg_index -
+						  PACKET3_SET_CONTEXT_REG_START);
+				for (i = 0; i < ext->reg_count; i++)
+					amdgpu_ring_write(ring, ext->extent[i]);
+			}
+		}
+	}
+
+	ctx_reg_offset =
+		SOC15_REG_OFFSET(GC, 0, regPA_SC_TILE_STEERING_OVERRIDE) - PACKET3_SET_CONTEXT_REG_START;
+	amdgpu_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 1));
+	amdgpu_ring_write(ring, ctx_reg_offset);
+	amdgpu_ring_write(ring, adev->gfx.config.pa_sc_tile_steering_override);
+
+	amdgpu_ring_write(ring, PACKET3(PACKET3_PREAMBLE_CNTL, 0));
+	amdgpu_ring_write(ring, PACKET3_PREAMBLE_END_CLEAR_STATE);
+
+	amdgpu_ring_write(ring, PACKET3(PACKET3_CLEAR_STATE, 0));
+	amdgpu_ring_write(ring, 0);
+
+	amdgpu_ring_commit(ring);
+
+	/* submit cs packet to copy state 0 to next available state */
+	if (adev->gfx.num_gfx_rings > 1) {
+		/* maximum supported gfx ring is 2 */
+		ring = &adev->gfx.gfx_ring[1];
+		r = amdgpu_ring_alloc(ring, 2);
+		if (r) {
+			DRM_ERROR("amdgpu: cp failed to lock ring (%d).\n", r);
+			return r;
+		}
+
+		amdgpu_ring_write(ring, PACKET3(PACKET3_CLEAR_STATE, 0));
+		amdgpu_ring_write(ring, 0);
+
+		amdgpu_ring_commit(ring);
+	}
+	return 0;
+}
+
+static void gfx_v11_0_cp_gfx_switch_pipe(struct amdgpu_device *adev,
+					 CP_PIPE_ID pipe)
+{
+	u32 tmp;
+
+	tmp = RREG32_SOC15(GC, 0, regGRBM_GFX_CNTL);
+	tmp = REG_SET_FIELD(tmp, GRBM_GFX_CNTL, PIPEID, pipe);
+
+	WREG32_SOC15(GC, 0, regGRBM_GFX_CNTL, tmp);
+}
+
+static void gfx_v11_0_cp_gfx_set_doorbell(struct amdgpu_device *adev,
+					  struct amdgpu_ring *ring)
+{
+	u32 tmp;
+
+	tmp = RREG32_SOC15(GC, 0, regCP_RB_DOORBELL_CONTROL);
+	if (ring->use_doorbell) {
+		tmp = REG_SET_FIELD(tmp, CP_RB_DOORBELL_CONTROL,
+				    DOORBELL_OFFSET, ring->doorbell_index);
+		tmp = REG_SET_FIELD(tmp, CP_RB_DOORBELL_CONTROL,
+				    DOORBELL_EN, 1);
+	} else {
+		tmp = REG_SET_FIELD(tmp, CP_RB_DOORBELL_CONTROL,
+				    DOORBELL_EN, 0);
+	}
+	WREG32_SOC15(GC, 0, regCP_RB_DOORBELL_CONTROL, tmp);
+
+	tmp = REG_SET_FIELD(0, CP_RB_DOORBELL_RANGE_LOWER,
+			    DOORBELL_RANGE_LOWER, ring->doorbell_index);
+	WREG32_SOC15(GC, 0, regCP_RB_DOORBELL_RANGE_LOWER, tmp);
+
+	WREG32_SOC15(GC, 0, regCP_RB_DOORBELL_RANGE_UPPER,
+		     CP_RB_DOORBELL_RANGE_UPPER__DOORBELL_RANGE_UPPER_MASK);
+}
+
+static int gfx_v11_0_cp_gfx_resume(struct amdgpu_device *adev)
+{
+	struct amdgpu_ring *ring;
+	u32 tmp;
+	u32 rb_bufsz;
+	u64 rb_addr, rptr_addr, wptr_gpu_addr;
+
+	/* Set the write pointer delay */
+	WREG32_SOC15(GC, 0, regCP_RB_WPTR_DELAY, 0);
+
+	/* set the RB to use vmid 0 */
+	WREG32_SOC15(GC, 0, regCP_RB_VMID, 0);
+
+	/* Init gfx ring 0 for pipe 0 */
+	mutex_lock(&adev->srbm_mutex);
+	gfx_v11_0_cp_gfx_switch_pipe(adev, PIPE_ID0);
+
+	/* Set ring buffer size */
+	ring = &adev->gfx.gfx_ring[0];
+	rb_bufsz = order_base_2(ring->ring_size / 8);
+	tmp = REG_SET_FIELD(0, CP_RB0_CNTL, RB_BUFSZ, rb_bufsz);
+	tmp = REG_SET_FIELD(tmp, CP_RB0_CNTL, RB_BLKSZ, rb_bufsz - 2);
+	WREG32_SOC15(GC, 0, regCP_RB0_CNTL, tmp);
+
+	/* Initialize the ring buffer's write pointers */
+	ring->wptr = 0;
+	WREG32_SOC15(GC, 0, regCP_RB0_WPTR, lower_32_bits(ring->wptr));
+	WREG32_SOC15(GC, 0, regCP_RB0_WPTR_HI, upper_32_bits(ring->wptr));
+
+	/* set the wb address whether it's enabled or not */
+	rptr_addr = ring->rptr_gpu_addr;
+	WREG32_SOC15(GC, 0, regCP_RB0_RPTR_ADDR, lower_32_bits(rptr_addr));
+	WREG32_SOC15(GC, 0, regCP_RB0_RPTR_ADDR_HI, upper_32_bits(rptr_addr) &
+		     CP_RB_RPTR_ADDR_HI__RB_RPTR_ADDR_HI_MASK);
+
+	wptr_gpu_addr = ring->wptr_gpu_addr;
+	WREG32_SOC15(GC, 0, regCP_RB_WPTR_POLL_ADDR_LO,
+		     lower_32_bits(wptr_gpu_addr));
+	WREG32_SOC15(GC, 0, regCP_RB_WPTR_POLL_ADDR_HI,
+		     upper_32_bits(wptr_gpu_addr));
+
+	mdelay(1);
+	WREG32_SOC15(GC, 0, regCP_RB0_CNTL, tmp);
+
+	rb_addr = ring->gpu_addr >> 8;
+	WREG32_SOC15(GC, 0, regCP_RB0_BASE, rb_addr);
+	WREG32_SOC15(GC, 0, regCP_RB0_BASE_HI, upper_32_bits(rb_addr));
+
+	WREG32_SOC15(GC, 0, regCP_RB_ACTIVE, 1);
+
+	gfx_v11_0_cp_gfx_set_doorbell(adev, ring);
+	mutex_unlock(&adev->srbm_mutex);
+
+	/* Init gfx ring 1 for pipe 1 */
+	if (adev->gfx.num_gfx_rings > 1) {
+		mutex_lock(&adev->srbm_mutex);
+		gfx_v11_0_cp_gfx_switch_pipe(adev, PIPE_ID1);
+		/* maximum supported gfx ring is 2 */
+		ring = &adev->gfx.gfx_ring[1];
+		rb_bufsz = order_base_2(ring->ring_size / 8);
+		tmp = REG_SET_FIELD(0, CP_RB1_CNTL, RB_BUFSZ, rb_bufsz);
+		tmp = REG_SET_FIELD(tmp, CP_RB1_CNTL, RB_BLKSZ, rb_bufsz - 2);
+		WREG32_SOC15(GC, 0, regCP_RB1_CNTL, tmp);
+		/* Initialize the ring buffer's write pointers */
+		ring->wptr = 0;
+		WREG32_SOC15(GC, 0, regCP_RB1_WPTR, lower_32_bits(ring->wptr));
+		WREG32_SOC15(GC, 0, regCP_RB1_WPTR_HI, upper_32_bits(ring->wptr));
+		/* Set the wb address whether it's enabled or not */
+		rptr_addr = ring->rptr_gpu_addr;
+		WREG32_SOC15(GC, 0, regCP_RB1_RPTR_ADDR, lower_32_bits(rptr_addr));
+		WREG32_SOC15(GC, 0, regCP_RB1_RPTR_ADDR_HI, upper_32_bits(rptr_addr) &
+			     CP_RB1_RPTR_ADDR_HI__RB_RPTR_ADDR_HI_MASK);
+		wptr_gpu_addr = ring->wptr_gpu_addr;
+		WREG32_SOC15(GC, 0, regCP_RB_WPTR_POLL_ADDR_LO,
+			     lower_32_bits(wptr_gpu_addr));
+		WREG32_SOC15(GC, 0, regCP_RB_WPTR_POLL_ADDR_HI,
+			     upper_32_bits(wptr_gpu_addr));
+
+		mdelay(1);
+		WREG32_SOC15(GC, 0, regCP_RB1_CNTL, tmp);
+
+		rb_addr = ring->gpu_addr >> 8;
+		WREG32_SOC15(GC, 0, regCP_RB1_BASE, rb_addr);
+		WREG32_SOC15(GC, 0, regCP_RB1_BASE_HI, upper_32_bits(rb_addr));
+		WREG32_SOC15(GC, 0, regCP_RB1_ACTIVE, 1);
+
+		gfx_v11_0_cp_gfx_set_doorbell(adev, ring);
+		mutex_unlock(&adev->srbm_mutex);
+	}
+	/* Switch to pipe 0 */
+	mutex_lock(&adev->srbm_mutex);
+	gfx_v11_0_cp_gfx_switch_pipe(adev, PIPE_ID0);
+	mutex_unlock(&adev->srbm_mutex);
+
+	/* start the ring */
+	gfx_v11_0_cp_gfx_start(adev);
+
+	return 0;
+}
+
+static void gfx_v11_0_cp_compute_enable(struct amdgpu_device *adev, bool enable)
+{
+	u32 data;
+
+	if (adev->gfx.rs64_enable) {
+		data = RREG32_SOC15(GC, 0, regCP_MEC_RS64_CNTL);
+		data = REG_SET_FIELD(data, CP_MEC_RS64_CNTL, MEC_INVALIDATE_ICACHE,
+							 enable ? 0 : 1);
+		data = REG_SET_FIELD(data, CP_MEC_RS64_CNTL, MEC_PIPE0_RESET,
+							 enable ? 0 : 1);
+		data = REG_SET_FIELD(data, CP_MEC_RS64_CNTL, MEC_PIPE1_RESET,
+							 enable ? 0 : 1);
+		data = REG_SET_FIELD(data, CP_MEC_RS64_CNTL, MEC_PIPE2_RESET,
+							 enable ? 0 : 1);
+		data = REG_SET_FIELD(data, CP_MEC_RS64_CNTL, MEC_PIPE3_RESET,
+							 enable ? 0 : 1);
+		data = REG_SET_FIELD(data, CP_MEC_RS64_CNTL, MEC_PIPE0_ACTIVE,
+							 enable ? 1 : 0);
+		data = REG_SET_FIELD(data, CP_MEC_RS64_CNTL, MEC_PIPE1_ACTIVE,
+				                         enable ? 1 : 0);
+		data = REG_SET_FIELD(data, CP_MEC_RS64_CNTL, MEC_PIPE2_ACTIVE,
+							 enable ? 1 : 0);
+		data = REG_SET_FIELD(data, CP_MEC_RS64_CNTL, MEC_PIPE3_ACTIVE,
+							 enable ? 1 : 0);
+		data = REG_SET_FIELD(data, CP_MEC_RS64_CNTL, MEC_HALT,
+							 enable ? 0 : 1);
+		WREG32_SOC15(GC, 0, regCP_MEC_RS64_CNTL, data);
+	} else {
+		data = RREG32_SOC15(GC, 0, regCP_MEC_CNTL);
+
+		if (enable) {
+			data = REG_SET_FIELD(data, CP_MEC_CNTL, MEC_ME1_HALT, 0);
+			if (!adev->enable_mes_kiq)
+				data = REG_SET_FIELD(data, CP_MEC_CNTL,
+						     MEC_ME2_HALT, 0);
+		} else {
+			data = REG_SET_FIELD(data, CP_MEC_CNTL, MEC_ME1_HALT, 1);
+			data = REG_SET_FIELD(data, CP_MEC_CNTL, MEC_ME2_HALT, 1);
+		}
+		WREG32_SOC15(GC, 0, regCP_MEC_CNTL, data);
+	}
+
+	udelay(50);
+}
+
+static int gfx_v11_0_cp_compute_load_microcode(struct amdgpu_device *adev)
+{
+	const struct gfx_firmware_header_v1_0 *mec_hdr;
+	const __le32 *fw_data;
+	unsigned i, fw_size;
+	u32 *fw = NULL;
+	int r;
+
+	if (!adev->gfx.mec_fw)
+		return -EINVAL;
+
+	gfx_v11_0_cp_compute_enable(adev, false);
+
+	mec_hdr = (const struct gfx_firmware_header_v1_0 *)adev->gfx.mec_fw->data;
+	amdgpu_ucode_print_gfx_hdr(&mec_hdr->header);
+
+	fw_data = (const __le32 *)
+		(adev->gfx.mec_fw->data +
+		 le32_to_cpu(mec_hdr->header.ucode_array_offset_bytes));
+	fw_size = le32_to_cpu(mec_hdr->header.ucode_size_bytes);
+
+	r = amdgpu_bo_create_reserved(adev, mec_hdr->header.ucode_size_bytes,
+					  PAGE_SIZE, AMDGPU_GEM_DOMAIN_GTT,
+					  &adev->gfx.mec.mec_fw_obj,
+					  &adev->gfx.mec.mec_fw_gpu_addr,
+					  (void **)&fw);
+	if (r) {
+		dev_err(adev->dev, "(%d) failed to create mec fw bo\n", r);
+		gfx_v11_0_mec_fini(adev);
+		return r;
+	}
+
+	memcpy(fw, fw_data, fw_size);
+
+	amdgpu_bo_kunmap(adev->gfx.mec.mec_fw_obj);
+	amdgpu_bo_unreserve(adev->gfx.mec.mec_fw_obj);
+
+	gfx_v11_0_config_mec_cache(adev, adev->gfx.mec.mec_fw_gpu_addr);
+
+	/* MEC1 */
+	WREG32_SOC15(GC, 0, regCP_MEC_ME1_UCODE_ADDR, 0);
+
+	for (i = 0; i < mec_hdr->jt_size; i++)
+		WREG32_SOC15(GC, 0, regCP_MEC_ME1_UCODE_DATA,
+			     le32_to_cpup(fw_data + mec_hdr->jt_offset + i));
+
+	WREG32_SOC15(GC, 0, regCP_MEC_ME1_UCODE_ADDR, adev->gfx.mec_fw_version);
+
+	return 0;
+}
+
+static int gfx_v11_0_cp_compute_load_microcode_rs64(struct amdgpu_device *adev)
+{
+	const struct gfx_firmware_header_v2_0 *mec_hdr;
+	const __le32 *fw_ucode, *fw_data;
+	u32 tmp, fw_ucode_size, fw_data_size;
+	u32 i, usec_timeout = 50000; /* Wait for 50 ms */
+	u32 *fw_ucode_ptr, *fw_data_ptr;
+	int r;
+
+	if (!adev->gfx.mec_fw)
+		return -EINVAL;
+
+	gfx_v11_0_cp_compute_enable(adev, false);
+
+	mec_hdr = (const struct gfx_firmware_header_v2_0 *)adev->gfx.mec_fw->data;
+	amdgpu_ucode_print_gfx_hdr(&mec_hdr->header);
+
+	fw_ucode = (const __le32 *) (adev->gfx.mec_fw->data +
+				le32_to_cpu(mec_hdr->ucode_offset_bytes));
+	fw_ucode_size = le32_to_cpu(mec_hdr->ucode_size_bytes);
+
+	fw_data = (const __le32 *) (adev->gfx.mec_fw->data +
+				le32_to_cpu(mec_hdr->data_offset_bytes));
+	fw_data_size = le32_to_cpu(mec_hdr->data_size_bytes);
+
+	r = amdgpu_bo_create_reserved(adev, fw_ucode_size,
+				      64 * 1024,
+				      AMDGPU_GEM_DOMAIN_VRAM |
+				      AMDGPU_GEM_DOMAIN_GTT,
+				      &adev->gfx.mec.mec_fw_obj,
+				      &adev->gfx.mec.mec_fw_gpu_addr,
+				      (void **)&fw_ucode_ptr);
+	if (r) {
+		dev_err(adev->dev, "(%d) failed to create mec fw ucode bo\n", r);
+		gfx_v11_0_mec_fini(adev);
+		return r;
+	}
+
+	r = amdgpu_bo_create_reserved(adev, fw_data_size,
+				      64 * 1024,
+				      AMDGPU_GEM_DOMAIN_VRAM |
+				      AMDGPU_GEM_DOMAIN_GTT,
+				      &adev->gfx.mec.mec_fw_data_obj,
+				      &adev->gfx.mec.mec_fw_data_gpu_addr,
+				      (void **)&fw_data_ptr);
+	if (r) {
+		dev_err(adev->dev, "(%d) failed to create mec fw ucode bo\n", r);
+		gfx_v11_0_mec_fini(adev);
+		return r;
+	}
+
+	memcpy(fw_ucode_ptr, fw_ucode, fw_ucode_size);
+	memcpy(fw_data_ptr, fw_data, fw_data_size);
+
+	amdgpu_bo_kunmap(adev->gfx.mec.mec_fw_obj);
+	amdgpu_bo_kunmap(adev->gfx.mec.mec_fw_data_obj);
+	amdgpu_bo_unreserve(adev->gfx.mec.mec_fw_obj);
+	amdgpu_bo_unreserve(adev->gfx.mec.mec_fw_data_obj);
+
+	tmp = RREG32_SOC15(GC, 0, regCP_CPC_IC_BASE_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_CPC_IC_BASE_CNTL, VMID, 0);
+	tmp = REG_SET_FIELD(tmp, CP_CPC_IC_BASE_CNTL, EXE_DISABLE, 0);
+	tmp = REG_SET_FIELD(tmp, CP_CPC_IC_BASE_CNTL, CACHE_POLICY, 0);
+	WREG32_SOC15(GC, 0, regCP_CPC_IC_BASE_CNTL, tmp);
+
+	tmp = RREG32_SOC15(GC, 0, regCP_MEC_DC_BASE_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_MEC_DC_BASE_CNTL, VMID, 0);
+	tmp = REG_SET_FIELD(tmp, CP_MEC_DC_BASE_CNTL, CACHE_POLICY, 0);
+	WREG32_SOC15(GC, 0, regCP_MEC_DC_BASE_CNTL, tmp);
+
+	mutex_lock(&adev->srbm_mutex);
+	for (i = 0; i < adev->gfx.mec.num_pipe_per_mec; i++) {
+		soc21_grbm_select(adev, 1, i, 0, 0);
+
+		WREG32_SOC15(GC, 0, regCP_MEC_MDBASE_LO, adev->gfx.mec.mec_fw_data_gpu_addr);
+		WREG32_SOC15(GC, 0, regCP_MEC_MDBASE_HI,
+		     upper_32_bits(adev->gfx.mec.mec_fw_data_gpu_addr));
+
+		WREG32_SOC15(GC, 0, regCP_MEC_RS64_PRGRM_CNTR_START,
+					mec_hdr->ucode_start_addr_lo >> 2 |
+					mec_hdr->ucode_start_addr_hi << 30);
+		WREG32_SOC15(GC, 0, regCP_MEC_RS64_PRGRM_CNTR_START_HI,
+					mec_hdr->ucode_start_addr_hi >> 2);
+
+		WREG32_SOC15(GC, 0, regCP_CPC_IC_BASE_LO, adev->gfx.mec.mec_fw_gpu_addr);
+		WREG32_SOC15(GC, 0, regCP_CPC_IC_BASE_HI,
+		     upper_32_bits(adev->gfx.mec.mec_fw_gpu_addr));
+	}
+	mutex_unlock(&adev->srbm_mutex);
+	soc21_grbm_select(adev, 0, 0, 0, 0);
+
+	/* Trigger an invalidation of the L1 instruction caches */
+	tmp = RREG32_SOC15(GC, 0, regCP_MEC_DC_OP_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_MEC_DC_OP_CNTL, INVALIDATE_DCACHE, 1);
+	WREG32_SOC15(GC, 0, regCP_MEC_DC_OP_CNTL, tmp);
+
+	/* Wait for invalidation complete */
+	for (i = 0; i < usec_timeout; i++) {
+		tmp = RREG32_SOC15(GC, 0, regCP_MEC_DC_OP_CNTL);
+		if (1 == REG_GET_FIELD(tmp, CP_MEC_DC_OP_CNTL,
+				       INVALIDATE_DCACHE_COMPLETE))
+			break;
+		udelay(1);
+	}
+
+	if (i >= usec_timeout) {
+		dev_err(adev->dev, "failed to invalidate instruction cache\n");
+		return -EINVAL;
+	}
+
+	/* Trigger an invalidation of the L1 instruction caches */
+	tmp = RREG32_SOC15(GC, 0, regCP_CPC_IC_OP_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_CPC_IC_OP_CNTL, INVALIDATE_CACHE, 1);
+	WREG32_SOC15(GC, 0, regCP_CPC_IC_OP_CNTL, tmp);
+
+	/* Wait for invalidation complete */
+	for (i = 0; i < usec_timeout; i++) {
+		tmp = RREG32_SOC15(GC, 0, regCP_CPC_IC_OP_CNTL);
+		if (1 == REG_GET_FIELD(tmp, CP_CPC_IC_OP_CNTL,
+				       INVALIDATE_CACHE_COMPLETE))
+			break;
+		udelay(1);
+	}
+
+	if (i >= usec_timeout) {
+		dev_err(adev->dev, "failed to invalidate instruction cache\n");
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static void gfx_v11_0_kiq_setting(struct amdgpu_ring *ring)
+{
+	uint32_t tmp;
+	struct amdgpu_device *adev = ring->adev;
+
+	/* tell RLC which is KIQ queue */
+	tmp = RREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS);
+	tmp &= 0xffffff00;
+	tmp |= (ring->me << 5) | (ring->pipe << 3) | (ring->queue);
+	WREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS, tmp | 0x80);
+}
+
+static void gfx_v11_0_cp_set_doorbell_range(struct amdgpu_device *adev)
+{
+	/* set graphics engine doorbell range */
+	WREG32_SOC15(GC, 0, regCP_RB_DOORBELL_RANGE_LOWER,
+		     (adev->doorbell_index.gfx_ring0 * 2) << 2);
+	WREG32_SOC15(GC, 0, regCP_RB_DOORBELL_RANGE_UPPER,
+		     (adev->doorbell_index.gfx_userqueue_end * 2) << 2);
+
+	/* set compute engine doorbell range */
+	WREG32_SOC15(GC, 0, regCP_MEC_DOORBELL_RANGE_LOWER,
+		     (adev->doorbell_index.kiq * 2) << 2);
+	WREG32_SOC15(GC, 0, regCP_MEC_DOORBELL_RANGE_UPPER,
+		     (adev->doorbell_index.userqueue_end * 2) << 2);
+}
+
+static void gfx_v11_0_gfx_mqd_set_priority(struct amdgpu_device *adev,
+					   struct v11_gfx_mqd *mqd,
+					   struct amdgpu_mqd_prop *prop)
+{
+	bool priority = 0;
+	u32 tmp;
+
+	/* set up default queue priority level
+	 * 0x0 = low priority, 0x1 = high priority
+	 */
+	if (prop->hqd_pipe_priority == AMDGPU_GFX_PIPE_PRIO_HIGH)
+		priority = 1;
+
+	tmp = regCP_GFX_HQD_QUEUE_PRIORITY_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, CP_GFX_HQD_QUEUE_PRIORITY, PRIORITY_LEVEL, priority);
+	mqd->cp_gfx_hqd_queue_priority = tmp;
+}
+
+static int gfx_v11_0_gfx_mqd_init(struct amdgpu_device *adev, void *m,
+				  struct amdgpu_mqd_prop *prop)
+{
+	struct v11_gfx_mqd *mqd = m;
+	uint64_t hqd_gpu_addr, wb_gpu_addr;
+	uint32_t tmp;
+	uint32_t rb_bufsz;
+
+	/* set up gfx hqd wptr */
+	mqd->cp_gfx_hqd_wptr = 0;
+	mqd->cp_gfx_hqd_wptr_hi = 0;
+
+	/* set the pointer to the MQD */
+	mqd->cp_mqd_base_addr = prop->mqd_gpu_addr & 0xfffffffc;
+	mqd->cp_mqd_base_addr_hi = upper_32_bits(prop->mqd_gpu_addr);
+
+	/* set up mqd control */
+	tmp = regCP_GFX_MQD_CONTROL_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, CP_GFX_MQD_CONTROL, VMID, 0);
+	tmp = REG_SET_FIELD(tmp, CP_GFX_MQD_CONTROL, PRIV_STATE, 1);
+	tmp = REG_SET_FIELD(tmp, CP_GFX_MQD_CONTROL, CACHE_POLICY, 0);
+	mqd->cp_gfx_mqd_control = tmp;
+
+	/* set up gfx_hqd_vimd with 0x0 to indicate the ring buffer's vmid */
+	tmp = regCP_GFX_HQD_VMID_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, CP_GFX_HQD_VMID, VMID, 0);
+	mqd->cp_gfx_hqd_vmid = 0;
+
+	/* set up gfx queue priority */
+	gfx_v11_0_gfx_mqd_set_priority(adev, mqd, prop);
+
+	/* set up time quantum */
+	tmp = regCP_GFX_HQD_QUANTUM_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, CP_GFX_HQD_QUANTUM, QUANTUM_EN, 1);
+	mqd->cp_gfx_hqd_quantum = tmp;
+
+	/* set up gfx hqd base. this is similar as CP_RB_BASE */
+	hqd_gpu_addr = prop->hqd_base_gpu_addr >> 8;
+	mqd->cp_gfx_hqd_base = hqd_gpu_addr;
+	mqd->cp_gfx_hqd_base_hi = upper_32_bits(hqd_gpu_addr);
+
+	/* set up hqd_rptr_addr/_hi, similar as CP_RB_RPTR */
+	wb_gpu_addr = prop->rptr_gpu_addr;
+	mqd->cp_gfx_hqd_rptr_addr = wb_gpu_addr & 0xfffffffc;
+	mqd->cp_gfx_hqd_rptr_addr_hi =
+		upper_32_bits(wb_gpu_addr) & 0xffff;
+
+	/* set up rb_wptr_poll addr */
+	wb_gpu_addr = prop->wptr_gpu_addr;
+	mqd->cp_rb_wptr_poll_addr_lo = wb_gpu_addr & 0xfffffffc;
+	mqd->cp_rb_wptr_poll_addr_hi = upper_32_bits(wb_gpu_addr) & 0xffff;
+
+	/* set up the gfx_hqd_control, similar as CP_RB0_CNTL */
+	rb_bufsz = order_base_2(prop->queue_size / 4) - 1;
+	tmp = regCP_GFX_HQD_CNTL_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, CP_GFX_HQD_CNTL, RB_BUFSZ, rb_bufsz);
+	tmp = REG_SET_FIELD(tmp, CP_GFX_HQD_CNTL, RB_BLKSZ, rb_bufsz - 2);
+#ifdef __BIG_ENDIAN
+	tmp = REG_SET_FIELD(tmp, CP_GFX_HQD_CNTL, BUF_SWAP, 1);
+#endif
+	if (prop->tmz_queue)
+		tmp = REG_SET_FIELD(tmp, CP_GFX_HQD_CNTL, TMZ_MATCH, 1);
+	if (!prop->kernel_queue)
+		tmp = REG_SET_FIELD(tmp, CP_GFX_HQD_CNTL, RB_NON_PRIV, 1);
+	mqd->cp_gfx_hqd_cntl = tmp;
+
+	/* set up cp_doorbell_control */
+	tmp = regCP_RB_DOORBELL_CONTROL_DEFAULT;
+	if (prop->use_doorbell) {
+		tmp = REG_SET_FIELD(tmp, CP_RB_DOORBELL_CONTROL,
+				    DOORBELL_OFFSET, prop->doorbell_index);
+		tmp = REG_SET_FIELD(tmp, CP_RB_DOORBELL_CONTROL,
+				    DOORBELL_EN, 1);
+	} else
+		tmp = REG_SET_FIELD(tmp, CP_RB_DOORBELL_CONTROL,
+				    DOORBELL_EN, 0);
+	mqd->cp_rb_doorbell_control = tmp;
+
+	/* reset read and write pointers, similar to CP_RB0_WPTR/_RPTR */
+	mqd->cp_gfx_hqd_rptr = regCP_GFX_HQD_RPTR_DEFAULT;
+
+	/* active the queue */
+	mqd->cp_gfx_hqd_active = 1;
+
+	/* set gfx UQ items */
+	mqd->shadow_base_lo = lower_32_bits(prop->shadow_addr);
+	mqd->shadow_base_hi = upper_32_bits(prop->shadow_addr);
+	mqd->gds_bkup_base_lo = lower_32_bits(prop->gds_bkup_addr);
+	mqd->gds_bkup_base_hi = upper_32_bits(prop->gds_bkup_addr);
+	mqd->fw_work_area_base_lo = lower_32_bits(prop->csa_addr);
+	mqd->fw_work_area_base_hi = upper_32_bits(prop->csa_addr);
+	mqd->fence_address_lo = lower_32_bits(prop->fence_address);
+	mqd->fence_address_hi = upper_32_bits(prop->fence_address);
+
+	return 0;
+}
+
+static int gfx_v11_0_kgq_init_queue(struct amdgpu_ring *ring, bool reset)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct v11_gfx_mqd *mqd = ring->mqd_ptr;
+	int mqd_idx = ring - &adev->gfx.gfx_ring[0];
+
+	if (!reset && !amdgpu_in_reset(adev) && !adev->in_suspend) {
+		memset((void *)mqd, 0, sizeof(*mqd));
+		mutex_lock(&adev->srbm_mutex);
+		soc21_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0);
+		amdgpu_ring_init_mqd(ring);
+		soc21_grbm_select(adev, 0, 0, 0, 0);
+		mutex_unlock(&adev->srbm_mutex);
+		if (adev->gfx.me.mqd_backup[mqd_idx])
+			memcpy_fromio(adev->gfx.me.mqd_backup[mqd_idx], mqd, sizeof(*mqd));
+	} else {
+		/* restore mqd with the backup copy */
+		if (adev->gfx.me.mqd_backup[mqd_idx])
+			memcpy_toio(mqd, adev->gfx.me.mqd_backup[mqd_idx], sizeof(*mqd));
+		/* reset the ring */
+		ring->wptr = 0;
+		*ring->wptr_cpu_addr = 0;
+		amdgpu_ring_clear_ring(ring);
+	}
+
+	return 0;
+}
+
+static int gfx_v11_0_cp_async_gfx_ring_resume(struct amdgpu_device *adev)
+{
+	int r, i;
+
+	for (i = 0; i < adev->gfx.num_gfx_rings; i++) {
+		r = gfx_v11_0_kgq_init_queue(&adev->gfx.gfx_ring[i], false);
+		if (r)
+			return r;
+	}
+
+	r = amdgpu_gfx_enable_kgq(adev, 0);
+	if (r)
+		return r;
+
+	return gfx_v11_0_cp_gfx_start(adev);
+}
+
+static int gfx_v11_0_compute_mqd_init(struct amdgpu_device *adev, void *m,
+				      struct amdgpu_mqd_prop *prop)
+{
+	struct v11_compute_mqd *mqd = m;
+	uint64_t hqd_gpu_addr, wb_gpu_addr, eop_base_addr;
+	uint32_t tmp;
+
+	mqd->header = 0xC0310800;
+	mqd->compute_pipelinestat_enable = 0x00000001;
+	mqd->compute_static_thread_mgmt_se0 = 0xffffffff;
+	mqd->compute_static_thread_mgmt_se1 = 0xffffffff;
+	mqd->compute_static_thread_mgmt_se2 = 0xffffffff;
+	mqd->compute_static_thread_mgmt_se3 = 0xffffffff;
+	mqd->compute_misc_reserved = 0x00000007;
+
+	eop_base_addr = prop->eop_gpu_addr >> 8;
+	mqd->cp_hqd_eop_base_addr_lo = eop_base_addr;
+	mqd->cp_hqd_eop_base_addr_hi = upper_32_bits(eop_base_addr);
+
+	/* set the EOP size, register value is 2^(EOP_SIZE+1) dwords */
+	tmp = regCP_HQD_EOP_CONTROL_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, CP_HQD_EOP_CONTROL, EOP_SIZE,
+			(order_base_2(GFX11_MEC_HPD_SIZE / 4) - 1));
+
+	mqd->cp_hqd_eop_control = tmp;
+
+	/* enable doorbell? */
+	tmp = regCP_HQD_PQ_DOORBELL_CONTROL_DEFAULT;
+
+	if (prop->use_doorbell) {
+		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
+				    DOORBELL_OFFSET, prop->doorbell_index);
+		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
+				    DOORBELL_EN, 1);
+		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
+				    DOORBELL_SOURCE, 0);
+		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
+				    DOORBELL_HIT, 0);
+	} else {
+		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
+				    DOORBELL_EN, 0);
+	}
+
+	mqd->cp_hqd_pq_doorbell_control = tmp;
+
+	/* disable the queue if it's active */
+	mqd->cp_hqd_dequeue_request = 0;
+	mqd->cp_hqd_pq_rptr = 0;
+	mqd->cp_hqd_pq_wptr_lo = 0;
+	mqd->cp_hqd_pq_wptr_hi = 0;
+
+	/* set the pointer to the MQD */
+	mqd->cp_mqd_base_addr_lo = prop->mqd_gpu_addr & 0xfffffffc;
+	mqd->cp_mqd_base_addr_hi = upper_32_bits(prop->mqd_gpu_addr);
+
+	/* set MQD vmid to 0 */
+	tmp = regCP_MQD_CONTROL_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, CP_MQD_CONTROL, VMID, 0);
+	mqd->cp_mqd_control = tmp;
+
+	/* set the pointer to the HQD, this is similar CP_RB0_BASE/_HI */
+	hqd_gpu_addr = prop->hqd_base_gpu_addr >> 8;
+	mqd->cp_hqd_pq_base_lo = hqd_gpu_addr;
+	mqd->cp_hqd_pq_base_hi = upper_32_bits(hqd_gpu_addr);
+
+	/* set up the HQD, this is similar to CP_RB0_CNTL */
+	tmp = regCP_HQD_PQ_CONTROL_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, QUEUE_SIZE,
+			    (order_base_2(prop->queue_size / 4) - 1));
+	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, RPTR_BLOCK_SIZE,
+			    (order_base_2(AMDGPU_GPU_PAGE_SIZE / 4) - 1));
+	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, UNORD_DISPATCH, 1);
+	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, TUNNEL_DISPATCH,
+			    prop->allow_tunneling);
+	if (prop->kernel_queue) {
+		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, PRIV_STATE, 1);
+		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, KMD_QUEUE, 1);
+	}
+	if (prop->tmz_queue)
+		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, TMZ, 1);
+	mqd->cp_hqd_pq_control = tmp;
+
+	/* set the wb address whether it's enabled or not */
+	wb_gpu_addr = prop->rptr_gpu_addr;
+	mqd->cp_hqd_pq_rptr_report_addr_lo = wb_gpu_addr & 0xfffffffc;
+	mqd->cp_hqd_pq_rptr_report_addr_hi =
+		upper_32_bits(wb_gpu_addr) & 0xffff;
+
+	/* only used if CP_PQ_WPTR_POLL_CNTL.CP_PQ_WPTR_POLL_CNTL__EN_MASK=1 */
+	wb_gpu_addr = prop->wptr_gpu_addr;
+	mqd->cp_hqd_pq_wptr_poll_addr_lo = wb_gpu_addr & 0xfffffffc;
+	mqd->cp_hqd_pq_wptr_poll_addr_hi = upper_32_bits(wb_gpu_addr) & 0xffff;
+
+	tmp = 0;
+	/* enable the doorbell if requested */
+	if (prop->use_doorbell) {
+		tmp = regCP_HQD_PQ_DOORBELL_CONTROL_DEFAULT;
+		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
+				DOORBELL_OFFSET, prop->doorbell_index);
+
+		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
+				    DOORBELL_EN, 1);
+		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
+				    DOORBELL_SOURCE, 0);
+		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
+				    DOORBELL_HIT, 0);
+	}
+
+	mqd->cp_hqd_pq_doorbell_control = tmp;
+
+	/* reset read and write pointers, similar to CP_RB0_WPTR/_RPTR */
+	mqd->cp_hqd_pq_rptr = regCP_HQD_PQ_RPTR_DEFAULT;
+
+	/* set the vmid for the queue */
+	mqd->cp_hqd_vmid = 0;
+
+	tmp = regCP_HQD_PERSISTENT_STATE_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, CP_HQD_PERSISTENT_STATE, PRELOAD_SIZE, 0x55);
+	mqd->cp_hqd_persistent_state = tmp;
+
+	/* set MIN_IB_AVAIL_SIZE */
+	tmp = regCP_HQD_IB_CONTROL_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, CP_HQD_IB_CONTROL, MIN_IB_AVAIL_SIZE, 3);
+	mqd->cp_hqd_ib_control = tmp;
+
+	/* set static priority for a compute queue/ring */
+	mqd->cp_hqd_pipe_priority = prop->hqd_pipe_priority;
+	mqd->cp_hqd_queue_priority = prop->hqd_queue_priority;
+
+	mqd->cp_hqd_active = prop->hqd_active;
+
+	/* set UQ fenceaddress */
+	mqd->fence_address_lo = lower_32_bits(prop->fence_address);
+	mqd->fence_address_hi = upper_32_bits(prop->fence_address);
+
+	return 0;
+}
+
+static int gfx_v11_0_kiq_init_register(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct v11_compute_mqd *mqd = ring->mqd_ptr;
+	int j;
+
+	/* inactivate the queue */
+	if (amdgpu_sriov_vf(adev))
+		WREG32_SOC15(GC, 0, regCP_HQD_ACTIVE, 0);
+
+	/* disable wptr polling */
+	WREG32_FIELD15_PREREG(GC, 0, CP_PQ_WPTR_POLL_CNTL, EN, 0);
+
+	/* write the EOP addr */
+	WREG32_SOC15(GC, 0, regCP_HQD_EOP_BASE_ADDR,
+	       mqd->cp_hqd_eop_base_addr_lo);
+	WREG32_SOC15(GC, 0, regCP_HQD_EOP_BASE_ADDR_HI,
+	       mqd->cp_hqd_eop_base_addr_hi);
+
+	/* set the EOP size, register value is 2^(EOP_SIZE+1) dwords */
+	WREG32_SOC15(GC, 0, regCP_HQD_EOP_CONTROL,
+	       mqd->cp_hqd_eop_control);
+
+	/* enable doorbell? */
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_DOORBELL_CONTROL,
+	       mqd->cp_hqd_pq_doorbell_control);
+
+	/* disable the queue if it's active */
+	if (RREG32_SOC15(GC, 0, regCP_HQD_ACTIVE) & 1) {
+		WREG32_SOC15(GC, 0, regCP_HQD_DEQUEUE_REQUEST, 1);
+		for (j = 0; j < adev->usec_timeout; j++) {
+			if (!(RREG32_SOC15(GC, 0, regCP_HQD_ACTIVE) & 1))
+				break;
+			udelay(1);
+		}
+		WREG32_SOC15(GC, 0, regCP_HQD_DEQUEUE_REQUEST,
+		       mqd->cp_hqd_dequeue_request);
+		WREG32_SOC15(GC, 0, regCP_HQD_PQ_RPTR,
+		       mqd->cp_hqd_pq_rptr);
+		WREG32_SOC15(GC, 0, regCP_HQD_PQ_WPTR_LO,
+		       mqd->cp_hqd_pq_wptr_lo);
+		WREG32_SOC15(GC, 0, regCP_HQD_PQ_WPTR_HI,
+		       mqd->cp_hqd_pq_wptr_hi);
+	}
+
+	/* set the pointer to the MQD */
+	WREG32_SOC15(GC, 0, regCP_MQD_BASE_ADDR,
+	       mqd->cp_mqd_base_addr_lo);
+	WREG32_SOC15(GC, 0, regCP_MQD_BASE_ADDR_HI,
+	       mqd->cp_mqd_base_addr_hi);
+
+	/* set MQD vmid to 0 */
+	WREG32_SOC15(GC, 0, regCP_MQD_CONTROL,
+	       mqd->cp_mqd_control);
+
+	/* set the pointer to the HQD, this is similar CP_RB0_BASE/_HI */
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_BASE,
+	       mqd->cp_hqd_pq_base_lo);
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_BASE_HI,
+	       mqd->cp_hqd_pq_base_hi);
+
+	/* set up the HQD, this is similar to CP_RB0_CNTL */
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_CONTROL,
+	       mqd->cp_hqd_pq_control);
+
+	/* set the wb address whether it's enabled or not */
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_RPTR_REPORT_ADDR,
+		mqd->cp_hqd_pq_rptr_report_addr_lo);
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_RPTR_REPORT_ADDR_HI,
+		mqd->cp_hqd_pq_rptr_report_addr_hi);
+
+	/* only used if CP_PQ_WPTR_POLL_CNTL.CP_PQ_WPTR_POLL_CNTL__EN_MASK=1 */
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_WPTR_POLL_ADDR,
+	       mqd->cp_hqd_pq_wptr_poll_addr_lo);
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_WPTR_POLL_ADDR_HI,
+	       mqd->cp_hqd_pq_wptr_poll_addr_hi);
+
+	/* enable the doorbell if requested */
+	if (ring->use_doorbell) {
+		WREG32_SOC15(GC, 0, regCP_MEC_DOORBELL_RANGE_LOWER,
+			(adev->doorbell_index.kiq * 2) << 2);
+		WREG32_SOC15(GC, 0, regCP_MEC_DOORBELL_RANGE_UPPER,
+			(adev->doorbell_index.userqueue_end * 2) << 2);
+	}
+
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_DOORBELL_CONTROL,
+	       mqd->cp_hqd_pq_doorbell_control);
+
+	/* reset read and write pointers, similar to CP_RB0_WPTR/_RPTR */
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_WPTR_LO,
+	       mqd->cp_hqd_pq_wptr_lo);
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_WPTR_HI,
+	       mqd->cp_hqd_pq_wptr_hi);
+
+	/* set the vmid for the queue */
+	WREG32_SOC15(GC, 0, regCP_HQD_VMID, mqd->cp_hqd_vmid);
+
+	WREG32_SOC15(GC, 0, regCP_HQD_PERSISTENT_STATE,
+	       mqd->cp_hqd_persistent_state);
+
+	/* activate the queue */
+	WREG32_SOC15(GC, 0, regCP_HQD_ACTIVE,
+	       mqd->cp_hqd_active);
+
+	if (ring->use_doorbell)
+		WREG32_FIELD15_PREREG(GC, 0, CP_PQ_STATUS, DOORBELL_ENABLE, 1);
+
+	return 0;
+}
+
+static int gfx_v11_0_kiq_init_queue(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct v11_compute_mqd *mqd = ring->mqd_ptr;
+
+	gfx_v11_0_kiq_setting(ring);
+
+	if (amdgpu_in_reset(adev)) { /* for GPU_RESET case */
+		/* reset MQD to a clean status */
+		if (adev->gfx.kiq[0].mqd_backup)
+			memcpy_toio(mqd, adev->gfx.kiq[0].mqd_backup, sizeof(*mqd));
+
+		/* reset ring buffer */
+		ring->wptr = 0;
+		amdgpu_ring_clear_ring(ring);
+
+		mutex_lock(&adev->srbm_mutex);
+		soc21_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0);
+		gfx_v11_0_kiq_init_register(ring);
+		soc21_grbm_select(adev, 0, 0, 0, 0);
+		mutex_unlock(&adev->srbm_mutex);
+	} else {
+		memset((void *)mqd, 0, sizeof(*mqd));
+		if (amdgpu_sriov_vf(adev) && adev->in_suspend)
+			amdgpu_ring_clear_ring(ring);
+		mutex_lock(&adev->srbm_mutex);
+		soc21_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0);
+		amdgpu_ring_init_mqd(ring);
+		gfx_v11_0_kiq_init_register(ring);
+		soc21_grbm_select(adev, 0, 0, 0, 0);
+		mutex_unlock(&adev->srbm_mutex);
+
+		if (adev->gfx.kiq[0].mqd_backup)
+			memcpy_fromio(adev->gfx.kiq[0].mqd_backup, mqd, sizeof(*mqd));
+	}
+
+	return 0;
+}
+
+static int gfx_v11_0_kcq_init_queue(struct amdgpu_ring *ring, bool reset)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct v11_compute_mqd *mqd = ring->mqd_ptr;
+	int mqd_idx = ring - &adev->gfx.compute_ring[0];
+
+	if (!reset && !amdgpu_in_reset(adev) && !adev->in_suspend) {
+		memset((void *)mqd, 0, sizeof(*mqd));
+		mutex_lock(&adev->srbm_mutex);
+		soc21_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0);
+		amdgpu_ring_init_mqd(ring);
+		soc21_grbm_select(adev, 0, 0, 0, 0);
+		mutex_unlock(&adev->srbm_mutex);
+
+		if (adev->gfx.mec.mqd_backup[mqd_idx])
+			memcpy_fromio(adev->gfx.mec.mqd_backup[mqd_idx], mqd, sizeof(*mqd));
+	} else {
+		/* restore MQD to a clean status */
+		if (adev->gfx.mec.mqd_backup[mqd_idx])
+			memcpy_toio(mqd, adev->gfx.mec.mqd_backup[mqd_idx], sizeof(*mqd));
+		/* reset ring buffer */
+		ring->wptr = 0;
+		atomic64_set((atomic64_t *)ring->wptr_cpu_addr, 0);
+		amdgpu_ring_clear_ring(ring);
+	}
+
+	return 0;
+}
+
+static int gfx_v11_0_kiq_resume(struct amdgpu_device *adev)
+{
+	gfx_v11_0_kiq_init_queue(&adev->gfx.kiq[0].ring);
+	return 0;
+}
+
+static int gfx_v11_0_kcq_resume(struct amdgpu_device *adev)
+{
+	int i, r;
+
+	if (!amdgpu_async_gfx_ring)
+		gfx_v11_0_cp_compute_enable(adev, true);
+
+	for (i = 0; i < adev->gfx.num_compute_rings; i++) {
+		r = gfx_v11_0_kcq_init_queue(&adev->gfx.compute_ring[i], false);
+		if (r)
+			return r;
+	}
+
+	return amdgpu_gfx_enable_kcq(adev, 0);
+}
+
+static int gfx_v11_0_cp_resume(struct amdgpu_device *adev)
+{
+	int r, i;
+	struct amdgpu_ring *ring;
+
+	if (!(adev->flags & AMD_IS_APU))
+		gfx_v11_0_enable_gui_idle_interrupt(adev, false);
+
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT) {
+		/* legacy firmware loading */
+		r = gfx_v11_0_cp_gfx_load_microcode(adev);
+		if (r)
+			return r;
+
+		if (adev->gfx.rs64_enable)
+			r = gfx_v11_0_cp_compute_load_microcode_rs64(adev);
+		else
+			r = gfx_v11_0_cp_compute_load_microcode(adev);
+		if (r)
+			return r;
+	}
+
+	gfx_v11_0_cp_set_doorbell_range(adev);
+
+	if (amdgpu_async_gfx_ring) {
+		gfx_v11_0_cp_compute_enable(adev, true);
+		gfx_v11_0_cp_gfx_enable(adev, true);
+	}
+
+	if (adev->enable_mes_kiq && adev->mes.kiq_hw_init)
+		r = amdgpu_mes_kiq_hw_init(adev);
+	else
+		r = gfx_v11_0_kiq_resume(adev);
+	if (r)
+		return r;
+
+	r = gfx_v11_0_kcq_resume(adev);
+	if (r)
+		return r;
+
+	if (!amdgpu_async_gfx_ring) {
+		r = gfx_v11_0_cp_gfx_resume(adev);
+		if (r)
+			return r;
+	} else {
+		r = gfx_v11_0_cp_async_gfx_ring_resume(adev);
+		if (r)
+			return r;
+	}
+
+	if (adev->gfx.disable_kq) {
+		for (i = 0; i < adev->gfx.num_gfx_rings; i++) {
+			ring = &adev->gfx.gfx_ring[i];
+			/* we don't want to set ring->ready */
+			r = amdgpu_ring_test_ring(ring);
+			if (r)
+				return r;
+		}
+		if (amdgpu_async_gfx_ring)
+			amdgpu_gfx_disable_kgq(adev, 0);
+	} else {
+		for (i = 0; i < adev->gfx.num_gfx_rings; i++) {
+			ring = &adev->gfx.gfx_ring[i];
+			r = amdgpu_ring_test_helper(ring);
+			if (r)
+				return r;
+		}
+	}
+
+	for (i = 0; i < adev->gfx.num_compute_rings; i++) {
+		ring = &adev->gfx.compute_ring[i];
+		r = amdgpu_ring_test_helper(ring);
+		if (r)
+			return r;
+	}
+
+	return 0;
+}
+
+static void gfx_v11_0_cp_enable(struct amdgpu_device *adev, bool enable)
+{
+	gfx_v11_0_cp_gfx_enable(adev, enable);
+	gfx_v11_0_cp_compute_enable(adev, enable);
+}
+
+static int gfx_v11_0_gfxhub_enable(struct amdgpu_device *adev)
+{
+	int r;
+	bool value;
+
+	r = adev->gfxhub.funcs->gart_enable(adev);
+	if (r)
+		return r;
+
+	amdgpu_device_flush_hdp(adev, NULL);
+
+	value = amdgpu_vm_fault_stop != AMDGPU_VM_FAULT_STOP_ALWAYS;
+
+	adev->gfxhub.funcs->set_fault_enable_default(adev, value);
+	/* TODO investigate why this and the hdp flush above is needed,
+	 * are we missing a flush somewhere else? */
+	adev->gmc.gmc_funcs->flush_gpu_tlb(adev, 0, AMDGPU_GFXHUB(0), 0);
+
+	return 0;
+}
+
+static void gfx_v11_0_select_cp_fw_arch(struct amdgpu_device *adev)
+{
+	u32 tmp;
+
+	/* select RS64 */
+	if (adev->gfx.rs64_enable) {
+		tmp = RREG32_SOC15(GC, 0, regCP_GFX_CNTL);
+		tmp = REG_SET_FIELD(tmp, CP_GFX_CNTL, ENGINE_SEL, 1);
+		WREG32_SOC15(GC, 0, regCP_GFX_CNTL, tmp);
+
+		tmp = RREG32_SOC15(GC, 0, regCP_MEC_ISA_CNTL);
+		tmp = REG_SET_FIELD(tmp, CP_MEC_ISA_CNTL, ISA_MODE, 1);
+		WREG32_SOC15(GC, 0, regCP_MEC_ISA_CNTL, tmp);
+	}
+
+	if (amdgpu_emu_mode == 1)
+		msleep(100);
+}
+
+static int get_gb_addr_config(struct amdgpu_device * adev)
+{
+	u32 gb_addr_config;
+
+	gb_addr_config = RREG32_SOC15(GC, 0, regGB_ADDR_CONFIG);
+	if (gb_addr_config == 0)
+		return -EINVAL;
+
+	adev->gfx.config.gb_addr_config_fields.num_pkrs =
+		1 << REG_GET_FIELD(gb_addr_config, GB_ADDR_CONFIG, NUM_PKRS);
+
+	adev->gfx.config.gb_addr_config = gb_addr_config;
+
+	adev->gfx.config.gb_addr_config_fields.num_pipes = 1 <<
+			REG_GET_FIELD(adev->gfx.config.gb_addr_config,
+				      GB_ADDR_CONFIG, NUM_PIPES);
+
+	adev->gfx.config.max_tile_pipes =
+		adev->gfx.config.gb_addr_config_fields.num_pipes;
+
+	adev->gfx.config.gb_addr_config_fields.max_compress_frags = 1 <<
+			REG_GET_FIELD(adev->gfx.config.gb_addr_config,
+				      GB_ADDR_CONFIG, MAX_COMPRESSED_FRAGS);
+	adev->gfx.config.gb_addr_config_fields.num_rb_per_se = 1 <<
+			REG_GET_FIELD(adev->gfx.config.gb_addr_config,
+				      GB_ADDR_CONFIG, NUM_RB_PER_SE);
+	adev->gfx.config.gb_addr_config_fields.num_se = 1 <<
+			REG_GET_FIELD(adev->gfx.config.gb_addr_config,
+				      GB_ADDR_CONFIG, NUM_SHADER_ENGINES);
+	adev->gfx.config.gb_addr_config_fields.pipe_interleave_size = 1 << (8 +
+			REG_GET_FIELD(adev->gfx.config.gb_addr_config,
+				      GB_ADDR_CONFIG, PIPE_INTERLEAVE_SIZE));
+
+	return 0;
+}
+
+static void gfx_v11_0_disable_gpa_mode(struct amdgpu_device *adev)
+{
+	uint32_t data;
+
+	data = RREG32_SOC15(GC, 0, regCPC_PSP_DEBUG);
+	data |= CPC_PSP_DEBUG__GPA_OVERRIDE_MASK;
+	WREG32_SOC15(GC, 0, regCPC_PSP_DEBUG, data);
+
+	data = RREG32_SOC15(GC, 0, regCPG_PSP_DEBUG);
+	data |= CPG_PSP_DEBUG__GPA_OVERRIDE_MASK;
+	WREG32_SOC15(GC, 0, regCPG_PSP_DEBUG, data);
+}
+
+static int gfx_v11_0_hw_init(struct amdgpu_ip_block *ip_block)
+{
+	int r;
+	struct amdgpu_device *adev = ip_block->adev;
+
+	amdgpu_gfx_cleaner_shader_init(adev, adev->gfx.cleaner_shader_size,
+				       adev->gfx.cleaner_shader_ptr);
+
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_RLC_BACKDOOR_AUTO) {
+		if (adev->gfx.imu.funcs) {
+			/* RLC autoload sequence 1: Program rlc ram */
+			if (adev->gfx.imu.funcs->program_rlc_ram)
+				adev->gfx.imu.funcs->program_rlc_ram(adev);
+			/* rlc autoload firmware */
+			r = gfx_v11_0_rlc_backdoor_autoload_enable(adev);
+			if (r)
+				return r;
+		}
+	} else {
+		if (adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT) {
+			if (adev->gfx.imu.funcs && (amdgpu_dpm > 0)) {
+				if (adev->gfx.imu.funcs->load_microcode)
+					adev->gfx.imu.funcs->load_microcode(adev);
+				if (adev->gfx.imu.funcs->setup_imu)
+					adev->gfx.imu.funcs->setup_imu(adev);
+				if (adev->gfx.imu.funcs->start_imu)
+					adev->gfx.imu.funcs->start_imu(adev);
+			}
+
+			/* disable gpa mode in backdoor loading */
+			gfx_v11_0_disable_gpa_mode(adev);
+		}
+	}
+
+	if ((adev->firmware.load_type == AMDGPU_FW_LOAD_RLC_BACKDOOR_AUTO) ||
+	    (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP)) {
+		r = gfx_v11_0_wait_for_rlc_autoload_complete(adev);
+		if (r) {
+			dev_err(adev->dev, "(%d) failed to wait rlc autoload complete\n", r);
+			return r;
+		}
+	}
+
+	adev->gfx.is_poweron = true;
+
+	if(get_gb_addr_config(adev))
+		DRM_WARN("Invalid gb_addr_config !\n");
+
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP &&
+	    adev->gfx.rs64_enable)
+		gfx_v11_0_config_gfx_rs64(adev);
+
+	r = gfx_v11_0_gfxhub_enable(adev);
+	if (r)
+		return r;
+
+	if (!amdgpu_emu_mode)
+		gfx_v11_0_init_golden_registers(adev);
+
+	if ((adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT) ||
+	    (adev->firmware.load_type == AMDGPU_FW_LOAD_RLC_BACKDOOR_AUTO && amdgpu_dpm == 1)) {
+		/**
+		 * For gfx 11, rlc firmware loading relies on smu firmware is
+		 * loaded firstly, so in direct type, it has to load smc ucode
+		 * here before rlc.
+		 */
+		r = amdgpu_pm_load_smu_firmware(adev, NULL);
+		if (r)
+			return r;
+	}
+
+	gfx_v11_0_constants_init(adev);
+
+	if (adev->firmware.load_type != AMDGPU_FW_LOAD_PSP)
+		gfx_v11_0_select_cp_fw_arch(adev);
+
+	if (adev->nbio.funcs->gc_doorbell_init)
+		adev->nbio.funcs->gc_doorbell_init(adev);
+
+	r = gfx_v11_0_rlc_resume(adev);
+	if (r)
+		return r;
+
+	/*
+	 * init golden registers and rlc resume may override some registers,
+	 * reconfig them here
+	 */
+	gfx_v11_0_tcp_harvest(adev);
+
+	r = gfx_v11_0_cp_resume(adev);
+	if (r)
+		return r;
+
+	/* get IMU version from HW if it's not set */
+	if (!adev->gfx.imu_fw_version)
+		adev->gfx.imu_fw_version = RREG32_SOC15(GC, 0, regGFX_IMU_SCRATCH_0);
+
+	return r;
+}
+
+static int gfx_v11_0_set_userq_eop_interrupts(struct amdgpu_device *adev,
+					      bool enable)
+{
+	unsigned int irq_type;
+	int m, p, r;
+
+	if (adev->userq_funcs[AMDGPU_HW_IP_GFX]) {
+		for (m = 0; m < adev->gfx.me.num_me; m++) {
+			for (p = 0; p < adev->gfx.me.num_pipe_per_me; p++) {
+				irq_type = AMDGPU_CP_IRQ_GFX_ME0_PIPE0_EOP + p;
+				if (enable)
+					r = amdgpu_irq_get(adev, &adev->gfx.eop_irq,
+							   irq_type);
+				else
+					r = amdgpu_irq_put(adev, &adev->gfx.eop_irq,
+							   irq_type);
+				if (r)
+					return r;
+			}
+		}
+	}
+
+	if (adev->userq_funcs[AMDGPU_HW_IP_COMPUTE]) {
+		for (m = 0; m < adev->gfx.mec.num_mec; ++m) {
+			for (p = 0; p < adev->gfx.mec.num_pipe_per_mec; p++) {
+				irq_type = AMDGPU_CP_IRQ_COMPUTE_MEC1_PIPE0_EOP
+					+ (m * adev->gfx.mec.num_pipe_per_mec)
+					+ p;
+				if (enable)
+					r = amdgpu_irq_get(adev, &adev->gfx.eop_irq,
+							   irq_type);
+				else
+					r = amdgpu_irq_put(adev, &adev->gfx.eop_irq,
+							   irq_type);
+				if (r)
+					return r;
+			}
+		}
+	}
+
+	return 0;
+}
+
+static int gfx_v11_0_hw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	cancel_delayed_work_sync(&adev->gfx.idle_work);
+
+	amdgpu_irq_put(adev, &adev->gfx.priv_reg_irq, 0);
+	amdgpu_irq_put(adev, &adev->gfx.priv_inst_irq, 0);
+	amdgpu_irq_put(adev, &adev->gfx.bad_op_irq, 0);
+	gfx_v11_0_set_userq_eop_interrupts(adev, false);
+
+	if (!adev->no_hw_access) {
+		if (amdgpu_async_gfx_ring &&
+		    !adev->gfx.disable_kq) {
+			if (amdgpu_gfx_disable_kgq(adev, 0))
+				DRM_ERROR("KGQ disable failed\n");
+		}
+
+		if (amdgpu_gfx_disable_kcq(adev, 0))
+			DRM_ERROR("KCQ disable failed\n");
+
+		amdgpu_mes_kiq_hw_fini(adev);
+	}
+
+	if (amdgpu_sriov_vf(adev))
+		/* Remove the steps disabling CPG and clearing KIQ position,
+		 * so that CP could perform IDLE-SAVE during switch. Those
+		 * steps are necessary to avoid a DMAR error in gfx9 but it is
+		 * not reproduced on gfx11.
+		 */
+		return 0;
+
+	gfx_v11_0_cp_enable(adev, false);
+	gfx_v11_0_enable_gui_idle_interrupt(adev, false);
+
+	adev->gfxhub.funcs->gart_disable(adev);
+
+	adev->gfx.is_poweron = false;
+
+	return 0;
+}
+
+static int gfx_v11_0_suspend(struct amdgpu_ip_block *ip_block)
+{
+	return gfx_v11_0_hw_fini(ip_block);
+}
+
+static int gfx_v11_0_resume(struct amdgpu_ip_block *ip_block)
+{
+	return gfx_v11_0_hw_init(ip_block);
+}
+
+static bool gfx_v11_0_is_idle(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	if (REG_GET_FIELD(RREG32_SOC15(GC, 0, regGRBM_STATUS),
+				GRBM_STATUS, GUI_ACTIVE))
+		return false;
+	else
+		return true;
+}
+
+static int gfx_v11_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
+{
+	unsigned i;
+	u32 tmp;
+	struct amdgpu_device *adev = ip_block->adev;
+
+	for (i = 0; i < adev->usec_timeout; i++) {
+		/* read MC_STATUS */
+		tmp = RREG32_SOC15(GC, 0, regGRBM_STATUS) &
+			GRBM_STATUS__GUI_ACTIVE_MASK;
+
+		if (!REG_GET_FIELD(tmp, GRBM_STATUS, GUI_ACTIVE))
+			return 0;
+		udelay(1);
+	}
+	return -ETIMEDOUT;
+}
+
+int gfx_v11_0_request_gfx_index_mutex(struct amdgpu_device *adev,
+				      bool req)
+{
+	u32 i, tmp, val;
+
+	for (i = 0; i < adev->usec_timeout; i++) {
+		/* Request with MeId=2, PipeId=0 */
+		tmp = REG_SET_FIELD(0, CP_GFX_INDEX_MUTEX, REQUEST, req);
+		tmp = REG_SET_FIELD(tmp, CP_GFX_INDEX_MUTEX, CLIENTID, 4);
+		WREG32_SOC15(GC, 0, regCP_GFX_INDEX_MUTEX, tmp);
+
+		val = RREG32_SOC15(GC, 0, regCP_GFX_INDEX_MUTEX);
+		if (req) {
+			if (val == tmp)
+				break;
+		} else {
+			tmp = REG_SET_FIELD(tmp, CP_GFX_INDEX_MUTEX,
+					    REQUEST, 1);
+
+			/* unlocked or locked by firmware */
+			if (val != tmp)
+				break;
+		}
+		udelay(1);
+	}
+
+	if (i >= adev->usec_timeout)
+		return -EINVAL;
+
+	return 0;
+}
+
+static int gfx_v11_0_soft_reset(struct amdgpu_ip_block *ip_block)
+{
+	u32 grbm_soft_reset = 0;
+	u32 tmp;
+	int r, i, j, k;
+	struct amdgpu_device *adev = ip_block->adev;
+
+	amdgpu_gfx_rlc_enter_safe_mode(adev, 0);
+
+	tmp = RREG32_SOC15(GC, 0, regCP_INT_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_INT_CNTL, CMP_BUSY_INT_ENABLE, 0);
+	tmp = REG_SET_FIELD(tmp, CP_INT_CNTL, CNTX_BUSY_INT_ENABLE, 0);
+	tmp = REG_SET_FIELD(tmp, CP_INT_CNTL, CNTX_EMPTY_INT_ENABLE, 0);
+	tmp = REG_SET_FIELD(tmp, CP_INT_CNTL, GFX_IDLE_INT_ENABLE, 0);
+	WREG32_SOC15(GC, 0, regCP_INT_CNTL, tmp);
+
+	mutex_lock(&adev->srbm_mutex);
+	for (i = 0; i < adev->gfx.mec.num_mec; ++i) {
+		for (j = 0; j < adev->gfx.mec.num_queue_per_pipe; j++) {
+			for (k = 0; k < adev->gfx.mec.num_pipe_per_mec; k++) {
+				soc21_grbm_select(adev, i, k, j, 0);
+
+				WREG32_SOC15(GC, 0, regCP_HQD_DEQUEUE_REQUEST, 0x2);
+				WREG32_SOC15(GC, 0, regSPI_COMPUTE_QUEUE_RESET, 0x1);
+			}
+		}
+	}
+	for (i = 0; i < adev->gfx.me.num_me; ++i) {
+		for (j = 0; j < adev->gfx.me.num_queue_per_pipe; j++) {
+			for (k = 0; k < adev->gfx.me.num_pipe_per_me; k++) {
+				soc21_grbm_select(adev, i, k, j, 0);
+
+				WREG32_SOC15(GC, 0, regCP_GFX_HQD_DEQUEUE_REQUEST, 0x1);
+			}
+		}
+	}
+	soc21_grbm_select(adev, 0, 0, 0, 0);
+	mutex_unlock(&adev->srbm_mutex);
+
+	/* Try to acquire the gfx mutex before access to CP_VMID_RESET */
+	mutex_lock(&adev->gfx.reset_sem_mutex);
+	r = gfx_v11_0_request_gfx_index_mutex(adev, true);
+	if (r) {
+		mutex_unlock(&adev->gfx.reset_sem_mutex);
+		DRM_ERROR("Failed to acquire the gfx mutex during soft reset\n");
+		return r;
+	}
+
+	WREG32_SOC15(GC, 0, regCP_VMID_RESET, 0xfffffffe);
+
+	// Read CP_VMID_RESET register three times.
+	// to get sufficient time for GFX_HQD_ACTIVE reach 0
+	RREG32_SOC15(GC, 0, regCP_VMID_RESET);
+	RREG32_SOC15(GC, 0, regCP_VMID_RESET);
+	RREG32_SOC15(GC, 0, regCP_VMID_RESET);
+
+	/* release the gfx mutex */
+	r = gfx_v11_0_request_gfx_index_mutex(adev, false);
+	mutex_unlock(&adev->gfx.reset_sem_mutex);
+	if (r) {
+		DRM_ERROR("Failed to release the gfx mutex during soft reset\n");
+		return r;
+	}
+
+	for (i = 0; i < adev->usec_timeout; i++) {
+		if (!RREG32_SOC15(GC, 0, regCP_HQD_ACTIVE) &&
+		    !RREG32_SOC15(GC, 0, regCP_GFX_HQD_ACTIVE))
+			break;
+		udelay(1);
+	}
+	if (i >= adev->usec_timeout) {
+		printk("Failed to wait all pipes clean\n");
+		return -EINVAL;
+	}
+
+	/**********  trigger soft reset  ***********/
+	grbm_soft_reset = RREG32_SOC15(GC, 0, regGRBM_SOFT_RESET);
+	grbm_soft_reset = REG_SET_FIELD(grbm_soft_reset, GRBM_SOFT_RESET,
+					SOFT_RESET_CP, 1);
+	grbm_soft_reset = REG_SET_FIELD(grbm_soft_reset, GRBM_SOFT_RESET,
+					SOFT_RESET_GFX, 1);
+	grbm_soft_reset = REG_SET_FIELD(grbm_soft_reset, GRBM_SOFT_RESET,
+					SOFT_RESET_CPF, 1);
+	grbm_soft_reset = REG_SET_FIELD(grbm_soft_reset, GRBM_SOFT_RESET,
+					SOFT_RESET_CPC, 1);
+	grbm_soft_reset = REG_SET_FIELD(grbm_soft_reset, GRBM_SOFT_RESET,
+					SOFT_RESET_CPG, 1);
+	WREG32_SOC15(GC, 0, regGRBM_SOFT_RESET, grbm_soft_reset);
+	/**********  exit soft reset  ***********/
+	grbm_soft_reset = RREG32_SOC15(GC, 0, regGRBM_SOFT_RESET);
+	grbm_soft_reset = REG_SET_FIELD(grbm_soft_reset, GRBM_SOFT_RESET,
+					SOFT_RESET_CP, 0);
+	grbm_soft_reset = REG_SET_FIELD(grbm_soft_reset, GRBM_SOFT_RESET,
+					SOFT_RESET_GFX, 0);
+	grbm_soft_reset = REG_SET_FIELD(grbm_soft_reset, GRBM_SOFT_RESET,
+					SOFT_RESET_CPF, 0);
+	grbm_soft_reset = REG_SET_FIELD(grbm_soft_reset, GRBM_SOFT_RESET,
+					SOFT_RESET_CPC, 0);
+	grbm_soft_reset = REG_SET_FIELD(grbm_soft_reset, GRBM_SOFT_RESET,
+					SOFT_RESET_CPG, 0);
+	WREG32_SOC15(GC, 0, regGRBM_SOFT_RESET, grbm_soft_reset);
+
+	tmp = RREG32_SOC15(GC, 0, regCP_SOFT_RESET_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_SOFT_RESET_CNTL, CMP_HQD_REG_RESET, 0x1);
+	WREG32_SOC15(GC, 0, regCP_SOFT_RESET_CNTL, tmp);
+
+	WREG32_SOC15(GC, 0, regCP_ME_CNTL, 0x0);
+	WREG32_SOC15(GC, 0, regCP_MEC_RS64_CNTL, 0x0);
+
+	for (i = 0; i < adev->usec_timeout; i++) {
+		if (!RREG32_SOC15(GC, 0, regCP_VMID_RESET))
+			break;
+		udelay(1);
+	}
+	if (i >= adev->usec_timeout) {
+		printk("Failed to wait CP_VMID_RESET to 0\n");
+		return -EINVAL;
+	}
+
+	tmp = RREG32_SOC15(GC, 0, regCP_INT_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_INT_CNTL, CMP_BUSY_INT_ENABLE, 1);
+	tmp = REG_SET_FIELD(tmp, CP_INT_CNTL, CNTX_BUSY_INT_ENABLE, 1);
+	tmp = REG_SET_FIELD(tmp, CP_INT_CNTL, CNTX_EMPTY_INT_ENABLE, 1);
+	tmp = REG_SET_FIELD(tmp, CP_INT_CNTL, GFX_IDLE_INT_ENABLE, 1);
+	WREG32_SOC15(GC, 0, regCP_INT_CNTL, tmp);
+
+	amdgpu_gfx_rlc_exit_safe_mode(adev, 0);
+
+	return gfx_v11_0_cp_resume(adev);
+}
+
+static bool gfx_v11_0_check_soft_reset(struct amdgpu_ip_block *ip_block)
+{
+	int i, r;
+	struct amdgpu_device *adev = ip_block->adev;
+	struct amdgpu_ring *ring;
+	long tmo = msecs_to_jiffies(1000);
+
+	for (i = 0; i < adev->gfx.num_gfx_rings; i++) {
+		ring = &adev->gfx.gfx_ring[i];
+		r = amdgpu_ring_test_ib(ring, tmo);
+		if (r)
+			return true;
+	}
+
+	for (i = 0; i < adev->gfx.num_compute_rings; i++) {
+		ring = &adev->gfx.compute_ring[i];
+		r = amdgpu_ring_test_ib(ring, tmo);
+		if (r)
+			return true;
+	}
+
+	return false;
+}
+
+static int gfx_v11_0_post_soft_reset(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	/**
+	 * GFX soft reset will impact MES, need resume MES when do GFX soft reset
+	 */
+	return amdgpu_mes_resume(adev);
+}
+
+static uint64_t gfx_v11_0_get_gpu_clock_counter(struct amdgpu_device *adev)
+{
+	uint64_t clock;
+	uint64_t clock_counter_lo, clock_counter_hi_pre, clock_counter_hi_after;
+
+	if (amdgpu_sriov_vf(adev)) {
+		amdgpu_gfx_off_ctrl(adev, false);
+		mutex_lock(&adev->gfx.gpu_clock_mutex);
+		clock_counter_hi_pre = (uint64_t)RREG32_SOC15(GC, 0, regCP_MES_MTIME_HI);
+		clock_counter_lo = (uint64_t)RREG32_SOC15(GC, 0, regCP_MES_MTIME_LO);
+		clock_counter_hi_after = (uint64_t)RREG32_SOC15(GC, 0, regCP_MES_MTIME_HI);
+		if (clock_counter_hi_pre != clock_counter_hi_after)
+			clock_counter_lo = (uint64_t)RREG32_SOC15(GC, 0, regCP_MES_MTIME_LO);
+		mutex_unlock(&adev->gfx.gpu_clock_mutex);
+		amdgpu_gfx_off_ctrl(adev, true);
+	} else {
+		preempt_disable();
+		clock_counter_hi_pre = (uint64_t)RREG32_SOC15(SMUIO, 0, regGOLDEN_TSC_COUNT_UPPER);
+		clock_counter_lo = (uint64_t)RREG32_SOC15(SMUIO, 0, regGOLDEN_TSC_COUNT_LOWER);
+		clock_counter_hi_after = (uint64_t)RREG32_SOC15(SMUIO, 0, regGOLDEN_TSC_COUNT_UPPER);
+		if (clock_counter_hi_pre != clock_counter_hi_after)
+			clock_counter_lo = (uint64_t)RREG32_SOC15(SMUIO, 0, regGOLDEN_TSC_COUNT_LOWER);
+		preempt_enable();
+	}
+	clock = clock_counter_lo | (clock_counter_hi_after << 32ULL);
+
+	return clock;
+}
+
+static void gfx_v11_0_ring_emit_gds_switch(struct amdgpu_ring *ring,
+					   uint32_t vmid,
+					   uint32_t gds_base, uint32_t gds_size,
+					   uint32_t gws_base, uint32_t gws_size,
+					   uint32_t oa_base, uint32_t oa_size)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	/* GDS Base */
+	gfx_v11_0_write_data_to_reg(ring, 0, false,
+				    SOC15_REG_OFFSET(GC, 0, regGDS_VMID0_BASE) + 2 * vmid,
+				    gds_base);
+
+	/* GDS Size */
+	gfx_v11_0_write_data_to_reg(ring, 0, false,
+				    SOC15_REG_OFFSET(GC, 0, regGDS_VMID0_SIZE) + 2 * vmid,
+				    gds_size);
+
+	/* GWS */
+	gfx_v11_0_write_data_to_reg(ring, 0, false,
+				    SOC15_REG_OFFSET(GC, 0, regGDS_GWS_VMID0) + vmid,
+				    gws_size << GDS_GWS_VMID0__SIZE__SHIFT | gws_base);
+
+	/* OA */
+	gfx_v11_0_write_data_to_reg(ring, 0, false,
+				    SOC15_REG_OFFSET(GC, 0, regGDS_OA_VMID0) + vmid,
+				    (1 << (oa_size + oa_base)) - (1 << oa_base));
+}
+
+static int gfx_v11_0_early_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	switch (amdgpu_user_queue) {
+	case -1:
+	case 0:
+	default:
+		adev->gfx.disable_kq = false;
+		adev->gfx.disable_uq = true;
+		break;
+	case 1:
+		adev->gfx.disable_kq = false;
+		adev->gfx.disable_uq = false;
+		break;
+	case 2:
+		adev->gfx.disable_kq = true;
+		adev->gfx.disable_uq = false;
+		break;
+	}
+
+	adev->gfx.funcs = &gfx_v11_0_gfx_funcs;
+
+	if (adev->gfx.disable_kq) {
+		/* We need one GFX ring temporarily to set up
+		 * the clear state.
+		 */
+		adev->gfx.num_gfx_rings = 1;
+		adev->gfx.num_compute_rings = 0;
+	} else {
+		adev->gfx.num_gfx_rings = GFX11_NUM_GFX_RINGS;
+		adev->gfx.num_compute_rings = min(amdgpu_gfx_get_num_kcq(adev),
+						  AMDGPU_MAX_COMPUTE_RINGS);
+	}
+
+	gfx_v11_0_set_kiq_pm4_funcs(adev);
+	gfx_v11_0_set_ring_funcs(adev);
+	gfx_v11_0_set_irq_funcs(adev);
+	gfx_v11_0_set_gds_init(adev);
+	gfx_v11_0_set_rlc_funcs(adev);
+	gfx_v11_0_set_mqd_funcs(adev);
+	gfx_v11_0_set_imu_funcs(adev);
+
+	gfx_v11_0_init_rlcg_reg_access_ctrl(adev);
+
+	return gfx_v11_0_init_microcode(adev);
+}
+
+static int gfx_v11_0_late_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int r;
+
+	r = amdgpu_irq_get(adev, &adev->gfx.priv_reg_irq, 0);
+	if (r)
+		return r;
+
+	r = amdgpu_irq_get(adev, &adev->gfx.priv_inst_irq, 0);
+	if (r)
+		return r;
+
+	r = amdgpu_irq_get(adev, &adev->gfx.bad_op_irq, 0);
+	if (r)
+		return r;
+
+	r = gfx_v11_0_set_userq_eop_interrupts(adev, true);
+	if (r)
+		return r;
+
+	return 0;
+}
+
+static bool gfx_v11_0_is_rlc_enabled(struct amdgpu_device *adev)
+{
+	uint32_t rlc_cntl;
+
+	/* if RLC is not enabled, do nothing */
+	rlc_cntl = RREG32_SOC15(GC, 0, regRLC_CNTL);
+	return (REG_GET_FIELD(rlc_cntl, RLC_CNTL, RLC_ENABLE_F32)) ? true : false;
+}
+
+static void gfx_v11_0_set_safe_mode(struct amdgpu_device *adev, int xcc_id)
+{
+	uint32_t data;
+	unsigned i;
+
+	data = RLC_SAFE_MODE__CMD_MASK;
+	data |= (1 << RLC_SAFE_MODE__MESSAGE__SHIFT);
+
+	WREG32_SOC15(GC, 0, regRLC_SAFE_MODE, data);
+
+	/* wait for RLC_SAFE_MODE */
+	for (i = 0; i < adev->usec_timeout; i++) {
+		if (!REG_GET_FIELD(RREG32_SOC15(GC, 0, regRLC_SAFE_MODE),
+				   RLC_SAFE_MODE, CMD))
+			break;
+		udelay(1);
+	}
+}
+
+static void gfx_v11_0_unset_safe_mode(struct amdgpu_device *adev, int xcc_id)
+{
+	WREG32_SOC15(GC, 0, regRLC_SAFE_MODE, RLC_SAFE_MODE__CMD_MASK);
+}
+
+static void gfx_v11_0_update_perf_clk(struct amdgpu_device *adev,
+				      bool enable)
+{
+	uint32_t def, data;
+
+	if (!(adev->cg_flags & AMD_CG_SUPPORT_GFX_PERF_CLK))
+		return;
+
+	def = data = RREG32_SOC15(GC, 0, regRLC_CGTT_MGCG_OVERRIDE);
+
+	if (enable)
+		data &= ~RLC_CGTT_MGCG_OVERRIDE__PERFMON_CLOCK_STATE_MASK;
+	else
+		data |= RLC_CGTT_MGCG_OVERRIDE__PERFMON_CLOCK_STATE_MASK;
+
+	if (def != data)
+		WREG32_SOC15(GC, 0, regRLC_CGTT_MGCG_OVERRIDE, data);
+}
+
+static void gfx_v11_0_update_sram_fgcg(struct amdgpu_device *adev,
+				       bool enable)
+{
+	uint32_t def, data;
+
+	if (!(adev->cg_flags & AMD_CG_SUPPORT_GFX_FGCG))
+		return;
+
+	def = data = RREG32_SOC15(GC, 0, regRLC_CGTT_MGCG_OVERRIDE);
+
+	if (enable)
+		data &= ~RLC_CGTT_MGCG_OVERRIDE__GFXIP_FGCG_OVERRIDE_MASK;
+	else
+		data |= RLC_CGTT_MGCG_OVERRIDE__GFXIP_FGCG_OVERRIDE_MASK;
+
+	if (def != data)
+		WREG32_SOC15(GC, 0, regRLC_CGTT_MGCG_OVERRIDE, data);
+}
+
+static void gfx_v11_0_update_repeater_fgcg(struct amdgpu_device *adev,
+					   bool enable)
+{
+	uint32_t def, data;
+
+	if (!(adev->cg_flags & AMD_CG_SUPPORT_REPEATER_FGCG))
+		return;
+
+	def = data = RREG32_SOC15(GC, 0, regRLC_CGTT_MGCG_OVERRIDE);
+
+	if (enable)
+		data &= ~RLC_CGTT_MGCG_OVERRIDE__GFXIP_REPEATER_FGCG_OVERRIDE_MASK;
+	else
+		data |= RLC_CGTT_MGCG_OVERRIDE__GFXIP_REPEATER_FGCG_OVERRIDE_MASK;
+
+	if (def != data)
+		WREG32_SOC15(GC, 0, regRLC_CGTT_MGCG_OVERRIDE, data);
+}
+
+static void gfx_v11_0_update_medium_grain_clock_gating(struct amdgpu_device *adev,
+						       bool enable)
+{
+	uint32_t data, def;
+
+	if (!(adev->cg_flags & (AMD_CG_SUPPORT_GFX_MGCG | AMD_CG_SUPPORT_GFX_MGLS)))
+		return;
+
+	/* It is disabled by HW by default */
+	if (enable) {
+		if (adev->cg_flags & AMD_CG_SUPPORT_GFX_MGCG) {
+			/* 1 - RLC_CGTT_MGCG_OVERRIDE */
+			def = data = RREG32_SOC15(GC, 0, regRLC_CGTT_MGCG_OVERRIDE);
+
+			data &= ~(RLC_CGTT_MGCG_OVERRIDE__GRBM_CGTT_SCLK_OVERRIDE_MASK |
+				  RLC_CGTT_MGCG_OVERRIDE__RLC_CGTT_SCLK_OVERRIDE_MASK |
+				  RLC_CGTT_MGCG_OVERRIDE__GFXIP_MGCG_OVERRIDE_MASK);
+
+			if (def != data)
+				WREG32_SOC15(GC, 0, regRLC_CGTT_MGCG_OVERRIDE, data);
+		}
+	} else {
+		if (adev->cg_flags & AMD_CG_SUPPORT_GFX_MGCG) {
+			def = data = RREG32_SOC15(GC, 0, regRLC_CGTT_MGCG_OVERRIDE);
+
+			data |= (RLC_CGTT_MGCG_OVERRIDE__RLC_CGTT_SCLK_OVERRIDE_MASK |
+				 RLC_CGTT_MGCG_OVERRIDE__GRBM_CGTT_SCLK_OVERRIDE_MASK |
+				 RLC_CGTT_MGCG_OVERRIDE__GFXIP_MGCG_OVERRIDE_MASK);
+
+			if (def != data)
+				WREG32_SOC15(GC, 0, regRLC_CGTT_MGCG_OVERRIDE, data);
+		}
+	}
+}
+
+static void gfx_v11_0_update_coarse_grain_clock_gating(struct amdgpu_device *adev,
+						       bool enable)
+{
+	uint32_t def, data;
+
+	if (!(adev->cg_flags &
+	      (AMD_CG_SUPPORT_GFX_CGCG |
+	      AMD_CG_SUPPORT_GFX_CGLS |
+	      AMD_CG_SUPPORT_GFX_3D_CGCG |
+	      AMD_CG_SUPPORT_GFX_3D_CGLS)))
+		return;
+
+	if (enable) {
+		def = data = RREG32_SOC15(GC, 0, regRLC_CGTT_MGCG_OVERRIDE);
+
+		/* unset CGCG override */
+		if (adev->cg_flags & AMD_CG_SUPPORT_GFX_CGCG)
+			data &= ~RLC_CGTT_MGCG_OVERRIDE__GFXIP_CGCG_OVERRIDE_MASK;
+		if (adev->cg_flags & AMD_CG_SUPPORT_GFX_CGLS)
+			data &= ~RLC_CGTT_MGCG_OVERRIDE__GFXIP_CGLS_OVERRIDE_MASK;
+		if (adev->cg_flags & AMD_CG_SUPPORT_GFX_3D_CGCG ||
+		    adev->cg_flags & AMD_CG_SUPPORT_GFX_3D_CGLS)
+			data &= ~RLC_CGTT_MGCG_OVERRIDE__GFXIP_GFX3D_CG_OVERRIDE_MASK;
+
+		/* update CGCG override bits */
+		if (def != data)
+			WREG32_SOC15(GC, 0, regRLC_CGTT_MGCG_OVERRIDE, data);
+
+		/* enable cgcg FSM(0x0000363F) */
+		def = data = RREG32_SOC15(GC, 0, regRLC_CGCG_CGLS_CTRL);
+
+		if (adev->cg_flags & AMD_CG_SUPPORT_GFX_CGCG) {
+			data &= ~RLC_CGCG_CGLS_CTRL__CGCG_GFX_IDLE_THRESHOLD_MASK;
+			data |= (0x36 << RLC_CGCG_CGLS_CTRL__CGCG_GFX_IDLE_THRESHOLD__SHIFT) |
+				 RLC_CGCG_CGLS_CTRL__CGCG_EN_MASK;
+		}
+
+		if (adev->cg_flags & AMD_CG_SUPPORT_GFX_CGLS) {
+			data &= ~RLC_CGCG_CGLS_CTRL__CGLS_REP_COMPANSAT_DELAY_MASK;
+			data |= (0x000F << RLC_CGCG_CGLS_CTRL__CGLS_REP_COMPANSAT_DELAY__SHIFT) |
+				 RLC_CGCG_CGLS_CTRL__CGLS_EN_MASK;
+		}
+
+		if (def != data)
+			WREG32_SOC15(GC, 0, regRLC_CGCG_CGLS_CTRL, data);
+
+		/* Program RLC_CGCG_CGLS_CTRL_3D */
+		def = data = RREG32_SOC15(GC, 0, regRLC_CGCG_CGLS_CTRL_3D);
+
+		if (adev->cg_flags & AMD_CG_SUPPORT_GFX_3D_CGCG) {
+			data &= ~RLC_CGCG_CGLS_CTRL_3D__CGCG_GFX_IDLE_THRESHOLD_MASK;
+			data |= (0x36 << RLC_CGCG_CGLS_CTRL_3D__CGCG_GFX_IDLE_THRESHOLD__SHIFT) |
+				 RLC_CGCG_CGLS_CTRL_3D__CGCG_EN_MASK;
+		}
+
+		if (adev->cg_flags & AMD_CG_SUPPORT_GFX_3D_CGLS) {
+			data &= ~RLC_CGCG_CGLS_CTRL_3D__CGLS_REP_COMPANSAT_DELAY_MASK;
+			data |= (0xf << RLC_CGCG_CGLS_CTRL_3D__CGLS_REP_COMPANSAT_DELAY__SHIFT) |
+				 RLC_CGCG_CGLS_CTRL_3D__CGLS_EN_MASK;
+		}
+
+		if (def != data)
+			WREG32_SOC15(GC, 0, regRLC_CGCG_CGLS_CTRL_3D, data);
+
+		/* set IDLE_POLL_COUNT(0x00900100) */
+		def = data = RREG32_SOC15(GC, 0, regCP_RB_WPTR_POLL_CNTL);
+
+		data &= ~(CP_RB_WPTR_POLL_CNTL__POLL_FREQUENCY_MASK | CP_RB_WPTR_POLL_CNTL__IDLE_POLL_COUNT_MASK);
+		data |= (0x0100 << CP_RB_WPTR_POLL_CNTL__POLL_FREQUENCY__SHIFT) |
+			(0x0090 << CP_RB_WPTR_POLL_CNTL__IDLE_POLL_COUNT__SHIFT);
+
+		if (def != data)
+			WREG32_SOC15(GC, 0, regCP_RB_WPTR_POLL_CNTL, data);
+
+		data = RREG32_SOC15(GC, 0, regCP_INT_CNTL);
+		data = REG_SET_FIELD(data, CP_INT_CNTL, CNTX_BUSY_INT_ENABLE, 1);
+		data = REG_SET_FIELD(data, CP_INT_CNTL, CNTX_EMPTY_INT_ENABLE, 1);
+		data = REG_SET_FIELD(data, CP_INT_CNTL, CMP_BUSY_INT_ENABLE, 1);
+		data = REG_SET_FIELD(data, CP_INT_CNTL, GFX_IDLE_INT_ENABLE, 1);
+		WREG32_SOC15(GC, 0, regCP_INT_CNTL, data);
+
+		data = RREG32_SOC15(GC, 0, regSDMA0_RLC_CGCG_CTRL);
+		data = REG_SET_FIELD(data, SDMA0_RLC_CGCG_CTRL, CGCG_INT_ENABLE, 1);
+		WREG32_SOC15(GC, 0, regSDMA0_RLC_CGCG_CTRL, data);
+
+		/* Some ASICs only have one SDMA instance, not need to configure SDMA1 */
+		if (adev->sdma.num_instances > 1) {
+			data = RREG32_SOC15(GC, 0, regSDMA1_RLC_CGCG_CTRL);
+			data = REG_SET_FIELD(data, SDMA1_RLC_CGCG_CTRL, CGCG_INT_ENABLE, 1);
+			WREG32_SOC15(GC, 0, regSDMA1_RLC_CGCG_CTRL, data);
+		}
+	} else {
+		/* Program RLC_CGCG_CGLS_CTRL */
+		def = data = RREG32_SOC15(GC, 0, regRLC_CGCG_CGLS_CTRL);
+
+		if (adev->cg_flags & AMD_CG_SUPPORT_GFX_CGCG)
+			data &= ~RLC_CGCG_CGLS_CTRL__CGCG_EN_MASK;
+
+		if (adev->cg_flags & AMD_CG_SUPPORT_GFX_CGLS)
+			data &= ~RLC_CGCG_CGLS_CTRL__CGLS_EN_MASK;
+
+		if (def != data)
+			WREG32_SOC15(GC, 0, regRLC_CGCG_CGLS_CTRL, data);
+
+		/* Program RLC_CGCG_CGLS_CTRL_3D */
+		def = data = RREG32_SOC15(GC, 0, regRLC_CGCG_CGLS_CTRL_3D);
+
+		if (adev->cg_flags & AMD_CG_SUPPORT_GFX_3D_CGCG)
+			data &= ~RLC_CGCG_CGLS_CTRL_3D__CGCG_EN_MASK;
+		if (adev->cg_flags & AMD_CG_SUPPORT_GFX_3D_CGLS)
+			data &= ~RLC_CGCG_CGLS_CTRL_3D__CGLS_EN_MASK;
+
+		if (def != data)
+			WREG32_SOC15(GC, 0, regRLC_CGCG_CGLS_CTRL_3D, data);
+
+		data = RREG32_SOC15(GC, 0, regSDMA0_RLC_CGCG_CTRL);
+		data &= ~SDMA0_RLC_CGCG_CTRL__CGCG_INT_ENABLE_MASK;
+		WREG32_SOC15(GC, 0, regSDMA0_RLC_CGCG_CTRL, data);
+
+		/* Some ASICs only have one SDMA instance, not need to configure SDMA1 */
+		if (adev->sdma.num_instances > 1) {
+			data = RREG32_SOC15(GC, 0, regSDMA1_RLC_CGCG_CTRL);
+			data &= ~SDMA1_RLC_CGCG_CTRL__CGCG_INT_ENABLE_MASK;
+			WREG32_SOC15(GC, 0, regSDMA1_RLC_CGCG_CTRL, data);
+		}
+	}
+}
+
+static int gfx_v11_0_update_gfx_clock_gating(struct amdgpu_device *adev,
+					    bool enable)
+{
+	amdgpu_gfx_rlc_enter_safe_mode(adev, 0);
+
+	gfx_v11_0_update_coarse_grain_clock_gating(adev, enable);
+
+	gfx_v11_0_update_medium_grain_clock_gating(adev, enable);
+
+	gfx_v11_0_update_repeater_fgcg(adev, enable);
+
+	gfx_v11_0_update_sram_fgcg(adev, enable);
+
+	gfx_v11_0_update_perf_clk(adev, enable);
+
+	if (adev->cg_flags &
+	    (AMD_CG_SUPPORT_GFX_MGCG |
+	     AMD_CG_SUPPORT_GFX_CGLS |
+	     AMD_CG_SUPPORT_GFX_CGCG |
+	     AMD_CG_SUPPORT_GFX_3D_CGCG |
+	     AMD_CG_SUPPORT_GFX_3D_CGLS))
+	        gfx_v11_0_enable_gui_idle_interrupt(adev, enable);
+
+	amdgpu_gfx_rlc_exit_safe_mode(adev, 0);
+
+	return 0;
+}
+
+static void gfx_v11_0_update_spm_vmid(struct amdgpu_device *adev, struct amdgpu_ring *ring, unsigned vmid)
+{
+	u32 reg, pre_data, data;
+
+	amdgpu_gfx_off_ctrl(adev, false);
+	reg = SOC15_REG_OFFSET(GC, 0, regRLC_SPM_MC_CNTL);
+	if (amdgpu_sriov_is_pp_one_vf(adev) && !amdgpu_sriov_runtime(adev))
+		pre_data = RREG32_NO_KIQ(reg);
+	else
+		pre_data = RREG32(reg);
+
+	data = pre_data & (~RLC_SPM_MC_CNTL__RLC_SPM_VMID_MASK);
+	data |= (vmid & RLC_SPM_MC_CNTL__RLC_SPM_VMID_MASK) << RLC_SPM_MC_CNTL__RLC_SPM_VMID__SHIFT;
+
+	if (pre_data != data) {
+		if (amdgpu_sriov_is_pp_one_vf(adev) && !amdgpu_sriov_runtime(adev)) {
+			WREG32_SOC15_NO_KIQ(GC, 0, regRLC_SPM_MC_CNTL, data);
+		} else
+			WREG32_SOC15(GC, 0, regRLC_SPM_MC_CNTL, data);
+	}
+	amdgpu_gfx_off_ctrl(adev, true);
+
+	if (ring
+		&& amdgpu_sriov_is_pp_one_vf(adev)
+		&& (pre_data != data)
+		&& ((ring->funcs->type == AMDGPU_RING_TYPE_GFX)
+			|| (ring->funcs->type == AMDGPU_RING_TYPE_COMPUTE))) {
+		amdgpu_ring_emit_wreg(ring, reg, data);
+	}
+}
+
+static const struct amdgpu_rlc_funcs gfx_v11_0_rlc_funcs = {
+	.is_rlc_enabled = gfx_v11_0_is_rlc_enabled,
+	.set_safe_mode = gfx_v11_0_set_safe_mode,
+	.unset_safe_mode = gfx_v11_0_unset_safe_mode,
+	.init = gfx_v11_0_rlc_init,
+	.get_csb_size = gfx_v11_0_get_csb_size,
+	.get_csb_buffer = gfx_v11_0_get_csb_buffer,
+	.resume = gfx_v11_0_rlc_resume,
+	.stop = gfx_v11_0_rlc_stop,
+	.reset = gfx_v11_0_rlc_reset,
+	.start = gfx_v11_0_rlc_start,
+	.update_spm_vmid = gfx_v11_0_update_spm_vmid,
+};
+
+static void gfx_v11_cntl_power_gating(struct amdgpu_device *adev, bool enable)
+{
+	u32 data = RREG32_SOC15(GC, 0, regRLC_PG_CNTL);
+
+	if (enable && (adev->pg_flags & AMD_PG_SUPPORT_GFX_PG))
+		data |= RLC_PG_CNTL__GFX_POWER_GATING_ENABLE_MASK;
+	else
+		data &= ~RLC_PG_CNTL__GFX_POWER_GATING_ENABLE_MASK;
+
+	WREG32_SOC15(GC, 0, regRLC_PG_CNTL, data);
+
+	// Program RLC_PG_DELAY3 for CGPG hysteresis
+	if (enable && (adev->pg_flags & AMD_PG_SUPPORT_GFX_PG)) {
+		switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+		case IP_VERSION(11, 0, 1):
+		case IP_VERSION(11, 0, 4):
+		case IP_VERSION(11, 5, 0):
+		case IP_VERSION(11, 5, 1):
+		case IP_VERSION(11, 5, 2):
+		case IP_VERSION(11, 5, 3):
+			WREG32_SOC15(GC, 0, regRLC_PG_DELAY_3, RLC_PG_DELAY_3_DEFAULT_GC_11_0_1);
+			break;
+		default:
+			break;
+		}
+	}
+}
+
+static void gfx_v11_cntl_pg(struct amdgpu_device *adev, bool enable)
+{
+	amdgpu_gfx_rlc_enter_safe_mode(adev, 0);
+
+	gfx_v11_cntl_power_gating(adev, enable);
+
+	amdgpu_gfx_rlc_exit_safe_mode(adev, 0);
+}
+
+static int gfx_v11_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
+					   enum amd_powergating_state state)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	bool enable = (state == AMD_PG_STATE_GATE);
+
+	if (amdgpu_sriov_vf(adev))
+		return 0;
+
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	case IP_VERSION(11, 0, 0):
+	case IP_VERSION(11, 0, 2):
+	case IP_VERSION(11, 0, 3):
+		amdgpu_gfx_off_ctrl(adev, enable);
+		break;
+	case IP_VERSION(11, 0, 1):
+	case IP_VERSION(11, 0, 4):
+	case IP_VERSION(11, 5, 0):
+	case IP_VERSION(11, 5, 1):
+	case IP_VERSION(11, 5, 2):
+	case IP_VERSION(11, 5, 3):
+		if (!enable)
+			amdgpu_gfx_off_ctrl(adev, false);
+
+		gfx_v11_cntl_pg(adev, enable);
+
+		if (enable)
+			amdgpu_gfx_off_ctrl(adev, true);
+
+		break;
+	default:
+		break;
+	}
+
+	return 0;
+}
+
+static int gfx_v11_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
+					  enum amd_clockgating_state state)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	if (amdgpu_sriov_vf(adev))
+	        return 0;
+
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	case IP_VERSION(11, 0, 0):
+	case IP_VERSION(11, 0, 1):
+	case IP_VERSION(11, 0, 2):
+	case IP_VERSION(11, 0, 3):
+	case IP_VERSION(11, 0, 4):
+	case IP_VERSION(11, 5, 0):
+	case IP_VERSION(11, 5, 1):
+	case IP_VERSION(11, 5, 2):
+	case IP_VERSION(11, 5, 3):
+	        gfx_v11_0_update_gfx_clock_gating(adev,
+	                        state ==  AMD_CG_STATE_GATE);
+	        break;
+	default:
+	        break;
+	}
+
+	return 0;
+}
+
+static void gfx_v11_0_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int data;
+
+	/* AMD_CG_SUPPORT_GFX_MGCG */
+	data = RREG32_SOC15(GC, 0, regRLC_CGTT_MGCG_OVERRIDE);
+	if (!(data & RLC_CGTT_MGCG_OVERRIDE__GFXIP_MGCG_OVERRIDE_MASK))
+		*flags |= AMD_CG_SUPPORT_GFX_MGCG;
+
+	/* AMD_CG_SUPPORT_REPEATER_FGCG */
+	if (!(data & RLC_CGTT_MGCG_OVERRIDE__GFXIP_REPEATER_FGCG_OVERRIDE_MASK))
+		*flags |= AMD_CG_SUPPORT_REPEATER_FGCG;
+
+	/* AMD_CG_SUPPORT_GFX_FGCG */
+	if (!(data & RLC_CGTT_MGCG_OVERRIDE__GFXIP_FGCG_OVERRIDE_MASK))
+		*flags |= AMD_CG_SUPPORT_GFX_FGCG;
+
+	/* AMD_CG_SUPPORT_GFX_PERF_CLK */
+	if (!(data & RLC_CGTT_MGCG_OVERRIDE__PERFMON_CLOCK_STATE_MASK))
+		*flags |= AMD_CG_SUPPORT_GFX_PERF_CLK;
+
+	/* AMD_CG_SUPPORT_GFX_CGCG */
+	data = RREG32_SOC15(GC, 0, regRLC_CGCG_CGLS_CTRL);
+	if (data & RLC_CGCG_CGLS_CTRL__CGCG_EN_MASK)
+		*flags |= AMD_CG_SUPPORT_GFX_CGCG;
+
+	/* AMD_CG_SUPPORT_GFX_CGLS */
+	if (data & RLC_CGCG_CGLS_CTRL__CGLS_EN_MASK)
+		*flags |= AMD_CG_SUPPORT_GFX_CGLS;
+
+	/* AMD_CG_SUPPORT_GFX_3D_CGCG */
+	data = RREG32_SOC15(GC, 0, regRLC_CGCG_CGLS_CTRL_3D);
+	if (data & RLC_CGCG_CGLS_CTRL_3D__CGCG_EN_MASK)
+		*flags |= AMD_CG_SUPPORT_GFX_3D_CGCG;
+
+	/* AMD_CG_SUPPORT_GFX_3D_CGLS */
+	if (data & RLC_CGCG_CGLS_CTRL_3D__CGLS_EN_MASK)
+		*flags |= AMD_CG_SUPPORT_GFX_3D_CGLS;
+}
+
+static u64 gfx_v11_0_ring_get_rptr_gfx(struct amdgpu_ring *ring)
+{
+	/* gfx11 is 32bit rptr*/
+	return *(uint32_t *)ring->rptr_cpu_addr;
+}
+
+static u64 gfx_v11_0_ring_get_wptr_gfx(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	u64 wptr;
+
+	/* XXX check if swapping is necessary on BE */
+	if (ring->use_doorbell) {
+		wptr = atomic64_read((atomic64_t *)ring->wptr_cpu_addr);
+	} else {
+		wptr = RREG32_SOC15(GC, 0, regCP_RB0_WPTR);
+		wptr += (u64)RREG32_SOC15(GC, 0, regCP_RB0_WPTR_HI) << 32;
+	}
+
+	return wptr;
+}
+
+static void gfx_v11_0_ring_set_wptr_gfx(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	if (ring->use_doorbell) {
+		/* XXX check if swapping is necessary on BE */
+		atomic64_set((atomic64_t *)ring->wptr_cpu_addr,
+			     ring->wptr);
+		WDOORBELL64(ring->doorbell_index, ring->wptr);
+	} else {
+		WREG32_SOC15(GC, 0, regCP_RB0_WPTR,
+			     lower_32_bits(ring->wptr));
+		WREG32_SOC15(GC, 0, regCP_RB0_WPTR_HI,
+			     upper_32_bits(ring->wptr));
+	}
+}
+
+static u64 gfx_v11_0_ring_get_rptr_compute(struct amdgpu_ring *ring)
+{
+	/* gfx11 hardware is 32bit rptr */
+	return *(uint32_t *)ring->rptr_cpu_addr;
+}
+
+static u64 gfx_v11_0_ring_get_wptr_compute(struct amdgpu_ring *ring)
+{
+	u64 wptr;
+
+	/* XXX check if swapping is necessary on BE */
+	if (ring->use_doorbell)
+		wptr = atomic64_read((atomic64_t *)ring->wptr_cpu_addr);
+	else
+		BUG();
+	return wptr;
+}
+
+static void gfx_v11_0_ring_set_wptr_compute(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	/* XXX check if swapping is necessary on BE */
+	if (ring->use_doorbell) {
+		atomic64_set((atomic64_t *)ring->wptr_cpu_addr,
+			     ring->wptr);
+		WDOORBELL64(ring->doorbell_index, ring->wptr);
+	} else {
+		BUG(); /* only DOORBELL method supported on gfx11 now */
+	}
+}
+
+static void gfx_v11_0_ring_emit_hdp_flush(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	u32 ref_and_mask, reg_mem_engine;
+	const struct nbio_hdp_flush_reg *nbio_hf_reg = adev->nbio.hdp_flush_reg;
+
+	if (ring->funcs->type == AMDGPU_RING_TYPE_COMPUTE) {
+		switch (ring->me) {
+		case 1:
+			ref_and_mask = nbio_hf_reg->ref_and_mask_cp2 << ring->pipe;
+			break;
+		case 2:
+			ref_and_mask = nbio_hf_reg->ref_and_mask_cp6 << ring->pipe;
+			break;
+		default:
+			return;
+		}
+		reg_mem_engine = 0;
+	} else {
+		ref_and_mask = nbio_hf_reg->ref_and_mask_cp0 << ring->pipe;
+		reg_mem_engine = 1; /* pfp */
+	}
+
+	gfx_v11_0_wait_reg_mem(ring, reg_mem_engine, 0, 1,
+			       adev->nbio.funcs->get_hdp_flush_req_offset(adev),
+			       adev->nbio.funcs->get_hdp_flush_done_offset(adev),
+			       ref_and_mask, ref_and_mask, 0x20);
+}
+
+static void gfx_v11_0_ring_emit_ib_gfx(struct amdgpu_ring *ring,
+				       struct amdgpu_job *job,
+				       struct amdgpu_ib *ib,
+				       uint32_t flags)
+{
+	unsigned vmid = AMDGPU_JOB_GET_VMID(job);
+	u32 header, control = 0;
+
+	header = PACKET3(PACKET3_INDIRECT_BUFFER, 2);
+
+	control |= ib->length_dw | (vmid << 24);
+
+	if (ring->adev->gfx.mcbp && (ib->flags & AMDGPU_IB_FLAG_PREEMPT)) {
+		control |= INDIRECT_BUFFER_PRE_ENB(1);
+
+		if (flags & AMDGPU_IB_PREEMPTED)
+			control |= INDIRECT_BUFFER_PRE_RESUME(1);
+
+		if (vmid && !ring->adev->gfx.rs64_enable)
+			gfx_v11_0_ring_emit_de_meta(ring,
+				!amdgpu_sriov_vf(ring->adev) && (flags & AMDGPU_IB_PREEMPTED));
+	}
+
+	amdgpu_ring_write(ring, header);
+	BUG_ON(ib->gpu_addr & 0x3); /* Dword align */
+	amdgpu_ring_write(ring,
+#ifdef __BIG_ENDIAN
+		(2 << 0) |
+#endif
+		lower_32_bits(ib->gpu_addr));
+	amdgpu_ring_write(ring, upper_32_bits(ib->gpu_addr));
+	amdgpu_ring_write(ring, control);
+}
+
+static void gfx_v11_0_ring_emit_ib_compute(struct amdgpu_ring *ring,
+					   struct amdgpu_job *job,
+					   struct amdgpu_ib *ib,
+					   uint32_t flags)
+{
+	unsigned vmid = AMDGPU_JOB_GET_VMID(job);
+	u32 control = INDIRECT_BUFFER_VALID | ib->length_dw | (vmid << 24);
+
+	/* Currently, there is a high possibility to get wave ID mismatch
+	 * between ME and GDS, leading to a hw deadlock, because ME generates
+	 * different wave IDs than the GDS expects. This situation happens
+	 * randomly when at least 5 compute pipes use GDS ordered append.
+	 * The wave IDs generated by ME are also wrong after suspend/resume.
+	 * Those are probably bugs somewhere else in the kernel driver.
+	 *
+	 * Writing GDS_COMPUTE_MAX_WAVE_ID resets wave ID counters in ME and
+	 * GDS to 0 for this ring (me/pipe).
+	 */
+	if (ib->flags & AMDGPU_IB_FLAG_RESET_GDS_MAX_WAVE_ID) {
+		amdgpu_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1));
+		amdgpu_ring_write(ring, regGDS_COMPUTE_MAX_WAVE_ID);
+		amdgpu_ring_write(ring, ring->adev->gds.gds_compute_max_wave_id);
+	}
+
+	amdgpu_ring_write(ring, PACKET3(PACKET3_INDIRECT_BUFFER, 2));
+	BUG_ON(ib->gpu_addr & 0x3); /* Dword align */
+	amdgpu_ring_write(ring,
+#ifdef __BIG_ENDIAN
+				(2 << 0) |
+#endif
+				lower_32_bits(ib->gpu_addr));
+	amdgpu_ring_write(ring, upper_32_bits(ib->gpu_addr));
+	amdgpu_ring_write(ring, control);
+}
+
+static void gfx_v11_0_ring_emit_fence(struct amdgpu_ring *ring, u64 addr,
+				     u64 seq, unsigned flags)
+{
+	bool write64bit = flags & AMDGPU_FENCE_FLAG_64BIT;
+	bool int_sel = flags & AMDGPU_FENCE_FLAG_INT;
+
+	/* RELEASE_MEM - flush caches, send int */
+	amdgpu_ring_write(ring, PACKET3(PACKET3_RELEASE_MEM, 6));
+	amdgpu_ring_write(ring, (PACKET3_RELEASE_MEM_GCR_SEQ |
+				 PACKET3_RELEASE_MEM_GCR_GL2_WB |
+				 PACKET3_RELEASE_MEM_GCR_GLM_INV | /* must be set with GLM_WB */
+				 PACKET3_RELEASE_MEM_GCR_GLM_WB |
+				 PACKET3_RELEASE_MEM_CACHE_POLICY(3) |
+				 PACKET3_RELEASE_MEM_EVENT_TYPE(CACHE_FLUSH_AND_INV_TS_EVENT) |
+				 PACKET3_RELEASE_MEM_EVENT_INDEX(5)));
+	amdgpu_ring_write(ring, (PACKET3_RELEASE_MEM_DATA_SEL(write64bit ? 2 : 1) |
+				 PACKET3_RELEASE_MEM_INT_SEL(int_sel ? 2 : 0)));
+
+	/*
+	 * the address should be Qword aligned if 64bit write, Dword
+	 * aligned if only send 32bit data low (discard data high)
+	 */
+	if (write64bit)
+		BUG_ON(addr & 0x7);
+	else
+		BUG_ON(addr & 0x3);
+	amdgpu_ring_write(ring, lower_32_bits(addr));
+	amdgpu_ring_write(ring, upper_32_bits(addr));
+	amdgpu_ring_write(ring, lower_32_bits(seq));
+	amdgpu_ring_write(ring, upper_32_bits(seq));
+	amdgpu_ring_write(ring, 0);
+}
+
+static void gfx_v11_0_ring_emit_pipeline_sync(struct amdgpu_ring *ring)
+{
+	int usepfp = (ring->funcs->type == AMDGPU_RING_TYPE_GFX);
+	uint32_t seq = ring->fence_drv.sync_seq;
+	uint64_t addr = ring->fence_drv.gpu_addr;
+
+	gfx_v11_0_wait_reg_mem(ring, usepfp, 1, 0, lower_32_bits(addr),
+			       upper_32_bits(addr), seq, 0xffffffff, 4);
+}
+
+static void gfx_v11_0_ring_invalidate_tlbs(struct amdgpu_ring *ring,
+				   uint16_t pasid, uint32_t flush_type,
+				   bool all_hub, uint8_t dst_sel)
+{
+	amdgpu_ring_write(ring, PACKET3(PACKET3_INVALIDATE_TLBS, 0));
+	amdgpu_ring_write(ring,
+			  PACKET3_INVALIDATE_TLBS_DST_SEL(dst_sel) |
+			  PACKET3_INVALIDATE_TLBS_ALL_HUB(all_hub) |
+			  PACKET3_INVALIDATE_TLBS_PASID(pasid) |
+			  PACKET3_INVALIDATE_TLBS_FLUSH_TYPE(flush_type));
+}
+
+static void gfx_v11_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
+					 unsigned vmid, uint64_t pd_addr)
+{
+	amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pd_addr);
+
+	/* compute doesn't have PFP */
+	if (ring->funcs->type == AMDGPU_RING_TYPE_GFX) {
+		/* sync PFP to ME, otherwise we might get invalid PFP reads */
+		amdgpu_ring_write(ring, PACKET3(PACKET3_PFP_SYNC_ME, 0));
+		amdgpu_ring_write(ring, 0x0);
+	}
+
+	/* Make sure that we can't skip the SET_Q_MODE packets when the VM
+	 * changed in any way.
+	 */
+	ring->set_q_mode_offs = 0;
+	ring->set_q_mode_ptr = NULL;
+}
+
+static void gfx_v11_0_ring_emit_fence_kiq(struct amdgpu_ring *ring, u64 addr,
+					  u64 seq, unsigned int flags)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	/* we only allocate 32bit for each seq wb address */
+	BUG_ON(flags & AMDGPU_FENCE_FLAG_64BIT);
+
+	/* write fence seq to the "addr" */
+	amdgpu_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
+	amdgpu_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) |
+				 WRITE_DATA_DST_SEL(5) | WR_CONFIRM));
+	amdgpu_ring_write(ring, lower_32_bits(addr));
+	amdgpu_ring_write(ring, upper_32_bits(addr));
+	amdgpu_ring_write(ring, lower_32_bits(seq));
+
+	if (flags & AMDGPU_FENCE_FLAG_INT) {
+		/* set register to trigger INT */
+		amdgpu_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
+		amdgpu_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) |
+					 WRITE_DATA_DST_SEL(0) | WR_CONFIRM));
+		amdgpu_ring_write(ring, SOC15_REG_OFFSET(GC, 0, regCPC_INT_STATUS));
+		amdgpu_ring_write(ring, 0);
+		amdgpu_ring_write(ring, 0x20000000); /* src_id is 178 */
+	}
+}
+
+static void gfx_v11_0_ring_emit_cntxcntl(struct amdgpu_ring *ring,
+					 uint32_t flags)
+{
+	uint32_t dw2 = 0;
+
+	dw2 |= 0x80000000; /* set load_enable otherwise this package is just NOPs */
+	if (flags & AMDGPU_HAVE_CTX_SWITCH) {
+		/* set load_global_config & load_global_uconfig */
+		dw2 |= 0x8001;
+		/* set load_cs_sh_regs */
+		dw2 |= 0x01000000;
+		/* set load_per_context_state & load_gfx_sh_regs for GFX */
+		dw2 |= 0x10002;
+	}
+
+	amdgpu_ring_write(ring, PACKET3(PACKET3_CONTEXT_CONTROL, 1));
+	amdgpu_ring_write(ring, dw2);
+	amdgpu_ring_write(ring, 0);
+}
+
+static unsigned gfx_v11_0_ring_emit_init_cond_exec(struct amdgpu_ring *ring,
+						   uint64_t addr)
+{
+	unsigned ret;
+
+	amdgpu_ring_write(ring, PACKET3(PACKET3_COND_EXEC, 3));
+	amdgpu_ring_write(ring, lower_32_bits(addr));
+	amdgpu_ring_write(ring, upper_32_bits(addr));
+	/* discard following DWs if *cond_exec_gpu_addr==0 */
+	amdgpu_ring_write(ring, 0);
+	ret = ring->wptr & ring->buf_mask;
+	/* patch dummy value later */
+	amdgpu_ring_write(ring, 0);
+
+	return ret;
+}
+
+static void gfx_v11_0_ring_emit_gfx_shadow(struct amdgpu_ring *ring,
+					   u64 shadow_va, u64 csa_va,
+					   u64 gds_va, bool init_shadow,
+					   int vmid)
+{
+	struct amdgpu_device *adev = ring->adev;
+	unsigned int offs, end;
+
+	if (!adev->gfx.cp_gfx_shadow || !ring->ring_obj)
+		return;
+
+	/*
+	 * The logic here isn't easy to understand because we need to keep state
+	 * accross multiple executions of the function as well as between the
+	 * CPU and GPU. The general idea is that the newly written GPU command
+	 * has a condition on the previous one and only executed if really
+	 * necessary.
+	 */
+
+	/*
+	 * The dw in the NOP controls if the next SET_Q_MODE packet should be
+	 * executed or not. Reserve 64bits just to be on the save side.
+	 */
+	amdgpu_ring_write(ring, PACKET3(PACKET3_NOP, 1));
+	offs = ring->wptr & ring->buf_mask;
+
+	/*
+	 * We start with skipping the prefix SET_Q_MODE and always executing
+	 * the postfix SET_Q_MODE packet. This is changed below with a
+	 * WRITE_DATA command when the postfix executed.
+	 */
+	amdgpu_ring_write(ring, shadow_va ? 1 : 0);
+	amdgpu_ring_write(ring, 0);
+
+	if (ring->set_q_mode_offs) {
+		uint64_t addr;
+
+		addr = amdgpu_bo_gpu_offset(ring->ring_obj);
+		addr += ring->set_q_mode_offs << 2;
+		end = gfx_v11_0_ring_emit_init_cond_exec(ring, addr);
+	}
+
+	/*
+	 * When the postfix SET_Q_MODE packet executes we need to make sure that the
+	 * next prefix SET_Q_MODE packet executes as well.
+	 */
+	if (!shadow_va) {
+		uint64_t addr;
+
+		addr = amdgpu_bo_gpu_offset(ring->ring_obj);
+		addr += offs << 2;
+		amdgpu_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
+		amdgpu_ring_write(ring, WRITE_DATA_DST_SEL(5) | WR_CONFIRM);
+		amdgpu_ring_write(ring, lower_32_bits(addr));
+		amdgpu_ring_write(ring, upper_32_bits(addr));
+		amdgpu_ring_write(ring, 0x1);
+	}
+
+	amdgpu_ring_write(ring, PACKET3(PACKET3_SET_Q_PREEMPTION_MODE, 7));
+	amdgpu_ring_write(ring, lower_32_bits(shadow_va));
+	amdgpu_ring_write(ring, upper_32_bits(shadow_va));
+	amdgpu_ring_write(ring, lower_32_bits(gds_va));
+	amdgpu_ring_write(ring, upper_32_bits(gds_va));
+	amdgpu_ring_write(ring, lower_32_bits(csa_va));
+	amdgpu_ring_write(ring, upper_32_bits(csa_va));
+	amdgpu_ring_write(ring, shadow_va ?
+			  PACKET3_SET_Q_PREEMPTION_MODE_IB_VMID(vmid) : 0);
+	amdgpu_ring_write(ring, init_shadow ?
+			  PACKET3_SET_Q_PREEMPTION_MODE_INIT_SHADOW_MEM : 0);
+
+	if (ring->set_q_mode_offs)
+		amdgpu_ring_patch_cond_exec(ring, end);
+
+	if (shadow_va) {
+		uint64_t token = shadow_va ^ csa_va ^ gds_va ^ vmid;
+
+		/*
+		 * If the tokens match try to skip the last postfix SET_Q_MODE
+		 * packet to avoid saving/restoring the state all the time.
+		 */
+		if (ring->set_q_mode_ptr && ring->set_q_mode_token == token)
+			*ring->set_q_mode_ptr = 0;
+
+		ring->set_q_mode_token = token;
+	} else {
+		ring->set_q_mode_ptr = &ring->ring[ring->set_q_mode_offs];
+	}
+
+	ring->set_q_mode_offs = offs;
+}
+
+static int gfx_v11_0_ring_preempt_ib(struct amdgpu_ring *ring)
+{
+	int i, r = 0;
+	struct amdgpu_device *adev = ring->adev;
+	struct amdgpu_kiq *kiq = &adev->gfx.kiq[0];
+	struct amdgpu_ring *kiq_ring = &kiq->ring;
+	unsigned long flags;
+
+	if (adev->enable_mes)
+		return -EINVAL;
+
+	if (!kiq->pmf || !kiq->pmf->kiq_unmap_queues)
+		return -EINVAL;
+
+	spin_lock_irqsave(&kiq->ring_lock, flags);
+
+	if (amdgpu_ring_alloc(kiq_ring, kiq->pmf->unmap_queues_size)) {
+		spin_unlock_irqrestore(&kiq->ring_lock, flags);
+		return -ENOMEM;
+	}
+
+	/* assert preemption condition */
+	amdgpu_ring_set_preempt_cond_exec(ring, false);
+
+	/* assert IB preemption, emit the trailing fence */
+	kiq->pmf->kiq_unmap_queues(kiq_ring, ring, PREEMPT_QUEUES_NO_UNMAP,
+				   ring->trail_fence_gpu_addr,
+				   ++ring->trail_seq);
+	amdgpu_ring_commit(kiq_ring);
+
+	spin_unlock_irqrestore(&kiq->ring_lock, flags);
+
+	/* poll the trailing fence */
+	for (i = 0; i < adev->usec_timeout; i++) {
+		if (ring->trail_seq ==
+		    le32_to_cpu(*(ring->trail_fence_cpu_addr)))
+			break;
+		udelay(1);
+	}
+
+	if (i >= adev->usec_timeout) {
+		r = -EINVAL;
+		DRM_ERROR("ring %d failed to preempt ib\n", ring->idx);
+	}
+
+	/* deassert preemption condition */
+	amdgpu_ring_set_preempt_cond_exec(ring, true);
+	return r;
+}
+
+static void gfx_v11_0_ring_emit_de_meta(struct amdgpu_ring *ring, bool resume)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct v10_de_ib_state de_payload = {0};
+	uint64_t offset, gds_addr, de_payload_gpu_addr;
+	void *de_payload_cpu_addr;
+	int cnt;
+
+	offset = offsetof(struct v10_gfx_meta_data, de_payload);
+	de_payload_gpu_addr = amdgpu_csa_vaddr(ring->adev) + offset;
+	de_payload_cpu_addr = adev->virt.csa_cpu_addr + offset;
+
+	gds_addr = ALIGN(amdgpu_csa_vaddr(ring->adev) +
+			 AMDGPU_CSA_SIZE - adev->gds.gds_size,
+			 PAGE_SIZE);
+
+	de_payload.gds_backup_addrlo = lower_32_bits(gds_addr);
+	de_payload.gds_backup_addrhi = upper_32_bits(gds_addr);
+
+	cnt = (sizeof(de_payload) >> 2) + 4 - 2;
+	amdgpu_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, cnt));
+	amdgpu_ring_write(ring, (WRITE_DATA_ENGINE_SEL(1) |
+				 WRITE_DATA_DST_SEL(8) |
+				 WR_CONFIRM) |
+				 WRITE_DATA_CACHE_POLICY(0));
+	amdgpu_ring_write(ring, lower_32_bits(de_payload_gpu_addr));
+	amdgpu_ring_write(ring, upper_32_bits(de_payload_gpu_addr));
+
+	if (resume)
+		amdgpu_ring_write_multiple(ring, de_payload_cpu_addr,
+					   sizeof(de_payload) >> 2);
+	else
+		amdgpu_ring_write_multiple(ring, (void *)&de_payload,
+					   sizeof(de_payload) >> 2);
+}
+
+static void gfx_v11_0_ring_emit_frame_cntl(struct amdgpu_ring *ring, bool start,
+				    bool secure)
+{
+	uint32_t v = secure ? FRAME_TMZ : 0;
+
+	amdgpu_ring_write(ring, PACKET3(PACKET3_FRAME_CONTROL, 0));
+	amdgpu_ring_write(ring, v | FRAME_CMD(start ? 0 : 1));
+}
+
+static void gfx_v11_0_ring_emit_rreg(struct amdgpu_ring *ring, uint32_t reg,
+				     uint32_t reg_val_offs)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	amdgpu_ring_write(ring, PACKET3(PACKET3_COPY_DATA, 4));
+	amdgpu_ring_write(ring, 0 |	/* src: register*/
+				(5 << 8) |	/* dst: memory */
+				(1 << 20));	/* write confirm */
+	amdgpu_ring_write(ring, reg);
+	amdgpu_ring_write(ring, 0);
+	amdgpu_ring_write(ring, lower_32_bits(adev->wb.gpu_addr +
+				reg_val_offs * 4));
+	amdgpu_ring_write(ring, upper_32_bits(adev->wb.gpu_addr +
+				reg_val_offs * 4));
+}
+
+static void gfx_v11_0_ring_emit_wreg(struct amdgpu_ring *ring, uint32_t reg,
+				   uint32_t val)
+{
+	uint32_t cmd = 0;
+
+	switch (ring->funcs->type) {
+	case AMDGPU_RING_TYPE_GFX:
+		cmd = WRITE_DATA_ENGINE_SEL(1) | WR_CONFIRM;
+		break;
+	case AMDGPU_RING_TYPE_KIQ:
+		cmd = (1 << 16); /* no inc addr */
+		break;
+	default:
+		cmd = WR_CONFIRM;
+		break;
+	}
+	amdgpu_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
+	amdgpu_ring_write(ring, cmd);
+	amdgpu_ring_write(ring, reg);
+	amdgpu_ring_write(ring, 0);
+	amdgpu_ring_write(ring, val);
+}
+
+static void gfx_v11_0_ring_emit_reg_wait(struct amdgpu_ring *ring, uint32_t reg,
+					uint32_t val, uint32_t mask)
+{
+	gfx_v11_0_wait_reg_mem(ring, 0, 0, 0, reg, 0, val, mask, 0x20);
+}
+
+static void gfx_v11_0_ring_emit_reg_write_reg_wait(struct amdgpu_ring *ring,
+						   uint32_t reg0, uint32_t reg1,
+						   uint32_t ref, uint32_t mask)
+{
+	int usepfp = (ring->funcs->type == AMDGPU_RING_TYPE_GFX);
+
+	gfx_v11_0_wait_reg_mem(ring, usepfp, 0, 1, reg0, reg1,
+			       ref, mask, 0x20);
+}
+
+static void
+gfx_v11_0_set_gfx_eop_interrupt_state(struct amdgpu_device *adev,
+				      uint32_t me, uint32_t pipe,
+				      enum amdgpu_interrupt_state state)
+{
+	uint32_t cp_int_cntl, cp_int_cntl_reg;
+
+	if (!me) {
+		switch (pipe) {
+		case 0:
+			cp_int_cntl_reg = SOC15_REG_OFFSET(GC, 0, regCP_INT_CNTL_RING0);
+			break;
+		case 1:
+			cp_int_cntl_reg = SOC15_REG_OFFSET(GC, 0, regCP_INT_CNTL_RING1);
+			break;
+		default:
+			DRM_DEBUG("invalid pipe %d\n", pipe);
+			return;
+		}
+	} else {
+		DRM_DEBUG("invalid me %d\n", me);
+		return;
+	}
+
+	switch (state) {
+	case AMDGPU_IRQ_STATE_DISABLE:
+		cp_int_cntl = RREG32_SOC15_IP(GC, cp_int_cntl_reg);
+		cp_int_cntl = REG_SET_FIELD(cp_int_cntl, CP_INT_CNTL_RING0,
+					    TIME_STAMP_INT_ENABLE, 0);
+		cp_int_cntl = REG_SET_FIELD(cp_int_cntl, CP_INT_CNTL_RING0,
+					    GENERIC0_INT_ENABLE, 0);
+		WREG32_SOC15_IP(GC, cp_int_cntl_reg, cp_int_cntl);
+		break;
+	case AMDGPU_IRQ_STATE_ENABLE:
+		cp_int_cntl = RREG32_SOC15_IP(GC, cp_int_cntl_reg);
+		cp_int_cntl = REG_SET_FIELD(cp_int_cntl, CP_INT_CNTL_RING0,
+					    TIME_STAMP_INT_ENABLE, 1);
+		cp_int_cntl = REG_SET_FIELD(cp_int_cntl, CP_INT_CNTL_RING0,
+					    GENERIC0_INT_ENABLE, 1);
+		WREG32_SOC15_IP(GC, cp_int_cntl_reg, cp_int_cntl);
+		break;
+	default:
+		break;
+	}
+}
+
+static void gfx_v11_0_set_compute_eop_interrupt_state(struct amdgpu_device *adev,
+						     int me, int pipe,
+						     enum amdgpu_interrupt_state state)
+{
+	u32 mec_int_cntl, mec_int_cntl_reg;
+
+	/*
+	 * amdgpu controls only the first MEC. That's why this function only
+	 * handles the setting of interrupts for this specific MEC. All other
+	 * pipes' interrupts are set by amdkfd.
+	 */
+
+	if (me == 1) {
+		switch (pipe) {
+		case 0:
+			mec_int_cntl_reg = SOC15_REG_OFFSET(GC, 0, regCP_ME1_PIPE0_INT_CNTL);
+			break;
+		case 1:
+			mec_int_cntl_reg = SOC15_REG_OFFSET(GC, 0, regCP_ME1_PIPE1_INT_CNTL);
+			break;
+		case 2:
+			mec_int_cntl_reg = SOC15_REG_OFFSET(GC, 0, regCP_ME1_PIPE2_INT_CNTL);
+			break;
+		case 3:
+			mec_int_cntl_reg = SOC15_REG_OFFSET(GC, 0, regCP_ME1_PIPE3_INT_CNTL);
+			break;
+		default:
+			DRM_DEBUG("invalid pipe %d\n", pipe);
+			return;
+		}
+	} else {
+		DRM_DEBUG("invalid me %d\n", me);
+		return;
+	}
+
+	switch (state) {
+	case AMDGPU_IRQ_STATE_DISABLE:
+		mec_int_cntl = RREG32_SOC15_IP(GC, mec_int_cntl_reg);
+		mec_int_cntl = REG_SET_FIELD(mec_int_cntl, CP_ME1_PIPE0_INT_CNTL,
+					     TIME_STAMP_INT_ENABLE, 0);
+		mec_int_cntl = REG_SET_FIELD(mec_int_cntl, CP_ME1_PIPE0_INT_CNTL,
+					     GENERIC0_INT_ENABLE, 0);
+		WREG32_SOC15_IP(GC, mec_int_cntl_reg, mec_int_cntl);
+		break;
+	case AMDGPU_IRQ_STATE_ENABLE:
+		mec_int_cntl = RREG32_SOC15_IP(GC, mec_int_cntl_reg);
+		mec_int_cntl = REG_SET_FIELD(mec_int_cntl, CP_ME1_PIPE0_INT_CNTL,
+					     TIME_STAMP_INT_ENABLE, 1);
+		mec_int_cntl = REG_SET_FIELD(mec_int_cntl, CP_ME1_PIPE0_INT_CNTL,
+					     GENERIC0_INT_ENABLE, 1);
+		WREG32_SOC15_IP(GC, mec_int_cntl_reg, mec_int_cntl);
+		break;
+	default:
+		break;
+	}
+}
+
+static int gfx_v11_0_set_eop_interrupt_state(struct amdgpu_device *adev,
+					    struct amdgpu_irq_src *src,
+					    unsigned type,
+					    enum amdgpu_interrupt_state state)
+{
+	switch (type) {
+	case AMDGPU_CP_IRQ_GFX_ME0_PIPE0_EOP:
+		gfx_v11_0_set_gfx_eop_interrupt_state(adev, 0, 0, state);
+		break;
+	case AMDGPU_CP_IRQ_GFX_ME0_PIPE1_EOP:
+		gfx_v11_0_set_gfx_eop_interrupt_state(adev, 0, 1, state);
+		break;
+	case AMDGPU_CP_IRQ_COMPUTE_MEC1_PIPE0_EOP:
+		gfx_v11_0_set_compute_eop_interrupt_state(adev, 1, 0, state);
+		break;
+	case AMDGPU_CP_IRQ_COMPUTE_MEC1_PIPE1_EOP:
+		gfx_v11_0_set_compute_eop_interrupt_state(adev, 1, 1, state);
+		break;
+	case AMDGPU_CP_IRQ_COMPUTE_MEC1_PIPE2_EOP:
+		gfx_v11_0_set_compute_eop_interrupt_state(adev, 1, 2, state);
+		break;
+	case AMDGPU_CP_IRQ_COMPUTE_MEC1_PIPE3_EOP:
+		gfx_v11_0_set_compute_eop_interrupt_state(adev, 1, 3, state);
+		break;
+	default:
+		break;
+	}
+	return 0;
+}
+
+static int gfx_v11_0_eop_irq(struct amdgpu_device *adev,
+			     struct amdgpu_irq_src *source,
+			     struct amdgpu_iv_entry *entry)
+{
+	u32 doorbell_offset = entry->src_data[0];
+	u8 me_id, pipe_id, queue_id;
+	struct amdgpu_ring *ring;
+	int i;
+
+	DRM_DEBUG("IH: CP EOP\n");
+
+	if (adev->enable_mes && doorbell_offset) {
+		struct amdgpu_userq_fence_driver *fence_drv = NULL;
+		struct xarray *xa = &adev->userq_xa;
+		unsigned long flags;
+
+		xa_lock_irqsave(xa, flags);
+		fence_drv = xa_load(xa, doorbell_offset);
+		if (fence_drv)
+			amdgpu_userq_fence_driver_process(fence_drv);
+		xa_unlock_irqrestore(xa, flags);
+	} else {
+		me_id = (entry->ring_id & 0x0c) >> 2;
+		pipe_id = (entry->ring_id & 0x03) >> 0;
+		queue_id = (entry->ring_id & 0x70) >> 4;
+
+		switch (me_id) {
+		case 0:
+			if (pipe_id == 0)
+				amdgpu_fence_process(&adev->gfx.gfx_ring[0]);
+			else
+				amdgpu_fence_process(&adev->gfx.gfx_ring[1]);
+			break;
+		case 1:
+		case 2:
+			for (i = 0; i < adev->gfx.num_compute_rings; i++) {
+				ring = &adev->gfx.compute_ring[i];
+				/* Per-queue interrupt is supported for MEC starting from VI.
+				 * The interrupt can only be enabled/disabled per pipe instead
+				 * of per queue.
+				 */
+				if ((ring->me == me_id) &&
+				    (ring->pipe == pipe_id) &&
+				    (ring->queue == queue_id))
+					amdgpu_fence_process(ring);
+			}
+			break;
+		}
+	}
+
+	return 0;
+}
+
+static int gfx_v11_0_set_priv_reg_fault_state(struct amdgpu_device *adev,
+					      struct amdgpu_irq_src *source,
+					      unsigned int type,
+					      enum amdgpu_interrupt_state state)
+{
+	u32 cp_int_cntl_reg, cp_int_cntl;
+	int i, j;
+
+	switch (state) {
+	case AMDGPU_IRQ_STATE_DISABLE:
+	case AMDGPU_IRQ_STATE_ENABLE:
+		for (i = 0; i < adev->gfx.me.num_me; i++) {
+			for (j = 0; j < adev->gfx.me.num_pipe_per_me; j++) {
+				cp_int_cntl_reg = gfx_v11_0_get_cpg_int_cntl(adev, i, j);
+
+				if (cp_int_cntl_reg) {
+					cp_int_cntl = RREG32_SOC15_IP(GC, cp_int_cntl_reg);
+					cp_int_cntl = REG_SET_FIELD(cp_int_cntl, CP_INT_CNTL_RING0,
+								    PRIV_REG_INT_ENABLE,
+								    state == AMDGPU_IRQ_STATE_ENABLE ? 1 : 0);
+					WREG32_SOC15_IP(GC, cp_int_cntl_reg, cp_int_cntl);
+				}
+			}
+		}
+		for (i = 0; i < adev->gfx.mec.num_mec; i++) {
+			for (j = 0; j < adev->gfx.mec.num_pipe_per_mec; j++) {
+				/* MECs start at 1 */
+				cp_int_cntl_reg = gfx_v11_0_get_cpc_int_cntl(adev, i + 1, j);
+
+				if (cp_int_cntl_reg) {
+					cp_int_cntl = RREG32_SOC15_IP(GC, cp_int_cntl_reg);
+					cp_int_cntl = REG_SET_FIELD(cp_int_cntl, CP_ME1_PIPE0_INT_CNTL,
+								    PRIV_REG_INT_ENABLE,
+								    state == AMDGPU_IRQ_STATE_ENABLE ? 1 : 0);
+					WREG32_SOC15_IP(GC, cp_int_cntl_reg, cp_int_cntl);
+				}
+			}
+		}
+		break;
+	default:
+		break;
+	}
+
+	return 0;
+}
+
+static int gfx_v11_0_set_bad_op_fault_state(struct amdgpu_device *adev,
+					    struct amdgpu_irq_src *source,
+					    unsigned type,
+					    enum amdgpu_interrupt_state state)
+{
+	u32 cp_int_cntl_reg, cp_int_cntl;
+	int i, j;
+
+	switch (state) {
+	case AMDGPU_IRQ_STATE_DISABLE:
+	case AMDGPU_IRQ_STATE_ENABLE:
+		for (i = 0; i < adev->gfx.me.num_me; i++) {
+			for (j = 0; j < adev->gfx.me.num_pipe_per_me; j++) {
+				cp_int_cntl_reg = gfx_v11_0_get_cpg_int_cntl(adev, i, j);
+
+				if (cp_int_cntl_reg) {
+					cp_int_cntl = RREG32_SOC15_IP(GC, cp_int_cntl_reg);
+					cp_int_cntl = REG_SET_FIELD(cp_int_cntl, CP_INT_CNTL_RING0,
+								    OPCODE_ERROR_INT_ENABLE,
+								    state == AMDGPU_IRQ_STATE_ENABLE ? 1 : 0);
+					WREG32_SOC15_IP(GC, cp_int_cntl_reg, cp_int_cntl);
+				}
+			}
+		}
+		for (i = 0; i < adev->gfx.mec.num_mec; i++) {
+			for (j = 0; j < adev->gfx.mec.num_pipe_per_mec; j++) {
+				/* MECs start at 1 */
+				cp_int_cntl_reg = gfx_v11_0_get_cpc_int_cntl(adev, i + 1, j);
+
+				if (cp_int_cntl_reg) {
+					cp_int_cntl = RREG32_SOC15_IP(GC, cp_int_cntl_reg);
+					cp_int_cntl = REG_SET_FIELD(cp_int_cntl, CP_ME1_PIPE0_INT_CNTL,
+								    OPCODE_ERROR_INT_ENABLE,
+								    state == AMDGPU_IRQ_STATE_ENABLE ? 1 : 0);
+					WREG32_SOC15_IP(GC, cp_int_cntl_reg, cp_int_cntl);
+				}
+			}
+		}
+		break;
+	default:
+		break;
+	}
+	return 0;
+}
+
+static int gfx_v11_0_set_priv_inst_fault_state(struct amdgpu_device *adev,
+					       struct amdgpu_irq_src *source,
+					       unsigned int type,
+					       enum amdgpu_interrupt_state state)
+{
+	u32 cp_int_cntl_reg, cp_int_cntl;
+	int i, j;
+
+	switch (state) {
+	case AMDGPU_IRQ_STATE_DISABLE:
+	case AMDGPU_IRQ_STATE_ENABLE:
+		for (i = 0; i < adev->gfx.me.num_me; i++) {
+			for (j = 0; j < adev->gfx.me.num_pipe_per_me; j++) {
+				cp_int_cntl_reg = gfx_v11_0_get_cpg_int_cntl(adev, i, j);
+
+				if (cp_int_cntl_reg) {
+					cp_int_cntl = RREG32_SOC15_IP(GC, cp_int_cntl_reg);
+					cp_int_cntl = REG_SET_FIELD(cp_int_cntl, CP_INT_CNTL_RING0,
+								    PRIV_INSTR_INT_ENABLE,
+								    state == AMDGPU_IRQ_STATE_ENABLE ? 1 : 0);
+					WREG32_SOC15_IP(GC, cp_int_cntl_reg, cp_int_cntl);
+				}
+			}
+		}
+		break;
+	default:
+		break;
+	}
+
+	return 0;
+}
+
+static void gfx_v11_0_handle_priv_fault(struct amdgpu_device *adev,
+					struct amdgpu_iv_entry *entry)
+{
+	u8 me_id, pipe_id, queue_id;
+	struct amdgpu_ring *ring;
+	int i;
+
+	me_id = (entry->ring_id & 0x0c) >> 2;
+	pipe_id = (entry->ring_id & 0x03) >> 0;
+	queue_id = (entry->ring_id & 0x70) >> 4;
+
+	if (!adev->gfx.disable_kq) {
+		switch (me_id) {
+		case 0:
+			for (i = 0; i < adev->gfx.num_gfx_rings; i++) {
+				ring = &adev->gfx.gfx_ring[i];
+				if (ring->me == me_id && ring->pipe == pipe_id &&
+				    ring->queue == queue_id)
+					drm_sched_fault(&ring->sched);
+			}
+			break;
+		case 1:
+		case 2:
+			for (i = 0; i < adev->gfx.num_compute_rings; i++) {
+				ring = &adev->gfx.compute_ring[i];
+				if (ring->me == me_id && ring->pipe == pipe_id &&
+				    ring->queue == queue_id)
+					drm_sched_fault(&ring->sched);
+			}
+			break;
+		default:
+			BUG();
+			break;
+		}
+	}
+}
+
+static int gfx_v11_0_priv_reg_irq(struct amdgpu_device *adev,
+				  struct amdgpu_irq_src *source,
+				  struct amdgpu_iv_entry *entry)
+{
+	DRM_ERROR("Illegal register access in command stream\n");
+	gfx_v11_0_handle_priv_fault(adev, entry);
+	return 0;
+}
+
+static int gfx_v11_0_bad_op_irq(struct amdgpu_device *adev,
+				struct amdgpu_irq_src *source,
+				struct amdgpu_iv_entry *entry)
+{
+	DRM_ERROR("Illegal opcode in command stream \n");
+	gfx_v11_0_handle_priv_fault(adev, entry);
+	return 0;
+}
+
+static int gfx_v11_0_priv_inst_irq(struct amdgpu_device *adev,
+				   struct amdgpu_irq_src *source,
+				   struct amdgpu_iv_entry *entry)
+{
+	DRM_ERROR("Illegal instruction in command stream\n");
+	gfx_v11_0_handle_priv_fault(adev, entry);
+	return 0;
+}
+
+static int gfx_v11_0_rlc_gc_fed_irq(struct amdgpu_device *adev,
+				  struct amdgpu_irq_src *source,
+				  struct amdgpu_iv_entry *entry)
+{
+	if (adev->gfx.ras && adev->gfx.ras->rlc_gc_fed_irq)
+		return adev->gfx.ras->rlc_gc_fed_irq(adev, source, entry);
+
+	return 0;
+}
+
+#if 0
+static int gfx_v11_0_kiq_set_interrupt_state(struct amdgpu_device *adev,
+					     struct amdgpu_irq_src *src,
+					     unsigned int type,
+					     enum amdgpu_interrupt_state state)
+{
+	uint32_t tmp, target;
+	struct amdgpu_ring *ring = &(adev->gfx.kiq[0].ring);
+
+	target = SOC15_REG_OFFSET(GC, 0, regCP_ME1_PIPE0_INT_CNTL);
+	target += ring->pipe;
+
+	switch (type) {
+	case AMDGPU_CP_KIQ_IRQ_DRIVER0:
+		if (state == AMDGPU_IRQ_STATE_DISABLE) {
+			tmp = RREG32_SOC15(GC, 0, regCPC_INT_CNTL);
+			tmp = REG_SET_FIELD(tmp, CPC_INT_CNTL,
+					    GENERIC2_INT_ENABLE, 0);
+			WREG32_SOC15(GC, 0, regCPC_INT_CNTL, tmp);
+
+			tmp = RREG32_SOC15_IP(GC, target);
+			tmp = REG_SET_FIELD(tmp, CP_ME1_PIPE0_INT_CNTL,
+					    GENERIC2_INT_ENABLE, 0);
+			WREG32_SOC15_IP(GC, target, tmp);
+		} else {
+			tmp = RREG32_SOC15(GC, 0, regCPC_INT_CNTL);
+			tmp = REG_SET_FIELD(tmp, CPC_INT_CNTL,
+					    GENERIC2_INT_ENABLE, 1);
+			WREG32_SOC15(GC, 0, regCPC_INT_CNTL, tmp);
+
+			tmp = RREG32_SOC15_IP(GC, target);
+			tmp = REG_SET_FIELD(tmp, CP_ME1_PIPE0_INT_CNTL,
+					    GENERIC2_INT_ENABLE, 1);
+			WREG32_SOC15_IP(GC, target, tmp);
+		}
+		break;
+	default:
+		BUG(); /* kiq only support GENERIC2_INT now */
+		break;
+	}
+	return 0;
+}
+#endif
+
+static void gfx_v11_0_emit_mem_sync(struct amdgpu_ring *ring)
+{
+	const unsigned int gcr_cntl =
+			PACKET3_ACQUIRE_MEM_GCR_CNTL_GL2_INV(1) |
+			PACKET3_ACQUIRE_MEM_GCR_CNTL_GL2_WB(1) |
+			PACKET3_ACQUIRE_MEM_GCR_CNTL_GLM_INV(1) |
+			PACKET3_ACQUIRE_MEM_GCR_CNTL_GLM_WB(1) |
+			PACKET3_ACQUIRE_MEM_GCR_CNTL_GL1_INV(1) |
+			PACKET3_ACQUIRE_MEM_GCR_CNTL_GLV_INV(1) |
+			PACKET3_ACQUIRE_MEM_GCR_CNTL_GLK_INV(1) |
+			PACKET3_ACQUIRE_MEM_GCR_CNTL_GLI_INV(1);
+
+	/* ACQUIRE_MEM - make one or more surfaces valid for use by the subsequent operations */
+	amdgpu_ring_write(ring, PACKET3(PACKET3_ACQUIRE_MEM, 6));
+	amdgpu_ring_write(ring, 0); /* CP_COHER_CNTL */
+	amdgpu_ring_write(ring, 0xffffffff);  /* CP_COHER_SIZE */
+	amdgpu_ring_write(ring, 0xffffff);  /* CP_COHER_SIZE_HI */
+	amdgpu_ring_write(ring, 0); /* CP_COHER_BASE */
+	amdgpu_ring_write(ring, 0);  /* CP_COHER_BASE_HI */
+	amdgpu_ring_write(ring, 0x0000000A); /* POLL_INTERVAL */
+	amdgpu_ring_write(ring, gcr_cntl); /* GCR_CNTL */
+}
+
+static bool gfx_v11_pipe_reset_support(struct amdgpu_device *adev)
+{
+	/* Disable the pipe reset until the CPFW fully support it.*/
+	dev_warn_once(adev->dev, "The CPFW hasn't support pipe reset yet.\n");
+	return false;
+}
+
+
+static int gfx_v11_reset_gfx_pipe(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	uint32_t reset_pipe = 0, clean_pipe = 0;
+	int r;
+
+	if (!gfx_v11_pipe_reset_support(adev))
+		return -EOPNOTSUPP;
+
+	gfx_v11_0_set_safe_mode(adev, 0);
+	mutex_lock(&adev->srbm_mutex);
+	soc21_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0);
+
+	switch (ring->pipe) {
+	case 0:
+		reset_pipe = REG_SET_FIELD(reset_pipe, CP_ME_CNTL,
+					   PFP_PIPE0_RESET, 1);
+		reset_pipe = REG_SET_FIELD(reset_pipe, CP_ME_CNTL,
+					   ME_PIPE0_RESET, 1);
+		clean_pipe = REG_SET_FIELD(clean_pipe, CP_ME_CNTL,
+					   PFP_PIPE0_RESET, 0);
+		clean_pipe = REG_SET_FIELD(clean_pipe, CP_ME_CNTL,
+					   ME_PIPE0_RESET, 0);
+		break;
+	case 1:
+		reset_pipe = REG_SET_FIELD(reset_pipe, CP_ME_CNTL,
+					   PFP_PIPE1_RESET, 1);
+		reset_pipe = REG_SET_FIELD(reset_pipe, CP_ME_CNTL,
+					   ME_PIPE1_RESET, 1);
+		clean_pipe = REG_SET_FIELD(clean_pipe, CP_ME_CNTL,
+					   PFP_PIPE1_RESET, 0);
+		clean_pipe = REG_SET_FIELD(clean_pipe, CP_ME_CNTL,
+					   ME_PIPE1_RESET, 0);
+		break;
+	default:
+		break;
+	}
+
+	WREG32_SOC15(GC, 0, regCP_ME_CNTL, reset_pipe);
+	WREG32_SOC15(GC, 0, regCP_ME_CNTL, clean_pipe);
+
+	r = (RREG32(SOC15_REG_OFFSET(GC, 0, regCP_GFX_RS64_INSTR_PNTR1)) << 2) -
+						RS64_FW_UC_START_ADDR_LO;
+	soc21_grbm_select(adev, 0, 0, 0, 0);
+	mutex_unlock(&adev->srbm_mutex);
+	gfx_v11_0_unset_safe_mode(adev, 0);
+
+	dev_info(adev->dev, "The ring %s pipe reset to the ME firmware start PC: %s\n", ring->name,
+			r == 0 ? "successfully" : "failed");
+	/* FIXME: Sometimes driver can't cache the ME firmware start PC correctly,
+	 * so the pipe reset status relies on the later gfx ring test result.
+	 */
+	return 0;
+}
+
+static int gfx_v11_0_reset_kgq(struct amdgpu_ring *ring,
+			       unsigned int vmid,
+			       struct amdgpu_fence *timedout_fence)
+{
+	struct amdgpu_device *adev = ring->adev;
+	int r;
+
+	amdgpu_ring_reset_helper_begin(ring, timedout_fence);
+
+	r = amdgpu_mes_reset_legacy_queue(ring->adev, ring, vmid, false);
+	if (r) {
+
+		dev_warn(adev->dev, "reset via MES failed and try pipe reset %d\n", r);
+		r = gfx_v11_reset_gfx_pipe(ring);
+		if (r)
+			return r;
+	}
+
+	r = gfx_v11_0_kgq_init_queue(ring, true);
+	if (r) {
+		dev_err(adev->dev, "failed to init kgq\n");
+		return r;
+	}
+
+	r = amdgpu_mes_map_legacy_queue(adev, ring);
+	if (r) {
+		dev_err(adev->dev, "failed to remap kgq\n");
+		return r;
+	}
+
+	return amdgpu_ring_reset_helper_end(ring, timedout_fence);
+}
+
+static int gfx_v11_0_reset_compute_pipe(struct amdgpu_ring *ring)
+{
+
+	struct amdgpu_device *adev = ring->adev;
+	uint32_t reset_pipe = 0, clean_pipe = 0;
+	int r;
+
+	if (!gfx_v11_pipe_reset_support(adev))
+		return -EOPNOTSUPP;
+
+	gfx_v11_0_set_safe_mode(adev, 0);
+	mutex_lock(&adev->srbm_mutex);
+	soc21_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0);
+
+	reset_pipe = RREG32_SOC15(GC, 0, regCP_MEC_RS64_CNTL);
+	clean_pipe = reset_pipe;
+
+	if (adev->gfx.rs64_enable) {
+
+		switch (ring->pipe) {
+		case 0:
+			reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_RS64_CNTL,
+						   MEC_PIPE0_RESET, 1);
+			clean_pipe = REG_SET_FIELD(clean_pipe, CP_MEC_RS64_CNTL,
+						   MEC_PIPE0_RESET, 0);
+			break;
+		case 1:
+			reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_RS64_CNTL,
+						   MEC_PIPE1_RESET, 1);
+			clean_pipe = REG_SET_FIELD(clean_pipe, CP_MEC_RS64_CNTL,
+						   MEC_PIPE1_RESET, 0);
+			break;
+		case 2:
+			reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_RS64_CNTL,
+						   MEC_PIPE2_RESET, 1);
+			clean_pipe = REG_SET_FIELD(clean_pipe, CP_MEC_RS64_CNTL,
+						   MEC_PIPE2_RESET, 0);
+			break;
+		case 3:
+			reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_RS64_CNTL,
+						   MEC_PIPE3_RESET, 1);
+			clean_pipe = REG_SET_FIELD(clean_pipe, CP_MEC_RS64_CNTL,
+						   MEC_PIPE3_RESET, 0);
+			break;
+		default:
+			break;
+		}
+		WREG32_SOC15(GC, 0, regCP_MEC_RS64_CNTL, reset_pipe);
+		WREG32_SOC15(GC, 0, regCP_MEC_RS64_CNTL, clean_pipe);
+		r = (RREG32_SOC15(GC, 0, regCP_MEC_RS64_INSTR_PNTR) << 2) -
+					RS64_FW_UC_START_ADDR_LO;
+	} else {
+		if (ring->me == 1) {
+			switch (ring->pipe) {
+			case 0:
+				reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_CNTL,
+							   MEC_ME1_PIPE0_RESET, 1);
+				clean_pipe = REG_SET_FIELD(clean_pipe, CP_MEC_CNTL,
+							   MEC_ME1_PIPE0_RESET, 0);
+				break;
+			case 1:
+				reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_CNTL,
+							   MEC_ME1_PIPE1_RESET, 1);
+				clean_pipe = REG_SET_FIELD(clean_pipe, CP_MEC_CNTL,
+							   MEC_ME1_PIPE1_RESET, 0);
+				break;
+			case 2:
+				reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_CNTL,
+							   MEC_ME1_PIPE2_RESET, 1);
+				clean_pipe = REG_SET_FIELD(clean_pipe, CP_MEC_CNTL,
+							   MEC_ME1_PIPE2_RESET, 0);
+				break;
+			case 3:
+				reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_CNTL,
+							   MEC_ME1_PIPE3_RESET, 1);
+				clean_pipe = REG_SET_FIELD(clean_pipe, CP_MEC_CNTL,
+							   MEC_ME1_PIPE3_RESET, 0);
+				break;
+			default:
+				break;
+			}
+			/* mec1 fw pc: CP_MEC1_INSTR_PNTR */
+		} else {
+			switch (ring->pipe) {
+			case 0:
+				reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_CNTL,
+							   MEC_ME2_PIPE0_RESET, 1);
+				clean_pipe = REG_SET_FIELD(clean_pipe, CP_MEC_CNTL,
+							   MEC_ME2_PIPE0_RESET, 0);
+				break;
+			case 1:
+				reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_CNTL,
+							   MEC_ME2_PIPE1_RESET, 1);
+				clean_pipe = REG_SET_FIELD(clean_pipe, CP_MEC_CNTL,
+							   MEC_ME2_PIPE1_RESET, 0);
+				break;
+			case 2:
+				reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_CNTL,
+							   MEC_ME2_PIPE2_RESET, 1);
+				clean_pipe = REG_SET_FIELD(clean_pipe, CP_MEC_CNTL,
+							   MEC_ME2_PIPE2_RESET, 0);
+				break;
+			case 3:
+				reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_CNTL,
+							   MEC_ME2_PIPE3_RESET, 1);
+				clean_pipe = REG_SET_FIELD(clean_pipe, CP_MEC_CNTL,
+							   MEC_ME2_PIPE3_RESET, 0);
+				break;
+			default:
+				break;
+			}
+			/* mec2 fw pc: CP:CP_MEC2_INSTR_PNTR */
+		}
+		WREG32_SOC15(GC, 0, regCP_MEC_CNTL, reset_pipe);
+		WREG32_SOC15(GC, 0, regCP_MEC_CNTL, clean_pipe);
+		r = RREG32(SOC15_REG_OFFSET(GC, 0, regCP_MEC1_INSTR_PNTR));
+	}
+
+	soc21_grbm_select(adev, 0, 0, 0, 0);
+	mutex_unlock(&adev->srbm_mutex);
+	gfx_v11_0_unset_safe_mode(adev, 0);
+
+	dev_info(adev->dev, "The ring %s pipe resets to MEC FW start PC: %s\n", ring->name,
+			r == 0 ? "successfully" : "failed");
+	/*FIXME:Sometimes driver can't cache the MEC firmware start PC correctly, so the pipe
+	 * reset status relies on the compute ring test result.
+	 */
+	return 0;
+}
+
+static int gfx_v11_0_reset_kcq(struct amdgpu_ring *ring,
+			       unsigned int vmid,
+			       struct amdgpu_fence *timedout_fence)
+{
+	struct amdgpu_device *adev = ring->adev;
+	int r = 0;
+
+	amdgpu_ring_reset_helper_begin(ring, timedout_fence);
+
+	r = amdgpu_mes_reset_legacy_queue(ring->adev, ring, vmid, true);
+	if (r) {
+		dev_warn(adev->dev, "fail(%d) to reset kcq and try pipe reset\n", r);
+		r = gfx_v11_0_reset_compute_pipe(ring);
+		if (r)
+			return r;
+	}
+
+	r = gfx_v11_0_kcq_init_queue(ring, true);
+	if (r) {
+		dev_err(adev->dev, "fail to init kcq\n");
+		return r;
+	}
+	r = amdgpu_mes_map_legacy_queue(adev, ring);
+	if (r) {
+		dev_err(adev->dev, "failed to remap kcq\n");
+		return r;
+	}
+
+	return amdgpu_ring_reset_helper_end(ring, timedout_fence);
+}
+
+static void gfx_v11_ip_print(struct amdgpu_ip_block *ip_block, struct drm_printer *p)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	uint32_t i, j, k, reg, index = 0;
+	uint32_t reg_count = ARRAY_SIZE(gc_reg_list_11_0);
+
+	if (!adev->gfx.ip_dump_core)
+		return;
+
+	for (i = 0; i < reg_count; i++)
+		drm_printf(p, "%-50s \t 0x%08x\n",
+			   gc_reg_list_11_0[i].reg_name,
+			   adev->gfx.ip_dump_core[i]);
+
+	/* print compute queue registers for all instances */
+	if (!adev->gfx.ip_dump_compute_queues)
+		return;
+
+	reg_count = ARRAY_SIZE(gc_cp_reg_list_11);
+	drm_printf(p, "\nnum_mec: %d num_pipe: %d num_queue: %d\n",
+		   adev->gfx.mec.num_mec,
+		   adev->gfx.mec.num_pipe_per_mec,
+		   adev->gfx.mec.num_queue_per_pipe);
+
+	for (i = 0; i < adev->gfx.mec.num_mec; i++) {
+		for (j = 0; j < adev->gfx.mec.num_pipe_per_mec; j++) {
+			for (k = 0; k < adev->gfx.mec.num_queue_per_pipe; k++) {
+				drm_printf(p, "\nmec %d, pipe %d, queue %d\n", i, j, k);
+				for (reg = 0; reg < reg_count; reg++) {
+					if (i && gc_cp_reg_list_11[reg].reg_offset == regCP_MEC_ME1_HEADER_DUMP)
+						drm_printf(p, "%-50s \t 0x%08x\n",
+							   "regCP_MEC_ME2_HEADER_DUMP",
+							   adev->gfx.ip_dump_compute_queues[index + reg]);
+					else
+						drm_printf(p, "%-50s \t 0x%08x\n",
+							   gc_cp_reg_list_11[reg].reg_name,
+							   adev->gfx.ip_dump_compute_queues[index + reg]);
+				}
+				index += reg_count;
+			}
+		}
+	}
+
+	/* print gfx queue registers for all instances */
+	if (!adev->gfx.ip_dump_gfx_queues)
+		return;
+
+	index = 0;
+	reg_count = ARRAY_SIZE(gc_gfx_queue_reg_list_11);
+	drm_printf(p, "\nnum_me: %d num_pipe: %d num_queue: %d\n",
+		   adev->gfx.me.num_me,
+		   adev->gfx.me.num_pipe_per_me,
+		   adev->gfx.me.num_queue_per_pipe);
+
+	for (i = 0; i < adev->gfx.me.num_me; i++) {
+		for (j = 0; j < adev->gfx.me.num_pipe_per_me; j++) {
+			for (k = 0; k < adev->gfx.me.num_queue_per_pipe; k++) {
+				drm_printf(p, "\nme %d, pipe %d, queue %d\n", i, j, k);
+				for (reg = 0; reg < reg_count; reg++) {
+					drm_printf(p, "%-50s \t 0x%08x\n",
+						   gc_gfx_queue_reg_list_11[reg].reg_name,
+						   adev->gfx.ip_dump_gfx_queues[index + reg]);
+				}
+				index += reg_count;
+			}
+		}
+	}
+}
+
+static void gfx_v11_ip_dump(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	uint32_t i, j, k, reg, index = 0;
+	uint32_t reg_count = ARRAY_SIZE(gc_reg_list_11_0);
+
+	if (!adev->gfx.ip_dump_core)
+		return;
+
+	amdgpu_gfx_off_ctrl(adev, false);
+	for (i = 0; i < reg_count; i++)
+		adev->gfx.ip_dump_core[i] = RREG32(SOC15_REG_ENTRY_OFFSET(gc_reg_list_11_0[i]));
+	amdgpu_gfx_off_ctrl(adev, true);
+
+	/* dump compute queue registers for all instances */
+	if (!adev->gfx.ip_dump_compute_queues)
+		return;
+
+	reg_count = ARRAY_SIZE(gc_cp_reg_list_11);
+	amdgpu_gfx_off_ctrl(adev, false);
+	mutex_lock(&adev->srbm_mutex);
+	for (i = 0; i < adev->gfx.mec.num_mec; i++) {
+		for (j = 0; j < adev->gfx.mec.num_pipe_per_mec; j++) {
+			for (k = 0; k < adev->gfx.mec.num_queue_per_pipe; k++) {
+				/* ME0 is for GFX so start from 1 for CP */
+				soc21_grbm_select(adev, adev->gfx.me.num_me + i, j, k, 0);
+				for (reg = 0; reg < reg_count; reg++) {
+					if (i &&
+					    gc_cp_reg_list_11[reg].reg_offset ==
+						    regCP_MEC_ME1_HEADER_DUMP)
+						adev->gfx.ip_dump_compute_queues[index + reg] =
+							RREG32(SOC15_REG_OFFSET(GC, 0,
+							       regCP_MEC_ME2_HEADER_DUMP));
+					else
+						adev->gfx.ip_dump_compute_queues[index + reg] =
+							RREG32(SOC15_REG_ENTRY_OFFSET(
+								       gc_cp_reg_list_11[reg]));
+				}
+				index += reg_count;
+			}
+		}
+	}
+	soc21_grbm_select(adev, 0, 0, 0, 0);
+	mutex_unlock(&adev->srbm_mutex);
+	amdgpu_gfx_off_ctrl(adev, true);
+
+	/* dump gfx queue registers for all instances */
+	if (!adev->gfx.ip_dump_gfx_queues)
+		return;
+
+	index = 0;
+	reg_count = ARRAY_SIZE(gc_gfx_queue_reg_list_11);
+	amdgpu_gfx_off_ctrl(adev, false);
+	mutex_lock(&adev->srbm_mutex);
+	for (i = 0; i < adev->gfx.me.num_me; i++) {
+		for (j = 0; j < adev->gfx.me.num_pipe_per_me; j++) {
+			for (k = 0; k < adev->gfx.me.num_queue_per_pipe; k++) {
+				soc21_grbm_select(adev, i, j, k, 0);
+
+				for (reg = 0; reg < reg_count; reg++) {
+					adev->gfx.ip_dump_gfx_queues[index + reg] =
+						RREG32(SOC15_REG_ENTRY_OFFSET(
+							gc_gfx_queue_reg_list_11[reg]));
+				}
+				index += reg_count;
+			}
+		}
+	}
+	soc21_grbm_select(adev, 0, 0, 0, 0);
+	mutex_unlock(&adev->srbm_mutex);
+	amdgpu_gfx_off_ctrl(adev, true);
+}
+
+static void gfx_v11_0_ring_emit_cleaner_shader(struct amdgpu_ring *ring)
+{
+	/* Emit the cleaner shader */
+	amdgpu_ring_write(ring, PACKET3(PACKET3_RUN_CLEANER_SHADER, 0));
+	amdgpu_ring_write(ring, 0);  /* RESERVED field, programmed to zero */
+}
+
+static void gfx_v11_0_ring_begin_use(struct amdgpu_ring *ring)
+{
+	amdgpu_gfx_profile_ring_begin_use(ring);
+
+	amdgpu_gfx_enforce_isolation_ring_begin_use(ring);
+}
+
+static void gfx_v11_0_ring_end_use(struct amdgpu_ring *ring)
+{
+	amdgpu_gfx_profile_ring_end_use(ring);
+
+	amdgpu_gfx_enforce_isolation_ring_end_use(ring);
+}
+
+static const struct amd_ip_funcs gfx_v11_0_ip_funcs = {
+	.name = "gfx_v11_0",
+	.early_init = gfx_v11_0_early_init,
+	.late_init = gfx_v11_0_late_init,
+	.sw_init = gfx_v11_0_sw_init,
+	.sw_fini = gfx_v11_0_sw_fini,
+	.hw_init = gfx_v11_0_hw_init,
+	.hw_fini = gfx_v11_0_hw_fini,
+	.suspend = gfx_v11_0_suspend,
+	.resume = gfx_v11_0_resume,
+	.is_idle = gfx_v11_0_is_idle,
+	.wait_for_idle = gfx_v11_0_wait_for_idle,
+	.soft_reset = gfx_v11_0_soft_reset,
+	.check_soft_reset = gfx_v11_0_check_soft_reset,
+	.post_soft_reset = gfx_v11_0_post_soft_reset,
+	.set_clockgating_state = gfx_v11_0_set_clockgating_state,
+	.set_powergating_state = gfx_v11_0_set_powergating_state,
+	.get_clockgating_state = gfx_v11_0_get_clockgating_state,
+	.dump_ip_state = gfx_v11_ip_dump,
+	.print_ip_state = gfx_v11_ip_print,
+};
+
+static const struct amdgpu_ring_funcs gfx_v11_0_ring_funcs_gfx = {
+	.type = AMDGPU_RING_TYPE_GFX,
+	.align_mask = 0xff,
+	.nop = PACKET3(PACKET3_NOP, 0x3FFF),
+	.support_64bit_ptrs = true,
+	.secure_submission_supported = true,
+	.get_rptr = gfx_v11_0_ring_get_rptr_gfx,
+	.get_wptr = gfx_v11_0_ring_get_wptr_gfx,
+	.set_wptr = gfx_v11_0_ring_set_wptr_gfx,
+	.emit_frame_size = /* totally 247 maximum if 16 IBs */
+		5 + /* update_spm_vmid */
+		5 + /* COND_EXEC */
+		22 + /* SET_Q_PREEMPTION_MODE */
+		7 + /* PIPELINE_SYNC */
+		SOC15_FLUSH_GPU_TLB_NUM_WREG * 5 +
+		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 7 +
+		4 + /* VM_FLUSH */
+		8 + /* FENCE for VM_FLUSH */
+		20 + /* GDS switch */
+		5 + /* COND_EXEC */
+		7 + /* HDP_flush */
+		4 + /* VGT_flush */
+		31 + /*	DE_META */
+		3 + /* CNTX_CTRL */
+		5 + /* HDP_INVL */
+		22 + /* SET_Q_PREEMPTION_MODE */
+		8 + 8 + /* FENCE x2 */
+		8 + /* gfx_v11_0_emit_mem_sync */
+		2, /* gfx_v11_0_ring_emit_cleaner_shader */
+	.emit_ib_size =	4, /* gfx_v11_0_ring_emit_ib_gfx */
+	.emit_ib = gfx_v11_0_ring_emit_ib_gfx,
+	.emit_fence = gfx_v11_0_ring_emit_fence,
+	.emit_pipeline_sync = gfx_v11_0_ring_emit_pipeline_sync,
+	.emit_vm_flush = gfx_v11_0_ring_emit_vm_flush,
+	.emit_gds_switch = gfx_v11_0_ring_emit_gds_switch,
+	.emit_hdp_flush = gfx_v11_0_ring_emit_hdp_flush,
+	.test_ring = gfx_v11_0_ring_test_ring,
+	.test_ib = gfx_v11_0_ring_test_ib,
+	.insert_nop = gfx_v11_ring_insert_nop,
+	.pad_ib = amdgpu_ring_generic_pad_ib,
+	.emit_cntxcntl = gfx_v11_0_ring_emit_cntxcntl,
+	.emit_gfx_shadow = gfx_v11_0_ring_emit_gfx_shadow,
+	.init_cond_exec = gfx_v11_0_ring_emit_init_cond_exec,
+	.preempt_ib = gfx_v11_0_ring_preempt_ib,
+	.emit_frame_cntl = gfx_v11_0_ring_emit_frame_cntl,
+	.emit_wreg = gfx_v11_0_ring_emit_wreg,
+	.emit_reg_wait = gfx_v11_0_ring_emit_reg_wait,
+	.emit_reg_write_reg_wait = gfx_v11_0_ring_emit_reg_write_reg_wait,
+	.emit_mem_sync = gfx_v11_0_emit_mem_sync,
+	.reset = gfx_v11_0_reset_kgq,
+	.emit_cleaner_shader = gfx_v11_0_ring_emit_cleaner_shader,
+	.begin_use = gfx_v11_0_ring_begin_use,
+	.end_use = gfx_v11_0_ring_end_use,
+};
+
+static const struct amdgpu_ring_funcs gfx_v11_0_ring_funcs_compute = {
+	.type = AMDGPU_RING_TYPE_COMPUTE,
+	.align_mask = 0xff,
+	.nop = PACKET3(PACKET3_NOP, 0x3FFF),
+	.support_64bit_ptrs = true,
+	.get_rptr = gfx_v11_0_ring_get_rptr_compute,
+	.get_wptr = gfx_v11_0_ring_get_wptr_compute,
+	.set_wptr = gfx_v11_0_ring_set_wptr_compute,
+	.emit_frame_size =
+		5 + /* update_spm_vmid */
+		20 + /* gfx_v11_0_ring_emit_gds_switch */
+		7 + /* gfx_v11_0_ring_emit_hdp_flush */
+		5 + /* hdp invalidate */
+		7 + /* gfx_v11_0_ring_emit_pipeline_sync */
+		SOC15_FLUSH_GPU_TLB_NUM_WREG * 5 +
+		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 7 +
+		2 + /* gfx_v11_0_ring_emit_vm_flush */
+		8 + 8 + 8 + /* gfx_v11_0_ring_emit_fence x3 for user fence, vm fence */
+		8 + /* gfx_v11_0_emit_mem_sync */
+		2, /* gfx_v11_0_ring_emit_cleaner_shader */
+	.emit_ib_size =	7, /* gfx_v11_0_ring_emit_ib_compute */
+	.emit_ib = gfx_v11_0_ring_emit_ib_compute,
+	.emit_fence = gfx_v11_0_ring_emit_fence,
+	.emit_pipeline_sync = gfx_v11_0_ring_emit_pipeline_sync,
+	.emit_vm_flush = gfx_v11_0_ring_emit_vm_flush,
+	.emit_gds_switch = gfx_v11_0_ring_emit_gds_switch,
+	.emit_hdp_flush = gfx_v11_0_ring_emit_hdp_flush,
+	.test_ring = gfx_v11_0_ring_test_ring,
+	.test_ib = gfx_v11_0_ring_test_ib,
+	.insert_nop = gfx_v11_ring_insert_nop,
+	.pad_ib = amdgpu_ring_generic_pad_ib,
+	.emit_wreg = gfx_v11_0_ring_emit_wreg,
+	.emit_reg_wait = gfx_v11_0_ring_emit_reg_wait,
+	.emit_reg_write_reg_wait = gfx_v11_0_ring_emit_reg_write_reg_wait,
+	.emit_mem_sync = gfx_v11_0_emit_mem_sync,
+	.reset = gfx_v11_0_reset_kcq,
+	.emit_cleaner_shader = gfx_v11_0_ring_emit_cleaner_shader,
+	.begin_use = gfx_v11_0_ring_begin_use,
+	.end_use = gfx_v11_0_ring_end_use,
+};
+
+static const struct amdgpu_ring_funcs gfx_v11_0_ring_funcs_kiq = {
+	.type = AMDGPU_RING_TYPE_KIQ,
+	.align_mask = 0xff,
+	.nop = PACKET3(PACKET3_NOP, 0x3FFF),
+	.support_64bit_ptrs = true,
+	.get_rptr = gfx_v11_0_ring_get_rptr_compute,
+	.get_wptr = gfx_v11_0_ring_get_wptr_compute,
+	.set_wptr = gfx_v11_0_ring_set_wptr_compute,
+	.emit_frame_size =
+		20 + /* gfx_v11_0_ring_emit_gds_switch */
+		7 + /* gfx_v11_0_ring_emit_hdp_flush */
+		5 + /*hdp invalidate */
+		7 + /* gfx_v11_0_ring_emit_pipeline_sync */
+		SOC15_FLUSH_GPU_TLB_NUM_WREG * 5 +
+		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 7 +
+		8 + 8 + 8, /* gfx_v11_0_ring_emit_fence_kiq x3 for user fence, vm fence */
+	.emit_ib_size =	7, /* gfx_v11_0_ring_emit_ib_compute */
+	.emit_ib = gfx_v11_0_ring_emit_ib_compute,
+	.emit_fence = gfx_v11_0_ring_emit_fence_kiq,
+	.test_ring = gfx_v11_0_ring_test_ring,
+	.test_ib = gfx_v11_0_ring_test_ib,
+	.insert_nop = amdgpu_ring_insert_nop,
+	.pad_ib = amdgpu_ring_generic_pad_ib,
+	.emit_rreg = gfx_v11_0_ring_emit_rreg,
+	.emit_wreg = gfx_v11_0_ring_emit_wreg,
+	.emit_reg_wait = gfx_v11_0_ring_emit_reg_wait,
+	.emit_reg_write_reg_wait = gfx_v11_0_ring_emit_reg_write_reg_wait,
+	.emit_hdp_flush = gfx_v11_0_ring_emit_hdp_flush,
+};
+
+static void gfx_v11_0_set_ring_funcs(struct amdgpu_device *adev)
+{
+	int i;
+
+	adev->gfx.kiq[0].ring.funcs = &gfx_v11_0_ring_funcs_kiq;
+
+	for (i = 0; i < adev->gfx.num_gfx_rings; i++)
+		adev->gfx.gfx_ring[i].funcs = &gfx_v11_0_ring_funcs_gfx;
+
+	for (i = 0; i < adev->gfx.num_compute_rings; i++)
+		adev->gfx.compute_ring[i].funcs = &gfx_v11_0_ring_funcs_compute;
+}
+
+static const struct amdgpu_irq_src_funcs gfx_v11_0_eop_irq_funcs = {
+	.set = gfx_v11_0_set_eop_interrupt_state,
+	.process = gfx_v11_0_eop_irq,
+};
+
+static const struct amdgpu_irq_src_funcs gfx_v11_0_priv_reg_irq_funcs = {
+	.set = gfx_v11_0_set_priv_reg_fault_state,
+	.process = gfx_v11_0_priv_reg_irq,
+};
+
+static const struct amdgpu_irq_src_funcs gfx_v11_0_bad_op_irq_funcs = {
+	.set = gfx_v11_0_set_bad_op_fault_state,
+	.process = gfx_v11_0_bad_op_irq,
+};
+
+static const struct amdgpu_irq_src_funcs gfx_v11_0_priv_inst_irq_funcs = {
+	.set = gfx_v11_0_set_priv_inst_fault_state,
+	.process = gfx_v11_0_priv_inst_irq,
+};
+
+static const struct amdgpu_irq_src_funcs gfx_v11_0_rlc_gc_fed_irq_funcs = {
+	.process = gfx_v11_0_rlc_gc_fed_irq,
+};
+
+static void gfx_v11_0_set_irq_funcs(struct amdgpu_device *adev)
+{
+	adev->gfx.eop_irq.num_types = AMDGPU_CP_IRQ_LAST;
+	adev->gfx.eop_irq.funcs = &gfx_v11_0_eop_irq_funcs;
+
+	adev->gfx.priv_reg_irq.num_types = 1;
+	adev->gfx.priv_reg_irq.funcs = &gfx_v11_0_priv_reg_irq_funcs;
+
+	adev->gfx.bad_op_irq.num_types = 1;
+	adev->gfx.bad_op_irq.funcs = &gfx_v11_0_bad_op_irq_funcs;
+
+	adev->gfx.priv_inst_irq.num_types = 1;
+	adev->gfx.priv_inst_irq.funcs = &gfx_v11_0_priv_inst_irq_funcs;
+
+	adev->gfx.rlc_gc_fed_irq.num_types = 1; /* 0x80 FED error */
+	adev->gfx.rlc_gc_fed_irq.funcs = &gfx_v11_0_rlc_gc_fed_irq_funcs;
+
+}
+
+static void gfx_v11_0_set_imu_funcs(struct amdgpu_device *adev)
+{
+	if (adev->flags & AMD_IS_APU)
+		adev->gfx.imu.mode = MISSION_MODE;
+	else
+		adev->gfx.imu.mode = DEBUG_MODE;
+
+	adev->gfx.imu.funcs = &gfx_v11_0_imu_funcs;
+}
+
+static void gfx_v11_0_set_rlc_funcs(struct amdgpu_device *adev)
+{
+	adev->gfx.rlc.funcs = &gfx_v11_0_rlc_funcs;
+}
+
+static void gfx_v11_0_set_gds_init(struct amdgpu_device *adev)
+{
+	unsigned total_cu = adev->gfx.config.max_cu_per_sh *
+			    adev->gfx.config.max_sh_per_se *
+			    adev->gfx.config.max_shader_engines;
+
+	adev->gds.gds_size = 0x1000;
+	adev->gds.gds_compute_max_wave_id = total_cu * 32 - 1;
+	adev->gds.gws_size = 64;
+	adev->gds.oa_size = 16;
+}
+
+static void gfx_v11_0_set_mqd_funcs(struct amdgpu_device *adev)
+{
+	/* set gfx eng mqd */
+	adev->mqds[AMDGPU_HW_IP_GFX].mqd_size =
+		sizeof(struct v11_gfx_mqd);
+	adev->mqds[AMDGPU_HW_IP_GFX].init_mqd =
+		gfx_v11_0_gfx_mqd_init;
+	/* set compute eng mqd */
+	adev->mqds[AMDGPU_HW_IP_COMPUTE].mqd_size =
+		sizeof(struct v11_compute_mqd);
+	adev->mqds[AMDGPU_HW_IP_COMPUTE].init_mqd =
+		gfx_v11_0_compute_mqd_init;
+}
+
+static void gfx_v11_0_set_user_wgp_inactive_bitmap_per_sh(struct amdgpu_device *adev,
+							  u32 bitmap)
+{
+	u32 data;
+
+	if (!bitmap)
+		return;
+
+	data = bitmap << GC_USER_SHADER_ARRAY_CONFIG__INACTIVE_WGPS__SHIFT;
+	data &= GC_USER_SHADER_ARRAY_CONFIG__INACTIVE_WGPS_MASK;
+
+	WREG32_SOC15(GC, 0, regGC_USER_SHADER_ARRAY_CONFIG, data);
+}
+
+static u32 gfx_v11_0_get_wgp_active_bitmap_per_sh(struct amdgpu_device *adev)
+{
+	u32 data, wgp_bitmask;
+	data = RREG32_SOC15(GC, 0, regCC_GC_SHADER_ARRAY_CONFIG);
+	data |= RREG32_SOC15(GC, 0, regGC_USER_SHADER_ARRAY_CONFIG);
+
+	data &= CC_GC_SHADER_ARRAY_CONFIG__INACTIVE_WGPS_MASK;
+	data >>= CC_GC_SHADER_ARRAY_CONFIG__INACTIVE_WGPS__SHIFT;
+
+	wgp_bitmask =
+		amdgpu_gfx_create_bitmask(adev->gfx.config.max_cu_per_sh >> 1);
+
+	return (~data) & wgp_bitmask;
+}
+
+static u32 gfx_v11_0_get_cu_active_bitmap_per_sh(struct amdgpu_device *adev)
+{
+	u32 wgp_idx, wgp_active_bitmap;
+	u32 cu_bitmap_per_wgp, cu_active_bitmap;
+
+	wgp_active_bitmap = gfx_v11_0_get_wgp_active_bitmap_per_sh(adev);
+	cu_active_bitmap = 0;
+
+	for (wgp_idx = 0; wgp_idx < 16; wgp_idx++) {
+		/* if there is one WGP enabled, it means 2 CUs will be enabled */
+		cu_bitmap_per_wgp = 3 << (2 * wgp_idx);
+		if (wgp_active_bitmap & (1 << wgp_idx))
+			cu_active_bitmap |= cu_bitmap_per_wgp;
+	}
+
+	return cu_active_bitmap;
+}
+
+static int gfx_v11_0_get_cu_info(struct amdgpu_device *adev,
+				 struct amdgpu_cu_info *cu_info)
+{
+	int i, j, k, counter, active_cu_number = 0;
+	u32 mask, bitmap;
+	unsigned disable_masks[8 * 2];
+
+	if (!adev || !cu_info)
+		return -EINVAL;
+
+	amdgpu_gfx_parse_disable_cu(disable_masks, 8, 2);
+
+	mutex_lock(&adev->grbm_idx_mutex);
+	for (i = 0; i < adev->gfx.config.max_shader_engines; i++) {
+		for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) {
+			bitmap = i * adev->gfx.config.max_sh_per_se + j;
+			if (!((gfx_v11_0_get_sa_active_bitmap(adev) >> bitmap) & 1))
+				continue;
+			mask = 1;
+			counter = 0;
+			gfx_v11_0_select_se_sh(adev, i, j, 0xffffffff, 0);
+			if (i < 8 && j < 2)
+				gfx_v11_0_set_user_wgp_inactive_bitmap_per_sh(
+					adev, disable_masks[i * 2 + j]);
+			bitmap = gfx_v11_0_get_cu_active_bitmap_per_sh(adev);
+
+			/**
+			 * GFX11 could support more than 4 SEs, while the bitmap
+			 * in cu_info struct is 4x4 and ioctl interface struct
+			 * drm_amdgpu_info_device should keep stable.
+			 * So we use last two columns of bitmap to store cu mask for
+			 * SEs 4 to 7, the layout of the bitmap is as below:
+			 *    SE0: {SH0,SH1} --> {bitmap[0][0], bitmap[0][1]}
+			 *    SE1: {SH0,SH1} --> {bitmap[1][0], bitmap[1][1]}
+			 *    SE2: {SH0,SH1} --> {bitmap[2][0], bitmap[2][1]}
+			 *    SE3: {SH0,SH1} --> {bitmap[3][0], bitmap[3][1]}
+			 *    SE4: {SH0,SH1} --> {bitmap[0][2], bitmap[0][3]}
+			 *    SE5: {SH0,SH1} --> {bitmap[1][2], bitmap[1][3]}
+			 *    SE6: {SH0,SH1} --> {bitmap[2][2], bitmap[2][3]}
+			 *    SE7: {SH0,SH1} --> {bitmap[3][2], bitmap[3][3]}
+			 */
+			cu_info->bitmap[0][i % 4][j + (i / 4) * 2] = bitmap;
+
+			for (k = 0; k < adev->gfx.config.max_cu_per_sh; k++) {
+				if (bitmap & mask)
+					counter++;
+
+				mask <<= 1;
+			}
+			active_cu_number += counter;
+		}
+	}
+	gfx_v11_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
+	mutex_unlock(&adev->grbm_idx_mutex);
+
+	cu_info->number = active_cu_number;
+	cu_info->simd_per_cu = NUM_SIMD_PER_CU;
+
+	return 0;
+}
+
+const struct amdgpu_ip_block_version gfx_v11_0_ip_block =
+{
+	.type = AMD_IP_BLOCK_TYPE_GFX,
+	.major = 11,
+	.minor = 0,
+	.rev = 0,
+	.funcs = &gfx_v11_0_ip_funcs,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.h b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.h
new file mode 100644
index 000000000000..157a5c812259
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.h
@@ -0,0 +1,32 @@
+/*
+ * Copyright 2021 dvanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __GFX_V11_0_H__
+#define __GFX_V11_0_H__
+
+extern const struct amdgpu_ip_block_version gfx_v11_0_ip_block;
+
+int gfx_v11_0_request_gfx_index_mutex(struct amdgpu_device *adev,
+				      bool req);
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0_3.c b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0_3.c
new file mode 100644
index 000000000000..999bb3cc88b7
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0_3.c
@@ -0,0 +1,110 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "amdgpu.h"
+#include "soc21.h"
+#include "gc/gc_11_0_3_offset.h"
+#include "gc/gc_11_0_3_sh_mask.h"
+#include "ivsrcid/gfx/irqsrcs_gfx_11_0_0.h"
+#include "soc15.h"
+#include "soc15d.h"
+#include "gfx_v11_0.h"
+
+
+static int gfx_v11_0_3_rlc_gc_fed_irq(struct amdgpu_device *adev,
+				  struct amdgpu_irq_src *source,
+				  struct amdgpu_iv_entry *entry)
+{
+	uint32_t rlc_status0 = 0, rlc_status1 = 0;
+	struct ras_common_if *ras_if = NULL;
+	struct ras_dispatch_if ih_data = {
+		.entry = entry,
+	};
+
+	rlc_status0 = RREG32(SOC15_REG_OFFSET(GC, 0, regRLC_RLCS_FED_STATUS_0));
+	rlc_status1 = RREG32(SOC15_REG_OFFSET(GC, 0, regRLC_RLCS_FED_STATUS_1));
+
+	if (!rlc_status0 && !rlc_status1) {
+		dev_warn(adev->dev, "RLC_GC_FED irq is generated, but rlc_status0 and rlc_status1 are empty!\n");
+		return 0;
+	}
+
+	/* Use RLC_RLCS_FED_STATUS_0/1 to distinguish FED error block. */
+	if (REG_GET_FIELD(rlc_status0, RLC_RLCS_FED_STATUS_0, SDMA0_FED_ERR) ||
+	    REG_GET_FIELD(rlc_status0, RLC_RLCS_FED_STATUS_0, SDMA1_FED_ERR))
+		ras_if = adev->sdma.ras_if;
+	else
+		ras_if = adev->gfx.ras_if;
+
+	if (!ras_if) {
+		dev_err(adev->dev, "Gfx or sdma ras block not initialized, rlc_status0:0x%x.\n",
+				rlc_status0);
+		return -EINVAL;
+	}
+
+	dev_warn(adev->dev, "RLC %s FED IRQ\n", ras_if->name);
+
+	if (!amdgpu_sriov_vf(adev)) {
+		ih_data.head = *ras_if;
+		amdgpu_ras_interrupt_dispatch(adev, &ih_data);
+	} else {
+		if (adev->virt.ops && adev->virt.ops->ras_poison_handler)
+			adev->virt.ops->ras_poison_handler(adev, ras_if->block);
+		else
+			dev_warn(adev->dev,
+				"No ras_poison_handler interface in SRIOV for %s!\n", ras_if->name);
+	}
+
+	return 0;
+}
+
+static int gfx_v11_0_3_poison_consumption_handler(struct amdgpu_device *adev,
+					struct amdgpu_iv_entry *entry)
+{
+	/* Workaround: when vmid and pasid are both zero, trigger gpu reset in KGD. */
+	if (entry && (entry->client_id == SOC21_IH_CLIENTID_GFX) &&
+	    (entry->src_id == GFX_11_0_0__SRCID__RLC_GC_FED_INTERRUPT) &&
+	     !entry->vmid && !entry->pasid) {
+		struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+		uint32_t rlc_status0 = 0;
+
+		rlc_status0 = RREG32_SOC15(GC, 0, regRLC_RLCS_FED_STATUS_0);
+
+		if (REG_GET_FIELD(rlc_status0, RLC_RLCS_FED_STATUS_0, SDMA0_FED_ERR) ||
+		    REG_GET_FIELD(rlc_status0, RLC_RLCS_FED_STATUS_0, SDMA1_FED_ERR)) {
+			struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
+
+			ras->gpu_reset_flags |= AMDGPU_RAS_GPU_RESET_MODE2_RESET;
+		}
+
+		if (con && !amdgpu_ras_is_rma(adev))
+			amdgpu_ras_reset_gpu(adev);
+	}
+
+	return 0;
+}
+
+struct amdgpu_gfx_ras gfx_v11_0_3_ras = {
+	.rlc_gc_fed_irq = gfx_v11_0_3_rlc_gc_fed_irq,
+	.poison_consumption_handler = gfx_v11_0_3_poison_consumption_handler,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0_3.h b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0_3.h
new file mode 100644
index 000000000000..672c7920b3d0
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0_3.h
@@ -0,0 +1,29 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __GFX_V11_0_3_H__
+#define __GFX_V11_0_3_H__
+
+extern struct amdgpu_gfx_ras gfx_v11_0_3_ras;
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0_3_cleaner_shader.asm b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0_3_cleaner_shader.asm
new file mode 100644
index 000000000000..9b90b66368c7
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0_3_cleaner_shader.asm
@@ -0,0 +1,118 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2024 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+// This shader is to clean LDS, SGPRs and VGPRs. It is  first 64 Dwords or 256 bytes of 192 Dwords cleaner shader.
+//To turn this shader program on for complitaion change this to main and lower shader main to main_1
+
+// Navi3 : Clear SGPRs, VGPRs and LDS
+//   Launch 32 waves per CU (16 per SIMD) as a workgroup (threadgroup) to fill every wave slot
+//   Waves are "wave32" and have 64 VGPRs each, which uses all 1024 VGPRs per SIMD
+//   Waves are launched in "CU" mode, and the workgroup shares 64KB of LDS (half of the WGP's LDS)
+//      It takes 2 workgroups to use all of LDS: one on each CU of the WGP
+//   Each wave clears SGPRs 0 - 107
+//   Each wave clears VGPRs 0 - 63
+//   The first wave of the workgroup clears its 64KB of LDS
+//   The shader starts with "S_BARRIER" to ensure SPI has launched all waves of the workgroup
+//       before any wave in the workgroup could end.  Without this, it is possible not all SGPRs get cleared.
+
+shader main
+  asic(GFX11)
+  type(CS)
+  wave_size(32)
+// Note: original source code from SQ team
+
+// Takes about 2500 clocks to run.
+//   (theorhetical fastest = 1024clks vgpr + 640lds = 1660 clks)
+//
+  S_BARRIER
+
+  //
+  // CLEAR VGPRs
+  //
+  s_mov_b32     m0, 0x00000058  // Loop 96/8=12 times  (loop unrolled for performance)
+
+label_0005:
+  v_movreld_b32     v0, 0
+  v_movreld_b32     v1, 0
+  v_movreld_b32     v2, 0
+  v_movreld_b32     v3, 0
+  v_movreld_b32     v4, 0
+  v_movreld_b32     v5, 0
+  v_movreld_b32     v6, 0
+  v_movreld_b32     v7, 0
+  s_sub_u32     m0, m0, 8
+  s_cbranch_scc0  label_0005
+  //
+  //
+
+  s_mov_b32     s2, 0x80000000                      // Bit31 is first_wave
+  s_and_b32     s2, s2, s0                          // sgpr0 has tg_size (first_wave) term as in ucode only COMPUTE_PGM_RSRC2.tg_size_en is set
+  s_cbranch_scc0  label_0023                        // Clean LDS if its first wave of ThreadGroup/WorkGroup
+  // CLEAR LDS
+  //
+  s_mov_b32 exec_lo, 0xffffffff
+  s_mov_b32 exec_hi, 0xffffffff
+  v_mbcnt_lo_u32_b32  v1, exec_hi, 0          // Set V1 to thread-ID (0..63)
+  v_mbcnt_hi_u32_b32  v1, exec_lo, v1        // Set V1 to thread-ID (0..63)
+  v_mul_u32_u24  v1, 0x00000008, v1          // * 8, so each thread is a double-dword address (8byte)
+  s_mov_b32     s2, 0x00000003f                    // 64 loop iterations
+  s_mov_b32     m0, 0xffffffff
+  // Clear all of LDS space
+  // Each FirstWave of WorkGroup clears 64kbyte block
+
+label_001F:
+  ds_write2_b64  v1, v[2:3], v[2:3] offset1:32
+  ds_write2_b64  v1, v[4:5], v[4:5] offset0:64 offset1:96
+  v_add_co_u32     v1, vcc, 0x00000400, v1
+  s_sub_u32     s2, s2, 1
+  s_cbranch_scc0  label_001F
+  //
+  // CLEAR SGPRs
+  //
+label_0023:
+  s_mov_b32     m0, 0x00000068  // Loop 108/4=27 times  (loop unrolled for performance)
+label_sgpr_loop:
+  s_movreld_b32     s0, 0
+  s_movreld_b32     s1, 0
+  s_movreld_b32     s2, 0
+  s_movreld_b32     s3, 0
+  s_sub_u32         m0, m0, 4
+  s_cbranch_scc0  label_sgpr_loop
+
+  //clear vcc
+  s_mov_b64 vcc, 0          //clear vcc
+  s_mov_b32 flat_scratch_lo, 0   //clear  flat scratch lo SGPR
+  s_mov_b32 flat_scratch_hi, 0   //clear  flat scratch hi SGPR
+  s_mov_b64 ttmp0, 0        //Clear ttmp0 and ttmp1
+  s_mov_b64 ttmp2, 0        //Clear ttmp2 and ttmp3
+  s_mov_b64 ttmp4, 0        //Clear ttmp4 and ttmp5
+  s_mov_b64 ttmp6, 0        //Clear ttmp6 and ttmp7
+  s_mov_b64 ttmp8, 0        //Clear ttmp8 and ttmp9
+  s_mov_b64 ttmp10, 0       //Clear ttmp10 and ttmp11
+  s_mov_b64 ttmp12, 0       //Clear ttmp12 and ttmp13
+  s_mov_b64 ttmp14, 0       //Clear ttmp14 and ttmp15
+
+ s_endpgm
+
+end
+
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0_cleaner_shader.h b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0_cleaner_shader.h
new file mode 100644
index 000000000000..3218cc04f543
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0_cleaner_shader.h
@@ -0,0 +1,56 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2024 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+/* Define the cleaner shader gfx_11_0_3 */
+static const u32 gfx_11_0_3_cleaner_shader_hex[] = {
+	0xb0804006, 0xbe8200ff,
+	0x00000058, 0xbefd0080,
+	0x7e008480, 0x7e028480,
+	0x7e048480, 0x7e068480,
+	0x7e088480, 0x7e0a8480,
+	0x7e0c8480, 0x7e0e8480,
+	0xbefd0002, 0x80828802,
+	0xbfa1fff5, 0xbe8200ff,
+	0x80000000, 0x8b020002,
+	0xbfa10012, 0xbefe00c1,
+	0xbeff00c1, 0xd71f0001,
+	0x0001007f, 0xd7200001,
+	0x0002027e, 0x16020288,
+	0xbe8200bf, 0xbefd00c1,
+	0xd9382000, 0x00020201,
+	0xd9386040, 0x00040401,
+	0xd7006a01, 0x000202ff,
+	0x00000400, 0x80828102,
+	0xbfa1fff7, 0xbefd00ff,
+	0x00000068, 0xbe804280,
+	0xbe814280, 0xbe824280,
+	0xbe834280, 0x80fd847d,
+	0xbfa1fffa, 0xbeea0180,
+	0xbeec0180, 0xbeee0180,
+	0xbef00180, 0xbef20180,
+	0xbef40180, 0xbef60180,
+	0xbef80180, 0xbefa0180,
+	0xbfb00000, 0xbf9f0000,
+	0xbf9f0000, 0xbf9f0000,
+	0xbf9f0000, 0xbf9f0000,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c
new file mode 100644
index 000000000000..d01d2712cf57
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c
@@ -0,0 +1,5793 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include <linux/delay.h>
+#include <linux/kernel.h>
+#include <linux/firmware.h>
+#include <linux/module.h>
+#include <linux/pci.h>
+#include "amdgpu.h"
+#include "amdgpu_gfx.h"
+#include "amdgpu_psp.h"
+#include "amdgpu_smu.h"
+#include "imu_v12_0.h"
+#include "soc24.h"
+#include "nvd.h"
+
+#include "gc/gc_12_0_0_offset.h"
+#include "gc/gc_12_0_0_sh_mask.h"
+#include "soc24_enum.h"
+#include "ivsrcid/gfx/irqsrcs_gfx_12_0_0.h"
+
+#include "soc15.h"
+#include "clearstate_gfx12.h"
+#include "v12_structs.h"
+#include "gfx_v12_0.h"
+#include "nbif_v6_3_1.h"
+#include "mes_v12_0.h"
+#include "mes_userqueue.h"
+#include "amdgpu_userq_fence.h"
+
+#define GFX12_NUM_GFX_RINGS	1
+#define GFX12_MEC_HPD_SIZE	2048
+
+#define RLCG_UCODE_LOADING_START_ADDRESS	0x00002000L
+
+#define regCP_GFX_MQD_CONTROL_DEFAULT                                             0x00000100
+#define regCP_GFX_HQD_VMID_DEFAULT                                                0x00000000
+#define regCP_GFX_HQD_QUEUE_PRIORITY_DEFAULT                                      0x00000000
+#define regCP_GFX_HQD_QUANTUM_DEFAULT                                             0x00000a01
+#define regCP_GFX_HQD_CNTL_DEFAULT                                                0x00f00000
+#define regCP_RB_DOORBELL_CONTROL_DEFAULT                                         0x00000000
+#define regCP_GFX_HQD_RPTR_DEFAULT                                                0x00000000
+
+#define regCP_HQD_EOP_CONTROL_DEFAULT                                             0x00000006
+#define regCP_HQD_PQ_DOORBELL_CONTROL_DEFAULT                                     0x00000000
+#define regCP_MQD_CONTROL_DEFAULT                                                 0x00000100
+#define regCP_HQD_PQ_CONTROL_DEFAULT                                              0x00308509
+#define regCP_HQD_PQ_DOORBELL_CONTROL_DEFAULT                                     0x00000000
+#define regCP_HQD_PQ_RPTR_DEFAULT                                                 0x00000000
+#define regCP_HQD_PERSISTENT_STATE_DEFAULT                                        0x0be05501
+#define regCP_HQD_IB_CONTROL_DEFAULT                                              0x00300000
+
+
+MODULE_FIRMWARE("amdgpu/gc_12_0_0_pfp.bin");
+MODULE_FIRMWARE("amdgpu/gc_12_0_0_me.bin");
+MODULE_FIRMWARE("amdgpu/gc_12_0_0_mec.bin");
+MODULE_FIRMWARE("amdgpu/gc_12_0_0_rlc.bin");
+MODULE_FIRMWARE("amdgpu/gc_12_0_0_toc.bin");
+MODULE_FIRMWARE("amdgpu/gc_12_0_1_pfp.bin");
+MODULE_FIRMWARE("amdgpu/gc_12_0_1_me.bin");
+MODULE_FIRMWARE("amdgpu/gc_12_0_1_mec.bin");
+MODULE_FIRMWARE("amdgpu/gc_12_0_1_rlc.bin");
+MODULE_FIRMWARE("amdgpu/gc_12_0_1_rlc_kicker.bin");
+MODULE_FIRMWARE("amdgpu/gc_12_0_1_toc.bin");
+
+static const struct amdgpu_hwip_reg_entry gc_reg_list_12_0[] = {
+	SOC15_REG_ENTRY_STR(GC, 0, regGRBM_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regGRBM_STATUS2),
+	SOC15_REG_ENTRY_STR(GC, 0, regGRBM_STATUS3),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_STALLED_STAT1),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_STALLED_STAT2),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_STALLED_STAT3),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_CPC_STALLED_STAT1),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_CPF_STALLED_STAT1),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_BUSY_STAT),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_CPC_BUSY_STAT),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_CPF_BUSY_STAT),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_CPC_BUSY_STAT2),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_CPF_BUSY_STAT2),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_CPF_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_ERROR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HPD_STATUS0),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_RB_BASE),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_RB_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_RB_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_RB0_BASE),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_RB0_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_RB0_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_IB1_CMD_BUFSZ),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_IB2_CMD_BUFSZ),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_IB1_BASE_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_IB1_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_IB1_BUFSZ),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_IB2_BASE_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_IB2_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_IB2_BUFSZ),
+	SOC15_REG_ENTRY_STR(GC, 0, regCPF_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regCPC_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regCPG_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regIA_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regIA_UTCL1_STATUS_2),
+	SOC15_REG_ENTRY_STR(GC, 0, regPA_CL_CNTL_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regRMI_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regSQC_CACHES),
+	SOC15_REG_ENTRY_STR(GC, 0, regSQG_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regWD_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regGCVM_L2_PROTECTION_FAULT_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, regGCVM_L2_PROTECTION_FAULT_STATUS_LO32),
+	SOC15_REG_ENTRY_STR(GC, 0, regGCVM_L2_PROTECTION_FAULT_STATUS_HI32),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_DEBUG),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MES_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MES_INSTR_PNTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_INSTR_PNTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_PFP_INSTR_PNTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_CPC_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_RS64_INSTR_PNTR0),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_RS64_INSTR_PNTR1),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_RS64_INSTR_PNTR),
+	/* cp header registers */
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MES_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MES_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MES_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MES_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MES_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MES_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MES_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MES_HEADER_DUMP),
+	/* SE status registers */
+	SOC15_REG_ENTRY_STR(GC, 0, regGRBM_STATUS_SE0),
+	SOC15_REG_ENTRY_STR(GC, 0, regGRBM_STATUS_SE1),
+	SOC15_REG_ENTRY_STR(GC, 0, regGRBM_STATUS_SE2),
+	SOC15_REG_ENTRY_STR(GC, 0, regGRBM_STATUS_SE3)
+};
+
+static const struct amdgpu_hwip_reg_entry gc_cp_reg_list_12[] = {
+	/* compute registers */
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_VMID),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PERSISTENT_STATE),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PIPE_PRIORITY),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_QUEUE_PRIORITY),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_QUANTUM),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PQ_BASE),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PQ_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PQ_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PQ_WPTR_POLL_ADDR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PQ_WPTR_POLL_ADDR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PQ_DOORBELL_CONTROL),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PQ_CONTROL),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_IB_BASE_ADDR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_IB_BASE_ADDR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_IB_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_IB_CONTROL),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_DEQUEUE_REQUEST),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_EOP_BASE_ADDR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_EOP_BASE_ADDR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_EOP_CONTROL),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_EOP_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_EOP_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_EOP_EVENTS),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_CTX_SAVE_BASE_ADDR_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_CTX_SAVE_BASE_ADDR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_CTX_SAVE_CONTROL),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_CNTL_STACK_OFFSET),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_CNTL_STACK_SIZE),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_WG_STATE_OFFSET),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_CTX_SAVE_SIZE),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_GDS_RESOURCE_STATE),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_ERROR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_EOP_WPTR_MEM),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PQ_WPTR_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PQ_WPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_SUSPEND_CNTL_STACK_OFFSET),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_SUSPEND_CNTL_STACK_DW_CNT),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_SUSPEND_WG_STATE_OFFSET),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_DEQUEUE_STATUS),
+	/* cp header registers */
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_ME1_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_ME1_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_ME1_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_ME1_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_ME1_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_ME1_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_ME1_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_ME1_HEADER_DUMP),
+};
+
+static const struct amdgpu_hwip_reg_entry gc_gfx_queue_reg_list_12[] = {
+	/* gfx queue registers */
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_ACTIVE),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_VMID),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_QUEUE_PRIORITY),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_QUANTUM),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_BASE),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_OFFSET),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_CSMD_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_WPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_DEQUEUE_REQUEST),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_MAPPED),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_QUE_MGR_CONTROL),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_HQ_CONTROL0),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_HQ_STATUS0),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_MQD_BASE_ADDR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_MQD_BASE_ADDR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_RB_WPTR_POLL_ADDR_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_RB_WPTR_POLL_ADDR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_RB_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_IB1_BASE_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_IB1_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_IB1_CMD_BUFSZ),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_IB1_BUFSZ),
+	/* cp header registers */
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_PFP_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_PFP_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_PFP_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_PFP_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_PFP_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_PFP_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_PFP_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_PFP_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP),
+};
+
+static const struct soc15_reg_golden golden_settings_gc_12_0_rev0[] = {
+	SOC15_REG_GOLDEN_VALUE(GC, 0, regDB_MEM_CONFIG, 0x0000000f, 0x0000000f),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, regCB_HW_CONTROL_1, 0x03000000, 0x03000000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, regGL2C_CTRL5, 0x00000070, 0x00000020)
+};
+
+static const struct soc15_reg_golden golden_settings_gc_12_0[] = {
+	SOC15_REG_GOLDEN_VALUE(GC, 0, regDB_MEM_CONFIG, 0x00008000, 0x00008000),
+};
+
+#define DEFAULT_SH_MEM_CONFIG \
+	((SH_MEM_ADDRESS_MODE_64 << SH_MEM_CONFIG__ADDRESS_MODE__SHIFT) | \
+	 (SH_MEM_ALIGNMENT_MODE_UNALIGNED << SH_MEM_CONFIG__ALIGNMENT_MODE__SHIFT) | \
+	 (3 << SH_MEM_CONFIG__INITIAL_INST_PREFETCH__SHIFT))
+
+static void gfx_v12_0_disable_gpa_mode(struct amdgpu_device *adev);
+static void gfx_v12_0_set_ring_funcs(struct amdgpu_device *adev);
+static void gfx_v12_0_set_irq_funcs(struct amdgpu_device *adev);
+static void gfx_v12_0_set_rlc_funcs(struct amdgpu_device *adev);
+static void gfx_v12_0_set_mqd_funcs(struct amdgpu_device *adev);
+static void gfx_v12_0_set_imu_funcs(struct amdgpu_device *adev);
+static int gfx_v12_0_get_cu_info(struct amdgpu_device *adev,
+				 struct amdgpu_cu_info *cu_info);
+static uint64_t gfx_v12_0_get_gpu_clock_counter(struct amdgpu_device *adev);
+static void gfx_v12_0_select_se_sh(struct amdgpu_device *adev, u32 se_num,
+				   u32 sh_num, u32 instance, int xcc_id);
+static u32 gfx_v12_0_get_wgp_active_bitmap_per_sh(struct amdgpu_device *adev);
+
+static void gfx_v12_0_ring_emit_frame_cntl(struct amdgpu_ring *ring, bool start, bool secure);
+static void gfx_v12_0_ring_emit_wreg(struct amdgpu_ring *ring, uint32_t reg,
+				     uint32_t val);
+static int gfx_v12_0_wait_for_rlc_autoload_complete(struct amdgpu_device *adev);
+static void gfx_v12_0_ring_invalidate_tlbs(struct amdgpu_ring *ring,
+					   uint16_t pasid, uint32_t flush_type,
+					   bool all_hub, uint8_t dst_sel);
+static void gfx_v12_0_set_safe_mode(struct amdgpu_device *adev, int xcc_id);
+static void gfx_v12_0_unset_safe_mode(struct amdgpu_device *adev, int xcc_id);
+static void gfx_v12_0_update_perf_clk(struct amdgpu_device *adev,
+				      bool enable);
+
+static void gfx_v12_0_kiq_set_resources(struct amdgpu_ring *kiq_ring,
+					uint64_t queue_mask)
+{
+	amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_SET_RESOURCES, 6));
+	amdgpu_ring_write(kiq_ring, PACKET3_SET_RESOURCES_VMID_MASK(0) |
+			  PACKET3_SET_RESOURCES_QUEUE_TYPE(0));	/* vmid_mask:0 queue_type:0 (KIQ) */
+	amdgpu_ring_write(kiq_ring, lower_32_bits(queue_mask));	/* queue mask lo */
+	amdgpu_ring_write(kiq_ring, upper_32_bits(queue_mask));	/* queue mask hi */
+	amdgpu_ring_write(kiq_ring, 0);	/* gws mask lo */
+	amdgpu_ring_write(kiq_ring, 0);	/* gws mask hi */
+	amdgpu_ring_write(kiq_ring, 0);	/* oac mask */
+	amdgpu_ring_write(kiq_ring, 0);
+}
+
+static void gfx_v12_0_kiq_map_queues(struct amdgpu_ring *kiq_ring,
+				     struct amdgpu_ring *ring)
+{
+	uint64_t mqd_addr = amdgpu_bo_gpu_offset(ring->mqd_obj);
+	uint64_t wptr_addr = ring->wptr_gpu_addr;
+	uint32_t me = 0, eng_sel = 0;
+
+	switch (ring->funcs->type) {
+	case AMDGPU_RING_TYPE_COMPUTE:
+		me = 1;
+		eng_sel = 0;
+		break;
+	case AMDGPU_RING_TYPE_GFX:
+		me = 0;
+		eng_sel = 4;
+		break;
+	case AMDGPU_RING_TYPE_MES:
+		me = 2;
+		eng_sel = 5;
+		break;
+	default:
+		WARN_ON(1);
+	}
+
+	amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_MAP_QUEUES, 5));
+	/* Q_sel:0, vmid:0, vidmem: 1, engine:0, num_Q:1*/
+	amdgpu_ring_write(kiq_ring, /* Q_sel: 0, vmid: 0, engine: 0, num_Q: 1 */
+			  PACKET3_MAP_QUEUES_QUEUE_SEL(0) | /* Queue_Sel */
+			  PACKET3_MAP_QUEUES_VMID(0) | /* VMID */
+			  PACKET3_MAP_QUEUES_QUEUE(ring->queue) |
+			  PACKET3_MAP_QUEUES_PIPE(ring->pipe) |
+			  PACKET3_MAP_QUEUES_ME((me)) |
+			  PACKET3_MAP_QUEUES_QUEUE_TYPE(0) | /*queue_type: normal compute queue */
+			  PACKET3_MAP_QUEUES_ALLOC_FORMAT(0) | /* alloc format: all_on_one_pipe */
+			  PACKET3_MAP_QUEUES_ENGINE_SEL(eng_sel) |
+			  PACKET3_MAP_QUEUES_NUM_QUEUES(1)); /* num_queues: must be 1 */
+	amdgpu_ring_write(kiq_ring, PACKET3_MAP_QUEUES_DOORBELL_OFFSET(ring->doorbell_index));
+	amdgpu_ring_write(kiq_ring, lower_32_bits(mqd_addr));
+	amdgpu_ring_write(kiq_ring, upper_32_bits(mqd_addr));
+	amdgpu_ring_write(kiq_ring, lower_32_bits(wptr_addr));
+	amdgpu_ring_write(kiq_ring, upper_32_bits(wptr_addr));
+}
+
+static void gfx_v12_0_kiq_unmap_queues(struct amdgpu_ring *kiq_ring,
+				       struct amdgpu_ring *ring,
+				       enum amdgpu_unmap_queues_action action,
+				       u64 gpu_addr, u64 seq)
+{
+	struct amdgpu_device *adev = kiq_ring->adev;
+	uint32_t eng_sel = ring->funcs->type == AMDGPU_RING_TYPE_GFX ? 4 : 0;
+
+	if (adev->enable_mes && !adev->gfx.kiq[0].ring.sched.ready) {
+		amdgpu_mes_unmap_legacy_queue(adev, ring, action, gpu_addr, seq);
+		return;
+	}
+
+	amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_UNMAP_QUEUES, 4));
+	amdgpu_ring_write(kiq_ring, /* Q_sel: 0, vmid: 0, engine: 0, num_Q: 1 */
+			  PACKET3_UNMAP_QUEUES_ACTION(action) |
+			  PACKET3_UNMAP_QUEUES_QUEUE_SEL(0) |
+			  PACKET3_UNMAP_QUEUES_ENGINE_SEL(eng_sel) |
+			  PACKET3_UNMAP_QUEUES_NUM_QUEUES(1));
+	amdgpu_ring_write(kiq_ring,
+		  PACKET3_UNMAP_QUEUES_DOORBELL_OFFSET0(ring->doorbell_index));
+
+	if (action == PREEMPT_QUEUES_NO_UNMAP) {
+		amdgpu_ring_write(kiq_ring, lower_32_bits(gpu_addr));
+		amdgpu_ring_write(kiq_ring, upper_32_bits(gpu_addr));
+		amdgpu_ring_write(kiq_ring, seq);
+	} else {
+		amdgpu_ring_write(kiq_ring, 0);
+		amdgpu_ring_write(kiq_ring, 0);
+		amdgpu_ring_write(kiq_ring, 0);
+	}
+}
+
+static void gfx_v12_0_kiq_query_status(struct amdgpu_ring *kiq_ring,
+				       struct amdgpu_ring *ring,
+				       u64 addr, u64 seq)
+{
+	uint32_t eng_sel = ring->funcs->type == AMDGPU_RING_TYPE_GFX ? 4 : 0;
+
+	amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_QUERY_STATUS, 5));
+	amdgpu_ring_write(kiq_ring,
+			  PACKET3_QUERY_STATUS_CONTEXT_ID(0) |
+			  PACKET3_QUERY_STATUS_INTERRUPT_SEL(0) |
+			  PACKET3_QUERY_STATUS_COMMAND(2));
+	amdgpu_ring_write(kiq_ring, /* Q_sel: 0, vmid: 0, engine: 0, num_Q: 1 */
+			  PACKET3_QUERY_STATUS_DOORBELL_OFFSET(ring->doorbell_index) |
+			  PACKET3_QUERY_STATUS_ENG_SEL(eng_sel));
+	amdgpu_ring_write(kiq_ring, lower_32_bits(addr));
+	amdgpu_ring_write(kiq_ring, upper_32_bits(addr));
+	amdgpu_ring_write(kiq_ring, lower_32_bits(seq));
+	amdgpu_ring_write(kiq_ring, upper_32_bits(seq));
+}
+
+static void gfx_v12_0_kiq_invalidate_tlbs(struct amdgpu_ring *kiq_ring,
+					  uint16_t pasid,
+					  uint32_t flush_type,
+					  bool all_hub)
+{
+	gfx_v12_0_ring_invalidate_tlbs(kiq_ring, pasid, flush_type, all_hub, 1);
+}
+
+static const struct kiq_pm4_funcs gfx_v12_0_kiq_pm4_funcs = {
+	.kiq_set_resources = gfx_v12_0_kiq_set_resources,
+	.kiq_map_queues = gfx_v12_0_kiq_map_queues,
+	.kiq_unmap_queues = gfx_v12_0_kiq_unmap_queues,
+	.kiq_query_status = gfx_v12_0_kiq_query_status,
+	.kiq_invalidate_tlbs = gfx_v12_0_kiq_invalidate_tlbs,
+	.set_resources_size = 8,
+	.map_queues_size = 7,
+	.unmap_queues_size = 6,
+	.query_status_size = 7,
+	.invalidate_tlbs_size = 2,
+};
+
+static void gfx_v12_0_set_kiq_pm4_funcs(struct amdgpu_device *adev)
+{
+	adev->gfx.kiq[0].pmf = &gfx_v12_0_kiq_pm4_funcs;
+}
+
+static void gfx_v12_0_wait_reg_mem(struct amdgpu_ring *ring, int eng_sel,
+				   int mem_space, int opt, uint32_t addr0,
+				   uint32_t addr1, uint32_t ref,
+				   uint32_t mask, uint32_t inv)
+{
+	amdgpu_ring_write(ring, PACKET3(PACKET3_WAIT_REG_MEM, 5));
+	amdgpu_ring_write(ring,
+			  /* memory (1) or register (0) */
+			  (WAIT_REG_MEM_MEM_SPACE(mem_space) |
+			   WAIT_REG_MEM_OPERATION(opt) | /* wait */
+			   WAIT_REG_MEM_FUNCTION(3) |  /* equal */
+			   WAIT_REG_MEM_ENGINE(eng_sel)));
+
+	if (mem_space)
+		BUG_ON(addr0 & 0x3); /* Dword align */
+	amdgpu_ring_write(ring, addr0);
+	amdgpu_ring_write(ring, addr1);
+	amdgpu_ring_write(ring, ref);
+	amdgpu_ring_write(ring, mask);
+	amdgpu_ring_write(ring, inv); /* poll interval */
+}
+
+static int gfx_v12_0_ring_test_ring(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	uint32_t scratch = SOC15_REG_OFFSET(GC, 0, regSCRATCH_REG0);
+	uint32_t tmp = 0;
+	unsigned i;
+	int r;
+
+	WREG32(scratch, 0xCAFEDEAD);
+	r = amdgpu_ring_alloc(ring, 5);
+	if (r) {
+		dev_err(adev->dev,
+			"amdgpu: cp failed to lock ring %d (%d).\n",
+			ring->idx, r);
+		return r;
+	}
+
+	if (ring->funcs->type == AMDGPU_RING_TYPE_KIQ) {
+		gfx_v12_0_ring_emit_wreg(ring, scratch, 0xDEADBEEF);
+	} else {
+		amdgpu_ring_write(ring, PACKET3(PACKET3_SET_UCONFIG_REG, 1));
+		amdgpu_ring_write(ring, scratch -
+				  PACKET3_SET_UCONFIG_REG_START);
+		amdgpu_ring_write(ring, 0xDEADBEEF);
+	}
+	amdgpu_ring_commit(ring);
+
+	for (i = 0; i < adev->usec_timeout; i++) {
+		tmp = RREG32(scratch);
+		if (tmp == 0xDEADBEEF)
+			break;
+		if (amdgpu_emu_mode == 1)
+			msleep(1);
+		else
+			udelay(1);
+	}
+
+	if (i >= adev->usec_timeout)
+		r = -ETIMEDOUT;
+	return r;
+}
+
+static int gfx_v12_0_ring_test_ib(struct amdgpu_ring *ring, long timeout)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct amdgpu_ib ib;
+	struct dma_fence *f = NULL;
+	unsigned index;
+	uint64_t gpu_addr;
+	uint32_t *cpu_ptr;
+	long r;
+
+	/* MES KIQ fw hasn't indirect buffer support for now */
+	if (adev->enable_mes_kiq &&
+	    ring->funcs->type == AMDGPU_RING_TYPE_KIQ)
+		return 0;
+
+	memset(&ib, 0, sizeof(ib));
+
+	r = amdgpu_device_wb_get(adev, &index);
+	if (r)
+		return r;
+
+	gpu_addr = adev->wb.gpu_addr + (index * 4);
+	adev->wb.wb[index] = cpu_to_le32(0xCAFEDEAD);
+	cpu_ptr = &adev->wb.wb[index];
+
+	r = amdgpu_ib_get(adev, NULL, 16, AMDGPU_IB_POOL_DIRECT, &ib);
+	if (r) {
+		dev_err(adev->dev, "amdgpu: failed to get ib (%ld).\n", r);
+		goto err1;
+	}
+
+	ib.ptr[0] = PACKET3(PACKET3_WRITE_DATA, 3);
+	ib.ptr[1] = WRITE_DATA_DST_SEL(5) | WR_CONFIRM;
+	ib.ptr[2] = lower_32_bits(gpu_addr);
+	ib.ptr[3] = upper_32_bits(gpu_addr);
+	ib.ptr[4] = 0xDEADBEEF;
+	ib.length_dw = 5;
+
+	r = amdgpu_ib_schedule(ring, 1, &ib, NULL, &f);
+	if (r)
+		goto err2;
+
+	r = dma_fence_wait_timeout(f, false, timeout);
+	if (r == 0) {
+		r = -ETIMEDOUT;
+		goto err2;
+	} else if (r < 0) {
+		goto err2;
+	}
+
+	if (le32_to_cpu(*cpu_ptr) == 0xDEADBEEF)
+		r = 0;
+	else
+		r = -EINVAL;
+err2:
+	amdgpu_ib_free(&ib, NULL);
+	dma_fence_put(f);
+err1:
+	amdgpu_device_wb_free(adev, index);
+	return r;
+}
+
+static void gfx_v12_0_free_microcode(struct amdgpu_device *adev)
+{
+	amdgpu_ucode_release(&adev->gfx.pfp_fw);
+	amdgpu_ucode_release(&adev->gfx.me_fw);
+	amdgpu_ucode_release(&adev->gfx.rlc_fw);
+	amdgpu_ucode_release(&adev->gfx.mec_fw);
+
+	kfree(adev->gfx.rlc.register_list_format);
+}
+
+static int gfx_v12_0_init_toc_microcode(struct amdgpu_device *adev, const char *ucode_prefix)
+{
+	const struct psp_firmware_header_v1_0 *toc_hdr;
+	int err = 0;
+
+	err = amdgpu_ucode_request(adev, &adev->psp.toc_fw,
+				   AMDGPU_UCODE_REQUIRED,
+				   "amdgpu/%s_toc.bin", ucode_prefix);
+	if (err)
+		goto out;
+
+	toc_hdr = (const struct psp_firmware_header_v1_0 *)adev->psp.toc_fw->data;
+	adev->psp.toc.fw_version = le32_to_cpu(toc_hdr->header.ucode_version);
+	adev->psp.toc.feature_version = le32_to_cpu(toc_hdr->sos.fw_version);
+	adev->psp.toc.size_bytes = le32_to_cpu(toc_hdr->header.ucode_size_bytes);
+	adev->psp.toc.start_addr = (uint8_t *)toc_hdr +
+			le32_to_cpu(toc_hdr->header.ucode_array_offset_bytes);
+	return 0;
+out:
+	amdgpu_ucode_release(&adev->psp.toc_fw);
+	return err;
+}
+
+static int gfx_v12_0_init_microcode(struct amdgpu_device *adev)
+{
+	char ucode_prefix[30];
+	int err;
+	const struct rlc_firmware_header_v2_0 *rlc_hdr;
+	uint16_t version_major;
+	uint16_t version_minor;
+
+	DRM_DEBUG("\n");
+
+	amdgpu_ucode_ip_version_decode(adev, GC_HWIP, ucode_prefix, sizeof(ucode_prefix));
+
+	err = amdgpu_ucode_request(adev, &adev->gfx.pfp_fw,
+				   AMDGPU_UCODE_REQUIRED,
+				   "amdgpu/%s_pfp.bin", ucode_prefix);
+	if (err)
+		goto out;
+	amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_RS64_PFP);
+	amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_RS64_PFP_P0_STACK);
+
+	err = amdgpu_ucode_request(adev, &adev->gfx.me_fw,
+				   AMDGPU_UCODE_REQUIRED,
+				   "amdgpu/%s_me.bin", ucode_prefix);
+	if (err)
+		goto out;
+	amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_RS64_ME);
+	amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_RS64_ME_P0_STACK);
+
+	if (!amdgpu_sriov_vf(adev)) {
+		if (amdgpu_is_kicker_fw(adev))
+			err = amdgpu_ucode_request(adev, &adev->gfx.rlc_fw,
+						   AMDGPU_UCODE_REQUIRED,
+						   "amdgpu/%s_rlc_kicker.bin", ucode_prefix);
+		else
+			err = amdgpu_ucode_request(adev, &adev->gfx.rlc_fw,
+						   AMDGPU_UCODE_REQUIRED,
+						   "amdgpu/%s_rlc.bin", ucode_prefix);
+		if (err)
+			goto out;
+		rlc_hdr = (const struct rlc_firmware_header_v2_0 *)adev->gfx.rlc_fw->data;
+		version_major = le16_to_cpu(rlc_hdr->header.header_version_major);
+		version_minor = le16_to_cpu(rlc_hdr->header.header_version_minor);
+		err = amdgpu_gfx_rlc_init_microcode(adev, version_major, version_minor);
+		if (err)
+			goto out;
+	}
+
+	err = amdgpu_ucode_request(adev, &adev->gfx.mec_fw,
+				   AMDGPU_UCODE_REQUIRED,
+				   "amdgpu/%s_mec.bin", ucode_prefix);
+	if (err)
+		goto out;
+	amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_RS64_MEC);
+	amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_RS64_MEC_P0_STACK);
+	amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_RS64_MEC_P1_STACK);
+
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_RLC_BACKDOOR_AUTO)
+		err = gfx_v12_0_init_toc_microcode(adev, ucode_prefix);
+
+	/* only one MEC for gfx 12 */
+	adev->gfx.mec2_fw = NULL;
+
+	if (adev->gfx.imu.funcs) {
+		if (adev->gfx.imu.funcs->init_microcode) {
+			err = adev->gfx.imu.funcs->init_microcode(adev);
+			if (err)
+				dev_err(adev->dev, "Failed to load imu firmware!\n");
+		}
+	}
+
+out:
+	if (err) {
+		amdgpu_ucode_release(&adev->gfx.pfp_fw);
+		amdgpu_ucode_release(&adev->gfx.me_fw);
+		amdgpu_ucode_release(&adev->gfx.rlc_fw);
+		amdgpu_ucode_release(&adev->gfx.mec_fw);
+	}
+
+	return err;
+}
+
+static u32 gfx_v12_0_get_csb_size(struct amdgpu_device *adev)
+{
+	u32 count = 0;
+	const struct cs_section_def *sect = NULL;
+	const struct cs_extent_def *ext = NULL;
+
+	count += 1;
+
+	for (sect = gfx12_cs_data; sect->section != NULL; ++sect) {
+		if (sect->id == SECT_CONTEXT) {
+			for (ext = sect->section; ext->extent != NULL; ++ext)
+				count += 2 + ext->reg_count;
+		} else
+			return 0;
+	}
+
+	return count;
+}
+
+static void gfx_v12_0_get_csb_buffer(struct amdgpu_device *adev, u32 *buffer)
+{
+	u32 count = 0, clustercount = 0, i;
+	const struct cs_section_def *sect = NULL;
+	const struct cs_extent_def *ext = NULL;
+
+	if (adev->gfx.rlc.cs_data == NULL)
+		return;
+	if (buffer == NULL)
+		return;
+
+	count += 1;
+
+	for (sect = adev->gfx.rlc.cs_data; sect->section != NULL; ++sect) {
+		if (sect->id == SECT_CONTEXT) {
+			for (ext = sect->section; ext->extent != NULL; ++ext) {
+				clustercount++;
+				buffer[count++] = ext->reg_count;
+				buffer[count++] = ext->reg_index;
+
+				for (i = 0; i < ext->reg_count; i++)
+					buffer[count++] = cpu_to_le32(ext->extent[i]);
+			}
+		} else
+			return;
+	}
+
+	buffer[0] = clustercount;
+}
+
+static void gfx_v12_0_rlc_fini(struct amdgpu_device *adev)
+{
+	/* clear state block */
+	amdgpu_bo_free_kernel(&adev->gfx.rlc.clear_state_obj,
+			&adev->gfx.rlc.clear_state_gpu_addr,
+			(void **)&adev->gfx.rlc.cs_ptr);
+
+	/* jump table block */
+	amdgpu_bo_free_kernel(&adev->gfx.rlc.cp_table_obj,
+			&adev->gfx.rlc.cp_table_gpu_addr,
+			(void **)&adev->gfx.rlc.cp_table_ptr);
+}
+
+static void gfx_v12_0_init_rlcg_reg_access_ctrl(struct amdgpu_device *adev)
+{
+	struct amdgpu_rlcg_reg_access_ctrl *reg_access_ctrl;
+
+	reg_access_ctrl = &adev->gfx.rlc.reg_access_ctrl[0];
+	reg_access_ctrl->scratch_reg0 = SOC15_REG_OFFSET(GC, 0, regSCRATCH_REG0);
+	reg_access_ctrl->scratch_reg1 = SOC15_REG_OFFSET(GC, 0, regSCRATCH_REG1);
+	reg_access_ctrl->scratch_reg2 = SOC15_REG_OFFSET(GC, 0, regSCRATCH_REG2);
+	reg_access_ctrl->scratch_reg3 = SOC15_REG_OFFSET(GC, 0, regSCRATCH_REG3);
+	reg_access_ctrl->grbm_cntl = SOC15_REG_OFFSET(GC, 0, regGRBM_GFX_CNTL);
+	reg_access_ctrl->grbm_idx = SOC15_REG_OFFSET(GC, 0, regGRBM_GFX_INDEX);
+	reg_access_ctrl->spare_int = SOC15_REG_OFFSET(GC, 0, regRLC_SPARE_INT_0);
+	adev->gfx.rlc.rlcg_reg_access_supported = true;
+}
+
+static int gfx_v12_0_rlc_init(struct amdgpu_device *adev)
+{
+	const struct cs_section_def *cs_data;
+	int r;
+
+	adev->gfx.rlc.cs_data = gfx12_cs_data;
+
+	cs_data = adev->gfx.rlc.cs_data;
+
+	if (cs_data) {
+		/* init clear state block */
+		r = amdgpu_gfx_rlc_init_csb(adev);
+		if (r)
+			return r;
+	}
+
+	/* init spm vmid with 0xf */
+	if (adev->gfx.rlc.funcs->update_spm_vmid)
+		adev->gfx.rlc.funcs->update_spm_vmid(adev, NULL, 0xf);
+
+	return 0;
+}
+
+static void gfx_v12_0_mec_fini(struct amdgpu_device *adev)
+{
+	amdgpu_bo_free_kernel(&adev->gfx.mec.hpd_eop_obj, NULL, NULL);
+	amdgpu_bo_free_kernel(&adev->gfx.mec.mec_fw_obj, NULL, NULL);
+	amdgpu_bo_free_kernel(&adev->gfx.mec.mec_fw_data_obj, NULL, NULL);
+}
+
+static void gfx_v12_0_me_init(struct amdgpu_device *adev)
+{
+	bitmap_zero(adev->gfx.me.queue_bitmap, AMDGPU_MAX_GFX_QUEUES);
+
+	amdgpu_gfx_graphics_queue_acquire(adev);
+}
+
+static int gfx_v12_0_mec_init(struct amdgpu_device *adev)
+{
+	int r;
+	u32 *hpd;
+	size_t mec_hpd_size;
+
+	bitmap_zero(adev->gfx.mec_bitmap[0].queue_bitmap, AMDGPU_MAX_COMPUTE_QUEUES);
+
+	/* take ownership of the relevant compute queues */
+	amdgpu_gfx_compute_queue_acquire(adev);
+	mec_hpd_size = adev->gfx.num_compute_rings * GFX12_MEC_HPD_SIZE;
+
+	if (mec_hpd_size) {
+		r = amdgpu_bo_create_reserved(adev, mec_hpd_size, PAGE_SIZE,
+					      AMDGPU_GEM_DOMAIN_GTT,
+					      &adev->gfx.mec.hpd_eop_obj,
+					      &adev->gfx.mec.hpd_eop_gpu_addr,
+					      (void **)&hpd);
+		if (r) {
+			dev_warn(adev->dev, "(%d) create HDP EOP bo failed\n", r);
+			gfx_v12_0_mec_fini(adev);
+			return r;
+		}
+
+		memset(hpd, 0, mec_hpd_size);
+
+		amdgpu_bo_kunmap(adev->gfx.mec.hpd_eop_obj);
+		amdgpu_bo_unreserve(adev->gfx.mec.hpd_eop_obj);
+	}
+
+	return 0;
+}
+
+static uint32_t wave_read_ind(struct amdgpu_device *adev, uint32_t wave, uint32_t address)
+{
+	WREG32_SOC15(GC, 0, regSQ_IND_INDEX,
+		(wave << SQ_IND_INDEX__WAVE_ID__SHIFT) |
+		(address << SQ_IND_INDEX__INDEX__SHIFT));
+	return RREG32_SOC15(GC, 0, regSQ_IND_DATA);
+}
+
+static void wave_read_regs(struct amdgpu_device *adev, uint32_t wave,
+			   uint32_t thread, uint32_t regno,
+			   uint32_t num, uint32_t *out)
+{
+	WREG32_SOC15(GC, 0, regSQ_IND_INDEX,
+		(wave << SQ_IND_INDEX__WAVE_ID__SHIFT) |
+		(regno << SQ_IND_INDEX__INDEX__SHIFT) |
+		(thread << SQ_IND_INDEX__WORKITEM_ID__SHIFT) |
+		(SQ_IND_INDEX__AUTO_INCR_MASK));
+	while (num--)
+		*(out++) = RREG32_SOC15(GC, 0, regSQ_IND_DATA);
+}
+
+static void gfx_v12_0_read_wave_data(struct amdgpu_device *adev,
+				     uint32_t xcc_id,
+				     uint32_t simd, uint32_t wave,
+				     uint32_t *dst, int *no_fields)
+{
+	/* in gfx12 the SIMD_ID is specified as part of the INSTANCE
+	 * field when performing a select_se_sh so it should be
+	 * zero here */
+	WARN_ON(simd != 0);
+
+	/* type 4 wave data */
+	dst[(*no_fields)++] = 4;
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_STATUS);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_PC_LO);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_PC_HI);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_EXEC_LO);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_EXEC_HI);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_HW_ID1);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_HW_ID2);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_GPR_ALLOC);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_LDS_ALLOC);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_IB_STS);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_IB_STS2);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_IB_DBG1);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_M0);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_MODE);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_STATE_PRIV);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_EXCP_FLAG_PRIV);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_EXCP_FLAG_USER);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_TRAP_CTRL);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_ACTIVE);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_VALID_AND_IDLE);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_DVGPR_ALLOC_LO);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_DVGPR_ALLOC_HI);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_SCHED_MODE);
+}
+
+static void gfx_v12_0_read_wave_sgprs(struct amdgpu_device *adev,
+				      uint32_t xcc_id, uint32_t simd,
+				      uint32_t wave, uint32_t start,
+				      uint32_t size, uint32_t *dst)
+{
+	WARN_ON(simd != 0);
+
+	wave_read_regs(
+		adev, wave, 0, start + SQIND_WAVE_SGPRS_OFFSET, size,
+		dst);
+}
+
+static void gfx_v12_0_read_wave_vgprs(struct amdgpu_device *adev,
+				      uint32_t xcc_id, uint32_t simd,
+				      uint32_t wave, uint32_t thread,
+				      uint32_t start, uint32_t size,
+				      uint32_t *dst)
+{
+	wave_read_regs(
+		adev, wave, thread,
+		start + SQIND_WAVE_VGPRS_OFFSET, size, dst);
+}
+
+static void gfx_v12_0_select_me_pipe_q(struct amdgpu_device *adev,
+				       u32 me, u32 pipe, u32 q, u32 vm, u32 xcc_id)
+{
+	soc24_grbm_select(adev, me, pipe, q, vm);
+}
+
+/* all sizes are in bytes */
+#define MQD_SHADOW_BASE_SIZE      73728
+#define MQD_SHADOW_BASE_ALIGNMENT 256
+#define MQD_FWWORKAREA_SIZE       484
+#define MQD_FWWORKAREA_ALIGNMENT  256
+
+static void gfx_v12_0_get_gfx_shadow_info_nocheck(struct amdgpu_device *adev,
+						  struct amdgpu_gfx_shadow_info *shadow_info)
+{
+	shadow_info->shadow_size = MQD_SHADOW_BASE_SIZE;
+	shadow_info->shadow_alignment = MQD_SHADOW_BASE_ALIGNMENT;
+	shadow_info->csa_size = MQD_FWWORKAREA_SIZE;
+	shadow_info->csa_alignment = MQD_FWWORKAREA_ALIGNMENT;
+}
+
+static int gfx_v12_0_get_gfx_shadow_info(struct amdgpu_device *adev,
+					 struct amdgpu_gfx_shadow_info *shadow_info,
+					 bool skip_check)
+{
+	if (adev->gfx.cp_gfx_shadow || skip_check) {
+		gfx_v12_0_get_gfx_shadow_info_nocheck(adev, shadow_info);
+		return 0;
+	}
+
+	memset(shadow_info, 0, sizeof(struct amdgpu_gfx_shadow_info));
+	return -EINVAL;
+}
+
+static const struct amdgpu_gfx_funcs gfx_v12_0_gfx_funcs = {
+	.get_gpu_clock_counter = &gfx_v12_0_get_gpu_clock_counter,
+	.select_se_sh = &gfx_v12_0_select_se_sh,
+	.read_wave_data = &gfx_v12_0_read_wave_data,
+	.read_wave_sgprs = &gfx_v12_0_read_wave_sgprs,
+	.read_wave_vgprs = &gfx_v12_0_read_wave_vgprs,
+	.select_me_pipe_q = &gfx_v12_0_select_me_pipe_q,
+	.update_perfmon_mgcg = &gfx_v12_0_update_perf_clk,
+	.get_gfx_shadow_info = &gfx_v12_0_get_gfx_shadow_info,
+};
+
+static int gfx_v12_0_gpu_early_init(struct amdgpu_device *adev)
+{
+
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	case IP_VERSION(12, 0, 0):
+	case IP_VERSION(12, 0, 1):
+		adev->gfx.config.max_hw_contexts = 8;
+		adev->gfx.config.sc_prim_fifo_size_frontend = 0x20;
+		adev->gfx.config.sc_prim_fifo_size_backend = 0x100;
+		adev->gfx.config.sc_hiz_tile_fifo_size = 0;
+		adev->gfx.config.sc_earlyz_tile_fifo_size = 0x4C0;
+		break;
+	default:
+		BUG();
+		break;
+	}
+
+	return 0;
+}
+
+static int gfx_v12_0_gfx_ring_init(struct amdgpu_device *adev, int ring_id,
+				   int me, int pipe, int queue)
+{
+	int r;
+	struct amdgpu_ring *ring;
+	unsigned int irq_type;
+
+	ring = &adev->gfx.gfx_ring[ring_id];
+
+	ring->me = me;
+	ring->pipe = pipe;
+	ring->queue = queue;
+
+	ring->ring_obj = NULL;
+	ring->use_doorbell = true;
+
+	if (!ring_id)
+		ring->doorbell_index = adev->doorbell_index.gfx_ring0 << 1;
+	else
+		ring->doorbell_index = adev->doorbell_index.gfx_ring1 << 1;
+	ring->vm_hub = AMDGPU_GFXHUB(0);
+	sprintf(ring->name, "gfx_%d.%d.%d", ring->me, ring->pipe, ring->queue);
+
+	irq_type = AMDGPU_CP_IRQ_GFX_ME0_PIPE0_EOP + ring->pipe;
+	r = amdgpu_ring_init(adev, ring, 1024, &adev->gfx.eop_irq, irq_type,
+			     AMDGPU_RING_PRIO_DEFAULT, NULL);
+	if (r)
+		return r;
+	return 0;
+}
+
+static int gfx_v12_0_compute_ring_init(struct amdgpu_device *adev, int ring_id,
+				       int mec, int pipe, int queue)
+{
+	int r;
+	unsigned irq_type;
+	struct amdgpu_ring *ring;
+	unsigned int hw_prio;
+
+	ring = &adev->gfx.compute_ring[ring_id];
+
+	/* mec0 is me1 */
+	ring->me = mec + 1;
+	ring->pipe = pipe;
+	ring->queue = queue;
+
+	ring->ring_obj = NULL;
+	ring->use_doorbell = true;
+	ring->doorbell_index = (adev->doorbell_index.mec_ring0 + ring_id) << 1;
+	ring->eop_gpu_addr = adev->gfx.mec.hpd_eop_gpu_addr
+				+ (ring_id * GFX12_MEC_HPD_SIZE);
+	ring->vm_hub = AMDGPU_GFXHUB(0);
+	sprintf(ring->name, "comp_%d.%d.%d", ring->me, ring->pipe, ring->queue);
+
+	irq_type = AMDGPU_CP_IRQ_COMPUTE_MEC1_PIPE0_EOP
+		+ ((ring->me - 1) * adev->gfx.mec.num_pipe_per_mec)
+		+ ring->pipe;
+	hw_prio = amdgpu_gfx_is_high_priority_compute_queue(adev, ring) ?
+			AMDGPU_GFX_PIPE_PRIO_HIGH : AMDGPU_GFX_PIPE_PRIO_NORMAL;
+	/* type-2 packets are deprecated on MEC, use type-3 instead */
+	r = amdgpu_ring_init(adev, ring, 1024, &adev->gfx.eop_irq, irq_type,
+			     hw_prio, NULL);
+	if (r)
+		return r;
+
+	return 0;
+}
+
+static struct {
+	SOC24_FIRMWARE_ID	id;
+	unsigned int		offset;
+	unsigned int		size;
+	unsigned int		size_x16;
+} rlc_autoload_info[SOC24_FIRMWARE_ID_MAX];
+
+#define RLC_TOC_OFFSET_DWUNIT   8
+#define RLC_SIZE_MULTIPLE       1024
+#define RLC_TOC_UMF_SIZE_inM	23ULL
+#define RLC_TOC_FORMAT_API	165ULL
+
+static void gfx_v12_0_parse_rlc_toc(struct amdgpu_device *adev, void *rlc_toc)
+{
+	RLC_TABLE_OF_CONTENT_V2 *ucode = rlc_toc;
+
+	while (ucode && (ucode->id > SOC24_FIRMWARE_ID_INVALID)) {
+		rlc_autoload_info[ucode->id].id = ucode->id;
+		rlc_autoload_info[ucode->id].offset =
+			ucode->offset * RLC_TOC_OFFSET_DWUNIT * 4;
+		rlc_autoload_info[ucode->id].size =
+			ucode->size_x16 ? ucode->size * RLC_SIZE_MULTIPLE * 4 :
+					  ucode->size * 4;
+		ucode++;
+	}
+}
+
+static uint32_t gfx_v12_0_calc_toc_total_size(struct amdgpu_device *adev)
+{
+	uint32_t total_size = 0;
+	SOC24_FIRMWARE_ID id;
+
+	gfx_v12_0_parse_rlc_toc(adev, adev->psp.toc.start_addr);
+
+	for (id = SOC24_FIRMWARE_ID_RLC_G_UCODE; id < SOC24_FIRMWARE_ID_MAX; id++)
+		total_size += rlc_autoload_info[id].size;
+
+	/* In case the offset in rlc toc ucode is aligned */
+	if (total_size < rlc_autoload_info[SOC24_FIRMWARE_ID_MAX-1].offset)
+		total_size = rlc_autoload_info[SOC24_FIRMWARE_ID_MAX-1].offset +
+			rlc_autoload_info[SOC24_FIRMWARE_ID_MAX-1].size;
+	if (total_size < (RLC_TOC_UMF_SIZE_inM << 20))
+		total_size = RLC_TOC_UMF_SIZE_inM << 20;
+
+	return total_size;
+}
+
+static int gfx_v12_0_rlc_autoload_buffer_init(struct amdgpu_device *adev)
+{
+	int r;
+	uint32_t total_size;
+
+	total_size = gfx_v12_0_calc_toc_total_size(adev);
+
+	r = amdgpu_bo_create_reserved(adev, total_size, 64 * 1024,
+				      AMDGPU_GEM_DOMAIN_VRAM,
+				      &adev->gfx.rlc.rlc_autoload_bo,
+				      &adev->gfx.rlc.rlc_autoload_gpu_addr,
+				      (void **)&adev->gfx.rlc.rlc_autoload_ptr);
+
+	if (r) {
+		dev_err(adev->dev, "(%d) failed to create fw autoload bo\n", r);
+		return r;
+	}
+
+	return 0;
+}
+
+static void gfx_v12_0_rlc_backdoor_autoload_copy_ucode(struct amdgpu_device *adev,
+						       SOC24_FIRMWARE_ID id,
+						       const void *fw_data,
+						       uint32_t fw_size)
+{
+	uint32_t toc_offset;
+	uint32_t toc_fw_size;
+	char *ptr = adev->gfx.rlc.rlc_autoload_ptr;
+
+	if (id <= SOC24_FIRMWARE_ID_INVALID || id >= SOC24_FIRMWARE_ID_MAX)
+		return;
+
+	toc_offset = rlc_autoload_info[id].offset;
+	toc_fw_size = rlc_autoload_info[id].size;
+
+	if (fw_size == 0)
+		fw_size = toc_fw_size;
+
+	if (fw_size > toc_fw_size)
+		fw_size = toc_fw_size;
+
+	memcpy(ptr + toc_offset, fw_data, fw_size);
+
+	if (fw_size < toc_fw_size)
+		memset(ptr + toc_offset + fw_size, 0, toc_fw_size - fw_size);
+}
+
+static void
+gfx_v12_0_rlc_backdoor_autoload_copy_toc_ucode(struct amdgpu_device *adev)
+{
+	void *data;
+	uint32_t size;
+	uint32_t *toc_ptr;
+
+	data = adev->psp.toc.start_addr;
+	size = rlc_autoload_info[SOC24_FIRMWARE_ID_RLC_TOC].size;
+
+	toc_ptr = (uint32_t *)data + size / 4 - 2;
+	*toc_ptr = (RLC_TOC_FORMAT_API << 24) | 0x1;
+
+	gfx_v12_0_rlc_backdoor_autoload_copy_ucode(adev, SOC24_FIRMWARE_ID_RLC_TOC,
+						   data, size);
+}
+
+static void
+gfx_v12_0_rlc_backdoor_autoload_copy_gfx_ucode(struct amdgpu_device *adev)
+{
+	const __le32 *fw_data;
+	uint32_t fw_size;
+	const struct gfx_firmware_header_v2_0 *cpv2_hdr;
+	const struct rlc_firmware_header_v2_0 *rlc_hdr;
+	const struct rlc_firmware_header_v2_1 *rlcv21_hdr;
+	const struct rlc_firmware_header_v2_2 *rlcv22_hdr;
+	uint16_t version_major, version_minor;
+
+	/* pfp ucode */
+	cpv2_hdr = (const struct gfx_firmware_header_v2_0 *)
+		adev->gfx.pfp_fw->data;
+	/* instruction */
+	fw_data = (const __le32 *)(adev->gfx.pfp_fw->data +
+		le32_to_cpu(cpv2_hdr->ucode_offset_bytes));
+	fw_size = le32_to_cpu(cpv2_hdr->ucode_size_bytes);
+	gfx_v12_0_rlc_backdoor_autoload_copy_ucode(adev, SOC24_FIRMWARE_ID_RS64_PFP,
+						   fw_data, fw_size);
+	/* data */
+	fw_data = (const __le32 *)(adev->gfx.pfp_fw->data +
+		le32_to_cpu(cpv2_hdr->data_offset_bytes));
+	fw_size = le32_to_cpu(cpv2_hdr->data_size_bytes);
+	gfx_v12_0_rlc_backdoor_autoload_copy_ucode(adev, SOC24_FIRMWARE_ID_RS64_PFP_P0_STACK,
+						   fw_data, fw_size);
+	gfx_v12_0_rlc_backdoor_autoload_copy_ucode(adev, SOC24_FIRMWARE_ID_RS64_PFP_P1_STACK,
+						   fw_data, fw_size);
+	/* me ucode */
+	cpv2_hdr = (const struct gfx_firmware_header_v2_0 *)
+		adev->gfx.me_fw->data;
+	/* instruction */
+	fw_data = (const __le32 *)(adev->gfx.me_fw->data +
+		le32_to_cpu(cpv2_hdr->ucode_offset_bytes));
+	fw_size = le32_to_cpu(cpv2_hdr->ucode_size_bytes);
+	gfx_v12_0_rlc_backdoor_autoload_copy_ucode(adev, SOC24_FIRMWARE_ID_RS64_ME,
+						   fw_data, fw_size);
+	/* data */
+	fw_data = (const __le32 *)(adev->gfx.me_fw->data +
+		le32_to_cpu(cpv2_hdr->data_offset_bytes));
+	fw_size = le32_to_cpu(cpv2_hdr->data_size_bytes);
+	gfx_v12_0_rlc_backdoor_autoload_copy_ucode(adev, SOC24_FIRMWARE_ID_RS64_ME_P0_STACK,
+						   fw_data, fw_size);
+	gfx_v12_0_rlc_backdoor_autoload_copy_ucode(adev, SOC24_FIRMWARE_ID_RS64_ME_P1_STACK,
+						   fw_data, fw_size);
+	/* mec ucode */
+	cpv2_hdr = (const struct gfx_firmware_header_v2_0 *)
+		adev->gfx.mec_fw->data;
+	/* instruction */
+	fw_data = (const __le32 *) (adev->gfx.mec_fw->data +
+		le32_to_cpu(cpv2_hdr->ucode_offset_bytes));
+	fw_size = le32_to_cpu(cpv2_hdr->ucode_size_bytes);
+	gfx_v12_0_rlc_backdoor_autoload_copy_ucode(adev, SOC24_FIRMWARE_ID_RS64_MEC,
+						   fw_data, fw_size);
+	/* data */
+	fw_data = (const __le32 *) (adev->gfx.mec_fw->data +
+		le32_to_cpu(cpv2_hdr->data_offset_bytes));
+	fw_size = le32_to_cpu(cpv2_hdr->data_size_bytes);
+	gfx_v12_0_rlc_backdoor_autoload_copy_ucode(adev, SOC24_FIRMWARE_ID_RS64_MEC_P0_STACK,
+						   fw_data, fw_size);
+	gfx_v12_0_rlc_backdoor_autoload_copy_ucode(adev, SOC24_FIRMWARE_ID_RS64_MEC_P1_STACK,
+						   fw_data, fw_size);
+	gfx_v12_0_rlc_backdoor_autoload_copy_ucode(adev, SOC24_FIRMWARE_ID_RS64_MEC_P2_STACK,
+						   fw_data, fw_size);
+	gfx_v12_0_rlc_backdoor_autoload_copy_ucode(adev, SOC24_FIRMWARE_ID_RS64_MEC_P3_STACK,
+						   fw_data, fw_size);
+
+	/* rlc ucode */
+	rlc_hdr = (const struct rlc_firmware_header_v2_0 *)
+		adev->gfx.rlc_fw->data;
+	fw_data = (const __le32 *)(adev->gfx.rlc_fw->data +
+			le32_to_cpu(rlc_hdr->header.ucode_array_offset_bytes));
+	fw_size = le32_to_cpu(rlc_hdr->header.ucode_size_bytes);
+	gfx_v12_0_rlc_backdoor_autoload_copy_ucode(adev, SOC24_FIRMWARE_ID_RLC_G_UCODE,
+						   fw_data, fw_size);
+
+	version_major = le16_to_cpu(rlc_hdr->header.header_version_major);
+	version_minor = le16_to_cpu(rlc_hdr->header.header_version_minor);
+	if (version_major == 2) {
+		if (version_minor >= 1) {
+			rlcv21_hdr = (const struct rlc_firmware_header_v2_1 *)adev->gfx.rlc_fw->data;
+
+			fw_data = (const __le32 *)(adev->gfx.rlc_fw->data +
+					le32_to_cpu(rlcv21_hdr->save_restore_list_gpm_offset_bytes));
+			fw_size = le32_to_cpu(rlcv21_hdr->save_restore_list_gpm_size_bytes);
+			gfx_v12_0_rlc_backdoor_autoload_copy_ucode(adev, SOC24_FIRMWARE_ID_RLCG_SCRATCH,
+						   fw_data, fw_size);
+
+			fw_data = (const __le32 *)(adev->gfx.rlc_fw->data +
+					le32_to_cpu(rlcv21_hdr->save_restore_list_srm_offset_bytes));
+			fw_size = le32_to_cpu(rlcv21_hdr->save_restore_list_srm_size_bytes);
+			gfx_v12_0_rlc_backdoor_autoload_copy_ucode(adev, SOC24_FIRMWARE_ID_RLC_SRM_ARAM,
+						   fw_data, fw_size);
+		}
+		if (version_minor >= 2) {
+			rlcv22_hdr = (const struct rlc_firmware_header_v2_2 *)adev->gfx.rlc_fw->data;
+
+			fw_data = (const __le32 *)(adev->gfx.rlc_fw->data +
+					le32_to_cpu(rlcv22_hdr->rlc_iram_ucode_offset_bytes));
+			fw_size = le32_to_cpu(rlcv22_hdr->rlc_iram_ucode_size_bytes);
+			gfx_v12_0_rlc_backdoor_autoload_copy_ucode(adev, SOC24_FIRMWARE_ID_RLX6_UCODE,
+						   fw_data, fw_size);
+
+			fw_data = (const __le32 *)(adev->gfx.rlc_fw->data +
+					le32_to_cpu(rlcv22_hdr->rlc_dram_ucode_offset_bytes));
+			fw_size = le32_to_cpu(rlcv22_hdr->rlc_dram_ucode_size_bytes);
+			gfx_v12_0_rlc_backdoor_autoload_copy_ucode(adev, SOC24_FIRMWARE_ID_RLX6_DRAM_BOOT,
+						   fw_data, fw_size);
+		}
+	}
+}
+
+static void
+gfx_v12_0_rlc_backdoor_autoload_copy_sdma_ucode(struct amdgpu_device *adev)
+{
+	const __le32 *fw_data;
+	uint32_t fw_size;
+	const struct sdma_firmware_header_v3_0 *sdma_hdr;
+
+	sdma_hdr = (const struct sdma_firmware_header_v3_0 *)
+		adev->sdma.instance[0].fw->data;
+	fw_data = (const __le32 *) (adev->sdma.instance[0].fw->data +
+			le32_to_cpu(sdma_hdr->ucode_offset_bytes));
+	fw_size = le32_to_cpu(sdma_hdr->ucode_size_bytes);
+
+	gfx_v12_0_rlc_backdoor_autoload_copy_ucode(adev, SOC24_FIRMWARE_ID_SDMA_UCODE_TH0,
+						   fw_data, fw_size);
+}
+
+static void
+gfx_v12_0_rlc_backdoor_autoload_copy_mes_ucode(struct amdgpu_device *adev)
+{
+	const __le32 *fw_data;
+	unsigned fw_size;
+	const struct mes_firmware_header_v1_0 *mes_hdr;
+	int pipe, ucode_id, data_id;
+
+	for (pipe = 0; pipe < 2; pipe++) {
+		if (pipe == 0) {
+			ucode_id = SOC24_FIRMWARE_ID_RS64_MES_P0;
+			data_id  = SOC24_FIRMWARE_ID_RS64_MES_P0_STACK;
+		} else {
+			ucode_id = SOC24_FIRMWARE_ID_RS64_MES_P1;
+			data_id  = SOC24_FIRMWARE_ID_RS64_MES_P1_STACK;
+		}
+
+		mes_hdr = (const struct mes_firmware_header_v1_0 *)
+			adev->mes.fw[pipe]->data;
+
+		fw_data = (const __le32 *)(adev->mes.fw[pipe]->data +
+				le32_to_cpu(mes_hdr->mes_ucode_offset_bytes));
+		fw_size = le32_to_cpu(mes_hdr->mes_ucode_size_bytes);
+
+		gfx_v12_0_rlc_backdoor_autoload_copy_ucode(adev, ucode_id, fw_data, fw_size);
+
+		fw_data = (const __le32 *)(adev->mes.fw[pipe]->data +
+				le32_to_cpu(mes_hdr->mes_ucode_data_offset_bytes));
+		fw_size = le32_to_cpu(mes_hdr->mes_ucode_data_size_bytes);
+
+		gfx_v12_0_rlc_backdoor_autoload_copy_ucode(adev, data_id, fw_data, fw_size);
+	}
+}
+
+static int gfx_v12_0_rlc_backdoor_autoload_enable(struct amdgpu_device *adev)
+{
+	uint32_t rlc_g_offset, rlc_g_size;
+	uint64_t gpu_addr;
+	uint32_t data;
+
+	/* RLC autoload sequence 2: copy ucode */
+	gfx_v12_0_rlc_backdoor_autoload_copy_sdma_ucode(adev);
+	gfx_v12_0_rlc_backdoor_autoload_copy_gfx_ucode(adev);
+	gfx_v12_0_rlc_backdoor_autoload_copy_mes_ucode(adev);
+	gfx_v12_0_rlc_backdoor_autoload_copy_toc_ucode(adev);
+
+	rlc_g_offset = rlc_autoload_info[SOC24_FIRMWARE_ID_RLC_G_UCODE].offset;
+	rlc_g_size = rlc_autoload_info[SOC24_FIRMWARE_ID_RLC_G_UCODE].size;
+	gpu_addr = adev->gfx.rlc.rlc_autoload_gpu_addr + rlc_g_offset - adev->gmc.vram_start;
+
+	WREG32_SOC15(GC, 0, regGFX_IMU_RLC_BOOTLOADER_ADDR_HI, upper_32_bits(gpu_addr));
+	WREG32_SOC15(GC, 0, regGFX_IMU_RLC_BOOTLOADER_ADDR_LO, lower_32_bits(gpu_addr));
+
+	WREG32_SOC15(GC, 0, regGFX_IMU_RLC_BOOTLOADER_SIZE, rlc_g_size);
+
+	if (adev->gfx.imu.funcs && (amdgpu_dpm > 0)) {
+		/* RLC autoload sequence 3: load IMU fw */
+		if (adev->gfx.imu.funcs->load_microcode)
+			adev->gfx.imu.funcs->load_microcode(adev);
+		/* RLC autoload sequence 4 init IMU fw */
+		if (adev->gfx.imu.funcs->setup_imu)
+			adev->gfx.imu.funcs->setup_imu(adev);
+		if (adev->gfx.imu.funcs->start_imu)
+			adev->gfx.imu.funcs->start_imu(adev);
+
+		/* RLC autoload sequence 5 disable gpa mode */
+		gfx_v12_0_disable_gpa_mode(adev);
+	} else {
+		/* unhalt rlc to start autoload without imu */
+		data = RREG32_SOC15(GC, 0, regRLC_GPM_THREAD_ENABLE);
+		data = REG_SET_FIELD(data, RLC_GPM_THREAD_ENABLE, THREAD0_ENABLE, 1);
+		data = REG_SET_FIELD(data, RLC_GPM_THREAD_ENABLE, THREAD1_ENABLE, 1);
+		WREG32_SOC15(GC, 0, regRLC_GPM_THREAD_ENABLE, data);
+		WREG32_SOC15(GC, 0, regRLC_CNTL, RLC_CNTL__RLC_ENABLE_F32_MASK);
+	}
+
+	return 0;
+}
+
+static void gfx_v12_0_alloc_ip_dump(struct amdgpu_device *adev)
+{
+	uint32_t reg_count = ARRAY_SIZE(gc_reg_list_12_0);
+	uint32_t *ptr;
+	uint32_t inst;
+
+	ptr = kcalloc(reg_count, sizeof(uint32_t), GFP_KERNEL);
+	if (!ptr) {
+		DRM_ERROR("Failed to allocate memory for GFX IP Dump\n");
+		adev->gfx.ip_dump_core = NULL;
+	} else {
+		adev->gfx.ip_dump_core = ptr;
+	}
+
+	/* Allocate memory for compute queue registers for all the instances */
+	reg_count = ARRAY_SIZE(gc_cp_reg_list_12);
+	inst = adev->gfx.mec.num_mec * adev->gfx.mec.num_pipe_per_mec *
+		adev->gfx.mec.num_queue_per_pipe;
+
+	ptr = kcalloc(reg_count * inst, sizeof(uint32_t), GFP_KERNEL);
+	if (!ptr) {
+		DRM_ERROR("Failed to allocate memory for Compute Queues IP Dump\n");
+		adev->gfx.ip_dump_compute_queues = NULL;
+	} else {
+		adev->gfx.ip_dump_compute_queues = ptr;
+	}
+
+	/* Allocate memory for gfx queue registers for all the instances */
+	reg_count = ARRAY_SIZE(gc_gfx_queue_reg_list_12);
+	inst = adev->gfx.me.num_me * adev->gfx.me.num_pipe_per_me *
+		adev->gfx.me.num_queue_per_pipe;
+
+	ptr = kcalloc(reg_count * inst, sizeof(uint32_t), GFP_KERNEL);
+	if (!ptr) {
+		DRM_ERROR("Failed to allocate memory for GFX Queues IP Dump\n");
+		adev->gfx.ip_dump_gfx_queues = NULL;
+	} else {
+		adev->gfx.ip_dump_gfx_queues = ptr;
+	}
+}
+
+static int gfx_v12_0_sw_init(struct amdgpu_ip_block *ip_block)
+{
+	int i, j, k, r, ring_id = 0;
+	unsigned num_compute_rings;
+	int xcc_id = 0;
+	struct amdgpu_device *adev = ip_block->adev;
+	int num_queue_per_pipe = 1; /* we only enable 1 KGQ per pipe */
+
+	INIT_DELAYED_WORK(&adev->gfx.idle_work, amdgpu_gfx_profile_idle_work_handler);
+
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	case IP_VERSION(12, 0, 0):
+	case IP_VERSION(12, 0, 1):
+		adev->gfx.me.num_me = 1;
+		adev->gfx.me.num_pipe_per_me = 1;
+		adev->gfx.me.num_queue_per_pipe = 8;
+		adev->gfx.mec.num_mec = 1;
+		adev->gfx.mec.num_pipe_per_mec = 2;
+		adev->gfx.mec.num_queue_per_pipe = 4;
+		break;
+	default:
+		adev->gfx.me.num_me = 1;
+		adev->gfx.me.num_pipe_per_me = 1;
+		adev->gfx.me.num_queue_per_pipe = 1;
+		adev->gfx.mec.num_mec = 1;
+		adev->gfx.mec.num_pipe_per_mec = 4;
+		adev->gfx.mec.num_queue_per_pipe = 8;
+		break;
+	}
+
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	case IP_VERSION(12, 0, 0):
+	case IP_VERSION(12, 0, 1):
+		if (!adev->gfx.disable_uq &&
+		    adev->gfx.me_fw_version  >= 2780 &&
+		    adev->gfx.pfp_fw_version >= 2840 &&
+		    adev->gfx.mec_fw_version >= 3050 &&
+		    adev->mes.fw_version[0] >= 123) {
+			adev->userq_funcs[AMDGPU_HW_IP_GFX] = &userq_mes_funcs;
+			adev->userq_funcs[AMDGPU_HW_IP_COMPUTE] = &userq_mes_funcs;
+		}
+		break;
+	default:
+		break;
+	}
+
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	case IP_VERSION(12, 0, 0):
+	case IP_VERSION(12, 0, 1):
+		if (adev->gfx.me_fw_version  >= 2480 &&
+		    adev->gfx.pfp_fw_version >= 2530 &&
+		    adev->gfx.mec_fw_version >= 2680 &&
+		    adev->mes.fw_version[0] >= 100)
+			adev->gfx.enable_cleaner_shader = true;
+		break;
+	default:
+		adev->gfx.enable_cleaner_shader = false;
+		break;
+	}
+
+	if (adev->gfx.num_compute_rings) {
+		/* recalculate compute rings to use based on hardware configuration */
+		num_compute_rings = (adev->gfx.mec.num_pipe_per_mec *
+				     adev->gfx.mec.num_queue_per_pipe) / 2;
+		adev->gfx.num_compute_rings = min(adev->gfx.num_compute_rings,
+						  num_compute_rings);
+	}
+
+	/* EOP Event */
+	r = amdgpu_irq_add_id(adev, SOC21_IH_CLIENTID_GRBM_CP,
+			      GFX_12_0_0__SRCID__CP_EOP_INTERRUPT,
+			      &adev->gfx.eop_irq);
+	if (r)
+		return r;
+
+	/* Bad opcode Event */
+	r = amdgpu_irq_add_id(adev, SOC21_IH_CLIENTID_GRBM_CP,
+			      GFX_12_0_0__SRCID__CP_BAD_OPCODE_ERROR,
+			      &adev->gfx.bad_op_irq);
+	if (r)
+		return r;
+
+	/* Privileged reg */
+	r = amdgpu_irq_add_id(adev, SOC21_IH_CLIENTID_GRBM_CP,
+			      GFX_12_0_0__SRCID__CP_PRIV_REG_FAULT,
+			      &adev->gfx.priv_reg_irq);
+	if (r)
+		return r;
+
+	/* Privileged inst */
+	r = amdgpu_irq_add_id(adev, SOC21_IH_CLIENTID_GRBM_CP,
+			      GFX_12_0_0__SRCID__CP_PRIV_INSTR_FAULT,
+			      &adev->gfx.priv_inst_irq);
+	if (r)
+		return r;
+
+	adev->gfx.gfx_current_status = AMDGPU_GFX_NORMAL_MODE;
+
+	gfx_v12_0_me_init(adev);
+
+	r = gfx_v12_0_rlc_init(adev);
+	if (r) {
+		dev_err(adev->dev, "Failed to init rlc BOs!\n");
+		return r;
+	}
+
+	r = gfx_v12_0_mec_init(adev);
+	if (r) {
+		dev_err(adev->dev, "Failed to init MEC BOs!\n");
+		return r;
+	}
+
+	if (adev->gfx.num_gfx_rings) {
+		/* set up the gfx ring */
+		for (i = 0; i < adev->gfx.me.num_me; i++) {
+			for (j = 0; j < num_queue_per_pipe; j++) {
+				for (k = 0; k < adev->gfx.me.num_pipe_per_me; k++) {
+					if (!amdgpu_gfx_is_me_queue_enabled(adev, i, k, j))
+						continue;
+
+					r = gfx_v12_0_gfx_ring_init(adev, ring_id,
+								    i, k, j);
+					if (r)
+						return r;
+					ring_id++;
+				}
+			}
+		}
+	}
+
+	if (adev->gfx.num_compute_rings) {
+		ring_id = 0;
+		/* set up the compute queues - allocate horizontally across pipes */
+		for (i = 0; i < adev->gfx.mec.num_mec; ++i) {
+			for (j = 0; j < adev->gfx.mec.num_queue_per_pipe; j++) {
+				for (k = 0; k < adev->gfx.mec.num_pipe_per_mec; k++) {
+					if (!amdgpu_gfx_is_mec_queue_enabled(adev,
+									     0, i, k, j))
+						continue;
+
+					r = gfx_v12_0_compute_ring_init(adev, ring_id,
+									i, k, j);
+					if (r)
+						return r;
+
+					ring_id++;
+				}
+			}
+		}
+	}
+
+	adev->gfx.gfx_supported_reset =
+		amdgpu_get_soft_full_reset_mask(&adev->gfx.gfx_ring[0]);
+	adev->gfx.compute_supported_reset =
+		amdgpu_get_soft_full_reset_mask(&adev->gfx.compute_ring[0]);
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	case IP_VERSION(12, 0, 0):
+	case IP_VERSION(12, 0, 1):
+		if ((adev->gfx.me_fw_version >= 2660) &&
+		    (adev->gfx.mec_fw_version >= 2920) &&
+		    !amdgpu_sriov_vf(adev) &&
+		    !adev->debug_disable_gpu_ring_reset) {
+			adev->gfx.compute_supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE;
+			adev->gfx.gfx_supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE;
+		}
+		break;
+	default:
+		break;
+	}
+
+	if (!adev->enable_mes_kiq) {
+		r = amdgpu_gfx_kiq_init(adev, GFX12_MEC_HPD_SIZE, 0);
+		if (r) {
+			dev_err(adev->dev, "Failed to init KIQ BOs!\n");
+			return r;
+		}
+
+		r = amdgpu_gfx_kiq_init_ring(adev, xcc_id);
+		if (r)
+			return r;
+	}
+
+	r = amdgpu_gfx_mqd_sw_init(adev, sizeof(struct v12_compute_mqd), 0);
+	if (r)
+		return r;
+
+	/* allocate visible FB for rlc auto-loading fw */
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_RLC_BACKDOOR_AUTO) {
+		r = gfx_v12_0_rlc_autoload_buffer_init(adev);
+		if (r)
+			return r;
+	}
+
+	r = gfx_v12_0_gpu_early_init(adev);
+	if (r)
+		return r;
+
+	gfx_v12_0_alloc_ip_dump(adev);
+
+	r = amdgpu_gfx_sysfs_init(adev);
+	if (r)
+		return r;
+
+	return 0;
+}
+
+static void gfx_v12_0_pfp_fini(struct amdgpu_device *adev)
+{
+	amdgpu_bo_free_kernel(&adev->gfx.pfp.pfp_fw_obj,
+			      &adev->gfx.pfp.pfp_fw_gpu_addr,
+			      (void **)&adev->gfx.pfp.pfp_fw_ptr);
+
+	amdgpu_bo_free_kernel(&adev->gfx.pfp.pfp_fw_data_obj,
+			      &adev->gfx.pfp.pfp_fw_data_gpu_addr,
+			      (void **)&adev->gfx.pfp.pfp_fw_data_ptr);
+}
+
+static void gfx_v12_0_me_fini(struct amdgpu_device *adev)
+{
+	amdgpu_bo_free_kernel(&adev->gfx.me.me_fw_obj,
+			      &adev->gfx.me.me_fw_gpu_addr,
+			      (void **)&adev->gfx.me.me_fw_ptr);
+
+	amdgpu_bo_free_kernel(&adev->gfx.me.me_fw_data_obj,
+			       &adev->gfx.me.me_fw_data_gpu_addr,
+			       (void **)&adev->gfx.me.me_fw_data_ptr);
+}
+
+static void gfx_v12_0_rlc_autoload_buffer_fini(struct amdgpu_device *adev)
+{
+	amdgpu_bo_free_kernel(&adev->gfx.rlc.rlc_autoload_bo,
+			&adev->gfx.rlc.rlc_autoload_gpu_addr,
+			(void **)&adev->gfx.rlc.rlc_autoload_ptr);
+}
+
+static int gfx_v12_0_sw_fini(struct amdgpu_ip_block *ip_block)
+{
+	int i;
+	struct amdgpu_device *adev = ip_block->adev;
+
+	for (i = 0; i < adev->gfx.num_gfx_rings; i++)
+		amdgpu_ring_fini(&adev->gfx.gfx_ring[i]);
+	for (i = 0; i < adev->gfx.num_compute_rings; i++)
+		amdgpu_ring_fini(&adev->gfx.compute_ring[i]);
+
+	amdgpu_gfx_mqd_sw_fini(adev, 0);
+
+	if (!adev->enable_mes_kiq) {
+		amdgpu_gfx_kiq_free_ring(&adev->gfx.kiq[0].ring);
+		amdgpu_gfx_kiq_fini(adev, 0);
+	}
+
+	gfx_v12_0_pfp_fini(adev);
+	gfx_v12_0_me_fini(adev);
+	gfx_v12_0_rlc_fini(adev);
+	gfx_v12_0_mec_fini(adev);
+
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_RLC_BACKDOOR_AUTO)
+		gfx_v12_0_rlc_autoload_buffer_fini(adev);
+
+	gfx_v12_0_free_microcode(adev);
+
+	amdgpu_gfx_sysfs_fini(adev);
+
+	kfree(adev->gfx.ip_dump_core);
+	kfree(adev->gfx.ip_dump_compute_queues);
+	kfree(adev->gfx.ip_dump_gfx_queues);
+
+	return 0;
+}
+
+static void gfx_v12_0_select_se_sh(struct amdgpu_device *adev, u32 se_num,
+				   u32 sh_num, u32 instance, int xcc_id)
+{
+	u32 data;
+
+	if (instance == 0xffffffff)
+		data = REG_SET_FIELD(0, GRBM_GFX_INDEX,
+				     INSTANCE_BROADCAST_WRITES, 1);
+	else
+		data = REG_SET_FIELD(0, GRBM_GFX_INDEX, INSTANCE_INDEX,
+				     instance);
+
+	if (se_num == 0xffffffff)
+		data = REG_SET_FIELD(data, GRBM_GFX_INDEX, SE_BROADCAST_WRITES,
+				     1);
+	else
+		data = REG_SET_FIELD(data, GRBM_GFX_INDEX, SE_INDEX, se_num);
+
+	if (sh_num == 0xffffffff)
+		data = REG_SET_FIELD(data, GRBM_GFX_INDEX, SA_BROADCAST_WRITES,
+				     1);
+	else
+		data = REG_SET_FIELD(data, GRBM_GFX_INDEX, SA_INDEX, sh_num);
+
+	WREG32_SOC15(GC, 0, regGRBM_GFX_INDEX, data);
+}
+
+static u32 gfx_v12_0_get_sa_active_bitmap(struct amdgpu_device *adev)
+{
+	u32 gc_disabled_sa_mask, gc_user_disabled_sa_mask, sa_mask;
+
+	gc_disabled_sa_mask = RREG32_SOC15(GC, 0, regGRBM_CC_GC_SA_UNIT_DISABLE);
+	gc_disabled_sa_mask = REG_GET_FIELD(gc_disabled_sa_mask,
+					    GRBM_CC_GC_SA_UNIT_DISABLE,
+					    SA_DISABLE);
+	gc_user_disabled_sa_mask = RREG32_SOC15(GC, 0, regGRBM_GC_USER_SA_UNIT_DISABLE);
+	gc_user_disabled_sa_mask = REG_GET_FIELD(gc_user_disabled_sa_mask,
+						 GRBM_GC_USER_SA_UNIT_DISABLE,
+						 SA_DISABLE);
+	sa_mask = amdgpu_gfx_create_bitmask(adev->gfx.config.max_sh_per_se *
+					    adev->gfx.config.max_shader_engines);
+
+	return sa_mask & (~(gc_disabled_sa_mask | gc_user_disabled_sa_mask));
+}
+
+static u32 gfx_v12_0_get_rb_active_bitmap(struct amdgpu_device *adev)
+{
+	u32 gc_disabled_rb_mask, gc_user_disabled_rb_mask;
+	u32 rb_mask;
+
+	gc_disabled_rb_mask = RREG32_SOC15(GC, 0, regCC_RB_BACKEND_DISABLE);
+	gc_disabled_rb_mask = REG_GET_FIELD(gc_disabled_rb_mask,
+					    CC_RB_BACKEND_DISABLE,
+					    BACKEND_DISABLE);
+	gc_user_disabled_rb_mask = RREG32_SOC15(GC, 0, regGC_USER_RB_BACKEND_DISABLE);
+	gc_user_disabled_rb_mask = REG_GET_FIELD(gc_user_disabled_rb_mask,
+						 GC_USER_RB_BACKEND_DISABLE,
+						 BACKEND_DISABLE);
+	rb_mask = amdgpu_gfx_create_bitmask(adev->gfx.config.max_backends_per_se *
+					    adev->gfx.config.max_shader_engines);
+
+	return rb_mask & (~(gc_disabled_rb_mask | gc_user_disabled_rb_mask));
+}
+
+static void gfx_v12_0_setup_rb(struct amdgpu_device *adev)
+{
+	u32 rb_bitmap_per_sa;
+	u32 rb_bitmap_width_per_sa;
+	u32 max_sa;
+	u32 active_sa_bitmap;
+	u32 global_active_rb_bitmap;
+	u32 active_rb_bitmap = 0;
+	u32 i;
+
+	/* query sa bitmap from SA_UNIT_DISABLE registers */
+	active_sa_bitmap = gfx_v12_0_get_sa_active_bitmap(adev);
+	/* query rb bitmap from RB_BACKEND_DISABLE registers */
+	global_active_rb_bitmap = gfx_v12_0_get_rb_active_bitmap(adev);
+
+	/* generate active rb bitmap according to active sa bitmap */
+	max_sa = adev->gfx.config.max_shader_engines *
+		 adev->gfx.config.max_sh_per_se;
+	rb_bitmap_width_per_sa = adev->gfx.config.max_backends_per_se /
+				 adev->gfx.config.max_sh_per_se;
+	rb_bitmap_per_sa = amdgpu_gfx_create_bitmask(rb_bitmap_width_per_sa);
+
+	for (i = 0; i < max_sa; i++) {
+		if (active_sa_bitmap & (1 << i))
+			active_rb_bitmap |= (rb_bitmap_per_sa << (i * rb_bitmap_width_per_sa));
+	}
+
+	active_rb_bitmap &= global_active_rb_bitmap;
+	adev->gfx.config.backend_enable_mask = active_rb_bitmap;
+	adev->gfx.config.num_rbs = hweight32(active_rb_bitmap);
+}
+
+#define LDS_APP_BASE           0x1
+#define SCRATCH_APP_BASE       0x2
+
+static void gfx_v12_0_init_compute_vmid(struct amdgpu_device *adev)
+{
+	int i;
+	uint32_t sh_mem_bases;
+	uint32_t data;
+
+	/*
+	 * Configure apertures:
+	 * LDS:         0x60000000'00000000 - 0x60000001'00000000 (4GB)
+	 * Scratch:     0x60000001'00000000 - 0x60000002'00000000 (4GB)
+	 * GPUVM:       0x60010000'00000000 - 0x60020000'00000000 (1TB)
+	 */
+	sh_mem_bases = (LDS_APP_BASE << SH_MEM_BASES__SHARED_BASE__SHIFT) |
+			SCRATCH_APP_BASE;
+
+	mutex_lock(&adev->srbm_mutex);
+	for (i = adev->vm_manager.first_kfd_vmid; i < AMDGPU_NUM_VMID; i++) {
+		soc24_grbm_select(adev, 0, 0, 0, i);
+		/* CP and shaders */
+		WREG32_SOC15(GC, 0, regSH_MEM_CONFIG, DEFAULT_SH_MEM_CONFIG);
+		WREG32_SOC15(GC, 0, regSH_MEM_BASES, sh_mem_bases);
+
+		/* Enable trap for each kfd vmid. */
+		data = RREG32_SOC15(GC, 0, regSPI_GDBG_PER_VMID_CNTL);
+		data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, TRAP_EN, 1);
+		WREG32_SOC15(GC, 0, regSPI_GDBG_PER_VMID_CNTL, data);
+	}
+	soc24_grbm_select(adev, 0, 0, 0, 0);
+	mutex_unlock(&adev->srbm_mutex);
+}
+
+static void gfx_v12_0_tcp_harvest(struct amdgpu_device *adev)
+{
+	/* TODO: harvest feature to be added later. */
+}
+
+static void gfx_v12_0_get_tcc_info(struct amdgpu_device *adev)
+{
+}
+
+static void gfx_v12_0_constants_init(struct amdgpu_device *adev)
+{
+	u32 tmp;
+	int i;
+
+	if (!amdgpu_sriov_vf(adev))
+		WREG32_FIELD15_PREREG(GC, 0, GRBM_CNTL, READ_TIMEOUT, 0xff);
+
+	gfx_v12_0_setup_rb(adev);
+	gfx_v12_0_get_cu_info(adev, &adev->gfx.cu_info);
+	gfx_v12_0_get_tcc_info(adev);
+	adev->gfx.config.pa_sc_tile_steering_override = 0;
+
+	/* XXX SH_MEM regs */
+	/* where to put LDS, scratch, GPUVM in FSA64 space */
+	mutex_lock(&adev->srbm_mutex);
+	for (i = 0; i < adev->vm_manager.id_mgr[AMDGPU_GFXHUB(0)].num_ids; i++) {
+		soc24_grbm_select(adev, 0, 0, 0, i);
+		/* CP and shaders */
+		WREG32_SOC15(GC, 0, regSH_MEM_CONFIG, DEFAULT_SH_MEM_CONFIG);
+		if (i != 0) {
+			tmp = REG_SET_FIELD(0, SH_MEM_BASES, PRIVATE_BASE,
+				(adev->gmc.private_aperture_start >> 48));
+			tmp = REG_SET_FIELD(tmp, SH_MEM_BASES, SHARED_BASE,
+				(adev->gmc.shared_aperture_start >> 48));
+			WREG32_SOC15(GC, 0, regSH_MEM_BASES, tmp);
+		}
+	}
+	soc24_grbm_select(adev, 0, 0, 0, 0);
+
+	mutex_unlock(&adev->srbm_mutex);
+
+	gfx_v12_0_init_compute_vmid(adev);
+}
+
+static u32 gfx_v12_0_get_cpg_int_cntl(struct amdgpu_device *adev,
+				      int me, int pipe)
+{
+	if (me != 0)
+		return 0;
+
+	switch (pipe) {
+	case 0:
+		return SOC15_REG_OFFSET(GC, 0, regCP_INT_CNTL_RING0);
+	default:
+		return 0;
+	}
+}
+
+static u32 gfx_v12_0_get_cpc_int_cntl(struct amdgpu_device *adev,
+				      int me, int pipe)
+{
+	/*
+	 * amdgpu controls only the first MEC. That's why this function only
+	 * handles the setting of interrupts for this specific MEC. All other
+	 * pipes' interrupts are set by amdkfd.
+	 */
+	if (me != 1)
+		return 0;
+
+	switch (pipe) {
+	case 0:
+		return SOC15_REG_OFFSET(GC, 0, regCP_ME1_PIPE0_INT_CNTL);
+	case 1:
+		return SOC15_REG_OFFSET(GC, 0, regCP_ME1_PIPE1_INT_CNTL);
+	default:
+		return 0;
+	}
+}
+
+static void gfx_v12_0_enable_gui_idle_interrupt(struct amdgpu_device *adev,
+					       bool enable)
+{
+	u32 tmp, cp_int_cntl_reg;
+	int i, j;
+
+	if (amdgpu_sriov_vf(adev))
+		return;
+
+	for (i = 0; i < adev->gfx.me.num_me; i++) {
+		for (j = 0; j < adev->gfx.me.num_pipe_per_me; j++) {
+			cp_int_cntl_reg = gfx_v12_0_get_cpg_int_cntl(adev, i, j);
+
+			if (cp_int_cntl_reg) {
+				tmp = RREG32_SOC15_IP(GC, cp_int_cntl_reg);
+				tmp = REG_SET_FIELD(tmp, CP_INT_CNTL_RING0, CNTX_BUSY_INT_ENABLE,
+						    enable ? 1 : 0);
+				tmp = REG_SET_FIELD(tmp, CP_INT_CNTL_RING0, CNTX_EMPTY_INT_ENABLE,
+						    enable ? 1 : 0);
+				tmp = REG_SET_FIELD(tmp, CP_INT_CNTL_RING0, CMP_BUSY_INT_ENABLE,
+						    enable ? 1 : 0);
+				tmp = REG_SET_FIELD(tmp, CP_INT_CNTL_RING0, GFX_IDLE_INT_ENABLE,
+						    enable ? 1 : 0);
+				WREG32_SOC15_IP(GC, cp_int_cntl_reg, tmp);
+			}
+		}
+	}
+}
+
+static int gfx_v12_0_init_csb(struct amdgpu_device *adev)
+{
+	adev->gfx.rlc.funcs->get_csb_buffer(adev, adev->gfx.rlc.cs_ptr);
+
+	WREG32_SOC15(GC, 0, regRLC_CSIB_ADDR_HI,
+			adev->gfx.rlc.clear_state_gpu_addr >> 32);
+	WREG32_SOC15(GC, 0, regRLC_CSIB_ADDR_LO,
+			adev->gfx.rlc.clear_state_gpu_addr & 0xfffffffc);
+	WREG32_SOC15(GC, 0, regRLC_CSIB_LENGTH, adev->gfx.rlc.clear_state_size);
+
+	return 0;
+}
+
+static void gfx_v12_0_rlc_stop(struct amdgpu_device *adev)
+{
+	u32 tmp = RREG32_SOC15(GC, 0, regRLC_CNTL);
+
+	tmp = REG_SET_FIELD(tmp, RLC_CNTL, RLC_ENABLE_F32, 0);
+	WREG32_SOC15(GC, 0, regRLC_CNTL, tmp);
+}
+
+static void gfx_v12_0_rlc_reset(struct amdgpu_device *adev)
+{
+	WREG32_FIELD15_PREREG(GC, 0, GRBM_SOFT_RESET, SOFT_RESET_RLC, 1);
+	udelay(50);
+	WREG32_FIELD15_PREREG(GC, 0, GRBM_SOFT_RESET, SOFT_RESET_RLC, 0);
+	udelay(50);
+}
+
+static void gfx_v12_0_rlc_smu_handshake_cntl(struct amdgpu_device *adev,
+					     bool enable)
+{
+	uint32_t rlc_pg_cntl;
+
+	rlc_pg_cntl = RREG32_SOC15(GC, 0, regRLC_PG_CNTL);
+
+	if (!enable) {
+		/* RLC_PG_CNTL[23] = 0 (default)
+		 * RLC will wait for handshake acks with SMU
+		 * GFXOFF will be enabled
+		 * RLC_PG_CNTL[23] = 1
+		 * RLC will not issue any message to SMU
+		 * hence no handshake between SMU & RLC
+		 * GFXOFF will be disabled
+		 */
+		rlc_pg_cntl |= RLC_PG_CNTL__SMU_HANDSHAKE_DISABLE_MASK;
+	} else
+		rlc_pg_cntl &= ~RLC_PG_CNTL__SMU_HANDSHAKE_DISABLE_MASK;
+	WREG32_SOC15(GC, 0, regRLC_PG_CNTL, rlc_pg_cntl);
+}
+
+static void gfx_v12_0_rlc_start(struct amdgpu_device *adev)
+{
+	/* TODO: enable rlc & smu handshake until smu
+	 * and gfxoff feature works as expected */
+	if (!(amdgpu_pp_feature_mask & PP_GFXOFF_MASK))
+		gfx_v12_0_rlc_smu_handshake_cntl(adev, false);
+
+	WREG32_FIELD15_PREREG(GC, 0, RLC_CNTL, RLC_ENABLE_F32, 1);
+	udelay(50);
+}
+
+static void gfx_v12_0_rlc_enable_srm(struct amdgpu_device *adev)
+{
+	uint32_t tmp;
+
+	/* enable Save Restore Machine */
+	tmp = RREG32(SOC15_REG_OFFSET(GC, 0, regRLC_SRM_CNTL));
+	tmp |= RLC_SRM_CNTL__AUTO_INCR_ADDR_MASK;
+	tmp |= RLC_SRM_CNTL__SRM_ENABLE_MASK;
+	WREG32(SOC15_REG_OFFSET(GC, 0, regRLC_SRM_CNTL), tmp);
+}
+
+static void gfx_v12_0_load_rlcg_microcode(struct amdgpu_device *adev)
+{
+	const struct rlc_firmware_header_v2_0 *hdr;
+	const __le32 *fw_data;
+	unsigned i, fw_size;
+
+	hdr = (const struct rlc_firmware_header_v2_0 *)adev->gfx.rlc_fw->data;
+	fw_data = (const __le32 *)(adev->gfx.rlc_fw->data +
+			   le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+	fw_size = le32_to_cpu(hdr->header.ucode_size_bytes) / 4;
+
+	WREG32_SOC15(GC, 0, regRLC_GPM_UCODE_ADDR,
+		     RLCG_UCODE_LOADING_START_ADDRESS);
+
+	for (i = 0; i < fw_size; i++)
+		WREG32_SOC15(GC, 0, regRLC_GPM_UCODE_DATA,
+			     le32_to_cpup(fw_data++));
+
+	WREG32_SOC15(GC, 0, regRLC_GPM_UCODE_ADDR, adev->gfx.rlc_fw_version);
+}
+
+static void gfx_v12_0_load_rlc_iram_dram_microcode(struct amdgpu_device *adev)
+{
+	const struct rlc_firmware_header_v2_2 *hdr;
+	const __le32 *fw_data;
+	unsigned i, fw_size;
+	u32 tmp;
+
+	hdr = (const struct rlc_firmware_header_v2_2 *)adev->gfx.rlc_fw->data;
+
+	fw_data = (const __le32 *)(adev->gfx.rlc_fw->data +
+			le32_to_cpu(hdr->rlc_iram_ucode_offset_bytes));
+	fw_size = le32_to_cpu(hdr->rlc_iram_ucode_size_bytes) / 4;
+
+	WREG32_SOC15(GC, 0, regRLC_LX6_IRAM_ADDR, 0);
+
+	for (i = 0; i < fw_size; i++) {
+		if ((amdgpu_emu_mode == 1) && (i % 100 == 99))
+			msleep(1);
+		WREG32_SOC15(GC, 0, regRLC_LX6_IRAM_DATA,
+				le32_to_cpup(fw_data++));
+	}
+
+	WREG32_SOC15(GC, 0, regRLC_LX6_IRAM_ADDR, adev->gfx.rlc_fw_version);
+
+	fw_data = (const __le32 *)(adev->gfx.rlc_fw->data +
+			le32_to_cpu(hdr->rlc_dram_ucode_offset_bytes));
+	fw_size = le32_to_cpu(hdr->rlc_dram_ucode_size_bytes) / 4;
+
+	WREG32_SOC15(GC, 0, regRLC_LX6_DRAM_ADDR, 0);
+	for (i = 0; i < fw_size; i++) {
+		if ((amdgpu_emu_mode == 1) && (i % 100 == 99))
+			msleep(1);
+		WREG32_SOC15(GC, 0, regRLC_LX6_DRAM_DATA,
+				le32_to_cpup(fw_data++));
+	}
+
+	WREG32_SOC15(GC, 0, regRLC_LX6_IRAM_ADDR, adev->gfx.rlc_fw_version);
+
+	tmp = RREG32_SOC15(GC, 0, regRLC_LX6_CNTL);
+	tmp = REG_SET_FIELD(tmp, RLC_LX6_CNTL, PDEBUG_ENABLE, 1);
+	tmp = REG_SET_FIELD(tmp, RLC_LX6_CNTL, BRESET, 0);
+	WREG32_SOC15(GC, 0, regRLC_LX6_CNTL, tmp);
+}
+
+static int gfx_v12_0_rlc_load_microcode(struct amdgpu_device *adev)
+{
+	const struct rlc_firmware_header_v2_0 *hdr;
+	uint16_t version_major;
+	uint16_t version_minor;
+
+	if (!adev->gfx.rlc_fw)
+		return -EINVAL;
+
+	hdr = (const struct rlc_firmware_header_v2_0 *)adev->gfx.rlc_fw->data;
+	amdgpu_ucode_print_rlc_hdr(&hdr->header);
+
+	version_major = le16_to_cpu(hdr->header.header_version_major);
+	version_minor = le16_to_cpu(hdr->header.header_version_minor);
+
+	if (version_major == 2) {
+		gfx_v12_0_load_rlcg_microcode(adev);
+		if (amdgpu_dpm == 1) {
+			if (version_minor >= 2)
+				gfx_v12_0_load_rlc_iram_dram_microcode(adev);
+		}
+
+		return 0;
+	}
+
+	return -EINVAL;
+}
+
+static int gfx_v12_0_rlc_resume(struct amdgpu_device *adev)
+{
+	int r;
+
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
+		gfx_v12_0_init_csb(adev);
+
+		if (!amdgpu_sriov_vf(adev)) /* enable RLC SRM */
+			gfx_v12_0_rlc_enable_srm(adev);
+	} else {
+		if (amdgpu_sriov_vf(adev)) {
+			gfx_v12_0_init_csb(adev);
+			return 0;
+		}
+
+		adev->gfx.rlc.funcs->stop(adev);
+
+		/* disable CG */
+		WREG32_SOC15(GC, 0, regRLC_CGCG_CGLS_CTRL, 0);
+
+		/* disable PG */
+		WREG32_SOC15(GC, 0, regRLC_PG_CNTL, 0);
+
+		if (adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT) {
+			/* legacy rlc firmware loading */
+			r = gfx_v12_0_rlc_load_microcode(adev);
+			if (r)
+				return r;
+		}
+
+		gfx_v12_0_init_csb(adev);
+
+		adev->gfx.rlc.funcs->start(adev);
+	}
+
+	return 0;
+}
+
+static void gfx_v12_0_config_gfx_rs64(struct amdgpu_device *adev)
+{
+	const struct gfx_firmware_header_v2_0 *pfp_hdr;
+	const struct gfx_firmware_header_v2_0 *me_hdr;
+	const struct gfx_firmware_header_v2_0 *mec_hdr;
+	uint32_t pipe_id, tmp;
+
+	mec_hdr = (const struct gfx_firmware_header_v2_0 *)
+		adev->gfx.mec_fw->data;
+	me_hdr = (const struct gfx_firmware_header_v2_0 *)
+		adev->gfx.me_fw->data;
+	pfp_hdr = (const struct gfx_firmware_header_v2_0 *)
+		adev->gfx.pfp_fw->data;
+
+	/* config pfp program start addr */
+	for (pipe_id = 0; pipe_id < 2; pipe_id++) {
+		soc24_grbm_select(adev, 0, pipe_id, 0, 0);
+		WREG32_SOC15(GC, 0, regCP_PFP_PRGRM_CNTR_START,
+			(pfp_hdr->ucode_start_addr_hi << 30) |
+			(pfp_hdr->ucode_start_addr_lo >> 2));
+		WREG32_SOC15(GC, 0, regCP_PFP_PRGRM_CNTR_START_HI,
+			pfp_hdr->ucode_start_addr_hi >> 2);
+	}
+	soc24_grbm_select(adev, 0, 0, 0, 0);
+
+	/* reset pfp pipe */
+	tmp = RREG32_SOC15(GC, 0, regCP_ME_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, PFP_PIPE0_RESET, 1);
+	tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, PFP_PIPE1_RESET, 1);
+	WREG32_SOC15(GC, 0, regCP_ME_CNTL, tmp);
+
+	/* clear pfp pipe reset */
+	tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, PFP_PIPE0_RESET, 0);
+	tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, PFP_PIPE1_RESET, 0);
+	WREG32_SOC15(GC, 0, regCP_ME_CNTL, tmp);
+
+	/* config me program start addr */
+	for (pipe_id = 0; pipe_id < 2; pipe_id++) {
+		soc24_grbm_select(adev, 0, pipe_id, 0, 0);
+		WREG32_SOC15(GC, 0, regCP_ME_PRGRM_CNTR_START,
+			(me_hdr->ucode_start_addr_hi << 30) |
+			(me_hdr->ucode_start_addr_lo >> 2));
+		WREG32_SOC15(GC, 0, regCP_ME_PRGRM_CNTR_START_HI,
+			me_hdr->ucode_start_addr_hi>>2);
+	}
+	soc24_grbm_select(adev, 0, 0, 0, 0);
+
+	/* reset me pipe */
+	tmp = RREG32_SOC15(GC, 0, regCP_ME_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, ME_PIPE0_RESET, 1);
+	tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, ME_PIPE1_RESET, 1);
+	WREG32_SOC15(GC, 0, regCP_ME_CNTL, tmp);
+
+	/* clear me pipe reset */
+	tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, ME_PIPE0_RESET, 0);
+	tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, ME_PIPE1_RESET, 0);
+	WREG32_SOC15(GC, 0, regCP_ME_CNTL, tmp);
+
+	/* config mec program start addr */
+	for (pipe_id = 0; pipe_id < 4; pipe_id++) {
+		soc24_grbm_select(adev, 1, pipe_id, 0, 0);
+		WREG32_SOC15(GC, 0, regCP_MEC_RS64_PRGRM_CNTR_START,
+					mec_hdr->ucode_start_addr_lo >> 2 |
+					mec_hdr->ucode_start_addr_hi << 30);
+		WREG32_SOC15(GC, 0, regCP_MEC_RS64_PRGRM_CNTR_START_HI,
+					mec_hdr->ucode_start_addr_hi >> 2);
+	}
+	soc24_grbm_select(adev, 0, 0, 0, 0);
+
+	/* reset mec pipe */
+	tmp = RREG32_SOC15(GC, 0, regCP_MEC_RS64_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_MEC_RS64_CNTL, MEC_PIPE0_RESET, 1);
+	tmp = REG_SET_FIELD(tmp, CP_MEC_RS64_CNTL, MEC_PIPE1_RESET, 1);
+	tmp = REG_SET_FIELD(tmp, CP_MEC_RS64_CNTL, MEC_PIPE2_RESET, 1);
+	tmp = REG_SET_FIELD(tmp, CP_MEC_RS64_CNTL, MEC_PIPE3_RESET, 1);
+	WREG32_SOC15(GC, 0, regCP_MEC_RS64_CNTL, tmp);
+
+	/* clear mec pipe reset */
+	tmp = REG_SET_FIELD(tmp, CP_MEC_RS64_CNTL, MEC_PIPE0_RESET, 0);
+	tmp = REG_SET_FIELD(tmp, CP_MEC_RS64_CNTL, MEC_PIPE1_RESET, 0);
+	tmp = REG_SET_FIELD(tmp, CP_MEC_RS64_CNTL, MEC_PIPE2_RESET, 0);
+	tmp = REG_SET_FIELD(tmp, CP_MEC_RS64_CNTL, MEC_PIPE3_RESET, 0);
+	WREG32_SOC15(GC, 0, regCP_MEC_RS64_CNTL, tmp);
+}
+
+static void gfx_v12_0_set_pfp_ucode_start_addr(struct amdgpu_device *adev)
+{
+	const struct gfx_firmware_header_v2_0 *cp_hdr;
+	unsigned pipe_id, tmp;
+
+	cp_hdr = (const struct gfx_firmware_header_v2_0 *)
+		adev->gfx.pfp_fw->data;
+	mutex_lock(&adev->srbm_mutex);
+	for (pipe_id = 0; pipe_id < adev->gfx.me.num_pipe_per_me; pipe_id++) {
+		soc24_grbm_select(adev, 0, pipe_id, 0, 0);
+		WREG32_SOC15(GC, 0, regCP_PFP_PRGRM_CNTR_START,
+			     (cp_hdr->ucode_start_addr_hi << 30) |
+			     (cp_hdr->ucode_start_addr_lo >> 2));
+		WREG32_SOC15(GC, 0, regCP_PFP_PRGRM_CNTR_START_HI,
+			     cp_hdr->ucode_start_addr_hi>>2);
+
+		/*
+		 * Program CP_ME_CNTL to reset given PIPE to take
+		 * effect of CP_PFP_PRGRM_CNTR_START.
+		 */
+		tmp = RREG32_SOC15(GC, 0, regCP_ME_CNTL);
+		if (pipe_id == 0)
+			tmp = REG_SET_FIELD(tmp, CP_ME_CNTL,
+					PFP_PIPE0_RESET, 1);
+		else
+			tmp = REG_SET_FIELD(tmp, CP_ME_CNTL,
+					PFP_PIPE1_RESET, 1);
+		WREG32_SOC15(GC, 0, regCP_ME_CNTL, tmp);
+
+		/* Clear pfp pipe0 reset bit. */
+		if (pipe_id == 0)
+			tmp = REG_SET_FIELD(tmp, CP_ME_CNTL,
+					PFP_PIPE0_RESET, 0);
+		else
+			tmp = REG_SET_FIELD(tmp, CP_ME_CNTL,
+					PFP_PIPE1_RESET, 0);
+		WREG32_SOC15(GC, 0, regCP_ME_CNTL, tmp);
+	}
+	soc24_grbm_select(adev, 0, 0, 0, 0);
+	mutex_unlock(&adev->srbm_mutex);
+}
+
+static void gfx_v12_0_set_me_ucode_start_addr(struct amdgpu_device *adev)
+{
+	const struct gfx_firmware_header_v2_0 *cp_hdr;
+	unsigned pipe_id, tmp;
+
+	cp_hdr = (const struct gfx_firmware_header_v2_0 *)
+		adev->gfx.me_fw->data;
+	mutex_lock(&adev->srbm_mutex);
+	for (pipe_id = 0; pipe_id < adev->gfx.me.num_pipe_per_me; pipe_id++) {
+		soc24_grbm_select(adev, 0, pipe_id, 0, 0);
+		WREG32_SOC15(GC, 0, regCP_ME_PRGRM_CNTR_START,
+			     (cp_hdr->ucode_start_addr_hi << 30) |
+			     (cp_hdr->ucode_start_addr_lo >> 2) );
+		WREG32_SOC15(GC, 0, regCP_ME_PRGRM_CNTR_START_HI,
+			     cp_hdr->ucode_start_addr_hi>>2);
+
+		/*
+		 * Program CP_ME_CNTL to reset given PIPE to take
+		 * effect of CP_ME_PRGRM_CNTR_START.
+		 */
+		tmp = RREG32_SOC15(GC, 0, regCP_ME_CNTL);
+		if (pipe_id == 0)
+			tmp = REG_SET_FIELD(tmp, CP_ME_CNTL,
+					ME_PIPE0_RESET, 1);
+		else
+			tmp = REG_SET_FIELD(tmp, CP_ME_CNTL,
+					ME_PIPE1_RESET, 1);
+		WREG32_SOC15(GC, 0, regCP_ME_CNTL, tmp);
+
+		/* Clear pfp pipe0 reset bit. */
+		if (pipe_id == 0)
+			tmp = REG_SET_FIELD(tmp, CP_ME_CNTL,
+					ME_PIPE0_RESET, 0);
+		else
+			tmp = REG_SET_FIELD(tmp, CP_ME_CNTL,
+					ME_PIPE1_RESET, 0);
+		WREG32_SOC15(GC, 0, regCP_ME_CNTL, tmp);
+	}
+	soc24_grbm_select(adev, 0, 0, 0, 0);
+	mutex_unlock(&adev->srbm_mutex);
+}
+
+static void gfx_v12_0_set_mec_ucode_start_addr(struct amdgpu_device *adev)
+{
+	const struct gfx_firmware_header_v2_0 *cp_hdr;
+	unsigned pipe_id;
+
+	cp_hdr = (const struct gfx_firmware_header_v2_0 *)
+		adev->gfx.mec_fw->data;
+	mutex_lock(&adev->srbm_mutex);
+	for (pipe_id = 0; pipe_id < adev->gfx.mec.num_pipe_per_mec; pipe_id++) {
+		soc24_grbm_select(adev, 1, pipe_id, 0, 0);
+		WREG32_SOC15(GC, 0, regCP_MEC_RS64_PRGRM_CNTR_START,
+			     cp_hdr->ucode_start_addr_lo >> 2 |
+			     cp_hdr->ucode_start_addr_hi << 30);
+		WREG32_SOC15(GC, 0, regCP_MEC_RS64_PRGRM_CNTR_START_HI,
+			     cp_hdr->ucode_start_addr_hi >> 2);
+	}
+	soc24_grbm_select(adev, 0, 0, 0, 0);
+	mutex_unlock(&adev->srbm_mutex);
+}
+
+static int gfx_v12_0_wait_for_rlc_autoload_complete(struct amdgpu_device *adev)
+{
+	uint32_t cp_status;
+	uint32_t bootload_status;
+	int i;
+
+	for (i = 0; i < adev->usec_timeout; i++) {
+		cp_status = RREG32_SOC15(GC, 0, regCP_STAT);
+		bootload_status = RREG32_SOC15(GC, 0, regRLC_RLCS_BOOTLOAD_STATUS);
+
+		if ((cp_status == 0) &&
+		    (REG_GET_FIELD(bootload_status,
+			RLC_RLCS_BOOTLOAD_STATUS, BOOTLOAD_COMPLETE) == 1)) {
+			break;
+		}
+		udelay(1);
+		if (amdgpu_emu_mode)
+			msleep(10);
+	}
+
+	if (i >= adev->usec_timeout) {
+		dev_err(adev->dev, "rlc autoload: gc ucode autoload timeout\n");
+		return -ETIMEDOUT;
+	}
+
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_RLC_BACKDOOR_AUTO) {
+		gfx_v12_0_set_pfp_ucode_start_addr(adev);
+		gfx_v12_0_set_me_ucode_start_addr(adev);
+		gfx_v12_0_set_mec_ucode_start_addr(adev);
+	}
+
+	return 0;
+}
+
+static int gfx_v12_0_cp_gfx_enable(struct amdgpu_device *adev, bool enable)
+{
+	int i;
+	u32 tmp = RREG32_SOC15(GC, 0, regCP_ME_CNTL);
+
+	tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, ME_HALT, enable ? 0 : 1);
+	tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, PFP_HALT, enable ? 0 : 1);
+	WREG32_SOC15(GC, 0, regCP_ME_CNTL, tmp);
+
+	for (i = 0; i < adev->usec_timeout; i++) {
+		if (RREG32_SOC15(GC, 0, regCP_STAT) == 0)
+			break;
+		udelay(1);
+	}
+
+	if (i >= adev->usec_timeout)
+		DRM_ERROR("failed to %s cp gfx\n", enable ? "unhalt" : "halt");
+
+	return 0;
+}
+
+static int gfx_v12_0_cp_gfx_load_pfp_microcode_rs64(struct amdgpu_device *adev)
+{
+	int r;
+	const struct gfx_firmware_header_v2_0 *pfp_hdr;
+	const __le32 *fw_ucode, *fw_data;
+	unsigned i, pipe_id, fw_ucode_size, fw_data_size;
+	uint32_t tmp;
+	uint32_t usec_timeout = 50000;  /* wait for 50ms */
+
+	pfp_hdr = (const struct gfx_firmware_header_v2_0 *)
+		adev->gfx.pfp_fw->data;
+
+	amdgpu_ucode_print_gfx_hdr(&pfp_hdr->header);
+
+	/* instruction */
+	fw_ucode = (const __le32 *)(adev->gfx.pfp_fw->data +
+		le32_to_cpu(pfp_hdr->ucode_offset_bytes));
+	fw_ucode_size = le32_to_cpu(pfp_hdr->ucode_size_bytes);
+	/* data */
+	fw_data = (const __le32 *)(adev->gfx.pfp_fw->data +
+		le32_to_cpu(pfp_hdr->data_offset_bytes));
+	fw_data_size = le32_to_cpu(pfp_hdr->data_size_bytes);
+
+	/* 64kb align */
+	r = amdgpu_bo_create_reserved(adev, fw_ucode_size,
+				      64 * 1024, AMDGPU_GEM_DOMAIN_VRAM,
+				      &adev->gfx.pfp.pfp_fw_obj,
+				      &adev->gfx.pfp.pfp_fw_gpu_addr,
+				      (void **)&adev->gfx.pfp.pfp_fw_ptr);
+	if (r) {
+		dev_err(adev->dev, "(%d) failed to create pfp ucode fw bo\n", r);
+		gfx_v12_0_pfp_fini(adev);
+		return r;
+	}
+
+	r = amdgpu_bo_create_reserved(adev, fw_data_size,
+				      64 * 1024, AMDGPU_GEM_DOMAIN_VRAM,
+				      &adev->gfx.pfp.pfp_fw_data_obj,
+				      &adev->gfx.pfp.pfp_fw_data_gpu_addr,
+				      (void **)&adev->gfx.pfp.pfp_fw_data_ptr);
+	if (r) {
+		dev_err(adev->dev, "(%d) failed to create pfp data fw bo\n", r);
+		gfx_v12_0_pfp_fini(adev);
+		return r;
+	}
+
+	memcpy(adev->gfx.pfp.pfp_fw_ptr, fw_ucode, fw_ucode_size);
+	memcpy(adev->gfx.pfp.pfp_fw_data_ptr, fw_data, fw_data_size);
+
+	amdgpu_bo_kunmap(adev->gfx.pfp.pfp_fw_obj);
+	amdgpu_bo_kunmap(adev->gfx.pfp.pfp_fw_data_obj);
+	amdgpu_bo_unreserve(adev->gfx.pfp.pfp_fw_obj);
+	amdgpu_bo_unreserve(adev->gfx.pfp.pfp_fw_data_obj);
+
+	if (amdgpu_emu_mode == 1)
+		amdgpu_device_flush_hdp(adev, NULL);
+
+	WREG32_SOC15(GC, 0, regCP_PFP_IC_BASE_LO,
+		lower_32_bits(adev->gfx.pfp.pfp_fw_gpu_addr));
+	WREG32_SOC15(GC, 0, regCP_PFP_IC_BASE_HI,
+		upper_32_bits(adev->gfx.pfp.pfp_fw_gpu_addr));
+
+	tmp = RREG32_SOC15(GC, 0, regCP_PFP_IC_BASE_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_PFP_IC_BASE_CNTL, VMID, 0);
+	tmp = REG_SET_FIELD(tmp, CP_PFP_IC_BASE_CNTL, CACHE_POLICY, 0);
+	tmp = REG_SET_FIELD(tmp, CP_PFP_IC_BASE_CNTL, EXE_DISABLE, 0);
+	WREG32_SOC15(GC, 0, regCP_PFP_IC_BASE_CNTL, tmp);
+
+	/*
+	 * Programming any of the CP_PFP_IC_BASE registers
+	 * forces invalidation of the ME L1 I$. Wait for the
+	 * invalidation complete
+	 */
+	for (i = 0; i < usec_timeout; i++) {
+		tmp = RREG32_SOC15(GC, 0, regCP_PFP_IC_OP_CNTL);
+		if (1 == REG_GET_FIELD(tmp, CP_PFP_IC_OP_CNTL,
+			INVALIDATE_CACHE_COMPLETE))
+			break;
+		udelay(1);
+	}
+
+	if (i >= usec_timeout) {
+		dev_err(adev->dev, "failed to invalidate instruction cache\n");
+		return -EINVAL;
+	}
+
+	/* Prime the L1 instruction caches */
+	tmp = RREG32_SOC15(GC, 0, regCP_PFP_IC_OP_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_PFP_IC_OP_CNTL, PRIME_ICACHE, 1);
+	WREG32_SOC15(GC, 0, regCP_PFP_IC_OP_CNTL, tmp);
+	/* Waiting for cache primed*/
+	for (i = 0; i < usec_timeout; i++) {
+		tmp = RREG32_SOC15(GC, 0, regCP_PFP_IC_OP_CNTL);
+		if (1 == REG_GET_FIELD(tmp, CP_PFP_IC_OP_CNTL,
+			ICACHE_PRIMED))
+			break;
+		udelay(1);
+	}
+
+	if (i >= usec_timeout) {
+		dev_err(adev->dev, "failed to prime instruction cache\n");
+		return -EINVAL;
+	}
+
+	mutex_lock(&adev->srbm_mutex);
+	for (pipe_id = 0; pipe_id < adev->gfx.me.num_pipe_per_me; pipe_id++) {
+		soc24_grbm_select(adev, 0, pipe_id, 0, 0);
+
+		WREG32_SOC15(GC, 0, regCP_GFX_RS64_DC_BASE0_LO,
+			lower_32_bits(adev->gfx.pfp.pfp_fw_data_gpu_addr));
+		WREG32_SOC15(GC, 0, regCP_GFX_RS64_DC_BASE0_HI,
+			upper_32_bits(adev->gfx.pfp.pfp_fw_data_gpu_addr));
+	}
+	soc24_grbm_select(adev, 0, 0, 0, 0);
+	mutex_unlock(&adev->srbm_mutex);
+
+	tmp = RREG32_SOC15(GC, 0, regCP_GFX_RS64_DC_BASE_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_GFX_RS64_DC_BASE_CNTL, VMID, 0);
+	tmp = REG_SET_FIELD(tmp, CP_GFX_RS64_DC_BASE_CNTL, CACHE_POLICY, 0);
+	WREG32_SOC15(GC, 0, regCP_GFX_RS64_DC_BASE_CNTL, tmp);
+
+	/* Invalidate the data caches */
+	tmp = RREG32_SOC15(GC, 0, regCP_GFX_RS64_DC_OP_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_GFX_RS64_DC_OP_CNTL, INVALIDATE_DCACHE, 1);
+	WREG32_SOC15(GC, 0, regCP_GFX_RS64_DC_OP_CNTL, tmp);
+
+	for (i = 0; i < usec_timeout; i++) {
+		tmp = RREG32_SOC15(GC, 0, regCP_GFX_RS64_DC_OP_CNTL);
+		if (1 == REG_GET_FIELD(tmp, CP_GFX_RS64_DC_OP_CNTL,
+			INVALIDATE_DCACHE_COMPLETE))
+			break;
+		udelay(1);
+	}
+
+	if (i >= usec_timeout) {
+		dev_err(adev->dev, "failed to invalidate RS64 data cache\n");
+		return -EINVAL;
+	}
+
+	gfx_v12_0_set_pfp_ucode_start_addr(adev);
+
+	return 0;
+}
+
+static int gfx_v12_0_cp_gfx_load_me_microcode_rs64(struct amdgpu_device *adev)
+{
+	int r;
+	const struct gfx_firmware_header_v2_0 *me_hdr;
+	const __le32 *fw_ucode, *fw_data;
+	unsigned i, pipe_id, fw_ucode_size, fw_data_size;
+	uint32_t tmp;
+	uint32_t usec_timeout = 50000;  /* wait for 50ms */
+
+	me_hdr = (const struct gfx_firmware_header_v2_0 *)
+		adev->gfx.me_fw->data;
+
+	amdgpu_ucode_print_gfx_hdr(&me_hdr->header);
+
+	/* instruction */
+	fw_ucode = (const __le32 *)(adev->gfx.me_fw->data +
+		le32_to_cpu(me_hdr->ucode_offset_bytes));
+	fw_ucode_size = le32_to_cpu(me_hdr->ucode_size_bytes);
+	/* data */
+	fw_data = (const __le32 *)(adev->gfx.me_fw->data +
+		le32_to_cpu(me_hdr->data_offset_bytes));
+	fw_data_size = le32_to_cpu(me_hdr->data_size_bytes);
+
+	/* 64kb align*/
+	r = amdgpu_bo_create_reserved(adev, fw_ucode_size,
+				      64 * 1024, AMDGPU_GEM_DOMAIN_VRAM,
+				      &adev->gfx.me.me_fw_obj,
+				      &adev->gfx.me.me_fw_gpu_addr,
+				      (void **)&adev->gfx.me.me_fw_ptr);
+	if (r) {
+		dev_err(adev->dev, "(%d) failed to create me ucode bo\n", r);
+		gfx_v12_0_me_fini(adev);
+		return r;
+	}
+
+	r = amdgpu_bo_create_reserved(adev, fw_data_size,
+				      64 * 1024, AMDGPU_GEM_DOMAIN_VRAM,
+				      &adev->gfx.me.me_fw_data_obj,
+				      &adev->gfx.me.me_fw_data_gpu_addr,
+				      (void **)&adev->gfx.me.me_fw_data_ptr);
+	if (r) {
+		dev_err(adev->dev, "(%d) failed to create me data bo\n", r);
+		gfx_v12_0_me_fini(adev);
+		return r;
+	}
+
+	memcpy(adev->gfx.me.me_fw_ptr, fw_ucode, fw_ucode_size);
+	memcpy(adev->gfx.me.me_fw_data_ptr, fw_data, fw_data_size);
+
+	amdgpu_bo_kunmap(adev->gfx.me.me_fw_obj);
+	amdgpu_bo_kunmap(adev->gfx.me.me_fw_data_obj);
+	amdgpu_bo_unreserve(adev->gfx.me.me_fw_obj);
+	amdgpu_bo_unreserve(adev->gfx.me.me_fw_data_obj);
+
+	if (amdgpu_emu_mode == 1)
+		amdgpu_device_flush_hdp(adev, NULL);
+
+	WREG32_SOC15(GC, 0, regCP_ME_IC_BASE_LO,
+		lower_32_bits(adev->gfx.me.me_fw_gpu_addr));
+	WREG32_SOC15(GC, 0, regCP_ME_IC_BASE_HI,
+		upper_32_bits(adev->gfx.me.me_fw_gpu_addr));
+
+	tmp = RREG32_SOC15(GC, 0, regCP_ME_IC_BASE_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_ME_IC_BASE_CNTL, VMID, 0);
+	tmp = REG_SET_FIELD(tmp, CP_ME_IC_BASE_CNTL, CACHE_POLICY, 0);
+	tmp = REG_SET_FIELD(tmp, CP_ME_IC_BASE_CNTL, EXE_DISABLE, 0);
+	WREG32_SOC15(GC, 0, regCP_ME_IC_BASE_CNTL, tmp);
+
+	/*
+	 * Programming any of the CP_ME_IC_BASE registers
+	 * forces invalidation of the ME L1 I$. Wait for the
+	 * invalidation complete
+	 */
+	for (i = 0; i < usec_timeout; i++) {
+		tmp = RREG32_SOC15(GC, 0, regCP_ME_IC_OP_CNTL);
+		if (1 == REG_GET_FIELD(tmp, CP_ME_IC_OP_CNTL,
+			INVALIDATE_CACHE_COMPLETE))
+			break;
+		udelay(1);
+	}
+
+	if (i >= usec_timeout) {
+		dev_err(adev->dev, "failed to invalidate instruction cache\n");
+		return -EINVAL;
+	}
+
+	/* Prime the instruction caches */
+	tmp = RREG32_SOC15(GC, 0, regCP_ME_IC_OP_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_ME_IC_OP_CNTL, PRIME_ICACHE, 1);
+	WREG32_SOC15(GC, 0, regCP_ME_IC_OP_CNTL, tmp);
+
+	/* Waiting for instruction cache primed*/
+	for (i = 0; i < usec_timeout; i++) {
+		tmp = RREG32_SOC15(GC, 0, regCP_ME_IC_OP_CNTL);
+		if (1 == REG_GET_FIELD(tmp, CP_ME_IC_OP_CNTL,
+			ICACHE_PRIMED))
+			break;
+		udelay(1);
+	}
+
+	if (i >= usec_timeout) {
+		dev_err(adev->dev, "failed to prime instruction cache\n");
+		return -EINVAL;
+	}
+
+	mutex_lock(&adev->srbm_mutex);
+	for (pipe_id = 0; pipe_id < adev->gfx.me.num_pipe_per_me; pipe_id++) {
+		soc24_grbm_select(adev, 0, pipe_id, 0, 0);
+
+		WREG32_SOC15(GC, 0, regCP_GFX_RS64_DC_BASE1_LO,
+			lower_32_bits(adev->gfx.me.me_fw_data_gpu_addr));
+		WREG32_SOC15(GC, 0, regCP_GFX_RS64_DC_BASE1_HI,
+			upper_32_bits(adev->gfx.me.me_fw_data_gpu_addr));
+	}
+	soc24_grbm_select(adev, 0, 0, 0, 0);
+	mutex_unlock(&adev->srbm_mutex);
+
+	tmp = RREG32_SOC15(GC, 0, regCP_GFX_RS64_DC_BASE_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_GFX_RS64_DC_BASE_CNTL, VMID, 0);
+	tmp = REG_SET_FIELD(tmp, CP_GFX_RS64_DC_BASE_CNTL, CACHE_POLICY, 0);
+	WREG32_SOC15(GC, 0, regCP_GFX_RS64_DC_BASE_CNTL, tmp);
+
+	/* Invalidate the data caches */
+	tmp = RREG32_SOC15(GC, 0, regCP_GFX_RS64_DC_OP_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_GFX_RS64_DC_OP_CNTL, INVALIDATE_DCACHE, 1);
+	WREG32_SOC15(GC, 0, regCP_GFX_RS64_DC_OP_CNTL, tmp);
+
+	for (i = 0; i < usec_timeout; i++) {
+		tmp = RREG32_SOC15(GC, 0, regCP_GFX_RS64_DC_OP_CNTL);
+		if (1 == REG_GET_FIELD(tmp, CP_GFX_RS64_DC_OP_CNTL,
+			INVALIDATE_DCACHE_COMPLETE))
+			break;
+		udelay(1);
+	}
+
+	if (i >= usec_timeout) {
+		dev_err(adev->dev, "failed to invalidate RS64 data cache\n");
+		return -EINVAL;
+	}
+
+	gfx_v12_0_set_me_ucode_start_addr(adev);
+
+	return 0;
+}
+
+static int gfx_v12_0_cp_gfx_load_microcode(struct amdgpu_device *adev)
+{
+	int r;
+
+	if (!adev->gfx.me_fw || !adev->gfx.pfp_fw)
+		return -EINVAL;
+
+	gfx_v12_0_cp_gfx_enable(adev, false);
+
+	r = gfx_v12_0_cp_gfx_load_pfp_microcode_rs64(adev);
+	if (r) {
+		dev_err(adev->dev, "(%d) failed to load pfp fw\n", r);
+		return r;
+	}
+
+	r = gfx_v12_0_cp_gfx_load_me_microcode_rs64(adev);
+	if (r) {
+		dev_err(adev->dev, "(%d) failed to load me fw\n", r);
+		return r;
+	}
+
+	return 0;
+}
+
+static int gfx_v12_0_cp_gfx_start(struct amdgpu_device *adev)
+{
+	/* init the CP */
+	WREG32_SOC15(GC, 0, regCP_MAX_CONTEXT,
+		     adev->gfx.config.max_hw_contexts - 1);
+	WREG32_SOC15(GC, 0, regCP_DEVICE_ID, 1);
+
+	if (!amdgpu_async_gfx_ring)
+		gfx_v12_0_cp_gfx_enable(adev, true);
+
+	return 0;
+}
+
+static void gfx_v12_0_cp_gfx_switch_pipe(struct amdgpu_device *adev,
+					 CP_PIPE_ID pipe)
+{
+	u32 tmp;
+
+	tmp = RREG32_SOC15(GC, 0, regGRBM_GFX_CNTL);
+	tmp = REG_SET_FIELD(tmp, GRBM_GFX_CNTL, PIPEID, pipe);
+
+	WREG32_SOC15(GC, 0, regGRBM_GFX_CNTL, tmp);
+}
+
+static void gfx_v12_0_cp_gfx_set_doorbell(struct amdgpu_device *adev,
+					  struct amdgpu_ring *ring)
+{
+	u32 tmp;
+
+	tmp = RREG32_SOC15(GC, 0, regCP_RB_DOORBELL_CONTROL);
+	if (ring->use_doorbell) {
+		tmp = REG_SET_FIELD(tmp, CP_RB_DOORBELL_CONTROL,
+				    DOORBELL_OFFSET, ring->doorbell_index);
+		tmp = REG_SET_FIELD(tmp, CP_RB_DOORBELL_CONTROL,
+				    DOORBELL_EN, 1);
+	} else {
+		tmp = REG_SET_FIELD(tmp, CP_RB_DOORBELL_CONTROL,
+				    DOORBELL_EN, 0);
+	}
+	WREG32_SOC15(GC, 0, regCP_RB_DOORBELL_CONTROL, tmp);
+
+	tmp = REG_SET_FIELD(0, CP_RB_DOORBELL_RANGE_LOWER,
+			    DOORBELL_RANGE_LOWER, ring->doorbell_index);
+	WREG32_SOC15(GC, 0, regCP_RB_DOORBELL_RANGE_LOWER, tmp);
+
+	WREG32_SOC15(GC, 0, regCP_RB_DOORBELL_RANGE_UPPER,
+		     CP_RB_DOORBELL_RANGE_UPPER__DOORBELL_RANGE_UPPER_MASK);
+}
+
+static int gfx_v12_0_cp_gfx_resume(struct amdgpu_device *adev)
+{
+	struct amdgpu_ring *ring;
+	u32 tmp;
+	u32 rb_bufsz;
+	u64 rb_addr, rptr_addr, wptr_gpu_addr;
+
+	/* Set the write pointer delay */
+	WREG32_SOC15(GC, 0, regCP_RB_WPTR_DELAY, 0);
+
+	/* set the RB to use vmid 0 */
+	WREG32_SOC15(GC, 0, regCP_RB_VMID, 0);
+
+	/* Init gfx ring 0 for pipe 0 */
+	mutex_lock(&adev->srbm_mutex);
+	gfx_v12_0_cp_gfx_switch_pipe(adev, PIPE_ID0);
+
+	/* Set ring buffer size */
+	ring = &adev->gfx.gfx_ring[0];
+	rb_bufsz = order_base_2(ring->ring_size / 8);
+	tmp = REG_SET_FIELD(0, CP_RB0_CNTL, RB_BUFSZ, rb_bufsz);
+	tmp = REG_SET_FIELD(tmp, CP_RB0_CNTL, RB_BLKSZ, rb_bufsz - 2);
+	WREG32_SOC15(GC, 0, regCP_RB0_CNTL, tmp);
+
+	/* Initialize the ring buffer's write pointers */
+	ring->wptr = 0;
+	WREG32_SOC15(GC, 0, regCP_RB0_WPTR, lower_32_bits(ring->wptr));
+	WREG32_SOC15(GC, 0, regCP_RB0_WPTR_HI, upper_32_bits(ring->wptr));
+
+	/* set the wb address whether it's enabled or not */
+	rptr_addr = ring->rptr_gpu_addr;
+	WREG32_SOC15(GC, 0, regCP_RB0_RPTR_ADDR, lower_32_bits(rptr_addr));
+	WREG32_SOC15(GC, 0, regCP_RB0_RPTR_ADDR_HI, upper_32_bits(rptr_addr) &
+		     CP_RB_RPTR_ADDR_HI__RB_RPTR_ADDR_HI_MASK);
+
+	wptr_gpu_addr = ring->wptr_gpu_addr;
+	WREG32_SOC15(GC, 0, regCP_RB_WPTR_POLL_ADDR_LO,
+		     lower_32_bits(wptr_gpu_addr));
+	WREG32_SOC15(GC, 0, regCP_RB_WPTR_POLL_ADDR_HI,
+		     upper_32_bits(wptr_gpu_addr));
+
+	mdelay(1);
+	WREG32_SOC15(GC, 0, regCP_RB0_CNTL, tmp);
+
+	rb_addr = ring->gpu_addr >> 8;
+	WREG32_SOC15(GC, 0, regCP_RB0_BASE, rb_addr);
+	WREG32_SOC15(GC, 0, regCP_RB0_BASE_HI, upper_32_bits(rb_addr));
+
+	WREG32_SOC15(GC, 0, regCP_RB_ACTIVE, 1);
+
+	gfx_v12_0_cp_gfx_set_doorbell(adev, ring);
+	mutex_unlock(&adev->srbm_mutex);
+
+	/* Switch to pipe 0 */
+	mutex_lock(&adev->srbm_mutex);
+	gfx_v12_0_cp_gfx_switch_pipe(adev, PIPE_ID0);
+	mutex_unlock(&adev->srbm_mutex);
+
+	/* start the ring */
+	gfx_v12_0_cp_gfx_start(adev);
+	return 0;
+}
+
+static void gfx_v12_0_cp_compute_enable(struct amdgpu_device *adev, bool enable)
+{
+	u32 data;
+
+	data = RREG32_SOC15(GC, 0, regCP_MEC_RS64_CNTL);
+	data = REG_SET_FIELD(data, CP_MEC_RS64_CNTL, MEC_INVALIDATE_ICACHE,
+						 enable ? 0 : 1);
+	data = REG_SET_FIELD(data, CP_MEC_RS64_CNTL, MEC_PIPE0_RESET,
+						 enable ? 0 : 1);
+	data = REG_SET_FIELD(data, CP_MEC_RS64_CNTL, MEC_PIPE1_RESET,
+						 enable ? 0 : 1);
+	data = REG_SET_FIELD(data, CP_MEC_RS64_CNTL, MEC_PIPE2_RESET,
+						 enable ? 0 : 1);
+	data = REG_SET_FIELD(data, CP_MEC_RS64_CNTL, MEC_PIPE3_RESET,
+						 enable ? 0 : 1);
+	data = REG_SET_FIELD(data, CP_MEC_RS64_CNTL, MEC_PIPE0_ACTIVE,
+						 enable ? 1 : 0);
+	data = REG_SET_FIELD(data, CP_MEC_RS64_CNTL, MEC_PIPE1_ACTIVE,
+			                         enable ? 1 : 0);
+	data = REG_SET_FIELD(data, CP_MEC_RS64_CNTL, MEC_PIPE2_ACTIVE,
+						 enable ? 1 : 0);
+	data = REG_SET_FIELD(data, CP_MEC_RS64_CNTL, MEC_PIPE3_ACTIVE,
+						 enable ? 1 : 0);
+	data = REG_SET_FIELD(data, CP_MEC_RS64_CNTL, MEC_HALT,
+						 enable ? 0 : 1);
+	WREG32_SOC15(GC, 0, regCP_MEC_RS64_CNTL, data);
+
+	adev->gfx.kiq[0].ring.sched.ready = enable;
+
+	udelay(50);
+}
+
+static int gfx_v12_0_cp_compute_load_microcode_rs64(struct amdgpu_device *adev)
+{
+	const struct gfx_firmware_header_v2_0 *mec_hdr;
+	const __le32 *fw_ucode, *fw_data;
+	u32 tmp, fw_ucode_size, fw_data_size;
+	u32 i, usec_timeout = 50000; /* Wait for 50 ms */
+	u32 *fw_ucode_ptr, *fw_data_ptr;
+	int r;
+
+	if (!adev->gfx.mec_fw)
+		return -EINVAL;
+
+	gfx_v12_0_cp_compute_enable(adev, false);
+
+	mec_hdr = (const struct gfx_firmware_header_v2_0 *)adev->gfx.mec_fw->data;
+	amdgpu_ucode_print_gfx_hdr(&mec_hdr->header);
+
+	fw_ucode = (const __le32 *) (adev->gfx.mec_fw->data +
+				le32_to_cpu(mec_hdr->ucode_offset_bytes));
+	fw_ucode_size = le32_to_cpu(mec_hdr->ucode_size_bytes);
+
+	fw_data = (const __le32 *) (adev->gfx.mec_fw->data +
+				le32_to_cpu(mec_hdr->data_offset_bytes));
+	fw_data_size = le32_to_cpu(mec_hdr->data_size_bytes);
+
+	r = amdgpu_bo_create_reserved(adev, fw_ucode_size,
+				      64 * 1024, AMDGPU_GEM_DOMAIN_VRAM,
+				      &adev->gfx.mec.mec_fw_obj,
+				      &adev->gfx.mec.mec_fw_gpu_addr,
+				      (void **)&fw_ucode_ptr);
+	if (r) {
+		dev_err(adev->dev, "(%d) failed to create mec fw ucode bo\n", r);
+		gfx_v12_0_mec_fini(adev);
+		return r;
+	}
+
+	r = amdgpu_bo_create_reserved(adev,
+				      ALIGN(fw_data_size, 64 * 1024) *
+				      adev->gfx.mec.num_pipe_per_mec,
+				      64 * 1024, AMDGPU_GEM_DOMAIN_VRAM,
+				      &adev->gfx.mec.mec_fw_data_obj,
+				      &adev->gfx.mec.mec_fw_data_gpu_addr,
+				      (void **)&fw_data_ptr);
+	if (r) {
+		dev_err(adev->dev, "(%d) failed to create mec fw ucode bo\n", r);
+		gfx_v12_0_mec_fini(adev);
+		return r;
+	}
+
+	memcpy(fw_ucode_ptr, fw_ucode, fw_ucode_size);
+	for (i = 0; i < adev->gfx.mec.num_pipe_per_mec; i++) {
+		memcpy(fw_data_ptr + i * ALIGN(fw_data_size, 64 * 1024) / 4, fw_data, fw_data_size);
+	}
+
+	amdgpu_bo_kunmap(adev->gfx.mec.mec_fw_obj);
+	amdgpu_bo_kunmap(adev->gfx.mec.mec_fw_data_obj);
+	amdgpu_bo_unreserve(adev->gfx.mec.mec_fw_obj);
+	amdgpu_bo_unreserve(adev->gfx.mec.mec_fw_data_obj);
+
+	tmp = RREG32_SOC15(GC, 0, regCP_CPC_IC_BASE_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_CPC_IC_BASE_CNTL, VMID, 0);
+	tmp = REG_SET_FIELD(tmp, CP_CPC_IC_BASE_CNTL, EXE_DISABLE, 0);
+	tmp = REG_SET_FIELD(tmp, CP_CPC_IC_BASE_CNTL, CACHE_POLICY, 0);
+	WREG32_SOC15(GC, 0, regCP_CPC_IC_BASE_CNTL, tmp);
+
+	tmp = RREG32_SOC15(GC, 0, regCP_MEC_DC_BASE_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_MEC_DC_BASE_CNTL, VMID, 0);
+	tmp = REG_SET_FIELD(tmp, CP_MEC_DC_BASE_CNTL, CACHE_POLICY, 0);
+	WREG32_SOC15(GC, 0, regCP_MEC_DC_BASE_CNTL, tmp);
+
+	mutex_lock(&adev->srbm_mutex);
+	for (i = 0; i < adev->gfx.mec.num_pipe_per_mec; i++) {
+		soc24_grbm_select(adev, 1, i, 0, 0);
+
+		WREG32_SOC15(GC, 0, regCP_MEC_MDBASE_LO,
+			     lower_32_bits(adev->gfx.mec.mec_fw_data_gpu_addr +
+					   i * ALIGN(fw_data_size, 64 * 1024)));
+		WREG32_SOC15(GC, 0, regCP_MEC_MDBASE_HI,
+			     upper_32_bits(adev->gfx.mec.mec_fw_data_gpu_addr +
+					   i * ALIGN(fw_data_size, 64 * 1024)));
+
+		WREG32_SOC15(GC, 0, regCP_CPC_IC_BASE_LO,
+			     lower_32_bits(adev->gfx.mec.mec_fw_gpu_addr));
+		WREG32_SOC15(GC, 0, regCP_CPC_IC_BASE_HI,
+			     upper_32_bits(adev->gfx.mec.mec_fw_gpu_addr));
+	}
+	mutex_unlock(&adev->srbm_mutex);
+	soc24_grbm_select(adev, 0, 0, 0, 0);
+
+	/* Trigger an invalidation of the L1 instruction caches */
+	tmp = RREG32_SOC15(GC, 0, regCP_MEC_DC_OP_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_MEC_DC_OP_CNTL, INVALIDATE_DCACHE, 1);
+	WREG32_SOC15(GC, 0, regCP_MEC_DC_OP_CNTL, tmp);
+
+	/* Wait for invalidation complete */
+	for (i = 0; i < usec_timeout; i++) {
+		tmp = RREG32_SOC15(GC, 0, regCP_MEC_DC_OP_CNTL);
+		if (1 == REG_GET_FIELD(tmp, CP_MEC_DC_OP_CNTL,
+				       INVALIDATE_DCACHE_COMPLETE))
+			break;
+		udelay(1);
+	}
+
+	if (i >= usec_timeout) {
+		dev_err(adev->dev, "failed to invalidate instruction cache\n");
+		return -EINVAL;
+	}
+
+	/* Trigger an invalidation of the L1 instruction caches */
+	tmp = RREG32_SOC15(GC, 0, regCP_CPC_IC_OP_CNTL);
+	tmp = REG_SET_FIELD(tmp, CP_CPC_IC_OP_CNTL, INVALIDATE_CACHE, 1);
+	WREG32_SOC15(GC, 0, regCP_CPC_IC_OP_CNTL, tmp);
+
+	/* Wait for invalidation complete */
+	for (i = 0; i < usec_timeout; i++) {
+		tmp = RREG32_SOC15(GC, 0, regCP_CPC_IC_OP_CNTL);
+		if (1 == REG_GET_FIELD(tmp, CP_CPC_IC_OP_CNTL,
+				       INVALIDATE_CACHE_COMPLETE))
+			break;
+		udelay(1);
+	}
+
+	if (i >= usec_timeout) {
+		dev_err(adev->dev, "failed to invalidate instruction cache\n");
+		return -EINVAL;
+	}
+
+	gfx_v12_0_set_mec_ucode_start_addr(adev);
+
+	return 0;
+}
+
+static void gfx_v12_0_kiq_setting(struct amdgpu_ring *ring)
+{
+	uint32_t tmp;
+	struct amdgpu_device *adev = ring->adev;
+
+	/* tell RLC which is KIQ queue */
+	tmp = RREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS);
+	tmp &= 0xffffff00;
+	tmp |= (ring->me << 5) | (ring->pipe << 3) | (ring->queue);
+	WREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS, tmp | 0x80);
+}
+
+static void gfx_v12_0_cp_set_doorbell_range(struct amdgpu_device *adev)
+{
+	/* set graphics engine doorbell range */
+	WREG32_SOC15(GC, 0, regCP_RB_DOORBELL_RANGE_LOWER,
+		     (adev->doorbell_index.gfx_ring0 * 2) << 2);
+	WREG32_SOC15(GC, 0, regCP_RB_DOORBELL_RANGE_UPPER,
+		     (adev->doorbell_index.gfx_userqueue_end * 2) << 2);
+
+	/* set compute engine doorbell range */
+	WREG32_SOC15(GC, 0, regCP_MEC_DOORBELL_RANGE_LOWER,
+		     (adev->doorbell_index.kiq * 2) << 2);
+	WREG32_SOC15(GC, 0, regCP_MEC_DOORBELL_RANGE_UPPER,
+		     (adev->doorbell_index.userqueue_end * 2) << 2);
+}
+
+static int gfx_v12_0_gfx_mqd_init(struct amdgpu_device *adev, void *m,
+				  struct amdgpu_mqd_prop *prop)
+{
+	struct v12_gfx_mqd *mqd = m;
+	uint64_t hqd_gpu_addr, wb_gpu_addr;
+	uint32_t tmp;
+	uint32_t rb_bufsz;
+
+	/* set up gfx hqd wptr */
+	mqd->cp_gfx_hqd_wptr = 0;
+	mqd->cp_gfx_hqd_wptr_hi = 0;
+
+	/* set the pointer to the MQD */
+	mqd->cp_mqd_base_addr = prop->mqd_gpu_addr & 0xfffffffc;
+	mqd->cp_mqd_base_addr_hi = upper_32_bits(prop->mqd_gpu_addr);
+
+	/* set up mqd control */
+	tmp = regCP_GFX_MQD_CONTROL_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, CP_GFX_MQD_CONTROL, VMID, 0);
+	tmp = REG_SET_FIELD(tmp, CP_GFX_MQD_CONTROL, PRIV_STATE, 1);
+	tmp = REG_SET_FIELD(tmp, CP_GFX_MQD_CONTROL, CACHE_POLICY, 0);
+	mqd->cp_gfx_mqd_control = tmp;
+
+	/* set up gfx_hqd_vimd with 0x0 to indicate the ring buffer's vmid */
+	tmp = regCP_GFX_HQD_VMID_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, CP_GFX_HQD_VMID, VMID, 0);
+	mqd->cp_gfx_hqd_vmid = 0;
+
+	/* set up default queue priority level
+	 * 0x0 = low priority, 0x1 = high priority */
+	tmp = regCP_GFX_HQD_QUEUE_PRIORITY_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, CP_GFX_HQD_QUEUE_PRIORITY, PRIORITY_LEVEL, 0);
+	mqd->cp_gfx_hqd_queue_priority = tmp;
+
+	/* set up time quantum */
+	tmp = regCP_GFX_HQD_QUANTUM_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, CP_GFX_HQD_QUANTUM, QUANTUM_EN, 1);
+	mqd->cp_gfx_hqd_quantum = tmp;
+
+	/* set up gfx hqd base. this is similar as CP_RB_BASE */
+	hqd_gpu_addr = prop->hqd_base_gpu_addr >> 8;
+	mqd->cp_gfx_hqd_base = hqd_gpu_addr;
+	mqd->cp_gfx_hqd_base_hi = upper_32_bits(hqd_gpu_addr);
+
+	/* set up hqd_rptr_addr/_hi, similar as CP_RB_RPTR */
+	wb_gpu_addr = prop->rptr_gpu_addr;
+	mqd->cp_gfx_hqd_rptr_addr = wb_gpu_addr & 0xfffffffc;
+	mqd->cp_gfx_hqd_rptr_addr_hi =
+		upper_32_bits(wb_gpu_addr) & 0xffff;
+
+	/* set up rb_wptr_poll addr */
+	wb_gpu_addr = prop->wptr_gpu_addr;
+	mqd->cp_rb_wptr_poll_addr_lo = wb_gpu_addr & 0xfffffffc;
+	mqd->cp_rb_wptr_poll_addr_hi = upper_32_bits(wb_gpu_addr) & 0xffff;
+
+	/* set up the gfx_hqd_control, similar as CP_RB0_CNTL */
+	rb_bufsz = order_base_2(prop->queue_size / 4) - 1;
+	tmp = regCP_GFX_HQD_CNTL_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, CP_GFX_HQD_CNTL, RB_BUFSZ, rb_bufsz);
+	tmp = REG_SET_FIELD(tmp, CP_GFX_HQD_CNTL, RB_BLKSZ, rb_bufsz - 2);
+#ifdef __BIG_ENDIAN
+	tmp = REG_SET_FIELD(tmp, CP_GFX_HQD_CNTL, BUF_SWAP, 1);
+#endif
+	if (prop->tmz_queue)
+		tmp = REG_SET_FIELD(tmp, CP_GFX_HQD_CNTL, TMZ_MATCH, 1);
+	if (!prop->kernel_queue)
+		tmp = REG_SET_FIELD(tmp, CP_GFX_HQD_CNTL, RB_NON_PRIV, 1);
+	mqd->cp_gfx_hqd_cntl = tmp;
+
+	/* set up cp_doorbell_control */
+	tmp = regCP_RB_DOORBELL_CONTROL_DEFAULT;
+	if (prop->use_doorbell) {
+		tmp = REG_SET_FIELD(tmp, CP_RB_DOORBELL_CONTROL,
+				    DOORBELL_OFFSET, prop->doorbell_index);
+		tmp = REG_SET_FIELD(tmp, CP_RB_DOORBELL_CONTROL,
+				    DOORBELL_EN, 1);
+	} else
+		tmp = REG_SET_FIELD(tmp, CP_RB_DOORBELL_CONTROL,
+				    DOORBELL_EN, 0);
+	mqd->cp_rb_doorbell_control = tmp;
+
+	/* reset read and write pointers, similar to CP_RB0_WPTR/_RPTR */
+	mqd->cp_gfx_hqd_rptr = regCP_GFX_HQD_RPTR_DEFAULT;
+
+	/* active the queue */
+	mqd->cp_gfx_hqd_active = 1;
+
+	/* set gfx UQ items */
+	mqd->shadow_base_lo = lower_32_bits(prop->shadow_addr);
+	mqd->shadow_base_hi = upper_32_bits(prop->shadow_addr);
+	mqd->fw_work_area_base_lo = lower_32_bits(prop->csa_addr);
+	mqd->fw_work_area_base_hi = upper_32_bits(prop->csa_addr);
+	mqd->fence_address_lo = lower_32_bits(prop->fence_address);
+	mqd->fence_address_hi = upper_32_bits(prop->fence_address);
+
+	return 0;
+}
+
+static int gfx_v12_0_kgq_init_queue(struct amdgpu_ring *ring, bool reset)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct v12_gfx_mqd *mqd = ring->mqd_ptr;
+	int mqd_idx = ring - &adev->gfx.gfx_ring[0];
+
+	if (!reset && !amdgpu_in_reset(adev) && !adev->in_suspend) {
+		memset((void *)mqd, 0, sizeof(*mqd));
+		mutex_lock(&adev->srbm_mutex);
+		soc24_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0);
+		amdgpu_ring_init_mqd(ring);
+		soc24_grbm_select(adev, 0, 0, 0, 0);
+		mutex_unlock(&adev->srbm_mutex);
+		if (adev->gfx.me.mqd_backup[mqd_idx])
+			memcpy_fromio(adev->gfx.me.mqd_backup[mqd_idx], mqd, sizeof(*mqd));
+	} else {
+		/* restore mqd with the backup copy */
+		if (adev->gfx.me.mqd_backup[mqd_idx])
+			memcpy_toio(mqd, adev->gfx.me.mqd_backup[mqd_idx], sizeof(*mqd));
+		/* reset the ring */
+		ring->wptr = 0;
+		*ring->wptr_cpu_addr = 0;
+		amdgpu_ring_clear_ring(ring);
+	}
+
+	return 0;
+}
+
+static int gfx_v12_0_cp_async_gfx_ring_resume(struct amdgpu_device *adev)
+{
+	int i, r;
+
+	for (i = 0; i < adev->gfx.num_gfx_rings; i++) {
+		r = gfx_v12_0_kgq_init_queue(&adev->gfx.gfx_ring[i], false);
+		if (r)
+			return r;
+	}
+
+	r = amdgpu_gfx_enable_kgq(adev, 0);
+	if (r)
+		return r;
+
+	return gfx_v12_0_cp_gfx_start(adev);
+}
+
+static int gfx_v12_0_compute_mqd_init(struct amdgpu_device *adev, void *m,
+				      struct amdgpu_mqd_prop *prop)
+{
+	struct v12_compute_mqd *mqd = m;
+	uint64_t hqd_gpu_addr, wb_gpu_addr, eop_base_addr;
+	uint32_t tmp;
+
+	mqd->header = 0xC0310800;
+	mqd->compute_pipelinestat_enable = 0x00000001;
+	mqd->compute_static_thread_mgmt_se0 = 0xffffffff;
+	mqd->compute_static_thread_mgmt_se1 = 0xffffffff;
+	mqd->compute_static_thread_mgmt_se2 = 0xffffffff;
+	mqd->compute_static_thread_mgmt_se3 = 0xffffffff;
+	mqd->compute_misc_reserved = 0x00000007;
+
+	eop_base_addr = prop->eop_gpu_addr >> 8;
+	mqd->cp_hqd_eop_base_addr_lo = eop_base_addr;
+	mqd->cp_hqd_eop_base_addr_hi = upper_32_bits(eop_base_addr);
+
+	/* set the EOP size, register value is 2^(EOP_SIZE+1) dwords */
+	tmp = regCP_HQD_EOP_CONTROL_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, CP_HQD_EOP_CONTROL, EOP_SIZE,
+			(order_base_2(GFX12_MEC_HPD_SIZE / 4) - 1));
+
+	mqd->cp_hqd_eop_control = tmp;
+
+	/* enable doorbell? */
+	tmp = regCP_HQD_PQ_DOORBELL_CONTROL_DEFAULT;
+
+	if (prop->use_doorbell) {
+		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
+				    DOORBELL_OFFSET, prop->doorbell_index);
+		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
+				    DOORBELL_EN, 1);
+		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
+				    DOORBELL_SOURCE, 0);
+		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
+				    DOORBELL_HIT, 0);
+	} else {
+		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
+				    DOORBELL_EN, 0);
+	}
+
+	mqd->cp_hqd_pq_doorbell_control = tmp;
+
+	/* disable the queue if it's active */
+	mqd->cp_hqd_dequeue_request = 0;
+	mqd->cp_hqd_pq_rptr = 0;
+	mqd->cp_hqd_pq_wptr_lo = 0;
+	mqd->cp_hqd_pq_wptr_hi = 0;
+
+	/* set the pointer to the MQD */
+	mqd->cp_mqd_base_addr_lo = prop->mqd_gpu_addr & 0xfffffffc;
+	mqd->cp_mqd_base_addr_hi = upper_32_bits(prop->mqd_gpu_addr);
+
+	/* set MQD vmid to 0 */
+	tmp = regCP_MQD_CONTROL_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, CP_MQD_CONTROL, VMID, 0);
+	mqd->cp_mqd_control = tmp;
+
+	/* set the pointer to the HQD, this is similar CP_RB0_BASE/_HI */
+	hqd_gpu_addr = prop->hqd_base_gpu_addr >> 8;
+	mqd->cp_hqd_pq_base_lo = hqd_gpu_addr;
+	mqd->cp_hqd_pq_base_hi = upper_32_bits(hqd_gpu_addr);
+
+	/* set up the HQD, this is similar to CP_RB0_CNTL */
+	tmp = regCP_HQD_PQ_CONTROL_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, QUEUE_SIZE,
+			    (order_base_2(prop->queue_size / 4) - 1));
+	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, RPTR_BLOCK_SIZE,
+			    (order_base_2(AMDGPU_GPU_PAGE_SIZE / 4) - 1));
+	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, UNORD_DISPATCH, 1);
+	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, TUNNEL_DISPATCH, 0);
+	if (prop->kernel_queue) {
+		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, PRIV_STATE, 1);
+		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, KMD_QUEUE, 1);
+	}
+	if (prop->tmz_queue)
+		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, TMZ, 1);
+	mqd->cp_hqd_pq_control = tmp;
+
+	/* set the wb address whether it's enabled or not */
+	wb_gpu_addr = prop->rptr_gpu_addr;
+	mqd->cp_hqd_pq_rptr_report_addr_lo = wb_gpu_addr & 0xfffffffc;
+	mqd->cp_hqd_pq_rptr_report_addr_hi =
+		upper_32_bits(wb_gpu_addr) & 0xffff;
+
+	/* only used if CP_PQ_WPTR_POLL_CNTL.CP_PQ_WPTR_POLL_CNTL__EN_MASK=1 */
+	wb_gpu_addr = prop->wptr_gpu_addr;
+	mqd->cp_hqd_pq_wptr_poll_addr_lo = wb_gpu_addr & 0xfffffffc;
+	mqd->cp_hqd_pq_wptr_poll_addr_hi = upper_32_bits(wb_gpu_addr) & 0xffff;
+
+	tmp = 0;
+	/* enable the doorbell if requested */
+	if (prop->use_doorbell) {
+		tmp = regCP_HQD_PQ_DOORBELL_CONTROL_DEFAULT;
+		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
+				DOORBELL_OFFSET, prop->doorbell_index);
+
+		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
+				    DOORBELL_EN, 1);
+		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
+				    DOORBELL_SOURCE, 0);
+		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
+				    DOORBELL_HIT, 0);
+	}
+
+	mqd->cp_hqd_pq_doorbell_control = tmp;
+
+	/* reset read and write pointers, similar to CP_RB0_WPTR/_RPTR */
+	mqd->cp_hqd_pq_rptr = regCP_HQD_PQ_RPTR_DEFAULT;
+
+	/* set the vmid for the queue */
+	mqd->cp_hqd_vmid = 0;
+
+	tmp = regCP_HQD_PERSISTENT_STATE_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, CP_HQD_PERSISTENT_STATE, PRELOAD_SIZE, 0x55);
+	mqd->cp_hqd_persistent_state = tmp;
+
+	/* set MIN_IB_AVAIL_SIZE */
+	tmp = regCP_HQD_IB_CONTROL_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, CP_HQD_IB_CONTROL, MIN_IB_AVAIL_SIZE, 3);
+	mqd->cp_hqd_ib_control = tmp;
+
+	/* set static priority for a compute queue/ring */
+	mqd->cp_hqd_pipe_priority = prop->hqd_pipe_priority;
+	mqd->cp_hqd_queue_priority = prop->hqd_queue_priority;
+
+	mqd->cp_hqd_active = prop->hqd_active;
+
+	/* set UQ fenceaddress */
+	mqd->fence_address_lo = lower_32_bits(prop->fence_address);
+	mqd->fence_address_hi = upper_32_bits(prop->fence_address);
+
+	return 0;
+}
+
+static int gfx_v12_0_kiq_init_register(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct v12_compute_mqd *mqd = ring->mqd_ptr;
+	int j;
+
+	/* inactivate the queue */
+	if (amdgpu_sriov_vf(adev))
+		WREG32_SOC15(GC, 0, regCP_HQD_ACTIVE, 0);
+
+	/* disable wptr polling */
+	WREG32_FIELD15_PREREG(GC, 0, CP_PQ_WPTR_POLL_CNTL, EN, 0);
+
+	/* write the EOP addr */
+	WREG32_SOC15(GC, 0, regCP_HQD_EOP_BASE_ADDR,
+	       mqd->cp_hqd_eop_base_addr_lo);
+	WREG32_SOC15(GC, 0, regCP_HQD_EOP_BASE_ADDR_HI,
+	       mqd->cp_hqd_eop_base_addr_hi);
+
+	/* set the EOP size, register value is 2^(EOP_SIZE+1) dwords */
+	WREG32_SOC15(GC, 0, regCP_HQD_EOP_CONTROL,
+	       mqd->cp_hqd_eop_control);
+
+	/* enable doorbell? */
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_DOORBELL_CONTROL,
+	       mqd->cp_hqd_pq_doorbell_control);
+
+	/* disable the queue if it's active */
+	if (RREG32_SOC15(GC, 0, regCP_HQD_ACTIVE) & 1) {
+		WREG32_SOC15(GC, 0, regCP_HQD_DEQUEUE_REQUEST, 1);
+		for (j = 0; j < adev->usec_timeout; j++) {
+			if (!(RREG32_SOC15(GC, 0, regCP_HQD_ACTIVE) & 1))
+				break;
+			udelay(1);
+		}
+		WREG32_SOC15(GC, 0, regCP_HQD_DEQUEUE_REQUEST,
+		       mqd->cp_hqd_dequeue_request);
+		WREG32_SOC15(GC, 0, regCP_HQD_PQ_RPTR,
+		       mqd->cp_hqd_pq_rptr);
+		WREG32_SOC15(GC, 0, regCP_HQD_PQ_WPTR_LO,
+		       mqd->cp_hqd_pq_wptr_lo);
+		WREG32_SOC15(GC, 0, regCP_HQD_PQ_WPTR_HI,
+		       mqd->cp_hqd_pq_wptr_hi);
+	}
+
+	/* set the pointer to the MQD */
+	WREG32_SOC15(GC, 0, regCP_MQD_BASE_ADDR,
+	       mqd->cp_mqd_base_addr_lo);
+	WREG32_SOC15(GC, 0, regCP_MQD_BASE_ADDR_HI,
+	       mqd->cp_mqd_base_addr_hi);
+
+	/* set MQD vmid to 0 */
+	WREG32_SOC15(GC, 0, regCP_MQD_CONTROL,
+	       mqd->cp_mqd_control);
+
+	/* set the pointer to the HQD, this is similar CP_RB0_BASE/_HI */
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_BASE,
+	       mqd->cp_hqd_pq_base_lo);
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_BASE_HI,
+	       mqd->cp_hqd_pq_base_hi);
+
+	/* set up the HQD, this is similar to CP_RB0_CNTL */
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_CONTROL,
+	       mqd->cp_hqd_pq_control);
+
+	/* set the wb address whether it's enabled or not */
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_RPTR_REPORT_ADDR,
+		mqd->cp_hqd_pq_rptr_report_addr_lo);
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_RPTR_REPORT_ADDR_HI,
+		mqd->cp_hqd_pq_rptr_report_addr_hi);
+
+	/* only used if CP_PQ_WPTR_POLL_CNTL.CP_PQ_WPTR_POLL_CNTL__EN_MASK=1 */
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_WPTR_POLL_ADDR,
+	       mqd->cp_hqd_pq_wptr_poll_addr_lo);
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_WPTR_POLL_ADDR_HI,
+	       mqd->cp_hqd_pq_wptr_poll_addr_hi);
+
+	/* enable the doorbell if requested */
+	if (ring->use_doorbell) {
+		WREG32_SOC15(GC, 0, regCP_MEC_DOORBELL_RANGE_LOWER,
+			(adev->doorbell_index.kiq * 2) << 2);
+		WREG32_SOC15(GC, 0, regCP_MEC_DOORBELL_RANGE_UPPER,
+			(adev->doorbell_index.userqueue_end * 2) << 2);
+	}
+
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_DOORBELL_CONTROL,
+	       mqd->cp_hqd_pq_doorbell_control);
+
+	/* reset read and write pointers, similar to CP_RB0_WPTR/_RPTR */
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_WPTR_LO,
+	       mqd->cp_hqd_pq_wptr_lo);
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_WPTR_HI,
+	       mqd->cp_hqd_pq_wptr_hi);
+
+	/* set the vmid for the queue */
+	WREG32_SOC15(GC, 0, regCP_HQD_VMID, mqd->cp_hqd_vmid);
+
+	WREG32_SOC15(GC, 0, regCP_HQD_PERSISTENT_STATE,
+	       mqd->cp_hqd_persistent_state);
+
+	/* activate the queue */
+	WREG32_SOC15(GC, 0, regCP_HQD_ACTIVE,
+	       mqd->cp_hqd_active);
+
+	if (ring->use_doorbell)
+		WREG32_FIELD15_PREREG(GC, 0, CP_PQ_STATUS, DOORBELL_ENABLE, 1);
+
+	return 0;
+}
+
+static int gfx_v12_0_kiq_init_queue(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct v12_compute_mqd *mqd = ring->mqd_ptr;
+	int mqd_idx = AMDGPU_MAX_COMPUTE_RINGS;
+
+	gfx_v12_0_kiq_setting(ring);
+
+	if (amdgpu_in_reset(adev)) { /* for GPU_RESET case */
+		/* reset MQD to a clean status */
+		if (adev->gfx.mec.mqd_backup[mqd_idx])
+			memcpy(mqd, adev->gfx.mec.mqd_backup[mqd_idx], sizeof(*mqd));
+
+		/* reset ring buffer */
+		ring->wptr = 0;
+		amdgpu_ring_clear_ring(ring);
+
+		mutex_lock(&adev->srbm_mutex);
+		soc24_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0);
+		gfx_v12_0_kiq_init_register(ring);
+		soc24_grbm_select(adev, 0, 0, 0, 0);
+		mutex_unlock(&adev->srbm_mutex);
+	} else {
+		memset((void *)mqd, 0, sizeof(*mqd));
+		if (amdgpu_sriov_vf(adev) && adev->in_suspend)
+			amdgpu_ring_clear_ring(ring);
+		mutex_lock(&adev->srbm_mutex);
+		soc24_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0);
+		amdgpu_ring_init_mqd(ring);
+		gfx_v12_0_kiq_init_register(ring);
+		soc24_grbm_select(adev, 0, 0, 0, 0);
+		mutex_unlock(&adev->srbm_mutex);
+
+		if (adev->gfx.mec.mqd_backup[mqd_idx])
+			memcpy(adev->gfx.mec.mqd_backup[mqd_idx], mqd, sizeof(*mqd));
+	}
+
+	return 0;
+}
+
+static int gfx_v12_0_kcq_init_queue(struct amdgpu_ring *ring, bool reset)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct v12_compute_mqd *mqd = ring->mqd_ptr;
+	int mqd_idx = ring - &adev->gfx.compute_ring[0];
+
+	if (!reset && !amdgpu_in_reset(adev) && !adev->in_suspend) {
+		memset((void *)mqd, 0, sizeof(*mqd));
+		mutex_lock(&adev->srbm_mutex);
+		soc24_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0);
+		amdgpu_ring_init_mqd(ring);
+		soc24_grbm_select(adev, 0, 0, 0, 0);
+		mutex_unlock(&adev->srbm_mutex);
+
+		if (adev->gfx.mec.mqd_backup[mqd_idx])
+			memcpy_fromio(adev->gfx.mec.mqd_backup[mqd_idx], mqd, sizeof(*mqd));
+	} else {
+		/* restore MQD to a clean status */
+		if (adev->gfx.mec.mqd_backup[mqd_idx])
+			memcpy_toio(mqd, adev->gfx.mec.mqd_backup[mqd_idx], sizeof(*mqd));
+		/* reset ring buffer */
+		ring->wptr = 0;
+		atomic64_set((atomic64_t *)ring->wptr_cpu_addr, 0);
+		amdgpu_ring_clear_ring(ring);
+	}
+
+	return 0;
+}
+
+static int gfx_v12_0_kiq_resume(struct amdgpu_device *adev)
+{
+	gfx_v12_0_kiq_init_queue(&adev->gfx.kiq[0].ring);
+	adev->gfx.kiq[0].ring.sched.ready = true;
+	return 0;
+}
+
+static int gfx_v12_0_kcq_resume(struct amdgpu_device *adev)
+{
+	int i, r;
+
+	if (!amdgpu_async_gfx_ring)
+		gfx_v12_0_cp_compute_enable(adev, true);
+
+	for (i = 0; i < adev->gfx.num_compute_rings; i++) {
+		r = gfx_v12_0_kcq_init_queue(&adev->gfx.compute_ring[i], false);
+		if (r)
+			return r;
+	}
+
+	return amdgpu_gfx_enable_kcq(adev, 0);
+}
+
+static int gfx_v12_0_cp_resume(struct amdgpu_device *adev)
+{
+	int r, i;
+	struct amdgpu_ring *ring;
+
+	if (!(adev->flags & AMD_IS_APU))
+		gfx_v12_0_enable_gui_idle_interrupt(adev, false);
+
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT) {
+		/* legacy firmware loading */
+		r = gfx_v12_0_cp_gfx_load_microcode(adev);
+		if (r)
+			return r;
+
+		r = gfx_v12_0_cp_compute_load_microcode_rs64(adev);
+		if (r)
+			return r;
+	}
+
+	gfx_v12_0_cp_set_doorbell_range(adev);
+
+	if (amdgpu_async_gfx_ring) {
+		gfx_v12_0_cp_compute_enable(adev, true);
+		gfx_v12_0_cp_gfx_enable(adev, true);
+	}
+
+	if (adev->enable_mes_kiq && adev->mes.kiq_hw_init)
+		r = amdgpu_mes_kiq_hw_init(adev);
+	else
+		r = gfx_v12_0_kiq_resume(adev);
+	if (r)
+		return r;
+
+	r = gfx_v12_0_kcq_resume(adev);
+	if (r)
+		return r;
+
+	if (!amdgpu_async_gfx_ring) {
+		r = gfx_v12_0_cp_gfx_resume(adev);
+		if (r)
+			return r;
+	} else {
+		r = gfx_v12_0_cp_async_gfx_ring_resume(adev);
+		if (r)
+			return r;
+	}
+
+	for (i = 0; i < adev->gfx.num_gfx_rings; i++) {
+		ring = &adev->gfx.gfx_ring[i];
+		r = amdgpu_ring_test_helper(ring);
+		if (r)
+			return r;
+	}
+
+	for (i = 0; i < adev->gfx.num_compute_rings; i++) {
+		ring = &adev->gfx.compute_ring[i];
+		r = amdgpu_ring_test_helper(ring);
+		if (r)
+			return r;
+	}
+
+	return 0;
+}
+
+static void gfx_v12_0_cp_enable(struct amdgpu_device *adev, bool enable)
+{
+	gfx_v12_0_cp_gfx_enable(adev, enable);
+	gfx_v12_0_cp_compute_enable(adev, enable);
+}
+
+static int gfx_v12_0_gfxhub_enable(struct amdgpu_device *adev)
+{
+	int r;
+	bool value;
+
+	r = adev->gfxhub.funcs->gart_enable(adev);
+	if (r)
+		return r;
+
+	amdgpu_device_flush_hdp(adev, NULL);
+
+	value = amdgpu_vm_fault_stop != AMDGPU_VM_FAULT_STOP_ALWAYS;
+
+	adev->gfxhub.funcs->set_fault_enable_default(adev, value);
+	/* TODO investigate why this and the hdp flush above is needed,
+	 * are we missing a flush somewhere else? */
+	adev->gmc.gmc_funcs->flush_gpu_tlb(adev, 0, AMDGPU_GFXHUB(0), 0);
+
+	return 0;
+}
+
+static int get_gb_addr_config(struct amdgpu_device *adev)
+{
+	u32 gb_addr_config;
+
+	gb_addr_config = RREG32_SOC15(GC, 0, regGB_ADDR_CONFIG);
+	if (gb_addr_config == 0)
+		return -EINVAL;
+
+	adev->gfx.config.gb_addr_config_fields.num_pkrs =
+		1 << REG_GET_FIELD(gb_addr_config, GB_ADDR_CONFIG, NUM_PKRS);
+
+	adev->gfx.config.gb_addr_config = gb_addr_config;
+
+	adev->gfx.config.gb_addr_config_fields.num_pipes = 1 <<
+			REG_GET_FIELD(adev->gfx.config.gb_addr_config,
+				      GB_ADDR_CONFIG, NUM_PIPES);
+
+	adev->gfx.config.max_tile_pipes =
+		adev->gfx.config.gb_addr_config_fields.num_pipes;
+
+	adev->gfx.config.gb_addr_config_fields.max_compress_frags = 1 <<
+			REG_GET_FIELD(adev->gfx.config.gb_addr_config,
+				      GB_ADDR_CONFIG, MAX_COMPRESSED_FRAGS);
+	adev->gfx.config.gb_addr_config_fields.num_rb_per_se = 1 <<
+			REG_GET_FIELD(adev->gfx.config.gb_addr_config,
+				      GB_ADDR_CONFIG, NUM_RB_PER_SE);
+	adev->gfx.config.gb_addr_config_fields.num_se = 1 <<
+			REG_GET_FIELD(adev->gfx.config.gb_addr_config,
+				      GB_ADDR_CONFIG, NUM_SHADER_ENGINES);
+	adev->gfx.config.gb_addr_config_fields.pipe_interleave_size = 1 << (8 +
+			REG_GET_FIELD(adev->gfx.config.gb_addr_config,
+				      GB_ADDR_CONFIG, PIPE_INTERLEAVE_SIZE));
+
+	return 0;
+}
+
+static void gfx_v12_0_disable_gpa_mode(struct amdgpu_device *adev)
+{
+	uint32_t data;
+
+	data = RREG32_SOC15(GC, 0, regCPC_PSP_DEBUG);
+	data |= CPC_PSP_DEBUG__GPA_OVERRIDE_MASK;
+	WREG32_SOC15(GC, 0, regCPC_PSP_DEBUG, data);
+
+	data = RREG32_SOC15(GC, 0, regCPG_PSP_DEBUG);
+	data |= CPG_PSP_DEBUG__GPA_OVERRIDE_MASK;
+	WREG32_SOC15(GC, 0, regCPG_PSP_DEBUG, data);
+}
+
+static void gfx_v12_0_init_golden_registers(struct amdgpu_device *adev)
+{
+	if (amdgpu_sriov_vf(adev))
+		return;
+
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	case IP_VERSION(12, 0, 0):
+	case IP_VERSION(12, 0, 1):
+		soc15_program_register_sequence(adev,
+						golden_settings_gc_12_0,
+						(const u32)ARRAY_SIZE(golden_settings_gc_12_0));
+
+		if (adev->rev_id == 0)
+			soc15_program_register_sequence(adev,
+					golden_settings_gc_12_0_rev0,
+					(const u32)ARRAY_SIZE(golden_settings_gc_12_0_rev0));
+		break;
+	default:
+		break;
+	}
+}
+
+static int gfx_v12_0_hw_init(struct amdgpu_ip_block *ip_block)
+{
+	int r;
+	struct amdgpu_device *adev = ip_block->adev;
+
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_RLC_BACKDOOR_AUTO) {
+		if (adev->gfx.imu.funcs && (amdgpu_dpm > 0)) {
+			/* RLC autoload sequence 1: Program rlc ram */
+			if (adev->gfx.imu.funcs->program_rlc_ram)
+				adev->gfx.imu.funcs->program_rlc_ram(adev);
+		}
+		/* rlc autoload firmware */
+		r = gfx_v12_0_rlc_backdoor_autoload_enable(adev);
+		if (r)
+			return r;
+	} else {
+		if (adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT) {
+			if (adev->gfx.imu.funcs && (amdgpu_dpm > 0)) {
+				if (adev->gfx.imu.funcs->load_microcode)
+					adev->gfx.imu.funcs->load_microcode(adev);
+				if (adev->gfx.imu.funcs->setup_imu)
+					adev->gfx.imu.funcs->setup_imu(adev);
+				if (adev->gfx.imu.funcs->start_imu)
+					adev->gfx.imu.funcs->start_imu(adev);
+			}
+
+			/* disable gpa mode in backdoor loading */
+			gfx_v12_0_disable_gpa_mode(adev);
+		}
+	}
+
+	if ((adev->firmware.load_type == AMDGPU_FW_LOAD_RLC_BACKDOOR_AUTO) ||
+	    (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP)) {
+		r = gfx_v12_0_wait_for_rlc_autoload_complete(adev);
+		if (r) {
+			dev_err(adev->dev, "(%d) failed to wait rlc autoload complete\n", r);
+			return r;
+		}
+	}
+
+	if (!amdgpu_emu_mode)
+		gfx_v12_0_init_golden_registers(adev);
+
+	adev->gfx.is_poweron = true;
+
+	if (get_gb_addr_config(adev))
+		DRM_WARN("Invalid gb_addr_config !\n");
+
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP)
+		gfx_v12_0_config_gfx_rs64(adev);
+
+	r = gfx_v12_0_gfxhub_enable(adev);
+	if (r)
+		return r;
+
+	if ((adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT ||
+	     adev->firmware.load_type == AMDGPU_FW_LOAD_RLC_BACKDOOR_AUTO) &&
+	     (amdgpu_dpm == 1)) {
+		/**
+		 * For gfx 12, rlc firmware loading relies on smu firmware is
+		 * loaded firstly, so in direct type, it has to load smc ucode
+		 * here before rlc.
+		 */
+		r = amdgpu_pm_load_smu_firmware(adev, NULL);
+		if (r)
+			return r;
+	}
+
+	gfx_v12_0_constants_init(adev);
+
+	if (adev->nbio.funcs->gc_doorbell_init)
+		adev->nbio.funcs->gc_doorbell_init(adev);
+
+	r = gfx_v12_0_rlc_resume(adev);
+	if (r)
+		return r;
+
+	/*
+	 * init golden registers and rlc resume may override some registers,
+	 * reconfig them here
+	 */
+	gfx_v12_0_tcp_harvest(adev);
+
+	r = gfx_v12_0_cp_resume(adev);
+	if (r)
+		return r;
+
+	return r;
+}
+
+static int gfx_v12_0_set_userq_eop_interrupts(struct amdgpu_device *adev,
+					      bool enable)
+{
+	unsigned int irq_type;
+	int m, p, r;
+
+	if (adev->userq_funcs[AMDGPU_HW_IP_GFX]) {
+		for (m = 0; m < adev->gfx.me.num_me; m++) {
+			for (p = 0; p < adev->gfx.me.num_pipe_per_me; p++) {
+				irq_type = AMDGPU_CP_IRQ_GFX_ME0_PIPE0_EOP + p;
+				if (enable)
+					r = amdgpu_irq_get(adev, &adev->gfx.eop_irq,
+							   irq_type);
+				else
+					r = amdgpu_irq_put(adev, &adev->gfx.eop_irq,
+							   irq_type);
+				if (r)
+					return r;
+			}
+		}
+	}
+
+	if (adev->userq_funcs[AMDGPU_HW_IP_COMPUTE]) {
+		for (m = 0; m < adev->gfx.mec.num_mec; ++m) {
+			for (p = 0; p < adev->gfx.mec.num_pipe_per_mec; p++) {
+				irq_type = AMDGPU_CP_IRQ_COMPUTE_MEC1_PIPE0_EOP
+					+ (m * adev->gfx.mec.num_pipe_per_mec)
+					+ p;
+				if (enable)
+					r = amdgpu_irq_get(adev, &adev->gfx.eop_irq,
+							   irq_type);
+				else
+					r = amdgpu_irq_put(adev, &adev->gfx.eop_irq,
+							   irq_type);
+				if (r)
+					return r;
+			}
+		}
+	}
+
+	return 0;
+}
+
+static int gfx_v12_0_hw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	uint32_t tmp;
+
+	cancel_delayed_work_sync(&adev->gfx.idle_work);
+
+	amdgpu_irq_put(adev, &adev->gfx.priv_reg_irq, 0);
+	amdgpu_irq_put(adev, &adev->gfx.priv_inst_irq, 0);
+	amdgpu_irq_put(adev, &adev->gfx.bad_op_irq, 0);
+	gfx_v12_0_set_userq_eop_interrupts(adev, false);
+
+	if (!adev->no_hw_access) {
+		if (amdgpu_async_gfx_ring) {
+			if (amdgpu_gfx_disable_kgq(adev, 0))
+				DRM_ERROR("KGQ disable failed\n");
+		}
+
+		if (amdgpu_gfx_disable_kcq(adev, 0))
+			DRM_ERROR("KCQ disable failed\n");
+
+		amdgpu_mes_kiq_hw_fini(adev);
+	}
+
+	if (amdgpu_sriov_vf(adev)) {
+		gfx_v12_0_cp_gfx_enable(adev, false);
+		/* Program KIQ position of RLC_CP_SCHEDULERS during destroy */
+		tmp = RREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS);
+		tmp &= 0xffffff00;
+		WREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS, tmp);
+
+		return 0;
+	}
+	gfx_v12_0_cp_enable(adev, false);
+	gfx_v12_0_enable_gui_idle_interrupt(adev, false);
+
+	adev->gfxhub.funcs->gart_disable(adev);
+
+	adev->gfx.is_poweron = false;
+
+	return 0;
+}
+
+static int gfx_v12_0_suspend(struct amdgpu_ip_block *ip_block)
+{
+	return gfx_v12_0_hw_fini(ip_block);
+}
+
+static int gfx_v12_0_resume(struct amdgpu_ip_block *ip_block)
+{
+	return gfx_v12_0_hw_init(ip_block);
+}
+
+static bool gfx_v12_0_is_idle(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	if (REG_GET_FIELD(RREG32_SOC15(GC, 0, regGRBM_STATUS),
+				GRBM_STATUS, GUI_ACTIVE))
+		return false;
+	else
+		return true;
+}
+
+static int gfx_v12_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
+{
+	unsigned i;
+	u32 tmp;
+	struct amdgpu_device *adev = ip_block->adev;
+
+	for (i = 0; i < adev->usec_timeout; i++) {
+		/* read MC_STATUS */
+		tmp = RREG32_SOC15(GC, 0, regGRBM_STATUS) &
+			GRBM_STATUS__GUI_ACTIVE_MASK;
+
+		if (!REG_GET_FIELD(tmp, GRBM_STATUS, GUI_ACTIVE))
+			return 0;
+		udelay(1);
+	}
+	return -ETIMEDOUT;
+}
+
+static uint64_t gfx_v12_0_get_gpu_clock_counter(struct amdgpu_device *adev)
+{
+	uint64_t clock = 0;
+
+	if (adev->smuio.funcs &&
+	    adev->smuio.funcs->get_gpu_clock_counter)
+		clock = adev->smuio.funcs->get_gpu_clock_counter(adev);
+	else
+		dev_warn(adev->dev, "query gpu clock counter is not supported\n");
+
+	return clock;
+}
+
+static int gfx_v12_0_early_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	switch (amdgpu_user_queue) {
+	case -1:
+	case 0:
+	default:
+		adev->gfx.disable_kq = false;
+		adev->gfx.disable_uq = true;
+		break;
+	case 1:
+		adev->gfx.disable_kq = false;
+		adev->gfx.disable_uq = false;
+		break;
+	case 2:
+		adev->gfx.disable_kq = true;
+		adev->gfx.disable_uq = false;
+		break;
+	}
+
+	adev->gfx.funcs = &gfx_v12_0_gfx_funcs;
+
+	if (adev->gfx.disable_kq) {
+		adev->gfx.num_gfx_rings = 0;
+		adev->gfx.num_compute_rings = 0;
+	} else {
+		adev->gfx.num_gfx_rings = GFX12_NUM_GFX_RINGS;
+		adev->gfx.num_compute_rings = min(amdgpu_gfx_get_num_kcq(adev),
+						  AMDGPU_MAX_COMPUTE_RINGS);
+	}
+
+	gfx_v12_0_set_kiq_pm4_funcs(adev);
+	gfx_v12_0_set_ring_funcs(adev);
+	gfx_v12_0_set_irq_funcs(adev);
+	gfx_v12_0_set_rlc_funcs(adev);
+	gfx_v12_0_set_mqd_funcs(adev);
+	gfx_v12_0_set_imu_funcs(adev);
+
+	gfx_v12_0_init_rlcg_reg_access_ctrl(adev);
+
+	return gfx_v12_0_init_microcode(adev);
+}
+
+static int gfx_v12_0_late_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int r;
+
+	r = amdgpu_irq_get(adev, &adev->gfx.priv_reg_irq, 0);
+	if (r)
+		return r;
+
+	r = amdgpu_irq_get(adev, &adev->gfx.priv_inst_irq, 0);
+	if (r)
+		return r;
+
+	r = amdgpu_irq_get(adev, &adev->gfx.bad_op_irq, 0);
+	if (r)
+		return r;
+
+	r = gfx_v12_0_set_userq_eop_interrupts(adev, true);
+	if (r)
+		return r;
+
+	return 0;
+}
+
+static bool gfx_v12_0_is_rlc_enabled(struct amdgpu_device *adev)
+{
+	uint32_t rlc_cntl;
+
+	/* if RLC is not enabled, do nothing */
+	rlc_cntl = RREG32_SOC15(GC, 0, regRLC_CNTL);
+	return (REG_GET_FIELD(rlc_cntl, RLC_CNTL, RLC_ENABLE_F32)) ? true : false;
+}
+
+static void gfx_v12_0_set_safe_mode(struct amdgpu_device *adev,
+				    int xcc_id)
+{
+	uint32_t data;
+	unsigned i;
+
+	data = RLC_SAFE_MODE__CMD_MASK;
+	data |= (1 << RLC_SAFE_MODE__MESSAGE__SHIFT);
+
+	WREG32_SOC15(GC, 0, regRLC_SAFE_MODE, data);
+
+	/* wait for RLC_SAFE_MODE */
+	for (i = 0; i < adev->usec_timeout; i++) {
+		if (!REG_GET_FIELD(RREG32_SOC15(GC, 0, regRLC_SAFE_MODE),
+				   RLC_SAFE_MODE, CMD))
+			break;
+		udelay(1);
+	}
+}
+
+static void gfx_v12_0_unset_safe_mode(struct amdgpu_device *adev,
+				      int xcc_id)
+{
+	WREG32_SOC15(GC, 0, regRLC_SAFE_MODE, RLC_SAFE_MODE__CMD_MASK);
+}
+
+static void gfx_v12_0_update_perf_clk(struct amdgpu_device *adev,
+				      bool enable)
+{
+	uint32_t def, data;
+
+	if (!(adev->cg_flags & AMD_CG_SUPPORT_GFX_PERF_CLK))
+		return;
+
+	def = data = RREG32_SOC15(GC, 0, regRLC_CGTT_MGCG_OVERRIDE);
+
+	if (enable)
+		data &= ~RLC_CGTT_MGCG_OVERRIDE__PERFMON_CLOCK_STATE_MASK;
+	else
+		data |= RLC_CGTT_MGCG_OVERRIDE__PERFMON_CLOCK_STATE_MASK;
+
+	if (def != data)
+		WREG32_SOC15(GC, 0, regRLC_CGTT_MGCG_OVERRIDE, data);
+}
+
+static void gfx_v12_0_update_spm_vmid(struct amdgpu_device *adev,
+				      struct amdgpu_ring *ring,
+				      unsigned vmid)
+{
+	u32 reg, data;
+
+	reg = SOC15_REG_OFFSET(GC, 0, regRLC_SPM_MC_CNTL);
+	if (amdgpu_sriov_is_pp_one_vf(adev))
+		data = RREG32_NO_KIQ(reg);
+	else
+		data = RREG32(reg);
+
+	data &= ~RLC_SPM_MC_CNTL__RLC_SPM_VMID_MASK;
+	data |= (vmid & RLC_SPM_MC_CNTL__RLC_SPM_VMID_MASK) << RLC_SPM_MC_CNTL__RLC_SPM_VMID__SHIFT;
+
+	if (amdgpu_sriov_is_pp_one_vf(adev))
+		WREG32_SOC15_NO_KIQ(GC, 0, regRLC_SPM_MC_CNTL, data);
+	else
+		WREG32_SOC15(GC, 0, regRLC_SPM_MC_CNTL, data);
+
+	if (ring
+	    && amdgpu_sriov_is_pp_one_vf(adev)
+	    && ((ring->funcs->type == AMDGPU_RING_TYPE_GFX)
+		|| (ring->funcs->type == AMDGPU_RING_TYPE_COMPUTE))) {
+		uint32_t reg = SOC15_REG_OFFSET(GC, 0, regRLC_SPM_MC_CNTL);
+		amdgpu_ring_emit_wreg(ring, reg, data);
+	}
+}
+
+static const struct amdgpu_rlc_funcs gfx_v12_0_rlc_funcs = {
+	.is_rlc_enabled = gfx_v12_0_is_rlc_enabled,
+	.set_safe_mode = gfx_v12_0_set_safe_mode,
+	.unset_safe_mode = gfx_v12_0_unset_safe_mode,
+	.init = gfx_v12_0_rlc_init,
+	.get_csb_size = gfx_v12_0_get_csb_size,
+	.get_csb_buffer = gfx_v12_0_get_csb_buffer,
+	.resume = gfx_v12_0_rlc_resume,
+	.stop = gfx_v12_0_rlc_stop,
+	.reset = gfx_v12_0_rlc_reset,
+	.start = gfx_v12_0_rlc_start,
+	.update_spm_vmid = gfx_v12_0_update_spm_vmid,
+};
+
+#if 0
+static void gfx_v12_cntl_power_gating(struct amdgpu_device *adev, bool enable)
+{
+	/* TODO */
+}
+
+static void gfx_v12_cntl_pg(struct amdgpu_device *adev, bool enable)
+{
+	/* TODO */
+}
+#endif
+
+static int gfx_v12_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
+					   enum amd_powergating_state state)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	bool enable = (state == AMD_PG_STATE_GATE);
+
+	if (amdgpu_sriov_vf(adev))
+		return 0;
+
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	case IP_VERSION(12, 0, 0):
+	case IP_VERSION(12, 0, 1):
+		amdgpu_gfx_off_ctrl(adev, enable);
+		break;
+	default:
+		break;
+	}
+
+	return 0;
+}
+
+static void gfx_v12_0_update_coarse_grain_clock_gating(struct amdgpu_device *adev,
+						       bool enable)
+{
+	uint32_t def, data;
+
+	if (!(adev->cg_flags &
+	      (AMD_CG_SUPPORT_GFX_CGCG |
+	      AMD_CG_SUPPORT_GFX_CGLS |
+	      AMD_CG_SUPPORT_GFX_3D_CGCG |
+	      AMD_CG_SUPPORT_GFX_3D_CGLS)))
+		return;
+
+	if (enable) {
+		def = data = RREG32_SOC15(GC, 0, regRLC_CGTT_MGCG_OVERRIDE);
+
+		/* unset CGCG override */
+		if (adev->cg_flags & AMD_CG_SUPPORT_GFX_CGCG)
+			data &= ~RLC_CGTT_MGCG_OVERRIDE__GFXIP_CGCG_OVERRIDE_MASK;
+		if (adev->cg_flags & AMD_CG_SUPPORT_GFX_CGLS)
+			data &= ~RLC_CGTT_MGCG_OVERRIDE__GFXIP_CGLS_OVERRIDE_MASK;
+		if (adev->cg_flags & AMD_CG_SUPPORT_GFX_3D_CGCG ||
+		    adev->cg_flags & AMD_CG_SUPPORT_GFX_3D_CGLS)
+			data &= ~RLC_CGTT_MGCG_OVERRIDE__GFXIP_GFX3D_CG_OVERRIDE_MASK;
+
+		/* update CGCG override bits */
+		if (def != data)
+			WREG32_SOC15(GC, 0, regRLC_CGTT_MGCG_OVERRIDE, data);
+
+		/* enable cgcg FSM(0x0000363F) */
+		def = data = RREG32_SOC15(GC, 0, regRLC_CGCG_CGLS_CTRL);
+
+		if (adev->cg_flags & AMD_CG_SUPPORT_GFX_CGCG) {
+			data &= ~RLC_CGCG_CGLS_CTRL__CGCG_GFX_IDLE_THRESHOLD_MASK;
+			data |= (0x36 << RLC_CGCG_CGLS_CTRL__CGCG_GFX_IDLE_THRESHOLD__SHIFT) |
+				 RLC_CGCG_CGLS_CTRL__CGCG_EN_MASK;
+		}
+
+		if (adev->cg_flags & AMD_CG_SUPPORT_GFX_CGLS) {
+			data &= ~RLC_CGCG_CGLS_CTRL__CGLS_REP_COMPANSAT_DELAY_MASK;
+			data |= (0x000F << RLC_CGCG_CGLS_CTRL__CGLS_REP_COMPANSAT_DELAY__SHIFT) |
+				 RLC_CGCG_CGLS_CTRL__CGLS_EN_MASK;
+		}
+
+		if (def != data)
+			WREG32_SOC15(GC, 0, regRLC_CGCG_CGLS_CTRL, data);
+
+		/* Program RLC_CGCG_CGLS_CTRL_3D */
+		def = data = RREG32_SOC15(GC, 0, regRLC_CGCG_CGLS_CTRL_3D);
+
+		if (adev->cg_flags & AMD_CG_SUPPORT_GFX_3D_CGCG) {
+			data &= ~RLC_CGCG_CGLS_CTRL_3D__CGCG_GFX_IDLE_THRESHOLD_MASK;
+			data |= (0x36 << RLC_CGCG_CGLS_CTRL_3D__CGCG_GFX_IDLE_THRESHOLD__SHIFT) |
+				 RLC_CGCG_CGLS_CTRL_3D__CGCG_EN_MASK;
+		}
+
+		if (adev->cg_flags & AMD_CG_SUPPORT_GFX_3D_CGLS) {
+			data &= ~RLC_CGCG_CGLS_CTRL_3D__CGLS_REP_COMPANSAT_DELAY_MASK;
+			data |= (0xf << RLC_CGCG_CGLS_CTRL_3D__CGLS_REP_COMPANSAT_DELAY__SHIFT) |
+				 RLC_CGCG_CGLS_CTRL_3D__CGLS_EN_MASK;
+		}
+
+		if (def != data)
+			WREG32_SOC15(GC, 0, regRLC_CGCG_CGLS_CTRL_3D, data);
+
+		/* set IDLE_POLL_COUNT(0x00900100) */
+		def = data = RREG32_SOC15(GC, 0, regCP_RB_WPTR_POLL_CNTL);
+
+		data &= ~(CP_RB_WPTR_POLL_CNTL__POLL_FREQUENCY_MASK | CP_RB_WPTR_POLL_CNTL__IDLE_POLL_COUNT_MASK);
+		data |= (0x0100 << CP_RB_WPTR_POLL_CNTL__POLL_FREQUENCY__SHIFT) |
+			(0x0090 << CP_RB_WPTR_POLL_CNTL__IDLE_POLL_COUNT__SHIFT);
+
+		if (def != data)
+			WREG32_SOC15(GC, 0, regCP_RB_WPTR_POLL_CNTL, data);
+
+		data = RREG32_SOC15(GC, 0, regCP_INT_CNTL);
+		data = REG_SET_FIELD(data, CP_INT_CNTL, CNTX_BUSY_INT_ENABLE, 1);
+		data = REG_SET_FIELD(data, CP_INT_CNTL, CNTX_EMPTY_INT_ENABLE, 1);
+		data = REG_SET_FIELD(data, CP_INT_CNTL, CMP_BUSY_INT_ENABLE, 1);
+		data = REG_SET_FIELD(data, CP_INT_CNTL, GFX_IDLE_INT_ENABLE, 1);
+		WREG32_SOC15(GC, 0, regCP_INT_CNTL, data);
+
+		data = RREG32_SOC15(GC, 0, regSDMA0_RLC_CGCG_CTRL);
+		data = REG_SET_FIELD(data, SDMA0_RLC_CGCG_CTRL, CGCG_INT_ENABLE, 1);
+		WREG32_SOC15(GC, 0, regSDMA0_RLC_CGCG_CTRL, data);
+
+		/* Some ASICs only have one SDMA instance, not need to configure SDMA1 */
+		if (adev->sdma.num_instances > 1) {
+			data = RREG32_SOC15(GC, 0, regSDMA1_RLC_CGCG_CTRL);
+			data = REG_SET_FIELD(data, SDMA1_RLC_CGCG_CTRL, CGCG_INT_ENABLE, 1);
+			WREG32_SOC15(GC, 0, regSDMA1_RLC_CGCG_CTRL, data);
+		}
+	} else {
+		/* Program RLC_CGCG_CGLS_CTRL */
+		def = data = RREG32_SOC15(GC, 0, regRLC_CGCG_CGLS_CTRL);
+
+		if (adev->cg_flags & AMD_CG_SUPPORT_GFX_CGCG)
+			data &= ~RLC_CGCG_CGLS_CTRL__CGCG_EN_MASK;
+
+		if (adev->cg_flags & AMD_CG_SUPPORT_GFX_CGLS)
+			data &= ~RLC_CGCG_CGLS_CTRL__CGLS_EN_MASK;
+
+		if (def != data)
+			WREG32_SOC15(GC, 0, regRLC_CGCG_CGLS_CTRL, data);
+
+		/* Program RLC_CGCG_CGLS_CTRL_3D */
+		def = data = RREG32_SOC15(GC, 0, regRLC_CGCG_CGLS_CTRL_3D);
+
+		if (adev->cg_flags & AMD_CG_SUPPORT_GFX_3D_CGCG)
+			data &= ~RLC_CGCG_CGLS_CTRL_3D__CGCG_EN_MASK;
+		if (adev->cg_flags & AMD_CG_SUPPORT_GFX_3D_CGLS)
+			data &= ~RLC_CGCG_CGLS_CTRL_3D__CGLS_EN_MASK;
+
+		if (def != data)
+			WREG32_SOC15(GC, 0, regRLC_CGCG_CGLS_CTRL_3D, data);
+	}
+}
+
+static void gfx_v12_0_update_medium_grain_clock_gating(struct amdgpu_device *adev,
+						       bool enable)
+{
+	uint32_t data, def;
+	if (!(adev->cg_flags & (AMD_CG_SUPPORT_GFX_MGCG | AMD_CG_SUPPORT_GFX_MGLS)))
+		return;
+
+	/* It is disabled by HW by default */
+	if (enable) {
+		if (adev->cg_flags & AMD_CG_SUPPORT_GFX_MGCG) {
+			/* 1 - RLC_CGTT_MGCG_OVERRIDE */
+			def = data = RREG32_SOC15(GC, 0, regRLC_CGTT_MGCG_OVERRIDE);
+
+			data &= ~(RLC_CGTT_MGCG_OVERRIDE__GRBM_CGTT_SCLK_OVERRIDE_MASK |
+				  RLC_CGTT_MGCG_OVERRIDE__RLC_CGTT_SCLK_OVERRIDE_MASK |
+				  RLC_CGTT_MGCG_OVERRIDE__GFXIP_MGCG_OVERRIDE_MASK);
+
+			if (def != data)
+				WREG32_SOC15(GC, 0, regRLC_CGTT_MGCG_OVERRIDE, data);
+		}
+	} else {
+		if (adev->cg_flags & AMD_CG_SUPPORT_GFX_MGCG) {
+			def = data = RREG32_SOC15(GC, 0, regRLC_CGTT_MGCG_OVERRIDE);
+
+			data |= (RLC_CGTT_MGCG_OVERRIDE__RLC_CGTT_SCLK_OVERRIDE_MASK |
+				 RLC_CGTT_MGCG_OVERRIDE__GRBM_CGTT_SCLK_OVERRIDE_MASK |
+				 RLC_CGTT_MGCG_OVERRIDE__GFXIP_MGCG_OVERRIDE_MASK);
+
+			if (def != data)
+				WREG32_SOC15(GC, 0, regRLC_CGTT_MGCG_OVERRIDE, data);
+		}
+	}
+}
+
+static void gfx_v12_0_update_repeater_fgcg(struct amdgpu_device *adev,
+					   bool enable)
+{
+	uint32_t def, data;
+
+	if (!(adev->cg_flags & AMD_CG_SUPPORT_REPEATER_FGCG))
+		return;
+
+	def = data = RREG32_SOC15(GC, 0, regRLC_CGTT_MGCG_OVERRIDE);
+
+	if (enable)
+		data &= ~(RLC_CGTT_MGCG_OVERRIDE__GFXIP_REPEATER_FGCG_OVERRIDE_MASK |
+				  RLC_CGTT_MGCG_OVERRIDE__RLC_REPEATER_FGCG_OVERRIDE_MASK);
+	else
+		data |= RLC_CGTT_MGCG_OVERRIDE__GFXIP_REPEATER_FGCG_OVERRIDE_MASK |
+				RLC_CGTT_MGCG_OVERRIDE__RLC_REPEATER_FGCG_OVERRIDE_MASK;
+
+	if (def != data)
+		WREG32_SOC15(GC, 0, regRLC_CGTT_MGCG_OVERRIDE, data);
+}
+
+static void gfx_v12_0_update_sram_fgcg(struct amdgpu_device *adev,
+				       bool enable)
+{
+	uint32_t def, data;
+
+	if (!(adev->cg_flags & AMD_CG_SUPPORT_GFX_FGCG))
+		return;
+
+	def = data = RREG32_SOC15(GC, 0, regRLC_CGTT_MGCG_OVERRIDE);
+
+	if (enable)
+		data &= ~RLC_CGTT_MGCG_OVERRIDE__GFXIP_FGCG_OVERRIDE_MASK;
+	else
+		data |= RLC_CGTT_MGCG_OVERRIDE__GFXIP_FGCG_OVERRIDE_MASK;
+
+	if (def != data)
+		WREG32_SOC15(GC, 0, regRLC_CGTT_MGCG_OVERRIDE, data);
+}
+
+static int gfx_v12_0_update_gfx_clock_gating(struct amdgpu_device *adev,
+					    bool enable)
+{
+	amdgpu_gfx_rlc_enter_safe_mode(adev, 0);
+
+	gfx_v12_0_update_coarse_grain_clock_gating(adev, enable);
+
+	gfx_v12_0_update_medium_grain_clock_gating(adev, enable);
+
+	gfx_v12_0_update_repeater_fgcg(adev, enable);
+
+	gfx_v12_0_update_sram_fgcg(adev, enable);
+
+	gfx_v12_0_update_perf_clk(adev, enable);
+
+	if (adev->cg_flags &
+	    (AMD_CG_SUPPORT_GFX_MGCG |
+	     AMD_CG_SUPPORT_GFX_CGLS |
+	     AMD_CG_SUPPORT_GFX_CGCG |
+	     AMD_CG_SUPPORT_GFX_3D_CGCG |
+	     AMD_CG_SUPPORT_GFX_3D_CGLS))
+		gfx_v12_0_enable_gui_idle_interrupt(adev, enable);
+
+	amdgpu_gfx_rlc_exit_safe_mode(adev, 0);
+
+	return 0;
+}
+
+static int gfx_v12_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
+					   enum amd_clockgating_state state)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	if (amdgpu_sriov_vf(adev))
+		return 0;
+
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	case IP_VERSION(12, 0, 0):
+	case IP_VERSION(12, 0, 1):
+		gfx_v12_0_update_gfx_clock_gating(adev,
+						  state == AMD_CG_STATE_GATE);
+		break;
+	default:
+		break;
+	}
+
+	return 0;
+}
+
+static void gfx_v12_0_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int data;
+
+	/* AMD_CG_SUPPORT_GFX_MGCG */
+	data = RREG32_SOC15(GC, 0, regRLC_CGTT_MGCG_OVERRIDE);
+	if (!(data & RLC_CGTT_MGCG_OVERRIDE__GFXIP_MGCG_OVERRIDE_MASK))
+		*flags |= AMD_CG_SUPPORT_GFX_MGCG;
+
+	/* AMD_CG_SUPPORT_REPEATER_FGCG */
+	if (!(data & RLC_CGTT_MGCG_OVERRIDE__GFXIP_REPEATER_FGCG_OVERRIDE_MASK))
+		*flags |= AMD_CG_SUPPORT_REPEATER_FGCG;
+
+	/* AMD_CG_SUPPORT_GFX_FGCG */
+	if (!(data & RLC_CGTT_MGCG_OVERRIDE__GFXIP_FGCG_OVERRIDE_MASK))
+		*flags |= AMD_CG_SUPPORT_GFX_FGCG;
+
+	/* AMD_CG_SUPPORT_GFX_PERF_CLK */
+	if (!(data & RLC_CGTT_MGCG_OVERRIDE__PERFMON_CLOCK_STATE_MASK))
+		*flags |= AMD_CG_SUPPORT_GFX_PERF_CLK;
+
+	/* AMD_CG_SUPPORT_GFX_CGCG */
+	data = RREG32_SOC15(GC, 0, regRLC_CGCG_CGLS_CTRL);
+	if (data & RLC_CGCG_CGLS_CTRL__CGCG_EN_MASK)
+		*flags |= AMD_CG_SUPPORT_GFX_CGCG;
+
+	/* AMD_CG_SUPPORT_GFX_CGLS */
+	if (data & RLC_CGCG_CGLS_CTRL__CGLS_EN_MASK)
+		*flags |= AMD_CG_SUPPORT_GFX_CGLS;
+
+	/* AMD_CG_SUPPORT_GFX_3D_CGCG */
+	data = RREG32_SOC15(GC, 0, regRLC_CGCG_CGLS_CTRL_3D);
+	if (data & RLC_CGCG_CGLS_CTRL_3D__CGCG_EN_MASK)
+		*flags |= AMD_CG_SUPPORT_GFX_3D_CGCG;
+
+	/* AMD_CG_SUPPORT_GFX_3D_CGLS */
+	if (data & RLC_CGCG_CGLS_CTRL_3D__CGLS_EN_MASK)
+		*flags |= AMD_CG_SUPPORT_GFX_3D_CGLS;
+}
+
+static u64 gfx_v12_0_ring_get_rptr_gfx(struct amdgpu_ring *ring)
+{
+	/* gfx12 is 32bit rptr*/
+	return *(uint32_t *)ring->rptr_cpu_addr;
+}
+
+static u64 gfx_v12_0_ring_get_wptr_gfx(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	u64 wptr;
+
+	/* XXX check if swapping is necessary on BE */
+	if (ring->use_doorbell) {
+		wptr = atomic64_read((atomic64_t *)ring->wptr_cpu_addr);
+	} else {
+		wptr = RREG32_SOC15(GC, 0, regCP_RB0_WPTR);
+		wptr += (u64)RREG32_SOC15(GC, 0, regCP_RB0_WPTR_HI) << 32;
+	}
+
+	return wptr;
+}
+
+static void gfx_v12_0_ring_set_wptr_gfx(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	if (ring->use_doorbell) {
+		/* XXX check if swapping is necessary on BE */
+		atomic64_set((atomic64_t *)ring->wptr_cpu_addr,
+			     ring->wptr);
+		WDOORBELL64(ring->doorbell_index, ring->wptr);
+	} else {
+		WREG32_SOC15(GC, 0, regCP_RB0_WPTR,
+			     lower_32_bits(ring->wptr));
+		WREG32_SOC15(GC, 0, regCP_RB0_WPTR_HI,
+			     upper_32_bits(ring->wptr));
+	}
+}
+
+static u64 gfx_v12_0_ring_get_rptr_compute(struct amdgpu_ring *ring)
+{
+	/* gfx12 hardware is 32bit rptr */
+	return *(uint32_t *)ring->rptr_cpu_addr;
+}
+
+static u64 gfx_v12_0_ring_get_wptr_compute(struct amdgpu_ring *ring)
+{
+	u64 wptr;
+
+	/* XXX check if swapping is necessary on BE */
+	if (ring->use_doorbell)
+		wptr = atomic64_read((atomic64_t *)ring->wptr_cpu_addr);
+	else
+		BUG();
+	return wptr;
+}
+
+static void gfx_v12_0_ring_set_wptr_compute(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	/* XXX check if swapping is necessary on BE */
+	if (ring->use_doorbell) {
+		atomic64_set((atomic64_t *)ring->wptr_cpu_addr,
+			     ring->wptr);
+		WDOORBELL64(ring->doorbell_index, ring->wptr);
+	} else {
+		BUG(); /* only DOORBELL method supported on gfx12 now */
+	}
+}
+
+static void gfx_v12_0_ring_emit_hdp_flush(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	u32 ref_and_mask, reg_mem_engine;
+	const struct nbio_hdp_flush_reg *nbio_hf_reg = adev->nbio.hdp_flush_reg;
+
+	if (ring->funcs->type == AMDGPU_RING_TYPE_COMPUTE) {
+		switch (ring->me) {
+		case 1:
+			ref_and_mask = nbio_hf_reg->ref_and_mask_cp2 << ring->pipe;
+			break;
+		case 2:
+			ref_and_mask = nbio_hf_reg->ref_and_mask_cp6 << ring->pipe;
+			break;
+		default:
+			return;
+		}
+		reg_mem_engine = 0;
+	} else {
+		ref_and_mask = nbio_hf_reg->ref_and_mask_cp0;
+		reg_mem_engine = 1; /* pfp */
+	}
+
+	gfx_v12_0_wait_reg_mem(ring, reg_mem_engine, 0, 1,
+			       adev->nbio.funcs->get_hdp_flush_req_offset(adev),
+			       adev->nbio.funcs->get_hdp_flush_done_offset(adev),
+			       ref_and_mask, ref_and_mask, 0x20);
+}
+
+static void gfx_v12_0_ring_emit_ib_gfx(struct amdgpu_ring *ring,
+				       struct amdgpu_job *job,
+				       struct amdgpu_ib *ib,
+				       uint32_t flags)
+{
+	unsigned vmid = AMDGPU_JOB_GET_VMID(job);
+	u32 header, control = 0;
+
+	header = PACKET3(PACKET3_INDIRECT_BUFFER, 2);
+
+	control |= ib->length_dw | (vmid << 24);
+
+	amdgpu_ring_write(ring, header);
+	BUG_ON(ib->gpu_addr & 0x3); /* Dword align */
+	amdgpu_ring_write(ring,
+#ifdef __BIG_ENDIAN
+		(2 << 0) |
+#endif
+		lower_32_bits(ib->gpu_addr));
+	amdgpu_ring_write(ring, upper_32_bits(ib->gpu_addr));
+	amdgpu_ring_write(ring, control);
+}
+
+static void gfx_v12_0_ring_emit_ib_compute(struct amdgpu_ring *ring,
+					   struct amdgpu_job *job,
+					   struct amdgpu_ib *ib,
+					   uint32_t flags)
+{
+	unsigned vmid = AMDGPU_JOB_GET_VMID(job);
+	u32 control = INDIRECT_BUFFER_VALID | ib->length_dw | (vmid << 24);
+
+	amdgpu_ring_write(ring, PACKET3(PACKET3_INDIRECT_BUFFER, 2));
+	BUG_ON(ib->gpu_addr & 0x3); /* Dword align */
+	amdgpu_ring_write(ring,
+#ifdef __BIG_ENDIAN
+				(2 << 0) |
+#endif
+				lower_32_bits(ib->gpu_addr));
+	amdgpu_ring_write(ring, upper_32_bits(ib->gpu_addr));
+	amdgpu_ring_write(ring, control);
+}
+
+static void gfx_v12_0_ring_emit_fence(struct amdgpu_ring *ring, u64 addr,
+				     u64 seq, unsigned flags)
+{
+	bool write64bit = flags & AMDGPU_FENCE_FLAG_64BIT;
+	bool int_sel = flags & AMDGPU_FENCE_FLAG_INT;
+
+	/* RELEASE_MEM - flush caches, send int */
+	amdgpu_ring_write(ring, PACKET3(PACKET3_RELEASE_MEM, 6));
+	amdgpu_ring_write(ring, (PACKET3_RELEASE_MEM_GCR_SEQ |
+				 PACKET3_RELEASE_MEM_GCR_GL2_WB |
+				 PACKET3_RELEASE_MEM_CACHE_POLICY(3) |
+				 PACKET3_RELEASE_MEM_EVENT_TYPE(CACHE_FLUSH_AND_INV_TS_EVENT) |
+				 PACKET3_RELEASE_MEM_EVENT_INDEX(5)));
+	amdgpu_ring_write(ring, (PACKET3_RELEASE_MEM_DATA_SEL(write64bit ? 2 : 1) |
+				 PACKET3_RELEASE_MEM_INT_SEL(int_sel ? 2 : 0)));
+
+	/*
+	 * the address should be Qword aligned if 64bit write, Dword
+	 * aligned if only send 32bit data low (discard data high)
+	 */
+	if (write64bit)
+		BUG_ON(addr & 0x7);
+	else
+		BUG_ON(addr & 0x3);
+	amdgpu_ring_write(ring, lower_32_bits(addr));
+	amdgpu_ring_write(ring, upper_32_bits(addr));
+	amdgpu_ring_write(ring, lower_32_bits(seq));
+	amdgpu_ring_write(ring, upper_32_bits(seq));
+	amdgpu_ring_write(ring, 0);
+}
+
+static void gfx_v12_0_ring_emit_pipeline_sync(struct amdgpu_ring *ring)
+{
+	int usepfp = (ring->funcs->type == AMDGPU_RING_TYPE_GFX);
+	uint32_t seq = ring->fence_drv.sync_seq;
+	uint64_t addr = ring->fence_drv.gpu_addr;
+
+	gfx_v12_0_wait_reg_mem(ring, usepfp, 1, 0, lower_32_bits(addr),
+			       upper_32_bits(addr), seq, 0xffffffff, 4);
+}
+
+static void gfx_v12_0_ring_invalidate_tlbs(struct amdgpu_ring *ring,
+				   uint16_t pasid, uint32_t flush_type,
+				   bool all_hub, uint8_t dst_sel)
+{
+	amdgpu_ring_write(ring, PACKET3(PACKET3_INVALIDATE_TLBS, 0));
+	amdgpu_ring_write(ring,
+			  PACKET3_INVALIDATE_TLBS_DST_SEL(dst_sel) |
+			  PACKET3_INVALIDATE_TLBS_ALL_HUB(all_hub) |
+			  PACKET3_INVALIDATE_TLBS_PASID(pasid) |
+			  PACKET3_INVALIDATE_TLBS_FLUSH_TYPE(flush_type));
+}
+
+static void gfx_v12_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
+					 unsigned vmid, uint64_t pd_addr)
+{
+	amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pd_addr);
+
+	/* compute doesn't have PFP */
+	if (ring->funcs->type == AMDGPU_RING_TYPE_GFX) {
+		/* sync PFP to ME, otherwise we might get invalid PFP reads */
+		amdgpu_ring_write(ring, PACKET3(PACKET3_PFP_SYNC_ME, 0));
+		amdgpu_ring_write(ring, 0x0);
+	}
+}
+
+static void gfx_v12_0_ring_emit_fence_kiq(struct amdgpu_ring *ring, u64 addr,
+					  u64 seq, unsigned int flags)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	/* we only allocate 32bit for each seq wb address */
+	BUG_ON(flags & AMDGPU_FENCE_FLAG_64BIT);
+
+	/* write fence seq to the "addr" */
+	amdgpu_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
+	amdgpu_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) |
+				 WRITE_DATA_DST_SEL(5) | WR_CONFIRM));
+	amdgpu_ring_write(ring, lower_32_bits(addr));
+	amdgpu_ring_write(ring, upper_32_bits(addr));
+	amdgpu_ring_write(ring, lower_32_bits(seq));
+
+	if (flags & AMDGPU_FENCE_FLAG_INT) {
+		/* set register to trigger INT */
+		amdgpu_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
+		amdgpu_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) |
+					 WRITE_DATA_DST_SEL(0) | WR_CONFIRM));
+		amdgpu_ring_write(ring, SOC15_REG_OFFSET(GC, 0, regCPC_INT_STATUS));
+		amdgpu_ring_write(ring, 0);
+		amdgpu_ring_write(ring, 0x20000000); /* src_id is 178 */
+	}
+}
+
+static void gfx_v12_0_ring_emit_cntxcntl(struct amdgpu_ring *ring,
+					 uint32_t flags)
+{
+	uint32_t dw2 = 0;
+
+	dw2 |= 0x80000000; /* set load_enable otherwise this package is just NOPs */
+	if (flags & AMDGPU_HAVE_CTX_SWITCH) {
+		/* set load_global_config & load_global_uconfig */
+		dw2 |= 0x8001;
+		/* set load_cs_sh_regs */
+		dw2 |= 0x01000000;
+		/* set load_per_context_state & load_gfx_sh_regs for GFX */
+		dw2 |= 0x10002;
+	}
+
+	amdgpu_ring_write(ring, PACKET3(PACKET3_CONTEXT_CONTROL, 1));
+	amdgpu_ring_write(ring, dw2);
+	amdgpu_ring_write(ring, 0);
+}
+
+static unsigned gfx_v12_0_ring_emit_init_cond_exec(struct amdgpu_ring *ring,
+						   uint64_t addr)
+{
+	unsigned ret;
+
+	amdgpu_ring_write(ring, PACKET3(PACKET3_COND_EXEC, 3));
+	amdgpu_ring_write(ring, lower_32_bits(addr));
+	amdgpu_ring_write(ring, upper_32_bits(addr));
+	/* discard following DWs if *cond_exec_gpu_addr==0 */
+	amdgpu_ring_write(ring, 0);
+	ret = ring->wptr & ring->buf_mask;
+	/* patch dummy value later */
+	amdgpu_ring_write(ring, 0);
+
+	return ret;
+}
+
+static int gfx_v12_0_ring_preempt_ib(struct amdgpu_ring *ring)
+{
+	int i, r = 0;
+	struct amdgpu_device *adev = ring->adev;
+	struct amdgpu_kiq *kiq = &adev->gfx.kiq[0];
+	struct amdgpu_ring *kiq_ring = &kiq->ring;
+	unsigned long flags;
+
+	if (adev->enable_mes)
+		return -EINVAL;
+
+	if (!kiq->pmf || !kiq->pmf->kiq_unmap_queues)
+		return -EINVAL;
+
+	spin_lock_irqsave(&kiq->ring_lock, flags);
+
+	if (amdgpu_ring_alloc(kiq_ring, kiq->pmf->unmap_queues_size)) {
+		spin_unlock_irqrestore(&kiq->ring_lock, flags);
+		return -ENOMEM;
+	}
+
+	/* assert preemption condition */
+	amdgpu_ring_set_preempt_cond_exec(ring, false);
+
+	/* assert IB preemption, emit the trailing fence */
+	kiq->pmf->kiq_unmap_queues(kiq_ring, ring, PREEMPT_QUEUES_NO_UNMAP,
+				   ring->trail_fence_gpu_addr,
+				   ++ring->trail_seq);
+	amdgpu_ring_commit(kiq_ring);
+
+	spin_unlock_irqrestore(&kiq->ring_lock, flags);
+
+	/* poll the trailing fence */
+	for (i = 0; i < adev->usec_timeout; i++) {
+		if (ring->trail_seq ==
+		    le32_to_cpu(*(ring->trail_fence_cpu_addr)))
+			break;
+		udelay(1);
+	}
+
+	if (i >= adev->usec_timeout) {
+		r = -EINVAL;
+		DRM_ERROR("ring %d failed to preempt ib\n", ring->idx);
+	}
+
+	/* deassert preemption condition */
+	amdgpu_ring_set_preempt_cond_exec(ring, true);
+	return r;
+}
+
+static void gfx_v12_0_ring_emit_frame_cntl(struct amdgpu_ring *ring,
+					   bool start,
+					   bool secure)
+{
+	uint32_t v = secure ? FRAME_TMZ : 0;
+
+	amdgpu_ring_write(ring, PACKET3(PACKET3_FRAME_CONTROL, 0));
+	amdgpu_ring_write(ring, v | FRAME_CMD(start ? 0 : 1));
+}
+
+static void gfx_v12_0_ring_emit_rreg(struct amdgpu_ring *ring, uint32_t reg,
+				     uint32_t reg_val_offs)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	amdgpu_ring_write(ring, PACKET3(PACKET3_COPY_DATA, 4));
+	amdgpu_ring_write(ring, 0 |	/* src: register*/
+				(5 << 8) |	/* dst: memory */
+				(1 << 20));	/* write confirm */
+	amdgpu_ring_write(ring, reg);
+	amdgpu_ring_write(ring, 0);
+	amdgpu_ring_write(ring, lower_32_bits(adev->wb.gpu_addr +
+				reg_val_offs * 4));
+	amdgpu_ring_write(ring, upper_32_bits(adev->wb.gpu_addr +
+				reg_val_offs * 4));
+}
+
+static void gfx_v12_0_ring_emit_wreg(struct amdgpu_ring *ring,
+				     uint32_t reg,
+				     uint32_t val)
+{
+	uint32_t cmd = 0;
+
+	switch (ring->funcs->type) {
+	case AMDGPU_RING_TYPE_GFX:
+		cmd = WRITE_DATA_ENGINE_SEL(1) | WR_CONFIRM;
+		break;
+	case AMDGPU_RING_TYPE_KIQ:
+		cmd = (1 << 16); /* no inc addr */
+		break;
+	default:
+		cmd = WR_CONFIRM;
+		break;
+	}
+	amdgpu_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
+	amdgpu_ring_write(ring, cmd);
+	amdgpu_ring_write(ring, reg);
+	amdgpu_ring_write(ring, 0);
+	amdgpu_ring_write(ring, val);
+}
+
+static void gfx_v12_0_ring_emit_reg_wait(struct amdgpu_ring *ring, uint32_t reg,
+					uint32_t val, uint32_t mask)
+{
+	gfx_v12_0_wait_reg_mem(ring, 0, 0, 0, reg, 0, val, mask, 0x20);
+}
+
+static void gfx_v12_0_ring_emit_reg_write_reg_wait(struct amdgpu_ring *ring,
+						   uint32_t reg0, uint32_t reg1,
+						   uint32_t ref, uint32_t mask)
+{
+	int usepfp = (ring->funcs->type == AMDGPU_RING_TYPE_GFX);
+
+	gfx_v12_0_wait_reg_mem(ring, usepfp, 0, 1, reg0, reg1,
+			       ref, mask, 0x20);
+}
+
+static void
+gfx_v12_0_set_gfx_eop_interrupt_state(struct amdgpu_device *adev,
+				      uint32_t me, uint32_t pipe,
+				      enum amdgpu_interrupt_state state)
+{
+	uint32_t cp_int_cntl, cp_int_cntl_reg;
+
+	if (!me) {
+		switch (pipe) {
+		case 0:
+			cp_int_cntl_reg = SOC15_REG_OFFSET(GC, 0, regCP_INT_CNTL_RING0);
+			break;
+		default:
+			DRM_DEBUG("invalid pipe %d\n", pipe);
+			return;
+		}
+	} else {
+		DRM_DEBUG("invalid me %d\n", me);
+		return;
+	}
+
+	switch (state) {
+	case AMDGPU_IRQ_STATE_DISABLE:
+		cp_int_cntl = RREG32_SOC15_IP(GC, cp_int_cntl_reg);
+		cp_int_cntl = REG_SET_FIELD(cp_int_cntl, CP_INT_CNTL_RING0,
+					    TIME_STAMP_INT_ENABLE, 0);
+		cp_int_cntl = REG_SET_FIELD(cp_int_cntl, CP_INT_CNTL_RING0,
+					    GENERIC0_INT_ENABLE, 0);
+		WREG32_SOC15_IP(GC, cp_int_cntl_reg, cp_int_cntl);
+		break;
+	case AMDGPU_IRQ_STATE_ENABLE:
+		cp_int_cntl = RREG32_SOC15_IP(GC, cp_int_cntl_reg);
+		cp_int_cntl = REG_SET_FIELD(cp_int_cntl, CP_INT_CNTL_RING0,
+					    TIME_STAMP_INT_ENABLE, 1);
+		cp_int_cntl = REG_SET_FIELD(cp_int_cntl, CP_INT_CNTL_RING0,
+					    GENERIC0_INT_ENABLE, 1);
+		WREG32_SOC15_IP(GC, cp_int_cntl_reg, cp_int_cntl);
+		break;
+	default:
+		break;
+	}
+}
+
+static void gfx_v12_0_set_compute_eop_interrupt_state(struct amdgpu_device *adev,
+						     int me, int pipe,
+						     enum amdgpu_interrupt_state state)
+{
+	u32 mec_int_cntl, mec_int_cntl_reg;
+
+	/*
+	 * amdgpu controls only the first MEC. That's why this function only
+	 * handles the setting of interrupts for this specific MEC. All other
+	 * pipes' interrupts are set by amdkfd.
+	 */
+
+	if (me == 1) {
+		switch (pipe) {
+		case 0:
+			mec_int_cntl_reg = SOC15_REG_OFFSET(GC, 0, regCP_ME1_PIPE0_INT_CNTL);
+			break;
+		case 1:
+			mec_int_cntl_reg = SOC15_REG_OFFSET(GC, 0, regCP_ME1_PIPE1_INT_CNTL);
+			break;
+		default:
+			DRM_DEBUG("invalid pipe %d\n", pipe);
+			return;
+		}
+	} else {
+		DRM_DEBUG("invalid me %d\n", me);
+		return;
+	}
+
+	switch (state) {
+	case AMDGPU_IRQ_STATE_DISABLE:
+		mec_int_cntl = RREG32_SOC15_IP(GC, mec_int_cntl_reg);
+		mec_int_cntl = REG_SET_FIELD(mec_int_cntl, CP_ME1_PIPE0_INT_CNTL,
+					     TIME_STAMP_INT_ENABLE, 0);
+		mec_int_cntl = REG_SET_FIELD(mec_int_cntl, CP_ME1_PIPE0_INT_CNTL,
+					     GENERIC0_INT_ENABLE, 0);
+		WREG32_SOC15_IP(GC, mec_int_cntl_reg, mec_int_cntl);
+		break;
+	case AMDGPU_IRQ_STATE_ENABLE:
+		mec_int_cntl = RREG32_SOC15_IP(GC, mec_int_cntl_reg);
+		mec_int_cntl = REG_SET_FIELD(mec_int_cntl, CP_ME1_PIPE0_INT_CNTL,
+					     TIME_STAMP_INT_ENABLE, 1);
+		mec_int_cntl = REG_SET_FIELD(mec_int_cntl, CP_ME1_PIPE0_INT_CNTL,
+					     GENERIC0_INT_ENABLE, 1);
+		WREG32_SOC15_IP(GC, mec_int_cntl_reg, mec_int_cntl);
+		break;
+	default:
+		break;
+	}
+}
+
+static int gfx_v12_0_set_eop_interrupt_state(struct amdgpu_device *adev,
+					    struct amdgpu_irq_src *src,
+					    unsigned type,
+					    enum amdgpu_interrupt_state state)
+{
+	switch (type) {
+	case AMDGPU_CP_IRQ_GFX_ME0_PIPE0_EOP:
+		gfx_v12_0_set_gfx_eop_interrupt_state(adev, 0, 0, state);
+		break;
+	case AMDGPU_CP_IRQ_GFX_ME0_PIPE1_EOP:
+		gfx_v12_0_set_gfx_eop_interrupt_state(adev, 0, 1, state);
+		break;
+	case AMDGPU_CP_IRQ_COMPUTE_MEC1_PIPE0_EOP:
+		gfx_v12_0_set_compute_eop_interrupt_state(adev, 1, 0, state);
+		break;
+	case AMDGPU_CP_IRQ_COMPUTE_MEC1_PIPE1_EOP:
+		gfx_v12_0_set_compute_eop_interrupt_state(adev, 1, 1, state);
+		break;
+	case AMDGPU_CP_IRQ_COMPUTE_MEC1_PIPE2_EOP:
+		gfx_v12_0_set_compute_eop_interrupt_state(adev, 1, 2, state);
+		break;
+	case AMDGPU_CP_IRQ_COMPUTE_MEC1_PIPE3_EOP:
+		gfx_v12_0_set_compute_eop_interrupt_state(adev, 1, 3, state);
+		break;
+	default:
+		break;
+	}
+	return 0;
+}
+
+static int gfx_v12_0_eop_irq(struct amdgpu_device *adev,
+			     struct amdgpu_irq_src *source,
+			     struct amdgpu_iv_entry *entry)
+{
+	u32 doorbell_offset = entry->src_data[0];
+	u8 me_id, pipe_id, queue_id;
+	struct amdgpu_ring *ring;
+	int i;
+
+	DRM_DEBUG("IH: CP EOP\n");
+
+	if (adev->enable_mes && doorbell_offset) {
+		struct amdgpu_userq_fence_driver *fence_drv = NULL;
+		struct xarray *xa = &adev->userq_xa;
+		unsigned long flags;
+
+		xa_lock_irqsave(xa, flags);
+		fence_drv = xa_load(xa, doorbell_offset);
+		if (fence_drv)
+			amdgpu_userq_fence_driver_process(fence_drv);
+		xa_unlock_irqrestore(xa, flags);
+	} else {
+		me_id = (entry->ring_id & 0x0c) >> 2;
+		pipe_id = (entry->ring_id & 0x03) >> 0;
+		queue_id = (entry->ring_id & 0x70) >> 4;
+
+		switch (me_id) {
+		case 0:
+			if (pipe_id == 0)
+				amdgpu_fence_process(&adev->gfx.gfx_ring[0]);
+			else
+				amdgpu_fence_process(&adev->gfx.gfx_ring[1]);
+			break;
+		case 1:
+		case 2:
+			for (i = 0; i < adev->gfx.num_compute_rings; i++) {
+				ring = &adev->gfx.compute_ring[i];
+				/* Per-queue interrupt is supported for MEC starting from VI.
+				 * The interrupt can only be enabled/disabled per pipe instead
+				 * of per queue.
+				 */
+				if ((ring->me == me_id) &&
+				    (ring->pipe == pipe_id) &&
+				    (ring->queue == queue_id))
+					amdgpu_fence_process(ring);
+			}
+			break;
+		}
+	}
+
+	return 0;
+}
+
+static int gfx_v12_0_set_priv_reg_fault_state(struct amdgpu_device *adev,
+					      struct amdgpu_irq_src *source,
+					      unsigned int type,
+					      enum amdgpu_interrupt_state state)
+{
+	u32 cp_int_cntl_reg, cp_int_cntl;
+	int i, j;
+
+	switch (state) {
+	case AMDGPU_IRQ_STATE_DISABLE:
+	case AMDGPU_IRQ_STATE_ENABLE:
+		for (i = 0; i < adev->gfx.me.num_me; i++) {
+			for (j = 0; j < adev->gfx.me.num_pipe_per_me; j++) {
+				cp_int_cntl_reg = gfx_v12_0_get_cpg_int_cntl(adev, i, j);
+
+				if (cp_int_cntl_reg) {
+					cp_int_cntl = RREG32_SOC15_IP(GC, cp_int_cntl_reg);
+					cp_int_cntl = REG_SET_FIELD(cp_int_cntl, CP_INT_CNTL_RING0,
+								    PRIV_REG_INT_ENABLE,
+								    state == AMDGPU_IRQ_STATE_ENABLE ? 1 : 0);
+					WREG32_SOC15_IP(GC, cp_int_cntl_reg, cp_int_cntl);
+				}
+			}
+		}
+		for (i = 0; i < adev->gfx.mec.num_mec; i++) {
+			for (j = 0; j < adev->gfx.mec.num_pipe_per_mec; j++) {
+				/* MECs start at 1 */
+				cp_int_cntl_reg = gfx_v12_0_get_cpc_int_cntl(adev, i + 1, j);
+
+				if (cp_int_cntl_reg) {
+					cp_int_cntl = RREG32_SOC15_IP(GC, cp_int_cntl_reg);
+					cp_int_cntl = REG_SET_FIELD(cp_int_cntl, CP_ME1_PIPE0_INT_CNTL,
+								    PRIV_REG_INT_ENABLE,
+								    state == AMDGPU_IRQ_STATE_ENABLE ? 1 : 0);
+					WREG32_SOC15_IP(GC, cp_int_cntl_reg, cp_int_cntl);
+				}
+			}
+		}
+		break;
+	default:
+		break;
+	}
+
+	return 0;
+}
+
+static int gfx_v12_0_set_bad_op_fault_state(struct amdgpu_device *adev,
+					    struct amdgpu_irq_src *source,
+					    unsigned type,
+					    enum amdgpu_interrupt_state state)
+{
+	u32 cp_int_cntl_reg, cp_int_cntl;
+	int i, j;
+
+	switch (state) {
+	case AMDGPU_IRQ_STATE_DISABLE:
+	case AMDGPU_IRQ_STATE_ENABLE:
+		for (i = 0; i < adev->gfx.me.num_me; i++) {
+			for (j = 0; j < adev->gfx.me.num_pipe_per_me; j++) {
+				cp_int_cntl_reg = gfx_v12_0_get_cpg_int_cntl(adev, i, j);
+
+				if (cp_int_cntl_reg) {
+					cp_int_cntl = RREG32_SOC15_IP(GC, cp_int_cntl_reg);
+					cp_int_cntl = REG_SET_FIELD(cp_int_cntl, CP_INT_CNTL_RING0,
+								    OPCODE_ERROR_INT_ENABLE,
+								    state == AMDGPU_IRQ_STATE_ENABLE ? 1 : 0);
+					WREG32_SOC15_IP(GC, cp_int_cntl_reg, cp_int_cntl);
+				}
+			}
+		}
+		for (i = 0; i < adev->gfx.mec.num_mec; i++) {
+			for (j = 0; j < adev->gfx.mec.num_pipe_per_mec; j++) {
+				/* MECs start at 1 */
+				cp_int_cntl_reg = gfx_v12_0_get_cpc_int_cntl(adev, i + 1, j);
+
+				if (cp_int_cntl_reg) {
+					cp_int_cntl = RREG32_SOC15_IP(GC, cp_int_cntl_reg);
+					cp_int_cntl = REG_SET_FIELD(cp_int_cntl, CP_ME1_PIPE0_INT_CNTL,
+								    OPCODE_ERROR_INT_ENABLE,
+								    state == AMDGPU_IRQ_STATE_ENABLE ? 1 : 0);
+					WREG32_SOC15_IP(GC, cp_int_cntl_reg, cp_int_cntl);
+				}
+			}
+		}
+		break;
+	default:
+		break;
+	}
+	return 0;
+}
+
+static int gfx_v12_0_set_priv_inst_fault_state(struct amdgpu_device *adev,
+					       struct amdgpu_irq_src *source,
+					       unsigned int type,
+					       enum amdgpu_interrupt_state state)
+{
+	u32 cp_int_cntl_reg, cp_int_cntl;
+	int i, j;
+
+	switch (state) {
+	case AMDGPU_IRQ_STATE_DISABLE:
+	case AMDGPU_IRQ_STATE_ENABLE:
+		for (i = 0; i < adev->gfx.me.num_me; i++) {
+			for (j = 0; j < adev->gfx.me.num_pipe_per_me; j++) {
+				cp_int_cntl_reg = gfx_v12_0_get_cpg_int_cntl(adev, i, j);
+
+				if (cp_int_cntl_reg) {
+					cp_int_cntl = RREG32_SOC15_IP(GC, cp_int_cntl_reg);
+					cp_int_cntl = REG_SET_FIELD(cp_int_cntl, CP_INT_CNTL_RING0,
+								    PRIV_INSTR_INT_ENABLE,
+								    state == AMDGPU_IRQ_STATE_ENABLE ? 1 : 0);
+					WREG32_SOC15_IP(GC, cp_int_cntl_reg, cp_int_cntl);
+				}
+			}
+		}
+		break;
+	default:
+		break;
+	}
+
+	return 0;
+}
+
+static void gfx_v12_0_handle_priv_fault(struct amdgpu_device *adev,
+					struct amdgpu_iv_entry *entry)
+{
+	u8 me_id, pipe_id, queue_id;
+	struct amdgpu_ring *ring;
+	int i;
+
+	me_id = (entry->ring_id & 0x0c) >> 2;
+	pipe_id = (entry->ring_id & 0x03) >> 0;
+	queue_id = (entry->ring_id & 0x70) >> 4;
+
+	if (!adev->gfx.disable_kq) {
+		switch (me_id) {
+		case 0:
+			for (i = 0; i < adev->gfx.num_gfx_rings; i++) {
+				ring = &adev->gfx.gfx_ring[i];
+				if (ring->me == me_id && ring->pipe == pipe_id &&
+				    ring->queue == queue_id)
+					drm_sched_fault(&ring->sched);
+			}
+			break;
+		case 1:
+		case 2:
+			for (i = 0; i < adev->gfx.num_compute_rings; i++) {
+				ring = &adev->gfx.compute_ring[i];
+				if (ring->me == me_id && ring->pipe == pipe_id &&
+				    ring->queue == queue_id)
+					drm_sched_fault(&ring->sched);
+			}
+			break;
+		default:
+			BUG();
+			break;
+		}
+	}
+}
+
+static int gfx_v12_0_priv_reg_irq(struct amdgpu_device *adev,
+				  struct amdgpu_irq_src *source,
+				  struct amdgpu_iv_entry *entry)
+{
+	DRM_ERROR("Illegal register access in command stream\n");
+	gfx_v12_0_handle_priv_fault(adev, entry);
+	return 0;
+}
+
+static int gfx_v12_0_bad_op_irq(struct amdgpu_device *adev,
+				struct amdgpu_irq_src *source,
+				struct amdgpu_iv_entry *entry)
+{
+	DRM_ERROR("Illegal opcode in command stream \n");
+	gfx_v12_0_handle_priv_fault(adev, entry);
+	return 0;
+}
+
+static int gfx_v12_0_priv_inst_irq(struct amdgpu_device *adev,
+				   struct amdgpu_irq_src *source,
+				   struct amdgpu_iv_entry *entry)
+{
+	DRM_ERROR("Illegal instruction in command stream\n");
+	gfx_v12_0_handle_priv_fault(adev, entry);
+	return 0;
+}
+
+static void gfx_v12_0_emit_mem_sync(struct amdgpu_ring *ring)
+{
+	const unsigned int gcr_cntl =
+			PACKET3_ACQUIRE_MEM_GCR_CNTL_GL2_INV(1) |
+			PACKET3_ACQUIRE_MEM_GCR_CNTL_GL2_WB(1) |
+			PACKET3_ACQUIRE_MEM_GCR_CNTL_GLM_INV(1) |
+			PACKET3_ACQUIRE_MEM_GCR_CNTL_GLM_WB(1) |
+			PACKET3_ACQUIRE_MEM_GCR_CNTL_GL1_INV(1) |
+			PACKET3_ACQUIRE_MEM_GCR_CNTL_GLV_INV(1) |
+			PACKET3_ACQUIRE_MEM_GCR_CNTL_GLK_INV(1) |
+			PACKET3_ACQUIRE_MEM_GCR_CNTL_GLI_INV(1);
+
+	/* ACQUIRE_MEM - make one or more surfaces valid for use by the subsequent operations */
+	amdgpu_ring_write(ring, PACKET3(PACKET3_ACQUIRE_MEM, 6));
+	amdgpu_ring_write(ring, 0); /* CP_COHER_CNTL */
+	amdgpu_ring_write(ring, 0xffffffff);  /* CP_COHER_SIZE */
+	amdgpu_ring_write(ring, 0xffffff);  /* CP_COHER_SIZE_HI */
+	amdgpu_ring_write(ring, 0); /* CP_COHER_BASE */
+	amdgpu_ring_write(ring, 0);  /* CP_COHER_BASE_HI */
+	amdgpu_ring_write(ring, 0x0000000A); /* POLL_INTERVAL */
+	amdgpu_ring_write(ring, gcr_cntl); /* GCR_CNTL */
+}
+
+static void gfx_v12_ring_insert_nop(struct amdgpu_ring *ring, uint32_t num_nop)
+{
+	/* Header itself is a NOP packet */
+	if (num_nop == 1) {
+		amdgpu_ring_write(ring, ring->funcs->nop);
+		return;
+	}
+
+	/* Max HW optimization till 0x3ffe, followed by remaining one NOP at a time*/
+	amdgpu_ring_write(ring, PACKET3(PACKET3_NOP, min(num_nop - 2, 0x3ffe)));
+
+	/* Header is at index 0, followed by num_nops - 1 NOP packet's */
+	amdgpu_ring_insert_nop(ring, num_nop - 1);
+}
+
+static void gfx_v12_0_ring_emit_cleaner_shader(struct amdgpu_ring *ring)
+{
+	/* Emit the cleaner shader */
+	amdgpu_ring_write(ring, PACKET3(PACKET3_RUN_CLEANER_SHADER, 0));
+	amdgpu_ring_write(ring, 0);  /* RESERVED field, programmed to zero */
+}
+
+static void gfx_v12_ip_print(struct amdgpu_ip_block *ip_block, struct drm_printer *p)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	uint32_t i, j, k, reg, index = 0;
+	uint32_t reg_count = ARRAY_SIZE(gc_reg_list_12_0);
+
+	if (!adev->gfx.ip_dump_core)
+		return;
+
+	for (i = 0; i < reg_count; i++)
+		drm_printf(p, "%-50s \t 0x%08x\n",
+			   gc_reg_list_12_0[i].reg_name,
+			   adev->gfx.ip_dump_core[i]);
+
+	/* print compute queue registers for all instances */
+	if (!adev->gfx.ip_dump_compute_queues)
+		return;
+
+	reg_count = ARRAY_SIZE(gc_cp_reg_list_12);
+	drm_printf(p, "\nnum_mec: %d num_pipe: %d num_queue: %d\n",
+		   adev->gfx.mec.num_mec,
+		   adev->gfx.mec.num_pipe_per_mec,
+		   adev->gfx.mec.num_queue_per_pipe);
+
+	for (i = 0; i < adev->gfx.mec.num_mec; i++) {
+		for (j = 0; j < adev->gfx.mec.num_pipe_per_mec; j++) {
+			for (k = 0; k < adev->gfx.mec.num_queue_per_pipe; k++) {
+				drm_printf(p, "\nmec %d, pipe %d, queue %d\n", i, j, k);
+				for (reg = 0; reg < reg_count; reg++) {
+					drm_printf(p, "%-50s \t 0x%08x\n",
+						   gc_cp_reg_list_12[reg].reg_name,
+						   adev->gfx.ip_dump_compute_queues[index + reg]);
+				}
+				index += reg_count;
+			}
+		}
+	}
+
+	/* print gfx queue registers for all instances */
+	if (!adev->gfx.ip_dump_gfx_queues)
+		return;
+
+	index = 0;
+	reg_count = ARRAY_SIZE(gc_gfx_queue_reg_list_12);
+	drm_printf(p, "\nnum_me: %d num_pipe: %d num_queue: %d\n",
+		   adev->gfx.me.num_me,
+		   adev->gfx.me.num_pipe_per_me,
+		   adev->gfx.me.num_queue_per_pipe);
+
+	for (i = 0; i < adev->gfx.me.num_me; i++) {
+		for (j = 0; j < adev->gfx.me.num_pipe_per_me; j++) {
+			for (k = 0; k < adev->gfx.me.num_queue_per_pipe; k++) {
+				drm_printf(p, "\nme %d, pipe %d, queue %d\n", i, j, k);
+				for (reg = 0; reg < reg_count; reg++) {
+					drm_printf(p, "%-50s \t 0x%08x\n",
+						   gc_gfx_queue_reg_list_12[reg].reg_name,
+						   adev->gfx.ip_dump_gfx_queues[index + reg]);
+				}
+				index += reg_count;
+			}
+		}
+	}
+}
+
+static void gfx_v12_ip_dump(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	uint32_t i, j, k, reg, index = 0;
+	uint32_t reg_count = ARRAY_SIZE(gc_reg_list_12_0);
+
+	if (!adev->gfx.ip_dump_core)
+		return;
+
+	amdgpu_gfx_off_ctrl(adev, false);
+	for (i = 0; i < reg_count; i++)
+		adev->gfx.ip_dump_core[i] = RREG32(SOC15_REG_ENTRY_OFFSET(gc_reg_list_12_0[i]));
+	amdgpu_gfx_off_ctrl(adev, true);
+
+	/* dump compute queue registers for all instances */
+	if (!adev->gfx.ip_dump_compute_queues)
+		return;
+
+	reg_count = ARRAY_SIZE(gc_cp_reg_list_12);
+	amdgpu_gfx_off_ctrl(adev, false);
+	mutex_lock(&adev->srbm_mutex);
+	for (i = 0; i < adev->gfx.mec.num_mec; i++) {
+		for (j = 0; j < adev->gfx.mec.num_pipe_per_mec; j++) {
+			for (k = 0; k < adev->gfx.mec.num_queue_per_pipe; k++) {
+				/* ME0 is for GFX so start from 1 for CP */
+				soc24_grbm_select(adev, adev->gfx.me.num_me + i, j, k, 0);
+				for (reg = 0; reg < reg_count; reg++) {
+					adev->gfx.ip_dump_compute_queues[index + reg] =
+						RREG32(SOC15_REG_ENTRY_OFFSET(
+							gc_cp_reg_list_12[reg]));
+				}
+				index += reg_count;
+			}
+		}
+	}
+	soc24_grbm_select(adev, 0, 0, 0, 0);
+	mutex_unlock(&adev->srbm_mutex);
+	amdgpu_gfx_off_ctrl(adev, true);
+
+	/* dump gfx queue registers for all instances */
+	if (!adev->gfx.ip_dump_gfx_queues)
+		return;
+
+	index = 0;
+	reg_count = ARRAY_SIZE(gc_gfx_queue_reg_list_12);
+	amdgpu_gfx_off_ctrl(adev, false);
+	mutex_lock(&adev->srbm_mutex);
+	for (i = 0; i < adev->gfx.me.num_me; i++) {
+		for (j = 0; j < adev->gfx.me.num_pipe_per_me; j++) {
+			for (k = 0; k < adev->gfx.me.num_queue_per_pipe; k++) {
+				soc24_grbm_select(adev, i, j, k, 0);
+
+				for (reg = 0; reg < reg_count; reg++) {
+					adev->gfx.ip_dump_gfx_queues[index + reg] =
+						RREG32(SOC15_REG_ENTRY_OFFSET(
+							gc_gfx_queue_reg_list_12[reg]));
+				}
+				index += reg_count;
+			}
+		}
+	}
+	soc24_grbm_select(adev, 0, 0, 0, 0);
+	mutex_unlock(&adev->srbm_mutex);
+	amdgpu_gfx_off_ctrl(adev, true);
+}
+
+static bool gfx_v12_pipe_reset_support(struct amdgpu_device *adev)
+{
+	/* Disable the pipe reset until the CPFW fully support it.*/
+	dev_warn_once(adev->dev, "The CPFW hasn't support pipe reset yet.\n");
+	return false;
+}
+
+static int gfx_v12_reset_gfx_pipe(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	uint32_t reset_pipe = 0, clean_pipe = 0;
+	int r;
+
+	if (!gfx_v12_pipe_reset_support(adev))
+		return -EOPNOTSUPP;
+
+	gfx_v12_0_set_safe_mode(adev, 0);
+	mutex_lock(&adev->srbm_mutex);
+	soc24_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0);
+
+	switch (ring->pipe) {
+	case 0:
+		reset_pipe = REG_SET_FIELD(reset_pipe, CP_ME_CNTL,
+					   PFP_PIPE0_RESET, 1);
+		reset_pipe = REG_SET_FIELD(reset_pipe, CP_ME_CNTL,
+					   ME_PIPE0_RESET, 1);
+		clean_pipe = REG_SET_FIELD(clean_pipe, CP_ME_CNTL,
+					   PFP_PIPE0_RESET, 0);
+		clean_pipe = REG_SET_FIELD(clean_pipe, CP_ME_CNTL,
+					   ME_PIPE0_RESET, 0);
+		break;
+	case 1:
+		reset_pipe = REG_SET_FIELD(reset_pipe, CP_ME_CNTL,
+					   PFP_PIPE1_RESET, 1);
+		reset_pipe = REG_SET_FIELD(reset_pipe, CP_ME_CNTL,
+					   ME_PIPE1_RESET, 1);
+		clean_pipe = REG_SET_FIELD(clean_pipe, CP_ME_CNTL,
+					   PFP_PIPE1_RESET, 0);
+		clean_pipe = REG_SET_FIELD(clean_pipe, CP_ME_CNTL,
+					   ME_PIPE1_RESET, 0);
+		break;
+	default:
+		break;
+	}
+
+	WREG32_SOC15(GC, 0, regCP_ME_CNTL, reset_pipe);
+	WREG32_SOC15(GC, 0, regCP_ME_CNTL, clean_pipe);
+
+	r = (RREG32(SOC15_REG_OFFSET(GC, 0, regCP_GFX_RS64_INSTR_PNTR1)) << 2) -
+					RS64_FW_UC_START_ADDR_LO;
+	soc24_grbm_select(adev, 0, 0, 0, 0);
+	mutex_unlock(&adev->srbm_mutex);
+	gfx_v12_0_unset_safe_mode(adev, 0);
+
+	dev_info(adev->dev, "The ring %s pipe reset: %s\n", ring->name,
+			r == 0 ? "successfully" : "failed");
+	/* Sometimes the ME start pc counter can't cache correctly, so the
+	 * PC check only as a reference and pipe reset result rely on the
+	 * later ring test.
+	 */
+	return 0;
+}
+
+static int gfx_v12_0_reset_kgq(struct amdgpu_ring *ring,
+			       unsigned int vmid,
+			       struct amdgpu_fence *timedout_fence)
+{
+	struct amdgpu_device *adev = ring->adev;
+	int r;
+
+	amdgpu_ring_reset_helper_begin(ring, timedout_fence);
+
+	r = amdgpu_mes_reset_legacy_queue(ring->adev, ring, vmid, false);
+	if (r) {
+		dev_warn(adev->dev, "reset via MES failed and try pipe reset %d\n", r);
+		r = gfx_v12_reset_gfx_pipe(ring);
+		if (r)
+			return r;
+	}
+
+	r = gfx_v12_0_kgq_init_queue(ring, true);
+	if (r) {
+		dev_err(adev->dev, "failed to init kgq\n");
+		return r;
+	}
+
+	r = amdgpu_mes_map_legacy_queue(adev, ring);
+	if (r) {
+		dev_err(adev->dev, "failed to remap kgq\n");
+		return r;
+	}
+
+	return amdgpu_ring_reset_helper_end(ring, timedout_fence);
+}
+
+static int gfx_v12_0_reset_compute_pipe(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	uint32_t reset_pipe = 0, clean_pipe = 0;
+	int r = 0;
+
+	if (!gfx_v12_pipe_reset_support(adev))
+		return -EOPNOTSUPP;
+
+	gfx_v12_0_set_safe_mode(adev, 0);
+	mutex_lock(&adev->srbm_mutex);
+	soc24_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0);
+
+	reset_pipe = RREG32_SOC15(GC, 0, regCP_MEC_RS64_CNTL);
+	clean_pipe = reset_pipe;
+
+	if (adev->gfx.rs64_enable) {
+		switch (ring->pipe) {
+		case 0:
+			reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_RS64_CNTL,
+						   MEC_PIPE0_RESET, 1);
+			clean_pipe = REG_SET_FIELD(clean_pipe, CP_MEC_RS64_CNTL,
+						   MEC_PIPE0_RESET, 0);
+			break;
+		case 1:
+			reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_RS64_CNTL,
+						   MEC_PIPE1_RESET, 1);
+			clean_pipe = REG_SET_FIELD(clean_pipe, CP_MEC_RS64_CNTL,
+						   MEC_PIPE1_RESET, 0);
+			break;
+		case 2:
+			reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_RS64_CNTL,
+						   MEC_PIPE2_RESET, 1);
+			clean_pipe = REG_SET_FIELD(clean_pipe, CP_MEC_RS64_CNTL,
+						   MEC_PIPE2_RESET, 0);
+			break;
+		case 3:
+			reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_RS64_CNTL,
+						   MEC_PIPE3_RESET, 1);
+			clean_pipe = REG_SET_FIELD(clean_pipe, CP_MEC_RS64_CNTL,
+						   MEC_PIPE3_RESET, 0);
+			break;
+		default:
+			break;
+		}
+		WREG32_SOC15(GC, 0, regCP_MEC_RS64_CNTL, reset_pipe);
+		WREG32_SOC15(GC, 0, regCP_MEC_RS64_CNTL, clean_pipe);
+		r = (RREG32_SOC15(GC, 0, regCP_MEC_RS64_INSTR_PNTR) << 2) -
+				RS64_FW_UC_START_ADDR_LO;
+	} else {
+		switch (ring->pipe) {
+		case 0:
+			reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_CNTL,
+							   MEC_ME1_PIPE0_RESET, 1);
+			clean_pipe = REG_SET_FIELD(clean_pipe, CP_MEC_CNTL,
+							   MEC_ME1_PIPE0_RESET, 0);
+			break;
+		case 1:
+			reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_CNTL,
+							   MEC_ME1_PIPE1_RESET, 1);
+			clean_pipe = REG_SET_FIELD(clean_pipe, CP_MEC_CNTL,
+							   MEC_ME1_PIPE1_RESET, 0);
+			break;
+		default:
+		break;
+		}
+		WREG32_SOC15(GC, 0, regCP_MEC_CNTL, reset_pipe);
+		WREG32_SOC15(GC, 0, regCP_MEC_CNTL, clean_pipe);
+		/* Doesn't find the F32 MEC instruction pointer register, and suppose
+		 * the driver won't run into the F32 mode.
+		 */
+	}
+
+	soc24_grbm_select(adev, 0, 0, 0, 0);
+	mutex_unlock(&adev->srbm_mutex);
+	gfx_v12_0_unset_safe_mode(adev, 0);
+
+	dev_info(adev->dev, "The ring %s pipe resets: %s\n", ring->name,
+			r == 0 ? "successfully" : "failed");
+	/* Need the ring test to verify the pipe reset result.*/
+	return 0;
+}
+
+static int gfx_v12_0_reset_kcq(struct amdgpu_ring *ring,
+			       unsigned int vmid,
+			       struct amdgpu_fence *timedout_fence)
+{
+	struct amdgpu_device *adev = ring->adev;
+	int r;
+
+	amdgpu_ring_reset_helper_begin(ring, timedout_fence);
+
+	r = amdgpu_mes_reset_legacy_queue(ring->adev, ring, vmid, true);
+	if (r) {
+		dev_warn(adev->dev, "fail(%d) to reset kcq  and try pipe reset\n", r);
+		r = gfx_v12_0_reset_compute_pipe(ring);
+		if (r)
+			return r;
+	}
+
+	r = gfx_v12_0_kcq_init_queue(ring, true);
+	if (r) {
+		dev_err(adev->dev, "failed to init kcq\n");
+		return r;
+	}
+	r = amdgpu_mes_map_legacy_queue(adev, ring);
+	if (r) {
+		dev_err(adev->dev, "failed to remap kcq\n");
+		return r;
+	}
+
+	return amdgpu_ring_reset_helper_end(ring, timedout_fence);
+}
+
+static void gfx_v12_0_ring_begin_use(struct amdgpu_ring *ring)
+{
+	amdgpu_gfx_profile_ring_begin_use(ring);
+
+	amdgpu_gfx_enforce_isolation_ring_begin_use(ring);
+}
+
+static void gfx_v12_0_ring_end_use(struct amdgpu_ring *ring)
+{
+	amdgpu_gfx_profile_ring_end_use(ring);
+
+	amdgpu_gfx_enforce_isolation_ring_end_use(ring);
+}
+
+static const struct amd_ip_funcs gfx_v12_0_ip_funcs = {
+	.name = "gfx_v12_0",
+	.early_init = gfx_v12_0_early_init,
+	.late_init = gfx_v12_0_late_init,
+	.sw_init = gfx_v12_0_sw_init,
+	.sw_fini = gfx_v12_0_sw_fini,
+	.hw_init = gfx_v12_0_hw_init,
+	.hw_fini = gfx_v12_0_hw_fini,
+	.suspend = gfx_v12_0_suspend,
+	.resume = gfx_v12_0_resume,
+	.is_idle = gfx_v12_0_is_idle,
+	.wait_for_idle = gfx_v12_0_wait_for_idle,
+	.set_clockgating_state = gfx_v12_0_set_clockgating_state,
+	.set_powergating_state = gfx_v12_0_set_powergating_state,
+	.get_clockgating_state = gfx_v12_0_get_clockgating_state,
+	.dump_ip_state = gfx_v12_ip_dump,
+	.print_ip_state = gfx_v12_ip_print,
+};
+
+static const struct amdgpu_ring_funcs gfx_v12_0_ring_funcs_gfx = {
+	.type = AMDGPU_RING_TYPE_GFX,
+	.align_mask = 0xff,
+	.nop = PACKET3(PACKET3_NOP, 0x3FFF),
+	.support_64bit_ptrs = true,
+	.secure_submission_supported = true,
+	.get_rptr = gfx_v12_0_ring_get_rptr_gfx,
+	.get_wptr = gfx_v12_0_ring_get_wptr_gfx,
+	.set_wptr = gfx_v12_0_ring_set_wptr_gfx,
+	.emit_frame_size = /* totally 242 maximum if 16 IBs */
+		5 + /* COND_EXEC */
+		7 + /* PIPELINE_SYNC */
+		SOC15_FLUSH_GPU_TLB_NUM_WREG * 5 +
+		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 7 +
+		2 + /* VM_FLUSH */
+		8 + /* FENCE for VM_FLUSH */
+		5 + /* COND_EXEC */
+		7 + /* HDP_flush */
+		4 + /* VGT_flush */
+		31 + /*	DE_META */
+		3 + /* CNTX_CTRL */
+		5 + /* HDP_INVL */
+		8 + 8 + /* FENCE x2 */
+		8 + /* gfx_v12_0_emit_mem_sync */
+		2, /* gfx_v12_0_ring_emit_cleaner_shader */
+	.emit_ib_size =	4, /* gfx_v12_0_ring_emit_ib_gfx */
+	.emit_ib = gfx_v12_0_ring_emit_ib_gfx,
+	.emit_fence = gfx_v12_0_ring_emit_fence,
+	.emit_pipeline_sync = gfx_v12_0_ring_emit_pipeline_sync,
+	.emit_vm_flush = gfx_v12_0_ring_emit_vm_flush,
+	.emit_hdp_flush = gfx_v12_0_ring_emit_hdp_flush,
+	.test_ring = gfx_v12_0_ring_test_ring,
+	.test_ib = gfx_v12_0_ring_test_ib,
+	.insert_nop = gfx_v12_ring_insert_nop,
+	.pad_ib = amdgpu_ring_generic_pad_ib,
+	.emit_cntxcntl = gfx_v12_0_ring_emit_cntxcntl,
+	.init_cond_exec = gfx_v12_0_ring_emit_init_cond_exec,
+	.preempt_ib = gfx_v12_0_ring_preempt_ib,
+	.emit_frame_cntl = gfx_v12_0_ring_emit_frame_cntl,
+	.emit_wreg = gfx_v12_0_ring_emit_wreg,
+	.emit_reg_wait = gfx_v12_0_ring_emit_reg_wait,
+	.emit_reg_write_reg_wait = gfx_v12_0_ring_emit_reg_write_reg_wait,
+	.emit_mem_sync = gfx_v12_0_emit_mem_sync,
+	.reset = gfx_v12_0_reset_kgq,
+	.emit_cleaner_shader = gfx_v12_0_ring_emit_cleaner_shader,
+	.begin_use = gfx_v12_0_ring_begin_use,
+	.end_use = gfx_v12_0_ring_end_use,
+};
+
+static const struct amdgpu_ring_funcs gfx_v12_0_ring_funcs_compute = {
+	.type = AMDGPU_RING_TYPE_COMPUTE,
+	.align_mask = 0xff,
+	.nop = PACKET3(PACKET3_NOP, 0x3FFF),
+	.support_64bit_ptrs = true,
+	.get_rptr = gfx_v12_0_ring_get_rptr_compute,
+	.get_wptr = gfx_v12_0_ring_get_wptr_compute,
+	.set_wptr = gfx_v12_0_ring_set_wptr_compute,
+	.emit_frame_size =
+		7 + /* gfx_v12_0_ring_emit_hdp_flush */
+		5 + /* hdp invalidate */
+		7 + /* gfx_v12_0_ring_emit_pipeline_sync */
+		SOC15_FLUSH_GPU_TLB_NUM_WREG * 5 +
+		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 7 +
+		2 + /* gfx_v12_0_ring_emit_vm_flush */
+		8 + 8 + 8 + /* gfx_v12_0_ring_emit_fence x3 for user fence, vm fence */
+		8 + /* gfx_v12_0_emit_mem_sync */
+		2, /* gfx_v12_0_ring_emit_cleaner_shader */
+	.emit_ib_size =	7, /* gfx_v12_0_ring_emit_ib_compute */
+	.emit_ib = gfx_v12_0_ring_emit_ib_compute,
+	.emit_fence = gfx_v12_0_ring_emit_fence,
+	.emit_pipeline_sync = gfx_v12_0_ring_emit_pipeline_sync,
+	.emit_vm_flush = gfx_v12_0_ring_emit_vm_flush,
+	.emit_hdp_flush = gfx_v12_0_ring_emit_hdp_flush,
+	.test_ring = gfx_v12_0_ring_test_ring,
+	.test_ib = gfx_v12_0_ring_test_ib,
+	.insert_nop = gfx_v12_ring_insert_nop,
+	.pad_ib = amdgpu_ring_generic_pad_ib,
+	.emit_wreg = gfx_v12_0_ring_emit_wreg,
+	.emit_reg_wait = gfx_v12_0_ring_emit_reg_wait,
+	.emit_reg_write_reg_wait = gfx_v12_0_ring_emit_reg_write_reg_wait,
+	.emit_mem_sync = gfx_v12_0_emit_mem_sync,
+	.reset = gfx_v12_0_reset_kcq,
+	.emit_cleaner_shader = gfx_v12_0_ring_emit_cleaner_shader,
+	.begin_use = gfx_v12_0_ring_begin_use,
+	.end_use = gfx_v12_0_ring_end_use,
+};
+
+static const struct amdgpu_ring_funcs gfx_v12_0_ring_funcs_kiq = {
+	.type = AMDGPU_RING_TYPE_KIQ,
+	.align_mask = 0xff,
+	.nop = PACKET3(PACKET3_NOP, 0x3FFF),
+	.support_64bit_ptrs = true,
+	.get_rptr = gfx_v12_0_ring_get_rptr_compute,
+	.get_wptr = gfx_v12_0_ring_get_wptr_compute,
+	.set_wptr = gfx_v12_0_ring_set_wptr_compute,
+	.emit_frame_size =
+		7 + /* gfx_v12_0_ring_emit_hdp_flush */
+		5 + /*hdp invalidate */
+		7 + /* gfx_v12_0_ring_emit_pipeline_sync */
+		SOC15_FLUSH_GPU_TLB_NUM_WREG * 5 +
+		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 7 +
+		2 + /* gfx_v12_0_ring_emit_vm_flush */
+		8 + 8 + 8, /* gfx_v12_0_ring_emit_fence_kiq x3 for user fence, vm fence */
+	.emit_ib_size =	7, /* gfx_v12_0_ring_emit_ib_compute */
+	.emit_ib = gfx_v12_0_ring_emit_ib_compute,
+	.emit_fence = gfx_v12_0_ring_emit_fence_kiq,
+	.test_ring = gfx_v12_0_ring_test_ring,
+	.test_ib = gfx_v12_0_ring_test_ib,
+	.insert_nop = amdgpu_ring_insert_nop,
+	.pad_ib = amdgpu_ring_generic_pad_ib,
+	.emit_rreg = gfx_v12_0_ring_emit_rreg,
+	.emit_wreg = gfx_v12_0_ring_emit_wreg,
+	.emit_reg_wait = gfx_v12_0_ring_emit_reg_wait,
+	.emit_reg_write_reg_wait = gfx_v12_0_ring_emit_reg_write_reg_wait,
+	.emit_hdp_flush = gfx_v12_0_ring_emit_hdp_flush,
+};
+
+static void gfx_v12_0_set_ring_funcs(struct amdgpu_device *adev)
+{
+	int i;
+
+	adev->gfx.kiq[0].ring.funcs = &gfx_v12_0_ring_funcs_kiq;
+
+	for (i = 0; i < adev->gfx.num_gfx_rings; i++)
+		adev->gfx.gfx_ring[i].funcs = &gfx_v12_0_ring_funcs_gfx;
+
+	for (i = 0; i < adev->gfx.num_compute_rings; i++)
+		adev->gfx.compute_ring[i].funcs = &gfx_v12_0_ring_funcs_compute;
+}
+
+static const struct amdgpu_irq_src_funcs gfx_v12_0_eop_irq_funcs = {
+	.set = gfx_v12_0_set_eop_interrupt_state,
+	.process = gfx_v12_0_eop_irq,
+};
+
+static const struct amdgpu_irq_src_funcs gfx_v12_0_priv_reg_irq_funcs = {
+	.set = gfx_v12_0_set_priv_reg_fault_state,
+	.process = gfx_v12_0_priv_reg_irq,
+};
+
+static const struct amdgpu_irq_src_funcs gfx_v12_0_bad_op_irq_funcs = {
+	.set = gfx_v12_0_set_bad_op_fault_state,
+	.process = gfx_v12_0_bad_op_irq,
+};
+
+static const struct amdgpu_irq_src_funcs gfx_v12_0_priv_inst_irq_funcs = {
+	.set = gfx_v12_0_set_priv_inst_fault_state,
+	.process = gfx_v12_0_priv_inst_irq,
+};
+
+static void gfx_v12_0_set_irq_funcs(struct amdgpu_device *adev)
+{
+	adev->gfx.eop_irq.num_types = AMDGPU_CP_IRQ_LAST;
+	adev->gfx.eop_irq.funcs = &gfx_v12_0_eop_irq_funcs;
+
+	adev->gfx.priv_reg_irq.num_types = 1;
+	adev->gfx.priv_reg_irq.funcs = &gfx_v12_0_priv_reg_irq_funcs;
+
+	adev->gfx.bad_op_irq.num_types = 1;
+	adev->gfx.bad_op_irq.funcs = &gfx_v12_0_bad_op_irq_funcs;
+
+	adev->gfx.priv_inst_irq.num_types = 1;
+	adev->gfx.priv_inst_irq.funcs = &gfx_v12_0_priv_inst_irq_funcs;
+}
+
+static void gfx_v12_0_set_imu_funcs(struct amdgpu_device *adev)
+{
+	if (adev->flags & AMD_IS_APU)
+		adev->gfx.imu.mode = MISSION_MODE;
+	else
+		adev->gfx.imu.mode = DEBUG_MODE;
+
+	adev->gfx.imu.funcs = &gfx_v12_0_imu_funcs;
+}
+
+static void gfx_v12_0_set_rlc_funcs(struct amdgpu_device *adev)
+{
+	adev->gfx.rlc.funcs = &gfx_v12_0_rlc_funcs;
+}
+
+static void gfx_v12_0_set_mqd_funcs(struct amdgpu_device *adev)
+{
+	/* set gfx eng mqd */
+	adev->mqds[AMDGPU_HW_IP_GFX].mqd_size =
+		sizeof(struct v12_gfx_mqd);
+	adev->mqds[AMDGPU_HW_IP_GFX].init_mqd =
+		gfx_v12_0_gfx_mqd_init;
+	/* set compute eng mqd */
+	adev->mqds[AMDGPU_HW_IP_COMPUTE].mqd_size =
+		sizeof(struct v12_compute_mqd);
+	adev->mqds[AMDGPU_HW_IP_COMPUTE].init_mqd =
+		gfx_v12_0_compute_mqd_init;
+}
+
+static void gfx_v12_0_set_user_wgp_inactive_bitmap_per_sh(struct amdgpu_device *adev,
+							  u32 bitmap)
+{
+	u32 data;
+
+	if (!bitmap)
+		return;
+
+	data = bitmap << GC_USER_SHADER_ARRAY_CONFIG__INACTIVE_WGPS__SHIFT;
+	data &= GC_USER_SHADER_ARRAY_CONFIG__INACTIVE_WGPS_MASK;
+
+	WREG32_SOC15(GC, 0, regGC_USER_SHADER_ARRAY_CONFIG, data);
+}
+
+static u32 gfx_v12_0_get_wgp_active_bitmap_per_sh(struct amdgpu_device *adev)
+{
+	u32 data, wgp_bitmask;
+	data = RREG32_SOC15(GC, 0, regCC_GC_SHADER_ARRAY_CONFIG);
+	data |= RREG32_SOC15(GC, 0, regGC_USER_SHADER_ARRAY_CONFIG);
+
+	data &= CC_GC_SHADER_ARRAY_CONFIG__INACTIVE_WGPS_MASK;
+	data >>= CC_GC_SHADER_ARRAY_CONFIG__INACTIVE_WGPS__SHIFT;
+
+	wgp_bitmask =
+		amdgpu_gfx_create_bitmask(adev->gfx.config.max_cu_per_sh >> 1);
+
+	return (~data) & wgp_bitmask;
+}
+
+static u32 gfx_v12_0_get_cu_active_bitmap_per_sh(struct amdgpu_device *adev)
+{
+	u32 wgp_idx, wgp_active_bitmap;
+	u32 cu_bitmap_per_wgp, cu_active_bitmap;
+
+	wgp_active_bitmap = gfx_v12_0_get_wgp_active_bitmap_per_sh(adev);
+	cu_active_bitmap = 0;
+
+	for (wgp_idx = 0; wgp_idx < 16; wgp_idx++) {
+		/* if there is one WGP enabled, it means 2 CUs will be enabled */
+		cu_bitmap_per_wgp = 3 << (2 * wgp_idx);
+		if (wgp_active_bitmap & (1 << wgp_idx))
+			cu_active_bitmap |= cu_bitmap_per_wgp;
+	}
+
+	return cu_active_bitmap;
+}
+
+static int gfx_v12_0_get_cu_info(struct amdgpu_device *adev,
+				 struct amdgpu_cu_info *cu_info)
+{
+	int i, j, k, counter, active_cu_number = 0;
+	u32 mask, bitmap;
+	unsigned disable_masks[8 * 2];
+
+	if (!adev || !cu_info)
+		return -EINVAL;
+
+	amdgpu_gfx_parse_disable_cu(disable_masks, 8, 2);
+
+	mutex_lock(&adev->grbm_idx_mutex);
+	for (i = 0; i < adev->gfx.config.max_shader_engines; i++) {
+		for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) {
+			bitmap = i * adev->gfx.config.max_sh_per_se + j;
+			if (!((gfx_v12_0_get_sa_active_bitmap(adev) >> bitmap) & 1))
+				continue;
+			mask = 1;
+			counter = 0;
+			gfx_v12_0_select_se_sh(adev, i, j, 0xffffffff, 0);
+			if (i < 8 && j < 2)
+				gfx_v12_0_set_user_wgp_inactive_bitmap_per_sh(
+					adev, disable_masks[i * 2 + j]);
+			bitmap = gfx_v12_0_get_cu_active_bitmap_per_sh(adev);
+
+			/**
+			 * GFX12 could support more than 4 SEs, while the bitmap
+			 * in cu_info struct is 4x4 and ioctl interface struct
+			 * drm_amdgpu_info_device should keep stable.
+			 * So we use last two columns of bitmap to store cu mask for
+			 * SEs 4 to 7, the layout of the bitmap is as below:
+			 *    SE0: {SH0,SH1} --> {bitmap[0][0], bitmap[0][1]}
+			 *    SE1: {SH0,SH1} --> {bitmap[1][0], bitmap[1][1]}
+			 *    SE2: {SH0,SH1} --> {bitmap[2][0], bitmap[2][1]}
+			 *    SE3: {SH0,SH1} --> {bitmap[3][0], bitmap[3][1]}
+			 *    SE4: {SH0,SH1} --> {bitmap[0][2], bitmap[0][3]}
+			 *    SE5: {SH0,SH1} --> {bitmap[1][2], bitmap[1][3]}
+			 *    SE6: {SH0,SH1} --> {bitmap[2][2], bitmap[2][3]}
+			 *    SE7: {SH0,SH1} --> {bitmap[3][2], bitmap[3][3]}
+			 */
+			cu_info->bitmap[0][i % 4][j + (i / 4) * 2] = bitmap;
+
+			for (k = 0; k < adev->gfx.config.max_cu_per_sh; k++) {
+				if (bitmap & mask)
+					counter++;
+
+				mask <<= 1;
+			}
+			active_cu_number += counter;
+		}
+	}
+	gfx_v12_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
+	mutex_unlock(&adev->grbm_idx_mutex);
+
+	cu_info->number = active_cu_number;
+	cu_info->simd_per_cu = NUM_SIMD_PER_CU;
+
+	return 0;
+}
+
+const struct amdgpu_ip_block_version gfx_v12_0_ip_block = {
+	.type = AMD_IP_BLOCK_TYPE_GFX,
+	.major = 12,
+	.minor = 0,
+	.rev = 0,
+	.funcs = &gfx_v12_0_ip_funcs,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.h b/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.h
new file mode 100644
index 000000000000..f7184b2dc4e8
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.h
@@ -0,0 +1,31 @@
+/*
+ * Copyright 2023 dvanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __GFX_V12_0_H__
+#define __GFX_V12_0_H__
+
+extern const struct amdgpu_ip_block_version gfx_v12_0_ip_block;
+
+int gfx_v12_0_request_gfx_index_mutex(struct amdgpu_device *adev,
+				      bool req);
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v6_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v6_0.c
index 6a8dadea40f9..80565392313f 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v6_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v6_0.c
@@ -28,19 +28,33 @@
 #include "amdgpu_gfx.h"
 #include "amdgpu_ucode.h"
 #include "clearstate_si.h"
+#include "si.h"
+#include "sid.h"
+
 #include "bif/bif_3_0_d.h"
 #include "bif/bif_3_0_sh_mask.h"
+
 #include "oss/oss_1_0_d.h"
 #include "oss/oss_1_0_sh_mask.h"
+
 #include "gca/gfx_6_0_d.h"
 #include "gca/gfx_6_0_sh_mask.h"
+#include "gca/gfx_7_2_enum.h"
+
 #include "gmc/gmc_6_0_d.h"
 #include "gmc/gmc_6_0_sh_mask.h"
+
 #include "dce/dce_6_0_d.h"
 #include "dce/dce_6_0_sh_mask.h"
-#include "gca/gfx_7_2_enum.h"
+
 #include "si_enums.h"
-#include "si.h"
+
+#define TAHITI_GB_ADDR_CONFIG_GOLDEN        0x12011003
+#define VERDE_GB_ADDR_CONFIG_GOLDEN         0x12010002
+#define HAINAN_GB_ADDR_CONFIG_GOLDEN        0x02010001
+
+#define GFX6_NUM_GFX_RINGS     1
+#define GFX6_NUM_COMPUTE_RINGS 2
 
 static void gfx_v6_0_set_ring_funcs(struct amdgpu_device *adev);
 static void gfx_v6_0_set_irq_funcs(struct amdgpu_device *adev);
@@ -72,7 +86,7 @@ MODULE_FIRMWARE("amdgpu/hainan_ce.bin");
 MODULE_FIRMWARE("amdgpu/hainan_rlc.bin");
 
 static u32 gfx_v6_0_get_csb_size(struct amdgpu_device *adev);
-static void gfx_v6_0_get_csb_buffer(struct amdgpu_device *adev, volatile u32 *buffer);
+static void gfx_v6_0_get_csb_buffer(struct amdgpu_device *adev, u32 *buffer);
 //static void gfx_v6_0_init_cp_pg_table(struct amdgpu_device *adev);
 static void gfx_v6_0_init_pg(struct amdgpu_device *adev);
 
@@ -311,7 +325,6 @@ static const u32 verde_rlc_save_restore_register_list[] =
 static int gfx_v6_0_init_microcode(struct amdgpu_device *adev)
 {
 	const char *chip_name;
-	char fw_name[30];
 	int err;
 	const struct gfx_firmware_header_v1_0 *cp_hdr;
 	const struct rlc_firmware_header_v1_0 *rlc_hdr;
@@ -337,59 +350,49 @@ static int gfx_v6_0_init_microcode(struct amdgpu_device *adev)
 	default: BUG();
 	}
 
-	snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_pfp.bin", chip_name);
-	err = request_firmware(&adev->gfx.pfp_fw, fw_name, adev->dev);
-	if (err)
-		goto out;
-	err = amdgpu_ucode_validate(adev->gfx.pfp_fw);
+	err = amdgpu_ucode_request(adev, &adev->gfx.pfp_fw,
+				   AMDGPU_UCODE_REQUIRED,
+				   "amdgpu/%s_pfp.bin", chip_name);
 	if (err)
 		goto out;
 	cp_hdr = (const struct gfx_firmware_header_v1_0 *)adev->gfx.pfp_fw->data;
 	adev->gfx.pfp_fw_version = le32_to_cpu(cp_hdr->header.ucode_version);
 	adev->gfx.pfp_feature_version = le32_to_cpu(cp_hdr->ucode_feature_version);
 
-	snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_me.bin", chip_name);
-	err = request_firmware(&adev->gfx.me_fw, fw_name, adev->dev);
-	if (err)
-		goto out;
-	err = amdgpu_ucode_validate(adev->gfx.me_fw);
+	err = amdgpu_ucode_request(adev, &adev->gfx.me_fw,
+				   AMDGPU_UCODE_REQUIRED,
+				   "amdgpu/%s_me.bin", chip_name);
 	if (err)
 		goto out;
 	cp_hdr = (const struct gfx_firmware_header_v1_0 *)adev->gfx.me_fw->data;
 	adev->gfx.me_fw_version = le32_to_cpu(cp_hdr->header.ucode_version);
 	adev->gfx.me_feature_version = le32_to_cpu(cp_hdr->ucode_feature_version);
 
-	snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_ce.bin", chip_name);
-	err = request_firmware(&adev->gfx.ce_fw, fw_name, adev->dev);
-	if (err)
-		goto out;
-	err = amdgpu_ucode_validate(adev->gfx.ce_fw);
+	err = amdgpu_ucode_request(adev, &adev->gfx.ce_fw,
+				   AMDGPU_UCODE_REQUIRED,
+				   "amdgpu/%s_ce.bin", chip_name);
 	if (err)
 		goto out;
 	cp_hdr = (const struct gfx_firmware_header_v1_0 *)adev->gfx.ce_fw->data;
 	adev->gfx.ce_fw_version = le32_to_cpu(cp_hdr->header.ucode_version);
 	adev->gfx.ce_feature_version = le32_to_cpu(cp_hdr->ucode_feature_version);
 
-	snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_rlc.bin", chip_name);
-	err = request_firmware(&adev->gfx.rlc_fw, fw_name, adev->dev);
+	err = amdgpu_ucode_request(adev, &adev->gfx.rlc_fw,
+				   AMDGPU_UCODE_REQUIRED,
+				   "amdgpu/%s_rlc.bin", chip_name);
 	if (err)
 		goto out;
-	err = amdgpu_ucode_validate(adev->gfx.rlc_fw);
 	rlc_hdr = (const struct rlc_firmware_header_v1_0 *)adev->gfx.rlc_fw->data;
 	adev->gfx.rlc_fw_version = le32_to_cpu(rlc_hdr->header.ucode_version);
 	adev->gfx.rlc_feature_version = le32_to_cpu(rlc_hdr->ucode_feature_version);
 
 out:
 	if (err) {
-		pr_err("gfx6: Failed to load firmware \"%s\"\n", fw_name);
-		release_firmware(adev->gfx.pfp_fw);
-		adev->gfx.pfp_fw = NULL;
-		release_firmware(adev->gfx.me_fw);
-		adev->gfx.me_fw = NULL;
-		release_firmware(adev->gfx.ce_fw);
-		adev->gfx.ce_fw = NULL;
-		release_firmware(adev->gfx.rlc_fw);
-		adev->gfx.rlc_fw = NULL;
+		pr_err("gfx6: Failed to load firmware %s gfx firmware\n", chip_name);
+		amdgpu_ucode_release(&adev->gfx.pfp_fw);
+		amdgpu_ucode_release(&adev->gfx.me_fw);
+		amdgpu_ucode_release(&adev->gfx.ce_fw);
+		amdgpu_ucode_release(&adev->gfx.rlc_fw);
 	}
 	return err;
 }
@@ -1299,7 +1302,7 @@ static void gfx_v6_0_tiling_mode_table_init(struct amdgpu_device *adev)
 }
 
 static void gfx_v6_0_select_se_sh(struct amdgpu_device *adev, u32 se_num,
-				  u32 sh_num, u32 instance)
+				  u32 sh_num, u32 instance, int xcc_id)
 {
 	u32 data;
 
@@ -1452,12 +1455,12 @@ static void gfx_v6_0_write_harvested_raster_configs(struct amdgpu_device *adev,
 		}
 
 		/* GRBM_GFX_INDEX has a different offset on SI */
-		gfx_v6_0_select_se_sh(adev, se, 0xffffffff, 0xffffffff);
+		gfx_v6_0_select_se_sh(adev, se, 0xffffffff, 0xffffffff, 0);
 		WREG32(mmPA_SC_RASTER_CONFIG, raster_config_se);
 	}
 
 	/* GRBM_GFX_INDEX has a different offset on SI */
-	gfx_v6_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+	gfx_v6_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
 }
 
 static void gfx_v6_0_setup_rb(struct amdgpu_device *adev)
@@ -1473,14 +1476,14 @@ static void gfx_v6_0_setup_rb(struct amdgpu_device *adev)
 	mutex_lock(&adev->grbm_idx_mutex);
 	for (i = 0; i < adev->gfx.config.max_shader_engines; i++) {
 		for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) {
-			gfx_v6_0_select_se_sh(adev, i, j, 0xffffffff);
+			gfx_v6_0_select_se_sh(adev, i, j, 0xffffffff, 0);
 			data = gfx_v6_0_get_rb_active_bitmap(adev);
 			active_rbs |= data <<
 				((i * adev->gfx.config.max_sh_per_se + j) *
 				 rb_bitmap_width_per_sh);
 		}
 	}
-	gfx_v6_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+	gfx_v6_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
 
 	adev->gfx.config.backend_enable_mask = active_rbs;
 	adev->gfx.config.num_rbs = hweight32(active_rbs);
@@ -1501,7 +1504,7 @@ static void gfx_v6_0_setup_rb(struct amdgpu_device *adev)
 	/* cache the values for userspace */
 	for (i = 0; i < adev->gfx.config.max_shader_engines; i++) {
 		for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) {
-			gfx_v6_0_select_se_sh(adev, i, j, 0xffffffff);
+			gfx_v6_0_select_se_sh(adev, i, j, 0xffffffff, 0);
 			adev->gfx.config.rb_config[i][j].rb_backend_disable =
 				RREG32(mmCC_RB_BACKEND_DISABLE);
 			adev->gfx.config.rb_config[i][j].user_rb_backend_disable =
@@ -1510,7 +1513,7 @@ static void gfx_v6_0_setup_rb(struct amdgpu_device *adev)
 				RREG32(mmPA_SC_RASTER_CONFIG);
 		}
 	}
-	gfx_v6_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+	gfx_v6_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
 	mutex_unlock(&adev->grbm_idx_mutex);
 }
 
@@ -1549,7 +1552,7 @@ static void gfx_v6_0_setup_spi(struct amdgpu_device *adev)
 	mutex_lock(&adev->grbm_idx_mutex);
 	for (i = 0; i < adev->gfx.config.max_shader_engines; i++) {
 		for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) {
-			gfx_v6_0_select_se_sh(adev, i, j, 0xffffffff);
+			gfx_v6_0_select_se_sh(adev, i, j, 0xffffffff, 0);
 			data = RREG32(mmSPI_STATIC_THREAD_MGMT_3);
 			active_cu = gfx_v6_0_get_cu_enabled(adev);
 
@@ -1564,7 +1567,7 @@ static void gfx_v6_0_setup_spi(struct amdgpu_device *adev)
 			}
 		}
 	}
-	gfx_v6_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+	gfx_v6_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
 	mutex_unlock(&adev->grbm_idx_mutex);
 }
 
@@ -1732,10 +1735,14 @@ static void gfx_v6_0_constants_init(struct amdgpu_device *adev)
 	gfx_v6_0_get_cu_info(adev);
 	gfx_v6_0_config_init(adev);
 
-	WREG32(mmCP_QUEUE_THRESHOLDS, ((0x16 << CP_QUEUE_THRESHOLDS__ROQ_IB1_START__SHIFT) |
-				       (0x2b << CP_QUEUE_THRESHOLDS__ROQ_IB2_START__SHIFT)));
-	WREG32(mmCP_MEQ_THRESHOLDS, (0x30 << CP_MEQ_THRESHOLDS__MEQ1_START__SHIFT) |
-				    (0x60 << CP_MEQ_THRESHOLDS__MEQ2_START__SHIFT));
+	WREG32(mmCP_QUEUE_THRESHOLDS,
+		((0x16 << CP_QUEUE_THRESHOLDS__ROQ_IB1_START__SHIFT) |
+		(0x2b << CP_QUEUE_THRESHOLDS__ROQ_IB2_START__SHIFT)));
+
+	/* set HW defaults for 3D engine */
+	WREG32(mmCP_MEQ_THRESHOLDS,
+		(0x30 << CP_MEQ_THRESHOLDS__MEQ1_START__SHIFT) |
+		(0x60 << CP_MEQ_THRESHOLDS__MEQ2_START__SHIFT));
 
 	sx_debug_1 = RREG32(mmSX_DEBUG_1);
 	WREG32(mmSX_DEBUG_1, sx_debug_1);
@@ -1778,39 +1785,26 @@ static void gfx_v6_0_constants_init(struct amdgpu_device *adev)
 	udelay(50);
 }
 
-
-static void gfx_v6_0_scratch_init(struct amdgpu_device *adev)
-{
-	adev->gfx.scratch.num_reg = 8;
-	adev->gfx.scratch.reg_base = mmSCRATCH_REG0;
-	adev->gfx.scratch.free_mask = (1u << adev->gfx.scratch.num_reg) - 1;
-}
-
 static int gfx_v6_0_ring_test_ring(struct amdgpu_ring *ring)
 {
 	struct amdgpu_device *adev = ring->adev;
-	uint32_t scratch;
 	uint32_t tmp = 0;
 	unsigned i;
 	int r;
 
-	r = amdgpu_gfx_scratch_get(adev, &scratch);
-	if (r)
-		return r;
-
-	WREG32(scratch, 0xCAFEDEAD);
+	WREG32(mmSCRATCH_REG0, 0xCAFEDEAD);
 
 	r = amdgpu_ring_alloc(ring, 3);
 	if (r)
-		goto error_free_scratch;
+		return r;
 
 	amdgpu_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1));
-	amdgpu_ring_write(ring, (scratch - PACKET3_SET_CONFIG_REG_START));
+	amdgpu_ring_write(ring, mmSCRATCH_REG0 - PACKET3_SET_CONFIG_REG_START);
 	amdgpu_ring_write(ring, 0xDEADBEEF);
 	amdgpu_ring_commit(ring);
 
 	for (i = 0; i < adev->usec_timeout; i++) {
-		tmp = RREG32(scratch);
+		tmp = RREG32(mmSCRATCH_REG0);
 		if (tmp == 0xDEADBEEF)
 			break;
 		udelay(1);
@@ -1818,9 +1812,6 @@ static int gfx_v6_0_ring_test_ring(struct amdgpu_ring *ring)
 
 	if (i >= adev->usec_timeout)
 		r = -ETIMEDOUT;
-
-error_free_scratch:
-	amdgpu_gfx_scratch_free(adev, scratch);
 	return r;
 }
 
@@ -1903,50 +1894,42 @@ static void gfx_v6_0_ring_emit_ib(struct amdgpu_ring *ring,
 static int gfx_v6_0_ring_test_ib(struct amdgpu_ring *ring, long timeout)
 {
 	struct amdgpu_device *adev = ring->adev;
-	struct amdgpu_ib ib;
 	struct dma_fence *f = NULL;
-	uint32_t scratch;
+	struct amdgpu_ib ib;
 	uint32_t tmp = 0;
 	long r;
 
-	r = amdgpu_gfx_scratch_get(adev, &scratch);
-	if (r)
-		return r;
-
-	WREG32(scratch, 0xCAFEDEAD);
+	WREG32(mmSCRATCH_REG0, 0xCAFEDEAD);
 	memset(&ib, 0, sizeof(ib));
-	r = amdgpu_ib_get(adev, NULL, 256,
-					AMDGPU_IB_POOL_DIRECT, &ib);
+	r = amdgpu_ib_get(adev, NULL, 256, AMDGPU_IB_POOL_DIRECT, &ib);
 	if (r)
-		goto err1;
+		return r;
 
 	ib.ptr[0] = PACKET3(PACKET3_SET_CONFIG_REG, 1);
-	ib.ptr[1] = ((scratch - PACKET3_SET_CONFIG_REG_START));
+	ib.ptr[1] = mmSCRATCH_REG0 - PACKET3_SET_CONFIG_REG_START;
 	ib.ptr[2] = 0xDEADBEEF;
 	ib.length_dw = 3;
 
 	r = amdgpu_ib_schedule(ring, 1, &ib, NULL, &f);
 	if (r)
-		goto err2;
+		goto error;
 
 	r = dma_fence_wait_timeout(f, false, timeout);
 	if (r == 0) {
 		r = -ETIMEDOUT;
-		goto err2;
+		goto error;
 	} else if (r < 0) {
-		goto err2;
+		goto error;
 	}
-	tmp = RREG32(scratch);
+	tmp = RREG32(mmSCRATCH_REG0);
 	if (tmp == 0xDEADBEEF)
 		r = 0;
 	else
 		r = -EINVAL;
 
-err2:
-	amdgpu_ib_free(adev, &ib, NULL);
+error:
+	amdgpu_ib_free(&ib, NULL);
 	dma_fence_put(f);
-err1:
-	amdgpu_gfx_scratch_free(adev, scratch);
 	return r;
 }
 
@@ -2117,7 +2100,7 @@ static int gfx_v6_0_cp_gfx_resume(struct amdgpu_device *adev)
 	WREG32(mmCP_RB0_WPTR, ring->wptr);
 
 	/* set the wb address whether it's enabled or not */
-	rptr_addr = adev->wb.gpu_addr + (ring->rptr_offs * 4);
+	rptr_addr = ring->rptr_gpu_addr;
 	WREG32(mmCP_RB0_RPTR_ADDR, lower_32_bits(rptr_addr));
 	WREG32(mmCP_RB0_RPTR_ADDR_HI, upper_32_bits(rptr_addr) & 0xFF);
 
@@ -2139,7 +2122,7 @@ static int gfx_v6_0_cp_gfx_resume(struct amdgpu_device *adev)
 
 static u64 gfx_v6_0_ring_get_rptr(struct amdgpu_ring *ring)
 {
-	return ring->adev->wb.wb[ring->rptr_offs];
+	return *ring->rptr_cpu_addr;
 }
 
 static u64 gfx_v6_0_ring_get_wptr(struct amdgpu_ring *ring)
@@ -2203,7 +2186,7 @@ static int gfx_v6_0_cp_compute_resume(struct amdgpu_device *adev)
 	ring->wptr = 0;
 	WREG32(mmCP_RB1_WPTR, ring->wptr);
 
-	rptr_addr = adev->wb.gpu_addr + (ring->rptr_offs * 4);
+	rptr_addr = ring->rptr_gpu_addr;
 	WREG32(mmCP_RB1_RPTR_ADDR, lower_32_bits(rptr_addr));
 	WREG32(mmCP_RB1_RPTR_ADDR_HI, upper_32_bits(rptr_addr) & 0xFF);
 
@@ -2222,7 +2205,7 @@ static int gfx_v6_0_cp_compute_resume(struct amdgpu_device *adev)
 	WREG32(mmCP_RB2_CNTL, tmp | CP_RB2_CNTL__RB_RPTR_WR_ENA_MASK);
 	ring->wptr = 0;
 	WREG32(mmCP_RB2_WPTR, ring->wptr);
-	rptr_addr = adev->wb.gpu_addr + (ring->rptr_offs * 4);
+	rptr_addr = ring->rptr_gpu_addr;
 	WREG32(mmCP_RB2_RPTR_ADDR, lower_32_bits(rptr_addr));
 	WREG32(mmCP_RB2_RPTR_ADDR_HI, upper_32_bits(rptr_addr) & 0xFF);
 
@@ -2371,7 +2354,7 @@ static void gfx_v6_0_ring_emit_wreg(struct amdgpu_ring *ring,
 static int gfx_v6_0_rlc_init(struct amdgpu_device *adev)
 {
 	const u32 *src_ptr;
-	volatile u32 *dst_ptr;
+	u32 *dst_ptr;
 	u32 dws;
 	u64 reg_list_mc_addr;
 	const struct cs_section_def *cs_data;
@@ -2399,7 +2382,8 @@ static int gfx_v6_0_rlc_init(struct amdgpu_device *adev)
 		dws = adev->gfx.rlc.clear_state_size + (256 / 4);
 
 		r = amdgpu_bo_create_reserved(adev, dws * 4, PAGE_SIZE,
-					      AMDGPU_GEM_DOMAIN_VRAM,
+					      AMDGPU_GEM_DOMAIN_VRAM |
+					      AMDGPU_GEM_DOMAIN_GTT,
 					      &adev->gfx.rlc.clear_state_obj,
 					      &adev->gfx.rlc.clear_state_gpu_addr,
 					      (void **)&adev->gfx.rlc.cs_ptr);
@@ -2428,7 +2412,7 @@ static void gfx_v6_0_enable_lbpw(struct amdgpu_device *adev, bool enable)
 	WREG32_FIELD(RLC_LB_CNTL, LOAD_BALANCE_ENABLE, enable ? 1 : 0);
 
 	if (!enable) {
-		gfx_v6_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+		gfx_v6_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
 		WREG32(mmSPI_LB_CU_MASK, 0x00ff);
 	}
 }
@@ -2871,47 +2855,23 @@ static u32 gfx_v6_0_get_csb_size(struct amdgpu_device *adev)
 	return count;
 }
 
-static void gfx_v6_0_get_csb_buffer(struct amdgpu_device *adev,
-				    volatile u32 *buffer)
+static void gfx_v6_0_get_csb_buffer(struct amdgpu_device *adev, u32 *buffer)
 {
-	u32 count = 0, i;
-	const struct cs_section_def *sect = NULL;
-	const struct cs_extent_def *ext = NULL;
+	u32 count = 0;
 
 	if (adev->gfx.rlc.cs_data == NULL)
 		return;
 	if (buffer == NULL)
 		return;
 
-	buffer[count++] = cpu_to_le32(PACKET3(PACKET3_PREAMBLE_CNTL, 0));
-	buffer[count++] = cpu_to_le32(PACKET3_PREAMBLE_BEGIN_CLEAR_STATE);
-	buffer[count++] = cpu_to_le32(PACKET3(PACKET3_CONTEXT_CONTROL, 1));
-	buffer[count++] = cpu_to_le32(0x80000000);
-	buffer[count++] = cpu_to_le32(0x80000000);
-
-	for (sect = adev->gfx.rlc.cs_data; sect->section != NULL; ++sect) {
-		for (ext = sect->section; ext->extent != NULL; ++ext) {
-			if (sect->id == SECT_CONTEXT) {
-				buffer[count++] =
-					cpu_to_le32(PACKET3(PACKET3_SET_CONTEXT_REG, ext->reg_count));
-				buffer[count++] = cpu_to_le32(ext->reg_index - 0xa000);
-				for (i = 0; i < ext->reg_count; i++)
-					buffer[count++] = cpu_to_le32(ext->extent[i]);
-			} else {
-				return;
-			}
-		}
-	}
+	count = amdgpu_gfx_csb_preamble_start(buffer);
+	count = amdgpu_gfx_csb_data_parser(adev, buffer, count);
 
 	buffer[count++] = cpu_to_le32(PACKET3(PACKET3_SET_CONTEXT_REG, 1));
 	buffer[count++] = cpu_to_le32(mmPA_SC_RASTER_CONFIG - PACKET3_SET_CONTEXT_REG_START);
 	buffer[count++] = cpu_to_le32(adev->gfx.config.rb_config[0][0].raster_config);
 
-	buffer[count++] = cpu_to_le32(PACKET3(PACKET3_PREAMBLE_CNTL, 0));
-	buffer[count++] = cpu_to_le32(PACKET3_PREAMBLE_END_CLEAR_STATE);
-
-	buffer[count++] = cpu_to_le32(PACKET3(PACKET3_CLEAR_STATE, 0));
-	buffer[count++] = cpu_to_le32(0);
+	amdgpu_gfx_csb_preamble_end(buffer, count);
 }
 
 static void gfx_v6_0_init_pg(struct amdgpu_device *adev)
@@ -3005,7 +2965,7 @@ static void wave_read_regs(struct amdgpu_device *adev, uint32_t simd,
 		*(out++) = RREG32(mmSQ_IND_DATA);
 }
 
-static void gfx_v6_0_read_wave_data(struct amdgpu_device *adev, uint32_t simd, uint32_t wave, uint32_t *dst, int *no_fields)
+static void gfx_v6_0_read_wave_data(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd, uint32_t wave, uint32_t *dst, int *no_fields)
 {
 	/* type 0 wave data */
 	dst[(*no_fields)++] = 0;
@@ -3030,7 +2990,7 @@ static void gfx_v6_0_read_wave_data(struct amdgpu_device *adev, uint32_t simd, u
 	dst[(*no_fields)++] = wave_read_ind(adev, simd, wave, ixSQ_WAVE_MODE);
 }
 
-static void gfx_v6_0_read_wave_sgprs(struct amdgpu_device *adev, uint32_t simd,
+static void gfx_v6_0_read_wave_sgprs(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd,
 				     uint32_t wave, uint32_t start,
 				     uint32_t size, uint32_t *dst)
 {
@@ -3040,7 +3000,7 @@ static void gfx_v6_0_read_wave_sgprs(struct amdgpu_device *adev, uint32_t simd,
 }
 
 static void gfx_v6_0_select_me_pipe_q(struct amdgpu_device *adev,
-				  u32 me, u32 pipe, u32 q, u32 vm)
+				  u32 me, u32 pipe, u32 q, u32 vm, u32 xcc_id)
 {
 	DRM_INFO("Not implemented\n");
 }
@@ -3061,10 +3021,11 @@ static const struct amdgpu_rlc_funcs gfx_v6_0_rlc_funcs = {
 	.start = gfx_v6_0_rlc_start
 };
 
-static int gfx_v6_0_early_init(void *handle)
+static int gfx_v6_0_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
+	adev->gfx.xcc_mask = 1;
 	adev->gfx.num_gfx_rings = GFX6_NUM_GFX_RINGS;
 	adev->gfx.num_compute_rings = min(amdgpu_gfx_get_num_kcq(adev),
 					  GFX6_NUM_COMPUTE_RINGS);
@@ -3076,10 +3037,10 @@ static int gfx_v6_0_early_init(void *handle)
 	return 0;
 }
 
-static int gfx_v6_0_sw_init(void *handle)
+static int gfx_v6_0_sw_init(struct amdgpu_ip_block *ip_block)
 {
 	struct amdgpu_ring *ring;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int i, r;
 
 	r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 181, &adev->gfx.eop_irq);
@@ -3094,8 +3055,6 @@ static int gfx_v6_0_sw_init(void *handle)
 	if (r)
 		return r;
 
-	gfx_v6_0_scratch_init(adev);
-
 	r = gfx_v6_0_init_microcode(adev);
 	if (r) {
 		DRM_ERROR("Failed to load gfx firmware!\n");
@@ -3112,7 +3071,7 @@ static int gfx_v6_0_sw_init(void *handle)
 		ring = &adev->gfx.gfx_ring[i];
 		ring->ring_obj = NULL;
 		sprintf(ring->name, "gfx");
-		r = amdgpu_ring_init(adev, ring, 1024,
+		r = amdgpu_ring_init(adev, ring, 2048,
 				     &adev->gfx.eop_irq,
 				     AMDGPU_CP_IRQ_GFX_ME0_PIPE0_EOP,
 				     AMDGPU_RING_PRIO_DEFAULT, NULL);
@@ -3143,13 +3102,18 @@ static int gfx_v6_0_sw_init(void *handle)
 			return r;
 	}
 
+	adev->gfx.gfx_supported_reset =
+		amdgpu_get_soft_full_reset_mask(&adev->gfx.gfx_ring[0]);
+	adev->gfx.compute_supported_reset =
+		amdgpu_get_soft_full_reset_mask(&adev->gfx.compute_ring[0]);
+
 	return r;
 }
 
-static int gfx_v6_0_sw_fini(void *handle)
+static int gfx_v6_0_sw_fini(struct amdgpu_ip_block *ip_block)
 {
 	int i;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	for (i = 0; i < adev->gfx.num_gfx_rings; i++)
 		amdgpu_ring_fini(&adev->gfx.gfx_ring[i]);
@@ -3161,10 +3125,10 @@ static int gfx_v6_0_sw_fini(void *handle)
 	return 0;
 }
 
-static int gfx_v6_0_hw_init(void *handle)
+static int gfx_v6_0_hw_init(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	gfx_v6_0_constants_init(adev);
 
@@ -3181,9 +3145,9 @@ static int gfx_v6_0_hw_init(void *handle)
 	return r;
 }
 
-static int gfx_v6_0_hw_fini(void *handle)
+static int gfx_v6_0_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	gfx_v6_0_cp_enable(adev, false);
 	adev->gfx.rlc.funcs->stop(adev);
@@ -3192,23 +3156,19 @@ static int gfx_v6_0_hw_fini(void *handle)
 	return 0;
 }
 
-static int gfx_v6_0_suspend(void *handle)
+static int gfx_v6_0_suspend(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	return gfx_v6_0_hw_fini(adev);
+	return gfx_v6_0_hw_fini(ip_block);
 }
 
-static int gfx_v6_0_resume(void *handle)
+static int gfx_v6_0_resume(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	return gfx_v6_0_hw_init(adev);
+	return gfx_v6_0_hw_init(ip_block);
 }
 
-static bool gfx_v6_0_is_idle(void *handle)
+static bool gfx_v6_0_is_idle(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (RREG32(mmGRBM_STATUS) & GRBM_STATUS__GUI_ACTIVE_MASK)
 		return false;
@@ -3216,24 +3176,19 @@ static bool gfx_v6_0_is_idle(void *handle)
 		return true;
 }
 
-static int gfx_v6_0_wait_for_idle(void *handle)
+static int gfx_v6_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
 {
 	unsigned i;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	for (i = 0; i < adev->usec_timeout; i++) {
-		if (gfx_v6_0_is_idle(handle))
+		if (gfx_v6_0_is_idle(ip_block))
 			return 0;
 		udelay(1);
 	}
 	return -ETIMEDOUT;
 }
 
-static int gfx_v6_0_soft_reset(void *handle)
-{
-	return 0;
-}
-
 static void gfx_v6_0_set_gfx_eop_interrupt_state(struct amdgpu_device *adev,
 						 enum amdgpu_interrupt_state state)
 {
@@ -3421,11 +3376,11 @@ static int gfx_v6_0_priv_inst_irq(struct amdgpu_device *adev,
 	return 0;
 }
 
-static int gfx_v6_0_set_clockgating_state(void *handle,
+static int gfx_v6_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_clockgating_state state)
 {
 	bool gate = false;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (state == AMD_CG_STATE_GATE)
 		gate = true;
@@ -3443,11 +3398,11 @@ static int gfx_v6_0_set_clockgating_state(void *handle,
 	return 0;
 }
 
-static int gfx_v6_0_set_powergating_state(void *handle,
+static int gfx_v6_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_powergating_state state)
 {
 	bool gate = false;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (state == AMD_PG_STATE_GATE)
 		gate = true;
@@ -3483,7 +3438,6 @@ static void gfx_v6_0_emit_mem_sync(struct amdgpu_ring *ring)
 static const struct amd_ip_funcs gfx_v6_0_ip_funcs = {
 	.name = "gfx_v6_0",
 	.early_init = gfx_v6_0_early_init,
-	.late_init = NULL,
 	.sw_init = gfx_v6_0_sw_init,
 	.sw_fini = gfx_v6_0_sw_fini,
 	.hw_init = gfx_v6_0_hw_init,
@@ -3492,7 +3446,6 @@ static const struct amd_ip_funcs gfx_v6_0_ip_funcs = {
 	.resume = gfx_v6_0_resume,
 	.is_idle = gfx_v6_0_is_idle,
 	.wait_for_idle = gfx_v6_0_wait_for_idle,
-	.soft_reset = gfx_v6_0_soft_reset,
 	.set_clockgating_state = gfx_v6_0_set_clockgating_state,
 	.set_powergating_state = gfx_v6_0_set_powergating_state,
 };
@@ -3610,12 +3563,12 @@ static void gfx_v6_0_get_cu_info(struct amdgpu_device *adev)
 			mask = 1;
 			ao_bitmap = 0;
 			counter = 0;
-			gfx_v6_0_select_se_sh(adev, i, j, 0xffffffff);
+			gfx_v6_0_select_se_sh(adev, i, j, 0xffffffff, 0);
 			if (i < 4 && j < 2)
 				gfx_v6_0_set_user_cu_inactive_bitmap(
 					adev, disable_masks[i * 2 + j]);
 			bitmap = gfx_v6_0_get_cu_enabled(adev);
-			cu_info->bitmap[i][j] = bitmap;
+			cu_info->bitmap[0][i][j] = bitmap;
 
 			for (k = 0; k < adev->gfx.config.max_cu_per_sh; k++) {
 				if (bitmap & mask) {
@@ -3632,7 +3585,7 @@ static void gfx_v6_0_get_cu_info(struct amdgpu_device *adev)
 		}
 	}
 
-	gfx_v6_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+	gfx_v6_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
 	mutex_unlock(&adev->grbm_idx_mutex);
 
 	cu_info->number = active_cu_number;
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c
index d17a6f399347..2b7aba22ecc1 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c
@@ -55,6 +55,9 @@
 #define GFX7_NUM_GFX_RINGS     1
 #define GFX7_MEC_HPD_SIZE      2048
 
+#define BONAIRE_GB_ADDR_CONFIG_GOLDEN        0x12010001
+#define HAWAII_GB_ADDR_CONFIG_GOLDEN         0x12011003
+
 static void gfx_v7_0_set_ring_funcs(struct amdgpu_device *adev);
 static void gfx_v7_0_set_irq_funcs(struct amdgpu_device *adev);
 static void gfx_v7_0_set_gds_init(struct amdgpu_device *adev);
@@ -90,8 +93,7 @@ MODULE_FIRMWARE("amdgpu/mullins_ce.bin");
 MODULE_FIRMWARE("amdgpu/mullins_rlc.bin");
 MODULE_FIRMWARE("amdgpu/mullins_mec.bin");
 
-static const struct amdgpu_gds_reg_offset amdgpu_gds_reg_offset[] =
-{
+static const struct amdgpu_gds_reg_offset amdgpu_gds_reg_offset[] = {
 	{mmGDS_VMID0_BASE, mmGDS_VMID0_SIZE, mmGDS_GWS_VMID0, mmGDS_OA_VMID0},
 	{mmGDS_VMID1_BASE, mmGDS_VMID1_SIZE, mmGDS_GWS_VMID1, mmGDS_OA_VMID1},
 	{mmGDS_VMID2_BASE, mmGDS_VMID2_SIZE, mmGDS_GWS_VMID2, mmGDS_OA_VMID2},
@@ -110,8 +112,7 @@ static const struct amdgpu_gds_reg_offset amdgpu_gds_reg_offset[] =
 	{mmGDS_VMID15_BASE, mmGDS_VMID15_SIZE, mmGDS_GWS_VMID15, mmGDS_OA_VMID15}
 };
 
-static const u32 spectre_rlc_save_restore_register_list[] =
-{
+static const u32 spectre_rlc_save_restore_register_list[] = {
 	(0x0e00 << 16) | (0xc12c >> 2),
 	0x00000000,
 	(0x0e00 << 16) | (0xc140 >> 2),
@@ -557,8 +558,7 @@ static const u32 spectre_rlc_save_restore_register_list[] =
 	(0x0e00 << 16) | (0x9600 >> 2),
 };
 
-static const u32 kalindi_rlc_save_restore_register_list[] =
-{
+static const u32 kalindi_rlc_save_restore_register_list[] = {
 	(0x0e00 << 16) | (0xc12c >> 2),
 	0x00000000,
 	(0x0e00 << 16) | (0xc140 >> 2),
@@ -883,10 +883,20 @@ static const u32 kalindi_rlc_save_restore_register_list[] =
 };
 
 static u32 gfx_v7_0_get_csb_size(struct amdgpu_device *adev);
-static void gfx_v7_0_get_csb_buffer(struct amdgpu_device *adev, volatile u32 *buffer);
+static void gfx_v7_0_get_csb_buffer(struct amdgpu_device *adev, u32 *buffer);
 static void gfx_v7_0_init_pg(struct amdgpu_device *adev);
 static void gfx_v7_0_get_cu_info(struct amdgpu_device *adev);
 
+static void gfx_v7_0_free_microcode(struct amdgpu_device *adev)
+{
+	amdgpu_ucode_release(&adev->gfx.pfp_fw);
+	amdgpu_ucode_release(&adev->gfx.me_fw);
+	amdgpu_ucode_release(&adev->gfx.ce_fw);
+	amdgpu_ucode_release(&adev->gfx.mec_fw);
+	amdgpu_ucode_release(&adev->gfx.mec2_fw);
+	amdgpu_ucode_release(&adev->gfx.rlc_fw);
+}
+
 /*
  * Core functions
  */
@@ -902,7 +912,6 @@ static void gfx_v7_0_get_cu_info(struct amdgpu_device *adev);
 static int gfx_v7_0_init_microcode(struct amdgpu_device *adev)
 {
 	const char *chip_name;
-	char fw_name[30];
 	int err;
 
 	DRM_DEBUG("\n");
@@ -923,92 +932,53 @@ static int gfx_v7_0_init_microcode(struct amdgpu_device *adev)
 	case CHIP_MULLINS:
 		chip_name = "mullins";
 		break;
-	default: BUG();
+	default:
+		BUG();
 	}
 
-	snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_pfp.bin", chip_name);
-	err = request_firmware(&adev->gfx.pfp_fw, fw_name, adev->dev);
-	if (err)
-		goto out;
-	err = amdgpu_ucode_validate(adev->gfx.pfp_fw);
+	err = amdgpu_ucode_request(adev, &adev->gfx.pfp_fw,
+				   AMDGPU_UCODE_REQUIRED,
+				   "amdgpu/%s_pfp.bin", chip_name);
 	if (err)
 		goto out;
 
-	snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_me.bin", chip_name);
-	err = request_firmware(&adev->gfx.me_fw, fw_name, adev->dev);
-	if (err)
-		goto out;
-	err = amdgpu_ucode_validate(adev->gfx.me_fw);
+	err = amdgpu_ucode_request(adev, &adev->gfx.me_fw,
+				   AMDGPU_UCODE_REQUIRED,
+				   "amdgpu/%s_me.bin", chip_name);
 	if (err)
 		goto out;
 
-	snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_ce.bin", chip_name);
-	err = request_firmware(&adev->gfx.ce_fw, fw_name, adev->dev);
-	if (err)
-		goto out;
-	err = amdgpu_ucode_validate(adev->gfx.ce_fw);
+	err = amdgpu_ucode_request(adev, &adev->gfx.ce_fw,
+				   AMDGPU_UCODE_REQUIRED,
+				   "amdgpu/%s_ce.bin", chip_name);
 	if (err)
 		goto out;
 
-	snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_mec.bin", chip_name);
-	err = request_firmware(&adev->gfx.mec_fw, fw_name, adev->dev);
-	if (err)
-		goto out;
-	err = amdgpu_ucode_validate(adev->gfx.mec_fw);
+	err = amdgpu_ucode_request(adev, &adev->gfx.mec_fw,
+				   AMDGPU_UCODE_REQUIRED,
+				   "amdgpu/%s_mec.bin", chip_name);
 	if (err)
 		goto out;
 
 	if (adev->asic_type == CHIP_KAVERI) {
-		snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_mec2.bin", chip_name);
-		err = request_firmware(&adev->gfx.mec2_fw, fw_name, adev->dev);
-		if (err)
-			goto out;
-		err = amdgpu_ucode_validate(adev->gfx.mec2_fw);
+		err = amdgpu_ucode_request(adev, &adev->gfx.mec2_fw,
+					   AMDGPU_UCODE_REQUIRED,
+					   "amdgpu/%s_mec2.bin", chip_name);
 		if (err)
 			goto out;
 	}
 
-	snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_rlc.bin", chip_name);
-	err = request_firmware(&adev->gfx.rlc_fw, fw_name, adev->dev);
-	if (err)
-		goto out;
-	err = amdgpu_ucode_validate(adev->gfx.rlc_fw);
-
+	err = amdgpu_ucode_request(adev, &adev->gfx.rlc_fw,
+				   AMDGPU_UCODE_REQUIRED,
+				   "amdgpu/%s_rlc.bin", chip_name);
 out:
 	if (err) {
-		pr_err("gfx7: Failed to load firmware \"%s\"\n", fw_name);
-		release_firmware(adev->gfx.pfp_fw);
-		adev->gfx.pfp_fw = NULL;
-		release_firmware(adev->gfx.me_fw);
-		adev->gfx.me_fw = NULL;
-		release_firmware(adev->gfx.ce_fw);
-		adev->gfx.ce_fw = NULL;
-		release_firmware(adev->gfx.mec_fw);
-		adev->gfx.mec_fw = NULL;
-		release_firmware(adev->gfx.mec2_fw);
-		adev->gfx.mec2_fw = NULL;
-		release_firmware(adev->gfx.rlc_fw);
-		adev->gfx.rlc_fw = NULL;
+		pr_err("gfx7: Failed to load firmware %s gfx firmware\n", chip_name);
+		gfx_v7_0_free_microcode(adev);
 	}
 	return err;
 }
 
-static void gfx_v7_0_free_microcode(struct amdgpu_device *adev)
-{
-	release_firmware(adev->gfx.pfp_fw);
-	adev->gfx.pfp_fw = NULL;
-	release_firmware(adev->gfx.me_fw);
-	adev->gfx.me_fw = NULL;
-	release_firmware(adev->gfx.ce_fw);
-	adev->gfx.ce_fw = NULL;
-	release_firmware(adev->gfx.mec_fw);
-	adev->gfx.mec_fw = NULL;
-	release_firmware(adev->gfx.mec2_fw);
-	adev->gfx.mec2_fw = NULL;
-	release_firmware(adev->gfx.rlc_fw);
-	adev->gfx.rlc_fw = NULL;
-}
-
 /**
  * gfx_v7_0_tiling_mode_table_init - init the hw tiling table
  *
@@ -1582,11 +1552,12 @@ static void gfx_v7_0_tiling_mode_table_init(struct amdgpu_device *adev)
  * @sh_num: sh block to address
  * @instance: Certain registers are instanced per SE or SH.
  *            0xffffffff means broadcast to all SEs or SHs (CIK).
- *
+ * @xcc_id: xcc accelerated compute core id
  * Select which SE, SH combinations to address.
  */
 static void gfx_v7_0_select_se_sh(struct amdgpu_device *adev,
-				  u32 se_num, u32 sh_num, u32 instance)
+				  u32 se_num, u32 sh_num, u32 instance,
+				  int xcc_id)
 {
 	u32 data;
 
@@ -1766,13 +1737,13 @@ gfx_v7_0_write_harvested_raster_configs(struct amdgpu_device *adev,
 		}
 
 		/* GRBM_GFX_INDEX has a different offset on CI+ */
-		gfx_v7_0_select_se_sh(adev, se, 0xffffffff, 0xffffffff);
+		gfx_v7_0_select_se_sh(adev, se, 0xffffffff, 0xffffffff, 0);
 		WREG32(mmPA_SC_RASTER_CONFIG, raster_config_se);
 		WREG32(mmPA_SC_RASTER_CONFIG_1, raster_config_1);
 	}
 
 	/* GRBM_GFX_INDEX has a different offset on CI+ */
-	gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+	gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
 }
 
 /**
@@ -1795,13 +1766,13 @@ static void gfx_v7_0_setup_rb(struct amdgpu_device *adev)
 	mutex_lock(&adev->grbm_idx_mutex);
 	for (i = 0; i < adev->gfx.config.max_shader_engines; i++) {
 		for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) {
-			gfx_v7_0_select_se_sh(adev, i, j, 0xffffffff);
+			gfx_v7_0_select_se_sh(adev, i, j, 0xffffffff, 0);
 			data = gfx_v7_0_get_rb_active_bitmap(adev);
 			active_rbs |= data << ((i * adev->gfx.config.max_sh_per_se + j) *
 					       rb_bitmap_width_per_sh);
 		}
 	}
-	gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+	gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
 
 	adev->gfx.config.backend_enable_mask = active_rbs;
 	adev->gfx.config.num_rbs = hweight32(active_rbs);
@@ -1824,7 +1795,7 @@ static void gfx_v7_0_setup_rb(struct amdgpu_device *adev)
 	/* cache the values for userspace */
 	for (i = 0; i < adev->gfx.config.max_shader_engines; i++) {
 		for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) {
-			gfx_v7_0_select_se_sh(adev, i, j, 0xffffffff);
+			gfx_v7_0_select_se_sh(adev, i, j, 0xffffffff, 0);
 			adev->gfx.config.rb_config[i][j].rb_backend_disable =
 				RREG32(mmCC_RB_BACKEND_DISABLE);
 			adev->gfx.config.rb_config[i][j].user_rb_backend_disable =
@@ -1835,7 +1806,7 @@ static void gfx_v7_0_setup_rb(struct amdgpu_device *adev)
 				RREG32(mmPA_SC_RASTER_CONFIG_1);
 		}
 	}
-	gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+	gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
 	mutex_unlock(&adev->grbm_idx_mutex);
 }
 
@@ -1945,7 +1916,7 @@ static void gfx_v7_0_constants_init(struct amdgpu_device *adev)
 	 * making sure that the following register writes will be broadcasted
 	 * to all the shaders
 	 */
-	gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+	gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
 
 	/* XXX SH_MEM regs */
 	/* where to put LDS, scratch, GPUVM in FSA64 space */
@@ -2049,26 +2020,6 @@ static void gfx_v7_0_constants_init(struct amdgpu_device *adev)
 	udelay(50);
 }
 
-/*
- * GPU scratch registers helpers function.
- */
-/**
- * gfx_v7_0_scratch_init - setup driver info for CP scratch regs
- *
- * @adev: amdgpu_device pointer
- *
- * Set up the number and offset of the CP scratch registers.
- * NOTE: use of CP scratch registers is a legacy interface and
- * is not used by default on newer asics (r6xx+).  On newer asics,
- * memory buffers are used for fences rather than scratch regs.
- */
-static void gfx_v7_0_scratch_init(struct amdgpu_device *adev)
-{
-	adev->gfx.scratch.num_reg = 8;
-	adev->gfx.scratch.reg_base = mmSCRATCH_REG0;
-	adev->gfx.scratch.free_mask = (1u << adev->gfx.scratch.num_reg) - 1;
-}
-
 /**
  * gfx_v7_0_ring_test_ring - basic gfx ring test
  *
@@ -2082,36 +2033,28 @@ static void gfx_v7_0_scratch_init(struct amdgpu_device *adev)
 static int gfx_v7_0_ring_test_ring(struct amdgpu_ring *ring)
 {
 	struct amdgpu_device *adev = ring->adev;
-	uint32_t scratch;
 	uint32_t tmp = 0;
 	unsigned i;
 	int r;
 
-	r = amdgpu_gfx_scratch_get(adev, &scratch);
-	if (r)
-		return r;
-
-	WREG32(scratch, 0xCAFEDEAD);
+	WREG32(mmSCRATCH_REG0, 0xCAFEDEAD);
 	r = amdgpu_ring_alloc(ring, 3);
 	if (r)
-		goto error_free_scratch;
+		return r;
 
 	amdgpu_ring_write(ring, PACKET3(PACKET3_SET_UCONFIG_REG, 1));
-	amdgpu_ring_write(ring, (scratch - PACKET3_SET_UCONFIG_REG_START));
+	amdgpu_ring_write(ring, mmSCRATCH_REG0 - PACKET3_SET_UCONFIG_REG_START);
 	amdgpu_ring_write(ring, 0xDEADBEEF);
 	amdgpu_ring_commit(ring);
 
 	for (i = 0; i < adev->usec_timeout; i++) {
-		tmp = RREG32(scratch);
+		tmp = RREG32(mmSCRATCH_REG0);
 		if (tmp == 0xDEADBEEF)
 			break;
 		udelay(1);
 	}
 	if (i >= adev->usec_timeout)
 		r = -ETIMEDOUT;
-
-error_free_scratch:
-	amdgpu_gfx_scratch_free(adev, scratch);
 	return r;
 }
 
@@ -2180,6 +2123,8 @@ static void gfx_v7_0_ring_emit_fence_gfx(struct amdgpu_ring *ring, u64 addr,
 {
 	bool write64bit = flags & AMDGPU_FENCE_FLAG_64BIT;
 	bool int_sel = flags & AMDGPU_FENCE_FLAG_INT;
+	bool exec = flags & AMDGPU_FENCE_FLAG_EXEC;
+
 	/* Workaround for cache flush problems. First send a dummy EOP
 	 * event down the pipe with seq one below.
 	 */
@@ -2199,7 +2144,8 @@ static void gfx_v7_0_ring_emit_fence_gfx(struct amdgpu_ring *ring, u64 addr,
 	amdgpu_ring_write(ring, (EOP_TCL1_ACTION_EN |
 				 EOP_TC_ACTION_EN |
 				 EVENT_TYPE(CACHE_FLUSH_AND_INV_TS_EVENT) |
-				 EVENT_INDEX(5)));
+				 EVENT_INDEX(5) |
+				 (exec ? EOP_EXEC : 0)));
 	amdgpu_ring_write(ring, addr & 0xfffffffc);
 	amdgpu_ring_write(ring, (upper_32_bits(addr) & 0xffff) |
 				DATA_SEL(write64bit ? 2 : 1) | INT_SEL(int_sel ? 2 : 0));
@@ -2355,48 +2301,40 @@ static int gfx_v7_0_ring_test_ib(struct amdgpu_ring *ring, long timeout)
 	struct amdgpu_device *adev = ring->adev;
 	struct amdgpu_ib ib;
 	struct dma_fence *f = NULL;
-	uint32_t scratch;
 	uint32_t tmp = 0;
 	long r;
 
-	r = amdgpu_gfx_scratch_get(adev, &scratch);
-	if (r)
-		return r;
-
-	WREG32(scratch, 0xCAFEDEAD);
+	WREG32(mmSCRATCH_REG0, 0xCAFEDEAD);
 	memset(&ib, 0, sizeof(ib));
-	r = amdgpu_ib_get(adev, NULL, 256,
-					AMDGPU_IB_POOL_DIRECT, &ib);
+	r = amdgpu_ib_get(adev, NULL, 256, AMDGPU_IB_POOL_DIRECT, &ib);
 	if (r)
-		goto err1;
+		return r;
 
 	ib.ptr[0] = PACKET3(PACKET3_SET_UCONFIG_REG, 1);
-	ib.ptr[1] = ((scratch - PACKET3_SET_UCONFIG_REG_START));
+	ib.ptr[1] = mmSCRATCH_REG0 - PACKET3_SET_UCONFIG_REG_START;
 	ib.ptr[2] = 0xDEADBEEF;
 	ib.length_dw = 3;
 
 	r = amdgpu_ib_schedule(ring, 1, &ib, NULL, &f);
 	if (r)
-		goto err2;
+		goto error;
 
 	r = dma_fence_wait_timeout(f, false, timeout);
 	if (r == 0) {
 		r = -ETIMEDOUT;
-		goto err2;
+		goto error;
 	} else if (r < 0) {
-		goto err2;
+		goto error;
 	}
-	tmp = RREG32(scratch);
+	tmp = RREG32(mmSCRATCH_REG0);
 	if (tmp == 0xDEADBEEF)
 		r = 0;
 	else
 		r = -EINVAL;
 
-err2:
-	amdgpu_ib_free(adev, &ib, NULL);
+error:
+	amdgpu_ib_free(&ib, NULL);
 	dma_fence_put(f);
-err1:
-	amdgpu_gfx_scratch_free(adev, scratch);
 	return r;
 }
 
@@ -2631,7 +2569,7 @@ static int gfx_v7_0_cp_gfx_resume(struct amdgpu_device *adev)
 	WREG32(mmCP_RB0_WPTR, lower_32_bits(ring->wptr));
 
 	/* set the wb address whether it's enabled or not */
-	rptr_addr = adev->wb.gpu_addr + (ring->rptr_offs * 4);
+	rptr_addr = ring->rptr_gpu_addr;
 	WREG32(mmCP_RB0_RPTR_ADDR, lower_32_bits(rptr_addr));
 	WREG32(mmCP_RB0_RPTR_ADDR_HI, upper_32_bits(rptr_addr) & 0xFF);
 
@@ -2656,7 +2594,7 @@ static int gfx_v7_0_cp_gfx_resume(struct amdgpu_device *adev)
 
 static u64 gfx_v7_0_ring_get_rptr(struct amdgpu_ring *ring)
 {
-	return ring->adev->wb.wb[ring->rptr_offs];
+	return *ring->rptr_cpu_addr;
 }
 
 static u64 gfx_v7_0_ring_get_wptr_gfx(struct amdgpu_ring *ring)
@@ -2677,7 +2615,7 @@ static void gfx_v7_0_ring_set_wptr_gfx(struct amdgpu_ring *ring)
 static u64 gfx_v7_0_ring_get_wptr_compute(struct amdgpu_ring *ring)
 {
 	/* XXX check if swapping is necessary on BE */
-	return ring->adev->wb.wb[ring->wptr_offs];
+	return *ring->wptr_cpu_addr;
 }
 
 static void gfx_v7_0_ring_set_wptr_compute(struct amdgpu_ring *ring)
@@ -2685,7 +2623,7 @@ static void gfx_v7_0_ring_set_wptr_compute(struct amdgpu_ring *ring)
 	struct amdgpu_device *adev = ring->adev;
 
 	/* XXX check if swapping is necessary on BE */
-	adev->wb.wb[ring->wptr_offs] = lower_32_bits(ring->wptr);
+	*ring->wptr_cpu_addr = lower_32_bits(ring->wptr);
 	WDOORBELL32(ring->doorbell_index, lower_32_bits(ring->wptr));
 }
 
@@ -2798,7 +2736,7 @@ static int gfx_v7_0_mec_init(struct amdgpu_device *adev)
 	u32 *hpd;
 	size_t mec_hpd_size;
 
-	bitmap_zero(adev->gfx.mec.queue_bitmap, AMDGPU_MAX_COMPUTE_QUEUES);
+	bitmap_zero(adev->gfx.mec_bitmap[0].queue_bitmap, AMDGPU_MAX_COMPUTE_QUEUES);
 
 	/* take ownership of the relevant compute queues */
 	amdgpu_gfx_compute_queue_acquire(adev);
@@ -2808,7 +2746,8 @@ static int gfx_v7_0_mec_init(struct amdgpu_device *adev)
 		* GFX7_MEC_HPD_SIZE * 2;
 
 	r = amdgpu_bo_create_reserved(adev, mec_hpd_size, PAGE_SIZE,
-				      AMDGPU_GEM_DOMAIN_VRAM,
+				      AMDGPU_GEM_DOMAIN_VRAM |
+				      AMDGPU_GEM_DOMAIN_GTT,
 				      &adev->gfx.mec.hpd_eop_obj,
 				      &adev->gfx.mec.hpd_eop_gpu_addr,
 				      (void **)&hpd);
@@ -2827,45 +2766,6 @@ static int gfx_v7_0_mec_init(struct amdgpu_device *adev)
 	return 0;
 }
 
-struct hqd_registers
-{
-	u32 cp_mqd_base_addr;
-	u32 cp_mqd_base_addr_hi;
-	u32 cp_hqd_active;
-	u32 cp_hqd_vmid;
-	u32 cp_hqd_persistent_state;
-	u32 cp_hqd_pipe_priority;
-	u32 cp_hqd_queue_priority;
-	u32 cp_hqd_quantum;
-	u32 cp_hqd_pq_base;
-	u32 cp_hqd_pq_base_hi;
-	u32 cp_hqd_pq_rptr;
-	u32 cp_hqd_pq_rptr_report_addr;
-	u32 cp_hqd_pq_rptr_report_addr_hi;
-	u32 cp_hqd_pq_wptr_poll_addr;
-	u32 cp_hqd_pq_wptr_poll_addr_hi;
-	u32 cp_hqd_pq_doorbell_control;
-	u32 cp_hqd_pq_wptr;
-	u32 cp_hqd_pq_control;
-	u32 cp_hqd_ib_base_addr;
-	u32 cp_hqd_ib_base_addr_hi;
-	u32 cp_hqd_ib_rptr;
-	u32 cp_hqd_ib_control;
-	u32 cp_hqd_iq_timer;
-	u32 cp_hqd_iq_rptr;
-	u32 cp_hqd_dequeue_request;
-	u32 cp_hqd_dma_offload;
-	u32 cp_hqd_sema_cmd;
-	u32 cp_hqd_msg_type;
-	u32 cp_hqd_atomic0_preop_lo;
-	u32 cp_hqd_atomic0_preop_hi;
-	u32 cp_hqd_atomic1_preop_lo;
-	u32 cp_hqd_atomic1_preop_hi;
-	u32 cp_hqd_hq_scheduler0;
-	u32 cp_hqd_hq_scheduler1;
-	u32 cp_mqd_control;
-};
-
 static void gfx_v7_0_compute_pipe_init(struct amdgpu_device *adev,
 				       int mec, int pipe)
 {
@@ -2981,12 +2881,12 @@ static void gfx_v7_0_mqd_init(struct amdgpu_device *adev,
 		CP_HQD_PQ_CONTROL__KMD_QUEUE_MASK; /* assuming kernel queue control */
 
 	/* only used if CP_PQ_WPTR_POLL_CNTL.CP_PQ_WPTR_POLL_CNTL__EN_MASK=1 */
-	wb_gpu_addr = adev->wb.gpu_addr + (ring->wptr_offs * 4);
+	wb_gpu_addr = ring->wptr_gpu_addr;
 	mqd->cp_hqd_pq_wptr_poll_addr_lo = wb_gpu_addr & 0xfffffffc;
 	mqd->cp_hqd_pq_wptr_poll_addr_hi = upper_32_bits(wb_gpu_addr) & 0xffff;
 
 	/* set the wb address whether it's enabled or not */
-	wb_gpu_addr = adev->wb.gpu_addr + (ring->rptr_offs * 4);
+	wb_gpu_addr = ring->rptr_gpu_addr;
 	mqd->cp_hqd_pq_rptr_report_addr_lo = wb_gpu_addr & 0xfffffffc;
 	mqd->cp_hqd_pq_rptr_report_addr_hi =
 		upper_32_bits(wb_gpu_addr) & 0xffff;
@@ -3345,7 +3245,7 @@ static int gfx_v7_0_rlc_init(struct amdgpu_device *adev)
 
 	/* init spm vmid with 0xf */
 	if (adev->gfx.rlc.funcs->update_spm_vmid)
-		adev->gfx.rlc.funcs->update_spm_vmid(adev, 0xf);
+		adev->gfx.rlc.funcs->update_spm_vmid(adev, NULL, 0xf);
 
 	return 0;
 }
@@ -3370,7 +3270,7 @@ static void gfx_v7_0_wait_for_rlc_serdes(struct amdgpu_device *adev)
 	mutex_lock(&adev->grbm_idx_mutex);
 	for (i = 0; i < adev->gfx.config.max_shader_engines; i++) {
 		for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) {
-			gfx_v7_0_select_se_sh(adev, i, j, 0xffffffff);
+			gfx_v7_0_select_se_sh(adev, i, j, 0xffffffff, 0);
 			for (k = 0; k < adev->usec_timeout; k++) {
 				if (RREG32(mmRLC_SERDES_CU_MASTER_BUSY) == 0)
 					break;
@@ -3378,7 +3278,7 @@ static void gfx_v7_0_wait_for_rlc_serdes(struct amdgpu_device *adev)
 			}
 		}
 	}
-	gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+	gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
 	mutex_unlock(&adev->grbm_idx_mutex);
 
 	mask = RLC_SERDES_NONCU_MASTER_BUSY__SE_MASTER_BUSY_MASK |
@@ -3430,7 +3330,7 @@ static bool gfx_v7_0_is_rlc_enabled(struct amdgpu_device *adev)
 	return true;
 }
 
-static void gfx_v7_0_set_safe_mode(struct amdgpu_device *adev)
+static void gfx_v7_0_set_safe_mode(struct amdgpu_device *adev, int xcc_id)
 {
 	u32 tmp, i, mask;
 
@@ -3452,7 +3352,7 @@ static void gfx_v7_0_set_safe_mode(struct amdgpu_device *adev)
 	}
 }
 
-static void gfx_v7_0_unset_safe_mode(struct amdgpu_device *adev)
+static void gfx_v7_0_unset_safe_mode(struct amdgpu_device *adev, int xcc_id)
 {
 	u32 tmp;
 
@@ -3543,7 +3443,7 @@ static int gfx_v7_0_rlc_resume(struct amdgpu_device *adev)
 	WREG32(mmRLC_LB_CNTR_MAX, 0x00008000);
 
 	mutex_lock(&adev->grbm_idx_mutex);
-	gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+	gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
 	WREG32(mmRLC_LB_INIT_CU_MASK, 0xffffffff);
 	WREG32(mmRLC_LB_PARAMS, 0x00600408);
 	WREG32(mmRLC_LB_CNTL, 0x80000004);
@@ -3571,7 +3471,7 @@ static int gfx_v7_0_rlc_resume(struct amdgpu_device *adev)
 	return 0;
 }
 
-static void gfx_v7_0_update_spm_vmid(struct amdgpu_device *adev, unsigned vmid)
+static void gfx_v7_0_update_spm_vmid(struct amdgpu_device *adev, struct amdgpu_ring *ring, unsigned vmid)
 {
 	u32 data;
 
@@ -3599,7 +3499,7 @@ static void gfx_v7_0_enable_cgcg(struct amdgpu_device *adev, bool enable)
 		tmp = gfx_v7_0_halt_rlc(adev);
 
 		mutex_lock(&adev->grbm_idx_mutex);
-		gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+		gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
 		WREG32(mmRLC_SERDES_WR_CU_MASTER_MASK, 0xffffffff);
 		WREG32(mmRLC_SERDES_WR_NONCU_MASTER_MASK, 0xffffffff);
 		tmp2 = RLC_SERDES_WR_CTRL__BPM_ADDR_MASK |
@@ -3653,7 +3553,7 @@ static void gfx_v7_0_enable_mgcg(struct amdgpu_device *adev, bool enable)
 		tmp = gfx_v7_0_halt_rlc(adev);
 
 		mutex_lock(&adev->grbm_idx_mutex);
-		gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+		gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
 		WREG32(mmRLC_SERDES_WR_CU_MASTER_MASK, 0xffffffff);
 		WREG32(mmRLC_SERDES_WR_NONCU_MASTER_MASK, 0xffffffff);
 		data = RLC_SERDES_WR_CTRL__BPM_ADDR_MASK |
@@ -3704,7 +3604,7 @@ static void gfx_v7_0_enable_mgcg(struct amdgpu_device *adev, bool enable)
 		tmp = gfx_v7_0_halt_rlc(adev);
 
 		mutex_lock(&adev->grbm_idx_mutex);
-		gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+		gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
 		WREG32(mmRLC_SERDES_WR_CU_MASTER_MASK, 0xffffffff);
 		WREG32(mmRLC_SERDES_WR_NONCU_MASTER_MASK, 0xffffffff);
 		data = RLC_SERDES_WR_CTRL__BPM_ADDR_MASK | RLC_SERDES_WR_CTRL__MGCG_OVERRIDE_1_MASK;
@@ -3982,70 +3882,24 @@ static u32 gfx_v7_0_get_csb_size(struct amdgpu_device *adev)
 	return count;
 }
 
-static void gfx_v7_0_get_csb_buffer(struct amdgpu_device *adev,
-				    volatile u32 *buffer)
+static void gfx_v7_0_get_csb_buffer(struct amdgpu_device *adev, u32 *buffer)
 {
-	u32 count = 0, i;
-	const struct cs_section_def *sect = NULL;
-	const struct cs_extent_def *ext = NULL;
+	u32 count = 0;
 
 	if (adev->gfx.rlc.cs_data == NULL)
 		return;
 	if (buffer == NULL)
 		return;
 
-	buffer[count++] = cpu_to_le32(PACKET3(PACKET3_PREAMBLE_CNTL, 0));
-	buffer[count++] = cpu_to_le32(PACKET3_PREAMBLE_BEGIN_CLEAR_STATE);
-
-	buffer[count++] = cpu_to_le32(PACKET3(PACKET3_CONTEXT_CONTROL, 1));
-	buffer[count++] = cpu_to_le32(0x80000000);
-	buffer[count++] = cpu_to_le32(0x80000000);
-
-	for (sect = adev->gfx.rlc.cs_data; sect->section != NULL; ++sect) {
-		for (ext = sect->section; ext->extent != NULL; ++ext) {
-			if (sect->id == SECT_CONTEXT) {
-				buffer[count++] =
-					cpu_to_le32(PACKET3(PACKET3_SET_CONTEXT_REG, ext->reg_count));
-				buffer[count++] = cpu_to_le32(ext->reg_index - PACKET3_SET_CONTEXT_REG_START);
-				for (i = 0; i < ext->reg_count; i++)
-					buffer[count++] = cpu_to_le32(ext->extent[i]);
-			} else {
-				return;
-			}
-		}
-	}
+	count = amdgpu_gfx_csb_preamble_start(buffer);
+	count = amdgpu_gfx_csb_data_parser(adev, buffer, count);
 
 	buffer[count++] = cpu_to_le32(PACKET3(PACKET3_SET_CONTEXT_REG, 2));
 	buffer[count++] = cpu_to_le32(mmPA_SC_RASTER_CONFIG - PACKET3_SET_CONTEXT_REG_START);
-	switch (adev->asic_type) {
-	case CHIP_BONAIRE:
-		buffer[count++] = cpu_to_le32(0x16000012);
-		buffer[count++] = cpu_to_le32(0x00000000);
-		break;
-	case CHIP_KAVERI:
-		buffer[count++] = cpu_to_le32(0x00000000); /* XXX */
-		buffer[count++] = cpu_to_le32(0x00000000);
-		break;
-	case CHIP_KABINI:
-	case CHIP_MULLINS:
-		buffer[count++] = cpu_to_le32(0x00000000); /* XXX */
-		buffer[count++] = cpu_to_le32(0x00000000);
-		break;
-	case CHIP_HAWAII:
-		buffer[count++] = cpu_to_le32(0x3a00161a);
-		buffer[count++] = cpu_to_le32(0x0000002e);
-		break;
-	default:
-		buffer[count++] = cpu_to_le32(0x00000000);
-		buffer[count++] = cpu_to_le32(0x00000000);
-		break;
-	}
-
-	buffer[count++] = cpu_to_le32(PACKET3(PACKET3_PREAMBLE_CNTL, 0));
-	buffer[count++] = cpu_to_le32(PACKET3_PREAMBLE_END_CLEAR_STATE);
+	buffer[count++] = cpu_to_le32(adev->gfx.config.rb_config[0][0].raster_config);
+	buffer[count++] = cpu_to_le32(adev->gfx.config.rb_config[0][0].raster_config_1);
 
-	buffer[count++] = cpu_to_le32(PACKET3(PACKET3_CLEAR_STATE, 0));
-	buffer[count++] = cpu_to_le32(0);
+	amdgpu_gfx_csb_preamble_end(buffer, count);
 }
 
 static void gfx_v7_0_init_pg(struct amdgpu_device *adev)
@@ -4180,7 +4034,7 @@ static void wave_read_regs(struct amdgpu_device *adev, uint32_t simd,
 		*(out++) = RREG32(mmSQ_IND_DATA);
 }
 
-static void gfx_v7_0_read_wave_data(struct amdgpu_device *adev, uint32_t simd, uint32_t wave, uint32_t *dst, int *no_fields)
+static void gfx_v7_0_read_wave_data(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd, uint32_t wave, uint32_t *dst, int *no_fields)
 {
 	/* type 0 wave data */
 	dst[(*no_fields)++] = 0;
@@ -4205,7 +4059,7 @@ static void gfx_v7_0_read_wave_data(struct amdgpu_device *adev, uint32_t simd, u
 	dst[(*no_fields)++] = wave_read_ind(adev, simd, wave, ixSQ_WAVE_MODE);
 }
 
-static void gfx_v7_0_read_wave_sgprs(struct amdgpu_device *adev, uint32_t simd,
+static void gfx_v7_0_read_wave_sgprs(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd,
 				     uint32_t wave, uint32_t start,
 				     uint32_t size, uint32_t *dst)
 {
@@ -4215,7 +4069,7 @@ static void gfx_v7_0_read_wave_sgprs(struct amdgpu_device *adev, uint32_t simd,
 }
 
 static void gfx_v7_0_select_me_pipe_q(struct amdgpu_device *adev,
-				  u32 me, u32 pipe, u32 q, u32 vm)
+				  u32 me, u32 pipe, u32 q, u32 vm, u32 xcc_id)
 {
 	cik_srbm_select(adev, me, pipe, q, vm);
 }
@@ -4243,10 +4097,11 @@ static const struct amdgpu_rlc_funcs gfx_v7_0_rlc_funcs = {
 	.update_spm_vmid = gfx_v7_0_update_spm_vmid
 };
 
-static int gfx_v7_0_early_init(void *handle)
+static int gfx_v7_0_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
+	adev->gfx.xcc_mask = 1;
 	adev->gfx.num_gfx_rings = GFX7_NUM_GFX_RINGS;
 	adev->gfx.num_compute_rings = min(amdgpu_gfx_get_num_kcq(adev),
 					  AMDGPU_MAX_COMPUTE_RINGS);
@@ -4259,9 +4114,9 @@ static int gfx_v7_0_early_init(void *handle)
 	return 0;
 }
 
-static int gfx_v7_0_late_init(void *handle)
+static int gfx_v7_0_late_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int r;
 
 	r = amdgpu_irq_get(adev, &adev->gfx.priv_reg_irq, 0);
@@ -4451,10 +4306,10 @@ static int gfx_v7_0_compute_ring_init(struct amdgpu_device *adev, int ring_id,
 	return 0;
 }
 
-static int gfx_v7_0_sw_init(void *handle)
+static int gfx_v7_0_sw_init(struct amdgpu_ip_block *ip_block)
 {
 	struct amdgpu_ring *ring;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int i, j, k, r, ring_id;
 
 	switch (adev->asic_type) {
@@ -4489,8 +4344,6 @@ static int gfx_v7_0_sw_init(void *handle)
 	if (r)
 		return r;
 
-	gfx_v7_0_scratch_init(adev);
-
 	r = gfx_v7_0_init_microcode(adev);
 	if (r) {
 		DRM_ERROR("Failed to load gfx firmware!\n");
@@ -4527,7 +4380,8 @@ static int gfx_v7_0_sw_init(void *handle)
 	for (i = 0; i < adev->gfx.mec.num_mec; ++i) {
 		for (j = 0; j < adev->gfx.mec.num_queue_per_pipe; j++) {
 			for (k = 0; k < adev->gfx.mec.num_pipe_per_mec; k++) {
-				if (!amdgpu_gfx_is_mec_queue_enabled(adev, i, k, j))
+				if (!amdgpu_gfx_is_mec_queue_enabled(adev, 0, i,
+								     k, j))
 					continue;
 
 				r = gfx_v7_0_compute_ring_init(adev,
@@ -4545,12 +4399,17 @@ static int gfx_v7_0_sw_init(void *handle)
 
 	gfx_v7_0_gpu_early_init(adev);
 
+	adev->gfx.gfx_supported_reset =
+		amdgpu_get_soft_full_reset_mask(&adev->gfx.gfx_ring[0]);
+	adev->gfx.compute_supported_reset =
+		amdgpu_get_soft_full_reset_mask(&adev->gfx.compute_ring[0]);
+
 	return r;
 }
 
-static int gfx_v7_0_sw_fini(void *handle)
+static int gfx_v7_0_sw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int i;
 
 	for (i = 0; i < adev->gfx.num_gfx_rings; i++)
@@ -4574,10 +4433,10 @@ static int gfx_v7_0_sw_fini(void *handle)
 	return 0;
 }
 
-static int gfx_v7_0_hw_init(void *handle)
+static int gfx_v7_0_hw_init(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	gfx_v7_0_constants_init(adev);
 
@@ -4595,9 +4454,9 @@ static int gfx_v7_0_hw_init(void *handle)
 	return r;
 }
 
-static int gfx_v7_0_hw_fini(void *handle)
+static int gfx_v7_0_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	amdgpu_irq_put(adev, &adev->gfx.priv_reg_irq, 0);
 	amdgpu_irq_put(adev, &adev->gfx.priv_inst_irq, 0);
@@ -4608,23 +4467,19 @@ static int gfx_v7_0_hw_fini(void *handle)
 	return 0;
 }
 
-static int gfx_v7_0_suspend(void *handle)
+static int gfx_v7_0_suspend(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	return gfx_v7_0_hw_fini(adev);
+	return gfx_v7_0_hw_fini(ip_block);
 }
 
-static int gfx_v7_0_resume(void *handle)
+static int gfx_v7_0_resume(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	return gfx_v7_0_hw_init(adev);
+	return gfx_v7_0_hw_init(ip_block);
 }
 
-static bool gfx_v7_0_is_idle(void *handle)
+static bool gfx_v7_0_is_idle(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (RREG32(mmGRBM_STATUS) & GRBM_STATUS__GUI_ACTIVE_MASK)
 		return false;
@@ -4632,11 +4487,11 @@ static bool gfx_v7_0_is_idle(void *handle)
 		return true;
 }
 
-static int gfx_v7_0_wait_for_idle(void *handle)
+static int gfx_v7_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
 {
 	unsigned i;
 	u32 tmp;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	for (i = 0; i < adev->usec_timeout; i++) {
 		/* read MC_STATUS */
@@ -4649,11 +4504,11 @@ static int gfx_v7_0_wait_for_idle(void *handle)
 	return -ETIMEDOUT;
 }
 
-static int gfx_v7_0_soft_reset(void *handle)
+static int gfx_v7_0_soft_reset(struct amdgpu_ip_block *ip_block)
 {
 	u32 grbm_soft_reset = 0, srbm_soft_reset = 0;
 	u32 tmp;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	/* GRBM_STATUS */
 	tmp = RREG32(mmGRBM_STATUS);
@@ -4959,11 +4814,11 @@ static int gfx_v7_0_priv_inst_irq(struct amdgpu_device *adev,
 	return 0;
 }
 
-static int gfx_v7_0_set_clockgating_state(void *handle,
+static int gfx_v7_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_clockgating_state state)
 {
 	bool gate = false;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (state == AMD_CG_STATE_GATE)
 		gate = true;
@@ -4982,11 +4837,11 @@ static int gfx_v7_0_set_clockgating_state(void *handle,
 	return 0;
 }
 
-static int gfx_v7_0_set_powergating_state(void *handle,
+static int gfx_v7_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_powergating_state state)
 {
 	bool gate = false;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (state == AMD_PG_STATE_GATE)
 		gate = true;
@@ -5112,6 +4967,7 @@ static const struct amdgpu_ring_funcs gfx_v7_0_ring_funcs_compute = {
 	.insert_nop = amdgpu_ring_insert_nop,
 	.pad_ib = amdgpu_ring_generic_pad_ib,
 	.emit_wreg = gfx_v7_0_ring_emit_wreg,
+	.soft_recovery = gfx_v7_0_ring_soft_recovery,
 	.emit_mem_sync = gfx_v7_0_emit_mem_sync_compute,
 };
 
@@ -5185,18 +5041,18 @@ static void gfx_v7_0_get_cu_info(struct amdgpu_device *adev)
 			mask = 1;
 			ao_bitmap = 0;
 			counter = 0;
-			gfx_v7_0_select_se_sh(adev, i, j, 0xffffffff);
+			gfx_v7_0_select_se_sh(adev, i, j, 0xffffffff, 0);
 			if (i < 4 && j < 2)
 				gfx_v7_0_set_user_cu_inactive_bitmap(
 					adev, disable_masks[i * 2 + j]);
 			bitmap = gfx_v7_0_get_cu_active_bitmap(adev);
-			cu_info->bitmap[i][j] = bitmap;
+			cu_info->bitmap[0][i][j] = bitmap;
 
-			for (k = 0; k < adev->gfx.config.max_cu_per_sh; k ++) {
+			for (k = 0; k < adev->gfx.config.max_cu_per_sh; k++) {
 				if (bitmap & mask) {
 					if (counter < ao_cu_num)
 						ao_bitmap |= mask;
-					counter ++;
+					counter++;
 				}
 				mask <<= 1;
 			}
@@ -5206,7 +5062,7 @@ static void gfx_v7_0_get_cu_info(struct amdgpu_device *adev)
 			cu_info->ao_cu_bitmap[i][j] = ao_bitmap;
 		}
 	}
-	gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+	gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
 	mutex_unlock(&adev->grbm_idx_mutex);
 
 	cu_info->number = active_cu_number;
@@ -5218,8 +5074,7 @@ static void gfx_v7_0_get_cu_info(struct amdgpu_device *adev)
 	cu_info->lds_size = 64;
 }
 
-const struct amdgpu_ip_block_version gfx_v7_1_ip_block =
-{
+const struct amdgpu_ip_block_version gfx_v7_1_ip_block = {
 	.type = AMD_IP_BLOCK_TYPE_GFX,
 	.major = 7,
 	.minor = 1,
@@ -5227,8 +5082,7 @@ const struct amdgpu_ip_block_version gfx_v7_1_ip_block =
 	.funcs = &gfx_v7_0_ip_funcs,
 };
 
-const struct amdgpu_ip_block_version gfx_v7_2_ip_block =
-{
+const struct amdgpu_ip_block_version gfx_v7_2_ip_block = {
 	.type = AMD_IP_BLOCK_TYPE_GFX,
 	.major = 7,
 	.minor = 2,
@@ -5236,8 +5090,7 @@ const struct amdgpu_ip_block_version gfx_v7_2_ip_block =
 	.funcs = &gfx_v7_0_ip_funcs,
 };
 
-const struct amdgpu_ip_block_version gfx_v7_3_ip_block =
-{
+const struct amdgpu_ip_block_version gfx_v7_3_ip_block = {
 	.type = AMD_IP_BLOCK_TYPE_GFX,
 	.major = 7,
 	.minor = 3,
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
index 5f112efda634..1c87375e1dd5 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
@@ -835,37 +835,25 @@ static void gfx_v8_0_init_golden_registers(struct amdgpu_device *adev)
 	}
 }
 
-static void gfx_v8_0_scratch_init(struct amdgpu_device *adev)
-{
-	adev->gfx.scratch.num_reg = 8;
-	adev->gfx.scratch.reg_base = mmSCRATCH_REG0;
-	adev->gfx.scratch.free_mask = (1u << adev->gfx.scratch.num_reg) - 1;
-}
-
 static int gfx_v8_0_ring_test_ring(struct amdgpu_ring *ring)
 {
 	struct amdgpu_device *adev = ring->adev;
-	uint32_t scratch;
 	uint32_t tmp = 0;
 	unsigned i;
 	int r;
 
-	r = amdgpu_gfx_scratch_get(adev, &scratch);
-	if (r)
-		return r;
-
-	WREG32(scratch, 0xCAFEDEAD);
+	WREG32(mmSCRATCH_REG0, 0xCAFEDEAD);
 	r = amdgpu_ring_alloc(ring, 3);
 	if (r)
-		goto error_free_scratch;
+		return r;
 
 	amdgpu_ring_write(ring, PACKET3(PACKET3_SET_UCONFIG_REG, 1));
-	amdgpu_ring_write(ring, (scratch - PACKET3_SET_UCONFIG_REG_START));
+	amdgpu_ring_write(ring, mmSCRATCH_REG0 - PACKET3_SET_UCONFIG_REG_START);
 	amdgpu_ring_write(ring, 0xDEADBEEF);
 	amdgpu_ring_commit(ring);
 
 	for (i = 0; i < adev->usec_timeout; i++) {
-		tmp = RREG32(scratch);
+		tmp = RREG32(mmSCRATCH_REG0);
 		if (tmp == 0xDEADBEEF)
 			break;
 		udelay(1);
@@ -874,8 +862,6 @@ static int gfx_v8_0_ring_test_ring(struct amdgpu_ring *ring)
 	if (i >= adev->usec_timeout)
 		r = -ETIMEDOUT;
 
-error_free_scratch:
-	amdgpu_gfx_scratch_free(adev, scratch);
 	return r;
 }
 
@@ -897,8 +883,8 @@ static int gfx_v8_0_ring_test_ib(struct amdgpu_ring *ring, long timeout)
 	gpu_addr = adev->wb.gpu_addr + (index * 4);
 	adev->wb.wb[index] = cpu_to_le32(0xCAFEDEAD);
 	memset(&ib, 0, sizeof(ib));
-	r = amdgpu_ib_get(adev, NULL, 16,
-					AMDGPU_IB_POOL_DIRECT, &ib);
+
+	r = amdgpu_ib_get(adev, NULL, 20, AMDGPU_IB_POOL_DIRECT, &ib);
 	if (r)
 		goto err1;
 
@@ -928,7 +914,7 @@ static int gfx_v8_0_ring_test_ib(struct amdgpu_ring *ring, long timeout)
 		r = -EINVAL;
 
 err2:
-	amdgpu_ib_free(adev, &ib, NULL);
+	amdgpu_ib_free(&ib, NULL);
 	dma_fence_put(f);
 err1:
 	amdgpu_device_wb_free(adev, index);
@@ -938,20 +924,14 @@ err1:
 
 static void gfx_v8_0_free_microcode(struct amdgpu_device *adev)
 {
-	release_firmware(adev->gfx.pfp_fw);
-	adev->gfx.pfp_fw = NULL;
-	release_firmware(adev->gfx.me_fw);
-	adev->gfx.me_fw = NULL;
-	release_firmware(adev->gfx.ce_fw);
-	adev->gfx.ce_fw = NULL;
-	release_firmware(adev->gfx.rlc_fw);
-	adev->gfx.rlc_fw = NULL;
-	release_firmware(adev->gfx.mec_fw);
-	adev->gfx.mec_fw = NULL;
+	amdgpu_ucode_release(&adev->gfx.pfp_fw);
+	amdgpu_ucode_release(&adev->gfx.me_fw);
+	amdgpu_ucode_release(&adev->gfx.ce_fw);
+	amdgpu_ucode_release(&adev->gfx.rlc_fw);
+	amdgpu_ucode_release(&adev->gfx.mec_fw);
 	if ((adev->asic_type != CHIP_STONEY) &&
 	    (adev->asic_type != CHIP_TOPAZ))
-		release_firmware(adev->gfx.mec2_fw);
-	adev->gfx.mec2_fw = NULL;
+		amdgpu_ucode_release(&adev->gfx.mec2_fw);
 
 	kfree(adev->gfx.rlc.register_list_format);
 }
@@ -959,7 +939,6 @@ static void gfx_v8_0_free_microcode(struct amdgpu_device *adev)
 static int gfx_v8_0_init_microcode(struct amdgpu_device *adev)
 {
 	const char *chip_name;
-	char fw_name[30];
 	int err;
 	struct amdgpu_firmware_info *info = NULL;
 	const struct common_firmware_header *header = NULL;
@@ -1002,62 +981,62 @@ static int gfx_v8_0_init_microcode(struct amdgpu_device *adev)
 	}
 
 	if (adev->asic_type >= CHIP_POLARIS10 && adev->asic_type <= CHIP_POLARIS12) {
-		snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_pfp_2.bin", chip_name);
-		err = request_firmware(&adev->gfx.pfp_fw, fw_name, adev->dev);
-		if (err == -ENOENT) {
-			snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_pfp.bin", chip_name);
-			err = request_firmware(&adev->gfx.pfp_fw, fw_name, adev->dev);
+		err = amdgpu_ucode_request(adev, &adev->gfx.pfp_fw,
+					   AMDGPU_UCODE_OPTIONAL,
+					   "amdgpu/%s_pfp_2.bin", chip_name);
+		if (err == -ENODEV) {
+			err = amdgpu_ucode_request(adev, &adev->gfx.pfp_fw,
+						   AMDGPU_UCODE_REQUIRED,
+						   "amdgpu/%s_pfp.bin", chip_name);
 		}
 	} else {
-		snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_pfp.bin", chip_name);
-		err = request_firmware(&adev->gfx.pfp_fw, fw_name, adev->dev);
+		err = amdgpu_ucode_request(adev, &adev->gfx.pfp_fw,
+					   AMDGPU_UCODE_REQUIRED,
+					   "amdgpu/%s_pfp.bin", chip_name);
 	}
 	if (err)
 		goto out;
-	err = amdgpu_ucode_validate(adev->gfx.pfp_fw);
-	if (err)
-		goto out;
 	cp_hdr = (const struct gfx_firmware_header_v1_0 *)adev->gfx.pfp_fw->data;
 	adev->gfx.pfp_fw_version = le32_to_cpu(cp_hdr->header.ucode_version);
 	adev->gfx.pfp_feature_version = le32_to_cpu(cp_hdr->ucode_feature_version);
 
 	if (adev->asic_type >= CHIP_POLARIS10 && adev->asic_type <= CHIP_POLARIS12) {
-		snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_me_2.bin", chip_name);
-		err = request_firmware(&adev->gfx.me_fw, fw_name, adev->dev);
-		if (err == -ENOENT) {
-			snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_me.bin", chip_name);
-			err = request_firmware(&adev->gfx.me_fw, fw_name, adev->dev);
+		err = amdgpu_ucode_request(adev, &adev->gfx.me_fw,
+					   AMDGPU_UCODE_OPTIONAL,
+					   "amdgpu/%s_me_2.bin", chip_name);
+		if (err == -ENODEV) {
+			err = amdgpu_ucode_request(adev, &adev->gfx.me_fw,
+						   AMDGPU_UCODE_REQUIRED,
+						   "amdgpu/%s_me.bin", chip_name);
 		}
 	} else {
-		snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_me.bin", chip_name);
-		err = request_firmware(&adev->gfx.me_fw, fw_name, adev->dev);
+		err = amdgpu_ucode_request(adev, &adev->gfx.me_fw,
+					   AMDGPU_UCODE_REQUIRED,
+					   "amdgpu/%s_me.bin", chip_name);
 	}
 	if (err)
 		goto out;
-	err = amdgpu_ucode_validate(adev->gfx.me_fw);
-	if (err)
-		goto out;
 	cp_hdr = (const struct gfx_firmware_header_v1_0 *)adev->gfx.me_fw->data;
 	adev->gfx.me_fw_version = le32_to_cpu(cp_hdr->header.ucode_version);
 
 	adev->gfx.me_feature_version = le32_to_cpu(cp_hdr->ucode_feature_version);
 
 	if (adev->asic_type >= CHIP_POLARIS10 && adev->asic_type <= CHIP_POLARIS12) {
-		snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_ce_2.bin", chip_name);
-		err = request_firmware(&adev->gfx.ce_fw, fw_name, adev->dev);
-		if (err == -ENOENT) {
-			snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_ce.bin", chip_name);
-			err = request_firmware(&adev->gfx.ce_fw, fw_name, adev->dev);
+		err = amdgpu_ucode_request(adev, &adev->gfx.ce_fw,
+					   AMDGPU_UCODE_OPTIONAL,
+					   "amdgpu/%s_ce_2.bin", chip_name);
+		if (err == -ENODEV) {
+			err = amdgpu_ucode_request(adev, &adev->gfx.ce_fw,
+						   AMDGPU_UCODE_REQUIRED,
+						   "amdgpu/%s_ce.bin", chip_name);
 		}
 	} else {
-		snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_ce.bin", chip_name);
-		err = request_firmware(&adev->gfx.ce_fw, fw_name, adev->dev);
+		err = amdgpu_ucode_request(adev, &adev->gfx.ce_fw,
+					   AMDGPU_UCODE_REQUIRED,
+					   "amdgpu/%s_ce.bin", chip_name);
 	}
 	if (err)
 		goto out;
-	err = amdgpu_ucode_validate(adev->gfx.ce_fw);
-	if (err)
-		goto out;
 	cp_hdr = (const struct gfx_firmware_header_v1_0 *)adev->gfx.ce_fw->data;
 	adev->gfx.ce_fw_version = le32_to_cpu(cp_hdr->header.ucode_version);
 	adev->gfx.ce_feature_version = le32_to_cpu(cp_hdr->ucode_feature_version);
@@ -1073,11 +1052,11 @@ static int gfx_v8_0_init_microcode(struct amdgpu_device *adev)
 	} else
 		adev->virt.chained_ib_support = false;
 
-	snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_rlc.bin", chip_name);
-	err = request_firmware(&adev->gfx.rlc_fw, fw_name, adev->dev);
+	err = amdgpu_ucode_request(adev, &adev->gfx.rlc_fw,
+				   AMDGPU_UCODE_REQUIRED,
+				   "amdgpu/%s_rlc.bin", chip_name);
 	if (err)
 		goto out;
-	err = amdgpu_ucode_validate(adev->gfx.rlc_fw);
 	rlc_hdr = (const struct rlc_firmware_header_v2_0 *)adev->gfx.rlc_fw->data;
 	adev->gfx.rlc_fw_version = le32_to_cpu(rlc_hdr->header.ucode_version);
 	adev->gfx.rlc_feature_version = le32_to_cpu(rlc_hdr->ucode_feature_version);
@@ -1123,21 +1102,21 @@ static int gfx_v8_0_init_microcode(struct amdgpu_device *adev)
 		adev->gfx.rlc.register_restore[i] = le32_to_cpu(tmp[i]);
 
 	if (adev->asic_type >= CHIP_POLARIS10 && adev->asic_type <= CHIP_POLARIS12) {
-		snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_mec_2.bin", chip_name);
-		err = request_firmware(&adev->gfx.mec_fw, fw_name, adev->dev);
-		if (err == -ENOENT) {
-			snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_mec.bin", chip_name);
-			err = request_firmware(&adev->gfx.mec_fw, fw_name, adev->dev);
+		err = amdgpu_ucode_request(adev, &adev->gfx.mec_fw,
+					   AMDGPU_UCODE_OPTIONAL,
+					   "amdgpu/%s_mec_2.bin", chip_name);
+		if (err == -ENODEV) {
+			err = amdgpu_ucode_request(adev, &adev->gfx.mec_fw,
+						   AMDGPU_UCODE_REQUIRED,
+						   "amdgpu/%s_mec.bin", chip_name);
 		}
 	} else {
-		snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_mec.bin", chip_name);
-		err = request_firmware(&adev->gfx.mec_fw, fw_name, adev->dev);
+		err = amdgpu_ucode_request(adev, &adev->gfx.mec_fw,
+					   AMDGPU_UCODE_REQUIRED,
+					   "amdgpu/%s_mec.bin", chip_name);
 	}
 	if (err)
 		goto out;
-	err = amdgpu_ucode_validate(adev->gfx.mec_fw);
-	if (err)
-		goto out;
 	cp_hdr = (const struct gfx_firmware_header_v1_0 *)adev->gfx.mec_fw->data;
 	adev->gfx.mec_fw_version = le32_to_cpu(cp_hdr->header.ucode_version);
 	adev->gfx.mec_feature_version = le32_to_cpu(cp_hdr->ucode_feature_version);
@@ -1145,20 +1124,20 @@ static int gfx_v8_0_init_microcode(struct amdgpu_device *adev)
 	if ((adev->asic_type != CHIP_STONEY) &&
 	    (adev->asic_type != CHIP_TOPAZ)) {
 		if (adev->asic_type >= CHIP_POLARIS10 && adev->asic_type <= CHIP_POLARIS12) {
-			snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_mec2_2.bin", chip_name);
-			err = request_firmware(&adev->gfx.mec2_fw, fw_name, adev->dev);
-			if (err == -ENOENT) {
-				snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_mec2.bin", chip_name);
-				err = request_firmware(&adev->gfx.mec2_fw, fw_name, adev->dev);
+			err = amdgpu_ucode_request(adev, &adev->gfx.mec2_fw,
+						   AMDGPU_UCODE_OPTIONAL,
+						   "amdgpu/%s_mec2_2.bin", chip_name);
+			if (err == -ENODEV) {
+				err = amdgpu_ucode_request(adev, &adev->gfx.mec2_fw,
+							   AMDGPU_UCODE_REQUIRED,
+							   "amdgpu/%s_mec2.bin", chip_name);
 			}
 		} else {
-			snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_mec2.bin", chip_name);
-			err = request_firmware(&adev->gfx.mec2_fw, fw_name, adev->dev);
+			err = amdgpu_ucode_request(adev, &adev->gfx.mec2_fw,
+						   AMDGPU_UCODE_REQUIRED,
+						   "amdgpu/%s_mec2.bin", chip_name);
 		}
 		if (!err) {
-			err = amdgpu_ucode_validate(adev->gfx.mec2_fw);
-			if (err)
-				goto out;
 			cp_hdr = (const struct gfx_firmware_header_v1_0 *)
 				adev->gfx.mec2_fw->data;
 			adev->gfx.mec2_fw_version =
@@ -1230,70 +1209,35 @@ static int gfx_v8_0_init_microcode(struct amdgpu_device *adev)
 
 out:
 	if (err) {
-		dev_err(adev->dev,
-			"gfx8: Failed to load firmware \"%s\"\n",
-			fw_name);
-		release_firmware(adev->gfx.pfp_fw);
-		adev->gfx.pfp_fw = NULL;
-		release_firmware(adev->gfx.me_fw);
-		adev->gfx.me_fw = NULL;
-		release_firmware(adev->gfx.ce_fw);
-		adev->gfx.ce_fw = NULL;
-		release_firmware(adev->gfx.rlc_fw);
-		adev->gfx.rlc_fw = NULL;
-		release_firmware(adev->gfx.mec_fw);
-		adev->gfx.mec_fw = NULL;
-		release_firmware(adev->gfx.mec2_fw);
-		adev->gfx.mec2_fw = NULL;
+		dev_err(adev->dev, "gfx8: Failed to load firmware %s gfx firmware\n", chip_name);
+		amdgpu_ucode_release(&adev->gfx.pfp_fw);
+		amdgpu_ucode_release(&adev->gfx.me_fw);
+		amdgpu_ucode_release(&adev->gfx.ce_fw);
+		amdgpu_ucode_release(&adev->gfx.rlc_fw);
+		amdgpu_ucode_release(&adev->gfx.mec_fw);
+		amdgpu_ucode_release(&adev->gfx.mec2_fw);
 	}
 	return err;
 }
 
-static void gfx_v8_0_get_csb_buffer(struct amdgpu_device *adev,
-				    volatile u32 *buffer)
+static void gfx_v8_0_get_csb_buffer(struct amdgpu_device *adev, u32 *buffer)
 {
-	u32 count = 0, i;
-	const struct cs_section_def *sect = NULL;
-	const struct cs_extent_def *ext = NULL;
+	u32 count = 0;
 
 	if (adev->gfx.rlc.cs_data == NULL)
 		return;
 	if (buffer == NULL)
 		return;
 
-	buffer[count++] = cpu_to_le32(PACKET3(PACKET3_PREAMBLE_CNTL, 0));
-	buffer[count++] = cpu_to_le32(PACKET3_PREAMBLE_BEGIN_CLEAR_STATE);
-
-	buffer[count++] = cpu_to_le32(PACKET3(PACKET3_CONTEXT_CONTROL, 1));
-	buffer[count++] = cpu_to_le32(0x80000000);
-	buffer[count++] = cpu_to_le32(0x80000000);
-
-	for (sect = adev->gfx.rlc.cs_data; sect->section != NULL; ++sect) {
-		for (ext = sect->section; ext->extent != NULL; ++ext) {
-			if (sect->id == SECT_CONTEXT) {
-				buffer[count++] =
-					cpu_to_le32(PACKET3(PACKET3_SET_CONTEXT_REG, ext->reg_count));
-				buffer[count++] = cpu_to_le32(ext->reg_index -
-						PACKET3_SET_CONTEXT_REG_START);
-				for (i = 0; i < ext->reg_count; i++)
-					buffer[count++] = cpu_to_le32(ext->extent[i]);
-			} else {
-				return;
-			}
-		}
-	}
+	count = amdgpu_gfx_csb_preamble_start(buffer);
+	count = amdgpu_gfx_csb_data_parser(adev, buffer, count);
 
 	buffer[count++] = cpu_to_le32(PACKET3(PACKET3_SET_CONTEXT_REG, 2));
-	buffer[count++] = cpu_to_le32(mmPA_SC_RASTER_CONFIG -
-			PACKET3_SET_CONTEXT_REG_START);
+	buffer[count++] = cpu_to_le32(mmPA_SC_RASTER_CONFIG - PACKET3_SET_CONTEXT_REG_START);
 	buffer[count++] = cpu_to_le32(adev->gfx.config.rb_config[0][0].raster_config);
 	buffer[count++] = cpu_to_le32(adev->gfx.config.rb_config[0][0].raster_config_1);
 
-	buffer[count++] = cpu_to_le32(PACKET3(PACKET3_PREAMBLE_CNTL, 0));
-	buffer[count++] = cpu_to_le32(PACKET3_PREAMBLE_END_CLEAR_STATE);
-
-	buffer[count++] = cpu_to_le32(PACKET3(PACKET3_CLEAR_STATE, 0));
-	buffer[count++] = cpu_to_le32(0);
+	amdgpu_gfx_csb_preamble_end(buffer, count);
 }
 
 static int gfx_v8_0_cp_jump_table_num(struct amdgpu_device *adev)
@@ -1330,7 +1274,7 @@ static int gfx_v8_0_rlc_init(struct amdgpu_device *adev)
 
 	/* init spm vmid with 0xf */
 	if (adev->gfx.rlc.funcs->update_spm_vmid)
-		adev->gfx.rlc.funcs->update_spm_vmid(adev, 0xf);
+		adev->gfx.rlc.funcs->update_spm_vmid(adev, NULL, 0xf);
 
 	return 0;
 }
@@ -1346,7 +1290,7 @@ static int gfx_v8_0_mec_init(struct amdgpu_device *adev)
 	u32 *hpd;
 	size_t mec_hpd_size;
 
-	bitmap_zero(adev->gfx.mec.queue_bitmap, AMDGPU_MAX_COMPUTE_QUEUES);
+	bitmap_zero(adev->gfx.mec_bitmap[0].queue_bitmap, AMDGPU_MAX_COMPUTE_QUEUES);
 
 	/* take ownership of the relevant compute queues */
 	amdgpu_gfx_compute_queue_acquire(adev);
@@ -1354,7 +1298,8 @@ static int gfx_v8_0_mec_init(struct amdgpu_device *adev)
 	mec_hpd_size = adev->gfx.num_compute_rings * GFX8_MEC_HPD_SIZE;
 	if (mec_hpd_size) {
 		r = amdgpu_bo_create_reserved(adev, mec_hpd_size, PAGE_SIZE,
-					      AMDGPU_GEM_DOMAIN_VRAM,
+					      AMDGPU_GEM_DOMAIN_VRAM |
+					      AMDGPU_GEM_DOMAIN_GTT,
 					      &adev->gfx.mec.hpd_eop_obj,
 					      &adev->gfx.mec.hpd_eop_gpu_addr,
 					      (void **)&hpd);
@@ -1684,7 +1629,7 @@ static int gfx_v8_0_do_edc_gpr_workarounds(struct amdgpu_device *adev)
 		RREG32(sec_ded_counter_registers[i]);
 
 fail:
-	amdgpu_ib_free(adev, &ib, NULL);
+	amdgpu_ib_free(&ib, NULL);
 	dma_fence_put(f);
 
 	return r;
@@ -1925,7 +1870,7 @@ static int gfx_v8_0_compute_ring_init(struct amdgpu_device *adev, int ring_id,
 		+ ring->pipe;
 
 	hw_prio = amdgpu_gfx_is_high_priority_compute_queue(adev, ring) ?
-			AMDGPU_GFX_PIPE_PRIO_HIGH : AMDGPU_RING_PRIO_DEFAULT;
+			AMDGPU_RING_PRIO_2 : AMDGPU_RING_PRIO_DEFAULT;
 	/* type-2 packets are deprecated on MEC, use type-3 instead */
 	r = amdgpu_ring_init(adev, ring, 1024, &adev->gfx.eop_irq, irq_type,
 			     hw_prio, NULL);
@@ -1938,12 +1883,12 @@ static int gfx_v8_0_compute_ring_init(struct amdgpu_device *adev, int ring_id,
 
 static void gfx_v8_0_sq_irq_work_func(struct work_struct *work);
 
-static int gfx_v8_0_sw_init(void *handle)
+static int gfx_v8_0_sw_init(struct amdgpu_ip_block *ip_block)
 {
 	int i, j, k, r, ring_id;
+	int xcc_id = 0;
 	struct amdgpu_ring *ring;
-	struct amdgpu_kiq *kiq;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	switch (adev->asic_type) {
 	case CHIP_TONGA:
@@ -2000,8 +1945,6 @@ static int gfx_v8_0_sw_init(void *handle)
 
 	adev->gfx.gfx_current_status = AMDGPU_GFX_NORMAL_MODE;
 
-	gfx_v8_0_scratch_init(adev);
-
 	r = gfx_v8_0_init_microcode(adev);
 	if (r) {
 		DRM_ERROR("Failed to load gfx firmware!\n");
@@ -2044,7 +1987,8 @@ static int gfx_v8_0_sw_init(void *handle)
 	for (i = 0; i < adev->gfx.mec.num_mec; ++i) {
 		for (j = 0; j < adev->gfx.mec.num_queue_per_pipe; j++) {
 			for (k = 0; k < adev->gfx.mec.num_pipe_per_mec; k++) {
-				if (!amdgpu_gfx_is_mec_queue_enabled(adev, i, k, j))
+				if (!amdgpu_gfx_is_mec_queue_enabled(adev, 0, i,
+								     k, j))
 					continue;
 
 				r = gfx_v8_0_compute_ring_init(adev,
@@ -2058,19 +2002,18 @@ static int gfx_v8_0_sw_init(void *handle)
 		}
 	}
 
-	r = amdgpu_gfx_kiq_init(adev, GFX8_MEC_HPD_SIZE);
+	r = amdgpu_gfx_kiq_init(adev, GFX8_MEC_HPD_SIZE, 0);
 	if (r) {
 		DRM_ERROR("Failed to init KIQ BOs!\n");
 		return r;
 	}
 
-	kiq = &adev->gfx.kiq;
-	r = amdgpu_gfx_kiq_init_ring(adev, &kiq->ring, &kiq->irq);
+	r = amdgpu_gfx_kiq_init_ring(adev, xcc_id);
 	if (r)
 		return r;
 
 	/* create MQD for all compute queues as well as KIQ for SRIOV case */
-	r = amdgpu_gfx_mqd_sw_init(adev, sizeof(struct vi_mqd_allocation));
+	r = amdgpu_gfx_mqd_sw_init(adev, sizeof(struct vi_mqd_allocation), 0);
 	if (r)
 		return r;
 
@@ -2080,12 +2023,17 @@ static int gfx_v8_0_sw_init(void *handle)
 	if (r)
 		return r;
 
+	adev->gfx.gfx_supported_reset =
+		amdgpu_get_soft_full_reset_mask(&adev->gfx.gfx_ring[0]);
+	adev->gfx.compute_supported_reset =
+		amdgpu_get_soft_full_reset_mask(&adev->gfx.compute_ring[0]);
+
 	return 0;
 }
 
-static int gfx_v8_0_sw_fini(void *handle)
+static int gfx_v8_0_sw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int i;
 
 	for (i = 0; i < adev->gfx.num_gfx_rings; i++)
@@ -2093,9 +2041,9 @@ static int gfx_v8_0_sw_fini(void *handle)
 	for (i = 0; i < adev->gfx.num_compute_rings; i++)
 		amdgpu_ring_fini(&adev->gfx.compute_ring[i]);
 
-	amdgpu_gfx_mqd_sw_fini(adev);
-	amdgpu_gfx_kiq_free_ring(&adev->gfx.kiq.ring);
-	amdgpu_gfx_kiq_fini(adev);
+	amdgpu_gfx_mqd_sw_fini(adev, 0);
+	amdgpu_gfx_kiq_free_ring(&adev->gfx.kiq[0].ring);
+	amdgpu_gfx_kiq_fini(adev, 0);
 
 	gfx_v8_0_mec_fini(adev);
 	amdgpu_gfx_rlc_fini(adev);
@@ -3437,7 +3385,8 @@ static void gfx_v8_0_tiling_mode_table_init(struct amdgpu_device *adev)
 }
 
 static void gfx_v8_0_select_se_sh(struct amdgpu_device *adev,
-				  u32 se_num, u32 sh_num, u32 instance)
+				  u32 se_num, u32 sh_num, u32 instance,
+				  int xcc_id)
 {
 	u32 data;
 
@@ -3460,7 +3409,7 @@ static void gfx_v8_0_select_se_sh(struct amdgpu_device *adev,
 }
 
 static void gfx_v8_0_select_me_pipe_q(struct amdgpu_device *adev,
-				  u32 me, u32 pipe, u32 q, u32 vm)
+				  u32 me, u32 pipe, u32 q, u32 vm, u32 xcc_id)
 {
 	vi_srbm_select(adev, me, pipe, q, vm);
 }
@@ -3621,13 +3570,13 @@ gfx_v8_0_write_harvested_raster_configs(struct amdgpu_device *adev,
 		}
 
 		/* GRBM_GFX_INDEX has a different offset on VI */
-		gfx_v8_0_select_se_sh(adev, se, 0xffffffff, 0xffffffff);
+		gfx_v8_0_select_se_sh(adev, se, 0xffffffff, 0xffffffff, 0);
 		WREG32(mmPA_SC_RASTER_CONFIG, raster_config_se);
 		WREG32(mmPA_SC_RASTER_CONFIG_1, raster_config_1);
 	}
 
 	/* GRBM_GFX_INDEX has a different offset on VI */
-	gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+	gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
 }
 
 static void gfx_v8_0_setup_rb(struct amdgpu_device *adev)
@@ -3643,13 +3592,13 @@ static void gfx_v8_0_setup_rb(struct amdgpu_device *adev)
 	mutex_lock(&adev->grbm_idx_mutex);
 	for (i = 0; i < adev->gfx.config.max_shader_engines; i++) {
 		for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) {
-			gfx_v8_0_select_se_sh(adev, i, j, 0xffffffff);
+			gfx_v8_0_select_se_sh(adev, i, j, 0xffffffff, 0);
 			data = gfx_v8_0_get_rb_active_bitmap(adev);
 			active_rbs |= data << ((i * adev->gfx.config.max_sh_per_se + j) *
 					       rb_bitmap_width_per_sh);
 		}
 	}
-	gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+	gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
 
 	adev->gfx.config.backend_enable_mask = active_rbs;
 	adev->gfx.config.num_rbs = hweight32(active_rbs);
@@ -3672,7 +3621,7 @@ static void gfx_v8_0_setup_rb(struct amdgpu_device *adev)
 	/* cache the values for userspace */
 	for (i = 0; i < adev->gfx.config.max_shader_engines; i++) {
 		for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) {
-			gfx_v8_0_select_se_sh(adev, i, j, 0xffffffff);
+			gfx_v8_0_select_se_sh(adev, i, j, 0xffffffff, 0);
 			adev->gfx.config.rb_config[i][j].rb_backend_disable =
 				RREG32(mmCC_RB_BACKEND_DISABLE);
 			adev->gfx.config.rb_config[i][j].user_rb_backend_disable =
@@ -3683,7 +3632,7 @@ static void gfx_v8_0_setup_rb(struct amdgpu_device *adev)
 				RREG32(mmPA_SC_RASTER_CONFIG_1);
 		}
 	}
-	gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+	gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
 	mutex_unlock(&adev->grbm_idx_mutex);
 }
 
@@ -3730,7 +3679,7 @@ static void gfx_v8_0_init_compute_vmid(struct amdgpu_device *adev)
 	mutex_unlock(&adev->srbm_mutex);
 
 	/* Initialize all compute VMIDs to have no GDS, GWS, or OA
-	   acccess. These should be enabled by FW for target VMIDs. */
+	   access. These should be enabled by FW for target VMIDs. */
 	for (i = adev->vm_manager.first_kfd_vmid; i < AMDGPU_NUM_VMID; i++) {
 		WREG32(amdgpu_gds_reg_offset[i].mem_base, 0);
 		WREG32(amdgpu_gds_reg_offset[i].mem_size, 0);
@@ -3830,7 +3779,7 @@ static void gfx_v8_0_constants_init(struct amdgpu_device *adev)
 	 * making sure that the following register writes will be broadcasted
 	 * to all the shaders
 	 */
-	gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+	gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
 
 	WREG32(mmPA_SC_FIFO_SIZE,
 		   (adev->gfx.config.sc_prim_fifo_size_frontend <<
@@ -3861,7 +3810,7 @@ static void gfx_v8_0_wait_for_rlc_serdes(struct amdgpu_device *adev)
 	mutex_lock(&adev->grbm_idx_mutex);
 	for (i = 0; i < adev->gfx.config.max_shader_engines; i++) {
 		for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) {
-			gfx_v8_0_select_se_sh(adev, i, j, 0xffffffff);
+			gfx_v8_0_select_se_sh(adev, i, j, 0xffffffff, 0);
 			for (k = 0; k < adev->usec_timeout; k++) {
 				if (RREG32(mmRLC_SERDES_CU_MASTER_BUSY) == 0)
 					break;
@@ -3869,7 +3818,7 @@ static void gfx_v8_0_wait_for_rlc_serdes(struct amdgpu_device *adev)
 			}
 			if (k == adev->usec_timeout) {
 				gfx_v8_0_select_se_sh(adev, 0xffffffff,
-						      0xffffffff, 0xffffffff);
+						      0xffffffff, 0xffffffff, 0);
 				mutex_unlock(&adev->grbm_idx_mutex);
 				DRM_INFO("Timeout wait for RLC serdes %u,%u\n",
 					 i, j);
@@ -3877,7 +3826,7 @@ static void gfx_v8_0_wait_for_rlc_serdes(struct amdgpu_device *adev)
 			}
 		}
 	}
-	gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+	gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
 	mutex_unlock(&adev->grbm_idx_mutex);
 
 	mask = RLC_SERDES_NONCU_MASTER_BUSY__SE_MASTER_BUSY_MASK |
@@ -4305,12 +4254,12 @@ static int gfx_v8_0_cp_gfx_resume(struct amdgpu_device *adev)
 	ring->wptr = 0;
 	WREG32(mmCP_RB0_WPTR, lower_32_bits(ring->wptr));
 
-	/* set the wb address wether it's enabled or not */
-	rptr_addr = adev->wb.gpu_addr + (ring->rptr_offs * 4);
+	/* set the wb address whether it's enabled or not */
+	rptr_addr = ring->rptr_gpu_addr;
 	WREG32(mmCP_RB0_RPTR_ADDR, lower_32_bits(rptr_addr));
 	WREG32(mmCP_RB0_RPTR_ADDR_HI, upper_32_bits(rptr_addr) & 0xFF);
 
-	wptr_gpu_addr = adev->wb.gpu_addr + (ring->wptr_offs * 4);
+	wptr_gpu_addr = ring->wptr_gpu_addr;
 	WREG32(mmCP_RB_WPTR_POLL_ADDR_LO, lower_32_bits(wptr_gpu_addr));
 	WREG32(mmCP_RB_WPTR_POLL_ADDR_HI, upper_32_bits(wptr_gpu_addr));
 	mdelay(1);
@@ -4324,7 +4273,6 @@ static int gfx_v8_0_cp_gfx_resume(struct amdgpu_device *adev)
 	/* start the ring */
 	amdgpu_ring_clear_ring(ring);
 	gfx_v8_0_cp_gfx_start(adev);
-	ring->sched.ready = true;
 
 	return 0;
 }
@@ -4335,7 +4283,7 @@ static void gfx_v8_0_cp_compute_enable(struct amdgpu_device *adev, bool enable)
 		WREG32(mmCP_MEC_CNTL, 0);
 	} else {
 		WREG32(mmCP_MEC_CNTL, (CP_MEC_CNTL__MEC_ME1_HALT_MASK | CP_MEC_CNTL__MEC_ME2_HALT_MASK));
-		adev->gfx.kiq.ring.sched.ready = false;
+		adev->gfx.kiq[0].ring.sched.ready = false;
 	}
 	udelay(50);
 }
@@ -4350,19 +4298,17 @@ static void gfx_v8_0_kiq_setting(struct amdgpu_ring *ring)
 	tmp = RREG32(mmRLC_CP_SCHEDULERS);
 	tmp &= 0xffffff00;
 	tmp |= (ring->me << 5) | (ring->pipe << 3) | (ring->queue);
-	WREG32(mmRLC_CP_SCHEDULERS, tmp);
-	tmp |= 0x80;
-	WREG32(mmRLC_CP_SCHEDULERS, tmp);
+	WREG32(mmRLC_CP_SCHEDULERS, tmp | 0x80);
 }
 
 static int gfx_v8_0_kiq_kcq_enable(struct amdgpu_device *adev)
 {
-	struct amdgpu_ring *kiq_ring = &adev->gfx.kiq.ring;
+	struct amdgpu_ring *kiq_ring = &adev->gfx.kiq[0].ring;
 	uint64_t queue_mask = 0;
 	int r, i;
 
 	for (i = 0; i < AMDGPU_MAX_COMPUTE_QUEUES; ++i) {
-		if (!test_bit(i, adev->gfx.mec.queue_bitmap))
+		if (!test_bit(i, adev->gfx.mec_bitmap[0].queue_bitmap))
 			continue;
 
 		/* This situation may be hit in the future if a new HW
@@ -4393,7 +4339,7 @@ static int gfx_v8_0_kiq_kcq_enable(struct amdgpu_device *adev)
 	for (i = 0; i < adev->gfx.num_compute_rings; i++) {
 		struct amdgpu_ring *ring = &adev->gfx.compute_ring[i];
 		uint64_t mqd_addr = amdgpu_bo_gpu_offset(ring->mqd_obj);
-		uint64_t wptr_addr = adev->wb.gpu_addr + (ring->wptr_offs * 4);
+		uint64_t wptr_addr = ring->wptr_gpu_addr;
 
 		/* map queues */
 		amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_MAP_QUEUES, 5));
@@ -4506,7 +4452,7 @@ static int gfx_v8_0_mqd_init(struct amdgpu_ring *ring)
 	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, QUEUE_SIZE,
 			    (order_base_2(ring->ring_size / 4) - 1));
 	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, RPTR_BLOCK_SIZE,
-			((order_base_2(AMDGPU_GPU_PAGE_SIZE / 4) - 1) << 8));
+			(order_base_2(AMDGPU_GPU_PAGE_SIZE / 4) - 1));
 #ifdef __BIG_ENDIAN
 	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, ENDIAN_SWAP, 1);
 #endif
@@ -4517,13 +4463,13 @@ static int gfx_v8_0_mqd_init(struct amdgpu_ring *ring)
 	mqd->cp_hqd_pq_control = tmp;
 
 	/* set the wb address whether it's enabled or not */
-	wb_gpu_addr = adev->wb.gpu_addr + (ring->rptr_offs * 4);
+	wb_gpu_addr = ring->rptr_gpu_addr;
 	mqd->cp_hqd_pq_rptr_report_addr_lo = wb_gpu_addr & 0xfffffffc;
 	mqd->cp_hqd_pq_rptr_report_addr_hi =
 		upper_32_bits(wb_gpu_addr) & 0xffff;
 
 	/* only used if CP_PQ_WPTR_POLL_CNTL.CP_PQ_WPTR_POLL_CNTL__EN_MASK=1 */
-	wb_gpu_addr = adev->wb.gpu_addr + (ring->wptr_offs * 4);
+	wb_gpu_addr = ring->wptr_gpu_addr;
 	mqd->cp_hqd_pq_wptr_poll_addr_lo = wb_gpu_addr & 0xfffffffc;
 	mqd->cp_hqd_pq_wptr_poll_addr_hi = upper_32_bits(wb_gpu_addr) & 0xffff;
 
@@ -4638,14 +4584,13 @@ static int gfx_v8_0_kiq_init_queue(struct amdgpu_ring *ring)
 {
 	struct amdgpu_device *adev = ring->adev;
 	struct vi_mqd *mqd = ring->mqd_ptr;
-	int mqd_idx = AMDGPU_MAX_COMPUTE_RINGS;
 
 	gfx_v8_0_kiq_setting(ring);
 
 	if (amdgpu_in_reset(adev)) { /* for GPU_RESET case */
 		/* reset MQD to a clean status */
-		if (adev->gfx.mec.mqd_backup[mqd_idx])
-			memcpy(mqd, adev->gfx.mec.mqd_backup[mqd_idx], sizeof(struct vi_mqd_allocation));
+		if (adev->gfx.kiq[0].mqd_backup)
+			memcpy(mqd, adev->gfx.kiq[0].mqd_backup, sizeof(struct vi_mqd_allocation));
 
 		/* reset ring buffer */
 		ring->wptr = 0;
@@ -4659,6 +4604,8 @@ static int gfx_v8_0_kiq_init_queue(struct amdgpu_ring *ring)
 		memset((void *)mqd, 0, sizeof(struct vi_mqd_allocation));
 		((struct vi_mqd_allocation *)mqd)->dynamic_cu_mask = 0xFFFFFFFF;
 		((struct vi_mqd_allocation *)mqd)->dynamic_rb_mask = 0xFFFFFFFF;
+		if (amdgpu_sriov_vf(adev) && adev->in_suspend)
+			amdgpu_ring_clear_ring(ring);
 		mutex_lock(&adev->srbm_mutex);
 		vi_srbm_select(adev, ring->me, ring->pipe, ring->queue, 0);
 		gfx_v8_0_mqd_init(ring);
@@ -4666,8 +4613,8 @@ static int gfx_v8_0_kiq_init_queue(struct amdgpu_ring *ring)
 		vi_srbm_select(adev, 0, 0, 0, 0);
 		mutex_unlock(&adev->srbm_mutex);
 
-		if (adev->gfx.mec.mqd_backup[mqd_idx])
-			memcpy(adev->gfx.mec.mqd_backup[mqd_idx], mqd, sizeof(struct vi_mqd_allocation));
+		if (adev->gfx.kiq[0].mqd_backup)
+			memcpy(adev->gfx.kiq[0].mqd_backup, mqd, sizeof(struct vi_mqd_allocation));
 	}
 
 	return 0;
@@ -4691,14 +4638,13 @@ static int gfx_v8_0_kcq_init_queue(struct amdgpu_ring *ring)
 
 		if (adev->gfx.mec.mqd_backup[mqd_idx])
 			memcpy(adev->gfx.mec.mqd_backup[mqd_idx], mqd, sizeof(struct vi_mqd_allocation));
-	} else if (amdgpu_in_reset(adev)) { /* for GPU_RESET case */
-		/* reset MQD to a clean status */
+	} else {
+		/* restore MQD to a clean status */
 		if (adev->gfx.mec.mqd_backup[mqd_idx])
 			memcpy(mqd, adev->gfx.mec.mqd_backup[mqd_idx], sizeof(struct vi_mqd_allocation));
 		/* reset ring buffer */
 		ring->wptr = 0;
-		amdgpu_ring_clear_ring(ring);
-	} else {
+		atomic64_set((atomic64_t *)ring->wptr_cpu_addr, 0);
 		amdgpu_ring_clear_ring(ring);
 	}
 	return 0;
@@ -4716,59 +4662,25 @@ static void gfx_v8_0_set_mec_doorbell_range(struct amdgpu_device *adev)
 
 static int gfx_v8_0_kiq_resume(struct amdgpu_device *adev)
 {
-	struct amdgpu_ring *ring;
-	int r;
-
-	ring = &adev->gfx.kiq.ring;
-
-	r = amdgpu_bo_reserve(ring->mqd_obj, false);
-	if (unlikely(r != 0))
-		return r;
-
-	r = amdgpu_bo_kmap(ring->mqd_obj, &ring->mqd_ptr);
-	if (unlikely(r != 0))
-		return r;
-
-	gfx_v8_0_kiq_init_queue(ring);
-	amdgpu_bo_kunmap(ring->mqd_obj);
-	ring->mqd_ptr = NULL;
-	amdgpu_bo_unreserve(ring->mqd_obj);
-	ring->sched.ready = true;
+	gfx_v8_0_kiq_init_queue(&adev->gfx.kiq[0].ring);
 	return 0;
 }
 
 static int gfx_v8_0_kcq_resume(struct amdgpu_device *adev)
 {
-	struct amdgpu_ring *ring = NULL;
-	int r = 0, i;
+	int i, r;
 
 	gfx_v8_0_cp_compute_enable(adev, true);
 
 	for (i = 0; i < adev->gfx.num_compute_rings; i++) {
-		ring = &adev->gfx.compute_ring[i];
-
-		r = amdgpu_bo_reserve(ring->mqd_obj, false);
-		if (unlikely(r != 0))
-			goto done;
-		r = amdgpu_bo_kmap(ring->mqd_obj, &ring->mqd_ptr);
-		if (!r) {
-			r = gfx_v8_0_kcq_init_queue(ring);
-			amdgpu_bo_kunmap(ring->mqd_obj);
-			ring->mqd_ptr = NULL;
-		}
-		amdgpu_bo_unreserve(ring->mqd_obj);
+		r = gfx_v8_0_kcq_init_queue(&adev->gfx.compute_ring[i]);
 		if (r)
-			goto done;
+			return r;
 	}
 
 	gfx_v8_0_set_mec_doorbell_range(adev);
 
-	r = gfx_v8_0_kiq_kcq_enable(adev);
-	if (r)
-		goto done;
-
-done:
-	return r;
+	return gfx_v8_0_kiq_kcq_enable(adev);
 }
 
 static int gfx_v8_0_cp_test_all_rings(struct amdgpu_device *adev)
@@ -4782,7 +4694,7 @@ static int gfx_v8_0_cp_test_all_rings(struct amdgpu_device *adev)
 	if (r)
 		return r;
 
-	ring = &adev->gfx.kiq.ring;
+	ring = &adev->gfx.kiq[0].ring;
 	r = amdgpu_ring_test_helper(ring);
 	if (r)
 		return r;
@@ -4829,10 +4741,10 @@ static void gfx_v8_0_cp_enable(struct amdgpu_device *adev, bool enable)
 	gfx_v8_0_cp_compute_enable(adev, enable);
 }
 
-static int gfx_v8_0_hw_init(void *handle)
+static int gfx_v8_0_hw_init(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	gfx_v8_0_init_golden_registers(adev);
 	gfx_v8_0_constants_init(adev);
@@ -4849,7 +4761,7 @@ static int gfx_v8_0_hw_init(void *handle)
 static int gfx_v8_0_kcq_disable(struct amdgpu_device *adev)
 {
 	int r, i;
-	struct amdgpu_ring *kiq_ring = &adev->gfx.kiq.ring;
+	struct amdgpu_ring *kiq_ring = &adev->gfx.kiq[0].ring;
 
 	r = amdgpu_ring_alloc(kiq_ring, 6 * adev->gfx.num_compute_rings);
 	if (r)
@@ -4869,6 +4781,13 @@ static int gfx_v8_0_kcq_disable(struct amdgpu_device *adev)
 		amdgpu_ring_write(kiq_ring, 0);
 		amdgpu_ring_write(kiq_ring, 0);
 	}
+	/* Submit unmap queue packet */
+	amdgpu_ring_commit(kiq_ring);
+	/*
+	 * Ring test will do a basic scratch register change check. Just run
+	 * this to ensure that unmap queues that is submitted before got
+	 * processed successfully before returning.
+	 */
 	r = amdgpu_ring_test_helper(kiq_ring);
 	if (r)
 		DRM_ERROR("KCQ disable failed\n");
@@ -4876,9 +4795,9 @@ static int gfx_v8_0_kcq_disable(struct amdgpu_device *adev)
 	return r;
 }
 
-static bool gfx_v8_0_is_idle(void *handle)
+static bool gfx_v8_0_is_idle(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (REG_GET_FIELD(RREG32(mmGRBM_STATUS), GRBM_STATUS, GUI_ACTIVE)
 		|| RREG32(mmGRBM_STATUS2) != 0x8)
@@ -4911,13 +4830,13 @@ static int gfx_v8_0_wait_for_rlc_idle(void *handle)
 	return -ETIMEDOUT;
 }
 
-static int gfx_v8_0_wait_for_idle(void *handle)
+static int gfx_v8_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
 {
 	unsigned int i;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	for (i = 0; i < adev->usec_timeout; i++) {
-		if (gfx_v8_0_is_idle(handle))
+		if (gfx_v8_0_is_idle(ip_block))
 			return 0;
 
 		udelay(1);
@@ -4925,9 +4844,9 @@ static int gfx_v8_0_wait_for_idle(void *handle)
 	return -ETIMEDOUT;
 }
 
-static int gfx_v8_0_hw_fini(void *handle)
+static int gfx_v8_0_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	amdgpu_irq_put(adev, &adev->gfx.priv_reg_irq, 0);
 	amdgpu_irq_put(adev, &adev->gfx.priv_inst_irq, 0);
@@ -4943,8 +4862,9 @@ static int gfx_v8_0_hw_fini(void *handle)
 		pr_debug("For SRIOV client, shouldn't do anything.\n");
 		return 0;
 	}
-	amdgpu_gfx_rlc_enter_safe_mode(adev);
-	if (!gfx_v8_0_wait_for_idle(adev))
+
+	amdgpu_gfx_rlc_enter_safe_mode(adev, 0);
+	if (!gfx_v8_0_wait_for_idle(ip_block))
 		gfx_v8_0_cp_enable(adev, false);
 	else
 		pr_err("cp is busy, skip halt cp\n");
@@ -4952,24 +4872,24 @@ static int gfx_v8_0_hw_fini(void *handle)
 		adev->gfx.rlc.funcs->stop(adev);
 	else
 		pr_err("rlc is busy, skip halt rlc\n");
-	amdgpu_gfx_rlc_exit_safe_mode(adev);
+	amdgpu_gfx_rlc_exit_safe_mode(adev, 0);
 
 	return 0;
 }
 
-static int gfx_v8_0_suspend(void *handle)
+static int gfx_v8_0_suspend(struct amdgpu_ip_block *ip_block)
 {
-	return gfx_v8_0_hw_fini(handle);
+	return gfx_v8_0_hw_fini(ip_block);
 }
 
-static int gfx_v8_0_resume(void *handle)
+static int gfx_v8_0_resume(struct amdgpu_ip_block *ip_block)
 {
-	return gfx_v8_0_hw_init(handle);
+	return gfx_v8_0_hw_init(ip_block);
 }
 
-static bool gfx_v8_0_check_soft_reset(void *handle)
+static bool gfx_v8_0_check_soft_reset(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	u32 grbm_soft_reset = 0, srbm_soft_reset = 0;
 	u32 tmp;
 
@@ -5029,9 +4949,9 @@ static bool gfx_v8_0_check_soft_reset(void *handle)
 	}
 }
 
-static int gfx_v8_0_pre_soft_reset(void *handle)
+static int gfx_v8_0_pre_soft_reset(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	u32 grbm_soft_reset = 0;
 
 	if ((!adev->gfx.grbm_soft_reset) &&
@@ -5070,9 +4990,9 @@ static int gfx_v8_0_pre_soft_reset(void *handle)
 	return 0;
 }
 
-static int gfx_v8_0_soft_reset(void *handle)
+static int gfx_v8_0_soft_reset(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	u32 grbm_soft_reset = 0, srbm_soft_reset = 0;
 	u32 tmp;
 
@@ -5132,9 +5052,9 @@ static int gfx_v8_0_soft_reset(void *handle)
 	return 0;
 }
 
-static int gfx_v8_0_post_soft_reset(void *handle)
+static int gfx_v8_0_post_soft_reset(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	u32 grbm_soft_reset = 0;
 
 	if ((!adev->gfx.grbm_soft_reset) &&
@@ -5257,7 +5177,7 @@ static void wave_read_regs(struct amdgpu_device *adev, uint32_t simd,
 		*(out++) = RREG32(mmSQ_IND_DATA);
 }
 
-static void gfx_v8_0_read_wave_data(struct amdgpu_device *adev, uint32_t simd, uint32_t wave, uint32_t *dst, int *no_fields)
+static void gfx_v8_0_read_wave_data(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd, uint32_t wave, uint32_t *dst, int *no_fields)
 {
 	/* type 0 wave data */
 	dst[(*no_fields)++] = 0;
@@ -5282,7 +5202,7 @@ static void gfx_v8_0_read_wave_data(struct amdgpu_device *adev, uint32_t simd, u
 	dst[(*no_fields)++] = wave_read_ind(adev, simd, wave, ixSQ_WAVE_MODE);
 }
 
-static void gfx_v8_0_read_wave_sgprs(struct amdgpu_device *adev, uint32_t simd,
+static void gfx_v8_0_read_wave_sgprs(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd,
 				     uint32_t wave, uint32_t start,
 				     uint32_t size, uint32_t *dst)
 {
@@ -5300,10 +5220,11 @@ static const struct amdgpu_gfx_funcs gfx_v8_0_gfx_funcs = {
 	.select_me_pipe_q = &gfx_v8_0_select_me_pipe_q
 };
 
-static int gfx_v8_0_early_init(void *handle)
+static int gfx_v8_0_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
+	adev->gfx.xcc_mask = 1;
 	adev->gfx.num_gfx_rings = GFX8_NUM_GFX_RINGS;
 	adev->gfx.num_compute_rings = min(amdgpu_gfx_get_num_kcq(adev),
 					  AMDGPU_MAX_COMPUTE_RINGS);
@@ -5316,9 +5237,9 @@ static int gfx_v8_0_early_init(void *handle)
 	return 0;
 }
 
-static int gfx_v8_0_late_init(void *handle)
+static int gfx_v8_0_late_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int r;
 
 	r = amdgpu_irq_get(adev, &adev->gfx.priv_reg_irq, 0);
@@ -5358,7 +5279,7 @@ static void gfx_v8_0_enable_gfx_static_mg_power_gating(struct amdgpu_device *ade
 	    (adev->asic_type == CHIP_POLARIS12) ||
 	    (adev->asic_type == CHIP_VEGAM))
 		/* Send msg to SMU via Powerplay */
-		amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_GFX, enable);
+		amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_GFX, enable, 0);
 
 	WREG32_FIELD(RLC_PG_CNTL, STATIC_PER_CU_PG_ENABLE, enable ? 1 : 0);
 }
@@ -5404,10 +5325,10 @@ static void cz_update_gfx_cg_power_gating(struct amdgpu_device *adev,
 	}
 }
 
-static int gfx_v8_0_set_powergating_state(void *handle,
+static int gfx_v8_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_powergating_state state)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	bool enable = (state == AMD_PG_STATE_GATE);
 
 	if (amdgpu_sriov_vf(adev))
@@ -5417,7 +5338,7 @@ static int gfx_v8_0_set_powergating_state(void *handle,
 				AMD_PG_SUPPORT_RLC_SMU_HS |
 				AMD_PG_SUPPORT_CP |
 				AMD_PG_SUPPORT_GFX_DMG))
-		amdgpu_gfx_rlc_enter_safe_mode(adev);
+		amdgpu_gfx_rlc_enter_safe_mode(adev, 0);
 	switch (adev->asic_type) {
 	case CHIP_CARRIZO:
 	case CHIP_STONEY:
@@ -5471,13 +5392,13 @@ static int gfx_v8_0_set_powergating_state(void *handle,
 				AMD_PG_SUPPORT_RLC_SMU_HS |
 				AMD_PG_SUPPORT_CP |
 				AMD_PG_SUPPORT_GFX_DMG))
-		amdgpu_gfx_rlc_exit_safe_mode(adev);
+		amdgpu_gfx_rlc_exit_safe_mode(adev, 0);
 	return 0;
 }
 
-static void gfx_v8_0_get_clockgating_state(void *handle, u32 *flags)
+static void gfx_v8_0_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int data;
 
 	if (amdgpu_sriov_vf(adev))
@@ -5522,7 +5443,7 @@ static void gfx_v8_0_send_serdes_cmd(struct amdgpu_device *adev,
 {
 	uint32_t data;
 
-	gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+	gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
 
 	WREG32(mmRLC_SERDES_WR_CU_MASTER_MASK, 0xffffffff);
 	WREG32(mmRLC_SERDES_WR_NONCU_MASTER_MASK, 0xffffffff);
@@ -5576,7 +5497,7 @@ static bool gfx_v8_0_is_rlc_enabled(struct amdgpu_device *adev)
 	return true;
 }
 
-static void gfx_v8_0_set_safe_mode(struct amdgpu_device *adev)
+static void gfx_v8_0_set_safe_mode(struct amdgpu_device *adev, int xcc_id)
 {
 	uint32_t data;
 	unsigned i;
@@ -5603,7 +5524,7 @@ static void gfx_v8_0_set_safe_mode(struct amdgpu_device *adev)
 	}
 }
 
-static void gfx_v8_0_unset_safe_mode(struct amdgpu_device *adev)
+static void gfx_v8_0_unset_safe_mode(struct amdgpu_device *adev, int xcc_id)
 {
 	uint32_t data;
 	unsigned i;
@@ -5620,7 +5541,7 @@ static void gfx_v8_0_unset_safe_mode(struct amdgpu_device *adev)
 	}
 }
 
-static void gfx_v8_0_update_spm_vmid(struct amdgpu_device *adev, unsigned vmid)
+static void gfx_v8_0_update_spm_vmid(struct amdgpu_device *adev, struct amdgpu_ring *ring, unsigned vmid)
 {
 	u32 data;
 
@@ -5662,8 +5583,6 @@ static void gfx_v8_0_update_medium_grain_clock_gating(struct amdgpu_device *adev
 {
 	uint32_t temp, data;
 
-	amdgpu_gfx_rlc_enter_safe_mode(adev);
-
 	/* It is disabled by HW by default */
 	if (enable && (adev->cg_flags & AMD_CG_SUPPORT_GFX_MGCG)) {
 		if (adev->cg_flags & AMD_CG_SUPPORT_GFX_MGLS) {
@@ -5757,8 +5676,6 @@ static void gfx_v8_0_update_medium_grain_clock_gating(struct amdgpu_device *adev
 		/* 7- wait for RLC_SERDES_CU_MASTER & RLC_SERDES_NONCU_MASTER idle */
 		gfx_v8_0_wait_for_rlc_serdes(adev);
 	}
-
-	amdgpu_gfx_rlc_exit_safe_mode(adev);
 }
 
 static void gfx_v8_0_update_coarse_grain_clock_gating(struct amdgpu_device *adev,
@@ -5768,8 +5685,6 @@ static void gfx_v8_0_update_coarse_grain_clock_gating(struct amdgpu_device *adev
 
 	temp = data = RREG32(mmRLC_CGCG_CGLS_CTRL);
 
-	amdgpu_gfx_rlc_enter_safe_mode(adev);
-
 	if (enable && (adev->cg_flags & AMD_CG_SUPPORT_GFX_CGCG)) {
 		temp1 = data1 =	RREG32(mmRLC_CGTT_MGCG_OVERRIDE);
 		data1 &= ~RLC_CGTT_MGCG_OVERRIDE__CGCG_MASK;
@@ -5831,7 +5746,7 @@ static void gfx_v8_0_update_coarse_grain_clock_gating(struct amdgpu_device *adev
 		/* wait for RLC_SERDES_CU_MASTER & RLC_SERDES_NONCU_MASTER idle */
 		gfx_v8_0_wait_for_rlc_serdes(adev);
 
-		/* write cmd to Set CGCG Overrride */
+		/* write cmd to Set CGCG Override */
 		gfx_v8_0_send_serdes_cmd(adev, BPM_REG_CGCG_OVERRIDE, SET_BPM_SERDES_CMD);
 
 		/* wait for RLC_SERDES_CU_MASTER & RLC_SERDES_NONCU_MASTER idle */
@@ -5850,12 +5765,12 @@ static void gfx_v8_0_update_coarse_grain_clock_gating(struct amdgpu_device *adev
 	}
 
 	gfx_v8_0_wait_for_rlc_serdes(adev);
-
-	amdgpu_gfx_rlc_exit_safe_mode(adev);
 }
 static int gfx_v8_0_update_gfx_clock_gating(struct amdgpu_device *adev,
 					    bool enable)
 {
+	amdgpu_gfx_rlc_enter_safe_mode(adev, 0);
+
 	if (enable) {
 		/* CGCG/CGLS should be enabled after MGCG/MGLS/TS(CG/LS)
 		 * ===  MGCG + MGLS + TS(CG/LS) ===
@@ -5869,6 +5784,8 @@ static int gfx_v8_0_update_gfx_clock_gating(struct amdgpu_device *adev,
 		gfx_v8_0_update_coarse_grain_clock_gating(adev, enable);
 		gfx_v8_0_update_medium_grain_clock_gating(adev, enable);
 	}
+
+	amdgpu_gfx_rlc_exit_safe_mode(adev, 0);
 	return 0;
 }
 
@@ -6019,10 +5936,10 @@ static int gfx_v8_0_polaris_update_gfx_clock_gating(struct amdgpu_device *adev,
 	return 0;
 }
 
-static int gfx_v8_0_set_clockgating_state(void *handle,
+static int gfx_v8_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_clockgating_state state)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (amdgpu_sriov_vf(adev))
 		return 0;
@@ -6051,7 +5968,7 @@ static int gfx_v8_0_set_clockgating_state(void *handle,
 
 static u64 gfx_v8_0_ring_get_rptr(struct amdgpu_ring *ring)
 {
-	return ring->adev->wb.wb[ring->rptr_offs];
+	return *ring->rptr_cpu_addr;
 }
 
 static u64 gfx_v8_0_ring_get_wptr_gfx(struct amdgpu_ring *ring)
@@ -6060,7 +5977,7 @@ static u64 gfx_v8_0_ring_get_wptr_gfx(struct amdgpu_ring *ring)
 
 	if (ring->use_doorbell)
 		/* XXX check if swapping is necessary on BE */
-		return ring->adev->wb.wb[ring->wptr_offs];
+		return *ring->wptr_cpu_addr;
 	else
 		return RREG32(mmCP_RB0_WPTR);
 }
@@ -6071,7 +5988,7 @@ static void gfx_v8_0_ring_set_wptr_gfx(struct amdgpu_ring *ring)
 
 	if (ring->use_doorbell) {
 		/* XXX check if swapping is necessary on BE */
-		adev->wb.wb[ring->wptr_offs] = lower_32_bits(ring->wptr);
+		*ring->wptr_cpu_addr = lower_32_bits(ring->wptr);
 		WDOORBELL32(ring->doorbell_index, lower_32_bits(ring->wptr));
 	} else {
 		WREG32(mmCP_RB0_WPTR, lower_32_bits(ring->wptr));
@@ -6194,6 +6111,7 @@ static void gfx_v8_0_ring_emit_fence_gfx(struct amdgpu_ring *ring, u64 addr,
 {
 	bool write64bit = flags & AMDGPU_FENCE_FLAG_64BIT;
 	bool int_sel = flags & AMDGPU_FENCE_FLAG_INT;
+	bool exec = flags & AMDGPU_FENCE_FLAG_EXEC;
 
 	/* Workaround for cache flush problems. First send a dummy EOP
 	 * event down the pipe with seq one below.
@@ -6217,7 +6135,8 @@ static void gfx_v8_0_ring_emit_fence_gfx(struct amdgpu_ring *ring, u64 addr,
 				 EOP_TC_ACTION_EN |
 				 EOP_TC_WB_ACTION_EN |
 				 EVENT_TYPE(CACHE_FLUSH_AND_INV_TS_EVENT) |
-				 EVENT_INDEX(5)));
+				 EVENT_INDEX(5) |
+				 (exec ? EOP_EXEC : 0)));
 	amdgpu_ring_write(ring, addr & 0xfffffffc);
 	amdgpu_ring_write(ring, (upper_32_bits(addr) & 0xffff) |
 			  DATA_SEL(write64bit ? 2 : 1) | INT_SEL(int_sel ? 2 : 0));
@@ -6271,7 +6190,7 @@ static void gfx_v8_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
 
 static u64 gfx_v8_0_ring_get_wptr_compute(struct amdgpu_ring *ring)
 {
-	return ring->adev->wb.wb[ring->wptr_offs];
+	return *ring->wptr_cpu_addr;
 }
 
 static void gfx_v8_0_ring_set_wptr_compute(struct amdgpu_ring *ring)
@@ -6279,7 +6198,7 @@ static void gfx_v8_0_ring_set_wptr_compute(struct amdgpu_ring *ring)
 	struct amdgpu_device *adev = ring->adev;
 
 	/* XXX check if swapping is necessary on BE */
-	adev->wb.wb[ring->wptr_offs] = lower_32_bits(ring->wptr);
+	*ring->wptr_cpu_addr = lower_32_bits(ring->wptr);
 	WDOORBELL32(ring->doorbell_index, lower_32_bits(ring->wptr));
 }
 
@@ -6368,33 +6287,22 @@ static void gfx_v8_ring_emit_cntxcntl(struct amdgpu_ring *ring, uint32_t flags)
 	amdgpu_ring_write(ring, 0);
 }
 
-static unsigned gfx_v8_0_ring_emit_init_cond_exec(struct amdgpu_ring *ring)
+static unsigned gfx_v8_0_ring_emit_init_cond_exec(struct amdgpu_ring *ring,
+						  uint64_t addr)
 {
 	unsigned ret;
 
 	amdgpu_ring_write(ring, PACKET3(PACKET3_COND_EXEC, 3));
-	amdgpu_ring_write(ring, lower_32_bits(ring->cond_exe_gpu_addr));
-	amdgpu_ring_write(ring, upper_32_bits(ring->cond_exe_gpu_addr));
-	amdgpu_ring_write(ring, 0); /* discard following DWs if *cond_exec_gpu_addr==0 */
+	amdgpu_ring_write(ring, lower_32_bits(addr));
+	amdgpu_ring_write(ring, upper_32_bits(addr));
+	/* discard following DWs if *cond_exec_gpu_addr==0 */
+	amdgpu_ring_write(ring, 0);
 	ret = ring->wptr & ring->buf_mask;
-	amdgpu_ring_write(ring, 0x55aa55aa); /* patch dummy value later */
+	/* patch dummy value later */
+	amdgpu_ring_write(ring, 0);
 	return ret;
 }
 
-static void gfx_v8_0_ring_emit_patch_cond_exec(struct amdgpu_ring *ring, unsigned offset)
-{
-	unsigned cur;
-
-	BUG_ON(offset > ring->buf_mask);
-	BUG_ON(ring->ring[offset] != 0x55aa55aa);
-
-	cur = (ring->wptr & ring->buf_mask) - 1;
-	if (likely(cur > offset))
-		ring->ring[offset] = cur - offset;
-	else
-		ring->ring[offset] = (ring->ring_size >> 2) - offset + cur;
-}
-
 static void gfx_v8_0_ring_emit_rreg(struct amdgpu_ring *ring, uint32_t reg,
 				    uint32_t reg_val_offs)
 {
@@ -6764,11 +6672,11 @@ static void gfx_v8_0_parse_sq_irq(struct amdgpu_device *adev, unsigned ih_data,
 			 */
 			if (from_wq) {
 				mutex_lock(&adev->grbm_idx_mutex);
-				gfx_v8_0_select_se_sh(adev, se_id, sh_id, cu_id);
+				gfx_v8_0_select_se_sh(adev, se_id, sh_id, cu_id, 0);
 
 				sq_edc_source = REG_GET_FIELD(RREG32(mmSQ_EDC_INFO), SQ_EDC_INFO, SOURCE);
 
-				gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+				gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
 				mutex_unlock(&adev->grbm_idx_mutex);
 			}
 
@@ -6974,7 +6882,6 @@ static const struct amdgpu_ring_funcs gfx_v8_0_ring_funcs_gfx = {
 	.emit_switch_buffer = gfx_v8_ring_emit_sb,
 	.emit_cntxcntl = gfx_v8_ring_emit_cntxcntl,
 	.init_cond_exec = gfx_v8_0_ring_emit_init_cond_exec,
-	.patch_cond_exec = gfx_v8_0_ring_emit_patch_cond_exec,
 	.emit_wreg = gfx_v8_0_ring_emit_wreg,
 	.soft_recovery = gfx_v8_0_ring_soft_recovery,
 	.emit_mem_sync = gfx_v8_0_emit_mem_sync,
@@ -7010,6 +6917,7 @@ static const struct amdgpu_ring_funcs gfx_v8_0_ring_funcs_compute = {
 	.insert_nop = amdgpu_ring_insert_nop,
 	.pad_ib = amdgpu_ring_generic_pad_ib,
 	.emit_wreg = gfx_v8_0_ring_emit_wreg,
+	.soft_recovery = gfx_v8_0_ring_soft_recovery,
 	.emit_mem_sync = gfx_v8_0_emit_mem_sync_compute,
 	.emit_wave_limit = gfx_v8_0_emit_wave_limit,
 };
@@ -7036,13 +6944,14 @@ static const struct amdgpu_ring_funcs gfx_v8_0_ring_funcs_kiq = {
 	.pad_ib = amdgpu_ring_generic_pad_ib,
 	.emit_rreg = gfx_v8_0_ring_emit_rreg,
 	.emit_wreg = gfx_v8_0_ring_emit_wreg,
+	.emit_hdp_flush = gfx_v8_0_ring_emit_hdp_flush,
 };
 
 static void gfx_v8_0_set_ring_funcs(struct amdgpu_device *adev)
 {
 	int i;
 
-	adev->gfx.kiq.ring.funcs = &gfx_v8_0_ring_funcs_kiq;
+	adev->gfx.kiq[0].ring.funcs = &gfx_v8_0_ring_funcs_kiq;
 
 	for (i = 0; i < adev->gfx.num_gfx_rings; i++)
 		adev->gfx.gfx_ring[i].funcs = &gfx_v8_0_ring_funcs_gfx;
@@ -7157,12 +7066,12 @@ static void gfx_v8_0_get_cu_info(struct amdgpu_device *adev)
 			mask = 1;
 			ao_bitmap = 0;
 			counter = 0;
-			gfx_v8_0_select_se_sh(adev, i, j, 0xffffffff);
+			gfx_v8_0_select_se_sh(adev, i, j, 0xffffffff, 0);
 			if (i < 4 && j < 2)
 				gfx_v8_0_set_user_cu_inactive_bitmap(
 					adev, disable_masks[i * 2 + j]);
 			bitmap = gfx_v8_0_get_cu_active_bitmap(adev);
-			cu_info->bitmap[i][j] = bitmap;
+			cu_info->bitmap[0][i][j] = bitmap;
 
 			for (k = 0; k < adev->gfx.config.max_cu_per_sh; k ++) {
 				if (bitmap & mask) {
@@ -7178,7 +7087,7 @@ static void gfx_v8_0_get_cu_info(struct amdgpu_device *adev)
 			cu_info->ao_cu_bitmap[i][j] = ao_bitmap;
 		}
 	}
-	gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+	gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
 	mutex_unlock(&adev->grbm_idx_mutex);
 
 	cu_info->number = active_cu_number;
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
index edb3e3b08eed..0148d7ff34d9 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
@@ -47,8 +47,10 @@
 
 #include "amdgpu_ras.h"
 
+#include "amdgpu_ring_mux.h"
 #include "gfx_v9_4.h"
 #include "gfx_v9_0.h"
+#include "gfx_v9_0_cleaner_shader.h"
 #include "gfx_v9_4_2.h"
 
 #include "asic_reg/pwr/pwr_10_0_offset.h"
@@ -56,6 +58,7 @@
 #include "asic_reg/gc/gc_9_0_default.h"
 
 #define GFX9_NUM_GFX_RINGS     1
+#define GFX9_NUM_SW_GFX_RINGS  2
 #define GFX9_MEC_HPD_SIZE 4096
 #define RLCG_UCODE_LOADING_START_ADDRESS 0x00002000L
 #define RLC_SAVE_RESTORE_ADDR_STARTING_OFFSET 0x00000000L
@@ -126,6 +129,8 @@ MODULE_FIRMWARE("amdgpu/green_sardine_rlc.bin");
 MODULE_FIRMWARE("amdgpu/aldebaran_mec.bin");
 MODULE_FIRMWARE("amdgpu/aldebaran_mec2.bin");
 MODULE_FIRMWARE("amdgpu/aldebaran_rlc.bin");
+MODULE_FIRMWARE("amdgpu/aldebaran_sjt_mec.bin");
+MODULE_FIRMWARE("amdgpu/aldebaran_sjt_mec2.bin");
 
 #define mmTCP_CHAN_STEER_0_ARCT								0x0b03
 #define mmTCP_CHAN_STEER_0_ARCT_BASE_IDX							0
@@ -145,6 +150,162 @@ MODULE_FIRMWARE("amdgpu/aldebaran_rlc.bin");
 #define mmGOLDEN_TSC_COUNT_LOWER_Renoir                0x0026
 #define mmGOLDEN_TSC_COUNT_LOWER_Renoir_BASE_IDX       1
 
+static const struct amdgpu_hwip_reg_entry gc_reg_list_9[] = {
+	SOC15_REG_ENTRY_STR(GC, 0, mmGRBM_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmGRBM_STATUS2),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_STALLED_STAT1),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_STALLED_STAT2),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CPC_STALLED_STAT1),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CPF_STALLED_STAT1),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_BUSY_STAT),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CPC_BUSY_STAT),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CPF_BUSY_STAT),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CPF_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_GFX_ERROR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_RB_BASE),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_RB_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_RB_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_RB0_BASE),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_RB0_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_RB0_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_RB1_BASE),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_RB1_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_RB1_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_RB2_BASE),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_RB2_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_RB2_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_IB1_CMD_BUFSZ),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_IB2_CMD_BUFSZ),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_IB1_CMD_BUFSZ),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_IB2_CMD_BUFSZ),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_IB1_BASE_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_IB1_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_IB1_BUFSZ),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_IB2_BASE_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_IB2_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_IB2_BUFSZ),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_IB1_BASE_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_IB1_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_IB1_BUFSZ),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_IB2_BASE_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_IB2_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_IB2_BUFSZ),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCPF_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCPC_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCPG_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmGDS_PROTECTION_FAULT),
+	SOC15_REG_ENTRY_STR(GC, 0, mmGDS_VM_PROTECTION_FAULT),
+	SOC15_REG_ENTRY_STR(GC, 0, mmIA_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmIA_UTCL1_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, mmPA_CL_CNTL_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmRLC_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmRMI_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSQC_DCACHE_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSQC_ICACHE_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSQ_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmTCP_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmWD_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmVM_L2_PROTECTION_FAULT_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, mmVM_L2_PROTECTION_FAULT_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_DEBUG),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_MEC_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_INSTR_PNTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_MEC1_INSTR_PNTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_MEC2_INSTR_PNTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_ME_INSTR_PNTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_PFP_INSTR_PNTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CPC_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmRLC_STAT),
+	SOC15_REG_ENTRY_STR(GC, 0, mmRLC_SMU_COMMAND),
+	SOC15_REG_ENTRY_STR(GC, 0, mmRLC_SMU_MESSAGE),
+	SOC15_REG_ENTRY_STR(GC, 0, mmRLC_SMU_ARGUMENT_1),
+	SOC15_REG_ENTRY_STR(GC, 0, mmRLC_SMU_ARGUMENT_2),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSMU_RLC_RESPONSE),
+	SOC15_REG_ENTRY_STR(GC, 0, mmRLC_SAFE_MODE),
+	SOC15_REG_ENTRY_STR(GC, 0, mmRLC_SMU_SAFE_MODE),
+	SOC15_REG_ENTRY_STR(GC, 0, mmRLC_INT_STAT),
+	SOC15_REG_ENTRY_STR(GC, 0, mmRLC_GPM_GENERAL_6),
+	/* SE status registers */
+	SOC15_REG_ENTRY_STR(GC, 0, mmGRBM_STATUS_SE0),
+	SOC15_REG_ENTRY_STR(GC, 0, mmGRBM_STATUS_SE1),
+	SOC15_REG_ENTRY_STR(GC, 0, mmGRBM_STATUS_SE2),
+	SOC15_REG_ENTRY_STR(GC, 0, mmGRBM_STATUS_SE3),
+	/* packet headers */
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_CE_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_PFP_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_PFP_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_PFP_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_PFP_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_PFP_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_PFP_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_PFP_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_PFP_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_ME_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_ME_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_ME_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_ME_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_ME_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_ME_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_ME_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_ME_HEADER_DUMP)
+};
+
+static const struct amdgpu_hwip_reg_entry gc_cp_reg_list_9[] = {
+	/* compute queue registers */
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_VMID),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_ACTIVE),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_PERSISTENT_STATE),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_PIPE_PRIORITY),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_QUEUE_PRIORITY),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_QUANTUM),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_PQ_BASE),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_PQ_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_PQ_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_PQ_WPTR_POLL_ADDR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_PQ_WPTR_POLL_ADDR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_PQ_DOORBELL_CONTROL),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_PQ_CONTROL),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_IB_BASE_ADDR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_IB_BASE_ADDR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_IB_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_IB_CONTROL),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_DEQUEUE_REQUEST),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_EOP_BASE_ADDR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_EOP_BASE_ADDR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_EOP_CONTROL),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_EOP_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_EOP_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_EOP_EVENTS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_CTX_SAVE_BASE_ADDR_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_CTX_SAVE_BASE_ADDR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_CTX_SAVE_CONTROL),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_CNTL_STACK_OFFSET),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_CNTL_STACK_SIZE),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_WG_STATE_OFFSET),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_CTX_SAVE_SIZE),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_GDS_RESOURCE_STATE),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_ERROR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_EOP_WPTR_MEM),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_PQ_WPTR_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_PQ_WPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_GFX_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_MEC_ME1_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_MEC_ME1_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_MEC_ME1_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_MEC_ME1_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_MEC_ME1_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_MEC_ME1_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_MEC_ME1_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, mmCP_MEC_ME1_HEADER_DUMP)
+};
+
 enum ta_ras_gfx_subblock {
 	/*CPC*/
 	TA_RAS_BLOCK__GFX_CPC_INDEX_START = 0,
@@ -739,75 +900,6 @@ static const u32 GFX_RLC_SRM_INDEX_CNTL_DATA_OFFSETS[] =
 	mmRLC_SRM_INDEX_CNTL_DATA_7 - mmRLC_SRM_INDEX_CNTL_DATA_0,
 };
 
-static void gfx_v9_0_rlcg_w(struct amdgpu_device *adev, u32 offset, u32 v, u32 flag)
-{
-	static void *scratch_reg0;
-	static void *scratch_reg1;
-	static void *scratch_reg2;
-	static void *scratch_reg3;
-	static void *spare_int;
-	static uint32_t grbm_cntl;
-	static uint32_t grbm_idx;
-
-	scratch_reg0 = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG0_BASE_IDX] + mmSCRATCH_REG0)*4;
-	scratch_reg1 = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG1_BASE_IDX] + mmSCRATCH_REG1)*4;
-	scratch_reg2 = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG1_BASE_IDX] + mmSCRATCH_REG2)*4;
-	scratch_reg3 = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG1_BASE_IDX] + mmSCRATCH_REG3)*4;
-	spare_int = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmRLC_SPARE_INT_BASE_IDX] + mmRLC_SPARE_INT)*4;
-
-	grbm_cntl = adev->reg_offset[GC_HWIP][0][mmGRBM_GFX_CNTL_BASE_IDX] + mmGRBM_GFX_CNTL;
-	grbm_idx = adev->reg_offset[GC_HWIP][0][mmGRBM_GFX_INDEX_BASE_IDX] + mmGRBM_GFX_INDEX;
-
-	if (amdgpu_sriov_runtime(adev)) {
-		pr_err("shouldn't call rlcg write register during runtime\n");
-		return;
-	}
-
-	if (offset == grbm_cntl || offset == grbm_idx) {
-		if (offset  == grbm_cntl)
-			writel(v, scratch_reg2);
-		else if (offset == grbm_idx)
-			writel(v, scratch_reg3);
-
-		writel(v, ((void __iomem *)adev->rmmio) + (offset * 4));
-	} else {
-		uint32_t i = 0;
-		uint32_t retries = 50000;
-
-		writel(v, scratch_reg0);
-		writel(offset | 0x80000000, scratch_reg1);
-		writel(1, spare_int);
-		for (i = 0; i < retries; i++) {
-			u32 tmp;
-
-			tmp = readl(scratch_reg1);
-			if (!(tmp & 0x80000000))
-				break;
-
-			udelay(10);
-		}
-		if (i >= retries)
-			pr_err("timeout: rlcg program reg:0x%05x failed !\n", offset);
-	}
-
-}
-
-static void gfx_v9_0_sriov_wreg(struct amdgpu_device *adev, u32 offset,
-			       u32 v, u32 acc_flags, u32 hwip)
-{
-	if ((acc_flags & AMDGPU_REGS_RLC) &&
-	    amdgpu_sriov_fullaccess(adev)) {
-		gfx_v9_0_rlcg_w(adev, offset, v, acc_flags);
-
-		return;
-	}
-
-	if (acc_flags & AMDGPU_REGS_NO_KIQ)
-		WREG32_NO_KIQ(offset, v);
-	else
-		WREG32(offset, v);
-}
-
 #define VEGA10_GB_ADDR_CONFIG_GOLDEN 0x2a114042
 #define VEGA12_GB_ADDR_CONFIG_GOLDEN 0x24104041
 #define RAVEN_GB_ADDR_CONFIG_GOLDEN 0x24000042
@@ -820,17 +912,27 @@ static void gfx_v9_0_set_rlc_funcs(struct amdgpu_device *adev);
 static int gfx_v9_0_get_cu_info(struct amdgpu_device *adev,
 				struct amdgpu_cu_info *cu_info);
 static uint64_t gfx_v9_0_get_gpu_clock_counter(struct amdgpu_device *adev);
-static void gfx_v9_0_ring_emit_de_meta(struct amdgpu_ring *ring);
+static void gfx_v9_0_ring_emit_de_meta(struct amdgpu_ring *ring, bool resume, bool usegds);
 static u64 gfx_v9_0_ring_get_rptr_compute(struct amdgpu_ring *ring);
-static int gfx_v9_0_query_ras_error_count(struct amdgpu_device *adev,
+static void gfx_v9_0_query_ras_error_count(struct amdgpu_device *adev,
 					  void *ras_error_status);
 static int gfx_v9_0_ras_error_inject(struct amdgpu_device *adev,
-				     void *inject_if);
+				     void *inject_if, uint32_t instance_mask);
 static void gfx_v9_0_reset_ras_error_count(struct amdgpu_device *adev);
+static void gfx_v9_0_update_spm_vmid_internal(struct amdgpu_device *adev,
+					      unsigned int vmid);
+static void gfx_v9_0_set_safe_mode(struct amdgpu_device *adev, int xcc_id);
+static void gfx_v9_0_unset_safe_mode(struct amdgpu_device *adev, int xcc_id);
 
 static void gfx_v9_0_kiq_set_resources(struct amdgpu_ring *kiq_ring,
 				uint64_t queue_mask)
 {
+	struct amdgpu_device *adev = kiq_ring->adev;
+	u64 shader_mc_addr;
+
+	/* Cleaner shader MC address */
+	shader_mc_addr = adev->gfx.cleaner_shader_gpu_addr >> 8;
+
 	amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_SET_RESOURCES, 6));
 	amdgpu_ring_write(kiq_ring,
 		PACKET3_SET_RESOURCES_VMID_MASK(0) |
@@ -840,8 +942,8 @@ static void gfx_v9_0_kiq_set_resources(struct amdgpu_ring *kiq_ring,
 			lower_32_bits(queue_mask));	/* queue mask lo */
 	amdgpu_ring_write(kiq_ring,
 			upper_32_bits(queue_mask));	/* queue mask hi */
-	amdgpu_ring_write(kiq_ring, 0);	/* gws mask lo */
-	amdgpu_ring_write(kiq_ring, 0);	/* gws mask hi */
+	amdgpu_ring_write(kiq_ring, lower_32_bits(shader_mc_addr)); /* cleaner shader addr lo */
+	amdgpu_ring_write(kiq_ring, upper_32_bits(shader_mc_addr)); /* cleaner shader addr hi */
 	amdgpu_ring_write(kiq_ring, 0);	/* oac mask */
 	amdgpu_ring_write(kiq_ring, 0);	/* gds heap base:0, gds heap size:0 */
 }
@@ -849,9 +951,8 @@ static void gfx_v9_0_kiq_set_resources(struct amdgpu_ring *kiq_ring,
 static void gfx_v9_0_kiq_map_queues(struct amdgpu_ring *kiq_ring,
 				 struct amdgpu_ring *ring)
 {
-	struct amdgpu_device *adev = kiq_ring->adev;
 	uint64_t mqd_addr = amdgpu_bo_gpu_offset(ring->mqd_obj);
-	uint64_t wptr_addr = adev->wb.gpu_addr + (ring->wptr_offs * 4);
+	uint64_t wptr_addr = ring->wptr_gpu_addr;
 	uint32_t eng_sel = ring->funcs->type == AMDGPU_RING_TYPE_GFX ? 4 : 0;
 
 	amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_MAP_QUEUES, 5));
@@ -894,9 +995,10 @@ static void gfx_v9_0_kiq_unmap_queues(struct amdgpu_ring *kiq_ring,
 			PACKET3_UNMAP_QUEUES_DOORBELL_OFFSET0(ring->doorbell_index));
 
 	if (action == PREEMPT_QUEUES_NO_UNMAP) {
-		amdgpu_ring_write(kiq_ring, lower_32_bits(gpu_addr));
-		amdgpu_ring_write(kiq_ring, upper_32_bits(gpu_addr));
-		amdgpu_ring_write(kiq_ring, seq);
+		amdgpu_ring_write(kiq_ring, lower_32_bits(ring->wptr & ring->buf_mask));
+		amdgpu_ring_write(kiq_ring, 0);
+		amdgpu_ring_write(kiq_ring, 0);
+
 	} else {
 		amdgpu_ring_write(kiq_ring, 0);
 		amdgpu_ring_write(kiq_ring, 0);
@@ -938,12 +1040,47 @@ static void gfx_v9_0_kiq_invalidate_tlbs(struct amdgpu_ring *kiq_ring,
 			PACKET3_INVALIDATE_TLBS_FLUSH_TYPE(flush_type));
 }
 
+
+static void gfx_v9_0_kiq_reset_hw_queue(struct amdgpu_ring *kiq_ring, uint32_t queue_type,
+					uint32_t me_id, uint32_t pipe_id, uint32_t queue_id,
+					uint32_t xcc_id, uint32_t vmid)
+{
+	struct amdgpu_device *adev = kiq_ring->adev;
+	unsigned i;
+
+	/* enter save mode */
+	amdgpu_gfx_rlc_enter_safe_mode(adev, xcc_id);
+	mutex_lock(&adev->srbm_mutex);
+	soc15_grbm_select(adev, me_id, pipe_id, queue_id, 0, 0);
+
+	if (queue_type == AMDGPU_RING_TYPE_COMPUTE) {
+		WREG32_SOC15(GC, 0, mmCP_HQD_DEQUEUE_REQUEST, 0x2);
+		WREG32_SOC15(GC, 0, mmSPI_COMPUTE_QUEUE_RESET, 0x1);
+		/* wait till dequeue take effects */
+		for (i = 0; i < adev->usec_timeout; i++) {
+			if (!(RREG32_SOC15(GC, 0, mmCP_HQD_ACTIVE) & 1))
+				break;
+			udelay(1);
+		}
+		if (i >= adev->usec_timeout)
+			dev_err(adev->dev, "fail to wait on hqd deactive\n");
+	} else {
+		dev_err(adev->dev, "reset queue_type(%d) not supported\n", queue_type);
+	}
+
+	soc15_grbm_select(adev, 0, 0, 0, 0, 0);
+	mutex_unlock(&adev->srbm_mutex);
+	/* exit safe mode */
+	amdgpu_gfx_rlc_exit_safe_mode(adev, xcc_id);
+}
+
 static const struct kiq_pm4_funcs gfx_v9_0_kiq_pm4_funcs = {
 	.kiq_set_resources = gfx_v9_0_kiq_set_resources,
 	.kiq_map_queues = gfx_v9_0_kiq_map_queues,
 	.kiq_unmap_queues = gfx_v9_0_kiq_unmap_queues,
 	.kiq_query_status = gfx_v9_0_kiq_query_status,
 	.kiq_invalidate_tlbs = gfx_v9_0_kiq_invalidate_tlbs,
+	.kiq_reset_hw_queue = gfx_v9_0_kiq_reset_hw_queue,
 	.set_resources_size = 8,
 	.map_queues_size = 7,
 	.unmap_queues_size = 6,
@@ -953,12 +1090,12 @@ static const struct kiq_pm4_funcs gfx_v9_0_kiq_pm4_funcs = {
 
 static void gfx_v9_0_set_kiq_pm4_funcs(struct amdgpu_device *adev)
 {
-	adev->gfx.kiq.pmf = &gfx_v9_0_kiq_pm4_funcs;
+	adev->gfx.kiq[0].pmf = &gfx_v9_0_kiq_pm4_funcs;
 }
 
 static void gfx_v9_0_init_golden_registers(struct amdgpu_device *adev)
 {
-	switch (adev->ip_versions[GC_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 	case IP_VERSION(9, 0, 1):
 		soc15_program_register_sequence(adev,
 						golden_settings_gc_9_0,
@@ -1014,19 +1151,12 @@ static void gfx_v9_0_init_golden_registers(struct amdgpu_device *adev)
 		break;
 	}
 
-	if ((adev->ip_versions[GC_HWIP][0] != IP_VERSION(9, 4, 1)) &&
-	    (adev->ip_versions[GC_HWIP][0] != IP_VERSION(9, 4, 2)))
+	if ((amdgpu_ip_version(adev, GC_HWIP, 0) != IP_VERSION(9, 4, 1)) &&
+	    (amdgpu_ip_version(adev, GC_HWIP, 0) != IP_VERSION(9, 4, 2)))
 		soc15_program_register_sequence(adev, golden_settings_gc_9_x_common,
 						(const u32)ARRAY_SIZE(golden_settings_gc_9_x_common));
 }
 
-static void gfx_v9_0_scratch_init(struct amdgpu_device *adev)
-{
-	adev->gfx.scratch.num_reg = 8;
-	adev->gfx.scratch.reg_base = SOC15_REG_OFFSET(GC, 0, mmSCRATCH_REG0);
-	adev->gfx.scratch.free_mask = (1u << adev->gfx.scratch.num_reg) - 1;
-}
-
 static void gfx_v9_0_write_data_to_reg(struct amdgpu_ring *ring, int eng_sel,
 				       bool wc, uint32_t reg, uint32_t val)
 {
@@ -1064,22 +1194,18 @@ static void gfx_v9_0_wait_reg_mem(struct amdgpu_ring *ring, int eng_sel,
 static int gfx_v9_0_ring_test_ring(struct amdgpu_ring *ring)
 {
 	struct amdgpu_device *adev = ring->adev;
-	uint32_t scratch;
+	uint32_t scratch = SOC15_REG_OFFSET(GC, 0, mmSCRATCH_REG0);
 	uint32_t tmp = 0;
 	unsigned i;
 	int r;
 
-	r = amdgpu_gfx_scratch_get(adev, &scratch);
-	if (r)
-		return r;
-
 	WREG32(scratch, 0xCAFEDEAD);
 	r = amdgpu_ring_alloc(ring, 3);
 	if (r)
-		goto error_free_scratch;
+		return r;
 
 	amdgpu_ring_write(ring, PACKET3(PACKET3_SET_UCONFIG_REG, 1));
-	amdgpu_ring_write(ring, (scratch - PACKET3_SET_UCONFIG_REG_START));
+	amdgpu_ring_write(ring, scratch - PACKET3_SET_UCONFIG_REG_START);
 	amdgpu_ring_write(ring, 0xDEADBEEF);
 	amdgpu_ring_commit(ring);
 
@@ -1092,9 +1218,6 @@ static int gfx_v9_0_ring_test_ring(struct amdgpu_ring *ring)
 
 	if (i >= adev->usec_timeout)
 		r = -ETIMEDOUT;
-
-error_free_scratch:
-	amdgpu_gfx_scratch_free(adev, scratch);
 	return r;
 }
 
@@ -1116,8 +1239,8 @@ static int gfx_v9_0_ring_test_ib(struct amdgpu_ring *ring, long timeout)
 	gpu_addr = adev->wb.gpu_addr + (index * 4);
 	adev->wb.wb[index] = cpu_to_le32(0xCAFEDEAD);
 	memset(&ib, 0, sizeof(ib));
-	r = amdgpu_ib_get(adev, NULL, 16,
-					AMDGPU_IB_POOL_DIRECT, &ib);
+
+	r = amdgpu_ib_get(adev, NULL, 20, AMDGPU_IB_POOL_DIRECT, &ib);
 	if (r)
 		goto err1;
 
@@ -1147,7 +1270,7 @@ static int gfx_v9_0_ring_test_ib(struct amdgpu_ring *ring, long timeout)
 		r = -EINVAL;
 
 err2:
-	amdgpu_ib_free(adev, &ib, NULL);
+	amdgpu_ib_free(&ib, NULL);
 	dma_fence_put(f);
 err1:
 	amdgpu_device_wb_free(adev, index);
@@ -1157,56 +1280,30 @@ err1:
 
 static void gfx_v9_0_free_microcode(struct amdgpu_device *adev)
 {
-	release_firmware(adev->gfx.pfp_fw);
-	adev->gfx.pfp_fw = NULL;
-	release_firmware(adev->gfx.me_fw);
-	adev->gfx.me_fw = NULL;
-	release_firmware(adev->gfx.ce_fw);
-	adev->gfx.ce_fw = NULL;
-	release_firmware(adev->gfx.rlc_fw);
-	adev->gfx.rlc_fw = NULL;
-	release_firmware(adev->gfx.mec_fw);
-	adev->gfx.mec_fw = NULL;
-	release_firmware(adev->gfx.mec2_fw);
-	adev->gfx.mec2_fw = NULL;
+	amdgpu_ucode_release(&adev->gfx.pfp_fw);
+	amdgpu_ucode_release(&adev->gfx.me_fw);
+	amdgpu_ucode_release(&adev->gfx.ce_fw);
+	amdgpu_ucode_release(&adev->gfx.rlc_fw);
+	amdgpu_ucode_release(&adev->gfx.mec_fw);
+	amdgpu_ucode_release(&adev->gfx.mec2_fw);
 
 	kfree(adev->gfx.rlc.register_list_format);
 }
 
-static void gfx_v9_0_init_rlc_ext_microcode(struct amdgpu_device *adev)
-{
-	const struct rlc_firmware_header_v2_1 *rlc_hdr;
-
-	rlc_hdr = (const struct rlc_firmware_header_v2_1 *)adev->gfx.rlc_fw->data;
-	adev->gfx.rlc_srlc_fw_version = le32_to_cpu(rlc_hdr->save_restore_list_cntl_ucode_ver);
-	adev->gfx.rlc_srlc_feature_version = le32_to_cpu(rlc_hdr->save_restore_list_cntl_feature_ver);
-	adev->gfx.rlc.save_restore_list_cntl_size_bytes = le32_to_cpu(rlc_hdr->save_restore_list_cntl_size_bytes);
-	adev->gfx.rlc.save_restore_list_cntl = (u8 *)rlc_hdr + le32_to_cpu(rlc_hdr->save_restore_list_cntl_offset_bytes);
-	adev->gfx.rlc_srlg_fw_version = le32_to_cpu(rlc_hdr->save_restore_list_gpm_ucode_ver);
-	adev->gfx.rlc_srlg_feature_version = le32_to_cpu(rlc_hdr->save_restore_list_gpm_feature_ver);
-	adev->gfx.rlc.save_restore_list_gpm_size_bytes = le32_to_cpu(rlc_hdr->save_restore_list_gpm_size_bytes);
-	adev->gfx.rlc.save_restore_list_gpm = (u8 *)rlc_hdr + le32_to_cpu(rlc_hdr->save_restore_list_gpm_offset_bytes);
-	adev->gfx.rlc_srls_fw_version = le32_to_cpu(rlc_hdr->save_restore_list_srm_ucode_ver);
-	adev->gfx.rlc_srls_feature_version = le32_to_cpu(rlc_hdr->save_restore_list_srm_feature_ver);
-	adev->gfx.rlc.save_restore_list_srm_size_bytes = le32_to_cpu(rlc_hdr->save_restore_list_srm_size_bytes);
-	adev->gfx.rlc.save_restore_list_srm = (u8 *)rlc_hdr + le32_to_cpu(rlc_hdr->save_restore_list_srm_offset_bytes);
-	adev->gfx.rlc.reg_list_format_direct_reg_list_length =
-			le32_to_cpu(rlc_hdr->reg_list_format_direct_reg_list_length);
-}
-
 static void gfx_v9_0_check_fw_write_wait(struct amdgpu_device *adev)
 {
 	adev->gfx.me_fw_write_wait = false;
 	adev->gfx.mec_fw_write_wait = false;
 
-	if ((adev->ip_versions[GC_HWIP][0] != IP_VERSION(9, 4, 1)) &&
+	if ((amdgpu_ip_version(adev, GC_HWIP, 0) != IP_VERSION(9, 4, 1)) &&
+	    (amdgpu_ip_version(adev, GC_HWIP, 0) != IP_VERSION(9, 4, 2)) &&
 	    ((adev->gfx.mec_fw_version < 0x000001a5) ||
-	    (adev->gfx.mec_feature_version < 46) ||
-	    (adev->gfx.pfp_fw_version < 0x000000b7) ||
-	    (adev->gfx.pfp_feature_version < 46)))
+	     (adev->gfx.mec_feature_version < 46) ||
+	     (adev->gfx.pfp_fw_version < 0x000000b7) ||
+	     (adev->gfx.pfp_feature_version < 46)))
 		DRM_WARN_ONCE("CP firmware version too old, please update!");
 
-	switch (adev->ip_versions[GC_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 	case IP_VERSION(9, 0, 1):
 		if ((adev->gfx.me_fw_version >= 0x0000009c) &&
 		    (adev->gfx.me_feature_version >= 42) &&
@@ -1274,6 +1371,12 @@ static const struct amdgpu_gfxoff_quirk amdgpu_gfxoff_quirk_list[] = {
 	{ 0x1002, 0x15dd, 0x103c, 0x83e7, 0xd3 },
 	/* GFXOFF is unstable on C6 parts with a VBIOS 113-RAVEN-114 */
 	{ 0x1002, 0x15dd, 0x1002, 0x15dd, 0xc6 },
+	/* Apple MacBook Pro (15-inch, 2019) Radeon Pro Vega 20 4 GB */
+	{ 0x1002, 0x69af, 0x106b, 0x019a, 0xc0 },
+	/* https://bbs.openkylin.top/t/topic/171497 */
+	{ 0x1002, 0x15d8, 0x19e5, 0x3e14, 0xc2 },
+	/* HP 705G4 DM with R5 2400G */
+	{ 0x1002, 0x15dd, 0x103c, 0x8464, 0xd6 },
 	{ 0, 0, 0, 0, 0 },
 };
 
@@ -1304,7 +1407,7 @@ static bool is_raven_kicker(struct amdgpu_device *adev)
 
 static bool check_if_enlarge_doorbell_range(struct amdgpu_device *adev)
 {
-	if ((adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 3, 0)) &&
+	if ((amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 3, 0)) &&
 	    (adev->gfx.me_fw_version >= 0x000000a5) &&
 	    (adev->gfx.me_feature_version >= 52))
 		return true;
@@ -1317,7 +1420,7 @@ static void gfx_v9_0_check_if_need_gfxoff(struct amdgpu_device *adev)
 	if (gfx_v9_0_should_disable_gfxoff(adev->pdev))
 		adev->pm.pp_feature &= ~PP_GFXOFF_MASK;
 
-	switch (adev->ip_versions[GC_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 	case IP_VERSION(9, 0, 1):
 	case IP_VERSION(9, 2, 1):
 	case IP_VERSION(9, 4, 0):
@@ -1349,95 +1452,45 @@ static void gfx_v9_0_check_if_need_gfxoff(struct amdgpu_device *adev)
 }
 
 static int gfx_v9_0_init_cp_gfx_microcode(struct amdgpu_device *adev,
-					  const char *chip_name)
+					  char *chip_name)
 {
-	char fw_name[30];
 	int err;
-	struct amdgpu_firmware_info *info = NULL;
-	const struct common_firmware_header *header = NULL;
-	const struct gfx_firmware_header_v1_0 *cp_hdr;
 
-	snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_pfp.bin", chip_name);
-	err = request_firmware(&adev->gfx.pfp_fw, fw_name, adev->dev);
-	if (err)
-		goto out;
-	err = amdgpu_ucode_validate(adev->gfx.pfp_fw);
+	err = amdgpu_ucode_request(adev, &adev->gfx.pfp_fw,
+				   AMDGPU_UCODE_REQUIRED,
+				   "amdgpu/%s_pfp.bin", chip_name);
 	if (err)
 		goto out;
-	cp_hdr = (const struct gfx_firmware_header_v1_0 *)adev->gfx.pfp_fw->data;
-	adev->gfx.pfp_fw_version = le32_to_cpu(cp_hdr->header.ucode_version);
-	adev->gfx.pfp_feature_version = le32_to_cpu(cp_hdr->ucode_feature_version);
+	amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_PFP);
 
-	snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_me.bin", chip_name);
-	err = request_firmware(&adev->gfx.me_fw, fw_name, adev->dev);
-	if (err)
-		goto out;
-	err = amdgpu_ucode_validate(adev->gfx.me_fw);
+	err = amdgpu_ucode_request(adev, &adev->gfx.me_fw,
+				   AMDGPU_UCODE_REQUIRED,
+				   "amdgpu/%s_me.bin", chip_name);
 	if (err)
 		goto out;
-	cp_hdr = (const struct gfx_firmware_header_v1_0 *)adev->gfx.me_fw->data;
-	adev->gfx.me_fw_version = le32_to_cpu(cp_hdr->header.ucode_version);
-	adev->gfx.me_feature_version = le32_to_cpu(cp_hdr->ucode_feature_version);
+	amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_ME);
 
-	snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_ce.bin", chip_name);
-	err = request_firmware(&adev->gfx.ce_fw, fw_name, adev->dev);
+	err = amdgpu_ucode_request(adev, &adev->gfx.ce_fw,
+				   AMDGPU_UCODE_REQUIRED,
+				   "amdgpu/%s_ce.bin", chip_name);
 	if (err)
 		goto out;
-	err = amdgpu_ucode_validate(adev->gfx.ce_fw);
-	if (err)
-		goto out;
-	cp_hdr = (const struct gfx_firmware_header_v1_0 *)adev->gfx.ce_fw->data;
-	adev->gfx.ce_fw_version = le32_to_cpu(cp_hdr->header.ucode_version);
-	adev->gfx.ce_feature_version = le32_to_cpu(cp_hdr->ucode_feature_version);
-
-	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
-		info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_PFP];
-		info->ucode_id = AMDGPU_UCODE_ID_CP_PFP;
-		info->fw = adev->gfx.pfp_fw;
-		header = (const struct common_firmware_header *)info->fw->data;
-		adev->firmware.fw_size +=
-			ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE);
-
-		info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_ME];
-		info->ucode_id = AMDGPU_UCODE_ID_CP_ME;
-		info->fw = adev->gfx.me_fw;
-		header = (const struct common_firmware_header *)info->fw->data;
-		adev->firmware.fw_size +=
-			ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE);
-
-		info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_CE];
-		info->ucode_id = AMDGPU_UCODE_ID_CP_CE;
-		info->fw = adev->gfx.ce_fw;
-		header = (const struct common_firmware_header *)info->fw->data;
-		adev->firmware.fw_size +=
-			ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE);
-	}
+	amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_CE);
 
 out:
 	if (err) {
-		dev_err(adev->dev,
-			"gfx9: Failed to load firmware \"%s\"\n",
-			fw_name);
-		release_firmware(adev->gfx.pfp_fw);
-		adev->gfx.pfp_fw = NULL;
-		release_firmware(adev->gfx.me_fw);
-		adev->gfx.me_fw = NULL;
-		release_firmware(adev->gfx.ce_fw);
-		adev->gfx.ce_fw = NULL;
+		amdgpu_ucode_release(&adev->gfx.pfp_fw);
+		amdgpu_ucode_release(&adev->gfx.me_fw);
+		amdgpu_ucode_release(&adev->gfx.ce_fw);
 	}
 	return err;
 }
 
 static int gfx_v9_0_init_rlc_microcode(struct amdgpu_device *adev,
-					  const char *chip_name)
+				       char *chip_name)
 {
-	char fw_name[30];
 	int err;
-	struct amdgpu_firmware_info *info = NULL;
-	const struct common_firmware_header *header = NULL;
 	const struct rlc_firmware_header_v2_0 *rlc_hdr;
-	unsigned int *tmp = NULL;
-	unsigned int i = 0;
 	uint16_t version_major;
 	uint16_t version_minor;
 	uint32_t smu_version;
@@ -1453,271 +1506,114 @@ static int gfx_v9_0_init_rlc_microcode(struct amdgpu_device *adev,
 	if (!strcmp(chip_name, "picasso") &&
 		(((adev->pdev->revision >= 0xC8) && (adev->pdev->revision <= 0xCF)) ||
 		((adev->pdev->revision >= 0xD8) && (adev->pdev->revision <= 0xDF))))
-		snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_rlc_am4.bin", chip_name);
+		err = amdgpu_ucode_request(adev, &adev->gfx.rlc_fw,
+					   AMDGPU_UCODE_REQUIRED,
+					   "amdgpu/%s_rlc_am4.bin", chip_name);
 	else if (!strcmp(chip_name, "raven") && (amdgpu_pm_load_smu_firmware(adev, &smu_version) == 0) &&
 		(smu_version >= 0x41e2b))
 		/**
 		*SMC is loaded by SBIOS on APU and it's able to get the SMU version directly.
 		*/
-		snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_kicker_rlc.bin", chip_name);
+		err = amdgpu_ucode_request(adev, &adev->gfx.rlc_fw,
+					   AMDGPU_UCODE_REQUIRED,
+					   "amdgpu/%s_kicker_rlc.bin", chip_name);
 	else
-		snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_rlc.bin", chip_name);
-	err = request_firmware(&adev->gfx.rlc_fw, fw_name, adev->dev);
+		err = amdgpu_ucode_request(adev, &adev->gfx.rlc_fw,
+					   AMDGPU_UCODE_REQUIRED,
+					   "amdgpu/%s_rlc.bin", chip_name);
 	if (err)
 		goto out;
-	err = amdgpu_ucode_validate(adev->gfx.rlc_fw);
-	rlc_hdr = (const struct rlc_firmware_header_v2_0 *)adev->gfx.rlc_fw->data;
 
+	rlc_hdr = (const struct rlc_firmware_header_v2_0 *)adev->gfx.rlc_fw->data;
 	version_major = le16_to_cpu(rlc_hdr->header.header_version_major);
 	version_minor = le16_to_cpu(rlc_hdr->header.header_version_minor);
-	if (version_major == 2 && version_minor == 1)
-		adev->gfx.rlc.is_rlc_v2_1 = true;
-
-	adev->gfx.rlc_fw_version = le32_to_cpu(rlc_hdr->header.ucode_version);
-	adev->gfx.rlc_feature_version = le32_to_cpu(rlc_hdr->ucode_feature_version);
-	adev->gfx.rlc.save_and_restore_offset =
-			le32_to_cpu(rlc_hdr->save_and_restore_offset);
-	adev->gfx.rlc.clear_state_descriptor_offset =
-			le32_to_cpu(rlc_hdr->clear_state_descriptor_offset);
-	adev->gfx.rlc.avail_scratch_ram_locations =
-			le32_to_cpu(rlc_hdr->avail_scratch_ram_locations);
-	adev->gfx.rlc.reg_restore_list_size =
-			le32_to_cpu(rlc_hdr->reg_restore_list_size);
-	adev->gfx.rlc.reg_list_format_start =
-			le32_to_cpu(rlc_hdr->reg_list_format_start);
-	adev->gfx.rlc.reg_list_format_separate_start =
-			le32_to_cpu(rlc_hdr->reg_list_format_separate_start);
-	adev->gfx.rlc.starting_offsets_start =
-			le32_to_cpu(rlc_hdr->starting_offsets_start);
-	adev->gfx.rlc.reg_list_format_size_bytes =
-			le32_to_cpu(rlc_hdr->reg_list_format_size_bytes);
-	adev->gfx.rlc.reg_list_size_bytes =
-			le32_to_cpu(rlc_hdr->reg_list_size_bytes);
-	adev->gfx.rlc.register_list_format =
-			kmalloc(adev->gfx.rlc.reg_list_format_size_bytes +
-				adev->gfx.rlc.reg_list_size_bytes, GFP_KERNEL);
-	if (!adev->gfx.rlc.register_list_format) {
-		err = -ENOMEM;
-		goto out;
-	}
-
-	tmp = (unsigned int *)((uintptr_t)rlc_hdr +
-			le32_to_cpu(rlc_hdr->reg_list_format_array_offset_bytes));
-	for (i = 0 ; i < (adev->gfx.rlc.reg_list_format_size_bytes >> 2); i++)
-		adev->gfx.rlc.register_list_format[i] =	le32_to_cpu(tmp[i]);
-
-	adev->gfx.rlc.register_restore = adev->gfx.rlc.register_list_format + i;
-
-	tmp = (unsigned int *)((uintptr_t)rlc_hdr +
-			le32_to_cpu(rlc_hdr->reg_list_array_offset_bytes));
-	for (i = 0 ; i < (adev->gfx.rlc.reg_list_size_bytes >> 2); i++)
-		adev->gfx.rlc.register_restore[i] = le32_to_cpu(tmp[i]);
-
-	if (adev->gfx.rlc.is_rlc_v2_1)
-		gfx_v9_0_init_rlc_ext_microcode(adev);
-
-	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
-		info = &adev->firmware.ucode[AMDGPU_UCODE_ID_RLC_G];
-		info->ucode_id = AMDGPU_UCODE_ID_RLC_G;
-		info->fw = adev->gfx.rlc_fw;
-		header = (const struct common_firmware_header *)info->fw->data;
-		adev->firmware.fw_size +=
-			ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE);
-
-		if (adev->gfx.rlc.is_rlc_v2_1 &&
-		    adev->gfx.rlc.save_restore_list_cntl_size_bytes &&
-		    adev->gfx.rlc.save_restore_list_gpm_size_bytes &&
-		    adev->gfx.rlc.save_restore_list_srm_size_bytes) {
-			info = &adev->firmware.ucode[AMDGPU_UCODE_ID_RLC_RESTORE_LIST_CNTL];
-			info->ucode_id = AMDGPU_UCODE_ID_RLC_RESTORE_LIST_CNTL;
-			info->fw = adev->gfx.rlc_fw;
-			adev->firmware.fw_size +=
-				ALIGN(adev->gfx.rlc.save_restore_list_cntl_size_bytes, PAGE_SIZE);
-
-			info = &adev->firmware.ucode[AMDGPU_UCODE_ID_RLC_RESTORE_LIST_GPM_MEM];
-			info->ucode_id = AMDGPU_UCODE_ID_RLC_RESTORE_LIST_GPM_MEM;
-			info->fw = adev->gfx.rlc_fw;
-			adev->firmware.fw_size +=
-				ALIGN(adev->gfx.rlc.save_restore_list_gpm_size_bytes, PAGE_SIZE);
-
-			info = &adev->firmware.ucode[AMDGPU_UCODE_ID_RLC_RESTORE_LIST_SRM_MEM];
-			info->ucode_id = AMDGPU_UCODE_ID_RLC_RESTORE_LIST_SRM_MEM;
-			info->fw = adev->gfx.rlc_fw;
-			adev->firmware.fw_size +=
-				ALIGN(adev->gfx.rlc.save_restore_list_srm_size_bytes, PAGE_SIZE);
-		}
-	}
-
+	err = amdgpu_gfx_rlc_init_microcode(adev, version_major, version_minor);
 out:
-	if (err) {
-		dev_err(adev->dev,
-			"gfx9: Failed to load firmware \"%s\"\n",
-			fw_name);
-		release_firmware(adev->gfx.rlc_fw);
-		adev->gfx.rlc_fw = NULL;
-	}
+	if (err)
+		amdgpu_ucode_release(&adev->gfx.rlc_fw);
+
 	return err;
 }
 
 static bool gfx_v9_0_load_mec2_fw_bin_support(struct amdgpu_device *adev)
 {
-	if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 2) ||
-	    adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 1) ||
-	    adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 3, 0))
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 2) ||
+	    amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 1) ||
+	    amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 3, 0))
 		return false;
 
 	return true;
 }
 
 static int gfx_v9_0_init_cp_compute_microcode(struct amdgpu_device *adev,
-					  const char *chip_name)
+					      char *chip_name)
 {
-	char fw_name[30];
 	int err;
-	struct amdgpu_firmware_info *info = NULL;
-	const struct common_firmware_header *header = NULL;
-	const struct gfx_firmware_header_v1_0 *cp_hdr;
 
-	snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_mec.bin", chip_name);
-	err = request_firmware(&adev->gfx.mec_fw, fw_name, adev->dev);
-	if (err)
-		goto out;
-	err = amdgpu_ucode_validate(adev->gfx.mec_fw);
+	if (amdgpu_sriov_vf(adev) && (adev->asic_type == CHIP_ALDEBARAN))
+		err = amdgpu_ucode_request(adev, &adev->gfx.mec_fw,
+				   AMDGPU_UCODE_REQUIRED,
+				   "amdgpu/%s_sjt_mec.bin", chip_name);
+	else
+		err = amdgpu_ucode_request(adev, &adev->gfx.mec_fw,
+					   AMDGPU_UCODE_REQUIRED,
+					   "amdgpu/%s_mec.bin", chip_name);
 	if (err)
 		goto out;
-	cp_hdr = (const struct gfx_firmware_header_v1_0 *)adev->gfx.mec_fw->data;
-	adev->gfx.mec_fw_version = le32_to_cpu(cp_hdr->header.ucode_version);
-	adev->gfx.mec_feature_version = le32_to_cpu(cp_hdr->ucode_feature_version);
 
+	amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_MEC1);
+	amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_MEC1_JT);
 
 	if (gfx_v9_0_load_mec2_fw_bin_support(adev)) {
-		snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_mec2.bin", chip_name);
-		err = request_firmware(&adev->gfx.mec2_fw, fw_name, adev->dev);
+		if (amdgpu_sriov_vf(adev) && (adev->asic_type == CHIP_ALDEBARAN))
+			err = amdgpu_ucode_request(adev, &adev->gfx.mec2_fw,
+						   AMDGPU_UCODE_REQUIRED,
+						   "amdgpu/%s_sjt_mec2.bin", chip_name);
+		else
+			err = amdgpu_ucode_request(adev, &adev->gfx.mec2_fw,
+						   AMDGPU_UCODE_REQUIRED,
+						   "amdgpu/%s_mec2.bin", chip_name);
 		if (!err) {
-			err = amdgpu_ucode_validate(adev->gfx.mec2_fw);
-			if (err)
-				goto out;
-			cp_hdr = (const struct gfx_firmware_header_v1_0 *)
-			adev->gfx.mec2_fw->data;
-			adev->gfx.mec2_fw_version =
-			le32_to_cpu(cp_hdr->header.ucode_version);
-			adev->gfx.mec2_feature_version =
-			le32_to_cpu(cp_hdr->ucode_feature_version);
+			amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_MEC2);
+			amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_MEC2_JT);
 		} else {
 			err = 0;
-			adev->gfx.mec2_fw = NULL;
+			amdgpu_ucode_release(&adev->gfx.mec2_fw);
 		}
 	} else {
 		adev->gfx.mec2_fw_version = adev->gfx.mec_fw_version;
 		adev->gfx.mec2_feature_version = adev->gfx.mec_feature_version;
 	}
 
-	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
-		info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_MEC1];
-		info->ucode_id = AMDGPU_UCODE_ID_CP_MEC1;
-		info->fw = adev->gfx.mec_fw;
-		header = (const struct common_firmware_header *)info->fw->data;
-		cp_hdr = (const struct gfx_firmware_header_v1_0 *)info->fw->data;
-		adev->firmware.fw_size +=
-			ALIGN(le32_to_cpu(header->ucode_size_bytes) - le32_to_cpu(cp_hdr->jt_size) * 4, PAGE_SIZE);
-
-		info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_MEC1_JT];
-		info->ucode_id = AMDGPU_UCODE_ID_CP_MEC1_JT;
-		info->fw = adev->gfx.mec_fw;
-		adev->firmware.fw_size +=
-			ALIGN(le32_to_cpu(cp_hdr->jt_size) * 4, PAGE_SIZE);
-
-		if (adev->gfx.mec2_fw) {
-			info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_MEC2];
-			info->ucode_id = AMDGPU_UCODE_ID_CP_MEC2;
-			info->fw = adev->gfx.mec2_fw;
-			header = (const struct common_firmware_header *)info->fw->data;
-			cp_hdr = (const struct gfx_firmware_header_v1_0 *)info->fw->data;
-			adev->firmware.fw_size +=
-				ALIGN(le32_to_cpu(header->ucode_size_bytes) - le32_to_cpu(cp_hdr->jt_size) * 4, PAGE_SIZE);
-
-			/* TODO: Determine if MEC2 JT FW loading can be removed
-				 for all GFX V9 asic and above */
-			if (gfx_v9_0_load_mec2_fw_bin_support(adev)) {
-				info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_MEC2_JT];
-				info->ucode_id = AMDGPU_UCODE_ID_CP_MEC2_JT;
-				info->fw = adev->gfx.mec2_fw;
-				adev->firmware.fw_size +=
-					ALIGN(le32_to_cpu(cp_hdr->jt_size) * 4,
-					PAGE_SIZE);
-			}
-		}
-	}
-
-out:
 	gfx_v9_0_check_if_need_gfxoff(adev);
 	gfx_v9_0_check_fw_write_wait(adev);
-	if (err) {
-		dev_err(adev->dev,
-			"gfx9: Failed to load firmware \"%s\"\n",
-			fw_name);
-		release_firmware(adev->gfx.mec_fw);
-		adev->gfx.mec_fw = NULL;
-		release_firmware(adev->gfx.mec2_fw);
-		adev->gfx.mec2_fw = NULL;
-	}
+
+out:
+	if (err)
+		amdgpu_ucode_release(&adev->gfx.mec_fw);
 	return err;
 }
 
 static int gfx_v9_0_init_microcode(struct amdgpu_device *adev)
 {
-	const char *chip_name;
+	char ucode_prefix[30];
 	int r;
 
 	DRM_DEBUG("\n");
-
-	switch (adev->ip_versions[GC_HWIP][0]) {
-	case IP_VERSION(9, 0, 1):
-		chip_name = "vega10";
-		break;
-	case IP_VERSION(9, 2, 1):
-		chip_name = "vega12";
-		break;
-	case IP_VERSION(9, 4, 0):
-		chip_name = "vega20";
-		break;
-	case IP_VERSION(9, 2, 2):
-	case IP_VERSION(9, 1, 0):
-		if (adev->apu_flags & AMD_APU_IS_RAVEN2)
-			chip_name = "raven2";
-		else if (adev->apu_flags & AMD_APU_IS_PICASSO)
-			chip_name = "picasso";
-		else
-			chip_name = "raven";
-		break;
-	case IP_VERSION(9, 4, 1):
-		chip_name = "arcturus";
-		break;
-	case IP_VERSION(9, 3, 0):
-		if (adev->apu_flags & AMD_APU_IS_RENOIR)
-			chip_name = "renoir";
-		else
-			chip_name = "green_sardine";
-		break;
-	case IP_VERSION(9, 4, 2):
-		chip_name = "aldebaran";
-		break;
-	default:
-		BUG();
-	}
+	amdgpu_ucode_ip_version_decode(adev, GC_HWIP, ucode_prefix, sizeof(ucode_prefix));
 
 	/* No CPG in Arcturus */
 	if (adev->gfx.num_gfx_rings) {
-		r = gfx_v9_0_init_cp_gfx_microcode(adev, chip_name);
+		r = gfx_v9_0_init_cp_gfx_microcode(adev, ucode_prefix);
 		if (r)
 			return r;
 	}
 
-	r = gfx_v9_0_init_rlc_microcode(adev, chip_name);
+	r = gfx_v9_0_init_rlc_microcode(adev, ucode_prefix);
 	if (r)
 		return r;
 
-	r = gfx_v9_0_init_cp_compute_microcode(adev, chip_name);
+	r = gfx_v9_0_init_cp_compute_microcode(adev, ucode_prefix);
 	if (r)
 		return r;
 
@@ -1752,45 +1648,18 @@ static u32 gfx_v9_0_get_csb_size(struct amdgpu_device *adev)
 	return count;
 }
 
-static void gfx_v9_0_get_csb_buffer(struct amdgpu_device *adev,
-				    volatile u32 *buffer)
+static void gfx_v9_0_get_csb_buffer(struct amdgpu_device *adev, u32 *buffer)
 {
-	u32 count = 0, i;
-	const struct cs_section_def *sect = NULL;
-	const struct cs_extent_def *ext = NULL;
+	u32 count = 0;
 
 	if (adev->gfx.rlc.cs_data == NULL)
 		return;
 	if (buffer == NULL)
 		return;
 
-	buffer[count++] = cpu_to_le32(PACKET3(PACKET3_PREAMBLE_CNTL, 0));
-	buffer[count++] = cpu_to_le32(PACKET3_PREAMBLE_BEGIN_CLEAR_STATE);
-
-	buffer[count++] = cpu_to_le32(PACKET3(PACKET3_CONTEXT_CONTROL, 1));
-	buffer[count++] = cpu_to_le32(0x80000000);
-	buffer[count++] = cpu_to_le32(0x80000000);
-
-	for (sect = adev->gfx.rlc.cs_data; sect->section != NULL; ++sect) {
-		for (ext = sect->section; ext->extent != NULL; ++ext) {
-			if (sect->id == SECT_CONTEXT) {
-				buffer[count++] =
-					cpu_to_le32(PACKET3(PACKET3_SET_CONTEXT_REG, ext->reg_count));
-				buffer[count++] = cpu_to_le32(ext->reg_index -
-						PACKET3_SET_CONTEXT_REG_START);
-				for (i = 0; i < ext->reg_count; i++)
-					buffer[count++] = cpu_to_le32(ext->extent[i]);
-			} else {
-				return;
-			}
-		}
-	}
-
-	buffer[count++] = cpu_to_le32(PACKET3(PACKET3_PREAMBLE_CNTL, 0));
-	buffer[count++] = cpu_to_le32(PACKET3_PREAMBLE_END_CLEAR_STATE);
-
-	buffer[count++] = cpu_to_le32(PACKET3(PACKET3_CLEAR_STATE, 0));
-	buffer[count++] = cpu_to_le32(0);
+	count = amdgpu_gfx_csb_preamble_start(buffer);
+	count = amdgpu_gfx_csb_data_parser(adev, buffer, count);
+	amdgpu_gfx_csb_preamble_end(buffer, count);
 }
 
 static void gfx_v9_0_init_always_on_cu_mask(struct amdgpu_device *adev)
@@ -1803,7 +1672,7 @@ static void gfx_v9_0_init_always_on_cu_mask(struct amdgpu_device *adev)
 
 	if (adev->flags & AMD_IS_APU)
 		always_on_cu_num = 4;
-	else if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 2, 1))
+	else if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 2, 1))
 		always_on_cu_num = 8;
 	else
 		always_on_cu_num = 12;
@@ -1814,10 +1683,10 @@ static void gfx_v9_0_init_always_on_cu_mask(struct amdgpu_device *adev)
 			mask = 1;
 			cu_bitmap = 0;
 			counter = 0;
-			gfx_v9_0_select_se_sh(adev, i, j, 0xffffffff);
+			amdgpu_gfx_select_se_sh(adev, i, j, 0xffffffff, 0);
 
 			for (k = 0; k < adev->gfx.config.max_cu_per_sh; k ++) {
-				if (cu_info->bitmap[i][j] & mask) {
+				if (cu_info->bitmap[0][i][j] & mask) {
 					if (counter == pg_always_on_cu_num)
 						WREG32_SOC15(GC, 0, mmRLC_PG_ALWAYS_ON_CU_MASK, cu_bitmap);
 					if (counter < always_on_cu_num)
@@ -1833,7 +1702,7 @@ static void gfx_v9_0_init_always_on_cu_mask(struct amdgpu_device *adev)
 			cu_info->ao_cu_bitmap[i][j] = cu_bitmap;
 		}
 	}
-	gfx_v9_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+	amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
 	mutex_unlock(&adev->grbm_idx_mutex);
 }
 
@@ -1855,7 +1724,7 @@ static void gfx_v9_0_init_lbpw(struct amdgpu_device *adev)
 
 	mutex_lock(&adev->grbm_idx_mutex);
 	/* set mmRLC_LB_INIT_CU_MASK thru broadcast mode to enable all SE/SH*/
-	gfx_v9_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+	amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
 	WREG32_SOC15(GC, 0, mmRLC_LB_INIT_CU_MASK, 0xffffffff);
 
 	/* set mmRLC_LB_PARAMS = 0x003F_1006 */
@@ -1904,7 +1773,7 @@ static void gfx_v9_4_init_lbpw(struct amdgpu_device *adev)
 
 	mutex_lock(&adev->grbm_idx_mutex);
 	/* set mmRLC_LB_INIT_CU_MASK thru broadcast mode to enable all SE/SH*/
-	gfx_v9_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+	amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
 	WREG32_SOC15(GC, 0, mmRLC_LB_INIT_CU_MASK, 0xffffffff);
 
 	/* set mmRLC_LB_PARAMS = 0x003F_1006 */
@@ -1948,6 +1817,21 @@ static int gfx_v9_0_cp_jump_table_num(struct amdgpu_device *adev)
 		return 4;
 }
 
+static void gfx_v9_0_init_rlcg_reg_access_ctrl(struct amdgpu_device *adev)
+{
+	struct amdgpu_rlcg_reg_access_ctrl *reg_access_ctrl;
+
+	reg_access_ctrl = &adev->gfx.rlc.reg_access_ctrl[0];
+	reg_access_ctrl->scratch_reg0 = SOC15_REG_OFFSET(GC, 0, mmSCRATCH_REG0);
+	reg_access_ctrl->scratch_reg1 = SOC15_REG_OFFSET(GC, 0, mmSCRATCH_REG1);
+	reg_access_ctrl->scratch_reg2 = SOC15_REG_OFFSET(GC, 0, mmSCRATCH_REG2);
+	reg_access_ctrl->scratch_reg3 = SOC15_REG_OFFSET(GC, 0, mmSCRATCH_REG3);
+	reg_access_ctrl->grbm_cntl = SOC15_REG_OFFSET(GC, 0, mmGRBM_GFX_CNTL);
+	reg_access_ctrl->grbm_idx = SOC15_REG_OFFSET(GC, 0, mmGRBM_GFX_INDEX);
+	reg_access_ctrl->spare_int = SOC15_REG_OFFSET(GC, 0, mmRLC_SPARE_INT);
+	adev->gfx.rlc.rlcg_reg_access_supported = true;
+}
+
 static int gfx_v9_0_rlc_init(struct amdgpu_device *adev)
 {
 	const struct cs_section_def *cs_data;
@@ -1972,22 +1856,6 @@ static int gfx_v9_0_rlc_init(struct amdgpu_device *adev)
 			return r;
 	}
 
-	switch (adev->ip_versions[GC_HWIP][0]) {
-	case IP_VERSION(9, 2, 2):
-	case IP_VERSION(9, 1, 0):
-		gfx_v9_0_init_lbpw(adev);
-		break;
-	case IP_VERSION(9, 4, 0):
-		gfx_v9_4_init_lbpw(adev);
-		break;
-	default:
-		break;
-	}
-
-	/* init spm vmid with 0xf */
-	if (adev->gfx.rlc.funcs->update_spm_vmid)
-		adev->gfx.rlc.funcs->update_spm_vmid(adev, 0xf);
-
 	return 0;
 }
 
@@ -2008,14 +1876,15 @@ static int gfx_v9_0_mec_init(struct amdgpu_device *adev)
 
 	const struct gfx_firmware_header_v1_0 *mec_hdr;
 
-	bitmap_zero(adev->gfx.mec.queue_bitmap, AMDGPU_MAX_COMPUTE_QUEUES);
+	bitmap_zero(adev->gfx.mec_bitmap[0].queue_bitmap, AMDGPU_MAX_COMPUTE_QUEUES);
 
 	/* take ownership of the relevant compute queues */
 	amdgpu_gfx_compute_queue_acquire(adev);
 	mec_hpd_size = adev->gfx.num_compute_rings * GFX9_MEC_HPD_SIZE;
 	if (mec_hpd_size) {
 		r = amdgpu_bo_create_reserved(adev, mec_hpd_size, PAGE_SIZE,
-					      AMDGPU_GEM_DOMAIN_VRAM,
+					      AMDGPU_GEM_DOMAIN_VRAM |
+					      AMDGPU_GEM_DOMAIN_GTT,
 					      &adev->gfx.mec.hpd_eop_obj,
 					      &adev->gfx.mec.hpd_eop_gpu_addr,
 					      (void **)&hpd);
@@ -2082,7 +1951,7 @@ static void wave_read_regs(struct amdgpu_device *adev, uint32_t simd,
 		*(out++) = RREG32_SOC15(GC, 0, mmSQ_IND_DATA);
 }
 
-static void gfx_v9_0_read_wave_data(struct amdgpu_device *adev, uint32_t simd, uint32_t wave, uint32_t *dst, int *no_fields)
+static void gfx_v9_0_read_wave_data(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd, uint32_t wave, uint32_t *dst, int *no_fields)
 {
 	/* type 1 wave data */
 	dst[(*no_fields)++] = 1;
@@ -2103,7 +1972,7 @@ static void gfx_v9_0_read_wave_data(struct amdgpu_device *adev, uint32_t simd, u
 	dst[(*no_fields)++] = wave_read_ind(adev, simd, wave, ixSQ_WAVE_MODE);
 }
 
-static void gfx_v9_0_read_wave_sgprs(struct amdgpu_device *adev, uint32_t simd,
+static void gfx_v9_0_read_wave_sgprs(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd,
 				     uint32_t wave, uint32_t start,
 				     uint32_t size, uint32_t *dst)
 {
@@ -2112,7 +1981,7 @@ static void gfx_v9_0_read_wave_sgprs(struct amdgpu_device *adev, uint32_t simd,
 		start + SQIND_WAVE_SGPRS_OFFSET, size, dst);
 }
 
-static void gfx_v9_0_read_wave_vgprs(struct amdgpu_device *adev, uint32_t simd,
+static void gfx_v9_0_read_wave_vgprs(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd,
 				     uint32_t wave, uint32_t thread,
 				     uint32_t start, uint32_t size,
 				     uint32_t *dst)
@@ -2123,9 +1992,9 @@ static void gfx_v9_0_read_wave_vgprs(struct amdgpu_device *adev, uint32_t simd,
 }
 
 static void gfx_v9_0_select_me_pipe_q(struct amdgpu_device *adev,
-				  u32 me, u32 pipe, u32 q, u32 vm)
+				  u32 me, u32 pipe, u32 q, u32 vm, u32 xcc_id)
 {
-	soc15_grbm_select(adev, me, pipe, q, vm);
+	soc15_grbm_select(adev, me, pipe, q, vm, 0);
 }
 
 static const struct amdgpu_gfx_funcs gfx_v9_0_gfx_funcs = {
@@ -2137,12 +2006,16 @@ static const struct amdgpu_gfx_funcs gfx_v9_0_gfx_funcs = {
         .select_me_pipe_q = &gfx_v9_0_select_me_pipe_q,
 };
 
-static const struct amdgpu_gfx_ras_funcs gfx_v9_0_ras_funcs = {
-	.ras_late_init = amdgpu_gfx_ras_late_init,
-	.ras_fini = amdgpu_gfx_ras_fini,
-	.ras_error_inject = &gfx_v9_0_ras_error_inject,
-	.query_ras_error_count = &gfx_v9_0_query_ras_error_count,
-	.reset_ras_error_count = &gfx_v9_0_reset_ras_error_count,
+const struct amdgpu_ras_block_hw_ops  gfx_v9_0_ras_ops = {
+		.ras_error_inject = &gfx_v9_0_ras_error_inject,
+		.query_ras_error_count = &gfx_v9_0_query_ras_error_count,
+		.reset_ras_error_count = &gfx_v9_0_reset_ras_error_count,
+};
+
+static struct amdgpu_gfx_ras gfx_v9_0_ras = {
+	.ras_block = {
+		.hw_ops = &gfx_v9_0_ras_ops,
+	},
 };
 
 static int gfx_v9_0_gpu_early_init(struct amdgpu_device *adev)
@@ -2150,9 +2023,7 @@ static int gfx_v9_0_gpu_early_init(struct amdgpu_device *adev)
 	u32 gb_addr_config;
 	int err;
 
-	adev->gfx.funcs = &gfx_v9_0_gfx_funcs;
-
-	switch (adev->ip_versions[GC_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 	case IP_VERSION(9, 0, 1):
 		adev->gfx.config.max_hw_contexts = 8;
 		adev->gfx.config.sc_prim_fifo_size_frontend = 0x20;
@@ -2171,7 +2042,7 @@ static int gfx_v9_0_gpu_early_init(struct amdgpu_device *adev)
 		DRM_INFO("fix gfx.config for vega12\n");
 		break;
 	case IP_VERSION(9, 4, 0):
-		adev->gfx.ras_funcs = &gfx_v9_0_ras_funcs;
+		adev->gfx.ras = &gfx_v9_0_ras;
 		adev->gfx.config.max_hw_contexts = 8;
 		adev->gfx.config.sc_prim_fifo_size_frontend = 0x20;
 		adev->gfx.config.sc_prim_fifo_size_backend = 0x100;
@@ -2198,7 +2069,7 @@ static int gfx_v9_0_gpu_early_init(struct amdgpu_device *adev)
 			gb_addr_config = RAVEN_GB_ADDR_CONFIG_GOLDEN;
 		break;
 	case IP_VERSION(9, 4, 1):
-		adev->gfx.ras_funcs = &gfx_v9_4_ras_funcs;
+		adev->gfx.ras = &gfx_v9_4_ras;
 		adev->gfx.config.max_hw_contexts = 8;
 		adev->gfx.config.sc_prim_fifo_size_frontend = 0x20;
 		adev->gfx.config.sc_prim_fifo_size_backend = 0x100;
@@ -2219,7 +2090,7 @@ static int gfx_v9_0_gpu_early_init(struct amdgpu_device *adev)
 		gb_addr_config |= 0x22010042;
 		break;
 	case IP_VERSION(9, 4, 2):
-		adev->gfx.ras_funcs = &gfx_v9_4_2_ras_funcs;
+		adev->gfx.ras = &gfx_v9_4_2_ras;
 		adev->gfx.config.max_hw_contexts = 8;
 		adev->gfx.config.sc_prim_fifo_size_frontend = 0x20;
 		adev->gfx.config.sc_prim_fifo_size_backend = 0x100;
@@ -2297,26 +2168,56 @@ static int gfx_v9_0_compute_ring_init(struct amdgpu_device *adev, int ring_id,
 	ring->doorbell_index = (adev->doorbell_index.mec_ring0 + ring_id) << 1;
 	ring->eop_gpu_addr = adev->gfx.mec.hpd_eop_gpu_addr
 				+ (ring_id * GFX9_MEC_HPD_SIZE);
+	ring->vm_hub = AMDGPU_GFXHUB(0);
 	sprintf(ring->name, "comp_%d.%d.%d", ring->me, ring->pipe, ring->queue);
 
 	irq_type = AMDGPU_CP_IRQ_COMPUTE_MEC1_PIPE0_EOP
 		+ ((ring->me - 1) * adev->gfx.mec.num_pipe_per_mec)
 		+ ring->pipe;
 	hw_prio = amdgpu_gfx_is_high_priority_compute_queue(adev, ring) ?
-			AMDGPU_GFX_PIPE_PRIO_HIGH : AMDGPU_GFX_PIPE_PRIO_NORMAL;
+			AMDGPU_RING_PRIO_2 : AMDGPU_RING_PRIO_DEFAULT;
 	/* type-2 packets are deprecated on MEC, use type-3 instead */
 	return amdgpu_ring_init(adev, ring, 1024, &adev->gfx.eop_irq, irq_type,
 				hw_prio, NULL);
 }
 
-static int gfx_v9_0_sw_init(void *handle)
+static void gfx_v9_0_alloc_ip_dump(struct amdgpu_device *adev)
+{
+	uint32_t reg_count = ARRAY_SIZE(gc_reg_list_9);
+	uint32_t *ptr;
+	uint32_t inst;
+
+	ptr = kcalloc(reg_count, sizeof(uint32_t), GFP_KERNEL);
+	if (!ptr) {
+		DRM_ERROR("Failed to allocate memory for GFX IP Dump\n");
+		adev->gfx.ip_dump_core = NULL;
+	} else {
+		adev->gfx.ip_dump_core = ptr;
+	}
+
+	/* Allocate memory for compute queue registers for all the instances */
+	reg_count = ARRAY_SIZE(gc_cp_reg_list_9);
+	inst = adev->gfx.mec.num_mec * adev->gfx.mec.num_pipe_per_mec *
+		adev->gfx.mec.num_queue_per_pipe;
+
+	ptr = kcalloc(reg_count * inst, sizeof(uint32_t), GFP_KERNEL);
+	if (!ptr) {
+		DRM_ERROR("Failed to allocate memory for Compute Queues IP Dump\n");
+		adev->gfx.ip_dump_compute_queues = NULL;
+	} else {
+		adev->gfx.ip_dump_compute_queues = ptr;
+	}
+}
+
+static int gfx_v9_0_sw_init(struct amdgpu_ip_block *ip_block)
 {
 	int i, j, k, r, ring_id;
+	int xcc_id = 0;
 	struct amdgpu_ring *ring;
-	struct amdgpu_kiq *kiq;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
+	unsigned int hw_prio;
 
-	switch (adev->ip_versions[GC_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 	case IP_VERSION(9, 0, 1):
 	case IP_VERSION(9, 2, 1):
 	case IP_VERSION(9, 4, 0):
@@ -2332,6 +2233,43 @@ static int gfx_v9_0_sw_init(void *handle)
 		break;
 	}
 
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	case IP_VERSION(9, 0, 1):
+	case IP_VERSION(9, 2, 1):
+	case IP_VERSION(9, 4, 0):
+	case IP_VERSION(9, 2, 2):
+	case IP_VERSION(9, 1, 0):
+	case IP_VERSION(9, 3, 0):
+		adev->gfx.cleaner_shader_ptr = gfx_9_4_2_cleaner_shader_hex;
+		adev->gfx.cleaner_shader_size = sizeof(gfx_9_4_2_cleaner_shader_hex);
+		if (adev->gfx.me_fw_version  >= 167 &&
+		    adev->gfx.pfp_fw_version >= 196 &&
+		    adev->gfx.mec_fw_version >= 474) {
+			adev->gfx.enable_cleaner_shader = true;
+			r = amdgpu_gfx_cleaner_shader_sw_init(adev, adev->gfx.cleaner_shader_size);
+			if (r) {
+				adev->gfx.enable_cleaner_shader = false;
+				dev_err(adev->dev, "Failed to initialize cleaner shader\n");
+			}
+		}
+		break;
+	case IP_VERSION(9, 4, 2):
+		adev->gfx.cleaner_shader_ptr = gfx_9_4_2_cleaner_shader_hex;
+		adev->gfx.cleaner_shader_size = sizeof(gfx_9_4_2_cleaner_shader_hex);
+		if (adev->gfx.mec_fw_version >= 88) {
+			adev->gfx.enable_cleaner_shader = true;
+			r = amdgpu_gfx_cleaner_shader_sw_init(adev, adev->gfx.cleaner_shader_size);
+			if (r) {
+				adev->gfx.enable_cleaner_shader = false;
+				dev_err(adev->dev, "Failed to initialize cleaner shader\n");
+			}
+		}
+		break;
+	default:
+		adev->gfx.enable_cleaner_shader = false;
+		break;
+	}
+
 	adev->gfx.mec.num_pipe_per_mec = 4;
 	adev->gfx.mec.num_queue_per_pipe = 8;
 
@@ -2340,6 +2278,13 @@ static int gfx_v9_0_sw_init(void *handle)
 	if (r)
 		return r;
 
+	/* Bad opcode Event */
+	r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_GRBM_CP,
+			      GFX_9_0__SRCID__CP_BAD_OPCODE_ERROR,
+			      &adev->gfx.bad_op_irq);
+	if (r)
+		return r;
+
 	/* Privileged reg */
 	r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_GRBM_CP, GFX_9_0__SRCID__CP_PRIV_REG_FAULT,
 			      &adev->gfx.priv_reg_irq);
@@ -2366,18 +2311,14 @@ static int gfx_v9_0_sw_init(void *handle)
 
 	adev->gfx.gfx_current_status = AMDGPU_GFX_NORMAL_MODE;
 
-	gfx_v9_0_scratch_init(adev);
-
-	r = gfx_v9_0_init_microcode(adev);
-	if (r) {
-		DRM_ERROR("Failed to load gfx firmware!\n");
-		return r;
-	}
-
-	r = adev->gfx.rlc.funcs->init(adev);
-	if (r) {
-		DRM_ERROR("Failed to init rlc BOs!\n");
-		return r;
+	if (adev->gfx.rlc.funcs) {
+		if (adev->gfx.rlc.funcs->init) {
+			r = adev->gfx.rlc.funcs->init(adev);
+			if (r) {
+				dev_err(adev->dev, "Failed to init rlc BOs!\n");
+				return r;
+			}
+		}
 	}
 
 	r = gfx_v9_0_mec_init(adev);
@@ -2396,6 +2337,10 @@ static int gfx_v9_0_sw_init(void *handle)
 			sprintf(ring->name, "gfx_%d", i);
 		ring->use_doorbell = true;
 		ring->doorbell_index = adev->doorbell_index.gfx_ring0 << 1;
+
+		/* disable scheduler on the real ring */
+		ring->no_scheduler = adev->gfx.mcbp;
+		ring->vm_hub = AMDGPU_GFXHUB(0);
 		r = amdgpu_ring_init(adev, ring, 1024, &adev->gfx.eop_irq,
 				     AMDGPU_CP_IRQ_GFX_ME0_PIPE0_EOP,
 				     AMDGPU_RING_PRIO_DEFAULT, NULL);
@@ -2403,12 +2348,49 @@ static int gfx_v9_0_sw_init(void *handle)
 			return r;
 	}
 
+	/* set up the software rings */
+	if (adev->gfx.mcbp && adev->gfx.num_gfx_rings) {
+		for (i = 0; i < GFX9_NUM_SW_GFX_RINGS; i++) {
+			ring = &adev->gfx.sw_gfx_ring[i];
+			ring->ring_obj = NULL;
+			sprintf(ring->name, amdgpu_sw_ring_name(i));
+			ring->use_doorbell = true;
+			ring->doorbell_index = adev->doorbell_index.gfx_ring0 << 1;
+			ring->is_sw_ring = true;
+			hw_prio = amdgpu_sw_ring_priority(i);
+			ring->vm_hub = AMDGPU_GFXHUB(0);
+			r = amdgpu_ring_init(adev, ring, 1024, &adev->gfx.eop_irq,
+					     AMDGPU_CP_IRQ_GFX_ME0_PIPE0_EOP, hw_prio,
+					     NULL);
+			if (r)
+				return r;
+			ring->wptr = 0;
+		}
+
+		/* init the muxer and add software rings */
+		r = amdgpu_ring_mux_init(&adev->gfx.muxer, &adev->gfx.gfx_ring[0],
+					 GFX9_NUM_SW_GFX_RINGS);
+		if (r) {
+			DRM_ERROR("amdgpu_ring_mux_init failed(%d)\n", r);
+			return r;
+		}
+		for (i = 0; i < GFX9_NUM_SW_GFX_RINGS; i++) {
+			r = amdgpu_ring_mux_add_sw_ring(&adev->gfx.muxer,
+							&adev->gfx.sw_gfx_ring[i]);
+			if (r) {
+				DRM_ERROR("amdgpu_ring_mux_add_sw_ring failed(%d)\n", r);
+				return r;
+			}
+		}
+	}
+
 	/* set up the compute queues - allocate horizontally across pipes */
 	ring_id = 0;
 	for (i = 0; i < adev->gfx.mec.num_mec; ++i) {
 		for (j = 0; j < adev->gfx.mec.num_queue_per_pipe; j++) {
 			for (k = 0; k < adev->gfx.mec.num_pipe_per_mec; k++) {
-				if (!amdgpu_gfx_is_mec_queue_enabled(adev, i, k, j))
+				if (!amdgpu_gfx_is_mec_queue_enabled(adev, 0, i,
+								     k, j))
 					continue;
 
 				r = gfx_v9_0_compute_ring_init(adev,
@@ -2422,19 +2404,26 @@ static int gfx_v9_0_sw_init(void *handle)
 		}
 	}
 
-	r = amdgpu_gfx_kiq_init(adev, GFX9_MEC_HPD_SIZE);
+	/* TODO: Add queue reset mask when FW fully supports it */
+	adev->gfx.gfx_supported_reset =
+		amdgpu_get_soft_full_reset_mask(&adev->gfx.gfx_ring[0]);
+	adev->gfx.compute_supported_reset =
+		amdgpu_get_soft_full_reset_mask(&adev->gfx.compute_ring[0]);
+	if (!amdgpu_sriov_vf(adev) && !adev->debug_disable_gpu_ring_reset)
+		adev->gfx.compute_supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE;
+
+	r = amdgpu_gfx_kiq_init(adev, GFX9_MEC_HPD_SIZE, 0);
 	if (r) {
 		DRM_ERROR("Failed to init KIQ BOs!\n");
 		return r;
 	}
 
-	kiq = &adev->gfx.kiq;
-	r = amdgpu_gfx_kiq_init_ring(adev, &kiq->ring, &kiq->irq);
+	r = amdgpu_gfx_kiq_init_ring(adev, xcc_id);
 	if (r)
 		return r;
 
 	/* create MQD for all compute queues as wel as KIQ for SRIOV case */
-	r = amdgpu_gfx_mqd_sw_init(adev, sizeof(struct v9_mqd_allocation));
+	r = amdgpu_gfx_mqd_sw_init(adev, sizeof(struct v9_mqd_allocation), 0);
 	if (r)
 		return r;
 
@@ -2444,27 +2433,42 @@ static int gfx_v9_0_sw_init(void *handle)
 	if (r)
 		return r;
 
+	if (amdgpu_gfx_ras_sw_init(adev)) {
+		dev_err(adev->dev, "Failed to initialize gfx ras block!\n");
+		return -EINVAL;
+	}
+
+	gfx_v9_0_alloc_ip_dump(adev);
+
+	r = amdgpu_gfx_sysfs_init(adev);
+	if (r)
+		return r;
+
 	return 0;
 }
 
 
-static int gfx_v9_0_sw_fini(void *handle)
+static int gfx_v9_0_sw_fini(struct amdgpu_ip_block *ip_block)
 {
 	int i;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
-	if (adev->gfx.ras_funcs &&
-	    adev->gfx.ras_funcs->ras_fini)
-		adev->gfx.ras_funcs->ras_fini(adev);
+	if (adev->gfx.mcbp && adev->gfx.num_gfx_rings) {
+		for (i = 0; i < GFX9_NUM_SW_GFX_RINGS; i++)
+			amdgpu_ring_fini(&adev->gfx.sw_gfx_ring[i]);
+		amdgpu_ring_mux_fini(&adev->gfx.muxer);
+	}
 
 	for (i = 0; i < adev->gfx.num_gfx_rings; i++)
 		amdgpu_ring_fini(&adev->gfx.gfx_ring[i]);
 	for (i = 0; i < adev->gfx.num_compute_rings; i++)
 		amdgpu_ring_fini(&adev->gfx.compute_ring[i]);
 
-	amdgpu_gfx_mqd_sw_fini(adev);
-	amdgpu_gfx_kiq_free_ring(&adev->gfx.kiq.ring);
-	amdgpu_gfx_kiq_fini(adev);
+	amdgpu_gfx_mqd_sw_fini(adev, 0);
+	amdgpu_gfx_kiq_free_ring(&adev->gfx.kiq[0].ring);
+	amdgpu_gfx_kiq_fini(adev, 0);
+
+	amdgpu_gfx_cleaner_shader_sw_fini(adev);
 
 	gfx_v9_0_mec_fini(adev);
 	amdgpu_bo_free_kernel(&adev->gfx.rlc.clear_state_obj,
@@ -2477,6 +2481,11 @@ static int gfx_v9_0_sw_fini(void *handle)
 	}
 	gfx_v9_0_free_microcode(adev);
 
+	amdgpu_gfx_sysfs_fini(adev);
+
+	kfree(adev->gfx.ip_dump_core);
+	kfree(adev->gfx.ip_dump_compute_queues);
+
 	return 0;
 }
 
@@ -2487,7 +2496,7 @@ static void gfx_v9_0_tiling_mode_table_init(struct amdgpu_device *adev)
 }
 
 void gfx_v9_0_select_se_sh(struct amdgpu_device *adev, u32 se_num, u32 sh_num,
-			   u32 instance)
+			   u32 instance, int xcc_id)
 {
 	u32 data;
 
@@ -2536,19 +2545,42 @@ static void gfx_v9_0_setup_rb(struct amdgpu_device *adev)
 	mutex_lock(&adev->grbm_idx_mutex);
 	for (i = 0; i < adev->gfx.config.max_shader_engines; i++) {
 		for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) {
-			gfx_v9_0_select_se_sh(adev, i, j, 0xffffffff);
+			amdgpu_gfx_select_se_sh(adev, i, j, 0xffffffff, 0);
 			data = gfx_v9_0_get_rb_active_bitmap(adev);
 			active_rbs |= data << ((i * adev->gfx.config.max_sh_per_se + j) *
 					       rb_bitmap_width_per_sh);
 		}
 	}
-	gfx_v9_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+	amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
 	mutex_unlock(&adev->grbm_idx_mutex);
 
 	adev->gfx.config.backend_enable_mask = active_rbs;
 	adev->gfx.config.num_rbs = hweight32(active_rbs);
 }
 
+static void gfx_v9_0_debug_trap_config_init(struct amdgpu_device *adev,
+				uint32_t first_vmid,
+				uint32_t last_vmid)
+{
+	uint32_t data;
+	uint32_t trap_config_vmid_mask = 0;
+	int i;
+
+	/* Calculate trap config vmid mask */
+	for (i = first_vmid; i < last_vmid; i++)
+		trap_config_vmid_mask |= (1 << i);
+
+	data = REG_SET_FIELD(0, SPI_GDBG_TRAP_CONFIG,
+			VMID_SEL, trap_config_vmid_mask);
+	data = REG_SET_FIELD(data, SPI_GDBG_TRAP_CONFIG,
+			TRAP_EN, 1);
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_CONFIG), data);
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_MASK), 0);
+
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_DATA0), 0);
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_DATA1), 0);
+}
+
 #define DEFAULT_SH_MEM_BASES	(0x6000)
 static void gfx_v9_0_init_compute_vmid(struct amdgpu_device *adev)
 {
@@ -2570,16 +2602,16 @@ static void gfx_v9_0_init_compute_vmid(struct amdgpu_device *adev)
 
 	mutex_lock(&adev->srbm_mutex);
 	for (i = adev->vm_manager.first_kfd_vmid; i < AMDGPU_NUM_VMID; i++) {
-		soc15_grbm_select(adev, 0, 0, 0, i);
+		soc15_grbm_select(adev, 0, 0, 0, i, 0);
 		/* CP and shaders */
 		WREG32_SOC15_RLC(GC, 0, mmSH_MEM_CONFIG, sh_mem_config);
 		WREG32_SOC15_RLC(GC, 0, mmSH_MEM_BASES, sh_mem_bases);
 	}
-	soc15_grbm_select(adev, 0, 0, 0, 0);
+	soc15_grbm_select(adev, 0, 0, 0, 0, 0);
 	mutex_unlock(&adev->srbm_mutex);
 
 	/* Initialize all compute VMIDs to have no GDS, GWS, or OA
-	   acccess. These should be enabled by FW for target VMIDs. */
+	   access. These should be enabled by FW for target VMIDs. */
 	for (i = adev->vm_manager.first_kfd_vmid; i < AMDGPU_NUM_VMID; i++) {
 		WREG32_SOC15_OFFSET(GC, 0, mmGDS_VMID0_BASE, 2 * i, 0);
 		WREG32_SOC15_OFFSET(GC, 0, mmGDS_VMID0_SIZE, 2 * i, 0);
@@ -2610,13 +2642,16 @@ static void gfx_v9_0_init_sq_config(struct amdgpu_device *adev)
 {
 	uint32_t tmp;
 
-	switch (adev->ip_versions[GC_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 	case IP_VERSION(9, 4, 1):
 		tmp = RREG32_SOC15(GC, 0, mmSQ_CONFIG);
-		tmp = REG_SET_FIELD(tmp, SQ_CONFIG,
-					DISABLE_BARRIER_WAITCNT, 1);
+		tmp = REG_SET_FIELD(tmp, SQ_CONFIG, DISABLE_BARRIER_WAITCNT,
+				!READ_ONCE(adev->barrier_has_auto_waitcnt));
 		WREG32_SOC15(GC, 0, mmSQ_CONFIG, tmp);
 		break;
+	case IP_VERSION(9, 4, 2):
+		gfx_v9_4_2_init_sq(adev);
+		break;
 	default:
 		break;
 	}
@@ -2627,19 +2662,23 @@ static void gfx_v9_0_constants_init(struct amdgpu_device *adev)
 	u32 tmp;
 	int i;
 
-	WREG32_FIELD15_RLC(GC, 0, GRBM_CNTL, READ_TIMEOUT, 0xff);
+	if (!amdgpu_sriov_vf(adev) ||
+	    amdgpu_ip_version(adev, GC_HWIP, 0) != IP_VERSION(9, 4, 2)) {
+		WREG32_FIELD15_RLC(GC, 0, GRBM_CNTL, READ_TIMEOUT, 0xff);
+	}
 
 	gfx_v9_0_tiling_mode_table_init(adev);
 
-	gfx_v9_0_setup_rb(adev);
+	if (adev->gfx.num_gfx_rings)
+		gfx_v9_0_setup_rb(adev);
 	gfx_v9_0_get_cu_info(adev, &adev->gfx.cu_info);
 	adev->gfx.config.db_debug2 = RREG32_SOC15(GC, 0, mmDB_DEBUG2);
 
 	/* XXX SH_MEM regs */
 	/* where to put LDS, scratch, GPUVM in FSA64 space */
 	mutex_lock(&adev->srbm_mutex);
-	for (i = 0; i < adev->vm_manager.id_mgr[AMDGPU_GFXHUB_0].num_ids; i++) {
-		soc15_grbm_select(adev, 0, 0, 0, i);
+	for (i = 0; i < adev->vm_manager.id_mgr[AMDGPU_GFXHUB(0)].num_ids; i++) {
+		soc15_grbm_select(adev, 0, 0, 0, i, 0);
 		/* CP and shaders */
 		if (i == 0) {
 			tmp = REG_SET_FIELD(0, SH_MEM_CONFIG, ALIGNMENT_MODE,
@@ -2661,7 +2700,7 @@ static void gfx_v9_0_constants_init(struct amdgpu_device *adev)
 			WREG32_SOC15_RLC(GC, 0, mmSH_MEM_BASES, tmp);
 		}
 	}
-	soc15_grbm_select(adev, 0, 0, 0, 0);
+	soc15_grbm_select(adev, 0, 0, 0, 0, 0);
 
 	mutex_unlock(&adev->srbm_mutex);
 
@@ -2678,15 +2717,15 @@ static void gfx_v9_0_wait_for_rlc_serdes(struct amdgpu_device *adev)
 	mutex_lock(&adev->grbm_idx_mutex);
 	for (i = 0; i < adev->gfx.config.max_shader_engines; i++) {
 		for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) {
-			gfx_v9_0_select_se_sh(adev, i, j, 0xffffffff);
+			amdgpu_gfx_select_se_sh(adev, i, j, 0xffffffff, 0);
 			for (k = 0; k < adev->usec_timeout; k++) {
 				if (RREG32_SOC15(GC, 0, mmRLC_SERDES_CU_MASTER_BUSY) == 0)
 					break;
 				udelay(1);
 			}
 			if (k == adev->usec_timeout) {
-				gfx_v9_0_select_se_sh(adev, 0xffffffff,
-						      0xffffffff, 0xffffffff);
+				amdgpu_gfx_select_se_sh(adev, 0xffffffff,
+						      0xffffffff, 0xffffffff, 0);
 				mutex_unlock(&adev->grbm_idx_mutex);
 				DRM_INFO("Timeout wait for RLC serdes %u,%u\n",
 					 i, j);
@@ -2694,7 +2733,7 @@ static void gfx_v9_0_wait_for_rlc_serdes(struct amdgpu_device *adev)
 			}
 		}
 	}
-	gfx_v9_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+	amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
 	mutex_unlock(&adev->grbm_idx_mutex);
 
 	mask = RLC_SERDES_NONCU_MASTER_BUSY__SE_MASTER_BUSY_MASK |
@@ -2720,7 +2759,7 @@ static void gfx_v9_0_enable_gui_idle_interrupt(struct amdgpu_device *adev,
 	tmp = REG_SET_FIELD(tmp, CP_INT_CNTL_RING0, CNTX_BUSY_INT_ENABLE, enable ? 1 : 0);
 	tmp = REG_SET_FIELD(tmp, CP_INT_CNTL_RING0, CNTX_EMPTY_INT_ENABLE, enable ? 1 : 0);
 	tmp = REG_SET_FIELD(tmp, CP_INT_CNTL_RING0, CMP_BUSY_INT_ENABLE, enable ? 1 : 0);
-	if(adev->gfx.num_gfx_rings)
+	if (adev->gfx.num_gfx_rings)
 		tmp = REG_SET_FIELD(tmp, CP_INT_CNTL_RING0, GFX_IDLE_INT_ENABLE, enable ? 1 : 0);
 
 	WREG32_SOC15(GC, 0, mmCP_INT_CNTL_RING0, tmp);
@@ -2946,7 +2985,7 @@ static void gfx_v9_0_init_gfx_power_gating(struct amdgpu_device *adev)
 		/* program GRBM_REG_SAVE_GFX_IDLE_THRESHOLD to 0x55f0 */
 		data |= (0x55f0 << RLC_AUTO_PG_CTRL__GRBM_REG_SAVE_GFX_IDLE_THRESHOLD__SHIFT);
 		WREG32(SOC15_REG_OFFSET(GC, 0, mmRLC_AUTO_PG_CTRL), data);
-		if (adev->ip_versions[GC_HWIP][0] != IP_VERSION(9, 3, 0))
+		if (amdgpu_ip_version(adev, GC_HWIP, 0) != IP_VERSION(9, 3, 0))
 			pwr_10_0_gfxip_control_over_cgpg(adev, true);
 	}
 }
@@ -3058,7 +3097,8 @@ static void gfx_v9_0_init_pg(struct amdgpu_device *adev)
 	 * And it's needed by gfxoff feature.
 	 */
 	if (adev->gfx.rlc.is_rlc_v2_1) {
-		if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 2, 1) ||
+		if (amdgpu_ip_version(adev, GC_HWIP, 0) ==
+			    IP_VERSION(9, 2, 1) ||
 		    (adev->apu_flags & AMD_APU_IS_RAVEN2))
 			gfx_v9_1_init_rlc_save_restore_list(adev);
 		gfx_v9_0_enable_save_restore_machine(adev);
@@ -3171,15 +3211,17 @@ static int gfx_v9_0_rlc_resume(struct amdgpu_device *adev)
 			return r;
 	}
 
-	switch (adev->ip_versions[GC_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 	case IP_VERSION(9, 2, 2):
 	case IP_VERSION(9, 1, 0):
+		gfx_v9_0_init_lbpw(adev);
 		if (amdgpu_lbpw == 0)
 			gfx_v9_0_enable_lbpw(adev, false);
 		else
 			gfx_v9_0_enable_lbpw(adev, true);
 		break;
 	case IP_VERSION(9, 4, 0):
+		gfx_v9_4_init_lbpw(adev);
 		if (amdgpu_lbpw > 0)
 			gfx_v9_0_enable_lbpw(adev, true);
 		else
@@ -3189,6 +3231,8 @@ static int gfx_v9_0_rlc_resume(struct amdgpu_device *adev)
 		break;
 	}
 
+	gfx_v9_0_update_spm_vmid_internal(adev, 0xf);
+
 	adev->gfx.rlc.funcs->start(adev);
 
 	return 0;
@@ -3198,6 +3242,15 @@ static void gfx_v9_0_cp_gfx_enable(struct amdgpu_device *adev, bool enable)
 {
 	u32 tmp = RREG32_SOC15(GC, 0, mmCP_ME_CNTL);
 
+	tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, CE_INVALIDATE_ICACHE, enable ? 0 : 1);
+	tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, PFP_INVALIDATE_ICACHE, enable ? 0 : 1);
+	tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, ME_INVALIDATE_ICACHE, enable ? 0 : 1);
+	tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, CE_PIPE0_RESET, enable ? 0 : 1);
+	tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, CE_PIPE1_RESET, enable ? 0 : 1);
+	tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, PFP_PIPE0_RESET, enable ? 0 : 1);
+	tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, PFP_PIPE1_RESET, enable ? 0 : 1);
+	tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, ME_PIPE0_RESET, enable ? 0 : 1);
+	tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, ME_PIPE1_RESET, enable ? 0 : 1);
 	tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, ME_HALT, enable ? 0 : 1);
 	tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, PFP_HALT, enable ? 0 : 1);
 	tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, CE_HALT, enable ? 0 : 1);
@@ -3275,6 +3328,14 @@ static int gfx_v9_0_cp_gfx_start(struct amdgpu_device *adev)
 
 	gfx_v9_0_cp_gfx_enable(adev, true);
 
+	/* Now only limit the quirk on the APU gfx9 series and already
+	 * confirmed that the APU gfx10/gfx11 needn't such update.
+	 */
+	if (adev->flags & AMD_IS_APU &&
+			adev->in_s3 && !pm_resume_via_firmware()) {
+		DRM_INFO("Will skip the CSB packet resubmit\n");
+		return 0;
+	}
 	r = amdgpu_ring_alloc(ring, gfx_v9_0_get_csb_size(adev) + 4 + 3);
 	if (r) {
 		DRM_ERROR("amdgpu: cp failed to lock ring (%d).\n", r);
@@ -3352,12 +3413,12 @@ static int gfx_v9_0_cp_gfx_resume(struct amdgpu_device *adev)
 	WREG32_SOC15(GC, 0, mmCP_RB0_WPTR, lower_32_bits(ring->wptr));
 	WREG32_SOC15(GC, 0, mmCP_RB0_WPTR_HI, upper_32_bits(ring->wptr));
 
-	/* set the wb address wether it's enabled or not */
-	rptr_addr = adev->wb.gpu_addr + (ring->rptr_offs * 4);
+	/* set the wb address whether it's enabled or not */
+	rptr_addr = ring->rptr_gpu_addr;
 	WREG32_SOC15(GC, 0, mmCP_RB0_RPTR_ADDR, lower_32_bits(rptr_addr));
 	WREG32_SOC15(GC, 0, mmCP_RB0_RPTR_ADDR_HI, upper_32_bits(rptr_addr) & CP_RB_RPTR_ADDR_HI__RB_RPTR_ADDR_HI_MASK);
 
-	wptr_gpu_addr = adev->wb.gpu_addr + (ring->wptr_offs * 4);
+	wptr_gpu_addr = ring->wptr_gpu_addr;
 	WREG32_SOC15(GC, 0, mmCP_RB_WPTR_POLL_ADDR_LO, lower_32_bits(wptr_gpu_addr));
 	WREG32_SOC15(GC, 0, mmCP_RB_WPTR_POLL_ADDR_HI, upper_32_bits(wptr_gpu_addr));
 
@@ -3389,7 +3450,6 @@ static int gfx_v9_0_cp_gfx_resume(struct amdgpu_device *adev)
 
 	/* start the ring */
 	gfx_v9_0_cp_gfx_start(adev);
-	ring->sched.ready = true;
 
 	return 0;
 }
@@ -3400,8 +3460,16 @@ static void gfx_v9_0_cp_compute_enable(struct amdgpu_device *adev, bool enable)
 		WREG32_SOC15_RLC(GC, 0, mmCP_MEC_CNTL, 0);
 	} else {
 		WREG32_SOC15_RLC(GC, 0, mmCP_MEC_CNTL,
-			(CP_MEC_CNTL__MEC_ME1_HALT_MASK | CP_MEC_CNTL__MEC_ME2_HALT_MASK));
-		adev->gfx.kiq.ring.sched.ready = false;
+				 (CP_MEC_CNTL__MEC_INVALIDATE_ICACHE_MASK |
+				  CP_MEC_CNTL__MEC_ME1_PIPE0_RESET_MASK |
+				  CP_MEC_CNTL__MEC_ME1_PIPE1_RESET_MASK |
+				  CP_MEC_CNTL__MEC_ME1_PIPE2_RESET_MASK |
+				  CP_MEC_CNTL__MEC_ME1_PIPE3_RESET_MASK |
+				  CP_MEC_CNTL__MEC_ME2_PIPE0_RESET_MASK |
+				  CP_MEC_CNTL__MEC_ME2_PIPE1_RESET_MASK |
+				  CP_MEC_CNTL__MEC_ME1_HALT_MASK |
+				  CP_MEC_CNTL__MEC_ME2_HALT_MASK));
+		adev->gfx.kiq[0].ring.sched.ready = false;
 	}
 	udelay(50);
 }
@@ -3458,9 +3526,7 @@ static void gfx_v9_0_kiq_setting(struct amdgpu_ring *ring)
 	tmp = RREG32_SOC15(GC, 0, mmRLC_CP_SCHEDULERS);
 	tmp &= 0xffffff00;
 	tmp |= (ring->me << 5) | (ring->pipe << 3) | (ring->queue);
-	WREG32_SOC15_RLC(GC, 0, mmRLC_CP_SCHEDULERS, tmp);
-	tmp |= 0x80;
-	WREG32_SOC15_RLC(GC, 0, mmRLC_CP_SCHEDULERS, tmp);
+	WREG32_SOC15_RLC(GC, 0, mmRLC_CP_SCHEDULERS, tmp | 0x80);
 }
 
 static void gfx_v9_0_mqd_set_priority(struct amdgpu_ring *ring, struct v9_mqd *mqd)
@@ -3558,7 +3624,7 @@ static int gfx_v9_0_mqd_init(struct amdgpu_ring *ring)
 	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, QUEUE_SIZE,
 			    (order_base_2(ring->ring_size / 4) - 1));
 	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, RPTR_BLOCK_SIZE,
-			((order_base_2(AMDGPU_GPU_PAGE_SIZE / 4) - 1) << 8));
+			(order_base_2(AMDGPU_GPU_PAGE_SIZE / 4) - 1));
 #ifdef __BIG_ENDIAN
 	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, ENDIAN_SWAP, 1);
 #endif
@@ -3569,33 +3635,16 @@ static int gfx_v9_0_mqd_init(struct amdgpu_ring *ring)
 	mqd->cp_hqd_pq_control = tmp;
 
 	/* set the wb address whether it's enabled or not */
-	wb_gpu_addr = adev->wb.gpu_addr + (ring->rptr_offs * 4);
+	wb_gpu_addr = ring->rptr_gpu_addr;
 	mqd->cp_hqd_pq_rptr_report_addr_lo = wb_gpu_addr & 0xfffffffc;
 	mqd->cp_hqd_pq_rptr_report_addr_hi =
 		upper_32_bits(wb_gpu_addr) & 0xffff;
 
 	/* only used if CP_PQ_WPTR_POLL_CNTL.CP_PQ_WPTR_POLL_CNTL__EN_MASK=1 */
-	wb_gpu_addr = adev->wb.gpu_addr + (ring->wptr_offs * 4);
+	wb_gpu_addr = ring->wptr_gpu_addr;
 	mqd->cp_hqd_pq_wptr_poll_addr_lo = wb_gpu_addr & 0xfffffffc;
 	mqd->cp_hqd_pq_wptr_poll_addr_hi = upper_32_bits(wb_gpu_addr) & 0xffff;
 
-	tmp = 0;
-	/* enable the doorbell if requested */
-	if (ring->use_doorbell) {
-		tmp = RREG32_SOC15(GC, 0, mmCP_HQD_PQ_DOORBELL_CONTROL);
-		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
-				DOORBELL_OFFSET, ring->doorbell_index);
-
-		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
-					 DOORBELL_EN, 1);
-		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
-					 DOORBELL_SOURCE, 0);
-		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
-					 DOORBELL_HIT, 0);
-	}
-
-	mqd->cp_hqd_pq_doorbell_control = tmp;
-
 	/* reset read and write pointers, similar to CP_RB0_WPTR/_RPTR */
 	ring->wptr = 0;
 	mqd->cp_hqd_pq_rptr = RREG32_SOC15(GC, 0, mmCP_HQD_PQ_RPTR);
@@ -3782,7 +3831,6 @@ static int gfx_v9_0_kiq_init_queue(struct amdgpu_ring *ring)
 {
 	struct amdgpu_device *adev = ring->adev;
 	struct v9_mqd *mqd = ring->mqd_ptr;
-	int mqd_idx = AMDGPU_MAX_COMPUTE_RINGS;
 	struct v9_mqd *tmp_mqd;
 
 	gfx_v9_0_kiq_setting(ring);
@@ -3792,40 +3840,42 @@ static int gfx_v9_0_kiq_init_queue(struct amdgpu_ring *ring)
 	 * driver need to re-init the mqd.
 	 * check mqd->cp_hqd_pq_control since this value should not be 0
 	 */
-	tmp_mqd = (struct v9_mqd *)adev->gfx.mec.mqd_backup[mqd_idx];
+	tmp_mqd = (struct v9_mqd *)adev->gfx.kiq[0].mqd_backup;
 	if (amdgpu_in_reset(adev) && tmp_mqd->cp_hqd_pq_control){
 		/* for GPU_RESET case , reset MQD to a clean status */
-		if (adev->gfx.mec.mqd_backup[mqd_idx])
-			memcpy(mqd, adev->gfx.mec.mqd_backup[mqd_idx], sizeof(struct v9_mqd_allocation));
+		if (adev->gfx.kiq[0].mqd_backup)
+			memcpy(mqd, adev->gfx.kiq[0].mqd_backup, sizeof(struct v9_mqd_allocation));
 
 		/* reset ring buffer */
 		ring->wptr = 0;
 		amdgpu_ring_clear_ring(ring);
 
 		mutex_lock(&adev->srbm_mutex);
-		soc15_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0);
+		soc15_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0, 0);
 		gfx_v9_0_kiq_init_register(ring);
-		soc15_grbm_select(adev, 0, 0, 0, 0);
+		soc15_grbm_select(adev, 0, 0, 0, 0, 0);
 		mutex_unlock(&adev->srbm_mutex);
 	} else {
 		memset((void *)mqd, 0, sizeof(struct v9_mqd_allocation));
 		((struct v9_mqd_allocation *)mqd)->dynamic_cu_mask = 0xFFFFFFFF;
 		((struct v9_mqd_allocation *)mqd)->dynamic_rb_mask = 0xFFFFFFFF;
+		if (amdgpu_sriov_vf(adev) && adev->in_suspend)
+			amdgpu_ring_clear_ring(ring);
 		mutex_lock(&adev->srbm_mutex);
-		soc15_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0);
+		soc15_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0, 0);
 		gfx_v9_0_mqd_init(ring);
 		gfx_v9_0_kiq_init_register(ring);
-		soc15_grbm_select(adev, 0, 0, 0, 0);
+		soc15_grbm_select(adev, 0, 0, 0, 0, 0);
 		mutex_unlock(&adev->srbm_mutex);
 
-		if (adev->gfx.mec.mqd_backup[mqd_idx])
-			memcpy(adev->gfx.mec.mqd_backup[mqd_idx], mqd, sizeof(struct v9_mqd_allocation));
+		if (adev->gfx.kiq[0].mqd_backup)
+			memcpy(adev->gfx.kiq[0].mqd_backup, mqd, sizeof(struct v9_mqd_allocation));
 	}
 
 	return 0;
 }
 
-static int gfx_v9_0_kcq_init_queue(struct amdgpu_ring *ring)
+static int gfx_v9_0_kcq_init_queue(struct amdgpu_ring *ring, bool restore)
 {
 	struct amdgpu_device *adev = ring->adev;
 	struct v9_mqd *mqd = ring->mqd_ptr;
@@ -3837,29 +3887,26 @@ static int gfx_v9_0_kcq_init_queue(struct amdgpu_ring *ring)
 	 */
 	tmp_mqd = (struct v9_mqd *)adev->gfx.mec.mqd_backup[mqd_idx];
 
-	if (!tmp_mqd->cp_hqd_pq_control ||
-	    (!amdgpu_in_reset(adev) && !adev->in_suspend)) {
+	if (!restore && (!tmp_mqd->cp_hqd_pq_control ||
+	    (!amdgpu_in_reset(adev) && !adev->in_suspend))) {
 		memset((void *)mqd, 0, sizeof(struct v9_mqd_allocation));
 		((struct v9_mqd_allocation *)mqd)->dynamic_cu_mask = 0xFFFFFFFF;
 		((struct v9_mqd_allocation *)mqd)->dynamic_rb_mask = 0xFFFFFFFF;
 		mutex_lock(&adev->srbm_mutex);
-		soc15_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0);
+		soc15_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0, 0);
 		gfx_v9_0_mqd_init(ring);
-		soc15_grbm_select(adev, 0, 0, 0, 0);
+		soc15_grbm_select(adev, 0, 0, 0, 0, 0);
 		mutex_unlock(&adev->srbm_mutex);
 
 		if (adev->gfx.mec.mqd_backup[mqd_idx])
 			memcpy(adev->gfx.mec.mqd_backup[mqd_idx], mqd, sizeof(struct v9_mqd_allocation));
-	} else if (amdgpu_in_reset(adev)) { /* for GPU_RESET case */
-		/* reset MQD to a clean status */
+	} else {
+		/* restore MQD to a clean status */
 		if (adev->gfx.mec.mqd_backup[mqd_idx])
 			memcpy(mqd, adev->gfx.mec.mqd_backup[mqd_idx], sizeof(struct v9_mqd_allocation));
-
 		/* reset ring buffer */
 		ring->wptr = 0;
-		atomic64_set((atomic64_t *)&adev->wb.wb[ring->wptr_offs], 0);
-		amdgpu_ring_clear_ring(ring);
-	} else {
+		atomic64_set((atomic64_t *)ring->wptr_cpu_addr, 0);
 		amdgpu_ring_clear_ring(ring);
 	}
 
@@ -3868,54 +3915,23 @@ static int gfx_v9_0_kcq_init_queue(struct amdgpu_ring *ring)
 
 static int gfx_v9_0_kiq_resume(struct amdgpu_device *adev)
 {
-	struct amdgpu_ring *ring;
-	int r;
-
-	ring = &adev->gfx.kiq.ring;
-
-	r = amdgpu_bo_reserve(ring->mqd_obj, false);
-	if (unlikely(r != 0))
-		return r;
-
-	r = amdgpu_bo_kmap(ring->mqd_obj, (void **)&ring->mqd_ptr);
-	if (unlikely(r != 0))
-		return r;
-
-	gfx_v9_0_kiq_init_queue(ring);
-	amdgpu_bo_kunmap(ring->mqd_obj);
-	ring->mqd_ptr = NULL;
-	amdgpu_bo_unreserve(ring->mqd_obj);
-	ring->sched.ready = true;
+	gfx_v9_0_kiq_init_queue(&adev->gfx.kiq[0].ring);
 	return 0;
 }
 
 static int gfx_v9_0_kcq_resume(struct amdgpu_device *adev)
 {
-	struct amdgpu_ring *ring = NULL;
-	int r = 0, i;
+	int i, r;
 
 	gfx_v9_0_cp_compute_enable(adev, true);
 
 	for (i = 0; i < adev->gfx.num_compute_rings; i++) {
-		ring = &adev->gfx.compute_ring[i];
-
-		r = amdgpu_bo_reserve(ring->mqd_obj, false);
-		if (unlikely(r != 0))
-			goto done;
-		r = amdgpu_bo_kmap(ring->mqd_obj, (void **)&ring->mqd_ptr);
-		if (!r) {
-			r = gfx_v9_0_kcq_init_queue(ring);
-			amdgpu_bo_kunmap(ring->mqd_obj);
-			ring->mqd_ptr = NULL;
-		}
-		amdgpu_bo_unreserve(ring->mqd_obj);
+		r = gfx_v9_0_kcq_init_queue(&adev->gfx.compute_ring[i], false);
 		if (r)
-			goto done;
+			return r;
 	}
 
-	r = amdgpu_gfx_enable_kcq(adev);
-done:
-	return r;
+	return amdgpu_gfx_enable_kcq(adev, 0);
 }
 
 static int gfx_v9_0_cp_resume(struct amdgpu_device *adev)
@@ -3939,6 +3955,10 @@ static int gfx_v9_0_cp_resume(struct amdgpu_device *adev)
 			return r;
 	}
 
+	if (adev->gfx.num_gfx_rings)
+		gfx_v9_0_cp_gfx_enable(adev, false);
+	gfx_v9_0_cp_compute_enable(adev, false);
+
 	r = gfx_v9_0_kiq_resume(adev);
 	if (r)
 		return r;
@@ -3974,8 +3994,8 @@ static void gfx_v9_0_init_tcp_config(struct amdgpu_device *adev)
 {
 	u32 tmp;
 
-	if (adev->ip_versions[GC_HWIP][0] != IP_VERSION(9, 4, 1) &&
-	    adev->ip_versions[GC_HWIP][0] != IP_VERSION(9, 4, 2))
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) != IP_VERSION(9, 4, 1) &&
+	    amdgpu_ip_version(adev, GC_HWIP, 0) != IP_VERSION(9, 4, 2))
 		return;
 
 	tmp = RREG32_SOC15(GC, 0, mmTCP_ADDR_CONFIG);
@@ -3995,10 +4015,13 @@ static void gfx_v9_0_cp_enable(struct amdgpu_device *adev, bool enable)
 	gfx_v9_0_cp_compute_enable(adev, enable);
 }
 
-static int gfx_v9_0_hw_init(void *handle)
+static int gfx_v9_0_hw_init(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
+
+	amdgpu_gfx_cleaner_shader_init(adev, adev->gfx.cleaner_shader_size,
+				       adev->gfx.cleaner_shader_ptr);
 
 	if (!amdgpu_sriov_vf(adev))
 		gfx_v9_0_init_golden_registers(adev);
@@ -4015,24 +4038,27 @@ static int gfx_v9_0_hw_init(void *handle)
 	if (r)
 		return r;
 
-	if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 2))
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 2) &&
+	    !amdgpu_sriov_vf(adev))
 		gfx_v9_4_2_set_power_brake_sequence(adev);
 
 	return r;
 }
 
-static int gfx_v9_0_hw_fini(void *handle)
+static int gfx_v9_0_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
-	amdgpu_irq_put(adev, &adev->gfx.cp_ecc_error_irq, 0);
+	if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__GFX))
+		amdgpu_irq_put(adev, &adev->gfx.cp_ecc_error_irq, 0);
 	amdgpu_irq_put(adev, &adev->gfx.priv_reg_irq, 0);
 	amdgpu_irq_put(adev, &adev->gfx.priv_inst_irq, 0);
+	amdgpu_irq_put(adev, &adev->gfx.bad_op_irq, 0);
 
 	/* DF freeze and kcq disable will fail */
 	if (!amdgpu_ras_intr_triggered())
 		/* disable KCQ to avoid CPC touch memory not valid anymore */
-		amdgpu_gfx_disable_kcq(adev);
+		amdgpu_gfx_disable_kcq(adev, 0);
 
 	if (amdgpu_sriov_vf(adev)) {
 		gfx_v9_0_cp_gfx_enable(adev, false);
@@ -4050,11 +4076,11 @@ static int gfx_v9_0_hw_fini(void *handle)
 	 */
 	if (!amdgpu_in_reset(adev) && !adev->in_suspend) {
 		mutex_lock(&adev->srbm_mutex);
-		soc15_grbm_select(adev, adev->gfx.kiq.ring.me,
-				adev->gfx.kiq.ring.pipe,
-				adev->gfx.kiq.ring.queue, 0);
-		gfx_v9_0_kiq_fini_register(&adev->gfx.kiq.ring);
-		soc15_grbm_select(adev, 0, 0, 0, 0);
+		soc15_grbm_select(adev, adev->gfx.kiq[0].ring.me,
+				adev->gfx.kiq[0].ring.pipe,
+				adev->gfx.kiq[0].ring.queue, 0, 0);
+		gfx_v9_0_kiq_fini_register(&adev->gfx.kiq[0].ring);
+		soc15_grbm_select(adev, 0, 0, 0, 0, 0);
 		mutex_unlock(&adev->srbm_mutex);
 	}
 
@@ -4062,7 +4088,7 @@ static int gfx_v9_0_hw_fini(void *handle)
 
 	/* Skip stopping RLC with A+A reset or when RLC controls GFX clock */
 	if ((adev->gmc.xgmi.connected_to_cpu && amdgpu_in_reset(adev)) ||
-	    (adev->ip_versions[GC_HWIP][0] >= IP_VERSION(9, 4, 2))) {
+	    (amdgpu_ip_version(adev, GC_HWIP, 0) >= IP_VERSION(9, 4, 2))) {
 		dev_dbg(adev->dev, "Skipping RLC halt\n");
 		return 0;
 	}
@@ -4071,19 +4097,19 @@ static int gfx_v9_0_hw_fini(void *handle)
 	return 0;
 }
 
-static int gfx_v9_0_suspend(void *handle)
+static int gfx_v9_0_suspend(struct amdgpu_ip_block *ip_block)
 {
-	return gfx_v9_0_hw_fini(handle);
+	return gfx_v9_0_hw_fini(ip_block);
 }
 
-static int gfx_v9_0_resume(void *handle)
+static int gfx_v9_0_resume(struct amdgpu_ip_block *ip_block)
 {
-	return gfx_v9_0_hw_init(handle);
+	return gfx_v9_0_hw_init(ip_block);
 }
 
-static bool gfx_v9_0_is_idle(void *handle)
+static bool gfx_v9_0_is_idle(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (REG_GET_FIELD(RREG32_SOC15(GC, 0, mmGRBM_STATUS),
 				GRBM_STATUS, GUI_ACTIVE))
@@ -4092,24 +4118,24 @@ static bool gfx_v9_0_is_idle(void *handle)
 		return true;
 }
 
-static int gfx_v9_0_wait_for_idle(void *handle)
+static int gfx_v9_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
 {
 	unsigned i;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	for (i = 0; i < adev->usec_timeout; i++) {
-		if (gfx_v9_0_is_idle(handle))
+		if (gfx_v9_0_is_idle(ip_block))
 			return 0;
 		udelay(1);
 	}
 	return -ETIMEDOUT;
 }
 
-static int gfx_v9_0_soft_reset(void *handle)
+static int gfx_v9_0_soft_reset(struct amdgpu_ip_block *ip_block)
 {
 	u32 grbm_soft_reset = 0;
 	u32 tmp;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	/* GRBM_STATUS */
 	tmp = RREG32_SOC15(GC, 0, mmGRBM_STATUS);
@@ -4148,19 +4174,17 @@ static int gfx_v9_0_soft_reset(void *handle)
 		/* Disable MEC parsing/prefetching */
 		gfx_v9_0_cp_compute_enable(adev, false);
 
-		if (grbm_soft_reset) {
-			tmp = RREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET);
-			tmp |= grbm_soft_reset;
-			dev_info(adev->dev, "GRBM_SOFT_RESET=0x%08X\n", tmp);
-			WREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET, tmp);
-			tmp = RREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET);
+		tmp = RREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET);
+		tmp |= grbm_soft_reset;
+		dev_info(adev->dev, "GRBM_SOFT_RESET=0x%08X\n", tmp);
+		WREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET, tmp);
+		tmp = RREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET);
 
-			udelay(50);
+		udelay(50);
 
-			tmp &= ~grbm_soft_reset;
-			WREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET, tmp);
-			tmp = RREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET);
-		}
+		tmp &= ~grbm_soft_reset;
+		WREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET, tmp);
+		tmp = RREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET);
 
 		/* Wait a little for things to settle down */
 		udelay(50);
@@ -4174,7 +4198,7 @@ static uint64_t gfx_v9_0_kiq_read_clock(struct amdgpu_device *adev)
 	unsigned long flags;
 	uint32_t seq, reg_val_offs = 0;
 	uint64_t value = 0;
-	struct amdgpu_kiq *kiq = &adev->gfx.kiq;
+	struct amdgpu_kiq *kiq = &adev->gfx.kiq[0];
 	struct amdgpu_ring *ring = &kiq->ring;
 
 	BUG_ON(!ring->funcs->emit_rreg);
@@ -4246,7 +4270,7 @@ static uint64_t gfx_v9_0_get_gpu_clock_counter(struct amdgpu_device *adev)
 {
 	uint64_t clock, clock_lo, clock_hi, hi_check;
 
-	switch (adev->ip_versions[GC_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 	case IP_VERSION(9, 3, 0):
 		preempt_disable();
 		clock_hi = RREG32_SOC15_NO_KIQ(SMUIO, 0, mmGOLDEN_TSC_COUNT_UPPER_Renoir);
@@ -4265,7 +4289,9 @@ static uint64_t gfx_v9_0_get_gpu_clock_counter(struct amdgpu_device *adev)
 	default:
 		amdgpu_gfx_off_ctrl(adev, false);
 		mutex_lock(&adev->gfx.gpu_clock_mutex);
-		if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 0, 1) && amdgpu_sriov_runtime(adev)) {
+		if (amdgpu_ip_version(adev, GC_HWIP, 0) ==
+			    IP_VERSION(9, 0, 1) &&
+		    amdgpu_sriov_runtime(adev)) {
 			clock = gfx_v9_0_kiq_read_clock(adev);
 		} else {
 			WREG32_SOC15(GC, 0, mmRLC_CAPTURE_GPU_CLOCK_COUNT, 1);
@@ -4617,7 +4643,7 @@ static int gfx_v9_0_do_edc_gpr_workarounds(struct amdgpu_device *adev)
 	if (!ring->sched.ready)
 		return 0;
 
-	if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 1)) {
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 1)) {
 		vgpr_init_shader_ptr = vgpr_init_compute_shader_arcturus;
 		vgpr_init_shader_size = sizeof(vgpr_init_compute_shader_arcturus);
 		vgpr_init_regs_ptr = vgpr_init_regs_arcturus;
@@ -4757,21 +4783,24 @@ static int gfx_v9_0_do_edc_gpr_workarounds(struct amdgpu_device *adev)
 	}
 
 fail:
-	amdgpu_ib_free(adev, &ib, NULL);
+	amdgpu_ib_free(&ib, NULL);
 	dma_fence_put(f);
 
 	return r;
 }
 
-static int gfx_v9_0_early_init(void *handle)
+static int gfx_v9_0_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
+
+	adev->gfx.funcs = &gfx_v9_0_gfx_funcs;
 
-	if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 1) ||
-	    adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 2))
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 1) ||
+	    amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 2))
 		adev->gfx.num_gfx_rings = 0;
 	else
 		adev->gfx.num_gfx_rings = GFX9_NUM_GFX_RINGS;
+	adev->gfx.xcc_mask = 1;
 	adev->gfx.num_compute_rings = min(amdgpu_gfx_get_num_kcq(adev),
 					  AMDGPU_MAX_COMPUTE_RINGS);
 	gfx_v9_0_set_kiq_pm4_funcs(adev);
@@ -4780,12 +4809,15 @@ static int gfx_v9_0_early_init(void *handle)
 	gfx_v9_0_set_gds_init(adev);
 	gfx_v9_0_set_rlc_funcs(adev);
 
-	return 0;
+	/* init rlcg reg access ctrl */
+	gfx_v9_0_init_rlcg_reg_access_ctrl(adev);
+
+	return gfx_v9_0_init_microcode(adev);
 }
 
-static int gfx_v9_0_ecc_late_init(void *handle)
+static int gfx_v9_0_ecc_late_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int r;
 
 	/*
@@ -4802,7 +4834,7 @@ static int gfx_v9_0_ecc_late_init(void *handle)
 	}
 
 	/* requires IBs so do in late init after IB pool is initialized */
-	if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 2))
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 2))
 		r = gfx_v9_4_2_do_edc_gpr_workarounds(adev);
 	else
 		r = gfx_v9_0_do_edc_gpr_workarounds(adev);
@@ -4810,23 +4842,16 @@ static int gfx_v9_0_ecc_late_init(void *handle)
 	if (r)
 		return r;
 
-	if (adev->gfx.ras_funcs &&
-	    adev->gfx.ras_funcs->ras_late_init) {
-		r = adev->gfx.ras_funcs->ras_late_init(adev);
-		if (r)
-			return r;
-	}
-
-	if (adev->gfx.ras_funcs &&
-	    adev->gfx.ras_funcs->enable_watchdog_timer)
-		adev->gfx.ras_funcs->enable_watchdog_timer(adev);
+	if (adev->gfx.ras &&
+	    adev->gfx.ras->enable_watchdog_timer)
+		adev->gfx.ras->enable_watchdog_timer(adev);
 
 	return 0;
 }
 
-static int gfx_v9_0_late_init(void *handle)
+static int gfx_v9_0_late_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int r;
 
 	r = amdgpu_irq_get(adev, &adev->gfx.priv_reg_irq, 0);
@@ -4837,10 +4862,21 @@ static int gfx_v9_0_late_init(void *handle)
 	if (r)
 		return r;
 
-	r = gfx_v9_0_ecc_late_init(handle);
+	r = amdgpu_irq_get(adev, &adev->gfx.bad_op_irq, 0);
 	if (r)
 		return r;
 
+	r = gfx_v9_0_ecc_late_init(ip_block);
+	if (r)
+		return r;
+
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 2))
+		gfx_v9_4_2_debug_trap_config_init(adev,
+			adev->vm_manager.first_kfd_vmid, AMDGPU_NUM_VMID);
+	else
+		gfx_v9_0_debug_trap_config_init(adev,
+			adev->vm_manager.first_kfd_vmid, AMDGPU_NUM_VMID);
+
 	return 0;
 }
 
@@ -4856,7 +4892,7 @@ static bool gfx_v9_0_is_rlc_enabled(struct amdgpu_device *adev)
 	return true;
 }
 
-static void gfx_v9_0_set_safe_mode(struct amdgpu_device *adev)
+static void gfx_v9_0_set_safe_mode(struct amdgpu_device *adev, int xcc_id)
 {
 	uint32_t data;
 	unsigned i;
@@ -4873,7 +4909,7 @@ static void gfx_v9_0_set_safe_mode(struct amdgpu_device *adev)
 	}
 }
 
-static void gfx_v9_0_unset_safe_mode(struct amdgpu_device *adev)
+static void gfx_v9_0_unset_safe_mode(struct amdgpu_device *adev, int xcc_id)
 {
 	uint32_t data;
 
@@ -4884,7 +4920,7 @@ static void gfx_v9_0_unset_safe_mode(struct amdgpu_device *adev)
 static void gfx_v9_0_update_gfx_cg_power_gating(struct amdgpu_device *adev,
 						bool enable)
 {
-	amdgpu_gfx_rlc_enter_safe_mode(adev);
+	amdgpu_gfx_rlc_enter_safe_mode(adev, 0);
 
 	if ((adev->pg_flags & AMD_PG_SUPPORT_GFX_PG) && enable) {
 		gfx_v9_0_enable_gfx_cg_power_gating(adev, true);
@@ -4896,7 +4932,7 @@ static void gfx_v9_0_update_gfx_cg_power_gating(struct amdgpu_device *adev,
 			gfx_v9_0_enable_gfx_pipeline_powergating(adev, false);
 	}
 
-	amdgpu_gfx_rlc_exit_safe_mode(adev);
+	amdgpu_gfx_rlc_exit_safe_mode(adev, 0);
 }
 
 static void gfx_v9_0_update_gfx_mg_power_gating(struct amdgpu_device *adev,
@@ -4923,14 +4959,12 @@ static void gfx_v9_0_update_medium_grain_clock_gating(struct amdgpu_device *adev
 {
 	uint32_t data, def;
 
-	amdgpu_gfx_rlc_enter_safe_mode(adev);
-
 	/* It is disabled by HW by default */
 	if (enable && (adev->cg_flags & AMD_CG_SUPPORT_GFX_MGCG)) {
 		/* 1 - RLC_CGTT_MGCG_OVERRIDE */
 		def = data = RREG32_SOC15(GC, 0, mmRLC_CGTT_MGCG_OVERRIDE);
 
-		if (adev->ip_versions[GC_HWIP][0] != IP_VERSION(9, 2, 1))
+		if (amdgpu_ip_version(adev, GC_HWIP, 0) != IP_VERSION(9, 2, 1))
 			data &= ~RLC_CGTT_MGCG_OVERRIDE__CPF_CGTT_SCLK_OVERRIDE_MASK;
 
 		data &= ~(RLC_CGTT_MGCG_OVERRIDE__GRBM_CGTT_SCLK_OVERRIDE_MASK |
@@ -4964,7 +4998,7 @@ static void gfx_v9_0_update_medium_grain_clock_gating(struct amdgpu_device *adev
 		/* 1 - MGCG_OVERRIDE */
 		def = data = RREG32_SOC15(GC, 0, mmRLC_CGTT_MGCG_OVERRIDE);
 
-		if (adev->ip_versions[GC_HWIP][0] != IP_VERSION(9, 2, 1))
+		if (amdgpu_ip_version(adev, GC_HWIP, 0) != IP_VERSION(9, 2, 1))
 			data |= RLC_CGTT_MGCG_OVERRIDE__CPF_CGTT_SCLK_OVERRIDE_MASK;
 
 		data |= (RLC_CGTT_MGCG_OVERRIDE__RLC_CGTT_SCLK_OVERRIDE_MASK |
@@ -4989,8 +5023,6 @@ static void gfx_v9_0_update_medium_grain_clock_gating(struct amdgpu_device *adev
 			WREG32_SOC15(GC, 0, mmCP_MEM_SLP_CNTL, data);
 		}
 	}
-
-	amdgpu_gfx_rlc_exit_safe_mode(adev);
 }
 
 static void gfx_v9_0_update_3d_clock_gating(struct amdgpu_device *adev,
@@ -5001,8 +5033,6 @@ static void gfx_v9_0_update_3d_clock_gating(struct amdgpu_device *adev,
 	if (!adev->gfx.num_gfx_rings)
 		return;
 
-	amdgpu_gfx_rlc_enter_safe_mode(adev);
-
 	/* Enable 3D CGCG/CGLS */
 	if (enable) {
 		/* write cmd to clear cgcg/cgls ov */
@@ -5044,8 +5074,6 @@ static void gfx_v9_0_update_3d_clock_gating(struct amdgpu_device *adev,
 		if (def != data)
 			WREG32_SOC15(GC, 0, mmRLC_CGCG_CGLS_CTRL_3D, data);
 	}
-
-	amdgpu_gfx_rlc_exit_safe_mode(adev);
 }
 
 static void gfx_v9_0_update_coarse_grain_clock_gating(struct amdgpu_device *adev,
@@ -5053,8 +5081,6 @@ static void gfx_v9_0_update_coarse_grain_clock_gating(struct amdgpu_device *adev
 {
 	uint32_t def, data;
 
-	amdgpu_gfx_rlc_enter_safe_mode(adev);
-
 	if (enable && (adev->cg_flags & AMD_CG_SUPPORT_GFX_CGCG)) {
 		def = data = RREG32_SOC15(GC, 0, mmRLC_CGTT_MGCG_OVERRIDE);
 		/* unset CGCG override */
@@ -5070,7 +5096,7 @@ static void gfx_v9_0_update_coarse_grain_clock_gating(struct amdgpu_device *adev
 		/* enable cgcg FSM(0x0000363F) */
 		def = RREG32_SOC15(GC, 0, mmRLC_CGCG_CGLS_CTRL);
 
-		if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 1))
+		if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 1))
 			data = (0x2000 << RLC_CGCG_CGLS_CTRL__CGCG_GFX_IDLE_THRESHOLD__SHIFT) |
 				RLC_CGCG_CGLS_CTRL__CGCG_EN_MASK;
 		else
@@ -5096,13 +5122,12 @@ static void gfx_v9_0_update_coarse_grain_clock_gating(struct amdgpu_device *adev
 		if (def != data)
 			WREG32_SOC15(GC, 0, mmRLC_CGCG_CGLS_CTRL, data);
 	}
-
-	amdgpu_gfx_rlc_exit_safe_mode(adev);
 }
 
 static int gfx_v9_0_update_gfx_clock_gating(struct amdgpu_device *adev,
 					    bool enable)
 {
+	amdgpu_gfx_rlc_enter_safe_mode(adev, 0);
 	if (enable) {
 		/* CGCG/CGLS should be enabled after MGCG/MGLS
 		 * ===  MGCG + MGLS ===
@@ -5122,20 +5147,20 @@ static int gfx_v9_0_update_gfx_clock_gating(struct amdgpu_device *adev,
 		/* ===  MGCG + MGLS === */
 		gfx_v9_0_update_medium_grain_clock_gating(adev, enable);
 	}
+	amdgpu_gfx_rlc_exit_safe_mode(adev, 0);
 	return 0;
 }
 
-static void gfx_v9_0_update_spm_vmid(struct amdgpu_device *adev, unsigned vmid)
+static void gfx_v9_0_update_spm_vmid_internal(struct amdgpu_device *adev,
+					      unsigned int vmid)
 {
 	u32 reg, data;
 
-	amdgpu_gfx_off_ctrl(adev, false);
-
 	reg = SOC15_REG_OFFSET(GC, 0, mmRLC_SPM_MC_CNTL);
 	if (amdgpu_sriov_is_pp_one_vf(adev))
 		data = RREG32_NO_KIQ(reg);
 	else
-		data = RREG32(reg);
+		data = RREG32_SOC15(GC, 0, mmRLC_SPM_MC_CNTL);
 
 	data &= ~RLC_SPM_MC_CNTL__RLC_SPM_VMID_MASK;
 	data |= (vmid & RLC_SPM_MC_CNTL__RLC_SPM_VMID_MASK) << RLC_SPM_MC_CNTL__RLC_SPM_VMID__SHIFT;
@@ -5144,6 +5169,13 @@ static void gfx_v9_0_update_spm_vmid(struct amdgpu_device *adev, unsigned vmid)
 		WREG32_SOC15_NO_KIQ(GC, 0, mmRLC_SPM_MC_CNTL, data);
 	else
 		WREG32_SOC15(GC, 0, mmRLC_SPM_MC_CNTL, data);
+}
+
+static void gfx_v9_0_update_spm_vmid(struct amdgpu_device *adev, struct amdgpu_ring *ring, unsigned int vmid)
+{
+	amdgpu_gfx_off_ctrl(adev, false);
+
+	gfx_v9_0_update_spm_vmid_internal(adev, vmid);
 
 	amdgpu_gfx_off_ctrl(adev, true);
 }
@@ -5190,22 +5222,21 @@ static const struct amdgpu_rlc_funcs gfx_v9_0_rlc_funcs = {
 	.reset = gfx_v9_0_rlc_reset,
 	.start = gfx_v9_0_rlc_start,
 	.update_spm_vmid = gfx_v9_0_update_spm_vmid,
-	.sriov_wreg = gfx_v9_0_sriov_wreg,
 	.is_rlcg_access_range = gfx_v9_0_is_rlcg_access_range,
 };
 
-static int gfx_v9_0_set_powergating_state(void *handle,
+static int gfx_v9_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_powergating_state state)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	bool enable = (state == AMD_PG_STATE_GATE);
 
-	switch (adev->ip_versions[GC_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 	case IP_VERSION(9, 2, 2):
 	case IP_VERSION(9, 1, 0):
 	case IP_VERSION(9, 3, 0):
 		if (!enable)
-			amdgpu_gfx_off_ctrl(adev, false);
+			amdgpu_gfx_off_ctrl_immediate(adev, false);
 
 		if (adev->pg_flags & AMD_PG_SUPPORT_RLC_SMU_HS) {
 			gfx_v9_0_enable_sck_slow_down_on_power_up(adev, true);
@@ -5227,10 +5258,10 @@ static int gfx_v9_0_set_powergating_state(void *handle,
 		gfx_v9_0_update_gfx_mg_power_gating(adev, enable);
 
 		if (enable)
-			amdgpu_gfx_off_ctrl(adev, true);
+			amdgpu_gfx_off_ctrl_immediate(adev, true);
 		break;
 	case IP_VERSION(9, 2, 1):
-		amdgpu_gfx_off_ctrl(adev, enable);
+		amdgpu_gfx_off_ctrl_immediate(adev, enable);
 		break;
 	default:
 		break;
@@ -5239,15 +5270,15 @@ static int gfx_v9_0_set_powergating_state(void *handle,
 	return 0;
 }
 
-static int gfx_v9_0_set_clockgating_state(void *handle,
+static int gfx_v9_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_clockgating_state state)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (amdgpu_sriov_vf(adev))
 		return 0;
 
-	switch (adev->ip_versions[GC_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 	case IP_VERSION(9, 0, 1):
 	case IP_VERSION(9, 2, 1):
 	case IP_VERSION(9, 4, 0):
@@ -5265,9 +5296,9 @@ static int gfx_v9_0_set_clockgating_state(void *handle,
 	return 0;
 }
 
-static void gfx_v9_0_get_clockgating_state(void *handle, u32 *flags)
+static void gfx_v9_0_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int data;
 
 	if (amdgpu_sriov_vf(adev))
@@ -5297,7 +5328,7 @@ static void gfx_v9_0_get_clockgating_state(void *handle, u32 *flags)
 	if (data & CP_MEM_SLP_CNTL__CP_MEM_LS_EN_MASK)
 		*flags |= AMD_CG_SUPPORT_GFX_CP_LS | AMD_CG_SUPPORT_GFX_MGLS;
 
-	if (adev->ip_versions[GC_HWIP][0] != IP_VERSION(9, 4, 1)) {
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) != IP_VERSION(9, 4, 1)) {
 		/* AMD_CG_SUPPORT_GFX_3D_CGCG */
 		data = RREG32_KIQ(SOC15_REG_OFFSET(GC, 0, mmRLC_CGCG_CGLS_CTRL_3D));
 		if (data & RLC_CGCG_CGLS_CTRL_3D__CGCG_EN_MASK)
@@ -5311,7 +5342,7 @@ static void gfx_v9_0_get_clockgating_state(void *handle, u32 *flags)
 
 static u64 gfx_v9_0_ring_get_rptr_gfx(struct amdgpu_ring *ring)
 {
-	return ring->adev->wb.wb[ring->rptr_offs]; /* gfx9 is 32bit rptr*/
+	return *ring->rptr_cpu_addr; /* gfx9 is 32bit rptr*/
 }
 
 static u64 gfx_v9_0_ring_get_wptr_gfx(struct amdgpu_ring *ring)
@@ -5321,7 +5352,7 @@ static u64 gfx_v9_0_ring_get_wptr_gfx(struct amdgpu_ring *ring)
 
 	/* XXX check if swapping is necessary on BE */
 	if (ring->use_doorbell) {
-		wptr = atomic64_read((atomic64_t *)&adev->wb.wb[ring->wptr_offs]);
+		wptr = atomic64_read((atomic64_t *)ring->wptr_cpu_addr);
 	} else {
 		wptr = RREG32_SOC15(GC, 0, mmCP_RB0_WPTR);
 		wptr += (u64)RREG32_SOC15(GC, 0, mmCP_RB0_WPTR_HI) << 32;
@@ -5336,7 +5367,7 @@ static void gfx_v9_0_ring_set_wptr_gfx(struct amdgpu_ring *ring)
 
 	if (ring->use_doorbell) {
 		/* XXX check if swapping is necessary on BE */
-		atomic64_set((atomic64_t *)&adev->wb.wb[ring->wptr_offs], ring->wptr);
+		atomic64_set((atomic64_t *)ring->wptr_cpu_addr, ring->wptr);
 		WDOORBELL64(ring->doorbell_index, ring->wptr);
 	} else {
 		WREG32_SOC15(GC, 0, mmCP_RB0_WPTR, lower_32_bits(ring->wptr));
@@ -5388,11 +5419,18 @@ static void gfx_v9_0_ring_emit_ib_gfx(struct amdgpu_ring *ring,
 
 	control |= ib->length_dw | (vmid << 24);
 
-	if (amdgpu_sriov_vf(ring->adev) && (ib->flags & AMDGPU_IB_FLAG_PREEMPT)) {
+	if (ib->flags & AMDGPU_IB_FLAG_PREEMPT) {
 		control |= INDIRECT_BUFFER_PRE_ENB(1);
 
+		if (flags & AMDGPU_IB_PREEMPTED)
+			control |= INDIRECT_BUFFER_PRE_RESUME(1);
+
 		if (!(ib->flags & AMDGPU_IB_FLAG_CE) && vmid)
-			gfx_v9_0_ring_emit_de_meta(ring);
+			gfx_v9_0_ring_emit_de_meta(ring,
+						   (!amdgpu_sriov_vf(ring->adev) &&
+						   flags & AMDGPU_IB_PREEMPTED) ?
+						   true : false,
+						   job->gds_size > 0 && job->gds_base != 0);
 	}
 
 	amdgpu_ring_write(ring, header);
@@ -5403,9 +5441,70 @@ static void gfx_v9_0_ring_emit_ib_gfx(struct amdgpu_ring *ring,
 #endif
 		lower_32_bits(ib->gpu_addr));
 	amdgpu_ring_write(ring, upper_32_bits(ib->gpu_addr));
+	amdgpu_ring_ib_on_emit_cntl(ring);
 	amdgpu_ring_write(ring, control);
 }
 
+static void gfx_v9_0_ring_patch_cntl(struct amdgpu_ring *ring,
+				     unsigned offset)
+{
+	u32 control = ring->ring[offset];
+
+	control |= INDIRECT_BUFFER_PRE_RESUME(1);
+	ring->ring[offset] = control;
+}
+
+static void gfx_v9_0_ring_patch_ce_meta(struct amdgpu_ring *ring,
+					unsigned offset)
+{
+	struct amdgpu_device *adev = ring->adev;
+	void *ce_payload_cpu_addr;
+	uint64_t payload_offset, payload_size;
+
+	payload_size = sizeof(struct v9_ce_ib_state);
+
+	payload_offset = offsetof(struct v9_gfx_meta_data, ce_payload);
+	ce_payload_cpu_addr = adev->virt.csa_cpu_addr + payload_offset;
+
+	if (offset + (payload_size >> 2) <= ring->buf_mask + 1) {
+		memcpy((void *)&ring->ring[offset], ce_payload_cpu_addr, payload_size);
+	} else {
+		memcpy((void *)&ring->ring[offset], ce_payload_cpu_addr,
+		       (ring->buf_mask + 1 - offset) << 2);
+		payload_size -= (ring->buf_mask + 1 - offset) << 2;
+		memcpy((void *)&ring->ring[0],
+		       ce_payload_cpu_addr + ((ring->buf_mask + 1 - offset) << 2),
+		       payload_size);
+	}
+}
+
+static void gfx_v9_0_ring_patch_de_meta(struct amdgpu_ring *ring,
+					unsigned offset)
+{
+	struct amdgpu_device *adev = ring->adev;
+	void *de_payload_cpu_addr;
+	uint64_t payload_offset, payload_size;
+
+	payload_size = sizeof(struct v9_de_ib_state);
+
+	payload_offset = offsetof(struct v9_gfx_meta_data, de_payload);
+	de_payload_cpu_addr = adev->virt.csa_cpu_addr + payload_offset;
+
+	((struct v9_de_ib_state *)de_payload_cpu_addr)->ib_completion_status =
+		IB_COMPLETION_STATUS_PREEMPTED;
+
+	if (offset + (payload_size >> 2) <= ring->buf_mask + 1) {
+		memcpy((void *)&ring->ring[offset], de_payload_cpu_addr, payload_size);
+	} else {
+		memcpy((void *)&ring->ring[offset], de_payload_cpu_addr,
+		       (ring->buf_mask + 1 - offset) << 2);
+		payload_size -= (ring->buf_mask + 1 - offset) << 2;
+		memcpy((void *)&ring->ring[0],
+		       de_payload_cpu_addr + ((ring->buf_mask + 1 - offset) << 2),
+		       payload_size);
+	}
+}
+
 static void gfx_v9_0_ring_emit_ib_compute(struct amdgpu_ring *ring,
 					  struct amdgpu_job *job,
 					  struct amdgpu_ib *ib,
@@ -5447,17 +5546,24 @@ static void gfx_v9_0_ring_emit_fence(struct amdgpu_ring *ring, u64 addr,
 	bool write64bit = flags & AMDGPU_FENCE_FLAG_64BIT;
 	bool int_sel = flags & AMDGPU_FENCE_FLAG_INT;
 	bool writeback = flags & AMDGPU_FENCE_FLAG_TC_WB_ONLY;
+	bool exec = flags & AMDGPU_FENCE_FLAG_EXEC;
+	uint32_t dw2 = 0;
 
 	/* RELEASE_MEM - flush caches, send int */
 	amdgpu_ring_write(ring, PACKET3(PACKET3_RELEASE_MEM, 6));
-	amdgpu_ring_write(ring, ((writeback ? (EOP_TC_WB_ACTION_EN |
-					       EOP_TC_NC_ACTION_EN) :
-					      (EOP_TCL1_ACTION_EN |
-					       EOP_TC_ACTION_EN |
-					       EOP_TC_WB_ACTION_EN |
-					       EOP_TC_MD_ACTION_EN)) |
-				 EVENT_TYPE(CACHE_FLUSH_AND_INV_TS_EVENT) |
-				 EVENT_INDEX(5)));
+
+	if (writeback) {
+		dw2 = EOP_TC_NC_ACTION_EN;
+	} else {
+		dw2 = EOP_TCL1_ACTION_EN | EOP_TC_ACTION_EN |
+				EOP_TC_MD_ACTION_EN;
+	}
+	dw2 |= EOP_TC_WB_ACTION_EN | EVENT_TYPE(CACHE_FLUSH_AND_INV_TS_EVENT) |
+				EVENT_INDEX(5);
+	if (exec)
+		dw2 |= EOP_EXEC;
+
+	amdgpu_ring_write(ring, dw2);
 	amdgpu_ring_write(ring, DATA_SEL(write64bit ? 2 : 1) | INT_SEL(int_sel ? 2 : 0));
 
 	/*
@@ -5501,7 +5607,7 @@ static void gfx_v9_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
 
 static u64 gfx_v9_0_ring_get_rptr_compute(struct amdgpu_ring *ring)
 {
-	return ring->adev->wb.wb[ring->rptr_offs]; /* gfx9 hardware is 32bit rptr */
+	return *ring->rptr_cpu_addr; /* gfx9 hardware is 32bit rptr */
 }
 
 static u64 gfx_v9_0_ring_get_wptr_compute(struct amdgpu_ring *ring)
@@ -5510,7 +5616,7 @@ static u64 gfx_v9_0_ring_get_wptr_compute(struct amdgpu_ring *ring)
 
 	/* XXX check if swapping is necessary on BE */
 	if (ring->use_doorbell)
-		wptr = atomic64_read((atomic64_t *)&ring->adev->wb.wb[ring->wptr_offs]);
+		wptr = atomic64_read((atomic64_t *)ring->wptr_cpu_addr);
 	else
 		BUG();
 	return wptr;
@@ -5522,7 +5628,7 @@ static void gfx_v9_0_ring_set_wptr_compute(struct amdgpu_ring *ring)
 
 	/* XXX check if swapping is necessary on BE */
 	if (ring->use_doorbell) {
-		atomic64_set((atomic64_t *)&adev->wb.wb[ring->wptr_offs], ring->wptr);
+		atomic64_set((atomic64_t *)ring->wptr_cpu_addr, ring->wptr);
 		WDOORBELL64(ring->doorbell_index, ring->wptr);
 	} else{
 		BUG(); /* only DOORBELL method supported on gfx9 now */
@@ -5562,35 +5668,116 @@ static void gfx_v9_ring_emit_sb(struct amdgpu_ring *ring)
 	amdgpu_ring_write(ring, 0);
 }
 
-static void gfx_v9_0_ring_emit_ce_meta(struct amdgpu_ring *ring)
+static void gfx_v9_0_ring_emit_ce_meta(struct amdgpu_ring *ring, bool resume)
 {
+	struct amdgpu_device *adev = ring->adev;
 	struct v9_ce_ib_state ce_payload = {0};
-	uint64_t csa_addr;
+	uint64_t offset, ce_payload_gpu_addr;
+	void *ce_payload_cpu_addr;
 	int cnt;
 
 	cnt = (sizeof(ce_payload) >> 2) + 4 - 2;
-	csa_addr = amdgpu_csa_vaddr(ring->adev);
+
+	offset = offsetof(struct v9_gfx_meta_data, ce_payload);
+	ce_payload_gpu_addr = amdgpu_csa_vaddr(ring->adev) + offset;
+	ce_payload_cpu_addr = adev->virt.csa_cpu_addr + offset;
 
 	amdgpu_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, cnt));
 	amdgpu_ring_write(ring, (WRITE_DATA_ENGINE_SEL(2) |
 				 WRITE_DATA_DST_SEL(8) |
 				 WR_CONFIRM) |
 				 WRITE_DATA_CACHE_POLICY(0));
-	amdgpu_ring_write(ring, lower_32_bits(csa_addr + offsetof(struct v9_gfx_meta_data, ce_payload)));
-	amdgpu_ring_write(ring, upper_32_bits(csa_addr + offsetof(struct v9_gfx_meta_data, ce_payload)));
-	amdgpu_ring_write_multiple(ring, (void *)&ce_payload, sizeof(ce_payload) >> 2);
+	amdgpu_ring_write(ring, lower_32_bits(ce_payload_gpu_addr));
+	amdgpu_ring_write(ring, upper_32_bits(ce_payload_gpu_addr));
+
+	amdgpu_ring_ib_on_emit_ce(ring);
+
+	if (resume)
+		amdgpu_ring_write_multiple(ring, ce_payload_cpu_addr,
+					   sizeof(ce_payload) >> 2);
+	else
+		amdgpu_ring_write_multiple(ring, (void *)&ce_payload,
+					   sizeof(ce_payload) >> 2);
+}
+
+static int gfx_v9_0_ring_preempt_ib(struct amdgpu_ring *ring)
+{
+	int i, r = 0;
+	struct amdgpu_device *adev = ring->adev;
+	struct amdgpu_kiq *kiq = &adev->gfx.kiq[0];
+	struct amdgpu_ring *kiq_ring = &kiq->ring;
+	unsigned long flags;
+
+	if (!kiq->pmf || !kiq->pmf->kiq_unmap_queues)
+		return -EINVAL;
+
+	spin_lock_irqsave(&kiq->ring_lock, flags);
+
+	if (amdgpu_ring_alloc(kiq_ring, kiq->pmf->unmap_queues_size)) {
+		spin_unlock_irqrestore(&kiq->ring_lock, flags);
+		return -ENOMEM;
+	}
+
+	/* assert preemption condition */
+	amdgpu_ring_set_preempt_cond_exec(ring, false);
+
+	ring->trail_seq += 1;
+	amdgpu_ring_alloc(ring, 13);
+	gfx_v9_0_ring_emit_fence(ring, ring->trail_fence_gpu_addr,
+				 ring->trail_seq, AMDGPU_FENCE_FLAG_EXEC | AMDGPU_FENCE_FLAG_INT);
+
+	/* assert IB preemption, emit the trailing fence */
+	kiq->pmf->kiq_unmap_queues(kiq_ring, ring, PREEMPT_QUEUES_NO_UNMAP,
+				   ring->trail_fence_gpu_addr,
+				   ring->trail_seq);
+
+	amdgpu_ring_commit(kiq_ring);
+	spin_unlock_irqrestore(&kiq->ring_lock, flags);
+
+	/* poll the trailing fence */
+	for (i = 0; i < adev->usec_timeout; i++) {
+		if (ring->trail_seq ==
+			le32_to_cpu(*ring->trail_fence_cpu_addr))
+			break;
+		udelay(1);
+	}
+
+	if (i >= adev->usec_timeout) {
+		r = -EINVAL;
+		DRM_WARN("ring %d timeout to preempt ib\n", ring->idx);
+	}
+
+	/*reset the CP_VMID_PREEMPT after trailing fence*/
+	amdgpu_ring_emit_wreg(ring,
+			      SOC15_REG_OFFSET(GC, 0, mmCP_VMID_PREEMPT),
+			      0x0);
+	amdgpu_ring_commit(ring);
+
+	/* deassert preemption condition */
+	amdgpu_ring_set_preempt_cond_exec(ring, true);
+	return r;
 }
 
-static void gfx_v9_0_ring_emit_de_meta(struct amdgpu_ring *ring)
+static void gfx_v9_0_ring_emit_de_meta(struct amdgpu_ring *ring, bool resume, bool usegds)
 {
+	struct amdgpu_device *adev = ring->adev;
 	struct v9_de_ib_state de_payload = {0};
-	uint64_t csa_addr, gds_addr;
+	uint64_t offset, gds_addr, de_payload_gpu_addr;
+	void *de_payload_cpu_addr;
 	int cnt;
 
-	csa_addr = amdgpu_csa_vaddr(ring->adev);
-	gds_addr = csa_addr + 4096;
-	de_payload.gds_backup_addrlo = lower_32_bits(gds_addr);
-	de_payload.gds_backup_addrhi = upper_32_bits(gds_addr);
+	offset = offsetof(struct v9_gfx_meta_data, de_payload);
+	de_payload_gpu_addr = amdgpu_csa_vaddr(ring->adev) + offset;
+	de_payload_cpu_addr = adev->virt.csa_cpu_addr + offset;
+
+	gds_addr = ALIGN(amdgpu_csa_vaddr(ring->adev) +
+			 AMDGPU_CSA_SIZE - adev->gds.gds_size,
+			 PAGE_SIZE);
+
+	if (usegds) {
+		de_payload.gds_backup_addrlo = lower_32_bits(gds_addr);
+		de_payload.gds_backup_addrhi = upper_32_bits(gds_addr);
+	}
 
 	cnt = (sizeof(de_payload) >> 2) + 4 - 2;
 	amdgpu_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, cnt));
@@ -5598,9 +5785,16 @@ static void gfx_v9_0_ring_emit_de_meta(struct amdgpu_ring *ring)
 				 WRITE_DATA_DST_SEL(8) |
 				 WR_CONFIRM) |
 				 WRITE_DATA_CACHE_POLICY(0));
-	amdgpu_ring_write(ring, lower_32_bits(csa_addr + offsetof(struct v9_gfx_meta_data, de_payload)));
-	amdgpu_ring_write(ring, upper_32_bits(csa_addr + offsetof(struct v9_gfx_meta_data, de_payload)));
-	amdgpu_ring_write_multiple(ring, (void *)&de_payload, sizeof(de_payload) >> 2);
+	amdgpu_ring_write(ring, lower_32_bits(de_payload_gpu_addr));
+	amdgpu_ring_write(ring, upper_32_bits(de_payload_gpu_addr));
+
+	amdgpu_ring_ib_on_emit_de(ring);
+	if (resume)
+		amdgpu_ring_write_multiple(ring, de_payload_cpu_addr,
+					   sizeof(de_payload) >> 2);
+	else
+		amdgpu_ring_write_multiple(ring, (void *)&de_payload,
+					   sizeof(de_payload) >> 2);
 }
 
 static void gfx_v9_0_ring_emit_frame_cntl(struct amdgpu_ring *ring, bool start,
@@ -5616,8 +5810,9 @@ static void gfx_v9_ring_emit_cntxcntl(struct amdgpu_ring *ring, uint32_t flags)
 {
 	uint32_t dw2 = 0;
 
-	if (amdgpu_sriov_vf(ring->adev))
-		gfx_v9_0_ring_emit_ce_meta(ring);
+	gfx_v9_0_ring_emit_ce_meta(ring,
+				   (!amdgpu_sriov_vf(ring->adev) &&
+				   flags & AMDGPU_IB_PREEMPTED) ? true : false);
 
 	dw2 |= 0x80000000; /* set load_enable otherwise this package is just NOPs */
 	if (flags & AMDGPU_HAVE_CTX_SWITCH) {
@@ -5644,31 +5839,21 @@ static void gfx_v9_ring_emit_cntxcntl(struct amdgpu_ring *ring, uint32_t flags)
 	amdgpu_ring_write(ring, 0);
 }
 
-static unsigned gfx_v9_0_ring_emit_init_cond_exec(struct amdgpu_ring *ring)
+static unsigned gfx_v9_0_ring_emit_init_cond_exec(struct amdgpu_ring *ring,
+						  uint64_t addr)
 {
 	unsigned ret;
 	amdgpu_ring_write(ring, PACKET3(PACKET3_COND_EXEC, 3));
-	amdgpu_ring_write(ring, lower_32_bits(ring->cond_exe_gpu_addr));
-	amdgpu_ring_write(ring, upper_32_bits(ring->cond_exe_gpu_addr));
-	amdgpu_ring_write(ring, 0); /* discard following DWs if *cond_exec_gpu_addr==0 */
+	amdgpu_ring_write(ring, lower_32_bits(addr));
+	amdgpu_ring_write(ring, upper_32_bits(addr));
+	/* discard following DWs if *cond_exec_gpu_addr==0 */
+	amdgpu_ring_write(ring, 0);
 	ret = ring->wptr & ring->buf_mask;
-	amdgpu_ring_write(ring, 0x55aa55aa); /* patch dummy value later */
+	/* patch dummy value later */
+	amdgpu_ring_write(ring, 0);
 	return ret;
 }
 
-static void gfx_v9_0_ring_emit_patch_cond_exec(struct amdgpu_ring *ring, unsigned offset)
-{
-	unsigned cur;
-	BUG_ON(offset > ring->buf_mask);
-	BUG_ON(ring->ring[offset] != 0x55aa55aa);
-
-	cur = (ring->wptr & ring->buf_mask) - 1;
-	if (likely(cur > offset))
-		ring->ring[offset] = cur - offset;
-	else
-		ring->ring[offset] = (ring->ring_size>>2) - offset + cur;
-}
-
 static void gfx_v9_0_ring_emit_rreg(struct amdgpu_ring *ring, uint32_t reg,
 				    uint32_t reg_val_offs)
 {
@@ -5741,7 +5926,9 @@ static void gfx_v9_0_ring_soft_recovery(struct amdgpu_ring *ring, unsigned vmid)
 	value = REG_SET_FIELD(value, SQ_CMD, MODE, 0x01);
 	value = REG_SET_FIELD(value, SQ_CMD, CHECK_VMID, 1);
 	value = REG_SET_FIELD(value, SQ_CMD, VM_ID, vmid);
+	amdgpu_gfx_rlc_enter_safe_mode(adev, 0);
 	WREG32_SOC15(GC, 0, mmSQ_CMD, value);
+	amdgpu_gfx_rlc_exit_safe_mode(adev, 0);
 }
 
 static void gfx_v9_0_set_gfx_eop_interrupt_state(struct amdgpu_device *adev,
@@ -5796,33 +5983,111 @@ static void gfx_v9_0_set_compute_eop_interrupt_state(struct amdgpu_device *adev,
 
 	switch (state) {
 	case AMDGPU_IRQ_STATE_DISABLE:
-		mec_int_cntl = RREG32(mec_int_cntl_reg);
+		mec_int_cntl = RREG32_SOC15_IP(GC,mec_int_cntl_reg);
 		mec_int_cntl = REG_SET_FIELD(mec_int_cntl, CP_ME1_PIPE0_INT_CNTL,
 					     TIME_STAMP_INT_ENABLE, 0);
-		WREG32(mec_int_cntl_reg, mec_int_cntl);
+		WREG32_SOC15_IP(GC, mec_int_cntl_reg, mec_int_cntl);
 		break;
 	case AMDGPU_IRQ_STATE_ENABLE:
-		mec_int_cntl = RREG32(mec_int_cntl_reg);
+		mec_int_cntl = RREG32_SOC15_IP(GC, mec_int_cntl_reg);
 		mec_int_cntl = REG_SET_FIELD(mec_int_cntl, CP_ME1_PIPE0_INT_CNTL,
 					     TIME_STAMP_INT_ENABLE, 1);
-		WREG32(mec_int_cntl_reg, mec_int_cntl);
+		WREG32_SOC15_IP(GC, mec_int_cntl_reg, mec_int_cntl);
 		break;
 	default:
 		break;
 	}
 }
 
+static u32 gfx_v9_0_get_cpc_int_cntl(struct amdgpu_device *adev,
+				     int me, int pipe)
+{
+	/*
+	 * amdgpu controls only the first MEC. That's why this function only
+	 * handles the setting of interrupts for this specific MEC. All other
+	 * pipes' interrupts are set by amdkfd.
+	 */
+	if (me != 1)
+		return 0;
+
+	switch (pipe) {
+	case 0:
+		return SOC15_REG_OFFSET(GC, 0, mmCP_ME1_PIPE0_INT_CNTL);
+	case 1:
+		return SOC15_REG_OFFSET(GC, 0, mmCP_ME1_PIPE1_INT_CNTL);
+	case 2:
+		return SOC15_REG_OFFSET(GC, 0, mmCP_ME1_PIPE2_INT_CNTL);
+	case 3:
+		return SOC15_REG_OFFSET(GC, 0, mmCP_ME1_PIPE3_INT_CNTL);
+	default:
+		return 0;
+	}
+}
+
 static int gfx_v9_0_set_priv_reg_fault_state(struct amdgpu_device *adev,
 					     struct amdgpu_irq_src *source,
 					     unsigned type,
 					     enum amdgpu_interrupt_state state)
 {
+	u32 cp_int_cntl_reg, cp_int_cntl;
+	int i, j;
+
 	switch (state) {
 	case AMDGPU_IRQ_STATE_DISABLE:
 	case AMDGPU_IRQ_STATE_ENABLE:
 		WREG32_FIELD15(GC, 0, CP_INT_CNTL_RING0,
 			       PRIV_REG_INT_ENABLE,
 			       state == AMDGPU_IRQ_STATE_ENABLE ? 1 : 0);
+		for (i = 0; i < adev->gfx.mec.num_mec; i++) {
+			for (j = 0; j < adev->gfx.mec.num_pipe_per_mec; j++) {
+				/* MECs start at 1 */
+				cp_int_cntl_reg = gfx_v9_0_get_cpc_int_cntl(adev, i + 1, j);
+
+				if (cp_int_cntl_reg) {
+					cp_int_cntl = RREG32_SOC15_IP(GC, cp_int_cntl_reg);
+					cp_int_cntl = REG_SET_FIELD(cp_int_cntl, CP_ME1_PIPE0_INT_CNTL,
+								    PRIV_REG_INT_ENABLE,
+								    state == AMDGPU_IRQ_STATE_ENABLE ? 1 : 0);
+					WREG32_SOC15_IP(GC, cp_int_cntl_reg, cp_int_cntl);
+				}
+			}
+		}
+		break;
+	default:
+		break;
+	}
+
+	return 0;
+}
+
+static int gfx_v9_0_set_bad_op_fault_state(struct amdgpu_device *adev,
+					   struct amdgpu_irq_src *source,
+					   unsigned type,
+					   enum amdgpu_interrupt_state state)
+{
+	u32 cp_int_cntl_reg, cp_int_cntl;
+	int i, j;
+
+	switch (state) {
+	case AMDGPU_IRQ_STATE_DISABLE:
+	case AMDGPU_IRQ_STATE_ENABLE:
+		WREG32_FIELD15(GC, 0, CP_INT_CNTL_RING0,
+			       OPCODE_ERROR_INT_ENABLE,
+			       state == AMDGPU_IRQ_STATE_ENABLE ? 1 : 0);
+		for (i = 0; i < adev->gfx.mec.num_mec; i++) {
+			for (j = 0; j < adev->gfx.mec.num_pipe_per_mec; j++) {
+				/* MECs start at 1 */
+				cp_int_cntl_reg = gfx_v9_0_get_cpc_int_cntl(adev, i + 1, j);
+
+				if (cp_int_cntl_reg) {
+					cp_int_cntl = RREG32_SOC15_IP(GC, cp_int_cntl_reg);
+					cp_int_cntl = REG_SET_FIELD(cp_int_cntl, CP_ME1_PIPE0_INT_CNTL,
+								    OPCODE_ERROR_INT_ENABLE,
+								    state == AMDGPU_IRQ_STATE_ENABLE ? 1 : 0);
+					WREG32_SOC15_IP(GC, cp_int_cntl_reg, cp_int_cntl);
+				}
+			}
+		}
 		break;
 	default:
 		break;
@@ -5943,7 +6208,15 @@ static int gfx_v9_0_eop_irq(struct amdgpu_device *adev,
 
 	switch (me_id) {
 	case 0:
-		amdgpu_fence_process(&adev->gfx.gfx_ring[0]);
+		if (adev->gfx.num_gfx_rings) {
+			if (!adev->gfx.mcbp) {
+				amdgpu_fence_process(&adev->gfx.gfx_ring[0]);
+			} else if (!amdgpu_mcbp_handle_trailing_fence_irq(&adev->gfx.muxer)) {
+				/* Fence signals are handled on the software rings*/
+				for (i = 0; i < GFX9_NUM_SW_GFX_RINGS; i++)
+					amdgpu_fence_process(&adev->gfx.sw_gfx_ring[i]);
+			}
+		}
 		break;
 	case 1:
 	case 2:
@@ -5996,6 +6269,15 @@ static int gfx_v9_0_priv_reg_irq(struct amdgpu_device *adev,
 	return 0;
 }
 
+static int gfx_v9_0_bad_op_irq(struct amdgpu_device *adev,
+			       struct amdgpu_irq_src *source,
+			       struct amdgpu_iv_entry *entry)
+{
+	DRM_ERROR("Illegal opcode in command stream\n");
+	gfx_v9_0_fault(adev, entry);
+	return 0;
+}
+
 static int gfx_v9_0_priv_inst_irq(struct amdgpu_device *adev,
 				  struct amdgpu_irq_src *source,
 				  struct amdgpu_iv_entry *entry)
@@ -6448,7 +6730,7 @@ static const struct soc15_ras_field_entry gfx_v9_0_ras_fields[] = {
 };
 
 static int gfx_v9_0_ras_error_inject(struct amdgpu_device *adev,
-				     void *inject_if)
+				     void *inject_if, uint32_t instance_mask)
 {
 	struct ras_inject_if *info = (struct ras_inject_if *)inject_if;
 	int ret;
@@ -6487,13 +6769,13 @@ static int gfx_v9_0_ras_error_inject(struct amdgpu_device *adev,
 	block_info.value = info->value;
 
 	mutex_lock(&adev->grbm_idx_mutex);
-	ret = psp_ras_trigger_error(&adev->psp, &block_info);
+	ret = psp_ras_trigger_error(&adev->psp, &block_info, instance_mask);
 	mutex_unlock(&adev->grbm_idx_mutex);
 
 	return ret;
 }
 
-static const char *vml2_mems[] = {
+static const char * const vml2_mems[] = {
 	"UTC_VML2_BANK_CACHE_0_BIGK_MEM0",
 	"UTC_VML2_BANK_CACHE_0_BIGK_MEM1",
 	"UTC_VML2_BANK_CACHE_0_4K_MEM0",
@@ -6512,7 +6794,7 @@ static const char *vml2_mems[] = {
 	"UTC_VML2_BANK_CACHE_3_4K_MEM1",
 };
 
-static const char *vml2_walker_mems[] = {
+static const char * const vml2_walker_mems[] = {
 	"UTC_VML2_CACHE_PDE0_MEM0",
 	"UTC_VML2_CACHE_PDE0_MEM1",
 	"UTC_VML2_CACHE_PDE1_MEM0",
@@ -6522,7 +6804,7 @@ static const char *vml2_walker_mems[] = {
 	"UTC_VML2_RDIF_LOG_FIFO",
 };
 
-static const char *atc_l2_cache_2m_mems[] = {
+static const char * const atc_l2_cache_2m_mems[] = {
 	"UTC_ATCL2_CACHE_2M_BANK0_WAY0_MEM",
 	"UTC_ATCL2_CACHE_2M_BANK0_WAY1_MEM",
 	"UTC_ATCL2_CACHE_2M_BANK1_WAY0_MEM",
@@ -6715,7 +6997,7 @@ static void gfx_v9_0_reset_ras_error_count(struct amdgpu_device *adev)
 	for (i = 0; i < ARRAY_SIZE(gfx_v9_0_edc_counter_regs); i++) {
 		for (j = 0; j < gfx_v9_0_edc_counter_regs[i].se_num; j++) {
 			for (k = 0; k < gfx_v9_0_edc_counter_regs[i].instance; k++) {
-				gfx_v9_0_select_se_sh(adev, j, 0x0, k);
+				amdgpu_gfx_select_se_sh(adev, j, 0x0, k, 0);
 				RREG32(SOC15_REG_ENTRY_OFFSET(gfx_v9_0_edc_counter_regs[i]));
 			}
 		}
@@ -6758,7 +7040,7 @@ static void gfx_v9_0_reset_ras_error_count(struct amdgpu_device *adev)
 	WREG32_SOC15(GC, 0, mmATC_L2_CACHE_4K_EDC_INDEX, 255);
 }
 
-static int gfx_v9_0_query_ras_error_count(struct amdgpu_device *adev,
+static void gfx_v9_0_query_ras_error_count(struct amdgpu_device *adev,
 					  void *ras_error_status)
 {
 	struct ras_err_data *err_data = (struct ras_err_data *)ras_error_status;
@@ -6767,7 +7049,7 @@ static int gfx_v9_0_query_ras_error_count(struct amdgpu_device *adev,
 	uint32_t reg_value;
 
 	if (!amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__GFX))
-		return -EINVAL;
+		return;
 
 	err_data->ue_count = 0;
 	err_data->ce_count = 0;
@@ -6777,7 +7059,7 @@ static int gfx_v9_0_query_ras_error_count(struct amdgpu_device *adev,
 	for (i = 0; i < ARRAY_SIZE(gfx_v9_0_edc_counter_regs); i++) {
 		for (j = 0; j < gfx_v9_0_edc_counter_regs[i].se_num; j++) {
 			for (k = 0; k < gfx_v9_0_edc_counter_regs[i].instance; k++) {
-				gfx_v9_0_select_se_sh(adev, j, 0, k);
+				amdgpu_gfx_select_se_sh(adev, j, 0, k, 0);
 				reg_value =
 					RREG32(SOC15_REG_ENTRY_OFFSET(gfx_v9_0_edc_counter_regs[i]));
 				if (reg_value)
@@ -6792,12 +7074,10 @@ static int gfx_v9_0_query_ras_error_count(struct amdgpu_device *adev,
 	err_data->ce_count += sec_count;
 	err_data->ue_count += ded_count;
 
-	gfx_v9_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+	amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
 	mutex_unlock(&adev->grbm_idx_mutex);
 
 	gfx_v9_0_query_utc_edc_status(adev, err_data);
-
-	return 0;
 }
 
 static void gfx_v9_0_emit_mem_sync(struct amdgpu_ring *ring)
@@ -6878,6 +7158,230 @@ static void gfx_v9_0_emit_wave_limit(struct amdgpu_ring *ring, bool enable)
 	}
 }
 
+static void gfx_v9_ring_insert_nop(struct amdgpu_ring *ring, uint32_t num_nop)
+{
+	/* Header itself is a NOP packet */
+	if (num_nop == 1) {
+		amdgpu_ring_write(ring, ring->funcs->nop);
+		return;
+	}
+
+	/* Max HW optimization till 0x3ffe, followed by remaining one NOP at a time*/
+	amdgpu_ring_write(ring, PACKET3(PACKET3_NOP, min(num_nop - 2, 0x3ffe)));
+
+	/* Header is at index 0, followed by num_nops - 1 NOP packet's */
+	amdgpu_ring_insert_nop(ring, num_nop - 1);
+}
+
+static int gfx_v9_0_reset_kcq(struct amdgpu_ring *ring,
+			      unsigned int vmid,
+			      struct amdgpu_fence *timedout_fence)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct amdgpu_kiq *kiq = &adev->gfx.kiq[0];
+	struct amdgpu_ring *kiq_ring = &kiq->ring;
+	unsigned long flags;
+	int i, r;
+
+	if (!kiq->pmf || !kiq->pmf->kiq_unmap_queues)
+		return -EINVAL;
+
+	amdgpu_ring_reset_helper_begin(ring, timedout_fence);
+
+	spin_lock_irqsave(&kiq->ring_lock, flags);
+
+	if (amdgpu_ring_alloc(kiq_ring, kiq->pmf->unmap_queues_size)) {
+		spin_unlock_irqrestore(&kiq->ring_lock, flags);
+		return -ENOMEM;
+	}
+
+	kiq->pmf->kiq_unmap_queues(kiq_ring, ring, RESET_QUEUES,
+				   0, 0);
+	amdgpu_ring_commit(kiq_ring);
+
+	spin_unlock_irqrestore(&kiq->ring_lock, flags);
+
+	r = amdgpu_ring_test_ring(kiq_ring);
+	if (r)
+		return r;
+
+	/* make sure dequeue is complete*/
+	amdgpu_gfx_rlc_enter_safe_mode(adev, 0);
+	mutex_lock(&adev->srbm_mutex);
+	soc15_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0, 0);
+	for (i = 0; i < adev->usec_timeout; i++) {
+		if (!(RREG32_SOC15(GC, 0, mmCP_HQD_ACTIVE) & 1))
+			break;
+		udelay(1);
+	}
+	if (i >= adev->usec_timeout)
+		r = -ETIMEDOUT;
+	soc15_grbm_select(adev, 0, 0, 0, 0, 0);
+	mutex_unlock(&adev->srbm_mutex);
+	amdgpu_gfx_rlc_exit_safe_mode(adev, 0);
+	if (r) {
+		dev_err(adev->dev, "fail to wait on hqd deactive\n");
+		return r;
+	}
+
+	r = gfx_v9_0_kcq_init_queue(ring, true);
+	if (r) {
+		dev_err(adev->dev, "fail to init kcq\n");
+		return r;
+	}
+	spin_lock_irqsave(&kiq->ring_lock, flags);
+	r = amdgpu_ring_alloc(kiq_ring, kiq->pmf->map_queues_size);
+	if (r) {
+		spin_unlock_irqrestore(&kiq->ring_lock, flags);
+		return -ENOMEM;
+	}
+	kiq->pmf->kiq_map_queues(kiq_ring, ring);
+	amdgpu_ring_commit(kiq_ring);
+	r = amdgpu_ring_test_ring(kiq_ring);
+	spin_unlock_irqrestore(&kiq->ring_lock, flags);
+	if (r) {
+		DRM_ERROR("fail to remap queue\n");
+		return r;
+	}
+	return amdgpu_ring_reset_helper_end(ring, timedout_fence);
+}
+
+static void gfx_v9_ip_print(struct amdgpu_ip_block *ip_block, struct drm_printer *p)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	uint32_t i, j, k, reg, index = 0;
+	uint32_t reg_count = ARRAY_SIZE(gc_reg_list_9);
+
+	if (!adev->gfx.ip_dump_core)
+		return;
+
+	for (i = 0; i < reg_count; i++)
+		drm_printf(p, "%-50s \t 0x%08x\n",
+			   gc_reg_list_9[i].reg_name,
+			   adev->gfx.ip_dump_core[i]);
+
+	/* print compute queue registers for all instances */
+	if (!adev->gfx.ip_dump_compute_queues)
+		return;
+
+	reg_count = ARRAY_SIZE(gc_cp_reg_list_9);
+	drm_printf(p, "\nnum_mec: %d num_pipe: %d num_queue: %d\n",
+		   adev->gfx.mec.num_mec,
+		   adev->gfx.mec.num_pipe_per_mec,
+		   adev->gfx.mec.num_queue_per_pipe);
+
+	for (i = 0; i < adev->gfx.mec.num_mec; i++) {
+		for (j = 0; j < adev->gfx.mec.num_pipe_per_mec; j++) {
+			for (k = 0; k < adev->gfx.mec.num_queue_per_pipe; k++) {
+				drm_printf(p, "\nmec %d, pipe %d, queue %d\n", i, j, k);
+				for (reg = 0; reg < reg_count; reg++) {
+					if (i && gc_cp_reg_list_9[reg].reg_offset == mmCP_MEC_ME1_HEADER_DUMP)
+						drm_printf(p, "%-50s \t 0x%08x\n",
+							   "mmCP_MEC_ME2_HEADER_DUMP",
+							   adev->gfx.ip_dump_compute_queues[index + reg]);
+					else
+						drm_printf(p, "%-50s \t 0x%08x\n",
+							   gc_cp_reg_list_9[reg].reg_name,
+							   adev->gfx.ip_dump_compute_queues[index + reg]);
+				}
+				index += reg_count;
+			}
+		}
+	}
+
+}
+
+static void gfx_v9_ip_dump(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	uint32_t i, j, k, reg, index = 0;
+	uint32_t reg_count = ARRAY_SIZE(gc_reg_list_9);
+
+	if (!adev->gfx.ip_dump_core || !adev->gfx.num_gfx_rings)
+		return;
+
+	amdgpu_gfx_off_ctrl(adev, false);
+	for (i = 0; i < reg_count; i++)
+		adev->gfx.ip_dump_core[i] = RREG32(SOC15_REG_ENTRY_OFFSET(gc_reg_list_9[i]));
+	amdgpu_gfx_off_ctrl(adev, true);
+
+	/* dump compute queue registers for all instances */
+	if (!adev->gfx.ip_dump_compute_queues)
+		return;
+
+	reg_count = ARRAY_SIZE(gc_cp_reg_list_9);
+	amdgpu_gfx_off_ctrl(adev, false);
+	mutex_lock(&adev->srbm_mutex);
+	for (i = 0; i < adev->gfx.mec.num_mec; i++) {
+		for (j = 0; j < adev->gfx.mec.num_pipe_per_mec; j++) {
+			for (k = 0; k < adev->gfx.mec.num_queue_per_pipe; k++) {
+				/* ME0 is for GFX so start from 1 for CP */
+				soc15_grbm_select(adev, 1 + i, j, k, 0, 0);
+
+				for (reg = 0; reg < reg_count; reg++) {
+					if (i && gc_cp_reg_list_9[reg].reg_offset == mmCP_MEC_ME1_HEADER_DUMP)
+						adev->gfx.ip_dump_compute_queues[index + reg] =
+							RREG32(SOC15_REG_OFFSET(GC, 0, mmCP_MEC_ME2_HEADER_DUMP));
+					else
+						adev->gfx.ip_dump_compute_queues[index + reg] =
+							RREG32(SOC15_REG_ENTRY_OFFSET(
+								       gc_cp_reg_list_9[reg]));
+				}
+				index += reg_count;
+			}
+		}
+	}
+	soc15_grbm_select(adev, 0, 0, 0, 0, 0);
+	mutex_unlock(&adev->srbm_mutex);
+	amdgpu_gfx_off_ctrl(adev, true);
+
+}
+
+static void gfx_v9_0_ring_emit_cleaner_shader(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	/* Emit the cleaner shader */
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 2))
+		amdgpu_ring_write(ring, PACKET3(PACKET3_RUN_CLEANER_SHADER, 0));
+	else
+		amdgpu_ring_write(ring, PACKET3(PACKET3_RUN_CLEANER_SHADER_9_0, 0));
+
+	amdgpu_ring_write(ring, 0);  /* RESERVED field, programmed to zero */
+}
+
+static void gfx_v9_0_ring_begin_use_compute(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct amdgpu_ip_block *gfx_block =
+		amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_GFX);
+
+	amdgpu_gfx_enforce_isolation_ring_begin_use(ring);
+
+	/* Raven and PCO APUs seem to have stability issues
+	 * with compute and gfxoff and gfx pg.  Disable gfx pg during
+	 * submission and allow again afterwards.
+	 */
+	if (gfx_block && amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 1, 0))
+		gfx_v9_0_set_powergating_state(gfx_block, AMD_PG_STATE_UNGATE);
+}
+
+static void gfx_v9_0_ring_end_use_compute(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct amdgpu_ip_block *gfx_block =
+		amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_GFX);
+
+	/* Raven and PCO APUs seem to have stability issues
+	 * with compute and gfxoff and gfx pg.  Disable gfx pg during
+	 * submission and allow again afterwards.
+	 */
+	if (gfx_block && amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 1, 0))
+		gfx_v9_0_set_powergating_state(gfx_block, AMD_PG_STATE_GATE);
+
+	amdgpu_gfx_enforce_isolation_ring_end_use(ring);
+}
+
 static const struct amd_ip_funcs gfx_v9_0_ip_funcs = {
 	.name = "gfx_v9_0",
 	.early_init = gfx_v9_0_early_init,
@@ -6894,6 +7398,8 @@ static const struct amd_ip_funcs gfx_v9_0_ip_funcs = {
 	.set_clockgating_state = gfx_v9_0_set_clockgating_state,
 	.set_powergating_state = gfx_v9_0_set_powergating_state,
 	.get_clockgating_state = gfx_v9_0_get_clockgating_state,
+	.dump_ip_state = gfx_v9_ip_dump,
+	.print_ip_state = gfx_v9_ip_print,
 };
 
 static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_gfx = {
@@ -6901,7 +7407,7 @@ static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_gfx = {
 	.align_mask = 0xff,
 	.nop = PACKET3(PACKET3_NOP, 0x3FFF),
 	.support_64bit_ptrs = true,
-	.vmhub = AMDGPU_GFXHUB_0,
+	.secure_submission_supported = true,
 	.get_rptr = gfx_v9_0_ring_get_rptr_gfx,
 	.get_wptr = gfx_v9_0_ring_get_wptr_gfx,
 	.set_wptr = gfx_v9_0_ring_set_wptr_gfx,
@@ -6925,7 +7431,65 @@ static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_gfx = {
 		5 + /* HDP_INVL */
 		8 + 8 + /* FENCE x2 */
 		2 + /* SWITCH_BUFFER */
-		7, /* gfx_v9_0_emit_mem_sync */
+		7 + /* gfx_v9_0_emit_mem_sync */
+		2, /* gfx_v9_0_ring_emit_cleaner_shader */
+	.emit_ib_size =	4, /* gfx_v9_0_ring_emit_ib_gfx */
+	.emit_ib = gfx_v9_0_ring_emit_ib_gfx,
+	.emit_fence = gfx_v9_0_ring_emit_fence,
+	.emit_pipeline_sync = gfx_v9_0_ring_emit_pipeline_sync,
+	.emit_vm_flush = gfx_v9_0_ring_emit_vm_flush,
+	.emit_gds_switch = gfx_v9_0_ring_emit_gds_switch,
+	.emit_hdp_flush = gfx_v9_0_ring_emit_hdp_flush,
+	.test_ring = gfx_v9_0_ring_test_ring,
+	.insert_nop = gfx_v9_ring_insert_nop,
+	.pad_ib = amdgpu_ring_generic_pad_ib,
+	.emit_switch_buffer = gfx_v9_ring_emit_sb,
+	.emit_cntxcntl = gfx_v9_ring_emit_cntxcntl,
+	.init_cond_exec = gfx_v9_0_ring_emit_init_cond_exec,
+	.preempt_ib = gfx_v9_0_ring_preempt_ib,
+	.emit_frame_cntl = gfx_v9_0_ring_emit_frame_cntl,
+	.emit_wreg = gfx_v9_0_ring_emit_wreg,
+	.emit_reg_wait = gfx_v9_0_ring_emit_reg_wait,
+	.emit_reg_write_reg_wait = gfx_v9_0_ring_emit_reg_write_reg_wait,
+	.soft_recovery = gfx_v9_0_ring_soft_recovery,
+	.emit_mem_sync = gfx_v9_0_emit_mem_sync,
+	.emit_cleaner_shader = gfx_v9_0_ring_emit_cleaner_shader,
+	.begin_use = amdgpu_gfx_enforce_isolation_ring_begin_use,
+	.end_use = amdgpu_gfx_enforce_isolation_ring_end_use,
+};
+
+static const struct amdgpu_ring_funcs gfx_v9_0_sw_ring_funcs_gfx = {
+	.type = AMDGPU_RING_TYPE_GFX,
+	.align_mask = 0xff,
+	.nop = PACKET3(PACKET3_NOP, 0x3FFF),
+	.support_64bit_ptrs = true,
+	.secure_submission_supported = true,
+	.get_rptr = amdgpu_sw_ring_get_rptr_gfx,
+	.get_wptr = amdgpu_sw_ring_get_wptr_gfx,
+	.set_wptr = amdgpu_sw_ring_set_wptr_gfx,
+	.emit_frame_size = /* totally 242 maximum if 16 IBs */
+		5 +  /* COND_EXEC */
+		7 +  /* PIPELINE_SYNC */
+		SOC15_FLUSH_GPU_TLB_NUM_WREG * 5 +
+		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 7 +
+		2 + /* VM_FLUSH */
+		8 +  /* FENCE for VM_FLUSH */
+		20 + /* GDS switch */
+		4 + /* double SWITCH_BUFFER,
+		     * the first COND_EXEC jump to the place just
+		     * prior to this double SWITCH_BUFFER
+		     */
+		5 + /* COND_EXEC */
+		7 +	 /*	HDP_flush */
+		4 +	 /*	VGT_flush */
+		14 + /*	CE_META */
+		31 + /*	DE_META */
+		3 + /* CNTX_CTRL */
+		5 + /* HDP_INVL */
+		8 + 8 + /* FENCE x2 */
+		2 + /* SWITCH_BUFFER */
+		7 + /* gfx_v9_0_emit_mem_sync */
+		2, /* gfx_v9_0_ring_emit_cleaner_shader */
 	.emit_ib_size =	4, /* gfx_v9_0_ring_emit_ib_gfx */
 	.emit_ib = gfx_v9_0_ring_emit_ib_gfx,
 	.emit_fence = gfx_v9_0_ring_emit_fence,
@@ -6935,18 +7499,23 @@ static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_gfx = {
 	.emit_hdp_flush = gfx_v9_0_ring_emit_hdp_flush,
 	.test_ring = gfx_v9_0_ring_test_ring,
 	.test_ib = gfx_v9_0_ring_test_ib,
-	.insert_nop = amdgpu_ring_insert_nop,
+	.insert_nop = gfx_v9_ring_insert_nop,
 	.pad_ib = amdgpu_ring_generic_pad_ib,
 	.emit_switch_buffer = gfx_v9_ring_emit_sb,
 	.emit_cntxcntl = gfx_v9_ring_emit_cntxcntl,
 	.init_cond_exec = gfx_v9_0_ring_emit_init_cond_exec,
-	.patch_cond_exec = gfx_v9_0_ring_emit_patch_cond_exec,
 	.emit_frame_cntl = gfx_v9_0_ring_emit_frame_cntl,
 	.emit_wreg = gfx_v9_0_ring_emit_wreg,
 	.emit_reg_wait = gfx_v9_0_ring_emit_reg_wait,
 	.emit_reg_write_reg_wait = gfx_v9_0_ring_emit_reg_write_reg_wait,
 	.soft_recovery = gfx_v9_0_ring_soft_recovery,
 	.emit_mem_sync = gfx_v9_0_emit_mem_sync,
+	.patch_cntl = gfx_v9_0_ring_patch_cntl,
+	.patch_de = gfx_v9_0_ring_patch_de_meta,
+	.patch_ce = gfx_v9_0_ring_patch_ce_meta,
+	.emit_cleaner_shader = gfx_v9_0_ring_emit_cleaner_shader,
+	.begin_use = amdgpu_gfx_enforce_isolation_ring_begin_use,
+	.end_use = amdgpu_gfx_enforce_isolation_ring_end_use,
 };
 
 static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_compute = {
@@ -6954,7 +7523,6 @@ static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_compute = {
 	.align_mask = 0xff,
 	.nop = PACKET3(PACKET3_NOP, 0x3FFF),
 	.support_64bit_ptrs = true,
-	.vmhub = AMDGPU_GFXHUB_0,
 	.get_rptr = gfx_v9_0_ring_get_rptr_compute,
 	.get_wptr = gfx_v9_0_ring_get_wptr_compute,
 	.set_wptr = gfx_v9_0_ring_set_wptr_compute,
@@ -6965,11 +7533,11 @@ static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_compute = {
 		7 + /* gfx_v9_0_ring_emit_pipeline_sync */
 		SOC15_FLUSH_GPU_TLB_NUM_WREG * 5 +
 		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 7 +
-		2 + /* gfx_v9_0_ring_emit_vm_flush */
 		8 + 8 + 8 + /* gfx_v9_0_ring_emit_fence x3 for user fence, vm fence */
 		7 + /* gfx_v9_0_emit_mem_sync */
 		5 + /* gfx_v9_0_emit_wave_limit for updating mmSPI_WCL_PIPE_PERCENT_GFX register */
-		15, /* for updating 3 mmSPI_WCL_PIPE_PERCENT_CS registers */
+		15 + /* for updating 3 mmSPI_WCL_PIPE_PERCENT_CS registers */
+		2, /* gfx_v9_0_ring_emit_cleaner_shader */
 	.emit_ib_size =	7, /* gfx_v9_0_ring_emit_ib_compute */
 	.emit_ib = gfx_v9_0_ring_emit_ib_compute,
 	.emit_fence = gfx_v9_0_ring_emit_fence,
@@ -6979,13 +7547,18 @@ static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_compute = {
 	.emit_hdp_flush = gfx_v9_0_ring_emit_hdp_flush,
 	.test_ring = gfx_v9_0_ring_test_ring,
 	.test_ib = gfx_v9_0_ring_test_ib,
-	.insert_nop = amdgpu_ring_insert_nop,
+	.insert_nop = gfx_v9_ring_insert_nop,
 	.pad_ib = amdgpu_ring_generic_pad_ib,
 	.emit_wreg = gfx_v9_0_ring_emit_wreg,
 	.emit_reg_wait = gfx_v9_0_ring_emit_reg_wait,
 	.emit_reg_write_reg_wait = gfx_v9_0_ring_emit_reg_write_reg_wait,
+	.soft_recovery = gfx_v9_0_ring_soft_recovery,
 	.emit_mem_sync = gfx_v9_0_emit_mem_sync,
 	.emit_wave_limit = gfx_v9_0_emit_wave_limit,
+	.reset = gfx_v9_0_reset_kcq,
+	.emit_cleaner_shader = gfx_v9_0_ring_emit_cleaner_shader,
+	.begin_use = gfx_v9_0_ring_begin_use_compute,
+	.end_use = gfx_v9_0_ring_end_use_compute,
 };
 
 static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_kiq = {
@@ -6993,7 +7566,6 @@ static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_kiq = {
 	.align_mask = 0xff,
 	.nop = PACKET3(PACKET3_NOP, 0x3FFF),
 	.support_64bit_ptrs = true,
-	.vmhub = AMDGPU_GFXHUB_0,
 	.get_rptr = gfx_v9_0_ring_get_rptr_compute,
 	.get_wptr = gfx_v9_0_ring_get_wptr_compute,
 	.set_wptr = gfx_v9_0_ring_set_wptr_compute,
@@ -7004,7 +7576,6 @@ static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_kiq = {
 		7 + /* gfx_v9_0_ring_emit_pipeline_sync */
 		SOC15_FLUSH_GPU_TLB_NUM_WREG * 5 +
 		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 7 +
-		2 + /* gfx_v9_0_ring_emit_vm_flush */
 		8 + 8 + 8, /* gfx_v9_0_ring_emit_fence_kiq x3 for user fence, vm fence */
 	.emit_ib_size =	7, /* gfx_v9_0_ring_emit_ib_compute */
 	.emit_fence = gfx_v9_0_ring_emit_fence_kiq,
@@ -7015,17 +7586,23 @@ static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_kiq = {
 	.emit_wreg = gfx_v9_0_ring_emit_wreg,
 	.emit_reg_wait = gfx_v9_0_ring_emit_reg_wait,
 	.emit_reg_write_reg_wait = gfx_v9_0_ring_emit_reg_write_reg_wait,
+	.emit_hdp_flush = gfx_v9_0_ring_emit_hdp_flush,
 };
 
 static void gfx_v9_0_set_ring_funcs(struct amdgpu_device *adev)
 {
 	int i;
 
-	adev->gfx.kiq.ring.funcs = &gfx_v9_0_ring_funcs_kiq;
+	adev->gfx.kiq[0].ring.funcs = &gfx_v9_0_ring_funcs_kiq;
 
 	for (i = 0; i < adev->gfx.num_gfx_rings; i++)
 		adev->gfx.gfx_ring[i].funcs = &gfx_v9_0_ring_funcs_gfx;
 
+	if (adev->gfx.mcbp && adev->gfx.num_gfx_rings) {
+		for (i = 0; i < GFX9_NUM_SW_GFX_RINGS; i++)
+			adev->gfx.sw_gfx_ring[i].funcs = &gfx_v9_0_sw_ring_funcs_gfx;
+	}
+
 	for (i = 0; i < adev->gfx.num_compute_rings; i++)
 		adev->gfx.compute_ring[i].funcs = &gfx_v9_0_ring_funcs_compute;
 }
@@ -7040,6 +7617,11 @@ static const struct amdgpu_irq_src_funcs gfx_v9_0_priv_reg_irq_funcs = {
 	.process = gfx_v9_0_priv_reg_irq,
 };
 
+static const struct amdgpu_irq_src_funcs gfx_v9_0_bad_op_irq_funcs = {
+	.set = gfx_v9_0_set_bad_op_fault_state,
+	.process = gfx_v9_0_bad_op_irq,
+};
+
 static const struct amdgpu_irq_src_funcs gfx_v9_0_priv_inst_irq_funcs = {
 	.set = gfx_v9_0_set_priv_inst_fault_state,
 	.process = gfx_v9_0_priv_inst_irq,
@@ -7059,6 +7641,9 @@ static void gfx_v9_0_set_irq_funcs(struct amdgpu_device *adev)
 	adev->gfx.priv_reg_irq.num_types = 1;
 	adev->gfx.priv_reg_irq.funcs = &gfx_v9_0_priv_reg_irq_funcs;
 
+	adev->gfx.bad_op_irq.num_types = 1;
+	adev->gfx.bad_op_irq.funcs = &gfx_v9_0_bad_op_irq_funcs;
+
 	adev->gfx.priv_inst_irq.num_types = 1;
 	adev->gfx.priv_inst_irq.funcs = &gfx_v9_0_priv_inst_irq_funcs;
 
@@ -7068,7 +7653,7 @@ static void gfx_v9_0_set_irq_funcs(struct amdgpu_device *adev)
 
 static void gfx_v9_0_set_rlc_funcs(struct amdgpu_device *adev)
 {
-	switch (adev->ip_versions[GC_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 	case IP_VERSION(9, 0, 1):
 	case IP_VERSION(9, 2, 1):
 	case IP_VERSION(9, 4, 0):
@@ -7087,7 +7672,7 @@ static void gfx_v9_0_set_rlc_funcs(struct amdgpu_device *adev)
 static void gfx_v9_0_set_gds_init(struct amdgpu_device *adev)
 {
 	/* init asci gds info */
-	switch (adev->ip_versions[GC_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 	case IP_VERSION(9, 0, 1):
 	case IP_VERSION(9, 2, 1):
 	case IP_VERSION(9, 4, 0):
@@ -7109,7 +7694,7 @@ static void gfx_v9_0_set_gds_init(struct amdgpu_device *adev)
 		break;
 	}
 
-	switch (adev->ip_versions[GC_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 	case IP_VERSION(9, 0, 1):
 	case IP_VERSION(9, 4, 0):
 		adev->gds.gds_compute_max_wave_id = 0x7ff;
@@ -7197,7 +7782,7 @@ static int gfx_v9_0_get_cu_info(struct amdgpu_device *adev,
 			mask = 1;
 			ao_bitmap = 0;
 			counter = 0;
-			gfx_v9_0_select_se_sh(adev, i, j, 0xffffffff);
+			amdgpu_gfx_select_se_sh(adev, i, j, 0xffffffff, 0);
 			gfx_v9_0_set_user_cu_inactive_bitmap(
 				adev, disable_masks[i * adev->gfx.config.max_sh_per_se + j]);
 			bitmap = gfx_v9_0_get_cu_active_bitmap(adev);
@@ -7214,7 +7799,7 @@ static int gfx_v9_0_get_cu_info(struct amdgpu_device *adev,
 			 *    SE6,SH0 --> bitmap[2][1]
 			 *    SE7,SH0 --> bitmap[3][1]
 			 */
-			cu_info->bitmap[i % 4][j + i / 4] = bitmap;
+			cu_info->bitmap[0][i % 4][j + i / 4] = bitmap;
 
 			for (k = 0; k < adev->gfx.config.max_cu_per_sh; k ++) {
 				if (bitmap & mask) {
@@ -7230,7 +7815,7 @@ static int gfx_v9_0_get_cu_info(struct amdgpu_device *adev,
 			cu_info->ao_cu_bitmap[i % 4][j + i / 4] = ao_bitmap;
 		}
 	}
-	gfx_v9_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+	amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
 	mutex_unlock(&adev->grbm_idx_mutex);
 
 	cu_info->number = active_cu_number;
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.h b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.h
index dfe8d4841f58..f9f6edc5e558 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.h
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.h
@@ -27,6 +27,6 @@
 extern const struct amdgpu_ip_block_version gfx_v9_0_ip_block;
 
 void gfx_v9_0_select_se_sh(struct amdgpu_device *adev, u32 se_num, u32 sh_num,
-			   u32 instance);
+			   u32 instance, int xcc_id);
 
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0_cleaner_shader.h b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0_cleaner_shader.h
new file mode 100644
index 000000000000..0b6bd09b7529
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0_cleaner_shader.h
@@ -0,0 +1,68 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2024 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+/* Define the cleaner shader gfx_9_0 */
+static const u32 __maybe_unused gfx_9_0_cleaner_shader_hex[] = {
+	/* Add the cleaner shader code here */
+};
+
+/* Define the cleaner shader gfx_9_4_2 */
+static const u32 gfx_9_4_2_cleaner_shader_hex[] = {
+	0xbf068100, 0xbf84003b,
+	0xbf8a0000, 0xb07c0000,
+	0xbe8200ff, 0x00000078,
+	0xbf110802, 0x7e000280,
+	0x7e020280, 0x7e040280,
+	0x7e060280, 0x7e080280,
+	0x7e0a0280, 0x7e0c0280,
+	0x7e0e0280, 0x80828802,
+	0xbe803202, 0xbf84fff5,
+	0xbf9c0000, 0xbe8200ff,
+	0x80000000, 0x86020102,
+	0xbf840011, 0xbefe00c1,
+	0xbeff00c1, 0xd28c0001,
+	0x0001007f, 0xd28d0001,
+	0x0002027e, 0x10020288,
+	0xbe8200bf, 0xbefc00c1,
+	0xd89c2000, 0x00020201,
+	0xd89c6040, 0x00040401,
+	0x320202ff, 0x00000400,
+	0x80828102, 0xbf84fff8,
+	0xbefc00ff, 0x0000005c,
+	0xbf800000, 0xbe802c80,
+	0xbe812c80, 0xbe822c80,
+	0xbe832c80, 0x80fc847c,
+	0xbf84fffa, 0xbee60080,
+	0xbee70080, 0xbeea0180,
+	0xbeec0180, 0xbeee0180,
+	0xbef00180, 0xbef20180,
+	0xbef40180, 0xbef60180,
+	0xbef80180, 0xbefa0180,
+	0xbf810000, 0xbf8d0001,
+	0xbefc00ff, 0x0000005c,
+	0xbf800000, 0xbe802c80,
+	0xbe812c80, 0xbe822c80,
+	0xbe832c80, 0x80fc847c,
+	0xbf84fffa, 0xbee60080,
+	0xbee70080, 0xbeea01ff,
+	0x000000ee, 0xbf810000,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4.c
index b4789dfc2bb9..6028afd81690 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4.c
@@ -27,7 +27,6 @@
 #include "amdgpu_gfx.h"
 #include "soc15.h"
 #include "soc15d.h"
-#include "amdgpu_atomfirmware.h"
 #include "amdgpu_pm.h"
 
 #include "gc/gc_9_4_1_offset.h"
@@ -863,7 +862,7 @@ static int gfx_v9_4_ras_error_count(struct amdgpu_device *adev,
 	return 0;
 }
 
-static int gfx_v9_4_query_ras_error_count(struct amdgpu_device *adev,
+static void gfx_v9_4_query_ras_error_count(struct amdgpu_device *adev,
 					  void *ras_error_status)
 {
 	struct ras_err_data *err_data = (struct ras_err_data *)ras_error_status;
@@ -872,7 +871,7 @@ static int gfx_v9_4_query_ras_error_count(struct amdgpu_device *adev,
 	uint32_t reg_value;
 
 	if (!amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__GFX))
-		return -EINVAL;
+		return;
 
 	err_data->ue_count = 0;
 	err_data->ce_count = 0;
@@ -903,7 +902,6 @@ static int gfx_v9_4_query_ras_error_count(struct amdgpu_device *adev,
 
 	gfx_v9_4_query_utc_edc_status(adev, err_data);
 
-	return 0;
 }
 
 static void gfx_v9_4_reset_ras_error_count(struct amdgpu_device *adev)
@@ -971,31 +969,9 @@ static void gfx_v9_4_reset_ras_error_count(struct amdgpu_device *adev)
 	WREG32_SOC15(GC, 0, mmATC_L2_CACHE_4K_DSM_INDEX, 255);
 }
 
-static int gfx_v9_4_ras_error_inject(struct amdgpu_device *adev,
-				     void *inject_if)
-{
-	struct ras_inject_if *info = (struct ras_inject_if *)inject_if;
-	int ret;
-	struct ta_ras_trigger_error_input block_info = { 0 };
-
-	if (!amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__GFX))
-		return -EINVAL;
-
-	block_info.block_id = amdgpu_ras_block_to_ta(info->head.block);
-	block_info.sub_block_index = info->head.sub_block_index;
-	block_info.inject_error_type = amdgpu_ras_error_to_ta(info->head.type);
-	block_info.address = info->address;
-	block_info.value = info->value;
-
-	mutex_lock(&adev->grbm_idx_mutex);
-	ret = psp_ras_trigger_error(&adev->psp, &block_info);
-	mutex_unlock(&adev->grbm_idx_mutex);
-
-	return ret;
-}
-
-static const struct soc15_reg_entry gfx_v9_4_ea_err_status_regs =
-	{ SOC15_REG_ENTRY(GC, 0, mmGCEA_ERR_STATUS), 0, 1, 32 };
+static const struct soc15_reg_entry gfx_v9_4_ea_err_status_regs = {
+	SOC15_REG_ENTRY(GC, 0, mmGCEA_ERR_STATUS), 0, 1, 32
+};
 
 static void gfx_v9_4_query_ras_error_status(struct amdgpu_device *adev)
 {
@@ -1029,11 +1005,15 @@ static void gfx_v9_4_query_ras_error_status(struct amdgpu_device *adev)
 	mutex_unlock(&adev->grbm_idx_mutex);
 }
 
-const struct amdgpu_gfx_ras_funcs gfx_v9_4_ras_funcs = {
-        .ras_late_init = amdgpu_gfx_ras_late_init,
-        .ras_fini = amdgpu_gfx_ras_fini,
-        .ras_error_inject = &gfx_v9_4_ras_error_inject,
-        .query_ras_error_count = &gfx_v9_4_query_ras_error_count,
-        .reset_ras_error_count = &gfx_v9_4_reset_ras_error_count,
-        .query_ras_error_status = &gfx_v9_4_query_ras_error_status,
+
+const struct amdgpu_ras_block_hw_ops  gfx_v9_4_ras_ops = {
+	.query_ras_error_count = &gfx_v9_4_query_ras_error_count,
+	.reset_ras_error_count = &gfx_v9_4_reset_ras_error_count,
+	.query_ras_error_status = &gfx_v9_4_query_ras_error_status,
+};
+
+struct amdgpu_gfx_ras gfx_v9_4_ras = {
+	.ras_block = {
+		.hw_ops = &gfx_v9_4_ras_ops,
+	},
 };
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4.h b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4.h
index bdd16b568021..ca520a767267 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4.h
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4.h
@@ -24,6 +24,6 @@
 #ifndef __GFX_V9_4_H__
 #define __GFX_V9_4_H__
 
-extern const struct amdgpu_gfx_ras_funcs gfx_v9_4_ras_funcs;
+extern struct amdgpu_gfx_ras gfx_v9_4_ras;
 
 #endif /* __GFX_V9_4_H__ */
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_2.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_2.c
index c4f37a161875..8058ea91ecaf 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_2.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_2.c
@@ -412,7 +412,7 @@ static int gfx_v9_4_2_run_shader(struct amdgpu_device *adev,
 	r = amdgpu_ib_schedule(ring, 1, ib, NULL, fence_ptr);
 	if (r) {
 		dev_err(adev->dev, "ib submit failed (%d).\n", r);
-		amdgpu_ib_free(adev, ib, NULL);
+		amdgpu_ib_free(ib, NULL);
 	}
 	return r;
 }
@@ -611,16 +611,16 @@ static int gfx_v9_4_2_do_sgprs_init(struct amdgpu_device *adev)
 	}
 
 disp2_failed:
-	amdgpu_ib_free(adev, &disp_ibs[2], NULL);
+	amdgpu_ib_free(&disp_ibs[2], NULL);
 	dma_fence_put(fences[2]);
 disp1_failed:
-	amdgpu_ib_free(adev, &disp_ibs[1], NULL);
+	amdgpu_ib_free(&disp_ibs[1], NULL);
 	dma_fence_put(fences[1]);
 disp0_failed:
-	amdgpu_ib_free(adev, &disp_ibs[0], NULL);
+	amdgpu_ib_free(&disp_ibs[0], NULL);
 	dma_fence_put(fences[0]);
 pro_end:
-	amdgpu_ib_free(adev, &wb_ib, NULL);
+	amdgpu_ib_free(&wb_ib, NULL);
 
 	if (r)
 		dev_info(adev->dev, "Init SGPRS Failed\n");
@@ -687,10 +687,10 @@ static int gfx_v9_4_2_do_vgprs_init(struct amdgpu_device *adev)
 	}
 
 disp_failed:
-	amdgpu_ib_free(adev, &disp_ib, NULL);
+	amdgpu_ib_free(&disp_ib, NULL);
 	dma_fence_put(fence);
 pro_end:
-	amdgpu_ib_free(adev, &wb_ib, NULL);
+	amdgpu_ib_free(&wb_ib, NULL);
 
 	if (r)
 		dev_info(adev->dev, "Init VGPRS Failed\n");
@@ -746,8 +746,18 @@ void gfx_v9_4_2_init_golden_registers(struct amdgpu_device *adev,
 			 die_id);
 		break;
 	}
+}
+
+void gfx_v9_4_2_init_sq(struct amdgpu_device *adev)
+{
+	uint32_t data;
 
-	return;
+	if (adev->gfx.mec_fw_version >= 98) {
+		adev->gmc.xnack_flags |= AMDGPU_GMC_XNACK_FLAG_CHAIN;
+		data = RREG32_SOC15(GC, 0, regSQ_CONFIG1);
+		data = REG_SET_FIELD(data, SQ_CONFIG1, DISABLE_XNACK_CHECK_IN_RETRY_DISABLE, 1);
+		WREG32_SOC15(GC, 0, regSQ_CONFIG1, data);
+	}
 }
 
 void gfx_v9_4_2_debug_trap_config_init(struct amdgpu_device *adev,
@@ -761,7 +771,7 @@ void gfx_v9_4_2_debug_trap_config_init(struct amdgpu_device *adev,
 
 	for (i = first_vmid; i < last_vmid; i++) {
 		data = 0;
-		soc15_grbm_select(adev, 0, 0, 0, i);
+		soc15_grbm_select(adev, 0, 0, 0, i, 0);
 		data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, TRAP_EN, 1);
 		data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, EXCP_EN, 0);
 		data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, EXCP_REPLACE,
@@ -769,15 +779,18 @@ void gfx_v9_4_2_debug_trap_config_init(struct amdgpu_device *adev,
 		WREG32(SOC15_REG_OFFSET(GC, 0, regSPI_GDBG_PER_VMID_CNTL), data);
 	}
 
-	soc15_grbm_select(adev, 0, 0, 0, 0);
+	soc15_grbm_select(adev, 0, 0, 0, 0, 0);
 	mutex_unlock(&adev->srbm_mutex);
+
+	WREG32(SOC15_REG_OFFSET(GC, 0, regSPI_GDBG_TRAP_DATA0), 0);
+	WREG32(SOC15_REG_OFFSET(GC, 0, regSPI_GDBG_TRAP_DATA1), 0);
 }
 
 void gfx_v9_4_2_set_power_brake_sequence(struct amdgpu_device *adev)
 {
 	u32 tmp;
 
-	gfx_v9_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+	gfx_v9_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
 
 	tmp = 0;
 	tmp = REG_SET_FIELD(tmp, GC_THROTTLE_CTRL, PATTERN_MODE, 1);
@@ -1545,8 +1558,8 @@ static void gfx_v9_4_2_log_utc_edc_count(struct amdgpu_device *adev,
 					 uint32_t ded_cnt)
 {
 	uint32_t bank, way, mem;
-	static const char *vml2_way_str[] = { "BIGK", "4K" };
-	static const char *utcl2_rounter_str[] = { "VMC", "APT" };
+	static const char * const vml2_way_str[] = { "BIGK", "4K" };
+	static const char * const utcl2_router_str[] = { "VMC", "APT" };
 
 	mem = instance % blk->num_mem_blocks;
 	way = (instance / blk->num_mem_blocks) % blk->num_ways;
@@ -1567,7 +1580,7 @@ static void gfx_v9_4_2_log_utc_edc_count(struct amdgpu_device *adev,
 		dev_info(
 			adev->dev,
 			"GFX SubBlock UTCL2_ROUTER_IFIF%d_GROUP0_%s, SED %d, DED %d\n",
-			bank, utcl2_rounter_str[mem], sec_cnt, ded_cnt);
+			bank, utcl2_router_str[mem], sec_cnt, ded_cnt);
 		break;
 	case ATC_L2_CACHE_2M:
 		dev_info(
@@ -1641,14 +1654,14 @@ static int gfx_v9_4_2_query_utc_edc_count(struct amdgpu_device *adev,
 	return 0;
 }
 
-static int gfx_v9_4_2_query_ras_error_count(struct amdgpu_device *adev,
+static void gfx_v9_4_2_query_ras_error_count(struct amdgpu_device *adev,
 					    void *ras_error_status)
 {
 	struct ras_err_data *err_data = (struct ras_err_data *)ras_error_status;
 	uint32_t sec_count = 0, ded_count = 0;
 
 	if (!amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__GFX))
-		return -EINVAL;
+		return;
 
 	err_data->ue_count = 0;
 	err_data->ce_count = 0;
@@ -1661,7 +1674,6 @@ static int gfx_v9_4_2_query_ras_error_count(struct amdgpu_device *adev,
 	err_data->ce_count += sec_count;
 	err_data->ue_count += ded_count;
 
-	return 0;
 }
 
 static void gfx_v9_4_2_reset_utc_err_status(struct amdgpu_device *adev)
@@ -1700,28 +1712,6 @@ static void gfx_v9_4_2_reset_ras_error_count(struct amdgpu_device *adev)
 	gfx_v9_4_2_query_utc_edc_count(adev, NULL, NULL);
 }
 
-static int gfx_v9_4_2_ras_error_inject(struct amdgpu_device *adev, void *inject_if)
-{
-	struct ras_inject_if *info = (struct ras_inject_if *)inject_if;
-	int ret;
-	struct ta_ras_trigger_error_input block_info = { 0 };
-
-	if (!amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__GFX))
-		return -EINVAL;
-
-	block_info.block_id = amdgpu_ras_block_to_ta(info->head.block);
-	block_info.sub_block_index = info->head.sub_block_index;
-	block_info.inject_error_type = amdgpu_ras_error_to_ta(info->head.type);
-	block_info.address = info->address;
-	block_info.value = info->value;
-
-	mutex_lock(&adev->grbm_idx_mutex);
-	ret = psp_ras_trigger_error(&adev->psp, &block_info);
-	mutex_unlock(&adev->grbm_idx_mutex);
-
-	return ret;
-}
-
 static void gfx_v9_4_2_query_ea_err_status(struct amdgpu_device *adev)
 {
 	uint32_t i, j;
@@ -1931,13 +1921,18 @@ static void gfx_v9_4_2_reset_sq_timeout_status(struct amdgpu_device *adev)
 	mutex_unlock(&adev->grbm_idx_mutex);
 }
 
-const struct amdgpu_gfx_ras_funcs gfx_v9_4_2_ras_funcs = {
-	.ras_late_init = amdgpu_gfx_ras_late_init,
-	.ras_fini = amdgpu_gfx_ras_fini,
-	.ras_error_inject = &gfx_v9_4_2_ras_error_inject,
-	.query_ras_error_count = &gfx_v9_4_2_query_ras_error_count,
-	.reset_ras_error_count = &gfx_v9_4_2_reset_ras_error_count,
-	.query_ras_error_status = &gfx_v9_4_2_query_ras_error_status,
-	.reset_ras_error_status = &gfx_v9_4_2_reset_ras_error_status,
+
+
+struct amdgpu_ras_block_hw_ops  gfx_v9_4_2_ras_ops = {
+		.query_ras_error_count = &gfx_v9_4_2_query_ras_error_count,
+		.reset_ras_error_count = &gfx_v9_4_2_reset_ras_error_count,
+		.query_ras_error_status = &gfx_v9_4_2_query_ras_error_status,
+		.reset_ras_error_status = &gfx_v9_4_2_reset_ras_error_status,
+};
+
+struct amdgpu_gfx_ras gfx_v9_4_2_ras = {
+	.ras_block = {
+		.hw_ops = &gfx_v9_4_2_ras_ops,
+	},
 	.enable_watchdog_timer = &gfx_v9_4_2_enable_watchdog_timer,
 };
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_2.h b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_2.h
index 6db1f88509af..a603724c1dfc 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_2.h
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_2.h
@@ -28,9 +28,10 @@ void gfx_v9_4_2_debug_trap_config_init(struct amdgpu_device *adev,
 				uint32_t first_vmid, uint32_t last_vmid);
 void gfx_v9_4_2_init_golden_registers(struct amdgpu_device *adev,
 				      uint32_t die_id);
+void gfx_v9_4_2_init_sq(struct amdgpu_device *adev);
 void gfx_v9_4_2_set_power_brake_sequence(struct amdgpu_device *adev);
 int gfx_v9_4_2_do_edc_gpr_workarounds(struct amdgpu_device *adev);
 
-extern const struct amdgpu_gfx_ras_funcs gfx_v9_4_2_ras_funcs;
+extern struct amdgpu_gfx_ras gfx_v9_4_2_ras;
 
 #endif /* __GFX_V9_4_2_H__ */
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_2_cleaner_shader.asm b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_2_cleaner_shader.asm
new file mode 100644
index 000000000000..35b8cf9070bd
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_2_cleaner_shader.asm
@@ -0,0 +1,153 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2024 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+// This shader is to clean LDS, SGPRs and VGPRs. It is  first 64 Dwords or 256 bytes of 192 Dwords cleaner shader.
+//To turn this shader program on for complitaion change this to main and lower shader main to main_1
+
+// MI200 : Clear SGPRs, VGPRs and LDS
+//   Uses two kernels launched separately:
+//   1. Clean VGPRs, LDS, and lower SGPRs
+//        Launches one workgroup per CU, each workgroup with 4x wave64 per SIMD in the CU
+//        Waves are "wave64" and have 128 VGPRs each, which uses all 512 VGPRs per SIMD
+//        Waves in the workgroup share the 64KB of LDS
+//        Each wave clears SGPRs 0 - 95. Because there are 4 waves/SIMD, this is physical SGPRs 0-383
+//        Each wave clears 128 VGPRs, so all 512 in the SIMD
+//        The first wave of the workgroup clears its 64KB of LDS
+//        The shader starts with "S_BARRIER" to ensure SPI has launched all waves of the workgroup
+//          before any wave in the workgroup could end.  Without this, it is possible not all SGPRs get cleared.
+//    2. Clean remaining SGPRs
+//        Launches a workgroup with 24 waves per workgroup, yielding 6 waves per SIMD in each CU
+//        Waves are allocating 96 SGPRs
+//          CP sets up SPI_RESOURCE_RESERVE_* registers to prevent these waves from allocating SGPRs 0-223.
+//          As such, these 6 waves per SIMD are allocated physical SGPRs 224-799
+//        Barriers do not work for >16 waves per workgroup, so we cannot start with S_BARRIER
+//          Instead, the shader starts with an S_SETHALT 1. Once all waves are launched CP will send unhalt command
+//        The shader then clears all SGPRs allocated to it, cleaning out physical SGPRs 224-799
+
+shader main
+  asic(MI200)
+  type(CS)
+  wave_size(64)
+// Note: original source code from SQ team
+
+//   (theorhetical fastest = ~512clks vgpr + 1536 lds + ~128 sgpr  = 2176 clks)
+
+  s_cmp_eq_u32 s0, 1                                // Bit0 is set, sgpr0 is set then clear VGPRS and LDS as FW set COMPUTE_USER_DATA_3
+  s_cbranch_scc0  label_0023                        // Clean VGPRs and LDS if sgpr0 of wave is set, scc = (s3 == 1)
+  S_BARRIER
+
+  s_movk_i32    m0, 0x0000
+  s_mov_b32     s2, 0x00000078  // Loop 128/8=16 times  (loop unrolled for performance)
+  //
+  // CLEAR VGPRs
+  //
+  s_set_gpr_idx_on  s2, 0x8    // enable Dest VGPR indexing
+label_0005:
+  v_mov_b32     v0, 0
+  v_mov_b32     v1, 0
+  v_mov_b32     v2, 0
+  v_mov_b32     v3, 0
+  v_mov_b32     v4, 0
+  v_mov_b32     v5, 0
+  v_mov_b32     v6, 0
+  v_mov_b32     v7, 0
+  s_sub_u32     s2, s2, 8
+  s_set_gpr_idx_idx  s2
+  s_cbranch_scc0  label_0005
+  s_set_gpr_idx_off
+
+  //
+  //
+
+  s_mov_b32     s2, 0x80000000                      // Bit31 is first_wave
+  s_and_b32     s2, s2, s1                          // sgpr0 has tg_size (first_wave) term as in ucode only COMPUTE_PGM_RSRC2.tg_size_en is set
+  s_cbranch_scc0  label_clean_sgpr_1                // Clean LDS if its first wave of ThreadGroup/WorkGroup
+  // CLEAR LDS
+  //
+  s_mov_b32 exec_lo, 0xffffffff
+  s_mov_b32 exec_hi, 0xffffffff
+  v_mbcnt_lo_u32_b32  v1, exec_hi, 0          // Set V1 to thread-ID (0..63)
+  v_mbcnt_hi_u32_b32  v1, exec_lo, v1         // Set V1 to thread-ID (0..63)
+  v_mul_u32_u24  v1, 0x00000008, v1           // * 8, so each thread is a double-dword address (8byte)
+  s_mov_b32     s2, 0x00000003f               // 64 loop iterations
+  s_mov_b32     m0, 0xffffffff
+  // Clear all of LDS space
+  // Each FirstWave of WorkGroup clears 64kbyte block
+
+label_001F:
+  ds_write2_b64  v1, v[2:3], v[2:3] offset1:32
+  ds_write2_b64  v1, v[4:5], v[4:5] offset0:64 offset1:96
+  v_add_co_u32     v1, vcc, 0x00000400, v1
+  s_sub_u32     s2, s2, 1
+  s_cbranch_scc0  label_001F
+  //
+  // CLEAR SGPRs
+  //
+label_clean_sgpr_1:
+  s_mov_b32     m0, 0x0000005c   // Loop 96/4=24 times  (loop unrolled for performance)
+  s_nop 0
+label_sgpr_loop:
+  s_movreld_b32     s0, 0
+  s_movreld_b32     s1, 0
+  s_movreld_b32     s2, 0
+  s_movreld_b32     s3, 0
+  s_sub_u32         m0, m0, 4
+  s_cbranch_scc0  label_sgpr_loop
+
+  //clear vcc, flat scratch
+  s_mov_b32 flat_scratch_lo, 0   //clear  flat scratch lo SGPR
+  s_mov_b32 flat_scratch_hi, 0   //clear  flat scratch hi SGPR
+  s_mov_b64 vcc, 0               //clear vcc
+  s_mov_b64 ttmp0, 0             //Clear ttmp0 and ttmp1
+  s_mov_b64 ttmp2, 0             //Clear ttmp2 and ttmp3
+  s_mov_b64 ttmp4, 0             //Clear ttmp4 and ttmp5
+  s_mov_b64 ttmp6, 0             //Clear ttmp6 and ttmp7
+  s_mov_b64 ttmp8, 0             //Clear ttmp8 and ttmp9
+  s_mov_b64 ttmp10, 0            //Clear ttmp10 and ttmp11
+  s_mov_b64 ttmp12, 0            //Clear ttmp12 and ttmp13
+  s_mov_b64 ttmp14, 0            //Clear ttmp14 and ttmp15
+s_endpgm
+
+label_0023:
+
+  s_sethalt 1
+
+  s_mov_b32     m0, 0x0000005c   // Loop 96/4=24 times  (loop unrolled for performance)
+  s_nop 0
+label_sgpr_loop1:
+
+  s_movreld_b32     s0, 0
+  s_movreld_b32     s1, 0
+  s_movreld_b32     s2, 0
+  s_movreld_b32     s3, 0
+  s_sub_u32         m0, m0, 4
+  s_cbranch_scc0  label_sgpr_loop1
+
+  //clear vcc, flat scratch
+  s_mov_b32 flat_scratch_lo, 0   //clear  flat scratch lo SGPR
+  s_mov_b32 flat_scratch_hi, 0   //clear  flat scratch hi SGPR
+  s_mov_b64 vcc, 0xee            //clear vcc
+
+s_endpgm
+end
+
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c
new file mode 100644
index 000000000000..cbb74ffc4792
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c
@@ -0,0 +1,5062 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include <linux/firmware.h>
+
+#include "amdgpu.h"
+#include "amdgpu_gfx.h"
+#include "soc15.h"
+#include "soc15d.h"
+#include "soc15_common.h"
+#include "vega10_enum.h"
+
+#include "v9_structs.h"
+
+#include "ivsrcid/gfx/irqsrcs_gfx_9_0.h"
+
+#include "gc/gc_9_4_3_offset.h"
+#include "gc/gc_9_4_3_sh_mask.h"
+
+#include "gfx_v9_4_3.h"
+#include "gfx_v9_4_3_cleaner_shader.h"
+#include "amdgpu_xcp.h"
+#include "amdgpu_aca.h"
+
+MODULE_FIRMWARE("amdgpu/gc_9_4_3_mec.bin");
+MODULE_FIRMWARE("amdgpu/gc_9_4_4_mec.bin");
+MODULE_FIRMWARE("amdgpu/gc_9_5_0_mec.bin");
+MODULE_FIRMWARE("amdgpu/gc_9_4_3_rlc.bin");
+MODULE_FIRMWARE("amdgpu/gc_9_4_4_rlc.bin");
+MODULE_FIRMWARE("amdgpu/gc_9_5_0_rlc.bin");
+MODULE_FIRMWARE("amdgpu/gc_9_4_3_sjt_mec.bin");
+MODULE_FIRMWARE("amdgpu/gc_9_4_4_sjt_mec.bin");
+
+#define GFX9_MEC_HPD_SIZE 4096
+#define RLCG_UCODE_LOADING_START_ADDRESS 0x00002000L
+
+#define GOLDEN_GB_ADDR_CONFIG 0x2a114042
+#define CP_HQD_PERSISTENT_STATE_DEFAULT 0xbe05301
+
+#define XCC_REG_RANGE_0_LOW  0x2000     /* XCC gfxdec0 lower Bound */
+#define XCC_REG_RANGE_0_HIGH 0x3400     /* XCC gfxdec0 upper Bound */
+#define XCC_REG_RANGE_1_LOW  0xA000     /* XCC gfxdec1 lower Bound */
+#define XCC_REG_RANGE_1_HIGH 0x10000    /* XCC gfxdec1 upper Bound */
+
+#define NORMALIZE_XCC_REG_OFFSET(offset) \
+	(offset & 0xFFFF)
+
+static const struct amdgpu_hwip_reg_entry gc_reg_list_9_4_3[] = {
+	SOC15_REG_ENTRY_STR(GC, 0, regGRBM_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regGRBM_STATUS2),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_STALLED_STAT1),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_STALLED_STAT2),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_CPC_STALLED_STAT1),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_CPF_STALLED_STAT1),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_BUSY_STAT),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_CPC_BUSY_STAT),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_CPF_BUSY_STAT),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_CPF_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_ERROR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCPF_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regCPC_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regCPG_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regGDS_PROTECTION_FAULT),
+	SOC15_REG_ENTRY_STR(GC, 0, regGDS_VM_PROTECTION_FAULT),
+	SOC15_REG_ENTRY_STR(GC, 0, regRLC_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regRMI_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regSQC_DCACHE_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regSQC_ICACHE_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regSQ_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regTCP_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regWD_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regVM_L2_PROTECTION_FAULT_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, regVM_L2_PROTECTION_FAULT_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_DEBUG),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC1_INSTR_PNTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC2_INSTR_PNTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_CPC_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regRLC_STAT),
+	SOC15_REG_ENTRY_STR(GC, 0, regRLC_SMU_COMMAND),
+	SOC15_REG_ENTRY_STR(GC, 0, regRLC_SMU_MESSAGE),
+	SOC15_REG_ENTRY_STR(GC, 0, regRLC_SMU_ARGUMENT_1),
+	SOC15_REG_ENTRY_STR(GC, 0, regRLC_SMU_ARGUMENT_2),
+	SOC15_REG_ENTRY_STR(GC, 0, regSMU_RLC_RESPONSE),
+	SOC15_REG_ENTRY_STR(GC, 0, regRLC_SAFE_MODE),
+	SOC15_REG_ENTRY_STR(GC, 0, regRLC_SMU_SAFE_MODE),
+	SOC15_REG_ENTRY_STR(GC, 0, regRLC_INT_STAT),
+	SOC15_REG_ENTRY_STR(GC, 0, regRLC_GPM_GENERAL_6),
+	/* SE status registers */
+	SOC15_REG_ENTRY_STR(GC, 0, regGRBM_STATUS_SE0),
+	SOC15_REG_ENTRY_STR(GC, 0, regGRBM_STATUS_SE1),
+	SOC15_REG_ENTRY_STR(GC, 0, regGRBM_STATUS_SE2),
+	SOC15_REG_ENTRY_STR(GC, 0, regGRBM_STATUS_SE3)
+};
+
+static const struct amdgpu_hwip_reg_entry gc_cp_reg_list_9_4_3[] = {
+	/* compute queue registers */
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_VMID),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_ACTIVE),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PERSISTENT_STATE),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PIPE_PRIORITY),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_QUEUE_PRIORITY),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_QUANTUM),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PQ_BASE),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PQ_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PQ_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PQ_WPTR_POLL_ADDR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PQ_WPTR_POLL_ADDR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PQ_DOORBELL_CONTROL),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PQ_CONTROL),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_IB_BASE_ADDR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_IB_BASE_ADDR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_IB_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_IB_CONTROL),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_DEQUEUE_REQUEST),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_EOP_BASE_ADDR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_EOP_BASE_ADDR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_EOP_CONTROL),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_EOP_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_EOP_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_EOP_EVENTS),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_CTX_SAVE_BASE_ADDR_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_CTX_SAVE_BASE_ADDR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_CTX_SAVE_CONTROL),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_CNTL_STACK_OFFSET),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_CNTL_STACK_SIZE),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_WG_STATE_OFFSET),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_CTX_SAVE_SIZE),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_GDS_RESOURCE_STATE),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_ERROR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_EOP_WPTR_MEM),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PQ_WPTR_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PQ_WPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_GFX_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_ME1_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_ME1_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_ME1_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_ME1_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_ME1_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_ME1_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_ME1_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_ME1_HEADER_DUMP),
+};
+
+struct amdgpu_gfx_ras gfx_v9_4_3_ras;
+
+static void gfx_v9_4_3_set_ring_funcs(struct amdgpu_device *adev);
+static void gfx_v9_4_3_set_irq_funcs(struct amdgpu_device *adev);
+static void gfx_v9_4_3_set_gds_init(struct amdgpu_device *adev);
+static void gfx_v9_4_3_set_rlc_funcs(struct amdgpu_device *adev);
+static int gfx_v9_4_3_get_cu_info(struct amdgpu_device *adev,
+				struct amdgpu_cu_info *cu_info);
+static void gfx_v9_4_3_xcc_set_safe_mode(struct amdgpu_device *adev, int xcc_id);
+static void gfx_v9_4_3_xcc_unset_safe_mode(struct amdgpu_device *adev, int xcc_id);
+
+static void gfx_v9_4_3_kiq_set_resources(struct amdgpu_ring *kiq_ring,
+				uint64_t queue_mask)
+{
+	struct amdgpu_device *adev = kiq_ring->adev;
+	u64 shader_mc_addr;
+
+	/* Cleaner shader MC address */
+	shader_mc_addr = adev->gfx.cleaner_shader_gpu_addr >> 8;
+
+	amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_SET_RESOURCES, 6));
+	amdgpu_ring_write(kiq_ring,
+		PACKET3_SET_RESOURCES_VMID_MASK(0) |
+		/* vmid_mask:0* queue_type:0 (KIQ) */
+		PACKET3_SET_RESOURCES_QUEUE_TYPE(0));
+	amdgpu_ring_write(kiq_ring,
+			lower_32_bits(queue_mask));	/* queue mask lo */
+	amdgpu_ring_write(kiq_ring,
+			upper_32_bits(queue_mask));	/* queue mask hi */
+	amdgpu_ring_write(kiq_ring, lower_32_bits(shader_mc_addr)); /* cleaner shader addr lo */
+	amdgpu_ring_write(kiq_ring, upper_32_bits(shader_mc_addr)); /* cleaner shader addr hi */
+	amdgpu_ring_write(kiq_ring, 0);	/* oac mask */
+	amdgpu_ring_write(kiq_ring, 0);	/* gds heap base:0, gds heap size:0 */
+}
+
+static void gfx_v9_4_3_kiq_map_queues(struct amdgpu_ring *kiq_ring,
+				 struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = kiq_ring->adev;
+	uint64_t mqd_addr = amdgpu_bo_gpu_offset(ring->mqd_obj);
+	uint64_t wptr_addr = adev->wb.gpu_addr + (ring->wptr_offs * 4);
+	uint32_t eng_sel = ring->funcs->type == AMDGPU_RING_TYPE_GFX ? 4 : 0;
+
+	amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_MAP_QUEUES, 5));
+	/* Q_sel:0, vmid:0, vidmem: 1, engine:0, num_Q:1*/
+	amdgpu_ring_write(kiq_ring, /* Q_sel: 0, vmid: 0, engine: 0, num_Q: 1 */
+			 PACKET3_MAP_QUEUES_QUEUE_SEL(0) | /* Queue_Sel */
+			 PACKET3_MAP_QUEUES_VMID(0) | /* VMID */
+			 PACKET3_MAP_QUEUES_QUEUE(ring->queue) |
+			 PACKET3_MAP_QUEUES_PIPE(ring->pipe) |
+			 PACKET3_MAP_QUEUES_ME((ring->me == 1 ? 0 : 1)) |
+			 /*queue_type: normal compute queue */
+			 PACKET3_MAP_QUEUES_QUEUE_TYPE(0) |
+			 /* alloc format: all_on_one_pipe */
+			 PACKET3_MAP_QUEUES_ALLOC_FORMAT(0) |
+			 PACKET3_MAP_QUEUES_ENGINE_SEL(eng_sel) |
+			 /* num_queues: must be 1 */
+			 PACKET3_MAP_QUEUES_NUM_QUEUES(1));
+	amdgpu_ring_write(kiq_ring,
+			PACKET3_MAP_QUEUES_DOORBELL_OFFSET(ring->doorbell_index));
+	amdgpu_ring_write(kiq_ring, lower_32_bits(mqd_addr));
+	amdgpu_ring_write(kiq_ring, upper_32_bits(mqd_addr));
+	amdgpu_ring_write(kiq_ring, lower_32_bits(wptr_addr));
+	amdgpu_ring_write(kiq_ring, upper_32_bits(wptr_addr));
+}
+
+static void gfx_v9_4_3_kiq_unmap_queues(struct amdgpu_ring *kiq_ring,
+				   struct amdgpu_ring *ring,
+				   enum amdgpu_unmap_queues_action action,
+				   u64 gpu_addr, u64 seq)
+{
+	uint32_t eng_sel = ring->funcs->type == AMDGPU_RING_TYPE_GFX ? 4 : 0;
+
+	amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_UNMAP_QUEUES, 4));
+	amdgpu_ring_write(kiq_ring, /* Q_sel: 0, vmid: 0, engine: 0, num_Q: 1 */
+			  PACKET3_UNMAP_QUEUES_ACTION(action) |
+			  PACKET3_UNMAP_QUEUES_QUEUE_SEL(0) |
+			  PACKET3_UNMAP_QUEUES_ENGINE_SEL(eng_sel) |
+			  PACKET3_UNMAP_QUEUES_NUM_QUEUES(1));
+	amdgpu_ring_write(kiq_ring,
+			PACKET3_UNMAP_QUEUES_DOORBELL_OFFSET0(ring->doorbell_index));
+
+	if (action == PREEMPT_QUEUES_NO_UNMAP) {
+		amdgpu_ring_write(kiq_ring, lower_32_bits(gpu_addr));
+		amdgpu_ring_write(kiq_ring, upper_32_bits(gpu_addr));
+		amdgpu_ring_write(kiq_ring, seq);
+	} else {
+		amdgpu_ring_write(kiq_ring, 0);
+		amdgpu_ring_write(kiq_ring, 0);
+		amdgpu_ring_write(kiq_ring, 0);
+	}
+}
+
+static void gfx_v9_4_3_kiq_query_status(struct amdgpu_ring *kiq_ring,
+				   struct amdgpu_ring *ring,
+				   u64 addr,
+				   u64 seq)
+{
+	uint32_t eng_sel = ring->funcs->type == AMDGPU_RING_TYPE_GFX ? 4 : 0;
+
+	amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_QUERY_STATUS, 5));
+	amdgpu_ring_write(kiq_ring,
+			  PACKET3_QUERY_STATUS_CONTEXT_ID(0) |
+			  PACKET3_QUERY_STATUS_INTERRUPT_SEL(0) |
+			  PACKET3_QUERY_STATUS_COMMAND(2));
+	/* Q_sel: 0, vmid: 0, engine: 0, num_Q: 1 */
+	amdgpu_ring_write(kiq_ring,
+			PACKET3_QUERY_STATUS_DOORBELL_OFFSET(ring->doorbell_index) |
+			PACKET3_QUERY_STATUS_ENG_SEL(eng_sel));
+	amdgpu_ring_write(kiq_ring, lower_32_bits(addr));
+	amdgpu_ring_write(kiq_ring, upper_32_bits(addr));
+	amdgpu_ring_write(kiq_ring, lower_32_bits(seq));
+	amdgpu_ring_write(kiq_ring, upper_32_bits(seq));
+}
+
+static void gfx_v9_4_3_kiq_invalidate_tlbs(struct amdgpu_ring *kiq_ring,
+				uint16_t pasid, uint32_t flush_type,
+				bool all_hub)
+{
+	amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_INVALIDATE_TLBS, 0));
+	amdgpu_ring_write(kiq_ring,
+			PACKET3_INVALIDATE_TLBS_DST_SEL(1) |
+			PACKET3_INVALIDATE_TLBS_ALL_HUB(all_hub) |
+			PACKET3_INVALIDATE_TLBS_PASID(pasid) |
+			PACKET3_INVALIDATE_TLBS_FLUSH_TYPE(flush_type));
+}
+
+static void gfx_v9_4_3_kiq_reset_hw_queue(struct amdgpu_ring *kiq_ring, uint32_t queue_type,
+					  uint32_t me_id, uint32_t pipe_id, uint32_t queue_id,
+					  uint32_t xcc_id, uint32_t vmid)
+{
+	struct amdgpu_device *adev = kiq_ring->adev;
+	unsigned i;
+
+	/* enter save mode */
+	amdgpu_gfx_rlc_enter_safe_mode(adev, xcc_id);
+	mutex_lock(&adev->srbm_mutex);
+	soc15_grbm_select(adev, me_id, pipe_id, queue_id, 0, xcc_id);
+
+	if (queue_type == AMDGPU_RING_TYPE_COMPUTE) {
+		WREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_HQD_DEQUEUE_REQUEST, 0x2);
+		WREG32_SOC15(GC, GET_INST(GC, xcc_id), regSPI_COMPUTE_QUEUE_RESET, 0x1);
+		/* wait till dequeue take effects */
+		for (i = 0; i < adev->usec_timeout; i++) {
+			if (!(RREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_HQD_ACTIVE) & 1))
+				break;
+			udelay(1);
+		}
+		if (i >= adev->usec_timeout)
+			dev_err(adev->dev, "fail to wait on hqd deactive\n");
+	} else {
+		dev_err(adev->dev, "reset queue_type(%d) not supported\n\n", queue_type);
+	}
+
+	soc15_grbm_select(adev, 0, 0, 0, 0, 0);
+	mutex_unlock(&adev->srbm_mutex);
+	/* exit safe mode */
+	amdgpu_gfx_rlc_exit_safe_mode(adev, xcc_id);
+}
+
+static const struct kiq_pm4_funcs gfx_v9_4_3_kiq_pm4_funcs = {
+	.kiq_set_resources = gfx_v9_4_3_kiq_set_resources,
+	.kiq_map_queues = gfx_v9_4_3_kiq_map_queues,
+	.kiq_unmap_queues = gfx_v9_4_3_kiq_unmap_queues,
+	.kiq_query_status = gfx_v9_4_3_kiq_query_status,
+	.kiq_invalidate_tlbs = gfx_v9_4_3_kiq_invalidate_tlbs,
+	.kiq_reset_hw_queue = gfx_v9_4_3_kiq_reset_hw_queue,
+	.set_resources_size = 8,
+	.map_queues_size = 7,
+	.unmap_queues_size = 6,
+	.query_status_size = 7,
+	.invalidate_tlbs_size = 2,
+};
+
+static void gfx_v9_4_3_set_kiq_pm4_funcs(struct amdgpu_device *adev)
+{
+	int i, num_xcc;
+
+	num_xcc = NUM_XCC(adev->gfx.xcc_mask);
+	for (i = 0; i < num_xcc; i++)
+		adev->gfx.kiq[i].pmf = &gfx_v9_4_3_kiq_pm4_funcs;
+}
+
+static void gfx_v9_4_3_init_golden_registers(struct amdgpu_device *adev)
+{
+	int i, num_xcc, dev_inst;
+
+	num_xcc = NUM_XCC(adev->gfx.xcc_mask);
+	for (i = 0; i < num_xcc; i++) {
+		dev_inst = GET_INST(GC, i);
+
+		WREG32_SOC15(GC, dev_inst, regGB_ADDR_CONFIG,
+			     GOLDEN_GB_ADDR_CONFIG);
+		WREG32_FIELD15_PREREG(GC, dev_inst, TCP_UTCL1_CNTL2, SPARE, 0x1);
+	}
+}
+
+static uint32_t gfx_v9_4_3_normalize_xcc_reg_offset(uint32_t reg)
+{
+	uint32_t normalized_reg = NORMALIZE_XCC_REG_OFFSET(reg);
+
+	/* If it is an XCC reg, normalize the reg to keep
+	   lower 16 bits in local xcc */
+
+	if (((normalized_reg >= XCC_REG_RANGE_0_LOW) && (normalized_reg < XCC_REG_RANGE_0_HIGH)) ||
+		((normalized_reg >= XCC_REG_RANGE_1_LOW) && (normalized_reg < XCC_REG_RANGE_1_HIGH)))
+		return normalized_reg;
+	else
+		return reg;
+}
+
+static void gfx_v9_4_3_write_data_to_reg(struct amdgpu_ring *ring, int eng_sel,
+				       bool wc, uint32_t reg, uint32_t val)
+{
+	reg = gfx_v9_4_3_normalize_xcc_reg_offset(reg);
+	amdgpu_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
+	amdgpu_ring_write(ring, WRITE_DATA_ENGINE_SEL(eng_sel) |
+				WRITE_DATA_DST_SEL(0) |
+				(wc ? WR_CONFIRM : 0));
+	amdgpu_ring_write(ring, reg);
+	amdgpu_ring_write(ring, 0);
+	amdgpu_ring_write(ring, val);
+}
+
+static void gfx_v9_4_3_wait_reg_mem(struct amdgpu_ring *ring, int eng_sel,
+				  int mem_space, int opt, uint32_t addr0,
+				  uint32_t addr1, uint32_t ref, uint32_t mask,
+				  uint32_t inv)
+{
+	/* Only do the normalization on regspace */
+	if (mem_space == 0) {
+		addr0 = gfx_v9_4_3_normalize_xcc_reg_offset(addr0);
+		addr1 = gfx_v9_4_3_normalize_xcc_reg_offset(addr1);
+	}
+
+	amdgpu_ring_write(ring, PACKET3(PACKET3_WAIT_REG_MEM, 5));
+	amdgpu_ring_write(ring,
+				 /* memory (1) or register (0) */
+				 (WAIT_REG_MEM_MEM_SPACE(mem_space) |
+				 WAIT_REG_MEM_OPERATION(opt) | /* wait */
+				 WAIT_REG_MEM_FUNCTION(3) |  /* equal */
+				 WAIT_REG_MEM_ENGINE(eng_sel)));
+
+	if (mem_space)
+		BUG_ON(addr0 & 0x3); /* Dword align */
+	amdgpu_ring_write(ring, addr0);
+	amdgpu_ring_write(ring, addr1);
+	amdgpu_ring_write(ring, ref);
+	amdgpu_ring_write(ring, mask);
+	amdgpu_ring_write(ring, inv); /* poll interval */
+}
+
+static int gfx_v9_4_3_ring_test_ring(struct amdgpu_ring *ring)
+{
+	uint32_t scratch_reg0_offset, xcc_offset;
+	struct amdgpu_device *adev = ring->adev;
+	uint32_t tmp = 0;
+	unsigned i;
+	int r;
+
+	/* Use register offset which is local to XCC in the packet */
+	xcc_offset = SOC15_REG_OFFSET(GC, 0, regSCRATCH_REG0);
+	scratch_reg0_offset = SOC15_REG_OFFSET(GC, GET_INST(GC, ring->xcc_id), regSCRATCH_REG0);
+	WREG32(scratch_reg0_offset, 0xCAFEDEAD);
+	tmp = RREG32(scratch_reg0_offset);
+
+	r = amdgpu_ring_alloc(ring, 3);
+	if (r)
+		return r;
+
+	amdgpu_ring_write(ring, PACKET3(PACKET3_SET_UCONFIG_REG, 1));
+	amdgpu_ring_write(ring, xcc_offset - PACKET3_SET_UCONFIG_REG_START);
+	amdgpu_ring_write(ring, 0xDEADBEEF);
+	amdgpu_ring_commit(ring);
+
+	for (i = 0; i < adev->usec_timeout; i++) {
+		tmp = RREG32(scratch_reg0_offset);
+		if (tmp == 0xDEADBEEF)
+			break;
+		udelay(1);
+	}
+
+	if (i >= adev->usec_timeout)
+		r = -ETIMEDOUT;
+	return r;
+}
+
+static int gfx_v9_4_3_ring_test_ib(struct amdgpu_ring *ring, long timeout)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct amdgpu_ib ib;
+	struct dma_fence *f = NULL;
+
+	unsigned index;
+	uint64_t gpu_addr;
+	uint32_t tmp;
+	long r;
+
+	r = amdgpu_device_wb_get(adev, &index);
+	if (r)
+		return r;
+
+	gpu_addr = adev->wb.gpu_addr + (index * 4);
+	adev->wb.wb[index] = cpu_to_le32(0xCAFEDEAD);
+	memset(&ib, 0, sizeof(ib));
+
+	r = amdgpu_ib_get(adev, NULL, 20, AMDGPU_IB_POOL_DIRECT, &ib);
+	if (r)
+		goto err1;
+
+	ib.ptr[0] = PACKET3(PACKET3_WRITE_DATA, 3);
+	ib.ptr[1] = WRITE_DATA_DST_SEL(5) | WR_CONFIRM;
+	ib.ptr[2] = lower_32_bits(gpu_addr);
+	ib.ptr[3] = upper_32_bits(gpu_addr);
+	ib.ptr[4] = 0xDEADBEEF;
+	ib.length_dw = 5;
+
+	r = amdgpu_ib_schedule(ring, 1, &ib, NULL, &f);
+	if (r)
+		goto err2;
+
+	r = dma_fence_wait_timeout(f, false, timeout);
+	if (r == 0) {
+		r = -ETIMEDOUT;
+		goto err2;
+	} else if (r < 0) {
+		goto err2;
+	}
+
+	tmp = adev->wb.wb[index];
+	if (tmp == 0xDEADBEEF)
+		r = 0;
+	else
+		r = -EINVAL;
+
+err2:
+	amdgpu_ib_free(&ib, NULL);
+	dma_fence_put(f);
+err1:
+	amdgpu_device_wb_free(adev, index);
+	return r;
+}
+
+
+/* This value might differs per partition */
+static uint64_t gfx_v9_4_3_get_gpu_clock_counter(struct amdgpu_device *adev)
+{
+	uint64_t clock;
+
+	mutex_lock(&adev->gfx.gpu_clock_mutex);
+	WREG32_SOC15(GC, GET_INST(GC, 0), regRLC_CAPTURE_GPU_CLOCK_COUNT, 1);
+	clock = (uint64_t)RREG32_SOC15(GC, GET_INST(GC, 0), regRLC_GPU_CLOCK_COUNT_LSB) |
+		((uint64_t)RREG32_SOC15(GC, GET_INST(GC, 0), regRLC_GPU_CLOCK_COUNT_MSB) << 32ULL);
+	mutex_unlock(&adev->gfx.gpu_clock_mutex);
+
+	return clock;
+}
+
+static void gfx_v9_4_3_free_microcode(struct amdgpu_device *adev)
+{
+	amdgpu_ucode_release(&adev->gfx.pfp_fw);
+	amdgpu_ucode_release(&adev->gfx.me_fw);
+	amdgpu_ucode_release(&adev->gfx.ce_fw);
+	amdgpu_ucode_release(&adev->gfx.rlc_fw);
+	amdgpu_ucode_release(&adev->gfx.mec_fw);
+	amdgpu_ucode_release(&adev->gfx.mec2_fw);
+
+	kfree(adev->gfx.rlc.register_list_format);
+}
+
+static int gfx_v9_4_3_init_rlc_microcode(struct amdgpu_device *adev,
+					  const char *chip_name)
+{
+	int err;
+	const struct rlc_firmware_header_v2_0 *rlc_hdr;
+	uint16_t version_major;
+	uint16_t version_minor;
+
+
+	err = amdgpu_ucode_request(adev, &adev->gfx.rlc_fw,
+				   AMDGPU_UCODE_REQUIRED,
+				   "amdgpu/%s_rlc.bin", chip_name);
+	if (err)
+		goto out;
+	rlc_hdr = (const struct rlc_firmware_header_v2_0 *)adev->gfx.rlc_fw->data;
+
+	version_major = le16_to_cpu(rlc_hdr->header.header_version_major);
+	version_minor = le16_to_cpu(rlc_hdr->header.header_version_minor);
+	err = amdgpu_gfx_rlc_init_microcode(adev, version_major, version_minor);
+out:
+	if (err)
+		amdgpu_ucode_release(&adev->gfx.rlc_fw);
+
+	return err;
+}
+
+static int gfx_v9_4_3_init_cp_compute_microcode(struct amdgpu_device *adev,
+					  const char *chip_name)
+{
+	int err;
+
+	if (amdgpu_sriov_vf(adev)) {
+		err = amdgpu_ucode_request(adev, &adev->gfx.mec_fw,
+					   AMDGPU_UCODE_REQUIRED,
+					   "amdgpu/%s_sjt_mec.bin", chip_name);
+
+		if (err)
+			err = amdgpu_ucode_request(adev, &adev->gfx.mec_fw,
+							AMDGPU_UCODE_REQUIRED,
+							"amdgpu/%s_mec.bin", chip_name);
+	} else
+		err = amdgpu_ucode_request(adev, &adev->gfx.mec_fw,
+					   AMDGPU_UCODE_REQUIRED,
+					   "amdgpu/%s_mec.bin", chip_name);
+	if (err)
+		goto out;
+	amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_MEC1);
+	amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_MEC1_JT);
+
+	adev->gfx.mec2_fw_version = adev->gfx.mec_fw_version;
+	adev->gfx.mec2_feature_version = adev->gfx.mec_feature_version;
+
+out:
+	if (err)
+		amdgpu_ucode_release(&adev->gfx.mec_fw);
+	return err;
+}
+
+static int gfx_v9_4_3_init_microcode(struct amdgpu_device *adev)
+{
+	char ucode_prefix[15];
+	int r;
+
+	amdgpu_ucode_ip_version_decode(adev, GC_HWIP, ucode_prefix, sizeof(ucode_prefix));
+
+	r = gfx_v9_4_3_init_rlc_microcode(adev, ucode_prefix);
+	if (r)
+		return r;
+
+	r = gfx_v9_4_3_init_cp_compute_microcode(adev, ucode_prefix);
+	if (r)
+		return r;
+
+	return r;
+}
+
+static void gfx_v9_4_3_mec_fini(struct amdgpu_device *adev)
+{
+	amdgpu_bo_free_kernel(&adev->gfx.mec.hpd_eop_obj, NULL, NULL);
+	amdgpu_bo_free_kernel(&adev->gfx.mec.mec_fw_obj, NULL, NULL);
+}
+
+static int gfx_v9_4_3_mec_init(struct amdgpu_device *adev)
+{
+	int r, i, num_xcc;
+	u32 *hpd;
+	const __le32 *fw_data;
+	unsigned fw_size;
+	u32 *fw;
+	size_t mec_hpd_size;
+
+	const struct gfx_firmware_header_v1_0 *mec_hdr;
+
+	num_xcc = NUM_XCC(adev->gfx.xcc_mask);
+	for (i = 0; i < num_xcc; i++)
+		bitmap_zero(adev->gfx.mec_bitmap[i].queue_bitmap,
+			AMDGPU_MAX_COMPUTE_QUEUES);
+
+	/* take ownership of the relevant compute queues */
+	amdgpu_gfx_compute_queue_acquire(adev);
+	mec_hpd_size =
+		adev->gfx.num_compute_rings * num_xcc * GFX9_MEC_HPD_SIZE;
+	if (mec_hpd_size) {
+		r = amdgpu_bo_create_reserved(adev, mec_hpd_size, PAGE_SIZE,
+					      AMDGPU_GEM_DOMAIN_VRAM |
+					      AMDGPU_GEM_DOMAIN_GTT,
+					      &adev->gfx.mec.hpd_eop_obj,
+					      &adev->gfx.mec.hpd_eop_gpu_addr,
+					      (void **)&hpd);
+		if (r) {
+			dev_warn(adev->dev, "(%d) create HDP EOP bo failed\n", r);
+			gfx_v9_4_3_mec_fini(adev);
+			return r;
+		}
+
+		if (amdgpu_emu_mode == 1) {
+			for (i = 0; i < mec_hpd_size / 4; i++) {
+				memset((void *)(hpd + i), 0, 4);
+				if (i % 50 == 0)
+					msleep(1);
+			}
+		} else {
+			memset(hpd, 0, mec_hpd_size);
+		}
+
+		amdgpu_bo_kunmap(adev->gfx.mec.hpd_eop_obj);
+		amdgpu_bo_unreserve(adev->gfx.mec.hpd_eop_obj);
+	}
+
+	mec_hdr = (const struct gfx_firmware_header_v1_0 *)adev->gfx.mec_fw->data;
+
+	fw_data = (const __le32 *)
+		(adev->gfx.mec_fw->data +
+		 le32_to_cpu(mec_hdr->header.ucode_array_offset_bytes));
+	fw_size = le32_to_cpu(mec_hdr->header.ucode_size_bytes);
+
+	r = amdgpu_bo_create_reserved(adev, mec_hdr->header.ucode_size_bytes,
+				      PAGE_SIZE, AMDGPU_GEM_DOMAIN_GTT,
+				      &adev->gfx.mec.mec_fw_obj,
+				      &adev->gfx.mec.mec_fw_gpu_addr,
+				      (void **)&fw);
+	if (r) {
+		dev_warn(adev->dev, "(%d) create mec firmware bo failed\n", r);
+		gfx_v9_4_3_mec_fini(adev);
+		return r;
+	}
+
+	memcpy(fw, fw_data, fw_size);
+
+	amdgpu_bo_kunmap(adev->gfx.mec.mec_fw_obj);
+	amdgpu_bo_unreserve(adev->gfx.mec.mec_fw_obj);
+
+	return 0;
+}
+
+static void gfx_v9_4_3_xcc_select_se_sh(struct amdgpu_device *adev, u32 se_num,
+					u32 sh_num, u32 instance, int xcc_id)
+{
+	u32 data;
+
+	if (instance == 0xffffffff)
+		data = REG_SET_FIELD(0, GRBM_GFX_INDEX,
+				     INSTANCE_BROADCAST_WRITES, 1);
+	else
+		data = REG_SET_FIELD(0, GRBM_GFX_INDEX,
+				     INSTANCE_INDEX, instance);
+
+	if (se_num == 0xffffffff)
+		data = REG_SET_FIELD(data, GRBM_GFX_INDEX,
+				     SE_BROADCAST_WRITES, 1);
+	else
+		data = REG_SET_FIELD(data, GRBM_GFX_INDEX, SE_INDEX, se_num);
+
+	if (sh_num == 0xffffffff)
+		data = REG_SET_FIELD(data, GRBM_GFX_INDEX,
+				     SH_BROADCAST_WRITES, 1);
+	else
+		data = REG_SET_FIELD(data, GRBM_GFX_INDEX, SH_INDEX, sh_num);
+
+	WREG32_SOC15_RLC_SHADOW_EX(reg, GC, GET_INST(GC, xcc_id), regGRBM_GFX_INDEX, data);
+}
+
+static uint32_t wave_read_ind(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd, uint32_t wave, uint32_t address)
+{
+	WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regSQ_IND_INDEX,
+		(wave << SQ_IND_INDEX__WAVE_ID__SHIFT) |
+		(simd << SQ_IND_INDEX__SIMD_ID__SHIFT) |
+		(address << SQ_IND_INDEX__INDEX__SHIFT) |
+		(SQ_IND_INDEX__FORCE_READ_MASK));
+	return RREG32_SOC15(GC, GET_INST(GC, xcc_id), regSQ_IND_DATA);
+}
+
+static void wave_read_regs(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd,
+			   uint32_t wave, uint32_t thread,
+			   uint32_t regno, uint32_t num, uint32_t *out)
+{
+	WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regSQ_IND_INDEX,
+		(wave << SQ_IND_INDEX__WAVE_ID__SHIFT) |
+		(simd << SQ_IND_INDEX__SIMD_ID__SHIFT) |
+		(regno << SQ_IND_INDEX__INDEX__SHIFT) |
+		(thread << SQ_IND_INDEX__THREAD_ID__SHIFT) |
+		(SQ_IND_INDEX__FORCE_READ_MASK) |
+		(SQ_IND_INDEX__AUTO_INCR_MASK));
+	while (num--)
+		*(out++) = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regSQ_IND_DATA);
+}
+
+static void gfx_v9_4_3_read_wave_data(struct amdgpu_device *adev,
+				      uint32_t xcc_id, uint32_t simd, uint32_t wave,
+				      uint32_t *dst, int *no_fields)
+{
+	/* type 1 wave data */
+	dst[(*no_fields)++] = 1;
+	dst[(*no_fields)++] = wave_read_ind(adev, xcc_id, simd, wave, ixSQ_WAVE_STATUS);
+	dst[(*no_fields)++] = wave_read_ind(adev, xcc_id, simd, wave, ixSQ_WAVE_PC_LO);
+	dst[(*no_fields)++] = wave_read_ind(adev, xcc_id, simd, wave, ixSQ_WAVE_PC_HI);
+	dst[(*no_fields)++] = wave_read_ind(adev, xcc_id, simd, wave, ixSQ_WAVE_EXEC_LO);
+	dst[(*no_fields)++] = wave_read_ind(adev, xcc_id, simd, wave, ixSQ_WAVE_EXEC_HI);
+	dst[(*no_fields)++] = wave_read_ind(adev, xcc_id, simd, wave, ixSQ_WAVE_HW_ID);
+	dst[(*no_fields)++] = wave_read_ind(adev, xcc_id, simd, wave, ixSQ_WAVE_INST_DW0);
+	dst[(*no_fields)++] = wave_read_ind(adev, xcc_id, simd, wave, ixSQ_WAVE_INST_DW1);
+	dst[(*no_fields)++] = wave_read_ind(adev, xcc_id, simd, wave, ixSQ_WAVE_GPR_ALLOC);
+	dst[(*no_fields)++] = wave_read_ind(adev, xcc_id, simd, wave, ixSQ_WAVE_LDS_ALLOC);
+	dst[(*no_fields)++] = wave_read_ind(adev, xcc_id, simd, wave, ixSQ_WAVE_TRAPSTS);
+	dst[(*no_fields)++] = wave_read_ind(adev, xcc_id, simd, wave, ixSQ_WAVE_IB_STS);
+	dst[(*no_fields)++] = wave_read_ind(adev, xcc_id, simd, wave, ixSQ_WAVE_IB_DBG0);
+	dst[(*no_fields)++] = wave_read_ind(adev, xcc_id, simd, wave, ixSQ_WAVE_M0);
+	dst[(*no_fields)++] = wave_read_ind(adev, xcc_id, simd, wave, ixSQ_WAVE_MODE);
+}
+
+static void gfx_v9_4_3_read_wave_sgprs(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd,
+				       uint32_t wave, uint32_t start,
+				       uint32_t size, uint32_t *dst)
+{
+	wave_read_regs(adev, xcc_id, simd, wave, 0,
+		       start + SQIND_WAVE_SGPRS_OFFSET, size, dst);
+}
+
+static void gfx_v9_4_3_read_wave_vgprs(struct amdgpu_device *adev, uint32_t xcc_id, uint32_t simd,
+				       uint32_t wave, uint32_t thread,
+				       uint32_t start, uint32_t size,
+				       uint32_t *dst)
+{
+	wave_read_regs(adev, xcc_id, simd, wave, thread,
+		       start + SQIND_WAVE_VGPRS_OFFSET, size, dst);
+}
+
+static void gfx_v9_4_3_select_me_pipe_q(struct amdgpu_device *adev,
+					u32 me, u32 pipe, u32 q, u32 vm, u32 xcc_id)
+{
+	soc15_grbm_select(adev, me, pipe, q, vm, GET_INST(GC, xcc_id));
+}
+
+static int gfx_v9_4_3_get_xccs_per_xcp(struct amdgpu_device *adev)
+{
+	u32 xcp_ctl;
+
+	/* Value is expected to be the same on all, fetch from first instance */
+	xcp_ctl = RREG32_SOC15(GC, GET_INST(GC, 0), regCP_HYP_XCP_CTL);
+
+	return REG_GET_FIELD(xcp_ctl, CP_HYP_XCP_CTL, NUM_XCC_IN_XCP);
+}
+
+static int gfx_v9_4_3_switch_compute_partition(struct amdgpu_device *adev,
+						int num_xccs_per_xcp)
+{
+	int ret, i, num_xcc;
+	u32 tmp = 0;
+
+	if (adev->psp.funcs) {
+		ret = psp_spatial_partition(&adev->psp,
+					    NUM_XCC(adev->gfx.xcc_mask) /
+						    num_xccs_per_xcp);
+		if (ret)
+			return ret;
+	} else {
+		num_xcc = NUM_XCC(adev->gfx.xcc_mask);
+
+		for (i = 0; i < num_xcc; i++) {
+			tmp = REG_SET_FIELD(tmp, CP_HYP_XCP_CTL, NUM_XCC_IN_XCP,
+					    num_xccs_per_xcp);
+			tmp = REG_SET_FIELD(tmp, CP_HYP_XCP_CTL, VIRTUAL_XCC_ID,
+					    i % num_xccs_per_xcp);
+			WREG32_SOC15(GC, GET_INST(GC, i), regCP_HYP_XCP_CTL,
+				     tmp);
+		}
+		ret = 0;
+	}
+
+	adev->gfx.num_xcc_per_xcp = num_xccs_per_xcp;
+
+	return ret;
+}
+
+static int gfx_v9_4_3_ih_to_xcc_inst(struct amdgpu_device *adev, int ih_node)
+{
+	int xcc;
+
+	xcc = hweight8(adev->gfx.xcc_mask & GENMASK(ih_node / 2, 0));
+	if (!xcc) {
+		dev_err(adev->dev, "Couldn't find xcc mapping from IH node");
+		return -EINVAL;
+	}
+
+	return xcc - 1;
+}
+
+static const struct amdgpu_gfx_funcs gfx_v9_4_3_gfx_funcs = {
+	.get_gpu_clock_counter = &gfx_v9_4_3_get_gpu_clock_counter,
+	.select_se_sh = &gfx_v9_4_3_xcc_select_se_sh,
+	.read_wave_data = &gfx_v9_4_3_read_wave_data,
+	.read_wave_sgprs = &gfx_v9_4_3_read_wave_sgprs,
+	.read_wave_vgprs = &gfx_v9_4_3_read_wave_vgprs,
+	.select_me_pipe_q = &gfx_v9_4_3_select_me_pipe_q,
+	.switch_partition_mode = &gfx_v9_4_3_switch_compute_partition,
+	.ih_node_to_logical_xcc = &gfx_v9_4_3_ih_to_xcc_inst,
+	.get_xccs_per_xcp = &gfx_v9_4_3_get_xccs_per_xcp,
+};
+
+static int gfx_v9_4_3_aca_bank_parser(struct aca_handle *handle,
+				      struct aca_bank *bank, enum aca_smu_type type,
+				      void *data)
+{
+	struct aca_bank_info info;
+	u64 misc0;
+	u32 instlo;
+	int ret;
+
+	ret = aca_bank_info_decode(bank, &info);
+	if (ret)
+		return ret;
+
+	/* NOTE: overwrite info.die_id with xcd id for gfx */
+	instlo = ACA_REG__IPID__INSTANCEIDLO(bank->regs[ACA_REG_IDX_IPID]);
+	instlo &= GENMASK(31, 1);
+	info.die_id = instlo == mmSMNAID_XCD0_MCA_SMU ? 0 : 1;
+
+	misc0 = bank->regs[ACA_REG_IDX_MISC0];
+
+	switch (type) {
+	case ACA_SMU_TYPE_UE:
+		bank->aca_err_type = ACA_ERROR_TYPE_UE;
+		ret = aca_error_cache_log_bank_error(handle, &info, bank->aca_err_type, 1ULL);
+		break;
+	case ACA_SMU_TYPE_CE:
+		bank->aca_err_type = ACA_ERROR_TYPE_CE;
+		ret = aca_error_cache_log_bank_error(handle, &info, bank->aca_err_type,
+						     ACA_REG__MISC0__ERRCNT(misc0));
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return ret;
+}
+
+static bool gfx_v9_4_3_aca_bank_is_valid(struct aca_handle *handle, struct aca_bank *bank,
+					 enum aca_smu_type type, void *data)
+{
+	u32 instlo;
+
+	instlo = ACA_REG__IPID__INSTANCEIDLO(bank->regs[ACA_REG_IDX_IPID]);
+	instlo &= GENMASK(31, 1);
+	switch (instlo) {
+	case mmSMNAID_XCD0_MCA_SMU:
+	case mmSMNAID_XCD1_MCA_SMU:
+	case mmSMNXCD_XCD0_MCA_SMU:
+		return true;
+	default:
+		break;
+	}
+
+	return false;
+}
+
+static const struct aca_bank_ops gfx_v9_4_3_aca_bank_ops = {
+	.aca_bank_parser = gfx_v9_4_3_aca_bank_parser,
+	.aca_bank_is_valid = gfx_v9_4_3_aca_bank_is_valid,
+};
+
+static const struct aca_info gfx_v9_4_3_aca_info = {
+	.hwip = ACA_HWIP_TYPE_SMU,
+	.mask = ACA_ERROR_UE_MASK | ACA_ERROR_CE_MASK,
+	.bank_ops = &gfx_v9_4_3_aca_bank_ops,
+};
+
+static int gfx_v9_4_3_gpu_early_init(struct amdgpu_device *adev)
+{
+	adev->gfx.funcs = &gfx_v9_4_3_gfx_funcs;
+	adev->gfx.ras = &gfx_v9_4_3_ras;
+
+	adev->gfx.config.max_hw_contexts = 8;
+	adev->gfx.config.sc_prim_fifo_size_frontend = 0x20;
+	adev->gfx.config.sc_prim_fifo_size_backend = 0x100;
+	adev->gfx.config.sc_hiz_tile_fifo_size = 0x30;
+	adev->gfx.config.sc_earlyz_tile_fifo_size = 0x4C0;
+	adev->gfx.config.gb_addr_config = GOLDEN_GB_ADDR_CONFIG;
+
+	adev->gfx.config.gb_addr_config_fields.num_pipes = 1 <<
+			REG_GET_FIELD(
+					adev->gfx.config.gb_addr_config,
+					GB_ADDR_CONFIG,
+					NUM_PIPES);
+
+	adev->gfx.config.max_tile_pipes =
+		adev->gfx.config.gb_addr_config_fields.num_pipes;
+
+	adev->gfx.config.gb_addr_config_fields.num_banks = 1 <<
+			REG_GET_FIELD(
+					adev->gfx.config.gb_addr_config,
+					GB_ADDR_CONFIG,
+					NUM_BANKS);
+	adev->gfx.config.gb_addr_config_fields.max_compress_frags = 1 <<
+			REG_GET_FIELD(
+					adev->gfx.config.gb_addr_config,
+					GB_ADDR_CONFIG,
+					MAX_COMPRESSED_FRAGS);
+	adev->gfx.config.gb_addr_config_fields.num_rb_per_se = 1 <<
+			REG_GET_FIELD(
+					adev->gfx.config.gb_addr_config,
+					GB_ADDR_CONFIG,
+					NUM_RB_PER_SE);
+	adev->gfx.config.gb_addr_config_fields.num_se = 1 <<
+			REG_GET_FIELD(
+					adev->gfx.config.gb_addr_config,
+					GB_ADDR_CONFIG,
+					NUM_SHADER_ENGINES);
+	adev->gfx.config.gb_addr_config_fields.pipe_interleave_size = 1 << (8 +
+			REG_GET_FIELD(
+					adev->gfx.config.gb_addr_config,
+					GB_ADDR_CONFIG,
+					PIPE_INTERLEAVE_SIZE));
+
+	return 0;
+}
+
+static int gfx_v9_4_3_compute_ring_init(struct amdgpu_device *adev, int ring_id,
+				        int xcc_id, int mec, int pipe, int queue)
+{
+	unsigned irq_type;
+	struct amdgpu_ring *ring = &adev->gfx.compute_ring[ring_id];
+	unsigned int hw_prio;
+	uint32_t xcc_doorbell_start;
+
+	ring = &adev->gfx.compute_ring[xcc_id * adev->gfx.num_compute_rings +
+				       ring_id];
+
+	/* mec0 is me1 */
+	ring->xcc_id = xcc_id;
+	ring->me = mec + 1;
+	ring->pipe = pipe;
+	ring->queue = queue;
+
+	ring->ring_obj = NULL;
+	ring->use_doorbell = true;
+	xcc_doorbell_start = adev->doorbell_index.mec_ring0 +
+			     xcc_id * adev->doorbell_index.xcc_doorbell_range;
+	ring->doorbell_index = (xcc_doorbell_start + ring_id) << 1;
+	ring->eop_gpu_addr = adev->gfx.mec.hpd_eop_gpu_addr +
+			     (ring_id + xcc_id * adev->gfx.num_compute_rings) *
+				     GFX9_MEC_HPD_SIZE;
+	ring->vm_hub = AMDGPU_GFXHUB(xcc_id);
+	sprintf(ring->name, "comp_%d.%d.%d.%d",
+			ring->xcc_id, ring->me, ring->pipe, ring->queue);
+
+	irq_type = AMDGPU_CP_IRQ_COMPUTE_MEC1_PIPE0_EOP
+		+ ((ring->me - 1) * adev->gfx.mec.num_pipe_per_mec)
+		+ ring->pipe;
+	hw_prio = amdgpu_gfx_is_high_priority_compute_queue(adev, ring) ?
+			AMDGPU_GFX_PIPE_PRIO_HIGH : AMDGPU_GFX_PIPE_PRIO_NORMAL;
+	/* type-2 packets are deprecated on MEC, use type-3 instead */
+	return amdgpu_ring_init(adev, ring, 1024, &adev->gfx.eop_irq, irq_type,
+				hw_prio, NULL);
+}
+
+static void gfx_v9_4_3_alloc_ip_dump(struct amdgpu_device *adev)
+{
+	uint32_t reg_count = ARRAY_SIZE(gc_reg_list_9_4_3);
+	uint32_t *ptr, num_xcc, inst;
+
+	num_xcc = NUM_XCC(adev->gfx.xcc_mask);
+
+	ptr = kcalloc(reg_count * num_xcc, sizeof(uint32_t), GFP_KERNEL);
+	if (!ptr) {
+		DRM_ERROR("Failed to allocate memory for GFX IP Dump\n");
+		adev->gfx.ip_dump_core = NULL;
+	} else {
+		adev->gfx.ip_dump_core = ptr;
+	}
+
+	/* Allocate memory for compute queue registers for all the instances */
+	reg_count = ARRAY_SIZE(gc_cp_reg_list_9_4_3);
+	inst = adev->gfx.mec.num_mec * adev->gfx.mec.num_pipe_per_mec *
+		adev->gfx.mec.num_queue_per_pipe;
+
+	ptr = kcalloc(reg_count * inst * num_xcc, sizeof(uint32_t), GFP_KERNEL);
+	if (!ptr) {
+		DRM_ERROR("Failed to allocate memory for Compute Queues IP Dump\n");
+		adev->gfx.ip_dump_compute_queues = NULL;
+	} else {
+		adev->gfx.ip_dump_compute_queues = ptr;
+	}
+}
+
+static int gfx_v9_4_3_sw_init(struct amdgpu_ip_block *ip_block)
+{
+	int i, j, k, r, ring_id, xcc_id, num_xcc;
+	struct amdgpu_device *adev = ip_block->adev;
+
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	case IP_VERSION(9, 4, 3):
+	case IP_VERSION(9, 4, 4):
+		adev->gfx.cleaner_shader_ptr = gfx_9_4_3_cleaner_shader_hex;
+		adev->gfx.cleaner_shader_size = sizeof(gfx_9_4_3_cleaner_shader_hex);
+		if (adev->gfx.mec_fw_version >= 153) {
+			adev->gfx.enable_cleaner_shader = true;
+			r = amdgpu_gfx_cleaner_shader_sw_init(adev, adev->gfx.cleaner_shader_size);
+			if (r) {
+				adev->gfx.enable_cleaner_shader = false;
+				dev_err(adev->dev, "Failed to initialize cleaner shader\n");
+			}
+		}
+		break;
+	default:
+		adev->gfx.enable_cleaner_shader = false;
+		break;
+	}
+
+	adev->gfx.mec.num_mec = 2;
+	adev->gfx.mec.num_pipe_per_mec = 4;
+	adev->gfx.mec.num_queue_per_pipe = 8;
+
+	num_xcc = NUM_XCC(adev->gfx.xcc_mask);
+
+	/* EOP Event */
+	r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_GRBM_CP, GFX_9_0__SRCID__CP_EOP_INTERRUPT, &adev->gfx.eop_irq);
+	if (r)
+		return r;
+
+	/* Bad opcode Event */
+	r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_GRBM_CP,
+			      GFX_9_0__SRCID__CP_BAD_OPCODE_ERROR,
+			      &adev->gfx.bad_op_irq);
+	if (r)
+		return r;
+
+	/* Privileged reg */
+	r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_GRBM_CP, GFX_9_0__SRCID__CP_PRIV_REG_FAULT,
+			      &adev->gfx.priv_reg_irq);
+	if (r)
+		return r;
+
+	/* Privileged inst */
+	r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_GRBM_CP, GFX_9_0__SRCID__CP_PRIV_INSTR_FAULT,
+			      &adev->gfx.priv_inst_irq);
+	if (r)
+		return r;
+
+	adev->gfx.gfx_current_status = AMDGPU_GFX_NORMAL_MODE;
+
+	r = adev->gfx.rlc.funcs->init(adev);
+	if (r) {
+		DRM_ERROR("Failed to init rlc BOs!\n");
+		return r;
+	}
+
+	r = gfx_v9_4_3_mec_init(adev);
+	if (r) {
+		DRM_ERROR("Failed to init MEC BOs!\n");
+		return r;
+	}
+
+	/* set up the compute queues - allocate horizontally across pipes */
+	for (xcc_id = 0; xcc_id < num_xcc; xcc_id++) {
+		ring_id = 0;
+		for (i = 0; i < adev->gfx.mec.num_mec; ++i) {
+			for (j = 0; j < adev->gfx.mec.num_queue_per_pipe; j++) {
+				for (k = 0; k < adev->gfx.mec.num_pipe_per_mec;
+				     k++) {
+					if (!amdgpu_gfx_is_mec_queue_enabled(
+							adev, xcc_id, i, k, j))
+						continue;
+
+					r = gfx_v9_4_3_compute_ring_init(adev,
+								       ring_id,
+								       xcc_id,
+								       i, k, j);
+					if (r)
+						return r;
+
+					ring_id++;
+				}
+			}
+		}
+
+		r = amdgpu_gfx_kiq_init(adev, GFX9_MEC_HPD_SIZE, xcc_id);
+		if (r) {
+			DRM_ERROR("Failed to init KIQ BOs!\n");
+			return r;
+		}
+
+		r = amdgpu_gfx_kiq_init_ring(adev, xcc_id);
+		if (r)
+			return r;
+
+		/* create MQD for all compute queues as wel as KIQ for SRIOV case */
+		r = amdgpu_gfx_mqd_sw_init(adev,
+				sizeof(struct v9_mqd_allocation), xcc_id);
+		if (r)
+			return r;
+	}
+
+	adev->gfx.compute_supported_reset =
+		amdgpu_get_soft_full_reset_mask(&adev->gfx.compute_ring[0]);
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	case IP_VERSION(9, 4, 3):
+	case IP_VERSION(9, 4, 4):
+		if ((adev->gfx.mec_fw_version >= 155) &&
+		    !amdgpu_sriov_vf(adev) &&
+		    !adev->debug_disable_gpu_ring_reset) {
+			adev->gfx.compute_supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE;
+			adev->gfx.compute_supported_reset |= AMDGPU_RESET_TYPE_PER_PIPE;
+		}
+		break;
+	case IP_VERSION(9, 5, 0):
+		if ((adev->gfx.mec_fw_version >= 21) &&
+		    !amdgpu_sriov_vf(adev) &&
+		    !adev->debug_disable_gpu_ring_reset) {
+			adev->gfx.compute_supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE;
+			adev->gfx.compute_supported_reset |= AMDGPU_RESET_TYPE_PER_PIPE;
+		}
+		break;
+	default:
+		break;
+	}
+	r = gfx_v9_4_3_gpu_early_init(adev);
+	if (r)
+		return r;
+
+	r = amdgpu_gfx_ras_sw_init(adev);
+	if (r)
+		return r;
+
+	r = amdgpu_gfx_sysfs_init(adev);
+	if (r)
+		return r;
+
+	gfx_v9_4_3_alloc_ip_dump(adev);
+
+	return 0;
+}
+
+static int gfx_v9_4_3_sw_fini(struct amdgpu_ip_block *ip_block)
+{
+	int i, num_xcc;
+	struct amdgpu_device *adev = ip_block->adev;
+
+	num_xcc = NUM_XCC(adev->gfx.xcc_mask);
+	for (i = 0; i < adev->gfx.num_compute_rings * num_xcc; i++)
+		amdgpu_ring_fini(&adev->gfx.compute_ring[i]);
+
+	for (i = 0; i < num_xcc; i++) {
+		amdgpu_gfx_mqd_sw_fini(adev, i);
+		amdgpu_gfx_kiq_free_ring(&adev->gfx.kiq[i].ring);
+		amdgpu_gfx_kiq_fini(adev, i);
+	}
+
+	amdgpu_gfx_cleaner_shader_sw_fini(adev);
+
+	gfx_v9_4_3_mec_fini(adev);
+	amdgpu_bo_unref(&adev->gfx.rlc.clear_state_obj);
+	gfx_v9_4_3_free_microcode(adev);
+	amdgpu_gfx_sysfs_fini(adev);
+
+	kfree(adev->gfx.ip_dump_core);
+	kfree(adev->gfx.ip_dump_compute_queues);
+
+	return 0;
+}
+
+#define DEFAULT_SH_MEM_BASES	(0x6000)
+static void gfx_v9_4_3_xcc_init_compute_vmid(struct amdgpu_device *adev,
+					     int xcc_id)
+{
+	int i;
+	uint32_t sh_mem_config;
+	uint32_t sh_mem_bases;
+	uint32_t data;
+
+	/*
+	 * Configure apertures:
+	 * LDS:         0x60000000'00000000 - 0x60000001'00000000 (4GB)
+	 * Scratch:     0x60000001'00000000 - 0x60000002'00000000 (4GB)
+	 * GPUVM:       0x60010000'00000000 - 0x60020000'00000000 (1TB)
+	 */
+	sh_mem_bases = DEFAULT_SH_MEM_BASES | (DEFAULT_SH_MEM_BASES << 16);
+
+	sh_mem_config = SH_MEM_ADDRESS_MODE_64 |
+			SH_MEM_ALIGNMENT_MODE_UNALIGNED <<
+			SH_MEM_CONFIG__ALIGNMENT_MODE__SHIFT;
+
+	mutex_lock(&adev->srbm_mutex);
+	for (i = adev->vm_manager.first_kfd_vmid; i < AMDGPU_NUM_VMID; i++) {
+		soc15_grbm_select(adev, 0, 0, 0, i, GET_INST(GC, xcc_id));
+		/* CP and shaders */
+		WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regSH_MEM_CONFIG, sh_mem_config);
+		WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regSH_MEM_BASES, sh_mem_bases);
+
+		/* Enable trap for each kfd vmid. */
+		data = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regSPI_GDBG_PER_VMID_CNTL);
+		data = REG_SET_FIELD(data, SPI_GDBG_PER_VMID_CNTL, TRAP_EN, 1);
+		WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regSPI_GDBG_PER_VMID_CNTL, data);
+	}
+	soc15_grbm_select(adev, 0, 0, 0, 0, GET_INST(GC, xcc_id));
+	mutex_unlock(&adev->srbm_mutex);
+
+	/*
+	 * Initialize all compute VMIDs to have no GDS, GWS, or OA
+	 * access. These should be enabled by FW for target VMIDs.
+	 */
+	for (i = adev->vm_manager.first_kfd_vmid; i < AMDGPU_NUM_VMID; i++) {
+		WREG32_SOC15_OFFSET(GC, GET_INST(GC, xcc_id), regGDS_VMID0_BASE, 2 * i, 0);
+		WREG32_SOC15_OFFSET(GC, GET_INST(GC, xcc_id), regGDS_VMID0_SIZE, 2 * i, 0);
+		WREG32_SOC15_OFFSET(GC, GET_INST(GC, xcc_id), regGDS_GWS_VMID0, i, 0);
+		WREG32_SOC15_OFFSET(GC, GET_INST(GC, xcc_id), regGDS_OA_VMID0, i, 0);
+	}
+}
+
+static void gfx_v9_4_3_xcc_init_gds_vmid(struct amdgpu_device *adev, int xcc_id)
+{
+	int vmid;
+
+	/*
+	 * Initialize all compute and user-gfx VMIDs to have no GDS, GWS, or OA
+	 * access. Compute VMIDs should be enabled by FW for target VMIDs,
+	 * the driver can enable them for graphics. VMID0 should maintain
+	 * access so that HWS firmware can save/restore entries.
+	 */
+	for (vmid = 1; vmid < AMDGPU_NUM_VMID; vmid++) {
+		WREG32_SOC15_OFFSET(GC, GET_INST(GC, xcc_id), regGDS_VMID0_BASE, 2 * vmid, 0);
+		WREG32_SOC15_OFFSET(GC, GET_INST(GC, xcc_id), regGDS_VMID0_SIZE, 2 * vmid, 0);
+		WREG32_SOC15_OFFSET(GC, GET_INST(GC, xcc_id), regGDS_GWS_VMID0, vmid, 0);
+		WREG32_SOC15_OFFSET(GC, GET_INST(GC, xcc_id), regGDS_OA_VMID0, vmid, 0);
+	}
+}
+
+/* For ASICs that needs xnack chain and MEC version supports, set SG_CONFIG1
+ * DISABLE_XNACK_CHECK_IN_RETRY_DISABLE bit and inform KFD to set xnack_chain
+ * bit in SET_RESOURCES
+ */
+static void gfx_v9_4_3_xcc_init_sq(struct amdgpu_device *adev, int xcc_id)
+{
+	uint32_t data;
+
+	if (!(adev->gmc.xnack_flags & AMDGPU_GMC_XNACK_FLAG_CHAIN))
+		return;
+
+	data = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regSQ_CONFIG1);
+	data = REG_SET_FIELD(data, SQ_CONFIG1, DISABLE_XNACK_CHECK_IN_RETRY_DISABLE, 1);
+	WREG32_SOC15(GC, xcc_id, regSQ_CONFIG1, data);
+}
+
+static void gfx_v9_4_3_xcc_constants_init(struct amdgpu_device *adev,
+					  int xcc_id)
+{
+	u32 tmp;
+	int i;
+
+	/* XXX SH_MEM regs */
+	/* where to put LDS, scratch, GPUVM in FSA64 space */
+	mutex_lock(&adev->srbm_mutex);
+	for (i = 0; i < adev->vm_manager.id_mgr[AMDGPU_GFXHUB(0)].num_ids; i++) {
+		soc15_grbm_select(adev, 0, 0, 0, i, GET_INST(GC, xcc_id));
+		/* CP and shaders */
+		if (i == 0) {
+			tmp = REG_SET_FIELD(0, SH_MEM_CONFIG, ALIGNMENT_MODE,
+					    SH_MEM_ALIGNMENT_MODE_UNALIGNED);
+			tmp = REG_SET_FIELD(tmp, SH_MEM_CONFIG, RETRY_DISABLE,
+					    !!adev->gmc.noretry);
+			WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id),
+					 regSH_MEM_CONFIG, tmp);
+			WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id),
+					 regSH_MEM_BASES, 0);
+		} else {
+			tmp = REG_SET_FIELD(0, SH_MEM_CONFIG, ALIGNMENT_MODE,
+					    SH_MEM_ALIGNMENT_MODE_UNALIGNED);
+			tmp = REG_SET_FIELD(tmp, SH_MEM_CONFIG, RETRY_DISABLE,
+					    !!adev->gmc.noretry);
+			WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id),
+					 regSH_MEM_CONFIG, tmp);
+			tmp = REG_SET_FIELD(0, SH_MEM_BASES, PRIVATE_BASE,
+					    (adev->gmc.private_aperture_start >>
+					     48));
+			tmp = REG_SET_FIELD(tmp, SH_MEM_BASES, SHARED_BASE,
+					    (adev->gmc.shared_aperture_start >>
+					     48));
+			WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id),
+					 regSH_MEM_BASES, tmp);
+		}
+	}
+	soc15_grbm_select(adev, 0, 0, 0, 0, GET_INST(GC, 0));
+
+	mutex_unlock(&adev->srbm_mutex);
+
+	gfx_v9_4_3_xcc_init_compute_vmid(adev, xcc_id);
+	gfx_v9_4_3_xcc_init_gds_vmid(adev, xcc_id);
+	gfx_v9_4_3_xcc_init_sq(adev, xcc_id);
+}
+
+static void gfx_v9_4_3_constants_init(struct amdgpu_device *adev)
+{
+	int i, num_xcc;
+
+	num_xcc = NUM_XCC(adev->gfx.xcc_mask);
+
+	gfx_v9_4_3_get_cu_info(adev, &adev->gfx.cu_info);
+	adev->gfx.config.db_debug2 =
+		RREG32_SOC15(GC, GET_INST(GC, 0), regDB_DEBUG2);
+
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	/* ToDo: GC 9.4.4 */
+	case IP_VERSION(9, 4, 3):
+		if (adev->gfx.mec_fw_version >= 184 &&
+		    (amdgpu_sriov_reg_access_sq_config(adev) ||
+		     !amdgpu_sriov_vf(adev)))
+			adev->gmc.xnack_flags |= AMDGPU_GMC_XNACK_FLAG_CHAIN;
+		break;
+	case IP_VERSION(9, 5, 0):
+		if (adev->gfx.mec_fw_version >= 23)
+			adev->gmc.xnack_flags |= AMDGPU_GMC_XNACK_FLAG_CHAIN;
+		break;
+	default:
+		break;
+	}
+
+	for (i = 0; i < num_xcc; i++)
+		gfx_v9_4_3_xcc_constants_init(adev, i);
+}
+
+static void
+gfx_v9_4_3_xcc_enable_save_restore_machine(struct amdgpu_device *adev,
+					   int xcc_id)
+{
+	WREG32_FIELD15_PREREG(GC, GET_INST(GC, xcc_id), RLC_SRM_CNTL, SRM_ENABLE, 1);
+}
+
+static void gfx_v9_4_3_xcc_init_pg(struct amdgpu_device *adev, int xcc_id)
+{
+	/*
+	 * Rlc save restore list is workable since v2_1.
+	 */
+	gfx_v9_4_3_xcc_enable_save_restore_machine(adev, xcc_id);
+}
+
+static void gfx_v9_4_3_xcc_disable_gpa_mode(struct amdgpu_device *adev, int xcc_id)
+{
+	uint32_t data;
+
+	data = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regCPC_PSP_DEBUG);
+	data |= CPC_PSP_DEBUG__UTCL2IUGPAOVERRIDE_MASK;
+	WREG32_SOC15(GC, GET_INST(GC, xcc_id), regCPC_PSP_DEBUG, data);
+}
+
+static bool gfx_v9_4_3_is_rlc_enabled(struct amdgpu_device *adev)
+{
+	uint32_t rlc_setting;
+
+	/* if RLC is not enabled, do nothing */
+	rlc_setting = RREG32_SOC15(GC, GET_INST(GC, 0), regRLC_CNTL);
+	if (!(rlc_setting & RLC_CNTL__RLC_ENABLE_F32_MASK))
+		return false;
+
+	return true;
+}
+
+static void gfx_v9_4_3_xcc_set_safe_mode(struct amdgpu_device *adev, int xcc_id)
+{
+	uint32_t data;
+	unsigned i;
+
+	data = RLC_SAFE_MODE__CMD_MASK;
+	data |= (1 << RLC_SAFE_MODE__MESSAGE__SHIFT);
+	WREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_SAFE_MODE, data);
+
+	/* wait for RLC_SAFE_MODE */
+	for (i = 0; i < adev->usec_timeout; i++) {
+		if (!REG_GET_FIELD(RREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_SAFE_MODE), RLC_SAFE_MODE, CMD))
+			break;
+		udelay(1);
+	}
+}
+
+static void gfx_v9_4_3_xcc_unset_safe_mode(struct amdgpu_device *adev,
+					   int xcc_id)
+{
+	uint32_t data;
+
+	data = RLC_SAFE_MODE__CMD_MASK;
+	WREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_SAFE_MODE, data);
+}
+
+static void gfx_v9_4_3_init_rlcg_reg_access_ctrl(struct amdgpu_device *adev)
+{
+	int xcc_id, num_xcc;
+	struct amdgpu_rlcg_reg_access_ctrl *reg_access_ctrl;
+
+	num_xcc = NUM_XCC(adev->gfx.xcc_mask);
+	for (xcc_id = 0; xcc_id < num_xcc; xcc_id++) {
+		reg_access_ctrl = &adev->gfx.rlc.reg_access_ctrl[GET_INST(GC, xcc_id)];
+		reg_access_ctrl->scratch_reg0 = SOC15_REG_OFFSET(GC, GET_INST(GC, xcc_id), regSCRATCH_REG0);
+		reg_access_ctrl->scratch_reg1 = SOC15_REG_OFFSET(GC, GET_INST(GC, xcc_id), regSCRATCH_REG1);
+		reg_access_ctrl->scratch_reg2 = SOC15_REG_OFFSET(GC, GET_INST(GC, xcc_id), regSCRATCH_REG2);
+		reg_access_ctrl->scratch_reg3 = SOC15_REG_OFFSET(GC, GET_INST(GC, xcc_id), regSCRATCH_REG3);
+		reg_access_ctrl->grbm_cntl = SOC15_REG_OFFSET(GC, GET_INST(GC, xcc_id), regGRBM_GFX_CNTL);
+		reg_access_ctrl->grbm_idx = SOC15_REG_OFFSET(GC, GET_INST(GC, xcc_id), regGRBM_GFX_INDEX);
+		reg_access_ctrl->spare_int = SOC15_REG_OFFSET(GC, GET_INST(GC, xcc_id), regRLC_SPARE_INT);
+	}
+	adev->gfx.rlc.rlcg_reg_access_supported = true;
+}
+
+static int gfx_v9_4_3_rlc_init(struct amdgpu_device *adev)
+{
+	/* init spm vmid with 0xf */
+	if (adev->gfx.rlc.funcs->update_spm_vmid)
+		adev->gfx.rlc.funcs->update_spm_vmid(adev, NULL, 0xf);
+
+	return 0;
+}
+
+static void gfx_v9_4_3_xcc_wait_for_rlc_serdes(struct amdgpu_device *adev,
+					       int xcc_id)
+{
+	u32 i, j, k;
+	u32 mask;
+
+	mutex_lock(&adev->grbm_idx_mutex);
+	for (i = 0; i < adev->gfx.config.max_shader_engines; i++) {
+		for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) {
+			gfx_v9_4_3_xcc_select_se_sh(adev, i, j, 0xffffffff,
+						    xcc_id);
+			for (k = 0; k < adev->usec_timeout; k++) {
+				if (RREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_SERDES_CU_MASTER_BUSY) == 0)
+					break;
+				udelay(1);
+			}
+			if (k == adev->usec_timeout) {
+				gfx_v9_4_3_xcc_select_se_sh(adev, 0xffffffff,
+							    0xffffffff,
+							    0xffffffff, xcc_id);
+				mutex_unlock(&adev->grbm_idx_mutex);
+				DRM_INFO("Timeout wait for RLC serdes %u,%u\n",
+					 i, j);
+				return;
+			}
+		}
+	}
+	gfx_v9_4_3_xcc_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff,
+				    xcc_id);
+	mutex_unlock(&adev->grbm_idx_mutex);
+
+	mask = RLC_SERDES_NONCU_MASTER_BUSY__SE_MASTER_BUSY_MASK |
+		RLC_SERDES_NONCU_MASTER_BUSY__GC_MASTER_BUSY_MASK |
+		RLC_SERDES_NONCU_MASTER_BUSY__TC0_MASTER_BUSY_MASK |
+		RLC_SERDES_NONCU_MASTER_BUSY__TC1_MASTER_BUSY_MASK;
+	for (k = 0; k < adev->usec_timeout; k++) {
+		if ((RREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_SERDES_NONCU_MASTER_BUSY) & mask) == 0)
+			break;
+		udelay(1);
+	}
+}
+
+static void gfx_v9_4_3_xcc_enable_gui_idle_interrupt(struct amdgpu_device *adev,
+						     bool enable, int xcc_id)
+{
+	u32 tmp;
+
+	/* These interrupts should be enabled to drive DS clock */
+
+	tmp = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_INT_CNTL_RING0);
+
+	tmp = REG_SET_FIELD(tmp, CP_INT_CNTL_RING0, CNTX_BUSY_INT_ENABLE, enable ? 1 : 0);
+	tmp = REG_SET_FIELD(tmp, CP_INT_CNTL_RING0, CNTX_EMPTY_INT_ENABLE, enable ? 1 : 0);
+	tmp = REG_SET_FIELD(tmp, CP_INT_CNTL_RING0, CMP_BUSY_INT_ENABLE, enable ? 1 : 0);
+
+	WREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_INT_CNTL_RING0, tmp);
+}
+
+static void gfx_v9_4_3_xcc_rlc_stop(struct amdgpu_device *adev, int xcc_id)
+{
+	WREG32_FIELD15_PREREG(GC, GET_INST(GC, xcc_id), RLC_CNTL,
+			      RLC_ENABLE_F32, 0);
+	gfx_v9_4_3_xcc_enable_gui_idle_interrupt(adev, false, xcc_id);
+	gfx_v9_4_3_xcc_wait_for_rlc_serdes(adev, xcc_id);
+}
+
+static void gfx_v9_4_3_rlc_stop(struct amdgpu_device *adev)
+{
+	int i, num_xcc;
+
+	num_xcc = NUM_XCC(adev->gfx.xcc_mask);
+	for (i = 0; i < num_xcc; i++)
+		gfx_v9_4_3_xcc_rlc_stop(adev, i);
+}
+
+static void gfx_v9_4_3_xcc_rlc_reset(struct amdgpu_device *adev, int xcc_id)
+{
+	WREG32_FIELD15_PREREG(GC, GET_INST(GC, xcc_id), GRBM_SOFT_RESET,
+			      SOFT_RESET_RLC, 1);
+	udelay(50);
+	WREG32_FIELD15_PREREG(GC, GET_INST(GC, xcc_id), GRBM_SOFT_RESET,
+			      SOFT_RESET_RLC, 0);
+	udelay(50);
+}
+
+static void gfx_v9_4_3_rlc_reset(struct amdgpu_device *adev)
+{
+	int i, num_xcc;
+
+	num_xcc = NUM_XCC(adev->gfx.xcc_mask);
+	for (i = 0; i < num_xcc; i++)
+		gfx_v9_4_3_xcc_rlc_reset(adev, i);
+}
+
+static void gfx_v9_4_3_xcc_rlc_start(struct amdgpu_device *adev, int xcc_id)
+{
+	WREG32_FIELD15_PREREG(GC, GET_INST(GC, xcc_id), RLC_CNTL,
+			      RLC_ENABLE_F32, 1);
+	udelay(50);
+
+	/* carrizo do enable cp interrupt after cp inited */
+	if (!(adev->flags & AMD_IS_APU)) {
+		gfx_v9_4_3_xcc_enable_gui_idle_interrupt(adev, true, xcc_id);
+		udelay(50);
+	}
+}
+
+static void gfx_v9_4_3_rlc_start(struct amdgpu_device *adev)
+{
+#ifdef AMDGPU_RLC_DEBUG_RETRY
+	u32 rlc_ucode_ver;
+#endif
+	int i, num_xcc;
+
+	num_xcc = NUM_XCC(adev->gfx.xcc_mask);
+	for (i = 0; i < num_xcc; i++) {
+		gfx_v9_4_3_xcc_rlc_start(adev, i);
+#ifdef AMDGPU_RLC_DEBUG_RETRY
+		/* RLC_GPM_GENERAL_6 : RLC Ucode version */
+		rlc_ucode_ver = RREG32_SOC15(GC, GET_INST(GC, i), regRLC_GPM_GENERAL_6);
+		if (rlc_ucode_ver == 0x108) {
+			dev_info(adev->dev,
+				 "Using rlc debug ucode. regRLC_GPM_GENERAL_6 ==0x08%x / fw_ver == %i \n",
+				 rlc_ucode_ver, adev->gfx.rlc_fw_version);
+			/* RLC_GPM_TIMER_INT_3 : Timer interval in RefCLK cycles,
+			 * default is 0x9C4 to create a 100us interval */
+			WREG32_SOC15(GC, GET_INST(GC, i), regRLC_GPM_TIMER_INT_3, 0x9C4);
+			/* RLC_GPM_GENERAL_12 : Minimum gap between wptr and rptr
+			 * to disable the page fault retry interrupts, default is
+			 * 0x100 (256) */
+			WREG32_SOC15(GC, GET_INST(GC, i), regRLC_GPM_GENERAL_12, 0x100);
+		}
+#endif
+	}
+}
+
+static int gfx_v9_4_3_xcc_rlc_load_microcode(struct amdgpu_device *adev,
+					     int xcc_id)
+{
+	const struct rlc_firmware_header_v2_0 *hdr;
+	const __le32 *fw_data;
+	unsigned i, fw_size;
+
+	if (!adev->gfx.rlc_fw)
+		return -EINVAL;
+
+	hdr = (const struct rlc_firmware_header_v2_0 *)adev->gfx.rlc_fw->data;
+	amdgpu_ucode_print_rlc_hdr(&hdr->header);
+
+	fw_data = (const __le32 *)(adev->gfx.rlc_fw->data +
+			   le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+	fw_size = le32_to_cpu(hdr->header.ucode_size_bytes) / 4;
+
+	WREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_GPM_UCODE_ADDR,
+			RLCG_UCODE_LOADING_START_ADDRESS);
+	for (i = 0; i < fw_size; i++) {
+		if (amdgpu_emu_mode == 1 && i % 100 == 0) {
+			dev_info(adev->dev, "Write RLC ucode data %u DWs\n", i);
+			msleep(1);
+		}
+		WREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_GPM_UCODE_DATA, le32_to_cpup(fw_data++));
+	}
+	WREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_GPM_UCODE_ADDR, adev->gfx.rlc_fw_version);
+
+	return 0;
+}
+
+static int gfx_v9_4_3_xcc_rlc_resume(struct amdgpu_device *adev, int xcc_id)
+{
+	int r;
+
+	if (adev->firmware.load_type != AMDGPU_FW_LOAD_PSP) {
+		gfx_v9_4_3_xcc_rlc_stop(adev, xcc_id);
+		/* legacy rlc firmware loading */
+		r = gfx_v9_4_3_xcc_rlc_load_microcode(adev, xcc_id);
+		if (r)
+			return r;
+		gfx_v9_4_3_xcc_rlc_start(adev, xcc_id);
+	}
+
+	amdgpu_gfx_rlc_enter_safe_mode(adev, xcc_id);
+	/* disable CG */
+	WREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_CGCG_CGLS_CTRL, 0);
+	gfx_v9_4_3_xcc_init_pg(adev, xcc_id);
+	amdgpu_gfx_rlc_exit_safe_mode(adev, xcc_id);
+
+	return 0;
+}
+
+static int gfx_v9_4_3_rlc_resume(struct amdgpu_device *adev)
+{
+	int r, i, num_xcc;
+
+	if (amdgpu_sriov_vf(adev))
+		return 0;
+
+	num_xcc = NUM_XCC(adev->gfx.xcc_mask);
+	for (i = 0; i < num_xcc; i++) {
+		r = gfx_v9_4_3_xcc_rlc_resume(adev, i);
+		if (r)
+			return r;
+	}
+
+	return 0;
+}
+
+static void gfx_v9_4_3_update_spm_vmid(struct amdgpu_device *adev, struct amdgpu_ring *ring,
+				       unsigned vmid)
+{
+	u32 reg, pre_data, data;
+
+	reg = SOC15_REG_OFFSET(GC, GET_INST(GC, 0), regRLC_SPM_MC_CNTL);
+	if (amdgpu_sriov_is_pp_one_vf(adev) && !amdgpu_sriov_runtime(adev))
+		pre_data = RREG32_NO_KIQ(reg);
+	else
+		pre_data = RREG32(reg);
+
+	data =	pre_data & (~RLC_SPM_MC_CNTL__RLC_SPM_VMID_MASK);
+	data |= (vmid & RLC_SPM_MC_CNTL__RLC_SPM_VMID_MASK) << RLC_SPM_MC_CNTL__RLC_SPM_VMID__SHIFT;
+
+	if (pre_data != data) {
+		if (amdgpu_sriov_is_pp_one_vf(adev) && !amdgpu_sriov_runtime(adev)) {
+			WREG32_SOC15_NO_KIQ(GC, GET_INST(GC, 0), regRLC_SPM_MC_CNTL, data);
+		} else
+			WREG32_SOC15(GC, GET_INST(GC, 0), regRLC_SPM_MC_CNTL, data);
+	}
+}
+
+static const struct soc15_reg_rlcg rlcg_access_gc_9_4_3[] = {
+	{SOC15_REG_ENTRY(GC, 0, regGRBM_GFX_INDEX)},
+	{SOC15_REG_ENTRY(GC, 0, regSQ_IND_INDEX)},
+};
+
+static bool gfx_v9_4_3_check_rlcg_range(struct amdgpu_device *adev,
+					uint32_t offset,
+					struct soc15_reg_rlcg *entries, int arr_size)
+{
+	int i, inst;
+	uint32_t reg;
+
+	if (!entries)
+		return false;
+
+	for (i = 0; i < arr_size; i++) {
+		const struct soc15_reg_rlcg *entry;
+
+		entry = &entries[i];
+		inst = adev->ip_map.logical_to_dev_inst ?
+			       adev->ip_map.logical_to_dev_inst(
+				       adev, entry->hwip, entry->instance) :
+			       entry->instance;
+		reg = adev->reg_offset[entry->hwip][inst][entry->segment] +
+		      entry->reg;
+		if (offset == reg)
+			return true;
+	}
+
+	return false;
+}
+
+static bool gfx_v9_4_3_is_rlcg_access_range(struct amdgpu_device *adev, u32 offset)
+{
+	return gfx_v9_4_3_check_rlcg_range(adev, offset,
+					(void *)rlcg_access_gc_9_4_3,
+					ARRAY_SIZE(rlcg_access_gc_9_4_3));
+}
+
+static void gfx_v9_4_3_xcc_cp_compute_enable(struct amdgpu_device *adev,
+					     bool enable, int xcc_id)
+{
+	if (enable) {
+		WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_MEC_CNTL, 0);
+	} else {
+		WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_MEC_CNTL,
+			(CP_MEC_CNTL__MEC_INVALIDATE_ICACHE_MASK |
+			 CP_MEC_CNTL__MEC_ME1_PIPE0_RESET_MASK |
+			 CP_MEC_CNTL__MEC_ME1_PIPE1_RESET_MASK |
+			 CP_MEC_CNTL__MEC_ME1_PIPE2_RESET_MASK |
+			 CP_MEC_CNTL__MEC_ME1_PIPE3_RESET_MASK |
+			 CP_MEC_CNTL__MEC_ME2_PIPE0_RESET_MASK |
+			 CP_MEC_CNTL__MEC_ME2_PIPE1_RESET_MASK |
+			 CP_MEC_CNTL__MEC_ME1_HALT_MASK |
+			 CP_MEC_CNTL__MEC_ME2_HALT_MASK));
+		adev->gfx.kiq[xcc_id].ring.sched.ready = false;
+	}
+	udelay(50);
+}
+
+static int gfx_v9_4_3_xcc_cp_compute_load_microcode(struct amdgpu_device *adev,
+						    int xcc_id)
+{
+	const struct gfx_firmware_header_v1_0 *mec_hdr;
+	const __le32 *fw_data;
+	unsigned i;
+	u32 tmp;
+	u32 mec_ucode_addr_offset;
+	u32 mec_ucode_data_offset;
+
+	if (!adev->gfx.mec_fw)
+		return -EINVAL;
+
+	gfx_v9_4_3_xcc_cp_compute_enable(adev, false, xcc_id);
+
+	mec_hdr = (const struct gfx_firmware_header_v1_0 *)adev->gfx.mec_fw->data;
+	amdgpu_ucode_print_gfx_hdr(&mec_hdr->header);
+
+	fw_data = (const __le32 *)
+		(adev->gfx.mec_fw->data +
+		 le32_to_cpu(mec_hdr->header.ucode_array_offset_bytes));
+	tmp = 0;
+	tmp = REG_SET_FIELD(tmp, CP_CPC_IC_BASE_CNTL, VMID, 0);
+	tmp = REG_SET_FIELD(tmp, CP_CPC_IC_BASE_CNTL, CACHE_POLICY, 0);
+	WREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_CPC_IC_BASE_CNTL, tmp);
+
+	WREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_CPC_IC_BASE_LO,
+		adev->gfx.mec.mec_fw_gpu_addr & 0xFFFFF000);
+	WREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_CPC_IC_BASE_HI,
+		upper_32_bits(adev->gfx.mec.mec_fw_gpu_addr));
+
+	mec_ucode_addr_offset =
+		SOC15_REG_OFFSET(GC, GET_INST(GC, xcc_id), regCP_MEC_ME1_UCODE_ADDR);
+	mec_ucode_data_offset =
+		SOC15_REG_OFFSET(GC, GET_INST(GC, xcc_id), regCP_MEC_ME1_UCODE_DATA);
+
+	/* MEC1 */
+	WREG32(mec_ucode_addr_offset, mec_hdr->jt_offset);
+	for (i = 0; i < mec_hdr->jt_size; i++)
+		WREG32(mec_ucode_data_offset,
+		       le32_to_cpup(fw_data + mec_hdr->jt_offset + i));
+
+	WREG32(mec_ucode_addr_offset, adev->gfx.mec_fw_version);
+	/* Todo : Loading MEC2 firmware is only necessary if MEC2 should run different microcode than MEC1. */
+
+	return 0;
+}
+
+/* KIQ functions */
+static void gfx_v9_4_3_xcc_kiq_setting(struct amdgpu_ring *ring, int xcc_id)
+{
+	uint32_t tmp;
+	struct amdgpu_device *adev = ring->adev;
+
+	/* tell RLC which is KIQ queue */
+	tmp = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_CP_SCHEDULERS);
+	tmp &= 0xffffff00;
+	tmp |= (ring->me << 5) | (ring->pipe << 3) | (ring->queue);
+	WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regRLC_CP_SCHEDULERS, tmp | 0x80);
+}
+
+static void gfx_v9_4_3_mqd_set_priority(struct amdgpu_ring *ring, struct v9_mqd *mqd)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	if (ring->funcs->type == AMDGPU_RING_TYPE_COMPUTE) {
+		if (amdgpu_gfx_is_high_priority_compute_queue(adev, ring)) {
+			mqd->cp_hqd_pipe_priority = AMDGPU_GFX_PIPE_PRIO_HIGH;
+			mqd->cp_hqd_queue_priority =
+				AMDGPU_GFX_QUEUE_PRIORITY_MAXIMUM;
+		}
+	}
+}
+
+static int gfx_v9_4_3_xcc_mqd_init(struct amdgpu_ring *ring, int xcc_id)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct v9_mqd *mqd = ring->mqd_ptr;
+	uint64_t hqd_gpu_addr, wb_gpu_addr, eop_base_addr;
+	uint32_t tmp;
+
+	mqd->header = 0xC0310800;
+	mqd->compute_pipelinestat_enable = 0x00000001;
+	mqd->compute_static_thread_mgmt_se0 = 0xffffffff;
+	mqd->compute_static_thread_mgmt_se1 = 0xffffffff;
+	mqd->compute_static_thread_mgmt_se2 = 0xffffffff;
+	mqd->compute_static_thread_mgmt_se3 = 0xffffffff;
+	mqd->compute_misc_reserved = 0x00000003;
+
+	mqd->dynamic_cu_mask_addr_lo =
+		lower_32_bits(ring->mqd_gpu_addr
+			      + offsetof(struct v9_mqd_allocation, dynamic_cu_mask));
+	mqd->dynamic_cu_mask_addr_hi =
+		upper_32_bits(ring->mqd_gpu_addr
+			      + offsetof(struct v9_mqd_allocation, dynamic_cu_mask));
+
+	eop_base_addr = ring->eop_gpu_addr >> 8;
+	mqd->cp_hqd_eop_base_addr_lo = eop_base_addr;
+	mqd->cp_hqd_eop_base_addr_hi = upper_32_bits(eop_base_addr);
+
+	/* set the EOP size, register value is 2^(EOP_SIZE+1) dwords */
+	tmp = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_HQD_EOP_CONTROL);
+	tmp = REG_SET_FIELD(tmp, CP_HQD_EOP_CONTROL, EOP_SIZE,
+			(order_base_2(GFX9_MEC_HPD_SIZE / 4) - 1));
+
+	mqd->cp_hqd_eop_control = tmp;
+
+	/* enable doorbell? */
+	tmp = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_DOORBELL_CONTROL);
+
+	if (ring->use_doorbell) {
+		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
+				    DOORBELL_OFFSET, ring->doorbell_index);
+		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
+				    DOORBELL_EN, 1);
+		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
+				    DOORBELL_SOURCE, 0);
+		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
+				    DOORBELL_HIT, 0);
+		if (amdgpu_sriov_multi_vf_mode(adev))
+			tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
+					    DOORBELL_MODE, 1);
+	} else {
+		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
+					 DOORBELL_EN, 0);
+	}
+
+	mqd->cp_hqd_pq_doorbell_control = tmp;
+
+	/* disable the queue if it's active */
+	ring->wptr = 0;
+	mqd->cp_hqd_dequeue_request = 0;
+	mqd->cp_hqd_pq_rptr = 0;
+	mqd->cp_hqd_pq_wptr_lo = 0;
+	mqd->cp_hqd_pq_wptr_hi = 0;
+
+	/* set the pointer to the MQD */
+	mqd->cp_mqd_base_addr_lo = ring->mqd_gpu_addr & 0xfffffffc;
+	mqd->cp_mqd_base_addr_hi = upper_32_bits(ring->mqd_gpu_addr);
+
+	/* set MQD vmid to 0 */
+	tmp = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_MQD_CONTROL);
+	tmp = REG_SET_FIELD(tmp, CP_MQD_CONTROL, VMID, 0);
+	mqd->cp_mqd_control = tmp;
+
+	/* set the pointer to the HQD, this is similar CP_RB0_BASE/_HI */
+	hqd_gpu_addr = ring->gpu_addr >> 8;
+	mqd->cp_hqd_pq_base_lo = hqd_gpu_addr;
+	mqd->cp_hqd_pq_base_hi = upper_32_bits(hqd_gpu_addr);
+
+	/* set up the HQD, this is similar to CP_RB0_CNTL */
+	tmp = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_CONTROL);
+	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, QUEUE_SIZE,
+			    (order_base_2(ring->ring_size / 4) - 1));
+	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, RPTR_BLOCK_SIZE,
+			((order_base_2(AMDGPU_GPU_PAGE_SIZE / 4) - 1) << 8));
+#ifdef __BIG_ENDIAN
+	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, ENDIAN_SWAP, 1);
+#endif
+	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, UNORD_DISPATCH, 0);
+	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, ROQ_PQ_IB_FLIP, 0);
+	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, PRIV_STATE, 1);
+	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, KMD_QUEUE, 1);
+	mqd->cp_hqd_pq_control = tmp;
+
+	/* set the wb address whether it's enabled or not */
+	wb_gpu_addr = adev->wb.gpu_addr + (ring->rptr_offs * 4);
+	mqd->cp_hqd_pq_rptr_report_addr_lo = wb_gpu_addr & 0xfffffffc;
+	mqd->cp_hqd_pq_rptr_report_addr_hi =
+		upper_32_bits(wb_gpu_addr) & 0xffff;
+
+	/* only used if CP_PQ_WPTR_POLL_CNTL.CP_PQ_WPTR_POLL_CNTL__EN_MASK=1 */
+	wb_gpu_addr = adev->wb.gpu_addr + (ring->wptr_offs * 4);
+	mqd->cp_hqd_pq_wptr_poll_addr_lo = wb_gpu_addr & 0xfffffffc;
+	mqd->cp_hqd_pq_wptr_poll_addr_hi = upper_32_bits(wb_gpu_addr) & 0xffff;
+
+	/* reset read and write pointers, similar to CP_RB0_WPTR/_RPTR */
+	ring->wptr = 0;
+	mqd->cp_hqd_pq_rptr = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_RPTR);
+
+	/* set the vmid for the queue */
+	mqd->cp_hqd_vmid = 0;
+
+	tmp = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_HQD_PERSISTENT_STATE);
+	tmp = REG_SET_FIELD(tmp, CP_HQD_PERSISTENT_STATE, PRELOAD_SIZE, 0x53);
+	mqd->cp_hqd_persistent_state = tmp;
+
+	/* set MIN_IB_AVAIL_SIZE */
+	tmp = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_HQD_IB_CONTROL);
+	tmp = REG_SET_FIELD(tmp, CP_HQD_IB_CONTROL, MIN_IB_AVAIL_SIZE, 3);
+	mqd->cp_hqd_ib_control = tmp;
+
+	/* set static priority for a queue/ring */
+	gfx_v9_4_3_mqd_set_priority(ring, mqd);
+	mqd->cp_hqd_quantum = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_HQD_QUANTUM);
+
+	/* map_queues packet doesn't need activate the queue,
+	 * so only kiq need set this field.
+	 */
+	if (ring->funcs->type == AMDGPU_RING_TYPE_KIQ)
+		mqd->cp_hqd_active = 1;
+
+	return 0;
+}
+
+static int gfx_v9_4_3_xcc_kiq_init_register(struct amdgpu_ring *ring,
+					    int xcc_id)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct v9_mqd *mqd = ring->mqd_ptr;
+	int j;
+
+	/* disable wptr polling */
+	WREG32_FIELD15_PREREG(GC, GET_INST(GC, xcc_id), CP_PQ_WPTR_POLL_CNTL, EN, 0);
+
+	WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_EOP_BASE_ADDR,
+	       mqd->cp_hqd_eop_base_addr_lo);
+	WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_EOP_BASE_ADDR_HI,
+	       mqd->cp_hqd_eop_base_addr_hi);
+
+	/* set the EOP size, register value is 2^(EOP_SIZE+1) dwords */
+	WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_EOP_CONTROL,
+	       mqd->cp_hqd_eop_control);
+
+	/* enable doorbell? */
+	WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_DOORBELL_CONTROL,
+	       mqd->cp_hqd_pq_doorbell_control);
+
+	/* disable the queue if it's active */
+	if (RREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_HQD_ACTIVE) & 1) {
+		WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_DEQUEUE_REQUEST, 1);
+		for (j = 0; j < adev->usec_timeout; j++) {
+			if (!(RREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_HQD_ACTIVE) & 1))
+				break;
+			udelay(1);
+		}
+		WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_DEQUEUE_REQUEST,
+		       mqd->cp_hqd_dequeue_request);
+		WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_RPTR,
+		       mqd->cp_hqd_pq_rptr);
+		WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_WPTR_LO,
+		       mqd->cp_hqd_pq_wptr_lo);
+		WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_WPTR_HI,
+		       mqd->cp_hqd_pq_wptr_hi);
+	}
+
+	/* set the pointer to the MQD */
+	WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_MQD_BASE_ADDR,
+	       mqd->cp_mqd_base_addr_lo);
+	WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_MQD_BASE_ADDR_HI,
+	       mqd->cp_mqd_base_addr_hi);
+
+	/* set MQD vmid to 0 */
+	WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_MQD_CONTROL,
+	       mqd->cp_mqd_control);
+
+	/* set the pointer to the HQD, this is similar CP_RB0_BASE/_HI */
+	WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_BASE,
+	       mqd->cp_hqd_pq_base_lo);
+	WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_BASE_HI,
+	       mqd->cp_hqd_pq_base_hi);
+
+	/* set up the HQD, this is similar to CP_RB0_CNTL */
+	WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_CONTROL,
+	       mqd->cp_hqd_pq_control);
+
+	/* set the wb address whether it's enabled or not */
+	WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_RPTR_REPORT_ADDR,
+				mqd->cp_hqd_pq_rptr_report_addr_lo);
+	WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_RPTR_REPORT_ADDR_HI,
+				mqd->cp_hqd_pq_rptr_report_addr_hi);
+
+	/* only used if CP_PQ_WPTR_POLL_CNTL.CP_PQ_WPTR_POLL_CNTL__EN_MASK=1 */
+	WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_WPTR_POLL_ADDR,
+	       mqd->cp_hqd_pq_wptr_poll_addr_lo);
+	WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_WPTR_POLL_ADDR_HI,
+	       mqd->cp_hqd_pq_wptr_poll_addr_hi);
+
+	/* enable the doorbell if requested */
+	if (ring->use_doorbell) {
+		WREG32_SOC15(
+			GC, GET_INST(GC, xcc_id),
+			regCP_MEC_DOORBELL_RANGE_LOWER,
+			((adev->doorbell_index.kiq +
+			  xcc_id * adev->doorbell_index.xcc_doorbell_range) *
+			 2) << 2);
+		WREG32_SOC15(
+			GC, GET_INST(GC, xcc_id),
+			regCP_MEC_DOORBELL_RANGE_UPPER,
+			((adev->doorbell_index.userqueue_end +
+			  xcc_id * adev->doorbell_index.xcc_doorbell_range) *
+			 2) << 2);
+	}
+
+	WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_DOORBELL_CONTROL,
+	       mqd->cp_hqd_pq_doorbell_control);
+
+	/* reset read and write pointers, similar to CP_RB0_WPTR/_RPTR */
+	WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_WPTR_LO,
+	       mqd->cp_hqd_pq_wptr_lo);
+	WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_WPTR_HI,
+	       mqd->cp_hqd_pq_wptr_hi);
+
+	/* set the vmid for the queue */
+	WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_VMID, mqd->cp_hqd_vmid);
+
+	WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_PERSISTENT_STATE,
+	       mqd->cp_hqd_persistent_state);
+
+	/* activate the queue */
+	WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_ACTIVE,
+	       mqd->cp_hqd_active);
+
+	if (ring->use_doorbell)
+		WREG32_FIELD15_PREREG(GC, GET_INST(GC, xcc_id), CP_PQ_STATUS, DOORBELL_ENABLE, 1);
+
+	return 0;
+}
+
+static int gfx_v9_4_3_xcc_q_fini_register(struct amdgpu_ring *ring,
+					    int xcc_id)
+{
+	struct amdgpu_device *adev = ring->adev;
+	int j;
+
+	/* disable the queue if it's active */
+	if (RREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_HQD_ACTIVE) & 1) {
+
+		WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_DEQUEUE_REQUEST, 1);
+
+		for (j = 0; j < adev->usec_timeout; j++) {
+			if (!(RREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_HQD_ACTIVE) & 1))
+				break;
+			udelay(1);
+		}
+
+		if (j == AMDGPU_MAX_USEC_TIMEOUT) {
+			DRM_DEBUG("%s dequeue request failed.\n", ring->name);
+
+			/* Manual disable if dequeue request times out */
+			WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_ACTIVE, 0);
+		}
+
+		WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_DEQUEUE_REQUEST,
+		      0);
+	}
+
+	WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_IQ_TIMER, 0);
+	WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_IB_CONTROL, 0);
+	WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_PERSISTENT_STATE, CP_HQD_PERSISTENT_STATE_DEFAULT);
+	WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_DOORBELL_CONTROL, 0x40000000);
+	WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_DOORBELL_CONTROL, 0);
+	WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_RPTR, 0);
+	WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_WPTR_HI, 0);
+	WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_WPTR_LO, 0);
+
+	return 0;
+}
+
+static int gfx_v9_4_3_xcc_kiq_init_queue(struct amdgpu_ring *ring, int xcc_id)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct v9_mqd *mqd = ring->mqd_ptr;
+	struct v9_mqd *tmp_mqd;
+
+	gfx_v9_4_3_xcc_kiq_setting(ring, xcc_id);
+
+	/* GPU could be in bad state during probe, driver trigger the reset
+	 * after load the SMU, in this case , the mqd is not be initialized.
+	 * driver need to re-init the mqd.
+	 * check mqd->cp_hqd_pq_control since this value should not be 0
+	 */
+	tmp_mqd = (struct v9_mqd *)adev->gfx.kiq[xcc_id].mqd_backup;
+	if (amdgpu_in_reset(adev) && tmp_mqd->cp_hqd_pq_control) {
+		/* for GPU_RESET case , reset MQD to a clean status */
+		if (adev->gfx.kiq[xcc_id].mqd_backup)
+			memcpy(mqd, adev->gfx.kiq[xcc_id].mqd_backup, sizeof(struct v9_mqd_allocation));
+
+		/* reset ring buffer */
+		ring->wptr = 0;
+		amdgpu_ring_clear_ring(ring);
+		mutex_lock(&adev->srbm_mutex);
+		soc15_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0, GET_INST(GC, xcc_id));
+		gfx_v9_4_3_xcc_kiq_init_register(ring, xcc_id);
+		soc15_grbm_select(adev, 0, 0, 0, 0, GET_INST(GC, xcc_id));
+		mutex_unlock(&adev->srbm_mutex);
+	} else {
+		memset((void *)mqd, 0, sizeof(struct v9_mqd_allocation));
+		((struct v9_mqd_allocation *)mqd)->dynamic_cu_mask = 0xFFFFFFFF;
+		((struct v9_mqd_allocation *)mqd)->dynamic_rb_mask = 0xFFFFFFFF;
+		mutex_lock(&adev->srbm_mutex);
+		if (amdgpu_sriov_vf(adev) && adev->in_suspend)
+			amdgpu_ring_clear_ring(ring);
+		soc15_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0, GET_INST(GC, xcc_id));
+		gfx_v9_4_3_xcc_mqd_init(ring, xcc_id);
+		gfx_v9_4_3_xcc_kiq_init_register(ring, xcc_id);
+		soc15_grbm_select(adev, 0, 0, 0, 0, GET_INST(GC, xcc_id));
+		mutex_unlock(&adev->srbm_mutex);
+
+		if (adev->gfx.kiq[xcc_id].mqd_backup)
+			memcpy(adev->gfx.kiq[xcc_id].mqd_backup, mqd, sizeof(struct v9_mqd_allocation));
+	}
+
+	return 0;
+}
+
+static void gfx_v9_4_3_xcc_kcq_init_queue(struct amdgpu_ring *ring, int xcc_id,
+					  bool restore)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct v9_mqd *mqd = ring->mqd_ptr;
+	int mqd_idx = ring - &adev->gfx.compute_ring[0];
+	struct v9_mqd *tmp_mqd;
+
+	/* Same as above kiq init, driver need to re-init the mqd if mqd->cp_hqd_pq_control
+	 * is not be initialized before
+	 */
+	tmp_mqd = (struct v9_mqd *)adev->gfx.mec.mqd_backup[mqd_idx];
+
+	if (!restore && (!tmp_mqd->cp_hqd_pq_control ||
+	    (!amdgpu_in_reset(adev) && !adev->in_suspend))) {
+		memset((void *)mqd, 0, sizeof(struct v9_mqd_allocation));
+		((struct v9_mqd_allocation *)mqd)->dynamic_cu_mask = 0xFFFFFFFF;
+		((struct v9_mqd_allocation *)mqd)->dynamic_rb_mask = 0xFFFFFFFF;
+		mutex_lock(&adev->srbm_mutex);
+		soc15_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0, GET_INST(GC, xcc_id));
+		gfx_v9_4_3_xcc_mqd_init(ring, xcc_id);
+		soc15_grbm_select(adev, 0, 0, 0, 0, GET_INST(GC, xcc_id));
+		mutex_unlock(&adev->srbm_mutex);
+
+		if (adev->gfx.mec.mqd_backup[mqd_idx])
+			memcpy(adev->gfx.mec.mqd_backup[mqd_idx], mqd, sizeof(struct v9_mqd_allocation));
+	} else {
+		/* restore MQD to a clean status */
+		if (adev->gfx.mec.mqd_backup[mqd_idx])
+			memcpy(mqd, adev->gfx.mec.mqd_backup[mqd_idx], sizeof(struct v9_mqd_allocation));
+		/* reset ring buffer */
+		ring->wptr = 0;
+		atomic64_set((atomic64_t *)&adev->wb.wb[ring->wptr_offs], 0);
+		amdgpu_ring_clear_ring(ring);
+	}
+}
+
+static int gfx_v9_4_3_xcc_kcq_fini_register(struct amdgpu_device *adev, int xcc_id)
+{
+	struct amdgpu_ring *ring;
+	int j;
+
+	for (j = 0; j < adev->gfx.num_compute_rings; j++) {
+		ring = &adev->gfx.compute_ring[j +  xcc_id * adev->gfx.num_compute_rings];
+		if (!amdgpu_in_reset(adev) && !adev->in_suspend) {
+			mutex_lock(&adev->srbm_mutex);
+			soc15_grbm_select(adev, ring->me,
+					ring->pipe,
+					ring->queue, 0, GET_INST(GC, xcc_id));
+			gfx_v9_4_3_xcc_q_fini_register(ring, xcc_id);
+			soc15_grbm_select(adev, 0, 0, 0, 0, GET_INST(GC, xcc_id));
+			mutex_unlock(&adev->srbm_mutex);
+		}
+	}
+
+	return 0;
+}
+
+static int gfx_v9_4_3_xcc_kiq_resume(struct amdgpu_device *adev, int xcc_id)
+{
+	gfx_v9_4_3_xcc_kiq_init_queue(&adev->gfx.kiq[xcc_id].ring, xcc_id);
+	return 0;
+}
+
+static int gfx_v9_4_3_xcc_kcq_resume(struct amdgpu_device *adev, int xcc_id)
+{
+	struct amdgpu_ring *ring;
+	int i;
+
+	gfx_v9_4_3_xcc_cp_compute_enable(adev, true, xcc_id);
+
+	for (i = 0; i < adev->gfx.num_compute_rings; i++) {
+		ring = &adev->gfx.compute_ring[i + xcc_id *
+			adev->gfx.num_compute_rings];
+
+		gfx_v9_4_3_xcc_kcq_init_queue(ring, xcc_id, false);
+	}
+
+	return amdgpu_gfx_enable_kcq(adev, xcc_id);
+}
+
+static int gfx_v9_4_3_xcc_cp_resume(struct amdgpu_device *adev, int xcc_id)
+{
+	struct amdgpu_ring *ring;
+	int r, j;
+
+	gfx_v9_4_3_xcc_enable_gui_idle_interrupt(adev, false, xcc_id);
+
+	if (adev->firmware.load_type != AMDGPU_FW_LOAD_PSP) {
+		gfx_v9_4_3_xcc_disable_gpa_mode(adev, xcc_id);
+
+		r = gfx_v9_4_3_xcc_cp_compute_load_microcode(adev, xcc_id);
+		if (r)
+			return r;
+	} else {
+		gfx_v9_4_3_xcc_cp_compute_enable(adev, false, xcc_id);
+	}
+
+	r = gfx_v9_4_3_xcc_kiq_resume(adev, xcc_id);
+	if (r)
+		return r;
+
+	r = gfx_v9_4_3_xcc_kcq_resume(adev, xcc_id);
+	if (r)
+		return r;
+
+	for (j = 0; j < adev->gfx.num_compute_rings; j++) {
+		ring = &adev->gfx.compute_ring
+				[j + xcc_id * adev->gfx.num_compute_rings];
+		r = amdgpu_ring_test_helper(ring);
+		if (r)
+			return r;
+	}
+
+	gfx_v9_4_3_xcc_enable_gui_idle_interrupt(adev, true, xcc_id);
+
+	return 0;
+}
+
+static int gfx_v9_4_3_cp_resume(struct amdgpu_device *adev)
+{
+	int r = 0, i, num_xcc, num_xcp, num_xcc_per_xcp;
+
+	num_xcc = NUM_XCC(adev->gfx.xcc_mask);
+	if (amdgpu_sriov_vf(adev)) {
+		enum amdgpu_gfx_partition mode;
+
+		mode = amdgpu_xcp_query_partition_mode(adev->xcp_mgr,
+						       AMDGPU_XCP_FL_NONE);
+		if (mode == AMDGPU_UNKNOWN_COMPUTE_PARTITION_MODE)
+			return -EINVAL;
+		num_xcc_per_xcp = gfx_v9_4_3_get_xccs_per_xcp(adev);
+		adev->gfx.num_xcc_per_xcp = num_xcc_per_xcp;
+		num_xcp = num_xcc / num_xcc_per_xcp;
+		r = amdgpu_xcp_init(adev->xcp_mgr, num_xcp, mode);
+
+	} else {
+		if (adev->in_suspend)
+			amdgpu_xcp_restore_partition_mode(adev->xcp_mgr);
+		else if (amdgpu_xcp_query_partition_mode(adev->xcp_mgr,
+						    AMDGPU_XCP_FL_NONE) ==
+		    AMDGPU_UNKNOWN_COMPUTE_PARTITION_MODE)
+			r = amdgpu_xcp_switch_partition_mode(
+				adev->xcp_mgr, amdgpu_user_partt_mode);
+	}
+	if (r)
+		return r;
+
+	for (i = 0; i < num_xcc; i++) {
+		r = gfx_v9_4_3_xcc_cp_resume(adev, i);
+		if (r)
+			return r;
+	}
+
+	return 0;
+}
+
+static void gfx_v9_4_3_xcc_fini(struct amdgpu_device *adev, int xcc_id)
+{
+	if (amdgpu_gfx_disable_kcq(adev, xcc_id))
+		DRM_ERROR("XCD %d KCQ disable failed\n", xcc_id);
+
+	if (amdgpu_sriov_vf(adev)) {
+		/* must disable polling for SRIOV when hw finished, otherwise
+		 * CPC engine may still keep fetching WB address which is already
+		 * invalid after sw finished and trigger DMAR reading error in
+		 * hypervisor side.
+		 */
+		WREG32_FIELD15_PREREG(GC, GET_INST(GC, xcc_id), CP_PQ_WPTR_POLL_CNTL, EN, 0);
+		return;
+	}
+
+	/* Use deinitialize sequence from CAIL when unbinding device
+	 * from driver, otherwise KIQ is hanging when binding back
+	 */
+	if (!amdgpu_in_reset(adev) && !adev->in_suspend) {
+		mutex_lock(&adev->srbm_mutex);
+		soc15_grbm_select(adev, adev->gfx.kiq[xcc_id].ring.me,
+				  adev->gfx.kiq[xcc_id].ring.pipe,
+				  adev->gfx.kiq[xcc_id].ring.queue, 0,
+				  GET_INST(GC, xcc_id));
+		gfx_v9_4_3_xcc_q_fini_register(&adev->gfx.kiq[xcc_id].ring,
+						 xcc_id);
+		soc15_grbm_select(adev, 0, 0, 0, 0, GET_INST(GC, xcc_id));
+		mutex_unlock(&adev->srbm_mutex);
+	}
+
+	gfx_v9_4_3_xcc_kcq_fini_register(adev, xcc_id);
+	gfx_v9_4_3_xcc_cp_compute_enable(adev, false, xcc_id);
+}
+
+static int gfx_v9_4_3_hw_init(struct amdgpu_ip_block *ip_block)
+{
+	int r;
+	struct amdgpu_device *adev = ip_block->adev;
+
+	amdgpu_gfx_cleaner_shader_init(adev, adev->gfx.cleaner_shader_size,
+				       adev->gfx.cleaner_shader_ptr);
+
+	if (!amdgpu_sriov_vf(adev))
+		gfx_v9_4_3_init_golden_registers(adev);
+
+	gfx_v9_4_3_constants_init(adev);
+
+	r = adev->gfx.rlc.funcs->resume(adev);
+	if (r)
+		return r;
+
+	r = gfx_v9_4_3_cp_resume(adev);
+	if (r)
+		return r;
+
+	return r;
+}
+
+static int gfx_v9_4_3_hw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, num_xcc;
+
+	amdgpu_irq_put(adev, &adev->gfx.priv_reg_irq, 0);
+	amdgpu_irq_put(adev, &adev->gfx.priv_inst_irq, 0);
+	amdgpu_irq_put(adev, &adev->gfx.bad_op_irq, 0);
+
+	num_xcc = NUM_XCC(adev->gfx.xcc_mask);
+	for (i = 0; i < num_xcc; i++) {
+		gfx_v9_4_3_xcc_fini(adev, i);
+	}
+
+	return 0;
+}
+
+static int gfx_v9_4_3_suspend(struct amdgpu_ip_block *ip_block)
+{
+	return gfx_v9_4_3_hw_fini(ip_block);
+}
+
+static int gfx_v9_4_3_resume(struct amdgpu_ip_block *ip_block)
+{
+	return gfx_v9_4_3_hw_init(ip_block);
+}
+
+static bool gfx_v9_4_3_is_idle(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, num_xcc;
+
+	num_xcc = NUM_XCC(adev->gfx.xcc_mask);
+	for (i = 0; i < num_xcc; i++) {
+		if (REG_GET_FIELD(RREG32_SOC15(GC, GET_INST(GC, i), regGRBM_STATUS),
+					GRBM_STATUS, GUI_ACTIVE))
+			return false;
+	}
+	return true;
+}
+
+static int gfx_v9_4_3_wait_for_idle(struct amdgpu_ip_block *ip_block)
+{
+	unsigned i;
+	struct amdgpu_device *adev = ip_block->adev;
+
+	for (i = 0; i < adev->usec_timeout; i++) {
+		if (gfx_v9_4_3_is_idle(ip_block))
+			return 0;
+		udelay(1);
+	}
+	return -ETIMEDOUT;
+}
+
+static int gfx_v9_4_3_soft_reset(struct amdgpu_ip_block *ip_block)
+{
+	u32 grbm_soft_reset = 0;
+	u32 tmp;
+	struct amdgpu_device *adev = ip_block->adev;
+
+	/* GRBM_STATUS */
+	tmp = RREG32_SOC15(GC, GET_INST(GC, 0), regGRBM_STATUS);
+	if (tmp & (GRBM_STATUS__PA_BUSY_MASK | GRBM_STATUS__SC_BUSY_MASK |
+		   GRBM_STATUS__BCI_BUSY_MASK | GRBM_STATUS__SX_BUSY_MASK |
+		   GRBM_STATUS__TA_BUSY_MASK | GRBM_STATUS__VGT_BUSY_MASK |
+		   GRBM_STATUS__DB_BUSY_MASK | GRBM_STATUS__CB_BUSY_MASK |
+		   GRBM_STATUS__GDS_BUSY_MASK | GRBM_STATUS__SPI_BUSY_MASK |
+		   GRBM_STATUS__IA_BUSY_MASK | GRBM_STATUS__IA_BUSY_NO_DMA_MASK)) {
+		grbm_soft_reset = REG_SET_FIELD(grbm_soft_reset,
+						GRBM_SOFT_RESET, SOFT_RESET_CP, 1);
+		grbm_soft_reset = REG_SET_FIELD(grbm_soft_reset,
+						GRBM_SOFT_RESET, SOFT_RESET_GFX, 1);
+	}
+
+	if (tmp & (GRBM_STATUS__CP_BUSY_MASK | GRBM_STATUS__CP_COHERENCY_BUSY_MASK)) {
+		grbm_soft_reset = REG_SET_FIELD(grbm_soft_reset,
+						GRBM_SOFT_RESET, SOFT_RESET_CP, 1);
+	}
+
+	/* GRBM_STATUS2 */
+	tmp = RREG32_SOC15(GC, GET_INST(GC, 0), regGRBM_STATUS2);
+	if (REG_GET_FIELD(tmp, GRBM_STATUS2, RLC_BUSY))
+		grbm_soft_reset = REG_SET_FIELD(grbm_soft_reset,
+						GRBM_SOFT_RESET, SOFT_RESET_RLC, 1);
+
+
+	if (grbm_soft_reset) {
+		/* stop the rlc */
+		adev->gfx.rlc.funcs->stop(adev);
+
+		/* Disable MEC parsing/prefetching */
+		gfx_v9_4_3_xcc_cp_compute_enable(adev, false, 0);
+
+		tmp = RREG32_SOC15(GC, GET_INST(GC, 0), regGRBM_SOFT_RESET);
+		tmp |= grbm_soft_reset;
+		dev_info(adev->dev, "GRBM_SOFT_RESET=0x%08X\n", tmp);
+		WREG32_SOC15(GC, GET_INST(GC, 0), regGRBM_SOFT_RESET, tmp);
+		tmp = RREG32_SOC15(GC, GET_INST(GC, 0), regGRBM_SOFT_RESET);
+
+		udelay(50);
+
+		tmp &= ~grbm_soft_reset;
+		WREG32_SOC15(GC, GET_INST(GC, 0), regGRBM_SOFT_RESET, tmp);
+		tmp = RREG32_SOC15(GC, GET_INST(GC, 0), regGRBM_SOFT_RESET);
+
+		/* Wait a little for things to settle down */
+		udelay(50);
+	}
+	return 0;
+}
+
+static void gfx_v9_4_3_ring_emit_gds_switch(struct amdgpu_ring *ring,
+					  uint32_t vmid,
+					  uint32_t gds_base, uint32_t gds_size,
+					  uint32_t gws_base, uint32_t gws_size,
+					  uint32_t oa_base, uint32_t oa_size)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	/* GDS Base */
+	gfx_v9_4_3_write_data_to_reg(ring, 0, false,
+				   SOC15_REG_OFFSET(GC, GET_INST(GC, 0), regGDS_VMID0_BASE) + 2 * vmid,
+				   gds_base);
+
+	/* GDS Size */
+	gfx_v9_4_3_write_data_to_reg(ring, 0, false,
+				   SOC15_REG_OFFSET(GC, GET_INST(GC, 0), regGDS_VMID0_SIZE) + 2 * vmid,
+				   gds_size);
+
+	/* GWS */
+	gfx_v9_4_3_write_data_to_reg(ring, 0, false,
+				   SOC15_REG_OFFSET(GC, GET_INST(GC, 0), regGDS_GWS_VMID0) + vmid,
+				   gws_size << GDS_GWS_VMID0__SIZE__SHIFT | gws_base);
+
+	/* OA */
+	gfx_v9_4_3_write_data_to_reg(ring, 0, false,
+				   SOC15_REG_OFFSET(GC, GET_INST(GC, 0), regGDS_OA_VMID0) + vmid,
+				   (1 << (oa_size + oa_base)) - (1 << oa_base));
+}
+
+static int gfx_v9_4_3_early_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	adev->gfx.num_compute_rings = min(amdgpu_gfx_get_num_kcq(adev),
+					  AMDGPU_MAX_COMPUTE_RINGS);
+	gfx_v9_4_3_set_kiq_pm4_funcs(adev);
+	gfx_v9_4_3_set_ring_funcs(adev);
+	gfx_v9_4_3_set_irq_funcs(adev);
+	gfx_v9_4_3_set_gds_init(adev);
+	gfx_v9_4_3_set_rlc_funcs(adev);
+
+	/* init rlcg reg access ctrl */
+	gfx_v9_4_3_init_rlcg_reg_access_ctrl(adev);
+
+	return gfx_v9_4_3_init_microcode(adev);
+}
+
+static int gfx_v9_4_3_late_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int r;
+
+	r = amdgpu_irq_get(adev, &adev->gfx.priv_reg_irq, 0);
+	if (r)
+		return r;
+
+	r = amdgpu_irq_get(adev, &adev->gfx.priv_inst_irq, 0);
+	if (r)
+		return r;
+
+	r = amdgpu_irq_get(adev, &adev->gfx.bad_op_irq, 0);
+	if (r)
+		return r;
+
+	if (adev->gfx.ras &&
+	    adev->gfx.ras->enable_watchdog_timer)
+		adev->gfx.ras->enable_watchdog_timer(adev);
+
+	return 0;
+}
+
+static void gfx_v9_4_3_xcc_update_sram_fgcg(struct amdgpu_device *adev,
+					    bool enable, int xcc_id)
+{
+	uint32_t def, data;
+
+	if (!(adev->cg_flags & AMD_CG_SUPPORT_GFX_FGCG))
+		return;
+
+	def = data = RREG32_SOC15(GC, GET_INST(GC, xcc_id),
+				  regRLC_CGTT_MGCG_OVERRIDE);
+
+	if (enable)
+		data &= ~RLC_CGTT_MGCG_OVERRIDE__GFXIP_FGCG_OVERRIDE_MASK;
+	else
+		data |= RLC_CGTT_MGCG_OVERRIDE__GFXIP_FGCG_OVERRIDE_MASK;
+
+	if (def != data)
+		WREG32_SOC15(GC, GET_INST(GC, xcc_id),
+			     regRLC_CGTT_MGCG_OVERRIDE, data);
+
+}
+
+static void gfx_v9_4_3_xcc_update_repeater_fgcg(struct amdgpu_device *adev,
+						bool enable, int xcc_id)
+{
+	uint32_t def, data;
+
+	if (!(adev->cg_flags & AMD_CG_SUPPORT_REPEATER_FGCG))
+		return;
+
+	def = data = RREG32_SOC15(GC, GET_INST(GC, xcc_id),
+				  regRLC_CGTT_MGCG_OVERRIDE);
+
+	if (enable)
+		data &= ~RLC_CGTT_MGCG_OVERRIDE__GFXIP_REP_FGCG_OVERRIDE_MASK;
+	else
+		data |= RLC_CGTT_MGCG_OVERRIDE__GFXIP_REP_FGCG_OVERRIDE_MASK;
+
+	if (def != data)
+		WREG32_SOC15(GC, GET_INST(GC, xcc_id),
+			     regRLC_CGTT_MGCG_OVERRIDE, data);
+}
+
+static void
+gfx_v9_4_3_xcc_update_medium_grain_clock_gating(struct amdgpu_device *adev,
+						bool enable, int xcc_id)
+{
+	uint32_t data, def;
+
+	/* It is disabled by HW by default */
+	if (enable && (adev->cg_flags & AMD_CG_SUPPORT_GFX_MGCG)) {
+		/* 1 - RLC_CGTT_MGCG_OVERRIDE */
+		def = data = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_CGTT_MGCG_OVERRIDE);
+
+		data &= ~(RLC_CGTT_MGCG_OVERRIDE__GRBM_CGTT_SCLK_OVERRIDE_MASK |
+			  RLC_CGTT_MGCG_OVERRIDE__GFXIP_MGCG_OVERRIDE_MASK |
+			  RLC_CGTT_MGCG_OVERRIDE__RLC_CGTT_SCLK_OVERRIDE_MASK |
+			  RLC_CGTT_MGCG_OVERRIDE__GFXIP_MGLS_OVERRIDE_MASK);
+
+		if (def != data)
+			WREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_CGTT_MGCG_OVERRIDE, data);
+
+		/* MGLS is a global flag to control all MGLS in GFX */
+		if (adev->cg_flags & AMD_CG_SUPPORT_GFX_MGLS) {
+			/* 2 - RLC memory Light sleep */
+			if (adev->cg_flags & AMD_CG_SUPPORT_GFX_RLC_LS) {
+				def = data = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_MEM_SLP_CNTL);
+				data |= RLC_MEM_SLP_CNTL__RLC_MEM_LS_EN_MASK;
+				if (def != data)
+					WREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_MEM_SLP_CNTL, data);
+			}
+			/* 3 - CP memory Light sleep */
+			if (adev->cg_flags & AMD_CG_SUPPORT_GFX_CP_LS) {
+				def = data = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_MEM_SLP_CNTL);
+				data |= CP_MEM_SLP_CNTL__CP_MEM_LS_EN_MASK;
+				if (def != data)
+					WREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_MEM_SLP_CNTL, data);
+			}
+		}
+	} else {
+		/* 1 - MGCG_OVERRIDE */
+		def = data = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_CGTT_MGCG_OVERRIDE);
+
+		data |= (RLC_CGTT_MGCG_OVERRIDE__RLC_CGTT_SCLK_OVERRIDE_MASK |
+			 RLC_CGTT_MGCG_OVERRIDE__GRBM_CGTT_SCLK_OVERRIDE_MASK |
+			 RLC_CGTT_MGCG_OVERRIDE__GFXIP_MGCG_OVERRIDE_MASK |
+			 RLC_CGTT_MGCG_OVERRIDE__GFXIP_MGLS_OVERRIDE_MASK);
+
+		if (def != data)
+			WREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_CGTT_MGCG_OVERRIDE, data);
+
+		/* 2 - disable MGLS in RLC */
+		data = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_MEM_SLP_CNTL);
+		if (data & RLC_MEM_SLP_CNTL__RLC_MEM_LS_EN_MASK) {
+			data &= ~RLC_MEM_SLP_CNTL__RLC_MEM_LS_EN_MASK;
+			WREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_MEM_SLP_CNTL, data);
+		}
+
+		/* 3 - disable MGLS in CP */
+		data = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_MEM_SLP_CNTL);
+		if (data & CP_MEM_SLP_CNTL__CP_MEM_LS_EN_MASK) {
+			data &= ~CP_MEM_SLP_CNTL__CP_MEM_LS_EN_MASK;
+			WREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_MEM_SLP_CNTL, data);
+		}
+	}
+
+}
+
+static void
+gfx_v9_4_3_xcc_update_coarse_grain_clock_gating(struct amdgpu_device *adev,
+						bool enable, int xcc_id)
+{
+	uint32_t def, data;
+
+	if (enable && (adev->cg_flags & AMD_CG_SUPPORT_GFX_CGCG)) {
+
+		def = data = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_CGTT_MGCG_OVERRIDE);
+		/* unset CGCG override */
+		data &= ~RLC_CGTT_MGCG_OVERRIDE__GFXIP_CGCG_OVERRIDE_MASK;
+		if (adev->cg_flags & AMD_CG_SUPPORT_GFX_CGLS)
+			data &= ~RLC_CGTT_MGCG_OVERRIDE__GFXIP_CGLS_OVERRIDE_MASK;
+		else
+			data |= RLC_CGTT_MGCG_OVERRIDE__GFXIP_CGLS_OVERRIDE_MASK;
+		/* update CGCG and CGLS override bits */
+		if (def != data)
+			WREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_CGTT_MGCG_OVERRIDE, data);
+
+		/* CGCG Hysteresis: 400us */
+		def = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_CGCG_CGLS_CTRL);
+
+		data = (0x2710
+			<< RLC_CGCG_CGLS_CTRL__CGCG_GFX_IDLE_THRESHOLD__SHIFT) |
+		       RLC_CGCG_CGLS_CTRL__CGCG_EN_MASK;
+		if (adev->cg_flags & AMD_CG_SUPPORT_GFX_CGLS)
+			data |= (0x000F << RLC_CGCG_CGLS_CTRL__CGLS_REP_COMPANSAT_DELAY__SHIFT) |
+				RLC_CGCG_CGLS_CTRL__CGLS_EN_MASK;
+		if (def != data)
+			WREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_CGCG_CGLS_CTRL, data);
+
+		/* set IDLE_POLL_COUNT(0x33450100)*/
+		def = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_RB_WPTR_POLL_CNTL);
+		data = (0x0100 << CP_RB_WPTR_POLL_CNTL__POLL_FREQUENCY__SHIFT) |
+			(0x3345 << CP_RB_WPTR_POLL_CNTL__IDLE_POLL_COUNT__SHIFT);
+		if (def != data)
+			WREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_RB_WPTR_POLL_CNTL, data);
+	} else {
+		def = data = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_CGCG_CGLS_CTRL);
+		/* reset CGCG/CGLS bits */
+		data &= ~(RLC_CGCG_CGLS_CTRL__CGCG_EN_MASK | RLC_CGCG_CGLS_CTRL__CGLS_EN_MASK);
+		/* disable cgcg and cgls in FSM */
+		if (def != data)
+			WREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_CGCG_CGLS_CTRL, data);
+	}
+
+}
+
+static int gfx_v9_4_3_xcc_update_gfx_clock_gating(struct amdgpu_device *adev,
+						  bool enable, int xcc_id)
+{
+	amdgpu_gfx_rlc_enter_safe_mode(adev, xcc_id);
+
+	if (enable) {
+		/* FGCG */
+		gfx_v9_4_3_xcc_update_sram_fgcg(adev, enable, xcc_id);
+		gfx_v9_4_3_xcc_update_repeater_fgcg(adev, enable, xcc_id);
+
+		/* CGCG/CGLS should be enabled after MGCG/MGLS
+		 * ===  MGCG + MGLS ===
+		 */
+		gfx_v9_4_3_xcc_update_medium_grain_clock_gating(adev, enable,
+								xcc_id);
+		/* ===  CGCG + CGLS === */
+		gfx_v9_4_3_xcc_update_coarse_grain_clock_gating(adev, enable,
+								xcc_id);
+	} else {
+		/* CGCG/CGLS should be disabled before MGCG/MGLS
+		 * ===  CGCG + CGLS ===
+		 */
+		gfx_v9_4_3_xcc_update_coarse_grain_clock_gating(adev, enable,
+								xcc_id);
+		/* ===  MGCG + MGLS === */
+		gfx_v9_4_3_xcc_update_medium_grain_clock_gating(adev, enable,
+								xcc_id);
+
+		/* FGCG */
+		gfx_v9_4_3_xcc_update_sram_fgcg(adev, enable, xcc_id);
+		gfx_v9_4_3_xcc_update_repeater_fgcg(adev, enable, xcc_id);
+	}
+
+	amdgpu_gfx_rlc_exit_safe_mode(adev, xcc_id);
+
+	return 0;
+}
+
+static const struct amdgpu_rlc_funcs gfx_v9_4_3_rlc_funcs = {
+	.is_rlc_enabled = gfx_v9_4_3_is_rlc_enabled,
+	.set_safe_mode = gfx_v9_4_3_xcc_set_safe_mode,
+	.unset_safe_mode = gfx_v9_4_3_xcc_unset_safe_mode,
+	.init = gfx_v9_4_3_rlc_init,
+	.resume = gfx_v9_4_3_rlc_resume,
+	.stop = gfx_v9_4_3_rlc_stop,
+	.reset = gfx_v9_4_3_rlc_reset,
+	.start = gfx_v9_4_3_rlc_start,
+	.update_spm_vmid = gfx_v9_4_3_update_spm_vmid,
+	.is_rlcg_access_range = gfx_v9_4_3_is_rlcg_access_range,
+};
+
+static int gfx_v9_4_3_set_powergating_state(struct amdgpu_ip_block *ip_block,
+					  enum amd_powergating_state state)
+{
+	return 0;
+}
+
+static int gfx_v9_4_3_set_clockgating_state(struct amdgpu_ip_block *ip_block,
+					  enum amd_clockgating_state state)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, num_xcc;
+
+	if (amdgpu_sriov_vf(adev))
+		return 0;
+
+	num_xcc = NUM_XCC(adev->gfx.xcc_mask);
+	for (i = 0; i < num_xcc; i++)
+		gfx_v9_4_3_xcc_update_gfx_clock_gating(
+			adev, state == AMD_CG_STATE_GATE, i);
+
+	return 0;
+}
+
+static void gfx_v9_4_3_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int data;
+
+	if (amdgpu_sriov_vf(adev))
+		*flags = 0;
+
+	/* AMD_CG_SUPPORT_GFX_MGCG */
+	data = RREG32_KIQ(SOC15_REG_OFFSET(GC, GET_INST(GC, 0), regRLC_CGTT_MGCG_OVERRIDE));
+	if (!(data & RLC_CGTT_MGCG_OVERRIDE__GFXIP_MGCG_OVERRIDE_MASK))
+		*flags |= AMD_CG_SUPPORT_GFX_MGCG;
+
+	/* AMD_CG_SUPPORT_GFX_CGCG */
+	data = RREG32_KIQ(SOC15_REG_OFFSET(GC, GET_INST(GC, 0), regRLC_CGCG_CGLS_CTRL));
+	if (data & RLC_CGCG_CGLS_CTRL__CGCG_EN_MASK)
+		*flags |= AMD_CG_SUPPORT_GFX_CGCG;
+
+	/* AMD_CG_SUPPORT_GFX_CGLS */
+	if (data & RLC_CGCG_CGLS_CTRL__CGLS_EN_MASK)
+		*flags |= AMD_CG_SUPPORT_GFX_CGLS;
+
+	/* AMD_CG_SUPPORT_GFX_RLC_LS */
+	data = RREG32_KIQ(SOC15_REG_OFFSET(GC, GET_INST(GC, 0), regRLC_MEM_SLP_CNTL));
+	if (data & RLC_MEM_SLP_CNTL__RLC_MEM_LS_EN_MASK)
+		*flags |= AMD_CG_SUPPORT_GFX_RLC_LS | AMD_CG_SUPPORT_GFX_MGLS;
+
+	/* AMD_CG_SUPPORT_GFX_CP_LS */
+	data = RREG32_KIQ(SOC15_REG_OFFSET(GC, GET_INST(GC, 0), regCP_MEM_SLP_CNTL));
+	if (data & CP_MEM_SLP_CNTL__CP_MEM_LS_EN_MASK)
+		*flags |= AMD_CG_SUPPORT_GFX_CP_LS | AMD_CG_SUPPORT_GFX_MGLS;
+}
+
+static void gfx_v9_4_3_ring_emit_hdp_flush(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	u32 ref_and_mask, reg_mem_engine;
+	const struct nbio_hdp_flush_reg *nbio_hf_reg = adev->nbio.hdp_flush_reg;
+
+	if (ring->funcs->type == AMDGPU_RING_TYPE_COMPUTE) {
+		switch (ring->me) {
+		case 1:
+			ref_and_mask = nbio_hf_reg->ref_and_mask_cp2 << ring->pipe;
+			break;
+		case 2:
+			ref_and_mask = nbio_hf_reg->ref_and_mask_cp6 << ring->pipe;
+			break;
+		default:
+			return;
+		}
+		reg_mem_engine = 0;
+	} else {
+		ref_and_mask = nbio_hf_reg->ref_and_mask_cp0;
+		reg_mem_engine = 1; /* pfp */
+	}
+
+	gfx_v9_4_3_wait_reg_mem(ring, reg_mem_engine, 0, 1,
+			      adev->nbio.funcs->get_hdp_flush_req_offset(adev),
+			      adev->nbio.funcs->get_hdp_flush_done_offset(adev),
+			      ref_and_mask, ref_and_mask, 0x20);
+}
+
+static void gfx_v9_4_3_ring_emit_ib_compute(struct amdgpu_ring *ring,
+					  struct amdgpu_job *job,
+					  struct amdgpu_ib *ib,
+					  uint32_t flags)
+{
+	unsigned vmid = AMDGPU_JOB_GET_VMID(job);
+	u32 control = INDIRECT_BUFFER_VALID | ib->length_dw | (vmid << 24);
+
+	/* Currently, there is a high possibility to get wave ID mismatch
+	 * between ME and GDS, leading to a hw deadlock, because ME generates
+	 * different wave IDs than the GDS expects. This situation happens
+	 * randomly when at least 5 compute pipes use GDS ordered append.
+	 * The wave IDs generated by ME are also wrong after suspend/resume.
+	 * Those are probably bugs somewhere else in the kernel driver.
+	 *
+	 * Writing GDS_COMPUTE_MAX_WAVE_ID resets wave ID counters in ME and
+	 * GDS to 0 for this ring (me/pipe).
+	 */
+	if (ib->flags & AMDGPU_IB_FLAG_RESET_GDS_MAX_WAVE_ID) {
+		amdgpu_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1));
+		amdgpu_ring_write(ring, regGDS_COMPUTE_MAX_WAVE_ID);
+		amdgpu_ring_write(ring, ring->adev->gds.gds_compute_max_wave_id);
+	}
+
+	amdgpu_ring_write(ring, PACKET3(PACKET3_INDIRECT_BUFFER, 2));
+	BUG_ON(ib->gpu_addr & 0x3); /* Dword align */
+	amdgpu_ring_write(ring,
+#ifdef __BIG_ENDIAN
+				(2 << 0) |
+#endif
+				lower_32_bits(ib->gpu_addr));
+	amdgpu_ring_write(ring, upper_32_bits(ib->gpu_addr));
+	amdgpu_ring_write(ring, control);
+}
+
+static void gfx_v9_4_3_ring_emit_fence(struct amdgpu_ring *ring, u64 addr,
+				     u64 seq, unsigned flags)
+{
+	bool write64bit = flags & AMDGPU_FENCE_FLAG_64BIT;
+	bool int_sel = flags & AMDGPU_FENCE_FLAG_INT;
+	bool writeback = flags & AMDGPU_FENCE_FLAG_TC_WB_ONLY;
+
+	/* RELEASE_MEM - flush caches, send int */
+	amdgpu_ring_write(ring, PACKET3(PACKET3_RELEASE_MEM, 6));
+	amdgpu_ring_write(ring, ((writeback ? (EOP_TC_WB_ACTION_EN |
+					       EOP_TC_NC_ACTION_EN) :
+					      (EOP_TCL1_ACTION_EN |
+					       EOP_TC_ACTION_EN |
+					       EOP_TC_WB_ACTION_EN |
+					       EOP_TC_MD_ACTION_EN)) |
+				 EVENT_TYPE(CACHE_FLUSH_AND_INV_TS_EVENT) |
+				 EVENT_INDEX(5)));
+	amdgpu_ring_write(ring, DATA_SEL(write64bit ? 2 : 1) | INT_SEL(int_sel ? 2 : 0));
+
+	/*
+	 * the address should be Qword aligned if 64bit write, Dword
+	 * aligned if only send 32bit data low (discard data high)
+	 */
+	if (write64bit)
+		BUG_ON(addr & 0x7);
+	else
+		BUG_ON(addr & 0x3);
+	amdgpu_ring_write(ring, lower_32_bits(addr));
+	amdgpu_ring_write(ring, upper_32_bits(addr));
+	amdgpu_ring_write(ring, lower_32_bits(seq));
+	amdgpu_ring_write(ring, upper_32_bits(seq));
+	amdgpu_ring_write(ring, 0);
+}
+
+static void gfx_v9_4_3_ring_emit_pipeline_sync(struct amdgpu_ring *ring)
+{
+	int usepfp = (ring->funcs->type == AMDGPU_RING_TYPE_GFX);
+	uint32_t seq = ring->fence_drv.sync_seq;
+	uint64_t addr = ring->fence_drv.gpu_addr;
+
+	gfx_v9_4_3_wait_reg_mem(ring, usepfp, 1, 0,
+			      lower_32_bits(addr), upper_32_bits(addr),
+			      seq, 0xffffffff, 4);
+}
+
+static void gfx_v9_4_3_ring_emit_vm_flush(struct amdgpu_ring *ring,
+					unsigned vmid, uint64_t pd_addr)
+{
+	amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pd_addr);
+}
+
+static u64 gfx_v9_4_3_ring_get_rptr_compute(struct amdgpu_ring *ring)
+{
+	return ring->adev->wb.wb[ring->rptr_offs]; /* gfx9 hardware is 32bit rptr */
+}
+
+static u64 gfx_v9_4_3_ring_get_wptr_compute(struct amdgpu_ring *ring)
+{
+	u64 wptr;
+
+	/* XXX check if swapping is necessary on BE */
+	if (ring->use_doorbell)
+		wptr = atomic64_read((atomic64_t *)&ring->adev->wb.wb[ring->wptr_offs]);
+	else
+		BUG();
+	return wptr;
+}
+
+static void gfx_v9_4_3_ring_set_wptr_compute(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	/* XXX check if swapping is necessary on BE */
+	if (ring->use_doorbell) {
+		atomic64_set((atomic64_t *)&adev->wb.wb[ring->wptr_offs], ring->wptr);
+		WDOORBELL64(ring->doorbell_index, ring->wptr);
+	} else {
+		BUG(); /* only DOORBELL method supported on gfx9 now */
+	}
+}
+
+static void gfx_v9_4_3_ring_emit_fence_kiq(struct amdgpu_ring *ring, u64 addr,
+					 u64 seq, unsigned int flags)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	/* we only allocate 32bit for each seq wb address */
+	BUG_ON(flags & AMDGPU_FENCE_FLAG_64BIT);
+
+	/* write fence seq to the "addr" */
+	amdgpu_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
+	amdgpu_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) |
+				 WRITE_DATA_DST_SEL(5) | WR_CONFIRM));
+	amdgpu_ring_write(ring, lower_32_bits(addr));
+	amdgpu_ring_write(ring, upper_32_bits(addr));
+	amdgpu_ring_write(ring, lower_32_bits(seq));
+
+	if (flags & AMDGPU_FENCE_FLAG_INT) {
+		/* set register to trigger INT */
+		amdgpu_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
+		amdgpu_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) |
+					 WRITE_DATA_DST_SEL(0) | WR_CONFIRM));
+		amdgpu_ring_write(ring, SOC15_REG_OFFSET(GC, GET_INST(GC, 0), regCPC_INT_STATUS));
+		amdgpu_ring_write(ring, 0);
+		amdgpu_ring_write(ring, 0x20000000); /* src_id is 178 */
+	}
+}
+
+static void gfx_v9_4_3_ring_emit_rreg(struct amdgpu_ring *ring, uint32_t reg,
+				    uint32_t reg_val_offs)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	reg = gfx_v9_4_3_normalize_xcc_reg_offset(reg);
+
+	amdgpu_ring_write(ring, PACKET3(PACKET3_COPY_DATA, 4));
+	amdgpu_ring_write(ring, 0 |	/* src: register*/
+				(5 << 8) |	/* dst: memory */
+				(1 << 20));	/* write confirm */
+	amdgpu_ring_write(ring, reg);
+	amdgpu_ring_write(ring, 0);
+	amdgpu_ring_write(ring, lower_32_bits(adev->wb.gpu_addr +
+				reg_val_offs * 4));
+	amdgpu_ring_write(ring, upper_32_bits(adev->wb.gpu_addr +
+				reg_val_offs * 4));
+}
+
+static void gfx_v9_4_3_ring_emit_wreg(struct amdgpu_ring *ring, uint32_t reg,
+				    uint32_t val)
+{
+	uint32_t cmd = 0;
+
+	reg = gfx_v9_4_3_normalize_xcc_reg_offset(reg);
+
+	switch (ring->funcs->type) {
+	case AMDGPU_RING_TYPE_GFX:
+		cmd = WRITE_DATA_ENGINE_SEL(1) | WR_CONFIRM;
+		break;
+	case AMDGPU_RING_TYPE_KIQ:
+		cmd = (1 << 16); /* no inc addr */
+		break;
+	default:
+		cmd = WR_CONFIRM;
+		break;
+	}
+	amdgpu_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
+	amdgpu_ring_write(ring, cmd);
+	amdgpu_ring_write(ring, reg);
+	amdgpu_ring_write(ring, 0);
+	amdgpu_ring_write(ring, val);
+}
+
+static void gfx_v9_4_3_ring_emit_reg_wait(struct amdgpu_ring *ring, uint32_t reg,
+					uint32_t val, uint32_t mask)
+{
+	gfx_v9_4_3_wait_reg_mem(ring, 0, 0, 0, reg, 0, val, mask, 0x20);
+}
+
+static void gfx_v9_4_3_ring_emit_reg_write_reg_wait(struct amdgpu_ring *ring,
+						  uint32_t reg0, uint32_t reg1,
+						  uint32_t ref, uint32_t mask)
+{
+	amdgpu_ring_emit_reg_write_reg_wait_helper(ring, reg0, reg1,
+						   ref, mask);
+}
+
+static void gfx_v9_4_3_ring_soft_recovery(struct amdgpu_ring *ring,
+					  unsigned vmid)
+{
+	struct amdgpu_device *adev = ring->adev;
+	uint32_t value = 0;
+
+	value = REG_SET_FIELD(value, SQ_CMD, CMD, 0x03);
+	value = REG_SET_FIELD(value, SQ_CMD, MODE, 0x01);
+	value = REG_SET_FIELD(value, SQ_CMD, CHECK_VMID, 1);
+	value = REG_SET_FIELD(value, SQ_CMD, VM_ID, vmid);
+	amdgpu_gfx_rlc_enter_safe_mode(adev, ring->xcc_id);
+	WREG32_SOC15(GC, GET_INST(GC, ring->xcc_id), regSQ_CMD, value);
+	amdgpu_gfx_rlc_exit_safe_mode(adev, ring->xcc_id);
+}
+
+static void gfx_v9_4_3_xcc_set_compute_eop_interrupt_state(
+	struct amdgpu_device *adev, int me, int pipe,
+	enum amdgpu_interrupt_state state, int xcc_id)
+{
+	u32 mec_int_cntl, mec_int_cntl_reg;
+
+	/*
+	 * amdgpu controls only the first MEC. That's why this function only
+	 * handles the setting of interrupts for this specific MEC. All other
+	 * pipes' interrupts are set by amdkfd.
+	 */
+
+	if (me == 1) {
+		switch (pipe) {
+		case 0:
+			mec_int_cntl_reg = SOC15_REG_OFFSET(GC, GET_INST(GC, xcc_id), regCP_ME1_PIPE0_INT_CNTL);
+			break;
+		case 1:
+			mec_int_cntl_reg = SOC15_REG_OFFSET(GC, GET_INST(GC, xcc_id), regCP_ME1_PIPE1_INT_CNTL);
+			break;
+		case 2:
+			mec_int_cntl_reg = SOC15_REG_OFFSET(GC, GET_INST(GC, xcc_id), regCP_ME1_PIPE2_INT_CNTL);
+			break;
+		case 3:
+			mec_int_cntl_reg = SOC15_REG_OFFSET(GC, GET_INST(GC, xcc_id), regCP_ME1_PIPE3_INT_CNTL);
+			break;
+		default:
+			DRM_DEBUG("invalid pipe %d\n", pipe);
+			return;
+		}
+	} else {
+		DRM_DEBUG("invalid me %d\n", me);
+		return;
+	}
+
+	switch (state) {
+	case AMDGPU_IRQ_STATE_DISABLE:
+		mec_int_cntl = RREG32_XCC(mec_int_cntl_reg, xcc_id);
+		mec_int_cntl = REG_SET_FIELD(mec_int_cntl, CP_ME1_PIPE0_INT_CNTL,
+					     TIME_STAMP_INT_ENABLE, 0);
+		WREG32_XCC(mec_int_cntl_reg, mec_int_cntl, xcc_id);
+		break;
+	case AMDGPU_IRQ_STATE_ENABLE:
+		mec_int_cntl = RREG32_XCC(mec_int_cntl_reg, xcc_id);
+		mec_int_cntl = REG_SET_FIELD(mec_int_cntl, CP_ME1_PIPE0_INT_CNTL,
+					     TIME_STAMP_INT_ENABLE, 1);
+		WREG32_XCC(mec_int_cntl_reg, mec_int_cntl, xcc_id);
+		break;
+	default:
+		break;
+	}
+}
+
+static u32 gfx_v9_4_3_get_cpc_int_cntl(struct amdgpu_device *adev,
+				     int xcc_id, int me, int pipe)
+{
+	/*
+	 * amdgpu controls only the first MEC. That's why this function only
+	 * handles the setting of interrupts for this specific MEC. All other
+	 * pipes' interrupts are set by amdkfd.
+	 */
+	if (me != 1)
+		return 0;
+
+	switch (pipe) {
+	case 0:
+		return SOC15_REG_OFFSET(GC, GET_INST(GC, xcc_id), regCP_ME1_PIPE0_INT_CNTL);
+	case 1:
+		return SOC15_REG_OFFSET(GC, GET_INST(GC, xcc_id), regCP_ME1_PIPE1_INT_CNTL);
+	case 2:
+		return SOC15_REG_OFFSET(GC, GET_INST(GC, xcc_id), regCP_ME1_PIPE2_INT_CNTL);
+	case 3:
+		return SOC15_REG_OFFSET(GC, GET_INST(GC, xcc_id), regCP_ME1_PIPE3_INT_CNTL);
+	default:
+		return 0;
+	}
+}
+
+static int gfx_v9_4_3_set_priv_reg_fault_state(struct amdgpu_device *adev,
+					     struct amdgpu_irq_src *source,
+					     unsigned type,
+					     enum amdgpu_interrupt_state state)
+{
+	u32 mec_int_cntl_reg, mec_int_cntl;
+	int i, j, k, num_xcc;
+
+	num_xcc = NUM_XCC(adev->gfx.xcc_mask);
+	switch (state) {
+	case AMDGPU_IRQ_STATE_DISABLE:
+	case AMDGPU_IRQ_STATE_ENABLE:
+		for (i = 0; i < num_xcc; i++) {
+			WREG32_FIELD15_PREREG(GC, GET_INST(GC, i), CP_INT_CNTL_RING0,
+					      PRIV_REG_INT_ENABLE,
+					      state == AMDGPU_IRQ_STATE_ENABLE ? 1 : 0);
+			for (j = 0; j < adev->gfx.mec.num_mec; j++) {
+				for (k = 0; k < adev->gfx.mec.num_pipe_per_mec; k++) {
+					/* MECs start at 1 */
+					mec_int_cntl_reg = gfx_v9_4_3_get_cpc_int_cntl(adev, i, j + 1, k);
+
+					if (mec_int_cntl_reg) {
+						mec_int_cntl = RREG32_XCC(mec_int_cntl_reg, i);
+						mec_int_cntl = REG_SET_FIELD(mec_int_cntl, CP_ME1_PIPE0_INT_CNTL,
+									     PRIV_REG_INT_ENABLE,
+									     state == AMDGPU_IRQ_STATE_ENABLE ?
+									     1 : 0);
+						WREG32_XCC(mec_int_cntl_reg, mec_int_cntl, i);
+					}
+				}
+			}
+		}
+		break;
+	default:
+		break;
+	}
+
+	return 0;
+}
+
+static int gfx_v9_4_3_set_bad_op_fault_state(struct amdgpu_device *adev,
+					     struct amdgpu_irq_src *source,
+					     unsigned type,
+					     enum amdgpu_interrupt_state state)
+{
+	u32 mec_int_cntl_reg, mec_int_cntl;
+	int i, j, k, num_xcc;
+
+	num_xcc = NUM_XCC(adev->gfx.xcc_mask);
+	switch (state) {
+	case AMDGPU_IRQ_STATE_DISABLE:
+	case AMDGPU_IRQ_STATE_ENABLE:
+		for (i = 0; i < num_xcc; i++) {
+			WREG32_FIELD15_PREREG(GC, GET_INST(GC, i), CP_INT_CNTL_RING0,
+					      OPCODE_ERROR_INT_ENABLE,
+					      state == AMDGPU_IRQ_STATE_ENABLE ? 1 : 0);
+			for (j = 0; j < adev->gfx.mec.num_mec; j++) {
+				for (k = 0; k < adev->gfx.mec.num_pipe_per_mec; k++) {
+					/* MECs start at 1 */
+					mec_int_cntl_reg = gfx_v9_4_3_get_cpc_int_cntl(adev, i, j + 1, k);
+
+					if (mec_int_cntl_reg) {
+						mec_int_cntl = RREG32_XCC(mec_int_cntl_reg, i);
+						mec_int_cntl = REG_SET_FIELD(mec_int_cntl, CP_ME1_PIPE0_INT_CNTL,
+									     OPCODE_ERROR_INT_ENABLE,
+									     state == AMDGPU_IRQ_STATE_ENABLE ?
+									     1 : 0);
+						WREG32_XCC(mec_int_cntl_reg, mec_int_cntl, i);
+					}
+				}
+			}
+		}
+		break;
+	default:
+		break;
+	}
+
+	return 0;
+}
+
+static int gfx_v9_4_3_set_priv_inst_fault_state(struct amdgpu_device *adev,
+					      struct amdgpu_irq_src *source,
+					      unsigned type,
+					      enum amdgpu_interrupt_state state)
+{
+	int i, num_xcc;
+
+	num_xcc = NUM_XCC(adev->gfx.xcc_mask);
+	switch (state) {
+	case AMDGPU_IRQ_STATE_DISABLE:
+	case AMDGPU_IRQ_STATE_ENABLE:
+		for (i = 0; i < num_xcc; i++)
+			WREG32_FIELD15_PREREG(GC, GET_INST(GC, i), CP_INT_CNTL_RING0,
+				PRIV_INSTR_INT_ENABLE,
+				state == AMDGPU_IRQ_STATE_ENABLE ? 1 : 0);
+		break;
+	default:
+		break;
+	}
+
+	return 0;
+}
+
+static int gfx_v9_4_3_set_eop_interrupt_state(struct amdgpu_device *adev,
+					    struct amdgpu_irq_src *src,
+					    unsigned type,
+					    enum amdgpu_interrupt_state state)
+{
+	int i, num_xcc;
+
+	num_xcc = NUM_XCC(adev->gfx.xcc_mask);
+	for (i = 0; i < num_xcc; i++) {
+		switch (type) {
+		case AMDGPU_CP_IRQ_COMPUTE_MEC1_PIPE0_EOP:
+			gfx_v9_4_3_xcc_set_compute_eop_interrupt_state(
+				adev, 1, 0, state, i);
+			break;
+		case AMDGPU_CP_IRQ_COMPUTE_MEC1_PIPE1_EOP:
+			gfx_v9_4_3_xcc_set_compute_eop_interrupt_state(
+				adev, 1, 1, state, i);
+			break;
+		case AMDGPU_CP_IRQ_COMPUTE_MEC1_PIPE2_EOP:
+			gfx_v9_4_3_xcc_set_compute_eop_interrupt_state(
+				adev, 1, 2, state, i);
+			break;
+		case AMDGPU_CP_IRQ_COMPUTE_MEC1_PIPE3_EOP:
+			gfx_v9_4_3_xcc_set_compute_eop_interrupt_state(
+				adev, 1, 3, state, i);
+			break;
+		case AMDGPU_CP_IRQ_COMPUTE_MEC2_PIPE0_EOP:
+			gfx_v9_4_3_xcc_set_compute_eop_interrupt_state(
+				adev, 2, 0, state, i);
+			break;
+		case AMDGPU_CP_IRQ_COMPUTE_MEC2_PIPE1_EOP:
+			gfx_v9_4_3_xcc_set_compute_eop_interrupt_state(
+				adev, 2, 1, state, i);
+			break;
+		case AMDGPU_CP_IRQ_COMPUTE_MEC2_PIPE2_EOP:
+			gfx_v9_4_3_xcc_set_compute_eop_interrupt_state(
+				adev, 2, 2, state, i);
+			break;
+		case AMDGPU_CP_IRQ_COMPUTE_MEC2_PIPE3_EOP:
+			gfx_v9_4_3_xcc_set_compute_eop_interrupt_state(
+				adev, 2, 3, state, i);
+			break;
+		default:
+			break;
+		}
+	}
+
+	return 0;
+}
+
+static int gfx_v9_4_3_eop_irq(struct amdgpu_device *adev,
+			    struct amdgpu_irq_src *source,
+			    struct amdgpu_iv_entry *entry)
+{
+	int i, xcc_id;
+	u8 me_id, pipe_id, queue_id;
+	struct amdgpu_ring *ring;
+
+	DRM_DEBUG("IH: CP EOP\n");
+	me_id = (entry->ring_id & 0x0c) >> 2;
+	pipe_id = (entry->ring_id & 0x03) >> 0;
+	queue_id = (entry->ring_id & 0x70) >> 4;
+
+	xcc_id = gfx_v9_4_3_ih_to_xcc_inst(adev, entry->node_id);
+
+	if (xcc_id == -EINVAL)
+		return -EINVAL;
+
+	switch (me_id) {
+	case 0:
+	case 1:
+	case 2:
+		for (i = 0; i < adev->gfx.num_compute_rings; i++) {
+			ring = &adev->gfx.compute_ring
+					[i +
+					 xcc_id * adev->gfx.num_compute_rings];
+			/* Per-queue interrupt is supported for MEC starting from VI.
+			  * The interrupt can only be enabled/disabled per pipe instead of per queue.
+			  */
+
+			if ((ring->me == me_id) && (ring->pipe == pipe_id) && (ring->queue == queue_id))
+				amdgpu_fence_process(ring);
+		}
+		break;
+	}
+	return 0;
+}
+
+static void gfx_v9_4_3_fault(struct amdgpu_device *adev,
+			   struct amdgpu_iv_entry *entry)
+{
+	u8 me_id, pipe_id, queue_id;
+	struct amdgpu_ring *ring;
+	int i, xcc_id;
+
+	me_id = (entry->ring_id & 0x0c) >> 2;
+	pipe_id = (entry->ring_id & 0x03) >> 0;
+	queue_id = (entry->ring_id & 0x70) >> 4;
+
+	xcc_id = gfx_v9_4_3_ih_to_xcc_inst(adev, entry->node_id);
+
+	if (xcc_id == -EINVAL)
+		return;
+
+	switch (me_id) {
+	case 0:
+	case 1:
+	case 2:
+		for (i = 0; i < adev->gfx.num_compute_rings; i++) {
+			ring = &adev->gfx.compute_ring
+					[i +
+					 xcc_id * adev->gfx.num_compute_rings];
+			if (ring->me == me_id && ring->pipe == pipe_id &&
+			    ring->queue == queue_id)
+				drm_sched_fault(&ring->sched);
+		}
+		break;
+	}
+}
+
+static int gfx_v9_4_3_priv_reg_irq(struct amdgpu_device *adev,
+				 struct amdgpu_irq_src *source,
+				 struct amdgpu_iv_entry *entry)
+{
+	DRM_ERROR("Illegal register access in command stream\n");
+	gfx_v9_4_3_fault(adev, entry);
+	return 0;
+}
+
+static int gfx_v9_4_3_bad_op_irq(struct amdgpu_device *adev,
+				 struct amdgpu_irq_src *source,
+				 struct amdgpu_iv_entry *entry)
+{
+	DRM_ERROR("Illegal opcode in command stream\n");
+	gfx_v9_4_3_fault(adev, entry);
+	return 0;
+}
+
+static int gfx_v9_4_3_priv_inst_irq(struct amdgpu_device *adev,
+				  struct amdgpu_irq_src *source,
+				  struct amdgpu_iv_entry *entry)
+{
+	DRM_ERROR("Illegal instruction in command stream\n");
+	gfx_v9_4_3_fault(adev, entry);
+	return 0;
+}
+
+static void gfx_v9_4_3_emit_mem_sync(struct amdgpu_ring *ring)
+{
+	const unsigned int cp_coher_cntl =
+			PACKET3_ACQUIRE_MEM_CP_COHER_CNTL_SH_ICACHE_ACTION_ENA(1) |
+			PACKET3_ACQUIRE_MEM_CP_COHER_CNTL_SH_KCACHE_ACTION_ENA(1) |
+			PACKET3_ACQUIRE_MEM_CP_COHER_CNTL_TC_ACTION_ENA(1) |
+			PACKET3_ACQUIRE_MEM_CP_COHER_CNTL_TCL1_ACTION_ENA(1) |
+			PACKET3_ACQUIRE_MEM_CP_COHER_CNTL_TC_WB_ACTION_ENA(1);
+
+	/* ACQUIRE_MEM -make one or more surfaces valid for use by the subsequent operations */
+	amdgpu_ring_write(ring, PACKET3(PACKET3_ACQUIRE_MEM, 5));
+	amdgpu_ring_write(ring, cp_coher_cntl); /* CP_COHER_CNTL */
+	amdgpu_ring_write(ring, 0xffffffff);  /* CP_COHER_SIZE */
+	amdgpu_ring_write(ring, 0xffffff);  /* CP_COHER_SIZE_HI */
+	amdgpu_ring_write(ring, 0); /* CP_COHER_BASE */
+	amdgpu_ring_write(ring, 0);  /* CP_COHER_BASE_HI */
+	amdgpu_ring_write(ring, 0x0000000A); /* POLL_INTERVAL */
+}
+
+static void gfx_v9_4_3_emit_wave_limit_cs(struct amdgpu_ring *ring,
+					uint32_t pipe, bool enable)
+{
+	struct amdgpu_device *adev = ring->adev;
+	uint32_t val;
+	uint32_t wcl_cs_reg;
+
+	/* regSPI_WCL_PIPE_PERCENT_CS[0-7]_DEFAULT values are same */
+	val = enable ? 0x1 : 0x7f;
+
+	switch (pipe) {
+	case 0:
+		wcl_cs_reg = SOC15_REG_OFFSET(GC, GET_INST(GC, 0), regSPI_WCL_PIPE_PERCENT_CS0);
+		break;
+	case 1:
+		wcl_cs_reg = SOC15_REG_OFFSET(GC, GET_INST(GC, 0), regSPI_WCL_PIPE_PERCENT_CS1);
+		break;
+	case 2:
+		wcl_cs_reg = SOC15_REG_OFFSET(GC, GET_INST(GC, 0), regSPI_WCL_PIPE_PERCENT_CS2);
+		break;
+	case 3:
+		wcl_cs_reg = SOC15_REG_OFFSET(GC, GET_INST(GC, 0), regSPI_WCL_PIPE_PERCENT_CS3);
+		break;
+	default:
+		DRM_DEBUG("invalid pipe %d\n", pipe);
+		return;
+	}
+
+	amdgpu_ring_emit_wreg(ring, wcl_cs_reg, val);
+
+}
+static void gfx_v9_4_3_emit_wave_limit(struct amdgpu_ring *ring, bool enable)
+{
+	struct amdgpu_device *adev = ring->adev;
+	uint32_t val;
+	int i;
+
+	/* regSPI_WCL_PIPE_PERCENT_GFX is 7 bit multiplier register to limit
+	 * number of gfx waves. Setting 5 bit will make sure gfx only gets
+	 * around 25% of gpu resources.
+	 */
+	val = enable ? 0x1f : 0x07ffffff;
+	amdgpu_ring_emit_wreg(ring,
+			      SOC15_REG_OFFSET(GC, GET_INST(GC, 0), regSPI_WCL_PIPE_PERCENT_GFX),
+			      val);
+
+	/* Restrict waves for normal/low priority compute queues as well
+	 * to get best QoS for high priority compute jobs.
+	 *
+	 * amdgpu controls only 1st ME(0-3 CS pipes).
+	 */
+	for (i = 0; i < adev->gfx.mec.num_pipe_per_mec; i++) {
+		if (i != ring->pipe)
+			gfx_v9_4_3_emit_wave_limit_cs(ring, i, enable);
+
+	}
+}
+
+static int gfx_v9_4_3_unmap_done(struct amdgpu_device *adev, uint32_t me,
+				uint32_t pipe, uint32_t queue,
+				uint32_t xcc_id)
+{
+	int i, r;
+	/* make sure dequeue is complete*/
+	gfx_v9_4_3_xcc_set_safe_mode(adev, xcc_id);
+	mutex_lock(&adev->srbm_mutex);
+	soc15_grbm_select(adev, me, pipe, queue, 0, GET_INST(GC, xcc_id));
+	for (i = 0; i < adev->usec_timeout; i++) {
+		if (!(RREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_HQD_ACTIVE) & 1))
+			break;
+		udelay(1);
+	}
+	if (i >= adev->usec_timeout)
+		r = -ETIMEDOUT;
+	else
+		r = 0;
+	soc15_grbm_select(adev, 0, 0, 0, 0, GET_INST(GC, xcc_id));
+	mutex_unlock(&adev->srbm_mutex);
+	gfx_v9_4_3_xcc_unset_safe_mode(adev, xcc_id);
+
+	return r;
+
+}
+
+static bool gfx_v9_4_3_pipe_reset_support(struct amdgpu_device *adev)
+{
+	if (!!(adev->gfx.compute_supported_reset & AMDGPU_RESET_TYPE_PER_PIPE))
+		return true;
+	else
+		dev_warn_once(adev->dev, "Please use the latest MEC version to see whether support pipe reset\n");
+
+	return false;
+}
+
+static int gfx_v9_4_3_reset_hw_pipe(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	uint32_t reset_pipe, clean_pipe;
+	int r;
+
+	if (!gfx_v9_4_3_pipe_reset_support(adev))
+		return -EINVAL;
+
+	gfx_v9_4_3_xcc_set_safe_mode(adev, ring->xcc_id);
+	mutex_lock(&adev->srbm_mutex);
+
+	reset_pipe = RREG32_SOC15(GC, GET_INST(GC, ring->xcc_id), regCP_MEC_CNTL);
+	clean_pipe = reset_pipe;
+
+	if (ring->me == 1) {
+		switch (ring->pipe) {
+		case 0:
+			reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_CNTL,
+						   MEC_ME1_PIPE0_RESET, 1);
+			break;
+		case 1:
+			reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_CNTL,
+						   MEC_ME1_PIPE1_RESET, 1);
+			break;
+		case 2:
+			reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_CNTL,
+						   MEC_ME1_PIPE2_RESET, 1);
+			break;
+		case 3:
+			reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_CNTL,
+						   MEC_ME1_PIPE3_RESET, 1);
+			break;
+		default:
+			break;
+		}
+	} else {
+		if (ring->pipe)
+			reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_CNTL,
+						   MEC_ME2_PIPE1_RESET, 1);
+		else
+			reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_CNTL,
+						   MEC_ME2_PIPE0_RESET, 1);
+	}
+
+	WREG32_SOC15(GC, GET_INST(GC, ring->xcc_id), regCP_MEC_CNTL, reset_pipe);
+	WREG32_SOC15(GC, GET_INST(GC, ring->xcc_id), regCP_MEC_CNTL, clean_pipe);
+	mutex_unlock(&adev->srbm_mutex);
+	gfx_v9_4_3_xcc_unset_safe_mode(adev, ring->xcc_id);
+
+	r = gfx_v9_4_3_unmap_done(adev, ring->me, ring->pipe, ring->queue, ring->xcc_id);
+	return r;
+}
+
+static int gfx_v9_4_3_reset_kcq(struct amdgpu_ring *ring,
+				unsigned int vmid,
+				struct amdgpu_fence *timedout_fence)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct amdgpu_kiq *kiq = &adev->gfx.kiq[ring->xcc_id];
+	struct amdgpu_ring *kiq_ring = &kiq->ring;
+	int reset_mode = AMDGPU_RESET_TYPE_PER_QUEUE;
+	unsigned long flags;
+	int r;
+
+	if (!kiq->pmf || !kiq->pmf->kiq_unmap_queues)
+		return -EINVAL;
+
+	amdgpu_ring_reset_helper_begin(ring, timedout_fence);
+
+	spin_lock_irqsave(&kiq->ring_lock, flags);
+
+	if (amdgpu_ring_alloc(kiq_ring, kiq->pmf->unmap_queues_size)) {
+		spin_unlock_irqrestore(&kiq->ring_lock, flags);
+		return -ENOMEM;
+	}
+
+	kiq->pmf->kiq_unmap_queues(kiq_ring, ring, RESET_QUEUES,
+				   0, 0);
+	amdgpu_ring_commit(kiq_ring);
+
+	spin_unlock_irqrestore(&kiq->ring_lock, flags);
+
+	r = amdgpu_ring_test_ring(kiq_ring);
+	if (r) {
+		dev_err(adev->dev, "kiq ring test failed after ring: %s queue reset\n",
+				ring->name);
+		goto pipe_reset;
+	}
+
+	r = gfx_v9_4_3_unmap_done(adev, ring->me, ring->pipe, ring->queue, ring->xcc_id);
+	if (r)
+		dev_err(adev->dev, "fail to wait on hqd deactive and will try pipe reset\n");
+
+pipe_reset:
+	if (r) {
+		if (!(adev->gfx.compute_supported_reset & AMDGPU_RESET_TYPE_PER_PIPE))
+			return -EOPNOTSUPP;
+		r = gfx_v9_4_3_reset_hw_pipe(ring);
+		reset_mode = AMDGPU_RESET_TYPE_PER_PIPE;
+		dev_info(adev->dev, "ring: %s pipe reset :%s\n", ring->name,
+				r ? "failed" : "successfully");
+		if (r)
+			return r;
+	}
+
+	gfx_v9_4_3_xcc_kcq_init_queue(ring, ring->xcc_id, true);
+
+	spin_lock_irqsave(&kiq->ring_lock, flags);
+	r = amdgpu_ring_alloc(kiq_ring, kiq->pmf->map_queues_size);
+	if (r) {
+		spin_unlock_irqrestore(&kiq->ring_lock, flags);
+		return -ENOMEM;
+	}
+	kiq->pmf->kiq_map_queues(kiq_ring, ring);
+	amdgpu_ring_commit(kiq_ring);
+	r = amdgpu_ring_test_ring(kiq_ring);
+	spin_unlock_irqrestore(&kiq->ring_lock, flags);
+	if (r) {
+		if (reset_mode == AMDGPU_RESET_TYPE_PER_QUEUE)
+			goto pipe_reset;
+
+		dev_err(adev->dev, "fail to remap queue\n");
+		return r;
+	}
+
+	if (reset_mode == AMDGPU_RESET_TYPE_PER_QUEUE) {
+		r = amdgpu_ring_test_ring(ring);
+		if (r)
+			goto pipe_reset;
+	}
+
+
+	return amdgpu_ring_reset_helper_end(ring, timedout_fence);
+}
+
+enum amdgpu_gfx_cp_ras_mem_id {
+	AMDGPU_GFX_CP_MEM1 = 1,
+	AMDGPU_GFX_CP_MEM2,
+	AMDGPU_GFX_CP_MEM3,
+	AMDGPU_GFX_CP_MEM4,
+	AMDGPU_GFX_CP_MEM5,
+};
+
+enum amdgpu_gfx_gcea_ras_mem_id {
+	AMDGPU_GFX_GCEA_IOWR_CMDMEM = 4,
+	AMDGPU_GFX_GCEA_IORD_CMDMEM,
+	AMDGPU_GFX_GCEA_GMIWR_CMDMEM,
+	AMDGPU_GFX_GCEA_GMIRD_CMDMEM,
+	AMDGPU_GFX_GCEA_DRAMWR_CMDMEM,
+	AMDGPU_GFX_GCEA_DRAMRD_CMDMEM,
+	AMDGPU_GFX_GCEA_MAM_DMEM0,
+	AMDGPU_GFX_GCEA_MAM_DMEM1,
+	AMDGPU_GFX_GCEA_MAM_DMEM2,
+	AMDGPU_GFX_GCEA_MAM_DMEM3,
+	AMDGPU_GFX_GCEA_MAM_AMEM0,
+	AMDGPU_GFX_GCEA_MAM_AMEM1,
+	AMDGPU_GFX_GCEA_MAM_AMEM2,
+	AMDGPU_GFX_GCEA_MAM_AMEM3,
+	AMDGPU_GFX_GCEA_MAM_AFLUSH_BUFFER,
+	AMDGPU_GFX_GCEA_WRET_TAGMEM,
+	AMDGPU_GFX_GCEA_RRET_TAGMEM,
+	AMDGPU_GFX_GCEA_IOWR_DATAMEM,
+	AMDGPU_GFX_GCEA_GMIWR_DATAMEM,
+	AMDGPU_GFX_GCEA_DRAM_DATAMEM,
+};
+
+enum amdgpu_gfx_gc_cane_ras_mem_id {
+	AMDGPU_GFX_GC_CANE_MEM0 = 0,
+};
+
+enum amdgpu_gfx_gcutcl2_ras_mem_id {
+	AMDGPU_GFX_GCUTCL2_MEM2P512X95 = 160,
+};
+
+enum amdgpu_gfx_gds_ras_mem_id {
+	AMDGPU_GFX_GDS_MEM0 = 0,
+};
+
+enum amdgpu_gfx_lds_ras_mem_id {
+	AMDGPU_GFX_LDS_BANK0 = 0,
+	AMDGPU_GFX_LDS_BANK1,
+	AMDGPU_GFX_LDS_BANK2,
+	AMDGPU_GFX_LDS_BANK3,
+	AMDGPU_GFX_LDS_BANK4,
+	AMDGPU_GFX_LDS_BANK5,
+	AMDGPU_GFX_LDS_BANK6,
+	AMDGPU_GFX_LDS_BANK7,
+	AMDGPU_GFX_LDS_BANK8,
+	AMDGPU_GFX_LDS_BANK9,
+	AMDGPU_GFX_LDS_BANK10,
+	AMDGPU_GFX_LDS_BANK11,
+	AMDGPU_GFX_LDS_BANK12,
+	AMDGPU_GFX_LDS_BANK13,
+	AMDGPU_GFX_LDS_BANK14,
+	AMDGPU_GFX_LDS_BANK15,
+	AMDGPU_GFX_LDS_BANK16,
+	AMDGPU_GFX_LDS_BANK17,
+	AMDGPU_GFX_LDS_BANK18,
+	AMDGPU_GFX_LDS_BANK19,
+	AMDGPU_GFX_LDS_BANK20,
+	AMDGPU_GFX_LDS_BANK21,
+	AMDGPU_GFX_LDS_BANK22,
+	AMDGPU_GFX_LDS_BANK23,
+	AMDGPU_GFX_LDS_BANK24,
+	AMDGPU_GFX_LDS_BANK25,
+	AMDGPU_GFX_LDS_BANK26,
+	AMDGPU_GFX_LDS_BANK27,
+	AMDGPU_GFX_LDS_BANK28,
+	AMDGPU_GFX_LDS_BANK29,
+	AMDGPU_GFX_LDS_BANK30,
+	AMDGPU_GFX_LDS_BANK31,
+	AMDGPU_GFX_LDS_SP_BUFFER_A,
+	AMDGPU_GFX_LDS_SP_BUFFER_B,
+};
+
+enum amdgpu_gfx_rlc_ras_mem_id {
+	AMDGPU_GFX_RLC_GPMF32 = 1,
+	AMDGPU_GFX_RLC_RLCVF32,
+	AMDGPU_GFX_RLC_SCRATCH,
+	AMDGPU_GFX_RLC_SRM_ARAM,
+	AMDGPU_GFX_RLC_SRM_DRAM,
+	AMDGPU_GFX_RLC_TCTAG,
+	AMDGPU_GFX_RLC_SPM_SE,
+	AMDGPU_GFX_RLC_SPM_GRBMT,
+};
+
+enum amdgpu_gfx_sp_ras_mem_id {
+	AMDGPU_GFX_SP_SIMDID0 = 0,
+};
+
+enum amdgpu_gfx_spi_ras_mem_id {
+	AMDGPU_GFX_SPI_MEM0 = 0,
+	AMDGPU_GFX_SPI_MEM1,
+	AMDGPU_GFX_SPI_MEM2,
+	AMDGPU_GFX_SPI_MEM3,
+};
+
+enum amdgpu_gfx_sqc_ras_mem_id {
+	AMDGPU_GFX_SQC_INST_CACHE_A = 100,
+	AMDGPU_GFX_SQC_INST_CACHE_B = 101,
+	AMDGPU_GFX_SQC_INST_CACHE_TAG_A = 102,
+	AMDGPU_GFX_SQC_INST_CACHE_TAG_B = 103,
+	AMDGPU_GFX_SQC_INST_CACHE_MISS_FIFO_A = 104,
+	AMDGPU_GFX_SQC_INST_CACHE_MISS_FIFO_B = 105,
+	AMDGPU_GFX_SQC_INST_CACHE_GATCL1_MISS_FIFO_A = 106,
+	AMDGPU_GFX_SQC_INST_CACHE_GATCL1_MISS_FIFO_B = 107,
+	AMDGPU_GFX_SQC_DATA_CACHE_A = 200,
+	AMDGPU_GFX_SQC_DATA_CACHE_B = 201,
+	AMDGPU_GFX_SQC_DATA_CACHE_TAG_A = 202,
+	AMDGPU_GFX_SQC_DATA_CACHE_TAG_B = 203,
+	AMDGPU_GFX_SQC_DATA_CACHE_MISS_FIFO_A = 204,
+	AMDGPU_GFX_SQC_DATA_CACHE_MISS_FIFO_B = 205,
+	AMDGPU_GFX_SQC_DATA_CACHE_HIT_FIFO_A = 206,
+	AMDGPU_GFX_SQC_DATA_CACHE_HIT_FIFO_B = 207,
+	AMDGPU_GFX_SQC_DIRTY_BIT_A = 208,
+	AMDGPU_GFX_SQC_DIRTY_BIT_B = 209,
+	AMDGPU_GFX_SQC_WRITE_DATA_BUFFER_CU0 = 210,
+	AMDGPU_GFX_SQC_WRITE_DATA_BUFFER_CU1 = 211,
+	AMDGPU_GFX_SQC_UTCL1_MISS_LFIFO_DATA_CACHE_A = 212,
+	AMDGPU_GFX_SQC_UTCL1_MISS_LFIFO_DATA_CACHE_B = 213,
+	AMDGPU_GFX_SQC_UTCL1_MISS_LFIFO_INST_CACHE = 108,
+};
+
+enum amdgpu_gfx_sq_ras_mem_id {
+	AMDGPU_GFX_SQ_SGPR_MEM0 = 0,
+	AMDGPU_GFX_SQ_SGPR_MEM1,
+	AMDGPU_GFX_SQ_SGPR_MEM2,
+	AMDGPU_GFX_SQ_SGPR_MEM3,
+};
+
+enum amdgpu_gfx_ta_ras_mem_id {
+	AMDGPU_GFX_TA_FS_AFIFO_RAM_LO = 1,
+	AMDGPU_GFX_TA_FS_AFIFO_RAM_HI,
+	AMDGPU_GFX_TA_FS_CFIFO_RAM,
+	AMDGPU_GFX_TA_FSX_LFIFO,
+	AMDGPU_GFX_TA_FS_DFIFO_RAM,
+};
+
+enum amdgpu_gfx_tcc_ras_mem_id {
+	AMDGPU_GFX_TCC_MEM1 = 1,
+};
+
+enum amdgpu_gfx_tca_ras_mem_id {
+	AMDGPU_GFX_TCA_MEM1 = 1,
+};
+
+enum amdgpu_gfx_tci_ras_mem_id {
+	AMDGPU_GFX_TCIW_MEM = 1,
+};
+
+enum amdgpu_gfx_tcp_ras_mem_id {
+	AMDGPU_GFX_TCP_LFIFO0 = 1,
+	AMDGPU_GFX_TCP_SET0BANK0_RAM,
+	AMDGPU_GFX_TCP_SET0BANK1_RAM,
+	AMDGPU_GFX_TCP_SET0BANK2_RAM,
+	AMDGPU_GFX_TCP_SET0BANK3_RAM,
+	AMDGPU_GFX_TCP_SET1BANK0_RAM,
+	AMDGPU_GFX_TCP_SET1BANK1_RAM,
+	AMDGPU_GFX_TCP_SET1BANK2_RAM,
+	AMDGPU_GFX_TCP_SET1BANK3_RAM,
+	AMDGPU_GFX_TCP_SET2BANK0_RAM,
+	AMDGPU_GFX_TCP_SET2BANK1_RAM,
+	AMDGPU_GFX_TCP_SET2BANK2_RAM,
+	AMDGPU_GFX_TCP_SET2BANK3_RAM,
+	AMDGPU_GFX_TCP_SET3BANK0_RAM,
+	AMDGPU_GFX_TCP_SET3BANK1_RAM,
+	AMDGPU_GFX_TCP_SET3BANK2_RAM,
+	AMDGPU_GFX_TCP_SET3BANK3_RAM,
+	AMDGPU_GFX_TCP_VM_FIFO,
+	AMDGPU_GFX_TCP_DB_TAGRAM0,
+	AMDGPU_GFX_TCP_DB_TAGRAM1,
+	AMDGPU_GFX_TCP_DB_TAGRAM2,
+	AMDGPU_GFX_TCP_DB_TAGRAM3,
+	AMDGPU_GFX_TCP_UTCL1_LFIFO_PROBE0,
+	AMDGPU_GFX_TCP_UTCL1_LFIFO_PROBE1,
+	AMDGPU_GFX_TCP_CMD_FIFO,
+};
+
+enum amdgpu_gfx_td_ras_mem_id {
+	AMDGPU_GFX_TD_UTD_CS_FIFO_MEM = 1,
+	AMDGPU_GFX_TD_UTD_SS_FIFO_LO_MEM,
+	AMDGPU_GFX_TD_UTD_SS_FIFO_HI_MEM,
+};
+
+enum amdgpu_gfx_tcx_ras_mem_id {
+	AMDGPU_GFX_TCX_FIFOD0 = 0,
+	AMDGPU_GFX_TCX_FIFOD1,
+	AMDGPU_GFX_TCX_FIFOD2,
+	AMDGPU_GFX_TCX_FIFOD3,
+	AMDGPU_GFX_TCX_FIFOD4,
+	AMDGPU_GFX_TCX_FIFOD5,
+	AMDGPU_GFX_TCX_FIFOD6,
+	AMDGPU_GFX_TCX_FIFOD7,
+	AMDGPU_GFX_TCX_FIFOB0,
+	AMDGPU_GFX_TCX_FIFOB1,
+	AMDGPU_GFX_TCX_FIFOB2,
+	AMDGPU_GFX_TCX_FIFOB3,
+	AMDGPU_GFX_TCX_FIFOB4,
+	AMDGPU_GFX_TCX_FIFOB5,
+	AMDGPU_GFX_TCX_FIFOB6,
+	AMDGPU_GFX_TCX_FIFOB7,
+	AMDGPU_GFX_TCX_FIFOA0,
+	AMDGPU_GFX_TCX_FIFOA1,
+	AMDGPU_GFX_TCX_FIFOA2,
+	AMDGPU_GFX_TCX_FIFOA3,
+	AMDGPU_GFX_TCX_FIFOA4,
+	AMDGPU_GFX_TCX_FIFOA5,
+	AMDGPU_GFX_TCX_FIFOA6,
+	AMDGPU_GFX_TCX_FIFOA7,
+	AMDGPU_GFX_TCX_CFIFO0,
+	AMDGPU_GFX_TCX_CFIFO1,
+	AMDGPU_GFX_TCX_CFIFO2,
+	AMDGPU_GFX_TCX_CFIFO3,
+	AMDGPU_GFX_TCX_CFIFO4,
+	AMDGPU_GFX_TCX_CFIFO5,
+	AMDGPU_GFX_TCX_CFIFO6,
+	AMDGPU_GFX_TCX_CFIFO7,
+	AMDGPU_GFX_TCX_FIFO_ACKB0,
+	AMDGPU_GFX_TCX_FIFO_ACKB1,
+	AMDGPU_GFX_TCX_FIFO_ACKB2,
+	AMDGPU_GFX_TCX_FIFO_ACKB3,
+	AMDGPU_GFX_TCX_FIFO_ACKB4,
+	AMDGPU_GFX_TCX_FIFO_ACKB5,
+	AMDGPU_GFX_TCX_FIFO_ACKB6,
+	AMDGPU_GFX_TCX_FIFO_ACKB7,
+	AMDGPU_GFX_TCX_FIFO_ACKD0,
+	AMDGPU_GFX_TCX_FIFO_ACKD1,
+	AMDGPU_GFX_TCX_FIFO_ACKD2,
+	AMDGPU_GFX_TCX_FIFO_ACKD3,
+	AMDGPU_GFX_TCX_FIFO_ACKD4,
+	AMDGPU_GFX_TCX_FIFO_ACKD5,
+	AMDGPU_GFX_TCX_FIFO_ACKD6,
+	AMDGPU_GFX_TCX_FIFO_ACKD7,
+	AMDGPU_GFX_TCX_DST_FIFOA0,
+	AMDGPU_GFX_TCX_DST_FIFOA1,
+	AMDGPU_GFX_TCX_DST_FIFOA2,
+	AMDGPU_GFX_TCX_DST_FIFOA3,
+	AMDGPU_GFX_TCX_DST_FIFOA4,
+	AMDGPU_GFX_TCX_DST_FIFOA5,
+	AMDGPU_GFX_TCX_DST_FIFOA6,
+	AMDGPU_GFX_TCX_DST_FIFOA7,
+	AMDGPU_GFX_TCX_DST_FIFOB0,
+	AMDGPU_GFX_TCX_DST_FIFOB1,
+	AMDGPU_GFX_TCX_DST_FIFOB2,
+	AMDGPU_GFX_TCX_DST_FIFOB3,
+	AMDGPU_GFX_TCX_DST_FIFOB4,
+	AMDGPU_GFX_TCX_DST_FIFOB5,
+	AMDGPU_GFX_TCX_DST_FIFOB6,
+	AMDGPU_GFX_TCX_DST_FIFOB7,
+	AMDGPU_GFX_TCX_DST_FIFOD0,
+	AMDGPU_GFX_TCX_DST_FIFOD1,
+	AMDGPU_GFX_TCX_DST_FIFOD2,
+	AMDGPU_GFX_TCX_DST_FIFOD3,
+	AMDGPU_GFX_TCX_DST_FIFOD4,
+	AMDGPU_GFX_TCX_DST_FIFOD5,
+	AMDGPU_GFX_TCX_DST_FIFOD6,
+	AMDGPU_GFX_TCX_DST_FIFOD7,
+	AMDGPU_GFX_TCX_DST_FIFO_ACKB0,
+	AMDGPU_GFX_TCX_DST_FIFO_ACKB1,
+	AMDGPU_GFX_TCX_DST_FIFO_ACKB2,
+	AMDGPU_GFX_TCX_DST_FIFO_ACKB3,
+	AMDGPU_GFX_TCX_DST_FIFO_ACKB4,
+	AMDGPU_GFX_TCX_DST_FIFO_ACKB5,
+	AMDGPU_GFX_TCX_DST_FIFO_ACKB6,
+	AMDGPU_GFX_TCX_DST_FIFO_ACKB7,
+	AMDGPU_GFX_TCX_DST_FIFO_ACKD0,
+	AMDGPU_GFX_TCX_DST_FIFO_ACKD1,
+	AMDGPU_GFX_TCX_DST_FIFO_ACKD2,
+	AMDGPU_GFX_TCX_DST_FIFO_ACKD3,
+	AMDGPU_GFX_TCX_DST_FIFO_ACKD4,
+	AMDGPU_GFX_TCX_DST_FIFO_ACKD5,
+	AMDGPU_GFX_TCX_DST_FIFO_ACKD6,
+	AMDGPU_GFX_TCX_DST_FIFO_ACKD7,
+};
+
+enum amdgpu_gfx_atc_l2_ras_mem_id {
+	AMDGPU_GFX_ATC_L2_MEM0 = 0,
+};
+
+enum amdgpu_gfx_utcl2_ras_mem_id {
+	AMDGPU_GFX_UTCL2_MEM0 = 0,
+};
+
+enum amdgpu_gfx_vml2_ras_mem_id {
+	AMDGPU_GFX_VML2_MEM0 = 0,
+};
+
+enum amdgpu_gfx_vml2_walker_ras_mem_id {
+	AMDGPU_GFX_VML2_WALKER_MEM0 = 0,
+};
+
+static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_cp_mem_list[] = {
+	{AMDGPU_GFX_CP_MEM1, "CP_MEM1"},
+	{AMDGPU_GFX_CP_MEM2, "CP_MEM2"},
+	{AMDGPU_GFX_CP_MEM3, "CP_MEM3"},
+	{AMDGPU_GFX_CP_MEM4, "CP_MEM4"},
+	{AMDGPU_GFX_CP_MEM5, "CP_MEM5"},
+};
+
+static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_gcea_mem_list[] = {
+	{AMDGPU_GFX_GCEA_IOWR_CMDMEM, "GCEA_IOWR_CMDMEM"},
+	{AMDGPU_GFX_GCEA_IORD_CMDMEM, "GCEA_IORD_CMDMEM"},
+	{AMDGPU_GFX_GCEA_GMIWR_CMDMEM, "GCEA_GMIWR_CMDMEM"},
+	{AMDGPU_GFX_GCEA_GMIRD_CMDMEM, "GCEA_GMIRD_CMDMEM"},
+	{AMDGPU_GFX_GCEA_DRAMWR_CMDMEM, "GCEA_DRAMWR_CMDMEM"},
+	{AMDGPU_GFX_GCEA_DRAMRD_CMDMEM, "GCEA_DRAMRD_CMDMEM"},
+	{AMDGPU_GFX_GCEA_MAM_DMEM0, "GCEA_MAM_DMEM0"},
+	{AMDGPU_GFX_GCEA_MAM_DMEM1, "GCEA_MAM_DMEM1"},
+	{AMDGPU_GFX_GCEA_MAM_DMEM2, "GCEA_MAM_DMEM2"},
+	{AMDGPU_GFX_GCEA_MAM_DMEM3, "GCEA_MAM_DMEM3"},
+	{AMDGPU_GFX_GCEA_MAM_AMEM0, "GCEA_MAM_AMEM0"},
+	{AMDGPU_GFX_GCEA_MAM_AMEM1, "GCEA_MAM_AMEM1"},
+	{AMDGPU_GFX_GCEA_MAM_AMEM2, "GCEA_MAM_AMEM2"},
+	{AMDGPU_GFX_GCEA_MAM_AMEM3, "GCEA_MAM_AMEM3"},
+	{AMDGPU_GFX_GCEA_MAM_AFLUSH_BUFFER, "GCEA_MAM_AFLUSH_BUFFER"},
+	{AMDGPU_GFX_GCEA_WRET_TAGMEM, "GCEA_WRET_TAGMEM"},
+	{AMDGPU_GFX_GCEA_RRET_TAGMEM, "GCEA_RRET_TAGMEM"},
+	{AMDGPU_GFX_GCEA_IOWR_DATAMEM, "GCEA_IOWR_DATAMEM"},
+	{AMDGPU_GFX_GCEA_GMIWR_DATAMEM, "GCEA_GMIWR_DATAMEM"},
+	{AMDGPU_GFX_GCEA_DRAM_DATAMEM, "GCEA_DRAM_DATAMEM"},
+};
+
+static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_gc_cane_mem_list[] = {
+	{AMDGPU_GFX_GC_CANE_MEM0, "GC_CANE_MEM0"},
+};
+
+static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_gcutcl2_mem_list[] = {
+	{AMDGPU_GFX_GCUTCL2_MEM2P512X95, "GCUTCL2_MEM2P512X95"},
+};
+
+static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_gds_mem_list[] = {
+	{AMDGPU_GFX_GDS_MEM0, "GDS_MEM"},
+};
+
+static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_lds_mem_list[] = {
+	{AMDGPU_GFX_LDS_BANK0, "LDS_BANK0"},
+	{AMDGPU_GFX_LDS_BANK1, "LDS_BANK1"},
+	{AMDGPU_GFX_LDS_BANK2, "LDS_BANK2"},
+	{AMDGPU_GFX_LDS_BANK3, "LDS_BANK3"},
+	{AMDGPU_GFX_LDS_BANK4, "LDS_BANK4"},
+	{AMDGPU_GFX_LDS_BANK5, "LDS_BANK5"},
+	{AMDGPU_GFX_LDS_BANK6, "LDS_BANK6"},
+	{AMDGPU_GFX_LDS_BANK7, "LDS_BANK7"},
+	{AMDGPU_GFX_LDS_BANK8, "LDS_BANK8"},
+	{AMDGPU_GFX_LDS_BANK9, "LDS_BANK9"},
+	{AMDGPU_GFX_LDS_BANK10, "LDS_BANK10"},
+	{AMDGPU_GFX_LDS_BANK11, "LDS_BANK11"},
+	{AMDGPU_GFX_LDS_BANK12, "LDS_BANK12"},
+	{AMDGPU_GFX_LDS_BANK13, "LDS_BANK13"},
+	{AMDGPU_GFX_LDS_BANK14, "LDS_BANK14"},
+	{AMDGPU_GFX_LDS_BANK15, "LDS_BANK15"},
+	{AMDGPU_GFX_LDS_BANK16, "LDS_BANK16"},
+	{AMDGPU_GFX_LDS_BANK17, "LDS_BANK17"},
+	{AMDGPU_GFX_LDS_BANK18, "LDS_BANK18"},
+	{AMDGPU_GFX_LDS_BANK19, "LDS_BANK19"},
+	{AMDGPU_GFX_LDS_BANK20, "LDS_BANK20"},
+	{AMDGPU_GFX_LDS_BANK21, "LDS_BANK21"},
+	{AMDGPU_GFX_LDS_BANK22, "LDS_BANK22"},
+	{AMDGPU_GFX_LDS_BANK23, "LDS_BANK23"},
+	{AMDGPU_GFX_LDS_BANK24, "LDS_BANK24"},
+	{AMDGPU_GFX_LDS_BANK25, "LDS_BANK25"},
+	{AMDGPU_GFX_LDS_BANK26, "LDS_BANK26"},
+	{AMDGPU_GFX_LDS_BANK27, "LDS_BANK27"},
+	{AMDGPU_GFX_LDS_BANK28, "LDS_BANK28"},
+	{AMDGPU_GFX_LDS_BANK29, "LDS_BANK29"},
+	{AMDGPU_GFX_LDS_BANK30, "LDS_BANK30"},
+	{AMDGPU_GFX_LDS_BANK31, "LDS_BANK31"},
+	{AMDGPU_GFX_LDS_SP_BUFFER_A, "LDS_SP_BUFFER_A"},
+	{AMDGPU_GFX_LDS_SP_BUFFER_B, "LDS_SP_BUFFER_B"},
+};
+
+static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_rlc_mem_list[] = {
+	{AMDGPU_GFX_RLC_GPMF32, "RLC_GPMF32"},
+	{AMDGPU_GFX_RLC_RLCVF32, "RLC_RLCVF32"},
+	{AMDGPU_GFX_RLC_SCRATCH, "RLC_SCRATCH"},
+	{AMDGPU_GFX_RLC_SRM_ARAM, "RLC_SRM_ARAM"},
+	{AMDGPU_GFX_RLC_SRM_DRAM, "RLC_SRM_DRAM"},
+	{AMDGPU_GFX_RLC_TCTAG, "RLC_TCTAG"},
+	{AMDGPU_GFX_RLC_SPM_SE, "RLC_SPM_SE"},
+	{AMDGPU_GFX_RLC_SPM_GRBMT, "RLC_SPM_GRBMT"},
+};
+
+static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_sp_mem_list[] = {
+	{AMDGPU_GFX_SP_SIMDID0, "SP_SIMDID0"},
+};
+
+static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_spi_mem_list[] = {
+	{AMDGPU_GFX_SPI_MEM0, "SPI_MEM0"},
+	{AMDGPU_GFX_SPI_MEM1, "SPI_MEM1"},
+	{AMDGPU_GFX_SPI_MEM2, "SPI_MEM2"},
+	{AMDGPU_GFX_SPI_MEM3, "SPI_MEM3"},
+};
+
+static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_sqc_mem_list[] = {
+	{AMDGPU_GFX_SQC_INST_CACHE_A, "SQC_INST_CACHE_A"},
+	{AMDGPU_GFX_SQC_INST_CACHE_B, "SQC_INST_CACHE_B"},
+	{AMDGPU_GFX_SQC_INST_CACHE_TAG_A, "SQC_INST_CACHE_TAG_A"},
+	{AMDGPU_GFX_SQC_INST_CACHE_TAG_B, "SQC_INST_CACHE_TAG_B"},
+	{AMDGPU_GFX_SQC_INST_CACHE_MISS_FIFO_A, "SQC_INST_CACHE_MISS_FIFO_A"},
+	{AMDGPU_GFX_SQC_INST_CACHE_MISS_FIFO_B, "SQC_INST_CACHE_MISS_FIFO_B"},
+	{AMDGPU_GFX_SQC_INST_CACHE_GATCL1_MISS_FIFO_A, "SQC_INST_CACHE_GATCL1_MISS_FIFO_A"},
+	{AMDGPU_GFX_SQC_INST_CACHE_GATCL1_MISS_FIFO_B, "SQC_INST_CACHE_GATCL1_MISS_FIFO_B"},
+	{AMDGPU_GFX_SQC_DATA_CACHE_A, "SQC_DATA_CACHE_A"},
+	{AMDGPU_GFX_SQC_DATA_CACHE_B, "SQC_DATA_CACHE_B"},
+	{AMDGPU_GFX_SQC_DATA_CACHE_TAG_A, "SQC_DATA_CACHE_TAG_A"},
+	{AMDGPU_GFX_SQC_DATA_CACHE_TAG_B, "SQC_DATA_CACHE_TAG_B"},
+	{AMDGPU_GFX_SQC_DATA_CACHE_MISS_FIFO_A, "SQC_DATA_CACHE_MISS_FIFO_A"},
+	{AMDGPU_GFX_SQC_DATA_CACHE_MISS_FIFO_B, "SQC_DATA_CACHE_MISS_FIFO_B"},
+	{AMDGPU_GFX_SQC_DATA_CACHE_HIT_FIFO_A, "SQC_DATA_CACHE_HIT_FIFO_A"},
+	{AMDGPU_GFX_SQC_DATA_CACHE_HIT_FIFO_B, "SQC_DATA_CACHE_HIT_FIFO_B"},
+	{AMDGPU_GFX_SQC_DIRTY_BIT_A, "SQC_DIRTY_BIT_A"},
+	{AMDGPU_GFX_SQC_DIRTY_BIT_B, "SQC_DIRTY_BIT_B"},
+	{AMDGPU_GFX_SQC_WRITE_DATA_BUFFER_CU0, "SQC_WRITE_DATA_BUFFER_CU0"},
+	{AMDGPU_GFX_SQC_WRITE_DATA_BUFFER_CU1, "SQC_WRITE_DATA_BUFFER_CU1"},
+	{AMDGPU_GFX_SQC_UTCL1_MISS_LFIFO_DATA_CACHE_A, "SQC_UTCL1_MISS_LFIFO_DATA_CACHE_A"},
+	{AMDGPU_GFX_SQC_UTCL1_MISS_LFIFO_DATA_CACHE_B, "SQC_UTCL1_MISS_LFIFO_DATA_CACHE_B"},
+	{AMDGPU_GFX_SQC_UTCL1_MISS_LFIFO_INST_CACHE, "SQC_UTCL1_MISS_LFIFO_INST_CACHE"},
+};
+
+static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_sq_mem_list[] = {
+	{AMDGPU_GFX_SQ_SGPR_MEM0, "SQ_SGPR_MEM0"},
+	{AMDGPU_GFX_SQ_SGPR_MEM1, "SQ_SGPR_MEM1"},
+	{AMDGPU_GFX_SQ_SGPR_MEM2, "SQ_SGPR_MEM2"},
+	{AMDGPU_GFX_SQ_SGPR_MEM3, "SQ_SGPR_MEM3"},
+};
+
+static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_ta_mem_list[] = {
+	{AMDGPU_GFX_TA_FS_AFIFO_RAM_LO, "TA_FS_AFIFO_RAM_LO"},
+	{AMDGPU_GFX_TA_FS_AFIFO_RAM_HI, "TA_FS_AFIFO_RAM_HI"},
+	{AMDGPU_GFX_TA_FS_CFIFO_RAM, "TA_FS_CFIFO_RAM"},
+	{AMDGPU_GFX_TA_FSX_LFIFO, "TA_FSX_LFIFO"},
+	{AMDGPU_GFX_TA_FS_DFIFO_RAM, "TA_FS_DFIFO_RAM"},
+};
+
+static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_tcc_mem_list[] = {
+	{AMDGPU_GFX_TCC_MEM1, "TCC_MEM1"},
+};
+
+static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_tca_mem_list[] = {
+	{AMDGPU_GFX_TCA_MEM1, "TCA_MEM1"},
+};
+
+static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_tci_mem_list[] = {
+	{AMDGPU_GFX_TCIW_MEM, "TCIW_MEM"},
+};
+
+static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_tcp_mem_list[] = {
+	{AMDGPU_GFX_TCP_LFIFO0, "TCP_LFIFO0"},
+	{AMDGPU_GFX_TCP_SET0BANK0_RAM, "TCP_SET0BANK0_RAM"},
+	{AMDGPU_GFX_TCP_SET0BANK1_RAM, "TCP_SET0BANK1_RAM"},
+	{AMDGPU_GFX_TCP_SET0BANK2_RAM, "TCP_SET0BANK2_RAM"},
+	{AMDGPU_GFX_TCP_SET0BANK3_RAM, "TCP_SET0BANK3_RAM"},
+	{AMDGPU_GFX_TCP_SET1BANK0_RAM, "TCP_SET1BANK0_RAM"},
+	{AMDGPU_GFX_TCP_SET1BANK1_RAM, "TCP_SET1BANK1_RAM"},
+	{AMDGPU_GFX_TCP_SET1BANK2_RAM, "TCP_SET1BANK2_RAM"},
+	{AMDGPU_GFX_TCP_SET1BANK3_RAM, "TCP_SET1BANK3_RAM"},
+	{AMDGPU_GFX_TCP_SET2BANK0_RAM, "TCP_SET2BANK0_RAM"},
+	{AMDGPU_GFX_TCP_SET2BANK1_RAM, "TCP_SET2BANK1_RAM"},
+	{AMDGPU_GFX_TCP_SET2BANK2_RAM, "TCP_SET2BANK2_RAM"},
+	{AMDGPU_GFX_TCP_SET2BANK3_RAM, "TCP_SET2BANK3_RAM"},
+	{AMDGPU_GFX_TCP_SET3BANK0_RAM, "TCP_SET3BANK0_RAM"},
+	{AMDGPU_GFX_TCP_SET3BANK1_RAM, "TCP_SET3BANK1_RAM"},
+	{AMDGPU_GFX_TCP_SET3BANK2_RAM, "TCP_SET3BANK2_RAM"},
+	{AMDGPU_GFX_TCP_SET3BANK3_RAM, "TCP_SET3BANK3_RAM"},
+	{AMDGPU_GFX_TCP_VM_FIFO, "TCP_VM_FIFO"},
+	{AMDGPU_GFX_TCP_DB_TAGRAM0, "TCP_DB_TAGRAM0"},
+	{AMDGPU_GFX_TCP_DB_TAGRAM1, "TCP_DB_TAGRAM1"},
+	{AMDGPU_GFX_TCP_DB_TAGRAM2, "TCP_DB_TAGRAM2"},
+	{AMDGPU_GFX_TCP_DB_TAGRAM3, "TCP_DB_TAGRAM3"},
+	{AMDGPU_GFX_TCP_UTCL1_LFIFO_PROBE0, "TCP_UTCL1_LFIFO_PROBE0"},
+	{AMDGPU_GFX_TCP_UTCL1_LFIFO_PROBE1, "TCP_UTCL1_LFIFO_PROBE1"},
+	{AMDGPU_GFX_TCP_CMD_FIFO, "TCP_CMD_FIFO"},
+};
+
+static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_td_mem_list[] = {
+	{AMDGPU_GFX_TD_UTD_CS_FIFO_MEM, "TD_UTD_CS_FIFO_MEM"},
+	{AMDGPU_GFX_TD_UTD_SS_FIFO_LO_MEM, "TD_UTD_SS_FIFO_LO_MEM"},
+	{AMDGPU_GFX_TD_UTD_SS_FIFO_HI_MEM, "TD_UTD_SS_FIFO_HI_MEM"},
+};
+
+static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_tcx_mem_list[] = {
+	{AMDGPU_GFX_TCX_FIFOD0, "TCX_FIFOD0"},
+	{AMDGPU_GFX_TCX_FIFOD1, "TCX_FIFOD1"},
+	{AMDGPU_GFX_TCX_FIFOD2, "TCX_FIFOD2"},
+	{AMDGPU_GFX_TCX_FIFOD3, "TCX_FIFOD3"},
+	{AMDGPU_GFX_TCX_FIFOD4, "TCX_FIFOD4"},
+	{AMDGPU_GFX_TCX_FIFOD5, "TCX_FIFOD5"},
+	{AMDGPU_GFX_TCX_FIFOD6, "TCX_FIFOD6"},
+	{AMDGPU_GFX_TCX_FIFOD7, "TCX_FIFOD7"},
+	{AMDGPU_GFX_TCX_FIFOB0, "TCX_FIFOB0"},
+	{AMDGPU_GFX_TCX_FIFOB1, "TCX_FIFOB1"},
+	{AMDGPU_GFX_TCX_FIFOB2, "TCX_FIFOB2"},
+	{AMDGPU_GFX_TCX_FIFOB3, "TCX_FIFOB3"},
+	{AMDGPU_GFX_TCX_FIFOB4, "TCX_FIFOB4"},
+	{AMDGPU_GFX_TCX_FIFOB5, "TCX_FIFOB5"},
+	{AMDGPU_GFX_TCX_FIFOB6, "TCX_FIFOB6"},
+	{AMDGPU_GFX_TCX_FIFOB7, "TCX_FIFOB7"},
+	{AMDGPU_GFX_TCX_FIFOA0, "TCX_FIFOA0"},
+	{AMDGPU_GFX_TCX_FIFOA1, "TCX_FIFOA1"},
+	{AMDGPU_GFX_TCX_FIFOA2, "TCX_FIFOA2"},
+	{AMDGPU_GFX_TCX_FIFOA3, "TCX_FIFOA3"},
+	{AMDGPU_GFX_TCX_FIFOA4, "TCX_FIFOA4"},
+	{AMDGPU_GFX_TCX_FIFOA5, "TCX_FIFOA5"},
+	{AMDGPU_GFX_TCX_FIFOA6, "TCX_FIFOA6"},
+	{AMDGPU_GFX_TCX_FIFOA7, "TCX_FIFOA7"},
+	{AMDGPU_GFX_TCX_CFIFO0, "TCX_CFIFO0"},
+	{AMDGPU_GFX_TCX_CFIFO1, "TCX_CFIFO1"},
+	{AMDGPU_GFX_TCX_CFIFO2, "TCX_CFIFO2"},
+	{AMDGPU_GFX_TCX_CFIFO3, "TCX_CFIFO3"},
+	{AMDGPU_GFX_TCX_CFIFO4, "TCX_CFIFO4"},
+	{AMDGPU_GFX_TCX_CFIFO5, "TCX_CFIFO5"},
+	{AMDGPU_GFX_TCX_CFIFO6, "TCX_CFIFO6"},
+	{AMDGPU_GFX_TCX_CFIFO7, "TCX_CFIFO7"},
+	{AMDGPU_GFX_TCX_FIFO_ACKB0, "TCX_FIFO_ACKB0"},
+	{AMDGPU_GFX_TCX_FIFO_ACKB1, "TCX_FIFO_ACKB1"},
+	{AMDGPU_GFX_TCX_FIFO_ACKB2, "TCX_FIFO_ACKB2"},
+	{AMDGPU_GFX_TCX_FIFO_ACKB3, "TCX_FIFO_ACKB3"},
+	{AMDGPU_GFX_TCX_FIFO_ACKB4, "TCX_FIFO_ACKB4"},
+	{AMDGPU_GFX_TCX_FIFO_ACKB5, "TCX_FIFO_ACKB5"},
+	{AMDGPU_GFX_TCX_FIFO_ACKB6, "TCX_FIFO_ACKB6"},
+	{AMDGPU_GFX_TCX_FIFO_ACKB7, "TCX_FIFO_ACKB7"},
+	{AMDGPU_GFX_TCX_FIFO_ACKD0, "TCX_FIFO_ACKD0"},
+	{AMDGPU_GFX_TCX_FIFO_ACKD1, "TCX_FIFO_ACKD1"},
+	{AMDGPU_GFX_TCX_FIFO_ACKD2, "TCX_FIFO_ACKD2"},
+	{AMDGPU_GFX_TCX_FIFO_ACKD3, "TCX_FIFO_ACKD3"},
+	{AMDGPU_GFX_TCX_FIFO_ACKD4, "TCX_FIFO_ACKD4"},
+	{AMDGPU_GFX_TCX_FIFO_ACKD5, "TCX_FIFO_ACKD5"},
+	{AMDGPU_GFX_TCX_FIFO_ACKD6, "TCX_FIFO_ACKD6"},
+	{AMDGPU_GFX_TCX_FIFO_ACKD7, "TCX_FIFO_ACKD7"},
+	{AMDGPU_GFX_TCX_DST_FIFOA0, "TCX_DST_FIFOA0"},
+	{AMDGPU_GFX_TCX_DST_FIFOA1, "TCX_DST_FIFOA1"},
+	{AMDGPU_GFX_TCX_DST_FIFOA2, "TCX_DST_FIFOA2"},
+	{AMDGPU_GFX_TCX_DST_FIFOA3, "TCX_DST_FIFOA3"},
+	{AMDGPU_GFX_TCX_DST_FIFOA4, "TCX_DST_FIFOA4"},
+	{AMDGPU_GFX_TCX_DST_FIFOA5, "TCX_DST_FIFOA5"},
+	{AMDGPU_GFX_TCX_DST_FIFOA6, "TCX_DST_FIFOA6"},
+	{AMDGPU_GFX_TCX_DST_FIFOA7, "TCX_DST_FIFOA7"},
+	{AMDGPU_GFX_TCX_DST_FIFOB0, "TCX_DST_FIFOB0"},
+	{AMDGPU_GFX_TCX_DST_FIFOB1, "TCX_DST_FIFOB1"},
+	{AMDGPU_GFX_TCX_DST_FIFOB2, "TCX_DST_FIFOB2"},
+	{AMDGPU_GFX_TCX_DST_FIFOB3, "TCX_DST_FIFOB3"},
+	{AMDGPU_GFX_TCX_DST_FIFOB4, "TCX_DST_FIFOB4"},
+	{AMDGPU_GFX_TCX_DST_FIFOB5, "TCX_DST_FIFOB5"},
+	{AMDGPU_GFX_TCX_DST_FIFOB6, "TCX_DST_FIFOB6"},
+	{AMDGPU_GFX_TCX_DST_FIFOB7, "TCX_DST_FIFOB7"},
+	{AMDGPU_GFX_TCX_DST_FIFOD0, "TCX_DST_FIFOD0"},
+	{AMDGPU_GFX_TCX_DST_FIFOD1, "TCX_DST_FIFOD1"},
+	{AMDGPU_GFX_TCX_DST_FIFOD2, "TCX_DST_FIFOD2"},
+	{AMDGPU_GFX_TCX_DST_FIFOD3, "TCX_DST_FIFOD3"},
+	{AMDGPU_GFX_TCX_DST_FIFOD4, "TCX_DST_FIFOD4"},
+	{AMDGPU_GFX_TCX_DST_FIFOD5, "TCX_DST_FIFOD5"},
+	{AMDGPU_GFX_TCX_DST_FIFOD6, "TCX_DST_FIFOD6"},
+	{AMDGPU_GFX_TCX_DST_FIFOD7, "TCX_DST_FIFOD7"},
+	{AMDGPU_GFX_TCX_DST_FIFO_ACKB0, "TCX_DST_FIFO_ACKB0"},
+	{AMDGPU_GFX_TCX_DST_FIFO_ACKB1, "TCX_DST_FIFO_ACKB1"},
+	{AMDGPU_GFX_TCX_DST_FIFO_ACKB2, "TCX_DST_FIFO_ACKB2"},
+	{AMDGPU_GFX_TCX_DST_FIFO_ACKB3, "TCX_DST_FIFO_ACKB3"},
+	{AMDGPU_GFX_TCX_DST_FIFO_ACKB4, "TCX_DST_FIFO_ACKB4"},
+	{AMDGPU_GFX_TCX_DST_FIFO_ACKB5, "TCX_DST_FIFO_ACKB5"},
+	{AMDGPU_GFX_TCX_DST_FIFO_ACKB6, "TCX_DST_FIFO_ACKB6"},
+	{AMDGPU_GFX_TCX_DST_FIFO_ACKB7, "TCX_DST_FIFO_ACKB7"},
+	{AMDGPU_GFX_TCX_DST_FIFO_ACKD0, "TCX_DST_FIFO_ACKD0"},
+	{AMDGPU_GFX_TCX_DST_FIFO_ACKD1, "TCX_DST_FIFO_ACKD1"},
+	{AMDGPU_GFX_TCX_DST_FIFO_ACKD2, "TCX_DST_FIFO_ACKD2"},
+	{AMDGPU_GFX_TCX_DST_FIFO_ACKD3, "TCX_DST_FIFO_ACKD3"},
+	{AMDGPU_GFX_TCX_DST_FIFO_ACKD4, "TCX_DST_FIFO_ACKD4"},
+	{AMDGPU_GFX_TCX_DST_FIFO_ACKD5, "TCX_DST_FIFO_ACKD5"},
+	{AMDGPU_GFX_TCX_DST_FIFO_ACKD6, "TCX_DST_FIFO_ACKD6"},
+	{AMDGPU_GFX_TCX_DST_FIFO_ACKD7, "TCX_DST_FIFO_ACKD7"},
+};
+
+static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_atc_l2_mem_list[] = {
+	{AMDGPU_GFX_ATC_L2_MEM, "ATC_L2_MEM"},
+};
+
+static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_utcl2_mem_list[] = {
+	{AMDGPU_GFX_UTCL2_MEM, "UTCL2_MEM"},
+};
+
+static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_vml2_mem_list[] = {
+	{AMDGPU_GFX_VML2_MEM, "VML2_MEM"},
+};
+
+static const struct amdgpu_ras_memory_id_entry gfx_v9_4_3_ras_vml2_walker_mem_list[] = {
+	{AMDGPU_GFX_VML2_WALKER_MEM, "VML2_WALKER_MEM"},
+};
+
+static const struct amdgpu_gfx_ras_mem_id_entry gfx_v9_4_3_ras_mem_list_array[AMDGPU_GFX_MEM_TYPE_NUM] = {
+	AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_cp_mem_list)
+	AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_gcea_mem_list)
+	AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_gc_cane_mem_list)
+	AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_gcutcl2_mem_list)
+	AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_gds_mem_list)
+	AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_lds_mem_list)
+	AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_rlc_mem_list)
+	AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_sp_mem_list)
+	AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_spi_mem_list)
+	AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_sqc_mem_list)
+	AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_sq_mem_list)
+	AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_ta_mem_list)
+	AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_tcc_mem_list)
+	AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_tca_mem_list)
+	AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_tci_mem_list)
+	AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_tcp_mem_list)
+	AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_td_mem_list)
+	AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_tcx_mem_list)
+	AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_atc_l2_mem_list)
+	AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_utcl2_mem_list)
+	AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_vml2_mem_list)
+	AMDGPU_GFX_MEMID_ENT(gfx_v9_4_3_ras_vml2_walker_mem_list)
+};
+
+static const struct amdgpu_gfx_ras_reg_entry gfx_v9_4_3_ce_reg_list[] = {
+	{{AMDGPU_RAS_REG_ENTRY(GC, 0, regRLC_CE_ERR_STATUS_LOW, regRLC_CE_ERR_STATUS_HIGH),
+	    1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "RLC"},
+	    AMDGPU_GFX_RLC_MEM, 1},
+	{{AMDGPU_RAS_REG_ENTRY(GC, 0, regCPC_CE_ERR_STATUS_LO, regCPC_CE_ERR_STATUS_HI),
+	    1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "CPC"},
+	    AMDGPU_GFX_CP_MEM, 1},
+	{{AMDGPU_RAS_REG_ENTRY(GC, 0, regCPF_CE_ERR_STATUS_LO, regCPF_CE_ERR_STATUS_HI),
+	    1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "CPF"},
+	    AMDGPU_GFX_CP_MEM, 1},
+	{{AMDGPU_RAS_REG_ENTRY(GC, 0, regCPG_CE_ERR_STATUS_LO, regCPG_CE_ERR_STATUS_HI),
+	    1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "CPG"},
+	    AMDGPU_GFX_CP_MEM, 1},
+	{{AMDGPU_RAS_REG_ENTRY(GC, 0, regGDS_CE_ERR_STATUS_LO, regGDS_CE_ERR_STATUS_HI),
+	    1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "GDS"},
+	    AMDGPU_GFX_GDS_MEM, 1},
+	{{AMDGPU_RAS_REG_ENTRY(GC, 0, regGC_CANE_CE_ERR_STATUS_LO, regGC_CANE_CE_ERR_STATUS_HI),
+	    1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "CANE"},
+	    AMDGPU_GFX_GC_CANE_MEM, 1},
+	{{AMDGPU_RAS_REG_ENTRY(GC, 0, regSPI_CE_ERR_STATUS_LO, regSPI_CE_ERR_STATUS_HI),
+	    1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "SPI"},
+	    AMDGPU_GFX_SPI_MEM, 1},
+	{{AMDGPU_RAS_REG_ENTRY(GC, 0, regSP0_CE_ERR_STATUS_LO, regSP0_CE_ERR_STATUS_HI),
+	    10, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "SP0"},
+	    AMDGPU_GFX_SP_MEM, 4},
+	{{AMDGPU_RAS_REG_ENTRY(GC, 0, regSP1_CE_ERR_STATUS_LO, regSP1_CE_ERR_STATUS_HI),
+	    10, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "SP1"},
+	    AMDGPU_GFX_SP_MEM, 4},
+	{{AMDGPU_RAS_REG_ENTRY(GC, 0, regSQ_CE_ERR_STATUS_LO, regSQ_CE_ERR_STATUS_HI),
+	    10, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "SQ"},
+	    AMDGPU_GFX_SQ_MEM, 4},
+	{{AMDGPU_RAS_REG_ENTRY(GC, 0, regSQC_CE_EDC_LO, regSQC_CE_EDC_HI),
+	    5, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "SQC"},
+	    AMDGPU_GFX_SQC_MEM, 4},
+	{{AMDGPU_RAS_REG_ENTRY(GC, 0, regTCX_CE_ERR_STATUS_LO, regTCX_CE_ERR_STATUS_HI),
+	    2, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "TCX"},
+	    AMDGPU_GFX_TCX_MEM, 1},
+	{{AMDGPU_RAS_REG_ENTRY(GC, 0, regTCC_CE_ERR_STATUS_LO, regTCC_CE_ERR_STATUS_HI),
+	    16, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "TCC"},
+	    AMDGPU_GFX_TCC_MEM, 1},
+	{{AMDGPU_RAS_REG_ENTRY(GC, 0, regTA_CE_EDC_LO, regTA_CE_EDC_HI),
+	    10, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "TA"},
+	    AMDGPU_GFX_TA_MEM, 4},
+	{{AMDGPU_RAS_REG_ENTRY(GC, 0, regTCI_CE_EDC_LO_REG, regTCI_CE_EDC_HI_REG),
+	    27, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "TCI"},
+	    AMDGPU_GFX_TCI_MEM, 1},
+	{{AMDGPU_RAS_REG_ENTRY(GC, 0, regTCP_CE_EDC_LO_REG, regTCP_CE_EDC_HI_REG),
+	    10, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "TCP"},
+	    AMDGPU_GFX_TCP_MEM, 4},
+	{{AMDGPU_RAS_REG_ENTRY(GC, 0, regTD_CE_EDC_LO, regTD_CE_EDC_HI),
+	    10, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "TD"},
+	    AMDGPU_GFX_TD_MEM, 4},
+	{{AMDGPU_RAS_REG_ENTRY(GC, 0, regGCEA_CE_ERR_STATUS_LO, regGCEA_CE_ERR_STATUS_HI),
+	    16, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "GCEA"},
+	    AMDGPU_GFX_GCEA_MEM, 1},
+	{{AMDGPU_RAS_REG_ENTRY(GC, 0, regLDS_CE_ERR_STATUS_LO, regLDS_CE_ERR_STATUS_HI),
+	    10, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "LDS"},
+	    AMDGPU_GFX_LDS_MEM, 4},
+};
+
+static const struct amdgpu_gfx_ras_reg_entry gfx_v9_4_3_ue_reg_list[] = {
+	{{AMDGPU_RAS_REG_ENTRY(GC, 0, regRLC_UE_ERR_STATUS_LOW, regRLC_UE_ERR_STATUS_HIGH),
+	    1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "RLC"},
+	    AMDGPU_GFX_RLC_MEM, 1},
+	{{AMDGPU_RAS_REG_ENTRY(GC, 0, regCPC_UE_ERR_STATUS_LO, regCPC_UE_ERR_STATUS_HI),
+	    1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "CPC"},
+	    AMDGPU_GFX_CP_MEM, 1},
+	{{AMDGPU_RAS_REG_ENTRY(GC, 0, regCPF_UE_ERR_STATUS_LO, regCPF_UE_ERR_STATUS_HI),
+	    1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "CPF"},
+	    AMDGPU_GFX_CP_MEM, 1},
+	{{AMDGPU_RAS_REG_ENTRY(GC, 0, regCPG_UE_ERR_STATUS_LO, regCPG_UE_ERR_STATUS_HI),
+	    1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "CPG"},
+	    AMDGPU_GFX_CP_MEM, 1},
+	{{AMDGPU_RAS_REG_ENTRY(GC, 0, regGDS_UE_ERR_STATUS_LO, regGDS_UE_ERR_STATUS_HI),
+	    1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "GDS"},
+	    AMDGPU_GFX_GDS_MEM, 1},
+	{{AMDGPU_RAS_REG_ENTRY(GC, 0, regGC_CANE_UE_ERR_STATUS_LO, regGC_CANE_UE_ERR_STATUS_HI),
+	    1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "CANE"},
+	    AMDGPU_GFX_GC_CANE_MEM, 1},
+	{{AMDGPU_RAS_REG_ENTRY(GC, 0, regSPI_UE_ERR_STATUS_LO, regSPI_UE_ERR_STATUS_HI),
+	    1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "SPI"},
+	    AMDGPU_GFX_SPI_MEM, 1},
+	{{AMDGPU_RAS_REG_ENTRY(GC, 0, regSP0_UE_ERR_STATUS_LO, regSP0_UE_ERR_STATUS_HI),
+	    10, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "SP0"},
+	    AMDGPU_GFX_SP_MEM, 4},
+	{{AMDGPU_RAS_REG_ENTRY(GC, 0, regSP1_UE_ERR_STATUS_LO, regSP1_UE_ERR_STATUS_HI),
+	    10, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "SP1"},
+	    AMDGPU_GFX_SP_MEM, 4},
+	{{AMDGPU_RAS_REG_ENTRY(GC, 0, regSQ_UE_ERR_STATUS_LO, regSQ_UE_ERR_STATUS_HI),
+	    10, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "SQ"},
+	    AMDGPU_GFX_SQ_MEM, 4},
+	{{AMDGPU_RAS_REG_ENTRY(GC, 0, regSQC_UE_EDC_LO, regSQC_UE_EDC_HI),
+	    5, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "SQC"},
+	    AMDGPU_GFX_SQC_MEM, 4},
+	{{AMDGPU_RAS_REG_ENTRY(GC, 0, regTCX_UE_ERR_STATUS_LO, regTCX_UE_ERR_STATUS_HI),
+	    2, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "TCX"},
+	    AMDGPU_GFX_TCX_MEM, 1},
+	{{AMDGPU_RAS_REG_ENTRY(GC, 0, regTCC_UE_ERR_STATUS_LO, regTCC_UE_ERR_STATUS_HI),
+	    16, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "TCC"},
+	    AMDGPU_GFX_TCC_MEM, 1},
+	{{AMDGPU_RAS_REG_ENTRY(GC, 0, regTA_UE_EDC_LO, regTA_UE_EDC_HI),
+	    10, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "TA"},
+	    AMDGPU_GFX_TA_MEM, 4},
+	{{AMDGPU_RAS_REG_ENTRY(GC, 0, regTCI_UE_EDC_LO_REG, regTCI_UE_EDC_HI_REG),
+	    27, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "TCI"},
+	    AMDGPU_GFX_TCI_MEM, 1},
+	{{AMDGPU_RAS_REG_ENTRY(GC, 0, regTCP_UE_EDC_LO_REG, regTCP_UE_EDC_HI_REG),
+	    10, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "TCP"},
+	    AMDGPU_GFX_TCP_MEM, 4},
+	{{AMDGPU_RAS_REG_ENTRY(GC, 0, regTD_UE_EDC_LO, regTD_UE_EDC_HI),
+	    10, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "TD"},
+	    AMDGPU_GFX_TD_MEM, 4},
+	{{AMDGPU_RAS_REG_ENTRY(GC, 0, regTCA_UE_ERR_STATUS_LO, regTCA_UE_ERR_STATUS_HI),
+	    2, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "TCA"},
+	    AMDGPU_GFX_TCA_MEM, 1},
+	{{AMDGPU_RAS_REG_ENTRY(GC, 0, regGCEA_UE_ERR_STATUS_LO, regGCEA_UE_ERR_STATUS_HI),
+	    16, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "GCEA"},
+	    AMDGPU_GFX_GCEA_MEM, 1},
+	{{AMDGPU_RAS_REG_ENTRY(GC, 0, regLDS_UE_ERR_STATUS_LO, regLDS_UE_ERR_STATUS_HI),
+	    10, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "LDS"},
+	    AMDGPU_GFX_LDS_MEM, 4},
+};
+
+static void gfx_v9_4_3_inst_query_ras_err_count(struct amdgpu_device *adev,
+					void *ras_error_status, int xcc_id)
+{
+	struct ras_err_data *err_data = (struct ras_err_data *)ras_error_status;
+	unsigned long ce_count = 0, ue_count = 0;
+	uint32_t i, j, k;
+
+	/* NOTE: convert xcc_id to physical XCD ID (XCD0 or XCD1) */
+	struct amdgpu_smuio_mcm_config_info mcm_info = {
+		.socket_id = adev->smuio.funcs->get_socket_id(adev),
+		.die_id = xcc_id & 0x01 ? 1 : 0,
+	};
+
+	mutex_lock(&adev->grbm_idx_mutex);
+
+	for (i = 0; i < ARRAY_SIZE(gfx_v9_4_3_ce_reg_list); i++) {
+		for (j = 0; j < gfx_v9_4_3_ce_reg_list[i].se_num; j++) {
+			for (k = 0; k < gfx_v9_4_3_ce_reg_list[i].reg_entry.reg_inst; k++) {
+				/* no need to select if instance number is 1 */
+				if (gfx_v9_4_3_ce_reg_list[i].se_num > 1 ||
+				    gfx_v9_4_3_ce_reg_list[i].reg_entry.reg_inst > 1)
+					gfx_v9_4_3_xcc_select_se_sh(adev, j, 0, k, xcc_id);
+
+				amdgpu_ras_inst_query_ras_error_count(adev,
+					&(gfx_v9_4_3_ce_reg_list[i].reg_entry),
+					1,
+					gfx_v9_4_3_ras_mem_list_array[gfx_v9_4_3_ce_reg_list[i].mem_id_type].mem_id_ent,
+					gfx_v9_4_3_ras_mem_list_array[gfx_v9_4_3_ce_reg_list[i].mem_id_type].size,
+					GET_INST(GC, xcc_id),
+					AMDGPU_RAS_ERROR__SINGLE_CORRECTABLE,
+					&ce_count);
+
+				amdgpu_ras_inst_query_ras_error_count(adev,
+					&(gfx_v9_4_3_ue_reg_list[i].reg_entry),
+					1,
+					gfx_v9_4_3_ras_mem_list_array[gfx_v9_4_3_ue_reg_list[i].mem_id_type].mem_id_ent,
+					gfx_v9_4_3_ras_mem_list_array[gfx_v9_4_3_ue_reg_list[i].mem_id_type].size,
+					GET_INST(GC, xcc_id),
+					AMDGPU_RAS_ERROR__MULTI_UNCORRECTABLE,
+					&ue_count);
+			}
+		}
+	}
+
+	/* handle extra register entries of UE */
+	for (; i < ARRAY_SIZE(gfx_v9_4_3_ue_reg_list); i++) {
+		for (j = 0; j < gfx_v9_4_3_ue_reg_list[i].se_num; j++) {
+			for (k = 0; k < gfx_v9_4_3_ue_reg_list[i].reg_entry.reg_inst; k++) {
+				/* no need to select if instance number is 1 */
+				if (gfx_v9_4_3_ue_reg_list[i].se_num > 1 ||
+					gfx_v9_4_3_ue_reg_list[i].reg_entry.reg_inst > 1)
+					gfx_v9_4_3_xcc_select_se_sh(adev, j, 0, k, xcc_id);
+
+				amdgpu_ras_inst_query_ras_error_count(adev,
+					&(gfx_v9_4_3_ue_reg_list[i].reg_entry),
+					1,
+					gfx_v9_4_3_ras_mem_list_array[gfx_v9_4_3_ue_reg_list[i].mem_id_type].mem_id_ent,
+					gfx_v9_4_3_ras_mem_list_array[gfx_v9_4_3_ue_reg_list[i].mem_id_type].size,
+					GET_INST(GC, xcc_id),
+					AMDGPU_RAS_ERROR__MULTI_UNCORRECTABLE,
+					&ue_count);
+			}
+		}
+	}
+
+	gfx_v9_4_3_xcc_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff,
+			xcc_id);
+	mutex_unlock(&adev->grbm_idx_mutex);
+
+	/* the caller should make sure initialize value of
+	 * err_data->ue_count and err_data->ce_count
+	 */
+	amdgpu_ras_error_statistic_ue_count(err_data, &mcm_info, ue_count);
+	amdgpu_ras_error_statistic_ce_count(err_data, &mcm_info, ce_count);
+}
+
+static void gfx_v9_4_3_inst_reset_ras_err_count(struct amdgpu_device *adev,
+					void *ras_error_status, int xcc_id)
+{
+	uint32_t i, j, k;
+
+	mutex_lock(&adev->grbm_idx_mutex);
+
+	for (i = 0; i < ARRAY_SIZE(gfx_v9_4_3_ce_reg_list); i++) {
+		for (j = 0; j < gfx_v9_4_3_ce_reg_list[i].se_num; j++) {
+			for (k = 0; k < gfx_v9_4_3_ce_reg_list[i].reg_entry.reg_inst; k++) {
+				/* no need to select if instance number is 1 */
+				if (gfx_v9_4_3_ce_reg_list[i].se_num > 1 ||
+				    gfx_v9_4_3_ce_reg_list[i].reg_entry.reg_inst > 1)
+					gfx_v9_4_3_xcc_select_se_sh(adev, j, 0, k, xcc_id);
+
+				amdgpu_ras_inst_reset_ras_error_count(adev,
+					&(gfx_v9_4_3_ce_reg_list[i].reg_entry),
+					1,
+					GET_INST(GC, xcc_id));
+
+				amdgpu_ras_inst_reset_ras_error_count(adev,
+					&(gfx_v9_4_3_ue_reg_list[i].reg_entry),
+					1,
+					GET_INST(GC, xcc_id));
+			}
+		}
+	}
+
+	/* handle extra register entries of UE */
+	for (; i < ARRAY_SIZE(gfx_v9_4_3_ue_reg_list); i++) {
+		for (j = 0; j < gfx_v9_4_3_ue_reg_list[i].se_num; j++) {
+			for (k = 0; k < gfx_v9_4_3_ue_reg_list[i].reg_entry.reg_inst; k++) {
+				/* no need to select if instance number is 1 */
+				if (gfx_v9_4_3_ue_reg_list[i].se_num > 1 ||
+					gfx_v9_4_3_ue_reg_list[i].reg_entry.reg_inst > 1)
+					gfx_v9_4_3_xcc_select_se_sh(adev, j, 0, k, xcc_id);
+
+				amdgpu_ras_inst_reset_ras_error_count(adev,
+					&(gfx_v9_4_3_ue_reg_list[i].reg_entry),
+					1,
+					GET_INST(GC, xcc_id));
+			}
+		}
+	}
+
+	gfx_v9_4_3_xcc_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff,
+			xcc_id);
+	mutex_unlock(&adev->grbm_idx_mutex);
+}
+
+static void gfx_v9_4_3_inst_enable_watchdog_timer(struct amdgpu_device *adev,
+					void *ras_error_status, int xcc_id)
+{
+	uint32_t i;
+	uint32_t data;
+
+	if (amdgpu_sriov_vf(adev))
+		return;
+
+	data = RREG32_SOC15(GC, GET_INST(GC, 0), regSQ_TIMEOUT_CONFIG);
+	data = REG_SET_FIELD(data, SQ_TIMEOUT_CONFIG, TIMEOUT_FATAL_DISABLE,
+			     amdgpu_watchdog_timer.timeout_fatal_disable ? 1 : 0);
+
+	if (amdgpu_watchdog_timer.timeout_fatal_disable &&
+	    (amdgpu_watchdog_timer.period < 1 ||
+	     amdgpu_watchdog_timer.period > 0x23)) {
+		dev_warn(adev->dev, "Watchdog period range is 1 to 0x23\n");
+		amdgpu_watchdog_timer.period = 0x23;
+	}
+	data = REG_SET_FIELD(data, SQ_TIMEOUT_CONFIG, PERIOD_SEL,
+			     amdgpu_watchdog_timer.period);
+
+	mutex_lock(&adev->grbm_idx_mutex);
+	for (i = 0; i < adev->gfx.config.max_shader_engines; i++) {
+		gfx_v9_4_3_xcc_select_se_sh(adev, i, 0xffffffff, 0xffffffff, xcc_id);
+		WREG32_SOC15(GC, GET_INST(GC, xcc_id), regSQ_TIMEOUT_CONFIG, data);
+	}
+	gfx_v9_4_3_xcc_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff,
+			xcc_id);
+	mutex_unlock(&adev->grbm_idx_mutex);
+}
+
+static void gfx_v9_4_3_query_ras_error_count(struct amdgpu_device *adev,
+					void *ras_error_status)
+{
+	amdgpu_gfx_ras_error_func(adev, ras_error_status,
+			gfx_v9_4_3_inst_query_ras_err_count);
+}
+
+static void gfx_v9_4_3_reset_ras_error_count(struct amdgpu_device *adev)
+{
+	amdgpu_gfx_ras_error_func(adev, NULL, gfx_v9_4_3_inst_reset_ras_err_count);
+}
+
+static void gfx_v9_4_3_enable_watchdog_timer(struct amdgpu_device *adev)
+{
+	amdgpu_gfx_ras_error_func(adev, NULL, gfx_v9_4_3_inst_enable_watchdog_timer);
+}
+
+static void gfx_v9_4_3_ring_insert_nop(struct amdgpu_ring *ring, uint32_t num_nop)
+{
+	/* Header itself is a NOP packet */
+	if (num_nop == 1) {
+		amdgpu_ring_write(ring, ring->funcs->nop);
+		return;
+	}
+
+	/* Max HW optimization till 0x3ffe, followed by remaining one NOP at a time*/
+	amdgpu_ring_write(ring, PACKET3(PACKET3_NOP, min(num_nop - 2, 0x3ffe)));
+
+	/* Header is at index 0, followed by num_nops - 1 NOP packet's */
+	amdgpu_ring_insert_nop(ring, num_nop - 1);
+}
+
+static void gfx_v9_4_3_ip_print(struct amdgpu_ip_block *ip_block, struct drm_printer *p)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	uint32_t i, j, k;
+	uint32_t xcc_id, xcc_offset, inst_offset;
+	uint32_t num_xcc, reg, num_inst;
+	uint32_t reg_count = ARRAY_SIZE(gc_reg_list_9_4_3);
+
+	if (!adev->gfx.ip_dump_core)
+		return;
+
+	num_xcc = NUM_XCC(adev->gfx.xcc_mask);
+	drm_printf(p, "Number of Instances:%d\n", num_xcc);
+	for (xcc_id = 0; xcc_id < num_xcc; xcc_id++) {
+		xcc_offset = xcc_id * reg_count;
+		drm_printf(p, "\nInstance id:%d\n", xcc_id);
+		for (i = 0; i < reg_count; i++)
+			drm_printf(p, "%-50s \t 0x%08x\n",
+				   gc_reg_list_9_4_3[i].reg_name,
+				   adev->gfx.ip_dump_core[xcc_offset + i]);
+	}
+
+	/* print compute queue registers for all instances */
+	if (!adev->gfx.ip_dump_compute_queues)
+		return;
+
+	num_inst = adev->gfx.mec.num_mec * adev->gfx.mec.num_pipe_per_mec *
+		adev->gfx.mec.num_queue_per_pipe;
+
+	reg_count = ARRAY_SIZE(gc_cp_reg_list_9_4_3);
+	drm_printf(p, "\nnum_xcc: %d num_mec: %d num_pipe: %d num_queue: %d\n",
+		   num_xcc,
+		   adev->gfx.mec.num_mec,
+		   adev->gfx.mec.num_pipe_per_mec,
+		   adev->gfx.mec.num_queue_per_pipe);
+
+	for (xcc_id = 0; xcc_id < num_xcc; xcc_id++) {
+		xcc_offset = xcc_id * reg_count * num_inst;
+		inst_offset = 0;
+		for (i = 0; i < adev->gfx.mec.num_mec; i++) {
+			for (j = 0; j < adev->gfx.mec.num_pipe_per_mec; j++) {
+				for (k = 0; k < adev->gfx.mec.num_queue_per_pipe; k++) {
+					drm_printf(p,
+						   "\nxcc:%d mec:%d, pipe:%d, queue:%d\n",
+						    xcc_id, i, j, k);
+					for (reg = 0; reg < reg_count; reg++) {
+						if (i && gc_cp_reg_list_9_4_3[reg].reg_offset ==
+						    regCP_MEC_ME1_HEADER_DUMP)
+							drm_printf(p,
+								   "%-50s \t 0x%08x\n",
+								   "regCP_MEC_ME2_HEADER_DUMP",
+								   adev->gfx.ip_dump_compute_queues
+								   [xcc_offset + inst_offset +
+								    reg]);
+						else
+							drm_printf(p,
+								   "%-50s \t 0x%08x\n",
+								   gc_cp_reg_list_9_4_3[reg].reg_name,
+								   adev->gfx.ip_dump_compute_queues
+								   [xcc_offset + inst_offset +
+								    reg]);
+					}
+					inst_offset += reg_count;
+				}
+			}
+		}
+	}
+}
+
+static void gfx_v9_4_3_ip_dump(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	uint32_t i, j, k;
+	uint32_t num_xcc, reg, num_inst;
+	uint32_t xcc_id, xcc_offset, inst_offset;
+	uint32_t reg_count = ARRAY_SIZE(gc_reg_list_9_4_3);
+
+	if (!adev->gfx.ip_dump_core)
+		return;
+
+	num_xcc = NUM_XCC(adev->gfx.xcc_mask);
+
+	for (xcc_id = 0; xcc_id < num_xcc; xcc_id++) {
+		xcc_offset = xcc_id * reg_count;
+		for (i = 0; i < reg_count; i++)
+			adev->gfx.ip_dump_core[xcc_offset + i] =
+				RREG32(SOC15_REG_ENTRY_OFFSET_INST(gc_reg_list_9_4_3[i],
+								   GET_INST(GC, xcc_id)));
+	}
+
+	/* dump compute queue registers for all instances */
+	if (!adev->gfx.ip_dump_compute_queues)
+		return;
+
+	num_inst = adev->gfx.mec.num_mec * adev->gfx.mec.num_pipe_per_mec *
+		adev->gfx.mec.num_queue_per_pipe;
+	reg_count = ARRAY_SIZE(gc_cp_reg_list_9_4_3);
+	mutex_lock(&adev->srbm_mutex);
+	for (xcc_id = 0; xcc_id < num_xcc; xcc_id++) {
+		xcc_offset = xcc_id * reg_count * num_inst;
+		inst_offset = 0;
+		for (i = 0; i < adev->gfx.mec.num_mec; i++) {
+			for (j = 0; j < adev->gfx.mec.num_pipe_per_mec; j++) {
+				for (k = 0; k < adev->gfx.mec.num_queue_per_pipe; k++) {
+					/* ME0 is for GFX so start from 1 for CP */
+					soc15_grbm_select(adev, 1 + i, j, k, 0,
+							  GET_INST(GC, xcc_id));
+
+					for (reg = 0; reg < reg_count; reg++) {
+						if (i && gc_cp_reg_list_9_4_3[reg].reg_offset ==
+						    regCP_MEC_ME1_HEADER_DUMP)
+							adev->gfx.ip_dump_compute_queues
+								[xcc_offset +
+								 inst_offset + reg] =
+								RREG32(SOC15_REG_OFFSET(GC, GET_INST(GC, xcc_id),
+											regCP_MEC_ME2_HEADER_DUMP));
+						else
+							adev->gfx.ip_dump_compute_queues
+								[xcc_offset +
+								 inst_offset + reg] =
+								RREG32(SOC15_REG_ENTRY_OFFSET_INST(
+									       gc_cp_reg_list_9_4_3[reg],
+									       GET_INST(GC, xcc_id)));
+					}
+					inst_offset += reg_count;
+				}
+			}
+		}
+	}
+	soc15_grbm_select(adev, 0, 0, 0, 0, 0);
+	mutex_unlock(&adev->srbm_mutex);
+}
+
+static void gfx_v9_4_3_ring_emit_cleaner_shader(struct amdgpu_ring *ring)
+{
+	/* Emit the cleaner shader */
+	amdgpu_ring_write(ring, PACKET3(PACKET3_RUN_CLEANER_SHADER, 0));
+	amdgpu_ring_write(ring, 0);  /* RESERVED field, programmed to zero */
+}
+
+static const struct amd_ip_funcs gfx_v9_4_3_ip_funcs = {
+	.name = "gfx_v9_4_3",
+	.early_init = gfx_v9_4_3_early_init,
+	.late_init = gfx_v9_4_3_late_init,
+	.sw_init = gfx_v9_4_3_sw_init,
+	.sw_fini = gfx_v9_4_3_sw_fini,
+	.hw_init = gfx_v9_4_3_hw_init,
+	.hw_fini = gfx_v9_4_3_hw_fini,
+	.suspend = gfx_v9_4_3_suspend,
+	.resume = gfx_v9_4_3_resume,
+	.is_idle = gfx_v9_4_3_is_idle,
+	.wait_for_idle = gfx_v9_4_3_wait_for_idle,
+	.soft_reset = gfx_v9_4_3_soft_reset,
+	.set_clockgating_state = gfx_v9_4_3_set_clockgating_state,
+	.set_powergating_state = gfx_v9_4_3_set_powergating_state,
+	.get_clockgating_state = gfx_v9_4_3_get_clockgating_state,
+	.dump_ip_state = gfx_v9_4_3_ip_dump,
+	.print_ip_state = gfx_v9_4_3_ip_print,
+};
+
+static const struct amdgpu_ring_funcs gfx_v9_4_3_ring_funcs_compute = {
+	.type = AMDGPU_RING_TYPE_COMPUTE,
+	.align_mask = 0xff,
+	.nop = PACKET3(PACKET3_NOP, 0x3FFF),
+	.support_64bit_ptrs = true,
+	.get_rptr = gfx_v9_4_3_ring_get_rptr_compute,
+	.get_wptr = gfx_v9_4_3_ring_get_wptr_compute,
+	.set_wptr = gfx_v9_4_3_ring_set_wptr_compute,
+	.emit_frame_size =
+		20 + /* gfx_v9_4_3_ring_emit_gds_switch */
+		7 + /* gfx_v9_4_3_ring_emit_hdp_flush */
+		5 + /* hdp invalidate */
+		7 + /* gfx_v9_4_3_ring_emit_pipeline_sync */
+		SOC15_FLUSH_GPU_TLB_NUM_WREG * 5 +
+		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 7 +
+		2 + /* gfx_v9_4_3_ring_emit_vm_flush */
+		8 + 8 + 8 + /* gfx_v9_4_3_ring_emit_fence x3 for user fence, vm fence */
+		7 + /* gfx_v9_4_3_emit_mem_sync */
+		5 + /* gfx_v9_4_3_emit_wave_limit for updating regSPI_WCL_PIPE_PERCENT_GFX register */
+		15 + /* for updating 3 regSPI_WCL_PIPE_PERCENT_CS registers */
+		2, /* gfx_v9_4_3_ring_emit_cleaner_shader */
+	.emit_ib_size =	7, /* gfx_v9_4_3_ring_emit_ib_compute */
+	.emit_ib = gfx_v9_4_3_ring_emit_ib_compute,
+	.emit_fence = gfx_v9_4_3_ring_emit_fence,
+	.emit_pipeline_sync = gfx_v9_4_3_ring_emit_pipeline_sync,
+	.emit_vm_flush = gfx_v9_4_3_ring_emit_vm_flush,
+	.emit_gds_switch = gfx_v9_4_3_ring_emit_gds_switch,
+	.emit_hdp_flush = gfx_v9_4_3_ring_emit_hdp_flush,
+	.test_ring = gfx_v9_4_3_ring_test_ring,
+	.test_ib = gfx_v9_4_3_ring_test_ib,
+	.insert_nop = gfx_v9_4_3_ring_insert_nop,
+	.pad_ib = amdgpu_ring_generic_pad_ib,
+	.emit_wreg = gfx_v9_4_3_ring_emit_wreg,
+	.emit_reg_wait = gfx_v9_4_3_ring_emit_reg_wait,
+	.emit_reg_write_reg_wait = gfx_v9_4_3_ring_emit_reg_write_reg_wait,
+	.soft_recovery = gfx_v9_4_3_ring_soft_recovery,
+	.emit_mem_sync = gfx_v9_4_3_emit_mem_sync,
+	.emit_wave_limit = gfx_v9_4_3_emit_wave_limit,
+	.reset = gfx_v9_4_3_reset_kcq,
+	.emit_cleaner_shader = gfx_v9_4_3_ring_emit_cleaner_shader,
+	.begin_use = amdgpu_gfx_enforce_isolation_ring_begin_use,
+	.end_use = amdgpu_gfx_enforce_isolation_ring_end_use,
+};
+
+static const struct amdgpu_ring_funcs gfx_v9_4_3_ring_funcs_kiq = {
+	.type = AMDGPU_RING_TYPE_KIQ,
+	.align_mask = 0xff,
+	.nop = PACKET3(PACKET3_NOP, 0x3FFF),
+	.support_64bit_ptrs = true,
+	.get_rptr = gfx_v9_4_3_ring_get_rptr_compute,
+	.get_wptr = gfx_v9_4_3_ring_get_wptr_compute,
+	.set_wptr = gfx_v9_4_3_ring_set_wptr_compute,
+	.emit_frame_size =
+		20 + /* gfx_v9_4_3_ring_emit_gds_switch */
+		7 + /* gfx_v9_4_3_ring_emit_hdp_flush */
+		5 + /* hdp invalidate */
+		7 + /* gfx_v9_4_3_ring_emit_pipeline_sync */
+		SOC15_FLUSH_GPU_TLB_NUM_WREG * 5 +
+		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 7 +
+		2 + /* gfx_v9_4_3_ring_emit_vm_flush */
+		8 + 8 + 8, /* gfx_v9_4_3_ring_emit_fence_kiq x3 for user fence, vm fence */
+	.emit_ib_size =	7, /* gfx_v9_4_3_ring_emit_ib_compute */
+	.emit_fence = gfx_v9_4_3_ring_emit_fence_kiq,
+	.test_ring = gfx_v9_4_3_ring_test_ring,
+	.insert_nop = amdgpu_ring_insert_nop,
+	.pad_ib = amdgpu_ring_generic_pad_ib,
+	.emit_rreg = gfx_v9_4_3_ring_emit_rreg,
+	.emit_wreg = gfx_v9_4_3_ring_emit_wreg,
+	.emit_reg_wait = gfx_v9_4_3_ring_emit_reg_wait,
+	.emit_reg_write_reg_wait = gfx_v9_4_3_ring_emit_reg_write_reg_wait,
+	.emit_hdp_flush = gfx_v9_4_3_ring_emit_hdp_flush,
+};
+
+static void gfx_v9_4_3_set_ring_funcs(struct amdgpu_device *adev)
+{
+	int i, j, num_xcc;
+
+	num_xcc = NUM_XCC(adev->gfx.xcc_mask);
+	for (i = 0; i < num_xcc; i++) {
+		adev->gfx.kiq[i].ring.funcs = &gfx_v9_4_3_ring_funcs_kiq;
+
+		for (j = 0; j < adev->gfx.num_compute_rings; j++)
+			adev->gfx.compute_ring[j + i * adev->gfx.num_compute_rings].funcs
+					= &gfx_v9_4_3_ring_funcs_compute;
+	}
+}
+
+static const struct amdgpu_irq_src_funcs gfx_v9_4_3_eop_irq_funcs = {
+	.set = gfx_v9_4_3_set_eop_interrupt_state,
+	.process = gfx_v9_4_3_eop_irq,
+};
+
+static const struct amdgpu_irq_src_funcs gfx_v9_4_3_priv_reg_irq_funcs = {
+	.set = gfx_v9_4_3_set_priv_reg_fault_state,
+	.process = gfx_v9_4_3_priv_reg_irq,
+};
+
+static const struct amdgpu_irq_src_funcs gfx_v9_4_3_bad_op_irq_funcs = {
+	.set = gfx_v9_4_3_set_bad_op_fault_state,
+	.process = gfx_v9_4_3_bad_op_irq,
+};
+
+static const struct amdgpu_irq_src_funcs gfx_v9_4_3_priv_inst_irq_funcs = {
+	.set = gfx_v9_4_3_set_priv_inst_fault_state,
+	.process = gfx_v9_4_3_priv_inst_irq,
+};
+
+static void gfx_v9_4_3_set_irq_funcs(struct amdgpu_device *adev)
+{
+	adev->gfx.eop_irq.num_types = AMDGPU_CP_IRQ_LAST;
+	adev->gfx.eop_irq.funcs = &gfx_v9_4_3_eop_irq_funcs;
+
+	adev->gfx.priv_reg_irq.num_types = 1;
+	adev->gfx.priv_reg_irq.funcs = &gfx_v9_4_3_priv_reg_irq_funcs;
+
+	adev->gfx.bad_op_irq.num_types = 1;
+	adev->gfx.bad_op_irq.funcs = &gfx_v9_4_3_bad_op_irq_funcs;
+
+	adev->gfx.priv_inst_irq.num_types = 1;
+	adev->gfx.priv_inst_irq.funcs = &gfx_v9_4_3_priv_inst_irq_funcs;
+}
+
+static void gfx_v9_4_3_set_rlc_funcs(struct amdgpu_device *adev)
+{
+	adev->gfx.rlc.funcs = &gfx_v9_4_3_rlc_funcs;
+}
+
+
+static void gfx_v9_4_3_set_gds_init(struct amdgpu_device *adev)
+{
+	/* 9.4.3 variants removed all the GDS internal memory,
+	 * only support GWS opcode in kernel, like barrier
+	 * semaphore.etc */
+
+	/* init asic gds info */
+	adev->gds.gds_size = 0;
+	adev->gds.gds_compute_max_wave_id = 0;
+	adev->gds.gws_size = 64;
+	adev->gds.oa_size = 16;
+}
+
+static void gfx_v9_4_3_set_user_cu_inactive_bitmap(struct amdgpu_device *adev,
+						 u32 bitmap, int xcc_id)
+{
+	u32 data;
+
+	if (!bitmap)
+		return;
+
+	data = bitmap << GC_USER_SHADER_ARRAY_CONFIG__INACTIVE_CUS__SHIFT;
+	data &= GC_USER_SHADER_ARRAY_CONFIG__INACTIVE_CUS_MASK;
+
+	WREG32_SOC15(GC, GET_INST(GC, xcc_id), regGC_USER_SHADER_ARRAY_CONFIG, data);
+}
+
+static u32 gfx_v9_4_3_get_cu_active_bitmap(struct amdgpu_device *adev, int xcc_id)
+{
+	u32 data, mask;
+
+	data = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regCC_GC_SHADER_ARRAY_CONFIG);
+	data |= RREG32_SOC15(GC, GET_INST(GC, xcc_id), regGC_USER_SHADER_ARRAY_CONFIG);
+
+	data &= CC_GC_SHADER_ARRAY_CONFIG__INACTIVE_CUS_MASK;
+	data >>= CC_GC_SHADER_ARRAY_CONFIG__INACTIVE_CUS__SHIFT;
+
+	mask = amdgpu_gfx_create_bitmask(adev->gfx.config.max_cu_per_sh);
+
+	return (~data) & mask;
+}
+
+static int gfx_v9_4_3_get_cu_info(struct amdgpu_device *adev,
+				 struct amdgpu_cu_info *cu_info)
+{
+	int i, j, k, prev_counter, counter, xcc_id, active_cu_number = 0;
+	u32 mask, bitmap, ao_bitmap, ao_cu_mask = 0, tmp;
+	unsigned disable_masks[4 * 4];
+	bool is_symmetric_cus;
+
+	if (!adev || !cu_info)
+		return -EINVAL;
+
+	/*
+	 * 16 comes from bitmap array size 4*4, and it can cover all gfx9 ASICs
+	 */
+	if (adev->gfx.config.max_shader_engines *
+		adev->gfx.config.max_sh_per_se > 16)
+		return -EINVAL;
+
+	amdgpu_gfx_parse_disable_cu(disable_masks,
+				    adev->gfx.config.max_shader_engines,
+				    adev->gfx.config.max_sh_per_se);
+
+	mutex_lock(&adev->grbm_idx_mutex);
+	for (xcc_id = 0; xcc_id < NUM_XCC(adev->gfx.xcc_mask); xcc_id++) {
+		is_symmetric_cus = true;
+		for (i = 0; i < adev->gfx.config.max_shader_engines; i++) {
+			for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) {
+				mask = 1;
+				ao_bitmap = 0;
+				counter = 0;
+				gfx_v9_4_3_xcc_select_se_sh(adev, i, j, 0xffffffff, xcc_id);
+				gfx_v9_4_3_set_user_cu_inactive_bitmap(
+					adev,
+					disable_masks[i * adev->gfx.config.max_sh_per_se + j],
+					xcc_id);
+				bitmap = gfx_v9_4_3_get_cu_active_bitmap(adev, xcc_id);
+
+				cu_info->bitmap[xcc_id][i][j] = bitmap;
+
+				for (k = 0; k < adev->gfx.config.max_cu_per_sh; k++) {
+					if (bitmap & mask) {
+						if (counter < adev->gfx.config.max_cu_per_sh)
+							ao_bitmap |= mask;
+						counter++;
+					}
+					mask <<= 1;
+				}
+				active_cu_number += counter;
+				if (i < 2 && j < 2)
+					ao_cu_mask |= (ao_bitmap << (i * 16 + j * 8));
+				cu_info->ao_cu_bitmap[i][j] = ao_bitmap;
+			}
+			if (i && is_symmetric_cus && prev_counter != counter)
+				is_symmetric_cus = false;
+			prev_counter = counter;
+		}
+		if (is_symmetric_cus) {
+			tmp = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_CPC_DEBUG);
+			tmp = REG_SET_FIELD(tmp, CP_CPC_DEBUG, CPC_HARVESTING_RELAUNCH_DISABLE, 1);
+			tmp = REG_SET_FIELD(tmp, CP_CPC_DEBUG, CPC_HARVESTING_DISPATCH_DISABLE, 1);
+			WREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_CPC_DEBUG, tmp);
+		}
+		gfx_v9_4_3_xcc_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff,
+					    xcc_id);
+	}
+	mutex_unlock(&adev->grbm_idx_mutex);
+
+	cu_info->number = active_cu_number;
+	cu_info->ao_cu_mask = ao_cu_mask;
+	cu_info->simd_per_cu = NUM_SIMD_PER_CU;
+
+	return 0;
+}
+
+const struct amdgpu_ip_block_version gfx_v9_4_3_ip_block = {
+	.type = AMD_IP_BLOCK_TYPE_GFX,
+	.major = 9,
+	.minor = 4,
+	.rev = 3,
+	.funcs = &gfx_v9_4_3_ip_funcs,
+};
+
+static int gfx_v9_4_3_xcp_resume(void *handle, uint32_t inst_mask)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	uint32_t tmp_mask;
+	int i, r;
+
+	/* TODO : Initialize golden regs */
+	/* gfx_v9_4_3_init_golden_registers(adev); */
+
+	tmp_mask = inst_mask;
+	for_each_inst(i, tmp_mask)
+		gfx_v9_4_3_xcc_constants_init(adev, i);
+
+	if (!amdgpu_sriov_vf(adev)) {
+		tmp_mask = inst_mask;
+		for_each_inst(i, tmp_mask) {
+			r = gfx_v9_4_3_xcc_rlc_resume(adev, i);
+			if (r)
+				return r;
+		}
+	}
+
+	tmp_mask = inst_mask;
+	for_each_inst(i, tmp_mask) {
+		r = gfx_v9_4_3_xcc_cp_resume(adev, i);
+		if (r)
+			return r;
+	}
+
+	return 0;
+}
+
+static int gfx_v9_4_3_xcp_suspend(void *handle, uint32_t inst_mask)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	int i;
+
+	for_each_inst(i, inst_mask)
+		gfx_v9_4_3_xcc_fini(adev, i);
+
+	return 0;
+}
+
+struct amdgpu_xcp_ip_funcs gfx_v9_4_3_xcp_funcs = {
+	.suspend = &gfx_v9_4_3_xcp_suspend,
+	.resume = &gfx_v9_4_3_xcp_resume
+};
+
+struct amdgpu_ras_block_hw_ops  gfx_v9_4_3_ras_ops = {
+	.query_ras_error_count = &gfx_v9_4_3_query_ras_error_count,
+	.reset_ras_error_count = &gfx_v9_4_3_reset_ras_error_count,
+};
+
+static int gfx_v9_4_3_ras_late_init(struct amdgpu_device *adev, struct ras_common_if *ras_block)
+{
+	int r;
+
+	r = amdgpu_ras_block_late_init(adev, ras_block);
+	if (r)
+		return r;
+
+	r = amdgpu_ras_bind_aca(adev, AMDGPU_RAS_BLOCK__GFX,
+				&gfx_v9_4_3_aca_info,
+				NULL);
+	if (r)
+		goto late_fini;
+
+	return 0;
+
+late_fini:
+	amdgpu_ras_block_late_fini(adev, ras_block);
+
+	return r;
+}
+
+struct amdgpu_gfx_ras gfx_v9_4_3_ras = {
+	.ras_block = {
+		.hw_ops = &gfx_v9_4_3_ras_ops,
+		.ras_late_init = &gfx_v9_4_3_ras_late_init,
+	},
+	.enable_watchdog_timer = &gfx_v9_4_3_enable_watchdog_timer,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.h b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.h
new file mode 100644
index 000000000000..42d67ee0e7ef
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.h
@@ -0,0 +1,31 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __GFX_V9_4_3_H__
+#define __GFX_V9_4_3_H__
+
+extern const struct amdgpu_ip_block_version gfx_v9_4_3_ip_block;
+
+extern struct amdgpu_xcp_ip_funcs gfx_v9_4_3_xcp_funcs;
+
+#endif /* __GFX_V9_4_3_H__ */
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3_cleaner_shader.asm b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3_cleaner_shader.asm
new file mode 100644
index 000000000000..d5325ef80ab0
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3_cleaner_shader.asm
@@ -0,0 +1,153 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2024 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+// This shader is to clean LDS, SGPRs and VGPRs. It is  first 64 Dwords or 256 bytes of 192 Dwords cleaner shader.
+//To turn this shader program on for complitaion change this to main and lower shader main to main_1
+ 
+// MI300 : Clear SGPRs, VGPRs and LDS
+//   Uses two kernels launched separately:
+//   1. Clean VGPRs, LDS, and lower SGPRs
+//        Launches one workgroup per CU, each workgroup with 4x wave64 per SIMD in the CU
+//        Waves are "wave64" and have 128 VGPRs each, which uses all 512 VGPRs per SIMD
+//        Waves in the workgroup share the 64KB of LDS
+//        Each wave clears SGPRs 0 - 95. Because there are 4 waves/SIMD, this is physical SGPRs 0-383
+//        Each wave clears 128 VGPRs, so all 512 in the SIMD
+//        The first wave of the workgroup clears its 64KB of LDS
+//        The shader starts with "S_BARRIER" to ensure SPI has launched all waves of the workgroup
+//          before any wave in the workgroup could end.  Without this, it is possible not all SGPRs get cleared.
+//    2. Clean remaining SGPRs
+//        Launches a workgroup with 24 waves per workgroup, yielding 6 waves per SIMD in each CU
+//        Waves are allocating 96 SGPRs
+//          CP sets up SPI_RESOURCE_RESERVE_* registers to prevent these waves from allocating SGPRs 0-223.
+//          As such, these 6 waves per SIMD are allocated physical SGPRs 224-799
+//        Barriers do not work for >16 waves per workgroup, so we cannot start with S_BARRIER
+//          Instead, the shader starts with an S_SETHALT 1. Once all waves are launched CP will send unhalt command
+//        The shader then clears all SGPRs allocated to it, cleaning out physical SGPRs 224-799
+ 
+shader main
+  asic(MI300)
+  type(CS)
+  wave_size(64)
+// Note: original source code from SQ team
+
+//   (theorhetical fastest = ~512clks vgpr + 1536 lds + ~128 sgpr  = 2176 clks)
+
+  s_cmp_eq_u32 s0, 1                                // Bit0 is set, sgpr0 is set then clear VGPRS and LDS as FW set COMPUTE_USER_DATA_3
+  s_cbranch_scc0  label_0023                        // Clean VGPRs and LDS if sgpr0 of wave is set, scc = (s3 == 1)
+  S_BARRIER
+
+  s_movk_i32    m0, 0x0000
+  s_mov_b32     s2, 0x00000078  // Loop 128/8=16 times  (loop unrolled for performance)
+  //
+  // CLEAR VGPRs
+  //
+  s_set_gpr_idx_on  s2, 0x8    // enable Dest VGPR indexing
+label_0005:
+  v_mov_b32     v0, 0
+  v_mov_b32     v1, 0
+  v_mov_b32     v2, 0
+  v_mov_b32     v3, 0
+  v_mov_b32     v4, 0
+  v_mov_b32     v5, 0
+  v_mov_b32     v6, 0
+  v_mov_b32     v7, 0
+  s_sub_u32     s2, s2, 8
+  s_set_gpr_idx_idx  s2
+  s_cbranch_scc0  label_0005
+  s_set_gpr_idx_off
+ 
+  //
+  //
+ 
+  s_mov_b32     s2, 0x80000000                      // Bit31 is first_wave
+  s_and_b32     s2, s2, s1                          // sgpr0 has tg_size (first_wave) term as in ucode only COMPUTE_PGM_RSRC2.tg_size_en is set
+  s_cbranch_scc0  label_clean_sgpr_1                // Clean LDS if its first wave of ThreadGroup/WorkGroup
+  // CLEAR LDS
+  //
+  s_mov_b32 exec_lo, 0xffffffff
+  s_mov_b32 exec_hi, 0xffffffff
+  v_mbcnt_lo_u32_b32  v1, exec_hi, 0          // Set V1 to thread-ID (0..63)
+  v_mbcnt_hi_u32_b32  v1, exec_lo, v1         // Set V1 to thread-ID (0..63)
+  v_mul_u32_u24  v1, 0x00000008, v1           // * 8, so each thread is a double-dword address (8byte)
+  s_mov_b32     s2, 0x00000003f               // 64 loop iteraions
+  s_mov_b32     m0, 0xffffffff
+  // Clear all of LDS space
+  // Each FirstWave of WorkGroup clears 64kbyte block
+ 
+label_001F:
+  ds_write2_b64  v1, v[2:3], v[2:3] offset1:32
+  ds_write2_b64  v1, v[4:5], v[4:5] offset0:64 offset1:96
+  v_add_co_u32     v1, vcc, 0x00000400, v1
+  s_sub_u32     s2, s2, 1
+  s_cbranch_scc0  label_001F
+  //
+  // CLEAR SGPRs
+  //
+label_clean_sgpr_1:
+  s_mov_b32     m0, 0x0000005c   // Loop 96/4=24 times  (loop unrolled for performance)
+  s_nop 0
+label_sgpr_loop:
+  s_movreld_b32     s0, 0
+  s_movreld_b32     s1, 0
+  s_movreld_b32     s2, 0
+  s_movreld_b32     s3, 0
+  s_sub_u32         m0, m0, 4
+  s_cbranch_scc0  label_sgpr_loop
+ 
+  //clear vcc, flat scratch
+  s_mov_b32 flat_scratch_lo, 0   //clear  flat scratch lo SGPR 
+  s_mov_b32 flat_scratch_hi, 0   //clear  flat scratch hi SGPR 
+  s_mov_b64 vcc, 0               //clear vcc
+  s_mov_b64 ttmp0, 0             //Clear ttmp0 and ttmp1 
+  s_mov_b64 ttmp2, 0             //Clear ttmp2 and ttmp3 
+  s_mov_b64 ttmp4, 0             //Clear ttmp4 and ttmp5 
+  s_mov_b64 ttmp6, 0             //Clear ttmp6 and ttmp7 
+  s_mov_b64 ttmp8, 0             //Clear ttmp8 and ttmp9 
+  s_mov_b64 ttmp10, 0            //Clear ttmp10 and ttmp11 
+  s_mov_b64 ttmp12, 0            //Clear ttmp12 and ttmp13 
+  s_mov_b64 ttmp14, 0            //Clear ttmp14 and ttmp15 
+s_endpgm
+
+label_0023:
+
+  s_sethalt 1
+
+  s_mov_b32     m0, 0x0000005c   // Loop 96/4=24 times  (loop unrolled for performance)
+  s_nop 0
+label_sgpr_loop1:
+
+  s_movreld_b32     s0, 0
+  s_movreld_b32     s1, 0
+  s_movreld_b32     s2, 0
+  s_movreld_b32     s3, 0
+  s_sub_u32         m0, m0, 4
+  s_cbranch_scc0  label_sgpr_loop1
+ 
+  //clear vcc, flat scratch
+  s_mov_b32 flat_scratch_lo, 0   //clear  flat scratch lo SGPR 
+  s_mov_b32 flat_scratch_hi, 0   //clear  flat scratch hi SGPR 
+  s_mov_b64 vcc, 0xee            //clear vcc
+
+s_endpgm
+end  
+  
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3_cleaner_shader.h b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3_cleaner_shader.h
new file mode 100644
index 000000000000..69aa567c6c1d
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3_cleaner_shader.h
@@ -0,0 +1,64 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2024 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+/* Define the cleaner shader gfx_9_4_3 */
+static const u32 gfx_9_4_3_cleaner_shader_hex[] = {
+	0xbf068100, 0xbf84003b,
+	0xbf8a0000, 0xb07c0000,
+	0xbe8200ff, 0x00000078,
+	0xbf110802, 0x7e000280,
+	0x7e020280, 0x7e040280,
+	0x7e060280, 0x7e080280,
+	0x7e0a0280, 0x7e0c0280,
+	0x7e0e0280, 0x80828802,
+	0xbe803202, 0xbf84fff5,
+	0xbf9c0000, 0xbe8200ff,
+	0x80000000, 0x86020102,
+	0xbf840011, 0xbefe00c1,
+	0xbeff00c1, 0xd28c0001,
+	0x0001007f, 0xd28d0001,
+	0x0002027e, 0x10020288,
+	0xbe8200bf, 0xbefc00c1,
+	0xd89c2000, 0x00020201,
+	0xd89c6040, 0x00040401,
+	0x320202ff, 0x00000400,
+	0x80828102, 0xbf84fff8,
+	0xbefc00ff, 0x0000005c,
+	0xbf800000, 0xbe802c80,
+	0xbe812c80, 0xbe822c80,
+	0xbe832c80, 0x80fc847c,
+	0xbf84fffa, 0xbee60080,
+	0xbee70080, 0xbeea0180,
+	0xbeec0180, 0xbeee0180,
+	0xbef00180, 0xbef20180,
+	0xbef40180, 0xbef60180,
+	0xbef80180, 0xbefa0180,
+	0xbf810000, 0xbf8d0001,
+	0xbefc00ff, 0x0000005c,
+	0xbf800000, 0xbe802c80,
+	0xbe812c80, 0xbe822c80,
+	0xbe832c80, 0x80fc847c,
+	0xbf84fffa, 0xbee60080,
+	0xbee70080, 0xbeea01ff,
+	0x000000ee, 0xbf810000,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/gfxhub_v11_5_0.c b/drivers/gpu/drm/amd/amdgpu/gfxhub_v11_5_0.c
new file mode 100644
index 000000000000..f9949fedfbb9
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/gfxhub_v11_5_0.c
@@ -0,0 +1,516 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "amdgpu.h"
+#include "gfxhub_v11_5_0.h"
+
+#include "gc/gc_11_5_0_offset.h"
+#include "gc/gc_11_5_0_sh_mask.h"
+
+#include "navi10_enum.h"
+#include "soc15_common.h"
+
+#define regGCVM_L2_CNTL3_DEFAULT                0x80100007
+#define regGCVM_L2_CNTL4_DEFAULT                0x000000c1
+#define regGCVM_L2_CNTL5_DEFAULT                0x00003fe0
+
+
+static const char *gfxhub_client_ids[] = {
+	"CB/DB",
+	"Reserved",
+	"GE1",
+	"GE2",
+	"CPF",
+	"CPC",
+	"CPG",
+	"RLC",
+	"TCP",
+	"SQC (inst)",
+	"SQC (data)",
+	"SQG",
+	"Reserved",
+	"SDMA0",
+	"SDMA1",
+	"GCR",
+	"SDMA2",
+	"SDMA3",
+};
+
+static uint32_t gfxhub_v11_5_0_get_invalidate_req(unsigned int vmid,
+					       uint32_t flush_type)
+{
+	u32 req = 0;
+
+	/* invalidate using legacy mode on vmid*/
+	req = REG_SET_FIELD(req, GCVM_INVALIDATE_ENG0_REQ,
+			    PER_VMID_INVALIDATE_REQ, 1 << vmid);
+	req = REG_SET_FIELD(req, GCVM_INVALIDATE_ENG0_REQ, FLUSH_TYPE, flush_type);
+	req = REG_SET_FIELD(req, GCVM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PTES, 1);
+	req = REG_SET_FIELD(req, GCVM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PDE0, 1);
+	req = REG_SET_FIELD(req, GCVM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PDE1, 1);
+	req = REG_SET_FIELD(req, GCVM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PDE2, 1);
+	req = REG_SET_FIELD(req, GCVM_INVALIDATE_ENG0_REQ, INVALIDATE_L1_PTES, 1);
+	req = REG_SET_FIELD(req, GCVM_INVALIDATE_ENG0_REQ,
+			    CLEAR_PROTECTION_FAULT_STATUS_ADDR,	0);
+
+	return req;
+}
+
+static void
+gfxhub_v11_5_0_print_l2_protection_fault_status(struct amdgpu_device *adev,
+					     uint32_t status)
+{
+	u32 cid = REG_GET_FIELD(status,
+				GCVM_L2_PROTECTION_FAULT_STATUS, CID);
+
+	dev_err(adev->dev,
+		"GCVM_L2_PROTECTION_FAULT_STATUS:0x%08X\n",
+		status);
+	dev_err(adev->dev, "\t Faulty UTCL2 client ID: %s (0x%x)\n",
+		cid >= ARRAY_SIZE(gfxhub_client_ids) ? "unknown" : gfxhub_client_ids[cid],
+		cid);
+	dev_err(adev->dev, "\t MORE_FAULTS: 0x%lx\n",
+		REG_GET_FIELD(status,
+		GCVM_L2_PROTECTION_FAULT_STATUS, MORE_FAULTS));
+	dev_err(adev->dev, "\t WALKER_ERROR: 0x%lx\n",
+		REG_GET_FIELD(status,
+		GCVM_L2_PROTECTION_FAULT_STATUS, WALKER_ERROR));
+	dev_err(adev->dev, "\t PERMISSION_FAULTS: 0x%lx\n",
+		REG_GET_FIELD(status,
+		GCVM_L2_PROTECTION_FAULT_STATUS, PERMISSION_FAULTS));
+	dev_err(adev->dev, "\t MAPPING_ERROR: 0x%lx\n",
+		REG_GET_FIELD(status,
+		GCVM_L2_PROTECTION_FAULT_STATUS, MAPPING_ERROR));
+	dev_err(adev->dev, "\t RW: 0x%lx\n",
+		REG_GET_FIELD(status,
+		GCVM_L2_PROTECTION_FAULT_STATUS, RW));
+}
+
+static u64 gfxhub_v11_5_0_get_fb_location(struct amdgpu_device *adev)
+{
+	u64 base = RREG32_SOC15(GC, 0, regGCMC_VM_FB_LOCATION_BASE);
+
+	base &= GCMC_VM_FB_LOCATION_BASE__FB_BASE_MASK;
+	base <<= 24;
+
+	return base;
+}
+
+static u64 gfxhub_v11_5_0_get_mc_fb_offset(struct amdgpu_device *adev)
+{
+	return (u64)RREG32_SOC15(GC, 0, regGCMC_VM_FB_OFFSET) << 24;
+}
+
+static void gfxhub_v11_5_0_setup_vm_pt_regs(struct amdgpu_device *adev, uint32_t vmid,
+				uint64_t page_table_base)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)];
+
+	WREG32_SOC15_OFFSET(GC, 0, regGCVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32,
+			    hub->ctx_addr_distance * vmid,
+			    lower_32_bits(page_table_base));
+
+	WREG32_SOC15_OFFSET(GC, 0, regGCVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32,
+			    hub->ctx_addr_distance * vmid,
+			    upper_32_bits(page_table_base));
+}
+
+static void gfxhub_v11_5_0_init_gart_aperture_regs(struct amdgpu_device *adev)
+{
+	uint64_t pt_base = amdgpu_gmc_pd_addr(adev->gart.bo);
+
+	gfxhub_v11_5_0_setup_vm_pt_regs(adev, 0, pt_base);
+
+	WREG32_SOC15(GC, 0, regGCVM_CONTEXT0_PAGE_TABLE_START_ADDR_LO32,
+		     (u32)(adev->gmc.gart_start >> 12));
+	WREG32_SOC15(GC, 0, regGCVM_CONTEXT0_PAGE_TABLE_START_ADDR_HI32,
+		     (u32)(adev->gmc.gart_start >> 44));
+
+	WREG32_SOC15(GC, 0, regGCVM_CONTEXT0_PAGE_TABLE_END_ADDR_LO32,
+		     (u32)(adev->gmc.gart_end >> 12));
+	WREG32_SOC15(GC, 0, regGCVM_CONTEXT0_PAGE_TABLE_END_ADDR_HI32,
+		     (u32)(adev->gmc.gart_end >> 44));
+}
+
+static void gfxhub_v11_5_0_init_system_aperture_regs(struct amdgpu_device *adev)
+{
+	uint64_t value;
+
+	WREG32_SOC15(GC, 0, regGCMC_VM_AGP_BASE, 0);
+	WREG32_SOC15(GC, 0, regGCMC_VM_AGP_BOT, adev->gmc.agp_start >> 24);
+	WREG32_SOC15(GC, 0, regGCMC_VM_AGP_TOP, adev->gmc.agp_end >> 24);
+
+	/* Program the system aperture low logical page number. */
+	WREG32_SOC15(GC, 0, regGCMC_VM_SYSTEM_APERTURE_LOW_ADDR,
+		     min(adev->gmc.fb_start, adev->gmc.agp_start) >> 18);
+
+	WREG32_SOC15(GC, 0, regGCMC_VM_SYSTEM_APERTURE_HIGH_ADDR,
+		     max(adev->gmc.fb_end, adev->gmc.agp_end) >> 18);
+
+	/* Set default page address. */
+	value = amdgpu_gmc_vram_mc2pa(adev, adev->mem_scratch.gpu_addr);
+	WREG32_SOC15(GC, 0, regGCMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB,
+		     (u32)(value >> 12));
+	WREG32_SOC15(GC, 0, regGCMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_MSB,
+		     (u32)(value >> 44));
+
+	/* Program "protection fault". */
+	WREG32_SOC15(GC, 0, regGCVM_L2_PROTECTION_FAULT_DEFAULT_ADDR_LO32,
+		     (u32)(adev->dummy_page_addr >> 12));
+	WREG32_SOC15(GC, 0, regGCVM_L2_PROTECTION_FAULT_DEFAULT_ADDR_HI32,
+		     (u32)((u64)adev->dummy_page_addr >> 44));
+
+	WREG32_FIELD15_PREREG(GC, 0, GCVM_L2_PROTECTION_FAULT_CNTL2,
+		       ACTIVE_PAGE_MIGRATION_PTE_READ_RETRY, 1);
+}
+
+static void gfxhub_v11_5_0_init_tlb_regs(struct amdgpu_device *adev)
+{
+	uint32_t tmp;
+
+	/* Setup TLB control */
+	tmp = RREG32_SOC15(GC, 0, regGCMC_VM_MX_L1_TLB_CNTL);
+
+	tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB, 1);
+	tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL, SYSTEM_ACCESS_MODE, 3);
+	tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL,
+			    ENABLE_ADVANCED_DRIVER_MODEL, 1);
+	tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL,
+			    SYSTEM_APERTURE_UNMAPPED_ACCESS, 0);
+	tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL, ECO_BITS, 0);
+	tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL,
+			    MTYPE, MTYPE_UC); /* UC, uncached */
+
+	WREG32_SOC15(GC, 0, regGCMC_VM_MX_L1_TLB_CNTL, tmp);
+}
+
+static void gfxhub_v11_5_0_init_cache_regs(struct amdgpu_device *adev)
+{
+	uint32_t tmp;
+
+	/* These registers are not accessible to VF-SRIOV.
+	 * The PF will program them instead.
+	 */
+	if (amdgpu_sriov_vf(adev))
+		return;
+
+	/* Setup L2 cache */
+	tmp = RREG32_SOC15(GC, 0, regGCVM_L2_CNTL);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL, ENABLE_L2_CACHE, 1);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL, ENABLE_L2_FRAGMENT_PROCESSING, 0);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL,
+			    ENABLE_DEFAULT_PAGE_OUT_TO_SYSTEM_MEMORY, 1);
+	/* XXX for emulation, Refer to closed source code.*/
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL,
+			    L2_PDE0_CACHE_TAG_GENERATION_MODE, 0);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL, PDE_FAULT_CLASSIFICATION, 0);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL, CONTEXT1_IDENTITY_ACCESS_MODE, 1);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL, IDENTITY_MODE_FRAGMENT_SIZE, 0);
+	WREG32_SOC15(GC, 0, regGCVM_L2_CNTL, tmp);
+
+	tmp = RREG32_SOC15(GC, 0, regGCVM_L2_CNTL2);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL2, INVALIDATE_ALL_L1_TLBS, 1);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL2, INVALIDATE_L2_CACHE, 1);
+	WREG32_SOC15(GC, 0, regGCVM_L2_CNTL2, tmp);
+
+	tmp = regGCVM_L2_CNTL3_DEFAULT;
+	if (adev->gmc.translate_further) {
+		tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL3, BANK_SELECT, 12);
+		tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL3,
+				    L2_CACHE_BIGK_FRAGMENT_SIZE, 9);
+	} else {
+		tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL3, BANK_SELECT, 9);
+		tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL3,
+				    L2_CACHE_BIGK_FRAGMENT_SIZE, 6);
+	}
+	WREG32_SOC15(GC, 0, regGCVM_L2_CNTL3, tmp);
+
+	tmp = regGCVM_L2_CNTL4_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL4, VMC_TAP_PDE_REQUEST_PHYSICAL, 0);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL4, VMC_TAP_PTE_REQUEST_PHYSICAL, 0);
+	WREG32_SOC15(GC, 0, regGCVM_L2_CNTL4, tmp);
+
+	tmp = regGCVM_L2_CNTL5_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL5, L2_CACHE_SMALLK_FRAGMENT_SIZE, 0);
+	WREG32_SOC15(GC, 0, regGCVM_L2_CNTL5, tmp);
+}
+
+static void gfxhub_v11_5_0_enable_system_domain(struct amdgpu_device *adev)
+{
+	uint32_t tmp;
+
+	tmp = RREG32_SOC15(GC, 0, regGCVM_CONTEXT0_CNTL);
+	tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT0_CNTL, ENABLE_CONTEXT, 1);
+	tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT0_CNTL, PAGE_TABLE_DEPTH, 0);
+	tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT0_CNTL,
+			    RETRY_PERMISSION_OR_INVALID_PAGE_FAULT, 0);
+	WREG32_SOC15(GC, 0, regGCVM_CONTEXT0_CNTL, tmp);
+}
+
+static void gfxhub_v11_5_0_disable_identity_aperture(struct amdgpu_device *adev)
+{
+	/* These registers are not accessible to VF-SRIOV.
+	 * The PF will program them instead.
+	 */
+	if (amdgpu_sriov_vf(adev))
+		return;
+
+	WREG32_SOC15(GC, 0, regGCVM_L2_CONTEXT1_IDENTITY_APERTURE_LOW_ADDR_LO32,
+		     0xFFFFFFFF);
+	WREG32_SOC15(GC, 0, regGCVM_L2_CONTEXT1_IDENTITY_APERTURE_LOW_ADDR_HI32,
+		     0x0000000F);
+
+	WREG32_SOC15(GC, 0, regGCVM_L2_CONTEXT1_IDENTITY_APERTURE_HIGH_ADDR_LO32,
+		     0);
+	WREG32_SOC15(GC, 0, regGCVM_L2_CONTEXT1_IDENTITY_APERTURE_HIGH_ADDR_HI32,
+		     0);
+
+	WREG32_SOC15(GC, 0, regGCVM_L2_CONTEXT_IDENTITY_PHYSICAL_OFFSET_LO32, 0);
+	WREG32_SOC15(GC, 0, regGCVM_L2_CONTEXT_IDENTITY_PHYSICAL_OFFSET_HI32, 0);
+
+}
+
+static void gfxhub_v11_5_0_setup_vmid_config(struct amdgpu_device *adev)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)];
+	int i;
+	uint32_t tmp;
+
+	for (i = 0; i <= 14; i++) {
+		tmp = RREG32_SOC15_OFFSET(GC, 0, regGCVM_CONTEXT1_CNTL, i * hub->ctx_distance);
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL, ENABLE_CONTEXT, 1);
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL, PAGE_TABLE_DEPTH,
+				    adev->vm_manager.num_level);
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
+				RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
+				DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
+				PDE0_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
+				VALID_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
+				READ_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
+				WRITE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
+				EXECUTE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
+				PAGE_TABLE_BLOCK_SIZE,
+				adev->vm_manager.block_size - 9);
+		/* Send no-retry XNACK on fault to suppress VM fault storm. */
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
+				    RETRY_PERMISSION_OR_INVALID_PAGE_FAULT,
+				    !amdgpu_noretry);
+		WREG32_SOC15_OFFSET(GC, 0, regGCVM_CONTEXT1_CNTL,
+				    i * hub->ctx_distance, tmp);
+		WREG32_SOC15_OFFSET(GC, 0, regGCVM_CONTEXT1_PAGE_TABLE_START_ADDR_LO32,
+				    i * hub->ctx_addr_distance, 0);
+		WREG32_SOC15_OFFSET(GC, 0, regGCVM_CONTEXT1_PAGE_TABLE_START_ADDR_HI32,
+				    i * hub->ctx_addr_distance, 0);
+		WREG32_SOC15_OFFSET(GC, 0, regGCVM_CONTEXT1_PAGE_TABLE_END_ADDR_LO32,
+				    i * hub->ctx_addr_distance,
+				    lower_32_bits(adev->vm_manager.max_pfn - 1));
+		WREG32_SOC15_OFFSET(GC, 0, regGCVM_CONTEXT1_PAGE_TABLE_END_ADDR_HI32,
+				    i * hub->ctx_addr_distance,
+				    upper_32_bits(adev->vm_manager.max_pfn - 1));
+	}
+
+	hub->vm_cntx_cntl = tmp;
+}
+
+static void gfxhub_v11_5_0_program_invalidation(struct amdgpu_device *adev)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)];
+	unsigned i;
+
+	for (i = 0 ; i < 18; ++i) {
+		WREG32_SOC15_OFFSET(GC, 0, regGCVM_INVALIDATE_ENG0_ADDR_RANGE_LO32,
+				    i * hub->eng_addr_distance, 0xffffffff);
+		WREG32_SOC15_OFFSET(GC, 0, regGCVM_INVALIDATE_ENG0_ADDR_RANGE_HI32,
+				    i * hub->eng_addr_distance, 0x1f);
+	}
+}
+
+static int gfxhub_v11_5_0_gart_enable(struct amdgpu_device *adev)
+{
+	if (amdgpu_sriov_vf(adev)) {
+		/*
+		 * GCMC_VM_FB_LOCATION_BASE/TOP is NULL for VF, becuase they are
+		 * VF copy registers so vbios post doesn't program them, for
+		 * SRIOV driver need to program them
+		 */
+		WREG32_SOC15(GC, 0, regGCMC_VM_FB_LOCATION_BASE,
+			     adev->gmc.vram_start >> 24);
+		WREG32_SOC15(GC, 0, regGCMC_VM_FB_LOCATION_TOP,
+			     adev->gmc.vram_end >> 24);
+	}
+
+	/* GART Enable. */
+	gfxhub_v11_5_0_init_gart_aperture_regs(adev);
+	gfxhub_v11_5_0_init_system_aperture_regs(adev);
+	gfxhub_v11_5_0_init_tlb_regs(adev);
+	gfxhub_v11_5_0_init_cache_regs(adev);
+
+	gfxhub_v11_5_0_enable_system_domain(adev);
+	gfxhub_v11_5_0_disable_identity_aperture(adev);
+	gfxhub_v11_5_0_setup_vmid_config(adev);
+	gfxhub_v11_5_0_program_invalidation(adev);
+
+	return 0;
+}
+
+static void gfxhub_v11_5_0_gart_disable(struct amdgpu_device *adev)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)];
+	u32 tmp;
+	u32 i;
+
+	/* Disable all tables */
+	for (i = 0; i < 16; i++)
+		WREG32_SOC15_OFFSET(GC, 0, regGCVM_CONTEXT0_CNTL,
+				    i * hub->ctx_distance, 0);
+
+	/* Setup TLB control */
+	tmp = RREG32_SOC15(GC, 0, regGCMC_VM_MX_L1_TLB_CNTL);
+	tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB, 0);
+	tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL,
+			    ENABLE_ADVANCED_DRIVER_MODEL, 0);
+	WREG32_SOC15(GC, 0, regGCMC_VM_MX_L1_TLB_CNTL, tmp);
+
+	/* Setup L2 cache */
+	WREG32_FIELD15_PREREG(GC, 0, GCVM_L2_CNTL, ENABLE_L2_CACHE, 0);
+	WREG32_SOC15(GC, 0, regGCVM_L2_CNTL3, 0);
+}
+
+/**
+ * gfxhub_v11_5_0_set_fault_enable_default - update GART/VM fault handling
+ *
+ * @adev: amdgpu_device pointer
+ * @value: true redirects VM faults to the default page
+ */
+static void gfxhub_v11_5_0_set_fault_enable_default(struct amdgpu_device *adev,
+					  bool value)
+{
+	u32 tmp;
+
+	/* NO halt CP when page fault */
+	tmp = RREG32_SOC15(GC, 0, regCP_DEBUG);
+	tmp = REG_SET_FIELD(tmp, CP_DEBUG, CPG_UTCL1_ERROR_HALT_DISABLE, 1);
+	WREG32_SOC15(GC, 0, regCP_DEBUG, tmp);
+
+	/* These registers are not accessible to VF-SRIOV.
+	 * The PF will program them instead.
+	 */
+	if (amdgpu_sriov_vf(adev))
+		return;
+
+	tmp = RREG32_SOC15(GC, 0, regGCVM_L2_PROTECTION_FAULT_CNTL);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    PDE0_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    PDE1_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    PDE2_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    TRANSLATE_FURTHER_PROTECTION_FAULT_ENABLE_DEFAULT,
+			    value);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    NACK_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    VALID_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    READ_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    WRITE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    EXECUTE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	if (!value) {
+		tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+				CRASH_ON_NO_RETRY_FAULT, 1);
+		tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+				CRASH_ON_RETRY_FAULT, 1);
+	}
+	WREG32_SOC15(GC, 0, regGCVM_L2_PROTECTION_FAULT_CNTL, tmp);
+}
+
+static const struct amdgpu_vmhub_funcs gfxhub_v11_5_0_vmhub_funcs = {
+	.print_l2_protection_fault_status = gfxhub_v11_5_0_print_l2_protection_fault_status,
+	.get_invalidate_req = gfxhub_v11_5_0_get_invalidate_req,
+};
+
+static void gfxhub_v11_5_0_init(struct amdgpu_device *adev)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)];
+
+	hub->ctx0_ptb_addr_lo32 =
+		SOC15_REG_OFFSET(GC, 0,
+				 regGCVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32);
+	hub->ctx0_ptb_addr_hi32 =
+		SOC15_REG_OFFSET(GC, 0,
+				 regGCVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32);
+	hub->vm_inv_eng0_sem =
+		SOC15_REG_OFFSET(GC, 0, regGCVM_INVALIDATE_ENG0_SEM);
+	hub->vm_inv_eng0_req =
+		SOC15_REG_OFFSET(GC, 0, regGCVM_INVALIDATE_ENG0_REQ);
+	hub->vm_inv_eng0_ack =
+		SOC15_REG_OFFSET(GC, 0, regGCVM_INVALIDATE_ENG0_ACK);
+	hub->vm_context0_cntl =
+		SOC15_REG_OFFSET(GC, 0, regGCVM_CONTEXT0_CNTL);
+	hub->vm_l2_pro_fault_status =
+		SOC15_REG_OFFSET(GC, 0, regGCVM_L2_PROTECTION_FAULT_STATUS);
+	hub->vm_l2_pro_fault_cntl =
+		SOC15_REG_OFFSET(GC, 0, regGCVM_L2_PROTECTION_FAULT_CNTL);
+
+	hub->ctx_distance = regGCVM_CONTEXT1_CNTL - regGCVM_CONTEXT0_CNTL;
+	hub->ctx_addr_distance = regGCVM_CONTEXT1_PAGE_TABLE_BASE_ADDR_LO32 -
+		regGCVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32;
+	hub->eng_distance = regGCVM_INVALIDATE_ENG1_REQ -
+		regGCVM_INVALIDATE_ENG0_REQ;
+	hub->eng_addr_distance = regGCVM_INVALIDATE_ENG1_ADDR_RANGE_LO32 -
+		regGCVM_INVALIDATE_ENG0_ADDR_RANGE_LO32;
+
+	hub->vm_cntx_cntl_vm_fault = GCVM_CONTEXT1_CNTL__RANGE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		GCVM_CONTEXT1_CNTL__DUMMY_PAGE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		GCVM_CONTEXT1_CNTL__PDE0_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		GCVM_CONTEXT1_CNTL__VALID_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		GCVM_CONTEXT1_CNTL__READ_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		GCVM_CONTEXT1_CNTL__WRITE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		GCVM_CONTEXT1_CNTL__EXECUTE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK;
+
+	hub->vmhub_funcs = &gfxhub_v11_5_0_vmhub_funcs;
+}
+
+const struct amdgpu_gfxhub_funcs gfxhub_v11_5_0_funcs = {
+	.get_fb_location = gfxhub_v11_5_0_get_fb_location,
+	.get_mc_fb_offset = gfxhub_v11_5_0_get_mc_fb_offset,
+	.setup_vm_pt_regs = gfxhub_v11_5_0_setup_vm_pt_regs,
+	.gart_enable = gfxhub_v11_5_0_gart_enable,
+	.gart_disable = gfxhub_v11_5_0_gart_disable,
+	.set_fault_enable_default = gfxhub_v11_5_0_set_fault_enable_default,
+	.init = gfxhub_v11_5_0_init,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/gfxhub_v11_5_0.h b/drivers/gpu/drm/amd/amdgpu/gfxhub_v11_5_0.h
new file mode 100644
index 000000000000..265ab631b3d0
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/gfxhub_v11_5_0.h
@@ -0,0 +1,29 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __GFXHUB_V11_5_0_H__
+#define __GFXHUB_V11_5_0_H__
+
+extern const struct amdgpu_gfxhub_funcs gfxhub_v11_5_0_funcs;
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/gfxhub_v12_0.c b/drivers/gpu/drm/amd/amdgpu/gfxhub_v12_0.c
new file mode 100644
index 000000000000..7609b9cecae8
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/gfxhub_v12_0.c
@@ -0,0 +1,521 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "amdgpu.h"
+#include "gfxhub_v12_0.h"
+
+#include "gc/gc_12_0_0_offset.h"
+#include "gc/gc_12_0_0_sh_mask.h"
+#include "soc24_enum.h"
+#include "soc15_common.h"
+
+#define regGCVM_L2_CNTL3_DEFAULT	0x80120007
+#define regGCVM_L2_CNTL4_DEFAULT	0x000000c1
+#define regGCVM_L2_CNTL5_DEFAULT	0x00003fe0
+#define regGRBM_GFX_INDEX_DEFAULT	0xe0000000
+
+static const char *gfxhub_client_ids[] = {
+	"CB",
+	"DB",
+	"GE1",
+	"GE2",
+	"CPF",
+	"CPC",
+	"CPG",
+	"RLC",
+	"TCP",
+	"SQC (inst)",
+	"SQC (data)",
+	"SQG/PC/SC",
+	"Reserved",
+	"SDMA0",
+	"SDMA1",
+	"GCR",
+	"Reserved",
+	"Reserved",
+	"WGS",
+	"DSM",
+	"PA"
+};
+
+static uint32_t gfxhub_v12_0_get_invalidate_req(unsigned int vmid,
+					        uint32_t flush_type)
+{
+	u32 req = 0;
+
+	/* invalidate using legacy mode on vmid*/
+	req = REG_SET_FIELD(req, GCVM_INVALIDATE_ENG0_REQ,
+			    PER_VMID_INVALIDATE_REQ, 1 << vmid);
+	req = REG_SET_FIELD(req, GCVM_INVALIDATE_ENG0_REQ, FLUSH_TYPE, flush_type);
+	req = REG_SET_FIELD(req, GCVM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PTES, 1);
+	req = REG_SET_FIELD(req, GCVM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PDE0, 1);
+	req = REG_SET_FIELD(req, GCVM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PDE1, 1);
+	req = REG_SET_FIELD(req, GCVM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PDE2, 1);
+	req = REG_SET_FIELD(req, GCVM_INVALIDATE_ENG0_REQ, INVALIDATE_L1_PTES, 1);
+	req = REG_SET_FIELD(req, GCVM_INVALIDATE_ENG0_REQ,
+			    CLEAR_PROTECTION_FAULT_STATUS_ADDR,	0);
+
+	return req;
+}
+
+static void
+gfxhub_v12_0_print_l2_protection_fault_status(struct amdgpu_device *adev,
+					      uint32_t status)
+{
+	u32 cid = REG_GET_FIELD(status,
+				GCVM_L2_PROTECTION_FAULT_STATUS_LO32, CID);
+
+	dev_err(adev->dev,
+		"GCVM_L2_PROTECTION_FAULT_STATUS:0x%08X\n",
+		status);
+	dev_err(adev->dev, "\t Faulty UTCL2 client ID: %s (0x%x)\n",
+		cid >= ARRAY_SIZE(gfxhub_client_ids) ? "unknown" : gfxhub_client_ids[cid],
+		cid);
+	dev_err(adev->dev, "\t MORE_FAULTS: 0x%lx\n",
+		REG_GET_FIELD(status,
+		GCVM_L2_PROTECTION_FAULT_STATUS_LO32, MORE_FAULTS));
+	dev_err(adev->dev, "\t WALKER_ERROR: 0x%lx\n",
+		REG_GET_FIELD(status,
+		GCVM_L2_PROTECTION_FAULT_STATUS_LO32, WALKER_ERROR));
+	dev_err(adev->dev, "\t PERMISSION_FAULTS: 0x%lx\n",
+		REG_GET_FIELD(status,
+		GCVM_L2_PROTECTION_FAULT_STATUS_LO32, PERMISSION_FAULTS));
+	dev_err(adev->dev, "\t MAPPING_ERROR: 0x%lx\n",
+		REG_GET_FIELD(status,
+		GCVM_L2_PROTECTION_FAULT_STATUS_LO32, MAPPING_ERROR));
+	dev_err(adev->dev, "\t RW: 0x%lx\n",
+		REG_GET_FIELD(status,
+		GCVM_L2_PROTECTION_FAULT_STATUS_LO32, RW));
+}
+
+static u64 gfxhub_v12_0_get_fb_location(struct amdgpu_device *adev)
+{
+	u64 base = RREG32_SOC15(GC, 0, regGCMC_VM_FB_LOCATION_BASE);
+
+	base &= GCMC_VM_FB_LOCATION_BASE__FB_BASE_MASK;
+	base <<= 24;
+
+	return base;
+}
+
+static u64 gfxhub_v12_0_get_mc_fb_offset(struct amdgpu_device *adev)
+{
+	return (u64)RREG32_SOC15(GC, 0, regGCMC_VM_FB_OFFSET) << 24;
+}
+
+static void gfxhub_v12_0_setup_vm_pt_regs(struct amdgpu_device *adev,
+					  uint32_t vmid,
+					  uint64_t page_table_base)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)];
+
+	WREG32_SOC15_OFFSET(GC, 0, regGCVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32,
+			    hub->ctx_addr_distance * vmid,
+			    lower_32_bits(page_table_base));
+
+	WREG32_SOC15_OFFSET(GC, 0, regGCVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32,
+			    hub->ctx_addr_distance * vmid,
+			    upper_32_bits(page_table_base));
+}
+
+static void gfxhub_v12_0_init_gart_aperture_regs(struct amdgpu_device *adev)
+{
+	uint64_t pt_base = amdgpu_gmc_pd_addr(adev->gart.bo);
+
+	gfxhub_v12_0_setup_vm_pt_regs(adev, 0, pt_base);
+
+	WREG32_SOC15(GC, 0, regGCVM_CONTEXT0_PAGE_TABLE_START_ADDR_LO32,
+		     (u32)(adev->gmc.gart_start >> 12));
+	WREG32_SOC15(GC, 0, regGCVM_CONTEXT0_PAGE_TABLE_START_ADDR_HI32,
+		     (u32)(adev->gmc.gart_start >> 44));
+
+	WREG32_SOC15(GC, 0, regGCVM_CONTEXT0_PAGE_TABLE_END_ADDR_LO32,
+		     (u32)(adev->gmc.gart_end >> 12));
+	WREG32_SOC15(GC, 0, regGCVM_CONTEXT0_PAGE_TABLE_END_ADDR_HI32,
+		     (u32)(adev->gmc.gart_end >> 44));
+}
+
+static void gfxhub_v12_0_init_system_aperture_regs(struct amdgpu_device *adev)
+{
+	uint64_t value;
+
+	/* Program the AGP BAR */
+	WREG32_SOC15(GC, 0, regGCMC_VM_AGP_BASE, 0);
+	WREG32_SOC15(GC, 0, regGCMC_VM_AGP_BOT, adev->gmc.agp_start >> 24);
+	WREG32_SOC15(GC, 0, regGCMC_VM_AGP_TOP, adev->gmc.agp_end >> 24);
+
+	/* Program the system aperture low logical page number. */
+	WREG32_SOC15(GC, 0, regGCMC_VM_SYSTEM_APERTURE_LOW_ADDR,
+		     min(adev->gmc.fb_start, adev->gmc.agp_start) >> 18);
+	WREG32_SOC15(GC, 0, regGCMC_VM_SYSTEM_APERTURE_HIGH_ADDR,
+		     max(adev->gmc.fb_end, adev->gmc.agp_end) >> 18);
+
+	/* Set default page address. */
+	value = adev->mem_scratch.gpu_addr - adev->gmc.vram_start
+		+ adev->vm_manager.vram_base_offset;
+	WREG32_SOC15(GC, 0, regGCMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB,
+		     (u32)(value >> 12));
+	WREG32_SOC15(GC, 0, regGCMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_MSB,
+		     (u32)(value >> 44));
+
+	/* Program "protection fault". */
+	WREG32_SOC15(GC, 0, regGCVM_L2_PROTECTION_FAULT_DEFAULT_ADDR_LO32,
+		     (u32)(adev->dummy_page_addr >> 12));
+	WREG32_SOC15(GC, 0, regGCVM_L2_PROTECTION_FAULT_DEFAULT_ADDR_HI32,
+		     (u32)((u64)adev->dummy_page_addr >> 44));
+
+	WREG32_FIELD15_PREREG(GC, 0, GCVM_L2_PROTECTION_FAULT_CNTL2,
+		       ACTIVE_PAGE_MIGRATION_PTE_READ_RETRY, 1);
+}
+
+
+static void gfxhub_v12_0_init_tlb_regs(struct amdgpu_device *adev)
+{
+	uint32_t tmp;
+
+	/* Setup TLB control */
+	tmp = RREG32_SOC15(GC, 0, regGCMC_VM_MX_L1_TLB_CNTL);
+
+	tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB, 1);
+	tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL, SYSTEM_ACCESS_MODE, 3);
+	tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL,
+			    ENABLE_ADVANCED_DRIVER_MODEL, 1);
+	tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL,
+			    SYSTEM_APERTURE_UNMAPPED_ACCESS, 0);
+	tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL, ECO_BITS, 0);
+	tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL,
+			    MTYPE, MTYPE_UC); /* UC, uncached */
+
+	WREG32_SOC15(GC, 0, regGCMC_VM_MX_L1_TLB_CNTL, tmp);
+}
+
+static void gfxhub_v12_0_init_cache_regs(struct amdgpu_device *adev)
+{
+	uint32_t tmp;
+
+	/* These registers are not accessible to VF-SRIOV.
+	 * The PF will program them instead.
+	 */
+	if (amdgpu_sriov_vf(adev))
+		return;
+
+	/* Setup L2 cache */
+	tmp = RREG32_SOC15(GC, 0, regGCVM_L2_CNTL);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL, ENABLE_L2_CACHE, 1);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL, ENABLE_L2_FRAGMENT_PROCESSING, 0);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL,
+			    ENABLE_DEFAULT_PAGE_OUT_TO_SYSTEM_MEMORY, 1);
+	/* XXX for emulation, Refer to closed source code.*/
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL,
+			    L2_PDE0_CACHE_TAG_GENERATION_MODE, 0);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL, PDE_FAULT_CLASSIFICATION, 0);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL, CONTEXT1_IDENTITY_ACCESS_MODE, 1);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL, IDENTITY_MODE_FRAGMENT_SIZE, 0);
+	WREG32_SOC15(GC, 0, regGCVM_L2_CNTL, tmp);
+
+	tmp = RREG32_SOC15(GC, 0, regGCVM_L2_CNTL2);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL2, INVALIDATE_ALL_L1_TLBS, 1);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL2, INVALIDATE_L2_CACHE, 1);
+	WREG32_SOC15(GC, 0, regGCVM_L2_CNTL2, tmp);
+
+	tmp = regGCVM_L2_CNTL3_DEFAULT;
+	if (adev->gmc.translate_further) {
+		tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL3, BANK_SELECT, 12);
+		tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL3,
+				    L2_CACHE_BIGK_FRAGMENT_SIZE, 9);
+	} else {
+		tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL3, BANK_SELECT, 9);
+		tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL3,
+				    L2_CACHE_BIGK_FRAGMENT_SIZE, 6);
+	}
+	WREG32_SOC15(GC, 0, regGCVM_L2_CNTL3, tmp);
+
+	tmp = regGCVM_L2_CNTL4_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL4, VMC_TAP_PDE_REQUEST_PHYSICAL, 0);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL4, VMC_TAP_PTE_REQUEST_PHYSICAL, 0);
+	WREG32_SOC15(GC, 0, regGCVM_L2_CNTL4, tmp);
+
+	tmp = regGCVM_L2_CNTL5_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL5, L2_CACHE_SMALLK_FRAGMENT_SIZE, 0);
+	WREG32_SOC15(GC, 0, regGCVM_L2_CNTL5, tmp);
+}
+
+static void gfxhub_v12_0_enable_system_domain(struct amdgpu_device *adev)
+{
+	uint32_t tmp;
+
+	tmp = RREG32_SOC15(GC, 0, regGCVM_CONTEXT0_CNTL);
+	tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT0_CNTL, ENABLE_CONTEXT, 1);
+	tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT0_CNTL, PAGE_TABLE_DEPTH, 0);
+	tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT0_CNTL,
+			    RETRY_PERMISSION_OR_INVALID_PAGE_FAULT, 0);
+	WREG32_SOC15(GC, 0, regGCVM_CONTEXT0_CNTL, tmp);
+}
+
+static void gfxhub_v12_0_disable_identity_aperture(struct amdgpu_device *adev)
+{
+	/* These registers are not accessible to VF-SRIOV.
+	 * The PF will program them instead.
+	 */
+	if (amdgpu_sriov_vf(adev))
+		return;
+
+	WREG32_SOC15(GC, 0, regGCVM_L2_CONTEXT1_IDENTITY_APERTURE_LOW_ADDR_LO32,
+		     0xFFFFFFFF);
+	WREG32_SOC15(GC, 0, regGCVM_L2_CONTEXT1_IDENTITY_APERTURE_LOW_ADDR_HI32,
+		     0x0000000F);
+
+	WREG32_SOC15(GC, 0, regGCVM_L2_CONTEXT1_IDENTITY_APERTURE_HIGH_ADDR_LO32,
+		     0);
+	WREG32_SOC15(GC, 0, regGCVM_L2_CONTEXT1_IDENTITY_APERTURE_HIGH_ADDR_HI32,
+		     0);
+
+	WREG32_SOC15(GC, 0, regGCVM_L2_CONTEXT_IDENTITY_PHYSICAL_OFFSET_LO32, 0);
+	WREG32_SOC15(GC, 0, regGCVM_L2_CONTEXT_IDENTITY_PHYSICAL_OFFSET_HI32, 0);
+
+}
+
+static void gfxhub_v12_0_setup_vmid_config(struct amdgpu_device *adev)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)];
+	int i;
+	uint32_t tmp;
+
+	for (i = 0; i <= 14; i++) {
+		tmp = RREG32_SOC15_OFFSET(GC, 0, regGCVM_CONTEXT1_CNTL, i);
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL, ENABLE_CONTEXT, 1);
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL, PAGE_TABLE_DEPTH,
+				    adev->vm_manager.num_level);
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
+				RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
+				DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
+				PDE0_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
+				VALID_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
+				READ_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
+				WRITE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
+				EXECUTE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
+				PAGE_TABLE_BLOCK_SIZE,
+				adev->vm_manager.block_size - 9);
+		/* Send no-retry XNACK on fault to suppress VM fault storm. */
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
+				    RETRY_PERMISSION_OR_INVALID_PAGE_FAULT,
+				    !amdgpu_noretry);
+		WREG32_SOC15_OFFSET(GC, 0, regGCVM_CONTEXT1_CNTL,
+				    i * hub->ctx_distance, tmp);
+		WREG32_SOC15_OFFSET(GC, 0, regGCVM_CONTEXT1_PAGE_TABLE_START_ADDR_LO32,
+				    i * hub->ctx_addr_distance, 0);
+		WREG32_SOC15_OFFSET(GC, 0, regGCVM_CONTEXT1_PAGE_TABLE_START_ADDR_HI32,
+				    i * hub->ctx_addr_distance, 0);
+		WREG32_SOC15_OFFSET(GC, 0, regGCVM_CONTEXT1_PAGE_TABLE_END_ADDR_LO32,
+				    i * hub->ctx_addr_distance,
+				    lower_32_bits(adev->vm_manager.max_pfn - 1));
+		WREG32_SOC15_OFFSET(GC, 0, regGCVM_CONTEXT1_PAGE_TABLE_END_ADDR_HI32,
+				    i * hub->ctx_addr_distance,
+				    upper_32_bits(adev->vm_manager.max_pfn - 1));
+	}
+
+	hub->vm_cntx_cntl = tmp;
+}
+
+static void gfxhub_v12_0_program_invalidation(struct amdgpu_device *adev)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)];
+	unsigned i;
+
+	for (i = 0 ; i < 18; ++i) {
+		WREG32_SOC15_OFFSET(GC, 0, regGCVM_INVALIDATE_ENG0_ADDR_RANGE_LO32,
+				    i * hub->eng_addr_distance, 0xffffffff);
+		WREG32_SOC15_OFFSET(GC, 0, regGCVM_INVALIDATE_ENG0_ADDR_RANGE_HI32,
+				    i * hub->eng_addr_distance, 0x1f);
+	}
+}
+
+static int gfxhub_v12_0_gart_enable(struct amdgpu_device *adev)
+{
+	if (amdgpu_sriov_vf(adev)) {
+		/*
+		 * GCMC_VM_FB_LOCATION_BASE/TOP is NULL for VF, becuase they are
+		 * VF copy registers so vbios post doesn't program them, for
+		 * SRIOV driver need to program them
+		 */
+		WREG32_SOC15(GC, 0, regGCMC_VM_FB_LOCATION_BASE,
+			     adev->gmc.vram_start >> 24);
+		WREG32_SOC15(GC, 0, regGCMC_VM_FB_LOCATION_TOP,
+			     adev->gmc.vram_end >> 24);
+	}
+
+	/* GART Enable. */
+	gfxhub_v12_0_init_gart_aperture_regs(adev);
+	gfxhub_v12_0_init_system_aperture_regs(adev);
+	gfxhub_v12_0_init_tlb_regs(adev);
+	gfxhub_v12_0_init_cache_regs(adev);
+
+	gfxhub_v12_0_enable_system_domain(adev);
+	gfxhub_v12_0_disable_identity_aperture(adev);
+	gfxhub_v12_0_setup_vmid_config(adev);
+	gfxhub_v12_0_program_invalidation(adev);
+
+	return 0;
+}
+
+static void gfxhub_v12_0_gart_disable(struct amdgpu_device *adev)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)];
+	u32 tmp;
+	u32 i;
+
+	/* Disable all tables */
+	for (i = 0; i < 16; i++)
+		WREG32_SOC15_OFFSET(GC, 0, regGCVM_CONTEXT0_CNTL,
+				    i * hub->ctx_distance, 0);
+
+	/* Setup TLB control */
+	tmp = RREG32_SOC15(GC, 0, regGCMC_VM_MX_L1_TLB_CNTL);
+	tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB, 0);
+	tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL,
+			    ENABLE_ADVANCED_DRIVER_MODEL, 0);
+	WREG32_SOC15(GC, 0, regGCMC_VM_MX_L1_TLB_CNTL, tmp);
+
+	/* Setup L2 cache */
+	WREG32_FIELD15_PREREG(GC, 0, GCVM_L2_CNTL, ENABLE_L2_CACHE, 0);
+	WREG32_SOC15(GC, 0, regGCVM_L2_CNTL3, 0);
+}
+
+/**
+ * gfxhub_v12_0_set_fault_enable_default - update GART/VM fault handling
+ *
+ * @adev: amdgpu_device pointer
+ * @value: true redirects VM faults to the default page
+ */
+static void gfxhub_v12_0_set_fault_enable_default(struct amdgpu_device *adev,
+						  bool value)
+{
+	u32 tmp;
+
+	/* NO halt CP when page fault */
+	tmp = RREG32_SOC15(GC, 0, regCP_DEBUG);
+	tmp = REG_SET_FIELD(tmp, CP_DEBUG, CPG_UTCL1_ERROR_HALT_DISABLE, 1);
+	WREG32_SOC15(GC, 0, regCP_DEBUG, tmp);
+
+	/* These registers are not accessible to VF-SRIOV.
+	 * The PF will program them instead.
+	 */
+	if (amdgpu_sriov_vf(adev))
+		return;
+
+	tmp = RREG32_SOC15(GC, 0, regGCVM_L2_PROTECTION_FAULT_CNTL);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    PDE0_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    PDE1_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    PDE2_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    TRANSLATE_FURTHER_PROTECTION_FAULT_ENABLE_DEFAULT,
+			    value);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    NACK_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    VALID_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    READ_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    WRITE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    EXECUTE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	if (!value) {
+		tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+				CRASH_ON_NO_RETRY_FAULT, 1);
+		tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+				CRASH_ON_RETRY_FAULT, 1);
+	}
+	WREG32_SOC15(GC, 0, regGCVM_L2_PROTECTION_FAULT_CNTL, tmp);
+}
+
+static const struct amdgpu_vmhub_funcs gfxhub_v12_0_vmhub_funcs = {
+	.print_l2_protection_fault_status = gfxhub_v12_0_print_l2_protection_fault_status,
+	.get_invalidate_req = gfxhub_v12_0_get_invalidate_req,
+};
+
+static void gfxhub_v12_0_init(struct amdgpu_device *adev)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)];
+
+	hub->ctx0_ptb_addr_lo32 =
+		SOC15_REG_OFFSET(GC, 0,
+				 regGCVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32);
+	hub->ctx0_ptb_addr_hi32 =
+		SOC15_REG_OFFSET(GC, 0,
+				 regGCVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32);
+	hub->vm_inv_eng0_sem =
+		SOC15_REG_OFFSET(GC, 0, regGCVM_INVALIDATE_ENG0_SEM);
+	hub->vm_inv_eng0_req =
+		SOC15_REG_OFFSET(GC, 0, regGCVM_INVALIDATE_ENG0_REQ);
+	hub->vm_inv_eng0_ack =
+		SOC15_REG_OFFSET(GC, 0, regGCVM_INVALIDATE_ENG0_ACK);
+	hub->vm_context0_cntl =
+		SOC15_REG_OFFSET(GC, 0, regGCVM_CONTEXT0_CNTL);
+	hub->vm_l2_pro_fault_status =
+		SOC15_REG_OFFSET(GC, 0, regGCVM_L2_PROTECTION_FAULT_STATUS_LO32);
+	hub->vm_l2_pro_fault_cntl =
+		SOC15_REG_OFFSET(GC, 0, regGCVM_L2_PROTECTION_FAULT_CNTL);
+
+	hub->ctx_distance = regGCVM_CONTEXT1_CNTL - regGCVM_CONTEXT0_CNTL;
+	hub->ctx_addr_distance = regGCVM_CONTEXT1_PAGE_TABLE_BASE_ADDR_LO32 -
+		regGCVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32;
+	hub->eng_distance = regGCVM_INVALIDATE_ENG1_REQ -
+		regGCVM_INVALIDATE_ENG0_REQ;
+	hub->eng_addr_distance = regGCVM_INVALIDATE_ENG1_ADDR_RANGE_LO32 -
+		regGCVM_INVALIDATE_ENG0_ADDR_RANGE_LO32;
+
+	hub->vm_cntx_cntl_vm_fault = GCVM_CONTEXT1_CNTL__RANGE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		GCVM_CONTEXT1_CNTL__DUMMY_PAGE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		GCVM_CONTEXT1_CNTL__PDE0_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		GCVM_CONTEXT1_CNTL__VALID_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		GCVM_CONTEXT1_CNTL__READ_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		GCVM_CONTEXT1_CNTL__WRITE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		GCVM_CONTEXT1_CNTL__EXECUTE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK;
+
+	hub->vmhub_funcs = &gfxhub_v12_0_vmhub_funcs;
+}
+
+const struct amdgpu_gfxhub_funcs gfxhub_v12_0_funcs = {
+	.get_fb_location = gfxhub_v12_0_get_fb_location,
+	.get_mc_fb_offset = gfxhub_v12_0_get_mc_fb_offset,
+	.setup_vm_pt_regs = gfxhub_v12_0_setup_vm_pt_regs,
+	.gart_enable = gfxhub_v12_0_gart_enable,
+	.gart_disable = gfxhub_v12_0_gart_disable,
+	.set_fault_enable_default = gfxhub_v12_0_set_fault_enable_default,
+	.init = gfxhub_v12_0_init,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/gfxhub_v12_0.h b/drivers/gpu/drm/amd/amdgpu/gfxhub_v12_0.h
new file mode 100644
index 000000000000..f1258265f802
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/gfxhub_v12_0.h
@@ -0,0 +1,29 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __GFXHUB_V12_0_H__
+#define __GFXHUB_V12_0_H__
+
+extern const struct amdgpu_gfxhub_funcs gfxhub_v12_0_funcs;
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c b/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c
index ec4d5e15b766..a7bfc9f41d0e 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c
@@ -40,7 +40,7 @@ static void gfxhub_v1_0_setup_vm_pt_regs(struct amdgpu_device *adev,
 					 uint32_t vmid,
 					 uint64_t page_table_base)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0];
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)];
 
 	WREG32_SOC15_OFFSET(GC, 0, mmVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32,
 			    hub->ctx_addr_distance * vmid,
@@ -92,18 +92,20 @@ static void gfxhub_v1_0_init_system_aperture_regs(struct amdgpu_device *adev)
 {
 	uint64_t value;
 
-	/* Program the AGP BAR */
-	WREG32_SOC15_RLC(GC, 0, mmMC_VM_AGP_BASE, 0);
-	WREG32_SOC15_RLC(GC, 0, mmMC_VM_AGP_BOT, adev->gmc.agp_start >> 24);
-	WREG32_SOC15_RLC(GC, 0, mmMC_VM_AGP_TOP, adev->gmc.agp_end >> 24);
-
 	if (!amdgpu_sriov_vf(adev) || adev->asic_type <= CHIP_VEGA10) {
+		/* Program the AGP BAR */
+		WREG32_SOC15_RLC(GC, 0, mmMC_VM_AGP_BASE, 0);
+		WREG32_SOC15_RLC(GC, 0, mmMC_VM_AGP_BOT, adev->gmc.agp_start >> 24);
+		WREG32_SOC15_RLC(GC, 0, mmMC_VM_AGP_TOP, adev->gmc.agp_end >> 24);
+
 		/* Program the system aperture low logical page number. */
 		WREG32_SOC15_RLC(GC, 0, mmMC_VM_SYSTEM_APERTURE_LOW_ADDR,
 			min(adev->gmc.fb_start, adev->gmc.agp_start) >> 18);
 
-		if (adev->apu_flags & AMD_APU_IS_RAVEN2)
-			/*
+		if (adev->apu_flags & (AMD_APU_IS_RAVEN2 |
+				       AMD_APU_IS_RENOIR |
+				       AMD_APU_IS_GREEN_SARDINE))
+		       /*
 			* Raven2 has a HW issue that it is unable to use the
 			* vram which is out of MC_VM_SYSTEM_APERTURE_HIGH_ADDR.
 			* So here is the workaround that increase system
@@ -120,7 +122,7 @@ static void gfxhub_v1_0_init_system_aperture_regs(struct amdgpu_device *adev)
 				max(adev->gmc.fb_end, adev->gmc.agp_end) >> 18);
 
 		/* Set default page address. */
-		value = amdgpu_gmc_vram_mc2pa(adev, adev->vram_scratch.gpu_addr);
+		value = amdgpu_gmc_vram_mc2pa(adev, adev->mem_scratch.gpu_addr);
 		WREG32_SOC15(GC, 0, mmMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB,
 			     (u32)(value >> 12));
 		WREG32_SOC15(GC, 0, mmMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_MSB,
@@ -247,8 +249,8 @@ static void gfxhub_v1_0_disable_identity_aperture(struct amdgpu_device *adev)
 
 static void gfxhub_v1_0_setup_vmid_config(struct amdgpu_device *adev)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0];
-	unsigned num_level, block_size;
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)];
+	unsigned int num_level, block_size;
 	uint32_t tmp;
 	int i;
 
@@ -260,7 +262,7 @@ static void gfxhub_v1_0_setup_vmid_config(struct amdgpu_device *adev)
 		block_size -= 9;
 
 	for (i = 0; i <= 14; i++) {
-		tmp = RREG32_SOC15_OFFSET(GC, 0, mmVM_CONTEXT1_CNTL, i);
+		tmp = RREG32_SOC15_OFFSET(GC, 0, mmVM_CONTEXT1_CNTL, i * hub->ctx_distance);
 		tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, ENABLE_CONTEXT, 1);
 		tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, PAGE_TABLE_DEPTH,
 				    num_level);
@@ -307,8 +309,8 @@ static void gfxhub_v1_0_setup_vmid_config(struct amdgpu_device *adev)
 
 static void gfxhub_v1_0_program_invalidation(struct amdgpu_device *adev)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0];
-	unsigned i;
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)];
+	unsigned int i;
 
 	for (i = 0 ; i < 18; ++i) {
 		WREG32_SOC15_OFFSET(GC, 0, mmVM_INVALIDATE_ENG0_ADDR_RANGE_LO32,
@@ -338,7 +340,7 @@ static int gfxhub_v1_0_gart_enable(struct amdgpu_device *adev)
 
 static void gfxhub_v1_0_gart_disable(struct amdgpu_device *adev)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0];
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)];
 	u32 tmp;
 	u32 i;
 
@@ -375,6 +377,7 @@ static void gfxhub_v1_0_set_fault_enable_default(struct amdgpu_device *adev,
 						 bool value)
 {
 	u32 tmp;
+
 	tmp = RREG32_SOC15(GC, 0, mmVM_L2_PROTECTION_FAULT_CNTL);
 	tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL,
 			RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
@@ -411,7 +414,7 @@ static void gfxhub_v1_0_set_fault_enable_default(struct amdgpu_device *adev,
 
 static void gfxhub_v1_0_init(struct amdgpu_device *adev)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0];
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)];
 
 	hub->ctx0_ptb_addr_lo32 =
 		SOC15_REG_OFFSET(GC, 0,
@@ -440,7 +443,6 @@ static void gfxhub_v1_0_init(struct amdgpu_device *adev)
 		mmVM_INVALIDATE_ENG0_ADDR_RANGE_LO32;
 }
 
-
 const struct amdgpu_gfxhub_funcs gfxhub_v1_0_funcs = {
 	.get_mc_fb_offset = gfxhub_v1_0_get_mc_fb_offset,
 	.setup_vm_pt_regs = gfxhub_v1_0_setup_vm_pt_regs,
diff --git a/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_2.c b/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_2.c
new file mode 100644
index 000000000000..6c03bf9f1ae8
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_2.c
@@ -0,0 +1,674 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include "amdgpu.h"
+#include "amdgpu_xcp.h"
+#include "gfxhub_v1_2.h"
+#include "gfxhub_v1_1.h"
+
+#include "gc/gc_9_4_3_offset.h"
+#include "gc/gc_9_4_3_sh_mask.h"
+#include "vega10_enum.h"
+
+#include "soc15_common.h"
+
+#define regVM_L2_CNTL3_DEFAULT	0x80100007
+#define regVM_L2_CNTL4_DEFAULT	0x000000c1
+
+static u64 gfxhub_v1_2_get_mc_fb_offset(struct amdgpu_device *adev)
+{
+	return (u64)RREG32_SOC15(GC, GET_INST(GC, 0), regMC_VM_FB_OFFSET) << 24;
+}
+
+static void gfxhub_v1_2_xcc_setup_vm_pt_regs(struct amdgpu_device *adev,
+					     uint32_t vmid,
+					     uint64_t page_table_base,
+					     uint32_t xcc_mask)
+{
+	struct amdgpu_vmhub *hub;
+	int i;
+
+	for_each_inst(i, xcc_mask) {
+		hub = &adev->vmhub[AMDGPU_GFXHUB(i)];
+		WREG32_SOC15_OFFSET(GC, GET_INST(GC, i),
+				    regVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32,
+				    hub->ctx_addr_distance * vmid,
+				    lower_32_bits(page_table_base));
+
+		WREG32_SOC15_OFFSET(GC, GET_INST(GC, i),
+				    regVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32,
+				    hub->ctx_addr_distance * vmid,
+				    upper_32_bits(page_table_base));
+	}
+}
+
+static void gfxhub_v1_2_setup_vm_pt_regs(struct amdgpu_device *adev,
+					 uint32_t vmid,
+					 uint64_t page_table_base)
+{
+	uint32_t xcc_mask;
+
+	xcc_mask = GENMASK(NUM_XCC(adev->gfx.xcc_mask) - 1, 0);
+	gfxhub_v1_2_xcc_setup_vm_pt_regs(adev, vmid, page_table_base, xcc_mask);
+}
+
+static void gfxhub_v1_2_xcc_init_gart_aperture_regs(struct amdgpu_device *adev,
+						    uint32_t xcc_mask)
+{
+	uint64_t gart_start = amdgpu_virt_xgmi_migrate_enabled(adev) ?
+			adev->gmc.vram_start : adev->gmc.fb_start;
+	uint64_t pt_base;
+	int i;
+
+	if (adev->gmc.pdb0_bo)
+		pt_base = amdgpu_gmc_pd_addr(adev->gmc.pdb0_bo);
+	else
+		pt_base = amdgpu_gmc_pd_addr(adev->gart.bo);
+
+	gfxhub_v1_2_xcc_setup_vm_pt_regs(adev, 0, pt_base, xcc_mask);
+
+	/* If use GART for FB translation, vmid0 page table covers both
+	 * vram and system memory (gart)
+	 */
+	for_each_inst(i, xcc_mask) {
+		if (adev->gmc.pdb0_bo) {
+			WREG32_SOC15(GC, GET_INST(GC, i),
+				     regVM_CONTEXT0_PAGE_TABLE_START_ADDR_LO32,
+				     (u32)(gart_start >> 12));
+			WREG32_SOC15(GC, GET_INST(GC, i),
+				     regVM_CONTEXT0_PAGE_TABLE_START_ADDR_HI32,
+				     (u32)(gart_start >> 44));
+
+			WREG32_SOC15(GC, GET_INST(GC, i),
+				     regVM_CONTEXT0_PAGE_TABLE_END_ADDR_LO32,
+				     (u32)(adev->gmc.gart_end >> 12));
+			WREG32_SOC15(GC, GET_INST(GC, i),
+				     regVM_CONTEXT0_PAGE_TABLE_END_ADDR_HI32,
+				     (u32)(adev->gmc.gart_end >> 44));
+		} else {
+			WREG32_SOC15(GC, GET_INST(GC, i),
+				     regVM_CONTEXT0_PAGE_TABLE_START_ADDR_LO32,
+				     (u32)(adev->gmc.gart_start >> 12));
+			WREG32_SOC15(GC, GET_INST(GC, i),
+				     regVM_CONTEXT0_PAGE_TABLE_START_ADDR_HI32,
+				     (u32)(adev->gmc.gart_start >> 44));
+
+			WREG32_SOC15(GC, GET_INST(GC, i),
+				     regVM_CONTEXT0_PAGE_TABLE_END_ADDR_LO32,
+				     (u32)(adev->gmc.gart_end >> 12));
+			WREG32_SOC15(GC, GET_INST(GC, i),
+				     regVM_CONTEXT0_PAGE_TABLE_END_ADDR_HI32,
+				     (u32)(adev->gmc.gart_end >> 44));
+		}
+	}
+}
+
+static void
+gfxhub_v1_2_xcc_init_system_aperture_regs(struct amdgpu_device *adev,
+					  uint32_t xcc_mask)
+{
+	uint64_t value;
+	uint32_t tmp;
+	int i;
+
+	for_each_inst(i, xcc_mask) {
+		/* Program the AGP BAR */
+		WREG32_SOC15_RLC(GC, GET_INST(GC, i), regMC_VM_AGP_BASE, 0);
+		WREG32_SOC15_RLC(GC, GET_INST(GC, i), regMC_VM_AGP_BOT, adev->gmc.agp_start >> 24);
+		WREG32_SOC15_RLC(GC, GET_INST(GC, i), regMC_VM_AGP_TOP, adev->gmc.agp_end >> 24);
+
+		if (!amdgpu_sriov_vf(adev) || adev->asic_type <= CHIP_VEGA10) {
+			/* Program the system aperture low logical page number. */
+			WREG32_SOC15_RLC(GC, GET_INST(GC, i), regMC_VM_SYSTEM_APERTURE_LOW_ADDR,
+				min(adev->gmc.fb_start, adev->gmc.agp_start) >> 18);
+
+			if (adev->apu_flags & (AMD_APU_IS_RAVEN2 |
+					       AMD_APU_IS_RENOIR |
+					       AMD_APU_IS_GREEN_SARDINE))
+			       /*
+				* Raven2 has a HW issue that it is unable to use the
+				* vram which is out of MC_VM_SYSTEM_APERTURE_HIGH_ADDR.
+				* So here is the workaround that increase system
+				* aperture high address (add 1) to get rid of the VM
+				* fault and hardware hang.
+				*/
+				WREG32_SOC15_RLC(GC, GET_INST(GC, i),
+						 regMC_VM_SYSTEM_APERTURE_HIGH_ADDR,
+						 max((adev->gmc.fb_end >> 18) + 0x1,
+						     adev->gmc.agp_end >> 18));
+			else
+				WREG32_SOC15_RLC(GC, GET_INST(GC, i),
+					regMC_VM_SYSTEM_APERTURE_HIGH_ADDR,
+					max(adev->gmc.fb_end, adev->gmc.agp_end) >> 18);
+
+			/* Set default page address. */
+			value = amdgpu_gmc_vram_mc2pa(adev, adev->mem_scratch.gpu_addr);
+			WREG32_SOC15(GC, GET_INST(GC, i), regMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB,
+				     (u32)(value >> 12));
+			WREG32_SOC15(GC, GET_INST(GC, i), regMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_MSB,
+				     (u32)(value >> 44));
+
+			/* Program "protection fault". */
+			WREG32_SOC15(GC, GET_INST(GC, i), regVM_L2_PROTECTION_FAULT_DEFAULT_ADDR_LO32,
+				     (u32)(adev->dummy_page_addr >> 12));
+			WREG32_SOC15(GC, GET_INST(GC, i), regVM_L2_PROTECTION_FAULT_DEFAULT_ADDR_HI32,
+				     (u32)((u64)adev->dummy_page_addr >> 44));
+
+			tmp = RREG32_SOC15(GC, GET_INST(GC, i), regVM_L2_PROTECTION_FAULT_CNTL2);
+			tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL2,
+					    ACTIVE_PAGE_MIGRATION_PTE_READ_RETRY, 1);
+			WREG32_SOC15(GC, GET_INST(GC, i), regVM_L2_PROTECTION_FAULT_CNTL2, tmp);
+		}
+
+		/* In the case squeezing vram into GART aperture, we don't use
+		 * FB aperture and AGP aperture. Disable them.
+		 */
+		if (adev->gmc.pdb0_bo && adev->gmc.xgmi.connected_to_cpu) {
+			WREG32_SOC15(GC, GET_INST(GC, i), regMC_VM_FB_LOCATION_TOP, 0);
+			WREG32_SOC15(GC, GET_INST(GC, i), regMC_VM_FB_LOCATION_BASE, 0x00FFFFFF);
+			WREG32_SOC15(GC, GET_INST(GC, i), regMC_VM_AGP_TOP, 0);
+			WREG32_SOC15(GC, GET_INST(GC, i), regMC_VM_AGP_BOT, 0xFFFFFF);
+			WREG32_SOC15(GC, GET_INST(GC, i), regMC_VM_SYSTEM_APERTURE_LOW_ADDR, 0x3FFFFFFF);
+			WREG32_SOC15(GC, GET_INST(GC, i), regMC_VM_SYSTEM_APERTURE_HIGH_ADDR, 0);
+		}
+	}
+}
+
+static void gfxhub_v1_2_xcc_init_tlb_regs(struct amdgpu_device *adev,
+					  uint32_t xcc_mask)
+{
+	uint32_t tmp;
+	int i;
+
+	for_each_inst(i, xcc_mask) {
+		/* Setup TLB control */
+		tmp = RREG32_SOC15(GC, GET_INST(GC, i), regMC_VM_MX_L1_TLB_CNTL);
+
+		tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL,
+				    ENABLE_L1_TLB, 1);
+		tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL,
+				    SYSTEM_ACCESS_MODE, 3);
+		tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL,
+				    ENABLE_ADVANCED_DRIVER_MODEL, 1);
+		tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL,
+				    SYSTEM_APERTURE_UNMAPPED_ACCESS, 0);
+		tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL,
+				    MTYPE, MTYPE_UC);/* XXX for emulation. */
+		tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ATC_EN, 1);
+
+		WREG32_SOC15_RLC(GC, GET_INST(GC, i), regMC_VM_MX_L1_TLB_CNTL, tmp);
+	}
+}
+
+static void gfxhub_v1_2_xcc_init_cache_regs(struct amdgpu_device *adev,
+					    uint32_t xcc_mask)
+{
+	uint32_t tmp;
+	int i;
+
+	for_each_inst(i, xcc_mask) {
+		/* Setup L2 cache */
+		tmp = RREG32_SOC15(GC, GET_INST(GC, i), regVM_L2_CNTL);
+		tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, ENABLE_L2_CACHE, 1);
+		tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, ENABLE_L2_FRAGMENT_PROCESSING, 1);
+		/* XXX for emulation, Refer to closed source code.*/
+		tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, L2_PDE0_CACHE_TAG_GENERATION_MODE,
+				    0);
+		tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, PDE_FAULT_CLASSIFICATION, 0);
+		tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, CONTEXT1_IDENTITY_ACCESS_MODE, 1);
+		tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, IDENTITY_MODE_FRAGMENT_SIZE, 0);
+		WREG32_SOC15_RLC(GC, GET_INST(GC, i), regVM_L2_CNTL, tmp);
+
+		tmp = RREG32_SOC15(GC, GET_INST(GC, i), regVM_L2_CNTL2);
+		tmp = REG_SET_FIELD(tmp, VM_L2_CNTL2, INVALIDATE_ALL_L1_TLBS, 1);
+		tmp = REG_SET_FIELD(tmp, VM_L2_CNTL2, INVALIDATE_L2_CACHE, 1);
+		WREG32_SOC15_RLC(GC, GET_INST(GC, i), regVM_L2_CNTL2, tmp);
+
+		tmp = regVM_L2_CNTL3_DEFAULT;
+		if (adev->gmc.translate_further) {
+			tmp = REG_SET_FIELD(tmp, VM_L2_CNTL3, BANK_SELECT, 12);
+			tmp = REG_SET_FIELD(tmp, VM_L2_CNTL3,
+					    L2_CACHE_BIGK_FRAGMENT_SIZE, 9);
+		} else {
+			tmp = REG_SET_FIELD(tmp, VM_L2_CNTL3, BANK_SELECT, 9);
+			tmp = REG_SET_FIELD(tmp, VM_L2_CNTL3,
+					    L2_CACHE_BIGK_FRAGMENT_SIZE, 6);
+		}
+		WREG32_SOC15_RLC(GC, GET_INST(GC, i), regVM_L2_CNTL3, tmp);
+
+		tmp = regVM_L2_CNTL4_DEFAULT;
+		/* For AMD APP APUs setup WC memory */
+		if (adev->gmc.xgmi.connected_to_cpu || adev->gmc.is_app_apu) {
+			tmp = REG_SET_FIELD(tmp, VM_L2_CNTL4, VMC_TAP_PDE_REQUEST_PHYSICAL, 1);
+			tmp = REG_SET_FIELD(tmp, VM_L2_CNTL4, VMC_TAP_PTE_REQUEST_PHYSICAL, 1);
+		} else {
+			tmp = REG_SET_FIELD(tmp, VM_L2_CNTL4, VMC_TAP_PDE_REQUEST_PHYSICAL, 0);
+			tmp = REG_SET_FIELD(tmp, VM_L2_CNTL4, VMC_TAP_PTE_REQUEST_PHYSICAL, 0);
+		}
+		WREG32_SOC15_RLC(GC, GET_INST(GC, i), regVM_L2_CNTL4, tmp);
+	}
+}
+
+static void gfxhub_v1_2_xcc_enable_system_domain(struct amdgpu_device *adev,
+						 uint32_t xcc_mask)
+{
+	uint32_t tmp;
+	int i;
+
+	for_each_inst(i, xcc_mask) {
+		tmp = RREG32_SOC15(GC, GET_INST(GC, i), regVM_CONTEXT0_CNTL);
+		tmp = REG_SET_FIELD(tmp, VM_CONTEXT0_CNTL, ENABLE_CONTEXT, 1);
+		tmp = REG_SET_FIELD(tmp, VM_CONTEXT0_CNTL, PAGE_TABLE_DEPTH,
+				adev->gmc.vmid0_page_table_depth);
+		tmp = REG_SET_FIELD(tmp, VM_CONTEXT0_CNTL, PAGE_TABLE_BLOCK_SIZE,
+				adev->gmc.vmid0_page_table_block_size);
+		tmp = REG_SET_FIELD(tmp, VM_CONTEXT0_CNTL,
+				    RETRY_PERMISSION_OR_INVALID_PAGE_FAULT, 0);
+		WREG32_SOC15(GC, GET_INST(GC, i), regVM_CONTEXT0_CNTL, tmp);
+	}
+}
+
+static void
+gfxhub_v1_2_xcc_disable_identity_aperture(struct amdgpu_device *adev,
+					  uint32_t xcc_mask)
+{
+	int i;
+
+	for_each_inst(i, xcc_mask) {
+		WREG32_SOC15(GC, GET_INST(GC, i),
+			     regVM_L2_CONTEXT1_IDENTITY_APERTURE_LOW_ADDR_LO32,
+			     0XFFFFFFFF);
+		WREG32_SOC15(GC, GET_INST(GC, i),
+			     regVM_L2_CONTEXT1_IDENTITY_APERTURE_LOW_ADDR_HI32,
+			     0x0000000F);
+
+		WREG32_SOC15(GC, GET_INST(GC, i),
+			     regVM_L2_CONTEXT1_IDENTITY_APERTURE_HIGH_ADDR_LO32,
+			     0);
+		WREG32_SOC15(GC, GET_INST(GC, i),
+			     regVM_L2_CONTEXT1_IDENTITY_APERTURE_HIGH_ADDR_HI32,
+			     0);
+
+		WREG32_SOC15(GC, GET_INST(GC, i),
+			     regVM_L2_CONTEXT_IDENTITY_PHYSICAL_OFFSET_LO32, 0);
+		WREG32_SOC15(GC, GET_INST(GC, i),
+			     regVM_L2_CONTEXT_IDENTITY_PHYSICAL_OFFSET_HI32, 0);
+	}
+}
+
+static inline bool
+gfxhub_v1_2_per_process_xnack_support(struct amdgpu_device *adev)
+{
+	/*
+	 * TODO: Check if this function is really needed, so far only 9.4.3
+	 * variants use GFXHUB 1.2
+	 */
+	return !!adev->aid_mask;
+}
+
+static void gfxhub_v1_2_xcc_setup_vmid_config(struct amdgpu_device *adev,
+					      uint32_t xcc_mask)
+{
+	struct amdgpu_vmhub *hub;
+	unsigned int num_level, block_size;
+	uint32_t tmp;
+	int i, j;
+
+	num_level = adev->vm_manager.num_level;
+	block_size = adev->vm_manager.block_size;
+	if (adev->gmc.translate_further)
+		num_level -= 1;
+	else
+		block_size -= 9;
+
+	for_each_inst(j, xcc_mask) {
+		hub = &adev->vmhub[AMDGPU_GFXHUB(j)];
+		for (i = 0; i <= 14; i++) {
+			tmp = RREG32_SOC15_OFFSET(GC, GET_INST(GC, j), regVM_CONTEXT1_CNTL,
+					i * hub->ctx_distance);
+			tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, ENABLE_CONTEXT, 1);
+			tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, PAGE_TABLE_DEPTH,
+					    num_level);
+			tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL,
+					    RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+			tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL,
+					    DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT,
+					    1);
+			tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL,
+					    PDE0_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+			tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL,
+					    VALID_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+			tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL,
+					    READ_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+			tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL,
+					    WRITE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+			tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL,
+					    EXECUTE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+			tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL,
+					    PAGE_TABLE_BLOCK_SIZE,
+					    block_size);
+			/* Send no-retry XNACK on fault to suppress VM fault storm.
+			 * On 9.4.3 variants, XNACK can be enabled in
+			 * the SQ per-process.
+			 * Retry faults need to be enabled for that to work.
+			 */
+			tmp = REG_SET_FIELD(
+				tmp, VM_CONTEXT1_CNTL,
+				RETRY_PERMISSION_OR_INVALID_PAGE_FAULT,
+				!adev->gmc.noretry ||
+					gfxhub_v1_2_per_process_xnack_support(
+						adev));
+			WREG32_SOC15_OFFSET(GC, GET_INST(GC, j), regVM_CONTEXT1_CNTL,
+					    i * hub->ctx_distance, tmp);
+			WREG32_SOC15_OFFSET(GC, GET_INST(GC, j),
+					    regVM_CONTEXT1_PAGE_TABLE_START_ADDR_LO32,
+					    i * hub->ctx_addr_distance, 0);
+			WREG32_SOC15_OFFSET(GC, GET_INST(GC, j),
+					    regVM_CONTEXT1_PAGE_TABLE_START_ADDR_HI32,
+					    i * hub->ctx_addr_distance, 0);
+			WREG32_SOC15_OFFSET(GC, GET_INST(GC, j),
+					    regVM_CONTEXT1_PAGE_TABLE_END_ADDR_LO32,
+					    i * hub->ctx_addr_distance,
+					    lower_32_bits(adev->vm_manager.max_pfn - 1));
+			WREG32_SOC15_OFFSET(GC, GET_INST(GC, j),
+					    regVM_CONTEXT1_PAGE_TABLE_END_ADDR_HI32,
+					    i * hub->ctx_addr_distance,
+					    upper_32_bits(adev->vm_manager.max_pfn - 1));
+		}
+	}
+}
+
+static void gfxhub_v1_2_xcc_program_invalidation(struct amdgpu_device *adev,
+						 uint32_t xcc_mask)
+{
+	struct amdgpu_vmhub *hub;
+	unsigned int i, j;
+
+	for_each_inst(j, xcc_mask) {
+		hub = &adev->vmhub[AMDGPU_GFXHUB(j)];
+
+		for (i = 0 ; i < 18; ++i) {
+			WREG32_SOC15_OFFSET(GC, GET_INST(GC, j), regVM_INVALIDATE_ENG0_ADDR_RANGE_LO32,
+					    i * hub->eng_addr_distance, 0xffffffff);
+			WREG32_SOC15_OFFSET(GC, GET_INST(GC, j), regVM_INVALIDATE_ENG0_ADDR_RANGE_HI32,
+					    i * hub->eng_addr_distance, 0x1f);
+		}
+	}
+}
+
+static int gfxhub_v1_2_xcc_gart_enable(struct amdgpu_device *adev,
+				       uint32_t xcc_mask)
+{
+	/* GART Enable. */
+	gfxhub_v1_2_xcc_init_gart_aperture_regs(adev, xcc_mask);
+	gfxhub_v1_2_xcc_init_system_aperture_regs(adev, xcc_mask);
+	gfxhub_v1_2_xcc_init_tlb_regs(adev, xcc_mask);
+	if (!amdgpu_sriov_vf(adev))
+		gfxhub_v1_2_xcc_init_cache_regs(adev, xcc_mask);
+
+	gfxhub_v1_2_xcc_enable_system_domain(adev, xcc_mask);
+	if (!amdgpu_sriov_vf(adev))
+		gfxhub_v1_2_xcc_disable_identity_aperture(adev, xcc_mask);
+	gfxhub_v1_2_xcc_setup_vmid_config(adev, xcc_mask);
+	gfxhub_v1_2_xcc_program_invalidation(adev, xcc_mask);
+
+	return 0;
+}
+
+static int gfxhub_v1_2_gart_enable(struct amdgpu_device *adev)
+{
+	uint32_t xcc_mask;
+
+	xcc_mask = GENMASK(NUM_XCC(adev->gfx.xcc_mask) - 1, 0);
+	return gfxhub_v1_2_xcc_gart_enable(adev, xcc_mask);
+}
+
+static void gfxhub_v1_2_xcc_gart_disable(struct amdgpu_device *adev,
+					 uint32_t xcc_mask)
+{
+	struct amdgpu_vmhub *hub;
+	u32 tmp;
+	u32 i, j;
+
+	for_each_inst(j, xcc_mask) {
+		hub = &adev->vmhub[AMDGPU_GFXHUB(j)];
+		/* Disable all tables */
+		for (i = 0; i < 16; i++)
+			WREG32_SOC15_OFFSET(GC, GET_INST(GC, j), regVM_CONTEXT0_CNTL,
+					    i * hub->ctx_distance, 0);
+
+		/* Setup TLB control */
+		tmp = RREG32_SOC15(GC, GET_INST(GC, j), regMC_VM_MX_L1_TLB_CNTL);
+		tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB, 0);
+		tmp = REG_SET_FIELD(tmp,
+					MC_VM_MX_L1_TLB_CNTL,
+					ENABLE_ADVANCED_DRIVER_MODEL,
+					0);
+		WREG32_SOC15_RLC(GC, GET_INST(GC, j), regMC_VM_MX_L1_TLB_CNTL, tmp);
+
+		/* Setup L2 cache */
+		if (!amdgpu_sriov_vf(adev)) {
+			tmp = RREG32_SOC15(GC, GET_INST(GC, j), regVM_L2_CNTL);
+			tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, ENABLE_L2_CACHE, 0);
+			WREG32_SOC15(GC, GET_INST(GC, j), regVM_L2_CNTL, tmp);
+			WREG32_SOC15(GC, GET_INST(GC, j), regVM_L2_CNTL3, 0);
+		}
+	}
+}
+
+static void gfxhub_v1_2_gart_disable(struct amdgpu_device *adev)
+{
+	uint32_t xcc_mask;
+
+	xcc_mask = GENMASK(NUM_XCC(adev->gfx.xcc_mask) - 1, 0);
+	gfxhub_v1_2_xcc_gart_disable(adev, xcc_mask);
+}
+
+static void gfxhub_v1_2_xcc_set_fault_enable_default(struct amdgpu_device *adev,
+						     bool value,
+						     uint32_t xcc_mask)
+{
+	u32 tmp;
+	int i;
+
+	for_each_inst(i, xcc_mask) {
+		tmp = RREG32_SOC15(GC, GET_INST(GC, i), regVM_L2_PROTECTION_FAULT_CNTL);
+		tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL,
+				RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+		tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL,
+				PDE0_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+		tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL,
+				PDE1_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+		tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL,
+				PDE2_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+		tmp = REG_SET_FIELD(tmp,
+				VM_L2_PROTECTION_FAULT_CNTL,
+				TRANSLATE_FURTHER_PROTECTION_FAULT_ENABLE_DEFAULT,
+				value);
+		tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL,
+				NACK_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+		tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL,
+				DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+		tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL,
+				VALID_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+		tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL,
+				READ_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+		tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL,
+				WRITE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+		tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL,
+				EXECUTE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+		if (!value) {
+			tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL,
+					CRASH_ON_NO_RETRY_FAULT, 1);
+			tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL,
+					CRASH_ON_RETRY_FAULT, 1);
+		}
+		WREG32_SOC15(GC, GET_INST(GC, i), regVM_L2_PROTECTION_FAULT_CNTL, tmp);
+	}
+}
+
+/**
+ * gfxhub_v1_2_set_fault_enable_default - update GART/VM fault handling
+ *
+ * @adev: amdgpu_device pointer
+ * @value: true redirects VM faults to the default page
+ */
+static void gfxhub_v1_2_set_fault_enable_default(struct amdgpu_device *adev,
+						 bool value)
+{
+	uint32_t xcc_mask;
+
+	xcc_mask = GENMASK(NUM_XCC(adev->gfx.xcc_mask) - 1, 0);
+	gfxhub_v1_2_xcc_set_fault_enable_default(adev, value, xcc_mask);
+}
+
+static void gfxhub_v1_2_xcc_init(struct amdgpu_device *adev, uint32_t xcc_mask)
+{
+	struct amdgpu_vmhub *hub;
+	int i;
+
+	for_each_inst(i, xcc_mask) {
+		hub = &adev->vmhub[AMDGPU_GFXHUB(i)];
+
+		hub->ctx0_ptb_addr_lo32 =
+			SOC15_REG_OFFSET(GC, GET_INST(GC, i),
+				regVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32);
+		hub->ctx0_ptb_addr_hi32 =
+			SOC15_REG_OFFSET(GC, GET_INST(GC, i),
+				regVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32);
+		hub->vm_inv_eng0_sem =
+			SOC15_REG_OFFSET(GC, GET_INST(GC, i), regVM_INVALIDATE_ENG0_SEM);
+		hub->vm_inv_eng0_req =
+			SOC15_REG_OFFSET(GC, GET_INST(GC, i), regVM_INVALIDATE_ENG0_REQ);
+		hub->vm_inv_eng0_ack =
+			SOC15_REG_OFFSET(GC, GET_INST(GC, i), regVM_INVALIDATE_ENG0_ACK);
+		hub->vm_context0_cntl =
+			SOC15_REG_OFFSET(GC, GET_INST(GC, i), regVM_CONTEXT0_CNTL);
+		hub->vm_l2_pro_fault_status =
+			SOC15_REG_OFFSET(GC, GET_INST(GC, i),
+				regVM_L2_PROTECTION_FAULT_STATUS);
+		hub->vm_l2_pro_fault_cntl =
+			SOC15_REG_OFFSET(GC, GET_INST(GC, i), regVM_L2_PROTECTION_FAULT_CNTL);
+
+		hub->ctx_distance = regVM_CONTEXT1_CNTL -
+				regVM_CONTEXT0_CNTL;
+		hub->ctx_addr_distance =
+				regVM_CONTEXT1_PAGE_TABLE_BASE_ADDR_LO32 -
+				regVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32;
+		hub->eng_distance = regVM_INVALIDATE_ENG1_REQ -
+				regVM_INVALIDATE_ENG0_REQ;
+		hub->eng_addr_distance =
+				regVM_INVALIDATE_ENG1_ADDR_RANGE_LO32 -
+				regVM_INVALIDATE_ENG0_ADDR_RANGE_LO32;
+	}
+}
+
+static void gfxhub_v1_2_init(struct amdgpu_device *adev)
+{
+	uint32_t xcc_mask;
+
+	xcc_mask = GENMASK(NUM_XCC(adev->gfx.xcc_mask) - 1, 0);
+	gfxhub_v1_2_xcc_init(adev, xcc_mask);
+}
+
+static int gfxhub_v1_2_get_xgmi_info(struct amdgpu_device *adev)
+{
+	u32 max_num_physical_nodes;
+	u32 max_physical_node_id;
+	u32 xgmi_lfb_cntl;
+	u32 max_region;
+	u64 seg_size;
+
+	xgmi_lfb_cntl = RREG32_SOC15(GC, GET_INST(GC, 0), regMC_VM_XGMI_LFB_CNTL);
+	seg_size = REG_GET_FIELD(
+		RREG32_SOC15(GC, GET_INST(GC, 0), regMC_VM_XGMI_LFB_SIZE),
+		MC_VM_XGMI_LFB_SIZE, PF_LFB_SIZE) << 24;
+	max_region =
+		REG_GET_FIELD(xgmi_lfb_cntl, MC_VM_XGMI_LFB_CNTL, PF_MAX_REGION);
+
+
+
+	max_num_physical_nodes   = 8;
+	max_physical_node_id     = 7;
+
+	/* PF_MAX_REGION=0 means xgmi is disabled */
+	if (max_region || adev->gmc.xgmi.connected_to_cpu) {
+		adev->gmc.xgmi.num_physical_nodes = max_region + 1;
+
+		if (adev->gmc.xgmi.num_physical_nodes > max_num_physical_nodes)
+			return -EINVAL;
+
+		adev->gmc.xgmi.physical_node_id =
+			REG_GET_FIELD(xgmi_lfb_cntl, MC_VM_XGMI_LFB_CNTL,
+					PF_LFB_REGION);
+
+		if (adev->gmc.xgmi.physical_node_id > max_physical_node_id)
+			return -EINVAL;
+
+		adev->gmc.xgmi.node_segment_size = seg_size;
+	}
+
+	return 0;
+}
+
+const struct amdgpu_gfxhub_funcs gfxhub_v1_2_funcs = {
+	.get_mc_fb_offset = gfxhub_v1_2_get_mc_fb_offset,
+	.setup_vm_pt_regs = gfxhub_v1_2_setup_vm_pt_regs,
+	.gart_enable = gfxhub_v1_2_gart_enable,
+	.gart_disable = gfxhub_v1_2_gart_disable,
+	.set_fault_enable_default = gfxhub_v1_2_set_fault_enable_default,
+	.init = gfxhub_v1_2_init,
+	.get_xgmi_info = gfxhub_v1_2_get_xgmi_info,
+};
+
+static int gfxhub_v1_2_xcp_resume(void *handle, uint32_t inst_mask)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	bool value;
+
+	if (amdgpu_vm_fault_stop == AMDGPU_VM_FAULT_STOP_ALWAYS)
+		value = false;
+	else
+		value = true;
+
+	gfxhub_v1_2_xcc_set_fault_enable_default(adev, value, inst_mask);
+
+	if (!amdgpu_sriov_vf(adev))
+		return gfxhub_v1_2_xcc_gart_enable(adev, inst_mask);
+
+	return 0;
+}
+
+static int gfxhub_v1_2_xcp_suspend(void *handle, uint32_t inst_mask)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+	if (!amdgpu_sriov_vf(adev))
+		gfxhub_v1_2_xcc_gart_disable(adev, inst_mask);
+
+	return 0;
+}
+
+struct amdgpu_xcp_ip_funcs gfxhub_v1_2_xcp_funcs = {
+	.suspend = &gfxhub_v1_2_xcp_suspend,
+	.resume = &gfxhub_v1_2_xcp_resume
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_2.h b/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_2.h
new file mode 100644
index 000000000000..997e9f90c990
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_2.h
@@ -0,0 +1,31 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __GFXHUB_V1_2_H__
+#define __GFXHUB_V1_2_H__
+
+extern const struct amdgpu_gfxhub_funcs gfxhub_v1_2_funcs;
+
+extern struct amdgpu_xcp_ip_funcs gfxhub_v1_2_xcp_funcs;
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/gfxhub_v2_0.c b/drivers/gpu/drm/amd/amdgpu/gfxhub_v2_0.c
index 6e0ace2fbfab..793faf62cb07 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfxhub_v2_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfxhub_v2_0.c
@@ -31,7 +31,7 @@
 
 #include "soc15_common.h"
 
-static const char *gfxhub_client_ids[] = {
+static const char * const gfxhub_client_ids[] = {
 	"CB/DB",
 	"Reserved",
 	"GE1",
@@ -120,7 +120,7 @@ static u64 gfxhub_v2_0_get_mc_fb_offset(struct amdgpu_device *adev)
 static void gfxhub_v2_0_setup_vm_pt_regs(struct amdgpu_device *adev, uint32_t vmid,
 				uint64_t page_table_base)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0];
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)];
 
 	WREG32_SOC15_OFFSET(GC, 0, mmGCVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32,
 			    hub->ctx_addr_distance * vmid,
@@ -165,7 +165,7 @@ static void gfxhub_v2_0_init_system_aperture_regs(struct amdgpu_device *adev)
 			     max(adev->gmc.fb_end, adev->gmc.agp_end) >> 18);
 
 		/* Set default page address. */
-		value = amdgpu_gmc_vram_mc2pa(adev, adev->vram_scratch.gpu_addr);
+		value = amdgpu_gmc_vram_mc2pa(adev, adev->mem_scratch.gpu_addr);
 		WREG32_SOC15(GC, 0, mmGCMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB,
 			     (u32)(value >> 12));
 		WREG32_SOC15(GC, 0, mmGCMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_MSB,
@@ -282,12 +282,12 @@ static void gfxhub_v2_0_disable_identity_aperture(struct amdgpu_device *adev)
 
 static void gfxhub_v2_0_setup_vmid_config(struct amdgpu_device *adev)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0];
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)];
 	int i;
 	uint32_t tmp;
 
 	for (i = 0; i <= 14; i++) {
-		tmp = RREG32_SOC15_OFFSET(GC, 0, mmGCVM_CONTEXT1_CNTL, i);
+		tmp = RREG32_SOC15_OFFSET(GC, 0, mmGCVM_CONTEXT1_CNTL, i * hub->ctx_distance);
 		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL, ENABLE_CONTEXT, 1);
 		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL, PAGE_TABLE_DEPTH,
 				    adev->vm_manager.num_level);
@@ -325,12 +325,14 @@ static void gfxhub_v2_0_setup_vmid_config(struct amdgpu_device *adev)
 				    i * hub->ctx_addr_distance,
 				    upper_32_bits(adev->vm_manager.max_pfn - 1));
 	}
+
+	hub->vm_cntx_cntl = tmp;
 }
 
 static void gfxhub_v2_0_program_invalidation(struct amdgpu_device *adev)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0];
-	unsigned i;
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)];
+	unsigned int i;
 
 	for (i = 0 ; i < 18; ++i) {
 		WREG32_SOC15_OFFSET(GC, 0, mmGCVM_INVALIDATE_ENG0_ADDR_RANGE_LO32,
@@ -358,7 +360,7 @@ static int gfxhub_v2_0_gart_enable(struct amdgpu_device *adev)
 
 static void gfxhub_v2_0_gart_disable(struct amdgpu_device *adev)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0];
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)];
 	u32 tmp;
 	u32 i;
 
@@ -391,6 +393,7 @@ static void gfxhub_v2_0_set_fault_enable_default(struct amdgpu_device *adev,
 					  bool value)
 {
 	u32 tmp;
+
 	tmp = RREG32_SOC15(GC, 0, mmGCVM_L2_PROTECTION_FAULT_CNTL);
 	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
 			    RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
@@ -431,7 +434,7 @@ static const struct amdgpu_vmhub_funcs gfxhub_v2_0_vmhub_funcs = {
 
 static void gfxhub_v2_0_init(struct amdgpu_device *adev)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0];
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)];
 
 	hub->ctx0_ptb_addr_lo32 =
 		SOC15_REG_OFFSET(GC, 0,
@@ -468,6 +471,9 @@ static void gfxhub_v2_0_init(struct amdgpu_device *adev)
 		GCVM_CONTEXT1_CNTL__WRITE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
 		GCVM_CONTEXT1_CNTL__EXECUTE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK;
 
+	/* TODO: This is only needed on some Navi 1x revisions */
+	hub->sdma_invalidation_workaround = true;
+
 	hub->vmhub_funcs = &gfxhub_v2_0_vmhub_funcs;
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/gfxhub_v2_1.c b/drivers/gpu/drm/amd/amdgpu/gfxhub_v2_1.c
index b4eddf6e98a6..deb95fab02df 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfxhub_v2_1.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfxhub_v2_1.c
@@ -34,7 +34,7 @@
 #define mmGCUTCL2_HARVEST_BYPASS_GROUPS_YELLOW_CARP				0x16f8
 #define mmGCUTCL2_HARVEST_BYPASS_GROUPS_YELLOW_CARP_BASE_IDX	0
 
-static const char *gfxhub_client_ids[] = {
+static const char * const gfxhub_client_ids[] = {
 	"CB/DB",
 	"Reserved",
 	"GE1",
@@ -123,7 +123,7 @@ static u64 gfxhub_v2_1_get_mc_fb_offset(struct amdgpu_device *adev)
 static void gfxhub_v2_1_setup_vm_pt_regs(struct amdgpu_device *adev, uint32_t vmid,
 				uint64_t page_table_base)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0];
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)];
 
 	WREG32_SOC15_OFFSET(GC, 0, mmGCVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32,
 			    hub->ctx_addr_distance * vmid,
@@ -155,6 +155,9 @@ static void gfxhub_v2_1_init_system_aperture_regs(struct amdgpu_device *adev)
 {
 	uint64_t value;
 
+	if (amdgpu_sriov_vf(adev))
+		return;
+
 	/* Program the AGP BAR */
 	WREG32_SOC15(GC, 0, mmGCMC_VM_AGP_BASE, 0);
 	WREG32_SOC15(GC, 0, mmGCMC_VM_AGP_BOT, adev->gmc.agp_start >> 24);
@@ -167,7 +170,7 @@ static void gfxhub_v2_1_init_system_aperture_regs(struct amdgpu_device *adev)
 		     max(adev->gmc.fb_end, adev->gmc.agp_end) >> 18);
 
 	/* Set default page address. */
-	value = amdgpu_gmc_vram_mc2pa(adev, adev->vram_scratch.gpu_addr);
+	value = amdgpu_gmc_vram_mc2pa(adev, adev->mem_scratch.gpu_addr);
 	WREG32_SOC15(GC, 0, mmGCMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB,
 		     (u32)(value >> 12));
 	WREG32_SOC15(GC, 0, mmGCMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_MSB,
@@ -291,12 +294,12 @@ static void gfxhub_v2_1_disable_identity_aperture(struct amdgpu_device *adev)
 
 static void gfxhub_v2_1_setup_vmid_config(struct amdgpu_device *adev)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0];
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)];
 	int i;
 	uint32_t tmp;
 
 	for (i = 0; i <= 14; i++) {
-		tmp = RREG32_SOC15_OFFSET(GC, 0, mmGCVM_CONTEXT1_CNTL, i);
+		tmp = RREG32_SOC15_OFFSET(GC, 0, mmGCVM_CONTEXT1_CNTL, i * hub->ctx_distance);
 		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL, ENABLE_CONTEXT, 1);
 		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL, PAGE_TABLE_DEPTH,
 				    adev->vm_manager.num_level);
@@ -334,12 +337,14 @@ static void gfxhub_v2_1_setup_vmid_config(struct amdgpu_device *adev)
 				    i * hub->ctx_addr_distance,
 				    upper_32_bits(adev->vm_manager.max_pfn - 1));
 	}
+
+	hub->vm_cntx_cntl = tmp;
 }
 
 static void gfxhub_v2_1_program_invalidation(struct amdgpu_device *adev)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0];
-	unsigned i;
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)];
+	unsigned int i;
 
 	for (i = 0 ; i < 18; ++i) {
 		WREG32_SOC15_OFFSET(GC, 0, mmGCVM_INVALIDATE_ENG0_ADDR_RANGE_LO32,
@@ -379,7 +384,7 @@ static int gfxhub_v2_1_gart_enable(struct amdgpu_device *adev)
 
 static void gfxhub_v2_1_gart_disable(struct amdgpu_device *adev)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0];
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)];
 	u32 tmp;
 	u32 i;
 
@@ -395,6 +400,9 @@ static void gfxhub_v2_1_gart_disable(struct amdgpu_device *adev)
 			    ENABLE_ADVANCED_DRIVER_MODEL, 0);
 	WREG32_SOC15(GC, 0, mmGCMC_VM_MX_L1_TLB_CNTL, tmp);
 
+	if (amdgpu_sriov_vf(adev))
+		return;
+
 	/* Setup L2 cache */
 	WREG32_FIELD15(GC, 0, GCVM_L2_CNTL, ENABLE_L2_CACHE, 0);
 	WREG32_SOC15(GC, 0, mmGCVM_L2_CNTL3, 0);
@@ -457,7 +465,7 @@ static const struct amdgpu_vmhub_funcs gfxhub_v2_1_vmhub_funcs = {
 
 static void gfxhub_v2_1_init(struct amdgpu_device *adev)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0];
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)];
 
 	hub->ctx0_ptb_addr_lo32 =
 		SOC15_REG_OFFSET(GC, 0,
@@ -497,42 +505,6 @@ static void gfxhub_v2_1_init(struct amdgpu_device *adev)
 	hub->vmhub_funcs = &gfxhub_v2_1_vmhub_funcs;
 }
 
-static int gfxhub_v2_1_get_xgmi_info(struct amdgpu_device *adev)
-{
-	u32 xgmi_lfb_cntl = RREG32_SOC15(GC, 0, mmGCMC_VM_XGMI_LFB_CNTL);
-	u32 max_region =
-		REG_GET_FIELD(xgmi_lfb_cntl, GCMC_VM_XGMI_LFB_CNTL, PF_MAX_REGION);
-	u32 max_num_physical_nodes   = 0;
-	u32 max_physical_node_id     = 0;
-
-	switch (adev->ip_versions[XGMI_HWIP][0]) {
-	case IP_VERSION(4, 8, 0):
-		max_num_physical_nodes   = 4;
-		max_physical_node_id     = 3;
-		break;
-	default:
-		return -EINVAL;
-	}
-
-	/* PF_MAX_REGION=0 means xgmi is disabled */
-	if (max_region) {
-		adev->gmc.xgmi.num_physical_nodes = max_region + 1;
-		if (adev->gmc.xgmi.num_physical_nodes > max_num_physical_nodes)
-			return -EINVAL;
-
-		adev->gmc.xgmi.physical_node_id =
-			REG_GET_FIELD(xgmi_lfb_cntl, GCMC_VM_XGMI_LFB_CNTL, PF_LFB_REGION);
-		if (adev->gmc.xgmi.physical_node_id > max_physical_node_id)
-			return -EINVAL;
-
-		adev->gmc.xgmi.node_segment_size = REG_GET_FIELD(
-			RREG32_SOC15(GC, 0, mmGCMC_VM_XGMI_LFB_SIZE),
-			GCMC_VM_XGMI_LFB_SIZE, PF_LFB_SIZE) << 24;
-	}
-
-	return 0;
-}
-
 static void gfxhub_v2_1_utcl2_harvest(struct amdgpu_device *adev)
 {
 	int i;
@@ -543,7 +515,9 @@ static void gfxhub_v2_1_utcl2_harvest(struct amdgpu_device *adev)
 		adev->gfx.config.max_sh_per_se *
 		adev->gfx.config.max_shader_engines);
 
-	if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 3, 3)) {
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	case IP_VERSION(10, 3, 1):
+	case IP_VERSION(10, 3, 3):
 		/* Get SA disabled bitmap from eFuse setting */
 		efuse_setting = RREG32_SOC15(GC, 0, mmCC_GC_SA_UNIT_DISABLE);
 		efuse_setting &= CC_GC_SA_UNIT_DISABLE__SA_DISABLE_MASK;
@@ -566,9 +540,118 @@ static void gfxhub_v2_1_utcl2_harvest(struct amdgpu_device *adev)
 		disabled_sa = tmp;
 
 		WREG32_SOC15(GC, 0, mmGCUTCL2_HARVEST_BYPASS_GROUPS_YELLOW_CARP, disabled_sa);
+		break;
+	default:
+		break;
 	}
 }
 
+static void gfxhub_v2_1_save_regs(struct amdgpu_device *adev)
+{
+	int i;
+
+	adev->gmc.VM_L2_CNTL = RREG32_SOC15(GC, 0, mmGCVM_L2_CNTL);
+	adev->gmc.VM_L2_CNTL2 = RREG32_SOC15(GC, 0, mmGCVM_L2_CNTL2);
+	adev->gmc.VM_DUMMY_PAGE_FAULT_CNTL = RREG32_SOC15(GC, 0, mmGCVM_DUMMY_PAGE_FAULT_CNTL);
+	adev->gmc.VM_DUMMY_PAGE_FAULT_ADDR_LO32 = RREG32_SOC15(GC, 0, mmGCVM_DUMMY_PAGE_FAULT_ADDR_LO32);
+	adev->gmc.VM_DUMMY_PAGE_FAULT_ADDR_HI32 = RREG32_SOC15(GC, 0, mmGCVM_DUMMY_PAGE_FAULT_ADDR_HI32);
+	adev->gmc.VM_L2_PROTECTION_FAULT_CNTL = RREG32_SOC15(GC, 0, mmGCVM_L2_PROTECTION_FAULT_CNTL);
+	adev->gmc.VM_L2_PROTECTION_FAULT_CNTL2 = RREG32_SOC15(GC, 0, mmGCVM_L2_PROTECTION_FAULT_CNTL2);
+	adev->gmc.VM_L2_PROTECTION_FAULT_MM_CNTL3 = RREG32_SOC15(GC, 0, mmGCVM_L2_PROTECTION_FAULT_MM_CNTL3);
+	adev->gmc.VM_L2_PROTECTION_FAULT_MM_CNTL4 = RREG32_SOC15(GC, 0, mmGCVM_L2_PROTECTION_FAULT_MM_CNTL4);
+	adev->gmc.VM_L2_PROTECTION_FAULT_ADDR_LO32 = RREG32_SOC15(GC, 0, mmGCVM_L2_PROTECTION_FAULT_ADDR_LO32);
+	adev->gmc.VM_L2_PROTECTION_FAULT_ADDR_HI32 = RREG32_SOC15(GC, 0, mmGCVM_L2_PROTECTION_FAULT_ADDR_HI32);
+	adev->gmc.VM_DEBUG = RREG32_SOC15(GC, 0, mmGCVM_DEBUG);
+	adev->gmc.VM_L2_MM_GROUP_RT_CLASSES = RREG32_SOC15(GC, 0, mmGCVM_L2_MM_GROUP_RT_CLASSES);
+	adev->gmc.VM_L2_BANK_SELECT_RESERVED_CID = RREG32_SOC15(GC, 0, mmGCVM_L2_BANK_SELECT_RESERVED_CID);
+	adev->gmc.VM_L2_BANK_SELECT_RESERVED_CID2 = RREG32_SOC15(GC, 0, mmGCVM_L2_BANK_SELECT_RESERVED_CID2);
+	adev->gmc.VM_L2_CACHE_PARITY_CNTL = RREG32_SOC15(GC, 0, mmGCVM_L2_CACHE_PARITY_CNTL);
+	adev->gmc.VM_L2_IH_LOG_CNTL = RREG32_SOC15(GC, 0, mmGCVM_L2_IH_LOG_CNTL);
+
+	for (i = 0; i <= 15; i++) {
+		adev->gmc.VM_CONTEXT_CNTL[i] = RREG32_SOC15_OFFSET(GC, 0, mmGCVM_CONTEXT0_CNTL, i);
+		adev->gmc.VM_CONTEXT_PAGE_TABLE_BASE_ADDR_LO32[i] = RREG32_SOC15_OFFSET(GC, 0, mmGCVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32, i * 2);
+		adev->gmc.VM_CONTEXT_PAGE_TABLE_BASE_ADDR_HI32[i] = RREG32_SOC15_OFFSET(GC, 0, mmGCVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32, i * 2);
+		adev->gmc.VM_CONTEXT_PAGE_TABLE_START_ADDR_LO32[i] = RREG32_SOC15_OFFSET(GC, 0, mmGCVM_CONTEXT0_PAGE_TABLE_START_ADDR_LO32, i * 2);
+		adev->gmc.VM_CONTEXT_PAGE_TABLE_START_ADDR_HI32[i] = RREG32_SOC15_OFFSET(GC, 0, mmGCVM_CONTEXT0_PAGE_TABLE_START_ADDR_HI32, i * 2);
+		adev->gmc.VM_CONTEXT_PAGE_TABLE_END_ADDR_LO32[i] = RREG32_SOC15_OFFSET(GC, 0, mmGCVM_CONTEXT0_PAGE_TABLE_END_ADDR_LO32, i * 2);
+		adev->gmc.VM_CONTEXT_PAGE_TABLE_END_ADDR_HI32[i] = RREG32_SOC15_OFFSET(GC, 0, mmGCVM_CONTEXT0_PAGE_TABLE_END_ADDR_HI32, i * 2);
+	}
+
+	adev->gmc.MC_VM_MX_L1_TLB_CNTL = RREG32_SOC15(GC, 0, mmGCMC_VM_MX_L1_TLB_CNTL);
+}
+
+static void gfxhub_v2_1_restore_regs(struct amdgpu_device *adev)
+{
+	int i;
+
+	WREG32_SOC15(GC, 0, mmGCVM_L2_CNTL, adev->gmc.VM_L2_CNTL);
+	WREG32_SOC15(GC, 0, mmGCVM_L2_CNTL2, adev->gmc.VM_L2_CNTL2);
+	WREG32_SOC15(GC, 0, mmGCVM_DUMMY_PAGE_FAULT_CNTL, adev->gmc.VM_DUMMY_PAGE_FAULT_CNTL);
+	WREG32_SOC15(GC, 0, mmGCVM_DUMMY_PAGE_FAULT_ADDR_LO32, adev->gmc.VM_DUMMY_PAGE_FAULT_ADDR_LO32);
+	WREG32_SOC15(GC, 0, mmGCVM_DUMMY_PAGE_FAULT_ADDR_HI32, adev->gmc.VM_DUMMY_PAGE_FAULT_ADDR_HI32);
+	WREG32_SOC15(GC, 0, mmGCVM_L2_PROTECTION_FAULT_CNTL, adev->gmc.VM_L2_PROTECTION_FAULT_CNTL);
+	WREG32_SOC15(GC, 0, mmGCVM_L2_PROTECTION_FAULT_CNTL2, adev->gmc.VM_L2_PROTECTION_FAULT_CNTL2);
+	WREG32_SOC15(GC, 0, mmGCVM_L2_PROTECTION_FAULT_MM_CNTL3, adev->gmc.VM_L2_PROTECTION_FAULT_MM_CNTL3);
+	WREG32_SOC15(GC, 0, mmGCVM_L2_PROTECTION_FAULT_MM_CNTL4, adev->gmc.VM_L2_PROTECTION_FAULT_MM_CNTL4);
+	WREG32_SOC15(GC, 0, mmGCVM_L2_PROTECTION_FAULT_ADDR_LO32, adev->gmc.VM_L2_PROTECTION_FAULT_ADDR_LO32);
+	WREG32_SOC15(GC, 0, mmGCVM_L2_PROTECTION_FAULT_ADDR_HI32, adev->gmc.VM_L2_PROTECTION_FAULT_ADDR_HI32);
+	WREG32_SOC15(GC, 0, mmGCVM_DEBUG, adev->gmc.VM_DEBUG);
+	WREG32_SOC15(GC, 0, mmGCVM_L2_MM_GROUP_RT_CLASSES, adev->gmc.VM_L2_MM_GROUP_RT_CLASSES);
+	WREG32_SOC15(GC, 0, mmGCVM_L2_BANK_SELECT_RESERVED_CID, adev->gmc.VM_L2_BANK_SELECT_RESERVED_CID);
+	WREG32_SOC15(GC, 0, mmGCVM_L2_BANK_SELECT_RESERVED_CID2, adev->gmc.VM_L2_BANK_SELECT_RESERVED_CID2);
+	WREG32_SOC15(GC, 0, mmGCVM_L2_CACHE_PARITY_CNTL, adev->gmc.VM_L2_CACHE_PARITY_CNTL);
+	WREG32_SOC15(GC, 0, mmGCVM_L2_IH_LOG_CNTL, adev->gmc.VM_L2_IH_LOG_CNTL);
+
+	for (i = 0; i <= 15; i++) {
+		WREG32_SOC15_OFFSET(GC, 0, mmGCVM_CONTEXT0_CNTL, i, adev->gmc.VM_CONTEXT_CNTL[i]);
+		WREG32_SOC15_OFFSET(GC, 0, mmGCVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32, i * 2, adev->gmc.VM_CONTEXT_PAGE_TABLE_BASE_ADDR_LO32[i]);
+		WREG32_SOC15_OFFSET(GC, 0, mmGCVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32, i * 2, adev->gmc.VM_CONTEXT_PAGE_TABLE_BASE_ADDR_HI32[i]);
+		WREG32_SOC15_OFFSET(GC, 0, mmGCVM_CONTEXT0_PAGE_TABLE_START_ADDR_LO32, i * 2, adev->gmc.VM_CONTEXT_PAGE_TABLE_START_ADDR_LO32[i]);
+		WREG32_SOC15_OFFSET(GC, 0, mmGCVM_CONTEXT0_PAGE_TABLE_START_ADDR_HI32, i * 2, adev->gmc.VM_CONTEXT_PAGE_TABLE_START_ADDR_HI32[i]);
+		WREG32_SOC15_OFFSET(GC, 0, mmGCVM_CONTEXT0_PAGE_TABLE_END_ADDR_LO32, i * 2, adev->gmc.VM_CONTEXT_PAGE_TABLE_END_ADDR_LO32[i]);
+		WREG32_SOC15_OFFSET(GC, 0, mmGCVM_CONTEXT0_PAGE_TABLE_END_ADDR_HI32, i * 2, adev->gmc.VM_CONTEXT_PAGE_TABLE_END_ADDR_HI32[i]);
+	}
+
+	WREG32_SOC15(GC, 0, mmGCMC_VM_FB_LOCATION_BASE, adev->gmc.vram_start >> 24);
+	WREG32_SOC15(GC, 0, mmGCMC_VM_FB_LOCATION_TOP, adev->gmc.vram_end >> 24);
+	WREG32_SOC15(GC, 0, mmGCMC_VM_MX_L1_TLB_CNTL, adev->gmc.MC_VM_MX_L1_TLB_CNTL);
+}
+
+static void gfxhub_v2_1_halt(struct amdgpu_device *adev)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)];
+	int i;
+	uint32_t tmp;
+	int time = 1000;
+
+	gfxhub_v2_1_set_fault_enable_default(adev, false);
+
+	for (i = 0; i <= 14; i++) {
+		WREG32_SOC15_OFFSET(GC, 0, mmGCVM_CONTEXT1_PAGE_TABLE_START_ADDR_LO32,
+				    i * hub->ctx_addr_distance, ~0);
+		WREG32_SOC15_OFFSET(GC, 0, mmGCVM_CONTEXT1_PAGE_TABLE_START_ADDR_HI32,
+				    i * hub->ctx_addr_distance, ~0);
+		WREG32_SOC15_OFFSET(GC, 0, mmGCVM_CONTEXT1_PAGE_TABLE_END_ADDR_LO32,
+				    i * hub->ctx_addr_distance,
+				    0);
+		WREG32_SOC15_OFFSET(GC, 0, mmGCVM_CONTEXT1_PAGE_TABLE_END_ADDR_HI32,
+				    i * hub->ctx_addr_distance,
+				    0);
+	}
+	tmp = RREG32_SOC15(GC, 0, mmGRBM_STATUS2);
+	while ((tmp & (GRBM_STATUS2__EA_BUSY_MASK |
+		      GRBM_STATUS2__EA_LINK_BUSY_MASK)) != 0 &&
+	       time) {
+		udelay(100);
+		time--;
+		tmp = RREG32_SOC15(GC, 0, mmGRBM_STATUS2);
+	}
+
+	if (!time)
+		DRM_WARN("failed to wait for GRBM(EA) idle\n");
+}
+
 const struct amdgpu_gfxhub_funcs gfxhub_v2_1_funcs = {
 	.get_fb_location = gfxhub_v2_1_get_fb_location,
 	.get_mc_fb_offset = gfxhub_v2_1_get_mc_fb_offset,
@@ -577,6 +660,8 @@ const struct amdgpu_gfxhub_funcs gfxhub_v2_1_funcs = {
 	.gart_disable = gfxhub_v2_1_gart_disable,
 	.set_fault_enable_default = gfxhub_v2_1_set_fault_enable_default,
 	.init = gfxhub_v2_1_init,
-	.get_xgmi_info = gfxhub_v2_1_get_xgmi_info,
 	.utcl2_harvest = gfxhub_v2_1_utcl2_harvest,
+	.mode2_save_regs = gfxhub_v2_1_save_regs,
+	.mode2_restore_regs = gfxhub_v2_1_restore_regs,
+	.halt = gfxhub_v2_1_halt,
 };
diff --git a/drivers/gpu/drm/amd/amdgpu/gfxhub_v3_0.c b/drivers/gpu/drm/amd/amdgpu/gfxhub_v3_0.c
new file mode 100644
index 000000000000..abe30c8bd2ba
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/gfxhub_v3_0.c
@@ -0,0 +1,513 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "amdgpu.h"
+#include "gfxhub_v3_0.h"
+
+#include "gc/gc_11_0_0_offset.h"
+#include "gc/gc_11_0_0_sh_mask.h"
+#include "gc/gc_11_0_0_default.h"
+#include "navi10_enum.h"
+#include "soc15_common.h"
+
+static const char * const gfxhub_client_ids[] = {
+	"CB/DB",
+	"Reserved",
+	"GE1",
+	"GE2",
+	"CPF",
+	"CPC",
+	"CPG",
+	"RLC",
+	"TCP",
+	"SQC (inst)",
+	"SQC (data)",
+	"SQG",
+	"Reserved",
+	"SDMA0",
+	"SDMA1",
+	"GCR",
+	"SDMA2",
+	"SDMA3",
+};
+
+static uint32_t gfxhub_v3_0_get_invalidate_req(unsigned int vmid,
+					       uint32_t flush_type)
+{
+	u32 req = 0;
+
+	/* invalidate using legacy mode on vmid*/
+	req = REG_SET_FIELD(req, GCVM_INVALIDATE_ENG0_REQ,
+			    PER_VMID_INVALIDATE_REQ, 1 << vmid);
+	req = REG_SET_FIELD(req, GCVM_INVALIDATE_ENG0_REQ, FLUSH_TYPE, flush_type);
+	req = REG_SET_FIELD(req, GCVM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PTES, 1);
+	req = REG_SET_FIELD(req, GCVM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PDE0, 1);
+	req = REG_SET_FIELD(req, GCVM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PDE1, 1);
+	req = REG_SET_FIELD(req, GCVM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PDE2, 1);
+	req = REG_SET_FIELD(req, GCVM_INVALIDATE_ENG0_REQ, INVALIDATE_L1_PTES, 1);
+	req = REG_SET_FIELD(req, GCVM_INVALIDATE_ENG0_REQ,
+			    CLEAR_PROTECTION_FAULT_STATUS_ADDR,	0);
+
+	return req;
+}
+
+static void
+gfxhub_v3_0_print_l2_protection_fault_status(struct amdgpu_device *adev,
+					     uint32_t status)
+{
+	u32 cid = REG_GET_FIELD(status,
+				GCVM_L2_PROTECTION_FAULT_STATUS, CID);
+
+	dev_err(adev->dev,
+		"GCVM_L2_PROTECTION_FAULT_STATUS:0x%08X\n",
+		status);
+	dev_err(adev->dev, "\t Faulty UTCL2 client ID: %s (0x%x)\n",
+		cid >= ARRAY_SIZE(gfxhub_client_ids) ? "unknown" : gfxhub_client_ids[cid],
+		cid);
+	dev_err(adev->dev, "\t MORE_FAULTS: 0x%lx\n",
+		REG_GET_FIELD(status,
+		GCVM_L2_PROTECTION_FAULT_STATUS, MORE_FAULTS));
+	dev_err(adev->dev, "\t WALKER_ERROR: 0x%lx\n",
+		REG_GET_FIELD(status,
+		GCVM_L2_PROTECTION_FAULT_STATUS, WALKER_ERROR));
+	dev_err(adev->dev, "\t PERMISSION_FAULTS: 0x%lx\n",
+		REG_GET_FIELD(status,
+		GCVM_L2_PROTECTION_FAULT_STATUS, PERMISSION_FAULTS));
+	dev_err(adev->dev, "\t MAPPING_ERROR: 0x%lx\n",
+		REG_GET_FIELD(status,
+		GCVM_L2_PROTECTION_FAULT_STATUS, MAPPING_ERROR));
+	dev_err(adev->dev, "\t RW: 0x%lx\n",
+		REG_GET_FIELD(status,
+		GCVM_L2_PROTECTION_FAULT_STATUS, RW));
+}
+
+static u64 gfxhub_v3_0_get_fb_location(struct amdgpu_device *adev)
+{
+	u64 base = RREG32_SOC15(GC, 0, regGCMC_VM_FB_LOCATION_BASE);
+
+	base &= GCMC_VM_FB_LOCATION_BASE__FB_BASE_MASK;
+	base <<= 24;
+
+	return base;
+}
+
+static u64 gfxhub_v3_0_get_mc_fb_offset(struct amdgpu_device *adev)
+{
+	return (u64)RREG32_SOC15(GC, 0, regGCMC_VM_FB_OFFSET) << 24;
+}
+
+static void gfxhub_v3_0_setup_vm_pt_regs(struct amdgpu_device *adev, uint32_t vmid,
+				uint64_t page_table_base)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)];
+
+	WREG32_SOC15_OFFSET(GC, 0, regGCVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32,
+			    hub->ctx_addr_distance * vmid,
+			    lower_32_bits(page_table_base));
+
+	WREG32_SOC15_OFFSET(GC, 0, regGCVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32,
+			    hub->ctx_addr_distance * vmid,
+			    upper_32_bits(page_table_base));
+}
+
+static void gfxhub_v3_0_init_gart_aperture_regs(struct amdgpu_device *adev)
+{
+	uint64_t pt_base = amdgpu_gmc_pd_addr(adev->gart.bo);
+
+	gfxhub_v3_0_setup_vm_pt_regs(adev, 0, pt_base);
+
+	WREG32_SOC15(GC, 0, regGCVM_CONTEXT0_PAGE_TABLE_START_ADDR_LO32,
+		     (u32)(adev->gmc.gart_start >> 12));
+	WREG32_SOC15(GC, 0, regGCVM_CONTEXT0_PAGE_TABLE_START_ADDR_HI32,
+		     (u32)(adev->gmc.gart_start >> 44));
+
+	WREG32_SOC15(GC, 0, regGCVM_CONTEXT0_PAGE_TABLE_END_ADDR_LO32,
+		     (u32)(adev->gmc.gart_end >> 12));
+	WREG32_SOC15(GC, 0, regGCVM_CONTEXT0_PAGE_TABLE_END_ADDR_HI32,
+		     (u32)(adev->gmc.gart_end >> 44));
+}
+
+static void gfxhub_v3_0_init_system_aperture_regs(struct amdgpu_device *adev)
+{
+	uint64_t value;
+
+	/* Program the AGP BAR */
+	WREG32_SOC15(GC, 0, regGCMC_VM_AGP_BASE, 0);
+	WREG32_SOC15(GC, 0, regGCMC_VM_AGP_BOT, adev->gmc.agp_start >> 24);
+	WREG32_SOC15(GC, 0, regGCMC_VM_AGP_TOP, adev->gmc.agp_end >> 24);
+
+
+	/* Program the system aperture low logical page number. */
+	WREG32_SOC15(GC, 0, regGCMC_VM_SYSTEM_APERTURE_LOW_ADDR,
+		     min(adev->gmc.fb_start, adev->gmc.agp_start) >> 18);
+	WREG32_SOC15(GC, 0, regGCMC_VM_SYSTEM_APERTURE_HIGH_ADDR,
+		     max(adev->gmc.fb_end, adev->gmc.agp_end) >> 18);
+
+	/* Set default page address. */
+	value = amdgpu_gmc_vram_mc2pa(adev, adev->mem_scratch.gpu_addr);
+	WREG32_SOC15(GC, 0, regGCMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB,
+		     (u32)(value >> 12));
+	WREG32_SOC15(GC, 0, regGCMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_MSB,
+		     (u32)(value >> 44));
+
+	/* Program "protection fault". */
+	WREG32_SOC15(GC, 0, regGCVM_L2_PROTECTION_FAULT_DEFAULT_ADDR_LO32,
+		     (u32)(adev->dummy_page_addr >> 12));
+	WREG32_SOC15(GC, 0, regGCVM_L2_PROTECTION_FAULT_DEFAULT_ADDR_HI32,
+		     (u32)((u64)adev->dummy_page_addr >> 44));
+
+	WREG32_FIELD15_PREREG(GC, 0, GCVM_L2_PROTECTION_FAULT_CNTL2,
+		       ACTIVE_PAGE_MIGRATION_PTE_READ_RETRY, 1);
+}
+
+
+static void gfxhub_v3_0_init_tlb_regs(struct amdgpu_device *adev)
+{
+	uint32_t tmp;
+
+	/* Setup TLB control */
+	tmp = RREG32_SOC15(GC, 0, regGCMC_VM_MX_L1_TLB_CNTL);
+
+	tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB, 1);
+	tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL, SYSTEM_ACCESS_MODE, 3);
+	tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL,
+			    ENABLE_ADVANCED_DRIVER_MODEL, 1);
+	tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL,
+			    SYSTEM_APERTURE_UNMAPPED_ACCESS, 0);
+	tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL, ECO_BITS, 0);
+	tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL,
+			    MTYPE, MTYPE_UC); /* UC, uncached */
+
+	WREG32_SOC15(GC, 0, regGCMC_VM_MX_L1_TLB_CNTL, tmp);
+}
+
+static void gfxhub_v3_0_init_cache_regs(struct amdgpu_device *adev)
+{
+	uint32_t tmp;
+
+	/* These registers are not accessible to VF-SRIOV.
+	 * The PF will program them instead.
+	 */
+	if (amdgpu_sriov_vf(adev))
+		return;
+
+	/* Setup L2 cache */
+	tmp = RREG32_SOC15(GC, 0, regGCVM_L2_CNTL);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL, ENABLE_L2_CACHE, 1);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL, ENABLE_L2_FRAGMENT_PROCESSING, 0);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL,
+			    ENABLE_DEFAULT_PAGE_OUT_TO_SYSTEM_MEMORY, 1);
+	/* XXX for emulation, Refer to closed source code.*/
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL,
+			    L2_PDE0_CACHE_TAG_GENERATION_MODE, 0);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL, PDE_FAULT_CLASSIFICATION, 0);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL, CONTEXT1_IDENTITY_ACCESS_MODE, 1);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL, IDENTITY_MODE_FRAGMENT_SIZE, 0);
+	WREG32_SOC15(GC, 0, regGCVM_L2_CNTL, tmp);
+
+	tmp = RREG32_SOC15(GC, 0, regGCVM_L2_CNTL2);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL2, INVALIDATE_ALL_L1_TLBS, 1);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL2, INVALIDATE_L2_CACHE, 1);
+	WREG32_SOC15(GC, 0, regGCVM_L2_CNTL2, tmp);
+
+	tmp = regGCVM_L2_CNTL3_DEFAULT;
+	if (adev->gmc.translate_further) {
+		tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL3, BANK_SELECT, 12);
+		tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL3,
+				    L2_CACHE_BIGK_FRAGMENT_SIZE, 9);
+	} else {
+		tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL3, BANK_SELECT, 9);
+		tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL3,
+				    L2_CACHE_BIGK_FRAGMENT_SIZE, 6);
+	}
+	WREG32_SOC15(GC, 0, regGCVM_L2_CNTL3, tmp);
+
+	tmp = regGCVM_L2_CNTL4_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL4, VMC_TAP_PDE_REQUEST_PHYSICAL, 0);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL4, VMC_TAP_PTE_REQUEST_PHYSICAL, 0);
+	WREG32_SOC15(GC, 0, regGCVM_L2_CNTL4, tmp);
+
+	tmp = regGCVM_L2_CNTL5_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL5, L2_CACHE_SMALLK_FRAGMENT_SIZE, 0);
+	WREG32_SOC15(GC, 0, regGCVM_L2_CNTL5, tmp);
+}
+
+static void gfxhub_v3_0_enable_system_domain(struct amdgpu_device *adev)
+{
+	uint32_t tmp;
+
+	tmp = RREG32_SOC15(GC, 0, regGCVM_CONTEXT0_CNTL);
+	tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT0_CNTL, ENABLE_CONTEXT, 1);
+	tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT0_CNTL, PAGE_TABLE_DEPTH, 0);
+	tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT0_CNTL,
+			    RETRY_PERMISSION_OR_INVALID_PAGE_FAULT, 0);
+	WREG32_SOC15(GC, 0, regGCVM_CONTEXT0_CNTL, tmp);
+}
+
+static void gfxhub_v3_0_disable_identity_aperture(struct amdgpu_device *adev)
+{
+	/* These registers are not accessible to VF-SRIOV.
+	 * The PF will program them instead.
+	 */
+	if (amdgpu_sriov_vf(adev))
+		return;
+
+	WREG32_SOC15(GC, 0, regGCVM_L2_CONTEXT1_IDENTITY_APERTURE_LOW_ADDR_LO32,
+		     0xFFFFFFFF);
+	WREG32_SOC15(GC, 0, regGCVM_L2_CONTEXT1_IDENTITY_APERTURE_LOW_ADDR_HI32,
+		     0x0000000F);
+
+	WREG32_SOC15(GC, 0, regGCVM_L2_CONTEXT1_IDENTITY_APERTURE_HIGH_ADDR_LO32,
+		     0);
+	WREG32_SOC15(GC, 0, regGCVM_L2_CONTEXT1_IDENTITY_APERTURE_HIGH_ADDR_HI32,
+		     0);
+
+	WREG32_SOC15(GC, 0, regGCVM_L2_CONTEXT_IDENTITY_PHYSICAL_OFFSET_LO32, 0);
+	WREG32_SOC15(GC, 0, regGCVM_L2_CONTEXT_IDENTITY_PHYSICAL_OFFSET_HI32, 0);
+
+}
+
+static void gfxhub_v3_0_setup_vmid_config(struct amdgpu_device *adev)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)];
+	int i;
+	uint32_t tmp;
+
+	for (i = 0; i <= 14; i++) {
+		tmp = RREG32_SOC15_OFFSET(GC, 0, regGCVM_CONTEXT1_CNTL, i * hub->ctx_distance);
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL, ENABLE_CONTEXT, 1);
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL, PAGE_TABLE_DEPTH,
+				    adev->vm_manager.num_level);
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
+				RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
+				DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
+				PDE0_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
+				VALID_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
+				READ_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
+				WRITE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
+				EXECUTE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
+				PAGE_TABLE_BLOCK_SIZE,
+				adev->vm_manager.block_size - 9);
+		/* Send no-retry XNACK on fault to suppress VM fault storm. */
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
+				    RETRY_PERMISSION_OR_INVALID_PAGE_FAULT,
+				    !amdgpu_noretry);
+		WREG32_SOC15_OFFSET(GC, 0, regGCVM_CONTEXT1_CNTL,
+				    i * hub->ctx_distance, tmp);
+		WREG32_SOC15_OFFSET(GC, 0, regGCVM_CONTEXT1_PAGE_TABLE_START_ADDR_LO32,
+				    i * hub->ctx_addr_distance, 0);
+		WREG32_SOC15_OFFSET(GC, 0, regGCVM_CONTEXT1_PAGE_TABLE_START_ADDR_HI32,
+				    i * hub->ctx_addr_distance, 0);
+		WREG32_SOC15_OFFSET(GC, 0, regGCVM_CONTEXT1_PAGE_TABLE_END_ADDR_LO32,
+				    i * hub->ctx_addr_distance,
+				    lower_32_bits(adev->vm_manager.max_pfn - 1));
+		WREG32_SOC15_OFFSET(GC, 0, regGCVM_CONTEXT1_PAGE_TABLE_END_ADDR_HI32,
+				    i * hub->ctx_addr_distance,
+				    upper_32_bits(adev->vm_manager.max_pfn - 1));
+	}
+
+	hub->vm_cntx_cntl = tmp;
+}
+
+static void gfxhub_v3_0_program_invalidation(struct amdgpu_device *adev)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)];
+	unsigned int i;
+
+	for (i = 0 ; i < 18; ++i) {
+		WREG32_SOC15_OFFSET(GC, 0, regGCVM_INVALIDATE_ENG0_ADDR_RANGE_LO32,
+				    i * hub->eng_addr_distance, 0xffffffff);
+		WREG32_SOC15_OFFSET(GC, 0, regGCVM_INVALIDATE_ENG0_ADDR_RANGE_HI32,
+				    i * hub->eng_addr_distance, 0x1f);
+	}
+}
+
+static int gfxhub_v3_0_gart_enable(struct amdgpu_device *adev)
+{
+	if (amdgpu_sriov_vf(adev)) {
+		/*
+		 * GCMC_VM_FB_LOCATION_BASE/TOP is NULL for VF, becuase they are
+		 * VF copy registers so vbios post doesn't program them, for
+		 * SRIOV driver need to program them
+		 */
+		WREG32_SOC15(GC, 0, regGCMC_VM_FB_LOCATION_BASE,
+			     adev->gmc.vram_start >> 24);
+		WREG32_SOC15(GC, 0, regGCMC_VM_FB_LOCATION_TOP,
+			     adev->gmc.vram_end >> 24);
+	}
+
+	/* GART Enable. */
+	gfxhub_v3_0_init_gart_aperture_regs(adev);
+	gfxhub_v3_0_init_system_aperture_regs(adev);
+	gfxhub_v3_0_init_tlb_regs(adev);
+	gfxhub_v3_0_init_cache_regs(adev);
+
+	gfxhub_v3_0_enable_system_domain(adev);
+	gfxhub_v3_0_disable_identity_aperture(adev);
+	gfxhub_v3_0_setup_vmid_config(adev);
+	gfxhub_v3_0_program_invalidation(adev);
+
+	return 0;
+}
+
+static void gfxhub_v3_0_gart_disable(struct amdgpu_device *adev)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)];
+	u32 tmp;
+	u32 i;
+
+	/* Disable all tables */
+	for (i = 0; i < 16; i++)
+		WREG32_SOC15_OFFSET(GC, 0, regGCVM_CONTEXT0_CNTL,
+				    i * hub->ctx_distance, 0);
+
+	/* Setup TLB control */
+	tmp = RREG32_SOC15(GC, 0, regGCMC_VM_MX_L1_TLB_CNTL);
+	tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB, 0);
+	tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL,
+			    ENABLE_ADVANCED_DRIVER_MODEL, 0);
+	WREG32_SOC15(GC, 0, regGCMC_VM_MX_L1_TLB_CNTL, tmp);
+
+	/* Setup L2 cache */
+	WREG32_FIELD15_PREREG(GC, 0, GCVM_L2_CNTL, ENABLE_L2_CACHE, 0);
+	WREG32_SOC15(GC, 0, regGCVM_L2_CNTL3, 0);
+}
+
+/**
+ * gfxhub_v3_0_set_fault_enable_default - update GART/VM fault handling
+ *
+ * @adev: amdgpu_device pointer
+ * @value: true redirects VM faults to the default page
+ */
+static void gfxhub_v3_0_set_fault_enable_default(struct amdgpu_device *adev,
+					  bool value)
+{
+	u32 tmp;
+
+	/* NO halt CP when page fault */
+	tmp = RREG32_SOC15(GC, 0, regCP_DEBUG);
+	tmp = REG_SET_FIELD(tmp, CP_DEBUG, CPG_UTCL1_ERROR_HALT_DISABLE, 1);
+	WREG32_SOC15(GC, 0, regCP_DEBUG, tmp);
+
+	/* These registers are not accessible to VF-SRIOV.
+	 * The PF will program them instead.
+	 */
+	if (amdgpu_sriov_vf(adev))
+		return;
+
+	tmp = RREG32_SOC15(GC, 0, regGCVM_L2_PROTECTION_FAULT_CNTL);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    PDE0_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    PDE1_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    PDE2_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    TRANSLATE_FURTHER_PROTECTION_FAULT_ENABLE_DEFAULT,
+			    value);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    NACK_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    VALID_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    READ_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    WRITE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    EXECUTE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	if (!value) {
+		tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+				CRASH_ON_NO_RETRY_FAULT, 1);
+		tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+				CRASH_ON_RETRY_FAULT, 1);
+	}
+	WREG32_SOC15(GC, 0, regGCVM_L2_PROTECTION_FAULT_CNTL, tmp);
+}
+
+static const struct amdgpu_vmhub_funcs gfxhub_v3_0_vmhub_funcs = {
+	.print_l2_protection_fault_status = gfxhub_v3_0_print_l2_protection_fault_status,
+	.get_invalidate_req = gfxhub_v3_0_get_invalidate_req,
+};
+
+static void gfxhub_v3_0_init(struct amdgpu_device *adev)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)];
+
+	hub->ctx0_ptb_addr_lo32 =
+		SOC15_REG_OFFSET(GC, 0,
+				 regGCVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32);
+	hub->ctx0_ptb_addr_hi32 =
+		SOC15_REG_OFFSET(GC, 0,
+				 regGCVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32);
+	hub->vm_inv_eng0_sem =
+		SOC15_REG_OFFSET(GC, 0, regGCVM_INVALIDATE_ENG0_SEM);
+	hub->vm_inv_eng0_req =
+		SOC15_REG_OFFSET(GC, 0, regGCVM_INVALIDATE_ENG0_REQ);
+	hub->vm_inv_eng0_ack =
+		SOC15_REG_OFFSET(GC, 0, regGCVM_INVALIDATE_ENG0_ACK);
+	hub->vm_context0_cntl =
+		SOC15_REG_OFFSET(GC, 0, regGCVM_CONTEXT0_CNTL);
+	hub->vm_l2_pro_fault_status =
+		SOC15_REG_OFFSET(GC, 0, regGCVM_L2_PROTECTION_FAULT_STATUS);
+	hub->vm_l2_pro_fault_cntl =
+		SOC15_REG_OFFSET(GC, 0, regGCVM_L2_PROTECTION_FAULT_CNTL);
+
+	hub->ctx_distance = regGCVM_CONTEXT1_CNTL - regGCVM_CONTEXT0_CNTL;
+	hub->ctx_addr_distance = regGCVM_CONTEXT1_PAGE_TABLE_BASE_ADDR_LO32 -
+		regGCVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32;
+	hub->eng_distance = regGCVM_INVALIDATE_ENG1_REQ -
+		regGCVM_INVALIDATE_ENG0_REQ;
+	hub->eng_addr_distance = regGCVM_INVALIDATE_ENG1_ADDR_RANGE_LO32 -
+		regGCVM_INVALIDATE_ENG0_ADDR_RANGE_LO32;
+
+	hub->vm_cntx_cntl_vm_fault = GCVM_CONTEXT1_CNTL__RANGE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		GCVM_CONTEXT1_CNTL__DUMMY_PAGE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		GCVM_CONTEXT1_CNTL__PDE0_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		GCVM_CONTEXT1_CNTL__VALID_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		GCVM_CONTEXT1_CNTL__READ_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		GCVM_CONTEXT1_CNTL__WRITE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		GCVM_CONTEXT1_CNTL__EXECUTE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK;
+
+	hub->vmhub_funcs = &gfxhub_v3_0_vmhub_funcs;
+}
+
+const struct amdgpu_gfxhub_funcs gfxhub_v3_0_funcs = {
+	.get_fb_location = gfxhub_v3_0_get_fb_location,
+	.get_mc_fb_offset = gfxhub_v3_0_get_mc_fb_offset,
+	.setup_vm_pt_regs = gfxhub_v3_0_setup_vm_pt_regs,
+	.gart_enable = gfxhub_v3_0_gart_enable,
+	.gart_disable = gfxhub_v3_0_gart_disable,
+	.set_fault_enable_default = gfxhub_v3_0_set_fault_enable_default,
+	.init = gfxhub_v3_0_init,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/gfxhub_v3_0.h b/drivers/gpu/drm/amd/amdgpu/gfxhub_v3_0.h
new file mode 100644
index 000000000000..ea345e4e072a
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/gfxhub_v3_0.h
@@ -0,0 +1,29 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __GFXHUB_V3_0_H__
+#define __GFXHUB_V3_0_H__
+
+extern const struct amdgpu_gfxhub_funcs gfxhub_v3_0_funcs;
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/gfxhub_v3_0_3.c b/drivers/gpu/drm/amd/amdgpu/gfxhub_v3_0_3.c
new file mode 100644
index 000000000000..b3ef6e71811f
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/gfxhub_v3_0_3.c
@@ -0,0 +1,501 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "amdgpu.h"
+#include "gfxhub_v3_0_3.h"
+
+#include "gc/gc_11_0_3_offset.h"
+#include "gc/gc_11_0_3_sh_mask.h"
+#include "navi10_enum.h"
+#include "soc15_common.h"
+
+#define regGCVM_L2_CNTL3_DEFAULT		0x80100007
+#define regGCVM_L2_CNTL4_DEFAULT		0x000000c1
+#define regGCVM_L2_CNTL5_DEFAULT		0x00003fe0
+
+static const char * const gfxhub_client_ids[] = {
+	"CB/DB",
+	"Reserved",
+	"GE1",
+	"GE2",
+	"CPF",
+	"CPC",
+	"CPG",
+	"RLC",
+	"TCP",
+	"SQC (inst)",
+	"SQC (data)",
+	"SQG",
+	"Reserved",
+	"SDMA0",
+	"SDMA1",
+	"GCR",
+	"SDMA2",
+	"SDMA3",
+};
+
+static uint32_t gfxhub_v3_0_3_get_invalidate_req(unsigned int vmid,
+					       uint32_t flush_type)
+{
+	u32 req = 0;
+
+	/* invalidate using legacy mode on vmid*/
+	req = REG_SET_FIELD(req, GCVM_INVALIDATE_ENG0_REQ,
+			    PER_VMID_INVALIDATE_REQ, 1 << vmid);
+	req = REG_SET_FIELD(req, GCVM_INVALIDATE_ENG0_REQ, FLUSH_TYPE, flush_type);
+	req = REG_SET_FIELD(req, GCVM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PTES, 1);
+	req = REG_SET_FIELD(req, GCVM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PDE0, 1);
+	req = REG_SET_FIELD(req, GCVM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PDE1, 1);
+	req = REG_SET_FIELD(req, GCVM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PDE2, 1);
+	req = REG_SET_FIELD(req, GCVM_INVALIDATE_ENG0_REQ, INVALIDATE_L1_PTES, 1);
+	req = REG_SET_FIELD(req, GCVM_INVALIDATE_ENG0_REQ,
+			    CLEAR_PROTECTION_FAULT_STATUS_ADDR,	0);
+
+	return req;
+}
+
+static void
+gfxhub_v3_0_3_print_l2_protection_fault_status(struct amdgpu_device *adev,
+					     uint32_t status)
+{
+	u32 cid = REG_GET_FIELD(status,
+				GCVM_L2_PROTECTION_FAULT_STATUS, CID);
+
+	dev_err(adev->dev,
+		"GCVM_L2_PROTECTION_FAULT_STATUS:0x%08X\n",
+		status);
+	dev_err(adev->dev, "\t Faulty UTCL2 client ID: %s (0x%x)\n",
+		cid >= ARRAY_SIZE(gfxhub_client_ids) ? "unknown" : gfxhub_client_ids[cid],
+		cid);
+	dev_err(adev->dev, "\t MORE_FAULTS: 0x%lx\n",
+		REG_GET_FIELD(status,
+		GCVM_L2_PROTECTION_FAULT_STATUS, MORE_FAULTS));
+	dev_err(adev->dev, "\t WALKER_ERROR: 0x%lx\n",
+		REG_GET_FIELD(status,
+		GCVM_L2_PROTECTION_FAULT_STATUS, WALKER_ERROR));
+	dev_err(adev->dev, "\t PERMISSION_FAULTS: 0x%lx\n",
+		REG_GET_FIELD(status,
+		GCVM_L2_PROTECTION_FAULT_STATUS, PERMISSION_FAULTS));
+	dev_err(adev->dev, "\t MAPPING_ERROR: 0x%lx\n",
+		REG_GET_FIELD(status,
+		GCVM_L2_PROTECTION_FAULT_STATUS, MAPPING_ERROR));
+	dev_err(adev->dev, "\t RW: 0x%lx\n",
+		REG_GET_FIELD(status,
+		GCVM_L2_PROTECTION_FAULT_STATUS, RW));
+}
+
+static u64 gfxhub_v3_0_3_get_fb_location(struct amdgpu_device *adev)
+{
+	u64 base = RREG32_SOC15(GC, 0, regGCMC_VM_FB_LOCATION_BASE);
+
+	base &= GCMC_VM_FB_LOCATION_BASE__FB_BASE_MASK;
+	base <<= 24;
+
+	return base;
+}
+
+static u64 gfxhub_v3_0_3_get_mc_fb_offset(struct amdgpu_device *adev)
+{
+	return (u64)RREG32_SOC15(GC, 0, regGCMC_VM_FB_OFFSET) << 24;
+}
+
+static void gfxhub_v3_0_3_setup_vm_pt_regs(struct amdgpu_device *adev, uint32_t vmid,
+				uint64_t page_table_base)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)];
+
+	WREG32_SOC15_OFFSET(GC, 0, regGCVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32,
+			    hub->ctx_addr_distance * vmid,
+			    lower_32_bits(page_table_base));
+
+	WREG32_SOC15_OFFSET(GC, 0, regGCVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32,
+			    hub->ctx_addr_distance * vmid,
+			    upper_32_bits(page_table_base));
+}
+
+static void gfxhub_v3_0_3_init_gart_aperture_regs(struct amdgpu_device *adev)
+{
+	uint64_t pt_base = amdgpu_gmc_pd_addr(adev->gart.bo);
+
+	gfxhub_v3_0_3_setup_vm_pt_regs(adev, 0, pt_base);
+
+	WREG32_SOC15(GC, 0, regGCVM_CONTEXT0_PAGE_TABLE_START_ADDR_LO32,
+		     (u32)(adev->gmc.gart_start >> 12));
+	WREG32_SOC15(GC, 0, regGCVM_CONTEXT0_PAGE_TABLE_START_ADDR_HI32,
+		     (u32)(adev->gmc.gart_start >> 44));
+
+	WREG32_SOC15(GC, 0, regGCVM_CONTEXT0_PAGE_TABLE_END_ADDR_LO32,
+		     (u32)(adev->gmc.gart_end >> 12));
+	WREG32_SOC15(GC, 0, regGCVM_CONTEXT0_PAGE_TABLE_END_ADDR_HI32,
+		     (u32)(adev->gmc.gart_end >> 44));
+}
+
+static void gfxhub_v3_0_3_init_system_aperture_regs(struct amdgpu_device *adev)
+{
+	uint64_t value;
+
+	if (amdgpu_sriov_vf(adev))
+		return;
+
+	/* Disable AGP. */
+	WREG32_SOC15(GC, 0, regGCMC_VM_AGP_BASE, 0);
+	WREG32_SOC15(GC, 0, regGCMC_VM_AGP_BOT, adev->gmc.agp_start >> 24);
+	WREG32_SOC15(GC, 0, regGCMC_VM_AGP_TOP, adev->gmc.agp_end >> 24);
+
+	/* Program the system aperture low logical page number. */
+	WREG32_SOC15(GC, 0, regGCMC_VM_SYSTEM_APERTURE_LOW_ADDR,
+		     min(adev->gmc.fb_start, adev->gmc.agp_start) >> 18);
+	WREG32_SOC15(GC, 0, regGCMC_VM_SYSTEM_APERTURE_HIGH_ADDR,
+		     max(adev->gmc.fb_end, adev->gmc.agp_end) >> 18);
+
+	/* Set default page address. */
+	value = amdgpu_gmc_vram_mc2pa(adev, adev->mem_scratch.gpu_addr);
+	WREG32_SOC15(GC, 0, regGCMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB,
+		     (u32)(value >> 12));
+	WREG32_SOC15(GC, 0, regGCMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_MSB,
+		     (u32)(value >> 44));
+
+	/* Program "protection fault". */
+	WREG32_SOC15(GC, 0, regGCVM_L2_PROTECTION_FAULT_DEFAULT_ADDR_LO32,
+		     (u32)(adev->dummy_page_addr >> 12));
+	WREG32_SOC15(GC, 0, regGCVM_L2_PROTECTION_FAULT_DEFAULT_ADDR_HI32,
+		     (u32)((u64)adev->dummy_page_addr >> 44));
+
+	WREG32_FIELD15_PREREG(GC, 0, GCVM_L2_PROTECTION_FAULT_CNTL2,
+		       ACTIVE_PAGE_MIGRATION_PTE_READ_RETRY, 1);
+}
+
+
+static void gfxhub_v3_0_3_init_tlb_regs(struct amdgpu_device *adev)
+{
+	uint32_t tmp;
+
+	/* Setup TLB control */
+	tmp = RREG32_SOC15(GC, 0, regGCMC_VM_MX_L1_TLB_CNTL);
+
+	tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB, 1);
+	tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL, SYSTEM_ACCESS_MODE, 3);
+	tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL,
+			    ENABLE_ADVANCED_DRIVER_MODEL, 1);
+	tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL,
+			    SYSTEM_APERTURE_UNMAPPED_ACCESS, 0);
+	tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL, ECO_BITS, 0);
+	tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL,
+			    MTYPE, MTYPE_UC); /* UC, uncached */
+
+	WREG32_SOC15(GC, 0, regGCMC_VM_MX_L1_TLB_CNTL, tmp);
+}
+
+static void gfxhub_v3_0_3_init_cache_regs(struct amdgpu_device *adev)
+{
+	uint32_t tmp;
+
+	/* These registers are not accessible to VF-SRIOV.
+	 * The PF will program them instead.
+	 */
+	if (amdgpu_sriov_vf(adev))
+		return;
+
+	/* Setup L2 cache */
+	tmp = RREG32_SOC15(GC, 0, regGCVM_L2_CNTL);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL, ENABLE_L2_CACHE, 1);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL, ENABLE_L2_FRAGMENT_PROCESSING, 0);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL,
+			    ENABLE_DEFAULT_PAGE_OUT_TO_SYSTEM_MEMORY, 1);
+	/* XXX for emulation, Refer to closed source code.*/
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL,
+			    L2_PDE0_CACHE_TAG_GENERATION_MODE, 0);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL, PDE_FAULT_CLASSIFICATION, 0);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL, CONTEXT1_IDENTITY_ACCESS_MODE, 1);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL, IDENTITY_MODE_FRAGMENT_SIZE, 0);
+	WREG32_SOC15(GC, 0, regGCVM_L2_CNTL, tmp);
+
+	tmp = RREG32_SOC15(GC, 0, regGCVM_L2_CNTL2);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL2, INVALIDATE_ALL_L1_TLBS, 1);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL2, INVALIDATE_L2_CACHE, 1);
+	WREG32_SOC15(GC, 0, regGCVM_L2_CNTL2, tmp);
+
+	tmp = regGCVM_L2_CNTL3_DEFAULT;
+	if (adev->gmc.translate_further) {
+		tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL3, BANK_SELECT, 12);
+		tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL3,
+				    L2_CACHE_BIGK_FRAGMENT_SIZE, 9);
+	} else {
+		tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL3, BANK_SELECT, 9);
+		tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL3,
+				    L2_CACHE_BIGK_FRAGMENT_SIZE, 6);
+	}
+	WREG32_SOC15(GC, 0, regGCVM_L2_CNTL3, tmp);
+
+	tmp = regGCVM_L2_CNTL4_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL4, VMC_TAP_PDE_REQUEST_PHYSICAL, 0);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL4, VMC_TAP_PTE_REQUEST_PHYSICAL, 0);
+	WREG32_SOC15(GC, 0, regGCVM_L2_CNTL4, tmp);
+
+	tmp = regGCVM_L2_CNTL5_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_CNTL5, L2_CACHE_SMALLK_FRAGMENT_SIZE, 0);
+	WREG32_SOC15(GC, 0, regGCVM_L2_CNTL5, tmp);
+}
+
+static void gfxhub_v3_0_3_enable_system_domain(struct amdgpu_device *adev)
+{
+	uint32_t tmp;
+
+	tmp = RREG32_SOC15(GC, 0, regGCVM_CONTEXT0_CNTL);
+	tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT0_CNTL, ENABLE_CONTEXT, 1);
+	tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT0_CNTL, PAGE_TABLE_DEPTH, 0);
+	tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT0_CNTL,
+			    RETRY_PERMISSION_OR_INVALID_PAGE_FAULT, 0);
+	WREG32_SOC15(GC, 0, regGCVM_CONTEXT0_CNTL, tmp);
+}
+
+static void gfxhub_v3_0_3_disable_identity_aperture(struct amdgpu_device *adev)
+{
+	/* These registers are not accessible to VF-SRIOV.
+	 * The PF will program them instead.
+	 */
+	if (amdgpu_sriov_vf(adev))
+		return;
+
+	WREG32_SOC15(GC, 0, regGCVM_L2_CONTEXT1_IDENTITY_APERTURE_LOW_ADDR_LO32,
+		     0xFFFFFFFF);
+	WREG32_SOC15(GC, 0, regGCVM_L2_CONTEXT1_IDENTITY_APERTURE_LOW_ADDR_HI32,
+		     0x0000000F);
+
+	WREG32_SOC15(GC, 0, regGCVM_L2_CONTEXT1_IDENTITY_APERTURE_HIGH_ADDR_LO32,
+		     0);
+	WREG32_SOC15(GC, 0, regGCVM_L2_CONTEXT1_IDENTITY_APERTURE_HIGH_ADDR_HI32,
+		     0);
+
+	WREG32_SOC15(GC, 0, regGCVM_L2_CONTEXT_IDENTITY_PHYSICAL_OFFSET_LO32, 0);
+	WREG32_SOC15(GC, 0, regGCVM_L2_CONTEXT_IDENTITY_PHYSICAL_OFFSET_HI32, 0);
+
+}
+
+static void gfxhub_v3_0_3_setup_vmid_config(struct amdgpu_device *adev)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)];
+	int i;
+	uint32_t tmp;
+
+	for (i = 0; i <= 14; i++) {
+		tmp = RREG32_SOC15_OFFSET(GC, 0, regGCVM_CONTEXT1_CNTL, i * hub->ctx_distance);
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL, ENABLE_CONTEXT, 1);
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL, PAGE_TABLE_DEPTH,
+				    adev->vm_manager.num_level);
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
+				RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
+				DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
+				PDE0_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
+				VALID_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
+				READ_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
+				WRITE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
+				EXECUTE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
+				PAGE_TABLE_BLOCK_SIZE,
+				adev->vm_manager.block_size - 9);
+		/* Send no-retry XNACK on fault to suppress VM fault storm. */
+		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
+				    RETRY_PERMISSION_OR_INVALID_PAGE_FAULT,
+				    !amdgpu_noretry);
+		WREG32_SOC15_OFFSET(GC, 0, regGCVM_CONTEXT1_CNTL,
+				    i * hub->ctx_distance, tmp);
+		WREG32_SOC15_OFFSET(GC, 0, regGCVM_CONTEXT1_PAGE_TABLE_START_ADDR_LO32,
+				    i * hub->ctx_addr_distance, 0);
+		WREG32_SOC15_OFFSET(GC, 0, regGCVM_CONTEXT1_PAGE_TABLE_START_ADDR_HI32,
+				    i * hub->ctx_addr_distance, 0);
+		WREG32_SOC15_OFFSET(GC, 0, regGCVM_CONTEXT1_PAGE_TABLE_END_ADDR_LO32,
+				    i * hub->ctx_addr_distance,
+				    lower_32_bits(adev->vm_manager.max_pfn - 1));
+		WREG32_SOC15_OFFSET(GC, 0, regGCVM_CONTEXT1_PAGE_TABLE_END_ADDR_HI32,
+				    i * hub->ctx_addr_distance,
+				    upper_32_bits(adev->vm_manager.max_pfn - 1));
+	}
+
+	hub->vm_cntx_cntl = tmp;
+}
+
+static void gfxhub_v3_0_3_program_invalidation(struct amdgpu_device *adev)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)];
+	unsigned int i;
+
+	for (i = 0 ; i < 18; ++i) {
+		WREG32_SOC15_OFFSET(GC, 0, regGCVM_INVALIDATE_ENG0_ADDR_RANGE_LO32,
+				    i * hub->eng_addr_distance, 0xffffffff);
+		WREG32_SOC15_OFFSET(GC, 0, regGCVM_INVALIDATE_ENG0_ADDR_RANGE_HI32,
+				    i * hub->eng_addr_distance, 0x1f);
+	}
+}
+
+static int gfxhub_v3_0_3_gart_enable(struct amdgpu_device *adev)
+{
+	/* GART Enable. */
+	gfxhub_v3_0_3_init_gart_aperture_regs(adev);
+	gfxhub_v3_0_3_init_system_aperture_regs(adev);
+	gfxhub_v3_0_3_init_tlb_regs(adev);
+	gfxhub_v3_0_3_init_cache_regs(adev);
+
+	gfxhub_v3_0_3_enable_system_domain(adev);
+	gfxhub_v3_0_3_disable_identity_aperture(adev);
+	gfxhub_v3_0_3_setup_vmid_config(adev);
+	gfxhub_v3_0_3_program_invalidation(adev);
+
+	return 0;
+}
+
+static void gfxhub_v3_0_3_gart_disable(struct amdgpu_device *adev)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)];
+	u32 tmp;
+	u32 i;
+
+	/* Disable all tables */
+	for (i = 0; i < 16; i++)
+		WREG32_SOC15_OFFSET(GC, 0, regGCVM_CONTEXT0_CNTL,
+				    i * hub->ctx_distance, 0);
+
+	/* Setup TLB control */
+	tmp = RREG32_SOC15(GC, 0, regGCMC_VM_MX_L1_TLB_CNTL);
+	tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB, 0);
+	tmp = REG_SET_FIELD(tmp, GCMC_VM_MX_L1_TLB_CNTL,
+			    ENABLE_ADVANCED_DRIVER_MODEL, 0);
+	WREG32_SOC15(GC, 0, regGCMC_VM_MX_L1_TLB_CNTL, tmp);
+
+	/* Setup L2 cache */
+	WREG32_FIELD15_PREREG(GC, 0, GCVM_L2_CNTL, ENABLE_L2_CACHE, 0);
+	WREG32_SOC15(GC, 0, regGCVM_L2_CNTL3, 0);
+}
+
+/**
+ * gfxhub_v3_0_3_set_fault_enable_default - update GART/VM fault handling
+ *
+ * @adev: amdgpu_device pointer
+ * @value: true redirects VM faults to the default page
+ */
+static void gfxhub_v3_0_3_set_fault_enable_default(struct amdgpu_device *adev,
+					  bool value)
+{
+	u32 tmp;
+
+	/* These registers are not accessible to VF-SRIOV.
+	 * The PF will program them instead.
+	 */
+	if (amdgpu_sriov_vf(adev))
+		return;
+
+	tmp = RREG32_SOC15(GC, 0, regGCVM_L2_PROTECTION_FAULT_CNTL);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    PDE0_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    PDE1_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    PDE2_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    TRANSLATE_FURTHER_PROTECTION_FAULT_ENABLE_DEFAULT,
+			    value);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    NACK_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    VALID_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    READ_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    WRITE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+			    EXECUTE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	if (!value) {
+		tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+				CRASH_ON_NO_RETRY_FAULT, 1);
+		tmp = REG_SET_FIELD(tmp, GCVM_L2_PROTECTION_FAULT_CNTL,
+				CRASH_ON_RETRY_FAULT, 1);
+	}
+	WREG32_SOC15(GC, 0, regGCVM_L2_PROTECTION_FAULT_CNTL, tmp);
+}
+
+static const struct amdgpu_vmhub_funcs gfxhub_v3_0_3_vmhub_funcs = {
+	.print_l2_protection_fault_status = gfxhub_v3_0_3_print_l2_protection_fault_status,
+	.get_invalidate_req = gfxhub_v3_0_3_get_invalidate_req,
+};
+
+static void gfxhub_v3_0_3_init(struct amdgpu_device *adev)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)];
+
+	hub->ctx0_ptb_addr_lo32 =
+		SOC15_REG_OFFSET(GC, 0,
+				 regGCVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32);
+	hub->ctx0_ptb_addr_hi32 =
+		SOC15_REG_OFFSET(GC, 0,
+				 regGCVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32);
+	hub->vm_inv_eng0_sem =
+		SOC15_REG_OFFSET(GC, 0, regGCVM_INVALIDATE_ENG0_SEM);
+	hub->vm_inv_eng0_req =
+		SOC15_REG_OFFSET(GC, 0, regGCVM_INVALIDATE_ENG0_REQ);
+	hub->vm_inv_eng0_ack =
+		SOC15_REG_OFFSET(GC, 0, regGCVM_INVALIDATE_ENG0_ACK);
+	hub->vm_context0_cntl =
+		SOC15_REG_OFFSET(GC, 0, regGCVM_CONTEXT0_CNTL);
+	hub->vm_l2_pro_fault_status =
+		SOC15_REG_OFFSET(GC, 0, regGCVM_L2_PROTECTION_FAULT_STATUS);
+	hub->vm_l2_pro_fault_cntl =
+		SOC15_REG_OFFSET(GC, 0, regGCVM_L2_PROTECTION_FAULT_CNTL);
+
+	hub->ctx_distance = regGCVM_CONTEXT1_CNTL - regGCVM_CONTEXT0_CNTL;
+	hub->ctx_addr_distance = regGCVM_CONTEXT1_PAGE_TABLE_BASE_ADDR_LO32 -
+		regGCVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32;
+	hub->eng_distance = regGCVM_INVALIDATE_ENG1_REQ -
+		regGCVM_INVALIDATE_ENG0_REQ;
+	hub->eng_addr_distance = regGCVM_INVALIDATE_ENG1_ADDR_RANGE_LO32 -
+		regGCVM_INVALIDATE_ENG0_ADDR_RANGE_LO32;
+
+	hub->vm_cntx_cntl_vm_fault = GCVM_CONTEXT1_CNTL__RANGE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		GCVM_CONTEXT1_CNTL__DUMMY_PAGE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		GCVM_CONTEXT1_CNTL__PDE0_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		GCVM_CONTEXT1_CNTL__VALID_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		GCVM_CONTEXT1_CNTL__READ_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		GCVM_CONTEXT1_CNTL__WRITE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		GCVM_CONTEXT1_CNTL__EXECUTE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK;
+
+	hub->vmhub_funcs = &gfxhub_v3_0_3_vmhub_funcs;
+}
+
+const struct amdgpu_gfxhub_funcs gfxhub_v3_0_3_funcs = {
+	.get_fb_location = gfxhub_v3_0_3_get_fb_location,
+	.get_mc_fb_offset = gfxhub_v3_0_3_get_mc_fb_offset,
+	.setup_vm_pt_regs = gfxhub_v3_0_3_setup_vm_pt_regs,
+	.gart_enable = gfxhub_v3_0_3_gart_enable,
+	.gart_disable = gfxhub_v3_0_3_gart_disable,
+	.set_fault_enable_default = gfxhub_v3_0_3_set_fault_enable_default,
+	.init = gfxhub_v3_0_3_init,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/gfxhub_v3_0_3.h b/drivers/gpu/drm/amd/amdgpu/gfxhub_v3_0_3.h
new file mode 100644
index 000000000000..6153bd5e3083
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/gfxhub_v3_0_3.h
@@ -0,0 +1,29 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __GFXHUB_V3_0_3_H__
+#define __GFXHUB_V3_0_3_H__
+
+extern const struct amdgpu_gfxhub_funcs gfxhub_v3_0_3_funcs;
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
index 3ec5ff5a6dbe..ce6e04242c52 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
@@ -22,6 +22,9 @@
  */
 #include <linux/firmware.h>
 #include <linux/pci.h>
+
+#include <drm/drm_cache.h>
+
 #include "amdgpu.h"
 #include "amdgpu_atomfirmware.h"
 #include "gmc_v10_0.h"
@@ -48,16 +51,9 @@
 #include "athub_v2_0.h"
 #include "athub_v2_1.h"
 
-#if 0
-static const struct soc15_reg_golden golden_settings_navi10_hdp[] =
-{
-	/* TODO add golden setting for hdp */
-};
-#endif
-
 static int gmc_v10_0_ecc_interrupt_state(struct amdgpu_device *adev,
 					 struct amdgpu_irq_src *src,
-					 unsigned type,
+					 unsigned int type,
 					 enum amdgpu_interrupt_state state)
 {
 	return 0;
@@ -65,21 +61,33 @@ static int gmc_v10_0_ecc_interrupt_state(struct amdgpu_device *adev,
 
 static int
 gmc_v10_0_vm_fault_interrupt_state(struct amdgpu_device *adev,
-				   struct amdgpu_irq_src *src, unsigned type,
+				   struct amdgpu_irq_src *src, unsigned int type,
 				   enum amdgpu_interrupt_state state)
 {
 	switch (state) {
 	case AMDGPU_IRQ_STATE_DISABLE:
 		/* MM HUB */
-		amdgpu_gmc_set_vm_fault_masks(adev, AMDGPU_MMHUB_0, false);
+		amdgpu_gmc_set_vm_fault_masks(adev, AMDGPU_MMHUB0(0), false);
 		/* GFX HUB */
-		amdgpu_gmc_set_vm_fault_masks(adev, AMDGPU_GFXHUB_0, false);
+		/* This works because this interrupt is only
+		 * enabled at init/resume and disabled in
+		 * fini/suspend, so the overall state doesn't
+		 * change over the course of suspend/resume.
+		 */
+		if (!adev->in_s0ix)
+			amdgpu_gmc_set_vm_fault_masks(adev, AMDGPU_GFXHUB(0), false);
 		break;
 	case AMDGPU_IRQ_STATE_ENABLE:
 		/* MM HUB */
-		amdgpu_gmc_set_vm_fault_masks(adev, AMDGPU_MMHUB_0, true);
+		amdgpu_gmc_set_vm_fault_masks(adev, AMDGPU_MMHUB0(0), true);
 		/* GFX HUB */
-		amdgpu_gmc_set_vm_fault_masks(adev, AMDGPU_GFXHUB_0, true);
+		/* This works because this interrupt is only
+		 * enabled at init/resume and disabled in
+		 * fini/suspend, so the overall state doesn't
+		 * change over the course of suspend/resume.
+		 */
+		if (!adev->in_s0ix)
+			amdgpu_gmc_set_vm_fault_masks(adev, AMDGPU_GFXHUB(0), true);
 		break;
 	default:
 		break;
@@ -92,10 +100,14 @@ static int gmc_v10_0_process_interrupt(struct amdgpu_device *adev,
 				       struct amdgpu_irq_src *source,
 				       struct amdgpu_iv_entry *entry)
 {
-	bool retry_fault = !!(entry->src_data[1] & 0x80);
-	bool write_fault = !!(entry->src_data[1] & 0x20);
-	struct amdgpu_vmhub *hub = &adev->vmhub[entry->vmid_src];
-	struct amdgpu_task_info task_info;
+	uint32_t vmhub_index = entry->client_id == SOC15_IH_CLIENTID_VMC ?
+			       AMDGPU_MMHUB0(0) : AMDGPU_GFXHUB(0);
+	struct amdgpu_vmhub *hub = &adev->vmhub[vmhub_index];
+	bool retry_fault = !!(entry->src_data[1] &
+			      AMDGPU_GMC9_FAULT_SOURCE_DATA_RETRY);
+	bool write_fault = !!(entry->src_data[1] &
+			      AMDGPU_GMC9_FAULT_SOURCE_DATA_WRITE);
+	struct amdgpu_task_info *task_info;
 	uint32_t status = 0;
 	u64 addr;
 
@@ -107,7 +119,7 @@ static int gmc_v10_0_process_interrupt(struct amdgpu_device *adev,
 
 		/* Process it onyl if it's the first fault for this address */
 		if (entry->ih != &adev->irq.ih_soft &&
-		    amdgpu_gmc_filter_faults(adev, addr, entry->pasid,
+		    amdgpu_gmc_filter_faults(adev, entry->ih, addr, entry->pasid,
 					     entry->timestamp))
 			return 1;
 
@@ -122,7 +134,8 @@ static int gmc_v10_0_process_interrupt(struct amdgpu_device *adev,
 		/* Try to handle the recoverable page faults by filling page
 		 * tables
 		 */
-		if (amdgpu_vm_handle_fault(adev, entry->pasid, addr, write_fault))
+		if (amdgpu_vm_handle_fault(adev, entry->pasid, 0, 0, addr,
+					   entry->timestamp, write_fault))
 			return 1;
 	}
 
@@ -132,32 +145,39 @@ static int gmc_v10_0_process_interrupt(struct amdgpu_device *adev,
 		 * be updated to avoid reading an incorrect value due to
 		 * the new fast GRBM interface.
 		 */
-		if ((entry->vmid_src == AMDGPU_GFXHUB_0) &&
-		    (adev->ip_versions[GC_HWIP][0] < IP_VERSION(10, 3, 0)))
+		if ((entry->vmid_src == AMDGPU_GFXHUB(0)) &&
+		    (amdgpu_ip_version(adev, GC_HWIP, 0) <
+		     IP_VERSION(10, 3, 0)))
 			RREG32(hub->vm_l2_pro_fault_status);
 
 		status = RREG32(hub->vm_l2_pro_fault_status);
 		WREG32_P(hub->vm_l2_pro_fault_cntl, 1, ~1);
+
+		amdgpu_vm_update_fault_cache(adev, entry->pasid, addr, status,
+					     entry->vmid_src ? AMDGPU_MMHUB0(0) : AMDGPU_GFXHUB(0));
 	}
 
 	if (!printk_ratelimit())
 		return 0;
 
-	memset(&task_info, 0, sizeof(struct amdgpu_task_info));
-	amdgpu_vm_get_task_info(adev, entry->pasid, &task_info);
-
 	dev_err(adev->dev,
-		"[%s] page fault (src_id:%u ring:%u vmid:%u pasid:%u, "
-		"for process %s pid %d thread %s pid %d)\n",
+		"[%s] page fault (src_id:%u ring:%u vmid:%u pasid:%u)\n",
 		entry->vmid_src ? "mmhub" : "gfxhub",
-		entry->src_id, entry->ring_id, entry->vmid,
-		entry->pasid, task_info.process_name, task_info.tgid,
-		task_info.task_name, task_info.pid);
+		entry->src_id, entry->ring_id, entry->vmid, entry->pasid);
+	task_info = amdgpu_vm_get_task_info_pasid(adev, entry->pasid);
+	if (task_info) {
+		amdgpu_vm_print_task_info(adev, task_info);
+		amdgpu_vm_put_task_info(task_info);
+	}
+
 	dev_err(adev->dev, "  in page starting at address 0x%016llx from client 0x%x (%s)\n",
-		addr, entry->client_id,
-		soc15_ih_clientid_name[entry->client_id]);
+			addr, entry->client_id,
+			soc15_ih_clientid_name[entry->client_id]);
 
-	if (!amdgpu_sriov_vf(adev))
+	/* Only print L2 fault status if the status register could be read and
+	 * contains useful information
+	 */
+	if (status != 0)
 		hub->vmhub_funcs->print_l2_protection_fault_status(adev,
 								   status);
 
@@ -195,8 +215,7 @@ static void gmc_v10_0_set_irq_funcs(struct amdgpu_device *adev)
 static bool gmc_v10_0_use_invalidate_semaphore(struct amdgpu_device *adev,
 				       uint32_t vmhub)
 {
-	return ((vmhub == AMDGPU_MMHUB_0 ||
-		 vmhub == AMDGPU_MMHUB_1) &&
+	return ((vmhub == AMDGPU_MMHUB0(0)) &&
 		(!amdgpu_sriov_vf(adev)));
 }
 
@@ -220,20 +239,48 @@ static bool gmc_v10_0_get_atc_vmid_pasid_mapping_info(
  * by the amdgpu vm/hsa code.
  */
 
-static void gmc_v10_0_flush_vm_hub(struct amdgpu_device *adev, uint32_t vmid,
-				   unsigned int vmhub, uint32_t flush_type)
+/**
+ * gmc_v10_0_flush_gpu_tlb - gart tlb flush callback
+ *
+ * @adev: amdgpu_device pointer
+ * @vmid: vm instance to flush
+ * @vmhub: vmhub type
+ * @flush_type: the flush type
+ *
+ * Flush the TLB for the requested page table.
+ */
+static void gmc_v10_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid,
+					uint32_t vmhub, uint32_t flush_type)
 {
 	bool use_semaphore = gmc_v10_0_use_invalidate_semaphore(adev, vmhub);
 	struct amdgpu_vmhub *hub = &adev->vmhub[vmhub];
 	u32 inv_req = hub->vmhub_funcs->get_invalidate_req(vmid, flush_type);
-	u32 tmp;
 	/* Use register 17 for GART */
-	const unsigned eng = 17;
-	unsigned int i;
+	const unsigned int eng = 17;
 	unsigned char hub_ip = 0;
+	u32 sem, req, ack;
+	unsigned int i;
+	u32 tmp;
+
+	sem = hub->vm_inv_eng0_sem + hub->eng_distance * eng;
+	req = hub->vm_inv_eng0_req + hub->eng_distance * eng;
+	ack = hub->vm_inv_eng0_ack + hub->eng_distance * eng;
 
-	hub_ip = (vmhub == AMDGPU_GFXHUB_0) ?
-		   GC_HWIP : MMHUB_HWIP;
+	/* flush hdp cache */
+	amdgpu_device_flush_hdp(adev, NULL);
+
+	/* This is necessary for SRIOV as well as for GFXOFF to function
+	 * properly under bare metal
+	 */
+	if (adev->gfx.kiq[0].ring.sched.ready && !adev->enable_mes &&
+	    (amdgpu_sriov_runtime(adev) || !amdgpu_sriov_vf(adev))) {
+		amdgpu_gmc_fw_reg_write_reg_wait(adev, req, ack, inv_req,
+						 1 << vmid, GET_INST(GC, 0));
+		return;
+	}
+
+	/* This path is needed before KIQ/MES/GFXOFF are set up */
+	hub_ip = (vmhub == AMDGPU_GFXHUB(0)) ? GC_HWIP : MMHUB_HWIP;
 
 	spin_lock(&adev->gmc.invalidate_lock);
 	/*
@@ -247,9 +294,7 @@ static void gmc_v10_0_flush_vm_hub(struct amdgpu_device *adev, uint32_t vmid,
 	if (use_semaphore) {
 		for (i = 0; i < adev->usec_timeout; i++) {
 			/* a read return value of 1 means semaphore acuqire */
-			tmp = RREG32_RLC_NO_KIQ(hub->vm_inv_eng0_sem +
-					 hub->eng_distance * eng, hub_ip);
-
+			tmp = RREG32_RLC_NO_KIQ(sem, hub_ip);
 			if (tmp & 0x1)
 				break;
 			udelay(1);
@@ -259,24 +304,19 @@ static void gmc_v10_0_flush_vm_hub(struct amdgpu_device *adev, uint32_t vmid,
 			DRM_ERROR("Timeout waiting for sem acquire in VM flush!\n");
 	}
 
-	WREG32_RLC_NO_KIQ(hub->vm_inv_eng0_req +
-			  hub->eng_distance * eng,
-			  inv_req, hub_ip);
+	WREG32_RLC_NO_KIQ(req, inv_req, hub_ip);
 
 	/*
 	 * Issue a dummy read to wait for the ACK register to be cleared
 	 * to avoid a false ACK due to the new fast GRBM interface.
 	 */
-	if ((vmhub == AMDGPU_GFXHUB_0) &&
-	    (adev->ip_versions[GC_HWIP][0] < IP_VERSION(10, 3, 0)))
-		RREG32_RLC_NO_KIQ(hub->vm_inv_eng0_req +
-				  hub->eng_distance * eng, hub_ip);
+	if ((vmhub == AMDGPU_GFXHUB(0)) &&
+	    (amdgpu_ip_version(adev, GC_HWIP, 0) < IP_VERSION(10, 3, 0)))
+		RREG32_RLC_NO_KIQ(req, hub_ip);
 
 	/* Wait for ACK with a delay.*/
 	for (i = 0; i < adev->usec_timeout; i++) {
-		tmp = RREG32_RLC_NO_KIQ(hub->vm_inv_eng0_ack +
-				  hub->eng_distance * eng, hub_ip);
-
+		tmp = RREG32_RLC_NO_KIQ(ack, hub_ip);
 		tmp &= 1 << vmid;
 		if (tmp)
 			break;
@@ -286,113 +326,13 @@ static void gmc_v10_0_flush_vm_hub(struct amdgpu_device *adev, uint32_t vmid,
 
 	/* TODO: It needs to continue working on debugging with semaphore for GFXHUB as well. */
 	if (use_semaphore)
-		/*
-		 * add semaphore release after invalidation,
-		 * write with 0 means semaphore release
-		 */
-		WREG32_RLC_NO_KIQ(hub->vm_inv_eng0_sem +
-				  hub->eng_distance * eng, 0, hub_ip);
+		WREG32_RLC_NO_KIQ(sem, 0, hub_ip);
 
 	spin_unlock(&adev->gmc.invalidate_lock);
 
-	if (i < adev->usec_timeout)
-		return;
-
-	DRM_ERROR("Timeout waiting for VM flush hub: %d!\n", vmhub);
-}
-
-/**
- * gmc_v10_0_flush_gpu_tlb - gart tlb flush callback
- *
- * @adev: amdgpu_device pointer
- * @vmid: vm instance to flush
- * @vmhub: vmhub type
- * @flush_type: the flush type
- *
- * Flush the TLB for the requested page table.
- */
-static void gmc_v10_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid,
-					uint32_t vmhub, uint32_t flush_type)
-{
-	struct amdgpu_ring *ring = adev->mman.buffer_funcs_ring;
-	struct dma_fence *fence;
-	struct amdgpu_job *job;
-
-	int r;
-
-	/* flush hdp cache */
-	adev->hdp.funcs->flush_hdp(adev, NULL);
-
-	/* For SRIOV run time, driver shouldn't access the register through MMIO
-	 * Directly use kiq to do the vm invalidation instead
-	 */
-	if (adev->gfx.kiq.ring.sched.ready &&
-	    (amdgpu_sriov_runtime(adev) || !amdgpu_sriov_vf(adev)) &&
-	    down_read_trylock(&adev->reset_sem)) {
-		struct amdgpu_vmhub *hub = &adev->vmhub[vmhub];
-		const unsigned eng = 17;
-		u32 inv_req = hub->vmhub_funcs->get_invalidate_req(vmid, flush_type);
-		u32 req = hub->vm_inv_eng0_req + hub->eng_distance * eng;
-		u32 ack = hub->vm_inv_eng0_ack + hub->eng_distance * eng;
-
-		amdgpu_virt_kiq_reg_write_reg_wait(adev, req, ack, inv_req,
-				1 << vmid);
-
-		up_read(&adev->reset_sem);
-		return;
-	}
-
-	mutex_lock(&adev->mman.gtt_window_lock);
-
-	if (vmhub == AMDGPU_MMHUB_0) {
-		gmc_v10_0_flush_vm_hub(adev, vmid, AMDGPU_MMHUB_0, 0);
-		mutex_unlock(&adev->mman.gtt_window_lock);
-		return;
-	}
-
-	BUG_ON(vmhub != AMDGPU_GFXHUB_0);
-
-	if (!adev->mman.buffer_funcs_enabled ||
-	    !adev->ib_pool_ready ||
-	    amdgpu_in_reset(adev) ||
-	    ring->sched.ready == false) {
-		gmc_v10_0_flush_vm_hub(adev, vmid, AMDGPU_GFXHUB_0, 0);
-		mutex_unlock(&adev->mman.gtt_window_lock);
-		return;
-	}
-
-	/* The SDMA on Navi has a bug which can theoretically result in memory
-	 * corruption if an invalidation happens at the same time as an VA
-	 * translation. Avoid this by doing the invalidation from the SDMA
-	 * itself.
-	 */
-	r = amdgpu_job_alloc_with_ib(adev, 16 * 4, AMDGPU_IB_POOL_IMMEDIATE,
-				     &job);
-	if (r)
-		goto error_alloc;
-
-	job->vm_pd_addr = amdgpu_gmc_pd_addr(adev->gart.bo);
-	job->vm_needs_flush = true;
-	job->ibs->ptr[job->ibs->length_dw++] = ring->funcs->nop;
-	amdgpu_ring_pad_ib(ring, &job->ibs[0]);
-	r = amdgpu_job_submit(job, &adev->mman.entity,
-			      AMDGPU_FENCE_OWNER_UNDEFINED, &fence);
-	if (r)
-		goto error_submit;
-
-	mutex_unlock(&adev->mman.gtt_window_lock);
-
-	dma_fence_wait(fence, false);
-	dma_fence_put(fence);
-
-	return;
-
-error_submit:
-	amdgpu_job_free(job);
-
-error_alloc:
-	mutex_unlock(&adev->mman.gtt_window_lock);
-	DRM_ERROR("Error flushing GPU TLB using the SDMA (%d)!\n", r);
+	if (i >= adev->usec_timeout)
+		dev_err(adev->dev, "Timeout waiting for VM flush hub: %d!\n",
+			vmhub);
 }
 
 /**
@@ -402,72 +342,44 @@ error_alloc:
  * @pasid: pasid to be flush
  * @flush_type: the flush type
  * @all_hub: Used with PACKET3_INVALIDATE_TLBS_ALL_HUB()
+ * @inst: is used to select which instance of KIQ to use for the invalidation
  *
  * Flush the TLB for the requested pasid.
  */
-static int gmc_v10_0_flush_gpu_tlb_pasid(struct amdgpu_device *adev,
-					uint16_t pasid, uint32_t flush_type,
-					bool all_hub)
+static void gmc_v10_0_flush_gpu_tlb_pasid(struct amdgpu_device *adev,
+					  uint16_t pasid, uint32_t flush_type,
+					  bool all_hub, uint32_t inst)
 {
+	uint16_t queried;
 	int vmid, i;
-	signed long r;
-	uint32_t seq;
-	uint16_t queried_pasid;
-	bool ret;
-	struct amdgpu_ring *ring = &adev->gfx.kiq.ring;
-	struct amdgpu_kiq *kiq = &adev->gfx.kiq;
-
-	if (amdgpu_emu_mode == 0 && ring->sched.ready) {
-		spin_lock(&adev->gfx.kiq.ring_lock);
-		/* 2 dwords flush + 8 dwords fence */
-		amdgpu_ring_alloc(ring, kiq->pmf->invalidate_tlbs_size + 8);
-		kiq->pmf->kiq_invalidate_tlbs(ring,
-					pasid, flush_type, all_hub);
-		r = amdgpu_fence_emit_polling(ring, &seq, MAX_KIQ_REG_WAIT);
-		if (r) {
-			amdgpu_ring_undo(ring);
-			spin_unlock(&adev->gfx.kiq.ring_lock);
-			return -ETIME;
-		}
-
-		amdgpu_ring_commit(ring);
-		spin_unlock(&adev->gfx.kiq.ring_lock);
-		r = amdgpu_fence_wait_polling(ring, seq, adev->usec_timeout);
-		if (r < 1) {
-			dev_err(adev->dev, "wait for kiq fence error: %ld.\n", r);
-			return -ETIME;
-		}
-
-		return 0;
-	}
 
 	for (vmid = 1; vmid < AMDGPU_NUM_VMID; vmid++) {
-
-		ret = gmc_v10_0_get_atc_vmid_pasid_mapping_info(adev, vmid,
-				&queried_pasid);
-		if (ret	&& queried_pasid == pasid) {
-			if (all_hub) {
-				for (i = 0; i < adev->num_vmhubs; i++)
-					gmc_v10_0_flush_gpu_tlb(adev, vmid,
-							i, flush_type);
-			} else {
-				gmc_v10_0_flush_gpu_tlb(adev, vmid,
-						AMDGPU_GFXHUB_0, flush_type);
-			}
-			break;
+		bool valid;
+
+		valid = gmc_v10_0_get_atc_vmid_pasid_mapping_info(adev, vmid,
+								  &queried);
+		if (!valid || queried != pasid)
+			continue;
+
+		if (all_hub) {
+			for_each_set_bit(i, adev->vmhubs_mask,
+					 AMDGPU_MAX_VMHUBS)
+				gmc_v10_0_flush_gpu_tlb(adev, vmid, i,
+							flush_type);
+		} else {
+			gmc_v10_0_flush_gpu_tlb(adev, vmid, AMDGPU_GFXHUB(0),
+						flush_type);
 		}
 	}
-
-	return 0;
 }
 
 static uint64_t gmc_v10_0_emit_flush_gpu_tlb(struct amdgpu_ring *ring,
-					     unsigned vmid, uint64_t pd_addr)
+					     unsigned int vmid, uint64_t pd_addr)
 {
-	bool use_semaphore = gmc_v10_0_use_invalidate_semaphore(ring->adev, ring->funcs->vmhub);
-	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
+	bool use_semaphore = gmc_v10_0_use_invalidate_semaphore(ring->adev, ring->vm_hub);
+	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->vm_hub];
 	uint32_t req = hub->vmhub_funcs->get_invalidate_req(vmid, 0);
-	unsigned eng = ring->vm_inv_eng;
+	unsigned int eng = ring->vm_inv_eng;
 
 	/*
 	 * It may lose gpuvm invalidate acknowldege state across power-gating
@@ -509,13 +421,13 @@ static uint64_t gmc_v10_0_emit_flush_gpu_tlb(struct amdgpu_ring *ring,
 	return pd_addr;
 }
 
-static void gmc_v10_0_emit_pasid_mapping(struct amdgpu_ring *ring, unsigned vmid,
-					 unsigned pasid)
+static void gmc_v10_0_emit_pasid_mapping(struct amdgpu_ring *ring, unsigned int vmid,
+					 unsigned int pasid)
 {
 	struct amdgpu_device *adev = ring->adev;
 	uint32_t reg;
 
-	if (ring->funcs->vmhub == AMDGPU_GFXHUB_0)
+	if (ring->vm_hub == AMDGPU_GFXHUB(0))
 		reg = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_VMID_0_LUT) + vmid;
 	else
 		reg = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_VMID_0_LUT_MM) + vmid;
@@ -556,24 +468,6 @@ static void gmc_v10_0_emit_pasid_mapping(struct amdgpu_ring *ring, unsigned vmid
  * 0 valid
  */
 
-static uint64_t gmc_v10_0_map_mtype(struct amdgpu_device *adev, uint32_t flags)
-{
-	switch (flags) {
-	case AMDGPU_VM_MTYPE_DEFAULT:
-		return AMDGPU_PTE_MTYPE_NV10(MTYPE_NC);
-	case AMDGPU_VM_MTYPE_NC:
-		return AMDGPU_PTE_MTYPE_NV10(MTYPE_NC);
-	case AMDGPU_VM_MTYPE_WC:
-		return AMDGPU_PTE_MTYPE_NV10(MTYPE_WC);
-	case AMDGPU_VM_MTYPE_CC:
-		return AMDGPU_PTE_MTYPE_NV10(MTYPE_CC);
-	case AMDGPU_VM_MTYPE_UC:
-		return AMDGPU_PTE_MTYPE_NV10(MTYPE_UC);
-	default:
-		return AMDGPU_PTE_MTYPE_NV10(MTYPE_NC);
-	}
-}
-
 static void gmc_v10_0_get_vm_pde(struct amdgpu_device *adev, int level,
 				 uint64_t *addr, uint64_t *flags)
 {
@@ -598,28 +492,56 @@ static void gmc_v10_0_get_vm_pde(struct amdgpu_device *adev, int level,
 }
 
 static void gmc_v10_0_get_vm_pte(struct amdgpu_device *adev,
-				 struct amdgpu_bo_va_mapping *mapping,
+				 struct amdgpu_vm *vm,
+				 struct amdgpu_bo *bo,
+				 uint32_t vm_flags,
 				 uint64_t *flags)
 {
-	*flags &= ~AMDGPU_PTE_EXECUTABLE;
-	*flags |= mapping->flags & AMDGPU_PTE_EXECUTABLE;
+	if (vm_flags & AMDGPU_VM_PAGE_EXECUTABLE)
+		*flags |= AMDGPU_PTE_EXECUTABLE;
+	else
+		*flags &= ~AMDGPU_PTE_EXECUTABLE;
 
-	*flags &= ~AMDGPU_PTE_MTYPE_NV10_MASK;
-	*flags |= (mapping->flags & AMDGPU_PTE_MTYPE_NV10_MASK);
+	switch (vm_flags & AMDGPU_VM_MTYPE_MASK) {
+	case AMDGPU_VM_MTYPE_DEFAULT:
+	case AMDGPU_VM_MTYPE_NC:
+	default:
+		*flags = AMDGPU_PTE_MTYPE_NV10(*flags, MTYPE_NC);
+		break;
+	case AMDGPU_VM_MTYPE_WC:
+		*flags = AMDGPU_PTE_MTYPE_NV10(*flags, MTYPE_WC);
+		break;
+	case AMDGPU_VM_MTYPE_CC:
+		*flags = AMDGPU_PTE_MTYPE_NV10(*flags, MTYPE_CC);
+		break;
+	case AMDGPU_VM_MTYPE_UC:
+		*flags = AMDGPU_PTE_MTYPE_NV10(*flags, MTYPE_UC);
+		break;
+	}
 
-	if (mapping->flags & AMDGPU_PTE_PRT) {
+	if (vm_flags & AMDGPU_VM_PAGE_NOALLOC)
+		*flags |= AMDGPU_PTE_NOALLOC;
+	else
+		*flags &= ~AMDGPU_PTE_NOALLOC;
+
+	if (vm_flags & AMDGPU_VM_PAGE_PRT) {
 		*flags |= AMDGPU_PTE_PRT;
 		*flags |= AMDGPU_PTE_SNOOPED;
 		*flags |= AMDGPU_PTE_LOG;
 		*flags |= AMDGPU_PTE_SYSTEM;
 		*flags &= ~AMDGPU_PTE_VALID;
 	}
+
+	if (bo && bo->flags & (AMDGPU_GEM_CREATE_COHERENT |
+			       AMDGPU_GEM_CREATE_EXT_COHERENT |
+			       AMDGPU_GEM_CREATE_UNCACHED))
+		*flags = AMDGPU_PTE_MTYPE_NV10(*flags, MTYPE_UC);
 }
 
-static unsigned gmc_v10_0_get_vbios_fb_size(struct amdgpu_device *adev)
+static unsigned int gmc_v10_0_get_vbios_fb_size(struct amdgpu_device *adev)
 {
 	u32 d1vga_control = RREG32_SOC15(DCE, 0, mmD1VGA_CONTROL);
-	unsigned size;
+	unsigned int size;
 
 	if (REG_GET_FIELD(d1vga_control, D1VGA_CONTROL, D1VGA_MODE_ENABLE)) {
 		size = AMDGPU_VBIOS_VGA_ALLOCATION;
@@ -643,7 +565,6 @@ static const struct amdgpu_gmc_funcs gmc_v10_0_gmc_funcs = {
 	.flush_gpu_tlb_pasid = gmc_v10_0_flush_gpu_tlb_pasid,
 	.emit_flush_gpu_tlb = gmc_v10_0_emit_flush_gpu_tlb,
 	.emit_pasid_mapping = gmc_v10_0_emit_pasid_mapping,
-	.map_mtype = gmc_v10_0_map_mtype,
 	.get_vm_pde = gmc_v10_0_get_vm_pde,
 	.get_vm_pte = gmc_v10_0_get_vm_pte,
 	.get_vbios_fb_size = gmc_v10_0_get_vbios_fb_size,
@@ -657,26 +578,27 @@ static void gmc_v10_0_set_gmc_funcs(struct amdgpu_device *adev)
 
 static void gmc_v10_0_set_umc_funcs(struct amdgpu_device *adev)
 {
-	switch (adev->ip_versions[UMC_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, UMC_HWIP, 0)) {
 	case IP_VERSION(8, 7, 0):
 		adev->umc.max_ras_err_cnt_per_query = UMC_V8_7_TOTAL_CHANNEL_NUM;
 		adev->umc.channel_inst_num = UMC_V8_7_CHANNEL_INSTANCE_NUM;
 		adev->umc.umc_inst_num = UMC_V8_7_UMC_INSTANCE_NUM;
 		adev->umc.channel_offs = UMC_V8_7_PER_CHANNEL_OFFSET_SIENNA;
+		adev->umc.retire_unit = 1;
 		adev->umc.channel_idx_tbl = &umc_v8_7_channel_idx_tbl[0][0];
-		adev->umc.ras_funcs = &umc_v8_7_ras_funcs;
+		adev->umc.ras = &umc_v8_7_ras;
 		break;
 	default:
 		break;
 	}
 }
 
-
 static void gmc_v10_0_set_mmhub_funcs(struct amdgpu_device *adev)
 {
-	switch (adev->ip_versions[MMHUB_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, MMHUB_HWIP, 0)) {
 	case IP_VERSION(2, 3, 0):
 	case IP_VERSION(2, 4, 0):
+	case IP_VERSION(2, 4, 1):
 		adev->mmhub.funcs = &mmhub_v2_3_funcs;
 		break;
 	default:
@@ -687,13 +609,15 @@ static void gmc_v10_0_set_mmhub_funcs(struct amdgpu_device *adev)
 
 static void gmc_v10_0_set_gfxhub_funcs(struct amdgpu_device *adev)
 {
-	switch (adev->ip_versions[GC_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 	case IP_VERSION(10, 3, 0):
 	case IP_VERSION(10, 3, 2):
 	case IP_VERSION(10, 3, 1):
 	case IP_VERSION(10, 3, 4):
 	case IP_VERSION(10, 3, 5):
+	case IP_VERSION(10, 3, 6):
 	case IP_VERSION(10, 3, 3):
+	case IP_VERSION(10, 3, 7):
 		adev->gfxhub.funcs = &gfxhub_v2_1_funcs;
 		break;
 	default:
@@ -703,9 +627,9 @@ static void gmc_v10_0_set_gfxhub_funcs(struct amdgpu_device *adev)
 }
 
 
-static int gmc_v10_0_early_init(void *handle)
+static int gmc_v10_0_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	gmc_v10_0_set_mmhub_funcs(adev);
 	gmc_v10_0_set_gfxhub_funcs(adev);
@@ -719,13 +643,14 @@ static int gmc_v10_0_early_init(void *handle)
 	adev->gmc.private_aperture_start = 0x1000000000000000ULL;
 	adev->gmc.private_aperture_end =
 		adev->gmc.private_aperture_start + (4ULL << 30) - 1;
+	adev->gmc.noretry_flags = AMDGPU_VM_NORETRY_FLAGS_TF;
 
 	return 0;
 }
 
-static int gmc_v10_0_late_init(void *handle)
+static int gmc_v10_0_late_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int r;
 
 	r = amdgpu_gmc_allocate_vm_inv_eng(adev);
@@ -749,9 +674,11 @@ static void gmc_v10_0_vram_gtt_location(struct amdgpu_device *adev,
 	/* add the xgmi offset of the physical node */
 	base += adev->gmc.xgmi.physical_node_id * adev->gmc.xgmi.node_segment_size;
 
+	amdgpu_gmc_set_agp_default(adev, mc);
 	amdgpu_gmc_vram_location(adev, &adev->gmc, base);
-	amdgpu_gmc_gart_location(adev, mc);
-	amdgpu_gmc_agp_location(adev, mc);
+	amdgpu_gmc_gart_location(adev, mc, AMDGPU_GART_PLACEMENT_BEST_FIT);
+	if (!amdgpu_sriov_vf(adev) && (amdgpu_agp == 1))
+		amdgpu_gmc_agp_location(adev, mc);
 
 	/* base offset of vram pages */
 	adev->vm_manager.vram_base_offset = adev->gfxhub.funcs->get_mc_fb_offset(adev);
@@ -788,22 +715,30 @@ static int gmc_v10_0_mc_init(struct amdgpu_device *adev)
 	adev->gmc.aper_size = pci_resource_len(adev->pdev, 0);
 
 #ifdef CONFIG_X86_64
-	if (adev->flags & AMD_IS_APU) {
+	if ((adev->flags & AMD_IS_APU) && !amdgpu_passthrough(adev)) {
 		adev->gmc.aper_base = adev->gfxhub.funcs->get_mc_fb_offset(adev);
 		adev->gmc.aper_size = adev->gmc.real_vram_size;
 	}
 #endif
 
-	/* In case the PCI BAR is larger than the actual amount of vram */
 	adev->gmc.visible_vram_size = adev->gmc.aper_size;
-	if (adev->gmc.visible_vram_size > adev->gmc.real_vram_size)
-		adev->gmc.visible_vram_size = adev->gmc.real_vram_size;
 
 	/* set the gart size */
-	if (amdgpu_gart_size == -1)
-		adev->gmc.gart_size = 512ULL << 20;
-	else
+	if (amdgpu_gart_size == -1) {
+		switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+		default:
+			adev->gmc.gart_size = 512ULL << 20;
+			break;
+		case IP_VERSION(10, 3, 1):   /* DCE SG support */
+		case IP_VERSION(10, 3, 3):   /* DCE SG support */
+		case IP_VERSION(10, 3, 6):   /* DCE SG support */
+		case IP_VERSION(10, 3, 7):   /* DCE SG support */
+			adev->gmc.gart_size = 1024ULL << 20;
+			break;
+		}
+	} else {
 		adev->gmc.gart_size = (u64)amdgpu_gart_size << 20;
+	}
 
 	gmc_v10_0_vram_gtt_location(adev, &adev->gmc);
 
@@ -825,16 +760,16 @@ static int gmc_v10_0_gart_init(struct amdgpu_device *adev)
 		return r;
 
 	adev->gart.table_size = adev->gart.num_gpu_pages * 8;
-	adev->gart.gart_pte_flags = AMDGPU_PTE_MTYPE_NV10(MTYPE_UC) |
+	adev->gart.gart_pte_flags = AMDGPU_PTE_MTYPE_NV10(0ULL, MTYPE_UC) |
 				 AMDGPU_PTE_EXECUTABLE;
 
 	return amdgpu_gart_table_vram_alloc(adev);
 }
 
-static int gmc_v10_0_sw_init(void *handle)
+static int gmc_v10_0_sw_init(struct amdgpu_ip_block *ip_block)
 {
 	int r, vram_width = 0, vram_type = 0, vram_vendor = 0;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	adev->gfxhub.funcs->init(adev);
 
@@ -857,18 +792,40 @@ static int gmc_v10_0_sw_init(void *handle)
 		adev->gmc.vram_vendor = vram_vendor;
 	}
 
-	switch (adev->ip_versions[GC_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	case IP_VERSION(10, 3, 0):
+		adev->gmc.mall_size = 128 * 1024 * 1024;
+		break;
+	case IP_VERSION(10, 3, 2):
+		adev->gmc.mall_size = 96 * 1024 * 1024;
+		break;
+	case IP_VERSION(10, 3, 4):
+		adev->gmc.mall_size = 32 * 1024 * 1024;
+		break;
+	case IP_VERSION(10, 3, 5):
+		adev->gmc.mall_size = 16 * 1024 * 1024;
+		break;
+	default:
+		adev->gmc.mall_size = 0;
+		break;
+	}
+
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 	case IP_VERSION(10, 1, 10):
 	case IP_VERSION(10, 1, 1):
 	case IP_VERSION(10, 1, 2):
 	case IP_VERSION(10, 1, 3):
+	case IP_VERSION(10, 1, 4):
 	case IP_VERSION(10, 3, 0):
 	case IP_VERSION(10, 3, 2):
 	case IP_VERSION(10, 3, 1):
 	case IP_VERSION(10, 3, 4):
 	case IP_VERSION(10, 3, 5):
+	case IP_VERSION(10, 3, 6):
 	case IP_VERSION(10, 3, 3):
-		adev->num_vmhubs = 2;
+	case IP_VERSION(10, 3, 7):
+		set_bit(AMDGPU_GFXHUB(0), adev->vmhubs_mask);
+		set_bit(AMDGPU_MMHUB0(0), adev->vmhubs_mask);
 		/*
 		 * To fulfill 4-level page support,
 		 * vm size is 256TB (48bit), maximum size of Navi10/Navi14/Navi12,
@@ -910,22 +867,17 @@ static int gmc_v10_0_sw_init(void *handle)
 
 	r = dma_set_mask_and_coherent(adev->dev, DMA_BIT_MASK(44));
 	if (r) {
-		printk(KERN_WARNING "amdgpu: No suitable DMA available.\n");
+		dev_warn(adev->dev, "amdgpu: No suitable DMA available.\n");
 		return r;
 	}
 
-	if (adev->gmc.xgmi.supported) {
-		r = adev->gfxhub.funcs->get_xgmi_info(adev);
-		if (r)
-			return r;
-	}
+	adev->need_swiotlb = drm_need_swiotlb(44);
 
 	r = gmc_v10_0_mc_init(adev);
 	if (r)
 		return r;
 
 	amdgpu_gmc_get_vbios_allocations(adev);
-	amdgpu_gmc_get_reserved_allocation(adev);
 
 	/* Memory manager */
 	r = amdgpu_bo_init(adev);
@@ -946,6 +898,10 @@ static int gmc_v10_0_sw_init(void *handle)
 
 	amdgpu_vm_manager_init(adev);
 
+	r = amdgpu_gmc_ras_sw_init(adev);
+	if (r)
+		return r;
+
 	return 0;
 }
 
@@ -961,9 +917,9 @@ static void gmc_v10_0_gart_fini(struct amdgpu_device *adev)
 	amdgpu_gart_table_vram_free(adev);
 }
 
-static int gmc_v10_0_sw_fini(void *handle)
+static int gmc_v10_0_sw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	amdgpu_vm_manager_fini(adev);
 	gmc_v10_0_gart_fini(adev);
@@ -992,13 +948,13 @@ static int gmc_v10_0_gart_enable(struct amdgpu_device *adev)
 		return -EINVAL;
 	}
 
-	r = amdgpu_gart_table_vram_pin(adev);
-	if (r)
-		return r;
+	amdgpu_gtt_mgr_recover(&adev->mman.gtt_mgr);
 
-	r = adev->gfxhub.funcs->gart_enable(adev);
-	if (r)
-		return r;
+	if (!adev->in_s0ix) {
+		r = adev->gfxhub.funcs->gart_enable(adev);
+		if (r)
+			return r;
+	}
 
 	r = adev->mmhub.funcs->gart_enable(adev);
 	if (r)
@@ -1007,29 +963,30 @@ static int gmc_v10_0_gart_enable(struct amdgpu_device *adev)
 	adev->hdp.funcs->init_registers(adev);
 
 	/* Flush HDP after it is initialized */
-	adev->hdp.funcs->flush_hdp(adev, NULL);
+	amdgpu_device_flush_hdp(adev, NULL);
 
-	value = (amdgpu_vm_fault_stop == AMDGPU_VM_FAULT_STOP_ALWAYS) ?
-		false : true;
+	value = amdgpu_vm_fault_stop != AMDGPU_VM_FAULT_STOP_ALWAYS;
 
-	adev->gfxhub.funcs->set_fault_enable_default(adev, value);
+	if (!adev->in_s0ix)
+		adev->gfxhub.funcs->set_fault_enable_default(adev, value);
 	adev->mmhub.funcs->set_fault_enable_default(adev, value);
-	gmc_v10_0_flush_gpu_tlb(adev, 0, AMDGPU_MMHUB_0, 0);
-	gmc_v10_0_flush_gpu_tlb(adev, 0, AMDGPU_GFXHUB_0, 0);
+	gmc_v10_0_flush_gpu_tlb(adev, 0, AMDGPU_MMHUB0(0), 0);
+	if (!adev->in_s0ix)
+		gmc_v10_0_flush_gpu_tlb(adev, 0, AMDGPU_GFXHUB(0), 0);
 
 	DRM_INFO("PCIE GART of %uM enabled (table at 0x%016llX).\n",
-		 (unsigned)(adev->gmc.gart_size >> 20),
+		 (unsigned int)(adev->gmc.gart_size >> 20),
 		 (unsigned long long)amdgpu_bo_gpu_offset(adev->gart.bo));
 
-	adev->gart.ready = true;
-
 	return 0;
 }
 
-static int gmc_v10_0_hw_init(void *handle)
+static int gmc_v10_0_hw_init(struct amdgpu_ip_block *ip_block)
 {
+	struct amdgpu_device *adev = ip_block->adev;
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+	adev->gmc.flush_pasid_uses_kiq = !amdgpu_emu_mode;
 
 	/* The sequence of these two function calls matters.*/
 	gmc_v10_0_init_golden_registers(adev);
@@ -1038,13 +995,19 @@ static int gmc_v10_0_hw_init(void *handle)
 	 * harvestable groups in gc_utcl2 need to be programmed before any GFX block
 	 * register setup within GMC, or else system hang when harvesting SA.
 	 */
-	if (adev->gfxhub.funcs && adev->gfxhub.funcs->utcl2_harvest)
+	if (!adev->in_s0ix && adev->gfxhub.funcs && adev->gfxhub.funcs->utcl2_harvest)
 		adev->gfxhub.funcs->utcl2_harvest(adev);
 
 	r = gmc_v10_0_gart_enable(adev);
 	if (r)
 		return r;
 
+	if (amdgpu_emu_mode == 1) {
+		r = amdgpu_gmc_vram_checking(adev);
+		if (r)
+			return r;
+	}
+
 	if (adev->umc.funcs && adev->umc.funcs->init_registers)
 		adev->umc.funcs->init_registers(adev);
 
@@ -1060,14 +1023,14 @@ static int gmc_v10_0_hw_init(void *handle)
  */
 static void gmc_v10_0_gart_disable(struct amdgpu_device *adev)
 {
-	adev->gfxhub.funcs->gart_disable(adev);
+	if (!adev->in_s0ix)
+		adev->gfxhub.funcs->gart_disable(adev);
 	adev->mmhub.funcs->gart_disable(adev);
-	amdgpu_gart_table_vram_unpin(adev);
 }
 
-static int gmc_v10_0_hw_fini(void *handle)
+static int gmc_v10_0_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	gmc_v10_0_gart_disable(adev);
 
@@ -1077,81 +1040,91 @@ static int gmc_v10_0_hw_fini(void *handle)
 		return 0;
 	}
 
-	amdgpu_irq_put(adev, &adev->gmc.ecc_irq, 0);
 	amdgpu_irq_put(adev, &adev->gmc.vm_fault, 0);
 
+	if (adev->gmc.ecc_irq.funcs &&
+		amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__UMC))
+		amdgpu_irq_put(adev, &adev->gmc.ecc_irq, 0);
+
 	return 0;
 }
 
-static int gmc_v10_0_suspend(void *handle)
+static int gmc_v10_0_suspend(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	gmc_v10_0_hw_fini(adev);
+	gmc_v10_0_hw_fini(ip_block);
 
 	return 0;
 }
 
-static int gmc_v10_0_resume(void *handle)
+static int gmc_v10_0_resume(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
-	r = gmc_v10_0_hw_init(adev);
+	r = gmc_v10_0_hw_init(ip_block);
 	if (r)
 		return r;
 
-	amdgpu_vmid_reset_all(adev);
+	amdgpu_vmid_reset_all(ip_block->adev);
 
 	return 0;
 }
 
-static bool gmc_v10_0_is_idle(void *handle)
+static bool gmc_v10_0_is_idle(struct amdgpu_ip_block *ip_block)
 {
 	/* MC is always ready in GMC v10.*/
 	return true;
 }
 
-static int gmc_v10_0_wait_for_idle(void *handle)
+static int gmc_v10_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
 {
 	/* There is no need to wait for MC idle in GMC v10.*/
 	return 0;
 }
 
-static int gmc_v10_0_soft_reset(void *handle)
-{
-	return 0;
-}
-
-static int gmc_v10_0_set_clockgating_state(void *handle,
+static int gmc_v10_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					   enum amd_clockgating_state state)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
+
+	/*
+	 * The issue mmhub can't disconnect from DF with MMHUB clock gating being disabled
+	 * is a new problem observed at DF 3.0.3, however with the same suspend sequence not
+	 * seen any issue on the DF 3.0.2 series platform.
+	 */
+	if (adev->in_s0ix &&
+	    amdgpu_ip_version(adev, DF_HWIP, 0) > IP_VERSION(3, 0, 2)) {
+		dev_dbg(adev->dev, "keep mmhub clock gating being enabled for s0ix\n");
+		return 0;
+	}
 
 	r = adev->mmhub.funcs->set_clockgating(adev, state);
 	if (r)
 		return r;
 
-	if (adev->ip_versions[ATHUB_HWIP][0] >= IP_VERSION(2, 1, 0))
+	if (amdgpu_ip_version(adev, ATHUB_HWIP, 0) >= IP_VERSION(2, 1, 0))
 		return athub_v2_1_set_clockgating(adev, state);
 	else
 		return athub_v2_0_set_clockgating(adev, state);
 }
 
-static void gmc_v10_0_get_clockgating_state(void *handle, u32 *flags)
+static void gmc_v10_0_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
+
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(10, 1, 3) ||
+	    amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(10, 1, 4))
+		return;
 
 	adev->mmhub.funcs->get_clockgating(adev, flags);
 
-	if (adev->ip_versions[ATHUB_HWIP][0] >= IP_VERSION(2, 1, 0))
+	if (amdgpu_ip_version(adev, ATHUB_HWIP, 0) >= IP_VERSION(2, 1, 0))
 		athub_v2_1_get_clockgating(adev, flags);
 	else
 		athub_v2_0_get_clockgating(adev, flags);
 }
 
-static int gmc_v10_0_set_powergating_state(void *handle,
+static int gmc_v10_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					   enum amd_powergating_state state)
 {
 	return 0;
@@ -1169,14 +1142,12 @@ const struct amd_ip_funcs gmc_v10_0_ip_funcs = {
 	.resume = gmc_v10_0_resume,
 	.is_idle = gmc_v10_0_is_idle,
 	.wait_for_idle = gmc_v10_0_wait_for_idle,
-	.soft_reset = gmc_v10_0_soft_reset,
 	.set_clockgating_state = gmc_v10_0_set_clockgating_state,
 	.set_powergating_state = gmc_v10_0_set_powergating_state,
 	.get_clockgating_state = gmc_v10_0_get_clockgating_state,
 };
 
-const struct amdgpu_ip_block_version gmc_v10_0_ip_block =
-{
+const struct amdgpu_ip_block_version gmc_v10_0_ip_block = {
 	.type = AMD_IP_BLOCK_TYPE_GMC,
 	.major = 10,
 	.minor = 0,
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c
new file mode 100644
index 000000000000..ba59ee8e398a
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c
@@ -0,0 +1,1085 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include <linux/firmware.h>
+#include <linux/pci.h>
+
+#include <drm/drm_cache.h>
+
+#include "amdgpu.h"
+#include "amdgpu_atomfirmware.h"
+#include "gmc_v11_0.h"
+#include "umc_v8_10.h"
+#include "athub/athub_3_0_0_sh_mask.h"
+#include "athub/athub_3_0_0_offset.h"
+#include "dcn/dcn_3_2_0_offset.h"
+#include "dcn/dcn_3_2_0_sh_mask.h"
+#include "oss/osssys_6_0_0_offset.h"
+#include "ivsrcid/vmc/irqsrcs_vmc_1_0.h"
+#include "navi10_enum.h"
+#include "soc15.h"
+#include "soc15d.h"
+#include "soc15_common.h"
+#include "nbio_v4_3.h"
+#include "gfxhub_v3_0.h"
+#include "gfxhub_v3_0_3.h"
+#include "gfxhub_v11_5_0.h"
+#include "mmhub_v3_0.h"
+#include "mmhub_v3_0_1.h"
+#include "mmhub_v3_0_2.h"
+#include "mmhub_v3_3.h"
+#include "athub_v3_0.h"
+
+
+static int gmc_v11_0_ecc_interrupt_state(struct amdgpu_device *adev,
+					 struct amdgpu_irq_src *src,
+					 unsigned int type,
+					 enum amdgpu_interrupt_state state)
+{
+	return 0;
+}
+
+static int
+gmc_v11_0_vm_fault_interrupt_state(struct amdgpu_device *adev,
+				   struct amdgpu_irq_src *src, unsigned int type,
+				   enum amdgpu_interrupt_state state)
+{
+	switch (state) {
+	case AMDGPU_IRQ_STATE_DISABLE:
+		/* MM HUB */
+		amdgpu_gmc_set_vm_fault_masks(adev, AMDGPU_MMHUB0(0), false);
+		/* GFX HUB */
+		/* This works because this interrupt is only
+		 * enabled at init/resume and disabled in
+		 * fini/suspend, so the overall state doesn't
+		 * change over the course of suspend/resume.
+		 */
+		if (!adev->in_s0ix && (adev->in_runpm || adev->in_suspend ||
+							   amdgpu_in_reset(adev)))
+			amdgpu_gmc_set_vm_fault_masks(adev, AMDGPU_GFXHUB(0), false);
+		break;
+	case AMDGPU_IRQ_STATE_ENABLE:
+		/* MM HUB */
+		amdgpu_gmc_set_vm_fault_masks(adev, AMDGPU_MMHUB0(0), true);
+		/* GFX HUB */
+		/* This works because this interrupt is only
+		 * enabled at init/resume and disabled in
+		 * fini/suspend, so the overall state doesn't
+		 * change over the course of suspend/resume.
+		 */
+		if (!adev->in_s0ix)
+			amdgpu_gmc_set_vm_fault_masks(adev, AMDGPU_GFXHUB(0), true);
+		break;
+	default:
+		break;
+	}
+
+	return 0;
+}
+
+static int gmc_v11_0_process_interrupt(struct amdgpu_device *adev,
+				       struct amdgpu_irq_src *source,
+				       struct amdgpu_iv_entry *entry)
+{
+	uint32_t vmhub_index = entry->client_id == SOC21_IH_CLIENTID_VMC ?
+			       AMDGPU_MMHUB0(0) : AMDGPU_GFXHUB(0);
+	struct amdgpu_vmhub *hub = &adev->vmhub[vmhub_index];
+	bool retry_fault = !!(entry->src_data[1] &
+			      AMDGPU_GMC9_FAULT_SOURCE_DATA_RETRY);
+	bool write_fault = !!(entry->src_data[1] &
+			      AMDGPU_GMC9_FAULT_SOURCE_DATA_WRITE);
+	uint32_t status = 0;
+	u64 addr;
+
+	addr = (u64)entry->src_data[0] << 12;
+	addr |= ((u64)entry->src_data[1] & 0xf) << 44;
+
+	if (retry_fault) {
+		/* Returning 1 here also prevents sending the IV to the KFD */
+
+		/* Process it only if it's the first fault for this address */
+		if (entry->ih != &adev->irq.ih_soft &&
+		    amdgpu_gmc_filter_faults(adev, entry->ih, addr, entry->pasid,
+					     entry->timestamp))
+			return 1;
+
+		/* Delegate it to a different ring if the hardware hasn't
+		 * already done it.
+		 */
+		if (entry->ih == &adev->irq.ih) {
+			amdgpu_irq_delegate(adev, entry, 8);
+			return 1;
+		}
+
+		/* Try to handle the recoverable page faults by filling page
+		 * tables
+		 */
+		if (amdgpu_vm_handle_fault(adev, entry->pasid, 0, 0, addr,
+					   entry->timestamp, write_fault))
+			return 1;
+	}
+
+	if (!amdgpu_sriov_vf(adev)) {
+		/*
+		 * Issue a dummy read to wait for the status register to
+		 * be updated to avoid reading an incorrect value due to
+		 * the new fast GRBM interface.
+		 */
+		if (entry->vmid_src == AMDGPU_GFXHUB(0))
+			RREG32(hub->vm_l2_pro_fault_status);
+
+		status = RREG32(hub->vm_l2_pro_fault_status);
+		WREG32_P(hub->vm_l2_pro_fault_cntl, 1, ~1);
+
+		amdgpu_vm_update_fault_cache(adev, entry->pasid, addr, status,
+					     entry->vmid_src ? AMDGPU_MMHUB0(0) : AMDGPU_GFXHUB(0));
+	}
+
+	if (printk_ratelimit()) {
+		struct amdgpu_task_info *task_info;
+
+		dev_err(adev->dev,
+			"[%s] page fault (src_id:%u ring:%u vmid:%u pasid:%u)\n",
+			entry->vmid_src ? "mmhub" : "gfxhub",
+			entry->src_id, entry->ring_id, entry->vmid, entry->pasid);
+		task_info = amdgpu_vm_get_task_info_pasid(adev, entry->pasid);
+		if (task_info) {
+			amdgpu_vm_print_task_info(adev, task_info);
+			amdgpu_vm_put_task_info(task_info);
+		}
+
+		dev_err(adev->dev, "  in page starting at address 0x%016llx from client %d\n",
+				addr, entry->client_id);
+
+		/* Only print L2 fault status if the status register could be read and
+		 * contains useful information
+		 */
+		if (status != 0)
+			hub->vmhub_funcs->print_l2_protection_fault_status(adev, status);
+	}
+
+	return 0;
+}
+
+static const struct amdgpu_irq_src_funcs gmc_v11_0_irq_funcs = {
+	.set = gmc_v11_0_vm_fault_interrupt_state,
+	.process = gmc_v11_0_process_interrupt,
+};
+
+static const struct amdgpu_irq_src_funcs gmc_v11_0_ecc_funcs = {
+	.set = gmc_v11_0_ecc_interrupt_state,
+	.process = amdgpu_umc_process_ecc_irq,
+};
+
+static void gmc_v11_0_set_irq_funcs(struct amdgpu_device *adev)
+{
+	adev->gmc.vm_fault.num_types = 1;
+	adev->gmc.vm_fault.funcs = &gmc_v11_0_irq_funcs;
+
+	if (!amdgpu_sriov_vf(adev)) {
+		adev->gmc.ecc_irq.num_types = 1;
+		adev->gmc.ecc_irq.funcs = &gmc_v11_0_ecc_funcs;
+	}
+}
+
+/**
+ * gmc_v11_0_use_invalidate_semaphore - judge whether to use semaphore
+ *
+ * @adev: amdgpu_device pointer
+ * @vmhub: vmhub type
+ *
+ */
+static bool gmc_v11_0_use_invalidate_semaphore(struct amdgpu_device *adev,
+				       uint32_t vmhub)
+{
+	return ((vmhub == AMDGPU_MMHUB0(0)) &&
+		(!amdgpu_sriov_vf(adev)));
+}
+
+static bool gmc_v11_0_get_vmid_pasid_mapping_info(
+					struct amdgpu_device *adev,
+					uint8_t vmid, uint16_t *p_pasid)
+{
+	*p_pasid = RREG32(SOC15_REG_OFFSET(OSSSYS, 0, regIH_VMID_0_LUT) + vmid) & 0xffff;
+
+	return !!(*p_pasid);
+}
+
+/**
+ * gmc_v11_0_flush_gpu_tlb - gart tlb flush callback
+ *
+ * @adev: amdgpu_device pointer
+ * @vmid: vm instance to flush
+ * @vmhub: which hub to flush
+ * @flush_type: the flush type
+ *
+ * Flush the TLB for the requested page table.
+ */
+static void gmc_v11_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid,
+					uint32_t vmhub, uint32_t flush_type)
+{
+	bool use_semaphore = gmc_v11_0_use_invalidate_semaphore(adev, vmhub);
+	struct amdgpu_vmhub *hub = &adev->vmhub[vmhub];
+	u32 inv_req = hub->vmhub_funcs->get_invalidate_req(vmid, flush_type);
+	/* Use register 17 for GART */
+	const unsigned int eng = 17;
+	unsigned char hub_ip;
+	u32 sem, req, ack;
+	unsigned int i;
+	u32 tmp;
+
+	if ((vmhub == AMDGPU_GFXHUB(0)) && !adev->gfx.is_poweron)
+		return;
+
+	sem = hub->vm_inv_eng0_sem + hub->eng_distance * eng;
+	req = hub->vm_inv_eng0_req + hub->eng_distance * eng;
+	ack = hub->vm_inv_eng0_ack + hub->eng_distance * eng;
+
+	/* flush hdp cache */
+	amdgpu_device_flush_hdp(adev, NULL);
+
+	/* This is necessary for SRIOV as well as for GFXOFF to function
+	 * properly under bare metal
+	 */
+	if ((adev->gfx.kiq[0].ring.sched.ready || adev->mes.ring[0].sched.ready) &&
+	    (amdgpu_sriov_runtime(adev) || !amdgpu_sriov_vf(adev))) {
+		amdgpu_gmc_fw_reg_write_reg_wait(adev, req, ack, inv_req,
+						 1 << vmid, GET_INST(GC, 0));
+		return;
+	}
+
+	/* This path is needed before KIQ/MES/GFXOFF are set up */
+	hub_ip = (vmhub == AMDGPU_GFXHUB(0)) ? GC_HWIP : MMHUB_HWIP;
+
+	spin_lock(&adev->gmc.invalidate_lock);
+	/*
+	 * It may lose gpuvm invalidate acknowldege state across power-gating
+	 * off cycle, add semaphore acquire before invalidation and semaphore
+	 * release after invalidation to avoid entering power gated state
+	 * to WA the Issue
+	 */
+
+	/* TODO: It needs to continue working on debugging with semaphore for GFXHUB as well. */
+	if (use_semaphore) {
+		for (i = 0; i < adev->usec_timeout; i++) {
+			/* a read return value of 1 means semaphore acuqire */
+			tmp = RREG32_RLC_NO_KIQ(sem, hub_ip);
+			if (tmp & 0x1)
+				break;
+			udelay(1);
+		}
+
+		if (i >= adev->usec_timeout)
+			DRM_ERROR("Timeout waiting for sem acquire in VM flush!\n");
+	}
+
+	WREG32_RLC_NO_KIQ(req, inv_req, hub_ip);
+
+	/* Wait for ACK with a delay.*/
+	for (i = 0; i < adev->usec_timeout; i++) {
+		tmp = RREG32_RLC_NO_KIQ(ack, hub_ip);
+		tmp &= 1 << vmid;
+		if (tmp)
+			break;
+
+		udelay(1);
+	}
+
+	/* TODO: It needs to continue working on debugging with semaphore for GFXHUB as well. */
+	if (use_semaphore)
+		WREG32_RLC_NO_KIQ(sem, 0, hub_ip);
+
+	/* Issue additional private vm invalidation to MMHUB */
+	if ((vmhub != AMDGPU_GFXHUB(0)) &&
+	    (hub->vm_l2_bank_select_reserved_cid2) &&
+		!amdgpu_sriov_vf(adev)) {
+		inv_req = RREG32_NO_KIQ(hub->vm_l2_bank_select_reserved_cid2);
+		/* bit 25: RSERVED_CACHE_PRIVATE_INVALIDATION */
+		inv_req |= (1 << 25);
+		/* Issue private invalidation */
+		WREG32_NO_KIQ(hub->vm_l2_bank_select_reserved_cid2, inv_req);
+		/* Read back to ensure invalidation is done*/
+		RREG32_NO_KIQ(hub->vm_l2_bank_select_reserved_cid2);
+	}
+
+	spin_unlock(&adev->gmc.invalidate_lock);
+
+	if (i >= adev->usec_timeout)
+		dev_err(adev->dev, "Timeout waiting for VM flush ACK!\n");
+}
+
+/**
+ * gmc_v11_0_flush_gpu_tlb_pasid - tlb flush via pasid
+ *
+ * @adev: amdgpu_device pointer
+ * @pasid: pasid to be flush
+ * @flush_type: the flush type
+ * @all_hub: flush all hubs
+ * @inst: is used to select which instance of KIQ to use for the invalidation
+ *
+ * Flush the TLB for the requested pasid.
+ */
+static void gmc_v11_0_flush_gpu_tlb_pasid(struct amdgpu_device *adev,
+					  uint16_t pasid, uint32_t flush_type,
+					  bool all_hub, uint32_t inst)
+{
+	uint16_t queried;
+	int vmid, i;
+
+	for (vmid = 1; vmid < 16; vmid++) {
+		bool valid;
+
+		valid = gmc_v11_0_get_vmid_pasid_mapping_info(adev, vmid,
+							      &queried);
+		if (!valid || queried != pasid)
+			continue;
+
+		if (all_hub) {
+			for_each_set_bit(i, adev->vmhubs_mask,
+					 AMDGPU_MAX_VMHUBS)
+				gmc_v11_0_flush_gpu_tlb(adev, vmid, i,
+							flush_type);
+		} else {
+			gmc_v11_0_flush_gpu_tlb(adev, vmid, AMDGPU_GFXHUB(0),
+						flush_type);
+		}
+	}
+}
+
+static uint64_t gmc_v11_0_emit_flush_gpu_tlb(struct amdgpu_ring *ring,
+					     unsigned int vmid, uint64_t pd_addr)
+{
+	bool use_semaphore = gmc_v11_0_use_invalidate_semaphore(ring->adev, ring->vm_hub);
+	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->vm_hub];
+	uint32_t req = hub->vmhub_funcs->get_invalidate_req(vmid, 0);
+	unsigned int eng = ring->vm_inv_eng;
+
+	/*
+	 * It may lose gpuvm invalidate acknowldege state across power-gating
+	 * off cycle, add semaphore acquire before invalidation and semaphore
+	 * release after invalidation to avoid entering power gated state
+	 * to WA the Issue
+	 */
+
+	/* TODO: It needs to continue working on debugging with semaphore for GFXHUB as well. */
+	if (use_semaphore)
+		/* a read return value of 1 means semaphore acuqire */
+		amdgpu_ring_emit_reg_wait(ring,
+					  hub->vm_inv_eng0_sem +
+					  hub->eng_distance * eng, 0x1, 0x1);
+
+	amdgpu_ring_emit_wreg(ring, hub->ctx0_ptb_addr_lo32 +
+			      (hub->ctx_addr_distance * vmid),
+			      lower_32_bits(pd_addr));
+
+	amdgpu_ring_emit_wreg(ring, hub->ctx0_ptb_addr_hi32 +
+			      (hub->ctx_addr_distance * vmid),
+			      upper_32_bits(pd_addr));
+
+	amdgpu_ring_emit_reg_write_reg_wait(ring, hub->vm_inv_eng0_req +
+					    hub->eng_distance * eng,
+					    hub->vm_inv_eng0_ack +
+					    hub->eng_distance * eng,
+					    req, 1 << vmid);
+
+	/* TODO: It needs to continue working on debugging with semaphore for GFXHUB as well. */
+	if (use_semaphore)
+		/*
+		 * add semaphore release after invalidation,
+		 * write with 0 means semaphore release
+		 */
+		amdgpu_ring_emit_wreg(ring, hub->vm_inv_eng0_sem +
+				      hub->eng_distance * eng, 0);
+
+	return pd_addr;
+}
+
+static void gmc_v11_0_emit_pasid_mapping(struct amdgpu_ring *ring, unsigned int vmid,
+					 unsigned int pasid)
+{
+	struct amdgpu_device *adev = ring->adev;
+	uint32_t reg;
+
+	if (ring->vm_hub == AMDGPU_GFXHUB(0))
+		reg = SOC15_REG_OFFSET(OSSSYS, 0, regIH_VMID_0_LUT) + vmid;
+	else
+		reg = SOC15_REG_OFFSET(OSSSYS, 0, regIH_VMID_0_LUT_MM) + vmid;
+
+	amdgpu_ring_emit_wreg(ring, reg, pasid);
+}
+
+/*
+ * PTE format:
+ * 63:59 reserved
+ * 58:57 reserved
+ * 56 F
+ * 55 L
+ * 54 reserved
+ * 53:52 SW
+ * 51 T
+ * 50:48 mtype
+ * 47:12 4k physical page base address
+ * 11:7 fragment
+ * 6 write
+ * 5 read
+ * 4 exe
+ * 3 Z
+ * 2 snooped
+ * 1 system
+ * 0 valid
+ *
+ * PDE format:
+ * 63:59 block fragment size
+ * 58:55 reserved
+ * 54 P
+ * 53:48 reserved
+ * 47:6 physical base address of PD or PTE
+ * 5:3 reserved
+ * 2 C
+ * 1 system
+ * 0 valid
+ */
+
+static void gmc_v11_0_get_vm_pde(struct amdgpu_device *adev, int level,
+				 uint64_t *addr, uint64_t *flags)
+{
+	if (!(*flags & AMDGPU_PDE_PTE) && !(*flags & AMDGPU_PTE_SYSTEM))
+		*addr = amdgpu_gmc_vram_mc2pa(adev, *addr);
+	BUG_ON(*addr & 0xFFFF00000000003FULL);
+
+	if (!adev->gmc.translate_further)
+		return;
+
+	if (level == AMDGPU_VM_PDB1) {
+		/* Set the block fragment size */
+		if (!(*flags & AMDGPU_PDE_PTE))
+			*flags |= AMDGPU_PDE_BFS(0x9);
+
+	} else if (level == AMDGPU_VM_PDB0) {
+		if (*flags & AMDGPU_PDE_PTE)
+			*flags &= ~AMDGPU_PDE_PTE;
+		else
+			*flags |= AMDGPU_PTE_TF;
+	}
+}
+
+static void gmc_v11_0_get_vm_pte(struct amdgpu_device *adev,
+				 struct amdgpu_vm *vm,
+				 struct amdgpu_bo *bo,
+				 uint32_t vm_flags,
+				 uint64_t *flags)
+{
+	if (vm_flags & AMDGPU_VM_PAGE_EXECUTABLE)
+		*flags |= AMDGPU_PTE_EXECUTABLE;
+	else
+		*flags &= ~AMDGPU_PTE_EXECUTABLE;
+
+	switch (vm_flags & AMDGPU_VM_MTYPE_MASK) {
+	case AMDGPU_VM_MTYPE_DEFAULT:
+	case AMDGPU_VM_MTYPE_NC:
+	default:
+		*flags = AMDGPU_PTE_MTYPE_NV10(*flags, MTYPE_NC);
+		break;
+	case AMDGPU_VM_MTYPE_WC:
+		*flags = AMDGPU_PTE_MTYPE_NV10(*flags, MTYPE_WC);
+		break;
+	case AMDGPU_VM_MTYPE_CC:
+		*flags = AMDGPU_PTE_MTYPE_NV10(*flags, MTYPE_CC);
+		break;
+	case AMDGPU_VM_MTYPE_UC:
+		*flags = AMDGPU_PTE_MTYPE_NV10(*flags, MTYPE_UC);
+		break;
+	}
+
+	if (vm_flags & AMDGPU_VM_PAGE_NOALLOC)
+		*flags |= AMDGPU_PTE_NOALLOC;
+	else
+		*flags &= ~AMDGPU_PTE_NOALLOC;
+
+	if (vm_flags & AMDGPU_VM_PAGE_PRT) {
+		*flags |= AMDGPU_PTE_PRT;
+		*flags |= AMDGPU_PTE_SNOOPED;
+		*flags |= AMDGPU_PTE_LOG;
+		*flags |= AMDGPU_PTE_SYSTEM;
+		*flags &= ~AMDGPU_PTE_VALID;
+	}
+
+	if (bo && bo->flags & (AMDGPU_GEM_CREATE_COHERENT |
+			       AMDGPU_GEM_CREATE_EXT_COHERENT |
+			       AMDGPU_GEM_CREATE_UNCACHED))
+		*flags = AMDGPU_PTE_MTYPE_NV10(*flags, MTYPE_UC);
+}
+
+static unsigned int gmc_v11_0_get_vbios_fb_size(struct amdgpu_device *adev)
+{
+	u32 d1vga_control = RREG32_SOC15(DCE, 0, regD1VGA_CONTROL);
+	unsigned int size;
+
+	if (REG_GET_FIELD(d1vga_control, D1VGA_CONTROL, D1VGA_MODE_ENABLE)) {
+		size = AMDGPU_VBIOS_VGA_ALLOCATION;
+	} else {
+		u32 viewport;
+		u32 pitch;
+
+		viewport = RREG32_SOC15(DCE, 0, regHUBP0_DCSURF_PRI_VIEWPORT_DIMENSION);
+		pitch = RREG32_SOC15(DCE, 0, regHUBPREQ0_DCSURF_SURFACE_PITCH);
+		size = (REG_GET_FIELD(viewport,
+					HUBP0_DCSURF_PRI_VIEWPORT_DIMENSION, PRI_VIEWPORT_HEIGHT) *
+				REG_GET_FIELD(pitch, HUBPREQ0_DCSURF_SURFACE_PITCH, PITCH) *
+				4);
+	}
+
+	return size;
+}
+
+static const struct amdgpu_gmc_funcs gmc_v11_0_gmc_funcs = {
+	.flush_gpu_tlb = gmc_v11_0_flush_gpu_tlb,
+	.flush_gpu_tlb_pasid = gmc_v11_0_flush_gpu_tlb_pasid,
+	.emit_flush_gpu_tlb = gmc_v11_0_emit_flush_gpu_tlb,
+	.emit_pasid_mapping = gmc_v11_0_emit_pasid_mapping,
+	.get_vm_pde = gmc_v11_0_get_vm_pde,
+	.get_vm_pte = gmc_v11_0_get_vm_pte,
+	.get_vbios_fb_size = gmc_v11_0_get_vbios_fb_size,
+};
+
+static void gmc_v11_0_set_gmc_funcs(struct amdgpu_device *adev)
+{
+	adev->gmc.gmc_funcs = &gmc_v11_0_gmc_funcs;
+}
+
+static void gmc_v11_0_set_umc_funcs(struct amdgpu_device *adev)
+{
+	switch (amdgpu_ip_version(adev, UMC_HWIP, 0)) {
+	case IP_VERSION(8, 10, 0):
+		adev->umc.channel_inst_num = UMC_V8_10_CHANNEL_INSTANCE_NUM;
+		adev->umc.umc_inst_num = UMC_V8_10_UMC_INSTANCE_NUM;
+		adev->umc.max_ras_err_cnt_per_query = UMC_V8_10_TOTAL_CHANNEL_NUM(adev);
+		adev->umc.channel_offs = UMC_V8_10_PER_CHANNEL_OFFSET;
+		adev->umc.retire_unit = UMC_V8_10_NA_COL_2BITS_POWER_OF_2_NUM;
+		if (adev->umc.node_inst_num == 4)
+			adev->umc.channel_idx_tbl = &umc_v8_10_channel_idx_tbl_ext0[0][0][0];
+		else
+			adev->umc.channel_idx_tbl = &umc_v8_10_channel_idx_tbl[0][0][0];
+		adev->umc.ras = &umc_v8_10_ras;
+		break;
+	case IP_VERSION(8, 11, 0):
+		break;
+	default:
+		break;
+	}
+}
+
+
+static void gmc_v11_0_set_mmhub_funcs(struct amdgpu_device *adev)
+{
+	switch (amdgpu_ip_version(adev, MMHUB_HWIP, 0)) {
+	case IP_VERSION(3, 0, 1):
+		adev->mmhub.funcs = &mmhub_v3_0_1_funcs;
+		break;
+	case IP_VERSION(3, 0, 2):
+		adev->mmhub.funcs = &mmhub_v3_0_2_funcs;
+		break;
+	case IP_VERSION(3, 3, 0):
+	case IP_VERSION(3, 3, 1):
+	case IP_VERSION(3, 3, 2):
+		adev->mmhub.funcs = &mmhub_v3_3_funcs;
+		break;
+	default:
+		adev->mmhub.funcs = &mmhub_v3_0_funcs;
+		break;
+	}
+}
+
+static void gmc_v11_0_set_gfxhub_funcs(struct amdgpu_device *adev)
+{
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	case IP_VERSION(11, 0, 3):
+		adev->gfxhub.funcs = &gfxhub_v3_0_3_funcs;
+		break;
+	case IP_VERSION(11, 5, 0):
+	case IP_VERSION(11, 5, 1):
+	case IP_VERSION(11, 5, 2):
+	case IP_VERSION(11, 5, 3):
+		adev->gfxhub.funcs = &gfxhub_v11_5_0_funcs;
+		break;
+	default:
+		adev->gfxhub.funcs = &gfxhub_v3_0_funcs;
+		break;
+	}
+}
+
+static int gmc_v11_0_early_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	gmc_v11_0_set_gfxhub_funcs(adev);
+	gmc_v11_0_set_mmhub_funcs(adev);
+	gmc_v11_0_set_gmc_funcs(adev);
+	gmc_v11_0_set_irq_funcs(adev);
+	gmc_v11_0_set_umc_funcs(adev);
+
+	adev->gmc.shared_aperture_start = 0x2000000000000000ULL;
+	adev->gmc.shared_aperture_end =
+		adev->gmc.shared_aperture_start + (4ULL << 30) - 1;
+	adev->gmc.private_aperture_start = 0x1000000000000000ULL;
+	adev->gmc.private_aperture_end =
+		adev->gmc.private_aperture_start + (4ULL << 30) - 1;
+	adev->gmc.noretry_flags = AMDGPU_VM_NORETRY_FLAGS_TF;
+
+	return 0;
+}
+
+static int gmc_v11_0_late_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int r;
+
+	r = amdgpu_gmc_allocate_vm_inv_eng(adev);
+	if (r)
+		return r;
+
+	r = amdgpu_gmc_ras_late_init(adev);
+	if (r)
+		return r;
+
+	return amdgpu_irq_get(adev, &adev->gmc.vm_fault, 0);
+}
+
+static void gmc_v11_0_vram_gtt_location(struct amdgpu_device *adev,
+					struct amdgpu_gmc *mc)
+{
+	u64 base = 0;
+
+	base = adev->mmhub.funcs->get_fb_location(adev);
+
+	amdgpu_gmc_set_agp_default(adev, mc);
+	amdgpu_gmc_vram_location(adev, &adev->gmc, base);
+	amdgpu_gmc_gart_location(adev, mc, AMDGPU_GART_PLACEMENT_HIGH);
+	if (!amdgpu_sriov_vf(adev) &&
+	    (amdgpu_ip_version(adev, GC_HWIP, 0) < IP_VERSION(11, 5, 0)) &&
+	    (amdgpu_agp == 1))
+		amdgpu_gmc_agp_location(adev, mc);
+
+	/* base offset of vram pages */
+	if (amdgpu_sriov_vf(adev))
+		adev->vm_manager.vram_base_offset = 0;
+	else
+		adev->vm_manager.vram_base_offset = adev->mmhub.funcs->get_mc_fb_offset(adev);
+}
+
+/**
+ * gmc_v11_0_mc_init - initialize the memory controller driver params
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Look up the amount of vram, vram width, and decide how to place
+ * vram and gart within the GPU's physical address space.
+ * Returns 0 for success.
+ */
+static int gmc_v11_0_mc_init(struct amdgpu_device *adev)
+{
+	int r;
+
+	/* size in MB on si */
+	adev->gmc.mc_vram_size =
+		adev->nbio.funcs->get_memsize(adev) * 1024ULL * 1024ULL;
+	adev->gmc.real_vram_size = adev->gmc.mc_vram_size;
+
+	if (!(adev->flags & AMD_IS_APU)) {
+		r = amdgpu_device_resize_fb_bar(adev);
+		if (r)
+			return r;
+	}
+	adev->gmc.aper_base = pci_resource_start(adev->pdev, 0);
+	adev->gmc.aper_size = pci_resource_len(adev->pdev, 0);
+
+#ifdef CONFIG_X86_64
+	if ((adev->flags & AMD_IS_APU) && !amdgpu_passthrough(adev)) {
+		adev->gmc.aper_base = adev->mmhub.funcs->get_mc_fb_offset(adev);
+		adev->gmc.aper_size = adev->gmc.real_vram_size;
+	}
+#endif
+	/* In case the PCI BAR is larger than the actual amount of vram */
+	adev->gmc.visible_vram_size = adev->gmc.aper_size;
+	if (adev->gmc.visible_vram_size > adev->gmc.real_vram_size)
+		adev->gmc.visible_vram_size = adev->gmc.real_vram_size;
+
+	/* set the gart size */
+	if (amdgpu_gart_size == -1)
+		adev->gmc.gart_size = 512ULL << 20;
+	else
+		adev->gmc.gart_size = (u64)amdgpu_gart_size << 20;
+
+	gmc_v11_0_vram_gtt_location(adev, &adev->gmc);
+
+	return 0;
+}
+
+static int gmc_v11_0_gart_init(struct amdgpu_device *adev)
+{
+	int r;
+
+	if (adev->gart.bo) {
+		WARN(1, "PCIE GART already initialized\n");
+		return 0;
+	}
+
+	/* Initialize common gart structure */
+	r = amdgpu_gart_init(adev);
+	if (r)
+		return r;
+
+	adev->gart.table_size = adev->gart.num_gpu_pages * 8;
+	adev->gart.gart_pte_flags = AMDGPU_PTE_MTYPE_NV10(0ULL, MTYPE_UC) |
+				 AMDGPU_PTE_EXECUTABLE;
+
+	return amdgpu_gart_table_vram_alloc(adev);
+}
+
+static int gmc_v11_0_sw_init(struct amdgpu_ip_block *ip_block)
+{
+	int r, vram_width = 0, vram_type = 0, vram_vendor = 0;
+	struct amdgpu_device *adev = ip_block->adev;
+
+	adev->mmhub.funcs->init(adev);
+
+	adev->gfxhub.funcs->init(adev);
+
+	spin_lock_init(&adev->gmc.invalidate_lock);
+
+	r = amdgpu_atomfirmware_get_vram_info(adev,
+					      &vram_width, &vram_type, &vram_vendor);
+	adev->gmc.vram_width = vram_width;
+
+	adev->gmc.vram_type = vram_type;
+	adev->gmc.vram_vendor = vram_vendor;
+
+	/* The mall_size is already calculated as mall_size_per_umc * num_umc.
+	 * However, for gfx1151, which features a 2-to-1 UMC mapping,
+	 * the result must be multiplied by 2 to determine the actual mall size.
+	 */
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	case IP_VERSION(11, 5, 1):
+		adev->gmc.mall_size *= 2;
+		break;
+	default:
+		break;
+	}
+
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	case IP_VERSION(11, 0, 0):
+	case IP_VERSION(11, 0, 1):
+	case IP_VERSION(11, 0, 2):
+	case IP_VERSION(11, 0, 3):
+	case IP_VERSION(11, 0, 4):
+	case IP_VERSION(11, 5, 0):
+	case IP_VERSION(11, 5, 1):
+	case IP_VERSION(11, 5, 2):
+	case IP_VERSION(11, 5, 3):
+		set_bit(AMDGPU_GFXHUB(0), adev->vmhubs_mask);
+		set_bit(AMDGPU_MMHUB0(0), adev->vmhubs_mask);
+		/*
+		 * To fulfill 4-level page support,
+		 * vm size is 256TB (48bit), maximum size,
+		 * block size 512 (9bit)
+		 */
+		amdgpu_vm_adjust_size(adev, 256 * 1024, 9, 3, 48);
+		break;
+	default:
+		break;
+	}
+
+	/* This interrupt is VMC page fault.*/
+	r = amdgpu_irq_add_id(adev, SOC21_IH_CLIENTID_VMC,
+			      VMC_1_0__SRCID__VM_FAULT,
+			      &adev->gmc.vm_fault);
+
+	if (r)
+		return r;
+
+	r = amdgpu_irq_add_id(adev, SOC21_IH_CLIENTID_GFX,
+			      UTCL2_1_0__SRCID__FAULT,
+			      &adev->gmc.vm_fault);
+	if (r)
+		return r;
+
+	if (!amdgpu_sriov_vf(adev)) {
+		/* interrupt sent to DF. */
+		r = amdgpu_irq_add_id(adev, SOC21_IH_CLIENTID_DF, 0,
+				      &adev->gmc.ecc_irq);
+		if (r)
+			return r;
+	}
+
+	/*
+	 * Set the internal MC address mask This is the max address of the GPU's
+	 * internal address space.
+	 */
+	adev->gmc.mc_mask = 0xffffffffffffULL; /* 48 bit MC */
+
+	r = dma_set_mask_and_coherent(adev->dev, DMA_BIT_MASK(44));
+	if (r) {
+		dev_warn(adev->dev, "amdgpu: No suitable DMA available.\n");
+		return r;
+	}
+
+	adev->need_swiotlb = drm_need_swiotlb(44);
+
+	r = gmc_v11_0_mc_init(adev);
+	if (r)
+		return r;
+
+	amdgpu_gmc_get_vbios_allocations(adev);
+
+	/* Memory manager */
+	r = amdgpu_bo_init(adev);
+	if (r)
+		return r;
+
+	r = gmc_v11_0_gart_init(adev);
+	if (r)
+		return r;
+
+	/*
+	 * number of VMs
+	 * VMID 0 is reserved for System
+	 * amdgpu graphics/compute will use VMIDs 1-7
+	 * amdkfd will use VMIDs 8-15
+	 */
+	adev->vm_manager.first_kfd_vmid = adev->gfx.disable_kq ? 1 : 8;
+
+	amdgpu_vm_manager_init(adev);
+
+	r = amdgpu_gmc_ras_sw_init(adev);
+	if (r)
+		return r;
+
+	return 0;
+}
+
+/**
+ * gmc_v11_0_gart_fini - vm fini callback
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Tears down the driver GART/VM setup (CIK).
+ */
+static void gmc_v11_0_gart_fini(struct amdgpu_device *adev)
+{
+	amdgpu_gart_table_vram_free(adev);
+}
+
+static int gmc_v11_0_sw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	amdgpu_vm_manager_fini(adev);
+	gmc_v11_0_gart_fini(adev);
+	amdgpu_gem_force_release(adev);
+	amdgpu_bo_fini(adev);
+
+	return 0;
+}
+
+static void gmc_v11_0_init_golden_registers(struct amdgpu_device *adev)
+{
+	if (amdgpu_sriov_vf(adev)) {
+		struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
+
+		WREG32(hub->vm_contexts_disable, 0);
+		return;
+	}
+}
+
+/**
+ * gmc_v11_0_gart_enable - gart enable
+ *
+ * @adev: amdgpu_device pointer
+ */
+static int gmc_v11_0_gart_enable(struct amdgpu_device *adev)
+{
+	int r;
+	bool value;
+
+	if (adev->gart.bo == NULL) {
+		dev_err(adev->dev, "No VRAM object for PCIE GART.\n");
+		return -EINVAL;
+	}
+
+	amdgpu_gtt_mgr_recover(&adev->mman.gtt_mgr);
+
+	r = adev->mmhub.funcs->gart_enable(adev);
+	if (r)
+		return r;
+
+	/* Flush HDP after it is initialized */
+	amdgpu_device_flush_hdp(adev, NULL);
+
+	value = amdgpu_vm_fault_stop != AMDGPU_VM_FAULT_STOP_ALWAYS;
+
+	adev->mmhub.funcs->set_fault_enable_default(adev, value);
+	gmc_v11_0_flush_gpu_tlb(adev, 0, AMDGPU_MMHUB0(0), 0);
+
+	DRM_INFO("PCIE GART of %uM enabled (table at 0x%016llX).\n",
+		 (unsigned int)(adev->gmc.gart_size >> 20),
+		 (unsigned long long)amdgpu_bo_gpu_offset(adev->gart.bo));
+
+	return 0;
+}
+
+static int gmc_v11_0_hw_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int r;
+
+	adev->gmc.flush_pasid_uses_kiq = !amdgpu_emu_mode;
+
+	/* The sequence of these two function calls matters.*/
+	gmc_v11_0_init_golden_registers(adev);
+
+	r = gmc_v11_0_gart_enable(adev);
+	if (r)
+		return r;
+
+	if (adev->umc.funcs && adev->umc.funcs->init_registers)
+		adev->umc.funcs->init_registers(adev);
+
+	return 0;
+}
+
+/**
+ * gmc_v11_0_gart_disable - gart disable
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * This disables all VM page table.
+ */
+static void gmc_v11_0_gart_disable(struct amdgpu_device *adev)
+{
+	adev->mmhub.funcs->gart_disable(adev);
+}
+
+static int gmc_v11_0_hw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	if (amdgpu_sriov_vf(adev)) {
+		/* full access mode, so don't touch any GMC register */
+		DRM_DEBUG("For SRIOV client, shouldn't do anything.\n");
+		return 0;
+	}
+
+	amdgpu_irq_put(adev, &adev->gmc.vm_fault, 0);
+
+	if (adev->gmc.ecc_irq.funcs &&
+		amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__UMC))
+		amdgpu_irq_put(adev, &adev->gmc.ecc_irq, 0);
+
+	gmc_v11_0_gart_disable(adev);
+
+	return 0;
+}
+
+static int gmc_v11_0_suspend(struct amdgpu_ip_block *ip_block)
+{
+	gmc_v11_0_hw_fini(ip_block);
+
+	return 0;
+}
+
+static int gmc_v11_0_resume(struct amdgpu_ip_block *ip_block)
+{
+	int r;
+
+	r = gmc_v11_0_hw_init(ip_block);
+	if (r)
+		return r;
+
+	amdgpu_vmid_reset_all(ip_block->adev);
+
+	return 0;
+}
+
+static bool gmc_v11_0_is_idle(struct amdgpu_ip_block *ip_block)
+{
+	/* MC is always ready in GMC v11.*/
+	return true;
+}
+
+static int gmc_v11_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
+{
+	/* There is no need to wait for MC idle in GMC v11.*/
+	return 0;
+}
+
+static int gmc_v11_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
+					   enum amd_clockgating_state state)
+{
+	int r;
+	struct amdgpu_device *adev = ip_block->adev;
+
+	r = adev->mmhub.funcs->set_clockgating(adev, state);
+	if (r)
+		return r;
+
+	return athub_v3_0_set_clockgating(adev, state);
+}
+
+static void gmc_v11_0_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	adev->mmhub.funcs->get_clockgating(adev, flags);
+
+	athub_v3_0_get_clockgating(adev, flags);
+}
+
+static int gmc_v11_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
+					   enum amd_powergating_state state)
+{
+	return 0;
+}
+
+const struct amd_ip_funcs gmc_v11_0_ip_funcs = {
+	.name = "gmc_v11_0",
+	.early_init = gmc_v11_0_early_init,
+	.sw_init = gmc_v11_0_sw_init,
+	.hw_init = gmc_v11_0_hw_init,
+	.late_init = gmc_v11_0_late_init,
+	.sw_fini = gmc_v11_0_sw_fini,
+	.hw_fini = gmc_v11_0_hw_fini,
+	.suspend = gmc_v11_0_suspend,
+	.resume = gmc_v11_0_resume,
+	.is_idle = gmc_v11_0_is_idle,
+	.wait_for_idle = gmc_v11_0_wait_for_idle,
+	.set_clockgating_state = gmc_v11_0_set_clockgating_state,
+	.set_powergating_state = gmc_v11_0_set_powergating_state,
+	.get_clockgating_state = gmc_v11_0_get_clockgating_state,
+};
+
+const struct amdgpu_ip_block_version gmc_v11_0_ip_block = {
+	.type = AMD_IP_BLOCK_TYPE_GMC,
+	.major = 11,
+	.minor = 0,
+	.rev = 0,
+	.funcs = &gmc_v11_0_ip_funcs,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.h b/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.h
new file mode 100644
index 000000000000..def4d5516f82
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.h
@@ -0,0 +1,30 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __GMC_V11_0_H__
+#define __GMC_V11_0_H__
+
+extern const struct amd_ip_funcs gmc_v11_0_ip_funcs;
+extern const struct amdgpu_ip_block_version gmc_v11_0_ip_block;
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v12_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v12_0.c
new file mode 100644
index 000000000000..7a9d6894e321
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v12_0.c
@@ -0,0 +1,1070 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include <linux/firmware.h>
+#include <linux/pci.h>
+
+#include <drm/drm_cache.h>
+
+#include "amdgpu.h"
+#include "amdgpu_atomfirmware.h"
+#include "gmc_v12_0.h"
+#include "athub/athub_4_1_0_sh_mask.h"
+#include "athub/athub_4_1_0_offset.h"
+#include "oss/osssys_7_0_0_offset.h"
+#include "ivsrcid/vmc/irqsrcs_vmc_1_0.h"
+#include "soc24_enum.h"
+#include "soc24.h"
+#include "soc15d.h"
+#include "soc15_common.h"
+#include "nbif_v6_3_1.h"
+#include "gfxhub_v12_0.h"
+#include "mmhub_v4_1_0.h"
+#include "athub_v4_1_0.h"
+#include "umc_v8_14.h"
+
+static int gmc_v12_0_ecc_interrupt_state(struct amdgpu_device *adev,
+					 struct amdgpu_irq_src *src,
+					 unsigned type,
+					 enum amdgpu_interrupt_state state)
+{
+	return 0;
+}
+
+static int gmc_v12_0_vm_fault_interrupt_state(struct amdgpu_device *adev,
+					      struct amdgpu_irq_src *src, unsigned type,
+					      enum amdgpu_interrupt_state state)
+{
+	switch (state) {
+	case AMDGPU_IRQ_STATE_DISABLE:
+		/* MM HUB */
+		amdgpu_gmc_set_vm_fault_masks(adev, AMDGPU_MMHUB0(0), false);
+		/* GFX HUB */
+		/* This works because this interrupt is only
+		 * enabled at init/resume and disabled in
+		 * fini/suspend, so the overall state doesn't
+		 * change over the course of suspend/resume.
+		 */
+		if (!adev->in_s0ix)
+			amdgpu_gmc_set_vm_fault_masks(adev, AMDGPU_GFXHUB(0), false);
+		break;
+	case AMDGPU_IRQ_STATE_ENABLE:
+		/* MM HUB */
+		amdgpu_gmc_set_vm_fault_masks(adev, AMDGPU_MMHUB0(0), true);
+		/* GFX HUB */
+		/* This works because this interrupt is only
+		 * enabled at init/resume and disabled in
+		 * fini/suspend, so the overall state doesn't
+		 * change over the course of suspend/resume.
+		 */
+		if (!adev->in_s0ix)
+			amdgpu_gmc_set_vm_fault_masks(adev, AMDGPU_GFXHUB(0), true);
+		break;
+	default:
+		break;
+	}
+
+	return 0;
+}
+
+static int gmc_v12_0_process_interrupt(struct amdgpu_device *adev,
+				       struct amdgpu_irq_src *source,
+				       struct amdgpu_iv_entry *entry)
+{
+	struct amdgpu_vmhub *hub;
+	bool retry_fault = !!(entry->src_data[1] &
+			      AMDGPU_GMC9_FAULT_SOURCE_DATA_RETRY);
+	bool write_fault = !!(entry->src_data[1] &
+			      AMDGPU_GMC9_FAULT_SOURCE_DATA_WRITE);
+	uint32_t status = 0;
+	u64 addr;
+
+	addr = (u64)entry->src_data[0] << 12;
+	addr |= ((u64)entry->src_data[1] & 0xf) << 44;
+
+	if (entry->client_id == SOC21_IH_CLIENTID_VMC)
+		hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
+	else
+		hub = &adev->vmhub[AMDGPU_GFXHUB(0)];
+
+	if (retry_fault) {
+		/* Returning 1 here also prevents sending the IV to the KFD */
+
+		/* Process it only if it's the first fault for this address */
+		if (entry->ih != &adev->irq.ih_soft &&
+		    amdgpu_gmc_filter_faults(adev, entry->ih, addr, entry->pasid,
+					     entry->timestamp))
+			return 1;
+
+		/* Delegate it to a different ring if the hardware hasn't
+		 * already done it.
+		 */
+		if (entry->ih == &adev->irq.ih) {
+			amdgpu_irq_delegate(adev, entry, 8);
+			return 1;
+		}
+
+		/* Try to handle the recoverable page faults by filling page
+		 * tables
+		 */
+		if (amdgpu_vm_handle_fault(adev, entry->pasid, 0, 0, addr,
+					   entry->timestamp, write_fault))
+			return 1;
+	}
+
+	if (!amdgpu_sriov_vf(adev)) {
+		/*
+		 * Issue a dummy read to wait for the status register to
+		 * be updated to avoid reading an incorrect value due to
+		 * the new fast GRBM interface.
+		 */
+		if (entry->vmid_src == AMDGPU_GFXHUB(0))
+			RREG32(hub->vm_l2_pro_fault_status);
+
+		status = RREG32(hub->vm_l2_pro_fault_status);
+		WREG32_P(hub->vm_l2_pro_fault_cntl, 1, ~1);
+
+		amdgpu_vm_update_fault_cache(adev, entry->pasid, addr, status,
+					     entry->vmid_src ? AMDGPU_MMHUB0(0) : AMDGPU_GFXHUB(0));
+	}
+
+	if (printk_ratelimit()) {
+		struct amdgpu_task_info *task_info;
+
+		dev_err(adev->dev,
+			"[%s] page fault (src_id:%u ring:%u vmid:%u pasid:%u)\n",
+			entry->vmid_src ? "mmhub" : "gfxhub",
+			entry->src_id, entry->ring_id, entry->vmid, entry->pasid);
+		task_info = amdgpu_vm_get_task_info_pasid(adev, entry->pasid);
+		if (task_info) {
+			amdgpu_vm_print_task_info(adev, task_info);
+			amdgpu_vm_put_task_info(task_info);
+		}
+
+		dev_err(adev->dev, "  in page starting at address 0x%016llx from client %d\n",
+				addr, entry->client_id);
+
+		/* Only print L2 fault status if the status register could be read and
+		 * contains useful information
+		 */
+		if (status != 0)
+			hub->vmhub_funcs->print_l2_protection_fault_status(adev, status);
+	}
+
+	return 0;
+}
+
+static const struct amdgpu_irq_src_funcs gmc_v12_0_irq_funcs = {
+	.set = gmc_v12_0_vm_fault_interrupt_state,
+	.process = gmc_v12_0_process_interrupt,
+};
+
+static const struct amdgpu_irq_src_funcs gmc_v12_0_ecc_funcs = {
+	.set = gmc_v12_0_ecc_interrupt_state,
+	.process = amdgpu_umc_process_ecc_irq,
+};
+
+static void gmc_v12_0_set_irq_funcs(struct amdgpu_device *adev)
+{
+	adev->gmc.vm_fault.num_types = 1;
+	adev->gmc.vm_fault.funcs = &gmc_v12_0_irq_funcs;
+
+	if (!amdgpu_sriov_vf(adev)) {
+		adev->gmc.ecc_irq.num_types = 1;
+		adev->gmc.ecc_irq.funcs = &gmc_v12_0_ecc_funcs;
+	}
+}
+
+/**
+ * gmc_v12_0_use_invalidate_semaphore - judge whether to use semaphore
+ *
+ * @adev: amdgpu_device pointer
+ * @vmhub: vmhub type
+ *
+ */
+static bool gmc_v12_0_use_invalidate_semaphore(struct amdgpu_device *adev,
+				       uint32_t vmhub)
+{
+	return ((vmhub == AMDGPU_MMHUB0(0)) &&
+		(!amdgpu_sriov_vf(adev)));
+}
+
+static bool gmc_v12_0_get_vmid_pasid_mapping_info(
+					struct amdgpu_device *adev,
+					uint8_t vmid, uint16_t *p_pasid)
+{
+	*p_pasid = RREG32(SOC15_REG_OFFSET(OSSSYS, 0, regIH_VMID_0_LUT) + vmid) & 0xffff;
+
+	return !!(*p_pasid);
+}
+
+/*
+ * GART
+ * VMID 0 is the physical GPU addresses as used by the kernel.
+ * VMIDs 1-15 are used for userspace clients and are handled
+ * by the amdgpu vm/hsa code.
+ */
+
+static void gmc_v12_0_flush_vm_hub(struct amdgpu_device *adev, uint32_t vmid,
+				   unsigned int vmhub, uint32_t flush_type)
+{
+	bool use_semaphore = gmc_v12_0_use_invalidate_semaphore(adev, vmhub);
+	struct amdgpu_vmhub *hub = &adev->vmhub[vmhub];
+	u32 inv_req = hub->vmhub_funcs->get_invalidate_req(vmid, flush_type);
+	u32 tmp;
+	/* Use register 17 for GART */
+	const unsigned eng = 17;
+	unsigned int i;
+	unsigned char hub_ip = 0;
+
+	hub_ip = (vmhub == AMDGPU_GFXHUB(0)) ?
+		   GC_HWIP : MMHUB_HWIP;
+
+	spin_lock(&adev->gmc.invalidate_lock);
+	/*
+	 * It may lose gpuvm invalidate acknowldege state across power-gating
+	 * off cycle, add semaphore acquire before invalidation and semaphore
+	 * release after invalidation to avoid entering power gated state
+	 * to WA the Issue
+	 */
+
+	/* TODO: It needs to continue working on debugging with semaphore for GFXHUB as well. */
+	if (use_semaphore) {
+		for (i = 0; i < adev->usec_timeout; i++) {
+			/* a read return value of 1 means semaphore acuqire */
+			tmp = RREG32_RLC_NO_KIQ(hub->vm_inv_eng0_sem +
+					    hub->eng_distance * eng, hub_ip);
+			if (tmp & 0x1)
+				break;
+			udelay(1);
+		}
+
+		if (i >= adev->usec_timeout)
+			dev_err(adev->dev,
+				"Timeout waiting for sem acquire in VM flush!\n");
+	}
+
+	WREG32_RLC_NO_KIQ(hub->vm_inv_eng0_req + hub->eng_distance * eng, inv_req, hub_ip);
+
+	/* Wait for ACK with a delay.*/
+	for (i = 0; i < adev->usec_timeout; i++) {
+		tmp = RREG32_RLC_NO_KIQ(hub->vm_inv_eng0_ack +
+				    hub->eng_distance * eng, hub_ip);
+		tmp &= 1 << vmid;
+		if (tmp)
+			break;
+
+		udelay(1);
+	}
+
+	/* TODO: It needs to continue working on debugging with semaphore for GFXHUB as well. */
+	if (use_semaphore)
+		/*
+		 * add semaphore release after invalidation,
+		 * write with 0 means semaphore release
+		 */
+		WREG32_RLC_NO_KIQ(hub->vm_inv_eng0_sem +
+			      hub->eng_distance * eng, 0, hub_ip);
+
+	/* Issue additional private vm invalidation to MMHUB */
+	if ((vmhub != AMDGPU_GFXHUB(0)) &&
+	    (hub->vm_l2_bank_select_reserved_cid2) &&
+		!amdgpu_sriov_vf(adev)) {
+		inv_req = RREG32_NO_KIQ(hub->vm_l2_bank_select_reserved_cid2);
+		/* bit 25: RSERVED_CACHE_PRIVATE_INVALIDATION */
+		inv_req |= (1 << 25);
+		/* Issue private invalidation */
+		WREG32_NO_KIQ(hub->vm_l2_bank_select_reserved_cid2, inv_req);
+		/* Read back to ensure invalidation is done*/
+		RREG32_NO_KIQ(hub->vm_l2_bank_select_reserved_cid2);
+	}
+
+	spin_unlock(&adev->gmc.invalidate_lock);
+
+	if (i < adev->usec_timeout)
+		return;
+
+	dev_err(adev->dev, "Timeout waiting for VM flush ACK!\n");
+}
+
+/**
+ * gmc_v12_0_flush_gpu_tlb - gart tlb flush callback
+ *
+ * @adev: amdgpu_device pointer
+ * @vmid: vm instance to flush
+ * @vmhub: which hub to flush
+ * @flush_type: the flush type
+ *
+ * Flush the TLB for the requested page table.
+ */
+static void gmc_v12_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid,
+					uint32_t vmhub, uint32_t flush_type)
+{
+	if ((vmhub == AMDGPU_GFXHUB(0)) && !adev->gfx.is_poweron)
+		return;
+
+	/* flush hdp cache */
+	amdgpu_device_flush_hdp(adev, NULL);
+
+	/* This is necessary for SRIOV as well as for GFXOFF to function
+	 * properly under bare metal
+	 */
+	if ((adev->gfx.kiq[0].ring.sched.ready || adev->mes.ring[0].sched.ready) &&
+	    (amdgpu_sriov_runtime(adev) || !amdgpu_sriov_vf(adev))) {
+		struct amdgpu_vmhub *hub = &adev->vmhub[vmhub];
+		const unsigned eng = 17;
+		u32 inv_req = hub->vmhub_funcs->get_invalidate_req(vmid, flush_type);
+		u32 req = hub->vm_inv_eng0_req + hub->eng_distance * eng;
+		u32 ack = hub->vm_inv_eng0_ack + hub->eng_distance * eng;
+
+		amdgpu_gmc_fw_reg_write_reg_wait(adev, req, ack, inv_req,
+				1 << vmid, GET_INST(GC, 0));
+		return;
+	}
+
+	gmc_v12_0_flush_vm_hub(adev, vmid, vmhub, 0);
+	return;
+}
+
+/**
+ * gmc_v12_0_flush_gpu_tlb_pasid - tlb flush via pasid
+ *
+ * @adev: amdgpu_device pointer
+ * @pasid: pasid to be flush
+ * @flush_type: the flush type
+ * @all_hub: flush all hubs
+ * @inst: is used to select which instance of KIQ to use for the invalidation
+ *
+ * Flush the TLB for the requested pasid.
+ */
+static void gmc_v12_0_flush_gpu_tlb_pasid(struct amdgpu_device *adev,
+					  uint16_t pasid, uint32_t flush_type,
+					  bool all_hub, uint32_t inst)
+{
+	uint16_t queried;
+	int vmid, i;
+
+	if (adev->enable_uni_mes && adev->mes.ring[AMDGPU_MES_SCHED_PIPE].sched.ready &&
+	    (adev->mes.sched_version & AMDGPU_MES_VERSION_MASK) >= 0x84) {
+		struct mes_inv_tlbs_pasid_input input = {0};
+		input.pasid = pasid;
+		input.flush_type = flush_type;
+		input.hub_id = AMDGPU_GFXHUB(0);
+		/* MES will invalidate all gc_hub for the device from master */
+		adev->mes.funcs->invalidate_tlbs_pasid(&adev->mes, &input);
+		if (all_hub) {
+			/* Only need to invalidate mm_hub now, gfx12 only support one mmhub */
+			input.hub_id = AMDGPU_MMHUB0(0);
+			adev->mes.funcs->invalidate_tlbs_pasid(&adev->mes, &input);
+		}
+		return;
+	}
+
+	for (vmid = 1; vmid < 16; vmid++) {
+		bool valid;
+
+		valid = gmc_v12_0_get_vmid_pasid_mapping_info(adev, vmid,
+							      &queried);
+		if (!valid || queried != pasid)
+			continue;
+
+		if (all_hub) {
+			for_each_set_bit(i, adev->vmhubs_mask,
+					 AMDGPU_MAX_VMHUBS)
+				gmc_v12_0_flush_gpu_tlb(adev, vmid, i,
+							flush_type);
+		} else {
+			gmc_v12_0_flush_gpu_tlb(adev, vmid, AMDGPU_GFXHUB(0),
+						flush_type);
+		}
+	}
+}
+
+static uint64_t gmc_v12_0_emit_flush_gpu_tlb(struct amdgpu_ring *ring,
+					     unsigned vmid, uint64_t pd_addr)
+{
+	bool use_semaphore = gmc_v12_0_use_invalidate_semaphore(ring->adev, ring->vm_hub);
+	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->vm_hub];
+	uint32_t req = hub->vmhub_funcs->get_invalidate_req(vmid, 0);
+	unsigned eng = ring->vm_inv_eng;
+
+	/*
+	 * It may lose gpuvm invalidate acknowldege state across power-gating
+	 * off cycle, add semaphore acquire before invalidation and semaphore
+	 * release after invalidation to avoid entering power gated state
+	 * to WA the Issue
+	 */
+
+	/* TODO: It needs to continue working on debugging with semaphore for GFXHUB as well. */
+	if (use_semaphore)
+		/* a read return value of 1 means semaphore acuqire */
+		amdgpu_ring_emit_reg_wait(ring,
+					  hub->vm_inv_eng0_sem +
+					  hub->eng_distance * eng, 0x1, 0x1);
+
+	amdgpu_ring_emit_wreg(ring, hub->ctx0_ptb_addr_lo32 +
+			      (hub->ctx_addr_distance * vmid),
+			      lower_32_bits(pd_addr));
+
+	amdgpu_ring_emit_wreg(ring, hub->ctx0_ptb_addr_hi32 +
+			      (hub->ctx_addr_distance * vmid),
+			      upper_32_bits(pd_addr));
+
+	amdgpu_ring_emit_reg_write_reg_wait(ring, hub->vm_inv_eng0_req +
+					    hub->eng_distance * eng,
+					    hub->vm_inv_eng0_ack +
+					    hub->eng_distance * eng,
+					    req, 1 << vmid);
+
+	/* TODO: It needs to continue working on debugging with semaphore for GFXHUB as well. */
+	if (use_semaphore)
+		/*
+		 * add semaphore release after invalidation,
+		 * write with 0 means semaphore release
+		 */
+		amdgpu_ring_emit_wreg(ring, hub->vm_inv_eng0_sem +
+				      hub->eng_distance * eng, 0);
+
+	return pd_addr;
+}
+
+static void gmc_v12_0_emit_pasid_mapping(struct amdgpu_ring *ring, unsigned vmid,
+					 unsigned pasid)
+{
+	struct amdgpu_device *adev = ring->adev;
+	uint32_t reg;
+
+	if (ring->vm_hub == AMDGPU_GFXHUB(0))
+		reg = SOC15_REG_OFFSET(OSSSYS, 0, regIH_VMID_0_LUT) + vmid;
+	else
+		reg = SOC15_REG_OFFSET(OSSSYS, 0, regIH_VMID_0_LUT_MM) + vmid;
+
+	amdgpu_ring_emit_wreg(ring, reg, pasid);
+}
+
+/*
+ * PTE format:
+ * 63 P
+ * 62:59 reserved
+ * 58 D
+ * 57 G
+ * 56 T
+ * 55:54 M
+ * 53:52 SW
+ * 51:48 reserved for future
+ * 47:12 4k physical page base address
+ * 11:7 fragment
+ * 6 write
+ * 5 read
+ * 4 exe
+ * 3 Z
+ * 2 snooped
+ * 1 system
+ * 0 valid
+ *
+ * PDE format:
+ * 63 P
+ * 62:58 block fragment size
+ * 57 reserved
+ * 56 A
+ * 55:54 M
+ * 53:52 reserved
+ * 51:48 reserved for future
+ * 47:6 physical base address of PD or PTE
+ * 5:3 reserved
+ * 2 C
+ * 1 system
+ * 0 valid
+ */
+
+static void gmc_v12_0_get_vm_pde(struct amdgpu_device *adev, int level,
+				 uint64_t *addr, uint64_t *flags)
+{
+	if (!(*flags & AMDGPU_PDE_PTE_GFX12) && !(*flags & AMDGPU_PTE_SYSTEM))
+		*addr = adev->vm_manager.vram_base_offset + *addr -
+			adev->gmc.vram_start;
+	BUG_ON(*addr & 0xFFFF00000000003FULL);
+
+	if (!adev->gmc.translate_further)
+		return;
+
+	if (level == AMDGPU_VM_PDB1) {
+		/* Set the block fragment size */
+		if (!(*flags & AMDGPU_PDE_PTE_GFX12))
+			*flags |= AMDGPU_PDE_BFS_GFX12(0x9);
+
+	} else if (level == AMDGPU_VM_PDB0) {
+		if (*flags & AMDGPU_PDE_PTE_GFX12)
+			*flags &= ~AMDGPU_PDE_PTE_GFX12;
+	}
+}
+
+static void gmc_v12_0_get_vm_pte(struct amdgpu_device *adev,
+				 struct amdgpu_vm *vm,
+				 struct amdgpu_bo *bo,
+				 uint32_t vm_flags,
+				 uint64_t *flags)
+{
+	if (vm_flags & AMDGPU_VM_PAGE_EXECUTABLE)
+		*flags |= AMDGPU_PTE_EXECUTABLE;
+	else
+		*flags &= ~AMDGPU_PTE_EXECUTABLE;
+
+	switch (vm_flags & AMDGPU_VM_MTYPE_MASK) {
+	case AMDGPU_VM_MTYPE_DEFAULT:
+		*flags = AMDGPU_PTE_MTYPE_GFX12(*flags, MTYPE_NC);
+		break;
+	case AMDGPU_VM_MTYPE_NC:
+	default:
+		*flags = AMDGPU_PTE_MTYPE_GFX12(*flags, MTYPE_NC);
+		break;
+	case AMDGPU_VM_MTYPE_UC:
+		*flags = AMDGPU_PTE_MTYPE_GFX12(*flags, MTYPE_UC);
+		break;
+	}
+
+	if (vm_flags & AMDGPU_VM_PAGE_NOALLOC)
+		*flags |= AMDGPU_PTE_NOALLOC;
+	else
+		*flags &= ~AMDGPU_PTE_NOALLOC;
+
+	if (vm_flags & AMDGPU_VM_PAGE_PRT) {
+		*flags |= AMDGPU_PTE_PRT_GFX12;
+		*flags |= AMDGPU_PTE_SNOOPED;
+		*flags |= AMDGPU_PTE_SYSTEM;
+		*flags |= AMDGPU_PTE_IS_PTE;
+		*flags &= ~AMDGPU_PTE_VALID;
+	}
+
+	if (bo && bo->flags & AMDGPU_GEM_CREATE_GFX12_DCC)
+		*flags |= AMDGPU_PTE_DCC;
+
+	if (bo && bo->flags & AMDGPU_GEM_CREATE_UNCACHED)
+		*flags = AMDGPU_PTE_MTYPE_GFX12(*flags, MTYPE_UC);
+}
+
+static unsigned gmc_v12_0_get_vbios_fb_size(struct amdgpu_device *adev)
+{
+	return 0;
+}
+
+static unsigned int gmc_v12_0_get_dcc_alignment(struct amdgpu_device *adev)
+{
+	unsigned int max_tex_channel_caches, alignment;
+
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) != IP_VERSION(12, 0, 0) &&
+	    amdgpu_ip_version(adev, GC_HWIP, 0) != IP_VERSION(12, 0, 1))
+		return 0;
+
+	max_tex_channel_caches = adev->gfx.config.max_texture_channel_caches;
+	if (is_power_of_2(max_tex_channel_caches))
+		alignment = (unsigned int)(max_tex_channel_caches / SZ_4);
+	else
+		alignment = roundup_pow_of_two(max_tex_channel_caches);
+
+	return (unsigned int)(alignment * max_tex_channel_caches * SZ_1K);
+}
+
+static const struct amdgpu_gmc_funcs gmc_v12_0_gmc_funcs = {
+	.flush_gpu_tlb = gmc_v12_0_flush_gpu_tlb,
+	.flush_gpu_tlb_pasid = gmc_v12_0_flush_gpu_tlb_pasid,
+	.emit_flush_gpu_tlb = gmc_v12_0_emit_flush_gpu_tlb,
+	.emit_pasid_mapping = gmc_v12_0_emit_pasid_mapping,
+	.get_vm_pde = gmc_v12_0_get_vm_pde,
+	.get_vm_pte = gmc_v12_0_get_vm_pte,
+	.get_vbios_fb_size = gmc_v12_0_get_vbios_fb_size,
+	.get_dcc_alignment = gmc_v12_0_get_dcc_alignment,
+};
+
+static void gmc_v12_0_set_gmc_funcs(struct amdgpu_device *adev)
+{
+	adev->gmc.gmc_funcs = &gmc_v12_0_gmc_funcs;
+}
+
+static void gmc_v12_0_set_umc_funcs(struct amdgpu_device *adev)
+{
+	switch (amdgpu_ip_version(adev, UMC_HWIP, 0)) {
+	case IP_VERSION(8, 14, 0):
+		adev->umc.channel_inst_num = UMC_V8_14_CHANNEL_INSTANCE_NUM;
+		adev->umc.umc_inst_num = UMC_V8_14_UMC_INSTANCE_NUM(adev);
+		adev->umc.node_inst_num = 0;
+		adev->umc.max_ras_err_cnt_per_query = UMC_V8_14_TOTAL_CHANNEL_NUM(adev);
+		adev->umc.channel_offs = UMC_V8_14_PER_CHANNEL_OFFSET;
+		adev->umc.ras = &umc_v8_14_ras;
+		break;
+	default:
+		break;
+	}
+}
+
+
+static void gmc_v12_0_set_mmhub_funcs(struct amdgpu_device *adev)
+{
+	switch (amdgpu_ip_version(adev, MMHUB_HWIP, 0)) {
+	case IP_VERSION(4, 1, 0):
+		adev->mmhub.funcs = &mmhub_v4_1_0_funcs;
+		break;
+	default:
+		break;
+	}
+}
+
+static void gmc_v12_0_set_gfxhub_funcs(struct amdgpu_device *adev)
+{
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	case IP_VERSION(12, 0, 0):
+	case IP_VERSION(12, 0, 1):
+		adev->gfxhub.funcs = &gfxhub_v12_0_funcs;
+		break;
+	default:
+		break;
+	}
+}
+
+static int gmc_v12_0_early_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	gmc_v12_0_set_gfxhub_funcs(adev);
+	gmc_v12_0_set_mmhub_funcs(adev);
+	gmc_v12_0_set_gmc_funcs(adev);
+	gmc_v12_0_set_irq_funcs(adev);
+	gmc_v12_0_set_umc_funcs(adev);
+
+	adev->gmc.shared_aperture_start = 0x2000000000000000ULL;
+	adev->gmc.shared_aperture_end =
+		adev->gmc.shared_aperture_start + (4ULL << 30) - 1;
+	adev->gmc.private_aperture_start = 0x1000000000000000ULL;
+	adev->gmc.private_aperture_end =
+		adev->gmc.private_aperture_start + (4ULL << 30) - 1;
+
+	return 0;
+}
+
+static int gmc_v12_0_late_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int r;
+
+	r = amdgpu_gmc_allocate_vm_inv_eng(adev);
+	if (r)
+		return r;
+
+	r = amdgpu_gmc_ras_late_init(adev);
+	if (r)
+		return r;
+
+	return amdgpu_irq_get(adev, &adev->gmc.vm_fault, 0);
+}
+
+static void gmc_v12_0_vram_gtt_location(struct amdgpu_device *adev,
+					struct amdgpu_gmc *mc)
+{
+	u64 base = 0;
+
+	base = adev->mmhub.funcs->get_fb_location(adev);
+
+	amdgpu_gmc_set_agp_default(adev, mc);
+	amdgpu_gmc_vram_location(adev, &adev->gmc, base);
+	amdgpu_gmc_gart_location(adev, mc, AMDGPU_GART_PLACEMENT_LOW);
+	if (!amdgpu_sriov_vf(adev) && (amdgpu_agp == 1))
+		amdgpu_gmc_agp_location(adev, mc);
+
+	/* base offset of vram pages */
+	if (amdgpu_sriov_vf(adev))
+		adev->vm_manager.vram_base_offset = 0;
+	else
+		adev->vm_manager.vram_base_offset = adev->mmhub.funcs->get_mc_fb_offset(adev);
+}
+
+/**
+ * gmc_v12_0_mc_init - initialize the memory controller driver params
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Look up the amount of vram, vram width, and decide how to place
+ * vram and gart within the GPU's physical address space.
+ * Returns 0 for success.
+ */
+static int gmc_v12_0_mc_init(struct amdgpu_device *adev)
+{
+	int r;
+
+	/* size in MB on si */
+	adev->gmc.mc_vram_size =
+		adev->nbio.funcs->get_memsize(adev) * 1024ULL * 1024ULL;
+	adev->gmc.real_vram_size = adev->gmc.mc_vram_size;
+
+	if (!(adev->flags & AMD_IS_APU)) {
+		r = amdgpu_device_resize_fb_bar(adev);
+		if (r)
+			return r;
+	}
+
+	adev->gmc.aper_base = pci_resource_start(adev->pdev, 0);
+	adev->gmc.aper_size = pci_resource_len(adev->pdev, 0);
+
+#ifdef CONFIG_X86_64
+	if ((adev->flags & AMD_IS_APU) && !amdgpu_passthrough(adev)) {
+		adev->gmc.aper_base = adev->mmhub.funcs->get_mc_fb_offset(adev);
+		adev->gmc.aper_size = adev->gmc.real_vram_size;
+	}
+#endif
+	/* In case the PCI BAR is larger than the actual amount of vram */
+	adev->gmc.visible_vram_size = adev->gmc.aper_size;
+	if (adev->gmc.visible_vram_size > adev->gmc.real_vram_size)
+		adev->gmc.visible_vram_size = adev->gmc.real_vram_size;
+
+	/* set the gart size */
+	if (amdgpu_gart_size == -1) {
+		adev->gmc.gart_size = 512ULL << 20;
+	} else
+		adev->gmc.gart_size = (u64)amdgpu_gart_size << 20;
+
+	gmc_v12_0_vram_gtt_location(adev, &adev->gmc);
+
+	return 0;
+}
+
+static int gmc_v12_0_gart_init(struct amdgpu_device *adev)
+{
+	int r;
+
+	if (adev->gart.bo) {
+		WARN(1, "PCIE GART already initialized\n");
+		return 0;
+	}
+
+	/* Initialize common gart structure */
+	r = amdgpu_gart_init(adev);
+	if (r)
+		return r;
+
+	adev->gart.table_size = adev->gart.num_gpu_pages * 8;
+	adev->gart.gart_pte_flags = AMDGPU_PTE_MTYPE_GFX12(0ULL, MTYPE_UC) |
+				    AMDGPU_PTE_EXECUTABLE |
+				    AMDGPU_PTE_IS_PTE;
+
+	return amdgpu_gart_table_vram_alloc(adev);
+}
+
+static int gmc_v12_0_sw_init(struct amdgpu_ip_block *ip_block)
+{
+	int r, vram_width = 0, vram_type = 0, vram_vendor = 0;
+	struct amdgpu_device *adev = ip_block->adev;
+
+	adev->mmhub.funcs->init(adev);
+
+	adev->gfxhub.funcs->init(adev);
+
+	spin_lock_init(&adev->gmc.invalidate_lock);
+
+	r = amdgpu_atomfirmware_get_vram_info(adev,
+					      &vram_width, &vram_type, &vram_vendor);
+	adev->gmc.vram_width = vram_width;
+
+	adev->gmc.vram_type = vram_type;
+	adev->gmc.vram_vendor = vram_vendor;
+
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	case IP_VERSION(12, 0, 0):
+	case IP_VERSION(12, 0, 1):
+		set_bit(AMDGPU_GFXHUB(0), adev->vmhubs_mask);
+		set_bit(AMDGPU_MMHUB0(0), adev->vmhubs_mask);
+		/*
+		 * To fulfill 4-level page support,
+		 * vm size is 256TB (48bit), maximum size,
+		 * block size 512 (9bit)
+		 */
+		amdgpu_vm_adjust_size(adev, 256 * 1024, 9, 3, 48);
+		break;
+	default:
+		break;
+	}
+
+	/* This interrupt is VMC page fault.*/
+	r = amdgpu_irq_add_id(adev, SOC21_IH_CLIENTID_VMC,
+			      VMC_1_0__SRCID__VM_FAULT,
+			      &adev->gmc.vm_fault);
+
+	if (r)
+		return r;
+
+	r = amdgpu_irq_add_id(adev, SOC21_IH_CLIENTID_GFX,
+			      UTCL2_1_0__SRCID__FAULT,
+			      &adev->gmc.vm_fault);
+	if (r)
+		return r;
+
+	if (!amdgpu_sriov_vf(adev)) {
+		/* interrupt sent to DF. */
+		r = amdgpu_irq_add_id(adev, SOC21_IH_CLIENTID_DF, 0,
+				      &adev->gmc.ecc_irq);
+		if (r)
+			return r;
+	}
+
+	/*
+	 * Set the internal MC address mask This is the max address of the GPU's
+	 * internal address space.
+	 */
+	adev->gmc.mc_mask = 0xffffffffffffULL; /* 48 bit MC */
+
+	r = dma_set_mask_and_coherent(adev->dev, DMA_BIT_MASK(44));
+	if (r) {
+		printk(KERN_WARNING "amdgpu: No suitable DMA available.\n");
+		return r;
+	}
+
+	adev->need_swiotlb = drm_need_swiotlb(44);
+
+	r = gmc_v12_0_mc_init(adev);
+	if (r)
+		return r;
+
+	amdgpu_gmc_get_vbios_allocations(adev);
+
+	/* Memory manager */
+	r = amdgpu_bo_init(adev);
+	if (r)
+		return r;
+
+	r = gmc_v12_0_gart_init(adev);
+	if (r)
+		return r;
+
+	/*
+	 * number of VMs
+	 * VMID 0 is reserved for System
+	 * amdgpu graphics/compute will use VMIDs 1-7
+	 * amdkfd will use VMIDs 8-15
+	 */
+	adev->vm_manager.first_kfd_vmid = adev->gfx.disable_kq ? 1 : 8;
+
+	amdgpu_vm_manager_init(adev);
+
+	r = amdgpu_gmc_ras_sw_init(adev);
+	if (r)
+		return r;
+
+	return 0;
+}
+
+/**
+ * gmc_v12_0_gart_fini - vm fini callback
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Tears down the driver GART/VM setup (CIK).
+ */
+static void gmc_v12_0_gart_fini(struct amdgpu_device *adev)
+{
+	amdgpu_gart_table_vram_free(adev);
+}
+
+static int gmc_v12_0_sw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	amdgpu_vm_manager_fini(adev);
+	gmc_v12_0_gart_fini(adev);
+	amdgpu_gem_force_release(adev);
+	amdgpu_bo_fini(adev);
+
+	return 0;
+}
+
+static void gmc_v12_0_init_golden_registers(struct amdgpu_device *adev)
+{
+}
+
+/**
+ * gmc_v12_0_gart_enable - gart enable
+ *
+ * @adev: amdgpu_device pointer
+ */
+static int gmc_v12_0_gart_enable(struct amdgpu_device *adev)
+{
+	int r;
+	bool value;
+
+	if (adev->gart.bo == NULL) {
+		dev_err(adev->dev, "No VRAM object for PCIE GART.\n");
+		return -EINVAL;
+	}
+
+	amdgpu_gtt_mgr_recover(&adev->mman.gtt_mgr);
+
+	r = adev->mmhub.funcs->gart_enable(adev);
+	if (r)
+		return r;
+
+	/* Flush HDP after it is initialized */
+	amdgpu_device_flush_hdp(adev, NULL);
+
+	value = amdgpu_vm_fault_stop != AMDGPU_VM_FAULT_STOP_ALWAYS;
+
+	adev->mmhub.funcs->set_fault_enable_default(adev, value);
+	gmc_v12_0_flush_gpu_tlb(adev, 0, AMDGPU_MMHUB0(0), 0);
+
+	dev_info(adev->dev, "PCIE GART of %uM enabled (table at 0x%016llX).\n",
+		 (unsigned)(adev->gmc.gart_size >> 20),
+		 (unsigned long long)amdgpu_bo_gpu_offset(adev->gart.bo));
+
+	return 0;
+}
+
+static int gmc_v12_0_hw_init(struct amdgpu_ip_block *ip_block)
+{
+	int r;
+	struct amdgpu_device *adev = ip_block->adev;
+
+	/* The sequence of these two function calls matters.*/
+	gmc_v12_0_init_golden_registers(adev);
+
+	r = gmc_v12_0_gart_enable(adev);
+	if (r)
+		return r;
+
+	if (adev->umc.funcs && adev->umc.funcs->init_registers)
+		adev->umc.funcs->init_registers(adev);
+
+	return 0;
+}
+
+/**
+ * gmc_v12_0_gart_disable - gart disable
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * This disables all VM page table.
+ */
+static void gmc_v12_0_gart_disable(struct amdgpu_device *adev)
+{
+	adev->mmhub.funcs->gart_disable(adev);
+}
+
+static int gmc_v12_0_hw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	if (amdgpu_sriov_vf(adev)) {
+		/* full access mode, so don't touch any GMC register */
+		DRM_DEBUG("For SRIOV client, shouldn't do anything.\n");
+		return 0;
+	}
+
+	amdgpu_irq_put(adev, &adev->gmc.vm_fault, 0);
+
+	if (adev->gmc.ecc_irq.funcs &&
+		amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__UMC))
+		amdgpu_irq_put(adev, &adev->gmc.ecc_irq, 0);
+
+	gmc_v12_0_gart_disable(adev);
+
+	return 0;
+}
+
+static int gmc_v12_0_suspend(struct amdgpu_ip_block *ip_block)
+{
+	gmc_v12_0_hw_fini(ip_block);
+
+	return 0;
+}
+
+static int gmc_v12_0_resume(struct amdgpu_ip_block *ip_block)
+{
+	int r;
+
+	r = gmc_v12_0_hw_init(ip_block);
+	if (r)
+		return r;
+
+	amdgpu_vmid_reset_all(ip_block->adev);
+
+	return 0;
+}
+
+static bool gmc_v12_0_is_idle(struct amdgpu_ip_block *ip_block)
+{
+	/* MC is always ready in GMC v11.*/
+	return true;
+}
+
+static int gmc_v12_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
+{
+	/* There is no need to wait for MC idle in GMC v11.*/
+	return 0;
+}
+
+static int gmc_v12_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
+					   enum amd_clockgating_state state)
+{
+	int r;
+	struct amdgpu_device *adev = ip_block->adev;
+
+	r = adev->mmhub.funcs->set_clockgating(adev, state);
+	if (r)
+		return r;
+
+	return athub_v4_1_0_set_clockgating(adev, state);
+}
+
+static void gmc_v12_0_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	adev->mmhub.funcs->get_clockgating(adev, flags);
+
+	athub_v4_1_0_get_clockgating(adev, flags);
+}
+
+static int gmc_v12_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
+					   enum amd_powergating_state state)
+{
+	return 0;
+}
+
+const struct amd_ip_funcs gmc_v12_0_ip_funcs = {
+	.name = "gmc_v12_0",
+	.early_init = gmc_v12_0_early_init,
+	.sw_init = gmc_v12_0_sw_init,
+	.hw_init = gmc_v12_0_hw_init,
+	.late_init = gmc_v12_0_late_init,
+	.sw_fini = gmc_v12_0_sw_fini,
+	.hw_fini = gmc_v12_0_hw_fini,
+	.suspend = gmc_v12_0_suspend,
+	.resume = gmc_v12_0_resume,
+	.is_idle = gmc_v12_0_is_idle,
+	.wait_for_idle = gmc_v12_0_wait_for_idle,
+	.set_clockgating_state = gmc_v12_0_set_clockgating_state,
+	.set_powergating_state = gmc_v12_0_set_powergating_state,
+	.get_clockgating_state = gmc_v12_0_get_clockgating_state,
+};
+
+const struct amdgpu_ip_block_version gmc_v12_0_ip_block = {
+	.type = AMD_IP_BLOCK_TYPE_GMC,
+	.major = 12,
+	.minor = 0,
+	.rev = 0,
+	.funcs = &gmc_v12_0_ip_funcs,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v12_0.h b/drivers/gpu/drm/amd/amdgpu/gmc_v12_0.h
new file mode 100644
index 000000000000..deca93e4a156
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v12_0.h
@@ -0,0 +1,30 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __GMC_V12_0_H__
+#define __GMC_V12_0_H__
+
+extern const struct amd_ip_funcs gmc_v12_0_ip_funcs;
+extern const struct amdgpu_ip_block_version gmc_v12_0_ip_block;
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c
index 0fe714f54cca..a8ec95f42926 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c
@@ -43,7 +43,7 @@
 
 static void gmc_v6_0_set_gmc_funcs(struct amdgpu_device *adev);
 static void gmc_v6_0_set_irq_funcs(struct amdgpu_device *adev);
-static int gmc_v6_0_wait_for_idle(void *handle);
+static int gmc_v6_0_wait_for_idle(struct amdgpu_ip_block *ip_block);
 
 MODULE_FIRMWARE("amdgpu/tahiti_mc.bin");
 MODULE_FIRMWARE("amdgpu/pitcairn_mc.bin");
@@ -64,8 +64,13 @@ MODULE_FIRMWARE("amdgpu/si58_mc.bin");
 static void gmc_v6_0_mc_stop(struct amdgpu_device *adev)
 {
 	u32 blackout;
+	struct amdgpu_ip_block *ip_block;
 
-	gmc_v6_0_wait_for_idle((void *)adev);
+	ip_block = amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_GMC);
+	if (!ip_block)
+		return;
+
+	gmc_v6_0_wait_for_idle(ip_block);
 
 	blackout = RREG32(mmMC_SHARED_BLACKOUT_CNTL);
 	if (REG_GET_FIELD(blackout, MC_SHARED_BLACKOUT_CNTL, BLACKOUT_MODE) != 1) {
@@ -98,9 +103,7 @@ static void gmc_v6_0_mc_resume(struct amdgpu_device *adev)
 static int gmc_v6_0_init_microcode(struct amdgpu_device *adev)
 {
 	const char *chip_name;
-	char fw_name[30];
 	int err;
-	bool is_58_fw = false;
 
 	DRM_DEBUG("\n");
 
@@ -120,30 +123,21 @@ static int gmc_v6_0_init_microcode(struct amdgpu_device *adev)
 	case CHIP_HAINAN:
 		chip_name = "hainan";
 		break;
-	default: BUG();
+	default:
+		BUG();
 	}
 
 	/* this memory configuration requires special firmware */
 	if (((RREG32(mmMC_SEQ_MISC0) & 0xff000000) >> 24) == 0x58)
-		is_58_fw = true;
-
-	if (is_58_fw)
-		snprintf(fw_name, sizeof(fw_name), "amdgpu/si58_mc.bin");
-	else
-		snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_mc.bin", chip_name);
-	err = request_firmware(&adev->gmc.fw, fw_name, adev->dev);
-	if (err)
-		goto out;
-
-	err = amdgpu_ucode_validate(adev->gmc.fw);
+		chip_name = "si58";
 
-out:
+	err = amdgpu_ucode_request(adev, &adev->gmc.fw, AMDGPU_UCODE_REQUIRED,
+				   "amdgpu/%s_mc.bin", chip_name);
 	if (err) {
 		dev_err(adev->dev,
-		       "si_mc: Failed to load firmware \"%s\"\n",
-		       fw_name);
-		release_firmware(adev->gmc.fw);
-		adev->gmc.fw = NULL;
+		       "si_mc: Failed to load firmware \"%s_mc.bin\"\n",
+		       chip_name);
+		amdgpu_ucode_release(&adev->gmc.fw);
 	}
 	return err;
 }
@@ -185,9 +179,8 @@ static int gmc_v6_0_mc_load_microcode(struct amdgpu_device *adev)
 			WREG32(mmMC_SEQ_IO_DEBUG_DATA, le32_to_cpup(new_io_mc_regs++));
 		}
 		/* load the MC ucode */
-		for (i = 0; i < ucode_size; i++) {
+		for (i = 0; i < ucode_size; i++)
 			WREG32(mmMC_SEQ_SUP_PGM, le32_to_cpup(new_fw_data++));
-		}
 
 		/* put the engine back into the active state */
 		WREG32(mmMC_SEQ_SUP_CNTL, 0x00000008);
@@ -215,15 +208,19 @@ static void gmc_v6_0_vram_gtt_location(struct amdgpu_device *adev,
 				       struct amdgpu_gmc *mc)
 {
 	u64 base = RREG32(mmMC_VM_FB_LOCATION) & 0xFFFF;
+
 	base <<= 24;
 
+	amdgpu_gmc_set_agp_default(adev, mc);
 	amdgpu_gmc_vram_location(adev, mc, base);
-	amdgpu_gmc_gart_location(adev, mc);
+	amdgpu_gmc_gart_location(adev, mc, AMDGPU_GART_PLACEMENT_LOW);
 }
 
 static void gmc_v6_0_mc_program(struct amdgpu_device *adev)
 {
 	int i, j;
+	struct amdgpu_ip_block *ip_block;
+
 
 	/* Initialize HDP */
 	for (i = 0, j = 0; i < 32; i++, j += 0x6) {
@@ -235,9 +232,12 @@ static void gmc_v6_0_mc_program(struct amdgpu_device *adev)
 	}
 	WREG32(mmHDP_REG_COHERENCY_FLUSH_CNTL, 0);
 
-	if (gmc_v6_0_wait_for_idle((void *)adev)) {
+	ip_block = amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_GMC);
+	if (!ip_block)
+		return;
+
+	if (gmc_v6_0_wait_for_idle(ip_block))
 		dev_warn(adev->dev, "Wait for MC idle timedout !\n");
-	}
 
 	if (adev->mode_info.num_crtc) {
 		u32 tmp;
@@ -249,7 +249,7 @@ static void gmc_v6_0_mc_program(struct amdgpu_device *adev)
 
 		/* disable VGA render */
 		tmp = RREG32(mmVGA_RENDER_CONTROL);
-		tmp &= ~VGA_VSTATUS_CNTL;
+		tmp &= VGA_RENDER_CONTROL__VGA_VSTATUS_CNTL_MASK;
 		WREG32(mmVGA_RENDER_CONTROL, tmp);
 	}
 	/* Update configuration */
@@ -258,14 +258,13 @@ static void gmc_v6_0_mc_program(struct amdgpu_device *adev)
 	WREG32(mmMC_VM_SYSTEM_APERTURE_HIGH_ADDR,
 	       adev->gmc.vram_end >> 12);
 	WREG32(mmMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR,
-	       adev->vram_scratch.gpu_addr >> 12);
+	       adev->mem_scratch.gpu_addr >> 12);
 	WREG32(mmMC_VM_AGP_BASE, 0);
-	WREG32(mmMC_VM_AGP_TOP, 0x0FFFFFFF);
-	WREG32(mmMC_VM_AGP_BOT, 0x0FFFFFFF);
+	WREG32(mmMC_VM_AGP_TOP, adev->gmc.agp_end >> 22);
+	WREG32(mmMC_VM_AGP_BOT, adev->gmc.agp_start >> 22);
 
-	if (gmc_v6_0_wait_for_idle((void *)adev)) {
+	if (gmc_v6_0_wait_for_idle(ip_block))
 		dev_warn(adev->dev, "Wait for MC idle timedout !\n");
-	}
 }
 
 static int gmc_v6_0_mc_init(struct amdgpu_device *adev)
@@ -276,13 +275,13 @@ static int gmc_v6_0_mc_init(struct amdgpu_device *adev)
 	int r;
 
 	tmp = RREG32(mmMC_ARB_RAMCFG);
-	if (tmp & (1 << 11)) {
+	if (tmp & (1 << 11))
 		chansize = 16;
-	} else if (tmp & MC_ARB_RAMCFG__CHANSIZE_MASK) {
+	else if (tmp & MC_ARB_RAMCFG__CHANSIZE_MASK)
 		chansize = 64;
-	} else {
+	else
 		chansize = 32;
-	}
+
 	tmp = RREG32(mmMC_SHARED_CHMAP);
 	switch ((tmp & MC_SHARED_CHMAP__NOOFCHAN_MASK) >> MC_SHARED_CHMAP__NOOFCHAN__SHIFT) {
 	case 0:
@@ -359,7 +358,7 @@ static void gmc_v6_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid,
 }
 
 static uint64_t gmc_v6_0_emit_flush_gpu_tlb(struct amdgpu_ring *ring,
-					    unsigned vmid, uint64_t pd_addr)
+					    unsigned int vmid, uint64_t pd_addr)
 {
 	uint32_t reg;
 
@@ -383,7 +382,9 @@ static void gmc_v6_0_get_vm_pde(struct amdgpu_device *adev, int level,
 }
 
 static void gmc_v6_0_get_vm_pte(struct amdgpu_device *adev,
-				struct amdgpu_bo_va_mapping *mapping,
+				struct amdgpu_vm *vm,
+				struct amdgpu_bo *bo,
+				uint32_t vm_flags,
 				uint64_t *flags)
 {
 	*flags &= ~AMDGPU_PTE_EXECUTABLE;
@@ -412,11 +413,11 @@ static void gmc_v6_0_set_fault_enable_default(struct amdgpu_device *adev,
 }
 
  /**
-   + * gmc_v8_0_set_prt - set PRT VM fault
-   + *
-   + * @adev: amdgpu_device pointer
-   + * @enable: enable/disable VM fault handling for PRT
-   +*/
+  * gmc_v8_0_set_prt() - set PRT VM fault
+  *
+  * @adev: amdgpu_device pointer
+  * @enable: enable/disable VM fault handling for PRT
+  */
 static void gmc_v6_0_set_prt(struct amdgpu_device *adev, bool enable)
 {
 	u32 tmp;
@@ -442,9 +443,10 @@ static void gmc_v6_0_set_prt(struct amdgpu_device *adev, bool enable)
 	WREG32(mmVM_PRT_CNTL, tmp);
 
 	if (enable) {
-		uint32_t low = AMDGPU_VA_RESERVED_SIZE >> AMDGPU_GPU_PAGE_SHIFT;
+		uint32_t low = AMDGPU_VA_RESERVED_BOTTOM >>
+			AMDGPU_GPU_PAGE_SHIFT;
 		uint32_t high = adev->vm_manager.max_pfn -
-			(AMDGPU_VA_RESERVED_SIZE >> AMDGPU_GPU_PAGE_SHIFT);
+			(AMDGPU_VA_RESERVED_TOP >> AMDGPU_GPU_PAGE_SHIFT);
 
 		WREG32(mmVM_PRT_APERTURE0_LOW_ADDR, low);
 		WREG32(mmVM_PRT_APERTURE1_LOW_ADDR, low);
@@ -469,16 +471,14 @@ static void gmc_v6_0_set_prt(struct amdgpu_device *adev, bool enable)
 static int gmc_v6_0_gart_enable(struct amdgpu_device *adev)
 {
 	uint64_t table_addr;
-	int r, i;
 	u32 field;
+	int i;
 
 	if (adev->gart.bo == NULL) {
 		dev_err(adev->dev, "No VRAM object for PCIE GART.\n");
 		return -EINVAL;
 	}
-	r = amdgpu_gart_table_vram_pin(adev);
-	if (r)
-		return r;
+	amdgpu_gtt_mgr_recover(&adev->mman.gtt_mgr);
 
 	table_addr = amdgpu_bo_gpu_offset(adev->gart.bo);
 
@@ -556,9 +556,8 @@ static int gmc_v6_0_gart_enable(struct amdgpu_device *adev)
 
 	gmc_v6_0_flush_gpu_tlb(adev, 0, 0, 0);
 	dev_info(adev->dev, "PCIE GART of %uM enabled (table at 0x%016llX).\n",
-		 (unsigned)(adev->gmc.gart_size >> 20),
+		 (unsigned int)(adev->gmc.gart_size >> 20),
 		 (unsigned long long)table_addr);
-	adev->gart.ready = true;
 	return 0;
 }
 
@@ -608,40 +607,36 @@ static void gmc_v6_0_gart_disable(struct amdgpu_device *adev)
 	WREG32(mmVM_L2_CNTL3,
 	       VM_L2_CNTL3__L2_CACHE_BIGK_ASSOCIATIVITY_MASK |
 	       (0UL << VM_L2_CNTL3__L2_CACHE_BIGK_FRAGMENT_SIZE__SHIFT));
-	amdgpu_gart_table_vram_unpin(adev);
 }
 
 static void gmc_v6_0_vm_decode_fault(struct amdgpu_device *adev,
-				     u32 status, u32 addr, u32 mc_client)
+				     u32 status, u32 addr)
 {
 	u32 mc_id;
 	u32 vmid = REG_GET_FIELD(status, VM_CONTEXT1_PROTECTION_FAULT_STATUS, VMID);
 	u32 protections = REG_GET_FIELD(status, VM_CONTEXT1_PROTECTION_FAULT_STATUS,
 					PROTECTIONS);
-	char block[5] = { mc_client >> 24, (mc_client >> 16) & 0xff,
-		(mc_client >> 8) & 0xff, mc_client & 0xff, 0 };
 
 	mc_id = REG_GET_FIELD(status, VM_CONTEXT1_PROTECTION_FAULT_STATUS,
 			      MEMORY_CLIENT_ID);
 
-	dev_err(adev->dev, "VM fault (0x%02x, vmid %d) at page %u, %s from '%s' (0x%08x) (%d)\n",
+	dev_err(adev->dev, "VM fault (0x%02x, vmid %d) at page %u, %s from %d\n",
 	       protections, vmid, addr,
 	       REG_GET_FIELD(status, VM_CONTEXT1_PROTECTION_FAULT_STATUS,
 			     MEMORY_CLIENT_RW) ?
-	       "write" : "read", block, mc_client, mc_id);
+	       "write" : "read", mc_id);
 }
 
-/*
 static const u32 mc_cg_registers[] = {
-	MC_HUB_MISC_HUB_CG,
-	MC_HUB_MISC_SIP_CG,
-	MC_HUB_MISC_VM_CG,
-	MC_XPB_CLK_GAT,
-	ATC_MISC_CG,
-	MC_CITF_MISC_WR_CG,
-	MC_CITF_MISC_RD_CG,
-	MC_CITF_MISC_VM_CG,
-	VM_L2_CG,
+	mmMC_HUB_MISC_HUB_CG,
+	mmMC_HUB_MISC_SIP_CG,
+	mmMC_HUB_MISC_VM_CG,
+	mmMC_XPB_CLK_GAT,
+	mmATC_MISC_CG,
+	mmMC_CITF_MISC_WR_CG,
+	mmMC_CITF_MISC_RD_CG,
+	mmMC_CITF_MISC_VM_CG,
+	mmVM_L2_CG,
 };
 
 static const u32 mc_cg_ls_en[] = {
@@ -676,7 +671,7 @@ static void gmc_v6_0_enable_mc_ls(struct amdgpu_device *adev,
 
 	for (i = 0; i < ARRAY_SIZE(mc_cg_registers); i++) {
 		orig = data = RREG32(mc_cg_registers[i]);
-		if (enable && (adev->cg_flags & AMDGPU_CG_SUPPORT_MC_LS))
+		if (enable && (adev->cg_flags & AMD_CG_SUPPORT_MC_LS))
 			data |= mc_cg_ls_en[i];
 		else
 			data &= ~mc_cg_ls_en[i];
@@ -693,7 +688,7 @@ static void gmc_v6_0_enable_mc_mgcg(struct amdgpu_device *adev,
 
 	for (i = 0; i < ARRAY_SIZE(mc_cg_registers); i++) {
 		orig = data = RREG32(mc_cg_registers[i]);
-		if (enable && (adev->cg_flags & AMDGPU_CG_SUPPORT_MC_MGCG))
+		if (enable && (adev->cg_flags & AMD_CG_SUPPORT_MC_MGCG))
 			data |= mc_cg_en[i];
 		else
 			data &= ~mc_cg_en[i];
@@ -709,7 +704,7 @@ static void gmc_v6_0_enable_bif_mgls(struct amdgpu_device *adev,
 
 	orig = data = RREG32_PCIE(ixPCIE_CNTL2);
 
-	if (enable && (adev->cg_flags & AMDGPU_CG_SUPPORT_BIF_LS)) {
+	if (enable && (adev->cg_flags & AMD_CG_SUPPORT_BIF_LS)) {
 		data = REG_SET_FIELD(data, PCIE_CNTL2, SLV_MEM_LS_EN, 1);
 		data = REG_SET_FIELD(data, PCIE_CNTL2, MST_MEM_LS_EN, 1);
 		data = REG_SET_FIELD(data, PCIE_CNTL2, REPLAY_MEM_LS_EN, 1);
@@ -732,7 +727,7 @@ static void gmc_v6_0_enable_hdp_mgcg(struct amdgpu_device *adev,
 
 	orig = data = RREG32(mmHDP_HOST_PATH_CNTL);
 
-	if (enable && (adev->cg_flags & AMDGPU_CG_SUPPORT_HDP_MGCG))
+	if (enable && (adev->cg_flags & AMD_CG_SUPPORT_HDP_MGCG))
 		data = REG_SET_FIELD(data, HDP_HOST_PATH_CNTL, CLOCK_GATING_DIS, 0);
 	else
 		data = REG_SET_FIELD(data, HDP_HOST_PATH_CNTL, CLOCK_GATING_DIS, 1);
@@ -748,7 +743,7 @@ static void gmc_v6_0_enable_hdp_ls(struct amdgpu_device *adev,
 
 	orig = data = RREG32(mmHDP_MEM_POWER_LS);
 
-	if (enable && (adev->cg_flags & AMDGPU_CG_SUPPORT_HDP_LS))
+	if (enable && (adev->cg_flags & AMD_CG_SUPPORT_HDP_LS))
 		data = REG_SET_FIELD(data, HDP_MEM_POWER_LS, LS_ENABLE, 1);
 	else
 		data = REG_SET_FIELD(data, HDP_MEM_POWER_LS, LS_ENABLE, 0);
@@ -756,7 +751,6 @@ static void gmc_v6_0_enable_hdp_ls(struct amdgpu_device *adev,
 	if (orig != data)
 		WREG32(mmHDP_MEM_POWER_LS, data);
 }
-*/
 
 static int gmc_v6_0_convert_vram_type(int mc_seq_vram_type)
 {
@@ -778,9 +772,9 @@ static int gmc_v6_0_convert_vram_type(int mc_seq_vram_type)
 	}
 }
 
-static int gmc_v6_0_early_init(void *handle)
+static int gmc_v6_0_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	gmc_v6_0_set_gmc_funcs(adev);
 	gmc_v6_0_set_irq_funcs(adev);
@@ -788,9 +782,9 @@ static int gmc_v6_0_early_init(void *handle)
 	return 0;
 }
 
-static int gmc_v6_0_late_init(void *handle)
+static int gmc_v6_0_late_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (amdgpu_vm_fault_stop != AMDGPU_VM_FAULT_STOP_ALWAYS)
 		return amdgpu_irq_get(adev, &adev->gmc.vm_fault, 0);
@@ -798,15 +792,16 @@ static int gmc_v6_0_late_init(void *handle)
 		return 0;
 }
 
-static unsigned gmc_v6_0_get_vbios_fb_size(struct amdgpu_device *adev)
+static unsigned int gmc_v6_0_get_vbios_fb_size(struct amdgpu_device *adev)
 {
 	u32 d1vga_control = RREG32(mmD1VGA_CONTROL);
-	unsigned size;
+	unsigned int size;
 
 	if (REG_GET_FIELD(d1vga_control, D1VGA_CONTROL, D1VGA_MODE_ENABLE)) {
 		size = AMDGPU_VBIOS_VGA_ALLOCATION;
 	} else {
 		u32 viewport = RREG32(mmVIEWPORT_SIZE);
+
 		size = (REG_GET_FIELD(viewport, VIEWPORT_SIZE, VIEWPORT_HEIGHT) *
 			REG_GET_FIELD(viewport, VIEWPORT_SIZE, VIEWPORT_WIDTH) *
 			4);
@@ -814,17 +809,18 @@ static unsigned gmc_v6_0_get_vbios_fb_size(struct amdgpu_device *adev)
 	return size;
 }
 
-static int gmc_v6_0_sw_init(void *handle)
+static int gmc_v6_0_sw_init(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
-	adev->num_vmhubs = 1;
+	set_bit(AMDGPU_GFXHUB(0), adev->vmhubs_mask);
 
 	if (adev->flags & AMD_IS_APU) {
 		adev->gmc.vram_type = AMDGPU_VRAM_TYPE_UNKNOWN;
 	} else {
 		u32 tmp = RREG32(mmMC_SEQ_MISC0);
+
 		tmp &= MC_SEQ_MISC0__MT__MASK;
 		adev->gmc.vram_type = gmc_v6_0_convert_vram_type(tmp);
 	}
@@ -890,24 +886,23 @@ static int gmc_v6_0_sw_init(void *handle)
 	return 0;
 }
 
-static int gmc_v6_0_sw_fini(void *handle)
+static int gmc_v6_0_sw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	amdgpu_gem_force_release(adev);
 	amdgpu_vm_manager_fini(adev);
 	amdgpu_gart_table_vram_free(adev);
 	amdgpu_bo_fini(adev);
-	release_firmware(adev->gmc.fw);
-	adev->gmc.fw = NULL;
+	amdgpu_ucode_release(&adev->gmc.fw);
 
 	return 0;
 }
 
-static int gmc_v6_0_hw_init(void *handle)
+static int gmc_v6_0_hw_init(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	gmc_v6_0_mc_program(adev);
 
@@ -923,12 +918,15 @@ static int gmc_v6_0_hw_init(void *handle)
 	if (r)
 		return r;
 
-	return r;
+	if (amdgpu_emu_mode == 1)
+		return amdgpu_gmc_vram_checking(adev);
+
+	return 0;
 }
 
-static int gmc_v6_0_hw_fini(void *handle)
+static int gmc_v6_0_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	amdgpu_irq_put(adev, &adev->gmc.vm_fault, 0);
 	gmc_v6_0_gart_disable(adev);
@@ -936,21 +934,19 @@ static int gmc_v6_0_hw_fini(void *handle)
 	return 0;
 }
 
-static int gmc_v6_0_suspend(void *handle)
+static int gmc_v6_0_suspend(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	gmc_v6_0_hw_fini(adev);
+	gmc_v6_0_hw_fini(ip_block);
 
 	return 0;
 }
 
-static int gmc_v6_0_resume(void *handle)
+static int gmc_v6_0_resume(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
-	r = gmc_v6_0_hw_init(adev);
+	r = gmc_v6_0_hw_init(ip_block);
 	if (r)
 		return r;
 
@@ -959,9 +955,10 @@ static int gmc_v6_0_resume(void *handle)
 	return 0;
 }
 
-static bool gmc_v6_0_is_idle(void *handle)
+static bool gmc_v6_0_is_idle(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
+
 	u32 tmp = RREG32(mmSRBM_STATUS);
 
 	if (tmp & (SRBM_STATUS__MCB_BUSY_MASK | SRBM_STATUS__MCB_NON_DISPLAY_BUSY_MASK |
@@ -971,13 +968,13 @@ static bool gmc_v6_0_is_idle(void *handle)
 	return true;
 }
 
-static int gmc_v6_0_wait_for_idle(void *handle)
+static int gmc_v6_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
 {
-	unsigned i;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	unsigned int i;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	for (i = 0; i < adev->usec_timeout; i++) {
-		if (gmc_v6_0_is_idle(handle))
+		if (gmc_v6_0_is_idle(ip_block))
 			return 0;
 		udelay(1);
 	}
@@ -985,9 +982,10 @@ static int gmc_v6_0_wait_for_idle(void *handle)
 
 }
 
-static int gmc_v6_0_soft_reset(void *handle)
+static int gmc_v6_0_soft_reset(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
+
 	u32 srbm_soft_reset = 0;
 	u32 tmp = RREG32(mmSRBM_STATUS);
 
@@ -1004,10 +1002,9 @@ static int gmc_v6_0_soft_reset(void *handle)
 
 	if (srbm_soft_reset) {
 		gmc_v6_0_mc_stop(adev);
-		if (gmc_v6_0_wait_for_idle(adev)) {
-			dev_warn(adev->dev, "Wait for GMC idle timed out !\n");
-		}
 
+		if (gmc_v6_0_wait_for_idle(ip_block))
+			dev_warn(adev->dev, "Wait for GMC idle timed out !\n");
 
 		tmp = RREG32(mmSRBM_SOFT_RESET);
 		tmp |= srbm_soft_reset;
@@ -1032,7 +1029,7 @@ static int gmc_v6_0_soft_reset(void *handle)
 
 static int gmc_v6_0_vm_fault_interrupt_state(struct amdgpu_device *adev,
 					     struct amdgpu_irq_src *src,
-					     unsigned type,
+					     unsigned int type,
 					     enum amdgpu_interrupt_state state)
 {
 	u32 tmp;
@@ -1073,6 +1070,12 @@ static int gmc_v6_0_process_interrupt(struct amdgpu_device *adev,
 {
 	u32 addr, status;
 
+	/* Delegate to the soft IRQ handler ring */
+	if (adev->irq.ih_soft.enabled && entry->ih != &adev->irq.ih_soft) {
+		amdgpu_irq_delegate(adev, entry, 4);
+		return 1;
+	}
+
 	addr = RREG32(mmVM_CONTEXT1_PROTECTION_FAULT_ADDR);
 	status = RREG32(mmVM_CONTEXT1_PROTECTION_FAULT_STATUS);
 	WREG32_P(mmVM_CONTEXT1_CNTL2, 1, ~1);
@@ -1080,6 +1083,10 @@ static int gmc_v6_0_process_interrupt(struct amdgpu_device *adev,
 	if (!addr && !status)
 		return 0;
 
+	amdgpu_vm_update_fault_cache(adev, entry->pasid,
+				     ((u64)addr) << AMDGPU_GPU_PAGE_SHIFT,
+				     status, AMDGPU_GFXHUB(0));
+
 	if (amdgpu_vm_fault_stop == AMDGPU_VM_FAULT_STOP_FIRST)
 		gmc_v6_0_set_fault_enable_default(adev, false);
 
@@ -1090,19 +1097,33 @@ static int gmc_v6_0_process_interrupt(struct amdgpu_device *adev,
 			addr);
 		dev_err(adev->dev, "  VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x%08X\n",
 			status);
-		gmc_v6_0_vm_decode_fault(adev, status, addr, 0);
+		gmc_v6_0_vm_decode_fault(adev, status, addr);
 	}
 
 	return 0;
 }
 
-static int gmc_v6_0_set_clockgating_state(void *handle,
+static int gmc_v6_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_clockgating_state state)
 {
+	struct amdgpu_device *adev = ip_block->adev;
+	bool gate = false;
+
+	if (state == AMD_CG_STATE_GATE)
+		gate = true;
+
+	if (!(adev->flags & AMD_IS_APU)) {
+		gmc_v6_0_enable_mc_mgcg(adev, gate);
+		gmc_v6_0_enable_mc_ls(adev, gate);
+	}
+	gmc_v6_0_enable_bif_mgls(adev, gate);
+	gmc_v6_0_enable_hdp_mgcg(adev, gate);
+	gmc_v6_0_enable_hdp_ls(adev, gate);
+
 	return 0;
 }
 
-static int gmc_v6_0_set_powergating_state(void *handle,
+static int gmc_v6_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_powergating_state state)
 {
 	return 0;
@@ -1150,8 +1171,7 @@ static void gmc_v6_0_set_irq_funcs(struct amdgpu_device *adev)
 	adev->gmc.vm_fault.funcs = &gmc_v6_0_irq_funcs;
 }
 
-const struct amdgpu_ip_block_version gmc_v6_0_ip_block =
-{
+const struct amdgpu_ip_block_version gmc_v6_0_ip_block = {
 	.type = AMD_IP_BLOCK_TYPE_GMC,
 	.major = 6,
 	.minor = 0,
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c
index 0a50fdaced7e..fbd0bf147f50 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c
@@ -52,22 +52,20 @@
 
 static void gmc_v7_0_set_gmc_funcs(struct amdgpu_device *adev);
 static void gmc_v7_0_set_irq_funcs(struct amdgpu_device *adev);
-static int gmc_v7_0_wait_for_idle(void *handle);
+static int gmc_v7_0_wait_for_idle(struct amdgpu_ip_block *ip_block);
 
 MODULE_FIRMWARE("amdgpu/bonaire_mc.bin");
 MODULE_FIRMWARE("amdgpu/hawaii_mc.bin");
 MODULE_FIRMWARE("amdgpu/topaz_mc.bin");
 
-static const u32 golden_settings_iceland_a11[] =
-{
+static const u32 golden_settings_iceland_a11[] = {
 	mmVM_PRT_APERTURE0_LOW_ADDR, 0x0fffffff, 0x0fffffff,
 	mmVM_PRT_APERTURE1_LOW_ADDR, 0x0fffffff, 0x0fffffff,
 	mmVM_PRT_APERTURE2_LOW_ADDR, 0x0fffffff, 0x0fffffff,
 	mmVM_PRT_APERTURE3_LOW_ADDR, 0x0fffffff, 0x0fffffff
 };
 
-static const u32 iceland_mgcg_cgcg_init[] =
-{
+static const u32 iceland_mgcg_cgcg_init[] = {
 	mmMC_MEM_POWER_LS, 0xffffffff, 0x00000104
 };
 
@@ -89,9 +87,14 @@ static void gmc_v7_0_init_golden_registers(struct amdgpu_device *adev)
 
 static void gmc_v7_0_mc_stop(struct amdgpu_device *adev)
 {
+	struct amdgpu_ip_block *ip_block;
 	u32 blackout;
 
-	gmc_v7_0_wait_for_idle((void *)adev);
+	ip_block = amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_GMC);
+	if (!ip_block)
+		return;
+
+	gmc_v7_0_wait_for_idle(ip_block);
 
 	blackout = RREG32(mmMC_SHARED_BLACKOUT_CNTL);
 	if (REG_GET_FIELD(blackout, MC_SHARED_BLACKOUT_CNTL, BLACKOUT_MODE) != 1) {
@@ -132,7 +135,6 @@ static void gmc_v7_0_mc_resume(struct amdgpu_device *adev)
 static int gmc_v7_0_init_microcode(struct amdgpu_device *adev)
 {
 	const char *chip_name;
-	char fw_name[30];
 	int err;
 
 	DRM_DEBUG("\n");
@@ -151,21 +153,15 @@ static int gmc_v7_0_init_microcode(struct amdgpu_device *adev)
 	case CHIP_KABINI:
 	case CHIP_MULLINS:
 		return 0;
-	default: BUG();
+	default:
+		return -EINVAL;
 	}
 
-	snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_mc.bin", chip_name);
-
-	err = request_firmware(&adev->gmc.fw, fw_name, adev->dev);
-	if (err)
-		goto out;
-	err = amdgpu_ucode_validate(adev->gmc.fw);
-
-out:
+	err = amdgpu_ucode_request(adev, &adev->gmc.fw, AMDGPU_UCODE_REQUIRED,
+				   "amdgpu/%s_mc.bin", chip_name);
 	if (err) {
-		pr_err("cik_mc: Failed to load firmware \"%s\"\n", fw_name);
-		release_firmware(adev->gmc.fw);
-		adev->gmc.fw = NULL;
+		pr_err("cik_mc: Failed to load firmware \"%s_mc.bin\"\n", chip_name);
+		amdgpu_ucode_release(&adev->gmc.fw);
 	}
 	return err;
 }
@@ -243,10 +239,12 @@ static void gmc_v7_0_vram_gtt_location(struct amdgpu_device *adev,
 				       struct amdgpu_gmc *mc)
 {
 	u64 base = RREG32(mmMC_VM_FB_LOCATION) & 0xFFFF;
+
 	base <<= 24;
 
+	amdgpu_gmc_set_agp_default(adev, mc);
 	amdgpu_gmc_vram_location(adev, mc, base);
-	amdgpu_gmc_gart_location(adev, mc);
+	amdgpu_gmc_gart_location(adev, mc, AMDGPU_GART_PLACEMENT_BEST_FIT);
 }
 
 /**
@@ -259,9 +257,14 @@ static void gmc_v7_0_vram_gtt_location(struct amdgpu_device *adev,
  */
 static void gmc_v7_0_mc_program(struct amdgpu_device *adev)
 {
+	struct amdgpu_ip_block *ip_block;
 	u32 tmp;
 	int i, j;
 
+	ip_block = amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_GMC);
+	if (!ip_block)
+		return;
+
 	/* Initialize HDP */
 	for (i = 0, j = 0; i < 32; i++, j += 0x6) {
 		WREG32((0xb05 + j), 0x00000000);
@@ -272,9 +275,9 @@ static void gmc_v7_0_mc_program(struct amdgpu_device *adev)
 	}
 	WREG32(mmHDP_REG_COHERENCY_FLUSH_CNTL, 0);
 
-	if (gmc_v7_0_wait_for_idle((void *)adev)) {
+	if (gmc_v7_0_wait_for_idle(ip_block))
 		dev_warn(adev->dev, "Wait for MC idle timedout !\n");
-	}
+
 	if (adev->mode_info.num_crtc) {
 		/* Lockout access through VGA aperture*/
 		tmp = RREG32(mmVGA_HDP_CONTROL);
@@ -292,13 +295,12 @@ static void gmc_v7_0_mc_program(struct amdgpu_device *adev)
 	WREG32(mmMC_VM_SYSTEM_APERTURE_HIGH_ADDR,
 	       adev->gmc.vram_end >> 12);
 	WREG32(mmMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR,
-	       adev->vram_scratch.gpu_addr >> 12);
+	       adev->mem_scratch.gpu_addr >> 12);
 	WREG32(mmMC_VM_AGP_BASE, 0);
-	WREG32(mmMC_VM_AGP_TOP, 0x0FFFFFFF);
-	WREG32(mmMC_VM_AGP_BOT, 0x0FFFFFFF);
-	if (gmc_v7_0_wait_for_idle((void *)adev)) {
+	WREG32(mmMC_VM_AGP_TOP, adev->gmc.agp_end >> 22);
+	WREG32(mmMC_VM_AGP_BOT, adev->gmc.agp_start >> 22);
+	if (gmc_v7_0_wait_for_idle(ip_block))
 		dev_warn(adev->dev, "Wait for MC idle timedout !\n");
-	}
 
 	WREG32(mmBIF_FB_EN, BIF_FB_EN__FB_READ_EN_MASK | BIF_FB_EN__FB_WRITE_EN_MASK);
 
@@ -330,11 +332,11 @@ static int gmc_v7_0_mc_init(struct amdgpu_device *adev)
 
 		/* Get VRAM informations */
 		tmp = RREG32(mmMC_ARB_RAMCFG);
-		if (REG_GET_FIELD(tmp, MC_ARB_RAMCFG, CHANSIZE)) {
+		if (REG_GET_FIELD(tmp, MC_ARB_RAMCFG, CHANSIZE))
 			chansize = 64;
-		} else {
+		else
 			chansize = 32;
-		}
+
 		tmp = RREG32(mmMC_SHARED_CHMAP);
 		switch (REG_GET_FIELD(tmp, MC_SHARED_CHMAP, NOOFCHAN)) {
 		case 0:
@@ -381,17 +383,15 @@ static int gmc_v7_0_mc_init(struct amdgpu_device *adev)
 	adev->gmc.aper_size = pci_resource_len(adev->pdev, 0);
 
 #ifdef CONFIG_X86_64
-	if (adev->flags & AMD_IS_APU &&
-	    adev->gmc.real_vram_size > adev->gmc.aper_size) {
+	if ((adev->flags & AMD_IS_APU) &&
+	    adev->gmc.real_vram_size > adev->gmc.aper_size &&
+	    !amdgpu_passthrough(adev)) {
 		adev->gmc.aper_base = ((u64)RREG32(mmMC_VM_FB_OFFSET)) << 22;
 		adev->gmc.aper_size = adev->gmc.real_vram_size;
 	}
 #endif
 
-	/* In case the PCI BAR is larger than the actual amount of vram */
 	adev->gmc.visible_vram_size = adev->gmc.aper_size;
-	if (adev->gmc.visible_vram_size > adev->gmc.real_vram_size)
-		adev->gmc.visible_vram_size = adev->gmc.real_vram_size;
 
 	/* set the gart size */
 	if (amdgpu_gart_size == -1) {
@@ -427,31 +427,27 @@ static int gmc_v7_0_mc_init(struct amdgpu_device *adev)
  * @pasid: pasid to be flush
  * @flush_type: type of flush
  * @all_hub: flush all hubs
+ * @inst: is used to select which instance of KIQ to use for the invalidation
  *
  * Flush the TLB for the requested pasid.
  */
-static int gmc_v7_0_flush_gpu_tlb_pasid(struct amdgpu_device *adev,
-					uint16_t pasid, uint32_t flush_type,
-					bool all_hub)
+static void gmc_v7_0_flush_gpu_tlb_pasid(struct amdgpu_device *adev,
+					 uint16_t pasid, uint32_t flush_type,
+					 bool all_hub, uint32_t inst)
 {
+	u32 mask = 0x0;
 	int vmid;
-	unsigned int tmp;
-
-	if (amdgpu_in_reset(adev))
-		return -EIO;
 
 	for (vmid = 1; vmid < 16; vmid++) {
+		u32 tmp = RREG32(mmATC_VMID0_PASID_MAPPING + vmid);
 
-		tmp = RREG32(mmATC_VMID0_PASID_MAPPING + vmid);
 		if ((tmp & ATC_VMID0_PASID_MAPPING__VALID_MASK) &&
-			(tmp & ATC_VMID0_PASID_MAPPING__PASID_MASK) == pasid) {
-			WREG32(mmVM_INVALIDATE_REQUEST, 1 << vmid);
-			RREG32(mmVM_INVALIDATE_RESPONSE);
-			break;
-		}
+		    (tmp & ATC_VMID0_PASID_MAPPING__PASID_MASK) == pasid)
+			mask |= 1 << vmid;
 	}
 
-	return 0;
+	WREG32(mmVM_INVALIDATE_REQUEST, mask);
+	RREG32(mmVM_INVALIDATE_RESPONSE);
 }
 
 /*
@@ -479,7 +475,7 @@ static void gmc_v7_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid,
 }
 
 static uint64_t gmc_v7_0_emit_flush_gpu_tlb(struct amdgpu_ring *ring,
-					    unsigned vmid, uint64_t pd_addr)
+					    unsigned int vmid, uint64_t pd_addr)
 {
 	uint32_t reg;
 
@@ -495,8 +491,8 @@ static uint64_t gmc_v7_0_emit_flush_gpu_tlb(struct amdgpu_ring *ring,
 	return pd_addr;
 }
 
-static void gmc_v7_0_emit_pasid_mapping(struct amdgpu_ring *ring, unsigned vmid,
-					unsigned pasid)
+static void gmc_v7_0_emit_pasid_mapping(struct amdgpu_ring *ring, unsigned int vmid,
+					unsigned int pasid)
 {
 	amdgpu_ring_emit_wreg(ring, mmIH_VMID_0_LUT + vmid, pasid);
 }
@@ -508,7 +504,9 @@ static void gmc_v7_0_get_vm_pde(struct amdgpu_device *adev, int level,
 }
 
 static void gmc_v7_0_get_vm_pte(struct amdgpu_device *adev,
-				struct amdgpu_bo_va_mapping *mapping,
+				struct amdgpu_vm *vm,
+				struct amdgpu_bo *bo,
+				uint32_t vm_flags,
 				uint64_t *flags)
 {
 	*flags &= ~AMDGPU_PTE_EXECUTABLE;
@@ -575,9 +573,10 @@ static void gmc_v7_0_set_prt(struct amdgpu_device *adev, bool enable)
 	WREG32(mmVM_PRT_CNTL, tmp);
 
 	if (enable) {
-		uint32_t low = AMDGPU_VA_RESERVED_SIZE >> AMDGPU_GPU_PAGE_SHIFT;
+		uint32_t low = AMDGPU_VA_RESERVED_BOTTOM >>
+			AMDGPU_GPU_PAGE_SHIFT;
 		uint32_t high = adev->vm_manager.max_pfn -
-			(AMDGPU_VA_RESERVED_SIZE >> AMDGPU_GPU_PAGE_SHIFT);
+			(AMDGPU_VA_RESERVED_TOP >> AMDGPU_GPU_PAGE_SHIFT);
 
 		WREG32(mmVM_PRT_APERTURE0_LOW_ADDR, low);
 		WREG32(mmVM_PRT_APERTURE1_LOW_ADDR, low);
@@ -613,17 +612,14 @@ static void gmc_v7_0_set_prt(struct amdgpu_device *adev, bool enable)
 static int gmc_v7_0_gart_enable(struct amdgpu_device *adev)
 {
 	uint64_t table_addr;
-	int r, i;
 	u32 tmp, field;
+	int i;
 
 	if (adev->gart.bo == NULL) {
 		dev_err(adev->dev, "No VRAM object for PCIE GART.\n");
 		return -EINVAL;
 	}
-	r = amdgpu_gart_table_vram_pin(adev);
-	if (r)
-		return r;
-
+	amdgpu_gtt_mgr_recover(&adev->mman.gtt_mgr);
 	table_addr = amdgpu_bo_gpu_offset(adev->gart.bo);
 
 	/* Setup TLB control */
@@ -710,9 +706,8 @@ static int gmc_v7_0_gart_enable(struct amdgpu_device *adev)
 
 	gmc_v7_0_flush_gpu_tlb(adev, 0, 0, 0);
 	DRM_INFO("PCIE GART of %uM enabled (table at 0x%016llX).\n",
-		 (unsigned)(adev->gmc.gart_size >> 20),
+		 (unsigned int)(adev->gmc.gart_size >> 20),
 		 (unsigned long long)table_addr);
-	adev->gart.ready = true;
 	return 0;
 }
 
@@ -758,7 +753,6 @@ static void gmc_v7_0_gart_disable(struct amdgpu_device *adev)
 	tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, ENABLE_L2_CACHE, 0);
 	WREG32(mmVM_L2_CNTL, tmp);
 	WREG32(mmVM_L2_CNTL2, 0);
-	amdgpu_gart_table_vram_unpin(adev);
 }
 
 /**
@@ -773,7 +767,7 @@ static void gmc_v7_0_gart_disable(struct amdgpu_device *adev)
  * Print human readable fault information (CIK).
  */
 static void gmc_v7_0_vm_decode_fault(struct amdgpu_device *adev, u32 status,
-				     u32 addr, u32 mc_client, unsigned pasid)
+				     u32 addr, u32 mc_client, unsigned int pasid)
 {
 	u32 vmid = REG_GET_FIELD(status, VM_CONTEXT1_PROTECTION_FAULT_STATUS, VMID);
 	u32 protections = REG_GET_FIELD(status, VM_CONTEXT1_PROTECTION_FAULT_STATUS,
@@ -940,9 +934,9 @@ static int gmc_v7_0_convert_vram_type(int mc_seq_vram_type)
 	}
 }
 
-static int gmc_v7_0_early_init(void *handle)
+static int gmc_v7_0_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	gmc_v7_0_set_gmc_funcs(adev);
 	gmc_v7_0_set_irq_funcs(adev);
@@ -954,13 +948,14 @@ static int gmc_v7_0_early_init(void *handle)
 		adev->gmc.shared_aperture_end + 1;
 	adev->gmc.private_aperture_end =
 		adev->gmc.private_aperture_start + (4ULL << 30) - 1;
+	adev->gmc.noretry_flags = AMDGPU_VM_NORETRY_FLAGS_TF;
 
 	return 0;
 }
 
-static int gmc_v7_0_late_init(void *handle)
+static int gmc_v7_0_late_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (amdgpu_vm_fault_stop != AMDGPU_VM_FAULT_STOP_ALWAYS)
 		return amdgpu_irq_get(adev, &adev->gmc.vm_fault, 0);
@@ -968,15 +963,16 @@ static int gmc_v7_0_late_init(void *handle)
 		return 0;
 }
 
-static unsigned gmc_v7_0_get_vbios_fb_size(struct amdgpu_device *adev)
+static unsigned int gmc_v7_0_get_vbios_fb_size(struct amdgpu_device *adev)
 {
 	u32 d1vga_control = RREG32(mmD1VGA_CONTROL);
-	unsigned size;
+	unsigned int size;
 
 	if (REG_GET_FIELD(d1vga_control, D1VGA_CONTROL, D1VGA_MODE_ENABLE)) {
 		size = AMDGPU_VBIOS_VGA_ALLOCATION;
 	} else {
 		u32 viewport = RREG32(mmVIEWPORT_SIZE);
+
 		size = (REG_GET_FIELD(viewport, VIEWPORT_SIZE, VIEWPORT_HEIGHT) *
 			REG_GET_FIELD(viewport, VIEWPORT_SIZE, VIEWPORT_WIDTH) *
 			4);
@@ -985,17 +981,18 @@ static unsigned gmc_v7_0_get_vbios_fb_size(struct amdgpu_device *adev)
 	return size;
 }
 
-static int gmc_v7_0_sw_init(void *handle)
+static int gmc_v7_0_sw_init(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
-	adev->num_vmhubs = 1;
+	set_bit(AMDGPU_GFXHUB(0), adev->vmhubs_mask);
 
 	if (adev->flags & AMD_IS_APU) {
 		adev->gmc.vram_type = AMDGPU_VRAM_TYPE_UNKNOWN;
 	} else {
 		u32 tmp = RREG32(mmMC_SEQ_MISC0);
+
 		tmp &= MC_SEQ_MISC0__MT__MASK;
 		adev->gmc.vram_type = gmc_v7_0_convert_vram_type(tmp);
 	}
@@ -1071,30 +1068,29 @@ static int gmc_v7_0_sw_init(void *handle)
 					GFP_KERNEL);
 	if (!adev->gmc.vm_fault_info)
 		return -ENOMEM;
-	atomic_set(&adev->gmc.vm_fault_info_updated, 0);
+	atomic_set_release(&adev->gmc.vm_fault_info_updated, 0);
 
 	return 0;
 }
 
-static int gmc_v7_0_sw_fini(void *handle)
+static int gmc_v7_0_sw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	amdgpu_gem_force_release(adev);
 	amdgpu_vm_manager_fini(adev);
 	kfree(adev->gmc.vm_fault_info);
 	amdgpu_gart_table_vram_free(adev);
 	amdgpu_bo_fini(adev);
-	release_firmware(adev->gmc.fw);
-	adev->gmc.fw = NULL;
+	amdgpu_ucode_release(&adev->gmc.fw);
 
 	return 0;
 }
 
-static int gmc_v7_0_hw_init(void *handle)
+static int gmc_v7_0_hw_init(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	gmc_v7_0_init_golden_registers(adev);
 
@@ -1112,12 +1108,15 @@ static int gmc_v7_0_hw_init(void *handle)
 	if (r)
 		return r;
 
-	return r;
+	if (amdgpu_emu_mode == 1)
+		return amdgpu_gmc_vram_checking(adev);
+
+	return 0;
 }
 
-static int gmc_v7_0_hw_fini(void *handle)
+static int gmc_v7_0_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	amdgpu_irq_put(adev, &adev->gmc.vm_fault, 0);
 	gmc_v7_0_gart_disable(adev);
@@ -1125,32 +1124,29 @@ static int gmc_v7_0_hw_fini(void *handle)
 	return 0;
 }
 
-static int gmc_v7_0_suspend(void *handle)
+static int gmc_v7_0_suspend(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	gmc_v7_0_hw_fini(adev);
+	gmc_v7_0_hw_fini(ip_block);
 
 	return 0;
 }
 
-static int gmc_v7_0_resume(void *handle)
+static int gmc_v7_0_resume(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
-	r = gmc_v7_0_hw_init(adev);
+	r = gmc_v7_0_hw_init(ip_block);
 	if (r)
 		return r;
 
-	amdgpu_vmid_reset_all(adev);
+	amdgpu_vmid_reset_all(ip_block->adev);
 
 	return 0;
 }
 
-static bool gmc_v7_0_is_idle(void *handle)
+static bool gmc_v7_0_is_idle(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	u32 tmp = RREG32(mmSRBM_STATUS);
 
 	if (tmp & (SRBM_STATUS__MCB_BUSY_MASK | SRBM_STATUS__MCB_NON_DISPLAY_BUSY_MASK |
@@ -1160,20 +1156,13 @@ static bool gmc_v7_0_is_idle(void *handle)
 	return true;
 }
 
-static int gmc_v7_0_wait_for_idle(void *handle)
+static int gmc_v7_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
 {
-	unsigned i;
-	u32 tmp;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	unsigned int i;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	for (i = 0; i < adev->usec_timeout; i++) {
-		/* read MC_STATUS */
-		tmp = RREG32(mmSRBM_STATUS) & (SRBM_STATUS__MCB_BUSY_MASK |
-					       SRBM_STATUS__MCB_NON_DISPLAY_BUSY_MASK |
-					       SRBM_STATUS__MCC_BUSY_MASK |
-					       SRBM_STATUS__MCD_BUSY_MASK |
-					       SRBM_STATUS__VMC_BUSY_MASK);
-		if (!tmp)
+		if (gmc_v7_0_is_idle(ip_block))
 			return 0;
 		udelay(1);
 	}
@@ -1181,9 +1170,9 @@ static int gmc_v7_0_wait_for_idle(void *handle)
 
 }
 
-static int gmc_v7_0_soft_reset(void *handle)
+static int gmc_v7_0_soft_reset(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	u32 srbm_soft_reset = 0;
 	u32 tmp = RREG32(mmSRBM_STATUS);
 
@@ -1200,10 +1189,8 @@ static int gmc_v7_0_soft_reset(void *handle)
 
 	if (srbm_soft_reset) {
 		gmc_v7_0_mc_stop(adev);
-		if (gmc_v7_0_wait_for_idle((void *)adev)) {
+		if (gmc_v7_0_wait_for_idle(ip_block))
 			dev_warn(adev->dev, "Wait for GMC idle timed out !\n");
-		}
-
 
 		tmp = RREG32(mmSRBM_SOFT_RESET);
 		tmp |= srbm_soft_reset;
@@ -1229,7 +1216,7 @@ static int gmc_v7_0_soft_reset(void *handle)
 
 static int gmc_v7_0_vm_fault_interrupt_state(struct amdgpu_device *adev,
 					     struct amdgpu_irq_src *src,
-					     unsigned type,
+					     unsigned int type,
 					     enum amdgpu_interrupt_state state)
 {
 	u32 tmp;
@@ -1274,6 +1261,12 @@ static int gmc_v7_0_process_interrupt(struct amdgpu_device *adev,
 {
 	u32 addr, status, mc_client, vmid;
 
+	/* Delegate to the soft IRQ handler ring */
+	if (adev->irq.ih_soft.enabled && entry->ih != &adev->irq.ih_soft) {
+		amdgpu_irq_delegate(adev, entry, 4);
+		return 1;
+	}
+
 	addr = RREG32(mmVM_CONTEXT1_PROTECTION_FAULT_ADDR);
 	status = RREG32(mmVM_CONTEXT1_PROTECTION_FAULT_STATUS);
 	mc_client = RREG32(mmVM_CONTEXT1_PROTECTION_FAULT_MCCLIENT);
@@ -1283,6 +1276,9 @@ static int gmc_v7_0_process_interrupt(struct amdgpu_device *adev,
 	if (!addr && !status)
 		return 0;
 
+	amdgpu_vm_update_fault_cache(adev, entry->pasid,
+				     ((u64)addr) << AMDGPU_GPU_PAGE_SHIFT, status, AMDGPU_GFXHUB(0));
+
 	if (amdgpu_vm_fault_stop == AMDGPU_VM_FAULT_STOP_FIRST)
 		gmc_v7_0_set_fault_enable_default(adev, false);
 
@@ -1300,7 +1296,7 @@ static int gmc_v7_0_process_interrupt(struct amdgpu_device *adev,
 	vmid = REG_GET_FIELD(status, VM_CONTEXT1_PROTECTION_FAULT_STATUS,
 			     VMID);
 	if (amdgpu_amdkfd_is_kfd_vmid(adev, vmid)
-		&& !atomic_read(&adev->gmc.vm_fault_info_updated)) {
+		&& !atomic_read_acquire(&adev->gmc.vm_fault_info_updated)) {
 		struct kfd_vm_fault_info *info = adev->gmc.vm_fault_info;
 		u32 protections = REG_GET_FIELD(status,
 					VM_CONTEXT1_PROTECTION_FAULT_STATUS,
@@ -1316,18 +1312,17 @@ static int gmc_v7_0_process_interrupt(struct amdgpu_device *adev,
 		info->prot_read = protections & 0x8 ? true : false;
 		info->prot_write = protections & 0x10 ? true : false;
 		info->prot_exec = protections & 0x20 ? true : false;
-		mb();
-		atomic_set(&adev->gmc.vm_fault_info_updated, 1);
+		atomic_set_release(&adev->gmc.vm_fault_info_updated, 1);
 	}
 
 	return 0;
 }
 
-static int gmc_v7_0_set_clockgating_state(void *handle,
+static int gmc_v7_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_clockgating_state state)
 {
 	bool gate = false;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (state == AMD_CG_STATE_GATE)
 		gate = true;
@@ -1343,7 +1338,7 @@ static int gmc_v7_0_set_clockgating_state(void *handle,
 	return 0;
 }
 
-static int gmc_v7_0_set_powergating_state(void *handle,
+static int gmc_v7_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_powergating_state state)
 {
 	return 0;
@@ -1393,8 +1388,7 @@ static void gmc_v7_0_set_irq_funcs(struct amdgpu_device *adev)
 	adev->gmc.vm_fault.funcs = &gmc_v7_0_irq_funcs;
 }
 
-const struct amdgpu_ip_block_version gmc_v7_0_ip_block =
-{
+const struct amdgpu_ip_block_version gmc_v7_0_ip_block = {
 	.type = AMD_IP_BLOCK_TYPE_GMC,
 	.major = 7,
 	.minor = 0,
@@ -1402,8 +1396,7 @@ const struct amdgpu_ip_block_version gmc_v7_0_ip_block =
 	.funcs = &gmc_v7_0_ip_funcs,
 };
 
-const struct amdgpu_ip_block_version gmc_v7_4_ip_block =
-{
+const struct amdgpu_ip_block_version gmc_v7_4_ip_block = {
 	.type = AMD_IP_BLOCK_TYPE_GMC,
 	.major = 7,
 	.minor = 4,
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c
index 492ebed2915b..6551b60f2584 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c
@@ -53,7 +53,7 @@
 
 static void gmc_v8_0_set_gmc_funcs(struct amdgpu_device *adev);
 static void gmc_v8_0_set_irq_funcs(struct amdgpu_device *adev);
-static int gmc_v8_0_wait_for_idle(void *handle);
+static int gmc_v8_0_wait_for_idle(struct amdgpu_ip_block *ip_block);
 
 MODULE_FIRMWARE("amdgpu/tonga_mc.bin");
 MODULE_FIRMWARE("amdgpu/polaris11_mc.bin");
@@ -64,8 +64,7 @@ MODULE_FIRMWARE("amdgpu/polaris11_k_mc.bin");
 MODULE_FIRMWARE("amdgpu/polaris10_k_mc.bin");
 MODULE_FIRMWARE("amdgpu/polaris12_k_mc.bin");
 
-static const u32 golden_settings_tonga_a11[] =
-{
+static const u32 golden_settings_tonga_a11[] = {
 	mmMC_ARB_WTM_GRPWT_RD, 0x00000003, 0x00000000,
 	mmMC_HUB_RDREQ_DMIF_LIMIT, 0x0000007f, 0x00000028,
 	mmMC_HUB_WDP_UMC, 0x00007fb6, 0x00000991,
@@ -75,34 +74,29 @@ static const u32 golden_settings_tonga_a11[] =
 	mmVM_PRT_APERTURE3_LOW_ADDR, 0x0fffffff, 0x0fffffff,
 };
 
-static const u32 tonga_mgcg_cgcg_init[] =
-{
+static const u32 tonga_mgcg_cgcg_init[] = {
 	mmMC_MEM_POWER_LS, 0xffffffff, 0x00000104
 };
 
-static const u32 golden_settings_fiji_a10[] =
-{
+static const u32 golden_settings_fiji_a10[] = {
 	mmVM_PRT_APERTURE0_LOW_ADDR, 0x0fffffff, 0x0fffffff,
 	mmVM_PRT_APERTURE1_LOW_ADDR, 0x0fffffff, 0x0fffffff,
 	mmVM_PRT_APERTURE2_LOW_ADDR, 0x0fffffff, 0x0fffffff,
 	mmVM_PRT_APERTURE3_LOW_ADDR, 0x0fffffff, 0x0fffffff,
 };
 
-static const u32 fiji_mgcg_cgcg_init[] =
-{
+static const u32 fiji_mgcg_cgcg_init[] = {
 	mmMC_MEM_POWER_LS, 0xffffffff, 0x00000104
 };
 
-static const u32 golden_settings_polaris11_a11[] =
-{
+static const u32 golden_settings_polaris11_a11[] = {
 	mmVM_PRT_APERTURE0_LOW_ADDR, 0x0fffffff, 0x0fffffff,
 	mmVM_PRT_APERTURE1_LOW_ADDR, 0x0fffffff, 0x0fffffff,
 	mmVM_PRT_APERTURE2_LOW_ADDR, 0x0fffffff, 0x0fffffff,
 	mmVM_PRT_APERTURE3_LOW_ADDR, 0x0fffffff, 0x0fffffff
 };
 
-static const u32 golden_settings_polaris10_a11[] =
-{
+static const u32 golden_settings_polaris10_a11[] = {
 	mmMC_ARB_WTM_GRPWT_RD, 0x00000003, 0x00000000,
 	mmVM_PRT_APERTURE0_LOW_ADDR, 0x0fffffff, 0x0fffffff,
 	mmVM_PRT_APERTURE1_LOW_ADDR, 0x0fffffff, 0x0fffffff,
@@ -110,19 +104,16 @@ static const u32 golden_settings_polaris10_a11[] =
 	mmVM_PRT_APERTURE3_LOW_ADDR, 0x0fffffff, 0x0fffffff
 };
 
-static const u32 cz_mgcg_cgcg_init[] =
-{
+static const u32 cz_mgcg_cgcg_init[] = {
 	mmMC_MEM_POWER_LS, 0xffffffff, 0x00000104
 };
 
-static const u32 stoney_mgcg_cgcg_init[] =
-{
+static const u32 stoney_mgcg_cgcg_init[] = {
 	mmATC_MISC_CG, 0xffffffff, 0x000c0200,
 	mmMC_MEM_POWER_LS, 0xffffffff, 0x00000104
 };
 
-static const u32 golden_settings_stoney_common[] =
-{
+static const u32 golden_settings_stoney_common[] = {
 	mmMC_HUB_RDREQ_UVD, MC_HUB_RDREQ_UVD__PRESCALE_MASK, 0x00000004,
 	mmMC_RD_GRP_OTH, MC_RD_GRP_OTH__UVD_MASK, 0x00600000
 };
@@ -179,8 +170,13 @@ static void gmc_v8_0_init_golden_registers(struct amdgpu_device *adev)
 static void gmc_v8_0_mc_stop(struct amdgpu_device *adev)
 {
 	u32 blackout;
+	struct amdgpu_ip_block *ip_block;
 
-	gmc_v8_0_wait_for_idle(adev);
+	ip_block = amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_GMC);
+	if (!ip_block)
+		return;
+
+	gmc_v8_0_wait_for_idle(ip_block);
 
 	blackout = RREG32(mmMC_SHARED_BLACKOUT_CNTL);
 	if (REG_GET_FIELD(blackout, MC_SHARED_BLACKOUT_CNTL, BLACKOUT_MODE) != 1) {
@@ -221,7 +217,6 @@ static void gmc_v8_0_mc_resume(struct amdgpu_device *adev)
 static int gmc_v8_0_init_microcode(struct amdgpu_device *adev)
 {
 	const char *chip_name;
-	char fw_name[30];
 	int err;
 
 	DRM_DEBUG("\n");
@@ -260,20 +255,15 @@ static int gmc_v8_0_init_microcode(struct amdgpu_device *adev)
 	case CHIP_STONEY:
 	case CHIP_VEGAM:
 		return 0;
-	default: BUG();
+	default:
+		return -EINVAL;
 	}
 
-	snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_mc.bin", chip_name);
-	err = request_firmware(&adev->gmc.fw, fw_name, adev->dev);
-	if (err)
-		goto out;
-	err = amdgpu_ucode_validate(adev->gmc.fw);
-
-out:
+	err = amdgpu_ucode_request(adev, &adev->gmc.fw, AMDGPU_UCODE_REQUIRED,
+				   "amdgpu/%s_mc.bin", chip_name);
 	if (err) {
-		pr_err("mc: Failed to load firmware \"%s\"\n", fw_name);
-		release_firmware(adev->gmc.fw);
-		adev->gmc.fw = NULL;
+		pr_err("mc: Failed to load firmware \"%s_mc.bin\"\n", chip_name);
+		amdgpu_ucode_release(&adev->gmc.fw);
 	}
 	return err;
 }
@@ -427,8 +417,9 @@ static void gmc_v8_0_vram_gtt_location(struct amdgpu_device *adev,
 		base = RREG32(mmMC_VM_FB_LOCATION) & 0xFFFF;
 	base <<= 24;
 
+	amdgpu_gmc_set_agp_default(adev, mc);
 	amdgpu_gmc_vram_location(adev, mc, base);
-	amdgpu_gmc_gart_location(adev, mc);
+	amdgpu_gmc_gart_location(adev, mc, AMDGPU_GART_PLACEMENT_BEST_FIT);
 }
 
 /**
@@ -441,6 +432,7 @@ static void gmc_v8_0_vram_gtt_location(struct amdgpu_device *adev,
  */
 static void gmc_v8_0_mc_program(struct amdgpu_device *adev)
 {
+	struct amdgpu_ip_block *ip_block;
 	u32 tmp;
 	int i, j;
 
@@ -454,9 +446,13 @@ static void gmc_v8_0_mc_program(struct amdgpu_device *adev)
 	}
 	WREG32(mmHDP_REG_COHERENCY_FLUSH_CNTL, 0);
 
-	if (gmc_v8_0_wait_for_idle((void *)adev)) {
+	ip_block = amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_GMC);
+	if (!ip_block)
+		return;
+
+	if (gmc_v8_0_wait_for_idle(ip_block))
 		dev_warn(adev->dev, "Wait for MC idle timedout !\n");
-	}
+
 	if (adev->mode_info.num_crtc) {
 		/* Lockout access through VGA aperture*/
 		tmp = RREG32(mmVGA_HDP_CONTROL);
@@ -474,7 +470,7 @@ static void gmc_v8_0_mc_program(struct amdgpu_device *adev)
 	WREG32(mmMC_VM_SYSTEM_APERTURE_HIGH_ADDR,
 	       adev->gmc.vram_end >> 12);
 	WREG32(mmMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR,
-	       adev->vram_scratch.gpu_addr >> 12);
+	       adev->mem_scratch.gpu_addr >> 12);
 
 	if (amdgpu_sriov_vf(adev)) {
 		tmp = ((adev->gmc.vram_end >> 24) & 0xFFFF) << 16;
@@ -487,11 +483,10 @@ static void gmc_v8_0_mc_program(struct amdgpu_device *adev)
 	}
 
 	WREG32(mmMC_VM_AGP_BASE, 0);
-	WREG32(mmMC_VM_AGP_TOP, 0x0FFFFFFF);
-	WREG32(mmMC_VM_AGP_BOT, 0x0FFFFFFF);
-	if (gmc_v8_0_wait_for_idle((void *)adev)) {
+	WREG32(mmMC_VM_AGP_TOP, adev->gmc.agp_end >> 22);
+	WREG32(mmMC_VM_AGP_BOT, adev->gmc.agp_start >> 22);
+	if (gmc_v8_0_wait_for_idle(ip_block))
 		dev_warn(adev->dev, "Wait for MC idle timedout !\n");
-	}
 
 	WREG32(mmBIF_FB_EN, BIF_FB_EN__FB_READ_EN_MASK | BIF_FB_EN__FB_WRITE_EN_MASK);
 
@@ -515,19 +510,19 @@ static void gmc_v8_0_mc_program(struct amdgpu_device *adev)
 static int gmc_v8_0_mc_init(struct amdgpu_device *adev)
 {
 	int r;
+	u32 tmp;
 
 	adev->gmc.vram_width = amdgpu_atombios_get_vram_width(adev);
 	if (!adev->gmc.vram_width) {
-		u32 tmp;
 		int chansize, numchan;
 
 		/* Get VRAM informations */
 		tmp = RREG32(mmMC_ARB_RAMCFG);
-		if (REG_GET_FIELD(tmp, MC_ARB_RAMCFG, CHANSIZE)) {
+		if (REG_GET_FIELD(tmp, MC_ARB_RAMCFG, CHANSIZE))
 			chansize = 64;
-		} else {
+		else
 			chansize = 32;
-		}
+
 		tmp = RREG32(mmMC_SHARED_CHMAP);
 		switch (REG_GET_FIELD(tmp, MC_SHARED_CHMAP, NOOFCHAN)) {
 		case 0:
@@ -562,8 +557,15 @@ static int gmc_v8_0_mc_init(struct amdgpu_device *adev)
 		adev->gmc.vram_width = numchan * chansize;
 	}
 	/* size in MB on si */
-	adev->gmc.mc_vram_size = RREG32(mmCONFIG_MEMSIZE) * 1024ULL * 1024ULL;
-	adev->gmc.real_vram_size = RREG32(mmCONFIG_MEMSIZE) * 1024ULL * 1024ULL;
+	tmp = RREG32(mmCONFIG_MEMSIZE);
+	/* some boards may have garbage in the upper 16 bits */
+	if (tmp & 0xffff0000) {
+		DRM_INFO("Probable bad vram size: 0x%08x\n", tmp);
+		if (tmp & 0xffff)
+			tmp &= 0xffff;
+	}
+	adev->gmc.mc_vram_size = tmp * 1024ULL * 1024ULL;
+	adev->gmc.real_vram_size = adev->gmc.mc_vram_size;
 
 	if (!(adev->flags & AMD_IS_APU)) {
 		r = amdgpu_device_resize_fb_bar(adev);
@@ -574,16 +576,13 @@ static int gmc_v8_0_mc_init(struct amdgpu_device *adev)
 	adev->gmc.aper_size = pci_resource_len(adev->pdev, 0);
 
 #ifdef CONFIG_X86_64
-	if (adev->flags & AMD_IS_APU) {
+	if ((adev->flags & AMD_IS_APU) && !amdgpu_passthrough(adev)) {
 		adev->gmc.aper_base = ((u64)RREG32(mmMC_VM_FB_OFFSET)) << 22;
 		adev->gmc.aper_size = adev->gmc.real_vram_size;
 	}
 #endif
 
-	/* In case the PCI BAR is larger than the actual amount of vram */
 	adev->gmc.visible_vram_size = adev->gmc.aper_size;
-	if (adev->gmc.visible_vram_size > adev->gmc.real_vram_size)
-		adev->gmc.visible_vram_size = adev->gmc.real_vram_size;
 
 	/* set the gart size */
 	if (amdgpu_gart_size == -1) {
@@ -619,32 +618,27 @@ static int gmc_v8_0_mc_init(struct amdgpu_device *adev)
  * @pasid: pasid to be flush
  * @flush_type: type of flush
  * @all_hub: flush all hubs
+ * @inst: is used to select which instance of KIQ to use for the invalidation
  *
  * Flush the TLB for the requested pasid.
  */
-static int gmc_v8_0_flush_gpu_tlb_pasid(struct amdgpu_device *adev,
-					uint16_t pasid, uint32_t flush_type,
-					bool all_hub)
+static void gmc_v8_0_flush_gpu_tlb_pasid(struct amdgpu_device *adev,
+					 uint16_t pasid, uint32_t flush_type,
+					 bool all_hub, uint32_t inst)
 {
+	u32 mask = 0x0;
 	int vmid;
-	unsigned int tmp;
-
-	if (amdgpu_in_reset(adev))
-		return -EIO;
 
 	for (vmid = 1; vmid < 16; vmid++) {
+		u32 tmp = RREG32(mmATC_VMID0_PASID_MAPPING + vmid);
 
-		tmp = RREG32(mmATC_VMID0_PASID_MAPPING + vmid);
 		if ((tmp & ATC_VMID0_PASID_MAPPING__VALID_MASK) &&
-			(tmp & ATC_VMID0_PASID_MAPPING__PASID_MASK) == pasid) {
-			WREG32(mmVM_INVALIDATE_REQUEST, 1 << vmid);
-			RREG32(mmVM_INVALIDATE_RESPONSE);
-			break;
-		}
+		    (tmp & ATC_VMID0_PASID_MAPPING__PASID_MASK) == pasid)
+			mask |= 1 << vmid;
 	}
 
-	return 0;
-
+	WREG32(mmVM_INVALIDATE_REQUEST, mask);
+	RREG32(mmVM_INVALIDATE_RESPONSE);
 }
 
 /*
@@ -672,7 +666,7 @@ static void gmc_v8_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid,
 }
 
 static uint64_t gmc_v8_0_emit_flush_gpu_tlb(struct amdgpu_ring *ring,
-					    unsigned vmid, uint64_t pd_addr)
+					    unsigned int vmid, uint64_t pd_addr)
 {
 	uint32_t reg;
 
@@ -688,8 +682,8 @@ static uint64_t gmc_v8_0_emit_flush_gpu_tlb(struct amdgpu_ring *ring,
 	return pd_addr;
 }
 
-static void gmc_v8_0_emit_pasid_mapping(struct amdgpu_ring *ring, unsigned vmid,
-					unsigned pasid)
+static void gmc_v8_0_emit_pasid_mapping(struct amdgpu_ring *ring, unsigned int vmid,
+					unsigned int pasid)
 {
 	amdgpu_ring_emit_wreg(ring, mmIH_VMID_0_LUT + vmid, pasid);
 }
@@ -722,11 +716,15 @@ static void gmc_v8_0_get_vm_pde(struct amdgpu_device *adev, int level,
 }
 
 static void gmc_v8_0_get_vm_pte(struct amdgpu_device *adev,
-				struct amdgpu_bo_va_mapping *mapping,
+				struct amdgpu_vm *vm,
+				struct amdgpu_bo *bo,
+				uint32_t vm_flags,
 				uint64_t *flags)
 {
-	*flags &= ~AMDGPU_PTE_EXECUTABLE;
-	*flags |= mapping->flags & AMDGPU_PTE_EXECUTABLE;
+	if (vm_flags & AMDGPU_VM_PAGE_EXECUTABLE)
+		*flags |= AMDGPU_PTE_EXECUTABLE;
+	else
+		*flags &= ~AMDGPU_PTE_EXECUTABLE;
 	*flags &= ~AMDGPU_PTE_PRT;
 }
 
@@ -760,11 +758,11 @@ static void gmc_v8_0_set_fault_enable_default(struct amdgpu_device *adev,
 }
 
 /**
- * gmc_v8_0_set_prt - set PRT VM fault
+ * gmc_v8_0_set_prt() - set PRT VM fault
  *
  * @adev: amdgpu_device pointer
  * @enable: enable/disable VM fault handling for PRT
-*/
+ */
 static void gmc_v8_0_set_prt(struct amdgpu_device *adev, bool enable)
 {
 	u32 tmp;
@@ -792,9 +790,10 @@ static void gmc_v8_0_set_prt(struct amdgpu_device *adev, bool enable)
 	WREG32(mmVM_PRT_CNTL, tmp);
 
 	if (enable) {
-		uint32_t low = AMDGPU_VA_RESERVED_SIZE >> AMDGPU_GPU_PAGE_SHIFT;
+		uint32_t low = AMDGPU_VA_RESERVED_BOTTOM >>
+			AMDGPU_GPU_PAGE_SHIFT;
 		uint32_t high = adev->vm_manager.max_pfn -
-			(AMDGPU_VA_RESERVED_SIZE >> AMDGPU_GPU_PAGE_SHIFT);
+			(AMDGPU_VA_RESERVED_TOP >> AMDGPU_GPU_PAGE_SHIFT);
 
 		WREG32(mmVM_PRT_APERTURE0_LOW_ADDR, low);
 		WREG32(mmVM_PRT_APERTURE1_LOW_ADDR, low);
@@ -830,17 +829,14 @@ static void gmc_v8_0_set_prt(struct amdgpu_device *adev, bool enable)
 static int gmc_v8_0_gart_enable(struct amdgpu_device *adev)
 {
 	uint64_t table_addr;
-	int r, i;
 	u32 tmp, field;
+	int i;
 
 	if (adev->gart.bo == NULL) {
 		dev_err(adev->dev, "No VRAM object for PCIE GART.\n");
 		return -EINVAL;
 	}
-	r = amdgpu_gart_table_vram_pin(adev);
-	if (r)
-		return r;
-
+	amdgpu_gtt_mgr_recover(&adev->mman.gtt_mgr);
 	table_addr = amdgpu_bo_gpu_offset(adev->gart.bo);
 
 	/* Setup TLB control */
@@ -944,9 +940,8 @@ static int gmc_v8_0_gart_enable(struct amdgpu_device *adev)
 
 	gmc_v8_0_flush_gpu_tlb(adev, 0, 0, 0);
 	DRM_INFO("PCIE GART of %uM enabled (table at 0x%016llX).\n",
-		 (unsigned)(adev->gmc.gart_size >> 20),
+		 (unsigned int)(adev->gmc.gart_size >> 20),
 		 (unsigned long long)table_addr);
-	adev->gart.ready = true;
 	return 0;
 }
 
@@ -992,7 +987,6 @@ static void gmc_v8_0_gart_disable(struct amdgpu_device *adev)
 	tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, ENABLE_L2_CACHE, 0);
 	WREG32(mmVM_L2_CNTL, tmp);
 	WREG32(mmVM_L2_CNTL2, 0);
-	amdgpu_gart_table_vram_unpin(adev);
 }
 
 /**
@@ -1007,7 +1001,7 @@ static void gmc_v8_0_gart_disable(struct amdgpu_device *adev)
  * Print human readable fault information (VI).
  */
 static void gmc_v8_0_vm_decode_fault(struct amdgpu_device *adev, u32 status,
-				     u32 addr, u32 mc_client, unsigned pasid)
+				     u32 addr, u32 mc_client, unsigned int pasid)
 {
 	u32 vmid = REG_GET_FIELD(status, VM_CONTEXT1_PROTECTION_FAULT_STATUS, VMID);
 	u32 protections = REG_GET_FIELD(status, VM_CONTEXT1_PROTECTION_FAULT_STATUS,
@@ -1048,9 +1042,9 @@ static int gmc_v8_0_convert_vram_type(int mc_seq_vram_type)
 	}
 }
 
-static int gmc_v8_0_early_init(void *handle)
+static int gmc_v8_0_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	gmc_v8_0_set_gmc_funcs(adev);
 	gmc_v8_0_set_irq_funcs(adev);
@@ -1062,13 +1056,14 @@ static int gmc_v8_0_early_init(void *handle)
 		adev->gmc.shared_aperture_end + 1;
 	adev->gmc.private_aperture_end =
 		adev->gmc.private_aperture_start + (4ULL << 30) - 1;
+	adev->gmc.noretry_flags = AMDGPU_VM_NORETRY_FLAGS_TF;
 
 	return 0;
 }
 
-static int gmc_v8_0_late_init(void *handle)
+static int gmc_v8_0_late_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (amdgpu_vm_fault_stop != AMDGPU_VM_FAULT_STOP_ALWAYS)
 		return amdgpu_irq_get(adev, &adev->gmc.vm_fault, 0);
@@ -1076,15 +1071,16 @@ static int gmc_v8_0_late_init(void *handle)
 		return 0;
 }
 
-static unsigned gmc_v8_0_get_vbios_fb_size(struct amdgpu_device *adev)
+static unsigned int gmc_v8_0_get_vbios_fb_size(struct amdgpu_device *adev)
 {
 	u32 d1vga_control = RREG32(mmD1VGA_CONTROL);
-	unsigned size;
+	unsigned int size;
 
 	if (REG_GET_FIELD(d1vga_control, D1VGA_CONTROL, D1VGA_MODE_ENABLE)) {
 		size = AMDGPU_VBIOS_VGA_ALLOCATION;
 	} else {
 		u32 viewport = RREG32(mmVIEWPORT_SIZE);
+
 		size = (REG_GET_FIELD(viewport, VIEWPORT_SIZE, VIEWPORT_HEIGHT) *
 			REG_GET_FIELD(viewport, VIEWPORT_SIZE, VIEWPORT_WIDTH) *
 			4);
@@ -1095,12 +1091,12 @@ static unsigned gmc_v8_0_get_vbios_fb_size(struct amdgpu_device *adev)
 
 #define mmMC_SEQ_MISC0_FIJI 0xA71
 
-static int gmc_v8_0_sw_init(void *handle)
+static int gmc_v8_0_sw_init(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
-	adev->num_vmhubs = 1;
+	set_bit(AMDGPU_GFXHUB(0), adev->vmhubs_mask);
 
 	if (adev->flags & AMD_IS_APU) {
 		adev->gmc.vram_type = AMDGPU_VRAM_TYPE_UNKNOWN;
@@ -1187,30 +1183,29 @@ static int gmc_v8_0_sw_init(void *handle)
 					GFP_KERNEL);
 	if (!adev->gmc.vm_fault_info)
 		return -ENOMEM;
-	atomic_set(&adev->gmc.vm_fault_info_updated, 0);
+	atomic_set_release(&adev->gmc.vm_fault_info_updated, 0);
 
 	return 0;
 }
 
-static int gmc_v8_0_sw_fini(void *handle)
+static int gmc_v8_0_sw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	amdgpu_gem_force_release(adev);
 	amdgpu_vm_manager_fini(adev);
 	kfree(adev->gmc.vm_fault_info);
 	amdgpu_gart_table_vram_free(adev);
 	amdgpu_bo_fini(adev);
-	release_firmware(adev->gmc.fw);
-	adev->gmc.fw = NULL;
+	amdgpu_ucode_release(&adev->gmc.fw);
 
 	return 0;
 }
 
-static int gmc_v8_0_hw_init(void *handle)
+static int gmc_v8_0_hw_init(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	gmc_v8_0_init_golden_registers(adev);
 
@@ -1236,12 +1231,15 @@ static int gmc_v8_0_hw_init(void *handle)
 	if (r)
 		return r;
 
-	return r;
+	if (amdgpu_emu_mode == 1)
+		return amdgpu_gmc_vram_checking(adev);
+
+	return 0;
 }
 
-static int gmc_v8_0_hw_fini(void *handle)
+static int gmc_v8_0_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	amdgpu_irq_put(adev, &adev->gmc.vm_fault, 0);
 	gmc_v8_0_gart_disable(adev);
@@ -1249,32 +1247,29 @@ static int gmc_v8_0_hw_fini(void *handle)
 	return 0;
 }
 
-static int gmc_v8_0_suspend(void *handle)
+static int gmc_v8_0_suspend(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	gmc_v8_0_hw_fini(adev);
+	gmc_v8_0_hw_fini(ip_block);
 
 	return 0;
 }
 
-static int gmc_v8_0_resume(void *handle)
+static int gmc_v8_0_resume(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
-	r = gmc_v8_0_hw_init(adev);
+	r = gmc_v8_0_hw_init(ip_block);
 	if (r)
 		return r;
 
-	amdgpu_vmid_reset_all(adev);
+	amdgpu_vmid_reset_all(ip_block->adev);
 
 	return 0;
 }
 
-static bool gmc_v8_0_is_idle(void *handle)
+static bool gmc_v8_0_is_idle(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	u32 tmp = RREG32(mmSRBM_STATUS);
 
 	if (tmp & (SRBM_STATUS__MCB_BUSY_MASK | SRBM_STATUS__MCB_NON_DISPLAY_BUSY_MASK |
@@ -1284,11 +1279,11 @@ static bool gmc_v8_0_is_idle(void *handle)
 	return true;
 }
 
-static int gmc_v8_0_wait_for_idle(void *handle)
+static int gmc_v8_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
 {
-	unsigned i;
+	unsigned int i;
 	u32 tmp;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	for (i = 0; i < adev->usec_timeout; i++) {
 		/* read MC_STATUS */
@@ -1306,10 +1301,10 @@ static int gmc_v8_0_wait_for_idle(void *handle)
 
 }
 
-static bool gmc_v8_0_check_soft_reset(void *handle)
+static bool gmc_v8_0_check_soft_reset(struct amdgpu_ip_block *ip_block)
 {
 	u32 srbm_soft_reset = 0;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	u32 tmp = RREG32(mmSRBM_STATUS);
 
 	if (tmp & SRBM_STATUS__VMC_BUSY_MASK)
@@ -1322,33 +1317,34 @@ static bool gmc_v8_0_check_soft_reset(void *handle)
 			srbm_soft_reset = REG_SET_FIELD(srbm_soft_reset,
 							SRBM_SOFT_RESET, SOFT_RESET_MC, 1);
 	}
+
 	if (srbm_soft_reset) {
 		adev->gmc.srbm_soft_reset = srbm_soft_reset;
 		return true;
-	} else {
-		adev->gmc.srbm_soft_reset = 0;
-		return false;
 	}
+
+	adev->gmc.srbm_soft_reset = 0;
+
+	return false;
 }
 
-static int gmc_v8_0_pre_soft_reset(void *handle)
+static int gmc_v8_0_pre_soft_reset(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (!adev->gmc.srbm_soft_reset)
 		return 0;
 
 	gmc_v8_0_mc_stop(adev);
-	if (gmc_v8_0_wait_for_idle(adev)) {
+	if (gmc_v8_0_wait_for_idle(ip_block))
 		dev_warn(adev->dev, "Wait for GMC idle timed out !\n");
-	}
 
 	return 0;
 }
 
-static int gmc_v8_0_soft_reset(void *handle)
+static int gmc_v8_0_soft_reset(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	u32 srbm_soft_reset;
 
 	if (!adev->gmc.srbm_soft_reset)
@@ -1377,9 +1373,9 @@ static int gmc_v8_0_soft_reset(void *handle)
 	return 0;
 }
 
-static int gmc_v8_0_post_soft_reset(void *handle)
+static int gmc_v8_0_post_soft_reset(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (!adev->gmc.srbm_soft_reset)
 		return 0;
@@ -1390,7 +1386,7 @@ static int gmc_v8_0_post_soft_reset(void *handle)
 
 static int gmc_v8_0_vm_fault_interrupt_state(struct amdgpu_device *adev,
 					     struct amdgpu_irq_src *src,
-					     unsigned type,
+					     unsigned int type,
 					     enum amdgpu_interrupt_state state)
 {
 	u32 tmp;
@@ -1443,6 +1439,12 @@ static int gmc_v8_0_process_interrupt(struct amdgpu_device *adev,
 		return 0;
 	}
 
+	/* Delegate to the soft IRQ handler ring */
+	if (adev->irq.ih_soft.enabled && entry->ih != &adev->irq.ih_soft) {
+		amdgpu_irq_delegate(adev, entry, 4);
+		return 1;
+	}
+
 	addr = RREG32(mmVM_CONTEXT1_PROTECTION_FAULT_ADDR);
 	status = RREG32(mmVM_CONTEXT1_PROTECTION_FAULT_STATUS);
 	mc_client = RREG32(mmVM_CONTEXT1_PROTECTION_FAULT_MCCLIENT);
@@ -1452,22 +1454,29 @@ static int gmc_v8_0_process_interrupt(struct amdgpu_device *adev,
 	if (!addr && !status)
 		return 0;
 
+	amdgpu_vm_update_fault_cache(adev, entry->pasid,
+				     ((u64)addr) << AMDGPU_GPU_PAGE_SHIFT, status, AMDGPU_GFXHUB(0));
+
 	if (amdgpu_vm_fault_stop == AMDGPU_VM_FAULT_STOP_FIRST)
 		gmc_v8_0_set_fault_enable_default(adev, false);
 
 	if (printk_ratelimit()) {
-		struct amdgpu_task_info task_info;
+		struct amdgpu_task_info *task_info;
+
+		dev_err(adev->dev, "GPU fault detected: %d 0x%08x\n",
+			entry->src_id, entry->src_data[0]);
 
-		memset(&task_info, 0, sizeof(struct amdgpu_task_info));
-		amdgpu_vm_get_task_info(adev, entry->pasid, &task_info);
+		task_info = amdgpu_vm_get_task_info_pasid(adev, entry->pasid);
+		if (task_info) {
+			amdgpu_vm_print_task_info(adev, task_info);
+			amdgpu_vm_put_task_info(task_info);
+		}
 
-		dev_err(adev->dev, "GPU fault detected: %d 0x%08x for process %s pid %d thread %s pid %d\n",
-			entry->src_id, entry->src_data[0], task_info.process_name,
-			task_info.tgid, task_info.task_name, task_info.pid);
 		dev_err(adev->dev, "  VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x%08X\n",
-			addr);
+				addr);
 		dev_err(adev->dev, "  VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x%08X\n",
 			status);
+
 		gmc_v8_0_vm_decode_fault(adev, status, addr, mc_client,
 					 entry->pasid);
 	}
@@ -1475,7 +1484,7 @@ static int gmc_v8_0_process_interrupt(struct amdgpu_device *adev,
 	vmid = REG_GET_FIELD(status, VM_CONTEXT1_PROTECTION_FAULT_STATUS,
 			     VMID);
 	if (amdgpu_amdkfd_is_kfd_vmid(adev, vmid)
-		&& !atomic_read(&adev->gmc.vm_fault_info_updated)) {
+		&& !atomic_read_acquire(&adev->gmc.vm_fault_info_updated)) {
 		struct kfd_vm_fault_info *info = adev->gmc.vm_fault_info;
 		u32 protections = REG_GET_FIELD(status,
 					VM_CONTEXT1_PROTECTION_FAULT_STATUS,
@@ -1491,8 +1500,7 @@ static int gmc_v8_0_process_interrupt(struct amdgpu_device *adev,
 		info->prot_read = protections & 0x8 ? true : false;
 		info->prot_write = protections & 0x10 ? true : false;
 		info->prot_exec = protections & 0x20 ? true : false;
-		mb();
-		atomic_set(&adev->gmc.vm_fault_info_updated, 1);
+		atomic_set_release(&adev->gmc.vm_fault_info_updated, 1);
 	}
 
 	return 0;
@@ -1658,10 +1666,10 @@ static void fiji_update_mc_light_sleep(struct amdgpu_device *adev,
 	}
 }
 
-static int gmc_v8_0_set_clockgating_state(void *handle,
+static int gmc_v8_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_clockgating_state state)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (amdgpu_sriov_vf(adev))
 		return 0;
@@ -1679,15 +1687,15 @@ static int gmc_v8_0_set_clockgating_state(void *handle,
 	return 0;
 }
 
-static int gmc_v8_0_set_powergating_state(void *handle,
+static int gmc_v8_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_powergating_state state)
 {
 	return 0;
 }
 
-static void gmc_v8_0_get_clockgating_state(void *handle, u32 *flags)
+static void gmc_v8_0_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int data;
 
 	if (amdgpu_sriov_vf(adev))
@@ -1751,8 +1759,7 @@ static void gmc_v8_0_set_irq_funcs(struct amdgpu_device *adev)
 	adev->gmc.vm_fault.funcs = &gmc_v8_0_irq_funcs;
 }
 
-const struct amdgpu_ip_block_version gmc_v8_0_ip_block =
-{
+const struct amdgpu_ip_block_version gmc_v8_0_ip_block = {
 	.type = AMD_IP_BLOCK_TYPE_GMC,
 	.major = 8,
 	.minor = 0,
@@ -1760,8 +1767,7 @@ const struct amdgpu_ip_block_version gmc_v8_0_ip_block =
 	.funcs = &gmc_v8_0_ip_funcs,
 };
 
-const struct amdgpu_ip_block_version gmc_v8_1_ip_block =
-{
+const struct amdgpu_ip_block_version gmc_v8_1_ip_block = {
 	.type = AMD_IP_BLOCK_TYPE_GMC,
 	.major = 8,
 	.minor = 1,
@@ -1769,8 +1775,7 @@ const struct amdgpu_ip_block_version gmc_v8_1_ip_block =
 	.funcs = &gmc_v8_0_ip_funcs,
 };
 
-const struct amdgpu_ip_block_version gmc_v8_5_ip_block =
-{
+const struct amdgpu_ip_block_version gmc_v8_5_ip_block = {
 	.type = AMD_IP_BLOCK_TYPE_GMC,
 	.major = 8,
 	.minor = 5,
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
index d84523cf5f75..8ad7519f7b58 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
@@ -49,11 +49,14 @@
 #include "mmhub_v1_0.h"
 #include "athub_v1_0.h"
 #include "gfxhub_v1_1.h"
+#include "gfxhub_v1_2.h"
 #include "mmhub_v9_4.h"
 #include "mmhub_v1_7.h"
+#include "mmhub_v1_8.h"
 #include "umc_v6_1.h"
 #include "umc_v6_0.h"
 #include "umc_v6_7.h"
+#include "umc_v12_0.h"
 #include "hdp_v4_0.h"
 #include "mca_v3_0.h"
 
@@ -72,8 +75,10 @@
 #define mmDCHUBBUB_SDPIF_MMIO_CNTRL_0                                                                  0x049d
 #define mmDCHUBBUB_SDPIF_MMIO_CNTRL_0_BASE_IDX                                                         2
 
+#define mmHUBP0_DCSURF_PRI_VIEWPORT_DIMENSION_DCN2                                                          0x05ea
+#define mmHUBP0_DCSURF_PRI_VIEWPORT_DIMENSION_DCN2_BASE_IDX                                                 2
 
-static const char *gfxhub_client_ids[] = {
+static const char * const gfxhub_client_ids[] = {
 	"CB",
 	"DB",
 	"IA",
@@ -324,14 +329,12 @@ static const char *mmhub_client_ids_aldebaran[][2] = {
 	[384+0][1] = "OSS",
 };
 
-static const struct soc15_reg_golden golden_settings_mmhub_1_0_0[] =
-{
+static const struct soc15_reg_golden golden_settings_mmhub_1_0_0[] = {
 	SOC15_REG_GOLDEN_VALUE(MMHUB, 0, mmDAGB1_WRCLI2, 0x00000007, 0xfe5fe0fa),
 	SOC15_REG_GOLDEN_VALUE(MMHUB, 0, mmMMEA1_DRAM_WR_CLI2GRP_MAP0, 0x00000030, 0x55555565)
 };
 
-static const struct soc15_reg_golden golden_settings_athub_1_0_0[] =
-{
+static const struct soc15_reg_golden golden_settings_athub_1_0_0[] = {
 	SOC15_REG_GOLDEN_VALUE(ATHUB, 0, mmRPB_ARB_CNTL, 0x0000ff00, 0x00000800),
 	SOC15_REG_GOLDEN_VALUE(ATHUB, 0, mmRPB_ARB_CNTL2, 0x00ff00ff, 0x00080008)
 };
@@ -408,13 +411,14 @@ static const uint32_t ecc_umc_mcumc_ctrl_mask_addrs[] = {
 
 static int gmc_v9_0_ecc_interrupt_state(struct amdgpu_device *adev,
 		struct amdgpu_irq_src *src,
-		unsigned type,
+		unsigned int type,
 		enum amdgpu_interrupt_state state)
 {
 	u32 bits, i, tmp, reg;
 
 	/* Devices newer then VEGA10/12 shall have these programming
-	     sequences performed by PSP BL */
+	 * sequences performed by PSP BL
+	 */
 	if (adev->asic_type >= CHIP_VEGA20)
 		return 0;
 
@@ -458,7 +462,7 @@ static int gmc_v9_0_ecc_interrupt_state(struct amdgpu_device *adev,
 
 static int gmc_v9_0_vm_fault_interrupt_state(struct amdgpu_device *adev,
 					struct amdgpu_irq_src *src,
-					unsigned type,
+					unsigned int type,
 					enum amdgpu_interrupt_state state)
 {
 	struct amdgpu_vmhub *hub;
@@ -474,24 +478,58 @@ static int gmc_v9_0_vm_fault_interrupt_state(struct amdgpu_device *adev,
 
 	switch (state) {
 	case AMDGPU_IRQ_STATE_DISABLE:
-		for (j = 0; j < adev->num_vmhubs; j++) {
+		for_each_set_bit(j, adev->vmhubs_mask, AMDGPU_MAX_VMHUBS) {
 			hub = &adev->vmhub[j];
 			for (i = 0; i < 16; i++) {
 				reg = hub->vm_context0_cntl + i;
-				tmp = RREG32(reg);
+
+				/* This works because this interrupt is only
+				 * enabled at init/resume and disabled in
+				 * fini/suspend, so the overall state doesn't
+				 * change over the course of suspend/resume.
+				 */
+				if (adev->in_s0ix && (j == AMDGPU_GFXHUB(0)))
+					continue;
+
+				if (j >= AMDGPU_MMHUB0(0))
+					tmp = RREG32_SOC15_IP(MMHUB, reg);
+				else
+					tmp = RREG32_XCC(reg, j);
+
 				tmp &= ~bits;
-				WREG32(reg, tmp);
+
+				if (j >= AMDGPU_MMHUB0(0))
+					WREG32_SOC15_IP(MMHUB, reg, tmp);
+				else
+					WREG32_XCC(reg, tmp, j);
 			}
 		}
 		break;
 	case AMDGPU_IRQ_STATE_ENABLE:
-		for (j = 0; j < adev->num_vmhubs; j++) {
+		for_each_set_bit(j, adev->vmhubs_mask, AMDGPU_MAX_VMHUBS) {
 			hub = &adev->vmhub[j];
 			for (i = 0; i < 16; i++) {
 				reg = hub->vm_context0_cntl + i;
-				tmp = RREG32(reg);
+
+				/* This works because this interrupt is only
+				 * enabled at init/resume and disabled in
+				 * fini/suspend, so the overall state doesn't
+				 * change over the course of suspend/resume.
+				 */
+				if (adev->in_s0ix && (j == AMDGPU_GFXHUB(0)))
+					continue;
+
+				if (j >= AMDGPU_MMHUB0(0))
+					tmp = RREG32_SOC15_IP(MMHUB, reg);
+				else
+					tmp = RREG32_XCC(reg, j);
+
 				tmp |= bits;
-				WREG32(reg, tmp);
+
+				if (j >= AMDGPU_MMHUB0(0))
+					WREG32_SOC15_IP(MMHUB, reg, tmp);
+				else
+					WREG32_XCC(reg, tmp, j);
 			}
 		}
 		break;
@@ -506,70 +544,111 @@ static int gmc_v9_0_process_interrupt(struct amdgpu_device *adev,
 				      struct amdgpu_irq_src *source,
 				      struct amdgpu_iv_entry *entry)
 {
-	bool retry_fault = !!(entry->src_data[1] & 0x80);
-	bool write_fault = !!(entry->src_data[1] & 0x20);
-	uint32_t status = 0, cid = 0, rw = 0;
-	struct amdgpu_task_info task_info;
+	bool retry_fault = !!(entry->src_data[1] &
+			      AMDGPU_GMC9_FAULT_SOURCE_DATA_RETRY);
+	bool write_fault = !!(entry->src_data[1] &
+			      AMDGPU_GMC9_FAULT_SOURCE_DATA_WRITE);
+	uint32_t status = 0, cid = 0, rw = 0, fed = 0;
+	struct amdgpu_task_info *task_info;
 	struct amdgpu_vmhub *hub;
 	const char *mmhub_cid;
 	const char *hub_name;
+	unsigned int vmhub;
 	u64 addr;
+	uint32_t cam_index = 0;
+	int ret, xcc_id = 0;
+	uint32_t node_id;
+
+	node_id = entry->node_id;
 
 	addr = (u64)entry->src_data[0] << 12;
 	addr |= ((u64)entry->src_data[1] & 0xf) << 44;
 
+	if (entry->client_id == SOC15_IH_CLIENTID_VMC) {
+		hub_name = "mmhub0";
+		vmhub = AMDGPU_MMHUB0(node_id / 4);
+	} else if (entry->client_id == SOC15_IH_CLIENTID_VMC1) {
+		hub_name = "mmhub1";
+		vmhub = AMDGPU_MMHUB1(0);
+	} else {
+		hub_name = "gfxhub0";
+		if (adev->gfx.funcs->ih_node_to_logical_xcc) {
+			xcc_id = adev->gfx.funcs->ih_node_to_logical_xcc(adev,
+				node_id);
+			if (xcc_id < 0)
+				xcc_id = 0;
+		}
+		vmhub = xcc_id;
+	}
+	hub = &adev->vmhub[vmhub];
+
 	if (retry_fault) {
-		/* Returning 1 here also prevents sending the IV to the KFD */
+		if (adev->irq.retry_cam_enabled) {
+			/* Delegate it to a different ring if the hardware hasn't
+			 * already done it.
+			 */
+			if (entry->ih == &adev->irq.ih) {
+				amdgpu_irq_delegate(adev, entry, 8);
+				return 1;
+			}
+
+			cam_index = entry->src_data[2] & 0x3ff;
 
-		/* Process it onyl if it's the first fault for this address */
-		if (entry->ih != &adev->irq.ih_soft &&
-		    amdgpu_gmc_filter_faults(adev, addr, entry->pasid,
+			ret = amdgpu_vm_handle_fault(adev, entry->pasid, entry->vmid, node_id,
+						     addr, entry->timestamp, write_fault);
+			WDOORBELL32(adev->irq.retry_cam_doorbell_index, cam_index);
+			if (ret)
+				return 1;
+		} else {
+			/* Process it onyl if it's the first fault for this address */
+			if (entry->ih != &adev->irq.ih_soft &&
+			    amdgpu_gmc_filter_faults(adev, entry->ih, addr, entry->pasid,
 					     entry->timestamp))
-			return 1;
+				return 1;
+
+			/* Delegate it to a different ring if the hardware hasn't
+			 * already done it.
+			 */
+			if (entry->ih == &adev->irq.ih) {
+				amdgpu_irq_delegate(adev, entry, 8);
+				return 1;
+			}
 
-		/* Delegate it to a different ring if the hardware hasn't
-		 * already done it.
-		 */
-		if (entry->ih == &adev->irq.ih) {
-			amdgpu_irq_delegate(adev, entry, 8);
-			return 1;
+			/* Try to handle the recoverable page faults by filling page
+			 * tables
+			 */
+			if (amdgpu_vm_handle_fault(adev, entry->pasid, entry->vmid, node_id,
+						   addr, entry->timestamp, write_fault))
+				return 1;
 		}
-
-		/* Try to handle the recoverable page faults by filling page
-		 * tables
-		 */
-		if (amdgpu_vm_handle_fault(adev, entry->pasid, addr, write_fault))
-			return 1;
 	}
 
+	if (kgd2kfd_vmfault_fast_path(adev, entry, retry_fault))
+		return 1;
+
 	if (!printk_ratelimit())
 		return 0;
 
-	if (entry->client_id == SOC15_IH_CLIENTID_VMC) {
-		hub_name = "mmhub0";
-		hub = &adev->vmhub[AMDGPU_MMHUB_0];
-	} else if (entry->client_id == SOC15_IH_CLIENTID_VMC1) {
-		hub_name = "mmhub1";
-		hub = &adev->vmhub[AMDGPU_MMHUB_1];
-	} else {
-		hub_name = "gfxhub0";
-		hub = &adev->vmhub[AMDGPU_GFXHUB_0];
+	dev_err(adev->dev,
+		"[%s] %s page fault (src_id:%u ring:%u vmid:%u pasid:%u)\n", hub_name,
+		retry_fault ? "retry" : "no-retry",
+		entry->src_id, entry->ring_id, entry->vmid, entry->pasid);
+
+	task_info = amdgpu_vm_get_task_info_pasid(adev, entry->pasid);
+	if (task_info) {
+		amdgpu_vm_print_task_info(adev, task_info);
+		amdgpu_vm_put_task_info(task_info);
 	}
 
-	memset(&task_info, 0, sizeof(struct amdgpu_task_info));
-	amdgpu_vm_get_task_info(adev, entry->pasid, &task_info);
-
-	dev_err(adev->dev,
-		"[%s] %s page fault (src_id:%u ring:%u vmid:%u "
-		"pasid:%u, for process %s pid %d thread %s pid %d)\n",
-		hub_name, retry_fault ? "retry" : "no-retry",
-		entry->src_id, entry->ring_id, entry->vmid,
-		entry->pasid, task_info.process_name, task_info.tgid,
-		task_info.task_name, task_info.pid);
 	dev_err(adev->dev, "  in page starting at address 0x%016llx from IH client 0x%x (%s)\n",
 		addr, entry->client_id,
 		soc15_ih_clientid_name[entry->client_id]);
 
+	if (amdgpu_is_multi_aid(adev))
+		dev_err(adev->dev, "  cookie node_id %d fault from die %s%d%s\n",
+			node_id, node_id % 4 == 3 ? "RSV" : "AID", node_id / 4,
+			node_id % 4 == 1 ? ".XCD0" : node_id % 4 == 2 ? ".XCD1" : "");
+
 	if (amdgpu_sriov_vf(adev))
 		return 0;
 
@@ -578,26 +657,41 @@ static int gmc_v9_0_process_interrupt(struct amdgpu_device *adev,
 	 * be updated to avoid reading an incorrect value due to
 	 * the new fast GRBM interface.
 	 */
-	if ((entry->vmid_src == AMDGPU_GFXHUB_0) &&
-	    (adev->ip_versions[GC_HWIP][0] < IP_VERSION(9, 4, 2)))
+	if ((entry->vmid_src == AMDGPU_GFXHUB(0)) &&
+	    (amdgpu_ip_version(adev, GC_HWIP, 0) < IP_VERSION(9, 4, 2)))
 		RREG32(hub->vm_l2_pro_fault_status);
 
 	status = RREG32(hub->vm_l2_pro_fault_status);
 	cid = REG_GET_FIELD(status, VM_L2_PROTECTION_FAULT_STATUS, CID);
 	rw = REG_GET_FIELD(status, VM_L2_PROTECTION_FAULT_STATUS, RW);
-	WREG32_P(hub->vm_l2_pro_fault_cntl, 1, ~1);
+	fed = REG_GET_FIELD(status, VM_L2_PROTECTION_FAULT_STATUS, FED);
+
+	/* for fed error, kfd will handle it, return directly */
+	if (fed && amdgpu_ras_is_poison_mode_supported(adev) &&
+	    (amdgpu_ip_version(adev, GC_HWIP, 0) >= IP_VERSION(9, 4, 2)))
+		return 0;
+
+	/* Only print L2 fault status if the status register could be read and
+	 * contains useful information
+	 */
+	if (!status)
+		return 0;
+
+	if (!amdgpu_sriov_vf(adev))
+		WREG32_P(hub->vm_l2_pro_fault_cntl, 1, ~1);
 
+	amdgpu_vm_update_fault_cache(adev, entry->pasid, addr, status, vmhub);
 
 	dev_err(adev->dev,
 		"VM_L2_PROTECTION_FAULT_STATUS:0x%08X\n",
 		status);
-	if (hub == &adev->vmhub[AMDGPU_GFXHUB_0]) {
+	if (entry->vmid_src == AMDGPU_GFXHUB(0)) {
 		dev_err(adev->dev, "\t Faulty UTCL2 client ID: %s (0x%x)\n",
 			cid >= ARRAY_SIZE(gfxhub_client_ids) ? "unknown" :
 			gfxhub_client_ids[cid],
 			cid);
 	} else {
-		switch (adev->ip_versions[MMHUB_HWIP][0]) {
+		switch (amdgpu_ip_version(adev, MMHUB_HWIP, 0)) {
 		case IP_VERSION(9, 0, 0):
 			mmhub_cid = mmhub_client_ids_vega10[cid][rw];
 			break;
@@ -618,6 +712,7 @@ static int gmc_v9_0_process_interrupt(struct amdgpu_device *adev,
 		case IP_VERSION(2, 4, 0):
 			mmhub_cid = mmhub_client_ids_renoir[cid][rw];
 			break;
+		case IP_VERSION(1, 8, 0):
 		case IP_VERSION(9, 4, 2):
 			mmhub_cid = mmhub_client_ids_aldebaran[cid][rw];
 			break;
@@ -661,7 +756,8 @@ static void gmc_v9_0_set_irq_funcs(struct amdgpu_device *adev)
 	adev->gmc.vm_fault.funcs = &gmc_v9_0_irq_funcs;
 
 	if (!amdgpu_sriov_vf(adev) &&
-	    !adev->gmc.xgmi.connected_to_cpu) {
+	    !adev->gmc.xgmi.connected_to_cpu &&
+	    !adev->gmc.is_app_apu) {
 		adev->gmc.ecc_irq.num_types = 1;
 		adev->gmc.ecc_irq.funcs = &gmc_v9_0_ecc_funcs;
 	}
@@ -696,11 +792,12 @@ static uint32_t gmc_v9_0_get_invalidate_req(unsigned int vmid,
 static bool gmc_v9_0_use_invalidate_semaphore(struct amdgpu_device *adev,
 				       uint32_t vmhub)
 {
-	if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 2))
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 2) ||
+	    amdgpu_is_multi_aid(adev))
 		return false;
 
-	return ((vmhub == AMDGPU_MMHUB_0 ||
-		 vmhub == AMDGPU_MMHUB_1) &&
+	return ((vmhub == AMDGPU_MMHUB0(0) ||
+		 vmhub == AMDGPU_MMHUB1(0)) &&
 		(!amdgpu_sriov_vf(adev)) &&
 		(!(!(adev->apu_flags & AMD_APU_IS_RAVEN2) &&
 		   (adev->apu_flags & AMD_APU_IS_PICASSO))));
@@ -739,43 +836,37 @@ static void gmc_v9_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid,
 					uint32_t vmhub, uint32_t flush_type)
 {
 	bool use_semaphore = gmc_v9_0_use_invalidate_semaphore(adev, vmhub);
-	const unsigned eng = 17;
-	u32 j, inv_req, inv_req2, tmp;
+	u32 j, inv_req, tmp, sem, req, ack, inst;
+	const unsigned int eng = 17;
 	struct amdgpu_vmhub *hub;
 
-	BUG_ON(vmhub >= adev->num_vmhubs);
+	BUG_ON(vmhub >= AMDGPU_MAX_VMHUBS);
 
 	hub = &adev->vmhub[vmhub];
-	if (adev->gmc.xgmi.num_physical_nodes &&
-	    adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 0)) {
-		/* Vega20+XGMI caches PTEs in TC and TLB. Add a
-		 * heavy-weight TLB flush (type 2), which flushes
-		 * both. Due to a race condition with concurrent
-		 * memory accesses using the same TLB cache line, we
-		 * still need a second TLB flush after this.
-		 */
-		inv_req = gmc_v9_0_get_invalidate_req(vmid, 2);
-		inv_req2 = gmc_v9_0_get_invalidate_req(vmid, flush_type);
-	} else {
-		inv_req = gmc_v9_0_get_invalidate_req(vmid, flush_type);
-		inv_req2 = 0;
-	}
+	inv_req = gmc_v9_0_get_invalidate_req(vmid, flush_type);
+	sem = hub->vm_inv_eng0_sem + hub->eng_distance * eng;
+	req = hub->vm_inv_eng0_req + hub->eng_distance * eng;
+	ack = hub->vm_inv_eng0_ack + hub->eng_distance * eng;
 
-	/* This is necessary for a HW workaround under SRIOV as well
-	 * as GFXOFF under bare metal
+	if (vmhub >= AMDGPU_MMHUB0(0))
+		inst = 0;
+	else
+		inst = vmhub;
+
+	/* This is necessary for SRIOV as well as for GFXOFF to function
+	 * properly under bare metal
 	 */
-	if (adev->gfx.kiq.ring.sched.ready &&
-	    (amdgpu_sriov_runtime(adev) || !amdgpu_sriov_vf(adev)) &&
-	    down_read_trylock(&adev->reset_sem)) {
+	if (adev->gfx.kiq[inst].ring.sched.ready &&
+	    (amdgpu_sriov_runtime(adev) || !amdgpu_sriov_vf(adev))) {
 		uint32_t req = hub->vm_inv_eng0_req + hub->eng_distance * eng;
 		uint32_t ack = hub->vm_inv_eng0_ack + hub->eng_distance * eng;
 
-		amdgpu_virt_kiq_reg_write_reg_wait(adev, req, ack, inv_req,
-						   1 << vmid);
-		up_read(&adev->reset_sem);
+		amdgpu_gmc_fw_reg_write_reg_wait(adev, req, ack, inv_req,
+						 1 << vmid, inst);
 		return;
 	}
 
+	/* This path is needed before KIQ/MES/GFXOFF are set up */
 	spin_lock(&adev->gmc.invalidate_lock);
 
 	/*
@@ -788,9 +879,11 @@ static void gmc_v9_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid,
 	/* TODO: It needs to continue working on debugging with semaphore for GFXHUB as well. */
 	if (use_semaphore) {
 		for (j = 0; j < adev->usec_timeout; j++) {
-			/* a read return value of 1 means semaphore acuqire */
-			tmp = RREG32_NO_KIQ(hub->vm_inv_eng0_sem +
-					    hub->eng_distance * eng);
+			/* a read return value of 1 means semaphore acquire */
+			if (vmhub >= AMDGPU_MMHUB0(0))
+				tmp = RREG32_SOC15_IP_NO_KIQ(MMHUB, sem, GET_INST(GC, inst));
+			else
+				tmp = RREG32_SOC15_IP_NO_KIQ(GC, sem, GET_INST(GC, inst));
 			if (tmp & 0x1)
 				break;
 			udelay(1);
@@ -800,40 +893,41 @@ static void gmc_v9_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid,
 			DRM_ERROR("Timeout waiting for sem acquire in VM flush!\n");
 	}
 
-	do {
-		WREG32_NO_KIQ(hub->vm_inv_eng0_req +
-			      hub->eng_distance * eng, inv_req);
-
-		/*
-		 * Issue a dummy read to wait for the ACK register to
-		 * be cleared to avoid a false ACK due to the new fast
-		 * GRBM interface.
-		 */
-		if ((vmhub == AMDGPU_GFXHUB_0) &&
-		    (adev->ip_versions[GC_HWIP][0] < IP_VERSION(9, 4, 2)))
-			RREG32_NO_KIQ(hub->vm_inv_eng0_req +
-				      hub->eng_distance * eng);
+	if (vmhub >= AMDGPU_MMHUB0(0))
+		WREG32_SOC15_IP_NO_KIQ(MMHUB, req, inv_req, GET_INST(GC, inst));
+	else
+		WREG32_SOC15_IP_NO_KIQ(GC, req, inv_req, GET_INST(GC, inst));
 
-		for (j = 0; j < adev->usec_timeout; j++) {
-			tmp = RREG32_NO_KIQ(hub->vm_inv_eng0_ack +
-					    hub->eng_distance * eng);
-			if (tmp & (1 << vmid))
-				break;
-			udelay(1);
-		}
+	/*
+	 * Issue a dummy read to wait for the ACK register to
+	 * be cleared to avoid a false ACK due to the new fast
+	 * GRBM interface.
+	 */
+	if ((vmhub == AMDGPU_GFXHUB(0)) &&
+	    (amdgpu_ip_version(adev, GC_HWIP, 0) < IP_VERSION(9, 4, 2)))
+		RREG32_NO_KIQ(req);
 
-		inv_req = inv_req2;
-		inv_req2 = 0;
-	} while (inv_req);
+	for (j = 0; j < adev->usec_timeout; j++) {
+		if (vmhub >= AMDGPU_MMHUB0(0))
+			tmp = RREG32_SOC15_IP_NO_KIQ(MMHUB, ack, GET_INST(GC, inst));
+		else
+			tmp = RREG32_SOC15_IP_NO_KIQ(GC, ack, GET_INST(GC, inst));
+		if (tmp & (1 << vmid))
+			break;
+		udelay(1);
+	}
 
 	/* TODO: It needs to continue working on debugging with semaphore for GFXHUB as well. */
-	if (use_semaphore)
+	if (use_semaphore) {
 		/*
 		 * add semaphore release after invalidation,
 		 * write with 0 means semaphore release
 		 */
-		WREG32_NO_KIQ(hub->vm_inv_eng0_sem +
-			      hub->eng_distance * eng, 0);
+		if (vmhub >= AMDGPU_MMHUB0(0))
+			WREG32_SOC15_IP_NO_KIQ(MMHUB, sem, 0, GET_INST(GC, inst));
+		else
+			WREG32_SOC15_IP_NO_KIQ(GC, sem, 0, GET_INST(GC, inst));
+	}
 
 	spin_unlock(&adev->gmc.invalidate_lock);
 
@@ -850,96 +944,46 @@ static void gmc_v9_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid,
  * @pasid: pasid to be flush
  * @flush_type: the flush type
  * @all_hub: flush all hubs
+ * @inst: is used to select which instance of KIQ to use for the invalidation
  *
  * Flush the TLB for the requested pasid.
  */
-static int gmc_v9_0_flush_gpu_tlb_pasid(struct amdgpu_device *adev,
-					uint16_t pasid, uint32_t flush_type,
-					bool all_hub)
+static void gmc_v9_0_flush_gpu_tlb_pasid(struct amdgpu_device *adev,
+					 uint16_t pasid, uint32_t flush_type,
+					 bool all_hub, uint32_t inst)
 {
-	int vmid, i;
-	signed long r;
-	uint32_t seq;
-	uint16_t queried_pasid;
-	bool ret;
-	struct amdgpu_ring *ring = &adev->gfx.kiq.ring;
-	struct amdgpu_kiq *kiq = &adev->gfx.kiq;
-
-	if (amdgpu_in_reset(adev))
-		return -EIO;
-
-	if (ring->sched.ready && down_read_trylock(&adev->reset_sem)) {
-		/* Vega20+XGMI caches PTEs in TC and TLB. Add a
-		 * heavy-weight TLB flush (type 2), which flushes
-		 * both. Due to a race condition with concurrent
-		 * memory accesses using the same TLB cache line, we
-		 * still need a second TLB flush after this.
-		 */
-		bool vega20_xgmi_wa = (adev->gmc.xgmi.num_physical_nodes &&
-				       adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 0));
-		/* 2 dwords flush + 8 dwords fence */
-		unsigned int ndw = kiq->pmf->invalidate_tlbs_size + 8;
-
-		if (vega20_xgmi_wa)
-			ndw += kiq->pmf->invalidate_tlbs_size;
-
-		spin_lock(&adev->gfx.kiq.ring_lock);
-		/* 2 dwords flush + 8 dwords fence */
-		amdgpu_ring_alloc(ring, ndw);
-		if (vega20_xgmi_wa)
-			kiq->pmf->kiq_invalidate_tlbs(ring,
-						      pasid, 2, all_hub);
-		kiq->pmf->kiq_invalidate_tlbs(ring,
-					pasid, flush_type, all_hub);
-		r = amdgpu_fence_emit_polling(ring, &seq, MAX_KIQ_REG_WAIT);
-		if (r) {
-			amdgpu_ring_undo(ring);
-			spin_unlock(&adev->gfx.kiq.ring_lock);
-			up_read(&adev->reset_sem);
-			return -ETIME;
-		}
-
-		amdgpu_ring_commit(ring);
-		spin_unlock(&adev->gfx.kiq.ring_lock);
-		r = amdgpu_fence_wait_polling(ring, seq, adev->usec_timeout);
-		if (r < 1) {
-			dev_err(adev->dev, "wait for kiq fence error: %ld.\n", r);
-			up_read(&adev->reset_sem);
-			return -ETIME;
-		}
-		up_read(&adev->reset_sem);
-		return 0;
-	}
+	uint16_t queried;
+	int i, vmid;
 
 	for (vmid = 1; vmid < 16; vmid++) {
-
-		ret = gmc_v9_0_get_atc_vmid_pasid_mapping_info(adev, vmid,
-				&queried_pasid);
-		if (ret && queried_pasid == pasid) {
-			if (all_hub) {
-				for (i = 0; i < adev->num_vmhubs; i++)
-					gmc_v9_0_flush_gpu_tlb(adev, vmid,
-							i, flush_type);
-			} else {
-				gmc_v9_0_flush_gpu_tlb(adev, vmid,
-						AMDGPU_GFXHUB_0, flush_type);
-			}
-			break;
+		bool valid;
+
+		valid = gmc_v9_0_get_atc_vmid_pasid_mapping_info(adev, vmid,
+								 &queried);
+		if (!valid || queried != pasid)
+			continue;
+
+		if (all_hub) {
+			for_each_set_bit(i, adev->vmhubs_mask,
+					 AMDGPU_MAX_VMHUBS)
+				gmc_v9_0_flush_gpu_tlb(adev, vmid, i,
+						       flush_type);
+		} else {
+			gmc_v9_0_flush_gpu_tlb(adev, vmid,
+					       AMDGPU_GFXHUB(0),
+					       flush_type);
 		}
 	}
-
-	return 0;
-
 }
 
 static uint64_t gmc_v9_0_emit_flush_gpu_tlb(struct amdgpu_ring *ring,
-					    unsigned vmid, uint64_t pd_addr)
+					    unsigned int vmid, uint64_t pd_addr)
 {
-	bool use_semaphore = gmc_v9_0_use_invalidate_semaphore(ring->adev, ring->funcs->vmhub);
+	bool use_semaphore = gmc_v9_0_use_invalidate_semaphore(ring->adev, ring->vm_hub);
 	struct amdgpu_device *adev = ring->adev;
-	struct amdgpu_vmhub *hub = &adev->vmhub[ring->funcs->vmhub];
+	struct amdgpu_vmhub *hub = &adev->vmhub[ring->vm_hub];
 	uint32_t req = gmc_v9_0_get_invalidate_req(vmid, 0);
-	unsigned eng = ring->vm_inv_eng;
+	unsigned int eng = ring->vm_inv_eng;
 
 	/*
 	 * It may lose gpuvm invalidate acknowldege state across power-gating
@@ -981,17 +1025,17 @@ static uint64_t gmc_v9_0_emit_flush_gpu_tlb(struct amdgpu_ring *ring,
 	return pd_addr;
 }
 
-static void gmc_v9_0_emit_pasid_mapping(struct amdgpu_ring *ring, unsigned vmid,
-					unsigned pasid)
+static void gmc_v9_0_emit_pasid_mapping(struct amdgpu_ring *ring, unsigned int vmid,
+					unsigned int pasid)
 {
 	struct amdgpu_device *adev = ring->adev;
 	uint32_t reg;
 
 	/* Do nothing because there's no lut register for mmhub1. */
-	if (ring->funcs->vmhub == AMDGPU_MMHUB_1)
+	if (ring->vm_hub == AMDGPU_MMHUB1(0))
 		return;
 
-	if (ring->funcs->vmhub == AMDGPU_GFXHUB_0)
+	if (ring->vm_hub == AMDGPU_GFXHUB(0))
 		reg = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_VMID_0_LUT) + vmid;
 	else
 		reg = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_VMID_0_LUT_MM) + vmid;
@@ -1031,27 +1075,6 @@ static void gmc_v9_0_emit_pasid_mapping(struct amdgpu_ring *ring, unsigned vmid,
  * 0 valid
  */
 
-static uint64_t gmc_v9_0_map_mtype(struct amdgpu_device *adev, uint32_t flags)
-
-{
-	switch (flags) {
-	case AMDGPU_VM_MTYPE_DEFAULT:
-		return AMDGPU_PTE_MTYPE_VG10(MTYPE_NC);
-	case AMDGPU_VM_MTYPE_NC:
-		return AMDGPU_PTE_MTYPE_VG10(MTYPE_NC);
-	case AMDGPU_VM_MTYPE_WC:
-		return AMDGPU_PTE_MTYPE_VG10(MTYPE_WC);
-	case AMDGPU_VM_MTYPE_RW:
-		return AMDGPU_PTE_MTYPE_VG10(MTYPE_RW);
-	case AMDGPU_VM_MTYPE_CC:
-		return AMDGPU_PTE_MTYPE_VG10(MTYPE_CC);
-	case AMDGPU_VM_MTYPE_UC:
-		return AMDGPU_PTE_MTYPE_VG10(MTYPE_UC);
-	default:
-		return AMDGPU_PTE_MTYPE_VG10(MTYPE_NC);
-	}
-}
-
 static void gmc_v9_0_get_vm_pde(struct amdgpu_device *adev, int level,
 				uint64_t *addr, uint64_t *flags)
 {
@@ -1068,52 +1091,263 @@ static void gmc_v9_0_get_vm_pde(struct amdgpu_device *adev, int level,
 			*flags |= AMDGPU_PDE_BFS(0x9);
 
 	} else if (level == AMDGPU_VM_PDB0) {
-		if (*flags & AMDGPU_PDE_PTE)
+		if (*flags & AMDGPU_PDE_PTE) {
 			*flags &= ~AMDGPU_PDE_PTE;
-		else
+			if (!(*flags & AMDGPU_PTE_VALID))
+				*addr |= 1 << PAGE_SHIFT;
+		} else {
 			*flags |= AMDGPU_PTE_TF;
+		}
+	}
+}
+
+static void gmc_v9_0_get_coherence_flags(struct amdgpu_device *adev,
+					 struct amdgpu_vm *vm,
+					 struct amdgpu_bo *bo,
+					 uint32_t vm_flags,
+					 uint64_t *flags)
+{
+	struct amdgpu_device *bo_adev = amdgpu_ttm_adev(bo->tbo.bdev);
+	bool is_vram = bo->tbo.resource &&
+		bo->tbo.resource->mem_type == TTM_PL_VRAM;
+	bool coherent = bo->flags & (AMDGPU_GEM_CREATE_COHERENT |
+				     AMDGPU_GEM_CREATE_EXT_COHERENT);
+	bool ext_coherent = bo->flags & AMDGPU_GEM_CREATE_EXT_COHERENT;
+	bool uncached = bo->flags & AMDGPU_GEM_CREATE_UNCACHED;
+	unsigned int mtype_local, mtype;
+	uint32_t gc_ip_version = amdgpu_ip_version(adev, GC_HWIP, 0);
+	bool snoop = false;
+	bool is_local;
+
+	dma_resv_assert_held(bo->tbo.base.resv);
+
+	switch (gc_ip_version) {
+	case IP_VERSION(9, 4, 1):
+	case IP_VERSION(9, 4, 2):
+		if (is_vram) {
+			if (bo_adev == adev) {
+				if (uncached)
+					mtype = MTYPE_UC;
+				else if (coherent)
+					mtype = MTYPE_CC;
+				else
+					mtype = MTYPE_RW;
+				/* FIXME: is this still needed? Or does
+				 * amdgpu_ttm_tt_pde_flags already handle this?
+				 */
+				if (gc_ip_version == IP_VERSION(9, 4, 2) &&
+				    adev->gmc.xgmi.connected_to_cpu)
+					snoop = true;
+			} else {
+				if (uncached || coherent)
+					mtype = MTYPE_UC;
+				else
+					mtype = MTYPE_NC;
+				if (amdgpu_xgmi_same_hive(adev, bo_adev))
+					snoop = true;
+			}
+		} else {
+			if (uncached || coherent)
+				mtype = MTYPE_UC;
+			else
+				mtype = MTYPE_NC;
+			/* FIXME: is this still needed? Or does
+			 * amdgpu_ttm_tt_pde_flags already handle this?
+			 */
+			snoop = true;
+		}
+		break;
+	case IP_VERSION(9, 4, 3):
+	case IP_VERSION(9, 4, 4):
+	case IP_VERSION(9, 5, 0):
+		/* Only local VRAM BOs or system memory on non-NUMA APUs
+		 * can be assumed to be local in their entirety. Choose
+		 * MTYPE_NC as safe fallback for all system memory BOs on
+		 * NUMA systems. Their MTYPE can be overridden per-page in
+		 * gmc_v9_0_override_vm_pte_flags.
+		 */
+		mtype_local = MTYPE_RW;
+		if (amdgpu_mtype_local == 1) {
+			DRM_INFO_ONCE("Using MTYPE_NC for local memory\n");
+			mtype_local = MTYPE_NC;
+		} else if (amdgpu_mtype_local == 2) {
+			DRM_INFO_ONCE("Using MTYPE_CC for local memory\n");
+			mtype_local = MTYPE_CC;
+		} else {
+			DRM_INFO_ONCE("Using MTYPE_RW for local memory\n");
+		}
+		is_local = (!is_vram && (adev->flags & AMD_IS_APU) &&
+			    num_possible_nodes() <= 1) ||
+			   (is_vram && adev == bo_adev &&
+			    KFD_XCP_MEM_ID(adev, bo->xcp_id) == vm->mem_id);
+		snoop = true;
+		if (uncached) {
+			mtype = MTYPE_UC;
+		} else if (ext_coherent) {
+			mtype = is_local ? MTYPE_CC : MTYPE_UC;
+		} else if (adev->flags & AMD_IS_APU) {
+			mtype = is_local ? mtype_local : MTYPE_NC;
+		} else {
+			/* dGPU */
+			if (is_local)
+				mtype = mtype_local;
+			else if (gc_ip_version < IP_VERSION(9, 5, 0) && !is_vram)
+				mtype = MTYPE_UC;
+			else
+				mtype = MTYPE_NC;
+		}
+
+		break;
+	default:
+		if (uncached || coherent)
+			mtype = MTYPE_UC;
+		else
+			mtype = MTYPE_NC;
+
+		/* FIXME: is this still needed? Or does
+		 * amdgpu_ttm_tt_pde_flags already handle this?
+		 */
+		if (!is_vram)
+			snoop = true;
 	}
+
+	if (mtype != MTYPE_NC)
+		*flags = AMDGPU_PTE_MTYPE_VG10(*flags, mtype);
+
+	*flags |= snoop ? AMDGPU_PTE_SNOOPED : 0;
 }
 
 static void gmc_v9_0_get_vm_pte(struct amdgpu_device *adev,
-				struct amdgpu_bo_va_mapping *mapping,
+				struct amdgpu_vm *vm,
+				struct amdgpu_bo *bo,
+				uint32_t vm_flags,
 				uint64_t *flags)
 {
-	*flags &= ~AMDGPU_PTE_EXECUTABLE;
-	*flags |= mapping->flags & AMDGPU_PTE_EXECUTABLE;
+	if (vm_flags & AMDGPU_VM_PAGE_EXECUTABLE)
+		*flags |= AMDGPU_PTE_EXECUTABLE;
+	else
+		*flags &= ~AMDGPU_PTE_EXECUTABLE;
 
-	*flags &= ~AMDGPU_PTE_MTYPE_VG10_MASK;
-	*flags |= mapping->flags & AMDGPU_PTE_MTYPE_VG10_MASK;
+	switch (vm_flags & AMDGPU_VM_MTYPE_MASK) {
+	case AMDGPU_VM_MTYPE_DEFAULT:
+	case AMDGPU_VM_MTYPE_NC:
+	default:
+		*flags = AMDGPU_PTE_MTYPE_VG10(*flags, MTYPE_NC);
+		break;
+	case AMDGPU_VM_MTYPE_WC:
+		*flags |= AMDGPU_PTE_MTYPE_VG10(*flags, MTYPE_WC);
+		break;
+	case AMDGPU_VM_MTYPE_RW:
+		*flags |= AMDGPU_PTE_MTYPE_VG10(*flags, MTYPE_RW);
+		break;
+	case AMDGPU_VM_MTYPE_CC:
+		*flags |= AMDGPU_PTE_MTYPE_VG10(*flags, MTYPE_CC);
+		break;
+	case AMDGPU_VM_MTYPE_UC:
+		*flags |= AMDGPU_PTE_MTYPE_VG10(*flags, MTYPE_UC);
+		break;
+	}
 
-	if (mapping->flags & AMDGPU_PTE_PRT) {
+	if (vm_flags & AMDGPU_VM_PAGE_PRT) {
 		*flags |= AMDGPU_PTE_PRT;
 		*flags &= ~AMDGPU_PTE_VALID;
 	}
 
-	if ((adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 1) ||
-	     adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 2)) &&
-	    !(*flags & AMDGPU_PTE_SYSTEM) &&
-	    mapping->bo_va->is_xgmi)
-		*flags |= AMDGPU_PTE_SNOOPED;
+	if ((*flags & AMDGPU_PTE_VALID) && bo)
+		gmc_v9_0_get_coherence_flags(adev, vm, bo, vm_flags, flags);
+}
+
+static void gmc_v9_0_override_vm_pte_flags(struct amdgpu_device *adev,
+					   struct amdgpu_vm *vm,
+					   uint64_t addr, uint64_t *flags)
+{
+	int local_node, nid;
+
+	/* Only GFX 9.4.3 APUs associate GPUs with NUMA nodes. Local system
+	 * memory can use more efficient MTYPEs.
+	 */
+	if (!(adev->flags & AMD_IS_APU) ||
+	    amdgpu_ip_version(adev, GC_HWIP, 0) != IP_VERSION(9, 4, 3))
+		return;
+
+	/* Only direct-mapped memory allows us to determine the NUMA node from
+	 * the DMA address.
+	 */
+	if (!adev->ram_is_direct_mapped) {
+		dev_dbg_ratelimited(adev->dev, "RAM is not direct mapped\n");
+		return;
+	}
+
+	/* MTYPE_NC is the same default and can be overridden.
+	 * MTYPE_UC will be present if the memory is extended-coherent
+	 * and can also be overridden.
+	 */
+	if ((*flags & AMDGPU_PTE_MTYPE_VG10_MASK) !=
+	    AMDGPU_PTE_MTYPE_VG10(0ULL, MTYPE_NC) &&
+	    (*flags & AMDGPU_PTE_MTYPE_VG10_MASK) !=
+	    AMDGPU_PTE_MTYPE_VG10(0ULL, MTYPE_UC)) {
+		dev_dbg_ratelimited(adev->dev, "MTYPE is not NC or UC\n");
+		return;
+	}
+
+	/* FIXME: Only supported on native mode for now. For carve-out, the
+	 * NUMA affinity of the GPU/VM needs to come from the PCI info because
+	 * memory partitions are not associated with different NUMA nodes.
+	 */
+	if (adev->gmc.is_app_apu && vm->mem_id >= 0) {
+		local_node = adev->gmc.mem_partitions[vm->mem_id].numa.node;
+	} else {
+		dev_dbg_ratelimited(adev->dev, "Only native mode APU is supported.\n");
+		return;
+	}
 
-	if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 2))
-		*flags |= mapping->flags & AMDGPU_PTE_SNOOPED;
+	/* Only handle real RAM. Mappings of PCIe resources don't have struct
+	 * page or NUMA nodes.
+	 */
+	if (!page_is_ram(addr >> PAGE_SHIFT)) {
+		dev_dbg_ratelimited(adev->dev, "Page is not RAM.\n");
+		return;
+	}
+	nid = pfn_to_nid(addr >> PAGE_SHIFT);
+	dev_dbg_ratelimited(adev->dev, "vm->mem_id=%d, local_node=%d, nid=%d\n",
+			    vm->mem_id, local_node, nid);
+	if (nid == local_node) {
+		uint64_t old_flags = *flags;
+		if ((*flags & AMDGPU_PTE_MTYPE_VG10_MASK) ==
+			AMDGPU_PTE_MTYPE_VG10(0ULL, MTYPE_NC)) {
+			unsigned int mtype_local = MTYPE_RW;
+
+			if (amdgpu_mtype_local == 1)
+				mtype_local = MTYPE_NC;
+			else if (amdgpu_mtype_local == 2)
+				mtype_local = MTYPE_CC;
+
+			*flags = AMDGPU_PTE_MTYPE_VG10(*flags, mtype_local);
+		} else {
+			/* MTYPE_UC case */
+			*flags = AMDGPU_PTE_MTYPE_VG10(*flags, MTYPE_CC);
+		}
+
+		dev_dbg_ratelimited(adev->dev, "flags updated from %llx to %llx\n",
+				    old_flags, *flags);
+	}
 }
 
-static unsigned gmc_v9_0_get_vbios_fb_size(struct amdgpu_device *adev)
+static unsigned int gmc_v9_0_get_vbios_fb_size(struct amdgpu_device *adev)
 {
 	u32 d1vga_control = RREG32_SOC15(DCE, 0, mmD1VGA_CONTROL);
-	unsigned size;
+	unsigned int size;
+
+	/* TODO move to DC so GMC doesn't need to hard-code DCN registers */
 
 	if (REG_GET_FIELD(d1vga_control, D1VGA_CONTROL, D1VGA_MODE_ENABLE)) {
 		size = AMDGPU_VBIOS_VGA_ALLOCATION;
 	} else {
 		u32 viewport;
 
-		switch (adev->ip_versions[DCE_HWIP][0]) {
+		switch (amdgpu_ip_version(adev, DCE_HWIP, 0)) {
 		case IP_VERSION(1, 0, 0):
 		case IP_VERSION(1, 0, 1):
-		case IP_VERSION(2, 1, 0):
 			viewport = RREG32_SOC15(DCE, 0, mmHUBP0_DCSURF_PRI_VIEWPORT_DIMENSION);
 			size = (REG_GET_FIELD(viewport,
 					      HUBP0_DCSURF_PRI_VIEWPORT_DIMENSION, PRI_VIEWPORT_HEIGHT) *
@@ -1121,6 +1355,14 @@ static unsigned gmc_v9_0_get_vbios_fb_size(struct amdgpu_device *adev)
 					      HUBP0_DCSURF_PRI_VIEWPORT_DIMENSION, PRI_VIEWPORT_WIDTH) *
 				4);
 			break;
+		case IP_VERSION(2, 1, 0):
+			viewport = RREG32_SOC15(DCE, 0, mmHUBP0_DCSURF_PRI_VIEWPORT_DIMENSION_DCN2);
+			size = (REG_GET_FIELD(viewport,
+					      HUBP0_DCSURF_PRI_VIEWPORT_DIMENSION, PRI_VIEWPORT_HEIGHT) *
+				REG_GET_FIELD(viewport,
+					      HUBP0_DCSURF_PRI_VIEWPORT_DIMENSION, PRI_VIEWPORT_WIDTH) *
+				4);
+			break;
 		default:
 			viewport = RREG32_SOC15(DCE, 0, mmSCL0_VIEWPORT_SIZE);
 			size = (REG_GET_FIELD(viewport, SCL0_VIEWPORT_SIZE, VIEWPORT_HEIGHT) *
@@ -1133,15 +1375,29 @@ static unsigned gmc_v9_0_get_vbios_fb_size(struct amdgpu_device *adev)
 	return size;
 }
 
+static bool gmc_v9_0_need_reset_on_init(struct amdgpu_device *adev)
+{
+	if (adev->nbio.funcs && adev->nbio.funcs->is_nps_switch_requested &&
+	    adev->nbio.funcs->is_nps_switch_requested(adev)) {
+		adev->gmc.reset_flags |= AMDGPU_GMC_INIT_RESET_NPS;
+		return true;
+	}
+
+	return false;
+}
+
 static const struct amdgpu_gmc_funcs gmc_v9_0_gmc_funcs = {
 	.flush_gpu_tlb = gmc_v9_0_flush_gpu_tlb,
 	.flush_gpu_tlb_pasid = gmc_v9_0_flush_gpu_tlb_pasid,
 	.emit_flush_gpu_tlb = gmc_v9_0_emit_flush_gpu_tlb,
 	.emit_pasid_mapping = gmc_v9_0_emit_pasid_mapping,
-	.map_mtype = gmc_v9_0_map_mtype,
 	.get_vm_pde = gmc_v9_0_get_vm_pde,
 	.get_vm_pte = gmc_v9_0_get_vm_pte,
+	.override_vm_pte_flags = gmc_v9_0_override_vm_pte_flags,
 	.get_vbios_fb_size = gmc_v9_0_get_vbios_fb_size,
+	.query_mem_partition_mode = &amdgpu_gmc_query_memory_partition,
+	.request_mem_partition_mode = &amdgpu_gmc_request_memory_partition,
+	.need_reset_on_init = &gmc_v9_0_need_reset_on_init,
 };
 
 static void gmc_v9_0_set_gmc_funcs(struct amdgpu_device *adev)
@@ -1151,7 +1407,7 @@ static void gmc_v9_0_set_gmc_funcs(struct amdgpu_device *adev)
 
 static void gmc_v9_0_set_umc_funcs(struct amdgpu_device *adev)
 {
-	switch (adev->ip_versions[UMC_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, UMC_HWIP, 0)) {
 	case IP_VERSION(6, 0, 0):
 		adev->umc.funcs = &umc_v6_0_funcs;
 		break;
@@ -1160,29 +1416,44 @@ static void gmc_v9_0_set_umc_funcs(struct amdgpu_device *adev)
 		adev->umc.channel_inst_num = UMC_V6_1_CHANNEL_INSTANCE_NUM;
 		adev->umc.umc_inst_num = UMC_V6_1_UMC_INSTANCE_NUM;
 		adev->umc.channel_offs = UMC_V6_1_PER_CHANNEL_OFFSET_VG20;
+		adev->umc.retire_unit = 1;
 		adev->umc.channel_idx_tbl = &umc_v6_1_channel_idx_tbl[0][0];
-		adev->umc.ras_funcs = &umc_v6_1_ras_funcs;
+		adev->umc.ras = &umc_v6_1_ras;
 		break;
 	case IP_VERSION(6, 1, 2):
 		adev->umc.max_ras_err_cnt_per_query = UMC_V6_1_TOTAL_CHANNEL_NUM;
 		adev->umc.channel_inst_num = UMC_V6_1_CHANNEL_INSTANCE_NUM;
 		adev->umc.umc_inst_num = UMC_V6_1_UMC_INSTANCE_NUM;
 		adev->umc.channel_offs = UMC_V6_1_PER_CHANNEL_OFFSET_ARCT;
+		adev->umc.retire_unit = 1;
 		adev->umc.channel_idx_tbl = &umc_v6_1_channel_idx_tbl[0][0];
-		adev->umc.ras_funcs = &umc_v6_1_ras_funcs;
+		adev->umc.ras = &umc_v6_1_ras;
 		break;
 	case IP_VERSION(6, 7, 0):
-		adev->umc.max_ras_err_cnt_per_query = UMC_V6_7_TOTAL_CHANNEL_NUM;
+		adev->umc.max_ras_err_cnt_per_query =
+			UMC_V6_7_TOTAL_CHANNEL_NUM * UMC_V6_7_BAD_PAGE_NUM_PER_CHANNEL;
 		adev->umc.channel_inst_num = UMC_V6_7_CHANNEL_INSTANCE_NUM;
 		adev->umc.umc_inst_num = UMC_V6_7_UMC_INSTANCE_NUM;
 		adev->umc.channel_offs = UMC_V6_7_PER_CHANNEL_OFFSET;
+		adev->umc.retire_unit = (UMC_V6_7_NA_MAP_PA_NUM * 2);
 		if (!adev->gmc.xgmi.connected_to_cpu)
-			adev->umc.ras_funcs = &umc_v6_7_ras_funcs;
+			adev->umc.ras = &umc_v6_7_ras;
 		if (1 & adev->smuio.funcs->get_die_id(adev))
 			adev->umc.channel_idx_tbl = &umc_v6_7_channel_idx_tbl_first[0][0];
 		else
 			adev->umc.channel_idx_tbl = &umc_v6_7_channel_idx_tbl_second[0][0];
 		break;
+	case IP_VERSION(12, 0, 0):
+	case IP_VERSION(12, 5, 0):
+		adev->umc.max_ras_err_cnt_per_query =
+			UMC_V12_0_TOTAL_CHANNEL_NUM(adev) * UMC_V12_0_BAD_PAGE_NUM_PER_CHANNEL;
+		adev->umc.channel_inst_num = UMC_V12_0_CHANNEL_INSTANCE_NUM;
+		adev->umc.umc_inst_num = UMC_V12_0_UMC_INSTANCE_NUM;
+		adev->umc.node_inst_num /= UMC_V12_0_UMC_INSTANCE_NUM;
+		adev->umc.channel_offs = UMC_V12_0_PER_CHANNEL_OFFSET;
+		if (!adev->gmc.xgmi.connected_to_cpu && !adev->gmc.is_app_apu)
+			adev->umc.ras = &umc_v12_0_ras;
+		break;
 	default:
 		break;
 	}
@@ -1190,13 +1461,17 @@ static void gmc_v9_0_set_umc_funcs(struct amdgpu_device *adev)
 
 static void gmc_v9_0_set_mmhub_funcs(struct amdgpu_device *adev)
 {
-	switch (adev->ip_versions[MMHUB_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, MMHUB_HWIP, 0)) {
 	case IP_VERSION(9, 4, 1):
 		adev->mmhub.funcs = &mmhub_v9_4_funcs;
 		break;
 	case IP_VERSION(9, 4, 2):
 		adev->mmhub.funcs = &mmhub_v1_7_funcs;
 		break;
+	case IP_VERSION(1, 8, 0):
+	case IP_VERSION(1, 8, 1):
+		adev->mmhub.funcs = &mmhub_v1_8_funcs;
+		break;
 	default:
 		adev->mmhub.funcs = &mmhub_v1_0_funcs;
 		break;
@@ -1205,15 +1480,19 @@ static void gmc_v9_0_set_mmhub_funcs(struct amdgpu_device *adev)
 
 static void gmc_v9_0_set_mmhub_ras_funcs(struct amdgpu_device *adev)
 {
-	switch (adev->ip_versions[MMHUB_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, MMHUB_HWIP, 0)) {
 	case IP_VERSION(9, 4, 0):
-		adev->mmhub.ras_funcs = &mmhub_v1_0_ras_funcs;
+		adev->mmhub.ras = &mmhub_v1_0_ras;
 		break;
 	case IP_VERSION(9, 4, 1):
-		adev->mmhub.ras_funcs = &mmhub_v9_4_ras_funcs;
+		adev->mmhub.ras = &mmhub_v9_4_ras;
 		break;
 	case IP_VERSION(9, 4, 2):
-		adev->mmhub.ras_funcs = &mmhub_v1_7_ras_funcs;
+		adev->mmhub.ras = &mmhub_v1_7_ras;
+		break;
+	case IP_VERSION(1, 8, 0):
+	case IP_VERSION(1, 8, 1):
+		adev->mmhub.ras = &mmhub_v1_8_ras;
 		break;
 	default:
 		/* mmhub ras is not available */
@@ -1223,42 +1502,111 @@ static void gmc_v9_0_set_mmhub_ras_funcs(struct amdgpu_device *adev)
 
 static void gmc_v9_0_set_gfxhub_funcs(struct amdgpu_device *adev)
 {
-	adev->gfxhub.funcs = &gfxhub_v1_0_funcs;
+	if (amdgpu_is_multi_aid(adev))
+		adev->gfxhub.funcs = &gfxhub_v1_2_funcs;
+	else
+		adev->gfxhub.funcs = &gfxhub_v1_0_funcs;
 }
 
 static void gmc_v9_0_set_hdp_ras_funcs(struct amdgpu_device *adev)
 {
-	adev->hdp.ras_funcs = &hdp_v4_0_ras_funcs;
+	adev->hdp.ras = &hdp_v4_0_ras;
 }
 
-static void gmc_v9_0_set_mca_funcs(struct amdgpu_device *adev)
+static void gmc_v9_0_set_mca_ras_funcs(struct amdgpu_device *adev)
 {
+	struct amdgpu_mca *mca = &adev->mca;
+
 	/* is UMC the right IP to check for MCA?  Maybe DF? */
-	switch (adev->ip_versions[UMC_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, UMC_HWIP, 0)) {
 	case IP_VERSION(6, 7, 0):
-		if (!adev->gmc.xgmi.connected_to_cpu)
-			adev->mca.funcs = &mca_v3_0_funcs;
+		if (!adev->gmc.xgmi.connected_to_cpu) {
+			mca->mp0.ras = &mca_v3_0_mp0_ras;
+			mca->mp1.ras = &mca_v3_0_mp1_ras;
+			mca->mpio.ras = &mca_v3_0_mpio_ras;
+		}
 		break;
 	default:
 		break;
 	}
 }
 
-static int gmc_v9_0_early_init(void *handle)
+static void gmc_v9_0_set_xgmi_ras_funcs(struct amdgpu_device *adev)
+{
+	if (!adev->gmc.xgmi.connected_to_cpu)
+		adev->gmc.xgmi.ras = &xgmi_ras;
+}
+
+static void gmc_v9_0_init_nps_details(struct amdgpu_device *adev)
+{
+	enum amdgpu_memory_partition mode;
+	uint32_t supp_modes;
+	int i;
+
+	adev->gmc.supported_nps_modes = 0;
+
+	if (amdgpu_sriov_vf(adev) || (adev->flags & AMD_IS_APU))
+		return;
+
+	mode = amdgpu_gmc_get_memory_partition(adev, &supp_modes);
+
+	/* Mode detected by hardware and supported modes available */
+	if ((mode != UNKNOWN_MEMORY_PARTITION_MODE) && supp_modes) {
+		while ((i = ffs(supp_modes))) {
+			if (AMDGPU_ALL_NPS_MASK & BIT(i))
+				adev->gmc.supported_nps_modes |= BIT(i);
+			supp_modes &= supp_modes - 1;
+		}
+	} else {
+		/*TODO: Check PSP version also which supports NPS switch. Otherwise keep
+	 * supported modes as 0.
+	 */
+		switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+		case IP_VERSION(9, 4, 3):
+		case IP_VERSION(9, 4, 4):
+			adev->gmc.supported_nps_modes =
+				BIT(AMDGPU_NPS1_PARTITION_MODE) |
+				BIT(AMDGPU_NPS4_PARTITION_MODE);
+			break;
+		default:
+			break;
+		}
+	}
+}
+
+static int gmc_v9_0_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
-	/* ARCT and VEGA20 don't have XGMI defined in their IP discovery tables */
-	if (adev->asic_type == CHIP_VEGA20 ||
-	    adev->asic_type == CHIP_ARCTURUS)
+	/*
+	 * 9.4.0, 9.4.1 and 9.4.3 don't have XGMI defined
+	 * in their IP discovery tables
+	 */
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 0) ||
+	    amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 1) ||
+	    amdgpu_is_multi_aid(adev))
 		adev->gmc.xgmi.supported = true;
 
-	if (adev->ip_versions[XGMI_HWIP][0] == IP_VERSION(6, 1, 0)) {
+	if (amdgpu_ip_version(adev, XGMI_HWIP, 0) == IP_VERSION(6, 1, 0)) {
 		adev->gmc.xgmi.supported = true;
 		adev->gmc.xgmi.connected_to_cpu =
 			adev->smuio.funcs->is_host_gpu_xgmi_supported(adev);
 	}
 
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3)) {
+		enum amdgpu_pkg_type pkg_type =
+			adev->smuio.funcs->get_pkg_type(adev);
+		/* On GFXIP 9.4.3. APU, there is no physical VRAM domain present
+		 * and the APU, can be in used two possible modes:
+		 *  - carveout mode
+		 *  - native APU mode
+		 * "is_app_apu" can be used to identify the APU in the native
+		 * mode.
+		 */
+		adev->gmc.is_app_apu = (pkg_type == AMDGPU_PKG_TYPE_APU &&
+					!pci_resource_len(adev->pdev, 0));
+	}
+
 	gmc_v9_0_set_gmc_funcs(adev);
 	gmc_v9_0_set_irq_funcs(adev);
 	gmc_v9_0_set_umc_funcs(adev);
@@ -1266,7 +1614,8 @@ static int gmc_v9_0_early_init(void *handle)
 	gmc_v9_0_set_mmhub_ras_funcs(adev);
 	gmc_v9_0_set_gfxhub_funcs(adev);
 	gmc_v9_0_set_hdp_ras_funcs(adev);
-	gmc_v9_0_set_mca_funcs(adev);
+	gmc_v9_0_set_mca_ras_funcs(adev);
+	gmc_v9_0_set_xgmi_ras_funcs(adev);
 
 	adev->gmc.shared_aperture_start = 0x2000000000000000ULL;
 	adev->gmc.shared_aperture_end =
@@ -1274,13 +1623,14 @@ static int gmc_v9_0_early_init(void *handle)
 	adev->gmc.private_aperture_start = 0x1000000000000000ULL;
 	adev->gmc.private_aperture_end =
 		adev->gmc.private_aperture_start + (4ULL << 30) - 1;
+	adev->gmc.noretry_flags = AMDGPU_VM_NORETRY_FLAGS_TF;
 
 	return 0;
 }
 
-static int gmc_v9_0_late_init(void *handle)
+static int gmc_v9_0_late_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int r;
 
 	r = amdgpu_gmc_allocate_vm_inv_eng(adev);
@@ -1292,21 +1642,17 @@ static int gmc_v9_0_late_init(void *handle)
 	 * writes, while disables HBM ECC for vega10.
 	 */
 	if (!amdgpu_sriov_vf(adev) &&
-	    (adev->ip_versions[UMC_HWIP][0] == IP_VERSION(6, 0, 0))) {
+	    (amdgpu_ip_version(adev, UMC_HWIP, 0) == IP_VERSION(6, 0, 0))) {
 		if (!(adev->ras_enabled & (1 << AMDGPU_RAS_BLOCK__UMC))) {
-			if (adev->df.funcs->enable_ecc_force_par_wr_rmw)
+			if (adev->df.funcs &&
+			    adev->df.funcs->enable_ecc_force_par_wr_rmw)
 				adev->df.funcs->enable_ecc_force_par_wr_rmw(adev, false);
 		}
 	}
 
 	if (!amdgpu_persistent_edc_harvesting_supported(adev)) {
-		if (adev->mmhub.ras_funcs &&
-		    adev->mmhub.ras_funcs->reset_ras_error_count)
-			adev->mmhub.ras_funcs->reset_ras_error_count(adev);
-
-		if (adev->hdp.ras_funcs &&
-		    adev->hdp.ras_funcs->reset_ras_error_count)
-			adev->hdp.ras_funcs->reset_ras_error_count(adev);
+		amdgpu_ras_reset_error_count(adev, AMDGPU_RAS_BLOCK__MMHUB);
+		amdgpu_ras_reset_error_count(adev, AMDGPU_RAS_BLOCK__HDP);
 	}
 
 	r = amdgpu_gmc_ras_late_init(adev);
@@ -1321,14 +1667,17 @@ static void gmc_v9_0_vram_gtt_location(struct amdgpu_device *adev,
 {
 	u64 base = adev->mmhub.funcs->get_fb_location(adev);
 
+	amdgpu_gmc_set_agp_default(adev, mc);
+
 	/* add the xgmi offset of the physical node */
 	base += adev->gmc.xgmi.physical_node_id * adev->gmc.xgmi.node_segment_size;
-	if (adev->gmc.xgmi.connected_to_cpu) {
+	if (amdgpu_gmc_is_pdb0_enabled(adev)) {
 		amdgpu_gmc_sysvm_location(adev, mc);
 	} else {
 		amdgpu_gmc_vram_location(adev, mc, base);
-		amdgpu_gmc_gart_location(adev, mc);
-		amdgpu_gmc_agp_location(adev, mc);
+		amdgpu_gmc_gart_location(adev, mc, AMDGPU_GART_PLACEMENT_BEST_FIT);
+		if (!amdgpu_sriov_vf(adev) && (amdgpu_agp == 1))
+			amdgpu_gmc_agp_location(adev, mc);
 	}
 	/* base offset of vram pages */
 	adev->vm_manager.vram_base_offset = adev->gfxhub.funcs->get_mc_fb_offset(adev);
@@ -1352,8 +1701,13 @@ static int gmc_v9_0_mc_init(struct amdgpu_device *adev)
 	int r;
 
 	/* size in MB on si */
-	adev->gmc.mc_vram_size =
-		adev->nbio.funcs->get_memsize(adev) * 1024ULL * 1024ULL;
+	if (!adev->gmc.is_app_apu) {
+		adev->gmc.mc_vram_size =
+			adev->nbio.funcs->get_memsize(adev) * 1024ULL * 1024ULL;
+	} else {
+		DRM_DEBUG("Set mc_vram_size = 0 for APP APU\n");
+		adev->gmc.mc_vram_size = 0;
+	}
 	adev->gmc.real_vram_size = adev->gmc.mc_vram_size;
 
 	if (!(adev->flags & AMD_IS_APU) &&
@@ -1378,7 +1732,8 @@ static int gmc_v9_0_mc_init(struct amdgpu_device *adev)
 	 */
 
 	/* check whether both host-gpu and gpu-gpu xgmi links exist */
-	if ((adev->flags & AMD_IS_APU) ||
+	if ((!amdgpu_sriov_vf(adev) &&
+		(adev->flags & AMD_IS_APU) && !amdgpu_passthrough(adev)) ||
 	    (adev->gmc.xgmi.supported &&
 	     adev->gmc.xgmi.connected_to_cpu)) {
 		adev->gmc.aper_base =
@@ -1389,19 +1744,19 @@ static int gmc_v9_0_mc_init(struct amdgpu_device *adev)
 	}
 
 #endif
-	/* In case the PCI BAR is larger than the actual amount of vram */
 	adev->gmc.visible_vram_size = adev->gmc.aper_size;
-	if (adev->gmc.visible_vram_size > adev->gmc.real_vram_size)
-		adev->gmc.visible_vram_size = adev->gmc.real_vram_size;
 
 	/* set the gart size */
 	if (amdgpu_gart_size == -1) {
-		switch (adev->ip_versions[GC_HWIP][0]) {
+		switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 		case IP_VERSION(9, 0, 1):  /* all engines support GPUVM */
 		case IP_VERSION(9, 2, 1):  /* all engines support GPUVM */
 		case IP_VERSION(9, 4, 0):
 		case IP_VERSION(9, 4, 1):
 		case IP_VERSION(9, 4, 2):
+		case IP_VERSION(9, 4, 3):
+		case IP_VERSION(9, 4, 4):
+		case IP_VERSION(9, 5, 0):
 		default:
 			adev->gmc.gart_size = 512ULL << 20;
 			break;
@@ -1431,7 +1786,7 @@ static int gmc_v9_0_gart_init(struct amdgpu_device *adev)
 		return 0;
 	}
 
-	if (adev->gmc.xgmi.connected_to_cpu) {
+	if (amdgpu_gmc_is_pdb0_enabled(adev)) {
 		adev->gmc.vmid0_page_table_depth = 1;
 		adev->gmc.vmid0_page_table_block_size = 12;
 	} else {
@@ -1444,15 +1799,21 @@ static int gmc_v9_0_gart_init(struct amdgpu_device *adev)
 	if (r)
 		return r;
 	adev->gart.table_size = adev->gart.num_gpu_pages * 8;
-	adev->gart.gart_pte_flags = AMDGPU_PTE_MTYPE_VG10(MTYPE_UC) |
+	adev->gart.gart_pte_flags = AMDGPU_PTE_MTYPE_VG10(0ULL, MTYPE_UC) |
 				 AMDGPU_PTE_EXECUTABLE;
 
-	r = amdgpu_gart_table_vram_alloc(adev);
-	if (r)
-		return r;
+	if (!adev->gmc.real_vram_size) {
+		dev_info(adev->dev, "Put GART in system memory for APU\n");
+		r = amdgpu_gart_table_ram_alloc(adev);
+		if (r)
+			dev_err(adev->dev, "Failed to allocate GART in system memory\n");
+	} else {
+		r = amdgpu_gart_table_vram_alloc(adev);
+		if (r)
+			return r;
 
-	if (adev->gmc.xgmi.connected_to_cpu) {
-		r = amdgpu_gmc_pdb0_alloc(adev);
+		if (amdgpu_gmc_is_pdb0_enabled(adev))
+			r = amdgpu_gmc_pdb0_alloc(adev);
 	}
 
 	return r;
@@ -1468,54 +1829,89 @@ static int gmc_v9_0_gart_init(struct amdgpu_device *adev)
  */
 static void gmc_v9_0_save_registers(struct amdgpu_device *adev)
 {
-	if ((adev->ip_versions[DCE_HWIP][0] == IP_VERSION(1, 0, 0)) ||
-	    (adev->ip_versions[DCE_HWIP][0] == IP_VERSION(1, 0, 1)))
+	if ((amdgpu_ip_version(adev, DCE_HWIP, 0) == IP_VERSION(1, 0, 0)) ||
+	    (amdgpu_ip_version(adev, DCE_HWIP, 0) == IP_VERSION(1, 0, 1)))
 		adev->gmc.sdpif_register = RREG32_SOC15(DCE, 0, mmDCHUBBUB_SDPIF_MMIO_CNTRL_0);
 }
 
-static int gmc_v9_0_sw_init(void *handle)
+static void gmc_v9_4_3_init_vram_info(struct amdgpu_device *adev)
 {
-	int r, vram_width = 0, vram_type = 0, vram_vendor = 0;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	static const u32 regBIF_BIOS_SCRATCH_4 = 0x50;
+	u32 vram_info;
+
+	adev->gmc.vram_type = AMDGPU_VRAM_TYPE_HBM;
+	adev->gmc.vram_width = 128 * 64;
+
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 5, 0))
+		adev->gmc.vram_type = AMDGPU_VRAM_TYPE_HBM3E;
+
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4) &&
+		adev->rev_id == 0x3)
+		adev->gmc.vram_type = AMDGPU_VRAM_TYPE_HBM3E;
+
+	if (!(adev->flags & AMD_IS_APU) && !amdgpu_sriov_vf(adev)) {
+		vram_info = RREG32(regBIF_BIOS_SCRATCH_4);
+		adev->gmc.vram_vendor = vram_info & 0xF;
+	}
+}
+
+static int gmc_v9_0_sw_init(struct amdgpu_ip_block *ip_block)
+{
+	int r, vram_width = 0, vram_type = 0, vram_vendor = 0, dma_addr_bits;
+	struct amdgpu_device *adev = ip_block->adev;
+	unsigned long inst_mask = adev->aid_mask;
 
 	adev->gfxhub.funcs->init(adev);
 
 	adev->mmhub.funcs->init(adev);
-	if (adev->mca.funcs)
-		adev->mca.funcs->init(adev);
 
 	spin_lock_init(&adev->gmc.invalidate_lock);
 
-	r = amdgpu_atomfirmware_get_vram_info(adev,
-		&vram_width, &vram_type, &vram_vendor);
-	if (amdgpu_sriov_vf(adev))
-		/* For Vega10 SR-IOV, vram_width can't be read from ATOM as RAVEN,
-		 * and DF related registers is not readable, seems hardcord is the
-		 * only way to set the correct vram_width
-		 */
-		adev->gmc.vram_width = 2048;
-	else if (amdgpu_emu_mode != 1)
-		adev->gmc.vram_width = vram_width;
-
-	if (!adev->gmc.vram_width) {
-		int chansize, numchan;
-
-		/* hbm memory channel size */
-		if (adev->flags & AMD_IS_APU)
-			chansize = 64;
-		else
-			chansize = 128;
+	if (amdgpu_is_multi_aid(adev)) {
+		gmc_v9_4_3_init_vram_info(adev);
+	} else if (!adev->bios) {
+		if (adev->flags & AMD_IS_APU) {
+			adev->gmc.vram_type = AMDGPU_VRAM_TYPE_DDR4;
+			adev->gmc.vram_width = 64 * 64;
+		} else {
+			adev->gmc.vram_type = AMDGPU_VRAM_TYPE_HBM;
+			adev->gmc.vram_width = 128 * 64;
+		}
+	} else {
+		r = amdgpu_atomfirmware_get_vram_info(adev,
+			&vram_width, &vram_type, &vram_vendor);
+		if (amdgpu_sriov_vf(adev))
+			/* For Vega10 SR-IOV, vram_width can't be read from ATOM as RAVEN,
+			 * and DF related registers is not readable, seems hardcord is the
+			 * only way to set the correct vram_width
+			 */
+			adev->gmc.vram_width = 2048;
+		else if (amdgpu_emu_mode != 1)
+			adev->gmc.vram_width = vram_width;
+
+		if (!adev->gmc.vram_width) {
+			int chansize, numchan;
+
+			/* hbm memory channel size */
+			if (adev->flags & AMD_IS_APU)
+				chansize = 64;
+			else
+				chansize = 128;
+			if (adev->df.funcs &&
+			    adev->df.funcs->get_hbm_channel_number) {
+				numchan = adev->df.funcs->get_hbm_channel_number(adev);
+				adev->gmc.vram_width = numchan * chansize;
+			}
+		}
 
-		numchan = adev->df.funcs->get_hbm_channel_number(adev);
-		adev->gmc.vram_width = numchan * chansize;
+		adev->gmc.vram_type = vram_type;
+		adev->gmc.vram_vendor = vram_vendor;
 	}
-
-	adev->gmc.vram_type = vram_type;
-	adev->gmc.vram_vendor = vram_vendor;
-	switch (adev->ip_versions[GC_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 	case IP_VERSION(9, 1, 0):
 	case IP_VERSION(9, 2, 2):
-		adev->num_vmhubs = 2;
+		set_bit(AMDGPU_GFXHUB(0), adev->vmhubs_mask);
+		set_bit(AMDGPU_MMHUB0(0), adev->vmhubs_mask);
 
 		if (adev->rev_id == 0x0 || adev->rev_id == 0x1) {
 			amdgpu_vm_adjust_size(adev, 256 * 1024, 9, 3, 48);
@@ -1531,25 +1927,39 @@ static int gmc_v9_0_sw_init(void *handle)
 	case IP_VERSION(9, 4, 0):
 	case IP_VERSION(9, 3, 0):
 	case IP_VERSION(9, 4, 2):
-		adev->num_vmhubs = 2;
-
+		set_bit(AMDGPU_GFXHUB(0), adev->vmhubs_mask);
+		set_bit(AMDGPU_MMHUB0(0), adev->vmhubs_mask);
 
 		/*
 		 * To fulfill 4-level page support,
 		 * vm size is 256TB (48bit), maximum size of Vega10,
 		 * block size 512 (9bit)
 		 */
-		/* sriov restrict max_pfn below AMDGPU_GMC_HOLE */
-		if (amdgpu_sriov_vf(adev))
-			amdgpu_vm_adjust_size(adev, 256 * 1024, 9, 3, 47);
-		else
-			amdgpu_vm_adjust_size(adev, 256 * 1024, 9, 3, 48);
+
+		amdgpu_vm_adjust_size(adev, 256 * 1024, 9, 3, 48);
+		if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 2))
+			adev->gmc.translate_further = adev->vm_manager.num_level > 1;
 		break;
 	case IP_VERSION(9, 4, 1):
-		adev->num_vmhubs = 3;
+		set_bit(AMDGPU_GFXHUB(0), adev->vmhubs_mask);
+		set_bit(AMDGPU_MMHUB0(0), adev->vmhubs_mask);
+		set_bit(AMDGPU_MMHUB1(0), adev->vmhubs_mask);
 
 		/* Keep the vm size same with Vega20 */
 		amdgpu_vm_adjust_size(adev, 256 * 1024, 9, 3, 48);
+		adev->gmc.translate_further = adev->vm_manager.num_level > 1;
+		break;
+	case IP_VERSION(9, 4, 3):
+	case IP_VERSION(9, 4, 4):
+	case IP_VERSION(9, 5, 0):
+		bitmap_set(adev->vmhubs_mask, AMDGPU_GFXHUB(0),
+				  NUM_XCC(adev->gfx.xcc_mask));
+
+		inst_mask <<= AMDGPU_MMHUB0(0);
+		bitmap_or(adev->vmhubs_mask, adev->vmhubs_mask, &inst_mask, 32);
+
+		amdgpu_vm_adjust_size(adev, 256 * 1024, 9, 3, 48);
+		adev->gmc.translate_further = adev->vm_manager.num_level > 1;
 		break;
 	default:
 		break;
@@ -1561,7 +1971,7 @@ static int gmc_v9_0_sw_init(void *handle)
 	if (r)
 		return r;
 
-	if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 1)) {
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 1)) {
 		r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_VMC1, VMC_1_0__SRCID__VM_FAULT,
 					&adev->gmc.vm_fault);
 		if (r)
@@ -1575,7 +1985,8 @@ static int gmc_v9_0_sw_init(void *handle)
 		return r;
 
 	if (!amdgpu_sriov_vf(adev) &&
-	    !adev->gmc.xgmi.connected_to_cpu) {
+	    !adev->gmc.xgmi.connected_to_cpu &&
+	    !adev->gmc.is_app_apu) {
 		/* interrupt sent to DF. */
 		r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_DF, 0,
 				      &adev->gmc.ecc_irq);
@@ -1589,18 +2000,16 @@ static int gmc_v9_0_sw_init(void *handle)
 	 */
 	adev->gmc.mc_mask = 0xffffffffffffULL; /* 48 bit MC */
 
-	r = dma_set_mask_and_coherent(adev->dev, DMA_BIT_MASK(44));
+	dma_addr_bits = amdgpu_ip_version(adev, GC_HWIP, 0) >=
+					IP_VERSION(9, 4, 2) ?
+				48 :
+				44;
+	r = dma_set_mask_and_coherent(adev->dev, DMA_BIT_MASK(dma_addr_bits));
 	if (r) {
-		printk(KERN_WARNING "amdgpu: No suitable DMA available.\n");
+		dev_warn(adev->dev, "amdgpu: No suitable DMA available.\n");
 		return r;
 	}
-	adev->need_swiotlb = drm_need_swiotlb(44);
-
-	if (adev->gmc.xgmi.supported) {
-		r = adev->gfxhub.funcs->get_xgmi_info(adev);
-		if (r)
-			return r;
-	}
+	adev->need_swiotlb = drm_need_swiotlb(dma_addr_bits);
 
 	r = gmc_v9_0_mc_init(adev);
 	if (r)
@@ -1608,6 +2017,12 @@ static int gmc_v9_0_sw_init(void *handle)
 
 	amdgpu_gmc_get_vbios_allocations(adev);
 
+	if (amdgpu_is_multi_aid(adev)) {
+		r = amdgpu_gmc_init_mem_ranges(adev);
+		if (r)
+			return r;
+	}
+
 	/* Memory manager */
 	r = amdgpu_bo_init(adev);
 	if (r)
@@ -1617,6 +2032,7 @@ static int gmc_v9_0_sw_init(void *handle)
 	if (r)
 		return r;
 
+	gmc_v9_0_init_nps_details(adev);
 	/*
 	 * number of VMs
 	 * VMID 0 is reserved for System
@@ -1628,34 +2044,54 @@ static int gmc_v9_0_sw_init(void *handle)
 	 * for video processing.
 	 */
 	adev->vm_manager.first_kfd_vmid =
-		(adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 1) ||
-		 adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 2)) ? 3 : 8;
+		(amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 1) ||
+		 amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 2) ||
+		 amdgpu_is_multi_aid(adev)) ?
+			3 :
+			8;
 
 	amdgpu_vm_manager_init(adev);
 
 	gmc_v9_0_save_registers(adev);
 
+	r = amdgpu_gmc_ras_sw_init(adev);
+	if (r)
+		return r;
+
+	if (amdgpu_is_multi_aid(adev))
+		amdgpu_gmc_sysfs_init(adev);
+
 	return 0;
 }
 
-static int gmc_v9_0_sw_fini(void *handle)
+static int gmc_v9_0_sw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
+
+	if (amdgpu_is_multi_aid(adev))
+		amdgpu_gmc_sysfs_fini(adev);
 
 	amdgpu_gmc_ras_fini(adev);
 	amdgpu_gem_force_release(adev);
 	amdgpu_vm_manager_fini(adev);
-	amdgpu_gart_table_vram_free(adev);
-	amdgpu_bo_unref(&adev->gmc.pdb0_bo);
+	if (!adev->gmc.real_vram_size) {
+		dev_info(adev->dev, "Put GART in system memory for APU free\n");
+		amdgpu_gart_table_ram_free(adev);
+	} else {
+		amdgpu_gart_table_vram_free(adev);
+	}
+	amdgpu_bo_free_kernel(&adev->gmc.pdb0_bo, NULL, &adev->gmc.ptr_pdb0);
 	amdgpu_bo_fini(adev);
 
+	adev->gmc.num_mem_partitions = 0;
+	kfree(adev->gmc.mem_partitions);
+
 	return 0;
 }
 
 static void gmc_v9_0_init_golden_registers(struct amdgpu_device *adev)
 {
-
-	switch (adev->ip_versions[MMHUB_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, MMHUB_HWIP, 0)) {
 	case IP_VERSION(9, 0, 0):
 		if (amdgpu_sriov_vf(adev))
 			break;
@@ -1689,8 +2125,8 @@ static void gmc_v9_0_init_golden_registers(struct amdgpu_device *adev)
  */
 void gmc_v9_0_restore_registers(struct amdgpu_device *adev)
 {
-	if ((adev->ip_versions[DCE_HWIP][0] == IP_VERSION(1, 0, 0)) ||
-	    (adev->ip_versions[DCE_HWIP][0] == IP_VERSION(1, 0, 1))) {
+	if ((amdgpu_ip_version(adev, DCE_HWIP, 0) == IP_VERSION(1, 0, 0)) ||
+	    (amdgpu_ip_version(adev, DCE_HWIP, 0) == IP_VERSION(1, 0, 1))) {
 		WREG32_SOC15(DCE, 0, mmDCHUBBUB_SDPIF_MMIO_CNTRL_0, adev->gmc.sdpif_register);
 		WARN_ON(adev->gmc.sdpif_register !=
 			RREG32_SOC15(DCE, 0, mmDCHUBBUB_SDPIF_MMIO_CNTRL_0));
@@ -1706,7 +2142,7 @@ static int gmc_v9_0_gart_enable(struct amdgpu_device *adev)
 {
 	int r;
 
-	if (adev->gmc.xgmi.connected_to_cpu)
+	if (amdgpu_gmc_is_pdb0_enabled(adev))
 		amdgpu_gmc_init_pdb0(adev);
 
 	if (adev->gart.bo == NULL) {
@@ -1714,35 +2150,45 @@ static int gmc_v9_0_gart_enable(struct amdgpu_device *adev)
 		return -EINVAL;
 	}
 
-	r = amdgpu_gart_table_vram_pin(adev);
-	if (r)
-		return r;
+	amdgpu_gtt_mgr_recover(&adev->mman.gtt_mgr);
 
-	r = adev->gfxhub.funcs->gart_enable(adev);
-	if (r)
-		return r;
+	if (!adev->in_s0ix) {
+		r = adev->gfxhub.funcs->gart_enable(adev);
+		if (r)
+			return r;
+	}
 
 	r = adev->mmhub.funcs->gart_enable(adev);
 	if (r)
 		return r;
 
 	DRM_INFO("PCIE GART of %uM enabled.\n",
-		 (unsigned)(adev->gmc.gart_size >> 20));
+		 (unsigned int)(adev->gmc.gart_size >> 20));
 	if (adev->gmc.pdb0_bo)
 		DRM_INFO("PDB0 located at 0x%016llX\n",
 				(unsigned long long)amdgpu_bo_gpu_offset(adev->gmc.pdb0_bo));
 	DRM_INFO("PTB located at 0x%016llX\n",
 			(unsigned long long)amdgpu_bo_gpu_offset(adev->gart.bo));
 
-	adev->gart.ready = true;
 	return 0;
 }
 
-static int gmc_v9_0_hw_init(void *handle)
+static int gmc_v9_0_hw_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	bool value;
-	int r, i;
+	int i, r;
+
+	adev->gmc.flush_pasid_uses_kiq = true;
+
+	/* Vega20+XGMI caches PTEs in TC and TLB. Add a heavy-weight TLB flush
+	 * (type 2), which flushes both. Due to a race condition with
+	 * concurrent memory accesses using the same TLB cache line, we still
+	 * need a second TLB flush after this.
+	 */
+	adev->gmc.flush_tlb_needs_extra_type_2 =
+		amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 0) &&
+		adev->gmc.xgmi.num_physical_nodes;
 
 	/* The sequence of these two function calls matters.*/
 	gmc_v9_0_init_golden_registers(adev);
@@ -1760,7 +2206,7 @@ static int gmc_v9_0_hw_init(void *handle)
 	adev->hdp.funcs->init_registers(adev);
 
 	/* After HDP is initialized, flush HDP.*/
-	adev->hdp.funcs->flush_hdp(adev, NULL);
+	amdgpu_device_flush_hdp(adev, NULL);
 
 	if (amdgpu_vm_fault_stop == AMDGPU_VM_FAULT_STOP_ALWAYS)
 		value = false;
@@ -1768,18 +2214,27 @@ static int gmc_v9_0_hw_init(void *handle)
 		value = true;
 
 	if (!amdgpu_sriov_vf(adev)) {
-		adev->gfxhub.funcs->set_fault_enable_default(adev, value);
+		if (!adev->in_s0ix)
+			adev->gfxhub.funcs->set_fault_enable_default(adev, value);
 		adev->mmhub.funcs->set_fault_enable_default(adev, value);
 	}
-	for (i = 0; i < adev->num_vmhubs; ++i)
+	for_each_set_bit(i, adev->vmhubs_mask, AMDGPU_MAX_VMHUBS) {
+		if (adev->in_s0ix && (i == AMDGPU_GFXHUB(0)))
+			continue;
 		gmc_v9_0_flush_gpu_tlb(adev, 0, i, 0);
+	}
 
 	if (adev->umc.funcs && adev->umc.funcs->init_registers)
 		adev->umc.funcs->init_registers(adev);
 
 	r = gmc_v9_0_gart_enable(adev);
+	if (r)
+		return r;
 
-	return r;
+	if (amdgpu_emu_mode == 1)
+		return amdgpu_gmc_vram_checking(adev);
+
+	return 0;
 }
 
 /**
@@ -1791,14 +2246,14 @@ static int gmc_v9_0_hw_init(void *handle)
  */
 static void gmc_v9_0_gart_disable(struct amdgpu_device *adev)
 {
-	adev->gfxhub.funcs->gart_disable(adev);
+	if (!adev->in_s0ix)
+		adev->gfxhub.funcs->gart_disable(adev);
 	adev->mmhub.funcs->gart_disable(adev);
-	amdgpu_gart_table_vram_unpin(adev);
 }
 
-static int gmc_v9_0_hw_fini(void *handle)
+static int gmc_v9_0_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	gmc_v9_0_gart_disable(adev);
 
@@ -1816,55 +2271,70 @@ static int gmc_v9_0_hw_fini(void *handle)
 	if (adev->mmhub.funcs->update_power_gating)
 		adev->mmhub.funcs->update_power_gating(adev, false);
 
-	amdgpu_irq_put(adev, &adev->gmc.ecc_irq, 0);
-	amdgpu_irq_put(adev, &adev->gmc.vm_fault, 0);
+	/*
+	 * For minimal init, late_init is not called, hence VM fault/RAS irqs
+	 * are not enabled.
+	 */
+	if (adev->init_lvl->level != AMDGPU_INIT_LEVEL_MINIMAL_XGMI) {
+		amdgpu_irq_put(adev, &adev->gmc.vm_fault, 0);
+
+		if (adev->gmc.ecc_irq.funcs &&
+		    amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__UMC))
+			amdgpu_irq_put(adev, &adev->gmc.ecc_irq, 0);
+	}
 
 	return 0;
 }
 
-static int gmc_v9_0_suspend(void *handle)
+static int gmc_v9_0_suspend(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	return gmc_v9_0_hw_fini(adev);
+	return gmc_v9_0_hw_fini(ip_block);
 }
 
-static int gmc_v9_0_resume(void *handle)
+static int gmc_v9_0_resume(struct amdgpu_ip_block *ip_block)
 {
+	struct amdgpu_device *adev = ip_block->adev;
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
-	r = gmc_v9_0_hw_init(adev);
+	/* If a reset is done for NPS mode switch, read the memory range
+	 * information again.
+	 */
+	if (adev->gmc.reset_flags & AMDGPU_GMC_INIT_RESET_NPS) {
+		amdgpu_gmc_init_sw_mem_ranges(adev, adev->gmc.mem_partitions);
+		adev->gmc.reset_flags &= ~AMDGPU_GMC_INIT_RESET_NPS;
+	}
+
+	r = gmc_v9_0_hw_init(ip_block);
 	if (r)
 		return r;
 
-	amdgpu_vmid_reset_all(adev);
+	amdgpu_vmid_reset_all(ip_block->adev);
 
 	return 0;
 }
 
-static bool gmc_v9_0_is_idle(void *handle)
+static bool gmc_v9_0_is_idle(struct amdgpu_ip_block *ip_block)
 {
 	/* MC is always ready in GMC v9.*/
 	return true;
 }
 
-static int gmc_v9_0_wait_for_idle(void *handle)
+static int gmc_v9_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
 {
 	/* There is no need to wait for MC idle in GMC v9.*/
 	return 0;
 }
 
-static int gmc_v9_0_soft_reset(void *handle)
+static int gmc_v9_0_soft_reset(struct amdgpu_ip_block *ip_block)
 {
 	/* XXX for emulation.*/
 	return 0;
 }
 
-static int gmc_v9_0_set_clockgating_state(void *handle,
+static int gmc_v9_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					enum amd_clockgating_state state)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	adev->mmhub.funcs->set_clockgating(adev, state);
 
@@ -1873,16 +2343,16 @@ static int gmc_v9_0_set_clockgating_state(void *handle,
 	return 0;
 }
 
-static void gmc_v9_0_get_clockgating_state(void *handle, u32 *flags)
+static void gmc_v9_0_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	adev->mmhub.funcs->get_clockgating(adev, flags);
 
 	athub_v1_0_get_clockgating(adev, flags);
 }
 
-static int gmc_v9_0_set_powergating_state(void *handle,
+static int gmc_v9_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					enum amd_powergating_state state)
 {
 	return 0;
@@ -1906,8 +2376,7 @@ const struct amd_ip_funcs gmc_v9_0_ip_funcs = {
 	.get_clockgating_state = gmc_v9_0_get_clockgating_state,
 };
 
-const struct amdgpu_ip_block_version gmc_v9_0_ip_block =
-{
+const struct amdgpu_ip_block_version gmc_v9_0_ip_block = {
 	.type = AMD_IP_BLOCK_TYPE_GMC,
 	.major = 9,
 	.minor = 0,
diff --git a/drivers/gpu/drm/amd/amdgpu/hdp_v4_0.c b/drivers/gpu/drm/amd/amdgpu/hdp_v4_0.c
index eecfb1545c1e..e6c0d86d3486 100644
--- a/drivers/gpu/drm/amd/amdgpu/hdp_v4_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/hdp_v4_0.c
@@ -21,7 +21,6 @@
  *
  */
 #include "amdgpu.h"
-#include "amdgpu_atombios.h"
 #include "hdp_v4_0.h"
 #include "amdgpu_ras.h"
 
@@ -37,26 +36,21 @@
 #define HDP_MEM_POWER_CTRL__RC_MEM_POWER_LS_EN_MASK     0x00020000L
 #define mmHDP_MEM_POWER_CTRL_BASE_IDX   0
 
-static void hdp_v4_0_flush_hdp(struct amdgpu_device *adev,
-				struct amdgpu_ring *ring)
-{
-	if (!ring || !ring->funcs->emit_wreg)
-		WREG32_NO_KIQ((adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2, 0);
-	else
-		amdgpu_ring_emit_wreg(ring, (adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2, 0);
-}
-
 static void hdp_v4_0_invalidate_hdp(struct amdgpu_device *adev,
 				    struct amdgpu_ring *ring)
 {
-	if (adev->ip_versions[HDP_HWIP][0] == IP_VERSION(4, 4, 0))
+	if (amdgpu_ip_version(adev, HDP_HWIP, 0) == IP_VERSION(4, 4, 0) ||
+	    amdgpu_ip_version(adev, HDP_HWIP, 0) == IP_VERSION(4, 4, 2) ||
+	    amdgpu_ip_version(adev, HDP_HWIP, 0) == IP_VERSION(4, 4, 5))
 		return;
 
-	if (!ring || !ring->funcs->emit_wreg)
+	if (!ring || !ring->funcs->emit_wreg) {
 		WREG32_SOC15_NO_KIQ(HDP, 0, mmHDP_READ_CACHE_INVALIDATE, 1);
-	else
+		RREG32_SOC15_NO_KIQ(HDP, 0, mmHDP_READ_CACHE_INVALIDATE);
+	} else {
 		amdgpu_ring_emit_wreg(ring, SOC15_REG_OFFSET(
 			HDP, 0, mmHDP_READ_CACHE_INVALIDATE), 1);
+	}
 }
 
 static void hdp_v4_0_query_ras_error_count(struct amdgpu_device *adev,
@@ -79,7 +73,7 @@ static void hdp_v4_0_reset_ras_error_count(struct amdgpu_device *adev)
 	if (!amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__HDP))
 		return;
 
-	if (adev->ip_versions[HDP_HWIP][0] >= IP_VERSION(4, 4, 0))
+	if (amdgpu_ip_version(adev, HDP_HWIP, 0) >= IP_VERSION(4, 4, 0))
 		WREG32_SOC15(HDP, 0, mmHDP_EDC_CNT, 0);
 	else
 		/*read back hdp ras counter to reset it to 0 */
@@ -91,10 +85,10 @@ static void hdp_v4_0_update_clock_gating(struct amdgpu_device *adev,
 {
 	uint32_t def, data;
 
-	if (adev->ip_versions[HDP_HWIP][0] == IP_VERSION(4, 0, 0) ||
-	    adev->ip_versions[HDP_HWIP][0] == IP_VERSION(4, 0, 1) ||
-	    adev->ip_versions[HDP_HWIP][0] == IP_VERSION(4, 1, 1) ||
-	    adev->ip_versions[HDP_HWIP][0] == IP_VERSION(4, 1, 0)) {
+	if (amdgpu_ip_version(adev, HDP_HWIP, 0) == IP_VERSION(4, 0, 0) ||
+	    amdgpu_ip_version(adev, HDP_HWIP, 0) == IP_VERSION(4, 0, 1) ||
+	    amdgpu_ip_version(adev, HDP_HWIP, 0) == IP_VERSION(4, 1, 1) ||
+	    amdgpu_ip_version(adev, HDP_HWIP, 0) == IP_VERSION(4, 1, 0)) {
 		def = data = RREG32(SOC15_REG_OFFSET(HDP, 0, mmHDP_MEM_POWER_LS));
 
 		if (enable && (adev->cg_flags & AMD_CG_SUPPORT_HDP_LS))
@@ -124,10 +118,16 @@ static void hdp_v4_0_update_clock_gating(struct amdgpu_device *adev,
 }
 
 static void hdp_v4_0_get_clockgating_state(struct amdgpu_device *adev,
-					    u32 *flags)
+					    u64 *flags)
 {
 	int data;
 
+	if (amdgpu_ip_version(adev, HDP_HWIP, 0) == IP_VERSION(4, 4, 2) ||
+	    amdgpu_ip_version(adev, HDP_HWIP, 0) == IP_VERSION(4, 4, 5)) {
+		/* Default enabled */
+		*flags |= AMD_CG_SUPPORT_HDP_MGCG;
+		return;
+	}
 	/* AMD_CG_SUPPORT_HDP_LS */
 	data = RREG32(SOC15_REG_OFFSET(HDP, 0, mmHDP_MEM_POWER_LS));
 	if (data & HDP_MEM_POWER_LS__LS_ENABLE_MASK)
@@ -136,7 +136,7 @@ static void hdp_v4_0_get_clockgating_state(struct amdgpu_device *adev,
 
 static void hdp_v4_0_init_registers(struct amdgpu_device *adev)
 {
-	switch (adev->ip_versions[HDP_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, HDP_HWIP, 0)) {
 	case IP_VERSION(4, 2, 1):
 		WREG32_FIELD15(HDP, 0, HDP_MMHUB_CNTL, HDP_MMHUB_GCC, 1);
 		break;
@@ -144,21 +144,32 @@ static void hdp_v4_0_init_registers(struct amdgpu_device *adev)
 		break;
 	}
 
+	/* Do not program registers if VF */
+	if (amdgpu_sriov_vf(adev))
+		return;
+
 	WREG32_FIELD15(HDP, 0, HDP_MISC_CNTL, FLUSH_INVALIDATE_CACHE, 1);
 
+	if (amdgpu_ip_version(adev, HDP_HWIP, 0) == IP_VERSION(4, 4, 0))
+		WREG32_FIELD15(HDP, 0, HDP_MISC_CNTL, READ_BUFFER_WATERMARK, 2);
+
 	WREG32_SOC15(HDP, 0, mmHDP_NONSURFACE_BASE, (adev->gmc.vram_start >> 8));
 	WREG32_SOC15(HDP, 0, mmHDP_NONSURFACE_BASE_HI, (adev->gmc.vram_start >> 40));
 }
 
-const struct amdgpu_hdp_ras_funcs hdp_v4_0_ras_funcs = {
-	.ras_late_init = amdgpu_hdp_ras_late_init,
-	.ras_fini = amdgpu_hdp_ras_fini,
+struct amdgpu_ras_block_hw_ops hdp_v4_0_ras_hw_ops = {
 	.query_ras_error_count = hdp_v4_0_query_ras_error_count,
 	.reset_ras_error_count = hdp_v4_0_reset_ras_error_count,
 };
 
+struct amdgpu_hdp_ras hdp_v4_0_ras = {
+	.ras_block = {
+		.hw_ops = &hdp_v4_0_ras_hw_ops,
+	},
+};
+
 const struct amdgpu_hdp_funcs hdp_v4_0_funcs = {
-	.flush_hdp = hdp_v4_0_flush_hdp,
+	.flush_hdp = amdgpu_hdp_generic_flush,
 	.invalidate_hdp = hdp_v4_0_invalidate_hdp,
 	.update_clock_gating = hdp_v4_0_update_clock_gating,
 	.get_clock_gating_state = hdp_v4_0_get_clockgating_state,
diff --git a/drivers/gpu/drm/amd/amdgpu/hdp_v4_0.h b/drivers/gpu/drm/amd/amdgpu/hdp_v4_0.h
index dc3a1b81dd62..c44eee9282ab 100644
--- a/drivers/gpu/drm/amd/amdgpu/hdp_v4_0.h
+++ b/drivers/gpu/drm/amd/amdgpu/hdp_v4_0.h
@@ -27,6 +27,6 @@
 #include "soc15_common.h"
 
 extern const struct amdgpu_hdp_funcs hdp_v4_0_funcs;
-extern const struct amdgpu_hdp_ras_funcs hdp_v4_0_ras_funcs;
+extern struct amdgpu_hdp_ras  hdp_v4_0_ras;
 
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/hdp_v5_0.c b/drivers/gpu/drm/amd/amdgpu/hdp_v5_0.c
index 5793977953cc..8bc001dc9f63 100644
--- a/drivers/gpu/drm/amd/amdgpu/hdp_v5_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/hdp_v5_0.c
@@ -21,27 +21,18 @@
  *
  */
 #include "amdgpu.h"
-#include "amdgpu_atombios.h"
 #include "hdp_v5_0.h"
 
 #include "hdp/hdp_5_0_0_offset.h"
 #include "hdp/hdp_5_0_0_sh_mask.h"
 #include <uapi/linux/kfd_ioctl.h>
 
-static void hdp_v5_0_flush_hdp(struct amdgpu_device *adev,
-				struct amdgpu_ring *ring)
-{
-	if (!ring || !ring->funcs->emit_wreg)
-		WREG32_NO_KIQ((adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2, 0);
-	else
-		amdgpu_ring_emit_wreg(ring, (adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2, 0);
-}
-
 static void hdp_v5_0_invalidate_hdp(struct amdgpu_device *adev,
 				    struct amdgpu_ring *ring)
 {
 	if (!ring || !ring->funcs->emit_wreg) {
 		WREG32_SOC15_NO_KIQ(HDP, 0, mmHDP_READ_CACHE_INVALIDATE, 1);
+		RREG32_SOC15_NO_KIQ(HDP, 0, mmHDP_READ_CACHE_INVALIDATE);
 	} else {
 		amdgpu_ring_emit_wreg(ring, SOC15_REG_OFFSET(
 					HDP, 0, mmHDP_READ_CACHE_INVALIDATE), 1);
@@ -181,7 +172,7 @@ static void hdp_v5_0_update_clock_gating(struct amdgpu_device *adev,
 }
 
 static void hdp_v5_0_get_clockgating_state(struct amdgpu_device *adev,
-					    u32 *flags)
+					    u64 *flags)
 {
 	uint32_t tmp;
 
@@ -215,7 +206,7 @@ static void hdp_v5_0_init_registers(struct amdgpu_device *adev)
 }
 
 const struct amdgpu_hdp_funcs hdp_v5_0_funcs = {
-	.flush_hdp = hdp_v5_0_flush_hdp,
+	.flush_hdp = amdgpu_hdp_generic_flush,
 	.invalidate_hdp = hdp_v5_0_invalidate_hdp,
 	.update_clock_gating = hdp_v5_0_update_clock_gating,
 	.get_clock_gating_state = hdp_v5_0_get_clockgating_state,
diff --git a/drivers/gpu/drm/amd/amdgpu/hdp_v5_2.c b/drivers/gpu/drm/amd/amdgpu/hdp_v5_2.c
new file mode 100644
index 000000000000..40940b4ab400
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/hdp_v5_2.c
@@ -0,0 +1,206 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include "amdgpu.h"
+#include "hdp_v5_2.h"
+
+#include "hdp/hdp_5_2_1_offset.h"
+#include "hdp/hdp_5_2_1_sh_mask.h"
+#include <uapi/linux/kfd_ioctl.h>
+
+static void hdp_v5_2_flush_hdp(struct amdgpu_device *adev,
+				struct amdgpu_ring *ring)
+{
+	if (!ring || !ring->funcs->emit_wreg) {
+		WREG32_NO_KIQ((adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2,
+			0);
+		if (amdgpu_sriov_vf(adev)) {
+			/* this is fine because SR_IOV doesn't remap the register */
+			RREG32_NO_KIQ((adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2);
+		} else {
+			/* We just need to read back a register to post the write.
+			 * Reading back the remapped register causes problems on
+			 * some platforms so just read back the memory size register.
+			 */
+			if (adev->nbio.funcs->get_memsize)
+				adev->nbio.funcs->get_memsize(adev);
+		}
+	} else {
+		amdgpu_ring_emit_wreg(ring,
+			(adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2,
+			0);
+	}
+}
+
+static void hdp_v5_2_update_mem_power_gating(struct amdgpu_device *adev,
+					     bool enable)
+{
+	uint32_t hdp_clk_cntl;
+	uint32_t hdp_mem_pwr_cntl;
+
+	if (!(adev->cg_flags & (AMD_CG_SUPPORT_HDP_LS |
+				AMD_CG_SUPPORT_HDP_DS |
+				AMD_CG_SUPPORT_HDP_SD)))
+		return;
+
+	hdp_clk_cntl = RREG32_SOC15(HDP, 0, regHDP_CLK_CNTL);
+	hdp_mem_pwr_cntl = RREG32_SOC15(HDP, 0, regHDP_MEM_POWER_CTRL);
+
+	/* Before doing clock/power mode switch, forced on MEM clock */
+	hdp_clk_cntl = REG_SET_FIELD(hdp_clk_cntl, HDP_CLK_CNTL,
+				     ATOMIC_MEM_CLK_SOFT_OVERRIDE, 1);
+	hdp_clk_cntl = REG_SET_FIELD(hdp_clk_cntl, HDP_CLK_CNTL,
+				     RC_MEM_CLK_SOFT_OVERRIDE, 1);
+	WREG32_SOC15(HDP, 0, regHDP_CLK_CNTL, hdp_clk_cntl);
+
+	/* disable clock and power gating before any changing */
+	hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
+					 ATOMIC_MEM_POWER_CTRL_EN, 0);
+	hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
+					 ATOMIC_MEM_POWER_LS_EN, 0);
+	hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
+					 ATOMIC_MEM_POWER_DS_EN, 0);
+	hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
+					 ATOMIC_MEM_POWER_SD_EN, 0);
+	hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
+					 RC_MEM_POWER_CTRL_EN, 0);
+	hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
+					 RC_MEM_POWER_LS_EN, 0);
+	hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
+					 RC_MEM_POWER_DS_EN, 0);
+	hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
+					 RC_MEM_POWER_SD_EN, 0);
+	WREG32_SOC15(HDP, 0, regHDP_MEM_POWER_CTRL, hdp_mem_pwr_cntl);
+
+	/* Already disabled above. The actions below are for "enabled" only */
+	if (enable) {
+		/* only one clock gating mode (LS/DS/SD) can be enabled */
+		if (adev->cg_flags & AMD_CG_SUPPORT_HDP_SD) {
+			hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl,
+							 HDP_MEM_POWER_CTRL,
+							 ATOMIC_MEM_POWER_SD_EN, 1);
+			hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl,
+							 HDP_MEM_POWER_CTRL,
+							 RC_MEM_POWER_SD_EN, 1);
+		} else if (adev->cg_flags & AMD_CG_SUPPORT_HDP_LS) {
+			hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl,
+							 HDP_MEM_POWER_CTRL,
+							 ATOMIC_MEM_POWER_LS_EN, 1);
+			hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl,
+							 HDP_MEM_POWER_CTRL,
+							 RC_MEM_POWER_LS_EN, 1);
+		} else if (adev->cg_flags & AMD_CG_SUPPORT_HDP_DS) {
+			hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl,
+							 HDP_MEM_POWER_CTRL,
+							 ATOMIC_MEM_POWER_DS_EN, 1);
+			hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl,
+							 HDP_MEM_POWER_CTRL,
+							 RC_MEM_POWER_DS_EN, 1);
+		}
+
+		/* confirmed that ATOMIC/RC_MEM_POWER_CTRL_EN have to be set for SRAM LS/DS/SD */
+		if (adev->cg_flags & (AMD_CG_SUPPORT_HDP_LS | AMD_CG_SUPPORT_HDP_DS |
+				      AMD_CG_SUPPORT_HDP_SD)) {
+			hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
+							 ATOMIC_MEM_POWER_CTRL_EN, 1);
+			hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
+							 RC_MEM_POWER_CTRL_EN, 1);
+			WREG32_SOC15(HDP, 0, regHDP_MEM_POWER_CTRL, hdp_mem_pwr_cntl);
+		}
+	}
+
+	/* disable MEM clock override after clock/power mode changing */
+	hdp_clk_cntl = REG_SET_FIELD(hdp_clk_cntl, HDP_CLK_CNTL,
+				     ATOMIC_MEM_CLK_SOFT_OVERRIDE, 0);
+	hdp_clk_cntl = REG_SET_FIELD(hdp_clk_cntl, HDP_CLK_CNTL,
+				     RC_MEM_CLK_SOFT_OVERRIDE, 0);
+	WREG32_SOC15(HDP, 0, regHDP_CLK_CNTL, hdp_clk_cntl);
+}
+
+static void hdp_v5_2_update_medium_grain_clock_gating(struct amdgpu_device *adev,
+						      bool enable)
+{
+	uint32_t hdp_clk_cntl;
+
+	if (!(adev->cg_flags & AMD_CG_SUPPORT_HDP_MGCG))
+		return;
+
+	hdp_clk_cntl = RREG32_SOC15(HDP, 0, regHDP_CLK_CNTL);
+
+	if (enable) {
+		hdp_clk_cntl &=
+			~(uint32_t)
+			(HDP_CLK_CNTL__ATOMIC_MEM_CLK_SOFT_OVERRIDE_MASK |
+			 HDP_CLK_CNTL__RC_MEM_CLK_SOFT_OVERRIDE_MASK |
+			 HDP_CLK_CNTL__DBUS_CLK_SOFT_OVERRIDE_MASK |
+			 HDP_CLK_CNTL__DYN_CLK_SOFT_OVERRIDE_MASK |
+			 HDP_CLK_CNTL__XDP_REG_CLK_SOFT_OVERRIDE_MASK |
+			 HDP_CLK_CNTL__HDP_REG_CLK_SOFT_OVERRIDE_MASK);
+	} else {
+		hdp_clk_cntl |= HDP_CLK_CNTL__ATOMIC_MEM_CLK_SOFT_OVERRIDE_MASK |
+			HDP_CLK_CNTL__RC_MEM_CLK_SOFT_OVERRIDE_MASK |
+			HDP_CLK_CNTL__DBUS_CLK_SOFT_OVERRIDE_MASK |
+			HDP_CLK_CNTL__DYN_CLK_SOFT_OVERRIDE_MASK |
+			HDP_CLK_CNTL__XDP_REG_CLK_SOFT_OVERRIDE_MASK |
+			HDP_CLK_CNTL__HDP_REG_CLK_SOFT_OVERRIDE_MASK;
+	}
+
+	WREG32_SOC15(HDP, 0, regHDP_CLK_CNTL, hdp_clk_cntl);
+}
+
+static void hdp_v5_2_get_clockgating_state(struct amdgpu_device *adev,
+					   u64 *flags)
+{
+	uint32_t tmp;
+
+	/* AMD_CG_SUPPORT_HDP_MGCG */
+	tmp = RREG32_SOC15(HDP, 0, regHDP_CLK_CNTL);
+	if (!(tmp & (HDP_CLK_CNTL__ATOMIC_MEM_CLK_SOFT_OVERRIDE_MASK |
+		     HDP_CLK_CNTL__RC_MEM_CLK_SOFT_OVERRIDE_MASK |
+		     HDP_CLK_CNTL__DBUS_CLK_SOFT_OVERRIDE_MASK |
+		     HDP_CLK_CNTL__DYN_CLK_SOFT_OVERRIDE_MASK |
+		     HDP_CLK_CNTL__XDP_REG_CLK_SOFT_OVERRIDE_MASK |
+		     HDP_CLK_CNTL__HDP_REG_CLK_SOFT_OVERRIDE_MASK)))
+		*flags |= AMD_CG_SUPPORT_HDP_MGCG;
+
+	/* AMD_CG_SUPPORT_HDP_LS/DS/SD */
+	tmp = RREG32_SOC15(HDP, 0, regHDP_MEM_POWER_CTRL);
+	if (tmp & HDP_MEM_POWER_CTRL__ATOMIC_MEM_POWER_LS_EN_MASK)
+		*flags |= AMD_CG_SUPPORT_HDP_LS;
+	else if (tmp & HDP_MEM_POWER_CTRL__ATOMIC_MEM_POWER_DS_EN_MASK)
+		*flags |= AMD_CG_SUPPORT_HDP_DS;
+	else if (tmp & HDP_MEM_POWER_CTRL__ATOMIC_MEM_POWER_SD_EN_MASK)
+		*flags |= AMD_CG_SUPPORT_HDP_SD;
+}
+
+static void hdp_v5_2_update_clock_gating(struct amdgpu_device *adev,
+					      bool enable)
+{
+	hdp_v5_2_update_mem_power_gating(adev, enable);
+	hdp_v5_2_update_medium_grain_clock_gating(adev, enable);
+}
+
+const struct amdgpu_hdp_funcs hdp_v5_2_funcs = {
+	.flush_hdp = hdp_v5_2_flush_hdp,
+	.update_clock_gating = hdp_v5_2_update_clock_gating,
+	.get_clock_gating_state = hdp_v5_2_get_clockgating_state,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/hdp_v5_2.h b/drivers/gpu/drm/amd/amdgpu/hdp_v5_2.h
new file mode 100644
index 000000000000..cb2abc0c80ee
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/hdp_v5_2.h
@@ -0,0 +1,31 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __HDP_V5_2_H__
+#define __HDP_V5_2_H__
+
+#include "soc15_common.h"
+
+extern const struct amdgpu_hdp_funcs hdp_v5_2_funcs;
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/hdp_v6_0.c b/drivers/gpu/drm/amd/amdgpu/hdp_v6_0.c
new file mode 100644
index 000000000000..ec20daf4272c
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/hdp_v6_0.c
@@ -0,0 +1,144 @@
+/*
+ * Copyright 2020 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include "amdgpu.h"
+#include "hdp_v6_0.h"
+
+#include "hdp/hdp_6_0_0_offset.h"
+#include "hdp/hdp_6_0_0_sh_mask.h"
+#include <uapi/linux/kfd_ioctl.h>
+
+#define regHDP_CLK_CNTL_V6_1	0xd5
+#define regHDP_CLK_CNTL_V6_1_BASE_IDX 0
+
+static void hdp_v6_0_update_clock_gating(struct amdgpu_device *adev,
+					 bool enable)
+{
+	uint32_t hdp_clk_cntl;
+	uint32_t hdp_mem_pwr_cntl;
+
+	if (!(adev->cg_flags & (AMD_CG_SUPPORT_HDP_LS |
+				AMD_CG_SUPPORT_HDP_DS |
+				AMD_CG_SUPPORT_HDP_SD)))
+		return;
+
+	if (amdgpu_ip_version(adev, HDP_HWIP, 0) == IP_VERSION(6, 1, 0))
+		hdp_clk_cntl = RREG32_SOC15(HDP, 0, regHDP_CLK_CNTL_V6_1);
+	else
+		hdp_clk_cntl = RREG32_SOC15(HDP, 0, regHDP_CLK_CNTL);
+	hdp_mem_pwr_cntl = RREG32_SOC15(HDP, 0, regHDP_MEM_POWER_CTRL);
+
+	/* Before doing clock/power mode switch,
+	 * forced on IPH & RC clock */
+	hdp_clk_cntl = REG_SET_FIELD(hdp_clk_cntl, HDP_CLK_CNTL,
+				     RC_MEM_CLK_SOFT_OVERRIDE, 1);
+	if (amdgpu_ip_version(adev, HDP_HWIP, 0) == IP_VERSION(6, 1, 0))
+		WREG32_SOC15(HDP, 0, regHDP_CLK_CNTL_V6_1, hdp_clk_cntl);
+	else
+		WREG32_SOC15(HDP, 0, regHDP_CLK_CNTL, hdp_clk_cntl);
+
+	/* disable clock and power gating before any changing */
+	hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
+					 ATOMIC_MEM_POWER_CTRL_EN, 0);
+	hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
+					 ATOMIC_MEM_POWER_LS_EN, 0);
+	hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
+					 ATOMIC_MEM_POWER_DS_EN, 0);
+	hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
+					 ATOMIC_MEM_POWER_SD_EN, 0);
+	hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
+					 RC_MEM_POWER_CTRL_EN, 0);
+	hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
+					 RC_MEM_POWER_LS_EN, 0);
+	hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
+					 RC_MEM_POWER_DS_EN, 0);
+	hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
+					 RC_MEM_POWER_SD_EN, 0);
+	WREG32_SOC15(HDP, 0, regHDP_MEM_POWER_CTRL, hdp_mem_pwr_cntl);
+
+	/* Already disabled above. The actions below are for "enabled" only */
+	if (enable) {
+		/* only one clock gating mode (LS/DS/SD) can be enabled */
+		if (adev->cg_flags & AMD_CG_SUPPORT_HDP_SD) {
+			hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl,
+							 HDP_MEM_POWER_CTRL,
+							 ATOMIC_MEM_POWER_SD_EN, 1);
+			hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl,
+							 HDP_MEM_POWER_CTRL,
+							 RC_MEM_POWER_SD_EN, 1);
+		} else if (adev->cg_flags & AMD_CG_SUPPORT_HDP_LS) {
+			hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl,
+							 HDP_MEM_POWER_CTRL,
+							 ATOMIC_MEM_POWER_LS_EN, 1);
+			hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl,
+							 HDP_MEM_POWER_CTRL,
+							 RC_MEM_POWER_LS_EN, 1);
+		} else if (adev->cg_flags & AMD_CG_SUPPORT_HDP_DS) {
+			hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl,
+							 HDP_MEM_POWER_CTRL,
+							 ATOMIC_MEM_POWER_DS_EN, 1);
+			hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl,
+							 HDP_MEM_POWER_CTRL,
+							 RC_MEM_POWER_DS_EN, 1);
+		}
+
+		/* confirmed that IPH_MEM_POWER_CTRL_EN and RC_MEM_POWER_CTRL_EN have to
+		 * be set for SRAM LS/DS/SD */
+		if (adev->cg_flags & (AMD_CG_SUPPORT_HDP_LS | AMD_CG_SUPPORT_HDP_DS |
+				      AMD_CG_SUPPORT_HDP_SD)) {
+			hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
+							 ATOMIC_MEM_POWER_CTRL_EN, 1);
+			hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
+							 RC_MEM_POWER_CTRL_EN, 1);
+			WREG32_SOC15(HDP, 0, regHDP_MEM_POWER_CTRL, hdp_mem_pwr_cntl);
+		}
+	}
+
+	/* disable IPH & RC clock override after clock/power mode changing */
+	hdp_clk_cntl = REG_SET_FIELD(hdp_clk_cntl, HDP_CLK_CNTL,
+				     RC_MEM_CLK_SOFT_OVERRIDE, 0);
+	if (amdgpu_ip_version(adev, HDP_HWIP, 0) == IP_VERSION(6, 1, 0))
+		WREG32_SOC15(HDP, 0, regHDP_CLK_CNTL_V6_1, hdp_clk_cntl);
+	else
+		WREG32_SOC15(HDP, 0, regHDP_CLK_CNTL, hdp_clk_cntl);
+}
+
+static void hdp_v6_0_get_clockgating_state(struct amdgpu_device *adev,
+					    u64 *flags)
+{
+	uint32_t tmp;
+
+	/* AMD_CG_SUPPORT_HDP_LS/DS/SD */
+	tmp = RREG32_SOC15(HDP, 0, regHDP_MEM_POWER_CTRL);
+	if (tmp & HDP_MEM_POWER_CTRL__ATOMIC_MEM_POWER_LS_EN_MASK)
+		*flags |= AMD_CG_SUPPORT_HDP_LS;
+	else if (tmp & HDP_MEM_POWER_CTRL__ATOMIC_MEM_POWER_DS_EN_MASK)
+		*flags |= AMD_CG_SUPPORT_HDP_DS;
+	else if (tmp & HDP_MEM_POWER_CTRL__ATOMIC_MEM_POWER_SD_EN_MASK)
+		*flags |= AMD_CG_SUPPORT_HDP_SD;
+}
+
+const struct amdgpu_hdp_funcs hdp_v6_0_funcs = {
+	.flush_hdp = amdgpu_hdp_generic_flush,
+	.update_clock_gating = hdp_v6_0_update_clock_gating,
+	.get_clock_gating_state = hdp_v6_0_get_clockgating_state,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/hdp_v6_0.h b/drivers/gpu/drm/amd/amdgpu/hdp_v6_0.h
new file mode 100644
index 000000000000..533ecd8c0800
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/hdp_v6_0.h
@@ -0,0 +1,31 @@
+/*
+ * Copyright 2020 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __HDP_V6_0_H__
+#define __HDP_V6_0_H__
+
+#include "soc15_common.h"
+
+extern const struct amdgpu_hdp_funcs hdp_v6_0_funcs;
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/hdp_v7_0.c b/drivers/gpu/drm/amd/amdgpu/hdp_v7_0.c
new file mode 100644
index 000000000000..ed1debc03507
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/hdp_v7_0.c
@@ -0,0 +1,132 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include "amdgpu.h"
+#include "hdp_v7_0.h"
+
+#include "hdp/hdp_7_0_0_offset.h"
+#include "hdp/hdp_7_0_0_sh_mask.h"
+#include <uapi/linux/kfd_ioctl.h>
+
+static void hdp_v7_0_update_clock_gating(struct amdgpu_device *adev,
+					 bool enable)
+{
+	uint32_t hdp_clk_cntl, hdp_clk_cntl1;
+	uint32_t hdp_mem_pwr_cntl;
+
+	if (!(adev->cg_flags & (AMD_CG_SUPPORT_HDP_LS |
+				AMD_CG_SUPPORT_HDP_DS |
+				AMD_CG_SUPPORT_HDP_SD)))
+		return;
+
+	hdp_clk_cntl = hdp_clk_cntl1 = RREG32_SOC15(HDP, 0,regHDP_CLK_CNTL);
+	hdp_mem_pwr_cntl = RREG32_SOC15(HDP, 0, regHDP_MEM_POWER_CTRL);
+
+	/* Before doing clock/power mode switch,
+	 * forced on IPH & RC clock */
+	hdp_clk_cntl = REG_SET_FIELD(hdp_clk_cntl, HDP_CLK_CNTL,
+				     RC_MEM_CLK_SOFT_OVERRIDE, 1);
+	WREG32_SOC15(HDP, 0, regHDP_CLK_CNTL, hdp_clk_cntl);
+
+	/* disable clock and power gating before any changing */
+	hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
+					 ATOMIC_MEM_POWER_CTRL_EN, 0);
+	hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
+					 ATOMIC_MEM_POWER_LS_EN, 0);
+	hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
+					 ATOMIC_MEM_POWER_DS_EN, 0);
+	hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
+					 ATOMIC_MEM_POWER_SD_EN, 0);
+	hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
+					 RC_MEM_POWER_CTRL_EN, 0);
+	hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
+					 RC_MEM_POWER_LS_EN, 0);
+	hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
+					 RC_MEM_POWER_DS_EN, 0);
+	hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
+					 RC_MEM_POWER_SD_EN, 0);
+	WREG32_SOC15(HDP, 0, regHDP_MEM_POWER_CTRL, hdp_mem_pwr_cntl);
+
+	/* Already disabled above. The actions below are for "enabled" only */
+	if (enable) {
+		/* only one clock gating mode (LS/DS/SD) can be enabled */
+		if (adev->cg_flags & AMD_CG_SUPPORT_HDP_SD) {
+			hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl,
+							 HDP_MEM_POWER_CTRL,
+							 ATOMIC_MEM_POWER_SD_EN, 1);
+			hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl,
+							 HDP_MEM_POWER_CTRL,
+							 RC_MEM_POWER_SD_EN, 1);
+		} else if (adev->cg_flags & AMD_CG_SUPPORT_HDP_LS) {
+			hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl,
+							 HDP_MEM_POWER_CTRL,
+							 ATOMIC_MEM_POWER_LS_EN, 1);
+			hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl,
+							 HDP_MEM_POWER_CTRL,
+							 RC_MEM_POWER_LS_EN, 1);
+		} else if (adev->cg_flags & AMD_CG_SUPPORT_HDP_DS) {
+			hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl,
+							 HDP_MEM_POWER_CTRL,
+							 ATOMIC_MEM_POWER_DS_EN, 1);
+			hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl,
+							 HDP_MEM_POWER_CTRL,
+							 RC_MEM_POWER_DS_EN, 1);
+		}
+
+		/* confirmed that IPH_MEM_POWER_CTRL_EN and RC_MEM_POWER_CTRL_EN have to
+		 * be set for SRAM LS/DS/SD */
+		if (adev->cg_flags & (AMD_CG_SUPPORT_HDP_LS | AMD_CG_SUPPORT_HDP_DS |
+				      AMD_CG_SUPPORT_HDP_SD)) {
+			hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
+							 ATOMIC_MEM_POWER_CTRL_EN, 1);
+			hdp_mem_pwr_cntl = REG_SET_FIELD(hdp_mem_pwr_cntl, HDP_MEM_POWER_CTRL,
+							 RC_MEM_POWER_CTRL_EN, 1);
+			WREG32_SOC15(HDP, 0, regHDP_MEM_POWER_CTRL, hdp_mem_pwr_cntl);
+		}
+	}
+
+	/* disable IPH & RC clock override after clock/power mode changing */
+	hdp_clk_cntl = REG_SET_FIELD(hdp_clk_cntl, HDP_CLK_CNTL,
+				     RC_MEM_CLK_SOFT_OVERRIDE, 0);
+	WREG32_SOC15(HDP, 0, regHDP_CLK_CNTL, hdp_clk_cntl);
+}
+
+static void hdp_v7_0_get_clockgating_state(struct amdgpu_device *adev,
+					    u64 *flags)
+{
+	uint32_t tmp;
+
+	/* AMD_CG_SUPPORT_HDP_LS/DS/SD */
+	tmp = RREG32_SOC15(HDP, 0, regHDP_MEM_POWER_CTRL);
+	if (tmp & HDP_MEM_POWER_CTRL__ATOMIC_MEM_POWER_LS_EN_MASK)
+		*flags |= AMD_CG_SUPPORT_HDP_LS;
+	else if (tmp & HDP_MEM_POWER_CTRL__ATOMIC_MEM_POWER_DS_EN_MASK)
+		*flags |= AMD_CG_SUPPORT_HDP_DS;
+	else if (tmp & HDP_MEM_POWER_CTRL__ATOMIC_MEM_POWER_SD_EN_MASK)
+		*flags |= AMD_CG_SUPPORT_HDP_SD;
+}
+
+const struct amdgpu_hdp_funcs hdp_v7_0_funcs = {
+	.flush_hdp = amdgpu_hdp_generic_flush,
+	.update_clock_gating = hdp_v7_0_update_clock_gating,
+	.get_clock_gating_state = hdp_v7_0_get_clockgating_state,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/hdp_v7_0.h b/drivers/gpu/drm/amd/amdgpu/hdp_v7_0.h
new file mode 100644
index 000000000000..25b69201402d
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/hdp_v7_0.h
@@ -0,0 +1,31 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __HDP_V7_0_H__
+#define __HDP_V7_0_H__
+
+#include "soc15_common.h"
+
+extern const struct amdgpu_hdp_funcs hdp_v7_0_funcs;
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/iceland_ih.c b/drivers/gpu/drm/amd/amdgpu/iceland_ih.c
index ddfe4eaeea05..01cadf898c00 100644
--- a/drivers/gpu/drm/amd/amdgpu/iceland_ih.c
+++ b/drivers/gpu/drm/amd/amdgpu/iceland_ih.c
@@ -157,6 +157,9 @@ static int iceland_ih_irq_init(struct amdgpu_device *adev)
 	/* enable interrupts */
 	iceland_ih_enable_interrupts(adev);
 
+	if (adev->irq.ih_soft.ring_size)
+		adev->irq.ih_soft.enabled = true;
+
 	return 0;
 }
 
@@ -194,6 +197,9 @@ static u32 iceland_ih_get_wptr(struct amdgpu_device *adev,
 
 	wptr = le32_to_cpu(*ih->wptr_cpu);
 
+	if (ih == &adev->irq.ih_soft)
+		goto out;
+
 	if (!REG_GET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW))
 		goto out;
 
@@ -215,6 +221,11 @@ static u32 iceland_ih_get_wptr(struct amdgpu_device *adev,
 	tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 1);
 	WREG32(mmIH_RB_CNTL, tmp);
 
+	/* Unset the CLEAR_OVERFLOW bit immediately so new overflows
+	 * can be detected.
+	 */
+	tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 0);
+	WREG32(mmIH_RB_CNTL, tmp);
 
 out:
 	return (wptr & ih->ptr_mask);
@@ -268,9 +279,9 @@ static void iceland_ih_set_rptr(struct amdgpu_device *adev,
 	WREG32(mmIH_RB_RPTR, ih->rptr);
 }
 
-static int iceland_ih_early_init(void *handle)
+static int iceland_ih_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int ret;
 
 	ret = amdgpu_irq_add_domain(adev);
@@ -282,23 +293,27 @@ static int iceland_ih_early_init(void *handle)
 	return 0;
 }
 
-static int iceland_ih_sw_init(void *handle)
+static int iceland_ih_sw_init(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	r = amdgpu_ih_ring_init(adev, &adev->irq.ih, 64 * 1024, false);
 	if (r)
 		return r;
 
+	r = amdgpu_ih_ring_init(adev, &adev->irq.ih_soft, IH_SW_RING_SIZE, true);
+	if (r)
+		return r;
+
 	r = amdgpu_irq_init(adev);
 
 	return r;
 }
 
-static int iceland_ih_sw_fini(void *handle)
+static int iceland_ih_sw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	amdgpu_irq_fini_sw(adev);
 	amdgpu_irq_remove_domain(adev);
@@ -306,44 +321,33 @@ static int iceland_ih_sw_fini(void *handle)
 	return 0;
 }
 
-static int iceland_ih_hw_init(void *handle)
+static int iceland_ih_hw_init(struct amdgpu_ip_block *ip_block)
 {
-	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	r = iceland_ih_irq_init(adev);
-	if (r)
-		return r;
+	struct amdgpu_device *adev = ip_block->adev;
 
-	return 0;
+	return iceland_ih_irq_init(adev);
 }
 
-static int iceland_ih_hw_fini(void *handle)
+static int iceland_ih_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	iceland_ih_irq_disable(adev);
+	iceland_ih_irq_disable(ip_block->adev);
 
 	return 0;
 }
 
-static int iceland_ih_suspend(void *handle)
+static int iceland_ih_suspend(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	return iceland_ih_hw_fini(adev);
+	return iceland_ih_hw_fini(ip_block);
 }
 
-static int iceland_ih_resume(void *handle)
+static int iceland_ih_resume(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	return iceland_ih_hw_init(adev);
+	return iceland_ih_hw_init(ip_block);
 }
 
-static bool iceland_ih_is_idle(void *handle)
+static bool iceland_ih_is_idle(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	u32 tmp = RREG32(mmSRBM_STATUS);
 
 	if (REG_GET_FIELD(tmp, SRBM_STATUS, IH_BUSY))
@@ -352,11 +356,11 @@ static bool iceland_ih_is_idle(void *handle)
 	return true;
 }
 
-static int iceland_ih_wait_for_idle(void *handle)
+static int iceland_ih_wait_for_idle(struct amdgpu_ip_block *ip_block)
 {
 	unsigned i;
 	u32 tmp;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	for (i = 0; i < adev->usec_timeout; i++) {
 		/* read MC_STATUS */
@@ -368,10 +372,10 @@ static int iceland_ih_wait_for_idle(void *handle)
 	return -ETIMEDOUT;
 }
 
-static int iceland_ih_soft_reset(void *handle)
+static int iceland_ih_soft_reset(struct amdgpu_ip_block *ip_block)
 {
 	u32 srbm_soft_reset = 0;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	u32 tmp = RREG32(mmSRBM_STATUS);
 
 	if (tmp & SRBM_STATUS__IH_BUSY_MASK)
@@ -398,13 +402,13 @@ static int iceland_ih_soft_reset(void *handle)
 	return 0;
 }
 
-static int iceland_ih_set_clockgating_state(void *handle,
+static int iceland_ih_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_clockgating_state state)
 {
 	return 0;
 }
 
-static int iceland_ih_set_powergating_state(void *handle,
+static int iceland_ih_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_powergating_state state)
 {
 	return 0;
@@ -413,7 +417,6 @@ static int iceland_ih_set_powergating_state(void *handle,
 static const struct amd_ip_funcs iceland_ih_ip_funcs = {
 	.name = "iceland_ih",
 	.early_init = iceland_ih_early_init,
-	.late_init = NULL,
 	.sw_init = iceland_ih_sw_init,
 	.sw_fini = iceland_ih_sw_fini,
 	.hw_init = iceland_ih_hw_init,
diff --git a/drivers/gpu/drm/amd/amdgpu/ih_v6_0.c b/drivers/gpu/drm/amd/amdgpu/ih_v6_0.c
new file mode 100644
index 000000000000..333e9c30c091
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/ih_v6_0.c
@@ -0,0 +1,817 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/pci.h>
+
+#include "amdgpu.h"
+#include "amdgpu_ih.h"
+
+#include "oss/osssys_6_0_0_offset.h"
+#include "oss/osssys_6_0_0_sh_mask.h"
+
+#include "soc15_common.h"
+#include "ih_v6_0.h"
+
+#define MAX_REARM_RETRY 10
+
+static void ih_v6_0_set_interrupt_funcs(struct amdgpu_device *adev);
+
+/**
+ * ih_v6_0_init_register_offset - Initialize register offset for ih rings
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Initialize register offset ih rings (IH_V6_0).
+ */
+static void ih_v6_0_init_register_offset(struct amdgpu_device *adev)
+{
+	struct amdgpu_ih_regs *ih_regs;
+
+	/* ih ring 2 is removed
+	 * ih ring and ih ring 1 are available */
+	if (adev->irq.ih.ring_size) {
+		ih_regs = &adev->irq.ih.ih_regs;
+		ih_regs->ih_rb_base = SOC15_REG_OFFSET(OSSSYS, 0, regIH_RB_BASE);
+		ih_regs->ih_rb_base_hi = SOC15_REG_OFFSET(OSSSYS, 0, regIH_RB_BASE_HI);
+		ih_regs->ih_rb_cntl = SOC15_REG_OFFSET(OSSSYS, 0, regIH_RB_CNTL);
+		ih_regs->ih_rb_wptr = SOC15_REG_OFFSET(OSSSYS, 0, regIH_RB_WPTR);
+		ih_regs->ih_rb_rptr = SOC15_REG_OFFSET(OSSSYS, 0, regIH_RB_RPTR);
+		ih_regs->ih_doorbell_rptr = SOC15_REG_OFFSET(OSSSYS, 0, regIH_DOORBELL_RPTR);
+		ih_regs->ih_rb_wptr_addr_lo = SOC15_REG_OFFSET(OSSSYS, 0, regIH_RB_WPTR_ADDR_LO);
+		ih_regs->ih_rb_wptr_addr_hi = SOC15_REG_OFFSET(OSSSYS, 0, regIH_RB_WPTR_ADDR_HI);
+		ih_regs->psp_reg_id = PSP_REG_IH_RB_CNTL;
+	}
+
+	if (adev->irq.ih1.ring_size) {
+		ih_regs = &adev->irq.ih1.ih_regs;
+		ih_regs->ih_rb_base = SOC15_REG_OFFSET(OSSSYS, 0, regIH_RB_BASE_RING1);
+		ih_regs->ih_rb_base_hi = SOC15_REG_OFFSET(OSSSYS, 0, regIH_RB_BASE_HI_RING1);
+		ih_regs->ih_rb_cntl = SOC15_REG_OFFSET(OSSSYS, 0, regIH_RB_CNTL_RING1);
+		ih_regs->ih_rb_wptr = SOC15_REG_OFFSET(OSSSYS, 0, regIH_RB_WPTR_RING1);
+		ih_regs->ih_rb_rptr = SOC15_REG_OFFSET(OSSSYS, 0, regIH_RB_RPTR_RING1);
+		ih_regs->ih_doorbell_rptr = SOC15_REG_OFFSET(OSSSYS, 0, regIH_DOORBELL_RPTR_RING1);
+		ih_regs->psp_reg_id = PSP_REG_IH_RB_CNTL_RING1;
+	}
+}
+
+/**
+ * force_update_wptr_for_self_int - Force update the wptr for self interrupt
+ *
+ * @adev: amdgpu_device pointer
+ * @threshold: threshold to trigger the wptr reporting
+ * @timeout: timeout to trigger the wptr reporting
+ * @enabled: Enable/disable timeout flush mechanism
+ *
+ * threshold input range: 0 ~ 15, default 0,
+ * real_threshold = 2^threshold
+ * timeout input range: 0 ~ 20, default 8,
+ * real_timeout = (2^timeout) * 1024 / (socclk_freq)
+ *
+ * Force update wptr for self interrupt ( >= SIENNA_CICHLID).
+ */
+static void
+force_update_wptr_for_self_int(struct amdgpu_device *adev,
+			       u32 threshold, u32 timeout, bool enabled)
+{
+	u32 ih_cntl, ih_rb_cntl;
+
+	ih_cntl = RREG32_SOC15(OSSSYS, 0, regIH_CNTL2);
+	ih_rb_cntl = RREG32_SOC15(OSSSYS, 0, regIH_RB_CNTL_RING1);
+
+	ih_cntl = REG_SET_FIELD(ih_cntl, IH_CNTL2,
+				SELF_IV_FORCE_WPTR_UPDATE_TIMEOUT, timeout);
+	ih_cntl = REG_SET_FIELD(ih_cntl, IH_CNTL2,
+				SELF_IV_FORCE_WPTR_UPDATE_ENABLE, enabled);
+	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL_RING1,
+				   RB_USED_INT_THRESHOLD, threshold);
+
+	if (amdgpu_sriov_vf(adev) && amdgpu_sriov_reg_indirect_ih(adev)) {
+		if (psp_reg_program(&adev->psp, PSP_REG_IH_RB_CNTL_RING1, ih_rb_cntl))
+			return;
+	} else {
+		WREG32_SOC15(OSSSYS, 0, regIH_RB_CNTL_RING1, ih_rb_cntl);
+	}
+
+	WREG32_SOC15(OSSSYS, 0, regIH_CNTL2, ih_cntl);
+}
+
+/**
+ * ih_v6_0_toggle_ring_interrupts - toggle the interrupt ring buffer
+ *
+ * @adev: amdgpu_device pointer
+ * @ih: amdgpu_ih_ring pointer
+ * @enable: true - enable the interrupts, false - disable the interrupts
+ *
+ * Toggle the interrupt ring buffer (IH_V6_0)
+ */
+static int ih_v6_0_toggle_ring_interrupts(struct amdgpu_device *adev,
+					  struct amdgpu_ih_ring *ih,
+					  bool enable)
+{
+	struct amdgpu_ih_regs *ih_regs;
+	uint32_t tmp;
+
+	ih_regs = &ih->ih_regs;
+
+	tmp = RREG32(ih_regs->ih_rb_cntl);
+	tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, RB_ENABLE, (enable ? 1 : 0));
+
+	if (enable) {
+		/* Unset the CLEAR_OVERFLOW bit to make sure the next step
+		 * is switching the bit from 0 to 1
+		 */
+		tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 0);
+		if (amdgpu_sriov_vf(adev) && amdgpu_sriov_reg_indirect_ih(adev)) {
+			if (psp_reg_program(&adev->psp, ih_regs->psp_reg_id, tmp))
+				return -ETIMEDOUT;
+		} else {
+			WREG32_NO_KIQ(ih_regs->ih_rb_cntl, tmp);
+		}
+
+		/* Clear RB_OVERFLOW bit */
+		tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 1);
+		if (amdgpu_sriov_vf(adev) && amdgpu_sriov_reg_indirect_ih(adev)) {
+			if (psp_reg_program(&adev->psp, ih_regs->psp_reg_id, tmp))
+				return -ETIMEDOUT;
+		} else {
+			WREG32_NO_KIQ(ih_regs->ih_rb_cntl, tmp);
+		}
+
+		/* Unset the CLEAR_OVERFLOW bit immediately so new overflows
+		 * can be detected.
+		 */
+		tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 0);
+	}
+
+	/* enable_intr field is only valid in ring0 */
+	if (ih == &adev->irq.ih)
+		tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, ENABLE_INTR, (enable ? 1 : 0));
+
+	if (amdgpu_sriov_vf(adev) && amdgpu_sriov_reg_indirect_ih(adev)) {
+		if (psp_reg_program(&adev->psp, ih_regs->psp_reg_id, tmp))
+			return -ETIMEDOUT;
+	} else {
+		WREG32(ih_regs->ih_rb_cntl, tmp);
+	}
+
+	if (enable) {
+		ih->enabled = true;
+	} else {
+		/* set rptr, wptr to 0 */
+		WREG32(ih_regs->ih_rb_rptr, 0);
+		WREG32(ih_regs->ih_rb_wptr, 0);
+		ih->enabled = false;
+		ih->rptr = 0;
+	}
+
+	return 0;
+}
+
+/**
+ * ih_v6_0_toggle_interrupts - Toggle all the available interrupt ring buffers
+ *
+ * @adev: amdgpu_device pointer
+ * @enable: enable or disable interrupt ring buffers
+ *
+ * Toggle all the available interrupt ring buffers (IH_V6_0).
+ */
+static int ih_v6_0_toggle_interrupts(struct amdgpu_device *adev, bool enable)
+{
+	struct amdgpu_ih_ring *ih[] = {&adev->irq.ih, &adev->irq.ih1};
+	int i;
+	int r;
+
+	for (i = 0; i < ARRAY_SIZE(ih); i++) {
+		if (ih[i]->ring_size) {
+			r = ih_v6_0_toggle_ring_interrupts(adev, ih[i], enable);
+			if (r)
+				return r;
+		}
+	}
+
+	return 0;
+}
+
+static uint32_t ih_v6_0_rb_cntl(struct amdgpu_ih_ring *ih, uint32_t ih_rb_cntl)
+{
+	int rb_bufsz = order_base_2(ih->ring_size / 4);
+
+	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL,
+				   MC_SPACE, ih->use_bus_addr ? 2 : 4);
+	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL,
+				   WPTR_OVERFLOW_CLEAR, 1);
+	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL,
+				   WPTR_OVERFLOW_ENABLE, 1);
+	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, RB_SIZE, rb_bufsz);
+	/* Ring Buffer write pointer writeback. If enabled, IH_RB_WPTR register
+	 * value is written to memory
+	 */
+	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL,
+				   WPTR_WRITEBACK_ENABLE, 1);
+	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, MC_SNOOP, 1);
+	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, MC_RO, 0);
+	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, MC_VMID, 0);
+
+	return ih_rb_cntl;
+}
+
+static uint32_t ih_v6_0_doorbell_rptr(struct amdgpu_ih_ring *ih)
+{
+	u32 ih_doorbell_rtpr = 0;
+
+	if (ih->use_doorbell) {
+		ih_doorbell_rtpr = REG_SET_FIELD(ih_doorbell_rtpr,
+						 IH_DOORBELL_RPTR, OFFSET,
+						 ih->doorbell_index);
+		ih_doorbell_rtpr = REG_SET_FIELD(ih_doorbell_rtpr,
+						 IH_DOORBELL_RPTR,
+						 ENABLE, 1);
+	} else {
+		ih_doorbell_rtpr = REG_SET_FIELD(ih_doorbell_rtpr,
+						 IH_DOORBELL_RPTR,
+						 ENABLE, 0);
+	}
+	return ih_doorbell_rtpr;
+}
+
+/**
+ * ih_v6_0_enable_ring - enable an ih ring buffer
+ *
+ * @adev: amdgpu_device pointer
+ * @ih: amdgpu_ih_ring pointer
+ *
+ * Enable an ih ring buffer (IH_V6_0)
+ */
+static int ih_v6_0_enable_ring(struct amdgpu_device *adev,
+				      struct amdgpu_ih_ring *ih)
+{
+	struct amdgpu_ih_regs *ih_regs;
+	uint32_t tmp;
+
+	ih_regs = &ih->ih_regs;
+
+	/* Ring Buffer base. [39:8] of 40-bit address of the beginning of the ring buffer*/
+	WREG32(ih_regs->ih_rb_base, ih->gpu_addr >> 8);
+	WREG32(ih_regs->ih_rb_base_hi, (ih->gpu_addr >> 40) & 0xff);
+
+	tmp = RREG32(ih_regs->ih_rb_cntl);
+	tmp = ih_v6_0_rb_cntl(ih, tmp);
+	if (ih == &adev->irq.ih)
+		tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, RPTR_REARM, !!adev->irq.msi_enabled);
+	if (ih == &adev->irq.ih1) {
+		tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_ENABLE, 0);
+		tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, RB_FULL_DRAIN_ENABLE, 1);
+	}
+
+	if (amdgpu_sriov_vf(adev) && amdgpu_sriov_reg_indirect_ih(adev)) {
+		if (psp_reg_program(&adev->psp, ih_regs->psp_reg_id, tmp)) {
+			DRM_ERROR("PSP program IH_RB_CNTL failed!\n");
+			return -ETIMEDOUT;
+		}
+	} else {
+		WREG32(ih_regs->ih_rb_cntl, tmp);
+	}
+
+	if (ih == &adev->irq.ih) {
+		/* set the ih ring 0 writeback address whether it's enabled or not */
+		WREG32(ih_regs->ih_rb_wptr_addr_lo, lower_32_bits(ih->wptr_addr));
+		WREG32(ih_regs->ih_rb_wptr_addr_hi, upper_32_bits(ih->wptr_addr) & 0xFFFF);
+	}
+
+	/* set rptr, wptr to 0 */
+	WREG32(ih_regs->ih_rb_wptr, 0);
+	WREG32(ih_regs->ih_rb_rptr, 0);
+
+	WREG32(ih_regs->ih_doorbell_rptr, ih_v6_0_doorbell_rptr(ih));
+
+	return 0;
+}
+
+/**
+ * ih_v6_0_irq_init - init and enable the interrupt ring
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Allocate a ring buffer for the interrupt controller,
+ * enable the RLC, disable interrupts, enable the IH
+ * ring buffer and enable it.
+ * Called at device load and reume.
+ * Returns 0 for success, errors for failure.
+ */
+static int ih_v6_0_irq_init(struct amdgpu_device *adev)
+{
+	struct amdgpu_ih_ring *ih[] = {&adev->irq.ih, &adev->irq.ih1};
+	u32 ih_chicken;
+	u32 tmp;
+	int ret;
+	int i;
+
+	/* disable irqs */
+	ret = ih_v6_0_toggle_interrupts(adev, false);
+	if (ret)
+		return ret;
+
+	adev->nbio.funcs->ih_control(adev);
+
+	if (unlikely((adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT) ||
+		     (adev->firmware.load_type == AMDGPU_FW_LOAD_RLC_BACKDOOR_AUTO))) {
+		if (ih[0]->use_bus_addr) {
+			ih_chicken = RREG32_SOC15(OSSSYS, 0, regIH_CHICKEN);
+			ih_chicken = REG_SET_FIELD(ih_chicken,
+					IH_CHICKEN, MC_SPACE_GPA_ENABLE, 1);
+			WREG32_SOC15(OSSSYS, 0, regIH_CHICKEN, ih_chicken);
+		}
+	}
+
+	for (i = 0; i < ARRAY_SIZE(ih); i++) {
+		if (ih[i]->ring_size) {
+			ret = ih_v6_0_enable_ring(adev, ih[i]);
+			if (ret)
+				return ret;
+		}
+		ih[i]->overflow = false;
+	}
+
+	/* update doorbell range for ih ring 0 */
+	adev->nbio.funcs->ih_doorbell_range(adev, ih[0]->use_doorbell,
+					    ih[0]->doorbell_index);
+
+	tmp = RREG32_SOC15(OSSSYS, 0, regIH_STORM_CLIENT_LIST_CNTL);
+	tmp = REG_SET_FIELD(tmp, IH_STORM_CLIENT_LIST_CNTL,
+			    CLIENT18_IS_STORM_CLIENT, 1);
+	WREG32_SOC15(OSSSYS, 0, regIH_STORM_CLIENT_LIST_CNTL, tmp);
+
+	tmp = RREG32_SOC15(OSSSYS, 0, regIH_INT_FLOOD_CNTL);
+	tmp = REG_SET_FIELD(tmp, IH_INT_FLOOD_CNTL, FLOOD_CNTL_ENABLE, 1);
+	WREG32_SOC15(OSSSYS, 0, regIH_INT_FLOOD_CNTL, tmp);
+
+	/* GC/MMHUB UTCL2 page fault interrupts are configured as
+	 * MSI storm capable interrupts by deafult. The delay is
+	 * used to avoid ISR being called too frequently
+	 * when page fault happens on several continuous page
+	 * and thus avoid MSI storm */
+	tmp = RREG32_SOC15(OSSSYS, 0, regIH_MSI_STORM_CTRL);
+	tmp = REG_SET_FIELD(tmp, IH_MSI_STORM_CTRL,
+			    DELAY, 3);
+	WREG32_SOC15(OSSSYS, 0, regIH_MSI_STORM_CTRL, tmp);
+
+	/* Redirect the interrupts to IH RB1 for dGPU */
+	if (adev->irq.ih1.ring_size) {
+		tmp = RREG32_SOC15(OSSSYS, 0, regIH_RING1_CLIENT_CFG_INDEX);
+		tmp = REG_SET_FIELD(tmp, IH_RING1_CLIENT_CFG_INDEX, INDEX, 0);
+		WREG32_SOC15(OSSSYS, 0, regIH_RING1_CLIENT_CFG_INDEX, tmp);
+
+		tmp = RREG32_SOC15(OSSSYS, 0, regIH_RING1_CLIENT_CFG_DATA);
+		tmp = REG_SET_FIELD(tmp, IH_RING1_CLIENT_CFG_DATA, CLIENT_ID, 0xa);
+		tmp = REG_SET_FIELD(tmp, IH_RING1_CLIENT_CFG_DATA, SOURCE_ID, 0x0);
+		tmp = REG_SET_FIELD(tmp, IH_RING1_CLIENT_CFG_DATA,
+				    SOURCE_ID_MATCH_ENABLE, 0x1);
+
+		WREG32_SOC15(OSSSYS, 0, regIH_RING1_CLIENT_CFG_DATA, tmp);
+	}
+
+	pci_set_master(adev->pdev);
+
+	/* enable interrupts */
+	ret = ih_v6_0_toggle_interrupts(adev, true);
+	if (ret)
+		return ret;
+	/* enable wptr force update for self int */
+	force_update_wptr_for_self_int(adev, 0, 8, true);
+
+	if (adev->irq.ih_soft.ring_size)
+		adev->irq.ih_soft.enabled = true;
+
+	return 0;
+}
+
+/**
+ * ih_v6_0_irq_disable - disable interrupts
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Disable interrupts on the hw.
+ */
+static void ih_v6_0_irq_disable(struct amdgpu_device *adev)
+{
+	force_update_wptr_for_self_int(adev, 0, 8, false);
+	ih_v6_0_toggle_interrupts(adev, false);
+
+	/* Wait and acknowledge irq */
+	mdelay(1);
+}
+
+/**
+ * ih_v6_0_get_wptr - get the IH ring buffer wptr
+ *
+ * @adev: amdgpu_device pointer
+ * @ih: amdgpu_ih_ring pointer
+ *
+ * Get the IH ring buffer wptr from either the register
+ * or the writeback memory buffer.  Also check for
+ * ring buffer overflow and deal with it.
+ * Returns the value of the wptr.
+ */
+static u32 ih_v6_0_get_wptr(struct amdgpu_device *adev,
+			      struct amdgpu_ih_ring *ih)
+{
+	u32 wptr, tmp;
+	struct amdgpu_ih_regs *ih_regs;
+
+	wptr = le32_to_cpu(*ih->wptr_cpu);
+	ih_regs = &ih->ih_regs;
+
+	if (!REG_GET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW))
+		goto out;
+
+	wptr = RREG32_NO_KIQ(ih_regs->ih_rb_wptr);
+	if (!REG_GET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW))
+		goto out;
+	if (!amdgpu_sriov_vf(adev))
+		wptr = REG_SET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW, 0);
+	else
+		ih->overflow = true;
+
+	/* When a ring buffer overflow happen start parsing interrupt
+	 * from the last not overwritten vector (wptr + 32). Hopefully
+	 * this should allow us to catch up.
+	 */
+	tmp = (wptr + 32) & ih->ptr_mask;
+	dev_warn(adev->dev, "IH ring buffer overflow "
+		 "(0x%08X, 0x%08X, 0x%08X)\n",
+		 wptr, ih->rptr, tmp);
+	ih->rptr = tmp;
+
+	tmp = RREG32_NO_KIQ(ih_regs->ih_rb_cntl);
+	tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 1);
+	WREG32_NO_KIQ(ih_regs->ih_rb_cntl, tmp);
+
+	/* Unset the CLEAR_OVERFLOW bit immediately so new overflows
+	 * can be detected.
+	 */
+	tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 0);
+	WREG32_NO_KIQ(ih_regs->ih_rb_cntl, tmp);
+out:
+	return (wptr & ih->ptr_mask);
+}
+
+/**
+ * ih_v6_0_irq_rearm - rearm IRQ if lost
+ *
+ * @adev: amdgpu_device pointer
+ * @ih: amdgpu_ih_ring pointer
+ *
+ */
+static void ih_v6_0_irq_rearm(struct amdgpu_device *adev,
+			       struct amdgpu_ih_ring *ih)
+{
+	uint32_t v = 0;
+	uint32_t i = 0;
+	struct amdgpu_ih_regs *ih_regs;
+
+	ih_regs = &ih->ih_regs;
+
+	/* Rearm IRQ / re-write doorbell if doorbell write is lost */
+	for (i = 0; i < MAX_REARM_RETRY; i++) {
+		v = RREG32_NO_KIQ(ih_regs->ih_rb_rptr);
+		if ((v < ih->ring_size) && (v != ih->rptr))
+			WDOORBELL32(ih->doorbell_index, ih->rptr);
+		else
+			break;
+	}
+}
+
+/**
+ * ih_v6_0_set_rptr - set the IH ring buffer rptr
+ *
+ * @adev: amdgpu_device pointer
+ * @ih: amdgpu_ih_ring pointer
+ *
+ * Set the IH ring buffer rptr.
+ */
+static void ih_v6_0_set_rptr(struct amdgpu_device *adev,
+			       struct amdgpu_ih_ring *ih)
+{
+	struct amdgpu_ih_regs *ih_regs;
+
+	if (ih->use_doorbell) {
+		/* XXX check if swapping is necessary on BE */
+		*ih->rptr_cpu = ih->rptr;
+		WDOORBELL32(ih->doorbell_index, ih->rptr);
+
+		if (amdgpu_sriov_vf(adev))
+			ih_v6_0_irq_rearm(adev, ih);
+	} else {
+		ih_regs = &ih->ih_regs;
+		WREG32(ih_regs->ih_rb_rptr, ih->rptr);
+	}
+}
+
+/**
+ * ih_v6_0_self_irq - dispatch work for ring 1
+ *
+ * @adev: amdgpu_device pointer
+ * @source: irq source
+ * @entry: IV with WPTR update
+ *
+ * Update the WPTR from the IV and schedule work to handle the entries.
+ */
+static int ih_v6_0_self_irq(struct amdgpu_device *adev,
+			      struct amdgpu_irq_src *source,
+			      struct amdgpu_iv_entry *entry)
+{
+	uint32_t wptr = cpu_to_le32(entry->src_data[0]);
+
+	switch (entry->ring_id) {
+	case 1:
+		*adev->irq.ih1.wptr_cpu = wptr;
+		schedule_work(&adev->irq.ih1_work);
+		break;
+	default:
+		break;
+	}
+	return 0;
+}
+
+static const struct amdgpu_irq_src_funcs ih_v6_0_self_irq_funcs = {
+	.process = ih_v6_0_self_irq,
+};
+
+static void ih_v6_0_set_self_irq_funcs(struct amdgpu_device *adev)
+{
+	adev->irq.self_irq.num_types = 0;
+	adev->irq.self_irq.funcs = &ih_v6_0_self_irq_funcs;
+}
+
+static int ih_v6_0_early_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	ih_v6_0_set_interrupt_funcs(adev);
+	ih_v6_0_set_self_irq_funcs(adev);
+	return 0;
+}
+
+static int ih_v6_0_sw_init(struct amdgpu_ip_block *ip_block)
+{
+	int r;
+	struct amdgpu_device *adev = ip_block->adev;
+	bool use_bus_addr;
+
+	r = amdgpu_irq_add_id(adev, SOC21_IH_CLIENTID_IH, 0,
+			      &adev->irq.self_irq);
+
+	if (r)
+		return r;
+
+	/* use gpu virtual address for ih ring
+	 * until ih_checken is programmed to allow
+	 * use bus address for ih ring by psp bl */
+	use_bus_addr = adev->firmware.load_type != AMDGPU_FW_LOAD_PSP;
+	r = amdgpu_ih_ring_init(adev, &adev->irq.ih, IH_RING_SIZE, use_bus_addr);
+	if (r)
+		return r;
+
+	adev->irq.ih.use_doorbell = true;
+	adev->irq.ih.doorbell_index = adev->doorbell_index.ih << 1;
+
+	if (!(adev->flags & AMD_IS_APU)) {
+		r = amdgpu_ih_ring_init(adev, &adev->irq.ih1, IH_RING_SIZE,
+					use_bus_addr);
+		if (r)
+			return r;
+
+		adev->irq.ih1.use_doorbell = true;
+		adev->irq.ih1.doorbell_index = (adev->doorbell_index.ih + 1) << 1;
+	}
+
+	/* initialize ih control register offset */
+	ih_v6_0_init_register_offset(adev);
+
+	r = amdgpu_ih_ring_init(adev, &adev->irq.ih_soft, IH_SW_RING_SIZE, true);
+	if (r)
+		return r;
+
+	r = amdgpu_irq_init(adev);
+
+	return r;
+}
+
+static int ih_v6_0_sw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	amdgpu_irq_fini_sw(adev);
+
+	return 0;
+}
+
+static int ih_v6_0_hw_init(struct amdgpu_ip_block *ip_block)
+{
+	int r;
+	struct amdgpu_device *adev = ip_block->adev;
+
+	r = ih_v6_0_irq_init(adev);
+	if (r)
+		return r;
+
+	return 0;
+}
+
+static int ih_v6_0_hw_fini(struct amdgpu_ip_block *ip_block)
+{
+	ih_v6_0_irq_disable(ip_block->adev);
+
+	return 0;
+}
+
+static int ih_v6_0_suspend(struct amdgpu_ip_block *ip_block)
+{
+	return ih_v6_0_hw_fini(ip_block);
+}
+
+static int ih_v6_0_resume(struct amdgpu_ip_block *ip_block)
+{
+	return ih_v6_0_hw_init(ip_block);
+}
+
+static bool ih_v6_0_is_idle(struct amdgpu_ip_block *ip_block)
+{
+	/* todo */
+	return true;
+}
+
+static int ih_v6_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
+{
+	/* todo */
+	return -ETIMEDOUT;
+}
+
+static int ih_v6_0_soft_reset(struct amdgpu_ip_block *ip_block)
+{
+	/* todo */
+	return 0;
+}
+
+static void ih_v6_0_update_clockgating_state(struct amdgpu_device *adev,
+					       bool enable)
+{
+	uint32_t data, def, field_val;
+
+	if (adev->cg_flags & AMD_CG_SUPPORT_IH_CG) {
+		def = data = RREG32_SOC15(OSSSYS, 0, regIH_CLK_CTRL);
+		field_val = enable ? 0 : 1;
+		data = REG_SET_FIELD(data, IH_CLK_CTRL,
+				     DBUS_MUX_CLK_SOFT_OVERRIDE, field_val);
+		data = REG_SET_FIELD(data, IH_CLK_CTRL,
+				     OSSSYS_SHARE_CLK_SOFT_OVERRIDE, field_val);
+		data = REG_SET_FIELD(data, IH_CLK_CTRL,
+				     LIMIT_SMN_CLK_SOFT_OVERRIDE, field_val);
+		data = REG_SET_FIELD(data, IH_CLK_CTRL,
+				     DYN_CLK_SOFT_OVERRIDE, field_val);
+		data = REG_SET_FIELD(data, IH_CLK_CTRL,
+				     REG_CLK_SOFT_OVERRIDE, field_val);
+		if (def != data)
+			WREG32_SOC15(OSSSYS, 0, regIH_CLK_CTRL, data);
+	}
+}
+
+static int ih_v6_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
+					   enum amd_clockgating_state state)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	ih_v6_0_update_clockgating_state(adev,
+				state == AMD_CG_STATE_GATE);
+	return 0;
+}
+
+static void ih_v6_0_update_ih_mem_power_gating(struct amdgpu_device *adev,
+					       bool enable)
+{
+	uint32_t ih_mem_pwr_cntl;
+
+	/* Disable ih sram power cntl before switch powergating mode */
+	ih_mem_pwr_cntl = RREG32_SOC15(OSSSYS, 0, regIH_MEM_POWER_CTRL);
+	ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+					IH_BUFFER_MEM_POWER_CTRL_EN, 0);
+	WREG32_SOC15(OSSSYS, 0, regIH_MEM_POWER_CTRL, ih_mem_pwr_cntl);
+
+	/* It is recommended to set mem powergating mode to DS mode */
+	if (enable) {
+		/* mem power mode */
+		ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+						IH_BUFFER_MEM_POWER_LS_EN, 0);
+		ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+						IH_BUFFER_MEM_POWER_DS_EN, 1);
+		ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+						IH_BUFFER_MEM_POWER_SD_EN, 0);
+		/* cam mem power mode */
+		ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+						IH_RETRY_INT_CAM_MEM_POWER_LS_EN, 0);
+		ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+						IH_RETRY_INT_CAM_MEM_POWER_DS_EN, 1);
+		ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+						IH_RETRY_INT_CAM_MEM_POWER_SD_EN, 0);
+		/* re-enable power cntl */
+		ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+						IH_BUFFER_MEM_POWER_CTRL_EN, 1);
+	} else {
+		/* mem power mode */
+		ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+						IH_BUFFER_MEM_POWER_LS_EN, 0);
+		ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+						IH_BUFFER_MEM_POWER_DS_EN, 0);
+		ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+						IH_BUFFER_MEM_POWER_SD_EN, 0);
+		/* cam mem power mode */
+		ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+						IH_RETRY_INT_CAM_MEM_POWER_LS_EN, 0);
+		ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+						IH_RETRY_INT_CAM_MEM_POWER_DS_EN, 0);
+		ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+						IH_RETRY_INT_CAM_MEM_POWER_SD_EN, 0);
+		/* re-enable power cntl*/
+		ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+						IH_BUFFER_MEM_POWER_CTRL_EN, 1);
+	}
+
+	WREG32_SOC15(OSSSYS, 0, regIH_MEM_POWER_CTRL, ih_mem_pwr_cntl);
+}
+
+static int ih_v6_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
+					 enum amd_powergating_state state)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	bool enable = (state == AMD_PG_STATE_GATE);
+
+	if (adev->pg_flags & AMD_PG_SUPPORT_IH_SRAM_PG)
+		ih_v6_0_update_ih_mem_power_gating(adev, enable);
+
+	return 0;
+}
+
+static void ih_v6_0_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	if (!RREG32_SOC15(OSSSYS, 0, regIH_CLK_CTRL))
+		*flags |= AMD_CG_SUPPORT_IH_CG;
+}
+
+static const struct amd_ip_funcs ih_v6_0_ip_funcs = {
+	.name = "ih_v6_0",
+	.early_init = ih_v6_0_early_init,
+	.sw_init = ih_v6_0_sw_init,
+	.sw_fini = ih_v6_0_sw_fini,
+	.hw_init = ih_v6_0_hw_init,
+	.hw_fini = ih_v6_0_hw_fini,
+	.suspend = ih_v6_0_suspend,
+	.resume = ih_v6_0_resume,
+	.is_idle = ih_v6_0_is_idle,
+	.wait_for_idle = ih_v6_0_wait_for_idle,
+	.soft_reset = ih_v6_0_soft_reset,
+	.set_clockgating_state = ih_v6_0_set_clockgating_state,
+	.set_powergating_state = ih_v6_0_set_powergating_state,
+	.get_clockgating_state = ih_v6_0_get_clockgating_state,
+};
+
+static const struct amdgpu_ih_funcs ih_v6_0_funcs = {
+	.get_wptr = ih_v6_0_get_wptr,
+	.decode_iv = amdgpu_ih_decode_iv_helper,
+	.decode_iv_ts = amdgpu_ih_decode_iv_ts_helper,
+	.set_rptr = ih_v6_0_set_rptr
+};
+
+static void ih_v6_0_set_interrupt_funcs(struct amdgpu_device *adev)
+{
+	adev->irq.ih_funcs = &ih_v6_0_funcs;
+}
+
+const struct amdgpu_ip_block_version ih_v6_0_ip_block = {
+	.type = AMD_IP_BLOCK_TYPE_IH,
+	.major = 6,
+	.minor = 0,
+	.rev = 0,
+	.funcs = &ih_v6_0_ip_funcs,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/ih_v6_0.h b/drivers/gpu/drm/amd/amdgpu/ih_v6_0.h
new file mode 100644
index 000000000000..f27b55580716
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/ih_v6_0.h
@@ -0,0 +1,28 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __IH_V6_0_IH_H__
+#define __IH_V6_0_IH_H__
+
+extern const struct amdgpu_ip_block_version ih_v6_0_ip_block;
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/ih_v6_1.c b/drivers/gpu/drm/amd/amdgpu/ih_v6_1.c
new file mode 100644
index 000000000000..95b3f4e55ec3
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/ih_v6_1.c
@@ -0,0 +1,796 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/pci.h>
+
+#include "amdgpu.h"
+#include "amdgpu_ih.h"
+
+#include "oss/osssys_6_1_0_offset.h"
+#include "oss/osssys_6_1_0_sh_mask.h"
+
+#include "soc15_common.h"
+#include "ih_v6_1.h"
+
+#define MAX_REARM_RETRY 10
+
+static void ih_v6_1_set_interrupt_funcs(struct amdgpu_device *adev);
+
+/**
+ * ih_v6_1_init_register_offset - Initialize register offset for ih rings
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Initialize register offset ih rings (IH_V6_0).
+ */
+static void ih_v6_1_init_register_offset(struct amdgpu_device *adev)
+{
+	struct amdgpu_ih_regs *ih_regs;
+
+	/* ih ring 2 is removed
+	 * ih ring and ih ring 1 are available */
+	if (adev->irq.ih.ring_size) {
+		ih_regs = &adev->irq.ih.ih_regs;
+		ih_regs->ih_rb_base = SOC15_REG_OFFSET(OSSSYS, 0, regIH_RB_BASE);
+		ih_regs->ih_rb_base_hi = SOC15_REG_OFFSET(OSSSYS, 0, regIH_RB_BASE_HI);
+		ih_regs->ih_rb_cntl = SOC15_REG_OFFSET(OSSSYS, 0, regIH_RB_CNTL);
+		ih_regs->ih_rb_wptr = SOC15_REG_OFFSET(OSSSYS, 0, regIH_RB_WPTR);
+		ih_regs->ih_rb_rptr = SOC15_REG_OFFSET(OSSSYS, 0, regIH_RB_RPTR);
+		ih_regs->ih_doorbell_rptr = SOC15_REG_OFFSET(OSSSYS, 0, regIH_DOORBELL_RPTR);
+		ih_regs->ih_rb_wptr_addr_lo = SOC15_REG_OFFSET(OSSSYS, 0, regIH_RB_WPTR_ADDR_LO);
+		ih_regs->ih_rb_wptr_addr_hi = SOC15_REG_OFFSET(OSSSYS, 0, regIH_RB_WPTR_ADDR_HI);
+		ih_regs->psp_reg_id = PSP_REG_IH_RB_CNTL;
+	}
+
+	if (adev->irq.ih1.ring_size) {
+		ih_regs = &adev->irq.ih1.ih_regs;
+		ih_regs->ih_rb_base = SOC15_REG_OFFSET(OSSSYS, 0, regIH_RB_BASE_RING1);
+		ih_regs->ih_rb_base_hi = SOC15_REG_OFFSET(OSSSYS, 0, regIH_RB_BASE_HI_RING1);
+		ih_regs->ih_rb_cntl = SOC15_REG_OFFSET(OSSSYS, 0, regIH_RB_CNTL_RING1);
+		ih_regs->ih_rb_wptr = SOC15_REG_OFFSET(OSSSYS, 0, regIH_RB_WPTR_RING1);
+		ih_regs->ih_rb_rptr = SOC15_REG_OFFSET(OSSSYS, 0, regIH_RB_RPTR_RING1);
+		ih_regs->ih_doorbell_rptr = SOC15_REG_OFFSET(OSSSYS, 0, regIH_DOORBELL_RPTR_RING1);
+		ih_regs->psp_reg_id = PSP_REG_IH_RB_CNTL_RING1;
+	}
+}
+
+/**
+ * force_update_wptr_for_self_int - Force update the wptr for self interrupt
+ *
+ * @adev: amdgpu_device pointer
+ * @threshold: threshold to trigger the wptr reporting
+ * @timeout: timeout to trigger the wptr reporting
+ * @enabled: Enable/disable timeout flush mechanism
+ *
+ * threshold input range: 0 ~ 15, default 0,
+ * real_threshold = 2^threshold
+ * timeout input range: 0 ~ 20, default 8,
+ * real_timeout = (2^timeout) * 1024 / (socclk_freq)
+ *
+ * Force update wptr for self interrupt ( >= SIENNA_CICHLID).
+ */
+static void
+force_update_wptr_for_self_int(struct amdgpu_device *adev,
+			       u32 threshold, u32 timeout, bool enabled)
+{
+	u32 ih_cntl, ih_rb_cntl;
+
+	ih_cntl = RREG32_SOC15(OSSSYS, 0, regIH_CNTL2);
+	ih_rb_cntl = RREG32_SOC15(OSSSYS, 0, regIH_RB_CNTL_RING1);
+
+	ih_cntl = REG_SET_FIELD(ih_cntl, IH_CNTL2,
+				SELF_IV_FORCE_WPTR_UPDATE_TIMEOUT, timeout);
+	ih_cntl = REG_SET_FIELD(ih_cntl, IH_CNTL2,
+				SELF_IV_FORCE_WPTR_UPDATE_ENABLE, enabled);
+	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL_RING1,
+				   RB_USED_INT_THRESHOLD, threshold);
+
+	if (amdgpu_sriov_vf(adev) && amdgpu_sriov_reg_indirect_ih(adev)) {
+		if (psp_reg_program(&adev->psp, PSP_REG_IH_RB_CNTL_RING1, ih_rb_cntl))
+			return;
+	} else {
+		WREG32_SOC15(OSSSYS, 0, regIH_RB_CNTL_RING1, ih_rb_cntl);
+	}
+
+	WREG32_SOC15(OSSSYS, 0, regIH_CNTL2, ih_cntl);
+}
+
+/**
+ * ih_v6_1_toggle_ring_interrupts - toggle the interrupt ring buffer
+ *
+ * @adev: amdgpu_device pointer
+ * @ih: amdgpu_ih_ring pointer
+ * @enable: true - enable the interrupts, false - disable the interrupts
+ *
+ * Toggle the interrupt ring buffer (IH_V6_0)
+ */
+static int ih_v6_1_toggle_ring_interrupts(struct amdgpu_device *adev,
+					  struct amdgpu_ih_ring *ih,
+					  bool enable)
+{
+	struct amdgpu_ih_regs *ih_regs;
+	uint32_t tmp;
+
+	ih_regs = &ih->ih_regs;
+
+	tmp = RREG32(ih_regs->ih_rb_cntl);
+	tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, RB_ENABLE, (enable ? 1 : 0));
+	/* enable_intr field is only valid in ring0 */
+	if (ih == &adev->irq.ih)
+		tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, ENABLE_INTR, (enable ? 1 : 0));
+
+	if (amdgpu_sriov_vf(adev) && amdgpu_sriov_reg_indirect_ih(adev)) {
+		if (psp_reg_program(&adev->psp, ih_regs->psp_reg_id, tmp))
+			return -ETIMEDOUT;
+	} else {
+		WREG32(ih_regs->ih_rb_cntl, tmp);
+	}
+
+	if (enable) {
+		ih->enabled = true;
+	} else {
+		/* set rptr, wptr to 0 */
+		WREG32(ih_regs->ih_rb_rptr, 0);
+		WREG32(ih_regs->ih_rb_wptr, 0);
+		ih->enabled = false;
+		ih->rptr = 0;
+	}
+
+	return 0;
+}
+
+/**
+ * ih_v6_1_toggle_interrupts - Toggle all the available interrupt ring buffers
+ *
+ * @adev: amdgpu_device pointer
+ * @enable: enable or disable interrupt ring buffers
+ *
+ * Toggle all the available interrupt ring buffers (IH_V6_0).
+ */
+static int ih_v6_1_toggle_interrupts(struct amdgpu_device *adev, bool enable)
+{
+	struct amdgpu_ih_ring *ih[] = {&adev->irq.ih, &adev->irq.ih1};
+	int i;
+	int r;
+
+	for (i = 0; i < ARRAY_SIZE(ih); i++) {
+		if (ih[i]->ring_size) {
+			r = ih_v6_1_toggle_ring_interrupts(adev, ih[i], enable);
+			if (r)
+				return r;
+		}
+	}
+
+	return 0;
+}
+
+static uint32_t ih_v6_1_rb_cntl(struct amdgpu_ih_ring *ih, uint32_t ih_rb_cntl)
+{
+	int rb_bufsz = order_base_2(ih->ring_size / 4);
+
+	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL,
+				   MC_SPACE, ih->use_bus_addr ? 2 : 4);
+	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL,
+				   WPTR_OVERFLOW_CLEAR, 1);
+	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL,
+				   WPTR_OVERFLOW_ENABLE, 1);
+	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, RB_SIZE, rb_bufsz);
+	/* Ring Buffer write pointer writeback. If enabled, IH_RB_WPTR register
+	 * value is written to memory
+	 */
+	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL,
+				   WPTR_WRITEBACK_ENABLE, 1);
+	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, MC_SNOOP, 1);
+	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, MC_RO, 0);
+	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, MC_VMID, 0);
+
+	return ih_rb_cntl;
+}
+
+static uint32_t ih_v6_1_doorbell_rptr(struct amdgpu_ih_ring *ih)
+{
+	u32 ih_doorbell_rtpr = 0;
+
+	if (ih->use_doorbell) {
+		ih_doorbell_rtpr = REG_SET_FIELD(ih_doorbell_rtpr,
+						 IH_DOORBELL_RPTR, OFFSET,
+						 ih->doorbell_index);
+		ih_doorbell_rtpr = REG_SET_FIELD(ih_doorbell_rtpr,
+						 IH_DOORBELL_RPTR,
+						 ENABLE, 1);
+	} else {
+		ih_doorbell_rtpr = REG_SET_FIELD(ih_doorbell_rtpr,
+						 IH_DOORBELL_RPTR,
+						 ENABLE, 0);
+	}
+	return ih_doorbell_rtpr;
+}
+
+/**
+ * ih_v6_1_enable_ring - enable an ih ring buffer
+ *
+ * @adev: amdgpu_device pointer
+ * @ih: amdgpu_ih_ring pointer
+ *
+ * Enable an ih ring buffer (IH_V6_0)
+ */
+static int ih_v6_1_enable_ring(struct amdgpu_device *adev,
+				      struct amdgpu_ih_ring *ih)
+{
+	struct amdgpu_ih_regs *ih_regs;
+	uint32_t tmp;
+
+	ih_regs = &ih->ih_regs;
+
+	/* Ring Buffer base. [39:8] of 40-bit address of the beginning of the ring buffer*/
+	WREG32(ih_regs->ih_rb_base, ih->gpu_addr >> 8);
+	WREG32(ih_regs->ih_rb_base_hi, (ih->gpu_addr >> 40) & 0xff);
+
+	tmp = RREG32(ih_regs->ih_rb_cntl);
+	tmp = ih_v6_1_rb_cntl(ih, tmp);
+	if (ih == &adev->irq.ih)
+		tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, RPTR_REARM, !!adev->irq.msi_enabled);
+	if (ih == &adev->irq.ih1) {
+		tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_ENABLE, 0);
+		tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, RB_FULL_DRAIN_ENABLE, 1);
+	}
+
+	if (amdgpu_sriov_vf(adev) && amdgpu_sriov_reg_indirect_ih(adev)) {
+		if (psp_reg_program(&adev->psp, ih_regs->psp_reg_id, tmp)) {
+			DRM_ERROR("PSP program IH_RB_CNTL failed!\n");
+			return -ETIMEDOUT;
+		}
+	} else {
+		WREG32(ih_regs->ih_rb_cntl, tmp);
+	}
+
+	if (ih == &adev->irq.ih) {
+		/* set the ih ring 0 writeback address whether it's enabled or not */
+		WREG32(ih_regs->ih_rb_wptr_addr_lo, lower_32_bits(ih->wptr_addr));
+		WREG32(ih_regs->ih_rb_wptr_addr_hi, upper_32_bits(ih->wptr_addr) & 0xFFFF);
+	}
+
+	/* set rptr, wptr to 0 */
+	WREG32(ih_regs->ih_rb_wptr, 0);
+	WREG32(ih_regs->ih_rb_rptr, 0);
+
+	WREG32(ih_regs->ih_doorbell_rptr, ih_v6_1_doorbell_rptr(ih));
+
+	return 0;
+}
+
+/**
+ * ih_v6_1_irq_init - init and enable the interrupt ring
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Allocate a ring buffer for the interrupt controller,
+ * enable the RLC, disable interrupts, enable the IH
+ * ring buffer and enable it.
+ * Called at device load and reume.
+ * Returns 0 for success, errors for failure.
+ */
+static int ih_v6_1_irq_init(struct amdgpu_device *adev)
+{
+	struct amdgpu_ih_ring *ih[] = {&adev->irq.ih, &adev->irq.ih1};
+	u32 ih_chicken;
+	u32 tmp;
+	int ret;
+	int i;
+
+	/* disable irqs */
+	ret = ih_v6_1_toggle_interrupts(adev, false);
+	if (ret)
+		return ret;
+
+	adev->nbio.funcs->ih_control(adev);
+
+	if (unlikely((adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT) ||
+		     (adev->firmware.load_type == AMDGPU_FW_LOAD_RLC_BACKDOOR_AUTO))) {
+		if (ih[0]->use_bus_addr) {
+			ih_chicken = RREG32_SOC15(OSSSYS, 0, regIH_CHICKEN);
+			ih_chicken = REG_SET_FIELD(ih_chicken,
+					IH_CHICKEN, MC_SPACE_GPA_ENABLE, 1);
+			WREG32_SOC15(OSSSYS, 0, regIH_CHICKEN, ih_chicken);
+		}
+	}
+
+	for (i = 0; i < ARRAY_SIZE(ih); i++) {
+		if (ih[i]->ring_size) {
+			ret = ih_v6_1_enable_ring(adev, ih[i]);
+			if (ret)
+				return ret;
+		}
+	}
+
+	/* update doorbell range for ih ring 0 */
+	adev->nbio.funcs->ih_doorbell_range(adev, ih[0]->use_doorbell,
+					    ih[0]->doorbell_index);
+
+	tmp = RREG32_SOC15(OSSSYS, 0, regIH_STORM_CLIENT_LIST_CNTL);
+	tmp = REG_SET_FIELD(tmp, IH_STORM_CLIENT_LIST_CNTL,
+			    CLIENT18_IS_STORM_CLIENT, 1);
+	WREG32_SOC15(OSSSYS, 0, regIH_STORM_CLIENT_LIST_CNTL, tmp);
+
+	tmp = RREG32_SOC15(OSSSYS, 0, regIH_INT_FLOOD_CNTL);
+	tmp = REG_SET_FIELD(tmp, IH_INT_FLOOD_CNTL, FLOOD_CNTL_ENABLE, 1);
+	WREG32_SOC15(OSSSYS, 0, regIH_INT_FLOOD_CNTL, tmp);
+
+	/* GC/MMHUB UTCL2 page fault interrupts are configured as
+	 * MSI storm capable interrupts by deafult. The delay is
+	 * used to avoid ISR being called too frequently
+	 * when page fault happens on several continuous page
+	 * and thus avoid MSI storm */
+	tmp = RREG32_SOC15(OSSSYS, 0, regIH_MSI_STORM_CTRL);
+	tmp = REG_SET_FIELD(tmp, IH_MSI_STORM_CTRL,
+			    DELAY, 3);
+	WREG32_SOC15(OSSSYS, 0, regIH_MSI_STORM_CTRL, tmp);
+
+	/* Redirect the interrupts to IH RB1 for dGPU */
+	if (adev->irq.ih1.ring_size) {
+		tmp = RREG32_SOC15(OSSSYS, 0, regIH_RING1_CLIENT_CFG_INDEX);
+		tmp = REG_SET_FIELD(tmp, IH_RING1_CLIENT_CFG_INDEX, INDEX, 0);
+		WREG32_SOC15(OSSSYS, 0, regIH_RING1_CLIENT_CFG_INDEX, tmp);
+
+		tmp = RREG32_SOC15(OSSSYS, 0, regIH_RING1_CLIENT_CFG_DATA);
+		tmp = REG_SET_FIELD(tmp, IH_RING1_CLIENT_CFG_DATA, CLIENT_ID, 0xa);
+		tmp = REG_SET_FIELD(tmp, IH_RING1_CLIENT_CFG_DATA, SOURCE_ID, 0x0);
+		tmp = REG_SET_FIELD(tmp, IH_RING1_CLIENT_CFG_DATA,
+				    SOURCE_ID_MATCH_ENABLE, 0x1);
+
+		WREG32_SOC15(OSSSYS, 0, regIH_RING1_CLIENT_CFG_DATA, tmp);
+	}
+
+	pci_set_master(adev->pdev);
+
+	/* enable interrupts */
+	ret = ih_v6_1_toggle_interrupts(adev, true);
+	if (ret)
+		return ret;
+	/* enable wptr force update for self int */
+	force_update_wptr_for_self_int(adev, 0, 8, true);
+
+	if (adev->irq.ih_soft.ring_size)
+		adev->irq.ih_soft.enabled = true;
+
+	return 0;
+}
+
+/**
+ * ih_v6_1_irq_disable - disable interrupts
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Disable interrupts on the hw.
+ */
+static void ih_v6_1_irq_disable(struct amdgpu_device *adev)
+{
+	force_update_wptr_for_self_int(adev, 0, 8, false);
+	ih_v6_1_toggle_interrupts(adev, false);
+
+	/* Wait and acknowledge irq */
+	mdelay(1);
+}
+
+/**
+ * ih_v6_1_get_wptr - get the IH ring buffer wptr
+ *
+ * @adev: amdgpu_device pointer
+ * @ih: amdgpu_ih_ring pointer
+ *
+ * Get the IH ring buffer wptr from either the register
+ * or the writeback memory buffer.  Also check for
+ * ring buffer overflow and deal with it.
+ * Returns the value of the wptr.
+ */
+static u32 ih_v6_1_get_wptr(struct amdgpu_device *adev,
+			      struct amdgpu_ih_ring *ih)
+{
+	u32 wptr, tmp;
+	struct amdgpu_ih_regs *ih_regs;
+
+	wptr = le32_to_cpu(*ih->wptr_cpu);
+	ih_regs = &ih->ih_regs;
+
+	if (!REG_GET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW))
+		goto out;
+
+	wptr = RREG32_NO_KIQ(ih_regs->ih_rb_wptr);
+	if (!REG_GET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW))
+		goto out;
+	wptr = REG_SET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW, 0);
+
+	/* When a ring buffer overflow happen start parsing interrupt
+	 * from the last not overwritten vector (wptr + 32). Hopefully
+	 * this should allow us to catch up.
+	 */
+	tmp = (wptr + 32) & ih->ptr_mask;
+	dev_warn(adev->dev, "IH ring buffer overflow "
+		 "(0x%08X, 0x%08X, 0x%08X)\n",
+		 wptr, ih->rptr, tmp);
+	ih->rptr = tmp;
+
+	tmp = RREG32_NO_KIQ(ih_regs->ih_rb_cntl);
+	tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 1);
+	WREG32_NO_KIQ(ih_regs->ih_rb_cntl, tmp);
+
+	/* Unset the CLEAR_OVERFLOW bit immediately so new overflows
+	 * can be detected.
+	 */
+	tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 0);
+	WREG32_NO_KIQ(ih_regs->ih_rb_cntl, tmp);
+
+out:
+	return (wptr & ih->ptr_mask);
+}
+
+/**
+ * ih_v6_1_irq_rearm - rearm IRQ if lost
+ *
+ * @adev: amdgpu_device pointer
+ * @ih: amdgpu_ih_ring pointer
+ *
+ */
+static void ih_v6_1_irq_rearm(struct amdgpu_device *adev,
+			       struct amdgpu_ih_ring *ih)
+{
+	uint32_t v = 0;
+	uint32_t i = 0;
+	struct amdgpu_ih_regs *ih_regs;
+
+	ih_regs = &ih->ih_regs;
+
+	/* Rearm IRQ / re-write doorbell if doorbell write is lost */
+	for (i = 0; i < MAX_REARM_RETRY; i++) {
+		v = RREG32_NO_KIQ(ih_regs->ih_rb_rptr);
+		if ((v < ih->ring_size) && (v != ih->rptr))
+			WDOORBELL32(ih->doorbell_index, ih->rptr);
+		else
+			break;
+	}
+}
+
+/**
+ * ih_v6_1_set_rptr - set the IH ring buffer rptr
+ *
+ * @adev: amdgpu_device pointer
+ * @ih: amdgpu_ih_ring pointer
+ *
+ * Set the IH ring buffer rptr.
+ */
+static void ih_v6_1_set_rptr(struct amdgpu_device *adev,
+			       struct amdgpu_ih_ring *ih)
+{
+	struct amdgpu_ih_regs *ih_regs;
+
+	if (ih->use_doorbell) {
+		/* XXX check if swapping is necessary on BE */
+		*ih->rptr_cpu = ih->rptr;
+		WDOORBELL32(ih->doorbell_index, ih->rptr);
+
+		if (amdgpu_sriov_vf(adev))
+			ih_v6_1_irq_rearm(adev, ih);
+	} else {
+		ih_regs = &ih->ih_regs;
+		WREG32(ih_regs->ih_rb_rptr, ih->rptr);
+	}
+}
+
+/**
+ * ih_v6_1_self_irq - dispatch work for ring 1
+ *
+ * @adev: amdgpu_device pointer
+ * @source: irq source
+ * @entry: IV with WPTR update
+ *
+ * Update the WPTR from the IV and schedule work to handle the entries.
+ */
+static int ih_v6_1_self_irq(struct amdgpu_device *adev,
+			      struct amdgpu_irq_src *source,
+			      struct amdgpu_iv_entry *entry)
+{
+	uint32_t wptr = cpu_to_le32(entry->src_data[0]);
+
+	switch (entry->ring_id) {
+	case 1:
+		*adev->irq.ih1.wptr_cpu = wptr;
+		schedule_work(&adev->irq.ih1_work);
+		break;
+	default:
+		break;
+	}
+	return 0;
+}
+
+static const struct amdgpu_irq_src_funcs ih_v6_1_self_irq_funcs = {
+	.process = ih_v6_1_self_irq,
+};
+
+static void ih_v6_1_set_self_irq_funcs(struct amdgpu_device *adev)
+{
+	adev->irq.self_irq.num_types = 0;
+	adev->irq.self_irq.funcs = &ih_v6_1_self_irq_funcs;
+}
+
+static int ih_v6_1_early_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int ret;
+
+	ret = amdgpu_irq_add_domain(adev);
+	if (ret) {
+		return ret;
+	}
+
+	ih_v6_1_set_interrupt_funcs(adev);
+	ih_v6_1_set_self_irq_funcs(adev);
+	return 0;
+}
+
+static int ih_v6_1_sw_init(struct amdgpu_ip_block *ip_block)
+{
+	int r;
+	struct amdgpu_device *adev = ip_block->adev;
+	bool use_bus_addr;
+
+	r = amdgpu_irq_add_id(adev, SOC21_IH_CLIENTID_IH, 0,
+			      &adev->irq.self_irq);
+
+	if (r)
+		return r;
+
+	/* use gpu virtual address for ih ring
+	 * until ih_checken is programmed to allow
+	 * use bus address for ih ring by psp bl */
+	use_bus_addr = adev->firmware.load_type != AMDGPU_FW_LOAD_PSP;
+	r = amdgpu_ih_ring_init(adev, &adev->irq.ih, 256 * 1024, use_bus_addr);
+	if (r)
+		return r;
+
+	adev->irq.ih.use_doorbell = true;
+	adev->irq.ih.doorbell_index = adev->doorbell_index.ih << 1;
+
+	if (!(adev->flags & AMD_IS_APU)) {
+		r = amdgpu_ih_ring_init(adev, &adev->irq.ih1, IH_RING_SIZE,
+					use_bus_addr);
+		if (r)
+			return r;
+
+		adev->irq.ih1.use_doorbell = true;
+		adev->irq.ih1.doorbell_index = (adev->doorbell_index.ih + 1) << 1;
+	}
+
+	/* initialize ih control register offset */
+	ih_v6_1_init_register_offset(adev);
+
+	r = amdgpu_ih_ring_init(adev, &adev->irq.ih_soft, PAGE_SIZE, true);
+	if (r)
+		return r;
+
+	r = amdgpu_irq_init(adev);
+
+	return r;
+}
+
+static int ih_v6_1_sw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	amdgpu_irq_fini_sw(adev);
+
+	return 0;
+}
+
+static int ih_v6_1_hw_init(struct amdgpu_ip_block *ip_block)
+{
+	int r;
+	struct amdgpu_device *adev = ip_block->adev;
+
+	r = ih_v6_1_irq_init(adev);
+	if (r)
+		return r;
+
+	return 0;
+}
+
+static int ih_v6_1_hw_fini(struct amdgpu_ip_block *ip_block)
+{
+	ih_v6_1_irq_disable(ip_block->adev);
+
+	return 0;
+}
+
+static int ih_v6_1_suspend(struct amdgpu_ip_block *ip_block)
+{
+	return ih_v6_1_hw_fini(ip_block);
+}
+
+static int ih_v6_1_resume(struct amdgpu_ip_block *ip_block)
+{
+	return ih_v6_1_hw_init(ip_block);
+}
+
+static bool ih_v6_1_is_idle(struct amdgpu_ip_block *ip_block)
+{
+	/* todo */
+	return true;
+}
+
+static int ih_v6_1_wait_for_idle(struct amdgpu_ip_block *ip_block)
+{
+	/* todo */
+	return -ETIMEDOUT;
+}
+
+static int ih_v6_1_soft_reset(struct amdgpu_ip_block *ip_block)
+{
+	/* todo */
+	return 0;
+}
+
+static void ih_v6_1_update_clockgating_state(struct amdgpu_device *adev,
+					       bool enable)
+{
+	uint32_t data, def, field_val;
+
+	if (adev->cg_flags & AMD_CG_SUPPORT_IH_CG) {
+		def = data = RREG32_SOC15(OSSSYS, 0, regIH_CLK_CTRL);
+		field_val = enable ? 0 : 1;
+		data = REG_SET_FIELD(data, IH_CLK_CTRL,
+				     DBUS_MUX_CLK_SOFT_OVERRIDE, field_val);
+		data = REG_SET_FIELD(data, IH_CLK_CTRL,
+				     OSSSYS_SHARE_CLK_SOFT_OVERRIDE, field_val);
+		data = REG_SET_FIELD(data, IH_CLK_CTRL,
+				     LIMIT_SMN_CLK_SOFT_OVERRIDE, field_val);
+		data = REG_SET_FIELD(data, IH_CLK_CTRL,
+				     DYN_CLK_SOFT_OVERRIDE, field_val);
+		data = REG_SET_FIELD(data, IH_CLK_CTRL,
+				     REG_CLK_SOFT_OVERRIDE, field_val);
+		if (def != data)
+			WREG32_SOC15(OSSSYS, 0, regIH_CLK_CTRL, data);
+	}
+
+	return;
+}
+
+static int ih_v6_1_set_clockgating_state(struct amdgpu_ip_block *ip_block,
+					   enum amd_clockgating_state state)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	ih_v6_1_update_clockgating_state(adev,
+				state == AMD_CG_STATE_GATE);
+	return 0;
+}
+
+static void ih_v6_1_update_ih_mem_power_gating(struct amdgpu_device *adev,
+					       bool enable)
+{
+	uint32_t ih_mem_pwr_cntl;
+
+	/* Disable ih sram power cntl before switch powergating mode */
+	ih_mem_pwr_cntl = RREG32_SOC15(OSSSYS, 0, regIH_MEM_POWER_CTRL);
+	ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+					IH_BUFFER_MEM_POWER_CTRL_EN, 0);
+	WREG32_SOC15(OSSSYS, 0, regIH_MEM_POWER_CTRL, ih_mem_pwr_cntl);
+
+	/* It is recommended to set mem powergating mode to DS mode */
+	if (enable) {
+		/* mem power mode */
+		ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+						IH_BUFFER_MEM_POWER_LS_EN, 0);
+		ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+						IH_BUFFER_MEM_POWER_DS_EN, 1);
+		ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+						IH_BUFFER_MEM_POWER_SD_EN, 0);
+		/* cam mem power mode */
+		ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+						IH_RETRY_INT_CAM_MEM_POWER_LS_EN, 0);
+		ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+						IH_RETRY_INT_CAM_MEM_POWER_DS_EN, 1);
+		ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+						IH_RETRY_INT_CAM_MEM_POWER_SD_EN, 0);
+		/* re-enable power cntl */
+		ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+						IH_BUFFER_MEM_POWER_CTRL_EN, 1);
+	} else {
+		/* mem power mode */
+		ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+						IH_BUFFER_MEM_POWER_LS_EN, 0);
+		ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+						IH_BUFFER_MEM_POWER_DS_EN, 0);
+		ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+						IH_BUFFER_MEM_POWER_SD_EN, 0);
+		/* cam mem power mode */
+		ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+						IH_RETRY_INT_CAM_MEM_POWER_LS_EN, 0);
+		ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+						IH_RETRY_INT_CAM_MEM_POWER_DS_EN, 0);
+		ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+						IH_RETRY_INT_CAM_MEM_POWER_SD_EN, 0);
+		/* re-enable power cntl*/
+		ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+						IH_BUFFER_MEM_POWER_CTRL_EN, 1);
+	}
+
+	WREG32_SOC15(OSSSYS, 0, regIH_MEM_POWER_CTRL, ih_mem_pwr_cntl);
+}
+
+static int ih_v6_1_set_powergating_state(struct amdgpu_ip_block *ip_block,
+					 enum amd_powergating_state state)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	bool enable = (state == AMD_PG_STATE_GATE);
+
+	if (adev->pg_flags & AMD_PG_SUPPORT_IH_SRAM_PG)
+		ih_v6_1_update_ih_mem_power_gating(adev, enable);
+
+	return 0;
+}
+
+static void ih_v6_1_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	if (!RREG32_SOC15(OSSSYS, 0, regIH_CLK_CTRL))
+		*flags |= AMD_CG_SUPPORT_IH_CG;
+
+	return;
+}
+
+static const struct amd_ip_funcs ih_v6_1_ip_funcs = {
+	.name = "ih_v6_1",
+	.early_init = ih_v6_1_early_init,
+	.sw_init = ih_v6_1_sw_init,
+	.sw_fini = ih_v6_1_sw_fini,
+	.hw_init = ih_v6_1_hw_init,
+	.hw_fini = ih_v6_1_hw_fini,
+	.suspend = ih_v6_1_suspend,
+	.resume = ih_v6_1_resume,
+	.is_idle = ih_v6_1_is_idle,
+	.wait_for_idle = ih_v6_1_wait_for_idle,
+	.soft_reset = ih_v6_1_soft_reset,
+	.set_clockgating_state = ih_v6_1_set_clockgating_state,
+	.set_powergating_state = ih_v6_1_set_powergating_state,
+	.get_clockgating_state = ih_v6_1_get_clockgating_state,
+};
+
+static const struct amdgpu_ih_funcs ih_v6_1_funcs = {
+	.get_wptr = ih_v6_1_get_wptr,
+	.decode_iv = amdgpu_ih_decode_iv_helper,
+	.decode_iv_ts = amdgpu_ih_decode_iv_ts_helper,
+	.set_rptr = ih_v6_1_set_rptr
+};
+
+static void ih_v6_1_set_interrupt_funcs(struct amdgpu_device *adev)
+{
+	adev->irq.ih_funcs = &ih_v6_1_funcs;
+}
+
+const struct amdgpu_ip_block_version ih_v6_1_ip_block = {
+	.type = AMD_IP_BLOCK_TYPE_IH,
+	.major = 6,
+	.minor = 0,
+	.rev = 0,
+	.funcs = &ih_v6_1_ip_funcs,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/ih_v6_1.h b/drivers/gpu/drm/amd/amdgpu/ih_v6_1.h
new file mode 100644
index 000000000000..2232bc5cbd09
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/ih_v6_1.h
@@ -0,0 +1,28 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __IH_V6_1_IH_H__
+#define __IH_V6_1_IH_H__
+
+extern const struct amdgpu_ip_block_version ih_v6_1_ip_block;
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/ih_v7_0.c b/drivers/gpu/drm/amd/amdgpu/ih_v7_0.c
new file mode 100644
index 000000000000..b32ea4129c61
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/ih_v7_0.c
@@ -0,0 +1,787 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/pci.h>
+
+#include "amdgpu.h"
+#include "amdgpu_ih.h"
+
+#include "oss/osssys_7_0_0_offset.h"
+#include "oss/osssys_7_0_0_sh_mask.h"
+
+#include "soc15_common.h"
+#include "ih_v7_0.h"
+
+#define MAX_REARM_RETRY 10
+
+static void ih_v7_0_set_interrupt_funcs(struct amdgpu_device *adev);
+
+/**
+ * ih_v7_0_init_register_offset - Initialize register offset for ih rings
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Initialize register offset ih rings (IH_V7_0).
+ */
+static void ih_v7_0_init_register_offset(struct amdgpu_device *adev)
+{
+	struct amdgpu_ih_regs *ih_regs;
+
+	/* ih ring 2 is removed
+	 * ih ring and ih ring 1 are available */
+	if (adev->irq.ih.ring_size) {
+		ih_regs = &adev->irq.ih.ih_regs;
+		ih_regs->ih_rb_base = SOC15_REG_OFFSET(OSSSYS, 0, regIH_RB_BASE);
+		ih_regs->ih_rb_base_hi = SOC15_REG_OFFSET(OSSSYS, 0, regIH_RB_BASE_HI);
+		ih_regs->ih_rb_cntl = SOC15_REG_OFFSET(OSSSYS, 0, regIH_RB_CNTL);
+		ih_regs->ih_rb_wptr = SOC15_REG_OFFSET(OSSSYS, 0, regIH_RB_WPTR);
+		ih_regs->ih_rb_rptr = SOC15_REG_OFFSET(OSSSYS, 0, regIH_RB_RPTR);
+		ih_regs->ih_doorbell_rptr = SOC15_REG_OFFSET(OSSSYS, 0, regIH_DOORBELL_RPTR);
+		ih_regs->ih_rb_wptr_addr_lo = SOC15_REG_OFFSET(OSSSYS, 0, regIH_RB_WPTR_ADDR_LO);
+		ih_regs->ih_rb_wptr_addr_hi = SOC15_REG_OFFSET(OSSSYS, 0, regIH_RB_WPTR_ADDR_HI);
+		ih_regs->psp_reg_id = PSP_REG_IH_RB_CNTL;
+	}
+
+	if (adev->irq.ih1.ring_size) {
+		ih_regs = &adev->irq.ih1.ih_regs;
+		ih_regs->ih_rb_base = SOC15_REG_OFFSET(OSSSYS, 0, regIH_RB_BASE_RING1);
+		ih_regs->ih_rb_base_hi = SOC15_REG_OFFSET(OSSSYS, 0, regIH_RB_BASE_HI_RING1);
+		ih_regs->ih_rb_cntl = SOC15_REG_OFFSET(OSSSYS, 0, regIH_RB_CNTL_RING1);
+		ih_regs->ih_rb_wptr = SOC15_REG_OFFSET(OSSSYS, 0, regIH_RB_WPTR_RING1);
+		ih_regs->ih_rb_rptr = SOC15_REG_OFFSET(OSSSYS, 0, regIH_RB_RPTR_RING1);
+		ih_regs->ih_doorbell_rptr = SOC15_REG_OFFSET(OSSSYS, 0, regIH_DOORBELL_RPTR_RING1);
+		ih_regs->psp_reg_id = PSP_REG_IH_RB_CNTL_RING1;
+	}
+}
+
+/**
+ * force_update_wptr_for_self_int - Force update the wptr for self interrupt
+ *
+ * @adev: amdgpu_device pointer
+ * @threshold: threshold to trigger the wptr reporting
+ * @timeout: timeout to trigger the wptr reporting
+ * @enabled: Enable/disable timeout flush mechanism
+ *
+ * threshold input range: 0 ~ 15, default 0,
+ * real_threshold = 2^threshold
+ * timeout input range: 0 ~ 20, default 8,
+ * real_timeout = (2^timeout) * 1024 / (socclk_freq)
+ *
+ * Force update wptr for self interrupt ( >= SIENNA_CICHLID).
+ */
+static void
+force_update_wptr_for_self_int(struct amdgpu_device *adev,
+			       u32 threshold, u32 timeout, bool enabled)
+{
+	u32 ih_cntl, ih_rb_cntl;
+
+	ih_cntl = RREG32_SOC15(OSSSYS, 0, regIH_CNTL2);
+	ih_rb_cntl = RREG32_SOC15(OSSSYS, 0, regIH_RB_CNTL_RING1);
+
+	ih_cntl = REG_SET_FIELD(ih_cntl, IH_CNTL2,
+				SELF_IV_FORCE_WPTR_UPDATE_TIMEOUT, timeout);
+	ih_cntl = REG_SET_FIELD(ih_cntl, IH_CNTL2,
+				SELF_IV_FORCE_WPTR_UPDATE_ENABLE, enabled);
+	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL_RING1,
+				   RB_USED_INT_THRESHOLD, threshold);
+
+	if (amdgpu_sriov_vf(adev) && amdgpu_sriov_reg_indirect_ih(adev)) {
+		if (psp_reg_program(&adev->psp, PSP_REG_IH_RB_CNTL_RING1, ih_rb_cntl))
+			return;
+	} else {
+		WREG32_SOC15(OSSSYS, 0, regIH_RB_CNTL_RING1, ih_rb_cntl);
+	}
+
+	WREG32_SOC15(OSSSYS, 0, regIH_CNTL2, ih_cntl);
+}
+
+/**
+ * ih_v7_0_toggle_ring_interrupts - toggle the interrupt ring buffer
+ *
+ * @adev: amdgpu_device pointer
+ * @ih: amdgpu_ih_ring pointet
+ * @enable: true - enable the interrupts, false - disable the interrupts
+ *
+ * Toggle the interrupt ring buffer (IH_V7_0)
+ */
+static int ih_v7_0_toggle_ring_interrupts(struct amdgpu_device *adev,
+					  struct amdgpu_ih_ring *ih,
+					  bool enable)
+{
+	struct amdgpu_ih_regs *ih_regs;
+	uint32_t tmp;
+
+	ih_regs = &ih->ih_regs;
+
+	tmp = RREG32(ih_regs->ih_rb_cntl);
+	tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, RB_ENABLE, (enable ? 1 : 0));
+	/* enable_intr field is only valid in ring0 */
+	if (ih == &adev->irq.ih)
+		tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, ENABLE_INTR, (enable ? 1 : 0));
+
+	if (amdgpu_sriov_vf(adev) && amdgpu_sriov_reg_indirect_ih(adev)) {
+		if (psp_reg_program(&adev->psp, ih_regs->psp_reg_id, tmp))
+			return -ETIMEDOUT;
+	} else {
+		WREG32(ih_regs->ih_rb_cntl, tmp);
+	}
+
+	if (enable) {
+		ih->enabled = true;
+	} else {
+		/* set rptr, wptr to 0 */
+		WREG32(ih_regs->ih_rb_rptr, 0);
+		WREG32(ih_regs->ih_rb_wptr, 0);
+		ih->enabled = false;
+		ih->rptr = 0;
+	}
+
+	return 0;
+}
+
+/**
+ * ih_v7_0_toggle_interrupts - Toggle all the available interrupt ring buffers
+ *
+ * @adev: amdgpu_device pointer
+ * @enable: enable or disable interrupt ring buffers
+ *
+ * Toggle all the available interrupt ring buffers (IH_V7_0).
+ */
+static int ih_v7_0_toggle_interrupts(struct amdgpu_device *adev, bool enable)
+{
+	struct amdgpu_ih_ring *ih[] = {&adev->irq.ih, &adev->irq.ih1};
+	int i;
+	int r;
+
+	for (i = 0; i < ARRAY_SIZE(ih); i++) {
+		if (ih[i]->ring_size) {
+			r = ih_v7_0_toggle_ring_interrupts(adev, ih[i], enable);
+			if (r)
+				return r;
+		}
+	}
+
+	return 0;
+}
+
+static uint32_t ih_v7_0_rb_cntl(struct amdgpu_ih_ring *ih, uint32_t ih_rb_cntl)
+{
+	int rb_bufsz = order_base_2(ih->ring_size / 4);
+
+	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL,
+				   MC_SPACE, ih->use_bus_addr ? 2 : 4);
+	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL,
+				   WPTR_OVERFLOW_CLEAR, 1);
+	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL,
+				   WPTR_OVERFLOW_ENABLE, 1);
+	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, RB_SIZE, rb_bufsz);
+	/* Ring Buffer write pointer writeback. If enabled, IH_RB_WPTR register
+	 * value is written to memory
+	 */
+	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL,
+				   WPTR_WRITEBACK_ENABLE, 1);
+	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, MC_SNOOP, 1);
+	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, MC_RO, 0);
+	ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, MC_VMID, 0);
+
+	return ih_rb_cntl;
+}
+
+static uint32_t ih_v7_0_doorbell_rptr(struct amdgpu_ih_ring *ih)
+{
+	u32 ih_doorbell_rtpr = 0;
+
+	if (ih->use_doorbell) {
+		ih_doorbell_rtpr = REG_SET_FIELD(ih_doorbell_rtpr,
+						 IH_DOORBELL_RPTR, OFFSET,
+						 ih->doorbell_index);
+		ih_doorbell_rtpr = REG_SET_FIELD(ih_doorbell_rtpr,
+						 IH_DOORBELL_RPTR,
+						 ENABLE, 1);
+	} else {
+		ih_doorbell_rtpr = REG_SET_FIELD(ih_doorbell_rtpr,
+						 IH_DOORBELL_RPTR,
+						 ENABLE, 0);
+	}
+	return ih_doorbell_rtpr;
+}
+
+/**
+ * ih_v7_0_enable_ring - enable an ih ring buffer
+ *
+ * @adev: amdgpu_device pointer
+ * @ih: amdgpu_ih_ring pointer
+ *
+ * Enable an ih ring buffer (IH_V7_0)
+ */
+static int ih_v7_0_enable_ring(struct amdgpu_device *adev,
+				      struct amdgpu_ih_ring *ih)
+{
+	struct amdgpu_ih_regs *ih_regs;
+	uint32_t tmp;
+
+	ih_regs = &ih->ih_regs;
+
+	/* Ring Buffer base. [39:8] of 40-bit address of the beginning of the ring buffer*/
+	WREG32(ih_regs->ih_rb_base, ih->gpu_addr >> 8);
+	WREG32(ih_regs->ih_rb_base_hi, (ih->gpu_addr >> 40) & 0xff);
+
+	tmp = RREG32(ih_regs->ih_rb_cntl);
+	tmp = ih_v7_0_rb_cntl(ih, tmp);
+	if (ih == &adev->irq.ih)
+		tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, RPTR_REARM, !!adev->irq.msi_enabled);
+	if (ih == &adev->irq.ih1) {
+		tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_ENABLE, 0);
+		tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, RB_FULL_DRAIN_ENABLE, 1);
+	}
+
+	if (amdgpu_sriov_vf(adev) && amdgpu_sriov_reg_indirect_ih(adev)) {
+		if (psp_reg_program(&adev->psp, ih_regs->psp_reg_id, tmp)) {
+			DRM_ERROR("PSP program IH_RB_CNTL failed!\n");
+			return -ETIMEDOUT;
+		}
+	} else {
+		WREG32(ih_regs->ih_rb_cntl, tmp);
+	}
+
+	if (ih == &adev->irq.ih) {
+		/* set the ih ring 0 writeback address whether it's enabled or not */
+		WREG32(ih_regs->ih_rb_wptr_addr_lo, lower_32_bits(ih->wptr_addr));
+		WREG32(ih_regs->ih_rb_wptr_addr_hi, upper_32_bits(ih->wptr_addr) & 0xFFFF);
+	}
+
+	/* set rptr, wptr to 0 */
+	WREG32(ih_regs->ih_rb_wptr, 0);
+	WREG32(ih_regs->ih_rb_rptr, 0);
+
+	WREG32(ih_regs->ih_doorbell_rptr, ih_v7_0_doorbell_rptr(ih));
+
+	return 0;
+}
+
+/**
+ * ih_v7_0_irq_init - init and enable the interrupt ring
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Allocate a ring buffer for the interrupt controller,
+ * enable the RLC, disable interrupts, enable the IH
+ * ring buffer and enable it.
+ * Called at device load and reume.
+ * Returns 0 for success, errors for failure.
+ */
+static int ih_v7_0_irq_init(struct amdgpu_device *adev)
+{
+	struct amdgpu_ih_ring *ih[] = {&adev->irq.ih, &adev->irq.ih1};
+	u32 ih_chicken;
+	u32 tmp;
+	int ret;
+	int i;
+
+	/* disable irqs */
+	ret = ih_v7_0_toggle_interrupts(adev, false);
+	if (ret)
+		return ret;
+
+	adev->nbio.funcs->ih_control(adev);
+
+	if (unlikely((adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT) ||
+		     (adev->firmware.load_type == AMDGPU_FW_LOAD_RLC_BACKDOOR_AUTO))) {
+		if (ih[0]->use_bus_addr) {
+			ih_chicken = RREG32_SOC15(OSSSYS, 0, regIH_CHICKEN);
+			ih_chicken = REG_SET_FIELD(ih_chicken,
+					IH_CHICKEN, MC_SPACE_GPA_ENABLE, 1);
+			WREG32_SOC15(OSSSYS, 0, regIH_CHICKEN, ih_chicken);
+		}
+	}
+
+	for (i = 0; i < ARRAY_SIZE(ih); i++) {
+		if (ih[i]->ring_size) {
+			ret = ih_v7_0_enable_ring(adev, ih[i]);
+			if (ret)
+				return ret;
+		}
+	}
+
+	/* update doorbell range for ih ring 0 */
+	adev->nbio.funcs->ih_doorbell_range(adev, ih[0]->use_doorbell,
+					    ih[0]->doorbell_index);
+
+	tmp = RREG32_SOC15(OSSSYS, 0, regIH_STORM_CLIENT_LIST_CNTL);
+	tmp = REG_SET_FIELD(tmp, IH_STORM_CLIENT_LIST_CNTL,
+			    CLIENT18_IS_STORM_CLIENT, 1);
+	WREG32_SOC15(OSSSYS, 0, regIH_STORM_CLIENT_LIST_CNTL, tmp);
+
+	tmp = RREG32_SOC15(OSSSYS, 0, regIH_INT_FLOOD_CNTL);
+	tmp = REG_SET_FIELD(tmp, IH_INT_FLOOD_CNTL, FLOOD_CNTL_ENABLE, 1);
+	WREG32_SOC15(OSSSYS, 0, regIH_INT_FLOOD_CNTL, tmp);
+
+	/* GC/MMHUB UTCL2 page fault interrupts are configured as
+	 * MSI storm capable interrupts by deafult. The delay is
+	 * used to avoid ISR being called too frequently
+	 * when page fault happens on several continuous page
+	 * and thus avoid MSI storm */
+	tmp = RREG32_SOC15(OSSSYS, 0, regIH_MSI_STORM_CTRL);
+	tmp = REG_SET_FIELD(tmp, IH_MSI_STORM_CTRL,
+			    DELAY, 3);
+	WREG32_SOC15(OSSSYS, 0, regIH_MSI_STORM_CTRL, tmp);
+
+	/* Redirect the interrupts to IH RB1 for dGPU */
+	if (adev->irq.ih1.ring_size) {
+		tmp = RREG32_SOC15(OSSSYS, 0, regIH_RING1_CLIENT_CFG_INDEX);
+		tmp = REG_SET_FIELD(tmp, IH_RING1_CLIENT_CFG_INDEX, INDEX, 0);
+		WREG32_SOC15(OSSSYS, 0, regIH_RING1_CLIENT_CFG_INDEX, tmp);
+
+		tmp = RREG32_SOC15(OSSSYS, 0, regIH_RING1_CLIENT_CFG_DATA);
+		tmp = REG_SET_FIELD(tmp, IH_RING1_CLIENT_CFG_DATA, CLIENT_ID, 0xa);
+		tmp = REG_SET_FIELD(tmp, IH_RING1_CLIENT_CFG_DATA, SOURCE_ID, 0x0);
+		tmp = REG_SET_FIELD(tmp, IH_RING1_CLIENT_CFG_DATA,
+				    SOURCE_ID_MATCH_ENABLE, 0x1);
+
+		WREG32_SOC15(OSSSYS, 0, regIH_RING1_CLIENT_CFG_DATA, tmp);
+	}
+
+	pci_set_master(adev->pdev);
+
+	/* enable interrupts */
+	ret = ih_v7_0_toggle_interrupts(adev, true);
+	if (ret)
+		return ret;
+	/* enable wptr force update for self int */
+	force_update_wptr_for_self_int(adev, 0, 8, true);
+
+	if (adev->irq.ih_soft.ring_size)
+		adev->irq.ih_soft.enabled = true;
+
+	return 0;
+}
+
+/**
+ * ih_v7_0_irq_disable - disable interrupts
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Disable interrupts on the hw.
+ */
+static void ih_v7_0_irq_disable(struct amdgpu_device *adev)
+{
+	force_update_wptr_for_self_int(adev, 0, 8, false);
+	ih_v7_0_toggle_interrupts(adev, false);
+
+	/* Wait and acknowledge irq */
+	mdelay(1);
+}
+
+/**
+ * ih_v7_0_get_wptr() - get the IH ring buffer wptr
+ *
+ * @adev: amdgpu_device pointer
+ * @ih: IH ring buffer to fetch wptr
+ *
+ * Get the IH ring buffer wptr from either the register
+ * or the writeback memory buffer.  Also check for
+ * ring buffer overflow and deal with it.
+ * Returns the value of the wptr.
+ */
+static u32 ih_v7_0_get_wptr(struct amdgpu_device *adev,
+			      struct amdgpu_ih_ring *ih)
+{
+	u32 wptr, tmp;
+	struct amdgpu_ih_regs *ih_regs;
+
+	wptr = le32_to_cpu(*ih->wptr_cpu);
+	ih_regs = &ih->ih_regs;
+
+	if (!REG_GET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW))
+		goto out;
+
+	wptr = RREG32_NO_KIQ(ih_regs->ih_rb_wptr);
+	if (!REG_GET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW))
+		goto out;
+	wptr = REG_SET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW, 0);
+
+	/* When a ring buffer overflow happen start parsing interrupt
+	 * from the last not overwritten vector (wptr + 32). Hopefully
+	 * this should allow us to catch up.
+	 */
+	tmp = (wptr + 32) & ih->ptr_mask;
+	dev_warn(adev->dev, "IH ring buffer overflow "
+		 "(0x%08X, 0x%08X, 0x%08X)\n",
+		 wptr, ih->rptr, tmp);
+	ih->rptr = tmp;
+
+	tmp = RREG32_NO_KIQ(ih_regs->ih_rb_cntl);
+	tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 1);
+	WREG32_NO_KIQ(ih_regs->ih_rb_cntl, tmp);
+
+	/* Unset the CLEAR_OVERFLOW bit immediately so new overflows
+	 * can be detected.
+	 */
+	tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 0);
+	WREG32_NO_KIQ(ih_regs->ih_rb_cntl, tmp);
+out:
+	return (wptr & ih->ptr_mask);
+}
+
+/**
+ * ih_v7_0_irq_rearm - rearm IRQ if lost
+ *
+ * @adev: amdgpu_device pointer
+ * @ih: IH ring to match
+ *
+ */
+static void ih_v7_0_irq_rearm(struct amdgpu_device *adev,
+			       struct amdgpu_ih_ring *ih)
+{
+	uint32_t v = 0;
+	uint32_t i = 0;
+	struct amdgpu_ih_regs *ih_regs;
+
+	ih_regs = &ih->ih_regs;
+
+	/* Rearm IRQ / re-write doorbell if doorbell write is lost */
+	for (i = 0; i < MAX_REARM_RETRY; i++) {
+		v = RREG32_NO_KIQ(ih_regs->ih_rb_rptr);
+		if ((v < ih->ring_size) && (v != ih->rptr))
+			WDOORBELL32(ih->doorbell_index, ih->rptr);
+		else
+			break;
+	}
+}
+
+/**
+ * ih_v7_0_set_rptr - set the IH ring buffer rptr
+ *
+ * @adev: amdgpu_device pointer
+ * @ih: IH ring buffer to set rptr
+ */
+static void ih_v7_0_set_rptr(struct amdgpu_device *adev,
+			       struct amdgpu_ih_ring *ih)
+{
+	struct amdgpu_ih_regs *ih_regs;
+
+	if (ih->use_doorbell) {
+		/* XXX check if swapping is necessary on BE */
+		*ih->rptr_cpu = ih->rptr;
+		WDOORBELL32(ih->doorbell_index, ih->rptr);
+
+		if (amdgpu_sriov_vf(adev))
+			ih_v7_0_irq_rearm(adev, ih);
+	} else {
+		ih_regs = &ih->ih_regs;
+		WREG32(ih_regs->ih_rb_rptr, ih->rptr);
+	}
+}
+
+/**
+ * ih_v7_0_self_irq - dispatch work for ring 1
+ *
+ * @adev: amdgpu_device pointer
+ * @source: irq source
+ * @entry: IV with WPTR update
+ *
+ * Update the WPTR from the IV and schedule work to handle the entries.
+ */
+static int ih_v7_0_self_irq(struct amdgpu_device *adev,
+			      struct amdgpu_irq_src *source,
+			      struct amdgpu_iv_entry *entry)
+{
+	uint32_t wptr = cpu_to_le32(entry->src_data[0]);
+
+	switch (entry->ring_id) {
+	case 1:
+		*adev->irq.ih1.wptr_cpu = wptr;
+		schedule_work(&adev->irq.ih1_work);
+		break;
+	default: break;
+	}
+	return 0;
+}
+
+static const struct amdgpu_irq_src_funcs ih_v7_0_self_irq_funcs = {
+	.process = ih_v7_0_self_irq,
+};
+
+static void ih_v7_0_set_self_irq_funcs(struct amdgpu_device *adev)
+{
+	adev->irq.self_irq.num_types = 0;
+	adev->irq.self_irq.funcs = &ih_v7_0_self_irq_funcs;
+}
+
+static int ih_v7_0_early_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	ih_v7_0_set_interrupt_funcs(adev);
+	ih_v7_0_set_self_irq_funcs(adev);
+	return 0;
+}
+
+static int ih_v7_0_sw_init(struct amdgpu_ip_block *ip_block)
+{
+	int r;
+	struct amdgpu_device *adev = ip_block->adev;
+	bool use_bus_addr;
+
+	r = amdgpu_irq_add_id(adev, SOC21_IH_CLIENTID_IH, 0,
+			      &adev->irq.self_irq);
+
+	if (r)
+		return r;
+
+	/* use gpu virtual address for ih ring
+	 * until ih_checken is programmed to allow
+	 * use bus address for ih ring by psp bl */
+	use_bus_addr = adev->firmware.load_type != AMDGPU_FW_LOAD_PSP;
+	r = amdgpu_ih_ring_init(adev, &adev->irq.ih, 256 * 1024, use_bus_addr);
+	if (r)
+		return r;
+
+	adev->irq.ih.use_doorbell = true;
+	adev->irq.ih.doorbell_index = adev->doorbell_index.ih << 1;
+
+	if (!(adev->flags & AMD_IS_APU)) {
+		r = amdgpu_ih_ring_init(adev, &adev->irq.ih1, IH_RING_SIZE,
+					use_bus_addr);
+		if (r)
+			return r;
+
+		adev->irq.ih1.use_doorbell = true;
+		adev->irq.ih1.doorbell_index = (adev->doorbell_index.ih + 1) << 1;
+	}
+
+	/* initialize ih control register offset */
+	ih_v7_0_init_register_offset(adev);
+
+	r = amdgpu_ih_ring_init(adev, &adev->irq.ih_soft, PAGE_SIZE, true);
+	if (r)
+		return r;
+
+	r = amdgpu_irq_init(adev);
+
+	return r;
+}
+
+static int ih_v7_0_sw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	amdgpu_irq_fini_sw(adev);
+
+	return 0;
+}
+
+static int ih_v7_0_hw_init(struct amdgpu_ip_block *ip_block)
+{
+	int r;
+	struct amdgpu_device *adev = ip_block->adev;
+
+	r = ih_v7_0_irq_init(adev);
+	if (r)
+		return r;
+
+	return 0;
+}
+
+static int ih_v7_0_hw_fini(struct amdgpu_ip_block *ip_block)
+{
+	ih_v7_0_irq_disable(ip_block->adev);
+
+	return 0;
+}
+
+static int ih_v7_0_suspend(struct amdgpu_ip_block *ip_block)
+{
+	return ih_v7_0_hw_fini(ip_block);
+}
+
+static int ih_v7_0_resume(struct amdgpu_ip_block *ip_block)
+{
+	return ih_v7_0_hw_init(ip_block);
+}
+
+static bool ih_v7_0_is_idle(struct amdgpu_ip_block *ip_block)
+{
+	/* todo */
+	return true;
+}
+
+static int ih_v7_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
+{
+	/* todo */
+	return -ETIMEDOUT;
+}
+
+static int ih_v7_0_soft_reset(struct amdgpu_ip_block *ip_block)
+{
+	/* todo */
+	return 0;
+}
+
+static void ih_v7_0_update_clockgating_state(struct amdgpu_device *adev,
+					       bool enable)
+{
+	uint32_t data, def, field_val;
+
+	if (adev->cg_flags & AMD_CG_SUPPORT_IH_CG) {
+		def = data = RREG32_SOC15(OSSSYS, 0, regIH_CLK_CTRL);
+		field_val = enable ? 0 : 1;
+		data = REG_SET_FIELD(data, IH_CLK_CTRL,
+				     DBUS_MUX_CLK_SOFT_OVERRIDE, field_val);
+		data = REG_SET_FIELD(data, IH_CLK_CTRL,
+				     OSSSYS_SHARE_CLK_SOFT_OVERRIDE, field_val);
+		data = REG_SET_FIELD(data, IH_CLK_CTRL,
+				     LIMIT_SMN_CLK_SOFT_OVERRIDE, field_val);
+		data = REG_SET_FIELD(data, IH_CLK_CTRL,
+				     DYN_CLK_SOFT_OVERRIDE, field_val);
+		data = REG_SET_FIELD(data, IH_CLK_CTRL,
+				     REG_CLK_SOFT_OVERRIDE, field_val);
+		if (def != data)
+			WREG32_SOC15(OSSSYS, 0, regIH_CLK_CTRL, data);
+	}
+
+	return;
+}
+
+static int ih_v7_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
+					   enum amd_clockgating_state state)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	ih_v7_0_update_clockgating_state(adev,
+				state == AMD_CG_STATE_GATE);
+	return 0;
+}
+
+static void ih_v7_0_update_ih_mem_power_gating(struct amdgpu_device *adev,
+					       bool enable)
+{
+	uint32_t ih_mem_pwr_cntl;
+
+	/* Disable ih sram power cntl before switch powergating mode */
+	ih_mem_pwr_cntl = RREG32_SOC15(OSSSYS, 0, regIH_MEM_POWER_CTRL);
+	ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+					IH_BUFFER_MEM_POWER_CTRL_EN, 0);
+	WREG32_SOC15(OSSSYS, 0, regIH_MEM_POWER_CTRL, ih_mem_pwr_cntl);
+
+	/* It is recommended to set mem powergating mode to DS mode */
+	if (enable) {
+		/* mem power mode */
+		ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+						IH_BUFFER_MEM_POWER_LS_EN, 0);
+		ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+						IH_BUFFER_MEM_POWER_DS_EN, 1);
+		ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+						IH_BUFFER_MEM_POWER_SD_EN, 0);
+		/* cam mem power mode */
+		ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+						IH_RETRY_INT_CAM_MEM_POWER_LS_EN, 0);
+		ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+						IH_RETRY_INT_CAM_MEM_POWER_DS_EN, 1);
+		ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+						IH_RETRY_INT_CAM_MEM_POWER_SD_EN, 0);
+		/* re-enable power cntl */
+		ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+						IH_BUFFER_MEM_POWER_CTRL_EN, 1);
+	} else {
+		/* mem power mode */
+		ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+						IH_BUFFER_MEM_POWER_LS_EN, 0);
+		ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+						IH_BUFFER_MEM_POWER_DS_EN, 0);
+		ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+						IH_BUFFER_MEM_POWER_SD_EN, 0);
+		/* cam mem power mode */
+		ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+						IH_RETRY_INT_CAM_MEM_POWER_LS_EN, 0);
+		ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+						IH_RETRY_INT_CAM_MEM_POWER_DS_EN, 0);
+		ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+						IH_RETRY_INT_CAM_MEM_POWER_SD_EN, 0);
+		/* re-enable power cntl*/
+		ih_mem_pwr_cntl = REG_SET_FIELD(ih_mem_pwr_cntl, IH_MEM_POWER_CTRL,
+						IH_BUFFER_MEM_POWER_CTRL_EN, 1);
+	}
+
+	WREG32_SOC15(OSSSYS, 0, regIH_MEM_POWER_CTRL, ih_mem_pwr_cntl);
+}
+
+static int ih_v7_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
+					 enum amd_powergating_state state)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	bool enable = (state == AMD_PG_STATE_GATE);
+
+	if (adev->pg_flags & AMD_PG_SUPPORT_IH_SRAM_PG)
+		ih_v7_0_update_ih_mem_power_gating(adev, enable);
+
+	return 0;
+}
+
+static void ih_v7_0_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	if (!RREG32_SOC15(OSSSYS, 0, regIH_CLK_CTRL))
+		*flags |= AMD_CG_SUPPORT_IH_CG;
+
+	return;
+}
+
+static const struct amd_ip_funcs ih_v7_0_ip_funcs = {
+	.name = "ih_v7_0",
+	.early_init = ih_v7_0_early_init,
+	.sw_init = ih_v7_0_sw_init,
+	.sw_fini = ih_v7_0_sw_fini,
+	.hw_init = ih_v7_0_hw_init,
+	.hw_fini = ih_v7_0_hw_fini,
+	.suspend = ih_v7_0_suspend,
+	.resume = ih_v7_0_resume,
+	.is_idle = ih_v7_0_is_idle,
+	.wait_for_idle = ih_v7_0_wait_for_idle,
+	.soft_reset = ih_v7_0_soft_reset,
+	.set_clockgating_state = ih_v7_0_set_clockgating_state,
+	.set_powergating_state = ih_v7_0_set_powergating_state,
+	.get_clockgating_state = ih_v7_0_get_clockgating_state,
+};
+
+static const struct amdgpu_ih_funcs ih_v7_0_funcs = {
+	.get_wptr = ih_v7_0_get_wptr,
+	.decode_iv = amdgpu_ih_decode_iv_helper,
+	.decode_iv_ts = amdgpu_ih_decode_iv_ts_helper,
+	.set_rptr = ih_v7_0_set_rptr
+};
+
+static void ih_v7_0_set_interrupt_funcs(struct amdgpu_device *adev)
+{
+	adev->irq.ih_funcs = &ih_v7_0_funcs;
+}
+
+const struct amdgpu_ip_block_version ih_v7_0_ip_block =
+{
+	.type = AMD_IP_BLOCK_TYPE_IH,
+	.major = 7,
+	.minor = 0,
+	.rev = 0,
+	.funcs = &ih_v7_0_ip_funcs,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/ih_v7_0.h b/drivers/gpu/drm/amd/amdgpu/ih_v7_0.h
new file mode 100644
index 000000000000..af9dcbc451fd
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/ih_v7_0.h
@@ -0,0 +1,28 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __IH_V7_0_IH_H__
+#define __IH_V7_0_IH_H__
+
+extern const struct amdgpu_ip_block_version ih_v7_0_ip_block;
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/imu_v11_0.c b/drivers/gpu/drm/amd/amdgpu/imu_v11_0.c
new file mode 100644
index 000000000000..cc626036ed9c
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/imu_v11_0.c
@@ -0,0 +1,393 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/firmware.h>
+#include "amdgpu.h"
+#include "amdgpu_imu.h"
+#include "amdgpu_dpm.h"
+
+#include "imu_v11_0_3.h"
+
+#include "gc/gc_11_0_0_offset.h"
+#include "gc/gc_11_0_0_sh_mask.h"
+
+MODULE_FIRMWARE("amdgpu/gc_11_0_0_imu.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_0_0_imu_kicker.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_0_1_imu.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_0_2_imu.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_0_3_imu.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_0_4_imu.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_5_0_imu.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_5_1_imu.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_5_2_imu.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_5_3_imu.bin");
+
+static int imu_v11_0_init_microcode(struct amdgpu_device *adev)
+{
+	char ucode_prefix[30];
+	int err;
+	const struct imu_firmware_header_v1_0 *imu_hdr;
+	struct amdgpu_firmware_info *info = NULL;
+
+	DRM_DEBUG("\n");
+
+	amdgpu_ucode_ip_version_decode(adev, GC_HWIP, ucode_prefix, sizeof(ucode_prefix));
+	if (amdgpu_is_kicker_fw(adev))
+		err = amdgpu_ucode_request(adev, &adev->gfx.imu_fw, AMDGPU_UCODE_REQUIRED,
+					   "amdgpu/%s_imu_kicker.bin", ucode_prefix);
+	else
+		err = amdgpu_ucode_request(adev, &adev->gfx.imu_fw, AMDGPU_UCODE_REQUIRED,
+					   "amdgpu/%s_imu.bin", ucode_prefix);
+	if (err)
+		goto out;
+
+	imu_hdr = (const struct imu_firmware_header_v1_0 *)adev->gfx.imu_fw->data;
+	//adev->gfx.imu_feature_version = le32_to_cpu(imu_hdr->ucode_feature_version);
+	
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
+		info = &adev->firmware.ucode[AMDGPU_UCODE_ID_IMU_I];
+		info->ucode_id = AMDGPU_UCODE_ID_IMU_I;
+		info->fw = adev->gfx.imu_fw;
+		adev->firmware.fw_size +=
+			ALIGN(le32_to_cpu(imu_hdr->imu_iram_ucode_size_bytes), PAGE_SIZE);
+		info = &adev->firmware.ucode[AMDGPU_UCODE_ID_IMU_D];
+		info->ucode_id = AMDGPU_UCODE_ID_IMU_D;
+		info->fw = adev->gfx.imu_fw;
+		adev->firmware.fw_size +=
+			ALIGN(le32_to_cpu(imu_hdr->imu_dram_ucode_size_bytes), PAGE_SIZE);
+	} else
+		adev->gfx.imu_fw_version = le32_to_cpu(imu_hdr->header.ucode_version);
+
+out:
+	if (err) {
+		dev_err(adev->dev,
+			"gfx11: Failed to load firmware \"%s_imu.bin\"\n",
+			ucode_prefix);
+		amdgpu_ucode_release(&adev->gfx.imu_fw);
+	}
+
+	return err;
+}
+
+static int imu_v11_0_load_microcode(struct amdgpu_device *adev)
+{
+	const struct imu_firmware_header_v1_0 *hdr;
+	const __le32 *fw_data;
+	unsigned i, fw_size;
+
+	if (!adev->gfx.imu_fw)
+		return -EINVAL;
+
+	hdr = (const struct imu_firmware_header_v1_0 *)adev->gfx.imu_fw->data;
+	//amdgpu_ucode_print_rlc_hdr(&hdr->header);
+
+	fw_data = (const __le32 *)(adev->gfx.imu_fw->data +
+			le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+	fw_size = le32_to_cpu(hdr->imu_iram_ucode_size_bytes) / 4;
+
+	WREG32_SOC15(GC, 0, regGFX_IMU_I_RAM_ADDR, 0);
+
+	for (i = 0; i < fw_size; i++)
+		WREG32_SOC15(GC, 0, regGFX_IMU_I_RAM_DATA, le32_to_cpup(fw_data++));
+
+	WREG32_SOC15(GC, 0, regGFX_IMU_I_RAM_ADDR, adev->gfx.imu_fw_version);
+
+	fw_data = (const __le32 *)(adev->gfx.imu_fw->data +
+			le32_to_cpu(hdr->header.ucode_array_offset_bytes) +
+			le32_to_cpu(hdr->imu_iram_ucode_size_bytes));
+	fw_size = le32_to_cpu(hdr->imu_dram_ucode_size_bytes) / 4;
+
+	WREG32_SOC15(GC, 0, regGFX_IMU_D_RAM_ADDR, 0);
+
+	for (i = 0; i < fw_size; i++)
+		WREG32_SOC15(GC, 0, regGFX_IMU_D_RAM_DATA, le32_to_cpup(fw_data++));
+
+	WREG32_SOC15(GC, 0, regGFX_IMU_D_RAM_ADDR, adev->gfx.imu_fw_version);
+
+	return 0;
+}
+
+static int imu_v11_0_wait_for_reset_status(struct amdgpu_device *adev)
+{
+	int i, imu_reg_val = 0;
+
+	for (i = 0; i < adev->usec_timeout; i++) {
+		imu_reg_val = RREG32_SOC15(GC, 0, regGFX_IMU_GFX_RESET_CTRL);
+		if ((imu_reg_val & 0x1f) == 0x1f)
+			break;
+		udelay(1);
+	}
+
+	if (i >= adev->usec_timeout) {
+		dev_err(adev->dev, "init imu: IMU start timeout\n");
+		return -ETIMEDOUT;
+	}
+
+	return 0;
+}
+
+static void imu_v11_0_setup(struct amdgpu_device *adev)
+{
+	int imu_reg_val;
+
+	//enable IMU debug mode
+	WREG32_SOC15(GC, 0, regGFX_IMU_C2PMSG_ACCESS_CTRL0, 0xffffff);
+	WREG32_SOC15(GC, 0, regGFX_IMU_C2PMSG_ACCESS_CTRL1, 0xffff);
+
+	if (adev->gfx.imu.mode == DEBUG_MODE) {
+		imu_reg_val = RREG32_SOC15(GC, 0, regGFX_IMU_C2PMSG_16);
+		imu_reg_val |= 0x1;
+		WREG32_SOC15(GC, 0, regGFX_IMU_C2PMSG_16, imu_reg_val);
+	}
+
+	//disable imu Rtavfs, SmsRepair, DfllBTC, and ClkB
+	imu_reg_val = RREG32_SOC15(GC, 0, regGFX_IMU_SCRATCH_10);
+	imu_reg_val |= 0x10007;
+	WREG32_SOC15(GC, 0, regGFX_IMU_SCRATCH_10, imu_reg_val);
+}
+
+static int imu_v11_0_start(struct amdgpu_device *adev)
+{
+	int imu_reg_val;
+
+	//Start IMU by set GFX_IMU_CORE_CTRL.CRESET = 0
+	imu_reg_val = RREG32_SOC15(GC, 0, regGFX_IMU_CORE_CTRL);
+	imu_reg_val &= 0xfffffffe;
+	WREG32_SOC15(GC, 0, regGFX_IMU_CORE_CTRL, imu_reg_val);
+
+	if (adev->flags & AMD_IS_APU)
+		amdgpu_dpm_set_gfx_power_up_by_imu(adev);
+
+	return imu_v11_0_wait_for_reset_status(adev);
+}
+
+static const struct imu_rlc_ram_golden imu_rlc_ram_golden_11[] =
+{
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGUS_IO_RD_COMBINE_FLUSH, 0x00055555, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGUS_IO_WR_COMBINE_FLUSH, 0x00055555, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGUS_DRAM_COMBINE_FLUSH, 0x00555555, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGUS_MISC2, 0x00001ffe, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGUS_SDP_CREDITS , 0x003f3fff, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGUS_SDP_TAG_RESERVE1, 0x00000000, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGUS_SDP_VCC_RESERVE0, 0x00041000, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGUS_SDP_VCC_RESERVE1, 0x00000000, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGUS_SDP_VCD_RESERVE0, 0x00040000, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGUS_SDP_VCD_RESERVE1, 0x00000000, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGUS_MISC, 0x00000017, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGUS_SDP_ENABLE, 0x00000001, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCEA_SDP_CREDITS , 0x003f3fbf, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCEA_SDP_TAG_RESERVE0, 0x10201000, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCEA_SDP_TAG_RESERVE1, 0x00000080, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCEA_SDP_VCC_RESERVE0, 0x1d041040, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCEA_SDP_VCC_RESERVE1, 0x80000000, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCEA_SDP_IO_PRIORITY, 0x88888888, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCEA_MAM_CTRL, 0x0000d800, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCEA_SDP_ARB_FINAL, 0x000003f7, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCEA_SDP_ENABLE, 0x00000001, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_L2_PROTECTION_FAULT_CNTL2, 0x00020000, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_APT_CNTL, 0x0000000c, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_CACHEABLE_DRAM_ADDRESS_END, 0x000fffff, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCEA_MISC, 0x0c48bff0, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regCC_GC_SA_UNIT_DISABLE, 0x00fffc01, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regCC_GC_PRIM_CONFIG, 0x000fffe1, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regCC_RB_BACKEND_DISABLE, 0x0fffff01, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regCC_GC_SHADER_ARRAY_CONFIG, 0xfffe0001, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_MX_L1_TLB_CNTL, 0x00000500, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_SYSTEM_APERTURE_LOW_ADDR, 0x00000001, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_SYSTEM_APERTURE_HIGH_ADDR, 0x00000000, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_LOCAL_FB_ADDRESS_START, 0x00000000, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_LOCAL_FB_ADDRESS_END, 0x000fffff, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_CONTEXT0_CNTL, 0x00000000, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_CONTEXT1_CNTL, 0x00000000, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_NB_TOP_OF_DRAM_SLOT1, 0xff800000, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_NB_LOWER_TOP_OF_DRAM2, 0x00000001, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_NB_UPPER_TOP_OF_DRAM2, 0x00000fff, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_L2_PROTECTION_FAULT_CNTL, 0x00001ffc, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_MX_L1_TLB_CNTL, 0x00000501, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_L2_CNTL, 0x00080603, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_L2_CNTL2, 0x00000003, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_L2_CNTL3, 0x00100003, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_L2_CNTL5, 0x00003fe0, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_CONTEXT0_CNTL, 0x00000001, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_L2_CONTEXT0_PER_PFVF_PTE_CACHE_FRAGMENT_SIZES, 0x00000c00, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_CONTEXT1_CNTL, 0x00000001, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_L2_CONTEXT1_PER_PFVF_PTE_CACHE_FRAGMENT_SIZES, 0x00000c00, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGB_ADDR_CONFIG, 0x00000545, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGL2_PIPE_STEER_0, 0x13455431, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGL2_PIPE_STEER_1, 0x13455431, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGL2_PIPE_STEER_2, 0x76027602, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGL2_PIPE_STEER_3, 0x76207620, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGB_ADDR_CONFIG, 0x00000345, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCUTCL2_HARVEST_BYPASS_GROUPS, 0x0000003e, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_FB_LOCATION_BASE, 0x00006000, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_FB_LOCATION_TOP, 0x000061ff, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_APT_CNTL, 0x0000000c, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_AGP_BASE, 0x00000000, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_AGP_BOT, 0x00000002, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_AGP_TOP, 0x00000000, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_L2_PROTECTION_FAULT_CNTL2, 0x00020000, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regSDMA0_UCODE_SELFLOAD_CONTROL, 0x00000210, 0), 
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regSDMA1_UCODE_SELFLOAD_CONTROL, 0x00000210, 0), 
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regCPC_PSP_DEBUG, CPC_PSP_DEBUG__GPA_OVERRIDE_MASK, 0), 
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regCPG_PSP_DEBUG, CPG_PSP_DEBUG__GPA_OVERRIDE_MASK, 0)
+};
+
+static const struct imu_rlc_ram_golden imu_rlc_ram_golden_11_0_2[] =
+{
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCEA_MISC, 0x0c48bff0, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCEA_SDP_CREDITS, 0x003f3fbf, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCEA_SDP_TAG_RESERVE0, 0x10200800, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCEA_SDP_TAG_RESERVE1, 0x00000088, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCEA_SDP_VCC_RESERVE0, 0x1d041040, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCEA_SDP_VCC_RESERVE1, 0x80000000, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCEA_SDP_IO_PRIORITY, 0x88888888, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCEA_MAM_CTRL, 0x0000d800, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCEA_SDP_ARB_FINAL, 0x000007ef, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCEA_DRAM_PAGE_BURST, 0x20080200, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCEA_SDP_ENABLE, 0x00000001, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_APT_CNTL, 0x0000000c, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_CACHEABLE_DRAM_ADDRESS_END, 0x000fffff, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGUS_IO_RD_COMBINE_FLUSH, 0x00055555, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGUS_IO_WR_COMBINE_FLUSH, 0x00055555, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGUS_DRAM_COMBINE_FLUSH, 0x00555555, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGUS_MISC2, 0x00001ffe, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGUS_SDP_CREDITS, 0x003f3fff, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGUS_SDP_TAG_RESERVE1, 0x00000000, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGUS_SDP_VCC_RESERVE0, 0x00041000, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGUS_SDP_VCC_RESERVE1, 0x00000000, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGUS_SDP_VCD_RESERVE0, 0x00040000, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGUS_SDP_VCD_RESERVE1, 0x00000000, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGUS_MISC, 0x00000017, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGUS_SDP_ENABLE, 0x00000001, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regCC_GC_SA_UNIT_DISABLE, 0x00fffc01, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regCC_GC_PRIM_CONFIG, 0x000fffe1, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regCC_RB_BACKEND_DISABLE, 0x00000f01, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regCC_GC_SHADER_ARRAY_CONFIG, 0xfffe0001, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGL1_PIPE_STEER, 0x000000e4, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regCH_PIPE_STEER, 0x000000e4, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGL2_PIPE_STEER_0, 0x01231023, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGB_ADDR_CONFIG, 0x00000243, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCUTCL2_HARVEST_BYPASS_GROUPS, 0x00000002, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_MX_L1_TLB_CNTL, 0x00000500, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_SYSTEM_APERTURE_LOW_ADDR, 0x00000001, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_SYSTEM_APERTURE_HIGH_ADDR, 0x00000000, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_LOCAL_FB_ADDRESS_START, 0x00000000, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_LOCAL_FB_ADDRESS_END, 0x000001ff, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_FB_LOCATION_BASE, 0x00006000, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_FB_LOCATION_TOP, 0x000061ff, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_CONTEXT0_CNTL, 0x00000000, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_CONTEXT1_CNTL, 0x00000000, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_APT_CNTL, 0x0000000c, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_NB_TOP_OF_DRAM_SLOT1, 0xff800000, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_NB_LOWER_TOP_OF_DRAM2, 0x00000001, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_NB_UPPER_TOP_OF_DRAM2, 0x00000fff, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_AGP_BASE, 0x00000000, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_AGP_BOT, 0x00000002, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_AGP_TOP, 0x00000000, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_L2_PROTECTION_FAULT_CNTL, 0x00001ffc, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_L2_PROTECTION_FAULT_CNTL2, 0x00002825, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_MX_L1_TLB_CNTL, 0x00000501, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_L2_CNTL, 0x00080603, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_L2_CNTL2, 0x00000003, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_L2_CNTL3, 0x00100003, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_L2_CNTL5, 0x00003fe0, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_CONTEXT0_CNTL, 0x00000001, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_L2_CONTEXT0_PER_PFVF_PTE_CACHE_FRAGMENT_SIZES, 0x00000c00, 0xe0000000),
+        IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_CONTEXT1_CNTL, 0x00000001, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_L2_CONTEXT1_PER_PFVF_PTE_CACHE_FRAGMENT_SIZES, 0x00000c00, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regSDMA0_UCODE_SELFLOAD_CONTROL, 0x00000210, 0),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regSDMA1_UCODE_SELFLOAD_CONTROL, 0x00000210, 0),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regCPC_PSP_DEBUG, CPC_PSP_DEBUG__GPA_OVERRIDE_MASK, 0),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regCPG_PSP_DEBUG, CPG_PSP_DEBUG__GPA_OVERRIDE_MASK, 0)
+};
+
+static void program_imu_rlc_ram(struct amdgpu_device *adev,
+				const struct imu_rlc_ram_golden *regs,
+				const u32 array_size)
+{
+	const struct imu_rlc_ram_golden *entry;
+	u32 reg, data;
+	int i;
+
+	for (i = 0; i < array_size; ++i) {
+		entry = &regs[i];
+		reg =  adev->reg_offset[entry->hwip][entry->instance][entry->segment] + entry->reg;
+		reg |= entry->addr_mask;
+
+		data = entry->data;
+		if (entry->reg == regGCMC_VM_AGP_BASE)
+			data = 0x00ffffff;
+		else if (entry->reg == regGCMC_VM_AGP_TOP)
+			data = 0x0;
+		else if (entry->reg == regGCMC_VM_FB_LOCATION_BASE)
+			data = adev->gmc.vram_start >> 24;
+		else if (entry->reg == regGCMC_VM_FB_LOCATION_TOP)
+			data = adev->gmc.vram_end >> 24;
+
+		WREG32_SOC15(GC, 0, regGFX_IMU_RLC_RAM_ADDR_HIGH, 0);
+		WREG32_SOC15(GC, 0, regGFX_IMU_RLC_RAM_ADDR_LOW, reg);
+		WREG32_SOC15(GC, 0, regGFX_IMU_RLC_RAM_DATA, data);
+	}
+	//Indicate the latest entry
+	WREG32_SOC15(GC, 0, regGFX_IMU_RLC_RAM_ADDR_HIGH, 0);
+	WREG32_SOC15(GC, 0, regGFX_IMU_RLC_RAM_ADDR_LOW, 0);
+	WREG32_SOC15(GC, 0, regGFX_IMU_RLC_RAM_DATA, 0);
+}
+
+static void imu_v11_0_program_rlc_ram(struct amdgpu_device *adev)
+{
+	u32 reg_data;
+
+	WREG32_SOC15(GC, 0, regGFX_IMU_RLC_RAM_INDEX, 0x2);
+
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	case IP_VERSION(11, 0, 0):
+		program_imu_rlc_ram(adev, imu_rlc_ram_golden_11,
+				(const u32)ARRAY_SIZE(imu_rlc_ram_golden_11));
+		break;
+	case IP_VERSION(11, 0, 2):
+		program_imu_rlc_ram(adev, imu_rlc_ram_golden_11_0_2,
+				(const u32)ARRAY_SIZE(imu_rlc_ram_golden_11_0_2));
+		break;
+	case IP_VERSION(11, 0, 3):
+		imu_v11_0_3_program_rlc_ram(adev);
+		break;
+	default:
+		BUG();
+		break;
+	}
+
+	//Indicate the contents of the RAM are valid
+	reg_data = RREG32_SOC15(GC, 0, regGFX_IMU_RLC_RAM_INDEX);
+	reg_data |= GFX_IMU_RLC_RAM_INDEX__RAM_VALID_MASK;
+	WREG32_SOC15(GC, 0, regGFX_IMU_RLC_RAM_INDEX, reg_data);
+}
+
+const struct amdgpu_imu_funcs gfx_v11_0_imu_funcs = {
+	.init_microcode = imu_v11_0_init_microcode,
+	.load_microcode = imu_v11_0_load_microcode,
+	.setup_imu = imu_v11_0_setup,
+	.start_imu = imu_v11_0_start,
+	.program_rlc_ram = imu_v11_0_program_rlc_ram,
+	.wait_for_reset_status = imu_v11_0_wait_for_reset_status,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/imu_v11_0.h b/drivers/gpu/drm/amd/amdgpu/imu_v11_0.h
new file mode 100644
index 000000000000..e71f96fc2f06
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/imu_v11_0.h
@@ -0,0 +1,30 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __IMU_V11_0_H__
+#define __IMU_V11_0_H__
+
+extern const struct amdgpu_imu_funcs gfx_v11_0_imu_funcs;
+
+#endif
+
diff --git a/drivers/gpu/drm/amd/amdgpu/imu_v11_0_3.c b/drivers/gpu/drm/amd/amdgpu/imu_v11_0_3.c
new file mode 100644
index 000000000000..fc69c1a29e23
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/imu_v11_0_3.c
@@ -0,0 +1,145 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include "amdgpu.h"
+#include "amdgpu_imu.h"
+#include "imu_v11_0_3.h"
+
+#include "gc/gc_11_0_3_offset.h"
+#include "gc/gc_11_0_3_sh_mask.h"
+
+static const struct imu_rlc_ram_golden imu_rlc_ram_golden_11_0_3[] = {
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGUS_IO_RD_COMBINE_FLUSH, 0x00055555, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGUS_IO_WR_COMBINE_FLUSH, 0x00055555, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGUS_DRAM_COMBINE_FLUSH, 0x00555555, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGUS_MISC2, 0x00001ffe, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGUS_SDP_CREDITS, 0x003f3fff, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGUS_SDP_TAG_RESERVE1, 0x00000000, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGUS_SDP_VCC_RESERVE0, 0x00041000, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGUS_SDP_VCC_RESERVE1, 0x00000000, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGUS_SDP_VCD_RESERVE0, 0x00040000, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGUS_SDP_VCD_RESERVE1, 0x00000000, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGUS_MISC, 0x00000017, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGUS_SDP_ENABLE, 0x00000001, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCEA_SDP_CREDITS, 0x003f3fbf, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCEA_SDP_TAG_RESERVE0, 0x10200800, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCEA_SDP_TAG_RESERVE1, 0x00000088, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCEA_SDP_VCC_RESERVE0, 0x1d041040, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCEA_SDP_VCC_RESERVE1, 0x80000000, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCEA_SDP_IO_PRIORITY, 0x88888888, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCEA_MAM_CTRL, 0x0000d800, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCEA_SDP_ARB_FINAL, 0x000007ff, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCEA_DRAM_PAGE_BURST, 0x20080200, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCEA_SDP_ENABLE, 0x00000001, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_L2_PROTECTION_FAULT_CNTL2, 0x00020000, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_APT_CNTL, 0x0000000c, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_CACHEABLE_DRAM_ADDRESS_END, 0x000fffff, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCEA_MISC, 0x0c48bff0, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regCC_GC_SA_UNIT_DISABLE, 0x00fffc01, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regCC_GC_PRIM_CONFIG, 0x000fffe1, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regCC_RB_BACKEND_DISABLE, 0xffffff01, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regCC_GC_SHADER_ARRAY_CONFIG, 0xfffe0001, 0x40000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regCC_GC_SHADER_ARRAY_CONFIG, 0xfffe0001, 0x42000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regCC_GC_SHADER_ARRAY_CONFIG, 0xffff0001, 0x44000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regCC_GC_SHADER_ARRAY_CONFIG, 0xffff0001, 0x46000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regCC_GC_SHADER_ARRAY_CONFIG, 0xffff0001, 0x48000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regCC_GC_SHADER_ARRAY_CONFIG, 0xffff0001, 0x4A000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regCGTS_TCC_DISABLE, 0x00000001, 0x00000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regCC_GC_SHADER_RATE_CONFIG, 0x00000001, 0x00000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regCC_GC_EDC_CONFIG, 0x00000001, 0x00000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_MX_L1_TLB_CNTL, 0x00000500, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_SYSTEM_APERTURE_LOW_ADDR, 0x00000001, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_SYSTEM_APERTURE_HIGH_ADDR, 0x00000000, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_LOCAL_FB_ADDRESS_START, 0x00000000, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_LOCAL_FB_ADDRESS_END, 0x000005ff, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_FB_LOCATION_BASE, 0x00006000, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_FB_LOCATION_TOP, 0x000065ff, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_CONTEXT0_CNTL, 0x00000000, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_CONTEXT1_CNTL, 0x00000000, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_NB_TOP_OF_DRAM_SLOT1, 0xff800000, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_NB_LOWER_TOP_OF_DRAM2, 0x00000001, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_NB_UPPER_TOP_OF_DRAM2, 0x00000fff, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_L2_PROTECTION_FAULT_CNTL, 0x00001ffc, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_MX_L1_TLB_CNTL, 0x00000551, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_L2_CNTL, 0x00080603, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_L2_CNTL2, 0x00000003, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_L2_CNTL3, 0x00100003, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_L2_CNTL5, 0x00003fe0, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_CONTEXT0_CNTL, 0x00000001, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_L2_CONTEXT0_PER_PFVF_PTE_CACHE_FRAGMENT_SIZES, 0x00000c00, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_CONTEXT1_CNTL, 0x00000001, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_L2_CONTEXT1_PER_PFVF_PTE_CACHE_FRAGMENT_SIZES, 0x00000c00, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGB_ADDR_CONFIG, 0x00000444, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGL2_PIPE_STEER_0, 0x54105410, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGL2_PIPE_STEER_2, 0x76323276, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGB_ADDR_CONFIG, 0x00000244, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCUTCL2_HARVEST_BYPASS_GROUPS, 0x00000006, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_APT_CNTL, 0x0000000c, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_AGP_BASE, 0x00000000, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_AGP_BOT, 0x00000002, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_AGP_TOP, 0x00000000, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_L2_PROTECTION_FAULT_CNTL2, 0x00020000, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regSDMA0_UCODE_SELFLOAD_CONTROL, 0x00000210, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regSDMA1_UCODE_SELFLOAD_CONTROL, 0x00000210, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regCPC_PSP_DEBUG, CPC_PSP_DEBUG__GPA_OVERRIDE_MASK, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regCPG_PSP_DEBUG, CPG_PSP_DEBUG__GPA_OVERRIDE_MASK, 0xe0000000),
+};
+
+static void program_rlc_ram_register_setting(struct amdgpu_device *adev,
+					     const struct imu_rlc_ram_golden *regs,
+					     const u32 array_size)
+{
+	const struct imu_rlc_ram_golden *entry;
+	u32 reg, data;
+	int i;
+
+	for (i = 0; i < array_size; ++i) {
+		entry = &regs[i];
+		reg =  adev->reg_offset[entry->hwip][entry->instance][entry->segment] + entry->reg;
+		reg |= entry->addr_mask;
+
+		data = entry->data;
+		if (entry->reg == regGCMC_VM_AGP_BASE)
+			data = 0x00ffffff;
+		else if (entry->reg == regGCMC_VM_AGP_TOP)
+			data = 0x0;
+		else if (entry->reg == regGCMC_VM_FB_LOCATION_BASE)
+			data = adev->gmc.vram_start >> 24;
+		else if (entry->reg == regGCMC_VM_FB_LOCATION_TOP)
+			data = adev->gmc.vram_end >> 24;
+
+		WREG32_SOC15(GC, 0, regGFX_IMU_RLC_RAM_ADDR_HIGH, 0);
+		WREG32_SOC15(GC, 0, regGFX_IMU_RLC_RAM_ADDR_LOW, reg);
+		WREG32_SOC15(GC, 0, regGFX_IMU_RLC_RAM_DATA, data);
+	}
+	//Indicate the latest entry
+	WREG32_SOC15(GC, 0, regGFX_IMU_RLC_RAM_ADDR_HIGH, 0);
+	WREG32_SOC15(GC, 0, regGFX_IMU_RLC_RAM_ADDR_LOW, 0);
+	WREG32_SOC15(GC, 0, regGFX_IMU_RLC_RAM_DATA, 0);
+}
+
+void imu_v11_0_3_program_rlc_ram(struct amdgpu_device *adev)
+{
+	program_rlc_ram_register_setting(adev,
+					 imu_rlc_ram_golden_11_0_3,
+					 (const u32)ARRAY_SIZE(imu_rlc_ram_golden_11_0_3));
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/imu_v11_0_3.h b/drivers/gpu/drm/amd/amdgpu/imu_v11_0_3.h
new file mode 100644
index 000000000000..702be568f26b
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/imu_v11_0_3.h
@@ -0,0 +1,29 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __IMU_V11_0_3_H__
+#define __IMU_V11_0_3_H__
+
+void imu_v11_0_3_program_rlc_ram(struct amdgpu_device *adev);
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/imu_v12_0.c b/drivers/gpu/drm/amd/amdgpu/imu_v12_0.c
new file mode 100644
index 000000000000..58cd87db8061
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/imu_v12_0.c
@@ -0,0 +1,406 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/firmware.h>
+#include "amdgpu.h"
+#include "amdgpu_imu.h"
+#include "amdgpu_dpm.h"
+
+#include "imu_v12_0.h"
+
+#include "gc/gc_12_0_0_offset.h"
+#include "gc/gc_12_0_0_sh_mask.h"
+#include "mmhub/mmhub_4_1_0_offset.h"
+
+MODULE_FIRMWARE("amdgpu/gc_12_0_0_imu.bin");
+MODULE_FIRMWARE("amdgpu/gc_12_0_1_imu.bin");
+MODULE_FIRMWARE("amdgpu/gc_12_0_1_imu_kicker.bin");
+
+#define TRANSFER_RAM_MASK	0x001c0000
+
+static int imu_v12_0_init_microcode(struct amdgpu_device *adev)
+{
+	char ucode_prefix[30];
+	int err;
+	const struct imu_firmware_header_v1_0 *imu_hdr;
+	struct amdgpu_firmware_info *info = NULL;
+
+	DRM_DEBUG("\n");
+
+	amdgpu_ucode_ip_version_decode(adev, GC_HWIP, ucode_prefix, sizeof(ucode_prefix));
+	if (amdgpu_is_kicker_fw(adev))
+		err = amdgpu_ucode_request(adev, &adev->gfx.imu_fw, AMDGPU_UCODE_REQUIRED,
+					   "amdgpu/%s_imu_kicker.bin", ucode_prefix);
+	else
+		err = amdgpu_ucode_request(adev, &adev->gfx.imu_fw, AMDGPU_UCODE_REQUIRED,
+					   "amdgpu/%s_imu.bin", ucode_prefix);
+	if (err)
+		goto out;
+
+	imu_hdr = (const struct imu_firmware_header_v1_0 *)adev->gfx.imu_fw->data;
+	adev->gfx.imu_fw_version = le32_to_cpu(imu_hdr->header.ucode_version);
+
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
+		info = &adev->firmware.ucode[AMDGPU_UCODE_ID_IMU_I];
+		info->ucode_id = AMDGPU_UCODE_ID_IMU_I;
+		info->fw = adev->gfx.imu_fw;
+		adev->firmware.fw_size +=
+			ALIGN(le32_to_cpu(imu_hdr->imu_iram_ucode_size_bytes), PAGE_SIZE);
+		info = &adev->firmware.ucode[AMDGPU_UCODE_ID_IMU_D];
+		info->ucode_id = AMDGPU_UCODE_ID_IMU_D;
+		info->fw = adev->gfx.imu_fw;
+		adev->firmware.fw_size +=
+			ALIGN(le32_to_cpu(imu_hdr->imu_dram_ucode_size_bytes), PAGE_SIZE);
+	}
+
+out:
+	if (err) {
+		dev_err(adev->dev,
+			"gfx12: Failed to load firmware \"%s_imu.bin\"\n",
+			ucode_prefix);
+		amdgpu_ucode_release(&adev->gfx.imu_fw);
+	}
+
+	return err;
+}
+
+static int imu_v12_0_load_microcode(struct amdgpu_device *adev)
+{
+	const struct imu_firmware_header_v1_0 *hdr;
+	const __le32 *fw_data;
+	unsigned i, fw_size;
+
+	if (!adev->gfx.imu_fw)
+		return -EINVAL;
+
+	hdr = (const struct imu_firmware_header_v1_0 *)adev->gfx.imu_fw->data;
+
+	fw_data = (const __le32 *)(adev->gfx.imu_fw->data +
+			le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+	fw_size = le32_to_cpu(hdr->imu_iram_ucode_size_bytes) / 4;
+
+	WREG32_SOC15(GC, 0, regGFX_IMU_I_RAM_ADDR, 0);
+
+	for (i = 0; i < fw_size; i++)
+		WREG32_SOC15(GC, 0, regGFX_IMU_I_RAM_DATA, le32_to_cpup(fw_data++));
+
+	WREG32_SOC15(GC, 0, regGFX_IMU_I_RAM_ADDR, adev->gfx.imu_fw_version);
+
+	fw_data = (const __le32 *)(adev->gfx.imu_fw->data +
+			le32_to_cpu(hdr->header.ucode_array_offset_bytes) +
+			le32_to_cpu(hdr->imu_iram_ucode_size_bytes));
+	fw_size = le32_to_cpu(hdr->imu_dram_ucode_size_bytes) / 4;
+
+	WREG32_SOC15(GC, 0, regGFX_IMU_D_RAM_ADDR, 0);
+
+	for (i = 0; i < fw_size; i++)
+		WREG32_SOC15(GC, 0, regGFX_IMU_D_RAM_DATA, le32_to_cpup(fw_data++));
+
+	WREG32_SOC15(GC, 0, regGFX_IMU_D_RAM_ADDR, adev->gfx.imu_fw_version);
+
+	return 0;
+}
+
+static int imu_v12_0_wait_for_reset_status(struct amdgpu_device *adev)
+{
+	u32 imu_reg_val = 0;
+	int i;
+
+	for (i = 0; i < adev->usec_timeout; i++) {
+		imu_reg_val = RREG32_SOC15(GC, 0, regGFX_IMU_GFX_RESET_CTRL);
+		if ((imu_reg_val & 0x1f) == 0x1f)
+			break;
+		udelay(1);
+	}
+
+	if (i >= adev->usec_timeout) {
+		dev_err(adev->dev, "init imu: IMU start timeout\n");
+		return -ETIMEDOUT;
+	}
+
+	return 0;
+}
+
+static void imu_v12_0_setup(struct amdgpu_device *adev)
+{
+	u32 imu_reg_val;
+
+	WREG32_SOC15(GC, 0, regGFX_IMU_C2PMSG_ACCESS_CTRL0, 0xffffff);
+	WREG32_SOC15(GC, 0, regGFX_IMU_C2PMSG_ACCESS_CTRL1, 0xffff);
+
+	if (adev->gfx.imu.mode == DEBUG_MODE) {
+		imu_reg_val = RREG32_SOC15(GC, 0, regGFX_IMU_C2PMSG_16);
+		imu_reg_val |= 0x1;
+		WREG32_SOC15(GC, 0, regGFX_IMU_C2PMSG_16, imu_reg_val);
+
+		imu_reg_val = RREG32_SOC15(GC, 0, regGFX_IMU_SCRATCH_10);
+		imu_reg_val |= 0x20010007;
+		WREG32_SOC15(GC, 0, regGFX_IMU_SCRATCH_10, imu_reg_val);
+
+	}
+}
+
+static int imu_v12_0_start(struct amdgpu_device *adev)
+{
+	u32 imu_reg_val;
+
+	imu_reg_val = RREG32_SOC15(GC, 0, regGFX_IMU_CORE_CTRL);
+	imu_reg_val &= 0xfffffffe;
+	WREG32_SOC15(GC, 0, regGFX_IMU_CORE_CTRL, imu_reg_val);
+
+	if (adev->flags & AMD_IS_APU)
+		amdgpu_dpm_set_gfx_power_up_by_imu(adev);
+
+	return imu_v12_0_wait_for_reset_status(adev);
+}
+
+static const struct imu_rlc_ram_golden imu_rlc_ram_golden_12_0_1[] = {
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regCH_PIPE_STEER, 0x1e4, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGL1X_PIPE_STEER, 0x1e4, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGL1_PIPE_STEER, 0x1e4, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGL2_PIPE_STEER_0, 0x13571357, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGL2_PIPE_STEER_1, 0x64206420, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGL2_PIPE_STEER_2, 0x2460246, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGL2_PIPE_STEER_3, 0x75317531, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGL2C_CTRL3, 0xc0d41183, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regSDMA0_CHICKEN_BITS, 0x507d1c0, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regSDMA1_CHICKEN_BITS, 0x507d1c0, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regCP_RB_WPTR_POLL_CNTL, 0x600100, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGC_EA_CPWD_SDP_CREDITS, 0x3f7fff, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGC_EA_SE_SDP_CREDITS, 0x3f7ebf, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGC_EA_CPWD_SDP_TAG_RESERVE0, 0x2e00000, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGC_EA_CPWD_SDP_TAG_RESERVE1, 0x1a078, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGC_EA_CPWD_SDP_TAG_RESERVE2, 0x0, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGC_EA_SE_SDP_TAG_RESERVE0, 0x0, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGC_EA_SE_SDP_TAG_RESERVE1, 0x12030, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGC_EA_SE_SDP_TAG_RESERVE2, 0x0, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGC_EA_CPWD_SDP_VCC_RESERVE0, 0x19041000, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGC_EA_CPWD_SDP_VCC_RESERVE1, 0x80000000, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGC_EA_SE_SDP_VCC_RESERVE0, 0x1e080000, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGC_EA_SE_SDP_VCC_RESERVE1, 0x80000000, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGC_EA_CPWD_SDP_PRIORITY, 0x880, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGC_EA_SE_SDP_PRIORITY, 0x8880, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGC_EA_CPWD_SDP_ARB_FINAL, 0x17, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGC_EA_SE_SDP_ARB_FINAL, 0x77, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGC_EA_CPWD_SDP_ENABLE, 0x00000001, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGC_EA_SE_SDP_ENABLE, 0x00000001, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_L2_PROTECTION_FAULT_CNTL2, 0x20000, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_APT_CNTL, 0x0c, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_CACHEABLE_DRAM_ADDRESS_END, 0xfffff, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGC_EA_CPWD_MISC, 0x0091, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGC_EA_SE_MISC, 0x0091, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGRBM_GFX_INDEX, 0xe0000000, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCR_GENERAL_CNTL, 0x00008500, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regPA_CL_ENHANCE, 0x00880007, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regTD_CNTL, 0x00000001, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGRBM_GFX_INDEX, 0x00000000, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regRMI_GENERAL_CNTL, 0x01e00000, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGRBM_GFX_INDEX, 0x00000001, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regRMI_GENERAL_CNTL, 0x01e00000, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGRBM_GFX_INDEX, 0x00000100, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regRMI_GENERAL_CNTL, 0x01e00000, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGRBM_GFX_INDEX, 0x00000101, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regRMI_GENERAL_CNTL, 0x01e00000, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGRBM_GFX_INDEX, 0xe0000000, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGB_ADDR_CONFIG, 0x08200545, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGRBMH_CP_PERFMON_CNTL, 0x00000000, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regCB_PERFCOUNTER0_SELECT1, 0x000fffff, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regCP_DEBUG_2, 0x00020000, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regCP_CPC_DEBUG, 0x00500010, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_MX_L1_TLB_CNTL, 0x00000500, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_SYSTEM_APERTURE_LOW_ADDR, 0x00000001, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_SYSTEM_APERTURE_HIGH_ADDR, 0x00000000, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_LOCAL_FB_ADDRESS_START, 0x00000000, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_LOCAL_FB_ADDRESS_END, 0x0000000f, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_FB_LOCATION_BASE, 0x00006000, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_FB_LOCATION_TOP, 0x0000600f, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_CONTEXT0_CNTL, 0x00000000, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_CONTEXT1_CNTL, 0x00000000, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_NB_TOP_OF_DRAM_SLOT1, 0xff800000, 0xe0000000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_NB_LOWER_TOP_OF_DRAM2, 0x00000001, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_NB_UPPER_TOP_OF_DRAM2, 0x0000ffff, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_AGP_BASE, 0x00000000, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_AGP_BOT, 0x00000002, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_AGP_TOP, 0x00000000, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_L2_PROTECTION_FAULT_CNTL, 0x00001ffc, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_MX_L1_TLB_CNTL, 0x00000551, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_L2_CNTL, 0x00080603, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_L2_CNTL2, 0x00000003, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_L2_CNTL3, 0x00100003, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCVM_L2_CNTL5, 0x00003fe0, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_SYSTEM_APERTURE_LOW_ADDR, 0x0003d000, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_SYSTEM_APERTURE_HIGH_ADDR, 0x0003d7ff, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB, 0, 0x1c0000),
+	IMU_RLC_RAM_GOLDEN_VALUE(GC, 0, regGCMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_MSB, 0, 0x1c0000)
+};
+
+static void program_imu_rlc_ram_old(struct amdgpu_device *adev,
+				    const struct imu_rlc_ram_golden *regs,
+				    const u32 array_size)
+{
+	const struct imu_rlc_ram_golden *entry;
+	u32 reg, data;
+	int i;
+
+	for (i = 0; i < array_size; ++i) {
+		entry = &regs[i];
+		reg =  adev->reg_offset[entry->hwip][entry->instance][entry->segment] + entry->reg;
+		reg |= entry->addr_mask;
+		data = entry->data;
+		if (entry->reg == regGCMC_VM_AGP_BASE)
+			data = 0x00ffffff;
+		else if (entry->reg == regGCMC_VM_AGP_TOP)
+			data = 0x0;
+		else if (entry->reg == regGCMC_VM_FB_LOCATION_BASE)
+			data = adev->gmc.vram_start >> 24;
+		else if (entry->reg == regGCMC_VM_FB_LOCATION_TOP)
+			data = adev->gmc.vram_end >> 24;
+
+		WREG32_SOC15(GC, 0, regGFX_IMU_RLC_RAM_ADDR_HIGH, 0);
+		WREG32_SOC15(GC, 0, regGFX_IMU_RLC_RAM_ADDR_LOW, reg);
+		WREG32_SOC15(GC, 0, regGFX_IMU_RLC_RAM_DATA, data);
+	}
+}
+
+static u32 imu_v12_0_grbm_gfx_index_remap(struct amdgpu_device *adev,
+					  u32 data, bool high)
+{
+	u32 val, inst_index;
+
+	inst_index = REG_GET_FIELD(data, GRBM_GFX_INDEX, INSTANCE_INDEX);
+
+	if (high)
+		val = inst_index >> 5;
+	else
+		val = REG_GET_FIELD(data, GRBM_GFX_INDEX, SE_BROADCAST_WRITES) << 18 |
+		      REG_GET_FIELD(data, GRBM_GFX_INDEX, SA_BROADCAST_WRITES) << 19 |
+		      REG_GET_FIELD(data, GRBM_GFX_INDEX, INSTANCE_BROADCAST_WRITES) << 20 |
+		      REG_GET_FIELD(data, GRBM_GFX_INDEX, SE_INDEX) << 21 |
+		      REG_GET_FIELD(data, GRBM_GFX_INDEX, SA_INDEX) << 25 |
+		      (inst_index & 0x1f);
+
+	return val;
+}
+
+static u32 imu_v12_init_gfxhub_settings(struct amdgpu_device *adev,
+					u32 reg, u32 data)
+{
+	if (reg == SOC15_REG_OFFSET(GC, 0, regGCMC_VM_FB_LOCATION_BASE))
+		return RREG32_SOC15(MMHUB, 0, regMMMC_VM_FB_LOCATION_BASE);
+	else if (reg == SOC15_REG_OFFSET(GC, 0, regGCMC_VM_FB_LOCATION_TOP))
+		return RREG32_SOC15(MMHUB, 0, regMMMC_VM_FB_LOCATION_TOP);
+	else if (reg == SOC15_REG_OFFSET(GC, 0, regGCMC_VM_FB_OFFSET))
+		return RREG32_SOC15(MMHUB, 0, regMMMC_VM_FB_OFFSET);
+	else if (reg == SOC15_REG_OFFSET(GC, 0, regGCMC_VM_AGP_BASE))
+		return RREG32_SOC15(MMHUB, 0, regMMMC_VM_AGP_BASE);
+	else if (reg == SOC15_REG_OFFSET(GC, 0, regGCMC_VM_AGP_BOT))
+		return RREG32_SOC15(MMHUB, 0, regMMMC_VM_AGP_BOT);
+	else if (reg == SOC15_REG_OFFSET(GC, 0, regGCMC_VM_AGP_TOP))
+		return RREG32_SOC15(MMHUB, 0, regMMMC_VM_AGP_TOP);
+	else if (reg == SOC15_REG_OFFSET(GC, 0, regGCMC_VM_MX_L1_TLB_CNTL))
+		return RREG32_SOC15(MMHUB, 0, regMMMC_VM_MX_L1_TLB_CNTL);
+	else if (reg == SOC15_REG_OFFSET(GC, 0, regGCMC_VM_SYSTEM_APERTURE_LOW_ADDR))
+		return RREG32_SOC15(MMHUB, 0, regMMMC_VM_SYSTEM_APERTURE_LOW_ADDR);
+	else if (reg == SOC15_REG_OFFSET(GC, 0, regGCMC_VM_SYSTEM_APERTURE_HIGH_ADDR))
+		return RREG32_SOC15(MMHUB, 0, regMMMC_VM_SYSTEM_APERTURE_HIGH_ADDR);
+	else if (reg == SOC15_REG_OFFSET(GC, 0, regGCMC_VM_LOCAL_FB_ADDRESS_START))
+		return RREG32_SOC15(MMHUB, 0, regMMMC_VM_LOCAL_FB_ADDRESS_START);
+	else if (reg == SOC15_REG_OFFSET(GC, 0, regGCMC_VM_LOCAL_FB_ADDRESS_END))
+		return RREG32_SOC15(MMHUB, 0, regMMMC_VM_LOCAL_FB_ADDRESS_END);
+	else if (reg == SOC15_REG_OFFSET(GC, 0, regGCMC_VM_LOCAL_SYSMEM_ADDRESS_START))
+		return RREG32_SOC15(MMHUB, 0, regMMMC_VM_LOCAL_SYSMEM_ADDRESS_START);
+	else if (reg == SOC15_REG_OFFSET(GC, 0, regGCMC_VM_LOCAL_SYSMEM_ADDRESS_END))
+		return RREG32_SOC15(MMHUB, 0, regMMMC_VM_LOCAL_SYSMEM_ADDRESS_END);
+	else if (reg == SOC15_REG_OFFSET(GC, 0, regGCMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB))
+		return RREG32_SOC15(MMHUB, 0, regMMMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB);
+	else if (reg == SOC15_REG_OFFSET(GC, 0, regGCMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_MSB))
+		return RREG32_SOC15(MMHUB, 0, regMMMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_MSB);
+	else
+		return data;
+}
+
+static void program_imu_rlc_ram(struct amdgpu_device *adev,
+				const u32 *regs,
+				const u32 array_size)
+{
+	u32 reg, data, val_h = 0, val_l = TRANSFER_RAM_MASK;
+	int i;
+
+	if (array_size % 3)
+		return;
+
+	for (i = 0; i < array_size; i += 3) {
+		reg = regs[i + 0];
+		data = regs[i + 2];
+		data = imu_v12_init_gfxhub_settings(adev, reg, data);
+		if (reg == SOC15_REG_OFFSET(GC, 0, regGRBM_GFX_INDEX)) {
+			val_l = imu_v12_0_grbm_gfx_index_remap(adev, data, false);
+			val_h = imu_v12_0_grbm_gfx_index_remap(adev, data, true);
+		} else {
+			WREG32_SOC15(GC, 0, regGFX_IMU_RLC_RAM_ADDR_HIGH, val_h);
+			WREG32_SOC15(GC, 0, regGFX_IMU_RLC_RAM_ADDR_LOW, reg | val_l);
+			WREG32_SOC15(GC, 0, regGFX_IMU_RLC_RAM_DATA, data);
+		}
+	}
+}
+
+static void imu_v12_0_program_rlc_ram(struct amdgpu_device *adev)
+{
+	u32 reg_data, size = 0;
+	const u32 *data = NULL;
+	int r = -EINVAL;
+
+	WREG32_SOC15(GC, 0, regGFX_IMU_RLC_RAM_INDEX, 0x2);
+
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	case IP_VERSION(12, 0, 0):
+	case IP_VERSION(12, 0, 1):
+		if (!r)
+			program_imu_rlc_ram(adev, data, (const u32)size);
+		else
+			program_imu_rlc_ram_old(adev, imu_rlc_ram_golden_12_0_1,
+				(const u32)ARRAY_SIZE(imu_rlc_ram_golden_12_0_1));
+		break;
+	default:
+		BUG();
+		break;
+	}
+
+	//Indicate the latest entry
+	WREG32_SOC15(GC, 0, regGFX_IMU_RLC_RAM_ADDR_HIGH, 0);
+	WREG32_SOC15(GC, 0, regGFX_IMU_RLC_RAM_ADDR_LOW, 0);
+	WREG32_SOC15(GC, 0, regGFX_IMU_RLC_RAM_DATA, 0);
+
+	reg_data = RREG32_SOC15(GC, 0, regGFX_IMU_RLC_RAM_INDEX);
+	reg_data |= GFX_IMU_RLC_RAM_INDEX__RAM_VALID_MASK;
+	WREG32_SOC15(GC, 0, regGFX_IMU_RLC_RAM_INDEX, reg_data);
+}
+
+const struct amdgpu_imu_funcs gfx_v12_0_imu_funcs = {
+	.init_microcode = imu_v12_0_init_microcode,
+	.load_microcode = imu_v12_0_load_microcode,
+	.setup_imu = imu_v12_0_setup,
+	.start_imu = imu_v12_0_start,
+	.program_rlc_ram = imu_v12_0_program_rlc_ram,
+	.wait_for_reset_status = imu_v12_0_wait_for_reset_status,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/imu_v12_0.h b/drivers/gpu/drm/amd/amdgpu/imu_v12_0.h
new file mode 100644
index 000000000000..a1f50cb1aeab
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/imu_v12_0.h
@@ -0,0 +1,30 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __IMU_V12_0_H__
+#define __IMU_V12_0_H__
+
+extern const struct amdgpu_imu_funcs gfx_v12_0_imu_funcs;
+
+#endif
+
diff --git a/drivers/gpu/drm/amd/amdgpu/isp_v4_1_0.c b/drivers/gpu/drm/amd/amdgpu/isp_v4_1_0.c
new file mode 100644
index 000000000000..0027a639c7e6
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/isp_v4_1_0.c
@@ -0,0 +1,191 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright (C) 2024 Advanced Micro Devices, Inc. All rights reserved.
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
+ * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
+ * USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ */
+
+#include "amdgpu.h"
+#include "isp_v4_1_0.h"
+
+static const unsigned int isp_4_1_0_int_srcid[MAX_ISP410_INT_SRC] = {
+	ISP_4_1__SRCID__ISP_RINGBUFFER_WPT9,
+	ISP_4_1__SRCID__ISP_RINGBUFFER_WPT10,
+	ISP_4_1__SRCID__ISP_RINGBUFFER_WPT11,
+	ISP_4_1__SRCID__ISP_RINGBUFFER_WPT12,
+	ISP_4_1__SRCID__ISP_RINGBUFFER_WPT13,
+	ISP_4_1__SRCID__ISP_RINGBUFFER_WPT14,
+	ISP_4_1__SRCID__ISP_RINGBUFFER_WPT15,
+	ISP_4_1__SRCID__ISP_RINGBUFFER_WPT16
+};
+
+static int isp_v4_1_0_hw_init(struct amdgpu_isp *isp)
+{
+	struct amdgpu_device *adev = isp->adev;
+	int idx, int_idx, num_res, r;
+	u64 isp_base;
+
+	if (adev->rmmio_size == 0 || adev->rmmio_size < 0x5289)
+		return -EINVAL;
+
+	isp_base = adev->rmmio_base;
+
+	isp->isp_cell = kcalloc(3, sizeof(struct mfd_cell), GFP_KERNEL);
+	if (!isp->isp_cell) {
+		r = -ENOMEM;
+		drm_err(&adev->ddev,
+			"%s: isp mfd cell alloc failed\n", __func__);
+		goto failure;
+	}
+
+	num_res = MAX_ISP410_MEM_RES + MAX_ISP410_INT_SRC;
+	isp->isp_res = kcalloc(num_res, sizeof(struct resource),
+			       GFP_KERNEL);
+	if (!isp->isp_res) {
+		r = -ENOMEM;
+		drm_err(&adev->ddev,
+			"%s: isp mfd res alloc failed\n", __func__);
+		goto failure;
+	}
+
+	isp->isp_pdata = kzalloc(sizeof(*isp->isp_pdata), GFP_KERNEL);
+	if (!isp->isp_pdata) {
+		r = -ENOMEM;
+		drm_err(&adev->ddev,
+			"%s: isp platform data alloc failed\n", __func__);
+		goto failure;
+	}
+
+	/* initialize isp platform data */
+	isp->isp_pdata->adev = (void *)adev;
+	isp->isp_pdata->asic_type = adev->asic_type;
+	isp->isp_pdata->base_rmmio_size = adev->rmmio_size;
+
+	isp->isp_res[0].name = "isp_4_1_0_reg";
+	isp->isp_res[0].flags = IORESOURCE_MEM;
+	isp->isp_res[0].start = isp_base;
+	isp->isp_res[0].end = isp_base + ISP_REGS_OFFSET_END;
+
+	isp->isp_res[1].name = "isp_4_1_phy0_reg";
+	isp->isp_res[1].flags = IORESOURCE_MEM;
+	isp->isp_res[1].start = isp_base + ISP410_PHY0_OFFSET;
+	isp->isp_res[1].end = isp_base + ISP410_PHY0_OFFSET + ISP410_PHY0_SIZE;
+
+	for (idx = MAX_ISP410_MEM_RES, int_idx = 0; idx < num_res; idx++, int_idx++) {
+		isp->isp_res[idx].name = "isp_4_1_0_irq";
+		isp->isp_res[idx].flags = IORESOURCE_IRQ;
+		isp->isp_res[idx].start =
+			amdgpu_irq_create_mapping(adev, isp_4_1_0_int_srcid[int_idx]);
+		isp->isp_res[idx].end =
+			isp->isp_res[idx].start;
+	}
+
+	isp->isp_cell[0].name = "amd_isp_capture";
+	isp->isp_cell[0].num_resources = num_res;
+	isp->isp_cell[0].resources = &isp->isp_res[0];
+	isp->isp_cell[0].platform_data = isp->isp_pdata;
+	isp->isp_cell[0].pdata_size = sizeof(struct isp_platform_data);
+
+	/* initialize isp i2c platform data */
+	isp->isp_i2c_res = kcalloc(1, sizeof(struct resource), GFP_KERNEL);
+	if (!isp->isp_i2c_res) {
+		r = -ENOMEM;
+		drm_err(&adev->ddev,
+			"%s: isp mfd res alloc failed\n", __func__);
+		goto failure;
+	}
+
+	isp->isp_i2c_res[0].name = "isp_i2c0_reg";
+	isp->isp_i2c_res[0].flags = IORESOURCE_MEM;
+	isp->isp_i2c_res[0].start = isp_base + ISP410_I2C0_OFFSET;
+	isp->isp_i2c_res[0].end = isp_base + ISP410_I2C0_OFFSET + ISP410_I2C0_SIZE;
+
+	isp->isp_cell[1].name = "amd_isp_i2c_designware";
+	isp->isp_cell[1].num_resources = 1;
+	isp->isp_cell[1].resources = &isp->isp_i2c_res[0];
+	isp->isp_cell[1].platform_data = isp->isp_pdata;
+	isp->isp_cell[1].pdata_size = sizeof(struct isp_platform_data);
+
+	/* initialize isp gpiochip platform data */
+	isp->isp_gpio_res = kcalloc(1, sizeof(struct resource), GFP_KERNEL);
+	if (!isp->isp_gpio_res) {
+		r = -ENOMEM;
+		drm_err(&adev->ddev,
+			"%s: isp gpio res alloc failed\n", __func__);
+		goto failure;
+	}
+
+	isp->isp_gpio_res[0].name = "isp_gpio_reg";
+	isp->isp_gpio_res[0].flags = IORESOURCE_MEM;
+	isp->isp_gpio_res[0].start = isp_base + ISP410_GPIO_SENSOR_OFFSET;
+	isp->isp_gpio_res[0].end = isp_base + ISP410_GPIO_SENSOR_OFFSET +
+				   ISP410_GPIO_SENSOR_SIZE;
+
+	isp->isp_cell[2].name = "amdisp-pinctrl";
+	isp->isp_cell[2].num_resources = 1;
+	isp->isp_cell[2].resources = &isp->isp_gpio_res[0];
+	isp->isp_cell[2].platform_data = isp->isp_pdata;
+	isp->isp_cell[2].pdata_size = sizeof(struct isp_platform_data);
+
+	r = mfd_add_hotplug_devices(isp->parent, isp->isp_cell, 3);
+	if (r) {
+		drm_err(&adev->ddev,
+			"%s: add mfd hotplug device failed\n", __func__);
+		goto failure;
+	}
+
+	return 0;
+
+failure:
+
+	kfree(isp->isp_pdata);
+	kfree(isp->isp_res);
+	kfree(isp->isp_cell);
+	kfree(isp->isp_i2c_res);
+	kfree(isp->isp_gpio_res);
+
+	return r;
+}
+
+static int isp_v4_1_0_hw_fini(struct amdgpu_isp *isp)
+{
+	mfd_remove_devices(isp->parent);
+
+	kfree(isp->isp_res);
+	kfree(isp->isp_cell);
+	kfree(isp->isp_pdata);
+	kfree(isp->isp_i2c_res);
+	kfree(isp->isp_gpio_res);
+
+	return 0;
+}
+
+static const struct isp_funcs isp_v4_1_0_funcs = {
+	.hw_init = isp_v4_1_0_hw_init,
+	.hw_fini = isp_v4_1_0_hw_fini,
+};
+
+void isp_v4_1_0_set_isp_funcs(struct amdgpu_isp *isp)
+{
+	isp->funcs = &isp_v4_1_0_funcs;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/isp_v4_1_0.h b/drivers/gpu/drm/amd/amdgpu/isp_v4_1_0.h
new file mode 100644
index 000000000000..4d239198edd0
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/isp_v4_1_0.h
@@ -0,0 +1,50 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright (C) 2024 Advanced Micro Devices, Inc. All rights reserved.
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
+ * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
+ * USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ */
+
+#ifndef __ISP_V4_1_0_H__
+#define __ISP_V4_1_0_H__
+
+#include "amdgpu_isp.h"
+
+#include "ivsrcid/isp/irqsrcs_isp_4_1.h"
+
+#define MAX_ISP410_MEM_RES 2
+#define MAX_ISP410_SENSOR_RES 1
+#define MAX_ISP410_INT_SRC 8
+
+#define ISP410_PHY0_OFFSET 0x66700
+#define ISP410_PHY0_SIZE   0xD30
+
+#define ISP410_I2C0_OFFSET 0x66400
+#define ISP410_I2C0_SIZE 0x100
+
+#define ISP410_GPIO_SENSOR_OFFSET 0x6613C
+#define ISP410_GPIO_SENSOR_SIZE 0x54
+
+void isp_v4_1_0_set_isp_funcs(struct amdgpu_isp *isp);
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/isp_v4_1_1.c b/drivers/gpu/drm/amd/amdgpu/isp_v4_1_1.c
new file mode 100644
index 000000000000..4258d3e0b706
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/isp_v4_1_1.c
@@ -0,0 +1,377 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright (C) 2024 Advanced Micro Devices, Inc. All rights reserved.
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
+ * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
+ * USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ */
+
+#include <linux/gpio/machine.h>
+#include "amdgpu.h"
+#include "isp_v4_1_1.h"
+
+MODULE_FIRMWARE("amdgpu/isp_4_1_1.bin");
+
+#define ISP_PERFORMANCE_STATE_LOW 0
+#define ISP_PERFORMANCE_STATE_HIGH 1
+
+#define ISP_HIGH_PERFORMANC_XCLK 788
+#define ISP_HIGH_PERFORMANC_ICLK 788
+
+static const unsigned int isp_4_1_1_int_srcid[MAX_ISP411_INT_SRC] = {
+	ISP_4_1__SRCID__ISP_RINGBUFFER_WPT9,
+	ISP_4_1__SRCID__ISP_RINGBUFFER_WPT10,
+	ISP_4_1__SRCID__ISP_RINGBUFFER_WPT11,
+	ISP_4_1__SRCID__ISP_RINGBUFFER_WPT12,
+	ISP_4_1__SRCID__ISP_RINGBUFFER_WPT13,
+	ISP_4_1__SRCID__ISP_RINGBUFFER_WPT14,
+	ISP_4_1__SRCID__ISP_RINGBUFFER_WPT15,
+	ISP_4_1__SRCID__ISP_RINGBUFFER_WPT16
+};
+
+static struct gpiod_lookup_table isp_gpio_table = {
+	.dev_id = "amd_isp_capture",
+	.table = {
+		GPIO_LOOKUP("AMDI0030:00", 85, "enable_isp", GPIO_ACTIVE_HIGH),
+		{ }
+	},
+};
+
+static struct gpiod_lookup_table isp_sensor_gpio_table = {
+	.dev_id = "i2c-ov05c10",
+	.table = {
+		GPIO_LOOKUP("amdisp-pinctrl", 0, "enable", GPIO_ACTIVE_HIGH),
+		{ }
+	},
+};
+
+static int isp_poweroff(struct generic_pm_domain *genpd)
+{
+	struct amdgpu_isp *isp = container_of(genpd, struct amdgpu_isp, ispgpd);
+	struct amdgpu_device *adev = isp->adev;
+
+	return amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_ISP, true, 0);
+}
+
+static int isp_poweron(struct generic_pm_domain *genpd)
+{
+	struct amdgpu_isp *isp = container_of(genpd, struct amdgpu_isp, ispgpd);
+	struct amdgpu_device *adev = isp->adev;
+
+	return amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_ISP, false, 0);
+}
+
+static int isp_set_performance_state(struct generic_pm_domain *genpd,
+				     unsigned int state)
+{
+	struct amdgpu_isp *isp = container_of(genpd, struct amdgpu_isp, ispgpd);
+	struct amdgpu_device *adev = isp->adev;
+	u32 iclk, xclk;
+	int ret;
+
+	switch (state) {
+	case ISP_PERFORMANCE_STATE_HIGH:
+		xclk = ISP_HIGH_PERFORMANC_XCLK;
+		iclk = ISP_HIGH_PERFORMANC_ICLK;
+		break;
+	case ISP_PERFORMANCE_STATE_LOW:
+		/* isp runs at default lowest clock-rate on power-on, do nothing */
+		return 0;
+	default:
+		return -EINVAL;
+	}
+
+	ret = amdgpu_dpm_set_soft_freq_range(adev, PP_ISPXCLK, xclk, 0);
+	if (ret) {
+		drm_err(&adev->ddev, "failed to set xclk %u to %u: %d\n",
+			xclk, state, ret);
+		return ret;
+	}
+
+	ret = amdgpu_dpm_set_soft_freq_range(adev, PP_ISPICLK, iclk, 0);
+	if (ret) {
+		drm_err(&adev->ddev, "failed to set iclk %u to %u: %d\n",
+			iclk, state, ret);
+		return ret;
+	}
+
+	return 0;
+}
+
+static int isp_genpd_add_device(struct device *dev, void *data)
+{
+	struct generic_pm_domain *gpd = data;
+	struct platform_device *pdev = container_of(dev, struct platform_device, dev);
+	struct amdgpu_isp *isp = container_of(gpd, struct amdgpu_isp, ispgpd);
+	struct amdgpu_device *adev = isp->adev;
+	int ret;
+
+	if (!pdev)
+		return -EINVAL;
+
+	if (!dev->type->name) {
+		drm_dbg(&adev->ddev, "Invalid device type to add\n");
+		goto exit;
+	}
+
+	if (strcmp(dev->type->name, "mfd_device")) {
+		drm_dbg(&adev->ddev, "Invalid isp mfd device %s to add\n", pdev->mfd_cell->name);
+		goto exit;
+	}
+
+	ret = pm_genpd_add_device(gpd, dev);
+	if (ret) {
+		drm_err(&adev->ddev, "Failed to add dev %s to genpd %d\n",
+			pdev->mfd_cell->name, ret);
+		return -ENODEV;
+	}
+
+exit:
+	/* Continue to add */
+	return 0;
+}
+
+static int isp_genpd_remove_device(struct device *dev, void *data)
+{
+	struct generic_pm_domain *gpd = data;
+	struct platform_device *pdev = container_of(dev, struct platform_device, dev);
+	struct amdgpu_isp *isp = container_of(gpd, struct amdgpu_isp, ispgpd);
+	struct amdgpu_device *adev = isp->adev;
+	int ret;
+
+	if (!pdev)
+		return -EINVAL;
+
+	if (!dev->type->name) {
+		drm_dbg(&adev->ddev, "Invalid device type to remove\n");
+		goto exit;
+	}
+
+	if (strcmp(dev->type->name, "mfd_device")) {
+		drm_dbg(&adev->ddev, "Invalid isp mfd device %s to remove\n",
+			pdev->mfd_cell->name);
+		goto exit;
+	}
+
+	ret = pm_genpd_remove_device(dev);
+	if (ret) {
+		drm_err(&adev->ddev, "Failed to remove dev from genpd %d\n", ret);
+		return -ENODEV;
+	}
+
+exit:
+	/* Continue to remove */
+	return 0;
+}
+
+static int isp_v4_1_1_hw_init(struct amdgpu_isp *isp)
+{
+	const struct software_node *amd_camera_node, *isp4_node;
+	struct amdgpu_device *adev = isp->adev;
+	struct acpi_device *acpi_dev;
+	int idx, int_idx, num_res, r;
+	u64 isp_base;
+
+	if (adev->rmmio_size == 0 || adev->rmmio_size < 0x5289)
+		return -EINVAL;
+
+	r = amdgpu_acpi_get_isp4_dev(&acpi_dev);
+	if (r) {
+		drm_dbg(&adev->ddev, "Invalid isp platform detected (%d)", r);
+		/* allow GPU init to progress */
+		return 0;
+	}
+
+	/* add GPIO resources required for OMNI5C10 sensor */
+	if (!strcmp("OMNI5C10", acpi_device_hid(acpi_dev))) {
+		gpiod_add_lookup_table(&isp_gpio_table);
+		gpiod_add_lookup_table(&isp_sensor_gpio_table);
+	}
+
+	isp_base = adev->rmmio_base;
+
+	isp->ispgpd.name = "ISP_v_4_1_1";
+	isp->ispgpd.power_off = isp_poweroff;
+	isp->ispgpd.power_on = isp_poweron;
+	isp->ispgpd.set_performance_state = isp_set_performance_state;
+
+	r = pm_genpd_init(&isp->ispgpd, NULL, true);
+	if (r) {
+		drm_err(&adev->ddev, "failed to initialize genpd (%d)\n", r);
+		return -EINVAL;
+	}
+
+	isp->isp_cell = kcalloc(3, sizeof(struct mfd_cell), GFP_KERNEL);
+	if (!isp->isp_cell) {
+		r = -ENOMEM;
+		drm_err(&adev->ddev, "isp mfd cell alloc failed (%d)\n", r);
+		goto failure;
+	}
+
+	num_res = MAX_ISP411_MEM_RES + MAX_ISP411_INT_SRC;
+
+	isp->isp_res = kcalloc(num_res, sizeof(struct resource),
+			       GFP_KERNEL);
+	if (!isp->isp_res) {
+		r = -ENOMEM;
+		drm_err(&adev->ddev, "isp mfd resource alloc failed (%d)\n", r);
+		goto failure;
+	}
+
+	isp->isp_pdata = kzalloc(sizeof(*isp->isp_pdata), GFP_KERNEL);
+	if (!isp->isp_pdata) {
+		r = -ENOMEM;
+		drm_err(&adev->ddev, "isp platform data alloc failed (%d)\n", r);
+		goto failure;
+	}
+
+	amd_camera_node = (const struct software_node *)acpi_dev->driver_data;
+	isp4_node = software_node_find_by_name(amd_camera_node, "isp4");
+
+	/* initialize isp platform data */
+	isp->isp_pdata->adev = (void *)adev;
+	isp->isp_pdata->asic_type = adev->asic_type;
+	isp->isp_pdata->base_rmmio_size = adev->rmmio_size;
+
+	isp->isp_res[0].name = "isp_4_1_1_reg";
+	isp->isp_res[0].flags = IORESOURCE_MEM;
+	isp->isp_res[0].start = isp_base;
+	isp->isp_res[0].end = isp_base + ISP_REGS_OFFSET_END;
+
+	isp->isp_res[1].name = "isp_4_1_1_phy0_reg";
+	isp->isp_res[1].flags = IORESOURCE_MEM;
+	isp->isp_res[1].start = isp_base + ISP411_PHY0_OFFSET;
+	isp->isp_res[1].end = isp_base + ISP411_PHY0_OFFSET + ISP411_PHY0_SIZE;
+
+	for (idx = MAX_ISP411_MEM_RES, int_idx = 0; idx < num_res; idx++, int_idx++) {
+		isp->isp_res[idx].name = "isp_4_1_1_irq";
+		isp->isp_res[idx].flags = IORESOURCE_IRQ;
+		isp->isp_res[idx].start =
+			amdgpu_irq_create_mapping(adev, isp_4_1_1_int_srcid[int_idx]);
+		isp->isp_res[idx].end =
+			isp->isp_res[idx].start;
+	}
+
+	isp->isp_cell[0].name = "amd_isp_capture";
+	isp->isp_cell[0].num_resources = num_res;
+	isp->isp_cell[0].resources = &isp->isp_res[0];
+	isp->isp_cell[0].platform_data = isp->isp_pdata;
+	isp->isp_cell[0].swnode = isp4_node;
+	isp->isp_cell[0].pdata_size = sizeof(struct isp_platform_data);
+
+	/* initialize isp i2c platform data */
+	isp->isp_i2c_res = kcalloc(1, sizeof(struct resource), GFP_KERNEL);
+	if (!isp->isp_i2c_res) {
+		r = -ENOMEM;
+		drm_err(&adev->ddev, "isp mfd res alloc failed (%d)\n", r);
+		goto failure;
+	}
+
+	isp->isp_i2c_res[0].name = "isp_i2c0_reg";
+	isp->isp_i2c_res[0].flags = IORESOURCE_MEM;
+	isp->isp_i2c_res[0].start = isp_base + ISP411_I2C0_OFFSET;
+	isp->isp_i2c_res[0].end = isp_base + ISP411_I2C0_OFFSET + ISP411_I2C0_SIZE;
+
+	isp->isp_cell[1].name = "amd_isp_i2c_designware";
+	isp->isp_cell[1].num_resources = 1;
+	isp->isp_cell[1].resources = &isp->isp_i2c_res[0];
+	isp->isp_cell[1].platform_data = isp->isp_pdata;
+	isp->isp_cell[1].pdata_size = sizeof(struct isp_platform_data);
+
+	/* initialize isp gpiochip platform data */
+	isp->isp_gpio_res = kcalloc(1, sizeof(struct resource), GFP_KERNEL);
+	if (!isp->isp_gpio_res) {
+		r = -ENOMEM;
+		drm_err(&adev->ddev, "isp gpio resource alloc failed (%d)\n", r);
+		goto failure;
+	}
+
+	isp->isp_gpio_res[0].name = "isp_gpio_reg";
+	isp->isp_gpio_res[0].flags = IORESOURCE_MEM;
+	isp->isp_gpio_res[0].start = isp_base + ISP411_GPIO_SENSOR_OFFSET;
+	isp->isp_gpio_res[0].end = isp_base + ISP411_GPIO_SENSOR_OFFSET +
+				   ISP411_GPIO_SENSOR_SIZE;
+
+	isp->isp_cell[2].name = "amdisp-pinctrl";
+	isp->isp_cell[2].num_resources = 1;
+	isp->isp_cell[2].resources = &isp->isp_gpio_res[0];
+	isp->isp_cell[2].platform_data = isp->isp_pdata;
+	isp->isp_cell[2].pdata_size = sizeof(struct isp_platform_data);
+
+	/* add only amd_isp_capture and amd_isp_i2c_designware to genpd */
+	r = mfd_add_hotplug_devices(isp->parent, isp->isp_cell, 2);
+	if (r) {
+		drm_err(&adev->ddev, "add mfd hotplug device failed (%d)\n", r);
+		goto failure;
+	}
+
+	r = device_for_each_child(isp->parent, &isp->ispgpd,
+				  isp_genpd_add_device);
+	if (r) {
+		drm_err(&adev->ddev, "failed to add devices to genpd (%d)\n", r);
+		goto failure;
+	}
+
+	r = mfd_add_hotplug_devices(isp->parent, &isp->isp_cell[2], 1);
+	if (r) {
+		drm_err(&adev->ddev, "add pinctl hotplug device failed (%d)\n", r);
+		goto failure;
+	}
+
+	return 0;
+
+failure:
+
+	kfree(isp->isp_pdata);
+	kfree(isp->isp_res);
+	kfree(isp->isp_cell);
+	kfree(isp->isp_i2c_res);
+	kfree(isp->isp_gpio_res);
+
+	return r;
+}
+
+static int isp_v4_1_1_hw_fini(struct amdgpu_isp *isp)
+{
+	device_for_each_child(isp->parent, NULL,
+			      isp_genpd_remove_device);
+
+	mfd_remove_devices(isp->parent);
+
+	kfree(isp->isp_res);
+	kfree(isp->isp_cell);
+	kfree(isp->isp_pdata);
+	kfree(isp->isp_i2c_res);
+	kfree(isp->isp_gpio_res);
+
+	return 0;
+}
+
+static const struct isp_funcs isp_v4_1_1_funcs = {
+	.hw_init = isp_v4_1_1_hw_init,
+	.hw_fini = isp_v4_1_1_hw_fini,
+};
+
+void isp_v4_1_1_set_isp_funcs(struct amdgpu_isp *isp)
+{
+	isp->funcs = &isp_v4_1_1_funcs;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/isp_v4_1_1.h b/drivers/gpu/drm/amd/amdgpu/isp_v4_1_1.h
new file mode 100644
index 000000000000..fe45d70d87f1
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/isp_v4_1_1.h
@@ -0,0 +1,49 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright (C) 2024 Advanced Micro Devices, Inc. All rights reserved.
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
+ * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
+ * USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ */
+
+#ifndef __ISP_V4_1_1_H__
+#define __ISP_V4_1_1_H__
+
+#include "amdgpu_isp.h"
+
+#include "ivsrcid/isp/irqsrcs_isp_4_1.h"
+
+#define MAX_ISP411_MEM_RES 2
+#define MAX_ISP411_INT_SRC 8
+
+#define ISP411_PHY0_OFFSET 0x66700
+#define ISP411_PHY0_SIZE   0xD30
+
+#define ISP411_I2C0_OFFSET 0x66400
+#define ISP411_I2C0_SIZE 0x100
+
+#define ISP411_GPIO_SENSOR_OFFSET 0x6613C
+#define ISP411_GPIO_SENSOR_SIZE 0x54
+
+void isp_v4_1_1_set_isp_funcs(struct amdgpu_isp *isp);
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v1_0.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v1_0.c
index 9360204da7fb..b5bb7f4d607c 100644
--- a/drivers/gpu/drm/amd/amdgpu/jpeg_v1_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v1_0.c
@@ -23,6 +23,7 @@
 
 #include "amdgpu.h"
 #include "amdgpu_jpeg.h"
+#include "amdgpu_cs.h"
 #include "soc15.h"
 #include "soc15d.h"
 #include "vcn_v1_0.h"
@@ -34,6 +35,9 @@
 static void jpeg_v1_0_set_dec_ring_funcs(struct amdgpu_device *adev);
 static void jpeg_v1_0_set_irq_funcs(struct amdgpu_device *adev);
 static void jpeg_v1_0_ring_begin_use(struct amdgpu_ring *ring);
+static int jpeg_v1_dec_ring_parse_cs(struct amdgpu_cs_parser *parser,
+				     struct amdgpu_job *job,
+				     struct amdgpu_ib *ib);
 
 static void jpeg_v1_0_decode_ring_patch_wreg(struct amdgpu_ring *ring, uint32_t *ptr, uint32_t reg_offset, uint32_t val)
 {
@@ -300,7 +304,10 @@ static void jpeg_v1_0_decode_ring_emit_ib(struct amdgpu_ring *ring,
 
 	amdgpu_ring_write(ring,
 		PACKETJ(SOC15_REG_OFFSET(JPEG, 0, mmUVD_LMI_JRBC_IB_VMID), 0, 0, PACKETJ_TYPE0));
-	amdgpu_ring_write(ring, (vmid | (vmid << 4)));
+	if (ring->funcs->parse_cs)
+		amdgpu_ring_write(ring, 0);
+	else
+		amdgpu_ring_write(ring, (vmid | (vmid << 4)));
 
 	amdgpu_ring_write(ring,
 		PACKETJ(SOC15_REG_OFFSET(JPEG, 0, mmUVD_LMI_JPEG_VMID), 0, 0, PACKETJ_TYPE0));
@@ -376,7 +383,7 @@ static void jpeg_v1_0_decode_ring_emit_reg_wait(struct amdgpu_ring *ring,
 static void jpeg_v1_0_decode_ring_emit_vm_flush(struct amdgpu_ring *ring,
 		unsigned vmid, uint64_t pd_addr)
 {
-	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
+	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->vm_hub];
 	uint32_t data0, data1, mask;
 
 	pd_addr = amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pd_addr);
@@ -437,7 +444,7 @@ static int jpeg_v1_0_process_interrupt(struct amdgpu_device *adev,
 
 	switch (entry->src_id) {
 	case 126:
-		amdgpu_fence_process(&adev->jpeg.inst->ring_dec);
+		amdgpu_fence_process(adev->jpeg.inst->ring_dec);
 		break;
 	default:
 		DRM_ERROR("Unhandled interrupt: %d %d\n",
@@ -451,15 +458,16 @@ static int jpeg_v1_0_process_interrupt(struct amdgpu_device *adev,
 /**
  * jpeg_v1_0_early_init - set function pointers
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Set ring and irq function pointers
  */
-int jpeg_v1_0_early_init(void *handle)
+int jpeg_v1_0_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	adev->jpeg.num_jpeg_inst = 1;
+	adev->jpeg.num_jpeg_rings = 1;
 
 	jpeg_v1_0_set_dec_ring_funcs(adev);
 	jpeg_v1_0_set_irq_funcs(adev);
@@ -470,12 +478,12 @@ int jpeg_v1_0_early_init(void *handle)
 /**
  * jpeg_v1_0_sw_init - sw init for JPEG block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  */
-int jpeg_v1_0_sw_init(void *handle)
+int jpeg_v1_0_sw_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	struct amdgpu_ring *ring;
 	int r;
 
@@ -484,14 +492,15 @@ int jpeg_v1_0_sw_init(void *handle)
 	if (r)
 		return r;
 
-	ring = &adev->jpeg.inst->ring_dec;
+	ring = adev->jpeg.inst->ring_dec;
+	ring->vm_hub = AMDGPU_MMHUB0(0);
 	sprintf(ring->name, "jpeg_dec");
 	r = amdgpu_ring_init(adev, ring, 512, &adev->jpeg.inst->irq,
 			     0, AMDGPU_RING_PRIO_DEFAULT, NULL);
 	if (r)
 		return r;
 
-	adev->jpeg.internal.jpeg_pitch = adev->jpeg.inst->external.jpeg_pitch =
+	adev->jpeg.internal.jpeg_pitch[0] = adev->jpeg.inst->external.jpeg_pitch[0] =
 		SOC15_REG_OFFSET(JPEG, 0, mmUVD_JPEG_PITCH);
 
 	return 0;
@@ -500,15 +509,15 @@ int jpeg_v1_0_sw_init(void *handle)
 /**
  * jpeg_v1_0_sw_fini - sw fini for JPEG block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * JPEG free up sw allocation
  */
-void jpeg_v1_0_sw_fini(void *handle)
+void jpeg_v1_0_sw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
-	amdgpu_ring_fini(&adev->jpeg.inst[0].ring_dec);
+	amdgpu_ring_fini(adev->jpeg.inst->ring_dec);
 }
 
 /**
@@ -521,7 +530,7 @@ void jpeg_v1_0_sw_fini(void *handle)
  */
 void jpeg_v1_0_start(struct amdgpu_device *adev, int mode)
 {
-	struct amdgpu_ring *ring = &adev->jpeg.inst->ring_dec;
+	struct amdgpu_ring *ring = adev->jpeg.inst->ring_dec;
 
 	if (mode == 0) {
 		WREG32_SOC15(JPEG, 0, mmUVD_LMI_JRBC_RB_VMID, 0);
@@ -548,11 +557,11 @@ static const struct amdgpu_ring_funcs jpeg_v1_0_decode_ring_vm_funcs = {
 	.nop = PACKET0(0x81ff, 0),
 	.support_64bit_ptrs = false,
 	.no_user_fence = true,
-	.vmhub = AMDGPU_MMHUB_0,
-	.extra_dw = 64,
+	.extra_bytes = 256,
 	.get_rptr = jpeg_v1_0_decode_ring_get_rptr,
 	.get_wptr = jpeg_v1_0_decode_ring_get_wptr,
 	.set_wptr = jpeg_v1_0_decode_ring_set_wptr,
+	.parse_cs = jpeg_v1_dec_ring_parse_cs,
 	.emit_frame_size =
 		6 + 6 + /* hdp invalidate / flush */
 		SOC15_FLUSH_GPU_TLB_NUM_WREG * 6 +
@@ -579,8 +588,7 @@ static const struct amdgpu_ring_funcs jpeg_v1_0_decode_ring_vm_funcs = {
 
 static void jpeg_v1_0_set_dec_ring_funcs(struct amdgpu_device *adev)
 {
-	adev->jpeg.inst->ring_dec.funcs = &jpeg_v1_0_decode_ring_vm_funcs;
-	DRM_INFO("JPEG decode is enabled in VM mode\n");
+	adev->jpeg.inst->ring_dec->funcs = &jpeg_v1_0_decode_ring_vm_funcs;
 }
 
 static const struct amdgpu_irq_src_funcs jpeg_v1_0_irq_funcs = {
@@ -596,18 +604,84 @@ static void jpeg_v1_0_set_irq_funcs(struct amdgpu_device *adev)
 static void jpeg_v1_0_ring_begin_use(struct amdgpu_ring *ring)
 {
 	struct	amdgpu_device *adev = ring->adev;
-	bool	set_clocks = !cancel_delayed_work_sync(&adev->vcn.idle_work);
+	bool	set_clocks = !cancel_delayed_work_sync(&adev->vcn.inst[0].idle_work);
 	int		cnt = 0;
 
-	mutex_lock(&adev->vcn.vcn1_jpeg1_workaround);
+	mutex_lock(&adev->vcn.inst[0].vcn1_jpeg1_workaround);
 
 	if (amdgpu_fence_wait_empty(&adev->vcn.inst->ring_dec))
 		DRM_ERROR("JPEG dec: vcn dec ring may not be empty\n");
 
-	for (cnt = 0; cnt < adev->vcn.num_enc_rings; cnt++) {
+	for (cnt = 0; cnt < adev->vcn.inst[0].num_enc_rings; cnt++) {
 		if (amdgpu_fence_wait_empty(&adev->vcn.inst->ring_enc[cnt]))
 			DRM_ERROR("JPEG dec: vcn enc ring[%d] may not be empty\n", cnt);
 	}
 
 	vcn_v1_0_set_pg_for_begin_use(ring, set_clocks);
 }
+
+/**
+ * jpeg_v1_dec_ring_parse_cs - command submission parser
+ *
+ * @parser: Command submission parser context
+ * @job: the job to parse
+ * @ib: the IB to parse
+ *
+ * Parse the command stream, return -EINVAL for invalid packet,
+ * 0 otherwise
+ */
+static int jpeg_v1_dec_ring_parse_cs(struct amdgpu_cs_parser *parser,
+				     struct amdgpu_job *job,
+				     struct amdgpu_ib *ib)
+{
+	u32 i, reg, res, cond, type;
+	int ret = 0;
+	struct amdgpu_device *adev = parser->adev;
+
+	for (i = 0; i < ib->length_dw ; i += 2) {
+		reg  = CP_PACKETJ_GET_REG(ib->ptr[i]);
+		res  = CP_PACKETJ_GET_RES(ib->ptr[i]);
+		cond = CP_PACKETJ_GET_COND(ib->ptr[i]);
+		type = CP_PACKETJ_GET_TYPE(ib->ptr[i]);
+
+		if (res || cond != PACKETJ_CONDITION_CHECK0) /* only allow 0 for now */
+			return -EINVAL;
+
+		if (reg >= JPEG_V1_REG_RANGE_START && reg <= JPEG_V1_REG_RANGE_END)
+			continue;
+
+		switch (type) {
+		case PACKETJ_TYPE0:
+			if (reg != JPEG_V1_LMI_JPEG_WRITE_64BIT_BAR_HIGH &&
+			    reg != JPEG_V1_LMI_JPEG_WRITE_64BIT_BAR_LOW &&
+			    reg != JPEG_V1_LMI_JPEG_READ_64BIT_BAR_HIGH &&
+			    reg != JPEG_V1_LMI_JPEG_READ_64BIT_BAR_LOW &&
+			    reg != JPEG_V1_REG_CTX_INDEX &&
+			    reg != JPEG_V1_REG_CTX_DATA) {
+				ret = -EINVAL;
+			}
+			break;
+		case PACKETJ_TYPE1:
+			if (reg != JPEG_V1_REG_CTX_DATA)
+				ret = -EINVAL;
+			break;
+		case PACKETJ_TYPE3:
+			if (reg != JPEG_V1_REG_SOFT_RESET)
+				ret = -EINVAL;
+			break;
+		case PACKETJ_TYPE6:
+			if (ib->ptr[i] != CP_PACKETJ_NOP)
+				ret = -EINVAL;
+			break;
+		default:
+			ret = -EINVAL;
+		}
+
+		if (ret) {
+			dev_err(adev->dev, "Invalid packet [0x%08x]!\n", ib->ptr[i]);
+			break;
+		}
+	}
+
+	return ret;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v1_0.h b/drivers/gpu/drm/amd/amdgpu/jpeg_v1_0.h
index bbf33a6a3972..097328635083 100644
--- a/drivers/gpu/drm/amd/amdgpu/jpeg_v1_0.h
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v1_0.h
@@ -24,9 +24,20 @@
 #ifndef __JPEG_V1_0_H__
 #define __JPEG_V1_0_H__
 
-int jpeg_v1_0_early_init(void *handle);
-int jpeg_v1_0_sw_init(void *handle);
-void jpeg_v1_0_sw_fini(void *handle);
+int jpeg_v1_0_early_init(struct amdgpu_ip_block *ip_block);
+int jpeg_v1_0_sw_init(struct amdgpu_ip_block *ip_block);
+void jpeg_v1_0_sw_fini(struct amdgpu_ip_block *ip_block);
 void jpeg_v1_0_start(struct amdgpu_device *adev, int mode);
 
+#define JPEG_V1_REG_RANGE_START	0x8000
+#define JPEG_V1_REG_RANGE_END	0x803f
+
+#define JPEG_V1_LMI_JPEG_WRITE_64BIT_BAR_HIGH	0x8238
+#define JPEG_V1_LMI_JPEG_WRITE_64BIT_BAR_LOW	0x8239
+#define JPEG_V1_LMI_JPEG_READ_64BIT_BAR_HIGH	0x825a
+#define JPEG_V1_LMI_JPEG_READ_64BIT_BAR_LOW	0x825b
+#define JPEG_V1_REG_CTX_INDEX			0x8328
+#define JPEG_V1_REG_CTX_DATA			0x8329
+#define JPEG_V1_REG_SOFT_RESET			0x83a0
+
 #endif /*__JPEG_V1_0_H__*/
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_0.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_0.c
index 299de1d131d8..27c76bd424cf 100644
--- a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_0.c
@@ -32,23 +32,40 @@
 #include "vcn/vcn_2_0_0_sh_mask.h"
 #include "ivsrcid/vcn/irqsrcs_vcn_2_0.h"
 
+static const struct amdgpu_hwip_reg_entry jpeg_reg_list_2_0[] = {
+	SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JPEG_POWER_STATUS),
+	SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JPEG_INT_STAT),
+	SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JRBC_RB_RPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JRBC_RB_WPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JRBC_RB_CNTL),
+	SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JRBC_RB_SIZE),
+	SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JRBC_STATUS),
+	SOC15_REG_ENTRY_STR(JPEG, 0, mmJPEG_DEC_ADDR_MODE),
+	SOC15_REG_ENTRY_STR(JPEG, 0, mmJPEG_DEC_GFX10_ADDR_CONFIG),
+	SOC15_REG_ENTRY_STR(JPEG, 0, mmJPEG_DEC_Y_GFX10_TILING_SURFACE),
+	SOC15_REG_ENTRY_STR(JPEG, 0, mmJPEG_DEC_UV_GFX10_TILING_SURFACE),
+	SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JPEG_PITCH),
+	SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JPEG_UV_PITCH),
+};
+
 static void jpeg_v2_0_set_dec_ring_funcs(struct amdgpu_device *adev);
 static void jpeg_v2_0_set_irq_funcs(struct amdgpu_device *adev);
-static int jpeg_v2_0_set_powergating_state(void *handle,
+static int jpeg_v2_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
 				enum amd_powergating_state state);
 
 /**
  * jpeg_v2_0_early_init - set function pointers
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Set ring and irq function pointers
  */
-static int jpeg_v2_0_early_init(void *handle)
+static int jpeg_v2_0_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	adev->jpeg.num_jpeg_inst = 1;
+	adev->jpeg.num_jpeg_rings = 1;
 
 	jpeg_v2_0_set_dec_ring_funcs(adev);
 	jpeg_v2_0_set_irq_funcs(adev);
@@ -59,13 +76,13 @@ static int jpeg_v2_0_early_init(void *handle)
 /**
  * jpeg_v2_0_sw_init - sw init for JPEG block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Load firmware and sw initialization
  */
-static int jpeg_v2_0_sw_init(void *handle)
+static int jpeg_v2_0_sw_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	struct amdgpu_ring *ring;
 	int r;
 
@@ -83,37 +100,50 @@ static int jpeg_v2_0_sw_init(void *handle)
 	if (r)
 		return r;
 
-	ring = &adev->jpeg.inst->ring_dec;
+	ring = adev->jpeg.inst->ring_dec;
 	ring->use_doorbell = true;
 	ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 1;
+	ring->vm_hub = AMDGPU_MMHUB0(0);
 	sprintf(ring->name, "jpeg_dec");
 	r = amdgpu_ring_init(adev, ring, 512, &adev->jpeg.inst->irq,
 			     0, AMDGPU_RING_PRIO_DEFAULT, NULL);
 	if (r)
 		return r;
 
-	adev->jpeg.internal.jpeg_pitch = mmUVD_JPEG_PITCH_INTERNAL_OFFSET;
-	adev->jpeg.inst->external.jpeg_pitch = SOC15_REG_OFFSET(JPEG, 0, mmUVD_JPEG_PITCH);
+	adev->jpeg.internal.jpeg_pitch[0] = mmUVD_JPEG_PITCH_INTERNAL_OFFSET;
+	adev->jpeg.inst->external.jpeg_pitch[0] = SOC15_REG_OFFSET(JPEG, 0, mmUVD_JPEG_PITCH);
 
-	return 0;
+	r = amdgpu_jpeg_reg_dump_init(adev, jpeg_reg_list_2_0, ARRAY_SIZE(jpeg_reg_list_2_0));
+	if (r)
+		return r;
+
+	adev->jpeg.supported_reset =
+		amdgpu_get_soft_full_reset_mask(adev->jpeg.inst[0].ring_dec);
+	if (!amdgpu_sriov_vf(adev))
+		adev->jpeg.supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE;
+	r = amdgpu_jpeg_sysfs_reset_mask_init(adev);
+
+	return r;
 }
 
 /**
  * jpeg_v2_0_sw_fini - sw fini for JPEG block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * JPEG suspend and free up sw allocation
  */
-static int jpeg_v2_0_sw_fini(void *handle)
+static int jpeg_v2_0_sw_fini(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	r = amdgpu_jpeg_suspend(adev);
 	if (r)
 		return r;
 
+	amdgpu_jpeg_sysfs_reset_mask_fini(adev);
+
 	r = amdgpu_jpeg_sw_fini(adev);
 
 	return r;
@@ -122,41 +152,36 @@ static int jpeg_v2_0_sw_fini(void *handle)
 /**
  * jpeg_v2_0_hw_init - start and test JPEG block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  */
-static int jpeg_v2_0_hw_init(void *handle)
+static int jpeg_v2_0_hw_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-	struct amdgpu_ring *ring = &adev->jpeg.inst->ring_dec;
-	int r;
+	struct amdgpu_device *adev = ip_block->adev;
+	struct amdgpu_ring *ring = adev->jpeg.inst->ring_dec;
 
 	adev->nbio.funcs->vcn_doorbell_range(adev, ring->use_doorbell,
 		(adev->doorbell_index.vcn.vcn_ring0_1 << 1), 0);
 
-	r = amdgpu_ring_test_helper(ring);
-	if (!r)
-		DRM_INFO("JPEG decode initialized successfully.\n");
-
-	return r;
+	return amdgpu_ring_test_helper(ring);
 }
 
 /**
  * jpeg_v2_0_hw_fini - stop the hardware block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Stop the JPEG block, mark ring as not ready any more
  */
-static int jpeg_v2_0_hw_fini(void *handle)
+static int jpeg_v2_0_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
-	cancel_delayed_work_sync(&adev->vcn.idle_work);
+	cancel_delayed_work_sync(&adev->jpeg.idle_work);
 
 	if (adev->jpeg.cur_state != AMD_PG_STATE_GATE &&
 	      RREG32_SOC15(JPEG, 0, mmUVD_JRBC_STATUS))
-		jpeg_v2_0_set_powergating_state(adev, AMD_PG_STATE_GATE);
+		jpeg_v2_0_set_powergating_state(ip_block, AMD_PG_STATE_GATE);
 
 	return 0;
 }
@@ -164,20 +189,19 @@ static int jpeg_v2_0_hw_fini(void *handle)
 /**
  * jpeg_v2_0_suspend - suspend JPEG block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * HW fini and suspend JPEG block
  */
-static int jpeg_v2_0_suspend(void *handle)
+static int jpeg_v2_0_suspend(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 	int r;
 
-	r = jpeg_v2_0_hw_fini(adev);
+	r = jpeg_v2_0_hw_fini(ip_block);
 	if (r)
 		return r;
 
-	r = amdgpu_jpeg_suspend(adev);
+	r = amdgpu_jpeg_suspend(ip_block->adev);
 
 	return r;
 }
@@ -185,20 +209,19 @@ static int jpeg_v2_0_suspend(void *handle)
 /**
  * jpeg_v2_0_resume - resume JPEG block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Resume firmware and hw init JPEG block
  */
-static int jpeg_v2_0_resume(void *handle)
+static int jpeg_v2_0_resume(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
-	r = amdgpu_jpeg_resume(adev);
+	r = amdgpu_jpeg_resume(ip_block->adev);
 	if (r)
 		return r;
 
-	r = jpeg_v2_0_hw_init(adev);
+	r = jpeg_v2_0_hw_init(ip_block);
 
 	return r;
 }
@@ -311,7 +334,7 @@ static void jpeg_v2_0_enable_clock_gating(struct amdgpu_device *adev)
  */
 static int jpeg_v2_0_start(struct amdgpu_device *adev)
 {
-	struct amdgpu_ring *ring = &adev->jpeg.inst->ring_dec;
+	struct amdgpu_ring *ring = adev->jpeg.inst->ring_dec;
 	int r;
 
 	if (adev->pm.dpm_enabled)
@@ -407,7 +430,7 @@ static uint64_t jpeg_v2_0_dec_ring_get_wptr(struct amdgpu_ring *ring)
 	struct amdgpu_device *adev = ring->adev;
 
 	if (ring->use_doorbell)
-		return adev->wb.wb[ring->wptr_offs];
+		return *ring->wptr_cpu_addr;
 	else
 		return RREG32_SOC15(JPEG, 0, mmUVD_JRBC_RB_WPTR);
 }
@@ -424,7 +447,7 @@ static void jpeg_v2_0_dec_ring_set_wptr(struct amdgpu_ring *ring)
 	struct amdgpu_device *adev = ring->adev;
 
 	if (ring->use_doorbell) {
-		adev->wb.wb[ring->wptr_offs] = lower_32_bits(ring->wptr);
+		*ring->wptr_cpu_addr = lower_32_bits(ring->wptr);
 		WDOORBELL32(ring->doorbell_index, lower_32_bits(ring->wptr));
 	} else {
 		WREG32_SOC15(JPEG, 0, mmUVD_JRBC_RB_WPTR, lower_32_bits(ring->wptr));
@@ -535,13 +558,21 @@ void jpeg_v2_0_dec_ring_emit_ib(struct amdgpu_ring *ring,
 {
 	unsigned vmid = AMDGPU_JOB_GET_VMID(job);
 
+	amdgpu_ring_write(ring,	PACKETJ(mmUVD_JPEG_IH_CTRL_INTERNAL_OFFSET,
+		0, 0, PACKETJ_TYPE0));
+	amdgpu_ring_write(ring, (vmid << JPEG_IH_CTRL__IH_VMID__SHIFT));
+
 	amdgpu_ring_write(ring, PACKETJ(mmUVD_LMI_JRBC_IB_VMID_INTERNAL_OFFSET,
 		0, 0, PACKETJ_TYPE0));
-	amdgpu_ring_write(ring, (vmid | (vmid << 4)));
+
+	if (ring->funcs->parse_cs)
+		amdgpu_ring_write(ring, 0);
+	else
+		amdgpu_ring_write(ring, (vmid | (vmid << 4) | (vmid << 8)));
 
 	amdgpu_ring_write(ring, PACKETJ(mmUVD_LMI_JPEG_VMID_INTERNAL_OFFSET,
 		0, 0, PACKETJ_TYPE0));
-	amdgpu_ring_write(ring, (vmid | (vmid << 4)));
+	amdgpu_ring_write(ring, (vmid | (vmid << 4) | (vmid << 8)));
 
 	amdgpu_ring_write(ring,	PACKETJ(mmUVD_LMI_JRBC_IB_64BIT_BAR_LOW_INTERNAL_OFFSET,
 		0, 0, PACKETJ_TYPE0));
@@ -609,7 +640,7 @@ void jpeg_v2_0_dec_ring_emit_reg_wait(struct amdgpu_ring *ring, uint32_t reg,
 void jpeg_v2_0_dec_ring_emit_vm_flush(struct amdgpu_ring *ring,
 				unsigned vmid, uint64_t pd_addr)
 {
-	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
+	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->vm_hub];
 	uint32_t data0, data1, mask;
 
 	pd_addr = amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pd_addr);
@@ -651,18 +682,18 @@ void jpeg_v2_0_dec_ring_nop(struct amdgpu_ring *ring, uint32_t count)
 	}
 }
 
-static bool jpeg_v2_0_is_idle(void *handle)
+static bool jpeg_v2_0_is_idle(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	return ((RREG32_SOC15(JPEG, 0, mmUVD_JRBC_STATUS) &
 		UVD_JRBC_STATUS__RB_JOB_DONE_MASK) ==
 		UVD_JRBC_STATUS__RB_JOB_DONE_MASK);
 }
 
-static int jpeg_v2_0_wait_for_idle(void *handle)
+static int jpeg_v2_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int ret;
 
 	ret = SOC15_WAIT_ON_RREG(JPEG, 0, mmUVD_JRBC_STATUS, UVD_JRBC_STATUS__RB_JOB_DONE_MASK,
@@ -671,14 +702,14 @@ static int jpeg_v2_0_wait_for_idle(void *handle)
 	return ret;
 }
 
-static int jpeg_v2_0_set_clockgating_state(void *handle,
+static int jpeg_v2_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_clockgating_state state)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	bool enable = (state == AMD_CG_STATE_GATE);
 
 	if (enable) {
-		if (!jpeg_v2_0_is_idle(handle))
+		if (!jpeg_v2_0_is_idle(ip_block))
 			return -EBUSY;
 		jpeg_v2_0_enable_clock_gating(adev);
 	} else {
@@ -688,10 +719,10 @@ static int jpeg_v2_0_set_clockgating_state(void *handle,
 	return 0;
 }
 
-static int jpeg_v2_0_set_powergating_state(void *handle,
+static int jpeg_v2_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					enum amd_powergating_state state)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int ret;
 
 	if (state == adev->jpeg.cur_state)
@@ -724,7 +755,7 @@ static int jpeg_v2_0_process_interrupt(struct amdgpu_device *adev,
 
 	switch (entry->src_id) {
 	case VCN_2_0__SRCID__JPEG_DECODE:
-		amdgpu_fence_process(&adev->jpeg.inst->ring_dec);
+		amdgpu_fence_process(adev->jpeg.inst->ring_dec);
 		break;
 	default:
 		DRM_ERROR("Unhandled interrupt: %d %d\n",
@@ -735,10 +766,25 @@ static int jpeg_v2_0_process_interrupt(struct amdgpu_device *adev,
 	return 0;
 }
 
+static int jpeg_v2_0_ring_reset(struct amdgpu_ring *ring,
+				unsigned int vmid,
+				struct amdgpu_fence *timedout_fence)
+{
+	int r;
+
+	amdgpu_ring_reset_helper_begin(ring, timedout_fence);
+	r = jpeg_v2_0_stop(ring->adev);
+	if (r)
+		return r;
+	r = jpeg_v2_0_start(ring->adev);
+	if (r)
+		return r;
+	return amdgpu_ring_reset_helper_end(ring, timedout_fence);
+}
+
 static const struct amd_ip_funcs jpeg_v2_0_ip_funcs = {
 	.name = "jpeg_v2_0",
 	.early_init = jpeg_v2_0_early_init,
-	.late_init = NULL,
 	.sw_init = jpeg_v2_0_sw_init,
 	.sw_fini = jpeg_v2_0_sw_fini,
 	.hw_init = jpeg_v2_0_hw_init,
@@ -747,28 +793,26 @@ static const struct amd_ip_funcs jpeg_v2_0_ip_funcs = {
 	.resume = jpeg_v2_0_resume,
 	.is_idle = jpeg_v2_0_is_idle,
 	.wait_for_idle = jpeg_v2_0_wait_for_idle,
-	.check_soft_reset = NULL,
-	.pre_soft_reset = NULL,
-	.soft_reset = NULL,
-	.post_soft_reset = NULL,
 	.set_clockgating_state = jpeg_v2_0_set_clockgating_state,
 	.set_powergating_state = jpeg_v2_0_set_powergating_state,
+	.dump_ip_state = amdgpu_jpeg_dump_ip_state,
+	.print_ip_state = amdgpu_jpeg_print_ip_state,
 };
 
 static const struct amdgpu_ring_funcs jpeg_v2_0_dec_ring_vm_funcs = {
 	.type = AMDGPU_RING_TYPE_VCN_JPEG,
 	.align_mask = 0xf,
-	.vmhub = AMDGPU_MMHUB_0,
 	.get_rptr = jpeg_v2_0_dec_ring_get_rptr,
 	.get_wptr = jpeg_v2_0_dec_ring_get_wptr,
 	.set_wptr = jpeg_v2_0_dec_ring_set_wptr,
+	.parse_cs = amdgpu_jpeg_dec_parse_cs,
 	.emit_frame_size =
 		SOC15_FLUSH_GPU_TLB_NUM_WREG * 6 +
 		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 8 +
 		8 + /* jpeg_v2_0_dec_ring_emit_vm_flush */
 		18 + 18 + /* jpeg_v2_0_dec_ring_emit_fence x2 vm fence */
 		8 + 16,
-	.emit_ib_size = 22, /* jpeg_v2_0_dec_ring_emit_ib */
+	.emit_ib_size = 24, /* jpeg_v2_0_dec_ring_emit_ib */
 	.emit_ib = jpeg_v2_0_dec_ring_emit_ib,
 	.emit_fence = jpeg_v2_0_dec_ring_emit_fence,
 	.emit_vm_flush = jpeg_v2_0_dec_ring_emit_vm_flush,
@@ -783,12 +827,12 @@ static const struct amdgpu_ring_funcs jpeg_v2_0_dec_ring_vm_funcs = {
 	.emit_wreg = jpeg_v2_0_dec_ring_emit_wreg,
 	.emit_reg_wait = jpeg_v2_0_dec_ring_emit_reg_wait,
 	.emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper,
+	.reset = jpeg_v2_0_ring_reset,
 };
 
 static void jpeg_v2_0_set_dec_ring_funcs(struct amdgpu_device *adev)
 {
-	adev->jpeg.inst->ring_dec.funcs = &jpeg_v2_0_dec_ring_vm_funcs;
-	DRM_INFO("JPEG decode is enabled in VM mode\n");
+	adev->jpeg.inst->ring_dec->funcs = &jpeg_v2_0_dec_ring_vm_funcs;
 }
 
 static const struct amdgpu_irq_src_funcs jpeg_v2_0_irq_funcs = {
@@ -802,8 +846,7 @@ static void jpeg_v2_0_set_irq_funcs(struct amdgpu_device *adev)
 	adev->jpeg.inst->irq.funcs = &jpeg_v2_0_irq_funcs;
 }
 
-const struct amdgpu_ip_block_version jpeg_v2_0_ip_block =
-{
+const struct amdgpu_ip_block_version jpeg_v2_0_ip_block = {
 		.type = AMD_IP_BLOCK_TYPE_JPEG,
 		.major = 2,
 		.minor = 0,
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_0.h b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_0.h
index 1a03baa59755..654e43e83e2c 100644
--- a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_0.h
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_0.h
@@ -41,6 +41,7 @@
 #define mmUVD_JRBC_RB_REF_DATA_INTERNAL_OFFSET				0x4084
 #define mmUVD_JRBC_STATUS_INTERNAL_OFFSET				0x4089
 #define mmUVD_JPEG_PITCH_INTERNAL_OFFSET				0x401f
+#define mmUVD_JPEG_IH_CTRL_INTERNAL_OFFSET				0x4149
 
 #define JRBC_DEC_EXTERNAL_REG_WRITE_ADDR				0x18000
 
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c
index a29c86617fb5..20983f126b49 100644
--- a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c
@@ -26,6 +26,7 @@
 #include "soc15.h"
 #include "soc15d.h"
 #include "jpeg_v2_0.h"
+#include "jpeg_v2_5.h"
 
 #include "vcn/vcn_2_5_offset.h"
 #include "vcn/vcn_2_5_sh_mask.h"
@@ -35,10 +36,27 @@
 
 #define JPEG25_MAX_HW_INSTANCES_ARCTURUS			2
 
+static const struct amdgpu_hwip_reg_entry jpeg_reg_list_2_5[] = {
+	SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JPEG_POWER_STATUS),
+	SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JPEG_INT_STAT),
+	SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JRBC_RB_RPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JRBC_RB_WPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JRBC_RB_CNTL),
+	SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JRBC_RB_SIZE),
+	SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JRBC_STATUS),
+	SOC15_REG_ENTRY_STR(JPEG, 0, mmJPEG_DEC_ADDR_MODE),
+	SOC15_REG_ENTRY_STR(JPEG, 0, mmJPEG_DEC_GFX10_ADDR_CONFIG),
+	SOC15_REG_ENTRY_STR(JPEG, 0, mmJPEG_DEC_Y_GFX10_TILING_SURFACE),
+	SOC15_REG_ENTRY_STR(JPEG, 0, mmJPEG_DEC_UV_GFX10_TILING_SURFACE),
+	SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JPEG_PITCH),
+	SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JPEG_UV_PITCH),
+};
+
 static void jpeg_v2_5_set_dec_ring_funcs(struct amdgpu_device *adev);
 static void jpeg_v2_5_set_irq_funcs(struct amdgpu_device *adev);
-static int jpeg_v2_5_set_powergating_state(void *handle,
+static int jpeg_v2_5_set_powergating_state(struct amdgpu_ip_block *ip_block,
 				enum amd_powergating_state state);
+static void jpeg_v2_5_set_ras_funcs(struct amdgpu_device *adev);
 
 static int amdgpu_ih_clientid_jpeg[] = {
 	SOC15_IH_CLIENTID_VCN,
@@ -48,16 +66,17 @@ static int amdgpu_ih_clientid_jpeg[] = {
 /**
  * jpeg_v2_5_early_init - set function pointers
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Set ring and irq function pointers
  */
-static int jpeg_v2_5_early_init(void *handle)
+static int jpeg_v2_5_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	u32 harvest;
 	int i;
 
+	adev->jpeg.num_jpeg_rings = 1;
 	adev->jpeg.num_jpeg_inst = JPEG25_MAX_HW_INSTANCES_ARCTURUS;
 	for (i = 0; i < adev->jpeg.num_jpeg_inst; i++) {
 		harvest = RREG32_SOC15(JPEG, i, mmCC_UVD_HARVESTING);
@@ -70,6 +89,7 @@ static int jpeg_v2_5_early_init(void *handle)
 
 	jpeg_v2_5_set_dec_ring_funcs(adev);
 	jpeg_v2_5_set_irq_funcs(adev);
+	jpeg_v2_5_set_ras_funcs(adev);
 
 	return 0;
 }
@@ -77,15 +97,15 @@ static int jpeg_v2_5_early_init(void *handle)
 /**
  * jpeg_v2_5_sw_init - sw init for JPEG block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Load firmware and sw initialization
  */
-static int jpeg_v2_5_sw_init(void *handle)
+static int jpeg_v2_5_sw_init(struct amdgpu_ip_block *ip_block)
 {
 	struct amdgpu_ring *ring;
 	int i, r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
 		if (adev->jpeg.harvest_config & (1 << i))
@@ -96,6 +116,18 @@ static int jpeg_v2_5_sw_init(void *handle)
 				VCN_2_0__SRCID__JPEG_DECODE, &adev->jpeg.inst[i].irq);
 		if (r)
 			return r;
+
+		/* JPEG DJPEG POISON EVENT */
+		r = amdgpu_irq_add_id(adev, amdgpu_ih_clientid_jpeg[i],
+			VCN_2_6__SRCID_DJPEG0_POISON, &adev->jpeg.inst[i].ras_poison_irq);
+		if (r)
+			return r;
+
+		/* JPEG EJPEG POISON EVENT */
+		r = amdgpu_irq_add_id(adev, amdgpu_ih_clientid_jpeg[i],
+			VCN_2_6__SRCID_EJPEG0_POISON, &adev->jpeg.inst[i].ras_poison_irq);
+		if (r)
+			return r;
 	}
 
 	r = amdgpu_jpeg_sw_init(adev);
@@ -110,8 +142,12 @@ static int jpeg_v2_5_sw_init(void *handle)
 		if (adev->jpeg.harvest_config & (1 << i))
 			continue;
 
-		ring = &adev->jpeg.inst[i].ring_dec;
+		ring = adev->jpeg.inst[i].ring_dec;
 		ring->use_doorbell = true;
+		if (amdgpu_ip_version(adev, UVD_HWIP, 0) == IP_VERSION(2, 5, 0))
+			ring->vm_hub = AMDGPU_MMHUB1(0);
+		else
+			ring->vm_hub = AMDGPU_MMHUB0(0);
 		ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 1 + 8 * i;
 		sprintf(ring->name, "jpeg_dec_%d", i);
 		r = amdgpu_ring_init(adev, ring, 512, &adev->jpeg.inst[i].irq,
@@ -119,29 +155,45 @@ static int jpeg_v2_5_sw_init(void *handle)
 		if (r)
 			return r;
 
-		adev->jpeg.internal.jpeg_pitch = mmUVD_JPEG_PITCH_INTERNAL_OFFSET;
-		adev->jpeg.inst[i].external.jpeg_pitch = SOC15_REG_OFFSET(JPEG, i, mmUVD_JPEG_PITCH);
+		adev->jpeg.internal.jpeg_pitch[0] = mmUVD_JPEG_PITCH_INTERNAL_OFFSET;
+		adev->jpeg.inst[i].external.jpeg_pitch[0] = SOC15_REG_OFFSET(JPEG, i, mmUVD_JPEG_PITCH);
 	}
 
-	return 0;
+	r = amdgpu_jpeg_ras_sw_init(adev);
+	if (r)
+		return r;
+
+	r = amdgpu_jpeg_reg_dump_init(adev, jpeg_reg_list_2_5, ARRAY_SIZE(jpeg_reg_list_2_5));
+	if (r)
+		return r;
+
+	adev->jpeg.supported_reset =
+		amdgpu_get_soft_full_reset_mask(adev->jpeg.inst[0].ring_dec);
+	if (!amdgpu_sriov_vf(adev))
+		adev->jpeg.supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE;
+	r = amdgpu_jpeg_sysfs_reset_mask_init(adev);
+
+	return r;
 }
 
 /**
  * jpeg_v2_5_sw_fini - sw fini for JPEG block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * JPEG suspend and free up sw allocation
  */
-static int jpeg_v2_5_sw_fini(void *handle)
+static int jpeg_v2_5_sw_fini(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	r = amdgpu_jpeg_suspend(adev);
 	if (r)
 		return r;
 
+	amdgpu_jpeg_sysfs_reset_mask_fini(adev);
+
 	r = amdgpu_jpeg_sw_fini(adev);
 
 	return r;
@@ -150,12 +202,12 @@ static int jpeg_v2_5_sw_fini(void *handle)
 /**
  * jpeg_v2_5_hw_init - start and test JPEG block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  */
-static int jpeg_v2_5_hw_init(void *handle)
+static int jpeg_v2_5_hw_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	struct amdgpu_ring *ring;
 	int i, r;
 
@@ -163,7 +215,7 @@ static int jpeg_v2_5_hw_init(void *handle)
 		if (adev->jpeg.harvest_config & (1 << i))
 			continue;
 
-		ring = &adev->jpeg.inst[i].ring_dec;
+		ring = adev->jpeg.inst[i].ring_dec;
 		adev->nbio.funcs->vcn_doorbell_range(adev, ring->use_doorbell,
 			(adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 8 * i, i);
 
@@ -172,24 +224,22 @@ static int jpeg_v2_5_hw_init(void *handle)
 			return r;
 	}
 
-	DRM_INFO("JPEG decode initialized successfully.\n");
-
 	return 0;
 }
 
 /**
  * jpeg_v2_5_hw_fini - stop the hardware block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Stop the JPEG block, mark ring as not ready any more
  */
-static int jpeg_v2_5_hw_fini(void *handle)
+static int jpeg_v2_5_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int i;
 
-	cancel_delayed_work_sync(&adev->vcn.idle_work);
+	cancel_delayed_work_sync(&adev->jpeg.idle_work);
 
 	for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
 		if (adev->jpeg.harvest_config & (1 << i))
@@ -197,7 +247,10 @@ static int jpeg_v2_5_hw_fini(void *handle)
 
 		if (adev->jpeg.cur_state != AMD_PG_STATE_GATE &&
 		      RREG32_SOC15(JPEG, i, mmUVD_JRBC_STATUS))
-			jpeg_v2_5_set_powergating_state(adev, AMD_PG_STATE_GATE);
+			jpeg_v2_5_set_powergating_state(ip_block, AMD_PG_STATE_GATE);
+
+		if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__JPEG))
+			amdgpu_irq_put(adev, &adev->jpeg.inst[i].ras_poison_irq, 0);
 	}
 
 	return 0;
@@ -206,20 +259,19 @@ static int jpeg_v2_5_hw_fini(void *handle)
 /**
  * jpeg_v2_5_suspend - suspend JPEG block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * HW fini and suspend JPEG block
  */
-static int jpeg_v2_5_suspend(void *handle)
+static int jpeg_v2_5_suspend(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 	int r;
 
-	r = jpeg_v2_5_hw_fini(adev);
+	r = jpeg_v2_5_hw_fini(ip_block);
 	if (r)
 		return r;
 
-	r = amdgpu_jpeg_suspend(adev);
+	r = amdgpu_jpeg_suspend(ip_block->adev);
 
 	return r;
 }
@@ -227,20 +279,19 @@ static int jpeg_v2_5_suspend(void *handle)
 /**
  * jpeg_v2_5_resume - resume JPEG block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Resume firmware and hw init JPEG block
  */
-static int jpeg_v2_5_resume(void *handle)
+static int jpeg_v2_5_resume(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 	int r;
 
-	r = amdgpu_jpeg_resume(adev);
+	r = amdgpu_jpeg_resume(ip_block->adev);
 	if (r)
 		return r;
 
-	r = jpeg_v2_5_hw_init(adev);
+	r = jpeg_v2_5_hw_init(ip_block);
 
 	return r;
 }
@@ -287,6 +338,44 @@ static void jpeg_v2_5_enable_clock_gating(struct amdgpu_device *adev, int inst)
 	WREG32_SOC15(JPEG, inst, mmJPEG_CGC_GATE, data);
 }
 
+static void jpeg_v2_5_start_inst(struct amdgpu_device *adev, int i)
+{
+	struct amdgpu_ring *ring = adev->jpeg.inst[i].ring_dec;
+	/* disable anti hang mechanism */
+	WREG32_P(SOC15_REG_OFFSET(JPEG, i, mmUVD_JPEG_POWER_STATUS), 0,
+		~UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK);
+
+	/* JPEG disable CGC */
+	jpeg_v2_5_disable_clock_gating(adev, i);
+
+	/* MJPEG global tiling registers */
+	WREG32_SOC15(JPEG, i, mmJPEG_DEC_GFX8_ADDR_CONFIG,
+		adev->gfx.config.gb_addr_config);
+	WREG32_SOC15(JPEG, i, mmJPEG_DEC_GFX10_ADDR_CONFIG,
+		adev->gfx.config.gb_addr_config);
+
+	/* enable JMI channel */
+	WREG32_P(SOC15_REG_OFFSET(JPEG, i, mmUVD_JMI_CNTL), 0,
+		~UVD_JMI_CNTL__SOFT_RESET_MASK);
+
+	/* enable System Interrupt for JRBC */
+	WREG32_P(SOC15_REG_OFFSET(JPEG, i, mmJPEG_SYS_INT_EN),
+		JPEG_SYS_INT_EN__DJRBC_MASK,
+		~JPEG_SYS_INT_EN__DJRBC_MASK);
+
+	WREG32_SOC15(JPEG, i, mmUVD_LMI_JRBC_RB_VMID, 0);
+	WREG32_SOC15(JPEG, i, mmUVD_JRBC_RB_CNTL, (0x00000001L | 0x00000002L));
+	WREG32_SOC15(JPEG, i, mmUVD_LMI_JRBC_RB_64BIT_BAR_LOW,
+		lower_32_bits(ring->gpu_addr));
+	WREG32_SOC15(JPEG, i, mmUVD_LMI_JRBC_RB_64BIT_BAR_HIGH,
+		upper_32_bits(ring->gpu_addr));
+	WREG32_SOC15(JPEG, i, mmUVD_JRBC_RB_RPTR, 0);
+	WREG32_SOC15(JPEG, i, mmUVD_JRBC_RB_WPTR, 0);
+	WREG32_SOC15(JPEG, i, mmUVD_JRBC_RB_CNTL, 0x00000002L);
+	WREG32_SOC15(JPEG, i, mmUVD_JRBC_RB_SIZE, ring->ring_size / 4);
+	ring->wptr = RREG32_SOC15(JPEG, i, mmUVD_JRBC_RB_WPTR);
+}
+
 /**
  * jpeg_v2_5_start - start JPEG block
  *
@@ -296,52 +385,33 @@ static void jpeg_v2_5_enable_clock_gating(struct amdgpu_device *adev, int inst)
  */
 static int jpeg_v2_5_start(struct amdgpu_device *adev)
 {
-	struct amdgpu_ring *ring;
 	int i;
 
 	for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
 		if (adev->jpeg.harvest_config & (1 << i))
 			continue;
+		jpeg_v2_5_start_inst(adev, i);
 
-		ring = &adev->jpeg.inst[i].ring_dec;
-		/* disable anti hang mechanism */
-		WREG32_P(SOC15_REG_OFFSET(JPEG, i, mmUVD_JPEG_POWER_STATUS), 0,
-			~UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK);
-
-		/* JPEG disable CGC */
-		jpeg_v2_5_disable_clock_gating(adev, i);
-
-		/* MJPEG global tiling registers */
-		WREG32_SOC15(JPEG, i, mmJPEG_DEC_GFX8_ADDR_CONFIG,
-			adev->gfx.config.gb_addr_config);
-		WREG32_SOC15(JPEG, i, mmJPEG_DEC_GFX10_ADDR_CONFIG,
-			adev->gfx.config.gb_addr_config);
-
-		/* enable JMI channel */
-		WREG32_P(SOC15_REG_OFFSET(JPEG, i, mmUVD_JMI_CNTL), 0,
-			~UVD_JMI_CNTL__SOFT_RESET_MASK);
-
-		/* enable System Interrupt for JRBC */
-		WREG32_P(SOC15_REG_OFFSET(JPEG, i, mmJPEG_SYS_INT_EN),
-			JPEG_SYS_INT_EN__DJRBC_MASK,
-			~JPEG_SYS_INT_EN__DJRBC_MASK);
-
-		WREG32_SOC15(JPEG, i, mmUVD_LMI_JRBC_RB_VMID, 0);
-		WREG32_SOC15(JPEG, i, mmUVD_JRBC_RB_CNTL, (0x00000001L | 0x00000002L));
-		WREG32_SOC15(JPEG, i, mmUVD_LMI_JRBC_RB_64BIT_BAR_LOW,
-			lower_32_bits(ring->gpu_addr));
-		WREG32_SOC15(JPEG, i, mmUVD_LMI_JRBC_RB_64BIT_BAR_HIGH,
-			upper_32_bits(ring->gpu_addr));
-		WREG32_SOC15(JPEG, i, mmUVD_JRBC_RB_RPTR, 0);
-		WREG32_SOC15(JPEG, i, mmUVD_JRBC_RB_WPTR, 0);
-		WREG32_SOC15(JPEG, i, mmUVD_JRBC_RB_CNTL, 0x00000002L);
-		WREG32_SOC15(JPEG, i, mmUVD_JRBC_RB_SIZE, ring->ring_size / 4);
-		ring->wptr = RREG32_SOC15(JPEG, i, mmUVD_JRBC_RB_WPTR);
 	}
 
 	return 0;
 }
 
+static void jpeg_v2_5_stop_inst(struct amdgpu_device *adev, int i)
+{
+	/* reset JMI */
+	WREG32_P(SOC15_REG_OFFSET(JPEG, i, mmUVD_JMI_CNTL),
+		UVD_JMI_CNTL__SOFT_RESET_MASK,
+		~UVD_JMI_CNTL__SOFT_RESET_MASK);
+
+	jpeg_v2_5_enable_clock_gating(adev, i);
+
+	/* enable anti hang mechanism */
+	WREG32_P(SOC15_REG_OFFSET(JPEG, i, mmUVD_JPEG_POWER_STATUS),
+		UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK,
+		~UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK);
+}
+
 /**
  * jpeg_v2_5_stop - stop JPEG block
  *
@@ -356,18 +426,7 @@ static int jpeg_v2_5_stop(struct amdgpu_device *adev)
 	for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
 		if (adev->jpeg.harvest_config & (1 << i))
 			continue;
-
-		/* reset JMI */
-		WREG32_P(SOC15_REG_OFFSET(JPEG, i, mmUVD_JMI_CNTL),
-			UVD_JMI_CNTL__SOFT_RESET_MASK,
-			~UVD_JMI_CNTL__SOFT_RESET_MASK);
-
-		jpeg_v2_5_enable_clock_gating(adev, i);
-
-		/* enable anti hang mechanism */
-		WREG32_P(SOC15_REG_OFFSET(JPEG, i, mmUVD_JPEG_POWER_STATUS),
-			UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK,
-			~UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK);
+		jpeg_v2_5_stop_inst(adev, i);
 	}
 
 	return 0;
@@ -399,7 +458,7 @@ static uint64_t jpeg_v2_5_dec_ring_get_wptr(struct amdgpu_ring *ring)
 	struct amdgpu_device *adev = ring->adev;
 
 	if (ring->use_doorbell)
-		return adev->wb.wb[ring->wptr_offs];
+		return *ring->wptr_cpu_addr;
 	else
 		return RREG32_SOC15(JPEG, ring->me, mmUVD_JRBC_RB_WPTR);
 }
@@ -416,7 +475,7 @@ static void jpeg_v2_5_dec_ring_set_wptr(struct amdgpu_ring *ring)
 	struct amdgpu_device *adev = ring->adev;
 
 	if (ring->use_doorbell) {
-		adev->wb.wb[ring->wptr_offs] = lower_32_bits(ring->wptr);
+		*ring->wptr_cpu_addr = lower_32_bits(ring->wptr);
 		WDOORBELL32(ring->doorbell_index, lower_32_bits(ring->wptr));
 	} else {
 		WREG32_SOC15(JPEG, ring->me, mmUVD_JRBC_RB_WPTR, lower_32_bits(ring->wptr));
@@ -459,9 +518,9 @@ static void jpeg_v2_6_dec_ring_insert_end(struct amdgpu_ring *ring)
 	amdgpu_ring_write(ring, (1 << (ring->me * 2 + 14)));
 }
 
-static bool jpeg_v2_5_is_idle(void *handle)
+static bool jpeg_v2_5_is_idle(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int i, ret = 1;
 
 	for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
@@ -476,9 +535,9 @@ static bool jpeg_v2_5_is_idle(void *handle)
 	return ret;
 }
 
-static int jpeg_v2_5_wait_for_idle(void *handle)
+static int jpeg_v2_5_wait_for_idle(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int i, ret;
 
 	for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
@@ -495,10 +554,10 @@ static int jpeg_v2_5_wait_for_idle(void *handle)
 	return 0;
 }
 
-static int jpeg_v2_5_set_clockgating_state(void *handle,
+static int jpeg_v2_5_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_clockgating_state state)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	bool enable = (state == AMD_CG_STATE_GATE);
 	int i;
 
@@ -507,7 +566,7 @@ static int jpeg_v2_5_set_clockgating_state(void *handle,
 			continue;
 
 		if (enable) {
-			if (!jpeg_v2_5_is_idle(handle))
+			if (!jpeg_v2_5_is_idle(ip_block))
 				return -EBUSY;
 			jpeg_v2_5_enable_clock_gating(adev, i);
 		} else {
@@ -518,13 +577,13 @@ static int jpeg_v2_5_set_clockgating_state(void *handle,
 	return 0;
 }
 
-static int jpeg_v2_5_set_powergating_state(void *handle,
+static int jpeg_v2_5_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_powergating_state state)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int ret;
 
-	if(state == adev->jpeg.cur_state)
+	if (state == adev->jpeg.cur_state)
 		return 0;
 
 	if (state == AMD_PG_STATE_GATE)
@@ -532,7 +591,7 @@ static int jpeg_v2_5_set_powergating_state(void *handle,
 	else
 		ret = jpeg_v2_5_start(adev);
 
-	if(!ret)
+	if (!ret)
 		adev->jpeg.cur_state = state;
 
 	return ret;
@@ -546,6 +605,14 @@ static int jpeg_v2_5_set_interrupt_state(struct amdgpu_device *adev,
 	return 0;
 }
 
+static int jpeg_v2_6_set_ras_interrupt_state(struct amdgpu_device *adev,
+					struct amdgpu_irq_src *source,
+					unsigned int type,
+					enum amdgpu_interrupt_state state)
+{
+	return 0;
+}
+
 static int jpeg_v2_5_process_interrupt(struct amdgpu_device *adev,
 				      struct amdgpu_irq_src *source,
 				      struct amdgpu_iv_entry *entry)
@@ -568,7 +635,7 @@ static int jpeg_v2_5_process_interrupt(struct amdgpu_device *adev,
 
 	switch (entry->src_id) {
 	case VCN_2_0__SRCID__JPEG_DECODE:
-		amdgpu_fence_process(&adev->jpeg.inst[ip_instance].ring_dec);
+		amdgpu_fence_process(adev->jpeg.inst[ip_instance].ring_dec);
 		break;
 	default:
 		DRM_ERROR("Unhandled interrupt: %d %d\n",
@@ -579,10 +646,19 @@ static int jpeg_v2_5_process_interrupt(struct amdgpu_device *adev,
 	return 0;
 }
 
+static int jpeg_v2_5_ring_reset(struct amdgpu_ring *ring,
+				unsigned int vmid,
+				struct amdgpu_fence *timedout_fence)
+{
+	amdgpu_ring_reset_helper_begin(ring, timedout_fence);
+	jpeg_v2_5_stop_inst(ring->adev, ring->me);
+	jpeg_v2_5_start_inst(ring->adev, ring->me);
+	return amdgpu_ring_reset_helper_end(ring, timedout_fence);
+}
+
 static const struct amd_ip_funcs jpeg_v2_5_ip_funcs = {
 	.name = "jpeg_v2_5",
 	.early_init = jpeg_v2_5_early_init,
-	.late_init = NULL,
 	.sw_init = jpeg_v2_5_sw_init,
 	.sw_fini = jpeg_v2_5_sw_fini,
 	.hw_init = jpeg_v2_5_hw_init,
@@ -591,18 +667,15 @@ static const struct amd_ip_funcs jpeg_v2_5_ip_funcs = {
 	.resume = jpeg_v2_5_resume,
 	.is_idle = jpeg_v2_5_is_idle,
 	.wait_for_idle = jpeg_v2_5_wait_for_idle,
-	.check_soft_reset = NULL,
-	.pre_soft_reset = NULL,
-	.soft_reset = NULL,
-	.post_soft_reset = NULL,
 	.set_clockgating_state = jpeg_v2_5_set_clockgating_state,
 	.set_powergating_state = jpeg_v2_5_set_powergating_state,
+	.dump_ip_state = amdgpu_jpeg_dump_ip_state,
+	.print_ip_state = amdgpu_jpeg_print_ip_state,
 };
 
 static const struct amd_ip_funcs jpeg_v2_6_ip_funcs = {
 	.name = "jpeg_v2_6",
 	.early_init = jpeg_v2_5_early_init,
-	.late_init = NULL,
 	.sw_init = jpeg_v2_5_sw_init,
 	.sw_fini = jpeg_v2_5_sw_fini,
 	.hw_init = jpeg_v2_5_hw_init,
@@ -611,21 +684,19 @@ static const struct amd_ip_funcs jpeg_v2_6_ip_funcs = {
 	.resume = jpeg_v2_5_resume,
 	.is_idle = jpeg_v2_5_is_idle,
 	.wait_for_idle = jpeg_v2_5_wait_for_idle,
-	.check_soft_reset = NULL,
-	.pre_soft_reset = NULL,
-	.soft_reset = NULL,
-	.post_soft_reset = NULL,
 	.set_clockgating_state = jpeg_v2_5_set_clockgating_state,
 	.set_powergating_state = jpeg_v2_5_set_powergating_state,
+	.dump_ip_state = amdgpu_jpeg_dump_ip_state,
+	.print_ip_state = amdgpu_jpeg_print_ip_state,
 };
 
 static const struct amdgpu_ring_funcs jpeg_v2_5_dec_ring_vm_funcs = {
 	.type = AMDGPU_RING_TYPE_VCN_JPEG,
 	.align_mask = 0xf,
-	.vmhub = AMDGPU_MMHUB_1,
 	.get_rptr = jpeg_v2_5_dec_ring_get_rptr,
 	.get_wptr = jpeg_v2_5_dec_ring_get_wptr,
 	.set_wptr = jpeg_v2_5_dec_ring_set_wptr,
+	.parse_cs = amdgpu_jpeg_dec_parse_cs,
 	.emit_frame_size =
 		SOC15_FLUSH_GPU_TLB_NUM_WREG * 6 +
 		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 8 +
@@ -647,15 +718,16 @@ static const struct amdgpu_ring_funcs jpeg_v2_5_dec_ring_vm_funcs = {
 	.emit_wreg = jpeg_v2_0_dec_ring_emit_wreg,
 	.emit_reg_wait = jpeg_v2_0_dec_ring_emit_reg_wait,
 	.emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper,
+	.reset = jpeg_v2_5_ring_reset,
 };
 
 static const struct amdgpu_ring_funcs jpeg_v2_6_dec_ring_vm_funcs = {
 	.type = AMDGPU_RING_TYPE_VCN_JPEG,
 	.align_mask = 0xf,
-	.vmhub = AMDGPU_MMHUB_0,
 	.get_rptr = jpeg_v2_5_dec_ring_get_rptr,
 	.get_wptr = jpeg_v2_5_dec_ring_get_wptr,
 	.set_wptr = jpeg_v2_5_dec_ring_set_wptr,
+	.parse_cs = amdgpu_jpeg_dec_parse_cs,
 	.emit_frame_size =
 		SOC15_FLUSH_GPU_TLB_NUM_WREG * 6 +
 		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 8 +
@@ -677,6 +749,7 @@ static const struct amdgpu_ring_funcs jpeg_v2_6_dec_ring_vm_funcs = {
 	.emit_wreg = jpeg_v2_0_dec_ring_emit_wreg,
 	.emit_reg_wait = jpeg_v2_0_dec_ring_emit_reg_wait,
 	.emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper,
+	.reset = jpeg_v2_5_ring_reset,
 };
 
 static void jpeg_v2_5_set_dec_ring_funcs(struct amdgpu_device *adev)
@@ -687,11 +760,10 @@ static void jpeg_v2_5_set_dec_ring_funcs(struct amdgpu_device *adev)
 		if (adev->jpeg.harvest_config & (1 << i))
 			continue;
 		if (adev->asic_type == CHIP_ARCTURUS)
-			adev->jpeg.inst[i].ring_dec.funcs = &jpeg_v2_5_dec_ring_vm_funcs;
+			adev->jpeg.inst[i].ring_dec->funcs = &jpeg_v2_5_dec_ring_vm_funcs;
 		else  /* CHIP_ALDEBARAN */
-			adev->jpeg.inst[i].ring_dec.funcs = &jpeg_v2_6_dec_ring_vm_funcs;
-		adev->jpeg.inst[i].ring_dec.me = i;
-		DRM_INFO("JPEG(%d) JPEG decode is enabled in VM mode\n", i);
+			adev->jpeg.inst[i].ring_dec->funcs = &jpeg_v2_6_dec_ring_vm_funcs;
+		adev->jpeg.inst[i].ring_dec->me = i;
 	}
 }
 
@@ -700,6 +772,11 @@ static const struct amdgpu_irq_src_funcs jpeg_v2_5_irq_funcs = {
 	.process = jpeg_v2_5_process_interrupt,
 };
 
+static const struct amdgpu_irq_src_funcs jpeg_v2_6_ras_irq_funcs = {
+	.set = jpeg_v2_6_set_ras_interrupt_state,
+	.process = amdgpu_jpeg_process_poison_irq,
+};
+
 static void jpeg_v2_5_set_irq_funcs(struct amdgpu_device *adev)
 {
 	int i;
@@ -710,11 +787,13 @@ static void jpeg_v2_5_set_irq_funcs(struct amdgpu_device *adev)
 
 		adev->jpeg.inst[i].irq.num_types = 1;
 		adev->jpeg.inst[i].irq.funcs = &jpeg_v2_5_irq_funcs;
+
+		adev->jpeg.inst[i].ras_poison_irq.num_types = 1;
+		adev->jpeg.inst[i].ras_poison_irq.funcs = &jpeg_v2_6_ras_irq_funcs;
 	}
 }
 
-const struct amdgpu_ip_block_version jpeg_v2_5_ip_block =
-{
+const struct amdgpu_ip_block_version jpeg_v2_5_ip_block = {
 		.type = AMD_IP_BLOCK_TYPE_JPEG,
 		.major = 2,
 		.minor = 5,
@@ -722,11 +801,69 @@ const struct amdgpu_ip_block_version jpeg_v2_5_ip_block =
 		.funcs = &jpeg_v2_5_ip_funcs,
 };
 
-const struct amdgpu_ip_block_version jpeg_v2_6_ip_block =
-{
+const struct amdgpu_ip_block_version jpeg_v2_6_ip_block = {
 		.type = AMD_IP_BLOCK_TYPE_JPEG,
 		.major = 2,
 		.minor = 6,
 		.rev = 0,
 		.funcs = &jpeg_v2_6_ip_funcs,
 };
+
+static uint32_t jpeg_v2_6_query_poison_by_instance(struct amdgpu_device *adev,
+		uint32_t instance, uint32_t sub_block)
+{
+	uint32_t poison_stat = 0, reg_value = 0;
+
+	switch (sub_block) {
+	case AMDGPU_JPEG_V2_6_JPEG0:
+		reg_value = RREG32_SOC15(JPEG, instance, mmUVD_RAS_JPEG0_STATUS);
+		poison_stat = REG_GET_FIELD(reg_value, UVD_RAS_JPEG0_STATUS, POISONED_PF);
+		break;
+	case AMDGPU_JPEG_V2_6_JPEG1:
+		reg_value = RREG32_SOC15(JPEG, instance, mmUVD_RAS_JPEG1_STATUS);
+		poison_stat = REG_GET_FIELD(reg_value, UVD_RAS_JPEG1_STATUS, POISONED_PF);
+		break;
+	default:
+		break;
+	}
+
+	if (poison_stat)
+		dev_info(adev->dev, "Poison detected in JPEG%d sub_block%d\n",
+			instance, sub_block);
+
+	return poison_stat;
+}
+
+static bool jpeg_v2_6_query_ras_poison_status(struct amdgpu_device *adev)
+{
+	uint32_t inst = 0, sub = 0, poison_stat = 0;
+
+	for (inst = 0; inst < adev->jpeg.num_jpeg_inst; inst++)
+		for (sub = 0; sub < AMDGPU_JPEG_V2_6_MAX_SUB_BLOCK; sub++)
+			poison_stat +=
+			jpeg_v2_6_query_poison_by_instance(adev, inst, sub);
+
+	return !!poison_stat;
+}
+
+const struct amdgpu_ras_block_hw_ops jpeg_v2_6_ras_hw_ops = {
+	.query_poison_status = jpeg_v2_6_query_ras_poison_status,
+};
+
+static struct amdgpu_jpeg_ras jpeg_v2_6_ras = {
+	.ras_block = {
+		.hw_ops = &jpeg_v2_6_ras_hw_ops,
+		.ras_late_init = amdgpu_jpeg_ras_late_init,
+	},
+};
+
+static void jpeg_v2_5_set_ras_funcs(struct amdgpu_device *adev)
+{
+	switch (amdgpu_ip_version(adev, JPEG_HWIP, 0)) {
+	case IP_VERSION(2, 6, 0):
+		adev->jpeg.ras = &jpeg_v2_6_ras;
+		break;
+	default:
+		break;
+	}
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.h b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.h
index 3b0aa29b9879..1e858c6cdf13 100644
--- a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.h
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.h
@@ -24,6 +24,13 @@
 #ifndef __JPEG_V2_5_H__
 #define __JPEG_V2_5_H__
 
+enum amdgpu_jpeg_v2_6_sub_block {
+	AMDGPU_JPEG_V2_6_JPEG0 = 0,
+	AMDGPU_JPEG_V2_6_JPEG1,
+
+	AMDGPU_JPEG_V2_6_MAX_SUB_BLOCK,
+};
+
 extern const struct amdgpu_ip_block_version jpeg_v2_5_ip_block;
 extern const struct amdgpu_ip_block_version jpeg_v2_6_ip_block;
 
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v3_0.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v3_0.c
index 01c242c5abc3..d1a011c40ba2 100644
--- a/drivers/gpu/drm/amd/amdgpu/jpeg_v3_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v3_0.c
@@ -34,30 +34,53 @@
 
 #define mmUVD_JPEG_PITCH_INTERNAL_OFFSET	0x401f
 
+static const struct amdgpu_hwip_reg_entry jpeg_reg_list_3_0[] = {
+	SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JPEG_POWER_STATUS),
+	SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JPEG_INT_STAT),
+	SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JRBC_RB_RPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JRBC_RB_WPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JRBC_RB_CNTL),
+	SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JRBC_RB_SIZE),
+	SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JRBC_STATUS),
+	SOC15_REG_ENTRY_STR(JPEG, 0, mmJPEG_DEC_ADDR_MODE),
+	SOC15_REG_ENTRY_STR(JPEG, 0, mmJPEG_DEC_GFX10_ADDR_CONFIG),
+	SOC15_REG_ENTRY_STR(JPEG, 0, mmJPEG_DEC_Y_GFX10_TILING_SURFACE),
+	SOC15_REG_ENTRY_STR(JPEG, 0, mmJPEG_DEC_UV_GFX10_TILING_SURFACE),
+	SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JPEG_PITCH),
+	SOC15_REG_ENTRY_STR(JPEG, 0, mmUVD_JPEG_UV_PITCH),
+};
+
 static void jpeg_v3_0_set_dec_ring_funcs(struct amdgpu_device *adev);
 static void jpeg_v3_0_set_irq_funcs(struct amdgpu_device *adev);
-static int jpeg_v3_0_set_powergating_state(void *handle,
+static int jpeg_v3_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
 				enum amd_powergating_state state);
 
 /**
  * jpeg_v3_0_early_init - set function pointers
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Set ring and irq function pointers
  */
-static int jpeg_v3_0_early_init(void *handle)
+static int jpeg_v3_0_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
-	if (adev->asic_type != CHIP_YELLOW_CARP) {
-		u32 harvest = RREG32_SOC15(JPEG, 0, mmCC_UVD_HARVESTING);
+	u32 harvest;
 
+	switch (amdgpu_ip_version(adev, UVD_HWIP, 0)) {
+	case IP_VERSION(3, 1, 1):
+	case IP_VERSION(3, 1, 2):
+		break;
+	default:
+		harvest = RREG32_SOC15(JPEG, 0, mmCC_UVD_HARVESTING);
 		if (harvest & CC_UVD_HARVESTING__UVD_DISABLE_MASK)
 			return -ENOENT;
+		break;
 	}
 
 	adev->jpeg.num_jpeg_inst = 1;
+	adev->jpeg.num_jpeg_rings = 1;
 
 	jpeg_v3_0_set_dec_ring_funcs(adev);
 	jpeg_v3_0_set_irq_funcs(adev);
@@ -68,13 +91,13 @@ static int jpeg_v3_0_early_init(void *handle)
 /**
  * jpeg_v3_0_sw_init - sw init for JPEG block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Load firmware and sw initialization
  */
-static int jpeg_v3_0_sw_init(void *handle)
+static int jpeg_v3_0_sw_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	struct amdgpu_ring *ring;
 	int r;
 
@@ -92,37 +115,50 @@ static int jpeg_v3_0_sw_init(void *handle)
 	if (r)
 		return r;
 
-	ring = &adev->jpeg.inst->ring_dec;
+	ring = adev->jpeg.inst->ring_dec;
 	ring->use_doorbell = true;
 	ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 1;
+	ring->vm_hub = AMDGPU_MMHUB0(0);
 	sprintf(ring->name, "jpeg_dec");
 	r = amdgpu_ring_init(adev, ring, 512, &adev->jpeg.inst->irq, 0,
 			     AMDGPU_RING_PRIO_DEFAULT, NULL);
 	if (r)
 		return r;
 
-	adev->jpeg.internal.jpeg_pitch = mmUVD_JPEG_PITCH_INTERNAL_OFFSET;
-	adev->jpeg.inst->external.jpeg_pitch = SOC15_REG_OFFSET(JPEG, 0, mmUVD_JPEG_PITCH);
+	adev->jpeg.internal.jpeg_pitch[0] = mmUVD_JPEG_PITCH_INTERNAL_OFFSET;
+	adev->jpeg.inst->external.jpeg_pitch[0] = SOC15_REG_OFFSET(JPEG, 0, mmUVD_JPEG_PITCH);
 
-	return 0;
+	r = amdgpu_jpeg_reg_dump_init(adev, jpeg_reg_list_3_0, ARRAY_SIZE(jpeg_reg_list_3_0));
+	if (r)
+		return r;
+
+	adev->jpeg.supported_reset =
+		amdgpu_get_soft_full_reset_mask(adev->jpeg.inst[0].ring_dec);
+	if (!amdgpu_sriov_vf(adev))
+		adev->jpeg.supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE;
+	r = amdgpu_jpeg_sysfs_reset_mask_init(adev);
+
+	return r;
 }
 
 /**
  * jpeg_v3_0_sw_fini - sw fini for JPEG block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * JPEG suspend and free up sw allocation
  */
-static int jpeg_v3_0_sw_fini(void *handle)
+static int jpeg_v3_0_sw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int r;
 
 	r = amdgpu_jpeg_suspend(adev);
 	if (r)
 		return r;
 
+	amdgpu_jpeg_sysfs_reset_mask_fini(adev);
+
 	r = amdgpu_jpeg_sw_fini(adev);
 
 	return r;
@@ -131,43 +167,36 @@ static int jpeg_v3_0_sw_fini(void *handle)
 /**
  * jpeg_v3_0_hw_init - start and test JPEG block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  */
-static int jpeg_v3_0_hw_init(void *handle)
+static int jpeg_v3_0_hw_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-	struct amdgpu_ring *ring = &adev->jpeg.inst->ring_dec;
-	int r;
+	struct amdgpu_device *adev = ip_block->adev;
+	struct amdgpu_ring *ring = adev->jpeg.inst->ring_dec;
 
 	adev->nbio.funcs->vcn_doorbell_range(adev, ring->use_doorbell,
 		(adev->doorbell_index.vcn.vcn_ring0_1 << 1), 0);
 
-	r = amdgpu_ring_test_helper(ring);
-	if (r)
-		return r;
-
-	DRM_INFO("JPEG decode initialized successfully.\n");
-
-	return 0;
+	return amdgpu_ring_test_helper(ring);
 }
 
 /**
  * jpeg_v3_0_hw_fini - stop the hardware block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Stop the JPEG block, mark ring as not ready any more
  */
-static int jpeg_v3_0_hw_fini(void *handle)
+static int jpeg_v3_0_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
-	cancel_delayed_work_sync(&adev->vcn.idle_work);
+	cancel_delayed_work_sync(&adev->jpeg.idle_work);
 
 	if (adev->jpeg.cur_state != AMD_PG_STATE_GATE &&
 	      RREG32_SOC15(JPEG, 0, mmUVD_JRBC_STATUS))
-		jpeg_v3_0_set_powergating_state(adev, AMD_PG_STATE_GATE);
+		jpeg_v3_0_set_powergating_state(ip_block, AMD_PG_STATE_GATE);
 
 	return 0;
 }
@@ -175,20 +204,19 @@ static int jpeg_v3_0_hw_fini(void *handle)
 /**
  * jpeg_v3_0_suspend - suspend JPEG block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * HW fini and suspend JPEG block
  */
-static int jpeg_v3_0_suspend(void *handle)
+static int jpeg_v3_0_suspend(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 	int r;
 
-	r = jpeg_v3_0_hw_fini(adev);
+	r = jpeg_v3_0_hw_fini(ip_block);
 	if (r)
 		return r;
 
-	r = amdgpu_jpeg_suspend(adev);
+	r = amdgpu_jpeg_suspend(ip_block->adev);
 
 	return r;
 }
@@ -196,20 +224,19 @@ static int jpeg_v3_0_suspend(void *handle)
 /**
  * jpeg_v3_0_resume - resume JPEG block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Resume firmware and hw init JPEG block
  */
-static int jpeg_v3_0_resume(void *handle)
+static int jpeg_v3_0_resume(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 	int r;
 
-	r = amdgpu_jpeg_resume(adev);
+	r = amdgpu_jpeg_resume(ip_block->adev);
 	if (r)
 		return r;
 
-	r = jpeg_v3_0_hw_init(adev);
+	r = jpeg_v3_0_hw_init(ip_block);
 
 	return r;
 }
@@ -323,7 +350,7 @@ static int jpeg_v3_0_enable_static_power_gating(struct amdgpu_device *adev)
  */
 static int jpeg_v3_0_start(struct amdgpu_device *adev)
 {
-	struct amdgpu_ring *ring = &adev->jpeg.inst->ring_dec;
+	struct amdgpu_ring *ring = adev->jpeg.inst->ring_dec;
 	int r;
 
 	if (adev->pm.dpm_enabled)
@@ -422,7 +449,7 @@ static uint64_t jpeg_v3_0_dec_ring_get_wptr(struct amdgpu_ring *ring)
 	struct amdgpu_device *adev = ring->adev;
 
 	if (ring->use_doorbell)
-		return adev->wb.wb[ring->wptr_offs];
+		return *ring->wptr_cpu_addr;
 	else
 		return RREG32_SOC15(JPEG, 0, mmUVD_JRBC_RB_WPTR);
 }
@@ -439,16 +466,16 @@ static void jpeg_v3_0_dec_ring_set_wptr(struct amdgpu_ring *ring)
 	struct amdgpu_device *adev = ring->adev;
 
 	if (ring->use_doorbell) {
-		adev->wb.wb[ring->wptr_offs] = lower_32_bits(ring->wptr);
+		*ring->wptr_cpu_addr = lower_32_bits(ring->wptr);
 		WDOORBELL32(ring->doorbell_index, lower_32_bits(ring->wptr));
 	} else {
 		WREG32_SOC15(JPEG, 0, mmUVD_JRBC_RB_WPTR, lower_32_bits(ring->wptr));
 	}
 }
 
-static bool jpeg_v3_0_is_idle(void *handle)
+static bool jpeg_v3_0_is_idle(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int ret = 1;
 
 	ret &= (((RREG32_SOC15(JPEG, 0, mmUVD_JRBC_STATUS) &
@@ -458,23 +485,23 @@ static bool jpeg_v3_0_is_idle(void *handle)
 	return ret;
 }
 
-static int jpeg_v3_0_wait_for_idle(void *handle)
+static int jpeg_v3_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	return SOC15_WAIT_ON_RREG(JPEG, 0, mmUVD_JRBC_STATUS,
 		UVD_JRBC_STATUS__RB_JOB_DONE_MASK,
 		UVD_JRBC_STATUS__RB_JOB_DONE_MASK);
 }
 
-static int jpeg_v3_0_set_clockgating_state(void *handle,
+static int jpeg_v3_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_clockgating_state state)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-	bool enable = (state == AMD_CG_STATE_GATE) ? true : false;
+	struct amdgpu_device *adev = ip_block->adev;
+	bool enable = state == AMD_CG_STATE_GATE;
 
 	if (enable) {
-		if (!jpeg_v3_0_is_idle(handle))
+		if (!jpeg_v3_0_is_idle(ip_block))
 			return -EBUSY;
 		jpeg_v3_0_enable_clock_gating(adev);
 	} else {
@@ -484,10 +511,10 @@ static int jpeg_v3_0_set_clockgating_state(void *handle,
 	return 0;
 }
 
-static int jpeg_v3_0_set_powergating_state(void *handle,
+static int jpeg_v3_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_powergating_state state)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int ret;
 
 	if(state == adev->jpeg.cur_state)
@@ -520,7 +547,7 @@ static int jpeg_v3_0_process_interrupt(struct amdgpu_device *adev,
 
 	switch (entry->src_id) {
 	case VCN_2_0__SRCID__JPEG_DECODE:
-		amdgpu_fence_process(&adev->jpeg.inst->ring_dec);
+		amdgpu_fence_process(adev->jpeg.inst->ring_dec);
 		break;
 	default:
 		DRM_ERROR("Unhandled interrupt: %d %d\n",
@@ -531,10 +558,25 @@ static int jpeg_v3_0_process_interrupt(struct amdgpu_device *adev,
 	return 0;
 }
 
+static int jpeg_v3_0_ring_reset(struct amdgpu_ring *ring,
+				unsigned int vmid,
+				struct amdgpu_fence *timedout_fence)
+{
+	int r;
+
+	amdgpu_ring_reset_helper_begin(ring, timedout_fence);
+	r = jpeg_v3_0_stop(ring->adev);
+	if (r)
+		return r;
+	r = jpeg_v3_0_start(ring->adev);
+	if (r)
+		return r;
+	return amdgpu_ring_reset_helper_end(ring, timedout_fence);
+}
+
 static const struct amd_ip_funcs jpeg_v3_0_ip_funcs = {
 	.name = "jpeg_v3_0",
 	.early_init = jpeg_v3_0_early_init,
-	.late_init = NULL,
 	.sw_init = jpeg_v3_0_sw_init,
 	.sw_fini = jpeg_v3_0_sw_fini,
 	.hw_init = jpeg_v3_0_hw_init,
@@ -543,21 +585,19 @@ static const struct amd_ip_funcs jpeg_v3_0_ip_funcs = {
 	.resume = jpeg_v3_0_resume,
 	.is_idle = jpeg_v3_0_is_idle,
 	.wait_for_idle = jpeg_v3_0_wait_for_idle,
-	.check_soft_reset = NULL,
-	.pre_soft_reset = NULL,
-	.soft_reset = NULL,
-	.post_soft_reset = NULL,
 	.set_clockgating_state = jpeg_v3_0_set_clockgating_state,
 	.set_powergating_state = jpeg_v3_0_set_powergating_state,
+	.dump_ip_state = amdgpu_jpeg_dump_ip_state,
+	.print_ip_state = amdgpu_jpeg_print_ip_state,
 };
 
 static const struct amdgpu_ring_funcs jpeg_v3_0_dec_ring_vm_funcs = {
 	.type = AMDGPU_RING_TYPE_VCN_JPEG,
 	.align_mask = 0xf,
-	.vmhub = AMDGPU_MMHUB_0,
 	.get_rptr = jpeg_v3_0_dec_ring_get_rptr,
 	.get_wptr = jpeg_v3_0_dec_ring_get_wptr,
 	.set_wptr = jpeg_v3_0_dec_ring_set_wptr,
+	.parse_cs = amdgpu_jpeg_dec_parse_cs,
 	.emit_frame_size =
 		SOC15_FLUSH_GPU_TLB_NUM_WREG * 6 +
 		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 8 +
@@ -579,12 +619,12 @@ static const struct amdgpu_ring_funcs jpeg_v3_0_dec_ring_vm_funcs = {
 	.emit_wreg = jpeg_v2_0_dec_ring_emit_wreg,
 	.emit_reg_wait = jpeg_v2_0_dec_ring_emit_reg_wait,
 	.emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper,
+	.reset = jpeg_v3_0_ring_reset,
 };
 
 static void jpeg_v3_0_set_dec_ring_funcs(struct amdgpu_device *adev)
 {
-	adev->jpeg.inst->ring_dec.funcs = &jpeg_v3_0_dec_ring_vm_funcs;
-	DRM_INFO("JPEG decode is enabled in VM mode\n");
+	adev->jpeg.inst->ring_dec->funcs = &jpeg_v3_0_dec_ring_vm_funcs;
 }
 
 static const struct amdgpu_irq_src_funcs jpeg_v3_0_irq_funcs = {
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0.c
new file mode 100644
index 000000000000..33db2c1ae6cc
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0.c
@@ -0,0 +1,878 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "amdgpu.h"
+#include "amdgpu_jpeg.h"
+#include "amdgpu_pm.h"
+#include "soc15.h"
+#include "soc15d.h"
+#include "jpeg_v2_0.h"
+#include "jpeg_v4_0.h"
+#include "mmsch_v4_0.h"
+
+#include "vcn/vcn_4_0_0_offset.h"
+#include "vcn/vcn_4_0_0_sh_mask.h"
+#include "ivsrcid/vcn/irqsrcs_vcn_4_0.h"
+
+#define regUVD_JPEG_PITCH_INTERNAL_OFFSET                  0x401f
+
+static const struct amdgpu_hwip_reg_entry jpeg_reg_list_4_0[] = {
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JPEG_POWER_STATUS),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JPEG_INT_STAT),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC_RB_RPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC_RB_WPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC_RB_CNTL),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC_RB_SIZE),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC_STATUS),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regJPEG_DEC_ADDR_MODE),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regJPEG_DEC_GFX10_ADDR_CONFIG),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regJPEG_DEC_Y_GFX10_TILING_SURFACE),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regJPEG_DEC_UV_GFX10_TILING_SURFACE),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JPEG_PITCH),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JPEG_UV_PITCH),
+};
+
+static int jpeg_v4_0_start_sriov(struct amdgpu_device *adev);
+static void jpeg_v4_0_set_dec_ring_funcs(struct amdgpu_device *adev);
+static void jpeg_v4_0_set_irq_funcs(struct amdgpu_device *adev);
+static int jpeg_v4_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
+				enum amd_powergating_state state);
+static void jpeg_v4_0_set_ras_funcs(struct amdgpu_device *adev);
+static void jpeg_v4_0_dec_ring_set_wptr(struct amdgpu_ring *ring);
+
+/**
+ * jpeg_v4_0_early_init - set function pointers
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Set ring and irq function pointers
+ */
+static int jpeg_v4_0_early_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+
+	adev->jpeg.num_jpeg_inst = 1;
+	adev->jpeg.num_jpeg_rings = 1;
+
+	jpeg_v4_0_set_dec_ring_funcs(adev);
+	jpeg_v4_0_set_irq_funcs(adev);
+	jpeg_v4_0_set_ras_funcs(adev);
+
+	return 0;
+}
+
+/**
+ * jpeg_v4_0_sw_init - sw init for JPEG block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Load firmware and sw initialization
+ */
+static int jpeg_v4_0_sw_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	struct amdgpu_ring *ring;
+	int r;
+
+	/* JPEG TRAP */
+	r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_VCN,
+		VCN_4_0__SRCID__JPEG_DECODE, &adev->jpeg.inst->irq);
+	if (r)
+		return r;
+
+	/* JPEG DJPEG POISON EVENT */
+	r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_VCN,
+			VCN_4_0__SRCID_DJPEG0_POISON, &adev->jpeg.inst->ras_poison_irq);
+	if (r)
+		return r;
+
+	/* JPEG EJPEG POISON EVENT */
+	r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_VCN,
+			VCN_4_0__SRCID_EJPEG0_POISON, &adev->jpeg.inst->ras_poison_irq);
+	if (r)
+		return r;
+
+	r = amdgpu_jpeg_sw_init(adev);
+	if (r)
+		return r;
+
+	r = amdgpu_jpeg_resume(adev);
+	if (r)
+		return r;
+
+	ring = adev->jpeg.inst->ring_dec;
+	ring->use_doorbell = true;
+	ring->doorbell_index = amdgpu_sriov_vf(adev) ? (((adev->doorbell_index.vcn.vcn_ring0_1) << 1) + 4) : ((adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 1);
+	ring->vm_hub = AMDGPU_MMHUB0(0);
+
+	sprintf(ring->name, "jpeg_dec");
+	r = amdgpu_ring_init(adev, ring, 512, &adev->jpeg.inst->irq, 0,
+			     AMDGPU_RING_PRIO_DEFAULT, NULL);
+	if (r)
+		return r;
+
+	adev->jpeg.internal.jpeg_pitch[0] = regUVD_JPEG_PITCH_INTERNAL_OFFSET;
+	adev->jpeg.inst->external.jpeg_pitch[0] = SOC15_REG_OFFSET(JPEG, 0, regUVD_JPEG_PITCH);
+
+	r = amdgpu_jpeg_ras_sw_init(adev);
+	if (r)
+		return r;
+
+	r = amdgpu_jpeg_reg_dump_init(adev, jpeg_reg_list_4_0, ARRAY_SIZE(jpeg_reg_list_4_0));
+	if (r)
+		return r;
+
+	adev->jpeg.supported_reset =
+		amdgpu_get_soft_full_reset_mask(adev->jpeg.inst[0].ring_dec);
+	if (!amdgpu_sriov_vf(adev))
+		adev->jpeg.supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE;
+	r = amdgpu_jpeg_sysfs_reset_mask_init(adev);
+
+	return r;
+}
+
+/**
+ * jpeg_v4_0_sw_fini - sw fini for JPEG block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * JPEG suspend and free up sw allocation
+ */
+static int jpeg_v4_0_sw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int r;
+
+	r = amdgpu_jpeg_suspend(adev);
+	if (r)
+		return r;
+
+	amdgpu_jpeg_sysfs_reset_mask_fini(adev);
+	r = amdgpu_jpeg_sw_fini(adev);
+
+	return r;
+}
+
+/**
+ * jpeg_v4_0_hw_init - start and test JPEG block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ */
+static int jpeg_v4_0_hw_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	struct amdgpu_ring *ring = adev->jpeg.inst->ring_dec;
+	int r;
+
+	if (amdgpu_sriov_vf(adev)) {
+		r = jpeg_v4_0_start_sriov(adev);
+		if (r)
+			return r;
+		ring->wptr = 0;
+		ring->wptr_old = 0;
+		jpeg_v4_0_dec_ring_set_wptr(ring);
+		ring->sched.ready = true;
+	} else {
+		adev->nbio.funcs->vcn_doorbell_range(adev, ring->use_doorbell,
+						(adev->doorbell_index.vcn.vcn_ring0_1 << 1), 0);
+
+		WREG32_SOC15(VCN, 0, regVCN_JPEG_DB_CTRL,
+			ring->doorbell_index << VCN_JPEG_DB_CTRL__OFFSET__SHIFT |
+			VCN_JPEG_DB_CTRL__EN_MASK);
+
+		r = amdgpu_ring_test_helper(ring);
+		if (r)
+			return r;
+	}
+
+	return 0;
+}
+
+/**
+ * jpeg_v4_0_hw_fini - stop the hardware block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Stop the JPEG block, mark ring as not ready any more
+ */
+static int jpeg_v4_0_hw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	cancel_delayed_work_sync(&adev->jpeg.idle_work);
+	if (!amdgpu_sriov_vf(adev)) {
+		if (adev->jpeg.cur_state != AMD_PG_STATE_GATE &&
+			RREG32_SOC15(JPEG, 0, regUVD_JRBC_STATUS))
+			jpeg_v4_0_set_powergating_state(ip_block, AMD_PG_STATE_GATE);
+	}
+	if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__JPEG))
+		amdgpu_irq_put(adev, &adev->jpeg.inst->ras_poison_irq, 0);
+
+	return 0;
+}
+
+/**
+ * jpeg_v4_0_suspend - suspend JPEG block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * HW fini and suspend JPEG block
+ */
+static int jpeg_v4_0_suspend(struct amdgpu_ip_block *ip_block)
+{
+	int r;
+
+	r = jpeg_v4_0_hw_fini(ip_block);
+	if (r)
+		return r;
+
+	r = amdgpu_jpeg_suspend(ip_block->adev);
+
+	return r;
+}
+
+/**
+ * jpeg_v4_0_resume - resume JPEG block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Resume firmware and hw init JPEG block
+ */
+static int jpeg_v4_0_resume(struct amdgpu_ip_block *ip_block)
+{
+	int r;
+
+	r = amdgpu_jpeg_resume(ip_block->adev);
+	if (r)
+		return r;
+
+	r = jpeg_v4_0_hw_init(ip_block);
+
+	return r;
+}
+
+static void jpeg_v4_0_disable_clock_gating(struct amdgpu_device *adev)
+{
+	uint32_t data = 0;
+
+	data = RREG32_SOC15(JPEG, 0, regJPEG_CGC_CTRL);
+	if (adev->cg_flags & AMD_CG_SUPPORT_JPEG_MGCG) {
+		data |= 1 << JPEG_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
+		data &= (~JPEG_CGC_CTRL__JPEG_DEC_MODE_MASK);
+	} else {
+		data &= ~JPEG_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
+	}
+
+	data |= 1 << JPEG_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT;
+	data |= 4 << JPEG_CGC_CTRL__CLK_OFF_DELAY__SHIFT;
+	WREG32_SOC15(JPEG, 0, regJPEG_CGC_CTRL, data);
+
+	data = RREG32_SOC15(JPEG, 0, regJPEG_CGC_GATE);
+	data &= ~(JPEG_CGC_GATE__JPEG_DEC_MASK
+		| JPEG_CGC_GATE__JPEG2_DEC_MASK
+		| JPEG_CGC_GATE__JMCIF_MASK
+		| JPEG_CGC_GATE__JRBBM_MASK);
+	WREG32_SOC15(JPEG, 0, regJPEG_CGC_GATE, data);
+}
+
+static void jpeg_v4_0_enable_clock_gating(struct amdgpu_device *adev)
+{
+	uint32_t data = 0;
+
+	data = RREG32_SOC15(JPEG, 0, regJPEG_CGC_CTRL);
+	if (adev->cg_flags & AMD_CG_SUPPORT_JPEG_MGCG) {
+		data |= 1 << JPEG_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
+		data |= JPEG_CGC_CTRL__JPEG_DEC_MODE_MASK;
+	} else {
+		data &= ~JPEG_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
+	}
+
+	data |= 1 << JPEG_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT;
+	data |= 4 << JPEG_CGC_CTRL__CLK_OFF_DELAY__SHIFT;
+	WREG32_SOC15(JPEG, 0, regJPEG_CGC_CTRL, data);
+
+	data = RREG32_SOC15(JPEG, 0, regJPEG_CGC_GATE);
+	data |= (JPEG_CGC_GATE__JPEG_DEC_MASK
+		|JPEG_CGC_GATE__JPEG2_DEC_MASK
+		|JPEG_CGC_GATE__JMCIF_MASK
+		|JPEG_CGC_GATE__JRBBM_MASK);
+	WREG32_SOC15(JPEG, 0, regJPEG_CGC_GATE, data);
+}
+
+static int jpeg_v4_0_disable_static_power_gating(struct amdgpu_device *adev)
+{
+	if (adev->pg_flags & AMD_PG_SUPPORT_JPEG) {
+		uint32_t data = 0;
+		int r = 0;
+
+		data = 1 << UVD_PGFSM_CONFIG__UVDJ_PWR_CONFIG__SHIFT;
+		WREG32(SOC15_REG_OFFSET(JPEG, 0, regUVD_PGFSM_CONFIG), data);
+
+		r = SOC15_WAIT_ON_RREG(JPEG, 0,
+			regUVD_PGFSM_STATUS, UVD_PGFSM_STATUS_UVDJ_PWR_ON,
+			UVD_PGFSM_STATUS__UVDJ_PWR_STATUS_MASK);
+
+		if (r) {
+			DRM_DEV_ERROR(adev->dev, "amdgpu: JPEG disable power gating failed\n");
+			return r;
+		}
+	}
+
+	/* disable anti hang mechanism */
+	WREG32_P(SOC15_REG_OFFSET(JPEG, 0, regUVD_JPEG_POWER_STATUS), 0,
+		~UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK);
+
+	/* keep the JPEG in static PG mode */
+	WREG32_P(SOC15_REG_OFFSET(JPEG, 0, regUVD_JPEG_POWER_STATUS), 0,
+		~UVD_JPEG_POWER_STATUS__JPEG_PG_MODE_MASK);
+
+	return 0;
+}
+
+static int jpeg_v4_0_enable_static_power_gating(struct amdgpu_device *adev)
+{
+	/* enable anti hang mechanism */
+	WREG32_P(SOC15_REG_OFFSET(JPEG, 0, regUVD_JPEG_POWER_STATUS),
+		UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK,
+		~UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK);
+
+	if (adev->pg_flags & AMD_PG_SUPPORT_JPEG) {
+		uint32_t data = 0;
+		int r = 0;
+
+		data = 2 << UVD_PGFSM_CONFIG__UVDJ_PWR_CONFIG__SHIFT;
+		WREG32(SOC15_REG_OFFSET(JPEG, 0, regUVD_PGFSM_CONFIG), data);
+
+		r = SOC15_WAIT_ON_RREG(JPEG, 0, regUVD_PGFSM_STATUS,
+			(2 << UVD_PGFSM_STATUS__UVDJ_PWR_STATUS__SHIFT),
+			UVD_PGFSM_STATUS__UVDJ_PWR_STATUS_MASK);
+
+		if (r) {
+			DRM_DEV_ERROR(adev->dev, "amdgpu: JPEG enable power gating failed\n");
+			return r;
+		}
+	}
+
+	return 0;
+}
+
+/**
+ * jpeg_v4_0_start - start JPEG block
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Setup and start the JPEG block
+ */
+static int jpeg_v4_0_start(struct amdgpu_device *adev)
+{
+	struct amdgpu_ring *ring = adev->jpeg.inst->ring_dec;
+	int r;
+
+	if (adev->pm.dpm_enabled)
+		amdgpu_dpm_enable_jpeg(adev, true);
+
+	/* disable power gating */
+	r = jpeg_v4_0_disable_static_power_gating(adev);
+	if (r)
+		return r;
+
+	/* JPEG disable CGC */
+	jpeg_v4_0_disable_clock_gating(adev);
+
+	/* MJPEG global tiling registers */
+	WREG32_SOC15(JPEG, 0, regJPEG_DEC_GFX10_ADDR_CONFIG,
+		adev->gfx.config.gb_addr_config);
+
+
+	/* enable JMI channel */
+	WREG32_P(SOC15_REG_OFFSET(JPEG, 0, regUVD_JMI_CNTL), 0,
+		~UVD_JMI_CNTL__SOFT_RESET_MASK);
+
+	/* enable System Interrupt for JRBC */
+	WREG32_P(SOC15_REG_OFFSET(JPEG, 0, regJPEG_SYS_INT_EN),
+		JPEG_SYS_INT_EN__DJRBC_MASK,
+		~JPEG_SYS_INT_EN__DJRBC_MASK);
+
+	WREG32_SOC15(JPEG, 0, regUVD_LMI_JRBC_RB_VMID, 0);
+	WREG32_SOC15(JPEG, 0, regUVD_JRBC_RB_CNTL, (0x00000001L | 0x00000002L));
+	WREG32_SOC15(JPEG, 0, regUVD_LMI_JRBC_RB_64BIT_BAR_LOW,
+		lower_32_bits(ring->gpu_addr));
+	WREG32_SOC15(JPEG, 0, regUVD_LMI_JRBC_RB_64BIT_BAR_HIGH,
+		upper_32_bits(ring->gpu_addr));
+	WREG32_SOC15(JPEG, 0, regUVD_JRBC_RB_RPTR, 0);
+	WREG32_SOC15(JPEG, 0, regUVD_JRBC_RB_WPTR, 0);
+	WREG32_SOC15(JPEG, 0, regUVD_JRBC_RB_CNTL, 0x00000002L);
+	WREG32_SOC15(JPEG, 0, regUVD_JRBC_RB_SIZE, ring->ring_size / 4);
+	ring->wptr = RREG32_SOC15(JPEG, 0, regUVD_JRBC_RB_WPTR);
+
+	return 0;
+}
+
+static int jpeg_v4_0_start_sriov(struct amdgpu_device *adev)
+{
+	struct amdgpu_ring *ring;
+	uint64_t ctx_addr;
+	uint32_t param, resp, expected;
+	uint32_t tmp, timeout;
+
+	struct amdgpu_mm_table *table = &adev->virt.mm_table;
+	uint32_t *table_loc;
+	uint32_t table_size;
+	uint32_t size, size_dw;
+	uint32_t init_status;
+
+	struct mmsch_v4_0_cmd_direct_write
+		direct_wt = { {0} };
+	struct mmsch_v4_0_cmd_end end = { {0} };
+	struct mmsch_v4_0_init_header header;
+
+	direct_wt.cmd_header.command_type =
+		MMSCH_COMMAND__DIRECT_REG_WRITE;
+	end.cmd_header.command_type =
+		MMSCH_COMMAND__END;
+
+	size = sizeof(struct mmsch_v4_0_init_header);
+	table_loc = (uint32_t *)table->cpu_addr;
+	memcpy(&header, (void *)table_loc, size);
+
+	header.version = MMSCH_VERSION;
+	header.total_size = RREG32_SOC15(VCN, 0, regMMSCH_VF_CTX_SIZE);
+
+	header.jpegdec.init_status = 0;
+	header.jpegdec.table_offset = 0;
+	header.jpegdec.table_size = 0;
+
+	table_loc = (uint32_t *)table->cpu_addr;
+	table_loc += header.total_size;
+
+	table_size = 0;
+
+	ring = adev->jpeg.inst->ring_dec;
+
+	MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(JPEG, 0,
+		regUVD_LMI_JRBC_RB_64BIT_BAR_LOW),
+		lower_32_bits(ring->gpu_addr));
+	MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(JPEG, 0,
+		regUVD_LMI_JRBC_RB_64BIT_BAR_HIGH),
+		upper_32_bits(ring->gpu_addr));
+	MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(JPEG, 0,
+		regUVD_JRBC_RB_SIZE), ring->ring_size / 4);
+
+	/* add end packet */
+	MMSCH_V4_0_INSERT_END();
+
+	/* refine header */
+	header.jpegdec.init_status = 0;
+	header.jpegdec.table_offset = header.total_size;
+	header.jpegdec.table_size = table_size;
+	header.total_size += table_size;
+
+	/* Update init table header in memory */
+	size = sizeof(struct mmsch_v4_0_init_header);
+	table_loc = (uint32_t *)table->cpu_addr;
+	memcpy((void *)table_loc, &header, size);
+
+	/* Perform HDP flush before writing to MMSCH registers */
+	amdgpu_device_flush_hdp(adev, NULL);
+
+	/* message MMSCH (in VCN[0]) to initialize this client
+	 * 1, write to mmsch_vf_ctx_addr_lo/hi register with GPU mc addr
+	 * of memory descriptor location
+	 */
+	ctx_addr = table->gpu_addr;
+	WREG32_SOC15(VCN, 0, regMMSCH_VF_CTX_ADDR_LO, lower_32_bits(ctx_addr));
+	WREG32_SOC15(VCN, 0, regMMSCH_VF_CTX_ADDR_HI, upper_32_bits(ctx_addr));
+
+	/* 2, update vmid of descriptor */
+	tmp = RREG32_SOC15(VCN, 0, regMMSCH_VF_VMID);
+	tmp &= ~MMSCH_VF_VMID__VF_CTX_VMID_MASK;
+	/* use domain0 for MM scheduler */
+	tmp |= (0 << MMSCH_VF_VMID__VF_CTX_VMID__SHIFT);
+	WREG32_SOC15(VCN, 0, regMMSCH_VF_VMID, tmp);
+
+	/* 3, notify mmsch about the size of this descriptor */
+	size = header.total_size;
+	WREG32_SOC15(VCN, 0, regMMSCH_VF_CTX_SIZE, size);
+
+	/* 4, set resp to zero */
+	WREG32_SOC15(VCN, 0, regMMSCH_VF_MAILBOX_RESP, 0);
+
+	/* 5, kick off the initialization and wait until
+	 * MMSCH_VF_MAILBOX_RESP becomes non-zero
+	 */
+	param = 0x00000001;
+	WREG32_SOC15(VCN, 0, regMMSCH_VF_MAILBOX_HOST, param);
+	tmp = 0;
+	timeout = 1000;
+	resp = 0;
+	expected = MMSCH_VF_MAILBOX_RESP__OK;
+	init_status = ((struct mmsch_v4_0_init_header *)(table_loc))->jpegdec.init_status;
+	while (resp != expected) {
+		resp = RREG32_SOC15(VCN, 0, regMMSCH_VF_MAILBOX_RESP);
+
+		if (resp != 0)
+			break;
+		udelay(10);
+		tmp = tmp + 10;
+		if (tmp >= timeout) {
+			DRM_ERROR("failed to init MMSCH. TIME-OUT after %d usec"\
+				" waiting for regMMSCH_VF_MAILBOX_RESP "\
+				"(expected=0x%08x, readback=0x%08x)\n",
+				tmp, expected, resp);
+			return -EBUSY;
+		}
+	}
+	if (resp != expected && resp != MMSCH_VF_MAILBOX_RESP__INCOMPLETE
+			&& init_status != MMSCH_VF_ENGINE_STATUS__PASS) {
+		DRM_ERROR("MMSCH init status is incorrect! readback=0x%08x, header init status for jpeg: %x\n", resp, init_status);
+		return -EINVAL;
+	}
+
+	return 0;
+
+}
+
+/**
+ * jpeg_v4_0_stop - stop JPEG block
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * stop the JPEG block
+ */
+static int jpeg_v4_0_stop(struct amdgpu_device *adev)
+{
+	int r;
+
+	/* reset JMI */
+	WREG32_P(SOC15_REG_OFFSET(JPEG, 0, regUVD_JMI_CNTL),
+		UVD_JMI_CNTL__SOFT_RESET_MASK,
+		~UVD_JMI_CNTL__SOFT_RESET_MASK);
+
+	jpeg_v4_0_enable_clock_gating(adev);
+
+	/* enable power gating */
+	r = jpeg_v4_0_enable_static_power_gating(adev);
+	if (r)
+		return r;
+
+	if (adev->pm.dpm_enabled)
+		amdgpu_dpm_enable_jpeg(adev, false);
+
+	return 0;
+}
+
+/**
+ * jpeg_v4_0_dec_ring_get_rptr - get read pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Returns the current hardware read pointer
+ */
+static uint64_t jpeg_v4_0_dec_ring_get_rptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	return RREG32_SOC15(JPEG, 0, regUVD_JRBC_RB_RPTR);
+}
+
+/**
+ * jpeg_v4_0_dec_ring_get_wptr - get write pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Returns the current hardware write pointer
+ */
+static uint64_t jpeg_v4_0_dec_ring_get_wptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	if (ring->use_doorbell)
+		return *ring->wptr_cpu_addr;
+	else
+		return RREG32_SOC15(JPEG, 0, regUVD_JRBC_RB_WPTR);
+}
+
+/**
+ * jpeg_v4_0_dec_ring_set_wptr - set write pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Commits the write pointer to the hardware
+ */
+static void jpeg_v4_0_dec_ring_set_wptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	if (ring->use_doorbell) {
+		*ring->wptr_cpu_addr = lower_32_bits(ring->wptr);
+		WDOORBELL32(ring->doorbell_index, lower_32_bits(ring->wptr));
+	} else {
+		WREG32_SOC15(JPEG, 0, regUVD_JRBC_RB_WPTR, lower_32_bits(ring->wptr));
+	}
+}
+
+static bool jpeg_v4_0_is_idle(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int ret = 1;
+
+	ret &= (((RREG32_SOC15(JPEG, 0, regUVD_JRBC_STATUS) &
+		UVD_JRBC_STATUS__RB_JOB_DONE_MASK) ==
+		UVD_JRBC_STATUS__RB_JOB_DONE_MASK));
+
+	return ret;
+}
+
+static int jpeg_v4_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	return SOC15_WAIT_ON_RREG(JPEG, 0, regUVD_JRBC_STATUS,
+		UVD_JRBC_STATUS__RB_JOB_DONE_MASK,
+		UVD_JRBC_STATUS__RB_JOB_DONE_MASK);
+}
+
+static int jpeg_v4_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
+					  enum amd_clockgating_state state)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	bool enable = state == AMD_CG_STATE_GATE;
+
+	if (enable) {
+		if (!jpeg_v4_0_is_idle(ip_block))
+			return -EBUSY;
+		jpeg_v4_0_enable_clock_gating(adev);
+	} else {
+		jpeg_v4_0_disable_clock_gating(adev);
+	}
+
+	return 0;
+}
+
+static int jpeg_v4_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
+					  enum amd_powergating_state state)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int ret;
+
+	if (amdgpu_sriov_vf(adev)) {
+		adev->jpeg.cur_state = AMD_PG_STATE_UNGATE;
+		return 0;
+	}
+
+	if (state == adev->jpeg.cur_state)
+		return 0;
+
+	if (state == AMD_PG_STATE_GATE)
+		ret = jpeg_v4_0_stop(adev);
+	else
+		ret = jpeg_v4_0_start(adev);
+
+	if (!ret)
+		adev->jpeg.cur_state = state;
+
+	return ret;
+}
+
+static int jpeg_v4_0_set_ras_interrupt_state(struct amdgpu_device *adev,
+					struct amdgpu_irq_src *source,
+					unsigned int type,
+					enum amdgpu_interrupt_state state)
+{
+	return 0;
+}
+
+static int jpeg_v4_0_process_interrupt(struct amdgpu_device *adev,
+				      struct amdgpu_irq_src *source,
+				      struct amdgpu_iv_entry *entry)
+{
+	DRM_DEBUG("IH: JPEG TRAP\n");
+
+	switch (entry->src_id) {
+	case VCN_4_0__SRCID__JPEG_DECODE:
+		amdgpu_fence_process(adev->jpeg.inst->ring_dec);
+		break;
+	default:
+		DRM_DEV_ERROR(adev->dev, "Unhandled interrupt: %d %d\n",
+			  entry->src_id, entry->src_data[0]);
+		break;
+	}
+
+	return 0;
+}
+
+static int jpeg_v4_0_ring_reset(struct amdgpu_ring *ring,
+				unsigned int vmid,
+				struct amdgpu_fence *timedout_fence)
+{
+	int r;
+
+	amdgpu_ring_reset_helper_begin(ring, timedout_fence);
+	r = jpeg_v4_0_stop(ring->adev);
+	if (r)
+		return r;
+	r = jpeg_v4_0_start(ring->adev);
+	if (r)
+		return r;
+	return amdgpu_ring_reset_helper_end(ring, timedout_fence);
+}
+
+static const struct amd_ip_funcs jpeg_v4_0_ip_funcs = {
+	.name = "jpeg_v4_0",
+	.early_init = jpeg_v4_0_early_init,
+	.sw_init = jpeg_v4_0_sw_init,
+	.sw_fini = jpeg_v4_0_sw_fini,
+	.hw_init = jpeg_v4_0_hw_init,
+	.hw_fini = jpeg_v4_0_hw_fini,
+	.suspend = jpeg_v4_0_suspend,
+	.resume = jpeg_v4_0_resume,
+	.is_idle = jpeg_v4_0_is_idle,
+	.wait_for_idle = jpeg_v4_0_wait_for_idle,
+	.set_clockgating_state = jpeg_v4_0_set_clockgating_state,
+	.set_powergating_state = jpeg_v4_0_set_powergating_state,
+	.dump_ip_state = amdgpu_jpeg_dump_ip_state,
+	.print_ip_state = amdgpu_jpeg_print_ip_state,
+};
+
+static const struct amdgpu_ring_funcs jpeg_v4_0_dec_ring_vm_funcs = {
+	.type = AMDGPU_RING_TYPE_VCN_JPEG,
+	.align_mask = 0xf,
+	.get_rptr = jpeg_v4_0_dec_ring_get_rptr,
+	.get_wptr = jpeg_v4_0_dec_ring_get_wptr,
+	.set_wptr = jpeg_v4_0_dec_ring_set_wptr,
+	.parse_cs = amdgpu_jpeg_dec_parse_cs,
+	.emit_frame_size =
+		SOC15_FLUSH_GPU_TLB_NUM_WREG * 6 +
+		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 8 +
+		8 + /* jpeg_v4_0_dec_ring_emit_vm_flush */
+		18 + 18 + /* jpeg_v4_0_dec_ring_emit_fence x2 vm fence */
+		8 + 16,
+	.emit_ib_size = 22, /* jpeg_v4_0_dec_ring_emit_ib */
+	.emit_ib = jpeg_v2_0_dec_ring_emit_ib,
+	.emit_fence = jpeg_v2_0_dec_ring_emit_fence,
+	.emit_vm_flush = jpeg_v2_0_dec_ring_emit_vm_flush,
+	.test_ring = amdgpu_jpeg_dec_ring_test_ring,
+	.test_ib = amdgpu_jpeg_dec_ring_test_ib,
+	.insert_nop = jpeg_v2_0_dec_ring_nop,
+	.insert_start = jpeg_v2_0_dec_ring_insert_start,
+	.insert_end = jpeg_v2_0_dec_ring_insert_end,
+	.pad_ib = amdgpu_ring_generic_pad_ib,
+	.begin_use = amdgpu_jpeg_ring_begin_use,
+	.end_use = amdgpu_jpeg_ring_end_use,
+	.emit_wreg = jpeg_v2_0_dec_ring_emit_wreg,
+	.emit_reg_wait = jpeg_v2_0_dec_ring_emit_reg_wait,
+	.emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper,
+	.reset = jpeg_v4_0_ring_reset,
+};
+
+static void jpeg_v4_0_set_dec_ring_funcs(struct amdgpu_device *adev)
+{
+	adev->jpeg.inst->ring_dec->funcs = &jpeg_v4_0_dec_ring_vm_funcs;
+}
+
+static const struct amdgpu_irq_src_funcs jpeg_v4_0_irq_funcs = {
+	.process = jpeg_v4_0_process_interrupt,
+};
+
+static const struct amdgpu_irq_src_funcs jpeg_v4_0_ras_irq_funcs = {
+	.set = jpeg_v4_0_set_ras_interrupt_state,
+	.process = amdgpu_jpeg_process_poison_irq,
+};
+
+static void jpeg_v4_0_set_irq_funcs(struct amdgpu_device *adev)
+{
+	adev->jpeg.inst->irq.num_types = 1;
+	adev->jpeg.inst->irq.funcs = &jpeg_v4_0_irq_funcs;
+
+	adev->jpeg.inst->ras_poison_irq.num_types = 1;
+	adev->jpeg.inst->ras_poison_irq.funcs = &jpeg_v4_0_ras_irq_funcs;
+}
+
+const struct amdgpu_ip_block_version jpeg_v4_0_ip_block = {
+	.type = AMD_IP_BLOCK_TYPE_JPEG,
+	.major = 4,
+	.minor = 0,
+	.rev = 0,
+	.funcs = &jpeg_v4_0_ip_funcs,
+};
+
+static uint32_t jpeg_v4_0_query_poison_by_instance(struct amdgpu_device *adev,
+		uint32_t instance, uint32_t sub_block)
+{
+	uint32_t poison_stat = 0, reg_value = 0;
+
+	switch (sub_block) {
+	case AMDGPU_JPEG_V4_0_JPEG0:
+		reg_value = RREG32_SOC15(JPEG, instance, regUVD_RAS_JPEG0_STATUS);
+		poison_stat = REG_GET_FIELD(reg_value, UVD_RAS_JPEG0_STATUS, POISONED_PF);
+		break;
+	case AMDGPU_JPEG_V4_0_JPEG1:
+		reg_value = RREG32_SOC15(JPEG, instance, regUVD_RAS_JPEG1_STATUS);
+		poison_stat = REG_GET_FIELD(reg_value, UVD_RAS_JPEG1_STATUS, POISONED_PF);
+		break;
+	default:
+		break;
+	}
+
+	if (poison_stat)
+		dev_info(adev->dev, "Poison detected in JPEG%d sub_block%d\n",
+			instance, sub_block);
+
+	return poison_stat;
+}
+
+static bool jpeg_v4_0_query_ras_poison_status(struct amdgpu_device *adev)
+{
+	uint32_t inst = 0, sub = 0, poison_stat = 0;
+
+	for (inst = 0; inst < adev->jpeg.num_jpeg_inst; inst++)
+		for (sub = 0; sub < AMDGPU_JPEG_V4_0_MAX_SUB_BLOCK; sub++)
+			poison_stat +=
+				jpeg_v4_0_query_poison_by_instance(adev, inst, sub);
+
+	return !!poison_stat;
+}
+
+const struct amdgpu_ras_block_hw_ops jpeg_v4_0_ras_hw_ops = {
+	.query_poison_status = jpeg_v4_0_query_ras_poison_status,
+};
+
+static struct amdgpu_jpeg_ras jpeg_v4_0_ras = {
+	.ras_block = {
+		.hw_ops = &jpeg_v4_0_ras_hw_ops,
+		.ras_late_init = amdgpu_jpeg_ras_late_init,
+	},
+};
+
+static void jpeg_v4_0_set_ras_funcs(struct amdgpu_device *adev)
+{
+	switch (amdgpu_ip_version(adev, JPEG_HWIP, 0)) {
+	case IP_VERSION(4, 0, 0):
+		adev->jpeg.ras = &jpeg_v4_0_ras;
+		break;
+	default:
+		break;
+	}
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0.h b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0.h
new file mode 100644
index 000000000000..47638fd4d4e2
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0.h
@@ -0,0 +1,35 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __JPEG_V4_0_H__
+#define __JPEG_V4_0_H__
+
+enum amdgpu_jpeg_v4_0_sub_block {
+	AMDGPU_JPEG_V4_0_JPEG0 = 0,
+	AMDGPU_JPEG_V4_0_JPEG1,
+
+	AMDGPU_JPEG_V4_0_MAX_SUB_BLOCK,
+};
+
+extern const struct amdgpu_ip_block_version jpeg_v4_0_ip_block;
+#endif /* __JPEG_V4_0_H__ */
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_3.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_3.c
new file mode 100644
index 000000000000..aae7328973d1
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_3.c
@@ -0,0 +1,1486 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "amdgpu.h"
+#include "amdgpu_jpeg.h"
+#include "soc15.h"
+#include "soc15d.h"
+#include "jpeg_v2_0.h"
+#include "jpeg_v4_0_3.h"
+#include "mmsch_v4_0_3.h"
+
+#include "vcn/vcn_4_0_3_offset.h"
+#include "vcn/vcn_4_0_3_sh_mask.h"
+#include "ivsrcid/vcn/irqsrcs_vcn_4_0.h"
+
+#define NORMALIZE_JPEG_REG_OFFSET(offset) \
+		(offset & 0x1FFFF)
+
+enum jpeg_engin_status {
+	UVD_PGFSM_STATUS__UVDJ_PWR_ON  = 0,
+	UVD_PGFSM_STATUS__UVDJ_PWR_OFF = 2,
+};
+
+static void jpeg_v4_0_3_set_dec_ring_funcs(struct amdgpu_device *adev);
+static void jpeg_v4_0_3_set_irq_funcs(struct amdgpu_device *adev);
+static int jpeg_v4_0_3_set_powergating_state(struct amdgpu_ip_block *ip_block,
+				enum amd_powergating_state state);
+static void jpeg_v4_0_3_set_ras_funcs(struct amdgpu_device *adev);
+static void jpeg_v4_0_3_dec_ring_set_wptr(struct amdgpu_ring *ring);
+
+static int amdgpu_ih_srcid_jpeg[] = {
+	VCN_4_0__SRCID__JPEG_DECODE,
+	VCN_4_0__SRCID__JPEG1_DECODE,
+	VCN_4_0__SRCID__JPEG2_DECODE,
+	VCN_4_0__SRCID__JPEG3_DECODE,
+	VCN_4_0__SRCID__JPEG4_DECODE,
+	VCN_4_0__SRCID__JPEG5_DECODE,
+	VCN_4_0__SRCID__JPEG6_DECODE,
+	VCN_4_0__SRCID__JPEG7_DECODE
+};
+
+static const struct amdgpu_hwip_reg_entry jpeg_reg_list_4_0_3[] = {
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JPEG_POWER_STATUS),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JPEG_INT_STAT),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regJPEG_SYS_INT_STATUS),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC0_UVD_JRBC_RB_RPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC0_UVD_JRBC_RB_WPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC0_UVD_JRBC_STATUS),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regJPEG_DEC_ADDR_MODE),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regJPEG_DEC_GFX10_ADDR_CONFIG),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regJPEG_DEC_Y_GFX10_TILING_SURFACE),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regJPEG_DEC_UV_GFX10_TILING_SURFACE),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JPEG_PITCH),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JPEG_UV_PITCH),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC1_UVD_JRBC_RB_RPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC1_UVD_JRBC_RB_WPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC1_UVD_JRBC_STATUS),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC2_UVD_JRBC_RB_RPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC2_UVD_JRBC_RB_WPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC2_UVD_JRBC_STATUS),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC3_UVD_JRBC_RB_RPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC3_UVD_JRBC_RB_WPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC3_UVD_JRBC_STATUS),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC4_UVD_JRBC_RB_RPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC4_UVD_JRBC_RB_WPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC4_UVD_JRBC_STATUS),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC5_UVD_JRBC_RB_RPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC5_UVD_JRBC_RB_WPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC5_UVD_JRBC_STATUS),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC6_UVD_JRBC_RB_RPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC6_UVD_JRBC_RB_WPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC6_UVD_JRBC_STATUS),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC7_UVD_JRBC_RB_RPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC7_UVD_JRBC_RB_WPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC7_UVD_JRBC_STATUS),
+};
+
+static inline bool jpeg_v4_0_3_normalizn_reqd(struct amdgpu_device *adev)
+{
+	return (adev->jpeg.caps & AMDGPU_JPEG_CAPS(RRMT_ENABLED)) == 0;
+}
+
+static inline int jpeg_v4_0_3_core_reg_offset(u32 pipe)
+{
+	if (pipe)
+		return ((0x40 * pipe) - 0xc80);
+	else
+		return 0;
+}
+
+/**
+ * jpeg_v4_0_3_early_init - set function pointers
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Set ring and irq function pointers
+ */
+static int jpeg_v4_0_3_early_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	adev->jpeg.num_jpeg_rings = AMDGPU_MAX_JPEG_RINGS_4_0_3;
+
+	jpeg_v4_0_3_set_dec_ring_funcs(adev);
+	jpeg_v4_0_3_set_irq_funcs(adev);
+	jpeg_v4_0_3_set_ras_funcs(adev);
+
+	return 0;
+}
+
+/**
+ * jpeg_v4_0_3_sw_init - sw init for JPEG block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Load firmware and sw initialization
+ */
+static int jpeg_v4_0_3_sw_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	struct amdgpu_ring *ring;
+	int i, j, r, jpeg_inst;
+
+	for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j) {
+		/* JPEG TRAP */
+		r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_VCN,
+				amdgpu_ih_srcid_jpeg[j], &adev->jpeg.inst->irq);
+		if (r)
+			return r;
+	}
+
+	/* JPEG DJPEG POISON EVENT */
+	r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_VCN,
+			VCN_4_0__SRCID_DJPEG0_POISON, &adev->jpeg.inst->ras_poison_irq);
+	if (r)
+		return r;
+
+	/* JPEG EJPEG POISON EVENT */
+	r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_VCN,
+			VCN_4_0__SRCID_EJPEG0_POISON, &adev->jpeg.inst->ras_poison_irq);
+	if (r)
+		return r;
+
+	r = amdgpu_jpeg_sw_init(adev);
+	if (r)
+		return r;
+
+	r = amdgpu_jpeg_resume(adev);
+	if (r)
+		return r;
+
+	for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
+		jpeg_inst = GET_INST(JPEG, i);
+
+		for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j) {
+			ring = &adev->jpeg.inst[i].ring_dec[j];
+			ring->use_doorbell = true;
+			ring->vm_hub = AMDGPU_MMHUB0(adev->jpeg.inst[i].aid_id);
+			if (!amdgpu_sriov_vf(adev)) {
+				ring->doorbell_index =
+					(adev->doorbell_index.vcn.vcn_ring0_1 << 1) +
+					1 + j + 9 * jpeg_inst;
+			} else {
+				if (j < 4)
+					ring->doorbell_index =
+						(adev->doorbell_index.vcn.vcn_ring0_1 << 1) +
+						4 + j + 32 * jpeg_inst;
+				else
+					ring->doorbell_index =
+						(adev->doorbell_index.vcn.vcn_ring0_1 << 1) +
+						8 + j + 32 * jpeg_inst;
+			}
+			sprintf(ring->name, "jpeg_dec_%d.%d", adev->jpeg.inst[i].aid_id, j);
+			r = amdgpu_ring_init(adev, ring, 512, &adev->jpeg.inst->irq, 0,
+						AMDGPU_RING_PRIO_DEFAULT, NULL);
+			if (r)
+				return r;
+
+			adev->jpeg.internal.jpeg_pitch[j] =
+				regUVD_JRBC0_UVD_JRBC_SCRATCH0_INTERNAL_OFFSET;
+			adev->jpeg.inst[i].external.jpeg_pitch[j] =
+				SOC15_REG_OFFSET1(JPEG, jpeg_inst, regUVD_JRBC0_UVD_JRBC_SCRATCH0,
+						  jpeg_v4_0_3_core_reg_offset(j));
+		}
+	}
+
+	if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__JPEG)) {
+		r = amdgpu_jpeg_ras_sw_init(adev);
+		if (r) {
+			dev_err(adev->dev, "Failed to initialize jpeg ras block!\n");
+			return r;
+		}
+	}
+
+	r = amdgpu_jpeg_reg_dump_init(adev, jpeg_reg_list_4_0_3, ARRAY_SIZE(jpeg_reg_list_4_0_3));
+	if (r)
+		return r;
+
+	adev->jpeg.supported_reset =
+		amdgpu_get_soft_full_reset_mask(adev->jpeg.inst[0].ring_dec);
+	if (!amdgpu_sriov_vf(adev))
+		adev->jpeg.supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE;
+	r = amdgpu_jpeg_sysfs_reset_mask_init(adev);
+
+	return 0;
+}
+
+/**
+ * jpeg_v4_0_3_sw_fini - sw fini for JPEG block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * JPEG suspend and free up sw allocation
+ */
+static int jpeg_v4_0_3_sw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int r;
+
+	r = amdgpu_jpeg_suspend(adev);
+	if (r)
+		return r;
+
+	amdgpu_jpeg_sysfs_reset_mask_fini(adev);
+
+	r = amdgpu_jpeg_sw_fini(adev);
+
+	return r;
+}
+
+static int jpeg_v4_0_3_start_sriov(struct amdgpu_device *adev)
+{
+	struct amdgpu_ring *ring;
+	uint64_t ctx_addr;
+	uint32_t param, resp, expected;
+	uint32_t tmp, timeout;
+
+	struct amdgpu_mm_table *table = &adev->virt.mm_table;
+	uint32_t *table_loc;
+	uint32_t table_size;
+	uint32_t size, size_dw, item_offset;
+	uint32_t init_status;
+	int i, j, jpeg_inst;
+
+	struct mmsch_v4_0_cmd_direct_write
+		direct_wt = { {0} };
+	struct mmsch_v4_0_cmd_end end = { {0} };
+	struct mmsch_v4_0_3_init_header header;
+
+	direct_wt.cmd_header.command_type =
+		MMSCH_COMMAND__DIRECT_REG_WRITE;
+	end.cmd_header.command_type =
+		MMSCH_COMMAND__END;
+
+	for (i = 0; i < adev->jpeg.num_jpeg_inst; i++) {
+		jpeg_inst = GET_INST(JPEG, i);
+
+		memset(&header, 0, sizeof(struct mmsch_v4_0_3_init_header));
+		header.version = MMSCH_VERSION;
+		header.total_size = sizeof(struct mmsch_v4_0_3_init_header) >> 2;
+
+		table_loc = (uint32_t *)table->cpu_addr;
+		table_loc += header.total_size;
+
+		item_offset = header.total_size;
+
+		for (j = 0; j < adev->jpeg.num_jpeg_rings; j++) {
+			ring = &adev->jpeg.inst[i].ring_dec[j];
+			table_size = 0;
+
+			tmp = SOC15_REG_OFFSET(JPEG, 0, regUVD_JMI0_UVD_LMI_JRBC_RB_64BIT_BAR_LOW);
+			MMSCH_V4_0_INSERT_DIRECT_WT(tmp, lower_32_bits(ring->gpu_addr));
+			tmp = SOC15_REG_OFFSET(JPEG, 0, regUVD_JMI0_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH);
+			MMSCH_V4_0_INSERT_DIRECT_WT(tmp, upper_32_bits(ring->gpu_addr));
+			tmp = SOC15_REG_OFFSET(JPEG, 0, regUVD_JRBC0_UVD_JRBC_RB_SIZE);
+			MMSCH_V4_0_INSERT_DIRECT_WT(tmp, ring->ring_size / 4);
+
+			if (j <= 3) {
+				header.mjpegdec0[j].table_offset = item_offset;
+				header.mjpegdec0[j].init_status = 0;
+				header.mjpegdec0[j].table_size = table_size;
+			} else {
+				header.mjpegdec1[j - 4].table_offset = item_offset;
+				header.mjpegdec1[j - 4].init_status = 0;
+				header.mjpegdec1[j - 4].table_size = table_size;
+			}
+			header.total_size += table_size;
+			item_offset += table_size;
+		}
+
+		MMSCH_V4_0_INSERT_END();
+
+		/* send init table to MMSCH */
+		size = sizeof(struct mmsch_v4_0_3_init_header);
+		table_loc = (uint32_t *)table->cpu_addr;
+		memcpy((void *)table_loc, &header, size);
+
+		ctx_addr = table->gpu_addr;
+		WREG32_SOC15(VCN, jpeg_inst, regMMSCH_VF_CTX_ADDR_LO, lower_32_bits(ctx_addr));
+		WREG32_SOC15(VCN, jpeg_inst, regMMSCH_VF_CTX_ADDR_HI, upper_32_bits(ctx_addr));
+
+		tmp = RREG32_SOC15(VCN, jpeg_inst, regMMSCH_VF_VMID);
+		tmp &= ~MMSCH_VF_VMID__VF_CTX_VMID_MASK;
+		tmp |= (0 << MMSCH_VF_VMID__VF_CTX_VMID__SHIFT);
+		WREG32_SOC15(VCN, jpeg_inst, regMMSCH_VF_VMID, tmp);
+
+		size = header.total_size;
+		WREG32_SOC15(VCN, jpeg_inst, regMMSCH_VF_CTX_SIZE, size);
+
+		WREG32_SOC15(VCN, jpeg_inst, regMMSCH_VF_MAILBOX_RESP, 0);
+
+		param = 0x00000001;
+		WREG32_SOC15(VCN, jpeg_inst, regMMSCH_VF_MAILBOX_HOST, param);
+		tmp = 0;
+		timeout = 1000;
+		resp = 0;
+		expected = MMSCH_VF_MAILBOX_RESP__OK;
+		init_status =
+			((struct mmsch_v4_0_3_init_header *)(table_loc))->mjpegdec0[i].init_status;
+		while (resp != expected) {
+			resp = RREG32_SOC15(VCN, jpeg_inst, regMMSCH_VF_MAILBOX_RESP);
+
+			if (resp != 0)
+				break;
+			udelay(10);
+			tmp = tmp + 10;
+			if (tmp >= timeout) {
+				DRM_ERROR("failed to init MMSCH. TIME-OUT after %d usec"\
+					" waiting for regMMSCH_VF_MAILBOX_RESP "\
+					"(expected=0x%08x, readback=0x%08x)\n",
+					tmp, expected, resp);
+				return -EBUSY;
+			}
+		}
+		if (resp != expected && resp != MMSCH_VF_MAILBOX_RESP__INCOMPLETE &&
+				init_status != MMSCH_VF_ENGINE_STATUS__PASS)
+			DRM_ERROR("MMSCH init status is incorrect! readback=0x%08x, header init status for jpeg: %x\n",
+					resp, init_status);
+
+	}
+	return 0;
+}
+
+/**
+ * jpeg_v4_0_3_hw_init - start and test JPEG block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ */
+static int jpeg_v4_0_3_hw_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	struct amdgpu_ring *ring;
+	int i, j, r, jpeg_inst;
+
+	if (amdgpu_sriov_vf(adev)) {
+		r = jpeg_v4_0_3_start_sriov(adev);
+		if (r)
+			return r;
+
+		for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
+			for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j) {
+				ring = &adev->jpeg.inst[i].ring_dec[j];
+				ring->wptr = 0;
+				ring->wptr_old = 0;
+				jpeg_v4_0_3_dec_ring_set_wptr(ring);
+				ring->sched.ready = true;
+			}
+		}
+	} else {
+		/* This flag is not set for VF, assumed to be disabled always */
+		if (RREG32_SOC15(VCN, GET_INST(VCN, 0), regVCN_RRMT_CNTL) &
+		    0x100)
+			adev->jpeg.caps |= AMDGPU_JPEG_CAPS(RRMT_ENABLED);
+
+		for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
+			jpeg_inst = GET_INST(JPEG, i);
+
+			ring = adev->jpeg.inst[i].ring_dec;
+
+			if (ring->use_doorbell)
+				adev->nbio.funcs->vcn_doorbell_range(
+					adev, ring->use_doorbell,
+					(adev->doorbell_index.vcn.vcn_ring0_1 << 1) +
+						9 * jpeg_inst,
+					adev->jpeg.inst[i].aid_id);
+
+			for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j) {
+				ring = &adev->jpeg.inst[i].ring_dec[j];
+				if (ring->use_doorbell)
+					WREG32_SOC15_OFFSET(
+						VCN, GET_INST(VCN, i),
+						regVCN_JPEG_DB_CTRL,
+						(ring->pipe ? (ring->pipe - 0x15) : 0),
+						ring->doorbell_index
+							<< VCN_JPEG_DB_CTRL__OFFSET__SHIFT |
+							VCN_JPEG_DB_CTRL__EN_MASK);
+				r = amdgpu_ring_test_helper(ring);
+				if (r)
+					return r;
+			}
+		}
+	}
+
+	return 0;
+}
+
+/**
+ * jpeg_v4_0_3_hw_fini - stop the hardware block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Stop the JPEG block, mark ring as not ready any more
+ */
+static int jpeg_v4_0_3_hw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int ret = 0;
+
+	cancel_delayed_work_sync(&adev->jpeg.idle_work);
+
+	if (!amdgpu_sriov_vf(adev)) {
+		if (adev->jpeg.cur_state != AMD_PG_STATE_GATE)
+			ret = jpeg_v4_0_3_set_powergating_state(ip_block, AMD_PG_STATE_GATE);
+	}
+
+	if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__JPEG) && !amdgpu_sriov_vf(adev))
+		amdgpu_irq_put(adev, &adev->jpeg.inst->ras_poison_irq, 0);
+
+	return ret;
+}
+
+/**
+ * jpeg_v4_0_3_suspend - suspend JPEG block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * HW fini and suspend JPEG block
+ */
+static int jpeg_v4_0_3_suspend(struct amdgpu_ip_block *ip_block)
+{
+	int r;
+
+	r = jpeg_v4_0_3_hw_fini(ip_block);
+	if (r)
+		return r;
+
+	r = amdgpu_jpeg_suspend(ip_block->adev);
+
+	return r;
+}
+
+/**
+ * jpeg_v4_0_3_resume - resume JPEG block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Resume firmware and hw init JPEG block
+ */
+static int jpeg_v4_0_3_resume(struct amdgpu_ip_block *ip_block)
+{
+	int r;
+
+	r = amdgpu_jpeg_resume(ip_block->adev);
+	if (r)
+		return r;
+
+	r = jpeg_v4_0_3_hw_init(ip_block);
+
+	return r;
+}
+
+static void jpeg_v4_0_3_disable_clock_gating(struct amdgpu_device *adev, int inst_idx)
+{
+	int i, jpeg_inst;
+	uint32_t data;
+
+	jpeg_inst = GET_INST(JPEG, inst_idx);
+	data = RREG32_SOC15(JPEG, jpeg_inst, regJPEG_CGC_CTRL);
+	if (adev->cg_flags & AMD_CG_SUPPORT_JPEG_MGCG) {
+		data |= 1 << JPEG_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
+		data &= (~(JPEG_CGC_CTRL__JPEG0_DEC_MODE_MASK << 1));
+	} else {
+		data &= ~JPEG_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
+	}
+
+	data |= 1 << JPEG_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT;
+	data |= 4 << JPEG_CGC_CTRL__CLK_OFF_DELAY__SHIFT;
+	WREG32_SOC15(JPEG, jpeg_inst, regJPEG_CGC_CTRL, data);
+
+	data = RREG32_SOC15(JPEG, jpeg_inst, regJPEG_CGC_GATE);
+	data &= ~(JPEG_CGC_GATE__JMCIF_MASK | JPEG_CGC_GATE__JRBBM_MASK);
+	for (i = 0; i < adev->jpeg.num_jpeg_rings; ++i)
+		data &= ~(JPEG_CGC_GATE__JPEG0_DEC_MASK << i);
+	WREG32_SOC15(JPEG, jpeg_inst, regJPEG_CGC_GATE, data);
+}
+
+static void jpeg_v4_0_3_enable_clock_gating(struct amdgpu_device *adev, int inst_idx)
+{
+	int i, jpeg_inst;
+	uint32_t data;
+
+	jpeg_inst = GET_INST(JPEG, inst_idx);
+	data = RREG32_SOC15(JPEG, jpeg_inst, regJPEG_CGC_CTRL);
+	if (adev->cg_flags & AMD_CG_SUPPORT_JPEG_MGCG) {
+		data |= 1 << JPEG_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
+		data |= (JPEG_CGC_CTRL__JPEG0_DEC_MODE_MASK << 1);
+	} else {
+		data &= ~JPEG_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
+	}
+
+	data |= 1 << JPEG_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT;
+	data |= 4 << JPEG_CGC_CTRL__CLK_OFF_DELAY__SHIFT;
+	WREG32_SOC15(JPEG, jpeg_inst, regJPEG_CGC_CTRL, data);
+
+	data = RREG32_SOC15(JPEG, jpeg_inst, regJPEG_CGC_GATE);
+	data |= (JPEG_CGC_GATE__JMCIF_MASK | JPEG_CGC_GATE__JRBBM_MASK);
+	for (i = 0; i < adev->jpeg.num_jpeg_rings; ++i)
+		data |= (JPEG_CGC_GATE__JPEG0_DEC_MASK << i);
+	WREG32_SOC15(JPEG, jpeg_inst, regJPEG_CGC_GATE, data);
+}
+
+static void jpeg_v4_0_3_start_inst(struct amdgpu_device *adev, int inst)
+{
+	int jpeg_inst = GET_INST(JPEG, inst);
+
+	WREG32_SOC15(JPEG, jpeg_inst, regUVD_PGFSM_CONFIG,
+		     1 << UVD_PGFSM_CONFIG__UVDJ_PWR_CONFIG__SHIFT);
+	SOC15_WAIT_ON_RREG(JPEG, jpeg_inst, regUVD_PGFSM_STATUS,
+			   UVD_PGFSM_STATUS__UVDJ_PWR_ON <<
+			   UVD_PGFSM_STATUS__UVDJ_PWR_STATUS__SHIFT,
+			   UVD_PGFSM_STATUS__UVDJ_PWR_STATUS_MASK);
+
+	/* disable anti hang mechanism */
+	WREG32_P(SOC15_REG_OFFSET(JPEG, jpeg_inst, regUVD_JPEG_POWER_STATUS),
+		 0, ~UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK);
+
+	/* JPEG disable CGC */
+	jpeg_v4_0_3_disable_clock_gating(adev, inst);
+
+	/* MJPEG global tiling registers */
+	WREG32_SOC15(JPEG, jpeg_inst, regJPEG_DEC_GFX8_ADDR_CONFIG,
+		     adev->gfx.config.gb_addr_config);
+	WREG32_SOC15(JPEG, jpeg_inst, regJPEG_DEC_GFX10_ADDR_CONFIG,
+		     adev->gfx.config.gb_addr_config);
+
+	/* enable JMI channel */
+	WREG32_P(SOC15_REG_OFFSET(JPEG, jpeg_inst, regUVD_JMI_CNTL), 0,
+		 ~UVD_JMI_CNTL__SOFT_RESET_MASK);
+}
+
+static void jpeg_v4_0_3_start_jrbc(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	int jpeg_inst = GET_INST(JPEG, ring->me);
+	int reg_offset = jpeg_v4_0_3_core_reg_offset(ring->pipe);
+
+	/* enable System Interrupt for JRBC */
+	WREG32_P(SOC15_REG_OFFSET(JPEG, jpeg_inst, regJPEG_SYS_INT_EN),
+		 JPEG_SYS_INT_EN__DJRBC0_MASK << ring->pipe,
+		 ~(JPEG_SYS_INT_EN__DJRBC0_MASK << ring->pipe));
+
+	WREG32_SOC15_OFFSET(JPEG, jpeg_inst,
+			    regUVD_JMI0_UVD_LMI_JRBC_RB_VMID,
+			    reg_offset, 0);
+	WREG32_SOC15_OFFSET(JPEG, jpeg_inst,
+			    regUVD_JRBC0_UVD_JRBC_RB_CNTL,
+			    reg_offset,
+			    (0x00000001L | 0x00000002L));
+	WREG32_SOC15_OFFSET(JPEG, jpeg_inst,
+			    regUVD_JMI0_UVD_LMI_JRBC_RB_64BIT_BAR_LOW,
+			    reg_offset, lower_32_bits(ring->gpu_addr));
+	WREG32_SOC15_OFFSET(JPEG, jpeg_inst,
+			    regUVD_JMI0_UVD_LMI_JRBC_RB_64BIT_BAR_HIGH,
+			    reg_offset, upper_32_bits(ring->gpu_addr));
+	WREG32_SOC15_OFFSET(JPEG, jpeg_inst,
+			    regUVD_JRBC0_UVD_JRBC_RB_RPTR,
+			    reg_offset, 0);
+	WREG32_SOC15_OFFSET(JPEG, jpeg_inst,
+			    regUVD_JRBC0_UVD_JRBC_RB_WPTR,
+			    reg_offset, 0);
+	WREG32_SOC15_OFFSET(JPEG, jpeg_inst,
+			    regUVD_JRBC0_UVD_JRBC_RB_CNTL,
+			    reg_offset, 0x00000002L);
+	WREG32_SOC15_OFFSET(JPEG, jpeg_inst,
+			    regUVD_JRBC0_UVD_JRBC_RB_SIZE,
+			    reg_offset, ring->ring_size / 4);
+	ring->wptr = RREG32_SOC15_OFFSET(JPEG, jpeg_inst, regUVD_JRBC0_UVD_JRBC_RB_WPTR,
+					 reg_offset);
+}
+
+/**
+ * jpeg_v4_0_3_start - start JPEG block
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Setup and start the JPEG block
+ */
+static int jpeg_v4_0_3_start(struct amdgpu_device *adev)
+{
+	struct amdgpu_ring *ring;
+	int i, j;
+
+	for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
+		jpeg_v4_0_3_start_inst(adev, i);
+		for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j) {
+			ring = &adev->jpeg.inst[i].ring_dec[j];
+			jpeg_v4_0_3_start_jrbc(ring);
+		}
+	}
+
+	return 0;
+}
+
+static void jpeg_v4_0_3_stop_inst(struct amdgpu_device *adev, int inst)
+{
+	int jpeg_inst = GET_INST(JPEG, inst);
+	/* reset JMI */
+	WREG32_P(SOC15_REG_OFFSET(JPEG, jpeg_inst, regUVD_JMI_CNTL),
+		 UVD_JMI_CNTL__SOFT_RESET_MASK,
+		 ~UVD_JMI_CNTL__SOFT_RESET_MASK);
+
+	jpeg_v4_0_3_enable_clock_gating(adev, inst);
+
+	/* enable anti hang mechanism */
+	WREG32_P(SOC15_REG_OFFSET(JPEG, jpeg_inst, regUVD_JPEG_POWER_STATUS),
+		 UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK,
+		 ~UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK);
+
+}
+
+/**
+ * jpeg_v4_0_3_stop - stop JPEG block
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * stop the JPEG block
+ */
+static int jpeg_v4_0_3_stop(struct amdgpu_device *adev)
+{
+	int i;
+
+	for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i)
+		jpeg_v4_0_3_stop_inst(adev, i);
+
+	return 0;
+}
+
+/**
+ * jpeg_v4_0_3_dec_ring_get_rptr - get read pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Returns the current hardware read pointer
+ */
+static uint64_t jpeg_v4_0_3_dec_ring_get_rptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	return RREG32_SOC15_OFFSET(JPEG, GET_INST(JPEG, ring->me), regUVD_JRBC0_UVD_JRBC_RB_RPTR,
+				   jpeg_v4_0_3_core_reg_offset(ring->pipe));
+}
+
+/**
+ * jpeg_v4_0_3_dec_ring_get_wptr - get write pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Returns the current hardware write pointer
+ */
+static uint64_t jpeg_v4_0_3_dec_ring_get_wptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	if (ring->use_doorbell)
+		return adev->wb.wb[ring->wptr_offs];
+
+	return RREG32_SOC15_OFFSET(JPEG, GET_INST(JPEG, ring->me), regUVD_JRBC0_UVD_JRBC_RB_WPTR,
+				   jpeg_v4_0_3_core_reg_offset(ring->pipe));
+}
+
+void jpeg_v4_0_3_ring_emit_hdp_flush(struct amdgpu_ring *ring)
+{
+	/* JPEG engine access for HDP flush doesn't work when RRMT is enabled.
+	 * This is a workaround to avoid any HDP flush through JPEG ring.
+	 */
+}
+
+/**
+ * jpeg_v4_0_3_dec_ring_set_wptr - set write pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Commits the write pointer to the hardware
+ */
+static void jpeg_v4_0_3_dec_ring_set_wptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	if (ring->use_doorbell) {
+		adev->wb.wb[ring->wptr_offs] = lower_32_bits(ring->wptr);
+		WDOORBELL32(ring->doorbell_index, lower_32_bits(ring->wptr));
+	} else {
+		WREG32_SOC15_OFFSET(JPEG, GET_INST(JPEG, ring->me), regUVD_JRBC0_UVD_JRBC_RB_WPTR,
+				    jpeg_v4_0_3_core_reg_offset(ring->pipe),
+				    lower_32_bits(ring->wptr));
+	}
+}
+
+/**
+ * jpeg_v4_0_3_dec_ring_insert_start - insert a start command
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Write a start command to the ring.
+ */
+void jpeg_v4_0_3_dec_ring_insert_start(struct amdgpu_ring *ring)
+{
+	if (!amdgpu_sriov_vf(ring->adev)) {
+		amdgpu_ring_write(ring, PACKETJ(regUVD_JRBC_EXTERNAL_REG_INTERNAL_OFFSET,
+			0, 0, PACKETJ_TYPE0));
+		amdgpu_ring_write(ring, 0x62a04); /* PCTL0_MMHUB_DEEPSLEEP_IB */
+
+		amdgpu_ring_write(ring,
+				  PACKETJ(JRBC_DEC_EXTERNAL_REG_WRITE_ADDR, 0,
+					  0, PACKETJ_TYPE0));
+		amdgpu_ring_write(ring, 0x80004000);
+	}
+}
+
+/**
+ * jpeg_v4_0_3_dec_ring_insert_end - insert a end command
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Write a end command to the ring.
+ */
+void jpeg_v4_0_3_dec_ring_insert_end(struct amdgpu_ring *ring)
+{
+	if (!amdgpu_sriov_vf(ring->adev)) {
+		amdgpu_ring_write(ring, PACKETJ(regUVD_JRBC_EXTERNAL_REG_INTERNAL_OFFSET,
+			0, 0, PACKETJ_TYPE0));
+		amdgpu_ring_write(ring, 0x62a04);
+
+		amdgpu_ring_write(ring,
+				  PACKETJ(JRBC_DEC_EXTERNAL_REG_WRITE_ADDR, 0,
+					  0, PACKETJ_TYPE0));
+		amdgpu_ring_write(ring, 0x00004000);
+	}
+}
+
+/**
+ * jpeg_v4_0_3_dec_ring_emit_fence - emit an fence & trap command
+ *
+ * @ring: amdgpu_ring pointer
+ * @addr: address
+ * @seq: sequence number
+ * @flags: fence related flags
+ *
+ * Write a fence and a trap command to the ring.
+ */
+void jpeg_v4_0_3_dec_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 seq,
+				unsigned int flags)
+{
+	WARN_ON(flags & AMDGPU_FENCE_FLAG_64BIT);
+
+	amdgpu_ring_write(ring, PACKETJ(regUVD_JPEG_GPCOM_DATA0_INTERNAL_OFFSET,
+		0, 0, PACKETJ_TYPE0));
+	amdgpu_ring_write(ring, seq);
+
+	amdgpu_ring_write(ring,	PACKETJ(regUVD_JPEG_GPCOM_DATA1_INTERNAL_OFFSET,
+		0, 0, PACKETJ_TYPE0));
+	amdgpu_ring_write(ring, seq);
+
+	amdgpu_ring_write(ring,	PACKETJ(regUVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW_INTERNAL_OFFSET,
+		0, 0, PACKETJ_TYPE0));
+	amdgpu_ring_write(ring, lower_32_bits(addr));
+
+	amdgpu_ring_write(ring,	PACKETJ(regUVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH_INTERNAL_OFFSET,
+		0, 0, PACKETJ_TYPE0));
+	amdgpu_ring_write(ring, upper_32_bits(addr));
+
+	amdgpu_ring_write(ring,	PACKETJ(regUVD_JPEG_GPCOM_CMD_INTERNAL_OFFSET,
+		0, 0, PACKETJ_TYPE0));
+	amdgpu_ring_write(ring, 0x8);
+
+	amdgpu_ring_write(ring,	PACKETJ(regUVD_JPEG_GPCOM_CMD_INTERNAL_OFFSET,
+		0, PACKETJ_CONDITION_CHECK0, PACKETJ_TYPE4));
+	amdgpu_ring_write(ring, 0);
+
+	amdgpu_ring_write(ring, PACKETJ(0, 0, 0, PACKETJ_TYPE6));
+	amdgpu_ring_write(ring, 0);
+
+	amdgpu_ring_write(ring, PACKETJ(0, 0, 0, PACKETJ_TYPE6));
+	amdgpu_ring_write(ring, 0);
+
+	amdgpu_ring_write(ring, PACKETJ(0, 0, 0, PACKETJ_TYPE7));
+	amdgpu_ring_write(ring, 0);
+}
+
+/**
+ * jpeg_v4_0_3_dec_ring_emit_ib - execute indirect buffer
+ *
+ * @ring: amdgpu_ring pointer
+ * @job: job to retrieve vmid from
+ * @ib: indirect buffer to execute
+ * @flags: unused
+ *
+ * Write ring commands to execute the indirect buffer.
+ */
+void jpeg_v4_0_3_dec_ring_emit_ib(struct amdgpu_ring *ring,
+				struct amdgpu_job *job,
+				struct amdgpu_ib *ib,
+				uint32_t flags)
+{
+	unsigned int vmid = AMDGPU_JOB_GET_VMID(job);
+
+	amdgpu_ring_write(ring, PACKETJ(regUVD_LMI_JRBC_IB_VMID_INTERNAL_OFFSET,
+		0, 0, PACKETJ_TYPE0));
+
+	if (ring->funcs->parse_cs)
+		amdgpu_ring_write(ring, 0);
+	else
+		amdgpu_ring_write(ring, (vmid | (vmid << 4) | (vmid << 8)));
+
+	amdgpu_ring_write(ring, PACKETJ(regUVD_LMI_JPEG_VMID_INTERNAL_OFFSET,
+		0, 0, PACKETJ_TYPE0));
+	amdgpu_ring_write(ring, (vmid | (vmid << 4) | (vmid << 8)));
+
+	amdgpu_ring_write(ring,	PACKETJ(regUVD_LMI_JRBC_IB_64BIT_BAR_LOW_INTERNAL_OFFSET,
+		0, 0, PACKETJ_TYPE0));
+	amdgpu_ring_write(ring, lower_32_bits(ib->gpu_addr));
+
+	amdgpu_ring_write(ring,	PACKETJ(regUVD_LMI_JRBC_IB_64BIT_BAR_HIGH_INTERNAL_OFFSET,
+		0, 0, PACKETJ_TYPE0));
+	amdgpu_ring_write(ring, upper_32_bits(ib->gpu_addr));
+
+	amdgpu_ring_write(ring,	PACKETJ(regUVD_JRBC_IB_SIZE_INTERNAL_OFFSET,
+		0, 0, PACKETJ_TYPE0));
+	amdgpu_ring_write(ring, ib->length_dw);
+
+	amdgpu_ring_write(ring,	PACKETJ(regUVD_LMI_JRBC_RB_MEM_RD_64BIT_BAR_LOW_INTERNAL_OFFSET,
+		0, 0, PACKETJ_TYPE0));
+	amdgpu_ring_write(ring, lower_32_bits(ring->gpu_addr));
+
+	amdgpu_ring_write(ring,	PACKETJ(regUVD_LMI_JRBC_RB_MEM_RD_64BIT_BAR_HIGH_INTERNAL_OFFSET,
+		0, 0, PACKETJ_TYPE0));
+	amdgpu_ring_write(ring, upper_32_bits(ring->gpu_addr));
+
+	amdgpu_ring_write(ring,	PACKETJ(0, 0, PACKETJ_CONDITION_CHECK0, PACKETJ_TYPE2));
+	amdgpu_ring_write(ring, 0);
+
+	amdgpu_ring_write(ring,	PACKETJ(regUVD_JRBC_RB_COND_RD_TIMER_INTERNAL_OFFSET,
+		0, 0, PACKETJ_TYPE0));
+	amdgpu_ring_write(ring, 0x01400200);
+
+	amdgpu_ring_write(ring, PACKETJ(regUVD_JRBC_RB_REF_DATA_INTERNAL_OFFSET,
+		0, 0, PACKETJ_TYPE0));
+	amdgpu_ring_write(ring, 0x2);
+
+	amdgpu_ring_write(ring,	PACKETJ(regUVD_JRBC_STATUS_INTERNAL_OFFSET,
+		0, PACKETJ_CONDITION_CHECK3, PACKETJ_TYPE3));
+	amdgpu_ring_write(ring, 0x2);
+}
+
+void jpeg_v4_0_3_dec_ring_emit_reg_wait(struct amdgpu_ring *ring, uint32_t reg,
+				uint32_t val, uint32_t mask)
+{
+	uint32_t reg_offset;
+
+	/* Use normalized offsets if required */
+	if (jpeg_v4_0_3_normalizn_reqd(ring->adev))
+		reg = NORMALIZE_JPEG_REG_OFFSET(reg);
+
+	reg_offset = (reg << 2);
+
+	amdgpu_ring_write(ring, PACKETJ(regUVD_JRBC_RB_COND_RD_TIMER_INTERNAL_OFFSET,
+		0, 0, PACKETJ_TYPE0));
+	amdgpu_ring_write(ring, 0x01400200);
+
+	amdgpu_ring_write(ring,	PACKETJ(regUVD_JRBC_RB_REF_DATA_INTERNAL_OFFSET,
+		0, 0, PACKETJ_TYPE0));
+	amdgpu_ring_write(ring, val);
+
+	amdgpu_ring_write(ring, PACKETJ(regUVD_JRBC_EXTERNAL_REG_INTERNAL_OFFSET,
+		0, 0, PACKETJ_TYPE0));
+	if (reg_offset >= 0x10000 && reg_offset <= 0x105ff) {
+		amdgpu_ring_write(ring, 0);
+		amdgpu_ring_write(ring,
+			PACKETJ((reg_offset >> 2), 0, 0, PACKETJ_TYPE3));
+	} else {
+		amdgpu_ring_write(ring, reg_offset);
+		amdgpu_ring_write(ring,	PACKETJ(JRBC_DEC_EXTERNAL_REG_WRITE_ADDR,
+			0, 0, PACKETJ_TYPE3));
+	}
+	amdgpu_ring_write(ring, mask);
+}
+
+void jpeg_v4_0_3_dec_ring_emit_vm_flush(struct amdgpu_ring *ring,
+				unsigned int vmid, uint64_t pd_addr)
+{
+	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->vm_hub];
+	uint32_t data0, data1, mask;
+
+	pd_addr = amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pd_addr);
+
+	/* wait for register write */
+	data0 = hub->ctx0_ptb_addr_lo32 + vmid * hub->ctx_addr_distance;
+	data1 = lower_32_bits(pd_addr);
+	mask = 0xffffffff;
+	jpeg_v4_0_3_dec_ring_emit_reg_wait(ring, data0, data1, mask);
+}
+
+void jpeg_v4_0_3_dec_ring_emit_wreg(struct amdgpu_ring *ring, uint32_t reg, uint32_t val)
+{
+	uint32_t reg_offset;
+
+	/* Use normalized offsets if required */
+	if (jpeg_v4_0_3_normalizn_reqd(ring->adev))
+		reg = NORMALIZE_JPEG_REG_OFFSET(reg);
+
+	reg_offset = (reg << 2);
+
+	amdgpu_ring_write(ring,	PACKETJ(regUVD_JRBC_EXTERNAL_REG_INTERNAL_OFFSET,
+		0, 0, PACKETJ_TYPE0));
+	if (reg_offset >= 0x10000 && reg_offset <= 0x105ff) {
+		amdgpu_ring_write(ring, 0);
+		amdgpu_ring_write(ring,
+			PACKETJ((reg_offset >> 2), 0, 0, PACKETJ_TYPE0));
+	} else {
+		amdgpu_ring_write(ring, reg_offset);
+		amdgpu_ring_write(ring,	PACKETJ(JRBC_DEC_EXTERNAL_REG_WRITE_ADDR,
+			0, 0, PACKETJ_TYPE0));
+	}
+	amdgpu_ring_write(ring, val);
+}
+
+void jpeg_v4_0_3_dec_ring_nop(struct amdgpu_ring *ring, uint32_t count)
+{
+	int i;
+
+	WARN_ON(ring->wptr % 2 || count % 2);
+
+	for (i = 0; i < count / 2; i++) {
+		amdgpu_ring_write(ring, PACKETJ(0, 0, 0, PACKETJ_TYPE6));
+		amdgpu_ring_write(ring, 0);
+	}
+}
+
+static bool jpeg_v4_0_3_is_idle(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	bool ret = false;
+	int i, j;
+
+	for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
+		for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j) {
+			ret &= ((RREG32_SOC15_OFFSET(JPEG, GET_INST(JPEG, i),
+				regUVD_JRBC0_UVD_JRBC_STATUS, jpeg_v4_0_3_core_reg_offset(j)) &
+				UVD_JRBC0_UVD_JRBC_STATUS__RB_JOB_DONE_MASK) ==
+				UVD_JRBC0_UVD_JRBC_STATUS__RB_JOB_DONE_MASK);
+		}
+	}
+
+	return ret;
+}
+
+static int jpeg_v4_0_3_wait_for_idle(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int ret = 0;
+	int i, j;
+
+	for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
+		for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j) {
+			ret &= (SOC15_WAIT_ON_RREG_OFFSET(JPEG, GET_INST(JPEG, i),
+				regUVD_JRBC0_UVD_JRBC_STATUS, jpeg_v4_0_3_core_reg_offset(j),
+				UVD_JRBC0_UVD_JRBC_STATUS__RB_JOB_DONE_MASK,
+				UVD_JRBC0_UVD_JRBC_STATUS__RB_JOB_DONE_MASK));
+		}
+	}
+	return ret;
+}
+
+static int jpeg_v4_0_3_set_clockgating_state(struct amdgpu_ip_block *ip_block,
+					  enum amd_clockgating_state state)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	bool enable = state == AMD_CG_STATE_GATE;
+	int i;
+
+	for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
+		if (enable) {
+			if (!jpeg_v4_0_3_is_idle(ip_block))
+				return -EBUSY;
+			jpeg_v4_0_3_enable_clock_gating(adev, i);
+		} else {
+			jpeg_v4_0_3_disable_clock_gating(adev, i);
+		}
+	}
+	return 0;
+}
+
+static int jpeg_v4_0_3_set_powergating_state(struct amdgpu_ip_block *ip_block,
+					  enum amd_powergating_state state)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int ret;
+
+	if (amdgpu_sriov_vf(adev)) {
+		adev->jpeg.cur_state = AMD_PG_STATE_UNGATE;
+		return 0;
+	}
+
+	if (state == adev->jpeg.cur_state)
+		return 0;
+
+	if (state == AMD_PG_STATE_GATE)
+		ret = jpeg_v4_0_3_stop(adev);
+	else
+		ret = jpeg_v4_0_3_start(adev);
+
+	if (!ret)
+		adev->jpeg.cur_state = state;
+
+	return ret;
+}
+
+static int jpeg_v4_0_3_set_interrupt_state(struct amdgpu_device *adev,
+					struct amdgpu_irq_src *source,
+					unsigned int type,
+					enum amdgpu_interrupt_state state)
+{
+	return 0;
+}
+
+static int jpeg_v4_0_3_set_ras_interrupt_state(struct amdgpu_device *adev,
+					struct amdgpu_irq_src *source,
+					unsigned int type,
+					enum amdgpu_interrupt_state state)
+{
+	return 0;
+}
+
+static int jpeg_v4_0_3_process_interrupt(struct amdgpu_device *adev,
+				      struct amdgpu_irq_src *source,
+				      struct amdgpu_iv_entry *entry)
+{
+	uint32_t i, inst;
+
+	i = node_id_to_phys_map[entry->node_id];
+	DRM_DEV_DEBUG(adev->dev, "IH: JPEG TRAP\n");
+
+	for (inst = 0; inst < adev->jpeg.num_jpeg_inst; ++inst)
+		if (adev->jpeg.inst[inst].aid_id == i)
+			break;
+
+	if (inst >= adev->jpeg.num_jpeg_inst) {
+		dev_WARN_ONCE(adev->dev, 1,
+			      "Interrupt received for unknown JPEG instance %d",
+			      entry->node_id);
+		return 0;
+	}
+
+	switch (entry->src_id) {
+	case VCN_4_0__SRCID__JPEG_DECODE:
+		amdgpu_fence_process(&adev->jpeg.inst[inst].ring_dec[0]);
+		break;
+	case VCN_4_0__SRCID__JPEG1_DECODE:
+		amdgpu_fence_process(&adev->jpeg.inst[inst].ring_dec[1]);
+		break;
+	case VCN_4_0__SRCID__JPEG2_DECODE:
+		amdgpu_fence_process(&adev->jpeg.inst[inst].ring_dec[2]);
+		break;
+	case VCN_4_0__SRCID__JPEG3_DECODE:
+		amdgpu_fence_process(&adev->jpeg.inst[inst].ring_dec[3]);
+		break;
+	case VCN_4_0__SRCID__JPEG4_DECODE:
+		amdgpu_fence_process(&adev->jpeg.inst[inst].ring_dec[4]);
+		break;
+	case VCN_4_0__SRCID__JPEG5_DECODE:
+		amdgpu_fence_process(&adev->jpeg.inst[inst].ring_dec[5]);
+		break;
+	case VCN_4_0__SRCID__JPEG6_DECODE:
+		amdgpu_fence_process(&adev->jpeg.inst[inst].ring_dec[6]);
+		break;
+	case VCN_4_0__SRCID__JPEG7_DECODE:
+		amdgpu_fence_process(&adev->jpeg.inst[inst].ring_dec[7]);
+		break;
+	default:
+		DRM_DEV_ERROR(adev->dev, "Unhandled interrupt: %d %d\n",
+			  entry->src_id, entry->src_data[0]);
+		break;
+	}
+
+	return 0;
+}
+
+static void jpeg_v4_0_3_core_stall_reset(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	int jpeg_inst = GET_INST(JPEG, ring->me);
+	int reg_offset = jpeg_v4_0_3_core_reg_offset(ring->pipe);
+
+	WREG32_SOC15_OFFSET(JPEG, jpeg_inst,
+			    regUVD_JMI0_UVD_JMI_CLIENT_STALL,
+			    reg_offset, 0x1F);
+	SOC15_WAIT_ON_RREG_OFFSET(JPEG, jpeg_inst,
+				  regUVD_JMI0_UVD_JMI_CLIENT_CLEAN_STATUS,
+				  reg_offset, 0x1F, 0x1F);
+	WREG32_SOC15_OFFSET(JPEG, jpeg_inst,
+			    regUVD_JMI0_JPEG_LMI_DROP,
+			    reg_offset, 0x1F);
+	WREG32_SOC15(JPEG, jpeg_inst, regJPEG_CORE_RST_CTRL, 1 << ring->pipe);
+	WREG32_SOC15_OFFSET(JPEG, jpeg_inst,
+			    regUVD_JMI0_UVD_JMI_CLIENT_STALL,
+			    reg_offset, 0x00);
+	WREG32_SOC15_OFFSET(JPEG, jpeg_inst,
+			    regUVD_JMI0_JPEG_LMI_DROP,
+			    reg_offset, 0x00);
+	WREG32_SOC15(JPEG, jpeg_inst, regJPEG_CORE_RST_CTRL, 0x00);
+}
+
+static int jpeg_v4_0_3_ring_reset(struct amdgpu_ring *ring,
+				  unsigned int vmid,
+				  struct amdgpu_fence *timedout_fence)
+{
+	if (amdgpu_sriov_vf(ring->adev))
+		return -EOPNOTSUPP;
+
+	amdgpu_ring_reset_helper_begin(ring, timedout_fence);
+	jpeg_v4_0_3_core_stall_reset(ring);
+	jpeg_v4_0_3_start_jrbc(ring);
+	return amdgpu_ring_reset_helper_end(ring, timedout_fence);
+}
+
+static const struct amd_ip_funcs jpeg_v4_0_3_ip_funcs = {
+	.name = "jpeg_v4_0_3",
+	.early_init = jpeg_v4_0_3_early_init,
+	.sw_init = jpeg_v4_0_3_sw_init,
+	.sw_fini = jpeg_v4_0_3_sw_fini,
+	.hw_init = jpeg_v4_0_3_hw_init,
+	.hw_fini = jpeg_v4_0_3_hw_fini,
+	.suspend = jpeg_v4_0_3_suspend,
+	.resume = jpeg_v4_0_3_resume,
+	.is_idle = jpeg_v4_0_3_is_idle,
+	.wait_for_idle = jpeg_v4_0_3_wait_for_idle,
+	.set_clockgating_state = jpeg_v4_0_3_set_clockgating_state,
+	.set_powergating_state = jpeg_v4_0_3_set_powergating_state,
+	.dump_ip_state = amdgpu_jpeg_dump_ip_state,
+	.print_ip_state = amdgpu_jpeg_print_ip_state,
+};
+
+static const struct amdgpu_ring_funcs jpeg_v4_0_3_dec_ring_vm_funcs = {
+	.type = AMDGPU_RING_TYPE_VCN_JPEG,
+	.align_mask = 0xf,
+	.get_rptr = jpeg_v4_0_3_dec_ring_get_rptr,
+	.get_wptr = jpeg_v4_0_3_dec_ring_get_wptr,
+	.set_wptr = jpeg_v4_0_3_dec_ring_set_wptr,
+	.parse_cs = amdgpu_jpeg_dec_parse_cs,
+	.emit_frame_size =
+		SOC15_FLUSH_GPU_TLB_NUM_WREG * 6 +
+		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 8 +
+		8 + /* jpeg_v4_0_3_dec_ring_emit_vm_flush */
+		18 + 18 + /* jpeg_v4_0_3_dec_ring_emit_fence x2 vm fence */
+		8 + 16,
+	.emit_ib_size = 22, /* jpeg_v4_0_3_dec_ring_emit_ib */
+	.emit_ib = jpeg_v4_0_3_dec_ring_emit_ib,
+	.emit_fence = jpeg_v4_0_3_dec_ring_emit_fence,
+	.emit_vm_flush = jpeg_v4_0_3_dec_ring_emit_vm_flush,
+	.emit_hdp_flush = jpeg_v4_0_3_ring_emit_hdp_flush,
+	.test_ring = amdgpu_jpeg_dec_ring_test_ring,
+	.test_ib = amdgpu_jpeg_dec_ring_test_ib,
+	.insert_nop = jpeg_v4_0_3_dec_ring_nop,
+	.insert_start = jpeg_v4_0_3_dec_ring_insert_start,
+	.insert_end = jpeg_v4_0_3_dec_ring_insert_end,
+	.pad_ib = amdgpu_ring_generic_pad_ib,
+	.begin_use = amdgpu_jpeg_ring_begin_use,
+	.end_use = amdgpu_jpeg_ring_end_use,
+	.emit_wreg = jpeg_v4_0_3_dec_ring_emit_wreg,
+	.emit_reg_wait = jpeg_v4_0_3_dec_ring_emit_reg_wait,
+	.emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper,
+	.reset = jpeg_v4_0_3_ring_reset,
+};
+
+static void jpeg_v4_0_3_set_dec_ring_funcs(struct amdgpu_device *adev)
+{
+	int i, j, jpeg_inst;
+
+	for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
+		for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j) {
+			adev->jpeg.inst[i].ring_dec[j].funcs = &jpeg_v4_0_3_dec_ring_vm_funcs;
+			adev->jpeg.inst[i].ring_dec[j].me = i;
+			adev->jpeg.inst[i].ring_dec[j].pipe = j;
+		}
+		jpeg_inst = GET_INST(JPEG, i);
+		adev->jpeg.inst[i].aid_id =
+			jpeg_inst / adev->jpeg.num_inst_per_aid;
+	}
+}
+
+static const struct amdgpu_irq_src_funcs jpeg_v4_0_3_irq_funcs = {
+	.set = jpeg_v4_0_3_set_interrupt_state,
+	.process = jpeg_v4_0_3_process_interrupt,
+};
+
+static const struct amdgpu_irq_src_funcs jpeg_v4_0_3_ras_irq_funcs = {
+	.set = jpeg_v4_0_3_set_ras_interrupt_state,
+	.process = amdgpu_jpeg_process_poison_irq,
+};
+
+static void jpeg_v4_0_3_set_irq_funcs(struct amdgpu_device *adev)
+{
+	int i;
+
+	for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
+		adev->jpeg.inst->irq.num_types += adev->jpeg.num_jpeg_rings;
+	}
+	adev->jpeg.inst->irq.funcs = &jpeg_v4_0_3_irq_funcs;
+
+	adev->jpeg.inst->ras_poison_irq.num_types = 1;
+	adev->jpeg.inst->ras_poison_irq.funcs = &jpeg_v4_0_3_ras_irq_funcs;
+}
+
+const struct amdgpu_ip_block_version jpeg_v4_0_3_ip_block = {
+	.type = AMD_IP_BLOCK_TYPE_JPEG,
+	.major = 4,
+	.minor = 0,
+	.rev = 3,
+	.funcs = &jpeg_v4_0_3_ip_funcs,
+};
+
+static const struct amdgpu_ras_err_status_reg_entry jpeg_v4_0_3_ue_reg_list[] = {
+	{AMDGPU_RAS_REG_ENTRY(JPEG, 0, regVCN_UE_ERR_STATUS_LO_JPEG0S, regVCN_UE_ERR_STATUS_HI_JPEG0S),
+	1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "JPEG0S"},
+	{AMDGPU_RAS_REG_ENTRY(JPEG, 0, regVCN_UE_ERR_STATUS_LO_JPEG0D, regVCN_UE_ERR_STATUS_HI_JPEG0D),
+	1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "JPEG0D"},
+	{AMDGPU_RAS_REG_ENTRY(JPEG, 0, regVCN_UE_ERR_STATUS_LO_JPEG1S, regVCN_UE_ERR_STATUS_HI_JPEG1S),
+	1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "JPEG1S"},
+	{AMDGPU_RAS_REG_ENTRY(JPEG, 0, regVCN_UE_ERR_STATUS_LO_JPEG1D, regVCN_UE_ERR_STATUS_HI_JPEG1D),
+	1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "JPEG1D"},
+	{AMDGPU_RAS_REG_ENTRY(JPEG, 0, regVCN_UE_ERR_STATUS_LO_JPEG2S, regVCN_UE_ERR_STATUS_HI_JPEG2S),
+	1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "JPEG2S"},
+	{AMDGPU_RAS_REG_ENTRY(JPEG, 0, regVCN_UE_ERR_STATUS_LO_JPEG2D, regVCN_UE_ERR_STATUS_HI_JPEG2D),
+	1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "JPEG2D"},
+	{AMDGPU_RAS_REG_ENTRY(JPEG, 0, regVCN_UE_ERR_STATUS_LO_JPEG3S, regVCN_UE_ERR_STATUS_HI_JPEG3S),
+	1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "JPEG3S"},
+	{AMDGPU_RAS_REG_ENTRY(JPEG, 0, regVCN_UE_ERR_STATUS_LO_JPEG3D, regVCN_UE_ERR_STATUS_HI_JPEG3D),
+	1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "JPEG3D"},
+	{AMDGPU_RAS_REG_ENTRY(JPEG, 0, regVCN_UE_ERR_STATUS_LO_JPEG4S, regVCN_UE_ERR_STATUS_HI_JPEG4S),
+	1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "JPEG4S"},
+	{AMDGPU_RAS_REG_ENTRY(JPEG, 0, regVCN_UE_ERR_STATUS_LO_JPEG4D, regVCN_UE_ERR_STATUS_HI_JPEG4D),
+	1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "JPEG4D"},
+	{AMDGPU_RAS_REG_ENTRY(JPEG, 0, regVCN_UE_ERR_STATUS_LO_JPEG5S, regVCN_UE_ERR_STATUS_HI_JPEG5S),
+	1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "JPEG5S"},
+	{AMDGPU_RAS_REG_ENTRY(JPEG, 0, regVCN_UE_ERR_STATUS_LO_JPEG5D, regVCN_UE_ERR_STATUS_HI_JPEG5D),
+	1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "JPEG5D"},
+	{AMDGPU_RAS_REG_ENTRY(JPEG, 0, regVCN_UE_ERR_STATUS_LO_JPEG6S, regVCN_UE_ERR_STATUS_HI_JPEG6S),
+	1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "JPEG6S"},
+	{AMDGPU_RAS_REG_ENTRY(JPEG, 0, regVCN_UE_ERR_STATUS_LO_JPEG6D, regVCN_UE_ERR_STATUS_HI_JPEG6D),
+	1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "JPEG6D"},
+	{AMDGPU_RAS_REG_ENTRY(JPEG, 0, regVCN_UE_ERR_STATUS_LO_JPEG7S, regVCN_UE_ERR_STATUS_HI_JPEG7S),
+	1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "JPEG7S"},
+	{AMDGPU_RAS_REG_ENTRY(JPEG, 0, regVCN_UE_ERR_STATUS_LO_JPEG7D, regVCN_UE_ERR_STATUS_HI_JPEG7D),
+	1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "JPEG7D"},
+};
+
+static void jpeg_v4_0_3_inst_query_ras_error_count(struct amdgpu_device *adev,
+						   uint32_t jpeg_inst,
+						   void *ras_err_status)
+{
+	struct ras_err_data *err_data = (struct ras_err_data *)ras_err_status;
+
+	/* jpeg v4_0_3 only support uncorrectable errors */
+	amdgpu_ras_inst_query_ras_error_count(adev,
+			jpeg_v4_0_3_ue_reg_list,
+			ARRAY_SIZE(jpeg_v4_0_3_ue_reg_list),
+			NULL, 0, GET_INST(VCN, jpeg_inst),
+			AMDGPU_RAS_ERROR__MULTI_UNCORRECTABLE,
+			&err_data->ue_count);
+}
+
+static void jpeg_v4_0_3_query_ras_error_count(struct amdgpu_device *adev,
+					      void *ras_err_status)
+{
+	uint32_t i;
+
+	if (!amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__JPEG)) {
+		dev_warn(adev->dev, "JPEG RAS is not supported\n");
+		return;
+	}
+
+	for (i = 0; i < adev->jpeg.num_jpeg_inst; i++)
+		jpeg_v4_0_3_inst_query_ras_error_count(adev, i, ras_err_status);
+}
+
+static void jpeg_v4_0_3_inst_reset_ras_error_count(struct amdgpu_device *adev,
+						   uint32_t jpeg_inst)
+{
+	amdgpu_ras_inst_reset_ras_error_count(adev,
+			jpeg_v4_0_3_ue_reg_list,
+			ARRAY_SIZE(jpeg_v4_0_3_ue_reg_list),
+			GET_INST(VCN, jpeg_inst));
+}
+
+static void jpeg_v4_0_3_reset_ras_error_count(struct amdgpu_device *adev)
+{
+	uint32_t i;
+
+	if (!amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__JPEG)) {
+		dev_warn(adev->dev, "JPEG RAS is not supported\n");
+		return;
+	}
+
+	for (i = 0; i < adev->jpeg.num_jpeg_inst; i++)
+		jpeg_v4_0_3_inst_reset_ras_error_count(adev, i);
+}
+
+static uint32_t jpeg_v4_0_3_query_poison_by_instance(struct amdgpu_device *adev,
+		uint32_t instance, uint32_t sub_block)
+{
+	uint32_t poison_stat = 0, reg_value = 0;
+
+	switch (sub_block) {
+	case AMDGPU_JPEG_V4_0_3_JPEG0:
+		reg_value = RREG32_SOC15(JPEG, instance, regUVD_RAS_JPEG0_STATUS);
+		poison_stat = REG_GET_FIELD(reg_value, UVD_RAS_JPEG0_STATUS, POISONED_PF);
+		break;
+	case AMDGPU_JPEG_V4_0_3_JPEG1:
+		reg_value = RREG32_SOC15(JPEG, instance, regUVD_RAS_JPEG1_STATUS);
+		poison_stat = REG_GET_FIELD(reg_value, UVD_RAS_JPEG1_STATUS, POISONED_PF);
+		break;
+	default:
+		break;
+	}
+
+	if (poison_stat)
+		dev_info(adev->dev, "Poison detected in JPEG%d sub_block%d\n",
+			instance, sub_block);
+
+	return poison_stat;
+}
+
+static bool jpeg_v4_0_3_query_ras_poison_status(struct amdgpu_device *adev)
+{
+	uint32_t inst = 0, sub = 0, poison_stat = 0;
+
+	for (inst = 0; inst < adev->jpeg.num_jpeg_inst; inst++)
+		for (sub = 0; sub < AMDGPU_JPEG_V4_0_3_MAX_SUB_BLOCK; sub++)
+			poison_stat +=
+			jpeg_v4_0_3_query_poison_by_instance(adev, inst, sub);
+
+	return !!poison_stat;
+}
+
+static const struct amdgpu_ras_block_hw_ops jpeg_v4_0_3_ras_hw_ops = {
+	.query_ras_error_count = jpeg_v4_0_3_query_ras_error_count,
+	.reset_ras_error_count = jpeg_v4_0_3_reset_ras_error_count,
+	.query_poison_status = jpeg_v4_0_3_query_ras_poison_status,
+};
+
+static int jpeg_v4_0_3_aca_bank_parser(struct aca_handle *handle, struct aca_bank *bank,
+				      enum aca_smu_type type, void *data)
+{
+	struct aca_bank_info info;
+	u64 misc0;
+	int ret;
+
+	ret = aca_bank_info_decode(bank, &info);
+	if (ret)
+		return ret;
+
+	misc0 = bank->regs[ACA_REG_IDX_MISC0];
+	switch (type) {
+	case ACA_SMU_TYPE_UE:
+		bank->aca_err_type = ACA_ERROR_TYPE_UE;
+		ret = aca_error_cache_log_bank_error(handle, &info, ACA_ERROR_TYPE_UE,
+						     1ULL);
+		break;
+	case ACA_SMU_TYPE_CE:
+		bank->aca_err_type = ACA_ERROR_TYPE_CE;
+		ret = aca_error_cache_log_bank_error(handle, &info, bank->aca_err_type,
+						     ACA_REG__MISC0__ERRCNT(misc0));
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return ret;
+}
+
+/* reference to smu driver if header file */
+static int jpeg_v4_0_3_err_codes[] = {
+	16, 17, 18, 19, 20, 21, 22, 23, /* JPEG[0-7][S|D] */
+	24, 25, 26, 27, 28, 29, 30, 31
+};
+
+static bool jpeg_v4_0_3_aca_bank_is_valid(struct aca_handle *handle, struct aca_bank *bank,
+					 enum aca_smu_type type, void *data)
+{
+	u32 instlo;
+
+	instlo = ACA_REG__IPID__INSTANCEIDLO(bank->regs[ACA_REG_IDX_IPID]);
+	instlo &= GENMASK(31, 1);
+
+	if (instlo != mmSMNAID_AID0_MCA_SMU)
+		return false;
+
+	if (aca_bank_check_error_codes(handle->adev, bank,
+				       jpeg_v4_0_3_err_codes,
+				       ARRAY_SIZE(jpeg_v4_0_3_err_codes)))
+		return false;
+
+	return true;
+}
+
+static const struct aca_bank_ops jpeg_v4_0_3_aca_bank_ops = {
+	.aca_bank_parser = jpeg_v4_0_3_aca_bank_parser,
+	.aca_bank_is_valid = jpeg_v4_0_3_aca_bank_is_valid,
+};
+
+static const struct aca_info jpeg_v4_0_3_aca_info = {
+	.hwip = ACA_HWIP_TYPE_SMU,
+	.mask = ACA_ERROR_UE_MASK,
+	.bank_ops = &jpeg_v4_0_3_aca_bank_ops,
+};
+
+static int jpeg_v4_0_3_ras_late_init(struct amdgpu_device *adev, struct ras_common_if *ras_block)
+{
+	int r;
+
+	r = amdgpu_ras_block_late_init(adev, ras_block);
+	if (r)
+		return r;
+
+	if (amdgpu_ras_is_supported(adev, ras_block->block) &&
+		adev->jpeg.inst->ras_poison_irq.funcs) {
+		r = amdgpu_irq_get(adev, &adev->jpeg.inst->ras_poison_irq, 0);
+		if (r)
+			goto late_fini;
+	}
+
+	r = amdgpu_ras_bind_aca(adev, AMDGPU_RAS_BLOCK__JPEG,
+				&jpeg_v4_0_3_aca_info, NULL);
+	if (r)
+		goto late_fini;
+
+	return 0;
+
+late_fini:
+	amdgpu_ras_block_late_fini(adev, ras_block);
+
+	return r;
+}
+
+static struct amdgpu_jpeg_ras jpeg_v4_0_3_ras = {
+	.ras_block = {
+		.hw_ops = &jpeg_v4_0_3_ras_hw_ops,
+		.ras_late_init = jpeg_v4_0_3_ras_late_init,
+	},
+};
+
+static void jpeg_v4_0_3_set_ras_funcs(struct amdgpu_device *adev)
+{
+	adev->jpeg.ras = &jpeg_v4_0_3_ras;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_3.h b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_3.h
new file mode 100644
index 000000000000..2e110d04af84
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_3.h
@@ -0,0 +1,74 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __JPEG_V4_0_3_H__
+#define __JPEG_V4_0_3_H__
+
+#define regUVD_JRBC_EXTERNAL_REG_INTERNAL_OFFSET			0x1bfff
+#define regUVD_JPEG_GPCOM_CMD_INTERNAL_OFFSET				0x404d
+#define regUVD_JPEG_GPCOM_DATA0_INTERNAL_OFFSET				0x404e
+#define regUVD_JPEG_GPCOM_DATA1_INTERNAL_OFFSET				0x404f
+#define regUVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW_INTERNAL_OFFSET		0x40ab
+#define regUVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH_INTERNAL_OFFSET	0x40ac
+#define regUVD_LMI_JRBC_IB_VMID_INTERNAL_OFFSET				0x40a4
+#define regUVD_LMI_JPEG_VMID_INTERNAL_OFFSET				0x40a6
+#define regUVD_LMI_JRBC_IB_64BIT_BAR_LOW_INTERNAL_OFFSET		0x40b6
+#define regUVD_LMI_JRBC_IB_64BIT_BAR_HIGH_INTERNAL_OFFSET		0x40b7
+#define regUVD_JRBC_IB_SIZE_INTERNAL_OFFSET				0x4082
+#define regUVD_LMI_JRBC_RB_MEM_RD_64BIT_BAR_LOW_INTERNAL_OFFSET		0x42d4
+#define regUVD_LMI_JRBC_RB_MEM_RD_64BIT_BAR_HIGH_INTERNAL_OFFSET	0x42d5
+#define regUVD_JRBC_RB_COND_RD_TIMER_INTERNAL_OFFSET			0x4085
+#define regUVD_JRBC_RB_REF_DATA_INTERNAL_OFFSET				0x4084
+#define regUVD_JRBC_STATUS_INTERNAL_OFFSET				0x4089
+#define regUVD_JPEG_PITCH_INTERNAL_OFFSET				0x4043
+#define regUVD_JRBC0_UVD_JRBC_SCRATCH0_INTERNAL_OFFSET			0x4094
+#define regUVD_JRBC_EXTERNAL_MCM_ADDR_INTERNAL_OFFSET			0x1bffe
+
+#define JRBC_DEC_EXTERNAL_REG_WRITE_ADDR				0x18000
+
+enum amdgpu_jpeg_v4_0_3_sub_block {
+	AMDGPU_JPEG_V4_0_3_JPEG0 = 0,
+	AMDGPU_JPEG_V4_0_3_JPEG1,
+
+	AMDGPU_JPEG_V4_0_3_MAX_SUB_BLOCK,
+};
+
+extern const struct amdgpu_ip_block_version jpeg_v4_0_3_ip_block;
+
+void jpeg_v4_0_3_dec_ring_emit_ib(struct amdgpu_ring *ring,
+				struct amdgpu_job *job,
+				struct amdgpu_ib *ib,
+				uint32_t flags);
+void jpeg_v4_0_3_dec_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 seq,
+				unsigned int flags);
+void jpeg_v4_0_3_dec_ring_emit_vm_flush(struct amdgpu_ring *ring,
+					unsigned int vmid, uint64_t pd_addr);
+void jpeg_v4_0_3_ring_emit_hdp_flush(struct amdgpu_ring *ring);
+void jpeg_v4_0_3_dec_ring_nop(struct amdgpu_ring *ring, uint32_t count);
+void jpeg_v4_0_3_dec_ring_insert_start(struct amdgpu_ring *ring);
+void jpeg_v4_0_3_dec_ring_insert_end(struct amdgpu_ring *ring);
+void jpeg_v4_0_3_dec_ring_emit_wreg(struct amdgpu_ring *ring, uint32_t reg, uint32_t val);
+void jpeg_v4_0_3_dec_ring_emit_reg_wait(struct amdgpu_ring *ring, uint32_t reg,
+					uint32_t val, uint32_t mask);
+
+#endif /* __JPEG_V4_0_3_H__ */
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_5.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_5.c
new file mode 100644
index 000000000000..54fd9c800c40
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_5.c
@@ -0,0 +1,872 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "amdgpu.h"
+#include "amdgpu_jpeg.h"
+#include "amdgpu_pm.h"
+#include "soc15.h"
+#include "soc15d.h"
+#include "jpeg_v2_0.h"
+#include "jpeg_v4_0_5.h"
+#include "mmsch_v4_0.h"
+
+#include "vcn/vcn_4_0_5_offset.h"
+#include "vcn/vcn_4_0_5_sh_mask.h"
+#include "ivsrcid/vcn/irqsrcs_vcn_4_0.h"
+
+#define mmUVD_DPG_LMA_CTL						regUVD_DPG_LMA_CTL
+#define mmUVD_DPG_LMA_CTL_BASE_IDX					regUVD_DPG_LMA_CTL_BASE_IDX
+#define mmUVD_DPG_LMA_DATA						regUVD_DPG_LMA_DATA
+#define mmUVD_DPG_LMA_DATA_BASE_IDX					regUVD_DPG_LMA_DATA_BASE_IDX
+
+#define regUVD_JPEG_PITCH_INTERNAL_OFFSET		0x401f
+#define regJPEG_DEC_GFX10_ADDR_CONFIG_INTERNAL_OFFSET	0x4026
+#define regJPEG_SYS_INT_EN_INTERNAL_OFFSET		0x4141
+#define regJPEG_CGC_CTRL_INTERNAL_OFFSET		0x4161
+#define regJPEG_CGC_GATE_INTERNAL_OFFSET		0x4160
+#define regUVD_NO_OP_INTERNAL_OFFSET			0x0029
+
+static const struct amdgpu_hwip_reg_entry jpeg_reg_list_4_0_5[] = {
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JPEG_POWER_STATUS),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JPEG_INT_STAT),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC_RB_RPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC_RB_WPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC_RB_CNTL),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC_RB_SIZE),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC_STATUS),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regJPEG_DEC_ADDR_MODE),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regJPEG_DEC_GFX10_ADDR_CONFIG),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regJPEG_DEC_Y_GFX10_TILING_SURFACE),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regJPEG_DEC_UV_GFX10_TILING_SURFACE),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JPEG_PITCH),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JPEG_UV_PITCH),
+};
+
+static void jpeg_v4_0_5_set_dec_ring_funcs(struct amdgpu_device *adev);
+static void jpeg_v4_0_5_set_irq_funcs(struct amdgpu_device *adev);
+static int jpeg_v4_0_5_set_powergating_state(struct amdgpu_ip_block *ip_block,
+				enum amd_powergating_state state);
+static void jpeg_v4_0_5_dec_ring_set_wptr(struct amdgpu_ring *ring);
+
+static int amdgpu_ih_clientid_jpeg[] = {
+	SOC15_IH_CLIENTID_VCN,
+	SOC15_IH_CLIENTID_VCN1
+};
+
+
+
+/**
+ * jpeg_v4_0_5_early_init - set function pointers
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Set ring and irq function pointers
+ */
+static int jpeg_v4_0_5_early_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	switch (amdgpu_ip_version(adev, UVD_HWIP, 0)) {
+	case IP_VERSION(4, 0, 5):
+		adev->jpeg.num_jpeg_inst = 1;
+		break;
+	case IP_VERSION(4, 0, 6):
+		adev->jpeg.num_jpeg_inst = 2;
+		break;
+	default:
+		DRM_DEV_ERROR(adev->dev,
+			"Failed to init vcn ip block(UVD_HWIP:0x%x)\n",
+			amdgpu_ip_version(adev, UVD_HWIP, 0));
+		return -EINVAL;
+	}
+
+	adev->jpeg.num_jpeg_rings = 1;
+
+	jpeg_v4_0_5_set_dec_ring_funcs(adev);
+	jpeg_v4_0_5_set_irq_funcs(adev);
+
+	return 0;
+}
+
+/**
+ * jpeg_v4_0_5_sw_init - sw init for JPEG block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Load firmware and sw initialization
+ */
+static int jpeg_v4_0_5_sw_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	struct amdgpu_ring *ring;
+	int r, i;
+
+	for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
+		if (adev->jpeg.harvest_config & (1 << i))
+			continue;
+
+		/* JPEG TRAP */
+		r = amdgpu_irq_add_id(adev, amdgpu_ih_clientid_jpeg[i],
+				VCN_4_0__SRCID__JPEG_DECODE, &adev->jpeg.inst[i].irq);
+		if (r)
+			return r;
+
+		/* JPEG DJPEG POISON EVENT */
+		r = amdgpu_irq_add_id(adev, amdgpu_ih_clientid_jpeg[i],
+			VCN_4_0__SRCID_DJPEG0_POISON, &adev->jpeg.inst[i].irq);
+		if (r)
+			return r;
+
+		/* JPEG EJPEG POISON EVENT */
+		r = amdgpu_irq_add_id(adev, amdgpu_ih_clientid_jpeg[i],
+			VCN_4_0__SRCID_EJPEG0_POISON, &adev->jpeg.inst[i].irq);
+		if (r)
+			return r;
+	}
+
+	r = amdgpu_jpeg_sw_init(adev);
+	if (r)
+		return r;
+
+	r = amdgpu_jpeg_resume(adev);
+	if (r)
+		return r;
+
+	for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
+		if (adev->jpeg.harvest_config & (1 << i))
+			continue;
+
+		ring = adev->jpeg.inst[i].ring_dec;
+		ring->use_doorbell = true;
+		ring->vm_hub = AMDGPU_MMHUB0(0);
+		ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 1 + 8 * i;
+		sprintf(ring->name, "jpeg_dec_%d", i);
+		r = amdgpu_ring_init(adev, ring, 512, &adev->jpeg.inst[i].irq,
+				     0, AMDGPU_RING_PRIO_DEFAULT, NULL);
+		if (r)
+			return r;
+
+		adev->jpeg.internal.jpeg_pitch[0] = regUVD_JPEG_PITCH_INTERNAL_OFFSET;
+		adev->jpeg.inst[i].external.jpeg_pitch[0] = SOC15_REG_OFFSET(JPEG, i, regUVD_JPEG_PITCH);
+	}
+
+	r = amdgpu_jpeg_reg_dump_init(adev, jpeg_reg_list_4_0_5, ARRAY_SIZE(jpeg_reg_list_4_0_5));
+	if (r)
+		return r;
+
+	adev->jpeg.supported_reset =
+		amdgpu_get_soft_full_reset_mask(&adev->jpeg.inst[0].ring_dec[0]);
+	if (!amdgpu_sriov_vf(adev))
+		adev->jpeg.supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE;
+	r = amdgpu_jpeg_sysfs_reset_mask_init(adev);
+	if (r)
+		return r;
+
+	return 0;
+}
+
+/**
+ * jpeg_v4_0_5_sw_fini - sw fini for JPEG block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * JPEG suspend and free up sw allocation
+ */
+static int jpeg_v4_0_5_sw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int r;
+
+	r = amdgpu_jpeg_suspend(adev);
+	if (r)
+		return r;
+
+	amdgpu_jpeg_sysfs_reset_mask_fini(adev);
+	r = amdgpu_jpeg_sw_fini(adev);
+
+	return r;
+}
+
+/**
+ * jpeg_v4_0_5_hw_init - start and test JPEG block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ */
+static int jpeg_v4_0_5_hw_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	struct amdgpu_ring *ring;
+	int i, r = 0;
+
+	// TODO: Enable ring test with DPG support
+	if (adev->pg_flags & AMD_PG_SUPPORT_JPEG_DPG) {
+		return 0;
+	}
+
+	for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
+		if (adev->jpeg.harvest_config & (1 << i))
+			continue;
+
+		ring = adev->jpeg.inst[i].ring_dec;
+		r = amdgpu_ring_test_helper(ring);
+		if (r)
+			return r;
+	}
+
+	return 0;
+}
+
+/**
+ * jpeg_v4_0_5_hw_fini - stop the hardware block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Stop the JPEG block, mark ring as not ready any more
+ */
+static int jpeg_v4_0_5_hw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i;
+
+	cancel_delayed_work_sync(&adev->jpeg.idle_work);
+
+	for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
+		if (adev->jpeg.harvest_config & (1 << i))
+			continue;
+
+		if (!amdgpu_sriov_vf(adev)) {
+			if (adev->jpeg.cur_state != AMD_PG_STATE_GATE &&
+			    RREG32_SOC15(JPEG, i, regUVD_JRBC_STATUS))
+				jpeg_v4_0_5_set_powergating_state(ip_block, AMD_PG_STATE_GATE);
+		}
+	}
+	return 0;
+}
+
+/**
+ * jpeg_v4_0_5_suspend - suspend JPEG block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * HW fini and suspend JPEG block
+ */
+static int jpeg_v4_0_5_suspend(struct amdgpu_ip_block *ip_block)
+{
+	int r;
+
+	r = jpeg_v4_0_5_hw_fini(ip_block);
+	if (r)
+		return r;
+
+	r = amdgpu_jpeg_suspend(ip_block->adev);
+
+	return r;
+}
+
+/**
+ * jpeg_v4_0_5_resume - resume JPEG block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Resume firmware and hw init JPEG block
+ */
+static int jpeg_v4_0_5_resume(struct amdgpu_ip_block *ip_block)
+{
+	int r;
+
+	r = amdgpu_jpeg_resume(ip_block->adev);
+	if (r)
+		return r;
+
+	r = jpeg_v4_0_5_hw_init(ip_block);
+
+	return r;
+}
+
+static void jpeg_v4_0_5_disable_clock_gating(struct amdgpu_device *adev, int inst)
+{
+	uint32_t data = 0;
+
+	data = RREG32_SOC15(JPEG, inst, regJPEG_CGC_CTRL);
+	if (adev->cg_flags & AMD_CG_SUPPORT_JPEG_MGCG) {
+		data |= 1 << JPEG_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
+		data &= (~JPEG_CGC_CTRL__JPEG_DEC_MODE_MASK);
+	} else {
+		data &= ~JPEG_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
+	}
+
+	data |= 1 << JPEG_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT;
+	data |= 4 << JPEG_CGC_CTRL__CLK_OFF_DELAY__SHIFT;
+	WREG32_SOC15(JPEG, inst, regJPEG_CGC_CTRL, data);
+
+	data = RREG32_SOC15(JPEG, inst, regJPEG_CGC_GATE);
+	data &= ~(JPEG_CGC_GATE__JPEG_DEC_MASK
+		| JPEG_CGC_GATE__JPEG2_DEC_MASK
+		| JPEG_CGC_GATE__JMCIF_MASK
+		| JPEG_CGC_GATE__JRBBM_MASK);
+	WREG32_SOC15(JPEG, inst, regJPEG_CGC_GATE, data);
+}
+
+static void jpeg_v4_0_5_enable_clock_gating(struct amdgpu_device *adev, int inst)
+{
+	uint32_t data = 0;
+
+	data = RREG32_SOC15(JPEG, inst, regJPEG_CGC_CTRL);
+	if (adev->cg_flags & AMD_CG_SUPPORT_JPEG_MGCG) {
+		data |= 1 << JPEG_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
+		data |= JPEG_CGC_CTRL__JPEG_DEC_MODE_MASK;
+	} else {
+		data &= ~JPEG_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
+	}
+
+	data |= 1 << JPEG_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT;
+	data |= 4 << JPEG_CGC_CTRL__CLK_OFF_DELAY__SHIFT;
+	WREG32_SOC15(JPEG, inst, regJPEG_CGC_CTRL, data);
+
+	data = RREG32_SOC15(JPEG, inst, regJPEG_CGC_GATE);
+	data |= (JPEG_CGC_GATE__JPEG_DEC_MASK
+		|JPEG_CGC_GATE__JPEG2_DEC_MASK
+		|JPEG_CGC_GATE__JMCIF_MASK
+		|JPEG_CGC_GATE__JRBBM_MASK);
+	WREG32_SOC15(JPEG, inst, regJPEG_CGC_GATE, data);
+}
+
+static void jpeg_engine_4_0_5_dpg_clock_gating_mode(struct amdgpu_device *adev,
+			int inst_idx, uint8_t indirect)
+{
+	uint32_t data = 0;
+
+	if (adev->cg_flags & AMD_CG_SUPPORT_JPEG_MGCG)
+		data |= 1 << JPEG_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
+	else
+		data |= 0 << JPEG_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
+
+	data |= 1 << JPEG_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT;
+	data |= 4 << JPEG_CGC_CTRL__CLK_OFF_DELAY__SHIFT;
+	WREG32_SOC15_JPEG_DPG_MODE(inst_idx, regJPEG_CGC_CTRL_INTERNAL_OFFSET, data, indirect);
+
+	data = 0;
+	WREG32_SOC15_JPEG_DPG_MODE(inst_idx, regJPEG_CGC_GATE_INTERNAL_OFFSET,
+				data, indirect);
+}
+
+static int jpeg_v4_0_5_disable_static_power_gating(struct amdgpu_device *adev, int inst)
+{
+	if (adev->pg_flags & AMD_PG_SUPPORT_JPEG) {
+		WREG32(SOC15_REG_OFFSET(JPEG, inst, regUVD_IPX_DLDO_CONFIG),
+			1 << UVD_IPX_DLDO_CONFIG__ONO1_PWR_CONFIG__SHIFT);
+		SOC15_WAIT_ON_RREG(JPEG, inst, regUVD_IPX_DLDO_STATUS,
+			0, UVD_IPX_DLDO_STATUS__ONO1_PWR_STATUS_MASK);
+	}
+
+	/* disable anti hang mechanism */
+	WREG32_P(SOC15_REG_OFFSET(JPEG, inst, regUVD_JPEG_POWER_STATUS), 0,
+		~UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK);
+
+	/* keep the JPEG in static PG mode */
+	WREG32_P(SOC15_REG_OFFSET(JPEG, inst, regUVD_JPEG_POWER_STATUS), 0,
+		~UVD_JPEG_POWER_STATUS__JPEG_PG_MODE_MASK);
+
+	return 0;
+}
+
+static int jpeg_v4_0_5_enable_static_power_gating(struct amdgpu_device *adev, int inst)
+{
+	/* enable anti hang mechanism */
+	WREG32_P(SOC15_REG_OFFSET(JPEG, inst, regUVD_JPEG_POWER_STATUS),
+		UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK,
+		~UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK);
+
+	if (adev->pg_flags & AMD_PG_SUPPORT_JPEG) {
+		WREG32(SOC15_REG_OFFSET(JPEG, inst, regUVD_IPX_DLDO_CONFIG),
+			2 << UVD_IPX_DLDO_CONFIG__ONO1_PWR_CONFIG__SHIFT);
+		SOC15_WAIT_ON_RREG(JPEG, inst, regUVD_IPX_DLDO_STATUS,
+			1 << UVD_IPX_DLDO_STATUS__ONO1_PWR_STATUS__SHIFT,
+			UVD_IPX_DLDO_STATUS__ONO1_PWR_STATUS_MASK);
+	}
+
+	return 0;
+}
+
+/**
+ * jpeg_v4_0_5_start_dpg_mode - Jpeg start with dpg mode
+ *
+ * @adev: amdgpu_device pointer
+ * @inst_idx: instance number index
+ * @indirect: indirectly write sram
+ *
+ * Start JPEG block with dpg mode
+ */
+static void jpeg_v4_0_5_start_dpg_mode(struct amdgpu_device *adev, int inst_idx, bool indirect)
+{
+	struct amdgpu_ring *ring = adev->jpeg.inst[inst_idx].ring_dec;
+	uint32_t reg_data = 0;
+
+	/* enable anti hang mechanism */
+	reg_data = RREG32_SOC15(JPEG, inst_idx, regUVD_JPEG_POWER_STATUS);
+	reg_data &= ~UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK;
+	reg_data |=  0x1;
+	WREG32_SOC15(JPEG, inst_idx, regUVD_JPEG_POWER_STATUS, reg_data);
+
+	if (adev->pg_flags & AMD_PG_SUPPORT_JPEG) {
+		WREG32(SOC15_REG_OFFSET(JPEG, inst_idx, regUVD_IPX_DLDO_CONFIG),
+			2 << UVD_IPX_DLDO_CONFIG__ONO1_PWR_CONFIG__SHIFT);
+		SOC15_WAIT_ON_RREG(JPEG, inst_idx, regUVD_IPX_DLDO_STATUS,
+			1 << UVD_IPX_DLDO_STATUS__ONO1_PWR_STATUS__SHIFT,
+			UVD_IPX_DLDO_STATUS__ONO1_PWR_STATUS_MASK);
+	}
+
+	reg_data = RREG32_SOC15(JPEG, inst_idx, regUVD_JPEG_POWER_STATUS);
+	reg_data |= UVD_JPEG_POWER_STATUS__JPEG_PG_MODE_MASK;
+	WREG32_SOC15(JPEG, inst_idx, regUVD_JPEG_POWER_STATUS, reg_data);
+
+	if (indirect)
+		adev->jpeg.inst[inst_idx].dpg_sram_curr_addr =
+					(uint32_t *)adev->jpeg.inst[inst_idx].dpg_sram_cpu_addr;
+
+	jpeg_engine_4_0_5_dpg_clock_gating_mode(adev, inst_idx, indirect);
+
+	/* MJPEG global tiling registers */
+	WREG32_SOC15_JPEG_DPG_MODE(inst_idx, regJPEG_DEC_GFX10_ADDR_CONFIG_INTERNAL_OFFSET,
+	adev->gfx.config.gb_addr_config, indirect);
+	/* enable System Interrupt for JRBC */
+	WREG32_SOC15_JPEG_DPG_MODE(inst_idx, regJPEG_SYS_INT_EN_INTERNAL_OFFSET,
+	JPEG_SYS_INT_EN__DJRBC_MASK, indirect);
+
+	/* add nop to workaround PSP size check */
+	WREG32_SOC15_JPEG_DPG_MODE(inst_idx, regUVD_NO_OP_INTERNAL_OFFSET, 0, indirect);
+
+	if (indirect)
+		amdgpu_jpeg_psp_update_sram(adev, inst_idx, 0);
+
+	WREG32_SOC15(JPEG, inst_idx, regUVD_LMI_JRBC_RB_VMID, 0);
+	WREG32_SOC15(JPEG, inst_idx, regUVD_JRBC_RB_CNTL, (0x00000001L | 0x00000002L));
+	WREG32_SOC15(JPEG, inst_idx, regUVD_LMI_JRBC_RB_64BIT_BAR_LOW,
+		lower_32_bits(ring->gpu_addr));
+	WREG32_SOC15(JPEG, inst_idx, regUVD_LMI_JRBC_RB_64BIT_BAR_HIGH,
+		upper_32_bits(ring->gpu_addr));
+	WREG32_SOC15(JPEG, inst_idx, regUVD_JRBC_RB_RPTR, 0);
+	WREG32_SOC15(JPEG, inst_idx, regUVD_JRBC_RB_WPTR, 0);
+	WREG32_SOC15(JPEG, inst_idx, regUVD_JRBC_RB_CNTL, 0x00000002L);
+	WREG32_SOC15(JPEG, inst_idx, regUVD_JRBC_RB_SIZE, ring->ring_size / 4);
+	ring->wptr = RREG32_SOC15(JPEG, inst_idx, regUVD_JRBC_RB_WPTR);
+}
+
+/**
+ * jpeg_v4_0_5_stop_dpg_mode - Jpeg stop with dpg mode
+ *
+ * @adev: amdgpu_device pointer
+ * @inst_idx: instance number index
+ *
+ * Stop JPEG block with dpg mode
+ */
+static void jpeg_v4_0_5_stop_dpg_mode(struct amdgpu_device *adev, int inst_idx)
+{
+	uint32_t reg_data = 0;
+
+	reg_data = RREG32_SOC15(JPEG, inst_idx, regUVD_JPEG_POWER_STATUS);
+	reg_data &= ~UVD_JPEG_POWER_STATUS__JPEG_PG_MODE_MASK;
+	WREG32_SOC15(JPEG, inst_idx, regUVD_JPEG_POWER_STATUS, reg_data);
+
+}
+
+/**
+ * jpeg_v4_0_5_start - start JPEG block
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Setup and start the JPEG block
+ */
+static int jpeg_v4_0_5_start(struct amdgpu_device *adev)
+{
+	struct amdgpu_ring *ring;
+	int r, i;
+
+	if (adev->pm.dpm_enabled)
+		amdgpu_dpm_enable_jpeg(adev, true);
+
+	for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
+		if (adev->jpeg.harvest_config & (1 << i))
+			continue;
+
+		ring = adev->jpeg.inst[i].ring_dec;
+		/* doorbell programming is done for every playback */
+		adev->nbio.funcs->vcn_doorbell_range(adev, ring->use_doorbell,
+				(adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 8 * i, i);
+
+		WREG32_SOC15(VCN, i, regVCN_JPEG_DB_CTRL,
+			ring->doorbell_index << VCN_JPEG_DB_CTRL__OFFSET__SHIFT |
+			VCN_JPEG_DB_CTRL__EN_MASK);
+
+		if (adev->pg_flags & AMD_PG_SUPPORT_JPEG_DPG) {
+			jpeg_v4_0_5_start_dpg_mode(adev, i, adev->jpeg.indirect_sram);
+			continue;
+		}
+
+		/* disable power gating */
+		r = jpeg_v4_0_5_disable_static_power_gating(adev, i);
+		if (r)
+			return r;
+
+		/* JPEG disable CGC */
+		jpeg_v4_0_5_disable_clock_gating(adev, i);
+
+		/* MJPEG global tiling registers */
+		WREG32_SOC15(JPEG, i, regJPEG_DEC_GFX10_ADDR_CONFIG,
+			adev->gfx.config.gb_addr_config);
+
+		/* enable JMI channel */
+		WREG32_P(SOC15_REG_OFFSET(JPEG, i, regUVD_JMI_CNTL), 0,
+			~UVD_JMI_CNTL__SOFT_RESET_MASK);
+
+		/* enable System Interrupt for JRBC */
+		WREG32_P(SOC15_REG_OFFSET(JPEG, i, regJPEG_SYS_INT_EN),
+			JPEG_SYS_INT_EN__DJRBC_MASK,
+			~JPEG_SYS_INT_EN__DJRBC_MASK);
+
+		WREG32_SOC15(JPEG, i, regUVD_LMI_JRBC_RB_VMID, 0);
+		WREG32_SOC15(JPEG, i, regUVD_JRBC_RB_CNTL, (0x00000001L | 0x00000002L));
+		WREG32_SOC15(JPEG, i, regUVD_LMI_JRBC_RB_64BIT_BAR_LOW,
+			lower_32_bits(ring->gpu_addr));
+		WREG32_SOC15(JPEG, i, regUVD_LMI_JRBC_RB_64BIT_BAR_HIGH,
+			upper_32_bits(ring->gpu_addr));
+		WREG32_SOC15(JPEG, i, regUVD_JRBC_RB_RPTR, 0);
+		WREG32_SOC15(JPEG, i, regUVD_JRBC_RB_WPTR, 0);
+		WREG32_SOC15(JPEG, i, regUVD_JRBC_RB_CNTL, 0x00000002L);
+		WREG32_SOC15(JPEG, i, regUVD_JRBC_RB_SIZE, ring->ring_size / 4);
+		ring->wptr = RREG32_SOC15(JPEG, i, regUVD_JRBC_RB_WPTR);
+	}
+
+	return 0;
+}
+
+/**
+ * jpeg_v4_0_5_stop - stop JPEG block
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * stop the JPEG block
+ */
+static int jpeg_v4_0_5_stop(struct amdgpu_device *adev)
+{
+	int r, i;
+
+	for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
+		if (adev->jpeg.harvest_config & (1 << i))
+			continue;
+
+		if (adev->pg_flags & AMD_PG_SUPPORT_JPEG_DPG) {
+			jpeg_v4_0_5_stop_dpg_mode(adev, i);
+			continue;
+		}
+
+		/* reset JMI */
+		WREG32_P(SOC15_REG_OFFSET(JPEG, i, regUVD_JMI_CNTL),
+			UVD_JMI_CNTL__SOFT_RESET_MASK,
+			~UVD_JMI_CNTL__SOFT_RESET_MASK);
+
+		jpeg_v4_0_5_enable_clock_gating(adev, i);
+
+		/* enable power gating */
+		r = jpeg_v4_0_5_enable_static_power_gating(adev, i);
+		if (r)
+			return r;
+	}
+	if (adev->pm.dpm_enabled)
+		amdgpu_dpm_enable_jpeg(adev, false);
+
+	return 0;
+}
+
+/**
+ * jpeg_v4_0_5_dec_ring_get_rptr - get read pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Returns the current hardware read pointer
+ */
+static uint64_t jpeg_v4_0_5_dec_ring_get_rptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	return RREG32_SOC15(JPEG, ring->me, regUVD_JRBC_RB_RPTR);
+}
+
+/**
+ * jpeg_v4_0_5_dec_ring_get_wptr - get write pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Returns the current hardware write pointer
+ */
+static uint64_t jpeg_v4_0_5_dec_ring_get_wptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	if (ring->use_doorbell)
+		return *ring->wptr_cpu_addr;
+	else
+		return RREG32_SOC15(JPEG, ring->me, regUVD_JRBC_RB_WPTR);
+}
+
+/**
+ * jpeg_v4_0_5_dec_ring_set_wptr - set write pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Commits the write pointer to the hardware
+ */
+static void jpeg_v4_0_5_dec_ring_set_wptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	if (ring->use_doorbell) {
+		*ring->wptr_cpu_addr = lower_32_bits(ring->wptr);
+		WDOORBELL32(ring->doorbell_index, lower_32_bits(ring->wptr));
+	} else {
+		WREG32_SOC15(JPEG, ring->me, regUVD_JRBC_RB_WPTR, lower_32_bits(ring->wptr));
+	}
+}
+
+static bool jpeg_v4_0_5_is_idle(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, ret = 1;
+
+	for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
+		if (adev->jpeg.harvest_config & (1 << i))
+			continue;
+
+		ret &= (((RREG32_SOC15(JPEG, i, regUVD_JRBC_STATUS) &
+			UVD_JRBC_STATUS__RB_JOB_DONE_MASK) ==
+			UVD_JRBC_STATUS__RB_JOB_DONE_MASK));
+	}
+	return ret;
+}
+
+static int jpeg_v4_0_5_wait_for_idle(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i;
+
+	for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
+		if (adev->jpeg.harvest_config & (1 << i))
+			continue;
+
+		return SOC15_WAIT_ON_RREG(JPEG, i, regUVD_JRBC_STATUS,
+			UVD_JRBC_STATUS__RB_JOB_DONE_MASK,
+			UVD_JRBC_STATUS__RB_JOB_DONE_MASK);
+	}
+
+	return 0;
+}
+
+static int jpeg_v4_0_5_set_clockgating_state(struct amdgpu_ip_block *ip_block,
+					  enum amd_clockgating_state state)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	bool enable = state == AMD_CG_STATE_GATE;
+	int i;
+
+	for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
+		if (adev->jpeg.harvest_config & (1 << i))
+			continue;
+
+		if (enable) {
+			if (!jpeg_v4_0_5_is_idle(ip_block))
+				return -EBUSY;
+
+			jpeg_v4_0_5_enable_clock_gating(adev, i);
+		} else {
+			jpeg_v4_0_5_disable_clock_gating(adev, i);
+		}
+	}
+
+	return 0;
+}
+
+static int jpeg_v4_0_5_set_powergating_state(struct amdgpu_ip_block *ip_block,
+					  enum amd_powergating_state state)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int ret;
+
+	if (amdgpu_sriov_vf(adev)) {
+		adev->jpeg.cur_state = AMD_PG_STATE_UNGATE;
+		return 0;
+	}
+
+	if (state == adev->jpeg.cur_state)
+		return 0;
+
+	if (state == AMD_PG_STATE_GATE)
+		ret = jpeg_v4_0_5_stop(adev);
+	else
+		ret = jpeg_v4_0_5_start(adev);
+
+	if (!ret)
+		adev->jpeg.cur_state = state;
+
+	return ret;
+}
+
+static int jpeg_v4_0_5_process_interrupt(struct amdgpu_device *adev,
+				      struct amdgpu_irq_src *source,
+				      struct amdgpu_iv_entry *entry)
+{
+	uint32_t ip_instance;
+
+	DRM_DEBUG("IH: JPEG TRAP\n");
+
+	switch (entry->client_id) {
+	case SOC15_IH_CLIENTID_VCN:
+		ip_instance = 0;
+		break;
+	case SOC15_IH_CLIENTID_VCN1:
+		ip_instance = 1;
+		break;
+	default:
+		DRM_ERROR("Unhandled client id: %d\n", entry->client_id);
+		return 0;
+	}
+
+	switch (entry->src_id) {
+	case VCN_4_0__SRCID__JPEG_DECODE:
+		amdgpu_fence_process(adev->jpeg.inst[ip_instance].ring_dec);
+		break;
+	case VCN_4_0__SRCID_DJPEG0_POISON:
+	case VCN_4_0__SRCID_EJPEG0_POISON:
+		amdgpu_jpeg_process_poison_irq(adev, source, entry);
+		break;
+	default:
+		DRM_DEV_ERROR(adev->dev, "Unhandled interrupt: %d %d\n",
+			  entry->src_id, entry->src_data[0]);
+		break;
+	}
+
+	return 0;
+}
+
+static int jpeg_v4_0_5_ring_reset(struct amdgpu_ring *ring,
+				  unsigned int vmid,
+				  struct amdgpu_fence *timedout_fence)
+{
+	int r;
+
+	amdgpu_ring_reset_helper_begin(ring, timedout_fence);
+	r = jpeg_v4_0_5_stop(ring->adev);
+	if (r)
+		return r;
+	r = jpeg_v4_0_5_start(ring->adev);
+	if (r)
+		return r;
+	return amdgpu_ring_reset_helper_end(ring, timedout_fence);
+}
+
+static const struct amd_ip_funcs jpeg_v4_0_5_ip_funcs = {
+	.name = "jpeg_v4_0_5",
+	.early_init = jpeg_v4_0_5_early_init,
+	.sw_init = jpeg_v4_0_5_sw_init,
+	.sw_fini = jpeg_v4_0_5_sw_fini,
+	.hw_init = jpeg_v4_0_5_hw_init,
+	.hw_fini = jpeg_v4_0_5_hw_fini,
+	.suspend = jpeg_v4_0_5_suspend,
+	.resume = jpeg_v4_0_5_resume,
+	.is_idle = jpeg_v4_0_5_is_idle,
+	.wait_for_idle = jpeg_v4_0_5_wait_for_idle,
+	.set_clockgating_state = jpeg_v4_0_5_set_clockgating_state,
+	.set_powergating_state = jpeg_v4_0_5_set_powergating_state,
+	.dump_ip_state = amdgpu_jpeg_dump_ip_state,
+	.print_ip_state = amdgpu_jpeg_print_ip_state,
+};
+
+static const struct amdgpu_ring_funcs jpeg_v4_0_5_dec_ring_vm_funcs = {
+	.type = AMDGPU_RING_TYPE_VCN_JPEG,
+	.align_mask = 0xf,
+	.get_rptr = jpeg_v4_0_5_dec_ring_get_rptr,
+	.get_wptr = jpeg_v4_0_5_dec_ring_get_wptr,
+	.set_wptr = jpeg_v4_0_5_dec_ring_set_wptr,
+	.parse_cs = amdgpu_jpeg_dec_parse_cs,
+	.emit_frame_size =
+		SOC15_FLUSH_GPU_TLB_NUM_WREG * 6 +
+		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 8 +
+		8 + /* jpeg_v4_0_5_dec_ring_emit_vm_flush */
+		18 + 18 + /* jpeg_v4_0_5_dec_ring_emit_fence x2 vm fence */
+		8 + 16,
+	.emit_ib_size = 22, /* jpeg_v4_0_5_dec_ring_emit_ib */
+	.emit_ib = jpeg_v2_0_dec_ring_emit_ib,
+	.emit_fence = jpeg_v2_0_dec_ring_emit_fence,
+	.emit_vm_flush = jpeg_v2_0_dec_ring_emit_vm_flush,
+	.test_ring = amdgpu_jpeg_dec_ring_test_ring,
+	.test_ib = amdgpu_jpeg_dec_ring_test_ib,
+	.insert_nop = jpeg_v2_0_dec_ring_nop,
+	.insert_start = jpeg_v2_0_dec_ring_insert_start,
+	.insert_end = jpeg_v2_0_dec_ring_insert_end,
+	.pad_ib = amdgpu_ring_generic_pad_ib,
+	.begin_use = amdgpu_jpeg_ring_begin_use,
+	.end_use = amdgpu_jpeg_ring_end_use,
+	.emit_wreg = jpeg_v2_0_dec_ring_emit_wreg,
+	.emit_reg_wait = jpeg_v2_0_dec_ring_emit_reg_wait,
+	.emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper,
+	.reset = jpeg_v4_0_5_ring_reset,
+};
+
+static void jpeg_v4_0_5_set_dec_ring_funcs(struct amdgpu_device *adev)
+{
+	int i;
+
+	for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
+		if (adev->jpeg.harvest_config & (1 << i))
+			continue;
+
+		adev->jpeg.inst[i].ring_dec->funcs = &jpeg_v4_0_5_dec_ring_vm_funcs;
+		adev->jpeg.inst[i].ring_dec->me = i;
+	}
+}
+
+static const struct amdgpu_irq_src_funcs jpeg_v4_0_5_irq_funcs = {
+	.process = jpeg_v4_0_5_process_interrupt,
+};
+
+static void jpeg_v4_0_5_set_irq_funcs(struct amdgpu_device *adev)
+{
+	int i;
+
+	for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
+		if (adev->jpeg.harvest_config & (1 << i))
+			continue;
+
+		adev->jpeg.inst[i].irq.num_types = 1;
+		adev->jpeg.inst[i].irq.funcs = &jpeg_v4_0_5_irq_funcs;
+	}
+}
+
+const struct amdgpu_ip_block_version jpeg_v4_0_5_ip_block = {
+	.type = AMD_IP_BLOCK_TYPE_JPEG,
+	.major = 4,
+	.minor = 0,
+	.rev = 5,
+	.funcs = &jpeg_v4_0_5_ip_funcs,
+};
+
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_5.h b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_5.h
new file mode 100644
index 000000000000..c5eee572079c
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_5.h
@@ -0,0 +1,35 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __JPEG_V4_0_5_H__
+#define __JPEG_V4_0_5_H__
+
+enum amdgpu_jpeg_v4_0_5_sub_block {
+	AMDGPU_JPEG_V4_0_5_JPEG0 = 0,
+
+	AMDGPU_JPEG_V4_0_5_MAX_SUB_BLOCK,
+};
+
+extern const struct amdgpu_ip_block_version jpeg_v4_0_5_ip_block;
+
+#endif /* __JPEG_V4_0_H__ */
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_0.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_0.c
new file mode 100644
index 000000000000..46bf15dce2bd
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_0.c
@@ -0,0 +1,733 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "amdgpu.h"
+#include "amdgpu_jpeg.h"
+#include "amdgpu_pm.h"
+#include "soc15.h"
+#include "soc15d.h"
+#include "jpeg_v2_0.h"
+#include "jpeg_v4_0_3.h"
+
+#include "vcn/vcn_5_0_0_offset.h"
+#include "vcn/vcn_5_0_0_sh_mask.h"
+#include "ivsrcid/vcn/irqsrcs_vcn_5_0.h"
+#include "jpeg_v5_0_0.h"
+
+static const struct amdgpu_hwip_reg_entry jpeg_reg_list_5_0[] = {
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JPEG_POWER_STATUS),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JPEG_INT_STAT),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC_RB_RPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC_RB_WPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC_RB_CNTL),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC_RB_SIZE),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC_STATUS),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regJPEG_DEC_ADDR_MODE),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regJPEG_DEC_GFX10_ADDR_CONFIG),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regJPEG_DEC_Y_GFX10_TILING_SURFACE),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regJPEG_DEC_UV_GFX10_TILING_SURFACE),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JPEG_PITCH),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JPEG_UV_PITCH),
+};
+
+static void jpeg_v5_0_0_set_dec_ring_funcs(struct amdgpu_device *adev);
+static void jpeg_v5_0_0_set_irq_funcs(struct amdgpu_device *adev);
+static int jpeg_v5_0_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
+				enum amd_powergating_state state);
+
+/**
+ * jpeg_v5_0_0_early_init - set function pointers
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Set ring and irq function pointers
+ */
+static int jpeg_v5_0_0_early_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	adev->jpeg.num_jpeg_inst = 1;
+	adev->jpeg.num_jpeg_rings = 1;
+
+	jpeg_v5_0_0_set_dec_ring_funcs(adev);
+	jpeg_v5_0_0_set_irq_funcs(adev);
+
+	return 0;
+}
+
+/**
+ * jpeg_v5_0_0_sw_init - sw init for JPEG block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Load firmware and sw initialization
+ */
+static int jpeg_v5_0_0_sw_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	struct amdgpu_ring *ring;
+	int r;
+
+	/* JPEG TRAP */
+	r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_VCN,
+		VCN_5_0__SRCID__JPEG_DECODE, &adev->jpeg.inst->irq);
+	if (r)
+		return r;
+
+	r = amdgpu_jpeg_sw_init(adev);
+	if (r)
+		return r;
+
+	r = amdgpu_jpeg_resume(adev);
+	if (r)
+		return r;
+
+	ring = adev->jpeg.inst->ring_dec;
+	ring->use_doorbell = true;
+	ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 1;
+	ring->vm_hub = AMDGPU_MMHUB0(0);
+
+	sprintf(ring->name, "jpeg_dec");
+	r = amdgpu_ring_init(adev, ring, 512, &adev->jpeg.inst->irq, 0,
+			     AMDGPU_RING_PRIO_DEFAULT, NULL);
+	if (r)
+		return r;
+
+	adev->jpeg.internal.jpeg_pitch[0] = regUVD_JPEG_PITCH_INTERNAL_OFFSET;
+	adev->jpeg.inst->external.jpeg_pitch[0] = SOC15_REG_OFFSET(JPEG, 0, regUVD_JPEG_PITCH);
+
+	r = amdgpu_jpeg_reg_dump_init(adev, jpeg_reg_list_5_0, ARRAY_SIZE(jpeg_reg_list_5_0));
+	if (r)
+		return r;
+
+	adev->jpeg.supported_reset =
+		amdgpu_get_soft_full_reset_mask(&adev->jpeg.inst[0].ring_dec[0]);
+	if (!amdgpu_sriov_vf(adev))
+		adev->jpeg.supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE;
+	r = amdgpu_jpeg_sysfs_reset_mask_init(adev);
+
+	return r;
+}
+
+/**
+ * jpeg_v5_0_0_sw_fini - sw fini for JPEG block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * JPEG suspend and free up sw allocation
+ */
+static int jpeg_v5_0_0_sw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int r;
+
+	r = amdgpu_jpeg_suspend(adev);
+	if (r)
+		return r;
+
+	amdgpu_jpeg_sysfs_reset_mask_fini(adev);
+	r = amdgpu_jpeg_sw_fini(adev);
+
+	return r;
+}
+
+/**
+ * jpeg_v5_0_0_hw_init - start and test JPEG block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ */
+static int jpeg_v5_0_0_hw_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	struct amdgpu_ring *ring = adev->jpeg.inst->ring_dec;
+	int r;
+
+	adev->nbio.funcs->vcn_doorbell_range(adev, ring->use_doorbell,
+			(adev->doorbell_index.vcn.vcn_ring0_1 << 1), 0);
+
+	/* Skip ring test because pause DPG is not implemented. */
+	if (adev->pg_flags & AMD_PG_SUPPORT_JPEG_DPG)
+		return 0;
+
+	r = amdgpu_ring_test_helper(ring);
+	if (r)
+		return r;
+
+	return 0;
+}
+
+/**
+ * jpeg_v5_0_0_hw_fini - stop the hardware block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Stop the JPEG block, mark ring as not ready any more
+ */
+static int jpeg_v5_0_0_hw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	cancel_delayed_work_sync(&adev->jpeg.idle_work);
+
+	if (adev->jpeg.cur_state != AMD_PG_STATE_GATE &&
+	      RREG32_SOC15(JPEG, 0, regUVD_JRBC_STATUS))
+		jpeg_v5_0_0_set_powergating_state(ip_block, AMD_PG_STATE_GATE);
+
+	return 0;
+}
+
+/**
+ * jpeg_v5_0_0_suspend - suspend JPEG block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * HW fini and suspend JPEG block
+ */
+static int jpeg_v5_0_0_suspend(struct amdgpu_ip_block *ip_block)
+{
+	int r;
+
+	r = jpeg_v5_0_0_hw_fini(ip_block);
+	if (r)
+		return r;
+
+	r = amdgpu_jpeg_suspend(ip_block->adev);
+
+	return r;
+}
+
+/**
+ * jpeg_v5_0_0_resume - resume JPEG block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Resume firmware and hw init JPEG block
+ */
+static int jpeg_v5_0_0_resume(struct amdgpu_ip_block *ip_block)
+{
+	int r;
+
+	r = amdgpu_jpeg_resume(ip_block->adev);
+	if (r)
+		return r;
+
+	r = jpeg_v5_0_0_hw_init(ip_block);
+
+	return r;
+}
+
+static void jpeg_v5_0_0_disable_clock_gating(struct amdgpu_device *adev)
+{
+	uint32_t data = 0;
+
+	WREG32_SOC15(JPEG, 0, regJPEG_CGC_GATE, data);
+
+	data = RREG32_SOC15(JPEG, 0, regJPEG_CGC_CTRL);
+	data &= ~(JPEG_CGC_CTRL__JPEG0_DEC_MODE_MASK
+		| JPEG_CGC_CTRL__JPEG_ENC_MODE_MASK);
+	WREG32_SOC15(JPEG, 0, regJPEG_CGC_CTRL, data);
+}
+
+static void jpeg_v5_0_0_enable_clock_gating(struct amdgpu_device *adev)
+{
+	uint32_t data = 0;
+
+	data = RREG32_SOC15(JPEG, 0, regJPEG_CGC_CTRL);
+
+	data |= 1 << JPEG_CGC_CTRL__JPEG0_DEC_MODE__SHIFT;
+	WREG32_SOC15(JPEG, 0, regJPEG_CGC_CTRL, data);
+
+	data = RREG32_SOC15(JPEG, 0, regJPEG_CGC_GATE);
+	data |= (JPEG_CGC_GATE__JPEG0_DEC_MASK
+		|JPEG_CGC_GATE__JPEG_ENC_MASK
+		|JPEG_CGC_GATE__JMCIF_MASK
+		|JPEG_CGC_GATE__JRBBM_MASK);
+	WREG32_SOC15(JPEG, 0, regJPEG_CGC_GATE, data);
+}
+
+static int jpeg_v5_0_0_disable_power_gating(struct amdgpu_device *adev)
+{
+	uint32_t data = 0;
+
+	data = 1 << UVD_IPX_DLDO_CONFIG__ONO1_PWR_CONFIG__SHIFT;
+	WREG32_SOC15(JPEG, 0, regUVD_IPX_DLDO_CONFIG, data);
+	SOC15_WAIT_ON_RREG(JPEG, 0, regUVD_IPX_DLDO_STATUS, 0,
+			UVD_IPX_DLDO_STATUS__ONO1_PWR_STATUS_MASK);
+
+	/* disable anti hang mechanism */
+	WREG32_P(SOC15_REG_OFFSET(JPEG, 0, regUVD_JPEG_POWER_STATUS), 0,
+		~UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK);
+
+	return 0;
+}
+
+static int jpeg_v5_0_0_enable_power_gating(struct amdgpu_device *adev)
+{
+	/* enable anti hang mechanism */
+	WREG32_P(SOC15_REG_OFFSET(JPEG, 0, regUVD_JPEG_POWER_STATUS),
+		UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK,
+		~UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK);
+
+	if (adev->pg_flags & AMD_PG_SUPPORT_JPEG) {
+		WREG32(SOC15_REG_OFFSET(JPEG, 0, regUVD_IPX_DLDO_CONFIG),
+			2 << UVD_IPX_DLDO_CONFIG__ONO1_PWR_CONFIG__SHIFT);
+		SOC15_WAIT_ON_RREG(JPEG, 0, regUVD_IPX_DLDO_STATUS,
+			1 << UVD_IPX_DLDO_STATUS__ONO1_PWR_STATUS__SHIFT,
+			UVD_IPX_DLDO_STATUS__ONO1_PWR_STATUS_MASK);
+	}
+
+	return 0;
+}
+
+static void jpeg_engine_5_0_0_dpg_clock_gating_mode(struct amdgpu_device *adev,
+	       int inst_idx, uint8_t indirect)
+{
+	uint32_t data = 0;
+
+	// JPEG disable CGC
+	if (adev->cg_flags & AMD_CG_SUPPORT_JPEG_MGCG)
+		data = 1 << JPEG_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
+	else
+		data = 0 << JPEG_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
+
+	data |= 1 << JPEG_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT;
+	data |= 4 << JPEG_CGC_CTRL__CLK_OFF_DELAY__SHIFT;
+
+	if (indirect) {
+		ADD_SOC24_JPEG_TO_DPG_SRAM(inst_idx, vcnipJPEG_CGC_CTRL, data, indirect);
+
+		// Turn on All JPEG clocks
+		data = 0;
+		ADD_SOC24_JPEG_TO_DPG_SRAM(inst_idx, vcnipJPEG_CGC_GATE, data, indirect);
+	} else {
+		WREG32_SOC24_JPEG_DPG_MODE(inst_idx, vcnipJPEG_CGC_CTRL, data, indirect);
+
+		// Turn on All JPEG clocks
+		data = 0;
+		WREG32_SOC24_JPEG_DPG_MODE(inst_idx, vcnipJPEG_CGC_GATE, data, indirect);
+	}
+}
+
+/**
+ * jpeg_v5_0_0_start_dpg_mode - Jpeg start with dpg mode
+ *
+ * @adev: amdgpu_device pointer
+ * @inst_idx: instance number index
+ * @indirect: indirectly write sram
+ *
+ * Start JPEG block with dpg mode
+ */
+static int jpeg_v5_0_0_start_dpg_mode(struct amdgpu_device *adev, int inst_idx, bool indirect)
+{
+	struct amdgpu_ring *ring = adev->jpeg.inst[inst_idx].ring_dec;
+	uint32_t reg_data = 0;
+
+	jpeg_v5_0_0_enable_power_gating(adev);
+
+	// enable dynamic power gating mode
+	reg_data = RREG32_SOC15(JPEG, inst_idx, regUVD_JPEG_POWER_STATUS);
+	reg_data |= UVD_JPEG_POWER_STATUS__JPEG_PG_MODE_MASK;
+	WREG32_SOC15(JPEG, inst_idx, regUVD_JPEG_POWER_STATUS, reg_data);
+
+	if (indirect)
+		adev->jpeg.inst[inst_idx].dpg_sram_curr_addr =
+			(uint32_t *)adev->jpeg.inst[inst_idx].dpg_sram_cpu_addr;
+
+	jpeg_engine_5_0_0_dpg_clock_gating_mode(adev, inst_idx, indirect);
+
+	/* MJPEG global tiling registers */
+	if (indirect)
+		ADD_SOC24_JPEG_TO_DPG_SRAM(inst_idx, vcnipJPEG_DEC_GFX10_ADDR_CONFIG,
+			adev->gfx.config.gb_addr_config, indirect);
+	else
+		WREG32_SOC24_JPEG_DPG_MODE(inst_idx, vcnipJPEG_DEC_GFX10_ADDR_CONFIG,
+			adev->gfx.config.gb_addr_config, 1);
+
+	/* enable System Interrupt for JRBC */
+	if (indirect)
+		ADD_SOC24_JPEG_TO_DPG_SRAM(inst_idx, vcnipJPEG_SYS_INT_EN,
+			JPEG_SYS_INT_EN__DJRBC0_MASK, indirect);
+	else
+		WREG32_SOC24_JPEG_DPG_MODE(inst_idx, vcnipJPEG_SYS_INT_EN,
+			JPEG_SYS_INT_EN__DJRBC0_MASK, 1);
+
+	if (indirect) {
+		/* add nop to workaround PSP size check */
+		ADD_SOC24_JPEG_TO_DPG_SRAM(inst_idx, vcnipUVD_NO_OP, 0, indirect);
+
+		amdgpu_jpeg_psp_update_sram(adev, inst_idx, 0);
+	}
+
+	WREG32_SOC15(VCN, 0, regVCN_JPEG_DB_CTRL,
+		ring->doorbell_index << VCN_JPEG_DB_CTRL__OFFSET__SHIFT |
+		VCN_JPEG_DB_CTRL__EN_MASK);
+
+	WREG32_SOC15(JPEG, inst_idx, regUVD_LMI_JRBC_RB_VMID, 0);
+	WREG32_SOC15(JPEG, inst_idx, regUVD_JRBC_RB_CNTL, (0x00000001L | 0x00000002L));
+	WREG32_SOC15(JPEG, inst_idx, regUVD_LMI_JRBC_RB_64BIT_BAR_LOW,
+		lower_32_bits(ring->gpu_addr));
+	WREG32_SOC15(JPEG, inst_idx, regUVD_LMI_JRBC_RB_64BIT_BAR_HIGH,
+		upper_32_bits(ring->gpu_addr));
+	WREG32_SOC15(JPEG, inst_idx, regUVD_JRBC_RB_RPTR, 0);
+	WREG32_SOC15(JPEG, inst_idx, regUVD_JRBC_RB_WPTR, 0);
+	WREG32_SOC15(JPEG, inst_idx, regUVD_JRBC_RB_CNTL, 0x00000002L);
+	WREG32_SOC15(JPEG, inst_idx, regUVD_JRBC_RB_SIZE, ring->ring_size / 4);
+	ring->wptr = RREG32_SOC15(JPEG, inst_idx, regUVD_JRBC_RB_WPTR);
+
+	return 0;
+}
+
+/**
+ * jpeg_v5_0_0_stop_dpg_mode - Jpeg stop with dpg mode
+ *
+ * @adev: amdgpu_device pointer
+ * @inst_idx: instance number index
+ *
+ * Stop JPEG block with dpg mode
+ */
+static void jpeg_v5_0_0_stop_dpg_mode(struct amdgpu_device *adev, int inst_idx)
+{
+	uint32_t reg_data = 0;
+
+	reg_data = RREG32_SOC15(JPEG, inst_idx, regUVD_JPEG_POWER_STATUS);
+	reg_data &= ~UVD_JPEG_POWER_STATUS__JPEG_PG_MODE_MASK;
+	WREG32_SOC15(JPEG, inst_idx, regUVD_JPEG_POWER_STATUS, reg_data);
+}
+
+/**
+ * jpeg_v5_0_0_start - start JPEG block
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Setup and start the JPEG block
+ */
+static int jpeg_v5_0_0_start(struct amdgpu_device *adev)
+{
+	struct amdgpu_ring *ring = adev->jpeg.inst->ring_dec;
+	int r;
+
+	if (adev->pm.dpm_enabled)
+		amdgpu_dpm_enable_jpeg(adev, true);
+
+	if (adev->pg_flags & AMD_PG_SUPPORT_JPEG_DPG) {
+		r = jpeg_v5_0_0_start_dpg_mode(adev, 0, adev->jpeg.indirect_sram);
+		return r;
+	}
+
+	/* disable power gating */
+	r = jpeg_v5_0_0_disable_power_gating(adev);
+	if (r)
+		return r;
+
+	/* JPEG disable CGC */
+	jpeg_v5_0_0_disable_clock_gating(adev);
+
+	/* MJPEG global tiling registers */
+	WREG32_SOC15(JPEG, 0, regJPEG_DEC_GFX10_ADDR_CONFIG,
+		adev->gfx.config.gb_addr_config);
+
+	/* enable JMI channel */
+	WREG32_P(SOC15_REG_OFFSET(JPEG, 0, regUVD_JMI_CNTL), 0,
+		~UVD_JMI_CNTL__SOFT_RESET_MASK);
+
+	/* enable System Interrupt for JRBC */
+	WREG32_P(SOC15_REG_OFFSET(JPEG, 0, regJPEG_SYS_INT_EN),
+		JPEG_SYS_INT_EN__DJRBC0_MASK,
+		~JPEG_SYS_INT_EN__DJRBC0_MASK);
+
+	WREG32_SOC15(VCN, 0, regVCN_JPEG_DB_CTRL,
+		ring->doorbell_index << VCN_JPEG_DB_CTRL__OFFSET__SHIFT |
+		VCN_JPEG_DB_CTRL__EN_MASK);
+
+	WREG32_SOC15(JPEG, 0, regUVD_LMI_JRBC_RB_VMID, 0);
+	WREG32_SOC15(JPEG, 0, regUVD_JRBC_RB_CNTL, (0x00000001L | 0x00000002L));
+	WREG32_SOC15(JPEG, 0, regUVD_LMI_JRBC_RB_64BIT_BAR_LOW,
+		lower_32_bits(ring->gpu_addr));
+	WREG32_SOC15(JPEG, 0, regUVD_LMI_JRBC_RB_64BIT_BAR_HIGH,
+		upper_32_bits(ring->gpu_addr));
+	WREG32_SOC15(JPEG, 0, regUVD_JRBC_RB_RPTR, 0);
+	WREG32_SOC15(JPEG, 0, regUVD_JRBC_RB_WPTR, 0);
+	WREG32_SOC15(JPEG, 0, regUVD_JRBC_RB_CNTL, 0x00000002L);
+	WREG32_SOC15(JPEG, 0, regUVD_JRBC_RB_SIZE, ring->ring_size / 4);
+	ring->wptr = RREG32_SOC15(JPEG, 0, regUVD_JRBC_RB_WPTR);
+
+	return 0;
+}
+
+/**
+ * jpeg_v5_0_0_stop - stop JPEG block
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * stop the JPEG block
+ */
+static int jpeg_v5_0_0_stop(struct amdgpu_device *adev)
+{
+	int r;
+
+	if (adev->pg_flags & AMD_PG_SUPPORT_JPEG_DPG) {
+		jpeg_v5_0_0_stop_dpg_mode(adev, 0);
+	} else {
+
+		/* reset JMI */
+		WREG32_P(SOC15_REG_OFFSET(JPEG, 0, regUVD_JMI_CNTL),
+			UVD_JMI_CNTL__SOFT_RESET_MASK,
+			~UVD_JMI_CNTL__SOFT_RESET_MASK);
+
+		jpeg_v5_0_0_enable_clock_gating(adev);
+
+		/* enable power gating */
+		r = jpeg_v5_0_0_enable_power_gating(adev);
+		if (r)
+			return r;
+	}
+
+	if (adev->pm.dpm_enabled)
+		amdgpu_dpm_enable_jpeg(adev, false);
+
+	return 0;
+}
+
+/**
+ * jpeg_v5_0_0_dec_ring_get_rptr - get read pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Returns the current hardware read pointer
+ */
+static uint64_t jpeg_v5_0_0_dec_ring_get_rptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	return RREG32_SOC15(JPEG, 0, regUVD_JRBC_RB_RPTR);
+}
+
+/**
+ * jpeg_v5_0_0_dec_ring_get_wptr - get write pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Returns the current hardware write pointer
+ */
+static uint64_t jpeg_v5_0_0_dec_ring_get_wptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	if (ring->use_doorbell)
+		return *ring->wptr_cpu_addr;
+	else
+		return RREG32_SOC15(JPEG, 0, regUVD_JRBC_RB_WPTR);
+}
+
+/**
+ * jpeg_v5_0_0_dec_ring_set_wptr - set write pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Commits the write pointer to the hardware
+ */
+static void jpeg_v5_0_0_dec_ring_set_wptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	if (ring->use_doorbell) {
+		*ring->wptr_cpu_addr = lower_32_bits(ring->wptr);
+		WDOORBELL32(ring->doorbell_index, lower_32_bits(ring->wptr));
+	} else {
+		WREG32_SOC15(JPEG, 0, regUVD_JRBC_RB_WPTR, lower_32_bits(ring->wptr));
+	}
+}
+
+static bool jpeg_v5_0_0_is_idle(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int ret = 1;
+
+	ret &= (((RREG32_SOC15(JPEG, 0, regUVD_JRBC_STATUS) &
+		UVD_JRBC_STATUS__RB_JOB_DONE_MASK) ==
+		UVD_JRBC_STATUS__RB_JOB_DONE_MASK));
+
+	return ret;
+}
+
+static int jpeg_v5_0_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	return SOC15_WAIT_ON_RREG(JPEG, 0, regUVD_JRBC_STATUS,
+		UVD_JRBC_STATUS__RB_JOB_DONE_MASK,
+		UVD_JRBC_STATUS__RB_JOB_DONE_MASK);
+}
+
+static int jpeg_v5_0_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
+					  enum amd_clockgating_state state)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	bool enable = state == AMD_CG_STATE_GATE;
+
+	if (enable) {
+		if (!jpeg_v5_0_0_is_idle(ip_block))
+			return -EBUSY;
+		jpeg_v5_0_0_enable_clock_gating(adev);
+	} else {
+		jpeg_v5_0_0_disable_clock_gating(adev);
+	}
+
+	return 0;
+}
+
+static int jpeg_v5_0_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
+					  enum amd_powergating_state state)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int ret;
+
+	if (state == adev->jpeg.cur_state)
+		return 0;
+
+	if (state == AMD_PG_STATE_GATE)
+		ret = jpeg_v5_0_0_stop(adev);
+	else
+		ret = jpeg_v5_0_0_start(adev);
+
+	if (!ret)
+		adev->jpeg.cur_state = state;
+
+	return ret;
+}
+
+static int jpeg_v5_0_0_set_interrupt_state(struct amdgpu_device *adev,
+					struct amdgpu_irq_src *source,
+					unsigned int type,
+					enum amdgpu_interrupt_state state)
+{
+	return 0;
+}
+
+static int jpeg_v5_0_0_process_interrupt(struct amdgpu_device *adev,
+				      struct amdgpu_irq_src *source,
+				      struct amdgpu_iv_entry *entry)
+{
+	DRM_DEBUG("IH: JPEG TRAP\n");
+
+	switch (entry->src_id) {
+	case VCN_5_0__SRCID__JPEG_DECODE:
+		amdgpu_fence_process(adev->jpeg.inst->ring_dec);
+		break;
+	default:
+		DRM_DEV_ERROR(adev->dev, "Unhandled interrupt: %d %d\n",
+			  entry->src_id, entry->src_data[0]);
+		break;
+	}
+
+	return 0;
+}
+
+static int jpeg_v5_0_0_ring_reset(struct amdgpu_ring *ring,
+				  unsigned int vmid,
+				  struct amdgpu_fence *timedout_fence)
+{
+	int r;
+
+	amdgpu_ring_reset_helper_begin(ring, timedout_fence);
+	r = jpeg_v5_0_0_stop(ring->adev);
+	if (r)
+		return r;
+	r = jpeg_v5_0_0_start(ring->adev);
+	if (r)
+		return r;
+	return amdgpu_ring_reset_helper_end(ring, timedout_fence);
+}
+
+static const struct amd_ip_funcs jpeg_v5_0_0_ip_funcs = {
+	.name = "jpeg_v5_0_0",
+	.early_init = jpeg_v5_0_0_early_init,
+	.sw_init = jpeg_v5_0_0_sw_init,
+	.sw_fini = jpeg_v5_0_0_sw_fini,
+	.hw_init = jpeg_v5_0_0_hw_init,
+	.hw_fini = jpeg_v5_0_0_hw_fini,
+	.suspend = jpeg_v5_0_0_suspend,
+	.resume = jpeg_v5_0_0_resume,
+	.is_idle = jpeg_v5_0_0_is_idle,
+	.wait_for_idle = jpeg_v5_0_0_wait_for_idle,
+	.set_clockgating_state = jpeg_v5_0_0_set_clockgating_state,
+	.set_powergating_state = jpeg_v5_0_0_set_powergating_state,
+	.dump_ip_state = amdgpu_jpeg_dump_ip_state,
+	.print_ip_state = amdgpu_jpeg_print_ip_state,
+};
+
+static const struct amdgpu_ring_funcs jpeg_v5_0_0_dec_ring_vm_funcs = {
+	.type = AMDGPU_RING_TYPE_VCN_JPEG,
+	.align_mask = 0xf,
+	.get_rptr = jpeg_v5_0_0_dec_ring_get_rptr,
+	.get_wptr = jpeg_v5_0_0_dec_ring_get_wptr,
+	.set_wptr = jpeg_v5_0_0_dec_ring_set_wptr,
+	.parse_cs = amdgpu_jpeg_dec_parse_cs,
+	.emit_frame_size =
+		SOC15_FLUSH_GPU_TLB_NUM_WREG * 6 +
+		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 8 +
+		8 + /* jpeg_v5_0_0_dec_ring_emit_vm_flush */
+		22 + 22 + /* jpeg_v5_0_0_dec_ring_emit_fence x2 vm fence */
+		8 + 16,
+	.emit_ib_size = 22, /* jpeg_v5_0_0_dec_ring_emit_ib */
+	.emit_ib = jpeg_v4_0_3_dec_ring_emit_ib,
+	.emit_fence = jpeg_v4_0_3_dec_ring_emit_fence,
+	.emit_vm_flush = jpeg_v4_0_3_dec_ring_emit_vm_flush,
+	.test_ring = amdgpu_jpeg_dec_ring_test_ring,
+	.test_ib = amdgpu_jpeg_dec_ring_test_ib,
+	.insert_nop = jpeg_v4_0_3_dec_ring_nop,
+	.insert_start = jpeg_v4_0_3_dec_ring_insert_start,
+	.insert_end = jpeg_v4_0_3_dec_ring_insert_end,
+	.pad_ib = amdgpu_ring_generic_pad_ib,
+	.begin_use = amdgpu_jpeg_ring_begin_use,
+	.end_use = amdgpu_jpeg_ring_end_use,
+	.emit_wreg = jpeg_v4_0_3_dec_ring_emit_wreg,
+	.emit_reg_wait = jpeg_v4_0_3_dec_ring_emit_reg_wait,
+	.emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper,
+	.reset = jpeg_v5_0_0_ring_reset,
+};
+
+static void jpeg_v5_0_0_set_dec_ring_funcs(struct amdgpu_device *adev)
+{
+	adev->jpeg.inst->ring_dec->funcs = &jpeg_v5_0_0_dec_ring_vm_funcs;
+}
+
+static const struct amdgpu_irq_src_funcs jpeg_v5_0_0_irq_funcs = {
+	.set = jpeg_v5_0_0_set_interrupt_state,
+	.process = jpeg_v5_0_0_process_interrupt,
+};
+
+static void jpeg_v5_0_0_set_irq_funcs(struct amdgpu_device *adev)
+{
+	adev->jpeg.inst->irq.num_types = 1;
+	adev->jpeg.inst->irq.funcs = &jpeg_v5_0_0_irq_funcs;
+}
+
+const struct amdgpu_ip_block_version jpeg_v5_0_0_ip_block = {
+	.type = AMD_IP_BLOCK_TYPE_JPEG,
+	.major = 5,
+	.minor = 0,
+	.rev = 0,
+	.funcs = &jpeg_v5_0_0_ip_funcs,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_0.h b/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_0.h
new file mode 100644
index 000000000000..5abb96159814
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_0.h
@@ -0,0 +1,35 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __JPEG_V5_0_0_H__
+#define __JPEG_V5_0_0_H__
+
+#define vcnipJPEG_CGC_GATE                                 0x4160
+#define vcnipJPEG_CGC_CTRL                                 0x4161
+#define vcnipJPEG_SYS_INT_EN                               0x4141
+#define vcnipUVD_NO_OP                                     0x0029
+#define vcnipJPEG_DEC_GFX10_ADDR_CONFIG                    0x404A
+
+extern const struct amdgpu_ip_block_version jpeg_v5_0_0_ip_block;
+
+#endif /* __JPEG_V5_0_0_H__ */
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_1.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_1.c
new file mode 100644
index 000000000000..ab0bf880d3d8
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_1.c
@@ -0,0 +1,1101 @@
+// SPDX-License-Identifier: GPL-2.0 OR MIT
+/*
+ * Copyright 2014-2024 Advanced Micro Devices, Inc. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#include "amdgpu.h"
+#include "amdgpu_jpeg.h"
+#include "amdgpu_pm.h"
+#include "soc15.h"
+#include "soc15d.h"
+#include "jpeg_v4_0_3.h"
+#include "jpeg_v5_0_1.h"
+#include "mmsch_v5_0.h"
+
+#include "vcn/vcn_5_0_0_offset.h"
+#include "vcn/vcn_5_0_0_sh_mask.h"
+#include "ivsrcid/vcn/irqsrcs_vcn_5_0.h"
+
+static int jpeg_v5_0_1_start_sriov(struct amdgpu_device *adev);
+static void jpeg_v5_0_1_set_dec_ring_funcs(struct amdgpu_device *adev);
+static void jpeg_v5_0_1_set_irq_funcs(struct amdgpu_device *adev);
+static int jpeg_v5_0_1_set_powergating_state(struct amdgpu_ip_block *ip_block,
+					     enum amd_powergating_state state);
+static void jpeg_v5_0_1_set_ras_funcs(struct amdgpu_device *adev);
+static void jpeg_v5_0_1_dec_ring_set_wptr(struct amdgpu_ring *ring);
+
+static int amdgpu_ih_srcid_jpeg[] = {
+	VCN_5_0__SRCID__JPEG_DECODE,
+	VCN_5_0__SRCID__JPEG1_DECODE,
+	VCN_5_0__SRCID__JPEG2_DECODE,
+	VCN_5_0__SRCID__JPEG3_DECODE,
+	VCN_5_0__SRCID__JPEG4_DECODE,
+	VCN_5_0__SRCID__JPEG5_DECODE,
+	VCN_5_0__SRCID__JPEG6_DECODE,
+	VCN_5_0__SRCID__JPEG7_DECODE,
+	VCN_5_0__SRCID__JPEG8_DECODE,
+	VCN_5_0__SRCID__JPEG9_DECODE,
+};
+
+static const struct amdgpu_hwip_reg_entry jpeg_reg_list_5_0_1[] = {
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JPEG_POWER_STATUS),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JPEG_INT_STAT),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC0_UVD_JRBC_RB_RPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC0_UVD_JRBC_RB_WPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC0_UVD_JRBC_STATUS),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regJPEG_DEC_ADDR_MODE),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regJPEG_DEC_GFX10_ADDR_CONFIG),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regJPEG_DEC_Y_GFX10_TILING_SURFACE),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regJPEG_DEC_UV_GFX10_TILING_SURFACE),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JPEG_PITCH),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JPEG_UV_PITCH),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC1_UVD_JRBC_RB_RPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC1_UVD_JRBC_RB_WPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC1_UVD_JRBC_STATUS),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC2_UVD_JRBC_RB_RPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC2_UVD_JRBC_RB_WPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC2_UVD_JRBC_STATUS),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC3_UVD_JRBC_RB_RPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC3_UVD_JRBC_RB_WPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC3_UVD_JRBC_STATUS),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC4_UVD_JRBC_RB_RPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC4_UVD_JRBC_RB_WPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC4_UVD_JRBC_STATUS),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC5_UVD_JRBC_RB_RPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC5_UVD_JRBC_RB_WPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC5_UVD_JRBC_STATUS),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC6_UVD_JRBC_RB_RPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC6_UVD_JRBC_RB_WPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC6_UVD_JRBC_STATUS),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC7_UVD_JRBC_RB_RPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC7_UVD_JRBC_RB_WPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC7_UVD_JRBC_STATUS),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC8_UVD_JRBC_RB_RPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC8_UVD_JRBC_RB_WPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC8_UVD_JRBC_STATUS),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC9_UVD_JRBC_RB_RPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC9_UVD_JRBC_RB_WPTR),
+	SOC15_REG_ENTRY_STR(JPEG, 0, regUVD_JRBC9_UVD_JRBC_STATUS),
+};
+
+static int jpeg_v5_0_1_core_reg_offset(u32 pipe)
+{
+	if (pipe <= AMDGPU_MAX_JPEG_RINGS_4_0_3)
+		return ((0x40 * pipe) - 0xc80);
+	else
+		return ((0x40 * pipe) - 0x440);
+}
+
+/**
+ * jpeg_v5_0_1_early_init - set function pointers
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Set ring and irq function pointers
+ */
+static int jpeg_v5_0_1_early_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	if (!adev->jpeg.num_jpeg_inst || adev->jpeg.num_jpeg_inst > AMDGPU_MAX_JPEG_INSTANCES)
+		return -ENOENT;
+
+	adev->jpeg.num_jpeg_rings = AMDGPU_MAX_JPEG_RINGS;
+	jpeg_v5_0_1_set_dec_ring_funcs(adev);
+	jpeg_v5_0_1_set_irq_funcs(adev);
+	jpeg_v5_0_1_set_ras_funcs(adev);
+
+	return 0;
+}
+
+/**
+ * jpeg_v5_0_1_sw_init - sw init for JPEG block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Load firmware and sw initialization
+ */
+static int jpeg_v5_0_1_sw_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	struct amdgpu_ring *ring;
+	int i, j, r, jpeg_inst;
+
+	for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j) {
+		/* JPEG TRAP */
+		r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_VCN,
+				      amdgpu_ih_srcid_jpeg[j], &adev->jpeg.inst->irq);
+		if (r)
+			return r;
+	}
+	/* JPEG DJPEG POISON EVENT */
+	r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_VCN,
+			VCN_5_0__SRCID_DJPEG0_POISON, &adev->jpeg.inst->ras_poison_irq);
+	if (r)
+		return r;
+
+	/* JPEG EJPEG POISON EVENT */
+	r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_VCN,
+			VCN_5_0__SRCID_EJPEG0_POISON, &adev->jpeg.inst->ras_poison_irq);
+	if (r)
+		return r;
+
+	r = amdgpu_jpeg_sw_init(adev);
+	if (r)
+		return r;
+
+	r = amdgpu_jpeg_resume(adev);
+	if (r)
+		return r;
+
+	for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
+		jpeg_inst = GET_INST(JPEG, i);
+
+		for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j) {
+			ring = &adev->jpeg.inst[i].ring_dec[j];
+			ring->use_doorbell = true;
+			ring->vm_hub = AMDGPU_MMHUB0(adev->jpeg.inst[i].aid_id);
+			if (!amdgpu_sriov_vf(adev)) {
+				ring->doorbell_index =
+					(adev->doorbell_index.vcn.vcn_ring0_1 << 1) +
+					1 + j + 11 * jpeg_inst;
+			} else {
+				ring->doorbell_index =
+					(adev->doorbell_index.vcn.vcn_ring0_1 << 1) +
+					2 + j + 32 * jpeg_inst;
+			}
+			sprintf(ring->name, "jpeg_dec_%d.%d", adev->jpeg.inst[i].aid_id, j);
+			r = amdgpu_ring_init(adev, ring, 512, &adev->jpeg.inst->irq, 0,
+					     AMDGPU_RING_PRIO_DEFAULT, NULL);
+			if (r)
+				return r;
+
+			adev->jpeg.internal.jpeg_pitch[j] =
+				regUVD_JRBC0_UVD_JRBC_SCRATCH0_INTERNAL_OFFSET;
+			adev->jpeg.inst[i].external.jpeg_pitch[j] =
+				SOC15_REG_OFFSET1(JPEG, jpeg_inst, regUVD_JRBC_SCRATCH0,
+						  (j ? jpeg_v5_0_1_core_reg_offset(j) : 0));
+		}
+	}
+
+	if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__JPEG)) {
+		r = amdgpu_jpeg_ras_sw_init(adev);
+		if (r) {
+			dev_err(adev->dev, "Failed to initialize jpeg ras block!\n");
+			return r;
+		}
+	}
+
+	r = amdgpu_jpeg_reg_dump_init(adev, jpeg_reg_list_5_0_1, ARRAY_SIZE(jpeg_reg_list_5_0_1));
+	if (r)
+		return r;
+
+	adev->jpeg.supported_reset =
+		amdgpu_get_soft_full_reset_mask(&adev->jpeg.inst[0].ring_dec[0]);
+	if (!amdgpu_sriov_vf(adev))
+		adev->jpeg.supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE;
+	r = amdgpu_jpeg_sysfs_reset_mask_init(adev);
+
+	return r;
+}
+
+/**
+ * jpeg_v5_0_1_sw_fini - sw fini for JPEG block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * JPEG suspend and free up sw allocation
+ */
+static int jpeg_v5_0_1_sw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int r;
+
+	r = amdgpu_jpeg_suspend(adev);
+	if (r)
+		return r;
+
+	amdgpu_jpeg_sysfs_reset_mask_fini(adev);
+
+	r = amdgpu_jpeg_sw_fini(adev);
+
+	return r;
+}
+
+/**
+ * jpeg_v5_0_1_hw_init - start and test JPEG block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ */
+static int jpeg_v5_0_1_hw_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	struct amdgpu_ring *ring;
+	int i, j, r, jpeg_inst;
+
+	if (amdgpu_sriov_vf(adev)) {
+		r = jpeg_v5_0_1_start_sriov(adev);
+		if (r)
+			return r;
+
+		for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
+			for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j) {
+				ring = &adev->jpeg.inst[i].ring_dec[j];
+				ring->wptr = 0;
+				ring->wptr_old = 0;
+				jpeg_v5_0_1_dec_ring_set_wptr(ring);
+				ring->sched.ready = true;
+			}
+		}
+		return 0;
+	}
+	if (RREG32_SOC15(VCN, GET_INST(VCN, 0), regVCN_RRMT_CNTL) & 0x100)
+		adev->jpeg.caps |= AMDGPU_JPEG_CAPS(RRMT_ENABLED);
+
+	for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
+		jpeg_inst = GET_INST(JPEG, i);
+		ring = adev->jpeg.inst[i].ring_dec;
+		if (ring->use_doorbell)
+			adev->nbio.funcs->vcn_doorbell_range(adev, ring->use_doorbell,
+				 (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 11 * jpeg_inst,
+				 adev->jpeg.inst[i].aid_id);
+
+		for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j) {
+			ring = &adev->jpeg.inst[i].ring_dec[j];
+			if (ring->use_doorbell)
+				WREG32_SOC15_OFFSET(VCN, GET_INST(VCN, i), regVCN_JPEG_DB_CTRL,
+						    ring->pipe,
+						    ring->doorbell_index <<
+						    VCN_JPEG_DB_CTRL__OFFSET__SHIFT |
+						    VCN_JPEG_DB_CTRL__EN_MASK);
+			r = amdgpu_ring_test_helper(ring);
+			if (r)
+				return r;
+		}
+	}
+
+	return 0;
+}
+
+/**
+ * jpeg_v5_0_1_hw_fini - stop the hardware block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Stop the JPEG block, mark ring as not ready any more
+ */
+static int jpeg_v5_0_1_hw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int ret = 0;
+
+	cancel_delayed_work_sync(&adev->jpeg.idle_work);
+
+	if (!amdgpu_sriov_vf(adev)) {
+		if (adev->jpeg.cur_state != AMD_PG_STATE_GATE)
+			ret = jpeg_v5_0_1_set_powergating_state(ip_block, AMD_PG_STATE_GATE);
+	}
+
+	if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__JPEG) && !amdgpu_sriov_vf(adev))
+		amdgpu_irq_put(adev, &adev->jpeg.inst->ras_poison_irq, 0);
+
+	return ret;
+}
+
+/**
+ * jpeg_v5_0_1_suspend - suspend JPEG block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * HW fini and suspend JPEG block
+ */
+static int jpeg_v5_0_1_suspend(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int r;
+
+	r = jpeg_v5_0_1_hw_fini(ip_block);
+	if (r)
+		return r;
+
+	r = amdgpu_jpeg_suspend(adev);
+
+	return r;
+}
+
+/**
+ * jpeg_v5_0_1_resume - resume JPEG block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Resume firmware and hw init JPEG block
+ */
+static int jpeg_v5_0_1_resume(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int r;
+
+	r = amdgpu_jpeg_resume(adev);
+	if (r)
+		return r;
+
+	r = jpeg_v5_0_1_hw_init(ip_block);
+
+	return r;
+}
+
+static void jpeg_v5_0_1_init_inst(struct amdgpu_device *adev, int i)
+{
+	int jpeg_inst = GET_INST(JPEG, i);
+
+	/* disable anti hang mechanism */
+	WREG32_P(SOC15_REG_OFFSET(JPEG, jpeg_inst, regUVD_JPEG_POWER_STATUS), 0,
+		 ~UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK);
+
+	/* keep the JPEG in static PG mode */
+	WREG32_P(SOC15_REG_OFFSET(JPEG, jpeg_inst, regUVD_JPEG_POWER_STATUS), 0,
+		 ~UVD_JPEG_POWER_STATUS__JPEG_PG_MODE_MASK);
+
+	/* MJPEG global tiling registers */
+	WREG32_SOC15(JPEG, 0, regJPEG_DEC_GFX10_ADDR_CONFIG,
+		     adev->gfx.config.gb_addr_config);
+
+	/* enable JMI channel */
+	WREG32_P(SOC15_REG_OFFSET(JPEG, jpeg_inst, regUVD_JMI_CNTL), 0,
+		 ~UVD_JMI_CNTL__SOFT_RESET_MASK);
+}
+
+static void jpeg_v5_0_1_deinit_inst(struct amdgpu_device *adev, int i)
+{
+	int jpeg_inst = GET_INST(JPEG, i);
+	/* reset JMI */
+	WREG32_P(SOC15_REG_OFFSET(JPEG, jpeg_inst, regUVD_JMI_CNTL),
+		 UVD_JMI_CNTL__SOFT_RESET_MASK,
+		 ~UVD_JMI_CNTL__SOFT_RESET_MASK);
+
+	/* enable anti hang mechanism */
+	WREG32_P(SOC15_REG_OFFSET(JPEG, jpeg_inst, regUVD_JPEG_POWER_STATUS),
+		 UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK,
+		 ~UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK);
+}
+
+static void jpeg_v5_0_1_init_jrbc(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	u32 reg, data, mask;
+	int jpeg_inst = GET_INST(JPEG, ring->me);
+	int reg_offset = ring->pipe ? jpeg_v5_0_1_core_reg_offset(ring->pipe) : 0;
+
+	/* enable System Interrupt for JRBC */
+	reg = SOC15_REG_OFFSET(JPEG, jpeg_inst, regJPEG_SYS_INT_EN);
+	if (ring->pipe < AMDGPU_MAX_JPEG_RINGS_4_0_3) {
+		data = JPEG_SYS_INT_EN__DJRBC0_MASK << ring->pipe;
+		mask = ~(JPEG_SYS_INT_EN__DJRBC0_MASK << ring->pipe);
+		WREG32_P(reg, data, mask);
+	} else {
+		data = JPEG_SYS_INT_EN__DJRBC0_MASK << (ring->pipe+12);
+		mask = ~(JPEG_SYS_INT_EN__DJRBC0_MASK << (ring->pipe+12));
+		WREG32_P(reg, data, mask);
+	}
+
+	WREG32_SOC15_OFFSET(JPEG, jpeg_inst,
+			    regUVD_LMI_JRBC_RB_VMID,
+			    reg_offset, 0);
+	WREG32_SOC15_OFFSET(JPEG, jpeg_inst,
+			    regUVD_JRBC_RB_CNTL,
+			    reg_offset,
+			    (0x00000001L | 0x00000002L));
+	WREG32_SOC15_OFFSET(JPEG, jpeg_inst,
+			    regUVD_LMI_JRBC_RB_64BIT_BAR_LOW,
+			    reg_offset, lower_32_bits(ring->gpu_addr));
+	WREG32_SOC15_OFFSET(JPEG, jpeg_inst,
+			    regUVD_LMI_JRBC_RB_64BIT_BAR_HIGH,
+			    reg_offset, upper_32_bits(ring->gpu_addr));
+	WREG32_SOC15_OFFSET(JPEG, jpeg_inst,
+			    regUVD_JRBC_RB_RPTR,
+			    reg_offset, 0);
+	WREG32_SOC15_OFFSET(JPEG, jpeg_inst,
+			    regUVD_JRBC_RB_WPTR,
+			    reg_offset, 0);
+	WREG32_SOC15_OFFSET(JPEG, jpeg_inst,
+			    regUVD_JRBC_RB_CNTL,
+			    reg_offset, 0x00000002L);
+	WREG32_SOC15_OFFSET(JPEG, jpeg_inst,
+			    regUVD_JRBC_RB_SIZE,
+			    reg_offset, ring->ring_size / 4);
+	ring->wptr = RREG32_SOC15_OFFSET(JPEG, jpeg_inst, regUVD_JRBC_RB_WPTR,
+					 reg_offset);
+}
+
+static int jpeg_v5_0_1_start_sriov(struct amdgpu_device *adev)
+{
+	struct amdgpu_ring *ring;
+	uint64_t ctx_addr;
+	uint32_t param, resp, expected;
+	uint32_t tmp, timeout;
+
+	struct amdgpu_mm_table *table = &adev->virt.mm_table;
+	uint32_t *table_loc;
+	uint32_t table_size;
+	uint32_t size, size_dw, item_offset;
+	uint32_t init_status;
+	int i, j, jpeg_inst;
+
+	struct mmsch_v5_0_cmd_direct_write
+		direct_wt = { {0} };
+	struct mmsch_v5_0_cmd_end end = { {0} };
+	struct mmsch_v5_0_init_header header;
+
+	direct_wt.cmd_header.command_type =
+		MMSCH_COMMAND__DIRECT_REG_WRITE;
+	end.cmd_header.command_type =
+		MMSCH_COMMAND__END;
+
+	for (i = 0; i < adev->jpeg.num_jpeg_inst; i++) {
+		jpeg_inst = GET_INST(JPEG, i);
+
+		memset(&header, 0, sizeof(struct mmsch_v5_0_init_header));
+		header.version = MMSCH_VERSION;
+		header.total_size = sizeof(struct mmsch_v5_0_init_header) >> 2;
+
+		table_loc = (uint32_t *)table->cpu_addr;
+		table_loc += header.total_size;
+
+		item_offset = header.total_size;
+
+		for (j = 0; j < adev->jpeg.num_jpeg_rings; j++) {
+			ring = &adev->jpeg.inst[i].ring_dec[j];
+			table_size = 0;
+
+			tmp = SOC15_REG_OFFSET(JPEG, 0, regUVD_LMI_JRBC_RB_64BIT_BAR_LOW);
+			MMSCH_V5_0_INSERT_DIRECT_WT(tmp, lower_32_bits(ring->gpu_addr));
+			tmp = SOC15_REG_OFFSET(JPEG, 0, regUVD_LMI_JRBC_RB_64BIT_BAR_HIGH);
+			MMSCH_V5_0_INSERT_DIRECT_WT(tmp, upper_32_bits(ring->gpu_addr));
+			tmp = SOC15_REG_OFFSET(JPEG, 0, regUVD_JRBC_RB_SIZE);
+			MMSCH_V5_0_INSERT_DIRECT_WT(tmp, ring->ring_size / 4);
+
+			if (j < 5) {
+				header.mjpegdec0[j].table_offset = item_offset;
+				header.mjpegdec0[j].init_status = 0;
+				header.mjpegdec0[j].table_size = table_size;
+			} else {
+				header.mjpegdec1[j - 5].table_offset = item_offset;
+				header.mjpegdec1[j - 5].init_status = 0;
+				header.mjpegdec1[j - 5].table_size = table_size;
+			}
+			header.total_size += table_size;
+			item_offset += table_size;
+		}
+
+		MMSCH_V5_0_INSERT_END();
+
+		/* send init table to MMSCH */
+		size = sizeof(struct mmsch_v5_0_init_header);
+		table_loc = (uint32_t *)table->cpu_addr;
+		memcpy((void *)table_loc, &header, size);
+
+		ctx_addr = table->gpu_addr;
+		WREG32_SOC15(VCN, jpeg_inst, regMMSCH_VF_CTX_ADDR_LO, lower_32_bits(ctx_addr));
+		WREG32_SOC15(VCN, jpeg_inst, regMMSCH_VF_CTX_ADDR_HI, upper_32_bits(ctx_addr));
+
+		tmp = RREG32_SOC15(VCN, jpeg_inst, regMMSCH_VF_VMID);
+		tmp &= ~MMSCH_VF_VMID__VF_CTX_VMID_MASK;
+		tmp |= (0 << MMSCH_VF_VMID__VF_CTX_VMID__SHIFT);
+		WREG32_SOC15(VCN, jpeg_inst, regMMSCH_VF_VMID, tmp);
+
+		size = header.total_size;
+		WREG32_SOC15(VCN, jpeg_inst, regMMSCH_VF_CTX_SIZE, size);
+
+		WREG32_SOC15(VCN, jpeg_inst, regMMSCH_VF_MAILBOX_RESP, 0);
+
+		param = 0x00000001;
+		WREG32_SOC15(VCN, jpeg_inst, regMMSCH_VF_MAILBOX_HOST, param);
+		tmp = 0;
+		timeout = 1000;
+		resp = 0;
+		expected = MMSCH_VF_MAILBOX_RESP__OK;
+		init_status =
+			((struct mmsch_v5_0_init_header *)(table_loc))->mjpegdec0[i].init_status;
+		while (resp != expected) {
+			resp = RREG32_SOC15(VCN, jpeg_inst, regMMSCH_VF_MAILBOX_RESP);
+
+			if (resp != 0)
+				break;
+			udelay(10);
+			tmp = tmp + 10;
+			if (tmp >= timeout) {
+				DRM_ERROR("failed to init MMSCH. TIME-OUT after %d usec"\
+					" waiting for regMMSCH_VF_MAILBOX_RESP "\
+					"(expected=0x%08x, readback=0x%08x)\n",
+					tmp, expected, resp);
+				return -EBUSY;
+			}
+		}
+		if (resp != expected && resp != MMSCH_VF_MAILBOX_RESP__INCOMPLETE &&
+				init_status != MMSCH_VF_ENGINE_STATUS__PASS)
+			DRM_ERROR("MMSCH init status is incorrect! readback=0x%08x, header init status for jpeg: %x\n",
+					resp, init_status);
+
+	}
+	return 0;
+}
+
+/**
+ * jpeg_v5_0_1_start - start JPEG block
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Setup and start the JPEG block
+ */
+static int jpeg_v5_0_1_start(struct amdgpu_device *adev)
+{
+	struct amdgpu_ring *ring;
+	int i, j;
+
+	for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
+		jpeg_v5_0_1_init_inst(adev, i);
+		for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j) {
+			ring = &adev->jpeg.inst[i].ring_dec[j];
+			jpeg_v5_0_1_init_jrbc(ring);
+		}
+	}
+
+	return 0;
+}
+
+/**
+ * jpeg_v5_0_1_stop - stop JPEG block
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * stop the JPEG block
+ */
+static int jpeg_v5_0_1_stop(struct amdgpu_device *adev)
+{
+	int i;
+
+	for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i)
+		jpeg_v5_0_1_deinit_inst(adev, i);
+
+	return 0;
+}
+
+/**
+ * jpeg_v5_0_1_dec_ring_get_rptr - get read pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Returns the current hardware read pointer
+ */
+static uint64_t jpeg_v5_0_1_dec_ring_get_rptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	return RREG32_SOC15_OFFSET(JPEG, GET_INST(JPEG, ring->me), regUVD_JRBC_RB_RPTR,
+				   ring->pipe ? jpeg_v5_0_1_core_reg_offset(ring->pipe) : 0);
+}
+
+/**
+ * jpeg_v5_0_1_dec_ring_get_wptr - get write pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Returns the current hardware write pointer
+ */
+static uint64_t jpeg_v5_0_1_dec_ring_get_wptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	if (ring->use_doorbell)
+		return adev->wb.wb[ring->wptr_offs];
+
+	return RREG32_SOC15_OFFSET(JPEG, GET_INST(JPEG, ring->me), regUVD_JRBC_RB_WPTR,
+				   ring->pipe ? jpeg_v5_0_1_core_reg_offset(ring->pipe) : 0);
+}
+
+/**
+ * jpeg_v5_0_1_dec_ring_set_wptr - set write pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Commits the write pointer to the hardware
+ */
+static void jpeg_v5_0_1_dec_ring_set_wptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	if (ring->use_doorbell) {
+		adev->wb.wb[ring->wptr_offs] = lower_32_bits(ring->wptr);
+		WDOORBELL32(ring->doorbell_index, lower_32_bits(ring->wptr));
+	} else {
+		WREG32_SOC15_OFFSET(JPEG, GET_INST(JPEG, ring->me),
+				    regUVD_JRBC_RB_WPTR,
+				    (ring->pipe ? jpeg_v5_0_1_core_reg_offset(ring->pipe) : 0),
+				    lower_32_bits(ring->wptr));
+	}
+}
+
+static bool jpeg_v5_0_1_is_idle(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	bool ret = false;
+	int i, j;
+
+	for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
+		for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j) {
+			int reg_offset = (j ? jpeg_v5_0_1_core_reg_offset(j) : 0);
+
+			ret &= ((RREG32_SOC15_OFFSET(JPEG, GET_INST(JPEG, i),
+				regUVD_JRBC_STATUS, reg_offset) &
+				UVD_JRBC_STATUS__RB_JOB_DONE_MASK) ==
+				UVD_JRBC_STATUS__RB_JOB_DONE_MASK);
+		}
+	}
+
+	return ret;
+}
+
+static int jpeg_v5_0_1_wait_for_idle(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int ret = 0;
+	int i, j;
+
+	for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
+		for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j) {
+			int reg_offset = (j ? jpeg_v5_0_1_core_reg_offset(j) : 0);
+
+			ret &= SOC15_WAIT_ON_RREG_OFFSET(JPEG, GET_INST(JPEG, i),
+							 regUVD_JRBC_STATUS, reg_offset,
+							 UVD_JRBC_STATUS__RB_JOB_DONE_MASK,
+							 UVD_JRBC_STATUS__RB_JOB_DONE_MASK);
+		}
+	}
+	return ret;
+}
+
+static int jpeg_v5_0_1_set_clockgating_state(struct amdgpu_ip_block *ip_block,
+					     enum amd_clockgating_state state)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	bool enable = state == AMD_CG_STATE_GATE;
+
+	int i;
+
+	if (!enable)
+		return 0;
+
+	for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
+		if (!jpeg_v5_0_1_is_idle(ip_block))
+			return -EBUSY;
+	}
+
+	return 0;
+}
+
+static int jpeg_v5_0_1_set_powergating_state(struct amdgpu_ip_block *ip_block,
+					     enum amd_powergating_state state)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int ret;
+
+	if (amdgpu_sriov_vf(adev)) {
+		adev->jpeg.cur_state = AMD_PG_STATE_UNGATE;
+		return 0;
+	}
+
+	if (state == adev->jpeg.cur_state)
+		return 0;
+
+	if (state == AMD_PG_STATE_GATE)
+		ret = jpeg_v5_0_1_stop(adev);
+	else
+		ret = jpeg_v5_0_1_start(adev);
+
+	if (!ret)
+		adev->jpeg.cur_state = state;
+
+	return ret;
+}
+
+static int jpeg_v5_0_1_set_interrupt_state(struct amdgpu_device *adev,
+					   struct amdgpu_irq_src *source,
+					   unsigned int type,
+					   enum amdgpu_interrupt_state state)
+{
+	return 0;
+}
+
+static int jpeg_v5_0_1_set_ras_interrupt_state(struct amdgpu_device *adev,
+					struct amdgpu_irq_src *source,
+					unsigned int type,
+					enum amdgpu_interrupt_state state)
+{
+	return 0;
+}
+
+
+
+static int jpeg_v5_0_1_process_interrupt(struct amdgpu_device *adev,
+					 struct amdgpu_irq_src *source,
+					 struct amdgpu_iv_entry *entry)
+{
+	u32 i, inst;
+
+	i = node_id_to_phys_map[entry->node_id];
+	DRM_DEV_DEBUG(adev->dev, "IH: JPEG TRAP\n");
+
+	for (inst = 0; inst < adev->jpeg.num_jpeg_inst; ++inst)
+		if (adev->jpeg.inst[inst].aid_id == i)
+			break;
+
+	if (inst >= adev->jpeg.num_jpeg_inst) {
+		dev_WARN_ONCE(adev->dev, 1,
+			      "Interrupt received for unknown JPEG instance %d",
+			      entry->node_id);
+		return 0;
+	}
+
+	switch (entry->src_id) {
+	case VCN_5_0__SRCID__JPEG_DECODE:
+		amdgpu_fence_process(&adev->jpeg.inst[inst].ring_dec[0]);
+		break;
+	case VCN_5_0__SRCID__JPEG1_DECODE:
+		amdgpu_fence_process(&adev->jpeg.inst[inst].ring_dec[1]);
+		break;
+	case VCN_5_0__SRCID__JPEG2_DECODE:
+		amdgpu_fence_process(&adev->jpeg.inst[inst].ring_dec[2]);
+		break;
+	case VCN_5_0__SRCID__JPEG3_DECODE:
+		amdgpu_fence_process(&adev->jpeg.inst[inst].ring_dec[3]);
+		break;
+	case VCN_5_0__SRCID__JPEG4_DECODE:
+		amdgpu_fence_process(&adev->jpeg.inst[inst].ring_dec[4]);
+		break;
+	case VCN_5_0__SRCID__JPEG5_DECODE:
+		amdgpu_fence_process(&adev->jpeg.inst[inst].ring_dec[5]);
+		break;
+	case VCN_5_0__SRCID__JPEG6_DECODE:
+		amdgpu_fence_process(&adev->jpeg.inst[inst].ring_dec[6]);
+		break;
+	case VCN_5_0__SRCID__JPEG7_DECODE:
+		amdgpu_fence_process(&adev->jpeg.inst[inst].ring_dec[7]);
+		break;
+	case VCN_5_0__SRCID__JPEG8_DECODE:
+		amdgpu_fence_process(&adev->jpeg.inst[inst].ring_dec[8]);
+		break;
+	case VCN_5_0__SRCID__JPEG9_DECODE:
+		amdgpu_fence_process(&adev->jpeg.inst[inst].ring_dec[9]);
+		break;
+	default:
+		DRM_DEV_ERROR(adev->dev, "Unhandled interrupt: %d %d\n",
+			      entry->src_id, entry->src_data[0]);
+		break;
+	}
+
+	return 0;
+}
+
+static void jpeg_v5_0_1_core_stall_reset(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	int jpeg_inst = GET_INST(JPEG, ring->me);
+	int reg_offset = ring->pipe ? jpeg_v5_0_1_core_reg_offset(ring->pipe) : 0;
+
+	WREG32_SOC15_OFFSET(JPEG, jpeg_inst,
+			    regUVD_JMI0_UVD_JMI_CLIENT_STALL,
+			    reg_offset, 0x1F);
+	SOC15_WAIT_ON_RREG_OFFSET(JPEG, jpeg_inst,
+				  regUVD_JMI0_UVD_JMI_CLIENT_CLEAN_STATUS,
+				  reg_offset, 0x1F, 0x1F);
+	WREG32_SOC15_OFFSET(JPEG, jpeg_inst,
+			    regUVD_JMI0_JPEG_LMI_DROP,
+			    reg_offset, 0x1F);
+	WREG32_SOC15(JPEG, jpeg_inst, regJPEG_CORE_RST_CTRL, 1 << ring->pipe);
+	WREG32_SOC15_OFFSET(JPEG, jpeg_inst,
+			    regUVD_JMI0_UVD_JMI_CLIENT_STALL,
+			    reg_offset, 0x00);
+	WREG32_SOC15_OFFSET(JPEG, jpeg_inst,
+			    regUVD_JMI0_JPEG_LMI_DROP,
+			    reg_offset, 0x00);
+	WREG32_SOC15(JPEG, jpeg_inst, regJPEG_CORE_RST_CTRL, 0x00);
+}
+
+static int jpeg_v5_0_1_ring_reset(struct amdgpu_ring *ring,
+				  unsigned int vmid,
+				  struct amdgpu_fence *timedout_fence)
+{
+	amdgpu_ring_reset_helper_begin(ring, timedout_fence);
+	jpeg_v5_0_1_core_stall_reset(ring);
+	jpeg_v5_0_1_init_jrbc(ring);
+	return amdgpu_ring_reset_helper_end(ring, timedout_fence);
+}
+
+static const struct amd_ip_funcs jpeg_v5_0_1_ip_funcs = {
+	.name = "jpeg_v5_0_1",
+	.early_init = jpeg_v5_0_1_early_init,
+	.late_init = NULL,
+	.sw_init = jpeg_v5_0_1_sw_init,
+	.sw_fini = jpeg_v5_0_1_sw_fini,
+	.hw_init = jpeg_v5_0_1_hw_init,
+	.hw_fini = jpeg_v5_0_1_hw_fini,
+	.suspend = jpeg_v5_0_1_suspend,
+	.resume = jpeg_v5_0_1_resume,
+	.is_idle = jpeg_v5_0_1_is_idle,
+	.wait_for_idle = jpeg_v5_0_1_wait_for_idle,
+	.check_soft_reset = NULL,
+	.pre_soft_reset = NULL,
+	.soft_reset = NULL,
+	.post_soft_reset = NULL,
+	.set_clockgating_state = jpeg_v5_0_1_set_clockgating_state,
+	.set_powergating_state = jpeg_v5_0_1_set_powergating_state,
+	.dump_ip_state = amdgpu_jpeg_dump_ip_state,
+	.print_ip_state = amdgpu_jpeg_print_ip_state,
+};
+
+static const struct amdgpu_ring_funcs jpeg_v5_0_1_dec_ring_vm_funcs = {
+	.type = AMDGPU_RING_TYPE_VCN_JPEG,
+	.align_mask = 0xf,
+	.get_rptr = jpeg_v5_0_1_dec_ring_get_rptr,
+	.get_wptr = jpeg_v5_0_1_dec_ring_get_wptr,
+	.set_wptr = jpeg_v5_0_1_dec_ring_set_wptr,
+	.parse_cs = amdgpu_jpeg_dec_parse_cs,
+	.emit_frame_size =
+		SOC15_FLUSH_GPU_TLB_NUM_WREG * 6 +
+		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 8 +
+		8 + /* jpeg_v5_0_1_dec_ring_emit_vm_flush */
+		22 + 22 + /* jpeg_v5_0_1_dec_ring_emit_fence x2 vm fence */
+		8 + 16,
+	.emit_ib_size = 22, /* jpeg_v5_0_1_dec_ring_emit_ib */
+	.emit_ib = jpeg_v4_0_3_dec_ring_emit_ib,
+	.emit_fence = jpeg_v4_0_3_dec_ring_emit_fence,
+	.emit_vm_flush = jpeg_v4_0_3_dec_ring_emit_vm_flush,
+	.emit_hdp_flush = jpeg_v4_0_3_ring_emit_hdp_flush,
+	.test_ring = amdgpu_jpeg_dec_ring_test_ring,
+	.test_ib = amdgpu_jpeg_dec_ring_test_ib,
+	.insert_nop = jpeg_v4_0_3_dec_ring_nop,
+	.insert_start = jpeg_v4_0_3_dec_ring_insert_start,
+	.insert_end = jpeg_v4_0_3_dec_ring_insert_end,
+	.pad_ib = amdgpu_ring_generic_pad_ib,
+	.begin_use = amdgpu_jpeg_ring_begin_use,
+	.end_use = amdgpu_jpeg_ring_end_use,
+	.emit_wreg = jpeg_v4_0_3_dec_ring_emit_wreg,
+	.emit_reg_wait = jpeg_v4_0_3_dec_ring_emit_reg_wait,
+	.emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper,
+	.reset = jpeg_v5_0_1_ring_reset,
+};
+
+static void jpeg_v5_0_1_set_dec_ring_funcs(struct amdgpu_device *adev)
+{
+	int i, j, jpeg_inst;
+
+	for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
+		for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j) {
+			adev->jpeg.inst[i].ring_dec[j].funcs = &jpeg_v5_0_1_dec_ring_vm_funcs;
+			adev->jpeg.inst[i].ring_dec[j].me = i;
+			adev->jpeg.inst[i].ring_dec[j].pipe = j;
+		}
+		jpeg_inst = GET_INST(JPEG, i);
+		adev->jpeg.inst[i].aid_id =
+			jpeg_inst / adev->jpeg.num_inst_per_aid;
+	}
+}
+
+static const struct amdgpu_irq_src_funcs jpeg_v5_0_1_irq_funcs = {
+	.set = jpeg_v5_0_1_set_interrupt_state,
+	.process = jpeg_v5_0_1_process_interrupt,
+};
+
+static const struct amdgpu_irq_src_funcs jpeg_v5_0_1_ras_irq_funcs = {
+	.set = jpeg_v5_0_1_set_ras_interrupt_state,
+	.process = amdgpu_jpeg_process_poison_irq,
+};
+
+static void jpeg_v5_0_1_set_irq_funcs(struct amdgpu_device *adev)
+{
+	int i;
+
+	for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i)
+		adev->jpeg.inst->irq.num_types += adev->jpeg.num_jpeg_rings;
+
+	adev->jpeg.inst->irq.funcs = &jpeg_v5_0_1_irq_funcs;
+
+	adev->jpeg.inst->ras_poison_irq.num_types = 1;
+	adev->jpeg.inst->ras_poison_irq.funcs = &jpeg_v5_0_1_ras_irq_funcs;
+
+}
+
+const struct amdgpu_ip_block_version jpeg_v5_0_1_ip_block = {
+	.type = AMD_IP_BLOCK_TYPE_JPEG,
+	.major = 5,
+	.minor = 0,
+	.rev = 1,
+	.funcs = &jpeg_v5_0_1_ip_funcs,
+};
+
+static uint32_t jpeg_v5_0_1_query_poison_by_instance(struct amdgpu_device *adev,
+		uint32_t instance, uint32_t sub_block)
+{
+	uint32_t poison_stat = 0, reg_value = 0;
+
+	switch (sub_block) {
+	case AMDGPU_JPEG_V5_0_1_JPEG0:
+		reg_value = RREG32_SOC15(JPEG, instance, regUVD_RAS_JPEG0_STATUS);
+		poison_stat = REG_GET_FIELD(reg_value, UVD_RAS_JPEG0_STATUS, POISONED_PF);
+		break;
+	case AMDGPU_JPEG_V5_0_1_JPEG1:
+		reg_value = RREG32_SOC15(JPEG, instance, regUVD_RAS_JPEG1_STATUS);
+		poison_stat = REG_GET_FIELD(reg_value, UVD_RAS_JPEG1_STATUS, POISONED_PF);
+		break;
+	default:
+		break;
+	}
+
+	if (poison_stat)
+		dev_info(adev->dev, "Poison detected in JPEG%d sub_block%d\n",
+			instance, sub_block);
+
+	return poison_stat;
+}
+
+static bool jpeg_v5_0_1_query_ras_poison_status(struct amdgpu_device *adev)
+{
+	uint32_t inst = 0, sub = 0, poison_stat = 0;
+
+	for (inst = 0; inst < adev->jpeg.num_jpeg_inst; inst++)
+		for (sub = 0; sub < AMDGPU_JPEG_V5_0_1_MAX_SUB_BLOCK; sub++)
+			poison_stat +=
+			jpeg_v5_0_1_query_poison_by_instance(adev, inst, sub);
+
+	return !!poison_stat;
+}
+
+static const struct amdgpu_ras_block_hw_ops jpeg_v5_0_1_ras_hw_ops = {
+	.query_poison_status = jpeg_v5_0_1_query_ras_poison_status,
+};
+
+static int jpeg_v5_0_1_aca_bank_parser(struct aca_handle *handle, struct aca_bank *bank,
+				      enum aca_smu_type type, void *data)
+{
+	struct aca_bank_info info;
+	u64 misc0;
+	int ret;
+
+	ret = aca_bank_info_decode(bank, &info);
+	if (ret)
+		return ret;
+
+	misc0 = bank->regs[ACA_REG_IDX_MISC0];
+	switch (type) {
+	case ACA_SMU_TYPE_UE:
+		bank->aca_err_type = ACA_ERROR_TYPE_UE;
+		ret = aca_error_cache_log_bank_error(handle, &info, ACA_ERROR_TYPE_UE,
+						     1ULL);
+		break;
+	case ACA_SMU_TYPE_CE:
+		bank->aca_err_type = ACA_ERROR_TYPE_CE;
+		ret = aca_error_cache_log_bank_error(handle, &info, bank->aca_err_type,
+						     ACA_REG__MISC0__ERRCNT(misc0));
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return ret;
+}
+
+/* reference to smu driver if header file */
+static int jpeg_v5_0_1_err_codes[] = {
+	16, 17, 18, 19, 20, 21, 22, 23, /* JPEG[0-9][S|D] */
+	24, 25, 26, 27, 28, 29, 30, 31,
+	48, 49, 50, 51,
+};
+
+static bool jpeg_v5_0_1_aca_bank_is_valid(struct aca_handle *handle, struct aca_bank *bank,
+					 enum aca_smu_type type, void *data)
+{
+	u32 instlo;
+
+	instlo = ACA_REG__IPID__INSTANCEIDLO(bank->regs[ACA_REG_IDX_IPID]);
+	instlo &= GENMASK(31, 1);
+
+	if (instlo != mmSMNAID_AID0_MCA_SMU)
+		return false;
+
+	if (aca_bank_check_error_codes(handle->adev, bank,
+				       jpeg_v5_0_1_err_codes,
+				       ARRAY_SIZE(jpeg_v5_0_1_err_codes)))
+		return false;
+
+	return true;
+}
+
+static const struct aca_bank_ops jpeg_v5_0_1_aca_bank_ops = {
+	.aca_bank_parser = jpeg_v5_0_1_aca_bank_parser,
+	.aca_bank_is_valid = jpeg_v5_0_1_aca_bank_is_valid,
+};
+
+static const struct aca_info jpeg_v5_0_1_aca_info = {
+	.hwip = ACA_HWIP_TYPE_SMU,
+	.mask = ACA_ERROR_UE_MASK,
+	.bank_ops = &jpeg_v5_0_1_aca_bank_ops,
+};
+
+static int jpeg_v5_0_1_ras_late_init(struct amdgpu_device *adev, struct ras_common_if *ras_block)
+{
+	int r;
+
+	r = amdgpu_ras_block_late_init(adev, ras_block);
+	if (r)
+		return r;
+
+	r = amdgpu_ras_bind_aca(adev, AMDGPU_RAS_BLOCK__JPEG,
+				&jpeg_v5_0_1_aca_info, NULL);
+	if (r)
+		goto late_fini;
+
+	if (amdgpu_ras_is_supported(adev, ras_block->block) &&
+		adev->jpeg.inst->ras_poison_irq.funcs) {
+		r = amdgpu_irq_get(adev, &adev->jpeg.inst->ras_poison_irq, 0);
+		if (r)
+			goto late_fini;
+	}
+
+	return 0;
+
+late_fini:
+	amdgpu_ras_block_late_fini(adev, ras_block);
+
+	return r;
+}
+
+static struct amdgpu_jpeg_ras jpeg_v5_0_1_ras = {
+	.ras_block = {
+		.hw_ops = &jpeg_v5_0_1_ras_hw_ops,
+		.ras_late_init = jpeg_v5_0_1_ras_late_init,
+	},
+};
+
+static void jpeg_v5_0_1_set_ras_funcs(struct amdgpu_device *adev)
+{
+	adev->jpeg.ras = &jpeg_v5_0_1_ras;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_1.h b/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_1.h
new file mode 100644
index 000000000000..a7e58d5fb246
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_1.h
@@ -0,0 +1,111 @@
+/*
+ * Copyright 2024 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __JPEG_V5_0_1_H__
+#define __JPEG_V5_0_1_H__
+
+extern const struct amdgpu_ip_block_version jpeg_v5_0_1_ip_block;
+
+#define regUVD_JRBC0_UVD_JRBC_SCRATCH0_INTERNAL_OFFSET			0x4094
+#define regUVD_JRBC_EXTERNAL_MCM_ADDR_INTERNAL_OFFSET			0x1bffe
+
+#define regUVD_JRBC0_UVD_JRBC_RB_WPTR                                                         0x0640
+#define regUVD_JRBC0_UVD_JRBC_RB_WPTR_BASE_IDX                                                1
+#define regUVD_JRBC0_UVD_JRBC_STATUS                                                          0x0649
+#define regUVD_JRBC0_UVD_JRBC_STATUS_BASE_IDX                                                 1
+#define regUVD_JRBC0_UVD_JRBC_RB_RPTR                                                         0x064a
+#define regUVD_JRBC0_UVD_JRBC_RB_RPTR_BASE_IDX                                                1
+#define regUVD_JRBC1_UVD_JRBC_RB_WPTR                                                         0x0000
+#define regUVD_JRBC1_UVD_JRBC_RB_WPTR_BASE_IDX                                                0
+#define regUVD_JRBC1_UVD_JRBC_STATUS                                                          0x0009
+#define regUVD_JRBC1_UVD_JRBC_STATUS_BASE_IDX                                                 0
+#define regUVD_JRBC1_UVD_JRBC_RB_RPTR                                                         0x000a
+#define regUVD_JRBC1_UVD_JRBC_RB_RPTR_BASE_IDX                                                0
+#define regUVD_JRBC2_UVD_JRBC_RB_WPTR                                                         0x0040
+#define regUVD_JRBC2_UVD_JRBC_RB_WPTR_BASE_IDX                                                0
+#define regUVD_JRBC2_UVD_JRBC_STATUS                                                          0x0049
+#define regUVD_JRBC2_UVD_JRBC_STATUS_BASE_IDX                                                 0
+#define regUVD_JRBC2_UVD_JRBC_RB_RPTR                                                         0x004a
+#define regUVD_JRBC2_UVD_JRBC_RB_RPTR_BASE_IDX                                                0
+#define regUVD_JRBC3_UVD_JRBC_RB_WPTR                                                         0x0080
+#define regUVD_JRBC3_UVD_JRBC_RB_WPTR_BASE_IDX                                                0
+#define regUVD_JRBC3_UVD_JRBC_STATUS                                                          0x0089
+#define regUVD_JRBC3_UVD_JRBC_STATUS_BASE_IDX                                                 0
+#define regUVD_JRBC3_UVD_JRBC_RB_RPTR                                                         0x008a
+#define regUVD_JRBC3_UVD_JRBC_RB_RPTR_BASE_IDX                                                0
+#define regUVD_JRBC4_UVD_JRBC_RB_WPTR                                                         0x00c0
+#define regUVD_JRBC4_UVD_JRBC_RB_WPTR_BASE_IDX                                                0
+#define regUVD_JRBC4_UVD_JRBC_STATUS                                                          0x00c9
+#define regUVD_JRBC4_UVD_JRBC_STATUS_BASE_IDX                                                 0
+#define regUVD_JRBC4_UVD_JRBC_RB_RPTR                                                         0x00ca
+#define regUVD_JRBC4_UVD_JRBC_RB_RPTR_BASE_IDX                                                0
+#define regUVD_JRBC5_UVD_JRBC_RB_WPTR                                                         0x0100
+#define regUVD_JRBC5_UVD_JRBC_RB_WPTR_BASE_IDX                                                0
+#define regUVD_JRBC5_UVD_JRBC_STATUS                                                          0x0109
+#define regUVD_JRBC5_UVD_JRBC_STATUS_BASE_IDX                                                 0
+#define regUVD_JRBC5_UVD_JRBC_RB_RPTR                                                         0x010a
+#define regUVD_JRBC5_UVD_JRBC_RB_RPTR_BASE_IDX                                                0
+#define regUVD_JRBC6_UVD_JRBC_RB_WPTR                                                         0x0140
+#define regUVD_JRBC6_UVD_JRBC_RB_WPTR_BASE_IDX                                                0
+#define regUVD_JRBC6_UVD_JRBC_STATUS                                                          0x0149
+#define regUVD_JRBC6_UVD_JRBC_STATUS_BASE_IDX                                                 0
+#define regUVD_JRBC6_UVD_JRBC_RB_RPTR                                                         0x014a
+#define regUVD_JRBC6_UVD_JRBC_RB_RPTR_BASE_IDX                                                0
+#define regUVD_JRBC7_UVD_JRBC_RB_WPTR                                                         0x0180
+#define regUVD_JRBC7_UVD_JRBC_RB_WPTR_BASE_IDX                                                0
+#define regUVD_JRBC7_UVD_JRBC_STATUS                                                          0x0189
+#define regUVD_JRBC7_UVD_JRBC_STATUS_BASE_IDX                                                 0
+#define regUVD_JRBC7_UVD_JRBC_RB_RPTR                                                         0x018a
+#define regUVD_JRBC7_UVD_JRBC_RB_RPTR_BASE_IDX                                                0
+#define regUVD_JRBC8_UVD_JRBC_RB_WPTR                                                         0x01c0
+#define regUVD_JRBC8_UVD_JRBC_RB_WPTR_BASE_IDX                                                0
+#define regUVD_JRBC8_UVD_JRBC_STATUS                                                          0x01c9
+#define regUVD_JRBC8_UVD_JRBC_STATUS_BASE_IDX                                                 0
+#define regUVD_JRBC8_UVD_JRBC_RB_RPTR                                                         0x01ca
+#define regUVD_JRBC8_UVD_JRBC_RB_RPTR_BASE_IDX                                                0
+#define regUVD_JRBC9_UVD_JRBC_RB_WPTR                                                         0x0440
+#define regUVD_JRBC9_UVD_JRBC_RB_WPTR_BASE_IDX                                                1
+#define regUVD_JRBC9_UVD_JRBC_STATUS                                                          0x0449
+#define regUVD_JRBC9_UVD_JRBC_STATUS_BASE_IDX                                                 1
+#define regUVD_JRBC9_UVD_JRBC_RB_RPTR                                                         0x044a
+#define regUVD_JRBC9_UVD_JRBC_RB_RPTR_BASE_IDX                                                1
+#define regUVD_JMI0_JPEG_LMI_DROP                                                             0x0663
+#define regUVD_JMI0_JPEG_LMI_DROP_BASE_IDX                                                    1
+#define regUVD_JMI0_UVD_JMI_CLIENT_STALL                                                      0x067a
+#define regUVD_JMI0_UVD_JMI_CLIENT_STALL_BASE_IDX                                             1
+#define regUVD_JMI0_UVD_JMI_CLIENT_CLEAN_STATUS                                               0x067b
+#define regUVD_JMI0_UVD_JMI_CLIENT_CLEAN_STATUS_BASE_IDX                                      1
+#define regJPEG_CORE_RST_CTRL                                                                 0x072e
+#define regJPEG_CORE_RST_CTRL_BASE_IDX                                                        1
+
+#define regVCN_RRMT_CNTL                          0x0940
+#define regVCN_RRMT_CNTL_BASE_IDX                 1
+
+enum amdgpu_jpeg_v5_0_1_sub_block {
+	AMDGPU_JPEG_V5_0_1_JPEG0 = 0,
+	AMDGPU_JPEG_V5_0_1_JPEG1,
+
+	AMDGPU_JPEG_V5_0_1_MAX_SUB_BLOCK,
+};
+
+#endif /* __JPEG_V5_0_1_H__ */
diff --git a/drivers/gpu/drm/amd/amdgpu/lsdma_v6_0.c b/drivers/gpu/drm/amd/amdgpu/lsdma_v6_0.c
new file mode 100644
index 000000000000..1a285b531881
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/lsdma_v6_0.c
@@ -0,0 +1,121 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/delay.h>
+#include "amdgpu.h"
+#include "lsdma_v6_0.h"
+#include "amdgpu_lsdma.h"
+
+#include "lsdma/lsdma_6_0_0_offset.h"
+#include "lsdma/lsdma_6_0_0_sh_mask.h"
+
+static int lsdma_v6_0_wait_pio_status(struct amdgpu_device *adev)
+{
+	return amdgpu_lsdma_wait_for(adev, SOC15_REG_OFFSET(LSDMA, 0, regLSDMA_PIO_STATUS),
+			LSDMA_PIO_STATUS__PIO_IDLE_MASK | LSDMA_PIO_STATUS__PIO_FIFO_EMPTY_MASK,
+			LSDMA_PIO_STATUS__PIO_IDLE_MASK | LSDMA_PIO_STATUS__PIO_FIFO_EMPTY_MASK);
+}
+
+static int lsdma_v6_0_copy_mem(struct amdgpu_device *adev,
+			       uint64_t src_addr,
+			       uint64_t dst_addr,
+			       uint64_t size)
+{
+	int ret;
+	uint32_t tmp;
+
+	WREG32_SOC15(LSDMA, 0, regLSDMA_PIO_SRC_ADDR_LO, lower_32_bits(src_addr));
+	WREG32_SOC15(LSDMA, 0, regLSDMA_PIO_SRC_ADDR_HI, upper_32_bits(src_addr));
+
+	WREG32_SOC15(LSDMA, 0, regLSDMA_PIO_DST_ADDR_LO, lower_32_bits(dst_addr));
+	WREG32_SOC15(LSDMA, 0, regLSDMA_PIO_DST_ADDR_HI, upper_32_bits(dst_addr));
+
+	WREG32_SOC15(LSDMA, 0, regLSDMA_PIO_CONTROL, 0x0);
+
+	tmp = RREG32_SOC15(LSDMA, 0, regLSDMA_PIO_COMMAND);
+	tmp = REG_SET_FIELD(tmp, LSDMA_PIO_COMMAND, BYTE_COUNT, size);
+	tmp = REG_SET_FIELD(tmp, LSDMA_PIO_COMMAND, SRC_LOCATION, 0);
+	tmp = REG_SET_FIELD(tmp, LSDMA_PIO_COMMAND, DST_LOCATION, 0);
+	tmp = REG_SET_FIELD(tmp, LSDMA_PIO_COMMAND, SRC_ADDR_INC, 0);
+	tmp = REG_SET_FIELD(tmp, LSDMA_PIO_COMMAND, DST_ADDR_INC, 0);
+	tmp = REG_SET_FIELD(tmp, LSDMA_PIO_COMMAND, OVERLAP_DISABLE, 0);
+	tmp = REG_SET_FIELD(tmp, LSDMA_PIO_COMMAND, CONSTANT_FILL, 0);
+	WREG32_SOC15(LSDMA, 0, regLSDMA_PIO_COMMAND, tmp);
+
+	ret = lsdma_v6_0_wait_pio_status(adev);
+	if (ret)
+		dev_err(adev->dev, "LSDMA PIO failed to copy memory!\n");
+
+	return ret;
+}
+
+static int lsdma_v6_0_fill_mem(struct amdgpu_device *adev,
+			       uint64_t dst_addr,
+			       uint32_t data,
+			       uint64_t size)
+{
+	int ret;
+	uint32_t tmp;
+
+	WREG32_SOC15(LSDMA, 0, regLSDMA_PIO_CONSTFILL_DATA, data);
+
+	WREG32_SOC15(LSDMA, 0, regLSDMA_PIO_DST_ADDR_LO, lower_32_bits(dst_addr));
+	WREG32_SOC15(LSDMA, 0, regLSDMA_PIO_DST_ADDR_HI, upper_32_bits(dst_addr));
+
+	WREG32_SOC15(LSDMA, 0, regLSDMA_PIO_CONTROL, 0x0);
+
+	tmp = RREG32_SOC15(LSDMA, 0, regLSDMA_PIO_COMMAND);
+	tmp = REG_SET_FIELD(tmp, LSDMA_PIO_COMMAND, BYTE_COUNT, size);
+	tmp = REG_SET_FIELD(tmp, LSDMA_PIO_COMMAND, SRC_LOCATION, 0);
+	tmp = REG_SET_FIELD(tmp, LSDMA_PIO_COMMAND, DST_LOCATION, 0);
+	tmp = REG_SET_FIELD(tmp, LSDMA_PIO_COMMAND, SRC_ADDR_INC, 0);
+	tmp = REG_SET_FIELD(tmp, LSDMA_PIO_COMMAND, DST_ADDR_INC, 0);
+	tmp = REG_SET_FIELD(tmp, LSDMA_PIO_COMMAND, OVERLAP_DISABLE, 0);
+	tmp = REG_SET_FIELD(tmp, LSDMA_PIO_COMMAND, CONSTANT_FILL, 1);
+	WREG32_SOC15(LSDMA, 0, regLSDMA_PIO_COMMAND, tmp);
+
+	ret = lsdma_v6_0_wait_pio_status(adev);
+	if (ret)
+		dev_err(adev->dev, "LSDMA PIO failed to fill memory!\n");
+
+	return ret;
+}
+
+static void lsdma_v6_0_update_memory_power_gating(struct amdgpu_device *adev,
+						 bool enable)
+{
+	uint32_t tmp;
+
+	tmp = RREG32_SOC15(LSDMA, 0, regLSDMA_MEM_POWER_CTRL);
+	tmp = REG_SET_FIELD(tmp, LSDMA_MEM_POWER_CTRL, MEM_POWER_CTRL_EN, 0);
+	WREG32_SOC15(LSDMA, 0, regLSDMA_MEM_POWER_CTRL, tmp);
+
+	tmp = REG_SET_FIELD(tmp, LSDMA_MEM_POWER_CTRL, MEM_POWER_CTRL_EN, enable);
+	WREG32_SOC15(LSDMA, 0, regLSDMA_MEM_POWER_CTRL, tmp);
+}
+
+const struct amdgpu_lsdma_funcs lsdma_v6_0_funcs = {
+	.copy_mem = lsdma_v6_0_copy_mem,
+	.fill_mem = lsdma_v6_0_fill_mem,
+	.update_memory_power_gating = lsdma_v6_0_update_memory_power_gating
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/lsdma_v6_0.h b/drivers/gpu/drm/amd/amdgpu/lsdma_v6_0.h
new file mode 100644
index 000000000000..3ef79be1a9bf
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/lsdma_v6_0.h
@@ -0,0 +1,31 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __LSDMA_V6_0_H__
+#define __LSDMA_V6_0_H__
+
+#include "soc15_common.h"
+
+extern const struct amdgpu_lsdma_funcs lsdma_v6_0_funcs;
+
+#endif /* __LSDMA_V6_0_H__ */
diff --git a/drivers/gpu/drm/amd/amdgpu/lsdma_v7_0.c b/drivers/gpu/drm/amd/amdgpu/lsdma_v7_0.c
new file mode 100644
index 000000000000..396262044ea8
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/lsdma_v7_0.c
@@ -0,0 +1,121 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/delay.h>
+#include "amdgpu.h"
+#include "lsdma_v7_0.h"
+#include "amdgpu_lsdma.h"
+
+#include "lsdma/lsdma_7_0_0_offset.h"
+#include "lsdma/lsdma_7_0_0_sh_mask.h"
+
+static int lsdma_v7_0_wait_pio_status(struct amdgpu_device *adev)
+{
+	return amdgpu_lsdma_wait_for(adev, SOC15_REG_OFFSET(LSDMA, 0, regLSDMA_PIO_STATUS),
+			LSDMA_PIO_STATUS__PIO_IDLE_MASK | LSDMA_PIO_STATUS__PIO_FIFO_EMPTY_MASK,
+			LSDMA_PIO_STATUS__PIO_IDLE_MASK | LSDMA_PIO_STATUS__PIO_FIFO_EMPTY_MASK);
+}
+
+static int lsdma_v7_0_copy_mem(struct amdgpu_device *adev,
+			       uint64_t src_addr,
+			       uint64_t dst_addr,
+			       uint64_t size)
+{
+	int ret;
+	uint32_t tmp;
+
+	WREG32_SOC15(LSDMA, 0, regLSDMA_PIO_SRC_ADDR_LO, lower_32_bits(src_addr));
+	WREG32_SOC15(LSDMA, 0, regLSDMA_PIO_SRC_ADDR_HI, upper_32_bits(src_addr));
+
+	WREG32_SOC15(LSDMA, 0, regLSDMA_PIO_DST_ADDR_LO, lower_32_bits(dst_addr));
+	WREG32_SOC15(LSDMA, 0, regLSDMA_PIO_DST_ADDR_HI, upper_32_bits(dst_addr));
+
+	WREG32_SOC15(LSDMA, 0, regLSDMA_PIO_CONTROL, 0x0);
+
+	tmp = RREG32_SOC15(LSDMA, 0, regLSDMA_PIO_COMMAND);
+	tmp = REG_SET_FIELD(tmp, LSDMA_PIO_COMMAND, BYTE_COUNT, size);
+	tmp = REG_SET_FIELD(tmp, LSDMA_PIO_COMMAND, SRC_LOCATION, 0);
+	tmp = REG_SET_FIELD(tmp, LSDMA_PIO_COMMAND, DST_LOCATION, 0);
+	tmp = REG_SET_FIELD(tmp, LSDMA_PIO_COMMAND, SRC_ADDR_INC, 0);
+	tmp = REG_SET_FIELD(tmp, LSDMA_PIO_COMMAND, DST_ADDR_INC, 0);
+	tmp = REG_SET_FIELD(tmp, LSDMA_PIO_COMMAND, OVERLAP_DISABLE, 0);
+	tmp = REG_SET_FIELD(tmp, LSDMA_PIO_COMMAND, CONSTANT_FILL, 0);
+	WREG32_SOC15(LSDMA, 0, regLSDMA_PIO_COMMAND, tmp);
+
+	ret = lsdma_v7_0_wait_pio_status(adev);
+	if (ret)
+		dev_err(adev->dev, "LSDMA PIO failed to copy memory!\n");
+
+	return ret;
+}
+
+static int lsdma_v7_0_fill_mem(struct amdgpu_device *adev,
+			       uint64_t dst_addr,
+			       uint32_t data,
+			       uint64_t size)
+{
+	int ret;
+	uint32_t tmp;
+
+	WREG32_SOC15(LSDMA, 0, regLSDMA_PIO_CONSTFILL_DATA, data);
+
+	WREG32_SOC15(LSDMA, 0, regLSDMA_PIO_DST_ADDR_LO, lower_32_bits(dst_addr));
+	WREG32_SOC15(LSDMA, 0, regLSDMA_PIO_DST_ADDR_HI, upper_32_bits(dst_addr));
+
+	WREG32_SOC15(LSDMA, 0, regLSDMA_PIO_CONTROL, 0x0);
+
+	tmp = RREG32_SOC15(LSDMA, 0, regLSDMA_PIO_COMMAND);
+	tmp = REG_SET_FIELD(tmp, LSDMA_PIO_COMMAND, BYTE_COUNT, size);
+	tmp = REG_SET_FIELD(tmp, LSDMA_PIO_COMMAND, SRC_LOCATION, 0);
+	tmp = REG_SET_FIELD(tmp, LSDMA_PIO_COMMAND, DST_LOCATION, 0);
+	tmp = REG_SET_FIELD(tmp, LSDMA_PIO_COMMAND, SRC_ADDR_INC, 0);
+	tmp = REG_SET_FIELD(tmp, LSDMA_PIO_COMMAND, DST_ADDR_INC, 0);
+	tmp = REG_SET_FIELD(tmp, LSDMA_PIO_COMMAND, OVERLAP_DISABLE, 0);
+	tmp = REG_SET_FIELD(tmp, LSDMA_PIO_COMMAND, CONSTANT_FILL, 1);
+	WREG32_SOC15(LSDMA, 0, regLSDMA_PIO_COMMAND, tmp);
+
+	ret = lsdma_v7_0_wait_pio_status(adev);
+	if (ret)
+		dev_err(adev->dev, "LSDMA PIO failed to fill memory!\n");
+
+	return ret;
+}
+
+static void lsdma_v7_0_update_memory_power_gating(struct amdgpu_device *adev,
+						 bool enable)
+{
+	uint32_t tmp;
+
+	tmp = RREG32_SOC15(LSDMA, 0, regLSDMA_MEM_POWER_CTRL);
+	tmp = REG_SET_FIELD(tmp, LSDMA_MEM_POWER_CTRL, MEM_POWER_CTRL_EN, 0);
+	WREG32_SOC15(LSDMA, 0, regLSDMA_MEM_POWER_CTRL, tmp);
+
+	tmp = REG_SET_FIELD(tmp, LSDMA_MEM_POWER_CTRL, MEM_POWER_CTRL_EN, enable);
+	WREG32_SOC15(LSDMA, 0, regLSDMA_MEM_POWER_CTRL, tmp);
+}
+
+const struct amdgpu_lsdma_funcs lsdma_v7_0_funcs = {
+	.copy_mem = lsdma_v7_0_copy_mem,
+	.fill_mem = lsdma_v7_0_fill_mem,
+	.update_memory_power_gating = lsdma_v7_0_update_memory_power_gating
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/lsdma_v7_0.h b/drivers/gpu/drm/amd/amdgpu/lsdma_v7_0.h
new file mode 100644
index 000000000000..52b4485cdd98
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/lsdma_v7_0.h
@@ -0,0 +1,31 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __LSDMA_V7_0_H__
+#define __LSDMA_V7_0_H__
+
+#include "soc15_common.h"
+
+extern const struct amdgpu_lsdma_funcs lsdma_v7_0_funcs;
+
+#endif /* __LSDMA_V7_0_H__ */
diff --git a/drivers/gpu/drm/amd/amdgpu/mca_v3_0.c b/drivers/gpu/drm/amd/amdgpu/mca_v3_0.c
index 8f7107d392af..6dae4a2e2767 100644
--- a/drivers/gpu/drm/amd/amdgpu/mca_v3_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/mca_v3_0.c
@@ -37,24 +37,30 @@ static void mca_v3_0_mp0_query_ras_error_count(struct amdgpu_device *adev,
 				         ras_error_status);
 }
 
-static int mca_v3_0_mp0_ras_late_init(struct amdgpu_device *adev)
+static int mca_v3_0_ras_block_match(struct amdgpu_ras_block_object *block_obj,
+				enum amdgpu_ras_block block, uint32_t sub_block_index)
 {
-	return amdgpu_mca_ras_late_init(adev, &adev->mca.mp0);
-}
+	if (!block_obj)
+		return -EINVAL;
 
-static void mca_v3_0_mp0_ras_fini(struct amdgpu_device *adev)
-{
-	amdgpu_mca_ras_fini(adev, &adev->mca.mp0);
+	if ((block_obj->ras_comm.block == block) &&
+		(block_obj->ras_comm.sub_block_index == sub_block_index)) {
+		return 0;
+	}
+
+	return -EINVAL;
 }
 
-const struct amdgpu_mca_ras_funcs mca_v3_0_mp0_ras_funcs = {
-	.ras_late_init = mca_v3_0_mp0_ras_late_init,
-	.ras_fini = mca_v3_0_mp0_ras_fini,
+static const struct amdgpu_ras_block_hw_ops mca_v3_0_mp0_hw_ops = {
 	.query_ras_error_count = mca_v3_0_mp0_query_ras_error_count,
 	.query_ras_error_address = NULL,
-	.ras_block = AMDGPU_RAS_BLOCK__MCA,
-	.ras_sub_block = AMDGPU_RAS_MCA_BLOCK__MP0,
-	.sysfs_name = "mp0_err_count",
+};
+
+struct amdgpu_mca_ras_block mca_v3_0_mp0_ras = {
+	.ras_block = {
+		.hw_ops = &mca_v3_0_mp0_hw_ops,
+		.ras_block_match = mca_v3_0_ras_block_match,
+	},
 };
 
 static void mca_v3_0_mp1_query_ras_error_count(struct amdgpu_device *adev,
@@ -65,24 +71,16 @@ static void mca_v3_0_mp1_query_ras_error_count(struct amdgpu_device *adev,
 				         ras_error_status);
 }
 
-static int mca_v3_0_mp1_ras_late_init(struct amdgpu_device *adev)
-{
-	return amdgpu_mca_ras_late_init(adev, &adev->mca.mp1);
-}
-
-static void mca_v3_0_mp1_ras_fini(struct amdgpu_device *adev)
-{
-	amdgpu_mca_ras_fini(adev, &adev->mca.mp1);
-}
-
-const struct amdgpu_mca_ras_funcs mca_v3_0_mp1_ras_funcs = {
-	.ras_late_init = mca_v3_0_mp1_ras_late_init,
-	.ras_fini = mca_v3_0_mp1_ras_fini,
+static const struct amdgpu_ras_block_hw_ops mca_v3_0_mp1_hw_ops = {
 	.query_ras_error_count = mca_v3_0_mp1_query_ras_error_count,
 	.query_ras_error_address = NULL,
-	.ras_block = AMDGPU_RAS_BLOCK__MCA,
-	.ras_sub_block = AMDGPU_RAS_MCA_BLOCK__MP1,
-	.sysfs_name = "mp1_err_count",
+};
+
+struct amdgpu_mca_ras_block mca_v3_0_mp1_ras = {
+	.ras_block = {
+		.hw_ops = &mca_v3_0_mp1_hw_ops,
+		.ras_block_match = mca_v3_0_ras_block_match,
+	},
 };
 
 static void mca_v3_0_mpio_query_ras_error_count(struct amdgpu_device *adev,
@@ -93,36 +91,14 @@ static void mca_v3_0_mpio_query_ras_error_count(struct amdgpu_device *adev,
 				         ras_error_status);
 }
 
-static int mca_v3_0_mpio_ras_late_init(struct amdgpu_device *adev)
-{
-	return amdgpu_mca_ras_late_init(adev, &adev->mca.mpio);
-}
-
-static void mca_v3_0_mpio_ras_fini(struct amdgpu_device *adev)
-{
-	amdgpu_mca_ras_fini(adev, &adev->mca.mpio);
-}
-
-const struct amdgpu_mca_ras_funcs mca_v3_0_mpio_ras_funcs = {
-	.ras_late_init = mca_v3_0_mpio_ras_late_init,
-	.ras_fini = mca_v3_0_mpio_ras_fini,
+static const struct amdgpu_ras_block_hw_ops mca_v3_0_mpio_hw_ops = {
 	.query_ras_error_count = mca_v3_0_mpio_query_ras_error_count,
 	.query_ras_error_address = NULL,
-	.ras_block = AMDGPU_RAS_BLOCK__MCA,
-	.ras_sub_block = AMDGPU_RAS_MCA_BLOCK__MPIO,
-	.sysfs_name = "mpio_err_count",
 };
 
-
-static void mca_v3_0_init(struct amdgpu_device *adev)
-{
-	struct amdgpu_mca *mca = &adev->mca;
-
-	mca->mp0.ras_funcs = &mca_v3_0_mp0_ras_funcs;
-	mca->mp1.ras_funcs = &mca_v3_0_mp1_ras_funcs;
-	mca->mpio.ras_funcs = &mca_v3_0_mpio_ras_funcs;
-}
-
-const struct amdgpu_mca_funcs mca_v3_0_funcs = {
-	.init = mca_v3_0_init,
-};
-\ No newline at end of file
+struct amdgpu_mca_ras_block mca_v3_0_mpio_ras = {
+	.ras_block = {
+		.hw_ops = &mca_v3_0_mpio_hw_ops,
+		.ras_block_match = mca_v3_0_ras_block_match,
+	},
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/mca_v3_0.h b/drivers/gpu/drm/amd/amdgpu/mca_v3_0.h
index b899b86194c2..d3eaef0d7f2d 100644
--- a/drivers/gpu/drm/amd/amdgpu/mca_v3_0.h
+++ b/drivers/gpu/drm/amd/amdgpu/mca_v3_0.h
@@ -21,6 +21,8 @@
 #ifndef __MCA_V3_0_H__
 #define __MCA_V3_0_H__
 
-extern const struct amdgpu_mca_funcs mca_v3_0_funcs;
+extern struct amdgpu_mca_ras_block mca_v3_0_mp0_ras;
+extern struct amdgpu_mca_ras_block mca_v3_0_mp1_ras;
+extern struct amdgpu_mca_ras_block mca_v3_0_mpio_ras;
 
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/mes_api_def.h b/drivers/gpu/drm/amd/amdgpu/mes_api_def.h
deleted file mode 100644
index 3f4fca5fd1da..000000000000
--- a/drivers/gpu/drm/amd/amdgpu/mes_api_def.h
+++ /dev/null
@@ -1,443 +0,0 @@
-/*
- * Copyright 2019 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- */
-
-#ifndef __MES_API_DEF_H__
-#define __MES_API_DEF_H__
-
-#pragma pack(push, 4)
-
-#define MES_API_VERSION 1
-
-/* Driver submits one API(cmd) as a single Frame and this command size is same
- * for all API to ease the debugging and parsing of ring buffer.
- */
-enum { API_FRAME_SIZE_IN_DWORDS = 64 };
-
-/* To avoid command in scheduler context to be overwritten whenenver mutilple
- * interrupts come in, this creates another queue.
- */
-enum { API_NUMBER_OF_COMMAND_MAX = 32 };
-
-enum MES_API_TYPE {
-	MES_API_TYPE_SCHEDULER = 1,
-	MES_API_TYPE_MAX
-};
-
-enum MES_SCH_API_OPCODE {
-	MES_SCH_API_SET_HW_RSRC			= 0,
-	MES_SCH_API_SET_SCHEDULING_CONFIG	= 1, /* agreegated db, quantums, etc */
-	MES_SCH_API_ADD_QUEUE			= 2,
-	MES_SCH_API_REMOVE_QUEUE		= 3,
-	MES_SCH_API_PERFORM_YIELD		= 4,
-	MES_SCH_API_SET_GANG_PRIORITY_LEVEL	= 5,
-	MES_SCH_API_SUSPEND			= 6,
-	MES_SCH_API_RESUME			= 7,
-	MES_SCH_API_RESET			= 8,
-	MES_SCH_API_SET_LOG_BUFFER		= 9,
-	MES_SCH_API_CHANGE_GANG_PRORITY		= 10,
-	MES_SCH_API_QUERY_SCHEDULER_STATUS	= 11,
-	MES_SCH_API_PROGRAM_GDS			= 12,
-	MES_SCH_API_SET_DEBUG_VMID		= 13,
-	MES_SCH_API_MISC			= 14,
-	MES_SCH_API_MAX				= 0xFF
-};
-
-union MES_API_HEADER {
-	struct {
-		uint32_t type		: 4; /* 0 - Invalid; 1 - Scheduling; 2 - TBD */
-		uint32_t opcode		: 8;
-		uint32_t dwsize		: 8; /* including header */
-		uint32_t reserved	: 12;
-	};
-
-	uint32_t	u32All;
-};
-
-enum MES_AMD_PRIORITY_LEVEL {
-	AMD_PRIORITY_LEVEL_LOW		= 0,
-	AMD_PRIORITY_LEVEL_NORMAL	= 1,
-	AMD_PRIORITY_LEVEL_MEDIUM	= 2,
-	AMD_PRIORITY_LEVEL_HIGH		= 3,
-	AMD_PRIORITY_LEVEL_REALTIME	= 4,
-	AMD_PRIORITY_NUM_LEVELS
-};
-
-enum MES_QUEUE_TYPE {
-	MES_QUEUE_TYPE_GFX,
-	MES_QUEUE_TYPE_COMPUTE,
-	MES_QUEUE_TYPE_SDMA,
-	MES_QUEUE_TYPE_MAX,
-};
-
-struct MES_API_STATUS {
-	uint64_t	api_completion_fence_addr;
-	uint64_t	api_completion_fence_value;
-};
-
-enum { MAX_COMPUTE_PIPES = 8 };
-enum { MAX_GFX_PIPES = 2 };
-enum { MAX_SDMA_PIPES = 2 };
-
-enum { MAX_COMPUTE_HQD_PER_PIPE = 8 };
-enum { MAX_GFX_HQD_PER_PIPE = 8 };
-enum { MAX_SDMA_HQD_PER_PIPE = 10 };
-
-enum { MAX_QUEUES_IN_A_GANG = 8 };
-
-enum VM_HUB_TYPE {
-	VM_HUB_TYPE_GC = 0,
-	VM_HUB_TYPE_MM = 1,
-	VM_HUB_TYPE_MAX,
-};
-
-enum { VMID_INVALID = 0xffff };
-
-enum { MAX_VMID_GCHUB = 16 };
-enum { MAX_VMID_MMHUB = 16 };
-
-enum MES_LOG_OPERATION {
-	MES_LOG_OPERATION_CONTEXT_STATE_CHANGE = 0
-};
-
-enum MES_LOG_CONTEXT_STATE {
-	MES_LOG_CONTEXT_STATE_IDLE		= 0,
-	MES_LOG_CONTEXT_STATE_RUNNING		= 1,
-	MES_LOG_CONTEXT_STATE_READY		= 2,
-	MES_LOG_CONTEXT_STATE_READY_STANDBY	= 3,
-};
-
-struct MES_LOG_CONTEXT_STATE_CHANGE {
-	void				*h_context;
-	enum MES_LOG_CONTEXT_STATE	new_context_state;
-};
-
-struct MES_LOG_ENTRY_HEADER {
-	uint32_t	first_free_entry_index;
-	uint32_t	wraparound_count;
-	uint64_t	number_of_entries;
-	uint64_t	reserved[2];
-};
-
-struct MES_LOG_ENTRY_DATA {
-	uint64_t	gpu_time_stamp;
-	uint32_t	operation_type; /* operation_type is of MES_LOG_OPERATION type */
-	uint32_t	reserved_operation_type_bits;
-	union {
-		struct MES_LOG_CONTEXT_STATE_CHANGE	context_state_change;
-		uint64_t				reserved_operation_data[2];
-	};
-};
-
-struct MES_LOG_BUFFER {
-	struct MES_LOG_ENTRY_HEADER	header;
-	struct MES_LOG_ENTRY_DATA	entries[1];
-};
-
-union MESAPI_SET_HW_RESOURCES {
-	struct {
-		union MES_API_HEADER	header;
-		uint32_t		vmid_mask_mmhub;
-		uint32_t		vmid_mask_gfxhub;
-		uint32_t		gds_size;
-		uint32_t		paging_vmid;
-		uint32_t		compute_hqd_mask[MAX_COMPUTE_PIPES];
-		uint32_t		gfx_hqd_mask[MAX_GFX_PIPES];
-		uint32_t		sdma_hqd_mask[MAX_SDMA_PIPES];
-		uint32_t		agreegated_doorbells[AMD_PRIORITY_NUM_LEVELS];
-		uint64_t		g_sch_ctx_gpu_mc_ptr;
-		uint64_t		query_status_fence_gpu_mc_ptr;
-		struct MES_API_STATUS	api_status;
-		union {
-			struct {
-				uint32_t disable_reset	: 1;
-				uint32_t reserved	: 31;
-			};
-			uint32_t	uint32_t_all;
-		};
-	};
-
-	uint32_t	max_dwords_in_api[API_FRAME_SIZE_IN_DWORDS];
-};
-
-union MESAPI__ADD_QUEUE {
-	struct {
-		union MES_API_HEADER		header;
-		uint32_t			process_id;
-		uint64_t			page_table_base_addr;
-		uint64_t			process_va_start;
-		uint64_t			process_va_end;
-		uint64_t			process_quantum;
-		uint64_t			process_context_addr;
-		uint64_t			gang_quantum;
-		uint64_t			gang_context_addr;
-		uint32_t			inprocess_gang_priority;
-		enum MES_AMD_PRIORITY_LEVEL	gang_global_priority_level;
-		uint32_t			doorbell_offset;
-		uint64_t			mqd_addr;
-		uint64_t			wptr_addr;
-		enum MES_QUEUE_TYPE		queue_type;
-		uint32_t			gds_base;
-		uint32_t			gds_size;
-		uint32_t			gws_base;
-		uint32_t			gws_size;
-		uint32_t			oa_mask;
-
-		struct {
-			uint32_t paging			: 1;
-			uint32_t debug_vmid		: 4;
-			uint32_t program_gds		: 1;
-			uint32_t is_gang_suspended	: 1;
-			uint32_t is_tmz_queue		: 1;
-			uint32_t reserved		: 24;
-		};
-		struct MES_API_STATUS		api_status;
-	};
-
-	uint32_t	max_dwords_in_api[API_FRAME_SIZE_IN_DWORDS];
-};
-
-union MESAPI__REMOVE_QUEUE {
-	struct {
-		union MES_API_HEADER	header;
-		uint32_t		doorbell_offset;
-		uint64_t		gang_context_addr;
-
-		struct {
-			uint32_t unmap_legacy_gfx_queue	: 1;
-			uint32_t reserved		: 31;
-		};
-		struct MES_API_STATUS	api_status;
-	};
-
-	uint32_t	max_dwords_in_api[API_FRAME_SIZE_IN_DWORDS];
-};
-
-union MESAPI__SET_SCHEDULING_CONFIG {
-	struct {
-		union MES_API_HEADER	header;
-		/* Grace period when preempting another priority band for this
-		 * priority band. The value for idle priority band is ignored,
-		 * as it never preempts other bands.
-		 */
-		uint64_t		grace_period_other_levels[AMD_PRIORITY_NUM_LEVELS];
-		/* Default quantum for scheduling across processes within
-		 * a priority band.
-		 */
-		uint64_t		process_quantum_for_level[AMD_PRIORITY_NUM_LEVELS];
-		/* Default grace period for processes that preempt each other
-		 * within a priority band.
-		 */
-		uint64_t		process_grace_period_same_level[AMD_PRIORITY_NUM_LEVELS];
-		/* For normal level this field specifies the target GPU
-		 * percentage in situations when it's starved by the high level.
-		 * Valid values are between 0 and 50, with the default being 10.
-		 */
-		uint32_t		normal_yield_percent;
-		struct MES_API_STATUS	api_status;
-	};
-
-	uint32_t	max_dwords_in_api[API_FRAME_SIZE_IN_DWORDS];
-};
-
-union MESAPI__PERFORM_YIELD {
-	struct {
-		union MES_API_HEADER	header;
-		uint32_t		dummy;
-		struct MES_API_STATUS	api_status;
-	};
-
-	uint32_t	max_dwords_in_api[API_FRAME_SIZE_IN_DWORDS];
-};
-
-union MESAPI__CHANGE_GANG_PRIORITY_LEVEL {
-	struct {
-		union MES_API_HEADER		header;
-		uint32_t			inprocess_gang_priority;
-		enum MES_AMD_PRIORITY_LEVEL	gang_global_priority_level;
-		uint64_t			gang_quantum;
-		uint64_t			gang_context_addr;
-		struct MES_API_STATUS		api_status;
-	};
-
-	uint32_t	max_dwords_in_api[API_FRAME_SIZE_IN_DWORDS];
-};
-
-union MESAPI__SUSPEND {
-	struct {
-		union MES_API_HEADER	header;
-		/* false - suspend all gangs; true - specific gang */
-		struct {
-			uint32_t suspend_all_gangs	: 1;
-			uint32_t reserved		: 31;
-		};
-		/* gang_context_addr is valid only if suspend_all = false */
-		uint64_t		gang_context_addr;
-
-		uint64_t		suspend_fence_addr;
-		uint32_t		suspend_fence_value;
-
-		struct MES_API_STATUS	api_status;
-	};
-
-	uint32_t	max_dwords_in_api[API_FRAME_SIZE_IN_DWORDS];
-};
-
-union MESAPI__RESUME {
-	struct {
-		union MES_API_HEADER	header;
-		/* false - resume all gangs; true - specified gang */
-		struct {
-			uint32_t resume_all_gangs	: 1;
-			uint32_t reserved		: 31;
-		};
-		/* valid only if resume_all_gangs = false */
-		uint64_t		gang_context_addr;
-
-		struct MES_API_STATUS	api_status;
-	};
-
-	uint32_t	max_dwords_in_api[API_FRAME_SIZE_IN_DWORDS];
-};
-
-union MESAPI__RESET {
-	struct {
-		union MES_API_HEADER	header;
-
-		struct {
-			uint32_t reset_queue	: 1;
-			uint32_t reserved	: 31;
-		};
-
-		uint64_t		gang_context_addr;
-		uint32_t		doorbell_offset; /* valid only if reset_queue = true */
-		struct MES_API_STATUS	api_status;
-	};
-
-	uint32_t	max_dwords_in_api[API_FRAME_SIZE_IN_DWORDS];
-};
-
-union MESAPI__SET_LOGGING_BUFFER {
-	struct {
-		union MES_API_HEADER	header;
-		/* There are separate log buffers for each queue type */
-		enum MES_QUEUE_TYPE	log_type;
-		/* Log buffer GPU Address */
-		uint64_t		logging_buffer_addr;
-		/* number of entries in the log buffer */
-		uint32_t		number_of_entries;
-		/* Entry index at which CPU interrupt needs to be signalled */
-		uint32_t		interrupt_entry;
-
-		struct MES_API_STATUS	api_status;
-	};
-
-	uint32_t	max_dwords_in_api[API_FRAME_SIZE_IN_DWORDS];
-};
-
-union MESAPI__QUERY_MES_STATUS {
-	struct {
-		union MES_API_HEADER	header;
-		bool			mes_healthy; /* 0 - not healthy, 1 - healthy */
-		struct MES_API_STATUS	api_status;
-	};
-
-	uint32_t	max_dwords_in_api[API_FRAME_SIZE_IN_DWORDS];
-};
-
-union MESAPI__PROGRAM_GDS {
-	struct {
-		union MES_API_HEADER	header;
-		uint64_t		process_context_addr;
-		uint32_t		gds_base;
-		uint32_t		gds_size;
-		uint32_t		gws_base;
-		uint32_t		gws_size;
-		uint32_t		oa_mask;
-		struct MES_API_STATUS	api_status;
-	};
-
-	uint32_t	max_dwords_in_api[API_FRAME_SIZE_IN_DWORDS];
-};
-
-union MESAPI__SET_DEBUG_VMID {
-	struct {
-		union MES_API_HEADER	header;
-		struct MES_API_STATUS	api_status;
-		union {
-			struct {
-				uint32_t use_gds	: 1;
-				uint32_t reserved	: 31;
-			} flags;
-			uint32_t	u32All;
-		};
-		uint32_t		reserved;
-		uint32_t		debug_vmid;
-		uint64_t		process_context_addr;
-		uint64_t		page_table_base_addr;
-		uint64_t		process_va_start;
-		uint64_t		process_va_end;
-		uint32_t		gds_base;
-		uint32_t		gds_size;
-		uint32_t		gws_base;
-		uint32_t		gws_size;
-		uint32_t		oa_mask;
-	};
-
-	uint32_t	max_dwords_in_api[API_FRAME_SIZE_IN_DWORDS];
-};
-
-enum MESAPI_MISC_OPCODE {
-	MESAPI_MISC__MODIFY_REG,
-	MESAPI_MISC__MAX,
-};
-
-enum MODIFY_REG_SUBCODE {
-	MODIFY_REG__OVERWRITE,
-	MODIFY_REG__RMW_OR,
-	MODIFY_REG__RMW_AND,
-	MODIFY_REG__MAX,
-};
-
-enum { MISC_DATA_MAX_SIZE_IN_DWORDS = 20 };
-
-union MESAPI__MISC {
-	struct {
-		union MES_API_HEADER	header;
-		enum MESAPI_MISC_OPCODE	opcode;
-		struct MES_API_STATUS	api_status;
-
-		union {
-			struct {
-				enum MODIFY_REG_SUBCODE	subcode;
-				uint32_t		reg_offset;
-				uint32_t		reg_value;
-			} modify_reg;
-			uint32_t	data[MISC_DATA_MAX_SIZE_IN_DWORDS];
-		};
-	};
-
-	uint32_t	max_dwords_in_api[API_FRAME_SIZE_IN_DWORDS];
-};
-
-#pragma pack(pop)
-#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/mes_userqueue.c b/drivers/gpu/drm/amd/amdgpu/mes_userqueue.c
new file mode 100644
index 000000000000..64cae89357b6
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/mes_userqueue.c
@@ -0,0 +1,501 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright 2024 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include <drm/drm_drv.h>
+#include "amdgpu.h"
+#include "amdgpu_gfx.h"
+#include "mes_userqueue.h"
+#include "amdgpu_userq_fence.h"
+
+#define AMDGPU_USERQ_PROC_CTX_SZ PAGE_SIZE
+#define AMDGPU_USERQ_GANG_CTX_SZ PAGE_SIZE
+
+static int
+mes_userq_map_gtt_bo_to_gart(struct amdgpu_bo *bo)
+{
+	int ret;
+
+	ret = amdgpu_bo_reserve(bo, true);
+	if (ret) {
+		DRM_ERROR("Failed to reserve bo. ret %d\n", ret);
+		goto err_reserve_bo_failed;
+	}
+
+	ret = amdgpu_ttm_alloc_gart(&bo->tbo);
+	if (ret) {
+		DRM_ERROR("Failed to bind bo to GART. ret %d\n", ret);
+		goto err_map_bo_gart_failed;
+	}
+
+	amdgpu_bo_unreserve(bo);
+	bo = amdgpu_bo_ref(bo);
+
+	return 0;
+
+err_map_bo_gart_failed:
+	amdgpu_bo_unreserve(bo);
+err_reserve_bo_failed:
+	return ret;
+}
+
+static int
+mes_userq_create_wptr_mapping(struct amdgpu_userq_mgr *uq_mgr,
+			      struct amdgpu_usermode_queue *queue,
+			      uint64_t wptr)
+{
+	struct amdgpu_bo_va_mapping *wptr_mapping;
+	struct amdgpu_vm *wptr_vm;
+	struct amdgpu_userq_obj *wptr_obj = &queue->wptr_obj;
+	int ret;
+
+	wptr_vm = queue->vm;
+	ret = amdgpu_bo_reserve(wptr_vm->root.bo, false);
+	if (ret)
+		return ret;
+
+	wptr &= AMDGPU_GMC_HOLE_MASK;
+	wptr_mapping = amdgpu_vm_bo_lookup_mapping(wptr_vm, wptr >> PAGE_SHIFT);
+	amdgpu_bo_unreserve(wptr_vm->root.bo);
+	if (!wptr_mapping) {
+		DRM_ERROR("Failed to lookup wptr bo\n");
+		return -EINVAL;
+	}
+
+	wptr_obj->obj = wptr_mapping->bo_va->base.bo;
+	if (wptr_obj->obj->tbo.base.size > PAGE_SIZE) {
+		DRM_ERROR("Requested GART mapping for wptr bo larger than one page\n");
+		return -EINVAL;
+	}
+
+	ret = mes_userq_map_gtt_bo_to_gart(wptr_obj->obj);
+	if (ret) {
+		DRM_ERROR("Failed to map wptr bo to GART\n");
+		return ret;
+	}
+
+	queue->wptr_obj.gpu_addr = amdgpu_bo_gpu_offset_no_check(wptr_obj->obj);
+	return 0;
+}
+
+static int convert_to_mes_priority(int priority)
+{
+	switch (priority) {
+	case AMDGPU_USERQ_CREATE_FLAGS_QUEUE_PRIORITY_NORMAL_LOW:
+	default:
+		return AMDGPU_MES_PRIORITY_LEVEL_NORMAL;
+	case AMDGPU_USERQ_CREATE_FLAGS_QUEUE_PRIORITY_LOW:
+		return AMDGPU_MES_PRIORITY_LEVEL_LOW;
+	case AMDGPU_USERQ_CREATE_FLAGS_QUEUE_PRIORITY_NORMAL_HIGH:
+		return AMDGPU_MES_PRIORITY_LEVEL_MEDIUM;
+	case AMDGPU_USERQ_CREATE_FLAGS_QUEUE_PRIORITY_HIGH:
+		return AMDGPU_MES_PRIORITY_LEVEL_HIGH;
+	}
+}
+
+static int mes_userq_map(struct amdgpu_userq_mgr *uq_mgr,
+			 struct amdgpu_usermode_queue *queue)
+{
+	struct amdgpu_device *adev = uq_mgr->adev;
+	struct amdgpu_userq_obj *ctx = &queue->fw_obj;
+	struct amdgpu_mqd_prop *userq_props = queue->userq_prop;
+	struct mes_add_queue_input queue_input;
+	int r;
+
+	memset(&queue_input, 0x0, sizeof(struct mes_add_queue_input));
+
+	queue_input.process_va_start = 0;
+	queue_input.process_va_end = adev->vm_manager.max_pfn - 1;
+
+	/* set process quantum to 10 ms and gang quantum to 1 ms as default */
+	queue_input.process_quantum = 100000;
+	queue_input.gang_quantum = 10000;
+	queue_input.paging = false;
+
+	queue_input.process_context_addr = ctx->gpu_addr;
+	queue_input.gang_context_addr = ctx->gpu_addr + AMDGPU_USERQ_PROC_CTX_SZ;
+	queue_input.inprocess_gang_priority = AMDGPU_MES_PRIORITY_LEVEL_NORMAL;
+	queue_input.gang_global_priority_level = convert_to_mes_priority(queue->priority);
+
+	queue_input.process_id = queue->vm->pasid;
+	queue_input.queue_type = queue->queue_type;
+	queue_input.mqd_addr = queue->mqd.gpu_addr;
+	queue_input.wptr_addr = userq_props->wptr_gpu_addr;
+	queue_input.queue_size = userq_props->queue_size >> 2;
+	queue_input.doorbell_offset = userq_props->doorbell_index;
+	queue_input.page_table_base_addr = amdgpu_gmc_pd_addr(queue->vm->root.bo);
+	queue_input.wptr_mc_addr = queue->wptr_obj.gpu_addr;
+
+	amdgpu_mes_lock(&adev->mes);
+	r = adev->mes.funcs->add_hw_queue(&adev->mes, &queue_input);
+	amdgpu_mes_unlock(&adev->mes);
+	if (r) {
+		DRM_ERROR("Failed to map queue in HW, err (%d)\n", r);
+		return r;
+	}
+
+	DRM_DEBUG_DRIVER("Queue (doorbell:%d) mapped successfully\n", userq_props->doorbell_index);
+	return 0;
+}
+
+static int mes_userq_unmap(struct amdgpu_userq_mgr *uq_mgr,
+			   struct amdgpu_usermode_queue *queue)
+{
+	struct amdgpu_device *adev = uq_mgr->adev;
+	struct mes_remove_queue_input queue_input;
+	struct amdgpu_userq_obj *ctx = &queue->fw_obj;
+	int r;
+
+	memset(&queue_input, 0x0, sizeof(struct mes_remove_queue_input));
+	queue_input.doorbell_offset = queue->doorbell_index;
+	queue_input.gang_context_addr = ctx->gpu_addr + AMDGPU_USERQ_PROC_CTX_SZ;
+
+	amdgpu_mes_lock(&adev->mes);
+	r = adev->mes.funcs->remove_hw_queue(&adev->mes, &queue_input);
+	amdgpu_mes_unlock(&adev->mes);
+	if (r)
+		DRM_ERROR("Failed to unmap queue in HW, err (%d)\n", r);
+	return r;
+}
+
+static int mes_userq_create_ctx_space(struct amdgpu_userq_mgr *uq_mgr,
+				      struct amdgpu_usermode_queue *queue,
+				      struct drm_amdgpu_userq_in *mqd_user)
+{
+	struct amdgpu_userq_obj *ctx = &queue->fw_obj;
+	int r, size;
+
+	/*
+	 * The FW expects at least one page space allocated for
+	 * process ctx and gang ctx each. Create an object
+	 * for the same.
+	 */
+	size = AMDGPU_USERQ_PROC_CTX_SZ + AMDGPU_USERQ_GANG_CTX_SZ;
+	r = amdgpu_userq_create_object(uq_mgr, ctx, size);
+	if (r) {
+		DRM_ERROR("Failed to allocate ctx space bo for userqueue, err:%d\n", r);
+		return r;
+	}
+
+	return 0;
+}
+
+static int mes_userq_detect_and_reset(struct amdgpu_device *adev,
+				      int queue_type)
+{
+	int db_array_size = amdgpu_mes_get_hung_queue_db_array_size(adev);
+	struct mes_detect_and_reset_queue_input input;
+	struct amdgpu_usermode_queue *queue;
+	unsigned int hung_db_num = 0;
+	unsigned long queue_id;
+	u32 db_array[8];
+	bool found_hung_queue = false;
+	int r, i;
+
+	if (db_array_size > 8) {
+		dev_err(adev->dev, "DB array size (%d vs 8) too small\n",
+			db_array_size);
+		return -EINVAL;
+	}
+
+	memset(&input, 0x0, sizeof(struct mes_detect_and_reset_queue_input));
+
+	input.queue_type = queue_type;
+
+	amdgpu_mes_lock(&adev->mes);
+	r = amdgpu_mes_detect_and_reset_hung_queues(adev, queue_type, false,
+						    &hung_db_num, db_array);
+	amdgpu_mes_unlock(&adev->mes);
+	if (r) {
+		dev_err(adev->dev, "Failed to detect and reset queues, err (%d)\n", r);
+	} else if (hung_db_num) {
+		xa_for_each(&adev->userq_doorbell_xa, queue_id, queue) {
+			if (queue->queue_type == queue_type) {
+				for (i = 0; i < hung_db_num; i++) {
+					if (queue->doorbell_index == db_array[i]) {
+						queue->state = AMDGPU_USERQ_STATE_HUNG;
+						found_hung_queue = true;
+						atomic_inc(&adev->gpu_reset_counter);
+						amdgpu_userq_fence_driver_force_completion(queue);
+						drm_dev_wedged_event(adev_to_drm(adev), DRM_WEDGE_RECOVERY_NONE, NULL);
+					}
+				}
+			}
+		}
+	}
+
+	if (found_hung_queue) {
+		/* Resume scheduling after hang recovery */
+		r = amdgpu_mes_resume(adev);
+	}
+
+	return r;
+}
+
+static int mes_userq_mqd_create(struct amdgpu_userq_mgr *uq_mgr,
+				struct drm_amdgpu_userq_in *args_in,
+				struct amdgpu_usermode_queue *queue)
+{
+	struct amdgpu_device *adev = uq_mgr->adev;
+	struct amdgpu_mqd *mqd_hw_default = &adev->mqds[queue->queue_type];
+	struct drm_amdgpu_userq_in *mqd_user = args_in;
+	struct amdgpu_mqd_prop *userq_props;
+	int r;
+
+	/* Structure to initialize MQD for userqueue using generic MQD init function */
+	userq_props = kzalloc(sizeof(struct amdgpu_mqd_prop), GFP_KERNEL);
+	if (!userq_props) {
+		DRM_ERROR("Failed to allocate memory for userq_props\n");
+		return -ENOMEM;
+	}
+
+	r = amdgpu_userq_create_object(uq_mgr, &queue->mqd, mqd_hw_default->mqd_size);
+	if (r) {
+		DRM_ERROR("Failed to create MQD object for userqueue\n");
+		goto free_props;
+	}
+
+	/* Initialize the MQD BO with user given values */
+	userq_props->wptr_gpu_addr = mqd_user->wptr_va;
+	userq_props->rptr_gpu_addr = mqd_user->rptr_va;
+	userq_props->queue_size = mqd_user->queue_size;
+	userq_props->hqd_base_gpu_addr = mqd_user->queue_va;
+	userq_props->mqd_gpu_addr = queue->mqd.gpu_addr;
+	userq_props->use_doorbell = true;
+	userq_props->doorbell_index = queue->doorbell_index;
+	userq_props->fence_address = queue->fence_drv->gpu_addr;
+
+	if (queue->queue_type == AMDGPU_HW_IP_COMPUTE) {
+		struct drm_amdgpu_userq_mqd_compute_gfx11 *compute_mqd;
+
+		if (mqd_user->mqd_size != sizeof(*compute_mqd)) {
+			DRM_ERROR("Invalid compute IP MQD size\n");
+			r = -EINVAL;
+			goto free_mqd;
+		}
+
+		compute_mqd = memdup_user(u64_to_user_ptr(mqd_user->mqd), mqd_user->mqd_size);
+		if (IS_ERR(compute_mqd)) {
+			DRM_ERROR("Failed to read user MQD\n");
+			r = -ENOMEM;
+			goto free_mqd;
+		}
+
+		r = amdgpu_userq_input_va_validate(queue, compute_mqd->eop_va,
+						   2048);
+		if (r)
+			goto free_mqd;
+
+		userq_props->eop_gpu_addr = compute_mqd->eop_va;
+		userq_props->hqd_pipe_priority = AMDGPU_GFX_PIPE_PRIO_NORMAL;
+		userq_props->hqd_queue_priority = AMDGPU_GFX_QUEUE_PRIORITY_MINIMUM;
+		userq_props->hqd_active = false;
+		userq_props->tmz_queue =
+			mqd_user->flags & AMDGPU_USERQ_CREATE_FLAGS_QUEUE_SECURE;
+		kfree(compute_mqd);
+	} else if (queue->queue_type == AMDGPU_HW_IP_GFX) {
+		struct drm_amdgpu_userq_mqd_gfx11 *mqd_gfx_v11;
+		struct amdgpu_gfx_shadow_info shadow_info;
+
+		if (adev->gfx.funcs->get_gfx_shadow_info) {
+			adev->gfx.funcs->get_gfx_shadow_info(adev, &shadow_info, true);
+		} else {
+			r = -EINVAL;
+			goto free_mqd;
+		}
+
+		if (mqd_user->mqd_size != sizeof(*mqd_gfx_v11) || !mqd_user->mqd) {
+			DRM_ERROR("Invalid GFX MQD\n");
+			r = -EINVAL;
+			goto free_mqd;
+		}
+
+		mqd_gfx_v11 = memdup_user(u64_to_user_ptr(mqd_user->mqd), mqd_user->mqd_size);
+		if (IS_ERR(mqd_gfx_v11)) {
+			DRM_ERROR("Failed to read user MQD\n");
+			r = -ENOMEM;
+			goto free_mqd;
+		}
+
+		userq_props->shadow_addr = mqd_gfx_v11->shadow_va;
+		userq_props->csa_addr = mqd_gfx_v11->csa_va;
+		userq_props->tmz_queue =
+			mqd_user->flags & AMDGPU_USERQ_CREATE_FLAGS_QUEUE_SECURE;
+
+		r = amdgpu_userq_input_va_validate(queue, mqd_gfx_v11->shadow_va,
+						   shadow_info.shadow_size);
+		if (r)
+			goto free_mqd;
+		r = amdgpu_userq_input_va_validate(queue, mqd_gfx_v11->csa_va,
+						   shadow_info.csa_size);
+		if (r)
+			goto free_mqd;
+
+		kfree(mqd_gfx_v11);
+	} else if (queue->queue_type == AMDGPU_HW_IP_DMA) {
+		struct drm_amdgpu_userq_mqd_sdma_gfx11 *mqd_sdma_v11;
+
+		if (mqd_user->mqd_size != sizeof(*mqd_sdma_v11) || !mqd_user->mqd) {
+			DRM_ERROR("Invalid SDMA MQD\n");
+			r = -EINVAL;
+			goto free_mqd;
+		}
+
+		mqd_sdma_v11 = memdup_user(u64_to_user_ptr(mqd_user->mqd), mqd_user->mqd_size);
+		if (IS_ERR(mqd_sdma_v11)) {
+			DRM_ERROR("Failed to read sdma user MQD\n");
+			r = -ENOMEM;
+			goto free_mqd;
+		}
+		r = amdgpu_userq_input_va_validate(queue, mqd_sdma_v11->csa_va,
+						   32);
+		if (r)
+			goto free_mqd;
+
+		userq_props->csa_addr = mqd_sdma_v11->csa_va;
+		kfree(mqd_sdma_v11);
+	}
+
+	queue->userq_prop = userq_props;
+
+	r = mqd_hw_default->init_mqd(adev, (void *)queue->mqd.cpu_ptr, userq_props);
+	if (r) {
+		DRM_ERROR("Failed to initialize MQD for userqueue\n");
+		goto free_mqd;
+	}
+
+	/* Create BO for FW operations */
+	r = mes_userq_create_ctx_space(uq_mgr, queue, mqd_user);
+	if (r) {
+		DRM_ERROR("Failed to allocate BO for userqueue (%d)", r);
+		goto free_mqd;
+	}
+
+	/* FW expects WPTR BOs to be mapped into GART */
+	r = mes_userq_create_wptr_mapping(uq_mgr, queue, userq_props->wptr_gpu_addr);
+	if (r) {
+		DRM_ERROR("Failed to create WPTR mapping\n");
+		goto free_ctx;
+	}
+
+	return 0;
+
+free_ctx:
+	amdgpu_userq_destroy_object(uq_mgr, &queue->fw_obj);
+
+free_mqd:
+	amdgpu_userq_destroy_object(uq_mgr, &queue->mqd);
+
+free_props:
+	kfree(userq_props);
+
+	return r;
+}
+
+static void
+mes_userq_mqd_destroy(struct amdgpu_userq_mgr *uq_mgr,
+		      struct amdgpu_usermode_queue *queue)
+{
+	amdgpu_userq_destroy_object(uq_mgr, &queue->fw_obj);
+	kfree(queue->userq_prop);
+	amdgpu_userq_destroy_object(uq_mgr, &queue->mqd);
+}
+
+static int mes_userq_preempt(struct amdgpu_userq_mgr *uq_mgr,
+				struct amdgpu_usermode_queue *queue)
+{
+	struct amdgpu_device *adev = uq_mgr->adev;
+	struct mes_suspend_gang_input queue_input;
+	struct amdgpu_userq_obj *ctx = &queue->fw_obj;
+	signed long timeout = 2100000; /* 2100 ms */
+	u64 fence_gpu_addr;
+	u32 fence_offset;
+	u64 *fence_ptr;
+	int i, r;
+
+	if (queue->state != AMDGPU_USERQ_STATE_MAPPED)
+		return 0;
+	r = amdgpu_device_wb_get(adev, &fence_offset);
+	if (r)
+		return r;
+
+	fence_gpu_addr = adev->wb.gpu_addr + (fence_offset * 4);
+	fence_ptr = (u64 *)&adev->wb.wb[fence_offset];
+	*fence_ptr = 0;
+
+	memset(&queue_input, 0x0, sizeof(struct mes_suspend_gang_input));
+	queue_input.gang_context_addr = ctx->gpu_addr + AMDGPU_USERQ_PROC_CTX_SZ;
+	queue_input.suspend_fence_addr = fence_gpu_addr;
+	queue_input.suspend_fence_value = 1;
+	amdgpu_mes_lock(&adev->mes);
+	r = adev->mes.funcs->suspend_gang(&adev->mes, &queue_input);
+	amdgpu_mes_unlock(&adev->mes);
+	if (r) {
+		DRM_ERROR("Failed to suspend gang: %d\n", r);
+		goto out;
+	}
+
+	for (i = 0; i < timeout; i++) {
+		if (*fence_ptr == 1)
+			goto out;
+		udelay(1);
+	}
+	r = -ETIMEDOUT;
+
+out:
+	amdgpu_device_wb_free(adev, fence_offset);
+	return r;
+}
+
+static int mes_userq_restore(struct amdgpu_userq_mgr *uq_mgr,
+				struct amdgpu_usermode_queue *queue)
+{
+	struct amdgpu_device *adev = uq_mgr->adev;
+	struct mes_resume_gang_input queue_input;
+	struct amdgpu_userq_obj *ctx = &queue->fw_obj;
+	int r;
+
+	if (queue->state == AMDGPU_USERQ_STATE_HUNG)
+		return -EINVAL;
+	if (queue->state != AMDGPU_USERQ_STATE_PREEMPTED)
+		return 0;
+
+	memset(&queue_input, 0x0, sizeof(struct mes_resume_gang_input));
+	queue_input.gang_context_addr = ctx->gpu_addr + AMDGPU_USERQ_PROC_CTX_SZ;
+
+	amdgpu_mes_lock(&adev->mes);
+	r = adev->mes.funcs->resume_gang(&adev->mes, &queue_input);
+	amdgpu_mes_unlock(&adev->mes);
+	if (r)
+		dev_err(adev->dev, "Failed to resume queue, err (%d)\n", r);
+	return r;
+}
+
+const struct amdgpu_userq_funcs userq_mes_funcs = {
+	.mqd_create = mes_userq_mqd_create,
+	.mqd_destroy = mes_userq_mqd_destroy,
+	.unmap = mes_userq_unmap,
+	.map = mes_userq_map,
+	.detect_and_reset = mes_userq_detect_and_reset,
+	.preempt = mes_userq_preempt,
+	.restore = mes_userq_restore,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/mes_userqueue.h b/drivers/gpu/drm/amd/amdgpu/mes_userqueue.h
new file mode 100644
index 000000000000..090ae8897770
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/mes_userqueue.h
@@ -0,0 +1,30 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2024 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef MES_USERQ_H
+#define MES_USERQ_H
+#include "amdgpu_userq.h"
+
+extern const struct amdgpu_userq_funcs userq_mes_funcs;
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/mes_v10_1.c b/drivers/gpu/drm/amd/amdgpu/mes_v10_1.c
deleted file mode 100644
index a7ec4ac89da5..000000000000
--- a/drivers/gpu/drm/amd/amdgpu/mes_v10_1.c
+++ /dev/null
@@ -1,1008 +0,0 @@
-/*
- * Copyright 2019 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- */
-
-#include <linux/firmware.h>
-#include <linux/module.h>
-#include "amdgpu.h"
-#include "soc15_common.h"
-#include "nv.h"
-#include "gc/gc_10_1_0_offset.h"
-#include "gc/gc_10_1_0_sh_mask.h"
-#include "v10_structs.h"
-#include "mes_api_def.h"
-
-#define mmCP_MES_IC_OP_CNTL_Sienna_Cichlid               0x2820
-#define mmCP_MES_IC_OP_CNTL_Sienna_Cichlid_BASE_IDX      1
-
-MODULE_FIRMWARE("amdgpu/navi10_mes.bin");
-MODULE_FIRMWARE("amdgpu/sienna_cichlid_mes.bin");
-
-static int mes_v10_1_hw_fini(void *handle);
-
-#define MES_EOP_SIZE   2048
-
-static void mes_v10_1_ring_set_wptr(struct amdgpu_ring *ring)
-{
-	struct amdgpu_device *adev = ring->adev;
-
-	if (ring->use_doorbell) {
-		atomic64_set((atomic64_t *)&adev->wb.wb[ring->wptr_offs],
-			     ring->wptr);
-		WDOORBELL64(ring->doorbell_index, ring->wptr);
-	} else {
-		BUG();
-	}
-}
-
-static u64 mes_v10_1_ring_get_rptr(struct amdgpu_ring *ring)
-{
-	return ring->adev->wb.wb[ring->rptr_offs];
-}
-
-static u64 mes_v10_1_ring_get_wptr(struct amdgpu_ring *ring)
-{
-	u64 wptr;
-
-	if (ring->use_doorbell)
-		wptr = atomic64_read((atomic64_t *)
-				     &ring->adev->wb.wb[ring->wptr_offs]);
-	else
-		BUG();
-	return wptr;
-}
-
-static const struct amdgpu_ring_funcs mes_v10_1_ring_funcs = {
-	.type = AMDGPU_RING_TYPE_MES,
-	.align_mask = 1,
-	.nop = 0,
-	.support_64bit_ptrs = true,
-	.get_rptr = mes_v10_1_ring_get_rptr,
-	.get_wptr = mes_v10_1_ring_get_wptr,
-	.set_wptr = mes_v10_1_ring_set_wptr,
-	.insert_nop = amdgpu_ring_insert_nop,
-};
-
-static int mes_v10_1_submit_pkt_and_poll_completion(struct amdgpu_mes *mes,
-						    void *pkt, int size)
-{
-	int ndw = size / 4;
-	signed long r;
-	union MESAPI__ADD_QUEUE *x_pkt = pkt;
-	struct amdgpu_device *adev = mes->adev;
-	struct amdgpu_ring *ring = &mes->ring;
-
-	BUG_ON(size % 4 != 0);
-
-	if (amdgpu_ring_alloc(ring, ndw))
-		return -ENOMEM;
-
-	amdgpu_ring_write_multiple(ring, pkt, ndw);
-	amdgpu_ring_commit(ring);
-
-	DRM_DEBUG("MES msg=%d was emitted\n", x_pkt->header.opcode);
-
-	r = amdgpu_fence_wait_polling(ring, ring->fence_drv.sync_seq,
-				      adev->usec_timeout);
-	if (r < 1) {
-		DRM_ERROR("MES failed to response msg=%d\n",
-			  x_pkt->header.opcode);
-		return -ETIMEDOUT;
-	}
-
-	return 0;
-}
-
-static int convert_to_mes_queue_type(int queue_type)
-{
-	if (queue_type == AMDGPU_RING_TYPE_GFX)
-		return MES_QUEUE_TYPE_GFX;
-	else if (queue_type == AMDGPU_RING_TYPE_COMPUTE)
-		return MES_QUEUE_TYPE_COMPUTE;
-	else if (queue_type == AMDGPU_RING_TYPE_SDMA)
-		return MES_QUEUE_TYPE_SDMA;
-	else
-		BUG();
-	return -1;
-}
-
-static int mes_v10_1_add_hw_queue(struct amdgpu_mes *mes,
-				  struct mes_add_queue_input *input)
-{
-	struct amdgpu_device *adev = mes->adev;
-	union MESAPI__ADD_QUEUE mes_add_queue_pkt;
-
-	memset(&mes_add_queue_pkt, 0, sizeof(mes_add_queue_pkt));
-
-	mes_add_queue_pkt.header.type = MES_API_TYPE_SCHEDULER;
-	mes_add_queue_pkt.header.opcode = MES_SCH_API_ADD_QUEUE;
-	mes_add_queue_pkt.header.dwsize = API_FRAME_SIZE_IN_DWORDS;
-
-	mes_add_queue_pkt.process_id = input->process_id;
-	mes_add_queue_pkt.page_table_base_addr =
-		input->page_table_base_addr - adev->gmc.vram_start;
-	mes_add_queue_pkt.process_va_start = input->process_va_start;
-	mes_add_queue_pkt.process_va_end = input->process_va_end;
-	mes_add_queue_pkt.process_quantum = input->process_quantum;
-	mes_add_queue_pkt.process_context_addr = input->process_context_addr;
-	mes_add_queue_pkt.gang_quantum = input->gang_quantum;
-	mes_add_queue_pkt.gang_context_addr = input->gang_context_addr;
-	mes_add_queue_pkt.inprocess_gang_priority =
-		input->inprocess_gang_priority;
-	mes_add_queue_pkt.gang_global_priority_level =
-		input->gang_global_priority_level;
-	mes_add_queue_pkt.doorbell_offset = input->doorbell_offset;
-	mes_add_queue_pkt.mqd_addr = input->mqd_addr;
-	mes_add_queue_pkt.wptr_addr = input->wptr_addr;
-	mes_add_queue_pkt.queue_type =
-		convert_to_mes_queue_type(input->queue_type);
-	mes_add_queue_pkt.paging = input->paging;
-
-	mes_add_queue_pkt.api_status.api_completion_fence_addr =
-		mes->ring.fence_drv.gpu_addr;
-	mes_add_queue_pkt.api_status.api_completion_fence_value =
-		++mes->ring.fence_drv.sync_seq;
-
-	return mes_v10_1_submit_pkt_and_poll_completion(mes,
-			&mes_add_queue_pkt, sizeof(mes_add_queue_pkt));
-}
-
-static int mes_v10_1_remove_hw_queue(struct amdgpu_mes *mes,
-				     struct mes_remove_queue_input *input)
-{
-	union MESAPI__REMOVE_QUEUE mes_remove_queue_pkt;
-
-	memset(&mes_remove_queue_pkt, 0, sizeof(mes_remove_queue_pkt));
-
-	mes_remove_queue_pkt.header.type = MES_API_TYPE_SCHEDULER;
-	mes_remove_queue_pkt.header.opcode = MES_SCH_API_REMOVE_QUEUE;
-	mes_remove_queue_pkt.header.dwsize = API_FRAME_SIZE_IN_DWORDS;
-
-	mes_remove_queue_pkt.doorbell_offset = input->doorbell_offset;
-	mes_remove_queue_pkt.gang_context_addr = input->gang_context_addr;
-
-	mes_remove_queue_pkt.api_status.api_completion_fence_addr =
-		mes->ring.fence_drv.gpu_addr;
-	mes_remove_queue_pkt.api_status.api_completion_fence_value =
-		++mes->ring.fence_drv.sync_seq;
-
-	return mes_v10_1_submit_pkt_and_poll_completion(mes,
-			&mes_remove_queue_pkt, sizeof(mes_remove_queue_pkt));
-}
-
-static int mes_v10_1_suspend_gang(struct amdgpu_mes *mes,
-				  struct mes_suspend_gang_input *input)
-{
-	return 0;
-}
-
-static int mes_v10_1_resume_gang(struct amdgpu_mes *mes,
-				 struct mes_resume_gang_input *input)
-{
-	return 0;
-}
-
-static int mes_v10_1_query_sched_status(struct amdgpu_mes *mes)
-{
-	union MESAPI__QUERY_MES_STATUS mes_status_pkt;
-
-	memset(&mes_status_pkt, 0, sizeof(mes_status_pkt));
-
-	mes_status_pkt.header.type = MES_API_TYPE_SCHEDULER;
-	mes_status_pkt.header.opcode = MES_SCH_API_QUERY_SCHEDULER_STATUS;
-	mes_status_pkt.header.dwsize = API_FRAME_SIZE_IN_DWORDS;
-
-	mes_status_pkt.api_status.api_completion_fence_addr =
-		mes->ring.fence_drv.gpu_addr;
-	mes_status_pkt.api_status.api_completion_fence_value =
-		++mes->ring.fence_drv.sync_seq;
-
-	return mes_v10_1_submit_pkt_and_poll_completion(mes,
-			&mes_status_pkt, sizeof(mes_status_pkt));
-}
-
-static int mes_v10_1_set_hw_resources(struct amdgpu_mes *mes)
-{
-	int i;
-	struct amdgpu_device *adev = mes->adev;
-	union MESAPI_SET_HW_RESOURCES mes_set_hw_res_pkt;
-
-	memset(&mes_set_hw_res_pkt, 0, sizeof(mes_set_hw_res_pkt));
-
-	mes_set_hw_res_pkt.header.type = MES_API_TYPE_SCHEDULER;
-	mes_set_hw_res_pkt.header.opcode = MES_SCH_API_SET_HW_RSRC;
-	mes_set_hw_res_pkt.header.dwsize = API_FRAME_SIZE_IN_DWORDS;
-
-	mes_set_hw_res_pkt.vmid_mask_mmhub = mes->vmid_mask_mmhub;
-	mes_set_hw_res_pkt.vmid_mask_gfxhub = mes->vmid_mask_gfxhub;
-	mes_set_hw_res_pkt.gds_size = adev->gds.gds_size;
-	mes_set_hw_res_pkt.paging_vmid = 0;
-	mes_set_hw_res_pkt.g_sch_ctx_gpu_mc_ptr = mes->sch_ctx_gpu_addr;
-	mes_set_hw_res_pkt.query_status_fence_gpu_mc_ptr =
-		mes->query_status_fence_gpu_addr;
-
-	for (i = 0; i < MAX_COMPUTE_PIPES; i++)
-		mes_set_hw_res_pkt.compute_hqd_mask[i] =
-			mes->compute_hqd_mask[i];
-
-	for (i = 0; i < MAX_GFX_PIPES; i++)
-		mes_set_hw_res_pkt.gfx_hqd_mask[i] = mes->gfx_hqd_mask[i];
-
-	for (i = 0; i < MAX_SDMA_PIPES; i++)
-		mes_set_hw_res_pkt.sdma_hqd_mask[i] = mes->sdma_hqd_mask[i];
-
-	for (i = 0; i < AMD_PRIORITY_NUM_LEVELS; i++)
-		mes_set_hw_res_pkt.agreegated_doorbells[i] =
-			mes->agreegated_doorbells[i];
-
-	mes_set_hw_res_pkt.api_status.api_completion_fence_addr =
-		mes->ring.fence_drv.gpu_addr;
-	mes_set_hw_res_pkt.api_status.api_completion_fence_value =
-		++mes->ring.fence_drv.sync_seq;
-
-	return mes_v10_1_submit_pkt_and_poll_completion(mes,
-			&mes_set_hw_res_pkt, sizeof(mes_set_hw_res_pkt));
-}
-
-static const struct amdgpu_mes_funcs mes_v10_1_funcs = {
-	.add_hw_queue = mes_v10_1_add_hw_queue,
-	.remove_hw_queue = mes_v10_1_remove_hw_queue,
-	.suspend_gang = mes_v10_1_suspend_gang,
-	.resume_gang = mes_v10_1_resume_gang,
-};
-
-static int mes_v10_1_init_microcode(struct amdgpu_device *adev)
-{
-	const char *chip_name;
-	char fw_name[30];
-	int err;
-	const struct mes_firmware_header_v1_0 *mes_hdr;
-	struct amdgpu_firmware_info *info;
-
-	switch (adev->asic_type) {
-	case CHIP_NAVI10:
-		chip_name = "navi10";
-		break;
-	case CHIP_SIENNA_CICHLID:
-		chip_name = "sienna_cichlid";
-		break;
-	default:
-		BUG();
-	}
-
-	snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_mes.bin", chip_name);
-	err = request_firmware(&adev->mes.fw, fw_name, adev->dev);
-	if (err)
-		return err;
-
-	err = amdgpu_ucode_validate(adev->mes.fw);
-	if (err) {
-		release_firmware(adev->mes.fw);
-		adev->mes.fw = NULL;
-		return err;
-	}
-
-	mes_hdr = (const struct mes_firmware_header_v1_0 *)adev->mes.fw->data;
-	adev->mes.ucode_fw_version = le32_to_cpu(mes_hdr->mes_ucode_version);
-	adev->mes.ucode_fw_version =
-		le32_to_cpu(mes_hdr->mes_ucode_data_version);
-	adev->mes.uc_start_addr =
-		le32_to_cpu(mes_hdr->mes_uc_start_addr_lo) |
-		((uint64_t)(le32_to_cpu(mes_hdr->mes_uc_start_addr_hi)) << 32);
-	adev->mes.data_start_addr =
-		le32_to_cpu(mes_hdr->mes_data_start_addr_lo) |
-		((uint64_t)(le32_to_cpu(mes_hdr->mes_data_start_addr_hi)) << 32);
-
-	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
-		info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_MES];
-		info->ucode_id = AMDGPU_UCODE_ID_CP_MES;
-		info->fw = adev->mes.fw;
-		adev->firmware.fw_size +=
-			ALIGN(le32_to_cpu(mes_hdr->mes_ucode_size_bytes),
-			      PAGE_SIZE);
-
-		info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_MES_DATA];
-		info->ucode_id = AMDGPU_UCODE_ID_CP_MES_DATA;
-		info->fw = adev->mes.fw;
-		adev->firmware.fw_size +=
-			ALIGN(le32_to_cpu(mes_hdr->mes_ucode_data_size_bytes),
-			      PAGE_SIZE);
-	}
-
-	return 0;
-}
-
-static void mes_v10_1_free_microcode(struct amdgpu_device *adev)
-{
-	release_firmware(adev->mes.fw);
-	adev->mes.fw = NULL;
-}
-
-static int mes_v10_1_allocate_ucode_buffer(struct amdgpu_device *adev)
-{
-	int r;
-	const struct mes_firmware_header_v1_0 *mes_hdr;
-	const __le32 *fw_data;
-	unsigned fw_size;
-
-	mes_hdr = (const struct mes_firmware_header_v1_0 *)
-		adev->mes.fw->data;
-
-	fw_data = (const __le32 *)(adev->mes.fw->data +
-		   le32_to_cpu(mes_hdr->mes_ucode_offset_bytes));
-	fw_size = le32_to_cpu(mes_hdr->mes_ucode_size_bytes);
-
-	r = amdgpu_bo_create_reserved(adev, fw_size,
-				      PAGE_SIZE, AMDGPU_GEM_DOMAIN_GTT,
-				      &adev->mes.ucode_fw_obj,
-				      &adev->mes.ucode_fw_gpu_addr,
-				      (void **)&adev->mes.ucode_fw_ptr);
-	if (r) {
-		dev_err(adev->dev, "(%d) failed to create mes fw bo\n", r);
-		return r;
-	}
-
-	memcpy(adev->mes.ucode_fw_ptr, fw_data, fw_size);
-
-	amdgpu_bo_kunmap(adev->mes.ucode_fw_obj);
-	amdgpu_bo_unreserve(adev->mes.ucode_fw_obj);
-
-	return 0;
-}
-
-static int mes_v10_1_allocate_ucode_data_buffer(struct amdgpu_device *adev)
-{
-	int r;
-	const struct mes_firmware_header_v1_0 *mes_hdr;
-	const __le32 *fw_data;
-	unsigned fw_size;
-
-	mes_hdr = (const struct mes_firmware_header_v1_0 *)
-		adev->mes.fw->data;
-
-	fw_data = (const __le32 *)(adev->mes.fw->data +
-		   le32_to_cpu(mes_hdr->mes_ucode_data_offset_bytes));
-	fw_size = le32_to_cpu(mes_hdr->mes_ucode_data_size_bytes);
-
-	r = amdgpu_bo_create_reserved(adev, fw_size,
-				      64 * 1024, AMDGPU_GEM_DOMAIN_GTT,
-				      &adev->mes.data_fw_obj,
-				      &adev->mes.data_fw_gpu_addr,
-				      (void **)&adev->mes.data_fw_ptr);
-	if (r) {
-		dev_err(adev->dev, "(%d) failed to create mes data fw bo\n", r);
-		return r;
-	}
-
-	memcpy(adev->mes.data_fw_ptr, fw_data, fw_size);
-
-	amdgpu_bo_kunmap(adev->mes.data_fw_obj);
-	amdgpu_bo_unreserve(adev->mes.data_fw_obj);
-
-	return 0;
-}
-
-static void mes_v10_1_free_ucode_buffers(struct amdgpu_device *adev)
-{
-	amdgpu_bo_free_kernel(&adev->mes.data_fw_obj,
-			      &adev->mes.data_fw_gpu_addr,
-			      (void **)&adev->mes.data_fw_ptr);
-
-	amdgpu_bo_free_kernel(&adev->mes.ucode_fw_obj,
-			      &adev->mes.ucode_fw_gpu_addr,
-			      (void **)&adev->mes.ucode_fw_ptr);
-}
-
-static void mes_v10_1_enable(struct amdgpu_device *adev, bool enable)
-{
-	uint32_t data = 0;
-
-	if (enable) {
-		data = RREG32_SOC15(GC, 0, mmCP_MES_CNTL);
-		data = REG_SET_FIELD(data, CP_MES_CNTL, MES_PIPE0_RESET, 1);
-		WREG32_SOC15(GC, 0, mmCP_MES_CNTL, data);
-
-		/* set ucode start address */
-		WREG32_SOC15(GC, 0, mmCP_MES_PRGRM_CNTR_START,
-			     (uint32_t)(adev->mes.uc_start_addr) >> 2);
-
-		/* clear BYPASS_UNCACHED to avoid hangs after interrupt. */
-		data = RREG32_SOC15(GC, 0, mmCP_MES_DC_OP_CNTL);
-		data = REG_SET_FIELD(data, CP_MES_DC_OP_CNTL,
-				     BYPASS_UNCACHED, 0);
-		WREG32_SOC15(GC, 0, mmCP_MES_DC_OP_CNTL, data);
-
-		/* unhalt MES and activate pipe0 */
-		data = REG_SET_FIELD(0, CP_MES_CNTL, MES_PIPE0_ACTIVE, 1);
-		WREG32_SOC15(GC, 0, mmCP_MES_CNTL, data);
-	} else {
-		data = RREG32_SOC15(GC, 0, mmCP_MES_CNTL);
-		data = REG_SET_FIELD(data, CP_MES_CNTL, MES_PIPE0_ACTIVE, 0);
-		data = REG_SET_FIELD(data, CP_MES_CNTL,
-				     MES_INVALIDATE_ICACHE, 1);
-		data = REG_SET_FIELD(data, CP_MES_CNTL, MES_PIPE0_RESET, 1);
-		data = REG_SET_FIELD(data, CP_MES_CNTL, MES_HALT, 1);
-		WREG32_SOC15(GC, 0, mmCP_MES_CNTL, data);
-	}
-}
-
-/* This function is for backdoor MES firmware */
-static int mes_v10_1_load_microcode(struct amdgpu_device *adev)
-{
-	int r;
-	uint32_t data;
-
-	if (!adev->mes.fw)
-		return -EINVAL;
-
-	r = mes_v10_1_allocate_ucode_buffer(adev);
-	if (r)
-		return r;
-
-	r = mes_v10_1_allocate_ucode_data_buffer(adev);
-	if (r) {
-		mes_v10_1_free_ucode_buffers(adev);
-		return r;
-	}
-
-	mes_v10_1_enable(adev, false);
-
-	WREG32_SOC15(GC, 0, mmCP_MES_IC_BASE_CNTL, 0);
-
-	mutex_lock(&adev->srbm_mutex);
-	/* me=3, pipe=0, queue=0 */
-	nv_grbm_select(adev, 3, 0, 0, 0);
-
-	/* set ucode start address */
-	WREG32_SOC15(GC, 0, mmCP_MES_PRGRM_CNTR_START,
-		     (uint32_t)(adev->mes.uc_start_addr) >> 2);
-
-	/* set ucode fimrware address */
-	WREG32_SOC15(GC, 0, mmCP_MES_IC_BASE_LO,
-		     lower_32_bits(adev->mes.ucode_fw_gpu_addr));
-	WREG32_SOC15(GC, 0, mmCP_MES_IC_BASE_HI,
-		     upper_32_bits(adev->mes.ucode_fw_gpu_addr));
-
-	/* set ucode instruction cache boundary to 2M-1 */
-	WREG32_SOC15(GC, 0, mmCP_MES_MIBOUND_LO, 0x1FFFFF);
-
-	/* set ucode data firmware address */
-	WREG32_SOC15(GC, 0, mmCP_MES_MDBASE_LO,
-		     lower_32_bits(adev->mes.data_fw_gpu_addr));
-	WREG32_SOC15(GC, 0, mmCP_MES_MDBASE_HI,
-		     upper_32_bits(adev->mes.data_fw_gpu_addr));
-
-	/* Set 0x3FFFF (256K-1) to CP_MES_MDBOUND_LO */
-	WREG32_SOC15(GC, 0, mmCP_MES_MDBOUND_LO, 0x3FFFF);
-
-	/* invalidate ICACHE */
-	switch (adev->asic_type) {
-	case CHIP_SIENNA_CICHLID:
-		data = RREG32_SOC15(GC, 0, mmCP_MES_IC_OP_CNTL_Sienna_Cichlid);
-		break;
-	default:
-		data = RREG32_SOC15(GC, 0, mmCP_MES_IC_OP_CNTL);
-		break;
-	}
-	data = REG_SET_FIELD(data, CP_MES_IC_OP_CNTL, PRIME_ICACHE, 0);
-	data = REG_SET_FIELD(data, CP_MES_IC_OP_CNTL, INVALIDATE_CACHE, 1);
-	switch (adev->asic_type) {
-	case CHIP_SIENNA_CICHLID:
-		WREG32_SOC15(GC, 0, mmCP_MES_IC_OP_CNTL_Sienna_Cichlid, data);
-		break;
-	default:
-		WREG32_SOC15(GC, 0, mmCP_MES_IC_OP_CNTL, data);
-		break;
-	}
-
-	/* prime the ICACHE. */
-	switch (adev->asic_type) {
-	case CHIP_SIENNA_CICHLID:
-		data = RREG32_SOC15(GC, 0, mmCP_MES_IC_OP_CNTL_Sienna_Cichlid);
-		break;
-	default:
-		data = RREG32_SOC15(GC, 0, mmCP_MES_IC_OP_CNTL);
-		break;
-	}
-	data = REG_SET_FIELD(data, CP_MES_IC_OP_CNTL, PRIME_ICACHE, 1);
-	switch (adev->asic_type) {
-	case CHIP_SIENNA_CICHLID:
-		WREG32_SOC15(GC, 0, mmCP_MES_IC_OP_CNTL_Sienna_Cichlid, data);
-		break;
-	default:
-		WREG32_SOC15(GC, 0, mmCP_MES_IC_OP_CNTL, data);
-		break;
-	}
-
-	nv_grbm_select(adev, 0, 0, 0, 0);
-	mutex_unlock(&adev->srbm_mutex);
-
-	return 0;
-}
-
-static int mes_v10_1_allocate_eop_buf(struct amdgpu_device *adev)
-{
-	int r;
-	u32 *eop;
-
-	r = amdgpu_bo_create_reserved(adev, MES_EOP_SIZE, PAGE_SIZE,
-				      AMDGPU_GEM_DOMAIN_GTT,
-				      &adev->mes.eop_gpu_obj,
-				      &adev->mes.eop_gpu_addr,
-				      (void **)&eop);
-	if (r) {
-		dev_warn(adev->dev, "(%d) create EOP bo failed\n", r);
-		return r;
-	}
-
-	memset(eop, 0, adev->mes.eop_gpu_obj->tbo.base.size);
-
-	amdgpu_bo_kunmap(adev->mes.eop_gpu_obj);
-	amdgpu_bo_unreserve(adev->mes.eop_gpu_obj);
-
-	return 0;
-}
-
-static int mes_v10_1_allocate_mem_slots(struct amdgpu_device *adev)
-{
-	int r;
-
-	r = amdgpu_device_wb_get(adev, &adev->mes.sch_ctx_offs);
-	if (r) {
-		dev_err(adev->dev,
-			"(%d) mes sch_ctx_offs wb alloc failed\n", r);
-		return r;
-	}
-	adev->mes.sch_ctx_gpu_addr =
-		adev->wb.gpu_addr + (adev->mes.sch_ctx_offs * 4);
-	adev->mes.sch_ctx_ptr =
-		(uint64_t *)&adev->wb.wb[adev->mes.sch_ctx_offs];
-
-	r = amdgpu_device_wb_get(adev, &adev->mes.query_status_fence_offs);
-	if (r) {
-		dev_err(adev->dev,
-			"(%d) query_status_fence_offs wb alloc failed\n", r);
-		return r;
-	}
-	adev->mes.query_status_fence_gpu_addr =
-		adev->wb.gpu_addr + (adev->mes.query_status_fence_offs * 4);
-	adev->mes.query_status_fence_ptr =
-		(uint64_t *)&adev->wb.wb[adev->mes.query_status_fence_offs];
-
-	return 0;
-}
-
-static int mes_v10_1_mqd_init(struct amdgpu_ring *ring)
-{
-	struct amdgpu_device *adev = ring->adev;
-	struct v10_compute_mqd *mqd = ring->mqd_ptr;
-	uint64_t hqd_gpu_addr, wb_gpu_addr, eop_base_addr;
-	uint32_t tmp;
-
-	mqd->header = 0xC0310800;
-	mqd->compute_pipelinestat_enable = 0x00000001;
-	mqd->compute_static_thread_mgmt_se0 = 0xffffffff;
-	mqd->compute_static_thread_mgmt_se1 = 0xffffffff;
-	mqd->compute_static_thread_mgmt_se2 = 0xffffffff;
-	mqd->compute_static_thread_mgmt_se3 = 0xffffffff;
-	mqd->compute_misc_reserved = 0x00000003;
-
-	eop_base_addr = ring->eop_gpu_addr >> 8;
-	mqd->cp_hqd_eop_base_addr_lo = eop_base_addr;
-	mqd->cp_hqd_eop_base_addr_hi = upper_32_bits(eop_base_addr);
-
-	/* set the EOP size, register value is 2^(EOP_SIZE+1) dwords */
-	tmp = RREG32_SOC15(GC, 0, mmCP_HQD_EOP_CONTROL);
-	tmp = REG_SET_FIELD(tmp, CP_HQD_EOP_CONTROL, EOP_SIZE,
-			(order_base_2(MES_EOP_SIZE / 4) - 1));
-
-	mqd->cp_hqd_eop_control = tmp;
-
-	/* enable doorbell? */
-	tmp = RREG32_SOC15(GC, 0, mmCP_HQD_PQ_DOORBELL_CONTROL);
-
-	if (ring->use_doorbell) {
-		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
-				    DOORBELL_OFFSET, ring->doorbell_index);
-		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
-				    DOORBELL_EN, 1);
-		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
-				    DOORBELL_SOURCE, 0);
-		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
-				    DOORBELL_HIT, 0);
-	}
-	else
-		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
-				    DOORBELL_EN, 0);
-
-	mqd->cp_hqd_pq_doorbell_control = tmp;
-
-	/* disable the queue if it's active */
-	ring->wptr = 0;
-	mqd->cp_hqd_dequeue_request = 0;
-	mqd->cp_hqd_pq_rptr = 0;
-	mqd->cp_hqd_pq_wptr_lo = 0;
-	mqd->cp_hqd_pq_wptr_hi = 0;
-
-	/* set the pointer to the MQD */
-	mqd->cp_mqd_base_addr_lo = ring->mqd_gpu_addr & 0xfffffffc;
-	mqd->cp_mqd_base_addr_hi = upper_32_bits(ring->mqd_gpu_addr);
-
-	/* set MQD vmid to 0 */
-	tmp = RREG32_SOC15(GC, 0, mmCP_MQD_CONTROL);
-	tmp = REG_SET_FIELD(tmp, CP_MQD_CONTROL, VMID, 0);
-	mqd->cp_mqd_control = tmp;
-
-	/* set the pointer to the HQD, this is similar CP_RB0_BASE/_HI */
-	hqd_gpu_addr = ring->gpu_addr >> 8;
-	mqd->cp_hqd_pq_base_lo = hqd_gpu_addr;
-	mqd->cp_hqd_pq_base_hi = upper_32_bits(hqd_gpu_addr);
-
-	/* set up the HQD, this is similar to CP_RB0_CNTL */
-	tmp = RREG32_SOC15(GC, 0, mmCP_HQD_PQ_CONTROL);
-	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, QUEUE_SIZE,
-			    (order_base_2(ring->ring_size / 4) - 1));
-	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, RPTR_BLOCK_SIZE,
-			    ((order_base_2(AMDGPU_GPU_PAGE_SIZE / 4) - 1) << 8));
-#ifdef __BIG_ENDIAN
-	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, ENDIAN_SWAP, 1);
-#endif
-	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, UNORD_DISPATCH, 0);
-	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, TUNNEL_DISPATCH, 0);
-	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, PRIV_STATE, 1);
-	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, KMD_QUEUE, 1);
-	mqd->cp_hqd_pq_control = tmp;
-
-	/* set the wb address whether it's enabled or not */
-	wb_gpu_addr = adev->wb.gpu_addr + (ring->rptr_offs * 4);
-	mqd->cp_hqd_pq_rptr_report_addr_lo = wb_gpu_addr & 0xfffffffc;
-	mqd->cp_hqd_pq_rptr_report_addr_hi =
-		upper_32_bits(wb_gpu_addr) & 0xffff;
-
-	/* only used if CP_PQ_WPTR_POLL_CNTL.CP_PQ_WPTR_POLL_CNTL__EN_MASK=1 */
-	wb_gpu_addr = adev->wb.gpu_addr + (ring->wptr_offs * 4);
-	mqd->cp_hqd_pq_wptr_poll_addr_lo = wb_gpu_addr & 0xfffffff8;
-	mqd->cp_hqd_pq_wptr_poll_addr_hi = upper_32_bits(wb_gpu_addr) & 0xffff;
-
-	tmp = 0;
-	/* enable the doorbell if requested */
-	if (ring->use_doorbell) {
-		tmp = RREG32_SOC15(GC, 0, mmCP_HQD_PQ_DOORBELL_CONTROL);
-		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
-				DOORBELL_OFFSET, ring->doorbell_index);
-
-		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
-				    DOORBELL_EN, 1);
-		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
-				    DOORBELL_SOURCE, 0);
-		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
-				    DOORBELL_HIT, 0);
-	}
-
-	mqd->cp_hqd_pq_doorbell_control = tmp;
-
-	/* reset read and write pointers, similar to CP_RB0_WPTR/_RPTR */
-	ring->wptr = 0;
-	mqd->cp_hqd_pq_rptr = RREG32_SOC15(GC, 0, mmCP_HQD_PQ_RPTR);
-
-	/* set the vmid for the queue */
-	mqd->cp_hqd_vmid = 0;
-
-	tmp = RREG32_SOC15(GC, 0, mmCP_HQD_PERSISTENT_STATE);
-	tmp = REG_SET_FIELD(tmp, CP_HQD_PERSISTENT_STATE, PRELOAD_SIZE, 0x53);
-	mqd->cp_hqd_persistent_state = tmp;
-
-	/* set MIN_IB_AVAIL_SIZE */
-	tmp = RREG32_SOC15(GC, 0, mmCP_HQD_IB_CONTROL);
-	tmp = REG_SET_FIELD(tmp, CP_HQD_IB_CONTROL, MIN_IB_AVAIL_SIZE, 3);
-	mqd->cp_hqd_ib_control = tmp;
-
-	/* activate the queue */
-	mqd->cp_hqd_active = 1;
-	return 0;
-}
-
-static void mes_v10_1_queue_init_register(struct amdgpu_ring *ring)
-{
-	struct v10_compute_mqd *mqd = ring->mqd_ptr;
-	struct amdgpu_device *adev = ring->adev;
-	uint32_t data = 0;
-
-	mutex_lock(&adev->srbm_mutex);
-	nv_grbm_select(adev, 3, 0, 0, 0);
-
-	/* set CP_HQD_VMID.VMID = 0. */
-	data = RREG32_SOC15(GC, 0, mmCP_HQD_VMID);
-	data = REG_SET_FIELD(data, CP_HQD_VMID, VMID, 0);
-	WREG32_SOC15(GC, 0, mmCP_HQD_VMID, data);
-
-	/* set CP_HQD_PQ_DOORBELL_CONTROL.DOORBELL_EN=0 */
-	data = RREG32_SOC15(GC, 0, mmCP_HQD_PQ_DOORBELL_CONTROL);
-	data = REG_SET_FIELD(data, CP_HQD_PQ_DOORBELL_CONTROL,
-			     DOORBELL_EN, 0);
-	WREG32_SOC15(GC, 0, mmCP_HQD_PQ_DOORBELL_CONTROL, data);
-
-	/* set CP_MQD_BASE_ADDR/HI with the MQD base address */
-	WREG32_SOC15(GC, 0, mmCP_MQD_BASE_ADDR, mqd->cp_mqd_base_addr_lo);
-	WREG32_SOC15(GC, 0, mmCP_MQD_BASE_ADDR_HI, mqd->cp_mqd_base_addr_hi);
-
-	/* set CP_MQD_CONTROL.VMID=0 */
-	data = RREG32_SOC15(GC, 0, mmCP_MQD_CONTROL);
-	data = REG_SET_FIELD(data, CP_MQD_CONTROL, VMID, 0);
-	WREG32_SOC15(GC, 0, mmCP_MQD_CONTROL, 0);
-
-	/* set CP_HQD_PQ_BASE/HI with the ring buffer base address */
-	WREG32_SOC15(GC, 0, mmCP_HQD_PQ_BASE, mqd->cp_hqd_pq_base_lo);
-	WREG32_SOC15(GC, 0, mmCP_HQD_PQ_BASE_HI, mqd->cp_hqd_pq_base_hi);
-
-	/* set CP_HQD_PQ_RPTR_REPORT_ADDR/HI */
-	WREG32_SOC15(GC, 0, mmCP_HQD_PQ_RPTR_REPORT_ADDR,
-		     mqd->cp_hqd_pq_rptr_report_addr_lo);
-	WREG32_SOC15(GC, 0, mmCP_HQD_PQ_RPTR_REPORT_ADDR_HI,
-		     mqd->cp_hqd_pq_rptr_report_addr_hi);
-
-	/* set CP_HQD_PQ_CONTROL */
-	WREG32_SOC15(GC, 0, mmCP_HQD_PQ_CONTROL, mqd->cp_hqd_pq_control);
-
-	/* set CP_HQD_PQ_WPTR_POLL_ADDR/HI */
-	WREG32_SOC15(GC, 0, mmCP_HQD_PQ_WPTR_POLL_ADDR,
-		     mqd->cp_hqd_pq_wptr_poll_addr_lo);
-	WREG32_SOC15(GC, 0, mmCP_HQD_PQ_WPTR_POLL_ADDR_HI,
-		     mqd->cp_hqd_pq_wptr_poll_addr_hi);
-
-	/* set CP_HQD_PQ_DOORBELL_CONTROL */
-	WREG32_SOC15(GC, 0, mmCP_HQD_PQ_DOORBELL_CONTROL,
-		     mqd->cp_hqd_pq_doorbell_control);
-
-	/* set CP_HQD_PERSISTENT_STATE.PRELOAD_SIZE=0x53 */
-	WREG32_SOC15(GC, 0, mmCP_HQD_PERSISTENT_STATE, mqd->cp_hqd_persistent_state);
-
-	/* set CP_HQD_ACTIVE.ACTIVE=1 */
-	WREG32_SOC15(GC, 0, mmCP_HQD_ACTIVE, mqd->cp_hqd_active);
-
-	nv_grbm_select(adev, 0, 0, 0, 0);
-	mutex_unlock(&adev->srbm_mutex);
-}
-
-#if 0
-static int mes_v10_1_kiq_enable_queue(struct amdgpu_device *adev)
-{
-	struct amdgpu_kiq *kiq = &adev->gfx.kiq;
-	struct amdgpu_ring *kiq_ring = &adev->gfx.kiq.ring;
-	int r;
-
-	if (!kiq->pmf || !kiq->pmf->kiq_map_queues)
-		return -EINVAL;
-
-	r = amdgpu_ring_alloc(kiq_ring, kiq->pmf->map_queues_size);
-	if (r) {
-		DRM_ERROR("Failed to lock KIQ (%d).\n", r);
-		return r;
-	}
-
-	kiq->pmf->kiq_map_queues(kiq_ring, &adev->mes.ring);
-
-	r = amdgpu_ring_test_ring(kiq_ring);
-	if (r) {
-		DRM_ERROR("kfq enable failed\n");
-		kiq_ring->sched.ready = false;
-	}
-	return r;
-}
-#endif
-
-static int mes_v10_1_queue_init(struct amdgpu_device *adev)
-{
-	int r;
-
-	r = mes_v10_1_mqd_init(&adev->mes.ring);
-	if (r)
-		return r;
-
-#if 0
-	r = mes_v10_1_kiq_enable_queue(adev);
-	if (r)
-		return r;
-#else
-	mes_v10_1_queue_init_register(&adev->mes.ring);
-#endif
-
-	return 0;
-}
-
-static int mes_v10_1_ring_init(struct amdgpu_device *adev)
-{
-	struct amdgpu_ring *ring;
-
-	ring = &adev->mes.ring;
-
-	ring->funcs = &mes_v10_1_ring_funcs;
-
-	ring->me = 3;
-	ring->pipe = 0;
-	ring->queue = 0;
-
-	ring->ring_obj = NULL;
-	ring->use_doorbell = true;
-	ring->doorbell_index = adev->doorbell_index.mes_ring << 1;
-	ring->eop_gpu_addr = adev->mes.eop_gpu_addr;
-	ring->no_scheduler = true;
-	sprintf(ring->name, "mes_%d.%d.%d", ring->me, ring->pipe, ring->queue);
-
-	return amdgpu_ring_init(adev, ring, 1024, NULL, 0,
-				AMDGPU_RING_PRIO_DEFAULT, NULL);
-}
-
-static int mes_v10_1_mqd_sw_init(struct amdgpu_device *adev)
-{
-	int r, mqd_size = sizeof(struct v10_compute_mqd);
-	struct amdgpu_ring *ring = &adev->mes.ring;
-
-	if (ring->mqd_obj)
-		return 0;
-
-	r = amdgpu_bo_create_kernel(adev, mqd_size, PAGE_SIZE,
-				    AMDGPU_GEM_DOMAIN_GTT, &ring->mqd_obj,
-				    &ring->mqd_gpu_addr, &ring->mqd_ptr);
-	if (r) {
-		dev_warn(adev->dev, "failed to create ring mqd bo (%d)", r);
-		return r;
-	}
-
-	/* prepare MQD backup */
-	adev->mes.mqd_backup = kmalloc(mqd_size, GFP_KERNEL);
-	if (!adev->mes.mqd_backup)
-		dev_warn(adev->dev,
-			 "no memory to create MQD backup for ring %s\n",
-			 ring->name);
-
-	return 0;
-}
-
-static int mes_v10_1_sw_init(void *handle)
-{
-	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	adev->mes.adev = adev;
-	adev->mes.funcs = &mes_v10_1_funcs;
-
-	r = mes_v10_1_init_microcode(adev);
-	if (r)
-		return r;
-
-	r = mes_v10_1_allocate_eop_buf(adev);
-	if (r)
-		return r;
-
-	r = mes_v10_1_mqd_sw_init(adev);
-	if (r)
-		return r;
-
-	r = mes_v10_1_ring_init(adev);
-	if (r)
-		return r;
-
-	r = mes_v10_1_allocate_mem_slots(adev);
-	if (r)
-		return r;
-
-	return 0;
-}
-
-static int mes_v10_1_sw_fini(void *handle)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	amdgpu_device_wb_free(adev, adev->mes.sch_ctx_offs);
-	amdgpu_device_wb_free(adev, adev->mes.query_status_fence_offs);
-
-	kfree(adev->mes.mqd_backup);
-
-	amdgpu_bo_free_kernel(&adev->mes.ring.mqd_obj,
-			      &adev->mes.ring.mqd_gpu_addr,
-			      &adev->mes.ring.mqd_ptr);
-
-	amdgpu_bo_free_kernel(&adev->mes.eop_gpu_obj,
-			      &adev->mes.eop_gpu_addr,
-			      NULL);
-
-	mes_v10_1_free_microcode(adev);
-
-	return 0;
-}
-
-static int mes_v10_1_hw_init(void *handle)
-{
-	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	if (adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT) {
-		r = mes_v10_1_load_microcode(adev);
-		if (r) {
-			DRM_ERROR("failed to MES fw, r=%d\n", r);
-			return r;
-		}
-	}
-
-	mes_v10_1_enable(adev, true);
-
-	r = mes_v10_1_queue_init(adev);
-	if (r)
-		goto failure;
-
-	r = mes_v10_1_set_hw_resources(&adev->mes);
-	if (r)
-		goto failure;
-
-	r = mes_v10_1_query_sched_status(&adev->mes);
-	if (r) {
-		DRM_ERROR("MES is busy\n");
-		goto failure;
-	}
-
-	return 0;
-
-failure:
-	mes_v10_1_hw_fini(adev);
-	return r;
-}
-
-static int mes_v10_1_hw_fini(void *handle)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	mes_v10_1_enable(adev, false);
-
-	if (adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT)
-		mes_v10_1_free_ucode_buffers(adev);
-
-	return 0;
-}
-
-static int mes_v10_1_suspend(void *handle)
-{
-	return 0;
-}
-
-static int mes_v10_1_resume(void *handle)
-{
-	return 0;
-}
-
-static const struct amd_ip_funcs mes_v10_1_ip_funcs = {
-	.name = "mes_v10_1",
-	.sw_init = mes_v10_1_sw_init,
-	.sw_fini = mes_v10_1_sw_fini,
-	.hw_init = mes_v10_1_hw_init,
-	.hw_fini = mes_v10_1_hw_fini,
-	.suspend = mes_v10_1_suspend,
-	.resume = mes_v10_1_resume,
-};
-
-const struct amdgpu_ip_block_version mes_v10_1_ip_block = {
-	.type = AMD_IP_BLOCK_TYPE_MES,
-	.major = 10,
-	.minor = 1,
-	.rev = 0,
-	.funcs = &mes_v10_1_ip_funcs,
-};
diff --git a/drivers/gpu/drm/amd/amdgpu/mes_v10_1.h b/drivers/gpu/drm/amd/amdgpu/mes_v10_1.h
deleted file mode 100644
index 9afd6ddb01e9..000000000000
--- a/drivers/gpu/drm/amd/amdgpu/mes_v10_1.h
+++ /dev/null
@@ -1,29 +0,0 @@
-/*
- * Copyright 2019 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- */
-
-#ifndef __MES_V10_1_H__
-#define __MES_V10_1_H__
-
-extern const struct amdgpu_ip_block_version mes_v10_1_ip_block;
-
-#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c b/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c
new file mode 100644
index 000000000000..3a52754b5cad
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c
@@ -0,0 +1,1760 @@
+/*
+ * Copyright 2019 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/firmware.h>
+#include <linux/module.h>
+#include "amdgpu.h"
+#include "soc15_common.h"
+#include "soc21.h"
+#include "gfx_v11_0.h"
+#include "gc/gc_11_0_0_offset.h"
+#include "gc/gc_11_0_0_sh_mask.h"
+#include "gc/gc_11_0_0_default.h"
+#include "v11_structs.h"
+#include "mes_v11_api_def.h"
+
+MODULE_FIRMWARE("amdgpu/gc_11_0_0_mes.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_0_0_mes_2.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_0_0_mes1.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_0_1_mes.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_0_1_mes_2.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_0_1_mes1.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_0_2_mes.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_0_2_mes_2.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_0_2_mes1.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_0_3_mes.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_0_3_mes_2.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_0_3_mes1.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_0_4_mes.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_0_4_mes_2.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_0_4_mes1.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_5_0_mes_2.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_5_0_mes1.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_5_1_mes_2.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_5_1_mes1.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_5_2_mes_2.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_5_2_mes1.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_5_3_mes_2.bin");
+MODULE_FIRMWARE("amdgpu/gc_11_5_3_mes1.bin");
+
+static int mes_v11_0_hw_init(struct amdgpu_ip_block *ip_block);
+static int mes_v11_0_hw_fini(struct amdgpu_ip_block *ip_block);
+static int mes_v11_0_kiq_hw_init(struct amdgpu_device *adev);
+static int mes_v11_0_kiq_hw_fini(struct amdgpu_device *adev);
+
+#define MES_EOP_SIZE   2048
+#define GFX_MES_DRAM_SIZE	0x80000
+#define MES11_HW_RESOURCE_1_SIZE (128 * AMDGPU_GPU_PAGE_SIZE)
+
+#define MES11_HUNG_DB_OFFSET_ARRAY_SIZE 8 /* [0:3] = db offset, [4:7] = hqd info */
+#define MES11_HUNG_HQD_INFO_OFFSET	4
+
+static void mes_v11_0_ring_set_wptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	if (ring->use_doorbell) {
+		atomic64_set((atomic64_t *)ring->wptr_cpu_addr,
+			     ring->wptr);
+		WDOORBELL64(ring->doorbell_index, ring->wptr);
+	} else {
+		BUG();
+	}
+}
+
+static u64 mes_v11_0_ring_get_rptr(struct amdgpu_ring *ring)
+{
+	return *ring->rptr_cpu_addr;
+}
+
+static u64 mes_v11_0_ring_get_wptr(struct amdgpu_ring *ring)
+{
+	u64 wptr;
+
+	if (ring->use_doorbell)
+		wptr = atomic64_read((atomic64_t *)ring->wptr_cpu_addr);
+	else
+		BUG();
+	return wptr;
+}
+
+static const struct amdgpu_ring_funcs mes_v11_0_ring_funcs = {
+	.type = AMDGPU_RING_TYPE_MES,
+	.align_mask = 1,
+	.nop = 0,
+	.support_64bit_ptrs = true,
+	.get_rptr = mes_v11_0_ring_get_rptr,
+	.get_wptr = mes_v11_0_ring_get_wptr,
+	.set_wptr = mes_v11_0_ring_set_wptr,
+	.insert_nop = amdgpu_ring_insert_nop,
+};
+
+static const char *mes_v11_0_opcodes[] = {
+	"SET_HW_RSRC",
+	"SET_SCHEDULING_CONFIG",
+	"ADD_QUEUE",
+	"REMOVE_QUEUE",
+	"PERFORM_YIELD",
+	"SET_GANG_PRIORITY_LEVEL",
+	"SUSPEND",
+	"RESUME",
+	"RESET",
+	"SET_LOG_BUFFER",
+	"CHANGE_GANG_PRORITY",
+	"QUERY_SCHEDULER_STATUS",
+	"PROGRAM_GDS",
+	"SET_DEBUG_VMID",
+	"MISC",
+	"UPDATE_ROOT_PAGE_TABLE",
+	"AMD_LOG",
+	"unused",
+	"unused",
+	"SET_HW_RSRC_1",
+};
+
+static const char *mes_v11_0_misc_opcodes[] = {
+	"WRITE_REG",
+	"INV_GART",
+	"QUERY_STATUS",
+	"READ_REG",
+	"WAIT_REG_MEM",
+	"SET_SHADER_DEBUGGER",
+};
+
+static const char *mes_v11_0_get_op_string(union MESAPI__MISC *x_pkt)
+{
+	const char *op_str = NULL;
+
+	if (x_pkt->header.opcode < ARRAY_SIZE(mes_v11_0_opcodes))
+		op_str = mes_v11_0_opcodes[x_pkt->header.opcode];
+
+	return op_str;
+}
+
+static const char *mes_v11_0_get_misc_op_string(union MESAPI__MISC *x_pkt)
+{
+	const char *op_str = NULL;
+
+	if ((x_pkt->header.opcode == MES_SCH_API_MISC) &&
+	    (x_pkt->opcode < ARRAY_SIZE(mes_v11_0_misc_opcodes)))
+		op_str = mes_v11_0_misc_opcodes[x_pkt->opcode];
+
+	return op_str;
+}
+
+static int mes_v11_0_submit_pkt_and_poll_completion(struct amdgpu_mes *mes,
+						    void *pkt, int size,
+						    int api_status_off)
+{
+	union MESAPI__QUERY_MES_STATUS mes_status_pkt;
+	signed long timeout = 2100000; /* 2100 ms */
+	struct amdgpu_device *adev = mes->adev;
+	struct amdgpu_ring *ring = &mes->ring[0];
+	struct MES_API_STATUS *api_status;
+	union MESAPI__MISC *x_pkt = pkt;
+	const char *op_str, *misc_op_str;
+	unsigned long flags;
+	u64 status_gpu_addr;
+	u32 seq, status_offset;
+	u64 *status_ptr;
+	signed long r;
+	int ret;
+
+	if (x_pkt->header.opcode >= MES_SCH_API_MAX)
+		return -EINVAL;
+
+	if (amdgpu_emu_mode) {
+		timeout *= 100;
+	} else if (amdgpu_sriov_vf(adev)) {
+		/* Worst case in sriov where all other 15 VF timeout, each VF needs about 600ms */
+		timeout = 15 * 600 * 1000;
+	}
+
+	ret = amdgpu_device_wb_get(adev, &status_offset);
+	if (ret)
+		return ret;
+
+	status_gpu_addr = adev->wb.gpu_addr + (status_offset * 4);
+	status_ptr = (u64 *)&adev->wb.wb[status_offset];
+	*status_ptr = 0;
+
+	spin_lock_irqsave(&mes->ring_lock[0], flags);
+	r = amdgpu_ring_alloc(ring, (size + sizeof(mes_status_pkt)) / 4);
+	if (r)
+		goto error_unlock_free;
+
+	seq = ++ring->fence_drv.sync_seq;
+	r = amdgpu_fence_wait_polling(ring,
+				      seq - ring->fence_drv.num_fences_mask,
+				      timeout);
+	if (r < 1)
+		goto error_undo;
+
+	api_status = (struct MES_API_STATUS *)((char *)pkt + api_status_off);
+	api_status->api_completion_fence_addr = status_gpu_addr;
+	api_status->api_completion_fence_value = 1;
+
+	amdgpu_ring_write_multiple(ring, pkt, size / 4);
+
+	memset(&mes_status_pkt, 0, sizeof(mes_status_pkt));
+	mes_status_pkt.header.type = MES_API_TYPE_SCHEDULER;
+	mes_status_pkt.header.opcode = MES_SCH_API_QUERY_SCHEDULER_STATUS;
+	mes_status_pkt.header.dwsize = API_FRAME_SIZE_IN_DWORDS;
+	mes_status_pkt.api_status.api_completion_fence_addr =
+		ring->fence_drv.gpu_addr;
+	mes_status_pkt.api_status.api_completion_fence_value = seq;
+
+	amdgpu_ring_write_multiple(ring, &mes_status_pkt,
+				   sizeof(mes_status_pkt) / 4);
+
+	amdgpu_ring_commit(ring);
+	spin_unlock_irqrestore(&mes->ring_lock[0], flags);
+
+	op_str = mes_v11_0_get_op_string(x_pkt);
+	misc_op_str = mes_v11_0_get_misc_op_string(x_pkt);
+
+	if (misc_op_str)
+		dev_dbg(adev->dev, "MES msg=%s (%s) was emitted\n", op_str,
+			misc_op_str);
+	else if (op_str)
+		dev_dbg(adev->dev, "MES msg=%s was emitted\n", op_str);
+	else
+		dev_dbg(adev->dev, "MES msg=%d was emitted\n",
+			x_pkt->header.opcode);
+
+	r = amdgpu_fence_wait_polling(ring, seq, timeout);
+	if (r < 1 || !*status_ptr) {
+
+		if (misc_op_str)
+			dev_err(adev->dev, "MES failed to respond to msg=%s (%s)\n",
+				op_str, misc_op_str);
+		else if (op_str)
+			dev_err(adev->dev, "MES failed to respond to msg=%s\n",
+				op_str);
+		else
+			dev_err(adev->dev, "MES failed to respond to msg=%d\n",
+				x_pkt->header.opcode);
+
+		while (halt_if_hws_hang)
+			schedule();
+
+		r = -ETIMEDOUT;
+		goto error_wb_free;
+	}
+
+	amdgpu_device_wb_free(adev, status_offset);
+	return 0;
+
+error_undo:
+	dev_err(adev->dev, "MES ring buffer is full.\n");
+	amdgpu_ring_undo(ring);
+
+error_unlock_free:
+	spin_unlock_irqrestore(&mes->ring_lock[0], flags);
+
+error_wb_free:
+	amdgpu_device_wb_free(adev, status_offset);
+	return r;
+}
+
+static int convert_to_mes_queue_type(int queue_type)
+{
+	if (queue_type == AMDGPU_RING_TYPE_GFX)
+		return MES_QUEUE_TYPE_GFX;
+	else if (queue_type == AMDGPU_RING_TYPE_COMPUTE)
+		return MES_QUEUE_TYPE_COMPUTE;
+	else if (queue_type == AMDGPU_RING_TYPE_SDMA)
+		return MES_QUEUE_TYPE_SDMA;
+	else
+		BUG();
+	return -1;
+}
+
+static int convert_to_mes_priority_level(int priority_level)
+{
+	switch (priority_level) {
+	case AMDGPU_MES_PRIORITY_LEVEL_LOW:
+		return AMD_PRIORITY_LEVEL_LOW;
+	case AMDGPU_MES_PRIORITY_LEVEL_NORMAL:
+	default:
+		return AMD_PRIORITY_LEVEL_NORMAL;
+	case AMDGPU_MES_PRIORITY_LEVEL_MEDIUM:
+		return AMD_PRIORITY_LEVEL_MEDIUM;
+	case AMDGPU_MES_PRIORITY_LEVEL_HIGH:
+		return AMD_PRIORITY_LEVEL_HIGH;
+	case AMDGPU_MES_PRIORITY_LEVEL_REALTIME:
+		return AMD_PRIORITY_LEVEL_REALTIME;
+	}
+}
+
+static int mes_v11_0_add_hw_queue(struct amdgpu_mes *mes,
+				  struct mes_add_queue_input *input)
+{
+	struct amdgpu_device *adev = mes->adev;
+	union MESAPI__ADD_QUEUE mes_add_queue_pkt;
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)];
+	uint32_t vm_cntx_cntl = hub->vm_cntx_cntl;
+
+	memset(&mes_add_queue_pkt, 0, sizeof(mes_add_queue_pkt));
+
+	mes_add_queue_pkt.header.type = MES_API_TYPE_SCHEDULER;
+	mes_add_queue_pkt.header.opcode = MES_SCH_API_ADD_QUEUE;
+	mes_add_queue_pkt.header.dwsize = API_FRAME_SIZE_IN_DWORDS;
+
+	mes_add_queue_pkt.process_id = input->process_id;
+	mes_add_queue_pkt.page_table_base_addr = input->page_table_base_addr;
+	mes_add_queue_pkt.process_va_start = input->process_va_start;
+	mes_add_queue_pkt.process_va_end = input->process_va_end;
+	mes_add_queue_pkt.process_quantum = input->process_quantum;
+	mes_add_queue_pkt.process_context_addr = input->process_context_addr;
+	mes_add_queue_pkt.gang_quantum = input->gang_quantum;
+	mes_add_queue_pkt.gang_context_addr = input->gang_context_addr;
+	mes_add_queue_pkt.inprocess_gang_priority =
+		convert_to_mes_priority_level(input->inprocess_gang_priority);
+	mes_add_queue_pkt.gang_global_priority_level =
+		convert_to_mes_priority_level(input->gang_global_priority_level);
+	mes_add_queue_pkt.doorbell_offset = input->doorbell_offset;
+	mes_add_queue_pkt.mqd_addr = input->mqd_addr;
+
+	if (((adev->mes.sched_version & AMDGPU_MES_API_VERSION_MASK) >>
+			AMDGPU_MES_API_VERSION_SHIFT) >= 2)
+		mes_add_queue_pkt.wptr_addr = input->wptr_mc_addr;
+	else
+		mes_add_queue_pkt.wptr_addr = input->wptr_addr;
+
+	mes_add_queue_pkt.queue_type =
+		convert_to_mes_queue_type(input->queue_type);
+	mes_add_queue_pkt.paging = input->paging;
+	mes_add_queue_pkt.vm_context_cntl = vm_cntx_cntl;
+	mes_add_queue_pkt.gws_base = input->gws_base;
+	mes_add_queue_pkt.gws_size = input->gws_size;
+	mes_add_queue_pkt.trap_handler_addr = input->tba_addr;
+	mes_add_queue_pkt.tma_addr = input->tma_addr;
+	mes_add_queue_pkt.trap_en = input->trap_en;
+	mes_add_queue_pkt.skip_process_ctx_clear = input->skip_process_ctx_clear;
+	mes_add_queue_pkt.is_kfd_process = input->is_kfd_process;
+
+	/* For KFD, gds_size is re-used for queue size (needed in MES for AQL queues) */
+	mes_add_queue_pkt.is_aql_queue = input->is_aql_queue;
+	mes_add_queue_pkt.gds_size = input->queue_size;
+
+	mes_add_queue_pkt.exclusively_scheduled = input->exclusively_scheduled;
+
+	return mes_v11_0_submit_pkt_and_poll_completion(mes,
+			&mes_add_queue_pkt, sizeof(mes_add_queue_pkt),
+			offsetof(union MESAPI__ADD_QUEUE, api_status));
+}
+
+static int mes_v11_0_remove_hw_queue(struct amdgpu_mes *mes,
+				     struct mes_remove_queue_input *input)
+{
+	union MESAPI__REMOVE_QUEUE mes_remove_queue_pkt;
+	uint32_t mes_rev = mes->sched_version & AMDGPU_MES_VERSION_MASK;
+
+	memset(&mes_remove_queue_pkt, 0, sizeof(mes_remove_queue_pkt));
+
+	mes_remove_queue_pkt.header.type = MES_API_TYPE_SCHEDULER;
+	mes_remove_queue_pkt.header.opcode = MES_SCH_API_REMOVE_QUEUE;
+	mes_remove_queue_pkt.header.dwsize = API_FRAME_SIZE_IN_DWORDS;
+
+	mes_remove_queue_pkt.doorbell_offset = input->doorbell_offset;
+	mes_remove_queue_pkt.gang_context_addr = input->gang_context_addr;
+
+	if (mes_rev >= 0x60)
+		mes_remove_queue_pkt.remove_queue_after_reset = input->remove_queue_after_reset;
+
+	return mes_v11_0_submit_pkt_and_poll_completion(mes,
+			&mes_remove_queue_pkt, sizeof(mes_remove_queue_pkt),
+			offsetof(union MESAPI__REMOVE_QUEUE, api_status));
+}
+
+static int mes_v11_0_reset_queue_mmio(struct amdgpu_mes *mes, uint32_t queue_type,
+				      uint32_t me_id, uint32_t pipe_id,
+				      uint32_t queue_id, uint32_t vmid)
+{
+	struct amdgpu_device *adev = mes->adev;
+	uint32_t value, reg;
+	int i, r = 0;
+
+	amdgpu_gfx_rlc_enter_safe_mode(adev, 0);
+
+	if (queue_type == AMDGPU_RING_TYPE_GFX) {
+		dev_info(adev->dev, "reset gfx queue (%d:%d:%d: vmid:%d)\n",
+			 me_id, pipe_id, queue_id, vmid);
+
+		mutex_lock(&adev->gfx.reset_sem_mutex);
+		gfx_v11_0_request_gfx_index_mutex(adev, true);
+		/* all se allow writes */
+		WREG32_SOC15(GC, 0, regGRBM_GFX_INDEX,
+			     (uint32_t)(0x1 << GRBM_GFX_INDEX__SE_BROADCAST_WRITES__SHIFT));
+		value = REG_SET_FIELD(0, CP_VMID_RESET, RESET_REQUEST, 1 << vmid);
+		if (pipe_id == 0)
+			value = REG_SET_FIELD(value, CP_VMID_RESET, PIPE0_QUEUES, 1 << queue_id);
+		else
+			value = REG_SET_FIELD(value, CP_VMID_RESET, PIPE1_QUEUES, 1 << queue_id);
+		WREG32_SOC15(GC, 0, regCP_VMID_RESET, value);
+		gfx_v11_0_request_gfx_index_mutex(adev, false);
+		mutex_unlock(&adev->gfx.reset_sem_mutex);
+
+		mutex_lock(&adev->srbm_mutex);
+		soc21_grbm_select(adev, me_id, pipe_id, queue_id, 0);
+		/* wait till dequeue take effects */
+		for (i = 0; i < adev->usec_timeout; i++) {
+			if (!(RREG32_SOC15(GC, 0, regCP_GFX_HQD_ACTIVE) & 1))
+				break;
+			udelay(1);
+		}
+		if (i >= adev->usec_timeout) {
+			dev_err(adev->dev, "failed to wait on gfx hqd deactivate\n");
+			r = -ETIMEDOUT;
+		}
+
+		soc21_grbm_select(adev, 0, 0, 0, 0);
+		mutex_unlock(&adev->srbm_mutex);
+	} else if (queue_type == AMDGPU_RING_TYPE_COMPUTE) {
+		dev_info(adev->dev, "reset compute queue (%d:%d:%d)\n",
+			 me_id, pipe_id, queue_id);
+		mutex_lock(&adev->srbm_mutex);
+		soc21_grbm_select(adev, me_id, pipe_id, queue_id, 0);
+		WREG32_SOC15(GC, 0, regCP_HQD_DEQUEUE_REQUEST, 0x2);
+		WREG32_SOC15(GC, 0, regSPI_COMPUTE_QUEUE_RESET, 0x1);
+
+		/* wait till dequeue take effects */
+		for (i = 0; i < adev->usec_timeout; i++) {
+			if (!(RREG32_SOC15(GC, 0, regCP_HQD_ACTIVE) & 1))
+				break;
+			udelay(1);
+		}
+		if (i >= adev->usec_timeout) {
+			dev_err(adev->dev, "failed to wait on hqd deactivate\n");
+			r = -ETIMEDOUT;
+		}
+		soc21_grbm_select(adev, 0, 0, 0, 0);
+		mutex_unlock(&adev->srbm_mutex);
+	} else if (queue_type == AMDGPU_RING_TYPE_SDMA) {
+		dev_info(adev->dev, "reset sdma queue (%d:%d:%d)\n",
+			 me_id, pipe_id, queue_id);
+		switch (me_id) {
+		case 1:
+			reg = SOC15_REG_OFFSET(GC, 0, regSDMA1_QUEUE_RESET_REQ);
+			break;
+		case 0:
+		default:
+			reg = SOC15_REG_OFFSET(GC, 0, regSDMA0_QUEUE_RESET_REQ);
+			break;
+		}
+
+		value = 1 << queue_id;
+		WREG32(reg, value);
+		/* wait for queue reset done */
+		for (i = 0; i < adev->usec_timeout; i++) {
+			if (!(RREG32(reg) & value))
+				break;
+			udelay(1);
+		}
+		if (i >= adev->usec_timeout) {
+			dev_err(adev->dev, "failed to wait on sdma queue reset done\n");
+			r = -ETIMEDOUT;
+		}
+	}
+
+	amdgpu_gfx_rlc_exit_safe_mode(adev, 0);
+	return r;
+}
+
+static int mes_v11_0_map_legacy_queue(struct amdgpu_mes *mes,
+				      struct mes_map_legacy_queue_input *input)
+{
+	union MESAPI__ADD_QUEUE mes_add_queue_pkt;
+
+	memset(&mes_add_queue_pkt, 0, sizeof(mes_add_queue_pkt));
+
+	mes_add_queue_pkt.header.type = MES_API_TYPE_SCHEDULER;
+	mes_add_queue_pkt.header.opcode = MES_SCH_API_ADD_QUEUE;
+	mes_add_queue_pkt.header.dwsize = API_FRAME_SIZE_IN_DWORDS;
+
+	mes_add_queue_pkt.pipe_id = input->pipe_id;
+	mes_add_queue_pkt.queue_id = input->queue_id;
+	mes_add_queue_pkt.doorbell_offset = input->doorbell_offset;
+	mes_add_queue_pkt.mqd_addr = input->mqd_addr;
+	mes_add_queue_pkt.wptr_addr = input->wptr_addr;
+	mes_add_queue_pkt.queue_type =
+		convert_to_mes_queue_type(input->queue_type);
+	mes_add_queue_pkt.map_legacy_kq = 1;
+
+	return mes_v11_0_submit_pkt_and_poll_completion(mes,
+			&mes_add_queue_pkt, sizeof(mes_add_queue_pkt),
+			offsetof(union MESAPI__ADD_QUEUE, api_status));
+}
+
+static int mes_v11_0_unmap_legacy_queue(struct amdgpu_mes *mes,
+			struct mes_unmap_legacy_queue_input *input)
+{
+	union MESAPI__REMOVE_QUEUE mes_remove_queue_pkt;
+
+	memset(&mes_remove_queue_pkt, 0, sizeof(mes_remove_queue_pkt));
+
+	mes_remove_queue_pkt.header.type = MES_API_TYPE_SCHEDULER;
+	mes_remove_queue_pkt.header.opcode = MES_SCH_API_REMOVE_QUEUE;
+	mes_remove_queue_pkt.header.dwsize = API_FRAME_SIZE_IN_DWORDS;
+
+	mes_remove_queue_pkt.doorbell_offset = input->doorbell_offset;
+	mes_remove_queue_pkt.gang_context_addr = 0;
+
+	mes_remove_queue_pkt.pipe_id = input->pipe_id;
+	mes_remove_queue_pkt.queue_id = input->queue_id;
+
+	if (input->action == PREEMPT_QUEUES_NO_UNMAP) {
+		mes_remove_queue_pkt.preempt_legacy_gfx_queue = 1;
+		mes_remove_queue_pkt.tf_addr = input->trail_fence_addr;
+		mes_remove_queue_pkt.tf_data =
+			lower_32_bits(input->trail_fence_data);
+	} else {
+		mes_remove_queue_pkt.unmap_legacy_queue = 1;
+		mes_remove_queue_pkt.queue_type =
+			convert_to_mes_queue_type(input->queue_type);
+	}
+
+	return mes_v11_0_submit_pkt_and_poll_completion(mes,
+			&mes_remove_queue_pkt, sizeof(mes_remove_queue_pkt),
+			offsetof(union MESAPI__REMOVE_QUEUE, api_status));
+}
+
+static int mes_v11_0_suspend_gang(struct amdgpu_mes *mes,
+				  struct mes_suspend_gang_input *input)
+{
+	union MESAPI__SUSPEND mes_suspend_gang_pkt;
+
+	memset(&mes_suspend_gang_pkt, 0, sizeof(mes_suspend_gang_pkt));
+
+	mes_suspend_gang_pkt.header.type = MES_API_TYPE_SCHEDULER;
+	mes_suspend_gang_pkt.header.opcode = MES_SCH_API_SUSPEND;
+	mes_suspend_gang_pkt.header.dwsize = API_FRAME_SIZE_IN_DWORDS;
+
+	mes_suspend_gang_pkt.suspend_all_gangs = input->suspend_all_gangs;
+	mes_suspend_gang_pkt.gang_context_addr = input->gang_context_addr;
+	mes_suspend_gang_pkt.suspend_fence_addr = input->suspend_fence_addr;
+	mes_suspend_gang_pkt.suspend_fence_value = input->suspend_fence_value;
+
+	return mes_v11_0_submit_pkt_and_poll_completion(mes,
+			&mes_suspend_gang_pkt, sizeof(mes_suspend_gang_pkt),
+			offsetof(union MESAPI__SUSPEND, api_status));
+}
+
+static int mes_v11_0_resume_gang(struct amdgpu_mes *mes,
+				 struct mes_resume_gang_input *input)
+{
+	union MESAPI__RESUME mes_resume_gang_pkt;
+
+	memset(&mes_resume_gang_pkt, 0, sizeof(mes_resume_gang_pkt));
+
+	mes_resume_gang_pkt.header.type = MES_API_TYPE_SCHEDULER;
+	mes_resume_gang_pkt.header.opcode = MES_SCH_API_RESUME;
+	mes_resume_gang_pkt.header.dwsize = API_FRAME_SIZE_IN_DWORDS;
+
+	mes_resume_gang_pkt.resume_all_gangs = input->resume_all_gangs;
+	mes_resume_gang_pkt.gang_context_addr = input->gang_context_addr;
+
+	return mes_v11_0_submit_pkt_and_poll_completion(mes,
+			&mes_resume_gang_pkt, sizeof(mes_resume_gang_pkt),
+			offsetof(union MESAPI__RESUME, api_status));
+}
+
+static int mes_v11_0_query_sched_status(struct amdgpu_mes *mes)
+{
+	union MESAPI__QUERY_MES_STATUS mes_status_pkt;
+
+	memset(&mes_status_pkt, 0, sizeof(mes_status_pkt));
+
+	mes_status_pkt.header.type = MES_API_TYPE_SCHEDULER;
+	mes_status_pkt.header.opcode = MES_SCH_API_QUERY_SCHEDULER_STATUS;
+	mes_status_pkt.header.dwsize = API_FRAME_SIZE_IN_DWORDS;
+
+	return mes_v11_0_submit_pkt_and_poll_completion(mes,
+			&mes_status_pkt, sizeof(mes_status_pkt),
+			offsetof(union MESAPI__QUERY_MES_STATUS, api_status));
+}
+
+static int mes_v11_0_misc_op(struct amdgpu_mes *mes,
+			     struct mes_misc_op_input *input)
+{
+	union MESAPI__MISC misc_pkt;
+
+	memset(&misc_pkt, 0, sizeof(misc_pkt));
+
+	misc_pkt.header.type = MES_API_TYPE_SCHEDULER;
+	misc_pkt.header.opcode = MES_SCH_API_MISC;
+	misc_pkt.header.dwsize = API_FRAME_SIZE_IN_DWORDS;
+
+	switch (input->op) {
+	case MES_MISC_OP_READ_REG:
+		misc_pkt.opcode = MESAPI_MISC__READ_REG;
+		misc_pkt.read_reg.reg_offset = input->read_reg.reg_offset;
+		misc_pkt.read_reg.buffer_addr = input->read_reg.buffer_addr;
+		break;
+	case MES_MISC_OP_WRITE_REG:
+		misc_pkt.opcode = MESAPI_MISC__WRITE_REG;
+		misc_pkt.write_reg.reg_offset = input->write_reg.reg_offset;
+		misc_pkt.write_reg.reg_value = input->write_reg.reg_value;
+		break;
+	case MES_MISC_OP_WRM_REG_WAIT:
+		misc_pkt.opcode = MESAPI_MISC__WAIT_REG_MEM;
+		misc_pkt.wait_reg_mem.op = WRM_OPERATION__WAIT_REG_MEM;
+		misc_pkt.wait_reg_mem.reference = input->wrm_reg.ref;
+		misc_pkt.wait_reg_mem.mask = input->wrm_reg.mask;
+		misc_pkt.wait_reg_mem.reg_offset1 = input->wrm_reg.reg0;
+		misc_pkt.wait_reg_mem.reg_offset2 = 0;
+		break;
+	case MES_MISC_OP_WRM_REG_WR_WAIT:
+		misc_pkt.opcode = MESAPI_MISC__WAIT_REG_MEM;
+		misc_pkt.wait_reg_mem.op = WRM_OPERATION__WR_WAIT_WR_REG;
+		misc_pkt.wait_reg_mem.reference = input->wrm_reg.ref;
+		misc_pkt.wait_reg_mem.mask = input->wrm_reg.mask;
+		misc_pkt.wait_reg_mem.reg_offset1 = input->wrm_reg.reg0;
+		misc_pkt.wait_reg_mem.reg_offset2 = input->wrm_reg.reg1;
+		break;
+	case MES_MISC_OP_SET_SHADER_DEBUGGER:
+		misc_pkt.opcode = MESAPI_MISC__SET_SHADER_DEBUGGER;
+		misc_pkt.set_shader_debugger.process_context_addr =
+				input->set_shader_debugger.process_context_addr;
+		misc_pkt.set_shader_debugger.flags.u32all =
+				input->set_shader_debugger.flags.u32all;
+		misc_pkt.set_shader_debugger.spi_gdbg_per_vmid_cntl =
+				input->set_shader_debugger.spi_gdbg_per_vmid_cntl;
+		memcpy(misc_pkt.set_shader_debugger.tcp_watch_cntl,
+				input->set_shader_debugger.tcp_watch_cntl,
+				sizeof(misc_pkt.set_shader_debugger.tcp_watch_cntl));
+		misc_pkt.set_shader_debugger.trap_en = input->set_shader_debugger.trap_en;
+		break;
+	case MES_MISC_OP_CHANGE_CONFIG:
+		if ((mes->adev->mes.sched_version & AMDGPU_MES_VERSION_MASK) < 0x63) {
+			dev_warn_once(mes->adev->dev,
+				      "MES FW version must be larger than 0x63 to support limit single process feature.\n");
+			return 0;
+		}
+		misc_pkt.opcode = MESAPI_MISC__CHANGE_CONFIG;
+		misc_pkt.change_config.opcode =
+				MESAPI_MISC__CHANGE_CONFIG_OPTION_LIMIT_SINGLE_PROCESS;
+		misc_pkt.change_config.option.bits.limit_single_process =
+				input->change_config.option.limit_single_process;
+		break;
+
+	default:
+		DRM_ERROR("unsupported misc op (%d) \n", input->op);
+		return -EINVAL;
+	}
+
+	return mes_v11_0_submit_pkt_and_poll_completion(mes,
+			&misc_pkt, sizeof(misc_pkt),
+			offsetof(union MESAPI__MISC, api_status));
+}
+
+static int mes_v11_0_set_hw_resources(struct amdgpu_mes *mes)
+{
+	int i;
+	struct amdgpu_device *adev = mes->adev;
+	union MESAPI_SET_HW_RESOURCES mes_set_hw_res_pkt;
+
+	memset(&mes_set_hw_res_pkt, 0, sizeof(mes_set_hw_res_pkt));
+
+	mes_set_hw_res_pkt.header.type = MES_API_TYPE_SCHEDULER;
+	mes_set_hw_res_pkt.header.opcode = MES_SCH_API_SET_HW_RSRC;
+	mes_set_hw_res_pkt.header.dwsize = API_FRAME_SIZE_IN_DWORDS;
+
+	mes_set_hw_res_pkt.vmid_mask_mmhub = mes->vmid_mask_mmhub;
+	mes_set_hw_res_pkt.vmid_mask_gfxhub = mes->vmid_mask_gfxhub;
+	mes_set_hw_res_pkt.gds_size = adev->gds.gds_size;
+	mes_set_hw_res_pkt.paging_vmid = 0;
+	mes_set_hw_res_pkt.g_sch_ctx_gpu_mc_ptr = mes->sch_ctx_gpu_addr[0];
+	mes_set_hw_res_pkt.query_status_fence_gpu_mc_ptr =
+		mes->query_status_fence_gpu_addr[0];
+
+	for (i = 0; i < MAX_COMPUTE_PIPES; i++)
+		mes_set_hw_res_pkt.compute_hqd_mask[i] =
+			mes->compute_hqd_mask[i];
+
+	for (i = 0; i < MAX_GFX_PIPES; i++)
+		mes_set_hw_res_pkt.gfx_hqd_mask[i] =
+			mes->gfx_hqd_mask[i];
+
+	for (i = 0; i < MAX_SDMA_PIPES; i++)
+		mes_set_hw_res_pkt.sdma_hqd_mask[i] = mes->sdma_hqd_mask[i];
+
+	for (i = 0; i < AMD_PRIORITY_NUM_LEVELS; i++)
+		mes_set_hw_res_pkt.aggregated_doorbells[i] =
+			mes->aggregated_doorbells[i];
+
+	for (i = 0; i < 5; i++) {
+		mes_set_hw_res_pkt.gc_base[i] = adev->reg_offset[GC_HWIP][0][i];
+		mes_set_hw_res_pkt.mmhub_base[i] =
+				adev->reg_offset[MMHUB_HWIP][0][i];
+		mes_set_hw_res_pkt.osssys_base[i] =
+		adev->reg_offset[OSSSYS_HWIP][0][i];
+	}
+
+	mes_set_hw_res_pkt.disable_reset = 1;
+	mes_set_hw_res_pkt.disable_mes_log = 1;
+	mes_set_hw_res_pkt.use_different_vmid_compute = 1;
+	mes_set_hw_res_pkt.enable_reg_active_poll = 1;
+	mes_set_hw_res_pkt.enable_level_process_quantum_check = 1;
+	mes_set_hw_res_pkt.oversubscription_timer = 50;
+	if ((mes->adev->mes.sched_version & AMDGPU_MES_VERSION_MASK) >= 0x7f)
+		mes_set_hw_res_pkt.enable_lr_compute_wa = 1;
+	else
+		dev_info_once(mes->adev->dev,
+			      "MES FW version must be >= 0x7f to enable LR compute workaround.\n");
+
+	if (amdgpu_mes_log_enable) {
+		mes_set_hw_res_pkt.enable_mes_event_int_logging = 1;
+		mes_set_hw_res_pkt.event_intr_history_gpu_mc_ptr =
+					mes->event_log_gpu_addr;
+	}
+
+	if (adev->enforce_isolation[0] == AMDGPU_ENFORCE_ISOLATION_ENABLE)
+		mes_set_hw_res_pkt.limit_single_process = 1;
+
+	return mes_v11_0_submit_pkt_and_poll_completion(mes,
+			&mes_set_hw_res_pkt, sizeof(mes_set_hw_res_pkt),
+			offsetof(union MESAPI_SET_HW_RESOURCES, api_status));
+}
+
+static int mes_v11_0_set_hw_resources_1(struct amdgpu_mes *mes)
+{
+	union MESAPI_SET_HW_RESOURCES_1 mes_set_hw_res_pkt;
+	memset(&mes_set_hw_res_pkt, 0, sizeof(mes_set_hw_res_pkt));
+
+	mes_set_hw_res_pkt.header.type = MES_API_TYPE_SCHEDULER;
+	mes_set_hw_res_pkt.header.opcode = MES_SCH_API_SET_HW_RSRC_1;
+	mes_set_hw_res_pkt.header.dwsize = API_FRAME_SIZE_IN_DWORDS;
+	mes_set_hw_res_pkt.enable_mes_info_ctx = 1;
+
+	mes_set_hw_res_pkt.cleaner_shader_fence_mc_addr = mes->resource_1_gpu_addr[0];
+	if (amdgpu_sriov_is_mes_info_enable(mes->adev)) {
+		mes_set_hw_res_pkt.mes_info_ctx_mc_addr =
+			mes->resource_1_gpu_addr[0] + AMDGPU_GPU_PAGE_SIZE;
+		mes_set_hw_res_pkt.mes_info_ctx_size = MES11_HW_RESOURCE_1_SIZE;
+	}
+
+	return mes_v11_0_submit_pkt_and_poll_completion(mes,
+			&mes_set_hw_res_pkt, sizeof(mes_set_hw_res_pkt),
+			offsetof(union MESAPI_SET_HW_RESOURCES_1, api_status));
+}
+
+static int mes_v11_0_reset_hw_queue(struct amdgpu_mes *mes,
+				    struct mes_reset_queue_input *input)
+{
+	union MESAPI__RESET mes_reset_queue_pkt;
+
+	if (input->use_mmio)
+		return mes_v11_0_reset_queue_mmio(mes, input->queue_type,
+						  input->me_id, input->pipe_id,
+						  input->queue_id, input->vmid);
+
+	memset(&mes_reset_queue_pkt, 0, sizeof(mes_reset_queue_pkt));
+
+	mes_reset_queue_pkt.header.type = MES_API_TYPE_SCHEDULER;
+	mes_reset_queue_pkt.header.opcode = MES_SCH_API_RESET;
+	mes_reset_queue_pkt.header.dwsize = API_FRAME_SIZE_IN_DWORDS;
+
+	mes_reset_queue_pkt.queue_type =
+		convert_to_mes_queue_type(input->queue_type);
+
+	if (input->legacy_gfx) {
+		mes_reset_queue_pkt.reset_legacy_gfx = 1;
+		mes_reset_queue_pkt.pipe_id_lp = input->pipe_id;
+		mes_reset_queue_pkt.queue_id_lp = input->queue_id;
+		mes_reset_queue_pkt.mqd_mc_addr_lp = input->mqd_addr;
+		mes_reset_queue_pkt.doorbell_offset_lp = input->doorbell_offset;
+		mes_reset_queue_pkt.wptr_addr_lp = input->wptr_addr;
+		mes_reset_queue_pkt.vmid_id_lp = input->vmid;
+	} else {
+		mes_reset_queue_pkt.reset_queue_only = 1;
+		mes_reset_queue_pkt.doorbell_offset = input->doorbell_offset;
+	}
+
+	return mes_v11_0_submit_pkt_and_poll_completion(mes,
+			&mes_reset_queue_pkt, sizeof(mes_reset_queue_pkt),
+			offsetof(union MESAPI__RESET, api_status));
+}
+
+static int mes_v11_0_detect_and_reset_hung_queues(struct amdgpu_mes *mes,
+						  struct mes_detect_and_reset_queue_input *input)
+{
+	union MESAPI__RESET mes_reset_queue_pkt;
+
+	memset(&mes_reset_queue_pkt, 0, sizeof(mes_reset_queue_pkt));
+
+	mes_reset_queue_pkt.header.type = MES_API_TYPE_SCHEDULER;
+	mes_reset_queue_pkt.header.opcode = MES_SCH_API_RESET;
+	mes_reset_queue_pkt.header.dwsize = API_FRAME_SIZE_IN_DWORDS;
+
+	mes_reset_queue_pkt.queue_type =
+		convert_to_mes_queue_type(input->queue_type);
+	mes_reset_queue_pkt.doorbell_offset_addr =
+		mes->hung_queue_db_array_gpu_addr;
+
+	if (input->detect_only)
+		mes_reset_queue_pkt.hang_detect_only = 1;
+	else
+		mes_reset_queue_pkt.hang_detect_then_reset = 1;
+
+	return mes_v11_0_submit_pkt_and_poll_completion(mes,
+			&mes_reset_queue_pkt, sizeof(mes_reset_queue_pkt),
+			offsetof(union MESAPI__RESET, api_status));
+}
+
+static const struct amdgpu_mes_funcs mes_v11_0_funcs = {
+	.add_hw_queue = mes_v11_0_add_hw_queue,
+	.remove_hw_queue = mes_v11_0_remove_hw_queue,
+	.map_legacy_queue = mes_v11_0_map_legacy_queue,
+	.unmap_legacy_queue = mes_v11_0_unmap_legacy_queue,
+	.suspend_gang = mes_v11_0_suspend_gang,
+	.resume_gang = mes_v11_0_resume_gang,
+	.misc_op = mes_v11_0_misc_op,
+	.reset_hw_queue = mes_v11_0_reset_hw_queue,
+	.detect_and_reset_hung_queues = mes_v11_0_detect_and_reset_hung_queues,
+};
+
+static int mes_v11_0_allocate_ucode_buffer(struct amdgpu_device *adev,
+					   enum amdgpu_mes_pipe pipe)
+{
+	int r;
+	const struct mes_firmware_header_v1_0 *mes_hdr;
+	const __le32 *fw_data;
+	unsigned fw_size;
+
+	mes_hdr = (const struct mes_firmware_header_v1_0 *)
+		adev->mes.fw[pipe]->data;
+
+	fw_data = (const __le32 *)(adev->mes.fw[pipe]->data +
+		   le32_to_cpu(mes_hdr->mes_ucode_offset_bytes));
+	fw_size = le32_to_cpu(mes_hdr->mes_ucode_size_bytes);
+
+	r = amdgpu_bo_create_reserved(adev, fw_size,
+				      PAGE_SIZE,
+				      AMDGPU_GEM_DOMAIN_VRAM |
+				      AMDGPU_GEM_DOMAIN_GTT,
+				      &adev->mes.ucode_fw_obj[pipe],
+				      &adev->mes.ucode_fw_gpu_addr[pipe],
+				      (void **)&adev->mes.ucode_fw_ptr[pipe]);
+	if (r) {
+		dev_err(adev->dev, "(%d) failed to create mes fw bo\n", r);
+		return r;
+	}
+
+	memcpy(adev->mes.ucode_fw_ptr[pipe], fw_data, fw_size);
+
+	amdgpu_bo_kunmap(adev->mes.ucode_fw_obj[pipe]);
+	amdgpu_bo_unreserve(adev->mes.ucode_fw_obj[pipe]);
+
+	return 0;
+}
+
+static int mes_v11_0_allocate_ucode_data_buffer(struct amdgpu_device *adev,
+						enum amdgpu_mes_pipe pipe)
+{
+	int r;
+	const struct mes_firmware_header_v1_0 *mes_hdr;
+	const __le32 *fw_data;
+	unsigned fw_size;
+
+	mes_hdr = (const struct mes_firmware_header_v1_0 *)
+		adev->mes.fw[pipe]->data;
+
+	fw_data = (const __le32 *)(adev->mes.fw[pipe]->data +
+		   le32_to_cpu(mes_hdr->mes_ucode_data_offset_bytes));
+	fw_size = le32_to_cpu(mes_hdr->mes_ucode_data_size_bytes);
+
+	if (fw_size > GFX_MES_DRAM_SIZE) {
+		dev_err(adev->dev, "PIPE%d ucode data fw size (%d) is greater than dram size (%d)\n",
+			pipe, fw_size, GFX_MES_DRAM_SIZE);
+		return -EINVAL;
+	}
+
+	r = amdgpu_bo_create_reserved(adev, GFX_MES_DRAM_SIZE,
+				      64 * 1024,
+				      AMDGPU_GEM_DOMAIN_VRAM |
+				      AMDGPU_GEM_DOMAIN_GTT,
+				      &adev->mes.data_fw_obj[pipe],
+				      &adev->mes.data_fw_gpu_addr[pipe],
+				      (void **)&adev->mes.data_fw_ptr[pipe]);
+	if (r) {
+		dev_err(adev->dev, "(%d) failed to create mes data fw bo\n", r);
+		return r;
+	}
+
+	memcpy(adev->mes.data_fw_ptr[pipe], fw_data, fw_size);
+
+	amdgpu_bo_kunmap(adev->mes.data_fw_obj[pipe]);
+	amdgpu_bo_unreserve(adev->mes.data_fw_obj[pipe]);
+
+	return 0;
+}
+
+static void mes_v11_0_free_ucode_buffers(struct amdgpu_device *adev,
+					 enum amdgpu_mes_pipe pipe)
+{
+	amdgpu_bo_free_kernel(&adev->mes.data_fw_obj[pipe],
+			      &adev->mes.data_fw_gpu_addr[pipe],
+			      (void **)&adev->mes.data_fw_ptr[pipe]);
+
+	amdgpu_bo_free_kernel(&adev->mes.ucode_fw_obj[pipe],
+			      &adev->mes.ucode_fw_gpu_addr[pipe],
+			      (void **)&adev->mes.ucode_fw_ptr[pipe]);
+}
+
+static void mes_v11_0_get_fw_version(struct amdgpu_device *adev)
+{
+	int pipe;
+
+	/* return early if we have already fetched these */
+	if (adev->mes.sched_version && adev->mes.kiq_version)
+		return;
+
+	/* get MES scheduler/KIQ versions */
+	mutex_lock(&adev->srbm_mutex);
+
+	for (pipe = 0; pipe < AMDGPU_MAX_MES_PIPES; pipe++) {
+		soc21_grbm_select(adev, 3, pipe, 0, 0);
+
+		if (pipe == AMDGPU_MES_SCHED_PIPE)
+			adev->mes.sched_version =
+				RREG32_SOC15(GC, 0, regCP_MES_GP3_LO);
+		else if (pipe == AMDGPU_MES_KIQ_PIPE && adev->enable_mes_kiq)
+			adev->mes.kiq_version =
+				RREG32_SOC15(GC, 0, regCP_MES_GP3_LO);
+	}
+
+	soc21_grbm_select(adev, 0, 0, 0, 0);
+	mutex_unlock(&adev->srbm_mutex);
+}
+
+static void mes_v11_0_enable(struct amdgpu_device *adev, bool enable)
+{
+	uint64_t ucode_addr;
+	uint32_t pipe, data = 0;
+
+	if (enable) {
+		if (amdgpu_mes_log_enable) {
+			WREG32_SOC15(GC, 0, regCP_MES_MSCRATCH_LO,
+				lower_32_bits(adev->mes.event_log_gpu_addr + AMDGPU_MES_LOG_BUFFER_SIZE));
+			WREG32_SOC15(GC, 0, regCP_MES_MSCRATCH_HI,
+				upper_32_bits(adev->mes.event_log_gpu_addr + AMDGPU_MES_LOG_BUFFER_SIZE));
+			dev_info(adev->dev, "Setup CP MES MSCRATCH address : 0x%x. 0x%x\n",
+				RREG32_SOC15(GC, 0, regCP_MES_MSCRATCH_HI),
+				RREG32_SOC15(GC, 0, regCP_MES_MSCRATCH_LO));
+		}
+
+		data = RREG32_SOC15(GC, 0, regCP_MES_CNTL);
+		data = REG_SET_FIELD(data, CP_MES_CNTL, MES_PIPE0_RESET, 1);
+		data = REG_SET_FIELD(data, CP_MES_CNTL,
+			     MES_PIPE1_RESET, adev->enable_mes_kiq ? 1 : 0);
+		WREG32_SOC15(GC, 0, regCP_MES_CNTL, data);
+
+		mutex_lock(&adev->srbm_mutex);
+		for (pipe = 0; pipe < AMDGPU_MAX_MES_PIPES; pipe++) {
+			if (!adev->enable_mes_kiq &&
+			    pipe == AMDGPU_MES_KIQ_PIPE)
+				continue;
+
+			soc21_grbm_select(adev, 3, pipe, 0, 0);
+
+			ucode_addr = adev->mes.uc_start_addr[pipe] >> 2;
+			WREG32_SOC15(GC, 0, regCP_MES_PRGRM_CNTR_START,
+				     lower_32_bits(ucode_addr));
+			WREG32_SOC15(GC, 0, regCP_MES_PRGRM_CNTR_START_HI,
+				     upper_32_bits(ucode_addr));
+		}
+		soc21_grbm_select(adev, 0, 0, 0, 0);
+		mutex_unlock(&adev->srbm_mutex);
+
+		/* unhalt MES and activate pipe0 */
+		data = REG_SET_FIELD(0, CP_MES_CNTL, MES_PIPE0_ACTIVE, 1);
+		data = REG_SET_FIELD(data, CP_MES_CNTL, MES_PIPE1_ACTIVE,
+				     adev->enable_mes_kiq ? 1 : 0);
+		WREG32_SOC15(GC, 0, regCP_MES_CNTL, data);
+
+		if (amdgpu_emu_mode)
+			msleep(100);
+		else
+			udelay(500);
+	} else {
+		data = RREG32_SOC15(GC, 0, regCP_MES_CNTL);
+		data = REG_SET_FIELD(data, CP_MES_CNTL, MES_PIPE0_ACTIVE, 0);
+		data = REG_SET_FIELD(data, CP_MES_CNTL, MES_PIPE1_ACTIVE, 0);
+		data = REG_SET_FIELD(data, CP_MES_CNTL,
+				     MES_INVALIDATE_ICACHE, 1);
+		data = REG_SET_FIELD(data, CP_MES_CNTL, MES_PIPE0_RESET, 1);
+		data = REG_SET_FIELD(data, CP_MES_CNTL, MES_PIPE1_RESET,
+				     adev->enable_mes_kiq ? 1 : 0);
+		data = REG_SET_FIELD(data, CP_MES_CNTL, MES_HALT, 1);
+		WREG32_SOC15(GC, 0, regCP_MES_CNTL, data);
+	}
+}
+
+/* This function is for backdoor MES firmware */
+static int mes_v11_0_load_microcode(struct amdgpu_device *adev,
+				    enum amdgpu_mes_pipe pipe, bool prime_icache)
+{
+	int r;
+	uint32_t data;
+	uint64_t ucode_addr;
+
+	mes_v11_0_enable(adev, false);
+
+	if (!adev->mes.fw[pipe])
+		return -EINVAL;
+
+	r = mes_v11_0_allocate_ucode_buffer(adev, pipe);
+	if (r)
+		return r;
+
+	r = mes_v11_0_allocate_ucode_data_buffer(adev, pipe);
+	if (r) {
+		mes_v11_0_free_ucode_buffers(adev, pipe);
+		return r;
+	}
+
+	mutex_lock(&adev->srbm_mutex);
+	/* me=3, pipe=0, queue=0 */
+	soc21_grbm_select(adev, 3, pipe, 0, 0);
+
+	WREG32_SOC15(GC, 0, regCP_MES_IC_BASE_CNTL, 0);
+
+	/* set ucode start address */
+	ucode_addr = adev->mes.uc_start_addr[pipe] >> 2;
+	WREG32_SOC15(GC, 0, regCP_MES_PRGRM_CNTR_START,
+		     lower_32_bits(ucode_addr));
+	WREG32_SOC15(GC, 0, regCP_MES_PRGRM_CNTR_START_HI,
+		     upper_32_bits(ucode_addr));
+
+	/* set ucode fimrware address */
+	WREG32_SOC15(GC, 0, regCP_MES_IC_BASE_LO,
+		     lower_32_bits(adev->mes.ucode_fw_gpu_addr[pipe]));
+	WREG32_SOC15(GC, 0, regCP_MES_IC_BASE_HI,
+		     upper_32_bits(adev->mes.ucode_fw_gpu_addr[pipe]));
+
+	/* set ucode instruction cache boundary to 2M-1 */
+	WREG32_SOC15(GC, 0, regCP_MES_MIBOUND_LO, 0x1FFFFF);
+
+	/* set ucode data firmware address */
+	WREG32_SOC15(GC, 0, regCP_MES_MDBASE_LO,
+		     lower_32_bits(adev->mes.data_fw_gpu_addr[pipe]));
+	WREG32_SOC15(GC, 0, regCP_MES_MDBASE_HI,
+		     upper_32_bits(adev->mes.data_fw_gpu_addr[pipe]));
+
+	/* Set 0x7FFFF (512K-1) to CP_MES_MDBOUND_LO */
+	WREG32_SOC15(GC, 0, regCP_MES_MDBOUND_LO, 0x7FFFF);
+
+	if (prime_icache) {
+		/* invalidate ICACHE */
+		data = RREG32_SOC15(GC, 0, regCP_MES_IC_OP_CNTL);
+		data = REG_SET_FIELD(data, CP_MES_IC_OP_CNTL, PRIME_ICACHE, 0);
+		data = REG_SET_FIELD(data, CP_MES_IC_OP_CNTL, INVALIDATE_CACHE, 1);
+		WREG32_SOC15(GC, 0, regCP_MES_IC_OP_CNTL, data);
+
+		/* prime the ICACHE. */
+		data = RREG32_SOC15(GC, 0, regCP_MES_IC_OP_CNTL);
+		data = REG_SET_FIELD(data, CP_MES_IC_OP_CNTL, PRIME_ICACHE, 1);
+		WREG32_SOC15(GC, 0, regCP_MES_IC_OP_CNTL, data);
+	}
+
+	soc21_grbm_select(adev, 0, 0, 0, 0);
+	mutex_unlock(&adev->srbm_mutex);
+
+	return 0;
+}
+
+static int mes_v11_0_allocate_eop_buf(struct amdgpu_device *adev,
+				      enum amdgpu_mes_pipe pipe)
+{
+	int r;
+	u32 *eop;
+
+	r = amdgpu_bo_create_reserved(adev, MES_EOP_SIZE, PAGE_SIZE,
+			      AMDGPU_GEM_DOMAIN_GTT,
+			      &adev->mes.eop_gpu_obj[pipe],
+			      &adev->mes.eop_gpu_addr[pipe],
+			      (void **)&eop);
+	if (r) {
+		dev_warn(adev->dev, "(%d) create EOP bo failed\n", r);
+		return r;
+	}
+
+	memset(eop, 0,
+	       adev->mes.eop_gpu_obj[pipe]->tbo.base.size);
+
+	amdgpu_bo_kunmap(adev->mes.eop_gpu_obj[pipe]);
+	amdgpu_bo_unreserve(adev->mes.eop_gpu_obj[pipe]);
+
+	return 0;
+}
+
+static int mes_v11_0_mqd_init(struct amdgpu_ring *ring)
+{
+	struct v11_compute_mqd *mqd = ring->mqd_ptr;
+	uint64_t hqd_gpu_addr, wb_gpu_addr, eop_base_addr;
+	uint32_t tmp;
+
+	memset(mqd, 0, sizeof(*mqd));
+
+	mqd->header = 0xC0310800;
+	mqd->compute_pipelinestat_enable = 0x00000001;
+	mqd->compute_static_thread_mgmt_se0 = 0xffffffff;
+	mqd->compute_static_thread_mgmt_se1 = 0xffffffff;
+	mqd->compute_static_thread_mgmt_se2 = 0xffffffff;
+	mqd->compute_static_thread_mgmt_se3 = 0xffffffff;
+	mqd->compute_misc_reserved = 0x00000007;
+
+	eop_base_addr = ring->eop_gpu_addr >> 8;
+
+	/* set the EOP size, register value is 2^(EOP_SIZE+1) dwords */
+	tmp = regCP_HQD_EOP_CONTROL_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, CP_HQD_EOP_CONTROL, EOP_SIZE,
+			(order_base_2(MES_EOP_SIZE / 4) - 1));
+
+	mqd->cp_hqd_eop_base_addr_lo = lower_32_bits(eop_base_addr);
+	mqd->cp_hqd_eop_base_addr_hi = upper_32_bits(eop_base_addr);
+	mqd->cp_hqd_eop_control = tmp;
+
+	/* disable the queue if it's active */
+	ring->wptr = 0;
+	mqd->cp_hqd_pq_rptr = 0;
+	mqd->cp_hqd_pq_wptr_lo = 0;
+	mqd->cp_hqd_pq_wptr_hi = 0;
+
+	/* set the pointer to the MQD */
+	mqd->cp_mqd_base_addr_lo = ring->mqd_gpu_addr & 0xfffffffc;
+	mqd->cp_mqd_base_addr_hi = upper_32_bits(ring->mqd_gpu_addr);
+
+	/* set MQD vmid to 0 */
+	tmp = regCP_MQD_CONTROL_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, CP_MQD_CONTROL, VMID, 0);
+	mqd->cp_mqd_control = tmp;
+
+	/* set the pointer to the HQD, this is similar CP_RB0_BASE/_HI */
+	hqd_gpu_addr = ring->gpu_addr >> 8;
+	mqd->cp_hqd_pq_base_lo = lower_32_bits(hqd_gpu_addr);
+	mqd->cp_hqd_pq_base_hi = upper_32_bits(hqd_gpu_addr);
+
+	/* set the wb address whether it's enabled or not */
+	wb_gpu_addr = ring->rptr_gpu_addr;
+	mqd->cp_hqd_pq_rptr_report_addr_lo = wb_gpu_addr & 0xfffffffc;
+	mqd->cp_hqd_pq_rptr_report_addr_hi =
+		upper_32_bits(wb_gpu_addr) & 0xffff;
+
+	/* only used if CP_PQ_WPTR_POLL_CNTL.CP_PQ_WPTR_POLL_CNTL__EN_MASK=1 */
+	wb_gpu_addr = ring->wptr_gpu_addr;
+	mqd->cp_hqd_pq_wptr_poll_addr_lo = wb_gpu_addr & 0xfffffff8;
+	mqd->cp_hqd_pq_wptr_poll_addr_hi = upper_32_bits(wb_gpu_addr) & 0xffff;
+
+	/* set up the HQD, this is similar to CP_RB0_CNTL */
+	tmp = regCP_HQD_PQ_CONTROL_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, QUEUE_SIZE,
+			    (order_base_2(ring->ring_size / 4) - 1));
+	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, RPTR_BLOCK_SIZE,
+			    ((order_base_2(AMDGPU_GPU_PAGE_SIZE / 4) - 1) << 8));
+	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, UNORD_DISPATCH, 1);
+	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, TUNNEL_DISPATCH, 0);
+	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, PRIV_STATE, 1);
+	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, KMD_QUEUE, 1);
+	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, NO_UPDATE_RPTR, 1);
+	mqd->cp_hqd_pq_control = tmp;
+
+	/* enable doorbell */
+	tmp = 0;
+	if (ring->use_doorbell) {
+		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
+				    DOORBELL_OFFSET, ring->doorbell_index);
+		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
+				    DOORBELL_EN, 1);
+		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
+				    DOORBELL_SOURCE, 0);
+		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
+				    DOORBELL_HIT, 0);
+	} else
+		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
+				    DOORBELL_EN, 0);
+	mqd->cp_hqd_pq_doorbell_control = tmp;
+
+	mqd->cp_hqd_vmid = 0;
+	/* activate the queue */
+	mqd->cp_hqd_active = 1;
+
+	tmp = regCP_HQD_PERSISTENT_STATE_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, CP_HQD_PERSISTENT_STATE,
+			    PRELOAD_SIZE, 0x55);
+	mqd->cp_hqd_persistent_state = tmp;
+
+	mqd->cp_hqd_ib_control = regCP_HQD_IB_CONTROL_DEFAULT;
+	mqd->cp_hqd_iq_timer = regCP_HQD_IQ_TIMER_DEFAULT;
+	mqd->cp_hqd_quantum = regCP_HQD_QUANTUM_DEFAULT;
+
+	amdgpu_device_flush_hdp(ring->adev, NULL);
+	return 0;
+}
+
+static void mes_v11_0_queue_init_register(struct amdgpu_ring *ring)
+{
+	struct v11_compute_mqd *mqd = ring->mqd_ptr;
+	struct amdgpu_device *adev = ring->adev;
+	uint32_t data = 0;
+
+	mutex_lock(&adev->srbm_mutex);
+	soc21_grbm_select(adev, 3, ring->pipe, 0, 0);
+
+	/* set CP_HQD_VMID.VMID = 0. */
+	data = RREG32_SOC15(GC, 0, regCP_HQD_VMID);
+	data = REG_SET_FIELD(data, CP_HQD_VMID, VMID, 0);
+	WREG32_SOC15(GC, 0, regCP_HQD_VMID, data);
+
+	/* set CP_HQD_PQ_DOORBELL_CONTROL.DOORBELL_EN=0 */
+	data = RREG32_SOC15(GC, 0, regCP_HQD_PQ_DOORBELL_CONTROL);
+	data = REG_SET_FIELD(data, CP_HQD_PQ_DOORBELL_CONTROL,
+			     DOORBELL_EN, 0);
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_DOORBELL_CONTROL, data);
+
+	/* set CP_MQD_BASE_ADDR/HI with the MQD base address */
+	WREG32_SOC15(GC, 0, regCP_MQD_BASE_ADDR, mqd->cp_mqd_base_addr_lo);
+	WREG32_SOC15(GC, 0, regCP_MQD_BASE_ADDR_HI, mqd->cp_mqd_base_addr_hi);
+
+	/* set CP_MQD_CONTROL.VMID=0 */
+	data = RREG32_SOC15(GC, 0, regCP_MQD_CONTROL);
+	data = REG_SET_FIELD(data, CP_MQD_CONTROL, VMID, 0);
+	WREG32_SOC15(GC, 0, regCP_MQD_CONTROL, 0);
+
+	/* set CP_HQD_PQ_BASE/HI with the ring buffer base address */
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_BASE, mqd->cp_hqd_pq_base_lo);
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_BASE_HI, mqd->cp_hqd_pq_base_hi);
+
+	/* set CP_HQD_PQ_RPTR_REPORT_ADDR/HI */
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_RPTR_REPORT_ADDR,
+		     mqd->cp_hqd_pq_rptr_report_addr_lo);
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_RPTR_REPORT_ADDR_HI,
+		     mqd->cp_hqd_pq_rptr_report_addr_hi);
+
+	/* set CP_HQD_PQ_CONTROL */
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_CONTROL, mqd->cp_hqd_pq_control);
+
+	/* set CP_HQD_PQ_WPTR_POLL_ADDR/HI */
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_WPTR_POLL_ADDR,
+		     mqd->cp_hqd_pq_wptr_poll_addr_lo);
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_WPTR_POLL_ADDR_HI,
+		     mqd->cp_hqd_pq_wptr_poll_addr_hi);
+
+	/* set CP_HQD_PQ_DOORBELL_CONTROL */
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_DOORBELL_CONTROL,
+		     mqd->cp_hqd_pq_doorbell_control);
+
+	/* set CP_HQD_PERSISTENT_STATE.PRELOAD_SIZE=0x53 */
+	WREG32_SOC15(GC, 0, regCP_HQD_PERSISTENT_STATE, mqd->cp_hqd_persistent_state);
+
+	/* set CP_HQD_ACTIVE.ACTIVE=1 */
+	WREG32_SOC15(GC, 0, regCP_HQD_ACTIVE, mqd->cp_hqd_active);
+
+	soc21_grbm_select(adev, 0, 0, 0, 0);
+	mutex_unlock(&adev->srbm_mutex);
+}
+
+static int mes_v11_0_kiq_enable_queue(struct amdgpu_device *adev)
+{
+	struct amdgpu_kiq *kiq = &adev->gfx.kiq[0];
+	struct amdgpu_ring *kiq_ring = &adev->gfx.kiq[0].ring;
+	int r;
+
+	if (!kiq->pmf || !kiq->pmf->kiq_map_queues)
+		return -EINVAL;
+
+	r = amdgpu_ring_alloc(kiq_ring, kiq->pmf->map_queues_size);
+	if (r) {
+		DRM_ERROR("Failed to lock KIQ (%d).\n", r);
+		return r;
+	}
+
+	kiq->pmf->kiq_map_queues(kiq_ring, &adev->mes.ring[0]);
+
+	return amdgpu_ring_test_helper(kiq_ring);
+}
+
+static int mes_v11_0_queue_init(struct amdgpu_device *adev,
+				enum amdgpu_mes_pipe pipe)
+{
+	struct amdgpu_ring *ring;
+	int r;
+
+	if (pipe == AMDGPU_MES_KIQ_PIPE)
+		ring = &adev->gfx.kiq[0].ring;
+	else if (pipe == AMDGPU_MES_SCHED_PIPE)
+		ring = &adev->mes.ring[0];
+	else
+		BUG();
+
+	if ((pipe == AMDGPU_MES_SCHED_PIPE) &&
+	    (amdgpu_in_reset(adev) || adev->in_suspend)) {
+		*(ring->wptr_cpu_addr) = 0;
+		*(ring->rptr_cpu_addr) = 0;
+		amdgpu_ring_clear_ring(ring);
+	}
+
+	r = mes_v11_0_mqd_init(ring);
+	if (r)
+		return r;
+
+	if (pipe == AMDGPU_MES_SCHED_PIPE) {
+		r = mes_v11_0_kiq_enable_queue(adev);
+		if (r)
+			return r;
+	} else {
+		mes_v11_0_queue_init_register(ring);
+	}
+
+	return 0;
+}
+
+static int mes_v11_0_ring_init(struct amdgpu_device *adev)
+{
+	struct amdgpu_ring *ring;
+
+	ring = &adev->mes.ring[0];
+
+	ring->funcs = &mes_v11_0_ring_funcs;
+
+	ring->me = 3;
+	ring->pipe = 0;
+	ring->queue = 0;
+
+	ring->ring_obj = NULL;
+	ring->use_doorbell = true;
+	ring->doorbell_index = adev->doorbell_index.mes_ring0 << 1;
+	ring->eop_gpu_addr = adev->mes.eop_gpu_addr[AMDGPU_MES_SCHED_PIPE];
+	ring->no_scheduler = true;
+	sprintf(ring->name, "mes_%d.%d.%d", ring->me, ring->pipe, ring->queue);
+
+	return amdgpu_ring_init(adev, ring, 1024, NULL, 0,
+				AMDGPU_RING_PRIO_DEFAULT, NULL);
+}
+
+static int mes_v11_0_kiq_ring_init(struct amdgpu_device *adev)
+{
+	struct amdgpu_ring *ring;
+
+	spin_lock_init(&adev->gfx.kiq[0].ring_lock);
+
+	ring = &adev->gfx.kiq[0].ring;
+
+	ring->me = 3;
+	ring->pipe = 1;
+	ring->queue = 0;
+
+	ring->adev = NULL;
+	ring->ring_obj = NULL;
+	ring->use_doorbell = true;
+	ring->doorbell_index = adev->doorbell_index.mes_ring1 << 1;
+	ring->eop_gpu_addr = adev->mes.eop_gpu_addr[AMDGPU_MES_KIQ_PIPE];
+	ring->no_scheduler = true;
+	sprintf(ring->name, "mes_kiq_%d.%d.%d",
+		ring->me, ring->pipe, ring->queue);
+
+	return amdgpu_ring_init(adev, ring, 1024, NULL, 0,
+				AMDGPU_RING_PRIO_DEFAULT, NULL);
+}
+
+static int mes_v11_0_mqd_sw_init(struct amdgpu_device *adev,
+				 enum amdgpu_mes_pipe pipe)
+{
+	int r, mqd_size = sizeof(struct v11_compute_mqd);
+	struct amdgpu_ring *ring;
+
+	if (pipe == AMDGPU_MES_KIQ_PIPE)
+		ring = &adev->gfx.kiq[0].ring;
+	else if (pipe == AMDGPU_MES_SCHED_PIPE)
+		ring = &adev->mes.ring[0];
+	else
+		BUG();
+
+	if (ring->mqd_obj)
+		return 0;
+
+	r = amdgpu_bo_create_kernel(adev, mqd_size, PAGE_SIZE,
+				    AMDGPU_GEM_DOMAIN_VRAM |
+				    AMDGPU_GEM_DOMAIN_GTT, &ring->mqd_obj,
+				    &ring->mqd_gpu_addr, &ring->mqd_ptr);
+	if (r) {
+		dev_warn(adev->dev, "failed to create ring mqd bo (%d)", r);
+		return r;
+	}
+
+	memset(ring->mqd_ptr, 0, mqd_size);
+
+	/* prepare MQD backup */
+	adev->mes.mqd_backup[pipe] = kmalloc(mqd_size, GFP_KERNEL);
+	if (!adev->mes.mqd_backup[pipe]) {
+		dev_warn(adev->dev,
+			 "no memory to create MQD backup for ring %s\n",
+			 ring->name);
+		return -ENOMEM;
+	}
+
+	return 0;
+}
+
+static int mes_v11_0_sw_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int pipe, r, bo_size;
+
+	adev->mes.funcs = &mes_v11_0_funcs;
+	adev->mes.kiq_hw_init = &mes_v11_0_kiq_hw_init;
+	adev->mes.kiq_hw_fini = &mes_v11_0_kiq_hw_fini;
+
+	adev->mes.event_log_size = AMDGPU_MES_LOG_BUFFER_SIZE + AMDGPU_MES_MSCRATCH_SIZE;
+
+	r = amdgpu_mes_init(adev);
+	if (r)
+		return r;
+
+	for (pipe = 0; pipe < AMDGPU_MAX_MES_PIPES; pipe++) {
+		if (!adev->enable_mes_kiq && pipe == AMDGPU_MES_KIQ_PIPE)
+			continue;
+
+		r = mes_v11_0_allocate_eop_buf(adev, pipe);
+		if (r)
+			return r;
+
+		r = mes_v11_0_mqd_sw_init(adev, pipe);
+		if (r)
+			return r;
+	}
+
+	if (adev->enable_mes_kiq) {
+		r = mes_v11_0_kiq_ring_init(adev);
+		if (r)
+			return r;
+	}
+
+	r = mes_v11_0_ring_init(adev);
+	if (r)
+		return r;
+
+	bo_size = AMDGPU_GPU_PAGE_SIZE;
+	if (amdgpu_sriov_is_mes_info_enable(adev))
+		bo_size += MES11_HW_RESOURCE_1_SIZE;
+
+	/* Only needed for AMDGPU_MES_SCHED_PIPE on MES 11*/
+	r = amdgpu_bo_create_kernel(adev,
+				    bo_size,
+				    PAGE_SIZE,
+				    AMDGPU_GEM_DOMAIN_VRAM,
+				    &adev->mes.resource_1[0],
+				    &adev->mes.resource_1_gpu_addr[0],
+				    &adev->mes.resource_1_addr[0]);
+	if (r) {
+		dev_err(adev->dev, "(%d) failed to create mes resource_1 bo\n", r);
+		return r;
+	}
+
+	return 0;
+}
+
+static int mes_v11_0_sw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int pipe;
+
+	amdgpu_bo_free_kernel(&adev->mes.resource_1[0], &adev->mes.resource_1_gpu_addr[0],
+			      &adev->mes.resource_1_addr[0]);
+
+	for (pipe = 0; pipe < AMDGPU_MAX_MES_PIPES; pipe++) {
+		kfree(adev->mes.mqd_backup[pipe]);
+
+		amdgpu_bo_free_kernel(&adev->mes.eop_gpu_obj[pipe],
+				      &adev->mes.eop_gpu_addr[pipe],
+				      NULL);
+		amdgpu_ucode_release(&adev->mes.fw[pipe]);
+	}
+
+	amdgpu_bo_free_kernel(&adev->gfx.kiq[0].ring.mqd_obj,
+			      &adev->gfx.kiq[0].ring.mqd_gpu_addr,
+			      &adev->gfx.kiq[0].ring.mqd_ptr);
+
+	amdgpu_bo_free_kernel(&adev->mes.ring[0].mqd_obj,
+			      &adev->mes.ring[0].mqd_gpu_addr,
+			      &adev->mes.ring[0].mqd_ptr);
+
+	amdgpu_ring_fini(&adev->gfx.kiq[0].ring);
+	amdgpu_ring_fini(&adev->mes.ring[0]);
+
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT) {
+		mes_v11_0_free_ucode_buffers(adev, AMDGPU_MES_KIQ_PIPE);
+		mes_v11_0_free_ucode_buffers(adev, AMDGPU_MES_SCHED_PIPE);
+	}
+
+	amdgpu_mes_fini(adev);
+	return 0;
+}
+
+static void mes_v11_0_kiq_dequeue(struct amdgpu_ring *ring)
+{
+	uint32_t data;
+	int i;
+	struct amdgpu_device *adev = ring->adev;
+
+	mutex_lock(&adev->srbm_mutex);
+	soc21_grbm_select(adev, 3, ring->pipe, 0, 0);
+
+	/* disable the queue if it's active */
+	if (RREG32_SOC15(GC, 0, regCP_HQD_ACTIVE) & 1) {
+		WREG32_SOC15(GC, 0, regCP_HQD_DEQUEUE_REQUEST, 1);
+		for (i = 0; i < adev->usec_timeout; i++) {
+			if (!(RREG32_SOC15(GC, 0, regCP_HQD_ACTIVE) & 1))
+				break;
+			udelay(1);
+		}
+	}
+	data = RREG32_SOC15(GC, 0, regCP_HQD_PQ_DOORBELL_CONTROL);
+	data = REG_SET_FIELD(data, CP_HQD_PQ_DOORBELL_CONTROL,
+				DOORBELL_EN, 0);
+	data = REG_SET_FIELD(data, CP_HQD_PQ_DOORBELL_CONTROL,
+				DOORBELL_HIT, 1);
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_DOORBELL_CONTROL, data);
+
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_DOORBELL_CONTROL, 0);
+
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_WPTR_LO, 0);
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_WPTR_HI, 0);
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_RPTR, 0);
+
+	soc21_grbm_select(adev, 0, 0, 0, 0);
+	mutex_unlock(&adev->srbm_mutex);
+}
+
+static void mes_v11_0_kiq_setting(struct amdgpu_ring *ring)
+{
+	uint32_t tmp;
+	struct amdgpu_device *adev = ring->adev;
+
+	/* tell RLC which is KIQ queue */
+	tmp = RREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS);
+	tmp &= 0xffffff00;
+	tmp |= (ring->me << 5) | (ring->pipe << 3) | (ring->queue);
+	WREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS, tmp | 0x80);
+}
+
+static void mes_v11_0_kiq_clear(struct amdgpu_device *adev)
+{
+	uint32_t tmp;
+
+	/* tell RLC which is KIQ dequeue */
+	tmp = RREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS);
+	tmp &= ~RLC_CP_SCHEDULERS__scheduler0_MASK;
+	WREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS, tmp);
+}
+
+static int mes_v11_0_kiq_hw_init(struct amdgpu_device *adev)
+{
+	int r = 0;
+	struct amdgpu_ip_block *ip_block;
+
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT) {
+
+		r = mes_v11_0_load_microcode(adev, AMDGPU_MES_SCHED_PIPE, false);
+		if (r) {
+			DRM_ERROR("failed to load MES fw, r=%d\n", r);
+			return r;
+		}
+
+		r = mes_v11_0_load_microcode(adev, AMDGPU_MES_KIQ_PIPE, true);
+		if (r) {
+			DRM_ERROR("failed to load MES kiq fw, r=%d\n", r);
+			return r;
+		}
+
+	}
+
+	mes_v11_0_enable(adev, true);
+
+	mes_v11_0_get_fw_version(adev);
+
+	mes_v11_0_kiq_setting(&adev->gfx.kiq[0].ring);
+
+	ip_block = amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_MES);
+	if (unlikely(!ip_block)) {
+		dev_err(adev->dev, "Failed to get MES handle\n");
+		return -EINVAL;
+	}
+
+	r = mes_v11_0_queue_init(adev, AMDGPU_MES_KIQ_PIPE);
+	if (r)
+		goto failure;
+
+	if ((adev->mes.sched_version & AMDGPU_MES_VERSION_MASK) >= 0x47)
+		adev->mes.enable_legacy_queue_map = true;
+	else
+		adev->mes.enable_legacy_queue_map = false;
+
+	if (adev->mes.enable_legacy_queue_map) {
+		r = mes_v11_0_hw_init(ip_block);
+		if (r)
+			goto failure;
+	}
+
+	return r;
+
+failure:
+	mes_v11_0_hw_fini(ip_block);
+	return r;
+}
+
+static int mes_v11_0_kiq_hw_fini(struct amdgpu_device *adev)
+{
+	if (adev->mes.ring[0].sched.ready) {
+		mes_v11_0_kiq_dequeue(&adev->mes.ring[0]);
+		adev->mes.ring[0].sched.ready = false;
+	}
+
+	if (amdgpu_sriov_vf(adev)) {
+		mes_v11_0_kiq_dequeue(&adev->gfx.kiq[0].ring);
+		mes_v11_0_kiq_clear(adev);
+	}
+
+	mes_v11_0_enable(adev, false);
+
+	return 0;
+}
+
+static int mes_v11_0_hw_init(struct amdgpu_ip_block *ip_block)
+{
+	int r;
+	struct amdgpu_device *adev = ip_block->adev;
+
+	if (adev->mes.ring[0].sched.ready)
+		goto out;
+
+	if (!adev->enable_mes_kiq) {
+		if (adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT) {
+			r = mes_v11_0_load_microcode(adev,
+					     AMDGPU_MES_SCHED_PIPE, true);
+			if (r) {
+				DRM_ERROR("failed to MES fw, r=%d\n", r);
+				return r;
+			}
+		}
+
+		mes_v11_0_enable(adev, true);
+	}
+
+	r = mes_v11_0_queue_init(adev, AMDGPU_MES_SCHED_PIPE);
+	if (r)
+		goto failure;
+
+	r = mes_v11_0_set_hw_resources(&adev->mes);
+	if (r)
+		goto failure;
+
+	if ((adev->mes.sched_version & AMDGPU_MES_VERSION_MASK) >= 0x50) {
+		r = mes_v11_0_set_hw_resources_1(&adev->mes);
+		if (r) {
+			DRM_ERROR("failed mes_v11_0_set_hw_resources_1, r=%d\n", r);
+			goto failure;
+		}
+	}
+
+	r = mes_v11_0_query_sched_status(&adev->mes);
+	if (r) {
+		DRM_ERROR("MES is busy\n");
+		goto failure;
+	}
+
+	r = amdgpu_mes_update_enforce_isolation(adev);
+	if (r)
+		goto failure;
+
+out:
+	/*
+	 * Disable KIQ ring usage from the driver once MES is enabled.
+	 * MES uses KIQ ring exclusively so driver cannot access KIQ ring
+	 * with MES enabled.
+	 */
+	adev->gfx.kiq[0].ring.sched.ready = false;
+	adev->mes.ring[0].sched.ready = true;
+
+	return 0;
+
+failure:
+	mes_v11_0_hw_fini(ip_block);
+	return r;
+}
+
+static int mes_v11_0_hw_fini(struct amdgpu_ip_block *ip_block)
+{
+	return 0;
+}
+
+static int mes_v11_0_suspend(struct amdgpu_ip_block *ip_block)
+{
+	return mes_v11_0_hw_fini(ip_block);
+}
+
+static int mes_v11_0_resume(struct amdgpu_ip_block *ip_block)
+{
+	return mes_v11_0_hw_init(ip_block);
+}
+
+static int mes_v11_0_early_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int pipe, r;
+
+	adev->mes.hung_queue_db_array_size = MES11_HUNG_DB_OFFSET_ARRAY_SIZE;
+	adev->mes.hung_queue_hqd_info_offset = MES11_HUNG_HQD_INFO_OFFSET;
+
+	for (pipe = 0; pipe < AMDGPU_MAX_MES_PIPES; pipe++) {
+		if (!adev->enable_mes_kiq && pipe == AMDGPU_MES_KIQ_PIPE)
+			continue;
+		r = amdgpu_mes_init_microcode(adev, pipe);
+		if (r)
+			return r;
+	}
+
+	return 0;
+}
+
+static const struct amd_ip_funcs mes_v11_0_ip_funcs = {
+	.name = "mes_v11_0",
+	.early_init = mes_v11_0_early_init,
+	.late_init = NULL,
+	.sw_init = mes_v11_0_sw_init,
+	.sw_fini = mes_v11_0_sw_fini,
+	.hw_init = mes_v11_0_hw_init,
+	.hw_fini = mes_v11_0_hw_fini,
+	.suspend = mes_v11_0_suspend,
+	.resume = mes_v11_0_resume,
+};
+
+const struct amdgpu_ip_block_version mes_v11_0_ip_block = {
+	.type = AMD_IP_BLOCK_TYPE_MES,
+	.major = 11,
+	.minor = 0,
+	.rev = 0,
+	.funcs = &mes_v11_0_ip_funcs,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/mes_v11_0.h b/drivers/gpu/drm/amd/amdgpu/mes_v11_0.h
new file mode 100644
index 000000000000..b3519e1df2b2
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/mes_v11_0.h
@@ -0,0 +1,29 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __MES_V11_0_H__
+#define __MES_V11_0_H__
+
+extern const struct amdgpu_ip_block_version mes_v11_0_ip_block;
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/mes_v12_0.c b/drivers/gpu/drm/amd/amdgpu/mes_v12_0.c
new file mode 100644
index 000000000000..744e95d3984a
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/mes_v12_0.c
@@ -0,0 +1,1942 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/firmware.h>
+#include <linux/module.h>
+#include "amdgpu.h"
+#include "gfx_v12_0.h"
+#include "soc15_common.h"
+#include "soc21.h"
+#include "gc/gc_12_0_0_offset.h"
+#include "gc/gc_12_0_0_sh_mask.h"
+#include "gc/gc_11_0_0_default.h"
+#include "v12_structs.h"
+#include "mes_v12_api_def.h"
+
+MODULE_FIRMWARE("amdgpu/gc_12_0_0_mes.bin");
+MODULE_FIRMWARE("amdgpu/gc_12_0_0_mes1.bin");
+MODULE_FIRMWARE("amdgpu/gc_12_0_0_uni_mes.bin");
+MODULE_FIRMWARE("amdgpu/gc_12_0_1_mes.bin");
+MODULE_FIRMWARE("amdgpu/gc_12_0_1_mes1.bin");
+MODULE_FIRMWARE("amdgpu/gc_12_0_1_uni_mes.bin");
+
+static int mes_v12_0_hw_init(struct amdgpu_ip_block *ip_block);
+static int mes_v12_0_hw_fini(struct amdgpu_ip_block *ip_block);
+static int mes_v12_0_kiq_hw_init(struct amdgpu_device *adev);
+static int mes_v12_0_kiq_hw_fini(struct amdgpu_device *adev);
+
+#define MES_EOP_SIZE   2048
+
+#define MES12_HUNG_DB_OFFSET_ARRAY_SIZE 8 /* [0:3] = db offset [4:7] hqd info */
+#define MES12_HUNG_HQD_INFO_OFFSET	4
+
+static void mes_v12_0_ring_set_wptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	if (ring->use_doorbell) {
+		atomic64_set((atomic64_t *)ring->wptr_cpu_addr,
+			     ring->wptr);
+		WDOORBELL64(ring->doorbell_index, ring->wptr);
+	} else {
+		BUG();
+	}
+}
+
+static u64 mes_v12_0_ring_get_rptr(struct amdgpu_ring *ring)
+{
+	return *ring->rptr_cpu_addr;
+}
+
+static u64 mes_v12_0_ring_get_wptr(struct amdgpu_ring *ring)
+{
+	u64 wptr;
+
+	if (ring->use_doorbell)
+		wptr = atomic64_read((atomic64_t *)ring->wptr_cpu_addr);
+	else
+		BUG();
+	return wptr;
+}
+
+static const struct amdgpu_ring_funcs mes_v12_0_ring_funcs = {
+	.type = AMDGPU_RING_TYPE_MES,
+	.align_mask = 1,
+	.nop = 0,
+	.support_64bit_ptrs = true,
+	.get_rptr = mes_v12_0_ring_get_rptr,
+	.get_wptr = mes_v12_0_ring_get_wptr,
+	.set_wptr = mes_v12_0_ring_set_wptr,
+	.insert_nop = amdgpu_ring_insert_nop,
+};
+
+static const char *mes_v12_0_opcodes[] = {
+	"SET_HW_RSRC",
+	"SET_SCHEDULING_CONFIG",
+	"ADD_QUEUE",
+	"REMOVE_QUEUE",
+	"PERFORM_YIELD",
+	"SET_GANG_PRIORITY_LEVEL",
+	"SUSPEND",
+	"RESUME",
+	"RESET",
+	"SET_LOG_BUFFER",
+	"CHANGE_GANG_PRORITY",
+	"QUERY_SCHEDULER_STATUS",
+	"unused",
+	"SET_DEBUG_VMID",
+	"MISC",
+	"UPDATE_ROOT_PAGE_TABLE",
+	"AMD_LOG",
+	"SET_SE_MODE",
+	"SET_GANG_SUBMIT",
+	"SET_HW_RSRC_1",
+	"INVALIDATE_TLBS",
+};
+
+static const char *mes_v12_0_misc_opcodes[] = {
+	"WRITE_REG",
+	"INV_GART",
+	"QUERY_STATUS",
+	"READ_REG",
+	"WAIT_REG_MEM",
+	"SET_SHADER_DEBUGGER",
+	"NOTIFY_WORK_ON_UNMAPPED_QUEUE",
+	"NOTIFY_TO_UNMAP_PROCESSES",
+};
+
+static const char *mes_v12_0_get_op_string(union MESAPI__MISC *x_pkt)
+{
+	const char *op_str = NULL;
+
+	if (x_pkt->header.opcode < ARRAY_SIZE(mes_v12_0_opcodes))
+		op_str = mes_v12_0_opcodes[x_pkt->header.opcode];
+
+	return op_str;
+}
+
+static const char *mes_v12_0_get_misc_op_string(union MESAPI__MISC *x_pkt)
+{
+	const char *op_str = NULL;
+
+	if ((x_pkt->header.opcode == MES_SCH_API_MISC) &&
+	    (x_pkt->opcode < ARRAY_SIZE(mes_v12_0_misc_opcodes)))
+		op_str = mes_v12_0_misc_opcodes[x_pkt->opcode];
+
+	return op_str;
+}
+
+static int mes_v12_0_submit_pkt_and_poll_completion(struct amdgpu_mes *mes,
+					    int pipe, void *pkt, int size,
+					    int api_status_off)
+{
+	union MESAPI__QUERY_MES_STATUS mes_status_pkt;
+	signed long timeout = 2100000; /* 2100 ms */
+	struct amdgpu_device *adev = mes->adev;
+	struct amdgpu_ring *ring = &mes->ring[pipe];
+	spinlock_t *ring_lock = &mes->ring_lock[pipe];
+	struct MES_API_STATUS *api_status;
+	union MESAPI__MISC *x_pkt = pkt;
+	const char *op_str, *misc_op_str;
+	unsigned long flags;
+	u64 status_gpu_addr;
+	u32 seq, status_offset;
+	u64 *status_ptr;
+	signed long r;
+	int ret;
+
+	if (x_pkt->header.opcode >= MES_SCH_API_MAX)
+		return -EINVAL;
+
+	if (amdgpu_emu_mode) {
+		timeout *= 100;
+	} else if (amdgpu_sriov_vf(adev)) {
+		/* Worst case in sriov where all other 15 VF timeout, each VF needs about 600ms */
+		timeout = 15 * 600 * 1000;
+	}
+
+	ret = amdgpu_device_wb_get(adev, &status_offset);
+	if (ret)
+		return ret;
+
+	status_gpu_addr = adev->wb.gpu_addr + (status_offset * 4);
+	status_ptr = (u64 *)&adev->wb.wb[status_offset];
+	*status_ptr = 0;
+
+	spin_lock_irqsave(ring_lock, flags);
+	r = amdgpu_ring_alloc(ring, (size + sizeof(mes_status_pkt)) / 4);
+	if (r)
+		goto error_unlock_free;
+
+	seq = ++ring->fence_drv.sync_seq;
+	r = amdgpu_fence_wait_polling(ring,
+				      seq - ring->fence_drv.num_fences_mask,
+				      timeout);
+	if (r < 1)
+		goto error_undo;
+
+	api_status = (struct MES_API_STATUS *)((char *)pkt + api_status_off);
+	api_status->api_completion_fence_addr = status_gpu_addr;
+	api_status->api_completion_fence_value = 1;
+
+	amdgpu_ring_write_multiple(ring, pkt, size / 4);
+
+	memset(&mes_status_pkt, 0, sizeof(mes_status_pkt));
+	mes_status_pkt.header.type = MES_API_TYPE_SCHEDULER;
+	mes_status_pkt.header.opcode = MES_SCH_API_QUERY_SCHEDULER_STATUS;
+	mes_status_pkt.header.dwsize = API_FRAME_SIZE_IN_DWORDS;
+	mes_status_pkt.api_status.api_completion_fence_addr =
+		ring->fence_drv.gpu_addr;
+	mes_status_pkt.api_status.api_completion_fence_value = seq;
+
+	amdgpu_ring_write_multiple(ring, &mes_status_pkt,
+				   sizeof(mes_status_pkt) / 4);
+
+	amdgpu_ring_commit(ring);
+	spin_unlock_irqrestore(ring_lock, flags);
+
+	op_str = mes_v12_0_get_op_string(x_pkt);
+	misc_op_str = mes_v12_0_get_misc_op_string(x_pkt);
+
+	if (misc_op_str)
+		dev_dbg(adev->dev, "MES(%d) msg=%s (%s) was emitted\n",
+			pipe, op_str, misc_op_str);
+	else if (op_str)
+		dev_dbg(adev->dev, "MES(%d) msg=%s was emitted\n",
+			pipe, op_str);
+	else
+		dev_dbg(adev->dev, "MES(%d) msg=%d was emitted\n",
+			pipe, x_pkt->header.opcode);
+
+	r = amdgpu_fence_wait_polling(ring, seq, timeout);
+
+	/*
+	 * status_ptr[31:0] == 0 (fail) or status_ptr[63:0] == 1 (success).
+	 * If status_ptr[31:0] == 0 then status_ptr[63:32] will have debug error information.
+	 */
+	if (r < 1 || !(lower_32_bits(*status_ptr))) {
+
+		if (misc_op_str)
+			dev_err(adev->dev, "MES(%d) failed to respond to msg=%s (%s)\n",
+				pipe, op_str, misc_op_str);
+		else if (op_str)
+			dev_err(adev->dev, "MES(%d) failed to respond to msg=%s\n",
+				pipe, op_str);
+		else
+			dev_err(adev->dev, "MES(%d) failed to respond to msg=%d\n",
+				pipe, x_pkt->header.opcode);
+
+		while (halt_if_hws_hang)
+			schedule();
+
+		r = -ETIMEDOUT;
+		goto error_wb_free;
+	}
+
+	amdgpu_device_wb_free(adev, status_offset);
+	return 0;
+
+error_undo:
+	dev_err(adev->dev, "MES ring buffer is full.\n");
+	amdgpu_ring_undo(ring);
+
+error_unlock_free:
+	spin_unlock_irqrestore(ring_lock, flags);
+
+error_wb_free:
+	amdgpu_device_wb_free(adev, status_offset);
+	return r;
+}
+
+static int convert_to_mes_queue_type(int queue_type)
+{
+	if (queue_type == AMDGPU_RING_TYPE_GFX)
+		return MES_QUEUE_TYPE_GFX;
+	else if (queue_type == AMDGPU_RING_TYPE_COMPUTE)
+		return MES_QUEUE_TYPE_COMPUTE;
+	else if (queue_type == AMDGPU_RING_TYPE_SDMA)
+		return MES_QUEUE_TYPE_SDMA;
+	else if (queue_type == AMDGPU_RING_TYPE_MES)
+		return MES_QUEUE_TYPE_SCHQ;
+	else
+		BUG();
+	return -1;
+}
+
+static int convert_to_mes_priority_level(int priority_level)
+{
+	switch (priority_level) {
+	case AMDGPU_MES_PRIORITY_LEVEL_LOW:
+		return AMD_PRIORITY_LEVEL_LOW;
+	case AMDGPU_MES_PRIORITY_LEVEL_NORMAL:
+	default:
+		return AMD_PRIORITY_LEVEL_NORMAL;
+	case AMDGPU_MES_PRIORITY_LEVEL_MEDIUM:
+		return AMD_PRIORITY_LEVEL_MEDIUM;
+	case AMDGPU_MES_PRIORITY_LEVEL_HIGH:
+		return AMD_PRIORITY_LEVEL_HIGH;
+	case AMDGPU_MES_PRIORITY_LEVEL_REALTIME:
+		return AMD_PRIORITY_LEVEL_REALTIME;
+	}
+}
+
+static int mes_v12_0_add_hw_queue(struct amdgpu_mes *mes,
+				  struct mes_add_queue_input *input)
+{
+	struct amdgpu_device *adev = mes->adev;
+	union MESAPI__ADD_QUEUE mes_add_queue_pkt;
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB(0)];
+	uint32_t vm_cntx_cntl = hub->vm_cntx_cntl;
+
+	memset(&mes_add_queue_pkt, 0, sizeof(mes_add_queue_pkt));
+
+	mes_add_queue_pkt.header.type = MES_API_TYPE_SCHEDULER;
+	mes_add_queue_pkt.header.opcode = MES_SCH_API_ADD_QUEUE;
+	mes_add_queue_pkt.header.dwsize = API_FRAME_SIZE_IN_DWORDS;
+
+	mes_add_queue_pkt.process_id = input->process_id;
+	mes_add_queue_pkt.page_table_base_addr = input->page_table_base_addr;
+	mes_add_queue_pkt.process_va_start = input->process_va_start;
+	mes_add_queue_pkt.process_va_end = input->process_va_end;
+	mes_add_queue_pkt.process_quantum = input->process_quantum;
+	mes_add_queue_pkt.process_context_addr = input->process_context_addr;
+	mes_add_queue_pkt.gang_quantum = input->gang_quantum;
+	mes_add_queue_pkt.gang_context_addr = input->gang_context_addr;
+	mes_add_queue_pkt.inprocess_gang_priority =
+		convert_to_mes_priority_level(input->inprocess_gang_priority);
+	mes_add_queue_pkt.gang_global_priority_level =
+		convert_to_mes_priority_level(input->gang_global_priority_level);
+	mes_add_queue_pkt.doorbell_offset = input->doorbell_offset;
+	mes_add_queue_pkt.mqd_addr = input->mqd_addr;
+
+	mes_add_queue_pkt.wptr_addr = input->wptr_mc_addr;
+
+	mes_add_queue_pkt.queue_type =
+		convert_to_mes_queue_type(input->queue_type);
+	mes_add_queue_pkt.paging = input->paging;
+	mes_add_queue_pkt.vm_context_cntl = vm_cntx_cntl;
+	mes_add_queue_pkt.gws_base = input->gws_base;
+	mes_add_queue_pkt.gws_size = input->gws_size;
+	mes_add_queue_pkt.trap_handler_addr = input->tba_addr;
+	mes_add_queue_pkt.tma_addr = input->tma_addr;
+	mes_add_queue_pkt.trap_en = input->trap_en;
+	mes_add_queue_pkt.skip_process_ctx_clear = input->skip_process_ctx_clear;
+	mes_add_queue_pkt.is_kfd_process = input->is_kfd_process;
+
+	/* For KFD, gds_size is re-used for queue size (needed in MES for AQL queues) */
+	mes_add_queue_pkt.is_aql_queue = input->is_aql_queue;
+	mes_add_queue_pkt.gds_size = input->queue_size;
+
+	/* For KFD, gds_size is re-used for queue size (needed in MES for AQL queues) */
+	mes_add_queue_pkt.is_aql_queue = input->is_aql_queue;
+	mes_add_queue_pkt.gds_size = input->queue_size;
+
+	return mes_v12_0_submit_pkt_and_poll_completion(mes,
+			AMDGPU_MES_SCHED_PIPE,
+			&mes_add_queue_pkt, sizeof(mes_add_queue_pkt),
+			offsetof(union MESAPI__ADD_QUEUE, api_status));
+}
+
+static int mes_v12_0_remove_hw_queue(struct amdgpu_mes *mes,
+				     struct mes_remove_queue_input *input)
+{
+	union MESAPI__REMOVE_QUEUE mes_remove_queue_pkt;
+	uint32_t mes_rev = mes->sched_version & AMDGPU_MES_VERSION_MASK;
+
+	memset(&mes_remove_queue_pkt, 0, sizeof(mes_remove_queue_pkt));
+
+	mes_remove_queue_pkt.header.type = MES_API_TYPE_SCHEDULER;
+	mes_remove_queue_pkt.header.opcode = MES_SCH_API_REMOVE_QUEUE;
+	mes_remove_queue_pkt.header.dwsize = API_FRAME_SIZE_IN_DWORDS;
+
+	mes_remove_queue_pkt.doorbell_offset = input->doorbell_offset;
+	mes_remove_queue_pkt.gang_context_addr = input->gang_context_addr;
+
+	if (mes_rev >= 0x5a)
+		mes_remove_queue_pkt.remove_queue_after_reset = input->remove_queue_after_reset;
+
+	return mes_v12_0_submit_pkt_and_poll_completion(mes,
+			AMDGPU_MES_SCHED_PIPE,
+			&mes_remove_queue_pkt, sizeof(mes_remove_queue_pkt),
+			offsetof(union MESAPI__REMOVE_QUEUE, api_status));
+}
+
+int gfx_v12_0_request_gfx_index_mutex(struct amdgpu_device *adev,
+				      bool req)
+{
+	u32 i, tmp, val;
+
+	for (i = 0; i < adev->usec_timeout; i++) {
+		/* Request with MeId=2, PipeId=0 */
+		tmp = REG_SET_FIELD(0, CP_GFX_INDEX_MUTEX, REQUEST, req);
+		tmp = REG_SET_FIELD(tmp, CP_GFX_INDEX_MUTEX, CLIENTID, 4);
+		WREG32_SOC15(GC, 0, regCP_GFX_INDEX_MUTEX, tmp);
+
+		val = RREG32_SOC15(GC, 0, regCP_GFX_INDEX_MUTEX);
+		if (req) {
+			if (val == tmp)
+				break;
+		} else {
+			tmp = REG_SET_FIELD(tmp, CP_GFX_INDEX_MUTEX,
+					    REQUEST, 1);
+
+			/* unlocked or locked by firmware */
+			if (val != tmp)
+				break;
+		}
+		udelay(1);
+	}
+
+	if (i >= adev->usec_timeout)
+		return -EINVAL;
+
+	return 0;
+}
+
+static int mes_v12_0_reset_queue_mmio(struct amdgpu_mes *mes, uint32_t queue_type,
+				      uint32_t me_id, uint32_t pipe_id,
+				      uint32_t queue_id, uint32_t vmid)
+{
+	struct amdgpu_device *adev = mes->adev;
+	uint32_t value, reg;
+	int i, r = 0;
+
+	amdgpu_gfx_rlc_enter_safe_mode(adev, 0);
+
+	if (queue_type == AMDGPU_RING_TYPE_GFX) {
+		dev_info(adev->dev, "reset gfx queue (%d:%d:%d: vmid:%d)\n",
+			 me_id, pipe_id, queue_id, vmid);
+
+		mutex_lock(&adev->gfx.reset_sem_mutex);
+		gfx_v12_0_request_gfx_index_mutex(adev, true);
+		/* all se allow writes */
+		WREG32_SOC15(GC, 0, regGRBM_GFX_INDEX,
+			     (uint32_t)(0x1 << GRBM_GFX_INDEX__SE_BROADCAST_WRITES__SHIFT));
+		value = REG_SET_FIELD(0, CP_VMID_RESET, RESET_REQUEST, 1 << vmid);
+		if (pipe_id == 0)
+			value = REG_SET_FIELD(value, CP_VMID_RESET, PIPE0_QUEUES, 1 << queue_id);
+		else
+			value = REG_SET_FIELD(value, CP_VMID_RESET, PIPE1_QUEUES, 1 << queue_id);
+		WREG32_SOC15(GC, 0, regCP_VMID_RESET, value);
+		gfx_v12_0_request_gfx_index_mutex(adev, false);
+		mutex_unlock(&adev->gfx.reset_sem_mutex);
+
+		mutex_lock(&adev->srbm_mutex);
+		soc21_grbm_select(adev, me_id, pipe_id, queue_id, 0);
+		/* wait till dequeue take effects */
+		for (i = 0; i < adev->usec_timeout; i++) {
+			if (!(RREG32_SOC15(GC, 0, regCP_GFX_HQD_ACTIVE) & 1))
+				break;
+			udelay(1);
+		}
+		if (i >= adev->usec_timeout) {
+			dev_err(adev->dev, "failed to wait on gfx hqd deactivate\n");
+			r = -ETIMEDOUT;
+		}
+
+		soc21_grbm_select(adev, 0, 0, 0, 0);
+		mutex_unlock(&adev->srbm_mutex);
+	} else if (queue_type == AMDGPU_RING_TYPE_COMPUTE) {
+		dev_info(adev->dev, "reset compute queue (%d:%d:%d)\n",
+			 me_id, pipe_id, queue_id);
+		mutex_lock(&adev->srbm_mutex);
+		soc21_grbm_select(adev, me_id, pipe_id, queue_id, 0);
+		WREG32_SOC15(GC, 0, regCP_HQD_DEQUEUE_REQUEST, 0x2);
+		WREG32_SOC15(GC, 0, regSPI_COMPUTE_QUEUE_RESET, 0x1);
+
+		/* wait till dequeue take effects */
+		for (i = 0; i < adev->usec_timeout; i++) {
+			if (!(RREG32_SOC15(GC, 0, regCP_HQD_ACTIVE) & 1))
+				break;
+			udelay(1);
+		}
+		if (i >= adev->usec_timeout) {
+			dev_err(adev->dev, "failed to wait on hqd deactivate\n");
+			r = -ETIMEDOUT;
+		}
+		soc21_grbm_select(adev, 0, 0, 0, 0);
+		mutex_unlock(&adev->srbm_mutex);
+	} else if (queue_type == AMDGPU_RING_TYPE_SDMA) {
+		dev_info(adev->dev, "reset sdma queue (%d:%d:%d)\n",
+			 me_id, pipe_id, queue_id);
+		switch (me_id) {
+		case 1:
+			reg = SOC15_REG_OFFSET(GC, 0, regSDMA1_QUEUE_RESET_REQ);
+			break;
+		case 0:
+		default:
+			reg = SOC15_REG_OFFSET(GC, 0, regSDMA0_QUEUE_RESET_REQ);
+			break;
+		}
+
+		value = 1 << queue_id;
+		WREG32(reg, value);
+		/* wait for queue reset done */
+		for (i = 0; i < adev->usec_timeout; i++) {
+			if (!(RREG32(reg) & value))
+				break;
+			udelay(1);
+		}
+		if (i >= adev->usec_timeout) {
+			dev_err(adev->dev, "failed to wait on sdma queue reset done\n");
+			r = -ETIMEDOUT;
+		}
+	}
+
+	amdgpu_gfx_rlc_exit_safe_mode(adev, 0);
+	return r;
+}
+
+static int mes_v12_0_map_legacy_queue(struct amdgpu_mes *mes,
+				      struct mes_map_legacy_queue_input *input)
+{
+	union MESAPI__ADD_QUEUE mes_add_queue_pkt;
+	int pipe;
+
+	memset(&mes_add_queue_pkt, 0, sizeof(mes_add_queue_pkt));
+
+	mes_add_queue_pkt.header.type = MES_API_TYPE_SCHEDULER;
+	mes_add_queue_pkt.header.opcode = MES_SCH_API_ADD_QUEUE;
+	mes_add_queue_pkt.header.dwsize = API_FRAME_SIZE_IN_DWORDS;
+
+	mes_add_queue_pkt.pipe_id = input->pipe_id;
+	mes_add_queue_pkt.queue_id = input->queue_id;
+	mes_add_queue_pkt.doorbell_offset = input->doorbell_offset;
+	mes_add_queue_pkt.mqd_addr = input->mqd_addr;
+	mes_add_queue_pkt.wptr_addr = input->wptr_addr;
+	mes_add_queue_pkt.queue_type =
+		convert_to_mes_queue_type(input->queue_type);
+	mes_add_queue_pkt.map_legacy_kq = 1;
+
+	if (mes->adev->enable_uni_mes)
+		pipe = AMDGPU_MES_KIQ_PIPE;
+	else
+		pipe = AMDGPU_MES_SCHED_PIPE;
+
+	return mes_v12_0_submit_pkt_and_poll_completion(mes, pipe,
+			&mes_add_queue_pkt, sizeof(mes_add_queue_pkt),
+			offsetof(union MESAPI__ADD_QUEUE, api_status));
+}
+
+static int mes_v12_0_unmap_legacy_queue(struct amdgpu_mes *mes,
+			struct mes_unmap_legacy_queue_input *input)
+{
+	union MESAPI__REMOVE_QUEUE mes_remove_queue_pkt;
+	int pipe;
+
+	memset(&mes_remove_queue_pkt, 0, sizeof(mes_remove_queue_pkt));
+
+	mes_remove_queue_pkt.header.type = MES_API_TYPE_SCHEDULER;
+	mes_remove_queue_pkt.header.opcode = MES_SCH_API_REMOVE_QUEUE;
+	mes_remove_queue_pkt.header.dwsize = API_FRAME_SIZE_IN_DWORDS;
+
+	mes_remove_queue_pkt.doorbell_offset = input->doorbell_offset;
+	mes_remove_queue_pkt.gang_context_addr = 0;
+
+	mes_remove_queue_pkt.pipe_id = input->pipe_id;
+	mes_remove_queue_pkt.queue_id = input->queue_id;
+
+	if (input->action == PREEMPT_QUEUES_NO_UNMAP) {
+		mes_remove_queue_pkt.preempt_legacy_gfx_queue = 1;
+		mes_remove_queue_pkt.tf_addr = input->trail_fence_addr;
+		mes_remove_queue_pkt.tf_data =
+			lower_32_bits(input->trail_fence_data);
+	} else {
+		mes_remove_queue_pkt.unmap_legacy_queue = 1;
+		mes_remove_queue_pkt.queue_type =
+			convert_to_mes_queue_type(input->queue_type);
+	}
+
+	if (mes->adev->enable_uni_mes)
+		pipe = AMDGPU_MES_KIQ_PIPE;
+	else
+		pipe = AMDGPU_MES_SCHED_PIPE;
+
+	return mes_v12_0_submit_pkt_and_poll_completion(mes, pipe,
+			&mes_remove_queue_pkt, sizeof(mes_remove_queue_pkt),
+			offsetof(union MESAPI__REMOVE_QUEUE, api_status));
+}
+
+static int mes_v12_0_suspend_gang(struct amdgpu_mes *mes,
+				  struct mes_suspend_gang_input *input)
+{
+	union MESAPI__SUSPEND mes_suspend_gang_pkt;
+
+	memset(&mes_suspend_gang_pkt, 0, sizeof(mes_suspend_gang_pkt));
+
+	mes_suspend_gang_pkt.header.type = MES_API_TYPE_SCHEDULER;
+	mes_suspend_gang_pkt.header.opcode = MES_SCH_API_SUSPEND;
+	mes_suspend_gang_pkt.header.dwsize = API_FRAME_SIZE_IN_DWORDS;
+
+	mes_suspend_gang_pkt.suspend_all_gangs = input->suspend_all_gangs;
+	mes_suspend_gang_pkt.gang_context_addr = input->gang_context_addr;
+	mes_suspend_gang_pkt.suspend_fence_addr = input->suspend_fence_addr;
+	mes_suspend_gang_pkt.suspend_fence_value = input->suspend_fence_value;
+
+	return mes_v12_0_submit_pkt_and_poll_completion(mes, AMDGPU_MES_SCHED_PIPE,
+			&mes_suspend_gang_pkt, sizeof(mes_suspend_gang_pkt),
+			offsetof(union MESAPI__SUSPEND, api_status));
+}
+
+static int mes_v12_0_resume_gang(struct amdgpu_mes *mes,
+				 struct mes_resume_gang_input *input)
+{
+	union MESAPI__RESUME mes_resume_gang_pkt;
+
+	memset(&mes_resume_gang_pkt, 0, sizeof(mes_resume_gang_pkt));
+
+	mes_resume_gang_pkt.header.type = MES_API_TYPE_SCHEDULER;
+	mes_resume_gang_pkt.header.opcode = MES_SCH_API_RESUME;
+	mes_resume_gang_pkt.header.dwsize = API_FRAME_SIZE_IN_DWORDS;
+
+	mes_resume_gang_pkt.resume_all_gangs = input->resume_all_gangs;
+	mes_resume_gang_pkt.gang_context_addr = input->gang_context_addr;
+
+	return mes_v12_0_submit_pkt_and_poll_completion(mes, AMDGPU_MES_SCHED_PIPE,
+			&mes_resume_gang_pkt, sizeof(mes_resume_gang_pkt),
+			offsetof(union MESAPI__RESUME, api_status));
+}
+
+static int mes_v12_0_query_sched_status(struct amdgpu_mes *mes, int pipe)
+{
+	union MESAPI__QUERY_MES_STATUS mes_status_pkt;
+
+	memset(&mes_status_pkt, 0, sizeof(mes_status_pkt));
+
+	mes_status_pkt.header.type = MES_API_TYPE_SCHEDULER;
+	mes_status_pkt.header.opcode = MES_SCH_API_QUERY_SCHEDULER_STATUS;
+	mes_status_pkt.header.dwsize = API_FRAME_SIZE_IN_DWORDS;
+
+	return mes_v12_0_submit_pkt_and_poll_completion(mes, pipe,
+			&mes_status_pkt, sizeof(mes_status_pkt),
+			offsetof(union MESAPI__QUERY_MES_STATUS, api_status));
+}
+
+static int mes_v12_0_misc_op(struct amdgpu_mes *mes,
+			     struct mes_misc_op_input *input)
+{
+	union MESAPI__MISC misc_pkt;
+	int pipe;
+
+	if (mes->adev->enable_uni_mes)
+		pipe = AMDGPU_MES_KIQ_PIPE;
+	else
+		pipe = AMDGPU_MES_SCHED_PIPE;
+
+	memset(&misc_pkt, 0, sizeof(misc_pkt));
+
+	misc_pkt.header.type = MES_API_TYPE_SCHEDULER;
+	misc_pkt.header.opcode = MES_SCH_API_MISC;
+	misc_pkt.header.dwsize = API_FRAME_SIZE_IN_DWORDS;
+
+	switch (input->op) {
+	case MES_MISC_OP_READ_REG:
+		misc_pkt.opcode = MESAPI_MISC__READ_REG;
+		misc_pkt.read_reg.reg_offset = input->read_reg.reg_offset;
+		misc_pkt.read_reg.buffer_addr = input->read_reg.buffer_addr;
+		break;
+	case MES_MISC_OP_WRITE_REG:
+		misc_pkt.opcode = MESAPI_MISC__WRITE_REG;
+		misc_pkt.write_reg.reg_offset = input->write_reg.reg_offset;
+		misc_pkt.write_reg.reg_value = input->write_reg.reg_value;
+		break;
+	case MES_MISC_OP_WRM_REG_WAIT:
+		misc_pkt.opcode = MESAPI_MISC__WAIT_REG_MEM;
+		misc_pkt.wait_reg_mem.op = WRM_OPERATION__WAIT_REG_MEM;
+		misc_pkt.wait_reg_mem.reference = input->wrm_reg.ref;
+		misc_pkt.wait_reg_mem.mask = input->wrm_reg.mask;
+		misc_pkt.wait_reg_mem.reg_offset1 = input->wrm_reg.reg0;
+		misc_pkt.wait_reg_mem.reg_offset2 = 0;
+		break;
+	case MES_MISC_OP_WRM_REG_WR_WAIT:
+		misc_pkt.opcode = MESAPI_MISC__WAIT_REG_MEM;
+		misc_pkt.wait_reg_mem.op = WRM_OPERATION__WR_WAIT_WR_REG;
+		misc_pkt.wait_reg_mem.reference = input->wrm_reg.ref;
+		misc_pkt.wait_reg_mem.mask = input->wrm_reg.mask;
+		misc_pkt.wait_reg_mem.reg_offset1 = input->wrm_reg.reg0;
+		misc_pkt.wait_reg_mem.reg_offset2 = input->wrm_reg.reg1;
+		break;
+	case MES_MISC_OP_SET_SHADER_DEBUGGER:
+		pipe = AMDGPU_MES_SCHED_PIPE;
+		misc_pkt.opcode = MESAPI_MISC__SET_SHADER_DEBUGGER;
+		misc_pkt.set_shader_debugger.process_context_addr =
+				input->set_shader_debugger.process_context_addr;
+		misc_pkt.set_shader_debugger.flags.u32all =
+				input->set_shader_debugger.flags.u32all;
+		misc_pkt.set_shader_debugger.spi_gdbg_per_vmid_cntl =
+				input->set_shader_debugger.spi_gdbg_per_vmid_cntl;
+		memcpy(misc_pkt.set_shader_debugger.tcp_watch_cntl,
+				input->set_shader_debugger.tcp_watch_cntl,
+				sizeof(misc_pkt.set_shader_debugger.tcp_watch_cntl));
+		misc_pkt.set_shader_debugger.trap_en = input->set_shader_debugger.trap_en;
+		break;
+	case MES_MISC_OP_CHANGE_CONFIG:
+		misc_pkt.opcode = MESAPI_MISC__CHANGE_CONFIG;
+		misc_pkt.change_config.opcode =
+				MESAPI_MISC__CHANGE_CONFIG_OPTION_LIMIT_SINGLE_PROCESS;
+		misc_pkt.change_config.option.bits.limit_single_process =
+				input->change_config.option.limit_single_process;
+		break;
+
+	default:
+		DRM_ERROR("unsupported misc op (%d) \n", input->op);
+		return -EINVAL;
+	}
+
+	return mes_v12_0_submit_pkt_and_poll_completion(mes, pipe,
+			&misc_pkt, sizeof(misc_pkt),
+			offsetof(union MESAPI__MISC, api_status));
+}
+
+static int mes_v12_0_set_hw_resources_1(struct amdgpu_mes *mes, int pipe)
+{
+	union MESAPI_SET_HW_RESOURCES_1 mes_set_hw_res_1_pkt;
+
+	memset(&mes_set_hw_res_1_pkt, 0, sizeof(mes_set_hw_res_1_pkt));
+
+	mes_set_hw_res_1_pkt.header.type = MES_API_TYPE_SCHEDULER;
+	mes_set_hw_res_1_pkt.header.opcode = MES_SCH_API_SET_HW_RSRC_1;
+	mes_set_hw_res_1_pkt.header.dwsize = API_FRAME_SIZE_IN_DWORDS;
+	mes_set_hw_res_1_pkt.mes_kiq_unmap_timeout = 0xa;
+	mes_set_hw_res_1_pkt.cleaner_shader_fence_mc_addr =
+		mes->resource_1_gpu_addr[pipe];
+
+	return mes_v12_0_submit_pkt_and_poll_completion(mes, pipe,
+			&mes_set_hw_res_1_pkt, sizeof(mes_set_hw_res_1_pkt),
+			offsetof(union MESAPI_SET_HW_RESOURCES_1, api_status));
+}
+
+static int mes_v12_0_set_hw_resources(struct amdgpu_mes *mes, int pipe)
+{
+	int i;
+	struct amdgpu_device *adev = mes->adev;
+	union MESAPI_SET_HW_RESOURCES mes_set_hw_res_pkt;
+
+	memset(&mes_set_hw_res_pkt, 0, sizeof(mes_set_hw_res_pkt));
+
+	mes_set_hw_res_pkt.header.type = MES_API_TYPE_SCHEDULER;
+	mes_set_hw_res_pkt.header.opcode = MES_SCH_API_SET_HW_RSRC;
+	mes_set_hw_res_pkt.header.dwsize = API_FRAME_SIZE_IN_DWORDS;
+
+	if (pipe == AMDGPU_MES_SCHED_PIPE) {
+		mes_set_hw_res_pkt.vmid_mask_mmhub = mes->vmid_mask_mmhub;
+		mes_set_hw_res_pkt.vmid_mask_gfxhub = mes->vmid_mask_gfxhub;
+		mes_set_hw_res_pkt.gds_size = adev->gds.gds_size;
+		mes_set_hw_res_pkt.paging_vmid = 0;
+
+		for (i = 0; i < MAX_COMPUTE_PIPES; i++)
+			mes_set_hw_res_pkt.compute_hqd_mask[i] =
+				mes->compute_hqd_mask[i];
+
+		for (i = 0; i < MAX_GFX_PIPES; i++)
+			mes_set_hw_res_pkt.gfx_hqd_mask[i] =
+				mes->gfx_hqd_mask[i];
+
+		for (i = 0; i < MAX_SDMA_PIPES; i++)
+			mes_set_hw_res_pkt.sdma_hqd_mask[i] =
+				mes->sdma_hqd_mask[i];
+
+		for (i = 0; i < AMD_PRIORITY_NUM_LEVELS; i++)
+			mes_set_hw_res_pkt.aggregated_doorbells[i] =
+				mes->aggregated_doorbells[i];
+	}
+
+	mes_set_hw_res_pkt.g_sch_ctx_gpu_mc_ptr =
+		mes->sch_ctx_gpu_addr[pipe];
+	mes_set_hw_res_pkt.query_status_fence_gpu_mc_ptr =
+		mes->query_status_fence_gpu_addr[pipe];
+
+	for (i = 0; i < 5; i++) {
+		mes_set_hw_res_pkt.gc_base[i] = adev->reg_offset[GC_HWIP][0][i];
+		mes_set_hw_res_pkt.mmhub_base[i] =
+				adev->reg_offset[MMHUB_HWIP][0][i];
+		mes_set_hw_res_pkt.osssys_base[i] =
+		adev->reg_offset[OSSSYS_HWIP][0][i];
+	}
+
+	mes_set_hw_res_pkt.disable_reset = 1;
+	mes_set_hw_res_pkt.disable_mes_log = 1;
+	mes_set_hw_res_pkt.use_different_vmid_compute = 1;
+	mes_set_hw_res_pkt.enable_reg_active_poll = 1;
+	mes_set_hw_res_pkt.enable_level_process_quantum_check = 1;
+	if ((mes->adev->mes.sched_version & AMDGPU_MES_VERSION_MASK) >= 0x82)
+		mes_set_hw_res_pkt.enable_lr_compute_wa = 1;
+	else
+		dev_info_once(adev->dev,
+			      "MES FW version must be >= 0x82 to enable LR compute workaround.\n");
+
+	/*
+	 * Keep oversubscribe timer for sdma . When we have unmapped doorbell
+	 * handling support, other queue will not use the oversubscribe timer.
+	 * handling  mode - 0: disabled; 1: basic version; 2: basic+ version
+	 */
+	mes_set_hw_res_pkt.oversubscription_timer = 50;
+	mes_set_hw_res_pkt.unmapped_doorbell_handling = 1;
+
+	if (amdgpu_mes_log_enable) {
+		mes_set_hw_res_pkt.enable_mes_event_int_logging = 1;
+		mes_set_hw_res_pkt.event_intr_history_gpu_mc_ptr = mes->event_log_gpu_addr +
+				pipe * (AMDGPU_MES_LOG_BUFFER_SIZE + AMDGPU_MES_MSCRATCH_SIZE);
+	}
+
+	if (adev->enforce_isolation[0] == AMDGPU_ENFORCE_ISOLATION_ENABLE)
+		mes_set_hw_res_pkt.limit_single_process = 1;
+
+	return mes_v12_0_submit_pkt_and_poll_completion(mes, pipe,
+			&mes_set_hw_res_pkt, sizeof(mes_set_hw_res_pkt),
+			offsetof(union MESAPI_SET_HW_RESOURCES, api_status));
+}
+
+static void mes_v12_0_init_aggregated_doorbell(struct amdgpu_mes *mes)
+{
+	struct amdgpu_device *adev = mes->adev;
+	uint32_t data;
+
+	data = RREG32_SOC15(GC, 0, regCP_MES_DOORBELL_CONTROL1);
+	data &= ~(CP_MES_DOORBELL_CONTROL1__DOORBELL_OFFSET_MASK |
+		  CP_MES_DOORBELL_CONTROL1__DOORBELL_EN_MASK |
+		  CP_MES_DOORBELL_CONTROL1__DOORBELL_HIT_MASK);
+	data |= mes->aggregated_doorbells[AMDGPU_MES_PRIORITY_LEVEL_LOW] <<
+		CP_MES_DOORBELL_CONTROL1__DOORBELL_OFFSET__SHIFT;
+	data |= 1 << CP_MES_DOORBELL_CONTROL1__DOORBELL_EN__SHIFT;
+	WREG32_SOC15(GC, 0, regCP_MES_DOORBELL_CONTROL1, data);
+
+	data = RREG32_SOC15(GC, 0, regCP_MES_DOORBELL_CONTROL2);
+	data &= ~(CP_MES_DOORBELL_CONTROL2__DOORBELL_OFFSET_MASK |
+		  CP_MES_DOORBELL_CONTROL2__DOORBELL_EN_MASK |
+		  CP_MES_DOORBELL_CONTROL2__DOORBELL_HIT_MASK);
+	data |= mes->aggregated_doorbells[AMDGPU_MES_PRIORITY_LEVEL_NORMAL] <<
+		CP_MES_DOORBELL_CONTROL2__DOORBELL_OFFSET__SHIFT;
+	data |= 1 << CP_MES_DOORBELL_CONTROL2__DOORBELL_EN__SHIFT;
+	WREG32_SOC15(GC, 0, regCP_MES_DOORBELL_CONTROL2, data);
+
+	data = RREG32_SOC15(GC, 0, regCP_MES_DOORBELL_CONTROL3);
+	data &= ~(CP_MES_DOORBELL_CONTROL3__DOORBELL_OFFSET_MASK |
+		  CP_MES_DOORBELL_CONTROL3__DOORBELL_EN_MASK |
+		  CP_MES_DOORBELL_CONTROL3__DOORBELL_HIT_MASK);
+	data |= mes->aggregated_doorbells[AMDGPU_MES_PRIORITY_LEVEL_MEDIUM] <<
+		CP_MES_DOORBELL_CONTROL3__DOORBELL_OFFSET__SHIFT;
+	data |= 1 << CP_MES_DOORBELL_CONTROL3__DOORBELL_EN__SHIFT;
+	WREG32_SOC15(GC, 0, regCP_MES_DOORBELL_CONTROL3, data);
+
+	data = RREG32_SOC15(GC, 0, regCP_MES_DOORBELL_CONTROL4);
+	data &= ~(CP_MES_DOORBELL_CONTROL4__DOORBELL_OFFSET_MASK |
+		  CP_MES_DOORBELL_CONTROL4__DOORBELL_EN_MASK |
+		  CP_MES_DOORBELL_CONTROL4__DOORBELL_HIT_MASK);
+	data |= mes->aggregated_doorbells[AMDGPU_MES_PRIORITY_LEVEL_HIGH] <<
+		CP_MES_DOORBELL_CONTROL4__DOORBELL_OFFSET__SHIFT;
+	data |= 1 << CP_MES_DOORBELL_CONTROL4__DOORBELL_EN__SHIFT;
+	WREG32_SOC15(GC, 0, regCP_MES_DOORBELL_CONTROL4, data);
+
+	data = RREG32_SOC15(GC, 0, regCP_MES_DOORBELL_CONTROL5);
+	data &= ~(CP_MES_DOORBELL_CONTROL5__DOORBELL_OFFSET_MASK |
+		  CP_MES_DOORBELL_CONTROL5__DOORBELL_EN_MASK |
+		  CP_MES_DOORBELL_CONTROL5__DOORBELL_HIT_MASK);
+	data |= mes->aggregated_doorbells[AMDGPU_MES_PRIORITY_LEVEL_REALTIME] <<
+		CP_MES_DOORBELL_CONTROL5__DOORBELL_OFFSET__SHIFT;
+	data |= 1 << CP_MES_DOORBELL_CONTROL5__DOORBELL_EN__SHIFT;
+	WREG32_SOC15(GC, 0, regCP_MES_DOORBELL_CONTROL5, data);
+
+	data = 1 << CP_HQD_GFX_CONTROL__DB_UPDATED_MSG_EN__SHIFT;
+	WREG32_SOC15(GC, 0, regCP_HQD_GFX_CONTROL, data);
+}
+
+
+static void mes_v12_0_enable_unmapped_doorbell_handling(
+		struct amdgpu_mes *mes, bool enable)
+{
+	struct amdgpu_device *adev = mes->adev;
+	uint32_t data = RREG32_SOC15(GC, 0, regCP_UNMAPPED_DOORBELL);
+
+	/*
+	 * The default PROC_LSB settng is 0xc which means doorbell
+	 * addr[16:12] gives the doorbell page number. For kfd, each
+	 * process will use 2 pages of doorbell, we need to change the
+	 * setting to 0xd
+	 */
+	data &= ~CP_UNMAPPED_DOORBELL__PROC_LSB_MASK;
+	data |= 0xd <<  CP_UNMAPPED_DOORBELL__PROC_LSB__SHIFT;
+
+	data |= (enable ? 1 : 0) << CP_UNMAPPED_DOORBELL__ENABLE__SHIFT;
+
+	WREG32_SOC15(GC, 0, regCP_UNMAPPED_DOORBELL, data);
+}
+
+static int mes_v12_0_reset_hw_queue(struct amdgpu_mes *mes,
+				    struct mes_reset_queue_input *input)
+{
+	union MESAPI__RESET mes_reset_queue_pkt;
+	int pipe;
+
+	if (input->use_mmio)
+		return mes_v12_0_reset_queue_mmio(mes, input->queue_type,
+						  input->me_id, input->pipe_id,
+						  input->queue_id, input->vmid);
+
+	memset(&mes_reset_queue_pkt, 0, sizeof(mes_reset_queue_pkt));
+
+	mes_reset_queue_pkt.header.type = MES_API_TYPE_SCHEDULER;
+	mes_reset_queue_pkt.header.opcode = MES_SCH_API_RESET;
+	mes_reset_queue_pkt.header.dwsize = API_FRAME_SIZE_IN_DWORDS;
+
+	mes_reset_queue_pkt.queue_type =
+		convert_to_mes_queue_type(input->queue_type);
+
+	if (input->legacy_gfx) {
+		mes_reset_queue_pkt.reset_legacy_gfx = 1;
+		mes_reset_queue_pkt.pipe_id_lp = input->pipe_id;
+		mes_reset_queue_pkt.queue_id_lp = input->queue_id;
+		mes_reset_queue_pkt.mqd_mc_addr_lp = input->mqd_addr;
+		mes_reset_queue_pkt.doorbell_offset_lp = input->doorbell_offset;
+		mes_reset_queue_pkt.wptr_addr_lp = input->wptr_addr;
+		mes_reset_queue_pkt.vmid_id_lp = input->vmid;
+	} else {
+		mes_reset_queue_pkt.reset_queue_only = 1;
+		mes_reset_queue_pkt.doorbell_offset = input->doorbell_offset;
+	}
+
+	if (input->is_kq)
+		pipe = AMDGPU_MES_KIQ_PIPE;
+	else
+		pipe = AMDGPU_MES_SCHED_PIPE;
+
+	return mes_v12_0_submit_pkt_and_poll_completion(mes, pipe,
+			&mes_reset_queue_pkt, sizeof(mes_reset_queue_pkt),
+			offsetof(union MESAPI__RESET, api_status));
+}
+
+static int mes_v12_0_detect_and_reset_hung_queues(struct amdgpu_mes *mes,
+						  struct mes_detect_and_reset_queue_input *input)
+{
+	union MESAPI__RESET mes_reset_queue_pkt;
+
+	memset(&mes_reset_queue_pkt, 0, sizeof(mes_reset_queue_pkt));
+
+	mes_reset_queue_pkt.header.type = MES_API_TYPE_SCHEDULER;
+	mes_reset_queue_pkt.header.opcode = MES_SCH_API_RESET;
+	mes_reset_queue_pkt.header.dwsize = API_FRAME_SIZE_IN_DWORDS;
+
+	mes_reset_queue_pkt.queue_type =
+		convert_to_mes_queue_type(input->queue_type);
+	mes_reset_queue_pkt.doorbell_offset_addr =
+		mes->hung_queue_db_array_gpu_addr;
+
+	if (input->detect_only)
+		mes_reset_queue_pkt.hang_detect_only = 1;
+	else
+		mes_reset_queue_pkt.hang_detect_then_reset = 1;
+
+	return mes_v12_0_submit_pkt_and_poll_completion(mes, AMDGPU_MES_SCHED_PIPE,
+			&mes_reset_queue_pkt, sizeof(mes_reset_queue_pkt),
+			offsetof(union MESAPI__RESET, api_status));
+}
+
+static int mes_v12_inv_tlb_convert_hub_id(uint8_t id)
+{
+	/*
+	 * MES doesn't support invalidate gc_hub on slave xcc individually
+	 * master xcc will invalidate all gc_hub for the partition
+	 */
+	if (AMDGPU_IS_GFXHUB(id))
+		return 0;
+	else if (AMDGPU_IS_MMHUB0(id))
+		return 1;
+	else
+		return -EINVAL;
+
+}
+
+static int mes_v12_0_inv_tlbs_pasid(struct amdgpu_mes *mes,
+				    struct mes_inv_tlbs_pasid_input *input)
+{
+	union MESAPI__INV_TLBS mes_inv_tlbs;
+	int ret;
+
+	memset(&mes_inv_tlbs, 0, sizeof(mes_inv_tlbs));
+
+	mes_inv_tlbs.header.type = MES_API_TYPE_SCHEDULER;
+	mes_inv_tlbs.header.opcode = MES_SCH_API_INV_TLBS;
+	mes_inv_tlbs.header.dwsize = API_FRAME_SIZE_IN_DWORDS;
+
+	mes_inv_tlbs.invalidate_tlbs.inv_sel = 0;
+	mes_inv_tlbs.invalidate_tlbs.flush_type = input->flush_type;
+	mes_inv_tlbs.invalidate_tlbs.inv_sel_id = input->pasid;
+
+	/*convert amdgpu_mes_hub_id to mes expected hub_id */
+	ret = mes_v12_inv_tlb_convert_hub_id(input->hub_id);
+	if (ret < 0)
+		return -EINVAL;
+	mes_inv_tlbs.invalidate_tlbs.hub_id = ret;
+	return mes_v12_0_submit_pkt_and_poll_completion(mes, AMDGPU_MES_KIQ_PIPE,
+			&mes_inv_tlbs, sizeof(mes_inv_tlbs),
+			offsetof(union MESAPI__INV_TLBS, api_status));
+
+}
+
+static const struct amdgpu_mes_funcs mes_v12_0_funcs = {
+	.add_hw_queue = mes_v12_0_add_hw_queue,
+	.remove_hw_queue = mes_v12_0_remove_hw_queue,
+	.map_legacy_queue = mes_v12_0_map_legacy_queue,
+	.unmap_legacy_queue = mes_v12_0_unmap_legacy_queue,
+	.suspend_gang = mes_v12_0_suspend_gang,
+	.resume_gang = mes_v12_0_resume_gang,
+	.misc_op = mes_v12_0_misc_op,
+	.reset_hw_queue = mes_v12_0_reset_hw_queue,
+	.invalidate_tlbs_pasid = mes_v12_0_inv_tlbs_pasid,
+	.detect_and_reset_hung_queues = mes_v12_0_detect_and_reset_hung_queues,
+};
+
+static int mes_v12_0_allocate_ucode_buffer(struct amdgpu_device *adev,
+					   enum amdgpu_mes_pipe pipe)
+{
+	int r;
+	const struct mes_firmware_header_v1_0 *mes_hdr;
+	const __le32 *fw_data;
+	unsigned fw_size;
+
+	mes_hdr = (const struct mes_firmware_header_v1_0 *)
+		adev->mes.fw[pipe]->data;
+
+	fw_data = (const __le32 *)(adev->mes.fw[pipe]->data +
+		   le32_to_cpu(mes_hdr->mes_ucode_offset_bytes));
+	fw_size = le32_to_cpu(mes_hdr->mes_ucode_size_bytes);
+
+	r = amdgpu_bo_create_reserved(adev, fw_size,
+				      PAGE_SIZE,
+				      AMDGPU_GEM_DOMAIN_VRAM,
+				      &adev->mes.ucode_fw_obj[pipe],
+				      &adev->mes.ucode_fw_gpu_addr[pipe],
+				      (void **)&adev->mes.ucode_fw_ptr[pipe]);
+	if (r) {
+		dev_err(adev->dev, "(%d) failed to create mes fw bo\n", r);
+		return r;
+	}
+
+	memcpy(adev->mes.ucode_fw_ptr[pipe], fw_data, fw_size);
+
+	amdgpu_bo_kunmap(adev->mes.ucode_fw_obj[pipe]);
+	amdgpu_bo_unreserve(adev->mes.ucode_fw_obj[pipe]);
+
+	return 0;
+}
+
+static int mes_v12_0_allocate_ucode_data_buffer(struct amdgpu_device *adev,
+						enum amdgpu_mes_pipe pipe)
+{
+	int r;
+	const struct mes_firmware_header_v1_0 *mes_hdr;
+	const __le32 *fw_data;
+	unsigned fw_size;
+
+	mes_hdr = (const struct mes_firmware_header_v1_0 *)
+		adev->mes.fw[pipe]->data;
+
+	fw_data = (const __le32 *)(adev->mes.fw[pipe]->data +
+		   le32_to_cpu(mes_hdr->mes_ucode_data_offset_bytes));
+	fw_size = le32_to_cpu(mes_hdr->mes_ucode_data_size_bytes);
+
+	r = amdgpu_bo_create_reserved(adev, fw_size,
+				      64 * 1024,
+				      AMDGPU_GEM_DOMAIN_VRAM,
+				      &adev->mes.data_fw_obj[pipe],
+				      &adev->mes.data_fw_gpu_addr[pipe],
+				      (void **)&adev->mes.data_fw_ptr[pipe]);
+	if (r) {
+		dev_err(adev->dev, "(%d) failed to create mes data fw bo\n", r);
+		return r;
+	}
+
+	memcpy(adev->mes.data_fw_ptr[pipe], fw_data, fw_size);
+
+	amdgpu_bo_kunmap(adev->mes.data_fw_obj[pipe]);
+	amdgpu_bo_unreserve(adev->mes.data_fw_obj[pipe]);
+
+	return 0;
+}
+
+static void mes_v12_0_free_ucode_buffers(struct amdgpu_device *adev,
+					 enum amdgpu_mes_pipe pipe)
+{
+	amdgpu_bo_free_kernel(&adev->mes.data_fw_obj[pipe],
+			      &adev->mes.data_fw_gpu_addr[pipe],
+			      (void **)&adev->mes.data_fw_ptr[pipe]);
+
+	amdgpu_bo_free_kernel(&adev->mes.ucode_fw_obj[pipe],
+			      &adev->mes.ucode_fw_gpu_addr[pipe],
+			      (void **)&adev->mes.ucode_fw_ptr[pipe]);
+}
+
+static void mes_v12_0_enable(struct amdgpu_device *adev, bool enable)
+{
+	uint64_t ucode_addr;
+	uint32_t pipe, data = 0;
+
+	if (enable) {
+		mutex_lock(&adev->srbm_mutex);
+		for (pipe = 0; pipe < AMDGPU_MAX_MES_PIPES; pipe++) {
+			soc21_grbm_select(adev, 3, pipe, 0, 0);
+			if (amdgpu_mes_log_enable) {
+				u32 log_size = AMDGPU_MES_LOG_BUFFER_SIZE + AMDGPU_MES_MSCRATCH_SIZE;
+				/* In case uni mes is not enabled, only program for pipe 0 */
+				if (adev->mes.event_log_size >= (pipe + 1) * log_size) {
+					WREG32_SOC15(GC, 0, regCP_MES_MSCRATCH_LO,
+						     lower_32_bits(adev->mes.event_log_gpu_addr +
+						     pipe * log_size + AMDGPU_MES_LOG_BUFFER_SIZE));
+					WREG32_SOC15(GC, 0, regCP_MES_MSCRATCH_HI,
+						     upper_32_bits(adev->mes.event_log_gpu_addr +
+						     pipe * log_size + AMDGPU_MES_LOG_BUFFER_SIZE));
+					dev_info(adev->dev, "Setup CP MES MSCRATCH address : 0x%x. 0x%x\n",
+						 RREG32_SOC15(GC, 0, regCP_MES_MSCRATCH_HI),
+						 RREG32_SOC15(GC, 0, regCP_MES_MSCRATCH_LO));
+				}
+			}
+
+			data = RREG32_SOC15(GC, 0, regCP_MES_CNTL);
+			if (pipe == 0)
+				data = REG_SET_FIELD(data, CP_MES_CNTL, MES_PIPE0_RESET, 1);
+			else
+				data = REG_SET_FIELD(data, CP_MES_CNTL, MES_PIPE1_RESET, 1);
+			WREG32_SOC15(GC, 0, regCP_MES_CNTL, data);
+
+			ucode_addr = adev->mes.uc_start_addr[pipe] >> 2;
+			WREG32_SOC15(GC, 0, regCP_MES_PRGRM_CNTR_START,
+				     lower_32_bits(ucode_addr));
+			WREG32_SOC15(GC, 0, regCP_MES_PRGRM_CNTR_START_HI,
+				     upper_32_bits(ucode_addr));
+
+			/* unhalt MES and activate one pipe each loop */
+			data = REG_SET_FIELD(0, CP_MES_CNTL, MES_PIPE0_ACTIVE, 1);
+			if (pipe)
+				data = REG_SET_FIELD(data, CP_MES_CNTL, MES_PIPE1_ACTIVE, 1);
+			dev_info(adev->dev, "program CP_MES_CNTL : 0x%x\n", data);
+
+			WREG32_SOC15(GC, 0, regCP_MES_CNTL, data);
+
+		}
+		soc21_grbm_select(adev, 0, 0, 0, 0);
+		mutex_unlock(&adev->srbm_mutex);
+
+		if (amdgpu_emu_mode)
+			msleep(100);
+		else if (adev->enable_uni_mes)
+			udelay(500);
+		else
+			udelay(50);
+	} else {
+		data = RREG32_SOC15(GC, 0, regCP_MES_CNTL);
+		data = REG_SET_FIELD(data, CP_MES_CNTL, MES_PIPE0_ACTIVE, 0);
+		data = REG_SET_FIELD(data, CP_MES_CNTL, MES_PIPE1_ACTIVE, 0);
+		data = REG_SET_FIELD(data, CP_MES_CNTL,
+				     MES_INVALIDATE_ICACHE, 1);
+		data = REG_SET_FIELD(data, CP_MES_CNTL, MES_PIPE0_RESET, 1);
+		data = REG_SET_FIELD(data, CP_MES_CNTL, MES_PIPE1_RESET, 1);
+		data = REG_SET_FIELD(data, CP_MES_CNTL, MES_HALT, 1);
+		WREG32_SOC15(GC, 0, regCP_MES_CNTL, data);
+	}
+}
+
+static void mes_v12_0_set_ucode_start_addr(struct amdgpu_device *adev)
+{
+	uint64_t ucode_addr;
+	int pipe;
+
+	mes_v12_0_enable(adev, false);
+
+	mutex_lock(&adev->srbm_mutex);
+	for (pipe = 0; pipe < AMDGPU_MAX_MES_PIPES; pipe++) {
+		/* me=3, queue=0 */
+		soc21_grbm_select(adev, 3, pipe, 0, 0);
+
+		/* set ucode start address */
+		ucode_addr = adev->mes.uc_start_addr[pipe] >> 2;
+		WREG32_SOC15(GC, 0, regCP_MES_PRGRM_CNTR_START,
+				lower_32_bits(ucode_addr));
+		WREG32_SOC15(GC, 0, regCP_MES_PRGRM_CNTR_START_HI,
+				upper_32_bits(ucode_addr));
+
+		soc21_grbm_select(adev, 0, 0, 0, 0);
+	}
+	mutex_unlock(&adev->srbm_mutex);
+}
+
+/* This function is for backdoor MES firmware */
+static int mes_v12_0_load_microcode(struct amdgpu_device *adev,
+				    enum amdgpu_mes_pipe pipe, bool prime_icache)
+{
+	int r;
+	uint32_t data;
+
+	mes_v12_0_enable(adev, false);
+
+	if (!adev->mes.fw[pipe])
+		return -EINVAL;
+
+	r = mes_v12_0_allocate_ucode_buffer(adev, pipe);
+	if (r)
+		return r;
+
+	r = mes_v12_0_allocate_ucode_data_buffer(adev, pipe);
+	if (r) {
+		mes_v12_0_free_ucode_buffers(adev, pipe);
+		return r;
+	}
+
+	mutex_lock(&adev->srbm_mutex);
+	/* me=3, pipe=0, queue=0 */
+	soc21_grbm_select(adev, 3, pipe, 0, 0);
+
+	WREG32_SOC15(GC, 0, regCP_MES_IC_BASE_CNTL, 0);
+
+	/* set ucode fimrware address */
+	WREG32_SOC15(GC, 0, regCP_MES_IC_BASE_LO,
+		     lower_32_bits(adev->mes.ucode_fw_gpu_addr[pipe]));
+	WREG32_SOC15(GC, 0, regCP_MES_IC_BASE_HI,
+		     upper_32_bits(adev->mes.ucode_fw_gpu_addr[pipe]));
+
+	/* set ucode instruction cache boundary to 2M-1 */
+	WREG32_SOC15(GC, 0, regCP_MES_MIBOUND_LO, 0x1FFFFF);
+
+	/* set ucode data firmware address */
+	WREG32_SOC15(GC, 0, regCP_MES_MDBASE_LO,
+		     lower_32_bits(adev->mes.data_fw_gpu_addr[pipe]));
+	WREG32_SOC15(GC, 0, regCP_MES_MDBASE_HI,
+		     upper_32_bits(adev->mes.data_fw_gpu_addr[pipe]));
+
+	/* Set data cache boundary CP_MES_MDBOUND_LO */
+	WREG32_SOC15(GC, 0, regCP_MES_MDBOUND_LO, 0x7FFFF);
+
+	if (prime_icache) {
+		/* invalidate ICACHE */
+		data = RREG32_SOC15(GC, 0, regCP_MES_IC_OP_CNTL);
+		data = REG_SET_FIELD(data, CP_MES_IC_OP_CNTL, PRIME_ICACHE, 0);
+		data = REG_SET_FIELD(data, CP_MES_IC_OP_CNTL, INVALIDATE_CACHE, 1);
+		WREG32_SOC15(GC, 0, regCP_MES_IC_OP_CNTL, data);
+
+		/* prime the ICACHE. */
+		data = RREG32_SOC15(GC, 0, regCP_MES_IC_OP_CNTL);
+		data = REG_SET_FIELD(data, CP_MES_IC_OP_CNTL, PRIME_ICACHE, 1);
+		WREG32_SOC15(GC, 0, regCP_MES_IC_OP_CNTL, data);
+	}
+
+	soc21_grbm_select(adev, 0, 0, 0, 0);
+	mutex_unlock(&adev->srbm_mutex);
+
+	return 0;
+}
+
+static int mes_v12_0_allocate_eop_buf(struct amdgpu_device *adev,
+				      enum amdgpu_mes_pipe pipe)
+{
+	int r;
+	u32 *eop;
+
+	r = amdgpu_bo_create_reserved(adev, MES_EOP_SIZE, PAGE_SIZE,
+			      AMDGPU_GEM_DOMAIN_GTT,
+			      &adev->mes.eop_gpu_obj[pipe],
+			      &adev->mes.eop_gpu_addr[pipe],
+			      (void **)&eop);
+	if (r) {
+		dev_warn(adev->dev, "(%d) create EOP bo failed\n", r);
+		return r;
+	}
+
+	memset(eop, 0,
+	       adev->mes.eop_gpu_obj[pipe]->tbo.base.size);
+
+	amdgpu_bo_kunmap(adev->mes.eop_gpu_obj[pipe]);
+	amdgpu_bo_unreserve(adev->mes.eop_gpu_obj[pipe]);
+
+	return 0;
+}
+
+static int mes_v12_0_mqd_init(struct amdgpu_ring *ring)
+{
+	struct v12_compute_mqd *mqd = ring->mqd_ptr;
+	uint64_t hqd_gpu_addr, wb_gpu_addr, eop_base_addr;
+	uint32_t tmp;
+
+	mqd->header = 0xC0310800;
+	mqd->compute_pipelinestat_enable = 0x00000001;
+	mqd->compute_static_thread_mgmt_se0 = 0xffffffff;
+	mqd->compute_static_thread_mgmt_se1 = 0xffffffff;
+	mqd->compute_static_thread_mgmt_se2 = 0xffffffff;
+	mqd->compute_static_thread_mgmt_se3 = 0xffffffff;
+	mqd->compute_misc_reserved = 0x00000007;
+
+	eop_base_addr = ring->eop_gpu_addr >> 8;
+
+	/* set the EOP size, register value is 2^(EOP_SIZE+1) dwords */
+	tmp = regCP_HQD_EOP_CONTROL_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, CP_HQD_EOP_CONTROL, EOP_SIZE,
+			(order_base_2(MES_EOP_SIZE / 4) - 1));
+
+	mqd->cp_hqd_eop_base_addr_lo = lower_32_bits(eop_base_addr);
+	mqd->cp_hqd_eop_base_addr_hi = upper_32_bits(eop_base_addr);
+	mqd->cp_hqd_eop_control = tmp;
+
+	/* disable the queue if it's active */
+	ring->wptr = 0;
+	mqd->cp_hqd_pq_rptr = 0;
+	mqd->cp_hqd_pq_wptr_lo = 0;
+	mqd->cp_hqd_pq_wptr_hi = 0;
+
+	/* set the pointer to the MQD */
+	mqd->cp_mqd_base_addr_lo = ring->mqd_gpu_addr & 0xfffffffc;
+	mqd->cp_mqd_base_addr_hi = upper_32_bits(ring->mqd_gpu_addr);
+
+	/* set MQD vmid to 0 */
+	tmp = regCP_MQD_CONTROL_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, CP_MQD_CONTROL, VMID, 0);
+	mqd->cp_mqd_control = tmp;
+
+	/* set the pointer to the HQD, this is similar CP_RB0_BASE/_HI */
+	hqd_gpu_addr = ring->gpu_addr >> 8;
+	mqd->cp_hqd_pq_base_lo = lower_32_bits(hqd_gpu_addr);
+	mqd->cp_hqd_pq_base_hi = upper_32_bits(hqd_gpu_addr);
+
+	/* set the wb address whether it's enabled or not */
+	wb_gpu_addr = ring->rptr_gpu_addr;
+	mqd->cp_hqd_pq_rptr_report_addr_lo = wb_gpu_addr & 0xfffffffc;
+	mqd->cp_hqd_pq_rptr_report_addr_hi =
+		upper_32_bits(wb_gpu_addr) & 0xffff;
+
+	/* only used if CP_PQ_WPTR_POLL_CNTL.CP_PQ_WPTR_POLL_CNTL__EN_MASK=1 */
+	wb_gpu_addr = ring->wptr_gpu_addr;
+	mqd->cp_hqd_pq_wptr_poll_addr_lo = wb_gpu_addr & 0xfffffff8;
+	mqd->cp_hqd_pq_wptr_poll_addr_hi = upper_32_bits(wb_gpu_addr) & 0xffff;
+
+	/* set up the HQD, this is similar to CP_RB0_CNTL */
+	tmp = regCP_HQD_PQ_CONTROL_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, QUEUE_SIZE,
+			    (order_base_2(ring->ring_size / 4) - 1));
+	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, RPTR_BLOCK_SIZE,
+			    ((order_base_2(AMDGPU_GPU_PAGE_SIZE / 4) - 1) << 8));
+	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, UNORD_DISPATCH, 1);
+	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, TUNNEL_DISPATCH, 0);
+	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, PRIV_STATE, 1);
+	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, KMD_QUEUE, 1);
+	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, NO_UPDATE_RPTR, 1);
+	mqd->cp_hqd_pq_control = tmp;
+
+	/* enable doorbell */
+	tmp = 0;
+	if (ring->use_doorbell) {
+		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
+				    DOORBELL_OFFSET, ring->doorbell_index);
+		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
+				    DOORBELL_EN, 1);
+		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
+				    DOORBELL_SOURCE, 0);
+		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
+				    DOORBELL_HIT, 0);
+	} else {
+		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
+				    DOORBELL_EN, 0);
+	}
+	mqd->cp_hqd_pq_doorbell_control = tmp;
+
+	mqd->cp_hqd_vmid = 0;
+	/* activate the queue */
+	mqd->cp_hqd_active = 1;
+
+	tmp = regCP_HQD_PERSISTENT_STATE_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, CP_HQD_PERSISTENT_STATE,
+			    PRELOAD_SIZE, 0x55);
+	mqd->cp_hqd_persistent_state = tmp;
+
+	mqd->cp_hqd_ib_control = regCP_HQD_IB_CONTROL_DEFAULT;
+	mqd->cp_hqd_iq_timer = regCP_HQD_IQ_TIMER_DEFAULT;
+	mqd->cp_hqd_quantum = regCP_HQD_QUANTUM_DEFAULT;
+
+	/*
+	 * Set CP_HQD_GFX_CONTROL.DB_UPDATED_MSG_EN[15] to enable unmapped
+	 * doorbell handling. This is a reserved CP internal register can
+	 * not be accesss by others
+	 */
+	mqd->reserved_184 = BIT(15);
+
+	return 0;
+}
+
+static void mes_v12_0_queue_init_register(struct amdgpu_ring *ring)
+{
+	struct v12_compute_mqd *mqd = ring->mqd_ptr;
+	struct amdgpu_device *adev = ring->adev;
+	uint32_t data = 0;
+
+	mutex_lock(&adev->srbm_mutex);
+	soc21_grbm_select(adev, 3, ring->pipe, 0, 0);
+
+	/* set CP_HQD_VMID.VMID = 0. */
+	data = RREG32_SOC15(GC, 0, regCP_HQD_VMID);
+	data = REG_SET_FIELD(data, CP_HQD_VMID, VMID, 0);
+	WREG32_SOC15(GC, 0, regCP_HQD_VMID, data);
+
+	/* set CP_HQD_PQ_DOORBELL_CONTROL.DOORBELL_EN=0 */
+	data = RREG32_SOC15(GC, 0, regCP_HQD_PQ_DOORBELL_CONTROL);
+	data = REG_SET_FIELD(data, CP_HQD_PQ_DOORBELL_CONTROL,
+			     DOORBELL_EN, 0);
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_DOORBELL_CONTROL, data);
+
+	/* set CP_MQD_BASE_ADDR/HI with the MQD base address */
+	WREG32_SOC15(GC, 0, regCP_MQD_BASE_ADDR, mqd->cp_mqd_base_addr_lo);
+	WREG32_SOC15(GC, 0, regCP_MQD_BASE_ADDR_HI, mqd->cp_mqd_base_addr_hi);
+
+	/* set CP_MQD_CONTROL.VMID=0 */
+	data = RREG32_SOC15(GC, 0, regCP_MQD_CONTROL);
+	data = REG_SET_FIELD(data, CP_MQD_CONTROL, VMID, 0);
+	WREG32_SOC15(GC, 0, regCP_MQD_CONTROL, 0);
+
+	/* set CP_HQD_PQ_BASE/HI with the ring buffer base address */
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_BASE, mqd->cp_hqd_pq_base_lo);
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_BASE_HI, mqd->cp_hqd_pq_base_hi);
+
+	/* set CP_HQD_PQ_RPTR_REPORT_ADDR/HI */
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_RPTR_REPORT_ADDR,
+		     mqd->cp_hqd_pq_rptr_report_addr_lo);
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_RPTR_REPORT_ADDR_HI,
+		     mqd->cp_hqd_pq_rptr_report_addr_hi);
+
+	/* set CP_HQD_PQ_CONTROL */
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_CONTROL, mqd->cp_hqd_pq_control);
+
+	/* set CP_HQD_PQ_WPTR_POLL_ADDR/HI */
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_WPTR_POLL_ADDR,
+		     mqd->cp_hqd_pq_wptr_poll_addr_lo);
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_WPTR_POLL_ADDR_HI,
+		     mqd->cp_hqd_pq_wptr_poll_addr_hi);
+
+	/* set CP_HQD_PQ_DOORBELL_CONTROL */
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_DOORBELL_CONTROL,
+		     mqd->cp_hqd_pq_doorbell_control);
+
+	/* set CP_HQD_PERSISTENT_STATE.PRELOAD_SIZE=0x53 */
+	WREG32_SOC15(GC, 0, regCP_HQD_PERSISTENT_STATE, mqd->cp_hqd_persistent_state);
+
+	/* set CP_HQD_ACTIVE.ACTIVE=1 */
+	WREG32_SOC15(GC, 0, regCP_HQD_ACTIVE, mqd->cp_hqd_active);
+
+	soc21_grbm_select(adev, 0, 0, 0, 0);
+	mutex_unlock(&adev->srbm_mutex);
+}
+
+static int mes_v12_0_kiq_enable_queue(struct amdgpu_device *adev)
+{
+	struct amdgpu_kiq *kiq = &adev->gfx.kiq[0];
+	struct amdgpu_ring *kiq_ring = &adev->gfx.kiq[0].ring;
+	int r;
+
+	if (!kiq->pmf || !kiq->pmf->kiq_map_queues)
+		return -EINVAL;
+
+	r = amdgpu_ring_alloc(kiq_ring, kiq->pmf->map_queues_size);
+	if (r) {
+		DRM_ERROR("Failed to lock KIQ (%d).\n", r);
+		return r;
+	}
+
+	kiq->pmf->kiq_map_queues(kiq_ring, &adev->mes.ring[0]);
+
+	r = amdgpu_ring_test_ring(kiq_ring);
+	if (r) {
+		DRM_ERROR("kfq enable failed\n");
+		kiq_ring->sched.ready = false;
+	}
+	return r;
+}
+
+static int mes_v12_0_queue_init(struct amdgpu_device *adev,
+				enum amdgpu_mes_pipe pipe)
+{
+	struct amdgpu_ring *ring;
+	int r;
+
+	if (!adev->enable_uni_mes && pipe == AMDGPU_MES_KIQ_PIPE)
+		ring = &adev->gfx.kiq[0].ring;
+	else
+		ring = &adev->mes.ring[pipe];
+
+	if ((adev->enable_uni_mes || pipe == AMDGPU_MES_SCHED_PIPE) &&
+	    (amdgpu_in_reset(adev) || adev->in_suspend)) {
+		*(ring->wptr_cpu_addr) = 0;
+		*(ring->rptr_cpu_addr) = 0;
+		amdgpu_ring_clear_ring(ring);
+	}
+
+	r = mes_v12_0_mqd_init(ring);
+	if (r)
+		return r;
+
+	if (pipe == AMDGPU_MES_SCHED_PIPE) {
+		if (adev->enable_uni_mes)
+			r = amdgpu_mes_map_legacy_queue(adev, ring);
+		else
+			r = mes_v12_0_kiq_enable_queue(adev);
+		if (r)
+			return r;
+	} else {
+		mes_v12_0_queue_init_register(ring);
+	}
+
+	if (((pipe == AMDGPU_MES_SCHED_PIPE) && !adev->mes.sched_version) ||
+	    ((pipe == AMDGPU_MES_KIQ_PIPE) && !adev->mes.kiq_version)) {
+		/* get MES scheduler/KIQ versions */
+		mutex_lock(&adev->srbm_mutex);
+		soc21_grbm_select(adev, 3, pipe, 0, 0);
+
+		if (pipe == AMDGPU_MES_SCHED_PIPE)
+			adev->mes.sched_version = RREG32_SOC15(GC, 0, regCP_MES_GP3_LO);
+		else if (pipe == AMDGPU_MES_KIQ_PIPE && adev->enable_mes_kiq)
+			adev->mes.kiq_version = RREG32_SOC15(GC, 0, regCP_MES_GP3_LO);
+
+		soc21_grbm_select(adev, 0, 0, 0, 0);
+		mutex_unlock(&adev->srbm_mutex);
+	}
+
+	return 0;
+}
+
+static int mes_v12_0_ring_init(struct amdgpu_device *adev, int pipe)
+{
+	struct amdgpu_ring *ring;
+
+	ring = &adev->mes.ring[pipe];
+
+	ring->funcs = &mes_v12_0_ring_funcs;
+
+	ring->me = 3;
+	ring->pipe = pipe;
+	ring->queue = 0;
+
+	ring->ring_obj = NULL;
+	ring->use_doorbell = true;
+	ring->eop_gpu_addr = adev->mes.eop_gpu_addr[pipe];
+	ring->no_scheduler = true;
+	sprintf(ring->name, "mes_%d.%d.%d", ring->me, ring->pipe, ring->queue);
+
+	if (pipe == AMDGPU_MES_SCHED_PIPE)
+		ring->doorbell_index = adev->doorbell_index.mes_ring0 << 1;
+	else
+		ring->doorbell_index = adev->doorbell_index.mes_ring1 << 1;
+
+	return amdgpu_ring_init(adev, ring, 1024, NULL, 0,
+				AMDGPU_RING_PRIO_DEFAULT, NULL);
+}
+
+static int mes_v12_0_kiq_ring_init(struct amdgpu_device *adev)
+{
+	struct amdgpu_ring *ring;
+
+	spin_lock_init(&adev->gfx.kiq[0].ring_lock);
+
+	ring = &adev->gfx.kiq[0].ring;
+
+	ring->me = 3;
+	ring->pipe = 1;
+	ring->queue = 0;
+
+	ring->adev = NULL;
+	ring->ring_obj = NULL;
+	ring->use_doorbell = true;
+	ring->doorbell_index = adev->doorbell_index.mes_ring1 << 1;
+	ring->eop_gpu_addr = adev->mes.eop_gpu_addr[AMDGPU_MES_KIQ_PIPE];
+	ring->no_scheduler = true;
+	sprintf(ring->name, "mes_kiq_%d.%d.%d",
+		ring->me, ring->pipe, ring->queue);
+
+	return amdgpu_ring_init(adev, ring, 1024, NULL, 0,
+				AMDGPU_RING_PRIO_DEFAULT, NULL);
+}
+
+static int mes_v12_0_mqd_sw_init(struct amdgpu_device *adev,
+				 enum amdgpu_mes_pipe pipe)
+{
+	int r, mqd_size = sizeof(struct v12_compute_mqd);
+	struct amdgpu_ring *ring;
+
+	if (!adev->enable_uni_mes && pipe == AMDGPU_MES_KIQ_PIPE)
+		ring = &adev->gfx.kiq[0].ring;
+	else
+		ring = &adev->mes.ring[pipe];
+
+	if (ring->mqd_obj)
+		return 0;
+
+	r = amdgpu_bo_create_kernel(adev, mqd_size, PAGE_SIZE,
+				    AMDGPU_GEM_DOMAIN_GTT, &ring->mqd_obj,
+				    &ring->mqd_gpu_addr, &ring->mqd_ptr);
+	if (r) {
+		dev_warn(adev->dev, "failed to create ring mqd bo (%d)", r);
+		return r;
+	}
+
+	memset(ring->mqd_ptr, 0, mqd_size);
+
+	/* prepare MQD backup */
+	adev->mes.mqd_backup[pipe] = kmalloc(mqd_size, GFP_KERNEL);
+	if (!adev->mes.mqd_backup[pipe])
+		dev_warn(adev->dev,
+			 "no memory to create MQD backup for ring %s\n",
+			 ring->name);
+
+	return 0;
+}
+
+static int mes_v12_0_sw_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int pipe, r;
+
+	adev->mes.funcs = &mes_v12_0_funcs;
+	adev->mes.kiq_hw_init = &mes_v12_0_kiq_hw_init;
+	adev->mes.kiq_hw_fini = &mes_v12_0_kiq_hw_fini;
+	adev->mes.enable_legacy_queue_map = true;
+
+	adev->mes.event_log_size = adev->enable_uni_mes ?
+		(AMDGPU_MAX_MES_PIPES * (AMDGPU_MES_LOG_BUFFER_SIZE + AMDGPU_MES_MSCRATCH_SIZE)) :
+		(AMDGPU_MES_LOG_BUFFER_SIZE + AMDGPU_MES_MSCRATCH_SIZE);
+	r = amdgpu_mes_init(adev);
+	if (r)
+		return r;
+
+	for (pipe = 0; pipe < AMDGPU_MAX_MES_PIPES; pipe++) {
+		r = mes_v12_0_allocate_eop_buf(adev, pipe);
+		if (r)
+			return r;
+
+		r = mes_v12_0_mqd_sw_init(adev, pipe);
+		if (r)
+			return r;
+
+		if (!adev->enable_uni_mes && pipe == AMDGPU_MES_KIQ_PIPE) {
+			r = mes_v12_0_kiq_ring_init(adev);
+		}
+		else {
+			r = mes_v12_0_ring_init(adev, pipe);
+			if (r)
+				return r;
+			r = amdgpu_bo_create_kernel(adev, AMDGPU_GPU_PAGE_SIZE, PAGE_SIZE,
+						    AMDGPU_GEM_DOMAIN_VRAM,
+						    &adev->mes.resource_1[pipe],
+						    &adev->mes.resource_1_gpu_addr[pipe],
+						    &adev->mes.resource_1_addr[pipe]);
+			if (r) {
+				dev_err(adev->dev, "(%d) failed to create mes resource_1 bo pipe[%d]\n", r, pipe);
+				return r;
+			}
+		}
+	}
+
+	return 0;
+}
+
+static int mes_v12_0_sw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int pipe;
+
+	for (pipe = 0; pipe < AMDGPU_MAX_MES_PIPES; pipe++) {
+		amdgpu_bo_free_kernel(&adev->mes.resource_1[pipe],
+				      &adev->mes.resource_1_gpu_addr[pipe],
+				      &adev->mes.resource_1_addr[pipe]);
+
+		kfree(adev->mes.mqd_backup[pipe]);
+
+		amdgpu_bo_free_kernel(&adev->mes.eop_gpu_obj[pipe],
+				      &adev->mes.eop_gpu_addr[pipe],
+				      NULL);
+		amdgpu_ucode_release(&adev->mes.fw[pipe]);
+
+		if (adev->enable_uni_mes || pipe == AMDGPU_MES_SCHED_PIPE) {
+			amdgpu_bo_free_kernel(&adev->mes.ring[pipe].mqd_obj,
+					      &adev->mes.ring[pipe].mqd_gpu_addr,
+					      &adev->mes.ring[pipe].mqd_ptr);
+			amdgpu_ring_fini(&adev->mes.ring[pipe]);
+		}
+	}
+
+	if (!adev->enable_uni_mes) {
+		amdgpu_bo_free_kernel(&adev->gfx.kiq[0].ring.mqd_obj,
+				      &adev->gfx.kiq[0].ring.mqd_gpu_addr,
+				      &adev->gfx.kiq[0].ring.mqd_ptr);
+		amdgpu_ring_fini(&adev->gfx.kiq[0].ring);
+	}
+
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT) {
+		mes_v12_0_free_ucode_buffers(adev, AMDGPU_MES_KIQ_PIPE);
+		mes_v12_0_free_ucode_buffers(adev, AMDGPU_MES_SCHED_PIPE);
+	}
+
+	amdgpu_mes_fini(adev);
+	return 0;
+}
+
+static void mes_v12_0_kiq_dequeue_sched(struct amdgpu_device *adev)
+{
+	uint32_t data;
+	int i;
+
+	mutex_lock(&adev->srbm_mutex);
+	soc21_grbm_select(adev, 3, AMDGPU_MES_SCHED_PIPE, 0, 0);
+
+	/* disable the queue if it's active */
+	if (RREG32_SOC15(GC, 0, regCP_HQD_ACTIVE) & 1) {
+		WREG32_SOC15(GC, 0, regCP_HQD_DEQUEUE_REQUEST, 1);
+		for (i = 0; i < adev->usec_timeout; i++) {
+			if (!(RREG32_SOC15(GC, 0, regCP_HQD_ACTIVE) & 1))
+				break;
+			udelay(1);
+		}
+	}
+	data = RREG32_SOC15(GC, 0, regCP_HQD_PQ_DOORBELL_CONTROL);
+	data = REG_SET_FIELD(data, CP_HQD_PQ_DOORBELL_CONTROL,
+				DOORBELL_EN, 0);
+	data = REG_SET_FIELD(data, CP_HQD_PQ_DOORBELL_CONTROL,
+				DOORBELL_HIT, 1);
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_DOORBELL_CONTROL, data);
+
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_DOORBELL_CONTROL, 0);
+
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_WPTR_LO, 0);
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_WPTR_HI, 0);
+	WREG32_SOC15(GC, 0, regCP_HQD_PQ_RPTR, 0);
+
+	soc21_grbm_select(adev, 0, 0, 0, 0);
+	mutex_unlock(&adev->srbm_mutex);
+
+	adev->mes.ring[0].sched.ready = false;
+}
+
+static void mes_v12_0_kiq_setting(struct amdgpu_ring *ring)
+{
+	uint32_t tmp;
+	struct amdgpu_device *adev = ring->adev;
+
+	/* tell RLC which is KIQ queue */
+	tmp = RREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS);
+	tmp &= 0xffffff00;
+	tmp |= (ring->me << 5) | (ring->pipe << 3) | (ring->queue);
+	WREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS, tmp | 0x80);
+}
+
+static int mes_v12_0_kiq_hw_init(struct amdgpu_device *adev)
+{
+	int r = 0;
+	struct amdgpu_ip_block *ip_block;
+
+	if (adev->enable_uni_mes)
+		mes_v12_0_kiq_setting(&adev->mes.ring[AMDGPU_MES_KIQ_PIPE]);
+	else
+		mes_v12_0_kiq_setting(&adev->gfx.kiq[0].ring);
+
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT) {
+
+		r = mes_v12_0_load_microcode(adev, AMDGPU_MES_SCHED_PIPE, false);
+		if (r) {
+			DRM_ERROR("failed to load MES fw, r=%d\n", r);
+			return r;
+		}
+
+		r = mes_v12_0_load_microcode(adev, AMDGPU_MES_KIQ_PIPE, true);
+		if (r) {
+			DRM_ERROR("failed to load MES kiq fw, r=%d\n", r);
+			return r;
+		}
+
+		mes_v12_0_set_ucode_start_addr(adev);
+
+	} else if (adev->firmware.load_type == AMDGPU_FW_LOAD_RLC_BACKDOOR_AUTO)
+		mes_v12_0_set_ucode_start_addr(adev);
+
+	mes_v12_0_enable(adev, true);
+
+	ip_block = amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_MES);
+	if (unlikely(!ip_block)) {
+		dev_err(adev->dev, "Failed to get MES handle\n");
+		return -EINVAL;
+	}
+
+	r = mes_v12_0_queue_init(adev, AMDGPU_MES_KIQ_PIPE);
+	if (r)
+		goto failure;
+
+	if (adev->enable_uni_mes) {
+		r = mes_v12_0_set_hw_resources(&adev->mes, AMDGPU_MES_KIQ_PIPE);
+		if (r)
+			goto failure;
+
+		mes_v12_0_set_hw_resources_1(&adev->mes, AMDGPU_MES_KIQ_PIPE);
+	}
+
+	if (adev->mes.enable_legacy_queue_map) {
+		r = mes_v12_0_hw_init(ip_block);
+		if (r)
+			goto failure;
+	}
+
+	return r;
+
+failure:
+	mes_v12_0_hw_fini(ip_block);
+	return r;
+}
+
+static int mes_v12_0_kiq_hw_fini(struct amdgpu_device *adev)
+{
+	if (adev->mes.ring[0].sched.ready) {
+		if (adev->enable_uni_mes)
+			amdgpu_mes_unmap_legacy_queue(adev,
+				      &adev->mes.ring[AMDGPU_MES_SCHED_PIPE],
+				      RESET_QUEUES, 0, 0);
+		else
+			mes_v12_0_kiq_dequeue_sched(adev);
+
+		adev->mes.ring[0].sched.ready = false;
+	}
+
+	mes_v12_0_enable(adev, false);
+
+	return 0;
+}
+
+static int mes_v12_0_hw_init(struct amdgpu_ip_block *ip_block)
+{
+	int r;
+	struct amdgpu_device *adev = ip_block->adev;
+
+	if (adev->mes.ring[0].sched.ready)
+		goto out;
+
+	if (!adev->enable_mes_kiq) {
+		if (adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT) {
+			r = mes_v12_0_load_microcode(adev,
+					     AMDGPU_MES_SCHED_PIPE, true);
+			if (r) {
+				DRM_ERROR("failed to MES fw, r=%d\n", r);
+				return r;
+			}
+
+			mes_v12_0_set_ucode_start_addr(adev);
+
+		} else if (adev->firmware.load_type ==
+			   AMDGPU_FW_LOAD_RLC_BACKDOOR_AUTO) {
+
+			mes_v12_0_set_ucode_start_addr(adev);
+		}
+
+		mes_v12_0_enable(adev, true);
+	}
+
+	/* Enable the MES to handle doorbell ring on unmapped queue */
+	mes_v12_0_enable_unmapped_doorbell_handling(&adev->mes, true);
+
+	r = mes_v12_0_queue_init(adev, AMDGPU_MES_SCHED_PIPE);
+	if (r)
+		goto failure;
+
+	r = mes_v12_0_set_hw_resources(&adev->mes, AMDGPU_MES_SCHED_PIPE);
+	if (r)
+		goto failure;
+
+	if ((adev->mes.sched_version & AMDGPU_MES_VERSION_MASK) >= 0x4b)
+		mes_v12_0_set_hw_resources_1(&adev->mes, AMDGPU_MES_SCHED_PIPE);
+
+	mes_v12_0_init_aggregated_doorbell(&adev->mes);
+
+	r = mes_v12_0_query_sched_status(&adev->mes, AMDGPU_MES_SCHED_PIPE);
+	if (r) {
+		DRM_ERROR("MES is busy\n");
+		goto failure;
+	}
+
+	r = amdgpu_mes_update_enforce_isolation(adev);
+	if (r)
+		goto failure;
+
+out:
+	/*
+	 * Disable KIQ ring usage from the driver once MES is enabled.
+	 * MES uses KIQ ring exclusively so driver cannot access KIQ ring
+	 * with MES enabled.
+	 */
+	adev->gfx.kiq[0].ring.sched.ready = false;
+	adev->mes.ring[0].sched.ready = true;
+
+	return 0;
+
+failure:
+	mes_v12_0_hw_fini(ip_block);
+	return r;
+}
+
+static int mes_v12_0_hw_fini(struct amdgpu_ip_block *ip_block)
+{
+	return 0;
+}
+
+static int mes_v12_0_suspend(struct amdgpu_ip_block *ip_block)
+{
+	return mes_v12_0_hw_fini(ip_block);
+}
+
+static int mes_v12_0_resume(struct amdgpu_ip_block *ip_block)
+{
+	return mes_v12_0_hw_init(ip_block);
+}
+
+static int mes_v12_0_early_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int pipe, r;
+
+	adev->mes.hung_queue_db_array_size = MES12_HUNG_DB_OFFSET_ARRAY_SIZE;
+	adev->mes.hung_queue_hqd_info_offset = MES12_HUNG_HQD_INFO_OFFSET;
+
+	for (pipe = 0; pipe < AMDGPU_MAX_MES_PIPES; pipe++) {
+		r = amdgpu_mes_init_microcode(adev, pipe);
+		if (r)
+			return r;
+	}
+
+	return 0;
+}
+
+static const struct amd_ip_funcs mes_v12_0_ip_funcs = {
+	.name = "mes_v12_0",
+	.early_init = mes_v12_0_early_init,
+	.late_init = NULL,
+	.sw_init = mes_v12_0_sw_init,
+	.sw_fini = mes_v12_0_sw_fini,
+	.hw_init = mes_v12_0_hw_init,
+	.hw_fini = mes_v12_0_hw_fini,
+	.suspend = mes_v12_0_suspend,
+	.resume = mes_v12_0_resume,
+};
+
+const struct amdgpu_ip_block_version mes_v12_0_ip_block = {
+	.type = AMD_IP_BLOCK_TYPE_MES,
+	.major = 12,
+	.minor = 0,
+	.rev = 0,
+	.funcs = &mes_v12_0_ip_funcs,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/mes_v12_0.h b/drivers/gpu/drm/amd/amdgpu/mes_v12_0.h
new file mode 100644
index 000000000000..ac3740f353aa
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/mes_v12_0.h
@@ -0,0 +1,29 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __MES_V12_0_H__
+#define __MES_V12_0_H__
+
+extern const struct amdgpu_ip_block_version mes_v12_0_ip_block;
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c
index 1da2ec692057..243eabda0607 100644
--- a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c
@@ -54,7 +54,7 @@ static u64 mmhub_v1_0_get_fb_location(struct amdgpu_device *adev)
 static void mmhub_v1_0_setup_vm_pt_regs(struct amdgpu_device *adev, uint32_t vmid,
 				uint64_t page_table_base)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0];
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
 
 	WREG32_SOC15_OFFSET(MMHUB, 0, mmVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32,
 			    hub->ctx_addr_distance * vmid,
@@ -96,7 +96,9 @@ static void mmhub_v1_0_init_system_aperture_regs(struct amdgpu_device *adev)
 	WREG32_SOC15(MMHUB, 0, mmMC_VM_SYSTEM_APERTURE_LOW_ADDR,
 		     min(adev->gmc.fb_start, adev->gmc.agp_start) >> 18);
 
-	if (adev->apu_flags & AMD_APU_IS_RAVEN2)
+	if (adev->apu_flags & (AMD_APU_IS_RAVEN2 |
+			       AMD_APU_IS_RENOIR |
+			       AMD_APU_IS_GREEN_SARDINE))
 		/*
 		 * Raven2 has a HW issue that it is unable to use the vram which
 		 * is out of MC_VM_SYSTEM_APERTURE_HIGH_ADDR. So here is the
@@ -114,7 +116,7 @@ static void mmhub_v1_0_init_system_aperture_regs(struct amdgpu_device *adev)
 		return;
 
 	/* Set default page address. */
-	value = amdgpu_gmc_vram_mc2pa(adev, adev->vram_scratch.gpu_addr);
+	value = amdgpu_gmc_vram_mc2pa(adev, adev->mem_scratch.gpu_addr);
 	WREG32_SOC15(MMHUB, 0, mmMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB,
 		     (u32)(value >> 12));
 	WREG32_SOC15(MMHUB, 0, mmMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_MSB,
@@ -176,6 +178,7 @@ static void mmhub_v1_0_init_cache_regs(struct amdgpu_device *adev)
 	tmp = REG_SET_FIELD(tmp, VM_L2_CNTL2, INVALIDATE_L2_CACHE, 1);
 	WREG32_SOC15(MMHUB, 0, mmVM_L2_CNTL2, tmp);
 
+	tmp = mmVM_L2_CNTL3_DEFAULT;
 	if (adev->gmc.translate_further) {
 		tmp = REG_SET_FIELD(tmp, VM_L2_CNTL3, BANK_SELECT, 12);
 		tmp = REG_SET_FIELD(tmp, VM_L2_CNTL3,
@@ -226,9 +229,55 @@ static void mmhub_v1_0_disable_identity_aperture(struct amdgpu_device *adev)
 		     0);
 }
 
+static void mmhub_v1_0_init_saw(struct amdgpu_device *adev)
+{
+	uint64_t pt_base = amdgpu_gmc_pd_addr(adev->gart.bo);
+	uint32_t tmp;
+
+	/* VM_9_X_REGISTER_VM_L2_SAW_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32 */
+	WREG32_SOC15(MMHUB, 0, mmVM_L2_SAW_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32,
+			    lower_32_bits(pt_base >> 12));
+
+	/* VM_9_X_REGISTER_VM_L2_SAW_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32 */
+	WREG32_SOC15(MMHUB, 0, mmVM_L2_SAW_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32,
+			    upper_32_bits(pt_base >> 12));
+
+	/* VM_9_X_REGISTER_VM_L2_SAW_CONTEXT0_PAGE_TABLE_START_ADDR_LO32 */
+	WREG32_SOC15(MMHUB, 0, mmVM_L2_SAW_CONTEXT0_PAGE_TABLE_START_ADDR_LO32,
+			    (u32)(adev->gmc.gart_start >> 12));
+
+	/* VM_9_X_REGISTER_VM_L2_SAW_CONTEXT0_PAGE_TABLE_START_ADDR_HI32 */
+	WREG32_SOC15(MMHUB, 0, mmVM_L2_SAW_CONTEXT0_PAGE_TABLE_START_ADDR_HI32,
+			    (u32)(adev->gmc.gart_start >> 44));
+
+	/* VM_9_X_REGISTER_VM_L2_SAW_CONTEXT0_PAGE_TABLE_END_ADDR_LO32 */
+	WREG32_SOC15(MMHUB, 0, mmVM_L2_SAW_CONTEXT0_PAGE_TABLE_END_ADDR_LO32,
+			    (u32)(adev->gmc.gart_end >> 12));
+
+	/* VM_9_X_REGISTER_VM_L2_SAW_CONTEXT0_PAGE_TABLE_END_ADDR_HI32 */
+	WREG32_SOC15(MMHUB, 0, mmVM_L2_SAW_CONTEXT0_PAGE_TABLE_END_ADDR_HI32,
+			    (u32)(adev->gmc.gart_end >> 44));
+
+	/* Program SAW CONTEXT0 CNTL */
+	tmp = RREG32_SOC15(MMHUB, 0, mmVM_L2_SAW_CONTEXT0_CNTL);
+	tmp |= 1 << CONTEXT0_CNTL_ENABLE_OFFSET;
+	tmp &= ~(3 << CONTEXT0_CNTL_PAGE_TABLE_DEPTH_OFFSET);
+	WREG32_SOC15(MMHUB, 0, mmVM_L2_SAW_CONTEXT0_CNTL, tmp);
+
+	/* Disable all Contexts except Context0 */
+	tmp = 0xfffe;
+	WREG32_SOC15(MMHUB, 0, mmVM_L2_SAW_CONTEXTS_DISABLE, tmp);
+
+	/* Program SAW CNTL4 */
+	tmp = RREG32_SOC15(MMHUB, 0, mmVM_L2_SAW_CNTL4);
+	tmp |= 1 << VMC_TAP_PDE_REQUEST_SNOOP_OFFSET;
+	tmp |= 1 << VMC_TAP_PTE_REQUEST_SNOOP_OFFSET;
+	WREG32_SOC15(MMHUB, 0, mmVM_L2_SAW_CNTL4, tmp);
+}
+
 static void mmhub_v1_0_setup_vmid_config(struct amdgpu_device *adev)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0];
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
 	unsigned num_level, block_size;
 	uint32_t tmp;
 	int i;
@@ -241,7 +290,7 @@ static void mmhub_v1_0_setup_vmid_config(struct amdgpu_device *adev)
 		block_size -= 9;
 
 	for (i = 0; i <= 14; i++) {
-		tmp = RREG32_SOC15_OFFSET(MMHUB, 0, mmVM_CONTEXT1_CNTL, i);
+		tmp = RREG32_SOC15_OFFSET(MMHUB, 0, mmVM_CONTEXT1_CNTL, i * hub->ctx_distance);
 		tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, ENABLE_CONTEXT, 1);
 		tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, PAGE_TABLE_DEPTH,
 				    num_level);
@@ -280,11 +329,14 @@ static void mmhub_v1_0_setup_vmid_config(struct amdgpu_device *adev)
 				    i * hub->ctx_addr_distance,
 				    upper_32_bits(adev->vm_manager.max_pfn - 1));
 	}
+
+	if (amdgpu_ip_version(adev, ISP_HWIP, 0))
+		mmhub_v1_0_init_saw(adev);
 }
 
 static void mmhub_v1_0_program_invalidation(struct amdgpu_device *adev)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0];
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
 	unsigned i;
 
 	for (i = 0; i < 18; ++i) {
@@ -304,7 +356,7 @@ static void mmhub_v1_0_update_power_gating(struct amdgpu_device *adev,
 	if (adev->pg_flags & AMD_PG_SUPPORT_MMHUB)
 		amdgpu_dpm_set_powergating_by_smu(adev,
 						  AMD_IP_BLOCK_TYPE_GMC,
-						  enable);
+						  enable, 0);
 }
 
 static int mmhub_v1_0_gart_enable(struct amdgpu_device *adev)
@@ -337,7 +389,7 @@ static int mmhub_v1_0_gart_enable(struct amdgpu_device *adev)
 
 static void mmhub_v1_0_gart_disable(struct amdgpu_device *adev)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0];
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
 	u32 tmp;
 	u32 i;
 
@@ -414,7 +466,7 @@ static void mmhub_v1_0_set_fault_enable_default(struct amdgpu_device *adev, bool
 
 static void mmhub_v1_0_init(struct amdgpu_device *adev)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0];
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
 
 	hub->ctx0_ptb_addr_lo32 =
 		SOC15_REG_OFFSET(MMHUB, 0,
@@ -546,7 +598,7 @@ static int mmhub_v1_0_set_clockgating(struct amdgpu_device *adev,
 	return 0;
 }
 
-static void mmhub_v1_0_get_clockgating(struct amdgpu_device *adev, u32 *flags)
+static void mmhub_v1_0_get_clockgating(struct amdgpu_device *adev, u64 *flags)
 {
 	int data, data1;
 
@@ -774,13 +826,17 @@ static void mmhub_v1_0_reset_ras_error_count(struct amdgpu_device *adev)
 	}
 }
 
-const struct amdgpu_mmhub_ras_funcs mmhub_v1_0_ras_funcs = {
-	.ras_late_init = amdgpu_mmhub_ras_late_init,
-	.ras_fini = amdgpu_mmhub_ras_fini,
+struct amdgpu_ras_block_hw_ops mmhub_v1_0_ras_hw_ops = {
 	.query_ras_error_count = mmhub_v1_0_query_ras_error_count,
 	.reset_ras_error_count = mmhub_v1_0_reset_ras_error_count,
 };
 
+struct amdgpu_mmhub_ras mmhub_v1_0_ras = {
+	.ras_block = {
+		.hw_ops = &mmhub_v1_0_ras_hw_ops,
+	},
+};
+
 const struct amdgpu_mmhub_funcs mmhub_v1_0_funcs = {
 	.get_fb_location = mmhub_v1_0_get_fb_location,
 	.init = mmhub_v1_0_init,
diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.h b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.h
index 4661b094e007..dae7ca48bd8b 100644
--- a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.h
+++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.h
@@ -24,6 +24,6 @@
 #define __MMHUB_V1_0_H__
 
 extern const struct amdgpu_mmhub_funcs mmhub_v1_0_funcs;
-extern const struct amdgpu_mmhub_ras_funcs mmhub_v1_0_ras_funcs;
+extern struct amdgpu_mmhub_ras mmhub_v1_0_ras;
 
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_7.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_7.c
index f5f7181f9af5..2adee2b94c37 100644
--- a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_7.c
+++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_7.c
@@ -54,7 +54,7 @@ static u64 mmhub_v1_7_get_fb_location(struct amdgpu_device *adev)
 static void mmhub_v1_7_setup_vm_pt_regs(struct amdgpu_device *adev, uint32_t vmid,
 				uint64_t page_table_base)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0];
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
 
 	WREG32_SOC15_OFFSET(MMHUB, 0, regVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32,
 			hub->ctx_addr_distance * vmid, lower_32_bits(page_table_base));
@@ -134,7 +134,7 @@ static void mmhub_v1_7_init_system_aperture_regs(struct amdgpu_device *adev)
 	}
 
 	/* Set default page address. */
-	value = amdgpu_gmc_vram_mc2pa(adev, adev->vram_scratch.gpu_addr);
+	value = amdgpu_gmc_vram_mc2pa(adev, adev->mem_scratch.gpu_addr);
 	WREG32_SOC15(MMHUB, 0, regMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB,
 		     (u32)(value >> 12));
 	WREG32_SOC15(MMHUB, 0, regMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_MSB,
@@ -172,6 +172,30 @@ static void mmhub_v1_7_init_tlb_regs(struct amdgpu_device *adev)
 	WREG32_SOC15(MMHUB, 0, regMC_VM_MX_L1_TLB_CNTL, tmp);
 }
 
+/* Set snoop bit for SDMA so that SDMA writes probe-invalidates RW lines */
+static void mmhub_v1_7_init_snoop_override_regs(struct amdgpu_device *adev)
+{
+	uint32_t tmp;
+	int i;
+	uint32_t distance = regDAGB1_WRCLI_GPU_SNOOP_OVERRIDE -
+			    regDAGB0_WRCLI_GPU_SNOOP_OVERRIDE;
+
+	for (i = 0; i < 5; i++) { /* DAGB instances */
+		tmp = RREG32_SOC15_OFFSET(MMHUB, 0,
+			regDAGB0_WRCLI_GPU_SNOOP_OVERRIDE, i * distance);
+		tmp |= (1 << 15); /* SDMA client is BIT15 */
+		WREG32_SOC15_OFFSET(MMHUB, 0,
+			regDAGB0_WRCLI_GPU_SNOOP_OVERRIDE, i * distance, tmp);
+
+		tmp = RREG32_SOC15_OFFSET(MMHUB, 0,
+			regDAGB0_WRCLI_GPU_SNOOP_OVERRIDE_VALUE, i * distance);
+		tmp |= (1 << 15);
+		WREG32_SOC15_OFFSET(MMHUB, 0,
+			regDAGB0_WRCLI_GPU_SNOOP_OVERRIDE_VALUE, i * distance, tmp);
+	}
+
+}
+
 static void mmhub_v1_7_init_cache_regs(struct amdgpu_device *adev)
 {
 	uint32_t tmp;
@@ -261,7 +285,7 @@ static void mmhub_v1_7_disable_identity_aperture(struct amdgpu_device *adev)
 
 static void mmhub_v1_7_setup_vmid_config(struct amdgpu_device *adev)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0];
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
 	unsigned num_level, block_size;
 	uint32_t tmp;
 	int i;
@@ -274,7 +298,7 @@ static void mmhub_v1_7_setup_vmid_config(struct amdgpu_device *adev)
 		block_size -= 9;
 
 	for (i = 0; i <= 14; i++) {
-		tmp = RREG32_SOC15_OFFSET(MMHUB, 0, regVM_CONTEXT1_CNTL, i);
+		tmp = RREG32_SOC15_OFFSET(MMHUB, 0, regVM_CONTEXT1_CNTL, i * hub->ctx_distance);
 		tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, ENABLE_CONTEXT, 1);
 		tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, PAGE_TABLE_DEPTH,
 				    num_level);
@@ -319,7 +343,7 @@ static void mmhub_v1_7_setup_vmid_config(struct amdgpu_device *adev)
 
 static void mmhub_v1_7_program_invalidation(struct amdgpu_device *adev)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0];
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
 	unsigned i;
 
 	for (i = 0; i < 18; ++i) {
@@ -337,6 +361,7 @@ static int mmhub_v1_7_gart_enable(struct amdgpu_device *adev)
 	mmhub_v1_7_init_system_aperture_regs(adev);
 	mmhub_v1_7_init_tlb_regs(adev);
 	mmhub_v1_7_init_cache_regs(adev);
+	mmhub_v1_7_init_snoop_override_regs(adev);
 
 	mmhub_v1_7_enable_system_domain(adev);
 	mmhub_v1_7_disable_identity_aperture(adev);
@@ -348,7 +373,7 @@ static int mmhub_v1_7_gart_enable(struct amdgpu_device *adev)
 
 static void mmhub_v1_7_gart_disable(struct amdgpu_device *adev)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0];
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
 	u32 tmp;
 	u32 i;
 
@@ -425,7 +450,7 @@ static void mmhub_v1_7_set_fault_enable_default(struct amdgpu_device *adev, bool
 
 static void mmhub_v1_7_init(struct amdgpu_device *adev)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0];
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
 
 	hub->ctx0_ptb_addr_lo32 =
 		SOC15_REG_OFFSET(MMHUB, 0,
@@ -542,9 +567,9 @@ static int mmhub_v1_7_set_clockgating(struct amdgpu_device *adev,
 	return 0;
 }
 
-static void mmhub_v1_7_get_clockgating(struct amdgpu_device *adev, u32 *flags)
+static void mmhub_v1_7_get_clockgating(struct amdgpu_device *adev, u64 *flags)
 {
-	int data, data1;
+	u32 data, data1;
 
 	if (amdgpu_sriov_vf(adev))
 		*flags = 0;
@@ -1321,15 +1346,19 @@ static void mmhub_v1_7_reset_ras_error_status(struct amdgpu_device *adev)
 	}
 }
 
-const struct amdgpu_mmhub_ras_funcs mmhub_v1_7_ras_funcs = {
-	.ras_late_init = amdgpu_mmhub_ras_late_init,
-	.ras_fini = amdgpu_mmhub_ras_fini,
+struct amdgpu_ras_block_hw_ops mmhub_v1_7_ras_hw_ops = {
 	.query_ras_error_count = mmhub_v1_7_query_ras_error_count,
 	.reset_ras_error_count = mmhub_v1_7_reset_ras_error_count,
 	.query_ras_error_status = mmhub_v1_7_query_ras_error_status,
 	.reset_ras_error_status = mmhub_v1_7_reset_ras_error_status,
 };
 
+struct amdgpu_mmhub_ras mmhub_v1_7_ras = {
+	.ras_block = {
+		.hw_ops = &mmhub_v1_7_ras_hw_ops,
+	},
+};
+
 const struct amdgpu_mmhub_funcs mmhub_v1_7_funcs = {
 	.get_fb_location = mmhub_v1_7_get_fb_location,
 	.init = mmhub_v1_7_init,
diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_7.h b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_7.h
index a7f9dfc24697..629f49052137 100644
--- a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_7.h
+++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_7.h
@@ -24,6 +24,6 @@
 #define __MMHUB_V1_7_H__
 
 extern const struct amdgpu_mmhub_funcs mmhub_v1_7_funcs;
-extern const struct amdgpu_mmhub_ras_funcs mmhub_v1_7_ras_funcs;
+extern struct amdgpu_mmhub_ras mmhub_v1_7_ras;
 
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_8.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_8.c
new file mode 100644
index 000000000000..cc688ae79e84
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_8.c
@@ -0,0 +1,871 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include "amdgpu.h"
+#include "mmhub_v1_8.h"
+
+#include "mmhub/mmhub_1_8_0_offset.h"
+#include "mmhub/mmhub_1_8_0_sh_mask.h"
+#include "vega10_enum.h"
+
+#include "soc15_common.h"
+#include "soc15.h"
+#include "amdgpu_ras.h"
+#include "amdgpu_psp.h"
+
+#define regVM_L2_CNTL3_DEFAULT	0x80100007
+#define regVM_L2_CNTL4_DEFAULT	0x000000c1
+
+static u64 mmhub_v1_8_get_fb_location(struct amdgpu_device *adev)
+{
+	u64 base = RREG32_SOC15(MMHUB, 0, regMC_VM_FB_LOCATION_BASE);
+	u64 top = RREG32_SOC15(MMHUB, 0, regMC_VM_FB_LOCATION_TOP);
+
+	base &= MC_VM_FB_LOCATION_BASE__FB_BASE_MASK;
+	base <<= 24;
+
+	top &= MC_VM_FB_LOCATION_TOP__FB_TOP_MASK;
+	top <<= 24;
+
+	adev->gmc.fb_start = base;
+	adev->gmc.fb_end = top;
+
+	return base;
+}
+
+static void mmhub_v1_8_setup_vm_pt_regs(struct amdgpu_device *adev, uint32_t vmid,
+				uint64_t page_table_base)
+{
+	struct amdgpu_vmhub *hub;
+	u32 inst_mask;
+	int i;
+
+	inst_mask = adev->aid_mask;
+	for_each_inst(i, inst_mask) {
+		hub = &adev->vmhub[AMDGPU_MMHUB0(i)];
+		WREG32_SOC15_OFFSET(MMHUB, i,
+				    regVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32,
+				    hub->ctx_addr_distance * vmid,
+				    lower_32_bits(page_table_base));
+
+		WREG32_SOC15_OFFSET(MMHUB, i,
+				    regVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32,
+				    hub->ctx_addr_distance * vmid,
+				    upper_32_bits(page_table_base));
+	}
+}
+
+static void mmhub_v1_8_init_gart_aperture_regs(struct amdgpu_device *adev)
+{
+	uint64_t gart_start = amdgpu_virt_xgmi_migrate_enabled(adev) ?
+			adev->gmc.vram_start : adev->gmc.fb_start;
+	uint64_t pt_base;
+	u32 inst_mask;
+	int i;
+
+	if (adev->gmc.pdb0_bo)
+		pt_base = amdgpu_gmc_pd_addr(adev->gmc.pdb0_bo);
+	else
+		pt_base = amdgpu_gmc_pd_addr(adev->gart.bo);
+
+	mmhub_v1_8_setup_vm_pt_regs(adev, 0, pt_base);
+
+	/* If use GART for FB translation, vmid0 page table covers both
+	 * vram and system memory (gart)
+	 */
+	inst_mask = adev->aid_mask;
+	for_each_inst(i, inst_mask) {
+		if (adev->gmc.pdb0_bo) {
+			WREG32_SOC15(MMHUB, i,
+				     regVM_CONTEXT0_PAGE_TABLE_START_ADDR_LO32,
+				     (u32)(gart_start >> 12));
+			WREG32_SOC15(MMHUB, i,
+				     regVM_CONTEXT0_PAGE_TABLE_START_ADDR_HI32,
+				     (u32)(gart_start >> 44));
+
+			WREG32_SOC15(MMHUB, i,
+				     regVM_CONTEXT0_PAGE_TABLE_END_ADDR_LO32,
+				     (u32)(adev->gmc.gart_end >> 12));
+			WREG32_SOC15(MMHUB, i,
+				     regVM_CONTEXT0_PAGE_TABLE_END_ADDR_HI32,
+				     (u32)(adev->gmc.gart_end >> 44));
+
+		} else {
+			WREG32_SOC15(MMHUB, i,
+				     regVM_CONTEXT0_PAGE_TABLE_START_ADDR_LO32,
+				     (u32)(adev->gmc.gart_start >> 12));
+			WREG32_SOC15(MMHUB, i,
+				     regVM_CONTEXT0_PAGE_TABLE_START_ADDR_HI32,
+				     (u32)(adev->gmc.gart_start >> 44));
+
+			WREG32_SOC15(MMHUB, i,
+				     regVM_CONTEXT0_PAGE_TABLE_END_ADDR_LO32,
+				     (u32)(adev->gmc.gart_end >> 12));
+			WREG32_SOC15(MMHUB, i,
+				     regVM_CONTEXT0_PAGE_TABLE_END_ADDR_HI32,
+				     (u32)(adev->gmc.gart_end >> 44));
+		}
+	}
+}
+
+static void mmhub_v1_8_init_system_aperture_regs(struct amdgpu_device *adev)
+{
+	uint32_t tmp, inst_mask;
+	uint64_t value;
+	int i;
+
+	if (amdgpu_sriov_vf(adev))
+		return;
+
+	inst_mask = adev->aid_mask;
+	for_each_inst(i, inst_mask) {
+		/* Program the AGP BAR */
+		WREG32_SOC15(MMHUB, i, regMC_VM_AGP_BASE, 0);
+		WREG32_SOC15(MMHUB, i, regMC_VM_AGP_BOT,
+			     adev->gmc.agp_start >> 24);
+		WREG32_SOC15(MMHUB, i, regMC_VM_AGP_TOP,
+			     adev->gmc.agp_end >> 24);
+
+		/* Program the system aperture low logical page number. */
+		WREG32_SOC15(MMHUB, i, regMC_VM_SYSTEM_APERTURE_LOW_ADDR,
+			min(adev->gmc.fb_start, adev->gmc.agp_start) >> 18);
+
+		WREG32_SOC15(MMHUB, i, regMC_VM_SYSTEM_APERTURE_HIGH_ADDR,
+			max(adev->gmc.fb_end, adev->gmc.agp_end) >> 18);
+
+		/* In the case squeezing vram into GART aperture, we don't use
+		 * FB aperture and AGP aperture. Disable them.
+		 */
+		if (adev->gmc.pdb0_bo) {
+			WREG32_SOC15(MMHUB, i, regMC_VM_AGP_BOT, 0xFFFFFF);
+			WREG32_SOC15(MMHUB, i, regMC_VM_AGP_TOP, 0);
+			WREG32_SOC15(MMHUB, i, regMC_VM_FB_LOCATION_TOP, 0);
+			WREG32_SOC15(MMHUB, i, regMC_VM_FB_LOCATION_BASE,
+				     0x00FFFFFF);
+			WREG32_SOC15(MMHUB, i,
+				     regMC_VM_SYSTEM_APERTURE_LOW_ADDR,
+				     0x3FFFFFFF);
+			WREG32_SOC15(MMHUB, i,
+				     regMC_VM_SYSTEM_APERTURE_HIGH_ADDR, 0);
+		}
+
+		/* Set default page address. */
+		value = amdgpu_gmc_vram_mc2pa(adev, adev->mem_scratch.gpu_addr);
+		WREG32_SOC15(MMHUB, i, regMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB,
+			     (u32)(value >> 12));
+		WREG32_SOC15(MMHUB, i, regMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_MSB,
+			     (u32)(value >> 44));
+
+		/* Program "protection fault". */
+		WREG32_SOC15(MMHUB, i,
+			     regVM_L2_PROTECTION_FAULT_DEFAULT_ADDR_LO32,
+			     (u32)(adev->dummy_page_addr >> 12));
+		WREG32_SOC15(MMHUB, i,
+			     regVM_L2_PROTECTION_FAULT_DEFAULT_ADDR_HI32,
+			     (u32)((u64)adev->dummy_page_addr >> 44));
+
+		tmp = RREG32_SOC15(MMHUB, i, regVM_L2_PROTECTION_FAULT_CNTL2);
+		tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL2,
+				    ACTIVE_PAGE_MIGRATION_PTE_READ_RETRY, 1);
+		WREG32_SOC15(MMHUB, i, regVM_L2_PROTECTION_FAULT_CNTL2, tmp);
+	}
+}
+
+static void mmhub_v1_8_init_tlb_regs(struct amdgpu_device *adev)
+{
+	uint32_t tmp, inst_mask;
+	int i;
+
+	if (amdgpu_sriov_reg_indirect_l1_tlb_cntl(adev)) {
+		tmp = RREG32_SOC15(MMHUB, 0, regMC_VM_MX_L1_TLB_CNTL);
+
+		tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB,
+				    1);
+		tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL,
+				    SYSTEM_ACCESS_MODE, 3);
+		tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL,
+				    ENABLE_ADVANCED_DRIVER_MODEL, 1);
+		tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL,
+				    SYSTEM_APERTURE_UNMAPPED_ACCESS, 0);
+		tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL,
+				    MTYPE, MTYPE_UC);/* XXX for emulation. */
+		tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ATC_EN, 1);
+
+		psp_reg_program_no_ring(&adev->psp, tmp, PSP_REG_MMHUB_L1_TLB_CNTL);
+	} else {
+		inst_mask = adev->aid_mask;
+		for_each_inst(i, inst_mask) {
+			tmp = RREG32_SOC15(MMHUB, i, regMC_VM_MX_L1_TLB_CNTL);
+
+			tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB,
+					    1);
+			tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL,
+					    SYSTEM_ACCESS_MODE, 3);
+			tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL,
+					    ENABLE_ADVANCED_DRIVER_MODEL, 1);
+			tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL,
+					    SYSTEM_APERTURE_UNMAPPED_ACCESS, 0);
+			tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL,
+					    MTYPE, MTYPE_UC);/* XXX for emulation. */
+			tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ATC_EN, 1);
+
+			WREG32_SOC15(MMHUB, i, regMC_VM_MX_L1_TLB_CNTL, tmp);
+		}
+	}
+}
+
+/* Set snoop bit for SDMA so that SDMA writes probe-invalidates RW lines */
+static void mmhub_v1_8_init_snoop_override_regs(struct amdgpu_device *adev)
+{
+	uint32_t tmp, inst_mask;
+	int i, j;
+	uint32_t distance = regDAGB1_WRCLI_GPU_SNOOP_OVERRIDE -
+			    regDAGB0_WRCLI_GPU_SNOOP_OVERRIDE;
+
+	if (amdgpu_sriov_vf(adev))
+		return;
+
+	inst_mask = adev->aid_mask;
+	for_each_inst(i, inst_mask) {
+		for (j = 0; j < 5; j++) { /* DAGB instances */
+			tmp = RREG32_SOC15_OFFSET(MMHUB, i,
+				regDAGB0_WRCLI_GPU_SNOOP_OVERRIDE, j * distance);
+			tmp |= (1 << 15); /* SDMA client is BIT15 */
+			WREG32_SOC15_OFFSET(MMHUB, i,
+				regDAGB0_WRCLI_GPU_SNOOP_OVERRIDE, j * distance, tmp);
+
+			tmp = RREG32_SOC15_OFFSET(MMHUB, i,
+				regDAGB0_WRCLI_GPU_SNOOP_OVERRIDE_VALUE, j * distance);
+			tmp |= (1 << 15);
+			WREG32_SOC15_OFFSET(MMHUB, i,
+				regDAGB0_WRCLI_GPU_SNOOP_OVERRIDE_VALUE, j * distance, tmp);
+		}
+	}
+}
+
+static void mmhub_v1_8_init_cache_regs(struct amdgpu_device *adev)
+{
+	uint32_t tmp, inst_mask;
+	int i;
+
+	if (amdgpu_sriov_vf(adev))
+		return;
+
+	/* Setup L2 cache */
+	inst_mask = adev->aid_mask;
+	for_each_inst(i, inst_mask) {
+		tmp = RREG32_SOC15(MMHUB, i, regVM_L2_CNTL);
+		tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, ENABLE_L2_CACHE, 1);
+		tmp = REG_SET_FIELD(tmp, VM_L2_CNTL,
+				    ENABLE_L2_FRAGMENT_PROCESSING, 1);
+		/* XXX for emulation, Refer to closed source code.*/
+		tmp = REG_SET_FIELD(tmp, VM_L2_CNTL,
+				    L2_PDE0_CACHE_TAG_GENERATION_MODE, 0);
+		tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, PDE_FAULT_CLASSIFICATION,
+				    0);
+		tmp = REG_SET_FIELD(tmp, VM_L2_CNTL,
+				    CONTEXT1_IDENTITY_ACCESS_MODE, 1);
+		tmp = REG_SET_FIELD(tmp, VM_L2_CNTL,
+				    IDENTITY_MODE_FRAGMENT_SIZE, 0);
+		WREG32_SOC15(MMHUB, i, regVM_L2_CNTL, tmp);
+
+		tmp = RREG32_SOC15(MMHUB, i, regVM_L2_CNTL2);
+		tmp = REG_SET_FIELD(tmp, VM_L2_CNTL2, INVALIDATE_ALL_L1_TLBS,
+				    1);
+		tmp = REG_SET_FIELD(tmp, VM_L2_CNTL2, INVALIDATE_L2_CACHE, 1);
+		WREG32_SOC15(MMHUB, i, regVM_L2_CNTL2, tmp);
+
+		tmp = regVM_L2_CNTL3_DEFAULT;
+		if (adev->gmc.translate_further) {
+			tmp = REG_SET_FIELD(tmp, VM_L2_CNTL3, BANK_SELECT, 12);
+			tmp = REG_SET_FIELD(tmp, VM_L2_CNTL3,
+					    L2_CACHE_BIGK_FRAGMENT_SIZE, 9);
+		} else {
+			tmp = REG_SET_FIELD(tmp, VM_L2_CNTL3, BANK_SELECT, 9);
+			tmp = REG_SET_FIELD(tmp, VM_L2_CNTL3,
+					    L2_CACHE_BIGK_FRAGMENT_SIZE, 6);
+		}
+		WREG32_SOC15(MMHUB, i, regVM_L2_CNTL3, tmp);
+
+		tmp = regVM_L2_CNTL4_DEFAULT;
+		/* For AMD APP APUs setup WC memory */
+		if (adev->gmc.xgmi.connected_to_cpu || adev->gmc.is_app_apu) {
+			tmp = REG_SET_FIELD(tmp, VM_L2_CNTL4,
+					    VMC_TAP_PDE_REQUEST_PHYSICAL, 1);
+			tmp = REG_SET_FIELD(tmp, VM_L2_CNTL4,
+					    VMC_TAP_PTE_REQUEST_PHYSICAL, 1);
+		} else {
+			tmp = REG_SET_FIELD(tmp, VM_L2_CNTL4,
+					    VMC_TAP_PDE_REQUEST_PHYSICAL, 0);
+			tmp = REG_SET_FIELD(tmp, VM_L2_CNTL4,
+					    VMC_TAP_PTE_REQUEST_PHYSICAL, 0);
+		}
+		WREG32_SOC15(MMHUB, i, regVM_L2_CNTL4, tmp);
+	}
+}
+
+static void mmhub_v1_8_enable_system_domain(struct amdgpu_device *adev)
+{
+	uint32_t tmp, inst_mask;
+	int i;
+
+	inst_mask = adev->aid_mask;
+	for_each_inst(i, inst_mask) {
+		tmp = RREG32_SOC15(MMHUB, i, regVM_CONTEXT0_CNTL);
+		tmp = REG_SET_FIELD(tmp, VM_CONTEXT0_CNTL, ENABLE_CONTEXT, 1);
+		tmp = REG_SET_FIELD(tmp, VM_CONTEXT0_CNTL, PAGE_TABLE_DEPTH,
+				adev->gmc.vmid0_page_table_depth);
+		tmp = REG_SET_FIELD(tmp,
+				    VM_CONTEXT0_CNTL, PAGE_TABLE_BLOCK_SIZE,
+				    adev->gmc.vmid0_page_table_block_size);
+		tmp = REG_SET_FIELD(tmp, VM_CONTEXT0_CNTL,
+				    RETRY_PERMISSION_OR_INVALID_PAGE_FAULT, 0);
+		WREG32_SOC15(MMHUB, i, regVM_CONTEXT0_CNTL, tmp);
+	}
+}
+
+static void mmhub_v1_8_disable_identity_aperture(struct amdgpu_device *adev)
+{
+	u32 inst_mask;
+	int i;
+
+	if (amdgpu_sriov_vf(adev))
+		return;
+
+	inst_mask = adev->aid_mask;
+	for_each_inst(i, inst_mask) {
+		WREG32_SOC15(MMHUB, i,
+			     regVM_L2_CONTEXT1_IDENTITY_APERTURE_LOW_ADDR_LO32,
+			     0XFFFFFFFF);
+		WREG32_SOC15(MMHUB, i,
+			     regVM_L2_CONTEXT1_IDENTITY_APERTURE_LOW_ADDR_HI32,
+			     0x0000000F);
+
+		WREG32_SOC15(MMHUB, i,
+			     regVM_L2_CONTEXT1_IDENTITY_APERTURE_HIGH_ADDR_LO32,
+			     0);
+		WREG32_SOC15(MMHUB, i,
+			     regVM_L2_CONTEXT1_IDENTITY_APERTURE_HIGH_ADDR_HI32,
+			     0);
+
+		WREG32_SOC15(MMHUB, i,
+			     regVM_L2_CONTEXT_IDENTITY_PHYSICAL_OFFSET_LO32, 0);
+		WREG32_SOC15(MMHUB, i,
+			     regVM_L2_CONTEXT_IDENTITY_PHYSICAL_OFFSET_HI32, 0);
+	}
+}
+
+static void mmhub_v1_8_setup_vmid_config(struct amdgpu_device *adev)
+{
+	struct amdgpu_vmhub *hub;
+	unsigned int num_level, block_size;
+	uint32_t tmp, inst_mask;
+	int i, j;
+
+	num_level = adev->vm_manager.num_level;
+	block_size = adev->vm_manager.block_size;
+	if (adev->gmc.translate_further)
+		num_level -= 1;
+	else
+		block_size -= 9;
+
+	inst_mask = adev->aid_mask;
+	for_each_inst(j, inst_mask) {
+		hub = &adev->vmhub[AMDGPU_MMHUB0(j)];
+		for (i = 0; i <= 14; i++) {
+			tmp = RREG32_SOC15_OFFSET(MMHUB, j, regVM_CONTEXT1_CNTL,
+						  i * hub->ctx_distance);
+			tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL,
+					    ENABLE_CONTEXT, 1);
+			tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL,
+					    PAGE_TABLE_DEPTH, num_level);
+			tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL,
+				RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+			tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL,
+				DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+			tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL,
+				PDE0_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+			tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL,
+				VALID_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+			tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL,
+				READ_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+			tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL,
+				WRITE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+			tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL,
+				EXECUTE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+			tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL,
+					    PAGE_TABLE_BLOCK_SIZE,
+					    block_size);
+			/* On 9.4.3, XNACK can be enabled in the SQ
+			 * per-process. Retry faults need to be enabled for
+			 * that to work.
+			 */
+			tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL,
+				RETRY_PERMISSION_OR_INVALID_PAGE_FAULT, 1);
+			WREG32_SOC15_OFFSET(MMHUB, j, regVM_CONTEXT1_CNTL,
+					    i * hub->ctx_distance, tmp);
+			WREG32_SOC15_OFFSET(MMHUB, j,
+				regVM_CONTEXT1_PAGE_TABLE_START_ADDR_LO32,
+				i * hub->ctx_addr_distance, 0);
+			WREG32_SOC15_OFFSET(MMHUB, j,
+				regVM_CONTEXT1_PAGE_TABLE_START_ADDR_HI32,
+				i * hub->ctx_addr_distance, 0);
+			WREG32_SOC15_OFFSET(MMHUB, j,
+				regVM_CONTEXT1_PAGE_TABLE_END_ADDR_LO32,
+				i * hub->ctx_addr_distance,
+				lower_32_bits(adev->vm_manager.max_pfn - 1));
+			WREG32_SOC15_OFFSET(MMHUB, j,
+				regVM_CONTEXT1_PAGE_TABLE_END_ADDR_HI32,
+				i * hub->ctx_addr_distance,
+				upper_32_bits(adev->vm_manager.max_pfn - 1));
+		}
+	}
+}
+
+static void mmhub_v1_8_program_invalidation(struct amdgpu_device *adev)
+{
+	struct amdgpu_vmhub *hub;
+	u32 i, j, inst_mask;
+
+	inst_mask = adev->aid_mask;
+	for_each_inst(j, inst_mask) {
+		hub = &adev->vmhub[AMDGPU_MMHUB0(j)];
+		for (i = 0; i < 18; ++i) {
+			WREG32_SOC15_OFFSET(MMHUB, j,
+					regVM_INVALIDATE_ENG0_ADDR_RANGE_LO32,
+					i * hub->eng_addr_distance, 0xffffffff);
+			WREG32_SOC15_OFFSET(MMHUB, j,
+					regVM_INVALIDATE_ENG0_ADDR_RANGE_HI32,
+					i * hub->eng_addr_distance, 0x1f);
+		}
+	}
+}
+
+static int mmhub_v1_8_gart_enable(struct amdgpu_device *adev)
+{
+	/* GART Enable. */
+	mmhub_v1_8_init_gart_aperture_regs(adev);
+	mmhub_v1_8_init_system_aperture_regs(adev);
+	mmhub_v1_8_init_tlb_regs(adev);
+	mmhub_v1_8_init_cache_regs(adev);
+	mmhub_v1_8_init_snoop_override_regs(adev);
+
+	mmhub_v1_8_enable_system_domain(adev);
+	mmhub_v1_8_disable_identity_aperture(adev);
+	mmhub_v1_8_setup_vmid_config(adev);
+	mmhub_v1_8_program_invalidation(adev);
+
+	return 0;
+}
+
+static void mmhub_v1_8_disable_l1_tlb(struct amdgpu_device *adev)
+{
+	u32 tmp;
+	u32 i, inst_mask;
+
+	if (amdgpu_sriov_reg_indirect_l1_tlb_cntl(adev)) {
+		tmp = RREG32_SOC15(MMHUB, 0, regMC_VM_MX_L1_TLB_CNTL);
+		tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB, 0);
+		tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL,
+				    ENABLE_ADVANCED_DRIVER_MODEL, 0);
+		psp_reg_program_no_ring(&adev->psp, tmp, PSP_REG_MMHUB_L1_TLB_CNTL);
+	} else {
+		inst_mask = adev->aid_mask;
+		for_each_inst(i, inst_mask) {
+			tmp = RREG32_SOC15(MMHUB, i, regMC_VM_MX_L1_TLB_CNTL);
+			tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB,
+					    0);
+			tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL,
+					    ENABLE_ADVANCED_DRIVER_MODEL, 0);
+			WREG32_SOC15(MMHUB, i, regMC_VM_MX_L1_TLB_CNTL, tmp);
+		}
+	}
+}
+
+static void mmhub_v1_8_gart_disable(struct amdgpu_device *adev)
+{
+	struct amdgpu_vmhub *hub;
+	u32 tmp;
+	u32 i, j, inst_mask;
+
+	/* Disable all tables */
+	inst_mask = adev->aid_mask;
+	for_each_inst(j, inst_mask) {
+		hub = &adev->vmhub[AMDGPU_MMHUB0(j)];
+		for (i = 0; i < 16; i++)
+			WREG32_SOC15_OFFSET(MMHUB, j, regVM_CONTEXT0_CNTL,
+					    i * hub->ctx_distance, 0);
+		if (!amdgpu_sriov_vf(adev)) {
+			/* Setup L2 cache */
+			tmp = RREG32_SOC15(MMHUB, j, regVM_L2_CNTL);
+			tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, ENABLE_L2_CACHE,
+					    0);
+			WREG32_SOC15(MMHUB, j, regVM_L2_CNTL, tmp);
+			WREG32_SOC15(MMHUB, j, regVM_L2_CNTL3, 0);
+		}
+	}
+
+	mmhub_v1_8_disable_l1_tlb(adev);
+}
+
+/**
+ * mmhub_v1_8_set_fault_enable_default - update GART/VM fault handling
+ *
+ * @adev: amdgpu_device pointer
+ * @value: true redirects VM faults to the default page
+ */
+static void mmhub_v1_8_set_fault_enable_default(struct amdgpu_device *adev, bool value)
+{
+	u32 tmp, inst_mask;
+	int i;
+
+	if (amdgpu_sriov_vf(adev))
+		return;
+
+	inst_mask = adev->aid_mask;
+	for_each_inst(i, inst_mask) {
+		tmp = RREG32_SOC15(MMHUB, i, regVM_L2_PROTECTION_FAULT_CNTL);
+		tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL,
+				RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+		tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL,
+				PDE0_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+		tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL,
+				PDE1_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+		tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL,
+				PDE2_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+		tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL,
+			TRANSLATE_FURTHER_PROTECTION_FAULT_ENABLE_DEFAULT,
+			value);
+		tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL,
+				NACK_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+		tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL,
+				DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+		tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL,
+				VALID_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+		tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL,
+				READ_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+		tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL,
+				WRITE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+		tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL,
+				EXECUTE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+		if (!value) {
+			tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL,
+					    CRASH_ON_NO_RETRY_FAULT, 1);
+			tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL,
+					    CRASH_ON_RETRY_FAULT, 1);
+		}
+
+		WREG32_SOC15(MMHUB, i, regVM_L2_PROTECTION_FAULT_CNTL, tmp);
+	}
+}
+
+static void mmhub_v1_8_init(struct amdgpu_device *adev)
+{
+	struct amdgpu_vmhub *hub;
+	u32 inst_mask;
+	int i;
+
+	inst_mask = adev->aid_mask;
+	for_each_inst(i, inst_mask) {
+		hub = &adev->vmhub[AMDGPU_MMHUB0(i)];
+
+		hub->ctx0_ptb_addr_lo32 = SOC15_REG_OFFSET(MMHUB, i,
+			regVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32);
+		hub->ctx0_ptb_addr_hi32 = SOC15_REG_OFFSET(MMHUB, i,
+			regVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32);
+		hub->vm_inv_eng0_req =
+			SOC15_REG_OFFSET(MMHUB, i, regVM_INVALIDATE_ENG0_REQ);
+		hub->vm_inv_eng0_ack =
+			SOC15_REG_OFFSET(MMHUB, i, regVM_INVALIDATE_ENG0_ACK);
+		hub->vm_context0_cntl =
+			SOC15_REG_OFFSET(MMHUB, i, regVM_CONTEXT0_CNTL);
+		hub->vm_l2_pro_fault_status = SOC15_REG_OFFSET(MMHUB, i,
+			regVM_L2_PROTECTION_FAULT_STATUS);
+		hub->vm_l2_pro_fault_cntl = SOC15_REG_OFFSET(MMHUB, i,
+			regVM_L2_PROTECTION_FAULT_CNTL);
+
+		hub->ctx_distance = regVM_CONTEXT1_CNTL - regVM_CONTEXT0_CNTL;
+		hub->ctx_addr_distance =
+			regVM_CONTEXT1_PAGE_TABLE_BASE_ADDR_LO32 -
+			regVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32;
+		hub->eng_distance = regVM_INVALIDATE_ENG1_REQ -
+			regVM_INVALIDATE_ENG0_REQ;
+		hub->eng_addr_distance = regVM_INVALIDATE_ENG1_ADDR_RANGE_LO32 -
+			regVM_INVALIDATE_ENG0_ADDR_RANGE_LO32;
+	}
+}
+
+static int mmhub_v1_8_set_clockgating(struct amdgpu_device *adev,
+				      enum amd_clockgating_state state)
+{
+	return 0;
+}
+
+static void mmhub_v1_8_get_clockgating(struct amdgpu_device *adev, u64 *flags)
+{
+
+}
+
+const struct amdgpu_mmhub_funcs mmhub_v1_8_funcs = {
+	.get_fb_location = mmhub_v1_8_get_fb_location,
+	.init = mmhub_v1_8_init,
+	.gart_enable = mmhub_v1_8_gart_enable,
+	.set_fault_enable_default = mmhub_v1_8_set_fault_enable_default,
+	.gart_disable = mmhub_v1_8_gart_disable,
+	.setup_vm_pt_regs = mmhub_v1_8_setup_vm_pt_regs,
+	.set_clockgating = mmhub_v1_8_set_clockgating,
+	.get_clockgating = mmhub_v1_8_get_clockgating,
+};
+
+static const struct amdgpu_ras_err_status_reg_entry mmhub_v1_8_ce_reg_list[] = {
+	{AMDGPU_RAS_REG_ENTRY(MMHUB, 0, regMMEA0_CE_ERR_STATUS_LO, regMMEA0_CE_ERR_STATUS_HI),
+	1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "MMEA0"},
+	{AMDGPU_RAS_REG_ENTRY(MMHUB, 0, regMMEA1_CE_ERR_STATUS_LO, regMMEA1_CE_ERR_STATUS_HI),
+	1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "MMEA1"},
+	{AMDGPU_RAS_REG_ENTRY(MMHUB, 0, regMMEA2_CE_ERR_STATUS_LO, regMMEA2_CE_ERR_STATUS_HI),
+	1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "MMEA2"},
+	{AMDGPU_RAS_REG_ENTRY(MMHUB, 0, regMMEA3_CE_ERR_STATUS_LO, regMMEA3_CE_ERR_STATUS_HI),
+	1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "MMEA3"},
+	{AMDGPU_RAS_REG_ENTRY(MMHUB, 0, regMMEA4_CE_ERR_STATUS_LO, regMMEA4_CE_ERR_STATUS_HI),
+	1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "MMEA4"},
+	{AMDGPU_RAS_REG_ENTRY(MMHUB, 0, regMM_CANE_CE_ERR_STATUS_LO, regMM_CANE_CE_ERR_STATUS_HI),
+	1, 0, "MM_CANE"},
+};
+
+static const struct amdgpu_ras_err_status_reg_entry mmhub_v1_8_ue_reg_list[] = {
+	{AMDGPU_RAS_REG_ENTRY(MMHUB, 0, regMMEA0_UE_ERR_STATUS_LO, regMMEA0_UE_ERR_STATUS_HI),
+	1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "MMEA0"},
+	{AMDGPU_RAS_REG_ENTRY(MMHUB, 0, regMMEA1_UE_ERR_STATUS_LO, regMMEA1_UE_ERR_STATUS_HI),
+	1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "MMEA1"},
+	{AMDGPU_RAS_REG_ENTRY(MMHUB, 0, regMMEA2_UE_ERR_STATUS_LO, regMMEA2_UE_ERR_STATUS_HI),
+	1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "MMEA2"},
+	{AMDGPU_RAS_REG_ENTRY(MMHUB, 0, regMMEA3_UE_ERR_STATUS_LO, regMMEA3_UE_ERR_STATUS_HI),
+	1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "MMEA3"},
+	{AMDGPU_RAS_REG_ENTRY(MMHUB, 0, regMMEA4_UE_ERR_STATUS_LO, regMMEA4_UE_ERR_STATUS_HI),
+	1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "MMEA4"},
+	{AMDGPU_RAS_REG_ENTRY(MMHUB, 0, regMM_CANE_UE_ERR_STATUS_LO, regMM_CANE_UE_ERR_STATUS_HI),
+	1, 0, "MM_CANE"},
+};
+
+static const struct amdgpu_ras_memory_id_entry mmhub_v1_8_ras_memory_list[] = {
+	{AMDGPU_MMHUB_WGMI_PAGEMEM, "MMEA_WGMI_PAGEMEM"},
+	{AMDGPU_MMHUB_RGMI_PAGEMEM, "MMEA_RGMI_PAGEMEM"},
+	{AMDGPU_MMHUB_WDRAM_PAGEMEM, "MMEA_WDRAM_PAGEMEM"},
+	{AMDGPU_MMHUB_RDRAM_PAGEMEM, "MMEA_RDRAM_PAGEMEM"},
+	{AMDGPU_MMHUB_WIO_CMDMEM, "MMEA_WIO_CMDMEM"},
+	{AMDGPU_MMHUB_RIO_CMDMEM, "MMEA_RIO_CMDMEM"},
+	{AMDGPU_MMHUB_WGMI_CMDMEM, "MMEA_WGMI_CMDMEM"},
+	{AMDGPU_MMHUB_RGMI_CMDMEM, "MMEA_RGMI_CMDMEM"},
+	{AMDGPU_MMHUB_WDRAM_CMDMEM, "MMEA_WDRAM_CMDMEM"},
+	{AMDGPU_MMHUB_RDRAM_CMDMEM, "MMEA_RDRAM_CMDMEM"},
+	{AMDGPU_MMHUB_MAM_DMEM0, "MMEA_MAM_DMEM0"},
+	{AMDGPU_MMHUB_MAM_DMEM1, "MMEA_MAM_DMEM1"},
+	{AMDGPU_MMHUB_MAM_DMEM2, "MMEA_MAM_DMEM2"},
+	{AMDGPU_MMHUB_MAM_DMEM3, "MMEA_MAM_DMEM3"},
+	{AMDGPU_MMHUB_WRET_TAGMEM, "MMEA_WRET_TAGMEM"},
+	{AMDGPU_MMHUB_RRET_TAGMEM, "MMEA_RRET_TAGMEM"},
+	{AMDGPU_MMHUB_WIO_DATAMEM, "MMEA_WIO_DATAMEM"},
+	{AMDGPU_MMHUB_WGMI_DATAMEM, "MMEA_WGMI_DATAMEM"},
+	{AMDGPU_MMHUB_WDRAM_DATAMEM, "MMEA_WDRAM_DATAMEM"},
+};
+
+static void mmhub_v1_8_inst_query_ras_error_count(struct amdgpu_device *adev,
+						  uint32_t mmhub_inst,
+						  void *ras_err_status)
+{
+	struct ras_err_data *err_data = (struct ras_err_data *)ras_err_status;
+	unsigned long ue_count = 0, ce_count = 0;
+
+	/* NOTE: mmhub is converted by aid_mask and the range is 0-3,
+	 * which can be used as die ID directly */
+	struct amdgpu_smuio_mcm_config_info mcm_info = {
+		.socket_id = adev->smuio.funcs->get_socket_id(adev),
+		.die_id = mmhub_inst,
+	};
+
+	amdgpu_ras_inst_query_ras_error_count(adev,
+					mmhub_v1_8_ce_reg_list,
+					ARRAY_SIZE(mmhub_v1_8_ce_reg_list),
+					mmhub_v1_8_ras_memory_list,
+					ARRAY_SIZE(mmhub_v1_8_ras_memory_list),
+					mmhub_inst,
+					AMDGPU_RAS_ERROR__SINGLE_CORRECTABLE,
+					&ce_count);
+	amdgpu_ras_inst_query_ras_error_count(adev,
+					mmhub_v1_8_ue_reg_list,
+					ARRAY_SIZE(mmhub_v1_8_ue_reg_list),
+					mmhub_v1_8_ras_memory_list,
+					ARRAY_SIZE(mmhub_v1_8_ras_memory_list),
+					mmhub_inst,
+					AMDGPU_RAS_ERROR__MULTI_UNCORRECTABLE,
+					&ue_count);
+
+	amdgpu_ras_error_statistic_ce_count(err_data, &mcm_info, ce_count);
+	amdgpu_ras_error_statistic_ue_count(err_data, &mcm_info, ue_count);
+}
+
+static void mmhub_v1_8_query_ras_error_count(struct amdgpu_device *adev,
+					     void *ras_err_status)
+{
+	uint32_t inst_mask;
+	uint32_t i;
+
+	if (!amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__MMHUB)) {
+		dev_warn(adev->dev, "MMHUB RAS is not supported\n");
+		return;
+	}
+
+	inst_mask = adev->aid_mask;
+	for_each_inst(i, inst_mask)
+		mmhub_v1_8_inst_query_ras_error_count(adev, i, ras_err_status);
+}
+
+static void mmhub_v1_8_inst_reset_ras_error_count(struct amdgpu_device *adev,
+						  uint32_t mmhub_inst)
+{
+	amdgpu_ras_inst_reset_ras_error_count(adev,
+					mmhub_v1_8_ce_reg_list,
+					ARRAY_SIZE(mmhub_v1_8_ce_reg_list),
+					mmhub_inst);
+	amdgpu_ras_inst_reset_ras_error_count(adev,
+					mmhub_v1_8_ue_reg_list,
+					ARRAY_SIZE(mmhub_v1_8_ue_reg_list),
+					mmhub_inst);
+}
+
+static void mmhub_v1_8_reset_ras_error_count(struct amdgpu_device *adev)
+{
+	uint32_t inst_mask;
+	uint32_t i;
+
+	if (!amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__MMHUB)) {
+		dev_warn(adev->dev, "MMHUB RAS is not supported\n");
+		return;
+	}
+
+	inst_mask = adev->aid_mask;
+	for_each_inst(i, inst_mask)
+		mmhub_v1_8_inst_reset_ras_error_count(adev, i);
+}
+
+static const struct amdgpu_ras_block_hw_ops mmhub_v1_8_ras_hw_ops = {
+	.query_ras_error_count = mmhub_v1_8_query_ras_error_count,
+	.reset_ras_error_count = mmhub_v1_8_reset_ras_error_count,
+};
+
+static int mmhub_v1_8_aca_bank_parser(struct aca_handle *handle, struct aca_bank *bank,
+				      enum aca_smu_type type, void *data)
+{
+	struct aca_bank_info info;
+	u64 misc0;
+	int ret;
+
+	ret = aca_bank_info_decode(bank, &info);
+	if (ret)
+		return ret;
+
+	misc0 = bank->regs[ACA_REG_IDX_MISC0];
+	switch (type) {
+	case ACA_SMU_TYPE_UE:
+		bank->aca_err_type = ACA_ERROR_TYPE_UE;
+		ret = aca_error_cache_log_bank_error(handle, &info, ACA_ERROR_TYPE_UE,
+						     1ULL);
+		break;
+	case ACA_SMU_TYPE_CE:
+		bank->aca_err_type = ACA_ERROR_TYPE_CE;
+		ret = aca_error_cache_log_bank_error(handle, &info, bank->aca_err_type,
+						     ACA_REG__MISC0__ERRCNT(misc0));
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return ret;
+}
+
+/* reference to smu driver if header file */
+static int mmhub_v1_8_err_codes[] = {
+	0, 1, 2, 3, 4, /* CODE_DAGB0 - 4 */
+	5, 6, 7, 8, 9, /* CODE_EA0 - 4 */
+	10, /* CODE_UTCL2_ROUTER */
+	11, /* CODE_VML2 */
+	12, /* CODE_VML2_WALKER */
+	13, /* CODE_MMCANE */
+};
+
+static bool mmhub_v1_8_aca_bank_is_valid(struct aca_handle *handle, struct aca_bank *bank,
+					 enum aca_smu_type type, void *data)
+{
+	u32 instlo;
+
+	instlo = ACA_REG__IPID__INSTANCEIDLO(bank->regs[ACA_REG_IDX_IPID]);
+	instlo &= GENMASK(31, 1);
+
+	if (instlo != mmSMNAID_AID0_MCA_SMU)
+		return false;
+
+	if (aca_bank_check_error_codes(handle->adev, bank,
+				       mmhub_v1_8_err_codes,
+				       ARRAY_SIZE(mmhub_v1_8_err_codes)))
+		return false;
+
+	return true;
+}
+
+static const struct aca_bank_ops mmhub_v1_8_aca_bank_ops = {
+	.aca_bank_parser = mmhub_v1_8_aca_bank_parser,
+	.aca_bank_is_valid = mmhub_v1_8_aca_bank_is_valid,
+};
+
+static const struct aca_info mmhub_v1_8_aca_info = {
+	.hwip = ACA_HWIP_TYPE_SMU,
+	.mask = ACA_ERROR_UE_MASK,
+	.bank_ops = &mmhub_v1_8_aca_bank_ops,
+};
+
+static int mmhub_v1_8_ras_late_init(struct amdgpu_device *adev, struct ras_common_if *ras_block)
+{
+	int r;
+
+	r = amdgpu_ras_block_late_init(adev, ras_block);
+	if (r)
+		return r;
+
+	r = amdgpu_ras_bind_aca(adev, AMDGPU_RAS_BLOCK__MMHUB,
+				&mmhub_v1_8_aca_info, NULL);
+	if (r)
+		goto late_fini;
+
+	return 0;
+
+late_fini:
+	amdgpu_ras_block_late_fini(adev, ras_block);
+
+	return r;
+}
+
+struct amdgpu_mmhub_ras mmhub_v1_8_ras = {
+	.ras_block = {
+		.hw_ops = &mmhub_v1_8_ras_hw_ops,
+		.ras_late_init = mmhub_v1_8_ras_late_init,
+	},
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_8.h b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_8.h
new file mode 100644
index 000000000000..126f0075ac50
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_8.h
@@ -0,0 +1,29 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __MMHUB_V1_8_H__
+#define __MMHUB_V1_8_H__
+
+extern const struct amdgpu_mmhub_funcs mmhub_v1_8_funcs;
+extern struct amdgpu_mmhub_ras mmhub_v1_8_ras;
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v2_0.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v2_0.c
index 3718ff610ab2..a0cc8e218ca1 100644
--- a/drivers/gpu/drm/amd/amdgpu/mmhub_v2_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v2_0.c
@@ -32,8 +32,6 @@
 #include "gc/gc_10_1_0_offset.h"
 #include "soc15_common.h"
 
-#define mmMM_ATC_L2_MISC_CG_Sienna_Cichlid                      0x064d
-#define mmMM_ATC_L2_MISC_CG_Sienna_Cichlid_BASE_IDX             0
 #define mmDAGB0_CNTL_MISC2_Sienna_Cichlid                       0x0070
 #define mmDAGB0_CNTL_MISC2_Sienna_Cichlid_BASE_IDX              0
 
@@ -153,7 +151,7 @@ mmhub_v2_0_print_l2_protection_fault_status(struct amdgpu_device *adev,
 	dev_err(adev->dev,
 		"MMVM_L2_PROTECTION_FAULT_STATUS:0x%08X\n",
 		status);
-	switch (adev->ip_versions[MMHUB_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, MMHUB_HWIP, 0)) {
 	case IP_VERSION(2, 0, 0):
 	case IP_VERSION(2, 0, 2):
 		mmhub_cid = mmhub_client_ids_navi1x[cid][rw];
@@ -189,7 +187,7 @@ mmhub_v2_0_print_l2_protection_fault_status(struct amdgpu_device *adev,
 static void mmhub_v2_0_setup_vm_pt_regs(struct amdgpu_device *adev, uint32_t vmid,
 				uint64_t page_table_base)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0];
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
 
 	WREG32_SOC15_OFFSET_RLC(MMHUB, 0, mmMMVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32,
 			    hub->ctx_addr_distance * vmid,
@@ -236,7 +234,7 @@ static void mmhub_v2_0_init_system_aperture_regs(struct amdgpu_device *adev)
 	}
 
 	/* Set default page address. */
-	value = amdgpu_gmc_vram_mc2pa(adev, adev->vram_scratch.gpu_addr);
+	value = amdgpu_gmc_vram_mc2pa(adev, adev->mem_scratch.gpu_addr);
 	WREG32_SOC15(MMHUB, 0, mmMMMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB,
 		     (u32)(value >> 12));
 	WREG32_SOC15(MMHUB, 0, mmMMMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_MSB,
@@ -321,7 +319,7 @@ static void mmhub_v2_0_init_cache_regs(struct amdgpu_device *adev)
 
 	tmp = mmMMVM_L2_CNTL5_DEFAULT;
 	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL5, L2_CACHE_SMALLK_FRAGMENT_SIZE, 0);
-	WREG32_SOC15(GC, 0, mmMMVM_L2_CNTL5, tmp);
+	WREG32_SOC15(MMHUB, 0, mmMMVM_L2_CNTL5, tmp);
 }
 
 static void mmhub_v2_0_enable_system_domain(struct amdgpu_device *adev)
@@ -364,12 +362,12 @@ static void mmhub_v2_0_disable_identity_aperture(struct amdgpu_device *adev)
 
 static void mmhub_v2_0_setup_vmid_config(struct amdgpu_device *adev)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0];
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
 	int i;
 	uint32_t tmp;
 
 	for (i = 0; i <= 14; i++) {
-		tmp = RREG32_SOC15_OFFSET(MMHUB, 0, mmMMVM_CONTEXT1_CNTL, i);
+		tmp = RREG32_SOC15_OFFSET(MMHUB, 0, mmMMVM_CONTEXT1_CNTL, i * hub->ctx_distance);
 		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL, ENABLE_CONTEXT, 1);
 		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL, PAGE_TABLE_DEPTH,
 				    adev->vm_manager.num_level);
@@ -408,11 +406,13 @@ static void mmhub_v2_0_setup_vmid_config(struct amdgpu_device *adev)
 				    i * hub->ctx_addr_distance,
 				    upper_32_bits(adev->vm_manager.max_pfn - 1));
 	}
+
+	hub->vm_cntx_cntl = tmp;
 }
 
 static void mmhub_v2_0_program_invalidation(struct amdgpu_device *adev)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0];
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
 	unsigned i;
 
 	for (i = 0; i < 18; ++i) {
@@ -441,7 +441,7 @@ static int mmhub_v2_0_gart_enable(struct amdgpu_device *adev)
 
 static void mmhub_v2_0_gart_disable(struct amdgpu_device *adev)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0];
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
 	u32 tmp;
 	u32 i;
 
@@ -520,7 +520,7 @@ static const struct amdgpu_vmhub_funcs mmhub_v2_0_vmhub_funcs = {
 
 static void mmhub_v2_0_init(struct amdgpu_device *adev)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0];
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
 
 	hub->ctx0_ptb_addr_lo32 =
 		SOC15_REG_OFFSET(MMHUB, 0,
@@ -568,11 +568,10 @@ static void mmhub_v2_0_update_medium_grain_clock_gating(struct amdgpu_device *ad
 	if (!(adev->cg_flags & AMD_CG_SUPPORT_MC_MGCG))
 		return;
 
-	switch (adev->ip_versions[MMHUB_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, MMHUB_HWIP, 0)) {
 	case IP_VERSION(2, 1, 0):
 	case IP_VERSION(2, 1, 1):
 	case IP_VERSION(2, 1, 2):
-		def  = data  = RREG32_SOC15(MMHUB, 0, mmMM_ATC_L2_MISC_CG_Sienna_Cichlid);
 		def1 = data1 = RREG32_SOC15(MMHUB, 0, mmDAGB0_CNTL_MISC2_Sienna_Cichlid);
 		break;
 	default:
@@ -602,12 +601,10 @@ static void mmhub_v2_0_update_medium_grain_clock_gating(struct amdgpu_device *ad
 			  DAGB0_CNTL_MISC2__DISABLE_TLBRD_CG_MASK);
 	}
 
-	switch (adev->ip_versions[MMHUB_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, MMHUB_HWIP, 0)) {
 	case IP_VERSION(2, 1, 0):
 	case IP_VERSION(2, 1, 1):
 	case IP_VERSION(2, 1, 2):
-		if (def != data)
-			WREG32_SOC15(MMHUB, 0, mmMM_ATC_L2_MISC_CG_Sienna_Cichlid, data);
 		if (def1 != data1)
 			WREG32_SOC15(MMHUB, 0, mmDAGB0_CNTL_MISC2_Sienna_Cichlid, data1);
 		break;
@@ -628,12 +625,12 @@ static void mmhub_v2_0_update_medium_grain_light_sleep(struct amdgpu_device *ade
 	if (!(adev->cg_flags & AMD_CG_SUPPORT_MC_LS))
 		return;
 
-	switch (adev->ip_versions[MMHUB_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, MMHUB_HWIP, 0)) {
 	case IP_VERSION(2, 1, 0):
 	case IP_VERSION(2, 1, 1):
 	case IP_VERSION(2, 1, 2):
-		def  = data  = RREG32_SOC15(MMHUB, 0, mmMM_ATC_L2_MISC_CG_Sienna_Cichlid);
-		break;
+		/* There is no ATCL2 in MMHUB for 2.1.x */
+		return;
 	default:
 		def  = data  = RREG32_SOC15(MMHUB, 0, mmMM_ATC_L2_MISC_CG);
 		break;
@@ -644,18 +641,8 @@ static void mmhub_v2_0_update_medium_grain_light_sleep(struct amdgpu_device *ade
 	else
 		data &= ~MM_ATC_L2_MISC_CG__MEM_LS_ENABLE_MASK;
 
-	if (def != data) {
-		switch (adev->ip_versions[MMHUB_HWIP][0]) {
-		case IP_VERSION(2, 1, 0):
-		case IP_VERSION(2, 1, 1):
-		case IP_VERSION(2, 1, 2):
-			WREG32_SOC15(MMHUB, 0, mmMM_ATC_L2_MISC_CG_Sienna_Cichlid, data);
-			break;
-		default:
-			WREG32_SOC15(MMHUB, 0, mmMM_ATC_L2_MISC_CG, data);
-			break;
-		}
-	}
+	if (def != data)
+		WREG32_SOC15(MMHUB, 0, mmMM_ATC_L2_MISC_CG, data);
 }
 
 static int mmhub_v2_0_set_clockgating(struct amdgpu_device *adev,
@@ -664,7 +651,7 @@ static int mmhub_v2_0_set_clockgating(struct amdgpu_device *adev,
 	if (amdgpu_sriov_vf(adev))
 		return 0;
 
-	switch (adev->ip_versions[MMHUB_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, MMHUB_HWIP, 0)) {
 	case IP_VERSION(2, 0, 0):
 	case IP_VERSION(2, 0, 2):
 	case IP_VERSION(2, 1, 0):
@@ -682,18 +669,21 @@ static int mmhub_v2_0_set_clockgating(struct amdgpu_device *adev,
 	return 0;
 }
 
-static void mmhub_v2_0_get_clockgating(struct amdgpu_device *adev, u32 *flags)
+static void mmhub_v2_0_get_clockgating(struct amdgpu_device *adev, u64 *flags)
 {
-	int data, data1;
+	u32 data, data1;
 
 	if (amdgpu_sriov_vf(adev))
 		*flags = 0;
 
-	switch (adev->ip_versions[MMHUB_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, MMHUB_HWIP, 0)) {
 	case IP_VERSION(2, 1, 0):
 	case IP_VERSION(2, 1, 1):
 	case IP_VERSION(2, 1, 2):
-		data  = RREG32_SOC15(MMHUB, 0, mmMM_ATC_L2_MISC_CG_Sienna_Cichlid);
+		/* There is no ATCL2 in MMHUB for 2.1.x. Keep the status
+		 * based on DAGB
+		 */
+		data = MM_ATC_L2_MISC_CG__ENABLE_MASK;
 		data1 = RREG32_SOC15(MMHUB, 0, mmDAGB0_CNTL_MISC2_Sienna_Cichlid);
 		break;
 	default:
diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v2_3.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v2_3.c
index 9e16da28505a..5eb8122e2746 100644
--- a/drivers/gpu/drm/amd/amdgpu/mmhub_v2_3.c
+++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v2_3.c
@@ -90,9 +90,10 @@ mmhub_v2_3_print_l2_protection_fault_status(struct amdgpu_device *adev,
 	dev_err(adev->dev,
 		"MMVM_L2_PROTECTION_FAULT_STATUS:0x%08X\n",
 		status);
-	switch (adev->ip_versions[MMHUB_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, MMHUB_HWIP, 0)) {
 	case IP_VERSION(2, 3, 0):
 	case IP_VERSION(2, 4, 0):
+	case IP_VERSION(2, 4, 1):
 		mmhub_cid = mmhub_client_ids_vangogh[cid][rw];
 		break;
 	default:
@@ -120,7 +121,7 @@ static void mmhub_v2_3_setup_vm_pt_regs(struct amdgpu_device *adev,
 					uint32_t vmid,
 					uint64_t page_table_base)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0];
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
 
 	WREG32_SOC15_OFFSET(MMHUB, 0, mmMMVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32,
 			    hub->ctx_addr_distance * vmid, lower_32_bits(page_table_base));
@@ -163,7 +164,7 @@ static void mmhub_v2_3_init_system_aperture_regs(struct amdgpu_device *adev)
 		     max(adev->gmc.fb_end, adev->gmc.agp_end) >> 18);
 
 	/* Set default page address. */
-	value = amdgpu_gmc_vram_mc2pa(adev, adev->vram_scratch.gpu_addr);
+	value = amdgpu_gmc_vram_mc2pa(adev, adev->mem_scratch.gpu_addr);
 	WREG32_SOC15(MMHUB, 0, mmMMMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB,
 		     (u32)(value >> 12));
 	WREG32_SOC15(MMHUB, 0, mmMMMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_MSB,
@@ -242,7 +243,7 @@ static void mmhub_v2_3_init_cache_regs(struct amdgpu_device *adev)
 
 	tmp = mmMMVM_L2_CNTL5_DEFAULT;
 	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL5, L2_CACHE_SMALLK_FRAGMENT_SIZE, 0);
-	WREG32_SOC15(GC, 0, mmMMVM_L2_CNTL5, tmp);
+	WREG32_SOC15(MMHUB, 0, mmMMVM_L2_CNTL5, tmp);
 }
 
 static void mmhub_v2_3_enable_system_domain(struct amdgpu_device *adev)
@@ -279,12 +280,12 @@ static void mmhub_v2_3_disable_identity_aperture(struct amdgpu_device *adev)
 
 static void mmhub_v2_3_setup_vmid_config(struct amdgpu_device *adev)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0];
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
 	int i;
 	uint32_t tmp;
 
 	for (i = 0; i <= 14; i++) {
-		tmp = RREG32_SOC15_OFFSET(MMHUB, 0, mmMMVM_CONTEXT1_CNTL, i);
+		tmp = RREG32_SOC15_OFFSET(MMHUB, 0, mmMMVM_CONTEXT1_CNTL, i * hub->ctx_distance);
 		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL, ENABLE_CONTEXT, 1);
 		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL, PAGE_TABLE_DEPTH,
 				    adev->vm_manager.num_level);
@@ -323,12 +324,14 @@ static void mmhub_v2_3_setup_vmid_config(struct amdgpu_device *adev)
 				    i * hub->ctx_addr_distance,
 				    upper_32_bits(adev->vm_manager.max_pfn - 1));
 	}
+
+	hub->vm_cntx_cntl = tmp;
 }
 
 static void mmhub_v2_3_program_invalidation(struct amdgpu_device *adev)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0];
-	unsigned i;
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
+	unsigned int i;
 
 	for (i = 0; i < 18; ++i) {
 		WREG32_SOC15_OFFSET(MMHUB, 0,
@@ -370,7 +373,7 @@ static int mmhub_v2_3_gart_enable(struct amdgpu_device *adev)
 
 static void mmhub_v2_3_gart_disable(struct amdgpu_device *adev)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0];
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
 	u32 tmp;
 	u32 i;
 
@@ -403,6 +406,7 @@ static void mmhub_v2_3_set_fault_enable_default(struct amdgpu_device *adev,
 						bool value)
 {
 	u32 tmp;
+
 	tmp = RREG32_SOC15(MMHUB, 0, mmMMVM_L2_PROTECTION_FAULT_CNTL);
 	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
 			    RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
@@ -443,7 +447,7 @@ static const struct amdgpu_vmhub_funcs mmhub_v2_3_vmhub_funcs = {
 
 static void mmhub_v2_3_init(struct amdgpu_device *adev)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0];
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
 
 	hub->ctx0_ptb_addr_lo32 =
 		SOC15_REG_OFFSET(MMHUB, 0,
@@ -496,11 +500,11 @@ mmhub_v2_3_update_medium_grain_clock_gating(struct amdgpu_device *adev,
 	if (enable && (adev->cg_flags & AMD_CG_SUPPORT_MC_MGCG)) {
 		data &= ~MM_ATC_L2_CGTT_CLK_CTRL__SOFT_OVERRIDE_MASK;
 		data1 &= ~(DAGB0_CNTL_MISC2__DISABLE_WRREQ_CG_MASK |
-		           DAGB0_CNTL_MISC2__DISABLE_WRRET_CG_MASK |
-		           DAGB0_CNTL_MISC2__DISABLE_RDREQ_CG_MASK |
-		           DAGB0_CNTL_MISC2__DISABLE_RDRET_CG_MASK |
-		           DAGB0_CNTL_MISC2__DISABLE_TLBWR_CG_MASK |
-		           DAGB0_CNTL_MISC2__DISABLE_TLBRD_CG_MASK);
+			   DAGB0_CNTL_MISC2__DISABLE_WRRET_CG_MASK |
+			   DAGB0_CNTL_MISC2__DISABLE_RDREQ_CG_MASK |
+			   DAGB0_CNTL_MISC2__DISABLE_RDRET_CG_MASK |
+			   DAGB0_CNTL_MISC2__DISABLE_TLBWR_CG_MASK |
+			   DAGB0_CNTL_MISC2__DISABLE_TLBRD_CG_MASK);
 
 	} else {
 		data |= MM_ATC_L2_CGTT_CLK_CTRL__SOFT_OVERRIDE_MASK;
@@ -576,7 +580,7 @@ static int mmhub_v2_3_set_clockgating(struct amdgpu_device *adev,
 	return 0;
 }
 
-static void mmhub_v2_3_get_clockgating(struct amdgpu_device *adev, u32 *flags)
+static void mmhub_v2_3_get_clockgating(struct amdgpu_device *adev, u64 *flags)
 {
 	int data, data1, data2, data3;
 
@@ -590,13 +594,13 @@ static void mmhub_v2_3_get_clockgating(struct amdgpu_device *adev, u32 *flags)
 
 	/* AMD_CG_SUPPORT_MC_MGCG */
 	if (!(data & (DAGB0_CNTL_MISC2__DISABLE_WRREQ_CG_MASK |
-		       DAGB0_CNTL_MISC2__DISABLE_WRRET_CG_MASK |
-		       DAGB0_CNTL_MISC2__DISABLE_RDREQ_CG_MASK |
-		       DAGB0_CNTL_MISC2__DISABLE_RDRET_CG_MASK |
-		       DAGB0_CNTL_MISC2__DISABLE_TLBWR_CG_MASK |
-		       DAGB0_CNTL_MISC2__DISABLE_TLBRD_CG_MASK))
+			DAGB0_CNTL_MISC2__DISABLE_WRRET_CG_MASK |
+			DAGB0_CNTL_MISC2__DISABLE_RDREQ_CG_MASK |
+			DAGB0_CNTL_MISC2__DISABLE_RDRET_CG_MASK |
+			DAGB0_CNTL_MISC2__DISABLE_TLBWR_CG_MASK |
+			DAGB0_CNTL_MISC2__DISABLE_TLBRD_CG_MASK))
 		&& !(data1 & MM_ATC_L2_CGTT_CLK_CTRL__SOFT_OVERRIDE_MASK)) {
-			*flags |= AMD_CG_SUPPORT_MC_MGCG;
+		*flags |= AMD_CG_SUPPORT_MC_MGCG;
 	}
 
 	/* AMD_CG_SUPPORT_MC_LS */
diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0.c
new file mode 100644
index 000000000000..7d5242df58a5
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0.c
@@ -0,0 +1,664 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "amdgpu.h"
+#include "mmhub_v3_0.h"
+
+#include "mmhub/mmhub_3_0_0_offset.h"
+#include "mmhub/mmhub_3_0_0_sh_mask.h"
+#include "navi10_enum.h"
+
+#include "soc15_common.h"
+
+#define regMMVM_L2_CNTL3_DEFAULT				0x80100007
+#define regMMVM_L2_CNTL4_DEFAULT				0x000000c1
+#define regMMVM_L2_CNTL5_DEFAULT				0x00003fe0
+
+static const char *mmhub_client_ids_v3_0_0[][2] = {
+	[0][0] = "VMC",
+	[4][0] = "DCEDMC",
+	[5][0] = "DCEVGA",
+	[6][0] = "MP0",
+	[7][0] = "MP1",
+	[8][0] = "MPIO",
+	[16][0] = "HDP",
+	[17][0] = "LSDMA",
+	[18][0] = "JPEG",
+	[19][0] = "VCNU0",
+	[21][0] = "VSCH",
+	[22][0] = "VCNU1",
+	[23][0] = "VCN1",
+	[32+20][0] = "VCN0",
+	[2][1] = "DBGUNBIO",
+	[3][1] = "DCEDWB",
+	[4][1] = "DCEDMC",
+	[5][1] = "DCEVGA",
+	[6][1] = "MP0",
+	[7][1] = "MP1",
+	[8][1] = "MPIO",
+	[10][1] = "DBGU0",
+	[11][1] = "DBGU1",
+	[12][1] = "DBGU2",
+	[13][1] = "DBGU3",
+	[14][1] = "XDP",
+	[15][1] = "OSSSYS",
+	[16][1] = "HDP",
+	[17][1] = "LSDMA",
+	[18][1] = "JPEG",
+	[19][1] = "VCNU0",
+	[20][1] = "VCN0",
+	[21][1] = "VSCH",
+	[22][1] = "VCNU1",
+	[23][1] = "VCN1",
+};
+
+static uint32_t mmhub_v3_0_get_invalidate_req(unsigned int vmid,
+					      uint32_t flush_type)
+{
+	u32 req = 0;
+
+	/* invalidate using legacy mode on vmid*/
+	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ,
+			    PER_VMID_INVALIDATE_REQ, 1 << vmid);
+	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ, FLUSH_TYPE, flush_type);
+	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PTES, 1);
+	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PDE0, 1);
+	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PDE1, 1);
+	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PDE2, 1);
+	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ, INVALIDATE_L1_PTES, 1);
+	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ,
+			    CLEAR_PROTECTION_FAULT_STATUS_ADDR,	0);
+
+	return req;
+}
+
+static void
+mmhub_v3_0_print_l2_protection_fault_status(struct amdgpu_device *adev,
+					     uint32_t status)
+{
+	uint32_t cid, rw;
+	const char *mmhub_cid = NULL;
+
+	cid = REG_GET_FIELD(status,
+			    MMVM_L2_PROTECTION_FAULT_STATUS, CID);
+	rw = REG_GET_FIELD(status,
+			   MMVM_L2_PROTECTION_FAULT_STATUS, RW);
+
+	dev_err(adev->dev,
+		"MMVM_L2_PROTECTION_FAULT_STATUS:0x%08X\n",
+		status);
+	switch (amdgpu_ip_version(adev, MMHUB_HWIP, 0)) {
+	case IP_VERSION(3, 0, 0):
+	case IP_VERSION(3, 0, 1):
+		mmhub_cid = mmhub_client_ids_v3_0_0[cid][rw];
+		break;
+	default:
+		mmhub_cid = NULL;
+		break;
+	}
+	dev_err(adev->dev, "\t Faulty UTCL2 client ID: %s (0x%x)\n",
+		mmhub_cid ? mmhub_cid : "unknown", cid);
+	dev_err(adev->dev, "\t MORE_FAULTS: 0x%lx\n",
+		REG_GET_FIELD(status,
+		MMVM_L2_PROTECTION_FAULT_STATUS, MORE_FAULTS));
+	dev_err(adev->dev, "\t WALKER_ERROR: 0x%lx\n",
+		REG_GET_FIELD(status,
+		MMVM_L2_PROTECTION_FAULT_STATUS, WALKER_ERROR));
+	dev_err(adev->dev, "\t PERMISSION_FAULTS: 0x%lx\n",
+		REG_GET_FIELD(status,
+		MMVM_L2_PROTECTION_FAULT_STATUS, PERMISSION_FAULTS));
+	dev_err(adev->dev, "\t MAPPING_ERROR: 0x%lx\n",
+		REG_GET_FIELD(status,
+		MMVM_L2_PROTECTION_FAULT_STATUS, MAPPING_ERROR));
+	dev_err(adev->dev, "\t RW: 0x%x\n", rw);
+}
+
+static void mmhub_v3_0_setup_vm_pt_regs(struct amdgpu_device *adev, uint32_t vmid,
+				uint64_t page_table_base)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
+
+	WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32,
+			    hub->ctx_addr_distance * vmid,
+			    lower_32_bits(page_table_base));
+
+	WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32,
+			    hub->ctx_addr_distance * vmid,
+			    upper_32_bits(page_table_base));
+}
+
+static void mmhub_v3_0_init_gart_aperture_regs(struct amdgpu_device *adev)
+{
+	uint64_t pt_base = amdgpu_gmc_pd_addr(adev->gart.bo);
+
+	mmhub_v3_0_setup_vm_pt_regs(adev, 0, pt_base);
+
+	WREG32_SOC15(MMHUB, 0, regMMVM_CONTEXT0_PAGE_TABLE_START_ADDR_LO32,
+		     (u32)(adev->gmc.gart_start >> 12));
+	WREG32_SOC15(MMHUB, 0, regMMVM_CONTEXT0_PAGE_TABLE_START_ADDR_HI32,
+		     (u32)(adev->gmc.gart_start >> 44));
+
+	WREG32_SOC15(MMHUB, 0, regMMVM_CONTEXT0_PAGE_TABLE_END_ADDR_LO32,
+		     (u32)(adev->gmc.gart_end >> 12));
+	WREG32_SOC15(MMHUB, 0, regMMVM_CONTEXT0_PAGE_TABLE_END_ADDR_HI32,
+		     (u32)(adev->gmc.gart_end >> 44));
+}
+
+static void mmhub_v3_0_init_system_aperture_regs(struct amdgpu_device *adev)
+{
+	uint64_t value;
+	uint32_t tmp;
+
+	if (amdgpu_sriov_vf(adev))
+		return;
+
+	/*
+	 * the new L1 policy will block SRIOV guest from writing
+	 * these regs, and they will be programed at host.
+	 * so skip programing these regs.
+	 */
+	/* Program the AGP BAR */
+	WREG32_SOC15(MMHUB, 0, regMMMC_VM_AGP_BASE, 0);
+	WREG32_SOC15(MMHUB, 0, regMMMC_VM_AGP_BOT, adev->gmc.agp_start >> 24);
+	WREG32_SOC15(MMHUB, 0, regMMMC_VM_AGP_TOP, adev->gmc.agp_end >> 24);
+
+	/* Program the system aperture low logical page number. */
+	WREG32_SOC15(MMHUB, 0, regMMMC_VM_SYSTEM_APERTURE_LOW_ADDR,
+		     min(adev->gmc.fb_start, adev->gmc.agp_start) >> 18);
+	WREG32_SOC15(MMHUB, 0, regMMMC_VM_SYSTEM_APERTURE_HIGH_ADDR,
+		     max(adev->gmc.fb_end, adev->gmc.agp_end) >> 18);
+
+	/* Set default page address. */
+	value = amdgpu_gmc_vram_mc2pa(adev, adev->mem_scratch.gpu_addr);
+	WREG32_SOC15(MMHUB, 0, regMMMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB,
+		     (u32)(value >> 12));
+	WREG32_SOC15(MMHUB, 0, regMMMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_MSB,
+		     (u32)(value >> 44));
+
+	/* Program "protection fault". */
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_PROTECTION_FAULT_DEFAULT_ADDR_LO32,
+		     (u32)(adev->dummy_page_addr >> 12));
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_PROTECTION_FAULT_DEFAULT_ADDR_HI32,
+		     (u32)((u64)adev->dummy_page_addr >> 44));
+
+	tmp = RREG32_SOC15(MMHUB, 0, regMMVM_L2_PROTECTION_FAULT_CNTL2);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL2,
+			    ACTIVE_PAGE_MIGRATION_PTE_READ_RETRY, 1);
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_PROTECTION_FAULT_CNTL2, tmp);
+}
+
+static void mmhub_v3_0_init_tlb_regs(struct amdgpu_device *adev)
+{
+	uint32_t tmp;
+
+	/* Setup TLB control */
+	tmp = RREG32_SOC15(MMHUB, 0, regMMMC_VM_MX_L1_TLB_CNTL);
+
+	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB, 1);
+	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL, SYSTEM_ACCESS_MODE, 3);
+	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL,
+			    ENABLE_ADVANCED_DRIVER_MODEL, 1);
+	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL,
+			    SYSTEM_APERTURE_UNMAPPED_ACCESS, 0);
+	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL, ECO_BITS, 0);
+	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL,
+			    MTYPE, MTYPE_UC); /* UC, uncached */
+
+	WREG32_SOC15(MMHUB, 0, regMMMC_VM_MX_L1_TLB_CNTL, tmp);
+}
+
+static void mmhub_v3_0_init_cache_regs(struct amdgpu_device *adev)
+{
+	uint32_t tmp;
+
+	/* These registers are not accessible to VF-SRIOV.
+	 * The PF will program them instead.
+	 */
+	if (amdgpu_sriov_vf(adev))
+		return;
+
+	/* Setup L2 cache */
+	tmp = RREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL, ENABLE_L2_CACHE, 1);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL, ENABLE_L2_FRAGMENT_PROCESSING, 0);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL,
+			    ENABLE_DEFAULT_PAGE_OUT_TO_SYSTEM_MEMORY, 1);
+	/* XXX for emulation, Refer to closed source code.*/
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL, L2_PDE0_CACHE_TAG_GENERATION_MODE,
+			    0);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL, PDE_FAULT_CLASSIFICATION, 0);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL, CONTEXT1_IDENTITY_ACCESS_MODE, 1);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL, IDENTITY_MODE_FRAGMENT_SIZE, 0);
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL, tmp);
+
+	tmp = RREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL2);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL2, INVALIDATE_ALL_L1_TLBS, 1);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL2, INVALIDATE_L2_CACHE, 1);
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL2, tmp);
+
+	tmp = regMMVM_L2_CNTL3_DEFAULT;
+	if (adev->gmc.translate_further) {
+		tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL3, BANK_SELECT, 12);
+		tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL3,
+				    L2_CACHE_BIGK_FRAGMENT_SIZE, 9);
+	} else {
+		tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL3, BANK_SELECT, 9);
+		tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL3,
+				    L2_CACHE_BIGK_FRAGMENT_SIZE, 6);
+	}
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL3, tmp);
+
+	tmp = regMMVM_L2_CNTL4_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL4, VMC_TAP_PDE_REQUEST_PHYSICAL, 0);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL4, VMC_TAP_PTE_REQUEST_PHYSICAL, 0);
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL4, tmp);
+
+	tmp = regMMVM_L2_CNTL5_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL5, L2_CACHE_SMALLK_FRAGMENT_SIZE, 0);
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL5, tmp);
+}
+
+static void mmhub_v3_0_enable_system_domain(struct amdgpu_device *adev)
+{
+	uint32_t tmp;
+
+	tmp = RREG32_SOC15(MMHUB, 0, regMMVM_CONTEXT0_CNTL);
+	tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT0_CNTL, ENABLE_CONTEXT, 1);
+	tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT0_CNTL, PAGE_TABLE_DEPTH, 0);
+	tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT0_CNTL,
+			    RETRY_PERMISSION_OR_INVALID_PAGE_FAULT, 0);
+	WREG32_SOC15(MMHUB, 0, regMMVM_CONTEXT0_CNTL, tmp);
+}
+
+static void mmhub_v3_0_disable_identity_aperture(struct amdgpu_device *adev)
+{
+	/* These registers are not accessible to VF-SRIOV.
+	 * The PF will program them instead.
+	 */
+	if (amdgpu_sriov_vf(adev))
+		return;
+
+	WREG32_SOC15(MMHUB, 0,
+		     regMMVM_L2_CONTEXT1_IDENTITY_APERTURE_LOW_ADDR_LO32,
+		     0xFFFFFFFF);
+	WREG32_SOC15(MMHUB, 0,
+		     regMMVM_L2_CONTEXT1_IDENTITY_APERTURE_LOW_ADDR_HI32,
+		     0x0000000F);
+
+	WREG32_SOC15(MMHUB, 0,
+		     regMMVM_L2_CONTEXT1_IDENTITY_APERTURE_HIGH_ADDR_LO32, 0);
+	WREG32_SOC15(MMHUB, 0,
+		     regMMVM_L2_CONTEXT1_IDENTITY_APERTURE_HIGH_ADDR_HI32, 0);
+
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CONTEXT_IDENTITY_PHYSICAL_OFFSET_LO32,
+		     0);
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CONTEXT_IDENTITY_PHYSICAL_OFFSET_HI32,
+		     0);
+}
+
+static void mmhub_v3_0_setup_vmid_config(struct amdgpu_device *adev)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
+	int i;
+	uint32_t tmp;
+
+	for (i = 0; i <= 14; i++) {
+		tmp = RREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT1_CNTL, i * hub->ctx_distance);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL, ENABLE_CONTEXT, 1);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL, PAGE_TABLE_DEPTH,
+				    adev->vm_manager.num_level);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT,
+				    1);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    PDE0_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    VALID_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    READ_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    WRITE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    EXECUTE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    PAGE_TABLE_BLOCK_SIZE,
+				    adev->vm_manager.block_size - 9);
+		/* Send no-retry XNACK on fault to suppress VM fault storm. */
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    RETRY_PERMISSION_OR_INVALID_PAGE_FAULT,
+				    !amdgpu_noretry);
+		WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT1_CNTL,
+				    i * hub->ctx_distance, tmp);
+		WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT1_PAGE_TABLE_START_ADDR_LO32,
+				    i * hub->ctx_addr_distance, 0);
+		WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT1_PAGE_TABLE_START_ADDR_HI32,
+				    i * hub->ctx_addr_distance, 0);
+		WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT1_PAGE_TABLE_END_ADDR_LO32,
+				    i * hub->ctx_addr_distance,
+				    lower_32_bits(adev->vm_manager.max_pfn - 1));
+		WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT1_PAGE_TABLE_END_ADDR_HI32,
+				    i * hub->ctx_addr_distance,
+				    upper_32_bits(adev->vm_manager.max_pfn - 1));
+	}
+
+	hub->vm_cntx_cntl = tmp;
+}
+
+static void mmhub_v3_0_program_invalidation(struct amdgpu_device *adev)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
+	unsigned i;
+
+	for (i = 0; i < 18; ++i) {
+		WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_INVALIDATE_ENG0_ADDR_RANGE_LO32,
+				    i * hub->eng_addr_distance, 0xffffffff);
+		WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_INVALIDATE_ENG0_ADDR_RANGE_HI32,
+				    i * hub->eng_addr_distance, 0x1f);
+	}
+}
+
+static int mmhub_v3_0_gart_enable(struct amdgpu_device *adev)
+{
+	/* GART Enable. */
+	mmhub_v3_0_init_gart_aperture_regs(adev);
+	mmhub_v3_0_init_system_aperture_regs(adev);
+	mmhub_v3_0_init_tlb_regs(adev);
+	mmhub_v3_0_init_cache_regs(adev);
+
+	mmhub_v3_0_enable_system_domain(adev);
+	mmhub_v3_0_disable_identity_aperture(adev);
+	mmhub_v3_0_setup_vmid_config(adev);
+	mmhub_v3_0_program_invalidation(adev);
+
+	return 0;
+}
+
+static void mmhub_v3_0_gart_disable(struct amdgpu_device *adev)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
+	u32 tmp;
+	u32 i;
+
+	/* Disable all tables */
+	for (i = 0; i < 16; i++)
+		WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT0_CNTL,
+				    i * hub->ctx_distance, 0);
+
+	/* Setup TLB control */
+	tmp = RREG32_SOC15(MMHUB, 0, regMMMC_VM_MX_L1_TLB_CNTL);
+	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB, 0);
+	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL,
+			    ENABLE_ADVANCED_DRIVER_MODEL, 0);
+	WREG32_SOC15(MMHUB, 0, regMMMC_VM_MX_L1_TLB_CNTL, tmp);
+
+	/* Setup L2 cache */
+	tmp = RREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL, ENABLE_L2_CACHE, 0);
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL, tmp);
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL3, 0);
+}
+
+/**
+ * mmhub_v3_0_set_fault_enable_default - update GART/VM fault handling
+ *
+ * @adev: amdgpu_device pointer
+ * @value: true redirects VM faults to the default page
+ */
+static void mmhub_v3_0_set_fault_enable_default(struct amdgpu_device *adev, bool value)
+{
+	u32 tmp;
+
+	/* These registers are not accessible to VF-SRIOV.
+	 * The PF will program them instead.
+	 */
+	if (amdgpu_sriov_vf(adev))
+		return;
+
+	tmp = RREG32_SOC15(MMHUB, 0, regMMVM_L2_PROTECTION_FAULT_CNTL);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    PDE0_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    PDE1_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    PDE2_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    TRANSLATE_FURTHER_PROTECTION_FAULT_ENABLE_DEFAULT,
+			    value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    NACK_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    VALID_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    READ_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    WRITE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    EXECUTE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	if (!value) {
+		tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+				CRASH_ON_NO_RETRY_FAULT, 1);
+		tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+				CRASH_ON_RETRY_FAULT, 1);
+	}
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_PROTECTION_FAULT_CNTL, tmp);
+}
+
+static const struct amdgpu_vmhub_funcs mmhub_v3_0_vmhub_funcs = {
+	.print_l2_protection_fault_status = mmhub_v3_0_print_l2_protection_fault_status,
+	.get_invalidate_req = mmhub_v3_0_get_invalidate_req,
+};
+
+static void mmhub_v3_0_init(struct amdgpu_device *adev)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
+
+	hub->ctx0_ptb_addr_lo32 =
+		SOC15_REG_OFFSET(MMHUB, 0,
+				 regMMVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32);
+	hub->ctx0_ptb_addr_hi32 =
+		SOC15_REG_OFFSET(MMHUB, 0,
+				 regMMVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32);
+	hub->vm_inv_eng0_sem =
+		SOC15_REG_OFFSET(MMHUB, 0, regMMVM_INVALIDATE_ENG0_SEM);
+	hub->vm_inv_eng0_req =
+		SOC15_REG_OFFSET(MMHUB, 0, regMMVM_INVALIDATE_ENG0_REQ);
+	hub->vm_inv_eng0_ack =
+		SOC15_REG_OFFSET(MMHUB, 0, regMMVM_INVALIDATE_ENG0_ACK);
+	hub->vm_context0_cntl =
+		SOC15_REG_OFFSET(MMHUB, 0, regMMVM_CONTEXT0_CNTL);
+	hub->vm_l2_pro_fault_status =
+		SOC15_REG_OFFSET(MMHUB, 0, regMMVM_L2_PROTECTION_FAULT_STATUS);
+	hub->vm_l2_pro_fault_cntl =
+		SOC15_REG_OFFSET(MMHUB, 0, regMMVM_L2_PROTECTION_FAULT_CNTL);
+
+	hub->ctx_distance = regMMVM_CONTEXT1_CNTL - regMMVM_CONTEXT0_CNTL;
+	hub->ctx_addr_distance = regMMVM_CONTEXT1_PAGE_TABLE_BASE_ADDR_LO32 -
+		regMMVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32;
+	hub->eng_distance = regMMVM_INVALIDATE_ENG1_REQ -
+		regMMVM_INVALIDATE_ENG0_REQ;
+	hub->eng_addr_distance = regMMVM_INVALIDATE_ENG1_ADDR_RANGE_LO32 -
+		regMMVM_INVALIDATE_ENG0_ADDR_RANGE_LO32;
+
+	hub->vm_cntx_cntl_vm_fault = MMVM_CONTEXT1_CNTL__RANGE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		MMVM_CONTEXT1_CNTL__DUMMY_PAGE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		MMVM_CONTEXT1_CNTL__PDE0_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		MMVM_CONTEXT1_CNTL__VALID_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		MMVM_CONTEXT1_CNTL__READ_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		MMVM_CONTEXT1_CNTL__WRITE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		MMVM_CONTEXT1_CNTL__EXECUTE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK;
+
+	hub->vm_l2_bank_select_reserved_cid2 =
+		SOC15_REG_OFFSET(MMHUB, 0, regMMVM_L2_BANK_SELECT_RESERVED_CID2);
+
+	hub->vm_contexts_disable =
+		SOC15_REG_OFFSET(MMHUB, 0, regMMVM_CONTEXTS_DISABLE);
+
+	hub->vmhub_funcs = &mmhub_v3_0_vmhub_funcs;
+}
+
+static u64 mmhub_v3_0_get_fb_location(struct amdgpu_device *adev)
+{
+	u64 base;
+
+	base = RREG32_SOC15(MMHUB, 0, regMMMC_VM_FB_LOCATION_BASE);
+
+	base &= MMMC_VM_FB_LOCATION_BASE__FB_BASE_MASK;
+	base <<= 24;
+
+	return base;
+}
+
+static u64 mmhub_v3_0_get_mc_fb_offset(struct amdgpu_device *adev)
+{
+	return (u64)RREG32_SOC15(MMHUB, 0, regMMMC_VM_FB_OFFSET) << 24;
+}
+
+static void mmhub_v3_0_update_medium_grain_clock_gating(struct amdgpu_device *adev,
+							bool enable)
+{
+	uint32_t def, data;
+#if 0
+	uint32_t def1, data1, def2 = 0, data2 = 0;
+#endif
+
+	def  = data  = RREG32_SOC15(MMHUB, 0, regMM_ATC_L2_MISC_CG);
+#if 0
+	def1 = data1 = RREG32_SOC15(MMHUB, 0, regDAGB0_CNTL_MISC2);
+	def2 = data2 = RREG32_SOC15(MMHUB, 0, regDAGB1_CNTL_MISC2);
+#endif
+
+	if (enable) {
+		data |= MM_ATC_L2_MISC_CG__ENABLE_MASK;
+#if 0
+		data1 &= ~(DAGB0_CNTL_MISC2__DISABLE_WRREQ_CG_MASK |
+		           DAGB0_CNTL_MISC2__DISABLE_WRRET_CG_MASK |
+		           DAGB0_CNTL_MISC2__DISABLE_RDREQ_CG_MASK |
+		           DAGB0_CNTL_MISC2__DISABLE_RDRET_CG_MASK |
+		           DAGB0_CNTL_MISC2__DISABLE_TLBWR_CG_MASK |
+		           DAGB0_CNTL_MISC2__DISABLE_TLBRD_CG_MASK);
+
+		data2 &= ~(DAGB1_CNTL_MISC2__DISABLE_WRREQ_CG_MASK |
+		           DAGB1_CNTL_MISC2__DISABLE_WRRET_CG_MASK |
+		           DAGB1_CNTL_MISC2__DISABLE_RDREQ_CG_MASK |
+		           DAGB1_CNTL_MISC2__DISABLE_RDRET_CG_MASK |
+		           DAGB1_CNTL_MISC2__DISABLE_TLBWR_CG_MASK |
+		           DAGB1_CNTL_MISC2__DISABLE_TLBRD_CG_MASK);
+#endif
+	} else {
+		data &= ~MM_ATC_L2_MISC_CG__ENABLE_MASK;
+#if 0
+		data1 |= (DAGB0_CNTL_MISC2__DISABLE_WRREQ_CG_MASK |
+			  DAGB0_CNTL_MISC2__DISABLE_WRRET_CG_MASK |
+			  DAGB0_CNTL_MISC2__DISABLE_RDREQ_CG_MASK |
+			  DAGB0_CNTL_MISC2__DISABLE_RDRET_CG_MASK |
+			  DAGB0_CNTL_MISC2__DISABLE_TLBWR_CG_MASK |
+			  DAGB0_CNTL_MISC2__DISABLE_TLBRD_CG_MASK);
+
+		data2 |= (DAGB1_CNTL_MISC2__DISABLE_WRREQ_CG_MASK |
+		          DAGB1_CNTL_MISC2__DISABLE_WRRET_CG_MASK |
+		          DAGB1_CNTL_MISC2__DISABLE_RDREQ_CG_MASK |
+		          DAGB1_CNTL_MISC2__DISABLE_RDRET_CG_MASK |
+		          DAGB1_CNTL_MISC2__DISABLE_TLBWR_CG_MASK |
+		          DAGB1_CNTL_MISC2__DISABLE_TLBRD_CG_MASK);
+#endif
+	}
+
+	if (def != data)
+		WREG32_SOC15(MMHUB, 0, regMM_ATC_L2_MISC_CG, data);
+#if 0
+	if (def1 != data1)
+		WREG32_SOC15(MMHUB, 0, regDAGB0_CNTL_MISC2, data1);
+
+	if (def2 != data2)
+		WREG32_SOC15(MMHUB, 0, regDAGB1_CNTL_MISC2, data2);
+#endif
+}
+
+static void mmhub_v3_0_update_medium_grain_light_sleep(struct amdgpu_device *adev,
+						       bool enable)
+{
+	uint32_t def, data;
+
+	def = data = RREG32_SOC15(MMHUB, 0, regMM_ATC_L2_MISC_CG);
+
+	if (enable)
+		data |= MM_ATC_L2_MISC_CG__MEM_LS_ENABLE_MASK;
+	else
+		data &= ~MM_ATC_L2_MISC_CG__MEM_LS_ENABLE_MASK;
+
+	if (def != data)
+		WREG32_SOC15(MMHUB, 0, regMM_ATC_L2_MISC_CG, data);
+}
+
+static int mmhub_v3_0_set_clockgating(struct amdgpu_device *adev,
+			       enum amd_clockgating_state state)
+{
+	if (amdgpu_sriov_vf(adev))
+		return 0;
+
+	if (adev->cg_flags & AMD_CG_SUPPORT_MC_MGCG)
+		mmhub_v3_0_update_medium_grain_clock_gating(adev,
+				state == AMD_CG_STATE_GATE);
+
+	if (adev->cg_flags & AMD_CG_SUPPORT_MC_LS)
+		mmhub_v3_0_update_medium_grain_light_sleep(adev,
+				state == AMD_CG_STATE_GATE);
+
+	return 0;
+}
+
+static void mmhub_v3_0_get_clockgating(struct amdgpu_device *adev, u64 *flags)
+{
+	int data;
+
+	if (amdgpu_sriov_vf(adev))
+		*flags = 0;
+
+	data = RREG32_SOC15(MMHUB, 0, regMM_ATC_L2_MISC_CG);
+
+	/* AMD_CG_SUPPORT_MC_MGCG */
+	if (data & MM_ATC_L2_MISC_CG__ENABLE_MASK)
+		*flags |= AMD_CG_SUPPORT_MC_MGCG;
+
+	/* AMD_CG_SUPPORT_MC_LS */
+	if (data & MM_ATC_L2_MISC_CG__MEM_LS_ENABLE_MASK)
+		*flags |= AMD_CG_SUPPORT_MC_LS;
+}
+
+const struct amdgpu_mmhub_funcs mmhub_v3_0_funcs = {
+	.init = mmhub_v3_0_init,
+	.get_fb_location = mmhub_v3_0_get_fb_location,
+	.get_mc_fb_offset = mmhub_v3_0_get_mc_fb_offset,
+	.gart_enable = mmhub_v3_0_gart_enable,
+	.set_fault_enable_default = mmhub_v3_0_set_fault_enable_default,
+	.gart_disable = mmhub_v3_0_gart_disable,
+	.set_clockgating = mmhub_v3_0_set_clockgating,
+	.get_clockgating = mmhub_v3_0_get_clockgating,
+	.setup_vm_pt_regs = mmhub_v3_0_setup_vm_pt_regs,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0.h b/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0.h
new file mode 100644
index 000000000000..3ced20f350bb
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0.h
@@ -0,0 +1,28 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __MMHUB_V3_0_H__
+#define __MMHUB_V3_0_H__
+
+extern const struct amdgpu_mmhub_funcs mmhub_v3_0_funcs;
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0_1.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0_1.c
new file mode 100644
index 000000000000..910337dc28d1
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0_1.c
@@ -0,0 +1,597 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "amdgpu.h"
+#include "mmhub_v3_0_1.h"
+
+#include "mmhub/mmhub_3_0_1_offset.h"
+#include "mmhub/mmhub_3_0_1_sh_mask.h"
+#include "navi10_enum.h"
+
+#include "soc15_common.h"
+
+#define regMMVM_L2_CNTL3_DEFAULT				0x80100007
+#define regMMVM_L2_CNTL4_DEFAULT				0x000000c1
+#define regMMVM_L2_CNTL5_DEFAULT				0x00003fe0
+
+static const char *mmhub_client_ids_v3_0_1[][2] = {
+	[0][0] = "VMC",
+	[1][0] = "ISPXT",
+	[2][0] = "ISPIXT",
+	[4][0] = "DCEDMC",
+	[5][0] = "DCEVGA",
+	[6][0] = "MP0",
+	[7][0] = "MP1",
+	[8][0] = "MPM",
+	[12][0] = "ISPTNR",
+	[14][0] = "ISPCRD0",
+	[15][0] = "ISPCRD1",
+	[16][0] = "ISPCRD2",
+	[22][0] = "HDP",
+	[23][0] = "LSDMA",
+	[24][0] = "JPEG",
+	[27][0] = "VSCH",
+	[28][0] = "VCNU",
+	[29][0] = "VCN",
+	[1][1] = "ISPXT",
+	[2][1] = "ISPIXT",
+	[3][1] = "DCEDWB",
+	[4][1] = "DCEDMC",
+	[5][1] = "DCEVGA",
+	[6][1] = "MP0",
+	[7][1] = "MP1",
+	[8][1] = "MPM",
+	[10][1] = "ISPMWR0",
+	[11][1] = "ISPMWR1",
+	[12][1] = "ISPTNR",
+	[13][1] = "ISPSWR",
+	[14][1] = "ISPCWR0",
+	[15][1] = "ISPCWR1",
+	[16][1] = "ISPCWR2",
+	[17][1] = "ISPCWR3",
+	[18][1] = "XDP",
+	[21][1] = "OSSSYS",
+	[22][1] = "HDP",
+	[23][1] = "LSDMA",
+	[24][1] = "JPEG",
+	[27][1] = "VSCH",
+	[28][1] = "VCNU",
+	[29][1] = "VCN",
+};
+
+static uint32_t mmhub_v3_0_1_get_invalidate_req(unsigned int vmid,
+						uint32_t flush_type)
+{
+	u32 req = 0;
+
+	/* invalidate using legacy mode on vmid*/
+	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ,
+			    PER_VMID_INVALIDATE_REQ, 1 << vmid);
+	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ, FLUSH_TYPE, flush_type);
+	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PTES, 1);
+	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PDE0, 1);
+	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PDE1, 1);
+	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PDE2, 1);
+	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ, INVALIDATE_L1_PTES, 1);
+	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ,
+			    CLEAR_PROTECTION_FAULT_STATUS_ADDR,	0);
+
+	return req;
+}
+
+static void
+mmhub_v3_0_1_print_l2_protection_fault_status(struct amdgpu_device *adev,
+					      uint32_t status)
+{
+	uint32_t cid, rw;
+	const char *mmhub_cid = NULL;
+
+	cid = REG_GET_FIELD(status,
+			    MMVM_L2_PROTECTION_FAULT_STATUS, CID);
+	rw = REG_GET_FIELD(status,
+			   MMVM_L2_PROTECTION_FAULT_STATUS, RW);
+
+	dev_err(adev->dev,
+		"MMVM_L2_PROTECTION_FAULT_STATUS:0x%08X\n",
+		status);
+
+	switch (amdgpu_ip_version(adev, MMHUB_HWIP, 0)) {
+	case IP_VERSION(3, 0, 1):
+		mmhub_cid = mmhub_client_ids_v3_0_1[cid][rw];
+		break;
+	default:
+		mmhub_cid = NULL;
+		break;
+	}
+
+	dev_err(adev->dev, "\t Faulty UTCL2 client ID: %s (0x%x)\n",
+		mmhub_cid ? mmhub_cid : "unknown", cid);
+	dev_err(adev->dev, "\t MORE_FAULTS: 0x%lx\n",
+		REG_GET_FIELD(status,
+		MMVM_L2_PROTECTION_FAULT_STATUS, MORE_FAULTS));
+	dev_err(adev->dev, "\t WALKER_ERROR: 0x%lx\n",
+		REG_GET_FIELD(status,
+		MMVM_L2_PROTECTION_FAULT_STATUS, WALKER_ERROR));
+	dev_err(adev->dev, "\t PERMISSION_FAULTS: 0x%lx\n",
+		REG_GET_FIELD(status,
+		MMVM_L2_PROTECTION_FAULT_STATUS, PERMISSION_FAULTS));
+	dev_err(adev->dev, "\t MAPPING_ERROR: 0x%lx\n",
+		REG_GET_FIELD(status,
+		MMVM_L2_PROTECTION_FAULT_STATUS, MAPPING_ERROR));
+	dev_err(adev->dev, "\t RW: 0x%x\n", rw);
+}
+
+static void mmhub_v3_0_1_setup_vm_pt_regs(struct amdgpu_device *adev,
+					  uint32_t vmid,
+					  uint64_t page_table_base)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
+
+	WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32,
+			    hub->ctx_addr_distance * vmid,
+			    lower_32_bits(page_table_base));
+
+	WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32,
+			    hub->ctx_addr_distance * vmid,
+			    upper_32_bits(page_table_base));
+}
+
+static void mmhub_v3_0_1_init_gart_aperture_regs(struct amdgpu_device *adev)
+{
+	uint64_t pt_base = amdgpu_gmc_pd_addr(adev->gart.bo);
+
+	mmhub_v3_0_1_setup_vm_pt_regs(adev, 0, pt_base);
+
+	WREG32_SOC15(MMHUB, 0, regMMVM_CONTEXT0_PAGE_TABLE_START_ADDR_LO32,
+		     (u32)(adev->gmc.gart_start >> 12));
+	WREG32_SOC15(MMHUB, 0, regMMVM_CONTEXT0_PAGE_TABLE_START_ADDR_HI32,
+		     (u32)(adev->gmc.gart_start >> 44));
+
+	WREG32_SOC15(MMHUB, 0, regMMVM_CONTEXT0_PAGE_TABLE_END_ADDR_LO32,
+		     (u32)(adev->gmc.gart_end >> 12));
+	WREG32_SOC15(MMHUB, 0, regMMVM_CONTEXT0_PAGE_TABLE_END_ADDR_HI32,
+		     (u32)(adev->gmc.gart_end >> 44));
+}
+
+static void mmhub_v3_0_1_init_system_aperture_regs(struct amdgpu_device *adev)
+{
+	uint64_t value;
+	uint32_t tmp;
+
+	/* Program the AGP BAR */
+	WREG32_SOC15(MMHUB, 0, regMMMC_VM_AGP_BASE, 0);
+	WREG32_SOC15(MMHUB, 0, regMMMC_VM_AGP_BOT, adev->gmc.agp_start >> 24);
+	WREG32_SOC15(MMHUB, 0, regMMMC_VM_AGP_TOP, adev->gmc.agp_end >> 24);
+
+	/*
+	 * the new L1 policy will block SRIOV guest from writing
+	 * these regs, and they will be programed at host.
+	 * so skip programing these regs.
+	 */
+	/* Program the system aperture low logical page number. */
+	WREG32_SOC15(MMHUB, 0, regMMMC_VM_SYSTEM_APERTURE_LOW_ADDR,
+		     min(adev->gmc.fb_start, adev->gmc.agp_start) >> 18);
+	WREG32_SOC15(MMHUB, 0, regMMMC_VM_SYSTEM_APERTURE_HIGH_ADDR,
+		     max(adev->gmc.fb_end, adev->gmc.agp_end) >> 18);
+
+	/* Set default page address. */
+	value = amdgpu_gmc_vram_mc2pa(adev, adev->mem_scratch.gpu_addr);
+	WREG32_SOC15(MMHUB, 0, regMMMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB,
+		     (u32)(value >> 12));
+	WREG32_SOC15(MMHUB, 0, regMMMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_MSB,
+		     (u32)(value >> 44));
+
+	/* Program "protection fault". */
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_PROTECTION_FAULT_DEFAULT_ADDR_LO32,
+		     (u32)(adev->dummy_page_addr >> 12));
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_PROTECTION_FAULT_DEFAULT_ADDR_HI32,
+		     (u32)((u64)adev->dummy_page_addr >> 44));
+
+	tmp = RREG32_SOC15(MMHUB, 0, regMMVM_L2_PROTECTION_FAULT_CNTL2);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL2,
+			    ACTIVE_PAGE_MIGRATION_PTE_READ_RETRY, 1);
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_PROTECTION_FAULT_CNTL2, tmp);
+}
+
+static void mmhub_v3_0_1_init_tlb_regs(struct amdgpu_device *adev)
+{
+	uint32_t tmp;
+
+	/* Setup TLB control */
+	tmp = RREG32_SOC15(MMHUB, 0, regMMMC_VM_MX_L1_TLB_CNTL);
+
+	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB, 1);
+	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL, SYSTEM_ACCESS_MODE, 3);
+	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL,
+			    ENABLE_ADVANCED_DRIVER_MODEL, 1);
+	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL,
+			    SYSTEM_APERTURE_UNMAPPED_ACCESS, 0);
+	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL, ECO_BITS, 0);
+	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL,
+			    MTYPE, MTYPE_UC); /* UC, uncached */
+
+	WREG32_SOC15(MMHUB, 0, regMMMC_VM_MX_L1_TLB_CNTL, tmp);
+}
+
+static void mmhub_v3_0_1_init_cache_regs(struct amdgpu_device *adev)
+{
+	uint32_t tmp;
+
+	/* Setup L2 cache */
+	tmp = RREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL, ENABLE_L2_CACHE, 1);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL, ENABLE_L2_FRAGMENT_PROCESSING, 0);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL,
+			    ENABLE_DEFAULT_PAGE_OUT_TO_SYSTEM_MEMORY, 1);
+	/* XXX for emulation, Refer to closed source code.*/
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL, L2_PDE0_CACHE_TAG_GENERATION_MODE,
+			    0);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL, PDE_FAULT_CLASSIFICATION, 0);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL, CONTEXT1_IDENTITY_ACCESS_MODE, 1);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL, IDENTITY_MODE_FRAGMENT_SIZE, 0);
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL, tmp);
+
+	tmp = RREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL2);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL2, INVALIDATE_ALL_L1_TLBS, 1);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL2, INVALIDATE_L2_CACHE, 1);
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL2, tmp);
+
+	tmp = regMMVM_L2_CNTL3_DEFAULT;
+	if (adev->gmc.translate_further) {
+		tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL3, BANK_SELECT, 12);
+		tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL3,
+				    L2_CACHE_BIGK_FRAGMENT_SIZE, 9);
+	} else {
+		tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL3, BANK_SELECT, 9);
+		tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL3,
+				    L2_CACHE_BIGK_FRAGMENT_SIZE, 6);
+	}
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL3, tmp);
+
+	tmp = regMMVM_L2_CNTL4_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL4, VMC_TAP_PDE_REQUEST_PHYSICAL, 0);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL4, VMC_TAP_PTE_REQUEST_PHYSICAL, 0);
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL4, tmp);
+
+	tmp = regMMVM_L2_CNTL5_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL5, L2_CACHE_SMALLK_FRAGMENT_SIZE, 0);
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL5, tmp);
+}
+
+static void mmhub_v3_0_1_enable_system_domain(struct amdgpu_device *adev)
+{
+	uint32_t tmp;
+
+	tmp = RREG32_SOC15(MMHUB, 0, regMMVM_CONTEXT0_CNTL);
+	tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT0_CNTL, ENABLE_CONTEXT, 1);
+	tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT0_CNTL, PAGE_TABLE_DEPTH, 0);
+	tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT0_CNTL,
+			    RETRY_PERMISSION_OR_INVALID_PAGE_FAULT, 0);
+	WREG32_SOC15(MMHUB, 0, regMMVM_CONTEXT0_CNTL, tmp);
+}
+
+static void mmhub_v3_0_1_disable_identity_aperture(struct amdgpu_device *adev)
+{
+	WREG32_SOC15(MMHUB, 0,
+		     regMMVM_L2_CONTEXT1_IDENTITY_APERTURE_LOW_ADDR_LO32,
+		     0xFFFFFFFF);
+	WREG32_SOC15(MMHUB, 0,
+		     regMMVM_L2_CONTEXT1_IDENTITY_APERTURE_LOW_ADDR_HI32,
+		     0x0000000F);
+
+	WREG32_SOC15(MMHUB, 0,
+		     regMMVM_L2_CONTEXT1_IDENTITY_APERTURE_HIGH_ADDR_LO32, 0);
+	WREG32_SOC15(MMHUB, 0,
+		     regMMVM_L2_CONTEXT1_IDENTITY_APERTURE_HIGH_ADDR_HI32, 0);
+
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CONTEXT_IDENTITY_PHYSICAL_OFFSET_LO32,
+		     0);
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CONTEXT_IDENTITY_PHYSICAL_OFFSET_HI32,
+		     0);
+}
+
+static void mmhub_v3_0_1_setup_vmid_config(struct amdgpu_device *adev)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
+	int i;
+	uint32_t tmp;
+
+	for (i = 0; i <= 14; i++) {
+		tmp = RREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT1_CNTL, i * hub->ctx_distance);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL, ENABLE_CONTEXT, 1);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL, PAGE_TABLE_DEPTH,
+				    adev->vm_manager.num_level);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT,
+				    1);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    PDE0_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    VALID_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    READ_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    WRITE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    EXECUTE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    PAGE_TABLE_BLOCK_SIZE,
+				    adev->vm_manager.block_size - 9);
+		/* Send no-retry XNACK on fault to suppress VM fault storm. */
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    RETRY_PERMISSION_OR_INVALID_PAGE_FAULT,
+				    !amdgpu_noretry);
+		WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT1_CNTL,
+				    i * hub->ctx_distance, tmp);
+		WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT1_PAGE_TABLE_START_ADDR_LO32,
+				    i * hub->ctx_addr_distance, 0);
+		WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT1_PAGE_TABLE_START_ADDR_HI32,
+				    i * hub->ctx_addr_distance, 0);
+		WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT1_PAGE_TABLE_END_ADDR_LO32,
+				    i * hub->ctx_addr_distance,
+				    lower_32_bits(adev->vm_manager.max_pfn - 1));
+		WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT1_PAGE_TABLE_END_ADDR_HI32,
+				    i * hub->ctx_addr_distance,
+				    upper_32_bits(adev->vm_manager.max_pfn - 1));
+	}
+
+	hub->vm_cntx_cntl = tmp;
+}
+
+static void mmhub_v3_0_1_program_invalidation(struct amdgpu_device *adev)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
+	unsigned i;
+
+	for (i = 0; i < 18; ++i) {
+		WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_INVALIDATE_ENG0_ADDR_RANGE_LO32,
+				    i * hub->eng_addr_distance, 0xffffffff);
+		WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_INVALIDATE_ENG0_ADDR_RANGE_HI32,
+				    i * hub->eng_addr_distance, 0x1f);
+	}
+}
+
+static int mmhub_v3_0_1_gart_enable(struct amdgpu_device *adev)
+{
+	/* GART Enable. */
+	mmhub_v3_0_1_init_gart_aperture_regs(adev);
+	mmhub_v3_0_1_init_system_aperture_regs(adev);
+	mmhub_v3_0_1_init_tlb_regs(adev);
+	mmhub_v3_0_1_init_cache_regs(adev);
+
+	mmhub_v3_0_1_enable_system_domain(adev);
+	mmhub_v3_0_1_disable_identity_aperture(adev);
+	mmhub_v3_0_1_setup_vmid_config(adev);
+	mmhub_v3_0_1_program_invalidation(adev);
+
+	return 0;
+}
+
+static void mmhub_v3_0_1_gart_disable(struct amdgpu_device *adev)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
+	u32 tmp;
+	u32 i;
+
+	/* Disable all tables */
+	for (i = 0; i < 16; i++)
+		WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT0_CNTL,
+				    i * hub->ctx_distance, 0);
+
+	/* Setup TLB control */
+	tmp = RREG32_SOC15(MMHUB, 0, regMMMC_VM_MX_L1_TLB_CNTL);
+	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB, 0);
+	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL,
+			    ENABLE_ADVANCED_DRIVER_MODEL, 0);
+	WREG32_SOC15(MMHUB, 0, regMMMC_VM_MX_L1_TLB_CNTL, tmp);
+
+	/* Setup L2 cache */
+	tmp = RREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL, ENABLE_L2_CACHE, 0);
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL, tmp);
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL3, 0);
+}
+
+/**
+ * mmhub_v3_0_1_set_fault_enable_default - update GART/VM fault handling
+ *
+ * @adev: amdgpu_device pointer
+ * @value: true redirects VM faults to the default page
+ */
+static void mmhub_v3_0_1_set_fault_enable_default(struct amdgpu_device *adev,
+						  bool value)
+{
+	u32 tmp;
+
+	tmp = RREG32_SOC15(MMHUB, 0, regMMVM_L2_PROTECTION_FAULT_CNTL);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    PDE0_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    PDE1_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    PDE2_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    TRANSLATE_FURTHER_PROTECTION_FAULT_ENABLE_DEFAULT,
+			    value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    NACK_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    VALID_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    READ_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    WRITE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    EXECUTE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	if (!value) {
+		tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+				CRASH_ON_NO_RETRY_FAULT, 1);
+		tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+				CRASH_ON_RETRY_FAULT, 1);
+	}
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_PROTECTION_FAULT_CNTL, tmp);
+}
+
+static const struct amdgpu_vmhub_funcs mmhub_v3_0_1_vmhub_funcs = {
+	.print_l2_protection_fault_status = mmhub_v3_0_1_print_l2_protection_fault_status,
+	.get_invalidate_req = mmhub_v3_0_1_get_invalidate_req,
+};
+
+static void mmhub_v3_0_1_init(struct amdgpu_device *adev)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
+
+	hub->ctx0_ptb_addr_lo32 =
+		SOC15_REG_OFFSET(MMHUB, 0,
+				 regMMVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32);
+	hub->ctx0_ptb_addr_hi32 =
+		SOC15_REG_OFFSET(MMHUB, 0,
+				 regMMVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32);
+	hub->vm_inv_eng0_sem =
+		SOC15_REG_OFFSET(MMHUB, 0, regMMVM_INVALIDATE_ENG0_SEM);
+	hub->vm_inv_eng0_req =
+		SOC15_REG_OFFSET(MMHUB, 0, regMMVM_INVALIDATE_ENG0_REQ);
+	hub->vm_inv_eng0_ack =
+		SOC15_REG_OFFSET(MMHUB, 0, regMMVM_INVALIDATE_ENG0_ACK);
+	hub->vm_context0_cntl =
+		SOC15_REG_OFFSET(MMHUB, 0, regMMVM_CONTEXT0_CNTL);
+	hub->vm_l2_pro_fault_status =
+		SOC15_REG_OFFSET(MMHUB, 0, regMMVM_L2_PROTECTION_FAULT_STATUS);
+	hub->vm_l2_pro_fault_cntl =
+		SOC15_REG_OFFSET(MMHUB, 0, regMMVM_L2_PROTECTION_FAULT_CNTL);
+
+	hub->ctx_distance = regMMVM_CONTEXT1_CNTL - regMMVM_CONTEXT0_CNTL;
+	hub->ctx_addr_distance = regMMVM_CONTEXT1_PAGE_TABLE_BASE_ADDR_LO32 -
+		regMMVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32;
+	hub->eng_distance = regMMVM_INVALIDATE_ENG1_REQ -
+		regMMVM_INVALIDATE_ENG0_REQ;
+	hub->eng_addr_distance = regMMVM_INVALIDATE_ENG1_ADDR_RANGE_LO32 -
+		regMMVM_INVALIDATE_ENG0_ADDR_RANGE_LO32;
+
+	hub->vm_cntx_cntl_vm_fault = MMVM_CONTEXT1_CNTL__RANGE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		MMVM_CONTEXT1_CNTL__DUMMY_PAGE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		MMVM_CONTEXT1_CNTL__PDE0_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		MMVM_CONTEXT1_CNTL__VALID_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		MMVM_CONTEXT1_CNTL__READ_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		MMVM_CONTEXT1_CNTL__WRITE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		MMVM_CONTEXT1_CNTL__EXECUTE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK;
+
+	hub->vmhub_funcs = &mmhub_v3_0_1_vmhub_funcs;
+}
+
+static u64 mmhub_v3_0_1_get_fb_location(struct amdgpu_device *adev)
+{
+	u64 base;
+
+	base = RREG32_SOC15(MMHUB, 0, regMMMC_VM_FB_LOCATION_BASE);
+	base &= MMMC_VM_FB_LOCATION_BASE__FB_BASE_MASK;
+	base <<= 24;
+
+	return base;
+}
+
+static u64 mmhub_v3_0_1_get_mc_fb_offset(struct amdgpu_device *adev)
+{
+	return (u64)RREG32_SOC15(MMHUB, 0, regMMMC_VM_FB_OFFSET) << 24;
+}
+
+static void mmhub_v3_0_1_update_medium_grain_clock_gating(struct amdgpu_device *adev,
+							  bool enable)
+{
+	uint32_t def, data;
+
+	def = data = RREG32_SOC15(MMHUB, 0, regMM_ATC_L2_MISC_CG);
+
+	if (enable)
+		data |= MM_ATC_L2_MISC_CG__ENABLE_MASK;
+	else
+		data &= ~MM_ATC_L2_MISC_CG__ENABLE_MASK;
+
+	if (def != data)
+		WREG32_SOC15(MMHUB, 0, regMM_ATC_L2_MISC_CG, data);
+}
+
+static void mmhub_v3_0_1_update_medium_grain_light_sleep(struct amdgpu_device *adev,
+							 bool enable)
+{
+	uint32_t def, data;
+
+	def = data = RREG32_SOC15(MMHUB, 0, regMM_ATC_L2_MISC_CG);
+
+	if (enable)
+		data |= MM_ATC_L2_MISC_CG__MEM_LS_ENABLE_MASK;
+	else
+		data &= ~MM_ATC_L2_MISC_CG__MEM_LS_ENABLE_MASK;
+
+	if (def != data)
+		WREG32_SOC15(MMHUB, 0, regMM_ATC_L2_MISC_CG, data);
+}
+
+static int mmhub_v3_0_1_set_clockgating(struct amdgpu_device *adev,
+					enum amd_clockgating_state state)
+{
+	if (amdgpu_sriov_vf(adev))
+		return 0;
+
+	mmhub_v3_0_1_update_medium_grain_clock_gating(adev,
+			state == AMD_CG_STATE_GATE);
+	mmhub_v3_0_1_update_medium_grain_light_sleep(adev,
+			state == AMD_CG_STATE_GATE);
+	return 0;
+}
+
+static void mmhub_v3_0_1_get_clockgating(struct amdgpu_device *adev, u64 *flags)
+{
+	int data;
+
+	if (amdgpu_sriov_vf(adev))
+		*flags = 0;
+
+	data = RREG32_SOC15(MMHUB, 0, regMM_ATC_L2_MISC_CG);
+
+	/* AMD_CG_SUPPORT_MC_MGCG */
+	if (data & MM_ATC_L2_MISC_CG__ENABLE_MASK)
+		*flags |= AMD_CG_SUPPORT_MC_MGCG;
+
+	/* AMD_CG_SUPPORT_MC_LS */
+	if (data & MM_ATC_L2_MISC_CG__MEM_LS_ENABLE_MASK)
+		*flags |= AMD_CG_SUPPORT_MC_LS;
+}
+
+const struct amdgpu_mmhub_funcs mmhub_v3_0_1_funcs = {
+	.init = mmhub_v3_0_1_init,
+	.get_fb_location = mmhub_v3_0_1_get_fb_location,
+	.get_mc_fb_offset = mmhub_v3_0_1_get_mc_fb_offset,
+	.gart_enable = mmhub_v3_0_1_gart_enable,
+	.set_fault_enable_default = mmhub_v3_0_1_set_fault_enable_default,
+	.gart_disable = mmhub_v3_0_1_gart_disable,
+	.set_clockgating = mmhub_v3_0_1_set_clockgating,
+	.get_clockgating = mmhub_v3_0_1_get_clockgating,
+	.setup_vm_pt_regs = mmhub_v3_0_1_setup_vm_pt_regs,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0_1.h b/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0_1.h
new file mode 100644
index 000000000000..4c1246735e7d
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0_1.h
@@ -0,0 +1,28 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __MMHUB_V3_0_1_H__
+#define __MMHUB_V3_0_1_H__
+
+extern const struct amdgpu_mmhub_funcs mmhub_v3_0_1_funcs;
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0_2.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0_2.c
new file mode 100644
index 000000000000..f0f182f033b9
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0_2.c
@@ -0,0 +1,570 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "amdgpu.h"
+#include "mmhub_v3_0_2.h"
+
+#include "mmhub/mmhub_3_0_2_offset.h"
+#include "mmhub/mmhub_3_0_2_sh_mask.h"
+#include "navi10_enum.h"
+
+#include "soc15_common.h"
+
+#define regMMVM_L2_CNTL3_DEFAULT				0x80100007
+#define regMMVM_L2_CNTL4_DEFAULT				0x000000c1
+#define regMMVM_L2_CNTL5_DEFAULT				0x00003fe0
+
+static const char *mmhub_client_ids_v3_0_2[][2] = {
+	[0][0] = "VMC",
+	[4][0] = "DCEDMC",
+	[5][0] = "DCEVGA",
+	[6][0] = "MP0",
+	[7][0] = "MP1",
+	[8][0] = "MPIO",
+	[16][0] = "HDP",
+	[17][0] = "LSDMA",
+	[18][0] = "JPEG",
+	[19][0] = "VCNU0",
+	[21][0] = "VSCH",
+	[22][0] = "VCNU1",
+	[23][0] = "VCN1",
+	[32+20][0] = "VCN0",
+	[2][1] = "DBGUNBIO",
+	[3][1] = "DCEDWB",
+	[4][1] = "DCEDMC",
+	[5][1] = "DCEVGA",
+	[6][1] = "MP0",
+	[7][1] = "MP1",
+	[8][1] = "MPIO",
+	[10][1] = "DBGU0",
+	[11][1] = "DBGU1",
+	[12][1] = "DBGU2",
+	[13][1] = "DBGU3",
+	[14][1] = "XDP",
+	[15][1] = "OSSSYS",
+	[16][1] = "HDP",
+	[17][1] = "LSDMA",
+	[18][1] = "JPEG",
+	[19][1] = "VCNU0",
+	[20][1] = "VCN0",
+	[21][1] = "VSCH",
+	[22][1] = "VCNU1",
+	[23][1] = "VCN1",
+};
+
+static uint32_t mmhub_v3_0_2_get_invalidate_req(unsigned int vmid,
+					      uint32_t flush_type)
+{
+	u32 req = 0;
+
+	/* invalidate using legacy mode on vmid*/
+	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ,
+			    PER_VMID_INVALIDATE_REQ, 1 << vmid);
+	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ, FLUSH_TYPE, flush_type);
+	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PTES, 1);
+	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PDE0, 1);
+	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PDE1, 1);
+	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PDE2, 1);
+	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ, INVALIDATE_L1_PTES, 1);
+	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ,
+			    CLEAR_PROTECTION_FAULT_STATUS_ADDR,	0);
+
+	return req;
+}
+
+static void
+mmhub_v3_0_2_print_l2_protection_fault_status(struct amdgpu_device *adev,
+					     uint32_t status)
+{
+	uint32_t cid, rw;
+	const char *mmhub_cid = NULL;
+
+	cid = REG_GET_FIELD(status,
+			    MMVM_L2_PROTECTION_FAULT_STATUS, CID);
+	rw = REG_GET_FIELD(status,
+			   MMVM_L2_PROTECTION_FAULT_STATUS, RW);
+
+	dev_err(adev->dev,
+		"MMVM_L2_PROTECTION_FAULT_STATUS:0x%08X\n",
+		status);
+
+	mmhub_cid = mmhub_client_ids_v3_0_2[cid][rw];
+	dev_err(adev->dev, "\t Faulty UTCL2 client ID: %s (0x%x)\n",
+		mmhub_cid ? mmhub_cid : "unknown", cid);
+	dev_err(adev->dev, "\t MORE_FAULTS: 0x%lx\n",
+		REG_GET_FIELD(status,
+		MMVM_L2_PROTECTION_FAULT_STATUS, MORE_FAULTS));
+	dev_err(adev->dev, "\t WALKER_ERROR: 0x%lx\n",
+		REG_GET_FIELD(status,
+		MMVM_L2_PROTECTION_FAULT_STATUS, WALKER_ERROR));
+	dev_err(adev->dev, "\t PERMISSION_FAULTS: 0x%lx\n",
+		REG_GET_FIELD(status,
+		MMVM_L2_PROTECTION_FAULT_STATUS, PERMISSION_FAULTS));
+	dev_err(adev->dev, "\t MAPPING_ERROR: 0x%lx\n",
+		REG_GET_FIELD(status,
+		MMVM_L2_PROTECTION_FAULT_STATUS, MAPPING_ERROR));
+	dev_err(adev->dev, "\t RW: 0x%x\n", rw);
+}
+
+static void mmhub_v3_0_2_setup_vm_pt_regs(struct amdgpu_device *adev, uint32_t vmid,
+				uint64_t page_table_base)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
+
+	WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32,
+			    hub->ctx_addr_distance * vmid,
+			    lower_32_bits(page_table_base));
+
+	WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32,
+			    hub->ctx_addr_distance * vmid,
+			    upper_32_bits(page_table_base));
+}
+
+static void mmhub_v3_0_2_init_gart_aperture_regs(struct amdgpu_device *adev)
+{
+	uint64_t pt_base = amdgpu_gmc_pd_addr(adev->gart.bo);
+
+	mmhub_v3_0_2_setup_vm_pt_regs(adev, 0, pt_base);
+
+	WREG32_SOC15(MMHUB, 0, regMMVM_CONTEXT0_PAGE_TABLE_START_ADDR_LO32,
+		     (u32)(adev->gmc.gart_start >> 12));
+	WREG32_SOC15(MMHUB, 0, regMMVM_CONTEXT0_PAGE_TABLE_START_ADDR_HI32,
+		     (u32)(adev->gmc.gart_start >> 44));
+
+	WREG32_SOC15(MMHUB, 0, regMMVM_CONTEXT0_PAGE_TABLE_END_ADDR_LO32,
+		     (u32)(adev->gmc.gart_end >> 12));
+	WREG32_SOC15(MMHUB, 0, regMMVM_CONTEXT0_PAGE_TABLE_END_ADDR_HI32,
+		     (u32)(adev->gmc.gart_end >> 44));
+}
+
+static void mmhub_v3_0_2_init_system_aperture_regs(struct amdgpu_device *adev)
+{
+	uint64_t value;
+	uint32_t tmp;
+
+	/* Program the AGP BAR */
+	WREG32_SOC15(MMHUB, 0, regMMMC_VM_AGP_BASE, 0);
+	WREG32_SOC15(MMHUB, 0, regMMMC_VM_AGP_BOT, adev->gmc.agp_start >> 24);
+	WREG32_SOC15(MMHUB, 0, regMMMC_VM_AGP_TOP, adev->gmc.agp_end >> 24);
+
+	if (!amdgpu_sriov_vf(adev)) {
+		/*
+		 * the new L1 policy will block SRIOV guest from writing
+		 * these regs, and they will be programed at host.
+		 * so skip programing these regs.
+		 */
+		/* Program the system aperture low logical page number. */
+		WREG32_SOC15(MMHUB, 0, regMMMC_VM_SYSTEM_APERTURE_LOW_ADDR,
+			     min(adev->gmc.fb_start, adev->gmc.agp_start) >> 18);
+		WREG32_SOC15(MMHUB, 0, regMMMC_VM_SYSTEM_APERTURE_HIGH_ADDR,
+			     max(adev->gmc.fb_end, adev->gmc.agp_end) >> 18);
+	}
+
+	/* Set default page address. */
+	value = amdgpu_gmc_vram_mc2pa(adev, adev->mem_scratch.gpu_addr);
+	WREG32_SOC15(MMHUB, 0, regMMMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB,
+		     (u32)(value >> 12));
+	WREG32_SOC15(MMHUB, 0, regMMMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_MSB,
+		     (u32)(value >> 44));
+
+	/* Program "protection fault". */
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_PROTECTION_FAULT_DEFAULT_ADDR_LO32,
+		     (u32)(adev->dummy_page_addr >> 12));
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_PROTECTION_FAULT_DEFAULT_ADDR_HI32,
+		     (u32)((u64)adev->dummy_page_addr >> 44));
+
+	tmp = RREG32_SOC15(MMHUB, 0, regMMVM_L2_PROTECTION_FAULT_CNTL2);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL2,
+			    ACTIVE_PAGE_MIGRATION_PTE_READ_RETRY, 1);
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_PROTECTION_FAULT_CNTL2, tmp);
+}
+
+static void mmhub_v3_0_2_init_tlb_regs(struct amdgpu_device *adev)
+{
+	uint32_t tmp;
+
+	/* Setup TLB control */
+	tmp = RREG32_SOC15(MMHUB, 0, regMMMC_VM_MX_L1_TLB_CNTL);
+
+	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB, 1);
+	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL, SYSTEM_ACCESS_MODE, 3);
+	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL,
+			    ENABLE_ADVANCED_DRIVER_MODEL, 1);
+	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL,
+			    SYSTEM_APERTURE_UNMAPPED_ACCESS, 0);
+	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL, ECO_BITS, 0);
+	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL,
+			    MTYPE, MTYPE_UC); /* UC, uncached */
+
+	WREG32_SOC15(MMHUB, 0, regMMMC_VM_MX_L1_TLB_CNTL, tmp);
+}
+
+static void mmhub_v3_0_2_init_cache_regs(struct amdgpu_device *adev)
+{
+	uint32_t tmp;
+
+	/* These registers are not accessible to VF-SRIOV.
+	 * The PF will program them instead.
+	 */
+	if (amdgpu_sriov_vf(adev))
+		return;
+
+	/* Setup L2 cache */
+	tmp = RREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL, ENABLE_L2_CACHE, 1);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL, ENABLE_L2_FRAGMENT_PROCESSING, 0);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL,
+			    ENABLE_DEFAULT_PAGE_OUT_TO_SYSTEM_MEMORY, 1);
+	/* XXX for emulation, Refer to closed source code.*/
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL, L2_PDE0_CACHE_TAG_GENERATION_MODE,
+			    0);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL, PDE_FAULT_CLASSIFICATION, 0);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL, CONTEXT1_IDENTITY_ACCESS_MODE, 1);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL, IDENTITY_MODE_FRAGMENT_SIZE, 0);
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL, tmp);
+
+	tmp = RREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL2);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL2, INVALIDATE_ALL_L1_TLBS, 1);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL2, INVALIDATE_L2_CACHE, 1);
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL2, tmp);
+
+	tmp = regMMVM_L2_CNTL3_DEFAULT;
+	if (adev->gmc.translate_further) {
+		tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL3, BANK_SELECT, 12);
+		tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL3,
+				    L2_CACHE_BIGK_FRAGMENT_SIZE, 9);
+	} else {
+		tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL3, BANK_SELECT, 9);
+		tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL3,
+				    L2_CACHE_BIGK_FRAGMENT_SIZE, 6);
+	}
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL3, tmp);
+
+	tmp = regMMVM_L2_CNTL4_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL4, VMC_TAP_PDE_REQUEST_PHYSICAL, 0);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL4, VMC_TAP_PTE_REQUEST_PHYSICAL, 0);
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL4, tmp);
+
+	tmp = regMMVM_L2_CNTL5_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL5, L2_CACHE_SMALLK_FRAGMENT_SIZE, 0);
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL5, tmp);
+}
+
+static void mmhub_v3_0_2_enable_system_domain(struct amdgpu_device *adev)
+{
+	uint32_t tmp;
+
+	tmp = RREG32_SOC15(MMHUB, 0, regMMVM_CONTEXT0_CNTL);
+	tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT0_CNTL, ENABLE_CONTEXT, 1);
+	tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT0_CNTL, PAGE_TABLE_DEPTH, 0);
+	tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT0_CNTL,
+			    RETRY_PERMISSION_OR_INVALID_PAGE_FAULT, 0);
+	WREG32_SOC15(MMHUB, 0, regMMVM_CONTEXT0_CNTL, tmp);
+}
+
+static void mmhub_v3_0_2_disable_identity_aperture(struct amdgpu_device *adev)
+{
+	/* These registers are not accessible to VF-SRIOV.
+	 * The PF will program them instead.
+	 */
+	if (amdgpu_sriov_vf(adev))
+		return;
+
+	WREG32_SOC15(MMHUB, 0,
+		     regMMVM_L2_CONTEXT1_IDENTITY_APERTURE_LOW_ADDR_LO32,
+		     0xFFFFFFFF);
+	WREG32_SOC15(MMHUB, 0,
+		     regMMVM_L2_CONTEXT1_IDENTITY_APERTURE_LOW_ADDR_HI32,
+		     0x0000000F);
+
+	WREG32_SOC15(MMHUB, 0,
+		     regMMVM_L2_CONTEXT1_IDENTITY_APERTURE_HIGH_ADDR_LO32, 0);
+	WREG32_SOC15(MMHUB, 0,
+		     regMMVM_L2_CONTEXT1_IDENTITY_APERTURE_HIGH_ADDR_HI32, 0);
+
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CONTEXT_IDENTITY_PHYSICAL_OFFSET_LO32,
+		     0);
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CONTEXT_IDENTITY_PHYSICAL_OFFSET_HI32,
+		     0);
+}
+
+static void mmhub_v3_0_2_setup_vmid_config(struct amdgpu_device *adev)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
+	int i;
+	uint32_t tmp;
+
+	for (i = 0; i <= 14; i++) {
+		tmp = RREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT1_CNTL, i * hub->ctx_distance);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL, ENABLE_CONTEXT, 1);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL, PAGE_TABLE_DEPTH,
+				    adev->vm_manager.num_level);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT,
+				    1);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    PDE0_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    VALID_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    READ_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    WRITE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    EXECUTE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    PAGE_TABLE_BLOCK_SIZE,
+				    adev->vm_manager.block_size - 9);
+		/* Send no-retry XNACK on fault to suppress VM fault storm. */
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    RETRY_PERMISSION_OR_INVALID_PAGE_FAULT,
+				    !amdgpu_noretry);
+		WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT1_CNTL,
+				    i * hub->ctx_distance, tmp);
+		WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT1_PAGE_TABLE_START_ADDR_LO32,
+				    i * hub->ctx_addr_distance, 0);
+		WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT1_PAGE_TABLE_START_ADDR_HI32,
+				    i * hub->ctx_addr_distance, 0);
+		WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT1_PAGE_TABLE_END_ADDR_LO32,
+				    i * hub->ctx_addr_distance,
+				    lower_32_bits(adev->vm_manager.max_pfn - 1));
+		WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT1_PAGE_TABLE_END_ADDR_HI32,
+				    i * hub->ctx_addr_distance,
+				    upper_32_bits(adev->vm_manager.max_pfn - 1));
+	}
+
+	hub->vm_cntx_cntl = tmp;
+}
+
+static void mmhub_v3_0_2_program_invalidation(struct amdgpu_device *adev)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
+	unsigned i;
+
+	for (i = 0; i < 18; ++i) {
+		WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_INVALIDATE_ENG0_ADDR_RANGE_LO32,
+				    i * hub->eng_addr_distance, 0xffffffff);
+		WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_INVALIDATE_ENG0_ADDR_RANGE_HI32,
+				    i * hub->eng_addr_distance, 0x1f);
+	}
+}
+
+static int mmhub_v3_0_2_gart_enable(struct amdgpu_device *adev)
+{
+	/* GART Enable. */
+	mmhub_v3_0_2_init_gart_aperture_regs(adev);
+	mmhub_v3_0_2_init_system_aperture_regs(adev);
+	mmhub_v3_0_2_init_tlb_regs(adev);
+	mmhub_v3_0_2_init_cache_regs(adev);
+
+	mmhub_v3_0_2_enable_system_domain(adev);
+	mmhub_v3_0_2_disable_identity_aperture(adev);
+	mmhub_v3_0_2_setup_vmid_config(adev);
+	mmhub_v3_0_2_program_invalidation(adev);
+
+	return 0;
+}
+
+static void mmhub_v3_0_2_gart_disable(struct amdgpu_device *adev)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
+	u32 tmp;
+	u32 i;
+
+	/* Disable all tables */
+	for (i = 0; i < 16; i++)
+		WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT0_CNTL,
+				    i * hub->ctx_distance, 0);
+
+	/* Setup TLB control */
+	tmp = RREG32_SOC15(MMHUB, 0, regMMMC_VM_MX_L1_TLB_CNTL);
+	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB, 0);
+	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL,
+			    ENABLE_ADVANCED_DRIVER_MODEL, 0);
+	WREG32_SOC15(MMHUB, 0, regMMMC_VM_MX_L1_TLB_CNTL, tmp);
+
+	/* Setup L2 cache */
+	tmp = RREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL, ENABLE_L2_CACHE, 0);
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL, tmp);
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL3, 0);
+}
+
+/**
+ * mmhub_v3_0_2_set_fault_enable_default - update GART/VM fault handling
+ *
+ * @adev: amdgpu_device pointer
+ * @value: true redirects VM faults to the default page
+ */
+static void mmhub_v3_0_2_set_fault_enable_default(struct amdgpu_device *adev, bool value)
+{
+	u32 tmp;
+
+	/* These registers are not accessible to VF-SRIOV.
+	 * The PF will program them instead.
+	 */
+	if (amdgpu_sriov_vf(adev))
+		return;
+
+	tmp = RREG32_SOC15(MMHUB, 0, regMMVM_L2_PROTECTION_FAULT_CNTL);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    PDE0_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    PDE1_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    PDE2_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    TRANSLATE_FURTHER_PROTECTION_FAULT_ENABLE_DEFAULT,
+			    value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    NACK_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    VALID_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    READ_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    WRITE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    EXECUTE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	if (!value) {
+		tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+				CRASH_ON_NO_RETRY_FAULT, 1);
+		tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+				CRASH_ON_RETRY_FAULT, 1);
+	}
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_PROTECTION_FAULT_CNTL, tmp);
+}
+
+static const struct amdgpu_vmhub_funcs mmhub_v3_0_2_vmhub_funcs = {
+	.print_l2_protection_fault_status = mmhub_v3_0_2_print_l2_protection_fault_status,
+	.get_invalidate_req = mmhub_v3_0_2_get_invalidate_req,
+};
+
+static void mmhub_v3_0_2_init(struct amdgpu_device *adev)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
+
+	hub->ctx0_ptb_addr_lo32 =
+		SOC15_REG_OFFSET(MMHUB, 0,
+				 regMMVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32);
+	hub->ctx0_ptb_addr_hi32 =
+		SOC15_REG_OFFSET(MMHUB, 0,
+				 regMMVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32);
+	hub->vm_inv_eng0_sem =
+		SOC15_REG_OFFSET(MMHUB, 0, regMMVM_INVALIDATE_ENG0_SEM);
+	hub->vm_inv_eng0_req =
+		SOC15_REG_OFFSET(MMHUB, 0, regMMVM_INVALIDATE_ENG0_REQ);
+	hub->vm_inv_eng0_ack =
+		SOC15_REG_OFFSET(MMHUB, 0, regMMVM_INVALIDATE_ENG0_ACK);
+	hub->vm_context0_cntl =
+		SOC15_REG_OFFSET(MMHUB, 0, regMMVM_CONTEXT0_CNTL);
+	hub->vm_l2_pro_fault_status =
+		SOC15_REG_OFFSET(MMHUB, 0, regMMVM_L2_PROTECTION_FAULT_STATUS);
+	hub->vm_l2_pro_fault_cntl =
+		SOC15_REG_OFFSET(MMHUB, 0, regMMVM_L2_PROTECTION_FAULT_CNTL);
+
+	hub->ctx_distance = regMMVM_CONTEXT1_CNTL - regMMVM_CONTEXT0_CNTL;
+	hub->ctx_addr_distance = regMMVM_CONTEXT1_PAGE_TABLE_BASE_ADDR_LO32 -
+		regMMVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32;
+	hub->eng_distance = regMMVM_INVALIDATE_ENG1_REQ -
+		regMMVM_INVALIDATE_ENG0_REQ;
+	hub->eng_addr_distance = regMMVM_INVALIDATE_ENG1_ADDR_RANGE_LO32 -
+		regMMVM_INVALIDATE_ENG0_ADDR_RANGE_LO32;
+
+	hub->vm_cntx_cntl_vm_fault = MMVM_CONTEXT1_CNTL__RANGE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		MMVM_CONTEXT1_CNTL__DUMMY_PAGE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		MMVM_CONTEXT1_CNTL__PDE0_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		MMVM_CONTEXT1_CNTL__VALID_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		MMVM_CONTEXT1_CNTL__READ_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		MMVM_CONTEXT1_CNTL__WRITE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		MMVM_CONTEXT1_CNTL__EXECUTE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK;
+
+	hub->vm_l2_bank_select_reserved_cid2 =
+		SOC15_REG_OFFSET(MMHUB, 0, regMMVM_L2_BANK_SELECT_RESERVED_CID2);
+
+	hub->vmhub_funcs = &mmhub_v3_0_2_vmhub_funcs;
+}
+
+static u64 mmhub_v3_0_2_get_fb_location(struct amdgpu_device *adev)
+{
+	u64 base;
+
+	base = RREG32_SOC15(MMHUB, 0, regMMMC_VM_FB_LOCATION_BASE);
+	base &= MMMC_VM_FB_LOCATION_BASE__FB_BASE_MASK;
+	base <<= 24;
+
+	return base;
+}
+
+static u64 mmhub_v3_0_2_get_mc_fb_offset(struct amdgpu_device *adev)
+{
+	return (u64)RREG32_SOC15(MMHUB, 0, regMMMC_VM_FB_OFFSET) << 24;
+}
+
+static void mmhub_v3_0_2_update_medium_grain_clock_gating(struct amdgpu_device *adev,
+							bool enable)
+{
+	//TODO
+}
+
+static void mmhub_v3_0_2_update_medium_grain_light_sleep(struct amdgpu_device *adev,
+						       bool enable)
+{
+	//TODO
+}
+
+static int mmhub_v3_0_2_set_clockgating(struct amdgpu_device *adev,
+			       enum amd_clockgating_state state)
+{
+	if (amdgpu_sriov_vf(adev))
+		return 0;
+
+	mmhub_v3_0_2_update_medium_grain_clock_gating(adev,
+			state == AMD_CG_STATE_GATE);
+	mmhub_v3_0_2_update_medium_grain_light_sleep(adev,
+			state == AMD_CG_STATE_GATE);
+	return 0;
+}
+
+static void mmhub_v3_0_2_get_clockgating(struct amdgpu_device *adev, u64 *flags)
+{
+	//TODO
+}
+
+const struct amdgpu_mmhub_funcs mmhub_v3_0_2_funcs = {
+	.init = mmhub_v3_0_2_init,
+	.get_fb_location = mmhub_v3_0_2_get_fb_location,
+	.get_mc_fb_offset = mmhub_v3_0_2_get_mc_fb_offset,
+	.gart_enable = mmhub_v3_0_2_gart_enable,
+	.set_fault_enable_default = mmhub_v3_0_2_set_fault_enable_default,
+	.gart_disable = mmhub_v3_0_2_gart_disable,
+	.set_clockgating = mmhub_v3_0_2_set_clockgating,
+	.get_clockgating = mmhub_v3_0_2_get_clockgating,
+	.setup_vm_pt_regs = mmhub_v3_0_2_setup_vm_pt_regs,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0_2.h b/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0_2.h
new file mode 100644
index 000000000000..23ad7b156cdb
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v3_0_2.h
@@ -0,0 +1,28 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __MMHUB_V3_0_2_H__
+#define __MMHUB_V3_0_2_H__
+
+extern const struct amdgpu_mmhub_funcs mmhub_v3_0_2_funcs;
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v3_3.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v3_3.c
new file mode 100644
index 000000000000..f6fc9778bc30
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v3_3.c
@@ -0,0 +1,746 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "amdgpu.h"
+#include "mmhub_v3_3.h"
+
+#include "mmhub/mmhub_3_3_0_offset.h"
+#include "mmhub/mmhub_3_3_0_sh_mask.h"
+
+#include "navi10_enum.h"
+#include "soc15_common.h"
+
+#define regMMVM_L2_CNTL3_DEFAULT				0x80100007
+#define regMMVM_L2_CNTL4_DEFAULT				0x000000c1
+#define regMMVM_L2_CNTL5_DEFAULT				0x00003fe0
+#define regDAGB0_L1TLB_REG_RW_3_3                   0x00a4
+#define regDAGB0_L1TLB_REG_RW_3_3_BASE_IDX          1
+#define regDAGB1_L1TLB_REG_RW_3_3                   0x0163
+#define regDAGB1_L1TLB_REG_RW_3_3_BASE_IDX          1
+
+static const char *mmhub_client_ids_v3_3[][2] = {
+	[0][0] = "VMC",
+	[1][0] = "ISPXT",
+	[2][0] = "ISPIXT",
+	[4][0] = "DCEDMC",
+	[6][0] = "MP0",
+	[7][0] = "MP1",
+	[8][0] = "MPM",
+	[9][0] = "ISPPDPRD",
+	[10][0] = "ISPCSTATRD",
+	[11][0] = "ISPBYRPRD",
+	[12][0] = "ISPRGBPRD",
+	[13][0] = "ISPMCFPRD",
+	[14][0] = "ISPMCFPRD1",
+	[15][0] = "ISPYUVPRD",
+	[16][0] = "ISPMCSCRD",
+	[17][0] = "ISPGDCRD",
+	[18][0] = "ISPLMERD",
+	[22][0] = "ISPXT1",
+	[23][0] = "ISPIXT1",
+	[24][0] = "HDP",
+	[25][0] = "LSDMA",
+	[26][0] = "JPEG",
+	[27][0] = "VPE",
+	[28][0] = "VSCH",
+	[29][0] = "VCNU",
+	[30][0] = "VCN",
+	[1][1] = "ISPXT",
+	[2][1] = "ISPIXT",
+	[3][1] = "DCEDWB",
+	[4][1] = "DCEDMC",
+	[5][1] = "ISPCSISWR",
+	[6][1] = "MP0",
+	[7][1] = "MP1",
+	[8][1] = "MPM",
+	[9][1] = "ISPPDPWR",
+	[10][1] = "ISPCSTATWR",
+	[11][1] = "ISPBYRPWR",
+	[12][1] = "ISPRGBPWR",
+	[13][1] = "ISPMCFPWR",
+	[14][1] = "ISPMWR0",
+	[15][1] = "ISPYUVPWR",
+	[16][1] = "ISPMCSCWR",
+	[17][1] = "ISPGDCWR",
+	[18][1] = "ISPLMEWR",
+	[20][1] = "ISPMWR2",
+	[21][1] = "OSSSYS",
+	[22][1] = "ISPXT1",
+	[23][1] = "ISPIXT1",
+	[24][1] = "HDP",
+	[25][1] = "LSDMA",
+	[26][1] = "JPEG",
+	[27][1] = "VPE",
+	[28][1] = "VSCH",
+	[29][1] = "VCNU",
+	[30][1] = "VCN",
+};
+
+static const char *mmhub_client_ids_v3_3_1[][2] = {
+	[0][0] = "VMC",
+	[4][0] = "DCEDMC",
+	[6][0] = "MP0",
+	[7][0] = "MP1",
+	[8][0] = "MPM",
+	[24][0] = "HDP",
+	[25][0] = "LSDMA",
+	[26][0] = "JPEG0",
+	[27][0] = "VPE0",
+	[28][0] = "VSCH",
+	[29][0] = "VCNU0",
+	[30][0] = "VCN0",
+	[32+1][0] = "ISPXT",
+	[32+2][0] = "ISPIXT",
+	[32+9][0] = "ISPPDPRD",
+	[32+10][0] = "ISPCSTATRD",
+	[32+11][0] = "ISPBYRPRD",
+	[32+12][0] = "ISPRGBPRD",
+	[32+13][0] = "ISPMCFPRD",
+	[32+14][0] = "ISPMCFPRD1",
+	[32+15][0] = "ISPYUVPRD",
+	[32+16][0] = "ISPMCSCRD",
+	[32+17][0] = "ISPGDCRD",
+	[32+18][0] = "ISPLMERD",
+	[32+22][0] = "ISPXT1",
+	[32+23][0] = "ISPIXT1",
+	[32+26][0] = "JPEG1",
+	[32+27][0] = "VPE1",
+	[32+29][0] = "VCNU1",
+	[32+30][0] = "VCN1",
+	[3][1] = "DCEDWB",
+	[4][1] = "DCEDMC",
+	[6][1] = "MP0",
+	[7][1] = "MP1",
+	[8][1] = "MPM",
+	[21][1] = "OSSSYS",
+	[24][1] = "HDP",
+	[25][1] = "LSDMA",
+	[26][1] = "JPEG0",
+	[27][1] = "VPE0",
+	[28][1] = "VSCH",
+	[29][1] = "VCNU0",
+	[30][1] = "VCN0",
+	[32+1][1] = "ISPXT",
+	[32+2][1] = "ISPIXT",
+	[32+5][1] = "ISPCSISWR",
+	[32+9][1] = "ISPPDPWR",
+	[32+10][1] = "ISPCSTATWR",
+	[32+11][1] = "ISPBYRPWR",
+	[32+12][1] = "ISPRGBPWR",
+	[32+13][1] = "ISPMCFPWR",
+	[32+14][1] = "ISPMWR0",
+	[32+15][1] = "ISPYUVPWR",
+	[32+16][1] = "ISPMCSCWR",
+	[32+17][1] = "ISPGDCWR",
+	[32+18][1] = "ISPLMEWR",
+	[32+19][1] = "ISPMWR1",
+	[32+20][1] = "ISPMWR2",
+	[32+22][1] = "ISPXT1",
+	[32+23][1] = "ISPIXT1",
+	[32+26][1] = "JPEG1",
+	[32+27][1] = "VPE1",
+	[32+29][1] = "VCNU1",
+	[32+30][1] = "VCN1",
+};
+
+static uint32_t mmhub_v3_3_get_invalidate_req(unsigned int vmid,
+						uint32_t flush_type)
+{
+	u32 req = 0;
+
+	/* invalidate using legacy mode on vmid*/
+	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ,
+			    PER_VMID_INVALIDATE_REQ, 1 << vmid);
+	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ, FLUSH_TYPE, flush_type ? : 1);
+	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PTES, 1);
+	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PDE0, 1);
+	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PDE1, 1);
+	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PDE2, 1);
+	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ, INVALIDATE_L1_PTES, 1);
+	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ,
+			    CLEAR_PROTECTION_FAULT_STATUS_ADDR,	0);
+
+	return req;
+}
+
+static void
+mmhub_v3_3_print_l2_protection_fault_status(struct amdgpu_device *adev,
+					      uint32_t status)
+{
+	uint32_t cid, rw;
+	const char *mmhub_cid = NULL;
+
+	cid = REG_GET_FIELD(status,
+			    MMVM_L2_PROTECTION_FAULT_STATUS, CID);
+	rw = REG_GET_FIELD(status,
+			   MMVM_L2_PROTECTION_FAULT_STATUS, RW);
+
+	dev_err(adev->dev,
+		"MMVM_L2_PROTECTION_FAULT_STATUS:0x%08X\n",
+		status);
+
+	switch (amdgpu_ip_version(adev, MMHUB_HWIP, 0)) {
+	case IP_VERSION(3, 3, 0):
+	case IP_VERSION(3, 3, 2):
+		mmhub_cid = cid < ARRAY_SIZE(mmhub_client_ids_v3_3) ?
+			    mmhub_client_ids_v3_3[cid][rw] :
+			    cid == 0x140 ? "UMSCH" : NULL;
+		break;
+	case IP_VERSION(3, 3, 1):
+		mmhub_cid = cid < ARRAY_SIZE(mmhub_client_ids_v3_3_1) ?
+			    mmhub_client_ids_v3_3_1[cid][rw] :
+			    cid == 0x140 ? "UMSCH" : NULL;
+		break;
+	default:
+		mmhub_cid = NULL;
+		break;
+	}
+
+	dev_err(adev->dev, "\t Faulty UTCL2 client ID: %s (0x%x)\n",
+		mmhub_cid ? mmhub_cid : "unknown", cid);
+	dev_err(adev->dev, "\t MORE_FAULTS: 0x%lx\n",
+		REG_GET_FIELD(status,
+		MMVM_L2_PROTECTION_FAULT_STATUS, MORE_FAULTS));
+	dev_err(adev->dev, "\t WALKER_ERROR: 0x%lx\n",
+		REG_GET_FIELD(status,
+		MMVM_L2_PROTECTION_FAULT_STATUS, WALKER_ERROR));
+	dev_err(adev->dev, "\t PERMISSION_FAULTS: 0x%lx\n",
+		REG_GET_FIELD(status,
+		MMVM_L2_PROTECTION_FAULT_STATUS, PERMISSION_FAULTS));
+	dev_err(adev->dev, "\t MAPPING_ERROR: 0x%lx\n",
+		REG_GET_FIELD(status,
+		MMVM_L2_PROTECTION_FAULT_STATUS, MAPPING_ERROR));
+	dev_err(adev->dev, "\t RW: 0x%x\n", rw);
+}
+
+static void mmhub_v3_3_setup_vm_pt_regs(struct amdgpu_device *adev,
+					  uint32_t vmid,
+					  uint64_t page_table_base)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
+
+	WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32,
+			    hub->ctx_addr_distance * vmid,
+			    lower_32_bits(page_table_base));
+
+	WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32,
+			    hub->ctx_addr_distance * vmid,
+			    upper_32_bits(page_table_base));
+
+}
+
+static void mmhub_v3_3_init_gart_aperture_regs(struct amdgpu_device *adev)
+{
+	uint64_t pt_base = amdgpu_gmc_pd_addr(adev->gart.bo);
+
+	mmhub_v3_3_setup_vm_pt_regs(adev, 0, pt_base);
+
+	WREG32_SOC15(MMHUB, 0, regMMVM_CONTEXT0_PAGE_TABLE_START_ADDR_LO32,
+		     (u32)(adev->gmc.gart_start >> 12));
+	WREG32_SOC15(MMHUB, 0, regMMVM_CONTEXT0_PAGE_TABLE_START_ADDR_HI32,
+		     (u32)(adev->gmc.gart_start >> 44));
+
+	WREG32_SOC15(MMHUB, 0, regMMVM_CONTEXT0_PAGE_TABLE_END_ADDR_LO32,
+		     (u32)(adev->gmc.gart_end >> 12));
+	WREG32_SOC15(MMHUB, 0, regMMVM_CONTEXT0_PAGE_TABLE_END_ADDR_HI32,
+		     (u32)(adev->gmc.gart_end >> 44));
+}
+
+static void mmhub_v3_3_init_system_aperture_regs(struct amdgpu_device *adev)
+{
+	uint64_t value;
+	uint32_t tmp;
+
+	/* Program the AGP BAR */
+	WREG32_SOC15(MMHUB, 0, regMMMC_VM_AGP_BASE, 0);
+	WREG32_SOC15(MMHUB, 0, regMMMC_VM_AGP_BOT, adev->gmc.agp_start >> 24);
+	WREG32_SOC15(MMHUB, 0, regMMMC_VM_AGP_TOP, adev->gmc.agp_end >> 24);
+
+	/*
+	 * the new L1 policy will block SRIOV guest from writing
+	 * these regs, and they will be programed at host.
+	 * so skip programing these regs.
+	 */
+	/* Program the system aperture low logical page number. */
+	WREG32_SOC15(MMHUB, 0, regMMMC_VM_SYSTEM_APERTURE_LOW_ADDR,
+		     min(adev->gmc.fb_start, adev->gmc.agp_start) >> 18);
+	WREG32_SOC15(MMHUB, 0, regMMMC_VM_SYSTEM_APERTURE_HIGH_ADDR,
+		     max(adev->gmc.fb_end, adev->gmc.agp_end) >> 18);
+
+	/* Set default page address. */
+	value = amdgpu_gmc_vram_mc2pa(adev, adev->mem_scratch.gpu_addr);
+	WREG32_SOC15(MMHUB, 0, regMMMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB,
+		     (u32)(value >> 12));
+	WREG32_SOC15(MMHUB, 0, regMMMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_MSB,
+		     (u32)(value >> 44));
+
+	/* Program "protection fault". */
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_PROTECTION_FAULT_DEFAULT_ADDR_LO32,
+		     (u32)(adev->dummy_page_addr >> 12));
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_PROTECTION_FAULT_DEFAULT_ADDR_HI32,
+		     (u32)((u64)adev->dummy_page_addr >> 44));
+
+	tmp = RREG32_SOC15(MMHUB, 0, regMMVM_L2_PROTECTION_FAULT_CNTL2);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL2,
+			    ACTIVE_PAGE_MIGRATION_PTE_READ_RETRY, 1);
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_PROTECTION_FAULT_CNTL2, tmp);
+}
+
+static void mmhub_v3_3_init_tlb_regs(struct amdgpu_device *adev)
+{
+	uint32_t tmp;
+
+	/* Setup TLB control */
+	tmp = RREG32_SOC15(MMHUB, 0, regMMMC_VM_MX_L1_TLB_CNTL);
+
+	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB, 1);
+	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL, SYSTEM_ACCESS_MODE, 3);
+	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL,
+			    ENABLE_ADVANCED_DRIVER_MODEL, 1);
+	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL,
+			    SYSTEM_APERTURE_UNMAPPED_ACCESS, 0);
+	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL, ECO_BITS, 0);
+	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL,
+			    MTYPE, MTYPE_UC); /* UC, uncached */
+
+	WREG32_SOC15(MMHUB, 0, regMMMC_VM_MX_L1_TLB_CNTL, tmp);
+}
+
+static void mmhub_v3_3_init_cache_regs(struct amdgpu_device *adev)
+{
+	uint32_t tmp;
+
+	/* Setup L2 cache */
+	tmp = RREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL, ENABLE_L2_CACHE, 1);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL, ENABLE_L2_FRAGMENT_PROCESSING, 0);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL,
+			    ENABLE_DEFAULT_PAGE_OUT_TO_SYSTEM_MEMORY, 1);
+	/* XXX for emulation, Refer to closed source code.*/
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL, L2_PDE0_CACHE_TAG_GENERATION_MODE,
+			    0);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL, PDE_FAULT_CLASSIFICATION, 0);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL, CONTEXT1_IDENTITY_ACCESS_MODE, 1);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL, IDENTITY_MODE_FRAGMENT_SIZE, 0);
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL, tmp);
+
+	tmp = RREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL2);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL2, INVALIDATE_ALL_L1_TLBS, 1);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL2, INVALIDATE_L2_CACHE, 1);
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL2, tmp);
+
+	tmp = regMMVM_L2_CNTL3_DEFAULT;
+	if (adev->gmc.translate_further) {
+		tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL3, BANK_SELECT, 12);
+		tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL3,
+				    L2_CACHE_BIGK_FRAGMENT_SIZE, 9);
+	} else {
+		tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL3, BANK_SELECT, 9);
+		tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL3,
+				    L2_CACHE_BIGK_FRAGMENT_SIZE, 6);
+	}
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL3, tmp);
+
+	tmp = regMMVM_L2_CNTL4_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL4, VMC_TAP_PDE_REQUEST_PHYSICAL, 0);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL4, VMC_TAP_PTE_REQUEST_PHYSICAL, 0);
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL4, tmp);
+
+	tmp = regMMVM_L2_CNTL5_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL5, L2_CACHE_SMALLK_FRAGMENT_SIZE, 0);
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL5, tmp);
+}
+
+static void mmhub_v3_3_enable_system_domain(struct amdgpu_device *adev)
+{
+	uint32_t tmp;
+
+	tmp = RREG32_SOC15(MMHUB, 0, regMMVM_CONTEXT0_CNTL);
+	tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT0_CNTL, ENABLE_CONTEXT, 1);
+	tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT0_CNTL, PAGE_TABLE_DEPTH, 0);
+	tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT0_CNTL,
+			    RETRY_PERMISSION_OR_INVALID_PAGE_FAULT, 0);
+
+	WREG32_SOC15(MMHUB, 0, regMMVM_CONTEXT0_CNTL, tmp);
+}
+
+static void mmhub_v3_3_disable_identity_aperture(struct amdgpu_device *adev)
+{
+	WREG32_SOC15(MMHUB, 0,
+		     regMMVM_L2_CONTEXT1_IDENTITY_APERTURE_LOW_ADDR_LO32,
+		     0xFFFFFFFF);
+	WREG32_SOC15(MMHUB, 0,
+		     regMMVM_L2_CONTEXT1_IDENTITY_APERTURE_LOW_ADDR_HI32,
+		     0x0000000F);
+
+	WREG32_SOC15(MMHUB, 0,
+		     regMMVM_L2_CONTEXT1_IDENTITY_APERTURE_HIGH_ADDR_LO32, 0);
+	WREG32_SOC15(MMHUB, 0,
+		     regMMVM_L2_CONTEXT1_IDENTITY_APERTURE_HIGH_ADDR_HI32, 0);
+
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CONTEXT_IDENTITY_PHYSICAL_OFFSET_LO32,
+		     0);
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CONTEXT_IDENTITY_PHYSICAL_OFFSET_HI32,
+		     0);
+}
+
+static void mmhub_v3_3_setup_vmid_config(struct amdgpu_device *adev)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
+	int i;
+	uint32_t tmp;
+
+	for (i = 0; i <= 14; i++) {
+		tmp = RREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT1_CNTL, i * hub->ctx_distance);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL, ENABLE_CONTEXT, 1);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL, PAGE_TABLE_DEPTH,
+				    adev->vm_manager.num_level);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT,
+				    1);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    PDE0_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    VALID_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    READ_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    WRITE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    EXECUTE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    PAGE_TABLE_BLOCK_SIZE,
+				    adev->vm_manager.block_size - 9);
+		/* Send no-retry XNACK on fault to suppress VM fault storm. */
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    RETRY_PERMISSION_OR_INVALID_PAGE_FAULT,
+				    !amdgpu_noretry);
+		WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT1_CNTL,
+				    i * hub->ctx_distance, tmp);
+		WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT1_PAGE_TABLE_START_ADDR_LO32,
+				    i * hub->ctx_addr_distance, 0);
+		WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT1_PAGE_TABLE_START_ADDR_HI32,
+				    i * hub->ctx_addr_distance, 0);
+		WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT1_PAGE_TABLE_END_ADDR_LO32,
+				    i * hub->ctx_addr_distance,
+				    lower_32_bits(adev->vm_manager.max_pfn - 1));
+		WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT1_PAGE_TABLE_END_ADDR_HI32,
+				    i * hub->ctx_addr_distance,
+				    upper_32_bits(adev->vm_manager.max_pfn - 1));
+	}
+
+	hub->vm_cntx_cntl = tmp;
+}
+
+static void mmhub_v3_3_program_invalidation(struct amdgpu_device *adev)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
+	unsigned int i;
+
+	for (i = 0; i < 18; ++i) {
+		WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_INVALIDATE_ENG0_ADDR_RANGE_LO32,
+				    i * hub->eng_addr_distance, 0xffffffff);
+		WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_INVALIDATE_ENG0_ADDR_RANGE_HI32,
+				    i * hub->eng_addr_distance, 0x1f);
+	}
+}
+
+static void mmhub_v3_3_init_saw_regs(struct amdgpu_device *adev)
+{
+	uint64_t pt_base = amdgpu_gmc_pd_addr(adev->gart.bo);
+	uint32_t tmp;
+
+	/* Program page table base, gart start, gart end */
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_SAW_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32,
+			lower_32_bits(pt_base >> 12));
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_SAW_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32,
+			upper_32_bits(pt_base >> 12));
+
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_SAW_CONTEXT0_PAGE_TABLE_START_ADDR_LO32,
+		     (u32)(adev->gmc.gart_start >> 12));
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_SAW_CONTEXT0_PAGE_TABLE_START_ADDR_HI32,
+		     (u32)(adev->gmc.gart_start >> 44));
+
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_SAW_CONTEXT0_PAGE_TABLE_END_ADDR_LO32,
+		     (u32)(adev->gmc.gart_end >> 12));
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_SAW_CONTEXT0_PAGE_TABLE_END_ADDR_HI32,
+		     (u32)(adev->gmc.gart_end >> 44));
+
+	tmp = RREG32_SOC15(MMHUB, 0, regMMVM_L2_SAW_CONTEXT0_CNTL);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_SAW_CONTEXT0_CNTL, ENABLE_CONTEXT, 1);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_SAW_CONTEXT0_CNTL, PAGE_TABLE_DEPTH, 0);
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_SAW_CONTEXT0_CNTL, tmp);
+
+	/* Disable all contexts except context 0 */
+	tmp = 0xfffe;
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_SAW_CONTEXTS_DISABLE, tmp);
+
+	/* Program saw cntl4 */
+	tmp = RREG32_SOC15(MMHUB, 0, regMMVM_L2_SAW_CNTL4);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_SAW_CNTL4, VMC_TAP_CONTEXT0_PDE_REQUEST_SNOOP, 1);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_SAW_CNTL4, VMC_TAP_CONTEXT0_PTE_REQUEST_SNOOP, 1);
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_SAW_CNTL4, tmp);
+}
+
+static void mmhub_v3_3_enable_tls(struct amdgpu_device *adev)
+{
+	WREG32_SOC15(MMHUB, 0, regDAGB0_L1TLB_REG_RW_3_3, 0);
+	WREG32_SOC15(MMHUB, 0, regDAGB1_L1TLB_REG_RW_3_3, 3);
+}
+
+static int mmhub_v3_3_gart_enable(struct amdgpu_device *adev)
+{
+	/* GART Enable. */
+	mmhub_v3_3_init_gart_aperture_regs(adev);
+	mmhub_v3_3_init_system_aperture_regs(adev);
+	mmhub_v3_3_init_tlb_regs(adev);
+	mmhub_v3_3_init_cache_regs(adev);
+
+	mmhub_v3_3_enable_system_domain(adev);
+	mmhub_v3_3_disable_identity_aperture(adev);
+	mmhub_v3_3_setup_vmid_config(adev);
+	mmhub_v3_3_program_invalidation(adev);
+
+	/* standalone alone walker init */
+	mmhub_v3_3_init_saw_regs(adev);
+
+	/* enable mmhub tls */
+	mmhub_v3_3_enable_tls(adev);
+
+	return 0;
+}
+
+static void mmhub_v3_3_gart_disable(struct amdgpu_device *adev)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
+	u32 tmp;
+	u32 i;
+
+	/* Disable all tables */
+	for (i = 0; i < 16; i++)
+		WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT0_CNTL,
+				    i * hub->ctx_distance, 0);
+
+	/* Setup TLB control */
+	tmp = RREG32_SOC15(MMHUB, 0, regMMMC_VM_MX_L1_TLB_CNTL);
+	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB, 0);
+	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL,
+			    ENABLE_ADVANCED_DRIVER_MODEL, 0);
+	WREG32_SOC15(MMHUB, 0, regMMMC_VM_MX_L1_TLB_CNTL, tmp);
+
+	/* Setup L2 cache */
+	tmp = RREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL, ENABLE_L2_CACHE, 0);
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL, tmp);
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL3, 0);
+}
+
+/**
+ * mmhub_v3_3_set_fault_enable_default - update GART/VM fault handling
+ *
+ * @adev: amdgpu_device pointer
+ * @value: true redirects VM faults to the default page
+ */
+static void mmhub_v3_3_set_fault_enable_default(struct amdgpu_device *adev,
+						  bool value)
+{
+	u32 tmp;
+
+	tmp = RREG32_SOC15(MMHUB, 0, regMMVM_L2_PROTECTION_FAULT_CNTL);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    PDE0_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    PDE1_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    PDE2_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    TRANSLATE_FURTHER_PROTECTION_FAULT_ENABLE_DEFAULT,
+			    value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    NACK_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    VALID_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    READ_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    WRITE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    EXECUTE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	if (!value) {
+		tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+				CRASH_ON_NO_RETRY_FAULT, 1);
+		tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+				CRASH_ON_RETRY_FAULT, 1);
+	}
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_PROTECTION_FAULT_CNTL, tmp);
+}
+
+static const struct amdgpu_vmhub_funcs mmhub_v3_3_vmhub_funcs = {
+	.print_l2_protection_fault_status = mmhub_v3_3_print_l2_protection_fault_status,
+	.get_invalidate_req = mmhub_v3_3_get_invalidate_req,
+};
+
+static void mmhub_v3_3_init(struct amdgpu_device *adev)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
+
+	hub->ctx0_ptb_addr_lo32 =
+		SOC15_REG_OFFSET(MMHUB, 0,
+				 regMMVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32);
+	hub->ctx0_ptb_addr_hi32 =
+		SOC15_REG_OFFSET(MMHUB, 0,
+				 regMMVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32);
+	hub->vm_inv_eng0_sem =
+		SOC15_REG_OFFSET(MMHUB, 0, regMMVM_INVALIDATE_ENG0_SEM);
+	hub->vm_inv_eng0_req =
+		SOC15_REG_OFFSET(MMHUB, 0, regMMVM_INVALIDATE_ENG0_REQ);
+	hub->vm_inv_eng0_ack =
+		SOC15_REG_OFFSET(MMHUB, 0, regMMVM_INVALIDATE_ENG0_ACK);
+	hub->vm_context0_cntl =
+		SOC15_REG_OFFSET(MMHUB, 0, regMMVM_CONTEXT0_CNTL);
+	hub->vm_l2_pro_fault_status =
+		SOC15_REG_OFFSET(MMHUB, 0, regMMVM_L2_PROTECTION_FAULT_STATUS);
+	hub->vm_l2_pro_fault_cntl =
+		SOC15_REG_OFFSET(MMHUB, 0, regMMVM_L2_PROTECTION_FAULT_CNTL);
+
+	hub->ctx_distance = regMMVM_CONTEXT1_CNTL - regMMVM_CONTEXT0_CNTL;
+	hub->ctx_addr_distance = regMMVM_CONTEXT1_PAGE_TABLE_BASE_ADDR_LO32 -
+		regMMVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32;
+	hub->eng_distance = regMMVM_INVALIDATE_ENG1_REQ -
+		regMMVM_INVALIDATE_ENG0_REQ;
+	hub->eng_addr_distance = regMMVM_INVALIDATE_ENG1_ADDR_RANGE_LO32 -
+		regMMVM_INVALIDATE_ENG0_ADDR_RANGE_LO32;
+
+	hub->vm_cntx_cntl_vm_fault = MMVM_CONTEXT1_CNTL__RANGE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		MMVM_CONTEXT1_CNTL__DUMMY_PAGE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		MMVM_CONTEXT1_CNTL__PDE0_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		MMVM_CONTEXT1_CNTL__VALID_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		MMVM_CONTEXT1_CNTL__READ_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		MMVM_CONTEXT1_CNTL__WRITE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		MMVM_CONTEXT1_CNTL__EXECUTE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK;
+
+	hub->vmhub_funcs = &mmhub_v3_3_vmhub_funcs;
+}
+
+static u64 mmhub_v3_3_get_fb_location(struct amdgpu_device *adev)
+{
+	u64 base;
+
+	base = RREG32_SOC15(MMHUB, 0, regMMMC_VM_FB_LOCATION_BASE);
+	base &= MMMC_VM_FB_LOCATION_BASE__FB_BASE_MASK;
+	base <<= 24;
+
+	return base;
+}
+
+static u64 mmhub_v3_3_get_mc_fb_offset(struct amdgpu_device *adev)
+{
+	u64 offset;
+
+	offset = RREG32_SOC15(MMHUB, 0, regMMMC_VM_FB_OFFSET);
+	offset &= MMMC_VM_FB_OFFSET__FB_OFFSET_MASK;
+	offset <<= 24;
+
+	return offset;
+}
+
+static void mmhub_v3_3_update_medium_grain_clock_gating(struct amdgpu_device *adev,
+							  bool enable)
+{
+	uint32_t def, data;
+
+	def = data = RREG32_SOC15(MMHUB, 0, regMM_ATC_L2_MISC_CG);
+
+	if (enable)
+		data |= MM_ATC_L2_MISC_CG__ENABLE_MASK;
+	else
+		data &= ~MM_ATC_L2_MISC_CG__ENABLE_MASK;
+
+	if (def != data)
+		WREG32_SOC15(MMHUB, 0, regMM_ATC_L2_MISC_CG, data);
+}
+
+static void mmhub_v3_3_update_medium_grain_light_sleep(struct amdgpu_device *adev,
+							 bool enable)
+{
+	uint32_t def, data;
+
+	def = data = RREG32_SOC15(MMHUB, 0, regMM_ATC_L2_MISC_CG);
+
+	if (enable)
+		data |= MM_ATC_L2_MISC_CG__MEM_LS_ENABLE_MASK;
+	else
+		data &= ~MM_ATC_L2_MISC_CG__MEM_LS_ENABLE_MASK;
+
+	if (def != data)
+		WREG32_SOC15(MMHUB, 0, regMM_ATC_L2_MISC_CG, data);
+}
+
+static int mmhub_v3_3_set_clockgating(struct amdgpu_device *adev,
+					enum amd_clockgating_state state)
+{
+	if (amdgpu_sriov_vf(adev))
+		return 0;
+
+	mmhub_v3_3_update_medium_grain_clock_gating(adev,
+			state == AMD_CG_STATE_GATE);
+	mmhub_v3_3_update_medium_grain_light_sleep(adev,
+			state == AMD_CG_STATE_GATE);
+	return 0;
+}
+
+static void mmhub_v3_3_get_clockgating(struct amdgpu_device *adev, u64 *flags)
+{
+	u32 data;
+
+	if (amdgpu_sriov_vf(adev))
+		*flags = 0;
+
+	data = RREG32_SOC15(MMHUB, 0, regMM_ATC_L2_MISC_CG);
+
+	/* AMD_CG_SUPPORT_MC_MGCG */
+	if (data & MM_ATC_L2_MISC_CG__ENABLE_MASK)
+		*flags |= AMD_CG_SUPPORT_MC_MGCG;
+
+	/* AMD_CG_SUPPORT_MC_LS */
+	if (data & MM_ATC_L2_MISC_CG__MEM_LS_ENABLE_MASK)
+		*flags |= AMD_CG_SUPPORT_MC_LS;
+}
+
+const struct amdgpu_mmhub_funcs mmhub_v3_3_funcs = {
+	.init = mmhub_v3_3_init,
+	.get_fb_location = mmhub_v3_3_get_fb_location,
+	.get_mc_fb_offset = mmhub_v3_3_get_mc_fb_offset,
+	.gart_enable = mmhub_v3_3_gart_enable,
+	.set_fault_enable_default = mmhub_v3_3_set_fault_enable_default,
+	.gart_disable = mmhub_v3_3_gart_disable,
+	.set_clockgating = mmhub_v3_3_set_clockgating,
+	.get_clockgating = mmhub_v3_3_get_clockgating,
+	.setup_vm_pt_regs = mmhub_v3_3_setup_vm_pt_regs,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v3_3.h b/drivers/gpu/drm/amd/amdgpu/mmhub_v3_3.h
new file mode 100644
index 000000000000..37b62c7e5a4a
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v3_3.h
@@ -0,0 +1,29 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __MMHUB_V3_3_H__
+#define __MMHUB_V3_3_H__
+
+extern const struct amdgpu_mmhub_funcs mmhub_v3_3_funcs;
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v4_1_0.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v4_1_0.c
new file mode 100644
index 000000000000..951998454b25
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v4_1_0.c
@@ -0,0 +1,647 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "amdgpu.h"
+#include "mmhub_v4_1_0.h"
+
+#include "mmhub/mmhub_4_1_0_offset.h"
+#include "mmhub/mmhub_4_1_0_sh_mask.h"
+
+#include "soc15_common.h"
+#include "soc24_enum.h"
+
+#define regMMVM_L2_CNTL3_DEFAULT				0x80100007
+#define regMMVM_L2_CNTL4_DEFAULT				0x000000c1
+#define regMMVM_L2_CNTL5_DEFAULT				0x00003fe0
+
+static const char *mmhub_client_ids_v4_1_0[][2] = {
+	[0][0] = "VMC",
+	[4][0] = "DCEDMC",
+	[6][0] = "MP0",
+	[7][0] = "MP1",
+	[8][0] = "MPIO",
+	[16][0] = "LSDMA",
+	[17][0] = "JPEG",
+	[19][0] = "VCNU",
+	[22][0] = "VSCH",
+	[23][0] = "HDP",
+	[32+23][0] = "VCNRD",
+	[3][1] = "DCEDWB",
+	[4][1] = "DCEDMC",
+	[6][1] = "MP0",
+	[7][1] = "MP1",
+	[8][1] = "MPIO",
+	[10][1] = "DBGU0",
+	[11][1] = "DBGU1",
+	[12][1] = "DBGUNBIO",
+	[14][1] = "XDP",
+	[15][1] = "OSSSYS",
+	[16][1] = "LSDMA",
+	[17][1] = "JPEG",
+	[18][1] = "VCNWR",
+	[19][1] = "VCNU",
+	[22][1] = "VSCH",
+	[23][1] = "HDP",
+};
+
+static uint32_t mmhub_v4_1_0_get_invalidate_req(unsigned int vmid,
+						uint32_t flush_type)
+{
+	u32 req = 0;
+
+	/* invalidate using legacy mode on vmid*/
+	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ,
+			    PER_VMID_INVALIDATE_REQ, 1 << vmid);
+	/* Only use legacy inv on mmhub side */
+	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ, FLUSH_TYPE, 0);
+	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PTES, 1);
+	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PDE0, 1);
+	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PDE1, 1);
+	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PDE2, 1);
+	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ, INVALIDATE_L1_PTES, 1);
+	req = REG_SET_FIELD(req, MMVM_INVALIDATE_ENG0_REQ,
+			    CLEAR_PROTECTION_FAULT_STATUS_ADDR,	0);
+
+	return req;
+}
+
+static void
+mmhub_v4_1_0_print_l2_protection_fault_status(struct amdgpu_device *adev,
+					      uint32_t status)
+{
+	uint32_t cid, rw;
+	const char *mmhub_cid = NULL;
+
+	cid = REG_GET_FIELD(status,
+			    MMVM_L2_PROTECTION_FAULT_STATUS_LO32, CID);
+	rw = REG_GET_FIELD(status,
+			   MMVM_L2_PROTECTION_FAULT_STATUS_LO32, RW);
+
+	dev_err(adev->dev,
+		"MMVM_L2_PROTECTION_FAULT_STATUS_LO32:0x%08X\n",
+		status);
+	switch (amdgpu_ip_version(adev, MMHUB_HWIP, 0)) {
+	case IP_VERSION(4, 1, 0):
+		mmhub_cid = mmhub_client_ids_v4_1_0[cid][rw];
+		break;
+	default:
+		mmhub_cid = NULL;
+		break;
+	}
+	dev_err(adev->dev, "\t Faulty UTCL2 client ID: %s (0x%x)\n",
+		mmhub_cid ? mmhub_cid : "unknown", cid);
+	dev_err(adev->dev, "\t MORE_FAULTS: 0x%lx\n",
+		REG_GET_FIELD(status,
+		MMVM_L2_PROTECTION_FAULT_STATUS_LO32, MORE_FAULTS));
+	dev_err(adev->dev, "\t WALKER_ERROR: 0x%lx\n",
+		REG_GET_FIELD(status,
+		MMVM_L2_PROTECTION_FAULT_STATUS_LO32, WALKER_ERROR));
+	dev_err(adev->dev, "\t PERMISSION_FAULTS: 0x%lx\n",
+		REG_GET_FIELD(status,
+		MMVM_L2_PROTECTION_FAULT_STATUS_LO32, PERMISSION_FAULTS));
+	dev_err(adev->dev, "\t MAPPING_ERROR: 0x%lx\n",
+		REG_GET_FIELD(status,
+		MMVM_L2_PROTECTION_FAULT_STATUS_LO32, MAPPING_ERROR));
+	dev_err(adev->dev, "\t RW: 0x%x\n", rw);
+}
+
+static void mmhub_v4_1_0_setup_vm_pt_regs(struct amdgpu_device *adev,
+					  uint32_t vmid, uint64_t page_table_base)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
+
+	WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32,
+			    hub->ctx_addr_distance * vmid,
+			    lower_32_bits(page_table_base));
+
+	WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32,
+			    hub->ctx_addr_distance * vmid,
+			    upper_32_bits(page_table_base));
+}
+
+static void mmhub_v4_1_0_init_gart_aperture_regs(struct amdgpu_device *adev)
+{
+	uint64_t pt_base = amdgpu_gmc_pd_addr(adev->gart.bo);
+
+	mmhub_v4_1_0_setup_vm_pt_regs(adev, 0, pt_base);
+
+	WREG32_SOC15(MMHUB, 0, regMMVM_CONTEXT0_PAGE_TABLE_START_ADDR_LO32,
+		     (u32)(adev->gmc.gart_start >> 12));
+	WREG32_SOC15(MMHUB, 0, regMMVM_CONTEXT0_PAGE_TABLE_START_ADDR_HI32,
+		     (u32)(adev->gmc.gart_start >> 44));
+
+	WREG32_SOC15(MMHUB, 0, regMMVM_CONTEXT0_PAGE_TABLE_END_ADDR_LO32,
+		     (u32)(adev->gmc.gart_end >> 12));
+	WREG32_SOC15(MMHUB, 0, regMMVM_CONTEXT0_PAGE_TABLE_END_ADDR_HI32,
+		     (u32)(adev->gmc.gart_end >> 44));
+}
+
+static void mmhub_v4_1_0_init_system_aperture_regs(struct amdgpu_device *adev)
+{
+	uint64_t value;
+	uint32_t tmp;
+
+	/*
+	 * the new L1 policy will block SRIOV guest from writing
+	 * these regs, and they will be programed at host.
+	 * so skip programing these regs.
+	 */
+	if (amdgpu_sriov_vf(adev))
+		return;
+
+	/* Program the AGP BAR */
+	WREG32_SOC15(MMHUB, 0, regMMMC_VM_AGP_BASE, 0);
+	WREG32_SOC15(MMHUB, 0, regMMMC_VM_AGP_BOT, adev->gmc.agp_start >> 24);
+	WREG32_SOC15(MMHUB, 0, regMMMC_VM_AGP_TOP, adev->gmc.agp_end >> 24);
+
+	/* Program the system aperture low logical page number. */
+	WREG32_SOC15(MMHUB, 0, regMMMC_VM_SYSTEM_APERTURE_LOW_ADDR,
+		     min(adev->gmc.fb_start, adev->gmc.agp_start) >> 18);
+	WREG32_SOC15(MMHUB, 0, regMMMC_VM_SYSTEM_APERTURE_HIGH_ADDR,
+		     max(adev->gmc.fb_end, adev->gmc.agp_end) >> 18);
+
+	/* Set default page address. */
+	value = adev->mem_scratch.gpu_addr - adev->gmc.vram_start +
+		adev->vm_manager.vram_base_offset;
+	WREG32_SOC15(MMHUB, 0, regMMMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB,
+		     (u32)(value >> 12));
+	WREG32_SOC15(MMHUB, 0, regMMMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_MSB,
+		     (u32)(value >> 44));
+
+	/* Program "protection fault". */
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_PROTECTION_FAULT_DEFAULT_ADDR_LO32,
+		     (u32)(adev->dummy_page_addr >> 12));
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_PROTECTION_FAULT_DEFAULT_ADDR_HI32,
+		     (u32)((u64)adev->dummy_page_addr >> 44));
+
+	tmp = RREG32_SOC15(MMHUB, 0, regMMVM_L2_PROTECTION_FAULT_CNTL2);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL2,
+			    ACTIVE_PAGE_MIGRATION_PTE_READ_RETRY, 1);
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_PROTECTION_FAULT_CNTL2, tmp);
+}
+
+static void mmhub_v4_1_0_init_tlb_regs(struct amdgpu_device *adev)
+{
+	uint32_t tmp;
+
+	/* Setup TLB control */
+	tmp = RREG32_SOC15(MMHUB, 0, regMMMC_VM_MX_L1_TLB_CNTL);
+
+	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB, 1);
+	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL, SYSTEM_ACCESS_MODE, 3);
+	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL,
+			    ENABLE_ADVANCED_DRIVER_MODEL, 1);
+	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL,
+			    SYSTEM_APERTURE_UNMAPPED_ACCESS, 0);
+	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL, ECO_BITS, 0);
+	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL,
+			    MTYPE, MTYPE_UC); /* UC, uncached */
+
+	WREG32_SOC15(MMHUB, 0, regMMMC_VM_MX_L1_TLB_CNTL, tmp);
+}
+
+static void mmhub_v4_1_0_init_cache_regs(struct amdgpu_device *adev)
+{
+	uint32_t tmp;
+
+	/* These registers are not accessible to VF-SRIOV.
+	 * The PF will program them instead.
+	 */
+	if (amdgpu_sriov_vf(adev))
+		return;
+
+	/* Setup L2 cache */
+	tmp = RREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL, ENABLE_L2_CACHE, 1);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL, ENABLE_L2_FRAGMENT_PROCESSING, 0);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL,
+			    ENABLE_DEFAULT_PAGE_OUT_TO_SYSTEM_MEMORY, 1);
+	/* XXX for emulation, Refer to closed source code.*/
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL, L2_PDE0_CACHE_TAG_GENERATION_MODE,
+			    0);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL, PDE_FAULT_CLASSIFICATION, 0);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL, CONTEXT1_IDENTITY_ACCESS_MODE, 1);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL, IDENTITY_MODE_FRAGMENT_SIZE, 0);
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL, tmp);
+
+	tmp = RREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL2);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL2, INVALIDATE_ALL_L1_TLBS, 1);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL2, INVALIDATE_L2_CACHE, 1);
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL2, tmp);
+
+	tmp = regMMVM_L2_CNTL3_DEFAULT;
+	if (adev->gmc.translate_further) {
+		tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL3, BANK_SELECT, 12);
+		tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL3,
+				    L2_CACHE_BIGK_FRAGMENT_SIZE, 9);
+	} else {
+		tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL3, BANK_SELECT, 9);
+		tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL3,
+				    L2_CACHE_BIGK_FRAGMENT_SIZE, 6);
+	}
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL3, tmp);
+
+	tmp = regMMVM_L2_CNTL4_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL4, VMC_TAP_PDE_REQUEST_PHYSICAL, 0);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL4, VMC_TAP_PTE_REQUEST_PHYSICAL, 0);
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL4, tmp);
+
+	tmp = regMMVM_L2_CNTL5_DEFAULT;
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL5, L2_CACHE_SMALLK_FRAGMENT_SIZE, 0);
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL5, tmp);
+}
+
+static void mmhub_v4_1_0_enable_system_domain(struct amdgpu_device *adev)
+{
+	uint32_t tmp;
+
+	tmp = RREG32_SOC15(MMHUB, 0, regMMVM_CONTEXT0_CNTL);
+	tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT0_CNTL, ENABLE_CONTEXT, 1);
+	tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT0_CNTL, PAGE_TABLE_DEPTH, 0);
+	tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT0_CNTL,
+			    RETRY_PERMISSION_OR_INVALID_PAGE_FAULT, 0);
+	WREG32_SOC15(MMHUB, 0, regMMVM_CONTEXT0_CNTL, tmp);
+}
+
+static void mmhub_v4_1_0_disable_identity_aperture(struct amdgpu_device *adev)
+{
+	/* These registers are not accessible to VF-SRIOV.
+	 * The PF will program them instead.
+	 */
+	if (amdgpu_sriov_vf(adev))
+		return;
+
+	WREG32_SOC15(MMHUB, 0,
+		     regMMVM_L2_CONTEXT1_IDENTITY_APERTURE_LOW_ADDR_LO32,
+		     0xFFFFFFFF);
+	WREG32_SOC15(MMHUB, 0,
+		     regMMVM_L2_CONTEXT1_IDENTITY_APERTURE_LOW_ADDR_HI32,
+		     0x0000000F);
+
+	WREG32_SOC15(MMHUB, 0,
+		     regMMVM_L2_CONTEXT1_IDENTITY_APERTURE_HIGH_ADDR_LO32, 0);
+	WREG32_SOC15(MMHUB, 0,
+		     regMMVM_L2_CONTEXT1_IDENTITY_APERTURE_HIGH_ADDR_HI32, 0);
+
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CONTEXT_IDENTITY_PHYSICAL_OFFSET_LO32,
+		     0);
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CONTEXT_IDENTITY_PHYSICAL_OFFSET_HI32,
+		     0);
+}
+
+static void mmhub_v4_1_0_setup_vmid_config(struct amdgpu_device *adev)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
+	int i;
+	uint32_t tmp;
+
+	for (i = 0; i <= 14; i++) {
+		tmp = RREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT1_CNTL, i);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL, ENABLE_CONTEXT, 1);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL, PAGE_TABLE_DEPTH,
+				    adev->vm_manager.num_level);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT,
+				    1);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    PDE0_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    VALID_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    READ_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    WRITE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    EXECUTE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    PAGE_TABLE_BLOCK_SIZE,
+				    adev->vm_manager.block_size - 9);
+		/* Send no-retry XNACK on fault to suppress VM fault storm. */
+		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
+				    RETRY_PERMISSION_OR_INVALID_PAGE_FAULT,
+				    !amdgpu_noretry);
+		WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT1_CNTL,
+				    i * hub->ctx_distance, tmp);
+		WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT1_PAGE_TABLE_START_ADDR_LO32,
+				    i * hub->ctx_addr_distance, 0);
+		WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT1_PAGE_TABLE_START_ADDR_HI32,
+				    i * hub->ctx_addr_distance, 0);
+		WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT1_PAGE_TABLE_END_ADDR_LO32,
+				    i * hub->ctx_addr_distance,
+				    lower_32_bits(adev->vm_manager.max_pfn - 1));
+		WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT1_PAGE_TABLE_END_ADDR_HI32,
+				    i * hub->ctx_addr_distance,
+				    upper_32_bits(adev->vm_manager.max_pfn - 1));
+	}
+
+	hub->vm_cntx_cntl = tmp;
+}
+
+static void mmhub_v4_1_0_program_invalidation(struct amdgpu_device *adev)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
+	unsigned i;
+
+	for (i = 0; i < 18; ++i) {
+		WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_INVALIDATE_ENG0_ADDR_RANGE_LO32,
+				    i * hub->eng_addr_distance, 0xffffffff);
+		WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_INVALIDATE_ENG0_ADDR_RANGE_HI32,
+				    i * hub->eng_addr_distance, 0x1f);
+	}
+}
+
+static int mmhub_v4_1_0_gart_enable(struct amdgpu_device *adev)
+{
+	/* GART Enable. */
+	mmhub_v4_1_0_init_gart_aperture_regs(adev);
+	mmhub_v4_1_0_init_system_aperture_regs(adev);
+	mmhub_v4_1_0_init_tlb_regs(adev);
+	mmhub_v4_1_0_init_cache_regs(adev);
+
+	mmhub_v4_1_0_enable_system_domain(adev);
+	mmhub_v4_1_0_disable_identity_aperture(adev);
+	mmhub_v4_1_0_setup_vmid_config(adev);
+	mmhub_v4_1_0_program_invalidation(adev);
+
+	return 0;
+}
+
+static void mmhub_v4_1_0_gart_disable(struct amdgpu_device *adev)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
+	u32 tmp;
+	u32 i;
+
+	/* Disable all tables */
+	for (i = 0; i < 16; i++)
+		WREG32_SOC15_OFFSET(MMHUB, 0, regMMVM_CONTEXT0_CNTL,
+				    i * hub->ctx_distance, 0);
+
+	/* Setup TLB control */
+	tmp = RREG32_SOC15(MMHUB, 0, regMMMC_VM_MX_L1_TLB_CNTL);
+	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB, 0);
+	tmp = REG_SET_FIELD(tmp, MMMC_VM_MX_L1_TLB_CNTL,
+			    ENABLE_ADVANCED_DRIVER_MODEL, 0);
+	WREG32_SOC15(MMHUB, 0, regMMMC_VM_MX_L1_TLB_CNTL, tmp);
+
+	/* Setup L2 cache */
+	tmp = RREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_CNTL, ENABLE_L2_CACHE, 0);
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL, tmp);
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_CNTL3, 0);
+}
+
+/**
+ * mmhub_v4_1_0_set_fault_enable_default - update GART/VM fault handling
+ *
+ * @adev: amdgpu_device pointer
+ * @value: true redirects VM faults to the default page
+ */
+static void
+mmhub_v4_1_0_set_fault_enable_default(struct amdgpu_device *adev, bool value)
+{
+	u32 tmp;
+
+	/* These registers are not accessible to VF-SRIOV.
+	 * The PF will program them instead.
+	 */
+	if (amdgpu_sriov_vf(adev))
+		return;
+
+	tmp = RREG32_SOC15(MMHUB, 0, regMMVM_L2_PROTECTION_FAULT_CNTL);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    PDE0_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    PDE1_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    PDE2_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    TRANSLATE_FURTHER_PROTECTION_FAULT_ENABLE_DEFAULT,
+			    value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    NACK_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    VALID_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    READ_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    WRITE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+			    EXECUTE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
+	if (!value) {
+		tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+				CRASH_ON_NO_RETRY_FAULT, 1);
+		tmp = REG_SET_FIELD(tmp, MMVM_L2_PROTECTION_FAULT_CNTL,
+				CRASH_ON_RETRY_FAULT, 1);
+	}
+	WREG32_SOC15(MMHUB, 0, regMMVM_L2_PROTECTION_FAULT_CNTL, tmp);
+}
+
+static const struct amdgpu_vmhub_funcs mmhub_v4_1_0_vmhub_funcs = {
+	.print_l2_protection_fault_status = mmhub_v4_1_0_print_l2_protection_fault_status,
+	.get_invalidate_req = mmhub_v4_1_0_get_invalidate_req,
+};
+
+static void mmhub_v4_1_0_init(struct amdgpu_device *adev)
+{
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
+
+	hub->ctx0_ptb_addr_lo32 =
+		SOC15_REG_OFFSET(MMHUB, 0,
+				 regMMVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32);
+	hub->ctx0_ptb_addr_hi32 =
+		SOC15_REG_OFFSET(MMHUB, 0,
+				 regMMVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32);
+	hub->vm_inv_eng0_sem =
+		SOC15_REG_OFFSET(MMHUB, 0, regMMVM_INVALIDATE_ENG0_SEM);
+	hub->vm_inv_eng0_req =
+		SOC15_REG_OFFSET(MMHUB, 0, regMMVM_INVALIDATE_ENG0_REQ);
+	hub->vm_inv_eng0_ack =
+		SOC15_REG_OFFSET(MMHUB, 0, regMMVM_INVALIDATE_ENG0_ACK);
+	hub->vm_context0_cntl =
+		SOC15_REG_OFFSET(MMHUB, 0, regMMVM_CONTEXT0_CNTL);
+	hub->vm_l2_pro_fault_status =
+		SOC15_REG_OFFSET(MMHUB, 0, regMMVM_L2_PROTECTION_FAULT_STATUS_LO32);
+	hub->vm_l2_pro_fault_cntl =
+		SOC15_REG_OFFSET(MMHUB, 0, regMMVM_L2_PROTECTION_FAULT_CNTL);
+
+	hub->ctx_distance = regMMVM_CONTEXT1_CNTL - regMMVM_CONTEXT0_CNTL;
+	hub->ctx_addr_distance = regMMVM_CONTEXT1_PAGE_TABLE_BASE_ADDR_LO32 -
+		regMMVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32;
+	hub->eng_distance = regMMVM_INVALIDATE_ENG1_REQ -
+		regMMVM_INVALIDATE_ENG0_REQ;
+	hub->eng_addr_distance = regMMVM_INVALIDATE_ENG1_ADDR_RANGE_LO32 -
+		regMMVM_INVALIDATE_ENG0_ADDR_RANGE_LO32;
+
+	hub->vm_cntx_cntl_vm_fault = MMVM_CONTEXT1_CNTL__RANGE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		MMVM_CONTEXT1_CNTL__DUMMY_PAGE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		MMVM_CONTEXT1_CNTL__PDE0_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		MMVM_CONTEXT1_CNTL__VALID_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		MMVM_CONTEXT1_CNTL__READ_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		MMVM_CONTEXT1_CNTL__WRITE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+		MMVM_CONTEXT1_CNTL__EXECUTE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK;
+
+	hub->vm_l2_bank_select_reserved_cid2 =
+		SOC15_REG_OFFSET(MMHUB, 0, regMMVM_L2_BANK_SELECT_RESERVED_CID2);
+
+	hub->vm_contexts_disable =
+		SOC15_REG_OFFSET(MMHUB, 0, regMMVM_CONTEXTS_DISABLE);
+
+	hub->vmhub_funcs = &mmhub_v4_1_0_vmhub_funcs;
+}
+
+static u64 mmhub_v4_1_0_get_fb_location(struct amdgpu_device *adev)
+{
+	u64 base;
+
+	base = RREG32_SOC15(MMHUB, 0, regMMMC_VM_FB_LOCATION_BASE);
+
+	base &= MMMC_VM_FB_LOCATION_BASE__FB_BASE_MASK;
+	base <<= 24;
+
+	return base;
+}
+
+static u64 mmhub_v4_1_0_get_mc_fb_offset(struct amdgpu_device *adev)
+{
+	return (u64)RREG32_SOC15(MMHUB, 0, regMMMC_VM_FB_OFFSET) << 24;
+}
+
+static void
+mmhub_v4_1_0_update_medium_grain_clock_gating(struct amdgpu_device *adev,
+					      bool enable)
+{
+#if 0
+	uint32_t def, data;
+#endif
+	uint32_t def1, data1, def2 = 0, data2 = 0;
+#if 0
+	def  = data  = RREG32_SOC15(MMHUB, 0, regMM_ATC_L2_MISC_CG);
+#endif
+	def1 = data1 = RREG32_SOC15(MMHUB, 0, regDAGB0_CNTL_MISC2);
+	def2 = data2 = RREG32_SOC15(MMHUB, 0, regDAGB1_CNTL_MISC2);
+
+	if (enable) {
+#if 0
+		data |= MM_ATC_L2_MISC_CG__ENABLE_MASK;
+#endif
+		data1 &= ~(DAGB0_CNTL_MISC2__DISABLE_RDRET_TAP_CHAIN_FGCG_MASK |
+			   DAGB0_CNTL_MISC2__DISABLE_WRRET_TAP_CHAIN_FGCG_MASK);
+
+		data2 &= ~(DAGB1_CNTL_MISC2__DISABLE_RDRET_TAP_CHAIN_FGCG_MASK |
+			   DAGB1_CNTL_MISC2__DISABLE_WRRET_TAP_CHAIN_FGCG_MASK);
+	} else {
+#if 0
+		data &= ~MM_ATC_L2_MISC_CG__ENABLE_MASK;
+#endif
+		data1 |= (DAGB0_CNTL_MISC2__DISABLE_RDRET_TAP_CHAIN_FGCG_MASK |
+			  DAGB0_CNTL_MISC2__DISABLE_WRRET_TAP_CHAIN_FGCG_MASK);
+
+		data2 |= (DAGB1_CNTL_MISC2__DISABLE_RDRET_TAP_CHAIN_FGCG_MASK |
+			  DAGB1_CNTL_MISC2__DISABLE_WRRET_TAP_CHAIN_FGCG_MASK);
+	}
+
+#if 0
+	if (def != data)
+		WREG32_SOC15(MMHUB, 0, regMM_ATC_L2_MISC_CG, data);
+#endif
+	if (def1 != data1)
+		WREG32_SOC15(MMHUB, 0, regDAGB0_CNTL_MISC2, data1);
+
+	if (def2 != data2)
+		WREG32_SOC15(MMHUB, 0, regDAGB1_CNTL_MISC2, data2);
+}
+
+static void
+mmhub_v4_1_0_update_medium_grain_light_sleep(struct amdgpu_device *adev,
+					     bool enable)
+{
+#if 0
+	uint32_t def, data;
+
+	def = data = RREG32_SOC15(MMHUB, 0, regMM_ATC_L2_MISC_CG);
+
+	if (enable)
+		data |= MM_ATC_L2_MISC_CG__MEM_LS_ENABLE_MASK;
+	else
+		data &= ~MM_ATC_L2_MISC_CG__MEM_LS_ENABLE_MASK;
+
+	if (def != data)
+		WREG32_SOC15(MMHUB, 0, regMM_ATC_L2_MISC_CG, data);
+#endif
+}
+
+static int mmhub_v4_1_0_set_clockgating(struct amdgpu_device *adev,
+					enum amd_clockgating_state state)
+{
+	if (amdgpu_sriov_vf(adev))
+		return 0;
+
+	if (adev->cg_flags & AMD_CG_SUPPORT_MC_MGCG)
+		mmhub_v4_1_0_update_medium_grain_clock_gating(adev,
+				state == AMD_CG_STATE_GATE);
+
+	if (adev->cg_flags & AMD_CG_SUPPORT_MC_LS)
+		mmhub_v4_1_0_update_medium_grain_light_sleep(adev,
+				state == AMD_CG_STATE_GATE);
+
+	return 0;
+}
+
+static void mmhub_v4_1_0_get_clockgating(struct amdgpu_device *adev, u64 *flags)
+{
+#if 0
+	int data;
+
+	if (amdgpu_sriov_vf(adev))
+		*flags = 0;
+
+	data = RREG32_SOC15(MMHUB, 0, regMM_ATC_L2_MISC_CG);
+
+	/* AMD_CG_SUPPORT_MC_MGCG */
+	if (data & MM_ATC_L2_MISC_CG__ENABLE_MASK)
+		*flags |= AMD_CG_SUPPORT_MC_MGCG;
+
+	/* AMD_CG_SUPPORT_MC_LS */
+	if (data & MM_ATC_L2_MISC_CG__MEM_LS_ENABLE_MASK)
+		*flags |= AMD_CG_SUPPORT_MC_LS;
+#endif
+}
+
+const struct amdgpu_mmhub_funcs mmhub_v4_1_0_funcs = {
+	.init = mmhub_v4_1_0_init,
+	.get_fb_location = mmhub_v4_1_0_get_fb_location,
+	.get_mc_fb_offset = mmhub_v4_1_0_get_mc_fb_offset,
+	.gart_enable = mmhub_v4_1_0_gart_enable,
+	.set_fault_enable_default = mmhub_v4_1_0_set_fault_enable_default,
+	.gart_disable = mmhub_v4_1_0_gart_disable,
+	.set_clockgating = mmhub_v4_1_0_set_clockgating,
+	.get_clockgating = mmhub_v4_1_0_get_clockgating,
+	.setup_vm_pt_regs = mmhub_v4_1_0_setup_vm_pt_regs,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v4_1_0.h b/drivers/gpu/drm/amd/amdgpu/mmhub_v4_1_0.h
new file mode 100644
index 000000000000..3902d653353c
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v4_1_0.h
@@ -0,0 +1,28 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __MMHUB_V4_1_0_H__
+#define __MMHUB_V4_1_0_H__
+
+extern const struct amdgpu_mmhub_funcs mmhub_v4_1_0_funcs;
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c
index ff49eeaf7882..fe0710b55c3a 100644
--- a/drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c
+++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c
@@ -57,7 +57,7 @@ static u64 mmhub_v9_4_get_fb_location(struct amdgpu_device *adev)
 static void mmhub_v9_4_setup_hubid_vm_pt_regs(struct amdgpu_device *adev, int hubid,
 				uint32_t vmid, uint64_t value)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0];
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
 
 	WREG32_SOC15_OFFSET(MMHUB, 0,
 			    mmVML2VC0_VM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32,
@@ -108,7 +108,7 @@ static void mmhub_v9_4_setup_vm_pt_regs(struct amdgpu_device *adev, uint32_t vmi
 }
 
 static void mmhub_v9_4_init_system_aperture_regs(struct amdgpu_device *adev,
-					         int hubid)
+						int hubid)
 {
 	uint64_t value;
 	uint32_t tmp;
@@ -136,7 +136,7 @@ static void mmhub_v9_4_init_system_aperture_regs(struct amdgpu_device *adev,
 			max(adev->gmc.fb_end, adev->gmc.agp_end) >> 18);
 
 		/* Set default page address. */
-		value = amdgpu_gmc_vram_mc2pa(adev, adev->vram_scratch.gpu_addr);
+		value = amdgpu_gmc_vram_mc2pa(adev, adev->mem_scratch.gpu_addr);
 		WREG32_SOC15_OFFSET(
 			MMHUB, 0,
 			mmVMSHAREDPF0_MC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB,
@@ -198,6 +198,36 @@ static void mmhub_v9_4_init_tlb_regs(struct amdgpu_device *adev, int hubid)
 			    hubid * MMHUB_INSTANCE_REGISTER_OFFSET, tmp);
 }
 
+/* Set snoop bit for SDMA so that SDMA writes probe-invalidates RW lines */
+static void mmhub_v9_4_init_snoop_override_regs(struct amdgpu_device *adev, int hubid)
+{
+	uint32_t tmp;
+	int i;
+	uint32_t distance = mmDAGB1_WRCLI_GPU_SNOOP_OVERRIDE -
+			    mmDAGB0_WRCLI_GPU_SNOOP_OVERRIDE;
+	uint32_t huboffset = hubid * MMHUB_INSTANCE_REGISTER_OFFSET;
+
+	for (i = 0; i < 5 - (2 * hubid); i++) {
+		/* DAGB instances 0 to 4 are in hub0 and 5 to 7 are in hub1 */
+		tmp = RREG32_SOC15_OFFSET(MMHUB, 0,
+			mmDAGB0_WRCLI_GPU_SNOOP_OVERRIDE,
+			huboffset + i * distance);
+		tmp |= (1 << 15); /* SDMA client is BIT15 */
+		WREG32_SOC15_OFFSET(MMHUB, 0,
+			mmDAGB0_WRCLI_GPU_SNOOP_OVERRIDE,
+			huboffset + i * distance, tmp);
+
+		tmp = RREG32_SOC15_OFFSET(MMHUB, 0,
+			mmDAGB0_WRCLI_GPU_SNOOP_OVERRIDE_VALUE,
+			huboffset + i * distance);
+		tmp |= (1 << 15);
+		WREG32_SOC15_OFFSET(MMHUB, 0,
+			mmDAGB0_WRCLI_GPU_SNOOP_OVERRIDE_VALUE,
+			huboffset + i * distance, tmp);
+	}
+
+}
+
 static void mmhub_v9_4_init_cache_regs(struct amdgpu_device *adev, int hubid)
 {
 	uint32_t tmp;
@@ -294,18 +324,26 @@ static void mmhub_v9_4_disable_identity_aperture(struct amdgpu_device *adev,
 
 static void mmhub_v9_4_setup_vmid_config(struct amdgpu_device *adev, int hubid)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0];
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
+	unsigned int num_level, block_size;
 	uint32_t tmp;
 	int i;
 
+	num_level = adev->vm_manager.num_level;
+	block_size = adev->vm_manager.block_size;
+	if (adev->gmc.translate_further)
+		num_level -= 1;
+	else
+		block_size -= 9;
+
 	for (i = 0; i <= 14; i++) {
 		tmp = RREG32_SOC15_OFFSET(MMHUB, 0, mmVML2VC0_VM_CONTEXT1_CNTL,
-				hubid * MMHUB_INSTANCE_REGISTER_OFFSET + i);
+				hubid * MMHUB_INSTANCE_REGISTER_OFFSET + i * hub->ctx_distance);
 		tmp = REG_SET_FIELD(tmp, VML2VC0_VM_CONTEXT1_CNTL,
 				    ENABLE_CONTEXT, 1);
 		tmp = REG_SET_FIELD(tmp, VML2VC0_VM_CONTEXT1_CNTL,
 				    PAGE_TABLE_DEPTH,
-				    adev->vm_manager.num_level);
+				    num_level);
 		tmp = REG_SET_FIELD(tmp, VML2VC0_VM_CONTEXT1_CNTL,
 				    RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
 		tmp = REG_SET_FIELD(tmp, VML2VC0_VM_CONTEXT1_CNTL,
@@ -323,7 +361,7 @@ static void mmhub_v9_4_setup_vmid_config(struct amdgpu_device *adev, int hubid)
 				    EXECUTE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
 		tmp = REG_SET_FIELD(tmp, VML2VC0_VM_CONTEXT1_CNTL,
 				    PAGE_TABLE_BLOCK_SIZE,
-				    adev->vm_manager.block_size - 9);
+				    block_size);
 		/* Send no-retry XNACK on fault to suppress VM fault storm. */
 		tmp = REG_SET_FIELD(tmp, VML2VC0_VM_CONTEXT1_CNTL,
 				    RETRY_PERMISSION_OR_INVALID_PAGE_FAULT,
@@ -355,7 +393,7 @@ static void mmhub_v9_4_setup_vmid_config(struct amdgpu_device *adev, int hubid)
 static void mmhub_v9_4_program_invalidation(struct amdgpu_device *adev,
 					    int hubid)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0];
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
 	unsigned i;
 
 	for (i = 0; i < 18; ++i) {
@@ -384,6 +422,7 @@ static int mmhub_v9_4_gart_enable(struct amdgpu_device *adev)
 		if (!amdgpu_sriov_vf(adev))
 			mmhub_v9_4_init_cache_regs(adev, i);
 
+		mmhub_v9_4_init_snoop_override_regs(adev, i);
 		mmhub_v9_4_enable_system_domain(adev, i);
 		if (!amdgpu_sriov_vf(adev))
 			mmhub_v9_4_disable_identity_aperture(adev, i);
@@ -396,7 +435,7 @@ static int mmhub_v9_4_gart_enable(struct amdgpu_device *adev)
 
 static void mmhub_v9_4_gart_disable(struct amdgpu_device *adev)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0];
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB0(0)];
 	u32 tmp;
 	u32 i, j;
 
@@ -499,8 +538,8 @@ static void mmhub_v9_4_set_fault_enable_default(struct amdgpu_device *adev, bool
 
 static void mmhub_v9_4_init(struct amdgpu_device *adev)
 {
-	struct amdgpu_vmhub *hub[MMHUB_NUM_INSTANCES] =
-		{&adev->vmhub[AMDGPU_MMHUB_0], &adev->vmhub[AMDGPU_MMHUB_1]};
+	struct amdgpu_vmhub *hub[MMHUB_NUM_INSTANCES] = {
+		&adev->vmhub[AMDGPU_MMHUB0(0)], &adev->vmhub[AMDGPU_MMHUB1(0)]};
 	int i;
 
 	for (i = 0; i < MMHUB_NUM_INSTANCES; i++) {
@@ -647,9 +686,9 @@ static int mmhub_v9_4_set_clockgating(struct amdgpu_device *adev,
 	return 0;
 }
 
-static void mmhub_v9_4_get_clockgating(struct amdgpu_device *adev, u32 *flags)
+static void mmhub_v9_4_get_clockgating(struct amdgpu_device *adev, u64 *flags)
 {
-	int data, data1;
+	u32 data, data1;
 
 	if (amdgpu_sriov_vf(adev))
 		*flags = 0;
@@ -1560,7 +1599,7 @@ static int mmhub_v9_4_get_ras_error_count(struct amdgpu_device *adev,
 	uint32_t sec_cnt, ded_cnt;
 
 	for (i = 0; i < ARRAY_SIZE(mmhub_v9_4_ras_fields); i++) {
-		if(mmhub_v9_4_ras_fields[i].reg_offset != reg->reg_offset)
+		if (mmhub_v9_4_ras_fields[i].reg_offset != reg->reg_offset)
 			continue;
 
 		sec_cnt = (value &
@@ -1655,14 +1694,18 @@ static void mmhub_v9_4_query_ras_error_status(struct amdgpu_device *adev)
 	}
 }
 
-const struct amdgpu_mmhub_ras_funcs mmhub_v9_4_ras_funcs = {
-	.ras_late_init = amdgpu_mmhub_ras_late_init,
-	.ras_fini = amdgpu_mmhub_ras_fini,
+const struct amdgpu_ras_block_hw_ops mmhub_v9_4_ras_hw_ops = {
 	.query_ras_error_count = mmhub_v9_4_query_ras_error_count,
 	.reset_ras_error_count = mmhub_v9_4_reset_ras_error_count,
 	.query_ras_error_status = mmhub_v9_4_query_ras_error_status,
 };
 
+struct amdgpu_mmhub_ras mmhub_v9_4_ras = {
+	.ras_block = {
+		.hw_ops = &mmhub_v9_4_ras_hw_ops,
+	},
+};
+
 const struct amdgpu_mmhub_funcs mmhub_v9_4_funcs = {
 	.get_fb_location = mmhub_v9_4_get_fb_location,
 	.init = mmhub_v9_4_init,
diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.h b/drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.h
index 90436efa92ef..a48329d95f71 100644
--- a/drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.h
+++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.h
@@ -24,6 +24,6 @@
 #define __MMHUB_V9_4_H__
 
 extern const struct amdgpu_mmhub_funcs mmhub_v9_4_funcs;
-extern const struct amdgpu_mmhub_ras_funcs mmhub_v9_4_ras_funcs;
+extern struct amdgpu_mmhub_ras mmhub_v9_4_ras;
 
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/mmsch_v3_0.h b/drivers/gpu/drm/amd/amdgpu/mmsch_v3_0.h
index 3e4e858a6965..a773ef61b78c 100644
--- a/drivers/gpu/drm/amd/amdgpu/mmsch_v3_0.h
+++ b/drivers/gpu/drm/amd/amdgpu/mmsch_v3_0.h
@@ -30,6 +30,8 @@
 #define MMSCH_VERSION_MINOR	0
 #define MMSCH_VERSION	(MMSCH_VERSION_MAJOR << 16 | MMSCH_VERSION_MINOR)
 
+#define MMSCH_V3_0_VCN_INSTANCES 0x2
+
 enum mmsch_v3_0_command_type {
 	MMSCH_COMMAND__DIRECT_REG_WRITE = 0,
 	MMSCH_COMMAND__DIRECT_REG_POLLING = 2,
@@ -47,7 +49,7 @@ struct mmsch_v3_0_table_info {
 struct mmsch_v3_0_init_header {
 	uint32_t version;
 	uint32_t total_size;
-	struct mmsch_v3_0_table_info inst[AMDGPU_MAX_VCN_INSTANCES];
+	struct mmsch_v3_0_table_info inst[MMSCH_V3_0_VCN_INSTANCES];
 };
 
 struct mmsch_v3_0_cmd_direct_reg_header {
diff --git a/drivers/gpu/drm/amd/amdgpu/mmsch_v4_0.h b/drivers/gpu/drm/amd/amdgpu/mmsch_v4_0.h
new file mode 100644
index 000000000000..ced26cc5123a
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/mmsch_v4_0.h
@@ -0,0 +1,144 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __MMSCH_V4_0_H__
+#define __MMSCH_V4_0_H__
+
+#include "amdgpu_vcn.h"
+
+#define MMSCH_VERSION_MAJOR	4
+#define MMSCH_VERSION_MINOR	0
+#define MMSCH_VERSION	(MMSCH_VERSION_MAJOR << 16 | MMSCH_VERSION_MINOR)
+
+#define RB_ENABLED (1 << 0)
+#define RB4_ENABLED (1 << 1)
+
+#define MMSCH_VF_ENGINE_STATUS__PASS 0x1
+
+#define MMSCH_VF_MAILBOX_RESP__OK                      0x1
+#define MMSCH_VF_MAILBOX_RESP__INCOMPLETE              0x2
+#define MMSCH_VF_MAILBOX_RESP__FAILED                  0x3
+#define MMSCH_VF_MAILBOX_RESP__FAILED_SMALL_CTX_SIZE   0x4
+#define MMSCH_VF_MAILBOX_RESP__UNKNOWN_CMD             0x5
+
+#define MMSCH_V4_0_VCN_INSTANCES 0x2
+
+enum mmsch_v4_0_command_type {
+	MMSCH_COMMAND__DIRECT_REG_WRITE = 0,
+	MMSCH_COMMAND__DIRECT_REG_POLLING = 2,
+	MMSCH_COMMAND__DIRECT_REG_READ_MODIFY_WRITE = 3,
+	MMSCH_COMMAND__INDIRECT_REG_WRITE = 8,
+	MMSCH_COMMAND__END = 0xf
+};
+
+struct mmsch_v4_0_table_info {
+	uint32_t init_status;
+	uint32_t table_offset;
+	uint32_t table_size;
+};
+
+struct mmsch_v4_0_init_header {
+	uint32_t version;
+	uint32_t total_size;
+	struct mmsch_v4_0_table_info inst[MMSCH_V4_0_VCN_INSTANCES];
+	struct mmsch_v4_0_table_info jpegdec;
+};
+
+struct mmsch_v4_0_cmd_direct_reg_header {
+	uint32_t reg_offset   : 28;
+	uint32_t command_type : 4;
+};
+
+struct mmsch_v4_0_cmd_indirect_reg_header {
+	uint32_t reg_offset    : 20;
+	uint32_t reg_idx_space : 8;
+	uint32_t command_type  : 4;
+};
+
+struct mmsch_v4_0_cmd_direct_write {
+	struct mmsch_v4_0_cmd_direct_reg_header cmd_header;
+	uint32_t reg_value;
+};
+
+struct mmsch_v4_0_cmd_direct_read_modify_write {
+	struct mmsch_v4_0_cmd_direct_reg_header cmd_header;
+	uint32_t write_data;
+	uint32_t mask_value;
+};
+
+struct mmsch_v4_0_cmd_direct_polling {
+	struct mmsch_v4_0_cmd_direct_reg_header cmd_header;
+	uint32_t mask_value;
+	uint32_t wait_value;
+};
+
+struct mmsch_v4_0_cmd_end {
+	struct mmsch_v4_0_cmd_direct_reg_header cmd_header;
+};
+
+struct mmsch_v4_0_cmd_indirect_write {
+	struct mmsch_v4_0_cmd_indirect_reg_header cmd_header;
+	uint32_t reg_value;
+};
+
+#define MMSCH_V4_0_INSERT_DIRECT_RD_MOD_WT(reg, mask, data) { \
+	size = sizeof(struct mmsch_v4_0_cmd_direct_read_modify_write); \
+	size_dw = size / 4; \
+	direct_rd_mod_wt.cmd_header.reg_offset = reg; \
+	direct_rd_mod_wt.mask_value = mask; \
+	direct_rd_mod_wt.write_data = data; \
+	memcpy((void *)table_loc, &direct_rd_mod_wt, size); \
+	table_loc += size_dw; \
+	table_size += size_dw; \
+}
+
+#define MMSCH_V4_0_INSERT_DIRECT_WT(reg, value) { \
+	size = sizeof(struct mmsch_v4_0_cmd_direct_write); \
+	size_dw = size / 4; \
+	direct_wt.cmd_header.reg_offset = reg; \
+	direct_wt.reg_value = value; \
+	memcpy((void *)table_loc, &direct_wt, size); \
+	table_loc += size_dw; \
+	table_size += size_dw; \
+}
+
+#define MMSCH_V4_0_INSERT_DIRECT_POLL(reg, mask, wait) { \
+	size = sizeof(struct mmsch_v4_0_cmd_direct_polling); \
+	size_dw = size / 4; \
+	direct_poll.cmd_header.reg_offset = reg; \
+	direct_poll.mask_value = mask; \
+	direct_poll.wait_value = wait; \
+	memcpy((void *)table_loc, &direct_poll, size); \
+	table_loc += size_dw; \
+	table_size += size_dw; \
+}
+
+#define MMSCH_V4_0_INSERT_END() { \
+	size = sizeof(struct mmsch_v4_0_cmd_end); \
+	size_dw = size / 4; \
+	memcpy((void *)table_loc, &end, size); \
+	table_loc += size_dw; \
+	table_size += size_dw; \
+}
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/mmsch_v4_0_3.h b/drivers/gpu/drm/amd/amdgpu/mmsch_v4_0_3.h
new file mode 100644
index 000000000000..db7eb5260295
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/mmsch_v4_0_3.h
@@ -0,0 +1,37 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __MMSCH_V4_0_3_H__
+#define __MMSCH_V4_0_3_H__
+
+#include "amdgpu_vcn.h"
+#include "mmsch_v4_0.h"
+
+struct mmsch_v4_0_3_init_header {
+	uint32_t version;
+	uint32_t total_size;
+	struct mmsch_v4_0_table_info vcn0;
+	struct mmsch_v4_0_table_info mjpegdec0[4];
+	struct mmsch_v4_0_table_info mjpegdec1[4];
+};
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/mmsch_v5_0.h b/drivers/gpu/drm/amd/amdgpu/mmsch_v5_0.h
new file mode 100644
index 000000000000..6f749814929f
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/mmsch_v5_0.h
@@ -0,0 +1,144 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __MMSCH_V5_0_H__
+#define __MMSCH_V5_0_H__
+
+#include "amdgpu_vcn.h"
+
+#define MMSCH_VERSION_MAJOR	5
+#define MMSCH_VERSION_MINOR	0
+#define MMSCH_VERSION	(MMSCH_VERSION_MAJOR << 16 | MMSCH_VERSION_MINOR)
+
+#define RB_ENABLED (1 << 0)
+#define RB4_ENABLED (1 << 1)
+
+#define MMSCH_VF_ENGINE_STATUS__PASS 0x1
+
+#define MMSCH_VF_MAILBOX_RESP__OK                      0x1
+#define MMSCH_VF_MAILBOX_RESP__INCOMPLETE              0x2
+#define MMSCH_VF_MAILBOX_RESP__FAILED                  0x3
+#define MMSCH_VF_MAILBOX_RESP__FAILED_SMALL_CTX_SIZE   0x4
+#define MMSCH_VF_MAILBOX_RESP__UNKNOWN_CMD             0x5
+
+enum mmsch_v5_0_command_type {
+	MMSCH_COMMAND__DIRECT_REG_WRITE = 0,
+	MMSCH_COMMAND__DIRECT_REG_POLLING = 2,
+	MMSCH_COMMAND__DIRECT_REG_READ_MODIFY_WRITE = 3,
+	MMSCH_COMMAND__INDIRECT_REG_WRITE = 8,
+	MMSCH_COMMAND__END = 0xf
+};
+
+struct mmsch_v5_0_table_info {
+	uint32_t init_status;
+	uint32_t table_offset;
+	uint32_t table_size;
+};
+
+struct mmsch_v5_0_init_header {
+	uint32_t version;
+	uint32_t total_size;
+	struct mmsch_v5_0_table_info vcn0;
+	struct mmsch_v5_0_table_info mjpegdec0[5];
+	struct mmsch_v5_0_table_info mjpegdec1[5];
+};
+
+struct mmsch_v5_0_cmd_direct_reg_header {
+	uint32_t reg_offset   : 28;
+	uint32_t command_type : 4;
+};
+
+struct mmsch_v5_0_cmd_indirect_reg_header {
+	uint32_t reg_offset    : 20;
+	uint32_t reg_idx_space : 8;
+	uint32_t command_type  : 4;
+};
+
+struct mmsch_v5_0_cmd_direct_write {
+	struct mmsch_v5_0_cmd_direct_reg_header cmd_header;
+	uint32_t reg_value;
+};
+
+struct mmsch_v5_0_cmd_direct_read_modify_write {
+	struct mmsch_v5_0_cmd_direct_reg_header cmd_header;
+	uint32_t write_data;
+	uint32_t mask_value;
+};
+
+struct mmsch_v5_0_cmd_direct_polling {
+	struct mmsch_v5_0_cmd_direct_reg_header cmd_header;
+	uint32_t mask_value;
+	uint32_t wait_value;
+};
+
+struct mmsch_v5_0_cmd_end {
+	struct mmsch_v5_0_cmd_direct_reg_header cmd_header;
+};
+
+struct mmsch_v5_0_cmd_indirect_write {
+	struct mmsch_v5_0_cmd_indirect_reg_header cmd_header;
+	uint32_t reg_value;
+};
+
+#define MMSCH_V5_0_INSERT_DIRECT_RD_MOD_WT(reg, mask, data) { \
+	size = sizeof(struct mmsch_v5_0_cmd_direct_read_modify_write); \
+	size_dw = size / 4; \
+	direct_rd_mod_wt.cmd_header.reg_offset = reg; \
+	direct_rd_mod_wt.mask_value = mask; \
+	direct_rd_mod_wt.write_data = data; \
+	memcpy((void *)table_loc, &direct_rd_mod_wt, size); \
+	table_loc += size_dw; \
+	table_size += size_dw; \
+}
+
+#define MMSCH_V5_0_INSERT_DIRECT_WT(reg, value) { \
+	size = sizeof(struct mmsch_v5_0_cmd_direct_write); \
+	size_dw = size / 4; \
+	direct_wt.cmd_header.reg_offset = reg; \
+	direct_wt.reg_value = value; \
+	memcpy((void *)table_loc, &direct_wt, size); \
+	table_loc += size_dw; \
+	table_size += size_dw; \
+}
+
+#define MMSCH_V5_0_INSERT_DIRECT_POLL(reg, mask, wait) { \
+	size = sizeof(struct mmsch_v5_0_cmd_direct_polling); \
+	size_dw = size / 4; \
+	direct_poll.cmd_header.reg_offset = reg; \
+	direct_poll.mask_value = mask; \
+	direct_poll.wait_value = wait; \
+	memcpy((void *)table_loc, &direct_poll, size); \
+	table_loc += size_dw; \
+	table_size += size_dw; \
+}
+
+#define MMSCH_V5_0_INSERT_END() { \
+	size = sizeof(struct mmsch_v5_0_cmd_end); \
+	size_dw = size / 4; \
+	memcpy((void *)table_loc, &end, size); \
+	table_loc += size_dw; \
+	table_size += size_dw; \
+}
+
+#endif
+
diff --git a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c
index 23b066bcffb2..9a40107a0869 100644
--- a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c
+++ b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c
@@ -32,6 +32,8 @@
 #include "soc15_common.h"
 #include "mxgpu_ai.h"
 
+#include "amdgpu_reset.h"
+
 static void xgpu_ai_mailbox_send_ack(struct amdgpu_device *adev)
 {
 	WREG8(AI_MAIBOX_CONTROL_RCV_OFFSET_BYTE, 2);
@@ -91,7 +93,7 @@ static int xgpu_ai_poll_ack(struct amdgpu_device *adev)
 		timeout -= 5;
 	} while (timeout > 1);
 
-	pr_err("Doesn't get TRN_MSG_ACK from pf in %d msec\n", AI_MAILBOX_POLL_ACK_TIMEDOUT);
+	dev_err(adev->dev, "Doesn't get TRN_MSG_ACK from pf in %d msec\n", AI_MAILBOX_POLL_ACK_TIMEDOUT);
 
 	return -ETIME;
 }
@@ -109,7 +111,7 @@ static int xgpu_ai_poll_msg(struct amdgpu_device *adev, enum idh_event event)
 		timeout -= 10;
 	} while (timeout > 1);
 
-	pr_err("Doesn't get msg:%d from pf, error=%d\n", event, r);
+	dev_err(adev->dev, "Doesn't get msg:%d from pf, error=%d\n", event, r);
 
 	return -ETIME;
 }
@@ -130,7 +132,7 @@ static void xgpu_ai_mailbox_trans_msg (struct amdgpu_device *adev,
 		xgpu_ai_mailbox_set_valid(adev, false);
 		trn = xgpu_ai_peek_ack(adev);
 		if (trn) {
-			pr_err("trn=%x ACK should not assert! wait again !\n", trn);
+			dev_err_ratelimited(adev->dev, "trn=%x ACK should not assert! wait again !\n", trn);
 			msleep(1);
 		}
 	} while(trn);
@@ -153,7 +155,7 @@ static void xgpu_ai_mailbox_trans_msg (struct amdgpu_device *adev,
 	/* start to poll ack */
 	r = xgpu_ai_poll_ack(adev);
 	if (r)
-		pr_err("Doesn't get ack from pf, continue\n");
+		dev_err(adev->dev, "Doesn't get ack from pf, continue\n");
 
 	xgpu_ai_mailbox_set_valid(adev, false);
 }
@@ -171,7 +173,7 @@ static int xgpu_ai_send_access_requests(struct amdgpu_device *adev,
 		req == IDH_REQ_GPU_RESET_ACCESS) {
 		r = xgpu_ai_poll_msg(adev, IDH_READY_TO_ACCESS_GPU);
 		if (r) {
-			pr_err("Doesn't get READY_TO_ACCESS_GPU from pf, give up\n");
+			dev_err(adev->dev, "Doesn't get READY_TO_ACCESS_GPU from pf, give up\n");
 			return r;
 		}
 		/* Retrieve checksum from mailbox2 */
@@ -180,6 +182,11 @@ static int xgpu_ai_send_access_requests(struct amdgpu_device *adev,
 				RREG32_NO_KIQ(SOC15_REG_OFFSET(NBIO, 0,
 					mmBIF_BX_PF0_MAILBOX_MSGBUF_RCV_DW2));
 		}
+	} else if (req == IDH_REQ_GPU_INIT_DATA){
+		/* Dummy REQ_GPU_INIT_DATA handling */
+		r = xgpu_ai_poll_msg(adev, IDH_REQ_GPU_INIT_DATA_READY);
+		/* version set to 0 since dummy */
+		adev->virt.req_init_data_ver = 0;	
 	}
 
 	return 0;
@@ -224,7 +231,7 @@ static int xgpu_ai_mailbox_ack_irq(struct amdgpu_device *adev,
 					struct amdgpu_irq_src *source,
 					struct amdgpu_iv_entry *entry)
 {
-	DRM_DEBUG("get ack intr and do nothing.\n");
+	dev_dbg(adev->dev, "get ack intr and do nothing.\n");
 	return 0;
 }
 
@@ -242,41 +249,78 @@ static int xgpu_ai_set_mailbox_ack_irq(struct amdgpu_device *adev,
 	return 0;
 }
 
-static void xgpu_ai_mailbox_flr_work(struct work_struct *work)
+static void xgpu_ai_ready_to_reset(struct amdgpu_device *adev)
 {
-	struct amdgpu_virt *virt = container_of(work, struct amdgpu_virt, flr_work);
-	struct amdgpu_device *adev = container_of(virt, struct amdgpu_device, virt);
-	int timeout = AI_MAILBOX_POLL_FLR_TIMEDOUT;
-
-	/* block amdgpu_gpu_recover till msg FLR COMPLETE received,
-	 * otherwise the mailbox msg will be ruined/reseted by
-	 * the VF FLR.
-	 */
-	if (!down_write_trylock(&adev->reset_sem))
-		return;
-
-	amdgpu_virt_fini_data_exchange(adev);
-	atomic_set(&adev->in_gpu_reset, 1);
-
 	xgpu_ai_mailbox_trans_msg(adev, IDH_READY_TO_RESET, 0, 0, 0);
+}
 
+static int xgpu_ai_wait_reset(struct amdgpu_device *adev)
+{
+	int timeout = AI_MAILBOX_POLL_FLR_TIMEDOUT;
 	do {
-		if (xgpu_ai_mailbox_peek_msg(adev) == IDH_FLR_NOTIFICATION_CMPL)
-			goto flr_done;
-
+		if (xgpu_ai_mailbox_peek_msg(adev) == IDH_FLR_NOTIFICATION_CMPL) {
+			dev_dbg(adev->dev, "Got AI IDH_FLR_NOTIFICATION_CMPL after %d ms\n", AI_MAILBOX_POLL_FLR_TIMEDOUT - timeout);
+			return 0;
+		}
 		msleep(10);
 		timeout -= 10;
 	} while (timeout > 1);
 
-flr_done:
-	atomic_set(&adev->in_gpu_reset, 0);
-	up_write(&adev->reset_sem);
+	dev_dbg(adev->dev, "waiting AI IDH_FLR_NOTIFICATION_CMPL timeout\n");
+	return -ETIME;
+}
+
+static void xgpu_ai_mailbox_flr_work(struct work_struct *work)
+{
+	struct amdgpu_virt *virt = container_of(work, struct amdgpu_virt, flr_work);
+	struct amdgpu_device *adev = container_of(virt, struct amdgpu_device, virt);
+	struct amdgpu_reset_context reset_context = { 0 };
+
+	amdgpu_virt_fini_data_exchange(adev);
 
 	/* Trigger recovery for world switch failure if no TDR */
 	if (amdgpu_device_should_recover_gpu(adev)
 		&& (!amdgpu_device_has_job_running(adev) ||
-		adev->sdma_timeout == MAX_SCHEDULE_TIMEOUT))
-		amdgpu_device_gpu_recover(adev, NULL);
+			adev->sdma_timeout == MAX_SCHEDULE_TIMEOUT)) {
+
+		reset_context.method = AMD_RESET_METHOD_NONE;
+		reset_context.reset_req_dev = adev;
+		clear_bit(AMDGPU_NEED_FULL_RESET, &reset_context.flags);
+		set_bit(AMDGPU_HOST_FLR, &reset_context.flags);
+
+		amdgpu_device_gpu_recover(adev, NULL, &reset_context);
+	}
+}
+
+static void xgpu_ai_mailbox_req_bad_pages_work(struct work_struct *work)
+{
+	struct amdgpu_virt *virt = container_of(work, struct amdgpu_virt, req_bad_pages_work);
+	struct amdgpu_device *adev = container_of(virt, struct amdgpu_device, virt);
+
+	if (down_read_trylock(&adev->reset_domain->sem)) {
+		amdgpu_virt_fini_data_exchange(adev);
+		amdgpu_virt_request_bad_pages(adev);
+		up_read(&adev->reset_domain->sem);
+	}
+}
+
+/**
+ * xgpu_ai_mailbox_handle_bad_pages_work - Reinitialize the data exchange region to get fresh bad page information
+ * @work: pointer to the work_struct
+ *
+ * This work handler is triggered when bad pages are ready, and it reinitializes
+ * the data exchange region to retrieve updated bad page information from the host.
+ */
+static void xgpu_ai_mailbox_handle_bad_pages_work(struct work_struct *work)
+{
+	struct amdgpu_virt *virt = container_of(work, struct amdgpu_virt, handle_bad_pages_work);
+	struct amdgpu_device *adev = container_of(virt, struct amdgpu_device, virt);
+
+	if (down_read_trylock(&adev->reset_domain->sem)) {
+		amdgpu_virt_fini_data_exchange(adev);
+		amdgpu_virt_init_data_exchange(adev);
+		up_read(&adev->reset_domain->sem);
+	}
 }
 
 static int xgpu_ai_set_mailbox_rcv_irq(struct amdgpu_device *adev,
@@ -298,23 +342,47 @@ static int xgpu_ai_mailbox_rcv_irq(struct amdgpu_device *adev,
 				   struct amdgpu_iv_entry *entry)
 {
 	enum idh_event event = xgpu_ai_mailbox_peek_msg(adev);
+	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
 
 	switch (event) {
-		case IDH_FLR_NOTIFICATION:
+	case IDH_RAS_BAD_PAGES_READY:
+		xgpu_ai_mailbox_send_ack(adev);
 		if (amdgpu_sriov_runtime(adev))
-			schedule_work(&adev->virt.flr_work);
+			schedule_work(&adev->virt.handle_bad_pages_work);
 		break;
-		case IDH_QUERY_ALIVE:
-			xgpu_ai_mailbox_send_ack(adev);
-			break;
-		/* READY_TO_ACCESS_GPU is fetched by kernel polling, IRQ can ignore
-		 * it byfar since that polling thread will handle it,
-		 * other msg like flr complete is not handled here.
-		 */
-		case IDH_CLR_MSG_BUF:
-		case IDH_FLR_NOTIFICATION_CMPL:
-		case IDH_READY_TO_ACCESS_GPU:
-		default:
+	case IDH_RAS_BAD_PAGES_NOTIFICATION:
+		xgpu_ai_mailbox_send_ack(adev);
+		if (amdgpu_sriov_runtime(adev))
+			schedule_work(&adev->virt.req_bad_pages_work);
+		break;
+	case IDH_UNRECOV_ERR_NOTIFICATION:
+		xgpu_ai_mailbox_send_ack(adev);
+		ras->is_rma = true;
+		dev_err(adev->dev, "VF is in an unrecoverable state. Runtime Services are halted.\n");
+		if (amdgpu_sriov_runtime(adev))
+			WARN_ONCE(!amdgpu_reset_domain_schedule(adev->reset_domain,
+					&adev->virt.flr_work),
+					"Failed to queue work! at %s",
+					__func__);
+		break;
+	case IDH_FLR_NOTIFICATION:
+		if (amdgpu_sriov_runtime(adev))
+			WARN_ONCE(!amdgpu_reset_domain_schedule(adev->reset_domain,
+						&adev->virt.flr_work),
+					"Failed to queue work! at %s",
+					__func__);
+		break;
+	case IDH_QUERY_ALIVE:
+		xgpu_ai_mailbox_send_ack(adev);
+		break;
+	/* READY_TO_ACCESS_GPU is fetched by kernel polling, IRQ can ignore
+	 * it byfar since that polling thread will handle it,
+	 * other msg like flr complete is not handled here.
+	 */
+	case IDH_CLR_MSG_BUF:
+	case IDH_FLR_NOTIFICATION_CMPL:
+	case IDH_READY_TO_ACCESS_GPU:
+	default:
 		break;
 	}
 
@@ -370,6 +438,8 @@ int xgpu_ai_mailbox_get_irq(struct amdgpu_device *adev)
 	}
 
 	INIT_WORK(&adev->virt.flr_work, xgpu_ai_mailbox_flr_work);
+	INIT_WORK(&adev->virt.req_bad_pages_work, xgpu_ai_mailbox_req_bad_pages_work);
+	INIT_WORK(&adev->virt.handle_bad_pages_work, xgpu_ai_mailbox_handle_bad_pages_work);
 
 	return 0;
 }
@@ -380,10 +450,32 @@ void xgpu_ai_mailbox_put_irq(struct amdgpu_device *adev)
 	amdgpu_irq_put(adev, &adev->virt.rcv_irq, 0);
 }
 
+static int xgpu_ai_request_init_data(struct amdgpu_device *adev)
+{
+	return xgpu_ai_send_access_requests(adev, IDH_REQ_GPU_INIT_DATA);
+}
+
+static void xgpu_ai_ras_poison_handler(struct amdgpu_device *adev,
+					enum amdgpu_ras_block block)
+{
+	xgpu_ai_send_access_requests(adev, IDH_RAS_POISON);
+}
+
+static bool xgpu_ai_rcvd_ras_intr(struct amdgpu_device *adev)
+{
+	enum idh_event msg = xgpu_ai_mailbox_peek_msg(adev);
+
+	return (msg == IDH_RAS_ERROR_DETECTED || msg == 0xFFFFFFFF);
+}
+
 const struct amdgpu_virt_ops xgpu_ai_virt_ops = {
 	.req_full_gpu	= xgpu_ai_request_full_gpu_access,
 	.rel_full_gpu	= xgpu_ai_release_full_gpu_access,
 	.reset_gpu = xgpu_ai_request_reset,
-	.wait_reset = NULL,
+	.ready_to_reset = xgpu_ai_ready_to_reset,
+	.wait_reset = xgpu_ai_wait_reset,
 	.trans_msg = xgpu_ai_mailbox_trans_msg,
+	.req_init_data  = xgpu_ai_request_init_data,
+	.ras_poison_handler = xgpu_ai_ras_poison_handler,
+	.rcvd_ras_intr = xgpu_ai_rcvd_ras_intr,
 };
diff --git a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.h b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.h
index bd3b23171579..874b9f8f9804 100644
--- a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.h
+++ b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.h
@@ -26,7 +26,7 @@
 
 #define AI_MAILBOX_POLL_ACK_TIMEDOUT	500
 #define AI_MAILBOX_POLL_MSG_TIMEDOUT	6000
-#define AI_MAILBOX_POLL_FLR_TIMEDOUT	5000
+#define AI_MAILBOX_POLL_FLR_TIMEDOUT	10000
 #define AI_MAILBOX_POLL_MSG_REP_MAX	11
 
 enum idh_request {
@@ -35,9 +35,12 @@ enum idh_request {
 	IDH_REQ_GPU_FINI_ACCESS,
 	IDH_REL_GPU_FINI_ACCESS,
 	IDH_REQ_GPU_RESET_ACCESS,
+	IDH_REQ_GPU_INIT_DATA,
 
 	IDH_LOG_VF_ERROR       = 200,
 	IDH_READY_TO_RESET 	= 201,
+	IDH_RAS_POISON  = 202,
+	IDH_REQ_RAS_BAD_PAGES = 205,
 };
 
 enum idh_event {
@@ -48,7 +51,13 @@ enum idh_event {
 	IDH_SUCCESS,
 	IDH_FAIL,
 	IDH_QUERY_ALIVE,
-
+	IDH_REQ_GPU_INIT_DATA_READY,
+	IDH_RAS_POISON_READY,
+	IDH_PF_SOFT_FLR_NOTIFICATION,
+	IDH_RAS_ERROR_DETECTED,
+	IDH_RAS_BAD_PAGES_READY = 15,
+	IDH_RAS_BAD_PAGES_NOTIFICATION = 16,
+	IDH_UNRECOV_ERR_NOTIFICATION = 17,
 	IDH_TEXT_MESSAGE = 255,
 };
 
@@ -59,7 +68,9 @@ int xgpu_ai_mailbox_add_irq_id(struct amdgpu_device *adev);
 int xgpu_ai_mailbox_get_irq(struct amdgpu_device *adev);
 void xgpu_ai_mailbox_put_irq(struct amdgpu_device *adev);
 
-#define AI_MAIBOX_CONTROL_TRN_OFFSET_BYTE SOC15_REG_OFFSET(NBIO, 0, mmBIF_BX_PF0_MAILBOX_CONTROL) * 4
-#define AI_MAIBOX_CONTROL_RCV_OFFSET_BYTE SOC15_REG_OFFSET(NBIO, 0, mmBIF_BX_PF0_MAILBOX_CONTROL) * 4 + 1
+#define AI_MAIBOX_CONTROL_TRN_OFFSET_BYTE \
+	(SOC15_REG_OFFSET(NBIO, 0, mmBIF_BX_PF0_MAILBOX_CONTROL) * 4)
+#define AI_MAIBOX_CONTROL_RCV_OFFSET_BYTE \
+	(SOC15_REG_OFFSET(NBIO, 0, mmBIF_BX_PF0_MAILBOX_CONTROL) * 4 + 1)
 
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.c b/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.c
index a35e6d87e537..e7cd07383d56 100644
--- a/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.c
+++ b/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.c
@@ -31,6 +31,8 @@
 #include "soc15_common.h"
 #include "mxgpu_nv.h"
 
+#include "amdgpu_reset.h"
+
 static void xgpu_nv_mailbox_send_ack(struct amdgpu_device *adev)
 {
 	WREG8(NV_MAIBOX_CONTROL_RCV_OFFSET_BYTE, 2);
@@ -59,15 +61,20 @@ static enum idh_event xgpu_nv_mailbox_peek_msg(struct amdgpu_device *adev)
 static int xgpu_nv_mailbox_rcv_msg(struct amdgpu_device *adev,
 				   enum idh_event event)
 {
+	int r = 0;
 	u32 reg;
 
 	reg = RREG32_NO_KIQ(mmMAILBOX_MSGBUF_RCV_DW0);
-	if (reg != event)
+	if (reg == IDH_FAIL)
+		r = -EINVAL;
+	if (reg == IDH_UNRECOV_ERR_NOTIFICATION)
+		r = -ENODEV;
+	else if (reg != event)
 		return -ENOENT;
 
 	xgpu_nv_mailbox_send_ack(adev);
 
-	return 0;
+	return r;
 }
 
 static uint8_t xgpu_nv_peek_ack(struct amdgpu_device *adev)
@@ -89,7 +96,7 @@ static int xgpu_nv_poll_ack(struct amdgpu_device *adev)
 		timeout -= 5;
 	} while (timeout > 1);
 
-	pr_err("Doesn't get TRN_MSG_ACK from pf in %d msec\n", NV_MAILBOX_POLL_ACK_TIMEDOUT);
+	dev_err(adev->dev, "Doesn't get TRN_MSG_ACK from pf in %d msec \n", NV_MAILBOX_POLL_ACK_TIMEDOUT);
 
 	return -ETIME;
 }
@@ -98,19 +105,31 @@ static int xgpu_nv_poll_msg(struct amdgpu_device *adev, enum idh_event event)
 {
 	int r;
 	uint64_t timeout, now;
+	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
 
 	now = (uint64_t)ktime_to_ms(ktime_get());
 	timeout = now + NV_MAILBOX_POLL_MSG_TIMEDOUT;
 
 	do {
 		r = xgpu_nv_mailbox_rcv_msg(adev, event);
-		if (!r)
+		if (!r) {
+			dev_dbg(adev->dev, "rcv_msg 0x%x after %llu ms\n",
+					event, NV_MAILBOX_POLL_MSG_TIMEDOUT - timeout + now);
 			return 0;
+		} else if (r == -ENODEV) {
+			if (!amdgpu_ras_is_rma(adev)) {
+				ras->is_rma = true;
+				dev_err(adev->dev, "VF is in an unrecoverable state. "
+						"Runtime Services are halted.\n");
+			}
+			return r;
+		}
 
 		msleep(10);
 		now = (uint64_t)ktime_to_ms(ktime_get());
 	} while (timeout > now);
 
+	dev_dbg(adev->dev, "nv_poll_msg timed out\n");
 
 	return -ETIME;
 }
@@ -131,11 +150,12 @@ static void xgpu_nv_mailbox_trans_msg (struct amdgpu_device *adev,
 		xgpu_nv_mailbox_set_valid(adev, false);
 		trn = xgpu_nv_peek_ack(adev);
 		if (trn) {
-			pr_err("trn=%x ACK should not assert! wait again !\n", trn);
+			dev_err_ratelimited(adev->dev, "trn=%x ACK should not assert! wait again !\n", trn);
 			msleep(1);
 		}
 	} while (trn);
 
+	dev_dbg(adev->dev, "trans_msg req = 0x%x, data1 = 0x%x\n", req, data1);
 	WREG32_NO_KIQ(mmMAILBOX_MSGBUF_TRN_DW0, req);
 	WREG32_NO_KIQ(mmMAILBOX_MSGBUF_TRN_DW1, data1);
 	WREG32_NO_KIQ(mmMAILBOX_MSGBUF_TRN_DW2, data2);
@@ -145,19 +165,27 @@ static void xgpu_nv_mailbox_trans_msg (struct amdgpu_device *adev,
 	/* start to poll ack */
 	r = xgpu_nv_poll_ack(adev);
 	if (r)
-		pr_err("Doesn't get ack from pf, continue\n");
+		dev_err(adev->dev, "Doesn't get ack from pf, continue\n");
 
 	xgpu_nv_mailbox_set_valid(adev, false);
 }
 
-static int xgpu_nv_send_access_requests(struct amdgpu_device *adev,
-					enum idh_request req)
+static int xgpu_nv_send_access_requests_with_param(struct amdgpu_device *adev,
+			enum idh_request req, u32 data1, u32 data2, u32 data3)
 {
-	int r, retry = 1;
+	struct amdgpu_virt *virt = &adev->virt;
+	int r = 0, retry = 1;
 	enum idh_event event = -1;
 
+	mutex_lock(&virt->access_req_mutex);
 send_request:
-	xgpu_nv_mailbox_trans_msg(adev, req, 0, 0, 0);
+
+	if (amdgpu_ras_is_rma(adev)) {
+		r = -ENODEV;
+		goto out;
+	}
+
+	xgpu_nv_mailbox_trans_msg(adev, req, data1, data2, data3);
 
 	switch (req) {
 	case IDH_REQ_GPU_INIT_ACCESS:
@@ -168,6 +196,19 @@ send_request:
 	case IDH_REQ_GPU_INIT_DATA:
 		event = IDH_REQ_GPU_INIT_DATA_READY;
 		break;
+	case IDH_RAS_POISON:
+		if (data1 != 0)
+			event = IDH_RAS_POISON_READY;
+		break;
+	case IDH_REQ_RAS_ERROR_COUNT:
+		event = IDH_RAS_ERROR_COUNT_READY;
+		break;
+	case IDH_REQ_RAS_CPER_DUMP:
+		event = IDH_RAS_CPER_DUMP_READY;
+		break;
+	case IDH_REQ_RAS_CHK_CRITI:
+		event = IDH_REQ_RAS_CHK_CRITI_READY;
+		break;
 	default:
 		break;
 	}
@@ -175,24 +216,30 @@ send_request:
 	if (event != -1) {
 		r = xgpu_nv_poll_msg(adev, event);
 		if (r) {
-			if (retry++ < 2)
+			if (retry++ < 5)
 				goto send_request;
 
 			if (req != IDH_REQ_GPU_INIT_DATA) {
-				pr_err("Doesn't get msg:%d from pf, error=%d\n", event, r);
-				return r;
-			}
-			else /* host doesn't support REQ_GPU_INIT_DATA handshake */
+				dev_err(adev->dev, "Doesn't get msg:%d from pf, error=%d\n", event, r);
+				goto out;
+			} else /* host doesn't support REQ_GPU_INIT_DATA handshake */
 				adev->virt.req_init_data_ver = 0;
 		} else {
-			if (req == IDH_REQ_GPU_INIT_DATA)
-			{
-				adev->virt.req_init_data_ver =
-					RREG32_NO_KIQ(mmMAILBOX_MSGBUF_RCV_DW1);
-
-				/* assume V1 in case host doesn't set version number */
-				if (adev->virt.req_init_data_ver < 1)
-					adev->virt.req_init_data_ver = 1;
+			if (req == IDH_REQ_GPU_INIT_DATA) {
+				switch (RREG32_NO_KIQ(mmMAILBOX_MSGBUF_RCV_DW1)) {
+				case GPU_CRIT_REGION_V2:
+					adev->virt.req_init_data_ver = GPU_CRIT_REGION_V2;
+					adev->virt.init_data_header.offset =
+						RREG32_NO_KIQ(mmMAILBOX_MSGBUF_RCV_DW2);
+					adev->virt.init_data_header.size_kb =
+						RREG32_NO_KIQ(mmMAILBOX_MSGBUF_RCV_DW3);
+					break;
+				default:
+					adev->virt.req_init_data_ver = GPU_CRIT_REGION_V1;
+					adev->virt.init_data_header.offset = -1;
+					adev->virt.init_data_header.size_kb = 0;
+					break;
+				}
 			}
 		}
 
@@ -203,7 +250,17 @@ send_request:
 		}
 	}
 
-	return 0;
+out:
+	mutex_unlock(&virt->access_req_mutex);
+
+	return r;
+}
+
+static int xgpu_nv_send_access_requests(struct amdgpu_device *adev,
+					enum idh_request req)
+{
+	return xgpu_nv_send_access_requests_with_param(adev,
+						req, 0, 0, 0);
 }
 
 static int xgpu_nv_request_reset(struct amdgpu_device *adev)
@@ -243,14 +300,15 @@ static int xgpu_nv_release_full_gpu_access(struct amdgpu_device *adev,
 
 static int xgpu_nv_request_init_data(struct amdgpu_device *adev)
 {
-	return xgpu_nv_send_access_requests(adev, IDH_REQ_GPU_INIT_DATA);
+	return xgpu_nv_send_access_requests_with_param(adev, IDH_REQ_GPU_INIT_DATA,
+			0, GPU_CRIT_REGION_V2, 0);
 }
 
 static int xgpu_nv_mailbox_ack_irq(struct amdgpu_device *adev,
 					struct amdgpu_irq_src *source,
 					struct amdgpu_iv_entry *entry)
 {
-	DRM_DEBUG("get ack intr and do nothing.\n");
+	dev_dbg(adev->dev, "get ack intr and do nothing.\n");
 	return 0;
 }
 
@@ -271,35 +329,34 @@ static int xgpu_nv_set_mailbox_ack_irq(struct amdgpu_device *adev,
 	return 0;
 }
 
-static void xgpu_nv_mailbox_flr_work(struct work_struct *work)
+static void xgpu_nv_ready_to_reset(struct amdgpu_device *adev)
 {
-	struct amdgpu_virt *virt = container_of(work, struct amdgpu_virt, flr_work);
-	struct amdgpu_device *adev = container_of(virt, struct amdgpu_device, virt);
-	int timeout = NV_MAILBOX_POLL_FLR_TIMEDOUT;
-
-	/* block amdgpu_gpu_recover till msg FLR COMPLETE received,
-	 * otherwise the mailbox msg will be ruined/reseted by
-	 * the VF FLR.
-	 */
-	if (!down_write_trylock(&adev->reset_sem))
-		return;
-
-	amdgpu_virt_fini_data_exchange(adev);
-	atomic_set(&adev->in_gpu_reset, 1);
-
 	xgpu_nv_mailbox_trans_msg(adev, IDH_READY_TO_RESET, 0, 0, 0);
+}
 
+static int xgpu_nv_wait_reset(struct amdgpu_device *adev)
+{
+	int timeout = NV_MAILBOX_POLL_FLR_TIMEDOUT;
 	do {
-		if (xgpu_nv_mailbox_peek_msg(adev) == IDH_FLR_NOTIFICATION_CMPL)
-			goto flr_done;
-
+		if (xgpu_nv_mailbox_peek_msg(adev) == IDH_FLR_NOTIFICATION_CMPL) {
+			dev_dbg(adev->dev, "Got NV IDH_FLR_NOTIFICATION_CMPL after %d ms\n", NV_MAILBOX_POLL_FLR_TIMEDOUT - timeout);
+			return 0;
+		}
 		msleep(10);
 		timeout -= 10;
 	} while (timeout > 1);
 
-flr_done:
-	atomic_set(&adev->in_gpu_reset, 0);
-	up_write(&adev->reset_sem);
+	dev_dbg(adev->dev, "waiting NV IDH_FLR_NOTIFICATION_CMPL timeout\n");
+	return -ETIME;
+}
+
+static void xgpu_nv_mailbox_flr_work(struct work_struct *work)
+{
+	struct amdgpu_virt *virt = container_of(work, struct amdgpu_virt, flr_work);
+	struct amdgpu_device *adev = container_of(virt, struct amdgpu_device, virt);
+	struct amdgpu_reset_context reset_context = { 0 };
+
+	amdgpu_virt_fini_data_exchange(adev);
 
 	/* Trigger recovery for world switch failure if no TDR */
 	if (amdgpu_device_should_recover_gpu(adev)
@@ -307,8 +364,46 @@ flr_done:
 		adev->sdma_timeout == MAX_SCHEDULE_TIMEOUT ||
 		adev->gfx_timeout == MAX_SCHEDULE_TIMEOUT ||
 		adev->compute_timeout == MAX_SCHEDULE_TIMEOUT ||
-		adev->video_timeout == MAX_SCHEDULE_TIMEOUT))
-		amdgpu_device_gpu_recover(adev, NULL);
+		adev->video_timeout == MAX_SCHEDULE_TIMEOUT)) {
+
+		reset_context.method = AMD_RESET_METHOD_NONE;
+		reset_context.reset_req_dev = adev;
+		clear_bit(AMDGPU_NEED_FULL_RESET, &reset_context.flags);
+		set_bit(AMDGPU_HOST_FLR, &reset_context.flags);
+
+		amdgpu_device_gpu_recover(adev, NULL, &reset_context);
+	}
+}
+
+static void xgpu_nv_mailbox_req_bad_pages_work(struct work_struct *work)
+{
+	struct amdgpu_virt *virt = container_of(work, struct amdgpu_virt, req_bad_pages_work);
+	struct amdgpu_device *adev = container_of(virt, struct amdgpu_device, virt);
+
+	if (down_read_trylock(&adev->reset_domain->sem)) {
+		amdgpu_virt_fini_data_exchange(adev);
+		amdgpu_virt_request_bad_pages(adev);
+		up_read(&adev->reset_domain->sem);
+	}
+}
+
+/**
+ * xgpu_nv_mailbox_handle_bad_pages_work - Reinitialize the data exchange region to get fresh bad page information
+ * @work: pointer to the work_struct
+ *
+ * This work handler is triggered when bad pages are ready, and it reinitializes
+ * the data exchange region to retrieve updated bad page information from the host.
+ */
+static void xgpu_nv_mailbox_handle_bad_pages_work(struct work_struct *work)
+{
+	struct amdgpu_virt *virt = container_of(work, struct amdgpu_virt, handle_bad_pages_work);
+	struct amdgpu_device *adev = container_of(virt, struct amdgpu_device, virt);
+
+	if (down_read_trylock(&adev->reset_domain->sem)) {
+		amdgpu_virt_fini_data_exchange(adev);
+		amdgpu_virt_init_data_exchange(adev);
+		up_read(&adev->reset_domain->sem);
+	}
 }
 
 static int xgpu_nv_set_mailbox_rcv_irq(struct amdgpu_device *adev,
@@ -333,11 +428,38 @@ static int xgpu_nv_mailbox_rcv_irq(struct amdgpu_device *adev,
 				   struct amdgpu_iv_entry *entry)
 {
 	enum idh_event event = xgpu_nv_mailbox_peek_msg(adev);
+	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
 
 	switch (event) {
+	case IDH_RAS_BAD_PAGES_READY:
+		xgpu_nv_mailbox_send_ack(adev);
+		if (amdgpu_sriov_runtime(adev))
+			schedule_work(&adev->virt.handle_bad_pages_work);
+		break;
+	case IDH_RAS_BAD_PAGES_NOTIFICATION:
+		xgpu_nv_mailbox_send_ack(adev);
+		if (amdgpu_sriov_runtime(adev))
+			schedule_work(&adev->virt.req_bad_pages_work);
+		break;
+	case IDH_UNRECOV_ERR_NOTIFICATION:
+		xgpu_nv_mailbox_send_ack(adev);
+		if (!amdgpu_ras_is_rma(adev)) {
+			ras->is_rma = true;
+			dev_err(adev->dev, "VF is in an unrecoverable state. Runtime Services are halted.\n");
+		}
+
+		if (amdgpu_sriov_runtime(adev))
+			WARN_ONCE(!amdgpu_reset_domain_schedule(adev->reset_domain,
+						&adev->virt.flr_work),
+					"Failed to queue work! at %s",
+					__func__);
+		break;
 	case IDH_FLR_NOTIFICATION:
 		if (amdgpu_sriov_runtime(adev))
-			schedule_work(&adev->virt.flr_work);
+			WARN_ONCE(!amdgpu_reset_domain_schedule(adev->reset_domain,
+				   &adev->virt.flr_work),
+				  "Failed to queue work! at %s",
+				  __func__);
 		break;
 		/* READY_TO_ACCESS_GPU is fetched by kernel polling, IRQ can ignore
 		 * it byfar since that polling thread will handle it,
@@ -402,6 +524,8 @@ int xgpu_nv_mailbox_get_irq(struct amdgpu_device *adev)
 	}
 
 	INIT_WORK(&adev->virt.flr_work, xgpu_nv_mailbox_flr_work);
+	INIT_WORK(&adev->virt.req_bad_pages_work, xgpu_nv_mailbox_req_bad_pages_work);
+	INIT_WORK(&adev->virt.handle_bad_pages_work, xgpu_nv_mailbox_handle_bad_pages_work);
 
 	return 0;
 }
@@ -412,11 +536,67 @@ void xgpu_nv_mailbox_put_irq(struct amdgpu_device *adev)
 	amdgpu_irq_put(adev, &adev->virt.rcv_irq, 0);
 }
 
+static void xgpu_nv_ras_poison_handler(struct amdgpu_device *adev,
+		enum amdgpu_ras_block block)
+{
+	if (amdgpu_ip_version(adev, UMC_HWIP, 0) < IP_VERSION(12, 0, 0)) {
+		xgpu_nv_send_access_requests(adev, IDH_RAS_POISON);
+	} else {
+		amdgpu_virt_fini_data_exchange(adev);
+		xgpu_nv_send_access_requests_with_param(adev,
+					IDH_RAS_POISON,	block, 0, 0);
+	}
+}
+
+static bool xgpu_nv_rcvd_ras_intr(struct amdgpu_device *adev)
+{
+	enum idh_event msg = xgpu_nv_mailbox_peek_msg(adev);
+
+	return (msg == IDH_RAS_ERROR_DETECTED || msg == 0xFFFFFFFF);
+}
+
+static int xgpu_nv_req_ras_err_count(struct amdgpu_device *adev)
+{
+	return xgpu_nv_send_access_requests(adev, IDH_REQ_RAS_ERROR_COUNT);
+}
+
+static int xgpu_nv_req_ras_cper_dump(struct amdgpu_device *adev, u64 vf_rptr)
+{
+	uint32_t vf_rptr_hi, vf_rptr_lo;
+
+	vf_rptr_hi = (uint32_t)(vf_rptr >> 32);
+	vf_rptr_lo = (uint32_t)(vf_rptr & 0xFFFFFFFF);
+	return xgpu_nv_send_access_requests_with_param(
+		adev, IDH_REQ_RAS_CPER_DUMP, vf_rptr_hi, vf_rptr_lo, 0);
+}
+
+static int xgpu_nv_req_ras_bad_pages(struct amdgpu_device *adev)
+{
+	return xgpu_nv_send_access_requests(adev, IDH_REQ_RAS_BAD_PAGES);
+}
+
+static int xgpu_nv_check_vf_critical_region(struct amdgpu_device *adev, u64 addr)
+{
+	uint32_t addr_hi, addr_lo;
+
+	addr_hi = (uint32_t)(addr >> 32);
+	addr_lo = (uint32_t)(addr & 0xFFFFFFFF);
+	return xgpu_nv_send_access_requests_with_param(
+		adev, IDH_REQ_RAS_CHK_CRITI, addr_hi, addr_lo, 0);
+}
+
 const struct amdgpu_virt_ops xgpu_nv_virt_ops = {
 	.req_full_gpu	= xgpu_nv_request_full_gpu_access,
 	.rel_full_gpu	= xgpu_nv_release_full_gpu_access,
 	.req_init_data  = xgpu_nv_request_init_data,
 	.reset_gpu = xgpu_nv_request_reset,
-	.wait_reset = NULL,
+	.ready_to_reset = xgpu_nv_ready_to_reset,
+	.wait_reset = xgpu_nv_wait_reset,
 	.trans_msg = xgpu_nv_mailbox_trans_msg,
+	.ras_poison_handler = xgpu_nv_ras_poison_handler,
+	.rcvd_ras_intr = xgpu_nv_rcvd_ras_intr,
+	.req_ras_err_count = xgpu_nv_req_ras_err_count,
+	.req_ras_cper_dump = xgpu_nv_req_ras_cper_dump,
+	.req_bad_pages = xgpu_nv_req_ras_bad_pages,
+	.req_ras_chk_criti = xgpu_nv_check_vf_critical_region
 };
diff --git a/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.h b/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.h
index 73887b0aa1d6..c1083e5e41e0 100644
--- a/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.h
+++ b/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.h
@@ -25,8 +25,8 @@
 #define __MXGPU_NV_H__
 
 #define NV_MAILBOX_POLL_ACK_TIMEDOUT	500
-#define NV_MAILBOX_POLL_MSG_TIMEDOUT	6000
-#define NV_MAILBOX_POLL_FLR_TIMEDOUT	5000
+#define NV_MAILBOX_POLL_MSG_TIMEDOUT	15000
+#define NV_MAILBOX_POLL_FLR_TIMEDOUT	10000
 #define NV_MAILBOX_POLL_MSG_REP_MAX	11
 
 enum idh_request {
@@ -39,6 +39,11 @@ enum idh_request {
 
 	IDH_LOG_VF_ERROR	= 200,
 	IDH_READY_TO_RESET 	= 201,
+	IDH_RAS_POISON	= 202,
+	IDH_REQ_RAS_ERROR_COUNT = 203,
+	IDH_REQ_RAS_CPER_DUMP = 204,
+	IDH_REQ_RAS_BAD_PAGES = 205,
+	IDH_REQ_RAS_CHK_CRITI = 206
 };
 
 enum idh_event {
@@ -50,6 +55,15 @@ enum idh_event {
 	IDH_FAIL,
 	IDH_QUERY_ALIVE,
 	IDH_REQ_GPU_INIT_DATA_READY,
+	IDH_RAS_POISON_READY,
+	IDH_PF_SOFT_FLR_NOTIFICATION,
+	IDH_RAS_ERROR_DETECTED,
+	IDH_RAS_ERROR_COUNT_READY = 11,
+	IDH_RAS_CPER_DUMP_READY = 14,
+	IDH_RAS_BAD_PAGES_READY = 15,
+	IDH_RAS_BAD_PAGES_NOTIFICATION = 16,
+	IDH_UNRECOV_ERR_NOTIFICATION = 17,
+	IDH_REQ_RAS_CHK_CRITI_READY = 18,
 
 	IDH_TEXT_MESSAGE = 255,
 };
diff --git a/drivers/gpu/drm/amd/amdgpu/mxgpu_vi.c b/drivers/gpu/drm/amd/amdgpu/mxgpu_vi.c
index aef9d059ae52..e1d63bed84bf 100644
--- a/drivers/gpu/drm/amd/amdgpu/mxgpu_vi.c
+++ b/drivers/gpu/drm/amd/amdgpu/mxgpu_vi.c
@@ -42,6 +42,8 @@
 #include "smu/smu_7_1_3_d.h"
 #include "mxgpu_vi.h"
 
+#include "amdgpu_reset.h"
+
 /* VI golden setting */
 static const u32 xgpu_fiji_mgcg_cgcg_init[] = {
 	mmRLC_CGTT_MGCG_OVERRIDE, 0xffffffff, 0xffffffff,
@@ -332,7 +334,7 @@ static void xgpu_vi_mailbox_send_ack(struct amdgpu_device *adev)
 			break;
 		}
 		mdelay(1);
-		timeout -=1;
+		timeout -= 1;
 
 		reg = RREG32_NO_KIQ(mmMAILBOX_CONTROL);
 	}
@@ -513,15 +515,18 @@ static void xgpu_vi_mailbox_flr_work(struct work_struct *work)
 	struct amdgpu_virt *virt = container_of(work, struct amdgpu_virt, flr_work);
 	struct amdgpu_device *adev = container_of(virt, struct amdgpu_device, virt);
 
-	/* wait until RCV_MSG become 3 */
-	if (xgpu_vi_poll_msg(adev, IDH_FLR_NOTIFICATION_CMPL)) {
-		pr_err("failed to receive FLR_CMPL\n");
-		return;
-	}
-
 	/* Trigger recovery due to world switch failure */
-	if (amdgpu_device_should_recover_gpu(adev))
-		amdgpu_device_gpu_recover(adev, NULL);
+	if (amdgpu_device_should_recover_gpu(adev)) {
+		struct amdgpu_reset_context reset_context;
+		memset(&reset_context, 0, sizeof(reset_context));
+
+		reset_context.method = AMD_RESET_METHOD_NONE;
+		reset_context.reset_req_dev = adev;
+		clear_bit(AMDGPU_NEED_FULL_RESET, &reset_context.flags);
+		set_bit(AMDGPU_HOST_FLR, &reset_context.flags);
+
+		amdgpu_device_gpu_recover(adev, NULL, &reset_context);
+	}
 }
 
 static int xgpu_vi_set_mailbox_rcv_irq(struct amdgpu_device *adev,
@@ -544,14 +549,17 @@ static int xgpu_vi_mailbox_rcv_irq(struct amdgpu_device *adev,
 {
 	int r;
 
-	/* trigger gpu-reset by hypervisor only if TDR disbaled */
+	/* trigger gpu-reset by hypervisor only if TDR disabled */
 	if (!amdgpu_gpu_recovery) {
 		/* see what event we get */
 		r = xgpu_vi_mailbox_rcv_msg(adev, IDH_FLR_NOTIFICATION);
 
 		/* only handle FLR_NOTIFY now */
 		if (!r)
-			schedule_work(&adev->virt.flr_work);
+			WARN_ONCE(!amdgpu_reset_domain_schedule(adev->reset_domain,
+								&adev->virt.flr_work),
+				  "Failed to queue work! at %s",
+				  __func__);
 	}
 
 	return 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/navi10_ih.c b/drivers/gpu/drm/amd/amdgpu/navi10_ih.c
index 38241cf0e1f1..4cd325149b63 100644
--- a/drivers/gpu/drm/amd/amdgpu/navi10_ih.c
+++ b/drivers/gpu/drm/amd/amdgpu/navi10_ih.c
@@ -107,7 +107,7 @@ force_update_wptr_for_self_int(struct amdgpu_device *adev,
 {
 	u32 ih_cntl, ih_rb_cntl;
 
-	if (adev->ip_versions[OSSSYS_HWIP][0] < IP_VERSION(5, 0, 3))
+	if (amdgpu_ip_version(adev, OSSSYS_HWIP, 0) < IP_VERSION(5, 0, 3))
 		return;
 
 	ih_cntl = RREG32_SOC15(OSSSYS, 0, mmIH_CNTL2);
@@ -330,7 +330,7 @@ static int navi10_ih_irq_init(struct amdgpu_device *adev)
 
 	if (unlikely(adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT)) {
 		if (ih[0]->use_bus_addr) {
-			switch (adev->ip_versions[OSSSYS_HWIP][0]) {
+			switch (amdgpu_ip_version(adev, OSSSYS_HWIP, 0)) {
 			case IP_VERSION(5, 0, 3):
 			case IP_VERSION(5, 2, 0):
 			case IP_VERSION(5, 2, 1):
@@ -409,9 +409,11 @@ static u32 navi10_ih_get_wptr(struct amdgpu_device *adev,
 	u32 wptr, tmp;
 	struct amdgpu_ih_regs *ih_regs;
 
-	if (ih == &adev->irq.ih) {
+	if (ih == &adev->irq.ih || ih == &adev->irq.ih_soft) {
 		/* Only ring0 supports writeback. On other rings fall back
 		 * to register-based code with overflow checking below.
+		 * ih_soft ring doesn't have any backing hardware registers,
+		 * update wptr and return.
 		 */
 		wptr = le32_to_cpu(*ih->wptr_cpu);
 
@@ -432,14 +434,19 @@ static u32 navi10_ih_get_wptr(struct amdgpu_device *adev,
 	 * this should allow us to catch up.
 	 */
 	tmp = (wptr + 32) & ih->ptr_mask;
-	dev_warn(adev->dev, "IH ring buffer overflow "
-		 "(0x%08X, 0x%08X, 0x%08X)\n",
-		 wptr, ih->rptr, tmp);
+	dev_warn(adev->dev, "%s ring buffer overflow (0x%08X, 0x%08X, 0x%08X)\n",
+		 amdgpu_ih_ring_name(adev, ih), wptr, ih->rptr, tmp);
 	ih->rptr = tmp;
 
 	tmp = RREG32_NO_KIQ(ih_regs->ih_rb_cntl);
 	tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 1);
 	WREG32_NO_KIQ(ih_regs->ih_rb_cntl, tmp);
+
+	/* Unset the CLEAR_OVERFLOW bit immediately so new overflows
+	 * can be detected.
+	 */
+	tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 0);
+	WREG32_NO_KIQ(ih_regs->ih_rb_cntl, tmp);
 out:
 	return (wptr & ih->ptr_mask);
 }
@@ -483,6 +490,9 @@ static void navi10_ih_set_rptr(struct amdgpu_device *adev,
 {
 	struct amdgpu_ih_regs *ih_regs;
 
+	if (ih == &adev->irq.ih_soft)
+		return;
+
 	if (ih->use_doorbell) {
 		/* XXX check if swapping is necessary on BE */
 		*ih->rptr_cpu = ih->rptr;
@@ -531,19 +541,19 @@ static void navi10_ih_set_self_irq_funcs(struct amdgpu_device *adev)
 	adev->irq.self_irq.funcs = &navi10_ih_self_irq_funcs;
 }
 
-static int navi10_ih_early_init(void *handle)
+static int navi10_ih_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	navi10_ih_set_interrupt_funcs(adev);
 	navi10_ih_set_self_irq_funcs(adev);
 	return 0;
 }
 
-static int navi10_ih_sw_init(void *handle)
+static int navi10_ih_sw_init(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	bool use_bus_addr;
 
 	r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_IH, 0,
@@ -560,7 +570,7 @@ static int navi10_ih_sw_init(void *handle)
 		use_bus_addr = false;
 	else
 		use_bus_addr = true;
-	r = amdgpu_ih_ring_init(adev, &adev->irq.ih, 256 * 1024, use_bus_addr);
+	r = amdgpu_ih_ring_init(adev, &adev->irq.ih, IH_RING_SIZE, use_bus_addr);
 	if (r)
 		return r;
 
@@ -573,7 +583,7 @@ static int navi10_ih_sw_init(void *handle)
 	/* initialize ih control registers offset */
 	navi10_ih_init_register_offset(adev);
 
-	r = amdgpu_ih_ring_init(adev, &adev->irq.ih_soft, PAGE_SIZE, true);
+	r = amdgpu_ih_ring_init(adev, &adev->irq.ih_soft, IH_SW_RING_SIZE, true);
 	if (r)
 		return r;
 
@@ -582,63 +592,52 @@ static int navi10_ih_sw_init(void *handle)
 	return r;
 }
 
-static int navi10_ih_sw_fini(void *handle)
+static int navi10_ih_sw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	amdgpu_irq_fini_sw(adev);
 
 	return 0;
 }
 
-static int navi10_ih_hw_init(void *handle)
+static int navi10_ih_hw_init(struct amdgpu_ip_block *ip_block)
 {
-	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
-	r = navi10_ih_irq_init(adev);
-	if (r)
-		return r;
-
-	return 0;
+	return navi10_ih_irq_init(adev);
 }
 
-static int navi10_ih_hw_fini(void *handle)
+static int navi10_ih_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	navi10_ih_irq_disable(adev);
+	navi10_ih_irq_disable(ip_block->adev);
 
 	return 0;
 }
 
-static int navi10_ih_suspend(void *handle)
+static int navi10_ih_suspend(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	return navi10_ih_hw_fini(adev);
+	return navi10_ih_hw_fini(ip_block);
 }
 
-static int navi10_ih_resume(void *handle)
+static int navi10_ih_resume(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	return navi10_ih_hw_init(adev);
+	return navi10_ih_hw_init(ip_block);
 }
 
-static bool navi10_ih_is_idle(void *handle)
+static bool navi10_ih_is_idle(struct amdgpu_ip_block *ip_block)
 {
 	/* todo */
 	return true;
 }
 
-static int navi10_ih_wait_for_idle(void *handle)
+static int navi10_ih_wait_for_idle(struct amdgpu_ip_block *ip_block)
 {
 	/* todo */
 	return -ETIMEDOUT;
 }
 
-static int navi10_ih_soft_reset(void *handle)
+static int navi10_ih_soft_reset(struct amdgpu_ip_block *ip_block)
 {
 	/* todo */
 	return 0;
@@ -665,40 +664,35 @@ static void navi10_ih_update_clockgating_state(struct amdgpu_device *adev,
 		if (def != data)
 			WREG32_SOC15(OSSSYS, 0, mmIH_CLK_CTRL, data);
 	}
-
-	return;
 }
 
-static int navi10_ih_set_clockgating_state(void *handle,
+static int navi10_ih_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					   enum amd_clockgating_state state)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	navi10_ih_update_clockgating_state(adev,
 				state == AMD_CG_STATE_GATE);
 	return 0;
 }
 
-static int navi10_ih_set_powergating_state(void *handle,
+static int navi10_ih_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					   enum amd_powergating_state state)
 {
 	return 0;
 }
 
-static void navi10_ih_get_clockgating_state(void *handle, u32 *flags)
+static void navi10_ih_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (!RREG32_SOC15(OSSSYS, 0, mmIH_CLK_CTRL))
 		*flags |= AMD_CG_SUPPORT_IH_CG;
-
-	return;
 }
 
 static const struct amd_ip_funcs navi10_ih_ip_funcs = {
 	.name = "navi10_ih",
 	.early_init = navi10_ih_early_init,
-	.late_init = NULL,
 	.sw_init = navi10_ih_sw_init,
 	.sw_fini = navi10_ih_sw_fini,
 	.hw_init = navi10_ih_hw_init,
@@ -716,6 +710,7 @@ static const struct amd_ip_funcs navi10_ih_ip_funcs = {
 static const struct amdgpu_ih_funcs navi10_ih_funcs = {
 	.get_wptr = navi10_ih_get_wptr,
 	.decode_iv = amdgpu_ih_decode_iv_helper,
+	.decode_iv_ts = amdgpu_ih_decode_iv_ts_helper,
 	.set_rptr = navi10_ih_set_rptr
 };
 
@@ -725,8 +720,7 @@ static void navi10_ih_set_interrupt_funcs(struct amdgpu_device *adev)
 		adev->irq.ih_funcs = &navi10_ih_funcs;
 }
 
-const struct amdgpu_ip_block_version navi10_ih_ip_block =
-{
+const struct amdgpu_ip_block_version navi10_ih_ip_block = {
 	.type = AMD_IP_BLOCK_TYPE_IH,
 	.major = 5,
 	.minor = 0,
diff --git a/drivers/gpu/drm/amd/amdgpu/navi10_sdma_pkt_open.h b/drivers/gpu/drm/amd/amdgpu/navi10_sdma_pkt_open.h
index a5b60c9a2418..c88284ff92d8 100644
--- a/drivers/gpu/drm/amd/amdgpu/navi10_sdma_pkt_open.h
+++ b/drivers/gpu/drm/amd/amdgpu/navi10_sdma_pkt_open.h
@@ -68,6 +68,7 @@
 #define SDMA_SUBOP_POLL_REG_WRITE_MEM  1
 #define SDMA_SUBOP_POLL_DBIT_WRITE_MEM  2
 #define SDMA_SUBOP_POLL_MEM_VERIFY  3
+#define SDMA_SUBOP_VM_INVALIDATION  4
 #define HEADER_AGENT_DISPATCH  4
 #define HEADER_BARRIER  5
 #define SDMA_OP_AQL_COPY  0
@@ -4041,6 +4042,69 @@
 
 
 /*
+** Definitions for SDMA_PKT_VM_INVALIDATION packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_VM_INVALIDATION_HEADER_op_offset 0
+#define SDMA_PKT_VM_INVALIDATION_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_VM_INVALIDATION_HEADER_op_shift  0
+#define SDMA_PKT_VM_INVALIDATION_HEADER_OP(x) (((x) & SDMA_PKT_VM_INVALIDATION_HEADER_op_mask) << SDMA_PKT_VM_INVALIDATION_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_VM_INVALIDATION_HEADER_sub_op_offset 0
+#define SDMA_PKT_VM_INVALIDATION_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_VM_INVALIDATION_HEADER_sub_op_shift  8
+#define SDMA_PKT_VM_INVALIDATION_HEADER_SUB_OP(x) (((x) & SDMA_PKT_VM_INVALIDATION_HEADER_sub_op_mask) << SDMA_PKT_VM_INVALIDATION_HEADER_sub_op_shift)
+
+/*define for gfx_eng_id field*/
+#define SDMA_PKT_VM_INVALIDATION_HEADER_gfx_eng_id_offset 0
+#define SDMA_PKT_VM_INVALIDATION_HEADER_gfx_eng_id_mask   0x0000001F
+#define SDMA_PKT_VM_INVALIDATION_HEADER_gfx_eng_id_shift  16
+#define SDMA_PKT_VM_INVALIDATION_HEADER_GFX_ENG_ID(x) (((x) & SDMA_PKT_VM_INVALIDATION_HEADER_gfx_eng_id_mask) << SDMA_PKT_VM_INVALIDATION_HEADER_gfx_eng_id_shift)
+
+/*define for mm_eng_id field*/
+#define SDMA_PKT_VM_INVALIDATION_HEADER_mm_eng_id_offset 0
+#define SDMA_PKT_VM_INVALIDATION_HEADER_mm_eng_id_mask   0x0000001F
+#define SDMA_PKT_VM_INVALIDATION_HEADER_mm_eng_id_shift  24
+#define SDMA_PKT_VM_INVALIDATION_HEADER_MM_ENG_ID(x) (((x) & SDMA_PKT_VM_INVALIDATION_HEADER_mm_eng_id_mask) << SDMA_PKT_VM_INVALIDATION_HEADER_mm_eng_id_shift)
+
+/*define for INVALIDATEREQ word*/
+/*define for invalidatereq field*/
+#define SDMA_PKT_VM_INVALIDATION_INVALIDATEREQ_invalidatereq_offset 1
+#define SDMA_PKT_VM_INVALIDATION_INVALIDATEREQ_invalidatereq_mask   0xFFFFFFFF
+#define SDMA_PKT_VM_INVALIDATION_INVALIDATEREQ_invalidatereq_shift  0
+#define SDMA_PKT_VM_INVALIDATION_INVALIDATEREQ_INVALIDATEREQ(x) (((x) & SDMA_PKT_VM_INVALIDATION_INVALIDATEREQ_invalidatereq_mask) << SDMA_PKT_VM_INVALIDATION_INVALIDATEREQ_invalidatereq_shift)
+
+/*define for ADDRESSRANGELO word*/
+/*define for addressrangelo field*/
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGELO_addressrangelo_offset 2
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGELO_addressrangelo_mask   0xFFFFFFFF
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGELO_addressrangelo_shift  0
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGELO_ADDRESSRANGELO(x) (((x) & SDMA_PKT_VM_INVALIDATION_ADDRESSRANGELO_addressrangelo_mask) << SDMA_PKT_VM_INVALIDATION_ADDRESSRANGELO_addressrangelo_shift)
+
+/*define for ADDRESSRANGEHI word*/
+/*define for invalidateack field*/
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_invalidateack_offset 3
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_invalidateack_mask   0x0000FFFF
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_invalidateack_shift  0
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_INVALIDATEACK(x) (((x) & SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_invalidateack_mask) << SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_invalidateack_shift)
+
+/*define for addressrangehi field*/
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_addressrangehi_offset 3
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_addressrangehi_mask   0x0000001F
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_addressrangehi_shift  16
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_ADDRESSRANGEHI(x) (((x) & SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_addressrangehi_mask) << SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_addressrangehi_shift)
+
+/*define for reserved field*/
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_reserved_offset 3
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_reserved_mask   0x000001FF
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_reserved_shift  23
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_RESERVED(x) (((x) & SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_reserved_mask) << SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_reserved_shift)
+
+
+/*
 ** Definitions for SDMA_PKT_ATOMIC packet
 */
 
diff --git a/drivers/gpu/drm/amd/amdgpu/nbif_v6_3_1.c b/drivers/gpu/drm/amd/amdgpu/nbif_v6_3_1.c
new file mode 100644
index 000000000000..9b4025c39e44
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/nbif_v6_3_1.c
@@ -0,0 +1,554 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include "amdgpu.h"
+#include "nbif_v6_3_1.h"
+
+#include "nbif/nbif_6_3_1_offset.h"
+#include "nbif/nbif_6_3_1_sh_mask.h"
+#include "pcie/pcie_6_1_0_offset.h"
+#include "pcie/pcie_6_1_0_sh_mask.h"
+#include "ivsrcid/nbio/irqsrcs_nbif_7_4.h"
+#include <uapi/linux/kfd_ioctl.h>
+
+static void nbif_v6_3_1_remap_hdp_registers(struct amdgpu_device *adev)
+{
+	WREG32_SOC15(NBIO, 0, regBIF_BX0_REMAP_HDP_MEM_FLUSH_CNTL,
+		adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL);
+	WREG32_SOC15(NBIO, 0, regBIF_BX0_REMAP_HDP_REG_FLUSH_CNTL,
+		adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_REG_FLUSH_CNTL);
+}
+
+static u32 nbif_v6_3_1_get_rev_id(struct amdgpu_device *adev)
+{
+	u32 tmp = RREG32_SOC15(NBIO, 0, regRCC_STRAP0_RCC_DEV0_EPF0_STRAP0);
+
+	tmp &= RCC_STRAP0_RCC_DEV0_EPF0_STRAP0__STRAP_ATI_REV_ID_DEV0_F0_MASK;
+	tmp >>= RCC_STRAP0_RCC_DEV0_EPF0_STRAP0__STRAP_ATI_REV_ID_DEV0_F0__SHIFT;
+
+	return tmp;
+}
+
+static void nbif_v6_3_1_mc_access_enable(struct amdgpu_device *adev, bool enable)
+{
+	if (enable)
+		WREG32_SOC15(NBIO, 0, regBIF_BX0_BIF_FB_EN,
+			     BIF_BX0_BIF_FB_EN__FB_READ_EN_MASK |
+			     BIF_BX0_BIF_FB_EN__FB_WRITE_EN_MASK);
+	else
+		WREG32_SOC15(NBIO, 0, regBIF_BX0_BIF_FB_EN, 0);
+}
+
+static u32 nbif_v6_3_1_get_memsize(struct amdgpu_device *adev)
+{
+	return RREG32_SOC15(NBIO, 0, regRCC_DEV0_EPF0_RCC_CONFIG_MEMSIZE);
+}
+
+static void nbif_v6_3_1_sdma_doorbell_range(struct amdgpu_device *adev,
+					    int instance, bool use_doorbell,
+					    int doorbell_index,
+					    int doorbell_size)
+{
+	if (instance == 0) {
+		u32 doorbell_range = RREG32_SOC15(NBIO, 0, regGDC_S2A0_S2A_DOORBELL_ENTRY_2_CTRL);
+
+		if (use_doorbell) {
+			doorbell_range = REG_SET_FIELD(doorbell_range,
+						       GDC_S2A0_S2A_DOORBELL_ENTRY_2_CTRL,
+						       S2A_DOORBELL_PORT2_ENABLE,
+						       0x1);
+			doorbell_range = REG_SET_FIELD(doorbell_range,
+						       GDC_S2A0_S2A_DOORBELL_ENTRY_2_CTRL,
+						       S2A_DOORBELL_PORT2_AWID,
+						       0xe);
+			doorbell_range = REG_SET_FIELD(doorbell_range,
+						       GDC_S2A0_S2A_DOORBELL_ENTRY_2_CTRL,
+						       S2A_DOORBELL_PORT2_RANGE_OFFSET,
+						       doorbell_index);
+			doorbell_range = REG_SET_FIELD(doorbell_range,
+						       GDC_S2A0_S2A_DOORBELL_ENTRY_2_CTRL,
+						       S2A_DOORBELL_PORT2_RANGE_SIZE,
+						       doorbell_size);
+			doorbell_range = REG_SET_FIELD(doorbell_range,
+						       GDC_S2A0_S2A_DOORBELL_ENTRY_2_CTRL,
+						       S2A_DOORBELL_PORT2_AWADDR_31_28_VALUE,
+						       0x3);
+		} else
+			doorbell_range = REG_SET_FIELD(doorbell_range,
+						       GDC_S2A0_S2A_DOORBELL_ENTRY_2_CTRL,
+						       S2A_DOORBELL_PORT2_RANGE_SIZE,
+						       0);
+
+		WREG32_SOC15(NBIO, 0, regGDC_S2A0_S2A_DOORBELL_ENTRY_2_CTRL, doorbell_range);
+	}
+}
+
+static void nbif_v6_3_1_vcn_doorbell_range(struct amdgpu_device *adev,
+					   bool use_doorbell, int doorbell_index,
+					   int instance)
+{
+	u32 doorbell_range;
+
+	if (instance)
+		doorbell_range = RREG32_SOC15(NBIO, 0, regGDC_S2A0_S2A_DOORBELL_ENTRY_5_CTRL);
+	else
+		doorbell_range = RREG32_SOC15(NBIO, 0, regGDC_S2A0_S2A_DOORBELL_ENTRY_4_CTRL);
+
+	if (use_doorbell) {
+		doorbell_range = REG_SET_FIELD(doorbell_range,
+					       GDC_S2A0_S2A_DOORBELL_ENTRY_4_CTRL,
+					       S2A_DOORBELL_PORT4_ENABLE,
+					       0x1);
+		doorbell_range = REG_SET_FIELD(doorbell_range,
+					       GDC_S2A0_S2A_DOORBELL_ENTRY_4_CTRL,
+					       S2A_DOORBELL_PORT4_AWID,
+					       instance ? 0x7 : 0x4);
+		doorbell_range = REG_SET_FIELD(doorbell_range,
+					       GDC_S2A0_S2A_DOORBELL_ENTRY_4_CTRL,
+					       S2A_DOORBELL_PORT4_RANGE_OFFSET,
+					       doorbell_index);
+		doorbell_range = REG_SET_FIELD(doorbell_range,
+					       GDC_S2A0_S2A_DOORBELL_ENTRY_4_CTRL,
+					       S2A_DOORBELL_PORT4_RANGE_SIZE,
+					       8);
+		doorbell_range = REG_SET_FIELD(doorbell_range,
+					       GDC_S2A0_S2A_DOORBELL_ENTRY_4_CTRL,
+					       S2A_DOORBELL_PORT4_AWADDR_31_28_VALUE,
+					       instance ? 0x7 : 0x4);
+	} else
+		doorbell_range = REG_SET_FIELD(doorbell_range,
+					       GDC_S2A0_S2A_DOORBELL_ENTRY_4_CTRL,
+					       S2A_DOORBELL_PORT4_RANGE_SIZE,
+					       0);
+
+	if (instance)
+		WREG32_SOC15(NBIO, 0, regGDC_S2A0_S2A_DOORBELL_ENTRY_5_CTRL, doorbell_range);
+	else
+		WREG32_SOC15(NBIO, 0, regGDC_S2A0_S2A_DOORBELL_ENTRY_4_CTRL, doorbell_range);
+}
+
+static void nbif_v6_3_1_gc_doorbell_init(struct amdgpu_device *adev)
+{
+	WREG32_SOC15(NBIO, 0, regGDC_S2A0_S2A_DOORBELL_ENTRY_0_CTRL, 0x30000007);
+	WREG32_SOC15(NBIO, 0, regGDC_S2A0_S2A_DOORBELL_ENTRY_3_CTRL, 0x3000000d);
+}
+
+static void nbif_v6_3_1_enable_doorbell_aperture(struct amdgpu_device *adev,
+						 bool enable)
+{
+	WREG32_FIELD15_PREREG(NBIO, 0, RCC_DEV0_EPF0_RCC_DOORBELL_APER_EN,
+			BIF_DOORBELL_APER_EN, enable ? 1 : 0);
+}
+
+static void
+nbif_v6_3_1_enable_doorbell_selfring_aperture(struct amdgpu_device *adev,
+					      bool enable)
+{
+	u32 tmp = 0;
+
+	if (enable) {
+		tmp = REG_SET_FIELD(tmp, BIF_BX_PF0_DOORBELL_SELFRING_GPA_APER_CNTL,
+				    DOORBELL_SELFRING_GPA_APER_EN, 1) |
+		      REG_SET_FIELD(tmp, BIF_BX_PF0_DOORBELL_SELFRING_GPA_APER_CNTL,
+				    DOORBELL_SELFRING_GPA_APER_MODE, 1) |
+		      REG_SET_FIELD(tmp, BIF_BX_PF0_DOORBELL_SELFRING_GPA_APER_CNTL,
+				    DOORBELL_SELFRING_GPA_APER_SIZE, 0);
+
+		WREG32_SOC15(NBIO, 0, regBIF_BX_PF0_DOORBELL_SELFRING_GPA_APER_BASE_LOW,
+			     lower_32_bits(adev->doorbell.base));
+		WREG32_SOC15(NBIO, 0, regBIF_BX_PF0_DOORBELL_SELFRING_GPA_APER_BASE_HIGH,
+			     upper_32_bits(adev->doorbell.base));
+	}
+
+	WREG32_SOC15(NBIO, 0, regBIF_BX_PF0_DOORBELL_SELFRING_GPA_APER_CNTL, tmp);
+}
+
+static void nbif_v6_3_1_ih_doorbell_range(struct amdgpu_device *adev,
+					  bool use_doorbell, int doorbell_index)
+{
+	u32 ih_doorbell_range = RREG32_SOC15(NBIO, 0, regGDC_S2A0_S2A_DOORBELL_ENTRY_1_CTRL);
+
+	if (use_doorbell) {
+		ih_doorbell_range = REG_SET_FIELD(ih_doorbell_range,
+						  GDC_S2A0_S2A_DOORBELL_ENTRY_1_CTRL,
+						  S2A_DOORBELL_PORT1_ENABLE,
+						  0x1);
+		ih_doorbell_range = REG_SET_FIELD(ih_doorbell_range,
+						  GDC_S2A0_S2A_DOORBELL_ENTRY_1_CTRL,
+						  S2A_DOORBELL_PORT1_AWID,
+						  0x0);
+		ih_doorbell_range = REG_SET_FIELD(ih_doorbell_range,
+						  GDC_S2A0_S2A_DOORBELL_ENTRY_1_CTRL,
+						  S2A_DOORBELL_PORT1_RANGE_OFFSET,
+						  doorbell_index);
+		ih_doorbell_range = REG_SET_FIELD(ih_doorbell_range,
+						  GDC_S2A0_S2A_DOORBELL_ENTRY_1_CTRL,
+						  S2A_DOORBELL_PORT1_RANGE_SIZE,
+						  2);
+		ih_doorbell_range = REG_SET_FIELD(ih_doorbell_range,
+						  GDC_S2A0_S2A_DOORBELL_ENTRY_1_CTRL,
+						  S2A_DOORBELL_PORT1_AWADDR_31_28_VALUE,
+						  0x0);
+	} else
+		ih_doorbell_range = REG_SET_FIELD(ih_doorbell_range,
+						  GDC_S2A0_S2A_DOORBELL_ENTRY_1_CTRL,
+						  S2A_DOORBELL_PORT1_RANGE_SIZE,
+						  0);
+
+	WREG32_SOC15(NBIO, 0, regGDC_S2A0_S2A_DOORBELL_ENTRY_1_CTRL, ih_doorbell_range);
+}
+
+static void nbif_v6_3_1_ih_control(struct amdgpu_device *adev)
+{
+	u32 interrupt_cntl;
+
+	/* setup interrupt control */
+	WREG32_SOC15(NBIO, 0, regBIF_BX0_INTERRUPT_CNTL2, adev->dummy_page_addr >> 8);
+
+	interrupt_cntl = RREG32_SOC15(NBIO, 0, regBIF_BX0_INTERRUPT_CNTL);
+	/*
+	 * BIF_BX0_INTERRUPT_CNTL__IH_DUMMY_RD_OVERRIDE_MASK=0 - dummy read disabled with msi, enabled without msi
+	 * BIF_BX0_INTERRUPT_CNTL__IH_DUMMY_RD_OVERRIDE_MASK=1 - dummy read controlled by IH_DUMMY_RD_EN
+	 */
+	interrupt_cntl = REG_SET_FIELD(interrupt_cntl, BIF_BX0_INTERRUPT_CNTL,
+				       IH_DUMMY_RD_OVERRIDE, 0);
+
+	/* BIF_BX0_INTERRUPT_CNTL__IH_REQ_NONSNOOP_EN_MASK=1 if ring is in non-cacheable memory, e.g., vram */
+	interrupt_cntl = REG_SET_FIELD(interrupt_cntl, BIF_BX0_INTERRUPT_CNTL,
+				       IH_REQ_NONSNOOP_EN, 0);
+
+	WREG32_SOC15(NBIO, 0, regBIF_BX0_INTERRUPT_CNTL, interrupt_cntl);
+}
+
+static void
+nbif_v6_3_1_update_medium_grain_clock_gating(struct amdgpu_device *adev,
+					     bool enable)
+{
+}
+
+static void
+nbif_v6_3_1_update_medium_grain_light_sleep(struct amdgpu_device *adev,
+					    bool enable)
+{
+}
+
+static void
+nbif_v6_3_1_get_clockgating_state(struct amdgpu_device *adev,
+				  u64 *flags)
+{
+}
+
+static u32 nbif_v6_3_1_get_hdp_flush_req_offset(struct amdgpu_device *adev)
+{
+	return SOC15_REG_OFFSET(NBIO, 0, regBIF_BX_PF0_GPU_HDP_FLUSH_REQ);
+}
+
+static u32 nbif_v6_3_1_get_hdp_flush_done_offset(struct amdgpu_device *adev)
+{
+	return SOC15_REG_OFFSET(NBIO, 0, regBIF_BX_PF0_GPU_HDP_FLUSH_DONE);
+}
+
+static u32 nbif_v6_3_1_get_pcie_index_offset(struct amdgpu_device *adev)
+{
+	return SOC15_REG_OFFSET(NBIO, 0, regBIF_BX_PF0_RSMU_INDEX);
+}
+
+static u32 nbif_v6_3_1_get_pcie_data_offset(struct amdgpu_device *adev)
+{
+	return SOC15_REG_OFFSET(NBIO, 0, regBIF_BX_PF0_RSMU_DATA);
+}
+
+const struct nbio_hdp_flush_reg nbif_v6_3_1_hdp_flush_reg = {
+	.ref_and_mask_cp0 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__CP0_MASK,
+	.ref_and_mask_cp1 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__CP1_MASK,
+	.ref_and_mask_cp2 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__CP2_MASK,
+	.ref_and_mask_cp3 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__CP3_MASK,
+	.ref_and_mask_cp4 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__CP4_MASK,
+	.ref_and_mask_cp5 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__CP5_MASK,
+	.ref_and_mask_cp6 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__CP6_MASK,
+	.ref_and_mask_cp7 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__CP7_MASK,
+	.ref_and_mask_cp8 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__CP8_MASK,
+	.ref_and_mask_cp9 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__CP9_MASK,
+	.ref_and_mask_sdma0 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__SDMA0_MASK,
+	.ref_and_mask_sdma1 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__SDMA1_MASK,
+};
+
+static void nbif_v6_3_1_init_registers(struct amdgpu_device *adev)
+{
+	uint32_t data;
+
+	data = RREG32_SOC15(NBIO, 0, regRCC_DEV0_EPF2_STRAP2);
+	data &= ~RCC_DEV0_EPF2_STRAP2__STRAP_NO_SOFT_RESET_DEV0_F2_MASK;
+	WREG32_SOC15(NBIO, 0, regRCC_DEV0_EPF2_STRAP2, data);
+}
+
+static u32 nbif_v6_3_1_get_rom_offset(struct amdgpu_device *adev)
+{
+	u32 data, rom_offset;
+
+	data = RREG32_SOC15(NBIO, 0, regREGS_ROM_OFFSET_CTRL);
+	rom_offset = REG_GET_FIELD(data, REGS_ROM_OFFSET_CTRL, ROM_OFFSET);
+
+	return rom_offset;
+}
+
+#ifdef CONFIG_PCIEASPM
+static void nbif_v6_3_1_program_ltr(struct amdgpu_device *adev)
+{
+	uint32_t def, data;
+	u16 devctl2;
+
+	def = RREG32_SOC15(NBIO, 0, regRCC_EP_DEV0_0_EP_PCIE_TX_LTR_CNTL);
+	data = 0x35EB;
+	data &= ~RCC_EP_DEV0_0_EP_PCIE_TX_LTR_CNTL__LTR_PRIV_MSG_DIS_IN_PM_NON_D0_MASK;
+	data &= ~RCC_EP_DEV0_0_EP_PCIE_TX_LTR_CNTL__LTR_PRIV_RST_LTR_IN_DL_DOWN_MASK;
+	if (def != data)
+		WREG32_SOC15(NBIO, 0, regRCC_EP_DEV0_0_EP_PCIE_TX_LTR_CNTL, data);
+
+	def = data = RREG32_SOC15(NBIO, 0, regRCC_STRAP0_RCC_BIF_STRAP2);
+	data &= ~RCC_STRAP0_RCC_BIF_STRAP2__STRAP_LTR_IN_ASPML1_DIS_MASK;
+	if (def != data)
+		WREG32_SOC15(NBIO, 0, regRCC_STRAP0_RCC_BIF_STRAP2, data);
+
+	pcie_capability_read_word(adev->pdev, PCI_EXP_DEVCTL2, &devctl2);
+
+	if (adev->pdev->ltr_path == (devctl2 & PCI_EXP_DEVCTL2_LTR_EN))
+		return;
+
+	if (adev->pdev->ltr_path)
+		pcie_capability_set_word(adev->pdev, PCI_EXP_DEVCTL2, PCI_EXP_DEVCTL2_LTR_EN);
+	else
+		pcie_capability_clear_word(adev->pdev, PCI_EXP_DEVCTL2, PCI_EXP_DEVCTL2_LTR_EN);
+}
+#endif
+
+static void nbif_v6_3_1_program_aspm(struct amdgpu_device *adev)
+{
+#ifdef CONFIG_PCIEASPM
+	uint32_t def, data;
+	u16 devctl2, ltr;
+
+	def = data = RREG32_SOC15(PCIE, 0, regPCIE_LC_CNTL);
+	data &= ~PCIE_LC_CNTL__LC_L1_INACTIVITY_MASK;
+	data &= ~PCIE_LC_CNTL__LC_L0S_INACTIVITY_MASK;
+	data |= PCIE_LC_CNTL__LC_PMI_TO_L1_DIS_MASK;
+	if (def != data)
+		WREG32_SOC15(PCIE, 0, regPCIE_LC_CNTL, data);
+
+	def = data = RREG32_SOC15(PCIE, 0, regPCIE_LC_CNTL7);
+	data |= PCIE_LC_CNTL7__LC_NBIF_ASPM_INPUT_EN_MASK;
+	if (def != data)
+		WREG32_SOC15(PCIE, 0, regPCIE_LC_CNTL7, data);
+
+	def = data = RREG32_SOC15(PCIE, 0, regPCIE_LC_CNTL3);
+	data |= PCIE_LC_CNTL3__LC_DSC_DONT_ENTER_L23_AFTER_PME_ACK_MASK;
+	if (def != data)
+		WREG32_SOC15(PCIE, 0, regPCIE_LC_CNTL3, data);
+
+	def = data = RREG32_SOC15(NBIO, 0, regRCC_STRAP0_RCC_BIF_STRAP3);
+	data &= ~RCC_STRAP0_RCC_BIF_STRAP3__STRAP_VLINK_ASPM_IDLE_TIMER_MASK;
+	data &= ~RCC_STRAP0_RCC_BIF_STRAP3__STRAP_VLINK_PM_L1_ENTRY_TIMER_MASK;
+	if (def != data)
+		WREG32_SOC15(NBIO, 0, regRCC_STRAP0_RCC_BIF_STRAP3, data);
+
+	def = data = RREG32_SOC15(NBIO, 0, regRCC_STRAP0_RCC_BIF_STRAP5);
+	data &= ~RCC_STRAP0_RCC_BIF_STRAP5__STRAP_VLINK_LDN_ENTRY_TIMER_MASK;
+	if (def != data)
+		WREG32_SOC15(NBIO, 0, regRCC_STRAP0_RCC_BIF_STRAP5, data);
+
+	pcie_capability_read_word(adev->pdev, PCI_EXP_DEVCTL2, &devctl2);
+	data = def = devctl2;
+	data &= ~PCI_EXP_DEVCTL2_LTR_EN;
+	if (def != data)
+		pcie_capability_set_word(adev->pdev, PCI_EXP_DEVCTL2, (u16)data);
+
+	ltr = pci_find_ext_capability(adev->pdev, PCI_EXT_CAP_ID_LTR);
+
+	if (ltr) {
+		pci_write_config_dword(adev->pdev, ltr + PCI_LTR_MAX_SNOOP_LAT, 0x10011001);
+	}
+
+#if 0
+	/* regPSWUSP0_PCIE_LC_CNTL2 should be replace by PCIE_LC_CNTL2 or someone else ? */
+	def = data = RREG32_SOC15(NBIO, 0, regPSWUSP0_PCIE_LC_CNTL2);
+	data |= PSWUSP0_PCIE_LC_CNTL2__LC_ALLOW_PDWN_IN_L1_MASK |
+		PSWUSP0_PCIE_LC_CNTL2__LC_ALLOW_PDWN_IN_L23_MASK;
+	data &= ~PSWUSP0_PCIE_LC_CNTL2__LC_RCV_L0_TO_RCV_L0S_DIS_MASK;
+	if (def != data)
+		WREG32_SOC15(NBIO, 0, regPSWUSP0_PCIE_LC_CNTL2, data);
+#endif
+	def = data = RREG32_SOC15(PCIE, 0, regPCIE_LC_CNTL4);
+	data |= PCIE_LC_CNTL4__LC_L1_POWERDOWN_MASK;
+	if (def != data)
+		WREG32_SOC15(PCIE, 0, regPCIE_LC_CNTL4, data);
+
+	def = data = RREG32_SOC15(PCIE, 0, regPCIE_LC_RXRECOVER_RXSTANDBY_CNTL);
+	data |= PCIE_LC_RXRECOVER_RXSTANDBY_CNTL__LC_RX_L0S_STANDBY_EN_MASK;
+	if (def != data)
+		WREG32_SOC15(PCIE, 0, regPCIE_LC_RXRECOVER_RXSTANDBY_CNTL, data);
+
+	nbif_v6_3_1_program_ltr(adev);
+
+	def = data = RREG32_SOC15(NBIO, 0, regRCC_STRAP0_RCC_BIF_STRAP3);
+	data |= 0x5DE0 << RCC_STRAP0_RCC_BIF_STRAP3__STRAP_VLINK_ASPM_IDLE_TIMER__SHIFT;
+	data |= 0x0010 << RCC_STRAP0_RCC_BIF_STRAP3__STRAP_VLINK_PM_L1_ENTRY_TIMER__SHIFT;
+	if (def != data)
+		WREG32_SOC15(NBIO, 0, regRCC_STRAP0_RCC_BIF_STRAP3, data);
+
+	def = data = RREG32_SOC15(NBIO, 0, regRCC_STRAP0_RCC_BIF_STRAP5);
+	data |= 0x0010 << RCC_STRAP0_RCC_BIF_STRAP5__STRAP_VLINK_LDN_ENTRY_TIMER__SHIFT;
+	if (def != data)
+		WREG32_SOC15(NBIO, 0, regRCC_STRAP0_RCC_BIF_STRAP5, data);
+
+	def = data = RREG32_SOC15(PCIE, 0, regPCIE_LC_CNTL);
+	data |= 0x0 << PCIE_LC_CNTL__LC_L0S_INACTIVITY__SHIFT;
+	data |= 0x9 << PCIE_LC_CNTL__LC_L1_INACTIVITY__SHIFT;
+	data &= ~PCIE_LC_CNTL__LC_PMI_TO_L1_DIS_MASK;
+	if (def != data)
+		WREG32_SOC15(PCIE, 0, regPCIE_LC_CNTL, data);
+
+	def = data = RREG32_SOC15(PCIE, 0, regPCIE_LC_CNTL3);
+	data &= ~PCIE_LC_CNTL3__LC_DSC_DONT_ENTER_L23_AFTER_PME_ACK_MASK;
+	if (def != data)
+		WREG32_SOC15(PCIE, 0, regPCIE_LC_CNTL3, data);
+#endif
+}
+
+#define MMIO_REG_HOLE_OFFSET (0x80000 - PAGE_SIZE)
+
+static void nbif_v6_3_1_set_reg_remap(struct amdgpu_device *adev)
+{
+	if (!amdgpu_sriov_vf(adev) && (PAGE_SIZE <= 4096)) {
+		adev->rmmio_remap.reg_offset = MMIO_REG_HOLE_OFFSET;
+		adev->rmmio_remap.bus_addr = adev->rmmio_base + MMIO_REG_HOLE_OFFSET;
+	} else {
+		adev->rmmio_remap.reg_offset = SOC15_REG_OFFSET(NBIO, 0,
+			regBIF_BX_PF0_HDP_MEM_COHERENCY_FLUSH_CNTL) << 2;
+		adev->rmmio_remap.bus_addr = 0;
+	}
+}
+
+const struct amdgpu_nbio_funcs nbif_v6_3_1_funcs = {
+	.get_hdp_flush_req_offset = nbif_v6_3_1_get_hdp_flush_req_offset,
+	.get_hdp_flush_done_offset = nbif_v6_3_1_get_hdp_flush_done_offset,
+	.get_pcie_index_offset = nbif_v6_3_1_get_pcie_index_offset,
+	.get_pcie_data_offset = nbif_v6_3_1_get_pcie_data_offset,
+	.get_rev_id = nbif_v6_3_1_get_rev_id,
+	.mc_access_enable = nbif_v6_3_1_mc_access_enable,
+	.get_memsize = nbif_v6_3_1_get_memsize,
+	.sdma_doorbell_range = nbif_v6_3_1_sdma_doorbell_range,
+	.vcn_doorbell_range = nbif_v6_3_1_vcn_doorbell_range,
+	.gc_doorbell_init = nbif_v6_3_1_gc_doorbell_init,
+	.enable_doorbell_aperture = nbif_v6_3_1_enable_doorbell_aperture,
+	.enable_doorbell_selfring_aperture = nbif_v6_3_1_enable_doorbell_selfring_aperture,
+	.ih_doorbell_range = nbif_v6_3_1_ih_doorbell_range,
+	.update_medium_grain_clock_gating = nbif_v6_3_1_update_medium_grain_clock_gating,
+	.update_medium_grain_light_sleep = nbif_v6_3_1_update_medium_grain_light_sleep,
+	.get_clockgating_state = nbif_v6_3_1_get_clockgating_state,
+	.ih_control = nbif_v6_3_1_ih_control,
+	.init_registers = nbif_v6_3_1_init_registers,
+	.remap_hdp_registers = nbif_v6_3_1_remap_hdp_registers,
+	.get_rom_offset = nbif_v6_3_1_get_rom_offset,
+	.program_aspm = nbif_v6_3_1_program_aspm,
+	.set_reg_remap = nbif_v6_3_1_set_reg_remap,
+};
+
+
+static int nbif_v6_3_1_set_ras_err_event_athub_irq_state(struct amdgpu_device *adev,
+						       struct amdgpu_irq_src *src,
+						       unsigned type,
+						       enum amdgpu_interrupt_state state)
+{
+	/* The ras_controller_irq enablement should be done in psp bl when it
+	 * tries to enable ras feature. Driver only need to set the correct interrupt
+	 * vector for bare-metal and sriov use case respectively
+	 */
+	uint32_t bif_doorbell_int_cntl;
+
+	bif_doorbell_int_cntl = RREG32_SOC15(NBIO, 0, regBIF_BX0_BIF_DOORBELL_INT_CNTL);
+	bif_doorbell_int_cntl = REG_SET_FIELD(bif_doorbell_int_cntl,
+					      BIF_BX0_BIF_DOORBELL_INT_CNTL,
+					      RAS_ATHUB_ERR_EVENT_INTERRUPT_DISABLE,
+					      (state == AMDGPU_IRQ_STATE_ENABLE) ? 0 : 1);
+	WREG32_SOC15(NBIO, 0, regBIF_BX0_BIF_DOORBELL_INT_CNTL, bif_doorbell_int_cntl);
+
+	return 0;
+}
+
+static int nbif_v6_3_1_process_err_event_athub_irq(struct amdgpu_device *adev,
+						 struct amdgpu_irq_src *source,
+						 struct amdgpu_iv_entry *entry)
+{
+	/* By design, the ih cookie for err_event_athub_irq should be written
+	 * to bif ring. since bif ring is not enabled, just leave process callback
+	 * as a dummy one.
+	 */
+	return 0;
+}
+
+static const struct amdgpu_irq_src_funcs nbif_v6_3_1_ras_err_event_athub_irq_funcs = {
+	.set = nbif_v6_3_1_set_ras_err_event_athub_irq_state,
+	.process = nbif_v6_3_1_process_err_event_athub_irq,
+};
+
+static void nbif_v6_3_1_handle_ras_err_event_athub_intr_no_bifring(struct amdgpu_device *adev)
+{
+	uint32_t bif_doorbell_int_cntl;
+
+	bif_doorbell_int_cntl = RREG32_SOC15(NBIO, 0, regBIF_BX0_BIF_DOORBELL_INT_CNTL);
+	if (REG_GET_FIELD(bif_doorbell_int_cntl,
+			  BIF_BX0_BIF_DOORBELL_INT_CNTL,
+			  RAS_ATHUB_ERR_EVENT_INTERRUPT_STATUS)) {
+		/* driver has to clear the interrupt status when bif ring is disabled */
+		bif_doorbell_int_cntl = REG_SET_FIELD(bif_doorbell_int_cntl,
+						BIF_BX0_BIF_DOORBELL_INT_CNTL,
+						RAS_ATHUB_ERR_EVENT_INTERRUPT_CLEAR, 1);
+		WREG32_SOC15(NBIO, 0, regBIF_BX0_BIF_DOORBELL_INT_CNTL, bif_doorbell_int_cntl);
+		amdgpu_ras_global_ras_isr(adev);
+	}
+}
+
+static int nbif_v6_3_1_init_ras_err_event_athub_interrupt(struct amdgpu_device *adev)
+{
+	int r;
+
+	/* init the irq funcs */
+	adev->nbio.ras_err_event_athub_irq.funcs =
+		&nbif_v6_3_1_ras_err_event_athub_irq_funcs;
+	adev->nbio.ras_err_event_athub_irq.num_types = 1;
+
+	/* register ras err event athub interrupt
+	 * nbif v6_3_1 uses the same irq source as nbio v7_4
+	 */
+	r = amdgpu_irq_add_id(adev, SOC21_IH_CLIENTID_BIF,
+			      NBIF_7_4__SRCID__ERREVENT_ATHUB_INTERRUPT,
+			      &adev->nbio.ras_err_event_athub_irq);
+
+	return r;
+}
+
+struct amdgpu_nbio_ras nbif_v6_3_1_ras = {
+	.handle_ras_err_event_athub_intr_no_bifring =
+		nbif_v6_3_1_handle_ras_err_event_athub_intr_no_bifring,
+	.init_ras_err_event_athub_interrupt =
+		nbif_v6_3_1_init_ras_err_event_athub_interrupt,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/nbif_v6_3_1.h b/drivers/gpu/drm/amd/amdgpu/nbif_v6_3_1.h
new file mode 100644
index 000000000000..3afec715a9fe
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/nbif_v6_3_1.h
@@ -0,0 +1,33 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __NBIO_V6_3_1_H__
+#define __NBIO_V6_3_1_H__
+
+#include "soc15_common.h"
+
+extern const struct nbio_hdp_flush_reg nbif_v6_3_1_hdp_flush_reg;
+extern const struct amdgpu_nbio_funcs nbif_v6_3_1_funcs;
+extern struct amdgpu_nbio_ras nbif_v6_3_1_ras;
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v2_3.c b/drivers/gpu/drm/amd/amdgpu/nbio_v2_3.c
index ee7cab37dfd5..04041b398781 100644
--- a/drivers/gpu/drm/amd/amdgpu/nbio_v2_3.c
+++ b/drivers/gpu/drm/amd/amdgpu/nbio_v2_3.c
@@ -21,13 +21,13 @@
  *
  */
 #include "amdgpu.h"
-#include "amdgpu_atombios.h"
 #include "nbio_v2_3.h"
 
 #include "nbio/nbio_2_3_default.h"
 #include "nbio/nbio_2_3_offset.h"
 #include "nbio/nbio_2_3_sh_mask.h"
 #include <uapi/linux/kfd_ioctl.h>
+#include <linux/device.h>
 #include <linux/pci.h>
 
 #define smnPCIE_CONFIG_CNTL	0x11180044
@@ -278,7 +278,7 @@ static void nbio_v2_3_update_medium_grain_light_sleep(struct amdgpu_device *adev
 }
 
 static void nbio_v2_3_get_clockgating_state(struct amdgpu_device *adev,
-					    u32 *flags)
+					    u64 *flags)
 {
 	int data;
 
@@ -328,27 +328,6 @@ const struct nbio_hdp_flush_reg nbio_v2_3_hdp_flush_reg = {
 	.ref_and_mask_sdma1 = BIF_BX_PF_GPU_HDP_FLUSH_DONE__SDMA1_MASK,
 };
 
-const struct nbio_hdp_flush_reg nbio_v2_3_hdp_flush_reg_sc = {
-	.ref_and_mask_cp0 = BIF_BX_PF_GPU_HDP_FLUSH_DONE__CP0_MASK,
-	.ref_and_mask_cp1 = BIF_BX_PF_GPU_HDP_FLUSH_DONE__CP1_MASK,
-	.ref_and_mask_cp2 = BIF_BX_PF_GPU_HDP_FLUSH_DONE__CP2_MASK,
-	.ref_and_mask_cp3 = BIF_BX_PF_GPU_HDP_FLUSH_DONE__CP3_MASK,
-	.ref_and_mask_cp4 = BIF_BX_PF_GPU_HDP_FLUSH_DONE__CP4_MASK,
-	.ref_and_mask_cp5 = BIF_BX_PF_GPU_HDP_FLUSH_DONE__CP5_MASK,
-	.ref_and_mask_cp6 = BIF_BX_PF_GPU_HDP_FLUSH_DONE__CP6_MASK,
-	.ref_and_mask_cp7 = BIF_BX_PF_GPU_HDP_FLUSH_DONE__CP7_MASK,
-	.ref_and_mask_cp8 = BIF_BX_PF_GPU_HDP_FLUSH_DONE__CP8_MASK,
-	.ref_and_mask_cp9 = BIF_BX_PF_GPU_HDP_FLUSH_DONE__CP9_MASK,
-	.ref_and_mask_sdma0 = GPU_HDP_FLUSH_DONE__RSVD_ENG1_MASK,
-	.ref_and_mask_sdma1 = GPU_HDP_FLUSH_DONE__RSVD_ENG2_MASK,
-	.ref_and_mask_sdma2 = GPU_HDP_FLUSH_DONE__RSVD_ENG3_MASK,
-	.ref_and_mask_sdma3 = GPU_HDP_FLUSH_DONE__RSVD_ENG4_MASK,
-	.ref_and_mask_sdma4 = GPU_HDP_FLUSH_DONE__RSVD_ENG5_MASK,
-	.ref_and_mask_sdma5 = GPU_HDP_FLUSH_DONE__RSVD_ENG6_MASK,
-	.ref_and_mask_sdma6 = GPU_HDP_FLUSH_DONE__RSVD_ENG7_MASK,
-	.ref_and_mask_sdma7 = GPU_HDP_FLUSH_DONE__RSVD_ENG8_MASK,
-};
-
 static void nbio_v2_3_init_registers(struct amdgpu_device *adev)
 {
 	uint32_t def, data;
@@ -359,15 +338,11 @@ static void nbio_v2_3_init_registers(struct amdgpu_device *adev)
 
 	if (def != data)
 		WREG32_PCIE(smnPCIE_CONFIG_CNTL, data);
-
-	if (amdgpu_sriov_vf(adev))
-		adev->rmmio_remap.reg_offset = SOC15_REG_OFFSET(NBIO, 0,
-			mmBIF_BX_DEV0_EPF0_VF0_HDP_MEM_COHERENCY_FLUSH_CNTL) << 2;
 }
 
 #define NAVI10_PCIE__LC_L0S_INACTIVITY_DEFAULT		0x00000000 // off by default, no gains over L1
-#define NAVI10_PCIE__LC_L1_INACTIVITY_DEFAULT		0x00000009 // 1=1us, 9=1ms
-#define NAVI10_PCIE__LC_L1_INACTIVITY_TBT_DEFAULT	0x0000000E // 4ms
+#define NAVI10_PCIE__LC_L1_INACTIVITY_DEFAULT		0x0000000A // 1=1us, 9=1ms, 10=4ms
+#define NAVI10_PCIE__LC_L1_INACTIVITY_TBT_DEFAULT	0x0000000E // 400ms
 
 static void nbio_v2_3_enable_aspm(struct amdgpu_device *adev,
 				  bool enable)
@@ -382,14 +357,14 @@ static void nbio_v2_3_enable_aspm(struct amdgpu_device *adev,
 
 		data |= NAVI10_PCIE__LC_L0S_INACTIVITY_DEFAULT << PCIE_LC_CNTL__LC_L0S_INACTIVITY__SHIFT;
 
-		if (pci_is_thunderbolt_attached(adev->pdev))
+		if (dev_is_removable(&adev->pdev->dev))
 			data |= NAVI10_PCIE__LC_L1_INACTIVITY_TBT_DEFAULT  << PCIE_LC_CNTL__LC_L1_INACTIVITY__SHIFT;
 		else
 			data |= NAVI10_PCIE__LC_L1_INACTIVITY_DEFAULT << PCIE_LC_CNTL__LC_L1_INACTIVITY__SHIFT;
 
 		data &= ~PCIE_LC_CNTL__LC_PMI_TO_L1_DIS_MASK;
 	} else {
-		/* Disbale ASPM L1 */
+		/* Disable ASPM L1 */
 		data &= ~PCIE_LC_CNTL__LC_L1_INACTIVITY_MASK;
 		/* Disable ASPM TxL0s */
 		data &= ~PCIE_LC_CNTL__LC_L0S_INACTIVITY_MASK;
@@ -401,6 +376,7 @@ static void nbio_v2_3_enable_aspm(struct amdgpu_device *adev,
 		WREG32_PCIE(smnPCIE_LC_CNTL, data);
 }
 
+#ifdef CONFIG_PCIEASPM
 static void nbio_v2_3_program_ltr(struct amdgpu_device *adev)
 {
 	uint32_t def, data;
@@ -422,9 +398,11 @@ static void nbio_v2_3_program_ltr(struct amdgpu_device *adev)
 	if (def != data)
 		WREG32_PCIE(smnBIF_CFG_DEV0_EPF0_DEVICE_CNTL2, data);
 }
+#endif
 
 static void nbio_v2_3_program_aspm(struct amdgpu_device *adev)
 {
+#ifdef CONFIG_PCIEASPM
 	uint32_t def, data;
 
 	def = data = RREG32_PCIE(smnPCIE_LC_CNTL);
@@ -480,7 +458,10 @@ static void nbio_v2_3_program_aspm(struct amdgpu_device *adev)
 	if (def != data)
 		WREG32_PCIE(smnPCIE_LC_CNTL6, data);
 
-	nbio_v2_3_program_ltr(adev);
+	/* Don't bother about LTR if LTR is not enabled
+	 * in the path */
+	if (adev->pdev->ltr_path)
+		nbio_v2_3_program_ltr(adev);
 
 	def = data = RREG32_SOC15(NBIO, 0, mmRCC_BIF_STRAP3);
 	data |= 0x5DE0 << RCC_BIF_STRAP3__STRAP_VLINK_ASPM_IDLE_TIMER__SHIFT;
@@ -494,9 +475,12 @@ static void nbio_v2_3_program_aspm(struct amdgpu_device *adev)
 		WREG32_SOC15(NBIO, 0, mmRCC_BIF_STRAP5, data);
 
 	def = data = RREG32_PCIE(smnPCIE_LC_CNTL);
-	data &= ~PCIE_LC_CNTL__LC_L0S_INACTIVITY_MASK;
-	data |= 0x9 << PCIE_LC_CNTL__LC_L1_INACTIVITY__SHIFT;
-	data |= 0x1 << PCIE_LC_CNTL__LC_PMI_TO_L1_DIS__SHIFT;
+	data |= NAVI10_PCIE__LC_L0S_INACTIVITY_DEFAULT << PCIE_LC_CNTL__LC_L0S_INACTIVITY__SHIFT;
+	if (dev_is_removable(&adev->pdev->dev))
+		data |= NAVI10_PCIE__LC_L1_INACTIVITY_TBT_DEFAULT  << PCIE_LC_CNTL__LC_L1_INACTIVITY__SHIFT;
+	else
+		data |= NAVI10_PCIE__LC_L1_INACTIVITY_DEFAULT << PCIE_LC_CNTL__LC_L1_INACTIVITY__SHIFT;
+	data &= ~PCIE_LC_CNTL__LC_PMI_TO_L1_DIS_MASK;
 	if (def != data)
 		WREG32_PCIE(smnPCIE_LC_CNTL, data);
 
@@ -504,6 +488,7 @@ static void nbio_v2_3_program_aspm(struct amdgpu_device *adev)
 	data &= ~PCIE_LC_CNTL3__LC_DSC_DONT_ENTER_L23_AFTER_PME_ACK_MASK;
 	if (def != data)
 		WREG32_PCIE(smnPCIE_LC_CNTL3, data);
+#endif
 }
 
 static void nbio_v2_3_apply_lc_spc_mode_wa(struct amdgpu_device *adev)
@@ -547,7 +532,7 @@ static void nbio_v2_3_clear_doorbell_interrupt(struct amdgpu_device *adev)
 {
 	uint32_t reg, reg_data;
 
-	if (adev->asic_type != CHIP_SIENNA_CICHLID)
+	if (amdgpu_ip_version(adev, NBIO_HWIP, 0) != IP_VERSION(3, 3, 0))
 		return;
 
 	reg = RREG32_SOC15(NBIO, 0, mmBIF_RB_CNTL);
@@ -563,6 +548,20 @@ static void nbio_v2_3_clear_doorbell_interrupt(struct amdgpu_device *adev)
 	}
 }
 
+#define MMIO_REG_HOLE_OFFSET (0x80000 - PAGE_SIZE)
+
+static void nbio_v2_3_set_reg_remap(struct amdgpu_device *adev)
+{
+	if (!amdgpu_sriov_vf(adev) && (PAGE_SIZE <= 4096)) {
+		adev->rmmio_remap.reg_offset = MMIO_REG_HOLE_OFFSET;
+		adev->rmmio_remap.bus_addr = adev->rmmio_base + MMIO_REG_HOLE_OFFSET;
+	} else {
+		adev->rmmio_remap.reg_offset = SOC15_REG_OFFSET(NBIO, 0,
+			mmBIF_BX_DEV0_EPF0_VF0_HDP_MEM_COHERENCY_FLUSH_CNTL) << 2;
+		adev->rmmio_remap.bus_addr = 0;
+	}
+}
+
 const struct amdgpu_nbio_funcs nbio_v2_3_funcs = {
 	.get_hdp_flush_req_offset = nbio_v2_3_get_hdp_flush_req_offset,
 	.get_hdp_flush_done_offset = nbio_v2_3_get_hdp_flush_done_offset,
@@ -587,4 +586,5 @@ const struct amdgpu_nbio_funcs nbio_v2_3_funcs = {
 	.apply_lc_spc_mode_wa = nbio_v2_3_apply_lc_spc_mode_wa,
 	.apply_l1_link_width_reconfig_wa = nbio_v2_3_apply_l1_link_width_reconfig_wa,
 	.clear_doorbell_interrupt = nbio_v2_3_clear_doorbell_interrupt,
+	.set_reg_remap = nbio_v2_3_set_reg_remap,
 };
diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v2_3.h b/drivers/gpu/drm/amd/amdgpu/nbio_v2_3.h
index 6074dd3a1ed8..a43b60acf7f6 100644
--- a/drivers/gpu/drm/amd/amdgpu/nbio_v2_3.h
+++ b/drivers/gpu/drm/amd/amdgpu/nbio_v2_3.h
@@ -27,7 +27,6 @@
 #include "soc15_common.h"
 
 extern const struct nbio_hdp_flush_reg nbio_v2_3_hdp_flush_reg;
-extern const struct nbio_hdp_flush_reg nbio_v2_3_hdp_flush_reg_sc;
 extern const struct amdgpu_nbio_funcs nbio_v2_3_funcs;
 
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v4_3.c b/drivers/gpu/drm/amd/amdgpu/nbio_v4_3.c
new file mode 100644
index 000000000000..f89e5f40e1a5
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/nbio_v4_3.c
@@ -0,0 +1,634 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include "amdgpu.h"
+#include "nbio_v4_3.h"
+
+#include "nbio/nbio_4_3_0_offset.h"
+#include "nbio/nbio_4_3_0_sh_mask.h"
+#include "ivsrcid/nbio/irqsrcs_nbif_7_4.h"
+#include <uapi/linux/kfd_ioctl.h>
+
+static void nbio_v4_3_remap_hdp_registers(struct amdgpu_device *adev)
+{
+	WREG32_SOC15(NBIO, 0, regBIF_BX0_REMAP_HDP_MEM_FLUSH_CNTL,
+		adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL);
+	WREG32_SOC15(NBIO, 0, regBIF_BX0_REMAP_HDP_REG_FLUSH_CNTL,
+		adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_REG_FLUSH_CNTL);
+}
+
+static u32 nbio_v4_3_get_rev_id(struct amdgpu_device *adev)
+{
+	u32 tmp = RREG32_SOC15(NBIO, 0, regRCC_STRAP0_RCC_DEV0_EPF0_STRAP0);
+
+	tmp &= RCC_STRAP0_RCC_DEV0_EPF0_STRAP0__STRAP_ATI_REV_ID_DEV0_F0_MASK;
+	tmp >>= RCC_STRAP0_RCC_DEV0_EPF0_STRAP0__STRAP_ATI_REV_ID_DEV0_F0__SHIFT;
+
+	return tmp;
+}
+
+static void nbio_v4_3_mc_access_enable(struct amdgpu_device *adev, bool enable)
+{
+	if (enable)
+		WREG32_SOC15(NBIO, 0, regBIF_BX0_BIF_FB_EN,
+			     BIF_BX0_BIF_FB_EN__FB_READ_EN_MASK |
+			     BIF_BX0_BIF_FB_EN__FB_WRITE_EN_MASK);
+	else
+		WREG32_SOC15(NBIO, 0, regBIF_BX0_BIF_FB_EN, 0);
+}
+
+static u32 nbio_v4_3_get_memsize(struct amdgpu_device *adev)
+{
+	return RREG32_SOC15(NBIO, 0, regRCC_DEV0_EPF0_RCC_CONFIG_MEMSIZE);
+}
+
+static void nbio_v4_3_sdma_doorbell_range(struct amdgpu_device *adev, int instance,
+					  bool use_doorbell, int doorbell_index,
+					  int doorbell_size)
+{
+	if (instance == 0) {
+		u32 doorbell_range = RREG32_SOC15(NBIO, 0, regS2A_DOORBELL_ENTRY_2_CTRL);
+
+		if (use_doorbell) {
+			doorbell_range = REG_SET_FIELD(doorbell_range,
+						       S2A_DOORBELL_ENTRY_2_CTRL,
+						       S2A_DOORBELL_PORT2_ENABLE,
+						       0x1);
+			doorbell_range = REG_SET_FIELD(doorbell_range,
+						       S2A_DOORBELL_ENTRY_2_CTRL,
+						       S2A_DOORBELL_PORT2_AWID,
+						       0xe);
+			doorbell_range = REG_SET_FIELD(doorbell_range,
+						       S2A_DOORBELL_ENTRY_2_CTRL,
+						       S2A_DOORBELL_PORT2_RANGE_OFFSET,
+						       doorbell_index);
+			doorbell_range = REG_SET_FIELD(doorbell_range,
+						       S2A_DOORBELL_ENTRY_2_CTRL,
+						       S2A_DOORBELL_PORT2_RANGE_SIZE,
+						       doorbell_size);
+			doorbell_range = REG_SET_FIELD(doorbell_range,
+						       S2A_DOORBELL_ENTRY_2_CTRL,
+						       S2A_DOORBELL_PORT2_AWADDR_31_28_VALUE,
+						       0x3);
+		} else
+			doorbell_range = REG_SET_FIELD(doorbell_range,
+						       S2A_DOORBELL_ENTRY_2_CTRL,
+						       S2A_DOORBELL_PORT2_RANGE_SIZE,
+						       0);
+
+		WREG32_SOC15(NBIO, 0, regS2A_DOORBELL_ENTRY_2_CTRL, doorbell_range);
+	}
+}
+
+static void nbio_v4_3_vcn_doorbell_range(struct amdgpu_device *adev, bool use_doorbell,
+					 int doorbell_index, int instance)
+{
+	u32 doorbell_range;
+
+	if (instance)
+		doorbell_range = RREG32_SOC15(NBIO, 0, regS2A_DOORBELL_ENTRY_5_CTRL);
+	else
+		doorbell_range = RREG32_SOC15(NBIO, 0, regS2A_DOORBELL_ENTRY_4_CTRL);
+
+	if (use_doorbell) {
+		doorbell_range = REG_SET_FIELD(doorbell_range,
+					       S2A_DOORBELL_ENTRY_4_CTRL,
+					       S2A_DOORBELL_PORT4_ENABLE,
+					       0x1);
+		doorbell_range = REG_SET_FIELD(doorbell_range,
+					       S2A_DOORBELL_ENTRY_4_CTRL,
+					       S2A_DOORBELL_PORT4_AWID,
+					       instance ? 0x7 : 0x4);
+		doorbell_range = REG_SET_FIELD(doorbell_range,
+					       S2A_DOORBELL_ENTRY_4_CTRL,
+					       S2A_DOORBELL_PORT4_RANGE_OFFSET,
+					       doorbell_index);
+		doorbell_range = REG_SET_FIELD(doorbell_range,
+					       S2A_DOORBELL_ENTRY_4_CTRL,
+					       S2A_DOORBELL_PORT4_RANGE_SIZE,
+					       8);
+		doorbell_range = REG_SET_FIELD(doorbell_range,
+					       S2A_DOORBELL_ENTRY_4_CTRL,
+					       S2A_DOORBELL_PORT4_AWADDR_31_28_VALUE,
+					       instance ? 0x7 : 0x4);
+	} else
+		doorbell_range = REG_SET_FIELD(doorbell_range,
+					       S2A_DOORBELL_ENTRY_4_CTRL,
+					       S2A_DOORBELL_PORT4_RANGE_SIZE,
+					       0);
+
+	if (instance)
+		WREG32_SOC15(NBIO, 0, regS2A_DOORBELL_ENTRY_5_CTRL, doorbell_range);
+	else
+		WREG32_SOC15(NBIO, 0, regS2A_DOORBELL_ENTRY_4_CTRL, doorbell_range);
+}
+
+static void nbio_v4_3_gc_doorbell_init(struct amdgpu_device *adev)
+{
+	WREG32_SOC15(NBIO, 0, regS2A_DOORBELL_ENTRY_0_CTRL, 0x30000007);
+	WREG32_SOC15(NBIO, 0, regS2A_DOORBELL_ENTRY_3_CTRL, 0x3000000d);
+}
+
+static void nbio_v4_3_enable_doorbell_aperture(struct amdgpu_device *adev,
+					       bool enable)
+{
+	WREG32_FIELD15_PREREG(NBIO, 0, RCC_DEV0_EPF0_RCC_DOORBELL_APER_EN,
+			BIF_DOORBELL_APER_EN, enable ? 1 : 0);
+}
+
+static void nbio_v4_3_enable_doorbell_selfring_aperture(struct amdgpu_device *adev,
+							bool enable)
+{
+	u32 tmp = 0;
+
+	if (enable) {
+		tmp = REG_SET_FIELD(tmp, BIF_BX_PF0_DOORBELL_SELFRING_GPA_APER_CNTL,
+				    DOORBELL_SELFRING_GPA_APER_EN, 1) |
+		      REG_SET_FIELD(tmp, BIF_BX_PF0_DOORBELL_SELFRING_GPA_APER_CNTL,
+				    DOORBELL_SELFRING_GPA_APER_MODE, 1) |
+		      REG_SET_FIELD(tmp, BIF_BX_PF0_DOORBELL_SELFRING_GPA_APER_CNTL,
+				    DOORBELL_SELFRING_GPA_APER_SIZE, 0);
+
+		WREG32_SOC15(NBIO, 0, regBIF_BX_PF0_DOORBELL_SELFRING_GPA_APER_BASE_LOW,
+			     lower_32_bits(adev->doorbell.base));
+		WREG32_SOC15(NBIO, 0, regBIF_BX_PF0_DOORBELL_SELFRING_GPA_APER_BASE_HIGH,
+			     upper_32_bits(adev->doorbell.base));
+	}
+
+	WREG32_SOC15(NBIO, 0, regBIF_BX_PF0_DOORBELL_SELFRING_GPA_APER_CNTL,
+		     tmp);
+}
+
+static void nbio_v4_3_ih_doorbell_range(struct amdgpu_device *adev,
+					bool use_doorbell, int doorbell_index)
+{
+	u32 ih_doorbell_range = RREG32_SOC15(NBIO, 0, regS2A_DOORBELL_ENTRY_1_CTRL);
+
+	if (use_doorbell) {
+		ih_doorbell_range = REG_SET_FIELD(ih_doorbell_range,
+						  S2A_DOORBELL_ENTRY_1_CTRL,
+						  S2A_DOORBELL_PORT1_ENABLE,
+						  0x1);
+		ih_doorbell_range = REG_SET_FIELD(ih_doorbell_range,
+						  S2A_DOORBELL_ENTRY_1_CTRL,
+						  S2A_DOORBELL_PORT1_AWID,
+						  0x0);
+		ih_doorbell_range = REG_SET_FIELD(ih_doorbell_range,
+						  S2A_DOORBELL_ENTRY_1_CTRL,
+						  S2A_DOORBELL_PORT1_RANGE_OFFSET,
+						  doorbell_index);
+		ih_doorbell_range = REG_SET_FIELD(ih_doorbell_range,
+						  S2A_DOORBELL_ENTRY_1_CTRL,
+						  S2A_DOORBELL_PORT1_RANGE_SIZE,
+						  2);
+		ih_doorbell_range = REG_SET_FIELD(ih_doorbell_range,
+						  S2A_DOORBELL_ENTRY_1_CTRL,
+						  S2A_DOORBELL_PORT1_AWADDR_31_28_VALUE,
+						  0x0);
+	} else
+		ih_doorbell_range = REG_SET_FIELD(ih_doorbell_range,
+						  S2A_DOORBELL_ENTRY_1_CTRL,
+						  S2A_DOORBELL_PORT1_RANGE_SIZE,
+						  0);
+
+	WREG32_SOC15(NBIO, 0, regS2A_DOORBELL_ENTRY_1_CTRL, ih_doorbell_range);
+}
+
+static void nbio_v4_3_ih_control(struct amdgpu_device *adev)
+{
+	u32 interrupt_cntl;
+
+	/* setup interrupt control */
+	WREG32_SOC15(NBIO, 0, regBIF_BX0_INTERRUPT_CNTL2, adev->dummy_page_addr >> 8);
+
+	interrupt_cntl = RREG32_SOC15(NBIO, 0, regBIF_BX0_INTERRUPT_CNTL);
+	/*
+	 * BIF_BX0_INTERRUPT_CNTL__IH_DUMMY_RD_OVERRIDE_MASK=0 - dummy read disabled with msi, enabled without msi
+	 * BIF_BX0_INTERRUPT_CNTL__IH_DUMMY_RD_OVERRIDE_MASK=1 - dummy read controlled by IH_DUMMY_RD_EN
+	 */
+	interrupt_cntl = REG_SET_FIELD(interrupt_cntl, BIF_BX0_INTERRUPT_CNTL,
+				       IH_DUMMY_RD_OVERRIDE, 0);
+
+	/* BIF_BX0_INTERRUPT_CNTL__IH_REQ_NONSNOOP_EN_MASK=1 if ring is in non-cacheable memory, e.g., vram */
+	interrupt_cntl = REG_SET_FIELD(interrupt_cntl, BIF_BX0_INTERRUPT_CNTL,
+				       IH_REQ_NONSNOOP_EN, 0);
+
+	WREG32_SOC15(NBIO, 0, regBIF_BX0_INTERRUPT_CNTL, interrupt_cntl);
+}
+
+static void nbio_v4_3_update_medium_grain_clock_gating(struct amdgpu_device *adev,
+						       bool enable)
+{
+	uint32_t def, data;
+
+	if (enable && !(adev->cg_flags & AMD_CG_SUPPORT_BIF_MGCG))
+		return;
+
+	def = data = RREG32_SOC15(NBIO, 0, regCPM_CONTROL);
+	if (enable) {
+		data |= (CPM_CONTROL__LCLK_DYN_GATE_ENABLE_MASK |
+			 CPM_CONTROL__TXCLK_DYN_GATE_ENABLE_MASK |
+			 CPM_CONTROL__TXCLK_LCNT_GATE_ENABLE_MASK |
+			 CPM_CONTROL__TXCLK_REGS_GATE_ENABLE_MASK |
+			 CPM_CONTROL__TXCLK_PRBS_GATE_ENABLE_MASK |
+			 CPM_CONTROL__REFCLK_REGS_GATE_ENABLE_MASK);
+	} else {
+		data &= ~(CPM_CONTROL__LCLK_DYN_GATE_ENABLE_MASK |
+			  CPM_CONTROL__TXCLK_DYN_GATE_ENABLE_MASK |
+			  CPM_CONTROL__TXCLK_LCNT_GATE_ENABLE_MASK |
+			  CPM_CONTROL__TXCLK_REGS_GATE_ENABLE_MASK |
+			  CPM_CONTROL__TXCLK_PRBS_GATE_ENABLE_MASK |
+			  CPM_CONTROL__REFCLK_REGS_GATE_ENABLE_MASK);
+	}
+
+	if (def != data)
+		WREG32_SOC15(NBIO, 0, regCPM_CONTROL, data);
+}
+
+static void nbio_v4_3_update_medium_grain_light_sleep(struct amdgpu_device *adev,
+						      bool enable)
+{
+	uint32_t def, data;
+
+	if (enable && !(adev->cg_flags & AMD_CG_SUPPORT_BIF_LS))
+		return;
+
+	/* TODO: need update in future */
+	def = data = RREG32_SOC15(NBIO, 0, regPCIE_CNTL2);
+	if (enable) {
+		data |= PCIE_CNTL2__SLV_MEM_LS_EN_MASK;
+	} else {
+		data &= ~PCIE_CNTL2__SLV_MEM_LS_EN_MASK;
+	}
+
+	if (def != data)
+		WREG32_SOC15(NBIO, 0, regPCIE_CNTL2, data);
+}
+
+static void nbio_v4_3_get_clockgating_state(struct amdgpu_device *adev,
+					    u64 *flags)
+{
+	int data;
+
+	/* AMD_CG_SUPPORT_BIF_MGCG */
+	data = RREG32_SOC15(NBIO, 0, regCPM_CONTROL);
+	if (data & CPM_CONTROL__LCLK_DYN_GATE_ENABLE_MASK)
+		*flags |= AMD_CG_SUPPORT_BIF_MGCG;
+
+	/* AMD_CG_SUPPORT_BIF_LS */
+	data = RREG32_SOC15(NBIO, 0, regPCIE_CNTL2);
+	if (data & PCIE_CNTL2__SLV_MEM_LS_EN_MASK)
+		*flags |= AMD_CG_SUPPORT_BIF_LS;
+}
+
+static u32 nbio_v4_3_get_hdp_flush_req_offset(struct amdgpu_device *adev)
+{
+	return SOC15_REG_OFFSET(NBIO, 0, regBIF_BX_PF0_GPU_HDP_FLUSH_REQ);
+}
+
+static u32 nbio_v4_3_get_hdp_flush_done_offset(struct amdgpu_device *adev)
+{
+	return SOC15_REG_OFFSET(NBIO, 0, regBIF_BX_PF0_GPU_HDP_FLUSH_DONE);
+}
+
+static u32 nbio_v4_3_get_pcie_index_offset(struct amdgpu_device *adev)
+{
+	return SOC15_REG_OFFSET(NBIO, 0, regBIF_BX_PF0_RSMU_INDEX);
+}
+
+static u32 nbio_v4_3_get_pcie_data_offset(struct amdgpu_device *adev)
+{
+	return SOC15_REG_OFFSET(NBIO, 0, regBIF_BX_PF0_RSMU_DATA);
+}
+
+const struct nbio_hdp_flush_reg nbio_v4_3_hdp_flush_reg = {
+	.ref_and_mask_cp0 = BIF_BX_PF_GPU_HDP_FLUSH_DONE__CP0_MASK,
+	.ref_and_mask_cp1 = BIF_BX_PF_GPU_HDP_FLUSH_DONE__CP1_MASK,
+	.ref_and_mask_cp2 = BIF_BX_PF_GPU_HDP_FLUSH_DONE__CP2_MASK,
+	.ref_and_mask_cp3 = BIF_BX_PF_GPU_HDP_FLUSH_DONE__CP3_MASK,
+	.ref_and_mask_cp4 = BIF_BX_PF_GPU_HDP_FLUSH_DONE__CP4_MASK,
+	.ref_and_mask_cp5 = BIF_BX_PF_GPU_HDP_FLUSH_DONE__CP5_MASK,
+	.ref_and_mask_cp6 = BIF_BX_PF_GPU_HDP_FLUSH_DONE__CP6_MASK,
+	.ref_and_mask_cp7 = BIF_BX_PF_GPU_HDP_FLUSH_DONE__CP7_MASK,
+	.ref_and_mask_cp8 = BIF_BX_PF_GPU_HDP_FLUSH_DONE__CP8_MASK,
+	.ref_and_mask_cp9 = BIF_BX_PF_GPU_HDP_FLUSH_DONE__CP9_MASK,
+	.ref_and_mask_sdma0 = BIF_BX_PF_GPU_HDP_FLUSH_DONE__SDMA0_MASK,
+	.ref_and_mask_sdma1 = BIF_BX_PF_GPU_HDP_FLUSH_DONE__SDMA1_MASK,
+};
+
+static void nbio_v4_3_init_registers(struct amdgpu_device *adev)
+{
+	if (amdgpu_ip_version(adev, NBIO_HWIP, 0) == IP_VERSION(4, 3, 0)) {
+		uint32_t data;
+
+		data = RREG32_SOC15(NBIO, 0, regRCC_DEV0_EPF2_STRAP2);
+		data &= ~RCC_DEV0_EPF2_STRAP2__STRAP_NO_SOFT_RESET_DEV0_F2_MASK;
+		WREG32_SOC15(NBIO, 0, regRCC_DEV0_EPF2_STRAP2, data);
+	}
+}
+
+static u32 nbio_v4_3_get_rom_offset(struct amdgpu_device *adev)
+{
+	u32 data, rom_offset;
+
+	data = RREG32_SOC15(NBIO, 0, regREGS_ROM_OFFSET_CTRL);
+	rom_offset = REG_GET_FIELD(data, REGS_ROM_OFFSET_CTRL, ROM_OFFSET);
+
+	return rom_offset;
+}
+
+#ifdef CONFIG_PCIEASPM
+static void nbio_v4_3_program_ltr(struct amdgpu_device *adev)
+{
+	uint32_t def, data;
+
+	def = RREG32_SOC15(NBIO, 0, regRCC_EP_DEV0_0_EP_PCIE_TX_LTR_CNTL);
+	data = 0x35EB;
+	data &= ~EP_PCIE_TX_LTR_CNTL__LTR_PRIV_MSG_DIS_IN_PM_NON_D0_MASK;
+	data &= ~EP_PCIE_TX_LTR_CNTL__LTR_PRIV_RST_LTR_IN_DL_DOWN_MASK;
+	if (def != data)
+		WREG32_SOC15(NBIO, 0, regRCC_EP_DEV0_0_EP_PCIE_TX_LTR_CNTL, data);
+
+	def = data = RREG32_SOC15(NBIO, 0, regRCC_STRAP0_RCC_BIF_STRAP2);
+	data &= ~RCC_BIF_STRAP2__STRAP_LTR_IN_ASPML1_DIS_MASK;
+	if (def != data)
+		WREG32_SOC15(NBIO, 0, regRCC_STRAP0_RCC_BIF_STRAP2, data);
+
+	def = data = RREG32_SOC15(NBIO, 0, regBIF_CFG_DEV0_EPF0_DEVICE_CNTL2);
+	if (adev->pdev->ltr_path)
+		data |= BIF_CFG_DEV0_EPF0_DEVICE_CNTL2__LTR_EN_MASK;
+	else
+		data &= ~BIF_CFG_DEV0_EPF0_DEVICE_CNTL2__LTR_EN_MASK;
+	if (def != data)
+		WREG32_SOC15(NBIO, 0, regBIF_CFG_DEV0_EPF0_DEVICE_CNTL2, data);
+}
+#endif
+
+static void nbio_v4_3_program_aspm(struct amdgpu_device *adev)
+{
+#ifdef CONFIG_PCIEASPM
+	uint32_t def, data;
+
+	if (!(amdgpu_ip_version(adev, PCIE_HWIP, 0) == IP_VERSION(7, 4, 0)) &&
+	    !(amdgpu_ip_version(adev, PCIE_HWIP, 0) == IP_VERSION(7, 6, 0)))
+		return;
+
+	def = data = RREG32_SOC15(NBIO, 0, regPCIE_LC_CNTL);
+	data &= ~PCIE_LC_CNTL__LC_L1_INACTIVITY_MASK;
+	data &= ~PCIE_LC_CNTL__LC_L0S_INACTIVITY_MASK;
+	data |= PCIE_LC_CNTL__LC_PMI_TO_L1_DIS_MASK;
+	if (def != data)
+		WREG32_SOC15(NBIO, 0, regPCIE_LC_CNTL, data);
+
+	def = data = RREG32_SOC15(NBIO, 0, regPCIE_LC_CNTL7);
+	data |= PCIE_LC_CNTL7__LC_NBIF_ASPM_INPUT_EN_MASK;
+	if (def != data)
+		WREG32_SOC15(NBIO, 0, regPCIE_LC_CNTL7, data);
+
+	def = data = RREG32_SOC15(NBIO, 0, regPCIE_LC_CNTL3);
+	data |= PCIE_LC_CNTL3__LC_DSC_DONT_ENTER_L23_AFTER_PME_ACK_MASK;
+	if (def != data)
+		WREG32_SOC15(NBIO, 0, regPCIE_LC_CNTL3, data);
+
+	def = data = RREG32_SOC15(NBIO, 0, regRCC_STRAP0_RCC_BIF_STRAP3);
+	data &= ~RCC_BIF_STRAP3__STRAP_VLINK_ASPM_IDLE_TIMER_MASK;
+	data &= ~RCC_BIF_STRAP3__STRAP_VLINK_PM_L1_ENTRY_TIMER_MASK;
+	if (def != data)
+		WREG32_SOC15(NBIO, 0, regRCC_STRAP0_RCC_BIF_STRAP3, data);
+
+	def = data = RREG32_SOC15(NBIO, 0, regRCC_STRAP0_RCC_BIF_STRAP5);
+	data &= ~RCC_BIF_STRAP5__STRAP_VLINK_LDN_ENTRY_TIMER_MASK;
+	if (def != data)
+		WREG32_SOC15(NBIO, 0, regRCC_STRAP0_RCC_BIF_STRAP5, data);
+
+	def = data = RREG32_SOC15(NBIO, 0, regBIF_CFG_DEV0_EPF0_DEVICE_CNTL2);
+	data &= ~BIF_CFG_DEV0_EPF0_DEVICE_CNTL2__LTR_EN_MASK;
+	if (def != data)
+		WREG32_SOC15(NBIO, 0, regBIF_CFG_DEV0_EPF0_DEVICE_CNTL2, data);
+
+	WREG32_SOC15(NBIO, 0, regBIF_CFG_DEV0_EPF0_PCIE_LTR_CAP, 0x10011001);
+
+	def = data = RREG32_SOC15(NBIO, 0, regPSWUSP0_PCIE_LC_CNTL2);
+	data |= PSWUSP0_PCIE_LC_CNTL2__LC_ALLOW_PDWN_IN_L1_MASK |
+		PSWUSP0_PCIE_LC_CNTL2__LC_ALLOW_PDWN_IN_L23_MASK;
+	data &= ~PSWUSP0_PCIE_LC_CNTL2__LC_RCV_L0_TO_RCV_L0S_DIS_MASK;
+	if (def != data)
+		WREG32_SOC15(NBIO, 0, regPSWUSP0_PCIE_LC_CNTL2, data);
+
+	def = data = RREG32_SOC15(NBIO, 0, regPCIE_LC_CNTL4);
+	data |= PCIE_LC_CNTL4__LC_L1_POWERDOWN_MASK;
+	if (def != data)
+		WREG32_SOC15(NBIO, 0, regPCIE_LC_CNTL4, data);
+
+	def = data = RREG32_SOC15(NBIO, 0, regPCIE_LC_RXRECOVER_RXSTANDBY_CNTL);
+	data |= PCIE_LC_RXRECOVER_RXSTANDBY_CNTL__LC_RX_L0S_STANDBY_EN_MASK;
+	if (def != data)
+		WREG32_SOC15(NBIO, 0, regPCIE_LC_RXRECOVER_RXSTANDBY_CNTL, data);
+
+	nbio_v4_3_program_ltr(adev);
+
+	def = data = RREG32_SOC15(NBIO, 0, regRCC_STRAP0_RCC_BIF_STRAP3);
+	data |= 0x5DE0 << RCC_BIF_STRAP3__STRAP_VLINK_ASPM_IDLE_TIMER__SHIFT;
+	data |= 0x0010 << RCC_BIF_STRAP3__STRAP_VLINK_PM_L1_ENTRY_TIMER__SHIFT;
+	if (def != data)
+		WREG32_SOC15(NBIO, 0, regRCC_STRAP0_RCC_BIF_STRAP3, data);
+
+	def = data = RREG32_SOC15(NBIO, 0, regRCC_STRAP0_RCC_BIF_STRAP5);
+	data |= 0x0010 << RCC_BIF_STRAP5__STRAP_VLINK_LDN_ENTRY_TIMER__SHIFT;
+	if (def != data)
+		WREG32_SOC15(NBIO, 0, regRCC_STRAP0_RCC_BIF_STRAP5, data);
+
+	def = data = RREG32_SOC15(NBIO, 0, regPCIE_LC_CNTL);
+	data |= 0x0 << PCIE_LC_CNTL__LC_L0S_INACTIVITY__SHIFT;
+	data |= 0x9 << PCIE_LC_CNTL__LC_L1_INACTIVITY__SHIFT;
+	data &= ~PCIE_LC_CNTL__LC_PMI_TO_L1_DIS_MASK;
+	if (def != data)
+		WREG32_SOC15(NBIO, 0, regPCIE_LC_CNTL, data);
+
+	def = data = RREG32_SOC15(NBIO, 0, regPCIE_LC_CNTL3);
+	data &= ~PCIE_LC_CNTL3__LC_DSC_DONT_ENTER_L23_AFTER_PME_ACK_MASK;
+	if (def != data)
+		WREG32_SOC15(NBIO, 0, regPCIE_LC_CNTL3, data);
+#endif
+}
+
+#define MMIO_REG_HOLE_OFFSET (0x80000 - PAGE_SIZE)
+
+static void nbio_v4_3_set_reg_remap(struct amdgpu_device *adev)
+{
+	if (!amdgpu_sriov_vf(adev) && (PAGE_SIZE <= 4096)) {
+		adev->rmmio_remap.reg_offset = MMIO_REG_HOLE_OFFSET;
+		adev->rmmio_remap.bus_addr = adev->rmmio_base + MMIO_REG_HOLE_OFFSET;
+	} else {
+		adev->rmmio_remap.reg_offset = SOC15_REG_OFFSET(NBIO, 0,
+			regBIF_BX_DEV0_EPF0_VF0_HDP_MEM_COHERENCY_FLUSH_CNTL) << 2;
+		adev->rmmio_remap.bus_addr = 0;
+	}
+}
+
+const struct amdgpu_nbio_funcs nbio_v4_3_funcs = {
+	.get_hdp_flush_req_offset = nbio_v4_3_get_hdp_flush_req_offset,
+	.get_hdp_flush_done_offset = nbio_v4_3_get_hdp_flush_done_offset,
+	.get_pcie_index_offset = nbio_v4_3_get_pcie_index_offset,
+	.get_pcie_data_offset = nbio_v4_3_get_pcie_data_offset,
+	.get_rev_id = nbio_v4_3_get_rev_id,
+	.mc_access_enable = nbio_v4_3_mc_access_enable,
+	.get_memsize = nbio_v4_3_get_memsize,
+	.sdma_doorbell_range = nbio_v4_3_sdma_doorbell_range,
+	.vcn_doorbell_range = nbio_v4_3_vcn_doorbell_range,
+	.gc_doorbell_init = nbio_v4_3_gc_doorbell_init,
+	.enable_doorbell_aperture = nbio_v4_3_enable_doorbell_aperture,
+	.enable_doorbell_selfring_aperture = nbio_v4_3_enable_doorbell_selfring_aperture,
+	.ih_doorbell_range = nbio_v4_3_ih_doorbell_range,
+	.update_medium_grain_clock_gating = nbio_v4_3_update_medium_grain_clock_gating,
+	.update_medium_grain_light_sleep = nbio_v4_3_update_medium_grain_light_sleep,
+	.get_clockgating_state = nbio_v4_3_get_clockgating_state,
+	.ih_control = nbio_v4_3_ih_control,
+	.init_registers = nbio_v4_3_init_registers,
+	.remap_hdp_registers = nbio_v4_3_remap_hdp_registers,
+	.get_rom_offset = nbio_v4_3_get_rom_offset,
+	.program_aspm = nbio_v4_3_program_aspm,
+	.set_reg_remap = nbio_v4_3_set_reg_remap,
+};
+
+
+static void nbio_v4_3_sriov_ih_doorbell_range(struct amdgpu_device *adev,
+					bool use_doorbell, int doorbell_index)
+{
+}
+
+static void nbio_v4_3_sriov_sdma_doorbell_range(struct amdgpu_device *adev, int instance,
+					  bool use_doorbell, int doorbell_index,
+					  int doorbell_size)
+{
+}
+
+static void nbio_v4_3_sriov_vcn_doorbell_range(struct amdgpu_device *adev, bool use_doorbell,
+					 int doorbell_index, int instance)
+{
+}
+
+static void nbio_v4_3_sriov_gc_doorbell_init(struct amdgpu_device *adev)
+{
+}
+
+const struct amdgpu_nbio_funcs nbio_v4_3_sriov_funcs = {
+	.get_hdp_flush_req_offset = nbio_v4_3_get_hdp_flush_req_offset,
+	.get_hdp_flush_done_offset = nbio_v4_3_get_hdp_flush_done_offset,
+	.get_pcie_index_offset = nbio_v4_3_get_pcie_index_offset,
+	.get_pcie_data_offset = nbio_v4_3_get_pcie_data_offset,
+	.get_rev_id = nbio_v4_3_get_rev_id,
+	.mc_access_enable = nbio_v4_3_mc_access_enable,
+	.get_memsize = nbio_v4_3_get_memsize,
+	.sdma_doorbell_range = nbio_v4_3_sriov_sdma_doorbell_range,
+	.vcn_doorbell_range = nbio_v4_3_sriov_vcn_doorbell_range,
+	.gc_doorbell_init = nbio_v4_3_sriov_gc_doorbell_init,
+	.enable_doorbell_aperture = nbio_v4_3_enable_doorbell_aperture,
+	.enable_doorbell_selfring_aperture = nbio_v4_3_enable_doorbell_selfring_aperture,
+	.ih_doorbell_range = nbio_v4_3_sriov_ih_doorbell_range,
+	.update_medium_grain_clock_gating = nbio_v4_3_update_medium_grain_clock_gating,
+	.update_medium_grain_light_sleep = nbio_v4_3_update_medium_grain_light_sleep,
+	.get_clockgating_state = nbio_v4_3_get_clockgating_state,
+	.ih_control = nbio_v4_3_ih_control,
+	.init_registers = nbio_v4_3_init_registers,
+	.remap_hdp_registers = nbio_v4_3_remap_hdp_registers,
+	.get_rom_offset = nbio_v4_3_get_rom_offset,
+	.set_reg_remap = nbio_v4_3_set_reg_remap,
+};
+
+static int nbio_v4_3_set_ras_err_event_athub_irq_state(struct amdgpu_device *adev,
+						       struct amdgpu_irq_src *src,
+						       unsigned type,
+						       enum amdgpu_interrupt_state state)
+{
+	/* The ras_controller_irq enablement should be done in psp bl when it
+	 * tries to enable ras feature. Driver only need to set the correct interrupt
+	 * vector for bare-metal and sriov use case respectively
+	 */
+	uint32_t bif_doorbell_int_cntl;
+
+	bif_doorbell_int_cntl = RREG32_SOC15(NBIO, 0, regBIF_BX0_BIF_DOORBELL_INT_CNTL);
+	bif_doorbell_int_cntl = REG_SET_FIELD(bif_doorbell_int_cntl,
+					      BIF_BX0_BIF_DOORBELL_INT_CNTL,
+					      RAS_ATHUB_ERR_EVENT_INTERRUPT_DISABLE,
+					      (state == AMDGPU_IRQ_STATE_ENABLE) ? 0 : 1);
+	WREG32_SOC15(NBIO, 0, regBIF_BX0_BIF_DOORBELL_INT_CNTL, bif_doorbell_int_cntl);
+
+	return 0;
+}
+
+static int nbio_v4_3_process_err_event_athub_irq(struct amdgpu_device *adev,
+						 struct amdgpu_irq_src *source,
+						 struct amdgpu_iv_entry *entry)
+{
+	/* By design, the ih cookie for err_event_athub_irq should be written
+	 * to bif ring. since bif ring is not enabled, just leave process callback
+	 * as a dummy one.
+	 */
+	return 0;
+}
+
+static const struct amdgpu_irq_src_funcs nbio_v4_3_ras_err_event_athub_irq_funcs = {
+	.set = nbio_v4_3_set_ras_err_event_athub_irq_state,
+	.process = nbio_v4_3_process_err_event_athub_irq,
+};
+
+static void nbio_v4_3_handle_ras_err_event_athub_intr_no_bifring(struct amdgpu_device *adev)
+{
+	uint32_t bif_doorbell_int_cntl;
+
+	bif_doorbell_int_cntl = RREG32_SOC15(NBIO, 0, regBIF_BX0_BIF_DOORBELL_INT_CNTL);
+	if (REG_GET_FIELD(bif_doorbell_int_cntl,
+			  BIF_DOORBELL_INT_CNTL,
+			  RAS_ATHUB_ERR_EVENT_INTERRUPT_STATUS)) {
+		/* driver has to clear the interrupt status when bif ring is disabled */
+		bif_doorbell_int_cntl = REG_SET_FIELD(bif_doorbell_int_cntl,
+						BIF_DOORBELL_INT_CNTL,
+						RAS_ATHUB_ERR_EVENT_INTERRUPT_CLEAR, 1);
+		WREG32_SOC15(NBIO, 0, regBIF_BX0_BIF_DOORBELL_INT_CNTL, bif_doorbell_int_cntl);
+		amdgpu_ras_global_ras_isr(adev);
+	}
+}
+
+static int nbio_v4_3_init_ras_err_event_athub_interrupt(struct amdgpu_device *adev)
+{
+
+	int r;
+
+	/* init the irq funcs */
+	adev->nbio.ras_err_event_athub_irq.funcs =
+		&nbio_v4_3_ras_err_event_athub_irq_funcs;
+	adev->nbio.ras_err_event_athub_irq.num_types = 1;
+
+	/* register ras err event athub interrupt
+	 * nbio v4_3 uses the same irq source as nbio v7_4 */
+	r = amdgpu_irq_add_id(adev, SOC21_IH_CLIENTID_BIF,
+			      NBIF_7_4__SRCID__ERREVENT_ATHUB_INTERRUPT,
+			      &adev->nbio.ras_err_event_athub_irq);
+
+	return r;
+}
+
+struct amdgpu_nbio_ras nbio_v4_3_ras = {
+	.handle_ras_err_event_athub_intr_no_bifring = nbio_v4_3_handle_ras_err_event_athub_intr_no_bifring,
+	.init_ras_err_event_athub_interrupt = nbio_v4_3_init_ras_err_event_athub_interrupt,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v4_3.h b/drivers/gpu/drm/amd/amdgpu/nbio_v4_3.h
new file mode 100644
index 000000000000..399037cdf4fb
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/nbio_v4_3.h
@@ -0,0 +1,34 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __NBIO_V4_3_H__
+#define __NBIO_V4_3_H__
+
+#include "soc15_common.h"
+
+extern const struct nbio_hdp_flush_reg nbio_v4_3_hdp_flush_reg;
+extern const struct amdgpu_nbio_funcs nbio_v4_3_funcs;
+extern const struct amdgpu_nbio_funcs nbio_v4_3_sriov_funcs;
+extern struct amdgpu_nbio_ras nbio_v4_3_ras;
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v6_1.c b/drivers/gpu/drm/amd/amdgpu/nbio_v6_1.c
index 4bbacf1be25a..e911368c1aeb 100644
--- a/drivers/gpu/drm/amd/amdgpu/nbio_v6_1.c
+++ b/drivers/gpu/drm/amd/amdgpu/nbio_v6_1.c
@@ -21,7 +21,6 @@
  *
  */
 #include "amdgpu.h"
-#include "amdgpu_atombios.h"
 #include "nbio_v6_1.h"
 
 #include "nbio/nbio_6_1_default.h"
@@ -210,7 +209,7 @@ static void nbio_v6_1_update_medium_grain_light_sleep(struct amdgpu_device *adev
 }
 
 static void nbio_v6_1_get_clockgating_state(struct amdgpu_device *adev,
-					    u32 *flags)
+					    u64 *flags)
 {
 	int data;
 
@@ -276,12 +275,9 @@ static void nbio_v6_1_init_registers(struct amdgpu_device *adev)
 
 	if (def != data)
 		WREG32_PCIE(smnPCIE_CI_CNTL, data);
-
-	if (amdgpu_sriov_vf(adev))
-		adev->rmmio_remap.reg_offset = SOC15_REG_OFFSET(NBIO, 0,
-			mmBIF_BX_DEV0_EPF0_VF0_HDP_MEM_COHERENCY_FLUSH_CNTL) << 2;
 }
 
+#ifdef CONFIG_PCIEASPM
 static void nbio_v6_1_program_ltr(struct amdgpu_device *adev)
 {
 	uint32_t def, data;
@@ -303,9 +299,11 @@ static void nbio_v6_1_program_ltr(struct amdgpu_device *adev)
 	if (def != data)
 		WREG32_PCIE(smnBIF_CFG_DEV0_EPF0_DEVICE_CNTL2, data);
 }
+#endif
 
 static void nbio_v6_1_program_aspm(struct amdgpu_device *adev)
 {
+#ifdef CONFIG_PCIEASPM
 	uint32_t def, data;
 
 	def = data = RREG32_PCIE(smnPCIE_LC_CNTL);
@@ -361,7 +359,10 @@ static void nbio_v6_1_program_aspm(struct amdgpu_device *adev)
 	if (def != data)
 		WREG32_PCIE(smnPCIE_LC_CNTL6, data);
 
-	nbio_v6_1_program_ltr(adev);
+	/* Don't bother about LTR if LTR is not enabled
+	 * in the path */
+	if (adev->pdev->ltr_path)
+		nbio_v6_1_program_ltr(adev);
 
 	def = data = RREG32_PCIE(smnRCC_BIF_STRAP3);
 	data |= 0x5DE0 << RCC_BIF_STRAP3__STRAP_VLINK_ASPM_IDLE_TIMER__SHIFT;
@@ -385,6 +386,22 @@ static void nbio_v6_1_program_aspm(struct amdgpu_device *adev)
 	data &= ~PCIE_LC_CNTL3__LC_DSC_DONT_ENTER_L23_AFTER_PME_ACK_MASK;
 	if (def != data)
 		WREG32_PCIE(smnPCIE_LC_CNTL3, data);
+#endif
+}
+
+#define MMIO_REG_HOLE_OFFSET (0x80000 - PAGE_SIZE)
+
+static void nbio_v6_1_set_reg_remap(struct amdgpu_device *adev)
+{
+	if (!amdgpu_sriov_vf(adev) && (PAGE_SIZE <= 4096)) {
+		adev->rmmio_remap.reg_offset = MMIO_REG_HOLE_OFFSET;
+		adev->rmmio_remap.bus_addr = adev->rmmio_base + MMIO_REG_HOLE_OFFSET;
+	} else {
+		adev->rmmio_remap.reg_offset =
+			SOC15_REG_OFFSET(NBIO, 0,
+					 mmBIF_BX_DEV0_EPF0_VF0_HDP_MEM_COHERENCY_FLUSH_CNTL) << 2;
+		adev->rmmio_remap.bus_addr = 0;
+	}
 }
 
 const struct amdgpu_nbio_funcs nbio_v6_1_funcs = {
@@ -405,5 +422,6 @@ const struct amdgpu_nbio_funcs nbio_v6_1_funcs = {
 	.ih_control = nbio_v6_1_ih_control,
 	.init_registers = nbio_v6_1_init_registers,
 	.remap_hdp_registers = nbio_v6_1_remap_hdp_registers,
-	.program_aspm =  nbio_v6_1_program_aspm,
+	.program_aspm = nbio_v6_1_program_aspm,
+	.set_reg_remap = nbio_v6_1_set_reg_remap,
 };
diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v7_0.c b/drivers/gpu/drm/amd/amdgpu/nbio_v7_0.c
index 37a4039fdfc5..1569a1e934ec 100644
--- a/drivers/gpu/drm/amd/amdgpu/nbio_v7_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/nbio_v7_0.c
@@ -21,7 +21,6 @@
  *
  */
 #include "amdgpu.h"
-#include "amdgpu_atombios.h"
 #include "nbio_v7_0.h"
 
 #include "nbio/nbio_7_0_default.h"
@@ -205,7 +204,7 @@ static void nbio_v7_0_update_medium_grain_light_sleep(struct amdgpu_device *adev
 }
 
 static void nbio_v7_0_get_clockgating_state(struct amdgpu_device *adev,
-					    u32 *flags)
+					    u64 *flags)
 {
 	int data;
 
@@ -271,11 +270,33 @@ const struct nbio_hdp_flush_reg nbio_v7_0_hdp_flush_reg = {
 	.ref_and_mask_sdma1 = GPU_HDP_FLUSH_DONE__SDMA1_MASK,
 };
 
+#define regRCC_DEV0_EPF6_STRAP4                                                                         0xd304
+#define regRCC_DEV0_EPF6_STRAP4_BASE_IDX                                                                5
+
 static void nbio_v7_0_init_registers(struct amdgpu_device *adev)
 {
-	if (amdgpu_sriov_vf(adev))
+	uint32_t data;
+
+	switch (amdgpu_ip_version(adev, NBIO_HWIP, 0)) {
+	case IP_VERSION(2, 5, 0):
+		data = RREG32_SOC15(NBIO, 0, regRCC_DEV0_EPF6_STRAP4) & ~BIT(23);
+		WREG32_SOC15(NBIO, 0, regRCC_DEV0_EPF6_STRAP4, data);
+		break;
+	}
+}
+
+#define MMIO_REG_HOLE_OFFSET (0x80000 - PAGE_SIZE)
+
+static void nbio_v7_0_set_reg_remap(struct amdgpu_device *adev)
+{
+	if (!amdgpu_sriov_vf(adev) && (PAGE_SIZE <= 4096)) {
+		adev->rmmio_remap.reg_offset = MMIO_REG_HOLE_OFFSET;
+		adev->rmmio_remap.bus_addr = adev->rmmio_base + MMIO_REG_HOLE_OFFSET;
+	} else {
 		adev->rmmio_remap.reg_offset =
 			SOC15_REG_OFFSET(NBIO, 0, mmHDP_MEM_COHERENCY_FLUSH_CNTL) << 2;
+		adev->rmmio_remap.bus_addr = 0;
+	}
 }
 
 const struct amdgpu_nbio_funcs nbio_v7_0_funcs = {
@@ -297,4 +318,5 @@ const struct amdgpu_nbio_funcs nbio_v7_0_funcs = {
 	.ih_control = nbio_v7_0_ih_control,
 	.init_registers = nbio_v7_0_init_registers,
 	.remap_hdp_registers = nbio_v7_0_remap_hdp_registers,
+	.set_reg_remap = nbio_v7_0_set_reg_remap,
 };
diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v7_11.c b/drivers/gpu/drm/amd/amdgpu/nbio_v7_11.c
new file mode 100644
index 000000000000..bed5ef4d8788
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/nbio_v7_11.c
@@ -0,0 +1,400 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include "amdgpu.h"
+#include "nbio_v7_11.h"
+
+#include "nbio/nbio_7_11_0_offset.h"
+#include "nbio/nbio_7_11_0_sh_mask.h"
+#include <uapi/linux/kfd_ioctl.h>
+
+static void nbio_v7_11_remap_hdp_registers(struct amdgpu_device *adev)
+{
+	WREG32_SOC15(NBIO, 0, regBIF_BX0_REMAP_HDP_MEM_FLUSH_CNTL,
+		     adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL);
+	WREG32_SOC15(NBIO, 0, regBIF_BX0_REMAP_HDP_REG_FLUSH_CNTL,
+		     adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_REG_FLUSH_CNTL);
+}
+
+static u32 nbio_v7_11_get_rev_id(struct amdgpu_device *adev)
+{
+	u32 tmp;
+
+	tmp = RREG32_SOC15(NBIO, 0, regRCC_STRAP1_RCC_DEV0_EPF0_STRAP0);
+	tmp &= RCC_STRAP0_RCC_DEV0_EPF0_STRAP0__STRAP_ATI_REV_ID_DEV0_F0_MASK;
+	tmp >>= RCC_STRAP0_RCC_DEV0_EPF0_STRAP0__STRAP_ATI_REV_ID_DEV0_F0__SHIFT;
+
+	return tmp;
+}
+
+static void nbio_v7_11_mc_access_enable(struct amdgpu_device *adev, bool enable)
+{
+	if (enable)
+		WREG32_SOC15(NBIO, 0, regBIF_BX1_BIF_FB_EN,
+			BIF_BX1_BIF_FB_EN__FB_READ_EN_MASK |
+			BIF_BX1_BIF_FB_EN__FB_WRITE_EN_MASK);
+	else
+		WREG32_SOC15(NBIO, 0, regBIF_BX1_BIF_FB_EN, 0);
+}
+
+static u32 nbio_v7_11_get_memsize(struct amdgpu_device *adev)
+{
+	return RREG32_SOC15(NBIO, 0, regRCC_DEV0_EPF0_0_RCC_CONFIG_MEMSIZE);
+}
+
+static void nbio_v7_11_sdma_doorbell_range(struct amdgpu_device *adev, int instance,
+					  bool use_doorbell, int doorbell_index,
+					  int doorbell_size)
+{
+	u32 reg = SOC15_REG_OFFSET(NBIO, 0, regGDC0_BIF_CSDMA_DOORBELL_RANGE);
+	u32 doorbell_range = RREG32_PCIE_PORT(reg);
+
+	if (use_doorbell) {
+		doorbell_range = REG_SET_FIELD(doorbell_range,
+					       GDC0_BIF_CSDMA_DOORBELL_RANGE,
+					       OFFSET, doorbell_index);
+		doorbell_range = REG_SET_FIELD(doorbell_range,
+					       GDC0_BIF_CSDMA_DOORBELL_RANGE,
+					       SIZE, doorbell_size);
+	} else {
+		doorbell_range = REG_SET_FIELD(doorbell_range,
+					       GDC0_BIF_CSDMA_DOORBELL_RANGE,
+					       SIZE, 0);
+	}
+
+	WREG32_PCIE_PORT(reg, doorbell_range);
+}
+
+static void nbio_v7_11_vpe_doorbell_range(struct amdgpu_device *adev, int instance,
+					  bool use_doorbell, int doorbell_index,
+					  int doorbell_size)
+{
+	u32 reg = instance == 0 ?
+		  SOC15_REG_OFFSET(NBIO, 0, regGDC0_BIF_VPE_DOORBELL_RANGE) :
+		  SOC15_REG_OFFSET(NBIO, 0, regGDC0_BIF_VPE1_DOORBELL_RANGE);
+	u32 doorbell_range = RREG32_PCIE_PORT(reg);
+
+	if (use_doorbell) {
+		doorbell_range = REG_SET_FIELD(doorbell_range,
+					       GDC0_BIF_VPE_DOORBELL_RANGE,
+					       OFFSET, doorbell_index);
+		doorbell_range = REG_SET_FIELD(doorbell_range,
+					       GDC0_BIF_VPE_DOORBELL_RANGE,
+					       SIZE, doorbell_size);
+	} else {
+		doorbell_range = REG_SET_FIELD(doorbell_range,
+					       GDC0_BIF_VPE_DOORBELL_RANGE,
+					       SIZE, 0);
+	}
+
+	WREG32_PCIE_PORT(reg, doorbell_range);
+}
+
+static void nbio_v7_11_vcn_doorbell_range(struct amdgpu_device *adev,
+					  bool use_doorbell,
+					  int doorbell_index, int instance)
+{
+	u32 reg = instance == 0 ?
+		SOC15_REG_OFFSET(NBIO, 0, regGDC0_BIF_VCN0_DOORBELL_RANGE):
+		SOC15_REG_OFFSET(NBIO, 0, regGDC0_BIF_VCN1_DOORBELL_RANGE);
+
+	u32 doorbell_range = RREG32_PCIE_PORT(reg);
+
+	if (use_doorbell) {
+		doorbell_range = REG_SET_FIELD(doorbell_range,
+					       GDC0_BIF_VCN0_DOORBELL_RANGE, OFFSET,
+					       doorbell_index);
+		doorbell_range = REG_SET_FIELD(doorbell_range,
+					       GDC0_BIF_VCN0_DOORBELL_RANGE, SIZE, 8);
+	} else {
+		doorbell_range = REG_SET_FIELD(doorbell_range,
+					       GDC0_BIF_VCN0_DOORBELL_RANGE, SIZE, 0);
+	}
+
+	WREG32_PCIE_PORT(reg, doorbell_range);
+}
+
+static void nbio_v7_11_enable_doorbell_aperture(struct amdgpu_device *adev,
+					       bool enable)
+{
+	u32 reg;
+
+
+	reg = RREG32_SOC15(NBIO, 0, regRCC_DEV0_EPF0_0_RCC_DOORBELL_APER_EN);
+	reg = REG_SET_FIELD(reg, RCC_DEV0_EPF0_0_RCC_DOORBELL_APER_EN,
+			    BIF_DOORBELL_APER_EN, enable ? 1 : 0);
+
+	WREG32_SOC15(NBIO, 0, regRCC_DEV0_EPF0_0_RCC_DOORBELL_APER_EN, reg);
+}
+
+static void nbio_v7_11_enable_doorbell_selfring_aperture(struct amdgpu_device *adev,
+							bool enable)
+{
+	u32 tmp = 0;
+
+	if (enable) {
+		tmp = REG_SET_FIELD(tmp, BIF_BX_PF1_DOORBELL_SELFRING_GPA_APER_CNTL,
+				DOORBELL_SELFRING_GPA_APER_EN, 1) |
+		      REG_SET_FIELD(tmp, BIF_BX_PF1_DOORBELL_SELFRING_GPA_APER_CNTL,
+				DOORBELL_SELFRING_GPA_APER_MODE, 1) |
+		      REG_SET_FIELD(tmp, BIF_BX_PF1_DOORBELL_SELFRING_GPA_APER_CNTL,
+				DOORBELL_SELFRING_GPA_APER_SIZE, 0);
+
+		WREG32_SOC15(NBIO, 0,
+			regBIF_BX_PF1_DOORBELL_SELFRING_GPA_APER_BASE_LOW,
+			lower_32_bits(adev->doorbell.base));
+		WREG32_SOC15(NBIO, 0,
+			regBIF_BX_PF1_DOORBELL_SELFRING_GPA_APER_BASE_HIGH,
+			upper_32_bits(adev->doorbell.base));
+	}
+
+	WREG32_SOC15(NBIO, 0, regBIF_BX_PF1_DOORBELL_SELFRING_GPA_APER_CNTL, tmp);
+}
+
+
+static void nbio_v7_11_ih_doorbell_range(struct amdgpu_device *adev,
+					bool use_doorbell, int doorbell_index)
+{    
+	u32 ih_doorbell_range = RREG32_SOC15(NBIO, 0,regGDC0_BIF_IH_DOORBELL_RANGE);
+
+	if (use_doorbell) {
+		ih_doorbell_range = REG_SET_FIELD(ih_doorbell_range,
+						  GDC0_BIF_IH_DOORBELL_RANGE, OFFSET,
+						  doorbell_index);
+		ih_doorbell_range = REG_SET_FIELD(ih_doorbell_range,
+						  GDC0_BIF_IH_DOORBELL_RANGE, SIZE,
+						  2);
+	} else {
+		ih_doorbell_range = REG_SET_FIELD(ih_doorbell_range,
+						  GDC0_BIF_IH_DOORBELL_RANGE, SIZE,
+						  0);
+	}
+
+	WREG32_SOC15(NBIO, 0, regGDC0_BIF_IH_DOORBELL_RANGE,
+			 ih_doorbell_range);
+}
+
+static void nbio_v7_11_ih_control(struct amdgpu_device *adev)
+{
+	u32 interrupt_cntl;
+
+	/* setup interrupt control */
+	WREG32_SOC15(NBIO, 0, regBIF_BX1_INTERRUPT_CNTL2,
+		     adev->dummy_page_addr >> 8);
+
+	interrupt_cntl = RREG32_SOC15(NBIO, 0, regBIF_BX1_INTERRUPT_CNTL);
+	/*
+	 * INTERRUPT_CNTL__IH_DUMMY_RD_OVERRIDE_MASK=0 - dummy read disabled with msi, enabled without msi
+	 * INTERRUPT_CNTL__IH_DUMMY_RD_OVERRIDE_MASK=1 - dummy read controlled by IH_DUMMY_RD_EN
+	 */
+	interrupt_cntl = REG_SET_FIELD(interrupt_cntl, BIF_BX1_INTERRUPT_CNTL,
+				       IH_DUMMY_RD_OVERRIDE, 0);
+
+	/* INTERRUPT_CNTL__IH_REQ_NONSNOOP_EN_MASK=1 if ring is in non-cacheable memory, e.g., vram */
+	interrupt_cntl = REG_SET_FIELD(interrupt_cntl, BIF_BX1_INTERRUPT_CNTL,
+				       IH_REQ_NONSNOOP_EN, 0);
+
+	WREG32_SOC15(NBIO, 0, regBIF_BX1_INTERRUPT_CNTL, interrupt_cntl);
+}
+
+static u32 nbio_v7_11_get_hdp_flush_req_offset(struct amdgpu_device *adev)
+{
+	return SOC15_REG_OFFSET(NBIO, 0, regBIF_BX_PF1_GPU_HDP_FLUSH_REQ);
+}
+
+static u32 nbio_v7_11_get_hdp_flush_done_offset(struct amdgpu_device *adev)
+{
+	return SOC15_REG_OFFSET(NBIO, 0, regBIF_BX_PF1_GPU_HDP_FLUSH_DONE);
+}
+
+static u32 nbio_v7_11_get_pcie_index_offset(struct amdgpu_device *adev)
+{
+	return SOC15_REG_OFFSET(NBIO, 0, regBIF_BX1_PCIE_INDEX2);
+}
+
+static u32 nbio_v7_11_get_pcie_data_offset(struct amdgpu_device *adev)
+{
+	return SOC15_REG_OFFSET(NBIO, 0, regBIF_BX1_PCIE_DATA2);
+}
+
+static u32 nbio_v7_11_get_pcie_port_index_offset(struct amdgpu_device *adev)
+{
+	return SOC15_REG_OFFSET(NBIO, 0, regBIF_BX_PF1_RSMU_INDEX);
+}
+
+static u32 nbio_v7_11_get_pcie_port_data_offset(struct amdgpu_device *adev)
+{
+	return SOC15_REG_OFFSET(NBIO, 0, regBIF_BX_PF1_RSMU_DATA);
+}
+
+const struct nbio_hdp_flush_reg nbio_v7_11_hdp_flush_reg = {
+	.ref_and_mask_cp0 = BIF_BX_PF1_GPU_HDP_FLUSH_DONE__CP0_MASK,
+	.ref_and_mask_cp1 = BIF_BX_PF1_GPU_HDP_FLUSH_DONE__CP1_MASK,
+	.ref_and_mask_cp2 = BIF_BX_PF1_GPU_HDP_FLUSH_DONE__CP2_MASK,
+	.ref_and_mask_cp3 = BIF_BX_PF1_GPU_HDP_FLUSH_DONE__CP3_MASK,
+	.ref_and_mask_cp4 = BIF_BX_PF1_GPU_HDP_FLUSH_DONE__CP4_MASK,
+	.ref_and_mask_cp5 = BIF_BX_PF1_GPU_HDP_FLUSH_DONE__CP5_MASK,
+	.ref_and_mask_cp6 = BIF_BX_PF1_GPU_HDP_FLUSH_DONE__CP6_MASK,
+	.ref_and_mask_cp7 = BIF_BX_PF1_GPU_HDP_FLUSH_DONE__CP7_MASK,
+	.ref_and_mask_cp8 = BIF_BX_PF1_GPU_HDP_FLUSH_DONE__CP8_MASK,
+	.ref_and_mask_cp9 = BIF_BX_PF1_GPU_HDP_FLUSH_DONE__CP9_MASK,
+	.ref_and_mask_sdma0 = BIF_BX_PF1_GPU_HDP_FLUSH_DONE__SDMA0_MASK,
+	.ref_and_mask_sdma1 = BIF_BX_PF1_GPU_HDP_FLUSH_DONE__SDMA1_MASK,
+};
+
+static void nbio_v7_11_init_registers(struct amdgpu_device *adev)
+{
+	uint32_t def, data;
+
+	def = data = RREG32_SOC15(NBIO, 0, regBIF_BIF256_CI256_RC3X4_USB4_PCIE_MST_CTRL_3);
+	data = REG_SET_FIELD(data, BIF_BIF256_CI256_RC3X4_USB4_PCIE_MST_CTRL_3,
+				CI_SWUS_MAX_READ_REQUEST_SIZE_MODE, 1);
+	data = REG_SET_FIELD(data, BIF_BIF256_CI256_RC3X4_USB4_PCIE_MST_CTRL_3,
+				CI_SWUS_MAX_READ_REQUEST_SIZE_PRIV, 1);
+
+	if (def != data)
+		WREG32_SOC15(NBIO, 0, regBIF_BIF256_CI256_RC3X4_USB4_PCIE_MST_CTRL_3, data);
+
+	switch (amdgpu_ip_version(adev, NBIO_HWIP, 0)) {
+	case IP_VERSION(7, 11, 0):
+	case IP_VERSION(7, 11, 1):
+	case IP_VERSION(7, 11, 2):
+	case IP_VERSION(7, 11, 3):
+		data = RREG32_SOC15(NBIO, 0, regRCC_DEV0_EPF5_STRAP4) & ~BIT(23);
+		WREG32_SOC15(NBIO, 0, regRCC_DEV0_EPF5_STRAP4, data);
+		break;
+	}
+}
+
+static void nbio_v7_11_update_medium_grain_clock_gating(struct amdgpu_device *adev,
+						       bool enable)
+{
+	uint32_t def, data;
+
+	if (!(adev->cg_flags & AMD_CG_SUPPORT_BIF_MGCG))
+		return;
+
+	def = data = RREG32_SOC15(NBIO, 0, regBIF_BIF256_CI256_RC3X4_USB4_CPM_CONTROL);
+	if (enable) {
+		data |= (BIF_BIF256_CI256_RC3X4_USB4_CPM_CONTROL__LCLK_DYN_GATE_ENABLE_MASK |
+			 BIF_BIF256_CI256_RC3X4_USB4_CPM_CONTROL__TXCLK_DYN_GATE_ENABLE_MASK |
+			 BIF_BIF256_CI256_RC3X4_USB4_CPM_CONTROL__TXCLK_LCNT_GATE_ENABLE_MASK |
+			 BIF_BIF256_CI256_RC3X4_USB4_CPM_CONTROL__TXCLK_REGS_GATE_ENABLE_MASK |
+			 BIF_BIF256_CI256_RC3X4_USB4_CPM_CONTROL__TXCLK_PRBS_GATE_ENABLE_MASK |
+			 BIF_BIF256_CI256_RC3X4_USB4_CPM_CONTROL__REFCLK_REGS_GATE_ENABLE_MASK);
+	} else {
+		data &= ~(BIF_BIF256_CI256_RC3X4_USB4_CPM_CONTROL__LCLK_DYN_GATE_ENABLE_MASK |
+			  BIF_BIF256_CI256_RC3X4_USB4_CPM_CONTROL__TXCLK_DYN_GATE_ENABLE_MASK |
+			  BIF_BIF256_CI256_RC3X4_USB4_CPM_CONTROL__TXCLK_LCNT_GATE_ENABLE_MASK |
+			  BIF_BIF256_CI256_RC3X4_USB4_CPM_CONTROL__TXCLK_REGS_GATE_ENABLE_MASK |
+			  BIF_BIF256_CI256_RC3X4_USB4_CPM_CONTROL__TXCLK_PRBS_GATE_ENABLE_MASK |
+			  BIF_BIF256_CI256_RC3X4_USB4_CPM_CONTROL__REFCLK_REGS_GATE_ENABLE_MASK);
+	}
+
+	if (def != data)
+		WREG32_SOC15(NBIO, 0, regBIF_BIF256_CI256_RC3X4_USB4_CPM_CONTROL, data);
+}
+
+static void nbio_v7_11_update_medium_grain_light_sleep(struct amdgpu_device *adev,
+						      bool enable)
+{
+	uint32_t def, data;
+
+	if (!(adev->cg_flags & AMD_CG_SUPPORT_BIF_LS))
+		return;
+
+	def = data = RREG32_SOC15(NBIO, 0, regBIF_BIF256_CI256_RC3X4_USB4_PCIE_CNTL2);
+	if (enable)
+		data |= BIF_BIF256_CI256_RC3X4_USB4_PCIE_CNTL2__SLV_MEM_LS_EN_MASK;
+	else
+		data &= ~BIF_BIF256_CI256_RC3X4_USB4_PCIE_CNTL2__SLV_MEM_LS_EN_MASK;
+
+	if (def != data)
+		WREG32_SOC15(NBIO, 0, regBIF_BIF256_CI256_RC3X4_USB4_PCIE_CNTL2, data);
+
+	def = data = RREG32_SOC15(NBIO, 0, regBIF_BIF256_CI256_RC3X4_USB4_PCIE_TX_POWER_CTRL_1);
+	if (enable) {
+		data |= (BIF_BIF256_CI256_RC3X4_USB4_PCIE_TX_POWER_CTRL_1__MST_MEM_LS_EN_MASK |
+			BIF_BIF256_CI256_RC3X4_USB4_PCIE_TX_POWER_CTRL_1__REPLAY_MEM_LS_EN_MASK);
+	} else {
+		data &= ~(BIF_BIF256_CI256_RC3X4_USB4_PCIE_TX_POWER_CTRL_1__MST_MEM_LS_EN_MASK |
+			BIF_BIF256_CI256_RC3X4_USB4_PCIE_TX_POWER_CTRL_1__REPLAY_MEM_LS_EN_MASK);
+	}
+
+	if (def != data)
+		WREG32_SOC15(NBIO, 0, regBIF_BIF256_CI256_RC3X4_USB4_PCIE_TX_POWER_CTRL_1, data);
+}
+
+static void nbio_v7_11_get_clockgating_state(struct amdgpu_device *adev,
+					    u64 *flags)
+{
+	uint32_t data;
+
+	/* AMD_CG_SUPPORT_BIF_MGCG */
+	data = RREG32_SOC15(NBIO, 0, regBIF_BIF256_CI256_RC3X4_USB4_CPM_CONTROL);
+	if (data & BIF_BIF256_CI256_RC3X4_USB4_CPM_CONTROL__LCLK_DYN_GATE_ENABLE_MASK)
+		*flags |= AMD_CG_SUPPORT_BIF_MGCG;
+
+	/* AMD_CG_SUPPORT_BIF_LS */
+	data = RREG32_SOC15(NBIO, 0, regBIF_BIF256_CI256_RC3X4_USB4_PCIE_CNTL2);
+	if (data & BIF_BIF256_CI256_RC3X4_USB4_PCIE_CNTL2__SLV_MEM_LS_EN_MASK)
+		*flags |= AMD_CG_SUPPORT_BIF_LS;
+}
+
+#define MMIO_REG_HOLE_OFFSET 0x44000
+
+static void nbio_v7_11_set_reg_remap(struct amdgpu_device *adev)
+{
+	if (!amdgpu_sriov_vf(adev) && (PAGE_SIZE <= 4096)) {
+		adev->rmmio_remap.reg_offset = MMIO_REG_HOLE_OFFSET;
+		adev->rmmio_remap.bus_addr = adev->rmmio_base + MMIO_REG_HOLE_OFFSET;
+	} else {
+		adev->rmmio_remap.reg_offset =
+			SOC15_REG_OFFSET(NBIO, 0, regBIF_BX_PF1_HDP_MEM_COHERENCY_FLUSH_CNTL) << 2;
+		adev->rmmio_remap.bus_addr = 0;
+	}
+}
+
+const struct amdgpu_nbio_funcs nbio_v7_11_funcs = {
+	.get_hdp_flush_req_offset = nbio_v7_11_get_hdp_flush_req_offset,
+	.get_hdp_flush_done_offset = nbio_v7_11_get_hdp_flush_done_offset,
+	.get_pcie_index_offset = nbio_v7_11_get_pcie_index_offset,
+	.get_pcie_data_offset = nbio_v7_11_get_pcie_data_offset,
+	.get_pcie_port_index_offset = nbio_v7_11_get_pcie_port_index_offset,
+	.get_pcie_port_data_offset = nbio_v7_11_get_pcie_port_data_offset,
+	.get_rev_id = nbio_v7_11_get_rev_id,
+	.mc_access_enable = nbio_v7_11_mc_access_enable,
+	.get_memsize = nbio_v7_11_get_memsize,
+	.sdma_doorbell_range = nbio_v7_11_sdma_doorbell_range,
+	.vcn_doorbell_range = nbio_v7_11_vcn_doorbell_range,
+	.vpe_doorbell_range = nbio_v7_11_vpe_doorbell_range,
+	.enable_doorbell_aperture = nbio_v7_11_enable_doorbell_aperture,
+	.enable_doorbell_selfring_aperture = nbio_v7_11_enable_doorbell_selfring_aperture,
+	.ih_doorbell_range = nbio_v7_11_ih_doorbell_range,
+	.update_medium_grain_clock_gating = nbio_v7_11_update_medium_grain_clock_gating,
+	.update_medium_grain_light_sleep = nbio_v7_11_update_medium_grain_light_sleep,
+	.get_clockgating_state = nbio_v7_11_get_clockgating_state,
+	.ih_control = nbio_v7_11_ih_control,
+	.init_registers = nbio_v7_11_init_registers,
+	.remap_hdp_registers = nbio_v7_11_remap_hdp_registers,
+	.set_reg_remap = nbio_v7_11_set_reg_remap,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v7_11.h b/drivers/gpu/drm/amd/amdgpu/nbio_v7_11.h
new file mode 100644
index 000000000000..9d8258ed3f0a
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/nbio_v7_11.h
@@ -0,0 +1,33 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __NBIO_V7_11_H__
+#define __NBIO_V7_11_H__
+
+#include "soc15_common.h"
+
+extern const struct nbio_hdp_flush_reg nbio_v7_11_hdp_flush_reg;
+extern const struct amdgpu_nbio_funcs nbio_v7_11_funcs;
+extern const struct amdgpu_nbio_ras_funcs nbio_v7_11_ras_funcs;
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v7_2.c b/drivers/gpu/drm/amd/amdgpu/nbio_v7_2.c
index 3444332ea110..acc5f363684a 100644
--- a/drivers/gpu/drm/amd/amdgpu/nbio_v7_2.c
+++ b/drivers/gpu/drm/amd/amdgpu/nbio_v7_2.c
@@ -21,7 +21,6 @@
  *
  */
 #include "amdgpu.h"
-#include "amdgpu_atombios.h"
 #include "nbio_v7_2.h"
 
 #include "nbio/nbio_7_2_0_offset.h"
@@ -59,10 +58,16 @@ static u32 nbio_v7_2_get_rev_id(struct amdgpu_device *adev)
 {
 	u32 tmp;
 
-	if (adev->asic_type == CHIP_YELLOW_CARP)
+	switch (amdgpu_ip_version(adev, NBIO_HWIP, 0)) {
+	case IP_VERSION(7, 2, 1):
+	case IP_VERSION(7, 3, 0):
+	case IP_VERSION(7, 5, 0):
 		tmp = RREG32_SOC15(NBIO, 0, regRCC_STRAP0_RCC_DEV0_EPF0_STRAP0_YC);
-	else
+		break;
+	default:
 		tmp = RREG32_SOC15(NBIO, 0, regRCC_STRAP0_RCC_DEV0_EPF0_STRAP0);
+		break;
+	}
 
 	tmp &= RCC_STRAP0_RCC_DEV0_EPF0_STRAP0__STRAP_ATI_REV_ID_DEV0_F0_MASK;
 	tmp >>= RCC_STRAP0_RCC_DEV0_EPF0_STRAP0__STRAP_ATI_REV_ID_DEV0_F0__SHIFT;
@@ -72,20 +77,26 @@ static u32 nbio_v7_2_get_rev_id(struct amdgpu_device *adev)
 
 static void nbio_v7_2_mc_access_enable(struct amdgpu_device *adev, bool enable)
 {
-	if (enable)
-		if (adev->asic_type == CHIP_YELLOW_CARP)
+	switch (amdgpu_ip_version(adev, NBIO_HWIP, 0)) {
+	case IP_VERSION(7, 2, 1):
+	case IP_VERSION(7, 3, 0):
+	case IP_VERSION(7, 5, 0):
+		if (enable)
 			WREG32_SOC15(NBIO, 0, regBIF_BX0_BIF_FB_EN_YC,
 				BIF_BX0_BIF_FB_EN__FB_READ_EN_MASK |
 				BIF_BX0_BIF_FB_EN__FB_WRITE_EN_MASK);
 		else
+			WREG32_SOC15(NBIO, 0, regBIF_BX0_BIF_FB_EN_YC, 0);
+	break;
+	default:
+		if (enable)
 			WREG32_SOC15(NBIO, 0, regBIF_BX0_BIF_FB_EN,
 				BIF_BX0_BIF_FB_EN__FB_READ_EN_MASK |
 				BIF_BX0_BIF_FB_EN__FB_WRITE_EN_MASK);
-	else
-		if (adev->asic_type == CHIP_YELLOW_CARP)
-			WREG32_SOC15(NBIO, 0, regBIF_BX0_BIF_FB_EN_YC, 0);
 		else
 			WREG32_SOC15(NBIO, 0, regBIF_BX0_BIF_FB_EN, 0);
+		break;
+	}
 }
 
 static u32 nbio_v7_2_get_memsize(struct amdgpu_device *adev)
@@ -250,7 +261,10 @@ static void nbio_v7_2_update_medium_grain_light_sleep(struct amdgpu_device *adev
 {
 	uint32_t def, data;
 
-	if (adev->asic_type == CHIP_YELLOW_CARP) {
+	switch (amdgpu_ip_version(adev, NBIO_HWIP, 0)) {
+	case IP_VERSION(7, 2, 1):
+	case IP_VERSION(7, 3, 0):
+	case IP_VERSION(7, 5, 0):
 		def = data = RREG32_PCIE_PORT(SOC15_REG_OFFSET(NBIO, 0, regPCIE_CNTL2));
 		if (enable && (adev->cg_flags & AMD_CG_SUPPORT_BIF_LS))
 			data |= PCIE_CNTL2__SLV_MEM_LS_EN_MASK;
@@ -260,8 +274,8 @@ static void nbio_v7_2_update_medium_grain_light_sleep(struct amdgpu_device *adev
 		if (def != data)
 			WREG32_PCIE_PORT(SOC15_REG_OFFSET(NBIO, 0, regPCIE_CNTL2), data);
 
-		data = RREG32_PCIE_PORT(SOC15_REG_OFFSET(NBIO, 0, regBIF1_PCIE_TX_POWER_CTRL_1));
-		def = data;
+		def = data = RREG32_PCIE_PORT(SOC15_REG_OFFSET(NBIO, 0,
+			regBIF1_PCIE_TX_POWER_CTRL_1));
 		if (enable && (adev->cg_flags & AMD_CG_SUPPORT_BIF_LS))
 			data |= (BIF1_PCIE_TX_POWER_CTRL_1__MST_MEM_LS_EN_MASK |
 				BIF1_PCIE_TX_POWER_CTRL_1__REPLAY_MEM_LS_EN_MASK);
@@ -272,7 +286,8 @@ static void nbio_v7_2_update_medium_grain_light_sleep(struct amdgpu_device *adev
 		if (def != data)
 			WREG32_PCIE_PORT(SOC15_REG_OFFSET(NBIO, 0, regBIF1_PCIE_TX_POWER_CTRL_1),
 				data);
-	} else {
+		break;
+	default:
 		def = data = RREG32_PCIE_PORT(SOC15_REG_OFFSET(NBIO, 0, regPCIE_CNTL2));
 		if (enable && (adev->cg_flags & AMD_CG_SUPPORT_BIF_LS))
 			data |= (PCIE_CNTL2__SLV_MEM_LS_EN_MASK |
@@ -285,11 +300,12 @@ static void nbio_v7_2_update_medium_grain_light_sleep(struct amdgpu_device *adev
 
 		if (def != data)
 			WREG32_PCIE_PORT(SOC15_REG_OFFSET(NBIO, 0, regPCIE_CNTL2), data);
+		break;
 	}
 }
 
 static void nbio_v7_2_get_clockgating_state(struct amdgpu_device *adev,
-					    u32 *flags)
+					    u64 *flags)
 {
 	int data;
 
@@ -352,7 +368,10 @@ const struct nbio_hdp_flush_reg nbio_v7_2_hdp_flush_reg = {
 static void nbio_v7_2_init_registers(struct amdgpu_device *adev)
 {
 	uint32_t def, data;
-	if (adev->asic_type == CHIP_YELLOW_CARP) {
+	switch (amdgpu_ip_version(adev, NBIO_HWIP, 0)) {
+	case IP_VERSION(7, 2, 1):
+	case IP_VERSION(7, 3, 0):
+	case IP_VERSION(7, 5, 0):
 		def = data = RREG32_PCIE_PORT(SOC15_REG_OFFSET(NBIO, 0, regBIF1_PCIE_MST_CTRL_3));
 		data = REG_SET_FIELD(data, BIF1_PCIE_MST_CTRL_3,
 			CI_SWUS_MAX_READ_REQUEST_SIZE_MODE, 1);
@@ -361,7 +380,8 @@ static void nbio_v7_2_init_registers(struct amdgpu_device *adev)
 
 		if (def != data)
 			WREG32_PCIE_PORT(SOC15_REG_OFFSET(NBIO, 0, regBIF1_PCIE_MST_CTRL_3), data);
-	} else {
+		break;
+	default:
 		def = data = RREG32_PCIE_PORT(SOC15_REG_OFFSET(NBIO, 0, regPCIE_CONFIG_CNTL));
 		data = REG_SET_FIELD(data, PCIE_CONFIG_CNTL,
 			CI_SWUS_MAX_READ_REQUEST_SIZE_MODE, 1);
@@ -370,11 +390,32 @@ static void nbio_v7_2_init_registers(struct amdgpu_device *adev)
 
 		if (def != data)
 			WREG32_PCIE_PORT(SOC15_REG_OFFSET(NBIO, 0, regPCIE_CONFIG_CNTL), data);
+		break;
 	}
 
-	if (amdgpu_sriov_vf(adev))
-		adev->rmmio_remap.reg_offset = SOC15_REG_OFFSET(NBIO, 0,
-			regBIF_BX_PF0_HDP_MEM_COHERENCY_FLUSH_CNTL) << 2;
+	switch (amdgpu_ip_version(adev, NBIO_HWIP, 0)) {
+	case IP_VERSION(7, 3, 0):
+	case IP_VERSION(7, 5, 1):
+		data = RREG32_SOC15(NBIO, 0, regRCC_DEV2_EPF0_STRAP2);
+		data &= ~RCC_DEV2_EPF0_STRAP2__STRAP_NO_SOFT_RESET_DEV2_F0_MASK;
+		WREG32_SOC15(NBIO, 0, regRCC_DEV2_EPF0_STRAP2, data);
+		break;
+	}
+}
+
+#define MMIO_REG_HOLE_OFFSET (0x80000 - PAGE_SIZE)
+
+static void nbio_v7_2_set_reg_remap(struct amdgpu_device *adev)
+{
+	if (!amdgpu_sriov_vf(adev) && (PAGE_SIZE <= 4096)) {
+		adev->rmmio_remap.reg_offset = MMIO_REG_HOLE_OFFSET;
+		adev->rmmio_remap.bus_addr = adev->rmmio_base + MMIO_REG_HOLE_OFFSET;
+	} else {
+		adev->rmmio_remap.reg_offset =
+			SOC15_REG_OFFSET(NBIO, 0,
+					 regBIF_BX_PF0_HDP_MEM_COHERENCY_FLUSH_CNTL) << 2;
+		adev->rmmio_remap.bus_addr = 0;
+	}
 }
 
 const struct amdgpu_nbio_funcs nbio_v7_2_funcs = {
@@ -398,4 +439,5 @@ const struct amdgpu_nbio_funcs nbio_v7_2_funcs = {
 	.ih_control = nbio_v7_2_ih_control,
 	.init_registers = nbio_v7_2_init_registers,
 	.remap_hdp_registers = nbio_v7_2_remap_hdp_registers,
+	.set_reg_remap = nbio_v7_2_set_reg_remap,
 };
diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c b/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c
index dc5e93756fea..860bc5cb03c8 100644
--- a/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c
+++ b/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c
@@ -21,7 +21,6 @@
  *
  */
 #include "amdgpu.h"
-#include "amdgpu_atombios.h"
 #include "nbio_v7_4.h"
 #include "amdgpu_ras.h"
 
@@ -152,9 +151,9 @@ static void nbio_v7_4_sdma_doorbell_range(struct amdgpu_device *adev, int instan
 		 *   BIF_SDMA0_DOORBELL_RANGE:  0x3bc0
 		 *   BIF_SDMA1_DOORBELL_RANGE:  0x3bc4
 		 *   BIF_SDMA2_DOORBELL_RANGE:  0x3bd8
-+		 *   BIF_SDMA4_DOORBELL_RANGE:
-+		 *     ARCTURUS:  0x3be0
-+		 *     ALDEBARAN: 0x3be4
+		 *   BIF_SDMA4_DOORBELL_RANGE:
+		 *     ARCTURUS:  0x3be0
+		 *     ALDEBARAN: 0x3be4
 		 */
 		if (adev->asic_type == CHIP_ALDEBARAN && instance == 4)
 			reg = instance + 0x4 + 0x1 +
@@ -238,7 +237,7 @@ static void nbio_v7_4_ih_doorbell_range(struct amdgpu_device *adev,
 
 	if (use_doorbell) {
 		ih_doorbell_range = REG_SET_FIELD(ih_doorbell_range, BIF_IH_DOORBELL_RANGE, OFFSET, doorbell_index);
-		ih_doorbell_range = REG_SET_FIELD(ih_doorbell_range, BIF_IH_DOORBELL_RANGE, SIZE, 4);
+		ih_doorbell_range = REG_SET_FIELD(ih_doorbell_range, BIF_IH_DOORBELL_RANGE, SIZE, 8);
 	} else
 		ih_doorbell_range = REG_SET_FIELD(ih_doorbell_range, BIF_IH_DOORBELL_RANGE, SIZE, 0);
 
@@ -273,7 +272,7 @@ static void nbio_v7_4_update_medium_grain_light_sleep(struct amdgpu_device *adev
 }
 
 static void nbio_v7_4_get_clockgating_state(struct amdgpu_device *adev,
-					    u32 *flags)
+					    u64 *flags)
 {
 	int data;
 
@@ -339,41 +338,34 @@ const struct nbio_hdp_flush_reg nbio_v7_4_hdp_flush_reg = {
 	.ref_and_mask_sdma1 = GPU_HDP_FLUSH_DONE__SDMA1_MASK,
 };
 
-const struct nbio_hdp_flush_reg nbio_v7_4_hdp_flush_reg_ald = {
-	.ref_and_mask_cp0 = GPU_HDP_FLUSH_DONE__CP0_MASK,
-	.ref_and_mask_cp1 = GPU_HDP_FLUSH_DONE__CP1_MASK,
-	.ref_and_mask_cp2 = GPU_HDP_FLUSH_DONE__CP2_MASK,
-	.ref_and_mask_cp3 = GPU_HDP_FLUSH_DONE__CP3_MASK,
-	.ref_and_mask_cp4 = GPU_HDP_FLUSH_DONE__CP4_MASK,
-	.ref_and_mask_cp5 = GPU_HDP_FLUSH_DONE__CP5_MASK,
-	.ref_and_mask_cp6 = GPU_HDP_FLUSH_DONE__CP6_MASK,
-	.ref_and_mask_cp7 = GPU_HDP_FLUSH_DONE__CP7_MASK,
-	.ref_and_mask_cp8 = GPU_HDP_FLUSH_DONE__CP8_MASK,
-	.ref_and_mask_cp9 = GPU_HDP_FLUSH_DONE__CP9_MASK,
-	.ref_and_mask_sdma0 = GPU_HDP_FLUSH_DONE__RSVD_ENG1_MASK,
-	.ref_and_mask_sdma1 = GPU_HDP_FLUSH_DONE__RSVD_ENG2_MASK,
-	.ref_and_mask_sdma2 = GPU_HDP_FLUSH_DONE__RSVD_ENG3_MASK,
-	.ref_and_mask_sdma3 = GPU_HDP_FLUSH_DONE__RSVD_ENG4_MASK,
-	.ref_and_mask_sdma4 = GPU_HDP_FLUSH_DONE__RSVD_ENG5_MASK,
-	.ref_and_mask_sdma5 = GPU_HDP_FLUSH_DONE__RSVD_ENG6_MASK,
-	.ref_and_mask_sdma6 = GPU_HDP_FLUSH_DONE__RSVD_ENG7_MASK,
-	.ref_and_mask_sdma7 = GPU_HDP_FLUSH_DONE__RSVD_ENG8_MASK,
-};
-
 static void nbio_v7_4_init_registers(struct amdgpu_device *adev)
 {
-	if (amdgpu_sriov_vf(adev))
-		adev->rmmio_remap.reg_offset = SOC15_REG_OFFSET(NBIO, 0,
-			mmBIF_BX_DEV0_EPF0_VF0_HDP_MEM_COHERENCY_FLUSH_CNTL) << 2;
+	uint32_t baco_cntl;
+
+	if (amdgpu_ip_version(adev, NBIO_HWIP, 0) == IP_VERSION(7, 4, 4) &&
+	    !amdgpu_sriov_vf(adev)) {
+		baco_cntl = RREG32_SOC15(NBIO, 0, mmBACO_CNTL);
+		if (baco_cntl &
+		    (BACO_CNTL__BACO_DUMMY_EN_MASK | BACO_CNTL__BACO_EN_MASK)) {
+			baco_cntl &= ~(BACO_CNTL__BACO_DUMMY_EN_MASK |
+				       BACO_CNTL__BACO_EN_MASK);
+			dev_dbg(adev->dev, "Unsetting baco dummy mode %x",
+				baco_cntl);
+			WREG32_SOC15(NBIO, 0, mmBACO_CNTL, baco_cntl);
+		}
+	}
 }
 
 static void nbio_v7_4_handle_ras_controller_intr_no_bifring(struct amdgpu_device *adev)
 {
 	uint32_t bif_doorbell_intr_cntl;
 	struct ras_manager *obj = amdgpu_ras_find_obj(adev, adev->nbio.ras_if);
-	struct ras_err_data err_data = {0, 0, 0, NULL};
+	struct ras_err_data err_data;
 	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
 
+	if (amdgpu_ras_error_data_init(&err_data))
+		return;
+
 	if (adev->asic_type == CHIP_ALDEBARAN)
 		bif_doorbell_intr_cntl = RREG32_SOC15(NBIO, 0, mmBIF_DOORBELL_INT_CNTL_ALDE);
 	else
@@ -390,7 +382,7 @@ static void nbio_v7_4_handle_ras_controller_intr_no_bifring(struct amdgpu_device
 		else
 			WREG32_SOC15(NBIO, 0, mmBIF_DOORBELL_INT_CNTL, bif_doorbell_intr_cntl);
 
-		if (!ras->disable_ras_err_cnt_harvest) {
+		if (ras && !ras->disable_ras_err_cnt_harvest && obj) {
 			/*
 			 * clear error status after ras_controller_intr
 			 * according to hw team and count ue number
@@ -404,8 +396,7 @@ static void nbio_v7_4_handle_ras_controller_intr_no_bifring(struct amdgpu_device
 
 			if (err_data.ce_count)
 				dev_info(adev->dev, "%ld correctable hardware "
-						"errors detected in %s block, "
-						"no user action is needed.\n",
+						"errors detected in %s block\n",
 						obj->err_data.ce_count,
 						get_ras_block_str(adev->nbio.ras_if));
 
@@ -422,8 +413,10 @@ static void nbio_v7_4_handle_ras_controller_intr_no_bifring(struct amdgpu_device
 		/* ras_controller_int is dedicated for nbif ras error,
 		 * not the global interrupt for sync flood
 		 */
-		amdgpu_ras_reset_gpu(adev);
+		amdgpu_ras_global_ras_isr(adev);
 	}
+
+	amdgpu_ras_error_data_fini(&err_data);
 }
 
 static void nbio_v7_4_handle_ras_err_event_athub_intr_no_bifring(struct amdgpu_device *adev)
@@ -658,16 +651,28 @@ static void nbio_v7_4_enable_doorbell_interrupt(struct amdgpu_device *adev,
 		       DOORBELL_INTERRUPT_DISABLE, enable ? 0 : 1);
 }
 
-const struct amdgpu_nbio_ras_funcs nbio_v7_4_ras_funcs = {
+const struct amdgpu_ras_block_hw_ops nbio_v7_4_ras_hw_ops = {
+	.query_ras_error_count = nbio_v7_4_query_ras_error_count,
+};
+
+struct amdgpu_nbio_ras nbio_v7_4_ras = {
+	.ras_block = {
+		.ras_comm = {
+			.name = "pcie_bif",
+			.block = AMDGPU_RAS_BLOCK__PCIE_BIF,
+			.type = AMDGPU_RAS_ERROR__MULTI_UNCORRECTABLE,
+		},
+		.hw_ops = &nbio_v7_4_ras_hw_ops,
+		.ras_late_init = amdgpu_nbio_ras_late_init,
+	},
 	.handle_ras_controller_intr_no_bifring = nbio_v7_4_handle_ras_controller_intr_no_bifring,
 	.handle_ras_err_event_athub_intr_no_bifring = nbio_v7_4_handle_ras_err_event_athub_intr_no_bifring,
 	.init_ras_controller_interrupt = nbio_v7_4_init_ras_controller_interrupt,
 	.init_ras_err_event_athub_interrupt = nbio_v7_4_init_ras_err_event_athub_interrupt,
-	.query_ras_error_count = nbio_v7_4_query_ras_error_count,
-	.ras_late_init = amdgpu_nbio_ras_late_init,
-	.ras_fini = amdgpu_nbio_ras_fini,
 };
 
+
+#ifdef CONFIG_PCIEASPM
 static void nbio_v7_4_program_ltr(struct amdgpu_device *adev)
 {
 	uint32_t def, data;
@@ -689,12 +694,14 @@ static void nbio_v7_4_program_ltr(struct amdgpu_device *adev)
 	if (def != data)
 		WREG32_PCIE(smnBIF_CFG_DEV0_EPF0_DEVICE_CNTL2, data);
 }
+#endif
 
 static void nbio_v7_4_program_aspm(struct amdgpu_device *adev)
 {
+#ifdef CONFIG_PCIEASPM
 	uint32_t def, data;
 
-	if (adev->ip_versions[NBIO_HWIP][0] == IP_VERSION(7, 4, 4))
+	if (amdgpu_ip_version(adev, NBIO_HWIP, 0) == IP_VERSION(7, 4, 4))
 		return;
 
 	def = data = RREG32_PCIE(smnPCIE_LC_CNTL);
@@ -750,7 +757,10 @@ static void nbio_v7_4_program_aspm(struct amdgpu_device *adev)
 	if (def != data)
 		WREG32_PCIE(smnPCIE_LC_CNTL6, data);
 
-	nbio_v7_4_program_ltr(adev);
+	/* Don't bother about LTR if LTR is not enabled
+	 * in the path */
+	if (adev->pdev->ltr_path)
+		nbio_v7_4_program_ltr(adev);
 
 	def = data = RREG32_PCIE(smnRCC_BIF_STRAP3);
 	data |= 0x5DE0 << RCC_BIF_STRAP3__STRAP_VLINK_ASPM_IDLE_TIMER__SHIFT;
@@ -774,6 +784,22 @@ static void nbio_v7_4_program_aspm(struct amdgpu_device *adev)
 	data &= ~PCIE_LC_CNTL3__LC_DSC_DONT_ENTER_L23_AFTER_PME_ACK_MASK;
 	if (def != data)
 		WREG32_PCIE(smnPCIE_LC_CNTL3, data);
+#endif
+}
+
+#define MMIO_REG_HOLE_OFFSET (0x80000 - PAGE_SIZE)
+
+static void nbio_v7_4_set_reg_remap(struct amdgpu_device *adev)
+{
+	if (!amdgpu_sriov_vf(adev) && (PAGE_SIZE <= 4096)) {
+		adev->rmmio_remap.reg_offset = MMIO_REG_HOLE_OFFSET;
+		adev->rmmio_remap.bus_addr = adev->rmmio_base + MMIO_REG_HOLE_OFFSET;
+	} else {
+		adev->rmmio_remap.reg_offset =
+			SOC15_REG_OFFSET(NBIO, 0,
+					 mmBIF_BX_DEV0_EPF0_VF0_HDP_MEM_COHERENCY_FLUSH_CNTL) << 2;
+		adev->rmmio_remap.bus_addr = 0;
+	}
 }
 
 const struct amdgpu_nbio_funcs nbio_v7_4_funcs = {
@@ -797,4 +823,5 @@ const struct amdgpu_nbio_funcs nbio_v7_4_funcs = {
 	.init_registers = nbio_v7_4_init_registers,
 	.remap_hdp_registers = nbio_v7_4_remap_hdp_registers,
 	.program_aspm =  nbio_v7_4_program_aspm,
+	.set_reg_remap = nbio_v7_4_set_reg_remap,
 };
diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.h b/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.h
index cc5692db6f98..f27c41728822 100644
--- a/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.h
+++ b/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.h
@@ -27,8 +27,7 @@
 #include "soc15_common.h"
 
 extern const struct nbio_hdp_flush_reg nbio_v7_4_hdp_flush_reg;
-extern const struct nbio_hdp_flush_reg nbio_v7_4_hdp_flush_reg_ald;
 extern const struct amdgpu_nbio_funcs nbio_v7_4_funcs;
-extern const struct amdgpu_nbio_ras_funcs nbio_v7_4_ras_funcs;
+extern struct amdgpu_nbio_ras nbio_v7_4_ras;
 
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v7_7.c b/drivers/gpu/drm/amd/amdgpu/nbio_v7_7.c
new file mode 100644
index 000000000000..2ee60b8746a6
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/nbio_v7_7.c
@@ -0,0 +1,369 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include "amdgpu.h"
+#include "nbio_v7_7.h"
+
+#include "nbio/nbio_7_7_0_offset.h"
+#include "nbio/nbio_7_7_0_sh_mask.h"
+#include <uapi/linux/kfd_ioctl.h>
+
+static void nbio_v7_7_remap_hdp_registers(struct amdgpu_device *adev)
+{
+	WREG32_SOC15(NBIO, 0, regBIF_BX0_REMAP_HDP_MEM_FLUSH_CNTL,
+		     adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL);
+	WREG32_SOC15(NBIO, 0, regBIF_BX0_REMAP_HDP_REG_FLUSH_CNTL,
+		     adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_REG_FLUSH_CNTL);
+}
+
+static u32 nbio_v7_7_get_rev_id(struct amdgpu_device *adev)
+{
+	u32 tmp;
+
+	tmp = RREG32_SOC15(NBIO, 0, regRCC_STRAP0_RCC_DEV0_EPF0_STRAP0);
+	tmp &= RCC_STRAP0_RCC_DEV0_EPF0_STRAP0__STRAP_ATI_REV_ID_DEV0_F0_MASK;
+	tmp >>= RCC_STRAP0_RCC_DEV0_EPF0_STRAP0__STRAP_ATI_REV_ID_DEV0_F0__SHIFT;
+
+	return tmp;
+}
+
+static void nbio_v7_7_mc_access_enable(struct amdgpu_device *adev, bool enable)
+{
+	if (enable)
+		WREG32_SOC15(NBIO, 0, regBIF_BX1_BIF_FB_EN,
+			BIF_BX1_BIF_FB_EN__FB_READ_EN_MASK |
+			BIF_BX1_BIF_FB_EN__FB_WRITE_EN_MASK);
+	else
+		WREG32_SOC15(NBIO, 0, regBIF_BX1_BIF_FB_EN, 0);
+}
+
+static u32 nbio_v7_7_get_memsize(struct amdgpu_device *adev)
+{
+	return RREG32_SOC15(NBIO, 0, regRCC_DEV0_EPF0_0_RCC_CONFIG_MEMSIZE);
+}
+
+static void nbio_v7_7_sdma_doorbell_range(struct amdgpu_device *adev, int instance,
+					  bool use_doorbell, int doorbell_index,
+					  int doorbell_size)
+{
+	u32 reg = SOC15_REG_OFFSET(NBIO, 0, regGDC0_BIF_CSDMA_DOORBELL_RANGE);
+	u32 doorbell_range = RREG32_PCIE_PORT(reg);
+
+	if (use_doorbell) {
+		doorbell_range = REG_SET_FIELD(doorbell_range,
+					       GDC0_BIF_CSDMA_DOORBELL_RANGE,
+					       OFFSET, doorbell_index);
+		doorbell_range = REG_SET_FIELD(doorbell_range,
+					       GDC0_BIF_CSDMA_DOORBELL_RANGE,
+					       SIZE, doorbell_size);
+	} else {
+		doorbell_range = REG_SET_FIELD(doorbell_range,
+					       GDC0_BIF_SDMA0_DOORBELL_RANGE,
+					       SIZE, 0);
+	}
+
+	WREG32_PCIE_PORT(reg, doorbell_range);
+}
+
+static void nbio_v7_7_vcn_doorbell_range(struct amdgpu_device *adev, bool use_doorbell,
+					int doorbell_index, int instance)
+{
+	u32 reg = SOC15_REG_OFFSET(NBIO, 0, regGDC0_BIF_VCN0_DOORBELL_RANGE);
+	u32 doorbell_range = RREG32_PCIE_PORT(reg);
+
+	if (use_doorbell) {
+		doorbell_range = REG_SET_FIELD(doorbell_range,
+					       GDC0_BIF_VCN0_DOORBELL_RANGE, OFFSET,
+					       doorbell_index);
+		doorbell_range = REG_SET_FIELD(doorbell_range,
+					       GDC0_BIF_VCN0_DOORBELL_RANGE, SIZE, 8);
+	} else {
+		doorbell_range = REG_SET_FIELD(doorbell_range,
+					       GDC0_BIF_VCN0_DOORBELL_RANGE, SIZE, 0);
+	}
+
+	WREG32_PCIE_PORT(reg, doorbell_range);
+}
+
+static void nbio_v7_7_enable_doorbell_aperture(struct amdgpu_device *adev,
+					       bool enable)
+{
+	u32 reg;
+
+	reg = RREG32_SOC15(NBIO, 0, regRCC_DEV0_EPF0_0_RCC_DOORBELL_APER_EN);
+	reg = REG_SET_FIELD(reg, RCC_DEV0_EPF0_0_RCC_DOORBELL_APER_EN,
+			    BIF_DOORBELL_APER_EN, enable ? 1 : 0);
+
+	WREG32_SOC15(NBIO, 0, regRCC_DEV0_EPF0_0_RCC_DOORBELL_APER_EN, reg);
+}
+
+static void nbio_v7_7_enable_doorbell_selfring_aperture(struct amdgpu_device *adev,
+							bool enable)
+{
+	u32 tmp = 0;
+
+	if (enable) {
+		tmp = REG_SET_FIELD(tmp, BIF_BX_PF0_DOORBELL_SELFRING_GPA_APER_CNTL,
+				DOORBELL_SELFRING_GPA_APER_EN, 1) |
+			REG_SET_FIELD(tmp, BIF_BX_PF0_DOORBELL_SELFRING_GPA_APER_CNTL,
+				DOORBELL_SELFRING_GPA_APER_MODE, 1) |
+			REG_SET_FIELD(tmp, BIF_BX_PF0_DOORBELL_SELFRING_GPA_APER_CNTL,
+				DOORBELL_SELFRING_GPA_APER_SIZE, 0);
+
+		WREG32_SOC15(NBIO, 0,
+			regBIF_BX_PF0_DOORBELL_SELFRING_GPA_APER_BASE_LOW,
+			lower_32_bits(adev->doorbell.base));
+		WREG32_SOC15(NBIO, 0,
+			regBIF_BX_PF0_DOORBELL_SELFRING_GPA_APER_BASE_HIGH,
+			upper_32_bits(adev->doorbell.base));
+	}
+
+	WREG32_SOC15(NBIO, 0, regBIF_BX_PF0_DOORBELL_SELFRING_GPA_APER_CNTL,
+		tmp);
+}
+
+
+static void nbio_v7_7_ih_doorbell_range(struct amdgpu_device *adev,
+					bool use_doorbell, int doorbell_index)
+{
+	u32 ih_doorbell_range = RREG32_SOC15(NBIO, 0,
+								regGDC0_BIF_IH_DOORBELL_RANGE);
+
+	if (use_doorbell) {
+		ih_doorbell_range = REG_SET_FIELD(ih_doorbell_range,
+						  GDC0_BIF_IH_DOORBELL_RANGE, OFFSET,
+						  doorbell_index);
+		ih_doorbell_range = REG_SET_FIELD(ih_doorbell_range,
+						  GDC0_BIF_IH_DOORBELL_RANGE, SIZE,
+						  2);
+	} else {
+		ih_doorbell_range = REG_SET_FIELD(ih_doorbell_range,
+						  GDC0_BIF_IH_DOORBELL_RANGE, SIZE,
+						  0);
+	}
+
+	WREG32_SOC15(NBIO, 0, regGDC0_BIF_IH_DOORBELL_RANGE,
+			 ih_doorbell_range);
+}
+
+static void nbio_v7_7_ih_control(struct amdgpu_device *adev)
+{
+	u32 interrupt_cntl;
+
+	/* setup interrupt control */
+	WREG32_SOC15(NBIO, 0, regBIF_BX1_INTERRUPT_CNTL2,
+		     adev->dummy_page_addr >> 8);
+
+	interrupt_cntl = RREG32_SOC15(NBIO, 0, regBIF_BX1_INTERRUPT_CNTL);
+	/*
+	 * INTERRUPT_CNTL__IH_DUMMY_RD_OVERRIDE_MASK=0 - dummy read disabled with msi, enabled without msi
+	 * INTERRUPT_CNTL__IH_DUMMY_RD_OVERRIDE_MASK=1 - dummy read controlled by IH_DUMMY_RD_EN
+	 */
+	interrupt_cntl = REG_SET_FIELD(interrupt_cntl, BIF_BX1_INTERRUPT_CNTL,
+				       IH_DUMMY_RD_OVERRIDE, 0);
+
+	/* INTERRUPT_CNTL__IH_REQ_NONSNOOP_EN_MASK=1 if ring is in non-cacheable memory, e.g., vram */
+	interrupt_cntl = REG_SET_FIELD(interrupt_cntl, BIF_BX1_INTERRUPT_CNTL,
+				       IH_REQ_NONSNOOP_EN, 0);
+
+	WREG32_SOC15(NBIO, 0, regBIF_BX1_INTERRUPT_CNTL, interrupt_cntl);
+}
+
+static u32 nbio_v7_7_get_hdp_flush_req_offset(struct amdgpu_device *adev)
+{
+	return SOC15_REG_OFFSET(NBIO, 0, regBIF_BX_PF0_GPU_HDP_FLUSH_REQ);
+}
+
+static u32 nbio_v7_7_get_hdp_flush_done_offset(struct amdgpu_device *adev)
+{
+	return SOC15_REG_OFFSET(NBIO, 0, regBIF_BX_PF0_GPU_HDP_FLUSH_DONE);
+}
+
+static u32 nbio_v7_7_get_pcie_index_offset(struct amdgpu_device *adev)
+{
+	return SOC15_REG_OFFSET(NBIO, 0, regBIF_BX0_PCIE_INDEX2);
+}
+
+static u32 nbio_v7_7_get_pcie_data_offset(struct amdgpu_device *adev)
+{
+	return SOC15_REG_OFFSET(NBIO, 0, regBIF_BX0_PCIE_DATA2);
+}
+
+static u32 nbio_v7_7_get_pcie_port_index_offset(struct amdgpu_device *adev)
+{
+	return SOC15_REG_OFFSET(NBIO, 0, regBIF_BX_PF0_RSMU_INDEX);
+}
+
+static u32 nbio_v7_7_get_pcie_port_data_offset(struct amdgpu_device *adev)
+{
+	return SOC15_REG_OFFSET(NBIO, 0, regBIF_BX_PF0_RSMU_DATA);
+}
+
+const struct nbio_hdp_flush_reg nbio_v7_7_hdp_flush_reg = {
+	.ref_and_mask_cp0 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__CP0_MASK,
+	.ref_and_mask_cp1 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__CP1_MASK,
+	.ref_and_mask_cp2 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__CP2_MASK,
+	.ref_and_mask_cp3 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__CP3_MASK,
+	.ref_and_mask_cp4 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__CP4_MASK,
+	.ref_and_mask_cp5 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__CP5_MASK,
+	.ref_and_mask_cp6 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__CP6_MASK,
+	.ref_and_mask_cp7 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__CP7_MASK,
+	.ref_and_mask_cp8 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__CP8_MASK,
+	.ref_and_mask_cp9 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__CP9_MASK,
+	.ref_and_mask_sdma0 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__SDMA0_MASK,
+	.ref_and_mask_sdma1 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__SDMA1_MASK,
+};
+
+static void nbio_v7_7_init_registers(struct amdgpu_device *adev)
+{
+	uint32_t def, data;
+
+	def = data = RREG32_SOC15(NBIO, 0, regBIF0_PCIE_MST_CTRL_3);
+	data = REG_SET_FIELD(data, BIF0_PCIE_MST_CTRL_3,
+			     CI_SWUS_MAX_READ_REQUEST_SIZE_MODE, 1);
+	data = REG_SET_FIELD(data, BIF0_PCIE_MST_CTRL_3,
+			     CI_SWUS_MAX_READ_REQUEST_SIZE_PRIV, 1);
+
+	if (def != data)
+		WREG32_SOC15(NBIO, 0, regBIF0_PCIE_MST_CTRL_3, data);
+
+	switch (amdgpu_ip_version(adev, NBIO_HWIP, 0)) {
+	case IP_VERSION(7, 7, 0):
+		data = RREG32_SOC15(NBIO, 0, regRCC_DEV0_EPF5_STRAP4) & ~BIT(23);
+		WREG32_SOC15(NBIO, 0, regRCC_DEV0_EPF5_STRAP4, data);
+		break;
+	}
+}
+
+static void nbio_v7_7_update_medium_grain_clock_gating(struct amdgpu_device *adev,
+						       bool enable)
+{
+	uint32_t def, data;
+
+	if (!(adev->cg_flags & AMD_CG_SUPPORT_BIF_MGCG))
+		return;
+
+	def = data = RREG32_SOC15(NBIO, 0, regBIF0_CPM_CONTROL);
+	if (enable) {
+		data |= (BIF0_CPM_CONTROL__LCLK_DYN_GATE_ENABLE_MASK |
+			 BIF0_CPM_CONTROL__TXCLK_DYN_GATE_ENABLE_MASK |
+			 BIF0_CPM_CONTROL__TXCLK_LCNT_GATE_ENABLE_MASK |
+			 BIF0_CPM_CONTROL__TXCLK_REGS_GATE_ENABLE_MASK |
+			 BIF0_CPM_CONTROL__TXCLK_PRBS_GATE_ENABLE_MASK |
+			 BIF0_CPM_CONTROL__REFCLK_REGS_GATE_ENABLE_MASK);
+	} else {
+		data &= ~(BIF0_CPM_CONTROL__LCLK_DYN_GATE_ENABLE_MASK |
+			  BIF0_CPM_CONTROL__TXCLK_DYN_GATE_ENABLE_MASK |
+			  BIF0_CPM_CONTROL__TXCLK_LCNT_GATE_ENABLE_MASK |
+			  BIF0_CPM_CONTROL__TXCLK_REGS_GATE_ENABLE_MASK |
+			  BIF0_CPM_CONTROL__TXCLK_PRBS_GATE_ENABLE_MASK |
+			  BIF0_CPM_CONTROL__REFCLK_REGS_GATE_ENABLE_MASK);
+	}
+
+	if (def != data)
+		WREG32_SOC15(NBIO, 0, regBIF0_CPM_CONTROL, data);
+}
+
+static void nbio_v7_7_update_medium_grain_light_sleep(struct amdgpu_device *adev,
+						      bool enable)
+{
+	uint32_t def, data;
+
+	if (!(adev->cg_flags & AMD_CG_SUPPORT_BIF_LS))
+		return;
+
+	def = data = RREG32_SOC15(NBIO, 0, regBIF0_PCIE_CNTL2);
+	if (enable)
+		data |= BIF0_PCIE_CNTL2__SLV_MEM_LS_EN_MASK;
+	else
+		data &= ~BIF0_PCIE_CNTL2__SLV_MEM_LS_EN_MASK;
+
+	if (def != data)
+		WREG32_SOC15(NBIO, 0, regBIF0_PCIE_CNTL2, data);
+
+	def = data = RREG32_SOC15(NBIO, 0, regBIF0_PCIE_TX_POWER_CTRL_1);
+	if (enable) {
+		data |= (BIF0_PCIE_TX_POWER_CTRL_1__MST_MEM_LS_EN_MASK |
+			BIF0_PCIE_TX_POWER_CTRL_1__REPLAY_MEM_LS_EN_MASK);
+	} else {
+		data &= ~(BIF0_PCIE_TX_POWER_CTRL_1__MST_MEM_LS_EN_MASK |
+			BIF0_PCIE_TX_POWER_CTRL_1__REPLAY_MEM_LS_EN_MASK);
+	}
+
+	if (def != data)
+		WREG32_SOC15(NBIO, 0, regBIF0_PCIE_TX_POWER_CTRL_1, data);
+}
+
+static void nbio_v7_7_get_clockgating_state(struct amdgpu_device *adev,
+					    u64 *flags)
+{
+	uint32_t data;
+
+	/* AMD_CG_SUPPORT_BIF_MGCG */
+	data = RREG32_SOC15(NBIO, 0, regBIF0_CPM_CONTROL);
+	if (data & BIF0_CPM_CONTROL__LCLK_DYN_GATE_ENABLE_MASK)
+		*flags |= AMD_CG_SUPPORT_BIF_MGCG;
+
+	/* AMD_CG_SUPPORT_BIF_LS */
+	data = RREG32_SOC15(NBIO, 0, regBIF0_PCIE_CNTL2);
+	if (data & BIF0_PCIE_CNTL2__SLV_MEM_LS_EN_MASK)
+		*flags |= AMD_CG_SUPPORT_BIF_LS;
+}
+
+#define MMIO_REG_HOLE_OFFSET (0x80000 - PAGE_SIZE)
+
+static void nbio_v7_7_set_reg_remap(struct amdgpu_device *adev)
+{
+	if (!amdgpu_sriov_vf(adev) && (PAGE_SIZE <= 4096)) {
+		adev->rmmio_remap.reg_offset = MMIO_REG_HOLE_OFFSET;
+		adev->rmmio_remap.bus_addr = adev->rmmio_base + MMIO_REG_HOLE_OFFSET;
+	} else {
+		adev->rmmio_remap.reg_offset =
+			SOC15_REG_OFFSET(NBIO, 0,
+					 regBIF_BX_PF0_HDP_MEM_COHERENCY_FLUSH_CNTL) << 2;
+		adev->rmmio_remap.bus_addr = 0;
+	}
+}
+
+const struct amdgpu_nbio_funcs nbio_v7_7_funcs = {
+	.get_hdp_flush_req_offset = nbio_v7_7_get_hdp_flush_req_offset,
+	.get_hdp_flush_done_offset = nbio_v7_7_get_hdp_flush_done_offset,
+	.get_pcie_index_offset = nbio_v7_7_get_pcie_index_offset,
+	.get_pcie_data_offset = nbio_v7_7_get_pcie_data_offset,
+	.get_pcie_port_index_offset = nbio_v7_7_get_pcie_port_index_offset,
+	.get_pcie_port_data_offset = nbio_v7_7_get_pcie_port_data_offset,
+	.get_rev_id = nbio_v7_7_get_rev_id,
+	.mc_access_enable = nbio_v7_7_mc_access_enable,
+	.get_memsize = nbio_v7_7_get_memsize,
+	.sdma_doorbell_range = nbio_v7_7_sdma_doorbell_range,
+	.vcn_doorbell_range = nbio_v7_7_vcn_doorbell_range,
+	.enable_doorbell_aperture = nbio_v7_7_enable_doorbell_aperture,
+	.enable_doorbell_selfring_aperture = nbio_v7_7_enable_doorbell_selfring_aperture,
+	.ih_doorbell_range = nbio_v7_7_ih_doorbell_range,
+	.update_medium_grain_clock_gating = nbio_v7_7_update_medium_grain_clock_gating,
+	.update_medium_grain_light_sleep = nbio_v7_7_update_medium_grain_light_sleep,
+	.get_clockgating_state = nbio_v7_7_get_clockgating_state,
+	.ih_control = nbio_v7_7_ih_control,
+	.init_registers = nbio_v7_7_init_registers,
+	.remap_hdp_registers = nbio_v7_7_remap_hdp_registers,
+	.set_reg_remap = nbio_v7_7_set_reg_remap,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v7_7.h b/drivers/gpu/drm/amd/amdgpu/nbio_v7_7.h
new file mode 100644
index 000000000000..2a33b256ba81
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/nbio_v7_7.h
@@ -0,0 +1,33 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __NBIO_V7_7_H__
+#define __NBIO_V7_7_H__
+
+#include "soc15_common.h"
+
+extern const struct nbio_hdp_flush_reg nbio_v7_7_hdp_flush_reg;
+extern const struct amdgpu_nbio_funcs nbio_v7_7_funcs;
+extern const struct amdgpu_nbio_ras_funcs nbio_v7_7_ras_funcs;
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v7_9.c b/drivers/gpu/drm/amd/amdgpu/nbio_v7_9.c
new file mode 100644
index 000000000000..bdfd2917e3ca
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/nbio_v7_9.c
@@ -0,0 +1,696 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include "amdgpu.h"
+#include "nbio_v7_9.h"
+#include "amdgpu_ras.h"
+
+#include "nbio/nbio_7_9_0_offset.h"
+#include "nbio/nbio_7_9_0_sh_mask.h"
+#include "ivsrcid/nbio/irqsrcs_nbif_7_4.h"
+#include <uapi/linux/kfd_ioctl.h>
+
+#define NPS_MODE_MASK 0x000000FFL
+
+static void nbio_v7_9_remap_hdp_registers(struct amdgpu_device *adev)
+{
+	WREG32_SOC15(NBIO, 0, regBIF_BX0_REMAP_HDP_MEM_FLUSH_CNTL,
+		adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL);
+	WREG32_SOC15(NBIO, 0, regBIF_BX0_REMAP_HDP_REG_FLUSH_CNTL,
+		adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_REG_FLUSH_CNTL);
+}
+
+static u32 nbio_v7_9_get_rev_id(struct amdgpu_device *adev)
+{
+	u32 rev_id;
+
+	/*
+	 * fetch the sub-revision field from the IP-discovery table
+	 * (returns zero if the table entry is not populated).
+	 */
+	if (amdgpu_sriov_vf(adev)) {
+		rev_id = IP_VERSION_SUBREV(amdgpu_ip_version_full(adev, NBIO_HWIP, 0));
+	} else {
+		rev_id = RREG32_SOC15(NBIO, 0, regRCC_STRAP0_RCC_DEV0_EPF0_STRAP0);
+		rev_id = REG_GET_FIELD(rev_id, RCC_STRAP0_RCC_DEV0_EPF0_STRAP0,
+				STRAP_ATI_REV_ID_DEV0_F0);
+	}
+
+	return rev_id;
+}
+
+static void nbio_v7_9_mc_access_enable(struct amdgpu_device *adev, bool enable)
+{
+	if (enable)
+		WREG32_SOC15(NBIO, 0, regBIF_BX0_BIF_FB_EN,
+			BIF_BX0_BIF_FB_EN__FB_READ_EN_MASK | BIF_BX0_BIF_FB_EN__FB_WRITE_EN_MASK);
+	else
+		WREG32_SOC15(NBIO, 0, regBIF_BX0_BIF_FB_EN, 0);
+}
+
+static u32 nbio_v7_9_get_memsize(struct amdgpu_device *adev)
+{
+	return RREG32_SOC15(NBIO, 0, regRCC_DEV0_EPF0_RCC_CONFIG_MEMSIZE);
+}
+
+static void nbio_v7_9_sdma_doorbell_range(struct amdgpu_device *adev, int instance,
+			bool use_doorbell, int doorbell_index, int doorbell_size)
+{
+	u32 doorbell_range = 0, doorbell_ctrl = 0;
+	int aid_id, dev_inst;
+
+	dev_inst = GET_INST(SDMA0, instance);
+	aid_id = adev->sdma.instance[instance].aid_id;
+
+	if (use_doorbell == false)
+		return;
+
+	doorbell_range =
+		REG_SET_FIELD(doorbell_range, DOORBELL0_CTRL_ENTRY_0,
+			BIF_DOORBELL0_RANGE_OFFSET_ENTRY, doorbell_index);
+	doorbell_range =
+		REG_SET_FIELD(doorbell_range, DOORBELL0_CTRL_ENTRY_0,
+			BIF_DOORBELL0_RANGE_SIZE_ENTRY, doorbell_size);
+	doorbell_ctrl =
+		REG_SET_FIELD(doorbell_ctrl, S2A_DOORBELL_ENTRY_1_CTRL,
+			S2A_DOORBELL_PORT1_ENABLE, 1);
+	doorbell_ctrl =
+		REG_SET_FIELD(doorbell_ctrl, S2A_DOORBELL_ENTRY_1_CTRL,
+			S2A_DOORBELL_PORT1_RANGE_SIZE, doorbell_size);
+
+	switch (dev_inst % adev->sdma.num_inst_per_aid) {
+	case 0:
+		WREG32_SOC15_OFFSET(NBIO, 0, regDOORBELL0_CTRL_ENTRY_1,
+			4 * aid_id, doorbell_range);
+
+		doorbell_ctrl = REG_SET_FIELD(doorbell_ctrl,
+					S2A_DOORBELL_ENTRY_1_CTRL,
+					S2A_DOORBELL_PORT1_AWID, 0xe);
+		doorbell_ctrl = REG_SET_FIELD(doorbell_ctrl,
+					S2A_DOORBELL_ENTRY_1_CTRL,
+					S2A_DOORBELL_PORT1_RANGE_OFFSET, 0xe);
+		doorbell_ctrl = REG_SET_FIELD(doorbell_ctrl,
+					S2A_DOORBELL_ENTRY_1_CTRL,
+					S2A_DOORBELL_PORT1_AWADDR_31_28_VALUE,
+					0x1);
+		WREG32_SOC15_EXT(NBIO, aid_id, regS2A_DOORBELL_ENTRY_1_CTRL,
+			aid_id, doorbell_ctrl);
+		break;
+	case 1:
+		WREG32_SOC15_OFFSET(NBIO, 0, regDOORBELL0_CTRL_ENTRY_2,
+			4 * aid_id, doorbell_range);
+
+		doorbell_ctrl = REG_SET_FIELD(doorbell_ctrl,
+					S2A_DOORBELL_ENTRY_1_CTRL,
+					S2A_DOORBELL_PORT1_AWID, 0x8);
+		doorbell_ctrl = REG_SET_FIELD(doorbell_ctrl,
+					S2A_DOORBELL_ENTRY_1_CTRL,
+					S2A_DOORBELL_PORT1_RANGE_OFFSET, 0x8);
+		doorbell_ctrl = REG_SET_FIELD(doorbell_ctrl,
+					S2A_DOORBELL_ENTRY_1_CTRL,
+					S2A_DOORBELL_PORT1_AWADDR_31_28_VALUE,
+					0x2);
+		WREG32_SOC15_EXT(NBIO, aid_id, regS2A_DOORBELL_ENTRY_2_CTRL,
+			aid_id, doorbell_ctrl);
+		break;
+	case 2:
+		WREG32_SOC15_OFFSET(NBIO, 0, regDOORBELL0_CTRL_ENTRY_3,
+			4 * aid_id, doorbell_range);
+
+		doorbell_ctrl = REG_SET_FIELD(doorbell_ctrl,
+					S2A_DOORBELL_ENTRY_1_CTRL,
+					S2A_DOORBELL_PORT1_AWID, 0x9);
+		doorbell_ctrl = REG_SET_FIELD(doorbell_ctrl,
+					S2A_DOORBELL_ENTRY_1_CTRL,
+					S2A_DOORBELL_PORT1_RANGE_OFFSET, 0x9);
+		doorbell_ctrl = REG_SET_FIELD(doorbell_ctrl,
+					S2A_DOORBELL_ENTRY_1_CTRL,
+					S2A_DOORBELL_PORT1_AWADDR_31_28_VALUE,
+					0x8);
+		WREG32_SOC15_EXT(NBIO, aid_id, regS2A_DOORBELL_ENTRY_5_CTRL,
+			aid_id, doorbell_ctrl);
+		break;
+	case 3:
+		WREG32_SOC15_OFFSET(NBIO, 0, regDOORBELL0_CTRL_ENTRY_4,
+			4 * aid_id, doorbell_range);
+
+		doorbell_ctrl = REG_SET_FIELD(doorbell_ctrl,
+					S2A_DOORBELL_ENTRY_1_CTRL,
+					S2A_DOORBELL_PORT1_AWID, 0xa);
+		doorbell_ctrl = REG_SET_FIELD(doorbell_ctrl,
+					S2A_DOORBELL_ENTRY_1_CTRL,
+					S2A_DOORBELL_PORT1_RANGE_OFFSET, 0xa);
+		doorbell_ctrl = REG_SET_FIELD(doorbell_ctrl,
+					S2A_DOORBELL_ENTRY_1_CTRL,
+					S2A_DOORBELL_PORT1_AWADDR_31_28_VALUE,
+					0x9);
+		WREG32_SOC15_EXT(NBIO, aid_id, regS2A_DOORBELL_ENTRY_6_CTRL,
+			aid_id, doorbell_ctrl);
+		break;
+	default:
+		break;
+	}
+}
+
+static void nbio_v7_9_vcn_doorbell_range(struct amdgpu_device *adev, bool use_doorbell,
+					 int doorbell_index, int instance)
+{
+	u32 doorbell_range = 0, doorbell_ctrl = 0;
+	u32 aid_id = instance;
+	u32 range_size;
+
+	if (use_doorbell) {
+		range_size = (amdgpu_ip_version(adev, GC_HWIP, 0) ==
+						IP_VERSION(9, 5, 0)) ?
+						0xb : 0x9;
+		doorbell_range = REG_SET_FIELD(doorbell_range,
+				DOORBELL0_CTRL_ENTRY_0,
+				BIF_DOORBELL0_RANGE_OFFSET_ENTRY,
+				doorbell_index);
+		doorbell_range = REG_SET_FIELD(doorbell_range,
+				DOORBELL0_CTRL_ENTRY_0,
+				BIF_DOORBELL0_RANGE_SIZE_ENTRY,
+				range_size);
+		if (aid_id)
+			doorbell_range = REG_SET_FIELD(doorbell_range,
+					DOORBELL0_CTRL_ENTRY_0,
+					DOORBELL0_FENCE_ENABLE_ENTRY,
+					0x4);
+
+		doorbell_ctrl = REG_SET_FIELD(doorbell_ctrl,
+				S2A_DOORBELL_ENTRY_1_CTRL,
+				S2A_DOORBELL_PORT1_ENABLE, 1);
+		doorbell_ctrl = REG_SET_FIELD(doorbell_ctrl,
+				S2A_DOORBELL_ENTRY_1_CTRL,
+				S2A_DOORBELL_PORT1_AWID, 0x4);
+		doorbell_ctrl = REG_SET_FIELD(doorbell_ctrl,
+				S2A_DOORBELL_ENTRY_1_CTRL,
+				S2A_DOORBELL_PORT1_RANGE_OFFSET, 0x4);
+		doorbell_ctrl = REG_SET_FIELD(doorbell_ctrl,
+				S2A_DOORBELL_ENTRY_1_CTRL,
+				S2A_DOORBELL_PORT1_RANGE_SIZE, range_size);
+		doorbell_ctrl = REG_SET_FIELD(doorbell_ctrl,
+				S2A_DOORBELL_ENTRY_1_CTRL,
+				S2A_DOORBELL_PORT1_AWADDR_31_28_VALUE, 0x4);
+
+		WREG32_SOC15_OFFSET(NBIO, 0, regDOORBELL0_CTRL_ENTRY_17,
+					aid_id, doorbell_range);
+		WREG32_SOC15_EXT(NBIO, aid_id, regS2A_DOORBELL_ENTRY_4_CTRL,
+				aid_id, doorbell_ctrl);
+	} else {
+		doorbell_range = REG_SET_FIELD(doorbell_range,
+				DOORBELL0_CTRL_ENTRY_0,
+				BIF_DOORBELL0_RANGE_SIZE_ENTRY, 0);
+		doorbell_ctrl = REG_SET_FIELD(doorbell_ctrl,
+				S2A_DOORBELL_ENTRY_1_CTRL,
+				S2A_DOORBELL_PORT1_RANGE_SIZE, 0);
+
+		WREG32_SOC15_OFFSET(NBIO, 0, regDOORBELL0_CTRL_ENTRY_17,
+					aid_id, doorbell_range);
+		WREG32_SOC15_EXT(NBIO, aid_id, regS2A_DOORBELL_ENTRY_4_CTRL,
+				aid_id, doorbell_ctrl);
+	}
+}
+
+static void nbio_v7_9_enable_doorbell_aperture(struct amdgpu_device *adev,
+					       bool enable)
+{
+	/* Enable to allow doorbell pass thru on pre-silicon bare-metal */
+	WREG32_SOC15(NBIO, 0, regBIFC_DOORBELL_ACCESS_EN_PF, 0xfffff);
+	WREG32_FIELD15_PREREG(NBIO, 0, RCC_DEV0_EPF0_RCC_DOORBELL_APER_EN,
+			BIF_DOORBELL_APER_EN, enable ? 1 : 0);
+}
+
+static void nbio_v7_9_enable_doorbell_selfring_aperture(struct amdgpu_device *adev,
+							bool enable)
+{
+	u32 tmp = 0;
+
+	if (enable) {
+		tmp = REG_SET_FIELD(tmp, BIF_BX_PF0_DOORBELL_SELFRING_GPA_APER_CNTL,
+				    DOORBELL_SELFRING_GPA_APER_EN, 1) |
+		      REG_SET_FIELD(tmp, BIF_BX_PF0_DOORBELL_SELFRING_GPA_APER_CNTL,
+				    DOORBELL_SELFRING_GPA_APER_MODE, 1) |
+		      REG_SET_FIELD(tmp, BIF_BX_PF0_DOORBELL_SELFRING_GPA_APER_CNTL,
+				    DOORBELL_SELFRING_GPA_APER_SIZE, 0);
+
+		WREG32_SOC15(NBIO, 0, regBIF_BX_PF0_DOORBELL_SELFRING_GPA_APER_BASE_LOW,
+			     lower_32_bits(adev->doorbell.base));
+		WREG32_SOC15(NBIO, 0, regBIF_BX_PF0_DOORBELL_SELFRING_GPA_APER_BASE_HIGH,
+			     upper_32_bits(adev->doorbell.base));
+	}
+
+	WREG32_SOC15(NBIO, 0, regBIF_BX_PF0_DOORBELL_SELFRING_GPA_APER_CNTL, tmp);
+}
+
+static void nbio_v7_9_ih_doorbell_range(struct amdgpu_device *adev,
+					bool use_doorbell, int doorbell_index)
+{
+	u32 ih_doorbell_range = 0, ih_doorbell_ctrl = 0;
+
+	if (use_doorbell) {
+		ih_doorbell_range = REG_SET_FIELD(ih_doorbell_range,
+				DOORBELL0_CTRL_ENTRY_0,
+				BIF_DOORBELL0_RANGE_OFFSET_ENTRY,
+				doorbell_index);
+		ih_doorbell_range = REG_SET_FIELD(ih_doorbell_range,
+				DOORBELL0_CTRL_ENTRY_0,
+				BIF_DOORBELL0_RANGE_SIZE_ENTRY,
+				0x8);
+
+		ih_doorbell_ctrl = REG_SET_FIELD(ih_doorbell_ctrl,
+				S2A_DOORBELL_ENTRY_1_CTRL,
+				S2A_DOORBELL_PORT1_ENABLE, 1);
+		ih_doorbell_ctrl = REG_SET_FIELD(ih_doorbell_ctrl,
+				S2A_DOORBELL_ENTRY_1_CTRL,
+				S2A_DOORBELL_PORT1_AWID, 0);
+		ih_doorbell_ctrl = REG_SET_FIELD(ih_doorbell_ctrl,
+				S2A_DOORBELL_ENTRY_1_CTRL,
+				S2A_DOORBELL_PORT1_RANGE_OFFSET, 0);
+		ih_doorbell_ctrl = REG_SET_FIELD(ih_doorbell_ctrl,
+				S2A_DOORBELL_ENTRY_1_CTRL,
+				S2A_DOORBELL_PORT1_RANGE_SIZE, 0x8);
+		ih_doorbell_ctrl = REG_SET_FIELD(ih_doorbell_ctrl,
+				S2A_DOORBELL_ENTRY_1_CTRL,
+				S2A_DOORBELL_PORT1_AWADDR_31_28_VALUE, 0);
+	} else {
+		ih_doorbell_range = REG_SET_FIELD(ih_doorbell_range,
+				DOORBELL0_CTRL_ENTRY_0,
+				BIF_DOORBELL0_RANGE_SIZE_ENTRY, 0);
+		ih_doorbell_ctrl = REG_SET_FIELD(ih_doorbell_ctrl,
+				S2A_DOORBELL_ENTRY_1_CTRL,
+				S2A_DOORBELL_PORT1_RANGE_SIZE, 0);
+	}
+
+	WREG32_SOC15(NBIO, 0, regDOORBELL0_CTRL_ENTRY_0, ih_doorbell_range);
+	WREG32_SOC15(NBIO, 0, regS2A_DOORBELL_ENTRY_3_CTRL, ih_doorbell_ctrl);
+}
+
+
+static void nbio_v7_9_update_medium_grain_clock_gating(struct amdgpu_device *adev,
+						       bool enable)
+{
+}
+
+static void nbio_v7_9_update_medium_grain_light_sleep(struct amdgpu_device *adev,
+						      bool enable)
+{
+}
+
+static void nbio_v7_9_get_clockgating_state(struct amdgpu_device *adev,
+					    u64 *flags)
+{
+}
+
+static void nbio_v7_9_ih_control(struct amdgpu_device *adev)
+{
+	u32 interrupt_cntl;
+
+	/* setup interrupt control */
+	WREG32_SOC15(NBIO, 0, regBIF_BX0_INTERRUPT_CNTL2, adev->dummy_page_addr >> 8);
+	interrupt_cntl = RREG32_SOC15(NBIO, 0, regBIF_BX0_INTERRUPT_CNTL);
+	/* INTERRUPT_CNTL__IH_DUMMY_RD_OVERRIDE_MASK=0 - dummy read disabled with msi, enabled without msi
+	 * INTERRUPT_CNTL__IH_DUMMY_RD_OVERRIDE_MASK=1 - dummy read controlled by IH_DUMMY_RD_EN
+	 */
+	interrupt_cntl =
+		REG_SET_FIELD(interrupt_cntl, BIF_BX0_INTERRUPT_CNTL, IH_DUMMY_RD_OVERRIDE, 0);
+	/* INTERRUPT_CNTL__IH_REQ_NONSNOOP_EN_MASK=1 if ring is in non-cacheable memory, e.g., vram */
+	interrupt_cntl =
+		REG_SET_FIELD(interrupt_cntl, BIF_BX0_INTERRUPT_CNTL, IH_REQ_NONSNOOP_EN, 0);
+	WREG32_SOC15(NBIO, 0, regBIF_BX0_INTERRUPT_CNTL, interrupt_cntl);
+}
+
+static u32 nbio_v7_9_get_hdp_flush_req_offset(struct amdgpu_device *adev)
+{
+	return SOC15_REG_OFFSET(NBIO, 0, regBIF_BX_PF0_GPU_HDP_FLUSH_REQ);
+}
+
+static u32 nbio_v7_9_get_hdp_flush_done_offset(struct amdgpu_device *adev)
+{
+	return SOC15_REG_OFFSET(NBIO, 0, regBIF_BX_PF0_GPU_HDP_FLUSH_DONE);
+}
+
+static u32 nbio_v7_9_get_pcie_index_offset(struct amdgpu_device *adev)
+{
+	return SOC15_REG_OFFSET(NBIO, 0, regBIF_BX0_PCIE_INDEX2);
+}
+
+static u32 nbio_v7_9_get_pcie_data_offset(struct amdgpu_device *adev)
+{
+	return SOC15_REG_OFFSET(NBIO, 0, regBIF_BX0_PCIE_DATA2);
+}
+
+static u32 nbio_v7_9_get_pcie_index_hi_offset(struct amdgpu_device *adev)
+{
+	return SOC15_REG_OFFSET(NBIO, 0, regBIF_BX0_PCIE_INDEX2_HI);
+}
+
+const struct nbio_hdp_flush_reg nbio_v7_9_hdp_flush_reg = {
+	.ref_and_mask_cp0 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__CP0_MASK,
+	.ref_and_mask_cp1 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__CP1_MASK,
+	.ref_and_mask_cp2 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__CP2_MASK,
+	.ref_and_mask_cp3 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__CP3_MASK,
+	.ref_and_mask_cp4 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__CP4_MASK,
+	.ref_and_mask_cp5 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__CP5_MASK,
+	.ref_and_mask_cp6 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__CP6_MASK,
+	.ref_and_mask_cp7 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__CP7_MASK,
+	.ref_and_mask_cp8 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__CP8_MASK,
+	.ref_and_mask_cp9 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__CP9_MASK,
+	.ref_and_mask_sdma0 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__SDMA0_MASK,
+	.ref_and_mask_sdma1 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__SDMA1_MASK,
+	.ref_and_mask_sdma2 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__RSVD_ENG0_MASK,
+	.ref_and_mask_sdma3 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__RSVD_ENG1_MASK,
+	.ref_and_mask_sdma4 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__RSVD_ENG2_MASK,
+	.ref_and_mask_sdma5 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__RSVD_ENG3_MASK,
+	.ref_and_mask_sdma6 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__RSVD_ENG4_MASK,
+	.ref_and_mask_sdma7 = BIF_BX_PF0_GPU_HDP_FLUSH_DONE__RSVD_ENG5_MASK,
+};
+
+static void nbio_v7_9_enable_doorbell_interrupt(struct amdgpu_device *adev,
+						bool enable)
+{
+	WREG32_FIELD15_PREREG(NBIO, 0, BIF_BX0_BIF_DOORBELL_INT_CNTL,
+			      DOORBELL_INTERRUPT_DISABLE, enable ? 0 : 1);
+}
+
+static int nbio_v7_9_get_compute_partition_mode(struct amdgpu_device *adev)
+{
+	u32 tmp, px;
+
+	tmp = RREG32_SOC15(NBIO, 0, regBIF_BX_PF0_PARTITION_COMPUTE_STATUS);
+	px = REG_GET_FIELD(tmp, BIF_BX_PF0_PARTITION_COMPUTE_STATUS,
+			   PARTITION_MODE);
+
+	return px;
+}
+
+static bool nbio_v7_9_is_nps_switch_requested(struct amdgpu_device *adev)
+{
+	u32 tmp;
+
+	tmp = RREG32_SOC15(NBIO, 0, regBIF_BX_PF0_PARTITION_MEM_STATUS);
+	tmp = REG_GET_FIELD(tmp, BIF_BX_PF0_PARTITION_MEM_STATUS,
+			    CHANGE_STATUE);
+
+	/* 0x8 - NPS switch requested */
+	return (tmp == 0x8);
+}
+static u32 nbio_v7_9_get_memory_partition_mode(struct amdgpu_device *adev,
+					       u32 *supp_modes)
+{
+	u32 tmp;
+
+	tmp = RREG32_SOC15(NBIO, 0, regBIF_BX_PF0_PARTITION_MEM_STATUS);
+	tmp = REG_GET_FIELD(tmp, BIF_BX_PF0_PARTITION_MEM_STATUS, NPS_MODE);
+
+	if (supp_modes) {
+		*supp_modes =
+			RREG32_SOC15(NBIO, 0, regBIF_BX_PF0_PARTITION_MEM_CAP);
+	}
+
+	return ffs(tmp);
+}
+
+static void nbio_v7_9_init_registers(struct amdgpu_device *adev)
+{
+	u32 inst_mask;
+	int i;
+
+	WREG32_SOC15(NBIO, 0, regXCC_DOORBELL_FENCE,
+		0xff & ~(adev->gfx.xcc_mask));
+
+	WREG32_SOC15(NBIO, 0, regBIFC_GFX_INT_MONITOR_MASK, 0x7ff);
+
+	inst_mask = adev->aid_mask & ~1U;
+	for_each_inst(i, inst_mask) {
+		WREG32_SOC15_EXT(NBIO, i, regXCC_DOORBELL_FENCE, i,
+			XCC_DOORBELL_FENCE__SHUB_SLV_MODE_MASK);
+
+	}
+
+	if (!amdgpu_sriov_vf(adev)) {
+		u32 baco_cntl;
+		for_each_inst(i, adev->aid_mask) {
+			baco_cntl = RREG32_SOC15(NBIO, i, regBIF_BX0_BACO_CNTL);
+			if (baco_cntl & (BIF_BX0_BACO_CNTL__BACO_DUMMY_EN_MASK |
+					 BIF_BX0_BACO_CNTL__BACO_EN_MASK)) {
+				baco_cntl &= ~(
+					BIF_BX0_BACO_CNTL__BACO_DUMMY_EN_MASK |
+					BIF_BX0_BACO_CNTL__BACO_EN_MASK);
+				dev_dbg(adev->dev,
+					"Unsetting baco dummy mode %x",
+					baco_cntl);
+				WREG32_SOC15(NBIO, i, regBIF_BX0_BACO_CNTL,
+					     baco_cntl);
+			}
+		}
+	}
+}
+
+#define MMIO_REG_HOLE_OFFSET 0x1A000
+
+static void nbio_v7_9_set_reg_remap(struct amdgpu_device *adev)
+{
+	if (!amdgpu_sriov_vf(adev) && (PAGE_SIZE <= 4096)) {
+		adev->rmmio_remap.reg_offset = MMIO_REG_HOLE_OFFSET;
+		adev->rmmio_remap.bus_addr = adev->rmmio_base + MMIO_REG_HOLE_OFFSET;
+	} else {
+		adev->rmmio_remap.reg_offset =
+			SOC15_REG_OFFSET(
+				NBIO, 0,
+				regBIF_BX_DEV0_EPF0_VF0_HDP_MEM_COHERENCY_FLUSH_CNTL)
+			<< 2;
+		adev->rmmio_remap.bus_addr = 0;
+	}
+}
+
+const struct amdgpu_nbio_funcs nbio_v7_9_funcs = {
+	.get_hdp_flush_req_offset = nbio_v7_9_get_hdp_flush_req_offset,
+	.get_hdp_flush_done_offset = nbio_v7_9_get_hdp_flush_done_offset,
+	.get_pcie_index_offset = nbio_v7_9_get_pcie_index_offset,
+	.get_pcie_data_offset = nbio_v7_9_get_pcie_data_offset,
+	.get_pcie_index_hi_offset = nbio_v7_9_get_pcie_index_hi_offset,
+	.get_rev_id = nbio_v7_9_get_rev_id,
+	.mc_access_enable = nbio_v7_9_mc_access_enable,
+	.get_memsize = nbio_v7_9_get_memsize,
+	.sdma_doorbell_range = nbio_v7_9_sdma_doorbell_range,
+	.vcn_doorbell_range = nbio_v7_9_vcn_doorbell_range,
+	.enable_doorbell_aperture = nbio_v7_9_enable_doorbell_aperture,
+	.enable_doorbell_selfring_aperture = nbio_v7_9_enable_doorbell_selfring_aperture,
+	.ih_doorbell_range = nbio_v7_9_ih_doorbell_range,
+	.enable_doorbell_interrupt = nbio_v7_9_enable_doorbell_interrupt,
+	.update_medium_grain_clock_gating = nbio_v7_9_update_medium_grain_clock_gating,
+	.update_medium_grain_light_sleep = nbio_v7_9_update_medium_grain_light_sleep,
+	.get_clockgating_state = nbio_v7_9_get_clockgating_state,
+	.ih_control = nbio_v7_9_ih_control,
+	.remap_hdp_registers = nbio_v7_9_remap_hdp_registers,
+	.get_compute_partition_mode = nbio_v7_9_get_compute_partition_mode,
+	.get_memory_partition_mode = nbio_v7_9_get_memory_partition_mode,
+	.is_nps_switch_requested = nbio_v7_9_is_nps_switch_requested,
+	.init_registers = nbio_v7_9_init_registers,
+	.set_reg_remap = nbio_v7_9_set_reg_remap,
+};
+
+static void nbio_v7_9_query_ras_error_count(struct amdgpu_device *adev,
+					void *ras_error_status)
+{
+}
+
+static void nbio_v7_9_handle_ras_controller_intr_no_bifring(struct amdgpu_device *adev)
+{
+	uint32_t bif_doorbell_intr_cntl;
+	struct ras_manager *obj = amdgpu_ras_find_obj(adev, adev->nbio.ras_if);
+	struct ras_err_data err_data;
+	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
+
+	if (amdgpu_ras_error_data_init(&err_data))
+		return;
+
+	bif_doorbell_intr_cntl = RREG32_SOC15(NBIO, 0, regBIF_BX0_BIF_DOORBELL_INT_CNTL);
+
+	if (REG_GET_FIELD(bif_doorbell_intr_cntl,
+		BIF_BX0_BIF_DOORBELL_INT_CNTL, RAS_CNTLR_INTERRUPT_STATUS)) {
+		/* driver has to clear the interrupt status when bif ring is disabled */
+		bif_doorbell_intr_cntl = REG_SET_FIELD(bif_doorbell_intr_cntl,
+						BIF_BX0_BIF_DOORBELL_INT_CNTL,
+						RAS_CNTLR_INTERRUPT_CLEAR, 1);
+		WREG32_SOC15(NBIO, 0, regBIF_BX0_BIF_DOORBELL_INT_CNTL, bif_doorbell_intr_cntl);
+
+		if (!ras->disable_ras_err_cnt_harvest) {
+			/*
+			 * clear error status after ras_controller_intr
+			 * according to hw team and count ue number
+			 * for query
+			 */
+			nbio_v7_9_query_ras_error_count(adev, &err_data);
+
+			/* logging on error cnt and printing for awareness */
+			obj->err_data.ue_count += err_data.ue_count;
+			obj->err_data.ce_count += err_data.ce_count;
+
+			if (err_data.ce_count)
+				dev_info(adev->dev, "%ld correctable hardware "
+						"errors detected in %s block\n",
+						obj->err_data.ce_count,
+						get_ras_block_str(adev->nbio.ras_if));
+
+			if (err_data.ue_count)
+				dev_info(adev->dev, "%ld uncorrectable hardware "
+						"errors detected in %s block\n",
+						obj->err_data.ue_count,
+						get_ras_block_str(adev->nbio.ras_if));
+		}
+
+		dev_info(adev->dev, "RAS controller interrupt triggered "
+					"by NBIF error\n");
+	}
+
+	amdgpu_ras_error_data_fini(&err_data);
+}
+
+static void nbio_v7_9_handle_ras_err_event_athub_intr_no_bifring(struct amdgpu_device *adev)
+{
+	uint32_t bif_doorbell_intr_cntl;
+
+	bif_doorbell_intr_cntl = RREG32_SOC15(NBIO, 0, regBIF_BX0_BIF_DOORBELL_INT_CNTL);
+
+	if (REG_GET_FIELD(bif_doorbell_intr_cntl,
+		BIF_BX0_BIF_DOORBELL_INT_CNTL, RAS_ATHUB_ERR_EVENT_INTERRUPT_STATUS)) {
+		/* driver has to clear the interrupt status when bif ring is disabled */
+		bif_doorbell_intr_cntl = REG_SET_FIELD(bif_doorbell_intr_cntl,
+						BIF_BX0_BIF_DOORBELL_INT_CNTL,
+						RAS_ATHUB_ERR_EVENT_INTERRUPT_CLEAR, 1);
+
+		WREG32_SOC15(NBIO, 0, regBIF_BX0_BIF_DOORBELL_INT_CNTL, bif_doorbell_intr_cntl);
+
+		amdgpu_ras_global_ras_isr(adev);
+	}
+}
+
+static int nbio_v7_9_set_ras_controller_irq_state(struct amdgpu_device *adev,
+						  struct amdgpu_irq_src *src,
+						  unsigned type,
+						  enum amdgpu_interrupt_state state)
+{
+	/* Dummy function, there is no initialization operation in driver */
+
+	return 0;
+}
+
+static int nbio_v7_9_process_ras_controller_irq(struct amdgpu_device *adev,
+						struct amdgpu_irq_src *source,
+						struct amdgpu_iv_entry *entry)
+{
+	/* By design, the ih cookie for ras_controller_irq should be written
+	 * to BIFring instead of general iv ring. However, due to known bif ring
+	 * hw bug, it has to be disabled. There is no chance the process function
+	 * will be involked. Just left it as a dummy one.
+	 */
+	return 0;
+}
+
+static int nbio_v7_9_set_ras_err_event_athub_irq_state(struct amdgpu_device *adev,
+						       struct amdgpu_irq_src *src,
+						       unsigned type,
+						       enum amdgpu_interrupt_state state)
+{
+	/* Dummy function, there is no initialization operation in driver */
+
+	return 0;
+}
+
+static int nbio_v7_9_process_err_event_athub_irq(struct amdgpu_device *adev,
+						 struct amdgpu_irq_src *source,
+						 struct amdgpu_iv_entry *entry)
+{
+	/* By design, the ih cookie for err_event_athub_irq should be written
+	 * to BIFring instead of general iv ring. However, due to known bif ring
+	 * hw bug, it has to be disabled. There is no chance the process function
+	 * will be involked. Just left it as a dummy one.
+	 */
+	return 0;
+}
+
+static const struct amdgpu_irq_src_funcs nbio_v7_9_ras_controller_irq_funcs = {
+	.set = nbio_v7_9_set_ras_controller_irq_state,
+	.process = nbio_v7_9_process_ras_controller_irq,
+};
+
+static const struct amdgpu_irq_src_funcs nbio_v7_9_ras_err_event_athub_irq_funcs = {
+	.set = nbio_v7_9_set_ras_err_event_athub_irq_state,
+	.process = nbio_v7_9_process_err_event_athub_irq,
+};
+
+static int nbio_v7_9_init_ras_controller_interrupt (struct amdgpu_device *adev)
+{
+	int r;
+
+	/* init the irq funcs */
+	adev->nbio.ras_controller_irq.funcs =
+		&nbio_v7_9_ras_controller_irq_funcs;
+	adev->nbio.ras_controller_irq.num_types = 1;
+
+	/* register ras controller interrupt */
+	r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_BIF,
+			      NBIF_7_4__SRCID__RAS_CONTROLLER_INTERRUPT,
+			      &adev->nbio.ras_controller_irq);
+
+	return r;
+}
+
+static int nbio_v7_9_init_ras_err_event_athub_interrupt (struct amdgpu_device *adev)
+{
+
+	int r;
+
+	/* init the irq funcs */
+	adev->nbio.ras_err_event_athub_irq.funcs =
+		&nbio_v7_9_ras_err_event_athub_irq_funcs;
+	adev->nbio.ras_err_event_athub_irq.num_types = 1;
+
+	/* register ras err event athub interrupt */
+	r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_BIF,
+			      NBIF_7_4__SRCID__ERREVENT_ATHUB_INTERRUPT,
+			      &adev->nbio.ras_err_event_athub_irq);
+
+	return r;
+}
+
+const struct amdgpu_ras_block_hw_ops nbio_v7_9_ras_hw_ops = {
+	.query_ras_error_count = nbio_v7_9_query_ras_error_count,
+};
+
+struct amdgpu_nbio_ras nbio_v7_9_ras = {
+	.ras_block = {
+		.ras_comm = {
+			.name = "pcie_bif",
+			.block = AMDGPU_RAS_BLOCK__PCIE_BIF,
+			.type = AMDGPU_RAS_ERROR__MULTI_UNCORRECTABLE,
+		},
+		.hw_ops = &nbio_v7_9_ras_hw_ops,
+		.ras_late_init = amdgpu_nbio_ras_late_init,
+	},
+	.handle_ras_controller_intr_no_bifring = nbio_v7_9_handle_ras_controller_intr_no_bifring,
+	.handle_ras_err_event_athub_intr_no_bifring = nbio_v7_9_handle_ras_err_event_athub_intr_no_bifring,
+	.init_ras_controller_interrupt = nbio_v7_9_init_ras_controller_interrupt,
+	.init_ras_err_event_athub_interrupt = nbio_v7_9_init_ras_err_event_athub_interrupt,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v7_9.h b/drivers/gpu/drm/amd/amdgpu/nbio_v7_9.h
new file mode 100644
index 000000000000..73709771950d
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/nbio_v7_9.h
@@ -0,0 +1,33 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __NBIO_V7_9_H__
+#define __NBIO_V7_9_H__
+
+#include "soc15_common.h"
+
+extern const struct nbio_hdp_flush_reg nbio_v7_9_hdp_flush_reg;
+extern const struct amdgpu_nbio_funcs nbio_v7_9_funcs;
+extern struct amdgpu_nbio_ras nbio_v7_9_ras;
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/nv.c b/drivers/gpu/drm/amd/amdgpu/nv.c
index 2ec1ffb36b1f..50e77d9b30af 100644
--- a/drivers/gpu/drm/amd/amdgpu/nv.c
+++ b/drivers/gpu/drm/amd/amdgpu/nv.c
@@ -59,7 +59,6 @@
 #include "vcn_v3_0.h"
 #include "jpeg_v3_0.h"
 #include "amdgpu_vkms.h"
-#include "mes_v10_1.h"
 #include "mxgpu_nv.h"
 #include "smuio_v11_0.h"
 #include "smuio_v11_0_6.h"
@@ -67,89 +66,119 @@
 static const struct amd_ip_funcs nv_common_ip_funcs;
 
 /* Navi */
-static const struct amdgpu_video_codec_info nv_video_codecs_encode_array[] =
-{
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 2304, 0)},
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC, 4096, 2304, 0)},
+static const struct amdgpu_video_codec_info nv_video_codecs_encode_array[] = {
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4096, 0)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC, 4096, 4096, 0)},
 };
 
-static const struct amdgpu_video_codecs nv_video_codecs_encode =
-{
+static const struct amdgpu_video_codecs nv_video_codecs_encode = {
 	.codec_count = ARRAY_SIZE(nv_video_codecs_encode_array),
 	.codec_array = nv_video_codecs_encode_array,
 };
 
 /* Navi1x */
-static const struct amdgpu_video_codec_info nv_video_codecs_decode_array[] =
-{
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG2, 4096, 4906, 3)},
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4, 4096, 4906, 5)},
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4906, 52)},
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VC1, 4096, 4906, 4)},
+static const struct amdgpu_video_codec_info nv_video_codecs_decode_array[] = {
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG2, 1920, 1088, 3)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4, 1920, 1088, 5)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4096, 52)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VC1, 1920, 1088, 4)},
 	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC, 8192, 4352, 186)},
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG, 4096, 4096, 0)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG, 8192, 8192, 0)},
 	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VP9, 8192, 4352, 0)},
 };
 
-static const struct amdgpu_video_codecs nv_video_codecs_decode =
-{
+static const struct amdgpu_video_codecs nv_video_codecs_decode = {
 	.codec_count = ARRAY_SIZE(nv_video_codecs_decode_array),
 	.codec_array = nv_video_codecs_decode_array,
 };
 
 /* Sienna Cichlid */
-static const struct amdgpu_video_codec_info sc_video_codecs_decode_array[] =
-{
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG2, 4096, 4906, 3)},
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4, 4096, 4906, 5)},
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4906, 52)},
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VC1, 4096, 4906, 4)},
+static const struct amdgpu_video_codec_info sc_video_codecs_encode_array[] = {
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4096, 0)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC, 8192, 4352, 0)},
+};
+
+static const struct amdgpu_video_codecs sc_video_codecs_encode = {
+	.codec_count = ARRAY_SIZE(sc_video_codecs_encode_array),
+	.codec_array = sc_video_codecs_encode_array,
+};
+
+static const struct amdgpu_video_codec_info sc_video_codecs_decode_array_vcn0[] = {
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG2, 1920, 1088, 3)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4, 1920, 1088, 5)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4096, 52)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VC1, 1920, 1088, 4)},
 	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC, 8192, 4352, 186)},
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG, 4096, 4096, 0)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG, 16384, 16384, 0)},
 	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VP9, 8192, 4352, 0)},
 	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_AV1, 8192, 4352, 0)},
 };
 
-static const struct amdgpu_video_codecs sc_video_codecs_decode =
-{
-	.codec_count = ARRAY_SIZE(sc_video_codecs_decode_array),
-	.codec_array = sc_video_codecs_decode_array,
+static const struct amdgpu_video_codec_info sc_video_codecs_decode_array_vcn1[] = {
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG2, 1920, 1088, 3)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4, 1920, 1088, 5)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4096, 52)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VC1, 1920, 1088, 4)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC, 8192, 4352, 186)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG, 16384, 16384, 0)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VP9, 8192, 4352, 0)},
+};
+
+static const struct amdgpu_video_codecs sc_video_codecs_decode_vcn0 = {
+	.codec_count = ARRAY_SIZE(sc_video_codecs_decode_array_vcn0),
+	.codec_array = sc_video_codecs_decode_array_vcn0,
+};
+
+static const struct amdgpu_video_codecs sc_video_codecs_decode_vcn1 = {
+	.codec_count = ARRAY_SIZE(sc_video_codecs_decode_array_vcn1),
+	.codec_array = sc_video_codecs_decode_array_vcn1,
 };
 
 /* SRIOV Sienna Cichlid, not const since data is controlled by host */
-static struct amdgpu_video_codec_info sriov_sc_video_codecs_encode_array[] =
-{
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 2304, 0)},
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC, 4096, 2304, 0)},
+static struct amdgpu_video_codec_info sriov_sc_video_codecs_encode_array[] = {
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4096, 0)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC, 8192, 4352, 0)},
 };
 
-static struct amdgpu_video_codec_info sriov_sc_video_codecs_decode_array[] =
-{
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG2, 4096, 4906, 3)},
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4, 4096, 4906, 5)},
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4906, 52)},
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VC1, 4096, 4906, 4)},
+static struct amdgpu_video_codec_info sriov_sc_video_codecs_decode_array_vcn0[] = {
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG2, 1920, 1088, 3)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4, 1920, 1088, 5)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4096, 52)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VC1, 1920, 1088, 4)},
 	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC, 8192, 4352, 186)},
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG, 4096, 4096, 0)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG, 16384, 16384, 0)},
 	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VP9, 8192, 4352, 0)},
 	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_AV1, 8192, 4352, 0)},
 };
 
-static struct amdgpu_video_codecs sriov_sc_video_codecs_encode =
-{
+static struct amdgpu_video_codec_info sriov_sc_video_codecs_decode_array_vcn1[] = {
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG2, 1920, 1088, 3)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4, 1920, 1088, 5)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4096, 52)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VC1, 1920, 1088, 4)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC, 8192, 4352, 186)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG, 16384, 16384, 0)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VP9, 8192, 4352, 0)},
+};
+
+static struct amdgpu_video_codecs sriov_sc_video_codecs_encode = {
 	.codec_count = ARRAY_SIZE(sriov_sc_video_codecs_encode_array),
 	.codec_array = sriov_sc_video_codecs_encode_array,
 };
 
-static struct amdgpu_video_codecs sriov_sc_video_codecs_decode =
-{
-	.codec_count = ARRAY_SIZE(sriov_sc_video_codecs_decode_array),
-	.codec_array = sriov_sc_video_codecs_decode_array,
+static struct amdgpu_video_codecs sriov_sc_video_codecs_decode_vcn0 = {
+	.codec_count = ARRAY_SIZE(sriov_sc_video_codecs_decode_array_vcn0),
+	.codec_array = sriov_sc_video_codecs_decode_array_vcn0,
+};
+
+static struct amdgpu_video_codecs sriov_sc_video_codecs_decode_vcn1 = {
+	.codec_count = ARRAY_SIZE(sriov_sc_video_codecs_decode_array_vcn1),
+	.codec_array = sriov_sc_video_codecs_decode_array_vcn1,
 };
 
 /* Beige Goby*/
 static const struct amdgpu_video_codec_info bg_video_codecs_decode_array[] = {
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4906, 52)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4096, 52)},
 	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC, 8192, 4352, 186)},
 	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VP9, 8192, 4352, 0)},
 };
@@ -166,10 +195,11 @@ static const struct amdgpu_video_codecs bg_video_codecs_encode = {
 
 /* Yellow Carp*/
 static const struct amdgpu_video_codec_info yc_video_codecs_decode_array[] = {
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4906, 52)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4096, 52)},
 	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC, 8192, 4352, 186)},
 	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VP9, 8192, 4352, 0)},
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG, 4096, 4096, 0)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG, 16384, 16384, 0)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_AV1, 8192, 4352, 0)},
 };
 
 static const struct amdgpu_video_codecs yc_video_codecs_decode = {
@@ -180,32 +210,50 @@ static const struct amdgpu_video_codecs yc_video_codecs_decode = {
 static int nv_query_video_codecs(struct amdgpu_device *adev, bool encode,
 				 const struct amdgpu_video_codecs **codecs)
 {
-	switch (adev->ip_versions[UVD_HWIP][0]) {
+	if (adev->vcn.num_vcn_inst == hweight8(adev->vcn.harvest_config))
+		return -EINVAL;
+
+	switch (amdgpu_ip_version(adev, UVD_HWIP, 0)) {
 	case IP_VERSION(3, 0, 0):
 	case IP_VERSION(3, 0, 64):
 	case IP_VERSION(3, 0, 192):
 		if (amdgpu_sriov_vf(adev)) {
-			if (encode)
-				*codecs = &sriov_sc_video_codecs_encode;
-			else
-				*codecs = &sriov_sc_video_codecs_decode;
+			if (adev->vcn.harvest_config & AMDGPU_VCN_HARVEST_VCN0) {
+				if (encode)
+					*codecs = &sriov_sc_video_codecs_encode;
+				else
+					*codecs = &sriov_sc_video_codecs_decode_vcn1;
+			} else {
+				if (encode)
+					*codecs = &sriov_sc_video_codecs_encode;
+				else
+					*codecs = &sriov_sc_video_codecs_decode_vcn0;
+			}
 		} else {
-			if (encode)
-				*codecs = &nv_video_codecs_encode;
-			else
-				*codecs = &sc_video_codecs_decode;
+			if (adev->vcn.harvest_config & AMDGPU_VCN_HARVEST_VCN0) {
+				if (encode)
+					*codecs = &sc_video_codecs_encode;
+				else
+					*codecs = &sc_video_codecs_decode_vcn1;
+			} else {
+				if (encode)
+					*codecs = &sc_video_codecs_encode;
+				else
+					*codecs = &sc_video_codecs_decode_vcn0;
+			}
 		}
 		return 0;
 	case IP_VERSION(3, 0, 16):
 	case IP_VERSION(3, 0, 2):
 		if (encode)
-			*codecs = &nv_video_codecs_encode;
+			*codecs = &sc_video_codecs_encode;
 		else
-			*codecs = &sc_video_codecs_decode;
+			*codecs = &sc_video_codecs_decode_vcn0;
 		return 0;
 	case IP_VERSION(3, 1, 1):
+	case IP_VERSION(3, 1, 2):
 		if (encode)
-			*codecs = &nv_video_codecs_encode;
+			*codecs = &sc_video_codecs_encode;
 		else
 			*codecs = &yc_video_codecs_decode;
 		return 0;
@@ -227,77 +275,6 @@ static int nv_query_video_codecs(struct amdgpu_device *adev, bool encode,
 	}
 }
 
-/*
- * Indirect registers accessor
- */
-static u32 nv_pcie_rreg(struct amdgpu_device *adev, u32 reg)
-{
-	unsigned long address, data;
-	address = adev->nbio.funcs->get_pcie_index_offset(adev);
-	data = adev->nbio.funcs->get_pcie_data_offset(adev);
-
-	return amdgpu_device_indirect_rreg(adev, address, data, reg);
-}
-
-static void nv_pcie_wreg(struct amdgpu_device *adev, u32 reg, u32 v)
-{
-	unsigned long address, data;
-
-	address = adev->nbio.funcs->get_pcie_index_offset(adev);
-	data = adev->nbio.funcs->get_pcie_data_offset(adev);
-
-	amdgpu_device_indirect_wreg(adev, address, data, reg, v);
-}
-
-static u64 nv_pcie_rreg64(struct amdgpu_device *adev, u32 reg)
-{
-	unsigned long address, data;
-	address = adev->nbio.funcs->get_pcie_index_offset(adev);
-	data = adev->nbio.funcs->get_pcie_data_offset(adev);
-
-	return amdgpu_device_indirect_rreg64(adev, address, data, reg);
-}
-
-static u32 nv_pcie_port_rreg(struct amdgpu_device *adev, u32 reg)
-{
-	unsigned long flags, address, data;
-	u32 r;
-	address = adev->nbio.funcs->get_pcie_port_index_offset(adev);
-	data = adev->nbio.funcs->get_pcie_port_data_offset(adev);
-
-	spin_lock_irqsave(&adev->pcie_idx_lock, flags);
-	WREG32(address, reg * 4);
-	(void)RREG32(address);
-	r = RREG32(data);
-	spin_unlock_irqrestore(&adev->pcie_idx_lock, flags);
-	return r;
-}
-
-static void nv_pcie_wreg64(struct amdgpu_device *adev, u32 reg, u64 v)
-{
-	unsigned long address, data;
-
-	address = adev->nbio.funcs->get_pcie_index_offset(adev);
-	data = adev->nbio.funcs->get_pcie_data_offset(adev);
-
-	amdgpu_device_indirect_wreg64(adev, address, data, reg, v);
-}
-
-static void nv_pcie_port_wreg(struct amdgpu_device *adev, u32 reg, u32 v)
-{
-	unsigned long flags, address, data;
-
-	address = adev->nbio.funcs->get_pcie_port_index_offset(adev);
-	data = adev->nbio.funcs->get_pcie_port_data_offset(adev);
-
-	spin_lock_irqsave(&adev->pcie_idx_lock, flags);
-	WREG32(address, reg * 4);
-	(void)RREG32(address);
-	WREG32(data, v);
-	(void)RREG32(data);
-	spin_unlock_irqrestore(&adev->pcie_idx_lock, flags);
-}
-
 static u32 nv_didt_rreg(struct amdgpu_device *adev, u32 reg)
 {
 	unsigned long flags, address, data;
@@ -349,49 +326,12 @@ void nv_grbm_select(struct amdgpu_device *adev,
 	WREG32_SOC15(GC, 0, mmGRBM_GFX_CNTL, grbm_gfx_cntl);
 }
 
-static void nv_vga_set_state(struct amdgpu_device *adev, bool state)
-{
-	/* todo */
-}
-
 static bool nv_read_disabled_bios(struct amdgpu_device *adev)
 {
 	/* todo */
 	return false;
 }
 
-static bool nv_read_bios_from_rom(struct amdgpu_device *adev,
-				  u8 *bios, u32 length_bytes)
-{
-	u32 *dw_ptr;
-	u32 i, length_dw;
-	u32 rom_index_offset, rom_data_offset;
-
-	if (bios == NULL)
-		return false;
-	if (length_bytes == 0)
-		return false;
-	/* APU vbios image is part of sbios image */
-	if (adev->flags & AMD_IS_APU)
-		return false;
-
-	dw_ptr = (u32 *)bios;
-	length_dw = ALIGN(length_bytes, 4) / 4;
-
-	rom_index_offset =
-		adev->smuio.funcs->get_rom_index_offset(adev);
-	rom_data_offset =
-		adev->smuio.funcs->get_rom_data_offset(adev);
-
-	/* set rom index to 0 */
-	WREG32(rom_index_offset, 0);
-	/* read out the rom data */
-	for (i = 0; i < length_dw; i++)
-		dw_ptr[i] = RREG32(rom_data_offset);
-
-	return true;
-}
-
 static struct soc15_allowed_register_entry nv_allowed_read_registers[] = {
 	{ SOC15_REG_ENTRY(GC, 0, mmGRBM_STATUS)},
 	{ SOC15_REG_ENTRY(GC, 0, mmGRBM_STATUS2)},
@@ -421,12 +361,12 @@ static uint32_t nv_read_indexed_register(struct amdgpu_device *adev, u32 se_num,
 
 	mutex_lock(&adev->grbm_idx_mutex);
 	if (se_num != 0xffffffff || sh_num != 0xffffffff)
-		amdgpu_gfx_select_se_sh(adev, se_num, sh_num, 0xffffffff);
+		amdgpu_gfx_select_se_sh(adev, se_num, sh_num, 0xffffffff, 0);
 
 	val = RREG32(reg_offset);
 
 	if (se_num != 0xffffffff || sh_num != 0xffffffff)
-		amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+		amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
 	mutex_unlock(&adev->grbm_idx_mutex);
 	return val;
 }
@@ -453,9 +393,10 @@ static int nv_read_register(struct amdgpu_device *adev, u32 se_num,
 	*value = 0;
 	for (i = 0; i < ARRAY_SIZE(nv_allowed_read_registers); i++) {
 		en = &nv_allowed_read_registers[i];
-		if ((i == 7 && (adev->sdma.num_instances == 1)) || /* some asics don't have SDMA1 */
-		    reg_offset !=
-		    (adev->reg_offset[en->hwip][en->inst][en->seg] + en->reg_offset))
+		if (!adev->reg_offset[en->hwip][en->inst])
+			continue;
+		else if (reg_offset != (adev->reg_offset[en->hwip][en->inst][en->seg]
+					+ en->reg_offset))
 			continue;
 
 		*value = nv_get_register_value(adev,
@@ -511,10 +452,13 @@ nv_asic_reset_method(struct amdgpu_device *adev)
 		dev_warn(adev->dev, "Specified reset method:%d isn't supported, using AUTO instead.\n",
 				  amdgpu_reset_method);
 
-	switch (adev->ip_versions[MP1_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, MP1_HWIP, 0)) {
 	case IP_VERSION(11, 5, 0):
+	case IP_VERSION(11, 5, 2):
 	case IP_VERSION(13, 0, 1):
 	case IP_VERSION(13, 0, 3):
+	case IP_VERSION(13, 0, 5):
+	case IP_VERSION(13, 0, 8):
 		return AMD_RESET_METHOD_MODE2;
 	case IP_VERSION(11, 0, 7):
 	case IP_VERSION(11, 0, 11):
@@ -567,41 +511,17 @@ static int nv_set_vce_clocks(struct amdgpu_device *adev, u32 evclk, u32 ecclk)
 	return 0;
 }
 
-static void nv_pcie_gen3_enable(struct amdgpu_device *adev)
-{
-	if (pci_is_root_bus(adev->pdev->bus))
-		return;
-
-	if (amdgpu_pcie_gen2 == 0)
-		return;
-
-	if (!(adev->pm.pcie_gen_mask & (CAIL_PCIE_LINK_SPEED_SUPPORT_GEN2 |
-					CAIL_PCIE_LINK_SPEED_SUPPORT_GEN3)))
-		return;
-
-	/* todo */
-}
-
 static void nv_program_aspm(struct amdgpu_device *adev)
 {
-	if (!amdgpu_aspm)
+	if (!amdgpu_device_should_use_aspm(adev))
 		return;
 
-	if (!(adev->flags & AMD_IS_APU) &&
-	    (adev->nbio.funcs->program_aspm))
+	if (adev->nbio.funcs->program_aspm)
 		adev->nbio.funcs->program_aspm(adev);
 
 }
 
-static void nv_enable_doorbell_aperture(struct amdgpu_device *adev,
-					bool enable)
-{
-	adev->nbio.funcs->enable_doorbell_aperture(adev, enable);
-	adev->nbio.funcs->enable_doorbell_selfring_aperture(adev, enable);
-}
-
-const struct amdgpu_ip_block_version nv_common_ip_block =
-{
+const struct amdgpu_ip_block_version nv_common_ip_block = {
 	.type = AMD_IP_BLOCK_TYPE_COMMON,
 	.major = 1,
 	.minor = 0,
@@ -614,11 +534,6 @@ void nv_set_virt_ops(struct amdgpu_device *adev)
 	adev->virt.ops = &xgpu_nv_virt_ops;
 }
 
-static uint32_t nv_get_rev_id(struct amdgpu_device *adev)
-{
-	return adev->nbio.funcs->get_rev_id(adev);
-}
-
 static bool nv_need_full_reset(struct amdgpu_device *adev)
 {
 	return true;
@@ -641,16 +556,6 @@ static bool nv_need_reset_on_init(struct amdgpu_device *adev)
 	return false;
 }
 
-static uint64_t nv_get_pcie_replay_count(struct amdgpu_device *adev)
-{
-
-	/* TODO
-	 * dummy implement for pcie_replay_count sysfs interface
-	 * */
-
-	return 0;
-}
-
 static void nv_init_doorbell_index(struct amdgpu_device *adev)
 {
 	adev->doorbell_index.kiq = AMDGPU_NAVI10_DOORBELL_KIQ;
@@ -666,7 +571,12 @@ static void nv_init_doorbell_index(struct amdgpu_device *adev)
 	adev->doorbell_index.userqueue_end = AMDGPU_NAVI10_DOORBELL_USERQUEUE_END;
 	adev->doorbell_index.gfx_ring0 = AMDGPU_NAVI10_DOORBELL_GFX_RING0;
 	adev->doorbell_index.gfx_ring1 = AMDGPU_NAVI10_DOORBELL_GFX_RING1;
-	adev->doorbell_index.mes_ring = AMDGPU_NAVI10_DOORBELL_MES_RING;
+	adev->doorbell_index.gfx_userqueue_start =
+		AMDGPU_NAVI10_DOORBELL_GFX_USERQUEUE_START;
+	adev->doorbell_index.gfx_userqueue_end =
+		AMDGPU_NAVI10_DOORBELL_GFX_USERQUEUE_END;
+	adev->doorbell_index.mes_ring0 = AMDGPU_NAVI10_DOORBELL_MES_RING0;
+	adev->doorbell_index.mes_ring1 = AMDGPU_NAVI10_DOORBELL_MES_RING1;
 	adev->doorbell_index.sdma_engine[0] = AMDGPU_NAVI10_DOORBELL_sDMA_ENGINE0;
 	adev->doorbell_index.sdma_engine[1] = AMDGPU_NAVI10_DOORBELL_sDMA_ENGINE1;
 	adev->doorbell_index.sdma_engine[2] = AMDGPU_NAVI10_DOORBELL_sDMA_ENGINE2;
@@ -691,28 +601,26 @@ static int nv_update_umd_stable_pstate(struct amdgpu_device *adev,
 				       bool enter)
 {
 	if (enter)
-		amdgpu_gfx_rlc_enter_safe_mode(adev);
+		amdgpu_gfx_rlc_enter_safe_mode(adev, 0);
 	else
-		amdgpu_gfx_rlc_exit_safe_mode(adev);
+		amdgpu_gfx_rlc_exit_safe_mode(adev, 0);
 
 	if (adev->gfx.funcs->update_perfmon_mgcg)
 		adev->gfx.funcs->update_perfmon_mgcg(adev, !enter);
 
-	if (!(adev->flags & AMD_IS_APU) &&
-	    (adev->nbio.funcs->enable_aspm))
+	if (adev->nbio.funcs->enable_aspm &&
+	    amdgpu_device_should_use_aspm(adev))
 		adev->nbio.funcs->enable_aspm(adev, !enter);
 
 	return 0;
 }
 
-static const struct amdgpu_asic_funcs nv_asic_funcs =
-{
+static const struct amdgpu_asic_funcs nv_asic_funcs = {
 	.read_disabled_bios = &nv_read_disabled_bios,
-	.read_bios_from_rom = &nv_read_bios_from_rom,
+	.read_bios_from_rom = &amdgpu_soc15_read_bios_from_rom,
 	.read_register = &nv_read_register,
 	.reset = &nv_asic_reset,
 	.reset_method = &nv_asic_reset_method,
-	.set_vga_state = &nv_vga_set_state,
 	.get_xclk = &nv_get_xclk,
 	.set_uvd_clocks = &nv_set_uvd_clocks,
 	.set_vce_clocks = &nv_set_vce_clocks,
@@ -720,30 +628,26 @@ static const struct amdgpu_asic_funcs nv_asic_funcs =
 	.init_doorbell_index = &nv_init_doorbell_index,
 	.need_full_reset = &nv_need_full_reset,
 	.need_reset_on_init = &nv_need_reset_on_init,
-	.get_pcie_replay_count = &nv_get_pcie_replay_count,
+	.get_pcie_replay_count = &amdgpu_nbio_get_pcie_replay_count,
 	.supports_baco = &amdgpu_dpm_is_baco_supported,
 	.pre_asic_init = &nv_pre_asic_init,
 	.update_umd_stable_pstate = &nv_update_umd_stable_pstate,
 	.query_video_codecs = &nv_query_video_codecs,
 };
 
-static int nv_common_early_init(void *handle)
+static int nv_common_early_init(struct amdgpu_ip_block *ip_block)
 {
-#define MMIO_REG_HOLE_OFFSET (0x80000 - PAGE_SIZE)
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
-	if (!amdgpu_sriov_vf(adev)) {
-		adev->rmmio_remap.reg_offset = MMIO_REG_HOLE_OFFSET;
-		adev->rmmio_remap.bus_addr = adev->rmmio_base + MMIO_REG_HOLE_OFFSET;
-	}
+	adev->nbio.funcs->set_reg_remap(adev);
 	adev->smc_rreg = NULL;
 	adev->smc_wreg = NULL;
-	adev->pcie_rreg = &nv_pcie_rreg;
-	adev->pcie_wreg = &nv_pcie_wreg;
-	adev->pcie_rreg64 = &nv_pcie_rreg64;
-	adev->pcie_wreg64 = &nv_pcie_wreg64;
-	adev->pciep_rreg = &nv_pcie_port_rreg;
-	adev->pciep_wreg = &nv_pcie_port_wreg;
+	adev->pcie_rreg = &amdgpu_device_indirect_rreg;
+	adev->pcie_wreg = &amdgpu_device_indirect_wreg;
+	adev->pcie_rreg64 = &amdgpu_device_indirect_rreg64;
+	adev->pcie_wreg64 = &amdgpu_device_indirect_wreg64;
+	adev->pciep_rreg = amdgpu_device_pcie_port_rreg;
+	adev->pciep_wreg = amdgpu_device_pcie_port_wreg;
 
 	/* TODO: will add them during VCN v2 implementation */
 	adev->uvd_ctx_rreg = NULL;
@@ -754,12 +658,12 @@ static int nv_common_early_init(void *handle)
 
 	adev->asic_funcs = &nv_asic_funcs;
 
-	adev->rev_id = nv_get_rev_id(adev);
+	adev->rev_id = amdgpu_device_get_rev_id(adev);
 	adev->external_rev_id = 0xff;
 	/* TODO: split the GC and PG flags based on the relevant IP version for which
 	 * they are relevant.
 	 */
-	switch (adev->ip_versions[GC_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 	case IP_VERSION(10, 1, 10):
 		adev->cg_flags = AMD_CG_SUPPORT_GFX_MGCG |
 			AMD_CG_SUPPORT_GFX_CGCG |
@@ -953,7 +857,8 @@ static int nv_common_early_init(void *handle)
 			AMD_CG_SUPPORT_ATHUB_LS |
 			AMD_CG_SUPPORT_IH_CG |
 			AMD_CG_SUPPORT_VCN_MGCG |
-			AMD_CG_SUPPORT_JPEG_MGCG;
+			AMD_CG_SUPPORT_JPEG_MGCG |
+			AMD_CG_SUPPORT_SDMA_MGCG;
 		adev->pg_flags = AMD_PG_SUPPORT_GFX_PG |
 			AMD_PG_SUPPORT_VCN |
 			AMD_PG_SUPPORT_VCN_DPG |
@@ -964,10 +869,64 @@ static int nv_common_early_init(void *handle)
 			adev->external_rev_id = adev->rev_id + 0x01;
 		break;
 	case IP_VERSION(10, 1, 3):
+	case IP_VERSION(10, 1, 4):
 		adev->cg_flags = 0;
 		adev->pg_flags = 0;
 		adev->external_rev_id = adev->rev_id + 0x82;
 		break;
+	case IP_VERSION(10, 3, 6):
+		adev->cg_flags = AMD_CG_SUPPORT_GFX_MGCG |
+			AMD_CG_SUPPORT_GFX_MGLS |
+			AMD_CG_SUPPORT_GFX_CGCG |
+			AMD_CG_SUPPORT_GFX_CGLS |
+			AMD_CG_SUPPORT_GFX_3D_CGCG |
+			AMD_CG_SUPPORT_GFX_3D_CGLS |
+			AMD_CG_SUPPORT_GFX_RLC_LS |
+			AMD_CG_SUPPORT_GFX_CP_LS |
+			AMD_CG_SUPPORT_GFX_FGCG |
+			AMD_CG_SUPPORT_MC_MGCG |
+			AMD_CG_SUPPORT_MC_LS |
+			AMD_CG_SUPPORT_SDMA_LS |
+			AMD_CG_SUPPORT_HDP_MGCG |
+			AMD_CG_SUPPORT_HDP_LS |
+			AMD_CG_SUPPORT_ATHUB_MGCG |
+			AMD_CG_SUPPORT_ATHUB_LS |
+			AMD_CG_SUPPORT_IH_CG |
+			AMD_CG_SUPPORT_VCN_MGCG |
+			AMD_CG_SUPPORT_JPEG_MGCG;
+		adev->pg_flags = AMD_PG_SUPPORT_GFX_PG |
+			AMD_PG_SUPPORT_VCN |
+			AMD_PG_SUPPORT_VCN_DPG |
+			AMD_PG_SUPPORT_JPEG;
+		adev->external_rev_id = adev->rev_id + 0x01;
+		break;
+	case IP_VERSION(10, 3, 7):
+		adev->cg_flags =  AMD_CG_SUPPORT_GFX_MGCG |
+			AMD_CG_SUPPORT_GFX_MGLS |
+			AMD_CG_SUPPORT_GFX_CGCG |
+			AMD_CG_SUPPORT_GFX_CGLS |
+			AMD_CG_SUPPORT_GFX_3D_CGCG |
+			AMD_CG_SUPPORT_GFX_3D_CGLS |
+			AMD_CG_SUPPORT_GFX_RLC_LS |
+			AMD_CG_SUPPORT_GFX_CP_LS |
+			AMD_CG_SUPPORT_GFX_FGCG |
+			AMD_CG_SUPPORT_MC_MGCG |
+			AMD_CG_SUPPORT_MC_LS |
+			AMD_CG_SUPPORT_SDMA_LS |
+			AMD_CG_SUPPORT_HDP_MGCG |
+			AMD_CG_SUPPORT_HDP_LS |
+			AMD_CG_SUPPORT_ATHUB_MGCG |
+			AMD_CG_SUPPORT_ATHUB_LS |
+			AMD_CG_SUPPORT_IH_CG |
+			AMD_CG_SUPPORT_VCN_MGCG |
+			AMD_CG_SUPPORT_JPEG_MGCG |
+			AMD_CG_SUPPORT_SDMA_MGCG;
+		adev->pg_flags = AMD_PG_SUPPORT_VCN |
+			AMD_PG_SUPPORT_VCN_DPG |
+			AMD_PG_SUPPORT_JPEG |
+			AMD_PG_SUPPORT_GFX_PG;
+		adev->external_rev_id = adev->rev_id + 0x01;
+		break;
 	default:
 		/* FIXME: not supported yet */
 		return -EINVAL;
@@ -986,23 +945,38 @@ static int nv_common_early_init(void *handle)
 	return 0;
 }
 
-static int nv_common_late_init(void *handle)
+static int nv_common_late_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (amdgpu_sriov_vf(adev)) {
 		xgpu_nv_mailbox_get_irq(adev);
-		amdgpu_virt_update_sriov_video_codec(adev,
-				sriov_sc_video_codecs_encode_array, ARRAY_SIZE(sriov_sc_video_codecs_encode_array),
-				sriov_sc_video_codecs_decode_array, ARRAY_SIZE(sriov_sc_video_codecs_decode_array));
+		if (adev->vcn.harvest_config & AMDGPU_VCN_HARVEST_VCN0) {
+			amdgpu_virt_update_sriov_video_codec(adev,
+							     sriov_sc_video_codecs_encode_array,
+							     ARRAY_SIZE(sriov_sc_video_codecs_encode_array),
+							     sriov_sc_video_codecs_decode_array_vcn1,
+							     ARRAY_SIZE(sriov_sc_video_codecs_decode_array_vcn1));
+		} else {
+			amdgpu_virt_update_sriov_video_codec(adev,
+							     sriov_sc_video_codecs_encode_array,
+							     ARRAY_SIZE(sriov_sc_video_codecs_encode_array),
+							     sriov_sc_video_codecs_decode_array_vcn0,
+							     ARRAY_SIZE(sriov_sc_video_codecs_decode_array_vcn0));
+		}
 	}
 
+	/* Enable selfring doorbell aperture late because doorbell BAR
+	 * aperture will change if resize BAR successfully in gmc sw_init.
+	 */
+	adev->nbio.funcs->enable_doorbell_selfring_aperture(adev, true);
+
 	return 0;
 }
 
-static int nv_common_sw_init(void *handle)
+static int nv_common_sw_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (amdgpu_sriov_vf(adev))
 		xgpu_nv_mailbox_add_irq_id(adev);
@@ -1010,14 +984,9 @@ static int nv_common_sw_init(void *handle)
 	return 0;
 }
 
-static int nv_common_sw_fini(void *handle)
+static int nv_common_hw_init(struct amdgpu_ip_block *ip_block)
 {
-	return 0;
-}
-
-static int nv_common_hw_init(void *handle)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (adev->nbio.funcs->apply_lc_spc_mode_wa)
 		adev->nbio.funcs->apply_lc_spc_mode_wa(adev);
@@ -1025,8 +994,6 @@ static int nv_common_hw_init(void *handle)
 	if (adev->nbio.funcs->apply_l1_link_width_reconfig_wa)
 		adev->nbio.funcs->apply_l1_link_width_reconfig_wa(adev);
 
-	/* enable pcie gen2/3 link */
-	nv_pcie_gen3_enable(adev);
 	/* enable aspm */
 	nv_program_aspm(adev);
 	/* setup nbio registers */
@@ -1038,59 +1005,50 @@ static int nv_common_hw_init(void *handle)
 	if (adev->nbio.funcs->remap_hdp_registers && !amdgpu_sriov_vf(adev))
 		adev->nbio.funcs->remap_hdp_registers(adev);
 	/* enable the doorbell aperture */
-	nv_enable_doorbell_aperture(adev, true);
+	adev->nbio.funcs->enable_doorbell_aperture(adev, true);
 
 	return 0;
 }
 
-static int nv_common_hw_fini(void *handle)
+static int nv_common_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
-	/* disable the doorbell aperture */
-	nv_enable_doorbell_aperture(adev, false);
+	/* Disable the doorbell aperture and selfring doorbell aperture
+	 * separately in hw_fini because nv_enable_doorbell_aperture
+	 * has been removed and there is no need to delay disabling
+	 * selfring doorbell.
+	 */
+	adev->nbio.funcs->enable_doorbell_aperture(adev, false);
+	adev->nbio.funcs->enable_doorbell_selfring_aperture(adev, false);
 
 	return 0;
 }
 
-static int nv_common_suspend(void *handle)
+static int nv_common_suspend(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	return nv_common_hw_fini(adev);
+	return nv_common_hw_fini(ip_block);
 }
 
-static int nv_common_resume(void *handle)
+static int nv_common_resume(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	return nv_common_hw_init(adev);
+	return nv_common_hw_init(ip_block);
 }
 
-static bool nv_common_is_idle(void *handle)
+static bool nv_common_is_idle(struct amdgpu_ip_block *ip_block)
 {
 	return true;
 }
 
-static int nv_common_wait_for_idle(void *handle)
-{
-	return 0;
-}
-
-static int nv_common_soft_reset(void *handle)
-{
-	return 0;
-}
-
-static int nv_common_set_clockgating_state(void *handle,
+static int nv_common_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					   enum amd_clockgating_state state)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (amdgpu_sriov_vf(adev))
 		return 0;
 
-	switch (adev->ip_versions[NBIO_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, NBIO_HWIP, 0)) {
 	case IP_VERSION(2, 3, 0):
 	case IP_VERSION(2, 3, 1):
 	case IP_VERSION(2, 3, 2):
@@ -1113,16 +1071,16 @@ static int nv_common_set_clockgating_state(void *handle,
 	return 0;
 }
 
-static int nv_common_set_powergating_state(void *handle,
+static int nv_common_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					   enum amd_powergating_state state)
 {
 	/* TODO */
 	return 0;
 }
 
-static void nv_common_get_clockgating_state(void *handle, u32 *flags)
+static void nv_common_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (amdgpu_sriov_vf(adev))
 		*flags = 0;
@@ -1132,8 +1090,6 @@ static void nv_common_get_clockgating_state(void *handle, u32 *flags)
 	adev->hdp.funcs->get_clock_gating_state(adev, flags);
 
 	adev->smuio.funcs->get_clock_gating_state(adev, flags);
-
-	return;
 }
 
 static const struct amd_ip_funcs nv_common_ip_funcs = {
@@ -1141,14 +1097,11 @@ static const struct amd_ip_funcs nv_common_ip_funcs = {
 	.early_init = nv_common_early_init,
 	.late_init = nv_common_late_init,
 	.sw_init = nv_common_sw_init,
-	.sw_fini = nv_common_sw_fini,
 	.hw_init = nv_common_hw_init,
 	.hw_fini = nv_common_hw_fini,
 	.suspend = nv_common_suspend,
 	.resume = nv_common_resume,
 	.is_idle = nv_common_is_idle,
-	.wait_for_idle = nv_common_wait_for_idle,
-	.soft_reset = nv_common_soft_reset,
 	.set_clockgating_state = nv_common_set_clockgating_state,
 	.set_powergating_state = nv_common_set_powergating_state,
 	.get_clockgating_state = nv_common_get_clockgating_state,
diff --git a/drivers/gpu/drm/amd/amdgpu/nv.h b/drivers/gpu/drm/amd/amdgpu/nv.h
index 83e9782aef39..8f4817404f10 100644
--- a/drivers/gpu/drm/amd/amdgpu/nv.h
+++ b/drivers/gpu/drm/amd/amdgpu/nv.h
@@ -31,5 +31,6 @@ extern const struct amdgpu_ip_block_version nv_common_ip_block;
 void nv_grbm_select(struct amdgpu_device *adev,
 		    u32 me, u32 pipe, u32 queue, u32 vmid);
 void nv_set_virt_ops(struct amdgpu_device *adev);
+int cyan_skillfish_reg_base_init(struct amdgpu_device *adev);
 
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/nvd.h b/drivers/gpu/drm/amd/amdgpu/nvd.h
index fd6b58243b03..56f1bfac0b20 100644
--- a/drivers/gpu/drm/amd/amdgpu/nvd.h
+++ b/drivers/gpu/drm/amd/amdgpu/nvd.h
@@ -64,6 +64,24 @@
 #define	PACKET3_INDIRECT_BUFFER_CNST_END		0x19
 #define	PACKET3_ATOMIC_GDS				0x1D
 #define	PACKET3_ATOMIC_MEM				0x1E
+#define 	PACKET3_ATOMIC_MEM__ATOMIC(x) ((((unsigned)(x)) & 0x7F) << 0)
+#define 	PACKET3_ATOMIC_MEM__COMMAND(x) ((((unsigned)(x)) & 0xF) << 8)
+#define 	PACKET3_ATOMIC_MEM__CACHE_POLICY(x) ((((unsigned)(x)) & 0x3) << 25)
+#define 	PACKET3_ATOMIC_MEM__ADDR_LO(x) (((unsigned)(x)))
+#define 	PACKET3_ATOMIC_MEM__ADDR_HI(x) (((unsigned)(x)))
+#define 	PACKET3_ATOMIC_MEM__SRC_DATA_LO(x) (((unsigned)(x)))
+#define 	PACKET3_ATOMIC_MEM__SRC_DATA_HI(x) (((unsigned)(x)))
+#define 	PACKET3_ATOMIC_MEM__CMP_DATA_LO(x) (((unsigned)(x)))
+#define 	PACKET3_ATOMIC_MEM__CMP_DATA_HI(x) (((unsigned)(x)))
+#define 	PACKET3_ATOMIC_MEM__LOOP_INTERVAL(x) ((((unsigned)(x)) & 0x1FFF) << 0)
+#define 	PACKET3_ATOMIC_MEM__COMMAND__SINGLE_PASS_ATOMIC 0
+#define 	PACKET3_ATOMIC_MEM__COMMAND__LOOP_UNTIL_COMPARE_SATISFIED 1
+#define 	PACKET3_ATOMIC_MEM__COMMAND__WAIT_FOR_WRITE_CONFIRMATION 2
+#define 	PACKET3_ATOMIC_MEM__COMMAND__SEND_AND_CONTINUE 3
+#define 	PACKET3_ATOMIC_MEM__CACHE_POLICY__LRU 0
+#define 	PACKET3_ATOMIC_MEM__CACHE_POLICY__STREAM 1
+#define 	PACKET3_ATOMIC_MEM__CACHE_POLICY__NOA 2
+#define 	PACKET3_ATOMIC_MEM__CACHE_POLICY__BYPASS 3
 #define	PACKET3_OCCLUSION_QUERY				0x1F
 #define	PACKET3_SET_PREDICATION				0x20
 #define	PACKET3_REG_RMW					0x21
@@ -105,6 +123,38 @@
 		 * 1 - pfp
 		 * 2 - ce
 		 */
+#define 	PACKET3_WRITE_DATA__DST_SEL(x) ((((unsigned)(x)) & 0xF) << 8)
+#define 	PACKET3_WRITE_DATA__ADDR_INCR(x) ((((unsigned)(x)) & 0x1) << 16)
+#define 	PACKET3_WRITE_DATA__WR_CONFIRM(x) ((((unsigned)(x)) & 0x1) << 20)
+#define 	PACKET3_WRITE_DATA__CACHE_POLICY(x) ((((unsigned)(x)) & 0x3) << 25)
+#define 	PACKET3_WRITE_DATA__DST_MMREG_ADDR(x) ((((unsigned)(x)) & 0x3FFFF) << 0)
+#define 	PACKET3_WRITE_DATA__DST_GDS_ADDR(x) ((((unsigned)(x)) & 0xFFFF) << 0)
+#define 	PACKET3_WRITE_DATA__DST_MEM_ADDR_LO(x) ((((unsigned)(x)) & 0x3FFFFFFF) << 2)
+#define 	PACKET3_WRITE_DATA__DST_MEM_ADDR_HI(x) ((unsigned)(x))
+#define 	PACKET3_WRITE_DATA__MODE(x) ((((unsigned)(x)) & 0x1) << 21)
+#define 	PACKET3_WRITE_DATA__AID_ID(x) ((((unsigned)(x)) & 0x3) << 22)
+#define 	PACKET3_WRITE_DATA__TEMPORAL(x) ((((unsigned)(x)) & 0x3) << 24)
+#define 	PACKET3_WRITE_DATA__DST_MMREG_ADDR_LO(x) ((unsigned)(x))
+#define 	PACKET3_WRITE_DATA__DST_MMREG_ADDR_HI(x) ((((unsigned)(x)) & 0xFF) << 0)
+#define 	PACKET3_WRITE_DATA__DST_SEL__MEM_MAPPED_REGISTER 0
+#define 	PACKET3_WRITE_DATA__DST_SEL__TC_L2 2
+#define 	PACKET3_WRITE_DATA__DST_SEL__GDS 3
+#define 	PACKET3_WRITE_DATA__DST_SEL__MEMORY 5
+#define 	PACKET3_WRITE_DATA__DST_SEL__MEMORY_MAPPED_ADC_PERSISTENT_STATE 6
+#define 	PACKET3_WRITE_DATA__ADDR_INCR__INCREMENT_ADDRESS 0
+#define 	PACKET3_WRITE_DATA__ADDR_INCR__DO_NOT_INCREMENT_ADDRESS 1
+#define 	PACKET3_WRITE_DATA__WR_CONFIRM__DO_NOT_WAIT_FOR_WRITE_CONFIRMATION 0
+#define 	PACKET3_WRITE_DATA__WR_CONFIRM__WAIT_FOR_WRITE_CONFIRMATION 1
+#define 	PACKET3_WRITE_DATA__MODE__PF_VF_DISABLED 0
+#define 	PACKET3_WRITE_DATA__MODE__PF_VF_ENABLED 1
+#define 	PACKET3_WRITE_DATA__TEMPORAL__RT 0
+#define 	PACKET3_WRITE_DATA__TEMPORAL__NT 1
+#define 	PACKET3_WRITE_DATA__TEMPORAL__HT 2
+#define 	PACKET3_WRITE_DATA__TEMPORAL__LU 3
+#define 	PACKET3_WRITE_DATA__CACHE_POLICY__LRU 0
+#define 	PACKET3_WRITE_DATA__CACHE_POLICY__STREAM 1
+#define 	PACKET3_WRITE_DATA__CACHE_POLICY__NOA 2
+#define 	PACKET3_WRITE_DATA__CACHE_POLICY__BYPASS 3
 #define	PACKET3_DRAW_INDEX_INDIRECT_MULTI		0x38
 #define	PACKET3_MEM_SEMAPHORE				0x39
 #              define PACKET3_SEM_USE_MAILBOX       (0x1 << 16)
@@ -135,6 +185,42 @@
 		/* 0 - me
 		 * 1 - pfp
 		 */
+#define		PACKET3_WAIT_REG_MEM__FUNCTION(x) ((((unsigned)(x)) & 0x7) << 0)
+#define		PACKET3_WAIT_REG_MEM__MEM_SPACE(x) ((((unsigned)(x)) & 0x3) << 4)
+#define		PACKET3_WAIT_REG_MEM__OPERATION(x) ((((unsigned)(x)) & 0x3) << 6)
+#define		PACKET3_WAIT_REG_MEM__MES_INTR_PIPE(x) ((((unsigned)(x)) & 0x3) << 22)
+#define		PACKET3_WAIT_REG_MEM__MES_ACTION(x) ((((unsigned)(x)) & 0x1) << 24)
+#define		PACKET3_WAIT_REG_MEM__CACHE_POLICY(x) ((((unsigned)(x)) & 0x3) << 25)
+#define		PACKET3_WAIT_REG_MEM__TEMPORAL(x) ((((unsigned)(x)) & 0x3) << 25)
+#define		PACKET3_WAIT_REG_MEM__MEM_POLL_ADDR_LO(x) ((((unsigned)(x)) & 0x3FFFFFFF) << 2)
+#define		PACKET3_WAIT_REG_MEM__REG_POLL_ADDR(x) ((((unsigned)(x)) & 0X3FFFF) << 0)
+#define		PACKET3_WAIT_REG_MEM__REG_WRITE_ADDR1(x) ((((unsigned)(x)) & 0X3FFFF) << 0)
+#define		PACKET3_WAIT_REG_MEM__MEM_POLL_ADDR_HI(x) ((unsigned)(x))
+#define		PACKET3_WAIT_REG_MEM__REG_WRITE_ADDR2(x) ((((unsigned)(x)) & 0x3FFFF) << 0)
+#define		PACKET3_WAIT_REG_MEM__REFERENCE(x) ((unsigned)(x))
+#define		PACKET3_WAIT_REG_MEM__MASK(x) ((unsigned)(x))
+#define		PACKET3_WAIT_REG_MEM__POLL_INTERVAL(x) ((((unsigned)(x)) & 0xFFFF) << 0)
+#define		PACKET3_WAIT_REG_MEM__OPTIMIZE_ACE_OFFLOAD_MODE(x) ((((unsigned)(x)) & 0x1) << 31)
+#define 	PACKET3_WAIT_REG_MEM__FUNCTION__ALWAYS_PASS 0
+#define 	PACKET3_WAIT_REG_MEM__FUNCTION__LESS_THAN_REF_VALUE 1
+#define 	PACKET3_WAIT_REG_MEM__FUNCTION__LESS_THAN_EQUAL_TO_THE_REF_VALUE 2
+#define 	PACKET3_WAIT_REG_MEM__FUNCTION__EQUAL_TO_THE_REFERENCE_VALUE 3
+#define 	PACKET3_WAIT_REG_MEM__FUNCTION__NOT_EQUAL_REFERENCE_VALUE 4
+#define 	PACKET3_WAIT_REG_MEM__FUNCTION__GREATER_THAN_OR_EQUAL_REFERENCE_VALUE 5
+#define 	PACKET3_WAIT_REG_MEM__FUNCTION__GREATER_THAN_REFERENCE_VALUE 6
+#define 	PACKET3_WAIT_REG_MEM__MEM_SPACE__REGISTER_SPACE 0
+#define 	PACKET3_WAIT_REG_MEM__MEM_SPACE__MEMORY_SPACE 1
+#define 	PACKET3_WAIT_REG_MEM__OPERATION__WAIT_REG_MEM 0
+#define 	PACKET3_WAIT_REG_MEM__OPERATION__WR_WAIT_WR_REG 1
+#define 	PACKET3_WAIT_REG_MEM__OPERATION__WAIT_MEM_PREEMPTABLE 3
+#define 	PACKET3_WAIT_REG_MEM__CACHE_POLICY__LRU 0
+#define 	PACKET3_WAIT_REG_MEM__CACHE_POLICY__STREAM 1
+#define 	PACKET3_WAIT_REG_MEM__CACHE_POLICY__NOA 2
+#define 	PACKET3_WAIT_REG_MEM__CACHE_POLICY__BYPASS 3
+#define 	PACKET3_WAIT_REG_MEM__TEMPORAL__RT 0
+#define 	PACKET3_WAIT_REG_MEM__TEMPORAL__NT 1
+#define 	PACKET3_WAIT_REG_MEM__TEMPORAL__HT 2
+#define 	PACKET3_WAIT_REG_MEM__TEMPORAL__LU 3
 #define	PACKET3_INDIRECT_BUFFER				0x3F
 #define		INDIRECT_BUFFER_VALID                   (1 << 23)
 #define		INDIRECT_BUFFER_CACHE_POLICY(x)         ((x) << 28)
@@ -144,8 +230,94 @@
 		 */
 #define		INDIRECT_BUFFER_PRE_ENB(x)		((x) << 21)
 #define		INDIRECT_BUFFER_PRE_RESUME(x)           ((x) << 30)
+#define 	PACKET3_INDIRECT_BUFFER__IB_BASE_LO(x) ((((unsigned)(x)) & 0x3FFFFFFF) << 2)
+#define 	PACKET3_INDIRECT_BUFFER__IB_BASE_HI(x) ((unsigned)(x))
+#define 	PACKET3_INDIRECT_BUFFER__IB_SIZE(x) ((((unsigned)(x)) & 0xFFFFF) << 0)
+#define 	PACKET3_INDIRECT_BUFFER__CHAIN(x) ((((unsigned)(x)) & 0x1) << 20)
+#define 	PACKET3_INDIRECT_BUFFER__OFFLOAD_POLLING(x) ((((unsigned)(x)) & 0x1) << 21)
+#define 	PACKET3_INDIRECT_BUFFER__VALID(x) ((((unsigned)(x)) & 0x1) << 23)
+#define 	PACKET3_INDIRECT_BUFFER__VMID(x) ((((unsigned)(x)) & 0xF) << 24)
+#define 	PACKET3_INDIRECT_BUFFER__CACHE_POLICY(x) ((((unsigned)(x)) & 0x3) << 28)
+#define 	PACKET3_INDIRECT_BUFFER__TEMPORAL(x) ((((unsigned)(x)) & 0x3) << 28)
+#define 	PACKET3_INDIRECT_BUFFER__PRIV(x) ((((unsigned)(x)) & 0x1) << 31)
+#define 	PACKET3_INDIRECT_BUFFER__TEMPORAL__RT 0
+#define 	PACKET3_INDIRECT_BUFFER__TEMPORAL__NT 1
+#define 	PACKET3_INDIRECT_BUFFER__TEMPORAL__HT 2
+#define 	PACKET3_INDIRECT_BUFFER__TEMPORAL__LU 3
+#define 	PACKET3_INDIRECT_BUFFER__CACHE_POLICY__LRU 0
+#define 	PACKET3_INDIRECT_BUFFER__CACHE_POLICY__STREAM 1
+#define 	PACKET3_INDIRECT_BUFFER__CACHE_POLICY__NOA 2
+#define 	PACKET3_INDIRECT_BUFFER__CACHE_POLICY__BYPASS 3
 #define	PACKET3_COND_INDIRECT_BUFFER			0x3F
 #define	PACKET3_COPY_DATA				0x40
+#define 	PACKET3_COPY_DATA__SRC_SEL(x) ((((unsigned)(x)) & 0xF) << 0)
+#define 	PACKET3_COPY_DATA__DST_SEL(x) ((((unsigned)(x)) & 0xF) << 8)
+#define 	PACKET3_COPY_DATA__SRC_CACHE_POLICY(x) ((((unsigned)(x)) & 0x3) << 13)
+#define 	PACKET3_COPY_DATA__SRC_TEMPORAL(x) ((((unsigned)(x)) & 0x3) << 13)
+#define 	PACKET3_COPY_DATA__COUNT_SEL(x) ((((unsigned)(x)) & 0x1) << 16)
+#define 	PACKET3_COPY_DATA__WR_CONFIRM(x) ((((unsigned)(x)) & 0x1) << 20)
+#define 	PACKET3_COPY_DATA__DST_CACHE_POLICY(x) ((((unsigned)(x)) & 0x3) << 25)
+#define 	PACKET3_COPY_DATA__PQ_EXE_STATUS(x) ((((unsigned)(x)) & 0x1) << 29)
+#define 	PACKET3_COPY_DATA__SRC_REG_OFFSET(x) ((((unsigned)(x)) & 0x3FFFF) << 0)
+#define 	PACKET3_COPY_DATA__SRC_32B_ADDR_LO(x) ((((unsigned)(x)) & 0x3FFFFFFF) << 2)
+#define 	PACKET3_COPY_DATA__SRC_64B_ADDR_LO(x) ((((unsigned)(x)) & 0x1FFFFFFF) << 3)
+#define 	PACKET3_COPY_DATA__SRC_GDS_ADDR_LO(x) ((((unsigned)(x)) & 0xFFFF) << 0)
+#define 	PACKET3_COPY_DATA__IMM_DATA(x) ((unsigned)(x))
+#define 	PACKET3_COPY_DATA__SRC_MEMTC_ADDR_HI(x) ((unsigned)(x))
+#define 	PACKET3_COPY_DATA__SRC_IMM_DATA(x) ((unsigned)(x))
+#define 	PACKET3_COPY_DATA__DST_REG_OFFSET(x) ((((unsigned)(x)) & 0x3FFFF) << 0)
+#define 	PACKET3_COPY_DATA__DST_32B_ADDR_LO(x) ((((unsigned)(x)) & 0x3FFFFFFF) << 2)
+#define 	PACKET3_COPY_DATA__DST_64B_ADDR_LO(x) ((((unsigned)(x)) & 0x1FFFFFFF) << 3)
+#define 	PACKET3_COPY_DATA__DST_GDS_ADDR_LO(x) ((((unsigned)(x)) & 0xFFFF) << 0)
+#define 	PACKET3_COPY_DATA__DST_ADDR_HI(x) ((unsigned)(x))
+#define 	PACKET3_COPY_DATA__MODE(x) ((((unsigned)(x)) & 0x1) << 21)
+#define 	PACKET3_COPY_DATA__AID_ID(x) ((((unsigned)(x)) & 0x3) << 23)
+#define 	PACKET3_COPY_DATA__DST_TEMPORAL(x) ((((unsigned)(x)) & 0x3) << 25)
+#define 	PACKET3_COPY_DATA__SRC_REG_OFFSET_LO(x) ((unsigned)(x))
+#define 	PACKET3_COPY_DATA__SRC_REG_OFFSET_HI(x) ((((unsigned)(x)) & 0xFF) << 0)
+#define 	PACKET3_COPY_DATA__DST_REG_OFFSET_LO(x) ((unsigned)(x))
+#define 	PACKET3_COPY_DATA__DST_REG_OFFSET_HI(x) ((((unsigned)(x)) & 0xFF) << 0)
+#define 	PACKET3_COPY_DATA__SRC_SEL__MEM_MAPPED_REGISTER 0
+#define 	PACKET3_COPY_DATA__SRC_SEL__TC_L2_OBSOLETE 1
+#define 	PACKET3_COPY_DATA__SRC_SEL__TC_L2 2
+#define 	PACKET3_COPY_DATA__SRC_SEL__GDS 3
+#define 	PACKET3_COPY_DATA__SRC_SEL__PERFCOUNTERS 4
+#define 	PACKET3_COPY_DATA__SRC_SEL__IMMEDIATE_DATA 5
+#define 	PACKET3_COPY_DATA__SRC_SEL__ATOMIC_RETURN_DATA 6
+#define 	PACKET3_COPY_DATA__SRC_SEL__GDS_ATOMIC_RETURN_DATA0 7
+#define 	PACKET3_COPY_DATA__SRC_SEL__GDS_ATOMIC_RETURN_DATA1 8
+#define 	PACKET3_COPY_DATA__SRC_SEL__GPU_CLOCK_COUNT 9
+#define 	PACKET3_COPY_DATA__SRC_SEL__SYSTEM_CLOCK_COUNT 10
+#define 	PACKET3_COPY_DATA__DST_SEL__MEM_MAPPED_REGISTER 0
+#define 	PACKET3_COPY_DATA__DST_SEL__TC_L2 2
+#define 	PACKET3_COPY_DATA__DST_SEL__GDS 3
+#define 	PACKET3_COPY_DATA__DST_SEL__PERFCOUNTERS 4
+#define 	PACKET3_COPY_DATA__DST_SEL__TC_L2_OBSOLETE 5
+#define 	PACKET3_COPY_DATA__DST_SEL__MEM_MAPPED_REG_DC 6
+#define 	PACKET3_COPY_DATA__SRC_TEMPORAL__RT 0
+#define 	PACKET3_COPY_DATA__SRC_TEMPORAL__NT 1
+#define 	PACKET3_COPY_DATA__SRC_TEMPORAL__HT 2
+#define 	PACKET3_COPY_DATA__SRC_TEMPORAL__LU 3
+#define 	PACKET3_COPY_DATA__SRC_CACHE_POLICY__LRU 0
+#define 	PACKET3_COPY_DATA__SRC_CACHE_POLICY__STREAM 1
+#define 	PACKET3_COPY_DATA__SRC_CACHE_POLICY__NOA 2
+#define 	PACKET3_COPY_DATA__SRC_CACHE_POLICY__BYPASS 3
+#define 	PACKET3_COPY_DATA__COUNT_SEL__32_BITS_OF_DATA 0
+#define 	PACKET3_COPY_DATA__COUNT_SEL__64_BITS_OF_DATA 1
+#define 	PACKET3_COPY_DATA__WR_CONFIRM__DO_NOT_WAIT_FOR_CONFIRMATION 0
+#define 	PACKET3_COPY_DATA__WR_CONFIRM__WAIT_FOR_CONFIRMATION 1
+#define 	PACKET3_COPY_DATA__MODE__PF_VF_DISABLED 0
+#define 	PACKET3_COPY_DATA__MODE__PF_VF_ENABLED 1
+#define 	PACKET3_COPY_DATA__DST_TEMPORAL__RT 0
+#define 	PACKET3_COPY_DATA__DST_TEMPORAL__NT 1
+#define 	PACKET3_COPY_DATA__DST_TEMPORAL__HT 2
+#define 	PACKET3_COPY_DATA__DST_TEMPORAL__LU 3
+#define 	PACKET3_COPY_DATA__DST_CACHE_POLICY__LRU 0
+#define 	PACKET3_COPY_DATA__DST_CACHE_POLICY__STREAM 1
+#define 	PACKET3_COPY_DATA__DST_CACHE_POLICY__NOA 2
+#define 	PACKET3_COPY_DATA__DST_CACHE_POLICY__BYPASS 3
+#define 	PACKET3_COPY_DATA__PQ_EXE_STATUS__DEFAULT 0
+#define 	PACKET3_COPY_DATA__PQ_EXE_STATUS__PHASE_UPDATE 1
 #define	PACKET3_CP_DMA					0x41
 #define	PACKET3_PFP_SYNC_ME				0x42
 #define	PACKET3_SURFACE_SYNC				0x43
@@ -160,6 +332,23 @@
 		 * 3 - SAMPLE_STREAMOUTSTAT*
 		 * 4 - *S_PARTIAL_FLUSH
 		 */
+#define		PACKET3_EVENT_WRITE__EVENT_TYPE(x) ((((unsigned)(x)) & 0x3F) << 0)
+#define		PACKET3_EVENT_WRITE__EVENT_INDEX(x) ((((unsigned)(x)) & 0xF) << 8)
+#define		PACKET3_EVENT_WRITE__SAMP_PLST_CNTR_MODE(x) ((((unsigned)(x)) & 0x3) << 29)
+#define		PACKET3_EVENT_WRITE__OFFLOAD_ENABLE(x) ((((unsigned)(x)) & 0x1) << 0)
+#define 	PACKET3_EVENT_WRITE__ADDRESS_LO(x) ((((unsigned)(x)) & 0x1FFFFFFF) << 3)
+#define 	PACKET3_EVENT_WRITE__ADDRESS_HI(x) ((unsigned)(x))
+#define 	PACKET3_EVENT_WRITE__EVENT_INDEX__OTHER 0
+#define 	PACKET3_EVENT_WRITE__EVENT_INDEX__SAMPLE_PIPELINESTAT 2
+#define 	PACKET3_EVENT_WRITE__EVENT_INDEX__CS_PARTIAL_FLUSH 4
+#define 	PACKET3_EVENT_WRITE__EVENT_INDEX__SAMPLE_STREAMOUTSTATS 8
+#define 	PACKET3_EVENT_WRITE__EVENT_INDEX__SAMPLE_STREAMOUTSTATS1 9
+#define 	PACKET3_EVENT_WRITE__EVENT_INDEX__SAMPLE_STREAMOUTSTATS2 10
+#define 	PACKET3_EVENT_WRITE__EVENT_INDEX__SAMPLE_STREAMOUTSTATS3 11
+#define 	PACKET3_EVENT_WRITE__SAMP_PLST_CNTR_MODE__LEGACY_MODE 0
+#define 	PACKET3_EVENT_WRITE__SAMP_PLST_CNTR_MODE__MIXED_MODE1 1
+#define 	PACKET3_EVENT_WRITE__SAMP_PLST_CNTR_MODE__NEW_MODE 2
+#define 	PACKET3_EVENT_WRITE__SAMP_PLST_CNTR_MODE__MIXED_MODE3 3
 #define	PACKET3_EVENT_WRITE_EOP				0x47
 #define	PACKET3_EVENT_WRITE_EOS				0x48
 #define	PACKET3_RELEASE_MEM				0x49
@@ -304,6 +493,12 @@
 		 * 2: REVERSE
 		 */
 #define 	PACKET3_ACQUIRE_MEM_GCR_RANGE_IS_PA  (1 << 18)
+#define 	PACKET3_ACQUIRE_MEM__COHER_SIZE(x) ((unsigned)(x))
+#define 	PACKET3_ACQUIRE_MEM__COHER_SIZE_HI(x) ((((unsigned)(x)) & 0xFF) << 0)
+#define 	PACKET3_ACQUIRE_MEM__COHER_BASE_LO(x) ((unsigned)(x))
+#define 	PACKET3_ACQUIRE_MEM__COHER_BASE_HI(x) ((((unsigned)(x)) & 0xFFFFFF) << 0)
+#define 	PACKET3_ACQUIRE_MEM__POLL_INTERVAL(x) ((((unsigned)(x)) & 0xFFFF) << 0)
+#define 	PACKET3_ACQUIRE_MEM__GCR_CNTL(x) ((((unsigned)(x)) & 0x7FFFF) << 0)
 #define	PACKET3_REWIND					0x59
 #define	PACKET3_INTERRUPT				0x5A
 #define	PACKET3_GEN_PDEPTE				0x5B
@@ -330,11 +525,17 @@
 #define	PACKET3_SET_SH_REG				0x76
 #define		PACKET3_SET_SH_REG_START			0x00002c00
 #define		PACKET3_SET_SH_REG_END				0x00003000
+#define 	PACKET3_SET_SH_REG__REG_OFFSET(x) ((((unsigned)(x)) & 0xFFFF) << 0)
+#define 	PACKET3_SET_SH_REG__VMID_SHIFT(x) ((((unsigned)(x)) & 0x1F) << 23)
+#define 	PACKET3_SET_SH_REG__INDEX(x) ((((unsigned)(x)) & 0xF) << 28)
+#define 	PACKET3_SET_SH_REG__INDEX__DEFAULT 0
+#define 	PACKET3_SET_SH_REG__INDEX__INSERT_VMID 1
 #define	PACKET3_SET_SH_REG_OFFSET			0x77
 #define	PACKET3_SET_QUEUE_REG				0x78
 #define	PACKET3_SET_UCONFIG_REG				0x79
 #define		PACKET3_SET_UCONFIG_REG_START			0x0000c000
 #define		PACKET3_SET_UCONFIG_REG_END			0x0000c400
+#define 	PACKET3_SET_UCONFIG_REG__REG_OFFSET(x) ((((unsigned)(x)) & 0xFFFF) << 0)
 #define	PACKET3_SET_UCONFIG_REG_INDEX			0x7A
 #define	PACKET3_FORWARD_HEADER				0x7C
 #define	PACKET3_SCRATCH_RAM_WRITE			0x7D
@@ -369,6 +570,7 @@
 #              define PACKET3_INVALIDATE_TLBS_DST_SEL(x)     ((x) << 0)
 #              define PACKET3_INVALIDATE_TLBS_ALL_HUB(x)     ((x) << 4)
 #              define PACKET3_INVALIDATE_TLBS_PASID(x)       ((x) << 5)
+#              define PACKET3_INVALIDATE_TLBS_FLUSH_TYPE(x)  ((x) << 29)
 #define	PACKET3_AQL_PACKET				0x99
 #define	PACKET3_DMA_DATA_FILL_MULTI			0x9A
 #define	PACKET3_SET_SH_REG_INDEX			0x9B
@@ -463,5 +665,14 @@
 #define	PACKET3_RUN_LIST				0xA5
 #define	PACKET3_MAP_PROCESS_VM				0xA6
 
+#define PACKET3_RUN_CLEANER_SHADER                      0xD2
+/* 1. header
+ * 2. RESERVED [31:0]
+ */
+
+/* GFX11 */
+#define	PACKET3_SET_Q_PREEMPTION_MODE			0xF0
+#              define PACKET3_SET_Q_PREEMPTION_MODE_IB_VMID(x)  ((x) << 0)
+#              define PACKET3_SET_Q_PREEMPTION_MODE_INIT_SHADOW_MEM    (1 << 0)
 
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/psp_gfx_if.h b/drivers/gpu/drm/amd/amdgpu/psp_gfx_if.h
index dd0dce254901..73f87131a7e9 100644
--- a/drivers/gpu/drm/amd/amdgpu/psp_gfx_if.h
+++ b/drivers/gpu/drm/amd/amdgpu/psp_gfx_if.h
@@ -102,6 +102,13 @@ enum psp_gfx_cmd_id
     GFX_CMD_ID_LOAD_TOC           = 0x00000020,   /* Load TOC and obtain TMR size */
     GFX_CMD_ID_AUTOLOAD_RLC       = 0x00000021,   /* Indicates all graphics fw loaded, start RLC autoload */
     GFX_CMD_ID_BOOT_CFG           = 0x00000022,   /* Boot Config */
+    GFX_CMD_ID_SRIOV_SPATIAL_PART = 0x00000027,   /* Configure spatial partitioning mode */
+    /*IDs of performance monitoring/profiling*/
+    GFX_CMD_ID_CONFIG_SQ_PERFMON  = 0x00000046,   /* Config CGTT_SQ_CLK_CTRL */
+    /* Dynamic memory partitioninig (NPS mode change)*/
+    GFX_CMD_ID_FB_NPS_MODE        = 0x00000048,  /* Configure memory partitioning mode */
+    GFX_CMD_ID_FB_FW_RESERV_ADDR  = 0x00000050,  /* Query FW reservation addr */
+    GFX_CMD_ID_FB_FW_RESERV_EXT_ADDR = 0x00000051,  /* Query FW reservation extended addr */
 };
 
 /* PSP boot config sub-commands */
@@ -258,7 +265,45 @@ enum psp_gfx_fw_type {
 	GFX_FW_TYPE_SDMA6                           = 56,   /* SDMA6                    MI      */
 	GFX_FW_TYPE_SDMA7                           = 57,   /* SDMA7                    MI      */
 	GFX_FW_TYPE_VCN1                            = 58,   /* VCN1                     MI      */
+	GFX_FW_TYPE_CAP                             = 62,   /* CAP_FW                           */
+	GFX_FW_TYPE_SE2_TAP_DELAYS                  = 65,   /* SE2 TAP DELAYS           NV      */
+	GFX_FW_TYPE_SE3_TAP_DELAYS                  = 66,   /* SE3 TAP DELAYS           NV      */
 	GFX_FW_TYPE_REG_LIST                        = 67,   /* REG_LIST                 MI      */
+	GFX_FW_TYPE_IMU_I                           = 68,   /* IMU Instruction FW       SOC21   */
+	GFX_FW_TYPE_IMU_D                           = 69,   /* IMU Data FW              SOC21   */
+	GFX_FW_TYPE_LSDMA                           = 70,   /* LSDMA FW                 SOC21   */
+	GFX_FW_TYPE_SDMA_UCODE_TH0                  = 71,   /* SDMA Thread 0/CTX        SOC21   */
+	GFX_FW_TYPE_SDMA_UCODE_TH1                  = 72,   /* SDMA Thread 1/CTL        SOC21   */
+	GFX_FW_TYPE_PPTABLE                         = 73,   /* PPTABLE                  SOC21   */
+	GFX_FW_TYPE_DISCRETE_USB4                   = 74,   /* dUSB4 FW                 SOC21   */
+	GFX_FW_TYPE_TA                              = 75,   /* SRIOV TA FW UUID         SOC21   */
+	GFX_FW_TYPE_RS64_MES                        = 76,   /* RS64 MES ucode           SOC21   */
+	GFX_FW_TYPE_RS64_MES_STACK                  = 77,   /* RS64 MES stack ucode     SOC21   */
+	GFX_FW_TYPE_RS64_KIQ                        = 78,   /* RS64 KIQ ucode           SOC21   */
+	GFX_FW_TYPE_RS64_KIQ_STACK                  = 79,   /* RS64 KIQ Heap stack      SOC21   */
+	GFX_FW_TYPE_ISP_DATA                        = 80,   /* ISP DATA                 SOC21   */
+	GFX_FW_TYPE_CP_MES_KIQ                      = 81,   /* MES KIQ ucode            SOC21   */
+	GFX_FW_TYPE_MES_KIQ_STACK                   = 82,   /* MES KIQ stack            SOC21   */
+	GFX_FW_TYPE_UMSCH_DATA                      = 83,   /* User Mode Scheduler Data SOC21   */
+	GFX_FW_TYPE_UMSCH_UCODE                     = 84,   /* User Mode Scheduler Ucode SOC21  */
+	GFX_FW_TYPE_UMSCH_CMD_BUFFER                = 85,   /* User Mode Scheduler Command Buffer SOC21 */
+	GFX_FW_TYPE_USB_DP_COMBO_PHY                = 86,   /* USB-Display port Combo   SOC21   */
+	GFX_FW_TYPE_RS64_PFP                        = 87,   /* RS64 PFP                 SOC21   */
+	GFX_FW_TYPE_RS64_ME                         = 88,   /* RS64 ME                  SOC21   */
+	GFX_FW_TYPE_RS64_MEC                        = 89,   /* RS64 MEC                 SOC21   */
+	GFX_FW_TYPE_RS64_PFP_P0_STACK               = 90,   /* RS64 PFP stack P0        SOC21   */
+	GFX_FW_TYPE_RS64_PFP_P1_STACK               = 91,   /* RS64 PFP stack P1        SOC21   */
+	GFX_FW_TYPE_RS64_ME_P0_STACK                = 92,   /* RS64 ME stack P0         SOC21   */
+	GFX_FW_TYPE_RS64_ME_P1_STACK                = 93,   /* RS64 ME stack P1         SOC21   */
+	GFX_FW_TYPE_RS64_MEC_P0_STACK               = 94,   /* RS64 MEC stack P0        SOC21   */
+	GFX_FW_TYPE_RS64_MEC_P1_STACK               = 95,   /* RS64 MEC stack P1        SOC21   */
+	GFX_FW_TYPE_RS64_MEC_P2_STACK               = 96,   /* RS64 MEC stack P2        SOC21   */
+	GFX_FW_TYPE_RS64_MEC_P3_STACK               = 97,   /* RS64 MEC stack P3        SOC21   */
+	GFX_FW_TYPE_VPEC_FW1                        = 100,  /* VPEC FW1 To Save         VPE     */
+	GFX_FW_TYPE_VPEC_FW2                        = 101,  /* VPEC FW2 To Save         VPE     */
+	GFX_FW_TYPE_VPE                             = 102,
+	GFX_FW_TYPE_JPEG_RAM                        = 128,  /**< JPEG Command buffer */
+	GFX_FW_TYPE_P2S_TABLE                       = 129,
 	GFX_FW_TYPE_MAX
 };
 
@@ -305,6 +350,27 @@ struct psp_gfx_cmd_boot_cfg
     uint32_t                        boot_config_valid;    /* dynamic boot configuration valid bits bitmask */
 };
 
+struct psp_gfx_cmd_sriov_spatial_part {
+	uint32_t mode;
+	uint32_t override_ips;
+	uint32_t override_xcds_avail;
+	uint32_t override_this_aid;
+};
+
+/*Structure for sq performance monitoring/profiling enable/disable*/
+struct psp_gfx_cmd_config_sq_perfmon {
+	uint32_t        gfx_xcp_mask;
+	uint8_t         core_override;
+	uint8_t         reg_override;
+	uint8_t         perfmon_override;
+	uint8_t         reserved[5];
+};
+
+struct psp_gfx_cmd_fb_memory_part {
+	uint32_t mode; /* requested NPS mode */
+	uint32_t resvd;
+};
+
 /* All GFX ring buffer commands. */
 union psp_gfx_commands
 {
@@ -318,6 +384,9 @@ union psp_gfx_commands
     struct psp_gfx_cmd_setup_tmr        cmd_setup_vmr;
     struct psp_gfx_cmd_load_toc         cmd_load_toc;
     struct psp_gfx_cmd_boot_cfg         boot_cfg;
+    struct psp_gfx_cmd_sriov_spatial_part cmd_spatial_part;
+    struct psp_gfx_cmd_config_sq_perfmon config_sq_perfmon;
+    struct psp_gfx_cmd_fb_memory_part cmd_memory_part;
 };
 
 struct psp_gfx_uresp_reserved
@@ -337,11 +406,19 @@ struct psp_gfx_uresp_bootcfg {
 	uint32_t boot_cfg;	/* boot config data */
 };
 
+/* Command-specific response for fw reserve info */
+struct psp_gfx_uresp_fw_reserve_info {
+    uint32_t reserve_base_address_hi;
+    uint32_t reserve_base_address_lo;
+    uint32_t reserve_size;
+};
+
 /* Union of command-specific responses for GPCOM ring. */
 union psp_gfx_uresp {
 	struct psp_gfx_uresp_reserved		reserved;
 	struct psp_gfx_uresp_bootcfg		boot_cfg;
 	struct psp_gfx_uresp_fwar_db_info	fwar_db_info;
+	struct psp_gfx_uresp_fw_reserve_info	fw_reserve_info;
 };
 
 /* Structure of GFX Response buffer.
@@ -417,8 +494,9 @@ struct psp_gfx_rb_frame
 #define PSP_ERR_UNKNOWN_COMMAND 0x00000100
 
 enum tee_error_code {
-    TEE_SUCCESS                         = 0x00000000,
-    TEE_ERROR_NOT_SUPPORTED             = 0xFFFF000A,
+	TEE_SUCCESS			= 0x00000000,
+	TEE_ERROR_CANCEL		= 0xFFFF0002,
+	TEE_ERROR_NOT_SUPPORTED		= 0xFFFF000A,
 };
 
 #endif /* _PSP_TEE_GFX_IF_H_ */
diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v10_0.c b/drivers/gpu/drm/amd/amdgpu/psp_v10_0.c
index ed2293686f0d..3584b8c18fd9 100644
--- a/drivers/gpu/drm/amd/amdgpu/psp_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/psp_v10_0.c
@@ -47,111 +47,26 @@ MODULE_FIRMWARE("amdgpu/raven_ta.bin");
 static int psp_v10_0_init_microcode(struct psp_context *psp)
 {
 	struct amdgpu_device *adev = psp->adev;
-	const char *chip_name;
-	char fw_name[30];
+	char ucode_prefix[30];
 	int err = 0;
-	const struct ta_firmware_header_v1_0 *ta_hdr;
 	DRM_DEBUG("\n");
 
-	switch (adev->asic_type) {
-	case CHIP_RAVEN:
-		if (adev->apu_flags & AMD_APU_IS_RAVEN2)
-			chip_name = "raven2";
-		else if (adev->apu_flags & AMD_APU_IS_PICASSO)
-			chip_name = "picasso";
-		else
-			chip_name = "raven";
-		break;
-	default: BUG();
-	}
+	amdgpu_ucode_ip_version_decode(adev, MP0_HWIP, ucode_prefix, sizeof(ucode_prefix));
 
-	err = psp_init_asd_microcode(psp, chip_name);
+	err = psp_init_asd_microcode(psp, ucode_prefix);
 	if (err)
-		goto out;
-
-	snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_ta.bin", chip_name);
-	err = request_firmware(&adev->psp.ta_fw, fw_name, adev->dev);
-	if (err) {
-		release_firmware(adev->psp.ta_fw);
-		adev->psp.ta_fw = NULL;
-		dev_info(adev->dev,
-			 "psp v10.0: Failed to load firmware \"%s\"\n",
-			 fw_name);
-	} else {
-		err = amdgpu_ucode_validate(adev->psp.ta_fw);
-		if (err)
-			goto out2;
-
-		ta_hdr = (const struct ta_firmware_header_v1_0 *)
-				 adev->psp.ta_fw->data;
-		adev->psp.hdcp_context.context.bin_desc.fw_version =
-			le32_to_cpu(ta_hdr->hdcp.fw_version);
-		adev->psp.hdcp_context.context.bin_desc.size_bytes =
-			le32_to_cpu(ta_hdr->hdcp.size_bytes);
-		adev->psp.hdcp_context.context.bin_desc.start_addr =
-			(uint8_t *)ta_hdr +
-			le32_to_cpu(ta_hdr->header.ucode_array_offset_bytes);
-
-		adev->psp.dtm_context.context.bin_desc.fw_version =
-			le32_to_cpu(ta_hdr->dtm.fw_version);
-		adev->psp.dtm_context.context.bin_desc.size_bytes =
-			le32_to_cpu(ta_hdr->dtm.size_bytes);
-		adev->psp.dtm_context.context.bin_desc.start_addr =
-			(uint8_t *)adev->psp.hdcp_context.context.bin_desc.start_addr +
-			le32_to_cpu(ta_hdr->dtm.offset_bytes);
-
-		adev->psp.securedisplay_context.context.bin_desc.fw_version =
-			le32_to_cpu(ta_hdr->securedisplay.fw_version);
-		adev->psp.securedisplay_context.context.bin_desc.size_bytes =
-			le32_to_cpu(ta_hdr->securedisplay.size_bytes);
-		adev->psp.securedisplay_context.context.bin_desc.start_addr =
-			(uint8_t *)adev->psp.hdcp_context.context.bin_desc.start_addr +
-			le32_to_cpu(ta_hdr->securedisplay.offset_bytes);
-
-		adev->psp.ta_fw_version = le32_to_cpu(ta_hdr->header.ucode_version);
+		return err;
+
+	err = psp_init_ta_microcode(psp, ucode_prefix);
+	if ((amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 1, 0)) &&
+	    (adev->pdev->revision == 0xa1) &&
+	    (psp->securedisplay_context.context.bin_desc.fw_version >=
+	     0x27000008)) {
+		adev->psp.securedisplay_context.context.bin_desc.size_bytes = 0;
 	}
-
-	return 0;
-
-out2:
-	release_firmware(adev->psp.ta_fw);
-	adev->psp.ta_fw = NULL;
-out:
-	if (err) {
-		dev_err(adev->dev,
-			"psp v10.0: Failed to load firmware \"%s\"\n",
-			fw_name);
-	}
-
 	return err;
 }
 
-static int psp_v10_0_ring_init(struct psp_context *psp,
-			       enum psp_ring_type ring_type)
-{
-	int ret = 0;
-	struct psp_ring *ring;
-	struct amdgpu_device *adev = psp->adev;
-
-	ring = &psp->km_ring;
-
-	ring->ring_type = ring_type;
-
-	/* allocate 4k Page of Local Frame Buffer memory for ring */
-	ring->ring_size = 0x1000;
-	ret = amdgpu_bo_create_kernel(adev, ring->ring_size, PAGE_SIZE,
-				      AMDGPU_GEM_DOMAIN_VRAM,
-				      &adev->firmware.rbuf,
-				      &ring->ring_mem_mc_addr,
-				      (void **)&ring->ring_mem);
-	if (ret) {
-		ring->ring_size = 0;
-		return ret;
-	}
-
-	return 0;
-}
-
 static int psp_v10_0_ring_create(struct psp_context *psp,
 				 enum psp_ring_type ring_type)
 {
@@ -179,7 +94,7 @@ static int psp_v10_0_ring_create(struct psp_context *psp,
 
 	/* Wait for response flag (bit 31) in C2PMSG_64 */
 	ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_64),
-			   0x80000000, 0x8000FFFF, false);
+			   MBOX_TOS_RESP_FLAG, MBOX_TOS_RESP_MASK, 0);
 
 	return ret;
 }
@@ -200,7 +115,7 @@ static int psp_v10_0_ring_stop(struct psp_context *psp,
 
 	/* Wait for response flag (bit 31) in C2PMSG_64 */
 	ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_64),
-			   0x80000000, 0x80000000, false);
+			   MBOX_TOS_RESP_FLAG, MBOX_TOS_RESP_MASK, 0);
 
 	return ret;
 }
@@ -245,7 +160,6 @@ static void psp_v10_0_ring_set_wptr(struct psp_context *psp, uint32_t value)
 
 static const struct psp_funcs psp_v10_0_funcs = {
 	.init_microcode = psp_v10_0_init_microcode,
-	.ring_init = psp_v10_0_ring_init,
 	.ring_create = psp_v10_0_ring_create,
 	.ring_stop = psp_v10_0_ring_stop,
 	.ring_destroy = psp_v10_0_ring_destroy,
diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c b/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c
index 2176ef85f137..a9be7a505026 100644
--- a/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c
@@ -53,11 +53,13 @@ MODULE_FIRMWARE("amdgpu/navi14_ta.bin");
 MODULE_FIRMWARE("amdgpu/navi12_sos.bin");
 MODULE_FIRMWARE("amdgpu/navi12_asd.bin");
 MODULE_FIRMWARE("amdgpu/navi12_ta.bin");
+MODULE_FIRMWARE("amdgpu/navi12_cap.bin");
 MODULE_FIRMWARE("amdgpu/arcturus_sos.bin");
 MODULE_FIRMWARE("amdgpu/arcturus_asd.bin");
 MODULE_FIRMWARE("amdgpu/arcturus_ta.bin");
 MODULE_FIRMWARE("amdgpu/sienna_cichlid_sos.bin");
 MODULE_FIRMWARE("amdgpu/sienna_cichlid_ta.bin");
+MODULE_FIRMWARE("amdgpu/sienna_cichlid_cap.bin");
 MODULE_FIRMWARE("amdgpu/navy_flounder_sos.bin");
 MODULE_FIRMWARE("amdgpu/navy_flounder_ta.bin");
 MODULE_FIRMWARE("amdgpu/vangogh_asd.bin");
@@ -86,179 +88,97 @@ MODULE_FIRMWARE("amdgpu/beige_goby_ta.bin");
 static int psp_v11_0_init_microcode(struct psp_context *psp)
 {
 	struct amdgpu_device *adev = psp->adev;
-	const char *chip_name;
-	char fw_name[PSP_FW_NAME_LEN];
+	char ucode_prefix[30];
 	int err = 0;
-	const struct ta_firmware_header_v1_0 *ta_hdr;
 
 	DRM_DEBUG("\n");
 
-	switch (adev->ip_versions[MP0_HWIP][0]) {
-	case IP_VERSION(11, 0, 2):
-		chip_name = "vega20";
-		break;
-	case IP_VERSION(11, 0, 0):
-		chip_name = "navi10";
-		break;
-	case IP_VERSION(11, 0, 5):
-		chip_name = "navi14";
-		break;
-	case IP_VERSION(11, 0, 9):
-		chip_name = "navi12";
-		break;
-	case IP_VERSION(11, 0, 4):
-		chip_name = "arcturus";
-		break;
-	case IP_VERSION(11, 0, 7):
-		chip_name = "sienna_cichlid";
-		break;
-	case IP_VERSION(11, 0, 11):
-		chip_name = "navy_flounder";
-		break;
-	case IP_VERSION(11, 5, 0):
-		chip_name = "vangogh";
-		break;
-	case IP_VERSION(11, 0, 12):
-		chip_name = "dimgrey_cavefish";
-		break;
-	case IP_VERSION(11, 0, 13):
-		chip_name = "beige_goby";
-		break;
-	default:
-		BUG();
-	}
-
+	amdgpu_ucode_ip_version_decode(adev, MP0_HWIP, ucode_prefix, sizeof(ucode_prefix));
 
-	switch (adev->ip_versions[MP0_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, MP0_HWIP, 0)) {
 	case IP_VERSION(11, 0, 2):
 	case IP_VERSION(11, 0, 4):
-		err = psp_init_sos_microcode(psp, chip_name);
+		err = psp_init_sos_microcode(psp, ucode_prefix);
 		if (err)
 			return err;
-		err = psp_init_asd_microcode(psp, chip_name);
+		err = psp_init_asd_microcode(psp, ucode_prefix);
 		if (err)
 			return err;
-		snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_ta.bin", chip_name);
-		err = request_firmware(&adev->psp.ta_fw, fw_name, adev->dev);
-		if (err) {
-			release_firmware(adev->psp.ta_fw);
-			adev->psp.ta_fw = NULL;
-			dev_info(adev->dev,
-				 "psp v11.0: Failed to load firmware \"%s\"\n", fw_name);
-		} else {
-			err = amdgpu_ucode_validate(adev->psp.ta_fw);
-			if (err)
-				goto out2;
-
-			ta_hdr = (const struct ta_firmware_header_v1_0 *)adev->psp.ta_fw->data;
-			adev->psp.xgmi_context.context.bin_desc.fw_version =
-				le32_to_cpu(ta_hdr->xgmi.fw_version);
-			adev->psp.xgmi_context.context.bin_desc.size_bytes =
-				le32_to_cpu(ta_hdr->xgmi.size_bytes);
-			adev->psp.xgmi_context.context.bin_desc.start_addr =
-				(uint8_t *)ta_hdr +
-				le32_to_cpu(ta_hdr->header.ucode_array_offset_bytes);
-			adev->psp.ta_fw_version = le32_to_cpu(ta_hdr->header.ucode_version);
-			adev->psp.ras_context.context.bin_desc.fw_version =
-				le32_to_cpu(ta_hdr->ras.fw_version);
-			adev->psp.ras_context.context.bin_desc.size_bytes =
-				le32_to_cpu(ta_hdr->ras.size_bytes);
-			adev->psp.ras_context.context.bin_desc.start_addr =
-				(uint8_t *)adev->psp.xgmi_context.context.bin_desc.start_addr +
-				le32_to_cpu(ta_hdr->ras.offset_bytes);
-		}
+		err = psp_init_ta_microcode(psp, ucode_prefix);
+		adev->psp.securedisplay_context.context.bin_desc.size_bytes = 0;
 		break;
 	case IP_VERSION(11, 0, 0):
 	case IP_VERSION(11, 0, 5):
 	case IP_VERSION(11, 0, 9):
-		err = psp_init_sos_microcode(psp, chip_name);
+		err = psp_init_sos_microcode(psp, ucode_prefix);
 		if (err)
 			return err;
-		err = psp_init_asd_microcode(psp, chip_name);
+		err = psp_init_asd_microcode(psp, ucode_prefix);
 		if (err)
 			return err;
-		if (amdgpu_sriov_vf(adev))
-			break;
-		snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_ta.bin", chip_name);
-		err = request_firmware(&adev->psp.ta_fw, fw_name, adev->dev);
-		if (err) {
-			release_firmware(adev->psp.ta_fw);
-			adev->psp.ta_fw = NULL;
-			dev_info(adev->dev,
-				 "psp v11.0: Failed to load firmware \"%s\"\n", fw_name);
-		} else {
-			err = amdgpu_ucode_validate(adev->psp.ta_fw);
-			if (err)
-				goto out2;
-
-			ta_hdr = (const struct ta_firmware_header_v1_0 *)adev->psp.ta_fw->data;
-			adev->psp.hdcp_context.context.bin_desc.fw_version =
-				le32_to_cpu(ta_hdr->hdcp.fw_version);
-			adev->psp.hdcp_context.context.bin_desc.size_bytes =
-				le32_to_cpu(ta_hdr->hdcp.size_bytes);
-			adev->psp.hdcp_context.context.bin_desc.start_addr =
-				(uint8_t *)ta_hdr +
-				le32_to_cpu(
-					ta_hdr->header.ucode_array_offset_bytes);
-
-			adev->psp.ta_fw_version = le32_to_cpu(ta_hdr->header.ucode_version);
-
-			adev->psp.dtm_context.context.bin_desc.fw_version =
-				le32_to_cpu(ta_hdr->dtm.fw_version);
-			adev->psp.dtm_context.context.bin_desc.size_bytes =
-				le32_to_cpu(ta_hdr->dtm.size_bytes);
-			adev->psp.dtm_context.context.bin_desc.start_addr =
-				(uint8_t *)adev->psp.hdcp_context.context
-					.bin_desc.start_addr +
-				le32_to_cpu(ta_hdr->dtm.offset_bytes);
-		}
+		err = psp_init_ta_microcode(psp, ucode_prefix);
+		adev->psp.securedisplay_context.context.bin_desc.size_bytes = 0;
 		break;
 	case IP_VERSION(11, 0, 7):
 	case IP_VERSION(11, 0, 11):
 	case IP_VERSION(11, 0, 12):
 	case IP_VERSION(11, 0, 13):
-		err = psp_init_sos_microcode(psp, chip_name);
-		if (err)
-			return err;
-		err = psp_init_ta_microcode(psp, chip_name);
+		err = psp_init_sos_microcode(psp, ucode_prefix);
 		if (err)
 			return err;
+		err = psp_init_ta_microcode(psp, ucode_prefix);
 		break;
 	case IP_VERSION(11, 5, 0):
-		err = psp_init_asd_microcode(psp, chip_name);
-		if (err)
-			return err;
-		err = psp_init_toc_microcode(psp, chip_name);
+	case IP_VERSION(11, 5, 2):
+		err = psp_init_asd_microcode(psp, ucode_prefix);
 		if (err)
 			return err;
+		err = psp_init_toc_microcode(psp, ucode_prefix);
 		break;
 	default:
 		BUG();
 	}
 
-	return 0;
-
-out2:
-	release_firmware(adev->psp.ta_fw);
-	adev->psp.ta_fw = NULL;
 	return err;
 }
 
-static int psp_v11_0_wait_for_bootloader(struct psp_context *psp)
+static int psp_v11_wait_for_tos_unload(struct psp_context *psp)
 {
 	struct amdgpu_device *adev = psp->adev;
+	uint32_t sol_reg1, sol_reg2;
+	int retry_loop;
+
+	/* Wait for the TOS to be unloaded */
+	for (retry_loop = 0; retry_loop < 20; retry_loop++) {
+		sol_reg1 = RREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_81);
+		usleep_range(1000, 2000);
+		sol_reg2 = RREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_81);
+		if (sol_reg1 == sol_reg2)
+			return 0;
+	}
+	dev_err(adev->dev, "TOS unload failed, C2PMSG_33: %x C2PMSG_81: %x",
+		RREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_33),
+		RREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_81));
 
+	return -ETIME;
+}
+
+static int psp_v11_0_wait_for_bootloader(struct psp_context *psp)
+{
+	struct amdgpu_device *adev = psp->adev;
 	int ret;
 	int retry_loop;
 
-	for (retry_loop = 0; retry_loop < 10; retry_loop++) {
+	/* For a reset done at the end of S3, only wait for TOS to be unloaded */
+	if (adev->in_s3 && !(adev->flags & AMD_IS_APU) && amdgpu_in_reset(adev))
+		return psp_v11_wait_for_tos_unload(psp);
+
+	for (retry_loop = 0; retry_loop < 20; retry_loop++) {
 		/* Wait for bootloader to signify that is
 		    ready having bit 31 of C2PMSG_35 set to 1 */
-		ret = psp_wait_for(psp,
-				   SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_35),
-				   0x80000000,
-				   0x80000000,
-				   false);
+		ret = psp_wait_for(
+			psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_35),
+			0x80000000, 0x8000FFFF, PSP_WAITREG_NOVERBOSE);
 
 		if (ret == 0)
 			return 0;
@@ -277,13 +197,15 @@ static bool psp_v11_0_is_sos_alive(struct psp_context *psp)
 	return sol_reg != 0x0;
 }
 
-static int psp_v11_0_bootloader_load_kdb(struct psp_context *psp)
+static int psp_v11_0_bootloader_load_component(struct psp_context  	*psp,
+					       struct psp_bin_desc 	*bin_desc,
+					       enum psp_bootloader_cmd  bl_cmd)
 {
 	int ret;
 	uint32_t psp_gfxdrv_command_reg = 0;
 	struct amdgpu_device *adev = psp->adev;
 
-	/* Check tOS sign of life register to confirm sys driver and sOS
+	/* Check sOS sign of life register to confirm sys driver and sOS
 	 * are already been loaded.
 	 */
 	if (psp_v11_0_is_sos_alive(psp))
@@ -293,13 +215,13 @@ static int psp_v11_0_bootloader_load_kdb(struct psp_context *psp)
 	if (ret)
 		return ret;
 
-	/* Copy PSP KDB binary to memory */
-	psp_copy_fw(psp, psp->kdb.start_addr, psp->kdb.size_bytes);
+	/* Copy PSP System Driver binary to memory */
+	psp_copy_fw(psp, bin_desc->start_addr, bin_desc->size_bytes);
 
-	/* Provide the PSP KDB to bootloader */
+	/* Provide the sys driver to bootloader */
 	WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_36,
 	       (uint32_t)(psp->fw_pri_mc_addr >> 20));
-	psp_gfxdrv_command_reg = PSP_BL__LOAD_KEY_DATABASE;
+	psp_gfxdrv_command_reg = bl_cmd;
 	WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_35,
 	       psp_gfxdrv_command_reg);
 
@@ -308,69 +230,19 @@ static int psp_v11_0_bootloader_load_kdb(struct psp_context *psp)
 	return ret;
 }
 
-static int psp_v11_0_bootloader_load_spl(struct psp_context *psp)
+static int psp_v11_0_bootloader_load_kdb(struct psp_context *psp)
 {
-	int ret;
-	uint32_t psp_gfxdrv_command_reg = 0;
-	struct amdgpu_device *adev = psp->adev;
-
-	/* Check tOS sign of life register to confirm sys driver and sOS
-	 * are already been loaded.
-	 */
-	if (psp_v11_0_is_sos_alive(psp))
-		return 0;
-
-	ret = psp_v11_0_wait_for_bootloader(psp);
-	if (ret)
-		return ret;
-
-	/* Copy PSP SPL binary to memory */
-	psp_copy_fw(psp, psp->spl.start_addr, psp->spl.size_bytes);
-
-	/* Provide the PSP SPL to bootloader */
-	WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_36,
-	       (uint32_t)(psp->fw_pri_mc_addr >> 20));
-	psp_gfxdrv_command_reg = PSP_BL__LOAD_TOS_SPL_TABLE;
-	WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_35,
-	       psp_gfxdrv_command_reg);
-
-	ret = psp_v11_0_wait_for_bootloader(psp);
+	return psp_v11_0_bootloader_load_component(psp, &psp->kdb, PSP_BL__LOAD_KEY_DATABASE);
+}
 
-	return ret;
+static int psp_v11_0_bootloader_load_spl(struct psp_context *psp)
+{
+	return psp_v11_0_bootloader_load_component(psp, &psp->spl, PSP_BL__LOAD_TOS_SPL_TABLE);
 }
 
 static int psp_v11_0_bootloader_load_sysdrv(struct psp_context *psp)
 {
-	int ret;
-	uint32_t psp_gfxdrv_command_reg = 0;
-	struct amdgpu_device *adev = psp->adev;
-
-	/* Check sOS sign of life register to confirm sys driver and sOS
-	 * are already been loaded.
-	 */
-	if (psp_v11_0_is_sos_alive(psp))
-		return 0;
-
-	ret = psp_v11_0_wait_for_bootloader(psp);
-	if (ret)
-		return ret;
-
-	/* Copy PSP System Driver binary to memory */
-	psp_copy_fw(psp, psp->sys.start_addr, psp->sys.size_bytes);
-
-	/* Provide the sys driver to bootloader */
-	WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_36,
-	       (uint32_t)(psp->fw_pri_mc_addr >> 20));
-	psp_gfxdrv_command_reg = PSP_BL__LOAD_SYSDRV;
-	WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_35,
-	       psp_gfxdrv_command_reg);
-
-	/* there might be handshake issue with hardware which needs delay */
-	mdelay(20);
-
-	ret = psp_v11_0_wait_for_bootloader(psp);
-
-	return ret;
+	return psp_v11_0_bootloader_load_component(psp, &psp->sys, PSP_BL__LOAD_SYSDRV);
 }
 
 static int psp_v11_0_bootloader_load_sos(struct psp_context *psp)
@@ -402,38 +274,12 @@ static int psp_v11_0_bootloader_load_sos(struct psp_context *psp)
 	/* there might be handshake issue with hardware which needs delay */
 	mdelay(20);
 	ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_81),
-			   RREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_81),
-			   0, true);
+			   RREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_81), 0,
+			   PSP_WAITREG_CHANGED);
 
 	return ret;
 }
 
-static int psp_v11_0_ring_init(struct psp_context *psp,
-			      enum psp_ring_type ring_type)
-{
-	int ret = 0;
-	struct psp_ring *ring;
-	struct amdgpu_device *adev = psp->adev;
-
-	ring = &psp->km_ring;
-
-	ring->ring_type = ring_type;
-
-	/* allocate 4k Page of Local Frame Buffer memory for ring */
-	ring->ring_size = 0x1000;
-	ret = amdgpu_bo_create_kernel(adev, ring->ring_size, PAGE_SIZE,
-				      AMDGPU_GEM_DOMAIN_VRAM,
-				      &adev->firmware.rbuf,
-				      &ring->ring_mem_mc_addr,
-				      (void **)&ring->ring_mem);
-	if (ret) {
-		ring->ring_size = 0;
-		return ret;
-	}
-
-	return 0;
-}
-
 static int psp_v11_0_ring_stop(struct psp_context *psp,
 			      enum psp_ring_type ring_type)
 {
@@ -453,11 +299,13 @@ static int psp_v11_0_ring_stop(struct psp_context *psp,
 
 	/* Wait for response flag (bit 31) */
 	if (amdgpu_sriov_vf(adev))
-		ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_101),
-				   0x80000000, 0x80000000, false);
+		ret = psp_wait_for(
+			psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_101),
+			MBOX_TOS_RESP_FLAG, MBOX_TOS_RESP_MASK, 0);
 	else
-		ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_64),
-				   0x80000000, 0x80000000, false);
+		ret = psp_wait_for(
+			psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_64),
+			MBOX_TOS_RESP_FLAG, MBOX_TOS_RESP_MASK, 0);
 
 	return ret;
 }
@@ -493,13 +341,15 @@ static int psp_v11_0_ring_create(struct psp_context *psp,
 		mdelay(20);
 
 		/* Wait for response flag (bit 31) in C2PMSG_101 */
-		ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_101),
-				   0x80000000, 0x8000FFFF, false);
+		ret = psp_wait_for(
+			psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_101),
+			MBOX_TOS_RESP_FLAG, MBOX_TOS_RESP_MASK, 0);
 
 	} else {
 		/* Wait for sOS ready for ring creation */
-		ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_64),
-				   0x80000000, 0x80000000, false);
+		ret = psp_wait_for(
+			psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_64),
+			MBOX_TOS_READY_FLAG, MBOX_TOS_READY_MASK, 0);
 		if (ret) {
 			DRM_ERROR("Failed to wait for sOS ready for ring creation\n");
 			return ret;
@@ -523,8 +373,9 @@ static int psp_v11_0_ring_create(struct psp_context *psp,
 		mdelay(20);
 
 		/* Wait for response flag (bit 31) in C2PMSG_64 */
-		ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_64),
-				   0x80000000, 0x8000FFFF, false);
+		ret = psp_wait_for(
+			psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_64),
+			MBOX_TOS_RESP_FLAG, MBOX_TOS_RESP_MASK, 0);
 	}
 
 	return ret;
@@ -557,7 +408,8 @@ static int psp_v11_0_mode1_reset(struct psp_context *psp)
 
 	offset = SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_64);
 
-	ret = psp_wait_for(psp, offset, 0x80000000, 0x8000FFFF, false);
+	ret = psp_wait_for(psp, offset, MBOX_TOS_READY_FLAG,
+			   MBOX_TOS_READY_MASK, 0);
 
 	if (ret) {
 		DRM_INFO("psp is not working correctly before mode1 reset!\n");
@@ -569,17 +421,6 @@ static int psp_v11_0_mode1_reset(struct psp_context *psp)
 
 	msleep(500);
 
-	offset = SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_33);
-
-	ret = psp_wait_for(psp, offset, 0x80000000, 0x80000000, false);
-
-	if (ret) {
-		DRM_INFO("psp mode 1 reset failed!\n");
-		return -EINVAL;
-	}
-
-	DRM_INFO("psp mode1 reset succeed \n");
-
 	return 0;
 }
 
@@ -597,8 +438,9 @@ static int psp_v11_0_memory_training_send_msg(struct psp_context *psp, int msg)
 
 	max_wait = MEM_TRAIN_SEND_MSG_TIMEOUT_US / adev->usec_timeout;
 	for (i = 0; i < max_wait; i++) {
-		ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_35),
-				   0x80000000, 0x80000000, false);
+		ret = psp_wait_for(
+			psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_35),
+			0x80000000, 0x80000000, PSP_WAITREG_NOVERBOSE);
 		if (ret == 0)
 			break;
 	}
@@ -683,7 +525,7 @@ static int psp_v11_0_memory_training(struct psp_context *psp, uint32_t ops)
 		 * before training, and restore it after training to avoid
 		 * VRAM corruption.
 		 */
-		sz = GDDR6_MEM_TRAINING_ENCROACHED_SIZE;
+		sz = BIST_MEM_TRAINING_ENCROACHED_SIZE;
 
 		if (adev->gmc.visible_vram_size < sz || !adev->mman.aper_base_kaddr) {
 			DRM_ERROR("visible_vram_size %llx or aper_base_kaddr %p is not initialized.\n",
@@ -709,7 +551,7 @@ static int psp_v11_0_memory_training(struct psp_context *psp, uint32_t ops)
 			}
 
 			memcpy_toio(adev->mman.aper_base_kaddr, buf, sz);
-			adev->hdp.funcs->flush_hdp(adev, NULL);
+			amdgpu_device_flush_hdp(adev, NULL);
 			vfree(buf);
 			drm_dev_exit(idx);
 		} else {
@@ -777,7 +619,7 @@ static int psp_v11_0_load_usbc_pd_fw(struct psp_context *psp, uint64_t fw_pri_mc
 	WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_36, (fw_pri_mc_addr >> 20));
 
 	ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_35),
-			     0x80000000, 0x80000000, false);
+			   0x80000000, 0x80000000, 0);
 	if (ret)
 		return ret;
 
@@ -814,7 +656,7 @@ static int psp_v11_0_read_usbc_pd_fw(struct psp_context *psp, uint32_t *fw_ver)
 	WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_35, C2PMSG_CMD_GFX_USB_PD_FW_VER);
 
 	ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_35),
-				     0x80000000, 0x80000000, false);
+			   0x80000000, 0x80000000, 0);
 	if (!ret)
 		*fw_ver = RREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_36);
 
@@ -827,7 +669,6 @@ static const struct psp_funcs psp_v11_0_funcs = {
 	.bootloader_load_spl = psp_v11_0_bootloader_load_spl,
 	.bootloader_load_sysdrv = psp_v11_0_bootloader_load_sysdrv,
 	.bootloader_load_sos = psp_v11_0_bootloader_load_sos,
-	.ring_init = psp_v11_0_ring_init,
 	.ring_create = psp_v11_0_ring_create,
 	.ring_stop = psp_v11_0_ring_stop,
 	.ring_destroy = psp_v11_0_ring_destroy,
@@ -836,7 +677,8 @@ static const struct psp_funcs psp_v11_0_funcs = {
 	.ring_get_wptr = psp_v11_0_ring_get_wptr,
 	.ring_set_wptr = psp_v11_0_ring_set_wptr,
 	.load_usbc_pd_fw = psp_v11_0_load_usbc_pd_fw,
-	.read_usbc_pd_fw = psp_v11_0_read_usbc_pd_fw
+	.read_usbc_pd_fw = psp_v11_0_read_usbc_pd_fw,
+	.wait_for_bootloader = psp_v11_0_wait_for_bootloader
 };
 
 void psp_v11_0_set_psp_funcs(struct psp_context *psp)
diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v11_0_8.c b/drivers/gpu/drm/amd/amdgpu/psp_v11_0_8.c
index ff13e1beb49b..93787a90d598 100644
--- a/drivers/gpu/drm/amd/amdgpu/psp_v11_0_8.c
+++ b/drivers/gpu/drm/amd/amdgpu/psp_v11_0_8.c
@@ -28,32 +28,6 @@
 
 #include "mp/mp_11_0_8_offset.h"
 
-static int psp_v11_0_8_ring_init(struct psp_context *psp,
-			      enum psp_ring_type ring_type)
-{
-	int ret = 0;
-	struct psp_ring *ring;
-	struct amdgpu_device *adev = psp->adev;
-
-	ring = &psp->km_ring;
-
-	ring->ring_type = ring_type;
-
-	/* allocate 4k Page of Local Frame Buffer memory for ring */
-	ring->ring_size = 0x1000;
-	ret = amdgpu_bo_create_kernel(adev, ring->ring_size, PAGE_SIZE,
-				      AMDGPU_GEM_DOMAIN_VRAM,
-				      &adev->firmware.rbuf,
-				      &ring->ring_mem_mc_addr,
-				      (void **)&ring->ring_mem);
-	if (ret) {
-		ring->ring_size = 0;
-		return ret;
-	}
-
-	return 0;
-}
-
 static int psp_v11_0_8_ring_stop(struct psp_context *psp,
 			       enum psp_ring_type ring_type)
 {
@@ -67,8 +41,9 @@ static int psp_v11_0_8_ring_stop(struct psp_context *psp,
 		/* there might be handshake issue with hardware which needs delay */
 		mdelay(20);
 		/* Wait for response flag (bit 31) */
-		ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_101),
-				   0x80000000, 0x80000000, false);
+		ret = psp_wait_for(
+			psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_101),
+			MBOX_TOS_RESP_FLAG, MBOX_TOS_RESP_MASK, 0);
 	} else {
 		/* Write the ring destroy command*/
 		WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_64,
@@ -76,8 +51,9 @@ static int psp_v11_0_8_ring_stop(struct psp_context *psp,
 		/* there might be handshake issue with hardware which needs delay */
 		mdelay(20);
 		/* Wait for response flag (bit 31) */
-		ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_64),
-				   0x80000000, 0x80000000, false);
+		ret = psp_wait_for(
+			psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_64),
+			MBOX_TOS_RESP_FLAG, MBOX_TOS_RESP_MASK, 0);
 	}
 
 	return ret;
@@ -113,13 +89,15 @@ static int psp_v11_0_8_ring_create(struct psp_context *psp,
 		mdelay(20);
 
 		/* Wait for response flag (bit 31) in C2PMSG_101 */
-		ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_101),
-				   0x80000000, 0x8000FFFF, false);
+		ret = psp_wait_for(
+			psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_101),
+			MBOX_TOS_RESP_FLAG, MBOX_TOS_RESP_MASK, 0);
 
 	} else {
 		/* Wait for sOS ready for ring creation */
-		ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_64),
-				   0x80000000, 0x80000000, false);
+		ret = psp_wait_for(
+			psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_64),
+			MBOX_TOS_READY_FLAG, MBOX_TOS_READY_MASK, 0);
 		if (ret) {
 			DRM_ERROR("Failed to wait for trust OS ready for ring creation\n");
 			return ret;
@@ -143,8 +121,9 @@ static int psp_v11_0_8_ring_create(struct psp_context *psp,
 		mdelay(20);
 
 		/* Wait for response flag (bit 31) in C2PMSG_64 */
-		ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_64),
-				   0x80000000, 0x8000FFFF, false);
+		ret = psp_wait_for(
+			psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_64),
+			MBOX_TOS_RESP_FLAG, MBOX_TOS_RESP_MASK, 0);
 	}
 
 	return ret;
@@ -194,7 +173,6 @@ static void psp_v11_0_8_ring_set_wptr(struct psp_context *psp, uint32_t value)
 }
 
 static const struct psp_funcs psp_v11_0_8_funcs = {
-	.ring_init = psp_v11_0_8_ring_init,
 	.ring_create = psp_v11_0_8_ring_create,
 	.ring_stop = psp_v11_0_8_ring_stop,
 	.ring_destroy = psp_v11_0_8_ring_destroy,
diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v12_0.c b/drivers/gpu/drm/amd/amdgpu/psp_v12_0.c
index a2588200ea58..4c6450d62299 100644
--- a/drivers/gpu/drm/amd/amdgpu/psp_v12_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/psp_v12_0.c
@@ -34,9 +34,6 @@
 #include "sdma0/sdma0_4_0_offset.h"
 #include "nbio/nbio_7_4_offset.h"
 
-#include "oss/osssys_4_0_offset.h"
-#include "oss/osssys_4_0_sh_mask.h"
-
 MODULE_FIRMWARE("amdgpu/renoir_asd.bin");
 MODULE_FIRMWARE("amdgpu/renoir_ta.bin");
 MODULE_FIRMWARE("amdgpu/green_sardine_asd.bin");
@@ -48,73 +45,25 @@ MODULE_FIRMWARE("amdgpu/green_sardine_ta.bin");
 static int psp_v12_0_init_microcode(struct psp_context *psp)
 {
 	struct amdgpu_device *adev = psp->adev;
-	const char *chip_name;
-	char fw_name[30];
+	char ucode_prefix[30];
 	int err = 0;
-	const struct ta_firmware_header_v1_0 *ta_hdr;
 	DRM_DEBUG("\n");
 
-	switch (adev->asic_type) {
-	case CHIP_RENOIR:
-		if (adev->apu_flags & AMD_APU_IS_RENOIR)
-			chip_name = "renoir";
-		else
-			chip_name = "green_sardine";
-		break;
-	default:
-		BUG();
-	}
+	amdgpu_ucode_ip_version_decode(adev, MP0_HWIP, ucode_prefix, sizeof(ucode_prefix));
 
-	err = psp_init_asd_microcode(psp, chip_name);
+	err = psp_init_asd_microcode(psp, ucode_prefix);
 	if (err)
 		return err;
 
-	snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_ta.bin", chip_name);
-	err = request_firmware(&adev->psp.ta_fw, fw_name, adev->dev);
-	if (err) {
-		release_firmware(adev->psp.ta_fw);
-		adev->psp.ta_fw = NULL;
-		dev_info(adev->dev,
-			 "psp v12.0: Failed to load firmware \"%s\"\n",
-			 fw_name);
-	} else {
-		err = amdgpu_ucode_validate(adev->psp.ta_fw);
-		if (err)
-			goto out;
-
-		ta_hdr = (const struct ta_firmware_header_v1_0 *)
-				 adev->psp.ta_fw->data;
-		adev->psp.hdcp_context.context.bin_desc.fw_version =
-			le32_to_cpu(ta_hdr->hdcp.fw_version);
-		adev->psp.hdcp_context.context.bin_desc.size_bytes =
-			le32_to_cpu(ta_hdr->hdcp.size_bytes);
-		adev->psp.hdcp_context.context.bin_desc.start_addr =
-			(uint8_t *)ta_hdr +
-			le32_to_cpu(ta_hdr->header.ucode_array_offset_bytes);
-
-		adev->psp.ta_fw_version = le32_to_cpu(ta_hdr->header.ucode_version);
-
-		adev->psp.dtm_context.context.bin_desc.fw_version =
-			le32_to_cpu(ta_hdr->dtm.fw_version);
-		adev->psp.dtm_context.context.bin_desc.size_bytes =
-			le32_to_cpu(ta_hdr->dtm.size_bytes);
-		adev->psp.dtm_context.context.bin_desc.start_addr =
-			(uint8_t *)adev->psp.hdcp_context.context.bin_desc.start_addr +
-			le32_to_cpu(ta_hdr->dtm.offset_bytes);
-	}
-
-	return 0;
+	err = psp_init_ta_microcode(psp, ucode_prefix);
+	if (err)
+		return err;
 
-out:
-	release_firmware(adev->psp.ta_fw);
-	adev->psp.ta_fw = NULL;
-	if (err) {
-		dev_err(adev->dev,
-			"psp v12.0: Failed to load firmware \"%s\"\n",
-			fw_name);
-	}
+	/* only supported on renoir */
+	if (!(adev->apu_flags & AMD_APU_IS_RENOIR))
+		adev->psp.securedisplay_context.context.bin_desc.size_bytes = 0;
 
-	return err;
+	return 0;
 }
 
 static int psp_v12_0_bootloader_load_sysdrv(struct psp_context *psp)
@@ -133,7 +82,7 @@ static int psp_v12_0_bootloader_load_sysdrv(struct psp_context *psp)
 
 	/* Wait for bootloader to signify that is ready having bit 31 of C2PMSG_35 set to 1 */
 	ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_35),
-			   0x80000000, 0x80000000, false);
+			   0x80000000, 0x80000000, 0);
 	if (ret)
 		return ret;
 
@@ -147,11 +96,8 @@ static int psp_v12_0_bootloader_load_sysdrv(struct psp_context *psp)
 	WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_35,
 	       psp_gfxdrv_command_reg);
 
-	/* there might be handshake issue with hardware which needs delay */
-	mdelay(20);
-
 	ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_35),
-			   0x80000000, 0x80000000, false);
+			   0x80000000, 0x80000000, 0);
 
 	return ret;
 }
@@ -172,7 +118,7 @@ static int psp_v12_0_bootloader_load_sos(struct psp_context *psp)
 
 	/* Wait for bootloader to signify that is ready having bit 31 of C2PMSG_35 set to 1 */
 	ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_35),
-			   0x80000000, 0x80000000, false);
+			   0x80000000, 0x80000000, 0);
 	if (ret)
 		return ret;
 
@@ -186,74 +132,13 @@ static int psp_v12_0_bootloader_load_sos(struct psp_context *psp)
 	WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_35,
 	       psp_gfxdrv_command_reg);
 
-	/* there might be handshake issue with hardware which needs delay */
-	mdelay(20);
 	ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_81),
-			   RREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_81),
-			   0, true);
+			   RREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_81), 0,
+			   PSP_WAITREG_CHANGED);
 
 	return ret;
 }
 
-static void psp_v12_0_reroute_ih(struct psp_context *psp)
-{
-	struct amdgpu_device *adev = psp->adev;
-	uint32_t tmp;
-
-	/* Change IH ring for VMC */
-	tmp = REG_SET_FIELD(0, IH_CLIENT_CFG_DATA, CREDIT_RETURN_ADDR, 0x1244b);
-	tmp = REG_SET_FIELD(tmp, IH_CLIENT_CFG_DATA, CLIENT_TYPE, 1);
-	tmp = REG_SET_FIELD(tmp, IH_CLIENT_CFG_DATA, RING_ID, 1);
-
-	WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_69, 3);
-	WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_70, tmp);
-	WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_64, GFX_CTRL_CMD_ID_GBR_IH_SET);
-
-	mdelay(20);
-	psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_64),
-		     0x80000000, 0x8000FFFF, false);
-
-	/* Change IH ring for UMC */
-	tmp = REG_SET_FIELD(0, IH_CLIENT_CFG_DATA, CREDIT_RETURN_ADDR, 0x1216b);
-	tmp = REG_SET_FIELD(tmp, IH_CLIENT_CFG_DATA, RING_ID, 1);
-
-	WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_69, 4);
-	WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_70, tmp);
-	WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_64, GFX_CTRL_CMD_ID_GBR_IH_SET);
-
-	mdelay(20);
-	psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_64),
-		     0x80000000, 0x8000FFFF, false);
-}
-
-static int psp_v12_0_ring_init(struct psp_context *psp,
-			      enum psp_ring_type ring_type)
-{
-	int ret = 0;
-	struct psp_ring *ring;
-	struct amdgpu_device *adev = psp->adev;
-
-	psp_v12_0_reroute_ih(psp);
-
-	ring = &psp->km_ring;
-
-	ring->ring_type = ring_type;
-
-	/* allocate 4k Page of Local Frame Buffer memory for ring */
-	ring->ring_size = 0x1000;
-	ret = amdgpu_bo_create_kernel(adev, ring->ring_size, PAGE_SIZE,
-				      AMDGPU_GEM_DOMAIN_VRAM,
-				      &adev->firmware.rbuf,
-				      &ring->ring_mem_mc_addr,
-				      (void **)&ring->ring_mem);
-	if (ret) {
-		ring->ring_size = 0;
-		return ret;
-	}
-
-	return 0;
-}
-
 static int psp_v12_0_ring_create(struct psp_context *psp,
 				enum psp_ring_type ring_type)
 {
@@ -262,47 +147,23 @@ static int psp_v12_0_ring_create(struct psp_context *psp,
 	struct psp_ring *ring = &psp->km_ring;
 	struct amdgpu_device *adev = psp->adev;
 
-	if (amdgpu_sriov_vf(psp->adev)) {
-		/* Write low address of the ring to C2PMSG_102 */
-		psp_ring_reg = lower_32_bits(ring->ring_mem_mc_addr);
-		WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_102, psp_ring_reg);
-		/* Write high address of the ring to C2PMSG_103 */
-		psp_ring_reg = upper_32_bits(ring->ring_mem_mc_addr);
-		WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_103, psp_ring_reg);
-
-		/* Write the ring initialization command to C2PMSG_101 */
-		WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_101,
-					     GFX_CTRL_CMD_ID_INIT_GPCOM_RING);
-
-		/* there might be handshake issue with hardware which needs delay */
-		mdelay(20);
-
-		/* Wait for response flag (bit 31) in C2PMSG_101 */
-		ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_101),
-				   0x80000000, 0x8000FFFF, false);
-
-	} else {
-		/* Write low address of the ring to C2PMSG_69 */
-		psp_ring_reg = lower_32_bits(ring->ring_mem_mc_addr);
-		WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_69, psp_ring_reg);
-		/* Write high address of the ring to C2PMSG_70 */
-		psp_ring_reg = upper_32_bits(ring->ring_mem_mc_addr);
-		WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_70, psp_ring_reg);
-		/* Write size of ring to C2PMSG_71 */
-		psp_ring_reg = ring->ring_size;
-		WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_71, psp_ring_reg);
-		/* Write the ring initialization command to C2PMSG_64 */
-		psp_ring_reg = ring_type;
-		psp_ring_reg = psp_ring_reg << 16;
-		WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_64, psp_ring_reg);
-
-		/* there might be handshake issue with hardware which needs delay */
-		mdelay(20);
-
-		/* Wait for response flag (bit 31) in C2PMSG_64 */
-		ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_64),
-				   0x80000000, 0x8000FFFF, false);
-	}
+	/* Write low address of the ring to C2PMSG_69 */
+	psp_ring_reg = lower_32_bits(ring->ring_mem_mc_addr);
+	WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_69, psp_ring_reg);
+	/* Write high address of the ring to C2PMSG_70 */
+	psp_ring_reg = upper_32_bits(ring->ring_mem_mc_addr);
+	WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_70, psp_ring_reg);
+	/* Write size of ring to C2PMSG_71 */
+	psp_ring_reg = ring->ring_size;
+	WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_71, psp_ring_reg);
+	/* Write the ring initialization command to C2PMSG_64 */
+	psp_ring_reg = ring_type;
+	psp_ring_reg = psp_ring_reg << 16;
+	WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_64, psp_ring_reg);
+
+	/* Wait for response flag (bit 31) in C2PMSG_64 */
+	ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_64),
+			   MBOX_TOS_RESP_FLAG, MBOX_TOS_RESP_MASK, 0);
 
 	return ret;
 }
@@ -321,16 +182,15 @@ static int psp_v12_0_ring_stop(struct psp_context *psp,
 		WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_64,
 				     GFX_CTRL_CMD_ID_DESTROY_RINGS);
 
-	/* there might be handshake issue with hardware which needs delay */
-	mdelay(20);
-
 	/* Wait for response flag (bit 31) */
 	if (amdgpu_sriov_vf(adev))
-		ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_101),
-				   0x80000000, 0x80000000, false);
+		ret = psp_wait_for(
+			psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_101),
+			MBOX_TOS_RESP_FLAG, MBOX_TOS_RESP_MASK, 0);
 	else
-		ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_64),
-				   0x80000000, 0x80000000, false);
+		ret = psp_wait_for(
+			psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_64),
+			MBOX_TOS_RESP_FLAG, MBOX_TOS_RESP_MASK, 0);
 
 	return ret;
 }
@@ -361,7 +221,8 @@ static int psp_v12_0_mode1_reset(struct psp_context *psp)
 
 	offset = SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_64);
 
-	ret = psp_wait_for(psp, offset, 0x80000000, 0x8000FFFF, false);
+	ret = psp_wait_for(psp, offset, MBOX_TOS_READY_FLAG,
+			   MBOX_TOS_READY_MASK, 0);
 
 	if (ret) {
 		DRM_INFO("psp is not working correctly before mode1 reset!\n");
@@ -375,7 +236,8 @@ static int psp_v12_0_mode1_reset(struct psp_context *psp)
 
 	offset = SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_33);
 
-	ret = psp_wait_for(psp, offset, 0x80000000, 0x80000000, false);
+	ret = psp_wait_for(psp, offset, MBOX_TOS_RESP_FLAG, MBOX_TOS_RESP_MASK,
+			   0);
 
 	if (ret) {
 		DRM_INFO("psp mode 1 reset failed!\n");
@@ -415,7 +277,6 @@ static const struct psp_funcs psp_v12_0_funcs = {
 	.init_microcode = psp_v12_0_init_microcode,
 	.bootloader_load_sysdrv = psp_v12_0_bootloader_load_sysdrv,
 	.bootloader_load_sos = psp_v12_0_bootloader_load_sos,
-	.ring_init = psp_v12_0_ring_init,
 	.ring_create = psp_v12_0_ring_create,
 	.ring_stop = psp_v12_0_ring_stop,
 	.ring_destroy = psp_v12_0_ring_destroy,
diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c b/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c
index 17655bc6d2f1..af4a7d7c4abd 100644
--- a/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c
@@ -20,20 +20,49 @@
  * OTHER DEALINGS IN THE SOFTWARE.
  *
  */
+#include <drm/drm_drv.h>
+#include <linux/vmalloc.h>
 #include "amdgpu.h"
 #include "amdgpu_psp.h"
 #include "amdgpu_ucode.h"
 #include "soc15_common.h"
 #include "psp_v13_0.h"
+#include "amdgpu_ras.h"
 
 #include "mp/mp_13_0_2_offset.h"
 #include "mp/mp_13_0_2_sh_mask.h"
 
 MODULE_FIRMWARE("amdgpu/aldebaran_sos.bin");
 MODULE_FIRMWARE("amdgpu/aldebaran_ta.bin");
-MODULE_FIRMWARE("amdgpu/yellow_carp_asd.bin");
+MODULE_FIRMWARE("amdgpu/aldebaran_cap.bin");
 MODULE_FIRMWARE("amdgpu/yellow_carp_toc.bin");
 MODULE_FIRMWARE("amdgpu/yellow_carp_ta.bin");
+MODULE_FIRMWARE("amdgpu/psp_13_0_5_toc.bin");
+MODULE_FIRMWARE("amdgpu/psp_13_0_5_ta.bin");
+MODULE_FIRMWARE("amdgpu/psp_13_0_8_toc.bin");
+MODULE_FIRMWARE("amdgpu/psp_13_0_8_ta.bin");
+MODULE_FIRMWARE("amdgpu/psp_13_0_0_sos.bin");
+MODULE_FIRMWARE("amdgpu/psp_13_0_0_sos_kicker.bin");
+MODULE_FIRMWARE("amdgpu/psp_13_0_0_ta.bin");
+MODULE_FIRMWARE("amdgpu/psp_13_0_0_ta_kicker.bin");
+MODULE_FIRMWARE("amdgpu/psp_13_0_7_sos.bin");
+MODULE_FIRMWARE("amdgpu/psp_13_0_7_ta.bin");
+MODULE_FIRMWARE("amdgpu/psp_13_0_10_sos.bin");
+MODULE_FIRMWARE("amdgpu/psp_13_0_10_ta.bin");
+MODULE_FIRMWARE("amdgpu/psp_13_0_11_toc.bin");
+MODULE_FIRMWARE("amdgpu/psp_13_0_11_ta.bin");
+MODULE_FIRMWARE("amdgpu/psp_13_0_6_sos.bin");
+MODULE_FIRMWARE("amdgpu/psp_13_0_6_ta.bin");
+MODULE_FIRMWARE("amdgpu/psp_13_0_12_sos.bin");
+MODULE_FIRMWARE("amdgpu/psp_13_0_12_ta.bin");
+MODULE_FIRMWARE("amdgpu/psp_13_0_14_sos.bin");
+MODULE_FIRMWARE("amdgpu/psp_13_0_14_ta.bin");
+MODULE_FIRMWARE("amdgpu/psp_14_0_0_toc.bin");
+MODULE_FIRMWARE("amdgpu/psp_14_0_0_ta.bin");
+MODULE_FIRMWARE("amdgpu/psp_14_0_1_toc.bin");
+MODULE_FIRMWARE("amdgpu/psp_14_0_1_ta.bin");
+MODULE_FIRMWARE("amdgpu/psp_14_0_4_toc.bin");
+MODULE_FIRMWARE("amdgpu/psp_14_0_4_ta.bin");
 
 /* For large FW files the time to complete can be very long */
 #define USBC_PD_POLLING_LIMIT_S 240
@@ -41,41 +70,62 @@ MODULE_FIRMWARE("amdgpu/yellow_carp_ta.bin");
 /* Read USB-PD from LFB */
 #define GFX_CMD_USB_PD_USE_LFB 0x480
 
+/* Retry times for vmbx ready wait */
+#define PSP_VMBX_POLLING_LIMIT 3000
+
+/* memory training timeout define */
+#define MEM_TRAIN_SEND_MSG_TIMEOUT_US	3000000
+
+#define regMP1_PUB_SCRATCH0	0x3b10090
+
+#define PSP13_BL_STATUS_SIZE 100
+
 static int psp_v13_0_init_microcode(struct psp_context *psp)
 {
 	struct amdgpu_device *adev = psp->adev;
-	const char *chip_name;
+	char ucode_prefix[30];
 	int err = 0;
 
-	switch (adev->ip_versions[MP0_HWIP][0]) {
+	amdgpu_ucode_ip_version_decode(adev, MP0_HWIP, ucode_prefix, sizeof(ucode_prefix));
+
+	switch (amdgpu_ip_version(adev, MP0_HWIP, 0)) {
 	case IP_VERSION(13, 0, 2):
-		chip_name = "aldebaran";
+		err = psp_init_sos_microcode(psp, ucode_prefix);
+		if (err)
+			return err;
+		/* It's not necessary to load ras ta on Guest side */
+		if (!amdgpu_sriov_vf(adev)) {
+			err = psp_init_ta_microcode(psp, ucode_prefix);
+			if (err)
+				return err;
+		}
 		break;
 	case IP_VERSION(13, 0, 1):
 	case IP_VERSION(13, 0, 3):
-		chip_name = "yellow_carp";
-		break;
-	default:
-		BUG();
-	}
-	switch (adev->ip_versions[MP0_HWIP][0]) {
-	case IP_VERSION(13, 0, 2):
-		err = psp_init_sos_microcode(psp, chip_name);
+	case IP_VERSION(13, 0, 5):
+	case IP_VERSION(13, 0, 8):
+	case IP_VERSION(13, 0, 11):
+	case IP_VERSION(14, 0, 0):
+	case IP_VERSION(14, 0, 1):
+	case IP_VERSION(14, 0, 4):
+		err = psp_init_toc_microcode(psp, ucode_prefix);
 		if (err)
 			return err;
-		err = psp_init_ta_microcode(&adev->psp, chip_name);
+		err = psp_init_ta_microcode(psp, ucode_prefix);
 		if (err)
 			return err;
 		break;
-	case IP_VERSION(13, 0, 1):
-	case IP_VERSION(13, 0, 3):
-		err = psp_init_asd_microcode(psp, chip_name);
+	case IP_VERSION(13, 0, 0):
+	case IP_VERSION(13, 0, 6):
+	case IP_VERSION(13, 0, 7):
+	case IP_VERSION(13, 0, 10):
+	case IP_VERSION(13, 0, 12):
+	case IP_VERSION(13, 0, 14):
+		err = psp_init_sos_microcode(psp, ucode_prefix);
 		if (err)
 			return err;
-		err = psp_init_toc_microcode(psp, chip_name);
-		if (err)
-			return err;
-		err = psp_init_ta_microcode(psp, chip_name);
+		/* It's not necessary to load ras ta on Guest side */
+		err = psp_init_ta_microcode(psp, ucode_prefix);
 		if (err)
 			return err;
 		break;
@@ -96,29 +146,107 @@ static bool psp_v13_0_is_sos_alive(struct psp_context *psp)
 	return sol_reg != 0x0;
 }
 
-static int psp_v13_0_wait_for_bootloader(struct psp_context *psp)
+static void psp_v13_0_bootloader_print_status(struct psp_context *psp,
+					      const char *msg)
 {
 	struct amdgpu_device *adev = psp->adev;
+	u32 bl_status_reg;
+	char bl_status_msg[PSP13_BL_STATUS_SIZE];
+	int i, at;
+
+	if (amdgpu_ip_version(adev, MP0_HWIP, 0) == IP_VERSION(13, 0, 6) ||
+	    amdgpu_ip_version(adev, MP0_HWIP, 0) == IP_VERSION(13, 0, 12) ||
+	    amdgpu_ip_version(adev, MP0_HWIP, 0) == IP_VERSION(13, 0, 14)) {
+		at = 0;
+		for_each_inst(i, adev->aid_mask) {
+			bl_status_reg =
+				(SOC15_REG_OFFSET(MP0, 0, regMP0_SMN_C2PMSG_92)
+				 << 2) +
+				adev->asic_funcs->encode_ext_smn_addressing(i);
+			at += snprintf(bl_status_msg + at,
+				       PSP13_BL_STATUS_SIZE - at,
+				       " status(%02i): 0x%08x", i,
+				       RREG32_PCIE_EXT(bl_status_reg));
+		}
+		dev_info(adev->dev, "%s - %s", msg, bl_status_msg);
+	}
+}
 
-	int ret;
-	int retry_loop;
+static int psp_v13_0_wait_for_vmbx_ready(struct psp_context *psp)
+{
+	struct amdgpu_device *adev = psp->adev;
+	int retry_loop, ret;
 
-	for (retry_loop = 0; retry_loop < 10; retry_loop++) {
+	for (retry_loop = 0; retry_loop < PSP_VMBX_POLLING_LIMIT; retry_loop++) {
 		/* Wait for bootloader to signify that is
-		    ready having bit 31 of C2PMSG_35 set to 1 */
-		ret = psp_wait_for(psp,
-				   SOC15_REG_OFFSET(MP0, 0, regMP0_SMN_C2PMSG_35),
-				   0x80000000,
-				   0x80000000,
-				   false);
+		   ready having bit 31 of C2PMSG_33 set to 1 */
+		ret = psp_wait_for(
+			psp, SOC15_REG_OFFSET(MP0, 0, regMP0_SMN_C2PMSG_33),
+			0x80000000, 0xffffffff, PSP_WAITREG_NOVERBOSE);
+
+		if (ret == 0)
+			break;
+	}
+
+	if (ret)
+		dev_warn(adev->dev, "Bootloader wait timed out");
+
+	return ret;
+}
+
+static int psp_v13_0_wait_for_bootloader(struct psp_context *psp)
+{
+	struct amdgpu_device *adev = psp->adev;
+	int retry_loop, retry_cnt, ret;
+
+	retry_cnt =
+		((amdgpu_ip_version(adev, MP0_HWIP, 0) == IP_VERSION(13, 0, 6) ||
+		  amdgpu_ip_version(adev, MP0_HWIP, 0) == IP_VERSION(13, 0, 12) ||
+		  amdgpu_ip_version(adev, MP0_HWIP, 0) == IP_VERSION(13, 0, 14))) ?
+			PSP_VMBX_POLLING_LIMIT :
+			10;
+	/* Wait for bootloader to signify that it is ready having bit 31 of
+	 * C2PMSG_35 set to 1. All other bits are expected to be cleared.
+	 * If there is an error in processing command, bits[7:0] will be set.
+	 * This is applicable for PSP v13.0.6 and newer.
+	 */
+	for (retry_loop = 0; retry_loop < retry_cnt; retry_loop++) {
+		ret = psp_wait_for(
+			psp, SOC15_REG_OFFSET(MP0, 0, regMP0_SMN_C2PMSG_35),
+			0x80000000, 0xffffffff, PSP_WAITREG_NOVERBOSE);
 
 		if (ret == 0)
 			return 0;
+		if (retry_loop && !(retry_loop % 10))
+			psp_v13_0_bootloader_print_status(
+				psp, "Waiting for bootloader completion");
 	}
 
 	return ret;
 }
 
+static int psp_v13_0_wait_for_bootloader_steady_state(struct psp_context *psp)
+{
+	struct amdgpu_device *adev = psp->adev;
+	int ret;
+
+	if (amdgpu_ip_version(adev, MP0_HWIP, 0) == IP_VERSION(13, 0, 6) ||
+	    amdgpu_ip_version(adev, MP0_HWIP, 0) == IP_VERSION(13, 0, 12) ||
+	    amdgpu_ip_version(adev, MP0_HWIP, 0) == IP_VERSION(13, 0, 14)) {
+		ret = psp_v13_0_wait_for_vmbx_ready(psp);
+		if (ret)
+			amdgpu_ras_query_boot_status(adev, 4);
+
+		ret = psp_v13_0_wait_for_bootloader(psp);
+		if (ret)
+			amdgpu_ras_query_boot_status(adev, 4);
+
+		return ret;
+	}
+
+	return 0;
+}
+
 static int psp_v13_0_bootloader_load_component(struct psp_context  	*psp,
 					       struct psp_bin_desc 	*bin_desc,
 					       enum psp_bootloader_cmd  bl_cmd)
@@ -159,6 +287,11 @@ static int psp_v13_0_bootloader_load_kdb(struct psp_context *psp)
 	return psp_v13_0_bootloader_load_component(psp, &psp->kdb, PSP_BL__LOAD_KEY_DATABASE);
 }
 
+static int psp_v13_0_bootloader_load_spl(struct psp_context *psp)
+{
+	return psp_v13_0_bootloader_load_component(psp, &psp->kdb, PSP_BL__LOAD_TOS_SPL_TABLE);
+}
+
 static int psp_v13_0_bootloader_load_sysdrv(struct psp_context *psp)
 {
 	return psp_v13_0_bootloader_load_component(psp, &psp->sys, PSP_BL__LOAD_SYSDRV);
@@ -179,6 +312,23 @@ static int psp_v13_0_bootloader_load_dbg_drv(struct psp_context *psp)
 	return psp_v13_0_bootloader_load_component(psp, &psp->dbg_drv, PSP_BL__LOAD_DBGDRV);
 }
 
+static int psp_v13_0_bootloader_load_ras_drv(struct psp_context *psp)
+{
+	return psp_v13_0_bootloader_load_component(psp, &psp->ras_drv, PSP_BL__LOAD_RASDRV);
+}
+
+static int psp_v13_0_bootloader_load_spdm_drv(struct psp_context *psp)
+{
+	return psp_v13_0_bootloader_load_component(psp, &psp->spdm_drv, PSP_BL__LOAD_SPDMDRV);
+}
+
+static inline void psp_v13_0_init_sos_version(struct psp_context *psp)
+{
+	struct amdgpu_device *adev = psp->adev;
+
+	psp->sos.fw_version = RREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_58);
+}
+
 static int psp_v13_0_bootloader_load_sos(struct psp_context *psp)
 {
 	int ret;
@@ -188,8 +338,10 @@ static int psp_v13_0_bootloader_load_sos(struct psp_context *psp)
 	/* Check sOS sign of life register to confirm sys driver and sOS
 	 * are already been loaded.
 	 */
-	if (psp_v13_0_is_sos_alive(psp))
+	if (psp_v13_0_is_sos_alive(psp)) {
+		psp_v13_0_init_sos_version(psp);
 		return 0;
+	}
 
 	ret = psp_v13_0_wait_for_bootloader(psp);
 	if (ret)
@@ -210,36 +362,13 @@ static int psp_v13_0_bootloader_load_sos(struct psp_context *psp)
 	/* there might be handshake issue with hardware which needs delay */
 	mdelay(20);
 	ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, regMP0_SMN_C2PMSG_81),
-			   RREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_81),
-			   0, true);
-
-	return ret;
-}
+			   RREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_81), 0,
+			   PSP_WAITREG_CHANGED);
 
-static int psp_v13_0_ring_init(struct psp_context *psp,
-			      enum psp_ring_type ring_type)
-{
-	int ret = 0;
-	struct psp_ring *ring;
-	struct amdgpu_device *adev = psp->adev;
-
-	ring = &psp->km_ring;
-
-	ring->ring_type = ring_type;
-
-	/* allocate 4k Page of Local Frame Buffer memory for ring */
-	ring->ring_size = 0x1000;
-	ret = amdgpu_bo_create_kernel(adev, ring->ring_size, PAGE_SIZE,
-				      AMDGPU_GEM_DOMAIN_VRAM,
-				      &adev->firmware.rbuf,
-				      &ring->ring_mem_mc_addr,
-				      (void **)&ring->ring_mem);
-	if (ret) {
-		ring->ring_size = 0;
-		return ret;
-	}
+	if (!ret)
+		psp_v13_0_init_sos_version(psp);
 
-	return 0;
+	return ret;
 }
 
 static int psp_v13_0_ring_stop(struct psp_context *psp,
@@ -255,8 +384,9 @@ static int psp_v13_0_ring_stop(struct psp_context *psp,
 		/* there might be handshake issue with hardware which needs delay */
 		mdelay(20);
 		/* Wait for response flag (bit 31) */
-		ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, regMP0_SMN_C2PMSG_101),
-				   0x80000000, 0x80000000, false);
+		ret = psp_wait_for(
+			psp, SOC15_REG_OFFSET(MP0, 0, regMP0_SMN_C2PMSG_101),
+			MBOX_TOS_RESP_FLAG, MBOX_TOS_RESP_MASK, 0);
 	} else {
 		/* Write the ring destroy command*/
 		WREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_64,
@@ -264,8 +394,9 @@ static int psp_v13_0_ring_stop(struct psp_context *psp,
 		/* there might be handshake issue with hardware which needs delay */
 		mdelay(20);
 		/* Wait for response flag (bit 31) */
-		ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, regMP0_SMN_C2PMSG_64),
-				   0x80000000, 0x80000000, false);
+		ret = psp_wait_for(
+			psp, SOC15_REG_OFFSET(MP0, 0, regMP0_SMN_C2PMSG_64),
+			MBOX_TOS_RESP_FLAG, MBOX_TOS_RESP_MASK, 0);
 	}
 
 	return ret;
@@ -301,13 +432,15 @@ static int psp_v13_0_ring_create(struct psp_context *psp,
 		mdelay(20);
 
 		/* Wait for response flag (bit 31) in C2PMSG_101 */
-		ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, regMP0_SMN_C2PMSG_101),
-				   0x80000000, 0x8000FFFF, false);
+		ret = psp_wait_for(
+			psp, SOC15_REG_OFFSET(MP0, 0, regMP0_SMN_C2PMSG_101),
+			MBOX_TOS_RESP_FLAG, MBOX_TOS_RESP_MASK, 0);
 
 	} else {
 		/* Wait for sOS ready for ring creation */
-		ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, regMP0_SMN_C2PMSG_64),
-				   0x80000000, 0x80000000, false);
+		ret = psp_wait_for(
+			psp, SOC15_REG_OFFSET(MP0, 0, regMP0_SMN_C2PMSG_64),
+			MBOX_TOS_READY_FLAG, MBOX_TOS_READY_MASK, 0);
 		if (ret) {
 			DRM_ERROR("Failed to wait for trust OS ready for ring creation\n");
 			return ret;
@@ -331,8 +464,9 @@ static int psp_v13_0_ring_create(struct psp_context *psp,
 		mdelay(20);
 
 		/* Wait for response flag (bit 31) in C2PMSG_64 */
-		ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, regMP0_SMN_C2PMSG_64),
-				   0x80000000, 0x8000FFFF, false);
+		ret = psp_wait_for(
+			psp, SOC15_REG_OFFSET(MP0, 0, regMP0_SMN_C2PMSG_64),
+			MBOX_TOS_RESP_FLAG, MBOX_TOS_RESP_MASK, 0);
 	}
 
 	return ret;
@@ -381,6 +515,160 @@ static void psp_v13_0_ring_set_wptr(struct psp_context *psp, uint32_t value)
 		WREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_67, value);
 }
 
+static int psp_v13_0_memory_training_send_msg(struct psp_context *psp, int msg)
+{
+	int ret;
+	int i;
+	uint32_t data_32;
+	int max_wait;
+	struct amdgpu_device *adev = psp->adev;
+
+	data_32 = (psp->mem_train_ctx.c2p_train_data_offset >> 20);
+	WREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_36, data_32);
+	WREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_35, msg);
+
+	max_wait = MEM_TRAIN_SEND_MSG_TIMEOUT_US / adev->usec_timeout;
+	for (i = 0; i < max_wait; i++) {
+		ret = psp_wait_for(
+			psp, SOC15_REG_OFFSET(MP0, 0, regMP0_SMN_C2PMSG_35),
+			0x80000000, 0x80000000, PSP_WAITREG_NOVERBOSE);
+		if (ret == 0)
+			break;
+	}
+	if (i < max_wait)
+		ret = 0;
+	else
+		ret = -ETIME;
+
+	dev_dbg(adev->dev, "training %s %s, cost %d @ %d ms\n",
+		  (msg == PSP_BL__DRAM_SHORT_TRAIN) ? "short" : "long",
+		  (ret == 0) ? "succeed" : "failed",
+		  i, adev->usec_timeout/1000);
+	return ret;
+}
+
+
+static int psp_v13_0_memory_training(struct psp_context *psp, uint32_t ops)
+{
+	struct psp_memory_training_context *ctx = &psp->mem_train_ctx;
+	uint32_t *pcache = (uint32_t *)ctx->sys_cache;
+	struct amdgpu_device *adev = psp->adev;
+	uint32_t p2c_header[4];
+	uint32_t sz;
+	void *buf;
+	int ret, idx;
+
+	if (ctx->init == PSP_MEM_TRAIN_NOT_SUPPORT) {
+		dev_dbg(adev->dev, "Memory training is not supported.\n");
+		return 0;
+	} else if (ctx->init != PSP_MEM_TRAIN_INIT_SUCCESS) {
+		dev_err(adev->dev, "Memory training initialization failure.\n");
+		return -EINVAL;
+	}
+
+	if (psp_v13_0_is_sos_alive(psp)) {
+		dev_dbg(adev->dev, "SOS is alive, skip memory training.\n");
+		return 0;
+	}
+
+	amdgpu_device_vram_access(adev, ctx->p2c_train_data_offset, p2c_header, sizeof(p2c_header), false);
+	dev_dbg(adev->dev, "sys_cache[%08x,%08x,%08x,%08x] p2c_header[%08x,%08x,%08x,%08x]\n",
+		  pcache[0], pcache[1], pcache[2], pcache[3],
+		  p2c_header[0], p2c_header[1], p2c_header[2], p2c_header[3]);
+
+	if (ops & PSP_MEM_TRAIN_SEND_SHORT_MSG) {
+		dev_dbg(adev->dev, "Short training depends on restore.\n");
+		ops |= PSP_MEM_TRAIN_RESTORE;
+	}
+
+	if ((ops & PSP_MEM_TRAIN_RESTORE) &&
+	    pcache[0] != MEM_TRAIN_SYSTEM_SIGNATURE) {
+		dev_dbg(adev->dev, "sys_cache[0] is invalid, restore depends on save.\n");
+		ops |= PSP_MEM_TRAIN_SAVE;
+	}
+
+	if (p2c_header[0] == MEM_TRAIN_SYSTEM_SIGNATURE &&
+	    !(pcache[0] == MEM_TRAIN_SYSTEM_SIGNATURE &&
+	      pcache[3] == p2c_header[3])) {
+		dev_dbg(adev->dev, "sys_cache is invalid or out-of-date, need save training data to sys_cache.\n");
+		ops |= PSP_MEM_TRAIN_SAVE;
+	}
+
+	if ((ops & PSP_MEM_TRAIN_SAVE) &&
+	    p2c_header[0] != MEM_TRAIN_SYSTEM_SIGNATURE) {
+		dev_dbg(adev->dev, "p2c_header[0] is invalid, save depends on long training.\n");
+		ops |= PSP_MEM_TRAIN_SEND_LONG_MSG;
+	}
+
+	if (ops & PSP_MEM_TRAIN_SEND_LONG_MSG) {
+		ops &= ~PSP_MEM_TRAIN_SEND_SHORT_MSG;
+		ops |= PSP_MEM_TRAIN_SAVE;
+	}
+
+	dev_dbg(adev->dev, "Memory training ops:%x.\n", ops);
+
+	if (ops & PSP_MEM_TRAIN_SEND_LONG_MSG) {
+		/*
+		 * Long training will encroach a certain amount on the bottom of VRAM;
+		 * save the content from the bottom of VRAM to system memory
+		 * before training, and restore it after training to avoid
+		 * VRAM corruption.
+		 */
+		sz = BIST_MEM_TRAINING_ENCROACHED_SIZE;
+
+		if (adev->gmc.visible_vram_size < sz || !adev->mman.aper_base_kaddr) {
+			dev_err(adev->dev, "visible_vram_size %llx or aper_base_kaddr %p is not initialized.\n",
+				  adev->gmc.visible_vram_size,
+				  adev->mman.aper_base_kaddr);
+			return -EINVAL;
+		}
+
+		buf = vmalloc(sz);
+		if (!buf) {
+			dev_err(adev->dev, "failed to allocate system memory.\n");
+			return -ENOMEM;
+		}
+
+		if (drm_dev_enter(adev_to_drm(adev), &idx)) {
+			memcpy_fromio(buf, adev->mman.aper_base_kaddr, sz);
+			ret = psp_v13_0_memory_training_send_msg(psp, PSP_BL__DRAM_LONG_TRAIN);
+			if (ret) {
+				DRM_ERROR("Send long training msg failed.\n");
+				vfree(buf);
+				drm_dev_exit(idx);
+				return ret;
+			}
+
+			memcpy_toio(adev->mman.aper_base_kaddr, buf, sz);
+			amdgpu_device_flush_hdp(adev, NULL);
+			vfree(buf);
+			drm_dev_exit(idx);
+		} else {
+			vfree(buf);
+			return -ENODEV;
+		}
+	}
+
+	if (ops & PSP_MEM_TRAIN_SAVE) {
+		amdgpu_device_vram_access(psp->adev, ctx->p2c_train_data_offset, ctx->sys_cache, ctx->train_data_size, false);
+	}
+
+	if (ops & PSP_MEM_TRAIN_RESTORE) {
+		amdgpu_device_vram_access(psp->adev, ctx->c2p_train_data_offset, ctx->sys_cache, ctx->train_data_size, true);
+	}
+
+	if (ops & PSP_MEM_TRAIN_SEND_SHORT_MSG) {
+		ret = psp_v13_0_memory_training_send_msg(psp, (amdgpu_force_long_training > 0) ?
+							 PSP_BL__DRAM_LONG_TRAIN : PSP_BL__DRAM_SHORT_TRAIN);
+		if (ret) {
+			dev_err(adev->dev, "send training msg failed.\n");
+			return ret;
+		}
+	}
+	ctx->training_cnt++;
+	return 0;
+}
+
 static int psp_v13_0_load_usbc_pd_fw(struct psp_context *psp, uint64_t fw_pri_mc_addr)
 {
 	struct amdgpu_device *adev = psp->adev;
@@ -395,7 +683,7 @@ static int psp_v13_0_load_usbc_pd_fw(struct psp_context *psp, uint64_t fw_pri_mc
 	WREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_36, (fw_pri_mc_addr >> 20));
 
 	ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, regMP0_SMN_C2PMSG_35),
-			     0x80000000, 0x80000000, false);
+			   0x80000000, 0x80000000, 0);
 	if (ret)
 		return ret;
 
@@ -432,29 +720,255 @@ static int psp_v13_0_read_usbc_pd_fw(struct psp_context *psp, uint32_t *fw_ver)
 	WREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_35, C2PMSG_CMD_GFX_USB_PD_FW_VER);
 
 	ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, regMP0_SMN_C2PMSG_35),
-				     0x80000000, 0x80000000, false);
+			   0x80000000, 0x80000000, 0);
 	if (!ret)
 		*fw_ver = RREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_36);
 
 	return ret;
 }
 
+static int psp_v13_0_exec_spi_cmd(struct psp_context *psp, int cmd)
+{
+	uint32_t reg_status = 0, reg_val = 0;
+	struct amdgpu_device *adev = psp->adev;
+	int ret;
+
+	/* clear MBX ready (MBOX_READY_MASK bit is 0) and set update command */
+	reg_val |= (cmd << 16);
+	WREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_115,  reg_val);
+
+	/* Ring the doorbell */
+	WREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_73, 1);
+
+	if (cmd == C2PMSG_CMD_SPI_UPDATE_FLASH_IMAGE ||
+	    cmd == C2PMSG_CMD_SPI_GET_FLASH_IMAGE)
+		ret = psp_wait_for_spirom_update(psp, SOC15_REG_OFFSET(MP0, 0, regMP0_SMN_C2PMSG_115),
+						 MBOX_READY_FLAG, MBOX_READY_MASK, PSP_SPIROM_UPDATE_TIMEOUT);
+	else
+		ret = psp_wait_for(
+			psp, SOC15_REG_OFFSET(MP0, 0, regMP0_SMN_C2PMSG_115),
+			MBOX_READY_FLAG, MBOX_READY_MASK, 0);
+	if (ret) {
+		dev_err(adev->dev, "SPI cmd %x timed out, ret = %d", cmd, ret);
+		return ret;
+	}
+
+	reg_status = RREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_115);
+	if ((reg_status & 0xFFFF) != 0) {
+		dev_err(adev->dev, "SPI cmd %x failed, fail status = %04x\n",
+				cmd, reg_status & 0xFFFF);
+		return -EIO;
+	}
+
+	return 0;
+}
+
+static int psp_v13_0_update_spirom(struct psp_context *psp,
+				   uint64_t fw_pri_mc_addr)
+{
+	struct amdgpu_device *adev = psp->adev;
+	int ret;
+
+	/* Confirm PSP is ready to start */
+	ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, regMP0_SMN_C2PMSG_115),
+			   MBOX_READY_FLAG, MBOX_READY_MASK, 0);
+	if (ret) {
+		dev_err(adev->dev, "PSP Not ready to start processing, ret = %d", ret);
+		return ret;
+	}
+
+	WREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_116, lower_32_bits(fw_pri_mc_addr));
+
+	ret = psp_v13_0_exec_spi_cmd(psp, C2PMSG_CMD_SPI_UPDATE_ROM_IMAGE_ADDR_LO);
+	if (ret)
+		return ret;
+
+	WREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_116, upper_32_bits(fw_pri_mc_addr));
+
+	ret = psp_v13_0_exec_spi_cmd(psp, C2PMSG_CMD_SPI_UPDATE_ROM_IMAGE_ADDR_HI);
+	if (ret)
+		return ret;
+
+	psp->vbflash_done = true;
+
+	ret = psp_v13_0_exec_spi_cmd(psp, C2PMSG_CMD_SPI_UPDATE_FLASH_IMAGE);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+static int psp_v13_0_dump_spirom(struct psp_context *psp,
+				 uint64_t fw_pri_mc_addr)
+{
+	struct amdgpu_device *adev = psp->adev;
+	int ret;
+
+	/* Confirm PSP is ready to start */
+	ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, regMP0_SMN_C2PMSG_115),
+			   MBOX_READY_FLAG, MBOX_READY_MASK, 0);
+	if (ret) {
+		dev_err(adev->dev, "PSP Not ready to start processing, ret = %d", ret);
+		return ret;
+	}
+
+	WREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_116, lower_32_bits(fw_pri_mc_addr));
+
+	ret = psp_v13_0_exec_spi_cmd(psp, C2PMSG_CMD_SPI_GET_ROM_IMAGE_ADDR_LO);
+	if (ret)
+		return ret;
+
+	WREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_116, upper_32_bits(fw_pri_mc_addr));
+
+	ret = psp_v13_0_exec_spi_cmd(psp, C2PMSG_CMD_SPI_GET_ROM_IMAGE_ADDR_HI);
+	if (ret)
+		return ret;
+
+	ret = psp_v13_0_exec_spi_cmd(psp, C2PMSG_CMD_SPI_GET_FLASH_IMAGE);
+
+	return ret;
+}
+
+static int psp_v13_0_vbflash_status(struct psp_context *psp)
+{
+	struct amdgpu_device *adev = psp->adev;
+
+	return RREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_115);
+}
+
+static int psp_v13_0_fatal_error_recovery_quirk(struct psp_context *psp)
+{
+	struct amdgpu_device *adev = psp->adev;
+
+	if (amdgpu_ip_version(adev, MP0_HWIP, 0) == IP_VERSION(13, 0, 10)) {
+		uint32_t  reg_data;
+		/* MP1 fatal error: trigger PSP dram read to unhalt PSP
+		 * during MP1 triggered sync flood.
+		 */
+		reg_data = RREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_67);
+		WREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_67, reg_data + 0x10);
+
+		/* delay 1000ms for the mode1 reset for fatal error
+		 * to be recovered back.
+		 */
+		msleep(1000);
+	}
+
+	return 0;
+}
+
+static bool psp_v13_0_get_ras_capability(struct psp_context *psp)
+{
+	struct amdgpu_device *adev = psp->adev;
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+	u32 reg_data;
+
+	/* query ras cap should be done from host side */
+	if (amdgpu_sriov_vf(adev))
+		return false;
+
+	if (!con)
+		return false;
+
+	if ((amdgpu_ip_version(adev, MP0_HWIP, 0) == IP_VERSION(13, 0, 6) ||
+	     amdgpu_ip_version(adev, MP0_HWIP, 0) == IP_VERSION(13, 0, 12) ||
+	     amdgpu_ip_version(adev, MP0_HWIP, 0) == IP_VERSION(13, 0, 14)) &&
+	    (!(adev->flags & AMD_IS_APU))) {
+		reg_data = RREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_127);
+		adev->ras_hw_enabled = (reg_data & GENMASK_ULL(23, 0));
+		con->poison_supported = ((reg_data & GENMASK_ULL(24, 24)) >> 24) ? true : false;
+		return true;
+	} else {
+		return false;
+	}
+}
+
+static bool psp_v13_0_is_aux_sos_load_required(struct psp_context *psp)
+{
+	struct amdgpu_device *adev = psp->adev;
+	u32 pmfw_ver;
+
+	if (amdgpu_ip_version(adev, MP0_HWIP, 0) != IP_VERSION(13, 0, 6))
+		return false;
+
+	/* load 4e version of sos if pmfw version less than 85.115.0 */
+	pmfw_ver = RREG32(regMP1_PUB_SCRATCH0 / 4);
+
+	return (pmfw_ver < 0x557300);
+}
+
+static bool psp_v13_0_is_reload_needed(struct psp_context *psp)
+{
+	uint32_t ucode_ver;
+
+	if (!psp_v13_0_is_sos_alive(psp))
+		return false;
+
+	/* Restrict reload support only to specific IP versions */
+	switch (amdgpu_ip_version(psp->adev, MP0_HWIP, 0)) {
+	case IP_VERSION(13, 0, 2):
+	case IP_VERSION(13, 0, 6):
+	case IP_VERSION(13, 0, 14):
+		/* TOS version read from microcode header */
+		ucode_ver = psp->sos.fw_version;
+		/* Read TOS version from hardware */
+		psp_v13_0_init_sos_version(psp);
+		return (ucode_ver != psp->sos.fw_version);
+	default:
+		return false;
+	}
+
+	return false;
+}
+
+static int psp_v13_0_reg_program_no_ring(struct psp_context *psp, uint32_t val,
+					 enum psp_reg_prog_id id)
+{
+	struct amdgpu_device *adev = psp->adev;
+	int ret = -EOPNOTSUPP;
+
+	/* PSP will broadcast the value to all instances */
+	if (amdgpu_sriov_vf(adev)) {
+		WREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_101, GFX_CTRL_CMD_ID_GBR_IH_SET);
+		WREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_102, id);
+		WREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_103, val);
+
+		ret = psp_wait_for(
+			psp, SOC15_REG_OFFSET(MP0, 0, regMP0_SMN_C2PMSG_101),
+			0x80000000, 0x80000000, 0);
+	}
+
+	return ret;
+}
+
 static const struct psp_funcs psp_v13_0_funcs = {
 	.init_microcode = psp_v13_0_init_microcode,
+	.wait_for_bootloader = psp_v13_0_wait_for_bootloader_steady_state,
 	.bootloader_load_kdb = psp_v13_0_bootloader_load_kdb,
+	.bootloader_load_spl = psp_v13_0_bootloader_load_spl,
 	.bootloader_load_sysdrv = psp_v13_0_bootloader_load_sysdrv,
 	.bootloader_load_soc_drv = psp_v13_0_bootloader_load_soc_drv,
 	.bootloader_load_intf_drv = psp_v13_0_bootloader_load_intf_drv,
 	.bootloader_load_dbg_drv = psp_v13_0_bootloader_load_dbg_drv,
+	.bootloader_load_ras_drv = psp_v13_0_bootloader_load_ras_drv,
+	.bootloader_load_spdm_drv = psp_v13_0_bootloader_load_spdm_drv,
 	.bootloader_load_sos = psp_v13_0_bootloader_load_sos,
-	.ring_init = psp_v13_0_ring_init,
 	.ring_create = psp_v13_0_ring_create,
 	.ring_stop = psp_v13_0_ring_stop,
 	.ring_destroy = psp_v13_0_ring_destroy,
 	.ring_get_wptr = psp_v13_0_ring_get_wptr,
 	.ring_set_wptr = psp_v13_0_ring_set_wptr,
+	.mem_training = psp_v13_0_memory_training,
 	.load_usbc_pd_fw = psp_v13_0_load_usbc_pd_fw,
-	.read_usbc_pd_fw = psp_v13_0_read_usbc_pd_fw
+	.read_usbc_pd_fw = psp_v13_0_read_usbc_pd_fw,
+	.update_spirom = psp_v13_0_update_spirom,
+	.dump_spirom = psp_v13_0_dump_spirom,
+	.vbflash_stat = psp_v13_0_vbflash_status,
+	.fatal_error_recovery_quirk = psp_v13_0_fatal_error_recovery_quirk,
+	.get_ras_capability = psp_v13_0_get_ras_capability,
+	.is_aux_sos_load_required = psp_v13_0_is_aux_sos_load_required,
+	.is_reload_needed = psp_v13_0_is_reload_needed,
+	.reg_program_no_ring = psp_v13_0_reg_program_no_ring,
 };
 
 void psp_v13_0_set_psp_funcs(struct psp_context *psp)
diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v13_0.h b/drivers/gpu/drm/amd/amdgpu/psp_v13_0.h
index b2414a729ca1..de5677ce4330 100644
--- a/drivers/gpu/drm/amd/amdgpu/psp_v13_0.h
+++ b/drivers/gpu/drm/amd/amdgpu/psp_v13_0.h
@@ -25,6 +25,8 @@
 
 #include "amdgpu_psp.h"
 
+#define PSP_SPIROM_UPDATE_TIMEOUT   60000   /* 60s */
+
 void psp_v13_0_set_psp_funcs(struct psp_context *psp);
 
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v13_0_4.c b/drivers/gpu/drm/amd/amdgpu/psp_v13_0_4.c
new file mode 100644
index 000000000000..5f39a2edcc95
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/psp_v13_0_4.c
@@ -0,0 +1,355 @@
+/*
+ * Copyright 2020 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include "amdgpu.h"
+#include "amdgpu_psp.h"
+#include "amdgpu_ucode.h"
+#include "soc15_common.h"
+#include "psp_v13_0_4.h"
+
+#include "mp/mp_13_0_4_offset.h"
+#include "mp/mp_13_0_4_sh_mask.h"
+
+MODULE_FIRMWARE("amdgpu/psp_13_0_4_toc.bin");
+MODULE_FIRMWARE("amdgpu/psp_13_0_4_ta.bin");
+
+static int psp_v13_0_4_init_microcode(struct psp_context *psp)
+{
+	struct amdgpu_device *adev = psp->adev;
+	char ucode_prefix[30];
+	int err = 0;
+
+	amdgpu_ucode_ip_version_decode(adev, MP0_HWIP, ucode_prefix, sizeof(ucode_prefix));
+
+	switch (amdgpu_ip_version(adev, MP0_HWIP, 0)) {
+	case IP_VERSION(13, 0, 4):
+		err = psp_init_toc_microcode(psp, ucode_prefix);
+		if (err)
+			return err;
+		err = psp_init_ta_microcode(psp, ucode_prefix);
+		if (err)
+			return err;
+		break;
+	default:
+		BUG();
+	}
+
+	return 0;
+}
+
+static bool psp_v13_0_4_is_sos_alive(struct psp_context *psp)
+{
+	struct amdgpu_device *adev = psp->adev;
+	uint32_t sol_reg;
+
+	sol_reg = RREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_81);
+
+	return sol_reg != 0x0;
+}
+
+static int psp_v13_0_4_wait_for_bootloader(struct psp_context *psp)
+{
+	struct amdgpu_device *adev = psp->adev;
+
+	int ret;
+	int retry_loop;
+
+	for (retry_loop = 0; retry_loop < 10; retry_loop++) {
+		/* Wait for bootloader to signify that is
+		    ready having bit 31 of C2PMSG_35 set to 1 */
+		ret = psp_wait_for(
+			psp, SOC15_REG_OFFSET(MP0, 0, regMP0_SMN_C2PMSG_35),
+			0x80000000, 0x80000000, PSP_WAITREG_NOVERBOSE);
+
+		if (ret == 0)
+			return 0;
+	}
+
+	return ret;
+}
+
+static int psp_v13_0_4_bootloader_load_component(struct psp_context  	*psp,
+					       struct psp_bin_desc 	*bin_desc,
+					       enum psp_bootloader_cmd  bl_cmd)
+{
+	int ret;
+	uint32_t psp_gfxdrv_command_reg = 0;
+	struct amdgpu_device *adev = psp->adev;
+
+	/* Check tOS sign of life register to confirm sys driver and sOS
+	 * are already been loaded.
+	 */
+	if (psp_v13_0_4_is_sos_alive(psp))
+		return 0;
+
+	ret = psp_v13_0_4_wait_for_bootloader(psp);
+	if (ret)
+		return ret;
+
+	memset(psp->fw_pri_buf, 0, PSP_1_MEG);
+
+	/* Copy PSP KDB binary to memory */
+	memcpy(psp->fw_pri_buf, bin_desc->start_addr, bin_desc->size_bytes);
+
+	/* Provide the PSP KDB to bootloader */
+	WREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_36,
+	       (uint32_t)(psp->fw_pri_mc_addr >> 20));
+	psp_gfxdrv_command_reg = bl_cmd;
+	WREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_35,
+	       psp_gfxdrv_command_reg);
+
+	ret = psp_v13_0_4_wait_for_bootloader(psp);
+
+	return ret;
+}
+
+static int psp_v13_0_4_bootloader_load_kdb(struct psp_context *psp)
+{
+	return psp_v13_0_4_bootloader_load_component(psp, &psp->kdb, PSP_BL__LOAD_KEY_DATABASE);
+}
+
+static int psp_v13_0_4_bootloader_load_spl(struct psp_context *psp)
+{
+	return psp_v13_0_4_bootloader_load_component(psp, &psp->kdb, PSP_BL__LOAD_TOS_SPL_TABLE);
+}
+
+static int psp_v13_0_4_bootloader_load_sysdrv(struct psp_context *psp)
+{
+	return psp_v13_0_4_bootloader_load_component(psp, &psp->sys, PSP_BL__LOAD_SYSDRV);
+}
+
+static int psp_v13_0_4_bootloader_load_soc_drv(struct psp_context *psp)
+{
+	return psp_v13_0_4_bootloader_load_component(psp, &psp->soc_drv, PSP_BL__LOAD_SOCDRV);
+}
+
+static int psp_v13_0_4_bootloader_load_intf_drv(struct psp_context *psp)
+{
+	return psp_v13_0_4_bootloader_load_component(psp, &psp->intf_drv, PSP_BL__LOAD_INTFDRV);
+}
+
+static int psp_v13_0_4_bootloader_load_dbg_drv(struct psp_context *psp)
+{
+	return psp_v13_0_4_bootloader_load_component(psp, &psp->dbg_drv, PSP_BL__LOAD_DBGDRV);
+}
+
+static int psp_v13_0_4_bootloader_load_sos(struct psp_context *psp)
+{
+	int ret;
+	unsigned int psp_gfxdrv_command_reg = 0;
+	struct amdgpu_device *adev = psp->adev;
+
+	/* Check sOS sign of life register to confirm sys driver and sOS
+	 * are already been loaded.
+	 */
+	if (psp_v13_0_4_is_sos_alive(psp))
+		return 0;
+
+	ret = psp_v13_0_4_wait_for_bootloader(psp);
+	if (ret)
+		return ret;
+
+	memset(psp->fw_pri_buf, 0, PSP_1_MEG);
+
+	/* Copy Secure OS binary to PSP memory */
+	memcpy(psp->fw_pri_buf, psp->sos.start_addr, psp->sos.size_bytes);
+
+	/* Provide the PSP secure OS to bootloader */
+	WREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_36,
+	       (uint32_t)(psp->fw_pri_mc_addr >> 20));
+	psp_gfxdrv_command_reg = PSP_BL__LOAD_SOSDRV;
+	WREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_35,
+	       psp_gfxdrv_command_reg);
+
+	/* there might be handshake issue with hardware which needs delay */
+	mdelay(20);
+	ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, regMP0_SMN_C2PMSG_81),
+			   RREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_81), 0,
+			   PSP_WAITREG_CHANGED);
+
+	return ret;
+}
+
+static int psp_v13_0_4_ring_stop(struct psp_context *psp,
+			       enum psp_ring_type ring_type)
+{
+	int ret = 0;
+	struct amdgpu_device *adev = psp->adev;
+
+	if (amdgpu_sriov_vf(adev)) {
+		/* Write the ring destroy command*/
+		WREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_101,
+			     GFX_CTRL_CMD_ID_DESTROY_GPCOM_RING);
+		/* there might be handshake issue with hardware which needs delay */
+		mdelay(20);
+		/* Wait for response flag (bit 31) */
+		ret = psp_wait_for(
+			psp, SOC15_REG_OFFSET(MP0, 0, regMP0_SMN_C2PMSG_101),
+			MBOX_TOS_RESP_FLAG, MBOX_TOS_RESP_MASK, 0);
+	} else {
+		/* Write the ring destroy command*/
+		WREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_64,
+			     GFX_CTRL_CMD_ID_DESTROY_RINGS);
+		/* there might be handshake issue with hardware which needs delay */
+		mdelay(20);
+		/* Wait for response flag (bit 31) */
+		ret = psp_wait_for(
+			psp, SOC15_REG_OFFSET(MP0, 0, regMP0_SMN_C2PMSG_64),
+			MBOX_TOS_RESP_FLAG, MBOX_TOS_RESP_MASK, 0);
+	}
+
+	return ret;
+}
+
+static int psp_v13_0_4_ring_create(struct psp_context *psp,
+				 enum psp_ring_type ring_type)
+{
+	int ret = 0;
+	unsigned int psp_ring_reg = 0;
+	struct psp_ring *ring = &psp->km_ring;
+	struct amdgpu_device *adev = psp->adev;
+
+	if (amdgpu_sriov_vf(adev)) {
+		ret = psp_v13_0_4_ring_stop(psp, ring_type);
+		if (ret) {
+			DRM_ERROR("psp_v13_0_ring_stop_sriov failed!\n");
+			return ret;
+		}
+
+		/* Write low address of the ring to C2PMSG_102 */
+		psp_ring_reg = lower_32_bits(ring->ring_mem_mc_addr);
+		WREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_102, psp_ring_reg);
+		/* Write high address of the ring to C2PMSG_103 */
+		psp_ring_reg = upper_32_bits(ring->ring_mem_mc_addr);
+		WREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_103, psp_ring_reg);
+
+		/* Write the ring initialization command to C2PMSG_101 */
+		WREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_101,
+			     GFX_CTRL_CMD_ID_INIT_GPCOM_RING);
+
+		/* there might be handshake issue with hardware which needs delay */
+		mdelay(20);
+
+		/* Wait for response flag (bit 31) in C2PMSG_101 */
+		ret = psp_wait_for(
+			psp, SOC15_REG_OFFSET(MP0, 0, regMP0_SMN_C2PMSG_101),
+			MBOX_TOS_RESP_FLAG, MBOX_TOS_RESP_MASK, 0);
+
+	} else {
+		/* Wait for sOS ready for ring creation */
+		ret = psp_wait_for(
+			psp, SOC15_REG_OFFSET(MP0, 0, regMP0_SMN_C2PMSG_64),
+			MBOX_TOS_READY_FLAG, MBOX_TOS_READY_MASK, 0);
+		if (ret) {
+			DRM_ERROR("Failed to wait for trust OS ready for ring creation\n");
+			return ret;
+		}
+
+		/* Write low address of the ring to C2PMSG_69 */
+		psp_ring_reg = lower_32_bits(ring->ring_mem_mc_addr);
+		WREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_69, psp_ring_reg);
+		/* Write high address of the ring to C2PMSG_70 */
+		psp_ring_reg = upper_32_bits(ring->ring_mem_mc_addr);
+		WREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_70, psp_ring_reg);
+		/* Write size of ring to C2PMSG_71 */
+		psp_ring_reg = ring->ring_size;
+		WREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_71, psp_ring_reg);
+		/* Write the ring initialization command to C2PMSG_64 */
+		psp_ring_reg = ring_type;
+		psp_ring_reg = psp_ring_reg << 16;
+		WREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_64, psp_ring_reg);
+
+		/* there might be handshake issue with hardware which needs delay */
+		mdelay(20);
+
+		/* Wait for response flag (bit 31) in C2PMSG_64 */
+		ret = psp_wait_for(
+			psp, SOC15_REG_OFFSET(MP0, 0, regMP0_SMN_C2PMSG_64),
+			MBOX_TOS_RESP_FLAG, MBOX_TOS_RESP_MASK, 0);
+	}
+
+	return ret;
+}
+
+static int psp_v13_0_4_ring_destroy(struct psp_context *psp,
+				  enum psp_ring_type ring_type)
+{
+	int ret = 0;
+	struct psp_ring *ring = &psp->km_ring;
+	struct amdgpu_device *adev = psp->adev;
+
+	ret = psp_v13_0_4_ring_stop(psp, ring_type);
+	if (ret)
+		DRM_ERROR("Fail to stop psp ring\n");
+
+	amdgpu_bo_free_kernel(&adev->firmware.rbuf,
+			      &ring->ring_mem_mc_addr,
+			      (void **)&ring->ring_mem);
+
+	return ret;
+}
+
+static uint32_t psp_v13_0_4_ring_get_wptr(struct psp_context *psp)
+{
+	uint32_t data;
+	struct amdgpu_device *adev = psp->adev;
+
+	if (amdgpu_sriov_vf(adev))
+		data = RREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_102);
+	else
+		data = RREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_67);
+
+	return data;
+}
+
+static void psp_v13_0_4_ring_set_wptr(struct psp_context *psp, uint32_t value)
+{
+	struct amdgpu_device *adev = psp->adev;
+
+	if (amdgpu_sriov_vf(adev)) {
+		WREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_102, value);
+		WREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_101,
+			     GFX_CTRL_CMD_ID_CONSUME_CMD);
+	} else
+		WREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_67, value);
+}
+
+static const struct psp_funcs psp_v13_0_4_funcs = {
+	.init_microcode = psp_v13_0_4_init_microcode,
+	.bootloader_load_kdb = psp_v13_0_4_bootloader_load_kdb,
+	.bootloader_load_spl = psp_v13_0_4_bootloader_load_spl,
+	.bootloader_load_sysdrv = psp_v13_0_4_bootloader_load_sysdrv,
+	.bootloader_load_soc_drv = psp_v13_0_4_bootloader_load_soc_drv,
+	.bootloader_load_intf_drv = psp_v13_0_4_bootloader_load_intf_drv,
+	.bootloader_load_dbg_drv = psp_v13_0_4_bootloader_load_dbg_drv,
+	.bootloader_load_sos = psp_v13_0_4_bootloader_load_sos,
+	.ring_create = psp_v13_0_4_ring_create,
+	.ring_stop = psp_v13_0_4_ring_stop,
+	.ring_destroy = psp_v13_0_4_ring_destroy,
+	.ring_get_wptr = psp_v13_0_4_ring_get_wptr,
+	.ring_set_wptr = psp_v13_0_4_ring_set_wptr,
+};
+
+void psp_v13_0_4_set_psp_funcs(struct psp_context *psp)
+{
+	psp->funcs = &psp_v13_0_4_funcs;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v13_0_4.h b/drivers/gpu/drm/amd/amdgpu/psp_v13_0_4.h
new file mode 100644
index 000000000000..8547b8d514d5
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/psp_v13_0_4.h
@@ -0,0 +1,30 @@
+/*
+ * Copyright 2020 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __PSP_V13_0_4_H__
+#define __PSP_V13_0_4_H__
+
+#include "amdgpu_psp.h"
+
+void psp_v13_0_4_set_psp_funcs(struct psp_context *psp);
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v14_0.c b/drivers/gpu/drm/amd/amdgpu/psp_v14_0.c
new file mode 100644
index 000000000000..38dfc5c19f2a
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/psp_v14_0.c
@@ -0,0 +1,705 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include <drm/drm_drv.h>
+#include <linux/vmalloc.h>
+#include "amdgpu.h"
+#include "amdgpu_psp.h"
+#include "amdgpu_ucode.h"
+#include "soc15_common.h"
+#include "psp_v14_0.h"
+
+#include "mp/mp_14_0_2_offset.h"
+#include "mp/mp_14_0_2_sh_mask.h"
+
+MODULE_FIRMWARE("amdgpu/psp_14_0_2_sos.bin");
+MODULE_FIRMWARE("amdgpu/psp_14_0_2_ta.bin");
+MODULE_FIRMWARE("amdgpu/psp_14_0_3_sos.bin");
+MODULE_FIRMWARE("amdgpu/psp_14_0_3_sos_kicker.bin");
+MODULE_FIRMWARE("amdgpu/psp_14_0_3_ta.bin");
+MODULE_FIRMWARE("amdgpu/psp_14_0_3_ta_kicker.bin");
+MODULE_FIRMWARE("amdgpu/psp_14_0_5_toc.bin");
+MODULE_FIRMWARE("amdgpu/psp_14_0_5_ta.bin");
+
+/* For large FW files the time to complete can be very long */
+#define USBC_PD_POLLING_LIMIT_S 240
+
+/* Read USB-PD from LFB */
+#define GFX_CMD_USB_PD_USE_LFB 0x480
+
+/* VBIOS gfl defines */
+#define MBOX_READY_MASK 0x80000000
+#define MBOX_STATUS_MASK 0x0000FFFF
+#define MBOX_COMMAND_MASK 0x00FF0000
+#define MBOX_READY_FLAG 0x80000000
+#define C2PMSG_CMD_SPI_UPDATE_ROM_IMAGE_ADDR_LO 0x2
+#define C2PMSG_CMD_SPI_UPDATE_ROM_IMAGE_ADDR_HI 0x3
+#define C2PMSG_CMD_SPI_UPDATE_FLASH_IMAGE 0x4
+
+/* memory training timeout define */
+#define MEM_TRAIN_SEND_MSG_TIMEOUT_US	3000000
+
+static int psp_v14_0_init_microcode(struct psp_context *psp)
+{
+	struct amdgpu_device *adev = psp->adev;
+	char ucode_prefix[30];
+	int err = 0;
+
+	amdgpu_ucode_ip_version_decode(adev, MP0_HWIP, ucode_prefix, sizeof(ucode_prefix));
+
+	switch (amdgpu_ip_version(adev, MP0_HWIP, 0)) {
+	case IP_VERSION(14, 0, 2):
+	case IP_VERSION(14, 0, 3):
+		err = psp_init_sos_microcode(psp, ucode_prefix);
+		if (err)
+			return err;
+		err = psp_init_ta_microcode(psp, ucode_prefix);
+		if (err)
+			return err;
+		break;
+	case IP_VERSION(14, 0, 5):
+		err = psp_init_toc_microcode(psp, ucode_prefix);
+		if (err)
+			return err;
+		err = psp_init_ta_microcode(psp, ucode_prefix);
+		if (err)
+			return err;
+		break;
+	default:
+		BUG();
+	}
+
+	return 0;
+}
+
+static bool psp_v14_0_is_sos_alive(struct psp_context *psp)
+{
+	struct amdgpu_device *adev = psp->adev;
+	uint32_t sol_reg;
+
+	sol_reg = RREG32_SOC15(MP0, 0, regMPASP_SMN_C2PMSG_81);
+
+	return sol_reg != 0x0;
+}
+
+static int psp_v14_0_wait_for_bootloader(struct psp_context *psp)
+{
+	struct amdgpu_device *adev = psp->adev;
+
+	int ret;
+	int retry_loop;
+
+	for (retry_loop = 0; retry_loop < 10; retry_loop++) {
+		/* Wait for bootloader to signify that is
+		    ready having bit 31 of C2PMSG_35 set to 1 */
+		ret = psp_wait_for(
+			psp, SOC15_REG_OFFSET(MP0, 0, regMPASP_SMN_C2PMSG_35),
+			0x80000000, 0x80000000, PSP_WAITREG_NOVERBOSE);
+
+		if (ret == 0)
+			return 0;
+	}
+
+	return ret;
+}
+
+static int psp_v14_0_bootloader_load_component(struct psp_context  	*psp,
+					       struct psp_bin_desc 	*bin_desc,
+					       enum psp_bootloader_cmd  bl_cmd)
+{
+	int ret;
+	uint32_t psp_gfxdrv_command_reg = 0;
+	struct amdgpu_device *adev = psp->adev;
+
+	/* Check tOS sign of life register to confirm sys driver and sOS
+	 * are already been loaded.
+	 */
+	if (psp_v14_0_is_sos_alive(psp))
+		return 0;
+
+	ret = psp_v14_0_wait_for_bootloader(psp);
+	if (ret)
+		return ret;
+
+	memset(psp->fw_pri_buf, 0, PSP_1_MEG);
+
+	/* Copy PSP KDB binary to memory */
+	memcpy(psp->fw_pri_buf, bin_desc->start_addr, bin_desc->size_bytes);
+
+	/* Provide the PSP KDB to bootloader */
+	WREG32_SOC15(MP0, 0, regMPASP_SMN_C2PMSG_36,
+	       (uint32_t)(psp->fw_pri_mc_addr >> 20));
+	psp_gfxdrv_command_reg = bl_cmd;
+	WREG32_SOC15(MP0, 0, regMPASP_SMN_C2PMSG_35,
+	       psp_gfxdrv_command_reg);
+
+	ret = psp_v14_0_wait_for_bootloader(psp);
+
+	return ret;
+}
+
+static int psp_v14_0_bootloader_load_kdb(struct psp_context *psp)
+{
+	return psp_v14_0_bootloader_load_component(psp, &psp->kdb, PSP_BL__LOAD_KEY_DATABASE);
+}
+
+static int psp_v14_0_bootloader_load_spl(struct psp_context *psp)
+{
+	return psp_v14_0_bootloader_load_component(psp, &psp->spl, PSP_BL__LOAD_TOS_SPL_TABLE);
+}
+
+static int psp_v14_0_bootloader_load_sysdrv(struct psp_context *psp)
+{
+	return psp_v14_0_bootloader_load_component(psp, &psp->sys, PSP_BL__LOAD_SYSDRV);
+}
+
+static int psp_v14_0_bootloader_load_soc_drv(struct psp_context *psp)
+{
+	return psp_v14_0_bootloader_load_component(psp, &psp->soc_drv, PSP_BL__LOAD_SOCDRV);
+}
+
+static int psp_v14_0_bootloader_load_intf_drv(struct psp_context *psp)
+{
+	return psp_v14_0_bootloader_load_component(psp, &psp->intf_drv, PSP_BL__LOAD_INTFDRV);
+}
+
+static int psp_v14_0_bootloader_load_dbg_drv(struct psp_context *psp)
+{
+	/* dbg_drv was renamed to had_drv in psp v14 */
+	return psp_v14_0_bootloader_load_component(psp, &psp->dbg_drv, PSP_BL__LOAD_HADDRV);
+}
+
+static int psp_v14_0_bootloader_load_ras_drv(struct psp_context *psp)
+{
+	return psp_v14_0_bootloader_load_component(psp, &psp->ras_drv, PSP_BL__LOAD_RASDRV);
+}
+
+static int psp_v14_0_bootloader_load_ipkeymgr_drv(struct psp_context *psp)
+{
+	return psp_v14_0_bootloader_load_component(psp, &psp->ipkeymgr_drv, PSP_BL__LOAD_IPKEYMGRDRV);
+}
+
+static int psp_v14_0_bootloader_load_sos(struct psp_context *psp)
+{
+	int ret;
+	unsigned int psp_gfxdrv_command_reg = 0;
+	struct amdgpu_device *adev = psp->adev;
+
+	/* Check sOS sign of life register to confirm sys driver and sOS
+	 * are already been loaded.
+	 */
+	if (psp_v14_0_is_sos_alive(psp))
+		return 0;
+
+	ret = psp_v14_0_wait_for_bootloader(psp);
+	if (ret)
+		return ret;
+
+	memset(psp->fw_pri_buf, 0, PSP_1_MEG);
+
+	/* Copy Secure OS binary to PSP memory */
+	memcpy(psp->fw_pri_buf, psp->sos.start_addr, psp->sos.size_bytes);
+
+	/* Provide the PSP secure OS to bootloader */
+	WREG32_SOC15(MP0, 0, regMPASP_SMN_C2PMSG_36,
+	       (uint32_t)(psp->fw_pri_mc_addr >> 20));
+	psp_gfxdrv_command_reg = PSP_BL__LOAD_SOSDRV;
+	WREG32_SOC15(MP0, 0, regMPASP_SMN_C2PMSG_35,
+	       psp_gfxdrv_command_reg);
+
+	/* there might be handshake issue with hardware which needs delay */
+	mdelay(20);
+	ret = psp_wait_for(psp,
+			   SOC15_REG_OFFSET(MP0, 0, regMPASP_SMN_C2PMSG_81),
+			   RREG32_SOC15(MP0, 0, regMPASP_SMN_C2PMSG_81), 0,
+			   PSP_WAITREG_CHANGED);
+
+	return ret;
+}
+
+static int psp_v14_0_ring_stop(struct psp_context *psp,
+			       enum psp_ring_type ring_type)
+{
+	int ret = 0;
+	struct amdgpu_device *adev = psp->adev;
+
+	if (amdgpu_sriov_vf(adev)) {
+		/* Write the ring destroy command*/
+		WREG32_SOC15(MP0, 0, regMPASP_SMN_C2PMSG_101,
+			     GFX_CTRL_CMD_ID_DESTROY_GPCOM_RING);
+		/* there might be handshake issue with hardware which needs delay */
+		mdelay(20);
+		/* Wait for response flag (bit 31) */
+		ret = psp_wait_for(
+			psp, SOC15_REG_OFFSET(MP0, 0, regMPASP_SMN_C2PMSG_101),
+			MBOX_TOS_RESP_FLAG, MBOX_TOS_RESP_MASK, 0);
+	} else {
+		/* Write the ring destroy command*/
+		WREG32_SOC15(MP0, 0, regMPASP_SMN_C2PMSG_64,
+			     GFX_CTRL_CMD_ID_DESTROY_RINGS);
+		/* there might be handshake issue with hardware which needs delay */
+		mdelay(20);
+		/* Wait for response flag (bit 31) */
+		ret = psp_wait_for(
+			psp, SOC15_REG_OFFSET(MP0, 0, regMPASP_SMN_C2PMSG_64),
+			MBOX_TOS_RESP_FLAG, MBOX_TOS_RESP_MASK, 0);
+	}
+
+	return ret;
+}
+
+static int psp_v14_0_ring_create(struct psp_context *psp,
+				 enum psp_ring_type ring_type)
+{
+	int ret = 0;
+	unsigned int psp_ring_reg = 0;
+	struct psp_ring *ring = &psp->km_ring;
+	struct amdgpu_device *adev = psp->adev;
+
+	if (amdgpu_sriov_vf(adev)) {
+		ret = psp_v14_0_ring_stop(psp, ring_type);
+		if (ret) {
+			DRM_ERROR("psp_v14_0_ring_stop_sriov failed!\n");
+			return ret;
+		}
+
+		/* Write low address of the ring to C2PMSG_102 */
+		psp_ring_reg = lower_32_bits(ring->ring_mem_mc_addr);
+		WREG32_SOC15(MP0, 0, regMPASP_SMN_C2PMSG_102, psp_ring_reg);
+		/* Write high address of the ring to C2PMSG_103 */
+		psp_ring_reg = upper_32_bits(ring->ring_mem_mc_addr);
+		WREG32_SOC15(MP0, 0, regMPASP_SMN_C2PMSG_103, psp_ring_reg);
+
+		/* Write the ring initialization command to C2PMSG_101 */
+		WREG32_SOC15(MP0, 0, regMPASP_SMN_C2PMSG_101,
+			     GFX_CTRL_CMD_ID_INIT_GPCOM_RING);
+
+		/* there might be handshake issue with hardware which needs delay */
+		mdelay(20);
+
+		/* Wait for response flag (bit 31) in C2PMSG_101 */
+		ret = psp_wait_for(
+			psp, SOC15_REG_OFFSET(MP0, 0, regMPASP_SMN_C2PMSG_101),
+			MBOX_TOS_RESP_FLAG, MBOX_TOS_RESP_MASK, 0);
+
+	} else {
+		/* Wait for sOS ready for ring creation */
+		ret = psp_wait_for(
+			psp, SOC15_REG_OFFSET(MP0, 0, regMPASP_SMN_C2PMSG_64),
+			MBOX_TOS_READY_FLAG, MBOX_TOS_READY_MASK, 0);
+		if (ret) {
+			DRM_ERROR("Failed to wait for trust OS ready for ring creation\n");
+			return ret;
+		}
+
+		/* Write low address of the ring to C2PMSG_69 */
+		psp_ring_reg = lower_32_bits(ring->ring_mem_mc_addr);
+		WREG32_SOC15(MP0, 0, regMPASP_SMN_C2PMSG_69, psp_ring_reg);
+		/* Write high address of the ring to C2PMSG_70 */
+		psp_ring_reg = upper_32_bits(ring->ring_mem_mc_addr);
+		WREG32_SOC15(MP0, 0, regMPASP_SMN_C2PMSG_70, psp_ring_reg);
+		/* Write size of ring to C2PMSG_71 */
+		psp_ring_reg = ring->ring_size;
+		WREG32_SOC15(MP0, 0, regMPASP_SMN_C2PMSG_71, psp_ring_reg);
+		/* Write the ring initialization command to C2PMSG_64 */
+		psp_ring_reg = ring_type;
+		psp_ring_reg = psp_ring_reg << 16;
+		WREG32_SOC15(MP0, 0, regMPASP_SMN_C2PMSG_64, psp_ring_reg);
+
+		/* there might be handshake issue with hardware which needs delay */
+		mdelay(20);
+
+		/* Wait for response flag (bit 31) in C2PMSG_64 */
+		ret = psp_wait_for(
+			psp, SOC15_REG_OFFSET(MP0, 0, regMPASP_SMN_C2PMSG_64),
+			MBOX_TOS_RESP_FLAG, MBOX_TOS_RESP_MASK, 0);
+	}
+
+	return ret;
+}
+
+static int psp_v14_0_ring_destroy(struct psp_context *psp,
+				  enum psp_ring_type ring_type)
+{
+	int ret = 0;
+	struct psp_ring *ring = &psp->km_ring;
+	struct amdgpu_device *adev = psp->adev;
+
+	ret = psp_v14_0_ring_stop(psp, ring_type);
+	if (ret)
+		DRM_ERROR("Fail to stop psp ring\n");
+
+	amdgpu_bo_free_kernel(&adev->firmware.rbuf,
+			      &ring->ring_mem_mc_addr,
+			      (void **)&ring->ring_mem);
+
+	return ret;
+}
+
+static uint32_t psp_v14_0_ring_get_wptr(struct psp_context *psp)
+{
+	uint32_t data;
+	struct amdgpu_device *adev = psp->adev;
+
+	if (amdgpu_sriov_vf(adev))
+		data = RREG32_SOC15(MP0, 0, regMPASP_SMN_C2PMSG_102);
+	else
+		data = RREG32_SOC15(MP0, 0, regMPASP_SMN_C2PMSG_67);
+
+	return data;
+}
+
+static void psp_v14_0_ring_set_wptr(struct psp_context *psp, uint32_t value)
+{
+	struct amdgpu_device *adev = psp->adev;
+
+	if (amdgpu_sriov_vf(adev)) {
+		WREG32_SOC15(MP0, 0, regMPASP_SMN_C2PMSG_102, value);
+		WREG32_SOC15(MP0, 0, regMPASP_SMN_C2PMSG_101,
+			     GFX_CTRL_CMD_ID_CONSUME_CMD);
+	} else
+		WREG32_SOC15(MP0, 0, regMPASP_SMN_C2PMSG_67, value);
+}
+
+static int psp_v14_0_memory_training_send_msg(struct psp_context *psp, int msg)
+{
+	int ret;
+	int i;
+	uint32_t data_32;
+	int max_wait;
+	struct amdgpu_device *adev = psp->adev;
+
+	data_32 = (psp->mem_train_ctx.c2p_train_data_offset >> 20);
+	WREG32_SOC15(MP0, 0, regMPASP_SMN_C2PMSG_36, data_32);
+	WREG32_SOC15(MP0, 0, regMPASP_SMN_C2PMSG_35, msg);
+
+	max_wait = MEM_TRAIN_SEND_MSG_TIMEOUT_US / adev->usec_timeout;
+	for (i = 0; i < max_wait; i++) {
+		ret = psp_wait_for(
+			psp, SOC15_REG_OFFSET(MP0, 0, regMPASP_SMN_C2PMSG_35),
+			0x80000000, 0x80000000, PSP_WAITREG_NOVERBOSE);
+		if (ret == 0)
+			break;
+	}
+	if (i < max_wait)
+		ret = 0;
+	else
+		ret = -ETIME;
+
+	dev_dbg(adev->dev, "training %s %s, cost %d @ %d ms\n",
+		  (msg == PSP_BL__DRAM_SHORT_TRAIN) ? "short" : "long",
+		  (ret == 0) ? "succeed" : "failed",
+		  i, adev->usec_timeout/1000);
+	return ret;
+}
+
+
+static int psp_v14_0_memory_training(struct psp_context *psp, uint32_t ops)
+{
+	struct psp_memory_training_context *ctx = &psp->mem_train_ctx;
+	uint32_t *pcache = (uint32_t *)ctx->sys_cache;
+	struct amdgpu_device *adev = psp->adev;
+	uint32_t p2c_header[4];
+	uint32_t sz;
+	void *buf;
+	int ret, idx;
+
+	if (ctx->init == PSP_MEM_TRAIN_NOT_SUPPORT) {
+		dev_dbg(adev->dev, "Memory training is not supported.\n");
+		return 0;
+	} else if (ctx->init != PSP_MEM_TRAIN_INIT_SUCCESS) {
+		dev_err(adev->dev, "Memory training initialization failure.\n");
+		return -EINVAL;
+	}
+
+	if (psp_v14_0_is_sos_alive(psp)) {
+		dev_dbg(adev->dev, "SOS is alive, skip memory training.\n");
+		return 0;
+	}
+
+	amdgpu_device_vram_access(adev, ctx->p2c_train_data_offset, p2c_header, sizeof(p2c_header), false);
+	dev_dbg(adev->dev, "sys_cache[%08x,%08x,%08x,%08x] p2c_header[%08x,%08x,%08x,%08x]\n",
+		  pcache[0], pcache[1], pcache[2], pcache[3],
+		  p2c_header[0], p2c_header[1], p2c_header[2], p2c_header[3]);
+
+	if (ops & PSP_MEM_TRAIN_SEND_SHORT_MSG) {
+		dev_dbg(adev->dev, "Short training depends on restore.\n");
+		ops |= PSP_MEM_TRAIN_RESTORE;
+	}
+
+	if ((ops & PSP_MEM_TRAIN_RESTORE) &&
+	    pcache[0] != MEM_TRAIN_SYSTEM_SIGNATURE) {
+		dev_dbg(adev->dev, "sys_cache[0] is invalid, restore depends on save.\n");
+		ops |= PSP_MEM_TRAIN_SAVE;
+	}
+
+	if (p2c_header[0] == MEM_TRAIN_SYSTEM_SIGNATURE &&
+	    !(pcache[0] == MEM_TRAIN_SYSTEM_SIGNATURE &&
+	      pcache[3] == p2c_header[3])) {
+		dev_dbg(adev->dev, "sys_cache is invalid or out-of-date, need save training data to sys_cache.\n");
+		ops |= PSP_MEM_TRAIN_SAVE;
+	}
+
+	if ((ops & PSP_MEM_TRAIN_SAVE) &&
+	    p2c_header[0] != MEM_TRAIN_SYSTEM_SIGNATURE) {
+		dev_dbg(adev->dev, "p2c_header[0] is invalid, save depends on long training.\n");
+		ops |= PSP_MEM_TRAIN_SEND_LONG_MSG;
+	}
+
+	if (ops & PSP_MEM_TRAIN_SEND_LONG_MSG) {
+		ops &= ~PSP_MEM_TRAIN_SEND_SHORT_MSG;
+		ops |= PSP_MEM_TRAIN_SAVE;
+	}
+
+	dev_dbg(adev->dev, "Memory training ops:%x.\n", ops);
+
+	if (ops & PSP_MEM_TRAIN_SEND_LONG_MSG) {
+		/*
+		 * Long training will encroach a certain amount on the bottom of VRAM;
+		 * save the content from the bottom of VRAM to system memory
+		 * before training, and restore it after training to avoid
+		 * VRAM corruption.
+		 */
+		sz = BIST_MEM_TRAINING_ENCROACHED_SIZE;
+
+		if (adev->gmc.visible_vram_size < sz || !adev->mman.aper_base_kaddr) {
+			dev_err(adev->dev, "visible_vram_size %llx or aper_base_kaddr %p is not initialized.\n",
+				  adev->gmc.visible_vram_size,
+				  adev->mman.aper_base_kaddr);
+			return -EINVAL;
+		}
+
+		buf = vmalloc(sz);
+		if (!buf) {
+			dev_err(adev->dev, "failed to allocate system memory.\n");
+			return -ENOMEM;
+		}
+
+		if (drm_dev_enter(adev_to_drm(adev), &idx)) {
+			memcpy_fromio(buf, adev->mman.aper_base_kaddr, sz);
+			ret = psp_v14_0_memory_training_send_msg(psp, PSP_BL__DRAM_LONG_TRAIN);
+			if (ret) {
+				DRM_ERROR("Send long training msg failed.\n");
+				vfree(buf);
+				drm_dev_exit(idx);
+				return ret;
+			}
+
+			memcpy_toio(adev->mman.aper_base_kaddr, buf, sz);
+			amdgpu_device_flush_hdp(adev, NULL);
+			vfree(buf);
+			drm_dev_exit(idx);
+		} else {
+			vfree(buf);
+			return -ENODEV;
+		}
+	}
+
+	if (ops & PSP_MEM_TRAIN_SAVE) {
+		amdgpu_device_vram_access(psp->adev, ctx->p2c_train_data_offset, ctx->sys_cache, ctx->train_data_size, false);
+	}
+
+	if (ops & PSP_MEM_TRAIN_RESTORE) {
+		amdgpu_device_vram_access(psp->adev, ctx->c2p_train_data_offset, ctx->sys_cache, ctx->train_data_size, true);
+	}
+
+	if (ops & PSP_MEM_TRAIN_SEND_SHORT_MSG) {
+		ret = psp_v14_0_memory_training_send_msg(psp, (amdgpu_force_long_training > 0) ?
+							 PSP_BL__DRAM_LONG_TRAIN : PSP_BL__DRAM_SHORT_TRAIN);
+		if (ret) {
+			dev_err(adev->dev, "send training msg failed.\n");
+			return ret;
+		}
+	}
+	ctx->training_cnt++;
+	return 0;
+}
+
+static int psp_v14_0_load_usbc_pd_fw(struct psp_context *psp, uint64_t fw_pri_mc_addr)
+{
+	struct amdgpu_device *adev = psp->adev;
+	uint32_t reg_status;
+	int ret, i = 0;
+
+	/*
+	 * LFB address which is aligned to 1MB address and has to be
+	 * right-shifted by 20 so that LFB address can be passed on a 32-bit C2P
+	 * register
+	 */
+	WREG32_SOC15(MP0, 0, regMPASP_SMN_C2PMSG_36, (fw_pri_mc_addr >> 20));
+
+	ret = psp_wait_for(psp,
+			   SOC15_REG_OFFSET(MP0, 0, regMPASP_SMN_C2PMSG_35),
+			   0x80000000, 0x80000000, 0);
+	if (ret)
+		return ret;
+
+	/* Fireup interrupt so PSP can pick up the address */
+	WREG32_SOC15(MP0, 0, regMPASP_SMN_C2PMSG_35, (GFX_CMD_USB_PD_USE_LFB << 16));
+
+	/* FW load takes very long time */
+	do {
+		msleep(1000);
+		reg_status = RREG32_SOC15(MP0, 0, regMPASP_SMN_C2PMSG_35);
+
+		if (reg_status & 0x80000000)
+			goto done;
+
+	} while (++i < USBC_PD_POLLING_LIMIT_S);
+
+	return -ETIME;
+done:
+
+	if ((reg_status & 0xFFFF) != 0) {
+		DRM_ERROR("Address load failed - MP0_SMN_C2PMSG_35.Bits [15:0] = %04x\n",
+				reg_status & 0xFFFF);
+		return -EIO;
+	}
+
+	return 0;
+}
+
+static int psp_v14_0_read_usbc_pd_fw(struct psp_context *psp, uint32_t *fw_ver)
+{
+	struct amdgpu_device *adev = psp->adev;
+	int ret;
+
+	WREG32_SOC15(MP0, 0, regMPASP_SMN_C2PMSG_35, C2PMSG_CMD_GFX_USB_PD_FW_VER);
+
+	ret = psp_wait_for(psp,
+			   SOC15_REG_OFFSET(MP0, 0, regMPASP_SMN_C2PMSG_35),
+			   0x80000000, 0x80000000, 0);
+	if (!ret)
+		*fw_ver = RREG32_SOC15(MP0, 0, regMPASP_SMN_C2PMSG_36);
+
+	return ret;
+}
+
+static int psp_v14_0_exec_spi_cmd(struct psp_context *psp, int cmd)
+{
+	uint32_t reg_status = 0, reg_val = 0;
+	struct amdgpu_device *adev = psp->adev;
+	int ret;
+
+	/* clear MBX ready (MBOX_READY_MASK bit is 0) and set update command */
+	reg_val |= (cmd << 16);
+	WREG32_SOC15(MP0, 0, regMPASP_SMN_C2PMSG_115,  reg_val);
+
+	/* Ring the doorbell */
+	WREG32_SOC15(MP0, 0, regMPASP_SMN_C2PMSG_73, 1);
+
+	if (cmd == C2PMSG_CMD_SPI_UPDATE_FLASH_IMAGE)
+		ret = psp_wait_for_spirom_update(psp, SOC15_REG_OFFSET(MP0, 0, regMPASP_SMN_C2PMSG_115),
+						 MBOX_READY_FLAG, MBOX_READY_MASK, PSP_SPIROM_UPDATE_TIMEOUT);
+	else
+		ret = psp_wait_for(
+			psp, SOC15_REG_OFFSET(MP0, 0, regMPASP_SMN_C2PMSG_115),
+			MBOX_READY_FLAG, MBOX_READY_MASK, 0);
+
+	ret = psp_wait_for(psp,
+			   SOC15_REG_OFFSET(MP0, 0, regMPASP_SMN_C2PMSG_115),
+			   MBOX_READY_FLAG, MBOX_READY_MASK, 0);
+	if (ret) {
+		dev_err(adev->dev, "SPI cmd %x timed out, ret = %d", cmd, ret);
+		return ret;
+	}
+
+	reg_status = RREG32_SOC15(MP0, 0, regMPASP_SMN_C2PMSG_115);
+	if ((reg_status & 0xFFFF) != 0) {
+		dev_err(adev->dev, "SPI cmd %x failed, fail status = %04x\n",
+				cmd, reg_status & 0xFFFF);
+		return -EIO;
+	}
+
+	return 0;
+}
+
+static int psp_v14_0_update_spirom(struct psp_context *psp,
+				   uint64_t fw_pri_mc_addr)
+{
+	struct amdgpu_device *adev = psp->adev;
+	int ret;
+
+	/* Confirm PSP is ready to start */
+	ret = psp_wait_for(psp,
+			   SOC15_REG_OFFSET(MP0, 0, regMPASP_SMN_C2PMSG_115),
+			   MBOX_READY_FLAG, MBOX_READY_MASK, 0);
+	if (ret) {
+		dev_err(adev->dev, "PSP Not ready to start processing, ret = %d", ret);
+		return ret;
+	}
+
+	WREG32_SOC15(MP0, 0, regMPASP_SMN_C2PMSG_116, lower_32_bits(fw_pri_mc_addr));
+
+	ret = psp_v14_0_exec_spi_cmd(psp, C2PMSG_CMD_SPI_UPDATE_ROM_IMAGE_ADDR_LO);
+	if (ret)
+		return ret;
+
+	WREG32_SOC15(MP0, 0, regMPASP_SMN_C2PMSG_116, upper_32_bits(fw_pri_mc_addr));
+
+	ret = psp_v14_0_exec_spi_cmd(psp, C2PMSG_CMD_SPI_UPDATE_ROM_IMAGE_ADDR_HI);
+	if (ret)
+		return ret;
+
+	psp->vbflash_done = true;
+
+	ret = psp_v14_0_exec_spi_cmd(psp, C2PMSG_CMD_SPI_UPDATE_FLASH_IMAGE);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+static int psp_v14_0_vbflash_status(struct psp_context *psp)
+{
+	struct amdgpu_device *adev = psp->adev;
+
+	return RREG32_SOC15(MP0, 0, regMPASP_SMN_C2PMSG_115);
+}
+
+static const struct psp_funcs psp_v14_0_funcs = {
+	.init_microcode = psp_v14_0_init_microcode,
+	.bootloader_load_kdb = psp_v14_0_bootloader_load_kdb,
+	.bootloader_load_spl = psp_v14_0_bootloader_load_spl,
+	.bootloader_load_sysdrv = psp_v14_0_bootloader_load_sysdrv,
+	.bootloader_load_soc_drv = psp_v14_0_bootloader_load_soc_drv,
+	.bootloader_load_intf_drv = psp_v14_0_bootloader_load_intf_drv,
+	.bootloader_load_dbg_drv = psp_v14_0_bootloader_load_dbg_drv,
+	.bootloader_load_ras_drv = psp_v14_0_bootloader_load_ras_drv,
+	.bootloader_load_ipkeymgr_drv = psp_v14_0_bootloader_load_ipkeymgr_drv,
+	.bootloader_load_sos = psp_v14_0_bootloader_load_sos,
+	.ring_create = psp_v14_0_ring_create,
+	.ring_stop = psp_v14_0_ring_stop,
+	.ring_destroy = psp_v14_0_ring_destroy,
+	.ring_get_wptr = psp_v14_0_ring_get_wptr,
+	.ring_set_wptr = psp_v14_0_ring_set_wptr,
+	.mem_training = psp_v14_0_memory_training,
+	.load_usbc_pd_fw = psp_v14_0_load_usbc_pd_fw,
+	.read_usbc_pd_fw = psp_v14_0_read_usbc_pd_fw,
+	.update_spirom = psp_v14_0_update_spirom,
+	.vbflash_stat = psp_v14_0_vbflash_status
+};
+
+void psp_v14_0_set_psp_funcs(struct psp_context *psp)
+{
+	psp->funcs = &psp_v14_0_funcs;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v14_0.h b/drivers/gpu/drm/amd/amdgpu/psp_v14_0.h
new file mode 100644
index 000000000000..dd18ba2cfad5
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/psp_v14_0.h
@@ -0,0 +1,32 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __PSP_V14_0_H__
+#define __PSP_V14_0_H__
+
+#include "amdgpu_psp.h"
+
+#define PSP_SPIROM_UPDATE_TIMEOUT   60000   /* 60s */
+
+void psp_v14_0_set_psp_funcs(struct psp_context *psp);
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v3_1.c b/drivers/gpu/drm/amd/amdgpu/psp_v3_1.c
index 1ed357cb0f49..833830bc3e2e 100644
--- a/drivers/gpu/drm/amd/amdgpu/psp_v3_1.c
+++ b/drivers/gpu/drm/amd/amdgpu/psp_v3_1.c
@@ -44,6 +44,7 @@
 
 MODULE_FIRMWARE("amdgpu/vega10_sos.bin");
 MODULE_FIRMWARE("amdgpu/vega10_asd.bin");
+MODULE_FIRMWARE("amdgpu/vega10_cap.bin");
 MODULE_FIRMWARE("amdgpu/vega12_sos.bin");
 MODULE_FIRMWARE("amdgpu/vega12_asd.bin");
 
@@ -56,26 +57,18 @@ static int psp_v3_1_ring_stop(struct psp_context *psp,
 static int psp_v3_1_init_microcode(struct psp_context *psp)
 {
 	struct amdgpu_device *adev = psp->adev;
-	const char *chip_name;
+	char ucode_prefix[30];
 	int err = 0;
 
 	DRM_DEBUG("\n");
 
-	switch (adev->asic_type) {
-	case CHIP_VEGA10:
-		chip_name = "vega10";
-		break;
-	case CHIP_VEGA12:
-		chip_name = "vega12";
-		break;
-	default: BUG();
-	}
+	amdgpu_ucode_ip_version_decode(adev, MP0_HWIP, ucode_prefix, sizeof(ucode_prefix));
 
-	err = psp_init_sos_microcode(psp, chip_name);
+	err = psp_init_sos_microcode(psp, ucode_prefix);
 	if (err)
 		return err;
 
-	err = psp_init_asd_microcode(psp, chip_name);
+	err = psp_init_asd_microcode(psp, ucode_prefix);
 	if (err)
 		return err;
 
@@ -98,7 +91,7 @@ static int psp_v3_1_bootloader_load_sysdrv(struct psp_context *psp)
 
 	/* Wait for bootloader to signify that is ready having bit 31 of C2PMSG_35 set to 1 */
 	ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_35),
-			   0x80000000, 0x80000000, false);
+			   0x80000000, 0x80000000, 0);
 	if (ret)
 		return ret;
 
@@ -116,7 +109,7 @@ static int psp_v3_1_bootloader_load_sysdrv(struct psp_context *psp)
 	mdelay(20);
 
 	ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_35),
-			   0x80000000, 0x80000000, false);
+			   0x80000000, 0x80000000, 0);
 
 	return ret;
 }
@@ -137,7 +130,7 @@ static int psp_v3_1_bootloader_load_sos(struct psp_context *psp)
 
 	/* Wait for bootloader to signify that is ready having bit 31 of C2PMSG_35 set to 1 */
 	ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_35),
-			   0x80000000, 0x80000000, false);
+			   0x80000000, 0x80000000, 0);
 	if (ret)
 		return ret;
 
@@ -154,37 +147,11 @@ static int psp_v3_1_bootloader_load_sos(struct psp_context *psp)
 	/* there might be handshake issue with hardware which needs delay */
 	mdelay(20);
 	ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_81),
-			   RREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_81),
-			   0, true);
+			   RREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_81), 0,
+			   PSP_WAITREG_CHANGED);
 	return ret;
 }
 
-static int psp_v3_1_ring_init(struct psp_context *psp,
-			      enum psp_ring_type ring_type)
-{
-	int ret = 0;
-	struct psp_ring *ring;
-	struct amdgpu_device *adev = psp->adev;
-
-	ring = &psp->km_ring;
-
-	ring->ring_type = ring_type;
-
-	/* allocate 4k Page of Local Frame Buffer memory for ring */
-	ring->ring_size = 0x1000;
-	ret = amdgpu_bo_create_kernel(adev, ring->ring_size, PAGE_SIZE,
-				      AMDGPU_GEM_DOMAIN_VRAM,
-				      &adev->firmware.rbuf,
-				      &ring->ring_mem_mc_addr,
-				      (void **)&ring->ring_mem);
-	if (ret) {
-		ring->ring_size = 0;
-		return ret;
-	}
-
-	return 0;
-}
-
 static void psp_v3_1_reroute_ih(struct psp_context *psp)
 {
 	struct amdgpu_device *adev = psp->adev;
@@ -201,7 +168,7 @@ static void psp_v3_1_reroute_ih(struct psp_context *psp)
 
 	mdelay(20);
 	psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_64),
-		     0x80000000, 0x8000FFFF, false);
+		     0x80000000, 0x8000FFFF, 0);
 
 	/* Change IH ring for UMC */
 	tmp = REG_SET_FIELD(0, IH_CLIENT_CFG_DATA, CREDIT_RETURN_ADDR, 0x1216b);
@@ -213,7 +180,7 @@ static void psp_v3_1_reroute_ih(struct psp_context *psp)
 
 	mdelay(20);
 	psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_64),
-		     0x80000000, 0x8000FFFF, false);
+		     0x80000000, 0x8000FFFF, 0);
 }
 
 static int psp_v3_1_ring_create(struct psp_context *psp,
@@ -250,9 +217,9 @@ static int psp_v3_1_ring_create(struct psp_context *psp,
 		mdelay(20);
 
 		/* Wait for response flag (bit 31) in C2PMSG_101 */
-		ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0,
-					mmMP0_SMN_C2PMSG_101), 0x80000000,
-					0x8000FFFF, false);
+		ret = psp_wait_for(
+			psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_101),
+			0x80000000, 0x8000FFFF, 0);
 	} else {
 
 		/* Write low address of the ring to C2PMSG_69 */
@@ -273,10 +240,9 @@ static int psp_v3_1_ring_create(struct psp_context *psp,
 		mdelay(20);
 
 		/* Wait for response flag (bit 31) in C2PMSG_64 */
-		ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0,
-					mmMP0_SMN_C2PMSG_64), 0x80000000,
-					0x8000FFFF, false);
-
+		ret = psp_wait_for(
+			psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_64),
+			0x80000000, 0x8000FFFF, 0);
 	}
 	return ret;
 }
@@ -300,11 +266,13 @@ static int psp_v3_1_ring_stop(struct psp_context *psp,
 
 	/* Wait for response flag (bit 31) */
 	if (amdgpu_sriov_vf(adev))
-		ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_101),
-				   0x80000000, 0x80000000, false);
+		ret = psp_wait_for(
+			psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_101),
+			0x80000000, 0x80000000, 0);
 	else
-		ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_64),
-				   0x80000000, 0x80000000, false);
+		ret = psp_wait_for(
+			psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_64),
+			0x80000000, 0x80000000, 0);
 
 	return ret;
 }
@@ -344,7 +312,7 @@ static int psp_v3_1_mode1_reset(struct psp_context *psp)
 
 	offset = SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_64);
 
-	ret = psp_wait_for(psp, offset, 0x80000000, 0x8000FFFF, false);
+	ret = psp_wait_for(psp, offset, 0x80000000, 0x8000FFFF, 0);
 
 	if (ret) {
 		DRM_INFO("psp is not working correctly before mode1 reset!\n");
@@ -358,7 +326,7 @@ static int psp_v3_1_mode1_reset(struct psp_context *psp)
 
 	offset = SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_33);
 
-	ret = psp_wait_for(psp, offset, 0x80000000, 0x80000000, false);
+	ret = psp_wait_for(psp, offset, 0x80000000, 0x80000000, 0);
 
 	if (ret) {
 		DRM_INFO("psp mode 1 reset failed!\n");
@@ -400,7 +368,6 @@ static const struct psp_funcs psp_v3_1_funcs = {
 	.init_microcode = psp_v3_1_init_microcode,
 	.bootloader_load_sysdrv = psp_v3_1_bootloader_load_sysdrv,
 	.bootloader_load_sos = psp_v3_1_bootloader_load_sos,
-	.ring_init = psp_v3_1_ring_init,
 	.ring_create = psp_v3_1_ring_create,
 	.ring_stop = psp_v3_1_ring_stop,
 	.ring_destroy = psp_v3_1_ring_destroy,
diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c b/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c
index 4509bd4cce2d..92ce580647cd 100644
--- a/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c
@@ -57,22 +57,19 @@ static void sdma_v2_4_set_irq_funcs(struct amdgpu_device *adev);
 MODULE_FIRMWARE("amdgpu/topaz_sdma.bin");
 MODULE_FIRMWARE("amdgpu/topaz_sdma1.bin");
 
-static const u32 sdma_offsets[SDMA_MAX_INSTANCE] =
-{
+static const u32 sdma_offsets[SDMA_MAX_INSTANCE] = {
 	SDMA0_REGISTER_OFFSET,
 	SDMA1_REGISTER_OFFSET
 };
 
-static const u32 golden_settings_iceland_a11[] =
-{
+static const u32 golden_settings_iceland_a11[] = {
 	mmSDMA0_CHICKEN_BITS, 0xfc910007, 0x00810007,
 	mmSDMA0_CLK_CTRL, 0xff000fff, 0x00000000,
 	mmSDMA1_CHICKEN_BITS, 0xfc910007, 0x00810007,
 	mmSDMA1_CLK_CTRL, 0xff000fff, 0x00000000,
 };
 
-static const u32 iceland_mgcg_cgcg_init[] =
-{
+static const u32 iceland_mgcg_cgcg_init[] = {
 	mmSDMA0_CLK_CTRL, 0xff000ff0, 0x00000100,
 	mmSDMA1_CLK_CTRL, 0xff000ff0, 0x00000100
 };
@@ -113,10 +110,9 @@ static void sdma_v2_4_init_golden_registers(struct amdgpu_device *adev)
 static void sdma_v2_4_free_microcode(struct amdgpu_device *adev)
 {
 	int i;
-	for (i = 0; i < adev->sdma.num_instances; i++) {
-		release_firmware(adev->sdma.instance[i].fw);
-		adev->sdma.instance[i].fw = NULL;
-	}
+
+	for (i = 0; i < adev->sdma.num_instances; i++)
+		amdgpu_ucode_release(&adev->sdma.instance[i].fw);
 }
 
 /**
@@ -131,7 +127,6 @@ static void sdma_v2_4_free_microcode(struct amdgpu_device *adev)
 static int sdma_v2_4_init_microcode(struct amdgpu_device *adev)
 {
 	const char *chip_name;
-	char fw_name[30];
 	int err = 0, i;
 	struct amdgpu_firmware_info *info = NULL;
 	const struct common_firmware_header *header = NULL;
@@ -143,18 +138,19 @@ static int sdma_v2_4_init_microcode(struct amdgpu_device *adev)
 	case CHIP_TOPAZ:
 		chip_name = "topaz";
 		break;
-	default: BUG();
+	default:
+		BUG();
 	}
 
 	for (i = 0; i < adev->sdma.num_instances; i++) {
 		if (i == 0)
-			snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_sdma.bin", chip_name);
+			err = amdgpu_ucode_request(adev, &adev->sdma.instance[i].fw,
+						   AMDGPU_UCODE_REQUIRED,
+						   "amdgpu/%s_sdma.bin", chip_name);
 		else
-			snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_sdma1.bin", chip_name);
-		err = request_firmware(&adev->sdma.instance[i].fw, fw_name, adev->dev);
-		if (err)
-			goto out;
-		err = amdgpu_ucode_validate(adev->sdma.instance[i].fw);
+			err = amdgpu_ucode_request(adev, &adev->sdma.instance[i].fw,
+						   AMDGPU_UCODE_REQUIRED,
+						   "amdgpu/%s_sdma1.bin", chip_name);
 		if (err)
 			goto out;
 		hdr = (const struct sdma_firmware_header_v1_0 *)adev->sdma.instance[i].fw->data;
@@ -175,11 +171,10 @@ static int sdma_v2_4_init_microcode(struct amdgpu_device *adev)
 
 out:
 	if (err) {
-		pr_err("sdma_v2_4: Failed to load firmware \"%s\"\n", fw_name);
-		for (i = 0; i < adev->sdma.num_instances; i++) {
-			release_firmware(adev->sdma.instance[i].fw);
-			adev->sdma.instance[i].fw = NULL;
-		}
+		pr_err("sdma_v2_4: Failed to load firmware \"%s_sdma%s.bin\"\n",
+		       chip_name, i == 0 ? "" : "1");
+		for (i = 0; i < adev->sdma.num_instances; i++)
+			amdgpu_ucode_release(&adev->sdma.instance[i].fw);
 	}
 	return err;
 }
@@ -194,7 +189,7 @@ out:
 static uint64_t sdma_v2_4_ring_get_rptr(struct amdgpu_ring *ring)
 {
 	/* XXX check if swapping is necessary on BE */
-	return ring->adev->wb.wb[ring->rptr_offs] >> 2;
+	return *ring->rptr_cpu_addr >> 2;
 }
 
 /**
@@ -223,7 +218,7 @@ static void sdma_v2_4_ring_set_wptr(struct amdgpu_ring *ring)
 {
 	struct amdgpu_device *adev = ring->adev;
 
-	WREG32(mmSDMA0_GFX_RB_WPTR + sdma_offsets[ring->me], lower_32_bits(ring->wptr) << 2);
+	WREG32(mmSDMA0_GFX_RB_WPTR + sdma_offsets[ring->me], ring->wptr << 2);
 }
 
 static void sdma_v2_4_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count)
@@ -342,15 +337,9 @@ static void sdma_v2_4_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 se
  */
 static void sdma_v2_4_gfx_stop(struct amdgpu_device *adev)
 {
-	struct amdgpu_ring *sdma0 = &adev->sdma.instance[0].ring;
-	struct amdgpu_ring *sdma1 = &adev->sdma.instance[1].ring;
 	u32 rb_cntl, ib_cntl;
 	int i;
 
-	if ((adev->mman.buffer_funcs_ring == sdma0) ||
-	    (adev->mman.buffer_funcs_ring == sdma1))
-		amdgpu_ttm_set_buffer_funcs_status(adev, false);
-
 	for (i = 0; i < adev->sdma.num_instances; i++) {
 		rb_cntl = RREG32(mmSDMA0_GFX_RB_CNTL + sdma_offsets[i]);
 		rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL, RB_ENABLE, 0);
@@ -414,12 +403,10 @@ static int sdma_v2_4_gfx_resume(struct amdgpu_device *adev)
 	struct amdgpu_ring *ring;
 	u32 rb_cntl, ib_cntl;
 	u32 rb_bufsz;
-	u32 wb_offset;
 	int i, j, r;
 
 	for (i = 0; i < adev->sdma.num_instances; i++) {
 		ring = &adev->sdma.instance[i].ring;
-		wb_offset = (ring->rptr_offs * 4);
 
 		mutex_lock(&adev->srbm_mutex);
 		for (j = 0; j < 16; j++) {
@@ -455,9 +442,9 @@ static int sdma_v2_4_gfx_resume(struct amdgpu_device *adev)
 
 		/* set the wb address whether it's enabled or not */
 		WREG32(mmSDMA0_GFX_RB_RPTR_ADDR_HI + sdma_offsets[i],
-		       upper_32_bits(adev->wb.gpu_addr + wb_offset) & 0xFFFFFFFF);
+		       upper_32_bits(ring->rptr_gpu_addr) & 0xFFFFFFFF);
 		WREG32(mmSDMA0_GFX_RB_RPTR_ADDR_LO + sdma_offsets[i],
-		       lower_32_bits(adev->wb.gpu_addr + wb_offset) & 0xFFFFFFFC);
+		       lower_32_bits(ring->rptr_gpu_addr) & 0xFFFFFFFC);
 
 		rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL, RPTR_WRITEBACK_ENABLE, 1);
 
@@ -465,7 +452,7 @@ static int sdma_v2_4_gfx_resume(struct amdgpu_device *adev)
 		WREG32(mmSDMA0_GFX_RB_BASE_HI + sdma_offsets[i], ring->gpu_addr >> 40);
 
 		ring->wptr = 0;
-		WREG32(mmSDMA0_GFX_RB_WPTR + sdma_offsets[i], lower_32_bits(ring->wptr) << 2);
+		WREG32(mmSDMA0_GFX_RB_WPTR + sdma_offsets[i], ring->wptr << 2);
 
 		/* enable DMA RB */
 		rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL, RB_ENABLE, 1);
@@ -478,8 +465,6 @@ static int sdma_v2_4_gfx_resume(struct amdgpu_device *adev)
 #endif
 		/* enable DMA IBs */
 		WREG32(mmSDMA0_GFX_IB_CNTL + sdma_offsets[i], ib_cntl);
-
-		ring->sched.ready = true;
 	}
 
 	sdma_v2_4_enable(adev, true);
@@ -488,9 +473,6 @@ static int sdma_v2_4_gfx_resume(struct amdgpu_device *adev)
 		r = amdgpu_ring_test_helper(ring);
 		if (r)
 			return r;
-
-		if (adev->mman.buffer_funcs_ring == ring)
-			amdgpu_ttm_set_buffer_funcs_status(adev, true);
 	}
 
 	return 0;
@@ -651,7 +633,7 @@ static int sdma_v2_4_ring_test_ib(struct amdgpu_ring *ring, long timeout)
 		r = -EINVAL;
 
 err1:
-	amdgpu_ib_free(adev, &ib, NULL);
+	amdgpu_ib_free(&ib, NULL);
 	dma_fence_put(f);
 err0:
 	amdgpu_device_wb_free(adev, index);
@@ -827,12 +809,17 @@ static void sdma_v2_4_ring_emit_wreg(struct amdgpu_ring *ring,
 	amdgpu_ring_write(ring, val);
 }
 
-static int sdma_v2_4_early_init(void *handle)
+static int sdma_v2_4_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
+	int r;
 
 	adev->sdma.num_instances = SDMA_MAX_INSTANCE;
 
+	r = sdma_v2_4_init_microcode(adev);
+	if (r)
+		return r;
+
 	sdma_v2_4_set_ring_funcs(adev);
 	sdma_v2_4_set_buffer_funcs(adev);
 	sdma_v2_4_set_vm_pte_funcs(adev);
@@ -841,11 +828,11 @@ static int sdma_v2_4_early_init(void *handle)
 	return 0;
 }
 
-static int sdma_v2_4_sw_init(void *handle)
+static int sdma_v2_4_sw_init(struct amdgpu_ip_block *ip_block)
 {
 	struct amdgpu_ring *ring;
 	int r, i;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	/* SDMA trap event */
 	r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_SDMA_TRAP,
@@ -865,12 +852,6 @@ static int sdma_v2_4_sw_init(void *handle)
 	if (r)
 		return r;
 
-	r = sdma_v2_4_init_microcode(adev);
-	if (r) {
-		DRM_ERROR("Failed to load sdma firmware!\n");
-		return r;
-	}
-
 	for (i = 0; i < adev->sdma.num_instances; i++) {
 		ring = &adev->sdma.instance[i].ring;
 		ring->ring_obj = NULL;
@@ -887,9 +868,9 @@ static int sdma_v2_4_sw_init(void *handle)
 	return r;
 }
 
-static int sdma_v2_4_sw_fini(void *handle)
+static int sdma_v2_4_sw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int i;
 
 	for (i = 0; i < adev->sdma.num_instances; i++)
@@ -899,10 +880,10 @@ static int sdma_v2_4_sw_fini(void *handle)
 	return 0;
 }
 
-static int sdma_v2_4_hw_init(void *handle)
+static int sdma_v2_4_hw_init(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	sdma_v2_4_init_golden_registers(adev);
 
@@ -913,32 +894,26 @@ static int sdma_v2_4_hw_init(void *handle)
 	return r;
 }
 
-static int sdma_v2_4_hw_fini(void *handle)
+static int sdma_v2_4_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	sdma_v2_4_enable(adev, false);
+	sdma_v2_4_enable(ip_block->adev, false);
 
 	return 0;
 }
 
-static int sdma_v2_4_suspend(void *handle)
+static int sdma_v2_4_suspend(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	return sdma_v2_4_hw_fini(adev);
+	return sdma_v2_4_hw_fini(ip_block);
 }
 
-static int sdma_v2_4_resume(void *handle)
+static int sdma_v2_4_resume(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	return sdma_v2_4_hw_init(adev);
+	return sdma_v2_4_hw_init(ip_block);
 }
 
-static bool sdma_v2_4_is_idle(void *handle)
+static bool sdma_v2_4_is_idle(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	u32 tmp = RREG32(mmSRBM_STATUS2);
 
 	if (tmp & (SRBM_STATUS2__SDMA_BUSY_MASK |
@@ -948,11 +923,11 @@ static bool sdma_v2_4_is_idle(void *handle)
 	return true;
 }
 
-static int sdma_v2_4_wait_for_idle(void *handle)
+static int sdma_v2_4_wait_for_idle(struct amdgpu_ip_block *ip_block)
 {
 	unsigned i;
 	u32 tmp;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	for (i = 0; i < adev->usec_timeout; i++) {
 		tmp = RREG32(mmSRBM_STATUS2) & (SRBM_STATUS2__SDMA_BUSY_MASK |
@@ -965,10 +940,10 @@ static int sdma_v2_4_wait_for_idle(void *handle)
 	return -ETIMEDOUT;
 }
 
-static int sdma_v2_4_soft_reset(void *handle)
+static int sdma_v2_4_soft_reset(struct amdgpu_ip_block *ip_block)
 {
 	u32 srbm_soft_reset = 0;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	u32 tmp = RREG32(mmSRBM_STATUS2);
 
 	if (tmp & SRBM_STATUS2__SDMA_BUSY_MASK) {
@@ -1107,14 +1082,14 @@ static int sdma_v2_4_process_illegal_inst_irq(struct amdgpu_device *adev,
 	return 0;
 }
 
-static int sdma_v2_4_set_clockgating_state(void *handle,
+static int sdma_v2_4_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_clockgating_state state)
 {
 	/* XXX handled via the smc on VI */
 	return 0;
 }
 
-static int sdma_v2_4_set_powergating_state(void *handle,
+static int sdma_v2_4_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_powergating_state state)
 {
 	return 0;
@@ -1123,7 +1098,6 @@ static int sdma_v2_4_set_powergating_state(void *handle,
 static const struct amd_ip_funcs sdma_v2_4_ip_funcs = {
 	.name = "sdma_v2_4",
 	.early_init = sdma_v2_4_early_init,
-	.late_init = NULL,
 	.sw_init = sdma_v2_4_sw_init,
 	.sw_fini = sdma_v2_4_sw_fini,
 	.hw_init = sdma_v2_4_hw_init,
@@ -1142,6 +1116,7 @@ static const struct amdgpu_ring_funcs sdma_v2_4_ring_funcs = {
 	.align_mask = 0xf,
 	.nop = SDMA_PKT_NOP_HEADER_OP(SDMA_OP_NOP),
 	.support_64bit_ptrs = false,
+	.secure_submission_supported = true,
 	.get_rptr = sdma_v2_4_ring_get_rptr,
 	.get_wptr = sdma_v2_4_ring_get_wptr,
 	.set_wptr = sdma_v2_4_ring_set_wptr,
@@ -1197,7 +1172,7 @@ static void sdma_v2_4_set_irq_funcs(struct amdgpu_device *adev)
  * @src_offset: src GPU address
  * @dst_offset: dst GPU address
  * @byte_count: number of bytes to xfer
- * @tmz: unused
+ * @copy_flags: unused
  *
  * Copy GPU buffers using the DMA engine (VI).
  * Used by the amdgpu ttm implementation to move pages if
@@ -1207,7 +1182,7 @@ static void sdma_v2_4_emit_copy_buffer(struct amdgpu_ib *ib,
 				       uint64_t src_offset,
 				       uint64_t dst_offset,
 				       uint32_t byte_count,
-				       bool tmz)
+				       uint32_t copy_flags)
 {
 	ib->ptr[ib->length_dw++] = SDMA_PKT_HEADER_OP(SDMA_OP_COPY) |
 		SDMA_PKT_HEADER_SUB_OP(SDMA_SUBOP_COPY_LINEAR);
@@ -1277,8 +1252,7 @@ static void sdma_v2_4_set_vm_pte_funcs(struct amdgpu_device *adev)
 	adev->vm_manager.vm_pte_num_scheds = adev->sdma.num_instances;
 }
 
-const struct amdgpu_ip_block_version sdma_v2_4_ip_block =
-{
+const struct amdgpu_ip_block_version sdma_v2_4_ip_block = {
 	.type = AMD_IP_BLOCK_TYPE_SDMA,
 	.major = 2,
 	.minor = 4,
diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c
index 135727b59c41..1c076bd1cf73 100644
--- a/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c
@@ -250,10 +250,9 @@ static void sdma_v3_0_init_golden_registers(struct amdgpu_device *adev)
 static void sdma_v3_0_free_microcode(struct amdgpu_device *adev)
 {
 	int i;
-	for (i = 0; i < adev->sdma.num_instances; i++) {
-		release_firmware(adev->sdma.instance[i].fw);
-		adev->sdma.instance[i].fw = NULL;
-	}
+
+	for (i = 0; i < adev->sdma.num_instances; i++)
+		amdgpu_ucode_release(&adev->sdma.instance[i].fw);
 }
 
 /**
@@ -268,7 +267,6 @@ static void sdma_v3_0_free_microcode(struct amdgpu_device *adev)
 static int sdma_v3_0_init_microcode(struct amdgpu_device *adev)
 {
 	const char *chip_name;
-	char fw_name[30];
 	int err = 0, i;
 	struct amdgpu_firmware_info *info = NULL;
 	const struct common_firmware_header *header = NULL;
@@ -306,13 +304,13 @@ static int sdma_v3_0_init_microcode(struct amdgpu_device *adev)
 
 	for (i = 0; i < adev->sdma.num_instances; i++) {
 		if (i == 0)
-			snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_sdma.bin", chip_name);
+			err = amdgpu_ucode_request(adev, &adev->sdma.instance[i].fw,
+						   AMDGPU_UCODE_REQUIRED,
+						   "amdgpu/%s_sdma.bin", chip_name);
 		else
-			snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_sdma1.bin", chip_name);
-		err = request_firmware(&adev->sdma.instance[i].fw, fw_name, adev->dev);
-		if (err)
-			goto out;
-		err = amdgpu_ucode_validate(adev->sdma.instance[i].fw);
+			err = amdgpu_ucode_request(adev, &adev->sdma.instance[i].fw,
+						   AMDGPU_UCODE_REQUIRED,
+						   "amdgpu/%s_sdma1.bin", chip_name);
 		if (err)
 			goto out;
 		hdr = (const struct sdma_firmware_header_v1_0 *)adev->sdma.instance[i].fw->data;
@@ -331,11 +329,10 @@ static int sdma_v3_0_init_microcode(struct amdgpu_device *adev)
 	}
 out:
 	if (err) {
-		pr_err("sdma_v3_0: Failed to load firmware \"%s\"\n", fw_name);
-		for (i = 0; i < adev->sdma.num_instances; i++) {
-			release_firmware(adev->sdma.instance[i].fw);
-			adev->sdma.instance[i].fw = NULL;
-		}
+		pr_err("sdma_v3_0: Failed to load firmware \"%s_sdma%s.bin\"\n",
+		       chip_name, i == 0 ? "" : "1");
+		for (i = 0; i < adev->sdma.num_instances; i++)
+			amdgpu_ucode_release(&adev->sdma.instance[i].fw);
 	}
 	return err;
 }
@@ -350,7 +347,7 @@ out:
 static uint64_t sdma_v3_0_ring_get_rptr(struct amdgpu_ring *ring)
 {
 	/* XXX check if swapping is necessary on BE */
-	return ring->adev->wb.wb[ring->rptr_offs] >> 2;
+	return *ring->rptr_cpu_addr >> 2;
 }
 
 /**
@@ -367,7 +364,7 @@ static uint64_t sdma_v3_0_ring_get_wptr(struct amdgpu_ring *ring)
 
 	if (ring->use_doorbell || ring->use_pollmem) {
 		/* XXX check if swapping is necessary on BE */
-		wptr = ring->adev->wb.wb[ring->wptr_offs] >> 2;
+		wptr = *ring->wptr_cpu_addr >> 2;
 	} else {
 		wptr = RREG32(mmSDMA0_GFX_RB_WPTR + sdma_offsets[ring->me]) >> 2;
 	}
@@ -387,16 +384,16 @@ static void sdma_v3_0_ring_set_wptr(struct amdgpu_ring *ring)
 	struct amdgpu_device *adev = ring->adev;
 
 	if (ring->use_doorbell) {
-		u32 *wb = (u32 *)&adev->wb.wb[ring->wptr_offs];
+		u32 *wb = (u32 *)ring->wptr_cpu_addr;
 		/* XXX check if swapping is necessary on BE */
-		WRITE_ONCE(*wb, (lower_32_bits(ring->wptr) << 2));
-		WDOORBELL32(ring->doorbell_index, lower_32_bits(ring->wptr) << 2);
+		WRITE_ONCE(*wb, ring->wptr << 2);
+		WDOORBELL32(ring->doorbell_index, ring->wptr << 2);
 	} else if (ring->use_pollmem) {
-		u32 *wb = (u32 *)&adev->wb.wb[ring->wptr_offs];
+		u32 *wb = (u32 *)ring->wptr_cpu_addr;
 
-		WRITE_ONCE(*wb, (lower_32_bits(ring->wptr) << 2));
+		WRITE_ONCE(*wb, ring->wptr << 2);
 	} else {
-		WREG32(mmSDMA0_GFX_RB_WPTR + sdma_offsets[ring->me], lower_32_bits(ring->wptr) << 2);
+		WREG32(mmSDMA0_GFX_RB_WPTR + sdma_offsets[ring->me], ring->wptr << 2);
 	}
 }
 
@@ -516,15 +513,9 @@ static void sdma_v3_0_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 se
  */
 static void sdma_v3_0_gfx_stop(struct amdgpu_device *adev)
 {
-	struct amdgpu_ring *sdma0 = &adev->sdma.instance[0].ring;
-	struct amdgpu_ring *sdma1 = &adev->sdma.instance[1].ring;
 	u32 rb_cntl, ib_cntl;
 	int i;
 
-	if ((adev->mman.buffer_funcs_ring == sdma0) ||
-	    (adev->mman.buffer_funcs_ring == sdma1))
-		amdgpu_ttm_set_buffer_funcs_status(adev, false);
-
 	for (i = 0; i < adev->sdma.num_instances; i++) {
 		rb_cntl = RREG32(mmSDMA0_GFX_RB_CNTL + sdma_offsets[i]);
 		rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL, RB_ENABLE, 0);
@@ -649,7 +640,6 @@ static int sdma_v3_0_gfx_resume(struct amdgpu_device *adev)
 	struct amdgpu_ring *ring;
 	u32 rb_cntl, ib_cntl, wptr_poll_cntl;
 	u32 rb_bufsz;
-	u32 wb_offset;
 	u32 doorbell;
 	u64 wptr_gpu_addr;
 	int i, j, r;
@@ -657,7 +647,6 @@ static int sdma_v3_0_gfx_resume(struct amdgpu_device *adev)
 	for (i = 0; i < adev->sdma.num_instances; i++) {
 		ring = &adev->sdma.instance[i].ring;
 		amdgpu_ring_clear_ring(ring);
-		wb_offset = (ring->rptr_offs * 4);
 
 		mutex_lock(&adev->srbm_mutex);
 		for (j = 0; j < 16; j++) {
@@ -694,9 +683,9 @@ static int sdma_v3_0_gfx_resume(struct amdgpu_device *adev)
 
 		/* set the wb address whether it's enabled or not */
 		WREG32(mmSDMA0_GFX_RB_RPTR_ADDR_HI + sdma_offsets[i],
-		       upper_32_bits(adev->wb.gpu_addr + wb_offset) & 0xFFFFFFFF);
+		       upper_32_bits(ring->rptr_gpu_addr) & 0xFFFFFFFF);
 		WREG32(mmSDMA0_GFX_RB_RPTR_ADDR_LO + sdma_offsets[i],
-		       lower_32_bits(adev->wb.gpu_addr + wb_offset) & 0xFFFFFFFC);
+		       lower_32_bits(ring->rptr_gpu_addr) & 0xFFFFFFFC);
 
 		rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL, RPTR_WRITEBACK_ENABLE, 1);
 
@@ -715,7 +704,7 @@ static int sdma_v3_0_gfx_resume(struct amdgpu_device *adev)
 		WREG32(mmSDMA0_GFX_DOORBELL + sdma_offsets[i], doorbell);
 
 		/* setup the wptr shadow polling */
-		wptr_gpu_addr = adev->wb.gpu_addr + (ring->wptr_offs * 4);
+		wptr_gpu_addr = ring->wptr_gpu_addr;
 
 		WREG32(mmSDMA0_GFX_RB_WPTR_POLL_ADDR_LO + sdma_offsets[i],
 		       lower_32_bits(wptr_gpu_addr));
@@ -723,7 +712,7 @@ static int sdma_v3_0_gfx_resume(struct amdgpu_device *adev)
 		       upper_32_bits(wptr_gpu_addr));
 		wptr_poll_cntl = RREG32(mmSDMA0_GFX_RB_WPTR_POLL_CNTL + sdma_offsets[i]);
 		if (ring->use_pollmem) {
-			/*wptr polling is not enogh fast, directly clean the wptr register */
+			/*wptr polling is not enough fast, directly clean the wptr register */
 			WREG32(mmSDMA0_GFX_RB_WPTR + sdma_offsets[i], 0);
 			wptr_poll_cntl = REG_SET_FIELD(wptr_poll_cntl,
 						       SDMA0_GFX_RB_WPTR_POLL_CNTL,
@@ -746,8 +735,6 @@ static int sdma_v3_0_gfx_resume(struct amdgpu_device *adev)
 #endif
 		/* enable DMA IBs */
 		WREG32(mmSDMA0_GFX_IB_CNTL + sdma_offsets[i], ib_cntl);
-
-		ring->sched.ready = true;
 	}
 
 	/* unhalt the MEs */
@@ -760,9 +747,6 @@ static int sdma_v3_0_gfx_resume(struct amdgpu_device *adev)
 		r = amdgpu_ring_test_helper(ring);
 		if (r)
 			return r;
-
-		if (adev->mman.buffer_funcs_ring == ring)
-			amdgpu_ttm_set_buffer_funcs_status(adev, true);
 	}
 
 	return 0;
@@ -922,7 +906,7 @@ static int sdma_v3_0_ring_test_ib(struct amdgpu_ring *ring, long timeout)
 	else
 		r = -EINVAL;
 err1:
-	amdgpu_ib_free(adev, &ib, NULL);
+	amdgpu_ib_free(&ib, NULL);
 	dma_fence_put(f);
 err0:
 	amdgpu_device_wb_free(adev, index);
@@ -1098,9 +1082,10 @@ static void sdma_v3_0_ring_emit_wreg(struct amdgpu_ring *ring,
 	amdgpu_ring_write(ring, val);
 }
 
-static int sdma_v3_0_early_init(void *handle)
+static int sdma_v3_0_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
+	int r;
 
 	switch (adev->asic_type) {
 	case CHIP_STONEY:
@@ -1111,6 +1096,10 @@ static int sdma_v3_0_early_init(void *handle)
 		break;
 	}
 
+	r = sdma_v3_0_init_microcode(adev);
+	if (r)
+		return r;
+
 	sdma_v3_0_set_ring_funcs(adev);
 	sdma_v3_0_set_buffer_funcs(adev);
 	sdma_v3_0_set_vm_pte_funcs(adev);
@@ -1119,11 +1108,11 @@ static int sdma_v3_0_early_init(void *handle)
 	return 0;
 }
 
-static int sdma_v3_0_sw_init(void *handle)
+static int sdma_v3_0_sw_init(struct amdgpu_ip_block *ip_block)
 {
 	struct amdgpu_ring *ring;
 	int r, i;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	/* SDMA trap event */
 	r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_SDMA_TRAP,
@@ -1143,12 +1132,6 @@ static int sdma_v3_0_sw_init(void *handle)
 	if (r)
 		return r;
 
-	r = sdma_v3_0_init_microcode(adev);
-	if (r) {
-		DRM_ERROR("Failed to load sdma firmware!\n");
-		return r;
-	}
-
 	for (i = 0; i < adev->sdma.num_instances; i++) {
 		ring = &adev->sdma.instance[i].ring;
 		ring->ring_obj = NULL;
@@ -1171,9 +1154,9 @@ static int sdma_v3_0_sw_init(void *handle)
 	return r;
 }
 
-static int sdma_v3_0_sw_fini(void *handle)
+static int sdma_v3_0_sw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int i;
 
 	for (i = 0; i < adev->sdma.num_instances; i++)
@@ -1183,10 +1166,10 @@ static int sdma_v3_0_sw_fini(void *handle)
 	return 0;
 }
 
-static int sdma_v3_0_hw_init(void *handle)
+static int sdma_v3_0_hw_init(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	sdma_v3_0_init_golden_registers(adev);
 
@@ -1197,9 +1180,9 @@ static int sdma_v3_0_hw_init(void *handle)
 	return r;
 }
 
-static int sdma_v3_0_hw_fini(void *handle)
+static int sdma_v3_0_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	sdma_v3_0_ctx_switch_enable(adev, false);
 	sdma_v3_0_enable(adev, false);
@@ -1207,23 +1190,19 @@ static int sdma_v3_0_hw_fini(void *handle)
 	return 0;
 }
 
-static int sdma_v3_0_suspend(void *handle)
+static int sdma_v3_0_suspend(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	return sdma_v3_0_hw_fini(adev);
+	return sdma_v3_0_hw_fini(ip_block);
 }
 
-static int sdma_v3_0_resume(void *handle)
+static int sdma_v3_0_resume(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	return sdma_v3_0_hw_init(adev);
+	return sdma_v3_0_hw_init(ip_block);
 }
 
-static bool sdma_v3_0_is_idle(void *handle)
+static bool sdma_v3_0_is_idle(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	u32 tmp = RREG32(mmSRBM_STATUS2);
 
 	if (tmp & (SRBM_STATUS2__SDMA_BUSY_MASK |
@@ -1233,11 +1212,11 @@ static bool sdma_v3_0_is_idle(void *handle)
 	return true;
 }
 
-static int sdma_v3_0_wait_for_idle(void *handle)
+static int sdma_v3_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
 {
 	unsigned i;
 	u32 tmp;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	for (i = 0; i < adev->usec_timeout; i++) {
 		tmp = RREG32(mmSRBM_STATUS2) & (SRBM_STATUS2__SDMA_BUSY_MASK |
@@ -1250,9 +1229,9 @@ static int sdma_v3_0_wait_for_idle(void *handle)
 	return -ETIMEDOUT;
 }
 
-static bool sdma_v3_0_check_soft_reset(void *handle)
+static bool sdma_v3_0_check_soft_reset(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	u32 srbm_soft_reset = 0;
 	u32 tmp = RREG32(mmSRBM_STATUS2);
 
@@ -1271,9 +1250,9 @@ static bool sdma_v3_0_check_soft_reset(void *handle)
 	}
 }
 
-static int sdma_v3_0_pre_soft_reset(void *handle)
+static int sdma_v3_0_pre_soft_reset(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	u32 srbm_soft_reset = 0;
 
 	if (!adev->sdma.srbm_soft_reset)
@@ -1290,9 +1269,9 @@ static int sdma_v3_0_pre_soft_reset(void *handle)
 	return 0;
 }
 
-static int sdma_v3_0_post_soft_reset(void *handle)
+static int sdma_v3_0_post_soft_reset(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	u32 srbm_soft_reset = 0;
 
 	if (!adev->sdma.srbm_soft_reset)
@@ -1309,9 +1288,9 @@ static int sdma_v3_0_post_soft_reset(void *handle)
 	return 0;
 }
 
-static int sdma_v3_0_soft_reset(void *handle)
+static int sdma_v3_0_soft_reset(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	u32 srbm_soft_reset = 0;
 	u32 tmp;
 
@@ -1506,10 +1485,10 @@ static void sdma_v3_0_update_sdma_medium_grain_light_sleep(
 	}
 }
 
-static int sdma_v3_0_set_clockgating_state(void *handle,
+static int sdma_v3_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_clockgating_state state)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (amdgpu_sriov_vf(adev))
 		return 0;
@@ -1529,15 +1508,15 @@ static int sdma_v3_0_set_clockgating_state(void *handle,
 	return 0;
 }
 
-static int sdma_v3_0_set_powergating_state(void *handle,
+static int sdma_v3_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_powergating_state state)
 {
 	return 0;
 }
 
-static void sdma_v3_0_get_clockgating_state(void *handle, u32 *flags)
+static void sdma_v3_0_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int data;
 
 	if (amdgpu_sriov_vf(adev))
@@ -1557,7 +1536,6 @@ static void sdma_v3_0_get_clockgating_state(void *handle, u32 *flags)
 static const struct amd_ip_funcs sdma_v3_0_ip_funcs = {
 	.name = "sdma_v3_0",
 	.early_init = sdma_v3_0_early_init,
-	.late_init = NULL,
 	.sw_init = sdma_v3_0_sw_init,
 	.sw_fini = sdma_v3_0_sw_fini,
 	.hw_init = sdma_v3_0_hw_init,
@@ -1580,6 +1558,7 @@ static const struct amdgpu_ring_funcs sdma_v3_0_ring_funcs = {
 	.align_mask = 0xf,
 	.nop = SDMA_PKT_NOP_HEADER_OP(SDMA_OP_NOP),
 	.support_64bit_ptrs = false,
+	.secure_submission_supported = true,
 	.get_rptr = sdma_v3_0_ring_get_rptr,
 	.get_wptr = sdma_v3_0_ring_get_wptr,
 	.set_wptr = sdma_v3_0_ring_set_wptr,
@@ -1635,7 +1614,7 @@ static void sdma_v3_0_set_irq_funcs(struct amdgpu_device *adev)
  * @src_offset: src GPU address
  * @dst_offset: dst GPU address
  * @byte_count: number of bytes to xfer
- * @tmz: unused
+ * @copy_flags: unused
  *
  * Copy GPU buffers using the DMA engine (VI).
  * Used by the amdgpu ttm implementation to move pages if
@@ -1645,7 +1624,7 @@ static void sdma_v3_0_emit_copy_buffer(struct amdgpu_ib *ib,
 				       uint64_t src_offset,
 				       uint64_t dst_offset,
 				       uint32_t byte_count,
-				       bool tmz)
+				       uint32_t copy_flags)
 {
 	ib->ptr[ib->length_dw++] = SDMA_PKT_HEADER_OP(SDMA_OP_COPY) |
 		SDMA_PKT_HEADER_SUB_OP(SDMA_SUBOP_COPY_LINEAR);
diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c
index e8e4749e9c79..f38004e6064e 100644
--- a/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c
@@ -72,6 +72,53 @@ MODULE_FIRMWARE("amdgpu/renoir_sdma.bin");
 MODULE_FIRMWARE("amdgpu/green_sardine_sdma.bin");
 MODULE_FIRMWARE("amdgpu/aldebaran_sdma.bin");
 
+static const struct amdgpu_hwip_reg_entry sdma_reg_list_4_0[] = {
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_STATUS_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_STATUS1_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_STATUS2_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_STATUS3_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_UCODE_CHECKSUM),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RB_RPTR_FETCH_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RB_RPTR_FETCH),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_UTCL1_RD_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_UTCL1_WR_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_UTCL1_RD_XNACK0),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_UTCL1_RD_XNACK1),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_UTCL1_WR_XNACK0),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_UTCL1_WR_XNACK1),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_RB_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_RB_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_RB_RPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_RB_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_RB_WPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_IB_OFFSET),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_IB_BASE_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_IB_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_IB_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_IB_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_IB_SUB_REMAIN),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_DUMMY_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_PAGE_RB_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_PAGE_RB_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_PAGE_RB_RPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_PAGE_RB_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_PAGE_RB_WPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_PAGE_IB_OFFSET),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_PAGE_IB_BASE_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_PAGE_IB_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_PAGE_DUMMY_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RLC0_RB_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RLC0_RB_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RLC0_RB_RPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RLC0_RB_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RLC0_RB_WPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RLC0_IB_OFFSET),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RLC0_IB_BASE_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RLC0_IB_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RLC0_DUMMY_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_VM_CNTL)
+};
+
 #define SDMA0_POWER_CNTL__ON_OFF_CONDITION_HOLD_TIME_MASK  0x000000F8L
 #define SDMA0_POWER_CNTL__ON_OFF_STATUS_DURATION_TIME_MASK 0xFC000000L
 
@@ -469,7 +516,7 @@ static int sdma_v4_0_irq_id_to_seq(unsigned client_id)
 
 static void sdma_v4_0_init_golden_registers(struct amdgpu_device *adev)
 {
-	switch (adev->ip_versions[SDMA0_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, SDMA0_HWIP, 0)) {
 	case IP_VERSION(4, 0, 0):
 		soc15_program_register_sequence(adev,
 						golden_settings_sdma_4,
@@ -539,7 +586,7 @@ static void sdma_v4_0_setup_ulv(struct amdgpu_device *adev)
 	 * The only chips with SDMAv4 and ULV are VG10 and VG20.
 	 * Server SKUs take a different hysteresis setting from other SKUs.
 	 */
-	switch (adev->ip_versions[SDMA0_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, SDMA0_HWIP, 0)) {
 	case IP_VERSION(4, 0, 0):
 		if (adev->pdev->device == 0x6860)
 			break;
@@ -561,44 +608,6 @@ static void sdma_v4_0_setup_ulv(struct amdgpu_device *adev)
 	}
 }
 
-static int sdma_v4_0_init_inst_ctx(struct amdgpu_sdma_instance *sdma_inst)
-{
-	int err = 0;
-	const struct sdma_firmware_header_v1_0 *hdr;
-
-	err = amdgpu_ucode_validate(sdma_inst->fw);
-	if (err)
-		return err;
-
-	hdr = (const struct sdma_firmware_header_v1_0 *)sdma_inst->fw->data;
-	sdma_inst->fw_version = le32_to_cpu(hdr->header.ucode_version);
-	sdma_inst->feature_version = le32_to_cpu(hdr->ucode_feature_version);
-
-	if (sdma_inst->feature_version >= 20)
-		sdma_inst->burst_nop = true;
-
-	return 0;
-}
-
-static void sdma_v4_0_destroy_inst_ctx(struct amdgpu_device *adev)
-{
-	int i;
-
-	for (i = 0; i < adev->sdma.num_instances; i++) {
-		release_firmware(adev->sdma.instance[i].fw);
-		adev->sdma.instance[i].fw = NULL;
-
-		/* arcturus shares the same FW memory across
-		   all SDMA isntances */
-		if (adev->ip_versions[SDMA0_HWIP][0] == IP_VERSION(4, 2, 2) ||
-		    adev->ip_versions[SDMA0_HWIP][0] == IP_VERSION(4, 4, 0))
-			break;
-	}
-
-	memset((void *)adev->sdma.instance, 0,
-		sizeof(struct amdgpu_sdma_instance) * AMDGPU_MAX_SDMA_INSTANCES);
-}
-
 /**
  * sdma_v4_0_init_microcode - load ucode images from disk
  *
@@ -613,101 +622,25 @@ static void sdma_v4_0_destroy_inst_ctx(struct amdgpu_device *adev)
 // vega10 real chip need to use PSP to load firmware
 static int sdma_v4_0_init_microcode(struct amdgpu_device *adev)
 {
-	const char *chip_name;
-	char fw_name[30];
-	int err = 0, i;
-	struct amdgpu_firmware_info *info = NULL;
-	const struct common_firmware_header *header = NULL;
-
-	DRM_DEBUG("\n");
+	int ret, i;
 
-	switch (adev->ip_versions[SDMA0_HWIP][0]) {
-	case IP_VERSION(4, 0, 0):
-		chip_name = "vega10";
-		break;
-	case IP_VERSION(4, 0, 1):
-		chip_name = "vega12";
-		break;
-	case IP_VERSION(4, 2, 0):
-		chip_name = "vega20";
-		break;
-	case IP_VERSION(4, 1, 0):
-	case IP_VERSION(4, 1, 1):
-		if (adev->apu_flags & AMD_APU_IS_RAVEN2)
-			chip_name = "raven2";
-		else if (adev->apu_flags & AMD_APU_IS_PICASSO)
-			chip_name = "picasso";
-		else
-			chip_name = "raven";
-		break;
-	case IP_VERSION(4, 2, 2):
-		chip_name = "arcturus";
-		break;
-	case IP_VERSION(4, 1, 2):
-		if (adev->apu_flags & AMD_APU_IS_RENOIR)
-			chip_name = "renoir";
-		else
-			chip_name = "green_sardine";
-		break;
-	case IP_VERSION(4, 4, 0):
-		chip_name = "aldebaran";
-		break;
-	default:
-		BUG();
-	}
-
-	snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_sdma.bin", chip_name);
-
-	err = request_firmware(&adev->sdma.instance[0].fw, fw_name, adev->dev);
-	if (err)
-		goto out;
-
-	err = sdma_v4_0_init_inst_ctx(&adev->sdma.instance[0]);
-	if (err)
-		goto out;
-
-	for (i = 1; i < adev->sdma.num_instances; i++) {
-		if (adev->ip_versions[SDMA0_HWIP][0] == IP_VERSION(4, 2, 2) ||
-                    adev->ip_versions[SDMA0_HWIP][0] == IP_VERSION(4, 4, 0)) {
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		if (amdgpu_ip_version(adev, SDMA0_HWIP, 0) ==
+			    IP_VERSION(4, 2, 2) ||
+		    amdgpu_ip_version(adev, SDMA0_HWIP, 0) ==
+			    IP_VERSION(4, 4, 0)) {
 			/* Acturus & Aldebaran will leverage the same FW memory
 			   for every SDMA instance */
-			memcpy((void *)&adev->sdma.instance[i],
-			       (void *)&adev->sdma.instance[0],
-			       sizeof(struct amdgpu_sdma_instance));
-		}
-		else {
-			snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_sdma%d.bin", chip_name, i);
-
-			err = request_firmware(&adev->sdma.instance[i].fw, fw_name, adev->dev);
-			if (err)
-				goto out;
-
-			err = sdma_v4_0_init_inst_ctx(&adev->sdma.instance[i]);
-			if (err)
-				goto out;
-		}
-	}
-
-	DRM_DEBUG("psp_load == '%s'\n",
-		adev->firmware.load_type == AMDGPU_FW_LOAD_PSP ? "true" : "false");
-
-	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
-		for (i = 0; i < adev->sdma.num_instances; i++) {
-			info = &adev->firmware.ucode[AMDGPU_UCODE_ID_SDMA0 + i];
-			info->ucode_id = AMDGPU_UCODE_ID_SDMA0 + i;
-			info->fw = adev->sdma.instance[i].fw;
-			header = (const struct common_firmware_header *)info->fw->data;
-			adev->firmware.fw_size +=
-				ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE);
+			ret = amdgpu_sdma_init_microcode(adev, 0, true);
+			break;
+		} else {
+			ret = amdgpu_sdma_init_microcode(adev, i, false);
+			if (ret)
+				return ret;
 		}
 	}
 
-out:
-	if (err) {
-		DRM_ERROR("sdma_v4_0: Failed to load firmware \"%s\"\n", fw_name);
-		sdma_v4_0_destroy_inst_ctx(adev);
-	}
-	return err;
+	return ret;
 }
 
 /**
@@ -722,7 +655,7 @@ static uint64_t sdma_v4_0_ring_get_rptr(struct amdgpu_ring *ring)
 	u64 *rptr;
 
 	/* XXX check if swapping is necessary on BE */
-	rptr = ((u64 *)&ring->adev->wb.wb[ring->rptr_offs]);
+	rptr = ((u64 *)ring->rptr_cpu_addr);
 
 	DRM_DEBUG("rptr before shift == 0x%016llx\n", *rptr);
 	return ((*rptr) >> 2);
@@ -742,7 +675,7 @@ static uint64_t sdma_v4_0_ring_get_wptr(struct amdgpu_ring *ring)
 
 	if (ring->use_doorbell) {
 		/* XXX check if swapping is necessary on BE */
-		wptr = READ_ONCE(*((u64 *)&adev->wb.wb[ring->wptr_offs]));
+		wptr = READ_ONCE(*((u64 *)ring->wptr_cpu_addr));
 		DRM_DEBUG("wptr/doorbell before shift == 0x%016llx\n", wptr);
 	} else {
 		wptr = RREG32_SDMA(ring->me, mmSDMA0_GFX_RB_WPTR_HI);
@@ -768,12 +701,12 @@ static void sdma_v4_0_ring_set_wptr(struct amdgpu_ring *ring)
 
 	DRM_DEBUG("Setting write pointer\n");
 	if (ring->use_doorbell) {
-		u64 *wb = (u64 *)&adev->wb.wb[ring->wptr_offs];
+		u64 *wb = (u64 *)ring->wptr_cpu_addr;
 
 		DRM_DEBUG("Using doorbell -- "
 				"wptr_offs == 0x%08x "
-				"lower_32_bits(ring->wptr) << 2 == 0x%08x "
-				"upper_32_bits(ring->wptr) << 2 == 0x%08x\n",
+				"lower_32_bits(ring->wptr << 2) == 0x%08x "
+				"upper_32_bits(ring->wptr << 2) == 0x%08x\n",
 				ring->wptr_offs,
 				lower_32_bits(ring->wptr << 2),
 				upper_32_bits(ring->wptr << 2));
@@ -811,7 +744,7 @@ static uint64_t sdma_v4_0_page_ring_get_wptr(struct amdgpu_ring *ring)
 
 	if (ring->use_doorbell) {
 		/* XXX check if swapping is necessary on BE */
-		wptr = READ_ONCE(*((u64 *)&adev->wb.wb[ring->wptr_offs]));
+		wptr = READ_ONCE(*((u64 *)ring->wptr_cpu_addr));
 	} else {
 		wptr = RREG32_SDMA(ring->me, mmSDMA0_PAGE_RB_WPTR_HI);
 		wptr = wptr << 32;
@@ -833,7 +766,7 @@ static void sdma_v4_0_page_ring_set_wptr(struct amdgpu_ring *ring)
 	struct amdgpu_device *adev = ring->adev;
 
 	if (ring->use_doorbell) {
-		u64 *wb = (u64 *)&adev->wb.wb[ring->wptr_offs];
+		u64 *wb = (u64 *)ring->wptr_cpu_addr;
 
 		/* XXX check if swapping is necessary on BE */
 		WRITE_ONCE(*wb, (ring->wptr << 2));
@@ -980,31 +913,23 @@ static void sdma_v4_0_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 se
 
 
 /**
- * sdma_v4_0_gfx_stop - stop the gfx async dma engines
+ * sdma_v4_0_gfx_enable - enable the gfx async dma engines
  *
  * @adev: amdgpu_device pointer
- *
- * Stop the gfx async dma ring buffers (VEGA10).
+ * @enable: enable SDMA RB/IB
+ * control the gfx async dma ring buffers (VEGA10).
  */
-static void sdma_v4_0_gfx_stop(struct amdgpu_device *adev)
+static void sdma_v4_0_gfx_enable(struct amdgpu_device *adev, bool enable)
 {
-	struct amdgpu_ring *sdma[AMDGPU_MAX_SDMA_INSTANCES];
 	u32 rb_cntl, ib_cntl;
-	int i, unset = 0;
+	int i;
 
 	for (i = 0; i < adev->sdma.num_instances; i++) {
-		sdma[i] = &adev->sdma.instance[i].ring;
-
-		if ((adev->mman.buffer_funcs_ring == sdma[i]) && unset != 1) {
-			amdgpu_ttm_set_buffer_funcs_status(adev, false);
-			unset = 1;
-		}
-
 		rb_cntl = RREG32_SDMA(i, mmSDMA0_GFX_RB_CNTL);
-		rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL, RB_ENABLE, 0);
+		rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL, RB_ENABLE, enable ? 1 : 0);
 		WREG32_SDMA(i, mmSDMA0_GFX_RB_CNTL, rb_cntl);
 		ib_cntl = RREG32_SDMA(i, mmSDMA0_GFX_IB_CNTL);
-		ib_cntl = REG_SET_FIELD(ib_cntl, SDMA0_GFX_IB_CNTL, IB_ENABLE, 0);
+		ib_cntl = REG_SET_FIELD(ib_cntl, SDMA0_GFX_IB_CNTL, IB_ENABLE, enable ? 1 : 0);
 		WREG32_SDMA(i, mmSDMA0_GFX_IB_CNTL, ib_cntl);
 	}
 }
@@ -1030,20 +955,10 @@ static void sdma_v4_0_rlc_stop(struct amdgpu_device *adev)
  */
 static void sdma_v4_0_page_stop(struct amdgpu_device *adev)
 {
-	struct amdgpu_ring *sdma[AMDGPU_MAX_SDMA_INSTANCES];
 	u32 rb_cntl, ib_cntl;
 	int i;
-	bool unset = false;
 
 	for (i = 0; i < adev->sdma.num_instances; i++) {
-		sdma[i] = &adev->sdma.instance[i].page;
-
-		if ((adev->mman.buffer_funcs_ring == sdma[i]) &&
-			(!unset)) {
-			amdgpu_ttm_set_buffer_funcs_status(adev, false);
-			unset = true;
-		}
-
 		rb_cntl = RREG32_SDMA(i, mmSDMA0_PAGE_RB_CNTL);
 		rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_PAGE_RB_CNTL,
 					RB_ENABLE, 0);
@@ -1108,7 +1023,8 @@ static void sdma_v4_0_ctx_switch_enable(struct amdgpu_device *adev, bool enable)
 		 * Arcturus for the moment and firmware version 14
 		 * and above.
 		 */
-		if (adev->ip_versions[SDMA0_HWIP][0] == IP_VERSION(4, 2, 2) &&
+		if (amdgpu_ip_version(adev, SDMA0_HWIP, 0) ==
+			    IP_VERSION(4, 2, 2) &&
 		    adev->sdma.instance[i].fw_version >= 14)
 			WREG32_SDMA(i, mmSDMA0_PUB_DUMMY_REG2, enable);
 		/* Extend page fault timeout to avoid interrupt storm */
@@ -1131,7 +1047,7 @@ static void sdma_v4_0_enable(struct amdgpu_device *adev, bool enable)
 	int i;
 
 	if (!enable) {
-		sdma_v4_0_gfx_stop(adev);
+		sdma_v4_0_gfx_enable(adev, enable);
 		sdma_v4_0_rlc_stop(adev);
 		if (adev->sdma.has_page_queue)
 			sdma_v4_0_page_stop(adev);
@@ -1174,13 +1090,10 @@ static void sdma_v4_0_gfx_resume(struct amdgpu_device *adev, unsigned int i)
 {
 	struct amdgpu_ring *ring = &adev->sdma.instance[i].ring;
 	u32 rb_cntl, ib_cntl, wptr_poll_cntl;
-	u32 wb_offset;
 	u32 doorbell;
 	u32 doorbell_offset;
 	u64 wptr_gpu_addr;
 
-	wb_offset = (ring->rptr_offs * 4);
-
 	rb_cntl = RREG32_SDMA(i, mmSDMA0_GFX_RB_CNTL);
 	rb_cntl = sdma_v4_0_rb_cntl(ring, rb_cntl);
 	WREG32_SDMA(i, mmSDMA0_GFX_RB_CNTL, rb_cntl);
@@ -1193,9 +1106,9 @@ static void sdma_v4_0_gfx_resume(struct amdgpu_device *adev, unsigned int i)
 
 	/* set the wb address whether it's enabled or not */
 	WREG32_SDMA(i, mmSDMA0_GFX_RB_RPTR_ADDR_HI,
-	       upper_32_bits(adev->wb.gpu_addr + wb_offset) & 0xFFFFFFFF);
+	       upper_32_bits(ring->rptr_gpu_addr) & 0xFFFFFFFF);
 	WREG32_SDMA(i, mmSDMA0_GFX_RB_RPTR_ADDR_LO,
-	       lower_32_bits(adev->wb.gpu_addr + wb_offset) & 0xFFFFFFFC);
+	       lower_32_bits(ring->rptr_gpu_addr) & 0xFFFFFFFC);
 
 	rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL,
 				RPTR_WRITEBACK_ENABLE, 1);
@@ -1225,7 +1138,7 @@ static void sdma_v4_0_gfx_resume(struct amdgpu_device *adev, unsigned int i)
 	WREG32_SDMA(i, mmSDMA0_GFX_MINOR_PTR_UPDATE, 0);
 
 	/* setup the wptr shadow polling */
-	wptr_gpu_addr = adev->wb.gpu_addr + (ring->wptr_offs * 4);
+	wptr_gpu_addr = ring->wptr_gpu_addr;
 	WREG32_SDMA(i, mmSDMA0_GFX_RB_WPTR_POLL_ADDR_LO,
 		    lower_32_bits(wptr_gpu_addr));
 	WREG32_SDMA(i, mmSDMA0_GFX_RB_WPTR_POLL_ADDR_HI,
@@ -1247,8 +1160,6 @@ static void sdma_v4_0_gfx_resume(struct amdgpu_device *adev, unsigned int i)
 #endif
 	/* enable DMA IBs */
 	WREG32_SDMA(i, mmSDMA0_GFX_IB_CNTL, ib_cntl);
-
-	ring->sched.ready = true;
 }
 
 /**
@@ -1264,13 +1175,10 @@ static void sdma_v4_0_page_resume(struct amdgpu_device *adev, unsigned int i)
 {
 	struct amdgpu_ring *ring = &adev->sdma.instance[i].page;
 	u32 rb_cntl, ib_cntl, wptr_poll_cntl;
-	u32 wb_offset;
 	u32 doorbell;
 	u32 doorbell_offset;
 	u64 wptr_gpu_addr;
 
-	wb_offset = (ring->rptr_offs * 4);
-
 	rb_cntl = RREG32_SDMA(i, mmSDMA0_PAGE_RB_CNTL);
 	rb_cntl = sdma_v4_0_rb_cntl(ring, rb_cntl);
 	WREG32_SDMA(i, mmSDMA0_PAGE_RB_CNTL, rb_cntl);
@@ -1283,9 +1191,9 @@ static void sdma_v4_0_page_resume(struct amdgpu_device *adev, unsigned int i)
 
 	/* set the wb address whether it's enabled or not */
 	WREG32_SDMA(i, mmSDMA0_PAGE_RB_RPTR_ADDR_HI,
-	       upper_32_bits(adev->wb.gpu_addr + wb_offset) & 0xFFFFFFFF);
+	       upper_32_bits(ring->rptr_gpu_addr) & 0xFFFFFFFF);
 	WREG32_SDMA(i, mmSDMA0_PAGE_RB_RPTR_ADDR_LO,
-	       lower_32_bits(adev->wb.gpu_addr + wb_offset) & 0xFFFFFFFC);
+	       lower_32_bits(ring->rptr_gpu_addr) & 0xFFFFFFFC);
 
 	rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_PAGE_RB_CNTL,
 				RPTR_WRITEBACK_ENABLE, 1);
@@ -1316,7 +1224,7 @@ static void sdma_v4_0_page_resume(struct amdgpu_device *adev, unsigned int i)
 	WREG32_SDMA(i, mmSDMA0_PAGE_MINOR_PTR_UPDATE, 0);
 
 	/* setup the wptr shadow polling */
-	wptr_gpu_addr = adev->wb.gpu_addr + (ring->wptr_offs * 4);
+	wptr_gpu_addr = ring->wptr_gpu_addr;
 	WREG32_SDMA(i, mmSDMA0_PAGE_RB_WPTR_POLL_ADDR_LO,
 		    lower_32_bits(wptr_gpu_addr));
 	WREG32_SDMA(i, mmSDMA0_PAGE_RB_WPTR_POLL_ADDR_HI,
@@ -1338,8 +1246,6 @@ static void sdma_v4_0_page_resume(struct amdgpu_device *adev, unsigned int i)
 #endif
 	/* enable DMA IBs */
 	WREG32_SDMA(i, mmSDMA0_PAGE_IB_CNTL, ib_cntl);
-
-	ring->sched.ready = true;
 }
 
 static void
@@ -1395,7 +1301,7 @@ static void sdma_v4_0_init_pg(struct amdgpu_device *adev)
 	if (!(adev->pg_flags & AMD_PG_SUPPORT_SDMA))
 		return;
 
-	switch (adev->ip_versions[SDMA0_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, SDMA0_HWIP, 0)) {
 	case IP_VERSION(4, 1, 0):
         case IP_VERSION(4, 1, 1):
 	case IP_VERSION(4, 1, 2):
@@ -1539,13 +1445,7 @@ static int sdma_v4_0_start(struct amdgpu_device *adev)
 			r = amdgpu_ring_test_helper(page);
 			if (r)
 				return r;
-
-			if (adev->mman.buffer_funcs_ring == page)
-				amdgpu_ttm_set_buffer_funcs_status(adev, true);
 		}
-
-		if (adev->mman.buffer_funcs_ring == ring)
-			amdgpu_ttm_set_buffer_funcs_status(adev, true);
 	}
 
 	return r;
@@ -1665,7 +1565,7 @@ static int sdma_v4_0_ring_test_ib(struct amdgpu_ring *ring, long timeout)
 		r = -EINVAL;
 
 err1:
-	amdgpu_ib_free(adev, &ib, NULL);
+	amdgpu_ib_free(&ib, NULL);
 	dma_fence_put(f);
 err0:
 	amdgpu_device_wb_free(adev, index);
@@ -1838,7 +1738,7 @@ static bool sdma_v4_0_fw_support_paging_queue(struct amdgpu_device *adev)
 {
 	uint fw_version = adev->sdma.instance[0].fw_version;
 
-	switch (adev->ip_versions[SDMA0_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, SDMA0_HWIP, 0)) {
 	case IP_VERSION(4, 0, 0):
 		return fw_version >= 430;
 	case IP_VERSION(4, 0, 1):
@@ -1851,19 +1751,17 @@ static bool sdma_v4_0_fw_support_paging_queue(struct amdgpu_device *adev)
 	}
 }
 
-static int sdma_v4_0_early_init(void *handle)
+static int sdma_v4_0_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int r;
 
 	r = sdma_v4_0_init_microcode(adev);
-	if (r) {
-		DRM_ERROR("Failed to load sdma firmware!\n");
+	if (r)
 		return r;
-	}
 
 	/* TODO: Page queue breaks driver reload under SRIOV */
-	if ((adev->ip_versions[SDMA0_HWIP][0] == IP_VERSION(4, 0, 0)) &&
+	if ((amdgpu_ip_version(adev, SDMA0_HWIP, 0) == IP_VERSION(4, 0, 0)) &&
 	    amdgpu_sriov_vf((adev)))
 		adev->sdma.has_page_queue = false;
 	else if (sdma_v4_0_fw_support_paging_queue(adev))
@@ -1882,32 +1780,25 @@ static int sdma_v4_0_process_ras_data_cb(struct amdgpu_device *adev,
 		void *err_data,
 		struct amdgpu_iv_entry *entry);
 
-static int sdma_v4_0_late_init(void *handle)
+static int sdma_v4_0_late_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-	struct ras_ih_if ih_info = {
-		.cb = sdma_v4_0_process_ras_data_cb,
-	};
+	struct amdgpu_device *adev = ip_block->adev;
 
 	sdma_v4_0_setup_ulv(adev);
 
-	if (!amdgpu_persistent_edc_harvesting_supported(adev)) {
-		if (adev->sdma.funcs &&
-		    adev->sdma.funcs->reset_ras_error_count)
-			adev->sdma.funcs->reset_ras_error_count(adev);
-	}
+	if (!amdgpu_persistent_edc_harvesting_supported(adev))
+		amdgpu_ras_reset_error_count(adev, AMDGPU_RAS_BLOCK__SDMA);
 
-	if (adev->sdma.funcs && adev->sdma.funcs->ras_late_init)
-		return adev->sdma.funcs->ras_late_init(adev, &ih_info);
-	else
-		return 0;
+	return 0;
 }
 
-static int sdma_v4_0_sw_init(void *handle)
+static int sdma_v4_0_sw_init(struct amdgpu_ip_block *ip_block)
 {
 	struct amdgpu_ring *ring;
 	int r, i;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
+	uint32_t reg_count = ARRAY_SIZE(sdma_reg_list_4_0);
+	uint32_t *ptr;
 
 	/* SDMA trap event */
 	for (i = 0; i < adev->sdma.num_instances; i++) {
@@ -1965,6 +1856,17 @@ static int sdma_v4_0_sw_init(void *handle)
 		/* doorbell size is 2 dwords, get DWORD offset */
 		ring->doorbell_index = adev->doorbell_index.sdma_engine[i] << 1;
 
+		/*
+		 * On Arcturus, SDMA instance 5~7 has a different vmhub
+		 * type(AMDGPU_MMHUB1).
+		 */
+		if (amdgpu_ip_version(adev, SDMA0_HWIP, 0) ==
+			    IP_VERSION(4, 2, 2) &&
+		    i >= 5)
+			ring->vm_hub = AMDGPU_MMHUB1(0);
+		else
+			ring->vm_hub = AMDGPU_MMHUB0(0);
+
 		sprintf(ring->name, "sdma%d", i);
 		r = amdgpu_ring_init(adev, ring, 1024, &adev->sdma.trap_irq,
 				     AMDGPU_SDMA_IRQ_INSTANCE0 + i,
@@ -1980,8 +1882,27 @@ static int sdma_v4_0_sw_init(void *handle)
 			/* paging queue use same doorbell index/routing as gfx queue
 			 * with 0x400 (4096 dwords) offset on second doorbell page
 			 */
-			ring->doorbell_index = adev->doorbell_index.sdma_engine[i] << 1;
-			ring->doorbell_index += 0x400;
+			if (amdgpu_ip_version(adev, SDMA0_HWIP, 0) >=
+				    IP_VERSION(4, 0, 0) &&
+			    amdgpu_ip_version(adev, SDMA0_HWIP, 0) <
+				    IP_VERSION(4, 2, 0)) {
+				ring->doorbell_index =
+					adev->doorbell_index.sdma_engine[i] << 1;
+				ring->doorbell_index += 0x400;
+			} else {
+				/* From vega20, the sdma_doorbell_range in 1st
+				 * doorbell page is reserved for page queue.
+				 */
+				ring->doorbell_index =
+					(adev->doorbell_index.sdma_engine[i] + 1) << 1;
+			}
+
+			if (amdgpu_ip_version(adev, SDMA0_HWIP, 0) ==
+				    IP_VERSION(4, 2, 2) &&
+			    i >= 5)
+				ring->vm_hub = AMDGPU_MMHUB1(0);
+			else
+				ring->vm_hub = AMDGPU_MMHUB0(0);
 
 			sprintf(ring->name, "page%d", i);
 			r = amdgpu_ring_init(adev, ring, 1024,
@@ -1993,83 +1914,110 @@ static int sdma_v4_0_sw_init(void *handle)
 		}
 	}
 
+	if (amdgpu_sdma_ras_sw_init(adev)) {
+		dev_err(adev->dev, "Failed to initialize sdma ras block!\n");
+		return -EINVAL;
+	}
+
+	/* Allocate memory for SDMA IP Dump buffer */
+	ptr = kcalloc(adev->sdma.num_instances * reg_count, sizeof(uint32_t), GFP_KERNEL);
+	if (ptr)
+		adev->sdma.ip_dump = ptr;
+	else
+		DRM_ERROR("Failed to allocated memory for SDMA IP Dump\n");
+
 	return r;
 }
 
-static int sdma_v4_0_sw_fini(void *handle)
+static int sdma_v4_0_sw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int i;
 
-	if (adev->sdma.funcs && adev->sdma.funcs->ras_fini)
-		adev->sdma.funcs->ras_fini(adev);
-
 	for (i = 0; i < adev->sdma.num_instances; i++) {
 		amdgpu_ring_fini(&adev->sdma.instance[i].ring);
 		if (adev->sdma.has_page_queue)
 			amdgpu_ring_fini(&adev->sdma.instance[i].page);
 	}
 
-	sdma_v4_0_destroy_inst_ctx(adev);
+	if (amdgpu_ip_version(adev, SDMA0_HWIP, 0) == IP_VERSION(4, 2, 2) ||
+	    amdgpu_ip_version(adev, SDMA0_HWIP, 0) == IP_VERSION(4, 4, 0))
+		amdgpu_sdma_destroy_inst_ctx(adev, true);
+	else
+		amdgpu_sdma_destroy_inst_ctx(adev, false);
+
+	kfree(adev->sdma.ip_dump);
 
 	return 0;
 }
 
-static int sdma_v4_0_hw_init(void *handle)
+static int sdma_v4_0_hw_init(struct amdgpu_ip_block *ip_block)
 {
-	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (adev->flags & AMD_IS_APU)
-		amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_SDMA, false);
+		amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_SDMA, false, 0);
 
 	if (!amdgpu_sriov_vf(adev))
 		sdma_v4_0_init_golden_registers(adev);
 
-	r = sdma_v4_0_start(adev);
-
-	return r;
+	return sdma_v4_0_start(adev);
 }
 
-static int sdma_v4_0_hw_fini(void *handle)
+static int sdma_v4_0_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int i;
 
 	if (amdgpu_sriov_vf(adev))
 		return 0;
 
-	for (i = 0; i < adev->sdma.num_instances; i++) {
-		amdgpu_irq_put(adev, &adev->sdma.ecc_irq,
-			       AMDGPU_SDMA_IRQ_INSTANCE0 + i);
+	if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__SDMA)) {
+		for (i = 0; i < adev->sdma.num_instances; i++) {
+			amdgpu_irq_put(adev, &adev->sdma.ecc_irq,
+				       AMDGPU_SDMA_IRQ_INSTANCE0 + i);
+		}
 	}
 
 	sdma_v4_0_ctx_switch_enable(adev, false);
 	sdma_v4_0_enable(adev, false);
 
 	if (adev->flags & AMD_IS_APU)
-		amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_SDMA, true);
+		amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_SDMA, true, 0);
 
 	return 0;
 }
 
-static int sdma_v4_0_suspend(void *handle)
+static int sdma_v4_0_suspend(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
+
+	/* SMU saves SDMA state for us */
+	if (adev->in_s0ix) {
+		sdma_v4_0_gfx_enable(adev, false);
+		return 0;
+	}
 
-	return sdma_v4_0_hw_fini(adev);
+	return sdma_v4_0_hw_fini(ip_block);
 }
 
-static int sdma_v4_0_resume(void *handle)
+static int sdma_v4_0_resume(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
-	return sdma_v4_0_hw_init(adev);
+	/* SMU restores SDMA state for us */
+	if (adev->in_s0ix) {
+		sdma_v4_0_enable(adev, true);
+		sdma_v4_0_gfx_enable(adev, true);
+		return 0;
+	}
+
+	return sdma_v4_0_hw_init(ip_block);
 }
 
-static bool sdma_v4_0_is_idle(void *handle)
+static bool sdma_v4_0_is_idle(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	u32 i;
 
 	for (i = 0; i < adev->sdma.num_instances; i++) {
@@ -2082,11 +2030,11 @@ static bool sdma_v4_0_is_idle(void *handle)
 	return true;
 }
 
-static int sdma_v4_0_wait_for_idle(void *handle)
+static int sdma_v4_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
 {
 	unsigned i, j;
 	u32 sdma[AMDGPU_MAX_SDMA_INSTANCES];
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	for (i = 0; i < adev->usec_timeout; i++) {
 		for (j = 0; j < adev->sdma.num_instances; j++) {
@@ -2101,7 +2049,7 @@ static int sdma_v4_0_wait_for_idle(void *handle)
 	return -ETIMEDOUT;
 }
 
-static int sdma_v4_0_soft_reset(void *handle)
+static int sdma_v4_0_soft_reset(struct amdgpu_ip_block *ip_block)
 {
 	/* todo */
 
@@ -2127,23 +2075,28 @@ static int sdma_v4_0_process_trap_irq(struct amdgpu_device *adev,
 				      struct amdgpu_irq_src *source,
 				      struct amdgpu_iv_entry *entry)
 {
-	uint32_t instance;
+	int instance;
 
 	DRM_DEBUG("IH: SDMA trap\n");
 	instance = sdma_v4_0_irq_id_to_seq(entry->client_id);
+	if (instance < 0)
+		return instance;
+
 	switch (entry->ring_id) {
 	case 0:
 		amdgpu_fence_process(&adev->sdma.instance[instance].ring);
 		break;
 	case 1:
-		if (adev->ip_versions[SDMA0_HWIP][0] == IP_VERSION(4, 2, 0))
+		if (amdgpu_ip_version(adev, SDMA0_HWIP, 0) ==
+		    IP_VERSION(4, 2, 0))
 			amdgpu_fence_process(&adev->sdma.instance[instance].page);
 		break;
 	case 2:
 		/* XXX compute */
 		break;
 	case 3:
-		if (adev->ip_versions[SDMA0_HWIP][0] != IP_VERSION(4, 2, 0))
+		if (amdgpu_ip_version(adev, SDMA0_HWIP, 0) !=
+		    IP_VERSION(4, 2, 0))
 			amdgpu_fence_process(&adev->sdma.instance[instance].page);
 		break;
 	}
@@ -2212,7 +2165,7 @@ static int sdma_v4_0_print_iv_entry(struct amdgpu_device *adev,
 					      struct amdgpu_iv_entry *entry)
 {
 	int instance;
-	struct amdgpu_task_info task_info;
+	struct amdgpu_task_info *task_info;
 	u64 addr;
 
 	instance = sdma_v4_0_irq_id_to_seq(entry->client_id);
@@ -2224,15 +2177,20 @@ static int sdma_v4_0_print_iv_entry(struct amdgpu_device *adev,
 	addr = (u64)entry->src_data[0] << 12;
 	addr |= ((u64)entry->src_data[1] & 0xf) << 44;
 
-	memset(&task_info, 0, sizeof(struct amdgpu_task_info));
-	amdgpu_vm_get_task_info(adev, entry->pasid, &task_info);
-
 	dev_dbg_ratelimited(adev->dev,
-		   "[sdma%d] address:0x%016llx src_id:%u ring:%u vmid:%u "
-		   "pasid:%u, for process %s pid %d thread %s pid %d\n",
-		   instance, addr, entry->src_id, entry->ring_id, entry->vmid,
-		   entry->pasid, task_info.process_name, task_info.tgid,
-		   task_info.task_name, task_info.pid);
+			   "[sdma%d] address:0x%016llx src_id:%u ring:%u vmid:%u pasid:%u\n",
+			   instance, addr, entry->src_id, entry->ring_id, entry->vmid,
+			   entry->pasid);
+
+	task_info = amdgpu_vm_get_task_info_pasid(adev, entry->pasid);
+	if (task_info) {
+		dev_dbg_ratelimited(adev->dev,
+				    " for process %s pid %d thread %s pid %d\n",
+				    task_info->process_name, task_info->tgid,
+				    task_info->task.comm, task_info->task.pid);
+		amdgpu_vm_put_task_info(task_info);
+	}
+
 	return 0;
 }
 
@@ -2339,10 +2297,10 @@ static void sdma_v4_0_update_medium_grain_light_sleep(
 	}
 }
 
-static int sdma_v4_0_set_clockgating_state(void *handle,
+static int sdma_v4_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_clockgating_state state)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (amdgpu_sriov_vf(adev))
 		return 0;
@@ -2354,12 +2312,12 @@ static int sdma_v4_0_set_clockgating_state(void *handle,
 	return 0;
 }
 
-static int sdma_v4_0_set_powergating_state(void *handle,
+static int sdma_v4_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_powergating_state state)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
-	switch (adev->ip_versions[SDMA0_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, SDMA0_HWIP, 0)) {
 	case IP_VERSION(4, 1, 0):
 	case IP_VERSION(4, 1, 1):
 	case IP_VERSION(4, 1, 2):
@@ -2373,9 +2331,9 @@ static int sdma_v4_0_set_powergating_state(void *handle,
 	return 0;
 }
 
-static void sdma_v4_0_get_clockgating_state(void *handle, u32 *flags)
+static void sdma_v4_0_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int data;
 
 	if (amdgpu_sriov_vf(adev))
@@ -2392,6 +2350,46 @@ static void sdma_v4_0_get_clockgating_state(void *handle, u32 *flags)
 		*flags |= AMD_CG_SUPPORT_SDMA_LS;
 }
 
+static void sdma_v4_0_print_ip_state(struct amdgpu_ip_block *ip_block, struct drm_printer *p)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, j;
+	uint32_t reg_count = ARRAY_SIZE(sdma_reg_list_4_0);
+	uint32_t instance_offset;
+
+	if (!adev->sdma.ip_dump)
+		return;
+
+	drm_printf(p, "num_instances:%d\n", adev->sdma.num_instances);
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		instance_offset = i * reg_count;
+		drm_printf(p, "\nInstance:%d\n", i);
+
+		for (j = 0; j < reg_count; j++)
+			drm_printf(p, "%-50s \t 0x%08x\n", sdma_reg_list_4_0[j].reg_name,
+				   adev->sdma.ip_dump[instance_offset + j]);
+	}
+}
+
+static void sdma_v4_0_dump_ip_state(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, j;
+	uint32_t instance_offset;
+	uint32_t reg_count = ARRAY_SIZE(sdma_reg_list_4_0);
+
+	if (!adev->sdma.ip_dump)
+		return;
+
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		instance_offset = i * reg_count;
+		for (j = 0; j < reg_count; j++)
+			adev->sdma.ip_dump[instance_offset + j] =
+				RREG32(sdma_v4_0_get_reg_offset(adev, i,
+				       sdma_reg_list_4_0[j].reg_offset));
+	}
+}
+
 const struct amd_ip_funcs sdma_v4_0_ip_funcs = {
 	.name = "sdma_v4_0",
 	.early_init = sdma_v4_0_early_init,
@@ -2408,50 +2406,16 @@ const struct amd_ip_funcs sdma_v4_0_ip_funcs = {
 	.set_clockgating_state = sdma_v4_0_set_clockgating_state,
 	.set_powergating_state = sdma_v4_0_set_powergating_state,
 	.get_clockgating_state = sdma_v4_0_get_clockgating_state,
+	.dump_ip_state = sdma_v4_0_dump_ip_state,
+	.print_ip_state = sdma_v4_0_print_ip_state,
 };
 
 static const struct amdgpu_ring_funcs sdma_v4_0_ring_funcs = {
 	.type = AMDGPU_RING_TYPE_SDMA,
-	.align_mask = 0xf,
-	.nop = SDMA_PKT_NOP_HEADER_OP(SDMA_OP_NOP),
-	.support_64bit_ptrs = true,
-	.vmhub = AMDGPU_MMHUB_0,
-	.get_rptr = sdma_v4_0_ring_get_rptr,
-	.get_wptr = sdma_v4_0_ring_get_wptr,
-	.set_wptr = sdma_v4_0_ring_set_wptr,
-	.emit_frame_size =
-		6 + /* sdma_v4_0_ring_emit_hdp_flush */
-		3 + /* hdp invalidate */
-		6 + /* sdma_v4_0_ring_emit_pipeline_sync */
-		/* sdma_v4_0_ring_emit_vm_flush */
-		SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 +
-		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 6 +
-		10 + 10 + 10, /* sdma_v4_0_ring_emit_fence x3 for user fence, vm fence */
-	.emit_ib_size = 7 + 6, /* sdma_v4_0_ring_emit_ib */
-	.emit_ib = sdma_v4_0_ring_emit_ib,
-	.emit_fence = sdma_v4_0_ring_emit_fence,
-	.emit_pipeline_sync = sdma_v4_0_ring_emit_pipeline_sync,
-	.emit_vm_flush = sdma_v4_0_ring_emit_vm_flush,
-	.emit_hdp_flush = sdma_v4_0_ring_emit_hdp_flush,
-	.test_ring = sdma_v4_0_ring_test_ring,
-	.test_ib = sdma_v4_0_ring_test_ib,
-	.insert_nop = sdma_v4_0_ring_insert_nop,
-	.pad_ib = sdma_v4_0_ring_pad_ib,
-	.emit_wreg = sdma_v4_0_ring_emit_wreg,
-	.emit_reg_wait = sdma_v4_0_ring_emit_reg_wait,
-	.emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper,
-};
-
-/*
- * On Arcturus, SDMA instance 5~7 has a different vmhub type(AMDGPU_MMHUB_1).
- * So create a individual constant ring_funcs for those instances.
- */
-static const struct amdgpu_ring_funcs sdma_v4_0_ring_funcs_2nd_mmhub = {
-	.type = AMDGPU_RING_TYPE_SDMA,
-	.align_mask = 0xf,
+	.align_mask = 0xff,
 	.nop = SDMA_PKT_NOP_HEADER_OP(SDMA_OP_NOP),
 	.support_64bit_ptrs = true,
-	.vmhub = AMDGPU_MMHUB_1,
+	.secure_submission_supported = true,
 	.get_rptr = sdma_v4_0_ring_get_rptr,
 	.get_wptr = sdma_v4_0_ring_get_wptr,
 	.set_wptr = sdma_v4_0_ring_set_wptr,
@@ -2480,42 +2444,10 @@ static const struct amdgpu_ring_funcs sdma_v4_0_ring_funcs_2nd_mmhub = {
 
 static const struct amdgpu_ring_funcs sdma_v4_0_page_ring_funcs = {
 	.type = AMDGPU_RING_TYPE_SDMA,
-	.align_mask = 0xf,
+	.align_mask = 0xff,
 	.nop = SDMA_PKT_NOP_HEADER_OP(SDMA_OP_NOP),
 	.support_64bit_ptrs = true,
-	.vmhub = AMDGPU_MMHUB_0,
-	.get_rptr = sdma_v4_0_ring_get_rptr,
-	.get_wptr = sdma_v4_0_page_ring_get_wptr,
-	.set_wptr = sdma_v4_0_page_ring_set_wptr,
-	.emit_frame_size =
-		6 + /* sdma_v4_0_ring_emit_hdp_flush */
-		3 + /* hdp invalidate */
-		6 + /* sdma_v4_0_ring_emit_pipeline_sync */
-		/* sdma_v4_0_ring_emit_vm_flush */
-		SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 +
-		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 6 +
-		10 + 10 + 10, /* sdma_v4_0_ring_emit_fence x3 for user fence, vm fence */
-	.emit_ib_size = 7 + 6, /* sdma_v4_0_ring_emit_ib */
-	.emit_ib = sdma_v4_0_ring_emit_ib,
-	.emit_fence = sdma_v4_0_ring_emit_fence,
-	.emit_pipeline_sync = sdma_v4_0_ring_emit_pipeline_sync,
-	.emit_vm_flush = sdma_v4_0_ring_emit_vm_flush,
-	.emit_hdp_flush = sdma_v4_0_ring_emit_hdp_flush,
-	.test_ring = sdma_v4_0_ring_test_ring,
-	.test_ib = sdma_v4_0_ring_test_ib,
-	.insert_nop = sdma_v4_0_ring_insert_nop,
-	.pad_ib = sdma_v4_0_ring_pad_ib,
-	.emit_wreg = sdma_v4_0_ring_emit_wreg,
-	.emit_reg_wait = sdma_v4_0_ring_emit_reg_wait,
-	.emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper,
-};
-
-static const struct amdgpu_ring_funcs sdma_v4_0_page_ring_funcs_2nd_mmhub = {
-	.type = AMDGPU_RING_TYPE_SDMA,
-	.align_mask = 0xf,
-	.nop = SDMA_PKT_NOP_HEADER_OP(SDMA_OP_NOP),
-	.support_64bit_ptrs = true,
-	.vmhub = AMDGPU_MMHUB_1,
+	.secure_submission_supported = true,
 	.get_rptr = sdma_v4_0_ring_get_rptr,
 	.get_wptr = sdma_v4_0_page_ring_get_wptr,
 	.set_wptr = sdma_v4_0_page_ring_set_wptr,
@@ -2547,19 +2479,10 @@ static void sdma_v4_0_set_ring_funcs(struct amdgpu_device *adev)
 	int i;
 
 	for (i = 0; i < adev->sdma.num_instances; i++) {
-		if (adev->ip_versions[SDMA0_HWIP][0] == IP_VERSION(4, 2, 2) && i >= 5)
-			adev->sdma.instance[i].ring.funcs =
-					&sdma_v4_0_ring_funcs_2nd_mmhub;
-		else
-			adev->sdma.instance[i].ring.funcs =
-					&sdma_v4_0_ring_funcs;
+		adev->sdma.instance[i].ring.funcs = &sdma_v4_0_ring_funcs;
 		adev->sdma.instance[i].ring.me = i;
 		if (adev->sdma.has_page_queue) {
-			if (adev->ip_versions[SDMA0_HWIP][0] == IP_VERSION(4, 2, 2) && i >= 5)
-				adev->sdma.instance[i].page.funcs =
-					&sdma_v4_0_page_ring_funcs_2nd_mmhub;
-			else
-				adev->sdma.instance[i].page.funcs =
+			adev->sdma.instance[i].page.funcs =
 					&sdma_v4_0_page_ring_funcs;
 			adev->sdma.instance[i].page.me = i;
 		}
@@ -2628,7 +2551,7 @@ static void sdma_v4_0_set_irq_funcs(struct amdgpu_device *adev)
  * @src_offset: src GPU address
  * @dst_offset: dst GPU address
  * @byte_count: number of bytes to xfer
- * @tmz: if a secure copy should be used
+ * @copy_flags: copy flags for the buffers
  *
  * Copy GPU buffers using the DMA engine (VEGA10/12).
  * Used by the amdgpu ttm implementation to move pages if
@@ -2638,11 +2561,11 @@ static void sdma_v4_0_emit_copy_buffer(struct amdgpu_ib *ib,
 				       uint64_t src_offset,
 				       uint64_t dst_offset,
 				       uint32_t byte_count,
-				       bool tmz)
+				       uint32_t copy_flags)
 {
 	ib->ptr[ib->length_dw++] = SDMA_PKT_HEADER_OP(SDMA_OP_COPY) |
 		SDMA_PKT_HEADER_SUB_OP(SDMA_SUBOP_COPY_LINEAR) |
-		SDMA_PKT_COPY_LINEAR_HEADER_TMZ(tmz ? 1 : 0);
+		SDMA_PKT_COPY_LINEAR_HEADER_TMZ((copy_flags & AMDGPU_COPY_FLAGS_TMZ) ? 1 : 0);
 	ib->ptr[ib->length_dw++] = byte_count - 1;
 	ib->ptr[ib->length_dw++] = 0; /* src/dst endian swap */
 	ib->ptr[ib->length_dw++] = lower_32_bits(src_offset);
@@ -2740,7 +2663,7 @@ static void sdma_v4_0_get_ras_error_count(uint32_t value,
 	}
 }
 
-static int sdma_v4_0_query_ras_error_count(struct amdgpu_device *adev,
+static int sdma_v4_0_query_ras_error_count_by_instance(struct amdgpu_device *adev,
 			uint32_t instance, void *ras_error_status)
 {
 	struct ras_err_data *err_data = (struct ras_err_data *)ras_error_status;
@@ -2762,6 +2685,18 @@ static int sdma_v4_0_query_ras_error_count(struct amdgpu_device *adev,
 	return 0;
 };
 
+static void sdma_v4_0_query_ras_error_count(struct amdgpu_device *adev,  void *ras_error_status)
+{
+	int i = 0;
+
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		if (sdma_v4_0_query_ras_error_count_by_instance(adev, i, ras_error_status)) {
+			dev_err(adev->dev, "Query ras error count failed in SDMA%d\n", i);
+			return;
+		}
+	}
+}
+
 static void sdma_v4_0_reset_ras_error_count(struct amdgpu_device *adev)
 {
 	int i;
@@ -2773,22 +2708,27 @@ static void sdma_v4_0_reset_ras_error_count(struct amdgpu_device *adev)
 	}
 }
 
-static const struct amdgpu_sdma_ras_funcs sdma_v4_0_ras_funcs = {
-	.ras_late_init = amdgpu_sdma_ras_late_init,
-	.ras_fini = amdgpu_sdma_ras_fini,
+const struct amdgpu_ras_block_hw_ops sdma_v4_0_ras_hw_ops = {
 	.query_ras_error_count = sdma_v4_0_query_ras_error_count,
 	.reset_ras_error_count = sdma_v4_0_reset_ras_error_count,
 };
 
+static struct amdgpu_sdma_ras sdma_v4_0_ras = {
+	.ras_block = {
+		.hw_ops = &sdma_v4_0_ras_hw_ops,
+		.ras_cb = sdma_v4_0_process_ras_data_cb,
+	},
+};
+
 static void sdma_v4_0_set_ras_funcs(struct amdgpu_device *adev)
 {
-	switch (adev->ip_versions[SDMA0_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, SDMA0_HWIP, 0)) {
 	case IP_VERSION(4, 2, 0):
 	case IP_VERSION(4, 2, 2):
-		adev->sdma.funcs = &sdma_v4_0_ras_funcs;
+		adev->sdma.ras = &sdma_v4_0_ras;
 		break;
 	case IP_VERSION(4, 4, 0):
-		adev->sdma.funcs = &sdma_v4_4_ras_funcs;
+		adev->sdma.ras = &sdma_v4_4_ras;
 		break;
 	default:
 		break;
diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v4_4.c b/drivers/gpu/drm/amd/amdgpu/sdma_v4_4.c
index bf95007f0843..0ddb6955a6d3 100644
--- a/drivers/gpu/drm/amd/amdgpu/sdma_v4_4.c
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v4_4.c
@@ -141,6 +141,10 @@ static const struct soc15_ras_field_entry sdma_v4_4_ras_fields[] = {
 	SOC15_REG_FIELD(SDMA0_EDC_COUNTER2, SDMA_UTCL1_RDBST_FIFO_SED),
 	0, 0,
 	},
+	{ "SDMA_UTCL1_WR_FIFO_SED", SOC15_REG_ENTRY(SDMA0, 0, regSDMA0_EDC_COUNTER2),
+	SOC15_REG_FIELD(SDMA0_EDC_COUNTER2, SDMA_UTCL1_WR_FIFO_SED),
+	0, 0,
+	},
 	{ "SDMA_DATA_LUT_FIFO_SED", SOC15_REG_ENTRY(SDMA0, 0, regSDMA0_EDC_COUNTER2),
 	SOC15_REG_FIELD(SDMA0_EDC_COUNTER2, SDMA_DATA_LUT_FIFO_SED),
 	0, 0,
@@ -188,7 +192,7 @@ static void sdma_v4_4_get_ras_error_count(struct amdgpu_device *adev,
 	}
 }
 
-static int sdma_v4_4_query_ras_error_count(struct amdgpu_device *adev,
+static int sdma_v4_4_query_ras_error_count_by_instance(struct amdgpu_device *adev,
 					   uint32_t instance,
 					   void *ras_error_status)
 {
@@ -245,9 +249,26 @@ static void sdma_v4_4_reset_ras_error_count(struct amdgpu_device *adev)
 	}
 }
 
-const struct amdgpu_sdma_ras_funcs sdma_v4_4_ras_funcs = {
-	.ras_late_init = amdgpu_sdma_ras_late_init,
-	.ras_fini = amdgpu_sdma_ras_fini,
+static void sdma_v4_4_query_ras_error_count(struct amdgpu_device *adev,  void *ras_error_status)
+{
+	int i = 0;
+
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		if (sdma_v4_4_query_ras_error_count_by_instance(adev, i, ras_error_status)) {
+			dev_err(adev->dev, "Query ras error count failed in SDMA%d\n", i);
+			return;
+		}
+	}
+
+}
+
+const struct amdgpu_ras_block_hw_ops sdma_v4_4_ras_hw_ops = {
 	.query_ras_error_count = sdma_v4_4_query_ras_error_count,
 	.reset_ras_error_count = sdma_v4_4_reset_ras_error_count,
 };
+
+struct amdgpu_sdma_ras sdma_v4_4_ras = {
+	.ras_block = {
+		.hw_ops = &sdma_v4_4_ras_hw_ops,
+	},
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v4_4.h b/drivers/gpu/drm/amd/amdgpu/sdma_v4_4.h
index 74a6e5b5e949..a9f0c68359e0 100644
--- a/drivers/gpu/drm/amd/amdgpu/sdma_v4_4.h
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v4_4.h
@@ -23,6 +23,6 @@
 #ifndef __SDMA_V4_4_H__
 #define __SDMA_V4_4_H__
 
-extern const struct amdgpu_sdma_ras_funcs sdma_v4_4_ras_funcs;
+extern struct amdgpu_sdma_ras sdma_v4_4_ras;
 
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c b/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c
new file mode 100644
index 000000000000..a1443990d5c6
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c
@@ -0,0 +1,2613 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/delay.h>
+#include <linux/firmware.h>
+#include <linux/module.h>
+#include <linux/pci.h>
+
+#include "amdgpu.h"
+#include "amdgpu_xcp.h"
+#include "amdgpu_ucode.h"
+#include "amdgpu_trace.h"
+#include "amdgpu_reset.h"
+
+#include "sdma/sdma_4_4_2_offset.h"
+#include "sdma/sdma_4_4_2_sh_mask.h"
+
+#include "soc15_common.h"
+#include "soc15.h"
+#include "vega10_sdma_pkt_open.h"
+
+#include "ivsrcid/sdma0/irqsrcs_sdma0_4_0.h"
+#include "ivsrcid/sdma1/irqsrcs_sdma1_4_0.h"
+
+#include "amdgpu_ras.h"
+
+MODULE_FIRMWARE("amdgpu/sdma_4_4_2.bin");
+MODULE_FIRMWARE("amdgpu/sdma_4_4_4.bin");
+MODULE_FIRMWARE("amdgpu/sdma_4_4_5.bin");
+
+static const struct amdgpu_hwip_reg_entry sdma_reg_list_4_4_2[] = {
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_STATUS_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_STATUS1_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_STATUS2_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_STATUS3_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_UCODE_CHECKSUM),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_RB_RPTR_FETCH_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_RB_RPTR_FETCH),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_UTCL1_RD_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_UTCL1_WR_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_UTCL1_RD_XNACK0),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_UTCL1_RD_XNACK1),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_UTCL1_WR_XNACK0),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_UTCL1_WR_XNACK1),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_GFX_RB_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_GFX_RB_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_GFX_RB_RPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_GFX_RB_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_GFX_RB_WPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_GFX_IB_OFFSET),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_GFX_IB_BASE_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_GFX_IB_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_GFX_IB_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_GFX_IB_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_GFX_IB_SUB_REMAIN),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_GFX_DUMMY_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_PAGE_RB_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_PAGE_RB_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_PAGE_RB_RPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_PAGE_RB_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_PAGE_RB_WPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_PAGE_IB_OFFSET),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_PAGE_IB_BASE_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_PAGE_IB_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_PAGE_DUMMY_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_RLC0_RB_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_RLC0_RB_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_RLC0_RB_RPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_RLC0_RB_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_RLC0_RB_WPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_RLC0_IB_OFFSET),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_RLC0_IB_BASE_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_RLC0_IB_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_RLC0_DUMMY_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA_VM_CNTL)
+};
+
+#define mmSMNAID_AID0_MCA_SMU 0x03b30400
+
+#define WREG32_SDMA(instance, offset, value) \
+	WREG32(sdma_v4_4_2_get_reg_offset(adev, (instance), (offset)), value)
+#define RREG32_SDMA(instance, offset) \
+	RREG32(sdma_v4_4_2_get_reg_offset(adev, (instance), (offset)))
+
+static void sdma_v4_4_2_set_ring_funcs(struct amdgpu_device *adev);
+static void sdma_v4_4_2_set_buffer_funcs(struct amdgpu_device *adev);
+static void sdma_v4_4_2_set_vm_pte_funcs(struct amdgpu_device *adev);
+static void sdma_v4_4_2_set_irq_funcs(struct amdgpu_device *adev);
+static void sdma_v4_4_2_set_ras_funcs(struct amdgpu_device *adev);
+static void sdma_v4_4_2_update_reset_mask(struct amdgpu_device *adev);
+static int sdma_v4_4_2_stop_queue(struct amdgpu_ring *ring);
+static int sdma_v4_4_2_restore_queue(struct amdgpu_ring *ring);
+static int sdma_v4_4_2_soft_reset_engine(struct amdgpu_device *adev,
+					 u32 instance_id);
+
+static u32 sdma_v4_4_2_get_reg_offset(struct amdgpu_device *adev,
+		u32 instance, u32 offset)
+{
+	u32 dev_inst = GET_INST(SDMA0, instance);
+
+	return (adev->reg_offset[SDMA0_HWIP][dev_inst][0] + offset);
+}
+
+static unsigned sdma_v4_4_2_seq_to_irq_id(int seq_num)
+{
+	switch (seq_num) {
+	case 0:
+		return SOC15_IH_CLIENTID_SDMA0;
+	case 1:
+		return SOC15_IH_CLIENTID_SDMA1;
+	case 2:
+		return SOC15_IH_CLIENTID_SDMA2;
+	case 3:
+		return SOC15_IH_CLIENTID_SDMA3;
+	default:
+		return -EINVAL;
+	}
+}
+
+static int sdma_v4_4_2_irq_id_to_seq(struct amdgpu_device *adev, unsigned client_id)
+{
+	switch (client_id) {
+	case SOC15_IH_CLIENTID_SDMA0:
+		return 0;
+	case SOC15_IH_CLIENTID_SDMA1:
+		return 1;
+	case SOC15_IH_CLIENTID_SDMA2:
+		if (amdgpu_sriov_vf(adev) && (adev->gfx.xcc_mask == 0x1))
+			return 0;
+		else
+			return 2;
+	case SOC15_IH_CLIENTID_SDMA3:
+		if (amdgpu_sriov_vf(adev) && (adev->gfx.xcc_mask == 0x1))
+			return 1;
+		else
+			return 3;
+	default:
+		return -EINVAL;
+	}
+}
+
+static void sdma_v4_4_2_inst_init_golden_registers(struct amdgpu_device *adev,
+						   uint32_t inst_mask)
+{
+	u32 val;
+	int i;
+
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		val = RREG32_SDMA(i, regSDMA_GB_ADDR_CONFIG);
+		val = REG_SET_FIELD(val, SDMA_GB_ADDR_CONFIG, NUM_BANKS, 4);
+		val = REG_SET_FIELD(val, SDMA_GB_ADDR_CONFIG,
+				    PIPE_INTERLEAVE_SIZE, 0);
+		WREG32_SDMA(i, regSDMA_GB_ADDR_CONFIG, val);
+
+		val = RREG32_SDMA(i, regSDMA_GB_ADDR_CONFIG_READ);
+		val = REG_SET_FIELD(val, SDMA_GB_ADDR_CONFIG_READ, NUM_BANKS,
+				    4);
+		val = REG_SET_FIELD(val, SDMA_GB_ADDR_CONFIG_READ,
+				    PIPE_INTERLEAVE_SIZE, 0);
+		WREG32_SDMA(i, regSDMA_GB_ADDR_CONFIG_READ, val);
+	}
+}
+
+/**
+ * sdma_v4_4_2_init_microcode - load ucode images from disk
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Use the firmware interface to load the ucode images into
+ * the driver (not loaded into hw).
+ * Returns 0 on success, error on failure.
+ */
+static int sdma_v4_4_2_init_microcode(struct amdgpu_device *adev)
+{
+	int ret, i;
+
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		if (amdgpu_ip_version(adev, SDMA0_HWIP, 0) == IP_VERSION(4, 4, 2) ||
+		    amdgpu_ip_version(adev, SDMA0_HWIP, 0) == IP_VERSION(4, 4, 4) ||
+		    amdgpu_ip_version(adev, SDMA0_HWIP, 0) == IP_VERSION(4, 4, 5)) {
+			ret = amdgpu_sdma_init_microcode(adev, 0, true);
+			break;
+		} else {
+			ret = amdgpu_sdma_init_microcode(adev, i, false);
+			if (ret)
+				return ret;
+		}
+	}
+
+	return ret;
+}
+
+/**
+ * sdma_v4_4_2_ring_get_rptr - get the current read pointer
+ *
+ * @ring: amdgpu ring pointer
+ *
+ * Get the current rptr from the hardware.
+ */
+static uint64_t sdma_v4_4_2_ring_get_rptr(struct amdgpu_ring *ring)
+{
+	u64 rptr;
+
+	/* XXX check if swapping is necessary on BE */
+	rptr = READ_ONCE(*((u64 *)&ring->adev->wb.wb[ring->rptr_offs]));
+
+	DRM_DEBUG("rptr before shift == 0x%016llx\n", rptr);
+	return rptr >> 2;
+}
+
+/**
+ * sdma_v4_4_2_ring_get_wptr - get the current write pointer
+ *
+ * @ring: amdgpu ring pointer
+ *
+ * Get the current wptr from the hardware.
+ */
+static uint64_t sdma_v4_4_2_ring_get_wptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	u64 wptr;
+
+	if (ring->use_doorbell) {
+		/* XXX check if swapping is necessary on BE */
+		wptr = READ_ONCE(*((u64 *)&adev->wb.wb[ring->wptr_offs]));
+		DRM_DEBUG("wptr/doorbell before shift == 0x%016llx\n", wptr);
+	} else {
+		wptr = RREG32_SDMA(ring->me, regSDMA_GFX_RB_WPTR_HI);
+		wptr = wptr << 32;
+		wptr |= RREG32_SDMA(ring->me, regSDMA_GFX_RB_WPTR);
+		DRM_DEBUG("wptr before shift [%i] wptr == 0x%016llx\n",
+				ring->me, wptr);
+	}
+
+	return wptr >> 2;
+}
+
+/**
+ * sdma_v4_4_2_ring_set_wptr - commit the write pointer
+ *
+ * @ring: amdgpu ring pointer
+ *
+ * Write the wptr back to the hardware.
+ */
+static void sdma_v4_4_2_ring_set_wptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	DRM_DEBUG("Setting write pointer\n");
+	if (ring->use_doorbell) {
+		u64 *wb = (u64 *)&adev->wb.wb[ring->wptr_offs];
+
+		DRM_DEBUG("Using doorbell -- "
+				"wptr_offs == 0x%08x "
+				"lower_32_bits(ring->wptr) << 2 == 0x%08x "
+				"upper_32_bits(ring->wptr) << 2 == 0x%08x\n",
+				ring->wptr_offs,
+				lower_32_bits(ring->wptr << 2),
+				upper_32_bits(ring->wptr << 2));
+		/* XXX check if swapping is necessary on BE */
+		WRITE_ONCE(*wb, (ring->wptr << 2));
+		DRM_DEBUG("calling WDOORBELL64(0x%08x, 0x%016llx)\n",
+				ring->doorbell_index, ring->wptr << 2);
+		WDOORBELL64(ring->doorbell_index, ring->wptr << 2);
+	} else {
+		DRM_DEBUG("Not using doorbell -- "
+				"regSDMA%i_GFX_RB_WPTR == 0x%08x "
+				"regSDMA%i_GFX_RB_WPTR_HI == 0x%08x\n",
+				ring->me,
+				lower_32_bits(ring->wptr << 2),
+				ring->me,
+				upper_32_bits(ring->wptr << 2));
+		WREG32_SDMA(ring->me, regSDMA_GFX_RB_WPTR,
+			    lower_32_bits(ring->wptr << 2));
+		WREG32_SDMA(ring->me, regSDMA_GFX_RB_WPTR_HI,
+			    upper_32_bits(ring->wptr << 2));
+	}
+}
+
+/**
+ * sdma_v4_4_2_page_ring_get_wptr - get the current write pointer
+ *
+ * @ring: amdgpu ring pointer
+ *
+ * Get the current wptr from the hardware.
+ */
+static uint64_t sdma_v4_4_2_page_ring_get_wptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	u64 wptr;
+
+	if (ring->use_doorbell) {
+		/* XXX check if swapping is necessary on BE */
+		wptr = READ_ONCE(*((u64 *)&adev->wb.wb[ring->wptr_offs]));
+	} else {
+		wptr = RREG32_SDMA(ring->me, regSDMA_PAGE_RB_WPTR_HI);
+		wptr = wptr << 32;
+		wptr |= RREG32_SDMA(ring->me, regSDMA_PAGE_RB_WPTR);
+	}
+
+	return wptr >> 2;
+}
+
+/**
+ * sdma_v4_4_2_page_ring_set_wptr - commit the write pointer
+ *
+ * @ring: amdgpu ring pointer
+ *
+ * Write the wptr back to the hardware.
+ */
+static void sdma_v4_4_2_page_ring_set_wptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	if (ring->use_doorbell) {
+		u64 *wb = (u64 *)&adev->wb.wb[ring->wptr_offs];
+
+		/* XXX check if swapping is necessary on BE */
+		WRITE_ONCE(*wb, (ring->wptr << 2));
+		WDOORBELL64(ring->doorbell_index, ring->wptr << 2);
+	} else {
+		uint64_t wptr = ring->wptr << 2;
+
+		WREG32_SDMA(ring->me, regSDMA_PAGE_RB_WPTR,
+			    lower_32_bits(wptr));
+		WREG32_SDMA(ring->me, regSDMA_PAGE_RB_WPTR_HI,
+			    upper_32_bits(wptr));
+	}
+}
+
+static void sdma_v4_4_2_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count)
+{
+	struct amdgpu_sdma_instance *sdma = amdgpu_sdma_get_instance_from_ring(ring);
+	int i;
+
+	for (i = 0; i < count; i++)
+		if (sdma && sdma->burst_nop && (i == 0))
+			amdgpu_ring_write(ring, ring->funcs->nop |
+				SDMA_PKT_NOP_HEADER_COUNT(count - 1));
+		else
+			amdgpu_ring_write(ring, ring->funcs->nop);
+}
+
+/**
+ * sdma_v4_4_2_ring_emit_ib - Schedule an IB on the DMA engine
+ *
+ * @ring: amdgpu ring pointer
+ * @job: job to retrieve vmid from
+ * @ib: IB object to schedule
+ * @flags: unused
+ *
+ * Schedule an IB in the DMA ring.
+ */
+static void sdma_v4_4_2_ring_emit_ib(struct amdgpu_ring *ring,
+				   struct amdgpu_job *job,
+				   struct amdgpu_ib *ib,
+				   uint32_t flags)
+{
+	unsigned vmid = AMDGPU_JOB_GET_VMID(job);
+
+	/* IB packet must end on a 8 DW boundary */
+	sdma_v4_4_2_ring_insert_nop(ring, (2 - lower_32_bits(ring->wptr)) & 7);
+
+	amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_INDIRECT) |
+			  SDMA_PKT_INDIRECT_HEADER_VMID(vmid & 0xf));
+	/* base must be 32 byte aligned */
+	amdgpu_ring_write(ring, lower_32_bits(ib->gpu_addr) & 0xffffffe0);
+	amdgpu_ring_write(ring, upper_32_bits(ib->gpu_addr));
+	amdgpu_ring_write(ring, ib->length_dw);
+	amdgpu_ring_write(ring, 0);
+	amdgpu_ring_write(ring, 0);
+
+}
+
+static void sdma_v4_4_2_wait_reg_mem(struct amdgpu_ring *ring,
+				   int mem_space, int hdp,
+				   uint32_t addr0, uint32_t addr1,
+				   uint32_t ref, uint32_t mask,
+				   uint32_t inv)
+{
+	amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_POLL_REGMEM) |
+			  SDMA_PKT_POLL_REGMEM_HEADER_HDP_FLUSH(hdp) |
+			  SDMA_PKT_POLL_REGMEM_HEADER_MEM_POLL(mem_space) |
+			  SDMA_PKT_POLL_REGMEM_HEADER_FUNC(3)); /* == */
+	if (mem_space) {
+		/* memory */
+		amdgpu_ring_write(ring, addr0);
+		amdgpu_ring_write(ring, addr1);
+	} else {
+		/* registers */
+		amdgpu_ring_write(ring, addr0 << 2);
+		amdgpu_ring_write(ring, addr1 << 2);
+	}
+	amdgpu_ring_write(ring, ref); /* reference */
+	amdgpu_ring_write(ring, mask); /* mask */
+	amdgpu_ring_write(ring, SDMA_PKT_POLL_REGMEM_DW5_RETRY_COUNT(0xfff) |
+			  SDMA_PKT_POLL_REGMEM_DW5_INTERVAL(inv)); /* retry count, poll interval */
+}
+
+/**
+ * sdma_v4_4_2_ring_emit_hdp_flush - emit an hdp flush on the DMA ring
+ *
+ * @ring: amdgpu ring pointer
+ *
+ * Emit an hdp flush packet on the requested DMA ring.
+ */
+static void sdma_v4_4_2_ring_emit_hdp_flush(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	u32 ref_and_mask = 0;
+	const struct nbio_hdp_flush_reg *nbio_hf_reg = adev->nbio.hdp_flush_reg;
+
+	ref_and_mask = nbio_hf_reg->ref_and_mask_sdma0
+		       << (ring->me % adev->sdma.num_inst_per_aid);
+
+	sdma_v4_4_2_wait_reg_mem(ring, 0, 1,
+			       adev->nbio.funcs->get_hdp_flush_done_offset(adev),
+			       adev->nbio.funcs->get_hdp_flush_req_offset(adev),
+			       ref_and_mask, ref_and_mask, 10);
+}
+
+/**
+ * sdma_v4_4_2_ring_emit_fence - emit a fence on the DMA ring
+ *
+ * @ring: amdgpu ring pointer
+ * @addr: address
+ * @seq: sequence number
+ * @flags: fence related flags
+ *
+ * Add a DMA fence packet to the ring to write
+ * the fence seq number and DMA trap packet to generate
+ * an interrupt if needed.
+ */
+static void sdma_v4_4_2_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 seq,
+				      unsigned flags)
+{
+	bool write64bit = flags & AMDGPU_FENCE_FLAG_64BIT;
+	/* write the fence */
+	amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_FENCE));
+	/* zero in first two bits */
+	BUG_ON(addr & 0x3);
+	amdgpu_ring_write(ring, lower_32_bits(addr));
+	amdgpu_ring_write(ring, upper_32_bits(addr));
+	amdgpu_ring_write(ring, lower_32_bits(seq));
+
+	/* optionally write high bits as well */
+	if (write64bit) {
+		addr += 4;
+		amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_FENCE));
+		/* zero in first two bits */
+		BUG_ON(addr & 0x3);
+		amdgpu_ring_write(ring, lower_32_bits(addr));
+		amdgpu_ring_write(ring, upper_32_bits(addr));
+		amdgpu_ring_write(ring, upper_32_bits(seq));
+	}
+
+	/* generate an interrupt */
+	amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_TRAP));
+	amdgpu_ring_write(ring, SDMA_PKT_TRAP_INT_CONTEXT_INT_CONTEXT(0));
+}
+
+
+/**
+ * sdma_v4_4_2_inst_gfx_stop - stop the gfx async dma engines
+ *
+ * @adev: amdgpu_device pointer
+ * @inst_mask: mask of dma engine instances to be disabled
+ *
+ * Stop the gfx async dma ring buffers.
+ */
+static void sdma_v4_4_2_inst_gfx_stop(struct amdgpu_device *adev,
+				      uint32_t inst_mask)
+{
+	struct amdgpu_ring *sdma[AMDGPU_MAX_SDMA_INSTANCES];
+	u32 doorbell_offset, doorbell;
+	u32 rb_cntl, ib_cntl, sdma_cntl;
+	int i;
+
+	for_each_inst(i, inst_mask) {
+		sdma[i] = &adev->sdma.instance[i].ring;
+
+		rb_cntl = RREG32_SDMA(i, regSDMA_GFX_RB_CNTL);
+		rb_cntl = REG_SET_FIELD(rb_cntl, SDMA_GFX_RB_CNTL, RB_ENABLE, 0);
+		WREG32_SDMA(i, regSDMA_GFX_RB_CNTL, rb_cntl);
+		ib_cntl = RREG32_SDMA(i, regSDMA_GFX_IB_CNTL);
+		ib_cntl = REG_SET_FIELD(ib_cntl, SDMA_GFX_IB_CNTL, IB_ENABLE, 0);
+		WREG32_SDMA(i, regSDMA_GFX_IB_CNTL, ib_cntl);
+		sdma_cntl = RREG32_SDMA(i, regSDMA_CNTL);
+		sdma_cntl = REG_SET_FIELD(sdma_cntl, SDMA_CNTL, UTC_L1_ENABLE, 0);
+		WREG32_SDMA(i, regSDMA_CNTL, sdma_cntl);
+
+		if (sdma[i]->use_doorbell) {
+			doorbell = RREG32_SDMA(i, regSDMA_GFX_DOORBELL);
+			doorbell_offset = RREG32_SDMA(i, regSDMA_GFX_DOORBELL_OFFSET);
+
+			doorbell = REG_SET_FIELD(doorbell, SDMA_GFX_DOORBELL, ENABLE, 0);
+			doorbell_offset = REG_SET_FIELD(doorbell_offset,
+					SDMA_GFX_DOORBELL_OFFSET,
+					OFFSET, 0);
+			WREG32_SDMA(i, regSDMA_GFX_DOORBELL, doorbell);
+			WREG32_SDMA(i, regSDMA_GFX_DOORBELL_OFFSET, doorbell_offset);
+		}
+	}
+}
+
+/**
+ * sdma_v4_4_2_inst_rlc_stop - stop the compute async dma engines
+ *
+ * @adev: amdgpu_device pointer
+ * @inst_mask: mask of dma engine instances to be disabled
+ *
+ * Stop the compute async dma queues.
+ */
+static void sdma_v4_4_2_inst_rlc_stop(struct amdgpu_device *adev,
+				      uint32_t inst_mask)
+{
+	/* XXX todo */
+}
+
+/**
+ * sdma_v4_4_2_inst_page_stop - stop the page async dma engines
+ *
+ * @adev: amdgpu_device pointer
+ * @inst_mask: mask of dma engine instances to be disabled
+ *
+ * Stop the page async dma ring buffers.
+ */
+static void sdma_v4_4_2_inst_page_stop(struct amdgpu_device *adev,
+				       uint32_t inst_mask)
+{
+	u32 rb_cntl, ib_cntl;
+	int i;
+
+	for_each_inst(i, inst_mask) {
+		rb_cntl = RREG32_SDMA(i, regSDMA_PAGE_RB_CNTL);
+		rb_cntl = REG_SET_FIELD(rb_cntl, SDMA_PAGE_RB_CNTL,
+					RB_ENABLE, 0);
+		WREG32_SDMA(i, regSDMA_PAGE_RB_CNTL, rb_cntl);
+		ib_cntl = RREG32_SDMA(i, regSDMA_PAGE_IB_CNTL);
+		ib_cntl = REG_SET_FIELD(ib_cntl, SDMA_PAGE_IB_CNTL,
+					IB_ENABLE, 0);
+		WREG32_SDMA(i, regSDMA_PAGE_IB_CNTL, ib_cntl);
+	}
+}
+
+/**
+ * sdma_v4_4_2_inst_ctx_switch_enable - stop the async dma engines context switch
+ *
+ * @adev: amdgpu_device pointer
+ * @enable: enable/disable the DMA MEs context switch.
+ * @inst_mask: mask of dma engine instances to be enabled
+ *
+ * Halt or unhalt the async dma engines context switch.
+ */
+static void sdma_v4_4_2_inst_ctx_switch_enable(struct amdgpu_device *adev,
+					       bool enable, uint32_t inst_mask)
+{
+	u32 f32_cntl, phase_quantum = 0;
+	int i;
+
+	if (amdgpu_sdma_phase_quantum) {
+		unsigned value = amdgpu_sdma_phase_quantum;
+		unsigned unit = 0;
+
+		while (value > (SDMA_PHASE0_QUANTUM__VALUE_MASK >>
+				SDMA_PHASE0_QUANTUM__VALUE__SHIFT)) {
+			value = (value + 1) >> 1;
+			unit++;
+		}
+		if (unit > (SDMA_PHASE0_QUANTUM__UNIT_MASK >>
+			    SDMA_PHASE0_QUANTUM__UNIT__SHIFT)) {
+			value = (SDMA_PHASE0_QUANTUM__VALUE_MASK >>
+				 SDMA_PHASE0_QUANTUM__VALUE__SHIFT);
+			unit = (SDMA_PHASE0_QUANTUM__UNIT_MASK >>
+				SDMA_PHASE0_QUANTUM__UNIT__SHIFT);
+			WARN_ONCE(1,
+			"clamping sdma_phase_quantum to %uK clock cycles\n",
+				  value << unit);
+		}
+		phase_quantum =
+			value << SDMA_PHASE0_QUANTUM__VALUE__SHIFT |
+			unit  << SDMA_PHASE0_QUANTUM__UNIT__SHIFT;
+	}
+
+	for_each_inst(i, inst_mask) {
+		f32_cntl = RREG32_SDMA(i, regSDMA_CNTL);
+		f32_cntl = REG_SET_FIELD(f32_cntl, SDMA_CNTL,
+				AUTO_CTXSW_ENABLE, enable ? 1 : 0);
+		if (enable && amdgpu_sdma_phase_quantum) {
+			WREG32_SDMA(i, regSDMA_PHASE0_QUANTUM, phase_quantum);
+			WREG32_SDMA(i, regSDMA_PHASE1_QUANTUM, phase_quantum);
+			WREG32_SDMA(i, regSDMA_PHASE2_QUANTUM, phase_quantum);
+		}
+		WREG32_SDMA(i, regSDMA_CNTL, f32_cntl);
+
+		/* Extend page fault timeout to avoid interrupt storm */
+		WREG32_SDMA(i, regSDMA_UTCL1_TIMEOUT, 0x00800080);
+	}
+}
+
+/**
+ * sdma_v4_4_2_inst_enable - stop the async dma engines
+ *
+ * @adev: amdgpu_device pointer
+ * @enable: enable/disable the DMA MEs.
+ * @inst_mask: mask of dma engine instances to be enabled
+ *
+ * Halt or unhalt the async dma engines.
+ */
+static void sdma_v4_4_2_inst_enable(struct amdgpu_device *adev, bool enable,
+				    uint32_t inst_mask)
+{
+	u32 f32_cntl;
+	int i;
+
+	if (!enable) {
+		sdma_v4_4_2_inst_gfx_stop(adev, inst_mask);
+		sdma_v4_4_2_inst_rlc_stop(adev, inst_mask);
+		if (adev->sdma.has_page_queue)
+			sdma_v4_4_2_inst_page_stop(adev, inst_mask);
+
+		/* SDMA FW needs to respond to FREEZE requests during reset.
+		 * Keep it running during reset */
+		if (!amdgpu_sriov_vf(adev) && amdgpu_in_reset(adev))
+			return;
+	}
+
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP)
+		return;
+
+	for_each_inst(i, inst_mask) {
+		f32_cntl = RREG32_SDMA(i, regSDMA_F32_CNTL);
+		f32_cntl = REG_SET_FIELD(f32_cntl, SDMA_F32_CNTL, HALT, enable ? 0 : 1);
+		WREG32_SDMA(i, regSDMA_F32_CNTL, f32_cntl);
+	}
+}
+
+/*
+ * sdma_v4_4_2_rb_cntl - get parameters for rb_cntl
+ */
+static uint32_t sdma_v4_4_2_rb_cntl(struct amdgpu_ring *ring, uint32_t rb_cntl)
+{
+	/* Set ring buffer size in dwords */
+	uint32_t rb_bufsz = order_base_2(ring->ring_size / 4);
+
+	barrier(); /* work around https://llvm.org/pr42576 */
+	rb_cntl = REG_SET_FIELD(rb_cntl, SDMA_GFX_RB_CNTL, RB_SIZE, rb_bufsz);
+#ifdef __BIG_ENDIAN
+	rb_cntl = REG_SET_FIELD(rb_cntl, SDMA_GFX_RB_CNTL, RB_SWAP_ENABLE, 1);
+	rb_cntl = REG_SET_FIELD(rb_cntl, SDMA_GFX_RB_CNTL,
+				RPTR_WRITEBACK_SWAP_ENABLE, 1);
+#endif
+	return rb_cntl;
+}
+
+/**
+ * sdma_v4_4_2_gfx_resume - setup and start the async dma engines
+ *
+ * @adev: amdgpu_device pointer
+ * @i: instance to resume
+ * @restore: used to restore wptr when restart
+ *
+ * Set up the gfx DMA ring buffers and enable them.
+ * Returns 0 for success, error for failure.
+ */
+static void sdma_v4_4_2_gfx_resume(struct amdgpu_device *adev, unsigned int i, bool restore)
+{
+	struct amdgpu_ring *ring = &adev->sdma.instance[i].ring;
+	u32 rb_cntl, ib_cntl, wptr_poll_cntl;
+	u32 wb_offset;
+	u32 doorbell;
+	u32 doorbell_offset;
+	u64 wptr_gpu_addr;
+	u64 rwptr;
+
+	wb_offset = (ring->rptr_offs * 4);
+
+	rb_cntl = RREG32_SDMA(i, regSDMA_GFX_RB_CNTL);
+	rb_cntl = sdma_v4_4_2_rb_cntl(ring, rb_cntl);
+	WREG32_SDMA(i, regSDMA_GFX_RB_CNTL, rb_cntl);
+
+	/* set the wb address whether it's enabled or not */
+	WREG32_SDMA(i, regSDMA_GFX_RB_RPTR_ADDR_HI,
+	       upper_32_bits(adev->wb.gpu_addr + wb_offset) & 0xFFFFFFFF);
+	WREG32_SDMA(i, regSDMA_GFX_RB_RPTR_ADDR_LO,
+	       lower_32_bits(adev->wb.gpu_addr + wb_offset) & 0xFFFFFFFC);
+
+	rb_cntl = REG_SET_FIELD(rb_cntl, SDMA_GFX_RB_CNTL,
+				RPTR_WRITEBACK_ENABLE, 1);
+
+	WREG32_SDMA(i, regSDMA_GFX_RB_BASE, ring->gpu_addr >> 8);
+	WREG32_SDMA(i, regSDMA_GFX_RB_BASE_HI, ring->gpu_addr >> 40);
+
+	if (!restore)
+		ring->wptr = 0;
+
+	/* before programing wptr to a less value, need set minor_ptr_update first */
+	WREG32_SDMA(i, regSDMA_GFX_MINOR_PTR_UPDATE, 1);
+
+	/* For the guilty queue, set RPTR to the current wptr to skip bad commands,
+	 * It is not a guilty queue, restore cache_rptr and continue execution.
+	 */
+	if (adev->sdma.instance[i].gfx_guilty)
+		rwptr = ring->wptr;
+	else
+		rwptr = ring->cached_rptr;
+
+	/* Initialize the ring buffer's read and write pointers */
+	if (restore) {
+		WREG32_SDMA(i, regSDMA_GFX_RB_RPTR, lower_32_bits(rwptr << 2));
+		WREG32_SDMA(i, regSDMA_GFX_RB_RPTR_HI, upper_32_bits(rwptr << 2));
+		WREG32_SDMA(i, regSDMA_GFX_RB_WPTR, lower_32_bits(rwptr << 2));
+		WREG32_SDMA(i, regSDMA_GFX_RB_WPTR_HI, upper_32_bits(rwptr << 2));
+	} else {
+		WREG32_SDMA(i, regSDMA_GFX_RB_RPTR, 0);
+		WREG32_SDMA(i, regSDMA_GFX_RB_RPTR_HI, 0);
+		WREG32_SDMA(i, regSDMA_GFX_RB_WPTR, 0);
+		WREG32_SDMA(i, regSDMA_GFX_RB_WPTR_HI, 0);
+	}
+
+	doorbell = RREG32_SDMA(i, regSDMA_GFX_DOORBELL);
+	doorbell_offset = RREG32_SDMA(i, regSDMA_GFX_DOORBELL_OFFSET);
+
+	doorbell = REG_SET_FIELD(doorbell, SDMA_GFX_DOORBELL, ENABLE,
+				 ring->use_doorbell);
+	doorbell_offset = REG_SET_FIELD(doorbell_offset,
+					SDMA_GFX_DOORBELL_OFFSET,
+					OFFSET, ring->doorbell_index);
+	WREG32_SDMA(i, regSDMA_GFX_DOORBELL, doorbell);
+	WREG32_SDMA(i, regSDMA_GFX_DOORBELL_OFFSET, doorbell_offset);
+
+	sdma_v4_4_2_ring_set_wptr(ring);
+
+	/* set minor_ptr_update to 0 after wptr programed */
+	WREG32_SDMA(i, regSDMA_GFX_MINOR_PTR_UPDATE, 0);
+
+	/* setup the wptr shadow polling */
+	wptr_gpu_addr = adev->wb.gpu_addr + (ring->wptr_offs * 4);
+	WREG32_SDMA(i, regSDMA_GFX_RB_WPTR_POLL_ADDR_LO,
+		    lower_32_bits(wptr_gpu_addr));
+	WREG32_SDMA(i, regSDMA_GFX_RB_WPTR_POLL_ADDR_HI,
+		    upper_32_bits(wptr_gpu_addr));
+	wptr_poll_cntl = RREG32_SDMA(i, regSDMA_GFX_RB_WPTR_POLL_CNTL);
+	wptr_poll_cntl = REG_SET_FIELD(wptr_poll_cntl,
+				       SDMA_GFX_RB_WPTR_POLL_CNTL,
+				       F32_POLL_ENABLE, amdgpu_sriov_vf(adev)? 1 : 0);
+	WREG32_SDMA(i, regSDMA_GFX_RB_WPTR_POLL_CNTL, wptr_poll_cntl);
+
+	/* enable DMA RB */
+	rb_cntl = REG_SET_FIELD(rb_cntl, SDMA_GFX_RB_CNTL, RB_ENABLE, 1);
+	WREG32_SDMA(i, regSDMA_GFX_RB_CNTL, rb_cntl);
+
+	ib_cntl = RREG32_SDMA(i, regSDMA_GFX_IB_CNTL);
+	ib_cntl = REG_SET_FIELD(ib_cntl, SDMA_GFX_IB_CNTL, IB_ENABLE, 1);
+#ifdef __BIG_ENDIAN
+	ib_cntl = REG_SET_FIELD(ib_cntl, SDMA_GFX_IB_CNTL, IB_SWAP_ENABLE, 1);
+#endif
+	/* enable DMA IBs */
+	WREG32_SDMA(i, regSDMA_GFX_IB_CNTL, ib_cntl);
+}
+
+/**
+ * sdma_v4_4_2_page_resume - setup and start the async dma engines
+ *
+ * @adev: amdgpu_device pointer
+ * @i: instance to resume
+ * @restore: boolean to say restore needed or not
+ *
+ * Set up the page DMA ring buffers and enable them.
+ * Returns 0 for success, error for failure.
+ */
+static void sdma_v4_4_2_page_resume(struct amdgpu_device *adev, unsigned int i, bool restore)
+{
+	struct amdgpu_ring *ring = &adev->sdma.instance[i].page;
+	u32 rb_cntl, ib_cntl, wptr_poll_cntl;
+	u32 wb_offset;
+	u32 doorbell;
+	u32 doorbell_offset;
+	u64 wptr_gpu_addr;
+	u64 rwptr;
+
+	wb_offset = (ring->rptr_offs * 4);
+
+	rb_cntl = RREG32_SDMA(i, regSDMA_PAGE_RB_CNTL);
+	rb_cntl = sdma_v4_4_2_rb_cntl(ring, rb_cntl);
+	WREG32_SDMA(i, regSDMA_PAGE_RB_CNTL, rb_cntl);
+
+	/* For the guilty queue, set RPTR to the current wptr to skip bad commands,
+	 * It is not a guilty queue, restore cache_rptr and continue execution.
+	 */
+	if (adev->sdma.instance[i].page_guilty)
+		rwptr = ring->wptr;
+	else
+		rwptr = ring->cached_rptr;
+
+	/* Initialize the ring buffer's read and write pointers */
+	if (restore) {
+		WREG32_SDMA(i, regSDMA_PAGE_RB_RPTR, lower_32_bits(rwptr << 2));
+		WREG32_SDMA(i, regSDMA_PAGE_RB_RPTR_HI, upper_32_bits(rwptr << 2));
+		WREG32_SDMA(i, regSDMA_PAGE_RB_WPTR, lower_32_bits(rwptr << 2));
+		WREG32_SDMA(i, regSDMA_PAGE_RB_WPTR_HI, upper_32_bits(rwptr << 2));
+	} else {
+		WREG32_SDMA(i, regSDMA_PAGE_RB_RPTR, 0);
+		WREG32_SDMA(i, regSDMA_PAGE_RB_RPTR_HI, 0);
+		WREG32_SDMA(i, regSDMA_PAGE_RB_WPTR, 0);
+		WREG32_SDMA(i, regSDMA_PAGE_RB_WPTR_HI, 0);
+	}
+
+	/* set the wb address whether it's enabled or not */
+	WREG32_SDMA(i, regSDMA_PAGE_RB_RPTR_ADDR_HI,
+	       upper_32_bits(adev->wb.gpu_addr + wb_offset) & 0xFFFFFFFF);
+	WREG32_SDMA(i, regSDMA_PAGE_RB_RPTR_ADDR_LO,
+	       lower_32_bits(adev->wb.gpu_addr + wb_offset) & 0xFFFFFFFC);
+
+	rb_cntl = REG_SET_FIELD(rb_cntl, SDMA_PAGE_RB_CNTL,
+				RPTR_WRITEBACK_ENABLE, 1);
+
+	WREG32_SDMA(i, regSDMA_PAGE_RB_BASE, ring->gpu_addr >> 8);
+	WREG32_SDMA(i, regSDMA_PAGE_RB_BASE_HI, ring->gpu_addr >> 40);
+
+	if (!restore)
+		ring->wptr = 0;
+
+	/* before programing wptr to a less value, need set minor_ptr_update first */
+	WREG32_SDMA(i, regSDMA_PAGE_MINOR_PTR_UPDATE, 1);
+
+	doorbell = RREG32_SDMA(i, regSDMA_PAGE_DOORBELL);
+	doorbell_offset = RREG32_SDMA(i, regSDMA_PAGE_DOORBELL_OFFSET);
+
+	doorbell = REG_SET_FIELD(doorbell, SDMA_PAGE_DOORBELL, ENABLE,
+				 ring->use_doorbell);
+	doorbell_offset = REG_SET_FIELD(doorbell_offset,
+					SDMA_PAGE_DOORBELL_OFFSET,
+					OFFSET, ring->doorbell_index);
+	WREG32_SDMA(i, regSDMA_PAGE_DOORBELL, doorbell);
+	WREG32_SDMA(i, regSDMA_PAGE_DOORBELL_OFFSET, doorbell_offset);
+
+	/* paging queue doorbell range is setup at sdma_v4_4_2_gfx_resume */
+	sdma_v4_4_2_page_ring_set_wptr(ring);
+
+	/* set minor_ptr_update to 0 after wptr programed */
+	WREG32_SDMA(i, regSDMA_PAGE_MINOR_PTR_UPDATE, 0);
+
+	/* setup the wptr shadow polling */
+	wptr_gpu_addr = adev->wb.gpu_addr + (ring->wptr_offs * 4);
+	WREG32_SDMA(i, regSDMA_PAGE_RB_WPTR_POLL_ADDR_LO,
+		    lower_32_bits(wptr_gpu_addr));
+	WREG32_SDMA(i, regSDMA_PAGE_RB_WPTR_POLL_ADDR_HI,
+		    upper_32_bits(wptr_gpu_addr));
+	wptr_poll_cntl = RREG32_SDMA(i, regSDMA_PAGE_RB_WPTR_POLL_CNTL);
+	wptr_poll_cntl = REG_SET_FIELD(wptr_poll_cntl,
+				       SDMA_PAGE_RB_WPTR_POLL_CNTL,
+				       F32_POLL_ENABLE, amdgpu_sriov_vf(adev)? 1 : 0);
+	WREG32_SDMA(i, regSDMA_PAGE_RB_WPTR_POLL_CNTL, wptr_poll_cntl);
+
+	/* enable DMA RB */
+	rb_cntl = REG_SET_FIELD(rb_cntl, SDMA_PAGE_RB_CNTL, RB_ENABLE, 1);
+	WREG32_SDMA(i, regSDMA_PAGE_RB_CNTL, rb_cntl);
+
+	ib_cntl = RREG32_SDMA(i, regSDMA_PAGE_IB_CNTL);
+	ib_cntl = REG_SET_FIELD(ib_cntl, SDMA_PAGE_IB_CNTL, IB_ENABLE, 1);
+#ifdef __BIG_ENDIAN
+	ib_cntl = REG_SET_FIELD(ib_cntl, SDMA_PAGE_IB_CNTL, IB_SWAP_ENABLE, 1);
+#endif
+	/* enable DMA IBs */
+	WREG32_SDMA(i, regSDMA_PAGE_IB_CNTL, ib_cntl);
+}
+
+static void sdma_v4_4_2_init_pg(struct amdgpu_device *adev)
+{
+
+}
+
+/**
+ * sdma_v4_4_2_inst_rlc_resume - setup and start the async dma engines
+ *
+ * @adev: amdgpu_device pointer
+ * @inst_mask: mask of dma engine instances to be enabled
+ *
+ * Set up the compute DMA queues and enable them.
+ * Returns 0 for success, error for failure.
+ */
+static int sdma_v4_4_2_inst_rlc_resume(struct amdgpu_device *adev,
+				       uint32_t inst_mask)
+{
+	sdma_v4_4_2_init_pg(adev);
+
+	return 0;
+}
+
+/**
+ * sdma_v4_4_2_inst_load_microcode - load the sDMA ME ucode
+ *
+ * @adev: amdgpu_device pointer
+ * @inst_mask: mask of dma engine instances to be enabled
+ *
+ * Loads the sDMA0/1 ucode.
+ * Returns 0 for success, -EINVAL if the ucode is not available.
+ */
+static int sdma_v4_4_2_inst_load_microcode(struct amdgpu_device *adev,
+					   uint32_t inst_mask)
+{
+	const struct sdma_firmware_header_v1_0 *hdr;
+	const __le32 *fw_data;
+	u32 fw_size;
+	int i, j;
+
+	/* halt the MEs */
+	sdma_v4_4_2_inst_enable(adev, false, inst_mask);
+
+	for_each_inst(i, inst_mask) {
+		if (!adev->sdma.instance[i].fw)
+			return -EINVAL;
+
+		hdr = (const struct sdma_firmware_header_v1_0 *)adev->sdma.instance[i].fw->data;
+		amdgpu_ucode_print_sdma_hdr(&hdr->header);
+		fw_size = le32_to_cpu(hdr->header.ucode_size_bytes) / 4;
+
+		fw_data = (const __le32 *)
+			(adev->sdma.instance[i].fw->data +
+				le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+
+		WREG32_SDMA(i, regSDMA_UCODE_ADDR, 0);
+
+		for (j = 0; j < fw_size; j++)
+			WREG32_SDMA(i, regSDMA_UCODE_DATA,
+				    le32_to_cpup(fw_data++));
+
+		WREG32_SDMA(i, regSDMA_UCODE_ADDR,
+			    adev->sdma.instance[i].fw_version);
+	}
+
+	return 0;
+}
+
+/**
+ * sdma_v4_4_2_inst_start - setup and start the async dma engines
+ *
+ * @adev: amdgpu_device pointer
+ * @inst_mask: mask of dma engine instances to be enabled
+ * @restore: boolean to say restore needed or not
+ *
+ * Set up the DMA engines and enable them.
+ * Returns 0 for success, error for failure.
+ */
+static int sdma_v4_4_2_inst_start(struct amdgpu_device *adev,
+				  uint32_t inst_mask, bool restore)
+{
+	struct amdgpu_ring *ring;
+	uint32_t tmp_mask;
+	int i, r = 0;
+
+	if (amdgpu_sriov_vf(adev)) {
+		sdma_v4_4_2_inst_ctx_switch_enable(adev, false, inst_mask);
+		sdma_v4_4_2_inst_enable(adev, false, inst_mask);
+	} else {
+		/* bypass sdma microcode loading on Gopher */
+		if (!restore && adev->firmware.load_type != AMDGPU_FW_LOAD_PSP &&
+		    adev->sdma.instance[0].fw) {
+			r = sdma_v4_4_2_inst_load_microcode(adev, inst_mask);
+			if (r)
+				return r;
+		}
+
+		/* unhalt the MEs */
+		sdma_v4_4_2_inst_enable(adev, true, inst_mask);
+		/* enable sdma ring preemption */
+		sdma_v4_4_2_inst_ctx_switch_enable(adev, true, inst_mask);
+	}
+
+	/* start the gfx rings and rlc compute queues */
+	tmp_mask = inst_mask;
+	for_each_inst(i, tmp_mask) {
+		uint32_t temp;
+
+		WREG32_SDMA(i, regSDMA_SEM_WAIT_FAIL_TIMER_CNTL, 0);
+		sdma_v4_4_2_gfx_resume(adev, i, restore);
+		if (adev->sdma.has_page_queue)
+			sdma_v4_4_2_page_resume(adev, i, restore);
+
+		/* set utc l1 enable flag always to 1 */
+		temp = RREG32_SDMA(i, regSDMA_CNTL);
+		temp = REG_SET_FIELD(temp, SDMA_CNTL, UTC_L1_ENABLE, 1);
+		WREG32_SDMA(i, regSDMA_CNTL, temp);
+
+		if (amdgpu_ip_version(adev, SDMA0_HWIP, 0) < IP_VERSION(4, 4, 5)) {
+			/* enable context empty interrupt during initialization */
+			temp = REG_SET_FIELD(temp, SDMA_CNTL, CTXEMPTY_INT_ENABLE, 1);
+			WREG32_SDMA(i, regSDMA_CNTL, temp);
+		}
+		if (!amdgpu_sriov_vf(adev)) {
+			if (adev->firmware.load_type != AMDGPU_FW_LOAD_PSP) {
+				/* unhalt engine */
+				temp = RREG32_SDMA(i, regSDMA_F32_CNTL);
+				temp = REG_SET_FIELD(temp, SDMA_F32_CNTL, HALT, 0);
+				WREG32_SDMA(i, regSDMA_F32_CNTL, temp);
+			}
+		}
+	}
+
+	if (amdgpu_sriov_vf(adev)) {
+		sdma_v4_4_2_inst_ctx_switch_enable(adev, true, inst_mask);
+		sdma_v4_4_2_inst_enable(adev, true, inst_mask);
+	} else {
+		r = sdma_v4_4_2_inst_rlc_resume(adev, inst_mask);
+		if (r)
+			return r;
+	}
+
+	tmp_mask = inst_mask;
+	for_each_inst(i, tmp_mask) {
+		ring = &adev->sdma.instance[i].ring;
+
+		r = amdgpu_ring_test_helper(ring);
+		if (r)
+			return r;
+
+		if (adev->sdma.has_page_queue) {
+			struct amdgpu_ring *page = &adev->sdma.instance[i].page;
+
+			r = amdgpu_ring_test_helper(page);
+			if (r)
+				return r;
+		}
+	}
+
+	return r;
+}
+
+/**
+ * sdma_v4_4_2_ring_test_ring - simple async dma engine test
+ *
+ * @ring: amdgpu_ring structure holding ring information
+ *
+ * Test the DMA engine by writing using it to write an
+ * value to memory.
+ * Returns 0 for success, error for failure.
+ */
+static int sdma_v4_4_2_ring_test_ring(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	unsigned i;
+	unsigned index;
+	int r;
+	u32 tmp;
+	u64 gpu_addr;
+
+	r = amdgpu_device_wb_get(adev, &index);
+	if (r)
+		return r;
+
+	gpu_addr = adev->wb.gpu_addr + (index * 4);
+	tmp = 0xCAFEDEAD;
+	adev->wb.wb[index] = cpu_to_le32(tmp);
+
+	r = amdgpu_ring_alloc(ring, 5);
+	if (r)
+		goto error_free_wb;
+
+	amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_WRITE) |
+			  SDMA_PKT_HEADER_SUB_OP(SDMA_SUBOP_WRITE_LINEAR));
+	amdgpu_ring_write(ring, lower_32_bits(gpu_addr));
+	amdgpu_ring_write(ring, upper_32_bits(gpu_addr));
+	amdgpu_ring_write(ring, SDMA_PKT_WRITE_UNTILED_DW_3_COUNT(0));
+	amdgpu_ring_write(ring, 0xDEADBEEF);
+	amdgpu_ring_commit(ring);
+
+	for (i = 0; i < adev->usec_timeout; i++) {
+		tmp = le32_to_cpu(adev->wb.wb[index]);
+		if (tmp == 0xDEADBEEF)
+			break;
+		udelay(1);
+	}
+
+	if (i >= adev->usec_timeout)
+		r = -ETIMEDOUT;
+
+error_free_wb:
+	amdgpu_device_wb_free(adev, index);
+	return r;
+}
+
+/**
+ * sdma_v4_4_2_ring_test_ib - test an IB on the DMA engine
+ *
+ * @ring: amdgpu_ring structure holding ring information
+ * @timeout: timeout value in jiffies, or MAX_SCHEDULE_TIMEOUT
+ *
+ * Test a simple IB in the DMA ring.
+ * Returns 0 on success, error on failure.
+ */
+static int sdma_v4_4_2_ring_test_ib(struct amdgpu_ring *ring, long timeout)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct amdgpu_ib ib;
+	struct dma_fence *f = NULL;
+	unsigned index;
+	long r;
+	u32 tmp = 0;
+	u64 gpu_addr;
+
+	r = amdgpu_device_wb_get(adev, &index);
+	if (r)
+		return r;
+
+	gpu_addr = adev->wb.gpu_addr + (index * 4);
+	tmp = 0xCAFEDEAD;
+	adev->wb.wb[index] = cpu_to_le32(tmp);
+	memset(&ib, 0, sizeof(ib));
+	r = amdgpu_ib_get(adev, NULL, 256,
+					AMDGPU_IB_POOL_DIRECT, &ib);
+	if (r)
+		goto err0;
+
+	ib.ptr[0] = SDMA_PKT_HEADER_OP(SDMA_OP_WRITE) |
+		SDMA_PKT_HEADER_SUB_OP(SDMA_SUBOP_WRITE_LINEAR);
+	ib.ptr[1] = lower_32_bits(gpu_addr);
+	ib.ptr[2] = upper_32_bits(gpu_addr);
+	ib.ptr[3] = SDMA_PKT_WRITE_UNTILED_DW_3_COUNT(0);
+	ib.ptr[4] = 0xDEADBEEF;
+	ib.ptr[5] = SDMA_PKT_NOP_HEADER_OP(SDMA_OP_NOP);
+	ib.ptr[6] = SDMA_PKT_NOP_HEADER_OP(SDMA_OP_NOP);
+	ib.ptr[7] = SDMA_PKT_NOP_HEADER_OP(SDMA_OP_NOP);
+	ib.length_dw = 8;
+
+	r = amdgpu_ib_schedule(ring, 1, &ib, NULL, &f);
+	if (r)
+		goto err1;
+
+	r = dma_fence_wait_timeout(f, false, timeout);
+	if (r == 0) {
+		r = -ETIMEDOUT;
+		goto err1;
+	} else if (r < 0) {
+		goto err1;
+	}
+	tmp = le32_to_cpu(adev->wb.wb[index]);
+	if (tmp == 0xDEADBEEF)
+		r = 0;
+	else
+		r = -EINVAL;
+
+err1:
+	amdgpu_ib_free(&ib, NULL);
+	dma_fence_put(f);
+err0:
+	amdgpu_device_wb_free(adev, index);
+	return r;
+}
+
+
+/**
+ * sdma_v4_4_2_vm_copy_pte - update PTEs by copying them from the GART
+ *
+ * @ib: indirect buffer to fill with commands
+ * @pe: addr of the page entry
+ * @src: src addr to copy from
+ * @count: number of page entries to update
+ *
+ * Update PTEs by copying them from the GART using sDMA.
+ */
+static void sdma_v4_4_2_vm_copy_pte(struct amdgpu_ib *ib,
+				  uint64_t pe, uint64_t src,
+				  unsigned count)
+{
+	unsigned bytes = count * 8;
+
+	ib->ptr[ib->length_dw++] = SDMA_PKT_HEADER_OP(SDMA_OP_COPY) |
+		SDMA_PKT_HEADER_SUB_OP(SDMA_SUBOP_COPY_LINEAR);
+	ib->ptr[ib->length_dw++] = bytes - 1;
+	ib->ptr[ib->length_dw++] = 0; /* src/dst endian swap */
+	ib->ptr[ib->length_dw++] = lower_32_bits(src);
+	ib->ptr[ib->length_dw++] = upper_32_bits(src);
+	ib->ptr[ib->length_dw++] = lower_32_bits(pe);
+	ib->ptr[ib->length_dw++] = upper_32_bits(pe);
+
+}
+
+/**
+ * sdma_v4_4_2_vm_write_pte - update PTEs by writing them manually
+ *
+ * @ib: indirect buffer to fill with commands
+ * @pe: addr of the page entry
+ * @value: dst addr to write into pe
+ * @count: number of page entries to update
+ * @incr: increase next addr by incr bytes
+ *
+ * Update PTEs by writing them manually using sDMA.
+ */
+static void sdma_v4_4_2_vm_write_pte(struct amdgpu_ib *ib, uint64_t pe,
+				   uint64_t value, unsigned count,
+				   uint32_t incr)
+{
+	unsigned ndw = count * 2;
+
+	ib->ptr[ib->length_dw++] = SDMA_PKT_HEADER_OP(SDMA_OP_WRITE) |
+		SDMA_PKT_HEADER_SUB_OP(SDMA_SUBOP_WRITE_LINEAR);
+	ib->ptr[ib->length_dw++] = lower_32_bits(pe);
+	ib->ptr[ib->length_dw++] = upper_32_bits(pe);
+	ib->ptr[ib->length_dw++] = ndw - 1;
+	for (; ndw > 0; ndw -= 2) {
+		ib->ptr[ib->length_dw++] = lower_32_bits(value);
+		ib->ptr[ib->length_dw++] = upper_32_bits(value);
+		value += incr;
+	}
+}
+
+/**
+ * sdma_v4_4_2_vm_set_pte_pde - update the page tables using sDMA
+ *
+ * @ib: indirect buffer to fill with commands
+ * @pe: addr of the page entry
+ * @addr: dst addr to write into pe
+ * @count: number of page entries to update
+ * @incr: increase next addr by incr bytes
+ * @flags: access flags
+ *
+ * Update the page tables using sDMA.
+ */
+static void sdma_v4_4_2_vm_set_pte_pde(struct amdgpu_ib *ib,
+				     uint64_t pe,
+				     uint64_t addr, unsigned count,
+				     uint32_t incr, uint64_t flags)
+{
+	/* for physically contiguous pages (vram) */
+	ib->ptr[ib->length_dw++] = SDMA_PKT_HEADER_OP(SDMA_OP_PTEPDE);
+	ib->ptr[ib->length_dw++] = lower_32_bits(pe); /* dst addr */
+	ib->ptr[ib->length_dw++] = upper_32_bits(pe);
+	ib->ptr[ib->length_dw++] = lower_32_bits(flags); /* mask */
+	ib->ptr[ib->length_dw++] = upper_32_bits(flags);
+	ib->ptr[ib->length_dw++] = lower_32_bits(addr); /* value */
+	ib->ptr[ib->length_dw++] = upper_32_bits(addr);
+	ib->ptr[ib->length_dw++] = incr; /* increment size */
+	ib->ptr[ib->length_dw++] = 0;
+	ib->ptr[ib->length_dw++] = count - 1; /* number of entries */
+}
+
+/**
+ * sdma_v4_4_2_ring_pad_ib - pad the IB to the required number of dw
+ *
+ * @ring: amdgpu_ring structure holding ring information
+ * @ib: indirect buffer to fill with padding
+ */
+static void sdma_v4_4_2_ring_pad_ib(struct amdgpu_ring *ring, struct amdgpu_ib *ib)
+{
+	struct amdgpu_sdma_instance *sdma = amdgpu_sdma_get_instance_from_ring(ring);
+	u32 pad_count;
+	int i;
+
+	pad_count = (-ib->length_dw) & 7;
+	for (i = 0; i < pad_count; i++)
+		if (sdma && sdma->burst_nop && (i == 0))
+			ib->ptr[ib->length_dw++] =
+				SDMA_PKT_HEADER_OP(SDMA_OP_NOP) |
+				SDMA_PKT_NOP_HEADER_COUNT(pad_count - 1);
+		else
+			ib->ptr[ib->length_dw++] =
+				SDMA_PKT_HEADER_OP(SDMA_OP_NOP);
+}
+
+
+/**
+ * sdma_v4_4_2_ring_emit_pipeline_sync - sync the pipeline
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Make sure all previous operations are completed (CIK).
+ */
+static void sdma_v4_4_2_ring_emit_pipeline_sync(struct amdgpu_ring *ring)
+{
+	uint32_t seq = ring->fence_drv.sync_seq;
+	uint64_t addr = ring->fence_drv.gpu_addr;
+
+	/* wait for idle */
+	sdma_v4_4_2_wait_reg_mem(ring, 1, 0,
+			       addr & 0xfffffffc,
+			       upper_32_bits(addr) & 0xffffffff,
+			       seq, 0xffffffff, 4);
+}
+
+
+/**
+ * sdma_v4_4_2_ring_emit_vm_flush - vm flush using sDMA
+ *
+ * @ring: amdgpu_ring pointer
+ * @vmid: vmid number to use
+ * @pd_addr: address
+ *
+ * Update the page table base and flush the VM TLB
+ * using sDMA.
+ */
+static void sdma_v4_4_2_ring_emit_vm_flush(struct amdgpu_ring *ring,
+					 unsigned vmid, uint64_t pd_addr)
+{
+	amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pd_addr);
+}
+
+static void sdma_v4_4_2_ring_emit_wreg(struct amdgpu_ring *ring,
+				     uint32_t reg, uint32_t val)
+{
+	amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_SRBM_WRITE) |
+			  SDMA_PKT_SRBM_WRITE_HEADER_BYTE_EN(0xf));
+	amdgpu_ring_write(ring, reg);
+	amdgpu_ring_write(ring, val);
+}
+
+static void sdma_v4_4_2_ring_emit_reg_wait(struct amdgpu_ring *ring, uint32_t reg,
+					 uint32_t val, uint32_t mask)
+{
+	sdma_v4_4_2_wait_reg_mem(ring, 0, 0, reg, 0, val, mask, 10);
+}
+
+static bool sdma_v4_4_2_fw_support_paging_queue(struct amdgpu_device *adev)
+{
+	switch (amdgpu_ip_version(adev, SDMA0_HWIP, 0)) {
+	case IP_VERSION(4, 4, 2):
+	case IP_VERSION(4, 4, 5):
+		return false;
+	default:
+		return false;
+	}
+}
+
+static const struct amdgpu_sdma_funcs sdma_v4_4_2_sdma_funcs = {
+	.stop_kernel_queue = &sdma_v4_4_2_stop_queue,
+	.start_kernel_queue = &sdma_v4_4_2_restore_queue,
+	.soft_reset_kernel_queue = &sdma_v4_4_2_soft_reset_engine,
+};
+
+static int sdma_v4_4_2_early_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int r;
+
+	r = sdma_v4_4_2_init_microcode(adev);
+	if (r)
+		return r;
+
+	/* TODO: Page queue breaks driver reload under SRIOV */
+	if (sdma_v4_4_2_fw_support_paging_queue(adev))
+		adev->sdma.has_page_queue = true;
+
+	sdma_v4_4_2_set_ring_funcs(adev);
+	sdma_v4_4_2_set_buffer_funcs(adev);
+	sdma_v4_4_2_set_vm_pte_funcs(adev);
+	sdma_v4_4_2_set_irq_funcs(adev);
+	sdma_v4_4_2_set_ras_funcs(adev);
+	return 0;
+}
+
+#if 0
+static int sdma_v4_4_2_process_ras_data_cb(struct amdgpu_device *adev,
+		void *err_data,
+		struct amdgpu_iv_entry *entry);
+#endif
+
+static int sdma_v4_4_2_late_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+#if 0
+	struct ras_ih_if ih_info = {
+		.cb = sdma_v4_4_2_process_ras_data_cb,
+	};
+#endif
+	if (!amdgpu_persistent_edc_harvesting_supported(adev))
+		amdgpu_ras_reset_error_count(adev, AMDGPU_RAS_BLOCK__SDMA);
+
+	/* The initialization is done in the late_init stage to ensure that the SMU
+	 * initialization and capability setup are completed before we check the SDMA
+	 * reset capability
+	 */
+	sdma_v4_4_2_update_reset_mask(adev);
+
+	return 0;
+}
+
+static int sdma_v4_4_2_sw_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_ring *ring;
+	int r, i;
+	struct amdgpu_device *adev = ip_block->adev;
+	u32 aid_id;
+	uint32_t reg_count = ARRAY_SIZE(sdma_reg_list_4_4_2);
+	uint32_t *ptr;
+
+	/* SDMA trap event */
+	for (i = 0; i < adev->sdma.num_inst_per_aid; i++) {
+		r = amdgpu_irq_add_id(adev, sdma_v4_4_2_seq_to_irq_id(i),
+				      SDMA0_4_0__SRCID__SDMA_TRAP,
+				      &adev->sdma.trap_irq);
+		if (r)
+			return r;
+	}
+
+	/* SDMA SRAM ECC event */
+	for (i = 0; i < adev->sdma.num_inst_per_aid; i++) {
+		r = amdgpu_irq_add_id(adev, sdma_v4_4_2_seq_to_irq_id(i),
+				      SDMA0_4_0__SRCID__SDMA_SRAM_ECC,
+				      &adev->sdma.ecc_irq);
+		if (r)
+			return r;
+	}
+
+	/* SDMA VM_HOLE/DOORBELL_INV/POLL_TIMEOUT/SRBM_WRITE_PROTECTION event*/
+	for (i = 0; i < adev->sdma.num_inst_per_aid; i++) {
+		r = amdgpu_irq_add_id(adev, sdma_v4_4_2_seq_to_irq_id(i),
+				      SDMA0_4_0__SRCID__SDMA_VM_HOLE,
+				      &adev->sdma.vm_hole_irq);
+		if (r)
+			return r;
+
+		r = amdgpu_irq_add_id(adev, sdma_v4_4_2_seq_to_irq_id(i),
+				      SDMA0_4_0__SRCID__SDMA_DOORBELL_INVALID,
+				      &adev->sdma.doorbell_invalid_irq);
+		if (r)
+			return r;
+
+		r = amdgpu_irq_add_id(adev, sdma_v4_4_2_seq_to_irq_id(i),
+				      SDMA0_4_0__SRCID__SDMA_POLL_TIMEOUT,
+				      &adev->sdma.pool_timeout_irq);
+		if (r)
+			return r;
+
+		r = amdgpu_irq_add_id(adev, sdma_v4_4_2_seq_to_irq_id(i),
+				      SDMA0_4_0__SRCID__SDMA_SRBMWRITE,
+				      &adev->sdma.srbm_write_irq);
+		if (r)
+			return r;
+
+		r = amdgpu_irq_add_id(adev, sdma_v4_4_2_seq_to_irq_id(i),
+				      SDMA0_4_0__SRCID__SDMA_CTXEMPTY,
+				      &adev->sdma.ctxt_empty_irq);
+		if (r)
+			return r;
+	}
+
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		mutex_init(&adev->sdma.instance[i].engine_reset_mutex);
+		/* Initialize guilty flags for GFX and PAGE queues */
+		adev->sdma.instance[i].gfx_guilty = false;
+		adev->sdma.instance[i].page_guilty = false;
+		adev->sdma.instance[i].funcs = &sdma_v4_4_2_sdma_funcs;
+
+		ring = &adev->sdma.instance[i].ring;
+		ring->ring_obj = NULL;
+		ring->use_doorbell = true;
+		aid_id = adev->sdma.instance[i].aid_id;
+
+		DRM_DEBUG("SDMA %d use_doorbell being set to: [%s]\n", i,
+				ring->use_doorbell?"true":"false");
+
+		/* doorbell size is 2 dwords, get DWORD offset */
+		ring->doorbell_index = adev->doorbell_index.sdma_engine[i] << 1;
+		ring->vm_hub = AMDGPU_MMHUB0(aid_id);
+
+		sprintf(ring->name, "sdma%d.%d", aid_id,
+				i % adev->sdma.num_inst_per_aid);
+		r = amdgpu_ring_init(adev, ring, 1024, &adev->sdma.trap_irq,
+				     AMDGPU_SDMA_IRQ_INSTANCE0 + i,
+				     AMDGPU_RING_PRIO_DEFAULT, NULL);
+		if (r)
+			return r;
+
+		if (adev->sdma.has_page_queue) {
+			ring = &adev->sdma.instance[i].page;
+			ring->ring_obj = NULL;
+			ring->use_doorbell = true;
+
+			/* doorbell index of page queue is assigned right after
+			 * gfx queue on the same instance
+			 */
+			ring->doorbell_index =
+				(adev->doorbell_index.sdma_engine[i] + 1) << 1;
+			ring->vm_hub = AMDGPU_MMHUB0(aid_id);
+
+			sprintf(ring->name, "page%d.%d", aid_id,
+					i % adev->sdma.num_inst_per_aid);
+			r = amdgpu_ring_init(adev, ring, 1024,
+					     &adev->sdma.trap_irq,
+					     AMDGPU_SDMA_IRQ_INSTANCE0 + i,
+					     AMDGPU_RING_PRIO_DEFAULT, NULL);
+			if (r)
+				return r;
+		}
+	}
+
+	adev->sdma.supported_reset =
+		amdgpu_get_soft_full_reset_mask(&adev->sdma.instance[0].ring);
+
+	if (amdgpu_sdma_ras_sw_init(adev)) {
+		dev_err(adev->dev, "fail to initialize sdma ras block\n");
+		return -EINVAL;
+	}
+
+	/* Allocate memory for SDMA IP Dump buffer */
+	ptr = kcalloc(adev->sdma.num_instances * reg_count, sizeof(uint32_t), GFP_KERNEL);
+	if (ptr)
+		adev->sdma.ip_dump = ptr;
+	else
+		DRM_ERROR("Failed to allocated memory for SDMA IP Dump\n");
+
+	r = amdgpu_sdma_sysfs_reset_mask_init(adev);
+	if (r)
+		return r;
+
+	return r;
+}
+
+static int sdma_v4_4_2_sw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i;
+
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		amdgpu_ring_fini(&adev->sdma.instance[i].ring);
+		if (adev->sdma.has_page_queue)
+			amdgpu_ring_fini(&adev->sdma.instance[i].page);
+	}
+
+	amdgpu_sdma_sysfs_reset_mask_fini(adev);
+	if (amdgpu_ip_version(adev, SDMA0_HWIP, 0) == IP_VERSION(4, 4, 2) ||
+	    amdgpu_ip_version(adev, SDMA0_HWIP, 0) == IP_VERSION(4, 4, 4) ||
+	    amdgpu_ip_version(adev, SDMA0_HWIP, 0) == IP_VERSION(4, 4, 5))
+		amdgpu_sdma_destroy_inst_ctx(adev, true);
+	else
+		amdgpu_sdma_destroy_inst_ctx(adev, false);
+
+	kfree(adev->sdma.ip_dump);
+
+	return 0;
+}
+
+static int sdma_v4_4_2_hw_init(struct amdgpu_ip_block *ip_block)
+{
+	int r;
+	struct amdgpu_device *adev = ip_block->adev;
+	uint32_t inst_mask;
+
+	inst_mask = GENMASK(adev->sdma.num_instances - 1, 0);
+	if (!amdgpu_sriov_vf(adev))
+		sdma_v4_4_2_inst_init_golden_registers(adev, inst_mask);
+
+	r = sdma_v4_4_2_inst_start(adev, inst_mask, false);
+
+	return r;
+}
+
+static int sdma_v4_4_2_hw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	uint32_t inst_mask;
+	int i;
+
+	if (amdgpu_sriov_vf(adev))
+		return 0;
+
+	inst_mask = GENMASK(adev->sdma.num_instances - 1, 0);
+	if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__SDMA)) {
+		for (i = 0; i < adev->sdma.num_instances; i++) {
+			amdgpu_irq_put(adev, &adev->sdma.ecc_irq,
+				       AMDGPU_SDMA_IRQ_INSTANCE0 + i);
+		}
+	}
+
+	sdma_v4_4_2_inst_ctx_switch_enable(adev, false, inst_mask);
+	sdma_v4_4_2_inst_enable(adev, false, inst_mask);
+
+	return 0;
+}
+
+static int sdma_v4_4_2_set_clockgating_state(struct amdgpu_ip_block *ip_block,
+					     enum amd_clockgating_state state);
+
+static int sdma_v4_4_2_suspend(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	if (amdgpu_in_reset(adev))
+		sdma_v4_4_2_set_clockgating_state(ip_block, AMD_CG_STATE_UNGATE);
+
+	return sdma_v4_4_2_hw_fini(ip_block);
+}
+
+static int sdma_v4_4_2_resume(struct amdgpu_ip_block *ip_block)
+{
+	return sdma_v4_4_2_hw_init(ip_block);
+}
+
+static bool sdma_v4_4_2_is_idle(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	u32 i;
+
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		u32 tmp = RREG32_SDMA(i, regSDMA_STATUS_REG);
+
+		if (!(tmp & SDMA_STATUS_REG__IDLE_MASK))
+			return false;
+	}
+
+	return true;
+}
+
+static int sdma_v4_4_2_wait_for_idle(struct amdgpu_ip_block *ip_block)
+{
+	unsigned i, j;
+	u32 sdma[AMDGPU_MAX_SDMA_INSTANCES];
+	struct amdgpu_device *adev = ip_block->adev;
+
+	for (i = 0; i < adev->usec_timeout; i++) {
+		for (j = 0; j < adev->sdma.num_instances; j++) {
+			sdma[j] = RREG32_SDMA(j, regSDMA_STATUS_REG);
+			if (!(sdma[j] & SDMA_STATUS_REG__IDLE_MASK))
+				break;
+		}
+		if (j == adev->sdma.num_instances)
+			return 0;
+		udelay(1);
+	}
+	return -ETIMEDOUT;
+}
+
+static int sdma_v4_4_2_soft_reset(struct amdgpu_ip_block *ip_block)
+{
+	/* todo */
+
+	return 0;
+}
+
+static bool sdma_v4_4_2_is_queue_selected(struct amdgpu_device *adev, uint32_t instance_id, bool is_page_queue)
+{
+	uint32_t reg_offset = is_page_queue ? regSDMA_PAGE_CONTEXT_STATUS : regSDMA_GFX_CONTEXT_STATUS;
+	uint32_t context_status = RREG32(sdma_v4_4_2_get_reg_offset(adev, instance_id, reg_offset));
+
+	/* Check if the SELECTED bit is set */
+	return (context_status & SDMA_GFX_CONTEXT_STATUS__SELECTED_MASK) != 0;
+}
+
+static int sdma_v4_4_2_reset_queue(struct amdgpu_ring *ring,
+				   unsigned int vmid,
+				   struct amdgpu_fence *timedout_fence)
+{
+	struct amdgpu_device *adev = ring->adev;
+	u32 id = ring->me;
+	int r;
+
+	amdgpu_amdkfd_suspend(adev, true);
+	r = amdgpu_sdma_reset_engine(adev, id, false);
+	amdgpu_amdkfd_resume(adev, true);
+	return r;
+}
+
+static int sdma_v4_4_2_stop_queue(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	u32 instance_id = ring->me;
+	u32 inst_mask;
+	uint64_t rptr;
+
+	if (amdgpu_sriov_vf(adev))
+		return -EINVAL;
+
+	/* Check if this queue is the guilty one */
+	adev->sdma.instance[instance_id].gfx_guilty =
+		sdma_v4_4_2_is_queue_selected(adev, instance_id, false);
+	if (adev->sdma.has_page_queue)
+		adev->sdma.instance[instance_id].page_guilty =
+			sdma_v4_4_2_is_queue_selected(adev, instance_id, true);
+
+	/* Cache the rptr before reset, after the reset,
+	* all of the registers will be reset to 0
+	*/
+	rptr = amdgpu_ring_get_rptr(ring);
+	ring->cached_rptr = rptr;
+	/* Cache the rptr for the page queue if it exists */
+	if (adev->sdma.has_page_queue) {
+		struct amdgpu_ring *page_ring = &adev->sdma.instance[instance_id].page;
+		rptr = amdgpu_ring_get_rptr(page_ring);
+		page_ring->cached_rptr = rptr;
+	}
+
+	/* stop queue */
+	inst_mask = 1 << ring->me;
+	sdma_v4_4_2_inst_gfx_stop(adev, inst_mask);
+	if (adev->sdma.has_page_queue)
+		sdma_v4_4_2_inst_page_stop(adev, inst_mask);
+
+	return 0;
+}
+
+static int sdma_v4_4_2_restore_queue(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	u32 inst_mask;
+	int i, r;
+
+	inst_mask = 1 << ring->me;
+	udelay(50);
+
+	for (i = 0; i < adev->usec_timeout; i++) {
+		if (!REG_GET_FIELD(RREG32_SDMA(ring->me, regSDMA_F32_CNTL), SDMA_F32_CNTL, HALT))
+			break;
+		udelay(1);
+	}
+
+	if (i == adev->usec_timeout) {
+		dev_err(adev->dev, "timed out waiting for SDMA%d unhalt after reset\n",
+			ring->me);
+		return -ETIMEDOUT;
+	}
+
+	r = sdma_v4_4_2_inst_start(adev, inst_mask, true);
+
+	return r;
+}
+
+static int sdma_v4_4_2_soft_reset_engine(struct amdgpu_device *adev,
+					 u32 instance_id)
+{
+	/* For SDMA 4.x, use the existing DPM interface for backward compatibility
+	 * we need to convert the logical instance ID to physical instance ID before reset.
+	 */
+	return amdgpu_dpm_reset_sdma(adev, 1 << GET_INST(SDMA0, instance_id));
+}
+
+static int sdma_v4_4_2_set_trap_irq_state(struct amdgpu_device *adev,
+					struct amdgpu_irq_src *source,
+					unsigned type,
+					enum amdgpu_interrupt_state state)
+{
+	u32 sdma_cntl;
+
+	sdma_cntl = RREG32_SDMA(type, regSDMA_CNTL);
+	sdma_cntl = REG_SET_FIELD(sdma_cntl, SDMA_CNTL, TRAP_ENABLE,
+		       state == AMDGPU_IRQ_STATE_ENABLE ? 1 : 0);
+	WREG32_SDMA(type, regSDMA_CNTL, sdma_cntl);
+
+	return 0;
+}
+
+static int sdma_v4_4_2_process_trap_irq(struct amdgpu_device *adev,
+				      struct amdgpu_irq_src *source,
+				      struct amdgpu_iv_entry *entry)
+{
+	uint32_t instance, i;
+
+	DRM_DEBUG("IH: SDMA trap\n");
+	instance = sdma_v4_4_2_irq_id_to_seq(adev, entry->client_id);
+
+	/* Client id gives the SDMA instance in AID. To know the exact SDMA
+	 * instance, interrupt entry gives the node id which corresponds to the AID instance.
+	 * Match node id with the AID id associated with the SDMA instance. */
+	for (i = instance; i < adev->sdma.num_instances;
+	     i += adev->sdma.num_inst_per_aid) {
+		if (adev->sdma.instance[i].aid_id ==
+		    node_id_to_phys_map[entry->node_id])
+			break;
+	}
+
+	if (i >= adev->sdma.num_instances) {
+		dev_WARN_ONCE(
+			adev->dev, 1,
+			"Couldn't find the right sdma instance in trap handler");
+		return 0;
+	}
+
+	switch (entry->ring_id) {
+	case 0:
+		amdgpu_fence_process(&adev->sdma.instance[i].ring);
+		break;
+	case 1:
+		amdgpu_fence_process(&adev->sdma.instance[i].page);
+		break;
+	default:
+		break;
+	}
+	return 0;
+}
+
+#if 0
+static int sdma_v4_4_2_process_ras_data_cb(struct amdgpu_device *adev,
+		void *err_data,
+		struct amdgpu_iv_entry *entry)
+{
+	int instance;
+
+	/* When “Full RAS” is enabled, the per-IP interrupt sources should
+	 * be disabled and the driver should only look for the aggregated
+	 * interrupt via sync flood
+	 */
+	if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__SDMA))
+		goto out;
+
+	instance = sdma_v4_4_2_irq_id_to_seq(adev, entry->client_id);
+	if (instance < 0)
+		goto out;
+
+	amdgpu_sdma_process_ras_data_cb(adev, err_data, entry);
+
+out:
+	return AMDGPU_RAS_SUCCESS;
+}
+#endif
+
+static int sdma_v4_4_2_process_illegal_inst_irq(struct amdgpu_device *adev,
+					      struct amdgpu_irq_src *source,
+					      struct amdgpu_iv_entry *entry)
+{
+	int instance;
+
+	DRM_ERROR("Illegal instruction in SDMA command stream\n");
+
+	instance = sdma_v4_4_2_irq_id_to_seq(adev, entry->client_id);
+	if (instance < 0)
+		return 0;
+
+	switch (entry->ring_id) {
+	case 0:
+		drm_sched_fault(&adev->sdma.instance[instance].ring.sched);
+		break;
+	}
+	return 0;
+}
+
+static int sdma_v4_4_2_set_ecc_irq_state(struct amdgpu_device *adev,
+					struct amdgpu_irq_src *source,
+					unsigned type,
+					enum amdgpu_interrupt_state state)
+{
+	u32 sdma_cntl;
+
+	sdma_cntl = RREG32_SDMA(type, regSDMA_CNTL);
+	sdma_cntl = REG_SET_FIELD(sdma_cntl, SDMA_CNTL, DRAM_ECC_INT_ENABLE,
+					state == AMDGPU_IRQ_STATE_ENABLE ? 1 : 0);
+	WREG32_SDMA(type, regSDMA_CNTL, sdma_cntl);
+
+	return 0;
+}
+
+static int sdma_v4_4_2_print_iv_entry(struct amdgpu_device *adev,
+					      struct amdgpu_iv_entry *entry)
+{
+	int instance;
+	struct amdgpu_task_info *task_info;
+	u64 addr;
+
+	instance = sdma_v4_4_2_irq_id_to_seq(adev, entry->client_id);
+	if (instance < 0 || instance >= adev->sdma.num_instances) {
+		dev_err(adev->dev, "sdma instance invalid %d\n", instance);
+		return -EINVAL;
+	}
+
+	addr = (u64)entry->src_data[0] << 12;
+	addr |= ((u64)entry->src_data[1] & 0xf) << 44;
+
+	dev_dbg_ratelimited(adev->dev,
+			    "[sdma%d] address:0x%016llx src_id:%u ring:%u vmid:%u pasid:%u\n",
+			    instance, addr, entry->src_id, entry->ring_id, entry->vmid,
+			    entry->pasid);
+
+	task_info = amdgpu_vm_get_task_info_pasid(adev, entry->pasid);
+	if (task_info) {
+		dev_dbg_ratelimited(adev->dev, " for process %s pid %d thread %s pid %d\n",
+				    task_info->process_name, task_info->tgid,
+				    task_info->task.comm, task_info->task.pid);
+		amdgpu_vm_put_task_info(task_info);
+	}
+
+	return 0;
+}
+
+static int sdma_v4_4_2_process_vm_hole_irq(struct amdgpu_device *adev,
+					      struct amdgpu_irq_src *source,
+					      struct amdgpu_iv_entry *entry)
+{
+	dev_dbg_ratelimited(adev->dev, "MC or SEM address in VM hole\n");
+	sdma_v4_4_2_print_iv_entry(adev, entry);
+	return 0;
+}
+
+static int sdma_v4_4_2_process_doorbell_invalid_irq(struct amdgpu_device *adev,
+					      struct amdgpu_irq_src *source,
+					      struct amdgpu_iv_entry *entry)
+{
+
+	dev_dbg_ratelimited(adev->dev, "SDMA received a doorbell from BIF with byte_enable !=0xff\n");
+	sdma_v4_4_2_print_iv_entry(adev, entry);
+	return 0;
+}
+
+static int sdma_v4_4_2_process_pool_timeout_irq(struct amdgpu_device *adev,
+					      struct amdgpu_irq_src *source,
+					      struct amdgpu_iv_entry *entry)
+{
+	dev_dbg_ratelimited(adev->dev,
+		"Polling register/memory timeout executing POLL_REG/MEM with finite timer\n");
+	sdma_v4_4_2_print_iv_entry(adev, entry);
+	return 0;
+}
+
+static int sdma_v4_4_2_process_srbm_write_irq(struct amdgpu_device *adev,
+					      struct amdgpu_irq_src *source,
+					      struct amdgpu_iv_entry *entry)
+{
+	dev_dbg_ratelimited(adev->dev,
+		"SDMA gets an Register Write SRBM_WRITE command in non-privilege command buffer\n");
+	sdma_v4_4_2_print_iv_entry(adev, entry);
+	return 0;
+}
+
+static int sdma_v4_4_2_process_ctxt_empty_irq(struct amdgpu_device *adev,
+					      struct amdgpu_irq_src *source,
+					      struct amdgpu_iv_entry *entry)
+{
+	/* There is nothing useful to be done here, only kept for debug */
+	dev_dbg_ratelimited(adev->dev, "SDMA context empty interrupt");
+	sdma_v4_4_2_print_iv_entry(adev, entry);
+	return 0;
+}
+
+static void sdma_v4_4_2_inst_update_medium_grain_light_sleep(
+	struct amdgpu_device *adev, bool enable, uint32_t inst_mask)
+{
+	uint32_t data, def;
+	int i;
+
+	/* leave as default if it is not driver controlled */
+	if (!(adev->cg_flags & AMD_CG_SUPPORT_SDMA_LS))
+		return;
+
+	if (enable) {
+		for_each_inst(i, inst_mask) {
+			/* 1-not override: enable sdma mem light sleep */
+			def = data = RREG32_SDMA(i, regSDMA_POWER_CNTL);
+			data |= SDMA_POWER_CNTL__MEM_POWER_OVERRIDE_MASK;
+			if (def != data)
+				WREG32_SDMA(i, regSDMA_POWER_CNTL, data);
+		}
+	} else {
+		for_each_inst(i, inst_mask) {
+			/* 0-override:disable sdma mem light sleep */
+			def = data = RREG32_SDMA(i, regSDMA_POWER_CNTL);
+			data &= ~SDMA_POWER_CNTL__MEM_POWER_OVERRIDE_MASK;
+			if (def != data)
+				WREG32_SDMA(i, regSDMA_POWER_CNTL, data);
+		}
+	}
+}
+
+static void sdma_v4_4_2_inst_update_medium_grain_clock_gating(
+	struct amdgpu_device *adev, bool enable, uint32_t inst_mask)
+{
+	uint32_t data, def;
+	int i;
+
+	/* leave as default if it is not driver controlled */
+	if (!(adev->cg_flags & AMD_CG_SUPPORT_SDMA_MGCG))
+		return;
+
+	if (enable) {
+		for_each_inst(i, inst_mask) {
+			def = data = RREG32_SDMA(i, regSDMA_CLK_CTRL);
+			data &= ~(SDMA_CLK_CTRL__SOFT_OVERRIDE5_MASK |
+				  SDMA_CLK_CTRL__SOFT_OVERRIDE4_MASK |
+				  SDMA_CLK_CTRL__SOFT_OVERRIDE3_MASK |
+				  SDMA_CLK_CTRL__SOFT_OVERRIDE2_MASK |
+				  SDMA_CLK_CTRL__SOFT_OVERRIDE1_MASK |
+				  SDMA_CLK_CTRL__SOFT_OVERRIDE0_MASK);
+			if (def != data)
+				WREG32_SDMA(i, regSDMA_CLK_CTRL, data);
+		}
+	} else {
+		for_each_inst(i, inst_mask) {
+			def = data = RREG32_SDMA(i, regSDMA_CLK_CTRL);
+			data |= (SDMA_CLK_CTRL__SOFT_OVERRIDE5_MASK |
+				 SDMA_CLK_CTRL__SOFT_OVERRIDE4_MASK |
+				 SDMA_CLK_CTRL__SOFT_OVERRIDE3_MASK |
+				 SDMA_CLK_CTRL__SOFT_OVERRIDE2_MASK |
+				 SDMA_CLK_CTRL__SOFT_OVERRIDE1_MASK |
+				 SDMA_CLK_CTRL__SOFT_OVERRIDE0_MASK);
+			if (def != data)
+				WREG32_SDMA(i, regSDMA_CLK_CTRL, data);
+		}
+	}
+}
+
+static int sdma_v4_4_2_set_clockgating_state(struct amdgpu_ip_block *ip_block,
+					  enum amd_clockgating_state state)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	uint32_t inst_mask;
+
+	if (amdgpu_sriov_vf(adev))
+		return 0;
+
+	inst_mask = GENMASK(adev->sdma.num_instances - 1, 0);
+
+	sdma_v4_4_2_inst_update_medium_grain_clock_gating(
+		adev, state == AMD_CG_STATE_GATE, inst_mask);
+	sdma_v4_4_2_inst_update_medium_grain_light_sleep(
+		adev, state == AMD_CG_STATE_GATE, inst_mask);
+	return 0;
+}
+
+static int sdma_v4_4_2_set_powergating_state(struct amdgpu_ip_block *ip_block,
+					  enum amd_powergating_state state)
+{
+	return 0;
+}
+
+static void sdma_v4_4_2_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int data;
+
+	if (amdgpu_sriov_vf(adev))
+		*flags = 0;
+
+	/* AMD_CG_SUPPORT_SDMA_MGCG */
+	data = RREG32(SOC15_REG_OFFSET(SDMA0, GET_INST(SDMA0, 0), regSDMA_CLK_CTRL));
+	if (!(data & SDMA_CLK_CTRL__SOFT_OVERRIDE5_MASK))
+		*flags |= AMD_CG_SUPPORT_SDMA_MGCG;
+
+	/* AMD_CG_SUPPORT_SDMA_LS */
+	data = RREG32(SOC15_REG_OFFSET(SDMA0, GET_INST(SDMA0, 0), regSDMA_POWER_CNTL));
+	if (data & SDMA_POWER_CNTL__MEM_POWER_OVERRIDE_MASK)
+		*flags |= AMD_CG_SUPPORT_SDMA_LS;
+}
+
+static void sdma_v4_4_2_print_ip_state(struct amdgpu_ip_block *ip_block, struct drm_printer *p)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, j;
+	uint32_t reg_count = ARRAY_SIZE(sdma_reg_list_4_4_2);
+	uint32_t instance_offset;
+
+	if (!adev->sdma.ip_dump)
+		return;
+
+	drm_printf(p, "num_instances:%d\n", adev->sdma.num_instances);
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		instance_offset = i * reg_count;
+		drm_printf(p, "\nInstance:%d\n", i);
+
+		for (j = 0; j < reg_count; j++)
+			drm_printf(p, "%-50s \t 0x%08x\n", sdma_reg_list_4_4_2[j].reg_name,
+				   adev->sdma.ip_dump[instance_offset + j]);
+	}
+}
+
+static void sdma_v4_4_2_dump_ip_state(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, j;
+	uint32_t instance_offset;
+	uint32_t reg_count = ARRAY_SIZE(sdma_reg_list_4_4_2);
+
+	if (!adev->sdma.ip_dump)
+		return;
+
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		instance_offset = i * reg_count;
+		for (j = 0; j < reg_count; j++)
+			adev->sdma.ip_dump[instance_offset + j] =
+				RREG32(sdma_v4_4_2_get_reg_offset(adev, i,
+				       sdma_reg_list_4_4_2[j].reg_offset));
+	}
+}
+
+const struct amd_ip_funcs sdma_v4_4_2_ip_funcs = {
+	.name = "sdma_v4_4_2",
+	.early_init = sdma_v4_4_2_early_init,
+	.late_init = sdma_v4_4_2_late_init,
+	.sw_init = sdma_v4_4_2_sw_init,
+	.sw_fini = sdma_v4_4_2_sw_fini,
+	.hw_init = sdma_v4_4_2_hw_init,
+	.hw_fini = sdma_v4_4_2_hw_fini,
+	.suspend = sdma_v4_4_2_suspend,
+	.resume = sdma_v4_4_2_resume,
+	.is_idle = sdma_v4_4_2_is_idle,
+	.wait_for_idle = sdma_v4_4_2_wait_for_idle,
+	.soft_reset = sdma_v4_4_2_soft_reset,
+	.set_clockgating_state = sdma_v4_4_2_set_clockgating_state,
+	.set_powergating_state = sdma_v4_4_2_set_powergating_state,
+	.get_clockgating_state = sdma_v4_4_2_get_clockgating_state,
+	.dump_ip_state = sdma_v4_4_2_dump_ip_state,
+	.print_ip_state = sdma_v4_4_2_print_ip_state,
+};
+
+static const struct amdgpu_ring_funcs sdma_v4_4_2_ring_funcs = {
+	.type = AMDGPU_RING_TYPE_SDMA,
+	.align_mask = 0xff,
+	.nop = SDMA_PKT_NOP_HEADER_OP(SDMA_OP_NOP),
+	.support_64bit_ptrs = true,
+	.get_rptr = sdma_v4_4_2_ring_get_rptr,
+	.get_wptr = sdma_v4_4_2_ring_get_wptr,
+	.set_wptr = sdma_v4_4_2_ring_set_wptr,
+	.emit_frame_size =
+		6 + /* sdma_v4_4_2_ring_emit_hdp_flush */
+		3 + /* hdp invalidate */
+		6 + /* sdma_v4_4_2_ring_emit_pipeline_sync */
+		/* sdma_v4_4_2_ring_emit_vm_flush */
+		SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 +
+		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 6 +
+		10 + 10 + 10, /* sdma_v4_4_2_ring_emit_fence x3 for user fence, vm fence */
+	.emit_ib_size = 7 + 6, /* sdma_v4_4_2_ring_emit_ib */
+	.emit_ib = sdma_v4_4_2_ring_emit_ib,
+	.emit_fence = sdma_v4_4_2_ring_emit_fence,
+	.emit_pipeline_sync = sdma_v4_4_2_ring_emit_pipeline_sync,
+	.emit_vm_flush = sdma_v4_4_2_ring_emit_vm_flush,
+	.emit_hdp_flush = sdma_v4_4_2_ring_emit_hdp_flush,
+	.test_ring = sdma_v4_4_2_ring_test_ring,
+	.test_ib = sdma_v4_4_2_ring_test_ib,
+	.insert_nop = sdma_v4_4_2_ring_insert_nop,
+	.pad_ib = sdma_v4_4_2_ring_pad_ib,
+	.emit_wreg = sdma_v4_4_2_ring_emit_wreg,
+	.emit_reg_wait = sdma_v4_4_2_ring_emit_reg_wait,
+	.emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper,
+	.reset = sdma_v4_4_2_reset_queue,
+};
+
+static const struct amdgpu_ring_funcs sdma_v4_4_2_page_ring_funcs = {
+	.type = AMDGPU_RING_TYPE_SDMA,
+	.align_mask = 0xff,
+	.nop = SDMA_PKT_NOP_HEADER_OP(SDMA_OP_NOP),
+	.support_64bit_ptrs = true,
+	.get_rptr = sdma_v4_4_2_ring_get_rptr,
+	.get_wptr = sdma_v4_4_2_page_ring_get_wptr,
+	.set_wptr = sdma_v4_4_2_page_ring_set_wptr,
+	.emit_frame_size =
+		6 + /* sdma_v4_4_2_ring_emit_hdp_flush */
+		3 + /* hdp invalidate */
+		6 + /* sdma_v4_4_2_ring_emit_pipeline_sync */
+		/* sdma_v4_4_2_ring_emit_vm_flush */
+		SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 +
+		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 6 +
+		10 + 10 + 10, /* sdma_v4_4_2_ring_emit_fence x3 for user fence, vm fence */
+	.emit_ib_size = 7 + 6, /* sdma_v4_4_2_ring_emit_ib */
+	.emit_ib = sdma_v4_4_2_ring_emit_ib,
+	.emit_fence = sdma_v4_4_2_ring_emit_fence,
+	.emit_pipeline_sync = sdma_v4_4_2_ring_emit_pipeline_sync,
+	.emit_vm_flush = sdma_v4_4_2_ring_emit_vm_flush,
+	.emit_hdp_flush = sdma_v4_4_2_ring_emit_hdp_flush,
+	.test_ring = sdma_v4_4_2_ring_test_ring,
+	.test_ib = sdma_v4_4_2_ring_test_ib,
+	.insert_nop = sdma_v4_4_2_ring_insert_nop,
+	.pad_ib = sdma_v4_4_2_ring_pad_ib,
+	.emit_wreg = sdma_v4_4_2_ring_emit_wreg,
+	.emit_reg_wait = sdma_v4_4_2_ring_emit_reg_wait,
+	.emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper,
+	.reset = sdma_v4_4_2_reset_queue,
+};
+
+static void sdma_v4_4_2_set_ring_funcs(struct amdgpu_device *adev)
+{
+	int i, dev_inst;
+
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		adev->sdma.instance[i].ring.funcs = &sdma_v4_4_2_ring_funcs;
+		adev->sdma.instance[i].ring.me = i;
+		if (adev->sdma.has_page_queue) {
+			adev->sdma.instance[i].page.funcs =
+				&sdma_v4_4_2_page_ring_funcs;
+			adev->sdma.instance[i].page.me = i;
+		}
+
+		dev_inst = GET_INST(SDMA0, i);
+		/* AID to which SDMA belongs depends on physical instance */
+		adev->sdma.instance[i].aid_id =
+			dev_inst / adev->sdma.num_inst_per_aid;
+	}
+}
+
+static const struct amdgpu_irq_src_funcs sdma_v4_4_2_trap_irq_funcs = {
+	.set = sdma_v4_4_2_set_trap_irq_state,
+	.process = sdma_v4_4_2_process_trap_irq,
+};
+
+static const struct amdgpu_irq_src_funcs sdma_v4_4_2_illegal_inst_irq_funcs = {
+	.process = sdma_v4_4_2_process_illegal_inst_irq,
+};
+
+static const struct amdgpu_irq_src_funcs sdma_v4_4_2_ecc_irq_funcs = {
+	.set = sdma_v4_4_2_set_ecc_irq_state,
+	.process = amdgpu_sdma_process_ecc_irq,
+};
+
+static const struct amdgpu_irq_src_funcs sdma_v4_4_2_vm_hole_irq_funcs = {
+	.process = sdma_v4_4_2_process_vm_hole_irq,
+};
+
+static const struct amdgpu_irq_src_funcs sdma_v4_4_2_doorbell_invalid_irq_funcs = {
+	.process = sdma_v4_4_2_process_doorbell_invalid_irq,
+};
+
+static const struct amdgpu_irq_src_funcs sdma_v4_4_2_pool_timeout_irq_funcs = {
+	.process = sdma_v4_4_2_process_pool_timeout_irq,
+};
+
+static const struct amdgpu_irq_src_funcs sdma_v4_4_2_srbm_write_irq_funcs = {
+	.process = sdma_v4_4_2_process_srbm_write_irq,
+};
+
+static const struct amdgpu_irq_src_funcs sdma_v4_4_2_ctxt_empty_irq_funcs = {
+	.process = sdma_v4_4_2_process_ctxt_empty_irq,
+};
+
+static void sdma_v4_4_2_set_irq_funcs(struct amdgpu_device *adev)
+{
+	adev->sdma.trap_irq.num_types = adev->sdma.num_instances;
+	adev->sdma.ecc_irq.num_types = adev->sdma.num_instances;
+	adev->sdma.vm_hole_irq.num_types = adev->sdma.num_instances;
+	adev->sdma.doorbell_invalid_irq.num_types = adev->sdma.num_instances;
+	adev->sdma.pool_timeout_irq.num_types = adev->sdma.num_instances;
+	adev->sdma.srbm_write_irq.num_types = adev->sdma.num_instances;
+	adev->sdma.ctxt_empty_irq.num_types = adev->sdma.num_instances;
+
+	adev->sdma.trap_irq.funcs = &sdma_v4_4_2_trap_irq_funcs;
+	adev->sdma.illegal_inst_irq.funcs = &sdma_v4_4_2_illegal_inst_irq_funcs;
+	adev->sdma.ecc_irq.funcs = &sdma_v4_4_2_ecc_irq_funcs;
+	adev->sdma.vm_hole_irq.funcs = &sdma_v4_4_2_vm_hole_irq_funcs;
+	adev->sdma.doorbell_invalid_irq.funcs = &sdma_v4_4_2_doorbell_invalid_irq_funcs;
+	adev->sdma.pool_timeout_irq.funcs = &sdma_v4_4_2_pool_timeout_irq_funcs;
+	adev->sdma.srbm_write_irq.funcs = &sdma_v4_4_2_srbm_write_irq_funcs;
+	adev->sdma.ctxt_empty_irq.funcs = &sdma_v4_4_2_ctxt_empty_irq_funcs;
+}
+
+/**
+ * sdma_v4_4_2_emit_copy_buffer - copy buffer using the sDMA engine
+ *
+ * @ib: indirect buffer to copy to
+ * @src_offset: src GPU address
+ * @dst_offset: dst GPU address
+ * @byte_count: number of bytes to xfer
+ * @copy_flags: copy flags for the buffers
+ *
+ * Copy GPU buffers using the DMA engine.
+ * Used by the amdgpu ttm implementation to move pages if
+ * registered as the asic copy callback.
+ */
+static void sdma_v4_4_2_emit_copy_buffer(struct amdgpu_ib *ib,
+				       uint64_t src_offset,
+				       uint64_t dst_offset,
+				       uint32_t byte_count,
+				       uint32_t copy_flags)
+{
+	ib->ptr[ib->length_dw++] = SDMA_PKT_HEADER_OP(SDMA_OP_COPY) |
+		SDMA_PKT_HEADER_SUB_OP(SDMA_SUBOP_COPY_LINEAR) |
+		SDMA_PKT_COPY_LINEAR_HEADER_TMZ((copy_flags & AMDGPU_COPY_FLAGS_TMZ) ? 1 : 0);
+	ib->ptr[ib->length_dw++] = byte_count - 1;
+	ib->ptr[ib->length_dw++] = 0; /* src/dst endian swap */
+	ib->ptr[ib->length_dw++] = lower_32_bits(src_offset);
+	ib->ptr[ib->length_dw++] = upper_32_bits(src_offset);
+	ib->ptr[ib->length_dw++] = lower_32_bits(dst_offset);
+	ib->ptr[ib->length_dw++] = upper_32_bits(dst_offset);
+}
+
+/**
+ * sdma_v4_4_2_emit_fill_buffer - fill buffer using the sDMA engine
+ *
+ * @ib: indirect buffer to copy to
+ * @src_data: value to write to buffer
+ * @dst_offset: dst GPU address
+ * @byte_count: number of bytes to xfer
+ *
+ * Fill GPU buffers using the DMA engine.
+ */
+static void sdma_v4_4_2_emit_fill_buffer(struct amdgpu_ib *ib,
+				       uint32_t src_data,
+				       uint64_t dst_offset,
+				       uint32_t byte_count)
+{
+	ib->ptr[ib->length_dw++] = SDMA_PKT_HEADER_OP(SDMA_OP_CONST_FILL);
+	ib->ptr[ib->length_dw++] = lower_32_bits(dst_offset);
+	ib->ptr[ib->length_dw++] = upper_32_bits(dst_offset);
+	ib->ptr[ib->length_dw++] = src_data;
+	ib->ptr[ib->length_dw++] = byte_count - 1;
+}
+
+static const struct amdgpu_buffer_funcs sdma_v4_4_2_buffer_funcs = {
+	.copy_max_bytes = 0x400000,
+	.copy_num_dw = 7,
+	.emit_copy_buffer = sdma_v4_4_2_emit_copy_buffer,
+
+	.fill_max_bytes = 0x400000,
+	.fill_num_dw = 5,
+	.emit_fill_buffer = sdma_v4_4_2_emit_fill_buffer,
+};
+
+static void sdma_v4_4_2_set_buffer_funcs(struct amdgpu_device *adev)
+{
+	adev->mman.buffer_funcs = &sdma_v4_4_2_buffer_funcs;
+	if (adev->sdma.has_page_queue)
+		adev->mman.buffer_funcs_ring = &adev->sdma.instance[0].page;
+	else
+		adev->mman.buffer_funcs_ring = &adev->sdma.instance[0].ring;
+}
+
+static const struct amdgpu_vm_pte_funcs sdma_v4_4_2_vm_pte_funcs = {
+	.copy_pte_num_dw = 7,
+	.copy_pte = sdma_v4_4_2_vm_copy_pte,
+
+	.write_pte = sdma_v4_4_2_vm_write_pte,
+	.set_pte_pde = sdma_v4_4_2_vm_set_pte_pde,
+};
+
+static void sdma_v4_4_2_set_vm_pte_funcs(struct amdgpu_device *adev)
+{
+	struct drm_gpu_scheduler *sched;
+	unsigned i;
+
+	adev->vm_manager.vm_pte_funcs = &sdma_v4_4_2_vm_pte_funcs;
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		if (adev->sdma.has_page_queue)
+			sched = &adev->sdma.instance[i].page.sched;
+		else
+			sched = &adev->sdma.instance[i].ring.sched;
+		adev->vm_manager.vm_pte_scheds[i] = sched;
+	}
+	adev->vm_manager.vm_pte_num_scheds = adev->sdma.num_instances;
+}
+
+/**
+ * sdma_v4_4_2_update_reset_mask - update  reset mask for SDMA
+ * @adev: Pointer to the AMDGPU device structure
+ *
+ * This function update reset mask for SDMA and sets the supported
+ * reset types based on the IP version and firmware versions.
+ *
+ */
+static void sdma_v4_4_2_update_reset_mask(struct amdgpu_device *adev)
+{
+	/* per queue reset not supported for SRIOV */
+	if (amdgpu_sriov_vf(adev))
+		return;
+
+	/*
+	 * the user queue relies on MEC fw and pmfw when the sdma queue do reset.
+	 * it needs to check both of them at here to skip old mec and pmfw.
+	 */
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	case IP_VERSION(9, 4, 3):
+	case IP_VERSION(9, 4, 4):
+		if ((adev->gfx.mec_fw_version >= 0xb0) &&
+		    amdgpu_dpm_reset_sdma_is_supported(adev) &&
+		    !adev->debug_disable_gpu_ring_reset)
+			adev->sdma.supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE;
+		break;
+	case IP_VERSION(9, 5, 0):
+		if ((adev->gfx.mec_fw_version >= 0xf) &&
+		    amdgpu_dpm_reset_sdma_is_supported(adev) &&
+		    !adev->debug_disable_gpu_ring_reset)
+			adev->sdma.supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE;
+		break;
+	default:
+		break;
+	}
+
+}
+
+const struct amdgpu_ip_block_version sdma_v4_4_2_ip_block = {
+	.type = AMD_IP_BLOCK_TYPE_SDMA,
+	.major = 4,
+	.minor = 4,
+	.rev = 2,
+	.funcs = &sdma_v4_4_2_ip_funcs,
+};
+
+static int sdma_v4_4_2_xcp_resume(void *handle, uint32_t inst_mask)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	int r;
+
+	if (!amdgpu_sriov_vf(adev))
+		sdma_v4_4_2_inst_init_golden_registers(adev, inst_mask);
+
+	r = sdma_v4_4_2_inst_start(adev, inst_mask, false);
+
+	return r;
+}
+
+static int sdma_v4_4_2_xcp_suspend(void *handle, uint32_t inst_mask)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	uint32_t tmp_mask = inst_mask;
+	int i;
+
+	if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__SDMA)) {
+		for_each_inst(i, tmp_mask) {
+			amdgpu_irq_put(adev, &adev->sdma.ecc_irq,
+				       AMDGPU_SDMA_IRQ_INSTANCE0 + i);
+		}
+	}
+
+	sdma_v4_4_2_inst_ctx_switch_enable(adev, false, inst_mask);
+	sdma_v4_4_2_inst_enable(adev, false, inst_mask);
+
+	return 0;
+}
+
+struct amdgpu_xcp_ip_funcs sdma_v4_4_2_xcp_funcs = {
+	.suspend = &sdma_v4_4_2_xcp_suspend,
+	.resume = &sdma_v4_4_2_xcp_resume
+};
+
+static const struct amdgpu_ras_err_status_reg_entry sdma_v4_2_2_ue_reg_list[] = {
+	{AMDGPU_RAS_REG_ENTRY(SDMA0, 0, regSDMA_UE_ERR_STATUS_LO, regSDMA_UE_ERR_STATUS_HI),
+	1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "SDMA"},
+};
+
+static const struct amdgpu_ras_memory_id_entry sdma_v4_4_2_ras_memory_list[] = {
+	{AMDGPU_SDMA_MBANK_DATA_BUF0, "SDMA_MBANK_DATA_BUF0"},
+	{AMDGPU_SDMA_MBANK_DATA_BUF1, "SDMA_MBANK_DATA_BUF1"},
+	{AMDGPU_SDMA_MBANK_DATA_BUF2, "SDMA_MBANK_DATA_BUF2"},
+	{AMDGPU_SDMA_MBANK_DATA_BUF3, "SDMA_MBANK_DATA_BUF3"},
+	{AMDGPU_SDMA_MBANK_DATA_BUF4, "SDMA_MBANK_DATA_BUF4"},
+	{AMDGPU_SDMA_MBANK_DATA_BUF5, "SDMA_MBANK_DATA_BUF5"},
+	{AMDGPU_SDMA_MBANK_DATA_BUF6, "SDMA_MBANK_DATA_BUF6"},
+	{AMDGPU_SDMA_MBANK_DATA_BUF7, "SDMA_MBANK_DATA_BUF7"},
+	{AMDGPU_SDMA_MBANK_DATA_BUF8, "SDMA_MBANK_DATA_BUF8"},
+	{AMDGPU_SDMA_MBANK_DATA_BUF9, "SDMA_MBANK_DATA_BUF9"},
+	{AMDGPU_SDMA_MBANK_DATA_BUF10, "SDMA_MBANK_DATA_BUF10"},
+	{AMDGPU_SDMA_MBANK_DATA_BUF11, "SDMA_MBANK_DATA_BUF11"},
+	{AMDGPU_SDMA_MBANK_DATA_BUF12, "SDMA_MBANK_DATA_BUF12"},
+	{AMDGPU_SDMA_MBANK_DATA_BUF13, "SDMA_MBANK_DATA_BUF13"},
+	{AMDGPU_SDMA_MBANK_DATA_BUF14, "SDMA_MBANK_DATA_BUF14"},
+	{AMDGPU_SDMA_MBANK_DATA_BUF15, "SDMA_MBANK_DATA_BUF15"},
+	{AMDGPU_SDMA_UCODE_BUF, "SDMA_UCODE_BUF"},
+	{AMDGPU_SDMA_RB_CMD_BUF, "SDMA_RB_CMD_BUF"},
+	{AMDGPU_SDMA_IB_CMD_BUF, "SDMA_IB_CMD_BUF"},
+	{AMDGPU_SDMA_UTCL1_RD_FIFO, "SDMA_UTCL1_RD_FIFO"},
+	{AMDGPU_SDMA_UTCL1_RDBST_FIFO, "SDMA_UTCL1_RDBST_FIFO"},
+	{AMDGPU_SDMA_UTCL1_WR_FIFO, "SDMA_UTCL1_WR_FIFO"},
+	{AMDGPU_SDMA_DATA_LUT_FIFO, "SDMA_DATA_LUT_FIFO"},
+	{AMDGPU_SDMA_SPLIT_DAT_BUF, "SDMA_SPLIT_DAT_BUF"},
+};
+
+static void sdma_v4_4_2_inst_query_ras_error_count(struct amdgpu_device *adev,
+						   uint32_t sdma_inst,
+						   void *ras_err_status)
+{
+	struct ras_err_data *err_data = (struct ras_err_data *)ras_err_status;
+	uint32_t sdma_dev_inst = GET_INST(SDMA0, sdma_inst);
+	unsigned long ue_count = 0;
+	struct amdgpu_smuio_mcm_config_info mcm_info = {
+		.socket_id = adev->smuio.funcs->get_socket_id(adev),
+		.die_id = adev->sdma.instance[sdma_inst].aid_id,
+	};
+
+	/* sdma v4_4_2 doesn't support query ce counts */
+	amdgpu_ras_inst_query_ras_error_count(adev,
+					sdma_v4_2_2_ue_reg_list,
+					ARRAY_SIZE(sdma_v4_2_2_ue_reg_list),
+					sdma_v4_4_2_ras_memory_list,
+					ARRAY_SIZE(sdma_v4_4_2_ras_memory_list),
+					sdma_dev_inst,
+					AMDGPU_RAS_ERROR__MULTI_UNCORRECTABLE,
+					&ue_count);
+
+	amdgpu_ras_error_statistic_ue_count(err_data, &mcm_info, ue_count);
+}
+
+static void sdma_v4_4_2_query_ras_error_count(struct amdgpu_device *adev,
+					      void *ras_err_status)
+{
+	uint32_t inst_mask;
+	int i = 0;
+
+	inst_mask = GENMASK(adev->sdma.num_instances - 1, 0);
+	if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__SDMA)) {
+		for_each_inst(i, inst_mask)
+			sdma_v4_4_2_inst_query_ras_error_count(adev, i, ras_err_status);
+	} else {
+		dev_warn(adev->dev, "SDMA RAS is not supported\n");
+	}
+}
+
+static void sdma_v4_4_2_inst_reset_ras_error_count(struct amdgpu_device *adev,
+						   uint32_t sdma_inst)
+{
+	uint32_t sdma_dev_inst = GET_INST(SDMA0, sdma_inst);
+
+	amdgpu_ras_inst_reset_ras_error_count(adev,
+					sdma_v4_2_2_ue_reg_list,
+					ARRAY_SIZE(sdma_v4_2_2_ue_reg_list),
+					sdma_dev_inst);
+}
+
+static void sdma_v4_4_2_reset_ras_error_count(struct amdgpu_device *adev)
+{
+	uint32_t inst_mask;
+	int i = 0;
+
+	inst_mask = GENMASK(adev->sdma.num_instances - 1, 0);
+	if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__SDMA)) {
+		for_each_inst(i, inst_mask)
+			sdma_v4_4_2_inst_reset_ras_error_count(adev, i);
+	} else {
+		dev_warn(adev->dev, "SDMA RAS is not supported\n");
+	}
+}
+
+static const struct amdgpu_ras_block_hw_ops sdma_v4_4_2_ras_hw_ops = {
+	.query_ras_error_count = sdma_v4_4_2_query_ras_error_count,
+	.reset_ras_error_count = sdma_v4_4_2_reset_ras_error_count,
+};
+
+static int sdma_v4_4_2_aca_bank_parser(struct aca_handle *handle, struct aca_bank *bank,
+				       enum aca_smu_type type, void *data)
+{
+	struct aca_bank_info info;
+	u64 misc0;
+	int ret;
+
+	ret = aca_bank_info_decode(bank, &info);
+	if (ret)
+		return ret;
+
+	misc0 = bank->regs[ACA_REG_IDX_MISC0];
+	switch (type) {
+	case ACA_SMU_TYPE_UE:
+		bank->aca_err_type = ACA_ERROR_TYPE_UE;
+		ret = aca_error_cache_log_bank_error(handle, &info, ACA_ERROR_TYPE_UE,
+						     1ULL);
+		break;
+	case ACA_SMU_TYPE_CE:
+		bank->aca_err_type = ACA_ERROR_TYPE_CE;
+		ret = aca_error_cache_log_bank_error(handle, &info, bank->aca_err_type,
+						     ACA_REG__MISC0__ERRCNT(misc0));
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return ret;
+}
+
+/* CODE_SDMA0 - CODE_SDMA4, reference to smu driver if header file */
+static int sdma_v4_4_2_err_codes[] = { 33, 34, 35, 36 };
+
+static bool sdma_v4_4_2_aca_bank_is_valid(struct aca_handle *handle, struct aca_bank *bank,
+					  enum aca_smu_type type, void *data)
+{
+	u32 instlo;
+
+	instlo = ACA_REG__IPID__INSTANCEIDLO(bank->regs[ACA_REG_IDX_IPID]);
+	instlo &= GENMASK(31, 1);
+
+	if (instlo != mmSMNAID_AID0_MCA_SMU)
+		return false;
+
+	if (aca_bank_check_error_codes(handle->adev, bank,
+				       sdma_v4_4_2_err_codes,
+				       ARRAY_SIZE(sdma_v4_4_2_err_codes)))
+		return false;
+
+	return true;
+}
+
+static const struct aca_bank_ops sdma_v4_4_2_aca_bank_ops = {
+	.aca_bank_parser = sdma_v4_4_2_aca_bank_parser,
+	.aca_bank_is_valid = sdma_v4_4_2_aca_bank_is_valid,
+};
+
+static const struct aca_info sdma_v4_4_2_aca_info = {
+	.hwip = ACA_HWIP_TYPE_SMU,
+	.mask = ACA_ERROR_UE_MASK,
+	.bank_ops = &sdma_v4_4_2_aca_bank_ops,
+};
+
+static int sdma_v4_4_2_ras_late_init(struct amdgpu_device *adev, struct ras_common_if *ras_block)
+{
+	int r;
+
+	r = amdgpu_sdma_ras_late_init(adev, ras_block);
+	if (r)
+		return r;
+
+	return amdgpu_ras_bind_aca(adev, AMDGPU_RAS_BLOCK__SDMA,
+				   &sdma_v4_4_2_aca_info, NULL);
+}
+
+static struct amdgpu_sdma_ras sdma_v4_4_2_ras = {
+	.ras_block = {
+		.hw_ops = &sdma_v4_4_2_ras_hw_ops,
+		.ras_late_init = sdma_v4_4_2_ras_late_init,
+	},
+};
+
+static void sdma_v4_4_2_set_ras_funcs(struct amdgpu_device *adev)
+{
+	adev->sdma.ras = &sdma_v4_4_2_ras;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.h b/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.h
new file mode 100644
index 000000000000..d516145529bb
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.h
@@ -0,0 +1,32 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __SDMA_V4_4_2_H__
+#define __SDMA_V4_4_2_H__
+
+extern const struct amd_ip_funcs sdma_v4_4_2_ip_funcs;
+extern const struct amdgpu_ip_block_version sdma_v4_4_2_ip_block;
+
+extern struct amdgpu_xcp_ip_funcs sdma_v4_4_2_xcp_funcs;
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c
index 853d1511b889..8ddc4df06a1f 100644
--- a/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c
@@ -51,9 +51,6 @@ MODULE_FIRMWARE("amdgpu/navi14_sdma1.bin");
 MODULE_FIRMWARE("amdgpu/navi12_sdma.bin");
 MODULE_FIRMWARE("amdgpu/navi12_sdma1.bin");
 
-MODULE_FIRMWARE("amdgpu/cyan_skillfish_sdma.bin");
-MODULE_FIRMWARE("amdgpu/cyan_skillfish_sdma1.bin");
-
 MODULE_FIRMWARE("amdgpu/cyan_skillfish2_sdma.bin");
 MODULE_FIRMWARE("amdgpu/cyan_skillfish2_sdma1.bin");
 
@@ -62,10 +59,61 @@ MODULE_FIRMWARE("amdgpu/cyan_skillfish2_sdma1.bin");
 #define SDMA0_HYP_DEC_REG_END 0x5893
 #define SDMA1_HYP_DEC_REG_OFFSET 0x20
 
+static const struct amdgpu_hwip_reg_entry sdma_reg_list_5_0[] = {
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_STATUS_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_STATUS1_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_STATUS2_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_STATUS3_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_UCODE_CHECKSUM),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RB_RPTR_FETCH_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RB_RPTR_FETCH),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_UTCL1_RD_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_UTCL1_WR_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_UTCL1_RD_XNACK0),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_UTCL1_RD_XNACK1),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_UTCL1_WR_XNACK0),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_UTCL1_WR_XNACK1),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_RB_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_RB_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_RB_RPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_RB_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_RB_WPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_IB_OFFSET),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_IB_BASE_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_IB_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_IB_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_IB_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_IB_SUB_REMAIN),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_DUMMY_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_PAGE_RB_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_PAGE_RB_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_PAGE_RB_RPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_PAGE_RB_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_PAGE_RB_WPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_PAGE_IB_OFFSET),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_PAGE_IB_BASE_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_PAGE_IB_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_PAGE_DUMMY_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RLC0_RB_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RLC0_RB_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RLC0_RB_RPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RLC0_RB_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RLC0_RB_WPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RLC0_IB_OFFSET),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RLC0_IB_BASE_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RLC0_IB_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RLC0_DUMMY_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_INT_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_VM_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, mmGRBM_STATUS2)
+};
+
 static void sdma_v5_0_set_ring_funcs(struct amdgpu_device *adev);
 static void sdma_v5_0_set_buffer_funcs(struct amdgpu_device *adev);
 static void sdma_v5_0_set_vm_pte_funcs(struct amdgpu_device *adev);
 static void sdma_v5_0_set_irq_funcs(struct amdgpu_device *adev);
+static int sdma_v5_0_stop_queue(struct amdgpu_ring *ring);
+static int sdma_v5_0_restore_queue(struct amdgpu_ring *ring);
 
 static const struct soc15_reg_golden golden_settings_sdma_5[] = {
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmSDMA0_CHICKEN_BITS, 0xffbf1f0f, 0x03ab0107),
@@ -187,7 +235,7 @@ static u32 sdma_v5_0_get_reg_offset(struct amdgpu_device *adev, u32 instance, u3
 
 static void sdma_v5_0_init_golden_registers(struct amdgpu_device *adev)
 {
-	switch (adev->ip_versions[SDMA0_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, SDMA0_HWIP, 0)) {
 	case IP_VERSION(5, 0, 0):
 		soc15_program_register_sequence(adev,
 						golden_settings_sdma_5,
@@ -241,106 +289,34 @@ static void sdma_v5_0_init_golden_registers(struct amdgpu_device *adev)
 // navi10 real chip need to use PSP to load firmware
 static int sdma_v5_0_init_microcode(struct amdgpu_device *adev)
 {
-	const char *chip_name;
-	char fw_name[40];
-	int err = 0, i;
-	struct amdgpu_firmware_info *info = NULL;
-	const struct common_firmware_header *header = NULL;
-	const struct sdma_firmware_header_v1_0 *hdr;
-
-	if (amdgpu_sriov_vf(adev) && (adev->ip_versions[SDMA0_HWIP][0] == IP_VERSION(5, 0, 5)))
-		return 0;
-
-	DRM_DEBUG("\n");
-
-	switch (adev->ip_versions[SDMA0_HWIP][0]) {
-	case IP_VERSION(5, 0, 0):
-		chip_name = "navi10";
-		break;
-	case IP_VERSION(5, 0, 2):
-		chip_name = "navi14";
-		break;
-	case IP_VERSION(5, 0, 5):
-		chip_name = "navi12";
-		break;
-	case IP_VERSION(5, 0, 1):
-		if (adev->apu_flags & AMD_APU_IS_CYAN_SKILLFISH2)
-			chip_name = "cyan_skillfish2";
-		else
-			chip_name = "cyan_skillfish";
-		break;
-	default:
-		BUG();
-	}
+	int ret, i;
 
 	for (i = 0; i < adev->sdma.num_instances; i++) {
-		if (i == 0)
-			snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_sdma.bin", chip_name);
-		else
-			snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_sdma1.bin", chip_name);
-		err = request_firmware(&adev->sdma.instance[i].fw, fw_name, adev->dev);
-		if (err)
-			goto out;
-		err = amdgpu_ucode_validate(adev->sdma.instance[i].fw);
-		if (err)
-			goto out;
-		hdr = (const struct sdma_firmware_header_v1_0 *)adev->sdma.instance[i].fw->data;
-		adev->sdma.instance[i].fw_version = le32_to_cpu(hdr->header.ucode_version);
-		adev->sdma.instance[i].feature_version = le32_to_cpu(hdr->ucode_feature_version);
-		if (adev->sdma.instance[i].feature_version >= 20)
-			adev->sdma.instance[i].burst_nop = true;
-		DRM_DEBUG("psp_load == '%s'\n",
-				adev->firmware.load_type == AMDGPU_FW_LOAD_PSP ? "true" : "false");
-
-		if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
-			info = &adev->firmware.ucode[AMDGPU_UCODE_ID_SDMA0 + i];
-			info->ucode_id = AMDGPU_UCODE_ID_SDMA0 + i;
-			info->fw = adev->sdma.instance[i].fw;
-			header = (const struct common_firmware_header *)info->fw->data;
-			adev->firmware.fw_size +=
-				ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE);
-		}
+		ret = amdgpu_sdma_init_microcode(adev, i, false);
+		if (ret)
+			return ret;
 	}
-out:
-	if (err) {
-		DRM_ERROR("sdma_v5_0: Failed to load firmware \"%s\"\n", fw_name);
-		for (i = 0; i < adev->sdma.num_instances; i++) {
-			release_firmware(adev->sdma.instance[i].fw);
-			adev->sdma.instance[i].fw = NULL;
-		}
-	}
-	return err;
+
+	return ret;
 }
 
-static unsigned sdma_v5_0_ring_init_cond_exec(struct amdgpu_ring *ring)
+static unsigned sdma_v5_0_ring_init_cond_exec(struct amdgpu_ring *ring,
+					      uint64_t addr)
 {
 	unsigned ret;
 
 	amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_COND_EXE));
-	amdgpu_ring_write(ring, lower_32_bits(ring->cond_exe_gpu_addr));
-	amdgpu_ring_write(ring, upper_32_bits(ring->cond_exe_gpu_addr));
+	amdgpu_ring_write(ring, lower_32_bits(addr));
+	amdgpu_ring_write(ring, upper_32_bits(addr));
 	amdgpu_ring_write(ring, 1);
-	ret = ring->wptr & ring->buf_mask;/* this is the offset we need patch later */
-	amdgpu_ring_write(ring, 0x55aa55aa);/* insert dummy here and patch it later */
+	/* this is the offset we need patch later */
+	ret = ring->wptr & ring->buf_mask;
+	/* insert dummy here and patch it later */
+	amdgpu_ring_write(ring, 0);
 
 	return ret;
 }
 
-static void sdma_v5_0_ring_patch_cond_exec(struct amdgpu_ring *ring,
-					   unsigned offset)
-{
-	unsigned cur;
-
-	BUG_ON(offset > ring->buf_mask);
-	BUG_ON(ring->ring[offset] != 0x55aa55aa);
-
-	cur = (ring->wptr - 1) & ring->buf_mask;
-	if (cur > offset)
-		ring->ring[offset] = cur - offset;
-	else
-		ring->ring[offset] = (ring->buf_mask + 1) - offset + cur;
-}
-
 /**
  * sdma_v5_0_ring_get_rptr - get the current read pointer
  *
@@ -353,7 +329,7 @@ static uint64_t sdma_v5_0_ring_get_rptr(struct amdgpu_ring *ring)
 	u64 *rptr;
 
 	/* XXX check if swapping is necessary on BE */
-	rptr = ((u64 *)&ring->adev->wb.wb[ring->rptr_offs]);
+	rptr = (u64 *)ring->rptr_cpu_addr;
 
 	DRM_DEBUG("rptr before shift == 0x%016llx\n", *rptr);
 	return ((*rptr) >> 2);
@@ -373,7 +349,7 @@ static uint64_t sdma_v5_0_ring_get_wptr(struct amdgpu_ring *ring)
 
 	if (ring->use_doorbell) {
 		/* XXX check if swapping is necessary on BE */
-		wptr = READ_ONCE(*((u64 *)&adev->wb.wb[ring->wptr_offs]));
+		wptr = READ_ONCE(*((u64 *)ring->wptr_cpu_addr));
 		DRM_DEBUG("wptr/doorbell before shift == 0x%016llx\n", wptr);
 	} else {
 		wptr = RREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, ring->me, mmSDMA0_GFX_RB_WPTR_HI));
@@ -399,30 +375,32 @@ static void sdma_v5_0_ring_set_wptr(struct amdgpu_ring *ring)
 	DRM_DEBUG("Setting write pointer\n");
 	if (ring->use_doorbell) {
 		DRM_DEBUG("Using doorbell -- "
-				"wptr_offs == 0x%08x "
-				"lower_32_bits(ring->wptr) << 2 == 0x%08x "
-				"upper_32_bits(ring->wptr) << 2 == 0x%08x\n",
-				ring->wptr_offs,
-				lower_32_bits(ring->wptr << 2),
-				upper_32_bits(ring->wptr << 2));
+			  "wptr_offs == 0x%08x "
+			  "lower_32_bits(ring->wptr) << 2 == 0x%08x "
+			  "upper_32_bits(ring->wptr) << 2 == 0x%08x\n",
+			  ring->wptr_offs,
+			  lower_32_bits(ring->wptr << 2),
+			  upper_32_bits(ring->wptr << 2));
 		/* XXX check if swapping is necessary on BE */
-		adev->wb.wb[ring->wptr_offs] = lower_32_bits(ring->wptr << 2);
-		adev->wb.wb[ring->wptr_offs + 1] = upper_32_bits(ring->wptr << 2);
+		atomic64_set((atomic64_t *)ring->wptr_cpu_addr,
+			     ring->wptr << 2);
 		DRM_DEBUG("calling WDOORBELL64(0x%08x, 0x%016llx)\n",
-				ring->doorbell_index, ring->wptr << 2);
+			  ring->doorbell_index, ring->wptr << 2);
 		WDOORBELL64(ring->doorbell_index, ring->wptr << 2);
 	} else {
 		DRM_DEBUG("Not using doorbell -- "
-				"mmSDMA%i_GFX_RB_WPTR == 0x%08x "
-				"mmSDMA%i_GFX_RB_WPTR_HI == 0x%08x\n",
-				ring->me,
-				lower_32_bits(ring->wptr << 2),
-				ring->me,
+			  "mmSDMA%i_GFX_RB_WPTR == 0x%08x "
+			  "mmSDMA%i_GFX_RB_WPTR_HI == 0x%08x\n",
+			  ring->me,
+			  lower_32_bits(ring->wptr << 2),
+			  ring->me,
+			  upper_32_bits(ring->wptr << 2));
+		WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev,
+							     ring->me, mmSDMA0_GFX_RB_WPTR),
+				lower_32_bits(ring->wptr << 2));
+		WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev,
+							     ring->me, mmSDMA0_GFX_RB_WPTR_HI),
 				upper_32_bits(ring->wptr << 2));
-		WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, ring->me, mmSDMA0_GFX_RB_WPTR),
-			lower_32_bits(ring->wptr << 2));
-		WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, ring->me, mmSDMA0_GFX_RB_WPTR_HI),
-			upper_32_bits(ring->wptr << 2));
 	}
 }
 
@@ -481,8 +459,6 @@ static void sdma_v5_0_ring_emit_ib(struct amdgpu_ring *ring,
  * sdma_v5_0_ring_emit_mem_sync - flush the IB by graphics cache rinse
  *
  * @ring: amdgpu ring pointer
- * @job: job to retrieve vmid from
- * @ib: IB object to schedule
  *
  * flush the IB by graphics cache rinse.
  */
@@ -581,21 +557,15 @@ static void sdma_v5_0_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 se
  * sdma_v5_0_gfx_stop - stop the gfx async dma engines
  *
  * @adev: amdgpu_device pointer
- *
+ * @inst_mask: mask of dma engine instances to be disabled
  * Stop the gfx async dma ring buffers (NAVI10).
  */
-static void sdma_v5_0_gfx_stop(struct amdgpu_device *adev)
+static void sdma_v5_0_gfx_stop(struct amdgpu_device *adev, uint32_t inst_mask)
 {
-	struct amdgpu_ring *sdma0 = &adev->sdma.instance[0].ring;
-	struct amdgpu_ring *sdma1 = &adev->sdma.instance[1].ring;
 	u32 rb_cntl, ib_cntl;
 	int i;
 
-	if ((adev->mman.buffer_funcs_ring == sdma0) ||
-	    (adev->mman.buffer_funcs_ring == sdma1))
-		amdgpu_ttm_set_buffer_funcs_status(adev, false);
-
-	for (i = 0; i < adev->sdma.num_instances; i++) {
+	for_each_inst(i, inst_mask) {
 		rb_cntl = RREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_RB_CNTL));
 		rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL, RB_ENABLE, 0);
 		WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_RB_CNTL), rb_cntl);
@@ -687,9 +657,11 @@ static void sdma_v5_0_enable(struct amdgpu_device *adev, bool enable)
 {
 	u32 f32_cntl;
 	int i;
+	uint32_t inst_mask;
 
+	inst_mask = GENMASK(adev->sdma.num_instances - 1, 0);
 	if (!enable) {
-		sdma_v5_0_gfx_stop(adev);
+		sdma_v5_0_gfx_stop(adev, 1 << inst_mask);
 		sdma_v5_0_rlc_stop(adev);
 	}
 
@@ -704,168 +676,184 @@ static void sdma_v5_0_enable(struct amdgpu_device *adev, bool enable)
 }
 
 /**
- * sdma_v5_0_gfx_resume - setup and start the async dma engines
+ * sdma_v5_0_gfx_resume_instance - start/restart a certain sdma engine
  *
  * @adev: amdgpu_device pointer
+ * @i: instance
+ * @restore: used to restore wptr when restart
  *
- * Set up the gfx DMA ring buffers and enable them (NAVI10).
- * Returns 0 for success, error for failure.
+ * Set up the gfx DMA ring buffers and enable them. On restart, we will restore wptr and rptr.
+ * Return 0 for success.
  */
-static int sdma_v5_0_gfx_resume(struct amdgpu_device *adev)
+static int sdma_v5_0_gfx_resume_instance(struct amdgpu_device *adev, int i, bool restore)
 {
 	struct amdgpu_ring *ring;
 	u32 rb_cntl, ib_cntl;
 	u32 rb_bufsz;
-	u32 wb_offset;
 	u32 doorbell;
 	u32 doorbell_offset;
 	u32 temp;
 	u32 wptr_poll_cntl;
 	u64 wptr_gpu_addr;
-	int i, r;
 
-	for (i = 0; i < adev->sdma.num_instances; i++) {
-		ring = &adev->sdma.instance[i].ring;
-		wb_offset = (ring->rptr_offs * 4);
+	ring = &adev->sdma.instance[i].ring;
 
-		if (!amdgpu_sriov_vf(adev))
-			WREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_SEM_WAIT_FAIL_TIMER_CNTL), 0);
+	if (!amdgpu_sriov_vf(adev))
+		WREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_SEM_WAIT_FAIL_TIMER_CNTL), 0);
 
-		/* Set ring buffer size in dwords */
-		rb_bufsz = order_base_2(ring->ring_size / 4);
-		rb_cntl = RREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_RB_CNTL));
-		rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL, RB_SIZE, rb_bufsz);
+	/* Set ring buffer size in dwords */
+	rb_bufsz = order_base_2(ring->ring_size / 4);
+	rb_cntl = RREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_RB_CNTL));
+	rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL, RB_SIZE, rb_bufsz);
 #ifdef __BIG_ENDIAN
-		rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL, RB_SWAP_ENABLE, 1);
-		rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL,
-					RPTR_WRITEBACK_SWAP_ENABLE, 1);
+	rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL, RB_SWAP_ENABLE, 1);
+	rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL,
+				RPTR_WRITEBACK_SWAP_ENABLE, 1);
 #endif
-		WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_RB_CNTL), rb_cntl);
-
-		/* Initialize the ring buffer's read and write pointers */
+	WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_RB_CNTL), rb_cntl);
+
+	/* Initialize the ring buffer's read and write pointers */
+	if (restore) {
+		WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_RB_RPTR), lower_32_bits(ring->wptr << 2));
+		WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_RB_RPTR_HI), upper_32_bits(ring->wptr << 2));
+		WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_RB_WPTR), lower_32_bits(ring->wptr << 2));
+		WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_RB_WPTR_HI), upper_32_bits(ring->wptr << 2));
+	} else {
 		WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_RB_RPTR), 0);
 		WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_RB_RPTR_HI), 0);
 		WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_RB_WPTR), 0);
 		WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_RB_WPTR_HI), 0);
-
-		/* setup the wptr shadow polling */
-		wptr_gpu_addr = adev->wb.gpu_addr + (ring->wptr_offs * 4);
-		WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_RB_WPTR_POLL_ADDR_LO),
-		       lower_32_bits(wptr_gpu_addr));
-		WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_RB_WPTR_POLL_ADDR_HI),
-		       upper_32_bits(wptr_gpu_addr));
-		wptr_poll_cntl = RREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i,
-							 mmSDMA0_GFX_RB_WPTR_POLL_CNTL));
-		wptr_poll_cntl = REG_SET_FIELD(wptr_poll_cntl,
-					       SDMA0_GFX_RB_WPTR_POLL_CNTL,
-					       F32_POLL_ENABLE, 1);
-		WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_RB_WPTR_POLL_CNTL),
-		       wptr_poll_cntl);
-
-		/* set the wb address whether it's enabled or not */
-		WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_RB_RPTR_ADDR_HI),
-		       upper_32_bits(adev->wb.gpu_addr + wb_offset) & 0xFFFFFFFF);
-		WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_RB_RPTR_ADDR_LO),
-		       lower_32_bits(adev->wb.gpu_addr + wb_offset) & 0xFFFFFFFC);
-
-		rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL, RPTR_WRITEBACK_ENABLE, 1);
-
-		WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_RB_BASE),
-		       ring->gpu_addr >> 8);
-		WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_RB_BASE_HI),
-		       ring->gpu_addr >> 40);
-
+	}
+	/* setup the wptr shadow polling */
+	wptr_gpu_addr = ring->wptr_gpu_addr;
+	WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_RB_WPTR_POLL_ADDR_LO),
+	       lower_32_bits(wptr_gpu_addr));
+	WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_RB_WPTR_POLL_ADDR_HI),
+	       upper_32_bits(wptr_gpu_addr));
+	wptr_poll_cntl = RREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i,
+						 mmSDMA0_GFX_RB_WPTR_POLL_CNTL));
+	wptr_poll_cntl = REG_SET_FIELD(wptr_poll_cntl,
+				       SDMA0_GFX_RB_WPTR_POLL_CNTL,
+				       F32_POLL_ENABLE, 1);
+	WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_RB_WPTR_POLL_CNTL),
+	       wptr_poll_cntl);
+
+	/* set the wb address whether it's enabled or not */
+	WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_RB_RPTR_ADDR_HI),
+	       upper_32_bits(ring->rptr_gpu_addr) & 0xFFFFFFFF);
+	WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_RB_RPTR_ADDR_LO),
+	       lower_32_bits(ring->rptr_gpu_addr) & 0xFFFFFFFC);
+
+	rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL, RPTR_WRITEBACK_ENABLE, 1);
+
+	WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_RB_BASE),
+	       ring->gpu_addr >> 8);
+	WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_RB_BASE_HI),
+	       ring->gpu_addr >> 40);
+
+	if (!restore)
 		ring->wptr = 0;
 
-		/* before programing wptr to a less value, need set minor_ptr_update first */
-		WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_MINOR_PTR_UPDATE), 1);
+	/* before programing wptr to a less value, need set minor_ptr_update first */
+	WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_MINOR_PTR_UPDATE), 1);
 
-		if (!amdgpu_sriov_vf(adev)) { /* only bare-metal use register write for wptr */
-			WREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_RB_WPTR),
-			       lower_32_bits(ring->wptr) << 2);
-			WREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_RB_WPTR_HI),
-			       upper_32_bits(ring->wptr) << 2);
-		}
+	if (!amdgpu_sriov_vf(adev)) { /* only bare-metal use register write for wptr */
+		WREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_RB_WPTR),
+		       lower_32_bits(ring->wptr << 2));
+		WREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_RB_WPTR_HI),
+		       upper_32_bits(ring->wptr << 2));
+	}
 
-		doorbell = RREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_DOORBELL));
-		doorbell_offset = RREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i,
-						mmSDMA0_GFX_DOORBELL_OFFSET));
+	doorbell = RREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_DOORBELL));
+	doorbell_offset = RREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i,
+					mmSDMA0_GFX_DOORBELL_OFFSET));
 
-		if (ring->use_doorbell) {
-			doorbell = REG_SET_FIELD(doorbell, SDMA0_GFX_DOORBELL, ENABLE, 1);
-			doorbell_offset = REG_SET_FIELD(doorbell_offset, SDMA0_GFX_DOORBELL_OFFSET,
-					OFFSET, ring->doorbell_index);
-		} else {
-			doorbell = REG_SET_FIELD(doorbell, SDMA0_GFX_DOORBELL, ENABLE, 0);
-		}
-		WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_DOORBELL), doorbell);
-		WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_DOORBELL_OFFSET),
-		       doorbell_offset);
+	if (ring->use_doorbell) {
+		doorbell = REG_SET_FIELD(doorbell, SDMA0_GFX_DOORBELL, ENABLE, 1);
+		doorbell_offset = REG_SET_FIELD(doorbell_offset, SDMA0_GFX_DOORBELL_OFFSET,
+				OFFSET, ring->doorbell_index);
+	} else {
+		doorbell = REG_SET_FIELD(doorbell, SDMA0_GFX_DOORBELL, ENABLE, 0);
+	}
+	WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_DOORBELL), doorbell);
+	WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_DOORBELL_OFFSET),
+	       doorbell_offset);
 
-		adev->nbio.funcs->sdma_doorbell_range(adev, i, ring->use_doorbell,
-						      ring->doorbell_index, 20);
+	adev->nbio.funcs->sdma_doorbell_range(adev, i, ring->use_doorbell,
+					      ring->doorbell_index, 20);
 
-		if (amdgpu_sriov_vf(adev))
-			sdma_v5_0_ring_set_wptr(ring);
+	if (amdgpu_sriov_vf(adev))
+		sdma_v5_0_ring_set_wptr(ring);
 
-		/* set minor_ptr_update to 0 after wptr programed */
-		WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_MINOR_PTR_UPDATE), 0);
+	/* set minor_ptr_update to 0 after wptr programed */
+	WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_MINOR_PTR_UPDATE), 0);
 
-		if (!amdgpu_sriov_vf(adev)) {
-			/* set utc l1 enable flag always to 1 */
-			temp = RREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_CNTL));
-			temp = REG_SET_FIELD(temp, SDMA0_CNTL, UTC_L1_ENABLE, 1);
-
-			/* enable MCBP */
-			temp = REG_SET_FIELD(temp, SDMA0_CNTL, MIDCMD_PREEMPT_ENABLE, 1);
-			WREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_CNTL), temp);
-
-			/* Set up RESP_MODE to non-copy addresses */
-			temp = RREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_UTCL1_CNTL));
-			temp = REG_SET_FIELD(temp, SDMA0_UTCL1_CNTL, RESP_MODE, 3);
-			temp = REG_SET_FIELD(temp, SDMA0_UTCL1_CNTL, REDO_DELAY, 9);
-			WREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_UTCL1_CNTL), temp);
-
-			/* program default cache read and write policy */
-			temp = RREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_UTCL1_PAGE));
-			/* clean read policy and write policy bits */
-			temp &= 0xFF0FFF;
-			temp |= ((CACHE_READ_POLICY_L2__DEFAULT << 12) | (CACHE_WRITE_POLICY_L2__DEFAULT << 14));
-			WREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_UTCL1_PAGE), temp);
-		}
+	if (!amdgpu_sriov_vf(adev)) {
+		/* set utc l1 enable flag always to 1 */
+		temp = RREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_CNTL));
+		temp = REG_SET_FIELD(temp, SDMA0_CNTL, UTC_L1_ENABLE, 1);
+
+		/* enable MCBP */
+		temp = REG_SET_FIELD(temp, SDMA0_CNTL, MIDCMD_PREEMPT_ENABLE, 1);
+		WREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_CNTL), temp);
+
+		/* Set up RESP_MODE to non-copy addresses */
+		temp = RREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_UTCL1_CNTL));
+		temp = REG_SET_FIELD(temp, SDMA0_UTCL1_CNTL, RESP_MODE, 3);
+		temp = REG_SET_FIELD(temp, SDMA0_UTCL1_CNTL, REDO_DELAY, 9);
+		WREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_UTCL1_CNTL), temp);
+
+		/* program default cache read and write policy */
+		temp = RREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_UTCL1_PAGE));
+		/* clean read policy and write policy bits */
+		temp &= 0xFF0FFF;
+		temp |= ((CACHE_READ_POLICY_L2__DEFAULT << 12) | (CACHE_WRITE_POLICY_L2__DEFAULT << 14));
+		WREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_UTCL1_PAGE), temp);
+	}
 
-		if (!amdgpu_sriov_vf(adev)) {
-			/* unhalt engine */
-			temp = RREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_F32_CNTL));
-			temp = REG_SET_FIELD(temp, SDMA0_F32_CNTL, HALT, 0);
-			WREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_F32_CNTL), temp);
-		}
+	if (!amdgpu_sriov_vf(adev)) {
+		/* unhalt engine */
+		temp = RREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_F32_CNTL));
+		temp = REG_SET_FIELD(temp, SDMA0_F32_CNTL, HALT, 0);
+		WREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_F32_CNTL), temp);
+	}
 
-		/* enable DMA RB */
-		rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL, RB_ENABLE, 1);
-		WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_RB_CNTL), rb_cntl);
+	/* enable DMA RB */
+	rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL, RB_ENABLE, 1);
+	WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_RB_CNTL), rb_cntl);
 
-		ib_cntl = RREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_IB_CNTL));
-		ib_cntl = REG_SET_FIELD(ib_cntl, SDMA0_GFX_IB_CNTL, IB_ENABLE, 1);
+	ib_cntl = RREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_IB_CNTL));
+	ib_cntl = REG_SET_FIELD(ib_cntl, SDMA0_GFX_IB_CNTL, IB_ENABLE, 1);
 #ifdef __BIG_ENDIAN
-		ib_cntl = REG_SET_FIELD(ib_cntl, SDMA0_GFX_IB_CNTL, IB_SWAP_ENABLE, 1);
+	ib_cntl = REG_SET_FIELD(ib_cntl, SDMA0_GFX_IB_CNTL, IB_SWAP_ENABLE, 1);
 #endif
-		/* enable DMA IBs */
-		WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_IB_CNTL), ib_cntl);
+	/* enable DMA IBs */
+	WREG32_SOC15_IP(GC, sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_GFX_IB_CNTL), ib_cntl);
 
-		ring->sched.ready = true;
+	if (amdgpu_sriov_vf(adev)) { /* bare-metal sequence doesn't need below to lines */
+		sdma_v5_0_ctx_switch_enable(adev, true);
+		sdma_v5_0_enable(adev, true);
+	}
 
-		if (amdgpu_sriov_vf(adev)) { /* bare-metal sequence doesn't need below to lines */
-			sdma_v5_0_ctx_switch_enable(adev, true);
-			sdma_v5_0_enable(adev, true);
-		}
+	return amdgpu_ring_test_helper(ring);
+}
 
-		r = amdgpu_ring_test_helper(ring);
+/**
+ * sdma_v5_0_gfx_resume - setup and start the async dma engines
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Set up the gfx DMA ring buffers and enable them (NAVI10).
+ * Returns 0 for success, error for failure.
+ */
+static int sdma_v5_0_gfx_resume(struct amdgpu_device *adev)
+{
+	int i, r;
+
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		r = sdma_v5_0_gfx_resume_instance(adev, i, false);
 		if (r)
 			return r;
-
-		if (adev->mman.buffer_funcs_ring == ring)
-			amdgpu_ttm_set_buffer_funcs_status(adev, true);
 	}
 
 	return 0;
@@ -969,6 +957,49 @@ static int sdma_v5_0_start(struct amdgpu_device *adev)
 	return r;
 }
 
+static int sdma_v5_0_mqd_init(struct amdgpu_device *adev, void *mqd,
+			      struct amdgpu_mqd_prop *prop)
+{
+	struct v10_sdma_mqd *m = mqd;
+	uint64_t wb_gpu_addr;
+
+	m->sdmax_rlcx_rb_cntl =
+		order_base_2(prop->queue_size / 4) << SDMA0_RLC0_RB_CNTL__RB_SIZE__SHIFT |
+		1 << SDMA0_RLC0_RB_CNTL__RPTR_WRITEBACK_ENABLE__SHIFT |
+		6 << SDMA0_RLC0_RB_CNTL__RPTR_WRITEBACK_TIMER__SHIFT |
+		1 << SDMA0_RLC0_RB_CNTL__RB_PRIV__SHIFT;
+
+	m->sdmax_rlcx_rb_base = lower_32_bits(prop->hqd_base_gpu_addr >> 8);
+	m->sdmax_rlcx_rb_base_hi = upper_32_bits(prop->hqd_base_gpu_addr >> 8);
+
+	m->sdmax_rlcx_rb_wptr_poll_cntl = RREG32(sdma_v5_0_get_reg_offset(adev, 0,
+						  mmSDMA0_GFX_RB_WPTR_POLL_CNTL));
+
+	wb_gpu_addr = prop->wptr_gpu_addr;
+	m->sdmax_rlcx_rb_wptr_poll_addr_lo = lower_32_bits(wb_gpu_addr);
+	m->sdmax_rlcx_rb_wptr_poll_addr_hi = upper_32_bits(wb_gpu_addr);
+
+	wb_gpu_addr = prop->rptr_gpu_addr;
+	m->sdmax_rlcx_rb_rptr_addr_lo = lower_32_bits(wb_gpu_addr);
+	m->sdmax_rlcx_rb_rptr_addr_hi = upper_32_bits(wb_gpu_addr);
+
+	m->sdmax_rlcx_ib_cntl = RREG32(sdma_v5_0_get_reg_offset(adev, 0,
+							mmSDMA0_GFX_IB_CNTL));
+
+	m->sdmax_rlcx_doorbell_offset =
+		prop->doorbell_index << SDMA0_RLC0_DOORBELL_OFFSET__OFFSET__SHIFT;
+
+	m->sdmax_rlcx_doorbell = REG_SET_FIELD(0, SDMA0_RLC0_DOORBELL, ENABLE, 1);
+
+	return 0;
+}
+
+static void sdma_v5_0_set_mqd_funcs(struct amdgpu_device *adev)
+{
+	adev->mqds[AMDGPU_HW_IP_DMA].mqd_size = sizeof(struct v10_sdma_mqd);
+	adev->mqds[AMDGPU_HW_IP_DMA].init_mqd = sdma_v5_0_mqd_init;
+}
+
 /**
  * sdma_v5_0_ring_test_ring - simple async dma engine test
  *
@@ -987,6 +1018,8 @@ static int sdma_v5_0_ring_test_ring(struct amdgpu_ring *ring)
 	u32 tmp;
 	u64 gpu_addr;
 
+	tmp = 0xCAFEDEAD;
+
 	r = amdgpu_device_wb_get(adev, &index);
 	if (r) {
 		dev_err(adev->dev, "(%d) failed to allocate wb slot\n", r);
@@ -994,10 +1027,9 @@ static int sdma_v5_0_ring_test_ring(struct amdgpu_ring *ring)
 	}
 
 	gpu_addr = adev->wb.gpu_addr + (index * 4);
-	tmp = 0xCAFEDEAD;
 	adev->wb.wb[index] = cpu_to_le32(tmp);
 
-	r = amdgpu_ring_alloc(ring, 5);
+	r = amdgpu_ring_alloc(ring, 20);
 	if (r) {
 		DRM_ERROR("amdgpu: dma failed to lock ring %d (%d).\n", ring->idx, r);
 		amdgpu_device_wb_free(adev, index);
@@ -1049,6 +1081,9 @@ static int sdma_v5_0_ring_test_ib(struct amdgpu_ring *ring, long timeout)
 	u32 tmp = 0;
 	u64 gpu_addr;
 
+	tmp = 0xCAFEDEAD;
+	memset(&ib, 0, sizeof(ib));
+
 	r = amdgpu_device_wb_get(adev, &index);
 	if (r) {
 		dev_err(adev->dev, "(%ld) failed to allocate wb slot\n", r);
@@ -1056,11 +1091,10 @@ static int sdma_v5_0_ring_test_ib(struct amdgpu_ring *ring, long timeout)
 	}
 
 	gpu_addr = adev->wb.gpu_addr + (index * 4);
-	tmp = 0xCAFEDEAD;
 	adev->wb.wb[index] = cpu_to_le32(tmp);
-	memset(&ib, 0, sizeof(ib));
+
 	r = amdgpu_ib_get(adev, NULL, 256,
-					AMDGPU_IB_POOL_DIRECT, &ib);
+			  AMDGPU_IB_POOL_DIRECT, &ib);
 	if (r) {
 		DRM_ERROR("amdgpu: failed to get ib (%ld).\n", r);
 		goto err0;
@@ -1090,14 +1124,16 @@ static int sdma_v5_0_ring_test_ib(struct amdgpu_ring *ring, long timeout)
 		DRM_ERROR("amdgpu: fence wait failed (%ld).\n", r);
 		goto err1;
 	}
+
 	tmp = le32_to_cpu(adev->wb.wb[index]);
+
 	if (tmp == 0xDEADBEEF)
 		r = 0;
 	else
 		r = -EINVAL;
 
 err1:
-	amdgpu_ib_free(adev, &ib, NULL);
+	amdgpu_ib_free(&ib, NULL);
 	dma_fence_put(f);
 err0:
 	amdgpu_device_wb_free(adev, index);
@@ -1291,24 +1327,62 @@ static void sdma_v5_0_ring_emit_reg_write_reg_wait(struct amdgpu_ring *ring,
 	amdgpu_ring_emit_reg_wait(ring, reg1, mask, mask);
 }
 
-static int sdma_v5_0_early_init(void *handle)
+static int sdma_v5_0_soft_reset_engine(struct amdgpu_device *adev, u32 instance_id)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	u32 grbm_soft_reset;
+	u32 tmp;
+
+	grbm_soft_reset = REG_SET_FIELD(0,
+					GRBM_SOFT_RESET, SOFT_RESET_SDMA0,
+					1);
+	grbm_soft_reset <<= instance_id;
+
+	tmp = RREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET);
+	tmp |= grbm_soft_reset;
+	DRM_DEBUG("GRBM_SOFT_RESET=0x%08X\n", tmp);
+	WREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET, tmp);
+	tmp = RREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET);
+
+	udelay(50);
+
+	tmp &= ~grbm_soft_reset;
+	WREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET, tmp);
+	tmp = RREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET);
+	return 0;
+}
+
+static const struct amdgpu_sdma_funcs sdma_v5_0_sdma_funcs = {
+	.stop_kernel_queue = &sdma_v5_0_stop_queue,
+	.start_kernel_queue = &sdma_v5_0_restore_queue,
+	.soft_reset_kernel_queue = &sdma_v5_0_soft_reset_engine,
+};
+
+static int sdma_v5_0_early_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int r;
+
+	r = sdma_v5_0_init_microcode(adev);
+	if (r)
+		return r;
 
 	sdma_v5_0_set_ring_funcs(adev);
 	sdma_v5_0_set_buffer_funcs(adev);
 	sdma_v5_0_set_vm_pte_funcs(adev);
 	sdma_v5_0_set_irq_funcs(adev);
+	sdma_v5_0_set_mqd_funcs(adev);
 
 	return 0;
 }
 
 
-static int sdma_v5_0_sw_init(void *handle)
+static int sdma_v5_0_sw_init(struct amdgpu_ip_block *ip_block)
 {
 	struct amdgpu_ring *ring;
 	int r, i;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
+	uint32_t reg_count = ARRAY_SIZE(sdma_reg_list_5_0);
+	uint32_t *ptr;
 
 	/* SDMA trap event */
 	r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_SDMA0,
@@ -1324,13 +1398,9 @@ static int sdma_v5_0_sw_init(void *handle)
 	if (r)
 		return r;
 
-	r = sdma_v5_0_init_microcode(adev);
-	if (r) {
-		DRM_ERROR("Failed to load sdma firmware!\n");
-		return r;
-	}
-
 	for (i = 0; i < adev->sdma.num_instances; i++) {
+		mutex_init(&adev->sdma.instance[i].engine_reset_mutex);
+		adev->sdma.instance[i].funcs = &sdma_v5_0_sdma_funcs;
 		ring = &adev->sdma.instance[i].ring;
 		ring->ring_obj = NULL;
 		ring->use_doorbell = true;
@@ -1342,6 +1412,7 @@ static int sdma_v5_0_sw_init(void *handle)
 			(adev->doorbell_index.sdma_engine[0] << 1) //get DWORD offset
 			: (adev->doorbell_index.sdma_engine[1] << 1); // get DWORD offset
 
+		ring->vm_hub = AMDGPU_GFXHUB(0);
 		sprintf(ring->name, "sdma%d", i);
 		r = amdgpu_ring_init(adev, ring, 1024, &adev->sdma.trap_irq,
 				     (i == 0) ? AMDGPU_SDMA_IRQ_INSTANCE0 :
@@ -1351,28 +1422,55 @@ static int sdma_v5_0_sw_init(void *handle)
 			return r;
 	}
 
+	adev->sdma.supported_reset =
+		amdgpu_get_soft_full_reset_mask(&adev->sdma.instance[0].ring);
+	switch (amdgpu_ip_version(adev, SDMA0_HWIP, 0)) {
+	case IP_VERSION(5, 0, 0):
+	case IP_VERSION(5, 0, 2):
+	case IP_VERSION(5, 0, 5):
+		if ((adev->sdma.instance[0].fw_version >= 35) &&
+		    !amdgpu_sriov_vf(adev) &&
+		    !adev->debug_disable_gpu_ring_reset)
+			adev->sdma.supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE;
+		break;
+	default:
+		break;
+	}
+
+	/* Allocate memory for SDMA IP Dump buffer */
+	ptr = kcalloc(adev->sdma.num_instances * reg_count, sizeof(uint32_t), GFP_KERNEL);
+	if (ptr)
+		adev->sdma.ip_dump = ptr;
+	else
+		DRM_ERROR("Failed to allocated memory for SDMA IP Dump\n");
+
+	r = amdgpu_sdma_sysfs_reset_mask_init(adev);
+	if (r)
+		return r;
+
 	return r;
 }
 
-static int sdma_v5_0_sw_fini(void *handle)
+static int sdma_v5_0_sw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int i;
 
-	for (i = 0; i < adev->sdma.num_instances; i++) {
-		release_firmware(adev->sdma.instance[i].fw);
-		adev->sdma.instance[i].fw = NULL;
-
+	for (i = 0; i < adev->sdma.num_instances; i++)
 		amdgpu_ring_fini(&adev->sdma.instance[i].ring);
-	}
+
+	amdgpu_sdma_sysfs_reset_mask_fini(adev);
+	amdgpu_sdma_destroy_inst_ctx(adev, false);
+
+	kfree(adev->sdma.ip_dump);
 
 	return 0;
 }
 
-static int sdma_v5_0_hw_init(void *handle)
+static int sdma_v5_0_hw_init(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	sdma_v5_0_init_golden_registers(adev);
 
@@ -1381,9 +1479,9 @@ static int sdma_v5_0_hw_init(void *handle)
 	return r;
 }
 
-static int sdma_v5_0_hw_fini(void *handle)
+static int sdma_v5_0_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (amdgpu_sriov_vf(adev))
 		return 0;
@@ -1394,23 +1492,19 @@ static int sdma_v5_0_hw_fini(void *handle)
 	return 0;
 }
 
-static int sdma_v5_0_suspend(void *handle)
+static int sdma_v5_0_suspend(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	return sdma_v5_0_hw_fini(adev);
+	return sdma_v5_0_hw_fini(ip_block);
 }
 
-static int sdma_v5_0_resume(void *handle)
+static int sdma_v5_0_resume(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	return sdma_v5_0_hw_init(adev);
+	return sdma_v5_0_hw_init(ip_block);
 }
 
-static bool sdma_v5_0_is_idle(void *handle)
+static bool sdma_v5_0_is_idle(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	u32 i;
 
 	for (i = 0; i < adev->sdma.num_instances; i++) {
@@ -1423,11 +1517,11 @@ static bool sdma_v5_0_is_idle(void *handle)
 	return true;
 }
 
-static int sdma_v5_0_wait_for_idle(void *handle)
+static int sdma_v5_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
 {
 	unsigned i;
 	u32 sdma0, sdma1;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	for (i = 0; i < adev->usec_timeout; i++) {
 		sdma0 = RREG32(sdma_v5_0_get_reg_offset(adev, 0, mmSDMA0_STATUS_REG));
@@ -1440,13 +1534,104 @@ static int sdma_v5_0_wait_for_idle(void *handle)
 	return -ETIMEDOUT;
 }
 
-static int sdma_v5_0_soft_reset(void *handle)
+static int sdma_v5_0_soft_reset(struct amdgpu_ip_block *ip_block)
 {
 	/* todo */
 
 	return 0;
 }
 
+static int sdma_v5_0_reset_queue(struct amdgpu_ring *ring,
+				 unsigned int vmid,
+				 struct amdgpu_fence *timedout_fence)
+{
+	struct amdgpu_device *adev = ring->adev;
+	int r;
+
+	if (ring->me >= adev->sdma.num_instances) {
+		dev_err(adev->dev, "sdma instance not found\n");
+		return -EINVAL;
+	}
+
+	amdgpu_ring_reset_helper_begin(ring, timedout_fence);
+
+	amdgpu_amdkfd_suspend(adev, true);
+	r = amdgpu_sdma_reset_engine(adev, ring->me, true);
+	amdgpu_amdkfd_resume(adev, true);
+	if (r)
+		return r;
+
+	return amdgpu_ring_reset_helper_end(ring, timedout_fence);
+}
+
+static int sdma_v5_0_stop_queue(struct amdgpu_ring *ring)
+{
+	u32 f32_cntl, freeze, cntl, stat1_reg;
+	struct amdgpu_device *adev = ring->adev;
+	int i, j, r = 0;
+
+	if (amdgpu_sriov_vf(adev))
+		return -EINVAL;
+
+	i = ring->me;
+	amdgpu_gfx_rlc_enter_safe_mode(adev, 0);
+
+	/* stop queue */
+	sdma_v5_0_gfx_stop(adev, 1 << i);
+
+	/* engine stop SDMA1_F32_CNTL.HALT to 1 and SDMAx_FREEZE freeze bit to 1 */
+	freeze = RREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_FREEZE));
+	freeze = REG_SET_FIELD(freeze, SDMA0_FREEZE, FREEZE, 1);
+	WREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_FREEZE), freeze);
+
+	for (j = 0; j < adev->usec_timeout; j++) {
+		freeze = RREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_FREEZE));
+		if (REG_GET_FIELD(freeze, SDMA0_FREEZE, FROZEN) & 1)
+			break;
+		udelay(1);
+	}
+
+	/* check sdma copy engine all idle if frozen not received*/
+	if (j == adev->usec_timeout) {
+		stat1_reg = RREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_STATUS1_REG));
+		if ((stat1_reg & 0x3FF) != 0x3FF) {
+			DRM_ERROR("cannot soft reset as sdma not idle\n");
+			r = -ETIMEDOUT;
+			goto err0;
+		}
+	}
+
+	f32_cntl = RREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_F32_CNTL));
+	f32_cntl = REG_SET_FIELD(f32_cntl, SDMA0_F32_CNTL, HALT, 1);
+	WREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_F32_CNTL), f32_cntl);
+
+	cntl = RREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_CNTL));
+	cntl = REG_SET_FIELD(cntl, SDMA0_CNTL, UTC_L1_ENABLE, 0);
+	WREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_CNTL), cntl);
+err0:
+	amdgpu_gfx_rlc_exit_safe_mode(adev, 0);
+	return r;
+}
+
+static int sdma_v5_0_restore_queue(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	u32 inst_id = ring->me;
+	u32 freeze;
+	int r;
+
+	amdgpu_gfx_rlc_enter_safe_mode(adev, 0);
+	/* unfreeze*/
+	freeze = RREG32(sdma_v5_0_get_reg_offset(adev, inst_id, mmSDMA0_FREEZE));
+	freeze = REG_SET_FIELD(freeze, SDMA0_FREEZE, FREEZE, 0);
+	WREG32(sdma_v5_0_get_reg_offset(adev, inst_id, mmSDMA0_FREEZE), freeze);
+
+	r = sdma_v5_0_gfx_resume_instance(adev, inst_id, true);
+	amdgpu_gfx_rlc_exit_safe_mode(adev, 0);
+
+	return r;
+}
+
 static int sdma_v5_0_ring_preempt_ib(struct amdgpu_ring *ring)
 {
 	int i, r = 0;
@@ -1519,7 +1704,25 @@ static int sdma_v5_0_process_trap_irq(struct amdgpu_device *adev,
 				      struct amdgpu_irq_src *source,
 				      struct amdgpu_iv_entry *entry)
 {
+	uint32_t mes_queue_id = entry->src_data[0];
+
 	DRM_DEBUG("IH: SDMA trap\n");
+
+	if (adev->enable_mes && (mes_queue_id & AMDGPU_FENCE_MES_QUEUE_FLAG)) {
+		struct amdgpu_mes_queue *queue;
+
+		mes_queue_id &= AMDGPU_FENCE_MES_QUEUE_ID_MASK;
+
+		spin_lock(&adev->mes.queue_id_lock);
+		queue = idr_find(&adev->mes.queue_id_idr, mes_queue_id);
+		if (queue) {
+			DRM_DEBUG("process smda queue id = %d\n", mes_queue_id);
+			amdgpu_fence_process(queue->ring);
+		}
+		spin_unlock(&adev->mes.queue_id_lock);
+		return 0;
+	}
+
 	switch (entry->client_id) {
 	case SOC15_IH_CLIENTID_SDMA0:
 		switch (entry->ring_id) {
@@ -1626,15 +1829,15 @@ static void sdma_v5_0_update_medium_grain_light_sleep(struct amdgpu_device *adev
 	}
 }
 
-static int sdma_v5_0_set_clockgating_state(void *handle,
+static int sdma_v5_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					   enum amd_clockgating_state state)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (amdgpu_sriov_vf(adev))
 		return 0;
 
-	switch (adev->ip_versions[SDMA0_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, SDMA0_HWIP, 0)) {
 	case IP_VERSION(5, 0, 0):
 	case IP_VERSION(5, 0, 2):
 	case IP_VERSION(5, 0, 5):
@@ -1650,15 +1853,15 @@ static int sdma_v5_0_set_clockgating_state(void *handle,
 	return 0;
 }
 
-static int sdma_v5_0_set_powergating_state(void *handle,
+static int sdma_v5_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_powergating_state state)
 {
 	return 0;
 }
 
-static void sdma_v5_0_get_clockgating_state(void *handle, u32 *flags)
+static void sdma_v5_0_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int data;
 
 	if (amdgpu_sriov_vf(adev))
@@ -1675,10 +1878,51 @@ static void sdma_v5_0_get_clockgating_state(void *handle, u32 *flags)
 		*flags |= AMD_CG_SUPPORT_SDMA_LS;
 }
 
-const struct amd_ip_funcs sdma_v5_0_ip_funcs = {
+static void sdma_v5_0_print_ip_state(struct amdgpu_ip_block *ip_block, struct drm_printer *p)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, j;
+	uint32_t reg_count = ARRAY_SIZE(sdma_reg_list_5_0);
+	uint32_t instance_offset;
+
+	if (!adev->sdma.ip_dump)
+		return;
+
+	drm_printf(p, "num_instances:%d\n", adev->sdma.num_instances);
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		instance_offset = i * reg_count;
+		drm_printf(p, "\nInstance:%d\n", i);
+
+		for (j = 0; j < reg_count; j++)
+			drm_printf(p, "%-50s \t 0x%08x\n", sdma_reg_list_5_0[j].reg_name,
+				   adev->sdma.ip_dump[instance_offset + j]);
+	}
+}
+
+static void sdma_v5_0_dump_ip_state(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, j;
+	uint32_t instance_offset;
+	uint32_t reg_count = ARRAY_SIZE(sdma_reg_list_5_0);
+
+	if (!adev->sdma.ip_dump)
+		return;
+
+	amdgpu_gfx_off_ctrl(adev, false);
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		instance_offset = i * reg_count;
+		for (j = 0; j < reg_count; j++)
+			adev->sdma.ip_dump[instance_offset + j] =
+				RREG32(sdma_v5_0_get_reg_offset(adev, i,
+				       sdma_reg_list_5_0[j].reg_offset));
+	}
+	amdgpu_gfx_off_ctrl(adev, true);
+}
+
+static const struct amd_ip_funcs sdma_v5_0_ip_funcs = {
 	.name = "sdma_v5_0",
 	.early_init = sdma_v5_0_early_init,
-	.late_init = NULL,
 	.sw_init = sdma_v5_0_sw_init,
 	.sw_fini = sdma_v5_0_sw_fini,
 	.hw_init = sdma_v5_0_hw_init,
@@ -1691,6 +1935,8 @@ const struct amd_ip_funcs sdma_v5_0_ip_funcs = {
 	.set_clockgating_state = sdma_v5_0_set_clockgating_state,
 	.set_powergating_state = sdma_v5_0_set_powergating_state,
 	.get_clockgating_state = sdma_v5_0_get_clockgating_state,
+	.dump_ip_state = sdma_v5_0_dump_ip_state,
+	.print_ip_state = sdma_v5_0_print_ip_state,
 };
 
 static const struct amdgpu_ring_funcs sdma_v5_0_ring_funcs = {
@@ -1698,7 +1944,7 @@ static const struct amdgpu_ring_funcs sdma_v5_0_ring_funcs = {
 	.align_mask = 0xf,
 	.nop = SDMA_PKT_NOP_HEADER_OP(SDMA_OP_NOP),
 	.support_64bit_ptrs = true,
-	.vmhub = AMDGPU_GFXHUB_0,
+	.secure_submission_supported = true,
 	.get_rptr = sdma_v5_0_ring_get_rptr,
 	.get_wptr = sdma_v5_0_ring_get_wptr,
 	.set_wptr = sdma_v5_0_ring_set_wptr,
@@ -1726,8 +1972,8 @@ static const struct amdgpu_ring_funcs sdma_v5_0_ring_funcs = {
 	.emit_reg_wait = sdma_v5_0_ring_emit_reg_wait,
 	.emit_reg_write_reg_wait = sdma_v5_0_ring_emit_reg_write_reg_wait,
 	.init_cond_exec = sdma_v5_0_ring_init_cond_exec,
-	.patch_cond_exec = sdma_v5_0_ring_patch_cond_exec,
 	.preempt_ib = sdma_v5_0_ring_preempt_ib,
+	.reset = sdma_v5_0_reset_queue,
 };
 
 static void sdma_v5_0_set_ring_funcs(struct amdgpu_device *adev)
@@ -1764,7 +2010,7 @@ static void sdma_v5_0_set_irq_funcs(struct amdgpu_device *adev)
  * @src_offset: src GPU address
  * @dst_offset: dst GPU address
  * @byte_count: number of bytes to xfer
- * @tmz: if a secure copy should be used
+ * @copy_flags: copy flags for the buffers
  *
  * Copy GPU buffers using the DMA engine (NAVI10).
  * Used by the amdgpu ttm implementation to move pages if
@@ -1774,11 +2020,11 @@ static void sdma_v5_0_emit_copy_buffer(struct amdgpu_ib *ib,
 				       uint64_t src_offset,
 				       uint64_t dst_offset,
 				       uint32_t byte_count,
-				       bool tmz)
+				       uint32_t copy_flags)
 {
 	ib->ptr[ib->length_dw++] = SDMA_PKT_HEADER_OP(SDMA_OP_COPY) |
 		SDMA_PKT_HEADER_SUB_OP(SDMA_SUBOP_COPY_LINEAR) |
-		SDMA_PKT_COPY_LINEAR_HEADER_TMZ(tmz ? 1 : 0);
+		SDMA_PKT_COPY_LINEAR_HEADER_TMZ((copy_flags & AMDGPU_COPY_FLAGS_TMZ) ? 1 : 0);
 	ib->ptr[ib->length_dw++] = byte_count - 1;
 	ib->ptr[ib->length_dw++] = 0; /* src/dst endian swap */
 	ib->ptr[ib->length_dw++] = lower_32_bits(src_offset);
diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.h b/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.h
index d4e3c2e696f6..2ab71f21755a 100644
--- a/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.h
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.h
@@ -24,7 +24,6 @@
 #ifndef __SDMA_V5_0_H__
 #define __SDMA_V5_0_H__
 
-extern const struct amd_ip_funcs sdma_v5_0_ip_funcs;
 extern const struct amdgpu_ip_block_version sdma_v5_0_ip_block;
 
 #endif /* __SDMA_V5_0_H__ */
diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c b/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c
index 4d4d1aa51b8a..51101b0aa2fa 100644
--- a/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c
@@ -51,6 +51,8 @@ MODULE_FIRMWARE("amdgpu/beige_goby_sdma.bin");
 
 MODULE_FIRMWARE("amdgpu/vangogh_sdma.bin");
 MODULE_FIRMWARE("amdgpu/yellow_carp_sdma.bin");
+MODULE_FIRMWARE("amdgpu/sdma_5_2_6.bin");
+MODULE_FIRMWARE("amdgpu/sdma_5_2_7.bin");
 
 #define SDMA1_REG_OFFSET 0x600
 #define SDMA3_REG_OFFSET 0x400
@@ -58,10 +60,61 @@ MODULE_FIRMWARE("amdgpu/yellow_carp_sdma.bin");
 #define SDMA0_HYP_DEC_REG_END 0x5893
 #define SDMA1_HYP_DEC_REG_OFFSET 0x20
 
+static const struct amdgpu_hwip_reg_entry sdma_reg_list_5_2[] = {
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_STATUS_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_STATUS1_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_STATUS2_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_STATUS3_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_UCODE_CHECKSUM),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RB_RPTR_FETCH_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RB_RPTR_FETCH),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_UTCL1_RD_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_UTCL1_WR_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_UTCL1_RD_XNACK0),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_UTCL1_RD_XNACK1),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_UTCL1_WR_XNACK0),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_UTCL1_WR_XNACK1),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_RB_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_RB_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_RB_RPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_RB_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_RB_WPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_IB_OFFSET),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_IB_BASE_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_IB_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_IB_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_IB_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_IB_SUB_REMAIN),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_DUMMY_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_PAGE_RB_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_PAGE_RB_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_PAGE_RB_RPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_PAGE_RB_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_PAGE_RB_WPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_PAGE_IB_OFFSET),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_PAGE_IB_BASE_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_PAGE_IB_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_PAGE_DUMMY_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RLC0_RB_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RLC0_RB_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RLC0_RB_RPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RLC0_RB_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RLC0_RB_WPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RLC0_IB_OFFSET),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RLC0_IB_BASE_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RLC0_IB_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RLC0_DUMMY_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_INT_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_VM_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, mmGRBM_STATUS2)
+};
+
 static void sdma_v5_2_set_ring_funcs(struct amdgpu_device *adev);
 static void sdma_v5_2_set_buffer_funcs(struct amdgpu_device *adev);
 static void sdma_v5_2_set_vm_pte_funcs(struct amdgpu_device *adev);
 static void sdma_v5_2_set_irq_funcs(struct amdgpu_device *adev);
+static int sdma_v5_2_stop_queue(struct amdgpu_ring *ring);
+static int sdma_v5_2_restore_queue(struct amdgpu_ring *ring);
 
 static u32 sdma_v5_2_get_reg_offset(struct amdgpu_device *adev, u32 instance, u32 internal_offset)
 {
@@ -87,147 +140,23 @@ static u32 sdma_v5_2_get_reg_offset(struct amdgpu_device *adev, u32 instance, u3
 	return base + internal_offset;
 }
 
-static int sdma_v5_2_init_inst_ctx(struct amdgpu_sdma_instance *sdma_inst)
-{
-	int err = 0;
-	const struct sdma_firmware_header_v1_0 *hdr;
-
-	err = amdgpu_ucode_validate(sdma_inst->fw);
-	if (err)
-		return err;
-
-	hdr = (const struct sdma_firmware_header_v1_0 *)sdma_inst->fw->data;
-	sdma_inst->fw_version = le32_to_cpu(hdr->header.ucode_version);
-	sdma_inst->feature_version = le32_to_cpu(hdr->ucode_feature_version);
-
-	if (sdma_inst->feature_version >= 20)
-		sdma_inst->burst_nop = true;
-
-	return 0;
-}
-
-static void sdma_v5_2_destroy_inst_ctx(struct amdgpu_device *adev)
-{
-	release_firmware(adev->sdma.instance[0].fw);
-
-	memset((void *)adev->sdma.instance, 0,
-	       sizeof(struct amdgpu_sdma_instance) * AMDGPU_MAX_SDMA_INSTANCES);
-}
-
-/**
- * sdma_v5_2_init_microcode - load ucode images from disk
- *
- * @adev: amdgpu_device pointer
- *
- * Use the firmware interface to load the ucode images into
- * the driver (not loaded into hw).
- * Returns 0 on success, error on failure.
- */
-
-// emulation only, won't work on real chip
-// navi10 real chip need to use PSP to load firmware
-static int sdma_v5_2_init_microcode(struct amdgpu_device *adev)
-{
-	const char *chip_name;
-	char fw_name[40];
-	int err = 0, i;
-	struct amdgpu_firmware_info *info = NULL;
-	const struct common_firmware_header *header = NULL;
-
-	DRM_DEBUG("\n");
-
-	switch (adev->ip_versions[SDMA0_HWIP][0]) {
-	case IP_VERSION(5, 2, 0):
-		chip_name = "sienna_cichlid";
-		break;
-	case IP_VERSION(5, 2, 2):
-		chip_name = "navy_flounder";
-		break;
-	case IP_VERSION(5, 2, 1):
-		chip_name = "vangogh";
-		break;
-	case IP_VERSION(5, 2, 4):
-		chip_name = "dimgrey_cavefish";
-		break;
-	case IP_VERSION(5, 2, 5):
-		chip_name = "beige_goby";
-		break;
-	case IP_VERSION(5, 2, 3):
-		chip_name = "yellow_carp";
-		break;
-	default:
-		BUG();
-	}
-
-	snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_sdma.bin", chip_name);
-
-	err = request_firmware(&adev->sdma.instance[0].fw, fw_name, adev->dev);
-	if (err)
-		goto out;
-
-	err = sdma_v5_2_init_inst_ctx(&adev->sdma.instance[0]);
-	if (err)
-		goto out;
-
-	for (i = 1; i < adev->sdma.num_instances; i++)
-		memcpy((void *)&adev->sdma.instance[i],
-		       (void *)&adev->sdma.instance[0],
-		       sizeof(struct amdgpu_sdma_instance));
-
-	if (amdgpu_sriov_vf(adev) && (adev->ip_versions[SDMA0_HWIP][0] == IP_VERSION(5, 2, 0)))
-		return 0;
-
-	DRM_DEBUG("psp_load == '%s'\n",
-		  adev->firmware.load_type == AMDGPU_FW_LOAD_PSP ? "true" : "false");
-
-	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
-		for (i = 0; i < adev->sdma.num_instances; i++) {
-			info = &adev->firmware.ucode[AMDGPU_UCODE_ID_SDMA0 + i];
-			info->ucode_id = AMDGPU_UCODE_ID_SDMA0 + i;
-			info->fw = adev->sdma.instance[i].fw;
-			header = (const struct common_firmware_header *)info->fw->data;
-			adev->firmware.fw_size +=
-				ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE);
-		}
-	}
-
-out:
-	if (err) {
-		DRM_ERROR("sdma_v5_2: Failed to load firmware \"%s\"\n", fw_name);
-		sdma_v5_2_destroy_inst_ctx(adev);
-	}
-	return err;
-}
-
-static unsigned sdma_v5_2_ring_init_cond_exec(struct amdgpu_ring *ring)
+static unsigned sdma_v5_2_ring_init_cond_exec(struct amdgpu_ring *ring,
+					      uint64_t addr)
 {
 	unsigned ret;
 
 	amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_COND_EXE));
-	amdgpu_ring_write(ring, lower_32_bits(ring->cond_exe_gpu_addr));
-	amdgpu_ring_write(ring, upper_32_bits(ring->cond_exe_gpu_addr));
+	amdgpu_ring_write(ring, lower_32_bits(addr));
+	amdgpu_ring_write(ring, upper_32_bits(addr));
 	amdgpu_ring_write(ring, 1);
-	ret = ring->wptr & ring->buf_mask;/* this is the offset we need patch later */
-	amdgpu_ring_write(ring, 0x55aa55aa);/* insert dummy here and patch it later */
+	/* this is the offset we need patch later */
+	ret = ring->wptr & ring->buf_mask;
+	/* insert dummy here and patch it later */
+	amdgpu_ring_write(ring, 0);
 
 	return ret;
 }
 
-static void sdma_v5_2_ring_patch_cond_exec(struct amdgpu_ring *ring,
-					   unsigned offset)
-{
-	unsigned cur;
-
-	BUG_ON(offset > ring->buf_mask);
-	BUG_ON(ring->ring[offset] != 0x55aa55aa);
-
-	cur = (ring->wptr - 1) & ring->buf_mask;
-	if (cur > offset)
-		ring->ring[offset] = cur - offset;
-	else
-		ring->ring[offset] = (ring->buf_mask + 1) - offset + cur;
-}
-
 /**
  * sdma_v5_2_ring_get_rptr - get the current read pointer
  *
@@ -240,7 +169,7 @@ static uint64_t sdma_v5_2_ring_get_rptr(struct amdgpu_ring *ring)
 	u64 *rptr;
 
 	/* XXX check if swapping is necessary on BE */
-	rptr = ((u64 *)&ring->adev->wb.wb[ring->rptr_offs]);
+	rptr = (u64 *)ring->rptr_cpu_addr;
 
 	DRM_DEBUG("rptr before shift == 0x%016llx\n", *rptr);
 	return ((*rptr) >> 2);
@@ -260,7 +189,7 @@ static uint64_t sdma_v5_2_ring_get_wptr(struct amdgpu_ring *ring)
 
 	if (ring->use_doorbell) {
 		/* XXX check if swapping is necessary on BE */
-		wptr = READ_ONCE(*((u64 *)&adev->wb.wb[ring->wptr_offs]));
+		wptr = READ_ONCE(*((u64 *)ring->wptr_cpu_addr));
 		DRM_DEBUG("wptr/doorbell before shift == 0x%016llx\n", wptr);
 	} else {
 		wptr = RREG32(sdma_v5_2_get_reg_offset(adev, ring->me, mmSDMA0_GFX_RB_WPTR_HI));
@@ -287,17 +216,27 @@ static void sdma_v5_2_ring_set_wptr(struct amdgpu_ring *ring)
 	if (ring->use_doorbell) {
 		DRM_DEBUG("Using doorbell -- "
 				"wptr_offs == 0x%08x "
-				"lower_32_bits(ring->wptr) << 2 == 0x%08x "
-				"upper_32_bits(ring->wptr) << 2 == 0x%08x\n",
+				"lower_32_bits(ring->wptr << 2) == 0x%08x "
+				"upper_32_bits(ring->wptr << 2) == 0x%08x\n",
 				ring->wptr_offs,
 				lower_32_bits(ring->wptr << 2),
 				upper_32_bits(ring->wptr << 2));
 		/* XXX check if swapping is necessary on BE */
-		adev->wb.wb[ring->wptr_offs] = lower_32_bits(ring->wptr << 2);
-		adev->wb.wb[ring->wptr_offs + 1] = upper_32_bits(ring->wptr << 2);
+		atomic64_set((atomic64_t *)ring->wptr_cpu_addr,
+			     ring->wptr << 2);
 		DRM_DEBUG("calling WDOORBELL64(0x%08x, 0x%016llx)\n",
 				ring->doorbell_index, ring->wptr << 2);
 		WDOORBELL64(ring->doorbell_index, ring->wptr << 2);
+		if (amdgpu_ip_version(adev, SDMA0_HWIP, 0) == IP_VERSION(5, 2, 1)) {
+			/* SDMA seems to miss doorbells sometimes when powergating kicks in.
+			 * Updating the wptr directly will wake it. This is only safe because
+			 * we disallow gfxoff in begin_use() and then allow it again in end_use().
+			 */
+			WREG32(sdma_v5_2_get_reg_offset(adev, ring->me, mmSDMA0_GFX_RB_WPTR),
+			       lower_32_bits(ring->wptr << 2));
+			WREG32(sdma_v5_2_get_reg_offset(adev, ring->me, mmSDMA0_GFX_RB_WPTR_HI),
+			       upper_32_bits(ring->wptr << 2));
+		}
 	} else {
 		DRM_DEBUG("Not using doorbell -- "
 				"mmSDMA%i_GFX_RB_WPTR == 0x%08x "
@@ -368,8 +307,6 @@ static void sdma_v5_2_ring_emit_ib(struct amdgpu_ring *ring,
  * sdma_v5_2_ring_emit_mem_sync - flush the IB by graphics cache rinse
  *
  * @ring: amdgpu ring pointer
- * @job: job to retrieve vmid from
- * @ib: IB object to schedule
  *
  * flush the IB by graphics cache rinse.
  */
@@ -404,17 +341,21 @@ static void sdma_v5_2_ring_emit_hdp_flush(struct amdgpu_ring *ring)
 	u32 ref_and_mask = 0;
 	const struct nbio_hdp_flush_reg *nbio_hf_reg = adev->nbio.hdp_flush_reg;
 
-	ref_and_mask = nbio_hf_reg->ref_and_mask_sdma0 << ring->me;
-
-	amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_POLL_REGMEM) |
-			  SDMA_PKT_POLL_REGMEM_HEADER_HDP_FLUSH(1) |
-			  SDMA_PKT_POLL_REGMEM_HEADER_FUNC(3)); /* == */
-	amdgpu_ring_write(ring, (adev->nbio.funcs->get_hdp_flush_done_offset(adev)) << 2);
-	amdgpu_ring_write(ring, (adev->nbio.funcs->get_hdp_flush_req_offset(adev)) << 2);
-	amdgpu_ring_write(ring, ref_and_mask); /* reference */
-	amdgpu_ring_write(ring, ref_and_mask); /* mask */
-	amdgpu_ring_write(ring, SDMA_PKT_POLL_REGMEM_DW5_RETRY_COUNT(0xfff) |
-			  SDMA_PKT_POLL_REGMEM_DW5_INTERVAL(10)); /* retry count, poll interval */
+	if (ring->me > 1) {
+		amdgpu_hdp_flush(adev, ring);
+	} else {
+		ref_and_mask = nbio_hf_reg->ref_and_mask_sdma0 << ring->me;
+
+		amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_POLL_REGMEM) |
+				  SDMA_PKT_POLL_REGMEM_HEADER_HDP_FLUSH(1) |
+				  SDMA_PKT_POLL_REGMEM_HEADER_FUNC(3)); /* == */
+		amdgpu_ring_write(ring, (adev->nbio.funcs->get_hdp_flush_done_offset(adev)) << 2);
+		amdgpu_ring_write(ring, (adev->nbio.funcs->get_hdp_flush_req_offset(adev)) << 2);
+		amdgpu_ring_write(ring, ref_and_mask); /* reference */
+		amdgpu_ring_write(ring, ref_and_mask); /* mask */
+		amdgpu_ring_write(ring, SDMA_PKT_POLL_REGMEM_DW5_RETRY_COUNT(0xfff) |
+				  SDMA_PKT_POLL_REGMEM_DW5_INTERVAL(10)); /* retry count, poll interval */
+	}
 }
 
 /**
@@ -454,7 +395,7 @@ static void sdma_v5_2_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 se
 		amdgpu_ring_write(ring, upper_32_bits(seq));
 	}
 
-	if (flags & AMDGPU_FENCE_FLAG_INT) {
+	if ((flags & AMDGPU_FENCE_FLAG_INT)) {
 		/* generate an interrupt */
 		amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_TRAP));
 		amdgpu_ring_write(ring, SDMA_PKT_TRAP_INT_CONTEXT_INT_CONTEXT(0));
@@ -466,25 +407,15 @@ static void sdma_v5_2_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 se
  * sdma_v5_2_gfx_stop - stop the gfx async dma engines
  *
  * @adev: amdgpu_device pointer
- *
+ * @inst_mask: mask of dma engine instances to be disabled
  * Stop the gfx async dma ring buffers.
  */
-static void sdma_v5_2_gfx_stop(struct amdgpu_device *adev)
+static void sdma_v5_2_gfx_stop(struct amdgpu_device *adev,  uint32_t inst_mask)
 {
-	struct amdgpu_ring *sdma0 = &adev->sdma.instance[0].ring;
-	struct amdgpu_ring *sdma1 = &adev->sdma.instance[1].ring;
-	struct amdgpu_ring *sdma2 = &adev->sdma.instance[2].ring;
-	struct amdgpu_ring *sdma3 = &adev->sdma.instance[3].ring;
 	u32 rb_cntl, ib_cntl;
 	int i;
 
-	if ((adev->mman.buffer_funcs_ring == sdma0) ||
-	    (adev->mman.buffer_funcs_ring == sdma1) ||
-	    (adev->mman.buffer_funcs_ring == sdma2) ||
-	    (adev->mman.buffer_funcs_ring == sdma3))
-		amdgpu_ttm_set_buffer_funcs_status(adev, false);
-
-	for (i = 0; i < adev->sdma.num_instances; i++) {
+	for_each_inst(i, inst_mask) {
 		rb_cntl = RREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_CNTL));
 		rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL, RB_ENABLE, 0);
 		WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_CNTL), rb_cntl);
@@ -544,9 +475,6 @@ static void sdma_v5_2_ctx_switch_enable(struct amdgpu_device *adev, bool enable)
 	}
 
 	for (i = 0; i < adev->sdma.num_instances; i++) {
-		f32_cntl = RREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_CNTL));
-		f32_cntl = REG_SET_FIELD(f32_cntl, SDMA0_CNTL,
-				AUTO_CTXSW_ENABLE, enable ? 1 : 0);
 		if (enable && amdgpu_sdma_phase_quantum) {
 			WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_PHASE0_QUANTUM),
 			       phase_quantum);
@@ -555,7 +483,13 @@ static void sdma_v5_2_ctx_switch_enable(struct amdgpu_device *adev, bool enable)
 			WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_PHASE2_QUANTUM),
 			       phase_quantum);
 		}
-		WREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_CNTL), f32_cntl);
+
+		if (!amdgpu_sriov_vf(adev)) {
+			f32_cntl = RREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_CNTL));
+			f32_cntl = REG_SET_FIELD(f32_cntl, SDMA0_CNTL,
+					AUTO_CTXSW_ENABLE, enable ? 1 : 0);
+			WREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_CNTL), f32_cntl);
+		}
 	}
 
 }
@@ -572,121 +506,136 @@ static void sdma_v5_2_enable(struct amdgpu_device *adev, bool enable)
 {
 	u32 f32_cntl;
 	int i;
+	uint32_t inst_mask;
 
+	inst_mask = GENMASK(adev->sdma.num_instances - 1, 0);
 	if (!enable) {
-		sdma_v5_2_gfx_stop(adev);
+		sdma_v5_2_gfx_stop(adev, inst_mask);
 		sdma_v5_2_rlc_stop(adev);
 	}
 
-	for (i = 0; i < adev->sdma.num_instances; i++) {
-		f32_cntl = RREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_F32_CNTL));
-		f32_cntl = REG_SET_FIELD(f32_cntl, SDMA0_F32_CNTL, HALT, enable ? 0 : 1);
-		WREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_F32_CNTL), f32_cntl);
+	if (!amdgpu_sriov_vf(adev)) {
+		for (i = 0; i < adev->sdma.num_instances; i++) {
+			f32_cntl = RREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_F32_CNTL));
+			f32_cntl = REG_SET_FIELD(f32_cntl, SDMA0_F32_CNTL, HALT, enable ? 0 : 1);
+			WREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_F32_CNTL), f32_cntl);
+		}
 	}
 }
 
 /**
- * sdma_v5_2_gfx_resume - setup and start the async dma engines
+ * sdma_v5_2_gfx_resume_instance - start/restart a certain sdma engine
  *
  * @adev: amdgpu_device pointer
+ * @i: instance
+ * @restore: used to restore wptr when restart
  *
- * Set up the gfx DMA ring buffers and enable them.
- * Returns 0 for success, error for failure.
+ * Set up the gfx DMA ring buffers and enable them. On restart, we will restore wptr and rptr.
+ * Return 0 for success.
  */
-static int sdma_v5_2_gfx_resume(struct amdgpu_device *adev)
+
+static int sdma_v5_2_gfx_resume_instance(struct amdgpu_device *adev, int i, bool restore)
 {
 	struct amdgpu_ring *ring;
 	u32 rb_cntl, ib_cntl;
 	u32 rb_bufsz;
-	u32 wb_offset;
 	u32 doorbell;
 	u32 doorbell_offset;
 	u32 temp;
 	u32 wptr_poll_cntl;
 	u64 wptr_gpu_addr;
-	int i, r;
 
-	for (i = 0; i < adev->sdma.num_instances; i++) {
-		ring = &adev->sdma.instance[i].ring;
-		wb_offset = (ring->rptr_offs * 4);
+	ring = &adev->sdma.instance[i].ring;
 
+	if (!amdgpu_sriov_vf(adev))
 		WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_SEM_WAIT_FAIL_TIMER_CNTL), 0);
 
-		/* Set ring buffer size in dwords */
-		rb_bufsz = order_base_2(ring->ring_size / 4);
-		rb_cntl = RREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_CNTL));
-		rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL, RB_SIZE, rb_bufsz);
+	/* Set ring buffer size in dwords */
+	rb_bufsz = order_base_2(ring->ring_size / 4);
+	rb_cntl = RREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_CNTL));
+	rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL, RB_SIZE, rb_bufsz);
 #ifdef __BIG_ENDIAN
-		rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL, RB_SWAP_ENABLE, 1);
-		rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL,
-					RPTR_WRITEBACK_SWAP_ENABLE, 1);
+	rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL, RB_SWAP_ENABLE, 1);
+	rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL,
+				RPTR_WRITEBACK_SWAP_ENABLE, 1);
 #endif
-		WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_CNTL), rb_cntl);
-
-		/* Initialize the ring buffer's read and write pointers */
+	WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_CNTL), rb_cntl);
+
+	/* Initialize the ring buffer's read and write pointers */
+	if (restore) {
+		WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_RPTR), lower_32_bits(ring->wptr << 2));
+		WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_RPTR_HI), upper_32_bits(ring->wptr << 2));
+		WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_WPTR), lower_32_bits(ring->wptr << 2));
+		WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_WPTR_HI), upper_32_bits(ring->wptr << 2));
+	} else {
 		WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_RPTR), 0);
 		WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_RPTR_HI), 0);
 		WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_WPTR), 0);
 		WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_WPTR_HI), 0);
+	}
 
-		/* setup the wptr shadow polling */
-		wptr_gpu_addr = adev->wb.gpu_addr + (ring->wptr_offs * 4);
-		WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_WPTR_POLL_ADDR_LO),
-		       lower_32_bits(wptr_gpu_addr));
-		WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_WPTR_POLL_ADDR_HI),
-		       upper_32_bits(wptr_gpu_addr));
-		wptr_poll_cntl = RREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i,
-							 mmSDMA0_GFX_RB_WPTR_POLL_CNTL));
-		wptr_poll_cntl = REG_SET_FIELD(wptr_poll_cntl,
-					       SDMA0_GFX_RB_WPTR_POLL_CNTL,
-					       F32_POLL_ENABLE, 1);
-		WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_WPTR_POLL_CNTL),
-		       wptr_poll_cntl);
-
-		/* set the wb address whether it's enabled or not */
-		WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_RPTR_ADDR_HI),
-		       upper_32_bits(adev->wb.gpu_addr + wb_offset) & 0xFFFFFFFF);
-		WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_RPTR_ADDR_LO),
-		       lower_32_bits(adev->wb.gpu_addr + wb_offset) & 0xFFFFFFFC);
-
-		rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL, RPTR_WRITEBACK_ENABLE, 1);
-
-		WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_BASE), ring->gpu_addr >> 8);
-		WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_BASE_HI), ring->gpu_addr >> 40);
-
+	/* setup the wptr shadow polling */
+	wptr_gpu_addr = ring->wptr_gpu_addr;
+	WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_WPTR_POLL_ADDR_LO),
+	       lower_32_bits(wptr_gpu_addr));
+	WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_WPTR_POLL_ADDR_HI),
+	       upper_32_bits(wptr_gpu_addr));
+	wptr_poll_cntl = RREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i,
+						 mmSDMA0_GFX_RB_WPTR_POLL_CNTL));
+	wptr_poll_cntl = REG_SET_FIELD(wptr_poll_cntl,
+				       SDMA0_GFX_RB_WPTR_POLL_CNTL,
+				       F32_POLL_ENABLE, 1);
+	WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_WPTR_POLL_CNTL),
+	       wptr_poll_cntl);
+
+	/* set the wb address whether it's enabled or not */
+	WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_RPTR_ADDR_HI),
+	       upper_32_bits(ring->rptr_gpu_addr) & 0xFFFFFFFF);
+	WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_RPTR_ADDR_LO),
+	       lower_32_bits(ring->rptr_gpu_addr) & 0xFFFFFFFC);
+
+	rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL, RPTR_WRITEBACK_ENABLE, 1);
+
+	WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_BASE), ring->gpu_addr >> 8);
+	WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_BASE_HI), ring->gpu_addr >> 40);
+
+	if (!restore)
 		ring->wptr = 0;
 
-		/* before programing wptr to a less value, need set minor_ptr_update first */
-		WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_MINOR_PTR_UPDATE), 1);
+	/* before programing wptr to a less value, need set minor_ptr_update first */
+	WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_MINOR_PTR_UPDATE), 1);
 
-		if (!amdgpu_sriov_vf(adev)) { /* only bare-metal use register write for wptr */
-			WREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_WPTR), lower_32_bits(ring->wptr) << 2);
-			WREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_WPTR_HI), upper_32_bits(ring->wptr) << 2);
-		}
+	if (!amdgpu_sriov_vf(adev)) { /* only bare-metal use register write for wptr */
+		WREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_WPTR), lower_32_bits(ring->wptr << 2));
+		WREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_WPTR_HI), upper_32_bits(ring->wptr << 2));
+	}
 
-		doorbell = RREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_DOORBELL));
-		doorbell_offset = RREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_DOORBELL_OFFSET));
+	doorbell = RREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_DOORBELL));
+	doorbell_offset = RREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_DOORBELL_OFFSET));
 
-		if (ring->use_doorbell) {
-			doorbell = REG_SET_FIELD(doorbell, SDMA0_GFX_DOORBELL, ENABLE, 1);
-			doorbell_offset = REG_SET_FIELD(doorbell_offset, SDMA0_GFX_DOORBELL_OFFSET,
-					OFFSET, ring->doorbell_index);
-		} else {
-			doorbell = REG_SET_FIELD(doorbell, SDMA0_GFX_DOORBELL, ENABLE, 0);
-		}
-		WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_DOORBELL), doorbell);
-		WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_DOORBELL_OFFSET), doorbell_offset);
+	if (ring->use_doorbell) {
+		doorbell = REG_SET_FIELD(doorbell, SDMA0_GFX_DOORBELL, ENABLE, 1);
+		doorbell_offset = REG_SET_FIELD(doorbell_offset, SDMA0_GFX_DOORBELL_OFFSET,
+				OFFSET, ring->doorbell_index);
+	} else {
+		doorbell = REG_SET_FIELD(doorbell, SDMA0_GFX_DOORBELL, ENABLE, 0);
+	}
+	WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_DOORBELL), doorbell);
+	WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_DOORBELL_OFFSET), doorbell_offset);
 
-		adev->nbio.funcs->sdma_doorbell_range(adev, i, ring->use_doorbell,
-						      ring->doorbell_index,
-						      adev->doorbell_index.sdma_doorbell_range);
+	adev->nbio.funcs->sdma_doorbell_range(adev, i, ring->use_doorbell,
+					      ring->doorbell_index,
+					      adev->doorbell_index.sdma_doorbell_range);
 
-		if (amdgpu_sriov_vf(adev))
-			sdma_v5_2_ring_set_wptr(ring);
+	if (amdgpu_sriov_vf(adev))
+		sdma_v5_2_ring_set_wptr(ring);
 
-		/* set minor_ptr_update to 0 after wptr programed */
-		WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_MINOR_PTR_UPDATE), 0);
+	/* set minor_ptr_update to 0 after wptr programed */
 
+	WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_MINOR_PTR_UPDATE), 0);
+
+	/* SRIOV VF has no control of any of registers below */
+	if (!amdgpu_sriov_vf(adev)) {
 		/* set utc l1 enable flag always to 1 */
 		temp = RREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_CNTL));
 		temp = REG_SET_FIELD(temp, SDMA0_CNTL, UTC_L1_ENABLE, 1);
@@ -710,40 +659,48 @@ static int sdma_v5_2_gfx_resume(struct amdgpu_device *adev)
 			 SDMA0_UTCL1_PAGE__LLC_NOALLOC_MASK);
 		WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_UTCL1_PAGE), temp);
 
-		if (!amdgpu_sriov_vf(adev)) {
-			/* unhalt engine */
-			temp = RREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_F32_CNTL));
-			temp = REG_SET_FIELD(temp, SDMA0_F32_CNTL, HALT, 0);
-			WREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_F32_CNTL), temp);
-		}
+		/* unhalt engine */
+		temp = RREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_F32_CNTL));
+		temp = REG_SET_FIELD(temp, SDMA0_F32_CNTL, HALT, 0);
+		WREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_F32_CNTL), temp);
+	}
 
-		/* enable DMA RB */
-		rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL, RB_ENABLE, 1);
-		WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_CNTL), rb_cntl);
+	/* enable DMA RB */
+	rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL, RB_ENABLE, 1);
+	WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_CNTL), rb_cntl);
 
-		ib_cntl = RREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_IB_CNTL));
-		ib_cntl = REG_SET_FIELD(ib_cntl, SDMA0_GFX_IB_CNTL, IB_ENABLE, 1);
+	ib_cntl = RREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_IB_CNTL));
+	ib_cntl = REG_SET_FIELD(ib_cntl, SDMA0_GFX_IB_CNTL, IB_ENABLE, 1);
 #ifdef __BIG_ENDIAN
-		ib_cntl = REG_SET_FIELD(ib_cntl, SDMA0_GFX_IB_CNTL, IB_SWAP_ENABLE, 1);
+	ib_cntl = REG_SET_FIELD(ib_cntl, SDMA0_GFX_IB_CNTL, IB_SWAP_ENABLE, 1);
 #endif
-		/* enable DMA IBs */
-		WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_IB_CNTL), ib_cntl);
+	/* enable DMA IBs */
+	WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_IB_CNTL), ib_cntl);
 
-		ring->sched.ready = true;
+	if (amdgpu_sriov_vf(adev)) { /* bare-metal sequence doesn't need below to lines */
+		sdma_v5_2_ctx_switch_enable(adev, true);
+		sdma_v5_2_enable(adev, true);
+	}
 
-		if (amdgpu_sriov_vf(adev)) { /* bare-metal sequence doesn't need below to lines */
-			sdma_v5_2_ctx_switch_enable(adev, true);
-			sdma_v5_2_enable(adev, true);
-		}
+	return amdgpu_ring_test_helper(ring);
+}
 
-		r = amdgpu_ring_test_ring(ring);
-		if (r) {
-			ring->sched.ready = false;
-			return r;
-		}
+/**
+ * sdma_v5_2_gfx_resume - setup and start the async dma engines
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Set up the gfx DMA ring buffers and enable them.
+ * Returns 0 for success, error for failure.
+ */
+static int sdma_v5_2_gfx_resume(struct amdgpu_device *adev)
+{
+	int i, r;
 
-		if (adev->mman.buffer_funcs_ring == ring)
-			amdgpu_ttm_set_buffer_funcs_status(adev, true);
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		r = sdma_v5_2_gfx_resume_instance(adev, i, false);
+		if (r)
+			return r;
 	}
 
 	return 0;
@@ -806,37 +763,49 @@ static int sdma_v5_2_load_microcode(struct amdgpu_device *adev)
 	return 0;
 }
 
-static int sdma_v5_2_soft_reset(void *handle)
+static int sdma_v5_2_soft_reset_engine(struct amdgpu_device *adev, u32 instance_id)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 	u32 grbm_soft_reset;
 	u32 tmp;
-	int i;
 
-	for (i = 0; i < adev->sdma.num_instances; i++) {
-		grbm_soft_reset = REG_SET_FIELD(0,
-						GRBM_SOFT_RESET, SOFT_RESET_SDMA0,
-						1);
-		grbm_soft_reset <<= i;
+	grbm_soft_reset = REG_SET_FIELD(0,
+					GRBM_SOFT_RESET, SOFT_RESET_SDMA0,
+					1);
+	grbm_soft_reset <<= instance_id;
 
-		tmp = RREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET);
-		tmp |= grbm_soft_reset;
-		DRM_DEBUG("GRBM_SOFT_RESET=0x%08X\n", tmp);
-		WREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET, tmp);
-		tmp = RREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET);
+	tmp = RREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET);
+	tmp |= grbm_soft_reset;
+	DRM_DEBUG("GRBM_SOFT_RESET=0x%08X\n", tmp);
+	WREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET, tmp);
+	tmp = RREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET);
 
-		udelay(50);
+	udelay(50);
 
-		tmp &= ~grbm_soft_reset;
-		WREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET, tmp);
-		tmp = RREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET);
+	tmp &= ~grbm_soft_reset;
+	WREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET, tmp);
+	tmp = RREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET);
+	return 0;
+}
 
+static int sdma_v5_2_soft_reset(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i;
+
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		sdma_v5_2_soft_reset_engine(adev, i);
 		udelay(50);
 	}
 
 	return 0;
 }
 
+static const struct amdgpu_sdma_funcs sdma_v5_2_sdma_funcs = {
+	.stop_kernel_queue = &sdma_v5_2_stop_queue,
+	.start_kernel_queue = &sdma_v5_2_restore_queue,
+	.soft_reset_kernel_queue = &sdma_v5_2_soft_reset_engine,
+};
+
 /**
  * sdma_v5_2_start - setup and start the async dma engines
  *
@@ -848,6 +817,7 @@ static int sdma_v5_2_soft_reset(void *handle)
 static int sdma_v5_2_start(struct amdgpu_device *adev)
 {
 	int r = 0;
+	struct amdgpu_ip_block *ip_block;
 
 	if (amdgpu_sriov_vf(adev)) {
 		sdma_v5_2_ctx_switch_enable(adev, false);
@@ -868,13 +838,11 @@ static int sdma_v5_2_start(struct amdgpu_device *adev)
 			msleep(1000);
 	}
 
-	/* TODO: check whether can submit a doorbell request to raise
-	 * a doorbell fence to exit gfxoff.
-	 */
-	if (adev->in_s0ix)
-		amdgpu_gfx_off_ctrl(adev, false);
+	ip_block = amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_SDMA);
+	if (!ip_block)
+		return -EINVAL;
 
-	sdma_v5_2_soft_reset(adev);
+	sdma_v5_2_soft_reset(ip_block);
 	/* unhalt the MEs */
 	sdma_v5_2_enable(adev, true);
 	/* enable sdma ring preemption */
@@ -882,8 +850,6 @@ static int sdma_v5_2_start(struct amdgpu_device *adev)
 
 	/* start the gfx rings and rlc compute queues */
 	r = sdma_v5_2_gfx_resume(adev);
-	if (adev->in_s0ix)
-		amdgpu_gfx_off_ctrl(adev, true);
 	if (r)
 		return r;
 	r = sdma_v5_2_rlc_resume(adev);
@@ -891,6 +857,49 @@ static int sdma_v5_2_start(struct amdgpu_device *adev)
 	return r;
 }
 
+static int sdma_v5_2_mqd_init(struct amdgpu_device *adev, void *mqd,
+			      struct amdgpu_mqd_prop *prop)
+{
+	struct v10_sdma_mqd *m = mqd;
+	uint64_t wb_gpu_addr;
+
+	m->sdmax_rlcx_rb_cntl =
+		order_base_2(prop->queue_size / 4) << SDMA0_RLC0_RB_CNTL__RB_SIZE__SHIFT |
+		1 << SDMA0_RLC0_RB_CNTL__RPTR_WRITEBACK_ENABLE__SHIFT |
+		6 << SDMA0_RLC0_RB_CNTL__RPTR_WRITEBACK_TIMER__SHIFT |
+		1 << SDMA0_RLC0_RB_CNTL__RB_PRIV__SHIFT;
+
+	m->sdmax_rlcx_rb_base = lower_32_bits(prop->hqd_base_gpu_addr >> 8);
+	m->sdmax_rlcx_rb_base_hi = upper_32_bits(prop->hqd_base_gpu_addr >> 8);
+
+	m->sdmax_rlcx_rb_wptr_poll_cntl = RREG32(sdma_v5_2_get_reg_offset(adev, 0,
+						  mmSDMA0_GFX_RB_WPTR_POLL_CNTL));
+
+	wb_gpu_addr = prop->wptr_gpu_addr;
+	m->sdmax_rlcx_rb_wptr_poll_addr_lo = lower_32_bits(wb_gpu_addr);
+	m->sdmax_rlcx_rb_wptr_poll_addr_hi = upper_32_bits(wb_gpu_addr);
+
+	wb_gpu_addr = prop->rptr_gpu_addr;
+	m->sdmax_rlcx_rb_rptr_addr_lo = lower_32_bits(wb_gpu_addr);
+	m->sdmax_rlcx_rb_rptr_addr_hi = upper_32_bits(wb_gpu_addr);
+
+	m->sdmax_rlcx_ib_cntl = RREG32(sdma_v5_2_get_reg_offset(adev, 0,
+							mmSDMA0_GFX_IB_CNTL));
+
+	m->sdmax_rlcx_doorbell_offset =
+		prop->doorbell_index << SDMA0_RLC0_DOORBELL_OFFSET__OFFSET__SHIFT;
+
+	m->sdmax_rlcx_doorbell = REG_SET_FIELD(0, SDMA0_RLC0_DOORBELL, ENABLE, 1);
+
+	return 0;
+}
+
+static void sdma_v5_2_set_mqd_funcs(struct amdgpu_device *adev)
+{
+	adev->mqds[AMDGPU_HW_IP_DMA].mqd_size = sizeof(struct v10_sdma_mqd);
+	adev->mqds[AMDGPU_HW_IP_DMA].init_mqd = sdma_v5_2_mqd_init;
+}
+
 /**
  * sdma_v5_2_ring_test_ring - simple async dma engine test
  *
@@ -909,6 +918,8 @@ static int sdma_v5_2_ring_test_ring(struct amdgpu_ring *ring)
 	u32 tmp;
 	u64 gpu_addr;
 
+	tmp = 0xCAFEDEAD;
+
 	r = amdgpu_device_wb_get(adev, &index);
 	if (r) {
 		dev_err(adev->dev, "(%d) failed to allocate wb slot\n", r);
@@ -916,10 +927,9 @@ static int sdma_v5_2_ring_test_ring(struct amdgpu_ring *ring)
 	}
 
 	gpu_addr = adev->wb.gpu_addr + (index * 4);
-	tmp = 0xCAFEDEAD;
 	adev->wb.wb[index] = cpu_to_le32(tmp);
 
-	r = amdgpu_ring_alloc(ring, 5);
+	r = amdgpu_ring_alloc(ring, 20);
 	if (r) {
 		DRM_ERROR("amdgpu: dma failed to lock ring %d (%d).\n", ring->idx, r);
 		amdgpu_device_wb_free(adev, index);
@@ -971,6 +981,9 @@ static int sdma_v5_2_ring_test_ib(struct amdgpu_ring *ring, long timeout)
 	u32 tmp = 0;
 	u64 gpu_addr;
 
+	tmp = 0xCAFEDEAD;
+	memset(&ib, 0, sizeof(ib));
+
 	r = amdgpu_device_wb_get(adev, &index);
 	if (r) {
 		dev_err(adev->dev, "(%ld) failed to allocate wb slot\n", r);
@@ -978,9 +991,8 @@ static int sdma_v5_2_ring_test_ib(struct amdgpu_ring *ring, long timeout)
 	}
 
 	gpu_addr = adev->wb.gpu_addr + (index * 4);
-	tmp = 0xCAFEDEAD;
 	adev->wb.wb[index] = cpu_to_le32(tmp);
-	memset(&ib, 0, sizeof(ib));
+
 	r = amdgpu_ib_get(adev, NULL, 256, AMDGPU_IB_POOL_DIRECT, &ib);
 	if (r) {
 		DRM_ERROR("amdgpu: failed to get ib (%ld).\n", r);
@@ -1011,14 +1023,16 @@ static int sdma_v5_2_ring_test_ib(struct amdgpu_ring *ring, long timeout)
 		DRM_ERROR("amdgpu: fence wait failed (%ld).\n", r);
 		goto err1;
 	}
+
 	tmp = le32_to_cpu(adev->wb.wb[index]);
+
 	if (tmp == 0xDEADBEEF)
 		r = 0;
 	else
 		r = -EINVAL;
 
 err1:
-	amdgpu_ib_free(adev, &ib, NULL);
+	amdgpu_ib_free(&ib, NULL);
 	dma_fence_put(f);
 err0:
 	amdgpu_device_wb_free(adev, index);
@@ -1177,7 +1191,28 @@ static void sdma_v5_2_ring_emit_pipeline_sync(struct amdgpu_ring *ring)
 static void sdma_v5_2_ring_emit_vm_flush(struct amdgpu_ring *ring,
 					 unsigned vmid, uint64_t pd_addr)
 {
-	amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pd_addr);
+	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->vm_hub];
+	uint32_t req = hub->vmhub_funcs->get_invalidate_req(vmid, 0);
+
+	/* Update the PD address for this VMID. */
+	amdgpu_ring_emit_wreg(ring, hub->ctx0_ptb_addr_lo32 +
+			      (hub->ctx_addr_distance * vmid),
+			      lower_32_bits(pd_addr));
+	amdgpu_ring_emit_wreg(ring, hub->ctx0_ptb_addr_hi32 +
+			      (hub->ctx_addr_distance * vmid),
+			      upper_32_bits(pd_addr));
+
+	/* Trigger invalidation. */
+	amdgpu_ring_write(ring,
+			  SDMA_PKT_VM_INVALIDATION_HEADER_OP(SDMA_OP_POLL_REGMEM) |
+			  SDMA_PKT_VM_INVALIDATION_HEADER_SUB_OP(SDMA_SUBOP_VM_INVALIDATION) |
+			  SDMA_PKT_VM_INVALIDATION_HEADER_GFX_ENG_ID(ring->vm_inv_eng) |
+			  SDMA_PKT_VM_INVALIDATION_HEADER_MM_ENG_ID(0x1f));
+	amdgpu_ring_write(ring, req);
+	amdgpu_ring_write(ring, 0xFFFFFFFF);
+	amdgpu_ring_write(ring,
+			  SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_INVALIDATEACK(1 << vmid) |
+			  SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_ADDRESSRANGEHI(0x1F));
 }
 
 static void sdma_v5_2_ring_emit_wreg(struct amdgpu_ring *ring,
@@ -1213,14 +1248,20 @@ static void sdma_v5_2_ring_emit_reg_write_reg_wait(struct amdgpu_ring *ring,
 	amdgpu_ring_emit_reg_wait(ring, reg1, mask, mask);
 }
 
-static int sdma_v5_2_early_init(void *handle)
+static int sdma_v5_2_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
+	int r;
+
+	r = amdgpu_sdma_init_microcode(adev, 0, true);
+	if (r)
+		return r;
 
 	sdma_v5_2_set_ring_funcs(adev);
 	sdma_v5_2_set_buffer_funcs(adev);
 	sdma_v5_2_set_vm_pte_funcs(adev);
 	sdma_v5_2_set_irq_funcs(adev);
+	sdma_v5_2_set_mqd_funcs(adev);
 
 	return 0;
 }
@@ -1259,11 +1300,13 @@ static unsigned sdma_v5_2_seq_to_trap_id(int seq_num)
 	return -EINVAL;
 }
 
-static int sdma_v5_2_sw_init(void *handle)
+static int sdma_v5_2_sw_init(struct amdgpu_ip_block *ip_block)
 {
 	struct amdgpu_ring *ring;
 	int r, i;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
+	uint32_t reg_count = ARRAY_SIZE(sdma_reg_list_5_2);
+	uint32_t *ptr;
 
 	/* SDMA trap event */
 	for (i = 0; i < adev->sdma.num_instances; i++) {
@@ -1274,13 +1317,9 @@ static int sdma_v5_2_sw_init(void *handle)
 			return r;
 	}
 
-	r = sdma_v5_2_init_microcode(adev);
-	if (r) {
-		DRM_ERROR("Failed to load sdma firmware!\n");
-		return r;
-	}
-
 	for (i = 0; i < adev->sdma.num_instances; i++) {
+		mutex_init(&adev->sdma.instance[i].engine_reset_mutex);
+		adev->sdma.instance[i].funcs = &sdma_v5_2_sdma_funcs;
 		ring = &adev->sdma.instance[i].ring;
 		ring->ring_obj = NULL;
 		ring->use_doorbell = true;
@@ -1292,6 +1331,7 @@ static int sdma_v5_2_sw_init(void *handle)
 		ring->doorbell_index =
 			(adev->doorbell_index.sdma_engine[i] << 1); //get DWORD offset
 
+		ring->vm_hub = AMDGPU_GFXHUB(0);
 		sprintf(ring->name, "sdma%d", i);
 		r = amdgpu_ring_init(adev, ring, 1024, &adev->sdma.trap_irq,
 				     AMDGPU_SDMA_IRQ_INSTANCE0 + i,
@@ -1300,35 +1340,68 @@ static int sdma_v5_2_sw_init(void *handle)
 			return r;
 	}
 
+	adev->sdma.supported_reset =
+		amdgpu_get_soft_full_reset_mask(&adev->sdma.instance[0].ring);
+	switch (amdgpu_ip_version(adev, SDMA0_HWIP, 0)) {
+	case IP_VERSION(5, 2, 0):
+	case IP_VERSION(5, 2, 2):
+	case IP_VERSION(5, 2, 3):
+	case IP_VERSION(5, 2, 4):
+		if ((adev->sdma.instance[0].fw_version >= 76) &&
+		    !amdgpu_sriov_vf(adev) &&
+		    !adev->debug_disable_gpu_ring_reset)
+			adev->sdma.supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE;
+		break;
+	case IP_VERSION(5, 2, 5):
+		if ((adev->sdma.instance[0].fw_version >= 34) &&
+		    !amdgpu_sriov_vf(adev) &&
+		    !adev->debug_disable_gpu_ring_reset)
+			adev->sdma.supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE;
+		break;
+	default:
+		break;
+	}
+
+	/* Allocate memory for SDMA IP Dump buffer */
+	ptr = kcalloc(adev->sdma.num_instances * reg_count, sizeof(uint32_t), GFP_KERNEL);
+	if (ptr)
+		adev->sdma.ip_dump = ptr;
+	else
+		DRM_ERROR("Failed to allocated memory for SDMA IP Dump\n");
+
+	r = amdgpu_sdma_sysfs_reset_mask_init(adev);
+	if (r)
+		return r;
+
 	return r;
 }
 
-static int sdma_v5_2_sw_fini(void *handle)
+static int sdma_v5_2_sw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int i;
 
 	for (i = 0; i < adev->sdma.num_instances; i++)
 		amdgpu_ring_fini(&adev->sdma.instance[i].ring);
 
-	sdma_v5_2_destroy_inst_ctx(adev);
+	amdgpu_sdma_sysfs_reset_mask_fini(adev);
+	amdgpu_sdma_destroy_inst_ctx(adev, true);
+
+	kfree(adev->sdma.ip_dump);
 
 	return 0;
 }
 
-static int sdma_v5_2_hw_init(void *handle)
+static int sdma_v5_2_hw_init(struct amdgpu_ip_block *ip_block)
 {
-	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
-	r = sdma_v5_2_start(adev);
-
-	return r;
+	return sdma_v5_2_start(adev);
 }
 
-static int sdma_v5_2_hw_fini(void *handle)
+static int sdma_v5_2_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (amdgpu_sriov_vf(adev))
 		return 0;
@@ -1339,23 +1412,19 @@ static int sdma_v5_2_hw_fini(void *handle)
 	return 0;
 }
 
-static int sdma_v5_2_suspend(void *handle)
+static int sdma_v5_2_suspend(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	return sdma_v5_2_hw_fini(adev);
+	return sdma_v5_2_hw_fini(ip_block);
 }
 
-static int sdma_v5_2_resume(void *handle)
+static int sdma_v5_2_resume(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	return sdma_v5_2_hw_init(adev);
+	return sdma_v5_2_hw_init(ip_block);
 }
 
-static bool sdma_v5_2_is_idle(void *handle)
+static bool sdma_v5_2_is_idle(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	u32 i;
 
 	for (i = 0; i < adev->sdma.num_instances; i++) {
@@ -1368,11 +1437,11 @@ static bool sdma_v5_2_is_idle(void *handle)
 	return true;
 }
 
-static int sdma_v5_2_wait_for_idle(void *handle)
+static int sdma_v5_2_wait_for_idle(struct amdgpu_ip_block *ip_block)
 {
 	unsigned i;
 	u32 sdma0, sdma1, sdma2, sdma3;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	for (i = 0; i < adev->usec_timeout; i++) {
 		sdma0 = RREG32(sdma_v5_2_get_reg_offset(adev, 0, mmSDMA0_STATUS_REG));
@@ -1387,6 +1456,100 @@ static int sdma_v5_2_wait_for_idle(void *handle)
 	return -ETIMEDOUT;
 }
 
+static int sdma_v5_2_reset_queue(struct amdgpu_ring *ring,
+				 unsigned int vmid,
+				 struct amdgpu_fence *timedout_fence)
+{
+	struct amdgpu_device *adev = ring->adev;
+	int r;
+
+	if (ring->me >= adev->sdma.num_instances) {
+		dev_err(adev->dev, "sdma instance not found\n");
+		return -EINVAL;
+	}
+
+	amdgpu_ring_reset_helper_begin(ring, timedout_fence);
+
+	amdgpu_amdkfd_suspend(adev, true);
+	r = amdgpu_sdma_reset_engine(adev, ring->me, true);
+	amdgpu_amdkfd_resume(adev, true);
+	if (r)
+		return r;
+
+	return amdgpu_ring_reset_helper_end(ring, timedout_fence);
+}
+
+static int sdma_v5_2_stop_queue(struct amdgpu_ring *ring)
+{
+	u32 f32_cntl, freeze, cntl, stat1_reg;
+	struct amdgpu_device *adev = ring->adev;
+	int i, j, r = 0;
+
+	if (amdgpu_sriov_vf(adev))
+		return -EINVAL;
+
+	i = ring->me;
+	amdgpu_gfx_rlc_enter_safe_mode(adev, 0);
+
+	/* stop queue */
+	sdma_v5_2_gfx_stop(adev, 1 << i);
+
+	/*engine stop SDMA1_F32_CNTL.HALT to 1 and SDMAx_FREEZE freeze bit to 1 */
+	freeze = RREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_FREEZE));
+	freeze = REG_SET_FIELD(freeze, SDMA0_FREEZE, FREEZE, 1);
+	WREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_FREEZE), freeze);
+
+	for (j = 0; j < adev->usec_timeout; j++) {
+		freeze = RREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_FREEZE));
+
+		if (REG_GET_FIELD(freeze, SDMA0_FREEZE, FROZEN) & 1)
+			break;
+		udelay(1);
+	}
+
+
+	if (j == adev->usec_timeout) {
+		stat1_reg = RREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_STATUS1_REG));
+		if ((stat1_reg & 0x3FF) != 0x3FF) {
+			DRM_ERROR("cannot soft reset as sdma not idle\n");
+			r = -ETIMEDOUT;
+			goto err0;
+		}
+	}
+
+	f32_cntl = RREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_F32_CNTL));
+	f32_cntl = REG_SET_FIELD(f32_cntl, SDMA0_F32_CNTL, HALT, 1);
+	WREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_F32_CNTL), f32_cntl);
+
+	cntl = RREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_CNTL));
+	cntl = REG_SET_FIELD(cntl, SDMA0_CNTL, UTC_L1_ENABLE, 0);
+	WREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_CNTL), cntl);
+
+err0:
+	amdgpu_gfx_rlc_exit_safe_mode(adev, 0);
+	return r;
+}
+
+static int sdma_v5_2_restore_queue(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	u32 inst_id = ring->me;
+	u32 freeze;
+	int r;
+
+	amdgpu_gfx_rlc_enter_safe_mode(adev, 0);
+	/* unfreeze and unhalt */
+	freeze = RREG32(sdma_v5_2_get_reg_offset(adev, inst_id, mmSDMA0_FREEZE));
+	freeze = REG_SET_FIELD(freeze, SDMA0_FREEZE, FREEZE, 0);
+	WREG32(sdma_v5_2_get_reg_offset(adev, inst_id, mmSDMA0_FREEZE), freeze);
+
+	r = sdma_v5_2_gfx_resume_instance(adev, inst_id, true);
+
+	amdgpu_gfx_rlc_exit_safe_mode(adev, 0);
+
+	return r;
+}
+
 static int sdma_v5_2_ring_preempt_ib(struct amdgpu_ring *ring)
 {
 	int i, r = 0;
@@ -1438,13 +1601,14 @@ static int sdma_v5_2_set_trap_irq_state(struct amdgpu_device *adev,
 					enum amdgpu_interrupt_state state)
 {
 	u32 sdma_cntl;
-
 	u32 reg_offset = sdma_v5_2_get_reg_offset(adev, type, mmSDMA0_CNTL);
 
-	sdma_cntl = RREG32(reg_offset);
-	sdma_cntl = REG_SET_FIELD(sdma_cntl, SDMA0_CNTL, TRAP_ENABLE,
-		       state == AMDGPU_IRQ_STATE_ENABLE ? 1 : 0);
-	WREG32(reg_offset, sdma_cntl);
+	if (!amdgpu_sriov_vf(adev)) {
+		sdma_cntl = RREG32(reg_offset);
+		sdma_cntl = REG_SET_FIELD(sdma_cntl, SDMA0_CNTL, TRAP_ENABLE,
+			       state == AMDGPU_IRQ_STATE_ENABLE ? 1 : 0);
+		WREG32(reg_offset, sdma_cntl);
+	}
 
 	return 0;
 }
@@ -1453,7 +1617,25 @@ static int sdma_v5_2_process_trap_irq(struct amdgpu_device *adev,
 				      struct amdgpu_irq_src *source,
 				      struct amdgpu_iv_entry *entry)
 {
+	uint32_t mes_queue_id = entry->src_data[0];
+
 	DRM_DEBUG("IH: SDMA trap\n");
+
+	if (adev->enable_mes && (mes_queue_id & AMDGPU_FENCE_MES_QUEUE_FLAG)) {
+		struct amdgpu_mes_queue *queue;
+
+		mes_queue_id &= AMDGPU_FENCE_MES_QUEUE_ID_MASK;
+
+		spin_lock(&adev->mes.queue_id_lock);
+		queue = idr_find(&adev->mes.queue_id_idr, mes_queue_id);
+		if (queue) {
+			DRM_DEBUG("process smda queue id = %d\n", mes_queue_id);
+			amdgpu_fence_process(queue->ring);
+		}
+		spin_unlock(&adev->mes.queue_id_lock);
+		return 0;
+	}
+
 	switch (entry->client_id) {
 	case SOC15_IH_CLIENTID_SDMA0:
 		switch (entry->ring_id) {
@@ -1530,6 +1712,30 @@ static int sdma_v5_2_process_illegal_inst_irq(struct amdgpu_device *adev,
 	return 0;
 }
 
+static bool sdma_v5_2_firmware_mgcg_support(struct amdgpu_device *adev,
+						     int i)
+{
+	switch (amdgpu_ip_version(adev, SDMA0_HWIP, 0)) {
+	case IP_VERSION(5, 2, 1):
+		if (adev->sdma.instance[i].fw_version < 70)
+			return false;
+		break;
+	case IP_VERSION(5, 2, 3):
+		if (adev->sdma.instance[i].fw_version < 47)
+			return false;
+		break;
+	case IP_VERSION(5, 2, 7):
+		if (adev->sdma.instance[i].fw_version < 9)
+			return false;
+		break;
+	default:
+		return true;
+	}
+
+	return true;
+
+}
+
 static void sdma_v5_2_update_medium_grain_clock_gating(struct amdgpu_device *adev,
 						       bool enable)
 {
@@ -1538,7 +1744,7 @@ static void sdma_v5_2_update_medium_grain_clock_gating(struct amdgpu_device *ade
 
 	for (i = 0; i < adev->sdma.num_instances; i++) {
 
-		if (adev->sdma.instance[i].fw_version < 70 && adev->ip_versions[SDMA0_HWIP][0] == IP_VERSION(5, 2, 1))
+		if (!sdma_v5_2_firmware_mgcg_support(adev, i))
 			adev->cg_flags &= ~AMD_CG_SUPPORT_SDMA_MGCG;
 
 		if (enable && (adev->cg_flags & AMD_CG_SUPPORT_SDMA_MGCG)) {
@@ -1574,8 +1780,9 @@ static void sdma_v5_2_update_medium_grain_light_sleep(struct amdgpu_device *adev
 	int i;
 
 	for (i = 0; i < adev->sdma.num_instances; i++) {
-
-		if (adev->sdma.instance[i].fw_version < 70 && adev->ip_versions[SDMA0_HWIP][0] == IP_VERSION(5, 2, 1))
+		if (adev->sdma.instance[i].fw_version < 70 &&
+		    amdgpu_ip_version(adev, SDMA0_HWIP, 0) ==
+			    IP_VERSION(5, 2, 1))
 			adev->cg_flags &= ~AMD_CG_SUPPORT_SDMA_LS;
 
 		if (enable && (adev->cg_flags & AMD_CG_SUPPORT_SDMA_LS)) {
@@ -1596,21 +1803,23 @@ static void sdma_v5_2_update_medium_grain_light_sleep(struct amdgpu_device *adev
 	}
 }
 
-static int sdma_v5_2_set_clockgating_state(void *handle,
+static int sdma_v5_2_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					   enum amd_clockgating_state state)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (amdgpu_sriov_vf(adev))
 		return 0;
 
-	switch (adev->ip_versions[SDMA0_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, SDMA0_HWIP, 0)) {
 	case IP_VERSION(5, 2, 0):
 	case IP_VERSION(5, 2, 2):
 	case IP_VERSION(5, 2, 1):
 	case IP_VERSION(5, 2, 4):
 	case IP_VERSION(5, 2, 5):
+	case IP_VERSION(5, 2, 6):
 	case IP_VERSION(5, 2, 3):
+	case IP_VERSION(5, 2, 7):
 		sdma_v5_2_update_medium_grain_clock_gating(adev,
 				state == AMD_CG_STATE_GATE);
 		sdma_v5_2_update_medium_grain_light_sleep(adev,
@@ -1623,30 +1832,106 @@ static int sdma_v5_2_set_clockgating_state(void *handle,
 	return 0;
 }
 
-static int sdma_v5_2_set_powergating_state(void *handle,
+static int sdma_v5_2_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_powergating_state state)
 {
 	return 0;
 }
 
-static void sdma_v5_2_get_clockgating_state(void *handle, u32 *flags)
+static void sdma_v5_2_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int data;
 
 	if (amdgpu_sriov_vf(adev))
 		*flags = 0;
 
+	/* AMD_CG_SUPPORT_SDMA_MGCG */
+	data = RREG32(sdma_v5_2_get_reg_offset(adev, 0, mmSDMA0_CLK_CTRL));
+	if (!(data & SDMA0_CLK_CTRL__CGCG_EN_OVERRIDE_MASK))
+		*flags |= AMD_CG_SUPPORT_SDMA_MGCG;
+
 	/* AMD_CG_SUPPORT_SDMA_LS */
 	data = RREG32_KIQ(sdma_v5_2_get_reg_offset(adev, 0, mmSDMA0_POWER_CNTL));
 	if (data & SDMA0_POWER_CNTL__MEM_POWER_OVERRIDE_MASK)
 		*flags |= AMD_CG_SUPPORT_SDMA_LS;
 }
 
-const struct amd_ip_funcs sdma_v5_2_ip_funcs = {
+static void sdma_v5_2_ring_begin_use(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	/* SDMA 5.2.3 (RMB) FW doesn't seem to properly
+	 * disallow GFXOFF in some cases leading to
+	 * hangs in SDMA.  Disallow GFXOFF while SDMA is active.
+	 * We can probably just limit this to 5.2.3,
+	 * but it shouldn't hurt for other parts since
+	 * this GFXOFF will be disallowed anyway when SDMA is
+	 * active, this just makes it explicit.
+	 * sdma_v5_2_ring_set_wptr() takes advantage of this
+	 * to update the wptr because sometimes SDMA seems to miss
+	 * doorbells when entering PG.  If you remove this, update
+	 * sdma_v5_2_ring_set_wptr() as well!
+	 */
+	amdgpu_gfx_off_ctrl(adev, false);
+}
+
+static void sdma_v5_2_ring_end_use(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	/* SDMA 5.2.3 (RMB) FW doesn't seem to properly
+	 * disallow GFXOFF in some cases leading to
+	 * hangs in SDMA.  Allow GFXOFF when SDMA is complete.
+	 */
+	amdgpu_gfx_off_ctrl(adev, true);
+}
+
+static void sdma_v5_2_print_ip_state(struct amdgpu_ip_block *ip_block, struct drm_printer *p)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, j;
+	uint32_t reg_count = ARRAY_SIZE(sdma_reg_list_5_2);
+	uint32_t instance_offset;
+
+	if (!adev->sdma.ip_dump)
+		return;
+
+	drm_printf(p, "num_instances:%d\n", adev->sdma.num_instances);
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		instance_offset = i * reg_count;
+		drm_printf(p, "\nInstance:%d\n", i);
+
+		for (j = 0; j < reg_count; j++)
+			drm_printf(p, "%-50s \t 0x%08x\n", sdma_reg_list_5_2[j].reg_name,
+				   adev->sdma.ip_dump[instance_offset + j]);
+	}
+}
+
+static void sdma_v5_2_dump_ip_state(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, j;
+	uint32_t instance_offset;
+	uint32_t reg_count = ARRAY_SIZE(sdma_reg_list_5_2);
+
+	if (!adev->sdma.ip_dump)
+		return;
+
+	amdgpu_gfx_off_ctrl(adev, false);
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		instance_offset = i * reg_count;
+		for (j = 0; j < reg_count; j++)
+			adev->sdma.ip_dump[instance_offset + j] =
+				RREG32(sdma_v5_2_get_reg_offset(adev, i,
+				       sdma_reg_list_5_2[j].reg_offset));
+	}
+	amdgpu_gfx_off_ctrl(adev, true);
+}
+
+static const struct amd_ip_funcs sdma_v5_2_ip_funcs = {
 	.name = "sdma_v5_2",
 	.early_init = sdma_v5_2_early_init,
-	.late_init = NULL,
 	.sw_init = sdma_v5_2_sw_init,
 	.sw_fini = sdma_v5_2_sw_fini,
 	.hw_init = sdma_v5_2_hw_init,
@@ -1659,6 +1944,8 @@ const struct amd_ip_funcs sdma_v5_2_ip_funcs = {
 	.set_clockgating_state = sdma_v5_2_set_clockgating_state,
 	.set_powergating_state = sdma_v5_2_set_powergating_state,
 	.get_clockgating_state = sdma_v5_2_get_clockgating_state,
+	.dump_ip_state = sdma_v5_2_dump_ip_state,
+	.print_ip_state = sdma_v5_2_print_ip_state,
 };
 
 static const struct amdgpu_ring_funcs sdma_v5_2_ring_funcs = {
@@ -1666,7 +1953,7 @@ static const struct amdgpu_ring_funcs sdma_v5_2_ring_funcs = {
 	.align_mask = 0xf,
 	.nop = SDMA_PKT_NOP_HEADER_OP(SDMA_OP_NOP),
 	.support_64bit_ptrs = true,
-	.vmhub = AMDGPU_GFXHUB_0,
+	.secure_submission_supported = true,
 	.get_rptr = sdma_v5_2_ring_get_rptr,
 	.get_wptr = sdma_v5_2_ring_get_wptr,
 	.set_wptr = sdma_v5_2_ring_set_wptr,
@@ -1690,12 +1977,14 @@ static const struct amdgpu_ring_funcs sdma_v5_2_ring_funcs = {
 	.test_ib = sdma_v5_2_ring_test_ib,
 	.insert_nop = sdma_v5_2_ring_insert_nop,
 	.pad_ib = sdma_v5_2_ring_pad_ib,
+	.begin_use = sdma_v5_2_ring_begin_use,
+	.end_use = sdma_v5_2_ring_end_use,
 	.emit_wreg = sdma_v5_2_ring_emit_wreg,
 	.emit_reg_wait = sdma_v5_2_ring_emit_reg_wait,
 	.emit_reg_write_reg_wait = sdma_v5_2_ring_emit_reg_write_reg_wait,
 	.init_cond_exec = sdma_v5_2_ring_init_cond_exec,
-	.patch_cond_exec = sdma_v5_2_ring_patch_cond_exec,
 	.preempt_ib = sdma_v5_2_ring_preempt_ib,
+	.reset = sdma_v5_2_reset_queue,
 };
 
 static void sdma_v5_2_set_ring_funcs(struct amdgpu_device *adev)
@@ -1732,7 +2021,7 @@ static void sdma_v5_2_set_irq_funcs(struct amdgpu_device *adev)
  * @src_offset: src GPU address
  * @dst_offset: dst GPU address
  * @byte_count: number of bytes to xfer
- * @tmz: if a secure copy should be used
+ * @copy_flags: copy flags for the buffers
  *
  * Copy GPU buffers using the DMA engine.
  * Used by the amdgpu ttm implementation to move pages if
@@ -1742,11 +2031,11 @@ static void sdma_v5_2_emit_copy_buffer(struct amdgpu_ib *ib,
 				       uint64_t src_offset,
 				       uint64_t dst_offset,
 				       uint32_t byte_count,
-				       bool tmz)
+				       uint32_t copy_flags)
 {
 	ib->ptr[ib->length_dw++] = SDMA_PKT_HEADER_OP(SDMA_OP_COPY) |
 		SDMA_PKT_HEADER_SUB_OP(SDMA_SUBOP_COPY_LINEAR) |
-		SDMA_PKT_COPY_LINEAR_HEADER_TMZ(tmz ? 1 : 0);
+		SDMA_PKT_COPY_LINEAR_HEADER_TMZ((copy_flags & AMDGPU_COPY_FLAGS_TMZ) ? 1 : 0);
 	ib->ptr[ib->length_dw++] = byte_count - 1;
 	ib->ptr[ib->length_dw++] = 0; /* src/dst endian swap */
 	ib->ptr[ib->length_dw++] = lower_32_bits(src_offset);
diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.h b/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.h
index b70414fef2a1..863145b3a77e 100644
--- a/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.h
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.h
@@ -24,7 +24,6 @@
 #ifndef __SDMA_V5_2_H__
 #define __SDMA_V5_2_H__
 
-extern const struct amd_ip_funcs sdma_v5_2_ip_funcs;
 extern const struct amdgpu_ip_block_version sdma_v5_2_ip_block;
 
 #endif /* __SDMA_V5_2_H__ */
diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v6_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v6_0.c
new file mode 100644
index 000000000000..217040044987
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v6_0.c
@@ -0,0 +1,1917 @@
+/*
+ * Copyright 2020 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/delay.h>
+#include <linux/firmware.h>
+#include <linux/module.h>
+#include <linux/pci.h>
+
+#include "amdgpu.h"
+#include "amdgpu_ucode.h"
+#include "amdgpu_trace.h"
+
+#include "gc/gc_11_0_0_offset.h"
+#include "gc/gc_11_0_0_sh_mask.h"
+#include "gc/gc_11_0_0_default.h"
+#include "hdp/hdp_6_0_0_offset.h"
+#include "ivsrcid/gfx/irqsrcs_gfx_11_0_0.h"
+
+#include "soc15_common.h"
+#include "soc15.h"
+#include "sdma_v6_0_0_pkt_open.h"
+#include "nbio_v4_3.h"
+#include "sdma_common.h"
+#include "sdma_v6_0.h"
+#include "v11_structs.h"
+#include "mes_userqueue.h"
+#include "amdgpu_userq_fence.h"
+
+MODULE_FIRMWARE("amdgpu/sdma_6_0_0.bin");
+MODULE_FIRMWARE("amdgpu/sdma_6_0_1.bin");
+MODULE_FIRMWARE("amdgpu/sdma_6_0_2.bin");
+MODULE_FIRMWARE("amdgpu/sdma_6_0_3.bin");
+MODULE_FIRMWARE("amdgpu/sdma_6_1_0.bin");
+MODULE_FIRMWARE("amdgpu/sdma_6_1_1.bin");
+MODULE_FIRMWARE("amdgpu/sdma_6_1_2.bin");
+MODULE_FIRMWARE("amdgpu/sdma_6_1_3.bin");
+
+#define SDMA1_REG_OFFSET 0x600
+#define SDMA0_HYP_DEC_REG_START 0x5880
+#define SDMA0_HYP_DEC_REG_END 0x589a
+#define SDMA1_HYP_DEC_REG_OFFSET 0x20
+
+static const struct amdgpu_hwip_reg_entry sdma_reg_list_6_0[] = {
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_STATUS_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_STATUS1_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_STATUS2_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_STATUS3_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_STATUS4_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_STATUS5_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_STATUS6_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_UCODE_CHECKSUM),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_RB_RPTR_FETCH_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_RB_RPTR_FETCH),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_UTCL1_RD_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_UTCL1_WR_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_UTCL1_RD_XNACK0),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_UTCL1_RD_XNACK1),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_UTCL1_WR_XNACK0),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_UTCL1_WR_XNACK1),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE0_RB_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE0_RB_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE0_RB_RPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE0_RB_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE0_RB_WPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE0_IB_OFFSET),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE0_IB_BASE_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE0_IB_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE0_IB_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE0_IB_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE0_IB_SUB_REMAIN),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE0_DUMMY_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE_STATUS0),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE1_RB_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE1_RB_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE1_RB_RPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE1_RB_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE1_RB_WPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE1_IB_OFFSET),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE1_IB_BASE_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE1_IB_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE1_IB_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE1_IB_SUB_REMAIN),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE1_DUMMY_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE2_RB_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE2_RB_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE2_RB_RPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE2_RB_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE2_RB_WPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE2_IB_OFFSET),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE2_IB_BASE_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE2_IB_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE2_IB_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE2_IB_SUB_REMAIN),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE2_DUMMY_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_INT_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regGRBM_STATUS2),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_CHICKEN_BITS),
+};
+
+static void sdma_v6_0_set_ring_funcs(struct amdgpu_device *adev);
+static void sdma_v6_0_set_buffer_funcs(struct amdgpu_device *adev);
+static void sdma_v6_0_set_vm_pte_funcs(struct amdgpu_device *adev);
+static void sdma_v6_0_set_irq_funcs(struct amdgpu_device *adev);
+static int sdma_v6_0_start(struct amdgpu_device *adev);
+
+static u32 sdma_v6_0_get_reg_offset(struct amdgpu_device *adev, u32 instance, u32 internal_offset)
+{
+	u32 base;
+
+	if (internal_offset >= SDMA0_HYP_DEC_REG_START &&
+	    internal_offset <= SDMA0_HYP_DEC_REG_END) {
+		base = adev->reg_offset[GC_HWIP][0][1];
+		if (instance != 0)
+			internal_offset += SDMA1_HYP_DEC_REG_OFFSET * instance;
+	} else {
+		base = adev->reg_offset[GC_HWIP][0][0];
+		if (instance == 1)
+			internal_offset += SDMA1_REG_OFFSET;
+	}
+
+	return base + internal_offset;
+}
+
+static unsigned sdma_v6_0_ring_init_cond_exec(struct amdgpu_ring *ring,
+					      uint64_t addr)
+{
+	unsigned ret;
+
+	amdgpu_ring_write(ring, SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_COND_EXE));
+	amdgpu_ring_write(ring, lower_32_bits(addr));
+	amdgpu_ring_write(ring, upper_32_bits(addr));
+	amdgpu_ring_write(ring, 1);
+	/* this is the offset we need patch later */
+	ret = ring->wptr & ring->buf_mask;
+	/* insert dummy here and patch it later */
+	amdgpu_ring_write(ring, 0);
+
+	return ret;
+}
+
+/**
+ * sdma_v6_0_ring_get_rptr - get the current read pointer
+ *
+ * @ring: amdgpu ring pointer
+ *
+ * Get the current rptr from the hardware.
+ */
+static uint64_t sdma_v6_0_ring_get_rptr(struct amdgpu_ring *ring)
+{
+	u64 *rptr;
+
+	/* XXX check if swapping is necessary on BE */
+	rptr = (u64 *)ring->rptr_cpu_addr;
+
+	DRM_DEBUG("rptr before shift == 0x%016llx\n", *rptr);
+	return ((*rptr) >> 2);
+}
+
+/**
+ * sdma_v6_0_ring_get_wptr - get the current write pointer
+ *
+ * @ring: amdgpu ring pointer
+ *
+ * Get the current wptr from the hardware.
+ */
+static uint64_t sdma_v6_0_ring_get_wptr(struct amdgpu_ring *ring)
+{
+	u64 wptr = 0;
+
+	if (ring->use_doorbell) {
+		/* XXX check if swapping is necessary on BE */
+		wptr = READ_ONCE(*((u64 *)ring->wptr_cpu_addr));
+		DRM_DEBUG("wptr/doorbell before shift == 0x%016llx\n", wptr);
+	}
+
+	return wptr >> 2;
+}
+
+/**
+ * sdma_v6_0_ring_set_wptr - commit the write pointer
+ *
+ * @ring: amdgpu ring pointer
+ *
+ * Write the wptr back to the hardware.
+ */
+static void sdma_v6_0_ring_set_wptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	if (ring->use_doorbell) {
+		DRM_DEBUG("Using doorbell -- "
+			  "wptr_offs == 0x%08x "
+			  "lower_32_bits(ring->wptr) << 2 == 0x%08x "
+			  "upper_32_bits(ring->wptr) << 2 == 0x%08x\n",
+			  ring->wptr_offs,
+			  lower_32_bits(ring->wptr << 2),
+			  upper_32_bits(ring->wptr << 2));
+		/* XXX check if swapping is necessary on BE */
+		atomic64_set((atomic64_t *)ring->wptr_cpu_addr,
+			     ring->wptr << 2);
+		DRM_DEBUG("calling WDOORBELL64(0x%08x, 0x%016llx)\n",
+			  ring->doorbell_index, ring->wptr << 2);
+		WDOORBELL64(ring->doorbell_index, ring->wptr << 2);
+	} else {
+		DRM_DEBUG("Not using doorbell -- "
+			  "regSDMA%i_GFX_RB_WPTR == 0x%08x "
+			  "regSDMA%i_GFX_RB_WPTR_HI == 0x%08x\n",
+			  ring->me,
+			  lower_32_bits(ring->wptr << 2),
+			  ring->me,
+			  upper_32_bits(ring->wptr << 2));
+		WREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev,
+							     ring->me, regSDMA0_QUEUE0_RB_WPTR),
+				lower_32_bits(ring->wptr << 2));
+		WREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev,
+							     ring->me, regSDMA0_QUEUE0_RB_WPTR_HI),
+				upper_32_bits(ring->wptr << 2));
+	}
+}
+
+static void sdma_v6_0_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count)
+{
+	struct amdgpu_sdma_instance *sdma = amdgpu_sdma_get_instance_from_ring(ring);
+	int i;
+
+	for (i = 0; i < count; i++)
+		if (sdma && sdma->burst_nop && (i == 0))
+			amdgpu_ring_write(ring, ring->funcs->nop |
+				SDMA_PKT_NOP_HEADER_COUNT(count - 1));
+		else
+			amdgpu_ring_write(ring, ring->funcs->nop);
+}
+
+/*
+ * sdma_v6_0_ring_emit_ib - Schedule an IB on the DMA engine
+ *
+ * @ring: amdgpu ring pointer
+ * @ib: IB object to schedule
+ * @flags: unused
+ * @job: job to retrieve vmid from
+ *
+ * Schedule an IB in the DMA ring.
+ */
+static void sdma_v6_0_ring_emit_ib(struct amdgpu_ring *ring,
+				   struct amdgpu_job *job,
+				   struct amdgpu_ib *ib,
+				   uint32_t flags)
+{
+	unsigned vmid = AMDGPU_JOB_GET_VMID(job);
+	uint64_t csa_mc_addr = amdgpu_sdma_get_csa_mc_addr(ring, vmid);
+
+	/* An IB packet must end on a 8 DW boundary--the next dword
+	 * must be on a 8-dword boundary. Our IB packet below is 6
+	 * dwords long, thus add x number of NOPs, such that, in
+	 * modular arithmetic,
+	 * wptr + 6 + x = 8k, k >= 0, which in C is,
+	 * (wptr + 6 + x) % 8 = 0.
+	 * The expression below, is a solution of x.
+	 */
+	sdma_v6_0_ring_insert_nop(ring, (2 - lower_32_bits(ring->wptr)) & 7);
+
+	amdgpu_ring_write(ring, SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_INDIRECT) |
+			  SDMA_PKT_INDIRECT_HEADER_VMID(vmid & 0xf));
+	/* base must be 32 byte aligned */
+	amdgpu_ring_write(ring, lower_32_bits(ib->gpu_addr) & 0xffffffe0);
+	amdgpu_ring_write(ring, upper_32_bits(ib->gpu_addr));
+	amdgpu_ring_write(ring, ib->length_dw);
+	amdgpu_ring_write(ring, lower_32_bits(csa_mc_addr));
+	amdgpu_ring_write(ring, upper_32_bits(csa_mc_addr));
+}
+
+/**
+ * sdma_v6_0_ring_emit_mem_sync - flush the IB by graphics cache rinse
+ *
+ * @ring: amdgpu ring pointer
+ *
+ * flush the IB by graphics cache rinse.
+ */
+static void sdma_v6_0_ring_emit_mem_sync(struct amdgpu_ring *ring)
+{
+        uint32_t gcr_cntl = SDMA_GCR_GL2_INV | SDMA_GCR_GL2_WB | SDMA_GCR_GLM_INV |
+                            SDMA_GCR_GL1_INV | SDMA_GCR_GLV_INV | SDMA_GCR_GLK_INV |
+                            SDMA_GCR_GLI_INV(1);
+
+        /* flush entire cache L0/L1/L2, this can be optimized by performance requirement */
+        amdgpu_ring_write(ring, SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_GCR_REQ));
+        amdgpu_ring_write(ring, SDMA_PKT_GCR_REQ_PAYLOAD1_BASE_VA_31_7(0));
+        amdgpu_ring_write(ring, SDMA_PKT_GCR_REQ_PAYLOAD2_GCR_CONTROL_15_0(gcr_cntl) |
+                          SDMA_PKT_GCR_REQ_PAYLOAD2_BASE_VA_47_32(0));
+        amdgpu_ring_write(ring, SDMA_PKT_GCR_REQ_PAYLOAD3_LIMIT_VA_31_7(0) |
+                          SDMA_PKT_GCR_REQ_PAYLOAD3_GCR_CONTROL_18_16(gcr_cntl >> 16));
+        amdgpu_ring_write(ring, SDMA_PKT_GCR_REQ_PAYLOAD4_LIMIT_VA_47_32(0) |
+                          SDMA_PKT_GCR_REQ_PAYLOAD4_VMID(0));
+}
+
+
+/**
+ * sdma_v6_0_ring_emit_hdp_flush - emit an hdp flush on the DMA ring
+ *
+ * @ring: amdgpu ring pointer
+ *
+ * Emit an hdp flush packet on the requested DMA ring.
+ */
+static void sdma_v6_0_ring_emit_hdp_flush(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	u32 ref_and_mask = 0;
+	const struct nbio_hdp_flush_reg *nbio_hf_reg = adev->nbio.hdp_flush_reg;
+
+	ref_and_mask = nbio_hf_reg->ref_and_mask_sdma0 << ring->me;
+
+	amdgpu_ring_write(ring, SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_POLL_REGMEM) |
+			  SDMA_PKT_POLL_REGMEM_HEADER_HDP_FLUSH(1) |
+			  SDMA_PKT_POLL_REGMEM_HEADER_FUNC(3)); /* == */
+	amdgpu_ring_write(ring, (adev->nbio.funcs->get_hdp_flush_done_offset(adev)) << 2);
+	amdgpu_ring_write(ring, (adev->nbio.funcs->get_hdp_flush_req_offset(adev)) << 2);
+	amdgpu_ring_write(ring, ref_and_mask); /* reference */
+	amdgpu_ring_write(ring, ref_and_mask); /* mask */
+	amdgpu_ring_write(ring, SDMA_PKT_POLL_REGMEM_DW5_RETRY_COUNT(0xfff) |
+			  SDMA_PKT_POLL_REGMEM_DW5_INTERVAL(10)); /* retry count, poll interval */
+}
+
+/**
+ * sdma_v6_0_ring_emit_fence - emit a fence on the DMA ring
+ *
+ * @ring: amdgpu ring pointer
+ * @addr: address
+ * @seq: fence seq number
+ * @flags: fence flags
+ *
+ * Add a DMA fence packet to the ring to write
+ * the fence seq number and DMA trap packet to generate
+ * an interrupt if needed.
+ */
+static void sdma_v6_0_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 seq,
+				      unsigned flags)
+{
+	bool write64bit = flags & AMDGPU_FENCE_FLAG_64BIT;
+	/* write the fence */
+	amdgpu_ring_write(ring, SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_FENCE) |
+			  SDMA_PKT_FENCE_HEADER_MTYPE(0x3)); /* Ucached(UC) */
+	/* zero in first two bits */
+	BUG_ON(addr & 0x3);
+	amdgpu_ring_write(ring, lower_32_bits(addr));
+	amdgpu_ring_write(ring, upper_32_bits(addr));
+	amdgpu_ring_write(ring, lower_32_bits(seq));
+
+	/* optionally write high bits as well */
+	if (write64bit) {
+		addr += 4;
+		amdgpu_ring_write(ring, SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_FENCE) |
+				  SDMA_PKT_FENCE_HEADER_MTYPE(0x3));
+		/* zero in first two bits */
+		BUG_ON(addr & 0x3);
+		amdgpu_ring_write(ring, lower_32_bits(addr));
+		amdgpu_ring_write(ring, upper_32_bits(addr));
+		amdgpu_ring_write(ring, upper_32_bits(seq));
+	}
+
+	if (flags & AMDGPU_FENCE_FLAG_INT) {
+		/* generate an interrupt */
+		amdgpu_ring_write(ring, SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_TRAP));
+		amdgpu_ring_write(ring, SDMA_PKT_TRAP_INT_CONTEXT_INT_CONTEXT(0));
+	}
+}
+
+/**
+ * sdma_v6_0_gfx_stop - stop the gfx async dma engines
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Stop the gfx async dma ring buffers.
+ */
+static void sdma_v6_0_gfx_stop(struct amdgpu_device *adev)
+{
+	u32 rb_cntl, ib_cntl;
+	int i;
+
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		rb_cntl = RREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_CNTL));
+		rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_QUEUE0_RB_CNTL, RB_ENABLE, 0);
+		WREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_CNTL), rb_cntl);
+		ib_cntl = RREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_IB_CNTL));
+		ib_cntl = REG_SET_FIELD(ib_cntl, SDMA0_QUEUE0_IB_CNTL, IB_ENABLE, 0);
+		WREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_IB_CNTL), ib_cntl);
+	}
+}
+
+/**
+ * sdma_v6_0_rlc_stop - stop the compute async dma engines
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Stop the compute async dma queues.
+ */
+static void sdma_v6_0_rlc_stop(struct amdgpu_device *adev)
+{
+	/* XXX todo */
+}
+
+/**
+ * sdma_v6_0_ctxempty_int_enable - enable or disable context empty interrupts
+ *
+ * @adev: amdgpu_device pointer
+ * @enable: enable/disable context switching due to queue empty conditions
+ *
+ * Enable or disable the async dma engines queue empty context switch.
+ */
+static void sdma_v6_0_ctxempty_int_enable(struct amdgpu_device *adev, bool enable)
+{
+	u32 f32_cntl;
+	int i;
+
+	if (!amdgpu_sriov_vf(adev)) {
+		for (i = 0; i < adev->sdma.num_instances; i++) {
+			f32_cntl = RREG32(sdma_v6_0_get_reg_offset(adev, i, regSDMA0_CNTL));
+			f32_cntl = REG_SET_FIELD(f32_cntl, SDMA0_CNTL,
+					CTXEMPTY_INT_ENABLE, enable ? 1 : 0);
+			WREG32(sdma_v6_0_get_reg_offset(adev, i, regSDMA0_CNTL), f32_cntl);
+		}
+	}
+}
+
+/**
+ * sdma_v6_0_enable - stop the async dma engines
+ *
+ * @adev: amdgpu_device pointer
+ * @enable: enable/disable the DMA MEs.
+ *
+ * Halt or unhalt the async dma engines.
+ */
+static void sdma_v6_0_enable(struct amdgpu_device *adev, bool enable)
+{
+	u32 f32_cntl;
+	int i;
+
+	if (!enable) {
+		sdma_v6_0_gfx_stop(adev);
+		sdma_v6_0_rlc_stop(adev);
+	}
+
+	if (amdgpu_sriov_vf(adev))
+		return;
+
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		f32_cntl = RREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_F32_CNTL));
+		f32_cntl = REG_SET_FIELD(f32_cntl, SDMA0_F32_CNTL, HALT, enable ? 0 : 1);
+		WREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_F32_CNTL), f32_cntl);
+	}
+}
+
+/**
+ * sdma_v6_0_gfx_resume_instance - start/restart a certain sdma engine
+ *
+ * @adev: amdgpu_device pointer
+ * @i: instance
+ * @restore: used to restore wptr when restart
+ *
+ * Set up the gfx DMA ring buffers and enable them. On restart, we will restore wptr and rptr.
+ * Return 0 for success.
+ */
+static int sdma_v6_0_gfx_resume_instance(struct amdgpu_device *adev, int i, bool restore)
+{
+	struct amdgpu_ring *ring;
+	u32 rb_cntl, ib_cntl;
+	u32 rb_bufsz;
+	u32 doorbell;
+	u32 doorbell_offset;
+	u32 temp;
+	u64 wptr_gpu_addr;
+
+	ring = &adev->sdma.instance[i].ring;
+	if (!amdgpu_sriov_vf(adev))
+		WREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_SEM_WAIT_FAIL_TIMER_CNTL), 0);
+
+	/* Set ring buffer size in dwords */
+	rb_bufsz = order_base_2(ring->ring_size / 4);
+	rb_cntl = RREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_CNTL));
+	rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_QUEUE0_RB_CNTL, RB_SIZE, rb_bufsz);
+#ifdef __BIG_ENDIAN
+	rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_QUEUE0_RB_CNTL, RB_SWAP_ENABLE, 1);
+	rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_QUEUE0_RB_CNTL,
+				RPTR_WRITEBACK_SWAP_ENABLE, 1);
+#endif
+	rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_QUEUE0_RB_CNTL, RB_PRIV, 1);
+	WREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_CNTL), rb_cntl);
+
+	/* Initialize the ring buffer's read and write pointers */
+	if (restore) {
+		WREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_RPTR), lower_32_bits(ring->wptr << 2));
+		WREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_RPTR_HI), upper_32_bits(ring->wptr << 2));
+		WREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_WPTR), lower_32_bits(ring->wptr << 2));
+		WREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_WPTR_HI), upper_32_bits(ring->wptr << 2));
+	} else {
+		WREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_RPTR), 0);
+		WREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_RPTR_HI), 0);
+		WREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_WPTR), 0);
+		WREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_WPTR_HI), 0);
+	}
+	/* setup the wptr shadow polling */
+	wptr_gpu_addr = ring->wptr_gpu_addr;
+	WREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_WPTR_POLL_ADDR_LO),
+	       lower_32_bits(wptr_gpu_addr));
+	WREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_WPTR_POLL_ADDR_HI),
+	       upper_32_bits(wptr_gpu_addr));
+
+	/* set the wb address whether it's enabled or not */
+	WREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_RPTR_ADDR_HI),
+	       upper_32_bits(ring->rptr_gpu_addr) & 0xFFFFFFFF);
+	WREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_RPTR_ADDR_LO),
+	       lower_32_bits(ring->rptr_gpu_addr) & 0xFFFFFFFC);
+
+	rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_QUEUE0_RB_CNTL, RPTR_WRITEBACK_ENABLE, 1);
+	rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_QUEUE0_RB_CNTL, WPTR_POLL_ENABLE, 0);
+	rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_QUEUE0_RB_CNTL, F32_WPTR_POLL_ENABLE, 1);
+
+	WREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_BASE), ring->gpu_addr >> 8);
+	WREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_BASE_HI), ring->gpu_addr >> 40);
+
+	if (!restore)
+		ring->wptr = 0;
+
+	/* before programing wptr to a less value, need set minor_ptr_update first */
+	WREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_MINOR_PTR_UPDATE), 1);
+
+	if (!amdgpu_sriov_vf(adev)) { /* only bare-metal use register write for wptr */
+		WREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_WPTR), lower_32_bits(ring->wptr) << 2);
+		WREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_WPTR_HI), upper_32_bits(ring->wptr) << 2);
+	}
+
+	doorbell = RREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_DOORBELL));
+	doorbell_offset = RREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_DOORBELL_OFFSET));
+
+	if (ring->use_doorbell) {
+		doorbell = REG_SET_FIELD(doorbell, SDMA0_QUEUE0_DOORBELL, ENABLE, 1);
+		doorbell_offset = REG_SET_FIELD(doorbell_offset, SDMA0_QUEUE0_DOORBELL_OFFSET,
+				OFFSET, ring->doorbell_index);
+	} else {
+		doorbell = REG_SET_FIELD(doorbell, SDMA0_QUEUE0_DOORBELL, ENABLE, 0);
+	}
+	WREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_DOORBELL), doorbell);
+	WREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_DOORBELL_OFFSET), doorbell_offset);
+
+	if (i == 0)
+		adev->nbio.funcs->sdma_doorbell_range(adev, i, ring->use_doorbell,
+					      ring->doorbell_index,
+					      adev->doorbell_index.sdma_doorbell_range * adev->sdma.num_instances);
+
+	if (amdgpu_sriov_vf(adev))
+		sdma_v6_0_ring_set_wptr(ring);
+
+	/* set minor_ptr_update to 0 after wptr programed */
+	WREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_MINOR_PTR_UPDATE), 0);
+
+	/* Set up sdma hang watchdog */
+	temp = RREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_WATCHDOG_CNTL));
+	/* 100ms per unit */
+	temp = REG_SET_FIELD(temp, SDMA0_WATCHDOG_CNTL, QUEUE_HANG_COUNT,
+			     max(adev->usec_timeout/100000, 1));
+	WREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_WATCHDOG_CNTL), temp);
+
+	/* Set up RESP_MODE to non-copy addresses */
+	temp = RREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_UTCL1_CNTL));
+	temp = REG_SET_FIELD(temp, SDMA0_UTCL1_CNTL, RESP_MODE, 3);
+	temp = REG_SET_FIELD(temp, SDMA0_UTCL1_CNTL, REDO_DELAY, 9);
+	WREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_UTCL1_CNTL), temp);
+
+	/* program default cache read and write policy */
+	temp = RREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_UTCL1_PAGE));
+	/* clean read policy and write policy bits */
+	temp &= 0xFF0FFF;
+	temp |= ((CACHE_READ_POLICY_L2__DEFAULT << 12) |
+		 (CACHE_WRITE_POLICY_L2__DEFAULT << 14) |
+		 SDMA0_UTCL1_PAGE__LLC_NOALLOC_MASK);
+	WREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_UTCL1_PAGE), temp);
+
+	if (!amdgpu_sriov_vf(adev)) {
+		/* unhalt engine */
+		temp = RREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_F32_CNTL));
+		temp = REG_SET_FIELD(temp, SDMA0_F32_CNTL, HALT, 0);
+		temp = REG_SET_FIELD(temp, SDMA0_F32_CNTL, TH1_RESET, 0);
+		WREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_F32_CNTL), temp);
+	}
+
+	/* enable DMA RB */
+	rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_QUEUE0_RB_CNTL, RB_ENABLE, 1);
+	WREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_CNTL), rb_cntl);
+
+	ib_cntl = RREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_IB_CNTL));
+	ib_cntl = REG_SET_FIELD(ib_cntl, SDMA0_QUEUE0_IB_CNTL, IB_ENABLE, 1);
+#ifdef __BIG_ENDIAN
+	ib_cntl = REG_SET_FIELD(ib_cntl, SDMA0_QUEUE0_IB_CNTL, IB_SWAP_ENABLE, 1);
+#endif
+	/* enable DMA IBs */
+	WREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_IB_CNTL), ib_cntl);
+
+	if (amdgpu_sriov_vf(adev))
+		sdma_v6_0_enable(adev, true);
+
+	return amdgpu_ring_test_helper(ring);
+}
+
+/**
+ * sdma_v6_0_gfx_resume - setup and start the async dma engines
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Set up the gfx DMA ring buffers and enable them.
+ * Returns 0 for success, error for failure.
+ */
+static int sdma_v6_0_gfx_resume(struct amdgpu_device *adev)
+{
+	int i, r;
+
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		r = sdma_v6_0_gfx_resume_instance(adev, i, false);
+		if (r)
+			return r;
+	}
+
+	return 0;
+}
+
+/**
+ * sdma_v6_0_rlc_resume - setup and start the async dma engines
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Set up the compute DMA queues and enable them.
+ * Returns 0 for success, error for failure.
+ */
+static int sdma_v6_0_rlc_resume(struct amdgpu_device *adev)
+{
+	return 0;
+}
+
+/**
+ * sdma_v6_0_load_microcode - load the sDMA ME ucode
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Loads the sDMA0/1 ucode.
+ * Returns 0 for success, -EINVAL if the ucode is not available.
+ */
+static int sdma_v6_0_load_microcode(struct amdgpu_device *adev)
+{
+	const struct sdma_firmware_header_v2_0 *hdr;
+	const __le32 *fw_data;
+	u32 fw_size;
+	int i, j;
+	bool use_broadcast;
+
+	/* halt the MEs */
+	sdma_v6_0_enable(adev, false);
+
+	if (!adev->sdma.instance[0].fw)
+		return -EINVAL;
+
+	/* use broadcast mode to load SDMA microcode by default */
+	use_broadcast = true;
+
+	if (use_broadcast) {
+		dev_info(adev->dev, "Use broadcast method to load SDMA firmware\n");
+		/* load Control Thread microcode */
+		hdr = (const struct sdma_firmware_header_v2_0 *)adev->sdma.instance[0].fw->data;
+		amdgpu_ucode_print_sdma_hdr(&hdr->header);
+		fw_size = le32_to_cpu(hdr->ctx_jt_offset + hdr->ctx_jt_size) / 4;
+
+		fw_data = (const __le32 *)
+			(adev->sdma.instance[0].fw->data +
+				le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+
+		WREG32(sdma_v6_0_get_reg_offset(adev, 0, regSDMA0_BROADCAST_UCODE_ADDR), 0);
+
+		for (j = 0; j < fw_size; j++) {
+			if (amdgpu_emu_mode == 1 && j % 500 == 0)
+				msleep(1);
+			WREG32(sdma_v6_0_get_reg_offset(adev, 0, regSDMA0_BROADCAST_UCODE_DATA), le32_to_cpup(fw_data++));
+		}
+
+		/* load Context Switch microcode */
+		fw_size = le32_to_cpu(hdr->ctl_jt_offset + hdr->ctl_jt_size) / 4;
+
+		fw_data = (const __le32 *)
+			(adev->sdma.instance[0].fw->data +
+				le32_to_cpu(hdr->ctl_ucode_offset));
+
+		WREG32(sdma_v6_0_get_reg_offset(adev, 0, regSDMA0_BROADCAST_UCODE_ADDR), 0x8000);
+
+		for (j = 0; j < fw_size; j++) {
+			if (amdgpu_emu_mode == 1 && j % 500 == 0)
+				msleep(1);
+			WREG32(sdma_v6_0_get_reg_offset(adev, 0, regSDMA0_BROADCAST_UCODE_DATA), le32_to_cpup(fw_data++));
+		}
+	} else {
+		dev_info(adev->dev, "Use legacy method to load SDMA firmware\n");
+		for (i = 0; i < adev->sdma.num_instances; i++) {
+			/* load Control Thread microcode */
+			hdr = (const struct sdma_firmware_header_v2_0 *)adev->sdma.instance[0].fw->data;
+			amdgpu_ucode_print_sdma_hdr(&hdr->header);
+			fw_size = le32_to_cpu(hdr->ctx_jt_offset + hdr->ctx_jt_size) / 4;
+
+			fw_data = (const __le32 *)
+				(adev->sdma.instance[0].fw->data +
+					le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+
+			WREG32(sdma_v6_0_get_reg_offset(adev, i, regSDMA0_UCODE_ADDR), 0);
+
+			for (j = 0; j < fw_size; j++) {
+				if (amdgpu_emu_mode == 1 && j % 500 == 0)
+					msleep(1);
+				WREG32(sdma_v6_0_get_reg_offset(adev, i, regSDMA0_UCODE_DATA), le32_to_cpup(fw_data++));
+			}
+
+			WREG32(sdma_v6_0_get_reg_offset(adev, i, regSDMA0_UCODE_ADDR), adev->sdma.instance[0].fw_version);
+
+			/* load Context Switch microcode */
+			fw_size = le32_to_cpu(hdr->ctl_jt_offset + hdr->ctl_jt_size) / 4;
+
+			fw_data = (const __le32 *)
+				(adev->sdma.instance[0].fw->data +
+					le32_to_cpu(hdr->ctl_ucode_offset));
+
+			WREG32(sdma_v6_0_get_reg_offset(adev, i, regSDMA0_UCODE_ADDR), 0x8000);
+
+			for (j = 0; j < fw_size; j++) {
+				if (amdgpu_emu_mode == 1 && j % 500 == 0)
+					msleep(1);
+				WREG32(sdma_v6_0_get_reg_offset(adev, i, regSDMA0_UCODE_DATA), le32_to_cpup(fw_data++));
+			}
+
+			WREG32(sdma_v6_0_get_reg_offset(adev, i, regSDMA0_UCODE_ADDR), adev->sdma.instance[0].fw_version);
+		}
+	}
+
+	return 0;
+}
+
+static int sdma_v6_0_soft_reset(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	u32 tmp;
+	int i;
+
+	sdma_v6_0_gfx_stop(adev);
+
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		tmp = RREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_FREEZE));
+		tmp |= SDMA0_FREEZE__FREEZE_MASK;
+		WREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_FREEZE), tmp);
+		tmp = RREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_F32_CNTL));
+		tmp |= SDMA0_F32_CNTL__HALT_MASK;
+		tmp |= SDMA0_F32_CNTL__TH1_RESET_MASK;
+		WREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_F32_CNTL), tmp);
+
+		WREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_PREEMPT), 0);
+
+		udelay(100);
+
+		tmp = GRBM_SOFT_RESET__SOFT_RESET_SDMA0_MASK << i;
+		WREG32_SOC15(GC, 0, regGRBM_SOFT_RESET, tmp);
+		tmp = RREG32_SOC15(GC, 0, regGRBM_SOFT_RESET);
+
+		udelay(100);
+
+		WREG32_SOC15(GC, 0, regGRBM_SOFT_RESET, 0);
+		tmp = RREG32_SOC15(GC, 0, regGRBM_SOFT_RESET);
+
+		udelay(100);
+	}
+
+	return sdma_v6_0_start(adev);
+}
+
+static bool sdma_v6_0_check_soft_reset(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	struct amdgpu_ring *ring;
+	int i, r;
+	long tmo = msecs_to_jiffies(1000);
+
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		ring = &adev->sdma.instance[i].ring;
+		r = amdgpu_ring_test_ib(ring, tmo);
+		if (r)
+			return true;
+	}
+
+	return false;
+}
+
+/**
+ * sdma_v6_0_start - setup and start the async dma engines
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Set up the DMA engines and enable them.
+ * Returns 0 for success, error for failure.
+ */
+static int sdma_v6_0_start(struct amdgpu_device *adev)
+{
+	int r = 0;
+
+	if (amdgpu_sriov_vf(adev)) {
+		sdma_v6_0_enable(adev, false);
+
+		/* set RB registers */
+		r = sdma_v6_0_gfx_resume(adev);
+		return r;
+	}
+
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT) {
+		r = sdma_v6_0_load_microcode(adev);
+		if (r)
+			return r;
+
+		/* The value of regSDMA_F32_CNTL is invalid the moment after loading fw */
+		if (amdgpu_emu_mode == 1)
+			msleep(1000);
+	}
+
+	/* unhalt the MEs */
+	sdma_v6_0_enable(adev, true);
+	/* enable sdma ring preemption */
+	sdma_v6_0_ctxempty_int_enable(adev, true);
+
+	/* start the gfx rings and rlc compute queues */
+	r = sdma_v6_0_gfx_resume(adev);
+	if (r)
+		return r;
+	r = sdma_v6_0_rlc_resume(adev);
+
+	return r;
+}
+
+static int sdma_v6_0_mqd_init(struct amdgpu_device *adev, void *mqd,
+			      struct amdgpu_mqd_prop *prop)
+{
+	struct v11_sdma_mqd *m = mqd;
+	uint64_t wb_gpu_addr;
+
+	m->sdmax_rlcx_rb_cntl =
+		order_base_2(prop->queue_size / 4) << SDMA0_QUEUE0_RB_CNTL__RB_SIZE__SHIFT |
+		1 << SDMA0_QUEUE0_RB_CNTL__RPTR_WRITEBACK_ENABLE__SHIFT |
+		4 << SDMA0_QUEUE0_RB_CNTL__RPTR_WRITEBACK_TIMER__SHIFT |
+		1 << SDMA0_QUEUE0_RB_CNTL__F32_WPTR_POLL_ENABLE__SHIFT;
+
+	m->sdmax_rlcx_rb_base = lower_32_bits(prop->hqd_base_gpu_addr >> 8);
+	m->sdmax_rlcx_rb_base_hi = upper_32_bits(prop->hqd_base_gpu_addr >> 8);
+
+	wb_gpu_addr = prop->wptr_gpu_addr;
+	m->sdmax_rlcx_rb_wptr_poll_addr_lo = lower_32_bits(wb_gpu_addr);
+	m->sdmax_rlcx_rb_wptr_poll_addr_hi = upper_32_bits(wb_gpu_addr);
+
+	wb_gpu_addr = prop->rptr_gpu_addr;
+	m->sdmax_rlcx_rb_rptr_addr_lo = lower_32_bits(wb_gpu_addr);
+	m->sdmax_rlcx_rb_rptr_addr_hi = upper_32_bits(wb_gpu_addr);
+
+	m->sdmax_rlcx_ib_cntl = RREG32_SOC15_IP(GC, sdma_v6_0_get_reg_offset(adev, 0,
+							regSDMA0_QUEUE0_IB_CNTL));
+
+	m->sdmax_rlcx_doorbell_offset =
+		prop->doorbell_index << SDMA0_QUEUE0_DOORBELL_OFFSET__OFFSET__SHIFT;
+
+	m->sdmax_rlcx_doorbell = REG_SET_FIELD(0, SDMA0_QUEUE0_DOORBELL, ENABLE, 1);
+
+	m->sdmax_rlcx_skip_cntl = 0;
+	m->sdmax_rlcx_context_status = 0;
+	m->sdmax_rlcx_doorbell_log = 0;
+
+	m->sdmax_rlcx_rb_aql_cntl = regSDMA0_QUEUE0_RB_AQL_CNTL_DEFAULT;
+	m->sdmax_rlcx_dummy_reg = regSDMA0_QUEUE0_DUMMY_REG_DEFAULT;
+
+	m->sdmax_rlcx_csa_addr_lo = lower_32_bits(prop->csa_addr);
+	m->sdmax_rlcx_csa_addr_hi = upper_32_bits(prop->csa_addr);
+
+	m->sdmax_rlcx_f32_dbg0 = lower_32_bits(prop->fence_address);
+	m->sdmax_rlcx_f32_dbg1 = upper_32_bits(prop->fence_address);
+
+	return 0;
+}
+
+static void sdma_v6_0_set_mqd_funcs(struct amdgpu_device *adev)
+{
+	adev->mqds[AMDGPU_HW_IP_DMA].mqd_size = sizeof(struct v11_sdma_mqd);
+	adev->mqds[AMDGPU_HW_IP_DMA].init_mqd = sdma_v6_0_mqd_init;
+}
+
+/**
+ * sdma_v6_0_ring_test_ring - simple async dma engine test
+ *
+ * @ring: amdgpu_ring structure holding ring information
+ *
+ * Test the DMA engine by writing using it to write an
+ * value to memory.
+ * Returns 0 for success, error for failure.
+ */
+static int sdma_v6_0_ring_test_ring(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	unsigned i;
+	unsigned index;
+	int r;
+	u32 tmp;
+	u64 gpu_addr;
+
+	tmp = 0xCAFEDEAD;
+
+	r = amdgpu_device_wb_get(adev, &index);
+	if (r) {
+		dev_err(adev->dev, "(%d) failed to allocate wb slot\n", r);
+		return r;
+	}
+
+	gpu_addr = adev->wb.gpu_addr + (index * 4);
+	adev->wb.wb[index] = cpu_to_le32(tmp);
+
+	r = amdgpu_ring_alloc(ring, 5);
+	if (r) {
+		DRM_ERROR("amdgpu: dma failed to lock ring %d (%d).\n", ring->idx, r);
+		amdgpu_device_wb_free(adev, index);
+		return r;
+	}
+
+	amdgpu_ring_write(ring, SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_WRITE) |
+			  SDMA_PKT_COPY_LINEAR_HEADER_SUB_OP(SDMA_SUBOP_WRITE_LINEAR));
+	amdgpu_ring_write(ring, lower_32_bits(gpu_addr));
+	amdgpu_ring_write(ring, upper_32_bits(gpu_addr));
+	amdgpu_ring_write(ring, SDMA_PKT_WRITE_UNTILED_DW_3_COUNT(0));
+	amdgpu_ring_write(ring, 0xDEADBEEF);
+	amdgpu_ring_commit(ring);
+
+	for (i = 0; i < adev->usec_timeout; i++) {
+		tmp = le32_to_cpu(adev->wb.wb[index]);
+		if (tmp == 0xDEADBEEF)
+			break;
+		if (amdgpu_emu_mode == 1)
+			msleep(1);
+		else
+			udelay(1);
+	}
+
+	if (i >= adev->usec_timeout)
+		r = -ETIMEDOUT;
+
+	amdgpu_device_wb_free(adev, index);
+
+	return r;
+}
+
+/*
+ * sdma_v6_0_ring_test_ib - test an IB on the DMA engine
+ *
+ * @ring: amdgpu_ring structure holding ring information
+ * @timeout: timeout value in jiffies, or MAX_SCHEDULE_TIMEOUT
+ *
+ * Test a simple IB in the DMA ring.
+ * Returns 0 on success, error on failure.
+ */
+static int sdma_v6_0_ring_test_ib(struct amdgpu_ring *ring, long timeout)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct amdgpu_ib ib;
+	struct dma_fence *f = NULL;
+	unsigned index;
+	long r;
+	u32 tmp = 0;
+	u64 gpu_addr;
+
+	tmp = 0xCAFEDEAD;
+	memset(&ib, 0, sizeof(ib));
+
+	r = amdgpu_device_wb_get(adev, &index);
+	if (r) {
+		dev_err(adev->dev, "(%ld) failed to allocate wb slot\n", r);
+		return r;
+	}
+
+	gpu_addr = adev->wb.gpu_addr + (index * 4);
+	adev->wb.wb[index] = cpu_to_le32(tmp);
+
+	r = amdgpu_ib_get(adev, NULL, 256, AMDGPU_IB_POOL_DIRECT, &ib);
+	if (r) {
+		DRM_ERROR("amdgpu: failed to get ib (%ld).\n", r);
+		goto err0;
+	}
+
+	ib.ptr[0] = SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_WRITE) |
+		SDMA_PKT_COPY_LINEAR_HEADER_SUB_OP(SDMA_SUBOP_WRITE_LINEAR);
+	ib.ptr[1] = lower_32_bits(gpu_addr);
+	ib.ptr[2] = upper_32_bits(gpu_addr);
+	ib.ptr[3] = SDMA_PKT_WRITE_UNTILED_DW_3_COUNT(0);
+	ib.ptr[4] = 0xDEADBEEF;
+	ib.ptr[5] = SDMA_PKT_NOP_HEADER_OP(SDMA_OP_NOP);
+	ib.ptr[6] = SDMA_PKT_NOP_HEADER_OP(SDMA_OP_NOP);
+	ib.ptr[7] = SDMA_PKT_NOP_HEADER_OP(SDMA_OP_NOP);
+	ib.length_dw = 8;
+
+	r = amdgpu_ib_schedule(ring, 1, &ib, NULL, &f);
+	if (r)
+		goto err1;
+
+	r = dma_fence_wait_timeout(f, false, timeout);
+	if (r == 0) {
+		DRM_ERROR("amdgpu: IB test timed out\n");
+		r = -ETIMEDOUT;
+		goto err1;
+	} else if (r < 0) {
+		DRM_ERROR("amdgpu: fence wait failed (%ld).\n", r);
+		goto err1;
+	}
+
+	tmp = le32_to_cpu(adev->wb.wb[index]);
+
+	if (tmp == 0xDEADBEEF)
+		r = 0;
+	else
+		r = -EINVAL;
+
+err1:
+	amdgpu_ib_free(&ib, NULL);
+	dma_fence_put(f);
+err0:
+	amdgpu_device_wb_free(adev, index);
+	return r;
+}
+
+
+/**
+ * sdma_v6_0_vm_copy_pte - update PTEs by copying them from the GART
+ *
+ * @ib: indirect buffer to fill with commands
+ * @pe: addr of the page entry
+ * @src: src addr to copy from
+ * @count: number of page entries to update
+ *
+ * Update PTEs by copying them from the GART using sDMA.
+ */
+static void sdma_v6_0_vm_copy_pte(struct amdgpu_ib *ib,
+				  uint64_t pe, uint64_t src,
+				  unsigned count)
+{
+	unsigned bytes = count * 8;
+
+	ib->ptr[ib->length_dw++] = SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_COPY) |
+		SDMA_PKT_COPY_LINEAR_HEADER_SUB_OP(SDMA_SUBOP_COPY_LINEAR);
+	ib->ptr[ib->length_dw++] = bytes - 1;
+	ib->ptr[ib->length_dw++] = 0; /* src/dst endian swap */
+	ib->ptr[ib->length_dw++] = lower_32_bits(src);
+	ib->ptr[ib->length_dw++] = upper_32_bits(src);
+	ib->ptr[ib->length_dw++] = lower_32_bits(pe);
+	ib->ptr[ib->length_dw++] = upper_32_bits(pe);
+
+}
+
+/**
+ * sdma_v6_0_vm_write_pte - update PTEs by writing them manually
+ *
+ * @ib: indirect buffer to fill with commands
+ * @pe: addr of the page entry
+ * @value: dst addr to write into pe
+ * @count: number of page entries to update
+ * @incr: increase next addr by incr bytes
+ *
+ * Update PTEs by writing them manually using sDMA.
+ */
+static void sdma_v6_0_vm_write_pte(struct amdgpu_ib *ib, uint64_t pe,
+				   uint64_t value, unsigned count,
+				   uint32_t incr)
+{
+	unsigned ndw = count * 2;
+
+	ib->ptr[ib->length_dw++] = SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_WRITE) |
+		SDMA_PKT_COPY_LINEAR_HEADER_SUB_OP(SDMA_SUBOP_WRITE_LINEAR);
+	ib->ptr[ib->length_dw++] = lower_32_bits(pe);
+	ib->ptr[ib->length_dw++] = upper_32_bits(pe);
+	ib->ptr[ib->length_dw++] = ndw - 1;
+	for (; ndw > 0; ndw -= 2) {
+		ib->ptr[ib->length_dw++] = lower_32_bits(value);
+		ib->ptr[ib->length_dw++] = upper_32_bits(value);
+		value += incr;
+	}
+}
+
+/**
+ * sdma_v6_0_vm_set_pte_pde - update the page tables using sDMA
+ *
+ * @ib: indirect buffer to fill with commands
+ * @pe: addr of the page entry
+ * @addr: dst addr to write into pe
+ * @count: number of page entries to update
+ * @incr: increase next addr by incr bytes
+ * @flags: access flags
+ *
+ * Update the page tables using sDMA.
+ */
+static void sdma_v6_0_vm_set_pte_pde(struct amdgpu_ib *ib,
+				     uint64_t pe,
+				     uint64_t addr, unsigned count,
+				     uint32_t incr, uint64_t flags)
+{
+	/* for physically contiguous pages (vram) */
+	ib->ptr[ib->length_dw++] = SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_PTEPDE);
+	ib->ptr[ib->length_dw++] = lower_32_bits(pe); /* dst addr */
+	ib->ptr[ib->length_dw++] = upper_32_bits(pe);
+	ib->ptr[ib->length_dw++] = lower_32_bits(flags); /* mask */
+	ib->ptr[ib->length_dw++] = upper_32_bits(flags);
+	ib->ptr[ib->length_dw++] = lower_32_bits(addr); /* value */
+	ib->ptr[ib->length_dw++] = upper_32_bits(addr);
+	ib->ptr[ib->length_dw++] = incr; /* increment size */
+	ib->ptr[ib->length_dw++] = 0;
+	ib->ptr[ib->length_dw++] = count - 1; /* number of entries */
+}
+
+/*
+ * sdma_v6_0_ring_pad_ib - pad the IB
+ * @ib: indirect buffer to fill with padding
+ * @ring: amdgpu ring pointer
+ *
+ * Pad the IB with NOPs to a boundary multiple of 8.
+ */
+static void sdma_v6_0_ring_pad_ib(struct amdgpu_ring *ring, struct amdgpu_ib *ib)
+{
+	struct amdgpu_sdma_instance *sdma = amdgpu_sdma_get_instance_from_ring(ring);
+	u32 pad_count;
+	int i;
+
+	pad_count = (-ib->length_dw) & 0x7;
+	for (i = 0; i < pad_count; i++)
+		if (sdma && sdma->burst_nop && (i == 0))
+			ib->ptr[ib->length_dw++] =
+				SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_NOP) |
+				SDMA_PKT_NOP_HEADER_COUNT(pad_count - 1);
+		else
+			ib->ptr[ib->length_dw++] =
+				SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_NOP);
+}
+
+/**
+ * sdma_v6_0_ring_emit_pipeline_sync - sync the pipeline
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Make sure all previous operations are completed (CIK).
+ */
+static void sdma_v6_0_ring_emit_pipeline_sync(struct amdgpu_ring *ring)
+{
+	uint32_t seq = ring->fence_drv.sync_seq;
+	uint64_t addr = ring->fence_drv.gpu_addr;
+
+	/* wait for idle */
+	amdgpu_ring_write(ring, SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_POLL_REGMEM) |
+			  SDMA_PKT_POLL_REGMEM_HEADER_HDP_FLUSH(0) |
+			  SDMA_PKT_POLL_REGMEM_HEADER_FUNC(3) | /* equal */
+			  SDMA_PKT_POLL_REGMEM_HEADER_MEM_POLL(1));
+	amdgpu_ring_write(ring, addr & 0xfffffffc);
+	amdgpu_ring_write(ring, upper_32_bits(addr) & 0xffffffff);
+	amdgpu_ring_write(ring, seq); /* reference */
+	amdgpu_ring_write(ring, 0xffffffff); /* mask */
+	amdgpu_ring_write(ring, SDMA_PKT_POLL_REGMEM_DW5_RETRY_COUNT(0xfff) |
+			  SDMA_PKT_POLL_REGMEM_DW5_INTERVAL(4)); /* retry count, poll interval */
+}
+
+/*
+ * sdma_v6_0_ring_emit_vm_flush - vm flush using sDMA
+ *
+ * @ring: amdgpu_ring pointer
+ * @vmid: vmid number to use
+ * @pd_addr: address
+ *
+ * Update the page table base and flush the VM TLB
+ * using sDMA.
+ */
+static void sdma_v6_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
+					 unsigned vmid, uint64_t pd_addr)
+{
+	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->vm_hub];
+	uint32_t req = hub->vmhub_funcs->get_invalidate_req(vmid, 0);
+
+	/* Update the PD address for this VMID. */
+	amdgpu_ring_emit_wreg(ring, hub->ctx0_ptb_addr_lo32 +
+			      (hub->ctx_addr_distance * vmid),
+			      lower_32_bits(pd_addr));
+	amdgpu_ring_emit_wreg(ring, hub->ctx0_ptb_addr_hi32 +
+			      (hub->ctx_addr_distance * vmid),
+			      upper_32_bits(pd_addr));
+
+	/* Trigger invalidation. */
+	amdgpu_ring_write(ring,
+			  SDMA_PKT_VM_INVALIDATION_HEADER_OP(SDMA_OP_POLL_REGMEM) |
+			  SDMA_PKT_VM_INVALIDATION_HEADER_SUB_OP(SDMA_SUBOP_VM_INVALIDATION) |
+			  SDMA_PKT_VM_INVALIDATION_HEADER_GFX_ENG_ID(ring->vm_inv_eng) |
+			  SDMA_PKT_VM_INVALIDATION_HEADER_MM_ENG_ID(0x1f));
+	amdgpu_ring_write(ring, req);
+	amdgpu_ring_write(ring, 0xFFFFFFFF);
+	amdgpu_ring_write(ring,
+			  SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_INVALIDATEACK(1 << vmid) |
+			  SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_ADDRESSRANGEHI(0x1F));
+}
+
+static void sdma_v6_0_ring_emit_wreg(struct amdgpu_ring *ring,
+				     uint32_t reg, uint32_t val)
+{
+	amdgpu_ring_write(ring, SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_SRBM_WRITE) |
+			  SDMA_PKT_SRBM_WRITE_HEADER_BYTE_EN(0xf));
+	amdgpu_ring_write(ring, reg);
+	amdgpu_ring_write(ring, val);
+}
+
+static void sdma_v6_0_ring_emit_reg_wait(struct amdgpu_ring *ring, uint32_t reg,
+					 uint32_t val, uint32_t mask)
+{
+	amdgpu_ring_write(ring, SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_POLL_REGMEM) |
+			  SDMA_PKT_POLL_REGMEM_HEADER_HDP_FLUSH(0) |
+			  SDMA_PKT_POLL_REGMEM_HEADER_FUNC(3)); /* equal */
+	amdgpu_ring_write(ring, reg << 2);
+	amdgpu_ring_write(ring, 0);
+	amdgpu_ring_write(ring, val); /* reference */
+	amdgpu_ring_write(ring, mask); /* mask */
+	amdgpu_ring_write(ring, SDMA_PKT_POLL_REGMEM_DW5_RETRY_COUNT(0xfff) |
+			  SDMA_PKT_POLL_REGMEM_DW5_INTERVAL(10));
+}
+
+static void sdma_v6_0_ring_emit_reg_write_reg_wait(struct amdgpu_ring *ring,
+						   uint32_t reg0, uint32_t reg1,
+						   uint32_t ref, uint32_t mask)
+{
+	amdgpu_ring_emit_wreg(ring, reg0, ref);
+	/* wait for a cycle to reset vm_inv_eng*_ack */
+	amdgpu_ring_emit_reg_wait(ring, reg0, 0, 0);
+	amdgpu_ring_emit_reg_wait(ring, reg1, mask, mask);
+}
+
+static struct amdgpu_sdma_ras sdma_v6_0_3_ras = {
+	.ras_block = {
+		.ras_late_init = amdgpu_ras_block_late_init,
+	},
+};
+
+static void sdma_v6_0_set_ras_funcs(struct amdgpu_device *adev)
+{
+	switch (amdgpu_ip_version(adev, SDMA0_HWIP, 0)) {
+	case IP_VERSION(6, 0, 3):
+		adev->sdma.ras = &sdma_v6_0_3_ras;
+		break;
+	default:
+		break;
+	}
+}
+
+static int sdma_v6_0_early_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int r;
+
+	switch (amdgpu_user_queue) {
+	case -1:
+	case 0:
+	default:
+		adev->sdma.no_user_submission = false;
+		adev->sdma.disable_uq = true;
+		break;
+	case 1:
+		adev->sdma.no_user_submission = false;
+		adev->sdma.disable_uq = false;
+		break;
+	case 2:
+		adev->sdma.no_user_submission = true;
+		adev->sdma.disable_uq = false;
+		break;
+	}
+
+	r = amdgpu_sdma_init_microcode(adev, 0, true);
+	if (r)
+		return r;
+
+	sdma_v6_0_set_ring_funcs(adev);
+	sdma_v6_0_set_buffer_funcs(adev);
+	sdma_v6_0_set_vm_pte_funcs(adev);
+	sdma_v6_0_set_irq_funcs(adev);
+	sdma_v6_0_set_mqd_funcs(adev);
+	sdma_v6_0_set_ras_funcs(adev);
+
+	return 0;
+}
+
+static int sdma_v6_0_sw_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_ring *ring;
+	int r, i;
+	struct amdgpu_device *adev = ip_block->adev;
+	uint32_t reg_count = ARRAY_SIZE(sdma_reg_list_6_0);
+	uint32_t *ptr;
+
+	/* SDMA trap event */
+	r = amdgpu_irq_add_id(adev, SOC21_IH_CLIENTID_GFX,
+			      GFX_11_0_0__SRCID__SDMA_TRAP,
+			      &adev->sdma.trap_irq);
+	if (r)
+		return r;
+
+	/* SDMA user fence event */
+	r = amdgpu_irq_add_id(adev, SOC21_IH_CLIENTID_GFX,
+			      GFX_11_0_0__SRCID__SDMA_FENCE,
+			      &adev->sdma.fence_irq);
+	if (r)
+		return r;
+
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		ring = &adev->sdma.instance[i].ring;
+		ring->ring_obj = NULL;
+		ring->use_doorbell = true;
+		ring->me = i;
+		ring->no_user_submission = adev->sdma.no_user_submission;
+
+		DRM_DEBUG("SDMA %d use_doorbell being set to: [%s]\n", i,
+				ring->use_doorbell?"true":"false");
+
+		ring->doorbell_index =
+			(adev->doorbell_index.sdma_engine[i] << 1); // get DWORD offset
+
+		ring->vm_hub = AMDGPU_GFXHUB(0);
+		sprintf(ring->name, "sdma%d", i);
+		r = amdgpu_ring_init(adev, ring, 1024,
+				     &adev->sdma.trap_irq,
+				     AMDGPU_SDMA_IRQ_INSTANCE0 + i,
+				     AMDGPU_RING_PRIO_DEFAULT, NULL);
+		if (r)
+			return r;
+	}
+
+	adev->sdma.supported_reset =
+		amdgpu_get_soft_full_reset_mask(&adev->sdma.instance[0].ring);
+	switch (amdgpu_ip_version(adev, SDMA0_HWIP, 0)) {
+	case IP_VERSION(6, 0, 0):
+	case IP_VERSION(6, 0, 2):
+	case IP_VERSION(6, 0, 3):
+		if ((adev->sdma.instance[0].fw_version >= 21) &&
+		    !amdgpu_sriov_vf(adev) &&
+		    !adev->debug_disable_gpu_ring_reset)
+			adev->sdma.supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE;
+		break;
+	default:
+		break;
+	}
+
+	if (amdgpu_sdma_ras_sw_init(adev)) {
+		dev_err(adev->dev, "Failed to initialize sdma ras block!\n");
+		return -EINVAL;
+	}
+
+	/* Allocate memory for SDMA IP Dump buffer */
+	ptr = kcalloc(adev->sdma.num_instances * reg_count, sizeof(uint32_t), GFP_KERNEL);
+	if (ptr)
+		adev->sdma.ip_dump = ptr;
+	else
+		DRM_ERROR("Failed to allocated memory for SDMA IP Dump\n");
+
+	switch (amdgpu_ip_version(adev, SDMA0_HWIP, 0)) {
+	case IP_VERSION(6, 0, 0):
+		if ((adev->sdma.instance[0].fw_version >= 27) && !adev->sdma.disable_uq)
+			adev->userq_funcs[AMDGPU_HW_IP_DMA] = &userq_mes_funcs;
+		break;
+	case IP_VERSION(6, 0, 1):
+		if ((adev->sdma.instance[0].fw_version >= 18) && !adev->sdma.disable_uq)
+			adev->userq_funcs[AMDGPU_HW_IP_DMA] = &userq_mes_funcs;
+		break;
+	case IP_VERSION(6, 0, 2):
+		if ((adev->sdma.instance[0].fw_version >= 23) && !adev->sdma.disable_uq)
+			adev->userq_funcs[AMDGPU_HW_IP_DMA] = &userq_mes_funcs;
+		break;
+	case IP_VERSION(6, 0, 3):
+		if (adev->sdma.instance[0].fw_version >= 29 && !adev->sdma.disable_uq)
+			adev->userq_funcs[AMDGPU_HW_IP_DMA] = &userq_mes_funcs;
+		break;
+	case IP_VERSION(6, 1, 0):
+		if ((adev->sdma.instance[0].fw_version >= 14) && !adev->sdma.disable_uq)
+			adev->userq_funcs[AMDGPU_HW_IP_DMA] = &userq_mes_funcs;
+		break;
+	case IP_VERSION(6, 1, 1):
+		if ((adev->sdma.instance[0].fw_version >= 17) && !adev->sdma.disable_uq)
+			adev->userq_funcs[AMDGPU_HW_IP_DMA] = &userq_mes_funcs;
+		break;
+	case IP_VERSION(6, 1, 2):
+		if ((adev->sdma.instance[0].fw_version >= 15) && !adev->sdma.disable_uq)
+			adev->userq_funcs[AMDGPU_HW_IP_DMA] = &userq_mes_funcs;
+		break;
+	case IP_VERSION(6, 1, 3):
+		if ((adev->sdma.instance[0].fw_version >= 10) && !adev->sdma.disable_uq)
+			adev->userq_funcs[AMDGPU_HW_IP_DMA] = &userq_mes_funcs;
+		break;
+	default:
+		break;
+	}
+
+	r = amdgpu_sdma_sysfs_reset_mask_init(adev);
+	if (r)
+		return r;
+
+	return r;
+}
+
+static int sdma_v6_0_sw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i;
+
+	for (i = 0; i < adev->sdma.num_instances; i++)
+		amdgpu_ring_fini(&adev->sdma.instance[i].ring);
+
+	amdgpu_sdma_sysfs_reset_mask_fini(adev);
+	amdgpu_sdma_destroy_inst_ctx(adev, true);
+
+	kfree(adev->sdma.ip_dump);
+
+	return 0;
+}
+
+static int sdma_v6_0_set_userq_trap_interrupts(struct amdgpu_device *adev,
+					       bool enable)
+{
+	unsigned int irq_type;
+	int i, r;
+
+	if (adev->userq_funcs[AMDGPU_HW_IP_DMA]) {
+		for (i = 0; i < adev->sdma.num_instances; i++) {
+			irq_type = AMDGPU_SDMA_IRQ_INSTANCE0 + i;
+			if (enable)
+				r = amdgpu_irq_get(adev, &adev->sdma.trap_irq,
+						   irq_type);
+			else
+				r = amdgpu_irq_put(adev, &adev->sdma.trap_irq,
+						   irq_type);
+			if (r)
+				return r;
+		}
+	}
+
+	return 0;
+}
+
+static int sdma_v6_0_hw_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int r;
+
+	r = sdma_v6_0_start(adev);
+	if (r)
+		return r;
+
+	return sdma_v6_0_set_userq_trap_interrupts(adev, true);
+}
+
+static int sdma_v6_0_hw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	if (amdgpu_sriov_vf(adev))
+		return 0;
+
+	sdma_v6_0_ctxempty_int_enable(adev, false);
+	sdma_v6_0_enable(adev, false);
+	sdma_v6_0_set_userq_trap_interrupts(adev, false);
+
+	return 0;
+}
+
+static int sdma_v6_0_suspend(struct amdgpu_ip_block *ip_block)
+{
+	return sdma_v6_0_hw_fini(ip_block);
+}
+
+static int sdma_v6_0_resume(struct amdgpu_ip_block *ip_block)
+{
+	return sdma_v6_0_hw_init(ip_block);
+}
+
+static bool sdma_v6_0_is_idle(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	u32 i;
+
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		u32 tmp = RREG32(sdma_v6_0_get_reg_offset(adev, i, regSDMA0_STATUS_REG));
+
+		if (!(tmp & SDMA0_STATUS_REG__IDLE_MASK))
+			return false;
+	}
+
+	return true;
+}
+
+static int sdma_v6_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
+{
+	unsigned i;
+	u32 sdma0, sdma1;
+	struct amdgpu_device *adev = ip_block->adev;
+
+	for (i = 0; i < adev->usec_timeout; i++) {
+		sdma0 = RREG32(sdma_v6_0_get_reg_offset(adev, 0, regSDMA0_STATUS_REG));
+		sdma1 = RREG32(sdma_v6_0_get_reg_offset(adev, 1, regSDMA0_STATUS_REG));
+
+		if (sdma0 & sdma1 & SDMA0_STATUS_REG__IDLE_MASK)
+			return 0;
+		udelay(1);
+	}
+	return -ETIMEDOUT;
+}
+
+static int sdma_v6_0_ring_preempt_ib(struct amdgpu_ring *ring)
+{
+	int i, r = 0;
+	struct amdgpu_device *adev = ring->adev;
+	u32 index = 0;
+	u64 sdma_gfx_preempt;
+
+	amdgpu_sdma_get_index_from_ring(ring, &index);
+	sdma_gfx_preempt =
+		sdma_v6_0_get_reg_offset(adev, index, regSDMA0_QUEUE0_PREEMPT);
+
+	/* assert preemption condition */
+	amdgpu_ring_set_preempt_cond_exec(ring, false);
+
+	/* emit the trailing fence */
+	ring->trail_seq += 1;
+	amdgpu_ring_alloc(ring, 10);
+	sdma_v6_0_ring_emit_fence(ring, ring->trail_fence_gpu_addr,
+				  ring->trail_seq, 0);
+	amdgpu_ring_commit(ring);
+
+	/* assert IB preemption */
+	WREG32(sdma_gfx_preempt, 1);
+
+	/* poll the trailing fence */
+	for (i = 0; i < adev->usec_timeout; i++) {
+		if (ring->trail_seq ==
+		    le32_to_cpu(*(ring->trail_fence_cpu_addr)))
+			break;
+		udelay(1);
+	}
+
+	if (i >= adev->usec_timeout) {
+		r = -EINVAL;
+		DRM_ERROR("ring %d failed to be preempted\n", ring->idx);
+	}
+
+	/* deassert IB preemption */
+	WREG32(sdma_gfx_preempt, 0);
+
+	/* deassert the preemption condition */
+	amdgpu_ring_set_preempt_cond_exec(ring, true);
+	return r;
+}
+
+static int sdma_v6_0_reset_queue(struct amdgpu_ring *ring,
+				 unsigned int vmid,
+				 struct amdgpu_fence *timedout_fence)
+{
+	struct amdgpu_device *adev = ring->adev;
+	int r;
+
+	if (ring->me >= adev->sdma.num_instances) {
+		dev_err(adev->dev, "sdma instance not found\n");
+		return -EINVAL;
+	}
+
+	amdgpu_ring_reset_helper_begin(ring, timedout_fence);
+
+	r = amdgpu_mes_reset_legacy_queue(adev, ring, vmid, true);
+	if (r)
+		return r;
+
+	r = sdma_v6_0_gfx_resume_instance(adev, ring->me, true);
+	if (r)
+		return r;
+
+	return amdgpu_ring_reset_helper_end(ring, timedout_fence);
+}
+
+static int sdma_v6_0_set_trap_irq_state(struct amdgpu_device *adev,
+					struct amdgpu_irq_src *source,
+					unsigned type,
+					enum amdgpu_interrupt_state state)
+{
+	u32 sdma_cntl;
+
+	u32 reg_offset = sdma_v6_0_get_reg_offset(adev, type, regSDMA0_CNTL);
+
+	if (!amdgpu_sriov_vf(adev)) {
+		sdma_cntl = RREG32(reg_offset);
+		sdma_cntl = REG_SET_FIELD(sdma_cntl, SDMA0_CNTL, TRAP_ENABLE,
+				state == AMDGPU_IRQ_STATE_ENABLE ? 1 : 0);
+		WREG32(reg_offset, sdma_cntl);
+	}
+
+	return 0;
+}
+
+static int sdma_v6_0_process_trap_irq(struct amdgpu_device *adev,
+				      struct amdgpu_irq_src *source,
+				      struct amdgpu_iv_entry *entry)
+{
+	int instances, queue;
+
+	DRM_DEBUG("IH: SDMA trap\n");
+
+	queue = entry->ring_id & 0xf;
+	instances = (entry->ring_id & 0xf0) >> 4;
+	if (instances > 1) {
+		DRM_ERROR("IH: wrong ring_ID detected, as wrong sdma instance\n");
+		return -EINVAL;
+	}
+
+	switch (entry->client_id) {
+	case SOC21_IH_CLIENTID_GFX:
+		switch (queue) {
+		case 0:
+			amdgpu_fence_process(&adev->sdma.instance[instances].ring);
+			break;
+		default:
+			break;
+		}
+		break;
+	}
+	return 0;
+}
+
+static int sdma_v6_0_process_fence_irq(struct amdgpu_device *adev,
+				       struct amdgpu_irq_src *source,
+				       struct amdgpu_iv_entry *entry)
+{
+	u32 doorbell_offset = entry->src_data[0];
+
+	if (adev->enable_mes && doorbell_offset) {
+		struct amdgpu_userq_fence_driver *fence_drv = NULL;
+		struct xarray *xa = &adev->userq_xa;
+		unsigned long flags;
+
+		doorbell_offset >>= SDMA0_QUEUE0_DOORBELL_OFFSET__OFFSET__SHIFT;
+
+		xa_lock_irqsave(xa, flags);
+		fence_drv = xa_load(xa, doorbell_offset);
+		if (fence_drv)
+			amdgpu_userq_fence_driver_process(fence_drv);
+		xa_unlock_irqrestore(xa, flags);
+	}
+
+	return 0;
+}
+
+static int sdma_v6_0_process_illegal_inst_irq(struct amdgpu_device *adev,
+					      struct amdgpu_irq_src *source,
+					      struct amdgpu_iv_entry *entry)
+{
+	return 0;
+}
+
+static int sdma_v6_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
+					   enum amd_clockgating_state state)
+{
+	return 0;
+}
+
+static int sdma_v6_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
+					  enum amd_powergating_state state)
+{
+	return 0;
+}
+
+static void sdma_v6_0_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags)
+{
+}
+
+static void sdma_v6_0_print_ip_state(struct amdgpu_ip_block *ip_block, struct drm_printer *p)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, j;
+	uint32_t reg_count = ARRAY_SIZE(sdma_reg_list_6_0);
+	uint32_t instance_offset;
+
+	if (!adev->sdma.ip_dump)
+		return;
+
+	drm_printf(p, "num_instances:%d\n", adev->sdma.num_instances);
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		instance_offset = i * reg_count;
+		drm_printf(p, "\nInstance:%d\n", i);
+
+		for (j = 0; j < reg_count; j++)
+			drm_printf(p, "%-50s \t 0x%08x\n", sdma_reg_list_6_0[j].reg_name,
+				   adev->sdma.ip_dump[instance_offset + j]);
+	}
+}
+
+static void sdma_v6_0_dump_ip_state(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, j;
+	uint32_t instance_offset;
+	uint32_t reg_count = ARRAY_SIZE(sdma_reg_list_6_0);
+
+	if (!adev->sdma.ip_dump)
+		return;
+
+	amdgpu_gfx_off_ctrl(adev, false);
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		instance_offset = i * reg_count;
+		for (j = 0; j < reg_count; j++)
+			adev->sdma.ip_dump[instance_offset + j] =
+				RREG32(sdma_v6_0_get_reg_offset(adev, i,
+				       sdma_reg_list_6_0[j].reg_offset));
+	}
+	amdgpu_gfx_off_ctrl(adev, true);
+}
+
+const struct amd_ip_funcs sdma_v6_0_ip_funcs = {
+	.name = "sdma_v6_0",
+	.early_init = sdma_v6_0_early_init,
+	.sw_init = sdma_v6_0_sw_init,
+	.sw_fini = sdma_v6_0_sw_fini,
+	.hw_init = sdma_v6_0_hw_init,
+	.hw_fini = sdma_v6_0_hw_fini,
+	.suspend = sdma_v6_0_suspend,
+	.resume = sdma_v6_0_resume,
+	.is_idle = sdma_v6_0_is_idle,
+	.wait_for_idle = sdma_v6_0_wait_for_idle,
+	.soft_reset = sdma_v6_0_soft_reset,
+	.check_soft_reset = sdma_v6_0_check_soft_reset,
+	.set_clockgating_state = sdma_v6_0_set_clockgating_state,
+	.set_powergating_state = sdma_v6_0_set_powergating_state,
+	.get_clockgating_state = sdma_v6_0_get_clockgating_state,
+	.dump_ip_state = sdma_v6_0_dump_ip_state,
+	.print_ip_state = sdma_v6_0_print_ip_state,
+};
+
+static const struct amdgpu_ring_funcs sdma_v6_0_ring_funcs = {
+	.type = AMDGPU_RING_TYPE_SDMA,
+	.align_mask = 0xf,
+	.nop = SDMA_PKT_NOP_HEADER_OP(SDMA_OP_NOP),
+	.support_64bit_ptrs = true,
+	.secure_submission_supported = true,
+	.get_rptr = sdma_v6_0_ring_get_rptr,
+	.get_wptr = sdma_v6_0_ring_get_wptr,
+	.set_wptr = sdma_v6_0_ring_set_wptr,
+	.emit_frame_size =
+		5 + /* sdma_v6_0_ring_init_cond_exec */
+		6 + /* sdma_v6_0_ring_emit_hdp_flush */
+		6 + /* sdma_v6_0_ring_emit_pipeline_sync */
+		/* sdma_v6_0_ring_emit_vm_flush */
+		SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 +
+		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 6 +
+		10 + 10 + 10, /* sdma_v6_0_ring_emit_fence x3 for user fence, vm fence */
+	.emit_ib_size = 5 + 7 + 6, /* sdma_v6_0_ring_emit_ib */
+	.emit_ib = sdma_v6_0_ring_emit_ib,
+	.emit_mem_sync = sdma_v6_0_ring_emit_mem_sync,
+	.emit_fence = sdma_v6_0_ring_emit_fence,
+	.emit_pipeline_sync = sdma_v6_0_ring_emit_pipeline_sync,
+	.emit_vm_flush = sdma_v6_0_ring_emit_vm_flush,
+	.emit_hdp_flush = sdma_v6_0_ring_emit_hdp_flush,
+	.test_ring = sdma_v6_0_ring_test_ring,
+	.test_ib = sdma_v6_0_ring_test_ib,
+	.insert_nop = sdma_v6_0_ring_insert_nop,
+	.pad_ib = sdma_v6_0_ring_pad_ib,
+	.emit_wreg = sdma_v6_0_ring_emit_wreg,
+	.emit_reg_wait = sdma_v6_0_ring_emit_reg_wait,
+	.emit_reg_write_reg_wait = sdma_v6_0_ring_emit_reg_write_reg_wait,
+	.init_cond_exec = sdma_v6_0_ring_init_cond_exec,
+	.preempt_ib = sdma_v6_0_ring_preempt_ib,
+	.reset = sdma_v6_0_reset_queue,
+};
+
+static void sdma_v6_0_set_ring_funcs(struct amdgpu_device *adev)
+{
+	int i;
+
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		adev->sdma.instance[i].ring.funcs = &sdma_v6_0_ring_funcs;
+		adev->sdma.instance[i].ring.me = i;
+	}
+}
+
+static const struct amdgpu_irq_src_funcs sdma_v6_0_trap_irq_funcs = {
+	.set = sdma_v6_0_set_trap_irq_state,
+	.process = sdma_v6_0_process_trap_irq,
+};
+
+static const struct amdgpu_irq_src_funcs sdma_v6_0_fence_irq_funcs = {
+	.process = sdma_v6_0_process_fence_irq,
+};
+
+static const struct amdgpu_irq_src_funcs sdma_v6_0_illegal_inst_irq_funcs = {
+	.process = sdma_v6_0_process_illegal_inst_irq,
+};
+
+static void sdma_v6_0_set_irq_funcs(struct amdgpu_device *adev)
+{
+	adev->sdma.trap_irq.num_types = AMDGPU_SDMA_IRQ_INSTANCE0 +
+					adev->sdma.num_instances;
+	adev->sdma.trap_irq.funcs = &sdma_v6_0_trap_irq_funcs;
+	adev->sdma.fence_irq.funcs = &sdma_v6_0_fence_irq_funcs;
+	adev->sdma.illegal_inst_irq.funcs = &sdma_v6_0_illegal_inst_irq_funcs;
+}
+
+/**
+ * sdma_v6_0_emit_copy_buffer - copy buffer using the sDMA engine
+ *
+ * @ib: indirect buffer to fill with commands
+ * @src_offset: src GPU address
+ * @dst_offset: dst GPU address
+ * @byte_count: number of bytes to xfer
+ * @copy_flags: copy flags for the buffers
+ *
+ * Copy GPU buffers using the DMA engine.
+ * Used by the amdgpu ttm implementation to move pages if
+ * registered as the asic copy callback.
+ */
+static void sdma_v6_0_emit_copy_buffer(struct amdgpu_ib *ib,
+				       uint64_t src_offset,
+				       uint64_t dst_offset,
+				       uint32_t byte_count,
+				       uint32_t copy_flags)
+{
+	ib->ptr[ib->length_dw++] = SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_COPY) |
+		SDMA_PKT_COPY_LINEAR_HEADER_SUB_OP(SDMA_SUBOP_COPY_LINEAR) |
+		SDMA_PKT_COPY_LINEAR_HEADER_TMZ((copy_flags & AMDGPU_COPY_FLAGS_TMZ) ? 1 : 0);
+	ib->ptr[ib->length_dw++] = byte_count - 1;
+	ib->ptr[ib->length_dw++] = 0; /* src/dst endian swap */
+	ib->ptr[ib->length_dw++] = lower_32_bits(src_offset);
+	ib->ptr[ib->length_dw++] = upper_32_bits(src_offset);
+	ib->ptr[ib->length_dw++] = lower_32_bits(dst_offset);
+	ib->ptr[ib->length_dw++] = upper_32_bits(dst_offset);
+}
+
+/**
+ * sdma_v6_0_emit_fill_buffer - fill buffer using the sDMA engine
+ *
+ * @ib: indirect buffer to fill
+ * @src_data: value to write to buffer
+ * @dst_offset: dst GPU address
+ * @byte_count: number of bytes to xfer
+ *
+ * Fill GPU buffers using the DMA engine.
+ */
+static void sdma_v6_0_emit_fill_buffer(struct amdgpu_ib *ib,
+				       uint32_t src_data,
+				       uint64_t dst_offset,
+				       uint32_t byte_count)
+{
+	ib->ptr[ib->length_dw++] = SDMA_PKT_CONSTANT_FILL_HEADER_OP(SDMA_OP_CONST_FILL);
+	ib->ptr[ib->length_dw++] = lower_32_bits(dst_offset);
+	ib->ptr[ib->length_dw++] = upper_32_bits(dst_offset);
+	ib->ptr[ib->length_dw++] = src_data;
+	ib->ptr[ib->length_dw++] = byte_count - 1;
+}
+
+static const struct amdgpu_buffer_funcs sdma_v6_0_buffer_funcs = {
+	.copy_max_bytes = 0x400000,
+	.copy_num_dw = 7,
+	.emit_copy_buffer = sdma_v6_0_emit_copy_buffer,
+
+	.fill_max_bytes = 0x400000,
+	.fill_num_dw = 5,
+	.emit_fill_buffer = sdma_v6_0_emit_fill_buffer,
+};
+
+static void sdma_v6_0_set_buffer_funcs(struct amdgpu_device *adev)
+{
+	adev->mman.buffer_funcs = &sdma_v6_0_buffer_funcs;
+	adev->mman.buffer_funcs_ring = &adev->sdma.instance[0].ring;
+}
+
+static const struct amdgpu_vm_pte_funcs sdma_v6_0_vm_pte_funcs = {
+	.copy_pte_num_dw = 7,
+	.copy_pte = sdma_v6_0_vm_copy_pte,
+	.write_pte = sdma_v6_0_vm_write_pte,
+	.set_pte_pde = sdma_v6_0_vm_set_pte_pde,
+};
+
+static void sdma_v6_0_set_vm_pte_funcs(struct amdgpu_device *adev)
+{
+	unsigned i;
+
+	adev->vm_manager.vm_pte_funcs = &sdma_v6_0_vm_pte_funcs;
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		adev->vm_manager.vm_pte_scheds[i] =
+			&adev->sdma.instance[i].ring.sched;
+	}
+	adev->vm_manager.vm_pte_num_scheds = adev->sdma.num_instances;
+}
+
+const struct amdgpu_ip_block_version sdma_v6_0_ip_block = {
+	.type = AMD_IP_BLOCK_TYPE_SDMA,
+	.major = 6,
+	.minor = 0,
+	.rev = 0,
+	.funcs = &sdma_v6_0_ip_funcs,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v6_0.h b/drivers/gpu/drm/amd/amdgpu/sdma_v6_0.h
new file mode 100644
index 000000000000..e473ec7dfc8f
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v6_0.h
@@ -0,0 +1,30 @@
+/*
+ * Copyright 2020 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __SDMA_V6_0_H__
+#define __SDMA_V6_0_H__
+
+extern const struct amd_ip_funcs sdma_v6_0_ip_funcs;
+extern const struct amdgpu_ip_block_version sdma_v6_0_ip_block;
+
+#endif /* __SDMA_V6_0_H__ */
diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v6_0_0_pkt_open.h b/drivers/gpu/drm/amd/amdgpu/sdma_v6_0_0_pkt_open.h
new file mode 100644
index 000000000000..d8cf830916b9
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v6_0_0_pkt_open.h
@@ -0,0 +1,5672 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __SDMA_V6_0_0_PKT_OPEN_H_
+#define __SDMA_V6_0_0_PKT_OPEN_H_
+
+#define SDMA_OP_NOP  0
+#define SDMA_OP_COPY  1
+#define SDMA_OP_WRITE  2
+#define SDMA_OP_INDIRECT  4
+#define SDMA_OP_FENCE  5
+#define SDMA_OP_TRAP  6
+#define SDMA_OP_SEM  7
+#define SDMA_OP_POLL_REGMEM  8
+#define SDMA_OP_COND_EXE  9
+#define SDMA_OP_ATOMIC  10
+#define SDMA_OP_CONST_FILL  11
+#define SDMA_OP_PTEPDE  12
+#define SDMA_OP_TIMESTAMP  13
+#define SDMA_OP_SRBM_WRITE  14
+#define SDMA_OP_PRE_EXE  15
+#define SDMA_OP_GPUVM_INV  16
+#define SDMA_OP_GCR_REQ  17
+#define SDMA_OP_DUMMY_TRAP  32
+#define SDMA_SUBOP_TIMESTAMP_SET  0
+#define SDMA_SUBOP_TIMESTAMP_GET  1
+#define SDMA_SUBOP_TIMESTAMP_GET_GLOBAL  2
+#define SDMA_SUBOP_COPY_LINEAR  0
+#define SDMA_SUBOP_COPY_LINEAR_SUB_WIND  4
+#define SDMA_SUBOP_COPY_TILED  1
+#define SDMA_SUBOP_COPY_TILED_SUB_WIND  5
+#define SDMA_SUBOP_COPY_T2T_SUB_WIND  6
+#define SDMA_SUBOP_COPY_SOA  3
+#define SDMA_SUBOP_COPY_DIRTY_PAGE  7
+#define SDMA_SUBOP_COPY_LINEAR_PHY  8
+#define SDMA_SUBOP_COPY_LINEAR_SUB_WIND_LARGE  36
+#define SDMA_SUBOP_COPY_LINEAR_BC  16
+#define SDMA_SUBOP_COPY_TILED_BC  17
+#define SDMA_SUBOP_COPY_LINEAR_SUB_WIND_BC  20
+#define SDMA_SUBOP_COPY_TILED_SUB_WIND_BC  21
+#define SDMA_SUBOP_COPY_T2T_SUB_WIND_BC  22
+#define SDMA_SUBOP_WRITE_LINEAR  0
+#define SDMA_SUBOP_WRITE_TILED  1
+#define SDMA_SUBOP_WRITE_TILED_BC  17
+#define SDMA_SUBOP_PTEPDE_GEN  0
+#define SDMA_SUBOP_PTEPDE_COPY  1
+#define SDMA_SUBOP_PTEPDE_RMW  2
+#define SDMA_SUBOP_PTEPDE_COPY_BACKWARDS  3
+#define SDMA_SUBOP_MEM_INCR  1
+#define SDMA_SUBOP_DATA_FILL_MULTI  1
+#define SDMA_SUBOP_POLL_REG_WRITE_MEM  1
+#define SDMA_SUBOP_POLL_DBIT_WRITE_MEM  2
+#define SDMA_SUBOP_POLL_MEM_VERIFY  3
+#define SDMA_SUBOP_VM_INVALIDATION  4
+#define HEADER_AGENT_DISPATCH  4
+#define HEADER_BARRIER  5
+#define SDMA_OP_AQL_COPY  0
+#define SDMA_OP_AQL_BARRIER_OR  0
+
+#define SDMA_GCR_RANGE_IS_PA		(1 << 18)
+#define SDMA_GCR_SEQ(x)			(((x) & 0x3) << 16)
+#define SDMA_GCR_GL2_WB			(1 << 15)
+#define SDMA_GCR_GL2_INV		(1 << 14)
+#define SDMA_GCR_GL2_DISCARD		(1 << 13)
+#define SDMA_GCR_GL2_RANGE(x)		(((x) & 0x3) << 11)
+#define SDMA_GCR_GL2_US			(1 << 10)
+#define SDMA_GCR_GL1_INV		(1 << 9)
+#define SDMA_GCR_GLV_INV		(1 << 8)
+#define SDMA_GCR_GLK_INV		(1 << 7)
+#define SDMA_GCR_GLK_WB			(1 << 6)
+#define SDMA_GCR_GLM_INV		(1 << 5)
+#define SDMA_GCR_GLM_WB			(1 << 4)
+#define SDMA_GCR_GL1_RANGE(x)		(((x) & 0x3) << 2)
+#define SDMA_GCR_GLI_INV(x)		(((x) & 0x3) << 0)
+
+#define SDMA_DCC_DATA_FORMAT(x) ((x) & 0x3f)
+#define SDMA_DCC_NUM_TYPE(x) (((x) & 0x7) << 9)
+#define SDMA_DCC_READ_CM(x) (((x) & 0x3) << 16)
+#define SDMA_DCC_WRITE_CM(x) (((x) & 0x3) << 18)
+#define SDMA_DCC_MAX_COM(x) (((x) & 0x3) << 24)
+#define SDMA_DCC_MAX_UCOM(x) (((x) & 0x1) << 26)
+
+/*
+** Definitions for SDMA_PKT_COPY_LINEAR packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_COPY_LINEAR_HEADER_op_offset 0
+#define SDMA_PKT_COPY_LINEAR_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_COPY_LINEAR_HEADER_op_shift  0
+#define SDMA_PKT_COPY_LINEAR_HEADER_OP(x) (((x) & SDMA_PKT_COPY_LINEAR_HEADER_op_mask) << SDMA_PKT_COPY_LINEAR_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_COPY_LINEAR_HEADER_sub_op_offset 0
+#define SDMA_PKT_COPY_LINEAR_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_COPY_LINEAR_HEADER_sub_op_shift  8
+#define SDMA_PKT_COPY_LINEAR_HEADER_SUB_OP(x) (((x) & SDMA_PKT_COPY_LINEAR_HEADER_sub_op_mask) << SDMA_PKT_COPY_LINEAR_HEADER_sub_op_shift)
+
+/*define for encrypt field*/
+#define SDMA_PKT_COPY_LINEAR_HEADER_encrypt_offset 0
+#define SDMA_PKT_COPY_LINEAR_HEADER_encrypt_mask   0x00000001
+#define SDMA_PKT_COPY_LINEAR_HEADER_encrypt_shift  16
+#define SDMA_PKT_COPY_LINEAR_HEADER_ENCRYPT(x) (((x) & SDMA_PKT_COPY_LINEAR_HEADER_encrypt_mask) << SDMA_PKT_COPY_LINEAR_HEADER_encrypt_shift)
+
+/*define for tmz field*/
+#define SDMA_PKT_COPY_LINEAR_HEADER_tmz_offset 0
+#define SDMA_PKT_COPY_LINEAR_HEADER_tmz_mask   0x00000001
+#define SDMA_PKT_COPY_LINEAR_HEADER_tmz_shift  18
+#define SDMA_PKT_COPY_LINEAR_HEADER_TMZ(x) (((x) & SDMA_PKT_COPY_LINEAR_HEADER_tmz_mask) << SDMA_PKT_COPY_LINEAR_HEADER_tmz_shift)
+
+/*define for cpv field*/
+#define SDMA_PKT_COPY_LINEAR_HEADER_cpv_offset 0
+#define SDMA_PKT_COPY_LINEAR_HEADER_cpv_mask   0x00000001
+#define SDMA_PKT_COPY_LINEAR_HEADER_cpv_shift  19
+#define SDMA_PKT_COPY_LINEAR_HEADER_CPV(x) (((x) & SDMA_PKT_COPY_LINEAR_HEADER_cpv_mask) << SDMA_PKT_COPY_LINEAR_HEADER_cpv_shift)
+
+/*define for backwards field*/
+#define SDMA_PKT_COPY_LINEAR_HEADER_backwards_offset 0
+#define SDMA_PKT_COPY_LINEAR_HEADER_backwards_mask   0x00000001
+#define SDMA_PKT_COPY_LINEAR_HEADER_backwards_shift  25
+#define SDMA_PKT_COPY_LINEAR_HEADER_BACKWARDS(x) (((x) & SDMA_PKT_COPY_LINEAR_HEADER_backwards_mask) << SDMA_PKT_COPY_LINEAR_HEADER_backwards_shift)
+
+/*define for broadcast field*/
+#define SDMA_PKT_COPY_LINEAR_HEADER_broadcast_offset 0
+#define SDMA_PKT_COPY_LINEAR_HEADER_broadcast_mask   0x00000001
+#define SDMA_PKT_COPY_LINEAR_HEADER_broadcast_shift  27
+#define SDMA_PKT_COPY_LINEAR_HEADER_BROADCAST(x) (((x) & SDMA_PKT_COPY_LINEAR_HEADER_broadcast_mask) << SDMA_PKT_COPY_LINEAR_HEADER_broadcast_shift)
+
+/*define for COUNT word*/
+/*define for count field*/
+#define SDMA_PKT_COPY_LINEAR_COUNT_count_offset 1
+#define SDMA_PKT_COPY_LINEAR_COUNT_count_mask   0x3FFFFFFF
+#define SDMA_PKT_COPY_LINEAR_COUNT_count_shift  0
+#define SDMA_PKT_COPY_LINEAR_COUNT_COUNT(x) (((x) & SDMA_PKT_COPY_LINEAR_COUNT_count_mask) << SDMA_PKT_COPY_LINEAR_COUNT_count_shift)
+
+/*define for PARAMETER word*/
+/*define for dst_sw field*/
+#define SDMA_PKT_COPY_LINEAR_PARAMETER_dst_sw_offset 2
+#define SDMA_PKT_COPY_LINEAR_PARAMETER_dst_sw_mask   0x00000003
+#define SDMA_PKT_COPY_LINEAR_PARAMETER_dst_sw_shift  16
+#define SDMA_PKT_COPY_LINEAR_PARAMETER_DST_SW(x) (((x) & SDMA_PKT_COPY_LINEAR_PARAMETER_dst_sw_mask) << SDMA_PKT_COPY_LINEAR_PARAMETER_dst_sw_shift)
+
+/*define for dst_cache_policy field*/
+#define SDMA_PKT_COPY_LINEAR_PARAMETER_dst_cache_policy_offset 2
+#define SDMA_PKT_COPY_LINEAR_PARAMETER_dst_cache_policy_mask   0x00000007
+#define SDMA_PKT_COPY_LINEAR_PARAMETER_dst_cache_policy_shift  18
+#define SDMA_PKT_COPY_LINEAR_PARAMETER_DST_CACHE_POLICY(x) (((x) & SDMA_PKT_COPY_LINEAR_PARAMETER_dst_cache_policy_mask) << SDMA_PKT_COPY_LINEAR_PARAMETER_dst_cache_policy_shift)
+
+/*define for src_sw field*/
+#define SDMA_PKT_COPY_LINEAR_PARAMETER_src_sw_offset 2
+#define SDMA_PKT_COPY_LINEAR_PARAMETER_src_sw_mask   0x00000003
+#define SDMA_PKT_COPY_LINEAR_PARAMETER_src_sw_shift  24
+#define SDMA_PKT_COPY_LINEAR_PARAMETER_SRC_SW(x) (((x) & SDMA_PKT_COPY_LINEAR_PARAMETER_src_sw_mask) << SDMA_PKT_COPY_LINEAR_PARAMETER_src_sw_shift)
+
+/*define for src_cache_policy field*/
+#define SDMA_PKT_COPY_LINEAR_PARAMETER_src_cache_policy_offset 2
+#define SDMA_PKT_COPY_LINEAR_PARAMETER_src_cache_policy_mask   0x00000007
+#define SDMA_PKT_COPY_LINEAR_PARAMETER_src_cache_policy_shift  26
+#define SDMA_PKT_COPY_LINEAR_PARAMETER_SRC_CACHE_POLICY(x) (((x) & SDMA_PKT_COPY_LINEAR_PARAMETER_src_cache_policy_mask) << SDMA_PKT_COPY_LINEAR_PARAMETER_src_cache_policy_shift)
+
+/*define for SRC_ADDR_LO word*/
+/*define for src_addr_31_0 field*/
+#define SDMA_PKT_COPY_LINEAR_SRC_ADDR_LO_src_addr_31_0_offset 3
+#define SDMA_PKT_COPY_LINEAR_SRC_ADDR_LO_src_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_LINEAR_SRC_ADDR_LO_src_addr_31_0_shift  0
+#define SDMA_PKT_COPY_LINEAR_SRC_ADDR_LO_SRC_ADDR_31_0(x) (((x) & SDMA_PKT_COPY_LINEAR_SRC_ADDR_LO_src_addr_31_0_mask) << SDMA_PKT_COPY_LINEAR_SRC_ADDR_LO_src_addr_31_0_shift)
+
+/*define for SRC_ADDR_HI word*/
+/*define for src_addr_63_32 field*/
+#define SDMA_PKT_COPY_LINEAR_SRC_ADDR_HI_src_addr_63_32_offset 4
+#define SDMA_PKT_COPY_LINEAR_SRC_ADDR_HI_src_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_LINEAR_SRC_ADDR_HI_src_addr_63_32_shift  0
+#define SDMA_PKT_COPY_LINEAR_SRC_ADDR_HI_SRC_ADDR_63_32(x) (((x) & SDMA_PKT_COPY_LINEAR_SRC_ADDR_HI_src_addr_63_32_mask) << SDMA_PKT_COPY_LINEAR_SRC_ADDR_HI_src_addr_63_32_shift)
+
+/*define for DST_ADDR_LO word*/
+/*define for dst_addr_31_0 field*/
+#define SDMA_PKT_COPY_LINEAR_DST_ADDR_LO_dst_addr_31_0_offset 5
+#define SDMA_PKT_COPY_LINEAR_DST_ADDR_LO_dst_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_LINEAR_DST_ADDR_LO_dst_addr_31_0_shift  0
+#define SDMA_PKT_COPY_LINEAR_DST_ADDR_LO_DST_ADDR_31_0(x) (((x) & SDMA_PKT_COPY_LINEAR_DST_ADDR_LO_dst_addr_31_0_mask) << SDMA_PKT_COPY_LINEAR_DST_ADDR_LO_dst_addr_31_0_shift)
+
+/*define for DST_ADDR_HI word*/
+/*define for dst_addr_63_32 field*/
+#define SDMA_PKT_COPY_LINEAR_DST_ADDR_HI_dst_addr_63_32_offset 6
+#define SDMA_PKT_COPY_LINEAR_DST_ADDR_HI_dst_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_LINEAR_DST_ADDR_HI_dst_addr_63_32_shift  0
+#define SDMA_PKT_COPY_LINEAR_DST_ADDR_HI_DST_ADDR_63_32(x) (((x) & SDMA_PKT_COPY_LINEAR_DST_ADDR_HI_dst_addr_63_32_mask) << SDMA_PKT_COPY_LINEAR_DST_ADDR_HI_dst_addr_63_32_shift)
+
+
+/*
+** Definitions for SDMA_PKT_COPY_LINEAR_BC packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_COPY_LINEAR_BC_HEADER_op_offset 0
+#define SDMA_PKT_COPY_LINEAR_BC_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_COPY_LINEAR_BC_HEADER_op_shift  0
+#define SDMA_PKT_COPY_LINEAR_BC_HEADER_OP(x) (((x) & SDMA_PKT_COPY_LINEAR_BC_HEADER_op_mask) << SDMA_PKT_COPY_LINEAR_BC_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_COPY_LINEAR_BC_HEADER_sub_op_offset 0
+#define SDMA_PKT_COPY_LINEAR_BC_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_COPY_LINEAR_BC_HEADER_sub_op_shift  8
+#define SDMA_PKT_COPY_LINEAR_BC_HEADER_SUB_OP(x) (((x) & SDMA_PKT_COPY_LINEAR_BC_HEADER_sub_op_mask) << SDMA_PKT_COPY_LINEAR_BC_HEADER_sub_op_shift)
+
+/*define for COUNT word*/
+/*define for count field*/
+#define SDMA_PKT_COPY_LINEAR_BC_COUNT_count_offset 1
+#define SDMA_PKT_COPY_LINEAR_BC_COUNT_count_mask   0x003FFFFF
+#define SDMA_PKT_COPY_LINEAR_BC_COUNT_count_shift  0
+#define SDMA_PKT_COPY_LINEAR_BC_COUNT_COUNT(x) (((x) & SDMA_PKT_COPY_LINEAR_BC_COUNT_count_mask) << SDMA_PKT_COPY_LINEAR_BC_COUNT_count_shift)
+
+/*define for PARAMETER word*/
+/*define for dst_sw field*/
+#define SDMA_PKT_COPY_LINEAR_BC_PARAMETER_dst_sw_offset 2
+#define SDMA_PKT_COPY_LINEAR_BC_PARAMETER_dst_sw_mask   0x00000003
+#define SDMA_PKT_COPY_LINEAR_BC_PARAMETER_dst_sw_shift  16
+#define SDMA_PKT_COPY_LINEAR_BC_PARAMETER_DST_SW(x) (((x) & SDMA_PKT_COPY_LINEAR_BC_PARAMETER_dst_sw_mask) << SDMA_PKT_COPY_LINEAR_BC_PARAMETER_dst_sw_shift)
+
+/*define for dst_ha field*/
+#define SDMA_PKT_COPY_LINEAR_BC_PARAMETER_dst_ha_offset 2
+#define SDMA_PKT_COPY_LINEAR_BC_PARAMETER_dst_ha_mask   0x00000001
+#define SDMA_PKT_COPY_LINEAR_BC_PARAMETER_dst_ha_shift  19
+#define SDMA_PKT_COPY_LINEAR_BC_PARAMETER_DST_HA(x) (((x) & SDMA_PKT_COPY_LINEAR_BC_PARAMETER_dst_ha_mask) << SDMA_PKT_COPY_LINEAR_BC_PARAMETER_dst_ha_shift)
+
+/*define for src_sw field*/
+#define SDMA_PKT_COPY_LINEAR_BC_PARAMETER_src_sw_offset 2
+#define SDMA_PKT_COPY_LINEAR_BC_PARAMETER_src_sw_mask   0x00000003
+#define SDMA_PKT_COPY_LINEAR_BC_PARAMETER_src_sw_shift  24
+#define SDMA_PKT_COPY_LINEAR_BC_PARAMETER_SRC_SW(x) (((x) & SDMA_PKT_COPY_LINEAR_BC_PARAMETER_src_sw_mask) << SDMA_PKT_COPY_LINEAR_BC_PARAMETER_src_sw_shift)
+
+/*define for src_ha field*/
+#define SDMA_PKT_COPY_LINEAR_BC_PARAMETER_src_ha_offset 2
+#define SDMA_PKT_COPY_LINEAR_BC_PARAMETER_src_ha_mask   0x00000001
+#define SDMA_PKT_COPY_LINEAR_BC_PARAMETER_src_ha_shift  27
+#define SDMA_PKT_COPY_LINEAR_BC_PARAMETER_SRC_HA(x) (((x) & SDMA_PKT_COPY_LINEAR_BC_PARAMETER_src_ha_mask) << SDMA_PKT_COPY_LINEAR_BC_PARAMETER_src_ha_shift)
+
+/*define for SRC_ADDR_LO word*/
+/*define for src_addr_31_0 field*/
+#define SDMA_PKT_COPY_LINEAR_BC_SRC_ADDR_LO_src_addr_31_0_offset 3
+#define SDMA_PKT_COPY_LINEAR_BC_SRC_ADDR_LO_src_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_LINEAR_BC_SRC_ADDR_LO_src_addr_31_0_shift  0
+#define SDMA_PKT_COPY_LINEAR_BC_SRC_ADDR_LO_SRC_ADDR_31_0(x) (((x) & SDMA_PKT_COPY_LINEAR_BC_SRC_ADDR_LO_src_addr_31_0_mask) << SDMA_PKT_COPY_LINEAR_BC_SRC_ADDR_LO_src_addr_31_0_shift)
+
+/*define for SRC_ADDR_HI word*/
+/*define for src_addr_63_32 field*/
+#define SDMA_PKT_COPY_LINEAR_BC_SRC_ADDR_HI_src_addr_63_32_offset 4
+#define SDMA_PKT_COPY_LINEAR_BC_SRC_ADDR_HI_src_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_LINEAR_BC_SRC_ADDR_HI_src_addr_63_32_shift  0
+#define SDMA_PKT_COPY_LINEAR_BC_SRC_ADDR_HI_SRC_ADDR_63_32(x) (((x) & SDMA_PKT_COPY_LINEAR_BC_SRC_ADDR_HI_src_addr_63_32_mask) << SDMA_PKT_COPY_LINEAR_BC_SRC_ADDR_HI_src_addr_63_32_shift)
+
+/*define for DST_ADDR_LO word*/
+/*define for dst_addr_31_0 field*/
+#define SDMA_PKT_COPY_LINEAR_BC_DST_ADDR_LO_dst_addr_31_0_offset 5
+#define SDMA_PKT_COPY_LINEAR_BC_DST_ADDR_LO_dst_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_LINEAR_BC_DST_ADDR_LO_dst_addr_31_0_shift  0
+#define SDMA_PKT_COPY_LINEAR_BC_DST_ADDR_LO_DST_ADDR_31_0(x) (((x) & SDMA_PKT_COPY_LINEAR_BC_DST_ADDR_LO_dst_addr_31_0_mask) << SDMA_PKT_COPY_LINEAR_BC_DST_ADDR_LO_dst_addr_31_0_shift)
+
+/*define for DST_ADDR_HI word*/
+/*define for dst_addr_63_32 field*/
+#define SDMA_PKT_COPY_LINEAR_BC_DST_ADDR_HI_dst_addr_63_32_offset 6
+#define SDMA_PKT_COPY_LINEAR_BC_DST_ADDR_HI_dst_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_LINEAR_BC_DST_ADDR_HI_dst_addr_63_32_shift  0
+#define SDMA_PKT_COPY_LINEAR_BC_DST_ADDR_HI_DST_ADDR_63_32(x) (((x) & SDMA_PKT_COPY_LINEAR_BC_DST_ADDR_HI_dst_addr_63_32_mask) << SDMA_PKT_COPY_LINEAR_BC_DST_ADDR_HI_dst_addr_63_32_shift)
+
+
+/*
+** Definitions for SDMA_PKT_COPY_DIRTY_PAGE packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_COPY_DIRTY_PAGE_HEADER_op_offset 0
+#define SDMA_PKT_COPY_DIRTY_PAGE_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_COPY_DIRTY_PAGE_HEADER_op_shift  0
+#define SDMA_PKT_COPY_DIRTY_PAGE_HEADER_OP(x) (((x) & SDMA_PKT_COPY_DIRTY_PAGE_HEADER_op_mask) << SDMA_PKT_COPY_DIRTY_PAGE_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_COPY_DIRTY_PAGE_HEADER_sub_op_offset 0
+#define SDMA_PKT_COPY_DIRTY_PAGE_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_COPY_DIRTY_PAGE_HEADER_sub_op_shift  8
+#define SDMA_PKT_COPY_DIRTY_PAGE_HEADER_SUB_OP(x) (((x) & SDMA_PKT_COPY_DIRTY_PAGE_HEADER_sub_op_mask) << SDMA_PKT_COPY_DIRTY_PAGE_HEADER_sub_op_shift)
+
+/*define for tmz field*/
+#define SDMA_PKT_COPY_DIRTY_PAGE_HEADER_tmz_offset 0
+#define SDMA_PKT_COPY_DIRTY_PAGE_HEADER_tmz_mask   0x00000001
+#define SDMA_PKT_COPY_DIRTY_PAGE_HEADER_tmz_shift  18
+#define SDMA_PKT_COPY_DIRTY_PAGE_HEADER_TMZ(x) (((x) & SDMA_PKT_COPY_DIRTY_PAGE_HEADER_tmz_mask) << SDMA_PKT_COPY_DIRTY_PAGE_HEADER_tmz_shift)
+
+/*define for cpv field*/
+#define SDMA_PKT_COPY_DIRTY_PAGE_HEADER_cpv_offset 0
+#define SDMA_PKT_COPY_DIRTY_PAGE_HEADER_cpv_mask   0x00000001
+#define SDMA_PKT_COPY_DIRTY_PAGE_HEADER_cpv_shift  19
+#define SDMA_PKT_COPY_DIRTY_PAGE_HEADER_CPV(x) (((x) & SDMA_PKT_COPY_DIRTY_PAGE_HEADER_cpv_mask) << SDMA_PKT_COPY_DIRTY_PAGE_HEADER_cpv_shift)
+
+/*define for all field*/
+#define SDMA_PKT_COPY_DIRTY_PAGE_HEADER_all_offset 0
+#define SDMA_PKT_COPY_DIRTY_PAGE_HEADER_all_mask   0x00000001
+#define SDMA_PKT_COPY_DIRTY_PAGE_HEADER_all_shift  31
+#define SDMA_PKT_COPY_DIRTY_PAGE_HEADER_ALL(x) (((x) & SDMA_PKT_COPY_DIRTY_PAGE_HEADER_all_mask) << SDMA_PKT_COPY_DIRTY_PAGE_HEADER_all_shift)
+
+/*define for COUNT word*/
+/*define for count field*/
+#define SDMA_PKT_COPY_DIRTY_PAGE_COUNT_count_offset 1
+#define SDMA_PKT_COPY_DIRTY_PAGE_COUNT_count_mask   0x003FFFFF
+#define SDMA_PKT_COPY_DIRTY_PAGE_COUNT_count_shift  0
+#define SDMA_PKT_COPY_DIRTY_PAGE_COUNT_COUNT(x) (((x) & SDMA_PKT_COPY_DIRTY_PAGE_COUNT_count_mask) << SDMA_PKT_COPY_DIRTY_PAGE_COUNT_count_shift)
+
+/*define for PARAMETER word*/
+/*define for dst_mtype field*/
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_dst_mtype_offset 2
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_dst_mtype_mask   0x00000007
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_dst_mtype_shift  3
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_DST_MTYPE(x) (((x) & SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_dst_mtype_mask) << SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_dst_mtype_shift)
+
+/*define for dst_l2_policy field*/
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_dst_l2_policy_offset 2
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_dst_l2_policy_mask   0x00000003
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_dst_l2_policy_shift  6
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_DST_L2_POLICY(x) (((x) & SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_dst_l2_policy_mask) << SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_dst_l2_policy_shift)
+
+/*define for dst_llc field*/
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_dst_llc_offset 2
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_dst_llc_mask   0x00000001
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_dst_llc_shift  8
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_DST_LLC(x) (((x) & SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_dst_llc_mask) << SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_dst_llc_shift)
+
+/*define for src_mtype field*/
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_src_mtype_offset 2
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_src_mtype_mask   0x00000007
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_src_mtype_shift  11
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_SRC_MTYPE(x) (((x) & SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_src_mtype_mask) << SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_src_mtype_shift)
+
+/*define for src_l2_policy field*/
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_src_l2_policy_offset 2
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_src_l2_policy_mask   0x00000003
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_src_l2_policy_shift  14
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_SRC_L2_POLICY(x) (((x) & SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_src_l2_policy_mask) << SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_src_l2_policy_shift)
+
+/*define for src_llc field*/
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_src_llc_offset 2
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_src_llc_mask   0x00000001
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_src_llc_shift  16
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_SRC_LLC(x) (((x) & SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_src_llc_mask) << SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_src_llc_shift)
+
+/*define for dst_sw field*/
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_dst_sw_offset 2
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_dst_sw_mask   0x00000003
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_dst_sw_shift  17
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_DST_SW(x) (((x) & SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_dst_sw_mask) << SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_dst_sw_shift)
+
+/*define for dst_gcc field*/
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_dst_gcc_offset 2
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_dst_gcc_mask   0x00000001
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_dst_gcc_shift  19
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_DST_GCC(x) (((x) & SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_dst_gcc_mask) << SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_dst_gcc_shift)
+
+/*define for dst_sys field*/
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_dst_sys_offset 2
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_dst_sys_mask   0x00000001
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_dst_sys_shift  20
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_DST_SYS(x) (((x) & SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_dst_sys_mask) << SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_dst_sys_shift)
+
+/*define for dst_snoop field*/
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_dst_snoop_offset 2
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_dst_snoop_mask   0x00000001
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_dst_snoop_shift  22
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_DST_SNOOP(x) (((x) & SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_dst_snoop_mask) << SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_dst_snoop_shift)
+
+/*define for dst_gpa field*/
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_dst_gpa_offset 2
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_dst_gpa_mask   0x00000001
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_dst_gpa_shift  23
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_DST_GPA(x) (((x) & SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_dst_gpa_mask) << SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_dst_gpa_shift)
+
+/*define for src_sw field*/
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_src_sw_offset 2
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_src_sw_mask   0x00000003
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_src_sw_shift  24
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_SRC_SW(x) (((x) & SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_src_sw_mask) << SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_src_sw_shift)
+
+/*define for src_sys field*/
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_src_sys_offset 2
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_src_sys_mask   0x00000001
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_src_sys_shift  28
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_SRC_SYS(x) (((x) & SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_src_sys_mask) << SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_src_sys_shift)
+
+/*define for src_snoop field*/
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_src_snoop_offset 2
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_src_snoop_mask   0x00000001
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_src_snoop_shift  30
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_SRC_SNOOP(x) (((x) & SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_src_snoop_mask) << SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_src_snoop_shift)
+
+/*define for src_gpa field*/
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_src_gpa_offset 2
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_src_gpa_mask   0x00000001
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_src_gpa_shift  31
+#define SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_SRC_GPA(x) (((x) & SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_src_gpa_mask) << SDMA_PKT_COPY_DIRTY_PAGE_PARAMETER_src_gpa_shift)
+
+/*define for SRC_ADDR_LO word*/
+/*define for src_addr_31_0 field*/
+#define SDMA_PKT_COPY_DIRTY_PAGE_SRC_ADDR_LO_src_addr_31_0_offset 3
+#define SDMA_PKT_COPY_DIRTY_PAGE_SRC_ADDR_LO_src_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_DIRTY_PAGE_SRC_ADDR_LO_src_addr_31_0_shift  0
+#define SDMA_PKT_COPY_DIRTY_PAGE_SRC_ADDR_LO_SRC_ADDR_31_0(x) (((x) & SDMA_PKT_COPY_DIRTY_PAGE_SRC_ADDR_LO_src_addr_31_0_mask) << SDMA_PKT_COPY_DIRTY_PAGE_SRC_ADDR_LO_src_addr_31_0_shift)
+
+/*define for SRC_ADDR_HI word*/
+/*define for src_addr_63_32 field*/
+#define SDMA_PKT_COPY_DIRTY_PAGE_SRC_ADDR_HI_src_addr_63_32_offset 4
+#define SDMA_PKT_COPY_DIRTY_PAGE_SRC_ADDR_HI_src_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_DIRTY_PAGE_SRC_ADDR_HI_src_addr_63_32_shift  0
+#define SDMA_PKT_COPY_DIRTY_PAGE_SRC_ADDR_HI_SRC_ADDR_63_32(x) (((x) & SDMA_PKT_COPY_DIRTY_PAGE_SRC_ADDR_HI_src_addr_63_32_mask) << SDMA_PKT_COPY_DIRTY_PAGE_SRC_ADDR_HI_src_addr_63_32_shift)
+
+/*define for DST_ADDR_LO word*/
+/*define for dst_addr_31_0 field*/
+#define SDMA_PKT_COPY_DIRTY_PAGE_DST_ADDR_LO_dst_addr_31_0_offset 5
+#define SDMA_PKT_COPY_DIRTY_PAGE_DST_ADDR_LO_dst_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_DIRTY_PAGE_DST_ADDR_LO_dst_addr_31_0_shift  0
+#define SDMA_PKT_COPY_DIRTY_PAGE_DST_ADDR_LO_DST_ADDR_31_0(x) (((x) & SDMA_PKT_COPY_DIRTY_PAGE_DST_ADDR_LO_dst_addr_31_0_mask) << SDMA_PKT_COPY_DIRTY_PAGE_DST_ADDR_LO_dst_addr_31_0_shift)
+
+/*define for DST_ADDR_HI word*/
+/*define for dst_addr_63_32 field*/
+#define SDMA_PKT_COPY_DIRTY_PAGE_DST_ADDR_HI_dst_addr_63_32_offset 6
+#define SDMA_PKT_COPY_DIRTY_PAGE_DST_ADDR_HI_dst_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_DIRTY_PAGE_DST_ADDR_HI_dst_addr_63_32_shift  0
+#define SDMA_PKT_COPY_DIRTY_PAGE_DST_ADDR_HI_DST_ADDR_63_32(x) (((x) & SDMA_PKT_COPY_DIRTY_PAGE_DST_ADDR_HI_dst_addr_63_32_mask) << SDMA_PKT_COPY_DIRTY_PAGE_DST_ADDR_HI_dst_addr_63_32_shift)
+
+
+/*
+** Definitions for SDMA_PKT_COPY_PHYSICAL_LINEAR packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_HEADER_op_offset 0
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_HEADER_op_shift  0
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_HEADER_OP(x) (((x) & SDMA_PKT_COPY_PHYSICAL_LINEAR_HEADER_op_mask) << SDMA_PKT_COPY_PHYSICAL_LINEAR_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_HEADER_sub_op_offset 0
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_HEADER_sub_op_shift  8
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_HEADER_SUB_OP(x) (((x) & SDMA_PKT_COPY_PHYSICAL_LINEAR_HEADER_sub_op_mask) << SDMA_PKT_COPY_PHYSICAL_LINEAR_HEADER_sub_op_shift)
+
+/*define for tmz field*/
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_HEADER_tmz_offset 0
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_HEADER_tmz_mask   0x00000001
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_HEADER_tmz_shift  18
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_HEADER_TMZ(x) (((x) & SDMA_PKT_COPY_PHYSICAL_LINEAR_HEADER_tmz_mask) << SDMA_PKT_COPY_PHYSICAL_LINEAR_HEADER_tmz_shift)
+
+/*define for cpv field*/
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_HEADER_cpv_offset 0
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_HEADER_cpv_mask   0x00000001
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_HEADER_cpv_shift  19
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_HEADER_CPV(x) (((x) & SDMA_PKT_COPY_PHYSICAL_LINEAR_HEADER_cpv_mask) << SDMA_PKT_COPY_PHYSICAL_LINEAR_HEADER_cpv_shift)
+
+/*define for COUNT word*/
+/*define for count field*/
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_COUNT_count_offset 1
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_COUNT_count_mask   0x003FFFFF
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_COUNT_count_shift  0
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_COUNT_COUNT(x) (((x) & SDMA_PKT_COPY_PHYSICAL_LINEAR_COUNT_count_mask) << SDMA_PKT_COPY_PHYSICAL_LINEAR_COUNT_count_shift)
+
+/*define for addr_pair_num field*/
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_COUNT_addr_pair_num_offset 1
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_COUNT_addr_pair_num_mask   0x000000FF
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_COUNT_addr_pair_num_shift  24
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_COUNT_ADDR_PAIR_NUM(x) (((x) & SDMA_PKT_COPY_PHYSICAL_LINEAR_COUNT_addr_pair_num_mask) << SDMA_PKT_COPY_PHYSICAL_LINEAR_COUNT_addr_pair_num_shift)
+
+/*define for PARAMETER word*/
+/*define for dst_mtype field*/
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_mtype_offset 2
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_mtype_mask   0x00000007
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_mtype_shift  3
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_DST_MTYPE(x) (((x) & SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_mtype_mask) << SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_mtype_shift)
+
+/*define for dst_l2_policy field*/
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_l2_policy_offset 2
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_l2_policy_mask   0x00000003
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_l2_policy_shift  6
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_DST_L2_POLICY(x) (((x) & SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_l2_policy_mask) << SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_l2_policy_shift)
+
+/*define for dst_llc field*/
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_llc_offset 2
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_llc_mask   0x00000001
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_llc_shift  8
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_DST_LLC(x) (((x) & SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_llc_mask) << SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_llc_shift)
+
+/*define for src_mtype field*/
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_src_mtype_offset 2
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_src_mtype_mask   0x00000007
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_src_mtype_shift  11
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_SRC_MTYPE(x) (((x) & SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_src_mtype_mask) << SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_src_mtype_shift)
+
+/*define for src_l2_policy field*/
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_src_l2_policy_offset 2
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_src_l2_policy_mask   0x00000003
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_src_l2_policy_shift  14
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_SRC_L2_POLICY(x) (((x) & SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_src_l2_policy_mask) << SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_src_l2_policy_shift)
+
+/*define for src_llc field*/
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_src_llc_offset 2
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_src_llc_mask   0x00000001
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_src_llc_shift  16
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_SRC_LLC(x) (((x) & SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_src_llc_mask) << SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_src_llc_shift)
+
+/*define for dst_sw field*/
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_sw_offset 2
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_sw_mask   0x00000003
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_sw_shift  17
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_DST_SW(x) (((x) & SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_sw_mask) << SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_sw_shift)
+
+/*define for dst_gcc field*/
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_gcc_offset 2
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_gcc_mask   0x00000001
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_gcc_shift  19
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_DST_GCC(x) (((x) & SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_gcc_mask) << SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_gcc_shift)
+
+/*define for dst_sys field*/
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_sys_offset 2
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_sys_mask   0x00000001
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_sys_shift  20
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_DST_SYS(x) (((x) & SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_sys_mask) << SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_sys_shift)
+
+/*define for dst_log field*/
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_log_offset 2
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_log_mask   0x00000001
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_log_shift  21
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_DST_LOG(x) (((x) & SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_log_mask) << SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_log_shift)
+
+/*define for dst_snoop field*/
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_snoop_offset 2
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_snoop_mask   0x00000001
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_snoop_shift  22
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_DST_SNOOP(x) (((x) & SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_snoop_mask) << SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_snoop_shift)
+
+/*define for dst_gpa field*/
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_gpa_offset 2
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_gpa_mask   0x00000001
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_gpa_shift  23
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_DST_GPA(x) (((x) & SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_gpa_mask) << SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_dst_gpa_shift)
+
+/*define for src_sw field*/
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_src_sw_offset 2
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_src_sw_mask   0x00000003
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_src_sw_shift  24
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_SRC_SW(x) (((x) & SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_src_sw_mask) << SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_src_sw_shift)
+
+/*define for src_gcc field*/
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_src_gcc_offset 2
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_src_gcc_mask   0x00000001
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_src_gcc_shift  27
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_SRC_GCC(x) (((x) & SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_src_gcc_mask) << SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_src_gcc_shift)
+
+/*define for src_sys field*/
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_src_sys_offset 2
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_src_sys_mask   0x00000001
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_src_sys_shift  28
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_SRC_SYS(x) (((x) & SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_src_sys_mask) << SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_src_sys_shift)
+
+/*define for src_snoop field*/
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_src_snoop_offset 2
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_src_snoop_mask   0x00000001
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_src_snoop_shift  30
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_SRC_SNOOP(x) (((x) & SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_src_snoop_mask) << SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_src_snoop_shift)
+
+/*define for src_gpa field*/
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_src_gpa_offset 2
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_src_gpa_mask   0x00000001
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_src_gpa_shift  31
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_SRC_GPA(x) (((x) & SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_src_gpa_mask) << SDMA_PKT_COPY_PHYSICAL_LINEAR_PARAMETER_src_gpa_shift)
+
+/*define for SRC_ADDR_LO word*/
+/*define for src_addr_31_0 field*/
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_SRC_ADDR_LO_src_addr_31_0_offset 3
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_SRC_ADDR_LO_src_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_SRC_ADDR_LO_src_addr_31_0_shift  0
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_SRC_ADDR_LO_SRC_ADDR_31_0(x) (((x) & SDMA_PKT_COPY_PHYSICAL_LINEAR_SRC_ADDR_LO_src_addr_31_0_mask) << SDMA_PKT_COPY_PHYSICAL_LINEAR_SRC_ADDR_LO_src_addr_31_0_shift)
+
+/*define for SRC_ADDR_HI word*/
+/*define for src_addr_63_32 field*/
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_SRC_ADDR_HI_src_addr_63_32_offset 4
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_SRC_ADDR_HI_src_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_SRC_ADDR_HI_src_addr_63_32_shift  0
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_SRC_ADDR_HI_SRC_ADDR_63_32(x) (((x) & SDMA_PKT_COPY_PHYSICAL_LINEAR_SRC_ADDR_HI_src_addr_63_32_mask) << SDMA_PKT_COPY_PHYSICAL_LINEAR_SRC_ADDR_HI_src_addr_63_32_shift)
+
+/*define for DST_ADDR_LO word*/
+/*define for dst_addr_31_0 field*/
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_DST_ADDR_LO_dst_addr_31_0_offset 5
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_DST_ADDR_LO_dst_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_DST_ADDR_LO_dst_addr_31_0_shift  0
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_DST_ADDR_LO_DST_ADDR_31_0(x) (((x) & SDMA_PKT_COPY_PHYSICAL_LINEAR_DST_ADDR_LO_dst_addr_31_0_mask) << SDMA_PKT_COPY_PHYSICAL_LINEAR_DST_ADDR_LO_dst_addr_31_0_shift)
+
+/*define for DST_ADDR_HI word*/
+/*define for dst_addr_63_32 field*/
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_DST_ADDR_HI_dst_addr_63_32_offset 6
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_DST_ADDR_HI_dst_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_DST_ADDR_HI_dst_addr_63_32_shift  0
+#define SDMA_PKT_COPY_PHYSICAL_LINEAR_DST_ADDR_HI_DST_ADDR_63_32(x) (((x) & SDMA_PKT_COPY_PHYSICAL_LINEAR_DST_ADDR_HI_dst_addr_63_32_mask) << SDMA_PKT_COPY_PHYSICAL_LINEAR_DST_ADDR_HI_dst_addr_63_32_shift)
+
+
+/*
+** Definitions for SDMA_PKT_COPY_BROADCAST_LINEAR packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_HEADER_op_offset 0
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_HEADER_op_shift  0
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_HEADER_OP(x) (((x) & SDMA_PKT_COPY_BROADCAST_LINEAR_HEADER_op_mask) << SDMA_PKT_COPY_BROADCAST_LINEAR_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_HEADER_sub_op_offset 0
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_HEADER_sub_op_shift  8
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_HEADER_SUB_OP(x) (((x) & SDMA_PKT_COPY_BROADCAST_LINEAR_HEADER_sub_op_mask) << SDMA_PKT_COPY_BROADCAST_LINEAR_HEADER_sub_op_shift)
+
+/*define for encrypt field*/
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_HEADER_encrypt_offset 0
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_HEADER_encrypt_mask   0x00000001
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_HEADER_encrypt_shift  16
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_HEADER_ENCRYPT(x) (((x) & SDMA_PKT_COPY_BROADCAST_LINEAR_HEADER_encrypt_mask) << SDMA_PKT_COPY_BROADCAST_LINEAR_HEADER_encrypt_shift)
+
+/*define for tmz field*/
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_HEADER_tmz_offset 0
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_HEADER_tmz_mask   0x00000001
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_HEADER_tmz_shift  18
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_HEADER_TMZ(x) (((x) & SDMA_PKT_COPY_BROADCAST_LINEAR_HEADER_tmz_mask) << SDMA_PKT_COPY_BROADCAST_LINEAR_HEADER_tmz_shift)
+
+/*define for cpv field*/
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_HEADER_cpv_offset 0
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_HEADER_cpv_mask   0x00000001
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_HEADER_cpv_shift  19
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_HEADER_CPV(x) (((x) & SDMA_PKT_COPY_BROADCAST_LINEAR_HEADER_cpv_mask) << SDMA_PKT_COPY_BROADCAST_LINEAR_HEADER_cpv_shift)
+
+/*define for broadcast field*/
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_HEADER_broadcast_offset 0
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_HEADER_broadcast_mask   0x00000001
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_HEADER_broadcast_shift  27
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_HEADER_BROADCAST(x) (((x) & SDMA_PKT_COPY_BROADCAST_LINEAR_HEADER_broadcast_mask) << SDMA_PKT_COPY_BROADCAST_LINEAR_HEADER_broadcast_shift)
+
+/*define for COUNT word*/
+/*define for count field*/
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_COUNT_count_offset 1
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_COUNT_count_mask   0x3FFFFFFF
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_COUNT_count_shift  0
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_COUNT_COUNT(x) (((x) & SDMA_PKT_COPY_BROADCAST_LINEAR_COUNT_count_mask) << SDMA_PKT_COPY_BROADCAST_LINEAR_COUNT_count_shift)
+
+/*define for PARAMETER word*/
+/*define for dst2_sw field*/
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_PARAMETER_dst2_sw_offset 2
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_PARAMETER_dst2_sw_mask   0x00000003
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_PARAMETER_dst2_sw_shift  8
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_PARAMETER_DST2_SW(x) (((x) & SDMA_PKT_COPY_BROADCAST_LINEAR_PARAMETER_dst2_sw_mask) << SDMA_PKT_COPY_BROADCAST_LINEAR_PARAMETER_dst2_sw_shift)
+
+/*define for dst2_cache_policy field*/
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_PARAMETER_dst2_cache_policy_offset 2
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_PARAMETER_dst2_cache_policy_mask   0x00000007
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_PARAMETER_dst2_cache_policy_shift  10
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_PARAMETER_DST2_CACHE_POLICY(x) (((x) & SDMA_PKT_COPY_BROADCAST_LINEAR_PARAMETER_dst2_cache_policy_mask) << SDMA_PKT_COPY_BROADCAST_LINEAR_PARAMETER_dst2_cache_policy_shift)
+
+/*define for dst1_sw field*/
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_PARAMETER_dst1_sw_offset 2
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_PARAMETER_dst1_sw_mask   0x00000003
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_PARAMETER_dst1_sw_shift  16
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_PARAMETER_DST1_SW(x) (((x) & SDMA_PKT_COPY_BROADCAST_LINEAR_PARAMETER_dst1_sw_mask) << SDMA_PKT_COPY_BROADCAST_LINEAR_PARAMETER_dst1_sw_shift)
+
+/*define for dst1_cache_policy field*/
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_PARAMETER_dst1_cache_policy_offset 2
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_PARAMETER_dst1_cache_policy_mask   0x00000007
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_PARAMETER_dst1_cache_policy_shift  18
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_PARAMETER_DST1_CACHE_POLICY(x) (((x) & SDMA_PKT_COPY_BROADCAST_LINEAR_PARAMETER_dst1_cache_policy_mask) << SDMA_PKT_COPY_BROADCAST_LINEAR_PARAMETER_dst1_cache_policy_shift)
+
+/*define for src_sw field*/
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_PARAMETER_src_sw_offset 2
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_PARAMETER_src_sw_mask   0x00000003
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_PARAMETER_src_sw_shift  24
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_PARAMETER_SRC_SW(x) (((x) & SDMA_PKT_COPY_BROADCAST_LINEAR_PARAMETER_src_sw_mask) << SDMA_PKT_COPY_BROADCAST_LINEAR_PARAMETER_src_sw_shift)
+
+/*define for src_cache_policy field*/
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_PARAMETER_src_cache_policy_offset 2
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_PARAMETER_src_cache_policy_mask   0x00000007
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_PARAMETER_src_cache_policy_shift  26
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_PARAMETER_SRC_CACHE_POLICY(x) (((x) & SDMA_PKT_COPY_BROADCAST_LINEAR_PARAMETER_src_cache_policy_mask) << SDMA_PKT_COPY_BROADCAST_LINEAR_PARAMETER_src_cache_policy_shift)
+
+/*define for SRC_ADDR_LO word*/
+/*define for src_addr_31_0 field*/
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_SRC_ADDR_LO_src_addr_31_0_offset 3
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_SRC_ADDR_LO_src_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_SRC_ADDR_LO_src_addr_31_0_shift  0
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_SRC_ADDR_LO_SRC_ADDR_31_0(x) (((x) & SDMA_PKT_COPY_BROADCAST_LINEAR_SRC_ADDR_LO_src_addr_31_0_mask) << SDMA_PKT_COPY_BROADCAST_LINEAR_SRC_ADDR_LO_src_addr_31_0_shift)
+
+/*define for SRC_ADDR_HI word*/
+/*define for src_addr_63_32 field*/
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_SRC_ADDR_HI_src_addr_63_32_offset 4
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_SRC_ADDR_HI_src_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_SRC_ADDR_HI_src_addr_63_32_shift  0
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_SRC_ADDR_HI_SRC_ADDR_63_32(x) (((x) & SDMA_PKT_COPY_BROADCAST_LINEAR_SRC_ADDR_HI_src_addr_63_32_mask) << SDMA_PKT_COPY_BROADCAST_LINEAR_SRC_ADDR_HI_src_addr_63_32_shift)
+
+/*define for DST1_ADDR_LO word*/
+/*define for dst1_addr_31_0 field*/
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_DST1_ADDR_LO_dst1_addr_31_0_offset 5
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_DST1_ADDR_LO_dst1_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_DST1_ADDR_LO_dst1_addr_31_0_shift  0
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_DST1_ADDR_LO_DST1_ADDR_31_0(x) (((x) & SDMA_PKT_COPY_BROADCAST_LINEAR_DST1_ADDR_LO_dst1_addr_31_0_mask) << SDMA_PKT_COPY_BROADCAST_LINEAR_DST1_ADDR_LO_dst1_addr_31_0_shift)
+
+/*define for DST1_ADDR_HI word*/
+/*define for dst1_addr_63_32 field*/
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_DST1_ADDR_HI_dst1_addr_63_32_offset 6
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_DST1_ADDR_HI_dst1_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_DST1_ADDR_HI_dst1_addr_63_32_shift  0
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_DST1_ADDR_HI_DST1_ADDR_63_32(x) (((x) & SDMA_PKT_COPY_BROADCAST_LINEAR_DST1_ADDR_HI_dst1_addr_63_32_mask) << SDMA_PKT_COPY_BROADCAST_LINEAR_DST1_ADDR_HI_dst1_addr_63_32_shift)
+
+/*define for DST2_ADDR_LO word*/
+/*define for dst2_addr_31_0 field*/
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_DST2_ADDR_LO_dst2_addr_31_0_offset 7
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_DST2_ADDR_LO_dst2_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_DST2_ADDR_LO_dst2_addr_31_0_shift  0
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_DST2_ADDR_LO_DST2_ADDR_31_0(x) (((x) & SDMA_PKT_COPY_BROADCAST_LINEAR_DST2_ADDR_LO_dst2_addr_31_0_mask) << SDMA_PKT_COPY_BROADCAST_LINEAR_DST2_ADDR_LO_dst2_addr_31_0_shift)
+
+/*define for DST2_ADDR_HI word*/
+/*define for dst2_addr_63_32 field*/
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_DST2_ADDR_HI_dst2_addr_63_32_offset 8
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_DST2_ADDR_HI_dst2_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_DST2_ADDR_HI_dst2_addr_63_32_shift  0
+#define SDMA_PKT_COPY_BROADCAST_LINEAR_DST2_ADDR_HI_DST2_ADDR_63_32(x) (((x) & SDMA_PKT_COPY_BROADCAST_LINEAR_DST2_ADDR_HI_dst2_addr_63_32_mask) << SDMA_PKT_COPY_BROADCAST_LINEAR_DST2_ADDR_HI_dst2_addr_63_32_shift)
+
+
+/*
+** Definitions for SDMA_PKT_COPY_LINEAR_SUBWIN packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_HEADER_op_offset 0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_HEADER_op_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_HEADER_OP(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_HEADER_op_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_HEADER_sub_op_offset 0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_HEADER_sub_op_shift  8
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_HEADER_SUB_OP(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_HEADER_sub_op_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_HEADER_sub_op_shift)
+
+/*define for tmz field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_HEADER_tmz_offset 0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_HEADER_tmz_mask   0x00000001
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_HEADER_tmz_shift  18
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_HEADER_TMZ(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_HEADER_tmz_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_HEADER_tmz_shift)
+
+/*define for cpv field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_HEADER_cpv_offset 0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_HEADER_cpv_mask   0x00000001
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_HEADER_cpv_shift  19
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_HEADER_CPV(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_HEADER_cpv_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_HEADER_cpv_shift)
+
+/*define for elementsize field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_HEADER_elementsize_offset 0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_HEADER_elementsize_mask   0x00000007
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_HEADER_elementsize_shift  29
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_HEADER_ELEMENTSIZE(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_HEADER_elementsize_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_HEADER_elementsize_shift)
+
+/*define for SRC_ADDR_LO word*/
+/*define for src_addr_31_0 field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_SRC_ADDR_LO_src_addr_31_0_offset 1
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_SRC_ADDR_LO_src_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_SRC_ADDR_LO_src_addr_31_0_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_SRC_ADDR_LO_SRC_ADDR_31_0(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_SRC_ADDR_LO_src_addr_31_0_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_SRC_ADDR_LO_src_addr_31_0_shift)
+
+/*define for SRC_ADDR_HI word*/
+/*define for src_addr_63_32 field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_SRC_ADDR_HI_src_addr_63_32_offset 2
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_SRC_ADDR_HI_src_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_SRC_ADDR_HI_src_addr_63_32_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_SRC_ADDR_HI_SRC_ADDR_63_32(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_SRC_ADDR_HI_src_addr_63_32_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_SRC_ADDR_HI_src_addr_63_32_shift)
+
+/*define for DW_3 word*/
+/*define for src_x field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_3_src_x_offset 3
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_3_src_x_mask   0x00003FFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_3_src_x_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_3_SRC_X(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_DW_3_src_x_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_DW_3_src_x_shift)
+
+/*define for src_y field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_3_src_y_offset 3
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_3_src_y_mask   0x00003FFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_3_src_y_shift  16
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_3_SRC_Y(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_DW_3_src_y_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_DW_3_src_y_shift)
+
+/*define for DW_4 word*/
+/*define for src_z field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_4_src_z_offset 4
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_4_src_z_mask   0x00001FFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_4_src_z_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_4_SRC_Z(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_DW_4_src_z_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_DW_4_src_z_shift)
+
+/*define for src_pitch field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_4_src_pitch_offset 4
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_4_src_pitch_mask   0x0007FFFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_4_src_pitch_shift  13
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_4_SRC_PITCH(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_DW_4_src_pitch_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_DW_4_src_pitch_shift)
+
+/*define for DW_5 word*/
+/*define for src_slice_pitch field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_5_src_slice_pitch_offset 5
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_5_src_slice_pitch_mask   0x0FFFFFFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_5_src_slice_pitch_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_5_SRC_SLICE_PITCH(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_DW_5_src_slice_pitch_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_DW_5_src_slice_pitch_shift)
+
+/*define for DST_ADDR_LO word*/
+/*define for dst_addr_31_0 field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DST_ADDR_LO_dst_addr_31_0_offset 6
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DST_ADDR_LO_dst_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DST_ADDR_LO_dst_addr_31_0_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DST_ADDR_LO_DST_ADDR_31_0(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_DST_ADDR_LO_dst_addr_31_0_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_DST_ADDR_LO_dst_addr_31_0_shift)
+
+/*define for DST_ADDR_HI word*/
+/*define for dst_addr_63_32 field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DST_ADDR_HI_dst_addr_63_32_offset 7
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DST_ADDR_HI_dst_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DST_ADDR_HI_dst_addr_63_32_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DST_ADDR_HI_DST_ADDR_63_32(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_DST_ADDR_HI_dst_addr_63_32_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_DST_ADDR_HI_dst_addr_63_32_shift)
+
+/*define for DW_8 word*/
+/*define for dst_x field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_8_dst_x_offset 8
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_8_dst_x_mask   0x00003FFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_8_dst_x_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_8_DST_X(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_DW_8_dst_x_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_DW_8_dst_x_shift)
+
+/*define for dst_y field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_8_dst_y_offset 8
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_8_dst_y_mask   0x00003FFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_8_dst_y_shift  16
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_8_DST_Y(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_DW_8_dst_y_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_DW_8_dst_y_shift)
+
+/*define for DW_9 word*/
+/*define for dst_z field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_9_dst_z_offset 9
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_9_dst_z_mask   0x00001FFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_9_dst_z_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_9_DST_Z(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_DW_9_dst_z_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_DW_9_dst_z_shift)
+
+/*define for dst_pitch field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_9_dst_pitch_offset 9
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_9_dst_pitch_mask   0x0007FFFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_9_dst_pitch_shift  13
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_9_DST_PITCH(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_DW_9_dst_pitch_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_DW_9_dst_pitch_shift)
+
+/*define for DW_10 word*/
+/*define for dst_slice_pitch field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_10_dst_slice_pitch_offset 10
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_10_dst_slice_pitch_mask   0x0FFFFFFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_10_dst_slice_pitch_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_10_DST_SLICE_PITCH(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_DW_10_dst_slice_pitch_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_DW_10_dst_slice_pitch_shift)
+
+/*define for DW_11 word*/
+/*define for rect_x field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_11_rect_x_offset 11
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_11_rect_x_mask   0x00003FFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_11_rect_x_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_11_RECT_X(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_DW_11_rect_x_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_DW_11_rect_x_shift)
+
+/*define for rect_y field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_11_rect_y_offset 11
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_11_rect_y_mask   0x00003FFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_11_rect_y_shift  16
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_11_RECT_Y(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_DW_11_rect_y_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_DW_11_rect_y_shift)
+
+/*define for DW_12 word*/
+/*define for rect_z field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_12_rect_z_offset 12
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_12_rect_z_mask   0x00001FFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_12_rect_z_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_12_RECT_Z(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_DW_12_rect_z_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_DW_12_rect_z_shift)
+
+/*define for dst_sw field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_12_dst_sw_offset 12
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_12_dst_sw_mask   0x00000003
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_12_dst_sw_shift  16
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_12_DST_SW(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_DW_12_dst_sw_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_DW_12_dst_sw_shift)
+
+/*define for dst_cache_policy field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_12_dst_cache_policy_offset 12
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_12_dst_cache_policy_mask   0x00000007
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_12_dst_cache_policy_shift  18
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_12_DST_CACHE_POLICY(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_DW_12_dst_cache_policy_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_DW_12_dst_cache_policy_shift)
+
+/*define for src_sw field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_12_src_sw_offset 12
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_12_src_sw_mask   0x00000003
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_12_src_sw_shift  24
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_12_SRC_SW(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_DW_12_src_sw_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_DW_12_src_sw_shift)
+
+/*define for src_cache_policy field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_12_src_cache_policy_offset 12
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_12_src_cache_policy_mask   0x00000007
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_12_src_cache_policy_shift  26
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_DW_12_SRC_CACHE_POLICY(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_DW_12_src_cache_policy_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_DW_12_src_cache_policy_shift)
+
+
+/*
+** Definitions for SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_HEADER_op_offset 0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_HEADER_op_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_HEADER_OP(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_HEADER_op_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_HEADER_sub_op_offset 0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_HEADER_sub_op_shift  8
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_HEADER_SUB_OP(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_HEADER_sub_op_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_HEADER_sub_op_shift)
+
+/*define for tmz field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_HEADER_tmz_offset 0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_HEADER_tmz_mask   0x00000001
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_HEADER_tmz_shift  18
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_HEADER_TMZ(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_HEADER_tmz_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_HEADER_tmz_shift)
+
+/*define for cpv field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_HEADER_cpv_offset 0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_HEADER_cpv_mask   0x00000001
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_HEADER_cpv_shift  19
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_HEADER_CPV(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_HEADER_cpv_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_HEADER_cpv_shift)
+
+/*define for SRC_ADDR_LO word*/
+/*define for src_addr_31_0 field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_SRC_ADDR_LO_src_addr_31_0_offset 1
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_SRC_ADDR_LO_src_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_SRC_ADDR_LO_src_addr_31_0_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_SRC_ADDR_LO_SRC_ADDR_31_0(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_SRC_ADDR_LO_src_addr_31_0_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_SRC_ADDR_LO_src_addr_31_0_shift)
+
+/*define for SRC_ADDR_HI word*/
+/*define for src_addr_63_32 field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_SRC_ADDR_HI_src_addr_63_32_offset 2
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_SRC_ADDR_HI_src_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_SRC_ADDR_HI_src_addr_63_32_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_SRC_ADDR_HI_SRC_ADDR_63_32(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_SRC_ADDR_HI_src_addr_63_32_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_SRC_ADDR_HI_src_addr_63_32_shift)
+
+/*define for DW_3 word*/
+/*define for src_x field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_3_src_x_offset 3
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_3_src_x_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_3_src_x_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_3_SRC_X(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_3_src_x_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_3_src_x_shift)
+
+/*define for DW_4 word*/
+/*define for src_y field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_4_src_y_offset 4
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_4_src_y_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_4_src_y_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_4_SRC_Y(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_4_src_y_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_4_src_y_shift)
+
+/*define for DW_5 word*/
+/*define for src_z field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_5_src_z_offset 5
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_5_src_z_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_5_src_z_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_5_SRC_Z(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_5_src_z_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_5_src_z_shift)
+
+/*define for DW_6 word*/
+/*define for src_pitch field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_6_src_pitch_offset 6
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_6_src_pitch_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_6_src_pitch_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_6_SRC_PITCH(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_6_src_pitch_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_6_src_pitch_shift)
+
+/*define for DW_7 word*/
+/*define for src_slice_pitch_31_0 field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_7_src_slice_pitch_31_0_offset 7
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_7_src_slice_pitch_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_7_src_slice_pitch_31_0_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_7_SRC_SLICE_PITCH_31_0(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_7_src_slice_pitch_31_0_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_7_src_slice_pitch_31_0_shift)
+
+/*define for DW_8 word*/
+/*define for src_slice_pitch_47_32 field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_8_src_slice_pitch_47_32_offset 8
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_8_src_slice_pitch_47_32_mask   0x0000FFFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_8_src_slice_pitch_47_32_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_8_SRC_SLICE_PITCH_47_32(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_8_src_slice_pitch_47_32_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_8_src_slice_pitch_47_32_shift)
+
+/*define for DST_ADDR_LO word*/
+/*define for dst_addr_31_0 field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DST_ADDR_LO_dst_addr_31_0_offset 9
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DST_ADDR_LO_dst_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DST_ADDR_LO_dst_addr_31_0_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DST_ADDR_LO_DST_ADDR_31_0(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DST_ADDR_LO_dst_addr_31_0_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DST_ADDR_LO_dst_addr_31_0_shift)
+
+/*define for DST_ADDR_HI word*/
+/*define for dst_addr_63_32 field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DST_ADDR_HI_dst_addr_63_32_offset 10
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DST_ADDR_HI_dst_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DST_ADDR_HI_dst_addr_63_32_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DST_ADDR_HI_DST_ADDR_63_32(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DST_ADDR_HI_dst_addr_63_32_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DST_ADDR_HI_dst_addr_63_32_shift)
+
+/*define for DW_11 word*/
+/*define for dst_x field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_11_dst_x_offset 11
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_11_dst_x_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_11_dst_x_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_11_DST_X(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_11_dst_x_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_11_dst_x_shift)
+
+/*define for DW_12 word*/
+/*define for dst_y field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_12_dst_y_offset 12
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_12_dst_y_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_12_dst_y_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_12_DST_Y(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_12_dst_y_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_12_dst_y_shift)
+
+/*define for DW_13 word*/
+/*define for dst_z field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_13_dst_z_offset 13
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_13_dst_z_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_13_dst_z_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_13_DST_Z(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_13_dst_z_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_13_dst_z_shift)
+
+/*define for DW_14 word*/
+/*define for dst_pitch field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_14_dst_pitch_offset 14
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_14_dst_pitch_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_14_dst_pitch_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_14_DST_PITCH(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_14_dst_pitch_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_14_dst_pitch_shift)
+
+/*define for DW_15 word*/
+/*define for dst_slice_pitch_31_0 field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_15_dst_slice_pitch_31_0_offset 15
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_15_dst_slice_pitch_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_15_dst_slice_pitch_31_0_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_15_DST_SLICE_PITCH_31_0(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_15_dst_slice_pitch_31_0_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_15_dst_slice_pitch_31_0_shift)
+
+/*define for DW_16 word*/
+/*define for dst_slice_pitch_47_32 field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_16_dst_slice_pitch_47_32_offset 16
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_16_dst_slice_pitch_47_32_mask   0x0000FFFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_16_dst_slice_pitch_47_32_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_16_DST_SLICE_PITCH_47_32(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_16_dst_slice_pitch_47_32_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_16_dst_slice_pitch_47_32_shift)
+
+/*define for dst_sw field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_16_dst_sw_offset 16
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_16_dst_sw_mask   0x00000003
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_16_dst_sw_shift  16
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_16_DST_SW(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_16_dst_sw_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_16_dst_sw_shift)
+
+/*define for dst_policy field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_16_dst_policy_offset 16
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_16_dst_policy_mask   0x00000007
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_16_dst_policy_shift  18
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_16_DST_POLICY(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_16_dst_policy_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_16_dst_policy_shift)
+
+/*define for src_sw field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_16_src_sw_offset 16
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_16_src_sw_mask   0x00000003
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_16_src_sw_shift  24
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_16_SRC_SW(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_16_src_sw_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_16_src_sw_shift)
+
+/*define for src_policy field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_16_src_policy_offset 16
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_16_src_policy_mask   0x00000007
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_16_src_policy_shift  26
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_16_SRC_POLICY(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_16_src_policy_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_16_src_policy_shift)
+
+/*define for DW_17 word*/
+/*define for rect_x field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_17_rect_x_offset 17
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_17_rect_x_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_17_rect_x_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_17_RECT_X(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_17_rect_x_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_17_rect_x_shift)
+
+/*define for DW_18 word*/
+/*define for rect_y field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_18_rect_y_offset 18
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_18_rect_y_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_18_rect_y_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_18_RECT_Y(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_18_rect_y_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_18_rect_y_shift)
+
+/*define for DW_19 word*/
+/*define for rect_z field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_19_rect_z_offset 19
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_19_rect_z_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_19_rect_z_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_19_RECT_Z(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_19_rect_z_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_LARGE_DW_19_rect_z_shift)
+
+
+/*
+** Definitions for SDMA_PKT_COPY_LINEAR_SUBWIN_BC packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_HEADER_op_offset 0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_HEADER_op_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_HEADER_OP(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_BC_HEADER_op_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_BC_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_HEADER_sub_op_offset 0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_HEADER_sub_op_shift  8
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_HEADER_SUB_OP(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_BC_HEADER_sub_op_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_BC_HEADER_sub_op_shift)
+
+/*define for elementsize field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_HEADER_elementsize_offset 0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_HEADER_elementsize_mask   0x00000007
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_HEADER_elementsize_shift  29
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_HEADER_ELEMENTSIZE(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_BC_HEADER_elementsize_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_BC_HEADER_elementsize_shift)
+
+/*define for SRC_ADDR_LO word*/
+/*define for src_addr_31_0 field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_SRC_ADDR_LO_src_addr_31_0_offset 1
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_SRC_ADDR_LO_src_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_SRC_ADDR_LO_src_addr_31_0_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_SRC_ADDR_LO_SRC_ADDR_31_0(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_BC_SRC_ADDR_LO_src_addr_31_0_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_BC_SRC_ADDR_LO_src_addr_31_0_shift)
+
+/*define for SRC_ADDR_HI word*/
+/*define for src_addr_63_32 field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_SRC_ADDR_HI_src_addr_63_32_offset 2
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_SRC_ADDR_HI_src_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_SRC_ADDR_HI_src_addr_63_32_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_SRC_ADDR_HI_SRC_ADDR_63_32(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_BC_SRC_ADDR_HI_src_addr_63_32_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_BC_SRC_ADDR_HI_src_addr_63_32_shift)
+
+/*define for DW_3 word*/
+/*define for src_x field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_3_src_x_offset 3
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_3_src_x_mask   0x00003FFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_3_src_x_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_3_SRC_X(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_3_src_x_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_3_src_x_shift)
+
+/*define for src_y field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_3_src_y_offset 3
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_3_src_y_mask   0x00003FFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_3_src_y_shift  16
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_3_SRC_Y(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_3_src_y_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_3_src_y_shift)
+
+/*define for DW_4 word*/
+/*define for src_z field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_4_src_z_offset 4
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_4_src_z_mask   0x000007FF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_4_src_z_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_4_SRC_Z(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_4_src_z_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_4_src_z_shift)
+
+/*define for src_pitch field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_4_src_pitch_offset 4
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_4_src_pitch_mask   0x00003FFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_4_src_pitch_shift  13
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_4_SRC_PITCH(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_4_src_pitch_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_4_src_pitch_shift)
+
+/*define for DW_5 word*/
+/*define for src_slice_pitch field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_5_src_slice_pitch_offset 5
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_5_src_slice_pitch_mask   0x0FFFFFFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_5_src_slice_pitch_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_5_SRC_SLICE_PITCH(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_5_src_slice_pitch_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_5_src_slice_pitch_shift)
+
+/*define for DST_ADDR_LO word*/
+/*define for dst_addr_31_0 field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DST_ADDR_LO_dst_addr_31_0_offset 6
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DST_ADDR_LO_dst_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DST_ADDR_LO_dst_addr_31_0_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DST_ADDR_LO_DST_ADDR_31_0(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DST_ADDR_LO_dst_addr_31_0_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DST_ADDR_LO_dst_addr_31_0_shift)
+
+/*define for DST_ADDR_HI word*/
+/*define for dst_addr_63_32 field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DST_ADDR_HI_dst_addr_63_32_offset 7
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DST_ADDR_HI_dst_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DST_ADDR_HI_dst_addr_63_32_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DST_ADDR_HI_DST_ADDR_63_32(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DST_ADDR_HI_dst_addr_63_32_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DST_ADDR_HI_dst_addr_63_32_shift)
+
+/*define for DW_8 word*/
+/*define for dst_x field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_8_dst_x_offset 8
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_8_dst_x_mask   0x00003FFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_8_dst_x_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_8_DST_X(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_8_dst_x_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_8_dst_x_shift)
+
+/*define for dst_y field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_8_dst_y_offset 8
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_8_dst_y_mask   0x00003FFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_8_dst_y_shift  16
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_8_DST_Y(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_8_dst_y_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_8_dst_y_shift)
+
+/*define for DW_9 word*/
+/*define for dst_z field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_9_dst_z_offset 9
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_9_dst_z_mask   0x000007FF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_9_dst_z_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_9_DST_Z(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_9_dst_z_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_9_dst_z_shift)
+
+/*define for dst_pitch field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_9_dst_pitch_offset 9
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_9_dst_pitch_mask   0x00003FFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_9_dst_pitch_shift  13
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_9_DST_PITCH(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_9_dst_pitch_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_9_dst_pitch_shift)
+
+/*define for DW_10 word*/
+/*define for dst_slice_pitch field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_10_dst_slice_pitch_offset 10
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_10_dst_slice_pitch_mask   0x0FFFFFFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_10_dst_slice_pitch_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_10_DST_SLICE_PITCH(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_10_dst_slice_pitch_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_10_dst_slice_pitch_shift)
+
+/*define for DW_11 word*/
+/*define for rect_x field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_11_rect_x_offset 11
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_11_rect_x_mask   0x00003FFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_11_rect_x_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_11_RECT_X(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_11_rect_x_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_11_rect_x_shift)
+
+/*define for rect_y field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_11_rect_y_offset 11
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_11_rect_y_mask   0x00003FFF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_11_rect_y_shift  16
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_11_RECT_Y(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_11_rect_y_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_11_rect_y_shift)
+
+/*define for DW_12 word*/
+/*define for rect_z field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_12_rect_z_offset 12
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_12_rect_z_mask   0x000007FF
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_12_rect_z_shift  0
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_12_RECT_Z(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_12_rect_z_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_12_rect_z_shift)
+
+/*define for dst_sw field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_12_dst_sw_offset 12
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_12_dst_sw_mask   0x00000003
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_12_dst_sw_shift  16
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_12_DST_SW(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_12_dst_sw_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_12_dst_sw_shift)
+
+/*define for dst_ha field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_12_dst_ha_offset 12
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_12_dst_ha_mask   0x00000001
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_12_dst_ha_shift  19
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_12_DST_HA(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_12_dst_ha_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_12_dst_ha_shift)
+
+/*define for src_sw field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_12_src_sw_offset 12
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_12_src_sw_mask   0x00000003
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_12_src_sw_shift  24
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_12_SRC_SW(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_12_src_sw_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_12_src_sw_shift)
+
+/*define for src_ha field*/
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_12_src_ha_offset 12
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_12_src_ha_mask   0x00000001
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_12_src_ha_shift  27
+#define SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_12_SRC_HA(x) (((x) & SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_12_src_ha_mask) << SDMA_PKT_COPY_LINEAR_SUBWIN_BC_DW_12_src_ha_shift)
+
+
+/*
+** Definitions for SDMA_PKT_COPY_TILED packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_COPY_TILED_HEADER_op_offset 0
+#define SDMA_PKT_COPY_TILED_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_COPY_TILED_HEADER_op_shift  0
+#define SDMA_PKT_COPY_TILED_HEADER_OP(x) (((x) & SDMA_PKT_COPY_TILED_HEADER_op_mask) << SDMA_PKT_COPY_TILED_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_COPY_TILED_HEADER_sub_op_offset 0
+#define SDMA_PKT_COPY_TILED_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_COPY_TILED_HEADER_sub_op_shift  8
+#define SDMA_PKT_COPY_TILED_HEADER_SUB_OP(x) (((x) & SDMA_PKT_COPY_TILED_HEADER_sub_op_mask) << SDMA_PKT_COPY_TILED_HEADER_sub_op_shift)
+
+/*define for encrypt field*/
+#define SDMA_PKT_COPY_TILED_HEADER_encrypt_offset 0
+#define SDMA_PKT_COPY_TILED_HEADER_encrypt_mask   0x00000001
+#define SDMA_PKT_COPY_TILED_HEADER_encrypt_shift  16
+#define SDMA_PKT_COPY_TILED_HEADER_ENCRYPT(x) (((x) & SDMA_PKT_COPY_TILED_HEADER_encrypt_mask) << SDMA_PKT_COPY_TILED_HEADER_encrypt_shift)
+
+/*define for tmz field*/
+#define SDMA_PKT_COPY_TILED_HEADER_tmz_offset 0
+#define SDMA_PKT_COPY_TILED_HEADER_tmz_mask   0x00000001
+#define SDMA_PKT_COPY_TILED_HEADER_tmz_shift  18
+#define SDMA_PKT_COPY_TILED_HEADER_TMZ(x) (((x) & SDMA_PKT_COPY_TILED_HEADER_tmz_mask) << SDMA_PKT_COPY_TILED_HEADER_tmz_shift)
+
+/*define for cpv field*/
+#define SDMA_PKT_COPY_TILED_HEADER_cpv_offset 0
+#define SDMA_PKT_COPY_TILED_HEADER_cpv_mask   0x00000001
+#define SDMA_PKT_COPY_TILED_HEADER_cpv_shift  19
+#define SDMA_PKT_COPY_TILED_HEADER_CPV(x) (((x) & SDMA_PKT_COPY_TILED_HEADER_cpv_mask) << SDMA_PKT_COPY_TILED_HEADER_cpv_shift)
+
+/*define for detile field*/
+#define SDMA_PKT_COPY_TILED_HEADER_detile_offset 0
+#define SDMA_PKT_COPY_TILED_HEADER_detile_mask   0x00000001
+#define SDMA_PKT_COPY_TILED_HEADER_detile_shift  31
+#define SDMA_PKT_COPY_TILED_HEADER_DETILE(x) (((x) & SDMA_PKT_COPY_TILED_HEADER_detile_mask) << SDMA_PKT_COPY_TILED_HEADER_detile_shift)
+
+/*define for TILED_ADDR_LO word*/
+/*define for tiled_addr_31_0 field*/
+#define SDMA_PKT_COPY_TILED_TILED_ADDR_LO_tiled_addr_31_0_offset 1
+#define SDMA_PKT_COPY_TILED_TILED_ADDR_LO_tiled_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_TILED_TILED_ADDR_LO_tiled_addr_31_0_shift  0
+#define SDMA_PKT_COPY_TILED_TILED_ADDR_LO_TILED_ADDR_31_0(x) (((x) & SDMA_PKT_COPY_TILED_TILED_ADDR_LO_tiled_addr_31_0_mask) << SDMA_PKT_COPY_TILED_TILED_ADDR_LO_tiled_addr_31_0_shift)
+
+/*define for TILED_ADDR_HI word*/
+/*define for tiled_addr_63_32 field*/
+#define SDMA_PKT_COPY_TILED_TILED_ADDR_HI_tiled_addr_63_32_offset 2
+#define SDMA_PKT_COPY_TILED_TILED_ADDR_HI_tiled_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_TILED_TILED_ADDR_HI_tiled_addr_63_32_shift  0
+#define SDMA_PKT_COPY_TILED_TILED_ADDR_HI_TILED_ADDR_63_32(x) (((x) & SDMA_PKT_COPY_TILED_TILED_ADDR_HI_tiled_addr_63_32_mask) << SDMA_PKT_COPY_TILED_TILED_ADDR_HI_tiled_addr_63_32_shift)
+
+/*define for DW_3 word*/
+/*define for width field*/
+#define SDMA_PKT_COPY_TILED_DW_3_width_offset 3
+#define SDMA_PKT_COPY_TILED_DW_3_width_mask   0x00003FFF
+#define SDMA_PKT_COPY_TILED_DW_3_width_shift  0
+#define SDMA_PKT_COPY_TILED_DW_3_WIDTH(x) (((x) & SDMA_PKT_COPY_TILED_DW_3_width_mask) << SDMA_PKT_COPY_TILED_DW_3_width_shift)
+
+/*define for DW_4 word*/
+/*define for height field*/
+#define SDMA_PKT_COPY_TILED_DW_4_height_offset 4
+#define SDMA_PKT_COPY_TILED_DW_4_height_mask   0x00003FFF
+#define SDMA_PKT_COPY_TILED_DW_4_height_shift  0
+#define SDMA_PKT_COPY_TILED_DW_4_HEIGHT(x) (((x) & SDMA_PKT_COPY_TILED_DW_4_height_mask) << SDMA_PKT_COPY_TILED_DW_4_height_shift)
+
+/*define for depth field*/
+#define SDMA_PKT_COPY_TILED_DW_4_depth_offset 4
+#define SDMA_PKT_COPY_TILED_DW_4_depth_mask   0x00001FFF
+#define SDMA_PKT_COPY_TILED_DW_4_depth_shift  16
+#define SDMA_PKT_COPY_TILED_DW_4_DEPTH(x) (((x) & SDMA_PKT_COPY_TILED_DW_4_depth_mask) << SDMA_PKT_COPY_TILED_DW_4_depth_shift)
+
+/*define for DW_5 word*/
+/*define for element_size field*/
+#define SDMA_PKT_COPY_TILED_DW_5_element_size_offset 5
+#define SDMA_PKT_COPY_TILED_DW_5_element_size_mask   0x00000007
+#define SDMA_PKT_COPY_TILED_DW_5_element_size_shift  0
+#define SDMA_PKT_COPY_TILED_DW_5_ELEMENT_SIZE(x) (((x) & SDMA_PKT_COPY_TILED_DW_5_element_size_mask) << SDMA_PKT_COPY_TILED_DW_5_element_size_shift)
+
+/*define for swizzle_mode field*/
+#define SDMA_PKT_COPY_TILED_DW_5_swizzle_mode_offset 5
+#define SDMA_PKT_COPY_TILED_DW_5_swizzle_mode_mask   0x0000001F
+#define SDMA_PKT_COPY_TILED_DW_5_swizzle_mode_shift  3
+#define SDMA_PKT_COPY_TILED_DW_5_SWIZZLE_MODE(x) (((x) & SDMA_PKT_COPY_TILED_DW_5_swizzle_mode_mask) << SDMA_PKT_COPY_TILED_DW_5_swizzle_mode_shift)
+
+/*define for dimension field*/
+#define SDMA_PKT_COPY_TILED_DW_5_dimension_offset 5
+#define SDMA_PKT_COPY_TILED_DW_5_dimension_mask   0x00000003
+#define SDMA_PKT_COPY_TILED_DW_5_dimension_shift  9
+#define SDMA_PKT_COPY_TILED_DW_5_DIMENSION(x) (((x) & SDMA_PKT_COPY_TILED_DW_5_dimension_mask) << SDMA_PKT_COPY_TILED_DW_5_dimension_shift)
+
+/*define for mip_max field*/
+#define SDMA_PKT_COPY_TILED_DW_5_mip_max_offset 5
+#define SDMA_PKT_COPY_TILED_DW_5_mip_max_mask   0x0000000F
+#define SDMA_PKT_COPY_TILED_DW_5_mip_max_shift  16
+#define SDMA_PKT_COPY_TILED_DW_5_MIP_MAX(x) (((x) & SDMA_PKT_COPY_TILED_DW_5_mip_max_mask) << SDMA_PKT_COPY_TILED_DW_5_mip_max_shift)
+
+/*define for DW_6 word*/
+/*define for x field*/
+#define SDMA_PKT_COPY_TILED_DW_6_x_offset 6
+#define SDMA_PKT_COPY_TILED_DW_6_x_mask   0x00003FFF
+#define SDMA_PKT_COPY_TILED_DW_6_x_shift  0
+#define SDMA_PKT_COPY_TILED_DW_6_X(x) (((x) & SDMA_PKT_COPY_TILED_DW_6_x_mask) << SDMA_PKT_COPY_TILED_DW_6_x_shift)
+
+/*define for y field*/
+#define SDMA_PKT_COPY_TILED_DW_6_y_offset 6
+#define SDMA_PKT_COPY_TILED_DW_6_y_mask   0x00003FFF
+#define SDMA_PKT_COPY_TILED_DW_6_y_shift  16
+#define SDMA_PKT_COPY_TILED_DW_6_Y(x) (((x) & SDMA_PKT_COPY_TILED_DW_6_y_mask) << SDMA_PKT_COPY_TILED_DW_6_y_shift)
+
+/*define for DW_7 word*/
+/*define for z field*/
+#define SDMA_PKT_COPY_TILED_DW_7_z_offset 7
+#define SDMA_PKT_COPY_TILED_DW_7_z_mask   0x00001FFF
+#define SDMA_PKT_COPY_TILED_DW_7_z_shift  0
+#define SDMA_PKT_COPY_TILED_DW_7_Z(x) (((x) & SDMA_PKT_COPY_TILED_DW_7_z_mask) << SDMA_PKT_COPY_TILED_DW_7_z_shift)
+
+/*define for linear_sw field*/
+#define SDMA_PKT_COPY_TILED_DW_7_linear_sw_offset 7
+#define SDMA_PKT_COPY_TILED_DW_7_linear_sw_mask   0x00000003
+#define SDMA_PKT_COPY_TILED_DW_7_linear_sw_shift  16
+#define SDMA_PKT_COPY_TILED_DW_7_LINEAR_SW(x) (((x) & SDMA_PKT_COPY_TILED_DW_7_linear_sw_mask) << SDMA_PKT_COPY_TILED_DW_7_linear_sw_shift)
+
+/*define for linear_cache_policy field*/
+#define SDMA_PKT_COPY_TILED_DW_7_linear_cache_policy_offset 7
+#define SDMA_PKT_COPY_TILED_DW_7_linear_cache_policy_mask   0x00000007
+#define SDMA_PKT_COPY_TILED_DW_7_linear_cache_policy_shift  18
+#define SDMA_PKT_COPY_TILED_DW_7_LINEAR_CACHE_POLICY(x) (((x) & SDMA_PKT_COPY_TILED_DW_7_linear_cache_policy_mask) << SDMA_PKT_COPY_TILED_DW_7_linear_cache_policy_shift)
+
+/*define for tile_sw field*/
+#define SDMA_PKT_COPY_TILED_DW_7_tile_sw_offset 7
+#define SDMA_PKT_COPY_TILED_DW_7_tile_sw_mask   0x00000003
+#define SDMA_PKT_COPY_TILED_DW_7_tile_sw_shift  24
+#define SDMA_PKT_COPY_TILED_DW_7_TILE_SW(x) (((x) & SDMA_PKT_COPY_TILED_DW_7_tile_sw_mask) << SDMA_PKT_COPY_TILED_DW_7_tile_sw_shift)
+
+/*define for tile_cache_policy field*/
+#define SDMA_PKT_COPY_TILED_DW_7_tile_cache_policy_offset 7
+#define SDMA_PKT_COPY_TILED_DW_7_tile_cache_policy_mask   0x00000007
+#define SDMA_PKT_COPY_TILED_DW_7_tile_cache_policy_shift  26
+#define SDMA_PKT_COPY_TILED_DW_7_TILE_CACHE_POLICY(x) (((x) & SDMA_PKT_COPY_TILED_DW_7_tile_cache_policy_mask) << SDMA_PKT_COPY_TILED_DW_7_tile_cache_policy_shift)
+
+/*define for LINEAR_ADDR_LO word*/
+/*define for linear_addr_31_0 field*/
+#define SDMA_PKT_COPY_TILED_LINEAR_ADDR_LO_linear_addr_31_0_offset 8
+#define SDMA_PKT_COPY_TILED_LINEAR_ADDR_LO_linear_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_TILED_LINEAR_ADDR_LO_linear_addr_31_0_shift  0
+#define SDMA_PKT_COPY_TILED_LINEAR_ADDR_LO_LINEAR_ADDR_31_0(x) (((x) & SDMA_PKT_COPY_TILED_LINEAR_ADDR_LO_linear_addr_31_0_mask) << SDMA_PKT_COPY_TILED_LINEAR_ADDR_LO_linear_addr_31_0_shift)
+
+/*define for LINEAR_ADDR_HI word*/
+/*define for linear_addr_63_32 field*/
+#define SDMA_PKT_COPY_TILED_LINEAR_ADDR_HI_linear_addr_63_32_offset 9
+#define SDMA_PKT_COPY_TILED_LINEAR_ADDR_HI_linear_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_TILED_LINEAR_ADDR_HI_linear_addr_63_32_shift  0
+#define SDMA_PKT_COPY_TILED_LINEAR_ADDR_HI_LINEAR_ADDR_63_32(x) (((x) & SDMA_PKT_COPY_TILED_LINEAR_ADDR_HI_linear_addr_63_32_mask) << SDMA_PKT_COPY_TILED_LINEAR_ADDR_HI_linear_addr_63_32_shift)
+
+/*define for LINEAR_PITCH word*/
+/*define for linear_pitch field*/
+#define SDMA_PKT_COPY_TILED_LINEAR_PITCH_linear_pitch_offset 10
+#define SDMA_PKT_COPY_TILED_LINEAR_PITCH_linear_pitch_mask   0x0007FFFF
+#define SDMA_PKT_COPY_TILED_LINEAR_PITCH_linear_pitch_shift  0
+#define SDMA_PKT_COPY_TILED_LINEAR_PITCH_LINEAR_PITCH(x) (((x) & SDMA_PKT_COPY_TILED_LINEAR_PITCH_linear_pitch_mask) << SDMA_PKT_COPY_TILED_LINEAR_PITCH_linear_pitch_shift)
+
+/*define for LINEAR_SLICE_PITCH word*/
+/*define for linear_slice_pitch field*/
+#define SDMA_PKT_COPY_TILED_LINEAR_SLICE_PITCH_linear_slice_pitch_offset 11
+#define SDMA_PKT_COPY_TILED_LINEAR_SLICE_PITCH_linear_slice_pitch_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_TILED_LINEAR_SLICE_PITCH_linear_slice_pitch_shift  0
+#define SDMA_PKT_COPY_TILED_LINEAR_SLICE_PITCH_LINEAR_SLICE_PITCH(x) (((x) & SDMA_PKT_COPY_TILED_LINEAR_SLICE_PITCH_linear_slice_pitch_mask) << SDMA_PKT_COPY_TILED_LINEAR_SLICE_PITCH_linear_slice_pitch_shift)
+
+/*define for COUNT word*/
+/*define for count field*/
+#define SDMA_PKT_COPY_TILED_COUNT_count_offset 12
+#define SDMA_PKT_COPY_TILED_COUNT_count_mask   0x3FFFFFFF
+#define SDMA_PKT_COPY_TILED_COUNT_count_shift  0
+#define SDMA_PKT_COPY_TILED_COUNT_COUNT(x) (((x) & SDMA_PKT_COPY_TILED_COUNT_count_mask) << SDMA_PKT_COPY_TILED_COUNT_count_shift)
+
+
+/*
+** Definitions for SDMA_PKT_COPY_TILED_BC packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_COPY_TILED_BC_HEADER_op_offset 0
+#define SDMA_PKT_COPY_TILED_BC_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_COPY_TILED_BC_HEADER_op_shift  0
+#define SDMA_PKT_COPY_TILED_BC_HEADER_OP(x) (((x) & SDMA_PKT_COPY_TILED_BC_HEADER_op_mask) << SDMA_PKT_COPY_TILED_BC_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_COPY_TILED_BC_HEADER_sub_op_offset 0
+#define SDMA_PKT_COPY_TILED_BC_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_COPY_TILED_BC_HEADER_sub_op_shift  8
+#define SDMA_PKT_COPY_TILED_BC_HEADER_SUB_OP(x) (((x) & SDMA_PKT_COPY_TILED_BC_HEADER_sub_op_mask) << SDMA_PKT_COPY_TILED_BC_HEADER_sub_op_shift)
+
+/*define for detile field*/
+#define SDMA_PKT_COPY_TILED_BC_HEADER_detile_offset 0
+#define SDMA_PKT_COPY_TILED_BC_HEADER_detile_mask   0x00000001
+#define SDMA_PKT_COPY_TILED_BC_HEADER_detile_shift  31
+#define SDMA_PKT_COPY_TILED_BC_HEADER_DETILE(x) (((x) & SDMA_PKT_COPY_TILED_BC_HEADER_detile_mask) << SDMA_PKT_COPY_TILED_BC_HEADER_detile_shift)
+
+/*define for TILED_ADDR_LO word*/
+/*define for tiled_addr_31_0 field*/
+#define SDMA_PKT_COPY_TILED_BC_TILED_ADDR_LO_tiled_addr_31_0_offset 1
+#define SDMA_PKT_COPY_TILED_BC_TILED_ADDR_LO_tiled_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_TILED_BC_TILED_ADDR_LO_tiled_addr_31_0_shift  0
+#define SDMA_PKT_COPY_TILED_BC_TILED_ADDR_LO_TILED_ADDR_31_0(x) (((x) & SDMA_PKT_COPY_TILED_BC_TILED_ADDR_LO_tiled_addr_31_0_mask) << SDMA_PKT_COPY_TILED_BC_TILED_ADDR_LO_tiled_addr_31_0_shift)
+
+/*define for TILED_ADDR_HI word*/
+/*define for tiled_addr_63_32 field*/
+#define SDMA_PKT_COPY_TILED_BC_TILED_ADDR_HI_tiled_addr_63_32_offset 2
+#define SDMA_PKT_COPY_TILED_BC_TILED_ADDR_HI_tiled_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_TILED_BC_TILED_ADDR_HI_tiled_addr_63_32_shift  0
+#define SDMA_PKT_COPY_TILED_BC_TILED_ADDR_HI_TILED_ADDR_63_32(x) (((x) & SDMA_PKT_COPY_TILED_BC_TILED_ADDR_HI_tiled_addr_63_32_mask) << SDMA_PKT_COPY_TILED_BC_TILED_ADDR_HI_tiled_addr_63_32_shift)
+
+/*define for DW_3 word*/
+/*define for width field*/
+#define SDMA_PKT_COPY_TILED_BC_DW_3_width_offset 3
+#define SDMA_PKT_COPY_TILED_BC_DW_3_width_mask   0x00003FFF
+#define SDMA_PKT_COPY_TILED_BC_DW_3_width_shift  0
+#define SDMA_PKT_COPY_TILED_BC_DW_3_WIDTH(x) (((x) & SDMA_PKT_COPY_TILED_BC_DW_3_width_mask) << SDMA_PKT_COPY_TILED_BC_DW_3_width_shift)
+
+/*define for DW_4 word*/
+/*define for height field*/
+#define SDMA_PKT_COPY_TILED_BC_DW_4_height_offset 4
+#define SDMA_PKT_COPY_TILED_BC_DW_4_height_mask   0x00003FFF
+#define SDMA_PKT_COPY_TILED_BC_DW_4_height_shift  0
+#define SDMA_PKT_COPY_TILED_BC_DW_4_HEIGHT(x) (((x) & SDMA_PKT_COPY_TILED_BC_DW_4_height_mask) << SDMA_PKT_COPY_TILED_BC_DW_4_height_shift)
+
+/*define for depth field*/
+#define SDMA_PKT_COPY_TILED_BC_DW_4_depth_offset 4
+#define SDMA_PKT_COPY_TILED_BC_DW_4_depth_mask   0x000007FF
+#define SDMA_PKT_COPY_TILED_BC_DW_4_depth_shift  16
+#define SDMA_PKT_COPY_TILED_BC_DW_4_DEPTH(x) (((x) & SDMA_PKT_COPY_TILED_BC_DW_4_depth_mask) << SDMA_PKT_COPY_TILED_BC_DW_4_depth_shift)
+
+/*define for DW_5 word*/
+/*define for element_size field*/
+#define SDMA_PKT_COPY_TILED_BC_DW_5_element_size_offset 5
+#define SDMA_PKT_COPY_TILED_BC_DW_5_element_size_mask   0x00000007
+#define SDMA_PKT_COPY_TILED_BC_DW_5_element_size_shift  0
+#define SDMA_PKT_COPY_TILED_BC_DW_5_ELEMENT_SIZE(x) (((x) & SDMA_PKT_COPY_TILED_BC_DW_5_element_size_mask) << SDMA_PKT_COPY_TILED_BC_DW_5_element_size_shift)
+
+/*define for array_mode field*/
+#define SDMA_PKT_COPY_TILED_BC_DW_5_array_mode_offset 5
+#define SDMA_PKT_COPY_TILED_BC_DW_5_array_mode_mask   0x0000000F
+#define SDMA_PKT_COPY_TILED_BC_DW_5_array_mode_shift  3
+#define SDMA_PKT_COPY_TILED_BC_DW_5_ARRAY_MODE(x) (((x) & SDMA_PKT_COPY_TILED_BC_DW_5_array_mode_mask) << SDMA_PKT_COPY_TILED_BC_DW_5_array_mode_shift)
+
+/*define for mit_mode field*/
+#define SDMA_PKT_COPY_TILED_BC_DW_5_mit_mode_offset 5
+#define SDMA_PKT_COPY_TILED_BC_DW_5_mit_mode_mask   0x00000007
+#define SDMA_PKT_COPY_TILED_BC_DW_5_mit_mode_shift  8
+#define SDMA_PKT_COPY_TILED_BC_DW_5_MIT_MODE(x) (((x) & SDMA_PKT_COPY_TILED_BC_DW_5_mit_mode_mask) << SDMA_PKT_COPY_TILED_BC_DW_5_mit_mode_shift)
+
+/*define for tilesplit_size field*/
+#define SDMA_PKT_COPY_TILED_BC_DW_5_tilesplit_size_offset 5
+#define SDMA_PKT_COPY_TILED_BC_DW_5_tilesplit_size_mask   0x00000007
+#define SDMA_PKT_COPY_TILED_BC_DW_5_tilesplit_size_shift  11
+#define SDMA_PKT_COPY_TILED_BC_DW_5_TILESPLIT_SIZE(x) (((x) & SDMA_PKT_COPY_TILED_BC_DW_5_tilesplit_size_mask) << SDMA_PKT_COPY_TILED_BC_DW_5_tilesplit_size_shift)
+
+/*define for bank_w field*/
+#define SDMA_PKT_COPY_TILED_BC_DW_5_bank_w_offset 5
+#define SDMA_PKT_COPY_TILED_BC_DW_5_bank_w_mask   0x00000003
+#define SDMA_PKT_COPY_TILED_BC_DW_5_bank_w_shift  15
+#define SDMA_PKT_COPY_TILED_BC_DW_5_BANK_W(x) (((x) & SDMA_PKT_COPY_TILED_BC_DW_5_bank_w_mask) << SDMA_PKT_COPY_TILED_BC_DW_5_bank_w_shift)
+
+/*define for bank_h field*/
+#define SDMA_PKT_COPY_TILED_BC_DW_5_bank_h_offset 5
+#define SDMA_PKT_COPY_TILED_BC_DW_5_bank_h_mask   0x00000003
+#define SDMA_PKT_COPY_TILED_BC_DW_5_bank_h_shift  18
+#define SDMA_PKT_COPY_TILED_BC_DW_5_BANK_H(x) (((x) & SDMA_PKT_COPY_TILED_BC_DW_5_bank_h_mask) << SDMA_PKT_COPY_TILED_BC_DW_5_bank_h_shift)
+
+/*define for num_bank field*/
+#define SDMA_PKT_COPY_TILED_BC_DW_5_num_bank_offset 5
+#define SDMA_PKT_COPY_TILED_BC_DW_5_num_bank_mask   0x00000003
+#define SDMA_PKT_COPY_TILED_BC_DW_5_num_bank_shift  21
+#define SDMA_PKT_COPY_TILED_BC_DW_5_NUM_BANK(x) (((x) & SDMA_PKT_COPY_TILED_BC_DW_5_num_bank_mask) << SDMA_PKT_COPY_TILED_BC_DW_5_num_bank_shift)
+
+/*define for mat_aspt field*/
+#define SDMA_PKT_COPY_TILED_BC_DW_5_mat_aspt_offset 5
+#define SDMA_PKT_COPY_TILED_BC_DW_5_mat_aspt_mask   0x00000003
+#define SDMA_PKT_COPY_TILED_BC_DW_5_mat_aspt_shift  24
+#define SDMA_PKT_COPY_TILED_BC_DW_5_MAT_ASPT(x) (((x) & SDMA_PKT_COPY_TILED_BC_DW_5_mat_aspt_mask) << SDMA_PKT_COPY_TILED_BC_DW_5_mat_aspt_shift)
+
+/*define for pipe_config field*/
+#define SDMA_PKT_COPY_TILED_BC_DW_5_pipe_config_offset 5
+#define SDMA_PKT_COPY_TILED_BC_DW_5_pipe_config_mask   0x0000001F
+#define SDMA_PKT_COPY_TILED_BC_DW_5_pipe_config_shift  26
+#define SDMA_PKT_COPY_TILED_BC_DW_5_PIPE_CONFIG(x) (((x) & SDMA_PKT_COPY_TILED_BC_DW_5_pipe_config_mask) << SDMA_PKT_COPY_TILED_BC_DW_5_pipe_config_shift)
+
+/*define for DW_6 word*/
+/*define for x field*/
+#define SDMA_PKT_COPY_TILED_BC_DW_6_x_offset 6
+#define SDMA_PKT_COPY_TILED_BC_DW_6_x_mask   0x00003FFF
+#define SDMA_PKT_COPY_TILED_BC_DW_6_x_shift  0
+#define SDMA_PKT_COPY_TILED_BC_DW_6_X(x) (((x) & SDMA_PKT_COPY_TILED_BC_DW_6_x_mask) << SDMA_PKT_COPY_TILED_BC_DW_6_x_shift)
+
+/*define for y field*/
+#define SDMA_PKT_COPY_TILED_BC_DW_6_y_offset 6
+#define SDMA_PKT_COPY_TILED_BC_DW_6_y_mask   0x00003FFF
+#define SDMA_PKT_COPY_TILED_BC_DW_6_y_shift  16
+#define SDMA_PKT_COPY_TILED_BC_DW_6_Y(x) (((x) & SDMA_PKT_COPY_TILED_BC_DW_6_y_mask) << SDMA_PKT_COPY_TILED_BC_DW_6_y_shift)
+
+/*define for DW_7 word*/
+/*define for z field*/
+#define SDMA_PKT_COPY_TILED_BC_DW_7_z_offset 7
+#define SDMA_PKT_COPY_TILED_BC_DW_7_z_mask   0x000007FF
+#define SDMA_PKT_COPY_TILED_BC_DW_7_z_shift  0
+#define SDMA_PKT_COPY_TILED_BC_DW_7_Z(x) (((x) & SDMA_PKT_COPY_TILED_BC_DW_7_z_mask) << SDMA_PKT_COPY_TILED_BC_DW_7_z_shift)
+
+/*define for linear_sw field*/
+#define SDMA_PKT_COPY_TILED_BC_DW_7_linear_sw_offset 7
+#define SDMA_PKT_COPY_TILED_BC_DW_7_linear_sw_mask   0x00000003
+#define SDMA_PKT_COPY_TILED_BC_DW_7_linear_sw_shift  16
+#define SDMA_PKT_COPY_TILED_BC_DW_7_LINEAR_SW(x) (((x) & SDMA_PKT_COPY_TILED_BC_DW_7_linear_sw_mask) << SDMA_PKT_COPY_TILED_BC_DW_7_linear_sw_shift)
+
+/*define for tile_sw field*/
+#define SDMA_PKT_COPY_TILED_BC_DW_7_tile_sw_offset 7
+#define SDMA_PKT_COPY_TILED_BC_DW_7_tile_sw_mask   0x00000003
+#define SDMA_PKT_COPY_TILED_BC_DW_7_tile_sw_shift  24
+#define SDMA_PKT_COPY_TILED_BC_DW_7_TILE_SW(x) (((x) & SDMA_PKT_COPY_TILED_BC_DW_7_tile_sw_mask) << SDMA_PKT_COPY_TILED_BC_DW_7_tile_sw_shift)
+
+/*define for LINEAR_ADDR_LO word*/
+/*define for linear_addr_31_0 field*/
+#define SDMA_PKT_COPY_TILED_BC_LINEAR_ADDR_LO_linear_addr_31_0_offset 8
+#define SDMA_PKT_COPY_TILED_BC_LINEAR_ADDR_LO_linear_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_TILED_BC_LINEAR_ADDR_LO_linear_addr_31_0_shift  0
+#define SDMA_PKT_COPY_TILED_BC_LINEAR_ADDR_LO_LINEAR_ADDR_31_0(x) (((x) & SDMA_PKT_COPY_TILED_BC_LINEAR_ADDR_LO_linear_addr_31_0_mask) << SDMA_PKT_COPY_TILED_BC_LINEAR_ADDR_LO_linear_addr_31_0_shift)
+
+/*define for LINEAR_ADDR_HI word*/
+/*define for linear_addr_63_32 field*/
+#define SDMA_PKT_COPY_TILED_BC_LINEAR_ADDR_HI_linear_addr_63_32_offset 9
+#define SDMA_PKT_COPY_TILED_BC_LINEAR_ADDR_HI_linear_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_TILED_BC_LINEAR_ADDR_HI_linear_addr_63_32_shift  0
+#define SDMA_PKT_COPY_TILED_BC_LINEAR_ADDR_HI_LINEAR_ADDR_63_32(x) (((x) & SDMA_PKT_COPY_TILED_BC_LINEAR_ADDR_HI_linear_addr_63_32_mask) << SDMA_PKT_COPY_TILED_BC_LINEAR_ADDR_HI_linear_addr_63_32_shift)
+
+/*define for LINEAR_PITCH word*/
+/*define for linear_pitch field*/
+#define SDMA_PKT_COPY_TILED_BC_LINEAR_PITCH_linear_pitch_offset 10
+#define SDMA_PKT_COPY_TILED_BC_LINEAR_PITCH_linear_pitch_mask   0x0007FFFF
+#define SDMA_PKT_COPY_TILED_BC_LINEAR_PITCH_linear_pitch_shift  0
+#define SDMA_PKT_COPY_TILED_BC_LINEAR_PITCH_LINEAR_PITCH(x) (((x) & SDMA_PKT_COPY_TILED_BC_LINEAR_PITCH_linear_pitch_mask) << SDMA_PKT_COPY_TILED_BC_LINEAR_PITCH_linear_pitch_shift)
+
+/*define for LINEAR_SLICE_PITCH word*/
+/*define for linear_slice_pitch field*/
+#define SDMA_PKT_COPY_TILED_BC_LINEAR_SLICE_PITCH_linear_slice_pitch_offset 11
+#define SDMA_PKT_COPY_TILED_BC_LINEAR_SLICE_PITCH_linear_slice_pitch_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_TILED_BC_LINEAR_SLICE_PITCH_linear_slice_pitch_shift  0
+#define SDMA_PKT_COPY_TILED_BC_LINEAR_SLICE_PITCH_LINEAR_SLICE_PITCH(x) (((x) & SDMA_PKT_COPY_TILED_BC_LINEAR_SLICE_PITCH_linear_slice_pitch_mask) << SDMA_PKT_COPY_TILED_BC_LINEAR_SLICE_PITCH_linear_slice_pitch_shift)
+
+/*define for COUNT word*/
+/*define for count field*/
+#define SDMA_PKT_COPY_TILED_BC_COUNT_count_offset 12
+#define SDMA_PKT_COPY_TILED_BC_COUNT_count_mask   0x000FFFFF
+#define SDMA_PKT_COPY_TILED_BC_COUNT_count_shift  2
+#define SDMA_PKT_COPY_TILED_BC_COUNT_COUNT(x) (((x) & SDMA_PKT_COPY_TILED_BC_COUNT_count_mask) << SDMA_PKT_COPY_TILED_BC_COUNT_count_shift)
+
+
+/*
+** Definitions for SDMA_PKT_COPY_L2T_BROADCAST packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_COPY_L2T_BROADCAST_HEADER_op_offset 0
+#define SDMA_PKT_COPY_L2T_BROADCAST_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_COPY_L2T_BROADCAST_HEADER_op_shift  0
+#define SDMA_PKT_COPY_L2T_BROADCAST_HEADER_OP(x) (((x) & SDMA_PKT_COPY_L2T_BROADCAST_HEADER_op_mask) << SDMA_PKT_COPY_L2T_BROADCAST_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_COPY_L2T_BROADCAST_HEADER_sub_op_offset 0
+#define SDMA_PKT_COPY_L2T_BROADCAST_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_COPY_L2T_BROADCAST_HEADER_sub_op_shift  8
+#define SDMA_PKT_COPY_L2T_BROADCAST_HEADER_SUB_OP(x) (((x) & SDMA_PKT_COPY_L2T_BROADCAST_HEADER_sub_op_mask) << SDMA_PKT_COPY_L2T_BROADCAST_HEADER_sub_op_shift)
+
+/*define for encrypt field*/
+#define SDMA_PKT_COPY_L2T_BROADCAST_HEADER_encrypt_offset 0
+#define SDMA_PKT_COPY_L2T_BROADCAST_HEADER_encrypt_mask   0x00000001
+#define SDMA_PKT_COPY_L2T_BROADCAST_HEADER_encrypt_shift  16
+#define SDMA_PKT_COPY_L2T_BROADCAST_HEADER_ENCRYPT(x) (((x) & SDMA_PKT_COPY_L2T_BROADCAST_HEADER_encrypt_mask) << SDMA_PKT_COPY_L2T_BROADCAST_HEADER_encrypt_shift)
+
+/*define for tmz field*/
+#define SDMA_PKT_COPY_L2T_BROADCAST_HEADER_tmz_offset 0
+#define SDMA_PKT_COPY_L2T_BROADCAST_HEADER_tmz_mask   0x00000001
+#define SDMA_PKT_COPY_L2T_BROADCAST_HEADER_tmz_shift  18
+#define SDMA_PKT_COPY_L2T_BROADCAST_HEADER_TMZ(x) (((x) & SDMA_PKT_COPY_L2T_BROADCAST_HEADER_tmz_mask) << SDMA_PKT_COPY_L2T_BROADCAST_HEADER_tmz_shift)
+
+/*define for cpv field*/
+#define SDMA_PKT_COPY_L2T_BROADCAST_HEADER_cpv_offset 0
+#define SDMA_PKT_COPY_L2T_BROADCAST_HEADER_cpv_mask   0x00000001
+#define SDMA_PKT_COPY_L2T_BROADCAST_HEADER_cpv_shift  19
+#define SDMA_PKT_COPY_L2T_BROADCAST_HEADER_CPV(x) (((x) & SDMA_PKT_COPY_L2T_BROADCAST_HEADER_cpv_mask) << SDMA_PKT_COPY_L2T_BROADCAST_HEADER_cpv_shift)
+
+/*define for videocopy field*/
+#define SDMA_PKT_COPY_L2T_BROADCAST_HEADER_videocopy_offset 0
+#define SDMA_PKT_COPY_L2T_BROADCAST_HEADER_videocopy_mask   0x00000001
+#define SDMA_PKT_COPY_L2T_BROADCAST_HEADER_videocopy_shift  26
+#define SDMA_PKT_COPY_L2T_BROADCAST_HEADER_VIDEOCOPY(x) (((x) & SDMA_PKT_COPY_L2T_BROADCAST_HEADER_videocopy_mask) << SDMA_PKT_COPY_L2T_BROADCAST_HEADER_videocopy_shift)
+
+/*define for broadcast field*/
+#define SDMA_PKT_COPY_L2T_BROADCAST_HEADER_broadcast_offset 0
+#define SDMA_PKT_COPY_L2T_BROADCAST_HEADER_broadcast_mask   0x00000001
+#define SDMA_PKT_COPY_L2T_BROADCAST_HEADER_broadcast_shift  27
+#define SDMA_PKT_COPY_L2T_BROADCAST_HEADER_BROADCAST(x) (((x) & SDMA_PKT_COPY_L2T_BROADCAST_HEADER_broadcast_mask) << SDMA_PKT_COPY_L2T_BROADCAST_HEADER_broadcast_shift)
+
+/*define for TILED_ADDR_LO_0 word*/
+/*define for tiled_addr0_31_0 field*/
+#define SDMA_PKT_COPY_L2T_BROADCAST_TILED_ADDR_LO_0_tiled_addr0_31_0_offset 1
+#define SDMA_PKT_COPY_L2T_BROADCAST_TILED_ADDR_LO_0_tiled_addr0_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_L2T_BROADCAST_TILED_ADDR_LO_0_tiled_addr0_31_0_shift  0
+#define SDMA_PKT_COPY_L2T_BROADCAST_TILED_ADDR_LO_0_TILED_ADDR0_31_0(x) (((x) & SDMA_PKT_COPY_L2T_BROADCAST_TILED_ADDR_LO_0_tiled_addr0_31_0_mask) << SDMA_PKT_COPY_L2T_BROADCAST_TILED_ADDR_LO_0_tiled_addr0_31_0_shift)
+
+/*define for TILED_ADDR_HI_0 word*/
+/*define for tiled_addr0_63_32 field*/
+#define SDMA_PKT_COPY_L2T_BROADCAST_TILED_ADDR_HI_0_tiled_addr0_63_32_offset 2
+#define SDMA_PKT_COPY_L2T_BROADCAST_TILED_ADDR_HI_0_tiled_addr0_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_L2T_BROADCAST_TILED_ADDR_HI_0_tiled_addr0_63_32_shift  0
+#define SDMA_PKT_COPY_L2T_BROADCAST_TILED_ADDR_HI_0_TILED_ADDR0_63_32(x) (((x) & SDMA_PKT_COPY_L2T_BROADCAST_TILED_ADDR_HI_0_tiled_addr0_63_32_mask) << SDMA_PKT_COPY_L2T_BROADCAST_TILED_ADDR_HI_0_tiled_addr0_63_32_shift)
+
+/*define for TILED_ADDR_LO_1 word*/
+/*define for tiled_addr1_31_0 field*/
+#define SDMA_PKT_COPY_L2T_BROADCAST_TILED_ADDR_LO_1_tiled_addr1_31_0_offset 3
+#define SDMA_PKT_COPY_L2T_BROADCAST_TILED_ADDR_LO_1_tiled_addr1_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_L2T_BROADCAST_TILED_ADDR_LO_1_tiled_addr1_31_0_shift  0
+#define SDMA_PKT_COPY_L2T_BROADCAST_TILED_ADDR_LO_1_TILED_ADDR1_31_0(x) (((x) & SDMA_PKT_COPY_L2T_BROADCAST_TILED_ADDR_LO_1_tiled_addr1_31_0_mask) << SDMA_PKT_COPY_L2T_BROADCAST_TILED_ADDR_LO_1_tiled_addr1_31_0_shift)
+
+/*define for TILED_ADDR_HI_1 word*/
+/*define for tiled_addr1_63_32 field*/
+#define SDMA_PKT_COPY_L2T_BROADCAST_TILED_ADDR_HI_1_tiled_addr1_63_32_offset 4
+#define SDMA_PKT_COPY_L2T_BROADCAST_TILED_ADDR_HI_1_tiled_addr1_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_L2T_BROADCAST_TILED_ADDR_HI_1_tiled_addr1_63_32_shift  0
+#define SDMA_PKT_COPY_L2T_BROADCAST_TILED_ADDR_HI_1_TILED_ADDR1_63_32(x) (((x) & SDMA_PKT_COPY_L2T_BROADCAST_TILED_ADDR_HI_1_tiled_addr1_63_32_mask) << SDMA_PKT_COPY_L2T_BROADCAST_TILED_ADDR_HI_1_tiled_addr1_63_32_shift)
+
+/*define for DW_5 word*/
+/*define for width field*/
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_5_width_offset 5
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_5_width_mask   0x00003FFF
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_5_width_shift  0
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_5_WIDTH(x) (((x) & SDMA_PKT_COPY_L2T_BROADCAST_DW_5_width_mask) << SDMA_PKT_COPY_L2T_BROADCAST_DW_5_width_shift)
+
+/*define for DW_6 word*/
+/*define for height field*/
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_6_height_offset 6
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_6_height_mask   0x00003FFF
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_6_height_shift  0
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_6_HEIGHT(x) (((x) & SDMA_PKT_COPY_L2T_BROADCAST_DW_6_height_mask) << SDMA_PKT_COPY_L2T_BROADCAST_DW_6_height_shift)
+
+/*define for depth field*/
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_6_depth_offset 6
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_6_depth_mask   0x00001FFF
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_6_depth_shift  16
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_6_DEPTH(x) (((x) & SDMA_PKT_COPY_L2T_BROADCAST_DW_6_depth_mask) << SDMA_PKT_COPY_L2T_BROADCAST_DW_6_depth_shift)
+
+/*define for DW_7 word*/
+/*define for element_size field*/
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_7_element_size_offset 7
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_7_element_size_mask   0x00000007
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_7_element_size_shift  0
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_7_ELEMENT_SIZE(x) (((x) & SDMA_PKT_COPY_L2T_BROADCAST_DW_7_element_size_mask) << SDMA_PKT_COPY_L2T_BROADCAST_DW_7_element_size_shift)
+
+/*define for swizzle_mode field*/
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_7_swizzle_mode_offset 7
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_7_swizzle_mode_mask   0x0000001F
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_7_swizzle_mode_shift  3
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_7_SWIZZLE_MODE(x) (((x) & SDMA_PKT_COPY_L2T_BROADCAST_DW_7_swizzle_mode_mask) << SDMA_PKT_COPY_L2T_BROADCAST_DW_7_swizzle_mode_shift)
+
+/*define for dimension field*/
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_7_dimension_offset 7
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_7_dimension_mask   0x00000003
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_7_dimension_shift  9
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_7_DIMENSION(x) (((x) & SDMA_PKT_COPY_L2T_BROADCAST_DW_7_dimension_mask) << SDMA_PKT_COPY_L2T_BROADCAST_DW_7_dimension_shift)
+
+/*define for mip_max field*/
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_7_mip_max_offset 7
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_7_mip_max_mask   0x0000000F
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_7_mip_max_shift  16
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_7_MIP_MAX(x) (((x) & SDMA_PKT_COPY_L2T_BROADCAST_DW_7_mip_max_mask) << SDMA_PKT_COPY_L2T_BROADCAST_DW_7_mip_max_shift)
+
+/*define for DW_8 word*/
+/*define for x field*/
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_8_x_offset 8
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_8_x_mask   0x00003FFF
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_8_x_shift  0
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_8_X(x) (((x) & SDMA_PKT_COPY_L2T_BROADCAST_DW_8_x_mask) << SDMA_PKT_COPY_L2T_BROADCAST_DW_8_x_shift)
+
+/*define for y field*/
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_8_y_offset 8
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_8_y_mask   0x00003FFF
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_8_y_shift  16
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_8_Y(x) (((x) & SDMA_PKT_COPY_L2T_BROADCAST_DW_8_y_mask) << SDMA_PKT_COPY_L2T_BROADCAST_DW_8_y_shift)
+
+/*define for DW_9 word*/
+/*define for z field*/
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_9_z_offset 9
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_9_z_mask   0x00001FFF
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_9_z_shift  0
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_9_Z(x) (((x) & SDMA_PKT_COPY_L2T_BROADCAST_DW_9_z_mask) << SDMA_PKT_COPY_L2T_BROADCAST_DW_9_z_shift)
+
+/*define for DW_10 word*/
+/*define for dst2_sw field*/
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_10_dst2_sw_offset 10
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_10_dst2_sw_mask   0x00000003
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_10_dst2_sw_shift  8
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_10_DST2_SW(x) (((x) & SDMA_PKT_COPY_L2T_BROADCAST_DW_10_dst2_sw_mask) << SDMA_PKT_COPY_L2T_BROADCAST_DW_10_dst2_sw_shift)
+
+/*define for dst2_cache_policy field*/
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_10_dst2_cache_policy_offset 10
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_10_dst2_cache_policy_mask   0x00000007
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_10_dst2_cache_policy_shift  10
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_10_DST2_CACHE_POLICY(x) (((x) & SDMA_PKT_COPY_L2T_BROADCAST_DW_10_dst2_cache_policy_mask) << SDMA_PKT_COPY_L2T_BROADCAST_DW_10_dst2_cache_policy_shift)
+
+/*define for linear_sw field*/
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_10_linear_sw_offset 10
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_10_linear_sw_mask   0x00000003
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_10_linear_sw_shift  16
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_10_LINEAR_SW(x) (((x) & SDMA_PKT_COPY_L2T_BROADCAST_DW_10_linear_sw_mask) << SDMA_PKT_COPY_L2T_BROADCAST_DW_10_linear_sw_shift)
+
+/*define for linear_cache_policy field*/
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_10_linear_cache_policy_offset 10
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_10_linear_cache_policy_mask   0x00000007
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_10_linear_cache_policy_shift  18
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_10_LINEAR_CACHE_POLICY(x) (((x) & SDMA_PKT_COPY_L2T_BROADCAST_DW_10_linear_cache_policy_mask) << SDMA_PKT_COPY_L2T_BROADCAST_DW_10_linear_cache_policy_shift)
+
+/*define for tile_sw field*/
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_10_tile_sw_offset 10
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_10_tile_sw_mask   0x00000003
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_10_tile_sw_shift  24
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_10_TILE_SW(x) (((x) & SDMA_PKT_COPY_L2T_BROADCAST_DW_10_tile_sw_mask) << SDMA_PKT_COPY_L2T_BROADCAST_DW_10_tile_sw_shift)
+
+/*define for tile_cache_policy field*/
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_10_tile_cache_policy_offset 10
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_10_tile_cache_policy_mask   0x00000007
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_10_tile_cache_policy_shift  26
+#define SDMA_PKT_COPY_L2T_BROADCAST_DW_10_TILE_CACHE_POLICY(x) (((x) & SDMA_PKT_COPY_L2T_BROADCAST_DW_10_tile_cache_policy_mask) << SDMA_PKT_COPY_L2T_BROADCAST_DW_10_tile_cache_policy_shift)
+
+/*define for LINEAR_ADDR_LO word*/
+/*define for linear_addr_31_0 field*/
+#define SDMA_PKT_COPY_L2T_BROADCAST_LINEAR_ADDR_LO_linear_addr_31_0_offset 11
+#define SDMA_PKT_COPY_L2T_BROADCAST_LINEAR_ADDR_LO_linear_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_L2T_BROADCAST_LINEAR_ADDR_LO_linear_addr_31_0_shift  0
+#define SDMA_PKT_COPY_L2T_BROADCAST_LINEAR_ADDR_LO_LINEAR_ADDR_31_0(x) (((x) & SDMA_PKT_COPY_L2T_BROADCAST_LINEAR_ADDR_LO_linear_addr_31_0_mask) << SDMA_PKT_COPY_L2T_BROADCAST_LINEAR_ADDR_LO_linear_addr_31_0_shift)
+
+/*define for LINEAR_ADDR_HI word*/
+/*define for linear_addr_63_32 field*/
+#define SDMA_PKT_COPY_L2T_BROADCAST_LINEAR_ADDR_HI_linear_addr_63_32_offset 12
+#define SDMA_PKT_COPY_L2T_BROADCAST_LINEAR_ADDR_HI_linear_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_L2T_BROADCAST_LINEAR_ADDR_HI_linear_addr_63_32_shift  0
+#define SDMA_PKT_COPY_L2T_BROADCAST_LINEAR_ADDR_HI_LINEAR_ADDR_63_32(x) (((x) & SDMA_PKT_COPY_L2T_BROADCAST_LINEAR_ADDR_HI_linear_addr_63_32_mask) << SDMA_PKT_COPY_L2T_BROADCAST_LINEAR_ADDR_HI_linear_addr_63_32_shift)
+
+/*define for LINEAR_PITCH word*/
+/*define for linear_pitch field*/
+#define SDMA_PKT_COPY_L2T_BROADCAST_LINEAR_PITCH_linear_pitch_offset 13
+#define SDMA_PKT_COPY_L2T_BROADCAST_LINEAR_PITCH_linear_pitch_mask   0x0007FFFF
+#define SDMA_PKT_COPY_L2T_BROADCAST_LINEAR_PITCH_linear_pitch_shift  0
+#define SDMA_PKT_COPY_L2T_BROADCAST_LINEAR_PITCH_LINEAR_PITCH(x) (((x) & SDMA_PKT_COPY_L2T_BROADCAST_LINEAR_PITCH_linear_pitch_mask) << SDMA_PKT_COPY_L2T_BROADCAST_LINEAR_PITCH_linear_pitch_shift)
+
+/*define for LINEAR_SLICE_PITCH word*/
+/*define for linear_slice_pitch field*/
+#define SDMA_PKT_COPY_L2T_BROADCAST_LINEAR_SLICE_PITCH_linear_slice_pitch_offset 14
+#define SDMA_PKT_COPY_L2T_BROADCAST_LINEAR_SLICE_PITCH_linear_slice_pitch_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_L2T_BROADCAST_LINEAR_SLICE_PITCH_linear_slice_pitch_shift  0
+#define SDMA_PKT_COPY_L2T_BROADCAST_LINEAR_SLICE_PITCH_LINEAR_SLICE_PITCH(x) (((x) & SDMA_PKT_COPY_L2T_BROADCAST_LINEAR_SLICE_PITCH_linear_slice_pitch_mask) << SDMA_PKT_COPY_L2T_BROADCAST_LINEAR_SLICE_PITCH_linear_slice_pitch_shift)
+
+/*define for COUNT word*/
+/*define for count field*/
+#define SDMA_PKT_COPY_L2T_BROADCAST_COUNT_count_offset 15
+#define SDMA_PKT_COPY_L2T_BROADCAST_COUNT_count_mask   0x3FFFFFFF
+#define SDMA_PKT_COPY_L2T_BROADCAST_COUNT_count_shift  0
+#define SDMA_PKT_COPY_L2T_BROADCAST_COUNT_COUNT(x) (((x) & SDMA_PKT_COPY_L2T_BROADCAST_COUNT_count_mask) << SDMA_PKT_COPY_L2T_BROADCAST_COUNT_count_shift)
+
+
+/*
+** Definitions for SDMA_PKT_COPY_T2T packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_COPY_T2T_HEADER_op_offset 0
+#define SDMA_PKT_COPY_T2T_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_COPY_T2T_HEADER_op_shift  0
+#define SDMA_PKT_COPY_T2T_HEADER_OP(x) (((x) & SDMA_PKT_COPY_T2T_HEADER_op_mask) << SDMA_PKT_COPY_T2T_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_COPY_T2T_HEADER_sub_op_offset 0
+#define SDMA_PKT_COPY_T2T_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_COPY_T2T_HEADER_sub_op_shift  8
+#define SDMA_PKT_COPY_T2T_HEADER_SUB_OP(x) (((x) & SDMA_PKT_COPY_T2T_HEADER_sub_op_mask) << SDMA_PKT_COPY_T2T_HEADER_sub_op_shift)
+
+/*define for tmz field*/
+#define SDMA_PKT_COPY_T2T_HEADER_tmz_offset 0
+#define SDMA_PKT_COPY_T2T_HEADER_tmz_mask   0x00000001
+#define SDMA_PKT_COPY_T2T_HEADER_tmz_shift  18
+#define SDMA_PKT_COPY_T2T_HEADER_TMZ(x) (((x) & SDMA_PKT_COPY_T2T_HEADER_tmz_mask) << SDMA_PKT_COPY_T2T_HEADER_tmz_shift)
+
+/*define for dcc field*/
+#define SDMA_PKT_COPY_T2T_HEADER_dcc_offset 0
+#define SDMA_PKT_COPY_T2T_HEADER_dcc_mask   0x00000001
+#define SDMA_PKT_COPY_T2T_HEADER_dcc_shift  19
+#define SDMA_PKT_COPY_T2T_HEADER_DCC(x) (((x) & SDMA_PKT_COPY_T2T_HEADER_dcc_mask) << SDMA_PKT_COPY_T2T_HEADER_dcc_shift)
+
+/*define for cpv field*/
+#define SDMA_PKT_COPY_T2T_HEADER_cpv_offset 0
+#define SDMA_PKT_COPY_T2T_HEADER_cpv_mask   0x00000001
+#define SDMA_PKT_COPY_T2T_HEADER_cpv_shift  28
+#define SDMA_PKT_COPY_T2T_HEADER_CPV(x) (((x) & SDMA_PKT_COPY_T2T_HEADER_cpv_mask) << SDMA_PKT_COPY_T2T_HEADER_cpv_shift)
+
+/*define for dcc_dir field*/
+#define SDMA_PKT_COPY_T2T_HEADER_dcc_dir_offset 0
+#define SDMA_PKT_COPY_T2T_HEADER_dcc_dir_mask   0x00000001
+#define SDMA_PKT_COPY_T2T_HEADER_dcc_dir_shift  31
+#define SDMA_PKT_COPY_T2T_HEADER_DCC_DIR(x) (((x) & SDMA_PKT_COPY_T2T_HEADER_dcc_dir_mask) << SDMA_PKT_COPY_T2T_HEADER_dcc_dir_shift)
+
+/*define for SRC_ADDR_LO word*/
+/*define for src_addr_31_0 field*/
+#define SDMA_PKT_COPY_T2T_SRC_ADDR_LO_src_addr_31_0_offset 1
+#define SDMA_PKT_COPY_T2T_SRC_ADDR_LO_src_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_T2T_SRC_ADDR_LO_src_addr_31_0_shift  0
+#define SDMA_PKT_COPY_T2T_SRC_ADDR_LO_SRC_ADDR_31_0(x) (((x) & SDMA_PKT_COPY_T2T_SRC_ADDR_LO_src_addr_31_0_mask) << SDMA_PKT_COPY_T2T_SRC_ADDR_LO_src_addr_31_0_shift)
+
+/*define for SRC_ADDR_HI word*/
+/*define for src_addr_63_32 field*/
+#define SDMA_PKT_COPY_T2T_SRC_ADDR_HI_src_addr_63_32_offset 2
+#define SDMA_PKT_COPY_T2T_SRC_ADDR_HI_src_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_T2T_SRC_ADDR_HI_src_addr_63_32_shift  0
+#define SDMA_PKT_COPY_T2T_SRC_ADDR_HI_SRC_ADDR_63_32(x) (((x) & SDMA_PKT_COPY_T2T_SRC_ADDR_HI_src_addr_63_32_mask) << SDMA_PKT_COPY_T2T_SRC_ADDR_HI_src_addr_63_32_shift)
+
+/*define for DW_3 word*/
+/*define for src_x field*/
+#define SDMA_PKT_COPY_T2T_DW_3_src_x_offset 3
+#define SDMA_PKT_COPY_T2T_DW_3_src_x_mask   0x00003FFF
+#define SDMA_PKT_COPY_T2T_DW_3_src_x_shift  0
+#define SDMA_PKT_COPY_T2T_DW_3_SRC_X(x) (((x) & SDMA_PKT_COPY_T2T_DW_3_src_x_mask) << SDMA_PKT_COPY_T2T_DW_3_src_x_shift)
+
+/*define for src_y field*/
+#define SDMA_PKT_COPY_T2T_DW_3_src_y_offset 3
+#define SDMA_PKT_COPY_T2T_DW_3_src_y_mask   0x00003FFF
+#define SDMA_PKT_COPY_T2T_DW_3_src_y_shift  16
+#define SDMA_PKT_COPY_T2T_DW_3_SRC_Y(x) (((x) & SDMA_PKT_COPY_T2T_DW_3_src_y_mask) << SDMA_PKT_COPY_T2T_DW_3_src_y_shift)
+
+/*define for DW_4 word*/
+/*define for src_z field*/
+#define SDMA_PKT_COPY_T2T_DW_4_src_z_offset 4
+#define SDMA_PKT_COPY_T2T_DW_4_src_z_mask   0x00001FFF
+#define SDMA_PKT_COPY_T2T_DW_4_src_z_shift  0
+#define SDMA_PKT_COPY_T2T_DW_4_SRC_Z(x) (((x) & SDMA_PKT_COPY_T2T_DW_4_src_z_mask) << SDMA_PKT_COPY_T2T_DW_4_src_z_shift)
+
+/*define for src_width field*/
+#define SDMA_PKT_COPY_T2T_DW_4_src_width_offset 4
+#define SDMA_PKT_COPY_T2T_DW_4_src_width_mask   0x00003FFF
+#define SDMA_PKT_COPY_T2T_DW_4_src_width_shift  16
+#define SDMA_PKT_COPY_T2T_DW_4_SRC_WIDTH(x) (((x) & SDMA_PKT_COPY_T2T_DW_4_src_width_mask) << SDMA_PKT_COPY_T2T_DW_4_src_width_shift)
+
+/*define for DW_5 word*/
+/*define for src_height field*/
+#define SDMA_PKT_COPY_T2T_DW_5_src_height_offset 5
+#define SDMA_PKT_COPY_T2T_DW_5_src_height_mask   0x00003FFF
+#define SDMA_PKT_COPY_T2T_DW_5_src_height_shift  0
+#define SDMA_PKT_COPY_T2T_DW_5_SRC_HEIGHT(x) (((x) & SDMA_PKT_COPY_T2T_DW_5_src_height_mask) << SDMA_PKT_COPY_T2T_DW_5_src_height_shift)
+
+/*define for src_depth field*/
+#define SDMA_PKT_COPY_T2T_DW_5_src_depth_offset 5
+#define SDMA_PKT_COPY_T2T_DW_5_src_depth_mask   0x00001FFF
+#define SDMA_PKT_COPY_T2T_DW_5_src_depth_shift  16
+#define SDMA_PKT_COPY_T2T_DW_5_SRC_DEPTH(x) (((x) & SDMA_PKT_COPY_T2T_DW_5_src_depth_mask) << SDMA_PKT_COPY_T2T_DW_5_src_depth_shift)
+
+/*define for DW_6 word*/
+/*define for src_element_size field*/
+#define SDMA_PKT_COPY_T2T_DW_6_src_element_size_offset 6
+#define SDMA_PKT_COPY_T2T_DW_6_src_element_size_mask   0x00000007
+#define SDMA_PKT_COPY_T2T_DW_6_src_element_size_shift  0
+#define SDMA_PKT_COPY_T2T_DW_6_SRC_ELEMENT_SIZE(x) (((x) & SDMA_PKT_COPY_T2T_DW_6_src_element_size_mask) << SDMA_PKT_COPY_T2T_DW_6_src_element_size_shift)
+
+/*define for src_swizzle_mode field*/
+#define SDMA_PKT_COPY_T2T_DW_6_src_swizzle_mode_offset 6
+#define SDMA_PKT_COPY_T2T_DW_6_src_swizzle_mode_mask   0x0000001F
+#define SDMA_PKT_COPY_T2T_DW_6_src_swizzle_mode_shift  3
+#define SDMA_PKT_COPY_T2T_DW_6_SRC_SWIZZLE_MODE(x) (((x) & SDMA_PKT_COPY_T2T_DW_6_src_swizzle_mode_mask) << SDMA_PKT_COPY_T2T_DW_6_src_swizzle_mode_shift)
+
+/*define for src_dimension field*/
+#define SDMA_PKT_COPY_T2T_DW_6_src_dimension_offset 6
+#define SDMA_PKT_COPY_T2T_DW_6_src_dimension_mask   0x00000003
+#define SDMA_PKT_COPY_T2T_DW_6_src_dimension_shift  9
+#define SDMA_PKT_COPY_T2T_DW_6_SRC_DIMENSION(x) (((x) & SDMA_PKT_COPY_T2T_DW_6_src_dimension_mask) << SDMA_PKT_COPY_T2T_DW_6_src_dimension_shift)
+
+/*define for src_mip_max field*/
+#define SDMA_PKT_COPY_T2T_DW_6_src_mip_max_offset 6
+#define SDMA_PKT_COPY_T2T_DW_6_src_mip_max_mask   0x0000000F
+#define SDMA_PKT_COPY_T2T_DW_6_src_mip_max_shift  16
+#define SDMA_PKT_COPY_T2T_DW_6_SRC_MIP_MAX(x) (((x) & SDMA_PKT_COPY_T2T_DW_6_src_mip_max_mask) << SDMA_PKT_COPY_T2T_DW_6_src_mip_max_shift)
+
+/*define for src_mip_id field*/
+#define SDMA_PKT_COPY_T2T_DW_6_src_mip_id_offset 6
+#define SDMA_PKT_COPY_T2T_DW_6_src_mip_id_mask   0x0000000F
+#define SDMA_PKT_COPY_T2T_DW_6_src_mip_id_shift  20
+#define SDMA_PKT_COPY_T2T_DW_6_SRC_MIP_ID(x) (((x) & SDMA_PKT_COPY_T2T_DW_6_src_mip_id_mask) << SDMA_PKT_COPY_T2T_DW_6_src_mip_id_shift)
+
+/*define for DST_ADDR_LO word*/
+/*define for dst_addr_31_0 field*/
+#define SDMA_PKT_COPY_T2T_DST_ADDR_LO_dst_addr_31_0_offset 7
+#define SDMA_PKT_COPY_T2T_DST_ADDR_LO_dst_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_T2T_DST_ADDR_LO_dst_addr_31_0_shift  0
+#define SDMA_PKT_COPY_T2T_DST_ADDR_LO_DST_ADDR_31_0(x) (((x) & SDMA_PKT_COPY_T2T_DST_ADDR_LO_dst_addr_31_0_mask) << SDMA_PKT_COPY_T2T_DST_ADDR_LO_dst_addr_31_0_shift)
+
+/*define for DST_ADDR_HI word*/
+/*define for dst_addr_63_32 field*/
+#define SDMA_PKT_COPY_T2T_DST_ADDR_HI_dst_addr_63_32_offset 8
+#define SDMA_PKT_COPY_T2T_DST_ADDR_HI_dst_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_T2T_DST_ADDR_HI_dst_addr_63_32_shift  0
+#define SDMA_PKT_COPY_T2T_DST_ADDR_HI_DST_ADDR_63_32(x) (((x) & SDMA_PKT_COPY_T2T_DST_ADDR_HI_dst_addr_63_32_mask) << SDMA_PKT_COPY_T2T_DST_ADDR_HI_dst_addr_63_32_shift)
+
+/*define for DW_9 word*/
+/*define for dst_x field*/
+#define SDMA_PKT_COPY_T2T_DW_9_dst_x_offset 9
+#define SDMA_PKT_COPY_T2T_DW_9_dst_x_mask   0x00003FFF
+#define SDMA_PKT_COPY_T2T_DW_9_dst_x_shift  0
+#define SDMA_PKT_COPY_T2T_DW_9_DST_X(x) (((x) & SDMA_PKT_COPY_T2T_DW_9_dst_x_mask) << SDMA_PKT_COPY_T2T_DW_9_dst_x_shift)
+
+/*define for dst_y field*/
+#define SDMA_PKT_COPY_T2T_DW_9_dst_y_offset 9
+#define SDMA_PKT_COPY_T2T_DW_9_dst_y_mask   0x00003FFF
+#define SDMA_PKT_COPY_T2T_DW_9_dst_y_shift  16
+#define SDMA_PKT_COPY_T2T_DW_9_DST_Y(x) (((x) & SDMA_PKT_COPY_T2T_DW_9_dst_y_mask) << SDMA_PKT_COPY_T2T_DW_9_dst_y_shift)
+
+/*define for DW_10 word*/
+/*define for dst_z field*/
+#define SDMA_PKT_COPY_T2T_DW_10_dst_z_offset 10
+#define SDMA_PKT_COPY_T2T_DW_10_dst_z_mask   0x00001FFF
+#define SDMA_PKT_COPY_T2T_DW_10_dst_z_shift  0
+#define SDMA_PKT_COPY_T2T_DW_10_DST_Z(x) (((x) & SDMA_PKT_COPY_T2T_DW_10_dst_z_mask) << SDMA_PKT_COPY_T2T_DW_10_dst_z_shift)
+
+/*define for dst_width field*/
+#define SDMA_PKT_COPY_T2T_DW_10_dst_width_offset 10
+#define SDMA_PKT_COPY_T2T_DW_10_dst_width_mask   0x00003FFF
+#define SDMA_PKT_COPY_T2T_DW_10_dst_width_shift  16
+#define SDMA_PKT_COPY_T2T_DW_10_DST_WIDTH(x) (((x) & SDMA_PKT_COPY_T2T_DW_10_dst_width_mask) << SDMA_PKT_COPY_T2T_DW_10_dst_width_shift)
+
+/*define for DW_11 word*/
+/*define for dst_height field*/
+#define SDMA_PKT_COPY_T2T_DW_11_dst_height_offset 11
+#define SDMA_PKT_COPY_T2T_DW_11_dst_height_mask   0x00003FFF
+#define SDMA_PKT_COPY_T2T_DW_11_dst_height_shift  0
+#define SDMA_PKT_COPY_T2T_DW_11_DST_HEIGHT(x) (((x) & SDMA_PKT_COPY_T2T_DW_11_dst_height_mask) << SDMA_PKT_COPY_T2T_DW_11_dst_height_shift)
+
+/*define for dst_depth field*/
+#define SDMA_PKT_COPY_T2T_DW_11_dst_depth_offset 11
+#define SDMA_PKT_COPY_T2T_DW_11_dst_depth_mask   0x00001FFF
+#define SDMA_PKT_COPY_T2T_DW_11_dst_depth_shift  16
+#define SDMA_PKT_COPY_T2T_DW_11_DST_DEPTH(x) (((x) & SDMA_PKT_COPY_T2T_DW_11_dst_depth_mask) << SDMA_PKT_COPY_T2T_DW_11_dst_depth_shift)
+
+/*define for DW_12 word*/
+/*define for dst_element_size field*/
+#define SDMA_PKT_COPY_T2T_DW_12_dst_element_size_offset 12
+#define SDMA_PKT_COPY_T2T_DW_12_dst_element_size_mask   0x00000007
+#define SDMA_PKT_COPY_T2T_DW_12_dst_element_size_shift  0
+#define SDMA_PKT_COPY_T2T_DW_12_DST_ELEMENT_SIZE(x) (((x) & SDMA_PKT_COPY_T2T_DW_12_dst_element_size_mask) << SDMA_PKT_COPY_T2T_DW_12_dst_element_size_shift)
+
+/*define for dst_swizzle_mode field*/
+#define SDMA_PKT_COPY_T2T_DW_12_dst_swizzle_mode_offset 12
+#define SDMA_PKT_COPY_T2T_DW_12_dst_swizzle_mode_mask   0x0000001F
+#define SDMA_PKT_COPY_T2T_DW_12_dst_swizzle_mode_shift  3
+#define SDMA_PKT_COPY_T2T_DW_12_DST_SWIZZLE_MODE(x) (((x) & SDMA_PKT_COPY_T2T_DW_12_dst_swizzle_mode_mask) << SDMA_PKT_COPY_T2T_DW_12_dst_swizzle_mode_shift)
+
+/*define for dst_dimension field*/
+#define SDMA_PKT_COPY_T2T_DW_12_dst_dimension_offset 12
+#define SDMA_PKT_COPY_T2T_DW_12_dst_dimension_mask   0x00000003
+#define SDMA_PKT_COPY_T2T_DW_12_dst_dimension_shift  9
+#define SDMA_PKT_COPY_T2T_DW_12_DST_DIMENSION(x) (((x) & SDMA_PKT_COPY_T2T_DW_12_dst_dimension_mask) << SDMA_PKT_COPY_T2T_DW_12_dst_dimension_shift)
+
+/*define for dst_mip_max field*/
+#define SDMA_PKT_COPY_T2T_DW_12_dst_mip_max_offset 12
+#define SDMA_PKT_COPY_T2T_DW_12_dst_mip_max_mask   0x0000000F
+#define SDMA_PKT_COPY_T2T_DW_12_dst_mip_max_shift  16
+#define SDMA_PKT_COPY_T2T_DW_12_DST_MIP_MAX(x) (((x) & SDMA_PKT_COPY_T2T_DW_12_dst_mip_max_mask) << SDMA_PKT_COPY_T2T_DW_12_dst_mip_max_shift)
+
+/*define for dst_mip_id field*/
+#define SDMA_PKT_COPY_T2T_DW_12_dst_mip_id_offset 12
+#define SDMA_PKT_COPY_T2T_DW_12_dst_mip_id_mask   0x0000000F
+#define SDMA_PKT_COPY_T2T_DW_12_dst_mip_id_shift  20
+#define SDMA_PKT_COPY_T2T_DW_12_DST_MIP_ID(x) (((x) & SDMA_PKT_COPY_T2T_DW_12_dst_mip_id_mask) << SDMA_PKT_COPY_T2T_DW_12_dst_mip_id_shift)
+
+/*define for DW_13 word*/
+/*define for rect_x field*/
+#define SDMA_PKT_COPY_T2T_DW_13_rect_x_offset 13
+#define SDMA_PKT_COPY_T2T_DW_13_rect_x_mask   0x00003FFF
+#define SDMA_PKT_COPY_T2T_DW_13_rect_x_shift  0
+#define SDMA_PKT_COPY_T2T_DW_13_RECT_X(x) (((x) & SDMA_PKT_COPY_T2T_DW_13_rect_x_mask) << SDMA_PKT_COPY_T2T_DW_13_rect_x_shift)
+
+/*define for rect_y field*/
+#define SDMA_PKT_COPY_T2T_DW_13_rect_y_offset 13
+#define SDMA_PKT_COPY_T2T_DW_13_rect_y_mask   0x00003FFF
+#define SDMA_PKT_COPY_T2T_DW_13_rect_y_shift  16
+#define SDMA_PKT_COPY_T2T_DW_13_RECT_Y(x) (((x) & SDMA_PKT_COPY_T2T_DW_13_rect_y_mask) << SDMA_PKT_COPY_T2T_DW_13_rect_y_shift)
+
+/*define for DW_14 word*/
+/*define for rect_z field*/
+#define SDMA_PKT_COPY_T2T_DW_14_rect_z_offset 14
+#define SDMA_PKT_COPY_T2T_DW_14_rect_z_mask   0x00001FFF
+#define SDMA_PKT_COPY_T2T_DW_14_rect_z_shift  0
+#define SDMA_PKT_COPY_T2T_DW_14_RECT_Z(x) (((x) & SDMA_PKT_COPY_T2T_DW_14_rect_z_mask) << SDMA_PKT_COPY_T2T_DW_14_rect_z_shift)
+
+/*define for dst_sw field*/
+#define SDMA_PKT_COPY_T2T_DW_14_dst_sw_offset 14
+#define SDMA_PKT_COPY_T2T_DW_14_dst_sw_mask   0x00000003
+#define SDMA_PKT_COPY_T2T_DW_14_dst_sw_shift  16
+#define SDMA_PKT_COPY_T2T_DW_14_DST_SW(x) (((x) & SDMA_PKT_COPY_T2T_DW_14_dst_sw_mask) << SDMA_PKT_COPY_T2T_DW_14_dst_sw_shift)
+
+/*define for dst_cache_policy field*/
+#define SDMA_PKT_COPY_T2T_DW_14_dst_cache_policy_offset 14
+#define SDMA_PKT_COPY_T2T_DW_14_dst_cache_policy_mask   0x00000007
+#define SDMA_PKT_COPY_T2T_DW_14_dst_cache_policy_shift  18
+#define SDMA_PKT_COPY_T2T_DW_14_DST_CACHE_POLICY(x) (((x) & SDMA_PKT_COPY_T2T_DW_14_dst_cache_policy_mask) << SDMA_PKT_COPY_T2T_DW_14_dst_cache_policy_shift)
+
+/*define for src_sw field*/
+#define SDMA_PKT_COPY_T2T_DW_14_src_sw_offset 14
+#define SDMA_PKT_COPY_T2T_DW_14_src_sw_mask   0x00000003
+#define SDMA_PKT_COPY_T2T_DW_14_src_sw_shift  24
+#define SDMA_PKT_COPY_T2T_DW_14_SRC_SW(x) (((x) & SDMA_PKT_COPY_T2T_DW_14_src_sw_mask) << SDMA_PKT_COPY_T2T_DW_14_src_sw_shift)
+
+/*define for src_cache_policy field*/
+#define SDMA_PKT_COPY_T2T_DW_14_src_cache_policy_offset 14
+#define SDMA_PKT_COPY_T2T_DW_14_src_cache_policy_mask   0x00000007
+#define SDMA_PKT_COPY_T2T_DW_14_src_cache_policy_shift  26
+#define SDMA_PKT_COPY_T2T_DW_14_SRC_CACHE_POLICY(x) (((x) & SDMA_PKT_COPY_T2T_DW_14_src_cache_policy_mask) << SDMA_PKT_COPY_T2T_DW_14_src_cache_policy_shift)
+
+/*define for META_ADDR_LO word*/
+/*define for meta_addr_31_0 field*/
+#define SDMA_PKT_COPY_T2T_META_ADDR_LO_meta_addr_31_0_offset 15
+#define SDMA_PKT_COPY_T2T_META_ADDR_LO_meta_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_T2T_META_ADDR_LO_meta_addr_31_0_shift  0
+#define SDMA_PKT_COPY_T2T_META_ADDR_LO_META_ADDR_31_0(x) (((x) & SDMA_PKT_COPY_T2T_META_ADDR_LO_meta_addr_31_0_mask) << SDMA_PKT_COPY_T2T_META_ADDR_LO_meta_addr_31_0_shift)
+
+/*define for META_ADDR_HI word*/
+/*define for meta_addr_63_32 field*/
+#define SDMA_PKT_COPY_T2T_META_ADDR_HI_meta_addr_63_32_offset 16
+#define SDMA_PKT_COPY_T2T_META_ADDR_HI_meta_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_T2T_META_ADDR_HI_meta_addr_63_32_shift  0
+#define SDMA_PKT_COPY_T2T_META_ADDR_HI_META_ADDR_63_32(x) (((x) & SDMA_PKT_COPY_T2T_META_ADDR_HI_meta_addr_63_32_mask) << SDMA_PKT_COPY_T2T_META_ADDR_HI_meta_addr_63_32_shift)
+
+/*define for META_CONFIG word*/
+/*define for data_format field*/
+#define SDMA_PKT_COPY_T2T_META_CONFIG_data_format_offset 17
+#define SDMA_PKT_COPY_T2T_META_CONFIG_data_format_mask   0x0000007F
+#define SDMA_PKT_COPY_T2T_META_CONFIG_data_format_shift  0
+#define SDMA_PKT_COPY_T2T_META_CONFIG_DATA_FORMAT(x) (((x) & SDMA_PKT_COPY_T2T_META_CONFIG_data_format_mask) << SDMA_PKT_COPY_T2T_META_CONFIG_data_format_shift)
+
+/*define for color_transform_disable field*/
+#define SDMA_PKT_COPY_T2T_META_CONFIG_color_transform_disable_offset 17
+#define SDMA_PKT_COPY_T2T_META_CONFIG_color_transform_disable_mask   0x00000001
+#define SDMA_PKT_COPY_T2T_META_CONFIG_color_transform_disable_shift  7
+#define SDMA_PKT_COPY_T2T_META_CONFIG_COLOR_TRANSFORM_DISABLE(x) (((x) & SDMA_PKT_COPY_T2T_META_CONFIG_color_transform_disable_mask) << SDMA_PKT_COPY_T2T_META_CONFIG_color_transform_disable_shift)
+
+/*define for alpha_is_on_msb field*/
+#define SDMA_PKT_COPY_T2T_META_CONFIG_alpha_is_on_msb_offset 17
+#define SDMA_PKT_COPY_T2T_META_CONFIG_alpha_is_on_msb_mask   0x00000001
+#define SDMA_PKT_COPY_T2T_META_CONFIG_alpha_is_on_msb_shift  8
+#define SDMA_PKT_COPY_T2T_META_CONFIG_ALPHA_IS_ON_MSB(x) (((x) & SDMA_PKT_COPY_T2T_META_CONFIG_alpha_is_on_msb_mask) << SDMA_PKT_COPY_T2T_META_CONFIG_alpha_is_on_msb_shift)
+
+/*define for number_type field*/
+#define SDMA_PKT_COPY_T2T_META_CONFIG_number_type_offset 17
+#define SDMA_PKT_COPY_T2T_META_CONFIG_number_type_mask   0x00000007
+#define SDMA_PKT_COPY_T2T_META_CONFIG_number_type_shift  9
+#define SDMA_PKT_COPY_T2T_META_CONFIG_NUMBER_TYPE(x) (((x) & SDMA_PKT_COPY_T2T_META_CONFIG_number_type_mask) << SDMA_PKT_COPY_T2T_META_CONFIG_number_type_shift)
+
+/*define for surface_type field*/
+#define SDMA_PKT_COPY_T2T_META_CONFIG_surface_type_offset 17
+#define SDMA_PKT_COPY_T2T_META_CONFIG_surface_type_mask   0x00000003
+#define SDMA_PKT_COPY_T2T_META_CONFIG_surface_type_shift  12
+#define SDMA_PKT_COPY_T2T_META_CONFIG_SURFACE_TYPE(x) (((x) & SDMA_PKT_COPY_T2T_META_CONFIG_surface_type_mask) << SDMA_PKT_COPY_T2T_META_CONFIG_surface_type_shift)
+
+/*define for meta_llc field*/
+#define SDMA_PKT_COPY_T2T_META_CONFIG_meta_llc_offset 17
+#define SDMA_PKT_COPY_T2T_META_CONFIG_meta_llc_mask   0x00000001
+#define SDMA_PKT_COPY_T2T_META_CONFIG_meta_llc_shift  14
+#define SDMA_PKT_COPY_T2T_META_CONFIG_META_LLC(x) (((x) & SDMA_PKT_COPY_T2T_META_CONFIG_meta_llc_mask) << SDMA_PKT_COPY_T2T_META_CONFIG_meta_llc_shift)
+
+/*define for max_comp_block_size field*/
+#define SDMA_PKT_COPY_T2T_META_CONFIG_max_comp_block_size_offset 17
+#define SDMA_PKT_COPY_T2T_META_CONFIG_max_comp_block_size_mask   0x00000003
+#define SDMA_PKT_COPY_T2T_META_CONFIG_max_comp_block_size_shift  24
+#define SDMA_PKT_COPY_T2T_META_CONFIG_MAX_COMP_BLOCK_SIZE(x) (((x) & SDMA_PKT_COPY_T2T_META_CONFIG_max_comp_block_size_mask) << SDMA_PKT_COPY_T2T_META_CONFIG_max_comp_block_size_shift)
+
+/*define for max_uncomp_block_size field*/
+#define SDMA_PKT_COPY_T2T_META_CONFIG_max_uncomp_block_size_offset 17
+#define SDMA_PKT_COPY_T2T_META_CONFIG_max_uncomp_block_size_mask   0x00000003
+#define SDMA_PKT_COPY_T2T_META_CONFIG_max_uncomp_block_size_shift  26
+#define SDMA_PKT_COPY_T2T_META_CONFIG_MAX_UNCOMP_BLOCK_SIZE(x) (((x) & SDMA_PKT_COPY_T2T_META_CONFIG_max_uncomp_block_size_mask) << SDMA_PKT_COPY_T2T_META_CONFIG_max_uncomp_block_size_shift)
+
+/*define for write_compress_enable field*/
+#define SDMA_PKT_COPY_T2T_META_CONFIG_write_compress_enable_offset 17
+#define SDMA_PKT_COPY_T2T_META_CONFIG_write_compress_enable_mask   0x00000001
+#define SDMA_PKT_COPY_T2T_META_CONFIG_write_compress_enable_shift  28
+#define SDMA_PKT_COPY_T2T_META_CONFIG_WRITE_COMPRESS_ENABLE(x) (((x) & SDMA_PKT_COPY_T2T_META_CONFIG_write_compress_enable_mask) << SDMA_PKT_COPY_T2T_META_CONFIG_write_compress_enable_shift)
+
+/*define for meta_tmz field*/
+#define SDMA_PKT_COPY_T2T_META_CONFIG_meta_tmz_offset 17
+#define SDMA_PKT_COPY_T2T_META_CONFIG_meta_tmz_mask   0x00000001
+#define SDMA_PKT_COPY_T2T_META_CONFIG_meta_tmz_shift  29
+#define SDMA_PKT_COPY_T2T_META_CONFIG_META_TMZ(x) (((x) & SDMA_PKT_COPY_T2T_META_CONFIG_meta_tmz_mask) << SDMA_PKT_COPY_T2T_META_CONFIG_meta_tmz_shift)
+
+/*define for pipe_aligned field*/
+#define SDMA_PKT_COPY_T2T_META_CONFIG_pipe_aligned_offset 17
+#define SDMA_PKT_COPY_T2T_META_CONFIG_pipe_aligned_mask   0x00000001
+#define SDMA_PKT_COPY_T2T_META_CONFIG_pipe_aligned_shift  31
+#define SDMA_PKT_COPY_T2T_META_CONFIG_PIPE_ALIGNED(x) (((x) & SDMA_PKT_COPY_T2T_META_CONFIG_pipe_aligned_mask) << SDMA_PKT_COPY_T2T_META_CONFIG_pipe_aligned_shift)
+
+
+/*
+** Definitions for SDMA_PKT_COPY_T2T_BC packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_COPY_T2T_BC_HEADER_op_offset 0
+#define SDMA_PKT_COPY_T2T_BC_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_COPY_T2T_BC_HEADER_op_shift  0
+#define SDMA_PKT_COPY_T2T_BC_HEADER_OP(x) (((x) & SDMA_PKT_COPY_T2T_BC_HEADER_op_mask) << SDMA_PKT_COPY_T2T_BC_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_COPY_T2T_BC_HEADER_sub_op_offset 0
+#define SDMA_PKT_COPY_T2T_BC_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_COPY_T2T_BC_HEADER_sub_op_shift  8
+#define SDMA_PKT_COPY_T2T_BC_HEADER_SUB_OP(x) (((x) & SDMA_PKT_COPY_T2T_BC_HEADER_sub_op_mask) << SDMA_PKT_COPY_T2T_BC_HEADER_sub_op_shift)
+
+/*define for SRC_ADDR_LO word*/
+/*define for src_addr_31_0 field*/
+#define SDMA_PKT_COPY_T2T_BC_SRC_ADDR_LO_src_addr_31_0_offset 1
+#define SDMA_PKT_COPY_T2T_BC_SRC_ADDR_LO_src_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_T2T_BC_SRC_ADDR_LO_src_addr_31_0_shift  0
+#define SDMA_PKT_COPY_T2T_BC_SRC_ADDR_LO_SRC_ADDR_31_0(x) (((x) & SDMA_PKT_COPY_T2T_BC_SRC_ADDR_LO_src_addr_31_0_mask) << SDMA_PKT_COPY_T2T_BC_SRC_ADDR_LO_src_addr_31_0_shift)
+
+/*define for SRC_ADDR_HI word*/
+/*define for src_addr_63_32 field*/
+#define SDMA_PKT_COPY_T2T_BC_SRC_ADDR_HI_src_addr_63_32_offset 2
+#define SDMA_PKT_COPY_T2T_BC_SRC_ADDR_HI_src_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_T2T_BC_SRC_ADDR_HI_src_addr_63_32_shift  0
+#define SDMA_PKT_COPY_T2T_BC_SRC_ADDR_HI_SRC_ADDR_63_32(x) (((x) & SDMA_PKT_COPY_T2T_BC_SRC_ADDR_HI_src_addr_63_32_mask) << SDMA_PKT_COPY_T2T_BC_SRC_ADDR_HI_src_addr_63_32_shift)
+
+/*define for DW_3 word*/
+/*define for src_x field*/
+#define SDMA_PKT_COPY_T2T_BC_DW_3_src_x_offset 3
+#define SDMA_PKT_COPY_T2T_BC_DW_3_src_x_mask   0x00003FFF
+#define SDMA_PKT_COPY_T2T_BC_DW_3_src_x_shift  0
+#define SDMA_PKT_COPY_T2T_BC_DW_3_SRC_X(x) (((x) & SDMA_PKT_COPY_T2T_BC_DW_3_src_x_mask) << SDMA_PKT_COPY_T2T_BC_DW_3_src_x_shift)
+
+/*define for src_y field*/
+#define SDMA_PKT_COPY_T2T_BC_DW_3_src_y_offset 3
+#define SDMA_PKT_COPY_T2T_BC_DW_3_src_y_mask   0x00003FFF
+#define SDMA_PKT_COPY_T2T_BC_DW_3_src_y_shift  16
+#define SDMA_PKT_COPY_T2T_BC_DW_3_SRC_Y(x) (((x) & SDMA_PKT_COPY_T2T_BC_DW_3_src_y_mask) << SDMA_PKT_COPY_T2T_BC_DW_3_src_y_shift)
+
+/*define for DW_4 word*/
+/*define for src_z field*/
+#define SDMA_PKT_COPY_T2T_BC_DW_4_src_z_offset 4
+#define SDMA_PKT_COPY_T2T_BC_DW_4_src_z_mask   0x000007FF
+#define SDMA_PKT_COPY_T2T_BC_DW_4_src_z_shift  0
+#define SDMA_PKT_COPY_T2T_BC_DW_4_SRC_Z(x) (((x) & SDMA_PKT_COPY_T2T_BC_DW_4_src_z_mask) << SDMA_PKT_COPY_T2T_BC_DW_4_src_z_shift)
+
+/*define for src_width field*/
+#define SDMA_PKT_COPY_T2T_BC_DW_4_src_width_offset 4
+#define SDMA_PKT_COPY_T2T_BC_DW_4_src_width_mask   0x00003FFF
+#define SDMA_PKT_COPY_T2T_BC_DW_4_src_width_shift  16
+#define SDMA_PKT_COPY_T2T_BC_DW_4_SRC_WIDTH(x) (((x) & SDMA_PKT_COPY_T2T_BC_DW_4_src_width_mask) << SDMA_PKT_COPY_T2T_BC_DW_4_src_width_shift)
+
+/*define for DW_5 word*/
+/*define for src_height field*/
+#define SDMA_PKT_COPY_T2T_BC_DW_5_src_height_offset 5
+#define SDMA_PKT_COPY_T2T_BC_DW_5_src_height_mask   0x00003FFF
+#define SDMA_PKT_COPY_T2T_BC_DW_5_src_height_shift  0
+#define SDMA_PKT_COPY_T2T_BC_DW_5_SRC_HEIGHT(x) (((x) & SDMA_PKT_COPY_T2T_BC_DW_5_src_height_mask) << SDMA_PKT_COPY_T2T_BC_DW_5_src_height_shift)
+
+/*define for src_depth field*/
+#define SDMA_PKT_COPY_T2T_BC_DW_5_src_depth_offset 5
+#define SDMA_PKT_COPY_T2T_BC_DW_5_src_depth_mask   0x000007FF
+#define SDMA_PKT_COPY_T2T_BC_DW_5_src_depth_shift  16
+#define SDMA_PKT_COPY_T2T_BC_DW_5_SRC_DEPTH(x) (((x) & SDMA_PKT_COPY_T2T_BC_DW_5_src_depth_mask) << SDMA_PKT_COPY_T2T_BC_DW_5_src_depth_shift)
+
+/*define for DW_6 word*/
+/*define for src_element_size field*/
+#define SDMA_PKT_COPY_T2T_BC_DW_6_src_element_size_offset 6
+#define SDMA_PKT_COPY_T2T_BC_DW_6_src_element_size_mask   0x00000007
+#define SDMA_PKT_COPY_T2T_BC_DW_6_src_element_size_shift  0
+#define SDMA_PKT_COPY_T2T_BC_DW_6_SRC_ELEMENT_SIZE(x) (((x) & SDMA_PKT_COPY_T2T_BC_DW_6_src_element_size_mask) << SDMA_PKT_COPY_T2T_BC_DW_6_src_element_size_shift)
+
+/*define for src_array_mode field*/
+#define SDMA_PKT_COPY_T2T_BC_DW_6_src_array_mode_offset 6
+#define SDMA_PKT_COPY_T2T_BC_DW_6_src_array_mode_mask   0x0000000F
+#define SDMA_PKT_COPY_T2T_BC_DW_6_src_array_mode_shift  3
+#define SDMA_PKT_COPY_T2T_BC_DW_6_SRC_ARRAY_MODE(x) (((x) & SDMA_PKT_COPY_T2T_BC_DW_6_src_array_mode_mask) << SDMA_PKT_COPY_T2T_BC_DW_6_src_array_mode_shift)
+
+/*define for src_mit_mode field*/
+#define SDMA_PKT_COPY_T2T_BC_DW_6_src_mit_mode_offset 6
+#define SDMA_PKT_COPY_T2T_BC_DW_6_src_mit_mode_mask   0x00000007
+#define SDMA_PKT_COPY_T2T_BC_DW_6_src_mit_mode_shift  8
+#define SDMA_PKT_COPY_T2T_BC_DW_6_SRC_MIT_MODE(x) (((x) & SDMA_PKT_COPY_T2T_BC_DW_6_src_mit_mode_mask) << SDMA_PKT_COPY_T2T_BC_DW_6_src_mit_mode_shift)
+
+/*define for src_tilesplit_size field*/
+#define SDMA_PKT_COPY_T2T_BC_DW_6_src_tilesplit_size_offset 6
+#define SDMA_PKT_COPY_T2T_BC_DW_6_src_tilesplit_size_mask   0x00000007
+#define SDMA_PKT_COPY_T2T_BC_DW_6_src_tilesplit_size_shift  11
+#define SDMA_PKT_COPY_T2T_BC_DW_6_SRC_TILESPLIT_SIZE(x) (((x) & SDMA_PKT_COPY_T2T_BC_DW_6_src_tilesplit_size_mask) << SDMA_PKT_COPY_T2T_BC_DW_6_src_tilesplit_size_shift)
+
+/*define for src_bank_w field*/
+#define SDMA_PKT_COPY_T2T_BC_DW_6_src_bank_w_offset 6
+#define SDMA_PKT_COPY_T2T_BC_DW_6_src_bank_w_mask   0x00000003
+#define SDMA_PKT_COPY_T2T_BC_DW_6_src_bank_w_shift  15
+#define SDMA_PKT_COPY_T2T_BC_DW_6_SRC_BANK_W(x) (((x) & SDMA_PKT_COPY_T2T_BC_DW_6_src_bank_w_mask) << SDMA_PKT_COPY_T2T_BC_DW_6_src_bank_w_shift)
+
+/*define for src_bank_h field*/
+#define SDMA_PKT_COPY_T2T_BC_DW_6_src_bank_h_offset 6
+#define SDMA_PKT_COPY_T2T_BC_DW_6_src_bank_h_mask   0x00000003
+#define SDMA_PKT_COPY_T2T_BC_DW_6_src_bank_h_shift  18
+#define SDMA_PKT_COPY_T2T_BC_DW_6_SRC_BANK_H(x) (((x) & SDMA_PKT_COPY_T2T_BC_DW_6_src_bank_h_mask) << SDMA_PKT_COPY_T2T_BC_DW_6_src_bank_h_shift)
+
+/*define for src_num_bank field*/
+#define SDMA_PKT_COPY_T2T_BC_DW_6_src_num_bank_offset 6
+#define SDMA_PKT_COPY_T2T_BC_DW_6_src_num_bank_mask   0x00000003
+#define SDMA_PKT_COPY_T2T_BC_DW_6_src_num_bank_shift  21
+#define SDMA_PKT_COPY_T2T_BC_DW_6_SRC_NUM_BANK(x) (((x) & SDMA_PKT_COPY_T2T_BC_DW_6_src_num_bank_mask) << SDMA_PKT_COPY_T2T_BC_DW_6_src_num_bank_shift)
+
+/*define for src_mat_aspt field*/
+#define SDMA_PKT_COPY_T2T_BC_DW_6_src_mat_aspt_offset 6
+#define SDMA_PKT_COPY_T2T_BC_DW_6_src_mat_aspt_mask   0x00000003
+#define SDMA_PKT_COPY_T2T_BC_DW_6_src_mat_aspt_shift  24
+#define SDMA_PKT_COPY_T2T_BC_DW_6_SRC_MAT_ASPT(x) (((x) & SDMA_PKT_COPY_T2T_BC_DW_6_src_mat_aspt_mask) << SDMA_PKT_COPY_T2T_BC_DW_6_src_mat_aspt_shift)
+
+/*define for src_pipe_config field*/
+#define SDMA_PKT_COPY_T2T_BC_DW_6_src_pipe_config_offset 6
+#define SDMA_PKT_COPY_T2T_BC_DW_6_src_pipe_config_mask   0x0000001F
+#define SDMA_PKT_COPY_T2T_BC_DW_6_src_pipe_config_shift  26
+#define SDMA_PKT_COPY_T2T_BC_DW_6_SRC_PIPE_CONFIG(x) (((x) & SDMA_PKT_COPY_T2T_BC_DW_6_src_pipe_config_mask) << SDMA_PKT_COPY_T2T_BC_DW_6_src_pipe_config_shift)
+
+/*define for DST_ADDR_LO word*/
+/*define for dst_addr_31_0 field*/
+#define SDMA_PKT_COPY_T2T_BC_DST_ADDR_LO_dst_addr_31_0_offset 7
+#define SDMA_PKT_COPY_T2T_BC_DST_ADDR_LO_dst_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_T2T_BC_DST_ADDR_LO_dst_addr_31_0_shift  0
+#define SDMA_PKT_COPY_T2T_BC_DST_ADDR_LO_DST_ADDR_31_0(x) (((x) & SDMA_PKT_COPY_T2T_BC_DST_ADDR_LO_dst_addr_31_0_mask) << SDMA_PKT_COPY_T2T_BC_DST_ADDR_LO_dst_addr_31_0_shift)
+
+/*define for DST_ADDR_HI word*/
+/*define for dst_addr_63_32 field*/
+#define SDMA_PKT_COPY_T2T_BC_DST_ADDR_HI_dst_addr_63_32_offset 8
+#define SDMA_PKT_COPY_T2T_BC_DST_ADDR_HI_dst_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_T2T_BC_DST_ADDR_HI_dst_addr_63_32_shift  0
+#define SDMA_PKT_COPY_T2T_BC_DST_ADDR_HI_DST_ADDR_63_32(x) (((x) & SDMA_PKT_COPY_T2T_BC_DST_ADDR_HI_dst_addr_63_32_mask) << SDMA_PKT_COPY_T2T_BC_DST_ADDR_HI_dst_addr_63_32_shift)
+
+/*define for DW_9 word*/
+/*define for dst_x field*/
+#define SDMA_PKT_COPY_T2T_BC_DW_9_dst_x_offset 9
+#define SDMA_PKT_COPY_T2T_BC_DW_9_dst_x_mask   0x00003FFF
+#define SDMA_PKT_COPY_T2T_BC_DW_9_dst_x_shift  0
+#define SDMA_PKT_COPY_T2T_BC_DW_9_DST_X(x) (((x) & SDMA_PKT_COPY_T2T_BC_DW_9_dst_x_mask) << SDMA_PKT_COPY_T2T_BC_DW_9_dst_x_shift)
+
+/*define for dst_y field*/
+#define SDMA_PKT_COPY_T2T_BC_DW_9_dst_y_offset 9
+#define SDMA_PKT_COPY_T2T_BC_DW_9_dst_y_mask   0x00003FFF
+#define SDMA_PKT_COPY_T2T_BC_DW_9_dst_y_shift  16
+#define SDMA_PKT_COPY_T2T_BC_DW_9_DST_Y(x) (((x) & SDMA_PKT_COPY_T2T_BC_DW_9_dst_y_mask) << SDMA_PKT_COPY_T2T_BC_DW_9_dst_y_shift)
+
+/*define for DW_10 word*/
+/*define for dst_z field*/
+#define SDMA_PKT_COPY_T2T_BC_DW_10_dst_z_offset 10
+#define SDMA_PKT_COPY_T2T_BC_DW_10_dst_z_mask   0x000007FF
+#define SDMA_PKT_COPY_T2T_BC_DW_10_dst_z_shift  0
+#define SDMA_PKT_COPY_T2T_BC_DW_10_DST_Z(x) (((x) & SDMA_PKT_COPY_T2T_BC_DW_10_dst_z_mask) << SDMA_PKT_COPY_T2T_BC_DW_10_dst_z_shift)
+
+/*define for dst_width field*/
+#define SDMA_PKT_COPY_T2T_BC_DW_10_dst_width_offset 10
+#define SDMA_PKT_COPY_T2T_BC_DW_10_dst_width_mask   0x00003FFF
+#define SDMA_PKT_COPY_T2T_BC_DW_10_dst_width_shift  16
+#define SDMA_PKT_COPY_T2T_BC_DW_10_DST_WIDTH(x) (((x) & SDMA_PKT_COPY_T2T_BC_DW_10_dst_width_mask) << SDMA_PKT_COPY_T2T_BC_DW_10_dst_width_shift)
+
+/*define for DW_11 word*/
+/*define for dst_height field*/
+#define SDMA_PKT_COPY_T2T_BC_DW_11_dst_height_offset 11
+#define SDMA_PKT_COPY_T2T_BC_DW_11_dst_height_mask   0x00003FFF
+#define SDMA_PKT_COPY_T2T_BC_DW_11_dst_height_shift  0
+#define SDMA_PKT_COPY_T2T_BC_DW_11_DST_HEIGHT(x) (((x) & SDMA_PKT_COPY_T2T_BC_DW_11_dst_height_mask) << SDMA_PKT_COPY_T2T_BC_DW_11_dst_height_shift)
+
+/*define for dst_depth field*/
+#define SDMA_PKT_COPY_T2T_BC_DW_11_dst_depth_offset 11
+#define SDMA_PKT_COPY_T2T_BC_DW_11_dst_depth_mask   0x00000FFF
+#define SDMA_PKT_COPY_T2T_BC_DW_11_dst_depth_shift  16
+#define SDMA_PKT_COPY_T2T_BC_DW_11_DST_DEPTH(x) (((x) & SDMA_PKT_COPY_T2T_BC_DW_11_dst_depth_mask) << SDMA_PKT_COPY_T2T_BC_DW_11_dst_depth_shift)
+
+/*define for DW_12 word*/
+/*define for dst_element_size field*/
+#define SDMA_PKT_COPY_T2T_BC_DW_12_dst_element_size_offset 12
+#define SDMA_PKT_COPY_T2T_BC_DW_12_dst_element_size_mask   0x00000007
+#define SDMA_PKT_COPY_T2T_BC_DW_12_dst_element_size_shift  0
+#define SDMA_PKT_COPY_T2T_BC_DW_12_DST_ELEMENT_SIZE(x) (((x) & SDMA_PKT_COPY_T2T_BC_DW_12_dst_element_size_mask) << SDMA_PKT_COPY_T2T_BC_DW_12_dst_element_size_shift)
+
+/*define for dst_array_mode field*/
+#define SDMA_PKT_COPY_T2T_BC_DW_12_dst_array_mode_offset 12
+#define SDMA_PKT_COPY_T2T_BC_DW_12_dst_array_mode_mask   0x0000000F
+#define SDMA_PKT_COPY_T2T_BC_DW_12_dst_array_mode_shift  3
+#define SDMA_PKT_COPY_T2T_BC_DW_12_DST_ARRAY_MODE(x) (((x) & SDMA_PKT_COPY_T2T_BC_DW_12_dst_array_mode_mask) << SDMA_PKT_COPY_T2T_BC_DW_12_dst_array_mode_shift)
+
+/*define for dst_mit_mode field*/
+#define SDMA_PKT_COPY_T2T_BC_DW_12_dst_mit_mode_offset 12
+#define SDMA_PKT_COPY_T2T_BC_DW_12_dst_mit_mode_mask   0x00000007
+#define SDMA_PKT_COPY_T2T_BC_DW_12_dst_mit_mode_shift  8
+#define SDMA_PKT_COPY_T2T_BC_DW_12_DST_MIT_MODE(x) (((x) & SDMA_PKT_COPY_T2T_BC_DW_12_dst_mit_mode_mask) << SDMA_PKT_COPY_T2T_BC_DW_12_dst_mit_mode_shift)
+
+/*define for dst_tilesplit_size field*/
+#define SDMA_PKT_COPY_T2T_BC_DW_12_dst_tilesplit_size_offset 12
+#define SDMA_PKT_COPY_T2T_BC_DW_12_dst_tilesplit_size_mask   0x00000007
+#define SDMA_PKT_COPY_T2T_BC_DW_12_dst_tilesplit_size_shift  11
+#define SDMA_PKT_COPY_T2T_BC_DW_12_DST_TILESPLIT_SIZE(x) (((x) & SDMA_PKT_COPY_T2T_BC_DW_12_dst_tilesplit_size_mask) << SDMA_PKT_COPY_T2T_BC_DW_12_dst_tilesplit_size_shift)
+
+/*define for dst_bank_w field*/
+#define SDMA_PKT_COPY_T2T_BC_DW_12_dst_bank_w_offset 12
+#define SDMA_PKT_COPY_T2T_BC_DW_12_dst_bank_w_mask   0x00000003
+#define SDMA_PKT_COPY_T2T_BC_DW_12_dst_bank_w_shift  15
+#define SDMA_PKT_COPY_T2T_BC_DW_12_DST_BANK_W(x) (((x) & SDMA_PKT_COPY_T2T_BC_DW_12_dst_bank_w_mask) << SDMA_PKT_COPY_T2T_BC_DW_12_dst_bank_w_shift)
+
+/*define for dst_bank_h field*/
+#define SDMA_PKT_COPY_T2T_BC_DW_12_dst_bank_h_offset 12
+#define SDMA_PKT_COPY_T2T_BC_DW_12_dst_bank_h_mask   0x00000003
+#define SDMA_PKT_COPY_T2T_BC_DW_12_dst_bank_h_shift  18
+#define SDMA_PKT_COPY_T2T_BC_DW_12_DST_BANK_H(x) (((x) & SDMA_PKT_COPY_T2T_BC_DW_12_dst_bank_h_mask) << SDMA_PKT_COPY_T2T_BC_DW_12_dst_bank_h_shift)
+
+/*define for dst_num_bank field*/
+#define SDMA_PKT_COPY_T2T_BC_DW_12_dst_num_bank_offset 12
+#define SDMA_PKT_COPY_T2T_BC_DW_12_dst_num_bank_mask   0x00000003
+#define SDMA_PKT_COPY_T2T_BC_DW_12_dst_num_bank_shift  21
+#define SDMA_PKT_COPY_T2T_BC_DW_12_DST_NUM_BANK(x) (((x) & SDMA_PKT_COPY_T2T_BC_DW_12_dst_num_bank_mask) << SDMA_PKT_COPY_T2T_BC_DW_12_dst_num_bank_shift)
+
+/*define for dst_mat_aspt field*/
+#define SDMA_PKT_COPY_T2T_BC_DW_12_dst_mat_aspt_offset 12
+#define SDMA_PKT_COPY_T2T_BC_DW_12_dst_mat_aspt_mask   0x00000003
+#define SDMA_PKT_COPY_T2T_BC_DW_12_dst_mat_aspt_shift  24
+#define SDMA_PKT_COPY_T2T_BC_DW_12_DST_MAT_ASPT(x) (((x) & SDMA_PKT_COPY_T2T_BC_DW_12_dst_mat_aspt_mask) << SDMA_PKT_COPY_T2T_BC_DW_12_dst_mat_aspt_shift)
+
+/*define for dst_pipe_config field*/
+#define SDMA_PKT_COPY_T2T_BC_DW_12_dst_pipe_config_offset 12
+#define SDMA_PKT_COPY_T2T_BC_DW_12_dst_pipe_config_mask   0x0000001F
+#define SDMA_PKT_COPY_T2T_BC_DW_12_dst_pipe_config_shift  26
+#define SDMA_PKT_COPY_T2T_BC_DW_12_DST_PIPE_CONFIG(x) (((x) & SDMA_PKT_COPY_T2T_BC_DW_12_dst_pipe_config_mask) << SDMA_PKT_COPY_T2T_BC_DW_12_dst_pipe_config_shift)
+
+/*define for DW_13 word*/
+/*define for rect_x field*/
+#define SDMA_PKT_COPY_T2T_BC_DW_13_rect_x_offset 13
+#define SDMA_PKT_COPY_T2T_BC_DW_13_rect_x_mask   0x00003FFF
+#define SDMA_PKT_COPY_T2T_BC_DW_13_rect_x_shift  0
+#define SDMA_PKT_COPY_T2T_BC_DW_13_RECT_X(x) (((x) & SDMA_PKT_COPY_T2T_BC_DW_13_rect_x_mask) << SDMA_PKT_COPY_T2T_BC_DW_13_rect_x_shift)
+
+/*define for rect_y field*/
+#define SDMA_PKT_COPY_T2T_BC_DW_13_rect_y_offset 13
+#define SDMA_PKT_COPY_T2T_BC_DW_13_rect_y_mask   0x00003FFF
+#define SDMA_PKT_COPY_T2T_BC_DW_13_rect_y_shift  16
+#define SDMA_PKT_COPY_T2T_BC_DW_13_RECT_Y(x) (((x) & SDMA_PKT_COPY_T2T_BC_DW_13_rect_y_mask) << SDMA_PKT_COPY_T2T_BC_DW_13_rect_y_shift)
+
+/*define for DW_14 word*/
+/*define for rect_z field*/
+#define SDMA_PKT_COPY_T2T_BC_DW_14_rect_z_offset 14
+#define SDMA_PKT_COPY_T2T_BC_DW_14_rect_z_mask   0x000007FF
+#define SDMA_PKT_COPY_T2T_BC_DW_14_rect_z_shift  0
+#define SDMA_PKT_COPY_T2T_BC_DW_14_RECT_Z(x) (((x) & SDMA_PKT_COPY_T2T_BC_DW_14_rect_z_mask) << SDMA_PKT_COPY_T2T_BC_DW_14_rect_z_shift)
+
+/*define for dst_sw field*/
+#define SDMA_PKT_COPY_T2T_BC_DW_14_dst_sw_offset 14
+#define SDMA_PKT_COPY_T2T_BC_DW_14_dst_sw_mask   0x00000003
+#define SDMA_PKT_COPY_T2T_BC_DW_14_dst_sw_shift  16
+#define SDMA_PKT_COPY_T2T_BC_DW_14_DST_SW(x) (((x) & SDMA_PKT_COPY_T2T_BC_DW_14_dst_sw_mask) << SDMA_PKT_COPY_T2T_BC_DW_14_dst_sw_shift)
+
+/*define for src_sw field*/
+#define SDMA_PKT_COPY_T2T_BC_DW_14_src_sw_offset 14
+#define SDMA_PKT_COPY_T2T_BC_DW_14_src_sw_mask   0x00000003
+#define SDMA_PKT_COPY_T2T_BC_DW_14_src_sw_shift  24
+#define SDMA_PKT_COPY_T2T_BC_DW_14_SRC_SW(x) (((x) & SDMA_PKT_COPY_T2T_BC_DW_14_src_sw_mask) << SDMA_PKT_COPY_T2T_BC_DW_14_src_sw_shift)
+
+
+/*
+** Definitions for SDMA_PKT_COPY_TILED_SUBWIN packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_HEADER_op_offset 0
+#define SDMA_PKT_COPY_TILED_SUBWIN_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_COPY_TILED_SUBWIN_HEADER_op_shift  0
+#define SDMA_PKT_COPY_TILED_SUBWIN_HEADER_OP(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_HEADER_op_mask) << SDMA_PKT_COPY_TILED_SUBWIN_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_HEADER_sub_op_offset 0
+#define SDMA_PKT_COPY_TILED_SUBWIN_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_COPY_TILED_SUBWIN_HEADER_sub_op_shift  8
+#define SDMA_PKT_COPY_TILED_SUBWIN_HEADER_SUB_OP(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_HEADER_sub_op_mask) << SDMA_PKT_COPY_TILED_SUBWIN_HEADER_sub_op_shift)
+
+/*define for tmz field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_HEADER_tmz_offset 0
+#define SDMA_PKT_COPY_TILED_SUBWIN_HEADER_tmz_mask   0x00000001
+#define SDMA_PKT_COPY_TILED_SUBWIN_HEADER_tmz_shift  18
+#define SDMA_PKT_COPY_TILED_SUBWIN_HEADER_TMZ(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_HEADER_tmz_mask) << SDMA_PKT_COPY_TILED_SUBWIN_HEADER_tmz_shift)
+
+/*define for dcc field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_HEADER_dcc_offset 0
+#define SDMA_PKT_COPY_TILED_SUBWIN_HEADER_dcc_mask   0x00000001
+#define SDMA_PKT_COPY_TILED_SUBWIN_HEADER_dcc_shift  19
+#define SDMA_PKT_COPY_TILED_SUBWIN_HEADER_DCC(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_HEADER_dcc_mask) << SDMA_PKT_COPY_TILED_SUBWIN_HEADER_dcc_shift)
+
+/*define for cpv field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_HEADER_cpv_offset 0
+#define SDMA_PKT_COPY_TILED_SUBWIN_HEADER_cpv_mask   0x00000001
+#define SDMA_PKT_COPY_TILED_SUBWIN_HEADER_cpv_shift  28
+#define SDMA_PKT_COPY_TILED_SUBWIN_HEADER_CPV(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_HEADER_cpv_mask) << SDMA_PKT_COPY_TILED_SUBWIN_HEADER_cpv_shift)
+
+/*define for detile field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_HEADER_detile_offset 0
+#define SDMA_PKT_COPY_TILED_SUBWIN_HEADER_detile_mask   0x00000001
+#define SDMA_PKT_COPY_TILED_SUBWIN_HEADER_detile_shift  31
+#define SDMA_PKT_COPY_TILED_SUBWIN_HEADER_DETILE(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_HEADER_detile_mask) << SDMA_PKT_COPY_TILED_SUBWIN_HEADER_detile_shift)
+
+/*define for TILED_ADDR_LO word*/
+/*define for tiled_addr_31_0 field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_TILED_ADDR_LO_tiled_addr_31_0_offset 1
+#define SDMA_PKT_COPY_TILED_SUBWIN_TILED_ADDR_LO_tiled_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_TILED_SUBWIN_TILED_ADDR_LO_tiled_addr_31_0_shift  0
+#define SDMA_PKT_COPY_TILED_SUBWIN_TILED_ADDR_LO_TILED_ADDR_31_0(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_TILED_ADDR_LO_tiled_addr_31_0_mask) << SDMA_PKT_COPY_TILED_SUBWIN_TILED_ADDR_LO_tiled_addr_31_0_shift)
+
+/*define for TILED_ADDR_HI word*/
+/*define for tiled_addr_63_32 field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_TILED_ADDR_HI_tiled_addr_63_32_offset 2
+#define SDMA_PKT_COPY_TILED_SUBWIN_TILED_ADDR_HI_tiled_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_TILED_SUBWIN_TILED_ADDR_HI_tiled_addr_63_32_shift  0
+#define SDMA_PKT_COPY_TILED_SUBWIN_TILED_ADDR_HI_TILED_ADDR_63_32(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_TILED_ADDR_HI_tiled_addr_63_32_mask) << SDMA_PKT_COPY_TILED_SUBWIN_TILED_ADDR_HI_tiled_addr_63_32_shift)
+
+/*define for DW_3 word*/
+/*define for tiled_x field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_3_tiled_x_offset 3
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_3_tiled_x_mask   0x00003FFF
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_3_tiled_x_shift  0
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_3_TILED_X(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_DW_3_tiled_x_mask) << SDMA_PKT_COPY_TILED_SUBWIN_DW_3_tiled_x_shift)
+
+/*define for tiled_y field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_3_tiled_y_offset 3
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_3_tiled_y_mask   0x00003FFF
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_3_tiled_y_shift  16
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_3_TILED_Y(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_DW_3_tiled_y_mask) << SDMA_PKT_COPY_TILED_SUBWIN_DW_3_tiled_y_shift)
+
+/*define for DW_4 word*/
+/*define for tiled_z field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_4_tiled_z_offset 4
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_4_tiled_z_mask   0x00001FFF
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_4_tiled_z_shift  0
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_4_TILED_Z(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_DW_4_tiled_z_mask) << SDMA_PKT_COPY_TILED_SUBWIN_DW_4_tiled_z_shift)
+
+/*define for width field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_4_width_offset 4
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_4_width_mask   0x00003FFF
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_4_width_shift  16
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_4_WIDTH(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_DW_4_width_mask) << SDMA_PKT_COPY_TILED_SUBWIN_DW_4_width_shift)
+
+/*define for DW_5 word*/
+/*define for height field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_5_height_offset 5
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_5_height_mask   0x00003FFF
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_5_height_shift  0
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_5_HEIGHT(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_DW_5_height_mask) << SDMA_PKT_COPY_TILED_SUBWIN_DW_5_height_shift)
+
+/*define for depth field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_5_depth_offset 5
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_5_depth_mask   0x00001FFF
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_5_depth_shift  16
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_5_DEPTH(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_DW_5_depth_mask) << SDMA_PKT_COPY_TILED_SUBWIN_DW_5_depth_shift)
+
+/*define for DW_6 word*/
+/*define for element_size field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_6_element_size_offset 6
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_6_element_size_mask   0x00000007
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_6_element_size_shift  0
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_6_ELEMENT_SIZE(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_DW_6_element_size_mask) << SDMA_PKT_COPY_TILED_SUBWIN_DW_6_element_size_shift)
+
+/*define for swizzle_mode field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_6_swizzle_mode_offset 6
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_6_swizzle_mode_mask   0x0000001F
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_6_swizzle_mode_shift  3
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_6_SWIZZLE_MODE(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_DW_6_swizzle_mode_mask) << SDMA_PKT_COPY_TILED_SUBWIN_DW_6_swizzle_mode_shift)
+
+/*define for dimension field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_6_dimension_offset 6
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_6_dimension_mask   0x00000003
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_6_dimension_shift  9
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_6_DIMENSION(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_DW_6_dimension_mask) << SDMA_PKT_COPY_TILED_SUBWIN_DW_6_dimension_shift)
+
+/*define for mip_max field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_6_mip_max_offset 6
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_6_mip_max_mask   0x0000000F
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_6_mip_max_shift  16
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_6_MIP_MAX(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_DW_6_mip_max_mask) << SDMA_PKT_COPY_TILED_SUBWIN_DW_6_mip_max_shift)
+
+/*define for mip_id field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_6_mip_id_offset 6
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_6_mip_id_mask   0x0000000F
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_6_mip_id_shift  20
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_6_MIP_ID(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_DW_6_mip_id_mask) << SDMA_PKT_COPY_TILED_SUBWIN_DW_6_mip_id_shift)
+
+/*define for LINEAR_ADDR_LO word*/
+/*define for linear_addr_31_0 field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_LINEAR_ADDR_LO_linear_addr_31_0_offset 7
+#define SDMA_PKT_COPY_TILED_SUBWIN_LINEAR_ADDR_LO_linear_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_TILED_SUBWIN_LINEAR_ADDR_LO_linear_addr_31_0_shift  0
+#define SDMA_PKT_COPY_TILED_SUBWIN_LINEAR_ADDR_LO_LINEAR_ADDR_31_0(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_LINEAR_ADDR_LO_linear_addr_31_0_mask) << SDMA_PKT_COPY_TILED_SUBWIN_LINEAR_ADDR_LO_linear_addr_31_0_shift)
+
+/*define for LINEAR_ADDR_HI word*/
+/*define for linear_addr_63_32 field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_LINEAR_ADDR_HI_linear_addr_63_32_offset 8
+#define SDMA_PKT_COPY_TILED_SUBWIN_LINEAR_ADDR_HI_linear_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_TILED_SUBWIN_LINEAR_ADDR_HI_linear_addr_63_32_shift  0
+#define SDMA_PKT_COPY_TILED_SUBWIN_LINEAR_ADDR_HI_LINEAR_ADDR_63_32(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_LINEAR_ADDR_HI_linear_addr_63_32_mask) << SDMA_PKT_COPY_TILED_SUBWIN_LINEAR_ADDR_HI_linear_addr_63_32_shift)
+
+/*define for DW_9 word*/
+/*define for linear_x field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_9_linear_x_offset 9
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_9_linear_x_mask   0x00003FFF
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_9_linear_x_shift  0
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_9_LINEAR_X(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_DW_9_linear_x_mask) << SDMA_PKT_COPY_TILED_SUBWIN_DW_9_linear_x_shift)
+
+/*define for linear_y field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_9_linear_y_offset 9
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_9_linear_y_mask   0x00003FFF
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_9_linear_y_shift  16
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_9_LINEAR_Y(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_DW_9_linear_y_mask) << SDMA_PKT_COPY_TILED_SUBWIN_DW_9_linear_y_shift)
+
+/*define for DW_10 word*/
+/*define for linear_z field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_10_linear_z_offset 10
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_10_linear_z_mask   0x00001FFF
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_10_linear_z_shift  0
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_10_LINEAR_Z(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_DW_10_linear_z_mask) << SDMA_PKT_COPY_TILED_SUBWIN_DW_10_linear_z_shift)
+
+/*define for linear_pitch field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_10_linear_pitch_offset 10
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_10_linear_pitch_mask   0x00003FFF
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_10_linear_pitch_shift  16
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_10_LINEAR_PITCH(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_DW_10_linear_pitch_mask) << SDMA_PKT_COPY_TILED_SUBWIN_DW_10_linear_pitch_shift)
+
+/*define for DW_11 word*/
+/*define for linear_slice_pitch field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_11_linear_slice_pitch_offset 11
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_11_linear_slice_pitch_mask   0x0FFFFFFF
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_11_linear_slice_pitch_shift  0
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_11_LINEAR_SLICE_PITCH(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_DW_11_linear_slice_pitch_mask) << SDMA_PKT_COPY_TILED_SUBWIN_DW_11_linear_slice_pitch_shift)
+
+/*define for DW_12 word*/
+/*define for rect_x field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_12_rect_x_offset 12
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_12_rect_x_mask   0x00003FFF
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_12_rect_x_shift  0
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_12_RECT_X(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_DW_12_rect_x_mask) << SDMA_PKT_COPY_TILED_SUBWIN_DW_12_rect_x_shift)
+
+/*define for rect_y field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_12_rect_y_offset 12
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_12_rect_y_mask   0x00003FFF
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_12_rect_y_shift  16
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_12_RECT_Y(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_DW_12_rect_y_mask) << SDMA_PKT_COPY_TILED_SUBWIN_DW_12_rect_y_shift)
+
+/*define for DW_13 word*/
+/*define for rect_z field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_13_rect_z_offset 13
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_13_rect_z_mask   0x00001FFF
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_13_rect_z_shift  0
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_13_RECT_Z(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_DW_13_rect_z_mask) << SDMA_PKT_COPY_TILED_SUBWIN_DW_13_rect_z_shift)
+
+/*define for linear_sw field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_13_linear_sw_offset 13
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_13_linear_sw_mask   0x00000003
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_13_linear_sw_shift  16
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_13_LINEAR_SW(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_DW_13_linear_sw_mask) << SDMA_PKT_COPY_TILED_SUBWIN_DW_13_linear_sw_shift)
+
+/*define for linear_cache_policy field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_13_linear_cache_policy_offset 13
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_13_linear_cache_policy_mask   0x00000007
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_13_linear_cache_policy_shift  18
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_13_LINEAR_CACHE_POLICY(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_DW_13_linear_cache_policy_mask) << SDMA_PKT_COPY_TILED_SUBWIN_DW_13_linear_cache_policy_shift)
+
+/*define for tile_sw field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_13_tile_sw_offset 13
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_13_tile_sw_mask   0x00000003
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_13_tile_sw_shift  24
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_13_TILE_SW(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_DW_13_tile_sw_mask) << SDMA_PKT_COPY_TILED_SUBWIN_DW_13_tile_sw_shift)
+
+/*define for tile_cache_policy field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_13_tile_cache_policy_offset 13
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_13_tile_cache_policy_mask   0x00000007
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_13_tile_cache_policy_shift  26
+#define SDMA_PKT_COPY_TILED_SUBWIN_DW_13_TILE_CACHE_POLICY(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_DW_13_tile_cache_policy_mask) << SDMA_PKT_COPY_TILED_SUBWIN_DW_13_tile_cache_policy_shift)
+
+/*define for META_ADDR_LO word*/
+/*define for meta_addr_31_0 field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_ADDR_LO_meta_addr_31_0_offset 14
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_ADDR_LO_meta_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_ADDR_LO_meta_addr_31_0_shift  0
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_ADDR_LO_META_ADDR_31_0(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_META_ADDR_LO_meta_addr_31_0_mask) << SDMA_PKT_COPY_TILED_SUBWIN_META_ADDR_LO_meta_addr_31_0_shift)
+
+/*define for META_ADDR_HI word*/
+/*define for meta_addr_63_32 field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_ADDR_HI_meta_addr_63_32_offset 15
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_ADDR_HI_meta_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_ADDR_HI_meta_addr_63_32_shift  0
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_ADDR_HI_META_ADDR_63_32(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_META_ADDR_HI_meta_addr_63_32_mask) << SDMA_PKT_COPY_TILED_SUBWIN_META_ADDR_HI_meta_addr_63_32_shift)
+
+/*define for META_CONFIG word*/
+/*define for data_format field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_data_format_offset 16
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_data_format_mask   0x0000007F
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_data_format_shift  0
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_DATA_FORMAT(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_data_format_mask) << SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_data_format_shift)
+
+/*define for color_transform_disable field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_color_transform_disable_offset 16
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_color_transform_disable_mask   0x00000001
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_color_transform_disable_shift  7
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_COLOR_TRANSFORM_DISABLE(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_color_transform_disable_mask) << SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_color_transform_disable_shift)
+
+/*define for alpha_is_on_msb field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_alpha_is_on_msb_offset 16
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_alpha_is_on_msb_mask   0x00000001
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_alpha_is_on_msb_shift  8
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_ALPHA_IS_ON_MSB(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_alpha_is_on_msb_mask) << SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_alpha_is_on_msb_shift)
+
+/*define for number_type field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_number_type_offset 16
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_number_type_mask   0x00000007
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_number_type_shift  9
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_NUMBER_TYPE(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_number_type_mask) << SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_number_type_shift)
+
+/*define for surface_type field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_surface_type_offset 16
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_surface_type_mask   0x00000003
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_surface_type_shift  12
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_SURFACE_TYPE(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_surface_type_mask) << SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_surface_type_shift)
+
+/*define for meta_llc field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_meta_llc_offset 16
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_meta_llc_mask   0x00000001
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_meta_llc_shift  14
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_META_LLC(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_meta_llc_mask) << SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_meta_llc_shift)
+
+/*define for max_comp_block_size field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_max_comp_block_size_offset 16
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_max_comp_block_size_mask   0x00000003
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_max_comp_block_size_shift  24
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_MAX_COMP_BLOCK_SIZE(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_max_comp_block_size_mask) << SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_max_comp_block_size_shift)
+
+/*define for max_uncomp_block_size field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_max_uncomp_block_size_offset 16
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_max_uncomp_block_size_mask   0x00000003
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_max_uncomp_block_size_shift  26
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_MAX_UNCOMP_BLOCK_SIZE(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_max_uncomp_block_size_mask) << SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_max_uncomp_block_size_shift)
+
+/*define for write_compress_enable field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_write_compress_enable_offset 16
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_write_compress_enable_mask   0x00000001
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_write_compress_enable_shift  28
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_WRITE_COMPRESS_ENABLE(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_write_compress_enable_mask) << SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_write_compress_enable_shift)
+
+/*define for meta_tmz field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_meta_tmz_offset 16
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_meta_tmz_mask   0x00000001
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_meta_tmz_shift  29
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_META_TMZ(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_meta_tmz_mask) << SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_meta_tmz_shift)
+
+/*define for pipe_aligned field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_pipe_aligned_offset 16
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_pipe_aligned_mask   0x00000001
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_pipe_aligned_shift  31
+#define SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_PIPE_ALIGNED(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_pipe_aligned_mask) << SDMA_PKT_COPY_TILED_SUBWIN_META_CONFIG_pipe_aligned_shift)
+
+
+/*
+** Definitions for SDMA_PKT_COPY_TILED_SUBWIN_BC packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_HEADER_op_offset 0
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_HEADER_op_shift  0
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_HEADER_OP(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_BC_HEADER_op_mask) << SDMA_PKT_COPY_TILED_SUBWIN_BC_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_HEADER_sub_op_offset 0
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_HEADER_sub_op_shift  8
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_HEADER_SUB_OP(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_BC_HEADER_sub_op_mask) << SDMA_PKT_COPY_TILED_SUBWIN_BC_HEADER_sub_op_shift)
+
+/*define for detile field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_HEADER_detile_offset 0
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_HEADER_detile_mask   0x00000001
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_HEADER_detile_shift  31
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_HEADER_DETILE(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_BC_HEADER_detile_mask) << SDMA_PKT_COPY_TILED_SUBWIN_BC_HEADER_detile_shift)
+
+/*define for TILED_ADDR_LO word*/
+/*define for tiled_addr_31_0 field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_TILED_ADDR_LO_tiled_addr_31_0_offset 1
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_TILED_ADDR_LO_tiled_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_TILED_ADDR_LO_tiled_addr_31_0_shift  0
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_TILED_ADDR_LO_TILED_ADDR_31_0(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_BC_TILED_ADDR_LO_tiled_addr_31_0_mask) << SDMA_PKT_COPY_TILED_SUBWIN_BC_TILED_ADDR_LO_tiled_addr_31_0_shift)
+
+/*define for TILED_ADDR_HI word*/
+/*define for tiled_addr_63_32 field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_TILED_ADDR_HI_tiled_addr_63_32_offset 2
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_TILED_ADDR_HI_tiled_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_TILED_ADDR_HI_tiled_addr_63_32_shift  0
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_TILED_ADDR_HI_TILED_ADDR_63_32(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_BC_TILED_ADDR_HI_tiled_addr_63_32_mask) << SDMA_PKT_COPY_TILED_SUBWIN_BC_TILED_ADDR_HI_tiled_addr_63_32_shift)
+
+/*define for DW_3 word*/
+/*define for tiled_x field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_3_tiled_x_offset 3
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_3_tiled_x_mask   0x00003FFF
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_3_tiled_x_shift  0
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_3_TILED_X(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_3_tiled_x_mask) << SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_3_tiled_x_shift)
+
+/*define for tiled_y field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_3_tiled_y_offset 3
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_3_tiled_y_mask   0x00003FFF
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_3_tiled_y_shift  16
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_3_TILED_Y(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_3_tiled_y_mask) << SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_3_tiled_y_shift)
+
+/*define for DW_4 word*/
+/*define for tiled_z field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_4_tiled_z_offset 4
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_4_tiled_z_mask   0x000007FF
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_4_tiled_z_shift  0
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_4_TILED_Z(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_4_tiled_z_mask) << SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_4_tiled_z_shift)
+
+/*define for width field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_4_width_offset 4
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_4_width_mask   0x00003FFF
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_4_width_shift  16
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_4_WIDTH(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_4_width_mask) << SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_4_width_shift)
+
+/*define for DW_5 word*/
+/*define for height field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_5_height_offset 5
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_5_height_mask   0x00003FFF
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_5_height_shift  0
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_5_HEIGHT(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_5_height_mask) << SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_5_height_shift)
+
+/*define for depth field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_5_depth_offset 5
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_5_depth_mask   0x000007FF
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_5_depth_shift  16
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_5_DEPTH(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_5_depth_mask) << SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_5_depth_shift)
+
+/*define for DW_6 word*/
+/*define for element_size field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_element_size_offset 6
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_element_size_mask   0x00000007
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_element_size_shift  0
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_ELEMENT_SIZE(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_element_size_mask) << SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_element_size_shift)
+
+/*define for array_mode field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_array_mode_offset 6
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_array_mode_mask   0x0000000F
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_array_mode_shift  3
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_ARRAY_MODE(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_array_mode_mask) << SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_array_mode_shift)
+
+/*define for mit_mode field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_mit_mode_offset 6
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_mit_mode_mask   0x00000007
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_mit_mode_shift  8
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_MIT_MODE(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_mit_mode_mask) << SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_mit_mode_shift)
+
+/*define for tilesplit_size field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_tilesplit_size_offset 6
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_tilesplit_size_mask   0x00000007
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_tilesplit_size_shift  11
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_TILESPLIT_SIZE(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_tilesplit_size_mask) << SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_tilesplit_size_shift)
+
+/*define for bank_w field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_bank_w_offset 6
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_bank_w_mask   0x00000003
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_bank_w_shift  15
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_BANK_W(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_bank_w_mask) << SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_bank_w_shift)
+
+/*define for bank_h field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_bank_h_offset 6
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_bank_h_mask   0x00000003
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_bank_h_shift  18
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_BANK_H(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_bank_h_mask) << SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_bank_h_shift)
+
+/*define for num_bank field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_num_bank_offset 6
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_num_bank_mask   0x00000003
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_num_bank_shift  21
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_NUM_BANK(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_num_bank_mask) << SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_num_bank_shift)
+
+/*define for mat_aspt field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_mat_aspt_offset 6
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_mat_aspt_mask   0x00000003
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_mat_aspt_shift  24
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_MAT_ASPT(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_mat_aspt_mask) << SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_mat_aspt_shift)
+
+/*define for pipe_config field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_pipe_config_offset 6
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_pipe_config_mask   0x0000001F
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_pipe_config_shift  26
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_PIPE_CONFIG(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_pipe_config_mask) << SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_6_pipe_config_shift)
+
+/*define for LINEAR_ADDR_LO word*/
+/*define for linear_addr_31_0 field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_LINEAR_ADDR_LO_linear_addr_31_0_offset 7
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_LINEAR_ADDR_LO_linear_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_LINEAR_ADDR_LO_linear_addr_31_0_shift  0
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_LINEAR_ADDR_LO_LINEAR_ADDR_31_0(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_BC_LINEAR_ADDR_LO_linear_addr_31_0_mask) << SDMA_PKT_COPY_TILED_SUBWIN_BC_LINEAR_ADDR_LO_linear_addr_31_0_shift)
+
+/*define for LINEAR_ADDR_HI word*/
+/*define for linear_addr_63_32 field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_LINEAR_ADDR_HI_linear_addr_63_32_offset 8
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_LINEAR_ADDR_HI_linear_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_LINEAR_ADDR_HI_linear_addr_63_32_shift  0
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_LINEAR_ADDR_HI_LINEAR_ADDR_63_32(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_BC_LINEAR_ADDR_HI_linear_addr_63_32_mask) << SDMA_PKT_COPY_TILED_SUBWIN_BC_LINEAR_ADDR_HI_linear_addr_63_32_shift)
+
+/*define for DW_9 word*/
+/*define for linear_x field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_9_linear_x_offset 9
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_9_linear_x_mask   0x00003FFF
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_9_linear_x_shift  0
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_9_LINEAR_X(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_9_linear_x_mask) << SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_9_linear_x_shift)
+
+/*define for linear_y field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_9_linear_y_offset 9
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_9_linear_y_mask   0x00003FFF
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_9_linear_y_shift  16
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_9_LINEAR_Y(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_9_linear_y_mask) << SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_9_linear_y_shift)
+
+/*define for DW_10 word*/
+/*define for linear_z field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_10_linear_z_offset 10
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_10_linear_z_mask   0x000007FF
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_10_linear_z_shift  0
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_10_LINEAR_Z(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_10_linear_z_mask) << SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_10_linear_z_shift)
+
+/*define for linear_pitch field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_10_linear_pitch_offset 10
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_10_linear_pitch_mask   0x00003FFF
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_10_linear_pitch_shift  16
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_10_LINEAR_PITCH(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_10_linear_pitch_mask) << SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_10_linear_pitch_shift)
+
+/*define for DW_11 word*/
+/*define for linear_slice_pitch field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_11_linear_slice_pitch_offset 11
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_11_linear_slice_pitch_mask   0x0FFFFFFF
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_11_linear_slice_pitch_shift  0
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_11_LINEAR_SLICE_PITCH(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_11_linear_slice_pitch_mask) << SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_11_linear_slice_pitch_shift)
+
+/*define for DW_12 word*/
+/*define for rect_x field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_12_rect_x_offset 12
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_12_rect_x_mask   0x00003FFF
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_12_rect_x_shift  0
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_12_RECT_X(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_12_rect_x_mask) << SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_12_rect_x_shift)
+
+/*define for rect_y field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_12_rect_y_offset 12
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_12_rect_y_mask   0x00003FFF
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_12_rect_y_shift  16
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_12_RECT_Y(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_12_rect_y_mask) << SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_12_rect_y_shift)
+
+/*define for DW_13 word*/
+/*define for rect_z field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_13_rect_z_offset 13
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_13_rect_z_mask   0x000007FF
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_13_rect_z_shift  0
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_13_RECT_Z(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_13_rect_z_mask) << SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_13_rect_z_shift)
+
+/*define for linear_sw field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_13_linear_sw_offset 13
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_13_linear_sw_mask   0x00000003
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_13_linear_sw_shift  16
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_13_LINEAR_SW(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_13_linear_sw_mask) << SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_13_linear_sw_shift)
+
+/*define for tile_sw field*/
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_13_tile_sw_offset 13
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_13_tile_sw_mask   0x00000003
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_13_tile_sw_shift  24
+#define SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_13_TILE_SW(x) (((x) & SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_13_tile_sw_mask) << SDMA_PKT_COPY_TILED_SUBWIN_BC_DW_13_tile_sw_shift)
+
+
+/*
+** Definitions for SDMA_PKT_COPY_STRUCT packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_COPY_STRUCT_HEADER_op_offset 0
+#define SDMA_PKT_COPY_STRUCT_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_COPY_STRUCT_HEADER_op_shift  0
+#define SDMA_PKT_COPY_STRUCT_HEADER_OP(x) (((x) & SDMA_PKT_COPY_STRUCT_HEADER_op_mask) << SDMA_PKT_COPY_STRUCT_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_COPY_STRUCT_HEADER_sub_op_offset 0
+#define SDMA_PKT_COPY_STRUCT_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_COPY_STRUCT_HEADER_sub_op_shift  8
+#define SDMA_PKT_COPY_STRUCT_HEADER_SUB_OP(x) (((x) & SDMA_PKT_COPY_STRUCT_HEADER_sub_op_mask) << SDMA_PKT_COPY_STRUCT_HEADER_sub_op_shift)
+
+/*define for tmz field*/
+#define SDMA_PKT_COPY_STRUCT_HEADER_tmz_offset 0
+#define SDMA_PKT_COPY_STRUCT_HEADER_tmz_mask   0x00000001
+#define SDMA_PKT_COPY_STRUCT_HEADER_tmz_shift  18
+#define SDMA_PKT_COPY_STRUCT_HEADER_TMZ(x) (((x) & SDMA_PKT_COPY_STRUCT_HEADER_tmz_mask) << SDMA_PKT_COPY_STRUCT_HEADER_tmz_shift)
+
+/*define for cpv field*/
+#define SDMA_PKT_COPY_STRUCT_HEADER_cpv_offset 0
+#define SDMA_PKT_COPY_STRUCT_HEADER_cpv_mask   0x00000001
+#define SDMA_PKT_COPY_STRUCT_HEADER_cpv_shift  28
+#define SDMA_PKT_COPY_STRUCT_HEADER_CPV(x) (((x) & SDMA_PKT_COPY_STRUCT_HEADER_cpv_mask) << SDMA_PKT_COPY_STRUCT_HEADER_cpv_shift)
+
+/*define for detile field*/
+#define SDMA_PKT_COPY_STRUCT_HEADER_detile_offset 0
+#define SDMA_PKT_COPY_STRUCT_HEADER_detile_mask   0x00000001
+#define SDMA_PKT_COPY_STRUCT_HEADER_detile_shift  31
+#define SDMA_PKT_COPY_STRUCT_HEADER_DETILE(x) (((x) & SDMA_PKT_COPY_STRUCT_HEADER_detile_mask) << SDMA_PKT_COPY_STRUCT_HEADER_detile_shift)
+
+/*define for SB_ADDR_LO word*/
+/*define for sb_addr_31_0 field*/
+#define SDMA_PKT_COPY_STRUCT_SB_ADDR_LO_sb_addr_31_0_offset 1
+#define SDMA_PKT_COPY_STRUCT_SB_ADDR_LO_sb_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_STRUCT_SB_ADDR_LO_sb_addr_31_0_shift  0
+#define SDMA_PKT_COPY_STRUCT_SB_ADDR_LO_SB_ADDR_31_0(x) (((x) & SDMA_PKT_COPY_STRUCT_SB_ADDR_LO_sb_addr_31_0_mask) << SDMA_PKT_COPY_STRUCT_SB_ADDR_LO_sb_addr_31_0_shift)
+
+/*define for SB_ADDR_HI word*/
+/*define for sb_addr_63_32 field*/
+#define SDMA_PKT_COPY_STRUCT_SB_ADDR_HI_sb_addr_63_32_offset 2
+#define SDMA_PKT_COPY_STRUCT_SB_ADDR_HI_sb_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_STRUCT_SB_ADDR_HI_sb_addr_63_32_shift  0
+#define SDMA_PKT_COPY_STRUCT_SB_ADDR_HI_SB_ADDR_63_32(x) (((x) & SDMA_PKT_COPY_STRUCT_SB_ADDR_HI_sb_addr_63_32_mask) << SDMA_PKT_COPY_STRUCT_SB_ADDR_HI_sb_addr_63_32_shift)
+
+/*define for START_INDEX word*/
+/*define for start_index field*/
+#define SDMA_PKT_COPY_STRUCT_START_INDEX_start_index_offset 3
+#define SDMA_PKT_COPY_STRUCT_START_INDEX_start_index_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_STRUCT_START_INDEX_start_index_shift  0
+#define SDMA_PKT_COPY_STRUCT_START_INDEX_START_INDEX(x) (((x) & SDMA_PKT_COPY_STRUCT_START_INDEX_start_index_mask) << SDMA_PKT_COPY_STRUCT_START_INDEX_start_index_shift)
+
+/*define for COUNT word*/
+/*define for count field*/
+#define SDMA_PKT_COPY_STRUCT_COUNT_count_offset 4
+#define SDMA_PKT_COPY_STRUCT_COUNT_count_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_STRUCT_COUNT_count_shift  0
+#define SDMA_PKT_COPY_STRUCT_COUNT_COUNT(x) (((x) & SDMA_PKT_COPY_STRUCT_COUNT_count_mask) << SDMA_PKT_COPY_STRUCT_COUNT_count_shift)
+
+/*define for DW_5 word*/
+/*define for stride field*/
+#define SDMA_PKT_COPY_STRUCT_DW_5_stride_offset 5
+#define SDMA_PKT_COPY_STRUCT_DW_5_stride_mask   0x000007FF
+#define SDMA_PKT_COPY_STRUCT_DW_5_stride_shift  0
+#define SDMA_PKT_COPY_STRUCT_DW_5_STRIDE(x) (((x) & SDMA_PKT_COPY_STRUCT_DW_5_stride_mask) << SDMA_PKT_COPY_STRUCT_DW_5_stride_shift)
+
+/*define for linear_sw field*/
+#define SDMA_PKT_COPY_STRUCT_DW_5_linear_sw_offset 5
+#define SDMA_PKT_COPY_STRUCT_DW_5_linear_sw_mask   0x00000003
+#define SDMA_PKT_COPY_STRUCT_DW_5_linear_sw_shift  16
+#define SDMA_PKT_COPY_STRUCT_DW_5_LINEAR_SW(x) (((x) & SDMA_PKT_COPY_STRUCT_DW_5_linear_sw_mask) << SDMA_PKT_COPY_STRUCT_DW_5_linear_sw_shift)
+
+/*define for linear_cache_policy field*/
+#define SDMA_PKT_COPY_STRUCT_DW_5_linear_cache_policy_offset 5
+#define SDMA_PKT_COPY_STRUCT_DW_5_linear_cache_policy_mask   0x00000007
+#define SDMA_PKT_COPY_STRUCT_DW_5_linear_cache_policy_shift  18
+#define SDMA_PKT_COPY_STRUCT_DW_5_LINEAR_CACHE_POLICY(x) (((x) & SDMA_PKT_COPY_STRUCT_DW_5_linear_cache_policy_mask) << SDMA_PKT_COPY_STRUCT_DW_5_linear_cache_policy_shift)
+
+/*define for struct_sw field*/
+#define SDMA_PKT_COPY_STRUCT_DW_5_struct_sw_offset 5
+#define SDMA_PKT_COPY_STRUCT_DW_5_struct_sw_mask   0x00000003
+#define SDMA_PKT_COPY_STRUCT_DW_5_struct_sw_shift  24
+#define SDMA_PKT_COPY_STRUCT_DW_5_STRUCT_SW(x) (((x) & SDMA_PKT_COPY_STRUCT_DW_5_struct_sw_mask) << SDMA_PKT_COPY_STRUCT_DW_5_struct_sw_shift)
+
+/*define for struct_cache_policy field*/
+#define SDMA_PKT_COPY_STRUCT_DW_5_struct_cache_policy_offset 5
+#define SDMA_PKT_COPY_STRUCT_DW_5_struct_cache_policy_mask   0x00000007
+#define SDMA_PKT_COPY_STRUCT_DW_5_struct_cache_policy_shift  26
+#define SDMA_PKT_COPY_STRUCT_DW_5_STRUCT_CACHE_POLICY(x) (((x) & SDMA_PKT_COPY_STRUCT_DW_5_struct_cache_policy_mask) << SDMA_PKT_COPY_STRUCT_DW_5_struct_cache_policy_shift)
+
+/*define for LINEAR_ADDR_LO word*/
+/*define for linear_addr_31_0 field*/
+#define SDMA_PKT_COPY_STRUCT_LINEAR_ADDR_LO_linear_addr_31_0_offset 6
+#define SDMA_PKT_COPY_STRUCT_LINEAR_ADDR_LO_linear_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_STRUCT_LINEAR_ADDR_LO_linear_addr_31_0_shift  0
+#define SDMA_PKT_COPY_STRUCT_LINEAR_ADDR_LO_LINEAR_ADDR_31_0(x) (((x) & SDMA_PKT_COPY_STRUCT_LINEAR_ADDR_LO_linear_addr_31_0_mask) << SDMA_PKT_COPY_STRUCT_LINEAR_ADDR_LO_linear_addr_31_0_shift)
+
+/*define for LINEAR_ADDR_HI word*/
+/*define for linear_addr_63_32 field*/
+#define SDMA_PKT_COPY_STRUCT_LINEAR_ADDR_HI_linear_addr_63_32_offset 7
+#define SDMA_PKT_COPY_STRUCT_LINEAR_ADDR_HI_linear_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_COPY_STRUCT_LINEAR_ADDR_HI_linear_addr_63_32_shift  0
+#define SDMA_PKT_COPY_STRUCT_LINEAR_ADDR_HI_LINEAR_ADDR_63_32(x) (((x) & SDMA_PKT_COPY_STRUCT_LINEAR_ADDR_HI_linear_addr_63_32_mask) << SDMA_PKT_COPY_STRUCT_LINEAR_ADDR_HI_linear_addr_63_32_shift)
+
+
+/*
+** Definitions for SDMA_PKT_WRITE_UNTILED packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_WRITE_UNTILED_HEADER_op_offset 0
+#define SDMA_PKT_WRITE_UNTILED_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_WRITE_UNTILED_HEADER_op_shift  0
+#define SDMA_PKT_WRITE_UNTILED_HEADER_OP(x) (((x) & SDMA_PKT_WRITE_UNTILED_HEADER_op_mask) << SDMA_PKT_WRITE_UNTILED_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_WRITE_UNTILED_HEADER_sub_op_offset 0
+#define SDMA_PKT_WRITE_UNTILED_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_WRITE_UNTILED_HEADER_sub_op_shift  8
+#define SDMA_PKT_WRITE_UNTILED_HEADER_SUB_OP(x) (((x) & SDMA_PKT_WRITE_UNTILED_HEADER_sub_op_mask) << SDMA_PKT_WRITE_UNTILED_HEADER_sub_op_shift)
+
+/*define for encrypt field*/
+#define SDMA_PKT_WRITE_UNTILED_HEADER_encrypt_offset 0
+#define SDMA_PKT_WRITE_UNTILED_HEADER_encrypt_mask   0x00000001
+#define SDMA_PKT_WRITE_UNTILED_HEADER_encrypt_shift  16
+#define SDMA_PKT_WRITE_UNTILED_HEADER_ENCRYPT(x) (((x) & SDMA_PKT_WRITE_UNTILED_HEADER_encrypt_mask) << SDMA_PKT_WRITE_UNTILED_HEADER_encrypt_shift)
+
+/*define for tmz field*/
+#define SDMA_PKT_WRITE_UNTILED_HEADER_tmz_offset 0
+#define SDMA_PKT_WRITE_UNTILED_HEADER_tmz_mask   0x00000001
+#define SDMA_PKT_WRITE_UNTILED_HEADER_tmz_shift  18
+#define SDMA_PKT_WRITE_UNTILED_HEADER_TMZ(x) (((x) & SDMA_PKT_WRITE_UNTILED_HEADER_tmz_mask) << SDMA_PKT_WRITE_UNTILED_HEADER_tmz_shift)
+
+/*define for cpv field*/
+#define SDMA_PKT_WRITE_UNTILED_HEADER_cpv_offset 0
+#define SDMA_PKT_WRITE_UNTILED_HEADER_cpv_mask   0x00000001
+#define SDMA_PKT_WRITE_UNTILED_HEADER_cpv_shift  28
+#define SDMA_PKT_WRITE_UNTILED_HEADER_CPV(x) (((x) & SDMA_PKT_WRITE_UNTILED_HEADER_cpv_mask) << SDMA_PKT_WRITE_UNTILED_HEADER_cpv_shift)
+
+/*define for DST_ADDR_LO word*/
+/*define for dst_addr_31_0 field*/
+#define SDMA_PKT_WRITE_UNTILED_DST_ADDR_LO_dst_addr_31_0_offset 1
+#define SDMA_PKT_WRITE_UNTILED_DST_ADDR_LO_dst_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_WRITE_UNTILED_DST_ADDR_LO_dst_addr_31_0_shift  0
+#define SDMA_PKT_WRITE_UNTILED_DST_ADDR_LO_DST_ADDR_31_0(x) (((x) & SDMA_PKT_WRITE_UNTILED_DST_ADDR_LO_dst_addr_31_0_mask) << SDMA_PKT_WRITE_UNTILED_DST_ADDR_LO_dst_addr_31_0_shift)
+
+/*define for DST_ADDR_HI word*/
+/*define for dst_addr_63_32 field*/
+#define SDMA_PKT_WRITE_UNTILED_DST_ADDR_HI_dst_addr_63_32_offset 2
+#define SDMA_PKT_WRITE_UNTILED_DST_ADDR_HI_dst_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_WRITE_UNTILED_DST_ADDR_HI_dst_addr_63_32_shift  0
+#define SDMA_PKT_WRITE_UNTILED_DST_ADDR_HI_DST_ADDR_63_32(x) (((x) & SDMA_PKT_WRITE_UNTILED_DST_ADDR_HI_dst_addr_63_32_mask) << SDMA_PKT_WRITE_UNTILED_DST_ADDR_HI_dst_addr_63_32_shift)
+
+/*define for DW_3 word*/
+/*define for count field*/
+#define SDMA_PKT_WRITE_UNTILED_DW_3_count_offset 3
+#define SDMA_PKT_WRITE_UNTILED_DW_3_count_mask   0x000FFFFF
+#define SDMA_PKT_WRITE_UNTILED_DW_3_count_shift  0
+#define SDMA_PKT_WRITE_UNTILED_DW_3_COUNT(x) (((x) & SDMA_PKT_WRITE_UNTILED_DW_3_count_mask) << SDMA_PKT_WRITE_UNTILED_DW_3_count_shift)
+
+/*define for sw field*/
+#define SDMA_PKT_WRITE_UNTILED_DW_3_sw_offset 3
+#define SDMA_PKT_WRITE_UNTILED_DW_3_sw_mask   0x00000003
+#define SDMA_PKT_WRITE_UNTILED_DW_3_sw_shift  24
+#define SDMA_PKT_WRITE_UNTILED_DW_3_SW(x) (((x) & SDMA_PKT_WRITE_UNTILED_DW_3_sw_mask) << SDMA_PKT_WRITE_UNTILED_DW_3_sw_shift)
+
+/*define for cache_policy field*/
+#define SDMA_PKT_WRITE_UNTILED_DW_3_cache_policy_offset 3
+#define SDMA_PKT_WRITE_UNTILED_DW_3_cache_policy_mask   0x00000007
+#define SDMA_PKT_WRITE_UNTILED_DW_3_cache_policy_shift  26
+#define SDMA_PKT_WRITE_UNTILED_DW_3_CACHE_POLICY(x) (((x) & SDMA_PKT_WRITE_UNTILED_DW_3_cache_policy_mask) << SDMA_PKT_WRITE_UNTILED_DW_3_cache_policy_shift)
+
+/*define for DATA0 word*/
+/*define for data0 field*/
+#define SDMA_PKT_WRITE_UNTILED_DATA0_data0_offset 4
+#define SDMA_PKT_WRITE_UNTILED_DATA0_data0_mask   0xFFFFFFFF
+#define SDMA_PKT_WRITE_UNTILED_DATA0_data0_shift  0
+#define SDMA_PKT_WRITE_UNTILED_DATA0_DATA0(x) (((x) & SDMA_PKT_WRITE_UNTILED_DATA0_data0_mask) << SDMA_PKT_WRITE_UNTILED_DATA0_data0_shift)
+
+
+/*
+** Definitions for SDMA_PKT_WRITE_TILED packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_WRITE_TILED_HEADER_op_offset 0
+#define SDMA_PKT_WRITE_TILED_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_WRITE_TILED_HEADER_op_shift  0
+#define SDMA_PKT_WRITE_TILED_HEADER_OP(x) (((x) & SDMA_PKT_WRITE_TILED_HEADER_op_mask) << SDMA_PKT_WRITE_TILED_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_WRITE_TILED_HEADER_sub_op_offset 0
+#define SDMA_PKT_WRITE_TILED_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_WRITE_TILED_HEADER_sub_op_shift  8
+#define SDMA_PKT_WRITE_TILED_HEADER_SUB_OP(x) (((x) & SDMA_PKT_WRITE_TILED_HEADER_sub_op_mask) << SDMA_PKT_WRITE_TILED_HEADER_sub_op_shift)
+
+/*define for encrypt field*/
+#define SDMA_PKT_WRITE_TILED_HEADER_encrypt_offset 0
+#define SDMA_PKT_WRITE_TILED_HEADER_encrypt_mask   0x00000001
+#define SDMA_PKT_WRITE_TILED_HEADER_encrypt_shift  16
+#define SDMA_PKT_WRITE_TILED_HEADER_ENCRYPT(x) (((x) & SDMA_PKT_WRITE_TILED_HEADER_encrypt_mask) << SDMA_PKT_WRITE_TILED_HEADER_encrypt_shift)
+
+/*define for tmz field*/
+#define SDMA_PKT_WRITE_TILED_HEADER_tmz_offset 0
+#define SDMA_PKT_WRITE_TILED_HEADER_tmz_mask   0x00000001
+#define SDMA_PKT_WRITE_TILED_HEADER_tmz_shift  18
+#define SDMA_PKT_WRITE_TILED_HEADER_TMZ(x) (((x) & SDMA_PKT_WRITE_TILED_HEADER_tmz_mask) << SDMA_PKT_WRITE_TILED_HEADER_tmz_shift)
+
+/*define for cpv field*/
+#define SDMA_PKT_WRITE_TILED_HEADER_cpv_offset 0
+#define SDMA_PKT_WRITE_TILED_HEADER_cpv_mask   0x00000001
+#define SDMA_PKT_WRITE_TILED_HEADER_cpv_shift  28
+#define SDMA_PKT_WRITE_TILED_HEADER_CPV(x) (((x) & SDMA_PKT_WRITE_TILED_HEADER_cpv_mask) << SDMA_PKT_WRITE_TILED_HEADER_cpv_shift)
+
+/*define for DST_ADDR_LO word*/
+/*define for dst_addr_31_0 field*/
+#define SDMA_PKT_WRITE_TILED_DST_ADDR_LO_dst_addr_31_0_offset 1
+#define SDMA_PKT_WRITE_TILED_DST_ADDR_LO_dst_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_WRITE_TILED_DST_ADDR_LO_dst_addr_31_0_shift  0
+#define SDMA_PKT_WRITE_TILED_DST_ADDR_LO_DST_ADDR_31_0(x) (((x) & SDMA_PKT_WRITE_TILED_DST_ADDR_LO_dst_addr_31_0_mask) << SDMA_PKT_WRITE_TILED_DST_ADDR_LO_dst_addr_31_0_shift)
+
+/*define for DST_ADDR_HI word*/
+/*define for dst_addr_63_32 field*/
+#define SDMA_PKT_WRITE_TILED_DST_ADDR_HI_dst_addr_63_32_offset 2
+#define SDMA_PKT_WRITE_TILED_DST_ADDR_HI_dst_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_WRITE_TILED_DST_ADDR_HI_dst_addr_63_32_shift  0
+#define SDMA_PKT_WRITE_TILED_DST_ADDR_HI_DST_ADDR_63_32(x) (((x) & SDMA_PKT_WRITE_TILED_DST_ADDR_HI_dst_addr_63_32_mask) << SDMA_PKT_WRITE_TILED_DST_ADDR_HI_dst_addr_63_32_shift)
+
+/*define for DW_3 word*/
+/*define for width field*/
+#define SDMA_PKT_WRITE_TILED_DW_3_width_offset 3
+#define SDMA_PKT_WRITE_TILED_DW_3_width_mask   0x00003FFF
+#define SDMA_PKT_WRITE_TILED_DW_3_width_shift  0
+#define SDMA_PKT_WRITE_TILED_DW_3_WIDTH(x) (((x) & SDMA_PKT_WRITE_TILED_DW_3_width_mask) << SDMA_PKT_WRITE_TILED_DW_3_width_shift)
+
+/*define for DW_4 word*/
+/*define for height field*/
+#define SDMA_PKT_WRITE_TILED_DW_4_height_offset 4
+#define SDMA_PKT_WRITE_TILED_DW_4_height_mask   0x00003FFF
+#define SDMA_PKT_WRITE_TILED_DW_4_height_shift  0
+#define SDMA_PKT_WRITE_TILED_DW_4_HEIGHT(x) (((x) & SDMA_PKT_WRITE_TILED_DW_4_height_mask) << SDMA_PKT_WRITE_TILED_DW_4_height_shift)
+
+/*define for depth field*/
+#define SDMA_PKT_WRITE_TILED_DW_4_depth_offset 4
+#define SDMA_PKT_WRITE_TILED_DW_4_depth_mask   0x00001FFF
+#define SDMA_PKT_WRITE_TILED_DW_4_depth_shift  16
+#define SDMA_PKT_WRITE_TILED_DW_4_DEPTH(x) (((x) & SDMA_PKT_WRITE_TILED_DW_4_depth_mask) << SDMA_PKT_WRITE_TILED_DW_4_depth_shift)
+
+/*define for DW_5 word*/
+/*define for element_size field*/
+#define SDMA_PKT_WRITE_TILED_DW_5_element_size_offset 5
+#define SDMA_PKT_WRITE_TILED_DW_5_element_size_mask   0x00000007
+#define SDMA_PKT_WRITE_TILED_DW_5_element_size_shift  0
+#define SDMA_PKT_WRITE_TILED_DW_5_ELEMENT_SIZE(x) (((x) & SDMA_PKT_WRITE_TILED_DW_5_element_size_mask) << SDMA_PKT_WRITE_TILED_DW_5_element_size_shift)
+
+/*define for swizzle_mode field*/
+#define SDMA_PKT_WRITE_TILED_DW_5_swizzle_mode_offset 5
+#define SDMA_PKT_WRITE_TILED_DW_5_swizzle_mode_mask   0x0000001F
+#define SDMA_PKT_WRITE_TILED_DW_5_swizzle_mode_shift  3
+#define SDMA_PKT_WRITE_TILED_DW_5_SWIZZLE_MODE(x) (((x) & SDMA_PKT_WRITE_TILED_DW_5_swizzle_mode_mask) << SDMA_PKT_WRITE_TILED_DW_5_swizzle_mode_shift)
+
+/*define for dimension field*/
+#define SDMA_PKT_WRITE_TILED_DW_5_dimension_offset 5
+#define SDMA_PKT_WRITE_TILED_DW_5_dimension_mask   0x00000003
+#define SDMA_PKT_WRITE_TILED_DW_5_dimension_shift  9
+#define SDMA_PKT_WRITE_TILED_DW_5_DIMENSION(x) (((x) & SDMA_PKT_WRITE_TILED_DW_5_dimension_mask) << SDMA_PKT_WRITE_TILED_DW_5_dimension_shift)
+
+/*define for mip_max field*/
+#define SDMA_PKT_WRITE_TILED_DW_5_mip_max_offset 5
+#define SDMA_PKT_WRITE_TILED_DW_5_mip_max_mask   0x0000000F
+#define SDMA_PKT_WRITE_TILED_DW_5_mip_max_shift  16
+#define SDMA_PKT_WRITE_TILED_DW_5_MIP_MAX(x) (((x) & SDMA_PKT_WRITE_TILED_DW_5_mip_max_mask) << SDMA_PKT_WRITE_TILED_DW_5_mip_max_shift)
+
+/*define for DW_6 word*/
+/*define for x field*/
+#define SDMA_PKT_WRITE_TILED_DW_6_x_offset 6
+#define SDMA_PKT_WRITE_TILED_DW_6_x_mask   0x00003FFF
+#define SDMA_PKT_WRITE_TILED_DW_6_x_shift  0
+#define SDMA_PKT_WRITE_TILED_DW_6_X(x) (((x) & SDMA_PKT_WRITE_TILED_DW_6_x_mask) << SDMA_PKT_WRITE_TILED_DW_6_x_shift)
+
+/*define for y field*/
+#define SDMA_PKT_WRITE_TILED_DW_6_y_offset 6
+#define SDMA_PKT_WRITE_TILED_DW_6_y_mask   0x00003FFF
+#define SDMA_PKT_WRITE_TILED_DW_6_y_shift  16
+#define SDMA_PKT_WRITE_TILED_DW_6_Y(x) (((x) & SDMA_PKT_WRITE_TILED_DW_6_y_mask) << SDMA_PKT_WRITE_TILED_DW_6_y_shift)
+
+/*define for DW_7 word*/
+/*define for z field*/
+#define SDMA_PKT_WRITE_TILED_DW_7_z_offset 7
+#define SDMA_PKT_WRITE_TILED_DW_7_z_mask   0x00001FFF
+#define SDMA_PKT_WRITE_TILED_DW_7_z_shift  0
+#define SDMA_PKT_WRITE_TILED_DW_7_Z(x) (((x) & SDMA_PKT_WRITE_TILED_DW_7_z_mask) << SDMA_PKT_WRITE_TILED_DW_7_z_shift)
+
+/*define for sw field*/
+#define SDMA_PKT_WRITE_TILED_DW_7_sw_offset 7
+#define SDMA_PKT_WRITE_TILED_DW_7_sw_mask   0x00000003
+#define SDMA_PKT_WRITE_TILED_DW_7_sw_shift  24
+#define SDMA_PKT_WRITE_TILED_DW_7_SW(x) (((x) & SDMA_PKT_WRITE_TILED_DW_7_sw_mask) << SDMA_PKT_WRITE_TILED_DW_7_sw_shift)
+
+/*define for cache_policy field*/
+#define SDMA_PKT_WRITE_TILED_DW_7_cache_policy_offset 7
+#define SDMA_PKT_WRITE_TILED_DW_7_cache_policy_mask   0x00000007
+#define SDMA_PKT_WRITE_TILED_DW_7_cache_policy_shift  26
+#define SDMA_PKT_WRITE_TILED_DW_7_CACHE_POLICY(x) (((x) & SDMA_PKT_WRITE_TILED_DW_7_cache_policy_mask) << SDMA_PKT_WRITE_TILED_DW_7_cache_policy_shift)
+
+/*define for COUNT word*/
+/*define for count field*/
+#define SDMA_PKT_WRITE_TILED_COUNT_count_offset 8
+#define SDMA_PKT_WRITE_TILED_COUNT_count_mask   0x000FFFFF
+#define SDMA_PKT_WRITE_TILED_COUNT_count_shift  0
+#define SDMA_PKT_WRITE_TILED_COUNT_COUNT(x) (((x) & SDMA_PKT_WRITE_TILED_COUNT_count_mask) << SDMA_PKT_WRITE_TILED_COUNT_count_shift)
+
+/*define for DATA0 word*/
+/*define for data0 field*/
+#define SDMA_PKT_WRITE_TILED_DATA0_data0_offset 9
+#define SDMA_PKT_WRITE_TILED_DATA0_data0_mask   0xFFFFFFFF
+#define SDMA_PKT_WRITE_TILED_DATA0_data0_shift  0
+#define SDMA_PKT_WRITE_TILED_DATA0_DATA0(x) (((x) & SDMA_PKT_WRITE_TILED_DATA0_data0_mask) << SDMA_PKT_WRITE_TILED_DATA0_data0_shift)
+
+
+/*
+** Definitions for SDMA_PKT_WRITE_TILED_BC packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_WRITE_TILED_BC_HEADER_op_offset 0
+#define SDMA_PKT_WRITE_TILED_BC_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_WRITE_TILED_BC_HEADER_op_shift  0
+#define SDMA_PKT_WRITE_TILED_BC_HEADER_OP(x) (((x) & SDMA_PKT_WRITE_TILED_BC_HEADER_op_mask) << SDMA_PKT_WRITE_TILED_BC_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_WRITE_TILED_BC_HEADER_sub_op_offset 0
+#define SDMA_PKT_WRITE_TILED_BC_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_WRITE_TILED_BC_HEADER_sub_op_shift  8
+#define SDMA_PKT_WRITE_TILED_BC_HEADER_SUB_OP(x) (((x) & SDMA_PKT_WRITE_TILED_BC_HEADER_sub_op_mask) << SDMA_PKT_WRITE_TILED_BC_HEADER_sub_op_shift)
+
+/*define for DST_ADDR_LO word*/
+/*define for dst_addr_31_0 field*/
+#define SDMA_PKT_WRITE_TILED_BC_DST_ADDR_LO_dst_addr_31_0_offset 1
+#define SDMA_PKT_WRITE_TILED_BC_DST_ADDR_LO_dst_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_WRITE_TILED_BC_DST_ADDR_LO_dst_addr_31_0_shift  0
+#define SDMA_PKT_WRITE_TILED_BC_DST_ADDR_LO_DST_ADDR_31_0(x) (((x) & SDMA_PKT_WRITE_TILED_BC_DST_ADDR_LO_dst_addr_31_0_mask) << SDMA_PKT_WRITE_TILED_BC_DST_ADDR_LO_dst_addr_31_0_shift)
+
+/*define for DST_ADDR_HI word*/
+/*define for dst_addr_63_32 field*/
+#define SDMA_PKT_WRITE_TILED_BC_DST_ADDR_HI_dst_addr_63_32_offset 2
+#define SDMA_PKT_WRITE_TILED_BC_DST_ADDR_HI_dst_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_WRITE_TILED_BC_DST_ADDR_HI_dst_addr_63_32_shift  0
+#define SDMA_PKT_WRITE_TILED_BC_DST_ADDR_HI_DST_ADDR_63_32(x) (((x) & SDMA_PKT_WRITE_TILED_BC_DST_ADDR_HI_dst_addr_63_32_mask) << SDMA_PKT_WRITE_TILED_BC_DST_ADDR_HI_dst_addr_63_32_shift)
+
+/*define for DW_3 word*/
+/*define for width field*/
+#define SDMA_PKT_WRITE_TILED_BC_DW_3_width_offset 3
+#define SDMA_PKT_WRITE_TILED_BC_DW_3_width_mask   0x00003FFF
+#define SDMA_PKT_WRITE_TILED_BC_DW_3_width_shift  0
+#define SDMA_PKT_WRITE_TILED_BC_DW_3_WIDTH(x) (((x) & SDMA_PKT_WRITE_TILED_BC_DW_3_width_mask) << SDMA_PKT_WRITE_TILED_BC_DW_3_width_shift)
+
+/*define for DW_4 word*/
+/*define for height field*/
+#define SDMA_PKT_WRITE_TILED_BC_DW_4_height_offset 4
+#define SDMA_PKT_WRITE_TILED_BC_DW_4_height_mask   0x00003FFF
+#define SDMA_PKT_WRITE_TILED_BC_DW_4_height_shift  0
+#define SDMA_PKT_WRITE_TILED_BC_DW_4_HEIGHT(x) (((x) & SDMA_PKT_WRITE_TILED_BC_DW_4_height_mask) << SDMA_PKT_WRITE_TILED_BC_DW_4_height_shift)
+
+/*define for depth field*/
+#define SDMA_PKT_WRITE_TILED_BC_DW_4_depth_offset 4
+#define SDMA_PKT_WRITE_TILED_BC_DW_4_depth_mask   0x000007FF
+#define SDMA_PKT_WRITE_TILED_BC_DW_4_depth_shift  16
+#define SDMA_PKT_WRITE_TILED_BC_DW_4_DEPTH(x) (((x) & SDMA_PKT_WRITE_TILED_BC_DW_4_depth_mask) << SDMA_PKT_WRITE_TILED_BC_DW_4_depth_shift)
+
+/*define for DW_5 word*/
+/*define for element_size field*/
+#define SDMA_PKT_WRITE_TILED_BC_DW_5_element_size_offset 5
+#define SDMA_PKT_WRITE_TILED_BC_DW_5_element_size_mask   0x00000007
+#define SDMA_PKT_WRITE_TILED_BC_DW_5_element_size_shift  0
+#define SDMA_PKT_WRITE_TILED_BC_DW_5_ELEMENT_SIZE(x) (((x) & SDMA_PKT_WRITE_TILED_BC_DW_5_element_size_mask) << SDMA_PKT_WRITE_TILED_BC_DW_5_element_size_shift)
+
+/*define for array_mode field*/
+#define SDMA_PKT_WRITE_TILED_BC_DW_5_array_mode_offset 5
+#define SDMA_PKT_WRITE_TILED_BC_DW_5_array_mode_mask   0x0000000F
+#define SDMA_PKT_WRITE_TILED_BC_DW_5_array_mode_shift  3
+#define SDMA_PKT_WRITE_TILED_BC_DW_5_ARRAY_MODE(x) (((x) & SDMA_PKT_WRITE_TILED_BC_DW_5_array_mode_mask) << SDMA_PKT_WRITE_TILED_BC_DW_5_array_mode_shift)
+
+/*define for mit_mode field*/
+#define SDMA_PKT_WRITE_TILED_BC_DW_5_mit_mode_offset 5
+#define SDMA_PKT_WRITE_TILED_BC_DW_5_mit_mode_mask   0x00000007
+#define SDMA_PKT_WRITE_TILED_BC_DW_5_mit_mode_shift  8
+#define SDMA_PKT_WRITE_TILED_BC_DW_5_MIT_MODE(x) (((x) & SDMA_PKT_WRITE_TILED_BC_DW_5_mit_mode_mask) << SDMA_PKT_WRITE_TILED_BC_DW_5_mit_mode_shift)
+
+/*define for tilesplit_size field*/
+#define SDMA_PKT_WRITE_TILED_BC_DW_5_tilesplit_size_offset 5
+#define SDMA_PKT_WRITE_TILED_BC_DW_5_tilesplit_size_mask   0x00000007
+#define SDMA_PKT_WRITE_TILED_BC_DW_5_tilesplit_size_shift  11
+#define SDMA_PKT_WRITE_TILED_BC_DW_5_TILESPLIT_SIZE(x) (((x) & SDMA_PKT_WRITE_TILED_BC_DW_5_tilesplit_size_mask) << SDMA_PKT_WRITE_TILED_BC_DW_5_tilesplit_size_shift)
+
+/*define for bank_w field*/
+#define SDMA_PKT_WRITE_TILED_BC_DW_5_bank_w_offset 5
+#define SDMA_PKT_WRITE_TILED_BC_DW_5_bank_w_mask   0x00000003
+#define SDMA_PKT_WRITE_TILED_BC_DW_5_bank_w_shift  15
+#define SDMA_PKT_WRITE_TILED_BC_DW_5_BANK_W(x) (((x) & SDMA_PKT_WRITE_TILED_BC_DW_5_bank_w_mask) << SDMA_PKT_WRITE_TILED_BC_DW_5_bank_w_shift)
+
+/*define for bank_h field*/
+#define SDMA_PKT_WRITE_TILED_BC_DW_5_bank_h_offset 5
+#define SDMA_PKT_WRITE_TILED_BC_DW_5_bank_h_mask   0x00000003
+#define SDMA_PKT_WRITE_TILED_BC_DW_5_bank_h_shift  18
+#define SDMA_PKT_WRITE_TILED_BC_DW_5_BANK_H(x) (((x) & SDMA_PKT_WRITE_TILED_BC_DW_5_bank_h_mask) << SDMA_PKT_WRITE_TILED_BC_DW_5_bank_h_shift)
+
+/*define for num_bank field*/
+#define SDMA_PKT_WRITE_TILED_BC_DW_5_num_bank_offset 5
+#define SDMA_PKT_WRITE_TILED_BC_DW_5_num_bank_mask   0x00000003
+#define SDMA_PKT_WRITE_TILED_BC_DW_5_num_bank_shift  21
+#define SDMA_PKT_WRITE_TILED_BC_DW_5_NUM_BANK(x) (((x) & SDMA_PKT_WRITE_TILED_BC_DW_5_num_bank_mask) << SDMA_PKT_WRITE_TILED_BC_DW_5_num_bank_shift)
+
+/*define for mat_aspt field*/
+#define SDMA_PKT_WRITE_TILED_BC_DW_5_mat_aspt_offset 5
+#define SDMA_PKT_WRITE_TILED_BC_DW_5_mat_aspt_mask   0x00000003
+#define SDMA_PKT_WRITE_TILED_BC_DW_5_mat_aspt_shift  24
+#define SDMA_PKT_WRITE_TILED_BC_DW_5_MAT_ASPT(x) (((x) & SDMA_PKT_WRITE_TILED_BC_DW_5_mat_aspt_mask) << SDMA_PKT_WRITE_TILED_BC_DW_5_mat_aspt_shift)
+
+/*define for pipe_config field*/
+#define SDMA_PKT_WRITE_TILED_BC_DW_5_pipe_config_offset 5
+#define SDMA_PKT_WRITE_TILED_BC_DW_5_pipe_config_mask   0x0000001F
+#define SDMA_PKT_WRITE_TILED_BC_DW_5_pipe_config_shift  26
+#define SDMA_PKT_WRITE_TILED_BC_DW_5_PIPE_CONFIG(x) (((x) & SDMA_PKT_WRITE_TILED_BC_DW_5_pipe_config_mask) << SDMA_PKT_WRITE_TILED_BC_DW_5_pipe_config_shift)
+
+/*define for DW_6 word*/
+/*define for x field*/
+#define SDMA_PKT_WRITE_TILED_BC_DW_6_x_offset 6
+#define SDMA_PKT_WRITE_TILED_BC_DW_6_x_mask   0x00003FFF
+#define SDMA_PKT_WRITE_TILED_BC_DW_6_x_shift  0
+#define SDMA_PKT_WRITE_TILED_BC_DW_6_X(x) (((x) & SDMA_PKT_WRITE_TILED_BC_DW_6_x_mask) << SDMA_PKT_WRITE_TILED_BC_DW_6_x_shift)
+
+/*define for y field*/
+#define SDMA_PKT_WRITE_TILED_BC_DW_6_y_offset 6
+#define SDMA_PKT_WRITE_TILED_BC_DW_6_y_mask   0x00003FFF
+#define SDMA_PKT_WRITE_TILED_BC_DW_6_y_shift  16
+#define SDMA_PKT_WRITE_TILED_BC_DW_6_Y(x) (((x) & SDMA_PKT_WRITE_TILED_BC_DW_6_y_mask) << SDMA_PKT_WRITE_TILED_BC_DW_6_y_shift)
+
+/*define for DW_7 word*/
+/*define for z field*/
+#define SDMA_PKT_WRITE_TILED_BC_DW_7_z_offset 7
+#define SDMA_PKT_WRITE_TILED_BC_DW_7_z_mask   0x000007FF
+#define SDMA_PKT_WRITE_TILED_BC_DW_7_z_shift  0
+#define SDMA_PKT_WRITE_TILED_BC_DW_7_Z(x) (((x) & SDMA_PKT_WRITE_TILED_BC_DW_7_z_mask) << SDMA_PKT_WRITE_TILED_BC_DW_7_z_shift)
+
+/*define for sw field*/
+#define SDMA_PKT_WRITE_TILED_BC_DW_7_sw_offset 7
+#define SDMA_PKT_WRITE_TILED_BC_DW_7_sw_mask   0x00000003
+#define SDMA_PKT_WRITE_TILED_BC_DW_7_sw_shift  24
+#define SDMA_PKT_WRITE_TILED_BC_DW_7_SW(x) (((x) & SDMA_PKT_WRITE_TILED_BC_DW_7_sw_mask) << SDMA_PKT_WRITE_TILED_BC_DW_7_sw_shift)
+
+/*define for COUNT word*/
+/*define for count field*/
+#define SDMA_PKT_WRITE_TILED_BC_COUNT_count_offset 8
+#define SDMA_PKT_WRITE_TILED_BC_COUNT_count_mask   0x000FFFFF
+#define SDMA_PKT_WRITE_TILED_BC_COUNT_count_shift  2
+#define SDMA_PKT_WRITE_TILED_BC_COUNT_COUNT(x) (((x) & SDMA_PKT_WRITE_TILED_BC_COUNT_count_mask) << SDMA_PKT_WRITE_TILED_BC_COUNT_count_shift)
+
+/*define for DATA0 word*/
+/*define for data0 field*/
+#define SDMA_PKT_WRITE_TILED_BC_DATA0_data0_offset 9
+#define SDMA_PKT_WRITE_TILED_BC_DATA0_data0_mask   0xFFFFFFFF
+#define SDMA_PKT_WRITE_TILED_BC_DATA0_data0_shift  0
+#define SDMA_PKT_WRITE_TILED_BC_DATA0_DATA0(x) (((x) & SDMA_PKT_WRITE_TILED_BC_DATA0_data0_mask) << SDMA_PKT_WRITE_TILED_BC_DATA0_data0_shift)
+
+
+/*
+** Definitions for SDMA_PKT_PTEPDE_COPY packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_PTEPDE_COPY_HEADER_op_offset 0
+#define SDMA_PKT_PTEPDE_COPY_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_PTEPDE_COPY_HEADER_op_shift  0
+#define SDMA_PKT_PTEPDE_COPY_HEADER_OP(x) (((x) & SDMA_PKT_PTEPDE_COPY_HEADER_op_mask) << SDMA_PKT_PTEPDE_COPY_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_PTEPDE_COPY_HEADER_sub_op_offset 0
+#define SDMA_PKT_PTEPDE_COPY_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_PTEPDE_COPY_HEADER_sub_op_shift  8
+#define SDMA_PKT_PTEPDE_COPY_HEADER_SUB_OP(x) (((x) & SDMA_PKT_PTEPDE_COPY_HEADER_sub_op_mask) << SDMA_PKT_PTEPDE_COPY_HEADER_sub_op_shift)
+
+/*define for tmz field*/
+#define SDMA_PKT_PTEPDE_COPY_HEADER_tmz_offset 0
+#define SDMA_PKT_PTEPDE_COPY_HEADER_tmz_mask   0x00000001
+#define SDMA_PKT_PTEPDE_COPY_HEADER_tmz_shift  18
+#define SDMA_PKT_PTEPDE_COPY_HEADER_TMZ(x) (((x) & SDMA_PKT_PTEPDE_COPY_HEADER_tmz_mask) << SDMA_PKT_PTEPDE_COPY_HEADER_tmz_shift)
+
+/*define for cpv field*/
+#define SDMA_PKT_PTEPDE_COPY_HEADER_cpv_offset 0
+#define SDMA_PKT_PTEPDE_COPY_HEADER_cpv_mask   0x00000001
+#define SDMA_PKT_PTEPDE_COPY_HEADER_cpv_shift  28
+#define SDMA_PKT_PTEPDE_COPY_HEADER_CPV(x) (((x) & SDMA_PKT_PTEPDE_COPY_HEADER_cpv_mask) << SDMA_PKT_PTEPDE_COPY_HEADER_cpv_shift)
+
+/*define for ptepde_op field*/
+#define SDMA_PKT_PTEPDE_COPY_HEADER_ptepde_op_offset 0
+#define SDMA_PKT_PTEPDE_COPY_HEADER_ptepde_op_mask   0x00000001
+#define SDMA_PKT_PTEPDE_COPY_HEADER_ptepde_op_shift  31
+#define SDMA_PKT_PTEPDE_COPY_HEADER_PTEPDE_OP(x) (((x) & SDMA_PKT_PTEPDE_COPY_HEADER_ptepde_op_mask) << SDMA_PKT_PTEPDE_COPY_HEADER_ptepde_op_shift)
+
+/*define for SRC_ADDR_LO word*/
+/*define for src_addr_31_0 field*/
+#define SDMA_PKT_PTEPDE_COPY_SRC_ADDR_LO_src_addr_31_0_offset 1
+#define SDMA_PKT_PTEPDE_COPY_SRC_ADDR_LO_src_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_PTEPDE_COPY_SRC_ADDR_LO_src_addr_31_0_shift  0
+#define SDMA_PKT_PTEPDE_COPY_SRC_ADDR_LO_SRC_ADDR_31_0(x) (((x) & SDMA_PKT_PTEPDE_COPY_SRC_ADDR_LO_src_addr_31_0_mask) << SDMA_PKT_PTEPDE_COPY_SRC_ADDR_LO_src_addr_31_0_shift)
+
+/*define for SRC_ADDR_HI word*/
+/*define for src_addr_63_32 field*/
+#define SDMA_PKT_PTEPDE_COPY_SRC_ADDR_HI_src_addr_63_32_offset 2
+#define SDMA_PKT_PTEPDE_COPY_SRC_ADDR_HI_src_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_PTEPDE_COPY_SRC_ADDR_HI_src_addr_63_32_shift  0
+#define SDMA_PKT_PTEPDE_COPY_SRC_ADDR_HI_SRC_ADDR_63_32(x) (((x) & SDMA_PKT_PTEPDE_COPY_SRC_ADDR_HI_src_addr_63_32_mask) << SDMA_PKT_PTEPDE_COPY_SRC_ADDR_HI_src_addr_63_32_shift)
+
+/*define for DST_ADDR_LO word*/
+/*define for dst_addr_31_0 field*/
+#define SDMA_PKT_PTEPDE_COPY_DST_ADDR_LO_dst_addr_31_0_offset 3
+#define SDMA_PKT_PTEPDE_COPY_DST_ADDR_LO_dst_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_PTEPDE_COPY_DST_ADDR_LO_dst_addr_31_0_shift  0
+#define SDMA_PKT_PTEPDE_COPY_DST_ADDR_LO_DST_ADDR_31_0(x) (((x) & SDMA_PKT_PTEPDE_COPY_DST_ADDR_LO_dst_addr_31_0_mask) << SDMA_PKT_PTEPDE_COPY_DST_ADDR_LO_dst_addr_31_0_shift)
+
+/*define for DST_ADDR_HI word*/
+/*define for dst_addr_63_32 field*/
+#define SDMA_PKT_PTEPDE_COPY_DST_ADDR_HI_dst_addr_63_32_offset 4
+#define SDMA_PKT_PTEPDE_COPY_DST_ADDR_HI_dst_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_PTEPDE_COPY_DST_ADDR_HI_dst_addr_63_32_shift  0
+#define SDMA_PKT_PTEPDE_COPY_DST_ADDR_HI_DST_ADDR_63_32(x) (((x) & SDMA_PKT_PTEPDE_COPY_DST_ADDR_HI_dst_addr_63_32_mask) << SDMA_PKT_PTEPDE_COPY_DST_ADDR_HI_dst_addr_63_32_shift)
+
+/*define for MASK_DW0 word*/
+/*define for mask_dw0 field*/
+#define SDMA_PKT_PTEPDE_COPY_MASK_DW0_mask_dw0_offset 5
+#define SDMA_PKT_PTEPDE_COPY_MASK_DW0_mask_dw0_mask   0xFFFFFFFF
+#define SDMA_PKT_PTEPDE_COPY_MASK_DW0_mask_dw0_shift  0
+#define SDMA_PKT_PTEPDE_COPY_MASK_DW0_MASK_DW0(x) (((x) & SDMA_PKT_PTEPDE_COPY_MASK_DW0_mask_dw0_mask) << SDMA_PKT_PTEPDE_COPY_MASK_DW0_mask_dw0_shift)
+
+/*define for MASK_DW1 word*/
+/*define for mask_dw1 field*/
+#define SDMA_PKT_PTEPDE_COPY_MASK_DW1_mask_dw1_offset 6
+#define SDMA_PKT_PTEPDE_COPY_MASK_DW1_mask_dw1_mask   0xFFFFFFFF
+#define SDMA_PKT_PTEPDE_COPY_MASK_DW1_mask_dw1_shift  0
+#define SDMA_PKT_PTEPDE_COPY_MASK_DW1_MASK_DW1(x) (((x) & SDMA_PKT_PTEPDE_COPY_MASK_DW1_mask_dw1_mask) << SDMA_PKT_PTEPDE_COPY_MASK_DW1_mask_dw1_shift)
+
+/*define for COUNT word*/
+/*define for count field*/
+#define SDMA_PKT_PTEPDE_COPY_COUNT_count_offset 7
+#define SDMA_PKT_PTEPDE_COPY_COUNT_count_mask   0x0007FFFF
+#define SDMA_PKT_PTEPDE_COPY_COUNT_count_shift  0
+#define SDMA_PKT_PTEPDE_COPY_COUNT_COUNT(x) (((x) & SDMA_PKT_PTEPDE_COPY_COUNT_count_mask) << SDMA_PKT_PTEPDE_COPY_COUNT_count_shift)
+
+/*define for dst_cache_policy field*/
+#define SDMA_PKT_PTEPDE_COPY_COUNT_dst_cache_policy_offset 7
+#define SDMA_PKT_PTEPDE_COPY_COUNT_dst_cache_policy_mask   0x00000007
+#define SDMA_PKT_PTEPDE_COPY_COUNT_dst_cache_policy_shift  22
+#define SDMA_PKT_PTEPDE_COPY_COUNT_DST_CACHE_POLICY(x) (((x) & SDMA_PKT_PTEPDE_COPY_COUNT_dst_cache_policy_mask) << SDMA_PKT_PTEPDE_COPY_COUNT_dst_cache_policy_shift)
+
+/*define for src_cache_policy field*/
+#define SDMA_PKT_PTEPDE_COPY_COUNT_src_cache_policy_offset 7
+#define SDMA_PKT_PTEPDE_COPY_COUNT_src_cache_policy_mask   0x00000007
+#define SDMA_PKT_PTEPDE_COPY_COUNT_src_cache_policy_shift  29
+#define SDMA_PKT_PTEPDE_COPY_COUNT_SRC_CACHE_POLICY(x) (((x) & SDMA_PKT_PTEPDE_COPY_COUNT_src_cache_policy_mask) << SDMA_PKT_PTEPDE_COPY_COUNT_src_cache_policy_shift)
+
+
+/*
+** Definitions for SDMA_PKT_PTEPDE_COPY_BACKWARDS packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_HEADER_op_offset 0
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_HEADER_op_shift  0
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_HEADER_OP(x) (((x) & SDMA_PKT_PTEPDE_COPY_BACKWARDS_HEADER_op_mask) << SDMA_PKT_PTEPDE_COPY_BACKWARDS_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_HEADER_sub_op_offset 0
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_HEADER_sub_op_shift  8
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_HEADER_SUB_OP(x) (((x) & SDMA_PKT_PTEPDE_COPY_BACKWARDS_HEADER_sub_op_mask) << SDMA_PKT_PTEPDE_COPY_BACKWARDS_HEADER_sub_op_shift)
+
+/*define for pte_size field*/
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_HEADER_pte_size_offset 0
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_HEADER_pte_size_mask   0x00000003
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_HEADER_pte_size_shift  28
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_HEADER_PTE_SIZE(x) (((x) & SDMA_PKT_PTEPDE_COPY_BACKWARDS_HEADER_pte_size_mask) << SDMA_PKT_PTEPDE_COPY_BACKWARDS_HEADER_pte_size_shift)
+
+/*define for direction field*/
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_HEADER_direction_offset 0
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_HEADER_direction_mask   0x00000001
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_HEADER_direction_shift  30
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_HEADER_DIRECTION(x) (((x) & SDMA_PKT_PTEPDE_COPY_BACKWARDS_HEADER_direction_mask) << SDMA_PKT_PTEPDE_COPY_BACKWARDS_HEADER_direction_shift)
+
+/*define for ptepde_op field*/
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_HEADER_ptepde_op_offset 0
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_HEADER_ptepde_op_mask   0x00000001
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_HEADER_ptepde_op_shift  31
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_HEADER_PTEPDE_OP(x) (((x) & SDMA_PKT_PTEPDE_COPY_BACKWARDS_HEADER_ptepde_op_mask) << SDMA_PKT_PTEPDE_COPY_BACKWARDS_HEADER_ptepde_op_shift)
+
+/*define for SRC_ADDR_LO word*/
+/*define for src_addr_31_0 field*/
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_SRC_ADDR_LO_src_addr_31_0_offset 1
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_SRC_ADDR_LO_src_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_SRC_ADDR_LO_src_addr_31_0_shift  0
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_SRC_ADDR_LO_SRC_ADDR_31_0(x) (((x) & SDMA_PKT_PTEPDE_COPY_BACKWARDS_SRC_ADDR_LO_src_addr_31_0_mask) << SDMA_PKT_PTEPDE_COPY_BACKWARDS_SRC_ADDR_LO_src_addr_31_0_shift)
+
+/*define for SRC_ADDR_HI word*/
+/*define for src_addr_63_32 field*/
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_SRC_ADDR_HI_src_addr_63_32_offset 2
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_SRC_ADDR_HI_src_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_SRC_ADDR_HI_src_addr_63_32_shift  0
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_SRC_ADDR_HI_SRC_ADDR_63_32(x) (((x) & SDMA_PKT_PTEPDE_COPY_BACKWARDS_SRC_ADDR_HI_src_addr_63_32_mask) << SDMA_PKT_PTEPDE_COPY_BACKWARDS_SRC_ADDR_HI_src_addr_63_32_shift)
+
+/*define for DST_ADDR_LO word*/
+/*define for dst_addr_31_0 field*/
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_DST_ADDR_LO_dst_addr_31_0_offset 3
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_DST_ADDR_LO_dst_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_DST_ADDR_LO_dst_addr_31_0_shift  0
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_DST_ADDR_LO_DST_ADDR_31_0(x) (((x) & SDMA_PKT_PTEPDE_COPY_BACKWARDS_DST_ADDR_LO_dst_addr_31_0_mask) << SDMA_PKT_PTEPDE_COPY_BACKWARDS_DST_ADDR_LO_dst_addr_31_0_shift)
+
+/*define for DST_ADDR_HI word*/
+/*define for dst_addr_63_32 field*/
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_DST_ADDR_HI_dst_addr_63_32_offset 4
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_DST_ADDR_HI_dst_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_DST_ADDR_HI_dst_addr_63_32_shift  0
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_DST_ADDR_HI_DST_ADDR_63_32(x) (((x) & SDMA_PKT_PTEPDE_COPY_BACKWARDS_DST_ADDR_HI_dst_addr_63_32_mask) << SDMA_PKT_PTEPDE_COPY_BACKWARDS_DST_ADDR_HI_dst_addr_63_32_shift)
+
+/*define for MASK_BIT_FOR_DW word*/
+/*define for mask_first_xfer field*/
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_MASK_BIT_FOR_DW_mask_first_xfer_offset 5
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_MASK_BIT_FOR_DW_mask_first_xfer_mask   0x000000FF
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_MASK_BIT_FOR_DW_mask_first_xfer_shift  0
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_MASK_BIT_FOR_DW_MASK_FIRST_XFER(x) (((x) & SDMA_PKT_PTEPDE_COPY_BACKWARDS_MASK_BIT_FOR_DW_mask_first_xfer_mask) << SDMA_PKT_PTEPDE_COPY_BACKWARDS_MASK_BIT_FOR_DW_mask_first_xfer_shift)
+
+/*define for mask_last_xfer field*/
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_MASK_BIT_FOR_DW_mask_last_xfer_offset 5
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_MASK_BIT_FOR_DW_mask_last_xfer_mask   0x000000FF
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_MASK_BIT_FOR_DW_mask_last_xfer_shift  8
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_MASK_BIT_FOR_DW_MASK_LAST_XFER(x) (((x) & SDMA_PKT_PTEPDE_COPY_BACKWARDS_MASK_BIT_FOR_DW_mask_last_xfer_mask) << SDMA_PKT_PTEPDE_COPY_BACKWARDS_MASK_BIT_FOR_DW_mask_last_xfer_shift)
+
+/*define for COUNT_IN_32B_XFER word*/
+/*define for count field*/
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_COUNT_IN_32B_XFER_count_offset 6
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_COUNT_IN_32B_XFER_count_mask   0x0001FFFF
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_COUNT_IN_32B_XFER_count_shift  0
+#define SDMA_PKT_PTEPDE_COPY_BACKWARDS_COUNT_IN_32B_XFER_COUNT(x) (((x) & SDMA_PKT_PTEPDE_COPY_BACKWARDS_COUNT_IN_32B_XFER_count_mask) << SDMA_PKT_PTEPDE_COPY_BACKWARDS_COUNT_IN_32B_XFER_count_shift)
+
+
+/*
+** Definitions for SDMA_PKT_PTEPDE_RMW packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_PTEPDE_RMW_HEADER_op_offset 0
+#define SDMA_PKT_PTEPDE_RMW_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_PTEPDE_RMW_HEADER_op_shift  0
+#define SDMA_PKT_PTEPDE_RMW_HEADER_OP(x) (((x) & SDMA_PKT_PTEPDE_RMW_HEADER_op_mask) << SDMA_PKT_PTEPDE_RMW_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_PTEPDE_RMW_HEADER_sub_op_offset 0
+#define SDMA_PKT_PTEPDE_RMW_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_PTEPDE_RMW_HEADER_sub_op_shift  8
+#define SDMA_PKT_PTEPDE_RMW_HEADER_SUB_OP(x) (((x) & SDMA_PKT_PTEPDE_RMW_HEADER_sub_op_mask) << SDMA_PKT_PTEPDE_RMW_HEADER_sub_op_shift)
+
+/*define for mtype field*/
+#define SDMA_PKT_PTEPDE_RMW_HEADER_mtype_offset 0
+#define SDMA_PKT_PTEPDE_RMW_HEADER_mtype_mask   0x00000007
+#define SDMA_PKT_PTEPDE_RMW_HEADER_mtype_shift  16
+#define SDMA_PKT_PTEPDE_RMW_HEADER_MTYPE(x) (((x) & SDMA_PKT_PTEPDE_RMW_HEADER_mtype_mask) << SDMA_PKT_PTEPDE_RMW_HEADER_mtype_shift)
+
+/*define for gcc field*/
+#define SDMA_PKT_PTEPDE_RMW_HEADER_gcc_offset 0
+#define SDMA_PKT_PTEPDE_RMW_HEADER_gcc_mask   0x00000001
+#define SDMA_PKT_PTEPDE_RMW_HEADER_gcc_shift  19
+#define SDMA_PKT_PTEPDE_RMW_HEADER_GCC(x) (((x) & SDMA_PKT_PTEPDE_RMW_HEADER_gcc_mask) << SDMA_PKT_PTEPDE_RMW_HEADER_gcc_shift)
+
+/*define for sys field*/
+#define SDMA_PKT_PTEPDE_RMW_HEADER_sys_offset 0
+#define SDMA_PKT_PTEPDE_RMW_HEADER_sys_mask   0x00000001
+#define SDMA_PKT_PTEPDE_RMW_HEADER_sys_shift  20
+#define SDMA_PKT_PTEPDE_RMW_HEADER_SYS(x) (((x) & SDMA_PKT_PTEPDE_RMW_HEADER_sys_mask) << SDMA_PKT_PTEPDE_RMW_HEADER_sys_shift)
+
+/*define for snp field*/
+#define SDMA_PKT_PTEPDE_RMW_HEADER_snp_offset 0
+#define SDMA_PKT_PTEPDE_RMW_HEADER_snp_mask   0x00000001
+#define SDMA_PKT_PTEPDE_RMW_HEADER_snp_shift  22
+#define SDMA_PKT_PTEPDE_RMW_HEADER_SNP(x) (((x) & SDMA_PKT_PTEPDE_RMW_HEADER_snp_mask) << SDMA_PKT_PTEPDE_RMW_HEADER_snp_shift)
+
+/*define for gpa field*/
+#define SDMA_PKT_PTEPDE_RMW_HEADER_gpa_offset 0
+#define SDMA_PKT_PTEPDE_RMW_HEADER_gpa_mask   0x00000001
+#define SDMA_PKT_PTEPDE_RMW_HEADER_gpa_shift  23
+#define SDMA_PKT_PTEPDE_RMW_HEADER_GPA(x) (((x) & SDMA_PKT_PTEPDE_RMW_HEADER_gpa_mask) << SDMA_PKT_PTEPDE_RMW_HEADER_gpa_shift)
+
+/*define for l2_policy field*/
+#define SDMA_PKT_PTEPDE_RMW_HEADER_l2_policy_offset 0
+#define SDMA_PKT_PTEPDE_RMW_HEADER_l2_policy_mask   0x00000003
+#define SDMA_PKT_PTEPDE_RMW_HEADER_l2_policy_shift  24
+#define SDMA_PKT_PTEPDE_RMW_HEADER_L2_POLICY(x) (((x) & SDMA_PKT_PTEPDE_RMW_HEADER_l2_policy_mask) << SDMA_PKT_PTEPDE_RMW_HEADER_l2_policy_shift)
+
+/*define for llc_policy field*/
+#define SDMA_PKT_PTEPDE_RMW_HEADER_llc_policy_offset 0
+#define SDMA_PKT_PTEPDE_RMW_HEADER_llc_policy_mask   0x00000001
+#define SDMA_PKT_PTEPDE_RMW_HEADER_llc_policy_shift  26
+#define SDMA_PKT_PTEPDE_RMW_HEADER_LLC_POLICY(x) (((x) & SDMA_PKT_PTEPDE_RMW_HEADER_llc_policy_mask) << SDMA_PKT_PTEPDE_RMW_HEADER_llc_policy_shift)
+
+/*define for cpv field*/
+#define SDMA_PKT_PTEPDE_RMW_HEADER_cpv_offset 0
+#define SDMA_PKT_PTEPDE_RMW_HEADER_cpv_mask   0x00000001
+#define SDMA_PKT_PTEPDE_RMW_HEADER_cpv_shift  28
+#define SDMA_PKT_PTEPDE_RMW_HEADER_CPV(x) (((x) & SDMA_PKT_PTEPDE_RMW_HEADER_cpv_mask) << SDMA_PKT_PTEPDE_RMW_HEADER_cpv_shift)
+
+/*define for ADDR_LO word*/
+/*define for addr_31_0 field*/
+#define SDMA_PKT_PTEPDE_RMW_ADDR_LO_addr_31_0_offset 1
+#define SDMA_PKT_PTEPDE_RMW_ADDR_LO_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_PTEPDE_RMW_ADDR_LO_addr_31_0_shift  0
+#define SDMA_PKT_PTEPDE_RMW_ADDR_LO_ADDR_31_0(x) (((x) & SDMA_PKT_PTEPDE_RMW_ADDR_LO_addr_31_0_mask) << SDMA_PKT_PTEPDE_RMW_ADDR_LO_addr_31_0_shift)
+
+/*define for ADDR_HI word*/
+/*define for addr_63_32 field*/
+#define SDMA_PKT_PTEPDE_RMW_ADDR_HI_addr_63_32_offset 2
+#define SDMA_PKT_PTEPDE_RMW_ADDR_HI_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_PTEPDE_RMW_ADDR_HI_addr_63_32_shift  0
+#define SDMA_PKT_PTEPDE_RMW_ADDR_HI_ADDR_63_32(x) (((x) & SDMA_PKT_PTEPDE_RMW_ADDR_HI_addr_63_32_mask) << SDMA_PKT_PTEPDE_RMW_ADDR_HI_addr_63_32_shift)
+
+/*define for MASK_LO word*/
+/*define for mask_31_0 field*/
+#define SDMA_PKT_PTEPDE_RMW_MASK_LO_mask_31_0_offset 3
+#define SDMA_PKT_PTEPDE_RMW_MASK_LO_mask_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_PTEPDE_RMW_MASK_LO_mask_31_0_shift  0
+#define SDMA_PKT_PTEPDE_RMW_MASK_LO_MASK_31_0(x) (((x) & SDMA_PKT_PTEPDE_RMW_MASK_LO_mask_31_0_mask) << SDMA_PKT_PTEPDE_RMW_MASK_LO_mask_31_0_shift)
+
+/*define for MASK_HI word*/
+/*define for mask_63_32 field*/
+#define SDMA_PKT_PTEPDE_RMW_MASK_HI_mask_63_32_offset 4
+#define SDMA_PKT_PTEPDE_RMW_MASK_HI_mask_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_PTEPDE_RMW_MASK_HI_mask_63_32_shift  0
+#define SDMA_PKT_PTEPDE_RMW_MASK_HI_MASK_63_32(x) (((x) & SDMA_PKT_PTEPDE_RMW_MASK_HI_mask_63_32_mask) << SDMA_PKT_PTEPDE_RMW_MASK_HI_mask_63_32_shift)
+
+/*define for VALUE_LO word*/
+/*define for value_31_0 field*/
+#define SDMA_PKT_PTEPDE_RMW_VALUE_LO_value_31_0_offset 5
+#define SDMA_PKT_PTEPDE_RMW_VALUE_LO_value_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_PTEPDE_RMW_VALUE_LO_value_31_0_shift  0
+#define SDMA_PKT_PTEPDE_RMW_VALUE_LO_VALUE_31_0(x) (((x) & SDMA_PKT_PTEPDE_RMW_VALUE_LO_value_31_0_mask) << SDMA_PKT_PTEPDE_RMW_VALUE_LO_value_31_0_shift)
+
+/*define for VALUE_HI word*/
+/*define for value_63_32 field*/
+#define SDMA_PKT_PTEPDE_RMW_VALUE_HI_value_63_32_offset 6
+#define SDMA_PKT_PTEPDE_RMW_VALUE_HI_value_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_PTEPDE_RMW_VALUE_HI_value_63_32_shift  0
+#define SDMA_PKT_PTEPDE_RMW_VALUE_HI_VALUE_63_32(x) (((x) & SDMA_PKT_PTEPDE_RMW_VALUE_HI_value_63_32_mask) << SDMA_PKT_PTEPDE_RMW_VALUE_HI_value_63_32_shift)
+
+/*define for COUNT word*/
+/*define for num_of_pte field*/
+#define SDMA_PKT_PTEPDE_RMW_COUNT_num_of_pte_offset 7
+#define SDMA_PKT_PTEPDE_RMW_COUNT_num_of_pte_mask   0xFFFFFFFF
+#define SDMA_PKT_PTEPDE_RMW_COUNT_num_of_pte_shift  0
+#define SDMA_PKT_PTEPDE_RMW_COUNT_NUM_OF_PTE(x) (((x) & SDMA_PKT_PTEPDE_RMW_COUNT_num_of_pte_mask) << SDMA_PKT_PTEPDE_RMW_COUNT_num_of_pte_shift)
+
+
+/*
+** Definitions for SDMA_PKT_REGISTER_RMW packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_REGISTER_RMW_HEADER_op_offset 0
+#define SDMA_PKT_REGISTER_RMW_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_REGISTER_RMW_HEADER_op_shift  0
+#define SDMA_PKT_REGISTER_RMW_HEADER_OP(x) (((x) & SDMA_PKT_REGISTER_RMW_HEADER_op_mask) << SDMA_PKT_REGISTER_RMW_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_REGISTER_RMW_HEADER_sub_op_offset 0
+#define SDMA_PKT_REGISTER_RMW_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_REGISTER_RMW_HEADER_sub_op_shift  8
+#define SDMA_PKT_REGISTER_RMW_HEADER_SUB_OP(x) (((x) & SDMA_PKT_REGISTER_RMW_HEADER_sub_op_mask) << SDMA_PKT_REGISTER_RMW_HEADER_sub_op_shift)
+
+/*define for ADDR word*/
+/*define for addr field*/
+#define SDMA_PKT_REGISTER_RMW_ADDR_addr_offset 1
+#define SDMA_PKT_REGISTER_RMW_ADDR_addr_mask   0x000FFFFF
+#define SDMA_PKT_REGISTER_RMW_ADDR_addr_shift  0
+#define SDMA_PKT_REGISTER_RMW_ADDR_ADDR(x) (((x) & SDMA_PKT_REGISTER_RMW_ADDR_addr_mask) << SDMA_PKT_REGISTER_RMW_ADDR_addr_shift)
+
+/*define for aperture_id field*/
+#define SDMA_PKT_REGISTER_RMW_ADDR_aperture_id_offset 1
+#define SDMA_PKT_REGISTER_RMW_ADDR_aperture_id_mask   0x00000FFF
+#define SDMA_PKT_REGISTER_RMW_ADDR_aperture_id_shift  20
+#define SDMA_PKT_REGISTER_RMW_ADDR_APERTURE_ID(x) (((x) & SDMA_PKT_REGISTER_RMW_ADDR_aperture_id_mask) << SDMA_PKT_REGISTER_RMW_ADDR_aperture_id_shift)
+
+/*define for MASK word*/
+/*define for mask field*/
+#define SDMA_PKT_REGISTER_RMW_MASK_mask_offset 2
+#define SDMA_PKT_REGISTER_RMW_MASK_mask_mask   0xFFFFFFFF
+#define SDMA_PKT_REGISTER_RMW_MASK_mask_shift  0
+#define SDMA_PKT_REGISTER_RMW_MASK_MASK(x) (((x) & SDMA_PKT_REGISTER_RMW_MASK_mask_mask) << SDMA_PKT_REGISTER_RMW_MASK_mask_shift)
+
+/*define for VALUE word*/
+/*define for value field*/
+#define SDMA_PKT_REGISTER_RMW_VALUE_value_offset 3
+#define SDMA_PKT_REGISTER_RMW_VALUE_value_mask   0xFFFFFFFF
+#define SDMA_PKT_REGISTER_RMW_VALUE_value_shift  0
+#define SDMA_PKT_REGISTER_RMW_VALUE_VALUE(x) (((x) & SDMA_PKT_REGISTER_RMW_VALUE_value_mask) << SDMA_PKT_REGISTER_RMW_VALUE_value_shift)
+
+/*define for MISC word*/
+/*define for stride field*/
+#define SDMA_PKT_REGISTER_RMW_MISC_stride_offset 4
+#define SDMA_PKT_REGISTER_RMW_MISC_stride_mask   0x000FFFFF
+#define SDMA_PKT_REGISTER_RMW_MISC_stride_shift  0
+#define SDMA_PKT_REGISTER_RMW_MISC_STRIDE(x) (((x) & SDMA_PKT_REGISTER_RMW_MISC_stride_mask) << SDMA_PKT_REGISTER_RMW_MISC_stride_shift)
+
+/*define for num_of_reg field*/
+#define SDMA_PKT_REGISTER_RMW_MISC_num_of_reg_offset 4
+#define SDMA_PKT_REGISTER_RMW_MISC_num_of_reg_mask   0x00000FFF
+#define SDMA_PKT_REGISTER_RMW_MISC_num_of_reg_shift  20
+#define SDMA_PKT_REGISTER_RMW_MISC_NUM_OF_REG(x) (((x) & SDMA_PKT_REGISTER_RMW_MISC_num_of_reg_mask) << SDMA_PKT_REGISTER_RMW_MISC_num_of_reg_shift)
+
+
+/*
+** Definitions for SDMA_PKT_WRITE_INCR packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_WRITE_INCR_HEADER_op_offset 0
+#define SDMA_PKT_WRITE_INCR_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_WRITE_INCR_HEADER_op_shift  0
+#define SDMA_PKT_WRITE_INCR_HEADER_OP(x) (((x) & SDMA_PKT_WRITE_INCR_HEADER_op_mask) << SDMA_PKT_WRITE_INCR_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_WRITE_INCR_HEADER_sub_op_offset 0
+#define SDMA_PKT_WRITE_INCR_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_WRITE_INCR_HEADER_sub_op_shift  8
+#define SDMA_PKT_WRITE_INCR_HEADER_SUB_OP(x) (((x) & SDMA_PKT_WRITE_INCR_HEADER_sub_op_mask) << SDMA_PKT_WRITE_INCR_HEADER_sub_op_shift)
+
+/*define for cache_policy field*/
+#define SDMA_PKT_WRITE_INCR_HEADER_cache_policy_offset 0
+#define SDMA_PKT_WRITE_INCR_HEADER_cache_policy_mask   0x00000007
+#define SDMA_PKT_WRITE_INCR_HEADER_cache_policy_shift  24
+#define SDMA_PKT_WRITE_INCR_HEADER_CACHE_POLICY(x) (((x) & SDMA_PKT_WRITE_INCR_HEADER_cache_policy_mask) << SDMA_PKT_WRITE_INCR_HEADER_cache_policy_shift)
+
+/*define for cpv field*/
+#define SDMA_PKT_WRITE_INCR_HEADER_cpv_offset 0
+#define SDMA_PKT_WRITE_INCR_HEADER_cpv_mask   0x00000001
+#define SDMA_PKT_WRITE_INCR_HEADER_cpv_shift  28
+#define SDMA_PKT_WRITE_INCR_HEADER_CPV(x) (((x) & SDMA_PKT_WRITE_INCR_HEADER_cpv_mask) << SDMA_PKT_WRITE_INCR_HEADER_cpv_shift)
+
+/*define for DST_ADDR_LO word*/
+/*define for dst_addr_31_0 field*/
+#define SDMA_PKT_WRITE_INCR_DST_ADDR_LO_dst_addr_31_0_offset 1
+#define SDMA_PKT_WRITE_INCR_DST_ADDR_LO_dst_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_WRITE_INCR_DST_ADDR_LO_dst_addr_31_0_shift  0
+#define SDMA_PKT_WRITE_INCR_DST_ADDR_LO_DST_ADDR_31_0(x) (((x) & SDMA_PKT_WRITE_INCR_DST_ADDR_LO_dst_addr_31_0_mask) << SDMA_PKT_WRITE_INCR_DST_ADDR_LO_dst_addr_31_0_shift)
+
+/*define for DST_ADDR_HI word*/
+/*define for dst_addr_63_32 field*/
+#define SDMA_PKT_WRITE_INCR_DST_ADDR_HI_dst_addr_63_32_offset 2
+#define SDMA_PKT_WRITE_INCR_DST_ADDR_HI_dst_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_WRITE_INCR_DST_ADDR_HI_dst_addr_63_32_shift  0
+#define SDMA_PKT_WRITE_INCR_DST_ADDR_HI_DST_ADDR_63_32(x) (((x) & SDMA_PKT_WRITE_INCR_DST_ADDR_HI_dst_addr_63_32_mask) << SDMA_PKT_WRITE_INCR_DST_ADDR_HI_dst_addr_63_32_shift)
+
+/*define for MASK_DW0 word*/
+/*define for mask_dw0 field*/
+#define SDMA_PKT_WRITE_INCR_MASK_DW0_mask_dw0_offset 3
+#define SDMA_PKT_WRITE_INCR_MASK_DW0_mask_dw0_mask   0xFFFFFFFF
+#define SDMA_PKT_WRITE_INCR_MASK_DW0_mask_dw0_shift  0
+#define SDMA_PKT_WRITE_INCR_MASK_DW0_MASK_DW0(x) (((x) & SDMA_PKT_WRITE_INCR_MASK_DW0_mask_dw0_mask) << SDMA_PKT_WRITE_INCR_MASK_DW0_mask_dw0_shift)
+
+/*define for MASK_DW1 word*/
+/*define for mask_dw1 field*/
+#define SDMA_PKT_WRITE_INCR_MASK_DW1_mask_dw1_offset 4
+#define SDMA_PKT_WRITE_INCR_MASK_DW1_mask_dw1_mask   0xFFFFFFFF
+#define SDMA_PKT_WRITE_INCR_MASK_DW1_mask_dw1_shift  0
+#define SDMA_PKT_WRITE_INCR_MASK_DW1_MASK_DW1(x) (((x) & SDMA_PKT_WRITE_INCR_MASK_DW1_mask_dw1_mask) << SDMA_PKT_WRITE_INCR_MASK_DW1_mask_dw1_shift)
+
+/*define for INIT_DW0 word*/
+/*define for init_dw0 field*/
+#define SDMA_PKT_WRITE_INCR_INIT_DW0_init_dw0_offset 5
+#define SDMA_PKT_WRITE_INCR_INIT_DW0_init_dw0_mask   0xFFFFFFFF
+#define SDMA_PKT_WRITE_INCR_INIT_DW0_init_dw0_shift  0
+#define SDMA_PKT_WRITE_INCR_INIT_DW0_INIT_DW0(x) (((x) & SDMA_PKT_WRITE_INCR_INIT_DW0_init_dw0_mask) << SDMA_PKT_WRITE_INCR_INIT_DW0_init_dw0_shift)
+
+/*define for INIT_DW1 word*/
+/*define for init_dw1 field*/
+#define SDMA_PKT_WRITE_INCR_INIT_DW1_init_dw1_offset 6
+#define SDMA_PKT_WRITE_INCR_INIT_DW1_init_dw1_mask   0xFFFFFFFF
+#define SDMA_PKT_WRITE_INCR_INIT_DW1_init_dw1_shift  0
+#define SDMA_PKT_WRITE_INCR_INIT_DW1_INIT_DW1(x) (((x) & SDMA_PKT_WRITE_INCR_INIT_DW1_init_dw1_mask) << SDMA_PKT_WRITE_INCR_INIT_DW1_init_dw1_shift)
+
+/*define for INCR_DW0 word*/
+/*define for incr_dw0 field*/
+#define SDMA_PKT_WRITE_INCR_INCR_DW0_incr_dw0_offset 7
+#define SDMA_PKT_WRITE_INCR_INCR_DW0_incr_dw0_mask   0xFFFFFFFF
+#define SDMA_PKT_WRITE_INCR_INCR_DW0_incr_dw0_shift  0
+#define SDMA_PKT_WRITE_INCR_INCR_DW0_INCR_DW0(x) (((x) & SDMA_PKT_WRITE_INCR_INCR_DW0_incr_dw0_mask) << SDMA_PKT_WRITE_INCR_INCR_DW0_incr_dw0_shift)
+
+/*define for INCR_DW1 word*/
+/*define for incr_dw1 field*/
+#define SDMA_PKT_WRITE_INCR_INCR_DW1_incr_dw1_offset 8
+#define SDMA_PKT_WRITE_INCR_INCR_DW1_incr_dw1_mask   0xFFFFFFFF
+#define SDMA_PKT_WRITE_INCR_INCR_DW1_incr_dw1_shift  0
+#define SDMA_PKT_WRITE_INCR_INCR_DW1_INCR_DW1(x) (((x) & SDMA_PKT_WRITE_INCR_INCR_DW1_incr_dw1_mask) << SDMA_PKT_WRITE_INCR_INCR_DW1_incr_dw1_shift)
+
+/*define for COUNT word*/
+/*define for count field*/
+#define SDMA_PKT_WRITE_INCR_COUNT_count_offset 9
+#define SDMA_PKT_WRITE_INCR_COUNT_count_mask   0x0007FFFF
+#define SDMA_PKT_WRITE_INCR_COUNT_count_shift  0
+#define SDMA_PKT_WRITE_INCR_COUNT_COUNT(x) (((x) & SDMA_PKT_WRITE_INCR_COUNT_count_mask) << SDMA_PKT_WRITE_INCR_COUNT_count_shift)
+
+
+/*
+** Definitions for SDMA_PKT_INDIRECT packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_INDIRECT_HEADER_op_offset 0
+#define SDMA_PKT_INDIRECT_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_INDIRECT_HEADER_op_shift  0
+#define SDMA_PKT_INDIRECT_HEADER_OP(x) (((x) & SDMA_PKT_INDIRECT_HEADER_op_mask) << SDMA_PKT_INDIRECT_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_INDIRECT_HEADER_sub_op_offset 0
+#define SDMA_PKT_INDIRECT_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_INDIRECT_HEADER_sub_op_shift  8
+#define SDMA_PKT_INDIRECT_HEADER_SUB_OP(x) (((x) & SDMA_PKT_INDIRECT_HEADER_sub_op_mask) << SDMA_PKT_INDIRECT_HEADER_sub_op_shift)
+
+/*define for vmid field*/
+#define SDMA_PKT_INDIRECT_HEADER_vmid_offset 0
+#define SDMA_PKT_INDIRECT_HEADER_vmid_mask   0x0000000F
+#define SDMA_PKT_INDIRECT_HEADER_vmid_shift  16
+#define SDMA_PKT_INDIRECT_HEADER_VMID(x) (((x) & SDMA_PKT_INDIRECT_HEADER_vmid_mask) << SDMA_PKT_INDIRECT_HEADER_vmid_shift)
+
+/*define for priv field*/
+#define SDMA_PKT_INDIRECT_HEADER_priv_offset 0
+#define SDMA_PKT_INDIRECT_HEADER_priv_mask   0x00000001
+#define SDMA_PKT_INDIRECT_HEADER_priv_shift  31
+#define SDMA_PKT_INDIRECT_HEADER_PRIV(x) (((x) & SDMA_PKT_INDIRECT_HEADER_priv_mask) << SDMA_PKT_INDIRECT_HEADER_priv_shift)
+
+/*define for BASE_LO word*/
+/*define for ib_base_31_0 field*/
+#define SDMA_PKT_INDIRECT_BASE_LO_ib_base_31_0_offset 1
+#define SDMA_PKT_INDIRECT_BASE_LO_ib_base_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_INDIRECT_BASE_LO_ib_base_31_0_shift  0
+#define SDMA_PKT_INDIRECT_BASE_LO_IB_BASE_31_0(x) (((x) & SDMA_PKT_INDIRECT_BASE_LO_ib_base_31_0_mask) << SDMA_PKT_INDIRECT_BASE_LO_ib_base_31_0_shift)
+
+/*define for BASE_HI word*/
+/*define for ib_base_63_32 field*/
+#define SDMA_PKT_INDIRECT_BASE_HI_ib_base_63_32_offset 2
+#define SDMA_PKT_INDIRECT_BASE_HI_ib_base_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_INDIRECT_BASE_HI_ib_base_63_32_shift  0
+#define SDMA_PKT_INDIRECT_BASE_HI_IB_BASE_63_32(x) (((x) & SDMA_PKT_INDIRECT_BASE_HI_ib_base_63_32_mask) << SDMA_PKT_INDIRECT_BASE_HI_ib_base_63_32_shift)
+
+/*define for IB_SIZE word*/
+/*define for ib_size field*/
+#define SDMA_PKT_INDIRECT_IB_SIZE_ib_size_offset 3
+#define SDMA_PKT_INDIRECT_IB_SIZE_ib_size_mask   0x000FFFFF
+#define SDMA_PKT_INDIRECT_IB_SIZE_ib_size_shift  0
+#define SDMA_PKT_INDIRECT_IB_SIZE_IB_SIZE(x) (((x) & SDMA_PKT_INDIRECT_IB_SIZE_ib_size_mask) << SDMA_PKT_INDIRECT_IB_SIZE_ib_size_shift)
+
+/*define for CSA_ADDR_LO word*/
+/*define for csa_addr_31_0 field*/
+#define SDMA_PKT_INDIRECT_CSA_ADDR_LO_csa_addr_31_0_offset 4
+#define SDMA_PKT_INDIRECT_CSA_ADDR_LO_csa_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_INDIRECT_CSA_ADDR_LO_csa_addr_31_0_shift  0
+#define SDMA_PKT_INDIRECT_CSA_ADDR_LO_CSA_ADDR_31_0(x) (((x) & SDMA_PKT_INDIRECT_CSA_ADDR_LO_csa_addr_31_0_mask) << SDMA_PKT_INDIRECT_CSA_ADDR_LO_csa_addr_31_0_shift)
+
+/*define for CSA_ADDR_HI word*/
+/*define for csa_addr_63_32 field*/
+#define SDMA_PKT_INDIRECT_CSA_ADDR_HI_csa_addr_63_32_offset 5
+#define SDMA_PKT_INDIRECT_CSA_ADDR_HI_csa_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_INDIRECT_CSA_ADDR_HI_csa_addr_63_32_shift  0
+#define SDMA_PKT_INDIRECT_CSA_ADDR_HI_CSA_ADDR_63_32(x) (((x) & SDMA_PKT_INDIRECT_CSA_ADDR_HI_csa_addr_63_32_mask) << SDMA_PKT_INDIRECT_CSA_ADDR_HI_csa_addr_63_32_shift)
+
+
+/*
+** Definitions for SDMA_PKT_SEMAPHORE packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_SEMAPHORE_HEADER_op_offset 0
+#define SDMA_PKT_SEMAPHORE_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_SEMAPHORE_HEADER_op_shift  0
+#define SDMA_PKT_SEMAPHORE_HEADER_OP(x) (((x) & SDMA_PKT_SEMAPHORE_HEADER_op_mask) << SDMA_PKT_SEMAPHORE_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_SEMAPHORE_HEADER_sub_op_offset 0
+#define SDMA_PKT_SEMAPHORE_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_SEMAPHORE_HEADER_sub_op_shift  8
+#define SDMA_PKT_SEMAPHORE_HEADER_SUB_OP(x) (((x) & SDMA_PKT_SEMAPHORE_HEADER_sub_op_mask) << SDMA_PKT_SEMAPHORE_HEADER_sub_op_shift)
+
+/*define for write_one field*/
+#define SDMA_PKT_SEMAPHORE_HEADER_write_one_offset 0
+#define SDMA_PKT_SEMAPHORE_HEADER_write_one_mask   0x00000001
+#define SDMA_PKT_SEMAPHORE_HEADER_write_one_shift  29
+#define SDMA_PKT_SEMAPHORE_HEADER_WRITE_ONE(x) (((x) & SDMA_PKT_SEMAPHORE_HEADER_write_one_mask) << SDMA_PKT_SEMAPHORE_HEADER_write_one_shift)
+
+/*define for signal field*/
+#define SDMA_PKT_SEMAPHORE_HEADER_signal_offset 0
+#define SDMA_PKT_SEMAPHORE_HEADER_signal_mask   0x00000001
+#define SDMA_PKT_SEMAPHORE_HEADER_signal_shift  30
+#define SDMA_PKT_SEMAPHORE_HEADER_SIGNAL(x) (((x) & SDMA_PKT_SEMAPHORE_HEADER_signal_mask) << SDMA_PKT_SEMAPHORE_HEADER_signal_shift)
+
+/*define for mailbox field*/
+#define SDMA_PKT_SEMAPHORE_HEADER_mailbox_offset 0
+#define SDMA_PKT_SEMAPHORE_HEADER_mailbox_mask   0x00000001
+#define SDMA_PKT_SEMAPHORE_HEADER_mailbox_shift  31
+#define SDMA_PKT_SEMAPHORE_HEADER_MAILBOX(x) (((x) & SDMA_PKT_SEMAPHORE_HEADER_mailbox_mask) << SDMA_PKT_SEMAPHORE_HEADER_mailbox_shift)
+
+/*define for ADDR_LO word*/
+/*define for addr_31_0 field*/
+#define SDMA_PKT_SEMAPHORE_ADDR_LO_addr_31_0_offset 1
+#define SDMA_PKT_SEMAPHORE_ADDR_LO_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_SEMAPHORE_ADDR_LO_addr_31_0_shift  0
+#define SDMA_PKT_SEMAPHORE_ADDR_LO_ADDR_31_0(x) (((x) & SDMA_PKT_SEMAPHORE_ADDR_LO_addr_31_0_mask) << SDMA_PKT_SEMAPHORE_ADDR_LO_addr_31_0_shift)
+
+/*define for ADDR_HI word*/
+/*define for addr_63_32 field*/
+#define SDMA_PKT_SEMAPHORE_ADDR_HI_addr_63_32_offset 2
+#define SDMA_PKT_SEMAPHORE_ADDR_HI_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_SEMAPHORE_ADDR_HI_addr_63_32_shift  0
+#define SDMA_PKT_SEMAPHORE_ADDR_HI_ADDR_63_32(x) (((x) & SDMA_PKT_SEMAPHORE_ADDR_HI_addr_63_32_mask) << SDMA_PKT_SEMAPHORE_ADDR_HI_addr_63_32_shift)
+
+
+/*
+** Definitions for SDMA_PKT_MEM_INCR packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_MEM_INCR_HEADER_op_offset 0
+#define SDMA_PKT_MEM_INCR_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_MEM_INCR_HEADER_op_shift  0
+#define SDMA_PKT_MEM_INCR_HEADER_OP(x) (((x) & SDMA_PKT_MEM_INCR_HEADER_op_mask) << SDMA_PKT_MEM_INCR_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_MEM_INCR_HEADER_sub_op_offset 0
+#define SDMA_PKT_MEM_INCR_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_MEM_INCR_HEADER_sub_op_shift  8
+#define SDMA_PKT_MEM_INCR_HEADER_SUB_OP(x) (((x) & SDMA_PKT_MEM_INCR_HEADER_sub_op_mask) << SDMA_PKT_MEM_INCR_HEADER_sub_op_shift)
+
+/*define for l2_policy field*/
+#define SDMA_PKT_MEM_INCR_HEADER_l2_policy_offset 0
+#define SDMA_PKT_MEM_INCR_HEADER_l2_policy_mask   0x00000003
+#define SDMA_PKT_MEM_INCR_HEADER_l2_policy_shift  24
+#define SDMA_PKT_MEM_INCR_HEADER_L2_POLICY(x) (((x) & SDMA_PKT_MEM_INCR_HEADER_l2_policy_mask) << SDMA_PKT_MEM_INCR_HEADER_l2_policy_shift)
+
+/*define for llc_policy field*/
+#define SDMA_PKT_MEM_INCR_HEADER_llc_policy_offset 0
+#define SDMA_PKT_MEM_INCR_HEADER_llc_policy_mask   0x00000001
+#define SDMA_PKT_MEM_INCR_HEADER_llc_policy_shift  26
+#define SDMA_PKT_MEM_INCR_HEADER_LLC_POLICY(x) (((x) & SDMA_PKT_MEM_INCR_HEADER_llc_policy_mask) << SDMA_PKT_MEM_INCR_HEADER_llc_policy_shift)
+
+/*define for cpv field*/
+#define SDMA_PKT_MEM_INCR_HEADER_cpv_offset 0
+#define SDMA_PKT_MEM_INCR_HEADER_cpv_mask   0x00000001
+#define SDMA_PKT_MEM_INCR_HEADER_cpv_shift  28
+#define SDMA_PKT_MEM_INCR_HEADER_CPV(x) (((x) & SDMA_PKT_MEM_INCR_HEADER_cpv_mask) << SDMA_PKT_MEM_INCR_HEADER_cpv_shift)
+
+/*define for ADDR_LO word*/
+/*define for addr_31_0 field*/
+#define SDMA_PKT_MEM_INCR_ADDR_LO_addr_31_0_offset 1
+#define SDMA_PKT_MEM_INCR_ADDR_LO_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_MEM_INCR_ADDR_LO_addr_31_0_shift  0
+#define SDMA_PKT_MEM_INCR_ADDR_LO_ADDR_31_0(x) (((x) & SDMA_PKT_MEM_INCR_ADDR_LO_addr_31_0_mask) << SDMA_PKT_MEM_INCR_ADDR_LO_addr_31_0_shift)
+
+/*define for ADDR_HI word*/
+/*define for addr_63_32 field*/
+#define SDMA_PKT_MEM_INCR_ADDR_HI_addr_63_32_offset 2
+#define SDMA_PKT_MEM_INCR_ADDR_HI_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_MEM_INCR_ADDR_HI_addr_63_32_shift  0
+#define SDMA_PKT_MEM_INCR_ADDR_HI_ADDR_63_32(x) (((x) & SDMA_PKT_MEM_INCR_ADDR_HI_addr_63_32_mask) << SDMA_PKT_MEM_INCR_ADDR_HI_addr_63_32_shift)
+
+
+/*
+** Definitions for SDMA_PKT_VM_INVALIDATION packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_VM_INVALIDATION_HEADER_op_offset 0
+#define SDMA_PKT_VM_INVALIDATION_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_VM_INVALIDATION_HEADER_op_shift  0
+#define SDMA_PKT_VM_INVALIDATION_HEADER_OP(x) (((x) & SDMA_PKT_VM_INVALIDATION_HEADER_op_mask) << SDMA_PKT_VM_INVALIDATION_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_VM_INVALIDATION_HEADER_sub_op_offset 0
+#define SDMA_PKT_VM_INVALIDATION_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_VM_INVALIDATION_HEADER_sub_op_shift  8
+#define SDMA_PKT_VM_INVALIDATION_HEADER_SUB_OP(x) (((x) & SDMA_PKT_VM_INVALIDATION_HEADER_sub_op_mask) << SDMA_PKT_VM_INVALIDATION_HEADER_sub_op_shift)
+
+/*define for gfx_eng_id field*/
+#define SDMA_PKT_VM_INVALIDATION_HEADER_gfx_eng_id_offset 0
+#define SDMA_PKT_VM_INVALIDATION_HEADER_gfx_eng_id_mask   0x0000001F
+#define SDMA_PKT_VM_INVALIDATION_HEADER_gfx_eng_id_shift  16
+#define SDMA_PKT_VM_INVALIDATION_HEADER_GFX_ENG_ID(x) (((x) & SDMA_PKT_VM_INVALIDATION_HEADER_gfx_eng_id_mask) << SDMA_PKT_VM_INVALIDATION_HEADER_gfx_eng_id_shift)
+
+/*define for mm_eng_id field*/
+#define SDMA_PKT_VM_INVALIDATION_HEADER_mm_eng_id_offset 0
+#define SDMA_PKT_VM_INVALIDATION_HEADER_mm_eng_id_mask   0x0000001F
+#define SDMA_PKT_VM_INVALIDATION_HEADER_mm_eng_id_shift  24
+#define SDMA_PKT_VM_INVALIDATION_HEADER_MM_ENG_ID(x) (((x) & SDMA_PKT_VM_INVALIDATION_HEADER_mm_eng_id_mask) << SDMA_PKT_VM_INVALIDATION_HEADER_mm_eng_id_shift)
+
+/*define for INVALIDATEREQ word*/
+/*define for invalidatereq field*/
+#define SDMA_PKT_VM_INVALIDATION_INVALIDATEREQ_invalidatereq_offset 1
+#define SDMA_PKT_VM_INVALIDATION_INVALIDATEREQ_invalidatereq_mask   0xFFFFFFFF
+#define SDMA_PKT_VM_INVALIDATION_INVALIDATEREQ_invalidatereq_shift  0
+#define SDMA_PKT_VM_INVALIDATION_INVALIDATEREQ_INVALIDATEREQ(x) (((x) & SDMA_PKT_VM_INVALIDATION_INVALIDATEREQ_invalidatereq_mask) << SDMA_PKT_VM_INVALIDATION_INVALIDATEREQ_invalidatereq_shift)
+
+/*define for ADDRESSRANGELO word*/
+/*define for addressrangelo field*/
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGELO_addressrangelo_offset 2
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGELO_addressrangelo_mask   0xFFFFFFFF
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGELO_addressrangelo_shift  0
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGELO_ADDRESSRANGELO(x) (((x) & SDMA_PKT_VM_INVALIDATION_ADDRESSRANGELO_addressrangelo_mask) << SDMA_PKT_VM_INVALIDATION_ADDRESSRANGELO_addressrangelo_shift)
+
+/*define for ADDRESSRANGEHI word*/
+/*define for invalidateack field*/
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_invalidateack_offset 3
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_invalidateack_mask   0x0000FFFF
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_invalidateack_shift  0
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_INVALIDATEACK(x) (((x) & SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_invalidateack_mask) << SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_invalidateack_shift)
+
+/*define for addressrangehi field*/
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_addressrangehi_offset 3
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_addressrangehi_mask   0x0000001F
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_addressrangehi_shift  16
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_ADDRESSRANGEHI(x) (((x) & SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_addressrangehi_mask) << SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_addressrangehi_shift)
+
+/*define for reserved field*/
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_reserved_offset 3
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_reserved_mask   0x000001FF
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_reserved_shift  23
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_RESERVED(x) (((x) & SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_reserved_mask) << SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_reserved_shift)
+
+
+/*
+** Definitions for SDMA_PKT_FENCE packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_FENCE_HEADER_op_offset 0
+#define SDMA_PKT_FENCE_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_FENCE_HEADER_op_shift  0
+#define SDMA_PKT_FENCE_HEADER_OP(x) (((x) & SDMA_PKT_FENCE_HEADER_op_mask) << SDMA_PKT_FENCE_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_FENCE_HEADER_sub_op_offset 0
+#define SDMA_PKT_FENCE_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_FENCE_HEADER_sub_op_shift  8
+#define SDMA_PKT_FENCE_HEADER_SUB_OP(x) (((x) & SDMA_PKT_FENCE_HEADER_sub_op_mask) << SDMA_PKT_FENCE_HEADER_sub_op_shift)
+
+/*define for mtype field*/
+#define SDMA_PKT_FENCE_HEADER_mtype_offset 0
+#define SDMA_PKT_FENCE_HEADER_mtype_mask   0x00000007
+#define SDMA_PKT_FENCE_HEADER_mtype_shift  16
+#define SDMA_PKT_FENCE_HEADER_MTYPE(x) (((x) & SDMA_PKT_FENCE_HEADER_mtype_mask) << SDMA_PKT_FENCE_HEADER_mtype_shift)
+
+/*define for gcc field*/
+#define SDMA_PKT_FENCE_HEADER_gcc_offset 0
+#define SDMA_PKT_FENCE_HEADER_gcc_mask   0x00000001
+#define SDMA_PKT_FENCE_HEADER_gcc_shift  19
+#define SDMA_PKT_FENCE_HEADER_GCC(x) (((x) & SDMA_PKT_FENCE_HEADER_gcc_mask) << SDMA_PKT_FENCE_HEADER_gcc_shift)
+
+/*define for sys field*/
+#define SDMA_PKT_FENCE_HEADER_sys_offset 0
+#define SDMA_PKT_FENCE_HEADER_sys_mask   0x00000001
+#define SDMA_PKT_FENCE_HEADER_sys_shift  20
+#define SDMA_PKT_FENCE_HEADER_SYS(x) (((x) & SDMA_PKT_FENCE_HEADER_sys_mask) << SDMA_PKT_FENCE_HEADER_sys_shift)
+
+/*define for snp field*/
+#define SDMA_PKT_FENCE_HEADER_snp_offset 0
+#define SDMA_PKT_FENCE_HEADER_snp_mask   0x00000001
+#define SDMA_PKT_FENCE_HEADER_snp_shift  22
+#define SDMA_PKT_FENCE_HEADER_SNP(x) (((x) & SDMA_PKT_FENCE_HEADER_snp_mask) << SDMA_PKT_FENCE_HEADER_snp_shift)
+
+/*define for gpa field*/
+#define SDMA_PKT_FENCE_HEADER_gpa_offset 0
+#define SDMA_PKT_FENCE_HEADER_gpa_mask   0x00000001
+#define SDMA_PKT_FENCE_HEADER_gpa_shift  23
+#define SDMA_PKT_FENCE_HEADER_GPA(x) (((x) & SDMA_PKT_FENCE_HEADER_gpa_mask) << SDMA_PKT_FENCE_HEADER_gpa_shift)
+
+/*define for l2_policy field*/
+#define SDMA_PKT_FENCE_HEADER_l2_policy_offset 0
+#define SDMA_PKT_FENCE_HEADER_l2_policy_mask   0x00000003
+#define SDMA_PKT_FENCE_HEADER_l2_policy_shift  24
+#define SDMA_PKT_FENCE_HEADER_L2_POLICY(x) (((x) & SDMA_PKT_FENCE_HEADER_l2_policy_mask) << SDMA_PKT_FENCE_HEADER_l2_policy_shift)
+
+/*define for llc_policy field*/
+#define SDMA_PKT_FENCE_HEADER_llc_policy_offset 0
+#define SDMA_PKT_FENCE_HEADER_llc_policy_mask   0x00000001
+#define SDMA_PKT_FENCE_HEADER_llc_policy_shift  26
+#define SDMA_PKT_FENCE_HEADER_LLC_POLICY(x) (((x) & SDMA_PKT_FENCE_HEADER_llc_policy_mask) << SDMA_PKT_FENCE_HEADER_llc_policy_shift)
+
+/*define for cpv field*/
+#define SDMA_PKT_FENCE_HEADER_cpv_offset 0
+#define SDMA_PKT_FENCE_HEADER_cpv_mask   0x00000001
+#define SDMA_PKT_FENCE_HEADER_cpv_shift  28
+#define SDMA_PKT_FENCE_HEADER_CPV(x) (((x) & SDMA_PKT_FENCE_HEADER_cpv_mask) << SDMA_PKT_FENCE_HEADER_cpv_shift)
+
+/*define for ADDR_LO word*/
+/*define for addr_31_0 field*/
+#define SDMA_PKT_FENCE_ADDR_LO_addr_31_0_offset 1
+#define SDMA_PKT_FENCE_ADDR_LO_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_FENCE_ADDR_LO_addr_31_0_shift  0
+#define SDMA_PKT_FENCE_ADDR_LO_ADDR_31_0(x) (((x) & SDMA_PKT_FENCE_ADDR_LO_addr_31_0_mask) << SDMA_PKT_FENCE_ADDR_LO_addr_31_0_shift)
+
+/*define for ADDR_HI word*/
+/*define for addr_63_32 field*/
+#define SDMA_PKT_FENCE_ADDR_HI_addr_63_32_offset 2
+#define SDMA_PKT_FENCE_ADDR_HI_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_FENCE_ADDR_HI_addr_63_32_shift  0
+#define SDMA_PKT_FENCE_ADDR_HI_ADDR_63_32(x) (((x) & SDMA_PKT_FENCE_ADDR_HI_addr_63_32_mask) << SDMA_PKT_FENCE_ADDR_HI_addr_63_32_shift)
+
+/*define for DATA word*/
+/*define for data field*/
+#define SDMA_PKT_FENCE_DATA_data_offset 3
+#define SDMA_PKT_FENCE_DATA_data_mask   0xFFFFFFFF
+#define SDMA_PKT_FENCE_DATA_data_shift  0
+#define SDMA_PKT_FENCE_DATA_DATA(x) (((x) & SDMA_PKT_FENCE_DATA_data_mask) << SDMA_PKT_FENCE_DATA_data_shift)
+
+
+/*
+** Definitions for SDMA_PKT_SRBM_WRITE packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_SRBM_WRITE_HEADER_op_offset 0
+#define SDMA_PKT_SRBM_WRITE_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_SRBM_WRITE_HEADER_op_shift  0
+#define SDMA_PKT_SRBM_WRITE_HEADER_OP(x) (((x) & SDMA_PKT_SRBM_WRITE_HEADER_op_mask) << SDMA_PKT_SRBM_WRITE_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_SRBM_WRITE_HEADER_sub_op_offset 0
+#define SDMA_PKT_SRBM_WRITE_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_SRBM_WRITE_HEADER_sub_op_shift  8
+#define SDMA_PKT_SRBM_WRITE_HEADER_SUB_OP(x) (((x) & SDMA_PKT_SRBM_WRITE_HEADER_sub_op_mask) << SDMA_PKT_SRBM_WRITE_HEADER_sub_op_shift)
+
+/*define for byte_en field*/
+#define SDMA_PKT_SRBM_WRITE_HEADER_byte_en_offset 0
+#define SDMA_PKT_SRBM_WRITE_HEADER_byte_en_mask   0x0000000F
+#define SDMA_PKT_SRBM_WRITE_HEADER_byte_en_shift  28
+#define SDMA_PKT_SRBM_WRITE_HEADER_BYTE_EN(x) (((x) & SDMA_PKT_SRBM_WRITE_HEADER_byte_en_mask) << SDMA_PKT_SRBM_WRITE_HEADER_byte_en_shift)
+
+/*define for ADDR word*/
+/*define for addr field*/
+#define SDMA_PKT_SRBM_WRITE_ADDR_addr_offset 1
+#define SDMA_PKT_SRBM_WRITE_ADDR_addr_mask   0x0003FFFF
+#define SDMA_PKT_SRBM_WRITE_ADDR_addr_shift  0
+#define SDMA_PKT_SRBM_WRITE_ADDR_ADDR(x) (((x) & SDMA_PKT_SRBM_WRITE_ADDR_addr_mask) << SDMA_PKT_SRBM_WRITE_ADDR_addr_shift)
+
+/*define for apertureid field*/
+#define SDMA_PKT_SRBM_WRITE_ADDR_apertureid_offset 1
+#define SDMA_PKT_SRBM_WRITE_ADDR_apertureid_mask   0x00000FFF
+#define SDMA_PKT_SRBM_WRITE_ADDR_apertureid_shift  20
+#define SDMA_PKT_SRBM_WRITE_ADDR_APERTUREID(x) (((x) & SDMA_PKT_SRBM_WRITE_ADDR_apertureid_mask) << SDMA_PKT_SRBM_WRITE_ADDR_apertureid_shift)
+
+/*define for DATA word*/
+/*define for data field*/
+#define SDMA_PKT_SRBM_WRITE_DATA_data_offset 2
+#define SDMA_PKT_SRBM_WRITE_DATA_data_mask   0xFFFFFFFF
+#define SDMA_PKT_SRBM_WRITE_DATA_data_shift  0
+#define SDMA_PKT_SRBM_WRITE_DATA_DATA(x) (((x) & SDMA_PKT_SRBM_WRITE_DATA_data_mask) << SDMA_PKT_SRBM_WRITE_DATA_data_shift)
+
+
+/*
+** Definitions for SDMA_PKT_PRE_EXE packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_PRE_EXE_HEADER_op_offset 0
+#define SDMA_PKT_PRE_EXE_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_PRE_EXE_HEADER_op_shift  0
+#define SDMA_PKT_PRE_EXE_HEADER_OP(x) (((x) & SDMA_PKT_PRE_EXE_HEADER_op_mask) << SDMA_PKT_PRE_EXE_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_PRE_EXE_HEADER_sub_op_offset 0
+#define SDMA_PKT_PRE_EXE_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_PRE_EXE_HEADER_sub_op_shift  8
+#define SDMA_PKT_PRE_EXE_HEADER_SUB_OP(x) (((x) & SDMA_PKT_PRE_EXE_HEADER_sub_op_mask) << SDMA_PKT_PRE_EXE_HEADER_sub_op_shift)
+
+/*define for dev_sel field*/
+#define SDMA_PKT_PRE_EXE_HEADER_dev_sel_offset 0
+#define SDMA_PKT_PRE_EXE_HEADER_dev_sel_mask   0x000000FF
+#define SDMA_PKT_PRE_EXE_HEADER_dev_sel_shift  16
+#define SDMA_PKT_PRE_EXE_HEADER_DEV_SEL(x) (((x) & SDMA_PKT_PRE_EXE_HEADER_dev_sel_mask) << SDMA_PKT_PRE_EXE_HEADER_dev_sel_shift)
+
+/*define for EXEC_COUNT word*/
+/*define for exec_count field*/
+#define SDMA_PKT_PRE_EXE_EXEC_COUNT_exec_count_offset 1
+#define SDMA_PKT_PRE_EXE_EXEC_COUNT_exec_count_mask   0x00003FFF
+#define SDMA_PKT_PRE_EXE_EXEC_COUNT_exec_count_shift  0
+#define SDMA_PKT_PRE_EXE_EXEC_COUNT_EXEC_COUNT(x) (((x) & SDMA_PKT_PRE_EXE_EXEC_COUNT_exec_count_mask) << SDMA_PKT_PRE_EXE_EXEC_COUNT_exec_count_shift)
+
+
+/*
+** Definitions for SDMA_PKT_COND_EXE packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_COND_EXE_HEADER_op_offset 0
+#define SDMA_PKT_COND_EXE_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_COND_EXE_HEADER_op_shift  0
+#define SDMA_PKT_COND_EXE_HEADER_OP(x) (((x) & SDMA_PKT_COND_EXE_HEADER_op_mask) << SDMA_PKT_COND_EXE_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_COND_EXE_HEADER_sub_op_offset 0
+#define SDMA_PKT_COND_EXE_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_COND_EXE_HEADER_sub_op_shift  8
+#define SDMA_PKT_COND_EXE_HEADER_SUB_OP(x) (((x) & SDMA_PKT_COND_EXE_HEADER_sub_op_mask) << SDMA_PKT_COND_EXE_HEADER_sub_op_shift)
+
+/*define for cache_policy field*/
+#define SDMA_PKT_COND_EXE_HEADER_cache_policy_offset 0
+#define SDMA_PKT_COND_EXE_HEADER_cache_policy_mask   0x00000007
+#define SDMA_PKT_COND_EXE_HEADER_cache_policy_shift  24
+#define SDMA_PKT_COND_EXE_HEADER_CACHE_POLICY(x) (((x) & SDMA_PKT_COND_EXE_HEADER_cache_policy_mask) << SDMA_PKT_COND_EXE_HEADER_cache_policy_shift)
+
+/*define for cpv field*/
+#define SDMA_PKT_COND_EXE_HEADER_cpv_offset 0
+#define SDMA_PKT_COND_EXE_HEADER_cpv_mask   0x00000001
+#define SDMA_PKT_COND_EXE_HEADER_cpv_shift  28
+#define SDMA_PKT_COND_EXE_HEADER_CPV(x) (((x) & SDMA_PKT_COND_EXE_HEADER_cpv_mask) << SDMA_PKT_COND_EXE_HEADER_cpv_shift)
+
+/*define for ADDR_LO word*/
+/*define for addr_31_0 field*/
+#define SDMA_PKT_COND_EXE_ADDR_LO_addr_31_0_offset 1
+#define SDMA_PKT_COND_EXE_ADDR_LO_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_COND_EXE_ADDR_LO_addr_31_0_shift  0
+#define SDMA_PKT_COND_EXE_ADDR_LO_ADDR_31_0(x) (((x) & SDMA_PKT_COND_EXE_ADDR_LO_addr_31_0_mask) << SDMA_PKT_COND_EXE_ADDR_LO_addr_31_0_shift)
+
+/*define for ADDR_HI word*/
+/*define for addr_63_32 field*/
+#define SDMA_PKT_COND_EXE_ADDR_HI_addr_63_32_offset 2
+#define SDMA_PKT_COND_EXE_ADDR_HI_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_COND_EXE_ADDR_HI_addr_63_32_shift  0
+#define SDMA_PKT_COND_EXE_ADDR_HI_ADDR_63_32(x) (((x) & SDMA_PKT_COND_EXE_ADDR_HI_addr_63_32_mask) << SDMA_PKT_COND_EXE_ADDR_HI_addr_63_32_shift)
+
+/*define for REFERENCE word*/
+/*define for reference field*/
+#define SDMA_PKT_COND_EXE_REFERENCE_reference_offset 3
+#define SDMA_PKT_COND_EXE_REFERENCE_reference_mask   0xFFFFFFFF
+#define SDMA_PKT_COND_EXE_REFERENCE_reference_shift  0
+#define SDMA_PKT_COND_EXE_REFERENCE_REFERENCE(x) (((x) & SDMA_PKT_COND_EXE_REFERENCE_reference_mask) << SDMA_PKT_COND_EXE_REFERENCE_reference_shift)
+
+/*define for EXEC_COUNT word*/
+/*define for exec_count field*/
+#define SDMA_PKT_COND_EXE_EXEC_COUNT_exec_count_offset 4
+#define SDMA_PKT_COND_EXE_EXEC_COUNT_exec_count_mask   0x00003FFF
+#define SDMA_PKT_COND_EXE_EXEC_COUNT_exec_count_shift  0
+#define SDMA_PKT_COND_EXE_EXEC_COUNT_EXEC_COUNT(x) (((x) & SDMA_PKT_COND_EXE_EXEC_COUNT_exec_count_mask) << SDMA_PKT_COND_EXE_EXEC_COUNT_exec_count_shift)
+
+
+/*
+** Definitions for SDMA_PKT_CONSTANT_FILL packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_CONSTANT_FILL_HEADER_op_offset 0
+#define SDMA_PKT_CONSTANT_FILL_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_CONSTANT_FILL_HEADER_op_shift  0
+#define SDMA_PKT_CONSTANT_FILL_HEADER_OP(x) (((x) & SDMA_PKT_CONSTANT_FILL_HEADER_op_mask) << SDMA_PKT_CONSTANT_FILL_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_CONSTANT_FILL_HEADER_sub_op_offset 0
+#define SDMA_PKT_CONSTANT_FILL_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_CONSTANT_FILL_HEADER_sub_op_shift  8
+#define SDMA_PKT_CONSTANT_FILL_HEADER_SUB_OP(x) (((x) & SDMA_PKT_CONSTANT_FILL_HEADER_sub_op_mask) << SDMA_PKT_CONSTANT_FILL_HEADER_sub_op_shift)
+
+/*define for sw field*/
+#define SDMA_PKT_CONSTANT_FILL_HEADER_sw_offset 0
+#define SDMA_PKT_CONSTANT_FILL_HEADER_sw_mask   0x00000003
+#define SDMA_PKT_CONSTANT_FILL_HEADER_sw_shift  16
+#define SDMA_PKT_CONSTANT_FILL_HEADER_SW(x) (((x) & SDMA_PKT_CONSTANT_FILL_HEADER_sw_mask) << SDMA_PKT_CONSTANT_FILL_HEADER_sw_shift)
+
+/*define for cache_policy field*/
+#define SDMA_PKT_CONSTANT_FILL_HEADER_cache_policy_offset 0
+#define SDMA_PKT_CONSTANT_FILL_HEADER_cache_policy_mask   0x00000007
+#define SDMA_PKT_CONSTANT_FILL_HEADER_cache_policy_shift  24
+#define SDMA_PKT_CONSTANT_FILL_HEADER_CACHE_POLICY(x) (((x) & SDMA_PKT_CONSTANT_FILL_HEADER_cache_policy_mask) << SDMA_PKT_CONSTANT_FILL_HEADER_cache_policy_shift)
+
+/*define for cpv field*/
+#define SDMA_PKT_CONSTANT_FILL_HEADER_cpv_offset 0
+#define SDMA_PKT_CONSTANT_FILL_HEADER_cpv_mask   0x00000001
+#define SDMA_PKT_CONSTANT_FILL_HEADER_cpv_shift  28
+#define SDMA_PKT_CONSTANT_FILL_HEADER_CPV(x) (((x) & SDMA_PKT_CONSTANT_FILL_HEADER_cpv_mask) << SDMA_PKT_CONSTANT_FILL_HEADER_cpv_shift)
+
+/*define for fillsize field*/
+#define SDMA_PKT_CONSTANT_FILL_HEADER_fillsize_offset 0
+#define SDMA_PKT_CONSTANT_FILL_HEADER_fillsize_mask   0x00000003
+#define SDMA_PKT_CONSTANT_FILL_HEADER_fillsize_shift  30
+#define SDMA_PKT_CONSTANT_FILL_HEADER_FILLSIZE(x) (((x) & SDMA_PKT_CONSTANT_FILL_HEADER_fillsize_mask) << SDMA_PKT_CONSTANT_FILL_HEADER_fillsize_shift)
+
+/*define for DST_ADDR_LO word*/
+/*define for dst_addr_31_0 field*/
+#define SDMA_PKT_CONSTANT_FILL_DST_ADDR_LO_dst_addr_31_0_offset 1
+#define SDMA_PKT_CONSTANT_FILL_DST_ADDR_LO_dst_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_CONSTANT_FILL_DST_ADDR_LO_dst_addr_31_0_shift  0
+#define SDMA_PKT_CONSTANT_FILL_DST_ADDR_LO_DST_ADDR_31_0(x) (((x) & SDMA_PKT_CONSTANT_FILL_DST_ADDR_LO_dst_addr_31_0_mask) << SDMA_PKT_CONSTANT_FILL_DST_ADDR_LO_dst_addr_31_0_shift)
+
+/*define for DST_ADDR_HI word*/
+/*define for dst_addr_63_32 field*/
+#define SDMA_PKT_CONSTANT_FILL_DST_ADDR_HI_dst_addr_63_32_offset 2
+#define SDMA_PKT_CONSTANT_FILL_DST_ADDR_HI_dst_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_CONSTANT_FILL_DST_ADDR_HI_dst_addr_63_32_shift  0
+#define SDMA_PKT_CONSTANT_FILL_DST_ADDR_HI_DST_ADDR_63_32(x) (((x) & SDMA_PKT_CONSTANT_FILL_DST_ADDR_HI_dst_addr_63_32_mask) << SDMA_PKT_CONSTANT_FILL_DST_ADDR_HI_dst_addr_63_32_shift)
+
+/*define for DATA word*/
+/*define for src_data_31_0 field*/
+#define SDMA_PKT_CONSTANT_FILL_DATA_src_data_31_0_offset 3
+#define SDMA_PKT_CONSTANT_FILL_DATA_src_data_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_CONSTANT_FILL_DATA_src_data_31_0_shift  0
+#define SDMA_PKT_CONSTANT_FILL_DATA_SRC_DATA_31_0(x) (((x) & SDMA_PKT_CONSTANT_FILL_DATA_src_data_31_0_mask) << SDMA_PKT_CONSTANT_FILL_DATA_src_data_31_0_shift)
+
+/*define for COUNT word*/
+/*define for count field*/
+#define SDMA_PKT_CONSTANT_FILL_COUNT_count_offset 4
+#define SDMA_PKT_CONSTANT_FILL_COUNT_count_mask   0x3FFFFFFF
+#define SDMA_PKT_CONSTANT_FILL_COUNT_count_shift  0
+#define SDMA_PKT_CONSTANT_FILL_COUNT_COUNT(x) (((x) & SDMA_PKT_CONSTANT_FILL_COUNT_count_mask) << SDMA_PKT_CONSTANT_FILL_COUNT_count_shift)
+
+
+/*
+** Definitions for SDMA_PKT_DATA_FILL_MULTI packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_DATA_FILL_MULTI_HEADER_op_offset 0
+#define SDMA_PKT_DATA_FILL_MULTI_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_DATA_FILL_MULTI_HEADER_op_shift  0
+#define SDMA_PKT_DATA_FILL_MULTI_HEADER_OP(x) (((x) & SDMA_PKT_DATA_FILL_MULTI_HEADER_op_mask) << SDMA_PKT_DATA_FILL_MULTI_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_DATA_FILL_MULTI_HEADER_sub_op_offset 0
+#define SDMA_PKT_DATA_FILL_MULTI_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_DATA_FILL_MULTI_HEADER_sub_op_shift  8
+#define SDMA_PKT_DATA_FILL_MULTI_HEADER_SUB_OP(x) (((x) & SDMA_PKT_DATA_FILL_MULTI_HEADER_sub_op_mask) << SDMA_PKT_DATA_FILL_MULTI_HEADER_sub_op_shift)
+
+/*define for cache_policy field*/
+#define SDMA_PKT_DATA_FILL_MULTI_HEADER_cache_policy_offset 0
+#define SDMA_PKT_DATA_FILL_MULTI_HEADER_cache_policy_mask   0x00000007
+#define SDMA_PKT_DATA_FILL_MULTI_HEADER_cache_policy_shift  24
+#define SDMA_PKT_DATA_FILL_MULTI_HEADER_CACHE_POLICY(x) (((x) & SDMA_PKT_DATA_FILL_MULTI_HEADER_cache_policy_mask) << SDMA_PKT_DATA_FILL_MULTI_HEADER_cache_policy_shift)
+
+/*define for cpv field*/
+#define SDMA_PKT_DATA_FILL_MULTI_HEADER_cpv_offset 0
+#define SDMA_PKT_DATA_FILL_MULTI_HEADER_cpv_mask   0x00000001
+#define SDMA_PKT_DATA_FILL_MULTI_HEADER_cpv_shift  28
+#define SDMA_PKT_DATA_FILL_MULTI_HEADER_CPV(x) (((x) & SDMA_PKT_DATA_FILL_MULTI_HEADER_cpv_mask) << SDMA_PKT_DATA_FILL_MULTI_HEADER_cpv_shift)
+
+/*define for memlog_clr field*/
+#define SDMA_PKT_DATA_FILL_MULTI_HEADER_memlog_clr_offset 0
+#define SDMA_PKT_DATA_FILL_MULTI_HEADER_memlog_clr_mask   0x00000001
+#define SDMA_PKT_DATA_FILL_MULTI_HEADER_memlog_clr_shift  31
+#define SDMA_PKT_DATA_FILL_MULTI_HEADER_MEMLOG_CLR(x) (((x) & SDMA_PKT_DATA_FILL_MULTI_HEADER_memlog_clr_mask) << SDMA_PKT_DATA_FILL_MULTI_HEADER_memlog_clr_shift)
+
+/*define for BYTE_STRIDE word*/
+/*define for byte_stride field*/
+#define SDMA_PKT_DATA_FILL_MULTI_BYTE_STRIDE_byte_stride_offset 1
+#define SDMA_PKT_DATA_FILL_MULTI_BYTE_STRIDE_byte_stride_mask   0xFFFFFFFF
+#define SDMA_PKT_DATA_FILL_MULTI_BYTE_STRIDE_byte_stride_shift  0
+#define SDMA_PKT_DATA_FILL_MULTI_BYTE_STRIDE_BYTE_STRIDE(x) (((x) & SDMA_PKT_DATA_FILL_MULTI_BYTE_STRIDE_byte_stride_mask) << SDMA_PKT_DATA_FILL_MULTI_BYTE_STRIDE_byte_stride_shift)
+
+/*define for DMA_COUNT word*/
+/*define for dma_count field*/
+#define SDMA_PKT_DATA_FILL_MULTI_DMA_COUNT_dma_count_offset 2
+#define SDMA_PKT_DATA_FILL_MULTI_DMA_COUNT_dma_count_mask   0xFFFFFFFF
+#define SDMA_PKT_DATA_FILL_MULTI_DMA_COUNT_dma_count_shift  0
+#define SDMA_PKT_DATA_FILL_MULTI_DMA_COUNT_DMA_COUNT(x) (((x) & SDMA_PKT_DATA_FILL_MULTI_DMA_COUNT_dma_count_mask) << SDMA_PKT_DATA_FILL_MULTI_DMA_COUNT_dma_count_shift)
+
+/*define for DST_ADDR_LO word*/
+/*define for dst_addr_31_0 field*/
+#define SDMA_PKT_DATA_FILL_MULTI_DST_ADDR_LO_dst_addr_31_0_offset 3
+#define SDMA_PKT_DATA_FILL_MULTI_DST_ADDR_LO_dst_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_DATA_FILL_MULTI_DST_ADDR_LO_dst_addr_31_0_shift  0
+#define SDMA_PKT_DATA_FILL_MULTI_DST_ADDR_LO_DST_ADDR_31_0(x) (((x) & SDMA_PKT_DATA_FILL_MULTI_DST_ADDR_LO_dst_addr_31_0_mask) << SDMA_PKT_DATA_FILL_MULTI_DST_ADDR_LO_dst_addr_31_0_shift)
+
+/*define for DST_ADDR_HI word*/
+/*define for dst_addr_63_32 field*/
+#define SDMA_PKT_DATA_FILL_MULTI_DST_ADDR_HI_dst_addr_63_32_offset 4
+#define SDMA_PKT_DATA_FILL_MULTI_DST_ADDR_HI_dst_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_DATA_FILL_MULTI_DST_ADDR_HI_dst_addr_63_32_shift  0
+#define SDMA_PKT_DATA_FILL_MULTI_DST_ADDR_HI_DST_ADDR_63_32(x) (((x) & SDMA_PKT_DATA_FILL_MULTI_DST_ADDR_HI_dst_addr_63_32_mask) << SDMA_PKT_DATA_FILL_MULTI_DST_ADDR_HI_dst_addr_63_32_shift)
+
+/*define for BYTE_COUNT word*/
+/*define for count field*/
+#define SDMA_PKT_DATA_FILL_MULTI_BYTE_COUNT_count_offset 5
+#define SDMA_PKT_DATA_FILL_MULTI_BYTE_COUNT_count_mask   0x03FFFFFF
+#define SDMA_PKT_DATA_FILL_MULTI_BYTE_COUNT_count_shift  0
+#define SDMA_PKT_DATA_FILL_MULTI_BYTE_COUNT_COUNT(x) (((x) & SDMA_PKT_DATA_FILL_MULTI_BYTE_COUNT_count_mask) << SDMA_PKT_DATA_FILL_MULTI_BYTE_COUNT_count_shift)
+
+
+/*
+** Definitions for SDMA_PKT_POLL_REGMEM packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_POLL_REGMEM_HEADER_op_offset 0
+#define SDMA_PKT_POLL_REGMEM_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_POLL_REGMEM_HEADER_op_shift  0
+#define SDMA_PKT_POLL_REGMEM_HEADER_OP(x) (((x) & SDMA_PKT_POLL_REGMEM_HEADER_op_mask) << SDMA_PKT_POLL_REGMEM_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_POLL_REGMEM_HEADER_sub_op_offset 0
+#define SDMA_PKT_POLL_REGMEM_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_POLL_REGMEM_HEADER_sub_op_shift  8
+#define SDMA_PKT_POLL_REGMEM_HEADER_SUB_OP(x) (((x) & SDMA_PKT_POLL_REGMEM_HEADER_sub_op_mask) << SDMA_PKT_POLL_REGMEM_HEADER_sub_op_shift)
+
+/*define for cache_policy field*/
+#define SDMA_PKT_POLL_REGMEM_HEADER_cache_policy_offset 0
+#define SDMA_PKT_POLL_REGMEM_HEADER_cache_policy_mask   0x00000007
+#define SDMA_PKT_POLL_REGMEM_HEADER_cache_policy_shift  20
+#define SDMA_PKT_POLL_REGMEM_HEADER_CACHE_POLICY(x) (((x) & SDMA_PKT_POLL_REGMEM_HEADER_cache_policy_mask) << SDMA_PKT_POLL_REGMEM_HEADER_cache_policy_shift)
+
+/*define for cpv field*/
+#define SDMA_PKT_POLL_REGMEM_HEADER_cpv_offset 0
+#define SDMA_PKT_POLL_REGMEM_HEADER_cpv_mask   0x00000001
+#define SDMA_PKT_POLL_REGMEM_HEADER_cpv_shift  24
+#define SDMA_PKT_POLL_REGMEM_HEADER_CPV(x) (((x) & SDMA_PKT_POLL_REGMEM_HEADER_cpv_mask) << SDMA_PKT_POLL_REGMEM_HEADER_cpv_shift)
+
+/*define for hdp_flush field*/
+#define SDMA_PKT_POLL_REGMEM_HEADER_hdp_flush_offset 0
+#define SDMA_PKT_POLL_REGMEM_HEADER_hdp_flush_mask   0x00000001
+#define SDMA_PKT_POLL_REGMEM_HEADER_hdp_flush_shift  26
+#define SDMA_PKT_POLL_REGMEM_HEADER_HDP_FLUSH(x) (((x) & SDMA_PKT_POLL_REGMEM_HEADER_hdp_flush_mask) << SDMA_PKT_POLL_REGMEM_HEADER_hdp_flush_shift)
+
+/*define for func field*/
+#define SDMA_PKT_POLL_REGMEM_HEADER_func_offset 0
+#define SDMA_PKT_POLL_REGMEM_HEADER_func_mask   0x00000007
+#define SDMA_PKT_POLL_REGMEM_HEADER_func_shift  28
+#define SDMA_PKT_POLL_REGMEM_HEADER_FUNC(x) (((x) & SDMA_PKT_POLL_REGMEM_HEADER_func_mask) << SDMA_PKT_POLL_REGMEM_HEADER_func_shift)
+
+/*define for mem_poll field*/
+#define SDMA_PKT_POLL_REGMEM_HEADER_mem_poll_offset 0
+#define SDMA_PKT_POLL_REGMEM_HEADER_mem_poll_mask   0x00000001
+#define SDMA_PKT_POLL_REGMEM_HEADER_mem_poll_shift  31
+#define SDMA_PKT_POLL_REGMEM_HEADER_MEM_POLL(x) (((x) & SDMA_PKT_POLL_REGMEM_HEADER_mem_poll_mask) << SDMA_PKT_POLL_REGMEM_HEADER_mem_poll_shift)
+
+/*define for ADDR_LO word*/
+/*define for addr_31_0 field*/
+#define SDMA_PKT_POLL_REGMEM_ADDR_LO_addr_31_0_offset 1
+#define SDMA_PKT_POLL_REGMEM_ADDR_LO_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_POLL_REGMEM_ADDR_LO_addr_31_0_shift  0
+#define SDMA_PKT_POLL_REGMEM_ADDR_LO_ADDR_31_0(x) (((x) & SDMA_PKT_POLL_REGMEM_ADDR_LO_addr_31_0_mask) << SDMA_PKT_POLL_REGMEM_ADDR_LO_addr_31_0_shift)
+
+/*define for ADDR_HI word*/
+/*define for addr_63_32 field*/
+#define SDMA_PKT_POLL_REGMEM_ADDR_HI_addr_63_32_offset 2
+#define SDMA_PKT_POLL_REGMEM_ADDR_HI_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_POLL_REGMEM_ADDR_HI_addr_63_32_shift  0
+#define SDMA_PKT_POLL_REGMEM_ADDR_HI_ADDR_63_32(x) (((x) & SDMA_PKT_POLL_REGMEM_ADDR_HI_addr_63_32_mask) << SDMA_PKT_POLL_REGMEM_ADDR_HI_addr_63_32_shift)
+
+/*define for VALUE word*/
+/*define for value field*/
+#define SDMA_PKT_POLL_REGMEM_VALUE_value_offset 3
+#define SDMA_PKT_POLL_REGMEM_VALUE_value_mask   0xFFFFFFFF
+#define SDMA_PKT_POLL_REGMEM_VALUE_value_shift  0
+#define SDMA_PKT_POLL_REGMEM_VALUE_VALUE(x) (((x) & SDMA_PKT_POLL_REGMEM_VALUE_value_mask) << SDMA_PKT_POLL_REGMEM_VALUE_value_shift)
+
+/*define for MASK word*/
+/*define for mask field*/
+#define SDMA_PKT_POLL_REGMEM_MASK_mask_offset 4
+#define SDMA_PKT_POLL_REGMEM_MASK_mask_mask   0xFFFFFFFF
+#define SDMA_PKT_POLL_REGMEM_MASK_mask_shift  0
+#define SDMA_PKT_POLL_REGMEM_MASK_MASK(x) (((x) & SDMA_PKT_POLL_REGMEM_MASK_mask_mask) << SDMA_PKT_POLL_REGMEM_MASK_mask_shift)
+
+/*define for DW5 word*/
+/*define for interval field*/
+#define SDMA_PKT_POLL_REGMEM_DW5_interval_offset 5
+#define SDMA_PKT_POLL_REGMEM_DW5_interval_mask   0x0000FFFF
+#define SDMA_PKT_POLL_REGMEM_DW5_interval_shift  0
+#define SDMA_PKT_POLL_REGMEM_DW5_INTERVAL(x) (((x) & SDMA_PKT_POLL_REGMEM_DW5_interval_mask) << SDMA_PKT_POLL_REGMEM_DW5_interval_shift)
+
+/*define for retry_count field*/
+#define SDMA_PKT_POLL_REGMEM_DW5_retry_count_offset 5
+#define SDMA_PKT_POLL_REGMEM_DW5_retry_count_mask   0x00000FFF
+#define SDMA_PKT_POLL_REGMEM_DW5_retry_count_shift  16
+#define SDMA_PKT_POLL_REGMEM_DW5_RETRY_COUNT(x) (((x) & SDMA_PKT_POLL_REGMEM_DW5_retry_count_mask) << SDMA_PKT_POLL_REGMEM_DW5_retry_count_shift)
+
+
+/*
+** Definitions for SDMA_PKT_POLL_REG_WRITE_MEM packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_POLL_REG_WRITE_MEM_HEADER_op_offset 0
+#define SDMA_PKT_POLL_REG_WRITE_MEM_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_POLL_REG_WRITE_MEM_HEADER_op_shift  0
+#define SDMA_PKT_POLL_REG_WRITE_MEM_HEADER_OP(x) (((x) & SDMA_PKT_POLL_REG_WRITE_MEM_HEADER_op_mask) << SDMA_PKT_POLL_REG_WRITE_MEM_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_POLL_REG_WRITE_MEM_HEADER_sub_op_offset 0
+#define SDMA_PKT_POLL_REG_WRITE_MEM_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_POLL_REG_WRITE_MEM_HEADER_sub_op_shift  8
+#define SDMA_PKT_POLL_REG_WRITE_MEM_HEADER_SUB_OP(x) (((x) & SDMA_PKT_POLL_REG_WRITE_MEM_HEADER_sub_op_mask) << SDMA_PKT_POLL_REG_WRITE_MEM_HEADER_sub_op_shift)
+
+/*define for cache_policy field*/
+#define SDMA_PKT_POLL_REG_WRITE_MEM_HEADER_cache_policy_offset 0
+#define SDMA_PKT_POLL_REG_WRITE_MEM_HEADER_cache_policy_mask   0x00000007
+#define SDMA_PKT_POLL_REG_WRITE_MEM_HEADER_cache_policy_shift  24
+#define SDMA_PKT_POLL_REG_WRITE_MEM_HEADER_CACHE_POLICY(x) (((x) & SDMA_PKT_POLL_REG_WRITE_MEM_HEADER_cache_policy_mask) << SDMA_PKT_POLL_REG_WRITE_MEM_HEADER_cache_policy_shift)
+
+/*define for cpv field*/
+#define SDMA_PKT_POLL_REG_WRITE_MEM_HEADER_cpv_offset 0
+#define SDMA_PKT_POLL_REG_WRITE_MEM_HEADER_cpv_mask   0x00000001
+#define SDMA_PKT_POLL_REG_WRITE_MEM_HEADER_cpv_shift  28
+#define SDMA_PKT_POLL_REG_WRITE_MEM_HEADER_CPV(x) (((x) & SDMA_PKT_POLL_REG_WRITE_MEM_HEADER_cpv_mask) << SDMA_PKT_POLL_REG_WRITE_MEM_HEADER_cpv_shift)
+
+/*define for SRC_ADDR word*/
+/*define for addr_31_2 field*/
+#define SDMA_PKT_POLL_REG_WRITE_MEM_SRC_ADDR_addr_31_2_offset 1
+#define SDMA_PKT_POLL_REG_WRITE_MEM_SRC_ADDR_addr_31_2_mask   0x3FFFFFFF
+#define SDMA_PKT_POLL_REG_WRITE_MEM_SRC_ADDR_addr_31_2_shift  2
+#define SDMA_PKT_POLL_REG_WRITE_MEM_SRC_ADDR_ADDR_31_2(x) (((x) & SDMA_PKT_POLL_REG_WRITE_MEM_SRC_ADDR_addr_31_2_mask) << SDMA_PKT_POLL_REG_WRITE_MEM_SRC_ADDR_addr_31_2_shift)
+
+/*define for DST_ADDR_LO word*/
+/*define for addr_31_0 field*/
+#define SDMA_PKT_POLL_REG_WRITE_MEM_DST_ADDR_LO_addr_31_0_offset 2
+#define SDMA_PKT_POLL_REG_WRITE_MEM_DST_ADDR_LO_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_POLL_REG_WRITE_MEM_DST_ADDR_LO_addr_31_0_shift  0
+#define SDMA_PKT_POLL_REG_WRITE_MEM_DST_ADDR_LO_ADDR_31_0(x) (((x) & SDMA_PKT_POLL_REG_WRITE_MEM_DST_ADDR_LO_addr_31_0_mask) << SDMA_PKT_POLL_REG_WRITE_MEM_DST_ADDR_LO_addr_31_0_shift)
+
+/*define for DST_ADDR_HI word*/
+/*define for addr_63_32 field*/
+#define SDMA_PKT_POLL_REG_WRITE_MEM_DST_ADDR_HI_addr_63_32_offset 3
+#define SDMA_PKT_POLL_REG_WRITE_MEM_DST_ADDR_HI_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_POLL_REG_WRITE_MEM_DST_ADDR_HI_addr_63_32_shift  0
+#define SDMA_PKT_POLL_REG_WRITE_MEM_DST_ADDR_HI_ADDR_63_32(x) (((x) & SDMA_PKT_POLL_REG_WRITE_MEM_DST_ADDR_HI_addr_63_32_mask) << SDMA_PKT_POLL_REG_WRITE_MEM_DST_ADDR_HI_addr_63_32_shift)
+
+
+/*
+** Definitions for SDMA_PKT_POLL_DBIT_WRITE_MEM packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_POLL_DBIT_WRITE_MEM_HEADER_op_offset 0
+#define SDMA_PKT_POLL_DBIT_WRITE_MEM_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_POLL_DBIT_WRITE_MEM_HEADER_op_shift  0
+#define SDMA_PKT_POLL_DBIT_WRITE_MEM_HEADER_OP(x) (((x) & SDMA_PKT_POLL_DBIT_WRITE_MEM_HEADER_op_mask) << SDMA_PKT_POLL_DBIT_WRITE_MEM_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_POLL_DBIT_WRITE_MEM_HEADER_sub_op_offset 0
+#define SDMA_PKT_POLL_DBIT_WRITE_MEM_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_POLL_DBIT_WRITE_MEM_HEADER_sub_op_shift  8
+#define SDMA_PKT_POLL_DBIT_WRITE_MEM_HEADER_SUB_OP(x) (((x) & SDMA_PKT_POLL_DBIT_WRITE_MEM_HEADER_sub_op_mask) << SDMA_PKT_POLL_DBIT_WRITE_MEM_HEADER_sub_op_shift)
+
+/*define for ea field*/
+#define SDMA_PKT_POLL_DBIT_WRITE_MEM_HEADER_ea_offset 0
+#define SDMA_PKT_POLL_DBIT_WRITE_MEM_HEADER_ea_mask   0x00000003
+#define SDMA_PKT_POLL_DBIT_WRITE_MEM_HEADER_ea_shift  16
+#define SDMA_PKT_POLL_DBIT_WRITE_MEM_HEADER_EA(x) (((x) & SDMA_PKT_POLL_DBIT_WRITE_MEM_HEADER_ea_mask) << SDMA_PKT_POLL_DBIT_WRITE_MEM_HEADER_ea_shift)
+
+/*define for cache_policy field*/
+#define SDMA_PKT_POLL_DBIT_WRITE_MEM_HEADER_cache_policy_offset 0
+#define SDMA_PKT_POLL_DBIT_WRITE_MEM_HEADER_cache_policy_mask   0x00000007
+#define SDMA_PKT_POLL_DBIT_WRITE_MEM_HEADER_cache_policy_shift  24
+#define SDMA_PKT_POLL_DBIT_WRITE_MEM_HEADER_CACHE_POLICY(x) (((x) & SDMA_PKT_POLL_DBIT_WRITE_MEM_HEADER_cache_policy_mask) << SDMA_PKT_POLL_DBIT_WRITE_MEM_HEADER_cache_policy_shift)
+
+/*define for cpv field*/
+#define SDMA_PKT_POLL_DBIT_WRITE_MEM_HEADER_cpv_offset 0
+#define SDMA_PKT_POLL_DBIT_WRITE_MEM_HEADER_cpv_mask   0x00000001
+#define SDMA_PKT_POLL_DBIT_WRITE_MEM_HEADER_cpv_shift  28
+#define SDMA_PKT_POLL_DBIT_WRITE_MEM_HEADER_CPV(x) (((x) & SDMA_PKT_POLL_DBIT_WRITE_MEM_HEADER_cpv_mask) << SDMA_PKT_POLL_DBIT_WRITE_MEM_HEADER_cpv_shift)
+
+/*define for DST_ADDR_LO word*/
+/*define for addr_31_0 field*/
+#define SDMA_PKT_POLL_DBIT_WRITE_MEM_DST_ADDR_LO_addr_31_0_offset 1
+#define SDMA_PKT_POLL_DBIT_WRITE_MEM_DST_ADDR_LO_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_POLL_DBIT_WRITE_MEM_DST_ADDR_LO_addr_31_0_shift  0
+#define SDMA_PKT_POLL_DBIT_WRITE_MEM_DST_ADDR_LO_ADDR_31_0(x) (((x) & SDMA_PKT_POLL_DBIT_WRITE_MEM_DST_ADDR_LO_addr_31_0_mask) << SDMA_PKT_POLL_DBIT_WRITE_MEM_DST_ADDR_LO_addr_31_0_shift)
+
+/*define for DST_ADDR_HI word*/
+/*define for addr_63_32 field*/
+#define SDMA_PKT_POLL_DBIT_WRITE_MEM_DST_ADDR_HI_addr_63_32_offset 2
+#define SDMA_PKT_POLL_DBIT_WRITE_MEM_DST_ADDR_HI_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_POLL_DBIT_WRITE_MEM_DST_ADDR_HI_addr_63_32_shift  0
+#define SDMA_PKT_POLL_DBIT_WRITE_MEM_DST_ADDR_HI_ADDR_63_32(x) (((x) & SDMA_PKT_POLL_DBIT_WRITE_MEM_DST_ADDR_HI_addr_63_32_mask) << SDMA_PKT_POLL_DBIT_WRITE_MEM_DST_ADDR_HI_addr_63_32_shift)
+
+/*define for START_PAGE word*/
+/*define for addr_31_4 field*/
+#define SDMA_PKT_POLL_DBIT_WRITE_MEM_START_PAGE_addr_31_4_offset 3
+#define SDMA_PKT_POLL_DBIT_WRITE_MEM_START_PAGE_addr_31_4_mask   0x0FFFFFFF
+#define SDMA_PKT_POLL_DBIT_WRITE_MEM_START_PAGE_addr_31_4_shift  4
+#define SDMA_PKT_POLL_DBIT_WRITE_MEM_START_PAGE_ADDR_31_4(x) (((x) & SDMA_PKT_POLL_DBIT_WRITE_MEM_START_PAGE_addr_31_4_mask) << SDMA_PKT_POLL_DBIT_WRITE_MEM_START_PAGE_addr_31_4_shift)
+
+/*define for PAGE_NUM word*/
+/*define for page_num_31_0 field*/
+#define SDMA_PKT_POLL_DBIT_WRITE_MEM_PAGE_NUM_page_num_31_0_offset 4
+#define SDMA_PKT_POLL_DBIT_WRITE_MEM_PAGE_NUM_page_num_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_POLL_DBIT_WRITE_MEM_PAGE_NUM_page_num_31_0_shift  0
+#define SDMA_PKT_POLL_DBIT_WRITE_MEM_PAGE_NUM_PAGE_NUM_31_0(x) (((x) & SDMA_PKT_POLL_DBIT_WRITE_MEM_PAGE_NUM_page_num_31_0_mask) << SDMA_PKT_POLL_DBIT_WRITE_MEM_PAGE_NUM_page_num_31_0_shift)
+
+
+/*
+** Definitions for SDMA_PKT_POLL_MEM_VERIFY packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_POLL_MEM_VERIFY_HEADER_op_offset 0
+#define SDMA_PKT_POLL_MEM_VERIFY_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_POLL_MEM_VERIFY_HEADER_op_shift  0
+#define SDMA_PKT_POLL_MEM_VERIFY_HEADER_OP(x) (((x) & SDMA_PKT_POLL_MEM_VERIFY_HEADER_op_mask) << SDMA_PKT_POLL_MEM_VERIFY_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_POLL_MEM_VERIFY_HEADER_sub_op_offset 0
+#define SDMA_PKT_POLL_MEM_VERIFY_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_POLL_MEM_VERIFY_HEADER_sub_op_shift  8
+#define SDMA_PKT_POLL_MEM_VERIFY_HEADER_SUB_OP(x) (((x) & SDMA_PKT_POLL_MEM_VERIFY_HEADER_sub_op_mask) << SDMA_PKT_POLL_MEM_VERIFY_HEADER_sub_op_shift)
+
+/*define for cache_policy field*/
+#define SDMA_PKT_POLL_MEM_VERIFY_HEADER_cache_policy_offset 0
+#define SDMA_PKT_POLL_MEM_VERIFY_HEADER_cache_policy_mask   0x00000007
+#define SDMA_PKT_POLL_MEM_VERIFY_HEADER_cache_policy_shift  24
+#define SDMA_PKT_POLL_MEM_VERIFY_HEADER_CACHE_POLICY(x) (((x) & SDMA_PKT_POLL_MEM_VERIFY_HEADER_cache_policy_mask) << SDMA_PKT_POLL_MEM_VERIFY_HEADER_cache_policy_shift)
+
+/*define for cpv field*/
+#define SDMA_PKT_POLL_MEM_VERIFY_HEADER_cpv_offset 0
+#define SDMA_PKT_POLL_MEM_VERIFY_HEADER_cpv_mask   0x00000001
+#define SDMA_PKT_POLL_MEM_VERIFY_HEADER_cpv_shift  28
+#define SDMA_PKT_POLL_MEM_VERIFY_HEADER_CPV(x) (((x) & SDMA_PKT_POLL_MEM_VERIFY_HEADER_cpv_mask) << SDMA_PKT_POLL_MEM_VERIFY_HEADER_cpv_shift)
+
+/*define for mode field*/
+#define SDMA_PKT_POLL_MEM_VERIFY_HEADER_mode_offset 0
+#define SDMA_PKT_POLL_MEM_VERIFY_HEADER_mode_mask   0x00000001
+#define SDMA_PKT_POLL_MEM_VERIFY_HEADER_mode_shift  31
+#define SDMA_PKT_POLL_MEM_VERIFY_HEADER_MODE(x) (((x) & SDMA_PKT_POLL_MEM_VERIFY_HEADER_mode_mask) << SDMA_PKT_POLL_MEM_VERIFY_HEADER_mode_shift)
+
+/*define for PATTERN word*/
+/*define for pattern field*/
+#define SDMA_PKT_POLL_MEM_VERIFY_PATTERN_pattern_offset 1
+#define SDMA_PKT_POLL_MEM_VERIFY_PATTERN_pattern_mask   0xFFFFFFFF
+#define SDMA_PKT_POLL_MEM_VERIFY_PATTERN_pattern_shift  0
+#define SDMA_PKT_POLL_MEM_VERIFY_PATTERN_PATTERN(x) (((x) & SDMA_PKT_POLL_MEM_VERIFY_PATTERN_pattern_mask) << SDMA_PKT_POLL_MEM_VERIFY_PATTERN_pattern_shift)
+
+/*define for CMP0_ADDR_START_LO word*/
+/*define for cmp0_start_31_0 field*/
+#define SDMA_PKT_POLL_MEM_VERIFY_CMP0_ADDR_START_LO_cmp0_start_31_0_offset 2
+#define SDMA_PKT_POLL_MEM_VERIFY_CMP0_ADDR_START_LO_cmp0_start_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_POLL_MEM_VERIFY_CMP0_ADDR_START_LO_cmp0_start_31_0_shift  0
+#define SDMA_PKT_POLL_MEM_VERIFY_CMP0_ADDR_START_LO_CMP0_START_31_0(x) (((x) & SDMA_PKT_POLL_MEM_VERIFY_CMP0_ADDR_START_LO_cmp0_start_31_0_mask) << SDMA_PKT_POLL_MEM_VERIFY_CMP0_ADDR_START_LO_cmp0_start_31_0_shift)
+
+/*define for CMP0_ADDR_START_HI word*/
+/*define for cmp0_start_63_32 field*/
+#define SDMA_PKT_POLL_MEM_VERIFY_CMP0_ADDR_START_HI_cmp0_start_63_32_offset 3
+#define SDMA_PKT_POLL_MEM_VERIFY_CMP0_ADDR_START_HI_cmp0_start_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_POLL_MEM_VERIFY_CMP0_ADDR_START_HI_cmp0_start_63_32_shift  0
+#define SDMA_PKT_POLL_MEM_VERIFY_CMP0_ADDR_START_HI_CMP0_START_63_32(x) (((x) & SDMA_PKT_POLL_MEM_VERIFY_CMP0_ADDR_START_HI_cmp0_start_63_32_mask) << SDMA_PKT_POLL_MEM_VERIFY_CMP0_ADDR_START_HI_cmp0_start_63_32_shift)
+
+/*define for CMP0_ADDR_END_LO word*/
+/*define for cmp0_end_31_0 field*/
+#define SDMA_PKT_POLL_MEM_VERIFY_CMP0_ADDR_END_LO_cmp0_end_31_0_offset 4
+#define SDMA_PKT_POLL_MEM_VERIFY_CMP0_ADDR_END_LO_cmp0_end_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_POLL_MEM_VERIFY_CMP0_ADDR_END_LO_cmp0_end_31_0_shift  0
+#define SDMA_PKT_POLL_MEM_VERIFY_CMP0_ADDR_END_LO_CMP0_END_31_0(x) (((x) & SDMA_PKT_POLL_MEM_VERIFY_CMP0_ADDR_END_LO_cmp0_end_31_0_mask) << SDMA_PKT_POLL_MEM_VERIFY_CMP0_ADDR_END_LO_cmp0_end_31_0_shift)
+
+/*define for CMP0_ADDR_END_HI word*/
+/*define for cmp0_end_63_32 field*/
+#define SDMA_PKT_POLL_MEM_VERIFY_CMP0_ADDR_END_HI_cmp0_end_63_32_offset 5
+#define SDMA_PKT_POLL_MEM_VERIFY_CMP0_ADDR_END_HI_cmp0_end_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_POLL_MEM_VERIFY_CMP0_ADDR_END_HI_cmp0_end_63_32_shift  0
+#define SDMA_PKT_POLL_MEM_VERIFY_CMP0_ADDR_END_HI_CMP0_END_63_32(x) (((x) & SDMA_PKT_POLL_MEM_VERIFY_CMP0_ADDR_END_HI_cmp0_end_63_32_mask) << SDMA_PKT_POLL_MEM_VERIFY_CMP0_ADDR_END_HI_cmp0_end_63_32_shift)
+
+/*define for CMP1_ADDR_START_LO word*/
+/*define for cmp1_start_31_0 field*/
+#define SDMA_PKT_POLL_MEM_VERIFY_CMP1_ADDR_START_LO_cmp1_start_31_0_offset 6
+#define SDMA_PKT_POLL_MEM_VERIFY_CMP1_ADDR_START_LO_cmp1_start_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_POLL_MEM_VERIFY_CMP1_ADDR_START_LO_cmp1_start_31_0_shift  0
+#define SDMA_PKT_POLL_MEM_VERIFY_CMP1_ADDR_START_LO_CMP1_START_31_0(x) (((x) & SDMA_PKT_POLL_MEM_VERIFY_CMP1_ADDR_START_LO_cmp1_start_31_0_mask) << SDMA_PKT_POLL_MEM_VERIFY_CMP1_ADDR_START_LO_cmp1_start_31_0_shift)
+
+/*define for CMP1_ADDR_START_HI word*/
+/*define for cmp1_start_63_32 field*/
+#define SDMA_PKT_POLL_MEM_VERIFY_CMP1_ADDR_START_HI_cmp1_start_63_32_offset 7
+#define SDMA_PKT_POLL_MEM_VERIFY_CMP1_ADDR_START_HI_cmp1_start_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_POLL_MEM_VERIFY_CMP1_ADDR_START_HI_cmp1_start_63_32_shift  0
+#define SDMA_PKT_POLL_MEM_VERIFY_CMP1_ADDR_START_HI_CMP1_START_63_32(x) (((x) & SDMA_PKT_POLL_MEM_VERIFY_CMP1_ADDR_START_HI_cmp1_start_63_32_mask) << SDMA_PKT_POLL_MEM_VERIFY_CMP1_ADDR_START_HI_cmp1_start_63_32_shift)
+
+/*define for CMP1_ADDR_END_LO word*/
+/*define for cmp1_end_31_0 field*/
+#define SDMA_PKT_POLL_MEM_VERIFY_CMP1_ADDR_END_LO_cmp1_end_31_0_offset 8
+#define SDMA_PKT_POLL_MEM_VERIFY_CMP1_ADDR_END_LO_cmp1_end_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_POLL_MEM_VERIFY_CMP1_ADDR_END_LO_cmp1_end_31_0_shift  0
+#define SDMA_PKT_POLL_MEM_VERIFY_CMP1_ADDR_END_LO_CMP1_END_31_0(x) (((x) & SDMA_PKT_POLL_MEM_VERIFY_CMP1_ADDR_END_LO_cmp1_end_31_0_mask) << SDMA_PKT_POLL_MEM_VERIFY_CMP1_ADDR_END_LO_cmp1_end_31_0_shift)
+
+/*define for CMP1_ADDR_END_HI word*/
+/*define for cmp1_end_63_32 field*/
+#define SDMA_PKT_POLL_MEM_VERIFY_CMP1_ADDR_END_HI_cmp1_end_63_32_offset 9
+#define SDMA_PKT_POLL_MEM_VERIFY_CMP1_ADDR_END_HI_cmp1_end_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_POLL_MEM_VERIFY_CMP1_ADDR_END_HI_cmp1_end_63_32_shift  0
+#define SDMA_PKT_POLL_MEM_VERIFY_CMP1_ADDR_END_HI_CMP1_END_63_32(x) (((x) & SDMA_PKT_POLL_MEM_VERIFY_CMP1_ADDR_END_HI_cmp1_end_63_32_mask) << SDMA_PKT_POLL_MEM_VERIFY_CMP1_ADDR_END_HI_cmp1_end_63_32_shift)
+
+/*define for REC_ADDR_LO word*/
+/*define for rec_31_0 field*/
+#define SDMA_PKT_POLL_MEM_VERIFY_REC_ADDR_LO_rec_31_0_offset 10
+#define SDMA_PKT_POLL_MEM_VERIFY_REC_ADDR_LO_rec_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_POLL_MEM_VERIFY_REC_ADDR_LO_rec_31_0_shift  0
+#define SDMA_PKT_POLL_MEM_VERIFY_REC_ADDR_LO_REC_31_0(x) (((x) & SDMA_PKT_POLL_MEM_VERIFY_REC_ADDR_LO_rec_31_0_mask) << SDMA_PKT_POLL_MEM_VERIFY_REC_ADDR_LO_rec_31_0_shift)
+
+/*define for REC_ADDR_HI word*/
+/*define for rec_63_32 field*/
+#define SDMA_PKT_POLL_MEM_VERIFY_REC_ADDR_HI_rec_63_32_offset 11
+#define SDMA_PKT_POLL_MEM_VERIFY_REC_ADDR_HI_rec_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_POLL_MEM_VERIFY_REC_ADDR_HI_rec_63_32_shift  0
+#define SDMA_PKT_POLL_MEM_VERIFY_REC_ADDR_HI_REC_63_32(x) (((x) & SDMA_PKT_POLL_MEM_VERIFY_REC_ADDR_HI_rec_63_32_mask) << SDMA_PKT_POLL_MEM_VERIFY_REC_ADDR_HI_rec_63_32_shift)
+
+/*define for RESERVED word*/
+/*define for reserved field*/
+#define SDMA_PKT_POLL_MEM_VERIFY_RESERVED_reserved_offset 12
+#define SDMA_PKT_POLL_MEM_VERIFY_RESERVED_reserved_mask   0xFFFFFFFF
+#define SDMA_PKT_POLL_MEM_VERIFY_RESERVED_reserved_shift  0
+#define SDMA_PKT_POLL_MEM_VERIFY_RESERVED_RESERVED(x) (((x) & SDMA_PKT_POLL_MEM_VERIFY_RESERVED_reserved_mask) << SDMA_PKT_POLL_MEM_VERIFY_RESERVED_reserved_shift)
+
+
+/*
+** Definitions for SDMA_PKT_ATOMIC packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_ATOMIC_HEADER_op_offset 0
+#define SDMA_PKT_ATOMIC_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_ATOMIC_HEADER_op_shift  0
+#define SDMA_PKT_ATOMIC_HEADER_OP(x) (((x) & SDMA_PKT_ATOMIC_HEADER_op_mask) << SDMA_PKT_ATOMIC_HEADER_op_shift)
+
+/*define for loop field*/
+#define SDMA_PKT_ATOMIC_HEADER_loop_offset 0
+#define SDMA_PKT_ATOMIC_HEADER_loop_mask   0x00000001
+#define SDMA_PKT_ATOMIC_HEADER_loop_shift  16
+#define SDMA_PKT_ATOMIC_HEADER_LOOP(x) (((x) & SDMA_PKT_ATOMIC_HEADER_loop_mask) << SDMA_PKT_ATOMIC_HEADER_loop_shift)
+
+/*define for tmz field*/
+#define SDMA_PKT_ATOMIC_HEADER_tmz_offset 0
+#define SDMA_PKT_ATOMIC_HEADER_tmz_mask   0x00000001
+#define SDMA_PKT_ATOMIC_HEADER_tmz_shift  18
+#define SDMA_PKT_ATOMIC_HEADER_TMZ(x) (((x) & SDMA_PKT_ATOMIC_HEADER_tmz_mask) << SDMA_PKT_ATOMIC_HEADER_tmz_shift)
+
+/*define for cache_policy field*/
+#define SDMA_PKT_ATOMIC_HEADER_cache_policy_offset 0
+#define SDMA_PKT_ATOMIC_HEADER_cache_policy_mask   0x00000007
+#define SDMA_PKT_ATOMIC_HEADER_cache_policy_shift  20
+#define SDMA_PKT_ATOMIC_HEADER_CACHE_POLICY(x) (((x) & SDMA_PKT_ATOMIC_HEADER_cache_policy_mask) << SDMA_PKT_ATOMIC_HEADER_cache_policy_shift)
+
+/*define for cpv field*/
+#define SDMA_PKT_ATOMIC_HEADER_cpv_offset 0
+#define SDMA_PKT_ATOMIC_HEADER_cpv_mask   0x00000001
+#define SDMA_PKT_ATOMIC_HEADER_cpv_shift  24
+#define SDMA_PKT_ATOMIC_HEADER_CPV(x) (((x) & SDMA_PKT_ATOMIC_HEADER_cpv_mask) << SDMA_PKT_ATOMIC_HEADER_cpv_shift)
+
+/*define for atomic_op field*/
+#define SDMA_PKT_ATOMIC_HEADER_atomic_op_offset 0
+#define SDMA_PKT_ATOMIC_HEADER_atomic_op_mask   0x0000007F
+#define SDMA_PKT_ATOMIC_HEADER_atomic_op_shift  25
+#define SDMA_PKT_ATOMIC_HEADER_ATOMIC_OP(x) (((x) & SDMA_PKT_ATOMIC_HEADER_atomic_op_mask) << SDMA_PKT_ATOMIC_HEADER_atomic_op_shift)
+
+/*define for ADDR_LO word*/
+/*define for addr_31_0 field*/
+#define SDMA_PKT_ATOMIC_ADDR_LO_addr_31_0_offset 1
+#define SDMA_PKT_ATOMIC_ADDR_LO_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_ATOMIC_ADDR_LO_addr_31_0_shift  0
+#define SDMA_PKT_ATOMIC_ADDR_LO_ADDR_31_0(x) (((x) & SDMA_PKT_ATOMIC_ADDR_LO_addr_31_0_mask) << SDMA_PKT_ATOMIC_ADDR_LO_addr_31_0_shift)
+
+/*define for ADDR_HI word*/
+/*define for addr_63_32 field*/
+#define SDMA_PKT_ATOMIC_ADDR_HI_addr_63_32_offset 2
+#define SDMA_PKT_ATOMIC_ADDR_HI_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_ATOMIC_ADDR_HI_addr_63_32_shift  0
+#define SDMA_PKT_ATOMIC_ADDR_HI_ADDR_63_32(x) (((x) & SDMA_PKT_ATOMIC_ADDR_HI_addr_63_32_mask) << SDMA_PKT_ATOMIC_ADDR_HI_addr_63_32_shift)
+
+/*define for SRC_DATA_LO word*/
+/*define for src_data_31_0 field*/
+#define SDMA_PKT_ATOMIC_SRC_DATA_LO_src_data_31_0_offset 3
+#define SDMA_PKT_ATOMIC_SRC_DATA_LO_src_data_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_ATOMIC_SRC_DATA_LO_src_data_31_0_shift  0
+#define SDMA_PKT_ATOMIC_SRC_DATA_LO_SRC_DATA_31_0(x) (((x) & SDMA_PKT_ATOMIC_SRC_DATA_LO_src_data_31_0_mask) << SDMA_PKT_ATOMIC_SRC_DATA_LO_src_data_31_0_shift)
+
+/*define for SRC_DATA_HI word*/
+/*define for src_data_63_32 field*/
+#define SDMA_PKT_ATOMIC_SRC_DATA_HI_src_data_63_32_offset 4
+#define SDMA_PKT_ATOMIC_SRC_DATA_HI_src_data_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_ATOMIC_SRC_DATA_HI_src_data_63_32_shift  0
+#define SDMA_PKT_ATOMIC_SRC_DATA_HI_SRC_DATA_63_32(x) (((x) & SDMA_PKT_ATOMIC_SRC_DATA_HI_src_data_63_32_mask) << SDMA_PKT_ATOMIC_SRC_DATA_HI_src_data_63_32_shift)
+
+/*define for CMP_DATA_LO word*/
+/*define for cmp_data_31_0 field*/
+#define SDMA_PKT_ATOMIC_CMP_DATA_LO_cmp_data_31_0_offset 5
+#define SDMA_PKT_ATOMIC_CMP_DATA_LO_cmp_data_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_ATOMIC_CMP_DATA_LO_cmp_data_31_0_shift  0
+#define SDMA_PKT_ATOMIC_CMP_DATA_LO_CMP_DATA_31_0(x) (((x) & SDMA_PKT_ATOMIC_CMP_DATA_LO_cmp_data_31_0_mask) << SDMA_PKT_ATOMIC_CMP_DATA_LO_cmp_data_31_0_shift)
+
+/*define for CMP_DATA_HI word*/
+/*define for cmp_data_63_32 field*/
+#define SDMA_PKT_ATOMIC_CMP_DATA_HI_cmp_data_63_32_offset 6
+#define SDMA_PKT_ATOMIC_CMP_DATA_HI_cmp_data_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_ATOMIC_CMP_DATA_HI_cmp_data_63_32_shift  0
+#define SDMA_PKT_ATOMIC_CMP_DATA_HI_CMP_DATA_63_32(x) (((x) & SDMA_PKT_ATOMIC_CMP_DATA_HI_cmp_data_63_32_mask) << SDMA_PKT_ATOMIC_CMP_DATA_HI_cmp_data_63_32_shift)
+
+/*define for LOOP_INTERVAL word*/
+/*define for loop_interval field*/
+#define SDMA_PKT_ATOMIC_LOOP_INTERVAL_loop_interval_offset 7
+#define SDMA_PKT_ATOMIC_LOOP_INTERVAL_loop_interval_mask   0x00001FFF
+#define SDMA_PKT_ATOMIC_LOOP_INTERVAL_loop_interval_shift  0
+#define SDMA_PKT_ATOMIC_LOOP_INTERVAL_LOOP_INTERVAL(x) (((x) & SDMA_PKT_ATOMIC_LOOP_INTERVAL_loop_interval_mask) << SDMA_PKT_ATOMIC_LOOP_INTERVAL_loop_interval_shift)
+
+
+/*
+** Definitions for SDMA_PKT_TIMESTAMP_SET packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_TIMESTAMP_SET_HEADER_op_offset 0
+#define SDMA_PKT_TIMESTAMP_SET_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_TIMESTAMP_SET_HEADER_op_shift  0
+#define SDMA_PKT_TIMESTAMP_SET_HEADER_OP(x) (((x) & SDMA_PKT_TIMESTAMP_SET_HEADER_op_mask) << SDMA_PKT_TIMESTAMP_SET_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_TIMESTAMP_SET_HEADER_sub_op_offset 0
+#define SDMA_PKT_TIMESTAMP_SET_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_TIMESTAMP_SET_HEADER_sub_op_shift  8
+#define SDMA_PKT_TIMESTAMP_SET_HEADER_SUB_OP(x) (((x) & SDMA_PKT_TIMESTAMP_SET_HEADER_sub_op_mask) << SDMA_PKT_TIMESTAMP_SET_HEADER_sub_op_shift)
+
+/*define for INIT_DATA_LO word*/
+/*define for init_data_31_0 field*/
+#define SDMA_PKT_TIMESTAMP_SET_INIT_DATA_LO_init_data_31_0_offset 1
+#define SDMA_PKT_TIMESTAMP_SET_INIT_DATA_LO_init_data_31_0_mask   0xFFFFFFFF
+#define SDMA_PKT_TIMESTAMP_SET_INIT_DATA_LO_init_data_31_0_shift  0
+#define SDMA_PKT_TIMESTAMP_SET_INIT_DATA_LO_INIT_DATA_31_0(x) (((x) & SDMA_PKT_TIMESTAMP_SET_INIT_DATA_LO_init_data_31_0_mask) << SDMA_PKT_TIMESTAMP_SET_INIT_DATA_LO_init_data_31_0_shift)
+
+/*define for INIT_DATA_HI word*/
+/*define for init_data_63_32 field*/
+#define SDMA_PKT_TIMESTAMP_SET_INIT_DATA_HI_init_data_63_32_offset 2
+#define SDMA_PKT_TIMESTAMP_SET_INIT_DATA_HI_init_data_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_TIMESTAMP_SET_INIT_DATA_HI_init_data_63_32_shift  0
+#define SDMA_PKT_TIMESTAMP_SET_INIT_DATA_HI_INIT_DATA_63_32(x) (((x) & SDMA_PKT_TIMESTAMP_SET_INIT_DATA_HI_init_data_63_32_mask) << SDMA_PKT_TIMESTAMP_SET_INIT_DATA_HI_init_data_63_32_shift)
+
+
+/*
+** Definitions for SDMA_PKT_TIMESTAMP_GET packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_TIMESTAMP_GET_HEADER_op_offset 0
+#define SDMA_PKT_TIMESTAMP_GET_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_TIMESTAMP_GET_HEADER_op_shift  0
+#define SDMA_PKT_TIMESTAMP_GET_HEADER_OP(x) (((x) & SDMA_PKT_TIMESTAMP_GET_HEADER_op_mask) << SDMA_PKT_TIMESTAMP_GET_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_TIMESTAMP_GET_HEADER_sub_op_offset 0
+#define SDMA_PKT_TIMESTAMP_GET_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_TIMESTAMP_GET_HEADER_sub_op_shift  8
+#define SDMA_PKT_TIMESTAMP_GET_HEADER_SUB_OP(x) (((x) & SDMA_PKT_TIMESTAMP_GET_HEADER_sub_op_mask) << SDMA_PKT_TIMESTAMP_GET_HEADER_sub_op_shift)
+
+/*define for l2_policy field*/
+#define SDMA_PKT_TIMESTAMP_GET_HEADER_l2_policy_offset 0
+#define SDMA_PKT_TIMESTAMP_GET_HEADER_l2_policy_mask   0x00000003
+#define SDMA_PKT_TIMESTAMP_GET_HEADER_l2_policy_shift  24
+#define SDMA_PKT_TIMESTAMP_GET_HEADER_L2_POLICY(x) (((x) & SDMA_PKT_TIMESTAMP_GET_HEADER_l2_policy_mask) << SDMA_PKT_TIMESTAMP_GET_HEADER_l2_policy_shift)
+
+/*define for llc_policy field*/
+#define SDMA_PKT_TIMESTAMP_GET_HEADER_llc_policy_offset 0
+#define SDMA_PKT_TIMESTAMP_GET_HEADER_llc_policy_mask   0x00000001
+#define SDMA_PKT_TIMESTAMP_GET_HEADER_llc_policy_shift  26
+#define SDMA_PKT_TIMESTAMP_GET_HEADER_LLC_POLICY(x) (((x) & SDMA_PKT_TIMESTAMP_GET_HEADER_llc_policy_mask) << SDMA_PKT_TIMESTAMP_GET_HEADER_llc_policy_shift)
+
+/*define for cpv field*/
+#define SDMA_PKT_TIMESTAMP_GET_HEADER_cpv_offset 0
+#define SDMA_PKT_TIMESTAMP_GET_HEADER_cpv_mask   0x00000001
+#define SDMA_PKT_TIMESTAMP_GET_HEADER_cpv_shift  28
+#define SDMA_PKT_TIMESTAMP_GET_HEADER_CPV(x) (((x) & SDMA_PKT_TIMESTAMP_GET_HEADER_cpv_mask) << SDMA_PKT_TIMESTAMP_GET_HEADER_cpv_shift)
+
+/*define for WRITE_ADDR_LO word*/
+/*define for write_addr_31_3 field*/
+#define SDMA_PKT_TIMESTAMP_GET_WRITE_ADDR_LO_write_addr_31_3_offset 1
+#define SDMA_PKT_TIMESTAMP_GET_WRITE_ADDR_LO_write_addr_31_3_mask   0x1FFFFFFF
+#define SDMA_PKT_TIMESTAMP_GET_WRITE_ADDR_LO_write_addr_31_3_shift  3
+#define SDMA_PKT_TIMESTAMP_GET_WRITE_ADDR_LO_WRITE_ADDR_31_3(x) (((x) & SDMA_PKT_TIMESTAMP_GET_WRITE_ADDR_LO_write_addr_31_3_mask) << SDMA_PKT_TIMESTAMP_GET_WRITE_ADDR_LO_write_addr_31_3_shift)
+
+/*define for WRITE_ADDR_HI word*/
+/*define for write_addr_63_32 field*/
+#define SDMA_PKT_TIMESTAMP_GET_WRITE_ADDR_HI_write_addr_63_32_offset 2
+#define SDMA_PKT_TIMESTAMP_GET_WRITE_ADDR_HI_write_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_TIMESTAMP_GET_WRITE_ADDR_HI_write_addr_63_32_shift  0
+#define SDMA_PKT_TIMESTAMP_GET_WRITE_ADDR_HI_WRITE_ADDR_63_32(x) (((x) & SDMA_PKT_TIMESTAMP_GET_WRITE_ADDR_HI_write_addr_63_32_mask) << SDMA_PKT_TIMESTAMP_GET_WRITE_ADDR_HI_write_addr_63_32_shift)
+
+
+/*
+** Definitions for SDMA_PKT_TIMESTAMP_GET_GLOBAL packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_TIMESTAMP_GET_GLOBAL_HEADER_op_offset 0
+#define SDMA_PKT_TIMESTAMP_GET_GLOBAL_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_TIMESTAMP_GET_GLOBAL_HEADER_op_shift  0
+#define SDMA_PKT_TIMESTAMP_GET_GLOBAL_HEADER_OP(x) (((x) & SDMA_PKT_TIMESTAMP_GET_GLOBAL_HEADER_op_mask) << SDMA_PKT_TIMESTAMP_GET_GLOBAL_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_TIMESTAMP_GET_GLOBAL_HEADER_sub_op_offset 0
+#define SDMA_PKT_TIMESTAMP_GET_GLOBAL_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_TIMESTAMP_GET_GLOBAL_HEADER_sub_op_shift  8
+#define SDMA_PKT_TIMESTAMP_GET_GLOBAL_HEADER_SUB_OP(x) (((x) & SDMA_PKT_TIMESTAMP_GET_GLOBAL_HEADER_sub_op_mask) << SDMA_PKT_TIMESTAMP_GET_GLOBAL_HEADER_sub_op_shift)
+
+/*define for l2_policy field*/
+#define SDMA_PKT_TIMESTAMP_GET_GLOBAL_HEADER_l2_policy_offset 0
+#define SDMA_PKT_TIMESTAMP_GET_GLOBAL_HEADER_l2_policy_mask   0x00000003
+#define SDMA_PKT_TIMESTAMP_GET_GLOBAL_HEADER_l2_policy_shift  24
+#define SDMA_PKT_TIMESTAMP_GET_GLOBAL_HEADER_L2_POLICY(x) (((x) & SDMA_PKT_TIMESTAMP_GET_GLOBAL_HEADER_l2_policy_mask) << SDMA_PKT_TIMESTAMP_GET_GLOBAL_HEADER_l2_policy_shift)
+
+/*define for llc_policy field*/
+#define SDMA_PKT_TIMESTAMP_GET_GLOBAL_HEADER_llc_policy_offset 0
+#define SDMA_PKT_TIMESTAMP_GET_GLOBAL_HEADER_llc_policy_mask   0x00000001
+#define SDMA_PKT_TIMESTAMP_GET_GLOBAL_HEADER_llc_policy_shift  26
+#define SDMA_PKT_TIMESTAMP_GET_GLOBAL_HEADER_LLC_POLICY(x) (((x) & SDMA_PKT_TIMESTAMP_GET_GLOBAL_HEADER_llc_policy_mask) << SDMA_PKT_TIMESTAMP_GET_GLOBAL_HEADER_llc_policy_shift)
+
+/*define for cpv field*/
+#define SDMA_PKT_TIMESTAMP_GET_GLOBAL_HEADER_cpv_offset 0
+#define SDMA_PKT_TIMESTAMP_GET_GLOBAL_HEADER_cpv_mask   0x00000001
+#define SDMA_PKT_TIMESTAMP_GET_GLOBAL_HEADER_cpv_shift  28
+#define SDMA_PKT_TIMESTAMP_GET_GLOBAL_HEADER_CPV(x) (((x) & SDMA_PKT_TIMESTAMP_GET_GLOBAL_HEADER_cpv_mask) << SDMA_PKT_TIMESTAMP_GET_GLOBAL_HEADER_cpv_shift)
+
+/*define for WRITE_ADDR_LO word*/
+/*define for write_addr_31_3 field*/
+#define SDMA_PKT_TIMESTAMP_GET_GLOBAL_WRITE_ADDR_LO_write_addr_31_3_offset 1
+#define SDMA_PKT_TIMESTAMP_GET_GLOBAL_WRITE_ADDR_LO_write_addr_31_3_mask   0x1FFFFFFF
+#define SDMA_PKT_TIMESTAMP_GET_GLOBAL_WRITE_ADDR_LO_write_addr_31_3_shift  3
+#define SDMA_PKT_TIMESTAMP_GET_GLOBAL_WRITE_ADDR_LO_WRITE_ADDR_31_3(x) (((x) & SDMA_PKT_TIMESTAMP_GET_GLOBAL_WRITE_ADDR_LO_write_addr_31_3_mask) << SDMA_PKT_TIMESTAMP_GET_GLOBAL_WRITE_ADDR_LO_write_addr_31_3_shift)
+
+/*define for WRITE_ADDR_HI word*/
+/*define for write_addr_63_32 field*/
+#define SDMA_PKT_TIMESTAMP_GET_GLOBAL_WRITE_ADDR_HI_write_addr_63_32_offset 2
+#define SDMA_PKT_TIMESTAMP_GET_GLOBAL_WRITE_ADDR_HI_write_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_PKT_TIMESTAMP_GET_GLOBAL_WRITE_ADDR_HI_write_addr_63_32_shift  0
+#define SDMA_PKT_TIMESTAMP_GET_GLOBAL_WRITE_ADDR_HI_WRITE_ADDR_63_32(x) (((x) & SDMA_PKT_TIMESTAMP_GET_GLOBAL_WRITE_ADDR_HI_write_addr_63_32_mask) << SDMA_PKT_TIMESTAMP_GET_GLOBAL_WRITE_ADDR_HI_write_addr_63_32_shift)
+
+
+/*
+** Definitions for SDMA_PKT_TRAP packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_TRAP_HEADER_op_offset 0
+#define SDMA_PKT_TRAP_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_TRAP_HEADER_op_shift  0
+#define SDMA_PKT_TRAP_HEADER_OP(x) (((x) & SDMA_PKT_TRAP_HEADER_op_mask) << SDMA_PKT_TRAP_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_TRAP_HEADER_sub_op_offset 0
+#define SDMA_PKT_TRAP_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_TRAP_HEADER_sub_op_shift  8
+#define SDMA_PKT_TRAP_HEADER_SUB_OP(x) (((x) & SDMA_PKT_TRAP_HEADER_sub_op_mask) << SDMA_PKT_TRAP_HEADER_sub_op_shift)
+
+/*define for INT_CONTEXT word*/
+/*define for int_context field*/
+#define SDMA_PKT_TRAP_INT_CONTEXT_int_context_offset 1
+#define SDMA_PKT_TRAP_INT_CONTEXT_int_context_mask   0x0FFFFFFF
+#define SDMA_PKT_TRAP_INT_CONTEXT_int_context_shift  0
+#define SDMA_PKT_TRAP_INT_CONTEXT_INT_CONTEXT(x) (((x) & SDMA_PKT_TRAP_INT_CONTEXT_int_context_mask) << SDMA_PKT_TRAP_INT_CONTEXT_int_context_shift)
+
+
+/*
+** Definitions for SDMA_PKT_DUMMY_TRAP packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_DUMMY_TRAP_HEADER_op_offset 0
+#define SDMA_PKT_DUMMY_TRAP_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_DUMMY_TRAP_HEADER_op_shift  0
+#define SDMA_PKT_DUMMY_TRAP_HEADER_OP(x) (((x) & SDMA_PKT_DUMMY_TRAP_HEADER_op_mask) << SDMA_PKT_DUMMY_TRAP_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_DUMMY_TRAP_HEADER_sub_op_offset 0
+#define SDMA_PKT_DUMMY_TRAP_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_DUMMY_TRAP_HEADER_sub_op_shift  8
+#define SDMA_PKT_DUMMY_TRAP_HEADER_SUB_OP(x) (((x) & SDMA_PKT_DUMMY_TRAP_HEADER_sub_op_mask) << SDMA_PKT_DUMMY_TRAP_HEADER_sub_op_shift)
+
+/*define for INT_CONTEXT word*/
+/*define for int_context field*/
+#define SDMA_PKT_DUMMY_TRAP_INT_CONTEXT_int_context_offset 1
+#define SDMA_PKT_DUMMY_TRAP_INT_CONTEXT_int_context_mask   0x0FFFFFFF
+#define SDMA_PKT_DUMMY_TRAP_INT_CONTEXT_int_context_shift  0
+#define SDMA_PKT_DUMMY_TRAP_INT_CONTEXT_INT_CONTEXT(x) (((x) & SDMA_PKT_DUMMY_TRAP_INT_CONTEXT_int_context_mask) << SDMA_PKT_DUMMY_TRAP_INT_CONTEXT_int_context_shift)
+
+
+/*
+** Definitions for SDMA_PKT_GPUVM_INV packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_GPUVM_INV_HEADER_op_offset 0
+#define SDMA_PKT_GPUVM_INV_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_GPUVM_INV_HEADER_op_shift  0
+#define SDMA_PKT_GPUVM_INV_HEADER_OP(x) (((x) & SDMA_PKT_GPUVM_INV_HEADER_op_mask) << SDMA_PKT_GPUVM_INV_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_GPUVM_INV_HEADER_sub_op_offset 0
+#define SDMA_PKT_GPUVM_INV_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_GPUVM_INV_HEADER_sub_op_shift  8
+#define SDMA_PKT_GPUVM_INV_HEADER_SUB_OP(x) (((x) & SDMA_PKT_GPUVM_INV_HEADER_sub_op_mask) << SDMA_PKT_GPUVM_INV_HEADER_sub_op_shift)
+
+/*define for PAYLOAD1 word*/
+/*define for per_vmid_inv_req field*/
+#define SDMA_PKT_GPUVM_INV_PAYLOAD1_per_vmid_inv_req_offset 1
+#define SDMA_PKT_GPUVM_INV_PAYLOAD1_per_vmid_inv_req_mask   0x0000FFFF
+#define SDMA_PKT_GPUVM_INV_PAYLOAD1_per_vmid_inv_req_shift  0
+#define SDMA_PKT_GPUVM_INV_PAYLOAD1_PER_VMID_INV_REQ(x) (((x) & SDMA_PKT_GPUVM_INV_PAYLOAD1_per_vmid_inv_req_mask) << SDMA_PKT_GPUVM_INV_PAYLOAD1_per_vmid_inv_req_shift)
+
+/*define for flush_type field*/
+#define SDMA_PKT_GPUVM_INV_PAYLOAD1_flush_type_offset 1
+#define SDMA_PKT_GPUVM_INV_PAYLOAD1_flush_type_mask   0x00000007
+#define SDMA_PKT_GPUVM_INV_PAYLOAD1_flush_type_shift  16
+#define SDMA_PKT_GPUVM_INV_PAYLOAD1_FLUSH_TYPE(x) (((x) & SDMA_PKT_GPUVM_INV_PAYLOAD1_flush_type_mask) << SDMA_PKT_GPUVM_INV_PAYLOAD1_flush_type_shift)
+
+/*define for l2_ptes field*/
+#define SDMA_PKT_GPUVM_INV_PAYLOAD1_l2_ptes_offset 1
+#define SDMA_PKT_GPUVM_INV_PAYLOAD1_l2_ptes_mask   0x00000001
+#define SDMA_PKT_GPUVM_INV_PAYLOAD1_l2_ptes_shift  19
+#define SDMA_PKT_GPUVM_INV_PAYLOAD1_L2_PTES(x) (((x) & SDMA_PKT_GPUVM_INV_PAYLOAD1_l2_ptes_mask) << SDMA_PKT_GPUVM_INV_PAYLOAD1_l2_ptes_shift)
+
+/*define for l2_pde0 field*/
+#define SDMA_PKT_GPUVM_INV_PAYLOAD1_l2_pde0_offset 1
+#define SDMA_PKT_GPUVM_INV_PAYLOAD1_l2_pde0_mask   0x00000001
+#define SDMA_PKT_GPUVM_INV_PAYLOAD1_l2_pde0_shift  20
+#define SDMA_PKT_GPUVM_INV_PAYLOAD1_L2_PDE0(x) (((x) & SDMA_PKT_GPUVM_INV_PAYLOAD1_l2_pde0_mask) << SDMA_PKT_GPUVM_INV_PAYLOAD1_l2_pde0_shift)
+
+/*define for l2_pde1 field*/
+#define SDMA_PKT_GPUVM_INV_PAYLOAD1_l2_pde1_offset 1
+#define SDMA_PKT_GPUVM_INV_PAYLOAD1_l2_pde1_mask   0x00000001
+#define SDMA_PKT_GPUVM_INV_PAYLOAD1_l2_pde1_shift  21
+#define SDMA_PKT_GPUVM_INV_PAYLOAD1_L2_PDE1(x) (((x) & SDMA_PKT_GPUVM_INV_PAYLOAD1_l2_pde1_mask) << SDMA_PKT_GPUVM_INV_PAYLOAD1_l2_pde1_shift)
+
+/*define for l2_pde2 field*/
+#define SDMA_PKT_GPUVM_INV_PAYLOAD1_l2_pde2_offset 1
+#define SDMA_PKT_GPUVM_INV_PAYLOAD1_l2_pde2_mask   0x00000001
+#define SDMA_PKT_GPUVM_INV_PAYLOAD1_l2_pde2_shift  22
+#define SDMA_PKT_GPUVM_INV_PAYLOAD1_L2_PDE2(x) (((x) & SDMA_PKT_GPUVM_INV_PAYLOAD1_l2_pde2_mask) << SDMA_PKT_GPUVM_INV_PAYLOAD1_l2_pde2_shift)
+
+/*define for l1_ptes field*/
+#define SDMA_PKT_GPUVM_INV_PAYLOAD1_l1_ptes_offset 1
+#define SDMA_PKT_GPUVM_INV_PAYLOAD1_l1_ptes_mask   0x00000001
+#define SDMA_PKT_GPUVM_INV_PAYLOAD1_l1_ptes_shift  23
+#define SDMA_PKT_GPUVM_INV_PAYLOAD1_L1_PTES(x) (((x) & SDMA_PKT_GPUVM_INV_PAYLOAD1_l1_ptes_mask) << SDMA_PKT_GPUVM_INV_PAYLOAD1_l1_ptes_shift)
+
+/*define for clr_protection_fault_status_addr field*/
+#define SDMA_PKT_GPUVM_INV_PAYLOAD1_clr_protection_fault_status_addr_offset 1
+#define SDMA_PKT_GPUVM_INV_PAYLOAD1_clr_protection_fault_status_addr_mask   0x00000001
+#define SDMA_PKT_GPUVM_INV_PAYLOAD1_clr_protection_fault_status_addr_shift  24
+#define SDMA_PKT_GPUVM_INV_PAYLOAD1_CLR_PROTECTION_FAULT_STATUS_ADDR(x) (((x) & SDMA_PKT_GPUVM_INV_PAYLOAD1_clr_protection_fault_status_addr_mask) << SDMA_PKT_GPUVM_INV_PAYLOAD1_clr_protection_fault_status_addr_shift)
+
+/*define for log_request field*/
+#define SDMA_PKT_GPUVM_INV_PAYLOAD1_log_request_offset 1
+#define SDMA_PKT_GPUVM_INV_PAYLOAD1_log_request_mask   0x00000001
+#define SDMA_PKT_GPUVM_INV_PAYLOAD1_log_request_shift  25
+#define SDMA_PKT_GPUVM_INV_PAYLOAD1_LOG_REQUEST(x) (((x) & SDMA_PKT_GPUVM_INV_PAYLOAD1_log_request_mask) << SDMA_PKT_GPUVM_INV_PAYLOAD1_log_request_shift)
+
+/*define for four_kilobytes field*/
+#define SDMA_PKT_GPUVM_INV_PAYLOAD1_four_kilobytes_offset 1
+#define SDMA_PKT_GPUVM_INV_PAYLOAD1_four_kilobytes_mask   0x00000001
+#define SDMA_PKT_GPUVM_INV_PAYLOAD1_four_kilobytes_shift  26
+#define SDMA_PKT_GPUVM_INV_PAYLOAD1_FOUR_KILOBYTES(x) (((x) & SDMA_PKT_GPUVM_INV_PAYLOAD1_four_kilobytes_mask) << SDMA_PKT_GPUVM_INV_PAYLOAD1_four_kilobytes_shift)
+
+/*define for PAYLOAD2 word*/
+/*define for s field*/
+#define SDMA_PKT_GPUVM_INV_PAYLOAD2_s_offset 2
+#define SDMA_PKT_GPUVM_INV_PAYLOAD2_s_mask   0x00000001
+#define SDMA_PKT_GPUVM_INV_PAYLOAD2_s_shift  0
+#define SDMA_PKT_GPUVM_INV_PAYLOAD2_S(x) (((x) & SDMA_PKT_GPUVM_INV_PAYLOAD2_s_mask) << SDMA_PKT_GPUVM_INV_PAYLOAD2_s_shift)
+
+/*define for page_va_42_12 field*/
+#define SDMA_PKT_GPUVM_INV_PAYLOAD2_page_va_42_12_offset 2
+#define SDMA_PKT_GPUVM_INV_PAYLOAD2_page_va_42_12_mask   0x7FFFFFFF
+#define SDMA_PKT_GPUVM_INV_PAYLOAD2_page_va_42_12_shift  1
+#define SDMA_PKT_GPUVM_INV_PAYLOAD2_PAGE_VA_42_12(x) (((x) & SDMA_PKT_GPUVM_INV_PAYLOAD2_page_va_42_12_mask) << SDMA_PKT_GPUVM_INV_PAYLOAD2_page_va_42_12_shift)
+
+/*define for PAYLOAD3 word*/
+/*define for page_va_47_43 field*/
+#define SDMA_PKT_GPUVM_INV_PAYLOAD3_page_va_47_43_offset 3
+#define SDMA_PKT_GPUVM_INV_PAYLOAD3_page_va_47_43_mask   0x0000003F
+#define SDMA_PKT_GPUVM_INV_PAYLOAD3_page_va_47_43_shift  0
+#define SDMA_PKT_GPUVM_INV_PAYLOAD3_PAGE_VA_47_43(x) (((x) & SDMA_PKT_GPUVM_INV_PAYLOAD3_page_va_47_43_mask) << SDMA_PKT_GPUVM_INV_PAYLOAD3_page_va_47_43_shift)
+
+
+/*
+** Definitions for SDMA_PKT_GCR_REQ packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_GCR_REQ_HEADER_op_offset 0
+#define SDMA_PKT_GCR_REQ_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_GCR_REQ_HEADER_op_shift  0
+#define SDMA_PKT_GCR_REQ_HEADER_OP(x) (((x) & SDMA_PKT_GCR_REQ_HEADER_op_mask) << SDMA_PKT_GCR_REQ_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_GCR_REQ_HEADER_sub_op_offset 0
+#define SDMA_PKT_GCR_REQ_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_GCR_REQ_HEADER_sub_op_shift  8
+#define SDMA_PKT_GCR_REQ_HEADER_SUB_OP(x) (((x) & SDMA_PKT_GCR_REQ_HEADER_sub_op_mask) << SDMA_PKT_GCR_REQ_HEADER_sub_op_shift)
+
+/*define for PAYLOAD1 word*/
+/*define for base_va_31_7 field*/
+#define SDMA_PKT_GCR_REQ_PAYLOAD1_base_va_31_7_offset 1
+#define SDMA_PKT_GCR_REQ_PAYLOAD1_base_va_31_7_mask   0x01FFFFFF
+#define SDMA_PKT_GCR_REQ_PAYLOAD1_base_va_31_7_shift  7
+#define SDMA_PKT_GCR_REQ_PAYLOAD1_BASE_VA_31_7(x) (((x) & SDMA_PKT_GCR_REQ_PAYLOAD1_base_va_31_7_mask) << SDMA_PKT_GCR_REQ_PAYLOAD1_base_va_31_7_shift)
+
+/*define for PAYLOAD2 word*/
+/*define for base_va_47_32 field*/
+#define SDMA_PKT_GCR_REQ_PAYLOAD2_base_va_47_32_offset 2
+#define SDMA_PKT_GCR_REQ_PAYLOAD2_base_va_47_32_mask   0x0000FFFF
+#define SDMA_PKT_GCR_REQ_PAYLOAD2_base_va_47_32_shift  0
+#define SDMA_PKT_GCR_REQ_PAYLOAD2_BASE_VA_47_32(x) (((x) & SDMA_PKT_GCR_REQ_PAYLOAD2_base_va_47_32_mask) << SDMA_PKT_GCR_REQ_PAYLOAD2_base_va_47_32_shift)
+
+/*define for gcr_control_15_0 field*/
+#define SDMA_PKT_GCR_REQ_PAYLOAD2_gcr_control_15_0_offset 2
+#define SDMA_PKT_GCR_REQ_PAYLOAD2_gcr_control_15_0_mask   0x0000FFFF
+#define SDMA_PKT_GCR_REQ_PAYLOAD2_gcr_control_15_0_shift  16
+#define SDMA_PKT_GCR_REQ_PAYLOAD2_GCR_CONTROL_15_0(x) (((x) & SDMA_PKT_GCR_REQ_PAYLOAD2_gcr_control_15_0_mask) << SDMA_PKT_GCR_REQ_PAYLOAD2_gcr_control_15_0_shift)
+
+/*define for PAYLOAD3 word*/
+/*define for gcr_control_18_16 field*/
+#define SDMA_PKT_GCR_REQ_PAYLOAD3_gcr_control_18_16_offset 3
+#define SDMA_PKT_GCR_REQ_PAYLOAD3_gcr_control_18_16_mask   0x00000007
+#define SDMA_PKT_GCR_REQ_PAYLOAD3_gcr_control_18_16_shift  0
+#define SDMA_PKT_GCR_REQ_PAYLOAD3_GCR_CONTROL_18_16(x) (((x) & SDMA_PKT_GCR_REQ_PAYLOAD3_gcr_control_18_16_mask) << SDMA_PKT_GCR_REQ_PAYLOAD3_gcr_control_18_16_shift)
+
+/*define for limit_va_31_7 field*/
+#define SDMA_PKT_GCR_REQ_PAYLOAD3_limit_va_31_7_offset 3
+#define SDMA_PKT_GCR_REQ_PAYLOAD3_limit_va_31_7_mask   0x01FFFFFF
+#define SDMA_PKT_GCR_REQ_PAYLOAD3_limit_va_31_7_shift  7
+#define SDMA_PKT_GCR_REQ_PAYLOAD3_LIMIT_VA_31_7(x) (((x) & SDMA_PKT_GCR_REQ_PAYLOAD3_limit_va_31_7_mask) << SDMA_PKT_GCR_REQ_PAYLOAD3_limit_va_31_7_shift)
+
+/*define for PAYLOAD4 word*/
+/*define for limit_va_47_32 field*/
+#define SDMA_PKT_GCR_REQ_PAYLOAD4_limit_va_47_32_offset 4
+#define SDMA_PKT_GCR_REQ_PAYLOAD4_limit_va_47_32_mask   0x0000FFFF
+#define SDMA_PKT_GCR_REQ_PAYLOAD4_limit_va_47_32_shift  0
+#define SDMA_PKT_GCR_REQ_PAYLOAD4_LIMIT_VA_47_32(x) (((x) & SDMA_PKT_GCR_REQ_PAYLOAD4_limit_va_47_32_mask) << SDMA_PKT_GCR_REQ_PAYLOAD4_limit_va_47_32_shift)
+
+/*define for vmid field*/
+#define SDMA_PKT_GCR_REQ_PAYLOAD4_vmid_offset 4
+#define SDMA_PKT_GCR_REQ_PAYLOAD4_vmid_mask   0x0000000F
+#define SDMA_PKT_GCR_REQ_PAYLOAD4_vmid_shift  24
+#define SDMA_PKT_GCR_REQ_PAYLOAD4_VMID(x) (((x) & SDMA_PKT_GCR_REQ_PAYLOAD4_vmid_mask) << SDMA_PKT_GCR_REQ_PAYLOAD4_vmid_shift)
+
+
+/*
+** Definitions for SDMA_PKT_NOP packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_NOP_HEADER_op_offset 0
+#define SDMA_PKT_NOP_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_NOP_HEADER_op_shift  0
+#define SDMA_PKT_NOP_HEADER_OP(x) (((x) & SDMA_PKT_NOP_HEADER_op_mask) << SDMA_PKT_NOP_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_NOP_HEADER_sub_op_offset 0
+#define SDMA_PKT_NOP_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_NOP_HEADER_sub_op_shift  8
+#define SDMA_PKT_NOP_HEADER_SUB_OP(x) (((x) & SDMA_PKT_NOP_HEADER_sub_op_mask) << SDMA_PKT_NOP_HEADER_sub_op_shift)
+
+/*define for count field*/
+#define SDMA_PKT_NOP_HEADER_count_offset 0
+#define SDMA_PKT_NOP_HEADER_count_mask   0x00003FFF
+#define SDMA_PKT_NOP_HEADER_count_shift  16
+#define SDMA_PKT_NOP_HEADER_COUNT(x) (((x) & SDMA_PKT_NOP_HEADER_count_mask) << SDMA_PKT_NOP_HEADER_count_shift)
+
+/*define for DATA0 word*/
+/*define for data0 field*/
+#define SDMA_PKT_NOP_DATA0_data0_offset 1
+#define SDMA_PKT_NOP_DATA0_data0_mask   0xFFFFFFFF
+#define SDMA_PKT_NOP_DATA0_data0_shift  0
+#define SDMA_PKT_NOP_DATA0_DATA0(x) (((x) & SDMA_PKT_NOP_DATA0_data0_mask) << SDMA_PKT_NOP_DATA0_data0_shift)
+
+
+/*
+** Definitions for SDMA_AQL_PKT_HEADER packet
+*/
+
+/*define for HEADER word*/
+/*define for format field*/
+#define SDMA_AQL_PKT_HEADER_HEADER_format_offset 0
+#define SDMA_AQL_PKT_HEADER_HEADER_format_mask   0x000000FF
+#define SDMA_AQL_PKT_HEADER_HEADER_format_shift  0
+#define SDMA_AQL_PKT_HEADER_HEADER_FORMAT(x) (((x) & SDMA_AQL_PKT_HEADER_HEADER_format_mask) << SDMA_AQL_PKT_HEADER_HEADER_format_shift)
+
+/*define for barrier field*/
+#define SDMA_AQL_PKT_HEADER_HEADER_barrier_offset 0
+#define SDMA_AQL_PKT_HEADER_HEADER_barrier_mask   0x00000001
+#define SDMA_AQL_PKT_HEADER_HEADER_barrier_shift  8
+#define SDMA_AQL_PKT_HEADER_HEADER_BARRIER(x) (((x) & SDMA_AQL_PKT_HEADER_HEADER_barrier_mask) << SDMA_AQL_PKT_HEADER_HEADER_barrier_shift)
+
+/*define for acquire_fence_scope field*/
+#define SDMA_AQL_PKT_HEADER_HEADER_acquire_fence_scope_offset 0
+#define SDMA_AQL_PKT_HEADER_HEADER_acquire_fence_scope_mask   0x00000003
+#define SDMA_AQL_PKT_HEADER_HEADER_acquire_fence_scope_shift  9
+#define SDMA_AQL_PKT_HEADER_HEADER_ACQUIRE_FENCE_SCOPE(x) (((x) & SDMA_AQL_PKT_HEADER_HEADER_acquire_fence_scope_mask) << SDMA_AQL_PKT_HEADER_HEADER_acquire_fence_scope_shift)
+
+/*define for release_fence_scope field*/
+#define SDMA_AQL_PKT_HEADER_HEADER_release_fence_scope_offset 0
+#define SDMA_AQL_PKT_HEADER_HEADER_release_fence_scope_mask   0x00000003
+#define SDMA_AQL_PKT_HEADER_HEADER_release_fence_scope_shift  11
+#define SDMA_AQL_PKT_HEADER_HEADER_RELEASE_FENCE_SCOPE(x) (((x) & SDMA_AQL_PKT_HEADER_HEADER_release_fence_scope_mask) << SDMA_AQL_PKT_HEADER_HEADER_release_fence_scope_shift)
+
+/*define for reserved field*/
+#define SDMA_AQL_PKT_HEADER_HEADER_reserved_offset 0
+#define SDMA_AQL_PKT_HEADER_HEADER_reserved_mask   0x00000007
+#define SDMA_AQL_PKT_HEADER_HEADER_reserved_shift  13
+#define SDMA_AQL_PKT_HEADER_HEADER_RESERVED(x) (((x) & SDMA_AQL_PKT_HEADER_HEADER_reserved_mask) << SDMA_AQL_PKT_HEADER_HEADER_reserved_shift)
+
+/*define for op field*/
+#define SDMA_AQL_PKT_HEADER_HEADER_op_offset 0
+#define SDMA_AQL_PKT_HEADER_HEADER_op_mask   0x0000000F
+#define SDMA_AQL_PKT_HEADER_HEADER_op_shift  16
+#define SDMA_AQL_PKT_HEADER_HEADER_OP(x) (((x) & SDMA_AQL_PKT_HEADER_HEADER_op_mask) << SDMA_AQL_PKT_HEADER_HEADER_op_shift)
+
+/*define for subop field*/
+#define SDMA_AQL_PKT_HEADER_HEADER_subop_offset 0
+#define SDMA_AQL_PKT_HEADER_HEADER_subop_mask   0x00000007
+#define SDMA_AQL_PKT_HEADER_HEADER_subop_shift  20
+#define SDMA_AQL_PKT_HEADER_HEADER_SUBOP(x) (((x) & SDMA_AQL_PKT_HEADER_HEADER_subop_mask) << SDMA_AQL_PKT_HEADER_HEADER_subop_shift)
+
+/*define for cpv field*/
+#define SDMA_AQL_PKT_HEADER_HEADER_cpv_offset 0
+#define SDMA_AQL_PKT_HEADER_HEADER_cpv_mask   0x00000001
+#define SDMA_AQL_PKT_HEADER_HEADER_cpv_shift  28
+#define SDMA_AQL_PKT_HEADER_HEADER_CPV(x) (((x) & SDMA_AQL_PKT_HEADER_HEADER_cpv_mask) << SDMA_AQL_PKT_HEADER_HEADER_cpv_shift)
+
+
+/*
+** Definitions for SDMA_AQL_PKT_COPY_LINEAR packet
+*/
+
+/*define for HEADER word*/
+/*define for format field*/
+#define SDMA_AQL_PKT_COPY_LINEAR_HEADER_format_offset 0
+#define SDMA_AQL_PKT_COPY_LINEAR_HEADER_format_mask   0x000000FF
+#define SDMA_AQL_PKT_COPY_LINEAR_HEADER_format_shift  0
+#define SDMA_AQL_PKT_COPY_LINEAR_HEADER_FORMAT(x) (((x) & SDMA_AQL_PKT_COPY_LINEAR_HEADER_format_mask) << SDMA_AQL_PKT_COPY_LINEAR_HEADER_format_shift)
+
+/*define for barrier field*/
+#define SDMA_AQL_PKT_COPY_LINEAR_HEADER_barrier_offset 0
+#define SDMA_AQL_PKT_COPY_LINEAR_HEADER_barrier_mask   0x00000001
+#define SDMA_AQL_PKT_COPY_LINEAR_HEADER_barrier_shift  8
+#define SDMA_AQL_PKT_COPY_LINEAR_HEADER_BARRIER(x) (((x) & SDMA_AQL_PKT_COPY_LINEAR_HEADER_barrier_mask) << SDMA_AQL_PKT_COPY_LINEAR_HEADER_barrier_shift)
+
+/*define for acquire_fence_scope field*/
+#define SDMA_AQL_PKT_COPY_LINEAR_HEADER_acquire_fence_scope_offset 0
+#define SDMA_AQL_PKT_COPY_LINEAR_HEADER_acquire_fence_scope_mask   0x00000003
+#define SDMA_AQL_PKT_COPY_LINEAR_HEADER_acquire_fence_scope_shift  9
+#define SDMA_AQL_PKT_COPY_LINEAR_HEADER_ACQUIRE_FENCE_SCOPE(x) (((x) & SDMA_AQL_PKT_COPY_LINEAR_HEADER_acquire_fence_scope_mask) << SDMA_AQL_PKT_COPY_LINEAR_HEADER_acquire_fence_scope_shift)
+
+/*define for release_fence_scope field*/
+#define SDMA_AQL_PKT_COPY_LINEAR_HEADER_release_fence_scope_offset 0
+#define SDMA_AQL_PKT_COPY_LINEAR_HEADER_release_fence_scope_mask   0x00000003
+#define SDMA_AQL_PKT_COPY_LINEAR_HEADER_release_fence_scope_shift  11
+#define SDMA_AQL_PKT_COPY_LINEAR_HEADER_RELEASE_FENCE_SCOPE(x) (((x) & SDMA_AQL_PKT_COPY_LINEAR_HEADER_release_fence_scope_mask) << SDMA_AQL_PKT_COPY_LINEAR_HEADER_release_fence_scope_shift)
+
+/*define for reserved field*/
+#define SDMA_AQL_PKT_COPY_LINEAR_HEADER_reserved_offset 0
+#define SDMA_AQL_PKT_COPY_LINEAR_HEADER_reserved_mask   0x00000007
+#define SDMA_AQL_PKT_COPY_LINEAR_HEADER_reserved_shift  13
+#define SDMA_AQL_PKT_COPY_LINEAR_HEADER_RESERVED(x) (((x) & SDMA_AQL_PKT_COPY_LINEAR_HEADER_reserved_mask) << SDMA_AQL_PKT_COPY_LINEAR_HEADER_reserved_shift)
+
+/*define for op field*/
+#define SDMA_AQL_PKT_COPY_LINEAR_HEADER_op_offset 0
+#define SDMA_AQL_PKT_COPY_LINEAR_HEADER_op_mask   0x0000000F
+#define SDMA_AQL_PKT_COPY_LINEAR_HEADER_op_shift  16
+#define SDMA_AQL_PKT_COPY_LINEAR_HEADER_OP(x) (((x) & SDMA_AQL_PKT_COPY_LINEAR_HEADER_op_mask) << SDMA_AQL_PKT_COPY_LINEAR_HEADER_op_shift)
+
+/*define for subop field*/
+#define SDMA_AQL_PKT_COPY_LINEAR_HEADER_subop_offset 0
+#define SDMA_AQL_PKT_COPY_LINEAR_HEADER_subop_mask   0x00000007
+#define SDMA_AQL_PKT_COPY_LINEAR_HEADER_subop_shift  20
+#define SDMA_AQL_PKT_COPY_LINEAR_HEADER_SUBOP(x) (((x) & SDMA_AQL_PKT_COPY_LINEAR_HEADER_subop_mask) << SDMA_AQL_PKT_COPY_LINEAR_HEADER_subop_shift)
+
+/*define for cpv field*/
+#define SDMA_AQL_PKT_COPY_LINEAR_HEADER_cpv_offset 0
+#define SDMA_AQL_PKT_COPY_LINEAR_HEADER_cpv_mask   0x00000001
+#define SDMA_AQL_PKT_COPY_LINEAR_HEADER_cpv_shift  28
+#define SDMA_AQL_PKT_COPY_LINEAR_HEADER_CPV(x) (((x) & SDMA_AQL_PKT_COPY_LINEAR_HEADER_cpv_mask) << SDMA_AQL_PKT_COPY_LINEAR_HEADER_cpv_shift)
+
+/*define for RESERVED_DW1 word*/
+/*define for reserved_dw1 field*/
+#define SDMA_AQL_PKT_COPY_LINEAR_RESERVED_DW1_reserved_dw1_offset 1
+#define SDMA_AQL_PKT_COPY_LINEAR_RESERVED_DW1_reserved_dw1_mask   0xFFFFFFFF
+#define SDMA_AQL_PKT_COPY_LINEAR_RESERVED_DW1_reserved_dw1_shift  0
+#define SDMA_AQL_PKT_COPY_LINEAR_RESERVED_DW1_RESERVED_DW1(x) (((x) & SDMA_AQL_PKT_COPY_LINEAR_RESERVED_DW1_reserved_dw1_mask) << SDMA_AQL_PKT_COPY_LINEAR_RESERVED_DW1_reserved_dw1_shift)
+
+/*define for RETURN_ADDR_LO word*/
+/*define for return_addr_31_0 field*/
+#define SDMA_AQL_PKT_COPY_LINEAR_RETURN_ADDR_LO_return_addr_31_0_offset 2
+#define SDMA_AQL_PKT_COPY_LINEAR_RETURN_ADDR_LO_return_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_AQL_PKT_COPY_LINEAR_RETURN_ADDR_LO_return_addr_31_0_shift  0
+#define SDMA_AQL_PKT_COPY_LINEAR_RETURN_ADDR_LO_RETURN_ADDR_31_0(x) (((x) & SDMA_AQL_PKT_COPY_LINEAR_RETURN_ADDR_LO_return_addr_31_0_mask) << SDMA_AQL_PKT_COPY_LINEAR_RETURN_ADDR_LO_return_addr_31_0_shift)
+
+/*define for RETURN_ADDR_HI word*/
+/*define for return_addr_63_32 field*/
+#define SDMA_AQL_PKT_COPY_LINEAR_RETURN_ADDR_HI_return_addr_63_32_offset 3
+#define SDMA_AQL_PKT_COPY_LINEAR_RETURN_ADDR_HI_return_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_AQL_PKT_COPY_LINEAR_RETURN_ADDR_HI_return_addr_63_32_shift  0
+#define SDMA_AQL_PKT_COPY_LINEAR_RETURN_ADDR_HI_RETURN_ADDR_63_32(x) (((x) & SDMA_AQL_PKT_COPY_LINEAR_RETURN_ADDR_HI_return_addr_63_32_mask) << SDMA_AQL_PKT_COPY_LINEAR_RETURN_ADDR_HI_return_addr_63_32_shift)
+
+/*define for COUNT word*/
+/*define for count field*/
+#define SDMA_AQL_PKT_COPY_LINEAR_COUNT_count_offset 4
+#define SDMA_AQL_PKT_COPY_LINEAR_COUNT_count_mask   0x003FFFFF
+#define SDMA_AQL_PKT_COPY_LINEAR_COUNT_count_shift  0
+#define SDMA_AQL_PKT_COPY_LINEAR_COUNT_COUNT(x) (((x) & SDMA_AQL_PKT_COPY_LINEAR_COUNT_count_mask) << SDMA_AQL_PKT_COPY_LINEAR_COUNT_count_shift)
+
+/*define for PARAMETER word*/
+/*define for dst_sw field*/
+#define SDMA_AQL_PKT_COPY_LINEAR_PARAMETER_dst_sw_offset 5
+#define SDMA_AQL_PKT_COPY_LINEAR_PARAMETER_dst_sw_mask   0x00000003
+#define SDMA_AQL_PKT_COPY_LINEAR_PARAMETER_dst_sw_shift  16
+#define SDMA_AQL_PKT_COPY_LINEAR_PARAMETER_DST_SW(x) (((x) & SDMA_AQL_PKT_COPY_LINEAR_PARAMETER_dst_sw_mask) << SDMA_AQL_PKT_COPY_LINEAR_PARAMETER_dst_sw_shift)
+
+/*define for dst_cache_policy field*/
+#define SDMA_AQL_PKT_COPY_LINEAR_PARAMETER_dst_cache_policy_offset 5
+#define SDMA_AQL_PKT_COPY_LINEAR_PARAMETER_dst_cache_policy_mask   0x00000007
+#define SDMA_AQL_PKT_COPY_LINEAR_PARAMETER_dst_cache_policy_shift  18
+#define SDMA_AQL_PKT_COPY_LINEAR_PARAMETER_DST_CACHE_POLICY(x) (((x) & SDMA_AQL_PKT_COPY_LINEAR_PARAMETER_dst_cache_policy_mask) << SDMA_AQL_PKT_COPY_LINEAR_PARAMETER_dst_cache_policy_shift)
+
+/*define for src_sw field*/
+#define SDMA_AQL_PKT_COPY_LINEAR_PARAMETER_src_sw_offset 5
+#define SDMA_AQL_PKT_COPY_LINEAR_PARAMETER_src_sw_mask   0x00000003
+#define SDMA_AQL_PKT_COPY_LINEAR_PARAMETER_src_sw_shift  24
+#define SDMA_AQL_PKT_COPY_LINEAR_PARAMETER_SRC_SW(x) (((x) & SDMA_AQL_PKT_COPY_LINEAR_PARAMETER_src_sw_mask) << SDMA_AQL_PKT_COPY_LINEAR_PARAMETER_src_sw_shift)
+
+/*define for src_cache_policy field*/
+#define SDMA_AQL_PKT_COPY_LINEAR_PARAMETER_src_cache_policy_offset 5
+#define SDMA_AQL_PKT_COPY_LINEAR_PARAMETER_src_cache_policy_mask   0x00000007
+#define SDMA_AQL_PKT_COPY_LINEAR_PARAMETER_src_cache_policy_shift  26
+#define SDMA_AQL_PKT_COPY_LINEAR_PARAMETER_SRC_CACHE_POLICY(x) (((x) & SDMA_AQL_PKT_COPY_LINEAR_PARAMETER_src_cache_policy_mask) << SDMA_AQL_PKT_COPY_LINEAR_PARAMETER_src_cache_policy_shift)
+
+/*define for SRC_ADDR_LO word*/
+/*define for src_addr_31_0 field*/
+#define SDMA_AQL_PKT_COPY_LINEAR_SRC_ADDR_LO_src_addr_31_0_offset 6
+#define SDMA_AQL_PKT_COPY_LINEAR_SRC_ADDR_LO_src_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_AQL_PKT_COPY_LINEAR_SRC_ADDR_LO_src_addr_31_0_shift  0
+#define SDMA_AQL_PKT_COPY_LINEAR_SRC_ADDR_LO_SRC_ADDR_31_0(x) (((x) & SDMA_AQL_PKT_COPY_LINEAR_SRC_ADDR_LO_src_addr_31_0_mask) << SDMA_AQL_PKT_COPY_LINEAR_SRC_ADDR_LO_src_addr_31_0_shift)
+
+/*define for SRC_ADDR_HI word*/
+/*define for src_addr_63_32 field*/
+#define SDMA_AQL_PKT_COPY_LINEAR_SRC_ADDR_HI_src_addr_63_32_offset 7
+#define SDMA_AQL_PKT_COPY_LINEAR_SRC_ADDR_HI_src_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_AQL_PKT_COPY_LINEAR_SRC_ADDR_HI_src_addr_63_32_shift  0
+#define SDMA_AQL_PKT_COPY_LINEAR_SRC_ADDR_HI_SRC_ADDR_63_32(x) (((x) & SDMA_AQL_PKT_COPY_LINEAR_SRC_ADDR_HI_src_addr_63_32_mask) << SDMA_AQL_PKT_COPY_LINEAR_SRC_ADDR_HI_src_addr_63_32_shift)
+
+/*define for DST_ADDR_LO word*/
+/*define for dst_addr_31_0 field*/
+#define SDMA_AQL_PKT_COPY_LINEAR_DST_ADDR_LO_dst_addr_31_0_offset 8
+#define SDMA_AQL_PKT_COPY_LINEAR_DST_ADDR_LO_dst_addr_31_0_mask   0xFFFFFFFF
+#define SDMA_AQL_PKT_COPY_LINEAR_DST_ADDR_LO_dst_addr_31_0_shift  0
+#define SDMA_AQL_PKT_COPY_LINEAR_DST_ADDR_LO_DST_ADDR_31_0(x) (((x) & SDMA_AQL_PKT_COPY_LINEAR_DST_ADDR_LO_dst_addr_31_0_mask) << SDMA_AQL_PKT_COPY_LINEAR_DST_ADDR_LO_dst_addr_31_0_shift)
+
+/*define for DST_ADDR_HI word*/
+/*define for dst_addr_63_32 field*/
+#define SDMA_AQL_PKT_COPY_LINEAR_DST_ADDR_HI_dst_addr_63_32_offset 9
+#define SDMA_AQL_PKT_COPY_LINEAR_DST_ADDR_HI_dst_addr_63_32_mask   0xFFFFFFFF
+#define SDMA_AQL_PKT_COPY_LINEAR_DST_ADDR_HI_dst_addr_63_32_shift  0
+#define SDMA_AQL_PKT_COPY_LINEAR_DST_ADDR_HI_DST_ADDR_63_32(x) (((x) & SDMA_AQL_PKT_COPY_LINEAR_DST_ADDR_HI_dst_addr_63_32_mask) << SDMA_AQL_PKT_COPY_LINEAR_DST_ADDR_HI_dst_addr_63_32_shift)
+
+/*define for RESERVED_DW10 word*/
+/*define for reserved_dw10 field*/
+#define SDMA_AQL_PKT_COPY_LINEAR_RESERVED_DW10_reserved_dw10_offset 10
+#define SDMA_AQL_PKT_COPY_LINEAR_RESERVED_DW10_reserved_dw10_mask   0xFFFFFFFF
+#define SDMA_AQL_PKT_COPY_LINEAR_RESERVED_DW10_reserved_dw10_shift  0
+#define SDMA_AQL_PKT_COPY_LINEAR_RESERVED_DW10_RESERVED_DW10(x) (((x) & SDMA_AQL_PKT_COPY_LINEAR_RESERVED_DW10_reserved_dw10_mask) << SDMA_AQL_PKT_COPY_LINEAR_RESERVED_DW10_reserved_dw10_shift)
+
+/*define for RESERVED_DW11 word*/
+/*define for reserved_dw11 field*/
+#define SDMA_AQL_PKT_COPY_LINEAR_RESERVED_DW11_reserved_dw11_offset 11
+#define SDMA_AQL_PKT_COPY_LINEAR_RESERVED_DW11_reserved_dw11_mask   0xFFFFFFFF
+#define SDMA_AQL_PKT_COPY_LINEAR_RESERVED_DW11_reserved_dw11_shift  0
+#define SDMA_AQL_PKT_COPY_LINEAR_RESERVED_DW11_RESERVED_DW11(x) (((x) & SDMA_AQL_PKT_COPY_LINEAR_RESERVED_DW11_reserved_dw11_mask) << SDMA_AQL_PKT_COPY_LINEAR_RESERVED_DW11_reserved_dw11_shift)
+
+/*define for RESERVED_DW12 word*/
+/*define for reserved_dw12 field*/
+#define SDMA_AQL_PKT_COPY_LINEAR_RESERVED_DW12_reserved_dw12_offset 12
+#define SDMA_AQL_PKT_COPY_LINEAR_RESERVED_DW12_reserved_dw12_mask   0xFFFFFFFF
+#define SDMA_AQL_PKT_COPY_LINEAR_RESERVED_DW12_reserved_dw12_shift  0
+#define SDMA_AQL_PKT_COPY_LINEAR_RESERVED_DW12_RESERVED_DW12(x) (((x) & SDMA_AQL_PKT_COPY_LINEAR_RESERVED_DW12_reserved_dw12_mask) << SDMA_AQL_PKT_COPY_LINEAR_RESERVED_DW12_reserved_dw12_shift)
+
+/*define for RESERVED_DW13 word*/
+/*define for reserved_dw13 field*/
+#define SDMA_AQL_PKT_COPY_LINEAR_RESERVED_DW13_reserved_dw13_offset 13
+#define SDMA_AQL_PKT_COPY_LINEAR_RESERVED_DW13_reserved_dw13_mask   0xFFFFFFFF
+#define SDMA_AQL_PKT_COPY_LINEAR_RESERVED_DW13_reserved_dw13_shift  0
+#define SDMA_AQL_PKT_COPY_LINEAR_RESERVED_DW13_RESERVED_DW13(x) (((x) & SDMA_AQL_PKT_COPY_LINEAR_RESERVED_DW13_reserved_dw13_mask) << SDMA_AQL_PKT_COPY_LINEAR_RESERVED_DW13_reserved_dw13_shift)
+
+/*define for COMPLETION_SIGNAL_LO word*/
+/*define for completion_signal_31_0 field*/
+#define SDMA_AQL_PKT_COPY_LINEAR_COMPLETION_SIGNAL_LO_completion_signal_31_0_offset 14
+#define SDMA_AQL_PKT_COPY_LINEAR_COMPLETION_SIGNAL_LO_completion_signal_31_0_mask   0xFFFFFFFF
+#define SDMA_AQL_PKT_COPY_LINEAR_COMPLETION_SIGNAL_LO_completion_signal_31_0_shift  0
+#define SDMA_AQL_PKT_COPY_LINEAR_COMPLETION_SIGNAL_LO_COMPLETION_SIGNAL_31_0(x) (((x) & SDMA_AQL_PKT_COPY_LINEAR_COMPLETION_SIGNAL_LO_completion_signal_31_0_mask) << SDMA_AQL_PKT_COPY_LINEAR_COMPLETION_SIGNAL_LO_completion_signal_31_0_shift)
+
+/*define for COMPLETION_SIGNAL_HI word*/
+/*define for completion_signal_63_32 field*/
+#define SDMA_AQL_PKT_COPY_LINEAR_COMPLETION_SIGNAL_HI_completion_signal_63_32_offset 15
+#define SDMA_AQL_PKT_COPY_LINEAR_COMPLETION_SIGNAL_HI_completion_signal_63_32_mask   0xFFFFFFFF
+#define SDMA_AQL_PKT_COPY_LINEAR_COMPLETION_SIGNAL_HI_completion_signal_63_32_shift  0
+#define SDMA_AQL_PKT_COPY_LINEAR_COMPLETION_SIGNAL_HI_COMPLETION_SIGNAL_63_32(x) (((x) & SDMA_AQL_PKT_COPY_LINEAR_COMPLETION_SIGNAL_HI_completion_signal_63_32_mask) << SDMA_AQL_PKT_COPY_LINEAR_COMPLETION_SIGNAL_HI_completion_signal_63_32_shift)
+
+
+/*
+** Definitions for SDMA_AQL_PKT_BARRIER_OR packet
+*/
+
+/*define for HEADER word*/
+/*define for format field*/
+#define SDMA_AQL_PKT_BARRIER_OR_HEADER_format_offset 0
+#define SDMA_AQL_PKT_BARRIER_OR_HEADER_format_mask   0x000000FF
+#define SDMA_AQL_PKT_BARRIER_OR_HEADER_format_shift  0
+#define SDMA_AQL_PKT_BARRIER_OR_HEADER_FORMAT(x) (((x) & SDMA_AQL_PKT_BARRIER_OR_HEADER_format_mask) << SDMA_AQL_PKT_BARRIER_OR_HEADER_format_shift)
+
+/*define for barrier field*/
+#define SDMA_AQL_PKT_BARRIER_OR_HEADER_barrier_offset 0
+#define SDMA_AQL_PKT_BARRIER_OR_HEADER_barrier_mask   0x00000001
+#define SDMA_AQL_PKT_BARRIER_OR_HEADER_barrier_shift  8
+#define SDMA_AQL_PKT_BARRIER_OR_HEADER_BARRIER(x) (((x) & SDMA_AQL_PKT_BARRIER_OR_HEADER_barrier_mask) << SDMA_AQL_PKT_BARRIER_OR_HEADER_barrier_shift)
+
+/*define for acquire_fence_scope field*/
+#define SDMA_AQL_PKT_BARRIER_OR_HEADER_acquire_fence_scope_offset 0
+#define SDMA_AQL_PKT_BARRIER_OR_HEADER_acquire_fence_scope_mask   0x00000003
+#define SDMA_AQL_PKT_BARRIER_OR_HEADER_acquire_fence_scope_shift  9
+#define SDMA_AQL_PKT_BARRIER_OR_HEADER_ACQUIRE_FENCE_SCOPE(x) (((x) & SDMA_AQL_PKT_BARRIER_OR_HEADER_acquire_fence_scope_mask) << SDMA_AQL_PKT_BARRIER_OR_HEADER_acquire_fence_scope_shift)
+
+/*define for release_fence_scope field*/
+#define SDMA_AQL_PKT_BARRIER_OR_HEADER_release_fence_scope_offset 0
+#define SDMA_AQL_PKT_BARRIER_OR_HEADER_release_fence_scope_mask   0x00000003
+#define SDMA_AQL_PKT_BARRIER_OR_HEADER_release_fence_scope_shift  11
+#define SDMA_AQL_PKT_BARRIER_OR_HEADER_RELEASE_FENCE_SCOPE(x) (((x) & SDMA_AQL_PKT_BARRIER_OR_HEADER_release_fence_scope_mask) << SDMA_AQL_PKT_BARRIER_OR_HEADER_release_fence_scope_shift)
+
+/*define for reserved field*/
+#define SDMA_AQL_PKT_BARRIER_OR_HEADER_reserved_offset 0
+#define SDMA_AQL_PKT_BARRIER_OR_HEADER_reserved_mask   0x00000007
+#define SDMA_AQL_PKT_BARRIER_OR_HEADER_reserved_shift  13
+#define SDMA_AQL_PKT_BARRIER_OR_HEADER_RESERVED(x) (((x) & SDMA_AQL_PKT_BARRIER_OR_HEADER_reserved_mask) << SDMA_AQL_PKT_BARRIER_OR_HEADER_reserved_shift)
+
+/*define for op field*/
+#define SDMA_AQL_PKT_BARRIER_OR_HEADER_op_offset 0
+#define SDMA_AQL_PKT_BARRIER_OR_HEADER_op_mask   0x0000000F
+#define SDMA_AQL_PKT_BARRIER_OR_HEADER_op_shift  16
+#define SDMA_AQL_PKT_BARRIER_OR_HEADER_OP(x) (((x) & SDMA_AQL_PKT_BARRIER_OR_HEADER_op_mask) << SDMA_AQL_PKT_BARRIER_OR_HEADER_op_shift)
+
+/*define for subop field*/
+#define SDMA_AQL_PKT_BARRIER_OR_HEADER_subop_offset 0
+#define SDMA_AQL_PKT_BARRIER_OR_HEADER_subop_mask   0x00000007
+#define SDMA_AQL_PKT_BARRIER_OR_HEADER_subop_shift  20
+#define SDMA_AQL_PKT_BARRIER_OR_HEADER_SUBOP(x) (((x) & SDMA_AQL_PKT_BARRIER_OR_HEADER_subop_mask) << SDMA_AQL_PKT_BARRIER_OR_HEADER_subop_shift)
+
+/*define for cpv field*/
+#define SDMA_AQL_PKT_BARRIER_OR_HEADER_cpv_offset 0
+#define SDMA_AQL_PKT_BARRIER_OR_HEADER_cpv_mask   0x00000001
+#define SDMA_AQL_PKT_BARRIER_OR_HEADER_cpv_shift  28
+#define SDMA_AQL_PKT_BARRIER_OR_HEADER_CPV(x) (((x) & SDMA_AQL_PKT_BARRIER_OR_HEADER_cpv_mask) << SDMA_AQL_PKT_BARRIER_OR_HEADER_cpv_shift)
+
+/*define for RESERVED_DW1 word*/
+/*define for reserved_dw1 field*/
+#define SDMA_AQL_PKT_BARRIER_OR_RESERVED_DW1_reserved_dw1_offset 1
+#define SDMA_AQL_PKT_BARRIER_OR_RESERVED_DW1_reserved_dw1_mask   0xFFFFFFFF
+#define SDMA_AQL_PKT_BARRIER_OR_RESERVED_DW1_reserved_dw1_shift  0
+#define SDMA_AQL_PKT_BARRIER_OR_RESERVED_DW1_RESERVED_DW1(x) (((x) & SDMA_AQL_PKT_BARRIER_OR_RESERVED_DW1_reserved_dw1_mask) << SDMA_AQL_PKT_BARRIER_OR_RESERVED_DW1_reserved_dw1_shift)
+
+/*define for DEPENDENT_ADDR_0_LO word*/
+/*define for dependent_addr_0_31_0 field*/
+#define SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_0_LO_dependent_addr_0_31_0_offset 2
+#define SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_0_LO_dependent_addr_0_31_0_mask   0xFFFFFFFF
+#define SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_0_LO_dependent_addr_0_31_0_shift  0
+#define SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_0_LO_DEPENDENT_ADDR_0_31_0(x) (((x) & SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_0_LO_dependent_addr_0_31_0_mask) << SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_0_LO_dependent_addr_0_31_0_shift)
+
+/*define for DEPENDENT_ADDR_0_HI word*/
+/*define for dependent_addr_0_63_32 field*/
+#define SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_0_HI_dependent_addr_0_63_32_offset 3
+#define SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_0_HI_dependent_addr_0_63_32_mask   0xFFFFFFFF
+#define SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_0_HI_dependent_addr_0_63_32_shift  0
+#define SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_0_HI_DEPENDENT_ADDR_0_63_32(x) (((x) & SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_0_HI_dependent_addr_0_63_32_mask) << SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_0_HI_dependent_addr_0_63_32_shift)
+
+/*define for DEPENDENT_ADDR_1_LO word*/
+/*define for dependent_addr_1_31_0 field*/
+#define SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_1_LO_dependent_addr_1_31_0_offset 4
+#define SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_1_LO_dependent_addr_1_31_0_mask   0xFFFFFFFF
+#define SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_1_LO_dependent_addr_1_31_0_shift  0
+#define SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_1_LO_DEPENDENT_ADDR_1_31_0(x) (((x) & SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_1_LO_dependent_addr_1_31_0_mask) << SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_1_LO_dependent_addr_1_31_0_shift)
+
+/*define for DEPENDENT_ADDR_1_HI word*/
+/*define for dependent_addr_1_63_32 field*/
+#define SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_1_HI_dependent_addr_1_63_32_offset 5
+#define SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_1_HI_dependent_addr_1_63_32_mask   0xFFFFFFFF
+#define SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_1_HI_dependent_addr_1_63_32_shift  0
+#define SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_1_HI_DEPENDENT_ADDR_1_63_32(x) (((x) & SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_1_HI_dependent_addr_1_63_32_mask) << SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_1_HI_dependent_addr_1_63_32_shift)
+
+/*define for DEPENDENT_ADDR_2_LO word*/
+/*define for dependent_addr_2_31_0 field*/
+#define SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_2_LO_dependent_addr_2_31_0_offset 6
+#define SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_2_LO_dependent_addr_2_31_0_mask   0xFFFFFFFF
+#define SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_2_LO_dependent_addr_2_31_0_shift  0
+#define SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_2_LO_DEPENDENT_ADDR_2_31_0(x) (((x) & SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_2_LO_dependent_addr_2_31_0_mask) << SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_2_LO_dependent_addr_2_31_0_shift)
+
+/*define for DEPENDENT_ADDR_2_HI word*/
+/*define for dependent_addr_2_63_32 field*/
+#define SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_2_HI_dependent_addr_2_63_32_offset 7
+#define SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_2_HI_dependent_addr_2_63_32_mask   0xFFFFFFFF
+#define SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_2_HI_dependent_addr_2_63_32_shift  0
+#define SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_2_HI_DEPENDENT_ADDR_2_63_32(x) (((x) & SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_2_HI_dependent_addr_2_63_32_mask) << SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_2_HI_dependent_addr_2_63_32_shift)
+
+/*define for DEPENDENT_ADDR_3_LO word*/
+/*define for dependent_addr_3_31_0 field*/
+#define SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_3_LO_dependent_addr_3_31_0_offset 8
+#define SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_3_LO_dependent_addr_3_31_0_mask   0xFFFFFFFF
+#define SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_3_LO_dependent_addr_3_31_0_shift  0
+#define SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_3_LO_DEPENDENT_ADDR_3_31_0(x) (((x) & SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_3_LO_dependent_addr_3_31_0_mask) << SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_3_LO_dependent_addr_3_31_0_shift)
+
+/*define for DEPENDENT_ADDR_3_HI word*/
+/*define for dependent_addr_3_63_32 field*/
+#define SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_3_HI_dependent_addr_3_63_32_offset 9
+#define SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_3_HI_dependent_addr_3_63_32_mask   0xFFFFFFFF
+#define SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_3_HI_dependent_addr_3_63_32_shift  0
+#define SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_3_HI_DEPENDENT_ADDR_3_63_32(x) (((x) & SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_3_HI_dependent_addr_3_63_32_mask) << SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_3_HI_dependent_addr_3_63_32_shift)
+
+/*define for DEPENDENT_ADDR_4_LO word*/
+/*define for dependent_addr_4_31_0 field*/
+#define SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_4_LO_dependent_addr_4_31_0_offset 10
+#define SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_4_LO_dependent_addr_4_31_0_mask   0xFFFFFFFF
+#define SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_4_LO_dependent_addr_4_31_0_shift  0
+#define SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_4_LO_DEPENDENT_ADDR_4_31_0(x) (((x) & SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_4_LO_dependent_addr_4_31_0_mask) << SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_4_LO_dependent_addr_4_31_0_shift)
+
+/*define for DEPENDENT_ADDR_4_HI word*/
+/*define for dependent_addr_4_63_32 field*/
+#define SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_4_HI_dependent_addr_4_63_32_offset 11
+#define SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_4_HI_dependent_addr_4_63_32_mask   0xFFFFFFFF
+#define SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_4_HI_dependent_addr_4_63_32_shift  0
+#define SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_4_HI_DEPENDENT_ADDR_4_63_32(x) (((x) & SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_4_HI_dependent_addr_4_63_32_mask) << SDMA_AQL_PKT_BARRIER_OR_DEPENDENT_ADDR_4_HI_dependent_addr_4_63_32_shift)
+
+/*define for CACHE_POLICY word*/
+/*define for cache_policy0 field*/
+#define SDMA_AQL_PKT_BARRIER_OR_CACHE_POLICY_cache_policy0_offset 12
+#define SDMA_AQL_PKT_BARRIER_OR_CACHE_POLICY_cache_policy0_mask   0x00000007
+#define SDMA_AQL_PKT_BARRIER_OR_CACHE_POLICY_cache_policy0_shift  0
+#define SDMA_AQL_PKT_BARRIER_OR_CACHE_POLICY_CACHE_POLICY0(x) (((x) & SDMA_AQL_PKT_BARRIER_OR_CACHE_POLICY_cache_policy0_mask) << SDMA_AQL_PKT_BARRIER_OR_CACHE_POLICY_cache_policy0_shift)
+
+/*define for cache_policy1 field*/
+#define SDMA_AQL_PKT_BARRIER_OR_CACHE_POLICY_cache_policy1_offset 12
+#define SDMA_AQL_PKT_BARRIER_OR_CACHE_POLICY_cache_policy1_mask   0x00000007
+#define SDMA_AQL_PKT_BARRIER_OR_CACHE_POLICY_cache_policy1_shift  5
+#define SDMA_AQL_PKT_BARRIER_OR_CACHE_POLICY_CACHE_POLICY1(x) (((x) & SDMA_AQL_PKT_BARRIER_OR_CACHE_POLICY_cache_policy1_mask) << SDMA_AQL_PKT_BARRIER_OR_CACHE_POLICY_cache_policy1_shift)
+
+/*define for cache_policy2 field*/
+#define SDMA_AQL_PKT_BARRIER_OR_CACHE_POLICY_cache_policy2_offset 12
+#define SDMA_AQL_PKT_BARRIER_OR_CACHE_POLICY_cache_policy2_mask   0x00000007
+#define SDMA_AQL_PKT_BARRIER_OR_CACHE_POLICY_cache_policy2_shift  10
+#define SDMA_AQL_PKT_BARRIER_OR_CACHE_POLICY_CACHE_POLICY2(x) (((x) & SDMA_AQL_PKT_BARRIER_OR_CACHE_POLICY_cache_policy2_mask) << SDMA_AQL_PKT_BARRIER_OR_CACHE_POLICY_cache_policy2_shift)
+
+/*define for cache_policy3 field*/
+#define SDMA_AQL_PKT_BARRIER_OR_CACHE_POLICY_cache_policy3_offset 12
+#define SDMA_AQL_PKT_BARRIER_OR_CACHE_POLICY_cache_policy3_mask   0x00000007
+#define SDMA_AQL_PKT_BARRIER_OR_CACHE_POLICY_cache_policy3_shift  15
+#define SDMA_AQL_PKT_BARRIER_OR_CACHE_POLICY_CACHE_POLICY3(x) (((x) & SDMA_AQL_PKT_BARRIER_OR_CACHE_POLICY_cache_policy3_mask) << SDMA_AQL_PKT_BARRIER_OR_CACHE_POLICY_cache_policy3_shift)
+
+/*define for cache_policy4 field*/
+#define SDMA_AQL_PKT_BARRIER_OR_CACHE_POLICY_cache_policy4_offset 12
+#define SDMA_AQL_PKT_BARRIER_OR_CACHE_POLICY_cache_policy4_mask   0x00000007
+#define SDMA_AQL_PKT_BARRIER_OR_CACHE_POLICY_cache_policy4_shift  20
+#define SDMA_AQL_PKT_BARRIER_OR_CACHE_POLICY_CACHE_POLICY4(x) (((x) & SDMA_AQL_PKT_BARRIER_OR_CACHE_POLICY_cache_policy4_mask) << SDMA_AQL_PKT_BARRIER_OR_CACHE_POLICY_cache_policy4_shift)
+
+/*define for RESERVED_DW13 word*/
+/*define for reserved_dw13 field*/
+#define SDMA_AQL_PKT_BARRIER_OR_RESERVED_DW13_reserved_dw13_offset 13
+#define SDMA_AQL_PKT_BARRIER_OR_RESERVED_DW13_reserved_dw13_mask   0xFFFFFFFF
+#define SDMA_AQL_PKT_BARRIER_OR_RESERVED_DW13_reserved_dw13_shift  0
+#define SDMA_AQL_PKT_BARRIER_OR_RESERVED_DW13_RESERVED_DW13(x) (((x) & SDMA_AQL_PKT_BARRIER_OR_RESERVED_DW13_reserved_dw13_mask) << SDMA_AQL_PKT_BARRIER_OR_RESERVED_DW13_reserved_dw13_shift)
+
+/*define for COMPLETION_SIGNAL_LO word*/
+/*define for completion_signal_31_0 field*/
+#define SDMA_AQL_PKT_BARRIER_OR_COMPLETION_SIGNAL_LO_completion_signal_31_0_offset 14
+#define SDMA_AQL_PKT_BARRIER_OR_COMPLETION_SIGNAL_LO_completion_signal_31_0_mask   0xFFFFFFFF
+#define SDMA_AQL_PKT_BARRIER_OR_COMPLETION_SIGNAL_LO_completion_signal_31_0_shift  0
+#define SDMA_AQL_PKT_BARRIER_OR_COMPLETION_SIGNAL_LO_COMPLETION_SIGNAL_31_0(x) (((x) & SDMA_AQL_PKT_BARRIER_OR_COMPLETION_SIGNAL_LO_completion_signal_31_0_mask) << SDMA_AQL_PKT_BARRIER_OR_COMPLETION_SIGNAL_LO_completion_signal_31_0_shift)
+
+/*define for COMPLETION_SIGNAL_HI word*/
+/*define for completion_signal_63_32 field*/
+#define SDMA_AQL_PKT_BARRIER_OR_COMPLETION_SIGNAL_HI_completion_signal_63_32_offset 15
+#define SDMA_AQL_PKT_BARRIER_OR_COMPLETION_SIGNAL_HI_completion_signal_63_32_mask   0xFFFFFFFF
+#define SDMA_AQL_PKT_BARRIER_OR_COMPLETION_SIGNAL_HI_completion_signal_63_32_shift  0
+#define SDMA_AQL_PKT_BARRIER_OR_COMPLETION_SIGNAL_HI_COMPLETION_SIGNAL_63_32(x) (((x) & SDMA_AQL_PKT_BARRIER_OR_COMPLETION_SIGNAL_HI_completion_signal_63_32_mask) << SDMA_AQL_PKT_BARRIER_OR_COMPLETION_SIGNAL_HI_completion_signal_63_32_shift)
+
+
+#endif /* __SDMA_V6_0_0_PKT_OPEN_H_ */
diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v7_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v7_0.c
new file mode 100644
index 000000000000..2b81344dcd66
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v7_0.c
@@ -0,0 +1,1859 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/delay.h>
+#include <linux/firmware.h>
+#include <linux/module.h>
+#include <linux/pci.h>
+
+#include "amdgpu.h"
+#include "amdgpu_ucode.h"
+#include "amdgpu_trace.h"
+
+#include "gc/gc_12_0_0_offset.h"
+#include "gc/gc_12_0_0_sh_mask.h"
+#include "hdp/hdp_6_0_0_offset.h"
+#include "ivsrcid/gfx/irqsrcs_gfx_12_0_0.h"
+
+#include "soc15_common.h"
+#include "soc15.h"
+#include "sdma_v6_0_0_pkt_open.h"
+#include "nbio_v4_3.h"
+#include "sdma_common.h"
+#include "sdma_v7_0.h"
+#include "v12_structs.h"
+#include "mes_userqueue.h"
+#include "amdgpu_userq_fence.h"
+
+MODULE_FIRMWARE("amdgpu/sdma_7_0_0.bin");
+MODULE_FIRMWARE("amdgpu/sdma_7_0_1.bin");
+
+#define SDMA1_REG_OFFSET 0x600
+#define SDMA0_HYP_DEC_REG_START 0x5880
+#define SDMA0_HYP_DEC_REG_END 0x589a
+#define SDMA1_HYP_DEC_REG_OFFSET 0x20
+
+/*define for compression field for sdma7*/
+#define SDMA_PKT_CONSTANT_FILL_HEADER_compress_offset 0
+#define SDMA_PKT_CONSTANT_FILL_HEADER_compress_mask   0x00000001
+#define SDMA_PKT_CONSTANT_FILL_HEADER_compress_shift  16
+#define SDMA_PKT_CONSTANT_FILL_HEADER_COMPRESS(x) (((x) & SDMA_PKT_CONSTANT_FILL_HEADER_compress_mask) << SDMA_PKT_CONSTANT_FILL_HEADER_compress_shift)
+
+static const struct amdgpu_hwip_reg_entry sdma_reg_list_7_0[] = {
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_STATUS_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_STATUS1_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_STATUS2_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_STATUS3_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_STATUS4_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_STATUS5_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_STATUS6_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_UCODE_REV),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_RB_RPTR_FETCH_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_RB_RPTR_FETCH),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_UTCL1_RD_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_UTCL1_WR_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_UTCL1_RD_XNACK0),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_UTCL1_RD_XNACK1),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_UTCL1_WR_XNACK0),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_UTCL1_WR_XNACK1),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE0_RB_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE0_RB_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE0_RB_RPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE0_RB_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE0_RB_WPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE0_IB_OFFSET),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE0_IB_BASE_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE0_IB_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE0_IB_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE0_IB_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE0_IB_SUB_REMAIN),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE0_DUMMY_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE_STATUS0),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE1_RB_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE1_RB_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE1_RB_RPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE1_RB_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE1_RB_WPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE1_IB_OFFSET),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE1_IB_BASE_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE1_IB_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE1_IB_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE1_IB_SUB_REMAIN),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE1_DUMMY_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE2_RB_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE2_RB_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE2_RB_RPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE2_RB_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE2_RB_WPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE2_IB_OFFSET),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE2_IB_BASE_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE2_IB_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE2_IB_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE2_IB_SUB_REMAIN),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_QUEUE2_DUMMY_REG),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_INT_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_VM_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, regGRBM_STATUS2),
+	SOC15_REG_ENTRY_STR(GC, 0, regSDMA0_CHICKEN_BITS),
+};
+
+static void sdma_v7_0_set_ring_funcs(struct amdgpu_device *adev);
+static void sdma_v7_0_set_buffer_funcs(struct amdgpu_device *adev);
+static void sdma_v7_0_set_vm_pte_funcs(struct amdgpu_device *adev);
+static void sdma_v7_0_set_irq_funcs(struct amdgpu_device *adev);
+static int sdma_v7_0_start(struct amdgpu_device *adev);
+
+static u32 sdma_v7_0_get_reg_offset(struct amdgpu_device *adev, u32 instance, u32 internal_offset)
+{
+	u32 base;
+
+	if (internal_offset >= SDMA0_HYP_DEC_REG_START &&
+	    internal_offset <= SDMA0_HYP_DEC_REG_END) {
+		base = adev->reg_offset[GC_HWIP][0][1];
+		if (instance != 0)
+			internal_offset += SDMA1_HYP_DEC_REG_OFFSET * instance;
+	} else {
+		base = adev->reg_offset[GC_HWIP][0][0];
+		if (instance == 1)
+			internal_offset += SDMA1_REG_OFFSET;
+	}
+
+	return base + internal_offset;
+}
+
+static unsigned sdma_v7_0_ring_init_cond_exec(struct amdgpu_ring *ring,
+					      uint64_t addr)
+{
+	unsigned ret;
+
+	amdgpu_ring_write(ring, SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_COND_EXE));
+	amdgpu_ring_write(ring, lower_32_bits(addr));
+	amdgpu_ring_write(ring, upper_32_bits(addr));
+	amdgpu_ring_write(ring, 1);
+	/* this is the offset we need patch later */
+	ret = ring->wptr & ring->buf_mask;
+	/* insert dummy here and patch it later */
+	amdgpu_ring_write(ring, 0);
+
+	return ret;
+}
+
+/**
+ * sdma_v7_0_ring_get_rptr - get the current read pointer
+ *
+ * @ring: amdgpu ring pointer
+ *
+ * Get the current rptr from the hardware.
+ */
+static uint64_t sdma_v7_0_ring_get_rptr(struct amdgpu_ring *ring)
+{
+	u64 *rptr;
+
+	/* XXX check if swapping is necessary on BE */
+	rptr = (u64 *)ring->rptr_cpu_addr;
+
+	DRM_DEBUG("rptr before shift == 0x%016llx\n", *rptr);
+	return ((*rptr) >> 2);
+}
+
+/**
+ * sdma_v7_0_ring_get_wptr - get the current write pointer
+ *
+ * @ring: amdgpu ring pointer
+ *
+ * Get the current wptr from the hardware.
+ */
+static uint64_t sdma_v7_0_ring_get_wptr(struct amdgpu_ring *ring)
+{
+	u64 wptr = 0;
+
+	if (ring->use_doorbell) {
+		/* XXX check if swapping is necessary on BE */
+		wptr = READ_ONCE(*((u64 *)ring->wptr_cpu_addr));
+		DRM_DEBUG("wptr/doorbell before shift == 0x%016llx\n", wptr);
+	}
+
+	return wptr >> 2;
+}
+
+/**
+ * sdma_v7_0_ring_set_wptr - commit the write pointer
+ *
+ * @ring: amdgpu ring pointer
+ *
+ * Write the wptr back to the hardware.
+ */
+static void sdma_v7_0_ring_set_wptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	DRM_DEBUG("Setting write pointer\n");
+
+	if (ring->use_doorbell) {
+		DRM_DEBUG("Using doorbell -- "
+			  "wptr_offs == 0x%08x "
+			  "lower_32_bits(ring->wptr) << 2 == 0x%08x "
+			  "upper_32_bits(ring->wptr) << 2 == 0x%08x\n",
+			  ring->wptr_offs,
+			  lower_32_bits(ring->wptr << 2),
+			  upper_32_bits(ring->wptr << 2));
+		/* XXX check if swapping is necessary on BE */
+		atomic64_set((atomic64_t *)ring->wptr_cpu_addr,
+			     ring->wptr << 2);
+		DRM_DEBUG("calling WDOORBELL64(0x%08x, 0x%016llx)\n",
+			  ring->doorbell_index, ring->wptr << 2);
+		WDOORBELL64(ring->doorbell_index, ring->wptr << 2);
+	} else {
+		DRM_DEBUG("Not using doorbell -- "
+			  "regSDMA%i_GFX_RB_WPTR == 0x%08x "
+			  "regSDMA%i_GFX_RB_WPTR_HI == 0x%08x\n",
+			  ring->me,
+			  lower_32_bits(ring->wptr << 2),
+			  ring->me,
+			  upper_32_bits(ring->wptr << 2));
+		WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev,
+							     ring->me,
+							     regSDMA0_QUEUE0_RB_WPTR),
+				lower_32_bits(ring->wptr << 2));
+		WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev,
+							     ring->me,
+							     regSDMA0_QUEUE0_RB_WPTR_HI),
+				upper_32_bits(ring->wptr << 2));
+	}
+}
+
+static void sdma_v7_0_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count)
+{
+	struct amdgpu_sdma_instance *sdma = amdgpu_sdma_get_instance_from_ring(ring);
+	int i;
+
+	for (i = 0; i < count; i++)
+		if (sdma && sdma->burst_nop && (i == 0))
+			amdgpu_ring_write(ring, ring->funcs->nop |
+				SDMA_PKT_NOP_HEADER_COUNT(count - 1));
+		else
+			amdgpu_ring_write(ring, ring->funcs->nop);
+}
+
+/**
+ * sdma_v7_0_ring_emit_ib - Schedule an IB on the DMA engine
+ *
+ * @ring: amdgpu ring pointer
+ * @job: job to retrieve vmid from
+ * @ib: IB object to schedule
+ * @flags: unused
+ *
+ * Schedule an IB in the DMA ring.
+ */
+static void sdma_v7_0_ring_emit_ib(struct amdgpu_ring *ring,
+				   struct amdgpu_job *job,
+				   struct amdgpu_ib *ib,
+				   uint32_t flags)
+{
+	unsigned vmid = AMDGPU_JOB_GET_VMID(job);
+	uint64_t csa_mc_addr = amdgpu_sdma_get_csa_mc_addr(ring, vmid);
+
+	/* An IB packet must end on a 8 DW boundary--the next dword
+	 * must be on a 8-dword boundary. Our IB packet below is 6
+	 * dwords long, thus add x number of NOPs, such that, in
+	 * modular arithmetic,
+	 * wptr + 6 + x = 8k, k >= 0, which in C is,
+	 * (wptr + 6 + x) % 8 = 0.
+	 * The expression below, is a solution of x.
+	 */
+	sdma_v7_0_ring_insert_nop(ring, (2 - lower_32_bits(ring->wptr)) & 7);
+
+	amdgpu_ring_write(ring, SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_INDIRECT) |
+			  SDMA_PKT_INDIRECT_HEADER_VMID(vmid & 0xf));
+	/* base must be 32 byte aligned */
+	amdgpu_ring_write(ring, lower_32_bits(ib->gpu_addr) & 0xffffffe0);
+	amdgpu_ring_write(ring, upper_32_bits(ib->gpu_addr));
+	amdgpu_ring_write(ring, ib->length_dw);
+	amdgpu_ring_write(ring, lower_32_bits(csa_mc_addr));
+	amdgpu_ring_write(ring, upper_32_bits(csa_mc_addr));
+}
+
+/**
+ * sdma_v7_0_ring_emit_mem_sync - flush the IB by graphics cache rinse
+ *
+ * @ring: amdgpu ring pointer
+ *
+ * flush the IB by graphics cache rinse.
+ */
+static void sdma_v7_0_ring_emit_mem_sync(struct amdgpu_ring *ring)
+{
+	uint32_t gcr_cntl = SDMA_GCR_GL2_INV | SDMA_GCR_GL2_WB | SDMA_GCR_GLM_INV |
+		SDMA_GCR_GL1_INV | SDMA_GCR_GLV_INV | SDMA_GCR_GLK_INV |
+		SDMA_GCR_GLI_INV(1);
+
+	/* flush entire cache L0/L1/L2, this can be optimized by performance requirement */
+	amdgpu_ring_write(ring, SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_GCR_REQ));
+	amdgpu_ring_write(ring, SDMA_PKT_GCR_REQ_PAYLOAD1_BASE_VA_31_7(0));
+	amdgpu_ring_write(ring, SDMA_PKT_GCR_REQ_PAYLOAD2_GCR_CONTROL_15_0(gcr_cntl) |
+			  SDMA_PKT_GCR_REQ_PAYLOAD2_BASE_VA_47_32(0));
+	amdgpu_ring_write(ring, SDMA_PKT_GCR_REQ_PAYLOAD3_LIMIT_VA_31_7(0) |
+			  SDMA_PKT_GCR_REQ_PAYLOAD3_GCR_CONTROL_18_16(gcr_cntl >> 16));
+	amdgpu_ring_write(ring, SDMA_PKT_GCR_REQ_PAYLOAD4_LIMIT_VA_47_32(0) |
+			  SDMA_PKT_GCR_REQ_PAYLOAD4_VMID(0));
+}
+
+
+/**
+ * sdma_v7_0_ring_emit_hdp_flush - emit an hdp flush on the DMA ring
+ *
+ * @ring: amdgpu ring pointer
+ *
+ * Emit an hdp flush packet on the requested DMA ring.
+ */
+static void sdma_v7_0_ring_emit_hdp_flush(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	u32 ref_and_mask = 0;
+	const struct nbio_hdp_flush_reg *nbio_hf_reg = adev->nbio.hdp_flush_reg;
+
+	ref_and_mask = nbio_hf_reg->ref_and_mask_sdma0 << ring->me;
+
+	amdgpu_ring_write(ring, SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_POLL_REGMEM) |
+			  SDMA_PKT_POLL_REGMEM_HEADER_HDP_FLUSH(1) |
+			  SDMA_PKT_POLL_REGMEM_HEADER_FUNC(3)); /* == */
+	amdgpu_ring_write(ring, (adev->nbio.funcs->get_hdp_flush_done_offset(adev)) << 2);
+	amdgpu_ring_write(ring, (adev->nbio.funcs->get_hdp_flush_req_offset(adev)) << 2);
+	amdgpu_ring_write(ring, ref_and_mask); /* reference */
+	amdgpu_ring_write(ring, ref_and_mask); /* mask */
+	amdgpu_ring_write(ring, SDMA_PKT_POLL_REGMEM_DW5_RETRY_COUNT(0xfff) |
+			  SDMA_PKT_POLL_REGMEM_DW5_INTERVAL(10)); /* retry count, poll interval */
+}
+
+/**
+ * sdma_v7_0_ring_emit_fence - emit a fence on the DMA ring
+ *
+ * @ring: amdgpu ring pointer
+ * @addr: address
+ * @seq: fence seq number
+ * @flags: fence flags
+ *
+ * Add a DMA fence packet to the ring to write
+ * the fence seq number and DMA trap packet to generate
+ * an interrupt if needed.
+ */
+static void sdma_v7_0_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 seq,
+				      unsigned flags)
+{
+	bool write64bit = flags & AMDGPU_FENCE_FLAG_64BIT;
+	/* write the fence */
+	amdgpu_ring_write(ring, SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_FENCE) |
+			  SDMA_PKT_FENCE_HEADER_MTYPE(0x3)); /* Ucached(UC) */
+	/* zero in first two bits */
+	BUG_ON(addr & 0x3);
+	amdgpu_ring_write(ring, lower_32_bits(addr));
+	amdgpu_ring_write(ring, upper_32_bits(addr));
+	amdgpu_ring_write(ring, lower_32_bits(seq));
+
+	/* optionally write high bits as well */
+	if (write64bit) {
+		addr += 4;
+		amdgpu_ring_write(ring, SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_FENCE) |
+				  SDMA_PKT_FENCE_HEADER_MTYPE(0x3));
+		/* zero in first two bits */
+		BUG_ON(addr & 0x3);
+		amdgpu_ring_write(ring, lower_32_bits(addr));
+		amdgpu_ring_write(ring, upper_32_bits(addr));
+		amdgpu_ring_write(ring, upper_32_bits(seq));
+	}
+
+	if (flags & AMDGPU_FENCE_FLAG_INT) {
+		/* generate an interrupt */
+		amdgpu_ring_write(ring, SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_TRAP));
+		amdgpu_ring_write(ring, SDMA_PKT_TRAP_INT_CONTEXT_INT_CONTEXT(0));
+	}
+}
+
+/**
+ * sdma_v7_0_gfx_stop - stop the gfx async dma engines
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Stop the gfx async dma ring buffers.
+ */
+static void sdma_v7_0_gfx_stop(struct amdgpu_device *adev)
+{
+	u32 rb_cntl, ib_cntl;
+	int i;
+
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		rb_cntl = RREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_CNTL));
+		rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_QUEUE0_RB_CNTL, RB_ENABLE, 0);
+		WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_CNTL), rb_cntl);
+		ib_cntl = RREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_IB_CNTL));
+		ib_cntl = REG_SET_FIELD(ib_cntl, SDMA0_QUEUE0_IB_CNTL, IB_ENABLE, 0);
+		WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_IB_CNTL), ib_cntl);
+	}
+}
+
+/**
+ * sdma_v7_0_rlc_stop - stop the compute async dma engines
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Stop the compute async dma queues.
+ */
+static void sdma_v7_0_rlc_stop(struct amdgpu_device *adev)
+{
+	/* XXX todo */
+}
+
+/**
+ * sdma_v7_0_ctx_switch_enable - stop the async dma engines context switch
+ *
+ * @adev: amdgpu_device pointer
+ * @enable: enable/disable the DMA MEs context switch.
+ *
+ * Halt or unhalt the async dma engines context switch.
+ */
+static void sdma_v7_0_ctx_switch_enable(struct amdgpu_device *adev, bool enable)
+{
+}
+
+/**
+ * sdma_v7_0_enable - stop the async dma engines
+ *
+ * @adev: amdgpu_device pointer
+ * @enable: enable/disable the DMA MEs.
+ *
+ * Halt or unhalt the async dma engines.
+ */
+static void sdma_v7_0_enable(struct amdgpu_device *adev, bool enable)
+{
+	u32 mcu_cntl;
+	int i;
+
+	if (!enable) {
+		sdma_v7_0_gfx_stop(adev);
+		sdma_v7_0_rlc_stop(adev);
+	}
+
+	if (amdgpu_sriov_vf(adev))
+		return;
+
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		mcu_cntl = RREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_MCU_CNTL));
+		mcu_cntl = REG_SET_FIELD(mcu_cntl, SDMA0_MCU_CNTL, HALT, enable ? 0 : 1);
+		WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_MCU_CNTL), mcu_cntl);
+	}
+}
+
+/**
+ * sdma_v7_0_gfx_resume_instance - start/restart a certain sdma engine
+ *
+ * @adev: amdgpu_device pointer
+ * @i: instance
+ * @restore: used to restore wptr when restart
+ *
+ * Set up the gfx DMA ring buffers and enable them. On restart, we will restore wptr and rptr.
+ * Return 0 for success.
+ */
+static int sdma_v7_0_gfx_resume_instance(struct amdgpu_device *adev, int i, bool restore)
+{
+	struct amdgpu_ring *ring;
+	u32 rb_cntl, ib_cntl;
+	u32 rb_bufsz;
+	u32 doorbell;
+	u32 doorbell_offset;
+	u32 temp;
+	u64 wptr_gpu_addr;
+	int r;
+
+	ring = &adev->sdma.instance[i].ring;
+
+	/* Set ring buffer size in dwords */
+	rb_bufsz = order_base_2(ring->ring_size / 4);
+	rb_cntl = RREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_CNTL));
+	rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_QUEUE0_RB_CNTL, RB_SIZE, rb_bufsz);
+#ifdef __BIG_ENDIAN
+	rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_QUEUE0_RB_CNTL, RB_SWAP_ENABLE, 1);
+	rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_QUEUE0_RB_CNTL,
+				RPTR_WRITEBACK_SWAP_ENABLE, 1);
+#endif
+	rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_QUEUE0_RB_CNTL, RB_PRIV, 1);
+	WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_CNTL), rb_cntl);
+
+	/* Initialize the ring buffer's read and write pointers */
+	if (restore) {
+		WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_RPTR), lower_32_bits(ring->wptr << 2));
+		WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_RPTR_HI), upper_32_bits(ring->wptr << 2));
+		WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_WPTR), lower_32_bits(ring->wptr << 2));
+		WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_WPTR_HI), upper_32_bits(ring->wptr << 2));
+	} else {
+		WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_RPTR), 0);
+		WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_RPTR_HI), 0);
+		WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_WPTR), 0);
+		WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_WPTR_HI), 0);
+	}
+	/* setup the wptr shadow polling */
+	wptr_gpu_addr = ring->wptr_gpu_addr;
+	WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_WPTR_POLL_ADDR_LO),
+	       lower_32_bits(wptr_gpu_addr));
+	WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_WPTR_POLL_ADDR_HI),
+	       upper_32_bits(wptr_gpu_addr));
+
+	/* set the wb address whether it's enabled or not */
+	WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_RPTR_ADDR_HI),
+	       upper_32_bits(ring->rptr_gpu_addr) & 0xFFFFFFFF);
+	WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_RPTR_ADDR_LO),
+	       lower_32_bits(ring->rptr_gpu_addr) & 0xFFFFFFFC);
+
+	rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_QUEUE0_RB_CNTL, RPTR_WRITEBACK_ENABLE, 1);
+	if (amdgpu_sriov_vf(adev))
+		rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_QUEUE0_RB_CNTL, WPTR_POLL_ENABLE, 1);
+	else
+		rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_QUEUE0_RB_CNTL, WPTR_POLL_ENABLE, 0);
+
+	rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_QUEUE0_RB_CNTL, MCU_WPTR_POLL_ENABLE, 1);
+
+	WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_BASE), ring->gpu_addr >> 8);
+	WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_BASE_HI), ring->gpu_addr >> 40);
+
+	if (!restore)
+		ring->wptr = 0;
+
+	/* before programing wptr to a less value, need set minor_ptr_update first */
+	WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_MINOR_PTR_UPDATE), 1);
+
+	if (!amdgpu_sriov_vf(adev)) { /* only bare-metal use register write for wptr */
+		WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_WPTR), lower_32_bits(ring->wptr) << 2);
+		WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_WPTR_HI), upper_32_bits(ring->wptr) << 2);
+	}
+
+	doorbell = RREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_DOORBELL));
+	doorbell_offset = RREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_DOORBELL_OFFSET));
+
+	if (ring->use_doorbell) {
+		doorbell = REG_SET_FIELD(doorbell, SDMA0_QUEUE0_DOORBELL, ENABLE, 1);
+		doorbell_offset = REG_SET_FIELD(doorbell_offset, SDMA0_QUEUE0_DOORBELL_OFFSET,
+				OFFSET, ring->doorbell_index);
+	} else {
+		doorbell = REG_SET_FIELD(doorbell, SDMA0_QUEUE0_DOORBELL, ENABLE, 0);
+	}
+	WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_DOORBELL), doorbell);
+	WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_DOORBELL_OFFSET), doorbell_offset);
+
+	if (i == 0)
+		adev->nbio.funcs->sdma_doorbell_range(adev, i, ring->use_doorbell,
+					      ring->doorbell_index,
+					      adev->doorbell_index.sdma_doorbell_range * adev->sdma.num_instances);
+
+	if (amdgpu_sriov_vf(adev))
+		sdma_v7_0_ring_set_wptr(ring);
+
+	/* set minor_ptr_update to 0 after wptr programed */
+	WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_MINOR_PTR_UPDATE), 0);
+
+	/* Set up sdma hang watchdog */
+	temp = RREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_WATCHDOG_CNTL));
+	/* 100ms per unit */
+	temp = REG_SET_FIELD(temp, SDMA0_WATCHDOG_CNTL, QUEUE_HANG_COUNT,
+			     max(adev->usec_timeout/100000, 1));
+	WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_WATCHDOG_CNTL), temp);
+
+	/* Set up RESP_MODE to non-copy addresses */
+	temp = RREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_UTCL1_CNTL));
+	temp = REG_SET_FIELD(temp, SDMA0_UTCL1_CNTL, RESP_MODE, 3);
+	temp = REG_SET_FIELD(temp, SDMA0_UTCL1_CNTL, REDO_DELAY, 9);
+	WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_UTCL1_CNTL), temp);
+
+	/* program default cache read and write policy */
+	temp = RREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_UTCL1_PAGE));
+	/* clean read policy and write policy bits */
+	temp &= 0xFF0FFF;
+	temp |= ((CACHE_READ_POLICY_L2__DEFAULT << 12) |
+		 (CACHE_WRITE_POLICY_L2__DEFAULT << 14));
+	WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_UTCL1_PAGE), temp);
+
+	if (!amdgpu_sriov_vf(adev)) {
+		/* unhalt engine */
+		temp = RREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_MCU_CNTL));
+		temp = REG_SET_FIELD(temp, SDMA0_MCU_CNTL, HALT, 0);
+		temp = REG_SET_FIELD(temp, SDMA0_MCU_CNTL, RESET, 0);
+		WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_MCU_CNTL), temp);
+	}
+
+	/* enable DMA RB */
+	rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_QUEUE0_RB_CNTL, RB_ENABLE, 1);
+	WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_RB_CNTL), rb_cntl);
+
+	ib_cntl = RREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_IB_CNTL));
+	ib_cntl = REG_SET_FIELD(ib_cntl, SDMA0_QUEUE0_IB_CNTL, IB_ENABLE, 1);
+#ifdef __BIG_ENDIAN
+	ib_cntl = REG_SET_FIELD(ib_cntl, SDMA0_QUEUE0_IB_CNTL, IB_SWAP_ENABLE, 1);
+#endif
+	/* enable DMA IBs */
+	WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_IB_CNTL), ib_cntl);
+	ring->sched.ready = true;
+
+	if (amdgpu_sriov_vf(adev)) { /* bare-metal sequence doesn't need below to lines */
+		sdma_v7_0_ctx_switch_enable(adev, true);
+		sdma_v7_0_enable(adev, true);
+	}
+
+	r = amdgpu_ring_test_helper(ring);
+	if (r)
+		ring->sched.ready = false;
+
+	return r;
+}
+
+/**
+ * sdma_v7_0_gfx_resume - setup and start the async dma engines
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Set up the gfx DMA ring buffers and enable them.
+ * Returns 0 for success, error for failure.
+ */
+static int sdma_v7_0_gfx_resume(struct amdgpu_device *adev)
+{
+	int i, r;
+
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		r = sdma_v7_0_gfx_resume_instance(adev, i, false);
+		if (r)
+			return r;
+	}
+
+	return 0;
+
+}
+
+/**
+ * sdma_v7_0_rlc_resume - setup and start the async dma engines
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Set up the compute DMA queues and enable them.
+ * Returns 0 for success, error for failure.
+ */
+static int sdma_v7_0_rlc_resume(struct amdgpu_device *adev)
+{
+	return 0;
+}
+
+static void sdma_v12_0_free_ucode_buffer(struct amdgpu_device *adev)
+{
+	int i;
+
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		amdgpu_bo_free_kernel(&adev->sdma.instance[i].sdma_fw_obj,
+				      &adev->sdma.instance[i].sdma_fw_gpu_addr,
+				      (void **)&adev->sdma.instance[i].sdma_fw_ptr);
+	}
+}
+
+/**
+ * sdma_v7_0_load_microcode - load the sDMA ME ucode
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Loads the sDMA0/1 ucode.
+ * Returns 0 for success, -EINVAL if the ucode is not available.
+ */
+static int sdma_v7_0_load_microcode(struct amdgpu_device *adev)
+{
+	const struct sdma_firmware_header_v3_0 *hdr;
+	const __le32 *fw_data;
+	u32 fw_size;
+	uint32_t tmp, sdma_status, ic_op_cntl;
+	int i, r, j;
+
+	/* halt the MEs */
+	sdma_v7_0_enable(adev, false);
+
+	if (!adev->sdma.instance[0].fw)
+		return -EINVAL;
+
+	hdr = (const struct sdma_firmware_header_v3_0 *)
+		adev->sdma.instance[0].fw->data;
+	amdgpu_ucode_print_sdma_hdr(&hdr->header);
+
+	fw_data = (const __le32 *)(adev->sdma.instance[0].fw->data +
+			le32_to_cpu(hdr->ucode_offset_bytes));
+	fw_size = le32_to_cpu(hdr->ucode_size_bytes);
+
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		r = amdgpu_bo_create_reserved(adev, fw_size,
+					      PAGE_SIZE,
+					      AMDGPU_GEM_DOMAIN_VRAM,
+					      &adev->sdma.instance[i].sdma_fw_obj,
+					      &adev->sdma.instance[i].sdma_fw_gpu_addr,
+					      (void **)&adev->sdma.instance[i].sdma_fw_ptr);
+		if (r) {
+			dev_err(adev->dev, "(%d) failed to create sdma ucode bo\n", r);
+			return r;
+		}
+
+		memcpy(adev->sdma.instance[i].sdma_fw_ptr, fw_data, fw_size);
+
+		amdgpu_bo_kunmap(adev->sdma.instance[i].sdma_fw_obj);
+		amdgpu_bo_unreserve(adev->sdma.instance[i].sdma_fw_obj);
+
+		tmp = RREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_IC_CNTL));
+		tmp = REG_SET_FIELD(tmp, SDMA0_IC_CNTL, GPA, 0);
+		WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_IC_CNTL), tmp);
+
+		WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_IC_BASE_LO),
+			lower_32_bits(adev->sdma.instance[i].sdma_fw_gpu_addr));
+		WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_IC_BASE_HI),
+			upper_32_bits(adev->sdma.instance[i].sdma_fw_gpu_addr));
+
+		tmp = RREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_IC_OP_CNTL));
+		tmp = REG_SET_FIELD(tmp, SDMA0_IC_OP_CNTL, PRIME_ICACHE, 1);
+		WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_IC_OP_CNTL), tmp);
+
+		/* Wait for sdma ucode init complete */
+		for (j = 0; j < adev->usec_timeout; j++) {
+			ic_op_cntl = RREG32_SOC15_IP(GC,
+					sdma_v7_0_get_reg_offset(adev, i, regSDMA0_IC_OP_CNTL));
+			sdma_status = RREG32_SOC15_IP(GC,
+					sdma_v7_0_get_reg_offset(adev, i, regSDMA0_STATUS_REG));
+			if ((REG_GET_FIELD(ic_op_cntl, SDMA0_IC_OP_CNTL, ICACHE_PRIMED) == 1) &&
+			    (REG_GET_FIELD(sdma_status, SDMA0_STATUS_REG, UCODE_INIT_DONE) == 1))
+				break;
+			udelay(1);
+		}
+
+		if (j >= adev->usec_timeout) {
+			dev_err(adev->dev, "failed to init sdma ucode\n");
+			return -EINVAL;
+		}
+	}
+
+	return 0;
+}
+
+static int sdma_v7_0_soft_reset(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	u32 tmp;
+	int i;
+
+	sdma_v7_0_gfx_stop(adev);
+
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		//tmp = RREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_FREEZE));
+		//tmp |= SDMA0_FREEZE__FREEZE_MASK;
+		//WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_FREEZE), tmp);
+		tmp = RREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_MCU_CNTL));
+		tmp |= SDMA0_MCU_CNTL__HALT_MASK;
+		tmp |= SDMA0_MCU_CNTL__RESET_MASK;
+		WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_MCU_CNTL), tmp);
+
+		WREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, i, regSDMA0_QUEUE0_PREEMPT), 0);
+
+		udelay(100);
+
+		tmp = GRBM_SOFT_RESET__SOFT_RESET_SDMA0_MASK << i;
+		WREG32_SOC15(GC, 0, regGRBM_SOFT_RESET, tmp);
+		tmp = RREG32_SOC15(GC, 0, regGRBM_SOFT_RESET);
+
+		udelay(100);
+
+		WREG32_SOC15(GC, 0, regGRBM_SOFT_RESET, 0);
+		tmp = RREG32_SOC15(GC, 0, regGRBM_SOFT_RESET);
+
+		udelay(100);
+	}
+
+	return sdma_v7_0_start(adev);
+}
+
+static bool sdma_v7_0_check_soft_reset(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	struct amdgpu_ring *ring;
+	int i, r;
+	long tmo = msecs_to_jiffies(1000);
+
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		ring = &adev->sdma.instance[i].ring;
+		r = amdgpu_ring_test_ib(ring, tmo);
+		if (r)
+			return true;
+	}
+
+	return false;
+}
+
+static int sdma_v7_0_reset_queue(struct amdgpu_ring *ring,
+				 unsigned int vmid,
+				 struct amdgpu_fence *timedout_fence)
+{
+	struct amdgpu_device *adev = ring->adev;
+	int r;
+
+	if (ring->me >= adev->sdma.num_instances) {
+		dev_err(adev->dev, "sdma instance not found\n");
+		return -EINVAL;
+	}
+
+	amdgpu_ring_reset_helper_begin(ring, timedout_fence);
+
+	r = amdgpu_mes_reset_legacy_queue(adev, ring, vmid, true);
+	if (r)
+		return r;
+
+	r = sdma_v7_0_gfx_resume_instance(adev, ring->me, true);
+	if (r)
+		return r;
+
+	return amdgpu_ring_reset_helper_end(ring, timedout_fence);
+}
+
+/**
+ * sdma_v7_0_start - setup and start the async dma engines
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Set up the DMA engines and enable them.
+ * Returns 0 for success, error for failure.
+ */
+static int sdma_v7_0_start(struct amdgpu_device *adev)
+{
+	int r = 0;
+
+	if (amdgpu_sriov_vf(adev)) {
+		sdma_v7_0_ctx_switch_enable(adev, false);
+		sdma_v7_0_enable(adev, false);
+
+		/* set RB registers */
+		r = sdma_v7_0_gfx_resume(adev);
+		return r;
+	}
+
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT) {
+		r = sdma_v7_0_load_microcode(adev);
+		if (r) {
+			sdma_v12_0_free_ucode_buffer(adev);
+			return r;
+		}
+
+		if (amdgpu_emu_mode == 1)
+			msleep(1000);
+	}
+
+	/* unhalt the MEs */
+	sdma_v7_0_enable(adev, true);
+	/* enable sdma ring preemption */
+	sdma_v7_0_ctx_switch_enable(adev, true);
+
+	/* start the gfx rings and rlc compute queues */
+	r = sdma_v7_0_gfx_resume(adev);
+	if (r)
+		return r;
+	r = sdma_v7_0_rlc_resume(adev);
+
+	return r;
+}
+
+static int sdma_v7_0_mqd_init(struct amdgpu_device *adev, void *mqd,
+			      struct amdgpu_mqd_prop *prop)
+{
+	struct v12_sdma_mqd *m = mqd;
+	uint64_t wb_gpu_addr;
+
+	m->sdmax_rlcx_rb_cntl =
+		order_base_2(prop->queue_size / 4) << SDMA0_QUEUE0_RB_CNTL__RB_SIZE__SHIFT |
+		1 << SDMA0_QUEUE0_RB_CNTL__RPTR_WRITEBACK_ENABLE__SHIFT |
+		4 << SDMA0_QUEUE0_RB_CNTL__RPTR_WRITEBACK_TIMER__SHIFT |
+		1 << SDMA0_QUEUE0_RB_CNTL__MCU_WPTR_POLL_ENABLE__SHIFT;
+
+	m->sdmax_rlcx_rb_base = lower_32_bits(prop->hqd_base_gpu_addr >> 8);
+	m->sdmax_rlcx_rb_base_hi = upper_32_bits(prop->hqd_base_gpu_addr >> 8);
+
+	wb_gpu_addr = prop->wptr_gpu_addr;
+	m->sdmax_rlcx_rb_wptr_poll_addr_lo = lower_32_bits(wb_gpu_addr);
+	m->sdmax_rlcx_rb_wptr_poll_addr_hi = upper_32_bits(wb_gpu_addr);
+
+	wb_gpu_addr = prop->rptr_gpu_addr;
+	m->sdmax_rlcx_rb_rptr_addr_lo = lower_32_bits(wb_gpu_addr);
+	m->sdmax_rlcx_rb_rptr_addr_hi = upper_32_bits(wb_gpu_addr);
+
+	m->sdmax_rlcx_ib_cntl = RREG32_SOC15_IP(GC, sdma_v7_0_get_reg_offset(adev, 0,
+							regSDMA0_QUEUE0_IB_CNTL));
+
+	m->sdmax_rlcx_doorbell_offset =
+		prop->doorbell_index << SDMA0_QUEUE0_DOORBELL_OFFSET__OFFSET__SHIFT;
+
+	m->sdmax_rlcx_doorbell = REG_SET_FIELD(0, SDMA0_QUEUE0_DOORBELL, ENABLE, 1);
+
+	m->sdmax_rlcx_doorbell_log = 0;
+	m->sdmax_rlcx_rb_aql_cntl = 0x4000;	//regSDMA0_QUEUE0_RB_AQL_CNTL_DEFAULT;
+	m->sdmax_rlcx_dummy_reg = 0xf;	//regSDMA0_QUEUE0_DUMMY_REG_DEFAULT;
+
+	m->sdmax_rlcx_csa_addr_lo = lower_32_bits(prop->csa_addr);
+	m->sdmax_rlcx_csa_addr_hi = upper_32_bits(prop->csa_addr);
+
+	m->sdmax_rlcx_mcu_dbg0 = lower_32_bits(prop->fence_address);
+	m->sdmax_rlcx_mcu_dbg1 = upper_32_bits(prop->fence_address);
+
+	return 0;
+}
+
+static void sdma_v7_0_set_mqd_funcs(struct amdgpu_device *adev)
+{
+	adev->mqds[AMDGPU_HW_IP_DMA].mqd_size = sizeof(struct v12_sdma_mqd);
+	adev->mqds[AMDGPU_HW_IP_DMA].init_mqd = sdma_v7_0_mqd_init;
+}
+
+/**
+ * sdma_v7_0_ring_test_ring - simple async dma engine test
+ *
+ * @ring: amdgpu_ring structure holding ring information
+ *
+ * Test the DMA engine by writing using it to write an
+ * value to memory.
+ * Returns 0 for success, error for failure.
+ */
+static int sdma_v7_0_ring_test_ring(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	unsigned i;
+	unsigned index;
+	int r;
+	u32 tmp;
+	u64 gpu_addr;
+
+	tmp = 0xCAFEDEAD;
+
+	r = amdgpu_device_wb_get(adev, &index);
+	if (r) {
+		dev_err(adev->dev, "(%d) failed to allocate wb slot\n", r);
+		return r;
+	}
+
+	gpu_addr = adev->wb.gpu_addr + (index * 4);
+	adev->wb.wb[index] = cpu_to_le32(tmp);
+
+	r = amdgpu_ring_alloc(ring, 5);
+	if (r) {
+		DRM_ERROR("amdgpu: dma failed to lock ring %d (%d).\n", ring->idx, r);
+		amdgpu_device_wb_free(adev, index);
+		return r;
+	}
+
+	amdgpu_ring_write(ring, SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_WRITE) |
+			  SDMA_PKT_COPY_LINEAR_HEADER_SUB_OP(SDMA_SUBOP_WRITE_LINEAR));
+	amdgpu_ring_write(ring, lower_32_bits(gpu_addr));
+	amdgpu_ring_write(ring, upper_32_bits(gpu_addr));
+	amdgpu_ring_write(ring, SDMA_PKT_WRITE_UNTILED_DW_3_COUNT(0));
+	amdgpu_ring_write(ring, 0xDEADBEEF);
+	amdgpu_ring_commit(ring);
+
+	for (i = 0; i < adev->usec_timeout; i++) {
+		tmp = le32_to_cpu(adev->wb.wb[index]);
+		if (tmp == 0xDEADBEEF)
+			break;
+		if (amdgpu_emu_mode == 1)
+			msleep(1);
+		else
+			udelay(1);
+	}
+
+	if (i >= adev->usec_timeout)
+		r = -ETIMEDOUT;
+
+	amdgpu_device_wb_free(adev, index);
+
+	return r;
+}
+
+/**
+ * sdma_v7_0_ring_test_ib - test an IB on the DMA engine
+ *
+ * @ring: amdgpu_ring structure holding ring information
+ * @timeout: timeout value in jiffies, or MAX_SCHEDULE_TIMEOUT
+ *
+ * Test a simple IB in the DMA ring.
+ * Returns 0 on success, error on failure.
+ */
+static int sdma_v7_0_ring_test_ib(struct amdgpu_ring *ring, long timeout)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct amdgpu_ib ib;
+	struct dma_fence *f = NULL;
+	unsigned index;
+	long r;
+	u32 tmp = 0;
+	u64 gpu_addr;
+
+	tmp = 0xCAFEDEAD;
+	memset(&ib, 0, sizeof(ib));
+
+	r = amdgpu_device_wb_get(adev, &index);
+	if (r) {
+		dev_err(adev->dev, "(%ld) failed to allocate wb slot\n", r);
+		return r;
+	}
+
+	gpu_addr = adev->wb.gpu_addr + (index * 4);
+	adev->wb.wb[index] = cpu_to_le32(tmp);
+
+	r = amdgpu_ib_get(adev, NULL, 256, AMDGPU_IB_POOL_DIRECT, &ib);
+	if (r) {
+		DRM_ERROR("amdgpu: failed to get ib (%ld).\n", r);
+		goto err0;
+	}
+
+	ib.ptr[0] = SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_WRITE) |
+		SDMA_PKT_COPY_LINEAR_HEADER_SUB_OP(SDMA_SUBOP_WRITE_LINEAR);
+	ib.ptr[1] = lower_32_bits(gpu_addr);
+	ib.ptr[2] = upper_32_bits(gpu_addr);
+	ib.ptr[3] = SDMA_PKT_WRITE_UNTILED_DW_3_COUNT(0);
+	ib.ptr[4] = 0xDEADBEEF;
+	ib.ptr[5] = SDMA_PKT_NOP_HEADER_OP(SDMA_OP_NOP);
+	ib.ptr[6] = SDMA_PKT_NOP_HEADER_OP(SDMA_OP_NOP);
+	ib.ptr[7] = SDMA_PKT_NOP_HEADER_OP(SDMA_OP_NOP);
+	ib.length_dw = 8;
+
+	r = amdgpu_ib_schedule(ring, 1, &ib, NULL, &f);
+	if (r)
+		goto err1;
+
+	r = dma_fence_wait_timeout(f, false, timeout);
+	if (r == 0) {
+		DRM_ERROR("amdgpu: IB test timed out\n");
+		r = -ETIMEDOUT;
+		goto err1;
+	} else if (r < 0) {
+		DRM_ERROR("amdgpu: fence wait failed (%ld).\n", r);
+		goto err1;
+	}
+
+	tmp = le32_to_cpu(adev->wb.wb[index]);
+
+	if (tmp == 0xDEADBEEF)
+		r = 0;
+	else
+		r = -EINVAL;
+
+err1:
+	amdgpu_ib_free(&ib, NULL);
+	dma_fence_put(f);
+err0:
+	amdgpu_device_wb_free(adev, index);
+	return r;
+}
+
+
+/**
+ * sdma_v7_0_vm_copy_pte - update PTEs by copying them from the GART
+ *
+ * @ib: indirect buffer to fill with commands
+ * @pe: addr of the page entry
+ * @src: src addr to copy from
+ * @count: number of page entries to update
+ *
+ * Update PTEs by copying them from the GART using sDMA.
+ */
+static void sdma_v7_0_vm_copy_pte(struct amdgpu_ib *ib,
+				  uint64_t pe, uint64_t src,
+				  unsigned count)
+{
+	unsigned bytes = count * 8;
+
+	ib->ptr[ib->length_dw++] = SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_COPY) |
+		SDMA_PKT_COPY_LINEAR_HEADER_SUB_OP(SDMA_SUBOP_COPY_LINEAR) |
+		SDMA_PKT_COPY_LINEAR_HEADER_CPV(1);
+
+	ib->ptr[ib->length_dw++] = bytes - 1;
+	ib->ptr[ib->length_dw++] = 0; /* src/dst endian swap */
+	ib->ptr[ib->length_dw++] = lower_32_bits(src);
+	ib->ptr[ib->length_dw++] = upper_32_bits(src);
+	ib->ptr[ib->length_dw++] = lower_32_bits(pe);
+	ib->ptr[ib->length_dw++] = upper_32_bits(pe);
+	ib->ptr[ib->length_dw++] = 0;
+
+}
+
+/**
+ * sdma_v7_0_vm_write_pte - update PTEs by writing them manually
+ *
+ * @ib: indirect buffer to fill with commands
+ * @pe: addr of the page entry
+ * @value: dst addr to write into pe
+ * @count: number of page entries to update
+ * @incr: increase next addr by incr bytes
+ *
+ * Update PTEs by writing them manually using sDMA.
+ */
+static void sdma_v7_0_vm_write_pte(struct amdgpu_ib *ib, uint64_t pe,
+				   uint64_t value, unsigned count,
+				   uint32_t incr)
+{
+	unsigned ndw = count * 2;
+
+	ib->ptr[ib->length_dw++] = SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_WRITE) |
+		SDMA_PKT_COPY_LINEAR_HEADER_SUB_OP(SDMA_SUBOP_WRITE_LINEAR);
+	ib->ptr[ib->length_dw++] = lower_32_bits(pe);
+	ib->ptr[ib->length_dw++] = upper_32_bits(pe);
+	ib->ptr[ib->length_dw++] = ndw - 1;
+	for (; ndw > 0; ndw -= 2) {
+		ib->ptr[ib->length_dw++] = lower_32_bits(value);
+		ib->ptr[ib->length_dw++] = upper_32_bits(value);
+		value += incr;
+	}
+}
+
+/**
+ * sdma_v7_0_vm_set_pte_pde - update the page tables using sDMA
+ *
+ * @ib: indirect buffer to fill with commands
+ * @pe: addr of the page entry
+ * @addr: dst addr to write into pe
+ * @count: number of page entries to update
+ * @incr: increase next addr by incr bytes
+ * @flags: access flags
+ *
+ * Update the page tables using sDMA.
+ */
+static void sdma_v7_0_vm_set_pte_pde(struct amdgpu_ib *ib,
+				     uint64_t pe,
+				     uint64_t addr, unsigned count,
+				     uint32_t incr, uint64_t flags)
+{
+	/* for physically contiguous pages (vram) */
+	ib->ptr[ib->length_dw++] = SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_PTEPDE);
+	ib->ptr[ib->length_dw++] = lower_32_bits(pe); /* dst addr */
+	ib->ptr[ib->length_dw++] = upper_32_bits(pe);
+	ib->ptr[ib->length_dw++] = lower_32_bits(flags); /* mask */
+	ib->ptr[ib->length_dw++] = upper_32_bits(flags);
+	ib->ptr[ib->length_dw++] = lower_32_bits(addr); /* value */
+	ib->ptr[ib->length_dw++] = upper_32_bits(addr);
+	ib->ptr[ib->length_dw++] = incr; /* increment size */
+	ib->ptr[ib->length_dw++] = 0;
+	ib->ptr[ib->length_dw++] = count - 1; /* number of entries */
+}
+
+/**
+ * sdma_v7_0_ring_pad_ib - pad the IB
+ *
+ * @ring: amdgpu ring pointer
+ * @ib: indirect buffer to fill with padding
+ *
+ * Pad the IB with NOPs to a boundary multiple of 8.
+ */
+static void sdma_v7_0_ring_pad_ib(struct amdgpu_ring *ring, struct amdgpu_ib *ib)
+{
+	struct amdgpu_sdma_instance *sdma = amdgpu_sdma_get_instance_from_ring(ring);
+	u32 pad_count;
+	int i;
+
+	pad_count = (-ib->length_dw) & 0x7;
+	for (i = 0; i < pad_count; i++)
+		if (sdma && sdma->burst_nop && (i == 0))
+			ib->ptr[ib->length_dw++] =
+				SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_NOP) |
+				SDMA_PKT_NOP_HEADER_COUNT(pad_count - 1);
+		else
+			ib->ptr[ib->length_dw++] =
+				SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_NOP);
+}
+
+/**
+ * sdma_v7_0_ring_emit_pipeline_sync - sync the pipeline
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Make sure all previous operations are completed (CIK).
+ */
+static void sdma_v7_0_ring_emit_pipeline_sync(struct amdgpu_ring *ring)
+{
+	uint32_t seq = ring->fence_drv.sync_seq;
+	uint64_t addr = ring->fence_drv.gpu_addr;
+
+	/* wait for idle */
+	amdgpu_ring_write(ring, SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_POLL_REGMEM) |
+			  SDMA_PKT_POLL_REGMEM_HEADER_HDP_FLUSH(0) |
+			  SDMA_PKT_POLL_REGMEM_HEADER_FUNC(3) | /* equal */
+			  SDMA_PKT_POLL_REGMEM_HEADER_MEM_POLL(1));
+	amdgpu_ring_write(ring, addr & 0xfffffffc);
+	amdgpu_ring_write(ring, upper_32_bits(addr) & 0xffffffff);
+	amdgpu_ring_write(ring, seq); /* reference */
+	amdgpu_ring_write(ring, 0xffffffff); /* mask */
+	amdgpu_ring_write(ring, SDMA_PKT_POLL_REGMEM_DW5_RETRY_COUNT(0xfff) |
+			  SDMA_PKT_POLL_REGMEM_DW5_INTERVAL(4)); /* retry count, poll interval */
+}
+
+/**
+ * sdma_v7_0_ring_emit_vm_flush - vm flush using sDMA
+ *
+ * @ring: amdgpu_ring pointer
+ * @vmid: vmid number to use
+ * @pd_addr: address
+ *
+ * Update the page table base and flush the VM TLB
+ * using sDMA.
+ */
+static void sdma_v7_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
+					 unsigned vmid, uint64_t pd_addr)
+{
+	amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pd_addr);
+}
+
+static void sdma_v7_0_ring_emit_wreg(struct amdgpu_ring *ring,
+				     uint32_t reg, uint32_t val)
+{
+	/* SRBM WRITE command will not support on sdma v7.
+	 * Use Register WRITE command instead, which OPCODE is same as SRBM WRITE
+	 */
+	amdgpu_ring_write(ring, SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_SRBM_WRITE));
+	amdgpu_ring_write(ring, reg << 2);
+	amdgpu_ring_write(ring, val);
+}
+
+static void sdma_v7_0_ring_emit_reg_wait(struct amdgpu_ring *ring, uint32_t reg,
+					 uint32_t val, uint32_t mask)
+{
+	amdgpu_ring_write(ring, SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_POLL_REGMEM) |
+			  SDMA_PKT_POLL_REGMEM_HEADER_HDP_FLUSH(0) |
+			  SDMA_PKT_POLL_REGMEM_HEADER_FUNC(3)); /* equal */
+	amdgpu_ring_write(ring, reg << 2);
+	amdgpu_ring_write(ring, 0);
+	amdgpu_ring_write(ring, val); /* reference */
+	amdgpu_ring_write(ring, mask); /* mask */
+	amdgpu_ring_write(ring, SDMA_PKT_POLL_REGMEM_DW5_RETRY_COUNT(0xfff) |
+			  SDMA_PKT_POLL_REGMEM_DW5_INTERVAL(10));
+}
+
+static void sdma_v7_0_ring_emit_reg_write_reg_wait(struct amdgpu_ring *ring,
+						   uint32_t reg0, uint32_t reg1,
+						   uint32_t ref, uint32_t mask)
+{
+	amdgpu_ring_emit_wreg(ring, reg0, ref);
+	/* wait for a cycle to reset vm_inv_eng*_ack */
+	amdgpu_ring_emit_reg_wait(ring, reg0, 0, 0);
+	amdgpu_ring_emit_reg_wait(ring, reg1, mask, mask);
+}
+
+static int sdma_v7_0_early_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int r;
+
+	switch (amdgpu_user_queue) {
+	case -1:
+	case 0:
+	default:
+		adev->sdma.no_user_submission = false;
+		adev->sdma.disable_uq = true;
+		break;
+	case 1:
+		adev->sdma.no_user_submission = false;
+		adev->sdma.disable_uq = false;
+		break;
+	case 2:
+		adev->sdma.no_user_submission = true;
+		adev->sdma.disable_uq = false;
+		break;
+	}
+
+	r = amdgpu_sdma_init_microcode(adev, 0, true);
+	if (r) {
+		DRM_ERROR("Failed to init sdma firmware!\n");
+		return r;
+	}
+
+	sdma_v7_0_set_ring_funcs(adev);
+	sdma_v7_0_set_buffer_funcs(adev);
+	sdma_v7_0_set_vm_pte_funcs(adev);
+	sdma_v7_0_set_irq_funcs(adev);
+	sdma_v7_0_set_mqd_funcs(adev);
+
+	return 0;
+}
+
+static int sdma_v7_0_sw_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_ring *ring;
+	int r, i;
+	struct amdgpu_device *adev = ip_block->adev;
+	uint32_t reg_count = ARRAY_SIZE(sdma_reg_list_7_0);
+	uint32_t *ptr;
+
+	/* SDMA trap event */
+	r = amdgpu_irq_add_id(adev, SOC21_IH_CLIENTID_GFX,
+			      GFX_12_0_0__SRCID__SDMA_TRAP,
+			      &adev->sdma.trap_irq);
+	if (r)
+		return r;
+
+	/* SDMA user fence event */
+	r = amdgpu_irq_add_id(adev, SOC21_IH_CLIENTID_GFX,
+			      GFX_12_0_0__SRCID__SDMA_FENCE,
+			      &adev->sdma.fence_irq);
+	if (r)
+		return r;
+
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		ring = &adev->sdma.instance[i].ring;
+		ring->ring_obj = NULL;
+		ring->use_doorbell = true;
+		ring->me = i;
+		ring->no_user_submission = adev->sdma.no_user_submission;
+
+		DRM_DEBUG("SDMA %d use_doorbell being set to: [%s]\n", i,
+				ring->use_doorbell?"true":"false");
+
+		ring->doorbell_index =
+			(adev->doorbell_index.sdma_engine[i] << 1); // get DWORD offset
+
+		ring->vm_hub = AMDGPU_GFXHUB(0);
+		sprintf(ring->name, "sdma%d", i);
+		r = amdgpu_ring_init(adev, ring, 1024,
+				     &adev->sdma.trap_irq,
+				     AMDGPU_SDMA_IRQ_INSTANCE0 + i,
+				     AMDGPU_RING_PRIO_DEFAULT, NULL);
+		if (r)
+			return r;
+	}
+
+	adev->sdma.supported_reset =
+		amdgpu_get_soft_full_reset_mask(&adev->sdma.instance[0].ring);
+	if (!amdgpu_sriov_vf(adev) &&
+	    !adev->debug_disable_gpu_ring_reset)
+		adev->sdma.supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE;
+
+	r = amdgpu_sdma_sysfs_reset_mask_init(adev);
+	if (r)
+		return r;
+	/* Allocate memory for SDMA IP Dump buffer */
+	ptr = kcalloc(adev->sdma.num_instances * reg_count, sizeof(uint32_t), GFP_KERNEL);
+	if (ptr)
+		adev->sdma.ip_dump = ptr;
+	else
+		DRM_ERROR("Failed to allocated memory for SDMA IP Dump\n");
+
+	switch (amdgpu_ip_version(adev, SDMA0_HWIP, 0)) {
+	case IP_VERSION(7, 0, 0):
+	case IP_VERSION(7, 0, 1):
+		if ((adev->sdma.instance[0].fw_version >= 7966358) && !adev->sdma.disable_uq)
+			adev->userq_funcs[AMDGPU_HW_IP_DMA] = &userq_mes_funcs;
+		break;
+	default:
+		break;
+	}
+
+	return r;
+}
+
+static int sdma_v7_0_sw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i;
+
+	for (i = 0; i < adev->sdma.num_instances; i++)
+		amdgpu_ring_fini(&adev->sdma.instance[i].ring);
+
+	amdgpu_sdma_sysfs_reset_mask_fini(adev);
+	amdgpu_sdma_destroy_inst_ctx(adev, true);
+
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT)
+		sdma_v12_0_free_ucode_buffer(adev);
+
+	kfree(adev->sdma.ip_dump);
+
+	return 0;
+}
+
+static int sdma_v7_0_set_userq_trap_interrupts(struct amdgpu_device *adev,
+					       bool enable)
+{
+	unsigned int irq_type;
+	int i, r;
+
+	if (adev->userq_funcs[AMDGPU_HW_IP_DMA]) {
+		for (i = 0; i < adev->sdma.num_instances; i++) {
+			irq_type = AMDGPU_SDMA_IRQ_INSTANCE0 + i;
+			if (enable)
+				r = amdgpu_irq_get(adev, &adev->sdma.trap_irq,
+						   irq_type);
+			else
+				r = amdgpu_irq_put(adev, &adev->sdma.trap_irq,
+						   irq_type);
+			if (r)
+				return r;
+		}
+	}
+
+	return 0;
+}
+
+static int sdma_v7_0_hw_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int r;
+
+	r = sdma_v7_0_start(adev);
+	if (r)
+		return r;
+
+	return sdma_v7_0_set_userq_trap_interrupts(adev, true);
+}
+
+static int sdma_v7_0_hw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	if (amdgpu_sriov_vf(adev))
+		return 0;
+
+	sdma_v7_0_ctx_switch_enable(adev, false);
+	sdma_v7_0_enable(adev, false);
+	sdma_v7_0_set_userq_trap_interrupts(adev, false);
+
+	return 0;
+}
+
+static int sdma_v7_0_suspend(struct amdgpu_ip_block *ip_block)
+{
+	return sdma_v7_0_hw_fini(ip_block);
+}
+
+static int sdma_v7_0_resume(struct amdgpu_ip_block *ip_block)
+{
+	return sdma_v7_0_hw_init(ip_block);
+}
+
+static bool sdma_v7_0_is_idle(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	u32 i;
+
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		u32 tmp = RREG32(sdma_v7_0_get_reg_offset(adev, i, regSDMA0_STATUS_REG));
+
+		if (!(tmp & SDMA0_STATUS_REG__IDLE_MASK))
+			return false;
+	}
+
+	return true;
+}
+
+static int sdma_v7_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
+{
+	unsigned i;
+	u32 sdma0, sdma1;
+	struct amdgpu_device *adev = ip_block->adev;
+
+	for (i = 0; i < adev->usec_timeout; i++) {
+		sdma0 = RREG32(sdma_v7_0_get_reg_offset(adev, 0, regSDMA0_STATUS_REG));
+		sdma1 = RREG32(sdma_v7_0_get_reg_offset(adev, 1, regSDMA0_STATUS_REG));
+
+		if (sdma0 & sdma1 & SDMA0_STATUS_REG__IDLE_MASK)
+			return 0;
+		udelay(1);
+	}
+	return -ETIMEDOUT;
+}
+
+static int sdma_v7_0_ring_preempt_ib(struct amdgpu_ring *ring)
+{
+	int i, r = 0;
+	struct amdgpu_device *adev = ring->adev;
+	u32 index = 0;
+	u64 sdma_gfx_preempt;
+
+	amdgpu_sdma_get_index_from_ring(ring, &index);
+	sdma_gfx_preempt =
+		sdma_v7_0_get_reg_offset(adev, index, regSDMA0_QUEUE0_PREEMPT);
+
+	/* assert preemption condition */
+	amdgpu_ring_set_preempt_cond_exec(ring, false);
+
+	/* emit the trailing fence */
+	ring->trail_seq += 1;
+	r = amdgpu_ring_alloc(ring, 10);
+	if (r) {
+		DRM_ERROR("ring %d failed to be allocated \n", ring->idx);
+		return r;
+	}
+	sdma_v7_0_ring_emit_fence(ring, ring->trail_fence_gpu_addr,
+				  ring->trail_seq, 0);
+	amdgpu_ring_commit(ring);
+
+	/* assert IB preemption */
+	WREG32(sdma_gfx_preempt, 1);
+
+	/* poll the trailing fence */
+	for (i = 0; i < adev->usec_timeout; i++) {
+		if (ring->trail_seq ==
+		    le32_to_cpu(*(ring->trail_fence_cpu_addr)))
+			break;
+		udelay(1);
+	}
+
+	if (i >= adev->usec_timeout) {
+		r = -EINVAL;
+		DRM_ERROR("ring %d failed to be preempted\n", ring->idx);
+	}
+
+	/* deassert IB preemption */
+	WREG32(sdma_gfx_preempt, 0);
+
+	/* deassert the preemption condition */
+	amdgpu_ring_set_preempt_cond_exec(ring, true);
+	return r;
+}
+
+static int sdma_v7_0_set_trap_irq_state(struct amdgpu_device *adev,
+					struct amdgpu_irq_src *source,
+					unsigned type,
+					enum amdgpu_interrupt_state state)
+{
+	u32 sdma_cntl;
+
+	u32 reg_offset = sdma_v7_0_get_reg_offset(adev, type, regSDMA0_CNTL);
+
+	sdma_cntl = RREG32(reg_offset);
+	sdma_cntl = REG_SET_FIELD(sdma_cntl, SDMA0_CNTL, TRAP_ENABLE,
+		       state == AMDGPU_IRQ_STATE_ENABLE ? 1 : 0);
+	WREG32(reg_offset, sdma_cntl);
+
+	return 0;
+}
+
+static int sdma_v7_0_process_trap_irq(struct amdgpu_device *adev,
+				      struct amdgpu_irq_src *source,
+				      struct amdgpu_iv_entry *entry)
+{
+	int instances, queue;
+
+	DRM_DEBUG("IH: SDMA trap\n");
+
+	queue = entry->ring_id & 0xf;
+	instances = (entry->ring_id & 0xf0) >> 4;
+	if (instances > 1) {
+		DRM_ERROR("IH: wrong ring_ID detected, as wrong sdma instance\n");
+		return -EINVAL;
+	}
+
+	switch (entry->client_id) {
+	case SOC21_IH_CLIENTID_GFX:
+		switch (queue) {
+		case 0:
+			amdgpu_fence_process(&adev->sdma.instance[instances].ring);
+			break;
+		default:
+			break;
+		}
+		break;
+	}
+	return 0;
+}
+
+static int sdma_v7_0_process_fence_irq(struct amdgpu_device *adev,
+				       struct amdgpu_irq_src *source,
+				       struct amdgpu_iv_entry *entry)
+{
+	u32 doorbell_offset = entry->src_data[0];
+
+	if (adev->enable_mes && doorbell_offset) {
+		struct amdgpu_userq_fence_driver *fence_drv = NULL;
+		struct xarray *xa = &adev->userq_xa;
+		unsigned long flags;
+
+		doorbell_offset >>= SDMA0_QUEUE0_DOORBELL_OFFSET__OFFSET__SHIFT;
+
+		xa_lock_irqsave(xa, flags);
+		fence_drv = xa_load(xa, doorbell_offset);
+		if (fence_drv)
+			amdgpu_userq_fence_driver_process(fence_drv);
+		xa_unlock_irqrestore(xa, flags);
+	}
+
+	return 0;
+}
+
+static int sdma_v7_0_process_illegal_inst_irq(struct amdgpu_device *adev,
+					      struct amdgpu_irq_src *source,
+					      struct amdgpu_iv_entry *entry)
+{
+	return 0;
+}
+
+static int sdma_v7_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
+					   enum amd_clockgating_state state)
+{
+	return 0;
+}
+
+static int sdma_v7_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
+					  enum amd_powergating_state state)
+{
+	return 0;
+}
+
+static void sdma_v7_0_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags)
+{
+}
+
+static void sdma_v7_0_print_ip_state(struct amdgpu_ip_block *ip_block, struct drm_printer *p)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, j;
+	uint32_t reg_count = ARRAY_SIZE(sdma_reg_list_7_0);
+	uint32_t instance_offset;
+
+	if (!adev->sdma.ip_dump)
+		return;
+
+	drm_printf(p, "num_instances:%d\n", adev->sdma.num_instances);
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		instance_offset = i * reg_count;
+		drm_printf(p, "\nInstance:%d\n", i);
+
+		for (j = 0; j < reg_count; j++)
+			drm_printf(p, "%-50s \t 0x%08x\n", sdma_reg_list_7_0[j].reg_name,
+				   adev->sdma.ip_dump[instance_offset + j]);
+	}
+}
+
+static void sdma_v7_0_dump_ip_state(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, j;
+	uint32_t instance_offset;
+	uint32_t reg_count = ARRAY_SIZE(sdma_reg_list_7_0);
+
+	if (!adev->sdma.ip_dump)
+		return;
+
+	amdgpu_gfx_off_ctrl(adev, false);
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		instance_offset = i * reg_count;
+		for (j = 0; j < reg_count; j++)
+			adev->sdma.ip_dump[instance_offset + j] =
+				RREG32(sdma_v7_0_get_reg_offset(adev, i,
+				       sdma_reg_list_7_0[j].reg_offset));
+	}
+	amdgpu_gfx_off_ctrl(adev, true);
+}
+
+const struct amd_ip_funcs sdma_v7_0_ip_funcs = {
+	.name = "sdma_v7_0",
+	.early_init = sdma_v7_0_early_init,
+	.late_init = NULL,
+	.sw_init = sdma_v7_0_sw_init,
+	.sw_fini = sdma_v7_0_sw_fini,
+	.hw_init = sdma_v7_0_hw_init,
+	.hw_fini = sdma_v7_0_hw_fini,
+	.suspend = sdma_v7_0_suspend,
+	.resume = sdma_v7_0_resume,
+	.is_idle = sdma_v7_0_is_idle,
+	.wait_for_idle = sdma_v7_0_wait_for_idle,
+	.soft_reset = sdma_v7_0_soft_reset,
+	.check_soft_reset = sdma_v7_0_check_soft_reset,
+	.set_clockgating_state = sdma_v7_0_set_clockgating_state,
+	.set_powergating_state = sdma_v7_0_set_powergating_state,
+	.get_clockgating_state = sdma_v7_0_get_clockgating_state,
+	.dump_ip_state = sdma_v7_0_dump_ip_state,
+	.print_ip_state = sdma_v7_0_print_ip_state,
+};
+
+static const struct amdgpu_ring_funcs sdma_v7_0_ring_funcs = {
+	.type = AMDGPU_RING_TYPE_SDMA,
+	.align_mask = 0xf,
+	.nop = SDMA_PKT_NOP_HEADER_OP(SDMA_OP_NOP),
+	.support_64bit_ptrs = true,
+	.secure_submission_supported = true,
+	.get_rptr = sdma_v7_0_ring_get_rptr,
+	.get_wptr = sdma_v7_0_ring_get_wptr,
+	.set_wptr = sdma_v7_0_ring_set_wptr,
+	.emit_frame_size =
+		5 + /* sdma_v7_0_ring_init_cond_exec */
+		6 + /* sdma_v7_0_ring_emit_hdp_flush */
+		6 + /* sdma_v7_0_ring_emit_pipeline_sync */
+		/* sdma_v7_0_ring_emit_vm_flush */
+		SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 +
+		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 6 +
+		10 + 10 + 10, /* sdma_v7_0_ring_emit_fence x3 for user fence, vm fence */
+	.emit_ib_size = 5 + 7 + 6, /* sdma_v7_0_ring_emit_ib */
+	.emit_ib = sdma_v7_0_ring_emit_ib,
+	.emit_mem_sync = sdma_v7_0_ring_emit_mem_sync,
+	.emit_fence = sdma_v7_0_ring_emit_fence,
+	.emit_pipeline_sync = sdma_v7_0_ring_emit_pipeline_sync,
+	.emit_vm_flush = sdma_v7_0_ring_emit_vm_flush,
+	.emit_hdp_flush = sdma_v7_0_ring_emit_hdp_flush,
+	.test_ring = sdma_v7_0_ring_test_ring,
+	.test_ib = sdma_v7_0_ring_test_ib,
+	.insert_nop = sdma_v7_0_ring_insert_nop,
+	.pad_ib = sdma_v7_0_ring_pad_ib,
+	.emit_wreg = sdma_v7_0_ring_emit_wreg,
+	.emit_reg_wait = sdma_v7_0_ring_emit_reg_wait,
+	.emit_reg_write_reg_wait = sdma_v7_0_ring_emit_reg_write_reg_wait,
+	.init_cond_exec = sdma_v7_0_ring_init_cond_exec,
+	.preempt_ib = sdma_v7_0_ring_preempt_ib,
+	.reset = sdma_v7_0_reset_queue,
+};
+
+static void sdma_v7_0_set_ring_funcs(struct amdgpu_device *adev)
+{
+	int i;
+
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		adev->sdma.instance[i].ring.funcs = &sdma_v7_0_ring_funcs;
+		adev->sdma.instance[i].ring.me = i;
+	}
+}
+
+static const struct amdgpu_irq_src_funcs sdma_v7_0_trap_irq_funcs = {
+	.set = sdma_v7_0_set_trap_irq_state,
+	.process = sdma_v7_0_process_trap_irq,
+};
+
+static const struct amdgpu_irq_src_funcs sdma_v7_0_fence_irq_funcs = {
+	.process = sdma_v7_0_process_fence_irq,
+};
+
+static const struct amdgpu_irq_src_funcs sdma_v7_0_illegal_inst_irq_funcs = {
+	.process = sdma_v7_0_process_illegal_inst_irq,
+};
+
+static void sdma_v7_0_set_irq_funcs(struct amdgpu_device *adev)
+{
+	adev->sdma.trap_irq.num_types = AMDGPU_SDMA_IRQ_INSTANCE0 +
+					adev->sdma.num_instances;
+	adev->sdma.trap_irq.funcs = &sdma_v7_0_trap_irq_funcs;
+	adev->sdma.fence_irq.funcs = &sdma_v7_0_fence_irq_funcs;
+	adev->sdma.illegal_inst_irq.funcs = &sdma_v7_0_illegal_inst_irq_funcs;
+}
+
+/**
+ * sdma_v7_0_emit_copy_buffer - copy buffer using the sDMA engine
+ *
+ * @ib: indirect buffer to fill with commands
+ * @src_offset: src GPU address
+ * @dst_offset: dst GPU address
+ * @byte_count: number of bytes to xfer
+ * @copy_flags: copy flags for the buffers
+ *
+ * Copy GPU buffers using the DMA engine.
+ * Used by the amdgpu ttm implementation to move pages if
+ * registered as the asic copy callback.
+ */
+static void sdma_v7_0_emit_copy_buffer(struct amdgpu_ib *ib,
+				       uint64_t src_offset,
+				       uint64_t dst_offset,
+				       uint32_t byte_count,
+				       uint32_t copy_flags)
+{
+	uint32_t num_type, data_format, max_com, write_cm;
+
+	max_com = AMDGPU_COPY_FLAGS_GET(copy_flags, MAX_COMPRESSED);
+	data_format = AMDGPU_COPY_FLAGS_GET(copy_flags, DATA_FORMAT);
+	num_type = AMDGPU_COPY_FLAGS_GET(copy_flags, NUMBER_TYPE);
+	write_cm = AMDGPU_COPY_FLAGS_GET(copy_flags, WRITE_COMPRESS_DISABLE) ? 2 : 1;
+
+	ib->ptr[ib->length_dw++] = SDMA_PKT_COPY_LINEAR_HEADER_OP(SDMA_OP_COPY) |
+		SDMA_PKT_COPY_LINEAR_HEADER_SUB_OP(SDMA_SUBOP_COPY_LINEAR) |
+		SDMA_PKT_COPY_LINEAR_HEADER_TMZ((copy_flags & AMDGPU_COPY_FLAGS_TMZ) ? 1 : 0) |
+		SDMA_PKT_COPY_LINEAR_HEADER_CPV(1);
+
+	ib->ptr[ib->length_dw++] = byte_count - 1;
+	ib->ptr[ib->length_dw++] = 0; /* src/dst endian swap */
+	ib->ptr[ib->length_dw++] = lower_32_bits(src_offset);
+	ib->ptr[ib->length_dw++] = upper_32_bits(src_offset);
+	ib->ptr[ib->length_dw++] = lower_32_bits(dst_offset);
+	ib->ptr[ib->length_dw++] = upper_32_bits(dst_offset);
+
+	if ((copy_flags & (AMDGPU_COPY_FLAGS_READ_DECOMPRESSED | AMDGPU_COPY_FLAGS_WRITE_COMPRESSED)))
+		ib->ptr[ib->length_dw++] = SDMA_DCC_DATA_FORMAT(data_format) | SDMA_DCC_NUM_TYPE(num_type) |
+			((copy_flags & AMDGPU_COPY_FLAGS_READ_DECOMPRESSED) ? SDMA_DCC_READ_CM(2) : 0) |
+			((copy_flags & AMDGPU_COPY_FLAGS_WRITE_COMPRESSED) ? SDMA_DCC_WRITE_CM(write_cm) : 0) |
+			SDMA_DCC_MAX_COM(max_com) | SDMA_DCC_MAX_UCOM(1);
+	else
+		ib->ptr[ib->length_dw++] = 0;
+}
+
+/**
+ * sdma_v7_0_emit_fill_buffer - fill buffer using the sDMA engine
+ *
+ * @ib: indirect buffer to fill
+ * @src_data: value to write to buffer
+ * @dst_offset: dst GPU address
+ * @byte_count: number of bytes to xfer
+ *
+ * Fill GPU buffers using the DMA engine.
+ */
+static void sdma_v7_0_emit_fill_buffer(struct amdgpu_ib *ib,
+				       uint32_t src_data,
+				       uint64_t dst_offset,
+				       uint32_t byte_count)
+{
+	ib->ptr[ib->length_dw++] = SDMA_PKT_CONSTANT_FILL_HEADER_OP(SDMA_OP_CONST_FILL) |
+		SDMA_PKT_CONSTANT_FILL_HEADER_COMPRESS(1);
+	ib->ptr[ib->length_dw++] = lower_32_bits(dst_offset);
+	ib->ptr[ib->length_dw++] = upper_32_bits(dst_offset);
+	ib->ptr[ib->length_dw++] = src_data;
+	ib->ptr[ib->length_dw++] = byte_count - 1;
+}
+
+static const struct amdgpu_buffer_funcs sdma_v7_0_buffer_funcs = {
+	.copy_max_bytes = 0x400000,
+	.copy_num_dw = 8,
+	.emit_copy_buffer = sdma_v7_0_emit_copy_buffer,
+	.fill_max_bytes = 0x400000,
+	.fill_num_dw = 5,
+	.emit_fill_buffer = sdma_v7_0_emit_fill_buffer,
+};
+
+static void sdma_v7_0_set_buffer_funcs(struct amdgpu_device *adev)
+{
+	adev->mman.buffer_funcs = &sdma_v7_0_buffer_funcs;
+	adev->mman.buffer_funcs_ring = &adev->sdma.instance[0].ring;
+}
+
+static const struct amdgpu_vm_pte_funcs sdma_v7_0_vm_pte_funcs = {
+	.copy_pte_num_dw = 8,
+	.copy_pte = sdma_v7_0_vm_copy_pte,
+	.write_pte = sdma_v7_0_vm_write_pte,
+	.set_pte_pde = sdma_v7_0_vm_set_pte_pde,
+};
+
+static void sdma_v7_0_set_vm_pte_funcs(struct amdgpu_device *adev)
+{
+	unsigned i;
+
+	adev->vm_manager.vm_pte_funcs = &sdma_v7_0_vm_pte_funcs;
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		adev->vm_manager.vm_pte_scheds[i] =
+			&adev->sdma.instance[i].ring.sched;
+	}
+	adev->vm_manager.vm_pte_num_scheds = adev->sdma.num_instances;
+}
+
+const struct amdgpu_ip_block_version sdma_v7_0_ip_block = {
+	.type = AMD_IP_BLOCK_TYPE_SDMA,
+	.major = 7,
+	.minor = 0,
+	.rev = 0,
+	.funcs = &sdma_v7_0_ip_funcs,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v7_0.h b/drivers/gpu/drm/amd/amdgpu/sdma_v7_0.h
new file mode 100644
index 000000000000..5af863bb39c4
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v7_0.h
@@ -0,0 +1,30 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __SDMA_V7_0_H__
+#define __SDMA_V7_0_H__
+
+extern const struct amd_ip_funcs sdma_v7_0_ip_funcs;
+extern const struct amdgpu_ip_block_version sdma_v7_0_ip_block;
+
+#endif /* __SDMA_V7_0_H__ */
diff --git a/drivers/gpu/drm/amd/amdgpu/si.c b/drivers/gpu/drm/amd/amdgpu/si.c
index e6d2f74a7976..f7288372ee61 100644
--- a/drivers/gpu/drm/amd/amdgpu/si.c
+++ b/drivers/gpu/drm/amd/amdgpu/si.c
@@ -35,6 +35,7 @@
 #include "amdgpu_vce.h"
 #include "atom.h"
 #include "amd_pcie.h"
+
 #include "si_dpm.h"
 #include "sid.h"
 #include "si_ih.h"
@@ -44,17 +45,31 @@
 #include "dce_v6_0.h"
 #include "si.h"
 #include "uvd_v3_1.h"
-#include "amdgpu_vkms.h"
+#include "vce_v1_0.h"
+
+#include "uvd/uvd_4_0_d.h"
+
+#include "smu/smu_6_0_d.h"
+#include "smu/smu_6_0_sh_mask.h"
+
 #include "gca/gfx_6_0_d.h"
+#include "gca/gfx_6_0_sh_mask.h"
+
 #include "oss/oss_1_0_d.h"
 #include "oss/oss_1_0_sh_mask.h"
+
 #include "gmc/gmc_6_0_d.h"
+#include"gmc/gmc_6_0_sh_mask.h"
+
 #include "dce/dce_6_0_d.h"
-#include "uvd/uvd_4_0_d.h"
+#include "dce/dce_6_0_sh_mask.h"
+
 #include "bif/bif_3_0_d.h"
 #include "bif/bif_3_0_sh_mask.h"
+#include "si_enums.h"
 
 #include "amdgpu_dm.h"
+#include "amdgpu_vkms.h"
 
 static const u32 tahiti_golden_registers[] =
 {
@@ -907,9 +922,7 @@ static const u32 hainan_mgcg_cgcg_init[] =
 	0x3630, 0xfffffff0, 0x00000100,
 };
 
-/* XXX: update when we support VCE */
-#if 0
-/* tahiti, pitcarin, verde */
+/* tahiti, pitcairn, verde */
 static const struct amdgpu_video_codec_info tahiti_video_codecs_encode_array[] =
 {
 	{
@@ -926,13 +939,7 @@ static const struct amdgpu_video_codecs tahiti_video_codecs_encode =
 	.codec_count = ARRAY_SIZE(tahiti_video_codecs_encode_array),
 	.codec_array = tahiti_video_codecs_encode_array,
 };
-#else
-static const struct amdgpu_video_codecs tahiti_video_codecs_encode =
-{
-	.codec_count = 0,
-	.codec_array = NULL,
-};
-#endif
+
 /* oland and hainan don't support encode */
 static const struct amdgpu_video_codecs hainan_video_codecs_encode =
 {
@@ -940,7 +947,7 @@ static const struct amdgpu_video_codecs hainan_video_codecs_encode =
 	.codec_array = NULL,
 };
 
-/* tahiti, pitcarin, verde, oland */
+/* tahiti, pitcairn, verde, oland */
 static const struct amdgpu_video_codec_info tahiti_video_codecs_decode_array[] =
 {
 	{
@@ -1071,8 +1078,8 @@ static u32 si_smc_rreg(struct amdgpu_device *adev, u32 reg)
 	u32 r;
 
 	spin_lock_irqsave(&adev->smc_idx_lock, flags);
-	WREG32(SMC_IND_INDEX_0, (reg));
-	r = RREG32(SMC_IND_DATA_0);
+	WREG32(mmSMC_IND_INDEX_0, (reg));
+	r = RREG32(mmSMC_IND_DATA_0);
 	spin_unlock_irqrestore(&adev->smc_idx_lock, flags);
 	return r;
 }
@@ -1082,8 +1089,8 @@ static void si_smc_wreg(struct amdgpu_device *adev, u32 reg, u32 v)
 	unsigned long flags;
 
 	spin_lock_irqsave(&adev->smc_idx_lock, flags);
-	WREG32(SMC_IND_INDEX_0, (reg));
-	WREG32(SMC_IND_DATA_0, (v));
+	WREG32(mmSMC_IND_INDEX_0, (reg));
+	WREG32(mmSMC_IND_DATA_0, (v));
 	spin_unlock_irqrestore(&adev->smc_idx_lock, flags);
 }
 
@@ -1110,55 +1117,55 @@ static void si_uvd_ctx_wreg(struct amdgpu_device *adev, u32 reg, u32 v)
 }
 
 static struct amdgpu_allowed_register_entry si_allowed_read_registers[] = {
-	{GRBM_STATUS},
+	{mmGRBM_STATUS},
 	{mmGRBM_STATUS2},
 	{mmGRBM_STATUS_SE0},
 	{mmGRBM_STATUS_SE1},
 	{mmSRBM_STATUS},
 	{mmSRBM_STATUS2},
-	{DMA_STATUS_REG + DMA0_REGISTER_OFFSET},
-	{DMA_STATUS_REG + DMA1_REGISTER_OFFSET},
+	{mmDMA_STATUS_REG + DMA0_REGISTER_OFFSET},
+	{mmDMA_STATUS_REG + DMA1_REGISTER_OFFSET},
 	{mmCP_STAT},
 	{mmCP_STALLED_STAT1},
 	{mmCP_STALLED_STAT2},
 	{mmCP_STALLED_STAT3},
-	{GB_ADDR_CONFIG},
-	{MC_ARB_RAMCFG},
-	{GB_TILE_MODE0},
-	{GB_TILE_MODE1},
-	{GB_TILE_MODE2},
-	{GB_TILE_MODE3},
-	{GB_TILE_MODE4},
-	{GB_TILE_MODE5},
-	{GB_TILE_MODE6},
-	{GB_TILE_MODE7},
-	{GB_TILE_MODE8},
-	{GB_TILE_MODE9},
-	{GB_TILE_MODE10},
-	{GB_TILE_MODE11},
-	{GB_TILE_MODE12},
-	{GB_TILE_MODE13},
-	{GB_TILE_MODE14},
-	{GB_TILE_MODE15},
-	{GB_TILE_MODE16},
-	{GB_TILE_MODE17},
-	{GB_TILE_MODE18},
-	{GB_TILE_MODE19},
-	{GB_TILE_MODE20},
-	{GB_TILE_MODE21},
-	{GB_TILE_MODE22},
-	{GB_TILE_MODE23},
-	{GB_TILE_MODE24},
-	{GB_TILE_MODE25},
-	{GB_TILE_MODE26},
-	{GB_TILE_MODE27},
-	{GB_TILE_MODE28},
-	{GB_TILE_MODE29},
-	{GB_TILE_MODE30},
-	{GB_TILE_MODE31},
-	{CC_RB_BACKEND_DISABLE, true},
-	{GC_USER_RB_BACKEND_DISABLE, true},
-	{PA_SC_RASTER_CONFIG, true},
+	{mmGB_ADDR_CONFIG},
+	{mmMC_ARB_RAMCFG},
+	{mmGB_TILE_MODE0},
+	{mmGB_TILE_MODE1},
+	{mmGB_TILE_MODE2},
+	{mmGB_TILE_MODE3},
+	{mmGB_TILE_MODE4},
+	{mmGB_TILE_MODE5},
+	{mmGB_TILE_MODE6},
+	{mmGB_TILE_MODE7},
+	{mmGB_TILE_MODE8},
+	{mmGB_TILE_MODE9},
+	{mmGB_TILE_MODE10},
+	{mmGB_TILE_MODE11},
+	{mmGB_TILE_MODE12},
+	{mmGB_TILE_MODE13},
+	{mmGB_TILE_MODE14},
+	{mmGB_TILE_MODE15},
+	{mmGB_TILE_MODE16},
+	{mmGB_TILE_MODE17},
+	{mmGB_TILE_MODE18},
+	{mmGB_TILE_MODE19},
+	{mmGB_TILE_MODE20},
+	{mmGB_TILE_MODE21},
+	{mmGB_TILE_MODE22},
+	{mmGB_TILE_MODE23},
+	{mmGB_TILE_MODE24},
+	{mmGB_TILE_MODE25},
+	{mmGB_TILE_MODE26},
+	{mmGB_TILE_MODE27},
+	{mmGB_TILE_MODE28},
+	{mmGB_TILE_MODE29},
+	{mmGB_TILE_MODE30},
+	{mmGB_TILE_MODE31},
+	{mmCC_RB_BACKEND_DISABLE, true},
+	{mmGC_USER_RB_BACKEND_DISABLE, true},
+	{mmPA_SC_RASTER_CONFIG, true},
 };
 
 static uint32_t si_get_register_value(struct amdgpu_device *adev,
@@ -1181,12 +1188,12 @@ static uint32_t si_get_register_value(struct amdgpu_device *adev,
 
 		mutex_lock(&adev->grbm_idx_mutex);
 		if (se_num != 0xffffffff || sh_num != 0xffffffff)
-			amdgpu_gfx_select_se_sh(adev, se_num, sh_num, 0xffffffff);
+			amdgpu_gfx_select_se_sh(adev, se_num, sh_num, 0xffffffff, 0);
 
 		val = RREG32(reg_offset);
 
 		if (se_num != 0xffffffff || sh_num != 0xffffffff)
-			amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+			amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
 		mutex_unlock(&adev->grbm_idx_mutex);
 		return val;
 	} else {
@@ -1264,37 +1271,37 @@ static bool si_read_disabled_bios(struct amdgpu_device *adev)
 	u32 rom_cntl;
 	bool r;
 
-	bus_cntl = RREG32(R600_BUS_CNTL);
+	bus_cntl = RREG32(mmBUS_CNTL);
 	if (adev->mode_info.num_crtc) {
-		d1vga_control = RREG32(AVIVO_D1VGA_CONTROL);
-		d2vga_control = RREG32(AVIVO_D2VGA_CONTROL);
-		vga_render_control = RREG32(VGA_RENDER_CONTROL);
+		d1vga_control = RREG32(mmD1VGA_CONTROL);
+		d2vga_control = RREG32(mmD2VGA_CONTROL);
+		vga_render_control = RREG32(mmVGA_RENDER_CONTROL);
 	}
 	rom_cntl = RREG32(R600_ROM_CNTL);
 
 	/* enable the rom */
-	WREG32(R600_BUS_CNTL, (bus_cntl & ~R600_BIOS_ROM_DIS));
+	WREG32(mmBUS_CNTL, (bus_cntl & ~BUS_CNTL__BIOS_ROM_DIS_MASK));
 	if (adev->mode_info.num_crtc) {
 		/* Disable VGA mode */
-		WREG32(AVIVO_D1VGA_CONTROL,
-		       (d1vga_control & ~(AVIVO_DVGA_CONTROL_MODE_ENABLE |
-					  AVIVO_DVGA_CONTROL_TIMING_SELECT)));
-		WREG32(AVIVO_D2VGA_CONTROL,
-		       (d2vga_control & ~(AVIVO_DVGA_CONTROL_MODE_ENABLE |
-					  AVIVO_DVGA_CONTROL_TIMING_SELECT)));
-		WREG32(VGA_RENDER_CONTROL,
-		       (vga_render_control & C_000300_VGA_VSTATUS_CNTL));
+		WREG32(mmD1VGA_CONTROL,
+		       (d1vga_control & ~(D1VGA_CONTROL__D1VGA_MODE_ENABLE_MASK |
+					  D1VGA_CONTROL__D1VGA_TIMING_SELECT_MASK)));
+		WREG32(mmD2VGA_CONTROL,
+		       (d2vga_control & ~(D1VGA_CONTROL__D1VGA_MODE_ENABLE_MASK |
+					  D1VGA_CONTROL__D1VGA_TIMING_SELECT_MASK)));
+		WREG32(mmVGA_RENDER_CONTROL,
+		       (vga_render_control & ~VGA_RENDER_CONTROL__VGA_VSTATUS_CNTL_MASK));
 	}
 	WREG32(R600_ROM_CNTL, rom_cntl | R600_SCK_OVERWRITE);
 
 	r = amdgpu_read_bios(adev);
 
 	/* restore regs */
-	WREG32(R600_BUS_CNTL, bus_cntl);
+	WREG32(mmBUS_CNTL, bus_cntl);
 	if (adev->mode_info.num_crtc) {
-		WREG32(AVIVO_D1VGA_CONTROL, d1vga_control);
-		WREG32(AVIVO_D2VGA_CONTROL, d2vga_control);
-		WREG32(VGA_RENDER_CONTROL, vga_render_control);
+		WREG32(mmD1VGA_CONTROL, d1vga_control);
+		WREG32(mmD2VGA_CONTROL, d2vga_control);
+		WREG32(mmVGA_RENDER_CONTROL, vga_render_control);
 	}
 	WREG32(R600_ROM_CNTL, rom_cntl);
 	return r;
@@ -1331,23 +1338,24 @@ static void si_set_clk_bypass_mode(struct amdgpu_device *adev)
 {
 	u32 tmp, i;
 
-	tmp = RREG32(CG_SPLL_FUNC_CNTL);
-	tmp |= SPLL_BYPASS_EN;
-	WREG32(CG_SPLL_FUNC_CNTL, tmp);
+	tmp = RREG32(mmCG_SPLL_FUNC_CNTL);
+	tmp |= CG_SPLL_FUNC_CNTL__SPLL_BYPASS_EN_MASK;
+	WREG32(mmCG_SPLL_FUNC_CNTL, tmp);
 
-	tmp = RREG32(CG_SPLL_FUNC_CNTL_2);
-	tmp |= SPLL_CTLREQ_CHG;
-	WREG32(CG_SPLL_FUNC_CNTL_2, tmp);
+	tmp = RREG32(mmCG_SPLL_FUNC_CNTL_2);
+	tmp |= CG_SPLL_FUNC_CNTL_2__SPLL_CTLREQ_CHG_MASK;
+	WREG32(mmCG_SPLL_FUNC_CNTL_2, tmp);
 
 	for (i = 0; i < adev->usec_timeout; i++) {
-		if (RREG32(SPLL_STATUS) & SPLL_CHG_STATUS)
+		if (RREG32(mmCG_SPLL_STATUS) & CG_SPLL_STATUS__SPLL_CHG_STATUS_MASK)
 			break;
 		udelay(1);
 	}
 
-	tmp = RREG32(CG_SPLL_FUNC_CNTL_2);
-	tmp &= ~(SPLL_CTLREQ_CHG | SCLK_MUX_UPDATE);
-	WREG32(CG_SPLL_FUNC_CNTL_2, tmp);
+	tmp = RREG32(mmCG_SPLL_FUNC_CNTL_2);
+	tmp &= ~(CG_SPLL_FUNC_CNTL_2__SPLL_CTLREQ_CHG_MASK |
+		CG_SPLL_FUNC_CNTL_2__SCLK_MUX_UPDATE_MASK);
+	WREG32(mmCG_SPLL_FUNC_CNTL_2, tmp);
 
 	tmp = RREG32(MPLL_CNTL_MODE);
 	tmp &= ~MPLL_MCLK_SEL;
@@ -1358,21 +1366,21 @@ static void si_spll_powerdown(struct amdgpu_device *adev)
 {
 	u32 tmp;
 
-	tmp = RREG32(SPLL_CNTL_MODE);
-	tmp |= SPLL_SW_DIR_CONTROL;
-	WREG32(SPLL_CNTL_MODE, tmp);
+	tmp = RREG32(mmSPLL_CNTL_MODE);
+	tmp |= SPLL_CNTL_MODE__SPLL_SW_DIR_CONTROL_MASK;
+	WREG32(mmSPLL_CNTL_MODE, tmp);
 
-	tmp = RREG32(CG_SPLL_FUNC_CNTL);
-	tmp |= SPLL_RESET;
-	WREG32(CG_SPLL_FUNC_CNTL, tmp);
+	tmp = RREG32(mmCG_SPLL_FUNC_CNTL);
+	tmp |= CG_SPLL_FUNC_CNTL__SPLL_RESET_MASK;
+	WREG32(mmCG_SPLL_FUNC_CNTL, tmp);
 
-	tmp = RREG32(CG_SPLL_FUNC_CNTL);
-	tmp |= SPLL_SLEEP;
-	WREG32(CG_SPLL_FUNC_CNTL, tmp);
+	tmp = RREG32(mmCG_SPLL_FUNC_CNTL);
+	tmp |= CG_SPLL_FUNC_CNTL__SPLL_SLEEP_MASK;
+	WREG32(mmCG_SPLL_FUNC_CNTL, tmp);
 
-	tmp = RREG32(SPLL_CNTL_MODE);
-	tmp &= ~SPLL_SW_DIR_CONTROL;
-	WREG32(SPLL_CNTL_MODE, tmp);
+	tmp = RREG32(mmSPLL_CNTL_MODE);
+	tmp &= ~SPLL_CNTL_MODE__SPLL_SW_DIR_CONTROL_MASK;
+	WREG32(mmSPLL_CNTL_MODE, tmp);
 }
 
 static int si_gpu_pci_config_reset(struct amdgpu_device *adev)
@@ -1409,9 +1417,9 @@ static int si_gpu_pci_config_reset(struct amdgpu_device *adev)
 	return r;
 }
 
-static bool si_asic_supports_baco(struct amdgpu_device *adev)
+static int si_asic_supports_baco(struct amdgpu_device *adev)
 {
-	return false;
+	return 0;
 }
 
 static enum amd_reset_method
@@ -1454,14 +1462,14 @@ static void si_vga_set_state(struct amdgpu_device *adev, bool state)
 {
 	uint32_t temp;
 
-	temp = RREG32(CONFIG_CNTL);
+	temp = RREG32(mmCONFIG_CNTL);
 	if (!state) {
 		temp &= ~(1<<0);
 		temp |= (1<<1);
 	} else {
 		temp &= ~(1<<1);
 	}
-	WREG32(CONFIG_CNTL, temp);
+	WREG32(mmCONFIG_CNTL, temp);
 }
 
 static u32 si_get_xclk(struct amdgpu_device *adev)
@@ -1469,12 +1477,12 @@ static u32 si_get_xclk(struct amdgpu_device *adev)
 	u32 reference_clock = adev->clock.spll.reference_freq;
 	u32 tmp;
 
-	tmp = RREG32(CG_CLKPIN_CNTL_2);
-	if (tmp & MUX_TCLK_TO_XCLK)
+	tmp = RREG32(mmCG_CLKPIN_CNTL_2);
+	if (tmp & CG_CLKPIN_CNTL_2__MUX_TCLK_TO_XCLK_MASK)
 		return TCLK;
 
-	tmp = RREG32(CG_CLKPIN_CNTL);
-	if (tmp & XTALIN_DIVIDE)
+	tmp = RREG32(mmCG_CLKPIN_CNTL);
+	if (tmp & CG_CLKPIN_CNTL__XTALIN_DIVIDE_MASK)
 		return reference_clock / 4;
 
 	return reference_clock;
@@ -1519,9 +1527,9 @@ static int si_get_pcie_lanes(struct amdgpu_device *adev)
 	if (adev->flags & AMD_IS_APU)
 		return 0;
 
-	link_width_cntl = RREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL);
+	link_width_cntl = RREG32_PCIE_PORT(ixPCIE_LC_LINK_WIDTH_CNTL);
 
-	switch ((link_width_cntl & LC_LINK_WIDTH_RD_MASK) >> LC_LINK_WIDTH_RD_SHIFT) {
+	switch ((link_width_cntl & PCIE_LC_LINK_WIDTH_CNTL__LC_LINK_WIDTH_RD_MASK) >> PCIE_LC_LINK_WIDTH_CNTL__LC_LINK_WIDTH_RD__SHIFT) {
 	case LC_LINK_WIDTH_X1:
 		return 1;
 	case LC_LINK_WIDTH_X2:
@@ -1568,13 +1576,13 @@ static void si_set_pcie_lanes(struct amdgpu_device *adev, int lanes)
 		return;
 	}
 
-	link_width_cntl = RREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL);
-	link_width_cntl &= ~LC_LINK_WIDTH_MASK;
-	link_width_cntl |= mask << LC_LINK_WIDTH_SHIFT;
-	link_width_cntl |= (LC_RECONFIG_NOW |
-			    LC_RECONFIG_ARC_MISSING_ESCAPE);
+	link_width_cntl = RREG32_PCIE_PORT(ixPCIE_LC_LINK_WIDTH_CNTL);
+	link_width_cntl &= ~PCIE_LC_LINK_WIDTH_CNTL__LC_LINK_WIDTH_MASK;
+	link_width_cntl |= mask << PCIE_LC_LINK_WIDTH_CNTL__LC_LINK_WIDTH__SHIFT;
+	link_width_cntl |= (PCIE_LC_LINK_WIDTH_CNTL__LC_RECONFIG_NOW_MASK |
+		PCIE_LC_LINK_WIDTH_CNTL__LC_RECONFIG_ARC_MISSING_ESCAPE_MASK);
 
-	WREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL, link_width_cntl);
+	WREG32_PCIE_PORT(ixPCIE_LC_LINK_WIDTH_CNTL, link_width_cntl);
 }
 
 static void si_get_pcie_usage(struct amdgpu_device *adev, uint64_t *count0,
@@ -1888,7 +1896,7 @@ static int si_vce_send_vcepll_ctlreq(struct amdgpu_device *adev)
 	WREG32_SMC_P(CG_VCEPLL_FUNC_CNTL, 0, ~UPLL_CTLREQ_MASK);
 
 	if (i == SI_MAX_CTLACKS_ASSERTION_WAIT) {
-		DRM_ERROR("Timeout setting UVD clocks!\n");
+		DRM_ERROR("Timeout setting VCE clocks!\n");
 		return -ETIMEDOUT;
 	}
 
@@ -1910,6 +1918,14 @@ static int si_set_vce_clocks(struct amdgpu_device *adev, u32 evclk, u32 ecclk)
 		     ~VCEPLL_BYPASS_EN_MASK);
 
 	if (!evclk || !ecclk) {
+		/*
+		 * On some chips, the PLL takes way too long to get out of
+		 * sleep mode, causing a timeout waiting on CTLACK/CTLACK2.
+		 * Leave the PLL running in bypass mode.
+		 */
+		if (adev->pdev->device == 0x6780)
+			return 0;
+
 		/* Keep the Bypass mode, put PLL to sleep */
 		WREG32_SMC_P(CG_VCEPLL_FUNC_CNTL, VCEPLL_SLEEP_MASK,
 			     ~VCEPLL_SLEEP_MASK);
@@ -2018,13 +2034,13 @@ static const struct amdgpu_asic_funcs si_asic_funcs =
 
 static uint32_t si_get_rev_id(struct amdgpu_device *adev)
 {
-	return (RREG32(CC_DRM_ID_STRAPS) & CC_DRM_ID_STRAPS__ATI_REV_ID_MASK)
+	return (RREG32(mmCC_DRM_ID_STRAPS) & CC_DRM_ID_STRAPS__ATI_REV_ID_MASK)
 		>> CC_DRM_ID_STRAPS__ATI_REV_ID__SHIFT;
 }
 
-static int si_common_early_init(void *handle)
+static int si_common_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	adev->smc_rreg = &si_smc_rreg;
 	adev->smc_wreg = &si_smc_wreg;
@@ -2148,17 +2164,6 @@ static int si_common_early_init(void *handle)
 	return 0;
 }
 
-static int si_common_sw_init(void *handle)
-{
-	return 0;
-}
-
-static int si_common_sw_fini(void *handle)
-{
-	return 0;
-}
-
-
 static void si_init_golden_registers(struct amdgpu_device *adev)
 {
 	switch (adev->asic_type) {
@@ -2250,9 +2255,9 @@ static void si_pcie_gen3_enable(struct amdgpu_device *adev)
 					CAIL_PCIE_LINK_SPEED_SUPPORT_GEN3)))
 		return;
 
-	speed_cntl = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
-	current_data_rate = (speed_cntl & LC_CURRENT_DATA_RATE_MASK) >>
-		LC_CURRENT_DATA_RATE_SHIFT;
+	speed_cntl = RREG32_PCIE_PORT(ixPCIE_LC_SPEED_CNTL);
+	current_data_rate = (speed_cntl & PCIE_LC_SPEED_CNTL__LC_CURRENT_DATA_RATE_MASK) >>
+		PCIE_LC_SPEED_CNTL__LC_CURRENT_DATA_RATE__SHIFT;
 	if (adev->pm.pcie_gen_mask & CAIL_PCIE_LINK_SPEED_SUPPORT_GEN3) {
 		if (current_data_rate == 2) {
 			DRM_INFO("PCIE gen 3 link speeds already enabled\n");
@@ -2276,29 +2281,20 @@ static void si_pcie_gen3_enable(struct amdgpu_device *adev)
 			u16 bridge_cfg2, gpu_cfg2;
 			u32 max_lw, current_lw, tmp;
 
-			pcie_capability_read_word(root, PCI_EXP_LNKCTL,
-						  &bridge_cfg);
-			pcie_capability_read_word(adev->pdev, PCI_EXP_LNKCTL,
-						  &gpu_cfg);
+			pcie_capability_set_word(root, PCI_EXP_LNKCTL, PCI_EXP_LNKCTL_HAWD);
+			pcie_capability_set_word(adev->pdev, PCI_EXP_LNKCTL, PCI_EXP_LNKCTL_HAWD);
 
-			tmp16 = bridge_cfg | PCI_EXP_LNKCTL_HAWD;
-			pcie_capability_write_word(root, PCI_EXP_LNKCTL, tmp16);
-
-			tmp16 = gpu_cfg | PCI_EXP_LNKCTL_HAWD;
-			pcie_capability_write_word(adev->pdev, PCI_EXP_LNKCTL,
-						   tmp16);
-
-			tmp = RREG32_PCIE(PCIE_LC_STATUS1);
-			max_lw = (tmp & LC_DETECTED_LINK_WIDTH_MASK) >> LC_DETECTED_LINK_WIDTH_SHIFT;
-			current_lw = (tmp & LC_OPERATING_LINK_WIDTH_MASK) >> LC_OPERATING_LINK_WIDTH_SHIFT;
+			tmp = RREG32_PCIE(ixPCIE_LC_STATUS1);
+			max_lw = (tmp & PCIE_LC_STATUS1__LC_DETECTED_LINK_WIDTH_MASK) >> PCIE_LC_STATUS1__LC_DETECTED_LINK_WIDTH__SHIFT;
+			current_lw = (tmp & PCIE_LC_STATUS1__LC_OPERATING_LINK_WIDTH_MASK) >> PCIE_LC_STATUS1__LC_OPERATING_LINK_WIDTH__SHIFT;
 
 			if (current_lw < max_lw) {
-				tmp = RREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL);
-				if (tmp & LC_RENEGOTIATION_SUPPORT) {
-					tmp &= ~(LC_LINK_WIDTH_MASK | LC_UPCONFIGURE_DIS);
-					tmp |= (max_lw << LC_LINK_WIDTH_SHIFT);
-					tmp |= LC_UPCONFIGURE_SUPPORT | LC_RENEGOTIATE_EN | LC_RECONFIG_NOW;
-					WREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL, tmp);
+				tmp = RREG32_PCIE_PORT(ixPCIE_LC_LINK_WIDTH_CNTL);
+				if (tmp & PCIE_LC_LINK_WIDTH_CNTL__LC_RENEGOTIATION_SUPPORT_MASK) {
+					tmp &= ~(PCIE_LC_LINK_WIDTH_CNTL__LC_LINK_WIDTH_MASK | PCIE_LC_LINK_WIDTH_CNTL__LC_UPCONFIGURE_DIS_MASK);
+					tmp |= (max_lw << PCIE_LC_LINK_WIDTH_CNTL__LC_LINK_WIDTH__SHIFT);
+					tmp |= PCIE_LC_LINK_WIDTH_CNTL__LC_UPCONFIGURE_SUPPORT_MASK | PCIE_LC_LINK_WIDTH_CNTL__LC_RENEGOTIATE_EN_MASK | PCIE_LC_LINK_WIDTH_CNTL__LC_RECONFIG_NOW_MASK;
+					WREG32_PCIE_PORT(ixPCIE_LC_LINK_WIDTH_CNTL, tmp);
 				}
 			}
 
@@ -2321,84 +2317,66 @@ static void si_pcie_gen3_enable(struct amdgpu_device *adev)
 							  PCI_EXP_LNKCTL2,
 							  &gpu_cfg2);
 
-				tmp = RREG32_PCIE_PORT(PCIE_LC_CNTL4);
-				tmp |= LC_SET_QUIESCE;
-				WREG32_PCIE_PORT(PCIE_LC_CNTL4, tmp);
+				tmp = RREG32_PCIE_PORT(ixPCIE_LC_CNTL4);
+				tmp |= PCIE_LC_CNTL4__LC_SET_QUIESCE_MASK;
+				WREG32_PCIE_PORT(ixPCIE_LC_CNTL4, tmp);
 
-				tmp = RREG32_PCIE_PORT(PCIE_LC_CNTL4);
-				tmp |= LC_REDO_EQ;
-				WREG32_PCIE_PORT(PCIE_LC_CNTL4, tmp);
+				tmp = RREG32_PCIE_PORT(ixPCIE_LC_CNTL4);
+				tmp |= PCIE_LC_CNTL4__LC_REDO_EQ_MASK;
+				WREG32_PCIE_PORT(ixPCIE_LC_CNTL4, tmp);
 
 				mdelay(100);
 
-				pcie_capability_read_word(root, PCI_EXP_LNKCTL,
-							  &tmp16);
-				tmp16 &= ~PCI_EXP_LNKCTL_HAWD;
-				tmp16 |= (bridge_cfg & PCI_EXP_LNKCTL_HAWD);
-				pcie_capability_write_word(root, PCI_EXP_LNKCTL,
-							   tmp16);
-
-				pcie_capability_read_word(adev->pdev,
-							  PCI_EXP_LNKCTL,
-							  &tmp16);
-				tmp16 &= ~PCI_EXP_LNKCTL_HAWD;
-				tmp16 |= (gpu_cfg & PCI_EXP_LNKCTL_HAWD);
-				pcie_capability_write_word(adev->pdev,
-							   PCI_EXP_LNKCTL,
-							   tmp16);
-
-				pcie_capability_read_word(root, PCI_EXP_LNKCTL2,
-							  &tmp16);
-				tmp16 &= ~(PCI_EXP_LNKCTL2_ENTER_COMP |
-					   PCI_EXP_LNKCTL2_TX_MARGIN);
-				tmp16 |= (bridge_cfg2 &
-					  (PCI_EXP_LNKCTL2_ENTER_COMP |
-					   PCI_EXP_LNKCTL2_TX_MARGIN));
-				pcie_capability_write_word(root,
-							   PCI_EXP_LNKCTL2,
-							   tmp16);
-
-				pcie_capability_read_word(adev->pdev,
-							  PCI_EXP_LNKCTL2,
-							  &tmp16);
-				tmp16 &= ~(PCI_EXP_LNKCTL2_ENTER_COMP |
-					   PCI_EXP_LNKCTL2_TX_MARGIN);
-				tmp16 |= (gpu_cfg2 &
-					  (PCI_EXP_LNKCTL2_ENTER_COMP |
-					   PCI_EXP_LNKCTL2_TX_MARGIN));
-				pcie_capability_write_word(adev->pdev,
-							   PCI_EXP_LNKCTL2,
-							   tmp16);
-
-				tmp = RREG32_PCIE_PORT(PCIE_LC_CNTL4);
-				tmp &= ~LC_SET_QUIESCE;
-				WREG32_PCIE_PORT(PCIE_LC_CNTL4, tmp);
+				pcie_capability_clear_and_set_word(root, PCI_EXP_LNKCTL,
+								   PCI_EXP_LNKCTL_HAWD,
+								   bridge_cfg &
+								   PCI_EXP_LNKCTL_HAWD);
+				pcie_capability_clear_and_set_word(adev->pdev, PCI_EXP_LNKCTL,
+								   PCI_EXP_LNKCTL_HAWD,
+								   gpu_cfg &
+								   PCI_EXP_LNKCTL_HAWD);
+
+				pcie_capability_clear_and_set_word(root, PCI_EXP_LNKCTL2,
+								   PCI_EXP_LNKCTL2_ENTER_COMP |
+								   PCI_EXP_LNKCTL2_TX_MARGIN,
+								   bridge_cfg2 &
+								   (PCI_EXP_LNKCTL2_ENTER_COMP |
+								    PCI_EXP_LNKCTL2_TX_MARGIN));
+				pcie_capability_clear_and_set_word(adev->pdev, PCI_EXP_LNKCTL2,
+								   PCI_EXP_LNKCTL2_ENTER_COMP |
+								   PCI_EXP_LNKCTL2_TX_MARGIN,
+								   gpu_cfg2 &
+								   (PCI_EXP_LNKCTL2_ENTER_COMP |
+								    PCI_EXP_LNKCTL2_TX_MARGIN));
+
+				tmp = RREG32_PCIE_PORT(ixPCIE_LC_CNTL4);
+				tmp &= ~PCIE_LC_CNTL4__LC_SET_QUIESCE_MASK;
+				WREG32_PCIE_PORT(ixPCIE_LC_CNTL4, tmp);
 			}
 		}
 	}
 
-	speed_cntl |= LC_FORCE_EN_SW_SPEED_CHANGE | LC_FORCE_DIS_HW_SPEED_CHANGE;
-	speed_cntl &= ~LC_FORCE_DIS_SW_SPEED_CHANGE;
-	WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, speed_cntl);
-
-	pcie_capability_read_word(adev->pdev, PCI_EXP_LNKCTL2, &tmp16);
-	tmp16 &= ~PCI_EXP_LNKCTL2_TLS;
+	speed_cntl |= PCIE_LC_SPEED_CNTL__LC_FORCE_EN_SW_SPEED_CHANGE_MASK | PCIE_LC_SPEED_CNTL__LC_FORCE_DIS_HW_SPEED_CHANGE_MASK;
+	speed_cntl &= ~PCIE_LC_SPEED_CNTL__LC_FORCE_DIS_SW_SPEED_CHANGE_MASK;
+	WREG32_PCIE_PORT(ixPCIE_LC_SPEED_CNTL, speed_cntl);
 
+	tmp16 = 0;
 	if (adev->pm.pcie_gen_mask & CAIL_PCIE_LINK_SPEED_SUPPORT_GEN3)
 		tmp16 |= PCI_EXP_LNKCTL2_TLS_8_0GT; /* gen3 */
 	else if (adev->pm.pcie_gen_mask & CAIL_PCIE_LINK_SPEED_SUPPORT_GEN2)
 		tmp16 |= PCI_EXP_LNKCTL2_TLS_5_0GT; /* gen2 */
 	else
 		tmp16 |= PCI_EXP_LNKCTL2_TLS_2_5GT; /* gen1 */
-	pcie_capability_write_word(adev->pdev, PCI_EXP_LNKCTL2, tmp16);
+	pcie_capability_clear_and_set_word(adev->pdev, PCI_EXP_LNKCTL2,
+					   PCI_EXP_LNKCTL2_TLS, tmp16);
 
-	speed_cntl = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
-	speed_cntl |= LC_INITIATE_LINK_SPEED_CHANGE;
-	WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, speed_cntl);
+	speed_cntl = RREG32_PCIE_PORT(ixPCIE_LC_SPEED_CNTL);
+	speed_cntl |= PCIE_LC_SPEED_CNTL__LC_INITIATE_LINK_SPEED_CHANGE_MASK;
+	WREG32_PCIE_PORT(ixPCIE_LC_SPEED_CNTL, speed_cntl);
 
 	for (i = 0; i < adev->usec_timeout; i++) {
-		speed_cntl = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
-		if ((speed_cntl & LC_INITIATE_LINK_SPEED_CHANGE) == 0)
+		speed_cntl = RREG32_PCIE_PORT(ixPCIE_LC_SPEED_CNTL);
+		if ((speed_cntl & PCIE_LC_SPEED_CNTL__LC_INITIATE_LINK_SPEED_CHANGE_MASK) == 0)
 			break;
 		udelay(1);
 	}
@@ -2453,126 +2431,124 @@ static void si_program_aspm(struct amdgpu_device *adev)
 	bool disable_l0s = false, disable_l1 = false, disable_plloff_in_l1 = false;
 	bool disable_clkreq = false;
 
-	if (amdgpu_aspm == 0)
+	if (!amdgpu_device_should_use_aspm(adev))
 		return;
 
-	if (adev->flags & AMD_IS_APU)
-		return;
-	orig = data = RREG32_PCIE_PORT(PCIE_LC_N_FTS_CNTL);
-	data &= ~LC_XMIT_N_FTS_MASK;
-	data |= LC_XMIT_N_FTS(0x24) | LC_XMIT_N_FTS_OVERRIDE_EN;
+	orig = data = RREG32_PCIE_PORT(ixPCIE_LC_N_FTS_CNTL);
+	data &= ~PCIE_LC_N_FTS_CNTL__LC_XMIT_N_FTS_MASK;
+	data |= (0x24 << PCIE_LC_N_FTS_CNTL__LC_XMIT_N_FTS__SHIFT) | PCIE_LC_N_FTS_CNTL__LC_XMIT_N_FTS_OVERRIDE_EN_MASK;
 	if (orig != data)
-		WREG32_PCIE_PORT(PCIE_LC_N_FTS_CNTL, data);
+		WREG32_PCIE_PORT(ixPCIE_LC_N_FTS_CNTL, data);
 
-	orig = data = RREG32_PCIE_PORT(PCIE_LC_CNTL3);
-	data |= LC_GO_TO_RECOVERY;
+	orig = data = RREG32_PCIE_PORT(ixPCIE_LC_CNTL3);
+	data |= PCIE_LC_CNTL3__LC_GO_TO_RECOVERY_MASK;
 	if (orig != data)
-		WREG32_PCIE_PORT(PCIE_LC_CNTL3, data);
+		WREG32_PCIE_PORT(ixPCIE_LC_CNTL3, data);
 
-	orig = data = RREG32_PCIE(PCIE_P_CNTL);
-	data |= P_IGNORE_EDB_ERR;
+	orig = data = RREG32_PCIE(ixPCIE_P_CNTL);
+	data |= PCIE_P_CNTL__P_IGNORE_EDB_ERR_MASK;
 	if (orig != data)
-		WREG32_PCIE(PCIE_P_CNTL, data);
+		WREG32_PCIE(ixPCIE_P_CNTL, data);
 
-	orig = data = RREG32_PCIE_PORT(PCIE_LC_CNTL);
-	data &= ~(LC_L0S_INACTIVITY_MASK | LC_L1_INACTIVITY_MASK);
-	data |= LC_PMI_TO_L1_DIS;
+	orig = data = RREG32_PCIE_PORT(ixPCIE_LC_CNTL);
+	data &= ~(PCIE_LC_CNTL__LC_L0S_INACTIVITY_MASK | PCIE_LC_CNTL__LC_L1_INACTIVITY_MASK);
+	data |= PCIE_LC_CNTL__LC_PMI_TO_L1_DIS_MASK;
 	if (!disable_l0s)
-		data |= LC_L0S_INACTIVITY(7);
+		data |= (7 << PCIE_LC_CNTL__LC_L0S_INACTIVITY__SHIFT);
 
 	if (!disable_l1) {
-		data |= LC_L1_INACTIVITY(7);
-		data &= ~LC_PMI_TO_L1_DIS;
+		data |= (7 << PCIE_LC_CNTL__LC_L1_INACTIVITY__SHIFT);
+		data &= ~PCIE_LC_CNTL__LC_PMI_TO_L1_DIS_MASK;
 		if (orig != data)
-			WREG32_PCIE_PORT(PCIE_LC_CNTL, data);
+			WREG32_PCIE_PORT(ixPCIE_LC_CNTL, data);
 
 		if (!disable_plloff_in_l1) {
 			bool clk_req_support;
 
-			orig = data = si_pif_phy0_rreg(adev,PB0_PIF_PWRDOWN_0);
-			data &= ~(PLL_POWER_STATE_IN_OFF_0_MASK | PLL_POWER_STATE_IN_TXS2_0_MASK);
-			data |= PLL_POWER_STATE_IN_OFF_0(7) | PLL_POWER_STATE_IN_TXS2_0(7);
+			orig = data = si_pif_phy0_rreg(adev,ixPB0_PIF_PWRDOWN_0);
+			data &= ~(PB0_PIF_PWRDOWN_0__PLL_POWER_STATE_IN_OFF_0_MASK | PB0_PIF_PWRDOWN_0__PLL_POWER_STATE_IN_TXS2_0_MASK);
+			data |= (7 << PB0_PIF_PWRDOWN_0__PLL_POWER_STATE_IN_OFF_0__SHIFT) | (7 << PB0_PIF_PWRDOWN_0__PLL_POWER_STATE_IN_TXS2_0__SHIFT);
 			if (orig != data)
-				si_pif_phy0_wreg(adev,PB0_PIF_PWRDOWN_0, data);
+				si_pif_phy0_wreg(adev,ixPB0_PIF_PWRDOWN_0, data);
 
-			orig = data = si_pif_phy0_rreg(adev,PB0_PIF_PWRDOWN_1);
-			data &= ~(PLL_POWER_STATE_IN_OFF_1_MASK | PLL_POWER_STATE_IN_TXS2_1_MASK);
-			data |= PLL_POWER_STATE_IN_OFF_1(7) | PLL_POWER_STATE_IN_TXS2_1(7);
+			orig = data = si_pif_phy0_rreg(adev,ixPB0_PIF_PWRDOWN_1);
+			data &= ~(PB0_PIF_PWRDOWN_1__PLL_POWER_STATE_IN_OFF_1_MASK | PB0_PIF_PWRDOWN_1__PLL_POWER_STATE_IN_TXS2_1_MASK);
+			data |= (7 << PB0_PIF_PWRDOWN_1__PLL_POWER_STATE_IN_OFF_1__SHIFT) | (7 << PB0_PIF_PWRDOWN_1__PLL_POWER_STATE_IN_TXS2_1__SHIFT);
 			if (orig != data)
-				si_pif_phy0_wreg(adev,PB0_PIF_PWRDOWN_1, data);
+				si_pif_phy0_wreg(adev,ixPB0_PIF_PWRDOWN_1, data);
 
-			orig = data = si_pif_phy1_rreg(adev,PB1_PIF_PWRDOWN_0);
-			data &= ~(PLL_POWER_STATE_IN_OFF_0_MASK | PLL_POWER_STATE_IN_TXS2_0_MASK);
-			data |= PLL_POWER_STATE_IN_OFF_0(7) | PLL_POWER_STATE_IN_TXS2_0(7);
+			orig = data = si_pif_phy1_rreg(adev,ixPB1_PIF_PWRDOWN_0);
+			data &= ~(PB1_PIF_PWRDOWN_0__PLL_POWER_STATE_IN_OFF_0_MASK | PB1_PIF_PWRDOWN_0__PLL_POWER_STATE_IN_TXS2_0_MASK);
+			data |= (7 << PB1_PIF_PWRDOWN_0__PLL_POWER_STATE_IN_OFF_0__SHIFT) | (7 << PB1_PIF_PWRDOWN_0__PLL_POWER_STATE_IN_TXS2_0__SHIFT);
 			if (orig != data)
-				si_pif_phy1_wreg(adev,PB1_PIF_PWRDOWN_0, data);
+				si_pif_phy1_wreg(adev,ixPB1_PIF_PWRDOWN_0, data);
 
-			orig = data = si_pif_phy1_rreg(adev,PB1_PIF_PWRDOWN_1);
-			data &= ~(PLL_POWER_STATE_IN_OFF_1_MASK | PLL_POWER_STATE_IN_TXS2_1_MASK);
-			data |= PLL_POWER_STATE_IN_OFF_1(7) | PLL_POWER_STATE_IN_TXS2_1(7);
+			orig = data = si_pif_phy1_rreg(adev,ixPB1_PIF_PWRDOWN_1);
+			data &= ~(PB1_PIF_PWRDOWN_1__PLL_POWER_STATE_IN_OFF_1_MASK | PB1_PIF_PWRDOWN_1__PLL_POWER_STATE_IN_TXS2_1_MASK);
+			data |= (7 << PB1_PIF_PWRDOWN_1__PLL_POWER_STATE_IN_OFF_1__SHIFT) | (7 << PB1_PIF_PWRDOWN_1__PLL_POWER_STATE_IN_TXS2_1__SHIFT);
 			if (orig != data)
-				si_pif_phy1_wreg(adev,PB1_PIF_PWRDOWN_1, data);
+				si_pif_phy1_wreg(adev,ixPB1_PIF_PWRDOWN_1, data);
 
 			if ((adev->asic_type != CHIP_OLAND) && (adev->asic_type != CHIP_HAINAN)) {
-				orig = data = si_pif_phy0_rreg(adev,PB0_PIF_PWRDOWN_0);
-				data &= ~PLL_RAMP_UP_TIME_0_MASK;
+				orig = data = si_pif_phy0_rreg(adev,ixPB0_PIF_PWRDOWN_0);
+				data &= ~PB0_PIF_PWRDOWN_0__PLL_RAMP_UP_TIME_0_MASK;
 				if (orig != data)
-					si_pif_phy0_wreg(adev,PB0_PIF_PWRDOWN_0, data);
+					si_pif_phy0_wreg(adev,ixPB0_PIF_PWRDOWN_0, data);
 
-				orig = data = si_pif_phy0_rreg(adev,PB0_PIF_PWRDOWN_1);
-				data &= ~PLL_RAMP_UP_TIME_1_MASK;
+				orig = data = si_pif_phy0_rreg(adev,ixPB0_PIF_PWRDOWN_1);
+				data &= ~PB0_PIF_PWRDOWN_1__PLL_RAMP_UP_TIME_1_MASK;
 				if (orig != data)
-					si_pif_phy0_wreg(adev,PB0_PIF_PWRDOWN_1, data);
+					si_pif_phy0_wreg(adev,ixPB0_PIF_PWRDOWN_1, data);
 
-				orig = data = si_pif_phy0_rreg(adev,PB0_PIF_PWRDOWN_2);
-				data &= ~PLL_RAMP_UP_TIME_2_MASK;
+				orig = data = si_pif_phy0_rreg(adev,ixPB0_PIF_PWRDOWN_2);
+				data &= ~PB0_PIF_PWRDOWN_2__PLL_RAMP_UP_TIME_2_MASK;
 				if (orig != data)
-					si_pif_phy0_wreg(adev,PB0_PIF_PWRDOWN_2, data);
+					si_pif_phy0_wreg(adev,ixPB0_PIF_PWRDOWN_2, data);
 
-				orig = data = si_pif_phy0_rreg(adev,PB0_PIF_PWRDOWN_3);
-				data &= ~PLL_RAMP_UP_TIME_3_MASK;
+				orig = data = si_pif_phy0_rreg(adev,ixPB0_PIF_PWRDOWN_3);
+				data &= ~PB0_PIF_PWRDOWN_3__PLL_RAMP_UP_TIME_3_MASK;
 				if (orig != data)
-					si_pif_phy0_wreg(adev,PB0_PIF_PWRDOWN_3, data);
+					si_pif_phy0_wreg(adev,ixPB0_PIF_PWRDOWN_3, data);
 
-				orig = data = si_pif_phy1_rreg(adev,PB1_PIF_PWRDOWN_0);
-				data &= ~PLL_RAMP_UP_TIME_0_MASK;
+				orig = data = si_pif_phy1_rreg(adev,ixPB1_PIF_PWRDOWN_0);
+				data &= ~PB1_PIF_PWRDOWN_0__PLL_RAMP_UP_TIME_0_MASK;
 				if (orig != data)
-					si_pif_phy1_wreg(adev,PB1_PIF_PWRDOWN_0, data);
+					si_pif_phy1_wreg(adev,ixPB1_PIF_PWRDOWN_0, data);
 
-				orig = data = si_pif_phy1_rreg(adev,PB1_PIF_PWRDOWN_1);
-				data &= ~PLL_RAMP_UP_TIME_1_MASK;
+				orig = data = si_pif_phy1_rreg(adev,ixPB1_PIF_PWRDOWN_1);
+				data &= ~PB1_PIF_PWRDOWN_1__PLL_RAMP_UP_TIME_1_MASK;
 				if (orig != data)
-					si_pif_phy1_wreg(adev,PB1_PIF_PWRDOWN_1, data);
+					si_pif_phy1_wreg(adev,ixPB1_PIF_PWRDOWN_1, data);
 
-				orig = data = si_pif_phy1_rreg(adev,PB1_PIF_PWRDOWN_2);
-				data &= ~PLL_RAMP_UP_TIME_2_MASK;
+				orig = data = si_pif_phy1_rreg(adev,ixPB1_PIF_PWRDOWN_2);
+				data &= ~PB1_PIF_PWRDOWN_2__PLL_RAMP_UP_TIME_2_MASK;
 				if (orig != data)
-					si_pif_phy1_wreg(adev,PB1_PIF_PWRDOWN_2, data);
+					si_pif_phy1_wreg(adev,ixPB1_PIF_PWRDOWN_2, data);
 
-				orig = data = si_pif_phy1_rreg(adev,PB1_PIF_PWRDOWN_3);
-				data &= ~PLL_RAMP_UP_TIME_3_MASK;
+				orig = data = si_pif_phy1_rreg(adev,ixPB1_PIF_PWRDOWN_3);
+				data &= ~PB1_PIF_PWRDOWN_3__PLL_RAMP_UP_TIME_3_MASK;
 				if (orig != data)
-					si_pif_phy1_wreg(adev,PB1_PIF_PWRDOWN_3, data);
+					si_pif_phy1_wreg(adev,ixPB1_PIF_PWRDOWN_3, data);
 			}
-			orig = data = RREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL);
-			data &= ~LC_DYN_LANES_PWR_STATE_MASK;
-			data |= LC_DYN_LANES_PWR_STATE(3);
+			orig = data = RREG32_PCIE_PORT(ixPCIE_LC_LINK_WIDTH_CNTL);
+			data &= ~PCIE_LC_LINK_WIDTH_CNTL__LC_DYN_LANES_PWR_STATE_MASK;
+			data |= (3 << PCIE_LC_LINK_WIDTH_CNTL__LC_DYN_LANES_PWR_STATE__SHIFT);
 			if (orig != data)
-				WREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL, data);
+				WREG32_PCIE_PORT(ixPCIE_LC_LINK_WIDTH_CNTL, data);
 
-			orig = data = si_pif_phy0_rreg(adev,PB0_PIF_CNTL);
-			data &= ~LS2_EXIT_TIME_MASK;
+			orig = data = si_pif_phy0_rreg(adev,ixPB0_PIF_CNTL);
+			data &= ~PB0_PIF_CNTL__LS2_EXIT_TIME_MASK;
 			if ((adev->asic_type == CHIP_OLAND) || (adev->asic_type == CHIP_HAINAN))
-				data |= LS2_EXIT_TIME(5);
+				data |= (5 << PB0_PIF_CNTL__LS2_EXIT_TIME__SHIFT);
 			if (orig != data)
-				si_pif_phy0_wreg(adev,PB0_PIF_CNTL, data);
+				si_pif_phy0_wreg(adev,ixPB0_PIF_CNTL, data);
 
-			orig = data = si_pif_phy1_rreg(adev,PB1_PIF_CNTL);
-			data &= ~LS2_EXIT_TIME_MASK;
+			orig = data = si_pif_phy1_rreg(adev,ixPB1_PIF_CNTL);
+			data &= ~PB1_PIF_CNTL__LS2_EXIT_TIME_MASK;
 			if ((adev->asic_type == CHIP_OLAND) || (adev->asic_type == CHIP_HAINAN))
-				data |= LS2_EXIT_TIME(5);
+				data |= (5 << PB1_PIF_CNTL__LS2_EXIT_TIME__SHIFT);
 			if (orig != data)
-				si_pif_phy1_wreg(adev,PB1_PIF_CNTL, data);
+				si_pif_phy1_wreg(adev,ixPB1_PIF_CNTL, data);
 
 			if (!disable_clkreq &&
 			    !pci_is_root_bus(adev->pdev->bus)) {
@@ -2588,64 +2564,64 @@ static void si_program_aspm(struct amdgpu_device *adev)
 			}
 
 			if (clk_req_support) {
-				orig = data = RREG32_PCIE_PORT(PCIE_LC_CNTL2);
-				data |= LC_ALLOW_PDWN_IN_L1 | LC_ALLOW_PDWN_IN_L23;
+				orig = data = RREG32_PCIE_PORT(ixPCIE_LC_CNTL2);
+				data |= PCIE_LC_CNTL2__LC_ALLOW_PDWN_IN_L1_MASK | PCIE_LC_CNTL2__LC_ALLOW_PDWN_IN_L23_MASK;
 				if (orig != data)
-					WREG32_PCIE_PORT(PCIE_LC_CNTL2, data);
+					WREG32_PCIE_PORT(ixPCIE_LC_CNTL2, data);
 
-				orig = data = RREG32(THM_CLK_CNTL);
-				data &= ~(CMON_CLK_SEL_MASK | TMON_CLK_SEL_MASK);
-				data |= CMON_CLK_SEL(1) | TMON_CLK_SEL(1);
+				orig = data = RREG32(mmTHM_CLK_CNTL);
+				data &= ~(THM_CLK_CNTL__CMON_CLK_SEL_MASK | THM_CLK_CNTL__TMON_CLK_SEL_MASK);
+				data |= (1 << THM_CLK_CNTL__CMON_CLK_SEL__SHIFT) | (1 << THM_CLK_CNTL__TMON_CLK_SEL__SHIFT);
 				if (orig != data)
-					WREG32(THM_CLK_CNTL, data);
+					WREG32(mmTHM_CLK_CNTL, data);
 
-				orig = data = RREG32(MISC_CLK_CNTL);
-				data &= ~(DEEP_SLEEP_CLK_SEL_MASK | ZCLK_SEL_MASK);
-				data |= DEEP_SLEEP_CLK_SEL(1) | ZCLK_SEL(1);
+				orig = data = RREG32(mmMISC_CLK_CNTL);
+				data &= ~(MISC_CLK_CNTL__DEEP_SLEEP_CLK_SEL_MASK | MISC_CLK_CNTL__ZCLK_SEL_MASK);
+				data |= (1 << MISC_CLK_CNTL__DEEP_SLEEP_CLK_SEL__SHIFT) | (1 << MISC_CLK_CNTL__ZCLK_SEL__SHIFT);
 				if (orig != data)
-					WREG32(MISC_CLK_CNTL, data);
+					WREG32(mmMISC_CLK_CNTL, data);
 
-				orig = data = RREG32(CG_CLKPIN_CNTL);
-				data &= ~BCLK_AS_XCLK;
+				orig = data = RREG32(mmCG_CLKPIN_CNTL);
+				data &= ~CG_CLKPIN_CNTL__BCLK_AS_XCLK_MASK;
 				if (orig != data)
-					WREG32(CG_CLKPIN_CNTL, data);
+					WREG32(mmCG_CLKPIN_CNTL, data);
 
-				orig = data = RREG32(CG_CLKPIN_CNTL_2);
-				data &= ~FORCE_BIF_REFCLK_EN;
+				orig = data = RREG32(mmCG_CLKPIN_CNTL_2);
+				data &= ~CG_CLKPIN_CNTL_2__FORCE_BIF_REFCLK_EN_MASK;
 				if (orig != data)
-					WREG32(CG_CLKPIN_CNTL_2, data);
+					WREG32(mmCG_CLKPIN_CNTL_2, data);
 
-				orig = data = RREG32(MPLL_BYPASSCLK_SEL);
-				data &= ~MPLL_CLKOUT_SEL_MASK;
-				data |= MPLL_CLKOUT_SEL(4);
+				orig = data = RREG32(mmMPLL_BYPASSCLK_SEL);
+				data &= ~MPLL_BYPASSCLK_SEL__MPLL_CLKOUT_SEL_MASK;
+				data |= 4 << MPLL_BYPASSCLK_SEL__MPLL_CLKOUT_SEL__SHIFT;
 				if (orig != data)
-					WREG32(MPLL_BYPASSCLK_SEL, data);
+					WREG32(mmMPLL_BYPASSCLK_SEL, data);
 
-				orig = data = RREG32(SPLL_CNTL_MODE);
-				data &= ~SPLL_REFCLK_SEL_MASK;
+				orig = data = RREG32(mmSPLL_CNTL_MODE);
+				data &= ~SPLL_CNTL_MODE__SPLL_REFCLK_SEL_MASK;
 				if (orig != data)
-					WREG32(SPLL_CNTL_MODE, data);
+					WREG32(mmSPLL_CNTL_MODE, data);
 			}
 		}
 	} else {
 		if (orig != data)
-			WREG32_PCIE_PORT(PCIE_LC_CNTL, data);
+			WREG32_PCIE_PORT(ixPCIE_LC_CNTL, data);
 	}
 
-	orig = data = RREG32_PCIE(PCIE_CNTL2);
-	data |= SLV_MEM_LS_EN | MST_MEM_LS_EN | REPLAY_MEM_LS_EN;
+	orig = data = RREG32_PCIE(ixPCIE_CNTL2);
+	data |= PCIE_CNTL2__SLV_MEM_LS_EN_MASK | PCIE_CNTL2__MST_MEM_LS_EN_MASK | PCIE_CNTL2__REPLAY_MEM_LS_EN_MASK;
 	if (orig != data)
-		WREG32_PCIE(PCIE_CNTL2, data);
+		WREG32_PCIE(ixPCIE_CNTL2, data);
 
 	if (!disable_l0s) {
-		data = RREG32_PCIE_PORT(PCIE_LC_N_FTS_CNTL);
-		if((data & LC_N_FTS_MASK) == LC_N_FTS_MASK) {
-			data = RREG32_PCIE(PCIE_LC_STATUS1);
-			if ((data & LC_REVERSE_XMIT) && (data & LC_REVERSE_RCVR)) {
-				orig = data = RREG32_PCIE_PORT(PCIE_LC_CNTL);
-				data &= ~LC_L0S_INACTIVITY_MASK;
+		data = RREG32_PCIE_PORT(ixPCIE_LC_N_FTS_CNTL);
+		if((data & PCIE_LC_N_FTS_CNTL__LC_N_FTS_MASK) == PCIE_LC_N_FTS_CNTL__LC_N_FTS_MASK) {
+			data = RREG32_PCIE(ixPCIE_LC_STATUS1);
+			if ((data & PCIE_LC_STATUS1__LC_REVERSE_XMIT_MASK) && (data & PCIE_LC_STATUS1__LC_REVERSE_RCVR_MASK)) {
+				orig = data = RREG32_PCIE_PORT(ixPCIE_LC_CNTL);
+				data &= ~PCIE_LC_CNTL__LC_L0S_INACTIVITY_MASK;
 				if (orig != data)
-					WREG32_PCIE_PORT(PCIE_LC_CNTL, data);
+					WREG32_PCIE_PORT(ixPCIE_LC_CNTL, data);
 			}
 		}
 	}
@@ -2662,9 +2638,9 @@ static void si_fix_pci_max_read_req_size(struct amdgpu_device *adev)
 		pcie_set_readrq(adev->pdev, 512);
 }
 
-static int si_common_hw_init(void *handle)
+static int si_common_hw_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	si_fix_pci_max_read_req_size(adev);
 	si_init_golden_registers(adev);
@@ -2674,47 +2650,28 @@ static int si_common_hw_init(void *handle)
 	return 0;
 }
 
-static int si_common_hw_fini(void *handle)
+static int si_common_hw_fini(struct amdgpu_ip_block *ip_block)
 {
 	return 0;
 }
 
-static int si_common_suspend(void *handle)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	return si_common_hw_fini(adev);
-}
-
-static int si_common_resume(void *handle)
+static int si_common_resume(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	return si_common_hw_init(adev);
+	return si_common_hw_init(ip_block);
 }
 
-static bool si_common_is_idle(void *handle)
+static bool si_common_is_idle(struct amdgpu_ip_block *ip_block)
 {
 	return true;
 }
 
-static int si_common_wait_for_idle(void *handle)
-{
-	return 0;
-}
-
-static int si_common_soft_reset(void *handle)
-{
-	return 0;
-}
-
-static int si_common_set_clockgating_state(void *handle,
+static int si_common_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					    enum amd_clockgating_state state)
 {
 	return 0;
 }
 
-static int si_common_set_powergating_state(void *handle,
+static int si_common_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					    enum amd_powergating_state state)
 {
 	return 0;
@@ -2723,16 +2680,10 @@ static int si_common_set_powergating_state(void *handle,
 static const struct amd_ip_funcs si_common_ip_funcs = {
 	.name = "si_common",
 	.early_init = si_common_early_init,
-	.late_init = NULL,
-	.sw_init = si_common_sw_init,
-	.sw_fini = si_common_sw_fini,
 	.hw_init = si_common_hw_init,
 	.hw_fini = si_common_hw_fini,
-	.suspend = si_common_suspend,
 	.resume = si_common_resume,
 	.is_idle = si_common_is_idle,
-	.wait_for_idle = si_common_wait_for_idle,
-	.soft_reset = si_common_soft_reset,
 	.set_clockgating_state = si_common_set_clockgating_state,
 	.set_powergating_state = si_common_set_powergating_state,
 };
@@ -2767,7 +2718,7 @@ int si_set_ip_blocks(struct amdgpu_device *adev)
 		else
 			amdgpu_device_ip_block_add(adev, &dce_v6_0_ip_block);
 		amdgpu_device_ip_block_add(adev, &uvd_v3_1_ip_block);
-		/* amdgpu_device_ip_block_add(adev, &vce_v1_0_ip_block); */
+		amdgpu_device_ip_block_add(adev, &vce_v1_0_ip_block);
 		break;
 	case CHIP_OLAND:
 		amdgpu_device_ip_block_add(adev, &si_common_ip_block);
@@ -2785,7 +2736,6 @@ int si_set_ip_blocks(struct amdgpu_device *adev)
 		else
 			amdgpu_device_ip_block_add(adev, &dce_v6_4_ip_block);
 		amdgpu_device_ip_block_add(adev, &uvd_v3_1_ip_block);
-		/* amdgpu_device_ip_block_add(adev, &vce_v1_0_ip_block); */
 		break;
 	case CHIP_HAINAN:
 		amdgpu_device_ip_block_add(adev, &si_common_ip_block);
diff --git a/drivers/gpu/drm/amd/amdgpu/si_dma.c b/drivers/gpu/drm/amd/amdgpu/si_dma.c
index 195b45bcb8ad..7f18e4875287 100644
--- a/drivers/gpu/drm/amd/amdgpu/si_dma.c
+++ b/drivers/gpu/drm/amd/amdgpu/si_dma.c
@@ -27,6 +27,8 @@
 #include "si.h"
 #include "sid.h"
 
+#include "oss/oss_1_0_d.h"
+#include "oss/oss_1_0_sh_mask.h"
 const u32 sdma_offsets[SDMA_MAX_INSTANCE] =
 {
 	DMA0_REGISTER_OFFSET,
@@ -38,17 +40,31 @@ static void si_dma_set_buffer_funcs(struct amdgpu_device *adev);
 static void si_dma_set_vm_pte_funcs(struct amdgpu_device *adev);
 static void si_dma_set_irq_funcs(struct amdgpu_device *adev);
 
+/**
+ * si_dma_ring_get_rptr - get the current read pointer
+ *
+ * @ring: amdgpu ring pointer
+ *
+ * Get the current rptr from the hardware (SI).
+ */
 static uint64_t si_dma_ring_get_rptr(struct amdgpu_ring *ring)
 {
-	return ring->adev->wb.wb[ring->rptr_offs>>2];
+	return *ring->rptr_cpu_addr;
 }
 
+/**
+ * si_dma_ring_get_wptr - get the current write pointer
+ *
+ * @ring: amdgpu ring pointer
+ *
+ * Get the current wptr from the hardware (SI).
+ */
 static uint64_t si_dma_ring_get_wptr(struct amdgpu_ring *ring)
 {
 	struct amdgpu_device *adev = ring->adev;
 	u32 me = (ring == &adev->sdma.instance[0].ring) ? 0 : 1;
 
-	return (RREG32(DMA_RB_WPTR + sdma_offsets[me]) & 0x3fffc) >> 2;
+	return (RREG32(mmDMA_GFX_RB_WPTR + sdma_offsets[me]) & 0x3fffc) >> 2;
 }
 
 static void si_dma_ring_set_wptr(struct amdgpu_ring *ring)
@@ -56,8 +72,7 @@ static void si_dma_ring_set_wptr(struct amdgpu_ring *ring)
 	struct amdgpu_device *adev = ring->adev;
 	u32 me = (ring == &adev->sdma.instance[0].ring) ? 0 : 1;
 
-	WREG32(DMA_RB_WPTR + sdma_offsets[me],
-	       (lower_32_bits(ring->wptr) << 2) & 0x3fffc);
+	WREG32(mmDMA_GFX_RB_WPTR + sdma_offsets[me], (ring->wptr << 2) & 0x3fffc);
 }
 
 static void si_dma_ring_emit_ib(struct amdgpu_ring *ring,
@@ -113,19 +128,14 @@ static void si_dma_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 seq,
 
 static void si_dma_stop(struct amdgpu_device *adev)
 {
-	struct amdgpu_ring *ring;
 	u32 rb_cntl;
 	unsigned i;
 
 	for (i = 0; i < adev->sdma.num_instances; i++) {
-		ring = &adev->sdma.instance[i].ring;
 		/* dma0 */
-		rb_cntl = RREG32(DMA_RB_CNTL + sdma_offsets[i]);
-		rb_cntl &= ~DMA_RB_ENABLE;
-		WREG32(DMA_RB_CNTL + sdma_offsets[i], rb_cntl);
-
-		if (adev->mman.buffer_funcs_ring == ring)
-			amdgpu_ttm_set_buffer_funcs_status(adev, false);
+		rb_cntl = RREG32(mmDMA_GFX_RB_CNTL + sdma_offsets[i]);
+		rb_cntl &= ~DMA_GFX_RB_CNTL__RB_ENABLE_MASK;
+		WREG32(mmDMA_GFX_RB_CNTL + sdma_offsets[i], rb_cntl);
 	}
 }
 
@@ -139,53 +149,48 @@ static int si_dma_start(struct amdgpu_device *adev)
 	for (i = 0; i < adev->sdma.num_instances; i++) {
 		ring = &adev->sdma.instance[i].ring;
 
-		WREG32(DMA_SEM_INCOMPLETE_TIMER_CNTL + sdma_offsets[i], 0);
-		WREG32(DMA_SEM_WAIT_FAIL_TIMER_CNTL + sdma_offsets[i], 0);
+		WREG32(mmDMA_SEM_INCOMPLETE_TIMER_CNTL + sdma_offsets[i], 0);
+		WREG32(mmDMA_SEM_WAIT_FAIL_TIMER_CNTL + sdma_offsets[i], 0);
 
 		/* Set ring buffer size in dwords */
 		rb_bufsz = order_base_2(ring->ring_size / 4);
 		rb_cntl = rb_bufsz << 1;
 #ifdef __BIG_ENDIAN
-		rb_cntl |= DMA_RB_SWAP_ENABLE | DMA_RPTR_WRITEBACK_SWAP_ENABLE;
+		rb_cntl |= DMA_GFX_RB_CNTL__RB_SWAP_ENABLE_MASK | DMA_GFX_RB_CNTL__RPTR_WRITEBACK_SWAP_ENABLE_MASK;
 #endif
-		WREG32(DMA_RB_CNTL + sdma_offsets[i], rb_cntl);
+		WREG32(mmDMA_GFX_RB_CNTL + sdma_offsets[i], rb_cntl);
 
 		/* Initialize the ring buffer's read and write pointers */
-		WREG32(DMA_RB_RPTR + sdma_offsets[i], 0);
-		WREG32(DMA_RB_WPTR + sdma_offsets[i], 0);
+		WREG32(mmDMA_GFX_RB_RPTR + sdma_offsets[i], 0);
+		WREG32(mmDMA_GFX_RB_WPTR + sdma_offsets[i], 0);
 
-		rptr_addr = adev->wb.gpu_addr + (ring->rptr_offs * 4);
+		rptr_addr = ring->rptr_gpu_addr;
 
-		WREG32(DMA_RB_RPTR_ADDR_LO + sdma_offsets[i], lower_32_bits(rptr_addr));
-		WREG32(DMA_RB_RPTR_ADDR_HI + sdma_offsets[i], upper_32_bits(rptr_addr) & 0xFF);
+		WREG32(mmDMA_GFX_RB_RPTR_ADDR_LO + sdma_offsets[i], lower_32_bits(rptr_addr));
+		WREG32(mmDMA_GFX_RB_RPTR_ADDR_HI + sdma_offsets[i], upper_32_bits(rptr_addr) & 0xFF);
 
-		rb_cntl |= DMA_RPTR_WRITEBACK_ENABLE;
+		rb_cntl |= DMA_GFX_RB_CNTL__RPTR_WRITEBACK_ENABLE_MASK;
 
-		WREG32(DMA_RB_BASE + sdma_offsets[i], ring->gpu_addr >> 8);
+		WREG32(mmDMA_GFX_RB_BASE + sdma_offsets[i], ring->gpu_addr >> 8);
 
 		/* enable DMA IBs */
-		ib_cntl = DMA_IB_ENABLE | CMD_VMID_FORCE;
+		ib_cntl = DMA_GFX_IB_CNTL__IB_ENABLE_MASK | DMA_GFX_IB_CNTL__CMD_VMID_FORCE_MASK;
 #ifdef __BIG_ENDIAN
-		ib_cntl |= DMA_IB_SWAP_ENABLE;
+		ib_cntl |= DMA_GFX_IB_CNTL__IB_SWAP_ENABLE_MASK;
 #endif
-		WREG32(DMA_IB_CNTL + sdma_offsets[i], ib_cntl);
+		WREG32(mmDMA_GFX_IB_CNTL + sdma_offsets[i], ib_cntl);
 
-		dma_cntl = RREG32(DMA_CNTL + sdma_offsets[i]);
-		dma_cntl &= ~CTXEMPTY_INT_ENABLE;
-		WREG32(DMA_CNTL + sdma_offsets[i], dma_cntl);
+		dma_cntl = RREG32(mmDMA_CNTL + sdma_offsets[i]);
+		dma_cntl &= ~DMA_CNTL__CTXEMPTY_INT_ENABLE_MASK;
+		WREG32(mmDMA_CNTL + sdma_offsets[i], dma_cntl);
 
 		ring->wptr = 0;
-		WREG32(DMA_RB_WPTR + sdma_offsets[i], lower_32_bits(ring->wptr) << 2);
-		WREG32(DMA_RB_CNTL + sdma_offsets[i], rb_cntl | DMA_RB_ENABLE);
-
-		ring->sched.ready = true;
+		WREG32(mmDMA_GFX_RB_WPTR + sdma_offsets[i], ring->wptr << 2);
+		WREG32(mmDMA_GFX_RB_CNTL + sdma_offsets[i], rb_cntl | DMA_GFX_RB_CNTL__RB_ENABLE_MASK);
 
 		r = amdgpu_ring_test_helper(ring);
 		if (r)
 			return r;
-
-		if (adev->mman.buffer_funcs_ring == ring)
-			amdgpu_ttm_set_buffer_funcs_status(adev, true);
 	}
 
 	return 0;
@@ -297,7 +302,7 @@ static int si_dma_ring_test_ib(struct amdgpu_ring *ring, long timeout)
 		r = -EINVAL;
 
 err1:
-	amdgpu_ib_free(adev, &ib, NULL);
+	amdgpu_ib_free(&ib, NULL);
 	dma_fence_put(f);
 err0:
 	amdgpu_device_wb_free(adev, index);
@@ -468,11 +473,11 @@ static void si_dma_ring_emit_wreg(struct amdgpu_ring *ring,
 	amdgpu_ring_write(ring, val);
 }
 
-static int si_dma_early_init(void *handle)
+static int si_dma_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
-	adev->sdma.num_instances = 2;
+	adev->sdma.num_instances = SDMA_MAX_INSTANCE;
 
 	si_dma_set_ring_funcs(adev);
 	si_dma_set_buffer_funcs(adev);
@@ -482,11 +487,11 @@ static int si_dma_early_init(void *handle)
 	return 0;
 }
 
-static int si_dma_sw_init(void *handle)
+static int si_dma_sw_init(struct amdgpu_ip_block *ip_block)
 {
 	struct amdgpu_ring *ring;
 	int r, i;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	/* DMA0 trap event */
 	r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 224,
@@ -517,9 +522,9 @@ static int si_dma_sw_init(void *handle)
 	return r;
 }
 
-static int si_dma_sw_fini(void *handle)
+static int si_dma_sw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int i;
 
 	for (i = 0; i < adev->sdma.num_instances; i++)
@@ -528,61 +533,56 @@ static int si_dma_sw_fini(void *handle)
 	return 0;
 }
 
-static int si_dma_hw_init(void *handle)
+static int si_dma_hw_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	return si_dma_start(adev);
 }
 
-static int si_dma_hw_fini(void *handle)
+static int si_dma_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	si_dma_stop(adev);
+	si_dma_stop(ip_block->adev);
 
 	return 0;
 }
 
-static int si_dma_suspend(void *handle)
+static int si_dma_suspend(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	return si_dma_hw_fini(adev);
+	return si_dma_hw_fini(ip_block);
 }
 
-static int si_dma_resume(void *handle)
+static int si_dma_resume(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	return si_dma_hw_init(adev);
+	return si_dma_hw_init(ip_block);
 }
 
-static bool si_dma_is_idle(void *handle)
+static bool si_dma_is_idle(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-	u32 tmp = RREG32(SRBM_STATUS2);
+	struct amdgpu_device *adev = ip_block->adev;
+
+	u32 tmp = RREG32(mmSRBM_STATUS2);
 
-	if (tmp & (DMA_BUSY_MASK | DMA1_BUSY_MASK))
+	if (tmp & (SRBM_STATUS2__DMA_BUSY_MASK | SRBM_STATUS2__DMA1_BUSY_MASK))
 	    return false;
 
 	return true;
 }
 
-static int si_dma_wait_for_idle(void *handle)
+static int si_dma_wait_for_idle(struct amdgpu_ip_block *ip_block)
 {
 	unsigned i;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	for (i = 0; i < adev->usec_timeout; i++) {
-		if (si_dma_is_idle(handle))
+		if (si_dma_is_idle(ip_block))
 			return 0;
 		udelay(1);
 	}
 	return -ETIMEDOUT;
 }
 
-static int si_dma_soft_reset(void *handle)
+static int si_dma_soft_reset(struct amdgpu_ip_block *ip_block)
 {
 	DRM_INFO("si_dma_soft_reset --- not implemented !!!!!!!\n");
 	return 0;
@@ -599,14 +599,14 @@ static int si_dma_set_trap_irq_state(struct amdgpu_device *adev,
 	case AMDGPU_SDMA_IRQ_INSTANCE0:
 		switch (state) {
 		case AMDGPU_IRQ_STATE_DISABLE:
-			sdma_cntl = RREG32(DMA_CNTL + DMA0_REGISTER_OFFSET);
-			sdma_cntl &= ~TRAP_ENABLE;
-			WREG32(DMA_CNTL + DMA0_REGISTER_OFFSET, sdma_cntl);
+			sdma_cntl = RREG32(mmDMA_CNTL + DMA0_REGISTER_OFFSET);
+			sdma_cntl &= ~DMA_CNTL__TRAP_ENABLE_MASK;
+			WREG32(mmDMA_CNTL + DMA0_REGISTER_OFFSET, sdma_cntl);
 			break;
 		case AMDGPU_IRQ_STATE_ENABLE:
-			sdma_cntl = RREG32(DMA_CNTL + DMA0_REGISTER_OFFSET);
-			sdma_cntl |= TRAP_ENABLE;
-			WREG32(DMA_CNTL + DMA0_REGISTER_OFFSET, sdma_cntl);
+			sdma_cntl = RREG32(mmDMA_CNTL + DMA0_REGISTER_OFFSET);
+			sdma_cntl |= DMA_CNTL__TRAP_ENABLE_MASK;
+			WREG32(mmDMA_CNTL + DMA0_REGISTER_OFFSET, sdma_cntl);
 			break;
 		default:
 			break;
@@ -615,14 +615,14 @@ static int si_dma_set_trap_irq_state(struct amdgpu_device *adev,
 	case AMDGPU_SDMA_IRQ_INSTANCE1:
 		switch (state) {
 		case AMDGPU_IRQ_STATE_DISABLE:
-			sdma_cntl = RREG32(DMA_CNTL + DMA1_REGISTER_OFFSET);
-			sdma_cntl &= ~TRAP_ENABLE;
-			WREG32(DMA_CNTL + DMA1_REGISTER_OFFSET, sdma_cntl);
+			sdma_cntl = RREG32(mmDMA_CNTL + DMA1_REGISTER_OFFSET);
+			sdma_cntl &= ~DMA_CNTL__TRAP_ENABLE_MASK;
+			WREG32(mmDMA_CNTL + DMA1_REGISTER_OFFSET, sdma_cntl);
 			break;
 		case AMDGPU_IRQ_STATE_ENABLE:
-			sdma_cntl = RREG32(DMA_CNTL + DMA1_REGISTER_OFFSET);
-			sdma_cntl |= TRAP_ENABLE;
-			WREG32(DMA_CNTL + DMA1_REGISTER_OFFSET, sdma_cntl);
+			sdma_cntl = RREG32(mmDMA_CNTL + DMA1_REGISTER_OFFSET);
+			sdma_cntl |= DMA_CNTL__TRAP_ENABLE_MASK;
+			WREG32(mmDMA_CNTL + DMA1_REGISTER_OFFSET, sdma_cntl);
 			break;
 		default:
 			break;
@@ -645,13 +645,13 @@ static int si_dma_process_trap_irq(struct amdgpu_device *adev,
 	return 0;
 }
 
-static int si_dma_set_clockgating_state(void *handle,
+static int si_dma_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_clockgating_state state)
 {
 	u32 orig, data, offset;
 	int i;
 	bool enable;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	enable = (state == AMD_CG_STATE_GATE);
 
@@ -661,11 +661,11 @@ static int si_dma_set_clockgating_state(void *handle,
 				offset = DMA0_REGISTER_OFFSET;
 			else
 				offset = DMA1_REGISTER_OFFSET;
-			orig = data = RREG32(DMA_POWER_CNTL + offset);
-			data &= ~MEM_POWER_OVERRIDE;
+			orig = data = RREG32(mmDMA_POWER_CNTL + offset);
+			data &= ~DMA_POWER_CNTL__MEM_POWER_OVERRIDE_MASK;
 			if (data != orig)
-				WREG32(DMA_POWER_CNTL + offset, data);
-			WREG32(DMA_CLK_CTRL + offset, 0x00000100);
+				WREG32(mmDMA_POWER_CNTL + offset, data);
+			WREG32(mmDMA_CLK_CTRL + offset, 0x00000100);
 		}
 	} else {
 		for (i = 0; i < adev->sdma.num_instances; i++) {
@@ -673,33 +673,33 @@ static int si_dma_set_clockgating_state(void *handle,
 				offset = DMA0_REGISTER_OFFSET;
 			else
 				offset = DMA1_REGISTER_OFFSET;
-			orig = data = RREG32(DMA_POWER_CNTL + offset);
-			data |= MEM_POWER_OVERRIDE;
+			orig = data = RREG32(mmDMA_POWER_CNTL + offset);
+			data |= DMA_POWER_CNTL__MEM_POWER_OVERRIDE_MASK;
 			if (data != orig)
-				WREG32(DMA_POWER_CNTL + offset, data);
+				WREG32(mmDMA_POWER_CNTL + offset, data);
 
-			orig = data = RREG32(DMA_CLK_CTRL + offset);
+			orig = data = RREG32(mmDMA_CLK_CTRL + offset);
 			data = 0xff000000;
 			if (data != orig)
-				WREG32(DMA_CLK_CTRL + offset, data);
+				WREG32(mmDMA_CLK_CTRL + offset, data);
 		}
 	}
 
 	return 0;
 }
 
-static int si_dma_set_powergating_state(void *handle,
+static int si_dma_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_powergating_state state)
 {
 	u32 tmp;
 
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
-	WREG32(DMA_PGFSM_WRITE,  0x00002000);
-	WREG32(DMA_PGFSM_CONFIG, 0x100010ff);
+	WREG32(mmDMA_PGFSM_WRITE,  0x00002000);
+	WREG32(mmDMA_PGFSM_CONFIG, 0x100010ff);
 
 	for (tmp = 0; tmp < 5; tmp++)
-		WREG32(DMA_PGFSM_WRITE, 0);
+		WREG32(mmDMA_PGFSM_WRITE, 0);
 
 	return 0;
 }
@@ -707,7 +707,6 @@ static int si_dma_set_powergating_state(void *handle,
 static const struct amd_ip_funcs si_dma_ip_funcs = {
 	.name = "si_dma",
 	.early_init = si_dma_early_init,
-	.late_init = NULL,
 	.sw_init = si_dma_sw_init,
 	.sw_fini = si_dma_sw_fini,
 	.hw_init = si_dma_hw_init,
@@ -772,7 +771,7 @@ static void si_dma_set_irq_funcs(struct amdgpu_device *adev)
  * @src_offset: src GPU address
  * @dst_offset: dst GPU address
  * @byte_count: number of bytes to xfer
- * @tmz: is this a secure operation
+ * @copy_flags: unused
  *
  * Copy GPU buffers using the DMA engine (VI).
  * Used by the amdgpu ttm implementation to move pages if
@@ -782,7 +781,7 @@ static void si_dma_emit_copy_buffer(struct amdgpu_ib *ib,
 				       uint64_t src_offset,
 				       uint64_t dst_offset,
 				       uint32_t byte_count,
-				       bool tmz)
+				       uint32_t copy_flags)
 {
 	ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_COPY,
 					      1, 0, 0, byte_count);
diff --git a/drivers/gpu/drm/amd/amdgpu/si_enums.h b/drivers/gpu/drm/amd/amdgpu/si_enums.h
index 4e935baa7b91..6da65778292b 100644
--- a/drivers/gpu/drm/amd/amdgpu/si_enums.h
+++ b/drivers/gpu/drm/amd/amdgpu/si_enums.h
@@ -23,123 +23,15 @@
 #ifndef SI_ENUMS_H
 #define SI_ENUMS_H
 
-#define VBLANK_INT_MASK                (1 << 0)
-#define DC_HPDx_INT_EN                 (1 << 16)
-#define VBLANK_ACK                     (1 << 4)
-#define VLINE_ACK                      (1 << 4)
-
-#define CURSOR_WIDTH 64
-#define CURSOR_HEIGHT 64
-
-#define VGA_VSTATUS_CNTL               0xFFFCFFFF
 #define PRIORITY_MARK_MASK             0x7fff
 #define PRIORITY_OFF                   (1 << 16)
 #define PRIORITY_ALWAYS_ON             (1 << 20)
-#define INTERLEAVE_EN                  (1 << 0)
-
-#define LATENCY_WATERMARK_MASK(x)      ((x) << 16)
-#define DC_LB_MEMORY_CONFIG(x)         ((x) << 20)
-#define ICON_DEGAMMA_MODE(x)           (((x) & 0x3) << 8)
-
-#define GRPH_ENDIAN_SWAP(x)            (((x) & 0x3) << 0)
-#define GRPH_ENDIAN_NONE               0
-#define GRPH_ENDIAN_8IN16              1
-#define GRPH_ENDIAN_8IN32              2
-#define GRPH_ENDIAN_8IN64              3
-#define GRPH_RED_CROSSBAR(x)           (((x) & 0x3) << 4)
-#define GRPH_RED_SEL_R                 0
-#define GRPH_RED_SEL_G                 1
-#define GRPH_RED_SEL_B                 2
-#define GRPH_RED_SEL_A                 3
-#define GRPH_GREEN_CROSSBAR(x)         (((x) & 0x3) << 6)
-#define GRPH_GREEN_SEL_G               0
-#define GRPH_GREEN_SEL_B               1
-#define GRPH_GREEN_SEL_A               2
-#define GRPH_GREEN_SEL_R               3
-#define GRPH_BLUE_CROSSBAR(x)          (((x) & 0x3) << 8)
-#define GRPH_BLUE_SEL_B                0
-#define GRPH_BLUE_SEL_A                1
-#define GRPH_BLUE_SEL_R                2
-#define GRPH_BLUE_SEL_G                3
-#define GRPH_ALPHA_CROSSBAR(x)         (((x) & 0x3) << 10)
-#define GRPH_ALPHA_SEL_A               0
-#define GRPH_ALPHA_SEL_R               1
-#define GRPH_ALPHA_SEL_G               2
-#define GRPH_ALPHA_SEL_B               3
-
-#define GRPH_DEPTH(x)                  (((x) & 0x3) << 0)
-#define GRPH_DEPTH_8BPP                0
-#define GRPH_DEPTH_16BPP               1
-#define GRPH_DEPTH_32BPP               2
-
-#define GRPH_FORMAT(x)                 (((x) & 0x7) << 8)
-#define GRPH_FORMAT_INDEXED            0
-#define GRPH_FORMAT_ARGB1555           0
-#define GRPH_FORMAT_ARGB565            1
-#define GRPH_FORMAT_ARGB4444           2
-#define GRPH_FORMAT_AI88               3
-#define GRPH_FORMAT_MONO16             4
-#define GRPH_FORMAT_BGRA5551           5
-#define GRPH_FORMAT_ARGB8888           0
-#define GRPH_FORMAT_ARGB2101010        1
-#define GRPH_FORMAT_32BPP_DIG          2
-#define GRPH_FORMAT_8B_ARGB2101010     3
-#define GRPH_FORMAT_BGRA1010102        4
-#define GRPH_FORMAT_8B_BGRA1010102     5
-#define GRPH_FORMAT_RGB111110          6
-#define GRPH_FORMAT_BGR101111          7
-
-#define GRPH_NUM_BANKS(x)              (((x) & 0x3) << 2)
-#define GRPH_ARRAY_MODE(x)             (((x) & 0x7) << 20)
-#define GRPH_ARRAY_LINEAR_GENERAL      0
-#define GRPH_ARRAY_LINEAR_ALIGNED      1
-#define GRPH_ARRAY_1D_TILED_THIN1      2
-#define GRPH_ARRAY_2D_TILED_THIN1      4
-#define GRPH_TILE_SPLIT(x)             (((x) & 0x7) << 13)
-#define GRPH_BANK_WIDTH(x)             (((x) & 0x3) << 6)
-#define GRPH_BANK_HEIGHT(x)            (((x) & 0x3) << 11)
-#define GRPH_MACRO_TILE_ASPECT(x)      (((x) & 0x3) << 18)
-#define GRPH_ARRAY_MODE(x)             (((x) & 0x7) << 20)
-#define GRPH_PIPE_CONFIG(x)                   (((x) & 0x1f) << 24)
-
-#define CURSOR_EN                      (1 << 0)
-#define CURSOR_MODE(x)                 (((x) & 0x3) << 8)
-#define CURSOR_MONO                    0
-#define CURSOR_24_1                    1
-#define CURSOR_24_8_PRE_MULT           2
-#define CURSOR_24_8_UNPRE_MULT         3
-#define CURSOR_2X_MAGNIFY              (1 << 16)
-#define CURSOR_FORCE_MC_ON             (1 << 20)
-#define CURSOR_URGENT_CONTROL(x)       (((x) & 0x7) << 24)
-#define CURSOR_URGENT_ALWAYS           0
-#define CURSOR_URGENT_1_8              1
-#define CURSOR_URGENT_1_4              2
-#define CURSOR_URGENT_3_8              3
-#define CURSOR_URGENT_1_2              4
-#define CURSOR_UPDATE_PENDING          (1 << 0)
-#define CURSOR_UPDATE_TAKEN            (1 << 1)
-#define CURSOR_UPDATE_LOCK             (1 << 16)
-#define CURSOR_DISABLE_MULTIPLE_UPDATE (1 << 24)
-
-#define SI_CRTC0_REGISTER_OFFSET                0
-#define SI_CRTC1_REGISTER_OFFSET                0x300
-#define SI_CRTC2_REGISTER_OFFSET                0x2600
-#define SI_CRTC3_REGISTER_OFFSET                0x2900
-#define SI_CRTC4_REGISTER_OFFSET                0x2c00
-#define SI_CRTC5_REGISTER_OFFSET                0x2f00
-
-#define DMA0_REGISTER_OFFSET 0x000
-#define DMA1_REGISTER_OFFSET 0x200
-#define ES_AND_GS_AUTO       3
-#define RADEON_PACKET_TYPE3  3
-#define CE_PARTITION_BASE    3
-#define BUF_SWAP_32BIT       (2 << 16)
 
 #define GFX_POWER_STATUS                           (1 << 1)
 #define GFX_CLOCK_STATUS                           (1 << 2)
 #define GFX_LS_STATUS                              (1 << 3)
-#define RLC_BUSY_STATUS                            (1 << 0)
 
+#define RLC_BUSY_STATUS                            (1 << 0)
 #define RLC_PUD(x)                               ((x) << 0)
 #define RLC_PUD_MASK                             (0xff << 0)
 #define RLC_PDD(x)                               ((x) << 8)
@@ -148,144 +40,8 @@
 #define RLC_TTPD_MASK                            (0xff << 16)
 #define RLC_MSD(x)                               ((x) << 24)
 #define RLC_MSD_MASK                             (0xff << 24)
-#define WRITE_DATA_ENGINE_SEL(x) ((x) << 30)
-#define WRITE_DATA_DST_SEL(x) ((x) << 8)
-#define EVENT_TYPE(x) ((x) << 0)
-#define EVENT_INDEX(x) ((x) << 8)
-#define WAIT_REG_MEM_MEM_SPACE(x)               ((x) << 4)
-#define WAIT_REG_MEM_FUNCTION(x)                ((x) << 0)
-#define WAIT_REG_MEM_ENGINE(x)                  ((x) << 8)
 
-#define GFX6_NUM_GFX_RINGS     1
-#define GFX6_NUM_COMPUTE_RINGS 2
 #define RLC_SAVE_AND_RESTORE_STARTING_OFFSET 0x90
 #define RLC_CLEAR_STATE_DESCRIPTOR_OFFSET    0x3D
 
-#define TAHITI_GB_ADDR_CONFIG_GOLDEN        0x12011003
-#define VERDE_GB_ADDR_CONFIG_GOLDEN         0x02010002
-#define HAINAN_GB_ADDR_CONFIG_GOLDEN        0x02011003
-
-#define PACKET3(op, n)  ((RADEON_PACKET_TYPE3 << 30) |                  \
-                         (((op) & 0xFF) << 8) |                         \
-                         ((n) & 0x3FFF) << 16)
-#define PACKET3_COMPUTE(op, n) (PACKET3(op, n) | 1 << 1)
-#define	PACKET3_NOP					0x10
-#define	PACKET3_SET_BASE				0x11
-#define		PACKET3_BASE_INDEX(x)                  ((x) << 0)
-#define	PACKET3_CLEAR_STATE				0x12
-#define	PACKET3_INDEX_BUFFER_SIZE			0x13
-#define	PACKET3_DISPATCH_DIRECT				0x15
-#define	PACKET3_DISPATCH_INDIRECT			0x16
-#define	PACKET3_ALLOC_GDS				0x1B
-#define	PACKET3_WRITE_GDS_RAM				0x1C
-#define	PACKET3_ATOMIC_GDS				0x1D
-#define	PACKET3_ATOMIC					0x1E
-#define	PACKET3_OCCLUSION_QUERY				0x1F
-#define	PACKET3_SET_PREDICATION				0x20
-#define	PACKET3_REG_RMW					0x21
-#define	PACKET3_COND_EXEC				0x22
-#define	PACKET3_PRED_EXEC				0x23
-#define	PACKET3_DRAW_INDIRECT				0x24
-#define	PACKET3_DRAW_INDEX_INDIRECT			0x25
-#define	PACKET3_INDEX_BASE				0x26
-#define	PACKET3_DRAW_INDEX_2				0x27
-#define	PACKET3_CONTEXT_CONTROL				0x28
-#define	PACKET3_INDEX_TYPE				0x2A
-#define	PACKET3_DRAW_INDIRECT_MULTI			0x2C
-#define	PACKET3_DRAW_INDEX_AUTO				0x2D
-#define	PACKET3_DRAW_INDEX_IMMD				0x2E
-#define	PACKET3_NUM_INSTANCES				0x2F
-#define	PACKET3_DRAW_INDEX_MULTI_AUTO			0x30
-#define	PACKET3_INDIRECT_BUFFER_CONST			0x31
-#define	PACKET3_INDIRECT_BUFFER				0x3F
-#define	PACKET3_STRMOUT_BUFFER_UPDATE			0x34
-#define	PACKET3_DRAW_INDEX_OFFSET_2			0x35
-#define	PACKET3_DRAW_INDEX_MULTI_ELEMENT		0x36
-#define	PACKET3_WRITE_DATA				0x37
-#define	PACKET3_DRAW_INDEX_INDIRECT_MULTI		0x38
-#define	PACKET3_MEM_SEMAPHORE				0x39
-#define	PACKET3_MPEG_INDEX				0x3A
-#define	PACKET3_COPY_DW					0x3B
-#define	PACKET3_WAIT_REG_MEM				0x3C
-#define	PACKET3_MEM_WRITE				0x3D
-#define	PACKET3_COPY_DATA				0x40
-#define	PACKET3_CP_DMA					0x41
-#              define PACKET3_CP_DMA_DST_SEL(x)    ((x) << 20)
-#              define PACKET3_CP_DMA_ENGINE(x)     ((x) << 27)
-#              define PACKET3_CP_DMA_SRC_SEL(x)    ((x) << 29)
-#              define PACKET3_CP_DMA_CP_SYNC       (1 << 31)
-#              define PACKET3_CP_DMA_DIS_WC        (1 << 21)
-#              define PACKET3_CP_DMA_CMD_SRC_SWAP(x) ((x) << 22)
-#              define PACKET3_CP_DMA_CMD_DST_SWAP(x) ((x) << 24)
-#              define PACKET3_CP_DMA_CMD_SAS       (1 << 26)
-#              define PACKET3_CP_DMA_CMD_DAS       (1 << 27)
-#              define PACKET3_CP_DMA_CMD_SAIC      (1 << 28)
-#              define PACKET3_CP_DMA_CMD_DAIC      (1 << 29)
-#              define PACKET3_CP_DMA_CMD_RAW_WAIT  (1 << 30)
-#define	PACKET3_PFP_SYNC_ME				0x42
-#define	PACKET3_SURFACE_SYNC				0x43
-#              define PACKET3_DEST_BASE_0_ENA      (1 << 0)
-#              define PACKET3_DEST_BASE_1_ENA      (1 << 1)
-#              define PACKET3_CB0_DEST_BASE_ENA    (1 << 6)
-#              define PACKET3_CB1_DEST_BASE_ENA    (1 << 7)
-#              define PACKET3_CB2_DEST_BASE_ENA    (1 << 8)
-#              define PACKET3_CB3_DEST_BASE_ENA    (1 << 9)
-#              define PACKET3_CB4_DEST_BASE_ENA    (1 << 10)
-#              define PACKET3_CB5_DEST_BASE_ENA    (1 << 11)
-#              define PACKET3_CB6_DEST_BASE_ENA    (1 << 12)
-#              define PACKET3_CB7_DEST_BASE_ENA    (1 << 13)
-#              define PACKET3_DB_DEST_BASE_ENA     (1 << 14)
-#              define PACKET3_DEST_BASE_2_ENA      (1 << 19)
-#              define PACKET3_DEST_BASE_3_ENA      (1 << 21)
-#              define PACKET3_TCL1_ACTION_ENA      (1 << 22)
-#              define PACKET3_TC_ACTION_ENA        (1 << 23)
-#              define PACKET3_CB_ACTION_ENA        (1 << 25)
-#              define PACKET3_DB_ACTION_ENA        (1 << 26)
-#              define PACKET3_SH_KCACHE_ACTION_ENA (1 << 27)
-#              define PACKET3_SH_ICACHE_ACTION_ENA (1 << 29)
-#define	PACKET3_ME_INITIALIZE				0x44
-#define		PACKET3_ME_INITIALIZE_DEVICE_ID(x) ((x) << 16)
-#define	PACKET3_COND_WRITE				0x45
-#define	PACKET3_EVENT_WRITE				0x46
-#define	PACKET3_EVENT_WRITE_EOP				0x47
-#define	PACKET3_EVENT_WRITE_EOS				0x48
-#define	PACKET3_PREAMBLE_CNTL				0x4A
-#              define PACKET3_PREAMBLE_BEGIN_CLEAR_STATE     (2 << 28)
-#              define PACKET3_PREAMBLE_END_CLEAR_STATE       (3 << 28)
-#define	PACKET3_ONE_REG_WRITE				0x57
-#define	PACKET3_LOAD_CONFIG_REG				0x5F
-#define	PACKET3_LOAD_CONTEXT_REG			0x60
-#define	PACKET3_LOAD_SH_REG				0x61
-#define	PACKET3_SET_CONFIG_REG				0x68
-#define		PACKET3_SET_CONFIG_REG_START			0x00002000
-#define		PACKET3_SET_CONFIG_REG_END			0x00002c00
-#define	PACKET3_SET_CONTEXT_REG				0x69
-#define		PACKET3_SET_CONTEXT_REG_START			0x000a000
-#define		PACKET3_SET_CONTEXT_REG_END			0x000a400
-#define	PACKET3_SET_CONTEXT_REG_INDIRECT		0x73
-#define	PACKET3_SET_RESOURCE_INDIRECT			0x74
-#define	PACKET3_SET_SH_REG				0x76
-#define		PACKET3_SET_SH_REG_START			0x00002c00
-#define		PACKET3_SET_SH_REG_END				0x00003000
-#define	PACKET3_SET_SH_REG_OFFSET			0x77
-#define	PACKET3_ME_WRITE				0x7A
-#define	PACKET3_SCRATCH_RAM_WRITE			0x7D
-#define	PACKET3_SCRATCH_RAM_READ			0x7E
-#define	PACKET3_CE_WRITE				0x7F
-#define	PACKET3_LOAD_CONST_RAM				0x80
-#define	PACKET3_WRITE_CONST_RAM				0x81
-#define	PACKET3_WRITE_CONST_RAM_OFFSET			0x82
-#define	PACKET3_DUMP_CONST_RAM				0x83
-#define	PACKET3_INCREMENT_CE_COUNTER			0x84
-#define	PACKET3_INCREMENT_DE_COUNTER			0x85
-#define	PACKET3_WAIT_ON_CE_COUNTER			0x86
-#define	PACKET3_WAIT_ON_DE_COUNTER			0x87
-#define	PACKET3_WAIT_ON_DE_COUNTER_DIFF			0x88
-#define	PACKET3_SET_CE_DE_COUNTERS			0x89
-#define	PACKET3_WAIT_ON_AVAIL_BUFFER			0x8A
-#define	PACKET3_SWITCH_BUFFER				0x8B
-#define PACKET3_SEM_WAIT_ON_SIGNAL    (0x1 << 12)
-#define PACKET3_SEM_SEL_SIGNAL	    (0x6 << 29)
-#define PACKET3_SEM_SEL_WAIT	    (0x7 << 29)
-
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/si_ih.c b/drivers/gpu/drm/amd/amdgpu/si_ih.c
index 9a24f17a5750..66f650f87243 100644
--- a/drivers/gpu/drm/amd/amdgpu/si_ih.c
+++ b/drivers/gpu/drm/amd/amdgpu/si_ih.c
@@ -27,6 +27,7 @@
 #include "amdgpu_ih.h"
 #include "sid.h"
 #include "si_ih.h"
+
 #include "oss/oss_1_0_d.h"
 #include "oss/oss_1_0_sh_mask.h"
 
@@ -95,6 +96,9 @@ static int si_ih_irq_init(struct amdgpu_device *adev)
 	pci_set_master(adev->pdev);
 	si_ih_enable_interrupts(adev);
 
+	if (adev->irq.ih_soft.ring_size)
+		adev->irq.ih_soft.enabled = true;
+
 	return 0;
 }
 
@@ -111,6 +115,9 @@ static u32 si_ih_get_wptr(struct amdgpu_device *adev,
 
 	wptr = le32_to_cpu(*ih->wptr_cpu);
 
+	if (ih == &adev->irq.ih_soft)
+		goto out;
+
 	if (wptr & IH_RB_WPTR__RB_OVERFLOW_MASK) {
 		wptr &= ~IH_RB_WPTR__RB_OVERFLOW_MASK;
 		dev_warn(adev->dev, "IH ring buffer overflow (0x%08X, 0x%08X, 0x%08X)\n",
@@ -119,7 +126,15 @@ static u32 si_ih_get_wptr(struct amdgpu_device *adev,
 		tmp = RREG32(IH_RB_CNTL);
 		tmp |= IH_RB_CNTL__WPTR_OVERFLOW_CLEAR_MASK;
 		WREG32(IH_RB_CNTL, tmp);
+
+		/* Unset the CLEAR_OVERFLOW bit immediately so new overflows
+		 * can be detected.
+		 */
+		tmp &= ~IH_RB_CNTL__WPTR_OVERFLOW_CLEAR_MASK;
+		WREG32(IH_RB_CNTL, tmp);
 	}
+
+out:
 	return (wptr & ih->ptr_mask);
 }
 
@@ -150,70 +165,68 @@ static void si_ih_set_rptr(struct amdgpu_device *adev,
 	WREG32(IH_RB_RPTR, ih->rptr);
 }
 
-static int si_ih_early_init(void *handle)
+static int si_ih_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	si_ih_set_interrupt_funcs(adev);
 
 	return 0;
 }
 
-static int si_ih_sw_init(void *handle)
+static int si_ih_sw_init(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	r = amdgpu_ih_ring_init(adev, &adev->irq.ih, 64 * 1024, false);
 	if (r)
 		return r;
 
+	r = amdgpu_ih_ring_init(adev, &adev->irq.ih_soft, IH_SW_RING_SIZE, true);
+	if (r)
+		return r;
+
 	return amdgpu_irq_init(adev);
 }
 
-static int si_ih_sw_fini(void *handle)
+static int si_ih_sw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	amdgpu_irq_fini_sw(adev);
 
 	return 0;
 }
 
-static int si_ih_hw_init(void *handle)
+static int si_ih_hw_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	return si_ih_irq_init(adev);
 }
 
-static int si_ih_hw_fini(void *handle)
+static int si_ih_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	si_ih_irq_disable(adev);
+	si_ih_irq_disable(ip_block->adev);
 
 	return 0;
 }
 
-static int si_ih_suspend(void *handle)
+static int si_ih_suspend(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	return si_ih_hw_fini(adev);
+	return si_ih_hw_fini(ip_block);
 }
 
-static int si_ih_resume(void *handle)
+static int si_ih_resume(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	return si_ih_hw_init(adev);
+	return si_ih_hw_init(ip_block);
 }
 
-static bool si_ih_is_idle(void *handle)
+static bool si_ih_is_idle(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-	u32 tmp = RREG32(SRBM_STATUS);
+	struct amdgpu_device *adev = ip_block->adev;
+	u32 tmp = RREG32(mmSRBM_STATUS);
 
 	if (tmp & SRBM_STATUS__IH_BUSY_MASK)
 		return false;
@@ -221,41 +234,41 @@ static bool si_ih_is_idle(void *handle)
 	return true;
 }
 
-static int si_ih_wait_for_idle(void *handle)
+static int si_ih_wait_for_idle(struct amdgpu_ip_block *ip_block)
 {
 	unsigned i;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	for (i = 0; i < adev->usec_timeout; i++) {
-		if (si_ih_is_idle(handle))
+		if (si_ih_is_idle(ip_block))
 			return 0;
 		udelay(1);
 	}
 	return -ETIMEDOUT;
 }
 
-static int si_ih_soft_reset(void *handle)
+static int si_ih_soft_reset(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	u32 srbm_soft_reset = 0;
-	u32 tmp = RREG32(SRBM_STATUS);
+	u32 tmp = RREG32(mmSRBM_STATUS);
 
 	if (tmp & SRBM_STATUS__IH_BUSY_MASK)
 		srbm_soft_reset |= SRBM_SOFT_RESET__SOFT_RESET_IH_MASK;
 
 	if (srbm_soft_reset) {
-		tmp = RREG32(SRBM_SOFT_RESET);
+		tmp = RREG32(mmSRBM_SOFT_RESET);
 		tmp |= srbm_soft_reset;
-		dev_info(adev->dev, "SRBM_SOFT_RESET=0x%08X\n", tmp);
-		WREG32(SRBM_SOFT_RESET, tmp);
-		tmp = RREG32(SRBM_SOFT_RESET);
+		dev_info(adev->dev, "mmSRBM_SOFT_RESET=0x%08X\n", tmp);
+		WREG32(mmSRBM_SOFT_RESET, tmp);
+		tmp = RREG32(mmSRBM_SOFT_RESET);
 
 		udelay(50);
 
 		tmp &= ~srbm_soft_reset;
-		WREG32(SRBM_SOFT_RESET, tmp);
-		tmp = RREG32(SRBM_SOFT_RESET);
+		WREG32(mmSRBM_SOFT_RESET, tmp);
+		tmp = RREG32(mmSRBM_SOFT_RESET);
 
 		udelay(50);
 	}
@@ -263,13 +276,13 @@ static int si_ih_soft_reset(void *handle)
 	return 0;
 }
 
-static int si_ih_set_clockgating_state(void *handle,
+static int si_ih_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_clockgating_state state)
 {
 	return 0;
 }
 
-static int si_ih_set_powergating_state(void *handle,
+static int si_ih_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_powergating_state state)
 {
 	return 0;
@@ -278,7 +291,6 @@ static int si_ih_set_powergating_state(void *handle,
 static const struct amd_ip_funcs si_ih_ip_funcs = {
 	.name = "si_ih",
 	.early_init = si_ih_early_init,
-	.late_init = NULL,
 	.sw_init = si_ih_sw_init,
 	.sw_fini = si_ih_sw_fini,
 	.hw_init = si_ih_hw_init,
diff --git a/drivers/gpu/drm/amd/amdgpu/sid.h b/drivers/gpu/drm/amd/amdgpu/sid.h
index 9a39cbfe6db9..561462a8332e 100644
--- a/drivers/gpu/drm/amd/amdgpu/sid.h
+++ b/drivers/gpu/drm/amd/amdgpu/sid.h
@@ -24,47 +24,12 @@
 #ifndef SI_H
 #define SI_H
 
-#define TAHITI_RB_BITMAP_WIDTH_PER_SH  2
-
-#define TAHITI_GB_ADDR_CONFIG_GOLDEN        0x12011003
-#define VERDE_GB_ADDR_CONFIG_GOLDEN         0x12010002
-#define HAINAN_GB_ADDR_CONFIG_GOLDEN        0x02010001
-
-#define SI_MAX_SH_GPRS		 	256
-#define SI_MAX_TEMP_GPRS         	16
-#define SI_MAX_SH_THREADS        	256
-#define SI_MAX_SH_STACK_ENTRIES  	4096
-#define SI_MAX_FRC_EOV_CNT       	16384
-#define SI_MAX_BACKENDS          	8
-#define SI_MAX_BACKENDS_MASK     	0xFF
-#define SI_MAX_BACKENDS_PER_SE_MASK     0x0F
-#define SI_MAX_SIMDS             	12
-#define SI_MAX_SIMDS_MASK        	0x0FFF
-#define SI_MAX_SIMDS_PER_SE_MASK        0x00FF
-#define SI_MAX_PIPES            	8
-#define SI_MAX_PIPES_MASK        	0xFF
-#define SI_MAX_PIPES_PER_SIMD_MASK      0x3F
-#define SI_MAX_LDS_NUM           	0xFFFF
-#define SI_MAX_TCC               	16
-#define SI_MAX_TCC_MASK          	0xFFFF
 #define SI_MAX_CTLACKS_ASSERTION_WAIT   100
 
-/* SMC IND accessor regs */
-#define SMC_IND_INDEX_0                              0x80
-#define SMC_IND_DATA_0                               0x81
-
-#define SMC_IND_ACCESS_CNTL                          0x8A
-#       define AUTO_INCREMENT_IND_0                  (1 << 0)
-#define SMC_MESSAGE_0                                0x8B
-#define SMC_RESP_0                                   0x8C
-
 /* CG IND registers are accessed via SMC indirect space + SMC_CG_IND_START */
 #define SMC_CG_IND_START                    0xc0030000
 #define SMC_CG_IND_END                      0xc0040000
 
-#define	CG_CGTT_LOCAL_0				0x400
-#define	CG_CGTT_LOCAL_1				0x401
-
 /* SMC IND registers */
 #define	SMC_SYSCON_RESET_CNTL				0x80000000
 #       define RST_REG                                  (1 << 0)
@@ -72,9 +37,6 @@
 #       define CK_DISABLE                               (1 << 0)
 #       define CKEN                                     (1 << 24)
 
-#define VGA_HDP_CONTROL  				0xCA
-#define		VGA_MEMORY_DISABLE				(1 << 4)
-
 #define DCCG_DISP_SLOW_SELECT_REG                       0x13F
 #define		DCCG_DISP1_SLOW_SELECT(x)		((x) << 0)
 #define		DCCG_DISP1_SLOW_SELECT_MASK		(7 << 0)
@@ -83,47 +45,6 @@
 #define		DCCG_DISP2_SLOW_SELECT_MASK		(7 << 4)
 #define		DCCG_DISP2_SLOW_SELECT_SHIFT		4
 
-#define	CG_SPLL_FUNC_CNTL				0x180
-#define		SPLL_RESET				(1 << 0)
-#define		SPLL_SLEEP				(1 << 1)
-#define		SPLL_BYPASS_EN				(1 << 3)
-#define		SPLL_REF_DIV(x)				((x) << 4)
-#define		SPLL_REF_DIV_MASK			(0x3f << 4)
-#define		SPLL_PDIV_A(x)				((x) << 20)
-#define		SPLL_PDIV_A_MASK			(0x7f << 20)
-#define		SPLL_PDIV_A_SHIFT			20
-#define	CG_SPLL_FUNC_CNTL_2				0x181
-#define		SCLK_MUX_SEL(x)				((x) << 0)
-#define		SCLK_MUX_SEL_MASK			(0x1ff << 0)
-#define		SPLL_CTLREQ_CHG				(1 << 23)
-#define		SCLK_MUX_UPDATE				(1 << 26)
-#define	CG_SPLL_FUNC_CNTL_3				0x182
-#define		SPLL_FB_DIV(x)				((x) << 0)
-#define		SPLL_FB_DIV_MASK			(0x3ffffff << 0)
-#define		SPLL_FB_DIV_SHIFT			0
-#define		SPLL_DITHEN				(1 << 28)
-#define	CG_SPLL_FUNC_CNTL_4				0x183
-
-#define	SPLL_STATUS					0x185
-#define		SPLL_CHG_STATUS				(1 << 1)
-#define	SPLL_CNTL_MODE					0x186
-#define		SPLL_SW_DIR_CONTROL			(1 << 0)
-#	define SPLL_REFCLK_SEL(x)			((x) << 26)
-#	define SPLL_REFCLK_SEL_MASK			(3 << 26)
-
-#define	CG_SPLL_SPREAD_SPECTRUM				0x188
-#define		SSEN					(1 << 0)
-#define		CLK_S(x)				((x) << 4)
-#define		CLK_S_MASK				(0xfff << 4)
-#define		CLK_S_SHIFT				4
-#define	CG_SPLL_SPREAD_SPECTRUM_2			0x189
-#define		CLK_V(x)				((x) << 0)
-#define		CLK_V_MASK				(0x3ffffff << 0)
-#define		CLK_V_SHIFT				0
-
-#define	CG_SPLL_AUTOSCALE_CNTL				0x18b
-#       define AUTOSCALE_ON_SS_CLEAR                    (1 << 9)
-
 /* discrete uvd clocks */
 #define	CG_UPLL_FUNC_CNTL				0x18d
 #	define UPLL_RESET_MASK				0x00000001
@@ -153,317 +74,13 @@
 #define	CG_UPLL_SPREAD_SPECTRUM				0x194
 #	define SSEN_MASK				0x00000001
 
-#define	MPLL_BYPASSCLK_SEL				0x197
-#	define MPLL_CLKOUT_SEL(x)			((x) << 8)
-#	define MPLL_CLKOUT_SEL_MASK			0xFF00
-
-#define CG_CLKPIN_CNTL                                    0x198
-#       define XTALIN_DIVIDE                              (1 << 1)
-#       define BCLK_AS_XCLK                               (1 << 2)
-#define CG_CLKPIN_CNTL_2                                  0x199
-#       define FORCE_BIF_REFCLK_EN                        (1 << 3)
-#       define MUX_TCLK_TO_XCLK                           (1 << 8)
-
-#define	THM_CLK_CNTL					0x19b
-#	define CMON_CLK_SEL(x)				((x) << 0)
-#	define CMON_CLK_SEL_MASK			0xFF
-#	define TMON_CLK_SEL(x)				((x) << 8)
-#	define TMON_CLK_SEL_MASK			0xFF00
-#define	MISC_CLK_CNTL					0x19c
-#	define DEEP_SLEEP_CLK_SEL(x)			((x) << 0)
-#	define DEEP_SLEEP_CLK_SEL_MASK			0xFF
-#	define ZCLK_SEL(x)				((x) << 8)
-#	define ZCLK_SEL_MASK				0xFF00
-
-#define	CG_THERMAL_CTRL					0x1c0
-#define 	DPM_EVENT_SRC(x)			((x) << 0)
-#define 	DPM_EVENT_SRC_MASK			(7 << 0)
-#define		DIG_THERM_DPM(x)			((x) << 14)
-#define		DIG_THERM_DPM_MASK			0x003FC000
-#define		DIG_THERM_DPM_SHIFT			14
-#define	CG_THERMAL_STATUS				0x1c1
-#define		FDO_PWM_DUTY(x)				((x) << 9)
-#define		FDO_PWM_DUTY_MASK			(0xff << 9)
-#define		FDO_PWM_DUTY_SHIFT			9
-#define	CG_THERMAL_INT					0x1c2
-#define		DIG_THERM_INTH(x)			((x) << 8)
-#define		DIG_THERM_INTH_MASK			0x0000FF00
-#define		DIG_THERM_INTH_SHIFT			8
-#define		DIG_THERM_INTL(x)			((x) << 16)
-#define		DIG_THERM_INTL_MASK			0x00FF0000
-#define		DIG_THERM_INTL_SHIFT			16
-#define 	THERM_INT_MASK_HIGH			(1 << 24)
-#define 	THERM_INT_MASK_LOW			(1 << 25)
-
-#define	CG_MULT_THERMAL_CTRL					0x1c4
-#define		TEMP_SEL(x)					((x) << 20)
-#define		TEMP_SEL_MASK					(0xff << 20)
-#define		TEMP_SEL_SHIFT					20
-#define	CG_MULT_THERMAL_STATUS					0x1c5
-#define		ASIC_MAX_TEMP(x)				((x) << 0)
-#define		ASIC_MAX_TEMP_MASK				0x000001ff
-#define		ASIC_MAX_TEMP_SHIFT				0
-#define		CTF_TEMP(x)					((x) << 9)
-#define		CTF_TEMP_MASK					0x0003fe00
-#define		CTF_TEMP_SHIFT					9
-
-#define	CG_FDO_CTRL0					0x1d5
-#define		FDO_STATIC_DUTY(x)			((x) << 0)
-#define		FDO_STATIC_DUTY_MASK			0x000000FF
-#define		FDO_STATIC_DUTY_SHIFT			0
-#define	CG_FDO_CTRL1					0x1d6
-#define		FMAX_DUTY100(x)				((x) << 0)
-#define		FMAX_DUTY100_MASK			0x000000FF
-#define		FMAX_DUTY100_SHIFT			0
-#define	CG_FDO_CTRL2					0x1d7
-#define		TMIN(x)					((x) << 0)
-#define		TMIN_MASK				0x000000FF
-#define		TMIN_SHIFT				0
-#define		FDO_PWM_MODE(x)				((x) << 11)
-#define		FDO_PWM_MODE_MASK			(7 << 11)
-#define		FDO_PWM_MODE_SHIFT			11
-#define		TACH_PWM_RESP_RATE(x)			((x) << 25)
-#define		TACH_PWM_RESP_RATE_MASK			(0x7f << 25)
-#define		TACH_PWM_RESP_RATE_SHIFT		25
-
-#define CG_TACH_CTRL                                    0x1dc
-#       define EDGE_PER_REV(x)                          ((x) << 0)
-#       define EDGE_PER_REV_MASK                        (0x7 << 0)
-#       define EDGE_PER_REV_SHIFT                       0
-#       define TARGET_PERIOD(x)                         ((x) << 3)
-#       define TARGET_PERIOD_MASK                       0xfffffff8
-#       define TARGET_PERIOD_SHIFT                      3
-#define CG_TACH_STATUS                                  0x1dd
-#       define TACH_PERIOD(x)                           ((x) << 0)
-#       define TACH_PERIOD_MASK                         0xffffffff
-#       define TACH_PERIOD_SHIFT                        0
-
-#define GENERAL_PWRMGT                                  0x1e0
-#       define GLOBAL_PWRMGT_EN                         (1 << 0)
-#       define STATIC_PM_EN                             (1 << 1)
-#       define THERMAL_PROTECTION_DIS                   (1 << 2)
-#       define THERMAL_PROTECTION_TYPE                  (1 << 3)
-#       define SW_SMIO_INDEX(x)                         ((x) << 6)
-#       define SW_SMIO_INDEX_MASK                       (1 << 6)
-#       define SW_SMIO_INDEX_SHIFT                      6
-#       define VOLT_PWRMGT_EN                           (1 << 10)
-#       define DYN_SPREAD_SPECTRUM_EN                   (1 << 23)
-#define CG_TPC                                            0x1e1
-#define SCLK_PWRMGT_CNTL                                  0x1e2
-#       define SCLK_PWRMGT_OFF                            (1 << 0)
-#       define SCLK_LOW_D1                                (1 << 1)
-#       define FIR_RESET                                  (1 << 4)
-#       define FIR_FORCE_TREND_SEL                        (1 << 5)
-#       define FIR_TREND_MODE                             (1 << 6)
-#       define DYN_GFX_CLK_OFF_EN                         (1 << 7)
-#       define GFX_CLK_FORCE_ON                           (1 << 8)
-#       define GFX_CLK_REQUEST_OFF                        (1 << 9)
-#       define GFX_CLK_FORCE_OFF                          (1 << 10)
-#       define GFX_CLK_OFF_ACPI_D1                        (1 << 11)
-#       define GFX_CLK_OFF_ACPI_D2                        (1 << 12)
-#       define GFX_CLK_OFF_ACPI_D3                        (1 << 13)
-#       define DYN_LIGHT_SLEEP_EN                         (1 << 14)
-
-#define TARGET_AND_CURRENT_PROFILE_INDEX                  0x1e6
-#       define CURRENT_STATE_INDEX_MASK                   (0xf << 4)
-#       define CURRENT_STATE_INDEX_SHIFT                  4
-
-#define CG_FTV                                            0x1ef
-
-#define CG_FFCT_0                                         0x1f0
-#       define UTC_0(x)                                   ((x) << 0)
-#       define UTC_0_MASK                                 (0x3ff << 0)
-#       define DTC_0(x)                                   ((x) << 10)
-#       define DTC_0_MASK                                 (0x3ff << 10)
-
-#define CG_BSP                                          0x1ff
-#       define BSP(x)					((x) << 0)
-#       define BSP_MASK					(0xffff << 0)
-#       define BSU(x)					((x) << 16)
-#       define BSU_MASK					(0xf << 16)
-#define CG_AT                                           0x200
-#       define CG_R(x)					((x) << 0)
-#       define CG_R_MASK				(0xffff << 0)
-#       define CG_L(x)					((x) << 16)
-#       define CG_L_MASK				(0xffff << 16)
-
-#define CG_GIT                                          0x201
-#       define CG_GICST(x)                              ((x) << 0)
-#       define CG_GICST_MASK                            (0xffff << 0)
-#       define CG_GIPOT(x)                              ((x) << 16)
-#       define CG_GIPOT_MASK                            (0xffff << 16)
-
-#define CG_SSP                                            0x203
-#       define SST(x)                                     ((x) << 0)
-#       define SST_MASK                                   (0xffff << 0)
-#       define SSTU(x)                                    ((x) << 16)
-#       define SSTU_MASK                                  (0xf << 16)
-
-#define CG_DISPLAY_GAP_CNTL                               0x20a
-#       define DISP1_GAP(x)                               ((x) << 0)
-#       define DISP1_GAP_MASK                             (3 << 0)
-#       define DISP2_GAP(x)                               ((x) << 2)
-#       define DISP2_GAP_MASK                             (3 << 2)
-#       define VBI_TIMER_COUNT(x)                         ((x) << 4)
-#       define VBI_TIMER_COUNT_MASK                       (0x3fff << 4)
-#       define VBI_TIMER_UNIT(x)                          ((x) << 20)
-#       define VBI_TIMER_UNIT_MASK                        (7 << 20)
-#       define DISP1_GAP_MCHG(x)                          ((x) << 24)
-#       define DISP1_GAP_MCHG_MASK                        (3 << 24)
-#       define DISP2_GAP_MCHG(x)                          ((x) << 26)
-#       define DISP2_GAP_MCHG_MASK                        (3 << 26)
-
-#define	CG_ULV_CONTROL					0x21e
-#define	CG_ULV_PARAMETER				0x21f
-
-#define	SMC_SCRATCH0					0x221
-
-#define	CG_CAC_CTRL					0x22e
-#	define CAC_WINDOW(x)				((x) << 0)
-#	define CAC_WINDOW_MASK				0x00ffffff
-
-#define DMIF_ADDR_CONFIG  				0x2F5
-
-#define DMIF_ADDR_CALC  				0x300
-
-#define	PIPE0_DMIF_BUFFER_CONTROL			  0x0328
-#       define DMIF_BUFFERS_ALLOCATED(x)                  ((x) << 0)
-#       define DMIF_BUFFERS_ALLOCATED_COMPLETED           (1 << 4)
-
-#define	SRBM_STATUS				        0x394
-#define		GRBM_RQ_PENDING 			(1 << 5)
-#define		VMC_BUSY 				(1 << 8)
-#define		MCB_BUSY 				(1 << 9)
-#define		MCB_NON_DISPLAY_BUSY 			(1 << 10)
-#define		MCC_BUSY 				(1 << 11)
-#define		MCD_BUSY 				(1 << 12)
-#define		SEM_BUSY 				(1 << 14)
-#define		IH_BUSY 				(1 << 17)
-
-#define	SRBM_SOFT_RESET				        0x398
-#define		SOFT_RESET_BIF				(1 << 1)
-#define		SOFT_RESET_DC				(1 << 5)
-#define		SOFT_RESET_DMA1				(1 << 6)
-#define		SOFT_RESET_GRBM				(1 << 8)
-#define		SOFT_RESET_HDP				(1 << 9)
-#define		SOFT_RESET_IH				(1 << 10)
-#define		SOFT_RESET_MC				(1 << 11)
-#define		SOFT_RESET_ROM				(1 << 14)
-#define		SOFT_RESET_SEM				(1 << 15)
-#define		SOFT_RESET_VMC				(1 << 17)
-#define		SOFT_RESET_DMA				(1 << 20)
-#define		SOFT_RESET_TST				(1 << 21)
-#define		SOFT_RESET_REGBB			(1 << 22)
-#define		SOFT_RESET_ORB				(1 << 23)
-
-#define	CC_SYS_RB_BACKEND_DISABLE			0x3A0
-#define	GC_USER_SYS_RB_BACKEND_DISABLE			0x3A1
-
-#define SRBM_READ_ERROR					0x3A6
-#define SRBM_INT_CNTL					0x3A8
-#define SRBM_INT_ACK					0x3AA
-
-#define	SRBM_STATUS2				        0x3B1
-#define		DMA_BUSY 				(1 << 5)
-#define		DMA1_BUSY 				(1 << 6)
-
-#define VM_L2_CNTL					0x500
-#define		ENABLE_L2_CACHE					(1 << 0)
-#define		ENABLE_L2_FRAGMENT_PROCESSING			(1 << 1)
-#define		L2_CACHE_PTE_ENDIAN_SWAP_MODE(x)		((x) << 2)
-#define		L2_CACHE_PDE_ENDIAN_SWAP_MODE(x)		((x) << 4)
-#define		ENABLE_L2_PTE_CACHE_LRU_UPDATE_BY_WRITE		(1 << 9)
-#define		ENABLE_L2_PDE0_CACHE_LRU_UPDATE_BY_WRITE	(1 << 10)
-#define		EFFECTIVE_L2_QUEUE_SIZE(x)			(((x) & 7) << 15)
-#define		CONTEXT1_IDENTITY_ACCESS_MODE(x)		(((x) & 3) << 19)
-#define VM_L2_CNTL2					0x501
-#define		INVALIDATE_ALL_L1_TLBS				(1 << 0)
-#define		INVALIDATE_L2_CACHE				(1 << 1)
-#define		INVALIDATE_CACHE_MODE(x)			((x) << 26)
-#define			INVALIDATE_PTE_AND_PDE_CACHES		0
-#define			INVALIDATE_ONLY_PTE_CACHES		1
-#define			INVALIDATE_ONLY_PDE_CACHES		2
-#define VM_L2_CNTL3					0x502
-#define		BANK_SELECT(x)					((x) << 0)
-#define		L2_CACHE_UPDATE_MODE(x)				((x) << 6)
-#define		L2_CACHE_BIGK_FRAGMENT_SIZE(x)			((x) << 15)
-#define		L2_CACHE_BIGK_ASSOCIATIVITY			(1 << 20)
-#define	VM_L2_STATUS					0x503
-#define		L2_BUSY						(1 << 0)
-#define VM_CONTEXT0_CNTL				0x504
-#define		ENABLE_CONTEXT					(1 << 0)
-#define		PAGE_TABLE_DEPTH(x)				(((x) & 3) << 1)
-#define		RANGE_PROTECTION_FAULT_ENABLE_INTERRUPT		(1 << 3)
-#define		RANGE_PROTECTION_FAULT_ENABLE_DEFAULT		(1 << 4)
-#define		DUMMY_PAGE_PROTECTION_FAULT_ENABLE_INTERRUPT	(1 << 6)
-#define		DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT	(1 << 7)
-#define		PDE0_PROTECTION_FAULT_ENABLE_INTERRUPT		(1 << 9)
-#define		PDE0_PROTECTION_FAULT_ENABLE_DEFAULT		(1 << 10)
-#define		VALID_PROTECTION_FAULT_ENABLE_INTERRUPT		(1 << 12)
-#define		VALID_PROTECTION_FAULT_ENABLE_DEFAULT		(1 << 13)
-#define		READ_PROTECTION_FAULT_ENABLE_INTERRUPT		(1 << 15)
-#define		READ_PROTECTION_FAULT_ENABLE_DEFAULT		(1 << 16)
-#define		WRITE_PROTECTION_FAULT_ENABLE_INTERRUPT		(1 << 18)
-#define		WRITE_PROTECTION_FAULT_ENABLE_DEFAULT		(1 << 19)
-#define		PAGE_TABLE_BLOCK_SIZE(x)			(((x) & 0xF) << 24)
-#define VM_CONTEXT1_CNTL				0x505
-#define VM_CONTEXT0_CNTL2				0x50C
-#define VM_CONTEXT1_CNTL2				0x50D
-#define	VM_CONTEXT8_PAGE_TABLE_BASE_ADDR		0x50E
-#define	VM_CONTEXT9_PAGE_TABLE_BASE_ADDR		0x50F
-#define	VM_CONTEXT10_PAGE_TABLE_BASE_ADDR		0x510
-#define	VM_CONTEXT11_PAGE_TABLE_BASE_ADDR		0x511
-#define	VM_CONTEXT12_PAGE_TABLE_BASE_ADDR		0x512
-#define	VM_CONTEXT13_PAGE_TABLE_BASE_ADDR		0x513
-#define	VM_CONTEXT14_PAGE_TABLE_BASE_ADDR		0x514
-#define	VM_CONTEXT15_PAGE_TABLE_BASE_ADDR		0x515
-
-#define	VM_CONTEXT1_PROTECTION_FAULT_ADDR		0x53f
-#define	VM_CONTEXT1_PROTECTION_FAULT_STATUS		0x537
-#define		PROTECTIONS_MASK			(0xf << 0)
-#define		PROTECTIONS_SHIFT			0
-		/* bit 0: range
-		 * bit 1: pde0
-		 * bit 2: valid
-		 * bit 3: read
-		 * bit 4: write
-		 */
-#define		MEMORY_CLIENT_ID_MASK			(0xff << 12)
-#define		MEMORY_CLIENT_ID_SHIFT			12
-#define		MEMORY_CLIENT_RW_MASK			(1 << 24)
-#define		MEMORY_CLIENT_RW_SHIFT			24
-#define		FAULT_VMID_MASK				(0xf << 25)
-#define		FAULT_VMID_SHIFT			25
-
 #define VM_INVALIDATE_REQUEST				0x51E
 #define VM_INVALIDATE_RESPONSE				0x51F
 
-#define VM_CONTEXT0_PROTECTION_FAULT_DEFAULT_ADDR	0x546
-#define VM_CONTEXT1_PROTECTION_FAULT_DEFAULT_ADDR	0x547
-
-#define	VM_CONTEXT0_PAGE_TABLE_BASE_ADDR		0x54F
-#define	VM_CONTEXT1_PAGE_TABLE_BASE_ADDR		0x550
-#define	VM_CONTEXT2_PAGE_TABLE_BASE_ADDR		0x551
-#define	VM_CONTEXT3_PAGE_TABLE_BASE_ADDR		0x552
-#define	VM_CONTEXT4_PAGE_TABLE_BASE_ADDR		0x553
-#define	VM_CONTEXT5_PAGE_TABLE_BASE_ADDR		0x554
-#define	VM_CONTEXT6_PAGE_TABLE_BASE_ADDR		0x555
-#define	VM_CONTEXT7_PAGE_TABLE_BASE_ADDR		0x556
-#define	VM_CONTEXT0_PAGE_TABLE_START_ADDR		0x557
-#define	VM_CONTEXT1_PAGE_TABLE_START_ADDR		0x558
-
-#define	VM_CONTEXT0_PAGE_TABLE_END_ADDR			0x55F
-#define	VM_CONTEXT1_PAGE_TABLE_END_ADDR			0x560
-
 #define VM_L2_CG           				0x570
 #define		MC_CG_ENABLE				(1 << 18)
 #define		MC_LS_ENABLE				(1 << 19)
 
-#define MC_SHARED_CHMAP						0x801
-#define		NOOFCHAN_SHIFT					12
-#define		NOOFCHAN_MASK					0x0000f000
-#define MC_SHARED_CHREMAP					0x802
-
 #define	MC_VM_FB_LOCATION				0x809
 #define	MC_VM_AGP_TOP					0x80A
 #define	MC_VM_AGP_BOT					0x80B
@@ -495,21 +112,6 @@
 #define MC_CITF_MISC_WR_CG           			0x993
 #define MC_CITF_MISC_VM_CG           			0x994
 
-#define	MC_ARB_RAMCFG					0x9D8
-#define		NOOFBANK_SHIFT					0
-#define		NOOFBANK_MASK					0x00000003
-#define		NOOFRANK_SHIFT					2
-#define		NOOFRANK_MASK					0x00000004
-#define		NOOFROWS_SHIFT					3
-#define		NOOFROWS_MASK					0x00000038
-#define		NOOFCOLS_SHIFT					6
-#define		NOOFCOLS_MASK					0x000000C0
-#define		CHANSIZE_SHIFT					8
-#define		CHANSIZE_MASK					0x00000100
-#define		CHANSIZE_OVERRIDE				(1 << 11)
-#define		NOOFGROUPS_SHIFT				12
-#define		NOOFGROUPS_MASK					0x00001000
-
 #define	MC_ARB_DRAM_TIMING				0x9DD
 #define	MC_ARB_DRAM_TIMING2				0x9DE
 
@@ -635,20 +237,6 @@
 #define		CLKS(x)					((x) << 0)
 #define		CLKS_MASK				(0xfff << 0)
 
-#define	HDP_HOST_PATH_CNTL				0xB00
-#define 	CLOCK_GATING_DIS			(1 << 23)
-#define	HDP_NONSURFACE_BASE				0xB01
-#define	HDP_NONSURFACE_INFO				0xB02
-#define	HDP_NONSURFACE_SIZE				0xB03
-
-#define HDP_DEBUG0  					0xBCC
-
-#define HDP_ADDR_CONFIG  				0xBD2
-#define HDP_MISC_CNTL					0xBD3
-#define 	HDP_FLUSH_INVALIDATE_CACHE			(1 << 0)
-#define HDP_MEM_POWER_LS				0xBD4
-#define 	HDP_LS_ENABLE				(1 << 0)
-
 #define ATC_MISC_CG           				0xCD4
 
 #define IH_RB_CNTL                                        0xF80
@@ -678,8 +266,6 @@
 #       define MC_WR_CLEAN_CNT(x)                         ((x) << 20)
 #       define MC_VMID(x)                                 ((x) << 25)
 
-#define	CONFIG_MEMSIZE					0x150A
-
 #define INTERRUPT_CNTL                                    0x151A
 #       define IH_DUMMY_RD_OVERRIDE                       (1 << 0)
 #       define IH_DUMMY_RD_EN                             (1 << 1)
@@ -687,734 +273,28 @@
 #       define GEN_IH_INT_EN                              (1 << 8)
 #define INTERRUPT_CNTL2                                   0x151B
 
-#define HDP_MEM_COHERENCY_FLUSH_CNTL			0x1520
-
-#define	BIF_FB_EN						0x1524
-#define		FB_READ_EN					(1 << 0)
-#define		FB_WRITE_EN					(1 << 1)
-
-#define HDP_REG_COHERENCY_FLUSH_CNTL			0x1528
-
-/* DCE6 ELD audio interface */
-#define AZ_F0_CODEC_ENDPOINT_INDEX                       0x1780
-#       define AZ_ENDPOINT_REG_INDEX(x)                  (((x) & 0xff) << 0)
-#       define AZ_ENDPOINT_REG_WRITE_EN                  (1 << 8)
-#define AZ_F0_CODEC_ENDPOINT_DATA                        0x1781
-
-#define AZ_F0_CODEC_PIN_CONTROL_CHANNEL_SPEAKER          0x25
-#define		SPEAKER_ALLOCATION(x)			(((x) & 0x7f) << 0)
-#define		SPEAKER_ALLOCATION_MASK			(0x7f << 0)
-#define		SPEAKER_ALLOCATION_SHIFT		0
-#define		HDMI_CONNECTION				(1 << 16)
-#define		DP_CONNECTION				(1 << 17)
-
-#define AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR0        0x28 /* LPCM */
-#define AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR1        0x29 /* AC3 */
-#define AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR2        0x2A /* MPEG1 */
-#define AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR3        0x2B /* MP3 */
-#define AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR4        0x2C /* MPEG2 */
-#define AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR5        0x2D /* AAC */
-#define AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR6        0x2E /* DTS */
-#define AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR7        0x2F /* ATRAC */
-#define AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR8        0x30 /* one bit audio - leave at 0 (default) */
-#define AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR9        0x31 /* Dolby Digital */
-#define AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR10       0x32 /* DTS-HD */
-#define AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR11       0x33 /* MAT-MLP */
-#define AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR12       0x34 /* DTS */
-#define AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR13       0x35 /* WMA Pro */
-#       define MAX_CHANNELS(x)                            (((x) & 0x7) << 0)
-/* max channels minus one.  7 = 8 channels */
-#       define SUPPORTED_FREQUENCIES(x)                   (((x) & 0xff) << 8)
-#       define DESCRIPTOR_BYTE_2(x)                       (((x) & 0xff) << 16)
-#       define SUPPORTED_FREQUENCIES_STEREO(x)            (((x) & 0xff) << 24) /* LPCM only */
-/* SUPPORTED_FREQUENCIES, SUPPORTED_FREQUENCIES_STEREO
- * bit0 = 32 kHz
- * bit1 = 44.1 kHz
- * bit2 = 48 kHz
- * bit3 = 88.2 kHz
- * bit4 = 96 kHz
- * bit5 = 176.4 kHz
- * bit6 = 192 kHz
- */
-
-#define AZ_F0_CODEC_PIN_CONTROL_RESPONSE_LIPSYNC         0x37
-#       define VIDEO_LIPSYNC(x)                           (((x) & 0xff) << 0)
-#       define AUDIO_LIPSYNC(x)                           (((x) & 0xff) << 8)
-/* VIDEO_LIPSYNC, AUDIO_LIPSYNC
- * 0   = invalid
- * x   = legal delay value
- * 255 = sync not supported
- */
-#define AZ_F0_CODEC_PIN_CONTROL_RESPONSE_HBR             0x38
-#       define HBR_CAPABLE                                (1 << 0) /* enabled by default */
-
-#define AZ_F0_CODEC_PIN_CONTROL_SINK_INFO0               0x3a
-#       define MANUFACTURER_ID(x)                        (((x) & 0xffff) << 0)
-#       define PRODUCT_ID(x)                             (((x) & 0xffff) << 16)
-#define AZ_F0_CODEC_PIN_CONTROL_SINK_INFO1               0x3b
-#       define SINK_DESCRIPTION_LEN(x)                   (((x) & 0xff) << 0)
-#define AZ_F0_CODEC_PIN_CONTROL_SINK_INFO2               0x3c
-#       define PORT_ID0(x)                               (((x) & 0xffffffff) << 0)
-#define AZ_F0_CODEC_PIN_CONTROL_SINK_INFO3               0x3d
-#       define PORT_ID1(x)                               (((x) & 0xffffffff) << 0)
-#define AZ_F0_CODEC_PIN_CONTROL_SINK_INFO4               0x3e
-#       define DESCRIPTION0(x)                           (((x) & 0xff) << 0)
-#       define DESCRIPTION1(x)                           (((x) & 0xff) << 8)
-#       define DESCRIPTION2(x)                           (((x) & 0xff) << 16)
-#       define DESCRIPTION3(x)                           (((x) & 0xff) << 24)
-#define AZ_F0_CODEC_PIN_CONTROL_SINK_INFO5               0x3f
-#       define DESCRIPTION4(x)                           (((x) & 0xff) << 0)
-#       define DESCRIPTION5(x)                           (((x) & 0xff) << 8)
-#       define DESCRIPTION6(x)                           (((x) & 0xff) << 16)
-#       define DESCRIPTION7(x)                           (((x) & 0xff) << 24)
-#define AZ_F0_CODEC_PIN_CONTROL_SINK_INFO6               0x40
-#       define DESCRIPTION8(x)                           (((x) & 0xff) << 0)
-#       define DESCRIPTION9(x)                           (((x) & 0xff) << 8)
-#       define DESCRIPTION10(x)                          (((x) & 0xff) << 16)
-#       define DESCRIPTION11(x)                          (((x) & 0xff) << 24)
-#define AZ_F0_CODEC_PIN_CONTROL_SINK_INFO7               0x41
-#       define DESCRIPTION12(x)                          (((x) & 0xff) << 0)
-#       define DESCRIPTION13(x)                          (((x) & 0xff) << 8)
-#       define DESCRIPTION14(x)                          (((x) & 0xff) << 16)
-#       define DESCRIPTION15(x)                          (((x) & 0xff) << 24)
-#define AZ_F0_CODEC_PIN_CONTROL_SINK_INFO8               0x42
-#       define DESCRIPTION16(x)                          (((x) & 0xff) << 0)
-#       define DESCRIPTION17(x)                          (((x) & 0xff) << 8)
-
-#define AZ_F0_CODEC_PIN_CONTROL_HOT_PLUG_CONTROL         0x54
-#       define AUDIO_ENABLED                             (1 << 31)
-
-#define AZ_F0_CODEC_PIN_CONTROL_RESPONSE_CONFIGURATION_DEFAULT  0x56
-#define		PORT_CONNECTIVITY_MASK				(3 << 30)
-#define		PORT_CONNECTIVITY_SHIFT				30
-
-#define	DC_LB_MEMORY_SPLIT					0x1AC3
-#define		DC_LB_MEMORY_CONFIG(x)				((x) << 20)
-
-#define	PRIORITY_A_CNT						0x1AC6
-#define		PRIORITY_MARK_MASK				0x7fff
-#define		PRIORITY_OFF					(1 << 16)
-#define		PRIORITY_ALWAYS_ON				(1 << 20)
-#define	PRIORITY_B_CNT						0x1AC7
-
-#define	DPG_PIPE_ARBITRATION_CONTROL3				0x1B32
-#       define LATENCY_WATERMARK_MASK(x)			((x) << 16)
-#define	DPG_PIPE_LATENCY_CONTROL				0x1B33
-#       define LATENCY_LOW_WATERMARK(x)				((x) << 0)
-#       define LATENCY_HIGH_WATERMARK(x)			((x) << 16)
-
-/* 0x6bb8, 0x77b8, 0x103b8, 0x10fb8, 0x11bb8, 0x127b8 */
-#define VLINE_STATUS                                    0x1AEE
-#       define VLINE_OCCURRED                           (1 << 0)
-#       define VLINE_ACK                                (1 << 4)
-#       define VLINE_STAT                               (1 << 12)
-#       define VLINE_INTERRUPT                          (1 << 16)
-#       define VLINE_INTERRUPT_TYPE                     (1 << 17)
-/* 0x6bbc, 0x77bc, 0x103bc, 0x10fbc, 0x11bbc, 0x127bc */
-#define VBLANK_STATUS                                   0x1AEF
-#       define VBLANK_OCCURRED                          (1 << 0)
-#       define VBLANK_ACK                               (1 << 4)
-#       define VBLANK_STAT                              (1 << 12)
-#       define VBLANK_INTERRUPT                         (1 << 16)
-#       define VBLANK_INTERRUPT_TYPE                    (1 << 17)
-
-/* 0x6b40, 0x7740, 0x10340, 0x10f40, 0x11b40, 0x12740 */
-#define INT_MASK                                        0x1AD0
-#       define VBLANK_INT_MASK                          (1 << 0)
-#       define VLINE_INT_MASK                           (1 << 4)
-
-#define DISP_INTERRUPT_STATUS                           0x183D
-#       define LB_D1_VLINE_INTERRUPT                    (1 << 2)
-#       define LB_D1_VBLANK_INTERRUPT                   (1 << 3)
-#       define DC_HPD1_INTERRUPT                        (1 << 17)
-#       define DC_HPD1_RX_INTERRUPT                     (1 << 18)
-#       define DACA_AUTODETECT_INTERRUPT                (1 << 22)
-#       define DACB_AUTODETECT_INTERRUPT                (1 << 23)
-#       define DC_I2C_SW_DONE_INTERRUPT                 (1 << 24)
-#       define DC_I2C_HW_DONE_INTERRUPT                 (1 << 25)
-#define DISP_INTERRUPT_STATUS_CONTINUE                  0x183E
-#       define LB_D2_VLINE_INTERRUPT                    (1 << 2)
-#       define LB_D2_VBLANK_INTERRUPT                   (1 << 3)
-#       define DC_HPD2_INTERRUPT                        (1 << 17)
-#       define DC_HPD2_RX_INTERRUPT                     (1 << 18)
-#       define DISP_TIMER_INTERRUPT                     (1 << 24)
-#define DISP_INTERRUPT_STATUS_CONTINUE2                 0x183F
-#       define LB_D3_VLINE_INTERRUPT                    (1 << 2)
-#       define LB_D3_VBLANK_INTERRUPT                   (1 << 3)
-#       define DC_HPD3_INTERRUPT                        (1 << 17)
-#       define DC_HPD3_RX_INTERRUPT                     (1 << 18)
-#define DISP_INTERRUPT_STATUS_CONTINUE3                 0x1840
-#       define LB_D4_VLINE_INTERRUPT                    (1 << 2)
-#       define LB_D4_VBLANK_INTERRUPT                   (1 << 3)
-#       define DC_HPD4_INTERRUPT                        (1 << 17)
-#       define DC_HPD4_RX_INTERRUPT                     (1 << 18)
-#define DISP_INTERRUPT_STATUS_CONTINUE4                 0x1853
-#       define LB_D5_VLINE_INTERRUPT                    (1 << 2)
-#       define LB_D5_VBLANK_INTERRUPT                   (1 << 3)
-#       define DC_HPD5_INTERRUPT                        (1 << 17)
-#       define DC_HPD5_RX_INTERRUPT                     (1 << 18)
-#define DISP_INTERRUPT_STATUS_CONTINUE5                 0x1854
-#       define LB_D6_VLINE_INTERRUPT                    (1 << 2)
-#       define LB_D6_VBLANK_INTERRUPT                   (1 << 3)
-#       define DC_HPD6_INTERRUPT                        (1 << 17)
-#       define DC_HPD6_RX_INTERRUPT                     (1 << 18)
-
-/* 0x6858, 0x7458, 0x10058, 0x10c58, 0x11858, 0x12458 */
-#define GRPH_INT_STATUS                                 0x1A16
-#       define GRPH_PFLIP_INT_OCCURRED                  (1 << 0)
-#       define GRPH_PFLIP_INT_CLEAR                     (1 << 8)
-/* 0x685c, 0x745c, 0x1005c, 0x10c5c, 0x1185c, 0x1245c */
-#define	GRPH_INT_CONTROL			        0x1A17
-#       define GRPH_PFLIP_INT_MASK                      (1 << 0)
-#       define GRPH_PFLIP_INT_TYPE                      (1 << 8)
-
-#define	DAC_AUTODETECT_INT_CONTROL			0x19F2
-
-#define DC_HPD1_INT_STATUS                              0x1807
-#define DC_HPD2_INT_STATUS                              0x180A
-#define DC_HPD3_INT_STATUS                              0x180D
-#define DC_HPD4_INT_STATUS                              0x1810
-#define DC_HPD5_INT_STATUS                              0x1813
-#define DC_HPD6_INT_STATUS                              0x1816
-#       define DC_HPDx_INT_STATUS                       (1 << 0)
-#       define DC_HPDx_SENSE                            (1 << 1)
-#       define DC_HPDx_RX_INT_STATUS                    (1 << 8)
-
-#define DC_HPD1_INT_CONTROL                             0x1808
-#define DC_HPD2_INT_CONTROL                             0x180B
-#define DC_HPD3_INT_CONTROL                             0x180E
-#define DC_HPD4_INT_CONTROL                             0x1811
-#define DC_HPD5_INT_CONTROL                             0x1814
-#define DC_HPD6_INT_CONTROL                             0x1817
-#       define DC_HPDx_INT_ACK                          (1 << 0)
-#       define DC_HPDx_INT_POLARITY                     (1 << 8)
-#       define DC_HPDx_INT_EN                           (1 << 16)
-#       define DC_HPDx_RX_INT_ACK                       (1 << 20)
-#       define DC_HPDx_RX_INT_EN                        (1 << 24)
-
-#define DC_HPD1_CONTROL                                   0x1809
-#define DC_HPD2_CONTROL                                   0x180C
-#define DC_HPD3_CONTROL                                   0x180F
-#define DC_HPD4_CONTROL                                   0x1812
-#define DC_HPD5_CONTROL                                   0x1815
-#define DC_HPD6_CONTROL                                   0x1818
-#       define DC_HPDx_CONNECTION_TIMER(x)                ((x) << 0)
-#       define DC_HPDx_RX_INT_TIMER(x)                    ((x) << 16)
-#       define DC_HPDx_EN                                 (1 << 28)
-
-#define DPG_PIPE_STUTTER_CONTROL                          0x1B35
-#       define STUTTER_ENABLE                             (1 << 0)
-
-/* 0x6e98, 0x7a98, 0x10698, 0x11298, 0x11e98, 0x12a98 */
-#define CRTC_STATUS_FRAME_COUNT                         0x1BA6
-
-/* Audio clocks */
-#define DCCG_AUDIO_DTO_SOURCE                           0x05ac
-#       define DCCG_AUDIO_DTO0_SOURCE_SEL(x) ((x) << 0) /* crtc0 - crtc5 */
-#       define DCCG_AUDIO_DTO_SEL            (1 << 4)   /* 0=dto0 1=dto1 */
-
-#define DCCG_AUDIO_DTO0_PHASE                           0x05b0
-#define DCCG_AUDIO_DTO0_MODULE                          0x05b4
-#define DCCG_AUDIO_DTO1_PHASE                           0x05c0
-#define DCCG_AUDIO_DTO1_MODULE                          0x05c4
-
-#define AFMT_AUDIO_SRC_CONTROL                          0x1c4f
-#define		AFMT_AUDIO_SRC_SELECT(x)		(((x) & 7) << 0)
-/* AFMT_AUDIO_SRC_SELECT
- * 0 = stream0
- * 1 = stream1
- * 2 = stream2
- * 3 = stream3
- * 4 = stream4
- * 5 = stream5
- */
-
-#define	GRBM_CNTL					0x2000
-#define		GRBM_READ_TIMEOUT(x)				((x) << 0)
-
-#define	GRBM_STATUS2					0x2002
-#define		RLC_RQ_PENDING 					(1 << 0)
-#define		RLC_BUSY 					(1 << 8)
-#define		TC_BUSY 					(1 << 9)
-
-#define	GRBM_STATUS					0x2004
-#define		CMDFIFO_AVAIL_MASK				0x0000000F
-#define		RING2_RQ_PENDING				(1 << 4)
-#define		SRBM_RQ_PENDING					(1 << 5)
-#define		RING1_RQ_PENDING				(1 << 6)
-#define		CF_RQ_PENDING					(1 << 7)
-#define		PF_RQ_PENDING					(1 << 8)
-#define		GDS_DMA_RQ_PENDING				(1 << 9)
-#define		GRBM_EE_BUSY					(1 << 10)
-#define		DB_CLEAN					(1 << 12)
-#define		CB_CLEAN					(1 << 13)
-#define		TA_BUSY 					(1 << 14)
-#define		GDS_BUSY 					(1 << 15)
-#define		VGT_BUSY					(1 << 17)
-#define		IA_BUSY_NO_DMA					(1 << 18)
-#define		IA_BUSY						(1 << 19)
-#define		SX_BUSY 					(1 << 20)
-#define		SPI_BUSY					(1 << 22)
-#define		BCI_BUSY					(1 << 23)
-#define		SC_BUSY 					(1 << 24)
-#define		PA_BUSY 					(1 << 25)
-#define		DB_BUSY 					(1 << 26)
-#define		CP_COHERENCY_BUSY      				(1 << 28)
-#define		CP_BUSY 					(1 << 29)
-#define		CB_BUSY 					(1 << 30)
-#define		GUI_ACTIVE					(1 << 31)
-#define	GRBM_STATUS_SE0					0x2005
-#define	GRBM_STATUS_SE1					0x2006
-#define		SE_DB_CLEAN					(1 << 1)
-#define		SE_CB_CLEAN					(1 << 2)
-#define		SE_BCI_BUSY					(1 << 22)
-#define		SE_VGT_BUSY					(1 << 23)
-#define		SE_PA_BUSY					(1 << 24)
-#define		SE_TA_BUSY					(1 << 25)
-#define		SE_SX_BUSY					(1 << 26)
-#define		SE_SPI_BUSY					(1 << 27)
-#define		SE_SC_BUSY					(1 << 29)
-#define		SE_DB_BUSY					(1 << 30)
-#define		SE_CB_BUSY					(1 << 31)
-
-#define	GRBM_SOFT_RESET					0x2008
-#define		SOFT_RESET_CP					(1 << 0)
-#define		SOFT_RESET_CB					(1 << 1)
-#define		SOFT_RESET_RLC					(1 << 2)
-#define		SOFT_RESET_DB					(1 << 3)
-#define		SOFT_RESET_GDS					(1 << 4)
-#define		SOFT_RESET_PA					(1 << 5)
-#define		SOFT_RESET_SC					(1 << 6)
-#define		SOFT_RESET_BCI					(1 << 7)
-#define		SOFT_RESET_SPI					(1 << 8)
-#define		SOFT_RESET_SX					(1 << 10)
-#define		SOFT_RESET_TC					(1 << 11)
-#define		SOFT_RESET_TA					(1 << 12)
-#define		SOFT_RESET_VGT					(1 << 14)
-#define		SOFT_RESET_IA					(1 << 15)
-
-#define GRBM_GFX_INDEX          			0x200B
-#define		INSTANCE_INDEX(x)			((x) << 0)
-#define		SH_INDEX(x)     			((x) << 8)
-#define		SE_INDEX(x)     			((x) << 16)
-#define		SH_BROADCAST_WRITES      		(1 << 29)
-#define		INSTANCE_BROADCAST_WRITES      		(1 << 30)
-#define		SE_BROADCAST_WRITES      		(1 << 31)
-
-#define GRBM_INT_CNTL                                   0x2018
-#       define RDERR_INT_ENABLE                         (1 << 0)
-#       define GUI_IDLE_INT_ENABLE                      (1 << 19)
-
-#define	CP_STRMOUT_CNTL					0x213F
-#define	SCRATCH_REG0					0x2140
-#define	SCRATCH_REG1					0x2141
-#define	SCRATCH_REG2					0x2142
-#define	SCRATCH_REG3					0x2143
-#define	SCRATCH_REG4					0x2144
-#define	SCRATCH_REG5					0x2145
-#define	SCRATCH_REG6					0x2146
-#define	SCRATCH_REG7					0x2147
-
-#define	SCRATCH_UMSK					0x2150
-#define	SCRATCH_ADDR					0x2151
-
-#define	CP_SEM_WAIT_TIMER				0x216F
-
-#define	CP_SEM_INCOMPLETE_TIMER_CNTL			0x2172
-
-#define CP_ME_CNTL					0x21B6
-#define		CP_CE_HALT					(1 << 24)
-#define		CP_PFP_HALT					(1 << 26)
-#define		CP_ME_HALT					(1 << 28)
-
-#define	CP_COHER_CNTL2					0x217A
-
-#define	CP_RB2_RPTR					0x21BE
-#define	CP_RB1_RPTR					0x21BF
-#define	CP_RB0_RPTR					0x21C0
-#define	CP_RB_WPTR_DELAY				0x21C1
-
-#define	CP_QUEUE_THRESHOLDS				0x21D8
-#define		ROQ_IB1_START(x)				((x) << 0)
-#define		ROQ_IB2_START(x)				((x) << 8)
-#define CP_MEQ_THRESHOLDS				0x21D9
-#define		MEQ1_START(x)				((x) << 0)
-#define		MEQ2_START(x)				((x) << 8)
-
-#define	CP_PERFMON_CNTL					0x21FF
-
 #define	VGT_VTX_VECT_EJECT_REG				0x222C
-
-#define	VGT_CACHE_INVALIDATION				0x2231
-#define		CACHE_INVALIDATION(x)				((x) << 0)
-#define			VC_ONLY						0
-#define			TC_ONLY						1
-#define			VC_AND_TC					2
-#define		AUTO_INVLD_EN(x)				((x) << 6)
-#define			NO_AUTO						0
-#define			ES_AUTO						1
-#define			GS_AUTO						2
-#define			ES_AND_GS_AUTO					3
 #define	VGT_ESGS_RING_SIZE				0x2232
 #define	VGT_GSVS_RING_SIZE				0x2233
-
 #define	VGT_GS_VERTEX_REUSE				0x2235
-
 #define	VGT_PRIMITIVE_TYPE				0x2256
 #define	VGT_INDEX_TYPE					0x2257
-
 #define	VGT_NUM_INDICES					0x225C
 #define	VGT_NUM_INSTANCES				0x225D
-
 #define	VGT_TF_RING_SIZE				0x2262
-
 #define	VGT_HS_OFFCHIP_PARAM				0x226C
-
 #define	VGT_TF_MEMORY_BASE				0x226E
 
-#define CC_GC_SHADER_ARRAY_CONFIG			0x226F
-#define		INACTIVE_CUS_MASK			0xFFFF0000
-#define		INACTIVE_CUS_SHIFT			16
-#define GC_USER_SHADER_ARRAY_CONFIG			0x2270
-
-#define	PA_CL_ENHANCE					0x2285
-#define		CLIP_VTX_REORDER_ENA				(1 << 0)
-#define		NUM_CLIP_SEQ(x)					((x) << 1)
-
-#define	PA_SU_LINE_STIPPLE_VALUE			0x2298
-
-#define	PA_SC_LINE_STIPPLE_STATE			0x22C4
-
-#define	PA_SC_FORCE_EOV_MAX_CNTS			0x22C9
-#define		FORCE_EOV_MAX_CLK_CNT(x)			((x) << 0)
-#define		FORCE_EOV_MAX_REZ_CNT(x)			((x) << 16)
-
-#define	PA_SC_FIFO_SIZE					0x22F3
-#define		SC_FRONTEND_PRIM_FIFO_SIZE(x)			((x) << 0)
-#define		SC_BACKEND_PRIM_FIFO_SIZE(x)			((x) << 6)
-#define		SC_HIZ_TILE_FIFO_SIZE(x)			((x) << 15)
-#define		SC_EARLYZ_TILE_FIFO_SIZE(x)			((x) << 23)
-
 #define	PA_SC_ENHANCE					0x22FC
 
-#define	SQ_CONFIG					0x2300
-
-#define	SQC_CACHES					0x2302
-
-#define SQ_POWER_THROTTLE                               0x2396
-#define		MIN_POWER(x)				((x) << 0)
-#define		MIN_POWER_MASK				(0x3fff << 0)
-#define		MIN_POWER_SHIFT				0
-#define		MAX_POWER(x)				((x) << 16)
-#define		MAX_POWER_MASK				(0x3fff << 16)
-#define		MAX_POWER_SHIFT				0
-#define SQ_POWER_THROTTLE2                              0x2397
-#define		MAX_POWER_DELTA(x)			((x) << 0)
-#define		MAX_POWER_DELTA_MASK			(0x3fff << 0)
-#define		MAX_POWER_DELTA_SHIFT			0
-#define		STI_SIZE(x)				((x) << 16)
-#define		STI_SIZE_MASK				(0x3ff << 16)
-#define		STI_SIZE_SHIFT				16
-#define		LTI_RATIO(x)				((x) << 27)
-#define		LTI_RATIO_MASK				(0xf << 27)
-#define		LTI_RATIO_SHIFT				27
-
-#define	SX_DEBUG_1					0x2418
-
-#define	SPI_STATIC_THREAD_MGMT_1			0x2438
-#define	SPI_STATIC_THREAD_MGMT_2			0x2439
-#define	SPI_STATIC_THREAD_MGMT_3			0x243A
-#define	SPI_PS_MAX_WAVE_ID				0x243B
-
-#define	SPI_CONFIG_CNTL					0x2440
-
-#define	SPI_CONFIG_CNTL_1				0x244F
-#define		VTX_DONE_DELAY(x)				((x) << 0)
-#define		INTERP_ONE_PRIM_PER_ROW				(1 << 4)
-
-#define	CGTS_TCC_DISABLE				0x2452
-#define	CGTS_USER_TCC_DISABLE				0x2453
-#define		TCC_DISABLE_MASK				0xFFFF0000
-#define		TCC_DISABLE_SHIFT				16
-#define	CGTS_SM_CTRL_REG				0x2454
-#define		OVERRIDE				(1 << 21)
-#define		LS_OVERRIDE				(1 << 22)
-
-#define	SPI_LB_CU_MASK					0x24D5
-
 #define	TA_CNTL_AUX					0x2542
 
-#define CC_RB_BACKEND_DISABLE				0x263D
-#define		BACKEND_DISABLE(x)     			((x) << 16)
-#define GB_ADDR_CONFIG  				0x263E
-#define		NUM_PIPES(x)				((x) << 0)
-#define		NUM_PIPES_MASK				0x00000007
-#define		NUM_PIPES_SHIFT				0
-#define		PIPE_INTERLEAVE_SIZE(x)			((x) << 4)
-#define		PIPE_INTERLEAVE_SIZE_MASK		0x00000070
-#define		PIPE_INTERLEAVE_SIZE_SHIFT		4
-#define		NUM_SHADER_ENGINES(x)			((x) << 12)
-#define		NUM_SHADER_ENGINES_MASK			0x00003000
-#define		NUM_SHADER_ENGINES_SHIFT		12
-#define		SHADER_ENGINE_TILE_SIZE(x)     		((x) << 16)
-#define		SHADER_ENGINE_TILE_SIZE_MASK		0x00070000
-#define		SHADER_ENGINE_TILE_SIZE_SHIFT		16
-#define		NUM_GPUS(x)     			((x) << 20)
-#define		NUM_GPUS_MASK				0x00700000
-#define		NUM_GPUS_SHIFT				20
-#define		MULTI_GPU_TILE_SIZE(x)     		((x) << 24)
-#define		MULTI_GPU_TILE_SIZE_MASK		0x03000000
-#define		MULTI_GPU_TILE_SIZE_SHIFT		24
-#define		ROW_SIZE(x)             		((x) << 28)
-#define		ROW_SIZE_MASK				0x30000000
-#define		ROW_SIZE_SHIFT				28
-
-#define	GB_TILE_MODE0					0x2644
-#       define MICRO_TILE_MODE(x)				((x) << 0)
-#              define	ADDR_SURF_DISPLAY_MICRO_TILING		0
-#              define	ADDR_SURF_THIN_MICRO_TILING		1
-#              define	ADDR_SURF_DEPTH_MICRO_TILING		2
-#       define ARRAY_MODE(x)					((x) << 2)
-#              define	ARRAY_LINEAR_GENERAL			0
-#              define	ARRAY_LINEAR_ALIGNED			1
-#              define	ARRAY_1D_TILED_THIN1			2
-#              define	ARRAY_2D_TILED_THIN1			4
-#       define PIPE_CONFIG(x)					((x) << 6)
-#              define	ADDR_SURF_P2				0
-#              define	ADDR_SURF_P4_8x16			4
-#              define	ADDR_SURF_P4_16x16			5
-#              define	ADDR_SURF_P4_16x32			6
-#              define	ADDR_SURF_P4_32x32			7
-#              define	ADDR_SURF_P8_16x16_8x16			8
-#              define	ADDR_SURF_P8_16x32_8x16			9
-#              define	ADDR_SURF_P8_32x32_8x16			10
-#              define	ADDR_SURF_P8_16x32_16x16		11
-#              define	ADDR_SURF_P8_32x32_16x16		12
-#              define	ADDR_SURF_P8_32x32_16x32		13
-#              define	ADDR_SURF_P8_32x64_32x32		14
-#       define TILE_SPLIT(x)					((x) << 11)
-#              define	ADDR_SURF_TILE_SPLIT_64B		0
-#              define	ADDR_SURF_TILE_SPLIT_128B		1
-#              define	ADDR_SURF_TILE_SPLIT_256B		2
-#              define	ADDR_SURF_TILE_SPLIT_512B		3
-#              define	ADDR_SURF_TILE_SPLIT_1KB		4
-#              define	ADDR_SURF_TILE_SPLIT_2KB		5
-#              define	ADDR_SURF_TILE_SPLIT_4KB		6
-#       define BANK_WIDTH(x)					((x) << 14)
-#              define	ADDR_SURF_BANK_WIDTH_1			0
-#              define	ADDR_SURF_BANK_WIDTH_2			1
-#              define	ADDR_SURF_BANK_WIDTH_4			2
-#              define	ADDR_SURF_BANK_WIDTH_8			3
-#       define BANK_HEIGHT(x)					((x) << 16)
-#              define	ADDR_SURF_BANK_HEIGHT_1			0
-#              define	ADDR_SURF_BANK_HEIGHT_2			1
-#              define	ADDR_SURF_BANK_HEIGHT_4			2
-#              define	ADDR_SURF_BANK_HEIGHT_8			3
-#       define MACRO_TILE_ASPECT(x)				((x) << 18)
-#              define	ADDR_SURF_MACRO_ASPECT_1		0
-#              define	ADDR_SURF_MACRO_ASPECT_2		1
-#              define	ADDR_SURF_MACRO_ASPECT_4		2
-#              define	ADDR_SURF_MACRO_ASPECT_8		3
-#       define NUM_BANKS(x)					((x) << 20)
-#              define	ADDR_SURF_2_BANK			0
-#              define	ADDR_SURF_4_BANK			1
-#              define	ADDR_SURF_8_BANK			2
-#              define	ADDR_SURF_16_BANK			3
-#define	GB_TILE_MODE1					0x2645
-#define	GB_TILE_MODE2					0x2646
-#define	GB_TILE_MODE3					0x2647
-#define	GB_TILE_MODE4					0x2648
-#define	GB_TILE_MODE5					0x2649
-#define	GB_TILE_MODE6					0x264a
-#define	GB_TILE_MODE7					0x264b
-#define	GB_TILE_MODE8					0x264c
-#define	GB_TILE_MODE9					0x264d
-#define	GB_TILE_MODE10					0x264e
-#define	GB_TILE_MODE11					0x264f
-#define	GB_TILE_MODE12					0x2650
-#define	GB_TILE_MODE13					0x2651
-#define	GB_TILE_MODE14					0x2652
-#define	GB_TILE_MODE15					0x2653
-#define	GB_TILE_MODE16					0x2654
-#define	GB_TILE_MODE17					0x2655
-#define	GB_TILE_MODE18					0x2656
-#define	GB_TILE_MODE19					0x2657
-#define	GB_TILE_MODE20					0x2658
-#define	GB_TILE_MODE21					0x2659
-#define	GB_TILE_MODE22					0x265a
-#define	GB_TILE_MODE23					0x265b
-#define	GB_TILE_MODE24					0x265c
-#define	GB_TILE_MODE25					0x265d
-#define	GB_TILE_MODE26					0x265e
-#define	GB_TILE_MODE27					0x265f
-#define	GB_TILE_MODE28					0x2660
-#define	GB_TILE_MODE29					0x2661
-#define	GB_TILE_MODE30					0x2662
-#define	GB_TILE_MODE31					0x2663
-
-#define	CB_PERFCOUNTER0_SELECT0				0x2688
-#define	CB_PERFCOUNTER0_SELECT1				0x2689
-#define	CB_PERFCOUNTER1_SELECT0				0x268A
-#define	CB_PERFCOUNTER1_SELECT1				0x268B
-#define	CB_PERFCOUNTER2_SELECT0				0x268C
-#define	CB_PERFCOUNTER2_SELECT1				0x268D
-#define	CB_PERFCOUNTER3_SELECT0				0x268E
-#define	CB_PERFCOUNTER3_SELECT1				0x268F
-
-#define	CB_CGTT_SCLK_CTRL				0x2698
-
-#define	GC_USER_RB_BACKEND_DISABLE			0x26DF
-#define		BACKEND_DISABLE_MASK			0x00FF0000
-#define		BACKEND_DISABLE_SHIFT			16
-
-#define	TCP_CHAN_STEER_LO				0x2B03
-#define	TCP_CHAN_STEER_HI				0x2B94
-
-#define	CP_RB0_BASE					0x3040
-#define	CP_RB0_CNTL					0x3041
-#define		RB_BUFSZ(x)					((x) << 0)
-#define		RB_BLKSZ(x)					((x) << 8)
-#define		BUF_SWAP_32BIT					(2 << 16)
-#define		RB_NO_UPDATE					(1 << 27)
-#define		RB_RPTR_WR_ENA					(1 << 31)
-
-#define	CP_RB0_RPTR_ADDR				0x3043
-#define	CP_RB0_RPTR_ADDR_HI				0x3044
-#define	CP_RB0_WPTR					0x3045
-
-#define	CP_PFP_UCODE_ADDR				0x3054
-#define	CP_PFP_UCODE_DATA				0x3055
-#define	CP_ME_RAM_RADDR					0x3056
-#define	CP_ME_RAM_WADDR					0x3057
-#define	CP_ME_RAM_DATA					0x3058
-
-#define	CP_CE_UCODE_ADDR				0x305A
-#define	CP_CE_UCODE_DATA				0x305B
-
-#define	CP_RB1_BASE					0x3060
-#define	CP_RB1_CNTL					0x3061
-#define	CP_RB1_RPTR_ADDR				0x3062
-#define	CP_RB1_RPTR_ADDR_HI				0x3063
-#define	CP_RB1_WPTR					0x3064
-#define	CP_RB2_BASE					0x3065
-#define	CP_RB2_CNTL					0x3066
-#define	CP_RB2_RPTR_ADDR				0x3067
-#define	CP_RB2_RPTR_ADDR_HI				0x3068
-#define	CP_RB2_WPTR					0x3069
-#define CP_INT_CNTL_RING0                               0x306A
-#define CP_INT_CNTL_RING1                               0x306B
-#define CP_INT_CNTL_RING2                               0x306C
-#       define CNTX_BUSY_INT_ENABLE                     (1 << 19)
-#       define CNTX_EMPTY_INT_ENABLE                    (1 << 20)
-#       define WAIT_MEM_SEM_INT_ENABLE                  (1 << 21)
-#       define TIME_STAMP_INT_ENABLE                    (1 << 26)
-#       define CP_RINGID2_INT_ENABLE                    (1 << 29)
-#       define CP_RINGID1_INT_ENABLE                    (1 << 30)
-#       define CP_RINGID0_INT_ENABLE                    (1 << 31)
-#define CP_INT_STATUS_RING0                             0x306D
-#define CP_INT_STATUS_RING1                             0x306E
-#define CP_INT_STATUS_RING2                             0x306F
-#       define WAIT_MEM_SEM_INT_STAT                    (1 << 21)
-#       define TIME_STAMP_INT_STAT                      (1 << 26)
-#       define CP_RINGID2_INT_STAT                      (1 << 29)
-#       define CP_RINGID1_INT_STAT                      (1 << 30)
-#       define CP_RINGID0_INT_STAT                      (1 << 31)
-
-#define	CP_MEM_SLP_CNTL					0x3079
-#       define CP_MEM_LS_EN                             (1 << 0)
-
-#define	CP_DEBUG					0x307F
-
-#define RLC_CNTL                                          0x30C0
-#       define RLC_ENABLE                                 (1 << 0)
-#define RLC_RL_BASE                                       0x30C1
-#define RLC_RL_SIZE                                       0x30C2
-#define RLC_LB_CNTL                                       0x30C3
-#       define LOAD_BALANCE_ENABLE                        (1 << 0)
-#define RLC_SAVE_AND_RESTORE_BASE                         0x30C4
-#define RLC_LB_CNTR_MAX                                   0x30C5
-#define RLC_LB_CNTR_INIT                                  0x30C6
-
-#define RLC_CLEAR_STATE_RESTORE_BASE                      0x30C8
-
-#define RLC_UCODE_ADDR                                    0x30CB
-#define RLC_UCODE_DATA                                    0x30CC
-
-#define RLC_GPU_CLOCK_COUNT_LSB                           0x30CE
-#define RLC_GPU_CLOCK_COUNT_MSB                           0x30CF
-#define RLC_CAPTURE_GPU_CLOCK_COUNT                       0x30D0
-#define RLC_MC_CNTL                                       0x30D1
-#define RLC_UCODE_CNTL                                    0x30D2
-#define RLC_STAT                                          0x30D3
-#       define RLC_BUSY_STATUS                            (1 << 0)
-#       define GFX_POWER_STATUS                           (1 << 1)
-#       define GFX_CLOCK_STATUS                           (1 << 2)
-#       define GFX_LS_STATUS                              (1 << 3)
-
-#define	RLC_PG_CNTL					0x30D7
-#	define GFX_PG_ENABLE				(1 << 0)
-#	define GFX_PG_SRC				(1 << 1)
-
-#define	RLC_CGTT_MGCG_OVERRIDE				0x3100
-#define	RLC_CGCG_CGLS_CTRL				0x3101
-#	define CGCG_EN					(1 << 0)
-#	define CGLS_EN					(1 << 1)
-
-#define	RLC_TTOP_D					0x3105
-#	define RLC_PUD(x)				((x) << 0)
-#	define RLC_PUD_MASK				(0xff << 0)
-#	define RLC_PDD(x)				((x) << 8)
-#	define RLC_PDD_MASK				(0xff << 8)
-#	define RLC_TTPD(x)				((x) << 16)
-#	define RLC_TTPD_MASK				(0xff << 16)
-#	define RLC_MSD(x)				((x) << 24)
-#	define RLC_MSD_MASK				(0xff << 24)
-
-#define RLC_LB_INIT_CU_MASK                               0x3107
-
-#define	RLC_PG_AO_CU_MASK				0x310B
-#define	RLC_MAX_PG_CU					0x310C
-#	define MAX_PU_CU(x)				((x) << 0)
-#	define MAX_PU_CU_MASK				(0xff << 0)
-#define	RLC_AUTO_PG_CTRL				0x310C
-#	define AUTO_PG_EN				(1 << 0)
-#	define GRBM_REG_SGIT(x)				((x) << 3)
-#	define GRBM_REG_SGIT_MASK			(0xffff << 3)
-#	define PG_AFTER_GRBM_REG_ST(x)			((x) << 19)
-#	define PG_AFTER_GRBM_REG_ST_MASK		(0x1fff << 19)
-
-#define RLC_SERDES_WR_MASTER_MASK_0                       0x3115
-#define RLC_SERDES_WR_MASTER_MASK_1                       0x3116
-#define RLC_SERDES_WR_CTRL                                0x3117
-
-#define RLC_SERDES_MASTER_BUSY_0                          0x3119
-#define RLC_SERDES_MASTER_BUSY_1                          0x311A
-
-#define RLC_GCPM_GENERAL_3                                0x311E
-
-#define	DB_RENDER_CONTROL				0xA000
-
-#define DB_DEPTH_INFO                                   0xA00F
-
-#define PA_SC_RASTER_CONFIG                             0xA0D4
-#	define RB_MAP_PKR0(x)				((x) << 0)
-#	define RB_MAP_PKR0_MASK				(0x3 << 0)
-#	define RB_MAP_PKR1(x)				((x) << 2)
-#	define RB_MAP_PKR1_MASK				(0x3 << 2)
-#       define RASTER_CONFIG_RB_MAP_0                   0
-#       define RASTER_CONFIG_RB_MAP_1                   1
-#       define RASTER_CONFIG_RB_MAP_2                   2
-#       define RASTER_CONFIG_RB_MAP_3                   3
+// #define PA_SC_RASTER_CONFIG                             0xA0D4
 #	define RB_XSEL2(x)				((x) << 4)
 #	define RB_XSEL2_MASK				(0x3 << 4)
 #	define RB_XSEL					(1 << 6)
 #	define RB_YSEL					(1 << 7)
 #	define PKR_MAP(x)				((x) << 8)
-#	define PKR_MAP_MASK				(0x3 << 8)
-#       define RASTER_CONFIG_PKR_MAP_0			0
-#       define RASTER_CONFIG_PKR_MAP_1			1
-#       define RASTER_CONFIG_PKR_MAP_2			2
-#       define RASTER_CONFIG_PKR_MAP_3			3
 #	define PKR_XSEL(x)				((x) << 10)
 #	define PKR_XSEL_MASK				(0x3 << 10)
 #	define PKR_YSEL(x)				((x) << 12)
@@ -1426,221 +306,19 @@
 #	define SC_YSEL(x)				((x) << 20)
 #	define SC_YSEL_MASK				(0x3 << 20)
 #	define SE_MAP(x)				((x) << 24)
-#	define SE_MAP_MASK				(0x3 << 24)
-#       define RASTER_CONFIG_SE_MAP_0			0
-#       define RASTER_CONFIG_SE_MAP_1			1
-#       define RASTER_CONFIG_SE_MAP_2			2
-#       define RASTER_CONFIG_SE_MAP_3			3
 #	define SE_XSEL(x)				((x) << 26)
 #	define SE_XSEL_MASK				(0x3 << 26)
 #	define SE_YSEL(x)				((x) << 28)
 #	define SE_YSEL_MASK				(0x3 << 28)
 
-
-#define VGT_EVENT_INITIATOR                             0xA2A4
-#       define SAMPLE_STREAMOUTSTATS1                   (1 << 0)
-#       define SAMPLE_STREAMOUTSTATS2                   (2 << 0)
-#       define SAMPLE_STREAMOUTSTATS3                   (3 << 0)
-#       define CACHE_FLUSH_TS                           (4 << 0)
-#       define CACHE_FLUSH                              (6 << 0)
-#       define CS_PARTIAL_FLUSH                         (7 << 0)
-#       define VGT_STREAMOUT_RESET                      (10 << 0)
-#       define END_OF_PIPE_INCR_DE                      (11 << 0)
-#       define END_OF_PIPE_IB_END                       (12 << 0)
-#       define RST_PIX_CNT                              (13 << 0)
-#       define VS_PARTIAL_FLUSH                         (15 << 0)
-#       define PS_PARTIAL_FLUSH                         (16 << 0)
-#       define CACHE_FLUSH_AND_INV_TS_EVENT             (20 << 0)
-#       define ZPASS_DONE                               (21 << 0)
-#       define CACHE_FLUSH_AND_INV_EVENT                (22 << 0)
-#       define PERFCOUNTER_START                        (23 << 0)
-#       define PERFCOUNTER_STOP                         (24 << 0)
-#       define PIPELINESTAT_START                       (25 << 0)
-#       define PIPELINESTAT_STOP                        (26 << 0)
-#       define PERFCOUNTER_SAMPLE                       (27 << 0)
-#       define SAMPLE_PIPELINESTAT                      (30 << 0)
-#       define SAMPLE_STREAMOUTSTATS                    (32 << 0)
-#       define RESET_VTX_CNT                            (33 << 0)
-#       define VGT_FLUSH                                (36 << 0)
-#       define BOTTOM_OF_PIPE_TS                        (40 << 0)
-#       define DB_CACHE_FLUSH_AND_INV                   (42 << 0)
-#       define FLUSH_AND_INV_DB_DATA_TS                 (43 << 0)
-#       define FLUSH_AND_INV_DB_META                    (44 << 0)
-#       define FLUSH_AND_INV_CB_DATA_TS                 (45 << 0)
-#       define FLUSH_AND_INV_CB_META                    (46 << 0)
-#       define CS_DONE                                  (47 << 0)
-#       define PS_DONE                                  (48 << 0)
-#       define FLUSH_AND_INV_CB_PIXEL_DATA              (49 << 0)
-#       define THREAD_TRACE_START                       (51 << 0)
-#       define THREAD_TRACE_STOP                        (52 << 0)
-#       define THREAD_TRACE_FLUSH                       (54 << 0)
-#       define THREAD_TRACE_FINISH                      (55 << 0)
-
-/* PIF PHY0 registers idx/data 0x8/0xc */
-#define PB0_PIF_CNTL                                      0x10
-#       define LS2_EXIT_TIME(x)                           ((x) << 17)
-#       define LS2_EXIT_TIME_MASK                         (0x7 << 17)
-#       define LS2_EXIT_TIME_SHIFT                        17
-#define PB0_PIF_PAIRING                                   0x11
-#       define MULTI_PIF                                  (1 << 25)
-#define PB0_PIF_PWRDOWN_0                                 0x12
-#       define PLL_POWER_STATE_IN_TXS2_0(x)               ((x) << 7)
-#       define PLL_POWER_STATE_IN_TXS2_0_MASK             (0x7 << 7)
-#       define PLL_POWER_STATE_IN_TXS2_0_SHIFT            7
-#       define PLL_POWER_STATE_IN_OFF_0(x)                ((x) << 10)
-#       define PLL_POWER_STATE_IN_OFF_0_MASK              (0x7 << 10)
-#       define PLL_POWER_STATE_IN_OFF_0_SHIFT             10
-#       define PLL_RAMP_UP_TIME_0(x)                      ((x) << 24)
-#       define PLL_RAMP_UP_TIME_0_MASK                    (0x7 << 24)
-#       define PLL_RAMP_UP_TIME_0_SHIFT                   24
-#define PB0_PIF_PWRDOWN_1                                 0x13
-#       define PLL_POWER_STATE_IN_TXS2_1(x)               ((x) << 7)
-#       define PLL_POWER_STATE_IN_TXS2_1_MASK             (0x7 << 7)
-#       define PLL_POWER_STATE_IN_TXS2_1_SHIFT            7
-#       define PLL_POWER_STATE_IN_OFF_1(x)                ((x) << 10)
-#       define PLL_POWER_STATE_IN_OFF_1_MASK              (0x7 << 10)
-#       define PLL_POWER_STATE_IN_OFF_1_SHIFT             10
-#       define PLL_RAMP_UP_TIME_1(x)                      ((x) << 24)
-#       define PLL_RAMP_UP_TIME_1_MASK                    (0x7 << 24)
-#       define PLL_RAMP_UP_TIME_1_SHIFT                   24
-
-#define PB0_PIF_PWRDOWN_2                                 0x17
-#       define PLL_POWER_STATE_IN_TXS2_2(x)               ((x) << 7)
-#       define PLL_POWER_STATE_IN_TXS2_2_MASK             (0x7 << 7)
-#       define PLL_POWER_STATE_IN_TXS2_2_SHIFT            7
-#       define PLL_POWER_STATE_IN_OFF_2(x)                ((x) << 10)
-#       define PLL_POWER_STATE_IN_OFF_2_MASK              (0x7 << 10)
-#       define PLL_POWER_STATE_IN_OFF_2_SHIFT             10
-#       define PLL_RAMP_UP_TIME_2(x)                      ((x) << 24)
-#       define PLL_RAMP_UP_TIME_2_MASK                    (0x7 << 24)
-#       define PLL_RAMP_UP_TIME_2_SHIFT                   24
-#define PB0_PIF_PWRDOWN_3                                 0x18
-#       define PLL_POWER_STATE_IN_TXS2_3(x)               ((x) << 7)
-#       define PLL_POWER_STATE_IN_TXS2_3_MASK             (0x7 << 7)
-#       define PLL_POWER_STATE_IN_TXS2_3_SHIFT            7
-#       define PLL_POWER_STATE_IN_OFF_3(x)                ((x) << 10)
-#       define PLL_POWER_STATE_IN_OFF_3_MASK              (0x7 << 10)
-#       define PLL_POWER_STATE_IN_OFF_3_SHIFT             10
-#       define PLL_RAMP_UP_TIME_3(x)                      ((x) << 24)
-#       define PLL_RAMP_UP_TIME_3_MASK                    (0x7 << 24)
-#       define PLL_RAMP_UP_TIME_3_SHIFT                   24
-/* PIF PHY1 registers idx/data 0x10/0x14 */
-#define PB1_PIF_CNTL                                      0x10
-#define PB1_PIF_PAIRING                                   0x11
-#define PB1_PIF_PWRDOWN_0                                 0x12
-#define PB1_PIF_PWRDOWN_1                                 0x13
-
-#define PB1_PIF_PWRDOWN_2                                 0x17
-#define PB1_PIF_PWRDOWN_3                                 0x18
-/* PCIE registers idx/data 0x30/0x34 */
-#define PCIE_CNTL2                                        0x1c /* PCIE */
-#       define SLV_MEM_LS_EN                              (1 << 16)
-#       define SLV_MEM_AGGRESSIVE_LS_EN                   (1 << 17)
-#       define MST_MEM_LS_EN                              (1 << 18)
-#       define REPLAY_MEM_LS_EN                           (1 << 19)
-#define PCIE_LC_STATUS1                                   0x28 /* PCIE */
-#       define LC_REVERSE_RCVR                            (1 << 0)
-#       define LC_REVERSE_XMIT                            (1 << 1)
-#       define LC_OPERATING_LINK_WIDTH_MASK               (0x7 << 2)
-#       define LC_OPERATING_LINK_WIDTH_SHIFT              2
-#       define LC_DETECTED_LINK_WIDTH_MASK                (0x7 << 5)
-#       define LC_DETECTED_LINK_WIDTH_SHIFT               5
-
-#define PCIE_P_CNTL                                       0x40 /* PCIE */
-#       define P_IGNORE_EDB_ERR                           (1 << 6)
-
 /* PCIE PORT registers idx/data 0x38/0x3c */
-#define PCIE_LC_CNTL                                      0xa0
-#       define LC_L0S_INACTIVITY(x)                       ((x) << 8)
-#       define LC_L0S_INACTIVITY_MASK                     (0xf << 8)
-#       define LC_L0S_INACTIVITY_SHIFT                    8
-#       define LC_L1_INACTIVITY(x)                        ((x) << 12)
-#       define LC_L1_INACTIVITY_MASK                      (0xf << 12)
-#       define LC_L1_INACTIVITY_SHIFT                     12
-#       define LC_PMI_TO_L1_DIS                           (1 << 16)
-#       define LC_ASPM_TO_L1_DIS                          (1 << 24)
-#define PCIE_LC_LINK_WIDTH_CNTL                           0xa2 /* PCIE_P */
-#       define LC_LINK_WIDTH_SHIFT                        0
-#       define LC_LINK_WIDTH_MASK                         0x7
+// #define PCIE_LC_LINK_WIDTH_CNTL                           0xa2 /* PCIE_P */
 #       define LC_LINK_WIDTH_X0                           0
 #       define LC_LINK_WIDTH_X1                           1
 #       define LC_LINK_WIDTH_X2                           2
 #       define LC_LINK_WIDTH_X4                           3
 #       define LC_LINK_WIDTH_X8                           4
 #       define LC_LINK_WIDTH_X16                          6
-#       define LC_LINK_WIDTH_RD_SHIFT                     4
-#       define LC_LINK_WIDTH_RD_MASK                      0x70
-#       define LC_RECONFIG_ARC_MISSING_ESCAPE             (1 << 7)
-#       define LC_RECONFIG_NOW                            (1 << 8)
-#       define LC_RENEGOTIATION_SUPPORT                   (1 << 9)
-#       define LC_RENEGOTIATE_EN                          (1 << 10)
-#       define LC_SHORT_RECONFIG_EN                       (1 << 11)
-#       define LC_UPCONFIGURE_SUPPORT                     (1 << 12)
-#       define LC_UPCONFIGURE_DIS                         (1 << 13)
-#       define LC_DYN_LANES_PWR_STATE(x)                  ((x) << 21)
-#       define LC_DYN_LANES_PWR_STATE_MASK                (0x3 << 21)
-#       define LC_DYN_LANES_PWR_STATE_SHIFT               21
-#define PCIE_LC_N_FTS_CNTL                                0xa3 /* PCIE_P */
-#       define LC_XMIT_N_FTS(x)                           ((x) << 0)
-#       define LC_XMIT_N_FTS_MASK                         (0xff << 0)
-#       define LC_XMIT_N_FTS_SHIFT                        0
-#       define LC_XMIT_N_FTS_OVERRIDE_EN                  (1 << 8)
-#       define LC_N_FTS_MASK                              (0xff << 24)
-#define PCIE_LC_SPEED_CNTL                                0xa4 /* PCIE_P */
-#       define LC_GEN2_EN_STRAP                           (1 << 0)
-#       define LC_GEN3_EN_STRAP                           (1 << 1)
-#       define LC_TARGET_LINK_SPEED_OVERRIDE_EN           (1 << 2)
-#       define LC_TARGET_LINK_SPEED_OVERRIDE_MASK         (0x3 << 3)
-#       define LC_TARGET_LINK_SPEED_OVERRIDE_SHIFT        3
-#       define LC_FORCE_EN_SW_SPEED_CHANGE                (1 << 5)
-#       define LC_FORCE_DIS_SW_SPEED_CHANGE               (1 << 6)
-#       define LC_FORCE_EN_HW_SPEED_CHANGE                (1 << 7)
-#       define LC_FORCE_DIS_HW_SPEED_CHANGE               (1 << 8)
-#       define LC_INITIATE_LINK_SPEED_CHANGE              (1 << 9)
-#       define LC_SPEED_CHANGE_ATTEMPTS_ALLOWED_MASK      (0x3 << 10)
-#       define LC_SPEED_CHANGE_ATTEMPTS_ALLOWED_SHIFT     10
-#       define LC_CURRENT_DATA_RATE_MASK                  (0x3 << 13) /* 0/1/2 = gen1/2/3 */
-#       define LC_CURRENT_DATA_RATE_SHIFT                 13
-#       define LC_CLR_FAILED_SPD_CHANGE_CNT               (1 << 16)
-#       define LC_OTHER_SIDE_EVER_SENT_GEN2               (1 << 18)
-#       define LC_OTHER_SIDE_SUPPORTS_GEN2                (1 << 19)
-#       define LC_OTHER_SIDE_EVER_SENT_GEN3               (1 << 20)
-#       define LC_OTHER_SIDE_SUPPORTS_GEN3                (1 << 21)
-
-#define PCIE_LC_CNTL2                                     0xb1
-#       define LC_ALLOW_PDWN_IN_L1                        (1 << 17)
-#       define LC_ALLOW_PDWN_IN_L23                       (1 << 18)
-
-#define PCIE_LC_CNTL3                                     0xb5 /* PCIE_P */
-#       define LC_GO_TO_RECOVERY                          (1 << 30)
-#define PCIE_LC_CNTL4                                     0xb6 /* PCIE_P */
-#       define LC_REDO_EQ                                 (1 << 5)
-#       define LC_SET_QUIESCE                             (1 << 13)
-
-/*
- * UVD
- */
-#define UVD_UDEC_ADDR_CONFIG				0x3bd3
-#define UVD_UDEC_DB_ADDR_CONFIG				0x3bd4
-#define UVD_UDEC_DBW_ADDR_CONFIG			0x3bd5
-#define UVD_RBC_RB_RPTR					0x3da4
-#define UVD_RBC_RB_WPTR					0x3da5
-#define UVD_STATUS					0x3daf
-
-#define	UVD_CGC_CTRL					0x3dc2
-#	define DCM					(1 << 0)
-#	define CG_DT(x)					((x) << 2)
-#	define CG_DT_MASK				(0xf << 2)
-#	define CLK_OD(x)				((x) << 6)
-#	define CLK_OD_MASK				(0x1f << 6)
-
- /* UVD CTX indirect */
-#define	UVD_CGC_MEM_CTRL				0xC0
-#define	UVD_CGC_CTRL2					0xC1
-#	define DYN_OR_EN				(1 << 0)
-#	define DYN_RR_EN				(1 << 1)
-#	define G_DIV_ID(x)				((x) << 2)
-#	define G_DIV_ID_MASK				(0x7 << 2)
 
 /*
  * PM4
@@ -1874,45 +552,7 @@
 /* ASYNC DMA - first instance at 0xd000, second at 0xd800 */
 #define DMA0_REGISTER_OFFSET                              0x0 /* not a register */
 #define DMA1_REGISTER_OFFSET                              0x200 /* not a register */
-
-#define DMA_RB_CNTL                                       0x3400
-#       define DMA_RB_ENABLE                              (1 << 0)
-#       define DMA_RB_SIZE(x)                             ((x) << 1) /* log2 */
-#       define DMA_RB_SWAP_ENABLE                         (1 << 9) /* 8IN32 */
-#       define DMA_RPTR_WRITEBACK_ENABLE                  (1 << 12)
-#       define DMA_RPTR_WRITEBACK_SWAP_ENABLE             (1 << 13)  /* 8IN32 */
-#       define DMA_RPTR_WRITEBACK_TIMER(x)                ((x) << 16) /* log2 */
-#define DMA_RB_BASE                                       0x3401
-#define DMA_RB_RPTR                                       0x3402
-#define DMA_RB_WPTR                                       0x3403
-
-#define DMA_RB_RPTR_ADDR_HI                               0x3407
-#define DMA_RB_RPTR_ADDR_LO                               0x3408
-
-#define DMA_IB_CNTL                                       0x3409
-#       define DMA_IB_ENABLE                              (1 << 0)
-#       define DMA_IB_SWAP_ENABLE                         (1 << 4)
-#       define CMD_VMID_FORCE                             (1 << 31)
-#define DMA_IB_RPTR                                       0x340a
-#define DMA_CNTL                                          0x340b
-#       define TRAP_ENABLE                                (1 << 0)
-#       define SEM_INCOMPLETE_INT_ENABLE                  (1 << 1)
-#       define SEM_WAIT_INT_ENABLE                        (1 << 2)
-#       define DATA_SWAP_ENABLE                           (1 << 3)
-#       define FENCE_SWAP_ENABLE                          (1 << 4)
-#       define CTXEMPTY_INT_ENABLE                        (1 << 28)
-#define DMA_STATUS_REG                                    0x340d
-#       define DMA_IDLE                                   (1 << 0)
-#define DMA_TILING_CONFIG  				  0x342e
-
-#define	DMA_POWER_CNTL					0x342f
-#       define MEM_POWER_OVERRIDE                       (1 << 8)
-#define	DMA_CLK_CTRL					0x3430
-
-#define	DMA_PG						0x3435
-#	define PG_CNTL_ENABLE				(1 << 0)
-#define	DMA_PGFSM_CONFIG				0x3436
-#define	DMA_PGFSM_WRITE					0x3437
+#define SDMA_MAX_INSTANCE 2
 
 #define DMA_PACKET(cmd, b, t, s, n)	((((cmd) & 0xF) << 28) |	\
 					 (((b) & 0x1) << 26) |		\
@@ -1941,45 +581,7 @@
 #define	DMA_PACKET_POLL_REG_MEM				  0xe
 #define	DMA_PACKET_NOP					  0xf
 
-#define VCE_STATUS					0x20004
-#define VCE_VCPU_CNTL					0x20014
-#define		VCE_CLK_EN				(1 << 0)
-#define VCE_VCPU_CACHE_OFFSET0				0x20024
-#define VCE_VCPU_CACHE_SIZE0				0x20028
-#define VCE_VCPU_CACHE_OFFSET1				0x2002c
-#define VCE_VCPU_CACHE_SIZE1				0x20030
-#define VCE_VCPU_CACHE_OFFSET2				0x20034
-#define VCE_VCPU_CACHE_SIZE2				0x20038
-#define VCE_SOFT_RESET					0x20120
-#define 	VCE_ECPU_SOFT_RESET			(1 << 0)
-#define 	VCE_FME_SOFT_RESET			(1 << 2)
-#define VCE_RB_BASE_LO2					0x2016c
-#define VCE_RB_BASE_HI2					0x20170
-#define VCE_RB_SIZE2					0x20174
-#define VCE_RB_RPTR2					0x20178
-#define VCE_RB_WPTR2					0x2017c
-#define VCE_RB_BASE_LO					0x20180
-#define VCE_RB_BASE_HI					0x20184
-#define VCE_RB_SIZE					0x20188
-#define VCE_RB_RPTR					0x2018c
-#define VCE_RB_WPTR					0x20190
-#define VCE_CLOCK_GATING_A				0x202f8
-#define VCE_CLOCK_GATING_B				0x202fc
-#define VCE_UENC_CLOCK_GATING				0x205bc
-#define VCE_UENC_REG_CLOCK_GATING			0x205c0
-#define VCE_FW_REG_STATUS				0x20e10
-#	define VCE_FW_REG_STATUS_BUSY			(1 << 0)
-#	define VCE_FW_REG_STATUS_PASS			(1 << 3)
-#	define VCE_FW_REG_STATUS_DONE			(1 << 11)
-#define VCE_LMI_FW_START_KEYSEL				0x20e18
-#define VCE_LMI_FW_PERIODIC_CTRL			0x20e20
-#define VCE_LMI_CTRL2					0x20e74
-#define VCE_LMI_CTRL					0x20e98
-#define VCE_LMI_VM_CTRL					0x20ea0
-#define VCE_LMI_SWAP_CNTL				0x20eb4
-#define VCE_LMI_SWAP_CNTL1				0x20eb8
-#define VCE_LMI_CACHE_CTRL				0x20ef4
-
+/* VCE */
 #define VCE_CMD_NO_OP					0x00000000
 #define VCE_CMD_END					0x00000001
 #define VCE_CMD_IB					0x00000002
@@ -1988,434 +590,146 @@
 #define VCE_CMD_IB_AUTO					0x00000005
 #define VCE_CMD_SEMAPHORE				0x00000006
 
-
 //#dce stupp
 /* display controller offsets used for crtc/cur/lut/grph/viewport/etc. */
-#define SI_CRTC0_REGISTER_OFFSET                0 //(0x6df0 - 0x6df0)/4
-#define SI_CRTC1_REGISTER_OFFSET                0x300 //(0x79f0 - 0x6df0)/4
-#define SI_CRTC2_REGISTER_OFFSET                0x2600 //(0x105f0 - 0x6df0)/4
-#define SI_CRTC3_REGISTER_OFFSET                0x2900 //(0x111f0 - 0x6df0)/4
-#define SI_CRTC4_REGISTER_OFFSET                0x2c00 //(0x11df0 - 0x6df0)/4
-#define SI_CRTC5_REGISTER_OFFSET                0x2f00 //(0x129f0 - 0x6df0)/4
+#define CRTC0_REGISTER_OFFSET                 (0x1b7c - 0x1b7c) //(0x6df0 - 0x6df0)/4
+#define CRTC1_REGISTER_OFFSET                 (0x1e7c - 0x1b7c) //(0x79f0 - 0x6df0)/4
+#define CRTC2_REGISTER_OFFSET                 (0x417c - 0x1b7c) //(0x105f0 - 0x6df0)/4
+#define CRTC3_REGISTER_OFFSET                 (0x447c - 0x1b7c) //(0x111f0 - 0x6df0)/4
+#define CRTC4_REGISTER_OFFSET                 (0x477c - 0x1b7c) //(0x11df0 - 0x6df0)/4
+#define CRTC5_REGISTER_OFFSET                 (0x4a7c - 0x1b7c) //(0x129f0 - 0x6df0)/4
+
+/* hpd instance offsets */
+#define HPD0_REGISTER_OFFSET                 (0x1807 - 0x1807)
+#define HPD1_REGISTER_OFFSET                 (0x180a - 0x1807)
+#define HPD2_REGISTER_OFFSET                 (0x180d - 0x1807)
+#define HPD3_REGISTER_OFFSET                 (0x1810 - 0x1807)
+#define HPD4_REGISTER_OFFSET                 (0x1813 - 0x1807)
+#define HPD5_REGISTER_OFFSET                 (0x1816 - 0x1807)
+
+/* audio endpt instance offsets */
+#define AUD0_REGISTER_OFFSET                 (0x1780 - 0x1780)
+#define AUD1_REGISTER_OFFSET                 (0x1786 - 0x1780)
+#define AUD2_REGISTER_OFFSET                 (0x178c - 0x1780)
+#define AUD3_REGISTER_OFFSET                 (0x1792 - 0x1780)
+#define AUD4_REGISTER_OFFSET                 (0x1798 - 0x1780)
+#define AUD5_REGISTER_OFFSET                 (0x179d - 0x1780)
+#define AUD6_REGISTER_OFFSET                 (0x17a4 - 0x1780)
 
 #define CURSOR_WIDTH 64
 #define CURSOR_HEIGHT 64
-#define AMDGPU_MM_INDEX		        0x0000
-#define AMDGPU_MM_DATA		        0x0001
-
-#define VERDE_NUM_CRTC 6
-#define	BLACKOUT_MODE_MASK			0x00000007
-#define	VGA_RENDER_CONTROL			0xC0
-#define R_000300_VGA_RENDER_CONTROL             0xC0
-#define C_000300_VGA_VSTATUS_CNTL               0xFFFCFFFF
-#define EVERGREEN_CRTC_STATUS                   0x1BA3
-#define EVERGREEN_CRTC_V_BLANK                  (1 << 0)
-#define EVERGREEN_CRTC_STATUS_POSITION          0x1BA4
-/* CRTC blocks at 0x6df0, 0x79f0, 0x105f0, 0x111f0, 0x11df0, 0x129f0 */
-#define EVERGREEN_CRTC_V_BLANK_START_END                0x1b8d
-#define EVERGREEN_CRTC_CONTROL                          0x1b9c
-#define EVERGREEN_CRTC_MASTER_EN                 (1 << 0)
-#define EVERGREEN_CRTC_DISP_READ_REQUEST_DISABLE (1 << 24)
-#define EVERGREEN_CRTC_BLANK_CONTROL                    0x1b9d
-#define EVERGREEN_CRTC_BLANK_DATA_EN             (1 << 8)
-#define EVERGREEN_CRTC_V_BLANK                   (1 << 0)
-#define EVERGREEN_CRTC_STATUS_HV_COUNT                  0x1ba8
-#define EVERGREEN_CRTC_UPDATE_LOCK                      0x1bb5
-#define EVERGREEN_MASTER_UPDATE_LOCK                    0x1bbd
-#define EVERGREEN_MASTER_UPDATE_MODE                    0x1bbe
-#define EVERGREEN_GRPH_UPDATE_LOCK               (1 << 16)
-#define EVERGREEN_GRPH_PRIMARY_SURFACE_ADDRESS_HIGH     0x1a07
-#define EVERGREEN_GRPH_SECONDARY_SURFACE_ADDRESS_HIGH   0x1a08
-#define EVERGREEN_GRPH_PRIMARY_SURFACE_ADDRESS          0x1a04
-#define EVERGREEN_GRPH_SECONDARY_SURFACE_ADDRESS        0x1a05
-#define EVERGREEN_GRPH_UPDATE                           0x1a11
-#define EVERGREEN_VGA_MEMORY_BASE_ADDRESS               0xc4
-#define EVERGREEN_VGA_MEMORY_BASE_ADDRESS_HIGH          0xc9
-#define EVERGREEN_GRPH_SURFACE_UPDATE_PENDING    (1 << 2)
-
-#define EVERGREEN_DATA_FORMAT                           0x1ac0
-#       define EVERGREEN_INTERLEAVE_EN                  (1 << 0)
-
-#define MC_SHARED_CHMAP__NOOFCHAN_MASK 0xf000
-#define MC_SHARED_CHMAP__NOOFCHAN__SHIFT 0xc
-
-#define R600_D1GRPH_ARRAY_MODE_LINEAR_GENERAL            (0 << 20)
-#define R600_D1GRPH_ARRAY_MODE_LINEAR_ALIGNED            (1 << 20)
-#define R600_D1GRPH_ARRAY_MODE_1D_TILED_THIN1            (2 << 20)
-#define R600_D1GRPH_ARRAY_MODE_2D_TILED_THIN1            (4 << 20)
-
-#define R700_D1GRPH_PRIMARY_SURFACE_ADDRESS_HIGH                0x1a45
-#define R700_D2GRPH_PRIMARY_SURFACE_ADDRESS_HIGH                0x1845
-
-#define R700_D2GRPH_SECONDARY_SURFACE_ADDRESS_HIGH              0x1847
-#define R700_D1GRPH_SECONDARY_SURFACE_ADDRESS_HIGH              0x1a47
-
-#define DISP_INTERRUPT_STATUS__LB_D1_VBLANK_INTERRUPT_MASK 0x8
-#define DISP_INTERRUPT_STATUS_CONTINUE__LB_D2_VBLANK_INTERRUPT_MASK 0x8
-#define DISP_INTERRUPT_STATUS_CONTINUE2__LB_D3_VBLANK_INTERRUPT_MASK 0x8
-#define DISP_INTERRUPT_STATUS_CONTINUE3__LB_D4_VBLANK_INTERRUPT_MASK 0x8
-#define DISP_INTERRUPT_STATUS_CONTINUE4__LB_D5_VBLANK_INTERRUPT_MASK 0x8
-#define DISP_INTERRUPT_STATUS_CONTINUE5__LB_D6_VBLANK_INTERRUPT_MASK 0x8
-
-#define DISP_INTERRUPT_STATUS__LB_D1_VLINE_INTERRUPT_MASK 0x4
-#define DISP_INTERRUPT_STATUS_CONTINUE__LB_D2_VLINE_INTERRUPT_MASK 0x4
-#define DISP_INTERRUPT_STATUS_CONTINUE2__LB_D3_VLINE_INTERRUPT_MASK 0x4
-#define DISP_INTERRUPT_STATUS_CONTINUE3__LB_D4_VLINE_INTERRUPT_MASK 0x4
-#define DISP_INTERRUPT_STATUS_CONTINUE4__LB_D5_VLINE_INTERRUPT_MASK 0x4
-#define DISP_INTERRUPT_STATUS_CONTINUE5__LB_D6_VLINE_INTERRUPT_MASK 0x4
-
-#define DISP_INTERRUPT_STATUS__DC_HPD1_INTERRUPT_MASK 0x20000
-#define DISP_INTERRUPT_STATUS_CONTINUE__DC_HPD2_INTERRUPT_MASK 0x20000
-#define DISP_INTERRUPT_STATUS_CONTINUE2__DC_HPD3_INTERRUPT_MASK 0x20000
-#define DISP_INTERRUPT_STATUS_CONTINUE3__DC_HPD4_INTERRUPT_MASK 0x20000
-#define DISP_INTERRUPT_STATUS_CONTINUE4__DC_HPD5_INTERRUPT_MASK 0x20000
-#define DISP_INTERRUPT_STATUS_CONTINUE5__DC_HPD6_INTERRUPT_MASK 0x20000
-
-#define GRPH_INTERRUPT_STATUS__GRPH_PFLIP_INT_OCCURRED_MASK 0x1
-#define GRPH_INTERRUPT_STATUS__GRPH_PFLIP_INT_CLEAR_MASK 0x100
-
-#define DC_HPD1_INT_CONTROL__DC_HPD1_INT_ACK_MASK 0x1
-
-#define R600_D1GRPH_SWAP_CONTROL                               0x1843
-#define R600_D1GRPH_SWAP_ENDIAN_NONE                    (0 << 0)
-#define R600_D1GRPH_SWAP_ENDIAN_16BIT                   (1 << 0)
-#define R600_D1GRPH_SWAP_ENDIAN_32BIT                   (2 << 0)
-#define R600_D1GRPH_SWAP_ENDIAN_64BIT                   (3 << 0)
-
-#define AVIVO_D1VGA_CONTROL					0x00cc
-#       define AVIVO_DVGA_CONTROL_MODE_ENABLE            (1 << 0)
-#       define AVIVO_DVGA_CONTROL_TIMING_SELECT          (1 << 8)
-#       define AVIVO_DVGA_CONTROL_SYNC_POLARITY_SELECT   (1 << 9)
-#       define AVIVO_DVGA_CONTROL_OVERSCAN_TIMING_SELECT (1 << 10)
-#       define AVIVO_DVGA_CONTROL_OVERSCAN_COLOR_EN      (1 << 16)
-#       define AVIVO_DVGA_CONTROL_ROTATE                 (1 << 24)
-#define AVIVO_D2VGA_CONTROL					0x00ce
-
-#define R600_BUS_CNTL                                           0x1508
-#       define R600_BIOS_ROM_DIS                                (1 << 1)
+
 
 #define R600_ROM_CNTL                              0x580
 #       define R600_SCK_OVERWRITE                  (1 << 1)
 #       define R600_SCK_PRESCALE_CRYSTAL_CLK_SHIFT 28
 #       define R600_SCK_PRESCALE_CRYSTAL_CLK_MASK  (0xf << 28)
 
-#define GRPH_INTERRUPT_CONTROL__GRPH_PFLIP_INT_MASK_MASK 0x1
-
-#define FMT_BIT_DEPTH_CONTROL                0x1bf2
-#define FMT_TRUNCATE_EN               (1 << 0)
-#define FMT_TRUNCATE_DEPTH            (1 << 4)
-#define FMT_SPATIAL_DITHER_EN         (1 << 8)
-#define FMT_SPATIAL_DITHER_MODE(x)    ((x) << 9)
-#define FMT_SPATIAL_DITHER_DEPTH      (1 << 12)
-#define FMT_FRAME_RANDOM_ENABLE       (1 << 13)
-#define FMT_RGB_RANDOM_ENABLE         (1 << 14)
-#define FMT_HIGHPASS_RANDOM_ENABLE    (1 << 15)
-#define FMT_TEMPORAL_DITHER_EN        (1 << 16)
-#define FMT_TEMPORAL_DITHER_DEPTH     (1 << 20)
-#define FMT_TEMPORAL_DITHER_OFFSET(x) ((x) << 21)
-#define FMT_TEMPORAL_LEVEL            (1 << 24)
-#define FMT_TEMPORAL_DITHER_RESET     (1 << 25)
-#define FMT_25FRC_SEL(x)              ((x) << 26)
-#define FMT_50FRC_SEL(x)              ((x) << 28)
-#define FMT_75FRC_SEL(x)              ((x) << 30)
-
-#define EVERGREEN_DC_LUT_CONTROL                        0x1a80
-#define EVERGREEN_DC_LUT_BLACK_OFFSET_BLUE              0x1a81
-#define EVERGREEN_DC_LUT_BLACK_OFFSET_GREEN             0x1a82
-#define EVERGREEN_DC_LUT_BLACK_OFFSET_RED               0x1a83
-#define EVERGREEN_DC_LUT_WHITE_OFFSET_BLUE              0x1a84
-#define EVERGREEN_DC_LUT_WHITE_OFFSET_GREEN             0x1a85
-#define EVERGREEN_DC_LUT_WHITE_OFFSET_RED               0x1a86
-#define EVERGREEN_DC_LUT_30_COLOR                       0x1a7c
-#define EVERGREEN_DC_LUT_RW_INDEX                       0x1a79
-#define EVERGREEN_DC_LUT_WRITE_EN_MASK                  0x1a7e
-#define EVERGREEN_DC_LUT_RW_MODE                        0x1a78
-
-#define EVERGREEN_GRPH_ENABLE                           0x1a00
-#define EVERGREEN_GRPH_CONTROL                          0x1a01
-#define EVERGREEN_GRPH_DEPTH(x)                  (((x) & 0x3) << 0)
-#define EVERGREEN_GRPH_DEPTH_8BPP                0
-#define EVERGREEN_GRPH_DEPTH_16BPP               1
-#define EVERGREEN_GRPH_DEPTH_32BPP               2
-#define EVERGREEN_GRPH_NUM_BANKS(x)              (((x) & 0x3) << 2)
-#define EVERGREEN_ADDR_SURF_2_BANK               0
-#define EVERGREEN_ADDR_SURF_4_BANK               1
-#define EVERGREEN_ADDR_SURF_8_BANK               2
-#define EVERGREEN_ADDR_SURF_16_BANK              3
-#define EVERGREEN_GRPH_Z(x)                      (((x) & 0x3) << 4)
-#define EVERGREEN_GRPH_BANK_WIDTH(x)             (((x) & 0x3) << 6)
-#define EVERGREEN_ADDR_SURF_BANK_WIDTH_1         0
-#define EVERGREEN_ADDR_SURF_BANK_WIDTH_2         1
-#define EVERGREEN_ADDR_SURF_BANK_WIDTH_4         2
-#define EVERGREEN_ADDR_SURF_BANK_WIDTH_8         3
-#define EVERGREEN_GRPH_FORMAT(x)                 (((x) & 0x7) << 8)
-
-#define EVERGREEN_GRPH_FORMAT_INDEXED            0
-#define EVERGREEN_GRPH_FORMAT_ARGB1555           0
-#define EVERGREEN_GRPH_FORMAT_ARGB565            1
-#define EVERGREEN_GRPH_FORMAT_ARGB4444           2
-#define EVERGREEN_GRPH_FORMAT_AI88               3
-#define EVERGREEN_GRPH_FORMAT_MONO16             4
-#define EVERGREEN_GRPH_FORMAT_BGRA5551           5
+#define GRPH_ARRAY_LINEAR_GENERAL      0
+#define GRPH_ARRAY_LINEAR_ALIGNED      1
+#define GRPH_ARRAY_1D_TILED_THIN1      2
+#define GRPH_ARRAY_2D_TILED_THIN1      4
+
+#define ES_AND_GS_AUTO       3
+#define BUF_SWAP_32BIT       (2 << 16)
+
+#define GRPH_DEPTH_8BPP                0
+#define GRPH_DEPTH_16BPP               1
+#define GRPH_DEPTH_32BPP               2
+
+/* 8 BPP */
+#define GRPH_FORMAT_INDEXED            0
+
+/* 16 BPP */
+#define GRPH_FORMAT_ARGB1555           0
+#define GRPH_FORMAT_ARGB565            1
+#define GRPH_FORMAT_ARGB4444           2
+#define GRPH_FORMAT_AI88               3
+#define GRPH_FORMAT_MONO16             4
+#define GRPH_FORMAT_BGRA5551           5
 
 /* 32 BPP */
-#define EVERGREEN_GRPH_FORMAT_ARGB8888           0
-#define EVERGREEN_GRPH_FORMAT_ARGB2101010        1
-#define EVERGREEN_GRPH_FORMAT_32BPP_DIG          2
-#define EVERGREEN_GRPH_FORMAT_8B_ARGB2101010     3
-#define EVERGREEN_GRPH_FORMAT_BGRA1010102        4
-#define EVERGREEN_GRPH_FORMAT_8B_BGRA1010102     5
-#define EVERGREEN_GRPH_FORMAT_RGB111110          6
-#define EVERGREEN_GRPH_FORMAT_BGR101111          7
-#define EVERGREEN_GRPH_BANK_HEIGHT(x)            (((x) & 0x3) << 11)
-#define EVERGREEN_ADDR_SURF_BANK_HEIGHT_1        0
-#define EVERGREEN_ADDR_SURF_BANK_HEIGHT_2        1
-#define EVERGREEN_ADDR_SURF_BANK_HEIGHT_4        2
-#define EVERGREEN_ADDR_SURF_BANK_HEIGHT_8        3
-#define EVERGREEN_GRPH_TILE_SPLIT(x)             (((x) & 0x7) << 13)
-#define EVERGREEN_ADDR_SURF_TILE_SPLIT_64B       0
-#define EVERGREEN_ADDR_SURF_TILE_SPLIT_128B      1
-#define EVERGREEN_ADDR_SURF_TILE_SPLIT_256B      2
-#define EVERGREEN_ADDR_SURF_TILE_SPLIT_512B      3
-#define EVERGREEN_ADDR_SURF_TILE_SPLIT_1KB       4
-#define EVERGREEN_ADDR_SURF_TILE_SPLIT_2KB       5
-#define EVERGREEN_ADDR_SURF_TILE_SPLIT_4KB       6
-#define EVERGREEN_GRPH_MACRO_TILE_ASPECT(x)      (((x) & 0x3) << 18)
-#define EVERGREEN_ADDR_SURF_MACRO_TILE_ASPECT_1  0
-#define EVERGREEN_ADDR_SURF_MACRO_TILE_ASPECT_2  1
-#define EVERGREEN_ADDR_SURF_MACRO_TILE_ASPECT_4  2
-#define EVERGREEN_ADDR_SURF_MACRO_TILE_ASPECT_8  3
-#define EVERGREEN_GRPH_ARRAY_MODE(x)             (((x) & 0x7) << 20)
-#define EVERGREEN_GRPH_ARRAY_LINEAR_GENERAL      0
-#define EVERGREEN_GRPH_ARRAY_LINEAR_ALIGNED      1
-#define EVERGREEN_GRPH_ARRAY_1D_TILED_THIN1      2
-#define EVERGREEN_GRPH_ARRAY_2D_TILED_THIN1      4
-#define EVERGREEN_ADDR_SURF_MACRO_TILE_ASPECT_1  0
-#define EVERGREEN_ADDR_SURF_MACRO_TILE_ASPECT_2  1
-#define EVERGREEN_ADDR_SURF_MACRO_TILE_ASPECT_4  2
-#define EVERGREEN_ADDR_SURF_MACRO_TILE_ASPECT_8  3
-
-#define EVERGREEN_GRPH_SWAP_CONTROL                     0x1a03
-#define EVERGREEN_GRPH_ENDIAN_SWAP(x)            (((x) & 0x3) << 0)
-#       define EVERGREEN_GRPH_ENDIAN_NONE               0
-#       define EVERGREEN_GRPH_ENDIAN_8IN16              1
-#       define EVERGREEN_GRPH_ENDIAN_8IN32              2
-#       define EVERGREEN_GRPH_ENDIAN_8IN64              3
-#define EVERGREEN_GRPH_RED_CROSSBAR(x)           (((x) & 0x3) << 4)
-#       define EVERGREEN_GRPH_RED_SEL_R                 0
-#       define EVERGREEN_GRPH_RED_SEL_G                 1
-#       define EVERGREEN_GRPH_RED_SEL_B                 2
-#       define EVERGREEN_GRPH_RED_SEL_A                 3
-#define EVERGREEN_GRPH_GREEN_CROSSBAR(x)         (((x) & 0x3) << 6)
-#       define EVERGREEN_GRPH_GREEN_SEL_G               0
-#       define EVERGREEN_GRPH_GREEN_SEL_B               1
-#       define EVERGREEN_GRPH_GREEN_SEL_A               2
-#       define EVERGREEN_GRPH_GREEN_SEL_R               3
-#define EVERGREEN_GRPH_BLUE_CROSSBAR(x)          (((x) & 0x3) << 8)
-#       define EVERGREEN_GRPH_BLUE_SEL_B                0
-#       define EVERGREEN_GRPH_BLUE_SEL_A                1
-#       define EVERGREEN_GRPH_BLUE_SEL_R                2
-#       define EVERGREEN_GRPH_BLUE_SEL_G                3
-#define EVERGREEN_GRPH_ALPHA_CROSSBAR(x)         (((x) & 0x3) << 10)
-#       define EVERGREEN_GRPH_ALPHA_SEL_A               0
-#       define EVERGREEN_GRPH_ALPHA_SEL_R               1
-#       define EVERGREEN_GRPH_ALPHA_SEL_G               2
-#       define EVERGREEN_GRPH_ALPHA_SEL_B               3
-
-#define EVERGREEN_D3VGA_CONTROL                         0xf8
-#define EVERGREEN_D4VGA_CONTROL                         0xf9
-#define EVERGREEN_D5VGA_CONTROL                         0xfa
-#define EVERGREEN_D6VGA_CONTROL                         0xfb
-
-#define EVERGREEN_GRPH_SURFACE_ADDRESS_MASK      0xffffff00
-
-#define EVERGREEN_GRPH_LUT_10BIT_BYPASS_CONTROL         0x1a02
-#define EVERGREEN_LUT_10BIT_BYPASS_EN            (1 << 8)
-
-#define EVERGREEN_GRPH_PITCH                            0x1a06
-#define EVERGREEN_GRPH_PRIMARY_SURFACE_ADDRESS_HIGH     0x1a07
-#define EVERGREEN_GRPH_SECONDARY_SURFACE_ADDRESS_HIGH   0x1a08
-#define EVERGREEN_GRPH_SURFACE_OFFSET_X                 0x1a09
-#define EVERGREEN_GRPH_SURFACE_OFFSET_Y                 0x1a0a
-#define EVERGREEN_GRPH_X_START                          0x1a0b
-#define EVERGREEN_GRPH_Y_START                          0x1a0c
-#define EVERGREEN_GRPH_X_END                            0x1a0d
-#define EVERGREEN_GRPH_Y_END                            0x1a0e
-#define EVERGREEN_GRPH_UPDATE                           0x1a11
-#define EVERGREEN_GRPH_SURFACE_UPDATE_PENDING    (1 << 2)
-#define EVERGREEN_GRPH_UPDATE_LOCK               (1 << 16)
-#define EVERGREEN_GRPH_FLIP_CONTROL                     0x1a12
-#define EVERGREEN_GRPH_SURFACE_UPDATE_H_RETRACE_EN (1 << 0)
-
-#define EVERGREEN_VIEWPORT_START                        0x1b5c
-#define EVERGREEN_VIEWPORT_SIZE                         0x1b5d
-#define EVERGREEN_DESKTOP_HEIGHT                        0x1ac1
-
-/* CUR blocks at 0x6998, 0x7598, 0x10198, 0x10d98, 0x11998, 0x12598 */
-#define EVERGREEN_CUR_CONTROL                           0x1a66
-#       define EVERGREEN_CURSOR_EN                      (1 << 0)
-#       define EVERGREEN_CURSOR_MODE(x)                 (((x) & 0x3) << 8)
-#       define EVERGREEN_CURSOR_MONO                    0
-#       define EVERGREEN_CURSOR_24_1                    1
-#       define EVERGREEN_CURSOR_24_8_PRE_MULT           2
-#       define EVERGREEN_CURSOR_24_8_UNPRE_MULT         3
-#       define EVERGREEN_CURSOR_2X_MAGNIFY              (1 << 16)
-#       define EVERGREEN_CURSOR_FORCE_MC_ON             (1 << 20)
-#       define EVERGREEN_CURSOR_URGENT_CONTROL(x)       (((x) & 0x7) << 24)
-#       define EVERGREEN_CURSOR_URGENT_ALWAYS           0
-#       define EVERGREEN_CURSOR_URGENT_1_8              1
-#       define EVERGREEN_CURSOR_URGENT_1_4              2
-#       define EVERGREEN_CURSOR_URGENT_3_8              3
-#       define EVERGREEN_CURSOR_URGENT_1_2              4
-#define EVERGREEN_CUR_SURFACE_ADDRESS                   0x1a67
-#       define EVERGREEN_CUR_SURFACE_ADDRESS_MASK       0xfffff000
-#define EVERGREEN_CUR_SIZE                              0x1a68
-#define EVERGREEN_CUR_SURFACE_ADDRESS_HIGH              0x1a69
-#define EVERGREEN_CUR_POSITION                          0x1a6a
-#define EVERGREEN_CUR_HOT_SPOT                          0x1a6b
-#define EVERGREEN_CUR_COLOR1                            0x1a6c
-#define EVERGREEN_CUR_COLOR2                            0x1a6d
-#define EVERGREEN_CUR_UPDATE                            0x1a6e
-#       define EVERGREEN_CURSOR_UPDATE_PENDING          (1 << 0)
-#       define EVERGREEN_CURSOR_UPDATE_TAKEN            (1 << 1)
-#       define EVERGREEN_CURSOR_UPDATE_LOCK             (1 << 16)
-#       define EVERGREEN_CURSOR_DISABLE_MULTIPLE_UPDATE (1 << 24)
-
-
-#define NI_INPUT_CSC_CONTROL                           0x1a35
-#       define NI_INPUT_CSC_GRPH_MODE(x)               (((x) & 0x3) << 0)
-#       define NI_INPUT_CSC_BYPASS                     0
-#       define NI_INPUT_CSC_PROG_COEFF                 1
-#       define NI_INPUT_CSC_PROG_SHARED_MATRIXA        2
-#       define NI_INPUT_CSC_OVL_MODE(x)                (((x) & 0x3) << 4)
-
-#define NI_OUTPUT_CSC_CONTROL                          0x1a3c
-#       define NI_OUTPUT_CSC_GRPH_MODE(x)              (((x) & 0x7) << 0)
-#       define NI_OUTPUT_CSC_BYPASS                    0
-#       define NI_OUTPUT_CSC_TV_RGB                    1
-#       define NI_OUTPUT_CSC_YCBCR_601                 2
-#       define NI_OUTPUT_CSC_YCBCR_709                 3
-#       define NI_OUTPUT_CSC_PROG_COEFF                4
-#       define NI_OUTPUT_CSC_PROG_SHARED_MATRIXB       5
-#       define NI_OUTPUT_CSC_OVL_MODE(x)               (((x) & 0x7) << 4)
-
-#define NI_DEGAMMA_CONTROL                             0x1a58
-#       define NI_GRPH_DEGAMMA_MODE(x)                 (((x) & 0x3) << 0)
-#       define NI_DEGAMMA_BYPASS                       0
-#       define NI_DEGAMMA_SRGB_24                      1
-#       define NI_DEGAMMA_XVYCC_222                    2
-#       define NI_OVL_DEGAMMA_MODE(x)                  (((x) & 0x3) << 4)
-#       define NI_ICON_DEGAMMA_MODE(x)                 (((x) & 0x3) << 8)
-#       define NI_CURSOR_DEGAMMA_MODE(x)               (((x) & 0x3) << 12)
-
-#define NI_GAMUT_REMAP_CONTROL                         0x1a59
-#       define NI_GRPH_GAMUT_REMAP_MODE(x)             (((x) & 0x3) << 0)
-#       define NI_GAMUT_REMAP_BYPASS                   0
-#       define NI_GAMUT_REMAP_PROG_COEFF               1
-#       define NI_GAMUT_REMAP_PROG_SHARED_MATRIXA      2
-#       define NI_GAMUT_REMAP_PROG_SHARED_MATRIXB      3
-#       define NI_OVL_GAMUT_REMAP_MODE(x)              (((x) & 0x3) << 4)
-
-#define NI_REGAMMA_CONTROL                             0x1aa0
-#       define NI_GRPH_REGAMMA_MODE(x)                 (((x) & 0x7) << 0)
-#       define NI_REGAMMA_BYPASS                       0
-#       define NI_REGAMMA_SRGB_24                      1
-#       define NI_REGAMMA_XVYCC_222                    2
-#       define NI_REGAMMA_PROG_A                       3
-#       define NI_REGAMMA_PROG_B                       4
-#       define NI_OVL_REGAMMA_MODE(x)                  (((x) & 0x7) << 4)
-
-
-#define NI_PRESCALE_GRPH_CONTROL                       0x1a2d
-#       define NI_GRPH_PRESCALE_BYPASS                 (1 << 4)
-
-#define NI_PRESCALE_OVL_CONTROL                        0x1a31
-#       define NI_OVL_PRESCALE_BYPASS                  (1 << 4)
-
-#define NI_INPUT_GAMMA_CONTROL                         0x1a10
-#       define NI_GRPH_INPUT_GAMMA_MODE(x)             (((x) & 0x3) << 0)
-#       define NI_INPUT_GAMMA_USE_LUT                  0
-#       define NI_INPUT_GAMMA_BYPASS                   1
-#       define NI_INPUT_GAMMA_SRGB_24                  2
-#       define NI_INPUT_GAMMA_XVYCC_222                3
-#       define NI_OVL_INPUT_GAMMA_MODE(x)              (((x) & 0x3) << 4)
-
-#define	BLACKOUT_MODE_MASK			0x00000007
-#define	VGA_RENDER_CONTROL			0xC0
-#define R_000300_VGA_RENDER_CONTROL             0xC0
-#define C_000300_VGA_VSTATUS_CNTL               0xFFFCFFFF
-#define EVERGREEN_CRTC_STATUS                   0x1BA3
-#define EVERGREEN_CRTC_V_BLANK                  (1 << 0)
-#define EVERGREEN_CRTC_STATUS_POSITION          0x1BA4
-/* CRTC blocks at 0x6df0, 0x79f0, 0x105f0, 0x111f0, 0x11df0, 0x129f0 */
-#define EVERGREEN_CRTC_V_BLANK_START_END                0x1b8d
-#define EVERGREEN_CRTC_CONTROL                          0x1b9c
-#       define EVERGREEN_CRTC_MASTER_EN                 (1 << 0)
-#       define EVERGREEN_CRTC_DISP_READ_REQUEST_DISABLE (1 << 24)
-#define EVERGREEN_CRTC_BLANK_CONTROL                    0x1b9d
-#       define EVERGREEN_CRTC_BLANK_DATA_EN             (1 << 8)
-#       define EVERGREEN_CRTC_V_BLANK                   (1 << 0)
-#define EVERGREEN_CRTC_STATUS_HV_COUNT                  0x1ba8
-#define EVERGREEN_CRTC_UPDATE_LOCK                      0x1bb5
-#define EVERGREEN_MASTER_UPDATE_LOCK                    0x1bbd
-#define EVERGREEN_MASTER_UPDATE_MODE                    0x1bbe
-#define EVERGREEN_GRPH_UPDATE_LOCK               (1 << 16)
-#define EVERGREEN_GRPH_PRIMARY_SURFACE_ADDRESS_HIGH     0x1a07
-#define EVERGREEN_GRPH_SECONDARY_SURFACE_ADDRESS_HIGH   0x1a08
-#define EVERGREEN_GRPH_PRIMARY_SURFACE_ADDRESS          0x1a04
-#define EVERGREEN_GRPH_SECONDARY_SURFACE_ADDRESS        0x1a05
-#define EVERGREEN_GRPH_UPDATE                           0x1a11
-#define EVERGREEN_VGA_MEMORY_BASE_ADDRESS               0xc4
-#define EVERGREEN_VGA_MEMORY_BASE_ADDRESS_HIGH          0xc9
-#define EVERGREEN_GRPH_SURFACE_UPDATE_PENDING    (1 << 2)
-
-#define mmVM_CONTEXT1_CNTL__xxRANGE_PROTECTION_FAULT_ENABLE_DEFAULT_MASK 0x10
-#define mmVM_CONTEXT1_CNTL__xxRANGE_PROTECTION_FAULT_ENABLE_DEFAULT__SHIFT 0x4
-#define mmVM_CONTEXT1_CNTL__xxDUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT_MASK 0x80
-#define mmVM_CONTEXT1_CNTL__xxDUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT__SHIFT 0x7
-#define mmVM_CONTEXT1_CNTL__xxPDE0_PROTECTION_FAULT_ENABLE_DEFAULT_MASK 0x400
-#define mmVM_CONTEXT1_CNTL__xxPDE0_PROTECTION_FAULT_ENABLE_DEFAULT__SHIFT 0xa
-#define mmVM_CONTEXT1_CNTL__xxVALID_PROTECTION_FAULT_ENABLE_DEFAULT_MASK 0x2000
-#define mmVM_CONTEXT1_CNTL__xxVALID_PROTECTION_FAULT_ENABLE_DEFAULT__SHIFT 0xd
-#define mmVM_CONTEXT1_CNTL__xxREAD_PROTECTION_FAULT_ENABLE_DEFAULT_MASK 0x10000
-#define mmVM_CONTEXT1_CNTL__xxREAD_PROTECTION_FAULT_ENABLE_DEFAULT__SHIFT 0x10
-#define mmVM_CONTEXT1_CNTL__xxWRITE_PROTECTION_FAULT_ENABLE_DEFAULT_MASK 0x80000
-#define mmVM_CONTEXT1_CNTL__xxWRITE_PROTECTION_FAULT_ENABLE_DEFAULT__SHIFT 0x13
-
-#define mmVM_CONTEXT1_PROTECTION_FAULT_STATUS__xxVMID_MASK 0x1e000000
-#define mmVM_CONTEXT1_PROTECTION_FAULT_STATUS__xxVMID__SHIFT 0x19
-#define mmVM_CONTEXT1_PROTECTION_FAULT_STATUS__xxPROTECTIONS_MASK 0xff
-#define mmVM_CONTEXT1_PROTECTION_FAULT_STATUS__xxPROTECTIONS__SHIFT 0x0
-#define mmVM_CONTEXT1_PROTECTION_FAULT_STATUS__xxMEMORY_CLIENT_ID_MASK 0xff000
-#define mmVM_CONTEXT1_PROTECTION_FAULT_STATUS__xxMEMORY_CLIENT_ID__SHIFT 0xc
-#define mmVM_CONTEXT1_PROTECTION_FAULT_STATUS__xxMEMORY_CLIENT_RW_MASK 0x1000000
-#define mmVM_CONTEXT1_PROTECTION_FAULT_STATUS__xxMEMORY_CLIENT_RW__SHIFT 0x18
-
-#define mmMC_SHARED_BLACKOUT_CNTL__xxBLACKOUT_MODE_MASK 0x7
-#define mmMC_SHARED_BLACKOUT_CNTL__xxBLACKOUT_MODE__SHIFT 0x0
-
-#define mmBIF_FB_EN__xxFB_READ_EN_MASK 0x1
-#define mmBIF_FB_EN__xxFB_READ_EN__SHIFT 0x0
-#define mmBIF_FB_EN__xxFB_WRITE_EN_MASK 0x2
-#define mmBIF_FB_EN__xxFB_WRITE_EN__SHIFT 0x1
-
-#define mmSRBM_SOFT_RESET__xxSOFT_RESET_VMC_MASK 0x20000
-#define mmSRBM_SOFT_RESET__xxSOFT_RESET_VMC__SHIFT 0x11
-#define mmSRBM_SOFT_RESET__xxSOFT_RESET_MC_MASK 0x800
-#define mmSRBM_SOFT_RESET__xxSOFT_RESET_MC__SHIFT 0xb
-
-#define VM_CONTEXT1_CNTL__RANGE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK 0x8
-#define VM_CONTEXT1_CNTL__RANGE_PROTECTION_FAULT_ENABLE_INTERRUPT__SHIFT 0x3
-#define VM_CONTEXT1_CNTL__DUMMY_PAGE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK 0x40
-#define VM_CONTEXT1_CNTL__DUMMY_PAGE_PROTECTION_FAULT_ENABLE_INTERRUPT__SHIFT 0x6
-#define VM_CONTEXT1_CNTL__PDE0_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK 0x200
-#define VM_CONTEXT1_CNTL__PDE0_PROTECTION_FAULT_ENABLE_INTERRUPT__SHIFT 0x9
-#define VM_CONTEXT1_CNTL__VALID_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK 0x1000
-#define VM_CONTEXT1_CNTL__VALID_PROTECTION_FAULT_ENABLE_INTERRUPT__SHIFT 0xc
-#define VM_CONTEXT1_CNTL__READ_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK 0x8000
-#define VM_CONTEXT1_CNTL__READ_PROTECTION_FAULT_ENABLE_INTERRUPT__SHIFT 0xf
-#define VM_CONTEXT1_CNTL__WRITE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK 0x40000
-#define VM_CONTEXT1_CNTL__WRITE_PROTECTION_FAULT_ENABLE_INTERRUPT__SHIFT 0x12
+#define GRPH_FORMAT_ARGB8888           0
+#define GRPH_FORMAT_ARGB2101010        1
+#define GRPH_FORMAT_32BPP_DIG          2
+#define GRPH_FORMAT_8B_ARGB2101010     3
+#define GRPH_FORMAT_BGRA1010102        4
+#define GRPH_FORMAT_8B_BGRA1010102     5
+#define GRPH_FORMAT_RGB111110          6
+#define GRPH_FORMAT_BGR101111          7
+
+#define GRPH_ENDIAN_NONE               0
+#define GRPH_ENDIAN_8IN16              1
+#define GRPH_ENDIAN_8IN32              2
+#define GRPH_ENDIAN_8IN64              3
+#define GRPH_RED_SEL_R                 0
+#define GRPH_RED_SEL_G                 1
+#define GRPH_RED_SEL_B                 2
+#define GRPH_RED_SEL_A                 3
+
+#define GRPH_GREEN_SEL_G               0
+#define GRPH_GREEN_SEL_B               1
+#define GRPH_GREEN_SEL_A               2
+#define GRPH_GREEN_SEL_R               3
+
+#define GRPH_BLUE_SEL_B                0
+#define GRPH_BLUE_SEL_A                1
+#define GRPH_BLUE_SEL_R                2
+#define GRPH_BLUE_SEL_G                3
+
+#define GRPH_ALPHA_SEL_A               0
+#define GRPH_ALPHA_SEL_R               1
+#define GRPH_ALPHA_SEL_G               2
+#define GRPH_ALPHA_SEL_B               3
+
+/* CUR_CONTROL */
+	#define CURSOR_MONO                    0
+	#define CURSOR_24_1                    1
+	#define CURSOR_24_8_PRE_MULT           2
+	#define CURSOR_24_8_UNPRE_MULT         3
+	#define CURSOR_URGENT_ALWAYS           0
+	#define CURSOR_URGENT_1_8              1
+	#define CURSOR_URGENT_1_4              2
+	#define CURSOR_URGENT_3_8              3
+	#define CURSOR_URGENT_1_2              4
+
+/* INPUT_CSC_CONTROL */
+#       define INPUT_CSC_BYPASS                     0
+#       define INPUT_CSC_PROG_COEFF                 1
+#       define INPUT_CSC_PROG_SHARED_MATRIXA        2
+
+/* OUTPUT_CSC_CONTROL */
+#       define OUTPUT_CSC_BYPASS                    0
+#       define OUTPUT_CSC_TV_RGB                    1
+#       define OUTPUT_CSC_YCBCR_601                 2
+#       define OUTPUT_CSC_YCBCR_709                 3
+#       define OUTPUT_CSC_PROG_COEFF                4
+#       define OUTPUT_CSC_PROG_SHARED_MATRIXB       5
+
+/* DEGAMMA_CONTROL */
+#       define DEGAMMA_BYPASS                       0
+#       define DEGAMMA_SRGB_24                      1
+#       define DEGAMMA_XVYCC_222                    2
+
+/* GAMUT_REMAP_CONTROL */
+#       define GAMUT_REMAP_BYPASS                   0
+#       define GAMUT_REMAP_PROG_COEFF               1
+#       define GAMUT_REMAP_PROG_SHARED_MATRIXA      2
+#       define GAMUT_REMAP_PROG_SHARED_MATRIXB      3
+
+/* REGAMMA_CONTROL */
+#       define REGAMMA_BYPASS                       0
+#       define REGAMMA_SRGB_24                      1
+#       define REGAMMA_XVYCC_222                    2
+#       define REGAMMA_PROG_A                       3
+#       define REGAMMA_PROG_B                       4
+
+
+/* INPUT_GAMMA_CONTROL */
+#       define INPUT_GAMMA_USE_LUT                  0
+#       define INPUT_GAMMA_BYPASS                   1
+#       define INPUT_GAMMA_SRGB_24                  2
+#       define INPUT_GAMMA_XVYCC_222                3
 
 #define MC_SEQ_MISC0__MT__MASK	0xf0000000
 #define MC_SEQ_MISC0__MT__GDDR1  0x10000000
@@ -2426,28 +740,14 @@
 #define MC_SEQ_MISC0__MT__HBM    0x60000000
 #define MC_SEQ_MISC0__MT__DDR3   0xB0000000
 
-#define GRBM_STATUS__GUI_ACTIVE_MASK 0x80000000
 #define CP_INT_CNTL_RING__TIME_STAMP_INT_ENABLE_MASK 0x4000000
-#define CP_INT_CNTL_RING0__PRIV_REG_INT_ENABLE_MASK 0x800000
-#define CP_INT_CNTL_RING0__PRIV_INSTR_INT_ENABLE_MASK 0x400000
 #define PACKET3_SEM_WAIT_ON_SIGNAL    (0x1 << 12)
 #define PACKET3_SEM_SEL_SIGNAL	    (0x6 << 29)
 #define PACKET3_SEM_SEL_WAIT	    (0x7 << 29)
 
-#define CONFIG_CNTL	0x1509
-#define CC_DRM_ID_STRAPS	0X1559
 #define AMDGPU_PCIE_INDEX	0xc
 #define AMDGPU_PCIE_DATA	0xd
 
-#define DMA_SEM_INCOMPLETE_TIMER_CNTL                     0x3411
-#define DMA_SEM_WAIT_FAIL_TIMER_CNTL                      0x3412
-#define DMA_MODE                                          0x342f
-#define DMA_RB_RPTR_ADDR_HI                               0x3407
-#define DMA_RB_RPTR_ADDR_LO                               0x3408
-#define DMA_BUSY_MASK 0x20
-#define DMA1_BUSY_MASK 0X40
-#define SDMA_MAX_INSTANCE 2
-
 #define PCIE_BUS_CLK    10000
 #define TCLK            (PCIE_BUS_CLK / 10)
 #define	PCIE_PORT_INDEX					0xe
@@ -2457,8 +757,6 @@
 #define EVERGREEN_PIF_PHY1_INDEX                        0x10
 #define EVERGREEN_PIF_PHY1_DATA				0x14
 
-#define	MC_VM_FB_OFFSET					0x81a
-
 /* Discrete VCE clocks */
 #define CG_VCEPLL_FUNC_CNTL                             0xc0030600
 #define    VCEPLL_RESET_MASK                            0x00000001
diff --git a/drivers/gpu/drm/amd/amdgpu/sienna_cichlid.c b/drivers/gpu/drm/amd/amdgpu/sienna_cichlid.c
new file mode 100644
index 000000000000..2594467bdd87
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/sienna_cichlid.c
@@ -0,0 +1,298 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "sienna_cichlid.h"
+#include "amdgpu_reset.h"
+#include "amdgpu_amdkfd.h"
+#include "amdgpu_dpm.h"
+#include "amdgpu_job.h"
+#include "amdgpu_ring.h"
+#include "amdgpu_ras.h"
+#include "amdgpu_psp.h"
+#include "amdgpu_xgmi.h"
+
+static bool sienna_cichlid_is_mode2_default(struct amdgpu_reset_control *reset_ctl)
+{
+#if 0
+	struct amdgpu_device *adev = (struct amdgpu_device *)reset_ctl->handle;
+
+	if (amdgpu_ip_version(adev, MP1_HWIP, 0) == IP_VERSION(11, 0, 7) &&
+	    adev->pm.fw_version >= 0x3a5500 && !amdgpu_sriov_vf(adev))
+		return true;
+#endif
+	return amdgpu_reset_method == AMD_RESET_METHOD_MODE2;
+}
+
+static struct amdgpu_reset_handler *
+sienna_cichlid_get_reset_handler(struct amdgpu_reset_control *reset_ctl,
+			    struct amdgpu_reset_context *reset_context)
+{
+	struct amdgpu_reset_handler *handler;
+	int i;
+
+	if (reset_context->method != AMD_RESET_METHOD_NONE) {
+		for_each_handler(i, handler, reset_ctl)	{
+			if (handler->reset_method == reset_context->method)
+				return handler;
+		}
+	}
+
+	if (sienna_cichlid_is_mode2_default(reset_ctl)) {
+		for_each_handler(i, handler, reset_ctl)	{
+			if (handler->reset_method == AMD_RESET_METHOD_MODE2)
+				return handler;
+		}
+	}
+
+	return NULL;
+}
+
+static int sienna_cichlid_mode2_suspend_ip(struct amdgpu_device *adev)
+{
+	int r, i;
+
+	amdgpu_device_set_pg_state(adev, AMD_PG_STATE_UNGATE);
+	amdgpu_device_set_cg_state(adev, AMD_CG_STATE_UNGATE);
+
+	for (i = adev->num_ip_blocks - 1; i >= 0; i--) {
+		if (!(adev->ip_blocks[i].version->type ==
+			      AMD_IP_BLOCK_TYPE_GFX ||
+		      adev->ip_blocks[i].version->type ==
+			      AMD_IP_BLOCK_TYPE_SDMA))
+			continue;
+
+		r = amdgpu_ip_block_suspend(&adev->ip_blocks[i]);
+		if (r)
+			return r;
+	}
+
+	return 0;
+}
+
+static int
+sienna_cichlid_mode2_prepare_hwcontext(struct amdgpu_reset_control *reset_ctl,
+				  struct amdgpu_reset_context *reset_context)
+{
+	int r = 0;
+	struct amdgpu_device *adev = (struct amdgpu_device *)reset_ctl->handle;
+
+	if (!amdgpu_sriov_vf(adev)) {
+		if (adev->gfxhub.funcs->mode2_save_regs)
+			adev->gfxhub.funcs->mode2_save_regs(adev);
+		if (adev->gfxhub.funcs->halt)
+			adev->gfxhub.funcs->halt(adev);
+		r = sienna_cichlid_mode2_suspend_ip(adev);
+	}
+
+	return r;
+}
+
+static void sienna_cichlid_async_reset(struct work_struct *work)
+{
+	struct amdgpu_reset_handler *handler;
+	struct amdgpu_reset_control *reset_ctl =
+		container_of(work, struct amdgpu_reset_control, reset_work);
+	struct amdgpu_device *adev = (struct amdgpu_device *)reset_ctl->handle;
+	int i;
+
+	for_each_handler(i, handler, reset_ctl)	{
+		if (handler->reset_method == reset_ctl->active_reset) {
+			dev_dbg(adev->dev, "Resetting device\n");
+			handler->do_reset(adev);
+			break;
+		}
+	}
+}
+
+static int sienna_cichlid_mode2_reset(struct amdgpu_device *adev)
+{
+	/* disable BM */
+	pci_clear_master(adev->pdev);
+	return amdgpu_dpm_mode2_reset(adev);
+}
+
+static int
+sienna_cichlid_mode2_perform_reset(struct amdgpu_reset_control *reset_ctl,
+			      struct amdgpu_reset_context *reset_context)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)reset_ctl->handle;
+	int r;
+
+	r = sienna_cichlid_mode2_reset(adev);
+	if (r) {
+		dev_err(adev->dev,
+			"ASIC reset failed with error, %d ", r);
+	}
+	return r;
+}
+
+static int sienna_cichlid_mode2_restore_ip(struct amdgpu_device *adev)
+{
+	int i, r;
+	struct psp_context *psp = &adev->psp;
+
+	r = psp_rlc_autoload_start(psp);
+	if (r) {
+		dev_err(adev->dev, "Failed to start rlc autoload\n");
+		return r;
+	}
+
+	/* Reinit GFXHUB */
+	if (adev->gfxhub.funcs->mode2_restore_regs)
+		adev->gfxhub.funcs->mode2_restore_regs(adev);
+	adev->gfxhub.funcs->init(adev);
+	r = adev->gfxhub.funcs->gart_enable(adev);
+	if (r) {
+		dev_err(adev->dev, "GFXHUB gart reenable failed after reset\n");
+		return r;
+	}
+
+	for (i = 0; i < adev->num_ip_blocks; i++) {
+		if (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_IH) {
+			r = amdgpu_ip_block_resume(&adev->ip_blocks[i]);
+			if (r)
+				return r;
+		}
+	}
+
+	for (i = 0; i < adev->num_ip_blocks; i++) {
+		if (!(adev->ip_blocks[i].version->type ==
+			      AMD_IP_BLOCK_TYPE_GFX ||
+		      adev->ip_blocks[i].version->type ==
+			      AMD_IP_BLOCK_TYPE_SDMA))
+			continue;
+		r = amdgpu_ip_block_resume(&adev->ip_blocks[i]);
+		if (r)
+			return r;
+	}
+
+	for (i = 0; i < adev->num_ip_blocks; i++) {
+		if (!(adev->ip_blocks[i].version->type ==
+			      AMD_IP_BLOCK_TYPE_GFX ||
+		      adev->ip_blocks[i].version->type ==
+			      AMD_IP_BLOCK_TYPE_SDMA))
+			continue;
+
+		if (adev->ip_blocks[i].version->funcs->late_init) {
+			r = adev->ip_blocks[i].version->funcs->late_init(
+				&adev->ip_blocks[i]);
+			if (r) {
+				dev_err(adev->dev,
+					"late_init of IP block <%s> failed %d after reset\n",
+					adev->ip_blocks[i].version->funcs->name,
+					r);
+				return r;
+			}
+		}
+		adev->ip_blocks[i].status.late_initialized = true;
+	}
+
+	amdgpu_device_set_cg_state(adev, AMD_CG_STATE_GATE);
+	amdgpu_device_set_pg_state(adev, AMD_PG_STATE_GATE);
+
+	return r;
+}
+
+static int
+sienna_cichlid_mode2_restore_hwcontext(struct amdgpu_reset_control *reset_ctl,
+				  struct amdgpu_reset_context *reset_context)
+{
+	int r;
+	struct amdgpu_device *tmp_adev = (struct amdgpu_device *)reset_ctl->handle;
+
+	amdgpu_set_init_level(tmp_adev, AMDGPU_INIT_LEVEL_RESET_RECOVERY);
+	dev_info(tmp_adev->dev,
+			"GPU reset succeeded, trying to resume\n");
+	r = sienna_cichlid_mode2_restore_ip(tmp_adev);
+	if (r)
+		goto end;
+
+	/*
+	* Add this ASIC as tracked as reset was already
+	* complete successfully.
+	*/
+	amdgpu_register_gpu_instance(tmp_adev);
+
+	/* Resume RAS */
+	amdgpu_ras_resume(tmp_adev);
+
+	amdgpu_irq_gpu_reset_resume_helper(tmp_adev);
+
+	amdgpu_set_init_level(tmp_adev, AMDGPU_INIT_LEVEL_DEFAULT);
+	r = amdgpu_ib_ring_tests(tmp_adev);
+	if (r) {
+		dev_err(tmp_adev->dev,
+			"ib ring test failed (%d).\n", r);
+		r = -EAGAIN;
+		goto end;
+	}
+
+end:
+	if (r)
+		return -EAGAIN;
+	else
+		return r;
+}
+
+static struct amdgpu_reset_handler sienna_cichlid_mode2_handler = {
+	.reset_method		= AMD_RESET_METHOD_MODE2,
+	.prepare_env		= NULL,
+	.prepare_hwcontext	= sienna_cichlid_mode2_prepare_hwcontext,
+	.perform_reset		= sienna_cichlid_mode2_perform_reset,
+	.restore_hwcontext	= sienna_cichlid_mode2_restore_hwcontext,
+	.restore_env		= NULL,
+	.do_reset		= sienna_cichlid_mode2_reset,
+};
+
+static struct amdgpu_reset_handler
+	*sienna_cichlid_rst_handlers[AMDGPU_RESET_MAX_HANDLERS] = {
+		&sienna_cichlid_mode2_handler,
+	};
+
+int sienna_cichlid_reset_init(struct amdgpu_device *adev)
+{
+	struct amdgpu_reset_control *reset_ctl;
+
+	reset_ctl = kzalloc(sizeof(*reset_ctl), GFP_KERNEL);
+	if (!reset_ctl)
+		return -ENOMEM;
+
+	reset_ctl->handle = adev;
+	reset_ctl->async_reset = sienna_cichlid_async_reset;
+	reset_ctl->active_reset = AMD_RESET_METHOD_NONE;
+	reset_ctl->get_reset_handler = sienna_cichlid_get_reset_handler;
+
+	INIT_WORK(&reset_ctl->reset_work, reset_ctl->async_reset);
+	/* Only mode2 is handled through reset control now */
+	reset_ctl->reset_handlers = &sienna_cichlid_rst_handlers;
+	adev->reset_cntl = reset_ctl;
+
+	return 0;
+}
+
+int sienna_cichlid_reset_fini(struct amdgpu_device *adev)
+{
+	kfree(adev->reset_cntl);
+	adev->reset_cntl = NULL;
+	return 0;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/sienna_cichlid.h b/drivers/gpu/drm/amd/amdgpu/sienna_cichlid.h
new file mode 100644
index 000000000000..5213b162dacd
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/sienna_cichlid.h
@@ -0,0 +1,32 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __SIENNA_CICHLID_H__
+#define __SIENNA_CICHLID_H__
+
+#include "amdgpu.h"
+
+int sienna_cichlid_reset_init(struct amdgpu_device *adev);
+int sienna_cichlid_reset_fini(struct amdgpu_device *adev);
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/smu_v11_0_i2c.c b/drivers/gpu/drm/amd/amdgpu/smu_v11_0_i2c.c
index 73ffa8fde3df..68aef47254a9 100644
--- a/drivers/gpu/drm/amd/amdgpu/smu_v11_0_i2c.c
+++ b/drivers/gpu/drm/amd/amdgpu/smu_v11_0_i2c.c
@@ -26,6 +26,7 @@
 
 #include "smu_v11_0_i2c.h"
 #include "amdgpu.h"
+#include "amdgpu_dpm.h"
 #include "soc15_common.h"
 #include <drm/drm_fixed.h>
 #include <drm/drm_drv.h>
@@ -43,11 +44,10 @@
 
 #define I2C_X_RESTART         BIT(31)
 
-#define to_amdgpu_device(x) (container_of(x, struct amdgpu_device, pm.smu_i2c))
-
 static void smu_v11_0_i2c_set_clock_gating(struct i2c_adapter *control, bool en)
 {
-	struct amdgpu_device *adev = to_amdgpu_device(control);
+	struct amdgpu_smu_i2c_bus *smu_i2c = i2c_get_adapdata(control);
+	struct amdgpu_device *adev = smu_i2c->adev;
 	uint32_t reg = RREG32_SOC15(SMUIO, 0, mmSMUIO_PWRMGT);
 
 	reg = REG_SET_FIELD(reg, SMUIO_PWRMGT, i2c_clk_gate_en, en ? 1 : 0);
@@ -75,7 +75,8 @@ static void smu_v11_0_i2c_set_clock_gating(struct i2c_adapter *control, bool en)
 
 static int smu_v11_0_i2c_enable(struct i2c_adapter *control, bool enable)
 {
-	struct amdgpu_device *adev = to_amdgpu_device(control);
+	struct amdgpu_smu_i2c_bus *smu_i2c = i2c_get_adapdata(control);
+	struct amdgpu_device *adev = smu_i2c->adev;
 
 	WREG32_SOC15(SMUIO, 0, mmCKSVII2C_IC_ENABLE, enable ? 1 : 0);
 
@@ -100,7 +101,8 @@ static int smu_v11_0_i2c_enable(struct i2c_adapter *control, bool enable)
 
 static void smu_v11_0_i2c_clear_status(struct i2c_adapter *control)
 {
-	struct amdgpu_device *adev = to_amdgpu_device(control);
+	struct amdgpu_smu_i2c_bus *smu_i2c = i2c_get_adapdata(control);
+	struct amdgpu_device *adev = smu_i2c->adev;
 	/* do */
 	{
 		RREG32_SOC15(SMUIO, 0, mmCKSVII2C_IC_CLR_INTR);
@@ -110,7 +112,8 @@ static void smu_v11_0_i2c_clear_status(struct i2c_adapter *control)
 
 static void smu_v11_0_i2c_configure(struct i2c_adapter *control)
 {
-	struct amdgpu_device *adev = to_amdgpu_device(control);
+	struct amdgpu_smu_i2c_bus *smu_i2c = i2c_get_adapdata(control);
+	struct amdgpu_device *adev = smu_i2c->adev;
 	uint32_t reg = 0;
 
 	reg = REG_SET_FIELD(reg, CKSVII2C_IC_CON, IC_SLAVE_DISABLE, 1);
@@ -131,7 +134,8 @@ static void smu_v11_0_i2c_configure(struct i2c_adapter *control)
 
 static void smu_v11_0_i2c_set_clock(struct i2c_adapter *control)
 {
-	struct amdgpu_device *adev = to_amdgpu_device(control);
+	struct amdgpu_smu_i2c_bus *smu_i2c = i2c_get_adapdata(control);
+	struct amdgpu_device *adev = smu_i2c->adev;
 
 	/*
 	 * Standard mode speed, These values are taken from SMUIO MAS,
@@ -154,7 +158,8 @@ static void smu_v11_0_i2c_set_clock(struct i2c_adapter *control)
 
 static void smu_v11_0_i2c_set_address(struct i2c_adapter *control, u16 address)
 {
-	struct amdgpu_device *adev = to_amdgpu_device(control);
+	struct amdgpu_smu_i2c_bus *smu_i2c = i2c_get_adapdata(control);
+	struct amdgpu_device *adev = smu_i2c->adev;
 
 	/* The IC_TAR::IC_TAR field is 10-bits wide.
 	 * It takes a 7-bit or 10-bit addresses as an address,
@@ -165,7 +170,8 @@ static void smu_v11_0_i2c_set_address(struct i2c_adapter *control, u16 address)
 
 static uint32_t smu_v11_0_i2c_poll_tx_status(struct i2c_adapter *control)
 {
-	struct amdgpu_device *adev = to_amdgpu_device(control);
+	struct amdgpu_smu_i2c_bus *smu_i2c = i2c_get_adapdata(control);
+	struct amdgpu_device *adev = smu_i2c->adev;
 	uint32_t ret = I2C_OK;
 	uint32_t reg, reg_c_tx_abrt_source;
 
@@ -216,7 +222,8 @@ static uint32_t smu_v11_0_i2c_poll_tx_status(struct i2c_adapter *control)
 
 static uint32_t smu_v11_0_i2c_poll_rx_status(struct i2c_adapter *control)
 {
-	struct amdgpu_device *adev = to_amdgpu_device(control);
+	struct amdgpu_smu_i2c_bus *smu_i2c = i2c_get_adapdata(control);
+	struct amdgpu_device *adev = smu_i2c->adev;
 	uint32_t ret = I2C_OK;
 	uint32_t reg_ic_status, reg_c_tx_abrt_source;
 
@@ -262,7 +269,8 @@ static uint32_t smu_v11_0_i2c_transmit(struct i2c_adapter *control,
 				       u16 address, u8 *data,
 				       u32 numbytes, u32 i2c_flag)
 {
-	struct amdgpu_device *adev = to_amdgpu_device(control);
+	struct amdgpu_smu_i2c_bus *smu_i2c = i2c_get_adapdata(control);
+	struct amdgpu_device *adev = smu_i2c->adev;
 	u32 bytes_sent, reg, ret = I2C_OK;
 	unsigned long  timeout_counter;
 
@@ -360,7 +368,8 @@ static uint32_t smu_v11_0_i2c_receive(struct i2c_adapter *control,
 				      u16 address, u8 *data,
 				      u32 numbytes, u32 i2c_flag)
 {
-	struct amdgpu_device *adev = to_amdgpu_device(control);
+	struct amdgpu_smu_i2c_bus *smu_i2c = i2c_get_adapdata(control);
+	struct amdgpu_device *adev = smu_i2c->adev;
 	uint32_t bytes_received, ret = I2C_OK;
 
 	bytes_received = 0;
@@ -431,7 +440,8 @@ static uint32_t smu_v11_0_i2c_receive(struct i2c_adapter *control,
 
 static void smu_v11_0_i2c_abort(struct i2c_adapter *control)
 {
-	struct amdgpu_device *adev = to_amdgpu_device(control);
+	struct amdgpu_smu_i2c_bus *smu_i2c = i2c_get_adapdata(control);
+	struct amdgpu_device *adev = smu_i2c->adev;
 	uint32_t reg = 0;
 
 	/* Enable I2C engine; */
@@ -447,7 +457,8 @@ static void smu_v11_0_i2c_abort(struct i2c_adapter *control)
 
 static bool smu_v11_0_i2c_activity_done(struct i2c_adapter *control)
 {
-	struct amdgpu_device *adev = to_amdgpu_device(control);
+	struct amdgpu_smu_i2c_bus *smu_i2c = i2c_get_adapdata(control);
+	struct amdgpu_device *adev = smu_i2c->adev;
 
 	const uint32_t IDLE_TIMEOUT = 1024;
 	uint32_t timeout_count = 0;
@@ -508,7 +519,8 @@ static void smu_v11_0_i2c_init(struct i2c_adapter *control)
 
 static void smu_v11_0_i2c_fini(struct i2c_adapter *control)
 {
-	struct amdgpu_device *adev = to_amdgpu_device(control);
+	struct amdgpu_smu_i2c_bus *smu_i2c = i2c_get_adapdata(control);
+	struct amdgpu_device *adev = smu_i2c->adev;
 	u32 status, enable, en_stat;
 	int res;
 
@@ -543,7 +555,8 @@ static void smu_v11_0_i2c_fini(struct i2c_adapter *control)
 
 static bool smu_v11_0_i2c_bus_lock(struct i2c_adapter *control)
 {
-	struct amdgpu_device *adev = to_amdgpu_device(control);
+	struct amdgpu_smu_i2c_bus *smu_i2c = i2c_get_adapdata(control);
+	struct amdgpu_device *adev = smu_i2c->adev;
 
 	/* Send  PPSMC_MSG_RequestI2CBus */
 	if (!amdgpu_dpm_smu_i2c_bus_access(adev, true))
@@ -554,7 +567,8 @@ static bool smu_v11_0_i2c_bus_lock(struct i2c_adapter *control)
 
 static bool smu_v11_0_i2c_bus_unlock(struct i2c_adapter *control)
 {
-	struct amdgpu_device *adev = to_amdgpu_device(control);
+	struct amdgpu_smu_i2c_bus *smu_i2c = i2c_get_adapdata(control);
+	struct amdgpu_device *adev = smu_i2c->adev;
 
 	/* Send  PPSMC_MSG_ReleaseI2CBus */
 	if (!amdgpu_dpm_smu_i2c_bus_access(adev, false))
@@ -587,16 +601,17 @@ static uint32_t smu_v11_0_i2c_write_data(struct i2c_adapter *control,
 
 	if (ret != I2C_OK)
 		DRM_ERROR("WriteI2CData() - I2C error occurred :%x", ret);
-	
+
 	return ret;
 
 }
 
 static void lock_bus(struct i2c_adapter *i2c, unsigned int flags)
 {
-	struct amdgpu_device *adev = to_amdgpu_device(i2c);
+	struct amdgpu_smu_i2c_bus *smu_i2c = i2c_get_adapdata(i2c);
+	struct amdgpu_device *adev = smu_i2c->adev;
 
-	mutex_lock(&adev->pm.smu_i2c_mutex);
+	mutex_lock(&smu_i2c->mutex);
 	if (!smu_v11_0_i2c_bus_lock(i2c))
 		DRM_ERROR("Failed to lock the bus from SMU");
 	else
@@ -611,13 +626,14 @@ static int trylock_bus(struct i2c_adapter *i2c, unsigned int flags)
 
 static void unlock_bus(struct i2c_adapter *i2c, unsigned int flags)
 {
-	struct amdgpu_device *adev = to_amdgpu_device(i2c);
+	struct amdgpu_smu_i2c_bus *smu_i2c = i2c_get_adapdata(i2c);
+	struct amdgpu_device *adev = smu_i2c->adev;
 
 	if (!smu_v11_0_i2c_bus_unlock(i2c))
 		DRM_ERROR("Failed to unlock the bus from SMU");
 	else
 		adev->pm.bus_locked = false;
-	mutex_unlock(&adev->pm.smu_i2c_mutex);
+	mutex_unlock(&smu_i2c->mutex);
 }
 
 static const struct i2c_lock_operations smu_v11_0_i2c_i2c_lock_ops = {
@@ -706,30 +722,38 @@ static const struct i2c_adapter_quirks smu_v11_0_i2c_control_quirks = {
 	.flags = I2C_AQ_NO_ZERO_LEN,
 };
 
-int smu_v11_0_i2c_control_init(struct i2c_adapter *control)
+int smu_v11_0_i2c_control_init(struct amdgpu_device *adev)
 {
-	struct amdgpu_device *adev = to_amdgpu_device(control);
+	struct amdgpu_smu_i2c_bus *smu_i2c = &adev->pm.smu_i2c[0];
+	struct i2c_adapter *control = &smu_i2c->adapter;
 	int res;
 
-	mutex_init(&adev->pm.smu_i2c_mutex);
+	smu_i2c->adev = adev;
+	smu_i2c->port = 0;
+	mutex_init(&smu_i2c->mutex);
 	control->owner = THIS_MODULE;
 	control->class = I2C_CLASS_HWMON;
 	control->dev.parent = &adev->pdev->dev;
 	control->algo = &smu_v11_0_i2c_algo;
-	snprintf(control->name, sizeof(control->name), "AMDGPU SMU");
+	snprintf(control->name, sizeof(control->name), "AMDGPU SMU 0");
 	control->lock_ops = &smu_v11_0_i2c_i2c_lock_ops;
 	control->quirks = &smu_v11_0_i2c_control_quirks;
+	i2c_set_adapdata(control, smu_i2c);
+
+	adev->pm.ras_eeprom_i2c_bus = &adev->pm.smu_i2c[0].adapter;
+	adev->pm.fru_eeprom_i2c_bus = &adev->pm.smu_i2c[0].adapter;
 
-	res = i2c_add_adapter(control);
+	res = devm_i2c_add_adapter(adev->dev, control);
 	if (res)
 		DRM_ERROR("Failed to register hw i2c, err: %d\n", res);
 
 	return res;
 }
 
-void smu_v11_0_i2c_control_fini(struct i2c_adapter *control)
+void smu_v11_0_i2c_control_fini(struct amdgpu_device *adev)
 {
-	i2c_del_adapter(control);
+	adev->pm.ras_eeprom_i2c_bus = NULL;
+	adev->pm.fru_eeprom_i2c_bus = NULL;
 }
 
 /*
diff --git a/drivers/gpu/drm/amd/amdgpu/smu_v11_0_i2c.h b/drivers/gpu/drm/amd/amdgpu/smu_v11_0_i2c.h
index 44467c05f642..96ad14288a0c 100644
--- a/drivers/gpu/drm/amd/amdgpu/smu_v11_0_i2c.h
+++ b/drivers/gpu/drm/amd/amdgpu/smu_v11_0_i2c.h
@@ -26,9 +26,9 @@
 
 #include <linux/types.h>
 
-struct i2c_adapter;
+struct amdgpu_device;
 
-int smu_v11_0_i2c_control_init(struct i2c_adapter *control);
-void smu_v11_0_i2c_control_fini(struct i2c_adapter *control);
+int smu_v11_0_i2c_control_init(struct amdgpu_device *adev);
+void smu_v11_0_i2c_control_fini(struct amdgpu_device *adev);
 
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/smu_v13_0_10.c b/drivers/gpu/drm/amd/amdgpu/smu_v13_0_10.c
new file mode 100644
index 000000000000..70569ea906bc
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/smu_v13_0_10.c
@@ -0,0 +1,296 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "smu_v13_0_10.h"
+#include "amdgpu_reset.h"
+#include "amdgpu_dpm.h"
+#include "amdgpu_job.h"
+#include "amdgpu_ring.h"
+#include "amdgpu_ras.h"
+#include "amdgpu_psp.h"
+
+static bool smu_v13_0_10_is_mode2_default(struct amdgpu_reset_control *reset_ctl)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)reset_ctl->handle;
+	if (adev->pm.fw_version >= 0x00502005 && !amdgpu_sriov_vf(adev))
+		return true;
+
+	return false;
+}
+
+static struct amdgpu_reset_handler *
+smu_v13_0_10_get_reset_handler(struct amdgpu_reset_control *reset_ctl,
+			    struct amdgpu_reset_context *reset_context)
+{
+	struct amdgpu_reset_handler *handler;
+	struct amdgpu_device *adev = (struct amdgpu_device *)reset_ctl->handle;
+	int i;
+
+	if (reset_context->method != AMD_RESET_METHOD_NONE) {
+		for_each_handler(i, handler, reset_ctl) {
+			if (handler->reset_method == reset_context->method)
+				return handler;
+		}
+	}
+
+	if (smu_v13_0_10_is_mode2_default(reset_ctl) &&
+		amdgpu_asic_reset_method(adev) == AMD_RESET_METHOD_MODE2) {
+		for_each_handler(i, handler, reset_ctl)	{
+			if (handler->reset_method == AMD_RESET_METHOD_MODE2)
+				return handler;
+		}
+	}
+
+	return NULL;
+}
+
+static int smu_v13_0_10_mode2_suspend_ip(struct amdgpu_device *adev)
+{
+	int r, i;
+
+	amdgpu_device_set_pg_state(adev, AMD_PG_STATE_UNGATE);
+	amdgpu_device_set_cg_state(adev, AMD_CG_STATE_UNGATE);
+
+	for (i = adev->num_ip_blocks - 1; i >= 0; i--) {
+		if (!(adev->ip_blocks[i].version->type ==
+			      AMD_IP_BLOCK_TYPE_GFX ||
+		      adev->ip_blocks[i].version->type ==
+			      AMD_IP_BLOCK_TYPE_SDMA ||
+		      adev->ip_blocks[i].version->type ==
+			      AMD_IP_BLOCK_TYPE_MES))
+			continue;
+
+		r = amdgpu_ip_block_suspend(&adev->ip_blocks[i]);
+		if (r)
+			return r;
+	}
+
+	return 0;
+}
+
+static int
+smu_v13_0_10_mode2_prepare_hwcontext(struct amdgpu_reset_control *reset_ctl,
+				  struct amdgpu_reset_context *reset_context)
+{
+	int r = 0;
+	struct amdgpu_device *adev = (struct amdgpu_device *)reset_ctl->handle;
+
+	if (!amdgpu_sriov_vf(adev))
+		r = smu_v13_0_10_mode2_suspend_ip(adev);
+
+	return r;
+}
+
+static int smu_v13_0_10_mode2_reset(struct amdgpu_device *adev)
+{
+	return amdgpu_dpm_mode2_reset(adev);
+}
+
+static void smu_v13_0_10_async_reset(struct work_struct *work)
+{
+	struct amdgpu_reset_handler *handler;
+	struct amdgpu_reset_control *reset_ctl =
+		container_of(work, struct amdgpu_reset_control, reset_work);
+	struct amdgpu_device *adev = (struct amdgpu_device *)reset_ctl->handle;
+	int i;
+
+	for_each_handler(i, handler, reset_ctl)	{
+		if (handler->reset_method == reset_ctl->active_reset) {
+			dev_dbg(adev->dev, "Resetting device\n");
+			handler->do_reset(adev);
+			break;
+		}
+	}
+}
+static int
+smu_v13_0_10_mode2_perform_reset(struct amdgpu_reset_control *reset_ctl,
+			      struct amdgpu_reset_context *reset_context)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)reset_ctl->handle;
+	int r;
+
+	r = smu_v13_0_10_mode2_reset(adev);
+	if (r) {
+		dev_err(adev->dev,
+			"ASIC reset failed with error, %d ", r);
+	}
+	return r;
+}
+
+static int smu_v13_0_10_mode2_restore_ip(struct amdgpu_device *adev)
+{
+	int i, r;
+	struct psp_context *psp = &adev->psp;
+	struct amdgpu_firmware_info *ucode;
+	struct amdgpu_firmware_info *ucode_list[2];
+	int ucode_count = 0;
+
+	for (i = 0; i < adev->firmware.max_ucodes; i++) {
+		ucode = &adev->firmware.ucode[i];
+
+		switch (ucode->ucode_id) {
+		case AMDGPU_UCODE_ID_IMU_I:
+		case AMDGPU_UCODE_ID_IMU_D:
+			ucode_list[ucode_count++] = ucode;
+			break;
+		default:
+			break;
+		}
+	}
+
+	r = psp_load_fw_list(psp, ucode_list, ucode_count);
+	if (r) {
+		dev_err(adev->dev, "IMU ucode load failed after mode2 reset\n");
+		return r;
+	}
+
+	r = psp_rlc_autoload_start(psp);
+	if (r) {
+		DRM_ERROR("Failed to start rlc autoload after mode2 reset\n");
+		return r;
+	}
+
+	amdgpu_dpm_enable_gfx_features(adev);
+
+	for (i = 0; i < adev->num_ip_blocks; i++) {
+		if (!(adev->ip_blocks[i].version->type ==
+			      AMD_IP_BLOCK_TYPE_GFX ||
+		      adev->ip_blocks[i].version->type ==
+			      AMD_IP_BLOCK_TYPE_MES ||
+		      adev->ip_blocks[i].version->type ==
+			      AMD_IP_BLOCK_TYPE_SDMA))
+			continue;
+		r = amdgpu_ip_block_resume(&adev->ip_blocks[i]);
+		if (r)
+			return r;
+	}
+
+	for (i = 0; i < adev->num_ip_blocks; i++) {
+		if (!(adev->ip_blocks[i].version->type ==
+			      AMD_IP_BLOCK_TYPE_GFX ||
+		      adev->ip_blocks[i].version->type ==
+			      AMD_IP_BLOCK_TYPE_MES ||
+		      adev->ip_blocks[i].version->type ==
+			      AMD_IP_BLOCK_TYPE_SDMA))
+			continue;
+
+		if (adev->ip_blocks[i].version->funcs->late_init) {
+			r = adev->ip_blocks[i].version->funcs->late_init(
+				&adev->ip_blocks[i]);
+			if (r) {
+				dev_err(adev->dev,
+					"late_init of IP block <%s> failed %d after reset\n",
+					adev->ip_blocks[i].version->funcs->name,
+					r);
+				return r;
+			}
+		}
+		adev->ip_blocks[i].status.late_initialized = true;
+	}
+
+	amdgpu_device_set_cg_state(adev, AMD_CG_STATE_GATE);
+	amdgpu_device_set_pg_state(adev, AMD_PG_STATE_GATE);
+
+	return r;
+}
+
+static int
+smu_v13_0_10_mode2_restore_hwcontext(struct amdgpu_reset_control *reset_ctl,
+				  struct amdgpu_reset_context *reset_context)
+{
+	int r;
+	struct amdgpu_device *tmp_adev = (struct amdgpu_device *)reset_ctl->handle;
+
+	amdgpu_set_init_level(tmp_adev, AMDGPU_INIT_LEVEL_RESET_RECOVERY);
+	dev_info(tmp_adev->dev,
+			"GPU reset succeeded, trying to resume\n");
+	r = smu_v13_0_10_mode2_restore_ip(tmp_adev);
+	if (r)
+		goto end;
+
+	amdgpu_register_gpu_instance(tmp_adev);
+
+	/* Resume RAS */
+	amdgpu_ras_resume(tmp_adev);
+
+	amdgpu_irq_gpu_reset_resume_helper(tmp_adev);
+
+	amdgpu_set_init_level(tmp_adev, AMDGPU_INIT_LEVEL_DEFAULT);
+	r = amdgpu_ib_ring_tests(tmp_adev);
+	if (r) {
+		dev_err(tmp_adev->dev,
+			"ib ring test failed (%d).\n", r);
+		r = -EAGAIN;
+		goto end;
+	}
+
+end:
+	if (r)
+		return -EAGAIN;
+	else
+		return r;
+}
+
+static struct amdgpu_reset_handler smu_v13_0_10_mode2_handler = {
+	.reset_method		= AMD_RESET_METHOD_MODE2,
+	.prepare_env		= NULL,
+	.prepare_hwcontext	= smu_v13_0_10_mode2_prepare_hwcontext,
+	.perform_reset		= smu_v13_0_10_mode2_perform_reset,
+	.restore_hwcontext	= smu_v13_0_10_mode2_restore_hwcontext,
+	.restore_env		= NULL,
+	.do_reset		= smu_v13_0_10_mode2_reset,
+};
+
+static struct amdgpu_reset_handler
+	*smu_v13_0_10_rst_handlers[AMDGPU_RESET_MAX_HANDLERS] = {
+		&smu_v13_0_10_mode2_handler,
+	};
+
+int smu_v13_0_10_reset_init(struct amdgpu_device *adev)
+{
+	struct amdgpu_reset_control *reset_ctl;
+
+	reset_ctl = kzalloc(sizeof(*reset_ctl), GFP_KERNEL);
+	if (!reset_ctl)
+		return -ENOMEM;
+
+	reset_ctl->handle = adev;
+	reset_ctl->async_reset = smu_v13_0_10_async_reset;
+	reset_ctl->active_reset = AMD_RESET_METHOD_NONE;
+	reset_ctl->get_reset_handler = smu_v13_0_10_get_reset_handler;
+
+	INIT_WORK(&reset_ctl->reset_work, reset_ctl->async_reset);
+	/* Only mode2 is handled through reset control now */
+	reset_ctl->reset_handlers = &smu_v13_0_10_rst_handlers;
+
+	adev->reset_cntl = reset_ctl;
+
+	return 0;
+}
+
+int smu_v13_0_10_reset_fini(struct amdgpu_device *adev)
+{
+	kfree(adev->reset_cntl);
+	adev->reset_cntl = NULL;
+	return 0;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/smu_v13_0_10.h b/drivers/gpu/drm/amd/amdgpu/smu_v13_0_10.h
new file mode 100644
index 000000000000..e0cb72a0eec6
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/smu_v13_0_10.h
@@ -0,0 +1,32 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __SMU_V13_0_10_H__
+#define __SMU_V13_0_10_H__
+
+#include "amdgpu.h"
+
+int smu_v13_0_10_reset_init(struct amdgpu_device *adev);
+int smu_v13_0_10_reset_fini(struct amdgpu_device *adev);
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/smuio_v11_0.c b/drivers/gpu/drm/amd/amdgpu/smuio_v11_0.c
index b6f1322f908c..acdc40f99ab3 100644
--- a/drivers/gpu/drm/amd/amdgpu/smuio_v11_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/smuio_v11_0.c
@@ -59,7 +59,7 @@ static void smuio_v11_0_update_rom_clock_gating(struct amdgpu_device *adev, bool
 		WREG32_SOC15(SMUIO, 0, mmCGTT_ROM_CLK_CTRL0, data);
 }
 
-static void smuio_v11_0_get_clock_gating_state(struct amdgpu_device *adev, u32 *flags)
+static void smuio_v11_0_get_clock_gating_state(struct amdgpu_device *adev, u64 *flags)
 {
 	u32 data;
 
diff --git a/drivers/gpu/drm/amd/amdgpu/smuio_v11_0_6.c b/drivers/gpu/drm/amd/amdgpu/smuio_v11_0_6.c
index 3a18dbb55c32..2afeb8b37f62 100644
--- a/drivers/gpu/drm/amd/amdgpu/smuio_v11_0_6.c
+++ b/drivers/gpu/drm/amd/amdgpu/smuio_v11_0_6.c
@@ -56,7 +56,7 @@ static void smuio_v11_0_6_update_rom_clock_gating(struct amdgpu_device *adev, bo
 		WREG32_SOC15(SMUIO, 0, mmCGTT_ROM_CLK_CTRL0, data);
 }
 
-static void smuio_v11_0_6_get_clock_gating_state(struct amdgpu_device *adev, u32 *flags)
+static void smuio_v11_0_6_get_clock_gating_state(struct amdgpu_device *adev, u64 *flags)
 {
 	u32 data;
 
diff --git a/drivers/gpu/drm/amd/amdgpu/smuio_v13_0.c b/drivers/gpu/drm/amd/amdgpu/smuio_v13_0.c
index 39b7c206770f..bf8b8e5ddf5d 100644
--- a/drivers/gpu/drm/amd/amdgpu/smuio_v13_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/smuio_v13_0.c
@@ -58,7 +58,7 @@ static void smuio_v13_0_update_rom_clock_gating(struct amdgpu_device *adev, bool
 		WREG32_SOC15(SMUIO, 0, regCGTT_ROM_CLK_CTRL0, data);
 }
 
-static void smuio_v13_0_get_clock_gating_state(struct amdgpu_device *adev, u32 *flags)
+static void smuio_v13_0_get_clock_gating_state(struct amdgpu_device *adev, u64 *flags)
 {
 	u32 data;
 
@@ -128,6 +128,27 @@ static bool smuio_v13_0_is_host_gpu_xgmi_supported(struct amdgpu_device *adev)
 	return data ? true : false;
 }
 
+static enum amdgpu_pkg_type smuio_v13_0_get_pkg_type(struct amdgpu_device *adev)
+{
+	enum amdgpu_pkg_type pkg_type;
+	u32 data;
+
+	data = RREG32_SOC15(SMUIO, 0, regSMUIO_MCM_CONFIG);
+	data = REG_GET_FIELD(data, SMUIO_MCM_CONFIG, TOPOLOGY_ID);
+
+	switch (data) {
+	case 0x4:
+	case 0xC:
+		pkg_type = AMDGPU_PKG_TYPE_CEM;
+		break;
+	default:
+		pkg_type = AMDGPU_PKG_TYPE_OAM;
+		break;
+	}
+
+	return pkg_type;
+}
+
 const struct amdgpu_smuio_funcs smuio_v13_0_funcs = {
 	.get_rom_index_offset = smuio_v13_0_get_rom_index_offset,
 	.get_rom_data_offset = smuio_v13_0_get_rom_data_offset,
@@ -136,4 +157,5 @@ const struct amdgpu_smuio_funcs smuio_v13_0_funcs = {
 	.is_host_gpu_xgmi_supported = smuio_v13_0_is_host_gpu_xgmi_supported,
 	.update_rom_clock_gating = smuio_v13_0_update_rom_clock_gating,
 	.get_clock_gating_state = smuio_v13_0_get_clock_gating_state,
+	.get_pkg_type = smuio_v13_0_get_pkg_type,
 };
diff --git a/drivers/gpu/drm/amd/amdgpu/smuio_v13_0_3.c b/drivers/gpu/drm/amd/amdgpu/smuio_v13_0_3.c
new file mode 100644
index 000000000000..5461b5289793
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/smuio_v13_0_3.c
@@ -0,0 +1,109 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include "amdgpu.h"
+#include "smuio_v13_0_3.h"
+#include "soc15_common.h"
+#include "smuio/smuio_13_0_3_offset.h"
+#include "smuio/smuio_13_0_3_sh_mask.h"
+
+#define PKG_TYPE_MASK		0x00000003L
+
+/**
+ * smuio_v13_0_3_get_die_id - query die id from FCH.
+ *
+ * @adev: amdgpu device pointer
+ *
+ * Returns die id
+ */
+static u32 smuio_v13_0_3_get_die_id(struct amdgpu_device *adev)
+{
+	u32 data, die_id;
+
+	data = RREG32_SOC15(SMUIO, 0, regSMUIO_MCM_CONFIG);
+	die_id = REG_GET_FIELD(data, SMUIO_MCM_CONFIG, DIE_ID);
+
+	return die_id;
+}
+
+/**
+ * smuio_v13_0_3_get_socket_id - query socket id from FCH
+ *
+ * @adev: amdgpu device pointer
+ *
+ * Returns socket id
+ */
+static u32 smuio_v13_0_3_get_socket_id(struct amdgpu_device *adev)
+{
+	u32 data, socket_id;
+
+	data = RREG32_SOC15(SMUIO, 0, regSMUIO_MCM_CONFIG);
+	socket_id = REG_GET_FIELD(data, SMUIO_MCM_CONFIG, SOCKET_ID);
+
+	return socket_id;
+}
+
+/**
+ * smuio_v13_0_3_get_pkg_type - query package type set by MP1/bootcode
+ *
+ * @adev: amdgpu device pointer
+ *
+ * Returns package type
+ */
+
+static enum amdgpu_pkg_type smuio_v13_0_3_get_pkg_type(struct amdgpu_device *adev)
+{
+	enum amdgpu_pkg_type pkg_type;
+	u32 data;
+
+	data = RREG32_SOC15(SMUIO, 0, regSMUIO_MCM_CONFIG);
+	data = REG_GET_FIELD(data, SMUIO_MCM_CONFIG, PKG_TYPE);
+	/* pkg_type[4:0]
+	 *
+	 * bit 1 == 1 APU form factor
+	 *
+	 * b0100 - b1111 - Reserved
+	 */
+	switch (data & PKG_TYPE_MASK) {
+	case 0x0:
+		pkg_type = AMDGPU_PKG_TYPE_CEM;
+		break;
+	case 0x1:
+		pkg_type = AMDGPU_PKG_TYPE_OAM;
+		break;
+	case 0x2:
+		pkg_type = AMDGPU_PKG_TYPE_APU;
+		break;
+	default:
+		pkg_type = AMDGPU_PKG_TYPE_UNKNOWN;
+		break;
+	}
+
+	return pkg_type;
+}
+
+
+const struct amdgpu_smuio_funcs smuio_v13_0_3_funcs = {
+	.get_die_id = smuio_v13_0_3_get_die_id,
+	.get_socket_id = smuio_v13_0_3_get_socket_id,
+	.get_pkg_type = smuio_v13_0_3_get_pkg_type,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/smuio_v13_0_3.h b/drivers/gpu/drm/amd/amdgpu/smuio_v13_0_3.h
new file mode 100644
index 000000000000..795f66c5a58b
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/smuio_v13_0_3.h
@@ -0,0 +1,30 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __SMUIO_V13_0_3_H__
+#define __SMUIO_V13_0_3_H__
+
+#include "soc15_common.h"
+
+extern const struct amdgpu_smuio_funcs smuio_v13_0_3_funcs;
+
+#endif /* __SMUIO_V13_0_3_H__ */
diff --git a/drivers/gpu/drm/amd/amdgpu/smuio_v13_0_6.c b/drivers/gpu/drm/amd/amdgpu/smuio_v13_0_6.c
new file mode 100644
index 000000000000..de998e328b08
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/smuio_v13_0_6.c
@@ -0,0 +1,41 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include "amdgpu.h"
+#include "smuio_v13_0_6.h"
+#include "smuio/smuio_13_0_6_offset.h"
+#include "smuio/smuio_13_0_6_sh_mask.h"
+
+static u32 smuio_v13_0_6_get_rom_index_offset(struct amdgpu_device *adev)
+{
+	return SOC15_REG_OFFSET(SMUIO, 0, regROM_INDEX);
+}
+
+static u32 smuio_v13_0_6_get_rom_data_offset(struct amdgpu_device *adev)
+{
+	return SOC15_REG_OFFSET(SMUIO, 0, regROM_DATA);
+}
+
+const struct amdgpu_smuio_funcs smuio_v13_0_6_funcs = {
+	.get_rom_index_offset = smuio_v13_0_6_get_rom_index_offset,
+	.get_rom_data_offset = smuio_v13_0_6_get_rom_data_offset,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/smuio_v13_0_6.h b/drivers/gpu/drm/amd/amdgpu/smuio_v13_0_6.h
new file mode 100644
index 000000000000..c75621de5ab5
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/smuio_v13_0_6.h
@@ -0,0 +1,30 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __SMUIO_V13_0_6_H__
+#define __SMUIO_V13_0_6_H__
+
+#include "soc15_common.h"
+
+extern const struct amdgpu_smuio_funcs smuio_v13_0_6_funcs;
+
+#endif /* __SMUIO_V13_0_6_H__ */
diff --git a/drivers/gpu/drm/amd/amdgpu/smuio_v14_0_2.c b/drivers/gpu/drm/amd/amdgpu/smuio_v14_0_2.c
new file mode 100644
index 000000000000..2a51a70d4846
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/smuio_v14_0_2.c
@@ -0,0 +1,62 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include "amdgpu.h"
+#include "smuio_v14_0_2.h"
+#include "smuio/smuio_14_0_2_offset.h"
+#include "smuio/smuio_14_0_2_sh_mask.h"
+#include <linux/preempt.h>
+
+static u32 smuio_v14_0_2_get_rom_index_offset(struct amdgpu_device *adev)
+{
+	return SOC15_REG_OFFSET(SMUIO, 0, regROM_INDEX);
+}
+
+static u32 smuio_v14_0_2_get_rom_data_offset(struct amdgpu_device *adev)
+{
+	return SOC15_REG_OFFSET(SMUIO, 0, regROM_DATA);
+}
+
+static u64 smuio_v14_0_2_get_gpu_clock_counter(struct amdgpu_device *adev)
+{
+	u64 clock;
+	u64 clock_counter_lo, clock_counter_hi_pre, clock_counter_hi_after;
+
+	preempt_disable();
+	clock_counter_hi_pre = (u64)RREG32_SOC15(SMUIO, 0, regGOLDEN_TSC_COUNT_UPPER);
+	clock_counter_lo = (u64)RREG32_SOC15(SMUIO, 0, regGOLDEN_TSC_COUNT_LOWER);
+	/* the clock counter may be udpated during polling the counters */
+	clock_counter_hi_after = (u64)RREG32_SOC15(SMUIO, 0, regGOLDEN_TSC_COUNT_UPPER);
+	if (clock_counter_hi_pre != clock_counter_hi_after)
+		clock_counter_lo = (u64)RREG32_SOC15(SMUIO, 0, regGOLDEN_TSC_COUNT_LOWER);
+	preempt_enable();
+
+	clock = clock_counter_lo | (clock_counter_hi_after << 32ULL);
+
+	return clock;
+}
+
+const struct amdgpu_smuio_funcs smuio_v14_0_2_funcs = {
+	.get_rom_index_offset = smuio_v14_0_2_get_rom_index_offset,
+	.get_rom_data_offset = smuio_v14_0_2_get_rom_data_offset,
+	.get_gpu_clock_counter = smuio_v14_0_2_get_gpu_clock_counter,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/smuio_v14_0_2.h b/drivers/gpu/drm/amd/amdgpu/smuio_v14_0_2.h
new file mode 100644
index 000000000000..6e617f832d90
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/smuio_v14_0_2.h
@@ -0,0 +1,30 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __SMUIO_V14_0_2_H__
+#define __SMUIO_V14_0_2_H__
+
+#include "soc15_common.h"
+
+extern const struct amdgpu_smuio_funcs smuio_v14_0_2_funcs;
+
+#endif /* __SMUIO_V14_0_2_H__ */
diff --git a/drivers/gpu/drm/amd/amdgpu/smuio_v9_0.c b/drivers/gpu/drm/amd/amdgpu/smuio_v9_0.c
index 8417890af227..c04fdd2d5b38 100644
--- a/drivers/gpu/drm/amd/amdgpu/smuio_v9_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/smuio_v9_0.c
@@ -56,11 +56,11 @@ static void smuio_v9_0_update_rom_clock_gating(struct amdgpu_device *adev, bool
 		WREG32_SOC15(SMUIO, 0, mmCGTT_ROM_CLK_CTRL0, data);
 }
 
-static void smuio_v9_0_get_clock_gating_state(struct amdgpu_device *adev, u32 *flags)
+static void smuio_v9_0_get_clock_gating_state(struct amdgpu_device *adev, u64 *flags)
 {
 	u32 data;
 
-	/* CGTT_ROM_CLK_CTRL0 is not availabe for APUs */
+	/* CGTT_ROM_CLK_CTRL0 is not available for APUs */
 	if (adev->flags & AMD_IS_APU)
 		return;
 
diff --git a/drivers/gpu/drm/amd/amdgpu/soc15.c b/drivers/gpu/drm/amd/amdgpu/soc15.c
index de9b55383e9f..42f5d9c0e3af 100644
--- a/drivers/gpu/drm/amd/amdgpu/soc15.c
+++ b/drivers/gpu/drm/amd/amdgpu/soc15.c
@@ -28,7 +28,6 @@
 #include <drm/amdgpu_drm.h>
 
 #include "amdgpu.h"
-#include "amdgpu_atombios.h"
 #include "amdgpu_ih.h"
 #include "amdgpu_uvd.h"
 #include "amdgpu_vce.h"
@@ -90,8 +89,8 @@ static const struct amd_ip_funcs soc15_common_ip_funcs;
 /* Vega, Raven, Arcturus */
 static const struct amdgpu_video_codec_info vega_video_codecs_encode_array[] =
 {
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 2304, 0)},
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC, 4096, 2304, 0)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4096, 0)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC, 4096, 4096, 0)},
 };
 
 static const struct amdgpu_video_codecs vega_video_codecs_encode =
@@ -103,12 +102,11 @@ static const struct amdgpu_video_codecs vega_video_codecs_encode =
 /* Vega */
 static const struct amdgpu_video_codec_info vega_video_codecs_decode_array[] =
 {
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG2, 4096, 4906, 3)},
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4, 4096, 4906, 5)},
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4906, 52)},
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VC1, 4096, 4906, 4)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG2, 1920, 1088, 3)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4, 1920, 1088, 5)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4096, 52)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VC1, 1920, 1088, 4)},
 	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC, 4096, 4096, 186)},
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG, 4096, 4096, 0)},
 };
 
 static const struct amdgpu_video_codecs vega_video_codecs_decode =
@@ -120,12 +118,12 @@ static const struct amdgpu_video_codecs vega_video_codecs_decode =
 /* Raven */
 static const struct amdgpu_video_codec_info rv_video_codecs_decode_array[] =
 {
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG2, 4096, 4906, 3)},
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4, 4096, 4906, 5)},
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4906, 52)},
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VC1, 4096, 4906, 4)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG2, 1920, 1088, 3)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4, 1920, 1088, 5)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4096, 52)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VC1, 1920, 1088, 4)},
 	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC, 4096, 4096, 186)},
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG, 4096, 4096, 0)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG, 8192, 8192, 0)},
 	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VP9, 4096, 4096, 0)},
 };
 
@@ -138,12 +136,12 @@ static const struct amdgpu_video_codecs rv_video_codecs_decode =
 /* Renoir, Arcturus */
 static const struct amdgpu_video_codec_info rn_video_codecs_decode_array[] =
 {
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG2, 4096, 4906, 3)},
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4, 4096, 4906, 5)},
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4906, 52)},
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VC1, 4096, 4906, 4)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG2, 1920, 1088, 3)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4, 1920, 1088, 5)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4096, 52)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VC1, 1920, 1088, 4)},
 	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC, 8192, 4352, 186)},
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG, 4096, 4096, 0)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG, 16384, 16384, 0)},
 	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VP9, 8192, 4352, 0)},
 };
 
@@ -153,11 +151,47 @@ static const struct amdgpu_video_codecs rn_video_codecs_decode =
 	.codec_array = rn_video_codecs_decode_array,
 };
 
+static const struct amdgpu_video_codec_info vcn_4_0_3_video_codecs_decode_array[] = {
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4096, 52)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC, 8192, 4352, 186)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG, 16384, 16384, 0)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VP9, 8192, 4352, 0)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_AV1, 8192, 4352, 0)},
+};
+
+static const struct amdgpu_video_codecs vcn_4_0_3_video_codecs_decode = {
+	.codec_count = ARRAY_SIZE(vcn_4_0_3_video_codecs_decode_array),
+	.codec_array = vcn_4_0_3_video_codecs_decode_array,
+};
+
+static const struct amdgpu_video_codecs vcn_4_0_3_video_codecs_encode = {
+	.codec_count = 0,
+	.codec_array = NULL,
+};
+
+static const struct amdgpu_video_codecs vcn_5_0_1_video_codecs_encode_vcn0 = {
+	.codec_count = 0,
+	.codec_array = NULL,
+};
+
+static const struct amdgpu_video_codec_info vcn_5_0_1_video_codecs_decode_array_vcn0[] = {
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4096, 52)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC, 8192, 4352, 186)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG, 16384, 16384, 0)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VP9, 8192, 4352, 0)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_AV1, 8192, 4352, 0)},
+};
+
+static const struct amdgpu_video_codecs vcn_5_0_1_video_codecs_decode_vcn0 = {
+	.codec_count = ARRAY_SIZE(vcn_5_0_1_video_codecs_decode_array_vcn0),
+	.codec_array = vcn_5_0_1_video_codecs_decode_array_vcn0,
+};
+
 static int soc15_query_video_codecs(struct amdgpu_device *adev, bool encode,
 				    const struct amdgpu_video_codecs **codecs)
 {
-	if (adev->ip_versions[VCE_HWIP][0]) {
-		switch (adev->ip_versions[VCE_HWIP][0]) {
+	if (amdgpu_ip_version(adev, VCE_HWIP, 0)) {
+		switch (amdgpu_ip_version(adev, VCE_HWIP, 0)) {
 		case IP_VERSION(4, 0, 0):
 		case IP_VERSION(4, 1, 0):
 			if (encode)
@@ -169,7 +203,7 @@ static int soc15_query_video_codecs(struct amdgpu_device *adev, bool encode,
 			return -EINVAL;
 		}
 	} else {
-		switch (adev->ip_versions[UVD_HWIP][0]) {
+		switch (amdgpu_ip_version(adev, UVD_HWIP, 0)) {
 		case IP_VERSION(1, 0, 0):
 		case IP_VERSION(1, 0, 1):
 			if (encode)
@@ -185,53 +219,24 @@ static int soc15_query_video_codecs(struct amdgpu_device *adev, bool encode,
 			else
 				*codecs = &rn_video_codecs_decode;
 			return 0;
+		case IP_VERSION(4, 0, 3):
+			if (encode)
+				*codecs = &vcn_4_0_3_video_codecs_encode;
+			else
+				*codecs = &vcn_4_0_3_video_codecs_decode;
+			return 0;
+		case IP_VERSION(5, 0, 1):
+			if (encode)
+				*codecs = &vcn_5_0_1_video_codecs_encode_vcn0;
+			else
+				*codecs = &vcn_5_0_1_video_codecs_decode_vcn0;
+			return 0;
 		default:
 			return -EINVAL;
 		}
 	}
 }
 
-/*
- * Indirect registers accessor
- */
-static u32 soc15_pcie_rreg(struct amdgpu_device *adev, u32 reg)
-{
-	unsigned long address, data;
-	address = adev->nbio.funcs->get_pcie_index_offset(adev);
-	data = adev->nbio.funcs->get_pcie_data_offset(adev);
-
-	return amdgpu_device_indirect_rreg(adev, address, data, reg);
-}
-
-static void soc15_pcie_wreg(struct amdgpu_device *adev, u32 reg, u32 v)
-{
-	unsigned long address, data;
-
-	address = adev->nbio.funcs->get_pcie_index_offset(adev);
-	data = adev->nbio.funcs->get_pcie_data_offset(adev);
-
-	amdgpu_device_indirect_wreg(adev, address, data, reg, v);
-}
-
-static u64 soc15_pcie_rreg64(struct amdgpu_device *adev, u32 reg)
-{
-	unsigned long address, data;
-	address = adev->nbio.funcs->get_pcie_index_offset(adev);
-	data = adev->nbio.funcs->get_pcie_data_offset(adev);
-
-	return amdgpu_device_indirect_rreg64(adev, address, data, reg);
-}
-
-static void soc15_pcie_wreg64(struct amdgpu_device *adev, u32 reg, u64 v)
-{
-	unsigned long address, data;
-
-	address = adev->nbio.funcs->get_pcie_index_offset(adev);
-	data = adev->nbio.funcs->get_pcie_data_offset(adev);
-
-	amdgpu_device_indirect_wreg64(adev, address, data, reg, v);
-}
-
 static u32 soc15_uvd_ctx_rreg(struct amdgpu_device *adev, u32 reg)
 {
 	unsigned long flags, address, data;
@@ -341,11 +346,14 @@ static u32 soc15_get_xclk(struct amdgpu_device *adev)
 {
 	u32 reference_clock = adev->clock.spll.reference_freq;
 
-	if (adev->ip_versions[MP1_HWIP][0] == IP_VERSION(12, 0, 0) ||
-	    adev->ip_versions[MP1_HWIP][0] == IP_VERSION(12, 0, 1))
+	if (amdgpu_ip_version(adev, MP1_HWIP, 0) == IP_VERSION(12, 0, 0) ||
+	    amdgpu_ip_version(adev, MP1_HWIP, 0) == IP_VERSION(12, 0, 1) ||
+	    amdgpu_ip_version(adev, MP1_HWIP, 0) == IP_VERSION(13, 0, 6) ||
+	    amdgpu_ip_version(adev, MP1_HWIP, 0) == IP_VERSION(13, 0, 12) ||
+	    amdgpu_ip_version(adev, MP1_HWIP, 0) == IP_VERSION(13, 0, 14))
 		return 10000;
-	if (adev->ip_versions[MP1_HWIP][0] == IP_VERSION(10, 0, 0) ||
-	    adev->ip_versions[MP1_HWIP][0] == IP_VERSION(10, 0, 1))
+	if (amdgpu_ip_version(adev, MP1_HWIP, 0) == IP_VERSION(10, 0, 0) ||
+	    amdgpu_ip_version(adev, MP1_HWIP, 0) == IP_VERSION(10, 0, 1))
 		return reference_clock / 4;
 
 	return reference_clock;
@@ -353,7 +361,7 @@ static u32 soc15_get_xclk(struct amdgpu_device *adev)
 
 
 void soc15_grbm_select(struct amdgpu_device *adev,
-		     u32 me, u32 pipe, u32 queue, u32 vmid)
+		     u32 me, u32 pipe, u32 queue, u32 vmid, int xcc_id)
 {
 	u32 grbm_gfx_cntl = 0;
 	grbm_gfx_cntl = REG_SET_FIELD(grbm_gfx_cntl, GRBM_GFX_CNTL, PIPEID, pipe);
@@ -361,12 +369,7 @@ void soc15_grbm_select(struct amdgpu_device *adev,
 	grbm_gfx_cntl = REG_SET_FIELD(grbm_gfx_cntl, GRBM_GFX_CNTL, VMID, vmid);
 	grbm_gfx_cntl = REG_SET_FIELD(grbm_gfx_cntl, GRBM_GFX_CNTL, QUEUEID, queue);
 
-	WREG32_SOC15_RLC_SHADOW(GC, 0, mmGRBM_GFX_CNTL, grbm_gfx_cntl);
-}
-
-static void soc15_vga_set_state(struct amdgpu_device *adev, bool state)
-{
-	/* todo */
+	WREG32_SOC15_RLC_SHADOW(GC, xcc_id, mmGRBM_GFX_CNTL, grbm_gfx_cntl);
 }
 
 static bool soc15_read_disabled_bios(struct amdgpu_device *adev)
@@ -375,39 +378,6 @@ static bool soc15_read_disabled_bios(struct amdgpu_device *adev)
 	return false;
 }
 
-static bool soc15_read_bios_from_rom(struct amdgpu_device *adev,
-				     u8 *bios, u32 length_bytes)
-{
-	u32 *dw_ptr;
-	u32 i, length_dw;
-	uint32_t rom_index_offset;
-	uint32_t rom_data_offset;
-
-	if (bios == NULL)
-		return false;
-	if (length_bytes == 0)
-		return false;
-	/* APU vbios image is part of sbios image */
-	if (adev->flags & AMD_IS_APU)
-		return false;
-
-	dw_ptr = (u32 *)bios;
-	length_dw = ALIGN(length_bytes, 4) / 4;
-
-	rom_index_offset =
-		adev->smuio.funcs->get_rom_index_offset(adev);
-	rom_data_offset =
-		adev->smuio.funcs->get_rom_data_offset(adev);
-
-	/* set rom index to 0 */
-	WREG32(rom_index_offset, 0);
-	/* read out the rom data */
-	for (i = 0; i < length_dw; i++)
-		dw_ptr[i] = RREG32(rom_data_offset);
-
-	return true;
-}
-
 static struct soc15_allowed_register_entry soc15_allowed_read_registers[] = {
 	{ SOC15_REG_ENTRY(GC, 0, mmGRBM_STATUS)},
 	{ SOC15_REG_ENTRY(GC, 0, mmGRBM_STATUS2)},
@@ -438,12 +408,12 @@ static uint32_t soc15_read_indexed_register(struct amdgpu_device *adev, u32 se_n
 
 	mutex_lock(&adev->grbm_idx_mutex);
 	if (se_num != 0xffffffff || sh_num != 0xffffffff)
-		amdgpu_gfx_select_se_sh(adev, se_num, sh_num, 0xffffffff);
+		amdgpu_gfx_select_se_sh(adev, se_num, sh_num, 0xffffffff, 0);
 
 	val = RREG32(reg_offset);
 
 	if (se_num != 0xffffffff || sh_num != 0xffffffff)
-		amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+		amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
 	mutex_unlock(&adev->grbm_idx_mutex);
 	return val;
 }
@@ -472,8 +442,9 @@ static int soc15_read_register(struct amdgpu_device *adev, u32 se_num,
 	*value = 0;
 	for (i = 0; i < ARRAY_SIZE(soc15_allowed_read_registers); i++) {
 		en = &soc15_allowed_read_registers[i];
-		if (adev->reg_offset[en->hwip][en->inst] &&
-			reg_offset != (adev->reg_offset[en->hwip][en->inst][en->seg]
+		if (!adev->reg_offset[en->hwip][en->inst])
+			continue;
+		else if (reg_offset != (adev->reg_offset[en->hwip][en->inst][en->seg]
 					+ en->reg_offset))
 			continue;
 
@@ -555,7 +526,7 @@ static int soc15_asic_baco_reset(struct amdgpu_device *adev)
 static enum amd_reset_method
 soc15_asic_reset_method(struct amdgpu_device *adev)
 {
-	bool baco_reset = false;
+	int baco_reset = 0;
 	bool connected_to_cpu = false;
 	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
 
@@ -576,7 +547,7 @@ soc15_asic_reset_method(struct amdgpu_device *adev)
 		dev_warn(adev->dev, "Specified reset method:%d isn't supported, using AUTO instead.\n",
 				  amdgpu_reset_method);
 
-	switch (adev->ip_versions[MP1_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, MP1_HWIP, 0)) {
 	case IP_VERSION(10, 0, 0):
 	case IP_VERSION(10, 0, 1):
 	case IP_VERSION(12, 0, 0):
@@ -593,7 +564,7 @@ soc15_asic_reset_method(struct amdgpu_device *adev)
 			 */
 			if (ras && adev->ras_enabled &&
 			    adev->pm.fw_version <= 0x283400)
-				baco_reset = false;
+				baco_reset = 0;
 		} else {
 			baco_reset = amdgpu_dpm_is_baco_supported(adev);
 		}
@@ -606,6 +577,21 @@ soc15_asic_reset_method(struct amdgpu_device *adev)
 		if (connected_to_cpu)
 			return AMD_RESET_METHOD_MODE2;
 		break;
+	case IP_VERSION(13, 0, 6):
+	case IP_VERSION(13, 0, 14):
+	case IP_VERSION(13, 0, 12):
+		/* Use gpu_recovery param to target a reset method.
+		 * Enable triggering of GPU reset only if specified
+		 * by module parameter.
+		 */
+		if (adev->pcie_reset_ctx.in_link_reset)
+			return AMD_RESET_METHOD_LINK;
+		if (amdgpu_gpu_recovery == 4 || amdgpu_gpu_recovery == 5)
+			return AMD_RESET_METHOD_MODE2;
+		else if (!(adev->flags & AMD_IS_APU))
+			return AMD_RESET_METHOD_MODE1;
+		else
+			return AMD_RESET_METHOD_MODE2;
 	default:
 		break;
 	}
@@ -616,13 +602,36 @@ soc15_asic_reset_method(struct amdgpu_device *adev)
 		return AMD_RESET_METHOD_MODE1;
 }
 
+static bool soc15_need_reset_on_resume(struct amdgpu_device *adev)
+{
+	/* Will reset for the following suspend abort cases.
+	 * 1) S3 suspend aborted in the normal S3 suspend
+	 * 2) S3 suspend aborted in performing pm core test.
+	 */
+	if (adev->in_s3 && !pm_resume_via_firmware())
+		return true;
+	else
+		return false;
+}
+
 static int soc15_asic_reset(struct amdgpu_device *adev)
 {
 	/* original raven doesn't have full asic reset */
-	if ((adev->apu_flags & AMD_APU_IS_RAVEN) &&
-	    !(adev->apu_flags & AMD_APU_IS_RAVEN2))
+	/* On the latest Raven, the GPU reset can be performed
+	 * successfully. So now, temporarily enable it for the
+	 * S3 suspend abort case.
+	 */
+
+	if ((adev->apu_flags & AMD_APU_IS_PICASSO ||
+			!(adev->apu_flags & AMD_APU_IS_RAVEN)) &&
+			soc15_need_reset_on_resume(adev))
+		goto asic_reset;
+
+	if ((adev->apu_flags & AMD_APU_IS_RAVEN) ||
+			(adev->apu_flags & AMD_APU_IS_RAVEN2))
 		return 0;
 
+asic_reset:
 	switch (soc15_asic_reset_method(adev)) {
 	case AMD_RESET_METHOD_PCI:
 		dev_info(adev->dev, "PCI reset\n");
@@ -633,27 +642,30 @@ static int soc15_asic_reset(struct amdgpu_device *adev)
 	case AMD_RESET_METHOD_MODE2:
 		dev_info(adev->dev, "MODE2 reset\n");
 		return amdgpu_dpm_mode2_reset(adev);
+	case AMD_RESET_METHOD_LINK:
+		dev_info(adev->dev, "Link reset\n");
+		return amdgpu_device_link_reset(adev);
 	default:
 		dev_info(adev->dev, "MODE1 reset\n");
 		return amdgpu_device_mode1_reset(adev);
 	}
 }
 
-static bool soc15_supports_baco(struct amdgpu_device *adev)
+static int soc15_supports_baco(struct amdgpu_device *adev)
 {
-	switch (adev->ip_versions[MP1_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, MP1_HWIP, 0)) {
 	case IP_VERSION(9, 0, 0):
 	case IP_VERSION(11, 0, 2):
 		if (adev->asic_type == CHIP_VEGA20) {
 			if (adev->psp.sos.fw_version >= 0x80067)
 				return amdgpu_dpm_is_baco_supported(adev);
-			return false;
+			return 0;
 		} else {
 			return amdgpu_dpm_is_baco_supported(adev);
 		}
 		break;
 	default:
-		return false;
+		return 0;
 	}
 }
 
@@ -683,41 +695,15 @@ static int soc15_set_vce_clocks(struct amdgpu_device *adev, u32 evclk, u32 ecclk
 	return 0;
 }
 
-static void soc15_pcie_gen3_enable(struct amdgpu_device *adev)
-{
-	if (pci_is_root_bus(adev->pdev->bus))
-		return;
-
-	if (amdgpu_pcie_gen2 == 0)
-		return;
-
-	if (adev->flags & AMD_IS_APU)
-		return;
-
-	if (!(adev->pm.pcie_gen_mask & (CAIL_PCIE_LINK_SPEED_SUPPORT_GEN2 |
-					CAIL_PCIE_LINK_SPEED_SUPPORT_GEN3)))
-		return;
-
-	/* todo */
-}
-
 static void soc15_program_aspm(struct amdgpu_device *adev)
 {
-	if (!amdgpu_aspm)
+	if (!amdgpu_device_should_use_aspm(adev))
 		return;
 
-	if (!(adev->flags & AMD_IS_APU) &&
-	    (adev->nbio.funcs->program_aspm))
+	if (adev->nbio.funcs->program_aspm)
 		adev->nbio.funcs->program_aspm(adev);
 }
 
-static void soc15_enable_doorbell_aperture(struct amdgpu_device *adev,
-					   bool enable)
-{
-	adev->nbio.funcs->enable_doorbell_aperture(adev, enable);
-	adev->nbio.funcs->enable_doorbell_selfring_aperture(adev, enable);
-}
-
 const struct amdgpu_ip_block_version vega10_common_ip_block =
 {
 	.type = AMD_IP_BLOCK_TYPE_COMMON,
@@ -727,32 +713,14 @@ const struct amdgpu_ip_block_version vega10_common_ip_block =
 	.funcs = &soc15_common_ip_funcs,
 };
 
-static uint32_t soc15_get_rev_id(struct amdgpu_device *adev)
-{
-	return adev->nbio.funcs->get_rev_id(adev);
-}
-
 static void soc15_reg_base_init(struct amdgpu_device *adev)
 {
-	int r;
-
 	/* Set IP register base before any HW register access */
 	switch (adev->asic_type) {
 	case CHIP_VEGA10:
 	case CHIP_VEGA12:
 	case CHIP_RAVEN:
-		vega10_reg_base_init(adev);
-		break;
 	case CHIP_RENOIR:
-		/* It's safe to do ip discovery here for Renior,
-		 * it doesn't support SRIOV. */
-		if (amdgpu_discovery) {
-			r = amdgpu_discovery_reg_base_init(adev);
-			if (r == 0)
-				break;
-			DRM_WARN("failed to init reg base from ip discovery table, "
-				 "fallback to legacy init method\n");
-		}
 		vega10_reg_base_init(adev);
 		break;
 	case CHIP_VEGA20:
@@ -773,7 +741,6 @@ static void soc15_reg_base_init(struct amdgpu_device *adev)
 void soc15_set_virt_ops(struct amdgpu_device *adev)
 {
 	adev->virt.ops = &xgpu_ai_virt_ops;
-
 	/* init soc15 reg base early enough so we can
 	 * request request full access for sriov before
 	 * set_ip_blocks. */
@@ -886,6 +853,10 @@ static bool soc15_need_reset_on_init(struct amdgpu_device *adev)
 {
 	u32 sol_reg;
 
+	if (amdgpu_gmc_need_reset_on_init(adev))
+		return true;
+	if (amdgpu_psp_tos_reload_needed(adev))
+		return true;
 	/* Just return false for soc15 GPUs.  Reset does not seem to
 	 * be necessary.
 	 */
@@ -925,11 +896,10 @@ static void soc15_pre_asic_init(struct amdgpu_device *adev)
 static const struct amdgpu_asic_funcs soc15_asic_funcs =
 {
 	.read_disabled_bios = &soc15_read_disabled_bios,
-	.read_bios_from_rom = &soc15_read_bios_from_rom,
+	.read_bios_from_rom = &amdgpu_soc15_read_bios_from_rom,
 	.read_register = &soc15_read_register,
 	.reset = &soc15_asic_reset,
 	.reset_method = &soc15_asic_reset_method,
-	.set_vga_state = &soc15_vga_set_state,
 	.get_xclk = &soc15_get_xclk,
 	.set_uvd_clocks = &soc15_set_uvd_clocks,
 	.set_vce_clocks = &soc15_set_vce_clocks,
@@ -947,11 +917,10 @@ static const struct amdgpu_asic_funcs soc15_asic_funcs =
 static const struct amdgpu_asic_funcs vega20_asic_funcs =
 {
 	.read_disabled_bios = &soc15_read_disabled_bios,
-	.read_bios_from_rom = &soc15_read_bios_from_rom,
+	.read_bios_from_rom = &amdgpu_soc15_read_bios_from_rom,
 	.read_register = &soc15_read_register,
 	.reset = &soc15_asic_reset,
 	.reset_method = &soc15_asic_reset_method,
-	.set_vga_state = &soc15_vga_set_state,
 	.get_xclk = &soc15_get_xclk,
 	.set_uvd_clocks = &soc15_set_uvd_clocks,
 	.set_vce_clocks = &soc15_set_vce_clocks,
@@ -966,21 +935,43 @@ static const struct amdgpu_asic_funcs vega20_asic_funcs =
 	.query_video_codecs = &soc15_query_video_codecs,
 };
 
-static int soc15_common_early_init(void *handle)
+static const struct amdgpu_asic_funcs aqua_vanjaram_asic_funcs =
 {
-#define MMIO_REG_HOLE_OFFSET (0x80000 - PAGE_SIZE)
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	.read_disabled_bios = &soc15_read_disabled_bios,
+	.read_bios_from_rom = &amdgpu_soc15_read_bios_from_rom,
+	.read_register = &soc15_read_register,
+	.reset = &soc15_asic_reset,
+	.reset_method = &soc15_asic_reset_method,
+	.get_xclk = &soc15_get_xclk,
+	.set_uvd_clocks = &soc15_set_uvd_clocks,
+	.set_vce_clocks = &soc15_set_vce_clocks,
+	.get_config_memsize = &soc15_get_config_memsize,
+	.need_full_reset = &soc15_need_full_reset,
+	.init_doorbell_index = &aqua_vanjaram_doorbell_index_init,
+	.need_reset_on_init = &soc15_need_reset_on_init,
+	.get_pcie_replay_count = &amdgpu_nbio_get_pcie_replay_count,
+	.supports_baco = &soc15_supports_baco,
+	.pre_asic_init = &soc15_pre_asic_init,
+	.query_video_codecs = &soc15_query_video_codecs,
+	.encode_ext_smn_addressing = &aqua_vanjaram_encode_ext_smn_addressing,
+	.get_reg_state = &aqua_vanjaram_get_reg_state,
+};
 
-	if (!amdgpu_sriov_vf(adev)) {
-		adev->rmmio_remap.reg_offset = MMIO_REG_HOLE_OFFSET;
-		adev->rmmio_remap.bus_addr = adev->rmmio_base + MMIO_REG_HOLE_OFFSET;
-	}
+static int soc15_common_early_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	adev->nbio.funcs->set_reg_remap(adev);
 	adev->smc_rreg = NULL;
 	adev->smc_wreg = NULL;
-	adev->pcie_rreg = &soc15_pcie_rreg;
-	adev->pcie_wreg = &soc15_pcie_wreg;
-	adev->pcie_rreg64 = &soc15_pcie_rreg64;
-	adev->pcie_wreg64 = &soc15_pcie_wreg64;
+	adev->pcie_rreg = &amdgpu_device_indirect_rreg;
+	adev->pcie_wreg = &amdgpu_device_indirect_wreg;
+	adev->pcie_rreg_ext = &amdgpu_device_indirect_rreg_ext;
+	adev->pcie_wreg_ext = &amdgpu_device_indirect_wreg_ext;
+	adev->pcie_rreg64 = &amdgpu_device_indirect_rreg64;
+	adev->pcie_wreg64 = &amdgpu_device_indirect_wreg64;
+	adev->pcie_rreg64_ext = &amdgpu_device_indirect_rreg64_ext;
+	adev->pcie_wreg64_ext = &amdgpu_device_indirect_wreg64_ext;
 	adev->uvd_ctx_rreg = &soc15_uvd_ctx_rreg;
 	adev->uvd_ctx_wreg = &soc15_uvd_ctx_wreg;
 	adev->didt_rreg = &soc15_didt_rreg;
@@ -990,12 +981,12 @@ static int soc15_common_early_init(void *handle)
 	adev->se_cac_rreg = &soc15_se_cac_rreg;
 	adev->se_cac_wreg = &soc15_se_cac_wreg;
 
-	adev->rev_id = soc15_get_rev_id(adev);
+	adev->rev_id = amdgpu_device_get_rev_id(adev);
 	adev->external_rev_id = 0xFF;
 	/* TODO: split the GC and PG flags based on the relevant IP version for which
 	 * they are relevant.
 	 */
-	switch (adev->ip_versions[GC_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 	case IP_VERSION(9, 0, 1):
 		adev->asic_funcs = &soc15_asic_funcs;
 		adev->cg_flags = AMD_CG_SUPPORT_GFX_MGCG |
@@ -1114,8 +1105,11 @@ static int soc15_common_early_init(void *handle)
 				AMD_CG_SUPPORT_SDMA_LS |
 				AMD_CG_SUPPORT_VCN_MGCG;
 
+			/*
+			 * MMHUB PG needs to be disabled for Picasso for
+			 * stability reasons.
+			 */
 			adev->pg_flags = AMD_PG_SUPPORT_SDMA |
-				AMD_PG_SUPPORT_MMHUB |
 				AMD_PG_SUPPORT_VCN;
 		} else {
 			adev->cg_flags = AMD_CG_SUPPORT_GFX_MGCG |
@@ -1203,6 +1197,25 @@ static int soc15_common_early_init(void *handle)
 		adev->pg_flags = AMD_PG_SUPPORT_VCN_DPG;
 		adev->external_rev_id = adev->rev_id + 0x3c;
 		break;
+	case IP_VERSION(9, 4, 3):
+	case IP_VERSION(9, 4, 4):
+	case IP_VERSION(9, 5, 0):
+		adev->asic_funcs = &aqua_vanjaram_asic_funcs;
+		adev->cg_flags =
+			AMD_CG_SUPPORT_GFX_MGCG | AMD_CG_SUPPORT_GFX_CGCG |
+			AMD_CG_SUPPORT_GFX_CGLS | AMD_CG_SUPPORT_SDMA_MGCG |
+			AMD_CG_SUPPORT_GFX_FGCG | AMD_CG_SUPPORT_REPEATER_FGCG |
+			AMD_CG_SUPPORT_VCN_MGCG | AMD_CG_SUPPORT_JPEG_MGCG |
+			AMD_CG_SUPPORT_IH_CG;
+		adev->pg_flags =
+			AMD_PG_SUPPORT_VCN |
+			AMD_PG_SUPPORT_VCN_DPG |
+			AMD_PG_SUPPORT_JPEG;
+		/*TODO: need a new external_rev_id for GC 9.4.4? */
+		adev->external_rev_id = adev->rev_id + 0x46;
+		if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 5, 0))
+			adev->external_rev_id = adev->rev_id + 0x50;
+		break;
 	default:
 		/* FIXME: not supported yet */
 		return -EINVAL;
@@ -1216,69 +1229,63 @@ static int soc15_common_early_init(void *handle)
 	return 0;
 }
 
-static int soc15_common_late_init(void *handle)
+static int soc15_common_late_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-	int r = 0;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (amdgpu_sriov_vf(adev))
 		xgpu_ai_mailbox_get_irq(adev);
 
-	if (adev->nbio.ras_funcs &&
-	    adev->nbio.ras_funcs->ras_late_init)
-		r = adev->nbio.ras_funcs->ras_late_init(adev);
+	/* Enable selfring doorbell aperture late because doorbell BAR
+	 * aperture will change if resize BAR successfully in gmc sw_init.
+	 */
+	adev->nbio.funcs->enable_doorbell_selfring_aperture(adev, true);
 
-	return r;
+	return 0;
 }
 
-static int soc15_common_sw_init(void *handle)
+static int soc15_common_sw_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (amdgpu_sriov_vf(adev))
 		xgpu_ai_mailbox_add_irq_id(adev);
 
-	adev->df.funcs->sw_init(adev);
+	if (adev->df.funcs &&
+	    adev->df.funcs->sw_init)
+		adev->df.funcs->sw_init(adev);
 
 	return 0;
 }
 
-static int soc15_common_sw_fini(void *handle)
+static int soc15_common_sw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
-	if (adev->nbio.ras_funcs &&
-	    adev->nbio.ras_funcs->ras_fini)
-		adev->nbio.ras_funcs->ras_fini(adev);
-	adev->df.funcs->sw_fini(adev);
+	if (adev->df.funcs &&
+	    adev->df.funcs->sw_fini)
+		adev->df.funcs->sw_fini(adev);
 	return 0;
 }
 
-static void soc15_doorbell_range_init(struct amdgpu_device *adev)
+static void soc15_sdma_doorbell_range_init(struct amdgpu_device *adev)
 {
 	int i;
-	struct amdgpu_ring *ring;
 
-	/* sdma/ih doorbell range are programed by hypervisor */
+	/* sdma doorbell range is programed by hypervisor */
 	if (!amdgpu_sriov_vf(adev)) {
 		for (i = 0; i < adev->sdma.num_instances; i++) {
-			ring = &adev->sdma.instance[i].ring;
 			adev->nbio.funcs->sdma_doorbell_range(adev, i,
-				ring->use_doorbell, ring->doorbell_index,
+				true, adev->doorbell_index.sdma_engine[i] << 1,
 				adev->doorbell_index.sdma_doorbell_range);
 		}
-
-		adev->nbio.funcs->ih_doorbell_range(adev, adev->irq.ih.use_doorbell,
-						adev->irq.ih.doorbell_index);
 	}
 }
 
-static int soc15_common_hw_init(void *handle)
+static int soc15_common_hw_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
-	/* enable pcie gen2/3 link */
-	soc15_pcie_gen3_enable(adev);
 	/* enable aspm */
 	soc15_program_aspm(adev);
 	/* setup nbio registers */
@@ -1291,68 +1298,74 @@ static int soc15_common_hw_init(void *handle)
 		adev->nbio.funcs->remap_hdp_registers(adev);
 
 	/* enable the doorbell aperture */
-	soc15_enable_doorbell_aperture(adev, true);
+	adev->nbio.funcs->enable_doorbell_aperture(adev, true);
+
 	/* HW doorbell routing policy: doorbell writing not
 	 * in SDMA/IH/MM/ACV range will be routed to CP. So
-	 * we need to init SDMA/IH/MM/ACV doorbell range prior
-	 * to CP ip block init and ring test.
+	 * we need to init SDMA doorbell range prior
+	 * to CP ip block init and ring test.  IH already
+	 * happens before CP.
 	 */
-	soc15_doorbell_range_init(adev);
+	soc15_sdma_doorbell_range_init(adev);
 
 	return 0;
 }
 
-static int soc15_common_hw_fini(void *handle)
+static int soc15_common_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
+
+	/* Disable the doorbell aperture and selfring doorbell aperture
+	 * separately in hw_fini because soc15_enable_doorbell_aperture
+	 * has been removed and there is no need to delay disabling
+	 * selfring doorbell.
+	 */
+	adev->nbio.funcs->enable_doorbell_aperture(adev, false);
+	adev->nbio.funcs->enable_doorbell_selfring_aperture(adev, false);
 
-	/* disable the doorbell aperture */
-	soc15_enable_doorbell_aperture(adev, false);
 	if (amdgpu_sriov_vf(adev))
 		xgpu_ai_mailbox_put_irq(adev);
 
-	if (adev->nbio.ras_if &&
+	/*
+	 * For minimal init, late_init is not called, hence RAS irqs are not
+	 * enabled.
+	 */
+	if ((!amdgpu_sriov_vf(adev)) &&
+	    (adev->init_lvl->level != AMDGPU_INIT_LEVEL_MINIMAL_XGMI) &&
+	    adev->nbio.ras_if &&
 	    amdgpu_ras_is_supported(adev, adev->nbio.ras_if->block)) {
-		if (adev->nbio.ras_funcs &&
-		    adev->nbio.ras_funcs->init_ras_controller_interrupt)
+		if (adev->nbio.ras &&
+		    adev->nbio.ras->init_ras_controller_interrupt)
 			amdgpu_irq_put(adev, &adev->nbio.ras_controller_irq, 0);
-		if (adev->nbio.ras_funcs &&
-		    adev->nbio.ras_funcs->init_ras_err_event_athub_interrupt)
+		if (adev->nbio.ras &&
+		    adev->nbio.ras->init_ras_err_event_athub_interrupt)
 			amdgpu_irq_put(adev, &adev->nbio.ras_err_event_athub_irq, 0);
 	}
 
 	return 0;
 }
 
-static int soc15_common_suspend(void *handle)
+static int soc15_common_suspend(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	return soc15_common_hw_fini(adev);
+	return soc15_common_hw_fini(ip_block);
 }
 
-static int soc15_common_resume(void *handle)
+static int soc15_common_resume(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
-	return soc15_common_hw_init(adev);
+	if (soc15_need_reset_on_resume(adev)) {
+		dev_info(adev->dev, "S3 suspend abort case, let's reset ASIC.\n");
+		soc15_asic_reset(adev);
+	}
+	return soc15_common_hw_init(ip_block);
 }
 
-static bool soc15_common_is_idle(void *handle)
+static bool soc15_common_is_idle(struct amdgpu_ip_block *ip_block)
 {
 	return true;
 }
 
-static int soc15_common_wait_for_idle(void *handle)
-{
-	return 0;
-}
-
-static int soc15_common_soft_reset(void *handle)
-{
-	return 0;
-}
-
 static void soc15_update_drm_clock_gating(struct amdgpu_device *adev, bool enable)
 {
 	uint32_t def, data;
@@ -1397,15 +1410,15 @@ static void soc15_update_drm_light_sleep(struct amdgpu_device *adev, bool enable
 		WREG32(SOC15_REG_OFFSET(MP0, 0, mmMP0_MISC_LIGHT_SLEEP_CTRL), data);
 }
 
-static int soc15_common_set_clockgating_state(void *handle,
+static int soc15_common_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					    enum amd_clockgating_state state)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (amdgpu_sriov_vf(adev))
 		return 0;
 
-	switch (adev->ip_versions[NBIO_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, NBIO_HWIP, 0)) {
 	case IP_VERSION(6, 1, 0):
 	case IP_VERSION(6, 2, 0):
 	case IP_VERSION(7, 4, 0):
@@ -1449,20 +1462,24 @@ static int soc15_common_set_clockgating_state(void *handle,
 	return 0;
 }
 
-static void soc15_common_get_clockgating_state(void *handle, u32 *flags)
+static void soc15_common_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int data;
 
 	if (amdgpu_sriov_vf(adev))
 		*flags = 0;
 
-	adev->nbio.funcs->get_clockgating_state(adev, flags);
-
-	adev->hdp.funcs->get_clock_gating_state(adev, flags);
+	if (adev->nbio.funcs && adev->nbio.funcs->get_clockgating_state)
+		adev->nbio.funcs->get_clockgating_state(adev, flags);
 
-	if (adev->ip_versions[MP0_HWIP][0] != IP_VERSION(13, 0, 2)) {
+	if (adev->hdp.funcs && adev->hdp.funcs->get_clock_gating_state)
+		adev->hdp.funcs->get_clock_gating_state(adev, flags);
 
+	if ((amdgpu_ip_version(adev, MP0_HWIP, 0) != IP_VERSION(13, 0, 2)) &&
+	    (amdgpu_ip_version(adev, MP0_HWIP, 0) != IP_VERSION(13, 0, 6)) &&
+	    (amdgpu_ip_version(adev, MP0_HWIP, 0) != IP_VERSION(13, 0, 12)) &&
+	    (amdgpu_ip_version(adev, MP0_HWIP, 0) != IP_VERSION(13, 0, 14))) {
 		/* AMD_CG_SUPPORT_DRM_MGCG */
 		data = RREG32(SOC15_REG_OFFSET(MP0, 0, mmMP0_MISC_CGTT_CTRL0));
 		if (!(data & 0x01000000))
@@ -1475,12 +1492,14 @@ static void soc15_common_get_clockgating_state(void *handle, u32 *flags)
 	}
 
 	/* AMD_CG_SUPPORT_ROM_MGCG */
-	adev->smuio.funcs->get_clock_gating_state(adev, flags);
+	if (adev->smuio.funcs && adev->smuio.funcs->get_clock_gating_state)
+		adev->smuio.funcs->get_clock_gating_state(adev, flags);
 
-	adev->df.funcs->get_clockgating_state(adev, flags);
+	if (adev->df.funcs && adev->df.funcs->get_clockgating_state)
+		adev->df.funcs->get_clockgating_state(adev, flags);
 }
 
-static int soc15_common_set_powergating_state(void *handle,
+static int soc15_common_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					    enum amd_powergating_state state)
 {
 	/* todo */
@@ -1498,8 +1517,6 @@ static const struct amd_ip_funcs soc15_common_ip_funcs = {
 	.suspend = soc15_common_suspend,
 	.resume = soc15_common_resume,
 	.is_idle = soc15_common_is_idle,
-	.wait_for_idle = soc15_common_wait_for_idle,
-	.soft_reset = soc15_common_soft_reset,
 	.set_clockgating_state = soc15_common_set_clockgating_state,
 	.set_powergating_state = soc15_common_set_powergating_state,
 	.get_clockgating_state= soc15_common_get_clockgating_state,
diff --git a/drivers/gpu/drm/amd/amdgpu/soc15.h b/drivers/gpu/drm/amd/amdgpu/soc15.h
index efc2a253e8db..c8ac11a9cdef 100644
--- a/drivers/gpu/drm/amd/amdgpu/soc15.h
+++ b/drivers/gpu/drm/amd/amdgpu/soc15.h
@@ -27,6 +27,7 @@
 #include "nbio_v6_1.h"
 #include "nbio_v7_0.h"
 #include "nbio_v7_4.h"
+#include "amdgpu_reg_state.h"
 
 extern const struct amdgpu_ip_block_version vega10_common_ip_block;
 
@@ -87,9 +88,15 @@ struct soc15_ras_field_entry {
 };
 
 #define SOC15_REG_ENTRY(ip, inst, reg)	ip##_HWIP, inst, reg##_BASE_IDX, reg
+#define SOC15_REG_ENTRY_STR(ip, inst, reg) \
+	{ ip##_HWIP, inst, reg##_BASE_IDX, reg, #reg }
 
 #define SOC15_REG_ENTRY_OFFSET(entry)	(adev->reg_offset[entry.hwip][entry.inst][entry.seg] + entry.reg_offset)
 
+/* Over ride the instance id */
+#define SOC15_REG_ENTRY_OFFSET_INST(entry, inst) \
+	(adev->reg_offset[entry.hwip][inst][entry.seg] + entry.reg_offset)
+
 #define SOC15_REG_GOLDEN_VALUE(ip, inst, reg, and_mask, or_mask) \
 	{ ip##_HWIP, inst, reg##_BASE_IDX, reg, and_mask, or_mask }
 
@@ -100,7 +107,7 @@ struct soc15_ras_field_entry {
 #define SOC15_RAS_REG_FIELD_VAL(val, entry, field) SOC15_REG_FIELD_VAL((val), (entry).field##_count_mask, (entry).field##_count_shift)
 
 void soc15_grbm_select(struct amdgpu_device *adev,
-		    u32 me, u32 pipe, u32 queue, u32 vmid);
+		    u32 me, u32 pipe, u32 queue, u32 vmid, int xcc_id);
 void soc15_set_virt_ops(struct amdgpu_device *adev);
 
 void soc15_program_register_sequence(struct amdgpu_device *adev,
@@ -111,7 +118,13 @@ int vega10_reg_base_init(struct amdgpu_device *adev);
 int vega20_reg_base_init(struct amdgpu_device *adev);
 int arct_reg_base_init(struct amdgpu_device *adev);
 int aldebaran_reg_base_init(struct amdgpu_device *adev);
+u64 aqua_vanjaram_encode_ext_smn_addressing(int ext_id);
+int aqua_vanjaram_init_soc_config(struct amdgpu_device *adev);
+ssize_t aqua_vanjaram_get_reg_state(struct amdgpu_device *adev,
+				    enum amdgpu_reg_state reg_state, void *buf,
+				    size_t max_size);
 
 void vega10_doorbell_index_init(struct amdgpu_device *adev);
 void vega20_doorbell_index_init(struct amdgpu_device *adev);
+void aqua_vanjaram_doorbell_index_init(struct amdgpu_device *adev);
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/soc15_common.h b/drivers/gpu/drm/amd/amdgpu/soc15_common.h
index 8a9ca87d8663..242b24f73c17 100644
--- a/drivers/gpu/drm/amd/amdgpu/soc15_common.h
+++ b/drivers/gpu/drm/amd/amdgpu/soc15_common.h
@@ -24,92 +24,100 @@
 #ifndef __SOC15_COMMON_H__
 #define __SOC15_COMMON_H__
 
+/* GET_INST returns the physical instance corresponding to a logical instance */
+#define GET_INST(ip, inst) \
+	(adev->ip_map.logical_to_dev_inst ? \
+	adev->ip_map.logical_to_dev_inst(adev, ip##_HWIP, inst) : inst)
+#define GET_MASK(ip, mask) \
+	(adev->ip_map.logical_to_dev_mask ? \
+	adev->ip_map.logical_to_dev_mask(adev, ip##_HWIP, mask) : mask)
+
 /* Register Access Macros */
 #define SOC15_REG_OFFSET(ip, inst, reg)	(adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg)
+#define SOC15_REG_OFFSET1(ip, inst, reg, offset) \
+	(adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + (reg)+(offset))
 
-#define __WREG32_SOC15_RLC__(reg, value, flag, hwip) \
-	((amdgpu_sriov_vf(adev) && adev->gfx.rlc.funcs && adev->gfx.rlc.funcs->sriov_wreg) ? \
-	 adev->gfx.rlc.funcs->sriov_wreg(adev, reg, value, flag, hwip) : \
+#define __WREG32_SOC15_RLC__(reg, value, flag, hwip, inst) \
+	((amdgpu_sriov_vf(adev) && adev->gfx.rlc.funcs && adev->gfx.rlc.rlcg_reg_access_supported) ? \
+	 amdgpu_sriov_wreg(adev, reg, value, flag, hwip, inst) : \
 	 WREG32(reg, value))
 
-#define __RREG32_SOC15_RLC__(reg, flag, hwip) \
-	((amdgpu_sriov_vf(adev) && adev->gfx.rlc.funcs && adev->gfx.rlc.funcs->sriov_rreg) ? \
-	 adev->gfx.rlc.funcs->sriov_rreg(adev, reg, flag, hwip) : \
+#define __RREG32_SOC15_RLC__(reg, flag, hwip, inst) \
+	((amdgpu_sriov_vf(adev) && adev->gfx.rlc.funcs && adev->gfx.rlc.rlcg_reg_access_supported) ? \
+	 amdgpu_sriov_rreg(adev, reg, flag, hwip, inst) : \
 	 RREG32(reg))
 
 #define WREG32_FIELD15(ip, idx, reg, field, val)	\
 	 __WREG32_SOC15_RLC__(adev->reg_offset[ip##_HWIP][idx][mm##reg##_BASE_IDX] + mm##reg,	\
 				(__RREG32_SOC15_RLC__( \
 					adev->reg_offset[ip##_HWIP][idx][mm##reg##_BASE_IDX] + mm##reg, \
-					0, ip##_HWIP) & \
+					0, ip##_HWIP, idx) & \
 				~REG_FIELD_MASK(reg, field)) | (val) << REG_FIELD_SHIFT(reg, field), \
-			      0, ip##_HWIP)
+			      0, ip##_HWIP, idx)
+
+#define WREG32_FIELD15_PREREG(ip, idx, reg_name, field, val)        \
+	__WREG32_SOC15_RLC__(adev->reg_offset[ip##_HWIP][idx][reg##reg_name##_BASE_IDX] + reg##reg_name,   \
+			(__RREG32_SOC15_RLC__( \
+					adev->reg_offset[ip##_HWIP][idx][reg##reg_name##_BASE_IDX] + reg##reg_name, \
+					0, ip##_HWIP, idx) & \
+					~REG_FIELD_MASK(reg_name, field)) | (val) << REG_FIELD_SHIFT(reg_name, field), \
+			0, ip##_HWIP, idx)
 
 #define RREG32_SOC15(ip, inst, reg) \
 	__RREG32_SOC15_RLC__(adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg, \
-			 0, ip##_HWIP)
+			 0, ip##_HWIP, inst)
+
+#define RREG32_SOC15_IP(ip, reg) __RREG32_SOC15_RLC__(reg, 0, ip##_HWIP, 0)
 
-#define RREG32_SOC15_IP(ip, reg) __RREG32_SOC15_RLC__(reg, 0, ip##_HWIP)
+#define RREG32_SOC15_IP_NO_KIQ(ip, reg, inst) __RREG32_SOC15_RLC__(reg, AMDGPU_REGS_NO_KIQ, ip##_HWIP, inst)
 
 #define RREG32_SOC15_NO_KIQ(ip, inst, reg) \
 	__RREG32_SOC15_RLC__(adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg, \
-			 AMDGPU_REGS_NO_KIQ, ip##_HWIP)
+			 AMDGPU_REGS_NO_KIQ, ip##_HWIP, inst)
 
 #define RREG32_SOC15_OFFSET(ip, inst, reg, offset) \
-	 __RREG32_SOC15_RLC__((adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg) + offset, 0, ip##_HWIP)
+	 __RREG32_SOC15_RLC__((adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + (reg)) + \
+			 (offset), 0, ip##_HWIP, inst)
 
 #define WREG32_SOC15(ip, inst, reg, value) \
 	 __WREG32_SOC15_RLC__((adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg), \
-			  value, 0, ip##_HWIP)
+			  value, 0, ip##_HWIP, inst)
 
 #define WREG32_SOC15_IP(ip, reg, value) \
-	 __WREG32_SOC15_RLC__(reg, value, 0, ip##_HWIP)
+	 __WREG32_SOC15_RLC__(reg, value, 0, ip##_HWIP, 0)
+
+#define WREG32_SOC15_IP_NO_KIQ(ip, reg, value, inst) \
+	 __WREG32_SOC15_RLC__(reg, value, AMDGPU_REGS_NO_KIQ, ip##_HWIP, inst)
 
 #define WREG32_SOC15_NO_KIQ(ip, inst, reg, value) \
 	__WREG32_SOC15_RLC__(adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg, \
-			     value, AMDGPU_REGS_NO_KIQ, ip##_HWIP)
+			     value, AMDGPU_REGS_NO_KIQ, ip##_HWIP, inst)
 
 #define WREG32_SOC15_OFFSET(ip, inst, reg, offset, value) \
 	 __WREG32_SOC15_RLC__((adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg) + offset, \
-			  value, 0, ip##_HWIP)
+			  value, 0, ip##_HWIP, inst)
 
-#define SOC15_WAIT_ON_RREG(ip, inst, reg, expected_value, mask) \
-({	int ret = 0;						\
-	do {							\
-		uint32_t old_ = 0;				\
-		uint32_t tmp_ = RREG32(adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg); \
-		uint32_t loop = adev->usec_timeout;		\
-		ret = 0;					\
-		while ((tmp_ & (mask)) != (expected_value)) {	\
-			if (old_ != tmp_) {			\
-				loop = adev->usec_timeout;	\
-				old_ = tmp_;			\
-			} else					\
-				udelay(1);			\
-			tmp_ = RREG32(adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg); \
-			loop--;					\
-			if (!loop) {				\
-				DRM_WARN("Register(%d) [%s] failed to reach value 0x%08x != 0x%08x\n", \
-					  inst, #reg, (unsigned)expected_value, (unsigned)(tmp_ & (mask))); \
-				ret = -ETIMEDOUT;		\
-				break;				\
-			}					\
-		}						\
-	} while (0);						\
-	ret;							\
-})
+#define SOC15_WAIT_ON_RREG(ip, inst, reg, expected_value, mask)      \
+	amdgpu_device_wait_on_rreg(adev, inst,                       \
+	(adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + (reg)), \
+	#reg, expected_value, mask)
+
+#define SOC15_WAIT_ON_RREG_OFFSET(ip, inst, reg, offset, expected_value, mask)  \
+	amdgpu_device_wait_on_rreg(adev, inst,                                  \
+	(adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + (reg) + (offset)), \
+	#reg, expected_value, mask)
 
 #define WREG32_RLC(reg, value) \
-	__WREG32_SOC15_RLC__(reg, value, AMDGPU_REGS_RLC, GC_HWIP)
+	__WREG32_SOC15_RLC__(reg, value, AMDGPU_REGS_RLC, GC_HWIP, 0)
 
-#define WREG32_RLC_EX(prefix, reg, value) \
+#define WREG32_RLC_EX(prefix, reg, value, inst) \
 	do {							\
 		if (amdgpu_sriov_fullaccess(adev)) {    \
 			uint32_t i = 0;	\
 			uint32_t retries = 50000;	\
-			uint32_t r0 = adev->reg_offset[GC_HWIP][0][prefix##SCRATCH_REG0_BASE_IDX] + prefix##SCRATCH_REG0;	\
-			uint32_t r1 = adev->reg_offset[GC_HWIP][0][prefix##SCRATCH_REG1_BASE_IDX] + prefix##SCRATCH_REG1;	\
-			uint32_t spare_int = adev->reg_offset[GC_HWIP][0][prefix##RLC_SPARE_INT_BASE_IDX] + prefix##RLC_SPARE_INT;	\
+			uint32_t r0 = adev->reg_offset[GC_HWIP][inst][prefix##SCRATCH_REG0_BASE_IDX] + prefix##SCRATCH_REG0;	\
+			uint32_t r1 = adev->reg_offset[GC_HWIP][inst][prefix##SCRATCH_REG1_BASE_IDX] + prefix##SCRATCH_REG1;	\
+			uint32_t spare_int = adev->reg_offset[GC_HWIP][inst][prefix##RLC_SPARE_INT_BASE_IDX] + prefix##RLC_SPARE_INT;	\
 			WREG32(r0, value);	\
 			WREG32(r1, (reg | 0x80000000));	\
 			WREG32(spare_int, 0x1);	\
@@ -128,26 +136,26 @@
 
 /* shadow the registers in the callback function */
 #define WREG32_SOC15_RLC_SHADOW(ip, inst, reg, value) \
-	__WREG32_SOC15_RLC__((adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg), value, AMDGPU_REGS_RLC, GC_HWIP)
+	__WREG32_SOC15_RLC__((adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg), value, AMDGPU_REGS_RLC, GC_HWIP, inst)
 
 /* for GC only */
 #define RREG32_RLC(reg) \
-	__RREG32_SOC15_RLC__(reg, AMDGPU_REGS_RLC, GC_HWIP)
+	__RREG32_SOC15_RLC__(reg, AMDGPU_REGS_RLC, GC_HWIP, 0)
 
 #define WREG32_RLC_NO_KIQ(reg, value, hwip) \
-	__WREG32_SOC15_RLC__(reg, value, AMDGPU_REGS_NO_KIQ | AMDGPU_REGS_RLC, hwip)
+	__WREG32_SOC15_RLC__(reg, value, AMDGPU_REGS_NO_KIQ | AMDGPU_REGS_RLC, hwip, 0)
 
 #define RREG32_RLC_NO_KIQ(reg, hwip) \
-	__RREG32_SOC15_RLC__(reg, AMDGPU_REGS_NO_KIQ | AMDGPU_REGS_RLC, hwip)
+	__RREG32_SOC15_RLC__(reg, AMDGPU_REGS_NO_KIQ | AMDGPU_REGS_RLC, hwip, 0)
 
 #define WREG32_SOC15_RLC_SHADOW_EX(prefix, ip, inst, reg, value) \
 	do {							\
 		uint32_t target_reg = adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg;\
 		if (amdgpu_sriov_fullaccess(adev)) {    \
-			uint32_t r2 = adev->reg_offset[GC_HWIP][0][prefix##SCRATCH_REG1_BASE_IDX] + prefix##SCRATCH_REG2;	\
-			uint32_t r3 = adev->reg_offset[GC_HWIP][0][prefix##SCRATCH_REG1_BASE_IDX] + prefix##SCRATCH_REG3;	\
-			uint32_t grbm_cntl = adev->reg_offset[GC_HWIP][0][prefix##GRBM_GFX_CNTL_BASE_IDX] + prefix##GRBM_GFX_CNTL;   \
-			uint32_t grbm_idx = adev->reg_offset[GC_HWIP][0][prefix##GRBM_GFX_INDEX_BASE_IDX] + prefix##GRBM_GFX_INDEX;   \
+			uint32_t r2 = adev->reg_offset[GC_HWIP][inst][prefix##SCRATCH_REG1_BASE_IDX] + prefix##SCRATCH_REG2;	\
+			uint32_t r3 = adev->reg_offset[GC_HWIP][inst][prefix##SCRATCH_REG1_BASE_IDX] + prefix##SCRATCH_REG3;	\
+			uint32_t grbm_cntl = adev->reg_offset[GC_HWIP][inst][prefix##GRBM_GFX_CNTL_BASE_IDX] + prefix##GRBM_GFX_CNTL;   \
+			uint32_t grbm_idx = adev->reg_offset[GC_HWIP][inst][prefix##GRBM_GFX_INDEX_BASE_IDX] + prefix##GRBM_GFX_INDEX;   \
 			if (target_reg == grbm_cntl) \
 				WREG32(r2, value);	\
 			else if (target_reg == grbm_idx) \
@@ -159,31 +167,47 @@
 	} while (0)
 
 #define RREG32_SOC15_RLC(ip, inst, reg) \
-	__RREG32_SOC15_RLC__(adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg, AMDGPU_REGS_RLC, ip##_HWIP)
+	__RREG32_SOC15_RLC__(adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg, AMDGPU_REGS_RLC, ip##_HWIP, inst)
 
 #define WREG32_SOC15_RLC(ip, inst, reg, value) \
 	do {							\
-		uint32_t target_reg = adev->reg_offset[ip##_HWIP][0][reg##_BASE_IDX] + reg;\
-		__WREG32_SOC15_RLC__(target_reg, value, AMDGPU_REGS_RLC, ip##_HWIP); \
+		uint32_t target_reg = adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg;\
+		__WREG32_SOC15_RLC__(target_reg, value, AMDGPU_REGS_RLC, ip##_HWIP, inst); \
 	} while (0)
 
 #define WREG32_SOC15_RLC_EX(prefix, ip, inst, reg, value) \
 	do {							\
-			uint32_t target_reg = adev->reg_offset[GC_HWIP][0][reg##_BASE_IDX] + reg;\
-			WREG32_RLC_EX(prefix, target_reg, value); \
+			uint32_t target_reg = adev->reg_offset[GC_HWIP][inst][reg##_BASE_IDX] + reg;\
+			WREG32_RLC_EX(prefix, target_reg, value, inst); \
 	} while (0)
 
 #define WREG32_FIELD15_RLC(ip, idx, reg, field, val)   \
 	__WREG32_SOC15_RLC__((adev->reg_offset[ip##_HWIP][idx][mm##reg##_BASE_IDX] + mm##reg), \
 			     (__RREG32_SOC15_RLC__(adev->reg_offset[ip##_HWIP][idx][mm##reg##_BASE_IDX] + mm##reg, \
-						   AMDGPU_REGS_RLC, ip##_HWIP) & \
+						   AMDGPU_REGS_RLC, ip##_HWIP, idx) & \
 			      ~REG_FIELD_MASK(reg, field)) | (val) << REG_FIELD_SHIFT(reg, field), \
-			     AMDGPU_REGS_RLC, ip##_HWIP)
+			     AMDGPU_REGS_RLC, ip##_HWIP, idx)
 
 #define WREG32_SOC15_OFFSET_RLC(ip, inst, reg, offset, value) \
-	__WREG32_SOC15_RLC__((adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg) + offset, value, AMDGPU_REGS_RLC, ip##_HWIP)
+	__WREG32_SOC15_RLC__((adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg) + offset, value, AMDGPU_REGS_RLC, ip##_HWIP, inst)
 
 #define RREG32_SOC15_OFFSET_RLC(ip, inst, reg, offset) \
-	__RREG32_SOC15_RLC__((adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg) + offset, AMDGPU_REGS_RLC, ip##_HWIP)
+	__RREG32_SOC15_RLC__((adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg) + offset, AMDGPU_REGS_RLC, ip##_HWIP, inst)
+
+/* inst equals to ext for some IPs */
+#define RREG32_SOC15_EXT(ip, inst, reg, ext) \
+	RREG32_PCIE_EXT((adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg) * 4 \
+			+ adev->asic_funcs->encode_ext_smn_addressing(ext)) \
+
+#define WREG32_SOC15_EXT(ip, inst, reg, ext, value) \
+	WREG32_PCIE_EXT((adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg) * 4 \
+			+ adev->asic_funcs->encode_ext_smn_addressing(ext), \
+			value) \
+
+#define RREG64_MCA(ext, mca_base, idx) \
+	RREG64_PCIE_EXT(adev->asic_funcs->encode_ext_smn_addressing(ext) + mca_base + (idx * 8))
+
+#define WREG64_MCA(ext, mca_base, idx, val) \
+	WREG64_PCIE_EXT(adev->asic_funcs->encode_ext_smn_addressing(ext) + mca_base + (idx * 8), val)
 
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/soc15d.h b/drivers/gpu/drm/amd/amdgpu/soc15d.h
index 799925d22fc8..cf93fa477674 100644
--- a/drivers/gpu/drm/amd/amdgpu/soc15d.h
+++ b/drivers/gpu/drm/amd/amdgpu/soc15d.h
@@ -76,6 +76,12 @@
 			 ((cond & 0xF) << 24) |				\
 			 ((type & 0xF) << 28))
 
+#define CP_PACKETJ_NOP		0x60000000
+#define CP_PACKETJ_GET_REG(x)  ((x) & 0x3FFFF)
+#define CP_PACKETJ_GET_RES(x)  (((x) >> 18) & 0x3F)
+#define CP_PACKETJ_GET_COND(x) (((x) >> 24) & 0xF)
+#define CP_PACKETJ_GET_TYPE(x) (((x) >> 28) & 0xF)
+
 /* Packet 3 types */
 #define	PACKET3_NOP					0x10
 #define	PACKET3_SET_BASE				0x11
@@ -87,11 +93,25 @@
 #define	PACKET3_DISPATCH_INDIRECT			0x16
 #define	PACKET3_ATOMIC_GDS				0x1D
 #define	PACKET3_ATOMIC_MEM				0x1E
+#define		PACKET3_ATOMIC_MEM__ATOMIC(x) ((((unsigned)(x)) & 0x3F) << 0)
+#define		PACKET3_ATOMIC_MEM__COMMAND(x) ((((unsigned)(x)) & 0xF) << 8)
+#define		PACKET3_ATOMIC_MEM__CACHE_POLICY(x) ((((unsigned)(x)) & 0x3) << 25)
+#define		PACKET3_ATOMIC_MEM__ADDR_LO(x) (((unsigned)(x)) << 0)
+#define		PACKET3_ATOMIC_MEM__ADDR_HI(x) (((unsigned)(x)) << 0)
+#define		PACKET3_ATOMIC_MEM__SRC_DATA_LO(x) (((unsigned)(x)) << 0)
+#define		PACKET3_ATOMIC_MEM__SRC_DATA_HI(x) (((unsigned)(x)) << 0)
+#define		PACKET3_ATOMIC_MEM__CMP_DATA_LO(x) (((unsigned)(x)) << 0)
+#define		PACKET3_ATOMIC_MEM__CMP_DATA_HI(x) (((unsigned)(x)) << 0)
+#define		PACKET3_ATOMIC_MEM__LOOP_INTERVAL(x) ((((unsigned)(x)) & 0x1FFF) << 0)
+#define 	PACKET3_ATOMIC_MEM__COMMAND__SINGLE_PASS_ATOMIC	0
+#define 	PACKET3_ATOMIC_MEM__COMMAND__LOOP_UNTIL_COMPARE_SATISFIED 1
 #define	PACKET3_OCCLUSION_QUERY				0x1F
 #define	PACKET3_SET_PREDICATION				0x20
 #define	PACKET3_REG_RMW					0x21
 #define	PACKET3_COND_EXEC				0x22
 #define	PACKET3_PRED_EXEC				0x23
+#define		PACKET3_PRED_EXEC__EXEC_COUNT(x)	((((unsigned)(x)) & 0x3FFF) << 0)
+#define		PACKET3_PRED_EXEC__VIRTUAL_XCC_ID_SELECT(x)	((((unsigned)(x)) & 0xFF) << 24)
 #define	PACKET3_DRAW_INDIRECT				0x24
 #define	PACKET3_DRAW_INDEX_INDIRECT			0x25
 #define	PACKET3_INDEX_BASE				0x26
@@ -126,6 +146,28 @@
 		 * 1 - pfp
 		 * 2 - ce
 		 */
+#define		PACKET3_WRITE_DATA__DST_SEL(x) ((((unsigned)(x)) & 0xF) << 8)
+#define		PACKET3_WRITE_DATA__ADDR_INCR(x) ((((unsigned)(x)) & 0x1) << 16)
+#define		PACKET3_WRITE_DATA__RESUME_VF_MI300(x) ((((unsigned)(x)) & 0x1) << 19)
+#define		PACKET3_WRITE_DATA__WR_CONFIRM(x) ((((unsigned)(x)) & 0x1) << 20)
+#define		PACKET3_WRITE_DATA__CACHE_POLICY(x) ((((unsigned)(x)) & 0x3) << 25)
+#define		PACKET3_WRITE_DATA__DST_MMREG_ADDR(x) ((((unsigned)(x)) & 0x3FFFF) << 0)
+#define		PACKET3_WRITE_DATA__DST_GDS_ADDR(x) ((((unsigned)(x)) & 0xFFFF) << 0)
+#define		PACKET3_WRITE_DATA__DST_MEM_ADDR_LO(x) ((((unsigned)(x)) & 0x3FFFFFFF) << 2)
+#define		PACKET3_WRITE_DATA__DST_MEM_ADDR_HI(x) ((unsigned)(x))
+#define		PACKET3_WRITE_DATA__DST_SEL__MEM_MAPPED_REGISTER	0
+#define		PACKET3_WRITE_DATA__DST_SEL__TC_L2	2
+#define		PACKET3_WRITE_DATA__DST_SEL__GDS	3
+#define		PACKET3_WRITE_DATA__DST_SEL__MEMORY	5
+#define		PACKET3_WRITE_DATA__DST_SEL__MEMORY_MAPPED_ADC_PERSISTENT_STATE	6
+#define		PACKET3_WRITE_DATA__ADDR_INCR__INCREMENT_ADDRESS	0
+#define		PACKET3_WRITE_DATA__ADDR_INCR__DO_NOT_INCREMENT_ADDRESS	1
+#define		PACKET3_WRITE_DATA__WR_CONFIRM__DO_NOT_WAIT_FOR_WRITE_CONFIRMATION	0
+#define		PACKET3_WRITE_DATA__WR_CONFIRM__WAIT_FOR_WRITE_CONFIRMATION	1
+#define		PACKET3_WRITE_DATA__CACHE_POLICY__LRU	0
+#define		PACKET3_WRITE_DATA__CACHE_POLICY__STREAM	1
+#define		PACKET3_WRITE_DATA__CACHE_POLICY__NOA	2
+#define		PACKET3_WRITE_DATA__CACHE_POLICY__BYPASS	3
 #define	PACKET3_DRAW_INDEX_INDIRECT_MULTI		0x38
 #define	PACKET3_MEM_SEMAPHORE				0x39
 #              define PACKET3_SEM_USE_MAILBOX       (0x1 << 16)
@@ -154,6 +196,33 @@
 		/* 0 - me
 		 * 1 - pfp
 		 */
+#define		PACKET3_WAIT_REG_MEM__FUNCTION(x) ((((unsigned)(x)) & 0x7) << 0)
+#define		PACKET3_WAIT_REG_MEM__MEM_SPACE(x) ((((unsigned)(x)) & 0x3) << 4)
+#define		PACKET3_WAIT_REG_MEM__OPERATION(x) ((((unsigned)(x)) & 0x3) << 6)
+#define		PACKET3_WAIT_REG_MEM__MES_INTR_PIPE(x) ((((unsigned)(x)) & 0x3) << 22)
+#define		PACKET3_WAIT_REG_MEM__MES_ACTION(x) ((((unsigned)(x)) & 0x1) << 24)
+#define		PACKET3_WAIT_REG_MEM__CACHE_POLICY(x) ((((unsigned)(x)) & 0x3) << 25)
+#define		PACKET3_WAIT_REG_MEM__MEM_POLL_ADDR_LO(x) ((((unsigned)(x)) & 0x3FFFFFFF) << 2)
+#define		PACKET3_WAIT_REG_MEM__REG_POLL_ADDR(x) ((((unsigned)(x)) & 0x3FFFF) << 0)
+#define		PACKET3_WAIT_REG_MEM__REG_WRITE_ADDR1(x) ((((unsigned)(x)) & 0x3FFFF) << 0)
+#define		PACKET3_WAIT_REG_MEM__MEM_POLL_ADDR_HI(x) ((unsigned)(x))
+#define		PACKET3_WAIT_REG_MEM__REG_WRITE_ADDR2(x) ((((unsigned)(x)) & 0x3FFFF) << 0)
+#define		PACKET3_WAIT_REG_MEM__REFERENCE(x) ((unsigned)(x))
+#define		PACKET3_WAIT_REG_MEM__MASK(x) ((unsigned)(x))
+#define		PACKET3_WAIT_REG_MEM__POLL_INTERVAL(x) ((((unsigned)(x)) & 0xFFFF) << 0)
+#define		PACKET3_WAIT_REG_MEM__OPTIMIZE_ACE_OFFLOAD_MODE(x) ((((unsigned)(x)) & 0x1) << 31)
+#define		PACKET3_WAIT_REG_MEM__FUNCTION__ALWAYS_PASS	0
+#define		PACKET3_WAIT_REG_MEM__FUNCTION__LESS_THAN_REF_VALUE	1
+#define		PACKET3_WAIT_REG_MEM__FUNCTION__LESS_THAN_EQUAL_TO_THE_REF_VALUE	2
+#define		PACKET3_WAIT_REG_MEM__FUNCTION__EQUAL_TO_THE_REFERENCE_VALUE	3
+#define		PACKET3_WAIT_REG_MEM__FUNCTION__NOT_EQUAL_REFERENCE_VALUE	4
+#define		PACKET3_WAIT_REG_MEM__FUNCTION__GREATER_THAN_OR_EQUAL_REFERENCE_VALUE	5
+#define		PACKET3_WAIT_REG_MEM__FUNCTION__GREATER_THAN_REFERENCE_VALUE	6
+#define		PACKET3_WAIT_REG_MEM__MEM_SPACE__REGISTER_SPACE	0
+#define		PACKET3_WAIT_REG_MEM__MEM_SPACE__MEMORY_SPACE	1
+#define		PACKET3_WAIT_REG_MEM__OPERATION__WAIT_REG_MEM	0
+#define		PACKET3_WAIT_REG_MEM__OPERATION__WR_WAIT_WR_REG	1
+#define		PACKET3_WAIT_REG_MEM__OPERATION__WAIT_MEM_PREEMPTABLE	3
 #define	PACKET3_INDIRECT_BUFFER				0x3F
 #define		INDIRECT_BUFFER_VALID                   (1 << 23)
 #define		INDIRECT_BUFFER_CACHE_POLICY(x)         ((x) << 28)
@@ -162,7 +231,64 @@
 		 * 2 - Bypass
 		 */
 #define     INDIRECT_BUFFER_PRE_ENB(x)		 ((x) << 21)
+#define     INDIRECT_BUFFER_PRE_RESUME(x)               ((x) << 30)
+#define		PACKET3_INDIRECT_BUFFER__IB_BASE_LO(x) ((((unsigned)(x)) & 0x3FFFFFFF) << 2)
+#define		PACKET3_INDIRECT_BUFFER__IB_BASE_HI(x) ((unsigned)(x))
+#define		PACKET3_INDIRECT_BUFFER__IB_SIZE(x) ((((unsigned)(x)) & 0xFFFFF) << 0)
+#define		PACKET3_INDIRECT_BUFFER__CHAIN(x) ((((unsigned)(x)) & 0x1) << 20)
+#define		PACKET3_INDIRECT_BUFFER__OFFLOAD_POLLING(x) ((((unsigned)(x)) & 0x1) << 21)
+#define		PACKET3_INDIRECT_BUFFER__VALID(x) ((((unsigned)(x)) & 0x1) << 23)
+#define		PACKET3_INDIRECT_BUFFER__VMID(x) ((((unsigned)(x)) & 0xF) << 24)
+#define		PACKET3_INDIRECT_BUFFER__CACHE_POLICY(x) ((((unsigned)(x)) & 0x3) << 28)
+#define		PACKET3_INDIRECT_BUFFER__PRIV(x) ((((unsigned)(x)) & 0x1) << 31)
+#define		PACKET3_INDIRECT_BUFFER__CACHE_POLICY__LRU	0
+#define		PACKET3_INDIRECT_BUFFER__CACHE_POLICY__STREAM	1
 #define	PACKET3_COPY_DATA				0x40
+#define		PACKET3_COPY_DATA__SRC_SEL(x) ((((unsigned)(x)) & 0xF) << 0)
+#define		PACKET3_COPY_DATA__DST_SEL(x) ((((unsigned)(x)) & 0xF) << 8)
+#define		PACKET3_COPY_DATA__SRC_CACHE_POLICY(x) ((((unsigned)(x)) & 0x3) << 13)
+#define		PACKET3_COPY_DATA__COUNT_SEL(x) ((((unsigned)(x)) & 0x1) << 16)
+#define		PACKET3_COPY_DATA__WR_CONFIRM(x) ((((unsigned)(x)) & 0x1) << 20)
+#define		PACKET3_COPY_DATA__DST_CACHE_POLICY(x) ((((unsigned)(x)) & 0x3) << 25)
+#define		PACKET3_COPY_DATA__PQ_EXE_STATUS(x) ((((unsigned)(x)) & 0x1) << 29)
+#define		PACKET3_COPY_DATA__SRC_REG_OFFSET(x) ((((unsigned)(x)) & 0x3FFFF) << 0)
+#define		PACKET3_COPY_DATA__SRC_32B_ADDR_LO(x) ((((unsigned)(x)) & 0x3FFFFFFF) << 2)
+#define		PACKET3_COPY_DATA__SRC_64B_ADDR_LO(x) ((((unsigned)(x)) & 0x1FFFFFFF) << 3)
+#define		PACKET3_COPY_DATA__SRC_GDS_ADDR_LO(x) ((((unsigned)(x)) & 0xFFFF) << 0)
+#define		PACKET3_COPY_DATA__IMM_DATA(x) ((unsigned)(x))
+#define		PACKET3_COPY_DATA__SRC_MEMTC_ADDR_HI(x) ((unsigned)(x))
+#define		PACKET3_COPY_DATA__SRC_IMM_DATA(x) ((unsigned)(x))
+#define		PACKET3_COPY_DATA__DST_REG_OFFSET(x) ((((unsigned)(x)) & 0x3FFFF) << 0)
+#define		PACKET3_COPY_DATA__DST_32B_ADDR_LO(x) ((((unsigned)(x)) & 0x3FFFFFFF) << 2)
+#define		PACKET3_COPY_DATA__DST_64B_ADDR_LO(x) ((((unsigned)(x)) & 0x1FFFFFFF) << 3)
+#define		PACKET3_COPY_DATA__DST_GDS_ADDR_LO(x) ((((unsigned)(x)) & 0xFFFF) << 0)
+#define		PACKET3_COPY_DATA__DST_ADDR_HI(x) ((unsigned)(x))
+#define		PACKET3_COPY_DATA__SRC_SEL__MEM_MAPPED_REGISTER	0
+#define		PACKET3_COPY_DATA__SRC_SEL__MEMORY	1
+#define		PACKET3_COPY_DATA__SRC_SEL__TC_L2	2
+#define		PACKET3_COPY_DATA__SRC_SEL__GDS	3
+#define		PACKET3_COPY_DATA__SRC_SEL__PERFCOUNTERS	4
+#define		PACKET3_COPY_DATA__SRC_SEL__IMMEDIATE_DATA	5
+#define		PACKET3_COPY_DATA__SRC_SEL__ATOMIC_RETURN_DATA	6
+#define		PACKET3_COPY_DATA__SRC_SEL__GDS_ATOMIC_RETURN_DATA0	7
+#define		PACKET3_COPY_DATA__SRC_SEL__GDS_ATOMIC_RETURN_DATA1	8
+#define		PACKET3_COPY_DATA__SRC_SEL__GPU_CLOCK_COUNT	9
+#define		PACKET3_COPY_DATA__DST_SEL__MEM_MAPPED_REGISTER	0
+#define		PACKET3_COPY_DATA__DST_SEL__TC_L2	2
+#define		PACKET3_COPY_DATA__DST_SEL__GDS	3
+#define		PACKET3_COPY_DATA__DST_SEL__PERFCOUNTERS	4
+#define		PACKET3_COPY_DATA__DST_SEL__MEMORY	5
+#define		PACKET3_COPY_DATA__DST_SEL__MEM_MAPPED_REG_DC	6
+#define		PACKET3_COPY_DATA__SRC_CACHE_POLICY__LRU	0
+#define		PACKET3_COPY_DATA__SRC_CACHE_POLICY__STREAM 1
+#define		PACKET3_COPY_DATA__COUNT_SEL__32_BITS_OF_DATA	0
+#define		PACKET3_COPY_DATA__COUNT_SEL__64_BITS_OF_DATA	1
+#define		PACKET3_COPY_DATA__WR_CONFIRM__DO_NOT_WAIT_FOR_CONFIRMATION	0
+#define		PACKET3_COPY_DATA__WR_CONFIRM__WAIT_FOR_CONFIRMATION	1
+#define		PACKET3_COPY_DATA__DST_CACHE_POLICY__LRU	0
+#define		PACKET3_COPY_DATA__DST_CACHE_POLICY__STREAM	1
+#define		PACKET3_COPY_DATA__PQ_EXE_STATUS__DEFAULT	0
+#define		PACKET3_COPY_DATA__PQ_EXE_STATUS__PHASE_UPDATE	1
 #define	PACKET3_PFP_SYNC_ME				0x42
 #define	PACKET3_COND_WRITE				0x45
 #define	PACKET3_EVENT_WRITE				0x46
@@ -174,6 +300,15 @@
 		 * 3 - SAMPLE_STREAMOUTSTAT*
 		 * 4 - *S_PARTIAL_FLUSH
 		 */
+#define		PACKET3_EVENT_WRITE__EVENT_TYPE(x) ((((unsigned)(x)) & 0x3F) << 0)
+#define		PACKET3_EVENT_WRITE__EVENT_INDEX(x) ((((unsigned)(x)) & 0xF) << 8)
+#define		PACKET3_EVENT_WRITE__OFFLOAD_ENABLE(x) ((((unsigned)(x)) & 0x1) << 31)
+#define		PACKET3_EVENT_WRITE__SAMP_PLST_CNTR_MODE(x) ((((unsigned)(x)) & 0x3) << 29)
+#define		PACKET3_EVENT_WRITE__ADDRESS_LO(x) ((((unsigned)(x)) & 0x1FFFFFFF) << 3)
+#define		PACKET3_EVENT_WRITE__ADDRESS_HI(x) (((unsigned)(x)) << 0)
+#define		PACKET3_EVENT_WRITE__EVENT_INDEX__OTHER	0
+#define		PACKET3_EVENT_WRITE__EVENT_INDEX__SAMPLE_PIPELINESTATS	2
+#define		PACKET3_EVENT_WRITE__EVENT_INDEX__CS_PARTIAL_FLUSH	4
 #define	PACKET3_RELEASE_MEM				0x49
 #define		EVENT_TYPE(x)                           ((x) << 0)
 #define		EVENT_INDEX(x)                          ((x) << 8)
@@ -184,6 +319,7 @@
 #define		EOP_TC_ACTION_EN                        (1 << 17) /* L2 */
 #define		EOP_TC_NC_ACTION_EN			(1 << 19)
 #define		EOP_TC_MD_ACTION_EN			(1 << 21) /* L2 metadata */
+#define		EOP_EXEC				(1 << 28) /* For Trailing Fence */
 
 #define		DATA_SEL(x)                             ((x) << 29)
 		/* 0 - discard
@@ -278,6 +414,13 @@
 #define 	PACKET3_ACQUIRE_MEM_CP_COHER_CNTL_SH_ICACHE_ACTION_ENA(x) ((x) << 29)
 #define 	PACKET3_ACQUIRE_MEM_CP_COHER_CNTL_SH_KCACHE_WB_ACTION_ENA(x) ((x) << 30)
 #define	PACKET3_REWIND					0x59
+#define		PACKET3_ACQUIRE_MEM__COHER_SIZE(x) ((unsigned)(x))
+#define 	PACKET3_ACQUIRE_MEM__COHER_SIZE_HI(x) ((((unsigned)(x)) & 0xFF) << 0)
+#define 	PACKET3_ACQUIRE_MEM__COHER_SIZE_HI_VG10(x) ((((unsigned)(x)) & 0xFFFFFF) << 0)
+#define 	PACKET3_ACQUIRE_MEM__COHER_BASE_LO(x) ((unsigned)(x))
+#define 	PACKET3_ACQUIRE_MEM__COHER_BASE_HI(x) ((((unsigned)(x)) & 0xFFFFFF) << 0)
+#define 	PACKET3_ACQUIRE_MEM__POLL_INTERVAL(x) ((((unsigned)(x)) & 0xFFFF) << 0)
+#define 	PACKET3_ACQUIRE_MEM__GCR_CNTL(x) ((((unsigned)(x)) & 0x7FF) << 0)
 #define	PACKET3_LOAD_UCONFIG_REG			0x5E
 #define	PACKET3_LOAD_SH_REG				0x5F
 #define	PACKET3_LOAD_CONFIG_REG				0x60
@@ -292,12 +435,16 @@
 #define	PACKET3_SET_SH_REG				0x76
 #define		PACKET3_SET_SH_REG_START			0x00002c00
 #define		PACKET3_SET_SH_REG_END				0x00003000
+#define 	PACKET3_SET_SH_REG__REG_OFFSET(x) ((((unsigned)(x)) & 0xFFFF) << 0)
+#define 	PACKET3_SET_SH_REG__VMID_SHIFT(x) ((((unsigned)(x)) & 0x1F) << 23)
+#define 	PACKET3_SET_SH_REG__INDEX(x) ((((unsigned)(x)) & 0xF) << 28)
 #define	PACKET3_SET_SH_REG_OFFSET			0x77
 #define	PACKET3_SET_QUEUE_REG				0x78
 #define	PACKET3_SET_UCONFIG_REG				0x79
 #define		PACKET3_SET_UCONFIG_REG_START			0x0000c000
 #define		PACKET3_SET_UCONFIG_REG_END			0x0000c400
 #define		PACKET3_SET_UCONFIG_REG_INDEX_TYPE		(2 << 28)
+#define 	PACKET3_SET_UCONFIG_REG__REG_OFFSET(x) ((((unsigned)(x)) & 0xFFFF) << 0)
 #define	PACKET3_SCRATCH_RAM_WRITE			0x7D
 #define	PACKET3_SCRATCH_RAM_READ			0x7E
 #define	PACKET3_LOAD_CONST_RAM				0x80
@@ -405,6 +552,15 @@
 #              define PACKET3_QUERY_STATUS_DOORBELL_OFFSET(x)  ((x) << 2)
 #              define PACKET3_QUERY_STATUS_ENG_SEL(x)          ((x) << 25)
 
+#define PACKET3_RUN_CLEANER_SHADER_9_0                      0xD7
+/* 1. header
+ * 2. RESERVED [31:0]
+ */
+
+#define PACKET3_RUN_CLEANER_SHADER                      0xD2
+/* 1. header
+ * 2. RESERVED [31:0]
+ */
 
 #define VCE_CMD_NO_OP		0x00000000
 #define VCE_CMD_END		0x00000001
diff --git a/drivers/gpu/drm/amd/amdgpu/soc21.c b/drivers/gpu/drm/amd/amdgpu/soc21.c
new file mode 100644
index 000000000000..ad36c96478a8
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/soc21.c
@@ -0,0 +1,1021 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include <linux/firmware.h>
+#include <linux/slab.h>
+#include <linux/module.h>
+#include <linux/pci.h>
+
+#include "amdgpu.h"
+#include "amdgpu_atombios.h"
+#include "amdgpu_ih.h"
+#include "amdgpu_uvd.h"
+#include "amdgpu_vce.h"
+#include "amdgpu_ucode.h"
+#include "amdgpu_psp.h"
+#include "amdgpu_smu.h"
+#include "atom.h"
+#include "amd_pcie.h"
+
+#include "gc/gc_11_0_0_offset.h"
+#include "gc/gc_11_0_0_sh_mask.h"
+#include "mp/mp_13_0_0_offset.h"
+
+#include "soc15.h"
+#include "soc15_common.h"
+#include "soc21.h"
+#include "mxgpu_nv.h"
+
+static const struct amd_ip_funcs soc21_common_ip_funcs;
+
+/* SOC21 */
+static const struct amdgpu_video_codec_info vcn_4_0_0_video_codecs_encode_array_vcn0[] = {
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4096, 0)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC, 8192, 4352, 0)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_AV1, 8192, 4352, 0)},
+};
+
+static const struct amdgpu_video_codec_info vcn_4_0_0_video_codecs_encode_array_vcn1[] = {
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4096, 0)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC, 8192, 4352, 0)},
+};
+
+static const struct amdgpu_video_codecs vcn_4_0_0_video_codecs_encode_vcn0 = {
+	.codec_count = ARRAY_SIZE(vcn_4_0_0_video_codecs_encode_array_vcn0),
+	.codec_array = vcn_4_0_0_video_codecs_encode_array_vcn0,
+};
+
+static const struct amdgpu_video_codecs vcn_4_0_0_video_codecs_encode_vcn1 = {
+	.codec_count = ARRAY_SIZE(vcn_4_0_0_video_codecs_encode_array_vcn1),
+	.codec_array = vcn_4_0_0_video_codecs_encode_array_vcn1,
+};
+
+static const struct amdgpu_video_codec_info vcn_4_0_0_video_codecs_decode_array_vcn0[] = {
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4096, 52)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC, 8192, 4352, 186)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG, 16384, 16384, 0)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VP9, 8192, 4352, 0)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_AV1, 8192, 4352, 0)},
+};
+
+static const struct amdgpu_video_codec_info vcn_4_0_0_video_codecs_decode_array_vcn1[] = {
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4096, 52)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC, 8192, 4352, 186)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG, 16384, 16384, 0)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VP9, 8192, 4352, 0)},
+};
+
+static const struct amdgpu_video_codecs vcn_4_0_0_video_codecs_decode_vcn0 = {
+	.codec_count = ARRAY_SIZE(vcn_4_0_0_video_codecs_decode_array_vcn0),
+	.codec_array = vcn_4_0_0_video_codecs_decode_array_vcn0,
+};
+
+static const struct amdgpu_video_codecs vcn_4_0_0_video_codecs_decode_vcn1 = {
+	.codec_count = ARRAY_SIZE(vcn_4_0_0_video_codecs_decode_array_vcn1),
+	.codec_array = vcn_4_0_0_video_codecs_decode_array_vcn1,
+};
+
+/* SRIOV SOC21, not const since data is controlled by host */
+static struct amdgpu_video_codec_info sriov_vcn_4_0_0_video_codecs_encode_array_vcn0[] = {
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4096, 0)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC, 8192, 4352, 0)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_AV1, 8192, 4352, 0)},
+};
+
+static struct amdgpu_video_codec_info sriov_vcn_4_0_0_video_codecs_encode_array_vcn1[] = {
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4096, 0)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC, 8192, 4352, 0)},
+};
+
+static struct amdgpu_video_codecs sriov_vcn_4_0_0_video_codecs_encode_vcn0 = {
+	.codec_count = ARRAY_SIZE(sriov_vcn_4_0_0_video_codecs_encode_array_vcn0),
+	.codec_array = sriov_vcn_4_0_0_video_codecs_encode_array_vcn0,
+};
+
+static struct amdgpu_video_codecs sriov_vcn_4_0_0_video_codecs_encode_vcn1 = {
+	.codec_count = ARRAY_SIZE(sriov_vcn_4_0_0_video_codecs_encode_array_vcn1),
+	.codec_array = sriov_vcn_4_0_0_video_codecs_encode_array_vcn1,
+};
+
+static struct amdgpu_video_codec_info sriov_vcn_4_0_0_video_codecs_decode_array_vcn0[] = {
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4096, 52)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC, 8192, 4352, 186)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG, 16384, 16384, 0)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VP9, 8192, 4352, 0)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_AV1, 8192, 4352, 0)},
+};
+
+static struct amdgpu_video_codec_info sriov_vcn_4_0_0_video_codecs_decode_array_vcn1[] = {
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4096, 52)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC, 8192, 4352, 186)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG, 16384, 16384, 0)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VP9, 8192, 4352, 0)},
+};
+
+static struct amdgpu_video_codecs sriov_vcn_4_0_0_video_codecs_decode_vcn0 = {
+	.codec_count = ARRAY_SIZE(sriov_vcn_4_0_0_video_codecs_decode_array_vcn0),
+	.codec_array = sriov_vcn_4_0_0_video_codecs_decode_array_vcn0,
+};
+
+static struct amdgpu_video_codecs sriov_vcn_4_0_0_video_codecs_decode_vcn1 = {
+	.codec_count = ARRAY_SIZE(sriov_vcn_4_0_0_video_codecs_decode_array_vcn1),
+	.codec_array = sriov_vcn_4_0_0_video_codecs_decode_array_vcn1,
+};
+
+static int soc21_query_video_codecs(struct amdgpu_device *adev, bool encode,
+				 const struct amdgpu_video_codecs **codecs)
+{
+	if (adev->vcn.num_vcn_inst == hweight8(adev->vcn.harvest_config))
+		return -EINVAL;
+
+	switch (amdgpu_ip_version(adev, UVD_HWIP, 0)) {
+	case IP_VERSION(4, 0, 0):
+	case IP_VERSION(4, 0, 2):
+	case IP_VERSION(4, 0, 4):
+	case IP_VERSION(4, 0, 5):
+		if (amdgpu_sriov_vf(adev)) {
+			if ((adev->vcn.harvest_config & AMDGPU_VCN_HARVEST_VCN0) ||
+			!amdgpu_sriov_is_av1_support(adev)) {
+				if (encode)
+					*codecs = &sriov_vcn_4_0_0_video_codecs_encode_vcn1;
+				else
+					*codecs = &sriov_vcn_4_0_0_video_codecs_decode_vcn1;
+			} else {
+				if (encode)
+					*codecs = &sriov_vcn_4_0_0_video_codecs_encode_vcn0;
+				else
+					*codecs = &sriov_vcn_4_0_0_video_codecs_decode_vcn0;
+			}
+		} else {
+			if ((adev->vcn.harvest_config & AMDGPU_VCN_HARVEST_VCN0)) {
+				if (encode)
+					*codecs = &vcn_4_0_0_video_codecs_encode_vcn1;
+				else
+					*codecs = &vcn_4_0_0_video_codecs_decode_vcn1;
+			} else {
+				if (encode)
+					*codecs = &vcn_4_0_0_video_codecs_encode_vcn0;
+				else
+					*codecs = &vcn_4_0_0_video_codecs_decode_vcn0;
+			}
+		}
+		return 0;
+	case IP_VERSION(4, 0, 6):
+		if (encode)
+			*codecs = &vcn_4_0_0_video_codecs_encode_vcn0;
+		else
+			*codecs = &vcn_4_0_0_video_codecs_decode_vcn0;
+		return 0;
+	default:
+		return -EINVAL;
+	}
+}
+
+static u32 soc21_didt_rreg(struct amdgpu_device *adev, u32 reg)
+{
+	unsigned long flags, address, data;
+	u32 r;
+
+	address = SOC15_REG_OFFSET(GC, 0, regDIDT_IND_INDEX);
+	data = SOC15_REG_OFFSET(GC, 0, regDIDT_IND_DATA);
+
+	spin_lock_irqsave(&adev->didt_idx_lock, flags);
+	WREG32(address, (reg));
+	r = RREG32(data);
+	spin_unlock_irqrestore(&adev->didt_idx_lock, flags);
+	return r;
+}
+
+static void soc21_didt_wreg(struct amdgpu_device *adev, u32 reg, u32 v)
+{
+	unsigned long flags, address, data;
+
+	address = SOC15_REG_OFFSET(GC, 0, regDIDT_IND_INDEX);
+	data = SOC15_REG_OFFSET(GC, 0, regDIDT_IND_DATA);
+
+	spin_lock_irqsave(&adev->didt_idx_lock, flags);
+	WREG32(address, (reg));
+	WREG32(data, (v));
+	spin_unlock_irqrestore(&adev->didt_idx_lock, flags);
+}
+
+static u32 soc21_get_config_memsize(struct amdgpu_device *adev)
+{
+	return adev->nbio.funcs->get_memsize(adev);
+}
+
+static u32 soc21_get_xclk(struct amdgpu_device *adev)
+{
+	return adev->clock.spll.reference_freq;
+}
+
+
+void soc21_grbm_select(struct amdgpu_device *adev,
+		     u32 me, u32 pipe, u32 queue, u32 vmid)
+{
+	u32 grbm_gfx_cntl = 0;
+	grbm_gfx_cntl = REG_SET_FIELD(grbm_gfx_cntl, GRBM_GFX_CNTL, PIPEID, pipe);
+	grbm_gfx_cntl = REG_SET_FIELD(grbm_gfx_cntl, GRBM_GFX_CNTL, MEID, me);
+	grbm_gfx_cntl = REG_SET_FIELD(grbm_gfx_cntl, GRBM_GFX_CNTL, VMID, vmid);
+	grbm_gfx_cntl = REG_SET_FIELD(grbm_gfx_cntl, GRBM_GFX_CNTL, QUEUEID, queue);
+
+	WREG32_SOC15(GC, 0, regGRBM_GFX_CNTL, grbm_gfx_cntl);
+}
+
+static bool soc21_read_disabled_bios(struct amdgpu_device *adev)
+{
+	/* todo */
+	return false;
+}
+
+static struct soc15_allowed_register_entry soc21_allowed_read_registers[] = {
+	{ SOC15_REG_ENTRY(GC, 0, regGRBM_STATUS)},
+	{ SOC15_REG_ENTRY(GC, 0, regGRBM_STATUS2)},
+	{ SOC15_REG_ENTRY(GC, 0, regGRBM_STATUS_SE0)},
+	{ SOC15_REG_ENTRY(GC, 0, regGRBM_STATUS_SE1)},
+	{ SOC15_REG_ENTRY(GC, 0, regGRBM_STATUS_SE2)},
+	{ SOC15_REG_ENTRY(GC, 0, regGRBM_STATUS_SE3)},
+	{ SOC15_REG_ENTRY(SDMA0, 0, regSDMA0_STATUS_REG)},
+	{ SOC15_REG_ENTRY(SDMA1, 0, regSDMA1_STATUS_REG)},
+	{ SOC15_REG_ENTRY(GC, 0, regCP_STAT)},
+	{ SOC15_REG_ENTRY(GC, 0, regCP_STALLED_STAT1)},
+	{ SOC15_REG_ENTRY(GC, 0, regCP_STALLED_STAT2)},
+	{ SOC15_REG_ENTRY(GC, 0, regCP_STALLED_STAT3)},
+	{ SOC15_REG_ENTRY(GC, 0, regCP_CPF_BUSY_STAT)},
+	{ SOC15_REG_ENTRY(GC, 0, regCP_CPF_STALLED_STAT1)},
+	{ SOC15_REG_ENTRY(GC, 0, regCP_CPF_STATUS)},
+	{ SOC15_REG_ENTRY(GC, 0, regCP_CPC_BUSY_STAT)},
+	{ SOC15_REG_ENTRY(GC, 0, regCP_CPC_STALLED_STAT1)},
+	{ SOC15_REG_ENTRY(GC, 0, regCP_CPC_STATUS)},
+	{ SOC15_REG_ENTRY(GC, 0, regGB_ADDR_CONFIG)},
+};
+
+static uint32_t soc21_read_indexed_register(struct amdgpu_device *adev, u32 se_num,
+					 u32 sh_num, u32 reg_offset)
+{
+	uint32_t val;
+
+	mutex_lock(&adev->grbm_idx_mutex);
+	if (se_num != 0xffffffff || sh_num != 0xffffffff)
+		amdgpu_gfx_select_se_sh(adev, se_num, sh_num, 0xffffffff, 0);
+
+	val = RREG32(reg_offset);
+
+	if (se_num != 0xffffffff || sh_num != 0xffffffff)
+		amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
+	mutex_unlock(&adev->grbm_idx_mutex);
+	return val;
+}
+
+static uint32_t soc21_get_register_value(struct amdgpu_device *adev,
+				      bool indexed, u32 se_num,
+				      u32 sh_num, u32 reg_offset)
+{
+	if (indexed) {
+		return soc21_read_indexed_register(adev, se_num, sh_num, reg_offset);
+	} else {
+		if (reg_offset == SOC15_REG_OFFSET(GC, 0, regGB_ADDR_CONFIG) && adev->gfx.config.gb_addr_config)
+			return adev->gfx.config.gb_addr_config;
+		return RREG32(reg_offset);
+	}
+}
+
+static int soc21_read_register(struct amdgpu_device *adev, u32 se_num,
+			    u32 sh_num, u32 reg_offset, u32 *value)
+{
+	uint32_t i;
+	struct soc15_allowed_register_entry  *en;
+
+	*value = 0;
+	for (i = 0; i < ARRAY_SIZE(soc21_allowed_read_registers); i++) {
+		en = &soc21_allowed_read_registers[i];
+		if (!adev->reg_offset[en->hwip][en->inst])
+			continue;
+		else if (reg_offset != (adev->reg_offset[en->hwip][en->inst][en->seg]
+					+ en->reg_offset))
+			continue;
+
+		*value = soc21_get_register_value(adev,
+					       soc21_allowed_read_registers[i].grbm_indexed,
+					       se_num, sh_num, reg_offset);
+		return 0;
+	}
+	return -EINVAL;
+}
+
+#if 0
+static int soc21_asic_mode1_reset(struct amdgpu_device *adev)
+{
+	u32 i;
+	int ret = 0;
+
+	amdgpu_atombios_scratch_regs_engine_hung(adev, true);
+
+	/* disable BM */
+	pci_clear_master(adev->pdev);
+
+	amdgpu_device_cache_pci_state(adev->pdev);
+
+	if (amdgpu_dpm_is_mode1_reset_supported(adev)) {
+		dev_info(adev->dev, "GPU smu mode1 reset\n");
+		ret = amdgpu_dpm_mode1_reset(adev);
+	} else {
+		dev_info(adev->dev, "GPU psp mode1 reset\n");
+		ret = psp_gpu_reset(adev);
+	}
+
+	if (ret)
+		dev_err(adev->dev, "GPU mode1 reset failed\n");
+	amdgpu_device_load_pci_state(adev->pdev);
+
+	/* wait for asic to come out of reset */
+	for (i = 0; i < adev->usec_timeout; i++) {
+		u32 memsize = adev->nbio.funcs->get_memsize(adev);
+
+		if (memsize != 0xffffffff)
+			break;
+		udelay(1);
+	}
+
+	amdgpu_atombios_scratch_regs_engine_hung(adev, false);
+
+	return ret;
+}
+#endif
+
+static enum amd_reset_method
+soc21_asic_reset_method(struct amdgpu_device *adev)
+{
+	if (amdgpu_reset_method == AMD_RESET_METHOD_MODE1 ||
+	    amdgpu_reset_method == AMD_RESET_METHOD_MODE2 ||
+	    amdgpu_reset_method == AMD_RESET_METHOD_BACO)
+		return amdgpu_reset_method;
+
+	if (amdgpu_reset_method != -1)
+		dev_warn(adev->dev, "Specified reset method:%d isn't supported, using AUTO instead.\n",
+				  amdgpu_reset_method);
+
+	switch (amdgpu_ip_version(adev, MP1_HWIP, 0)) {
+	case IP_VERSION(13, 0, 0):
+	case IP_VERSION(13, 0, 7):
+	case IP_VERSION(13, 0, 10):
+		return AMD_RESET_METHOD_MODE1;
+	case IP_VERSION(13, 0, 4):
+	case IP_VERSION(13, 0, 11):
+	case IP_VERSION(14, 0, 0):
+	case IP_VERSION(14, 0, 1):
+	case IP_VERSION(14, 0, 4):
+	case IP_VERSION(14, 0, 5):
+		return AMD_RESET_METHOD_MODE2;
+	default:
+		if (amdgpu_dpm_is_baco_supported(adev))
+			return AMD_RESET_METHOD_BACO;
+		else
+			return AMD_RESET_METHOD_MODE1;
+	}
+}
+
+static int soc21_asic_reset(struct amdgpu_device *adev)
+{
+	int ret = 0;
+
+	switch (soc21_asic_reset_method(adev)) {
+	case AMD_RESET_METHOD_PCI:
+		dev_info(adev->dev, "PCI reset\n");
+		ret = amdgpu_device_pci_reset(adev);
+		break;
+	case AMD_RESET_METHOD_BACO:
+		dev_info(adev->dev, "BACO reset\n");
+		ret = amdgpu_dpm_baco_reset(adev);
+		break;
+	case AMD_RESET_METHOD_MODE2:
+		dev_info(adev->dev, "MODE2 reset\n");
+		ret = amdgpu_dpm_mode2_reset(adev);
+		break;
+	default:
+		dev_info(adev->dev, "MODE1 reset\n");
+		ret = amdgpu_device_mode1_reset(adev);
+		break;
+	}
+
+	return ret;
+}
+
+static int soc21_set_uvd_clocks(struct amdgpu_device *adev, u32 vclk, u32 dclk)
+{
+	/* todo */
+	return 0;
+}
+
+static int soc21_set_vce_clocks(struct amdgpu_device *adev, u32 evclk, u32 ecclk)
+{
+	/* todo */
+	return 0;
+}
+
+static void soc21_program_aspm(struct amdgpu_device *adev)
+{
+	if (!amdgpu_device_should_use_aspm(adev))
+		return;
+
+	if (adev->nbio.funcs->program_aspm)
+		adev->nbio.funcs->program_aspm(adev);
+}
+
+const struct amdgpu_ip_block_version soc21_common_ip_block = {
+	.type = AMD_IP_BLOCK_TYPE_COMMON,
+	.major = 1,
+	.minor = 0,
+	.rev = 0,
+	.funcs = &soc21_common_ip_funcs,
+};
+
+static bool soc21_need_full_reset(struct amdgpu_device *adev)
+{
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	case IP_VERSION(11, 0, 0):
+	case IP_VERSION(11, 0, 2):
+	case IP_VERSION(11, 0, 3):
+	default:
+		return true;
+	}
+}
+
+static bool soc21_need_reset_on_init(struct amdgpu_device *adev)
+{
+	u32 sol_reg;
+
+	if (adev->flags & AMD_IS_APU)
+		return false;
+
+	/* Check sOS sign of life register to confirm sys driver and sOS
+	 * are already been loaded.
+	 */
+	sol_reg = RREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_81);
+	if (sol_reg)
+		return true;
+
+	return false;
+}
+
+static void soc21_init_doorbell_index(struct amdgpu_device *adev)
+{
+	adev->doorbell_index.kiq = AMDGPU_NAVI10_DOORBELL_KIQ;
+	adev->doorbell_index.mec_ring0 = AMDGPU_NAVI10_DOORBELL_MEC_RING0;
+	adev->doorbell_index.mec_ring1 = AMDGPU_NAVI10_DOORBELL_MEC_RING1;
+	adev->doorbell_index.mec_ring2 = AMDGPU_NAVI10_DOORBELL_MEC_RING2;
+	adev->doorbell_index.mec_ring3 = AMDGPU_NAVI10_DOORBELL_MEC_RING3;
+	adev->doorbell_index.mec_ring4 = AMDGPU_NAVI10_DOORBELL_MEC_RING4;
+	adev->doorbell_index.mec_ring5 = AMDGPU_NAVI10_DOORBELL_MEC_RING5;
+	adev->doorbell_index.mec_ring6 = AMDGPU_NAVI10_DOORBELL_MEC_RING6;
+	adev->doorbell_index.mec_ring7 = AMDGPU_NAVI10_DOORBELL_MEC_RING7;
+	adev->doorbell_index.userqueue_start = AMDGPU_NAVI10_DOORBELL_USERQUEUE_START;
+	adev->doorbell_index.userqueue_end = AMDGPU_NAVI10_DOORBELL_USERQUEUE_END;
+	adev->doorbell_index.gfx_ring0 = AMDGPU_NAVI10_DOORBELL_GFX_RING0;
+	adev->doorbell_index.gfx_ring1 = AMDGPU_NAVI10_DOORBELL_GFX_RING1;
+	adev->doorbell_index.gfx_userqueue_start =
+		AMDGPU_NAVI10_DOORBELL_GFX_USERQUEUE_START;
+	adev->doorbell_index.gfx_userqueue_end =
+		AMDGPU_NAVI10_DOORBELL_GFX_USERQUEUE_END;
+	adev->doorbell_index.mes_ring0 = AMDGPU_NAVI10_DOORBELL_MES_RING0;
+	adev->doorbell_index.mes_ring1 = AMDGPU_NAVI10_DOORBELL_MES_RING1;
+	adev->doorbell_index.sdma_engine[0] = AMDGPU_NAVI10_DOORBELL_sDMA_ENGINE0;
+	adev->doorbell_index.sdma_engine[1] = AMDGPU_NAVI10_DOORBELL_sDMA_ENGINE1;
+	adev->doorbell_index.ih = AMDGPU_NAVI10_DOORBELL_IH;
+	adev->doorbell_index.vcn.vcn_ring0_1 = AMDGPU_NAVI10_DOORBELL64_VCN0_1;
+	adev->doorbell_index.vcn.vcn_ring2_3 = AMDGPU_NAVI10_DOORBELL64_VCN2_3;
+	adev->doorbell_index.vcn.vcn_ring4_5 = AMDGPU_NAVI10_DOORBELL64_VCN4_5;
+	adev->doorbell_index.vcn.vcn_ring6_7 = AMDGPU_NAVI10_DOORBELL64_VCN6_7;
+	adev->doorbell_index.vpe_ring = AMDGPU_NAVI10_DOORBELL64_VPE;
+	adev->doorbell_index.first_non_cp = AMDGPU_NAVI10_DOORBELL64_FIRST_NON_CP;
+	adev->doorbell_index.last_non_cp = AMDGPU_NAVI10_DOORBELL64_LAST_NON_CP;
+
+	adev->doorbell_index.max_assignment = AMDGPU_NAVI10_DOORBELL_MAX_ASSIGNMENT << 1;
+	adev->doorbell_index.sdma_doorbell_range = 20;
+}
+
+static void soc21_pre_asic_init(struct amdgpu_device *adev)
+{
+}
+
+static int soc21_update_umd_stable_pstate(struct amdgpu_device *adev,
+					  bool enter)
+{
+	if (enter)
+		amdgpu_gfx_rlc_enter_safe_mode(adev, 0);
+	else
+		amdgpu_gfx_rlc_exit_safe_mode(adev, 0);
+
+	if (adev->gfx.funcs->update_perfmon_mgcg)
+		adev->gfx.funcs->update_perfmon_mgcg(adev, !enter);
+
+	return 0;
+}
+
+static const struct amdgpu_asic_funcs soc21_asic_funcs = {
+	.read_disabled_bios = &soc21_read_disabled_bios,
+	.read_bios_from_rom = &amdgpu_soc15_read_bios_from_rom,
+	.read_register = &soc21_read_register,
+	.reset = &soc21_asic_reset,
+	.reset_method = &soc21_asic_reset_method,
+	.get_xclk = &soc21_get_xclk,
+	.set_uvd_clocks = &soc21_set_uvd_clocks,
+	.set_vce_clocks = &soc21_set_vce_clocks,
+	.get_config_memsize = &soc21_get_config_memsize,
+	.init_doorbell_index = &soc21_init_doorbell_index,
+	.need_full_reset = &soc21_need_full_reset,
+	.need_reset_on_init = &soc21_need_reset_on_init,
+	.get_pcie_replay_count = &amdgpu_nbio_get_pcie_replay_count,
+	.supports_baco = &amdgpu_dpm_is_baco_supported,
+	.pre_asic_init = &soc21_pre_asic_init,
+	.query_video_codecs = &soc21_query_video_codecs,
+	.update_umd_stable_pstate = &soc21_update_umd_stable_pstate,
+};
+
+static int soc21_common_early_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	adev->nbio.funcs->set_reg_remap(adev);
+	adev->smc_rreg = NULL;
+	adev->smc_wreg = NULL;
+	adev->pcie_rreg = &amdgpu_device_indirect_rreg;
+	adev->pcie_wreg = &amdgpu_device_indirect_wreg;
+	adev->pcie_rreg64 = &amdgpu_device_indirect_rreg64;
+	adev->pcie_wreg64 = &amdgpu_device_indirect_wreg64;
+	adev->pciep_rreg = amdgpu_device_pcie_port_rreg;
+	adev->pciep_wreg = amdgpu_device_pcie_port_wreg;
+
+	/* TODO: will add them during VCN v2 implementation */
+	adev->uvd_ctx_rreg = NULL;
+	adev->uvd_ctx_wreg = NULL;
+
+	adev->didt_rreg = &soc21_didt_rreg;
+	adev->didt_wreg = &soc21_didt_wreg;
+
+	adev->asic_funcs = &soc21_asic_funcs;
+
+	adev->rev_id = amdgpu_device_get_rev_id(adev);
+	adev->external_rev_id = 0xff;
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	case IP_VERSION(11, 0, 0):
+		adev->cg_flags = AMD_CG_SUPPORT_GFX_CGCG |
+			AMD_CG_SUPPORT_GFX_CGLS |
+#if 0
+			AMD_CG_SUPPORT_GFX_3D_CGCG |
+			AMD_CG_SUPPORT_GFX_3D_CGLS |
+#endif
+			AMD_CG_SUPPORT_GFX_MGCG |
+			AMD_CG_SUPPORT_REPEATER_FGCG |
+			AMD_CG_SUPPORT_GFX_FGCG |
+			AMD_CG_SUPPORT_GFX_PERF_CLK |
+			AMD_CG_SUPPORT_VCN_MGCG |
+			AMD_CG_SUPPORT_JPEG_MGCG |
+			AMD_CG_SUPPORT_ATHUB_MGCG |
+			AMD_CG_SUPPORT_ATHUB_LS |
+			AMD_CG_SUPPORT_MC_MGCG |
+			AMD_CG_SUPPORT_MC_LS |
+			AMD_CG_SUPPORT_IH_CG |
+			AMD_CG_SUPPORT_HDP_SD;
+		adev->pg_flags = AMD_PG_SUPPORT_VCN |
+			AMD_PG_SUPPORT_VCN_DPG |
+			AMD_PG_SUPPORT_JPEG |
+			AMD_PG_SUPPORT_ATHUB |
+			AMD_PG_SUPPORT_MMHUB;
+		adev->external_rev_id = adev->rev_id + 0x1; // TODO: need update
+		break;
+	case IP_VERSION(11, 0, 2):
+		adev->cg_flags =
+			AMD_CG_SUPPORT_GFX_CGCG |
+			AMD_CG_SUPPORT_GFX_CGLS |
+			AMD_CG_SUPPORT_REPEATER_FGCG |
+			AMD_CG_SUPPORT_VCN_MGCG |
+			AMD_CG_SUPPORT_JPEG_MGCG |
+			AMD_CG_SUPPORT_ATHUB_MGCG |
+			AMD_CG_SUPPORT_ATHUB_LS |
+			AMD_CG_SUPPORT_IH_CG |
+			AMD_CG_SUPPORT_HDP_SD;
+		adev->pg_flags =
+			AMD_PG_SUPPORT_VCN |
+			AMD_PG_SUPPORT_VCN_DPG |
+			AMD_PG_SUPPORT_JPEG |
+			AMD_PG_SUPPORT_ATHUB |
+			AMD_PG_SUPPORT_MMHUB;
+		adev->external_rev_id = adev->rev_id + 0x10;
+		break;
+	case IP_VERSION(11, 0, 1):
+		adev->cg_flags =
+			AMD_CG_SUPPORT_GFX_CGCG |
+			AMD_CG_SUPPORT_GFX_CGLS |
+			AMD_CG_SUPPORT_GFX_MGCG |
+			AMD_CG_SUPPORT_GFX_FGCG |
+			AMD_CG_SUPPORT_REPEATER_FGCG |
+			AMD_CG_SUPPORT_GFX_PERF_CLK |
+			AMD_CG_SUPPORT_MC_MGCG |
+			AMD_CG_SUPPORT_MC_LS |
+			AMD_CG_SUPPORT_HDP_MGCG |
+			AMD_CG_SUPPORT_HDP_LS |
+			AMD_CG_SUPPORT_ATHUB_MGCG |
+			AMD_CG_SUPPORT_ATHUB_LS |
+			AMD_CG_SUPPORT_IH_CG |
+			AMD_CG_SUPPORT_BIF_MGCG |
+			AMD_CG_SUPPORT_BIF_LS |
+			AMD_CG_SUPPORT_VCN_MGCG |
+			AMD_CG_SUPPORT_JPEG_MGCG;
+		adev->pg_flags =
+			AMD_PG_SUPPORT_GFX_PG |
+			AMD_PG_SUPPORT_VCN |
+			AMD_PG_SUPPORT_VCN_DPG |
+			AMD_PG_SUPPORT_JPEG;
+		adev->external_rev_id = adev->rev_id + 0x1;
+		break;
+	case IP_VERSION(11, 0, 3):
+		adev->cg_flags = AMD_CG_SUPPORT_VCN_MGCG |
+			AMD_CG_SUPPORT_JPEG_MGCG |
+			AMD_CG_SUPPORT_GFX_CGCG |
+			AMD_CG_SUPPORT_GFX_CGLS |
+			AMD_CG_SUPPORT_REPEATER_FGCG |
+			AMD_CG_SUPPORT_GFX_MGCG |
+			AMD_CG_SUPPORT_HDP_SD |
+			AMD_CG_SUPPORT_ATHUB_MGCG |
+			AMD_CG_SUPPORT_ATHUB_LS;
+		adev->pg_flags = AMD_PG_SUPPORT_VCN |
+			AMD_PG_SUPPORT_VCN_DPG |
+			AMD_PG_SUPPORT_JPEG;
+		adev->external_rev_id = adev->rev_id + 0x20;
+		break;
+	case IP_VERSION(11, 0, 4):
+		adev->cg_flags =
+			AMD_CG_SUPPORT_GFX_CGCG |
+			AMD_CG_SUPPORT_GFX_CGLS |
+			AMD_CG_SUPPORT_GFX_MGCG |
+			AMD_CG_SUPPORT_GFX_FGCG |
+			AMD_CG_SUPPORT_REPEATER_FGCG |
+			AMD_CG_SUPPORT_GFX_PERF_CLK |
+			AMD_CG_SUPPORT_MC_MGCG |
+			AMD_CG_SUPPORT_MC_LS |
+			AMD_CG_SUPPORT_HDP_MGCG |
+			AMD_CG_SUPPORT_HDP_LS |
+			AMD_CG_SUPPORT_ATHUB_MGCG |
+			AMD_CG_SUPPORT_ATHUB_LS |
+			AMD_CG_SUPPORT_IH_CG |
+			AMD_CG_SUPPORT_BIF_MGCG |
+			AMD_CG_SUPPORT_BIF_LS |
+			AMD_CG_SUPPORT_VCN_MGCG |
+			AMD_CG_SUPPORT_JPEG_MGCG;
+		adev->pg_flags = AMD_PG_SUPPORT_VCN |
+			AMD_PG_SUPPORT_VCN_DPG |
+			AMD_PG_SUPPORT_GFX_PG |
+			AMD_PG_SUPPORT_JPEG;
+		adev->external_rev_id = adev->rev_id + 0x80;
+		break;
+	case IP_VERSION(11, 5, 0):
+		adev->cg_flags = AMD_CG_SUPPORT_VCN_MGCG |
+			AMD_CG_SUPPORT_JPEG_MGCG |
+			AMD_CG_SUPPORT_GFX_CGCG |
+			AMD_CG_SUPPORT_GFX_CGLS |
+			AMD_CG_SUPPORT_GFX_MGCG |
+			AMD_CG_SUPPORT_GFX_FGCG |
+			AMD_CG_SUPPORT_REPEATER_FGCG |
+			AMD_CG_SUPPORT_GFX_PERF_CLK	|
+			AMD_CG_SUPPORT_GFX_3D_CGCG |
+			AMD_CG_SUPPORT_GFX_3D_CGLS	|
+			AMD_CG_SUPPORT_MC_MGCG |
+			AMD_CG_SUPPORT_MC_LS |
+			AMD_CG_SUPPORT_HDP_LS |
+			AMD_CG_SUPPORT_HDP_DS |
+			AMD_CG_SUPPORT_HDP_SD |
+			AMD_CG_SUPPORT_ATHUB_MGCG |
+			AMD_CG_SUPPORT_ATHUB_LS |
+			AMD_CG_SUPPORT_IH_CG |
+			AMD_CG_SUPPORT_BIF_MGCG |
+			AMD_CG_SUPPORT_BIF_LS;
+		adev->pg_flags = AMD_PG_SUPPORT_VCN_DPG |
+			AMD_PG_SUPPORT_JPEG_DPG |
+			AMD_PG_SUPPORT_VCN |
+			AMD_PG_SUPPORT_JPEG |
+			AMD_PG_SUPPORT_GFX_PG;
+		if (adev->rev_id == 0)
+			adev->external_rev_id = 0x1;
+		else
+			adev->external_rev_id = adev->rev_id + 0x10;
+		break;
+	case IP_VERSION(11, 5, 1):
+		adev->cg_flags =
+			AMD_CG_SUPPORT_GFX_CGCG |
+			AMD_CG_SUPPORT_GFX_CGLS |
+			AMD_CG_SUPPORT_GFX_MGCG |
+			AMD_CG_SUPPORT_GFX_FGCG |
+			AMD_CG_SUPPORT_REPEATER_FGCG |
+			AMD_CG_SUPPORT_GFX_PERF_CLK	|
+			AMD_CG_SUPPORT_GFX_3D_CGCG |
+			AMD_CG_SUPPORT_GFX_3D_CGLS	|
+			AMD_CG_SUPPORT_MC_MGCG |
+			AMD_CG_SUPPORT_MC_LS |
+			AMD_CG_SUPPORT_HDP_LS |
+			AMD_CG_SUPPORT_HDP_DS |
+			AMD_CG_SUPPORT_HDP_SD |
+			AMD_CG_SUPPORT_ATHUB_MGCG |
+			AMD_CG_SUPPORT_ATHUB_LS |
+			AMD_CG_SUPPORT_IH_CG |
+			AMD_CG_SUPPORT_BIF_MGCG |
+			AMD_CG_SUPPORT_BIF_LS |
+			AMD_CG_SUPPORT_VCN_MGCG |
+			AMD_CG_SUPPORT_JPEG_MGCG;
+		adev->pg_flags =
+			AMD_PG_SUPPORT_GFX_PG |
+			AMD_PG_SUPPORT_VCN |
+			AMD_PG_SUPPORT_VCN_DPG |
+			AMD_PG_SUPPORT_JPEG;
+		adev->external_rev_id = adev->rev_id + 0xc1;
+		break;
+	case IP_VERSION(11, 5, 2):
+		adev->cg_flags = AMD_CG_SUPPORT_VCN_MGCG |
+			AMD_CG_SUPPORT_JPEG_MGCG |
+			AMD_CG_SUPPORT_GFX_CGCG |
+			AMD_CG_SUPPORT_GFX_CGLS |
+			AMD_CG_SUPPORT_GFX_MGCG |
+			AMD_CG_SUPPORT_GFX_FGCG |
+			AMD_CG_SUPPORT_REPEATER_FGCG |
+			AMD_CG_SUPPORT_GFX_PERF_CLK	|
+			AMD_CG_SUPPORT_GFX_3D_CGCG |
+			AMD_CG_SUPPORT_GFX_3D_CGLS	|
+			AMD_CG_SUPPORT_MC_MGCG |
+			AMD_CG_SUPPORT_MC_LS |
+			AMD_CG_SUPPORT_HDP_LS |
+			AMD_CG_SUPPORT_HDP_DS |
+			AMD_CG_SUPPORT_HDP_SD |
+			AMD_CG_SUPPORT_ATHUB_MGCG |
+			AMD_CG_SUPPORT_ATHUB_LS |
+			AMD_CG_SUPPORT_IH_CG |
+			AMD_CG_SUPPORT_BIF_MGCG |
+			AMD_CG_SUPPORT_BIF_LS;
+		adev->pg_flags = AMD_PG_SUPPORT_VCN_DPG |
+			AMD_PG_SUPPORT_VCN |
+			AMD_PG_SUPPORT_JPEG_DPG |
+			AMD_PG_SUPPORT_JPEG |
+			AMD_PG_SUPPORT_GFX_PG;
+		adev->external_rev_id = adev->rev_id + 0x40;
+		break;
+	case IP_VERSION(11, 5, 3):
+		adev->cg_flags = AMD_CG_SUPPORT_VCN_MGCG |
+			AMD_CG_SUPPORT_JPEG_MGCG |
+			AMD_CG_SUPPORT_GFX_CGCG |
+			AMD_CG_SUPPORT_GFX_CGLS |
+			AMD_CG_SUPPORT_GFX_MGCG |
+			AMD_CG_SUPPORT_GFX_FGCG |
+			AMD_CG_SUPPORT_REPEATER_FGCG |
+			AMD_CG_SUPPORT_GFX_PERF_CLK |
+			AMD_CG_SUPPORT_GFX_3D_CGCG |
+			AMD_CG_SUPPORT_GFX_3D_CGLS |
+			AMD_CG_SUPPORT_MC_MGCG |
+			AMD_CG_SUPPORT_MC_LS |
+			AMD_CG_SUPPORT_HDP_LS |
+			AMD_CG_SUPPORT_HDP_DS |
+			AMD_CG_SUPPORT_HDP_SD |
+			AMD_CG_SUPPORT_ATHUB_MGCG |
+			AMD_CG_SUPPORT_ATHUB_LS |
+			AMD_CG_SUPPORT_IH_CG |
+			AMD_CG_SUPPORT_BIF_MGCG |
+			AMD_CG_SUPPORT_BIF_LS;
+		adev->pg_flags = AMD_PG_SUPPORT_VCN_DPG |
+			AMD_PG_SUPPORT_VCN |
+			AMD_PG_SUPPORT_JPEG_DPG |
+			AMD_PG_SUPPORT_JPEG |
+			AMD_PG_SUPPORT_GFX_PG;
+		adev->external_rev_id = adev->rev_id + 0x50;
+		break;
+	default:
+		/* FIXME: not supported yet */
+		return -EINVAL;
+	}
+
+	if (amdgpu_sriov_vf(adev)) {
+		amdgpu_virt_init_setting(adev);
+		xgpu_nv_mailbox_set_irq_funcs(adev);
+	}
+
+	return 0;
+}
+
+static int soc21_common_late_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	if (amdgpu_sriov_vf(adev)) {
+		xgpu_nv_mailbox_get_irq(adev);
+		if ((adev->vcn.harvest_config & AMDGPU_VCN_HARVEST_VCN0) ||
+		!amdgpu_sriov_is_av1_support(adev)) {
+			amdgpu_virt_update_sriov_video_codec(adev,
+							     sriov_vcn_4_0_0_video_codecs_encode_array_vcn1,
+							     ARRAY_SIZE(sriov_vcn_4_0_0_video_codecs_encode_array_vcn1),
+							     sriov_vcn_4_0_0_video_codecs_decode_array_vcn1,
+							     ARRAY_SIZE(sriov_vcn_4_0_0_video_codecs_decode_array_vcn1));
+		} else {
+			amdgpu_virt_update_sriov_video_codec(adev,
+							     sriov_vcn_4_0_0_video_codecs_encode_array_vcn0,
+							     ARRAY_SIZE(sriov_vcn_4_0_0_video_codecs_encode_array_vcn0),
+							     sriov_vcn_4_0_0_video_codecs_decode_array_vcn0,
+							     ARRAY_SIZE(sriov_vcn_4_0_0_video_codecs_decode_array_vcn0));
+		}
+	} else {
+		if (adev->nbio.ras &&
+		    adev->nbio.ras_err_event_athub_irq.funcs)
+			/* don't need to fail gpu late init
+			 * if enabling athub_err_event interrupt failed
+			 * nbio v4_3 only support fatal error hanlding
+			 * just enable the interrupt directly */
+			amdgpu_irq_get(adev, &adev->nbio.ras_err_event_athub_irq, 0);
+	}
+
+	/* Enable selfring doorbell aperture late because doorbell BAR
+	 * aperture will change if resize BAR successfully in gmc sw_init.
+	 */
+	adev->nbio.funcs->enable_doorbell_selfring_aperture(adev, true);
+
+	return 0;
+}
+
+static int soc21_common_sw_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	if (amdgpu_sriov_vf(adev))
+		xgpu_nv_mailbox_add_irq_id(adev);
+
+	return 0;
+}
+
+static int soc21_common_hw_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	/* enable aspm */
+	soc21_program_aspm(adev);
+	/* setup nbio registers */
+	adev->nbio.funcs->init_registers(adev);
+	/* remap HDP registers to a hole in mmio space,
+	 * for the purpose of expose those registers
+	 * to process space
+	 */
+	if (adev->nbio.funcs->remap_hdp_registers && !amdgpu_sriov_vf(adev))
+		adev->nbio.funcs->remap_hdp_registers(adev);
+	/* enable the doorbell aperture */
+	adev->nbio.funcs->enable_doorbell_aperture(adev, true);
+
+	return 0;
+}
+
+static int soc21_common_hw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	/* Disable the doorbell aperture and selfring doorbell aperture
+	 * separately in hw_fini because soc21_enable_doorbell_aperture
+	 * has been removed and there is no need to delay disabling
+	 * selfring doorbell.
+	 */
+	adev->nbio.funcs->enable_doorbell_aperture(adev, false);
+	adev->nbio.funcs->enable_doorbell_selfring_aperture(adev, false);
+
+	if (amdgpu_sriov_vf(adev)) {
+		xgpu_nv_mailbox_put_irq(adev);
+	} else {
+		if (adev->nbio.ras &&
+		    adev->nbio.ras_err_event_athub_irq.funcs)
+			amdgpu_irq_put(adev, &adev->nbio.ras_err_event_athub_irq, 0);
+	}
+
+	return 0;
+}
+
+static int soc21_common_suspend(struct amdgpu_ip_block *ip_block)
+{
+	return soc21_common_hw_fini(ip_block);
+}
+
+static bool soc21_need_reset_on_resume(struct amdgpu_device *adev)
+{
+	u32 sol_reg1, sol_reg2;
+
+	/* Will reset for the following suspend abort cases.
+	 * 1) Only reset dGPU side.
+	 * 2) S3 suspend got aborted and TOS is active.
+	 *    As for dGPU suspend abort cases the SOL value
+	 *    will be kept as zero at this resume point.
+	 */
+	if (!(adev->flags & AMD_IS_APU) && adev->in_s3) {
+		sol_reg1 = RREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_81);
+		msleep(100);
+		sol_reg2 = RREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_81);
+
+		return (sol_reg1 != sol_reg2);
+	}
+
+	return false;
+}
+
+static int soc21_common_resume(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	if (soc21_need_reset_on_resume(adev)) {
+		dev_info(adev->dev, "S3 suspend aborted, resetting...");
+		soc21_asic_reset(adev);
+	}
+
+	return soc21_common_hw_init(ip_block);
+}
+
+static bool soc21_common_is_idle(struct amdgpu_ip_block *ip_block)
+{
+	return true;
+}
+
+static int soc21_common_set_clockgating_state(struct amdgpu_ip_block *ip_block,
+					   enum amd_clockgating_state state)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	switch (amdgpu_ip_version(adev, NBIO_HWIP, 0)) {
+	case IP_VERSION(4, 3, 0):
+	case IP_VERSION(4, 3, 1):
+	case IP_VERSION(7, 7, 0):
+	case IP_VERSION(7, 7, 1):
+	case IP_VERSION(7, 11, 0):
+	case IP_VERSION(7, 11, 1):
+	case IP_VERSION(7, 11, 2):
+	case IP_VERSION(7, 11, 3):
+		adev->nbio.funcs->update_medium_grain_clock_gating(adev,
+				state == AMD_CG_STATE_GATE);
+		adev->nbio.funcs->update_medium_grain_light_sleep(adev,
+				state == AMD_CG_STATE_GATE);
+		adev->hdp.funcs->update_clock_gating(adev,
+				state == AMD_CG_STATE_GATE);
+		break;
+	default:
+		break;
+	}
+	return 0;
+}
+
+static int soc21_common_set_powergating_state(struct amdgpu_ip_block *ip_block,
+					   enum amd_powergating_state state)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	switch (amdgpu_ip_version(adev, LSDMA_HWIP, 0)) {
+	case IP_VERSION(6, 0, 0):
+	case IP_VERSION(6, 0, 2):
+		adev->lsdma.funcs->update_memory_power_gating(adev,
+				state == AMD_PG_STATE_GATE);
+		break;
+	default:
+		break;
+	}
+
+	return 0;
+}
+
+static void soc21_common_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	adev->nbio.funcs->get_clockgating_state(adev, flags);
+
+	adev->hdp.funcs->get_clock_gating_state(adev, flags);
+}
+
+static const struct amd_ip_funcs soc21_common_ip_funcs = {
+	.name = "soc21_common",
+	.early_init = soc21_common_early_init,
+	.late_init = soc21_common_late_init,
+	.sw_init = soc21_common_sw_init,
+	.hw_init = soc21_common_hw_init,
+	.hw_fini = soc21_common_hw_fini,
+	.suspend = soc21_common_suspend,
+	.resume = soc21_common_resume,
+	.is_idle = soc21_common_is_idle,
+	.set_clockgating_state = soc21_common_set_clockgating_state,
+	.set_powergating_state = soc21_common_set_powergating_state,
+	.get_clockgating_state = soc21_common_get_clockgating_state,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/soc21.h b/drivers/gpu/drm/amd/amdgpu/soc21.h
new file mode 100644
index 000000000000..4c8067af1b65
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/soc21.h
@@ -0,0 +1,30 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __SOC21_H__
+#define __SOC21_H__
+
+extern const struct amdgpu_ip_block_version soc21_common_ip_block;
+
+void soc21_grbm_select(struct amdgpu_device *adev,
+		    u32 me, u32 pipe, u32 queue, u32 vmid);
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/soc24.c b/drivers/gpu/drm/amd/amdgpu/soc24.c
new file mode 100644
index 000000000000..972b449ab89f
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/soc24.c
@@ -0,0 +1,601 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include <linux/firmware.h>
+#include <linux/slab.h>
+#include <linux/module.h>
+#include <linux/pci.h>
+
+#include "amdgpu.h"
+#include "amdgpu_ih.h"
+#include "amdgpu_uvd.h"
+#include "amdgpu_vce.h"
+#include "amdgpu_ucode.h"
+#include "amdgpu_psp.h"
+#include "amdgpu_smu.h"
+#include "atom.h"
+#include "amd_pcie.h"
+
+#include "gc/gc_12_0_0_offset.h"
+#include "gc/gc_12_0_0_sh_mask.h"
+#include "mp/mp_14_0_2_offset.h"
+
+#include "soc15.h"
+#include "soc15_common.h"
+#include "soc24.h"
+#include "mxgpu_nv.h"
+
+static const struct amd_ip_funcs soc24_common_ip_funcs;
+
+static const struct amdgpu_video_codec_info vcn_5_0_0_video_codecs_encode_array_vcn0[] = {
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4096, 0)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC, 8192, 4352, 0)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_AV1, 8192, 4352, 0)},
+};
+
+static const struct amdgpu_video_codecs vcn_5_0_0_video_codecs_encode_vcn0 = {
+	.codec_count = ARRAY_SIZE(vcn_5_0_0_video_codecs_encode_array_vcn0),
+	.codec_array = vcn_5_0_0_video_codecs_encode_array_vcn0,
+};
+
+static const struct amdgpu_video_codec_info vcn_5_0_0_video_codecs_decode_array_vcn0[] = {
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4096, 52)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC, 8192, 4352, 186)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG, 16384, 16384, 0)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VP9, 8192, 4352, 0)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_AV1, 8192, 4352, 0)},
+};
+
+static const struct amdgpu_video_codecs vcn_5_0_0_video_codecs_decode_vcn0 = {
+	.codec_count = ARRAY_SIZE(vcn_5_0_0_video_codecs_decode_array_vcn0),
+	.codec_array = vcn_5_0_0_video_codecs_decode_array_vcn0,
+};
+
+static int soc24_query_video_codecs(struct amdgpu_device *adev, bool encode,
+				 const struct amdgpu_video_codecs **codecs)
+{
+	if (adev->vcn.num_vcn_inst == hweight8(adev->vcn.harvest_config))
+		return -EINVAL;
+
+	switch (amdgpu_ip_version(adev, UVD_HWIP, 0)) {
+	case IP_VERSION(5, 0, 0):
+		if (encode)
+			*codecs = &vcn_5_0_0_video_codecs_encode_vcn0;
+		else
+			*codecs = &vcn_5_0_0_video_codecs_decode_vcn0;
+		return 0;
+	default:
+		return -EINVAL;
+	}
+}
+
+static u32 soc24_get_config_memsize(struct amdgpu_device *adev)
+{
+	return adev->nbio.funcs->get_memsize(adev);
+}
+
+static u32 soc24_get_xclk(struct amdgpu_device *adev)
+{
+	return adev->clock.spll.reference_freq;
+}
+
+void soc24_grbm_select(struct amdgpu_device *adev,
+		       u32 me, u32 pipe, u32 queue, u32 vmid)
+{
+	u32 grbm_gfx_cntl = 0;
+	grbm_gfx_cntl = REG_SET_FIELD(grbm_gfx_cntl, GRBM_GFX_CNTL, PIPEID, pipe);
+	grbm_gfx_cntl = REG_SET_FIELD(grbm_gfx_cntl, GRBM_GFX_CNTL, MEID, me);
+	grbm_gfx_cntl = REG_SET_FIELD(grbm_gfx_cntl, GRBM_GFX_CNTL, VMID, vmid);
+	grbm_gfx_cntl = REG_SET_FIELD(grbm_gfx_cntl, GRBM_GFX_CNTL, QUEUEID, queue);
+
+	WREG32_SOC15(GC, 0, regGRBM_GFX_CNTL, grbm_gfx_cntl);
+}
+
+static struct soc15_allowed_register_entry soc24_allowed_read_registers[] = {
+	{ SOC15_REG_ENTRY(GC, 0, regGRBM_STATUS)},
+	{ SOC15_REG_ENTRY(GC, 0, regGRBM_STATUS2)},
+	{ SOC15_REG_ENTRY(GC, 0, regGRBM_STATUS_SE0)},
+	{ SOC15_REG_ENTRY(GC, 0, regGRBM_STATUS_SE1)},
+	{ SOC15_REG_ENTRY(GC, 0, regGRBM_STATUS_SE2)},
+	{ SOC15_REG_ENTRY(GC, 0, regGRBM_STATUS_SE3)},
+	{ SOC15_REG_ENTRY(SDMA0, 0, regSDMA0_STATUS_REG)},
+	{ SOC15_REG_ENTRY(SDMA1, 0, regSDMA1_STATUS_REG)},
+	{ SOC15_REG_ENTRY(GC, 0, regCP_STAT)},
+	{ SOC15_REG_ENTRY(GC, 0, regCP_STALLED_STAT1)},
+	{ SOC15_REG_ENTRY(GC, 0, regCP_STALLED_STAT2)},
+	{ SOC15_REG_ENTRY(GC, 0, regCP_STALLED_STAT3)},
+	{ SOC15_REG_ENTRY(GC, 0, regCP_CPF_BUSY_STAT)},
+	{ SOC15_REG_ENTRY(GC, 0, regCP_CPF_STALLED_STAT1)},
+	{ SOC15_REG_ENTRY(GC, 0, regCP_CPF_STATUS)},
+	{ SOC15_REG_ENTRY(GC, 0, regCP_CPC_BUSY_STAT)},
+	{ SOC15_REG_ENTRY(GC, 0, regCP_CPC_STALLED_STAT1)},
+	{ SOC15_REG_ENTRY(GC, 0, regCP_CPC_STATUS)},
+	{ SOC15_REG_ENTRY(GC, 0, regGB_ADDR_CONFIG)},
+};
+
+static uint32_t soc24_read_indexed_register(struct amdgpu_device *adev,
+					    u32 se_num,
+					    u32 sh_num,
+					    u32 reg_offset)
+{
+	uint32_t val;
+
+	mutex_lock(&adev->grbm_idx_mutex);
+	if (se_num != 0xffffffff || sh_num != 0xffffffff)
+		amdgpu_gfx_select_se_sh(adev, se_num, sh_num, 0xffffffff, 0);
+
+	val = RREG32(reg_offset);
+
+	if (se_num != 0xffffffff || sh_num != 0xffffffff)
+		amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
+	mutex_unlock(&adev->grbm_idx_mutex);
+	return val;
+}
+
+static uint32_t soc24_get_register_value(struct amdgpu_device *adev,
+					 bool indexed, u32 se_num,
+					 u32 sh_num, u32 reg_offset)
+{
+	if (indexed) {
+		return soc24_read_indexed_register(adev, se_num, sh_num, reg_offset);
+	} else {
+		if (reg_offset == SOC15_REG_OFFSET(GC, 0, regGB_ADDR_CONFIG) &&
+		    adev->gfx.config.gb_addr_config)
+			return adev->gfx.config.gb_addr_config;
+		return RREG32(reg_offset);
+	}
+}
+
+static int soc24_read_register(struct amdgpu_device *adev, u32 se_num,
+			       u32 sh_num, u32 reg_offset, u32 *value)
+{
+	uint32_t i;
+	struct soc15_allowed_register_entry  *en;
+
+	*value = 0;
+	for (i = 0; i < ARRAY_SIZE(soc24_allowed_read_registers); i++) {
+		en = &soc24_allowed_read_registers[i];
+		if (!adev->reg_offset[en->hwip][en->inst])
+			continue;
+		else if (reg_offset != (adev->reg_offset[en->hwip][en->inst][en->seg]
+					+ en->reg_offset))
+			continue;
+
+		*value = soc24_get_register_value(adev,
+				soc24_allowed_read_registers[i].grbm_indexed,
+				se_num, sh_num, reg_offset);
+		return 0;
+	}
+	return -EINVAL;
+}
+
+static enum amd_reset_method
+soc24_asic_reset_method(struct amdgpu_device *adev)
+{
+	if (amdgpu_reset_method == AMD_RESET_METHOD_MODE1 ||
+	    amdgpu_reset_method == AMD_RESET_METHOD_MODE2 ||
+	    amdgpu_reset_method == AMD_RESET_METHOD_BACO)
+		return amdgpu_reset_method;
+
+	if (amdgpu_reset_method != -1)
+		dev_warn(adev->dev,
+			 "Specified reset method:%d isn't supported, using AUTO instead.\n",
+			 amdgpu_reset_method);
+
+	switch (amdgpu_ip_version(adev, MP1_HWIP, 0)) {
+	case IP_VERSION(14, 0, 2):
+	case IP_VERSION(14, 0, 3):
+		return AMD_RESET_METHOD_MODE1;
+	default:
+		if (amdgpu_dpm_is_baco_supported(adev))
+			return AMD_RESET_METHOD_BACO;
+		else
+			return AMD_RESET_METHOD_MODE1;
+	}
+}
+
+static int soc24_asic_reset(struct amdgpu_device *adev)
+{
+	int ret = 0;
+
+	switch (soc24_asic_reset_method(adev)) {
+	case AMD_RESET_METHOD_PCI:
+		dev_info(adev->dev, "PCI reset\n");
+		ret = amdgpu_device_pci_reset(adev);
+		break;
+	case AMD_RESET_METHOD_BACO:
+		dev_info(adev->dev, "BACO reset\n");
+		ret = amdgpu_dpm_baco_reset(adev);
+		break;
+	case AMD_RESET_METHOD_MODE2:
+		dev_info(adev->dev, "MODE2 reset\n");
+		ret = amdgpu_dpm_mode2_reset(adev);
+		break;
+	default:
+		dev_info(adev->dev, "MODE1 reset\n");
+		ret = amdgpu_device_mode1_reset(adev);
+		break;
+	}
+
+	return ret;
+}
+
+static void soc24_program_aspm(struct amdgpu_device *adev)
+{
+	if (!amdgpu_device_should_use_aspm(adev))
+		return;
+
+	if (!(adev->flags & AMD_IS_APU) &&
+	    (adev->nbio.funcs->program_aspm))
+		adev->nbio.funcs->program_aspm(adev);
+}
+
+const struct amdgpu_ip_block_version soc24_common_ip_block = {
+	.type = AMD_IP_BLOCK_TYPE_COMMON,
+	.major = 1,
+	.minor = 0,
+	.rev = 0,
+	.funcs = &soc24_common_ip_funcs,
+};
+
+static bool soc24_need_full_reset(struct amdgpu_device *adev)
+{
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	case IP_VERSION(12, 0, 0):
+	case IP_VERSION(12, 0, 1):
+	default:
+		return true;
+	}
+}
+
+static bool soc24_need_reset_on_init(struct amdgpu_device *adev)
+{
+	u32 sol_reg;
+
+	if (adev->flags & AMD_IS_APU)
+		return false;
+
+	/* Check sOS sign of life register to confirm sys driver and sOS
+	 * are already been loaded.
+	 */
+	sol_reg = RREG32_SOC15(MP0, 0, regMPASP_SMN_C2PMSG_81);
+	if (sol_reg)
+		return true;
+
+	return false;
+}
+
+static uint64_t soc24_get_pcie_replay_count(struct amdgpu_device *adev)
+{
+	/* TODO
+	 * dummy implement for pcie_replay_count sysfs interface
+	 * */
+	return 0;
+}
+
+static void soc24_init_doorbell_index(struct amdgpu_device *adev)
+{
+	adev->doorbell_index.kiq = AMDGPU_NAVI10_DOORBELL_KIQ;
+	adev->doorbell_index.mec_ring0 = AMDGPU_NAVI10_DOORBELL_MEC_RING0;
+	adev->doorbell_index.mec_ring1 = AMDGPU_NAVI10_DOORBELL_MEC_RING1;
+	adev->doorbell_index.mec_ring2 = AMDGPU_NAVI10_DOORBELL_MEC_RING2;
+	adev->doorbell_index.mec_ring3 = AMDGPU_NAVI10_DOORBELL_MEC_RING3;
+	adev->doorbell_index.mec_ring4 = AMDGPU_NAVI10_DOORBELL_MEC_RING4;
+	adev->doorbell_index.mec_ring5 = AMDGPU_NAVI10_DOORBELL_MEC_RING5;
+	adev->doorbell_index.mec_ring6 = AMDGPU_NAVI10_DOORBELL_MEC_RING6;
+	adev->doorbell_index.mec_ring7 = AMDGPU_NAVI10_DOORBELL_MEC_RING7;
+	adev->doorbell_index.userqueue_start = AMDGPU_NAVI10_DOORBELL_USERQUEUE_START;
+	adev->doorbell_index.userqueue_end = AMDGPU_NAVI10_DOORBELL_USERQUEUE_END;
+	adev->doorbell_index.gfx_ring0 = AMDGPU_NAVI10_DOORBELL_GFX_RING0;
+	adev->doorbell_index.gfx_ring1 = AMDGPU_NAVI10_DOORBELL_GFX_RING1;
+	adev->doorbell_index.gfx_userqueue_start =
+		AMDGPU_NAVI10_DOORBELL_GFX_USERQUEUE_START;
+	adev->doorbell_index.gfx_userqueue_end =
+		AMDGPU_NAVI10_DOORBELL_GFX_USERQUEUE_END;
+	adev->doorbell_index.mes_ring0 = AMDGPU_NAVI10_DOORBELL_MES_RING0;
+	adev->doorbell_index.mes_ring1 = AMDGPU_NAVI10_DOORBELL_MES_RING1;
+	adev->doorbell_index.sdma_engine[0] = AMDGPU_NAVI10_DOORBELL_sDMA_ENGINE0;
+	adev->doorbell_index.sdma_engine[1] = AMDGPU_NAVI10_DOORBELL_sDMA_ENGINE1;
+	adev->doorbell_index.ih = AMDGPU_NAVI10_DOORBELL_IH;
+	adev->doorbell_index.vcn.vcn_ring0_1 = AMDGPU_NAVI10_DOORBELL64_VCN0_1;
+	adev->doorbell_index.vcn.vcn_ring2_3 = AMDGPU_NAVI10_DOORBELL64_VCN2_3;
+	adev->doorbell_index.vcn.vcn_ring4_5 = AMDGPU_NAVI10_DOORBELL64_VCN4_5;
+	adev->doorbell_index.vcn.vcn_ring6_7 = AMDGPU_NAVI10_DOORBELL64_VCN6_7;
+	adev->doorbell_index.first_non_cp = AMDGPU_NAVI10_DOORBELL64_FIRST_NON_CP;
+	adev->doorbell_index.last_non_cp = AMDGPU_NAVI10_DOORBELL64_LAST_NON_CP;
+
+	adev->doorbell_index.max_assignment = AMDGPU_NAVI10_DOORBELL_MAX_ASSIGNMENT << 1;
+	adev->doorbell_index.sdma_doorbell_range = 20;
+}
+
+static void soc24_pre_asic_init(struct amdgpu_device *adev)
+{
+}
+
+static int soc24_update_umd_stable_pstate(struct amdgpu_device *adev,
+					  bool enter)
+{
+	if (enter)
+		amdgpu_gfx_rlc_enter_safe_mode(adev, 0);
+	else
+		amdgpu_gfx_rlc_exit_safe_mode(adev, 0);
+
+	if (adev->gfx.funcs->update_perfmon_mgcg)
+		adev->gfx.funcs->update_perfmon_mgcg(adev, !enter);
+
+	return 0;
+}
+
+static const struct amdgpu_asic_funcs soc24_asic_funcs = {
+	.read_bios_from_rom = &amdgpu_soc15_read_bios_from_rom,
+	.read_register = &soc24_read_register,
+	.reset = &soc24_asic_reset,
+	.reset_method = &soc24_asic_reset_method,
+	.get_xclk = &soc24_get_xclk,
+	.get_config_memsize = &soc24_get_config_memsize,
+	.init_doorbell_index = &soc24_init_doorbell_index,
+	.need_full_reset = &soc24_need_full_reset,
+	.need_reset_on_init = &soc24_need_reset_on_init,
+	.get_pcie_replay_count = &soc24_get_pcie_replay_count,
+	.supports_baco = &amdgpu_dpm_is_baco_supported,
+	.pre_asic_init = &soc24_pre_asic_init,
+	.query_video_codecs = &soc24_query_video_codecs,
+	.update_umd_stable_pstate = &soc24_update_umd_stable_pstate,
+};
+
+static int soc24_common_early_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	adev->nbio.funcs->set_reg_remap(adev);
+	adev->smc_rreg = NULL;
+	adev->smc_wreg = NULL;
+	adev->pcie_rreg = &amdgpu_device_indirect_rreg;
+	adev->pcie_wreg = &amdgpu_device_indirect_wreg;
+	adev->pcie_rreg64 = &amdgpu_device_indirect_rreg64;
+	adev->pcie_wreg64 = &amdgpu_device_indirect_wreg64;
+	adev->pciep_rreg = amdgpu_device_pcie_port_rreg;
+	adev->pciep_wreg = amdgpu_device_pcie_port_wreg;
+	adev->uvd_ctx_rreg = NULL;
+	adev->uvd_ctx_wreg = NULL;
+	adev->didt_rreg = NULL;
+	adev->didt_wreg = NULL;
+
+	adev->asic_funcs = &soc24_asic_funcs;
+
+	adev->rev_id = amdgpu_device_get_rev_id(adev);
+	adev->external_rev_id = 0xff;
+
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	case IP_VERSION(12, 0, 0):
+		adev->cg_flags = AMD_CG_SUPPORT_GFX_CGCG |
+			AMD_CG_SUPPORT_GFX_CGLS |
+			AMD_CG_SUPPORT_GFX_MGCG |
+			AMD_CG_SUPPORT_GFX_3D_CGCG |
+			AMD_CG_SUPPORT_GFX_3D_CGLS |
+			AMD_CG_SUPPORT_REPEATER_FGCG |
+			AMD_CG_SUPPORT_GFX_FGCG |
+			AMD_CG_SUPPORT_GFX_PERF_CLK |
+			AMD_CG_SUPPORT_ATHUB_MGCG |
+			AMD_CG_SUPPORT_ATHUB_LS |
+			AMD_CG_SUPPORT_MC_MGCG |
+			AMD_CG_SUPPORT_HDP_SD |
+			AMD_CG_SUPPORT_MC_LS;
+		adev->pg_flags = AMD_PG_SUPPORT_VCN |
+			AMD_PG_SUPPORT_JPEG |
+			AMD_PG_SUPPORT_VCN_DPG;
+		adev->external_rev_id = adev->rev_id + 0x40;
+		break;
+	case IP_VERSION(12, 0, 1):
+		adev->cg_flags = AMD_CG_SUPPORT_GFX_CGCG |
+			AMD_CG_SUPPORT_GFX_CGLS |
+			AMD_CG_SUPPORT_GFX_MGCG |
+			AMD_CG_SUPPORT_GFX_3D_CGCG |
+			AMD_CG_SUPPORT_GFX_3D_CGLS |
+			AMD_CG_SUPPORT_REPEATER_FGCG |
+			AMD_CG_SUPPORT_GFX_FGCG |
+			AMD_CG_SUPPORT_GFX_PERF_CLK |
+			AMD_CG_SUPPORT_ATHUB_MGCG |
+			AMD_CG_SUPPORT_ATHUB_LS |
+			AMD_CG_SUPPORT_MC_MGCG |
+			AMD_CG_SUPPORT_HDP_SD |
+			AMD_CG_SUPPORT_MC_LS;
+
+		adev->pg_flags = AMD_PG_SUPPORT_VCN |
+			AMD_PG_SUPPORT_JPEG |
+			AMD_PG_SUPPORT_JPEG_DPG |
+			AMD_PG_SUPPORT_VCN_DPG;
+		adev->external_rev_id = adev->rev_id + 0x50;
+		break;
+	default:
+		/* FIXME: not supported yet */
+		return -EINVAL;
+	}
+
+	if (amdgpu_sriov_vf(adev)) {
+		amdgpu_virt_init_setting(adev);
+		xgpu_nv_mailbox_set_irq_funcs(adev);
+	}
+
+	return 0;
+}
+
+static int soc24_common_late_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	if (amdgpu_sriov_vf(adev)) {
+		xgpu_nv_mailbox_get_irq(adev);
+	} else {
+		if (adev->nbio.ras &&
+		    adev->nbio.ras_err_event_athub_irq.funcs)
+			/* don't need to fail gpu late init
+			 * if enabling athub_err_event interrupt failed
+			 * nbif v6_3_1 only support fatal error hanlding
+			 * just enable the interrupt directly
+			 */
+			amdgpu_irq_get(adev, &adev->nbio.ras_err_event_athub_irq, 0);
+	}
+
+	/* Enable selfring doorbell aperture late because doorbell BAR
+	 * aperture will change if resize BAR successfully in gmc sw_init.
+	 */
+	adev->nbio.funcs->enable_doorbell_selfring_aperture(adev, true);
+
+	return 0;
+}
+
+static int soc24_common_sw_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	if (amdgpu_sriov_vf(adev))
+		xgpu_nv_mailbox_add_irq_id(adev);
+
+	return 0;
+}
+
+static int soc24_common_hw_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	/* enable aspm */
+	soc24_program_aspm(adev);
+	/* setup nbio registers */
+	adev->nbio.funcs->init_registers(adev);
+	/* remap HDP registers to a hole in mmio space,
+	 * for the purpose of expose those registers
+	 * to process space
+	 */
+	if (adev->nbio.funcs->remap_hdp_registers)
+		adev->nbio.funcs->remap_hdp_registers(adev);
+
+	if (adev->df.funcs->hw_init)
+		adev->df.funcs->hw_init(adev);
+
+	/* enable the doorbell aperture */
+	adev->nbio.funcs->enable_doorbell_aperture(adev, true);
+
+	return 0;
+}
+
+static int soc24_common_hw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	/* Disable the doorbell aperture and selfring doorbell aperture
+	 * separately in hw_fini because soc21_enable_doorbell_aperture
+	 * has been removed and there is no need to delay disabling
+	 * selfring doorbell.
+	 */
+	adev->nbio.funcs->enable_doorbell_aperture(adev, false);
+	adev->nbio.funcs->enable_doorbell_selfring_aperture(adev, false);
+
+	if (amdgpu_sriov_vf(adev)) {
+		xgpu_nv_mailbox_put_irq(adev);
+	} else {
+		if (adev->nbio.ras &&
+		    adev->nbio.ras_err_event_athub_irq.funcs)
+			amdgpu_irq_put(adev, &adev->nbio.ras_err_event_athub_irq, 0);
+	}
+
+	return 0;
+}
+
+static int soc24_common_suspend(struct amdgpu_ip_block *ip_block)
+{
+	return soc24_common_hw_fini(ip_block);
+}
+
+static int soc24_common_resume(struct amdgpu_ip_block *ip_block)
+{
+	return soc24_common_hw_init(ip_block);
+}
+
+static bool soc24_common_is_idle(struct amdgpu_ip_block *ip_block)
+{
+	return true;
+}
+
+static int soc24_common_set_clockgating_state(struct amdgpu_ip_block *ip_block,
+					      enum amd_clockgating_state state)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	switch (amdgpu_ip_version(adev, NBIO_HWIP, 0)) {
+	case IP_VERSION(6, 3, 1):
+		adev->nbio.funcs->update_medium_grain_clock_gating(adev,
+				state == AMD_CG_STATE_GATE);
+		adev->nbio.funcs->update_medium_grain_light_sleep(adev,
+				state == AMD_CG_STATE_GATE);
+		adev->hdp.funcs->update_clock_gating(adev,
+				state == AMD_CG_STATE_GATE);
+		break;
+	default:
+		break;
+	}
+	return 0;
+}
+
+static int soc24_common_set_powergating_state(struct amdgpu_ip_block *ip_block,
+					      enum amd_powergating_state state)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	switch (amdgpu_ip_version(adev, LSDMA_HWIP, 0)) {
+	case IP_VERSION(7, 0, 0):
+	case IP_VERSION(7, 0, 1):
+		adev->lsdma.funcs->update_memory_power_gating(adev,
+				state == AMD_PG_STATE_GATE);
+		break;
+	default:
+		break;
+	}
+
+	return 0;
+}
+
+static void soc24_common_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	adev->nbio.funcs->get_clockgating_state(adev, flags);
+
+	adev->hdp.funcs->get_clock_gating_state(adev, flags);
+
+	return;
+}
+
+static const struct amd_ip_funcs soc24_common_ip_funcs = {
+	.name = "soc24_common",
+	.early_init = soc24_common_early_init,
+	.late_init = soc24_common_late_init,
+	.sw_init = soc24_common_sw_init,
+	.hw_init = soc24_common_hw_init,
+	.hw_fini = soc24_common_hw_fini,
+	.suspend = soc24_common_suspend,
+	.resume = soc24_common_resume,
+	.is_idle = soc24_common_is_idle,
+	.set_clockgating_state = soc24_common_set_clockgating_state,
+	.set_powergating_state = soc24_common_set_powergating_state,
+	.get_clockgating_state = soc24_common_get_clockgating_state,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/soc24.h b/drivers/gpu/drm/amd/amdgpu/soc24.h
new file mode 100644
index 000000000000..fa7e442e0b62
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/soc24.h
@@ -0,0 +1,30 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __SOC24_H__
+#define __SOC24_H__
+
+extern const struct amdgpu_ip_block_version soc24_common_ip_block;
+
+void soc24_grbm_select(struct amdgpu_device *adev,
+		       u32 me, u32 pipe, u32 queue, u32 vmid);
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/ta_ras_if.h b/drivers/gpu/drm/amd/amdgpu/ta_ras_if.h
index 5093826a43d1..8a3f326474e5 100644
--- a/drivers/gpu/drm/amd/amdgpu/ta_ras_if.h
+++ b/drivers/gpu/drm/amd/amdgpu/ta_ras_if.h
@@ -30,12 +30,18 @@
 #define RSP_ID_MASK (1U << 31)
 #define RSP_ID(cmdId) (((uint32_t)(cmdId)) | RSP_ID_MASK)
 
+/* invalid node instance value */
+#define TA_RAS_INV_NODE 0xffff
+
 /* RAS related enumerations */
 /**********************************************************/
 enum ras_command {
 	TA_RAS_COMMAND__ENABLE_FEATURES = 0,
 	TA_RAS_COMMAND__DISABLE_FEATURES,
 	TA_RAS_COMMAND__TRIGGER_ERROR,
+	TA_RAS_COMMAND__QUERY_BLOCK_INFO,
+	TA_RAS_COMMAND__QUERY_SUB_BLOCK_INFO,
+	TA_RAS_COMMAND__QUERY_ADDRESS,
 };
 
 enum ta_ras_status {
@@ -64,7 +70,8 @@ enum ta_ras_status {
 	TA_RAS_STATUS__ERROR_PCS_STATE_ERROR            = 0xA016,
 	TA_RAS_STATUS__ERROR_PCS_STATE_HANG             = 0xA017,
 	TA_RAS_STATUS__ERROR_PCS_STATE_UNKNOWN          = 0xA018,
-	TA_RAS_STATUS__ERROR_UNSUPPORTED_ERROR_INJ      = 0xA019
+	TA_RAS_STATUS__ERROR_UNSUPPORTED_ERROR_INJ      = 0xA019,
+	TA_RAS_STATUS__TEE_ERROR_ACCESS_DENIED          = 0xA01A
 };
 
 enum ta_ras_block {
@@ -83,6 +90,11 @@ enum ta_ras_block {
 	TA_RAS_BLOCK__MP1,
 	TA_RAS_BLOCK__FUSE,
 	TA_RAS_BLOCK__MCA,
+	TA_RAS_BLOCK__VCN,
+	TA_RAS_BLOCK__JPEG,
+	TA_RAS_BLOCK__IH,
+	TA_RAS_BLOCK__MPIO,
+	TA_RAS_BLOCK__MMSCH,
 	TA_NUM_BLOCK_MAX
 };
 
@@ -102,6 +114,19 @@ enum ta_ras_error_type {
 	TA_RAS_ERROR__POISON			= 8,
 };
 
+enum ta_ras_address_type {
+	TA_RAS_MCA_TO_PA,
+	TA_RAS_PA_TO_MCA,
+};
+
+enum ta_ras_nps_mode {
+	TA_RAS_UNKNOWN_MODE = 0,
+	TA_RAS_NPS1_MODE = 1,
+	TA_RAS_NPS2_MODE = 2,
+	TA_RAS_NPS4_MODE = 4,
+	TA_RAS_NPS8_MODE = 8,
+};
+
 /* Input/output structures for RAS commands */
 /**********************************************************/
 
@@ -126,6 +151,30 @@ struct ta_ras_trigger_error_input {
 struct ta_ras_init_flags {
 	uint8_t poison_mode_en;
 	uint8_t dgpu_mode;
+	uint16_t xcc_mask;
+	uint8_t channel_dis_num;
+	uint8_t nps_mode;
+	uint32_t active_umc_mask;
+};
+
+struct ta_ras_mca_addr {
+	uint64_t err_addr;
+	uint32_t ch_inst;
+	uint32_t umc_inst;
+	uint32_t node_inst;
+	uint32_t socket_id;
+};
+
+struct ta_ras_phy_addr {
+	uint64_t pa;
+	uint32_t bank;
+	uint32_t channel_idx;
+};
+
+struct ta_ras_query_address_input {
+	enum ta_ras_address_type addr_type;
+	struct ta_ras_mca_addr ma;
+	struct ta_ras_phy_addr pa;
 };
 
 struct ta_ras_output_flags {
@@ -134,6 +183,13 @@ struct ta_ras_output_flags {
 	uint8_t reg_access_failure_flag;
 };
 
+struct ta_ras_query_address_output {
+	/* don't use the flags here */
+	struct ta_ras_output_flags flags;
+	struct ta_ras_mca_addr ma;
+	struct ta_ras_phy_addr pa;
+};
+
 /* Common input structure for RAS callbacks */
 /**********************************************************/
 union ta_ras_cmd_input {
@@ -141,12 +197,14 @@ union ta_ras_cmd_input {
 	struct ta_ras_enable_features_input	enable_features;
 	struct ta_ras_disable_features_input	disable_features;
 	struct ta_ras_trigger_error_input	trigger_error;
+	struct ta_ras_query_address_input	address;
 
 	uint32_t reserve_pad[256];
 };
 
 union ta_ras_cmd_output {
 	struct ta_ras_output_flags flags;
+	struct ta_ras_query_address_output address;
 
 	uint32_t reserve_pad[256];
 };
diff --git a/drivers/gpu/drm/amd/amdgpu/ta_secureDisplay_if.h b/drivers/gpu/drm/amd/amdgpu/ta_secureDisplay_if.h
index cf8ff064dc72..9ec2e03d41c7 100644
--- a/drivers/gpu/drm/amd/amdgpu/ta_secureDisplay_if.h
+++ b/drivers/gpu/drm/amd/amdgpu/ta_secureDisplay_if.h
@@ -31,10 +31,12 @@
  *    Secure Display Command ID
  */
 enum ta_securedisplay_command {
-	/* Query whether TA is responding used only for validation purpose */
+	/* Query whether TA is responding. It is used only for validation purpose */
 	TA_SECUREDISPLAY_COMMAND__QUERY_TA              = 1,
 	/* Send region of Interest and CRC value to I2C */
 	TA_SECUREDISPLAY_COMMAND__SEND_ROI_CRC          = 2,
+	/* V2 to send multiple regions of Interest and CRC value to I2C */
+	TA_SECUREDISPLAY_COMMAND__SEND_ROI_CRC_V2       = 3,
 	/* Maximum Command ID */
 	TA_SECUREDISPLAY_COMMAND__MAX_ID                = 0x7FFFFFFF,
 };
@@ -55,10 +57,10 @@ enum ta_securedisplay_status {
 	TA_SECUREDISPLAY_STATUS__MAX                     = 0x7FFFFFFF,/* Maximum Value for status*/
 };
 
-/** @enum ta_securedisplay_max_phy
+/** @enum ta_securedisplay_phy_ID
  *    Physical ID number to use for reading corresponding DIO Scratch register for ROI
  */
-enum  ta_securedisplay_max_phy {
+enum  ta_securedisplay_phy_ID {
 	TA_SECUREDISPLAY_PHY0                           = 0,
 	TA_SECUREDISPLAY_PHY1                           = 1,
 	TA_SECUREDISPLAY_PHY2                           = 2,
@@ -83,6 +85,8 @@ enum ta_securedisplay_ta_query_cmd_ret {
 enum ta_securedisplay_buffer_size {
 	/* 15 bytes = 8 byte (ROI) + 6 byte(CRC) + 1 byte(phy_id) */
 	TA_SECUREDISPLAY_I2C_BUFFER_SIZE                = 15,
+	/* 16 bytes = 8 byte (ROI) + 6 byte(CRC) + 1 byte(phy_id) + 1 byte(roi_idx) */
+	TA_SECUREDISPLAY_V2_I2C_BUFFER_SIZE             = 16,
 };
 
 /** Input/output structures for Secure Display commands */
@@ -95,7 +99,15 @@ enum ta_securedisplay_buffer_size {
  *    Physical ID to determine which DIO scratch register should be used to get ROI
  */
 struct ta_securedisplay_send_roi_crc_input {
-	uint32_t  phy_id;  /* Physical ID */
+	/* Physical ID */
+	uint32_t  phy_id;
+};
+
+struct ta_securedisplay_send_roi_crc_v2_input {
+	/* Physical ID */
+	uint32_t phy_id;
+	/* Region of interest index */
+	uint8_t  roi_idx;
 };
 
 /** @union ta_securedisplay_cmd_input
@@ -104,6 +116,8 @@ struct ta_securedisplay_send_roi_crc_input {
 union ta_securedisplay_cmd_input {
 	/* send ROI and CRC input buffer format */
 	struct ta_securedisplay_send_roi_crc_input        send_roi_crc;
+	/* send ROI and CRC input buffer format, v2 adds a ROI index */
+	struct ta_securedisplay_send_roi_crc_v2_input     send_roi_crc_v2;
 	uint32_t                                          reserved[4];
 };
 
@@ -128,6 +142,10 @@ struct ta_securedisplay_send_roi_crc_output {
 	uint8_t  reserved;
 };
 
+struct ta_securedisplay_send_roi_crc_v2_output {
+	uint8_t  i2c_buf[TA_SECUREDISPLAY_V2_I2C_BUFFER_SIZE];  /* I2C buffer */
+};
+
 /** @union ta_securedisplay_cmd_output
  *    Output buffer
  */
@@ -136,19 +154,21 @@ union ta_securedisplay_cmd_output {
 	struct ta_securedisplay_query_ta_output            query_ta;
 	/* Send ROI CRC output buffer format used only for validation purpose */
 	struct ta_securedisplay_send_roi_crc_output        send_roi_crc;
+	/* Send ROI CRC output buffer format used only for validation purpose */
+	struct ta_securedisplay_send_roi_crc_v2_output     send_roi_crc_v2;
 	uint32_t                                           reserved[4];
 };
 
-/** @struct securedisplay_cmd
- *    Secure Display Command which is shared buffer memory
- */
-struct securedisplay_cmd {
-	uint32_t                             cmd_id;                    /* +0  Bytes Command ID */
-	enum ta_securedisplay_status         status;     /* +4  Bytes Status of Secure Display TA */
-	uint32_t                             reserved[2];               /* +8  Bytes Reserved */
-	union ta_securedisplay_cmd_input     securedisplay_in_message;  /* +16 Bytes Input Buffer */
-	union ta_securedisplay_cmd_output    securedisplay_out_message;/* +32 Bytes Output Buffer */
-	/**@note Total 48 Bytes */
+/** @struct ta_securedisplay_cmd
+*    Secure display command which is shared buffer memory
+*/
+struct ta_securedisplay_cmd {
+    uint32_t                                           cmd_id;                         /**< +0  Bytes Command ID */
+    enum ta_securedisplay_status                       status;                         /**< +4  Bytes Status code returned by the secure display TA */
+    uint32_t                                           reserved[2];                    /**< +8  Bytes Reserved */
+    union ta_securedisplay_cmd_input                   securedisplay_in_message;       /**< +16 Bytes Command input buffer */
+    union ta_securedisplay_cmd_output                  securedisplay_out_message;      /**< +32 Bytes Command output buffer */
+    /**@note Total 48 Bytes */
 };
 
 #endif   //_TA_SECUREDISPLAY_IF_H
diff --git a/drivers/gpu/drm/amd/amdgpu/ta_xgmi_if.h b/drivers/gpu/drm/amd/amdgpu/ta_xgmi_if.h
index da815a93d46e..d5748032674e 100644
--- a/drivers/gpu/drm/amd/amdgpu/ta_xgmi_if.h
+++ b/drivers/gpu/drm/amd/amdgpu/ta_xgmi_if.h
@@ -1,5 +1,5 @@
 /*
- * Copyright 2018 Advanced Micro Devices, Inc.
+ * Copyright 2018-2022 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -20,7 +20,6 @@
  * OTHER DEALINGS IN THE SOFTWARE.
  *
  */
-
 #ifndef _TA_XGMI_IF_H
 #define _TA_XGMI_IF_H
 
@@ -28,20 +27,31 @@
 #define RSP_ID_MASK (1U << 31)
 #define RSP_ID(cmdId) (((uint32_t)(cmdId)) | RSP_ID_MASK)
 
+#define EXTEND_PEER_LINK_INFO_CMD_FLAG 1
+
 enum ta_command_xgmi {
+	/* Initialize the Context and Session Topology */
 	TA_COMMAND_XGMI__INITIALIZE			= 0x00,
+	/* Gets the current GPU's node ID */
 	TA_COMMAND_XGMI__GET_NODE_ID			= 0x01,
+	/* Gets the current GPU's hive ID */
 	TA_COMMAND_XGMI__GET_HIVE_ID			= 0x02,
-	TA_COMMAND_XGMI__GET_GET_TOPOLOGY_INFO		= 0x03,
+	/* Gets the Peer's topology Information */
+	TA_COMMAND_XGMI__GET_TOPOLOGY_INFO		= 0x03,
+	/* Sets the Peer's topology Information */
 	TA_COMMAND_XGMI__SET_TOPOLOGY_INFO		= 0x04,
-	TA_COMMAND_XGMI__GET_PEER_LINKS			= 0x0B
+	/* Gets the total links between adjacent peer dies in hive */
+	TA_COMMAND_XGMI__GET_PEER_LINKS			= 0x0B,
+	/* Gets the total links and connected port numbers between adjacent peer dies in hive */
+	TA_COMMAND_XGMI__GET_EXTEND_PEER_LINKS		= 0x0C
 };
 
 /* XGMI related enumerations */
 /**********************************************************/;
-enum ta_xgmi_connected_nodes {
-	TA_XGMI__MAX_CONNECTED_NODES			= 64
-};
+enum { TA_XGMI__MAX_CONNECTED_NODES = 64 };
+enum { TA_XGMI__MAX_INTERNAL_STATE = 32 };
+enum { TA_XGMI__MAX_INTERNAL_STATE_BUFFER = 128 };
+enum { TA_XGMI__MAX_PORT_NUM = 8 };
 
 enum ta_xgmi_status {
 	TA_XGMI_STATUS__SUCCESS				= 0x00,
@@ -81,6 +91,18 @@ struct ta_xgmi_peer_link_info {
 	uint8_t					num_links;
 };
 
+struct xgmi_connected_port_num {
+	uint8_t		dst_xgmi_port_num;
+	uint8_t		src_xgmi_port_num;
+};
+
+/* support both the port num and num_links */
+struct ta_xgmi_extend_peer_link_info {
+	uint64_t				node_id;
+	uint8_t					num_links;
+	struct xgmi_connected_port_num		port_num[TA_XGMI__MAX_PORT_NUM];
+};
+
 struct ta_xgmi_cmd_initialize_output {
 	uint32_t	status;
 };
@@ -103,16 +125,21 @@ struct ta_xgmi_cmd_get_topology_info_output {
 	struct ta_xgmi_node_info	nodes[TA_XGMI__MAX_CONNECTED_NODES];
 };
 
-struct ta_xgmi_cmd_get_peer_link_info_output {
+struct ta_xgmi_cmd_set_topology_info_input {
 	uint32_t			num_nodes;
-	struct ta_xgmi_peer_link_info	nodes[TA_XGMI__MAX_CONNECTED_NODES];
+	struct ta_xgmi_node_info	nodes[TA_XGMI__MAX_CONNECTED_NODES];
 };
 
-struct ta_xgmi_cmd_set_topology_info_input {
+/* support XGMI TA w/ and w/o port_num both so two similar structs defined */
+struct ta_xgmi_cmd_get_peer_link_info {
 	uint32_t			num_nodes;
-	struct ta_xgmi_node_info	nodes[TA_XGMI__MAX_CONNECTED_NODES];
+	struct ta_xgmi_peer_link_info	nodes[TA_XGMI__MAX_CONNECTED_NODES];
 };
 
+struct ta_xgmi_cmd_get_extend_peer_link_info {
+	uint32_t				num_nodes;
+	struct ta_xgmi_extend_peer_link_info nodes[TA_XGMI__MAX_CONNECTED_NODES];
+};
 /**********************************************************/
 /* Common input structure for XGMI callbacks */
 union ta_xgmi_cmd_input {
@@ -126,16 +153,23 @@ union ta_xgmi_cmd_output {
 	struct ta_xgmi_cmd_get_node_id_output		get_node_id;
 	struct ta_xgmi_cmd_get_hive_id_output		get_hive_id;
 	struct ta_xgmi_cmd_get_topology_info_output	get_topology_info;
-	struct ta_xgmi_cmd_get_peer_link_info_output	get_link_info;
+	struct ta_xgmi_cmd_get_peer_link_info		get_link_info;
+	struct ta_xgmi_cmd_get_extend_peer_link_info	get_extend_link_info;
 };
-/**********************************************************/
 
 struct ta_xgmi_shared_memory {
 	uint32_t			cmd_id;
 	uint32_t			resp_id;
 	enum ta_xgmi_status		xgmi_status;
+
+	/* if the number of xgmi link record is more than 128, driver will set the
+	 * flag 0 to get the first 128 of the link records and will set to 1, to get
+	 * the second set
+	 */
 	uint8_t				flag_extend_link_record;
-	uint8_t				reserved0[3];
+	/* bit0: port_num info support flag for GET_EXTEND_PEER_LINKS commmand */
+	uint8_t				caps_flag;
+	uint8_t				reserved[2];
 	union ta_xgmi_cmd_input		xgmi_in_message;
 	union ta_xgmi_cmd_output	xgmi_out_message;
 };
diff --git a/drivers/gpu/drm/amd/amdgpu/tonga_ih.c b/drivers/gpu/drm/amd/amdgpu/tonga_ih.c
index b08905d1c00f..ee8038df17e3 100644
--- a/drivers/gpu/drm/amd/amdgpu/tonga_ih.c
+++ b/drivers/gpu/drm/amd/amdgpu/tonga_ih.c
@@ -159,6 +159,9 @@ static int tonga_ih_irq_init(struct amdgpu_device *adev)
 	/* enable interrupts */
 	tonga_ih_enable_interrupts(adev);
 
+	if (adev->irq.ih_soft.ring_size)
+		adev->irq.ih_soft.enabled = true;
+
 	return 0;
 }
 
@@ -196,6 +199,9 @@ static u32 tonga_ih_get_wptr(struct amdgpu_device *adev,
 
 	wptr = le32_to_cpu(*ih->wptr_cpu);
 
+	if (ih == &adev->irq.ih_soft)
+		goto out;
+
 	if (!REG_GET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW))
 		goto out;
 
@@ -219,6 +225,12 @@ static u32 tonga_ih_get_wptr(struct amdgpu_device *adev,
 	tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 1);
 	WREG32(mmIH_RB_CNTL, tmp);
 
+	/* Unset the CLEAR_OVERFLOW bit immediately so new overflows
+	 * can be detected.
+	 */
+	tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 0);
+	WREG32(mmIH_RB_CNTL, tmp);
+
 out:
 	return (wptr & ih->ptr_mask);
 }
@@ -277,9 +289,9 @@ static void tonga_ih_set_rptr(struct amdgpu_device *adev,
 	}
 }
 
-static int tonga_ih_early_init(void *handle)
+static int tonga_ih_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int ret;
 
 	ret = amdgpu_irq_add_domain(adev);
@@ -291,15 +303,19 @@ static int tonga_ih_early_init(void *handle)
 	return 0;
 }
 
-static int tonga_ih_sw_init(void *handle)
+static int tonga_ih_sw_init(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	r = amdgpu_ih_ring_init(adev, &adev->irq.ih, 64 * 1024, true);
 	if (r)
 		return r;
 
+	r = amdgpu_ih_ring_init(adev, &adev->irq.ih_soft, IH_SW_RING_SIZE, true);
+	if (r)
+		return r;
+
 	adev->irq.ih.use_doorbell = true;
 	adev->irq.ih.doorbell_index = adev->doorbell_index.ih;
 
@@ -308,9 +324,9 @@ static int tonga_ih_sw_init(void *handle)
 	return r;
 }
 
-static int tonga_ih_sw_fini(void *handle)
+static int tonga_ih_sw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	amdgpu_irq_fini_sw(adev);
 	amdgpu_irq_remove_domain(adev);
@@ -318,10 +334,10 @@ static int tonga_ih_sw_fini(void *handle)
 	return 0;
 }
 
-static int tonga_ih_hw_init(void *handle)
+static int tonga_ih_hw_init(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	r = tonga_ih_irq_init(adev);
 	if (r)
@@ -330,32 +346,26 @@ static int tonga_ih_hw_init(void *handle)
 	return 0;
 }
 
-static int tonga_ih_hw_fini(void *handle)
+static int tonga_ih_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	tonga_ih_irq_disable(adev);
+	tonga_ih_irq_disable(ip_block->adev);
 
 	return 0;
 }
 
-static int tonga_ih_suspend(void *handle)
+static int tonga_ih_suspend(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	return tonga_ih_hw_fini(adev);
+	return tonga_ih_hw_fini(ip_block);
 }
 
-static int tonga_ih_resume(void *handle)
+static int tonga_ih_resume(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	return tonga_ih_hw_init(adev);
+	return tonga_ih_hw_init(ip_block);
 }
 
-static bool tonga_ih_is_idle(void *handle)
+static bool tonga_ih_is_idle(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	u32 tmp = RREG32(mmSRBM_STATUS);
 
 	if (REG_GET_FIELD(tmp, SRBM_STATUS, IH_BUSY))
@@ -364,11 +374,11 @@ static bool tonga_ih_is_idle(void *handle)
 	return true;
 }
 
-static int tonga_ih_wait_for_idle(void *handle)
+static int tonga_ih_wait_for_idle(struct amdgpu_ip_block *ip_block)
 {
 	unsigned i;
 	u32 tmp;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	for (i = 0; i < adev->usec_timeout; i++) {
 		/* read MC_STATUS */
@@ -380,9 +390,9 @@ static int tonga_ih_wait_for_idle(void *handle)
 	return -ETIMEDOUT;
 }
 
-static bool tonga_ih_check_soft_reset(void *handle)
+static bool tonga_ih_check_soft_reset(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	u32 srbm_soft_reset = 0;
 	u32 tmp = RREG32(mmSRBM_STATUS);
 
@@ -399,29 +409,27 @@ static bool tonga_ih_check_soft_reset(void *handle)
 	}
 }
 
-static int tonga_ih_pre_soft_reset(void *handle)
+static int tonga_ih_pre_soft_reset(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	if (!adev->irq.srbm_soft_reset)
+	if (!ip_block->adev->irq.srbm_soft_reset)
 		return 0;
 
-	return tonga_ih_hw_fini(adev);
+	return tonga_ih_hw_fini(ip_block);
 }
 
-static int tonga_ih_post_soft_reset(void *handle)
+static int tonga_ih_post_soft_reset(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (!adev->irq.srbm_soft_reset)
 		return 0;
 
-	return tonga_ih_hw_init(adev);
+	return tonga_ih_hw_init(ip_block);
 }
 
-static int tonga_ih_soft_reset(void *handle)
+static int tonga_ih_soft_reset(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	u32 srbm_soft_reset;
 
 	if (!adev->irq.srbm_soft_reset)
@@ -450,13 +458,13 @@ static int tonga_ih_soft_reset(void *handle)
 	return 0;
 }
 
-static int tonga_ih_set_clockgating_state(void *handle,
+static int tonga_ih_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_clockgating_state state)
 {
 	return 0;
 }
 
-static int tonga_ih_set_powergating_state(void *handle,
+static int tonga_ih_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_powergating_state state)
 {
 	return 0;
@@ -465,7 +473,6 @@ static int tonga_ih_set_powergating_state(void *handle,
 static const struct amd_ip_funcs tonga_ih_ip_funcs = {
 	.name = "tonga_ih",
 	.early_init = tonga_ih_early_init,
-	.late_init = NULL,
 	.sw_init = tonga_ih_sw_init,
 	.sw_fini = tonga_ih_sw_fini,
 	.hw_init = tonga_ih_hw_init,
@@ -493,8 +500,7 @@ static void tonga_ih_set_interrupt_funcs(struct amdgpu_device *adev)
 	adev->irq.ih_funcs = &tonga_ih_funcs;
 }
 
-const struct amdgpu_ip_block_version tonga_ih_ip_block =
-{
+const struct amdgpu_ip_block_version tonga_ih_ip_block = {
 	.type = AMD_IP_BLOCK_TYPE_IH,
 	.major = 3,
 	.minor = 0,
diff --git a/drivers/gpu/drm/amd/amdgpu/umc_v12_0.c b/drivers/gpu/drm/amd/amdgpu/umc_v12_0.c
new file mode 100644
index 000000000000..0f5b1719fda5
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/umc_v12_0.c
@@ -0,0 +1,742 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include "umc_v12_0.h"
+#include "amdgpu_ras.h"
+#include "amdgpu_umc.h"
+#include "amdgpu.h"
+#include "umc/umc_12_0_0_offset.h"
+#include "umc/umc_12_0_0_sh_mask.h"
+#include "mp/mp_13_0_6_sh_mask.h"
+
+#define MAX_ECC_NUM_PER_RETIREMENT  32
+#define DELAYED_TIME_FOR_GPU_RESET  1000  //ms
+
+static inline uint64_t get_umc_v12_0_reg_offset(struct amdgpu_device *adev,
+					    uint32_t node_inst,
+					    uint32_t umc_inst,
+					    uint32_t ch_inst)
+{
+	uint32_t index = umc_inst * adev->umc.channel_inst_num + ch_inst;
+	uint64_t cross_node_offset = (node_inst == 0) ? 0 : UMC_V12_0_CROSS_NODE_OFFSET;
+
+	umc_inst = index / 4;
+	ch_inst = index % 4;
+
+	return adev->umc.channel_offs * ch_inst + UMC_V12_0_INST_DIST * umc_inst +
+		UMC_V12_0_NODE_DIST * node_inst + cross_node_offset;
+}
+
+static int umc_v12_0_reset_error_count_per_channel(struct amdgpu_device *adev,
+					uint32_t node_inst, uint32_t umc_inst,
+					uint32_t ch_inst, void *data)
+{
+	uint64_t odecc_err_cnt_addr;
+	uint64_t umc_reg_offset =
+		get_umc_v12_0_reg_offset(adev, node_inst, umc_inst, ch_inst);
+
+	odecc_err_cnt_addr =
+		SOC15_REG_OFFSET(UMC, 0, regUMCCH0_OdEccErrCnt);
+
+	/* clear error count */
+	WREG32_PCIE_EXT((odecc_err_cnt_addr + umc_reg_offset) * 4,
+			UMC_V12_0_CE_CNT_INIT);
+
+	return 0;
+}
+
+static void umc_v12_0_reset_error_count(struct amdgpu_device *adev)
+{
+	amdgpu_umc_loop_channels(adev,
+		umc_v12_0_reset_error_count_per_channel, NULL);
+}
+
+bool umc_v12_0_is_deferred_error(struct amdgpu_device *adev, uint64_t mc_umc_status)
+{
+	dev_dbg(adev->dev,
+		"MCA_UMC_STATUS(0x%llx): Val:%llu, Poison:%llu, Deferred:%llu, PCC:%llu, UC:%llu, TCC:%llu\n",
+		mc_umc_status,
+		REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Val),
+		REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Poison),
+		REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Deferred),
+		REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, PCC),
+		REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UC),
+		REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, TCC)
+	);
+
+	return (amdgpu_ras_is_poison_mode_supported(adev) &&
+		(REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Val) == 1) &&
+		((REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Deferred) == 1) ||
+		(REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Poison) == 1)));
+}
+
+bool umc_v12_0_is_uncorrectable_error(struct amdgpu_device *adev, uint64_t mc_umc_status)
+{
+	if (umc_v12_0_is_deferred_error(adev, mc_umc_status))
+		return false;
+
+	return ((REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Val) == 1) &&
+		(REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, PCC) == 1 ||
+		REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UC) == 1 ||
+		REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, TCC) == 1));
+}
+
+bool umc_v12_0_is_correctable_error(struct amdgpu_device *adev, uint64_t mc_umc_status)
+{
+	if (umc_v12_0_is_deferred_error(adev, mc_umc_status))
+		return false;
+
+	return (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Val) == 1 &&
+		(REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, CECC) == 1 ||
+		(REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UECC) == 1 &&
+		REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UC) == 0) ||
+		/* Identify data parity error in replay mode */
+		((REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, ErrorCodeExt) == 0x5 ||
+		REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, ErrorCodeExt) == 0xb) &&
+		!(umc_v12_0_is_uncorrectable_error(adev, mc_umc_status)))));
+}
+
+static void umc_v12_0_query_error_count_per_type(struct amdgpu_device *adev,
+						   uint64_t umc_reg_offset,
+						   unsigned long *error_count,
+						   check_error_type_func error_type_func)
+{
+	uint64_t mc_umc_status;
+	uint64_t mc_umc_status_addr;
+
+	mc_umc_status_addr =
+		SOC15_REG_OFFSET(UMC, 0, regMCA_UMC_UMC0_MCUMC_STATUST0);
+
+	/* Check MCUMC_STATUS */
+	mc_umc_status =
+		RREG64_PCIE_EXT((mc_umc_status_addr + umc_reg_offset) * 4);
+
+	if (error_type_func(adev, mc_umc_status))
+		*error_count += 1;
+}
+
+static int umc_v12_0_query_error_count(struct amdgpu_device *adev,
+					uint32_t node_inst, uint32_t umc_inst,
+					uint32_t ch_inst, void *data)
+{
+	struct ras_err_data *err_data = (struct ras_err_data *)data;
+	unsigned long ue_count = 0, ce_count = 0, de_count = 0;
+
+	/* NOTE: node_inst is converted by adev->umc.active_mask and the range is [0-3],
+	 * which can be used as die ID directly */
+	struct amdgpu_smuio_mcm_config_info mcm_info = {
+		.socket_id = adev->smuio.funcs->get_socket_id(adev),
+		.die_id = node_inst,
+	};
+
+	uint64_t umc_reg_offset =
+		get_umc_v12_0_reg_offset(adev, node_inst, umc_inst, ch_inst);
+
+	umc_v12_0_query_error_count_per_type(adev, umc_reg_offset,
+					    &ce_count, umc_v12_0_is_correctable_error);
+	umc_v12_0_query_error_count_per_type(adev, umc_reg_offset,
+					    &ue_count, umc_v12_0_is_uncorrectable_error);
+	umc_v12_0_query_error_count_per_type(adev, umc_reg_offset,
+					    &de_count, umc_v12_0_is_deferred_error);
+
+	amdgpu_ras_error_statistic_ue_count(err_data, &mcm_info, ue_count);
+	amdgpu_ras_error_statistic_ce_count(err_data, &mcm_info, ce_count);
+	amdgpu_ras_error_statistic_de_count(err_data, &mcm_info, de_count);
+
+	return 0;
+}
+
+static void umc_v12_0_query_ras_error_count(struct amdgpu_device *adev,
+					   void *ras_error_status)
+{
+	amdgpu_umc_loop_channels(adev,
+		umc_v12_0_query_error_count, ras_error_status);
+
+	umc_v12_0_reset_error_count(adev);
+}
+
+static void umc_v12_0_get_retire_flip_bits(struct amdgpu_device *adev)
+{
+	enum amdgpu_memory_partition nps = AMDGPU_NPS1_PARTITION_MODE;
+	uint32_t vram_type = adev->gmc.vram_type;
+	struct amdgpu_umc_flip_bits *flip_bits = &(adev->umc.flip_bits);
+
+	if (adev->gmc.gmc_funcs->query_mem_partition_mode)
+		nps = adev->gmc.gmc_funcs->query_mem_partition_mode(adev);
+
+	/* default setting */
+	flip_bits->flip_bits_in_pa[0] = UMC_V12_0_PA_C2_BIT;
+	flip_bits->flip_bits_in_pa[1] = UMC_V12_0_PA_C3_BIT;
+	flip_bits->flip_bits_in_pa[2] = UMC_V12_0_PA_C4_BIT;
+	flip_bits->flip_bits_in_pa[3] = UMC_V12_0_PA_R13_BIT;
+	flip_bits->flip_row_bit = 13;
+	flip_bits->bit_num = 4;
+	flip_bits->r13_in_pa = UMC_V12_0_PA_R13_BIT;
+
+	if (nps == AMDGPU_NPS2_PARTITION_MODE) {
+		flip_bits->flip_bits_in_pa[0] = UMC_V12_0_PA_CH5_BIT;
+		flip_bits->flip_bits_in_pa[1] = UMC_V12_0_PA_C2_BIT;
+		flip_bits->flip_bits_in_pa[2] = UMC_V12_0_PA_B1_BIT;
+		flip_bits->r13_in_pa = UMC_V12_0_PA_R12_BIT;
+	} else if (nps == AMDGPU_NPS4_PARTITION_MODE) {
+		flip_bits->flip_bits_in_pa[0] = UMC_V12_0_PA_CH4_BIT;
+		flip_bits->flip_bits_in_pa[1] = UMC_V12_0_PA_CH5_BIT;
+		flip_bits->flip_bits_in_pa[2] = UMC_V12_0_PA_B0_BIT;
+		flip_bits->r13_in_pa = UMC_V12_0_PA_R11_BIT;
+	}
+
+	switch (vram_type) {
+	case AMDGPU_VRAM_TYPE_HBM:
+		/* other nps modes are taken as nps1 */
+		if (nps == AMDGPU_NPS2_PARTITION_MODE)
+			flip_bits->flip_bits_in_pa[3] = UMC_V12_0_PA_R12_BIT;
+		else if (nps == AMDGPU_NPS4_PARTITION_MODE)
+			flip_bits->flip_bits_in_pa[3] = UMC_V12_0_PA_R11_BIT;
+
+		break;
+	case AMDGPU_VRAM_TYPE_HBM3E:
+		flip_bits->flip_bits_in_pa[3] = UMC_V12_0_PA_R12_BIT;
+		flip_bits->flip_row_bit = 12;
+
+		if (nps == AMDGPU_NPS2_PARTITION_MODE)
+			flip_bits->flip_bits_in_pa[3] = UMC_V12_0_PA_R11_BIT;
+		else if (nps == AMDGPU_NPS4_PARTITION_MODE)
+			flip_bits->flip_bits_in_pa[3] = UMC_V12_0_PA_R10_BIT;
+
+		break;
+	default:
+		dev_warn(adev->dev,
+			"Unknown HBM type, set RAS retire flip bits to the value in NPS1 mode.\n");
+		break;
+	}
+
+	adev->umc.retire_unit = 0x1 << flip_bits->bit_num;
+}
+
+static int umc_v12_0_convert_error_address(struct amdgpu_device *adev,
+					struct ras_err_data *err_data,
+					struct ta_ras_query_address_input *addr_in,
+					struct ta_ras_query_address_output *addr_out,
+					bool dump_addr)
+{
+	uint32_t col, col_lower, row, row_lower, row_high, bank;
+	uint32_t channel_index = 0, umc_inst = 0;
+	uint32_t i, bit_num, retire_unit, *flip_bits;
+	uint64_t soc_pa, column, err_addr;
+	struct ta_ras_query_address_output addr_out_tmp;
+	struct ta_ras_query_address_output *paddr_out;
+	int ret = 0;
+
+	if (!addr_out)
+		paddr_out = &addr_out_tmp;
+	else
+		paddr_out = addr_out;
+
+	err_addr = bank = 0;
+	if (addr_in) {
+		err_addr = addr_in->ma.err_addr;
+		addr_in->addr_type = TA_RAS_MCA_TO_PA;
+		ret = psp_ras_query_address(&adev->psp, addr_in, paddr_out);
+		if (ret) {
+			dev_warn(adev->dev, "Failed to query RAS physical address for 0x%llx",
+				err_addr);
+
+			goto out;
+		}
+
+		bank = paddr_out->pa.bank;
+		/* no need to care about umc inst if addr_in is NULL */
+		umc_inst = addr_in->ma.umc_inst;
+	}
+
+	flip_bits = adev->umc.flip_bits.flip_bits_in_pa;
+	bit_num = adev->umc.flip_bits.bit_num;
+	retire_unit = adev->umc.retire_unit;
+
+	soc_pa = paddr_out->pa.pa;
+	channel_index = paddr_out->pa.channel_idx;
+	/* clear loop bits in soc physical address */
+	for (i = 0; i < bit_num; i++)
+		soc_pa &= ~BIT_ULL(flip_bits[i]);
+
+	paddr_out->pa.pa = soc_pa;
+	/* get column bit 0 and 1 in mca address */
+	col_lower = (err_addr >> 1) & 0x3ULL;
+	/* extra row bit will be handled later */
+	row_lower = (err_addr >> UMC_V12_0_MA_R0_BIT) & 0x1fffULL;
+	row_lower &= ~BIT_ULL(adev->umc.flip_bits.flip_row_bit);
+
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) >= IP_VERSION(9, 5, 0)) {
+		row_high = (soc_pa >> adev->umc.flip_bits.r13_in_pa) & 0x3ULL;
+		/* it's 2.25GB in each channel, from MCA address to PA
+		 * [R14 R13] is converted if the two bits value are 0x3,
+		 * get them from PA instead of MCA address.
+		 */
+		row_lower |= (row_high << 13);
+	}
+
+	if (!err_data && !dump_addr)
+		goto out;
+
+	/* loop for all possibilities of retired bits */
+	for (column = 0; column < retire_unit; column++) {
+		soc_pa = paddr_out->pa.pa;
+		for (i = 0; i < bit_num; i++)
+			soc_pa |= (((column >> i) & 0x1ULL) << flip_bits[i]);
+
+		col = ((column & 0x7) << 2) | col_lower;
+		/* handle extra row bit */
+		if (bit_num == RETIRE_FLIP_BITS_NUM)
+			row = ((column >> 3) << adev->umc.flip_bits.flip_row_bit) |
+					row_lower;
+
+		if (dump_addr)
+			dev_info(adev->dev,
+				"Error Address(PA):0x%-10llx Row:0x%-4x Col:0x%-2x Bank:0x%x Channel:0x%x\n",
+				soc_pa, row, col, bank, channel_index);
+
+		if (err_data)
+			amdgpu_umc_fill_error_record(err_data, err_addr,
+				soc_pa, channel_index, umc_inst);
+	}
+
+out:
+	return ret;
+}
+
+static int umc_v12_0_query_error_address(struct amdgpu_device *adev,
+					uint32_t node_inst, uint32_t umc_inst,
+					uint32_t ch_inst, void *data)
+{
+	struct ras_err_data *err_data = (struct ras_err_data *)data;
+	struct ta_ras_query_address_input addr_in;
+	uint64_t mc_umc_status_addr;
+	uint64_t mc_umc_status, err_addr;
+	uint64_t mc_umc_addrt0;
+	uint64_t umc_reg_offset =
+		get_umc_v12_0_reg_offset(adev, node_inst, umc_inst, ch_inst);
+
+	mc_umc_status_addr =
+		SOC15_REG_OFFSET(UMC, 0, regMCA_UMC_UMC0_MCUMC_STATUST0);
+
+	mc_umc_status = RREG64_PCIE_EXT((mc_umc_status_addr + umc_reg_offset) * 4);
+
+	if (mc_umc_status == 0)
+		return 0;
+
+	if (!err_data->err_addr) {
+		/* clear umc status */
+		WREG64_PCIE_EXT((mc_umc_status_addr + umc_reg_offset) * 4, 0x0ULL);
+
+		return 0;
+	}
+
+	/* calculate error address if ue error is detected */
+	if (umc_v12_0_is_uncorrectable_error(adev, mc_umc_status) ||
+	    umc_v12_0_is_deferred_error(adev, mc_umc_status)) {
+		mc_umc_addrt0 =
+			SOC15_REG_OFFSET(UMC, 0, regMCA_UMC_UMC0_MCUMC_ADDRT0);
+
+		err_addr = RREG64_PCIE_EXT((mc_umc_addrt0 + umc_reg_offset) * 4);
+
+		err_addr = REG_GET_FIELD(err_addr, MCA_UMC_UMC0_MCUMC_ADDRT0, ErrorAddr);
+
+		if (!adev->aid_mask &&
+		    adev->smuio.funcs &&
+		    adev->smuio.funcs->get_socket_id)
+			addr_in.ma.socket_id = adev->smuio.funcs->get_socket_id(adev);
+		else
+			addr_in.ma.socket_id = 0;
+
+		addr_in.ma.err_addr = err_addr;
+		addr_in.ma.ch_inst = ch_inst;
+		addr_in.ma.umc_inst = umc_inst;
+		addr_in.ma.node_inst = node_inst;
+
+		umc_v12_0_convert_error_address(adev, err_data, &addr_in, NULL, true);
+	}
+
+	/* clear umc status */
+	WREG64_PCIE_EXT((mc_umc_status_addr + umc_reg_offset) * 4, 0x0ULL);
+
+	return 0;
+}
+
+static void umc_v12_0_query_ras_error_address(struct amdgpu_device *adev,
+					     void *ras_error_status)
+{
+	amdgpu_umc_loop_channels(adev,
+		umc_v12_0_query_error_address, ras_error_status);
+}
+
+static int umc_v12_0_err_cnt_init_per_channel(struct amdgpu_device *adev,
+					uint32_t node_inst, uint32_t umc_inst,
+					uint32_t ch_inst, void *data)
+{
+	uint32_t odecc_cnt_sel;
+	uint64_t odecc_cnt_sel_addr, odecc_err_cnt_addr;
+	uint64_t umc_reg_offset =
+		get_umc_v12_0_reg_offset(adev, node_inst, umc_inst, ch_inst);
+
+	odecc_cnt_sel_addr =
+		SOC15_REG_OFFSET(UMC, 0, regUMCCH0_OdEccCntSel);
+	odecc_err_cnt_addr =
+		SOC15_REG_OFFSET(UMC, 0, regUMCCH0_OdEccErrCnt);
+
+	odecc_cnt_sel = RREG32_PCIE_EXT((odecc_cnt_sel_addr + umc_reg_offset) * 4);
+
+	/* set ce error interrupt type to APIC based interrupt */
+	odecc_cnt_sel = REG_SET_FIELD(odecc_cnt_sel, UMCCH0_OdEccCntSel,
+					OdEccErrInt, 0x1);
+	WREG32_PCIE_EXT((odecc_cnt_sel_addr + umc_reg_offset) * 4, odecc_cnt_sel);
+
+	/* set error count to initial value */
+	WREG32_PCIE_EXT((odecc_err_cnt_addr + umc_reg_offset) * 4, UMC_V12_0_CE_CNT_INIT);
+
+	return 0;
+}
+
+static bool umc_v12_0_check_ecc_err_status(struct amdgpu_device *adev,
+			enum amdgpu_mca_error_type type, void *ras_error_status)
+{
+	uint64_t mc_umc_status = *(uint64_t *)ras_error_status;
+
+	switch (type) {
+	case AMDGPU_MCA_ERROR_TYPE_UE:
+		return umc_v12_0_is_uncorrectable_error(adev, mc_umc_status);
+	case AMDGPU_MCA_ERROR_TYPE_CE:
+		return umc_v12_0_is_correctable_error(adev, mc_umc_status);
+	case AMDGPU_MCA_ERROR_TYPE_DE:
+		return umc_v12_0_is_deferred_error(adev, mc_umc_status);
+	default:
+		return false;
+	}
+
+	return false;
+}
+
+static void umc_v12_0_err_cnt_init(struct amdgpu_device *adev)
+{
+	amdgpu_umc_loop_channels(adev,
+		umc_v12_0_err_cnt_init_per_channel, NULL);
+}
+
+static bool umc_v12_0_query_ras_poison_mode(struct amdgpu_device *adev)
+{
+	/*
+	 * Force return true, because regUMCCH0_EccCtrl
+	 * is not accessible from host side
+	 */
+	return true;
+}
+
+const struct amdgpu_ras_block_hw_ops umc_v12_0_ras_hw_ops = {
+	.query_ras_error_count = umc_v12_0_query_ras_error_count,
+	.query_ras_error_address = umc_v12_0_query_ras_error_address,
+};
+
+static int umc_v12_0_aca_bank_parser(struct aca_handle *handle, struct aca_bank *bank,
+				     enum aca_smu_type type, void *data)
+{
+	struct amdgpu_device *adev = handle->adev;
+	struct aca_bank_info info;
+	enum aca_error_type err_type;
+	u64 status, count;
+	u32 ext_error_code;
+	int ret;
+
+	status = bank->regs[ACA_REG_IDX_STATUS];
+	if (umc_v12_0_is_deferred_error(adev, status))
+		err_type = ACA_ERROR_TYPE_DEFERRED;
+	else if (umc_v12_0_is_uncorrectable_error(adev, status))
+		err_type = ACA_ERROR_TYPE_UE;
+	else if (umc_v12_0_is_correctable_error(adev, status))
+		err_type = ACA_ERROR_TYPE_CE;
+	else
+		return 0;
+	bank->aca_err_type = err_type;
+
+	ret = aca_bank_info_decode(bank, &info);
+	if (ret)
+		return ret;
+
+	amdgpu_umc_update_ecc_status(adev,
+		bank->regs[ACA_REG_IDX_STATUS],
+		bank->regs[ACA_REG_IDX_IPID],
+		bank->regs[ACA_REG_IDX_ADDR]);
+
+	ext_error_code = ACA_REG__STATUS__ERRORCODEEXT(status);
+	if (umc_v12_0_is_deferred_error(adev, status))
+		count = ext_error_code == 0 ?
+			adev->umc.err_addr_cnt / adev->umc.retire_unit : 1ULL;
+	else
+		count = ext_error_code == 0 ?
+			ACA_REG__MISC0__ERRCNT(bank->regs[ACA_REG_IDX_MISC0]) : 1ULL;
+
+	return aca_error_cache_log_bank_error(handle, &info, err_type, count);
+}
+
+static const struct aca_bank_ops umc_v12_0_aca_bank_ops = {
+	.aca_bank_parser = umc_v12_0_aca_bank_parser,
+};
+
+const struct aca_info umc_v12_0_aca_info = {
+	.hwip = ACA_HWIP_TYPE_UMC,
+	.mask = ACA_ERROR_UE_MASK | ACA_ERROR_CE_MASK | ACA_ERROR_DEFERRED_MASK,
+	.bank_ops = &umc_v12_0_aca_bank_ops,
+};
+
+static int umc_v12_0_ras_late_init(struct amdgpu_device *adev, struct ras_common_if *ras_block)
+{
+	int ret;
+
+	ret = amdgpu_umc_ras_late_init(adev, ras_block);
+	if (ret)
+		return ret;
+
+	ret = amdgpu_ras_bind_aca(adev, AMDGPU_RAS_BLOCK__UMC,
+				  &umc_v12_0_aca_info, NULL);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+static int umc_v12_0_update_ecc_status(struct amdgpu_device *adev,
+			uint64_t status, uint64_t ipid, uint64_t addr)
+{
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+	uint16_t hwid, mcatype;
+	uint64_t page_pfn[UMC_V12_0_BAD_PAGE_NUM_PER_CHANNEL];
+	uint64_t err_addr, pa_addr = 0;
+	struct ras_ecc_err *ecc_err;
+	struct ta_ras_query_address_output addr_out;
+	uint32_t shift_bit = adev->umc.flip_bits.flip_bits_in_pa[2];
+	int count, ret, i;
+
+	hwid = REG_GET_FIELD(ipid, MCMP1_IPIDT0, HardwareID);
+	mcatype = REG_GET_FIELD(ipid, MCMP1_IPIDT0, McaType);
+
+	/* The IP block decode of consumption is SMU */
+	if (hwid != MCA_UMC_HWID_V12_0 || mcatype != MCA_UMC_MCATYPE_V12_0) {
+		con->umc_ecc_log.consumption_q_count++;
+		return 0;
+	}
+
+	if (!status)
+		return 0;
+
+	if (!umc_v12_0_is_deferred_error(adev, status))
+		return 0;
+
+	err_addr = REG_GET_FIELD(addr,
+				MCA_UMC_UMC0_MCUMC_ADDRT0, ErrorAddr);
+
+	dev_dbg(adev->dev,
+		"UMC:IPID:0x%llx, socket:%llu, aid:%llu, inst:%llu, ch:%llu, err_addr:0x%llx\n",
+		ipid,
+		MCA_IPID_2_SOCKET_ID(ipid),
+		MCA_IPID_2_DIE_ID(ipid),
+		MCA_IPID_2_UMC_INST(ipid),
+		MCA_IPID_2_UMC_CH(ipid),
+		err_addr);
+
+	ret = amdgpu_umc_mca_to_addr(adev,
+			err_addr, MCA_IPID_2_UMC_CH(ipid),
+			MCA_IPID_2_UMC_INST(ipid), MCA_IPID_2_DIE_ID(ipid),
+			MCA_IPID_2_SOCKET_ID(ipid), &addr_out, true);
+	if (ret)
+		return ret;
+
+	ecc_err = kzalloc(sizeof(*ecc_err), GFP_KERNEL);
+	if (!ecc_err)
+		return -ENOMEM;
+
+	pa_addr = addr_out.pa.pa;
+	ecc_err->status = status;
+	ecc_err->ipid = ipid;
+	ecc_err->addr = addr;
+	ecc_err->pa_pfn = pa_addr >> AMDGPU_GPU_PAGE_SHIFT;
+	ecc_err->channel_idx = addr_out.pa.channel_idx;
+
+	/* If converted pa_pfn is 0, use pa C4 pfn. */
+	if (!ecc_err->pa_pfn)
+		ecc_err->pa_pfn = BIT_ULL(shift_bit) >> AMDGPU_GPU_PAGE_SHIFT;
+
+	ret = amdgpu_umc_logs_ecc_err(adev, &con->umc_ecc_log.de_page_tree, ecc_err);
+	if (ret) {
+		if (ret == -EEXIST)
+			con->umc_ecc_log.de_queried_count++;
+		else
+			dev_err(adev->dev, "Fail to log ecc error! ret:%d\n", ret);
+
+		kfree(ecc_err);
+		return ret;
+	}
+
+	con->umc_ecc_log.de_queried_count++;
+
+	memset(page_pfn, 0, sizeof(page_pfn));
+	count = amdgpu_umc_lookup_bad_pages_in_a_row(adev,
+				pa_addr,
+				page_pfn, ARRAY_SIZE(page_pfn));
+	if (count <= 0) {
+		dev_warn(adev->dev, "Fail to convert error address! count:%d\n", count);
+		return 0;
+	}
+
+	/* Reserve memory */
+	for (i = 0; i < count; i++)
+		amdgpu_ras_reserve_page(adev, page_pfn[i]);
+
+	/* The problem case is as follows:
+	 * 1. GPU A triggers a gpu ras reset, and GPU A drives
+	 *    GPU B to also perform a gpu ras reset.
+	 * 2. After gpu B ras reset started, gpu B queried a DE
+	 *    data. Since the DE data was queried in the ras reset
+	 *    thread instead of the page retirement thread, bad
+	 *    page retirement work would not be triggered. Then
+	 *    even if all gpu resets are completed, the bad pages
+	 *    will be cached in RAM until GPU B's bad page retirement
+	 *    work is triggered again and then saved to eeprom.
+	 * Trigger delayed work to save the bad pages to eeprom in time
+	 * after gpu ras reset is completed.
+	 */
+	if (amdgpu_ras_in_recovery(adev))
+		schedule_delayed_work(&con->page_retirement_dwork,
+			msecs_to_jiffies(DELAYED_TIME_FOR_GPU_RESET));
+
+	return 0;
+}
+
+static int umc_v12_0_fill_error_record(struct amdgpu_device *adev,
+				struct ras_ecc_err *ecc_err, void *ras_error_status)
+{
+	struct ras_err_data *err_data = (struct ras_err_data *)ras_error_status;
+	uint64_t page_pfn[UMC_V12_0_BAD_PAGE_NUM_PER_CHANNEL];
+	int ret, i, count;
+
+	if (!err_data || !ecc_err)
+		return -EINVAL;
+
+	memset(page_pfn, 0, sizeof(page_pfn));
+	count = amdgpu_umc_lookup_bad_pages_in_a_row(adev,
+				ecc_err->pa_pfn << AMDGPU_GPU_PAGE_SHIFT,
+				page_pfn, ARRAY_SIZE(page_pfn));
+
+	for (i = 0; i < count; i++) {
+		ret = amdgpu_umc_fill_error_record(err_data,
+				ecc_err->addr,
+				page_pfn[i] << AMDGPU_GPU_PAGE_SHIFT,
+				ecc_err->channel_idx,
+				MCA_IPID_2_UMC_INST(ecc_err->ipid));
+		if (ret)
+			break;
+	}
+
+	err_data->de_count++;
+
+	return ret;
+}
+
+static void umc_v12_0_query_ras_ecc_err_addr(struct amdgpu_device *adev,
+					void *ras_error_status)
+{
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+	struct ras_ecc_err *entries[MAX_ECC_NUM_PER_RETIREMENT];
+	struct radix_tree_root *ecc_tree;
+	int new_detected, ret, i;
+
+	ecc_tree = &con->umc_ecc_log.de_page_tree;
+
+	mutex_lock(&con->umc_ecc_log.lock);
+	new_detected = radix_tree_gang_lookup_tag(ecc_tree, (void **)entries,
+			0, ARRAY_SIZE(entries), UMC_ECC_NEW_DETECTED_TAG);
+	for (i = 0; i < new_detected; i++) {
+		if (!entries[i])
+			continue;
+
+		ret = umc_v12_0_fill_error_record(adev, entries[i], ras_error_status);
+		if (ret) {
+			dev_err(adev->dev, "Fail to fill umc error record, ret:%d\n", ret);
+			break;
+		}
+		radix_tree_tag_clear(ecc_tree,
+				entries[i]->pa_pfn, UMC_ECC_NEW_DETECTED_TAG);
+	}
+	mutex_unlock(&con->umc_ecc_log.lock);
+}
+
+static uint32_t umc_v12_0_get_die_id(struct amdgpu_device *adev,
+		uint64_t mca_addr, uint64_t retired_page)
+{
+	uint32_t die = 0;
+
+	/* we only calculate die id for nps1 mode right now */
+	die += ((((retired_page >> 12) & 0x1ULL)^
+	    ((retired_page >> 20) & 0x1ULL) ^
+	    ((retired_page >> 27) & 0x1ULL) ^
+	    ((retired_page >> 34) & 0x1ULL) ^
+	    ((retired_page >> 41) & 0x1ULL)) << 0);
+
+	/* the original PA_C4 and PA_R13 may be cleared in retired_page, so
+	 * get them from mca_addr.
+	 */
+	die += ((((retired_page >> 13) & 0x1ULL) ^
+	    ((mca_addr >> 5) & 0x1ULL) ^
+	    ((retired_page >> 28) & 0x1ULL) ^
+	    ((mca_addr >> 23) & 0x1ULL) ^
+	    ((retired_page >> 42) & 0x1ULL)) << 1);
+	die &= 3;
+
+	return die;
+}
+
+static void umc_v12_0_mca_ipid_parse(struct amdgpu_device *adev, uint64_t ipid,
+		uint32_t *did, uint32_t *ch, uint32_t *umc_inst, uint32_t *sid)
+{
+	if (did)
+		*did = MCA_IPID_2_DIE_ID(ipid);
+	if (ch)
+		*ch = MCA_IPID_2_UMC_CH(ipid);
+	if (umc_inst)
+		*umc_inst = MCA_IPID_2_UMC_INST(ipid);
+	if (sid)
+		*sid = MCA_IPID_2_SOCKET_ID(ipid);
+}
+
+struct amdgpu_umc_ras umc_v12_0_ras = {
+	.ras_block = {
+		.hw_ops = &umc_v12_0_ras_hw_ops,
+		.ras_late_init = umc_v12_0_ras_late_init,
+	},
+	.err_cnt_init = umc_v12_0_err_cnt_init,
+	.query_ras_poison_mode = umc_v12_0_query_ras_poison_mode,
+	.ecc_info_query_ras_error_address = umc_v12_0_query_ras_ecc_err_addr,
+	.check_ecc_err_status = umc_v12_0_check_ecc_err_status,
+	.update_ecc_status = umc_v12_0_update_ecc_status,
+	.convert_ras_err_addr = umc_v12_0_convert_error_address,
+	.get_die_id_from_pa = umc_v12_0_get_die_id,
+	.get_retire_flip_bits = umc_v12_0_get_retire_flip_bits,
+	.mca_ipid_parse = umc_v12_0_mca_ipid_parse,
+};
+
diff --git a/drivers/gpu/drm/amd/amdgpu/umc_v12_0.h b/drivers/gpu/drm/amd/amdgpu/umc_v12_0.h
new file mode 100644
index 000000000000..63b7e7254526
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/umc_v12_0.h
@@ -0,0 +1,105 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __UMC_V12_0_H__
+#define __UMC_V12_0_H__
+
+#include "soc15_common.h"
+#include "amdgpu.h"
+
+#define UMC_V12_0_NODE_DIST		0x40000000
+#define UMC_V12_0_INST_DIST		0x40000
+
+/* UMC register per channel offset */
+#define UMC_V12_0_PER_CHANNEL_OFFSET	0x400
+
+/* UMC cross node offset */
+#define UMC_V12_0_CROSS_NODE_OFFSET		0x100000000
+
+/* OdEccErrCnt max value */
+#define UMC_V12_0_CE_CNT_MAX		0xffff
+/* umc ce interrupt threshold */
+#define UMC_V12_0_CE_INT_THRESHOLD	0xffff
+/* umc ce count initial value */
+#define UMC_V12_0_CE_CNT_INIT	(UMC_V12_0_CE_CNT_MAX - UMC_V12_0_CE_INT_THRESHOLD)
+
+/* number of umc channel instance with memory map register access */
+#define UMC_V12_0_CHANNEL_INSTANCE_NUM		8
+/* number of umc instance with memory map register access */
+#define UMC_V12_0_UMC_INSTANCE_NUM		4
+
+/* Total channel instances for all available umc nodes */
+#define UMC_V12_0_TOTAL_CHANNEL_NUM(adev) \
+	(UMC_V12_0_CHANNEL_INSTANCE_NUM * (adev)->gmc.num_umc)
+
+/* one piece of normalized address is mapped to 8 pieces of physical address */
+#define UMC_V12_0_NA_MAP_PA_NUM        8
+/* R13 bit shift should be considered, double the number */
+#define UMC_V12_0_BAD_PAGE_NUM_PER_CHANNEL (UMC_V12_0_NA_MAP_PA_NUM * 2)
+
+/* column bits in SOC physical address */
+#define UMC_V12_0_PA_C2_BIT 15
+#define UMC_V12_0_PA_C3_BIT 16
+#define UMC_V12_0_PA_C4_BIT 21
+/* row bits in SOC physical address */
+#define UMC_V12_0_PA_R0_BIT 22
+#define UMC_V12_0_PA_R10_BIT 32
+#define UMC_V12_0_PA_R11_BIT 33
+#define UMC_V12_0_PA_R12_BIT 34
+#define UMC_V12_0_PA_R13_BIT 35
+/* channel bit in SOC physical address */
+#define UMC_V12_0_PA_CH4_BIT 12
+#define UMC_V12_0_PA_CH5_BIT 13
+/* bank bit in SOC physical address */
+#define UMC_V12_0_PA_B0_BIT 19
+#define UMC_V12_0_PA_B1_BIT 20
+/* row bits in MCA address */
+#define UMC_V12_0_MA_R0_BIT 10
+
+#define MCA_UMC_HWID_V12_0     0x96
+#define MCA_UMC_MCATYPE_V12_0  0x0
+
+#define MCA_IPID_LO_2_UMC_CH(_ipid_lo) (((((_ipid_lo) >> 20) & 0x1) * 4) + \
+			(((_ipid_lo) >> 12) & 0xF))
+#define MCA_IPID_LO_2_UMC_INST(_ipid_lo) (((_ipid_lo) >> 21) & 0x7)
+
+#define MCA_IPID_2_DIE_ID(ipid)  ((REG_GET_FIELD(ipid, MCMP1_IPIDT0, InstanceIdHi) >> 2) & 0x03)
+
+#define MCA_IPID_2_UMC_CH(ipid) \
+	(MCA_IPID_LO_2_UMC_CH(REG_GET_FIELD(ipid, MCMP1_IPIDT0, InstanceIdLo)))
+
+#define MCA_IPID_2_UMC_INST(ipid) \
+	(MCA_IPID_LO_2_UMC_INST(REG_GET_FIELD(ipid, MCMP1_IPIDT0, InstanceIdLo)))
+
+#define MCA_IPID_2_SOCKET_ID(ipid) \
+	(((REG_GET_FIELD(ipid, MCMP1_IPIDT0, InstanceIdLo) & 0x1) << 2) | \
+	 (REG_GET_FIELD(ipid, MCMP1_IPIDT0, InstanceIdHi) & 0x03))
+
+bool umc_v12_0_is_deferred_error(struct amdgpu_device *adev, uint64_t mc_umc_status);
+bool umc_v12_0_is_uncorrectable_error(struct amdgpu_device *adev, uint64_t mc_umc_status);
+bool umc_v12_0_is_correctable_error(struct amdgpu_device *adev, uint64_t mc_umc_status);
+
+typedef bool (*check_error_type_func)(struct amdgpu_device *adev, uint64_t mc_umc_status);
+
+extern struct amdgpu_umc_ras umc_v12_0_ras;
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/umc_v6_0.c b/drivers/gpu/drm/amd/amdgpu/umc_v6_0.c
index 0d6b50528d76..97fa88ed770c 100644
--- a/drivers/gpu/drm/amd/amdgpu/umc_v6_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/umc_v6_0.c
@@ -25,7 +25,7 @@
 
 static void umc_v6_0_init_registers(struct amdgpu_device *adev)
 {
-	unsigned i,j;
+	unsigned i, j;
 
 	for (i = 0; i < 4; i++)
 		for (j = 0; j < 4; j++)
diff --git a/drivers/gpu/drm/amd/amdgpu/umc_v6_1.c b/drivers/gpu/drm/amd/amdgpu/umc_v6_1.c
index 20b44983ac94..f17d297b594b 100644
--- a/drivers/gpu/drm/amd/amdgpu/umc_v6_1.c
+++ b/drivers/gpu/drm/amd/amdgpu/umc_v6_1.c
@@ -300,7 +300,6 @@ static void umc_v6_1_query_error_address(struct amdgpu_device *adev,
 {
 	uint32_t lsb, mc_umc_status_addr;
 	uint64_t mc_umc_status, err_addr, retired_page, mc_umc_addrt0;
-	struct eeprom_table_record *err_rec;
 	uint32_t channel_index = adev->umc.channel_idx_tbl[umc_inst * adev->umc.channel_inst_num + ch_inst];
 
 	if (adev->asic_type == CHIP_ARCTURUS) {
@@ -328,12 +327,9 @@ static void umc_v6_1_query_error_address(struct amdgpu_device *adev,
 		return;
 	}
 
-	err_rec = &err_data->err_addr[err_data->err_addr_cnt];
-
-	/* calculate error address if ue/ce error is detected */
+	/* calculate error address if ue error is detected */
 	if (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Val) == 1 &&
-	    (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UECC) == 1 ||
-	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, CECC) == 1)) {
+	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UECC) == 1) {
 
 		err_addr = RREG64_PCIE((mc_umc_addrt0 + umc_reg_offset) * 4);
 		/* the lowest lsb bits should be ignored */
@@ -346,20 +342,8 @@ static void umc_v6_1_query_error_address(struct amdgpu_device *adev,
 				ADDR_OF_256B_BLOCK(channel_index) |
 				OFFSET_IN_256B_BLOCK(err_addr);
 
-		/* we only save ue error information currently, ce is skipped */
-		if (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UECC)
-				== 1) {
-			err_rec->address = err_addr;
-			/* page frame address is saved */
-			err_rec->retired_page = retired_page >> AMDGPU_GPU_PAGE_SHIFT;
-			err_rec->ts = (uint64_t)ktime_get_real_seconds();
-			err_rec->err_type = AMDGPU_RAS_EEPROM_ERR_NON_RECOVERABLE;
-			err_rec->cu = 0;
-			err_rec->mem_channel = channel_index;
-			err_rec->mcumc_id = umc_inst;
-
-			err_data->err_addr_cnt++;
-		}
+		amdgpu_umc_fill_error_record(err_data, err_addr,
+					retired_page, channel_index, umc_inst);
 	}
 
 	/* clear umc status */
@@ -465,10 +449,14 @@ static void umc_v6_1_err_cnt_init(struct amdgpu_device *adev)
 		umc_v6_1_enable_umc_index_mode(adev);
 }
 
-const struct amdgpu_umc_ras_funcs umc_v6_1_ras_funcs = {
-	.err_cnt_init = umc_v6_1_err_cnt_init,
-	.ras_late_init = amdgpu_umc_ras_late_init,
-	.ras_fini = amdgpu_umc_ras_fini,
+const struct amdgpu_ras_block_hw_ops umc_v6_1_ras_hw_ops = {
 	.query_ras_error_count = umc_v6_1_query_ras_error_count,
 	.query_ras_error_address = umc_v6_1_query_ras_error_address,
 };
+
+struct amdgpu_umc_ras umc_v6_1_ras = {
+	.ras_block = {
+		.hw_ops = &umc_v6_1_ras_hw_ops,
+	},
+	.err_cnt_init = umc_v6_1_err_cnt_init,
+};
+\ No newline at end of file
diff --git a/drivers/gpu/drm/amd/amdgpu/umc_v6_1.h b/drivers/gpu/drm/amd/amdgpu/umc_v6_1.h
index 5dc36c730bb2..50c632eb4cc6 100644
--- a/drivers/gpu/drm/amd/amdgpu/umc_v6_1.h
+++ b/drivers/gpu/drm/amd/amdgpu/umc_v6_1.h
@@ -45,7 +45,7 @@
 /* umc ce count initial value */
 #define UMC_V6_1_CE_CNT_INIT	(UMC_V6_1_CE_CNT_MAX - UMC_V6_1_CE_INT_THRESHOLD)
 
-extern const struct amdgpu_umc_ras_funcs umc_v6_1_ras_funcs;
+extern struct amdgpu_umc_ras umc_v6_1_ras;
 extern const uint32_t
 	umc_v6_1_channel_idx_tbl[UMC_V6_1_UMC_INSTANCE_NUM][UMC_V6_1_CHANNEL_INSTANCE_NUM];
 
diff --git a/drivers/gpu/drm/amd/amdgpu/umc_v6_7.c b/drivers/gpu/drm/amd/amdgpu/umc_v6_7.c
index f7ec3fe134e5..a3ee3c4c650f 100644
--- a/drivers/gpu/drm/amd/amdgpu/umc_v6_7.c
+++ b/drivers/gpu/drm/amd/amdgpu/umc_v6_7.c
@@ -47,12 +47,222 @@ static inline uint32_t get_umc_v6_7_reg_offset(struct amdgpu_device *adev,
 					      uint32_t umc_inst,
 					      uint32_t ch_inst)
 {
+	uint32_t index = umc_inst * adev->umc.channel_inst_num + ch_inst;
+
+	/* adjust umc and channel index offset,
+	 * the register address is not linear on each umc instace */
+	umc_inst = index / 4;
+	ch_inst = index % 4;
+
 	return adev->umc.channel_offs * ch_inst + UMC_V6_7_INST_DIST * umc_inst;
 }
 
+static void umc_v6_7_query_error_status_helper(struct amdgpu_device *adev,
+						  uint64_t mc_umc_status, uint32_t umc_reg_offset)
+{
+	uint32_t mc_umc_addr;
+	uint64_t reg_value;
+
+	if (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Deferred) == 1)
+		dev_info(adev->dev, "Deferred error\n");
+
+	if (mc_umc_status)
+		dev_info(adev->dev, "MCA STATUS 0x%llx, umc_reg_offset 0x%x\n", mc_umc_status, umc_reg_offset);
+
+	/* print IPID registers value */
+	mc_umc_addr =
+		SOC15_REG_OFFSET(UMC, 0, regMCA_UMC_UMC0_MCUMC_IPIDT0);
+	reg_value = RREG64_PCIE((mc_umc_addr + umc_reg_offset) * 4);
+	if (reg_value)
+		dev_info(adev->dev, "MCA IPID 0x%llx, umc_reg_offset 0x%x\n", reg_value, umc_reg_offset);
+
+	/* print SYND registers value */
+	mc_umc_addr =
+		SOC15_REG_OFFSET(UMC, 0, regMCA_UMC_UMC0_MCUMC_SYNDT0);
+	reg_value = RREG64_PCIE((mc_umc_addr + umc_reg_offset) * 4);
+	if (reg_value)
+		dev_info(adev->dev, "MCA SYND 0x%llx, umc_reg_offset 0x%x\n", reg_value, umc_reg_offset);
+
+	/* print MISC0 registers value */
+	mc_umc_addr =
+		SOC15_REG_OFFSET(UMC, 0, regMCA_UMC_UMC0_MCUMC_MISC0T0);
+	reg_value = RREG64_PCIE((mc_umc_addr + umc_reg_offset) * 4);
+	if (reg_value)
+		dev_info(adev->dev, "MCA MISC0 0x%llx, umc_reg_offset 0x%x\n", reg_value, umc_reg_offset);
+}
+
+static void umc_v6_7_ecc_info_query_correctable_error_count(struct amdgpu_device *adev,
+						   uint32_t umc_inst, uint32_t ch_inst,
+						   unsigned long *error_count)
+{
+	uint64_t mc_umc_status;
+	uint32_t eccinfo_table_idx;
+	uint32_t umc_reg_offset;
+	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
+
+	umc_reg_offset = get_umc_v6_7_reg_offset(adev,
+						umc_inst, ch_inst);
+
+	eccinfo_table_idx = umc_inst * adev->umc.channel_inst_num + ch_inst;
+	/* check for SRAM correctable error
+	  MCUMC_STATUS is a 64 bit register */
+	mc_umc_status = ras->umc_ecc.ecc[eccinfo_table_idx].mca_umc_status;
+	if (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Val) == 1 &&
+	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, CECC) == 1) {
+		*error_count += 1;
+
+		umc_v6_7_query_error_status_helper(adev, mc_umc_status, umc_reg_offset);
+
+		if (ras->umc_ecc.record_ce_addr_supported)	{
+			uint64_t err_addr, soc_pa;
+			uint32_t channel_index =
+				adev->umc.channel_idx_tbl[umc_inst * adev->umc.channel_inst_num + ch_inst];
+
+			err_addr = ras->umc_ecc.ecc[eccinfo_table_idx].mca_ceumc_addr;
+			err_addr = REG_GET_FIELD(err_addr, MCA_UMC_UMC0_MCUMC_ADDRT0, ErrorAddr);
+			/* translate umc channel address to soc pa, 3 parts are included */
+			soc_pa = ADDR_OF_8KB_BLOCK(err_addr) |
+					ADDR_OF_256B_BLOCK(channel_index) |
+					OFFSET_IN_256B_BLOCK(err_addr);
+
+			/* The umc channel bits are not original values, they are hashed */
+			SET_CHANNEL_HASH(channel_index, soc_pa);
+
+			dev_info(adev->dev, "Error Address(PA): 0x%llx\n", soc_pa);
+		}
+	}
+}
+
+static void umc_v6_7_ecc_info_querry_uncorrectable_error_count(struct amdgpu_device *adev,
+							  uint32_t umc_inst, uint32_t ch_inst,
+						      unsigned long *error_count)
+{
+	uint64_t mc_umc_status;
+	uint32_t eccinfo_table_idx;
+	uint32_t umc_reg_offset;
+	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
+
+	umc_reg_offset = get_umc_v6_7_reg_offset(adev,
+						umc_inst, ch_inst);
+
+	eccinfo_table_idx = umc_inst * adev->umc.channel_inst_num + ch_inst;
+	/* check the MCUMC_STATUS */
+	mc_umc_status = ras->umc_ecc.ecc[eccinfo_table_idx].mca_umc_status;
+	if ((REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Val) == 1) &&
+	    (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Deferred) == 1 ||
+	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UECC) == 1 ||
+	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, PCC) == 1 ||
+	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UC) == 1 ||
+	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, TCC) == 1)) {
+		*error_count += 1;
+
+		umc_v6_7_query_error_status_helper(adev, mc_umc_status, umc_reg_offset);
+	}
+}
+
+static int umc_v6_7_ecc_info_querry_ecc_error_count(struct amdgpu_device *adev,
+					uint32_t node_inst, uint32_t umc_inst,
+					uint32_t ch_inst, void *data)
+{
+	struct ras_err_data *err_data = (struct ras_err_data *)data;
+
+	umc_v6_7_ecc_info_query_correctable_error_count(adev,
+		umc_inst, ch_inst,
+		&(err_data->ce_count));
+
+	umc_v6_7_ecc_info_querry_uncorrectable_error_count(adev,
+		umc_inst, ch_inst,
+		&(err_data->ue_count));
+
+	return 0;
+}
+
+static void umc_v6_7_ecc_info_query_ras_error_count(struct amdgpu_device *adev,
+					   void *ras_error_status)
+{
+	amdgpu_umc_loop_channels(adev,
+		umc_v6_7_ecc_info_querry_ecc_error_count, ras_error_status);
+}
+
+void umc_v6_7_convert_error_address(struct amdgpu_device *adev,
+				    struct ras_err_data *err_data, uint64_t err_addr,
+				    uint32_t ch_inst, uint32_t umc_inst)
+{
+	uint32_t channel_index;
+	uint64_t soc_pa, retired_page, column;
+
+	channel_index =
+		adev->umc.channel_idx_tbl[umc_inst * adev->umc.channel_inst_num + ch_inst];
+	/* translate umc channel address to soc pa, 3 parts are included */
+	soc_pa = ADDR_OF_8KB_BLOCK(err_addr) |
+			ADDR_OF_256B_BLOCK(channel_index) |
+			OFFSET_IN_256B_BLOCK(err_addr);
+
+	/* The umc channel bits are not original values, they are hashed */
+	SET_CHANNEL_HASH(channel_index, soc_pa);
+
+	/* clear [C4 C3 C2] in soc physical address */
+	soc_pa &= ~(0x7ULL << UMC_V6_7_PA_C2_BIT);
+
+	/* loop for all possibilities of [C4 C3 C2] */
+	for (column = 0; column < UMC_V6_7_NA_MAP_PA_NUM; column++) {
+		retired_page = soc_pa | (column << UMC_V6_7_PA_C2_BIT);
+		dev_info(adev->dev, "Error Address(PA): 0x%llx\n", retired_page);
+		amdgpu_umc_fill_error_record(err_data, err_addr,
+			retired_page, channel_index, umc_inst);
+
+		/* shift R14 bit */
+		retired_page ^= (0x1ULL << UMC_V6_7_PA_R14_BIT);
+		dev_info(adev->dev, "Error Address(PA): 0x%llx\n", retired_page);
+		amdgpu_umc_fill_error_record(err_data, err_addr,
+			retired_page, channel_index, umc_inst);
+	}
+}
+
+static int umc_v6_7_ecc_info_query_error_address(struct amdgpu_device *adev,
+					uint32_t node_inst, uint32_t umc_inst,
+					uint32_t ch_inst, void *data)
+{
+	uint64_t mc_umc_status, err_addr;
+	uint32_t eccinfo_table_idx;
+	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
+	struct ras_err_data *err_data = (struct ras_err_data *)data;
+
+	eccinfo_table_idx = umc_inst * adev->umc.channel_inst_num + ch_inst;
+	mc_umc_status = ras->umc_ecc.ecc[eccinfo_table_idx].mca_umc_status;
+
+	if (mc_umc_status == 0)
+		return 0;
+
+	if (!err_data->err_addr)
+		return 0;
+
+	/* calculate error address if ue error is detected */
+	if (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Val) == 1 &&
+	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UECC) == 1) {
+
+		err_addr = ras->umc_ecc.ecc[eccinfo_table_idx].mca_umc_addr;
+		err_addr = REG_GET_FIELD(err_addr, MCA_UMC_UMC0_MCUMC_ADDRT0, ErrorAddr);
+
+		umc_v6_7_convert_error_address(adev, err_data, err_addr,
+					ch_inst, umc_inst);
+	}
+
+	return 0;
+}
+
+static void umc_v6_7_ecc_info_query_ras_error_address(struct amdgpu_device *adev,
+					     void *ras_error_status)
+{
+	amdgpu_umc_loop_channels(adev,
+	    umc_v6_7_ecc_info_query_error_address, ras_error_status);
+}
+
 static void umc_v6_7_query_correctable_error_count(struct amdgpu_device *adev,
 						   uint32_t umc_reg_offset,
-						   unsigned long *error_count)
+						   unsigned long *error_count,
+						   uint32_t ch_inst,
+						   uint32_t umc_inst)
 {
 	uint32_t ecc_err_cnt_sel, ecc_err_cnt_sel_addr;
 	uint32_t ecc_err_cnt, ecc_err_cnt_addr;
@@ -92,8 +302,36 @@ static void umc_v6_7_query_correctable_error_count(struct amdgpu_device *adev,
 	  MCUMC_STATUS is a 64 bit register */
 	mc_umc_status = RREG64_PCIE((mc_umc_status_addr + umc_reg_offset) * 4);
 	if (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Val) == 1 &&
-	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, CECC) == 1)
+	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, CECC) == 1) {
 		*error_count += 1;
+
+		umc_v6_7_query_error_status_helper(adev, mc_umc_status, umc_reg_offset);
+
+		{
+			uint64_t err_addr, soc_pa;
+			uint32_t mc_umc_addrt0;
+			uint32_t channel_index;
+
+			mc_umc_addrt0 =
+				SOC15_REG_OFFSET(UMC, 0, regMCA_UMC_UMC0_MCUMC_ADDRT0);
+
+			channel_index =
+				adev->umc.channel_idx_tbl[umc_inst * adev->umc.channel_inst_num + ch_inst];
+
+			err_addr = RREG64_PCIE((mc_umc_addrt0 + umc_reg_offset) * 4);
+			err_addr = REG_GET_FIELD(err_addr, MCA_UMC_UMC0_MCUMC_ADDRT0, ErrorAddr);
+
+			/* translate umc channel address to soc pa, 3 parts are included */
+			soc_pa = ADDR_OF_8KB_BLOCK(err_addr) |
+					ADDR_OF_256B_BLOCK(channel_index) |
+					OFFSET_IN_256B_BLOCK(err_addr);
+
+			/* The umc channel bits are not original values, they are hashed */
+			SET_CHANNEL_HASH(channel_index, soc_pa);
+
+			dev_info(adev->dev, "Error Address(PA): 0x%llx\n", soc_pa);
+		}
+	}
 }
 
 static void umc_v6_7_querry_uncorrectable_error_count(struct amdgpu_device *adev,
@@ -113,15 +351,21 @@ static void umc_v6_7_querry_uncorrectable_error_count(struct amdgpu_device *adev
 	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UECC) == 1 ||
 	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, PCC) == 1 ||
 	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UC) == 1 ||
-	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, TCC) == 1))
+	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, TCC) == 1)) {
 		*error_count += 1;
+
+		umc_v6_7_query_error_status_helper(adev, mc_umc_status, umc_reg_offset);
+	}
 }
 
-static void umc_v6_7_reset_error_count_per_channel(struct amdgpu_device *adev,
-						   uint32_t umc_reg_offset)
+static int umc_v6_7_reset_error_count_per_channel(struct amdgpu_device *adev,
+					uint32_t node_inst, uint32_t umc_inst,
+					uint32_t ch_inst, void *data)
 {
 	uint32_t ecc_err_cnt_addr;
 	uint32_t ecc_err_cnt_sel, ecc_err_cnt_sel_addr;
+	uint32_t umc_reg_offset =
+		get_umc_v6_7_reg_offset(adev, umc_inst, ch_inst);
 
 	ecc_err_cnt_sel_addr =
 		SOC15_REG_OFFSET(UMC, 0,
@@ -155,60 +399,54 @@ static void umc_v6_7_reset_error_count_per_channel(struct amdgpu_device *adev,
 	/* clear higher chip error count */
 	WREG32_PCIE((ecc_err_cnt_addr + umc_reg_offset) * 4,
 			UMC_V6_7_CE_CNT_INIT);
+
+	return 0;
 }
 
 static void umc_v6_7_reset_error_count(struct amdgpu_device *adev)
 {
-	uint32_t umc_inst        = 0;
-	uint32_t ch_inst         = 0;
-	uint32_t umc_reg_offset  = 0;
+	amdgpu_umc_loop_channels(adev,
+		umc_v6_7_reset_error_count_per_channel, NULL);
+}
 
-	LOOP_UMC_INST_AND_CH(umc_inst, ch_inst) {
-		umc_reg_offset = get_umc_v6_7_reg_offset(adev,
-							 umc_inst,
-							 ch_inst);
+static int umc_v6_7_query_ecc_error_count(struct amdgpu_device *adev,
+					uint32_t node_inst, uint32_t umc_inst,
+					uint32_t ch_inst, void *data)
+{
+	struct ras_err_data *err_data = (struct ras_err_data *)data;
+	uint32_t umc_reg_offset =
+		get_umc_v6_7_reg_offset(adev, umc_inst, ch_inst);
 
-		umc_v6_7_reset_error_count_per_channel(adev,
-						       umc_reg_offset);
-	}
+	umc_v6_7_query_correctable_error_count(adev,
+					umc_reg_offset,
+					&(err_data->ce_count),
+					ch_inst, umc_inst);
+
+	umc_v6_7_querry_uncorrectable_error_count(adev,
+					umc_reg_offset,
+					&(err_data->ue_count));
+
+	return 0;
 }
 
 static void umc_v6_7_query_ras_error_count(struct amdgpu_device *adev,
 					   void *ras_error_status)
 {
-	struct ras_err_data *err_data = (struct ras_err_data *)ras_error_status;
-
-	uint32_t umc_inst        = 0;
-	uint32_t ch_inst         = 0;
-	uint32_t umc_reg_offset  = 0;
-
-	/*TODO: driver needs to toggle DF Cstate to ensure
-	 * safe access of UMC registers. Will add the protection */
-	LOOP_UMC_INST_AND_CH(umc_inst, ch_inst) {
-		umc_reg_offset = get_umc_v6_7_reg_offset(adev,
-							 umc_inst,
-							 ch_inst);
-		umc_v6_7_query_correctable_error_count(adev,
-						       umc_reg_offset,
-						       &(err_data->ce_count));
-		umc_v6_7_querry_uncorrectable_error_count(adev,
-							  umc_reg_offset,
-							  &(err_data->ue_count));
-	}
+	amdgpu_umc_loop_channels(adev,
+		umc_v6_7_query_ecc_error_count, ras_error_status);
 
 	umc_v6_7_reset_error_count(adev);
 }
 
-static void umc_v6_7_query_error_address(struct amdgpu_device *adev,
-					 struct ras_err_data *err_data,
-					 uint32_t umc_reg_offset,
-					 uint32_t ch_inst,
-					 uint32_t umc_inst)
+static int umc_v6_7_query_error_address(struct amdgpu_device *adev,
+					uint32_t node_inst, uint32_t umc_inst,
+					uint32_t ch_inst, void *data)
 {
 	uint32_t mc_umc_status_addr;
-	uint64_t mc_umc_status, err_addr, retired_page, mc_umc_addrt0;
-	struct eeprom_table_record *err_rec;
-	uint32_t channel_index;
+	uint64_t mc_umc_status = 0, mc_umc_addrt0, err_addr;
+	struct ras_err_data *err_data = (struct ras_err_data *)data;
+	uint32_t umc_reg_offset =
+		get_umc_v6_7_reg_offset(adev, umc_inst, ch_inst);
 
 	mc_umc_status_addr =
 		SOC15_REG_OFFSET(UMC, 0, regMCA_UMC_UMC0_MCUMC_STATUST0);
@@ -218,74 +456,36 @@ static void umc_v6_7_query_error_address(struct amdgpu_device *adev,
 	mc_umc_status = RREG64_PCIE((mc_umc_status_addr + umc_reg_offset) * 4);
 
 	if (mc_umc_status == 0)
-		return;
+		return 0;
 
 	if (!err_data->err_addr) {
 		/* clear umc status */
 		WREG64_PCIE((mc_umc_status_addr + umc_reg_offset) * 4, 0x0ULL);
-		return;
+		return 0;
 	}
 
-	err_rec = &err_data->err_addr[err_data->err_addr_cnt];
-
-	channel_index =
-		adev->umc.channel_idx_tbl[umc_inst * adev->umc.channel_inst_num + ch_inst];
-
-	/* calculate error address if ue/ce error is detected */
+	/* calculate error address if ue error is detected */
 	if (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Val) == 1 &&
-	    (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UECC) == 1 ||
-	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, CECC) == 1)) {
-
+	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UECC) == 1) {
 		err_addr = RREG64_PCIE((mc_umc_addrt0 + umc_reg_offset) * 4);
-		err_addr = REG_GET_FIELD(err_addr, MCA_UMC_UMC0_MCUMC_ADDRT0, ErrorAddr);
+		err_addr =
+			REG_GET_FIELD(err_addr, MCA_UMC_UMC0_MCUMC_ADDRT0, ErrorAddr);
 
-		/* translate umc channel address to soc pa, 3 parts are included */
-		retired_page = ADDR_OF_8KB_BLOCK(err_addr) |
-				ADDR_OF_256B_BLOCK(channel_index) |
-				OFFSET_IN_256B_BLOCK(err_addr);
-
-		/* we only save ue error information currently, ce is skipped */
-		if (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UECC)
-				== 1) {
-			err_rec->address = err_addr;
-			/* page frame address is saved */
-			err_rec->retired_page = retired_page >> AMDGPU_GPU_PAGE_SHIFT;
-			err_rec->ts = (uint64_t)ktime_get_real_seconds();
-			err_rec->err_type = AMDGPU_RAS_EEPROM_ERR_NON_RECOVERABLE;
-			err_rec->cu = 0;
-			err_rec->mem_channel = channel_index;
-			err_rec->mcumc_id = umc_inst;
-
-			err_data->err_addr_cnt++;
-		}
+		umc_v6_7_convert_error_address(adev, err_data, err_addr,
+					ch_inst, umc_inst);
 	}
 
 	/* clear umc status */
 	WREG64_PCIE((mc_umc_status_addr + umc_reg_offset) * 4, 0x0ULL);
+
+	return 0;
 }
 
 static void umc_v6_7_query_ras_error_address(struct amdgpu_device *adev,
 					     void *ras_error_status)
 {
-	struct ras_err_data *err_data = (struct ras_err_data *)ras_error_status;
-
-	uint32_t umc_inst        = 0;
-	uint32_t ch_inst         = 0;
-	uint32_t umc_reg_offset  = 0;
-
-	/*TODO: driver needs to toggle DF Cstate to ensure
-	 * safe access of UMC resgisters. Will add the protection
-	 * when firmware interface is ready */
-	LOOP_UMC_INST_AND_CH(umc_inst, ch_inst) {
-		umc_reg_offset = get_umc_v6_7_reg_offset(adev,
-							 umc_inst,
-							 ch_inst);
-		umc_v6_7_query_error_address(adev,
-					     err_data,
-					     umc_reg_offset,
-					     ch_inst,
-					     umc_inst);
-	}
+	amdgpu_umc_loop_channels(adev,
+		umc_v6_7_query_error_address, ras_error_status);
 }
 
 static uint32_t umc_v6_7_query_ras_poison_mode_per_channel(
@@ -304,27 +504,25 @@ static uint32_t umc_v6_7_query_ras_poison_mode_per_channel(
 
 static bool umc_v6_7_query_ras_poison_mode(struct amdgpu_device *adev)
 {
-	uint32_t umc_inst        = 0;
-	uint32_t ch_inst         = 0;
 	uint32_t umc_reg_offset  = 0;
 
-	LOOP_UMC_INST_AND_CH(umc_inst, ch_inst) {
-		umc_reg_offset = get_umc_v6_7_reg_offset(adev,
-							umc_inst,
-							ch_inst);
-		/* Enabling fatal error in one channel will be considered
-		   as fatal error mode */
-		if (umc_v6_7_query_ras_poison_mode_per_channel(adev, umc_reg_offset))
-			return false;
-	}
-
-	return true;
+	/* Enabling fatal error in umc instance0 channel0 will be
+	 * considered as fatal error mode
+	 */
+	umc_reg_offset = get_umc_v6_7_reg_offset(adev, 0, 0);
+	return !umc_v6_7_query_ras_poison_mode_per_channel(adev, umc_reg_offset);
 }
 
-const struct amdgpu_umc_ras_funcs umc_v6_7_ras_funcs = {
-	.ras_late_init = amdgpu_umc_ras_late_init,
-	.ras_fini = amdgpu_umc_ras_fini,
+const struct amdgpu_ras_block_hw_ops umc_v6_7_ras_hw_ops = {
 	.query_ras_error_count = umc_v6_7_query_ras_error_count,
 	.query_ras_error_address = umc_v6_7_query_ras_error_address,
+};
+
+struct amdgpu_umc_ras umc_v6_7_ras = {
+	.ras_block = {
+		.hw_ops = &umc_v6_7_ras_hw_ops,
+	},
 	.query_ras_poison_mode = umc_v6_7_query_ras_poison_mode,
+	.ecc_info_query_ras_error_count = umc_v6_7_ecc_info_query_ras_error_count,
+	.ecc_info_query_ras_error_address = umc_v6_7_ecc_info_query_ras_error_address,
 };
diff --git a/drivers/gpu/drm/amd/amdgpu/umc_v6_7.h b/drivers/gpu/drm/amd/amdgpu/umc_v6_7.h
index 57f2557e7aca..105245d5b6e5 100644
--- a/drivers/gpu/drm/amd/amdgpu/umc_v6_7.h
+++ b/drivers/gpu/drm/amd/amdgpu/umc_v6_7.h
@@ -41,12 +41,37 @@
 #define UMC_V6_7_CHANNEL_INSTANCE_NUM		8
 /* total channel instances in one umc block */
 #define UMC_V6_7_TOTAL_CHANNEL_NUM	(UMC_V6_7_CHANNEL_INSTANCE_NUM * UMC_V6_7_UMC_INSTANCE_NUM)
+/* one piece of normalizing address is mapped to 8 pieces of physical address */
+#define UMC_V6_7_NA_MAP_PA_NUM	8
+/* R14 bit shift should be considered, double the number */
+#define UMC_V6_7_BAD_PAGE_NUM_PER_CHANNEL	(UMC_V6_7_NA_MAP_PA_NUM * 2)
+/* The CH4 bit in SOC physical address */
+#define UMC_V6_7_PA_CH4_BIT	12
+/* The C2 bit in SOC physical address */
+#define UMC_V6_7_PA_C2_BIT	17
+/* The R14 bit in SOC physical address */
+#define UMC_V6_7_PA_R14_BIT	34
 /* UMC regiser per channel offset */
 #define UMC_V6_7_PER_CHANNEL_OFFSET		0x400
-extern const struct amdgpu_umc_ras_funcs umc_v6_7_ras_funcs;
+
+/* XOR bit 20, 25, 34 of PA into CH4 bit (bit 12 of PA),
+ * hash bit is only effective when related setting is enabled
+ */
+#define CHANNEL_HASH(channel_idx, pa) (((channel_idx) >> 4) ^ \
+			(((pa)  >> 20) & 0x1ULL & adev->df.hash_status.hash_64k) ^ \
+			(((pa)  >> 25) & 0x1ULL & adev->df.hash_status.hash_2m) ^ \
+			(((pa)  >> 34) & 0x1ULL & adev->df.hash_status.hash_1g))
+#define SET_CHANNEL_HASH(channel_idx, pa) do { \
+		(pa) &= ~(0x1ULL << UMC_V6_7_PA_CH4_BIT); \
+		(pa) |= (CHANNEL_HASH(channel_idx, pa) << UMC_V6_7_PA_CH4_BIT); \
+	} while (0)
+
+extern struct amdgpu_umc_ras umc_v6_7_ras;
 extern const uint32_t
 	umc_v6_7_channel_idx_tbl_second[UMC_V6_7_UMC_INSTANCE_NUM][UMC_V6_7_CHANNEL_INSTANCE_NUM];
 extern const uint32_t
 	umc_v6_7_channel_idx_tbl_first[UMC_V6_7_UMC_INSTANCE_NUM][UMC_V6_7_CHANNEL_INSTANCE_NUM];
-
+void umc_v6_7_convert_error_address(struct amdgpu_device *adev,
+                                    struct ras_err_data *err_data, uint64_t err_addr,
+                                    uint32_t ch_inst, uint32_t umc_inst);
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/umc_v8_10.c b/drivers/gpu/drm/amd/amdgpu/umc_v8_10.c
new file mode 100644
index 000000000000..a32f87992f20
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/umc_v8_10.c
@@ -0,0 +1,458 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include "umc_v8_10.h"
+#include "amdgpu_ras.h"
+#include "amdgpu_umc.h"
+#include "amdgpu.h"
+#include "umc/umc_8_10_0_offset.h"
+#include "umc/umc_8_10_0_sh_mask.h"
+
+#define UMC_8_NODE_DIST   0x800000
+#define UMC_8_INST_DIST   0x4000
+
+struct channelnum_map_colbit {
+	uint32_t channel_num;
+	uint32_t col_bit;
+};
+
+const struct channelnum_map_colbit umc_v8_10_channelnum_map_colbit_table[] = {
+	{24, 13},
+	{20, 13},
+	{16, 12},
+	{14, 12},
+	{12, 12},
+	{10, 12},
+	{6,  11},
+};
+
+const uint32_t
+	umc_v8_10_channel_idx_tbl_ext0[]
+				[UMC_V8_10_UMC_INSTANCE_NUM]
+				[UMC_V8_10_CHANNEL_INSTANCE_NUM] = {
+	   {{1,   5}, {7,  3}},
+	   {{14, 15}, {13, 12}},
+	   {{10, 11}, {9,  8}},
+	   {{6,   2}, {0,  4}}
+	};
+
+const uint32_t
+	umc_v8_10_channel_idx_tbl[]
+				[UMC_V8_10_UMC_INSTANCE_NUM]
+				[UMC_V8_10_CHANNEL_INSTANCE_NUM] = {
+	   {{16, 18}, {17, 19}},
+	   {{15, 11}, {3,   7}},
+	   {{1,   5}, {13,  9}},
+	   {{23, 21}, {22, 20}},
+	   {{0,   4}, {12,  8}},
+	   {{14, 10}, {2,   6}}
+	};
+
+static inline uint32_t get_umc_v8_10_reg_offset(struct amdgpu_device *adev,
+					    uint32_t node_inst,
+					    uint32_t umc_inst,
+					    uint32_t ch_inst)
+{
+	return adev->umc.channel_offs * ch_inst + UMC_8_INST_DIST * umc_inst +
+		UMC_8_NODE_DIST * node_inst;
+}
+
+static int umc_v8_10_clear_error_count_per_channel(struct amdgpu_device *adev,
+					uint32_t node_inst, uint32_t umc_inst,
+					uint32_t ch_inst, void *data)
+{
+	uint32_t ecc_err_cnt_addr;
+	uint32_t umc_reg_offset =
+		get_umc_v8_10_reg_offset(adev, node_inst, umc_inst, ch_inst);
+
+	ecc_err_cnt_addr =
+		SOC15_REG_OFFSET(UMC, 0, regUMCCH0_0_GeccErrCnt);
+
+	/* clear error count */
+	WREG32_PCIE((ecc_err_cnt_addr + umc_reg_offset) * 4,
+			UMC_V8_10_CE_CNT_INIT);
+
+	return 0;
+}
+
+static void umc_v8_10_clear_error_count(struct amdgpu_device *adev)
+{
+	amdgpu_umc_loop_channels(adev,
+		umc_v8_10_clear_error_count_per_channel, NULL);
+}
+
+static void umc_v8_10_query_correctable_error_count(struct amdgpu_device *adev,
+						   uint32_t umc_reg_offset,
+						   unsigned long *error_count)
+{
+	uint64_t mc_umc_status;
+	uint32_t mc_umc_status_addr;
+
+	/* UMC 8_10 registers */
+	mc_umc_status_addr =
+		SOC15_REG_OFFSET(UMC, 0, regMCA_UMC_UMC0_MCUMC_STATUST0);
+
+	/* Rely on MCUMC_STATUS for correctable error counter
+	 * MCUMC_STATUS is a 64 bit register
+	 */
+	mc_umc_status = RREG64_PCIE((mc_umc_status_addr + umc_reg_offset) * 4);
+	if (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Val) == 1 &&
+	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, CECC) == 1)
+		*error_count += 1;
+}
+
+static void umc_v8_10_query_uncorrectable_error_count(struct amdgpu_device *adev,
+						      uint32_t umc_reg_offset,
+						      unsigned long *error_count)
+{
+	uint64_t mc_umc_status;
+	uint32_t mc_umc_status_addr;
+
+	mc_umc_status_addr = SOC15_REG_OFFSET(UMC, 0, regMCA_UMC_UMC0_MCUMC_STATUST0);
+
+	/* Check the MCUMC_STATUS. */
+	mc_umc_status = RREG64_PCIE((mc_umc_status_addr + umc_reg_offset) * 4);
+	if ((REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Val) == 1) &&
+	    (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Deferred) == 1 ||
+	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UECC) == 1 ||
+	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, PCC) == 1 ||
+	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UC) == 1 ||
+	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, TCC) == 1))
+		*error_count += 1;
+}
+
+static int umc_v8_10_query_ecc_error_count(struct amdgpu_device *adev,
+					uint32_t node_inst, uint32_t umc_inst,
+					uint32_t ch_inst, void *data)
+{
+	struct ras_err_data *err_data = (struct ras_err_data *)data;
+	uint32_t umc_reg_offset =
+		get_umc_v8_10_reg_offset(adev, node_inst, umc_inst, ch_inst);
+
+	umc_v8_10_query_correctable_error_count(adev,
+					umc_reg_offset,
+					&(err_data->ce_count));
+	umc_v8_10_query_uncorrectable_error_count(adev,
+					umc_reg_offset,
+					&(err_data->ue_count));
+
+	return 0;
+}
+
+static void umc_v8_10_query_ras_error_count(struct amdgpu_device *adev,
+					   void *ras_error_status)
+{
+	amdgpu_umc_loop_channels(adev,
+		umc_v8_10_query_ecc_error_count, ras_error_status);
+
+	umc_v8_10_clear_error_count(adev);
+}
+
+static uint32_t umc_v8_10_get_col_bit(uint32_t channel_num)
+{
+	uint32_t t = 0;
+
+	for (t = 0; t < ARRAY_SIZE(umc_v8_10_channelnum_map_colbit_table); t++)
+		if (channel_num == umc_v8_10_channelnum_map_colbit_table[t].channel_num)
+			return umc_v8_10_channelnum_map_colbit_table[t].col_bit;
+
+	/* Failed to get col_bit. */
+	return U32_MAX;
+}
+
+/*
+ * Mapping normal address to soc physical address in swizzle mode.
+ */
+static int umc_v8_10_swizzle_mode_na_to_pa(struct amdgpu_device *adev,
+					uint32_t channel_idx,
+					uint64_t na, uint64_t *soc_pa)
+{
+	uint32_t channel_num = UMC_V8_10_TOTAL_CHANNEL_NUM(adev);
+	uint32_t col_bit = umc_v8_10_get_col_bit(channel_num);
+	uint64_t tmp_addr;
+
+	if (col_bit == U32_MAX)
+		return -1;
+
+	tmp_addr = SWIZZLE_MODE_TMP_ADDR(na, channel_num, channel_idx);
+	*soc_pa = SWIZZLE_MODE_ADDR_HI(tmp_addr, col_bit) |
+		SWIZZLE_MODE_ADDR_MID(na, col_bit) |
+		SWIZZLE_MODE_ADDR_LOW(tmp_addr, col_bit) |
+		SWIZZLE_MODE_ADDR_LSB(na);
+
+	return 0;
+}
+
+static void umc_v8_10_convert_error_address(struct amdgpu_device *adev,
+					    struct ras_err_data *err_data, uint64_t err_addr,
+					    uint32_t ch_inst, uint32_t umc_inst,
+					    uint32_t node_inst, uint64_t mc_umc_status)
+{
+	uint64_t na_err_addr_base;
+	uint64_t na_err_addr, retired_page_addr;
+	uint32_t channel_index, addr_lsb, col = 0;
+	int ret = 0;
+
+	channel_index =
+		adev->umc.channel_idx_tbl[node_inst * adev->umc.umc_inst_num *
+					adev->umc.channel_inst_num +
+					umc_inst * adev->umc.channel_inst_num +
+					ch_inst];
+
+	/* the lowest lsb bits should be ignored */
+	addr_lsb = REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, AddrLsb);
+	err_addr &= ~((0x1ULL << addr_lsb) - 1);
+	na_err_addr_base = err_addr & ~(0x3ULL << UMC_V8_10_NA_C5_BIT);
+
+	/* loop for all possibilities of [C6 C5] in normal address. */
+	for (col = 0; col < UMC_V8_10_NA_COL_2BITS_POWER_OF_2_NUM; col++) {
+		na_err_addr = na_err_addr_base | (col << UMC_V8_10_NA_C5_BIT);
+
+		/* Mapping normal error address to retired soc physical address. */
+		ret = umc_v8_10_swizzle_mode_na_to_pa(adev, channel_index,
+						na_err_addr, &retired_page_addr);
+		if (ret) {
+			dev_err(adev->dev, "Failed to map pa from umc na.\n");
+			break;
+		}
+		dev_info(adev->dev, "Error Address(PA): 0x%llx\n",
+			retired_page_addr);
+		amdgpu_umc_fill_error_record(err_data, na_err_addr,
+				retired_page_addr, channel_index, umc_inst);
+	}
+}
+
+static int umc_v8_10_query_error_address(struct amdgpu_device *adev,
+					uint32_t node_inst, uint32_t umc_inst,
+					uint32_t ch_inst, void *data)
+{
+	uint64_t mc_umc_status_addr;
+	uint64_t mc_umc_status, err_addr;
+	uint64_t mc_umc_addrt0;
+	struct ras_err_data *err_data = (struct ras_err_data *)data;
+	uint32_t umc_reg_offset =
+		get_umc_v8_10_reg_offset(adev, node_inst, umc_inst, ch_inst);
+
+	mc_umc_status_addr =
+		SOC15_REG_OFFSET(UMC, 0, regMCA_UMC_UMC0_MCUMC_STATUST0);
+	mc_umc_status = RREG64_PCIE((mc_umc_status_addr + umc_reg_offset) * 4);
+
+	if (mc_umc_status == 0)
+		return 0;
+
+	if (!err_data->err_addr) {
+		/* clear umc status */
+		WREG64_PCIE((mc_umc_status_addr + umc_reg_offset) * 4, 0x0ULL);
+		return 0;
+	}
+
+	/* calculate error address if ue error is detected */
+	if (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Val) == 1 &&
+	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, AddrV) == 1 &&
+	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UECC) == 1) {
+
+		mc_umc_addrt0 = SOC15_REG_OFFSET(UMC, 0, regMCA_UMC_UMC0_MCUMC_ADDRT0);
+		err_addr = RREG64_PCIE((mc_umc_addrt0 + umc_reg_offset) * 4);
+		err_addr = REG_GET_FIELD(err_addr, MCA_UMC_UMC0_MCUMC_ADDRT0, ErrorAddr);
+
+		umc_v8_10_convert_error_address(adev, err_data, err_addr,
+					ch_inst, umc_inst, node_inst, mc_umc_status);
+	}
+
+	/* clear umc status */
+	WREG64_PCIE((mc_umc_status_addr + umc_reg_offset) * 4, 0x0ULL);
+
+	return 0;
+}
+
+static void umc_v8_10_query_ras_error_address(struct amdgpu_device *adev,
+					     void *ras_error_status)
+{
+	amdgpu_umc_loop_channels(adev,
+		umc_v8_10_query_error_address, ras_error_status);
+}
+
+static int umc_v8_10_err_cnt_init_per_channel(struct amdgpu_device *adev,
+					uint32_t node_inst, uint32_t umc_inst,
+					uint32_t ch_inst, void *data)
+{
+	uint32_t ecc_err_cnt_sel, ecc_err_cnt_sel_addr;
+	uint32_t ecc_err_cnt_addr;
+	uint32_t umc_reg_offset =
+		get_umc_v8_10_reg_offset(adev, node_inst, umc_inst, ch_inst);
+
+	ecc_err_cnt_sel_addr =
+		SOC15_REG_OFFSET(UMC, 0, regUMCCH0_0_GeccErrCntSel);
+	ecc_err_cnt_addr =
+		SOC15_REG_OFFSET(UMC, 0, regUMCCH0_0_GeccErrCnt);
+
+	ecc_err_cnt_sel = RREG32_PCIE((ecc_err_cnt_sel_addr + umc_reg_offset) * 4);
+
+	/* set ce error interrupt type to APIC based interrupt */
+	ecc_err_cnt_sel = REG_SET_FIELD(ecc_err_cnt_sel, UMCCH0_0_GeccErrCntSel,
+					GeccErrInt, 0x1);
+	WREG32_PCIE((ecc_err_cnt_sel_addr + umc_reg_offset) * 4, ecc_err_cnt_sel);
+	/* set error count to initial value */
+	WREG32_PCIE((ecc_err_cnt_addr + umc_reg_offset) * 4, UMC_V8_10_CE_CNT_INIT);
+
+	return 0;
+}
+
+static void umc_v8_10_err_cnt_init(struct amdgpu_device *adev)
+{
+	amdgpu_umc_loop_channels(adev,
+		umc_v8_10_err_cnt_init_per_channel, NULL);
+}
+
+static bool umc_v8_10_query_ras_poison_mode(struct amdgpu_device *adev)
+{
+	/*
+	 * Force return true, because UMCCH0_0_GeccCtrl
+	 * is not accessible from host side
+	 */
+	return true;
+}
+
+static void umc_v8_10_ecc_info_query_correctable_error_count(struct amdgpu_device *adev,
+				      uint32_t node_inst, uint32_t umc_inst, uint32_t ch_inst,
+				      unsigned long *error_count)
+{
+	uint16_t ecc_ce_cnt;
+	uint32_t eccinfo_table_idx;
+	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
+
+	eccinfo_table_idx = node_inst * adev->umc.umc_inst_num *
+				  adev->umc.channel_inst_num +
+				  umc_inst * adev->umc.channel_inst_num +
+				  ch_inst;
+
+	/* Retrieve CE count */
+	ecc_ce_cnt = ras->umc_ecc.ecc[eccinfo_table_idx].ce_count_lo_chip;
+	if (ecc_ce_cnt)
+		*error_count += ecc_ce_cnt;
+}
+
+static void umc_v8_10_ecc_info_query_uncorrectable_error_count(struct amdgpu_device *adev,
+				      uint32_t node_inst, uint32_t umc_inst, uint32_t ch_inst,
+				      unsigned long *error_count)
+{
+	uint64_t mc_umc_status;
+	uint32_t eccinfo_table_idx;
+	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
+
+	eccinfo_table_idx = node_inst * adev->umc.umc_inst_num *
+				  adev->umc.channel_inst_num +
+				  umc_inst * adev->umc.channel_inst_num +
+				  ch_inst;
+
+	/* check the MCUMC_STATUS */
+	mc_umc_status = ras->umc_ecc.ecc[eccinfo_table_idx].mca_umc_status;
+	if ((REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Val) == 1) &&
+	    (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Deferred) == 1 ||
+	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UECC) == 1 ||
+	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, PCC) == 1 ||
+	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UC) == 1 ||
+	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, TCC) == 1)) {
+		*error_count += 1;
+	}
+}
+
+static int umc_v8_10_ecc_info_query_ecc_error_count(struct amdgpu_device *adev,
+					uint32_t node_inst, uint32_t umc_inst,
+					uint32_t ch_inst, void *data)
+{
+	struct ras_err_data *err_data = (struct ras_err_data *)data;
+
+	umc_v8_10_ecc_info_query_correctable_error_count(adev,
+					node_inst, umc_inst, ch_inst,
+					&(err_data->ce_count));
+	umc_v8_10_ecc_info_query_uncorrectable_error_count(adev,
+					node_inst, umc_inst, ch_inst,
+					&(err_data->ue_count));
+	return 0;
+}
+
+static void umc_v8_10_ecc_info_query_ras_error_count(struct amdgpu_device *adev,
+					void *ras_error_status)
+{
+	amdgpu_umc_loop_channels(adev,
+		umc_v8_10_ecc_info_query_ecc_error_count, ras_error_status);
+}
+
+static int umc_v8_10_ecc_info_query_error_address(struct amdgpu_device *adev,
+					uint32_t node_inst, uint32_t umc_inst,
+					uint32_t ch_inst, void *data)
+{
+	uint32_t eccinfo_table_idx;
+	uint64_t mc_umc_status, err_addr;
+	struct ras_err_data *err_data = (struct ras_err_data *)data;
+	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
+
+	eccinfo_table_idx = node_inst * adev->umc.umc_inst_num *
+				  adev->umc.channel_inst_num +
+				  umc_inst * adev->umc.channel_inst_num +
+				  ch_inst;
+
+	mc_umc_status = ras->umc_ecc.ecc[eccinfo_table_idx].mca_umc_status;
+
+	if (mc_umc_status == 0)
+		return 0;
+
+	if (!err_data->err_addr)
+		return 0;
+
+	/* calculate error address if ue error is detected */
+	if (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Val) == 1 &&
+	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, AddrV) == 1 &&
+	    (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UECC) == 1)) {
+
+		err_addr = ras->umc_ecc.ecc[eccinfo_table_idx].mca_umc_addr;
+		err_addr = REG_GET_FIELD(err_addr, MCA_UMC_UMC0_MCUMC_ADDRT0, ErrorAddr);
+
+		umc_v8_10_convert_error_address(adev, err_data, err_addr,
+					ch_inst, umc_inst, node_inst, mc_umc_status);
+	}
+
+	return 0;
+}
+
+static void umc_v8_10_ecc_info_query_ras_error_address(struct amdgpu_device *adev,
+					void *ras_error_status)
+{
+	amdgpu_umc_loop_channels(adev,
+		umc_v8_10_ecc_info_query_error_address, ras_error_status);
+}
+
+const struct amdgpu_ras_block_hw_ops umc_v8_10_ras_hw_ops = {
+	.query_ras_error_count = umc_v8_10_query_ras_error_count,
+	.query_ras_error_address = umc_v8_10_query_ras_error_address,
+};
+
+struct amdgpu_umc_ras umc_v8_10_ras = {
+	.ras_block = {
+		.hw_ops = &umc_v8_10_ras_hw_ops,
+	},
+	.err_cnt_init = umc_v8_10_err_cnt_init,
+	.query_ras_poison_mode = umc_v8_10_query_ras_poison_mode,
+	.ecc_info_query_ras_error_count = umc_v8_10_ecc_info_query_ras_error_count,
+	.ecc_info_query_ras_error_address = umc_v8_10_ecc_info_query_ras_error_address,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/umc_v8_10.h b/drivers/gpu/drm/amd/amdgpu/umc_v8_10.h
new file mode 100644
index 000000000000..dc12e0af5451
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/umc_v8_10.h
@@ -0,0 +1,75 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __UMC_V8_10_H__
+#define __UMC_V8_10_H__
+
+#include "soc15_common.h"
+#include "amdgpu.h"
+
+/* number of umc channel instance with memory map register access */
+#define UMC_V8_10_CHANNEL_INSTANCE_NUM		2
+/* number of umc instance with memory map register access */
+#define UMC_V8_10_UMC_INSTANCE_NUM		2
+
+/* Total channel instances for all available umc nodes */
+#define UMC_V8_10_TOTAL_CHANNEL_NUM(adev) \
+	(UMC_V8_10_CHANNEL_INSTANCE_NUM * UMC_V8_10_UMC_INSTANCE_NUM * \
+	(adev)->gmc.num_umc - hweight32((adev)->gmc.m_half_use) * 2)
+
+/* UMC regiser per channel offset */
+#define UMC_V8_10_PER_CHANNEL_OFFSET	0x400
+
+/* EccErrCnt max value */
+#define UMC_V8_10_CE_CNT_MAX		0xffff
+/* umc ce interrupt threshold */
+#define UUMC_V8_10_CE_INT_THRESHOLD	0xffff
+/* umc ce count initial value */
+#define UMC_V8_10_CE_CNT_INIT	(UMC_V8_10_CE_CNT_MAX - UUMC_V8_10_CE_INT_THRESHOLD)
+
+#define UMC_V8_10_NA_COL_2BITS_POWER_OF_2_NUM	 4
+
+/* The C5 bit in NA  address */
+#define UMC_V8_10_NA_C5_BIT	14
+
+/* Map to swizzle mode address */
+#define SWIZZLE_MODE_TMP_ADDR(na, ch_num, ch_idx) \
+		((((na) >> 10) * (ch_num) + (ch_idx)) << 10)
+#define SWIZZLE_MODE_ADDR_HI(addr, col_bit)  \
+		(((addr) >> ((col_bit) + 2)) << ((col_bit) + 2))
+#define SWIZZLE_MODE_ADDR_MID(na, col_bit) ((((na) >> 8) & 0x3) << (col_bit))
+#define SWIZZLE_MODE_ADDR_LOW(addr, col_bit) \
+		((((addr) >> 10) & ((0x1ULL << (col_bit - 8)) - 1)) << 8)
+#define SWIZZLE_MODE_ADDR_LSB(na) ((na) & 0xFF)
+
+extern struct amdgpu_umc_ras umc_v8_10_ras;
+extern const uint32_t
+	umc_v8_10_channel_idx_tbl[]
+				[UMC_V8_10_UMC_INSTANCE_NUM]
+				[UMC_V8_10_CHANNEL_INSTANCE_NUM];
+
+extern const uint32_t
+	umc_v8_10_channel_idx_tbl_ext0[]
+				[UMC_V8_10_UMC_INSTANCE_NUM]
+				[UMC_V8_10_CHANNEL_INSTANCE_NUM];
+#endif
+
diff --git a/drivers/gpu/drm/amd/amdgpu/umc_v8_14.c b/drivers/gpu/drm/amd/amdgpu/umc_v8_14.c
new file mode 100644
index 000000000000..eaca10a3c4a9
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/umc_v8_14.c
@@ -0,0 +1,160 @@
+/*
+ * Copyright 2024 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include "umc_v8_14.h"
+#include "amdgpu_ras.h"
+#include "amdgpu_umc.h"
+#include "amdgpu.h"
+#include "umc/umc_8_14_0_offset.h"
+#include "umc/umc_8_14_0_sh_mask.h"
+
+static inline uint32_t get_umc_v8_14_reg_offset(struct amdgpu_device *adev,
+					    uint32_t umc_inst,
+					    uint32_t ch_inst)
+{
+	return adev->umc.channel_offs * ch_inst + UMC_V8_14_INST_DIST * umc_inst;
+}
+
+static int umc_v8_14_clear_error_count_per_channel(struct amdgpu_device *adev,
+					uint32_t node_inst, uint32_t umc_inst,
+					uint32_t ch_inst, void *data)
+{
+	uint32_t ecc_err_cnt_addr;
+	uint32_t umc_reg_offset =
+		get_umc_v8_14_reg_offset(adev, umc_inst, ch_inst);
+
+	ecc_err_cnt_addr =
+		SOC15_REG_OFFSET(UMC, 0, regUMCCH0_GeccErrCnt);
+
+	/* clear error count */
+	WREG32_PCIE((ecc_err_cnt_addr + umc_reg_offset) * 4,
+			UMC_V8_14_CE_CNT_INIT);
+
+	return 0;
+}
+
+static void umc_v8_14_clear_error_count(struct amdgpu_device *adev)
+{
+	amdgpu_umc_loop_channels(adev,
+		umc_v8_14_clear_error_count_per_channel, NULL);
+}
+
+static void umc_v8_14_query_correctable_error_count(struct amdgpu_device *adev,
+						   uint32_t umc_reg_offset,
+						   unsigned long *error_count)
+{
+	uint32_t ecc_err_cnt, ecc_err_cnt_addr;
+
+	/* UMC 8_14 registers */
+	ecc_err_cnt_addr =
+		SOC15_REG_OFFSET(UMC, 0, regUMCCH0_GeccErrCnt);
+
+	ecc_err_cnt = RREG32_PCIE((ecc_err_cnt_addr + umc_reg_offset) * 4);
+	*error_count +=
+		(REG_GET_FIELD(ecc_err_cnt, UMCCH0_GeccErrCnt, GeccErrCnt) -
+		 UMC_V8_14_CE_CNT_INIT);
+}
+
+static void umc_v8_14_query_uncorrectable_error_count(struct amdgpu_device *adev,
+						      uint32_t umc_reg_offset,
+						      unsigned long *error_count)
+{
+	uint32_t ecc_err_cnt, ecc_err_cnt_addr;
+	/* UMC 8_14 registers */
+	ecc_err_cnt_addr =
+		SOC15_REG_OFFSET(UMC, 0, regUMCCH0_GeccErrCnt);
+
+	ecc_err_cnt = RREG32_PCIE((ecc_err_cnt_addr + umc_reg_offset) * 4);
+	*error_count +=
+		(REG_GET_FIELD(ecc_err_cnt, UMCCH0_GeccErrCnt, GeccUnCorrErrCnt) -
+		 UMC_V8_14_CE_CNT_INIT);
+}
+
+static int umc_v8_14_query_error_count_per_channel(struct amdgpu_device *adev,
+					uint32_t node_inst, uint32_t umc_inst,
+					uint32_t ch_inst, void *data)
+{
+	struct ras_err_data *err_data = (struct ras_err_data *)data;
+	uint32_t umc_reg_offset =
+		get_umc_v8_14_reg_offset(adev, umc_inst, ch_inst);
+
+	umc_v8_14_query_correctable_error_count(adev,
+					umc_reg_offset,
+					&(err_data->ce_count));
+	umc_v8_14_query_uncorrectable_error_count(adev,
+					umc_reg_offset,
+					&(err_data->ue_count));
+
+	return 0;
+}
+
+static void umc_v8_14_query_ras_error_count(struct amdgpu_device *adev,
+					   void *ras_error_status)
+{
+	amdgpu_umc_loop_channels(adev,
+		umc_v8_14_query_error_count_per_channel, ras_error_status);
+
+	umc_v8_14_clear_error_count(adev);
+}
+
+static int umc_v8_14_err_cnt_init_per_channel(struct amdgpu_device *adev,
+					uint32_t node_inst, uint32_t umc_inst,
+					uint32_t ch_inst, void *data)
+{
+	uint32_t ecc_err_cnt_sel, ecc_err_cnt_sel_addr;
+	uint32_t ecc_err_cnt_addr;
+	uint32_t umc_reg_offset =
+		get_umc_v8_14_reg_offset(adev, umc_inst, ch_inst);
+
+	ecc_err_cnt_sel_addr =
+		SOC15_REG_OFFSET(UMC, 0, regUMCCH0_GeccErrCntSel);
+	ecc_err_cnt_addr =
+		SOC15_REG_OFFSET(UMC, 0, regUMCCH0_GeccErrCnt);
+
+	ecc_err_cnt_sel = RREG32_PCIE((ecc_err_cnt_sel_addr + umc_reg_offset) * 4);
+
+	/* set ce error interrupt type to APIC based interrupt */
+	ecc_err_cnt_sel = REG_SET_FIELD(ecc_err_cnt_sel, UMCCH0_GeccErrCntSel,
+					GeccErrInt, 0x1);
+	WREG32_PCIE((ecc_err_cnt_sel_addr + umc_reg_offset) * 4, ecc_err_cnt_sel);
+	/* set error count to initial value */
+	WREG32_PCIE((ecc_err_cnt_addr + umc_reg_offset) * 4, UMC_V8_14_CE_CNT_INIT);
+
+	return 0;
+}
+
+static void umc_v8_14_err_cnt_init(struct amdgpu_device *adev)
+{
+	amdgpu_umc_loop_channels(adev,
+		umc_v8_14_err_cnt_init_per_channel, NULL);
+}
+
+const struct amdgpu_ras_block_hw_ops umc_v8_14_ras_hw_ops = {
+	.query_ras_error_count = umc_v8_14_query_ras_error_count,
+};
+
+struct amdgpu_umc_ras umc_v8_14_ras = {
+	.ras_block = {
+		.hw_ops = &umc_v8_14_ras_hw_ops,
+	},
+	.err_cnt_init = umc_v8_14_err_cnt_init,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/umc_v8_14.h b/drivers/gpu/drm/amd/amdgpu/umc_v8_14.h
new file mode 100644
index 000000000000..20a258f0017a
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/umc_v8_14.h
@@ -0,0 +1,51 @@
+/*
+ * Copyright 2024 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __UMC_V8_14_H__
+#define __UMC_V8_14_H__
+
+#include "soc15_common.h"
+#include "amdgpu.h"
+
+/* number of umc channel instance with memory map register access */
+#define UMC_V8_14_CHANNEL_INSTANCE_NUM		2
+/* number of umc instance with memory map register access */
+#define UMC_V8_14_UMC_INSTANCE_NUM(adev)		((adev)->umc.node_inst_num)
+
+/* Total channel instances for all available umc nodes */
+#define UMC_V8_14_TOTAL_CHANNEL_NUM(adev) \
+	(UMC_V8_14_CHANNEL_INSTANCE_NUM * (adev)->gmc.num_umc)
+
+/* UMC register per channel offset */
+#define UMC_V8_14_PER_CHANNEL_OFFSET		0x400
+
+#define UMC_V8_14_INST_DIST		0x40000
+
+/* EccErrCnt max value */
+#define UMC_V8_14_CE_CNT_MAX		0xffff
+/* umc ce interrupt threshold */
+#define UMC_V8_14_CE_INT_THRESHOLD		0xffff
+/* umc ce count initial value */
+#define UMC_V8_14_CE_CNT_INIT	(UMC_V8_14_CE_CNT_MAX - UMC_V8_14_CE_INT_THRESHOLD)
+
+extern struct amdgpu_umc_ras umc_v8_14_ras;
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/umc_v8_7.c b/drivers/gpu/drm/amd/amdgpu/umc_v8_7.c
index af59a35788e3..b717fdaa46e4 100644
--- a/drivers/gpu/drm/amd/amdgpu/umc_v8_7.c
+++ b/drivers/gpu/drm/amd/amdgpu/umc_v8_7.c
@@ -40,13 +40,143 @@ const uint32_t
 		{9, 0},   {15, 6}
 };
 
-static inline uint32_t get_umc_8_reg_offset(struct amdgpu_device *adev,
+static inline uint32_t get_umc_v8_7_reg_offset(struct amdgpu_device *adev,
 					    uint32_t umc_inst,
 					    uint32_t ch_inst)
 {
 	return adev->umc.channel_offs*ch_inst + UMC_8_INST_DIST*umc_inst;
 }
 
+static void umc_v8_7_ecc_info_query_correctable_error_count(struct amdgpu_device *adev,
+						uint32_t umc_inst, uint32_t ch_inst,
+						unsigned long *error_count)
+{
+	uint64_t mc_umc_status;
+	uint32_t eccinfo_table_idx;
+	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
+
+	eccinfo_table_idx = umc_inst * adev->umc.channel_inst_num + ch_inst;
+
+	/* check for SRAM correctable error
+	 * MCUMC_STATUS is a 64 bit register
+	 */
+	mc_umc_status = ras->umc_ecc.ecc[eccinfo_table_idx].mca_umc_status;
+	if (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Val) == 1 &&
+	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, CECC) == 1)
+		*error_count += 1;
+}
+
+static void umc_v8_7_ecc_info_querry_uncorrectable_error_count(struct amdgpu_device *adev,
+							uint32_t umc_inst, uint32_t ch_inst,
+							unsigned long *error_count)
+{
+	uint64_t mc_umc_status;
+	uint32_t eccinfo_table_idx;
+	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
+
+	eccinfo_table_idx = umc_inst * adev->umc.channel_inst_num + ch_inst;
+
+	/* check the MCUMC_STATUS */
+	mc_umc_status = ras->umc_ecc.ecc[eccinfo_table_idx].mca_umc_status;
+	if ((REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Val) == 1) &&
+	    (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Deferred) == 1 ||
+	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UECC) == 1 ||
+	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, PCC) == 1 ||
+	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UC) == 1 ||
+	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, TCC) == 1))
+		*error_count += 1;
+}
+
+static void umc_v8_7_ecc_info_query_ras_error_count(struct amdgpu_device *adev,
+					void *ras_error_status)
+{
+	struct ras_err_data *err_data = (struct ras_err_data *)ras_error_status;
+
+	uint32_t umc_inst        = 0;
+	uint32_t ch_inst         = 0;
+
+	/* TODO: driver needs to toggle DF Cstate to ensure
+	 * safe access of UMC registers. Will add the protection
+	 */
+	LOOP_UMC_INST_AND_CH(umc_inst, ch_inst) {
+		umc_v8_7_ecc_info_query_correctable_error_count(adev,
+							umc_inst, ch_inst,
+							&(err_data->ce_count));
+		umc_v8_7_ecc_info_querry_uncorrectable_error_count(adev,
+							umc_inst, ch_inst,
+							&(err_data->ue_count));
+	}
+}
+
+static void umc_v8_7_convert_error_address(struct amdgpu_device *adev,
+					struct ras_err_data *err_data, uint64_t err_addr,
+					uint32_t ch_inst, uint32_t umc_inst)
+{
+	uint64_t retired_page;
+	uint32_t channel_index;
+
+	channel_index =
+		adev->umc.channel_idx_tbl[umc_inst * adev->umc.channel_inst_num + ch_inst];
+
+	/* translate umc channel address to soc pa, 3 parts are included */
+	retired_page = ADDR_OF_4KB_BLOCK(err_addr) |
+			ADDR_OF_256B_BLOCK(channel_index) |
+			OFFSET_IN_256B_BLOCK(err_addr);
+
+	amdgpu_umc_fill_error_record(err_data, err_addr,
+				retired_page, channel_index, umc_inst);
+}
+
+static void umc_v8_7_ecc_info_query_error_address(struct amdgpu_device *adev,
+					struct ras_err_data *err_data,
+					uint32_t ch_inst,
+					uint32_t umc_inst)
+{
+	uint64_t mc_umc_status, err_addr;
+	uint32_t eccinfo_table_idx;
+	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
+
+	eccinfo_table_idx = umc_inst * adev->umc.channel_inst_num + ch_inst;
+	mc_umc_status = ras->umc_ecc.ecc[eccinfo_table_idx].mca_umc_status;
+
+	if (mc_umc_status == 0)
+		return;
+
+	if (!err_data->err_addr)
+		return;
+
+	/* calculate error address if ue error is detected */
+	if (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Val) == 1 &&
+	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UECC) == 1) {
+
+		err_addr = ras->umc_ecc.ecc[eccinfo_table_idx].mca_umc_addr;
+		err_addr = REG_GET_FIELD(err_addr, MCA_UMC_UMC0_MCUMC_ADDRT0, ErrorAddr);
+
+		umc_v8_7_convert_error_address(adev, err_data, err_addr,
+						ch_inst, umc_inst);
+	}
+}
+
+static void umc_v8_7_ecc_info_query_ras_error_address(struct amdgpu_device *adev,
+					void *ras_error_status)
+{
+	struct ras_err_data *err_data = (struct ras_err_data *)ras_error_status;
+
+	uint32_t umc_inst        = 0;
+	uint32_t ch_inst         = 0;
+
+	/* TODO: driver needs to toggle DF Cstate to ensure
+	 * safe access of UMC resgisters. Will add the protection
+	 * when firmware interface is ready
+	 */
+	LOOP_UMC_INST_AND_CH(umc_inst, ch_inst) {
+		umc_v8_7_ecc_info_query_error_address(adev,
+						err_data,
+						ch_inst,
+						umc_inst);
+	}
+}
+
 static void umc_v8_7_clear_error_count_per_channel(struct amdgpu_device *adev,
 					uint32_t umc_reg_offset)
 {
@@ -92,7 +222,7 @@ static void umc_v8_7_clear_error_count(struct amdgpu_device *adev)
 	uint32_t umc_reg_offset  = 0;
 
 	LOOP_UMC_INST_AND_CH(umc_inst, ch_inst) {
-		umc_reg_offset = get_umc_8_reg_offset(adev,
+		umc_reg_offset = get_umc_v8_7_reg_offset(adev,
 						umc_inst,
 						ch_inst);
 
@@ -178,7 +308,7 @@ static void umc_v8_7_query_ras_error_count(struct amdgpu_device *adev,
 	uint32_t umc_reg_offset  = 0;
 
 	LOOP_UMC_INST_AND_CH(umc_inst, ch_inst) {
-		umc_reg_offset = get_umc_8_reg_offset(adev,
+		umc_reg_offset = get_umc_v8_7_reg_offset(adev,
 						      umc_inst,
 						      ch_inst);
 
@@ -200,15 +330,12 @@ static void umc_v8_7_query_error_address(struct amdgpu_device *adev,
 					 uint32_t umc_inst)
 {
 	uint32_t lsb, mc_umc_status_addr;
-	uint64_t mc_umc_status, err_addr, retired_page, mc_umc_addrt0;
-	struct eeprom_table_record *err_rec;
-	uint32_t channel_index = adev->umc.channel_idx_tbl[umc_inst * adev->umc.channel_inst_num + ch_inst];
+	uint64_t mc_umc_status, err_addr, mc_umc_addrt0;
 
 	mc_umc_status_addr =
 		SOC15_REG_OFFSET(UMC, 0, mmMCA_UMC_UMC0_MCUMC_STATUST0);
 	mc_umc_addrt0 =
 		SOC15_REG_OFFSET(UMC, 0, mmMCA_UMC_UMC0_MCUMC_ADDRT0);
-
 	mc_umc_status = RREG64_PCIE((mc_umc_status_addr + umc_reg_offset) * 4);
 
 	if (mc_umc_status == 0)
@@ -220,12 +347,9 @@ static void umc_v8_7_query_error_address(struct amdgpu_device *adev,
 		return;
 	}
 
-	err_rec = &err_data->err_addr[err_data->err_addr_cnt];
-
-	/* calculate error address if ue/ce error is detected */
+	/* calculate error address if ue error is detected */
 	if (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Val) == 1 &&
-	    (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UECC) == 1 ||
-	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, CECC) == 1)) {
+	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UECC) == 1) {
 
 		err_addr = RREG64_PCIE((mc_umc_addrt0 + umc_reg_offset) * 4);
 		/* the lowest lsb bits should be ignored */
@@ -233,25 +357,8 @@ static void umc_v8_7_query_error_address(struct amdgpu_device *adev,
 		err_addr = REG_GET_FIELD(err_addr, MCA_UMC_UMC0_MCUMC_ADDRT0, ErrorAddr);
 		err_addr &= ~((0x1ULL << lsb) - 1);
 
-		/* translate umc channel address to soc pa, 3 parts are included */
-		retired_page = ADDR_OF_4KB_BLOCK(err_addr) |
-				ADDR_OF_256B_BLOCK(channel_index) |
-				OFFSET_IN_256B_BLOCK(err_addr);
-
-		/* we only save ue error information currently, ce is skipped */
-		if (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UECC)
-				== 1) {
-			err_rec->address = err_addr;
-			/* page frame address is saved */
-			err_rec->retired_page = retired_page >> AMDGPU_GPU_PAGE_SHIFT;
-			err_rec->ts = (uint64_t)ktime_get_real_seconds();
-			err_rec->err_type = AMDGPU_RAS_EEPROM_ERR_NON_RECOVERABLE;
-			err_rec->cu = 0;
-			err_rec->mem_channel = channel_index;
-			err_rec->mcumc_id = umc_inst;
-
-			err_data->err_addr_cnt++;
-		}
+		umc_v8_7_convert_error_address(adev, err_data, err_addr,
+								ch_inst, umc_inst);
 	}
 
 	/* clear umc status */
@@ -268,7 +375,7 @@ static void umc_v8_7_query_ras_error_address(struct amdgpu_device *adev,
 	uint32_t umc_reg_offset  = 0;
 
 	LOOP_UMC_INST_AND_CH(umc_inst, ch_inst) {
-		umc_reg_offset = get_umc_8_reg_offset(adev,
+		umc_reg_offset = get_umc_v8_7_reg_offset(adev,
 						      umc_inst,
 						      ch_inst);
 
@@ -316,7 +423,7 @@ static void umc_v8_7_err_cnt_init(struct amdgpu_device *adev)
 	uint32_t umc_reg_offset  = 0;
 
 	LOOP_UMC_INST_AND_CH(umc_inst, ch_inst) {
-		umc_reg_offset = get_umc_8_reg_offset(adev,
+		umc_reg_offset = get_umc_v8_7_reg_offset(adev,
 						      umc_inst,
 						      ch_inst);
 
@@ -324,10 +431,16 @@ static void umc_v8_7_err_cnt_init(struct amdgpu_device *adev)
 	}
 }
 
-const struct amdgpu_umc_ras_funcs umc_v8_7_ras_funcs = {
-	.err_cnt_init = umc_v8_7_err_cnt_init,
-	.ras_late_init = amdgpu_umc_ras_late_init,
-	.ras_fini = amdgpu_umc_ras_fini,
+const struct amdgpu_ras_block_hw_ops umc_v8_7_ras_hw_ops = {
 	.query_ras_error_count = umc_v8_7_query_ras_error_count,
 	.query_ras_error_address = umc_v8_7_query_ras_error_address,
 };
+
+struct amdgpu_umc_ras umc_v8_7_ras = {
+	.ras_block = {
+		.hw_ops = &umc_v8_7_ras_hw_ops,
+	},
+	.err_cnt_init = umc_v8_7_err_cnt_init,
+	.ecc_info_query_ras_error_count = umc_v8_7_ecc_info_query_ras_error_count,
+	.ecc_info_query_ras_error_address = umc_v8_7_ecc_info_query_ras_error_address,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/umc_v8_7.h b/drivers/gpu/drm/amd/amdgpu/umc_v8_7.h
index 37e6dc7c28e0..dd4993f5f78f 100644
--- a/drivers/gpu/drm/amd/amdgpu/umc_v8_7.h
+++ b/drivers/gpu/drm/amd/amdgpu/umc_v8_7.h
@@ -44,7 +44,7 @@
 /* umc ce count initial value */
 #define UMC_V8_7_CE_CNT_INIT	(UMC_V8_7_CE_CNT_MAX - UMC_V8_7_CE_INT_THRESHOLD)
 
-extern const struct amdgpu_umc_ras_funcs umc_v8_7_ras_funcs;
+extern struct amdgpu_umc_ras umc_v8_7_ras;
 extern const uint32_t
 	umc_v8_7_channel_idx_tbl[UMC_V8_7_UMC_INSTANCE_NUM][UMC_V8_7_CHANNEL_INSTANCE_NUM];
 
diff --git a/drivers/gpu/drm/amd/amdgpu/umsch_mm_v4_0.c b/drivers/gpu/drm/amd/amdgpu/umsch_mm_v4_0.c
new file mode 100644
index 000000000000..ce3bb12e3572
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/umsch_mm_v4_0.c
@@ -0,0 +1,434 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/firmware.h>
+#include <linux/module.h>
+#include <linux/debugfs.h>
+#include "amdgpu.h"
+#include "soc15_common.h"
+#include "soc21.h"
+#include "vcn/vcn_4_0_0_offset.h"
+#include "vcn/vcn_4_0_0_sh_mask.h"
+
+#include "amdgpu_umsch_mm.h"
+#include "umsch_mm_4_0_api_def.h"
+#include "umsch_mm_v4_0.h"
+
+#define regUVD_IPX_DLDO_CONFIG                             0x0064
+#define regUVD_IPX_DLDO_CONFIG_BASE_IDX                    1
+#define regUVD_IPX_DLDO_STATUS                             0x0065
+#define regUVD_IPX_DLDO_STATUS_BASE_IDX                    1
+
+#define UVD_IPX_DLDO_CONFIG__ONO0_PWR_CONFIG__SHIFT        0x00000002
+#define UVD_IPX_DLDO_CONFIG__ONO0_PWR_CONFIG_MASK          0x0000000cUL
+#define UVD_IPX_DLDO_STATUS__ONO0_PWR_STATUS__SHIFT        0x00000001
+#define UVD_IPX_DLDO_STATUS__ONO0_PWR_STATUS_MASK          0x00000002UL
+
+static int umsch_mm_v4_0_load_microcode(struct amdgpu_umsch_mm *umsch)
+{
+	struct amdgpu_device *adev = umsch->ring.adev;
+	uint64_t data;
+	int r;
+
+	r = amdgpu_umsch_mm_allocate_ucode_buffer(umsch);
+	if (r)
+		return r;
+
+	r = amdgpu_umsch_mm_allocate_ucode_data_buffer(umsch);
+	if (r)
+		goto err_free_ucode_bo;
+
+	umsch->cmd_buf_curr_ptr = umsch->cmd_buf_ptr;
+
+	if (amdgpu_ip_version(adev, VCN_HWIP, 0) >= IP_VERSION(4, 0, 5)) {
+		WREG32_SOC15(VCN, 0, regUVD_IPX_DLDO_CONFIG,
+			1 << UVD_IPX_DLDO_CONFIG__ONO0_PWR_CONFIG__SHIFT);
+		SOC15_WAIT_ON_RREG(VCN, 0, regUVD_IPX_DLDO_STATUS,
+			0 << UVD_IPX_DLDO_STATUS__ONO0_PWR_STATUS__SHIFT,
+			UVD_IPX_DLDO_STATUS__ONO0_PWR_STATUS_MASK);
+	}
+
+	data = RREG32_SOC15(VCN, 0, regUMSCH_MES_RESET_CTRL);
+	data = REG_SET_FIELD(data, UMSCH_MES_RESET_CTRL, MES_CORE_SOFT_RESET, 0);
+	WREG32_SOC15_UMSCH(regUMSCH_MES_RESET_CTRL, data);
+
+	data = RREG32_SOC15(VCN, 0, regVCN_MES_CNTL);
+	data = REG_SET_FIELD(data, VCN_MES_CNTL, MES_INVALIDATE_ICACHE, 1);
+	data = REG_SET_FIELD(data, VCN_MES_CNTL, MES_PIPE0_RESET, 1);
+	data = REG_SET_FIELD(data, VCN_MES_CNTL, MES_PIPE0_ACTIVE, 0);
+	data = REG_SET_FIELD(data, VCN_MES_CNTL, MES_HALT, 1);
+	WREG32_SOC15_UMSCH(regVCN_MES_CNTL, data);
+
+	data = RREG32_SOC15(VCN, 0, regVCN_MES_IC_BASE_CNTL);
+	data = REG_SET_FIELD(data, VCN_MES_IC_BASE_CNTL, VMID, 0);
+	data = REG_SET_FIELD(data, VCN_MES_IC_BASE_CNTL, EXE_DISABLE, 0);
+	data = REG_SET_FIELD(data, VCN_MES_IC_BASE_CNTL, CACHE_POLICY, 0);
+	WREG32_SOC15_UMSCH(regVCN_MES_IC_BASE_CNTL, data);
+
+	WREG32_SOC15_UMSCH(regVCN_MES_INTR_ROUTINE_START,
+		lower_32_bits(adev->umsch_mm.irq_start_addr >> 2));
+	WREG32_SOC15_UMSCH(regVCN_MES_INTR_ROUTINE_START_HI,
+		upper_32_bits(adev->umsch_mm.irq_start_addr >> 2));
+
+	WREG32_SOC15_UMSCH(regVCN_MES_PRGRM_CNTR_START,
+		lower_32_bits(adev->umsch_mm.uc_start_addr >> 2));
+	WREG32_SOC15_UMSCH(regVCN_MES_PRGRM_CNTR_START_HI,
+		upper_32_bits(adev->umsch_mm.uc_start_addr >> 2));
+
+	WREG32_SOC15_UMSCH(regVCN_MES_LOCAL_INSTR_BASE_LO, 0);
+	WREG32_SOC15_UMSCH(regVCN_MES_LOCAL_INSTR_BASE_HI, 0);
+
+	data = adev->umsch_mm.uc_start_addr + adev->umsch_mm.ucode_size - 1;
+	WREG32_SOC15_UMSCH(regVCN_MES_LOCAL_INSTR_MASK_LO, lower_32_bits(data));
+	WREG32_SOC15_UMSCH(regVCN_MES_LOCAL_INSTR_MASK_HI, upper_32_bits(data));
+
+	data = adev->firmware.load_type == AMDGPU_FW_LOAD_PSP ?
+	       0 : adev->umsch_mm.ucode_fw_gpu_addr;
+	WREG32_SOC15_UMSCH(regVCN_MES_IC_BASE_LO, lower_32_bits(data));
+	WREG32_SOC15_UMSCH(regVCN_MES_IC_BASE_HI, upper_32_bits(data));
+
+	WREG32_SOC15_UMSCH(regVCN_MES_MIBOUND_LO, 0x1FFFFF);
+
+	WREG32_SOC15_UMSCH(regVCN_MES_LOCAL_BASE0_LO,
+		lower_32_bits(adev->umsch_mm.data_start_addr));
+	WREG32_SOC15_UMSCH(regVCN_MES_LOCAL_BASE0_HI,
+		upper_32_bits(adev->umsch_mm.data_start_addr));
+
+	WREG32_SOC15_UMSCH(regVCN_MES_LOCAL_MASK0_LO,
+		adev->umsch_mm.data_size - 1);
+	WREG32_SOC15_UMSCH(regVCN_MES_LOCAL_MASK0_HI, 0);
+
+	data = adev->firmware.load_type == AMDGPU_FW_LOAD_PSP ?
+	       0 : adev->umsch_mm.data_fw_gpu_addr;
+	WREG32_SOC15_UMSCH(regVCN_MES_DC_BASE_LO, lower_32_bits(data));
+	WREG32_SOC15_UMSCH(regVCN_MES_DC_BASE_HI, upper_32_bits(data));
+
+	WREG32_SOC15_UMSCH(regVCN_MES_MDBOUND_LO, 0x3FFFF);
+
+	data = RREG32_SOC15(VCN, 0, regUVD_UMSCH_FORCE);
+	data = REG_SET_FIELD(data, UVD_UMSCH_FORCE, IC_FORCE_GPUVM, 1);
+	data = REG_SET_FIELD(data, UVD_UMSCH_FORCE, DC_FORCE_GPUVM, 1);
+	WREG32_SOC15_UMSCH(regUVD_UMSCH_FORCE, data);
+
+	data = RREG32_SOC15(VCN, 0, regVCN_MES_IC_OP_CNTL);
+	data = REG_SET_FIELD(data, VCN_MES_IC_OP_CNTL, PRIME_ICACHE, 0);
+	data = REG_SET_FIELD(data, VCN_MES_IC_OP_CNTL, INVALIDATE_CACHE, 1);
+	WREG32_SOC15_UMSCH(regVCN_MES_IC_OP_CNTL, data);
+
+	data = RREG32_SOC15(VCN, 0, regVCN_MES_IC_OP_CNTL);
+	data = REG_SET_FIELD(data, VCN_MES_IC_OP_CNTL, PRIME_ICACHE, 1);
+	WREG32_SOC15_UMSCH(regVCN_MES_IC_OP_CNTL, data);
+
+	WREG32_SOC15_UMSCH(regVCN_MES_GP0_LO, 0);
+	WREG32_SOC15_UMSCH(regVCN_MES_GP0_HI, 0);
+
+#if defined(CONFIG_DEBUG_FS)
+	WREG32_SOC15_UMSCH(regVCN_MES_GP0_LO, lower_32_bits(umsch->log_gpu_addr));
+	WREG32_SOC15_UMSCH(regVCN_MES_GP0_HI, upper_32_bits(umsch->log_gpu_addr));
+#endif
+
+	WREG32_SOC15_UMSCH(regVCN_MES_GP1_LO, 0);
+	WREG32_SOC15_UMSCH(regVCN_MES_GP1_HI, 0);
+
+	data = RREG32_SOC15(VCN, 0, regVCN_MES_CNTL);
+	data = REG_SET_FIELD(data, VCN_MES_CNTL, MES_INVALIDATE_ICACHE, 0);
+	data = REG_SET_FIELD(data, VCN_MES_CNTL, MES_PIPE0_RESET, 0);
+	data = REG_SET_FIELD(data, VCN_MES_CNTL, MES_HALT, 0);
+	data = REG_SET_FIELD(data, VCN_MES_CNTL, MES_PIPE0_ACTIVE, 1);
+	WREG32_SOC15_UMSCH(regVCN_MES_CNTL, data);
+
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP)
+		amdgpu_umsch_mm_psp_execute_cmd_buf(umsch);
+
+	r = SOC15_WAIT_ON_RREG(VCN, 0, regVCN_MES_MSTATUS_LO, 0xAAAAAAAA, 0xFFFFFFFF);
+	if (r) {
+		dev_err(adev->dev, "UMSCH FW Load: Failed, regVCN_MES_MSTATUS_LO: 0x%08x\n",
+			RREG32_SOC15(VCN, 0, regVCN_MES_MSTATUS_LO));
+		goto err_free_data_bo;
+	}
+
+	return 0;
+
+err_free_data_bo:
+	amdgpu_bo_free_kernel(&adev->umsch_mm.data_fw_obj,
+			      &adev->umsch_mm.data_fw_gpu_addr,
+			      (void **)&adev->umsch_mm.data_fw_ptr);
+err_free_ucode_bo:
+	amdgpu_bo_free_kernel(&adev->umsch_mm.ucode_fw_obj,
+			      &adev->umsch_mm.ucode_fw_gpu_addr,
+			      (void **)&adev->umsch_mm.ucode_fw_ptr);
+	return r;
+}
+
+static void umsch_mm_v4_0_aggregated_doorbell_init(struct amdgpu_umsch_mm *umsch)
+{
+	struct amdgpu_device *adev = umsch->ring.adev;
+	uint32_t data;
+
+	data = RREG32_SOC15(VCN, 0, regVCN_AGDB_CTRL0);
+	data = REG_SET_FIELD(data, VCN_AGDB_CTRL0, OFFSET,
+	       umsch->agdb_index[CONTEXT_PRIORITY_LEVEL_REALTIME]);
+	data = REG_SET_FIELD(data, VCN_AGDB_CTRL0, EN, 1);
+	WREG32_SOC15(VCN, 0, regVCN_AGDB_CTRL0, data);
+
+	data = RREG32_SOC15(VCN, 0, regVCN_AGDB_CTRL1);
+	data = REG_SET_FIELD(data, VCN_AGDB_CTRL1, OFFSET,
+	       umsch->agdb_index[CONTEXT_PRIORITY_LEVEL_FOCUS]);
+	data = REG_SET_FIELD(data, VCN_AGDB_CTRL1, EN, 1);
+	WREG32_SOC15(VCN, 0, regVCN_AGDB_CTRL1, data);
+
+	data = RREG32_SOC15(VCN, 0, regVCN_AGDB_CTRL2);
+	data = REG_SET_FIELD(data, VCN_AGDB_CTRL2, OFFSET,
+	       umsch->agdb_index[CONTEXT_PRIORITY_LEVEL_NORMAL]);
+	data = REG_SET_FIELD(data, VCN_AGDB_CTRL2, EN, 1);
+	WREG32_SOC15(VCN, 0, regVCN_AGDB_CTRL2, data);
+
+	data = RREG32_SOC15(VCN, 0, regVCN_AGDB_CTRL3);
+	data = REG_SET_FIELD(data, VCN_AGDB_CTRL3, OFFSET,
+	       umsch->agdb_index[CONTEXT_PRIORITY_LEVEL_IDLE]);
+	data = REG_SET_FIELD(data, VCN_AGDB_CTRL3, EN, 1);
+	WREG32_SOC15(VCN, 0, regVCN_AGDB_CTRL3, data);
+}
+
+static int umsch_mm_v4_0_ring_start(struct amdgpu_umsch_mm *umsch)
+{
+	struct amdgpu_ring *ring = &umsch->ring;
+	struct amdgpu_device *adev = ring->adev;
+	uint32_t data;
+
+	data = RREG32_SOC15(VCN, 0, regVCN_UMSCH_RB_DB_CTRL);
+	data = REG_SET_FIELD(data, VCN_UMSCH_RB_DB_CTRL, OFFSET, ring->doorbell_index);
+	data = REG_SET_FIELD(data, VCN_UMSCH_RB_DB_CTRL, EN, 1);
+	WREG32_SOC15(VCN, 0, regVCN_UMSCH_RB_DB_CTRL, data);
+
+	adev->nbio.funcs->vcn_doorbell_range(adev, ring->use_doorbell,
+		(adev->doorbell_index.vcn.vcn_ring0_1 << 1), 0);
+
+	WREG32_SOC15(VCN, 0, regVCN_UMSCH_RB_BASE_LO, lower_32_bits(ring->gpu_addr));
+	WREG32_SOC15(VCN, 0, regVCN_UMSCH_RB_BASE_HI, upper_32_bits(ring->gpu_addr));
+
+	WREG32_SOC15(VCN, 0, regVCN_UMSCH_RB_SIZE, ring->ring_size);
+
+	ring->wptr = 0;
+
+	data = RREG32_SOC15(VCN, 0, regVCN_RB_ENABLE);
+	data &= ~(VCN_RB_ENABLE__AUDIO_RB_EN_MASK);
+	WREG32_SOC15(VCN, 0, regVCN_RB_ENABLE, data);
+
+	umsch_mm_v4_0_aggregated_doorbell_init(umsch);
+
+	return 0;
+}
+
+static int umsch_mm_v4_0_ring_stop(struct amdgpu_umsch_mm *umsch)
+{
+	struct amdgpu_ring *ring = &umsch->ring;
+	struct amdgpu_device *adev = ring->adev;
+	uint32_t data;
+
+	data = RREG32_SOC15(VCN, 0, regVCN_RB_ENABLE);
+	data = REG_SET_FIELD(data, VCN_RB_ENABLE, UMSCH_RB_EN, 0);
+	WREG32_SOC15(VCN, 0, regVCN_RB_ENABLE, data);
+
+	data = RREG32_SOC15(VCN, 0, regVCN_UMSCH_RB_DB_CTRL);
+	data = REG_SET_FIELD(data, VCN_UMSCH_RB_DB_CTRL, EN, 0);
+	WREG32_SOC15(VCN, 0, regVCN_UMSCH_RB_DB_CTRL, data);
+
+	if (amdgpu_ip_version(adev, VCN_HWIP, 0) >= IP_VERSION(4, 0, 5)) {
+		WREG32_SOC15(VCN, 0, regUVD_IPX_DLDO_CONFIG,
+			2 << UVD_IPX_DLDO_CONFIG__ONO0_PWR_CONFIG__SHIFT);
+		SOC15_WAIT_ON_RREG(VCN, 0, regUVD_IPX_DLDO_STATUS,
+			1 << UVD_IPX_DLDO_STATUS__ONO0_PWR_STATUS__SHIFT,
+			UVD_IPX_DLDO_STATUS__ONO0_PWR_STATUS_MASK);
+	}
+
+	return 0;
+}
+
+static int umsch_mm_v4_0_set_hw_resources(struct amdgpu_umsch_mm *umsch)
+{
+	union UMSCHAPI__SET_HW_RESOURCES set_hw_resources = {};
+	struct amdgpu_device *adev = umsch->ring.adev;
+	int r;
+
+	set_hw_resources.header.type = UMSCH_API_TYPE_SCHEDULER;
+	set_hw_resources.header.opcode = UMSCH_API_SET_HW_RSRC;
+	set_hw_resources.header.dwsize = API_FRAME_SIZE_IN_DWORDS;
+
+	set_hw_resources.vmid_mask_mm_vcn = umsch->vmid_mask_mm_vcn;
+	set_hw_resources.vmid_mask_mm_vpe = umsch->vmid_mask_mm_vpe;
+	set_hw_resources.collaboration_mask_vpe =
+		adev->vpe.collaborate_mode ? 0x3 : 0x0;
+	set_hw_resources.engine_mask = umsch->engine_mask;
+
+	set_hw_resources.vcn0_hqd_mask[0] = umsch->vcn0_hqd_mask;
+	set_hw_resources.vcn1_hqd_mask[0] = umsch->vcn1_hqd_mask;
+	set_hw_resources.vcn_hqd_mask[0] = umsch->vcn_hqd_mask[0];
+	set_hw_resources.vcn_hqd_mask[1] = umsch->vcn_hqd_mask[1];
+	set_hw_resources.vpe_hqd_mask[0] = umsch->vpe_hqd_mask;
+
+	set_hw_resources.g_sch_ctx_gpu_mc_ptr = umsch->sch_ctx_gpu_addr;
+
+	set_hw_resources.enable_level_process_quantum_check = 1;
+
+	memcpy(set_hw_resources.mmhub_base, adev->reg_offset[MMHUB_HWIP][0],
+	       sizeof(uint32_t) * 5);
+	set_hw_resources.mmhub_version =
+		IP_VERSION_MAJ_MIN_REV(amdgpu_ip_version(adev, MMHUB_HWIP, 0));
+
+	memcpy(set_hw_resources.osssys_base, adev->reg_offset[OSSSYS_HWIP][0],
+	       sizeof(uint32_t) * 5);
+	set_hw_resources.osssys_version =
+		IP_VERSION_MAJ_MIN_REV(amdgpu_ip_version(adev, OSSSYS_HWIP, 0));
+
+	set_hw_resources.vcn_version =
+		IP_VERSION_MAJ_MIN_REV(amdgpu_ip_version(adev, VCN_HWIP, 0));
+	set_hw_resources.vpe_version =
+		IP_VERSION_MAJ_MIN_REV(amdgpu_ip_version(adev, VPE_HWIP, 0));
+
+	set_hw_resources.api_status.api_completion_fence_addr = umsch->ring.fence_drv.gpu_addr;
+	set_hw_resources.api_status.api_completion_fence_value = ++umsch->ring.fence_drv.sync_seq;
+
+	r = amdgpu_umsch_mm_submit_pkt(umsch, &set_hw_resources.max_dwords_in_api,
+				       API_FRAME_SIZE_IN_DWORDS);
+	if (r)
+		return r;
+
+	r = amdgpu_umsch_mm_query_fence(umsch);
+	if (r) {
+		dev_err(adev->dev, "UMSCH SET_HW_RESOURCES: Failed\n");
+		return r;
+	}
+
+	return 0;
+}
+
+static int umsch_mm_v4_0_add_queue(struct amdgpu_umsch_mm *umsch,
+				   struct umsch_mm_add_queue_input *input_ptr)
+{
+	struct amdgpu_device *adev = umsch->ring.adev;
+	union UMSCHAPI__ADD_QUEUE add_queue = {};
+	int r;
+
+	add_queue.header.type = UMSCH_API_TYPE_SCHEDULER;
+	add_queue.header.opcode = UMSCH_API_ADD_QUEUE;
+	add_queue.header.dwsize = API_FRAME_SIZE_IN_DWORDS;
+
+	add_queue.process_id = input_ptr->process_id;
+	add_queue.page_table_base_addr = input_ptr->page_table_base_addr;
+	add_queue.process_va_start = input_ptr->process_va_start;
+	add_queue.process_va_end = input_ptr->process_va_end;
+	add_queue.process_quantum = input_ptr->process_quantum;
+	add_queue.process_csa_addr = input_ptr->process_csa_addr;
+	add_queue.context_quantum = input_ptr->context_quantum;
+	add_queue.context_csa_addr = input_ptr->context_csa_addr;
+	add_queue.inprocess_context_priority = input_ptr->inprocess_context_priority;
+	add_queue.context_global_priority_level =
+		(enum UMSCH_AMD_PRIORITY_LEVEL)input_ptr->context_global_priority_level;
+	add_queue.doorbell_offset_0 = input_ptr->doorbell_offset_0;
+	add_queue.doorbell_offset_1 = input_ptr->doorbell_offset_1;
+	add_queue.affinity.u32All = input_ptr->affinity;
+	add_queue.mqd_addr = input_ptr->mqd_addr;
+	add_queue.engine_type = (enum UMSCH_ENGINE_TYPE)input_ptr->engine_type;
+	add_queue.h_context = input_ptr->h_context;
+	add_queue.h_queue = input_ptr->h_queue;
+	add_queue.vm_context_cntl = input_ptr->vm_context_cntl;
+	add_queue.is_context_suspended = input_ptr->is_context_suspended;
+	add_queue.collaboration_mode = adev->vpe.collaborate_mode ? 1 : 0;
+
+	add_queue.api_status.api_completion_fence_addr = umsch->ring.fence_drv.gpu_addr;
+	add_queue.api_status.api_completion_fence_value = ++umsch->ring.fence_drv.sync_seq;
+
+	r = amdgpu_umsch_mm_submit_pkt(umsch, &add_queue.max_dwords_in_api,
+				       API_FRAME_SIZE_IN_DWORDS);
+	if (r)
+		return r;
+
+	r = amdgpu_umsch_mm_query_fence(umsch);
+	if (r) {
+		dev_err(adev->dev, "UMSCH ADD_QUEUE: Failed\n");
+		return r;
+	}
+
+	return 0;
+}
+
+static int umsch_mm_v4_0_remove_queue(struct amdgpu_umsch_mm *umsch,
+				      struct umsch_mm_remove_queue_input *input_ptr)
+{
+	union UMSCHAPI__REMOVE_QUEUE remove_queue = {};
+	struct amdgpu_device *adev = umsch->ring.adev;
+	int r;
+
+	remove_queue.header.type = UMSCH_API_TYPE_SCHEDULER;
+	remove_queue.header.opcode = UMSCH_API_REMOVE_QUEUE;
+	remove_queue.header.dwsize = API_FRAME_SIZE_IN_DWORDS;
+
+	remove_queue.doorbell_offset_0 = input_ptr->doorbell_offset_0;
+	remove_queue.doorbell_offset_1 = input_ptr->doorbell_offset_1;
+	remove_queue.context_csa_addr = input_ptr->context_csa_addr;
+
+	remove_queue.api_status.api_completion_fence_addr = umsch->ring.fence_drv.gpu_addr;
+	remove_queue.api_status.api_completion_fence_value = ++umsch->ring.fence_drv.sync_seq;
+
+	r = amdgpu_umsch_mm_submit_pkt(umsch, &remove_queue.max_dwords_in_api,
+				       API_FRAME_SIZE_IN_DWORDS);
+	if (r)
+		return r;
+
+	r = amdgpu_umsch_mm_query_fence(umsch);
+	if (r) {
+		dev_err(adev->dev, "UMSCH REMOVE_QUEUE: Failed\n");
+		return r;
+	}
+
+	return 0;
+}
+
+static int umsch_mm_v4_0_set_regs(struct amdgpu_umsch_mm *umsch)
+{
+	struct amdgpu_device *adev = container_of(umsch, struct amdgpu_device, umsch_mm);
+
+	umsch->rb_wptr = SOC15_REG_OFFSET(VCN, 0, regVCN_UMSCH_RB_WPTR);
+	umsch->rb_rptr = SOC15_REG_OFFSET(VCN, 0, regVCN_UMSCH_RB_RPTR);
+
+	return 0;
+}
+
+static const struct umsch_mm_funcs umsch_mm_v4_0_funcs = {
+	.set_hw_resources = umsch_mm_v4_0_set_hw_resources,
+	.add_queue = umsch_mm_v4_0_add_queue,
+	.remove_queue = umsch_mm_v4_0_remove_queue,
+	.set_regs = umsch_mm_v4_0_set_regs,
+	.init_microcode = amdgpu_umsch_mm_init_microcode,
+	.load_microcode = umsch_mm_v4_0_load_microcode,
+	.ring_init = amdgpu_umsch_mm_ring_init,
+	.ring_start = umsch_mm_v4_0_ring_start,
+	.ring_stop = umsch_mm_v4_0_ring_stop,
+};
+
+void umsch_mm_v4_0_set_funcs(struct amdgpu_umsch_mm *umsch)
+{
+	umsch->funcs = &umsch_mm_v4_0_funcs;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/umsch_mm_v4_0.h b/drivers/gpu/drm/amd/amdgpu/umsch_mm_v4_0.h
new file mode 100644
index 000000000000..06bc0fa74996
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/umsch_mm_v4_0.h
@@ -0,0 +1,30 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __UMSCH_MM_V4_0_H__
+#define __UMSCH_MM_V4_0_H__
+
+void umsch_mm_v4_0_set_funcs(struct amdgpu_umsch_mm *umsch);
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v3_1.c b/drivers/gpu/drm/amd/amdgpu/uvd_v3_1.c
index 0fef925b6602..2e79a3afc774 100644
--- a/drivers/gpu/drm/amd/amdgpu/uvd_v3_1.c
+++ b/drivers/gpu/drm/amd/amdgpu/uvd_v3_1.c
@@ -98,7 +98,7 @@ static void uvd_v3_1_ring_emit_ib(struct amdgpu_ring *ring,
 }
 
 /**
- * uvd_v3_1_ring_emit_fence - emit an fence & trap command
+ * uvd_v3_1_ring_emit_fence - emit a fence & trap command
  *
  * @ring: amdgpu_ring pointer
  * @addr: address
@@ -242,7 +242,7 @@ static void uvd_v3_1_mc_resume(struct amdgpu_device *adev)
 	uint64_t addr;
 	uint32_t size;
 
-	/* programm the VCPU memory controller bits 0-27 */
+	/* program the VCPU memory controller bits 0-27 */
 	addr = (adev->uvd.inst->gpu_addr + AMDGPU_UVD_FIRMWARE_OFFSET) >> 3;
 	size = AMDGPU_UVD_FIRMWARE_SIZE(adev) >> 3;
 	WREG32(mmUVD_VCPU_CACHE_OFFSET0, addr);
@@ -416,7 +416,7 @@ static int uvd_v3_1_start(struct amdgpu_device *adev)
 	/* Set the write pointer delay */
 	WREG32(mmUVD_RBC_RB_WPTR_CNTL, 0);
 
-	/* programm the 4GB memory segment for rptr and ring buffer */
+	/* Program the 4GB memory segment for rptr and ring buffer */
 	WREG32(mmUVD_LMI_EXT40_ADDR, upper_32_bits(ring->gpu_addr) |
 		   (0x7 << 16) | (0x1 << 31));
 
@@ -531,9 +531,9 @@ static void uvd_v3_1_set_irq_funcs(struct amdgpu_device *adev)
 }
 
 
-static int uvd_v3_1_early_init(void *handle)
+static int uvd_v3_1_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	adev->uvd.num_uvd_inst = 1;
 
 	uvd_v3_1_set_ring_funcs(adev);
@@ -542,10 +542,10 @@ static int uvd_v3_1_early_init(void *handle)
 	return 0;
 }
 
-static int uvd_v3_1_sw_init(void *handle)
+static int uvd_v3_1_sw_init(struct amdgpu_ip_block *ip_block)
 {
 	struct amdgpu_ring *ring;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int r;
 	void *ptr;
 	uint32_t ucode_len;
@@ -577,15 +577,13 @@ static int uvd_v3_1_sw_init(void *handle)
 	ptr += ucode_len;
 	memcpy(&adev->uvd.keyselect, ptr, 4);
 
-	r = amdgpu_uvd_entity_init(adev);
-
 	return r;
 }
 
-static int uvd_v3_1_sw_fini(void *handle)
+static int uvd_v3_1_sw_fini(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	r = amdgpu_uvd_suspend(adev);
 	if (r)
@@ -623,18 +621,42 @@ static void uvd_v3_1_enable_mgcg(struct amdgpu_device *adev,
 /**
  * uvd_v3_1_hw_init - start and test UVD block
  *
- * @handle: handle used to pass amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Initialize the hardware, boot up the VCPU and do some testing.
  *
- * Initialize the hardware, boot up the VCPU and do some testing
+ * On SI, the UVD is meant to be used in a specific power state,
+ * or alternatively the driver can manually enable its clock.
+ * In amdgpu we use the dedicated UVD power state when DPM is enabled.
+ * Calling amdgpu_dpm_enable_uvd makes DPM select the UVD power state
+ * for the SMU and afterwards enables the UVD clock.
+ * This is automatically done by amdgpu_uvd_ring_begin_use when work
+ * is submitted to the UVD ring. Here, we have to call it manually
+ * in order to power up UVD before firmware validation.
+ *
+ * Note that we must not disable the UVD clock here, as that would
+ * cause the ring test to fail. However, UVD is powered off
+ * automatically after the ring test: amdgpu_uvd_ring_end_use calls
+ * the UVD idle work handler which will disable the UVD clock when
+ * all fences are signalled.
  */
-static int uvd_v3_1_hw_init(void *handle)
+static int uvd_v3_1_hw_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	struct amdgpu_ring *ring = &adev->uvd.inst->ring;
 	uint32_t tmp;
 	int r;
 
 	uvd_v3_1_mc_resume(adev);
+	uvd_v3_1_enable_mgcg(adev, true);
+
+	/* Make sure UVD is powered during FW validation.
+	 * It's going to be automatically powered off after the ring test.
+	 */
+	if (adev->pm.dpm_enabled)
+		amdgpu_dpm_enable_uvd(adev, true);
+	else
+		amdgpu_asic_set_uvd_clocks(adev, 53300, 40000);
 
 	r = uvd_v3_1_fw_validate(adev);
 	if (r) {
@@ -642,9 +664,6 @@ static int uvd_v3_1_hw_init(void *handle)
 		return r;
 	}
 
-	uvd_v3_1_enable_mgcg(adev, true);
-	amdgpu_asic_set_uvd_clocks(adev, 53300, 40000);
-
 	uvd_v3_1_start(adev);
 
 	r = amdgpu_ring_test_helper(ring);
@@ -690,13 +709,13 @@ done:
 /**
  * uvd_v3_1_hw_fini - stop the hardware block
  *
- * @handle: handle used to pass amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Stop the UVD block, mark ring as not ready any more
  */
-static int uvd_v3_1_hw_fini(void *handle)
+static int uvd_v3_1_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	cancel_delayed_work_sync(&adev->uvd.idle_work);
 
@@ -706,10 +725,17 @@ static int uvd_v3_1_hw_fini(void *handle)
 	return 0;
 }
 
-static int uvd_v3_1_suspend(void *handle)
+static int uvd_v3_1_prepare_suspend(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	return amdgpu_uvd_prepare_suspend(adev);
+}
+
+static int uvd_v3_1_suspend(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	/*
 	 * Proper cleanups before halting the HW engine:
@@ -735,36 +761,35 @@ static int uvd_v3_1_suspend(void *handle)
 						       AMD_CG_STATE_GATE);
 	}
 
-	r = uvd_v3_1_hw_fini(adev);
+	r = uvd_v3_1_hw_fini(ip_block);
 	if (r)
 		return r;
 
 	return amdgpu_uvd_suspend(adev);
 }
 
-static int uvd_v3_1_resume(void *handle)
+static int uvd_v3_1_resume(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
-	r = amdgpu_uvd_resume(adev);
+	r = amdgpu_uvd_resume(ip_block->adev);
 	if (r)
 		return r;
 
-	return uvd_v3_1_hw_init(adev);
+	return uvd_v3_1_hw_init(ip_block);
 }
 
-static bool uvd_v3_1_is_idle(void *handle)
+static bool uvd_v3_1_is_idle(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	return !(RREG32(mmSRBM_STATUS) & SRBM_STATUS__UVD_BUSY_MASK);
 }
 
-static int uvd_v3_1_wait_for_idle(void *handle)
+static int uvd_v3_1_wait_for_idle(struct amdgpu_ip_block *ip_block)
 {
 	unsigned i;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	for (i = 0; i < adev->usec_timeout; i++) {
 		if (!(RREG32(mmSRBM_STATUS) & SRBM_STATUS__UVD_BUSY_MASK))
@@ -773,9 +798,9 @@ static int uvd_v3_1_wait_for_idle(void *handle)
 	return -ETIMEDOUT;
 }
 
-static int uvd_v3_1_soft_reset(void *handle)
+static int uvd_v3_1_soft_reset(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	uvd_v3_1_stop(adev);
 
@@ -786,13 +811,13 @@ static int uvd_v3_1_soft_reset(void *handle)
 	return uvd_v3_1_start(adev);
 }
 
-static int uvd_v3_1_set_clockgating_state(void *handle,
+static int uvd_v3_1_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_clockgating_state state)
 {
 	return 0;
 }
 
-static int uvd_v3_1_set_powergating_state(void *handle,
+static int uvd_v3_1_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_powergating_state state)
 {
 	return 0;
@@ -801,11 +826,11 @@ static int uvd_v3_1_set_powergating_state(void *handle,
 static const struct amd_ip_funcs uvd_v3_1_ip_funcs = {
 	.name = "uvd_v3_1",
 	.early_init = uvd_v3_1_early_init,
-	.late_init = NULL,
 	.sw_init = uvd_v3_1_sw_init,
 	.sw_fini = uvd_v3_1_sw_fini,
 	.hw_init = uvd_v3_1_hw_init,
 	.hw_fini = uvd_v3_1_hw_fini,
+	.prepare_suspend = uvd_v3_1_prepare_suspend,
 	.suspend = uvd_v3_1_suspend,
 	.resume = uvd_v3_1_resume,
 	.is_idle = uvd_v3_1_is_idle,
@@ -815,8 +840,7 @@ static const struct amd_ip_funcs uvd_v3_1_ip_funcs = {
 	.set_powergating_state = uvd_v3_1_set_powergating_state,
 };
 
-const struct amdgpu_ip_block_version uvd_v3_1_ip_block =
-{
+const struct amdgpu_ip_block_version uvd_v3_1_ip_block = {
 	.type = AMD_IP_BLOCK_TYPE_UVD,
 	.major = 3,
 	.minor = 1,
diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c b/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c
index c108b8381795..4b96fd583772 100644
--- a/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c
+++ b/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c
@@ -44,7 +44,7 @@ static void uvd_v4_2_set_ring_funcs(struct amdgpu_device *adev);
 static void uvd_v4_2_set_irq_funcs(struct amdgpu_device *adev);
 static int uvd_v4_2_start(struct amdgpu_device *adev);
 static void uvd_v4_2_stop(struct amdgpu_device *adev);
-static int uvd_v4_2_set_clockgating_state(void *handle,
+static int uvd_v4_2_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 				enum amd_clockgating_state state);
 static void uvd_v4_2_set_dcm(struct amdgpu_device *adev,
 			     bool sw_mode);
@@ -90,9 +90,9 @@ static void uvd_v4_2_ring_set_wptr(struct amdgpu_ring *ring)
 	WREG32(mmUVD_RBC_RB_WPTR, lower_32_bits(ring->wptr));
 }
 
-static int uvd_v4_2_early_init(void *handle)
+static int uvd_v4_2_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	adev->uvd.num_uvd_inst = 1;
 
 	uvd_v4_2_set_ring_funcs(adev);
@@ -101,10 +101,10 @@ static int uvd_v4_2_early_init(void *handle)
 	return 0;
 }
 
-static int uvd_v4_2_sw_init(void *handle)
+static int uvd_v4_2_sw_init(struct amdgpu_ip_block *ip_block)
 {
 	struct amdgpu_ring *ring;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int r;
 
 	/* UVD TRAP */
@@ -127,15 +127,13 @@ static int uvd_v4_2_sw_init(void *handle)
 	if (r)
 		return r;
 
-	r = amdgpu_uvd_entity_init(adev);
-
 	return r;
 }
 
-static int uvd_v4_2_sw_fini(void *handle)
+static int uvd_v4_2_sw_fini(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	r = amdgpu_uvd_suspend(adev);
 	if (r)
@@ -149,13 +147,13 @@ static void uvd_v4_2_enable_mgcg(struct amdgpu_device *adev,
 /**
  * uvd_v4_2_hw_init - start and test UVD block
  *
- * @handle: handle used to pass amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Initialize the hardware, boot up the VCPU and do some testing
  */
-static int uvd_v4_2_hw_init(void *handle)
+static int uvd_v4_2_hw_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	struct amdgpu_ring *ring = &adev->uvd.inst->ring;
 	uint32_t tmp;
 	int r;
@@ -204,13 +202,13 @@ done:
 /**
  * uvd_v4_2_hw_fini - stop the hardware block
  *
- * @handle: handle used to pass amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Stop the UVD block, mark ring as not ready any more
  */
-static int uvd_v4_2_hw_fini(void *handle)
+static int uvd_v4_2_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	cancel_delayed_work_sync(&adev->uvd.idle_work);
 
@@ -220,10 +218,17 @@ static int uvd_v4_2_hw_fini(void *handle)
 	return 0;
 }
 
-static int uvd_v4_2_suspend(void *handle)
+static int uvd_v4_2_prepare_suspend(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	return amdgpu_uvd_prepare_suspend(adev);
+}
+
+static int uvd_v4_2_suspend(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	/*
 	 * Proper cleanups before halting the HW engine:
@@ -249,23 +254,22 @@ static int uvd_v4_2_suspend(void *handle)
 						       AMD_CG_STATE_GATE);
 	}
 
-	r = uvd_v4_2_hw_fini(adev);
+	r = uvd_v4_2_hw_fini(ip_block);
 	if (r)
 		return r;
 
 	return amdgpu_uvd_suspend(adev);
 }
 
-static int uvd_v4_2_resume(void *handle)
+static int uvd_v4_2_resume(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
-	r = amdgpu_uvd_resume(adev);
+	r = amdgpu_uvd_resume(ip_block->adev);
 	if (r)
 		return r;
 
-	return uvd_v4_2_hw_init(adev);
+	return uvd_v4_2_hw_init(ip_block);
 }
 
 /**
@@ -298,7 +302,7 @@ static int uvd_v4_2_start(struct amdgpu_device *adev)
 	/* enable VCPU clock */
 	WREG32(mmUVD_VCPU_CNTL,  1 << 9);
 
-	/* disable interupt */
+	/* disable interrupt */
 	WREG32_P(mmUVD_MASTINT_EN, 0, ~(1 << 1));
 
 #ifdef __BIG_ENDIAN
@@ -308,6 +312,7 @@ static int uvd_v4_2_start(struct amdgpu_device *adev)
 #endif
 	WREG32(mmUVD_LMI_SWAP_CNTL, lmi_swap_cntl);
 	WREG32(mmUVD_MP_SWAP_CNTL, mp_swap_cntl);
+
 	/* initialize UVD memory controller */
 	WREG32(mmUVD_LMI_CTRL, 0x203108);
 
@@ -654,17 +659,17 @@ static void uvd_v4_2_set_dcm(struct amdgpu_device *adev,
 	WREG32_UVD_CTX(ixUVD_CGC_CTRL2, tmp2);
 }
 
-static bool uvd_v4_2_is_idle(void *handle)
+static bool uvd_v4_2_is_idle(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	return !(RREG32(mmSRBM_STATUS) & SRBM_STATUS__UVD_BUSY_MASK);
 }
 
-static int uvd_v4_2_wait_for_idle(void *handle)
+static int uvd_v4_2_wait_for_idle(struct amdgpu_ip_block *ip_block)
 {
 	unsigned i;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	for (i = 0; i < adev->usec_timeout; i++) {
 		if (!(RREG32(mmSRBM_STATUS) & SRBM_STATUS__UVD_BUSY_MASK))
@@ -673,9 +678,9 @@ static int uvd_v4_2_wait_for_idle(void *handle)
 	return -ETIMEDOUT;
 }
 
-static int uvd_v4_2_soft_reset(void *handle)
+static int uvd_v4_2_soft_reset(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	uvd_v4_2_stop(adev);
 
@@ -704,13 +709,13 @@ static int uvd_v4_2_process_interrupt(struct amdgpu_device *adev,
 	return 0;
 }
 
-static int uvd_v4_2_set_clockgating_state(void *handle,
+static int uvd_v4_2_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_clockgating_state state)
 {
 	return 0;
 }
 
-static int uvd_v4_2_set_powergating_state(void *handle,
+static int uvd_v4_2_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_powergating_state state)
 {
 	/* This doesn't actually powergate the UVD block.
@@ -720,7 +725,7 @@ static int uvd_v4_2_set_powergating_state(void *handle,
 	 * revisit this when there is a cleaner line between
 	 * the smc and the hw blocks
 	 */
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (state == AMD_PG_STATE_GATE) {
 		uvd_v4_2_stop(adev);
@@ -751,11 +756,11 @@ static int uvd_v4_2_set_powergating_state(void *handle,
 static const struct amd_ip_funcs uvd_v4_2_ip_funcs = {
 	.name = "uvd_v4_2",
 	.early_init = uvd_v4_2_early_init,
-	.late_init = NULL,
 	.sw_init = uvd_v4_2_sw_init,
 	.sw_fini = uvd_v4_2_sw_fini,
 	.hw_init = uvd_v4_2_hw_init,
 	.hw_fini = uvd_v4_2_hw_fini,
+	.prepare_suspend = uvd_v4_2_prepare_suspend,
 	.suspend = uvd_v4_2_suspend,
 	.resume = uvd_v4_2_resume,
 	.is_idle = uvd_v4_2_is_idle,
diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c
index 563493d1f830..71409ad8b7ed 100644
--- a/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c
@@ -42,7 +42,7 @@ static void uvd_v5_0_set_ring_funcs(struct amdgpu_device *adev);
 static void uvd_v5_0_set_irq_funcs(struct amdgpu_device *adev);
 static int uvd_v5_0_start(struct amdgpu_device *adev);
 static void uvd_v5_0_stop(struct amdgpu_device *adev);
-static int uvd_v5_0_set_clockgating_state(void *handle,
+static int uvd_v5_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_clockgating_state state);
 static void uvd_v5_0_enable_mgcg(struct amdgpu_device *adev,
 				 bool enable);
@@ -88,9 +88,9 @@ static void uvd_v5_0_ring_set_wptr(struct amdgpu_ring *ring)
 	WREG32(mmUVD_RBC_RB_WPTR, lower_32_bits(ring->wptr));
 }
 
-static int uvd_v5_0_early_init(void *handle)
+static int uvd_v5_0_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	adev->uvd.num_uvd_inst = 1;
 
 	uvd_v5_0_set_ring_funcs(adev);
@@ -99,10 +99,10 @@ static int uvd_v5_0_early_init(void *handle)
 	return 0;
 }
 
-static int uvd_v5_0_sw_init(void *handle)
+static int uvd_v5_0_sw_init(struct amdgpu_ip_block *ip_block)
 {
 	struct amdgpu_ring *ring;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int r;
 
 	/* UVD TRAP */
@@ -125,15 +125,13 @@ static int uvd_v5_0_sw_init(void *handle)
 	if (r)
 		return r;
 
-	r = amdgpu_uvd_entity_init(adev);
-
 	return r;
 }
 
-static int uvd_v5_0_sw_fini(void *handle)
+static int uvd_v5_0_sw_fini(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	r = amdgpu_uvd_suspend(adev);
 	if (r)
@@ -145,19 +143,19 @@ static int uvd_v5_0_sw_fini(void *handle)
 /**
  * uvd_v5_0_hw_init - start and test UVD block
  *
- * @handle: handle used to pass amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Initialize the hardware, boot up the VCPU and do some testing
  */
-static int uvd_v5_0_hw_init(void *handle)
+static int uvd_v5_0_hw_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	struct amdgpu_ring *ring = &adev->uvd.inst->ring;
 	uint32_t tmp;
 	int r;
 
 	amdgpu_asic_set_uvd_clocks(adev, 10000, 10000);
-	uvd_v5_0_set_clockgating_state(adev, AMD_CG_STATE_UNGATE);
+	uvd_v5_0_set_clockgating_state(ip_block, AMD_CG_STATE_UNGATE);
 	uvd_v5_0_enable_mgcg(adev, true);
 
 	r = amdgpu_ring_test_helper(ring);
@@ -202,13 +200,13 @@ done:
 /**
  * uvd_v5_0_hw_fini - stop the hardware block
  *
- * @handle: handle used to pass amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Stop the UVD block, mark ring as not ready any more
  */
-static int uvd_v5_0_hw_fini(void *handle)
+static int uvd_v5_0_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	cancel_delayed_work_sync(&adev->uvd.idle_work);
 
@@ -218,10 +216,17 @@ static int uvd_v5_0_hw_fini(void *handle)
 	return 0;
 }
 
-static int uvd_v5_0_suspend(void *handle)
+static int uvd_v5_0_prepare_suspend(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	return amdgpu_uvd_prepare_suspend(adev);
+}
+
+static int uvd_v5_0_suspend(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	/*
 	 * Proper cleanups before halting the HW engine:
@@ -247,23 +252,22 @@ static int uvd_v5_0_suspend(void *handle)
 						       AMD_CG_STATE_GATE);
 	}
 
-	r = uvd_v5_0_hw_fini(adev);
+	r = uvd_v5_0_hw_fini(ip_block);
 	if (r)
 		return r;
 
 	return amdgpu_uvd_suspend(adev);
 }
 
-static int uvd_v5_0_resume(void *handle)
+static int uvd_v5_0_resume(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
-	r = amdgpu_uvd_resume(adev);
+	r = amdgpu_uvd_resume(ip_block->adev);
 	if (r)
 		return r;
 
-	return uvd_v5_0_hw_init(adev);
+	return uvd_v5_0_hw_init(ip_block);
 }
 
 /**
@@ -576,17 +580,17 @@ static void uvd_v5_0_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count)
 	}
 }
 
-static bool uvd_v5_0_is_idle(void *handle)
+static bool uvd_v5_0_is_idle(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	return !(RREG32(mmSRBM_STATUS) & SRBM_STATUS__UVD_BUSY_MASK);
 }
 
-static int uvd_v5_0_wait_for_idle(void *handle)
+static int uvd_v5_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
 {
 	unsigned i;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	for (i = 0; i < adev->usec_timeout; i++) {
 		if (!(RREG32(mmSRBM_STATUS) & SRBM_STATUS__UVD_BUSY_MASK))
@@ -595,9 +599,9 @@ static int uvd_v5_0_wait_for_idle(void *handle)
 	return -ETIMEDOUT;
 }
 
-static int uvd_v5_0_soft_reset(void *handle)
+static int uvd_v5_0_soft_reset(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	uvd_v5_0_stop(adev);
 
@@ -786,15 +790,15 @@ static void uvd_v5_0_enable_mgcg(struct amdgpu_device *adev,
 	}
 }
 
-static int uvd_v5_0_set_clockgating_state(void *handle,
+static int uvd_v5_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_clockgating_state state)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	bool enable = (state == AMD_CG_STATE_GATE);
 
 	if (enable) {
 		/* wait for STATUS to clear */
-		if (uvd_v5_0_wait_for_idle(handle))
+		if (uvd_v5_0_wait_for_idle(ip_block))
 			return -EBUSY;
 		uvd_v5_0_enable_clock_gating(adev, true);
 
@@ -808,7 +812,7 @@ static int uvd_v5_0_set_clockgating_state(void *handle,
 	return 0;
 }
 
-static int uvd_v5_0_set_powergating_state(void *handle,
+static int uvd_v5_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_powergating_state state)
 {
 	/* This doesn't actually powergate the UVD block.
@@ -818,7 +822,7 @@ static int uvd_v5_0_set_powergating_state(void *handle,
 	 * revisit this when there is a cleaner line between
 	 * the smc and the hw blocks
 	 */
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int ret = 0;
 
 	if (state == AMD_PG_STATE_GATE) {
@@ -833,9 +837,9 @@ out:
 	return ret;
 }
 
-static void uvd_v5_0_get_clockgating_state(void *handle, u32 *flags)
+static void uvd_v5_0_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int data;
 
 	mutex_lock(&adev->pm.mutex);
@@ -858,11 +862,11 @@ out:
 static const struct amd_ip_funcs uvd_v5_0_ip_funcs = {
 	.name = "uvd_v5_0",
 	.early_init = uvd_v5_0_early_init,
-	.late_init = NULL,
 	.sw_init = uvd_v5_0_sw_init,
 	.sw_fini = uvd_v5_0_sw_fini,
 	.hw_init = uvd_v5_0_hw_init,
 	.hw_fini = uvd_v5_0_hw_fini,
+	.prepare_suspend = uvd_v5_0_prepare_suspend,
 	.suspend = uvd_v5_0_suspend,
 	.resume = uvd_v5_0_resume,
 	.is_idle = uvd_v5_0_is_idle,
diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c
index 2d558c2f417d..ceb94bbb03a4 100644
--- a/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c
@@ -48,7 +48,7 @@ static void uvd_v6_0_set_irq_funcs(struct amdgpu_device *adev);
 static int uvd_v6_0_start(struct amdgpu_device *adev);
 static void uvd_v6_0_stop(struct amdgpu_device *adev);
 static void uvd_v6_0_set_sw_clock_gating(struct amdgpu_device *adev);
-static int uvd_v6_0_set_clockgating_state(void *handle,
+static int uvd_v6_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_clockgating_state state);
 static void uvd_v6_0_enable_mgcg(struct amdgpu_device *adev,
 				 bool enable);
@@ -216,8 +216,9 @@ static int uvd_v6_0_enc_get_create_msg(struct amdgpu_ring *ring, uint32_t handle
 	uint64_t addr;
 	int i, r;
 
-	r = amdgpu_job_alloc_with_ib(ring->adev, ib_size_dw * 4,
-					AMDGPU_IB_POOL_DIRECT, &job);
+	r = amdgpu_job_alloc_with_ib(ring->adev, NULL, NULL, ib_size_dw * 4,
+				     AMDGPU_IB_POOL_DIRECT, &job,
+				     AMDGPU_KERNEL_JOB_ID_VCN_RING_TEST);
 	if (r)
 		return r;
 
@@ -280,8 +281,9 @@ static int uvd_v6_0_enc_get_destroy_msg(struct amdgpu_ring *ring,
 	uint64_t addr;
 	int i, r;
 
-	r = amdgpu_job_alloc_with_ib(ring->adev, ib_size_dw * 4,
-					AMDGPU_IB_POOL_DIRECT, &job);
+	r = amdgpu_job_alloc_with_ib(ring->adev, NULL, NULL, ib_size_dw * 4,
+				     AMDGPU_IB_POOL_DIRECT, &job,
+				     AMDGPU_KERNEL_JOB_ID_VCN_RING_TEST);
 	if (r)
 		return r;
 
@@ -354,9 +356,9 @@ error:
 	return r;
 }
 
-static int uvd_v6_0_early_init(void *handle)
+static int uvd_v6_0_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	adev->uvd.num_uvd_inst = 1;
 
 	if (!(adev->flags & AMD_IS_APU) &&
@@ -375,11 +377,11 @@ static int uvd_v6_0_early_init(void *handle)
 	return 0;
 }
 
-static int uvd_v6_0_sw_init(void *handle)
+static int uvd_v6_0_sw_init(struct amdgpu_ip_block *ip_block)
 {
 	struct amdgpu_ring *ring;
 	int i, r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	/* UVD TRAP */
 	r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_UVD_SYSTEM_MESSAGE, &adev->uvd.inst->irq);
@@ -432,15 +434,13 @@ static int uvd_v6_0_sw_init(void *handle)
 		}
 	}
 
-	r = amdgpu_uvd_entity_init(adev);
-
 	return r;
 }
 
-static int uvd_v6_0_sw_fini(void *handle)
+static int uvd_v6_0_sw_fini(struct amdgpu_ip_block *ip_block)
 {
 	int i, r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	r = amdgpu_uvd_suspend(adev);
 	if (r)
@@ -457,19 +457,19 @@ static int uvd_v6_0_sw_fini(void *handle)
 /**
  * uvd_v6_0_hw_init - start and test UVD block
  *
- * @handle: handle used to pass amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Initialize the hardware, boot up the VCPU and do some testing
  */
-static int uvd_v6_0_hw_init(void *handle)
+static int uvd_v6_0_hw_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	struct amdgpu_ring *ring = &adev->uvd.inst->ring;
 	uint32_t tmp;
 	int i, r;
 
 	amdgpu_asic_set_uvd_clocks(adev, 10000, 10000);
-	uvd_v6_0_set_clockgating_state(adev, AMD_CG_STATE_UNGATE);
+	uvd_v6_0_set_clockgating_state(ip_block, AMD_CG_STATE_UNGATE);
 	uvd_v6_0_enable_mgcg(adev, true);
 
 	r = amdgpu_ring_test_helper(ring);
@@ -526,13 +526,13 @@ done:
 /**
  * uvd_v6_0_hw_fini - stop the hardware block
  *
- * @handle: handle used to pass amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Stop the UVD block, mark ring as not ready any more
  */
-static int uvd_v6_0_hw_fini(void *handle)
+static int uvd_v6_0_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	cancel_delayed_work_sync(&adev->uvd.idle_work);
 
@@ -542,10 +542,17 @@ static int uvd_v6_0_hw_fini(void *handle)
 	return 0;
 }
 
-static int uvd_v6_0_suspend(void *handle)
+static int uvd_v6_0_prepare_suspend(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	return amdgpu_uvd_prepare_suspend(adev);
+}
+
+static int uvd_v6_0_suspend(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	/*
 	 * Proper cleanups before halting the HW engine:
@@ -571,23 +578,22 @@ static int uvd_v6_0_suspend(void *handle)
 						       AMD_CG_STATE_GATE);
 	}
 
-	r = uvd_v6_0_hw_fini(adev);
+	r = uvd_v6_0_hw_fini(ip_block);
 	if (r)
 		return r;
 
 	return amdgpu_uvd_suspend(adev);
 }
 
-static int uvd_v6_0_resume(void *handle)
+static int uvd_v6_0_resume(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
-	r = amdgpu_uvd_resume(adev);
+	r = amdgpu_uvd_resume(ip_block->adev);
 	if (r)
 		return r;
 
-	return uvd_v6_0_hw_init(adev);
+	return uvd_v6_0_hw_init(ip_block);
 }
 
 /**
@@ -1139,29 +1145,29 @@ static void uvd_v6_0_enc_ring_emit_vm_flush(struct amdgpu_ring *ring,
 	amdgpu_ring_write(ring, vmid);
 }
 
-static bool uvd_v6_0_is_idle(void *handle)
+static bool uvd_v6_0_is_idle(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	return !(RREG32(mmSRBM_STATUS) & SRBM_STATUS__UVD_BUSY_MASK);
 }
 
-static int uvd_v6_0_wait_for_idle(void *handle)
+static int uvd_v6_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
 {
 	unsigned i;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	for (i = 0; i < adev->usec_timeout; i++) {
-		if (uvd_v6_0_is_idle(handle))
+		if (uvd_v6_0_is_idle(ip_block))
 			return 0;
 	}
 	return -ETIMEDOUT;
 }
 
 #define AMDGPU_UVD_STATUS_BUSY_MASK    0xfd
-static bool uvd_v6_0_check_soft_reset(void *handle)
+static bool uvd_v6_0_check_soft_reset(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	u32 srbm_soft_reset = 0;
 	u32 tmp = RREG32(mmSRBM_STATUS);
 
@@ -1179,9 +1185,9 @@ static bool uvd_v6_0_check_soft_reset(void *handle)
 	}
 }
 
-static int uvd_v6_0_pre_soft_reset(void *handle)
+static int uvd_v6_0_pre_soft_reset(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (!adev->uvd.inst->srbm_soft_reset)
 		return 0;
@@ -1190,9 +1196,9 @@ static int uvd_v6_0_pre_soft_reset(void *handle)
 	return 0;
 }
 
-static int uvd_v6_0_soft_reset(void *handle)
+static int uvd_v6_0_soft_reset(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	u32 srbm_soft_reset;
 
 	if (!adev->uvd.inst->srbm_soft_reset)
@@ -1221,9 +1227,9 @@ static int uvd_v6_0_soft_reset(void *handle)
 	return 0;
 }
 
-static int uvd_v6_0_post_soft_reset(void *handle)
+static int uvd_v6_0_post_soft_reset(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (!adev->uvd.inst->srbm_soft_reset)
 		return 0;
@@ -1446,15 +1452,15 @@ static void uvd_v6_0_enable_mgcg(struct amdgpu_device *adev,
 	}
 }
 
-static int uvd_v6_0_set_clockgating_state(void *handle,
+static int uvd_v6_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_clockgating_state state)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	bool enable = (state == AMD_CG_STATE_GATE);
 
 	if (enable) {
 		/* wait for STATUS to clear */
-		if (uvd_v6_0_wait_for_idle(handle))
+		if (uvd_v6_0_wait_for_idle(ip_block))
 			return -EBUSY;
 		uvd_v6_0_enable_clock_gating(adev, true);
 		/* enable HW gates because UVD is idle */
@@ -1467,7 +1473,7 @@ static int uvd_v6_0_set_clockgating_state(void *handle,
 	return 0;
 }
 
-static int uvd_v6_0_set_powergating_state(void *handle,
+static int uvd_v6_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_powergating_state state)
 {
 	/* This doesn't actually powergate the UVD block.
@@ -1477,7 +1483,7 @@ static int uvd_v6_0_set_powergating_state(void *handle,
 	 * revisit this when there is a cleaner line between
 	 * the smc and the hw blocks
 	 */
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int ret = 0;
 
 	WREG32(mmUVD_POWER_STATUS, UVD_POWER_STATUS__UVD_PG_EN_MASK);
@@ -1494,9 +1500,9 @@ out:
 	return ret;
 }
 
-static void uvd_v6_0_get_clockgating_state(void *handle, u32 *flags)
+static void uvd_v6_0_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int data;
 
 	mutex_lock(&adev->pm.mutex);
@@ -1523,11 +1529,11 @@ out:
 static const struct amd_ip_funcs uvd_v6_0_ip_funcs = {
 	.name = "uvd_v6_0",
 	.early_init = uvd_v6_0_early_init,
-	.late_init = NULL,
 	.sw_init = uvd_v6_0_sw_init,
 	.sw_fini = uvd_v6_0_sw_fini,
 	.hw_init = uvd_v6_0_hw_init,
 	.hw_fini = uvd_v6_0_hw_fini,
+	.prepare_suspend = uvd_v6_0_prepare_suspend,
 	.suspend = uvd_v6_0_suspend,
 	.resume = uvd_v6_0_resume,
 	.is_idle = uvd_v6_0_is_idle,
diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c
index b483f03b4591..1f8866f3f63c 100644
--- a/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c
@@ -25,6 +25,7 @@
 
 #include "amdgpu.h"
 #include "amdgpu_uvd.h"
+#include "amdgpu_cs.h"
 #include "soc15.h"
 #include "soc15d.h"
 #include "soc15_common.h"
@@ -117,7 +118,7 @@ static uint64_t uvd_v7_0_enc_ring_get_wptr(struct amdgpu_ring *ring)
 	struct amdgpu_device *adev = ring->adev;
 
 	if (ring->use_doorbell)
-		return adev->wb.wb[ring->wptr_offs];
+		return *ring->wptr_cpu_addr;
 
 	if (ring == &adev->uvd.inst[ring->me].ring_enc[0])
 		return RREG32_SOC15(UVD, ring->me, mmUVD_RB_WPTR);
@@ -152,7 +153,7 @@ static void uvd_v7_0_enc_ring_set_wptr(struct amdgpu_ring *ring)
 
 	if (ring->use_doorbell) {
 		/* XXX check if swapping is necessary on BE */
-		adev->wb.wb[ring->wptr_offs] = lower_32_bits(ring->wptr);
+		*ring->wptr_cpu_addr = lower_32_bits(ring->wptr);
 		WDOORBELL32(ring->doorbell_index, lower_32_bits(ring->wptr));
 		return;
 	}
@@ -212,7 +213,7 @@ static int uvd_v7_0_enc_ring_test_ring(struct amdgpu_ring *ring)
  *
  * Open up a stream for HW test
  */
-static int uvd_v7_0_enc_get_create_msg(struct amdgpu_ring *ring, uint32_t handle,
+static int uvd_v7_0_enc_get_create_msg(struct amdgpu_ring *ring, u32 handle,
 				       struct amdgpu_bo *bo,
 				       struct dma_fence **fence)
 {
@@ -223,8 +224,9 @@ static int uvd_v7_0_enc_get_create_msg(struct amdgpu_ring *ring, uint32_t handle
 	uint64_t addr;
 	int i, r;
 
-	r = amdgpu_job_alloc_with_ib(ring->adev, ib_size_dw * 4,
-					AMDGPU_IB_POOL_DIRECT, &job);
+	r = amdgpu_job_alloc_with_ib(ring->adev, NULL, NULL, ib_size_dw * 4,
+				     AMDGPU_IB_POOL_DIRECT, &job,
+				     AMDGPU_KERNEL_JOB_ID_VCN_RING_TEST);
 	if (r)
 		return r;
 
@@ -275,7 +277,7 @@ err:
  *
  * Close up a stream for HW test or if userspace failed to do so
  */
-static int uvd_v7_0_enc_get_destroy_msg(struct amdgpu_ring *ring, uint32_t handle,
+static int uvd_v7_0_enc_get_destroy_msg(struct amdgpu_ring *ring, u32 handle,
 					struct amdgpu_bo *bo,
 					struct dma_fence **fence)
 {
@@ -286,8 +288,9 @@ static int uvd_v7_0_enc_get_destroy_msg(struct amdgpu_ring *ring, uint32_t handl
 	uint64_t addr;
 	int i, r;
 
-	r = amdgpu_job_alloc_with_ib(ring->adev, ib_size_dw * 4,
-					AMDGPU_IB_POOL_DIRECT, &job);
+	r = amdgpu_job_alloc_with_ib(ring->adev, NULL, NULL, ib_size_dw * 4,
+				     AMDGPU_IB_POOL_DIRECT, &job,
+				     AMDGPU_KERNEL_JOB_ID_VCN_RING_TEST);
 	if (r)
 		return r;
 
@@ -360,9 +363,9 @@ error:
 	return r;
 }
 
-static int uvd_v7_0_early_init(void *handle)
+static int uvd_v7_0_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (adev->asic_type == CHIP_VEGA20) {
 		u32 harvest;
@@ -394,12 +397,12 @@ static int uvd_v7_0_early_init(void *handle)
 	return 0;
 }
 
-static int uvd_v7_0_sw_init(void *handle)
+static int uvd_v7_0_sw_init(struct amdgpu_ip_block *ip_block)
 {
 	struct amdgpu_ring *ring;
 
 	int i, j, r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	for (j = 0; j < adev->uvd.num_uvd_inst; j++) {
 		if (adev->uvd.harvest_config & (1 << j))
@@ -443,6 +446,7 @@ static int uvd_v7_0_sw_init(void *handle)
 			continue;
 		if (!amdgpu_sriov_vf(adev)) {
 			ring = &adev->uvd.inst[j].ring;
+			ring->vm_hub = AMDGPU_MMHUB0(0);
 			sprintf(ring->name, "uvd_%d", ring->me);
 			r = amdgpu_ring_init(adev, ring, 512,
 					     &adev->uvd.inst[j].irq, 0,
@@ -453,6 +457,7 @@ static int uvd_v7_0_sw_init(void *handle)
 
 		for (i = 0; i < adev->uvd.num_enc_rings; ++i) {
 			ring = &adev->uvd.inst[j].ring_enc[i];
+			ring->vm_hub = AMDGPU_MMHUB0(0);
 			sprintf(ring->name, "uvd_enc_%d.%d", ring->me, i);
 			if (amdgpu_sriov_vf(adev)) {
 				ring->use_doorbell = true;
@@ -477,10 +482,6 @@ static int uvd_v7_0_sw_init(void *handle)
 	if (r)
 		return r;
 
-	r = amdgpu_uvd_entity_init(adev);
-	if (r)
-		return r;
-
 	r = amdgpu_virt_alloc_mm_table(adev);
 	if (r)
 		return r;
@@ -488,10 +489,10 @@ static int uvd_v7_0_sw_init(void *handle)
 	return r;
 }
 
-static int uvd_v7_0_sw_fini(void *handle)
+static int uvd_v7_0_sw_fini(struct amdgpu_ip_block *ip_block)
 {
 	int i, j, r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	amdgpu_virt_free_mm_table(adev);
 
@@ -511,13 +512,13 @@ static int uvd_v7_0_sw_fini(void *handle)
 /**
  * uvd_v7_0_hw_init - start and test UVD block
  *
- * @handle: handle used to pass amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Initialize the hardware, boot up the VCPU and do some testing
  */
-static int uvd_v7_0_hw_init(void *handle)
+static int uvd_v7_0_hw_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	struct amdgpu_ring *ring;
 	uint32_t tmp;
 	int i, j, r;
@@ -589,13 +590,13 @@ done:
 /**
  * uvd_v7_0_hw_fini - stop the hardware block
  *
- * @handle: handle used to pass amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Stop the UVD block, mark ring as not ready any more
  */
-static int uvd_v7_0_hw_fini(void *handle)
+static int uvd_v7_0_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	cancel_delayed_work_sync(&adev->uvd.idle_work);
 
@@ -609,10 +610,17 @@ static int uvd_v7_0_hw_fini(void *handle)
 	return 0;
 }
 
-static int uvd_v7_0_suspend(void *handle)
+static int uvd_v7_0_prepare_suspend(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	return amdgpu_uvd_prepare_suspend(adev);
+}
+
+static int uvd_v7_0_suspend(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	/*
 	 * Proper cleanups before halting the HW engine:
@@ -638,23 +646,22 @@ static int uvd_v7_0_suspend(void *handle)
 						       AMD_CG_STATE_GATE);
 	}
 
-	r = uvd_v7_0_hw_fini(adev);
+	r = uvd_v7_0_hw_fini(ip_block);
 	if (r)
 		return r;
 
 	return amdgpu_uvd_suspend(adev);
 }
 
-static int uvd_v7_0_resume(void *handle)
+static int uvd_v7_0_resume(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
-	r = amdgpu_uvd_resume(adev);
+	r = amdgpu_uvd_resume(ip_block->adev);
 	if (r)
 		return r;
 
-	return uvd_v7_0_hw_init(adev);
+	return uvd_v7_0_hw_init(ip_block);
 }
 
 /**
@@ -676,11 +683,11 @@ static void uvd_v7_0_mc_resume(struct amdgpu_device *adev)
 		if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
 			WREG32_SOC15(UVD, i, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW,
 				i == 0 ?
-				adev->firmware.ucode[AMDGPU_UCODE_ID_UVD].tmr_mc_addr_lo:
+				adev->firmware.ucode[AMDGPU_UCODE_ID_UVD].tmr_mc_addr_lo :
 				adev->firmware.ucode[AMDGPU_UCODE_ID_UVD1].tmr_mc_addr_lo);
 			WREG32_SOC15(UVD, i, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH,
 				i == 0 ?
-				adev->firmware.ucode[AMDGPU_UCODE_ID_UVD].tmr_mc_addr_hi:
+				adev->firmware.ucode[AMDGPU_UCODE_ID_UVD].tmr_mc_addr_hi :
 				adev->firmware.ucode[AMDGPU_UCODE_ID_UVD1].tmr_mc_addr_hi);
 			WREG32_SOC15(UVD, i, mmUVD_VCPU_CACHE_OFFSET0, 0);
 			offset = 0;
@@ -753,7 +760,7 @@ static int uvd_v7_0_mmsch_start(struct amdgpu_device *adev,
 		if (adev->uvd.harvest_config & (1 << i))
 			continue;
 		WDOORBELL32(adev->uvd.inst[i].ring_enc[0].doorbell_index, 0);
-		adev->wb.wb[adev->uvd.inst[i].ring_enc[0].wptr_offs] = 0;
+		*adev->uvd.inst[i].ring_enc[0].wptr_cpu_addr = 0;
 		adev->uvd.inst[i].ring_enc[0].wptr = 0;
 		adev->uvd.inst[i].ring_enc[0].wptr_old = 0;
 	}
@@ -1275,14 +1282,15 @@ static int uvd_v7_0_ring_test_ring(struct amdgpu_ring *ring)
  * uvd_v7_0_ring_patch_cs_in_place - Patch the IB for command submission.
  *
  * @p: the CS parser with the IBs
- * @ib_idx: which IB to patch
+ * @job: which job this ib is in
+ * @ib: which IB to patch
  *
  */
 static int uvd_v7_0_ring_patch_cs_in_place(struct amdgpu_cs_parser *p,
-					   uint32_t ib_idx)
+					   struct amdgpu_job *job,
+					   struct amdgpu_ib *ib)
 {
-	struct amdgpu_ring *ring = to_amdgpu_ring(p->entity->rq->sched);
-	struct amdgpu_ib *ib = &p->job->ibs[ib_idx];
+	struct amdgpu_ring *ring = amdgpu_job_ring(job);
 	unsigned i;
 
 	/* No patching necessary for the first instance */
@@ -1290,12 +1298,12 @@ static int uvd_v7_0_ring_patch_cs_in_place(struct amdgpu_cs_parser *p,
 		return 0;
 
 	for (i = 0; i < ib->length_dw; i += 2) {
-		uint32_t reg = amdgpu_get_ib_value(p, ib_idx, i);
+		uint32_t reg = amdgpu_ib_get_value(ib, i);
 
 		reg -= p->adev->reg_offset[UVD_HWIP][0][1];
 		reg += p->adev->reg_offset[UVD_HWIP][1][1];
 
-		amdgpu_set_ib_value(p, ib_idx, i, reg);
+		amdgpu_ib_set_value(ib, i, reg);
 	}
 	return 0;
 }
@@ -1395,7 +1403,7 @@ static void uvd_v7_0_ring_emit_reg_wait(struct amdgpu_ring *ring, uint32_t reg,
 static void uvd_v7_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
 					unsigned vmid, uint64_t pd_addr)
 {
-	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
+	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->vm_hub];
 	uint32_t data0, data1, mask;
 
 	pd_addr = amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pd_addr);
@@ -1438,7 +1446,7 @@ static void uvd_v7_0_enc_ring_emit_reg_wait(struct amdgpu_ring *ring,
 static void uvd_v7_0_enc_ring_emit_vm_flush(struct amdgpu_ring *ring,
 					    unsigned int vmid, uint64_t pd_addr)
 {
-	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
+	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->vm_hub];
 
 	pd_addr = amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pd_addr);
 
@@ -1456,104 +1464,6 @@ static void uvd_v7_0_enc_ring_emit_wreg(struct amdgpu_ring *ring,
 	amdgpu_ring_write(ring, val);
 }
 
-#if 0
-static bool uvd_v7_0_is_idle(void *handle)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	return !(RREG32(mmSRBM_STATUS) & SRBM_STATUS__UVD_BUSY_MASK);
-}
-
-static int uvd_v7_0_wait_for_idle(void *handle)
-{
-	unsigned i;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	for (i = 0; i < adev->usec_timeout; i++) {
-		if (uvd_v7_0_is_idle(handle))
-			return 0;
-	}
-	return -ETIMEDOUT;
-}
-
-#define AMDGPU_UVD_STATUS_BUSY_MASK    0xfd
-static bool uvd_v7_0_check_soft_reset(void *handle)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-	u32 srbm_soft_reset = 0;
-	u32 tmp = RREG32(mmSRBM_STATUS);
-
-	if (REG_GET_FIELD(tmp, SRBM_STATUS, UVD_RQ_PENDING) ||
-	    REG_GET_FIELD(tmp, SRBM_STATUS, UVD_BUSY) ||
-	    (RREG32_SOC15(UVD, ring->me, mmUVD_STATUS) &
-		    AMDGPU_UVD_STATUS_BUSY_MASK))
-		srbm_soft_reset = REG_SET_FIELD(srbm_soft_reset,
-				SRBM_SOFT_RESET, SOFT_RESET_UVD, 1);
-
-	if (srbm_soft_reset) {
-		adev->uvd.inst[ring->me].srbm_soft_reset = srbm_soft_reset;
-		return true;
-	} else {
-		adev->uvd.inst[ring->me].srbm_soft_reset = 0;
-		return false;
-	}
-}
-
-static int uvd_v7_0_pre_soft_reset(void *handle)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	if (!adev->uvd.inst[ring->me].srbm_soft_reset)
-		return 0;
-
-	uvd_v7_0_stop(adev);
-	return 0;
-}
-
-static int uvd_v7_0_soft_reset(void *handle)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-	u32 srbm_soft_reset;
-
-	if (!adev->uvd.inst[ring->me].srbm_soft_reset)
-		return 0;
-	srbm_soft_reset = adev->uvd.inst[ring->me].srbm_soft_reset;
-
-	if (srbm_soft_reset) {
-		u32 tmp;
-
-		tmp = RREG32(mmSRBM_SOFT_RESET);
-		tmp |= srbm_soft_reset;
-		dev_info(adev->dev, "SRBM_SOFT_RESET=0x%08X\n", tmp);
-		WREG32(mmSRBM_SOFT_RESET, tmp);
-		tmp = RREG32(mmSRBM_SOFT_RESET);
-
-		udelay(50);
-
-		tmp &= ~srbm_soft_reset;
-		WREG32(mmSRBM_SOFT_RESET, tmp);
-		tmp = RREG32(mmSRBM_SOFT_RESET);
-
-		/* Wait a little for things to settle down */
-		udelay(50);
-	}
-
-	return 0;
-}
-
-static int uvd_v7_0_post_soft_reset(void *handle)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	if (!adev->uvd.inst[ring->me].srbm_soft_reset)
-		return 0;
-
-	mdelay(5);
-
-	return uvd_v7_0_start(adev);
-}
-#endif
-
 static int uvd_v7_0_set_interrupt_state(struct amdgpu_device *adev,
 					struct amdgpu_irq_src *source,
 					unsigned type,
@@ -1603,172 +1513,7 @@ static int uvd_v7_0_process_interrupt(struct amdgpu_device *adev,
 	return 0;
 }
 
-#if 0
-static void uvd_v7_0_set_sw_clock_gating(struct amdgpu_device *adev)
-{
-	uint32_t data, data1, data2, suvd_flags;
-
-	data = RREG32_SOC15(UVD, ring->me, mmUVD_CGC_CTRL);
-	data1 = RREG32_SOC15(UVD, ring->me, mmUVD_SUVD_CGC_GATE);
-	data2 = RREG32_SOC15(UVD, ring->me, mmUVD_SUVD_CGC_CTRL);
-
-	data &= ~(UVD_CGC_CTRL__CLK_OFF_DELAY_MASK |
-		  UVD_CGC_CTRL__CLK_GATE_DLY_TIMER_MASK);
-
-	suvd_flags = UVD_SUVD_CGC_GATE__SRE_MASK |
-		     UVD_SUVD_CGC_GATE__SIT_MASK |
-		     UVD_SUVD_CGC_GATE__SMP_MASK |
-		     UVD_SUVD_CGC_GATE__SCM_MASK |
-		     UVD_SUVD_CGC_GATE__SDB_MASK;
-
-	data |= UVD_CGC_CTRL__DYN_CLOCK_MODE_MASK |
-		(1 << REG_FIELD_SHIFT(UVD_CGC_CTRL, CLK_GATE_DLY_TIMER)) |
-		(4 << REG_FIELD_SHIFT(UVD_CGC_CTRL, CLK_OFF_DELAY));
-
-	data &= ~(UVD_CGC_CTRL__UDEC_RE_MODE_MASK |
-			UVD_CGC_CTRL__UDEC_CM_MODE_MASK |
-			UVD_CGC_CTRL__UDEC_IT_MODE_MASK |
-			UVD_CGC_CTRL__UDEC_DB_MODE_MASK |
-			UVD_CGC_CTRL__UDEC_MP_MODE_MASK |
-			UVD_CGC_CTRL__SYS_MODE_MASK |
-			UVD_CGC_CTRL__UDEC_MODE_MASK |
-			UVD_CGC_CTRL__MPEG2_MODE_MASK |
-			UVD_CGC_CTRL__REGS_MODE_MASK |
-			UVD_CGC_CTRL__RBC_MODE_MASK |
-			UVD_CGC_CTRL__LMI_MC_MODE_MASK |
-			UVD_CGC_CTRL__LMI_UMC_MODE_MASK |
-			UVD_CGC_CTRL__IDCT_MODE_MASK |
-			UVD_CGC_CTRL__MPRD_MODE_MASK |
-			UVD_CGC_CTRL__MPC_MODE_MASK |
-			UVD_CGC_CTRL__LBSI_MODE_MASK |
-			UVD_CGC_CTRL__LRBBM_MODE_MASK |
-			UVD_CGC_CTRL__WCB_MODE_MASK |
-			UVD_CGC_CTRL__VCPU_MODE_MASK |
-			UVD_CGC_CTRL__JPEG_MODE_MASK |
-			UVD_CGC_CTRL__JPEG2_MODE_MASK |
-			UVD_CGC_CTRL__SCPU_MODE_MASK);
-	data2 &= ~(UVD_SUVD_CGC_CTRL__SRE_MODE_MASK |
-			UVD_SUVD_CGC_CTRL__SIT_MODE_MASK |
-			UVD_SUVD_CGC_CTRL__SMP_MODE_MASK |
-			UVD_SUVD_CGC_CTRL__SCM_MODE_MASK |
-			UVD_SUVD_CGC_CTRL__SDB_MODE_MASK);
-	data1 |= suvd_flags;
-
-	WREG32_SOC15(UVD, ring->me, mmUVD_CGC_CTRL, data);
-	WREG32_SOC15(UVD, ring->me, mmUVD_CGC_GATE, 0);
-	WREG32_SOC15(UVD, ring->me, mmUVD_SUVD_CGC_GATE, data1);
-	WREG32_SOC15(UVD, ring->me, mmUVD_SUVD_CGC_CTRL, data2);
-}
-
-static void uvd_v7_0_set_hw_clock_gating(struct amdgpu_device *adev)
-{
-	uint32_t data, data1, cgc_flags, suvd_flags;
-
-	data = RREG32_SOC15(UVD, ring->me, mmUVD_CGC_GATE);
-	data1 = RREG32_SOC15(UVD, ring->me, mmUVD_SUVD_CGC_GATE);
-
-	cgc_flags = UVD_CGC_GATE__SYS_MASK |
-		UVD_CGC_GATE__UDEC_MASK |
-		UVD_CGC_GATE__MPEG2_MASK |
-		UVD_CGC_GATE__RBC_MASK |
-		UVD_CGC_GATE__LMI_MC_MASK |
-		UVD_CGC_GATE__IDCT_MASK |
-		UVD_CGC_GATE__MPRD_MASK |
-		UVD_CGC_GATE__MPC_MASK |
-		UVD_CGC_GATE__LBSI_MASK |
-		UVD_CGC_GATE__LRBBM_MASK |
-		UVD_CGC_GATE__UDEC_RE_MASK |
-		UVD_CGC_GATE__UDEC_CM_MASK |
-		UVD_CGC_GATE__UDEC_IT_MASK |
-		UVD_CGC_GATE__UDEC_DB_MASK |
-		UVD_CGC_GATE__UDEC_MP_MASK |
-		UVD_CGC_GATE__WCB_MASK |
-		UVD_CGC_GATE__VCPU_MASK |
-		UVD_CGC_GATE__SCPU_MASK |
-		UVD_CGC_GATE__JPEG_MASK |
-		UVD_CGC_GATE__JPEG2_MASK;
-
-	suvd_flags = UVD_SUVD_CGC_GATE__SRE_MASK |
-				UVD_SUVD_CGC_GATE__SIT_MASK |
-				UVD_SUVD_CGC_GATE__SMP_MASK |
-				UVD_SUVD_CGC_GATE__SCM_MASK |
-				UVD_SUVD_CGC_GATE__SDB_MASK;
-
-	data |= cgc_flags;
-	data1 |= suvd_flags;
-
-	WREG32_SOC15(UVD, ring->me, mmUVD_CGC_GATE, data);
-	WREG32_SOC15(UVD, ring->me, mmUVD_SUVD_CGC_GATE, data1);
-}
-
-static void uvd_v7_0_set_bypass_mode(struct amdgpu_device *adev, bool enable)
-{
-	u32 tmp = RREG32_SMC(ixGCK_DFS_BYPASS_CNTL);
-
-	if (enable)
-		tmp |= (GCK_DFS_BYPASS_CNTL__BYPASSDCLK_MASK |
-			GCK_DFS_BYPASS_CNTL__BYPASSVCLK_MASK);
-	else
-		tmp &= ~(GCK_DFS_BYPASS_CNTL__BYPASSDCLK_MASK |
-			 GCK_DFS_BYPASS_CNTL__BYPASSVCLK_MASK);
-
-	WREG32_SMC(ixGCK_DFS_BYPASS_CNTL, tmp);
-}
-
-
-static int uvd_v7_0_set_clockgating_state(void *handle,
-					  enum amd_clockgating_state state)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-	bool enable = (state == AMD_CG_STATE_GATE);
-
-	uvd_v7_0_set_bypass_mode(adev, enable);
-
-	if (!(adev->cg_flags & AMD_CG_SUPPORT_UVD_MGCG))
-		return 0;
-
-	if (enable) {
-		/* disable HW gating and enable Sw gating */
-		uvd_v7_0_set_sw_clock_gating(adev);
-	} else {
-		/* wait for STATUS to clear */
-		if (uvd_v7_0_wait_for_idle(handle))
-			return -EBUSY;
-
-		/* enable HW gates because UVD is idle */
-		/* uvd_v7_0_set_hw_clock_gating(adev); */
-	}
-
-	return 0;
-}
-
-static int uvd_v7_0_set_powergating_state(void *handle,
-					  enum amd_powergating_state state)
-{
-	/* This doesn't actually powergate the UVD block.
-	 * That's done in the dpm code via the SMC.  This
-	 * just re-inits the block as necessary.  The actual
-	 * gating still happens in the dpm code.  We should
-	 * revisit this when there is a cleaner line between
-	 * the smc and the hw blocks
-	 */
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	if (!(adev->pg_flags & AMD_PG_SUPPORT_UVD))
-		return 0;
-
-	WREG32_SOC15(UVD, ring->me, mmUVD_POWER_STATUS, UVD_POWER_STATUS__UVD_PG_EN_MASK);
-
-	if (state == AMD_PG_STATE_GATE) {
-		uvd_v7_0_stop(adev);
-		return 0;
-	} else {
-		return uvd_v7_0_start(adev);
-	}
-}
-#endif
-
-static int uvd_v7_0_set_clockgating_state(void *handle,
+static int uvd_v7_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_clockgating_state state)
 {
 	/* needed for driver unload*/
@@ -1778,19 +1523,13 @@ static int uvd_v7_0_set_clockgating_state(void *handle,
 const struct amd_ip_funcs uvd_v7_0_ip_funcs = {
 	.name = "uvd_v7_0",
 	.early_init = uvd_v7_0_early_init,
-	.late_init = NULL,
 	.sw_init = uvd_v7_0_sw_init,
 	.sw_fini = uvd_v7_0_sw_fini,
 	.hw_init = uvd_v7_0_hw_init,
 	.hw_fini = uvd_v7_0_hw_fini,
+	.prepare_suspend = uvd_v7_0_prepare_suspend,
 	.suspend = uvd_v7_0_suspend,
 	.resume = uvd_v7_0_resume,
-	.is_idle = NULL /* uvd_v7_0_is_idle */,
-	.wait_for_idle = NULL /* uvd_v7_0_wait_for_idle */,
-	.check_soft_reset = NULL /* uvd_v7_0_check_soft_reset */,
-	.pre_soft_reset = NULL /* uvd_v7_0_pre_soft_reset */,
-	.soft_reset = NULL /* uvd_v7_0_soft_reset */,
-	.post_soft_reset = NULL /* uvd_v7_0_post_soft_reset */,
 	.set_clockgating_state = uvd_v7_0_set_clockgating_state,
 	.set_powergating_state = NULL /* uvd_v7_0_set_powergating_state */,
 };
@@ -1800,7 +1539,6 @@ static const struct amdgpu_ring_funcs uvd_v7_0_ring_vm_funcs = {
 	.align_mask = 0xf,
 	.support_64bit_ptrs = false,
 	.no_user_fence = true,
-	.vmhub = AMDGPU_MMHUB_0,
 	.get_rptr = uvd_v7_0_ring_get_rptr,
 	.get_wptr = uvd_v7_0_ring_get_wptr,
 	.set_wptr = uvd_v7_0_ring_set_wptr,
@@ -1833,7 +1571,6 @@ static const struct amdgpu_ring_funcs uvd_v7_0_enc_ring_vm_funcs = {
 	.nop = HEVC_ENC_CMD_NO_OP,
 	.support_64bit_ptrs = false,
 	.no_user_fence = true,
-	.vmhub = AMDGPU_MMHUB_0,
 	.get_rptr = uvd_v7_0_enc_ring_get_rptr,
 	.get_wptr = uvd_v7_0_enc_ring_get_wptr,
 	.set_wptr = uvd_v7_0_enc_ring_set_wptr,
@@ -1906,8 +1643,7 @@ static void uvd_v7_0_set_irq_funcs(struct amdgpu_device *adev)
 	}
 }
 
-const struct amdgpu_ip_block_version uvd_v7_0_ip_block =
-{
+const struct amdgpu_ip_block_version uvd_v7_0_ip_block = {
 		.type = AMD_IP_BLOCK_TYPE_UVD,
 		.major = 7,
 		.minor = 0,
diff --git a/drivers/gpu/drm/amd/amdgpu/vce_v1_0.c b/drivers/gpu/drm/amd/amdgpu/vce_v1_0.c
new file mode 100644
index 000000000000..9ae424618556
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/vce_v1_0.c
@@ -0,0 +1,839 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright 2013 Advanced Micro Devices, Inc.
+ * Copyright 2025 Valve Corporation
+ * Copyright 2025 Alexandre Demers
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
+ * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
+ * USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ * Authors: Christian König <christian.koenig@amd.com>
+ *          Timur Kristóf <timur.kristof@gmail.com>
+ *          Alexandre Demers <alexandre.f.demers@gmail.com>
+ */
+
+#include <linux/firmware.h>
+
+#include "amdgpu.h"
+#include "amdgpu_vce.h"
+#include "amdgpu_gart.h"
+#include "sid.h"
+#include "vce_v1_0.h"
+#include "vce/vce_1_0_d.h"
+#include "vce/vce_1_0_sh_mask.h"
+#include "oss/oss_1_0_d.h"
+#include "oss/oss_1_0_sh_mask.h"
+
+#define VCE_V1_0_FW_SIZE	(256 * 1024)
+#define VCE_V1_0_STACK_SIZE	(64 * 1024)
+#define VCE_V1_0_DATA_SIZE	(7808 * (AMDGPU_MAX_VCE_HANDLES + 1))
+#define VCE_STATUS_VCPU_REPORT_FW_LOADED_MASK	0x02
+
+#define VCE_V1_0_GART_PAGE_START \
+	(AMDGPU_GTT_MAX_TRANSFER_SIZE * AMDGPU_GTT_NUM_TRANSFER_WINDOWS)
+#define VCE_V1_0_GART_ADDR_START \
+	(VCE_V1_0_GART_PAGE_START * AMDGPU_GPU_PAGE_SIZE)
+
+static void vce_v1_0_set_ring_funcs(struct amdgpu_device *adev);
+static void vce_v1_0_set_irq_funcs(struct amdgpu_device *adev);
+
+struct vce_v1_0_fw_signature {
+	int32_t offset;
+	uint32_t length;
+	int32_t number;
+	struct {
+		uint32_t chip_id;
+		uint32_t keyselect;
+		uint32_t nonce[4];
+		uint32_t sigval[4];
+	} val[8];
+};
+
+/**
+ * vce_v1_0_ring_get_rptr - get read pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Returns the current hardware read pointer
+ */
+static uint64_t vce_v1_0_ring_get_rptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	if (ring->me == 0)
+		return RREG32(mmVCE_RB_RPTR);
+	else
+		return RREG32(mmVCE_RB_RPTR2);
+}
+
+/**
+ * vce_v1_0_ring_get_wptr - get write pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Returns the current hardware write pointer
+ */
+static uint64_t vce_v1_0_ring_get_wptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	if (ring->me == 0)
+		return RREG32(mmVCE_RB_WPTR);
+	else
+		return RREG32(mmVCE_RB_WPTR2);
+}
+
+/**
+ * vce_v1_0_ring_set_wptr - set write pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Commits the write pointer to the hardware
+ */
+static void vce_v1_0_ring_set_wptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	if (ring->me == 0)
+		WREG32(mmVCE_RB_WPTR, lower_32_bits(ring->wptr));
+	else
+		WREG32(mmVCE_RB_WPTR2, lower_32_bits(ring->wptr));
+}
+
+static int vce_v1_0_lmi_clean(struct amdgpu_device *adev)
+{
+	int i, j;
+
+	for (i = 0; i < 10; ++i) {
+		for (j = 0; j < 100; ++j) {
+			if (RREG32(mmVCE_LMI_STATUS) & 0x337f)
+				return 0;
+
+			mdelay(10);
+		}
+	}
+
+	return -ETIMEDOUT;
+}
+
+static int vce_v1_0_firmware_loaded(struct amdgpu_device *adev)
+{
+	int i, j;
+
+	for (i = 0; i < 10; ++i) {
+		for (j = 0; j < 100; ++j) {
+			if (RREG32(mmVCE_STATUS) & VCE_STATUS_VCPU_REPORT_FW_LOADED_MASK)
+				return 0;
+			mdelay(10);
+		}
+
+		dev_err(adev->dev, "VCE not responding, trying to reset the ECPU\n");
+
+		WREG32_P(mmVCE_SOFT_RESET,
+			VCE_SOFT_RESET__ECPU_SOFT_RESET_MASK,
+			~VCE_SOFT_RESET__ECPU_SOFT_RESET_MASK);
+		mdelay(10);
+		WREG32_P(mmVCE_SOFT_RESET, 0,
+			~VCE_SOFT_RESET__ECPU_SOFT_RESET_MASK);
+		mdelay(10);
+	}
+
+	return -ETIMEDOUT;
+}
+
+static void vce_v1_0_init_cg(struct amdgpu_device *adev)
+{
+	u32 tmp;
+
+	tmp = RREG32(mmVCE_CLOCK_GATING_A);
+	tmp |= VCE_CLOCK_GATING_A__CGC_DYN_CLOCK_MODE_MASK;
+	WREG32(mmVCE_CLOCK_GATING_A, tmp);
+
+	tmp = RREG32(mmVCE_CLOCK_GATING_B);
+	tmp |= 0x1e;
+	tmp &= ~0xe100e1;
+	WREG32(mmVCE_CLOCK_GATING_B, tmp);
+
+	tmp = RREG32(mmVCE_UENC_CLOCK_GATING);
+	tmp &= ~0xff9ff000;
+	WREG32(mmVCE_UENC_CLOCK_GATING, tmp);
+
+	tmp = RREG32(mmVCE_UENC_REG_CLOCK_GATING);
+	tmp &= ~0x3ff;
+	WREG32(mmVCE_UENC_REG_CLOCK_GATING, tmp);
+}
+
+/**
+ * vce_v1_0_load_fw_signature - load firmware signature into VCPU BO
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * The VCE1 firmware validation mechanism needs a firmware signature.
+ * This function finds the signature appropriate for the current
+ * ASIC and writes that into the VCPU BO.
+ */
+static int vce_v1_0_load_fw_signature(struct amdgpu_device *adev)
+{
+	const struct common_firmware_header *hdr;
+	struct vce_v1_0_fw_signature *sign;
+	unsigned int ucode_offset;
+	uint32_t chip_id;
+	u32 *cpu_addr;
+	int i;
+
+	hdr = (const struct common_firmware_header *)adev->vce.fw->data;
+	ucode_offset = le32_to_cpu(hdr->ucode_array_offset_bytes);
+	cpu_addr = adev->vce.cpu_addr;
+
+	sign = (void *)adev->vce.fw->data + ucode_offset;
+
+	switch (adev->asic_type) {
+	case CHIP_TAHITI:
+		chip_id = 0x01000014;
+		break;
+	case CHIP_VERDE:
+		chip_id = 0x01000015;
+		break;
+	case CHIP_PITCAIRN:
+		chip_id = 0x01000016;
+		break;
+	default:
+		dev_err(adev->dev, "asic_type %#010x was not found!", adev->asic_type);
+		return -EINVAL;
+	}
+
+	for (i = 0; i < le32_to_cpu(sign->number); ++i) {
+		if (le32_to_cpu(sign->val[i].chip_id) == chip_id)
+			break;
+	}
+
+	if (i == le32_to_cpu(sign->number)) {
+		dev_err(adev->dev, "chip_id 0x%x for %s was not found in VCE firmware",
+			chip_id, amdgpu_asic_name[adev->asic_type]);
+		return -EINVAL;
+	}
+
+	cpu_addr += (256 - 64) / 4;
+	memcpy_toio(&cpu_addr[0], &sign->val[i].nonce[0], 16);
+	cpu_addr[4] = cpu_to_le32(le32_to_cpu(sign->length) + 64);
+
+	memset_io(&cpu_addr[5], 0, 44);
+	memcpy_toio(&cpu_addr[16], &sign[1], hdr->ucode_size_bytes - sizeof(*sign));
+
+	cpu_addr += (le32_to_cpu(sign->length) + 64) / 4;
+	memcpy_toio(&cpu_addr[0], &sign->val[i].sigval[0], 16);
+
+	adev->vce.keyselect = le32_to_cpu(sign->val[i].keyselect);
+
+	return 0;
+}
+
+static int vce_v1_0_wait_for_fw_validation(struct amdgpu_device *adev)
+{
+	int i;
+
+	dev_dbg(adev->dev, "VCE keyselect: %d", adev->vce.keyselect);
+	WREG32(mmVCE_LMI_FW_START_KEYSEL, adev->vce.keyselect);
+
+	for (i = 0; i < 10; ++i) {
+		mdelay(10);
+		if (RREG32(mmVCE_FW_REG_STATUS) & VCE_FW_REG_STATUS__DONE_MASK)
+			break;
+	}
+
+	if (!(RREG32(mmVCE_FW_REG_STATUS) & VCE_FW_REG_STATUS__DONE_MASK)) {
+		dev_err(adev->dev, "VCE FW validation timeout\n");
+		return -ETIMEDOUT;
+	}
+
+	if (!(RREG32(mmVCE_FW_REG_STATUS) & VCE_FW_REG_STATUS__PASS_MASK)) {
+		dev_err(adev->dev, "VCE FW validation failed\n");
+		return -EINVAL;
+	}
+
+	for (i = 0; i < 10; ++i) {
+		mdelay(10);
+		if (!(RREG32(mmVCE_FW_REG_STATUS) & VCE_FW_REG_STATUS__BUSY_MASK))
+			break;
+	}
+
+	if (RREG32(mmVCE_FW_REG_STATUS) & VCE_FW_REG_STATUS__BUSY_MASK) {
+		dev_err(adev->dev, "VCE FW busy timeout\n");
+		return -ETIMEDOUT;
+	}
+
+	return 0;
+}
+
+static int vce_v1_0_mc_resume(struct amdgpu_device *adev)
+{
+	uint32_t offset;
+	uint32_t size;
+
+	/*
+	 * When the keyselect is already set, don't perturb VCE FW.
+	 * Validation seems to always fail the second time.
+	 */
+	if (RREG32(mmVCE_LMI_FW_START_KEYSEL)) {
+		dev_dbg(adev->dev, "keyselect already set: 0x%x (on CPU: 0x%x)\n",
+			RREG32(mmVCE_LMI_FW_START_KEYSEL), adev->vce.keyselect);
+
+		WREG32_P(mmVCE_LMI_CTRL2, 0x0, ~0x100);
+		return 0;
+	}
+
+	WREG32_P(mmVCE_CLOCK_GATING_A, 0, ~(1 << 16));
+	WREG32_P(mmVCE_UENC_CLOCK_GATING, 0x1FF000, ~0xFF9FF000);
+	WREG32_P(mmVCE_UENC_REG_CLOCK_GATING, 0x3F, ~0x3F);
+	WREG32(mmVCE_CLOCK_GATING_B, 0);
+
+	WREG32_P(mmVCE_LMI_FW_PERIODIC_CTRL, 0x4, ~0x4);
+
+	WREG32(mmVCE_LMI_CTRL, 0x00398000);
+
+	WREG32_P(mmVCE_LMI_CACHE_CTRL, 0x0, ~0x1);
+	WREG32(mmVCE_LMI_SWAP_CNTL, 0);
+	WREG32(mmVCE_LMI_SWAP_CNTL1, 0);
+	WREG32(mmVCE_LMI_VM_CTRL, 0);
+
+	WREG32(mmVCE_VCPU_SCRATCH7, AMDGPU_MAX_VCE_HANDLES);
+
+	offset =  adev->vce.gpu_addr + AMDGPU_VCE_FIRMWARE_OFFSET;
+	size = VCE_V1_0_FW_SIZE;
+	WREG32(mmVCE_VCPU_CACHE_OFFSET0, offset & 0x7fffffff);
+	WREG32(mmVCE_VCPU_CACHE_SIZE0, size);
+
+	offset += size;
+	size = VCE_V1_0_STACK_SIZE;
+	WREG32(mmVCE_VCPU_CACHE_OFFSET1, offset & 0x7fffffff);
+	WREG32(mmVCE_VCPU_CACHE_SIZE1, size);
+
+	offset += size;
+	size = VCE_V1_0_DATA_SIZE;
+	WREG32(mmVCE_VCPU_CACHE_OFFSET2, offset & 0x7fffffff);
+	WREG32(mmVCE_VCPU_CACHE_SIZE2, size);
+
+	WREG32_P(mmVCE_LMI_CTRL2, 0x0, ~0x100);
+
+	return vce_v1_0_wait_for_fw_validation(adev);
+}
+
+/**
+ * vce_v1_0_is_idle() - Check idle status of VCE1 IP block
+ *
+ * @ip_block: amdgpu_ip_block pointer
+ *
+ * Check whether VCE is busy according to VCE_STATUS.
+ * Also check whether the SRBM thinks VCE is busy, although
+ * SRBM_STATUS.VCE_BUSY seems to be bogus because it
+ * appears to mirror the VCE_STATUS.VCPU_REPORT_FW_LOADED bit.
+ */
+static bool vce_v1_0_is_idle(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	bool busy =
+		(RREG32(mmVCE_STATUS) & (VCE_STATUS__JOB_BUSY_MASK | VCE_STATUS__UENC_BUSY_MASK)) ||
+		(RREG32(mmSRBM_STATUS2) & SRBM_STATUS2__VCE_BUSY_MASK);
+
+	return !busy;
+}
+
+static int vce_v1_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	unsigned int i;
+
+	for (i = 0; i < adev->usec_timeout; i++) {
+		udelay(1);
+		if (vce_v1_0_is_idle(ip_block))
+			return 0;
+	}
+	return -ETIMEDOUT;
+}
+
+/**
+ * vce_v1_0_start - start VCE block
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Setup and start the VCE block
+ */
+static int vce_v1_0_start(struct amdgpu_device *adev)
+{
+	struct amdgpu_ring *ring;
+	int r;
+
+	WREG32_P(mmVCE_STATUS, 1, ~1);
+
+	r = vce_v1_0_mc_resume(adev);
+	if (r)
+		return r;
+
+	ring = &adev->vce.ring[0];
+	WREG32(mmVCE_RB_RPTR, lower_32_bits(ring->wptr));
+	WREG32(mmVCE_RB_WPTR, lower_32_bits(ring->wptr));
+	WREG32(mmVCE_RB_BASE_LO, lower_32_bits(ring->gpu_addr));
+	WREG32(mmVCE_RB_BASE_HI, upper_32_bits(ring->gpu_addr));
+	WREG32(mmVCE_RB_SIZE, ring->ring_size / 4);
+
+	ring = &adev->vce.ring[1];
+	WREG32(mmVCE_RB_RPTR2, lower_32_bits(ring->wptr));
+	WREG32(mmVCE_RB_WPTR2, lower_32_bits(ring->wptr));
+	WREG32(mmVCE_RB_BASE_LO2, lower_32_bits(ring->gpu_addr));
+	WREG32(mmVCE_RB_BASE_HI2, upper_32_bits(ring->gpu_addr));
+	WREG32(mmVCE_RB_SIZE2, ring->ring_size / 4);
+
+	WREG32_P(mmVCE_VCPU_CNTL, VCE_VCPU_CNTL__CLK_EN_MASK,
+		 ~VCE_VCPU_CNTL__CLK_EN_MASK);
+
+	WREG32_P(mmVCE_SOFT_RESET,
+		VCE_SOFT_RESET__ECPU_SOFT_RESET_MASK |
+		VCE_SOFT_RESET__FME_SOFT_RESET_MASK,
+		~(VCE_SOFT_RESET__ECPU_SOFT_RESET_MASK |
+		  VCE_SOFT_RESET__FME_SOFT_RESET_MASK));
+
+	mdelay(100);
+
+	WREG32_P(mmVCE_SOFT_RESET, 0,
+		~(VCE_SOFT_RESET__ECPU_SOFT_RESET_MASK |
+		  VCE_SOFT_RESET__FME_SOFT_RESET_MASK));
+
+	r = vce_v1_0_firmware_loaded(adev);
+
+	/* Clear VCE_STATUS, otherwise SRBM thinks VCE1 is busy. */
+	WREG32(mmVCE_STATUS, 0);
+
+	if (r) {
+		dev_err(adev->dev, "VCE not responding, giving up\n");
+		return r;
+	}
+
+	return 0;
+}
+
+static int vce_v1_0_stop(struct amdgpu_device *adev)
+{
+	struct amdgpu_ip_block *ip_block;
+	int status;
+	int i;
+
+	ip_block = amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_VCE);
+	if (!ip_block)
+		return -EINVAL;
+
+	if (vce_v1_0_lmi_clean(adev))
+		dev_warn(adev->dev, "VCE not idle\n");
+
+	if (vce_v1_0_wait_for_idle(ip_block))
+		dev_warn(adev->dev, "VCE busy: VCE_STATUS=0x%x, SRBM_STATUS2=0x%x\n",
+			RREG32(mmVCE_STATUS), RREG32(mmSRBM_STATUS2));
+
+	/* Stall UMC and register bus before resetting VCPU */
+	WREG32_P(mmVCE_LMI_CTRL2, 1 << 8, ~(1 << 8));
+
+	for (i = 0; i < 100; ++i) {
+		status = RREG32(mmVCE_LMI_STATUS);
+		if (status & 0x240)
+			break;
+		mdelay(1);
+	}
+
+	WREG32_P(mmVCE_VCPU_CNTL, 0, ~VCE_VCPU_CNTL__CLK_EN_MASK);
+
+	WREG32_P(mmVCE_SOFT_RESET,
+		VCE_SOFT_RESET__ECPU_SOFT_RESET_MASK |
+		VCE_SOFT_RESET__FME_SOFT_RESET_MASK,
+		~(VCE_SOFT_RESET__ECPU_SOFT_RESET_MASK |
+		  VCE_SOFT_RESET__FME_SOFT_RESET_MASK));
+
+	WREG32(mmVCE_STATUS, 0);
+
+	return 0;
+}
+
+static void vce_v1_0_enable_mgcg(struct amdgpu_device *adev, bool enable)
+{
+	u32 tmp;
+
+	if (enable && (adev->cg_flags & AMD_CG_SUPPORT_VCE_MGCG)) {
+		tmp = RREG32(mmVCE_CLOCK_GATING_A);
+		tmp |= VCE_CLOCK_GATING_A__CGC_DYN_CLOCK_MODE_MASK;
+		WREG32(mmVCE_CLOCK_GATING_A, tmp);
+
+		tmp = RREG32(mmVCE_UENC_CLOCK_GATING);
+		tmp &= ~0x1ff000;
+		tmp |= 0xff800000;
+		WREG32(mmVCE_UENC_CLOCK_GATING, tmp);
+
+		tmp = RREG32(mmVCE_UENC_REG_CLOCK_GATING);
+		tmp &= ~0x3ff;
+		WREG32(mmVCE_UENC_REG_CLOCK_GATING, tmp);
+	} else {
+		tmp = RREG32(mmVCE_CLOCK_GATING_A);
+		tmp &= ~VCE_CLOCK_GATING_A__CGC_DYN_CLOCK_MODE_MASK;
+		WREG32(mmVCE_CLOCK_GATING_A, tmp);
+
+		tmp = RREG32(mmVCE_UENC_CLOCK_GATING);
+		tmp |= 0x1ff000;
+		tmp &= ~0xff800000;
+		WREG32(mmVCE_UENC_CLOCK_GATING, tmp);
+
+		tmp = RREG32(mmVCE_UENC_REG_CLOCK_GATING);
+		tmp |= 0x3ff;
+		WREG32(mmVCE_UENC_REG_CLOCK_GATING, tmp);
+	}
+}
+
+static int vce_v1_0_early_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int r;
+
+	r = amdgpu_vce_early_init(adev);
+	if (r)
+		return r;
+
+	adev->vce.num_rings = 2;
+
+	vce_v1_0_set_ring_funcs(adev);
+	vce_v1_0_set_irq_funcs(adev);
+
+	return 0;
+}
+
+/**
+ * vce_v1_0_ensure_vcpu_bo_32bit_addr() - ensure the VCPU BO has a 32-bit address
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Due to various hardware limitations, the VCE1 requires
+ * the VCPU BO to be in the low 32 bit address range.
+ * Ensure that the VCPU BO has a 32-bit GPU address,
+ * or return an error code when that isn't possible.
+ *
+ * To accomodate that, we put GART to the LOW address range
+ * and reserve some GART pages where we map the VCPU BO,
+ * so that it gets a 32-bit address.
+ */
+static int vce_v1_0_ensure_vcpu_bo_32bit_addr(struct amdgpu_device *adev)
+{
+	u64 gpu_addr = amdgpu_bo_gpu_offset(adev->vce.vcpu_bo);
+	u64 bo_size = amdgpu_bo_size(adev->vce.vcpu_bo);
+	u64 max_vcpu_bo_addr = 0xffffffff - bo_size;
+	u64 num_pages = ALIGN(bo_size, AMDGPU_GPU_PAGE_SIZE) / AMDGPU_GPU_PAGE_SIZE;
+	u64 pa = amdgpu_gmc_vram_pa(adev, adev->vce.vcpu_bo);
+	u64 flags = AMDGPU_PTE_READABLE | AMDGPU_PTE_WRITEABLE | AMDGPU_PTE_VALID;
+
+	/*
+	 * Check if the VCPU BO already has a 32-bit address.
+	 * Eg. if MC is configured to put VRAM in the low address range.
+	 */
+	if (gpu_addr <= max_vcpu_bo_addr)
+		return 0;
+
+	/* Check if we can map the VCPU BO in GART to a 32-bit address. */
+	if (adev->gmc.gart_start + VCE_V1_0_GART_ADDR_START > max_vcpu_bo_addr)
+		return -EINVAL;
+
+	amdgpu_gart_map_vram_range(adev, pa, VCE_V1_0_GART_PAGE_START,
+				   num_pages, flags, adev->gart.ptr);
+	adev->vce.gpu_addr = adev->gmc.gart_start + VCE_V1_0_GART_ADDR_START;
+	if (adev->vce.gpu_addr > max_vcpu_bo_addr)
+		return -EINVAL;
+
+	return 0;
+}
+
+static int vce_v1_0_sw_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	struct amdgpu_ring *ring;
+	int r, i;
+
+	r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 167, &adev->vce.irq);
+	if (r)
+		return r;
+
+	r = amdgpu_vce_sw_init(adev, VCE_V1_0_FW_SIZE +
+		VCE_V1_0_STACK_SIZE + VCE_V1_0_DATA_SIZE);
+	if (r)
+		return r;
+
+	r = amdgpu_vce_resume(adev);
+	if (r)
+		return r;
+	r = vce_v1_0_load_fw_signature(adev);
+	if (r)
+		return r;
+	r = vce_v1_0_ensure_vcpu_bo_32bit_addr(adev);
+	if (r)
+		return r;
+
+	for (i = 0; i < adev->vce.num_rings; i++) {
+		enum amdgpu_ring_priority_level hw_prio = amdgpu_vce_get_ring_prio(i);
+
+		ring = &adev->vce.ring[i];
+		sprintf(ring->name, "vce%d", i);
+		r = amdgpu_ring_init(adev, ring, 512, &adev->vce.irq, 0,
+				     hw_prio, NULL);
+		if (r)
+			return r;
+	}
+
+	return r;
+}
+
+static int vce_v1_0_sw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int r;
+
+	r = amdgpu_vce_suspend(adev);
+	if (r)
+		return r;
+
+	return amdgpu_vce_sw_fini(adev);
+}
+
+/**
+ * vce_v1_0_hw_init - start and test VCE block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Initialize the hardware, boot up the VCPU and do some testing
+ */
+static int vce_v1_0_hw_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, r;
+
+	if (adev->pm.dpm_enabled)
+		amdgpu_dpm_enable_vce(adev, true);
+	else
+		amdgpu_asic_set_vce_clocks(adev, 10000, 10000);
+
+	for (i = 0; i < adev->vce.num_rings; i++) {
+		r = amdgpu_ring_test_helper(&adev->vce.ring[i]);
+		if (r)
+			return r;
+	}
+
+	dev_info(adev->dev, "VCE initialized successfully.\n");
+
+	return 0;
+}
+
+static int vce_v1_0_hw_fini(struct amdgpu_ip_block *ip_block)
+{
+	int r;
+
+	r = vce_v1_0_stop(ip_block->adev);
+	if (r)
+		return r;
+
+	cancel_delayed_work_sync(&ip_block->adev->vce.idle_work);
+	return 0;
+}
+
+static int vce_v1_0_suspend(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int r;
+
+	/*
+	 * Proper cleanups before halting the HW engine:
+	 *   - cancel the delayed idle work
+	 *   - enable powergating
+	 *   - enable clockgating
+	 *   - disable dpm
+	 *
+	 * TODO: to align with the VCN implementation, move the
+	 * jobs for clockgating/powergating/dpm setting to
+	 * ->set_powergating_state().
+	 */
+	cancel_delayed_work_sync(&adev->vce.idle_work);
+
+	if (adev->pm.dpm_enabled) {
+		amdgpu_dpm_enable_vce(adev, false);
+	} else {
+		amdgpu_asic_set_vce_clocks(adev, 0, 0);
+		amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_VCE,
+						       AMD_PG_STATE_GATE);
+		amdgpu_device_ip_set_clockgating_state(adev, AMD_IP_BLOCK_TYPE_VCE,
+						       AMD_CG_STATE_GATE);
+	}
+
+	r = vce_v1_0_hw_fini(ip_block);
+	if (r) {
+		dev_err(adev->dev, "vce_v1_0_hw_fini() failed with error %i", r);
+		return r;
+	}
+
+	return amdgpu_vce_suspend(adev);
+}
+
+static int vce_v1_0_resume(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int r;
+
+	r = amdgpu_vce_resume(adev);
+	if (r)
+		return r;
+	r = vce_v1_0_load_fw_signature(adev);
+	if (r)
+		return r;
+	r = vce_v1_0_ensure_vcpu_bo_32bit_addr(adev);
+	if (r)
+		return r;
+
+	return vce_v1_0_hw_init(ip_block);
+}
+
+static int vce_v1_0_set_interrupt_state(struct amdgpu_device *adev,
+					struct amdgpu_irq_src *source,
+					unsigned int type,
+					enum amdgpu_interrupt_state state)
+{
+	uint32_t val = 0;
+
+	if (state == AMDGPU_IRQ_STATE_ENABLE)
+		val |= VCE_SYS_INT_EN__VCE_SYS_INT_TRAP_INTERRUPT_EN_MASK;
+
+	WREG32_P(mmVCE_SYS_INT_EN, val,
+		 ~VCE_SYS_INT_EN__VCE_SYS_INT_TRAP_INTERRUPT_EN_MASK);
+	return 0;
+}
+
+static int vce_v1_0_process_interrupt(struct amdgpu_device *adev,
+				      struct amdgpu_irq_src *source,
+				      struct amdgpu_iv_entry *entry)
+{
+	dev_dbg(adev->dev, "IH: VCE\n");
+	switch (entry->src_data[0]) {
+	case 0:
+	case 1:
+		amdgpu_fence_process(&adev->vce.ring[entry->src_data[0]]);
+		break;
+	default:
+		dev_err(adev->dev, "Unhandled interrupt: %d %d\n",
+			  entry->src_id, entry->src_data[0]);
+		break;
+	}
+
+	return 0;
+}
+
+static int vce_v1_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
+					  enum amd_clockgating_state state)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	vce_v1_0_init_cg(adev);
+	vce_v1_0_enable_mgcg(adev, state == AMD_CG_STATE_GATE);
+
+	return 0;
+}
+
+static int vce_v1_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
+					  enum amd_powergating_state state)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	/*
+	 * This doesn't actually powergate the VCE block.
+	 * That's done in the dpm code via the SMC.  This
+	 * just re-inits the block as necessary.  The actual
+	 * gating still happens in the dpm code.  We should
+	 * revisit this when there is a cleaner line between
+	 * the smc and the hw blocks
+	 */
+	if (state == AMD_PG_STATE_GATE)
+		return vce_v1_0_stop(adev);
+	else
+		return vce_v1_0_start(adev);
+}
+
+static const struct amd_ip_funcs vce_v1_0_ip_funcs = {
+	.name = "vce_v1_0",
+	.early_init = vce_v1_0_early_init,
+	.sw_init = vce_v1_0_sw_init,
+	.sw_fini = vce_v1_0_sw_fini,
+	.hw_init = vce_v1_0_hw_init,
+	.hw_fini = vce_v1_0_hw_fini,
+	.suspend = vce_v1_0_suspend,
+	.resume = vce_v1_0_resume,
+	.is_idle = vce_v1_0_is_idle,
+	.wait_for_idle = vce_v1_0_wait_for_idle,
+	.set_clockgating_state = vce_v1_0_set_clockgating_state,
+	.set_powergating_state = vce_v1_0_set_powergating_state,
+};
+
+static const struct amdgpu_ring_funcs vce_v1_0_ring_funcs = {
+	.type = AMDGPU_RING_TYPE_VCE,
+	.align_mask = 0xf,
+	.nop = VCE_CMD_NO_OP,
+	.support_64bit_ptrs = false,
+	.no_user_fence = true,
+	.get_rptr = vce_v1_0_ring_get_rptr,
+	.get_wptr = vce_v1_0_ring_get_wptr,
+	.set_wptr = vce_v1_0_ring_set_wptr,
+	.parse_cs = amdgpu_vce_ring_parse_cs,
+	.emit_frame_size = 6, /* amdgpu_vce_ring_emit_fence  x1 no user fence */
+	.emit_ib_size = 4, /* amdgpu_vce_ring_emit_ib */
+	.emit_ib = amdgpu_vce_ring_emit_ib,
+	.emit_fence = amdgpu_vce_ring_emit_fence,
+	.test_ring = amdgpu_vce_ring_test_ring,
+	.test_ib = amdgpu_vce_ring_test_ib,
+	.insert_nop = amdgpu_ring_insert_nop,
+	.pad_ib = amdgpu_ring_generic_pad_ib,
+	.begin_use = amdgpu_vce_ring_begin_use,
+	.end_use = amdgpu_vce_ring_end_use,
+};
+
+static void vce_v1_0_set_ring_funcs(struct amdgpu_device *adev)
+{
+	int i;
+
+	for (i = 0; i < adev->vce.num_rings; i++) {
+		adev->vce.ring[i].funcs = &vce_v1_0_ring_funcs;
+		adev->vce.ring[i].me = i;
+	}
+};
+
+static const struct amdgpu_irq_src_funcs vce_v1_0_irq_funcs = {
+	.set = vce_v1_0_set_interrupt_state,
+	.process = vce_v1_0_process_interrupt,
+};
+
+static void vce_v1_0_set_irq_funcs(struct amdgpu_device *adev)
+{
+	adev->vce.irq.num_types = 1;
+	adev->vce.irq.funcs = &vce_v1_0_irq_funcs;
+};
+
+const struct amdgpu_ip_block_version vce_v1_0_ip_block = {
+	.type = AMD_IP_BLOCK_TYPE_VCE,
+	.major = 1,
+	.minor = 0,
+	.rev = 0,
+	.funcs = &vce_v1_0_ip_funcs,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/vce_v1_0.h b/drivers/gpu/drm/amd/amdgpu/vce_v1_0.h
new file mode 100644
index 000000000000..206e7bec897f
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/vce_v1_0.h
@@ -0,0 +1,32 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2025 Advanced Micro Devices, Inc.
+ * Copyright 2025 Valve Corporation
+ * Copyright 2025 Alexandre Demers
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __VCE_V1_0_H__
+#define __VCE_V1_0_H__
+
+extern const struct amdgpu_ip_block_version vce_v1_0_ip_block;
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/vce_v2_0.c b/drivers/gpu/drm/amd/amdgpu/vce_v2_0.c
index 67eb01fef789..8ea8a6193492 100644
--- a/drivers/gpu/drm/amd/amdgpu/vce_v2_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vce_v2_0.c
@@ -201,20 +201,20 @@ static void vce_v2_0_mc_resume(struct amdgpu_device *adev)
 	WREG32_FIELD(VCE_SYS_INT_EN, VCE_SYS_INT_TRAP_INTERRUPT_EN, 1);
 }
 
-static bool vce_v2_0_is_idle(void *handle)
+static bool vce_v2_0_is_idle(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	return !(RREG32(mmSRBM_STATUS2) & SRBM_STATUS2__VCE_BUSY_MASK);
 }
 
-static int vce_v2_0_wait_for_idle(void *handle)
+static int vce_v2_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	unsigned i;
 
 	for (i = 0; i < adev->usec_timeout; i++) {
-		if (vce_v2_0_is_idle(handle))
+		if (vce_v2_0_is_idle(ip_block))
 			return 0;
 	}
 	return -ETIMEDOUT;
@@ -274,15 +274,21 @@ static int vce_v2_0_start(struct amdgpu_device *adev)
 
 static int vce_v2_0_stop(struct amdgpu_device *adev)
 {
+	struct amdgpu_ip_block *ip_block;
 	int i;
 	int status;
 
+
 	if (vce_v2_0_lmi_clean(adev)) {
-		DRM_INFO("vce is not idle \n");
+		DRM_INFO("VCE is not idle \n");
 		return 0;
 	}
 
-	if (vce_v2_0_wait_for_idle(adev)) {
+	ip_block = amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_VCE);
+	if (!ip_block)
+		return -EINVAL;
+
+	if (vce_v2_0_wait_for_idle(ip_block)) {
 		DRM_INFO("VCE is busy, Can't set clock gating");
 		return 0;
 	}
@@ -398,9 +404,14 @@ static void vce_v2_0_enable_mgcg(struct amdgpu_device *adev, bool enable,
 	}
 }
 
-static int vce_v2_0_early_init(void *handle)
+static int vce_v2_0_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
+	int r;
+
+	r = amdgpu_vce_early_init(adev);
+	if (r)
+		return r;
 
 	adev->vce.num_rings = 2;
 
@@ -410,11 +421,11 @@ static int vce_v2_0_early_init(void *handle)
 	return 0;
 }
 
-static int vce_v2_0_sw_init(void *handle)
+static int vce_v2_0_sw_init(struct amdgpu_ip_block *ip_block)
 {
 	struct amdgpu_ring *ring;
 	int r, i;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	/* VCE */
 	r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 167, &adev->vce.irq);
@@ -441,15 +452,13 @@ static int vce_v2_0_sw_init(void *handle)
 			return r;
 	}
 
-	r = amdgpu_vce_entity_init(adev);
-
 	return r;
 }
 
-static int vce_v2_0_sw_fini(void *handle)
+static int vce_v2_0_sw_fini(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	r = amdgpu_vce_suspend(adev);
 	if (r)
@@ -458,10 +467,10 @@ static int vce_v2_0_sw_fini(void *handle)
 	return amdgpu_vce_sw_fini(adev);
 }
 
-static int vce_v2_0_hw_init(void *handle)
+static int vce_v2_0_hw_init(struct amdgpu_ip_block *ip_block)
 {
 	int r, i;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	amdgpu_asic_set_vce_clocks(adev, 10000, 10000);
 	vce_v2_0_enable_mgcg(adev, true, false);
@@ -477,19 +486,17 @@ static int vce_v2_0_hw_init(void *handle)
 	return 0;
 }
 
-static int vce_v2_0_hw_fini(void *handle)
+static int vce_v2_0_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	cancel_delayed_work_sync(&adev->vce.idle_work);
+	cancel_delayed_work_sync(&ip_block->adev->vce.idle_work);
 
 	return 0;
 }
 
-static int vce_v2_0_suspend(void *handle)
+static int vce_v2_0_suspend(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 
 	/*
@@ -515,28 +522,27 @@ static int vce_v2_0_suspend(void *handle)
 						       AMD_CG_STATE_GATE);
 	}
 
-	r = vce_v2_0_hw_fini(adev);
+	r = vce_v2_0_hw_fini(ip_block);
 	if (r)
 		return r;
 
 	return amdgpu_vce_suspend(adev);
 }
 
-static int vce_v2_0_resume(void *handle)
+static int vce_v2_0_resume(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
-	r = amdgpu_vce_resume(adev);
+	r = amdgpu_vce_resume(ip_block->adev);
 	if (r)
 		return r;
 
-	return vce_v2_0_hw_init(adev);
+	return vce_v2_0_hw_init(ip_block);
 }
 
-static int vce_v2_0_soft_reset(void *handle)
+static int vce_v2_0_soft_reset(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	WREG32_FIELD(SRBM_SOFT_RESET, SOFT_RESET_VCE, 1);
 	mdelay(5);
@@ -577,13 +583,13 @@ static int vce_v2_0_process_interrupt(struct amdgpu_device *adev,
 	return 0;
 }
 
-static int vce_v2_0_set_clockgating_state(void *handle,
+static int vce_v2_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_clockgating_state state)
 {
 	bool gate = false;
 	bool sw_cg = false;
 
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (state == AMD_CG_STATE_GATE) {
 		gate = true;
@@ -595,7 +601,7 @@ static int vce_v2_0_set_clockgating_state(void *handle,
 	return 0;
 }
 
-static int vce_v2_0_set_powergating_state(void *handle,
+static int vce_v2_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_powergating_state state)
 {
 	/* This doesn't actually powergate the VCE block.
@@ -605,7 +611,7 @@ static int vce_v2_0_set_powergating_state(void *handle,
 	 * revisit this when there is a cleaner line between
 	 * the smc and the hw blocks
 	 */
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (state == AMD_PG_STATE_GATE)
 		return vce_v2_0_stop(adev);
@@ -616,7 +622,6 @@ static int vce_v2_0_set_powergating_state(void *handle,
 static const struct amd_ip_funcs vce_v2_0_ip_funcs = {
 	.name = "vce_v2_0",
 	.early_init = vce_v2_0_early_init,
-	.late_init = NULL,
 	.sw_init = vce_v2_0_sw_init,
 	.sw_fini = vce_v2_0_sw_fini,
 	.hw_init = vce_v2_0_hw_init,
diff --git a/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c
index 142e291983b4..719e9643c43d 100644
--- a/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c
@@ -64,8 +64,8 @@
 static void vce_v3_0_mc_resume(struct amdgpu_device *adev, int idx);
 static void vce_v3_0_set_ring_funcs(struct amdgpu_device *adev);
 static void vce_v3_0_set_irq_funcs(struct amdgpu_device *adev);
-static int vce_v3_0_wait_for_idle(void *handle);
-static int vce_v3_0_set_clockgating_state(void *handle,
+static int vce_v3_0_wait_for_idle(struct amdgpu_ip_block *ip_block);
+static int vce_v3_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_clockgating_state state);
 /**
  * vce_v3_0_ring_get_rptr - get read pointer
@@ -396,9 +396,10 @@ static unsigned vce_v3_0_get_harvest_config(struct amdgpu_device *adev)
 	}
 }
 
-static int vce_v3_0_early_init(void *handle)
+static int vce_v3_0_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
+	int r;
 
 	adev->vce.harvest_config = vce_v3_0_get_harvest_config(adev);
 
@@ -407,6 +408,10 @@ static int vce_v3_0_early_init(void *handle)
 	    (AMDGPU_VCE_HARVEST_VCE0 | AMDGPU_VCE_HARVEST_VCE1))
 		return -ENOENT;
 
+	r = amdgpu_vce_early_init(adev);
+	if (r)
+		return r;
+
 	adev->vce.num_rings = 3;
 
 	vce_v3_0_set_ring_funcs(adev);
@@ -415,9 +420,9 @@ static int vce_v3_0_early_init(void *handle)
 	return 0;
 }
 
-static int vce_v3_0_sw_init(void *handle)
+static int vce_v3_0_sw_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	struct amdgpu_ring *ring;
 	int r, i;
 
@@ -450,15 +455,13 @@ static int vce_v3_0_sw_init(void *handle)
 			return r;
 	}
 
-	r = amdgpu_vce_entity_init(adev);
-
 	return r;
 }
 
-static int vce_v3_0_sw_fini(void *handle)
+static int vce_v3_0_sw_fini(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	r = amdgpu_vce_suspend(adev);
 	if (r)
@@ -467,10 +470,10 @@ static int vce_v3_0_sw_fini(void *handle)
 	return amdgpu_vce_sw_fini(adev);
 }
 
-static int vce_v3_0_hw_init(void *handle)
+static int vce_v3_0_hw_init(struct amdgpu_ip_block *ip_block)
 {
 	int r, i;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	vce_v3_0_override_vce_clock_gating(adev, true);
 
@@ -487,25 +490,25 @@ static int vce_v3_0_hw_init(void *handle)
 	return 0;
 }
 
-static int vce_v3_0_hw_fini(void *handle)
+static int vce_v3_0_hw_fini(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	cancel_delayed_work_sync(&adev->vce.idle_work);
 
-	r = vce_v3_0_wait_for_idle(handle);
+	r = vce_v3_0_wait_for_idle(ip_block);
 	if (r)
 		return r;
 
 	vce_v3_0_stop(adev);
-	return vce_v3_0_set_clockgating_state(adev, AMD_CG_STATE_GATE);
+	return vce_v3_0_set_clockgating_state(ip_block, AMD_CG_STATE_GATE);
 }
 
-static int vce_v3_0_suspend(void *handle)
+static int vce_v3_0_suspend(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	/*
 	 * Proper cleanups before halting the HW engine:
@@ -530,23 +533,22 @@ static int vce_v3_0_suspend(void *handle)
 						       AMD_CG_STATE_GATE);
 	}
 
-	r = vce_v3_0_hw_fini(adev);
+	r = vce_v3_0_hw_fini(ip_block);
 	if (r)
 		return r;
 
 	return amdgpu_vce_suspend(adev);
 }
 
-static int vce_v3_0_resume(void *handle)
+static int vce_v3_0_resume(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
-	r = amdgpu_vce_resume(adev);
+	r = amdgpu_vce_resume(ip_block->adev);
 	if (r)
 		return r;
 
-	return vce_v3_0_hw_init(adev);
+	return vce_v3_0_hw_init(ip_block);
 }
 
 static void vce_v3_0_mc_resume(struct amdgpu_device *adev, int idx)
@@ -600,9 +602,9 @@ static void vce_v3_0_mc_resume(struct amdgpu_device *adev, int idx)
 	WREG32_FIELD(VCE_SYS_INT_EN, VCE_SYS_INT_TRAP_INTERRUPT_EN, 1);
 }
 
-static bool vce_v3_0_is_idle(void *handle)
+static bool vce_v3_0_is_idle(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	u32 mask = 0;
 
 	mask |= (adev->vce.harvest_config & AMDGPU_VCE_HARVEST_VCE0) ? 0 : SRBM_STATUS2__VCE0_BUSY_MASK;
@@ -611,13 +613,13 @@ static bool vce_v3_0_is_idle(void *handle)
 	return !(RREG32(mmSRBM_STATUS2) & mask);
 }
 
-static int vce_v3_0_wait_for_idle(void *handle)
+static int vce_v3_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
 {
 	unsigned i;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	for (i = 0; i < adev->usec_timeout; i++)
-		if (vce_v3_0_is_idle(handle))
+		if (vce_v3_0_is_idle(ip_block))
 			return 0;
 
 	return -ETIMEDOUT;
@@ -629,9 +631,9 @@ static int vce_v3_0_wait_for_idle(void *handle)
 #define  AMDGPU_VCE_STATUS_BUSY_MASK (VCE_STATUS_VCPU_REPORT_AUTO_BUSY_MASK | \
 				      VCE_STATUS_VCPU_REPORT_RB0_BUSY_MASK)
 
-static bool vce_v3_0_check_soft_reset(void *handle)
+static bool vce_v3_0_check_soft_reset(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	u32 srbm_soft_reset = 0;
 
 	/* According to VCE team , we should use VCE_STATUS instead
@@ -670,9 +672,9 @@ static bool vce_v3_0_check_soft_reset(void *handle)
 	}
 }
 
-static int vce_v3_0_soft_reset(void *handle)
+static int vce_v3_0_soft_reset(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	u32 srbm_soft_reset;
 
 	if (!adev->vce.srbm_soft_reset)
@@ -701,29 +703,29 @@ static int vce_v3_0_soft_reset(void *handle)
 	return 0;
 }
 
-static int vce_v3_0_pre_soft_reset(void *handle)
+static int vce_v3_0_pre_soft_reset(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (!adev->vce.srbm_soft_reset)
 		return 0;
 
 	mdelay(5);
 
-	return vce_v3_0_suspend(adev);
+	return vce_v3_0_suspend(ip_block);
 }
 
 
-static int vce_v3_0_post_soft_reset(void *handle)
+static int vce_v3_0_post_soft_reset(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (!adev->vce.srbm_soft_reset)
 		return 0;
 
 	mdelay(5);
 
-	return vce_v3_0_resume(adev);
+	return vce_v3_0_resume(ip_block);
 }
 
 static int vce_v3_0_set_interrupt_state(struct amdgpu_device *adev,
@@ -763,10 +765,10 @@ static int vce_v3_0_process_interrupt(struct amdgpu_device *adev,
 	return 0;
 }
 
-static int vce_v3_0_set_clockgating_state(void *handle,
+static int vce_v3_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_clockgating_state state)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	bool enable = (state == AMD_CG_STATE_GATE);
 	int i;
 
@@ -804,7 +806,7 @@ static int vce_v3_0_set_clockgating_state(void *handle,
 	return 0;
 }
 
-static int vce_v3_0_set_powergating_state(void *handle,
+static int vce_v3_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_powergating_state state)
 {
 	/* This doesn't actually powergate the VCE block.
@@ -814,7 +816,7 @@ static int vce_v3_0_set_powergating_state(void *handle,
 	 * revisit this when there is a cleaner line between
 	 * the smc and the hw blocks
 	 */
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int ret = 0;
 
 	if (state == AMD_PG_STATE_GATE) {
@@ -831,9 +833,9 @@ out:
 	return ret;
 }
 
-static void vce_v3_0_get_clockgating_state(void *handle, u32 *flags)
+static void vce_v3_0_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int data;
 
 	mutex_lock(&adev->pm.mutex);
@@ -899,7 +901,6 @@ static void vce_v3_0_emit_pipeline_sync(struct amdgpu_ring *ring)
 static const struct amd_ip_funcs vce_v3_0_ip_funcs = {
 	.name = "vce_v3_0",
 	.early_init = vce_v3_0_early_init,
-	.late_init = NULL,
 	.sw_init = vce_v3_0_sw_init,
 	.sw_fini = vce_v3_0_sw_fini,
 	.hw_init = vce_v3_0_hw_init,
@@ -950,7 +951,7 @@ static const struct amdgpu_ring_funcs vce_v3_0_ring_vm_funcs = {
 	.get_rptr = vce_v3_0_ring_get_rptr,
 	.get_wptr = vce_v3_0_ring_get_wptr,
 	.set_wptr = vce_v3_0_ring_set_wptr,
-	.parse_cs = amdgpu_vce_ring_parse_cs_vm,
+	.patch_cs_in_place = amdgpu_vce_ring_parse_cs_vm,
 	.emit_frame_size =
 		6 + /* vce_v3_0_emit_vm_flush */
 		4 + /* vce_v3_0_emit_pipeline_sync */
@@ -998,8 +999,7 @@ static void vce_v3_0_set_irq_funcs(struct amdgpu_device *adev)
 	adev->vce.irq.funcs = &vce_v3_0_irq_funcs;
 };
 
-const struct amdgpu_ip_block_version vce_v3_0_ip_block =
-{
+const struct amdgpu_ip_block_version vce_v3_0_ip_block = {
 	.type = AMD_IP_BLOCK_TYPE_VCE,
 	.major = 3,
 	.minor = 0,
@@ -1007,8 +1007,7 @@ const struct amdgpu_ip_block_version vce_v3_0_ip_block =
 	.funcs = &vce_v3_0_ip_funcs,
 };
 
-const struct amdgpu_ip_block_version vce_v3_1_ip_block =
-{
+const struct amdgpu_ip_block_version vce_v3_1_ip_block = {
 	.type = AMD_IP_BLOCK_TYPE_VCE,
 	.major = 3,
 	.minor = 1,
@@ -1016,8 +1015,7 @@ const struct amdgpu_ip_block_version vce_v3_1_ip_block =
 	.funcs = &vce_v3_0_ip_funcs,
 };
 
-const struct amdgpu_ip_block_version vce_v3_4_ip_block =
-{
+const struct amdgpu_ip_block_version vce_v3_4_ip_block = {
 	.type = AMD_IP_BLOCK_TYPE_VCE,
 	.major = 3,
 	.minor = 4,
diff --git a/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c b/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c
index d1fc4e0b8265..2d64002bed61 100644
--- a/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c
@@ -83,7 +83,7 @@ static uint64_t vce_v4_0_ring_get_wptr(struct amdgpu_ring *ring)
 	struct amdgpu_device *adev = ring->adev;
 
 	if (ring->use_doorbell)
-		return adev->wb.wb[ring->wptr_offs];
+		return *ring->wptr_cpu_addr;
 
 	if (ring->me == 0)
 		return RREG32(SOC15_REG_OFFSET(VCE, 0, mmVCE_RB_WPTR));
@@ -106,7 +106,7 @@ static void vce_v4_0_ring_set_wptr(struct amdgpu_ring *ring)
 
 	if (ring->use_doorbell) {
 		/* XXX check if swapping is necessary on BE */
-		adev->wb.wb[ring->wptr_offs] = lower_32_bits(ring->wptr);
+		*ring->wptr_cpu_addr = lower_32_bits(ring->wptr);
 		WDOORBELL32(ring->doorbell_index, lower_32_bits(ring->wptr));
 		return;
 	}
@@ -177,7 +177,7 @@ static int vce_v4_0_mmsch_start(struct amdgpu_device *adev,
 	WREG32(SOC15_REG_OFFSET(VCE, 0, mmVCE_MMSCH_VF_MAILBOX_RESP), 0);
 
 	WDOORBELL32(adev->vce.ring[0].doorbell_index, 0);
-	adev->wb.wb[adev->vce.ring[0].wptr_offs] = 0;
+	*adev->vce.ring[0].wptr_cpu_addr = 0;
 	adev->vce.ring[0].wptr = 0;
 	adev->vce.ring[0].wptr_old = 0;
 
@@ -407,9 +407,14 @@ static int vce_v4_0_stop(struct amdgpu_device *adev)
 	return 0;
 }
 
-static int vce_v4_0_early_init(void *handle)
+static int vce_v4_0_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
+	int r;
+
+	r = amdgpu_vce_early_init(adev);
+	if (r)
+		return r;
 
 	if (amdgpu_sriov_vf(adev)) /* currently only VCN0 support SRIOV */
 		adev->vce.num_rings = 1;
@@ -422,9 +427,9 @@ static int vce_v4_0_early_init(void *handle)
 	return 0;
 }
 
-static int vce_v4_0_sw_init(void *handle)
+static int vce_v4_0_sw_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	struct amdgpu_ring *ring;
 
 	unsigned size;
@@ -466,6 +471,7 @@ static int vce_v4_0_sw_init(void *handle)
 		enum amdgpu_ring_priority_level hw_prio = amdgpu_vce_get_ring_prio(i);
 
 		ring = &adev->vce.ring[i];
+		ring->vm_hub = AMDGPU_MMHUB0(0);
 		sprintf(ring->name, "vce%d", i);
 		if (amdgpu_sriov_vf(adev)) {
 			/* DOORBELL only works under SRIOV */
@@ -485,11 +491,6 @@ static int vce_v4_0_sw_init(void *handle)
 			return r;
 	}
 
-
-	r = amdgpu_vce_entity_init(adev);
-	if (r)
-		return r;
-
 	r = amdgpu_virt_alloc_mm_table(adev);
 	if (r)
 		return r;
@@ -497,10 +498,10 @@ static int vce_v4_0_sw_init(void *handle)
 	return r;
 }
 
-static int vce_v4_0_sw_fini(void *handle)
+static int vce_v4_0_sw_fini(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	/* free MM table */
 	amdgpu_virt_free_mm_table(adev);
@@ -517,10 +518,10 @@ static int vce_v4_0_sw_fini(void *handle)
 	return amdgpu_vce_sw_fini(adev);
 }
 
-static int vce_v4_0_hw_init(void *handle)
+static int vce_v4_0_hw_init(struct amdgpu_ip_block *ip_block)
 {
 	int r, i;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (amdgpu_sriov_vf(adev))
 		r = vce_v4_0_sriov_start(adev);
@@ -540,14 +541,14 @@ static int vce_v4_0_hw_init(void *handle)
 	return 0;
 }
 
-static int vce_v4_0_hw_fini(void *handle)
+static int vce_v4_0_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	cancel_delayed_work_sync(&adev->vce.idle_work);
 
 	if (!amdgpu_sriov_vf(adev)) {
-		/* vce_v4_0_wait_for_idle(handle); */
+		/* vce_v4_0_wait_for_idle(ip_block); */
 		vce_v4_0_stop(adev);
 	} else {
 		/* full access mode, so don't touch any VCE register */
@@ -557,9 +558,9 @@ static int vce_v4_0_hw_fini(void *handle)
 	return 0;
 }
 
-static int vce_v4_0_suspend(void *handle)
+static int vce_v4_0_suspend(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int r, idx;
 
 	if (adev->vce.vcpu_bo == NULL)
@@ -598,16 +599,16 @@ static int vce_v4_0_suspend(void *handle)
 						       AMD_CG_STATE_GATE);
 	}
 
-	r = vce_v4_0_hw_fini(adev);
+	r = vce_v4_0_hw_fini(ip_block);
 	if (r)
 		return r;
 
 	return amdgpu_vce_suspend(adev);
 }
 
-static int vce_v4_0_resume(void *handle)
+static int vce_v4_0_resume(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int r, idx;
 
 	if (adev->vce.vcpu_bo == NULL)
@@ -628,7 +629,7 @@ static int vce_v4_0_resume(void *handle)
 			return r;
 	}
 
-	return vce_v4_0_hw_init(adev);
+	return vce_v4_0_hw_init(ip_block);
 }
 
 static void vce_v4_0_mc_resume(struct amdgpu_device *adev)
@@ -688,281 +689,14 @@ static void vce_v4_0_mc_resume(struct amdgpu_device *adev)
 			~VCE_SYS_INT_EN__VCE_SYS_INT_TRAP_INTERRUPT_EN_MASK);
 }
 
-static int vce_v4_0_set_clockgating_state(void *handle,
+static int vce_v4_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_clockgating_state state)
 {
 	/* needed for driver unload*/
 	return 0;
 }
 
-#if 0
-static bool vce_v4_0_is_idle(void *handle)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-	u32 mask = 0;
-
-	mask |= (adev->vce.harvest_config & AMDGPU_VCE_HARVEST_VCE0) ? 0 : SRBM_STATUS2__VCE0_BUSY_MASK;
-	mask |= (adev->vce.harvest_config & AMDGPU_VCE_HARVEST_VCE1) ? 0 : SRBM_STATUS2__VCE1_BUSY_MASK;
-
-	return !(RREG32(mmSRBM_STATUS2) & mask);
-}
-
-static int vce_v4_0_wait_for_idle(void *handle)
-{
-	unsigned i;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	for (i = 0; i < adev->usec_timeout; i++)
-		if (vce_v4_0_is_idle(handle))
-			return 0;
-
-	return -ETIMEDOUT;
-}
-
-#define  VCE_STATUS_VCPU_REPORT_AUTO_BUSY_MASK  0x00000008L   /* AUTO_BUSY */
-#define  VCE_STATUS_VCPU_REPORT_RB0_BUSY_MASK   0x00000010L   /* RB0_BUSY */
-#define  VCE_STATUS_VCPU_REPORT_RB1_BUSY_MASK   0x00000020L   /* RB1_BUSY */
-#define  AMDGPU_VCE_STATUS_BUSY_MASK (VCE_STATUS_VCPU_REPORT_AUTO_BUSY_MASK | \
-				      VCE_STATUS_VCPU_REPORT_RB0_BUSY_MASK)
-
-static bool vce_v4_0_check_soft_reset(void *handle)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-	u32 srbm_soft_reset = 0;
-
-	/* According to VCE team , we should use VCE_STATUS instead
-	 * SRBM_STATUS.VCE_BUSY bit for busy status checking.
-	 * GRBM_GFX_INDEX.INSTANCE_INDEX is used to specify which VCE
-	 * instance's registers are accessed
-	 * (0 for 1st instance, 10 for 2nd instance).
-	 *
-	 *VCE_STATUS
-	 *|UENC|ACPI|AUTO ACTIVE|RB1 |RB0 |RB2 |          |FW_LOADED|JOB |
-	 *|----+----+-----------+----+----+----+----------+---------+----|
-	 *|bit8|bit7|    bit6   |bit5|bit4|bit3|   bit2   |  bit1   |bit0|
-	 *
-	 * VCE team suggest use bit 3--bit 6 for busy status check
-	 */
-	mutex_lock(&adev->grbm_idx_mutex);
-	WREG32_FIELD(GRBM_GFX_INDEX, INSTANCE_INDEX, 0);
-	if (RREG32(SOC15_REG_OFFSET(VCE, 0, mmVCE_STATUS) & AMDGPU_VCE_STATUS_BUSY_MASK) {
-		srbm_soft_reset = REG_SET_FIELD(srbm_soft_reset, SRBM_SOFT_RESET, SOFT_RESET_VCE0, 1);
-		srbm_soft_reset = REG_SET_FIELD(srbm_soft_reset, SRBM_SOFT_RESET, SOFT_RESET_VCE1, 1);
-	}
-	WREG32_FIELD(GRBM_GFX_INDEX, INSTANCE_INDEX, 0x10);
-	if (RREG32(SOC15_REG_OFFSET(VCE, 0, mmVCE_STATUS) & AMDGPU_VCE_STATUS_BUSY_MASK) {
-		srbm_soft_reset = REG_SET_FIELD(srbm_soft_reset, SRBM_SOFT_RESET, SOFT_RESET_VCE0, 1);
-		srbm_soft_reset = REG_SET_FIELD(srbm_soft_reset, SRBM_SOFT_RESET, SOFT_RESET_VCE1, 1);
-	}
-	WREG32_FIELD(GRBM_GFX_INDEX, INSTANCE_INDEX, 0);
-	mutex_unlock(&adev->grbm_idx_mutex);
-
-	if (srbm_soft_reset) {
-		adev->vce.srbm_soft_reset = srbm_soft_reset;
-		return true;
-	} else {
-		adev->vce.srbm_soft_reset = 0;
-		return false;
-	}
-}
-
-static int vce_v4_0_soft_reset(void *handle)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-	u32 srbm_soft_reset;
-
-	if (!adev->vce.srbm_soft_reset)
-		return 0;
-	srbm_soft_reset = adev->vce.srbm_soft_reset;
-
-	if (srbm_soft_reset) {
-		u32 tmp;
-
-		tmp = RREG32(mmSRBM_SOFT_RESET);
-		tmp |= srbm_soft_reset;
-		dev_info(adev->dev, "SRBM_SOFT_RESET=0x%08X\n", tmp);
-		WREG32(mmSRBM_SOFT_RESET, tmp);
-		tmp = RREG32(mmSRBM_SOFT_RESET);
-
-		udelay(50);
-
-		tmp &= ~srbm_soft_reset;
-		WREG32(mmSRBM_SOFT_RESET, tmp);
-		tmp = RREG32(mmSRBM_SOFT_RESET);
-
-		/* Wait a little for things to settle down */
-		udelay(50);
-	}
-
-	return 0;
-}
-
-static int vce_v4_0_pre_soft_reset(void *handle)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	if (!adev->vce.srbm_soft_reset)
-		return 0;
-
-	mdelay(5);
-
-	return vce_v4_0_suspend(adev);
-}
-
-
-static int vce_v4_0_post_soft_reset(void *handle)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	if (!adev->vce.srbm_soft_reset)
-		return 0;
-
-	mdelay(5);
-
-	return vce_v4_0_resume(adev);
-}
-
-static void vce_v4_0_override_vce_clock_gating(struct amdgpu_device *adev, bool override)
-{
-	u32 tmp, data;
-
-	tmp = data = RREG32(SOC15_REG_OFFSET(VCE, 0, mmVCE_RB_ARB_CTRL));
-	if (override)
-		data |= VCE_RB_ARB_CTRL__VCE_CGTT_OVERRIDE_MASK;
-	else
-		data &= ~VCE_RB_ARB_CTRL__VCE_CGTT_OVERRIDE_MASK;
-
-	if (tmp != data)
-		WREG32(SOC15_REG_OFFSET(VCE, 0, mmVCE_RB_ARB_CTRL), data);
-}
-
-static void vce_v4_0_set_vce_sw_clock_gating(struct amdgpu_device *adev,
-					     bool gated)
-{
-	u32 data;
-
-	/* Set Override to disable Clock Gating */
-	vce_v4_0_override_vce_clock_gating(adev, true);
-
-	/* This function enables MGCG which is controlled by firmware.
-	   With the clocks in the gated state the core is still
-	   accessible but the firmware will throttle the clocks on the
-	   fly as necessary.
-	*/
-	if (gated) {
-		data = RREG32(SOC15_REG_OFFSET(VCE, 0, mmVCE_CLOCK_GATING_B));
-		data |= 0x1ff;
-		data &= ~0xef0000;
-		WREG32(SOC15_REG_OFFSET(VCE, 0, mmVCE_CLOCK_GATING_B), data);
-
-		data = RREG32(SOC15_REG_OFFSET(VCE, 0, mmVCE_UENC_CLOCK_GATING));
-		data |= 0x3ff000;
-		data &= ~0xffc00000;
-		WREG32(SOC15_REG_OFFSET(VCE, 0, mmVCE_UENC_CLOCK_GATING), data);
-
-		data = RREG32(SOC15_REG_OFFSET(VCE, 0, mmVCE_UENC_CLOCK_GATING_2));
-		data |= 0x2;
-		data &= ~0x00010000;
-		WREG32(SOC15_REG_OFFSET(VCE, 0, mmVCE_UENC_CLOCK_GATING_2), data);
-
-		data = RREG32(SOC15_REG_OFFSET(VCE, 0, mmVCE_UENC_REG_CLOCK_GATING));
-		data |= 0x37f;
-		WREG32(SOC15_REG_OFFSET(VCE, 0, mmVCE_UENC_REG_CLOCK_GATING), data);
-
-		data = RREG32(SOC15_REG_OFFSET(VCE, 0, mmVCE_UENC_DMA_DCLK_CTRL));
-		data |= VCE_UENC_DMA_DCLK_CTRL__WRDMCLK_FORCEON_MASK |
-			VCE_UENC_DMA_DCLK_CTRL__RDDMCLK_FORCEON_MASK |
-			VCE_UENC_DMA_DCLK_CTRL__REGCLK_FORCEON_MASK  |
-			0x8;
-		WREG32(SOC15_REG_OFFSET(VCE, 0, mmVCE_UENC_DMA_DCLK_CTRL), data);
-	} else {
-		data = RREG32(SOC15_REG_OFFSET(VCE, 0, mmVCE_CLOCK_GATING_B));
-		data &= ~0x80010;
-		data |= 0xe70008;
-		WREG32(SOC15_REG_OFFSET(VCE, 0, mmVCE_CLOCK_GATING_B), data);
-
-		data = RREG32(SOC15_REG_OFFSET(VCE, 0, mmVCE_UENC_CLOCK_GATING));
-		data |= 0xffc00000;
-		WREG32(SOC15_REG_OFFSET(VCE, 0, mmVCE_UENC_CLOCK_GATING), data);
-
-		data = RREG32(SOC15_REG_OFFSET(VCE, 0, mmVCE_UENC_CLOCK_GATING_2));
-		data |= 0x10000;
-		WREG32(SOC15_REG_OFFSET(VCE, 0, mmVCE_UENC_CLOCK_GATING_2), data);
-
-		data = RREG32(SOC15_REG_OFFSET(VCE, 0, mmVCE_UENC_REG_CLOCK_GATING));
-		data &= ~0xffc00000;
-		WREG32(SOC15_REG_OFFSET(VCE, 0, mmVCE_UENC_REG_CLOCK_GATING), data);
-
-		data = RREG32(SOC15_REG_OFFSET(VCE, 0, mmVCE_UENC_DMA_DCLK_CTRL));
-		data &= ~(VCE_UENC_DMA_DCLK_CTRL__WRDMCLK_FORCEON_MASK |
-			  VCE_UENC_DMA_DCLK_CTRL__RDDMCLK_FORCEON_MASK |
-			  VCE_UENC_DMA_DCLK_CTRL__REGCLK_FORCEON_MASK  |
-			  0x8);
-		WREG32(SOC15_REG_OFFSET(VCE, 0, mmVCE_UENC_DMA_DCLK_CTRL), data);
-	}
-	vce_v4_0_override_vce_clock_gating(adev, false);
-}
-
-static void vce_v4_0_set_bypass_mode(struct amdgpu_device *adev, bool enable)
-{
-	u32 tmp = RREG32_SMC(ixGCK_DFS_BYPASS_CNTL);
-
-	if (enable)
-		tmp |= GCK_DFS_BYPASS_CNTL__BYPASSECLK_MASK;
-	else
-		tmp &= ~GCK_DFS_BYPASS_CNTL__BYPASSECLK_MASK;
-
-	WREG32_SMC(ixGCK_DFS_BYPASS_CNTL, tmp);
-}
-
-static int vce_v4_0_set_clockgating_state(void *handle,
-					  enum amd_clockgating_state state)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-	bool enable = (state == AMD_CG_STATE_GATE);
-	int i;
-
-	if ((adev->asic_type == CHIP_POLARIS10) ||
-		(adev->asic_type == CHIP_TONGA) ||
-		(adev->asic_type == CHIP_FIJI))
-		vce_v4_0_set_bypass_mode(adev, enable);
-
-	if (!(adev->cg_flags & AMD_CG_SUPPORT_VCE_MGCG))
-		return 0;
-
-	mutex_lock(&adev->grbm_idx_mutex);
-	for (i = 0; i < 2; i++) {
-		/* Program VCE Instance 0 or 1 if not harvested */
-		if (adev->vce.harvest_config & (1 << i))
-			continue;
-
-		WREG32_FIELD(GRBM_GFX_INDEX, VCE_INSTANCE, i);
-
-		if (enable) {
-			/* initialize VCE_CLOCK_GATING_A: Clock ON/OFF delay */
-			uint32_t data = RREG32(SOC15_REG_OFFSET(VCE, 0, mmVCE_CLOCK_GATING_A);
-			data &= ~(0xf | 0xff0);
-			data |= ((0x0 << 0) | (0x04 << 4));
-			WREG32(SOC15_REG_OFFSET(VCE, 0, mmVCE_CLOCK_GATING_A, data);
-
-			/* initialize VCE_UENC_CLOCK_GATING: Clock ON/OFF delay */
-			data = RREG32(SOC15_REG_OFFSET(VCE, 0, mmVCE_UENC_CLOCK_GATING);
-			data &= ~(0xf | 0xff0);
-			data |= ((0x0 << 0) | (0x04 << 4));
-			WREG32(SOC15_REG_OFFSET(VCE, 0, mmVCE_UENC_CLOCK_GATING, data);
-		}
-
-		vce_v4_0_set_vce_sw_clock_gating(adev, enable);
-	}
-
-	WREG32_FIELD(GRBM_GFX_INDEX, VCE_INSTANCE, 0);
-	mutex_unlock(&adev->grbm_idx_mutex);
-
-	return 0;
-}
-#endif
-
-static int vce_v4_0_set_powergating_state(void *handle,
+static int vce_v4_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_powergating_state state)
 {
 	/* This doesn't actually powergate the VCE block.
@@ -972,7 +706,7 @@ static int vce_v4_0_set_powergating_state(void *handle,
 	 * revisit this when there is a cleaner line between
 	 * the smc and the hw blocks
 	 */
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (state == AMD_PG_STATE_GATE)
 		return vce_v4_0_stop(adev);
@@ -1021,7 +755,7 @@ static void vce_v4_0_emit_reg_wait(struct amdgpu_ring *ring, uint32_t reg,
 static void vce_v4_0_emit_vm_flush(struct amdgpu_ring *ring,
 				   unsigned int vmid, uint64_t pd_addr)
 {
-	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
+	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->vm_hub];
 
 	pd_addr = amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pd_addr);
 
@@ -1080,19 +814,12 @@ static int vce_v4_0_process_interrupt(struct amdgpu_device *adev,
 const struct amd_ip_funcs vce_v4_0_ip_funcs = {
 	.name = "vce_v4_0",
 	.early_init = vce_v4_0_early_init,
-	.late_init = NULL,
 	.sw_init = vce_v4_0_sw_init,
 	.sw_fini = vce_v4_0_sw_fini,
 	.hw_init = vce_v4_0_hw_init,
 	.hw_fini = vce_v4_0_hw_fini,
 	.suspend = vce_v4_0_suspend,
 	.resume = vce_v4_0_resume,
-	.is_idle = NULL /* vce_v4_0_is_idle */,
-	.wait_for_idle = NULL /* vce_v4_0_wait_for_idle */,
-	.check_soft_reset = NULL /* vce_v4_0_check_soft_reset */,
-	.pre_soft_reset = NULL /* vce_v4_0_pre_soft_reset */,
-	.soft_reset = NULL /* vce_v4_0_soft_reset */,
-	.post_soft_reset = NULL /* vce_v4_0_post_soft_reset */,
 	.set_clockgating_state = vce_v4_0_set_clockgating_state,
 	.set_powergating_state = vce_v4_0_set_powergating_state,
 };
@@ -1103,11 +830,10 @@ static const struct amdgpu_ring_funcs vce_v4_0_ring_vm_funcs = {
 	.nop = VCE_CMD_NO_OP,
 	.support_64bit_ptrs = false,
 	.no_user_fence = true,
-	.vmhub = AMDGPU_MMHUB_0,
 	.get_rptr = vce_v4_0_ring_get_rptr,
 	.get_wptr = vce_v4_0_ring_get_wptr,
 	.set_wptr = vce_v4_0_ring_set_wptr,
-	.parse_cs = amdgpu_vce_ring_parse_cs_vm,
+	.patch_cs_in_place = amdgpu_vce_ring_parse_cs_vm,
 	.emit_frame_size =
 		SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 +
 		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 4 +
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_sw_ring.c b/drivers/gpu/drm/amd/amdgpu/vcn_sw_ring.c
new file mode 100644
index 000000000000..2b9ddb3d2fe1
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_sw_ring.c
@@ -0,0 +1,86 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "amdgpu.h"
+#include "vcn_sw_ring.h"
+
+void vcn_dec_sw_ring_emit_fence(struct amdgpu_ring *ring, u64 addr,
+	u64 seq, uint32_t flags)
+{
+	WARN_ON(flags & AMDGPU_FENCE_FLAG_64BIT);
+
+	amdgpu_ring_write(ring, VCN_DEC_SW_CMD_FENCE);
+	amdgpu_ring_write(ring, addr);
+	amdgpu_ring_write(ring, upper_32_bits(addr));
+	amdgpu_ring_write(ring, seq);
+	amdgpu_ring_write(ring, VCN_DEC_SW_CMD_TRAP);
+}
+
+void vcn_dec_sw_ring_insert_end(struct amdgpu_ring *ring)
+{
+	amdgpu_ring_write(ring, VCN_DEC_SW_CMD_END);
+}
+
+void vcn_dec_sw_ring_emit_ib(struct amdgpu_ring *ring, struct amdgpu_job *job,
+	struct amdgpu_ib *ib, uint32_t flags)
+{
+	uint32_t vmid = AMDGPU_JOB_GET_VMID(job);
+
+	amdgpu_ring_write(ring, VCN_DEC_SW_CMD_IB);
+	amdgpu_ring_write(ring, vmid);
+	amdgpu_ring_write(ring, lower_32_bits(ib->gpu_addr));
+	amdgpu_ring_write(ring, upper_32_bits(ib->gpu_addr));
+	amdgpu_ring_write(ring, ib->length_dw);
+}
+
+void vcn_dec_sw_ring_emit_reg_wait(struct amdgpu_ring *ring, uint32_t reg,
+	uint32_t val, uint32_t mask)
+{
+	amdgpu_ring_write(ring, VCN_DEC_SW_CMD_REG_WAIT);
+	amdgpu_ring_write(ring, reg << 2);
+	amdgpu_ring_write(ring, mask);
+	amdgpu_ring_write(ring, val);
+}
+
+void vcn_dec_sw_ring_emit_vm_flush(struct amdgpu_ring *ring,
+	uint32_t vmid, uint64_t pd_addr)
+{
+	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->vm_hub];
+	uint32_t data0, data1, mask;
+
+	pd_addr = amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pd_addr);
+
+	/* wait for register write */
+	data0 = hub->ctx0_ptb_addr_lo32 + vmid * hub->ctx_addr_distance;
+	data1 = lower_32_bits(pd_addr);
+	mask = 0xffffffff;
+	vcn_dec_sw_ring_emit_reg_wait(ring, data0, data1, mask);
+}
+
+void vcn_dec_sw_ring_emit_wreg(struct amdgpu_ring *ring, uint32_t reg,
+	uint32_t val)
+{
+	amdgpu_ring_write(ring, VCN_DEC_SW_CMD_REG_WRITE);
+	amdgpu_ring_write(ring,	reg << 2);
+	amdgpu_ring_write(ring, val);
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_sw_ring.h b/drivers/gpu/drm/amd/amdgpu/vcn_sw_ring.h
new file mode 100644
index 000000000000..7e775725f120
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_sw_ring.h
@@ -0,0 +1,44 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __VCN_SW_RING_H__
+#define __VCN_SW_RING_H__
+
+#define VCN_SW_RING_EMIT_FRAME_SIZE \
+		(4 + /* vcn_dec_sw_ring_emit_vm_flush */ \
+		5 + 5 + /* vcn_dec_sw_ring_emit_fence x2 vm fence */ \
+		1) /* vcn_dec_sw_ring_insert_end */
+
+void vcn_dec_sw_ring_emit_fence(struct amdgpu_ring *ring, u64 addr,
+	u64 seq, uint32_t flags);
+void vcn_dec_sw_ring_insert_end(struct amdgpu_ring *ring);
+void vcn_dec_sw_ring_emit_ib(struct amdgpu_ring *ring, struct amdgpu_job *job,
+	struct amdgpu_ib *ib, uint32_t flags);
+void vcn_dec_sw_ring_emit_reg_wait(struct amdgpu_ring *ring, uint32_t reg,
+	uint32_t val, uint32_t mask);
+void vcn_dec_sw_ring_emit_vm_flush(struct amdgpu_ring *ring,
+	uint32_t vmid, uint64_t pd_addr);
+void vcn_dec_sw_ring_emit_wreg(struct amdgpu_ring *ring, uint32_t reg,
+	uint32_t val);
+
+#endif /* __VCN_SW_RING_H__ */
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c
index d54d720b3cf6..a316797875a8 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c
@@ -24,6 +24,7 @@
 #include <linux/firmware.h>
 
 #include "amdgpu.h"
+#include "amdgpu_cs.h"
 #include "amdgpu_vcn.h"
 #include "amdgpu_pm.h"
 #include "soc15.h"
@@ -44,51 +45,92 @@
 #define mmUVD_REG_XX_MASK_1_0			0x05ac
 #define mmUVD_REG_XX_MASK_1_0_BASE_IDX		1
 
-static int vcn_v1_0_stop(struct amdgpu_device *adev);
+static const struct amdgpu_hwip_reg_entry vcn_reg_list_1_0[] = {
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_POWER_STATUS),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_STATUS),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_CONTEXT_ID),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_CONTEXT_ID2),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_GPCOM_VCPU_DATA0),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_GPCOM_VCPU_DATA1),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_GPCOM_VCPU_CMD),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_BASE_HI),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_BASE_LO),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_BASE_HI2),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_BASE_LO2),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_BASE_HI3),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_BASE_LO3),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_BASE_HI4),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_BASE_LO4),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_RPTR),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_WPTR),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_RPTR2),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_WPTR2),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_RPTR3),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_WPTR3),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_RPTR4),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_WPTR4),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_SIZE),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_SIZE2),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_SIZE3),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_SIZE4),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_PGFSM_CONFIG),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_PGFSM_STATUS),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_DPG_LMA_CTL),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_DPG_LMA_DATA),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_DPG_LMA_MASK),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_DPG_PAUSE)
+};
+
+static int vcn_v1_0_stop(struct amdgpu_vcn_inst *vinst);
 static void vcn_v1_0_set_dec_ring_funcs(struct amdgpu_device *adev);
 static void vcn_v1_0_set_enc_ring_funcs(struct amdgpu_device *adev);
 static void vcn_v1_0_set_irq_funcs(struct amdgpu_device *adev);
-static int vcn_v1_0_set_powergating_state(void *handle, enum amd_powergating_state state);
-static int vcn_v1_0_pause_dpg_mode(struct amdgpu_device *adev,
-				int inst_idx, struct dpg_pause_state *new_state);
+static int vcn_v1_0_set_pg_state(struct amdgpu_vcn_inst *vinst,
+				 enum amd_powergating_state state);
+static int vcn_v1_0_pause_dpg_mode(struct amdgpu_vcn_inst *vinst,
+				   struct dpg_pause_state *new_state);
 
 static void vcn_v1_0_idle_work_handler(struct work_struct *work);
 static void vcn_v1_0_ring_begin_use(struct amdgpu_ring *ring);
 
 /**
- * vcn_v1_0_early_init - set function pointers
+ * vcn_v1_0_early_init - set function pointers and load microcode
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Set ring and irq function pointers
+ * Load microcode from filesystem
  */
-static int vcn_v1_0_early_init(void *handle)
+static int vcn_v1_0_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
-	adev->vcn.num_enc_rings = 2;
+	adev->vcn.inst[0].num_enc_rings = 2;
+	adev->vcn.inst[0].set_pg_state = vcn_v1_0_set_pg_state;
 
 	vcn_v1_0_set_dec_ring_funcs(adev);
 	vcn_v1_0_set_enc_ring_funcs(adev);
 	vcn_v1_0_set_irq_funcs(adev);
 
-	jpeg_v1_0_early_init(handle);
+	jpeg_v1_0_early_init(ip_block);
 
-	return 0;
+	return amdgpu_vcn_early_init(adev, 0);
 }
 
 /**
  * vcn_v1_0_sw_init - sw init for VCN block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Load firmware and sw initialization
  */
-static int vcn_v1_0_sw_init(void *handle)
+static int vcn_v1_0_sw_init(struct amdgpu_ip_block *ip_block)
 {
 	struct amdgpu_ring *ring;
 	int i, r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	uint32_t reg_count = ARRAY_SIZE(vcn_reg_list_1_0);
+	uint32_t *ptr;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	/* VCN DEC TRAP */
 	r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_VCN,
@@ -97,48 +139,50 @@ static int vcn_v1_0_sw_init(void *handle)
 		return r;
 
 	/* VCN ENC TRAP */
-	for (i = 0; i < adev->vcn.num_enc_rings; ++i) {
+	for (i = 0; i < adev->vcn.inst[0].num_enc_rings; ++i) {
 		r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_VCN, i + VCN_1_0__SRCID__UVD_ENC_GENERAL_PURPOSE,
 					&adev->vcn.inst->irq);
 		if (r)
 			return r;
 	}
 
-	r = amdgpu_vcn_sw_init(adev);
+	r = amdgpu_vcn_sw_init(adev, 0);
 	if (r)
 		return r;
 
 	/* Override the work func */
-	adev->vcn.idle_work.work.func = vcn_v1_0_idle_work_handler;
+	adev->vcn.inst[0].idle_work.work.func = vcn_v1_0_idle_work_handler;
 
-	amdgpu_vcn_setup_ucode(adev);
+	amdgpu_vcn_setup_ucode(adev, 0);
 
-	r = amdgpu_vcn_resume(adev);
+	r = amdgpu_vcn_resume(adev, 0);
 	if (r)
 		return r;
 
 	ring = &adev->vcn.inst->ring_dec;
+	ring->vm_hub = AMDGPU_MMHUB0(0);
 	sprintf(ring->name, "vcn_dec");
 	r = amdgpu_ring_init(adev, ring, 512, &adev->vcn.inst->irq, 0,
 			     AMDGPU_RING_PRIO_DEFAULT, NULL);
 	if (r)
 		return r;
 
-	adev->vcn.internal.scratch9 = adev->vcn.inst->external.scratch9 =
+	adev->vcn.inst[0].internal.scratch9 = adev->vcn.inst->external.scratch9 =
 		SOC15_REG_OFFSET(UVD, 0, mmUVD_SCRATCH9);
-	adev->vcn.internal.data0 = adev->vcn.inst->external.data0 =
+	adev->vcn.inst[0].internal.data0 = adev->vcn.inst->external.data0 =
 		SOC15_REG_OFFSET(UVD, 0, mmUVD_GPCOM_VCPU_DATA0);
-	adev->vcn.internal.data1 = adev->vcn.inst->external.data1 =
+	adev->vcn.inst[0].internal.data1 = adev->vcn.inst->external.data1 =
 		SOC15_REG_OFFSET(UVD, 0, mmUVD_GPCOM_VCPU_DATA1);
-	adev->vcn.internal.cmd = adev->vcn.inst->external.cmd =
+	adev->vcn.inst[0].internal.cmd = adev->vcn.inst->external.cmd =
 		SOC15_REG_OFFSET(UVD, 0, mmUVD_GPCOM_VCPU_CMD);
-	adev->vcn.internal.nop = adev->vcn.inst->external.nop =
+	adev->vcn.inst[0].internal.nop = adev->vcn.inst->external.nop =
 		SOC15_REG_OFFSET(UVD, 0, mmUVD_NO_OP);
 
-	for (i = 0; i < adev->vcn.num_enc_rings; ++i) {
+	for (i = 0; i < adev->vcn.inst[0].num_enc_rings; ++i) {
 		enum amdgpu_ring_priority_level hw_prio = amdgpu_vcn_get_enc_ring_prio(i);
 
 		ring = &adev->vcn.inst->ring_enc[i];
+		ring->vm_hub = AMDGPU_MMHUB0(0);
 		sprintf(ring->name, "vcn_enc%d", i);
 		r = amdgpu_ring_init(adev, ring, 512, &adev->vcn.inst->irq, 0,
 				     hw_prio, NULL);
@@ -146,69 +190,79 @@ static int vcn_v1_0_sw_init(void *handle)
 			return r;
 	}
 
-	adev->vcn.pause_dpg_mode = vcn_v1_0_pause_dpg_mode;
+	adev->vcn.inst[0].pause_dpg_mode = vcn_v1_0_pause_dpg_mode;
+
+	if (amdgpu_vcnfw_log) {
+		struct amdgpu_fw_shared *fw_shared = adev->vcn.inst->fw_shared.cpu_addr;
+
+		fw_shared->present_flag_0 = 0;
+		amdgpu_vcn_fwlog_init(adev->vcn.inst);
+	}
 
-	r = jpeg_v1_0_sw_init(handle);
+	r = jpeg_v1_0_sw_init(ip_block);
 
+	/* Allocate memory for VCN IP Dump buffer */
+	ptr = kcalloc(adev->vcn.num_vcn_inst * reg_count, sizeof(uint32_t), GFP_KERNEL);
+	if (!ptr) {
+		DRM_ERROR("Failed to allocate memory for VCN IP Dump\n");
+		adev->vcn.ip_dump = NULL;
+	} else {
+		adev->vcn.ip_dump = ptr;
+	}
 	return r;
 }
 
 /**
  * vcn_v1_0_sw_fini - sw fini for VCN block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * VCN suspend and free up sw allocation
  */
-static int vcn_v1_0_sw_fini(void *handle)
+static int vcn_v1_0_sw_fini(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
-	r = amdgpu_vcn_suspend(adev);
+	r = amdgpu_vcn_suspend(adev, 0);
 	if (r)
 		return r;
 
-	jpeg_v1_0_sw_fini(handle);
+	jpeg_v1_0_sw_fini(ip_block);
 
-	r = amdgpu_vcn_sw_fini(adev);
+	amdgpu_vcn_sw_fini(adev, 0);
 
-	return r;
+	kfree(adev->vcn.ip_dump);
+
+	return 0;
 }
 
 /**
  * vcn_v1_0_hw_init - start and test VCN block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Initialize the hardware, boot up the VCPU and do some testing
  */
-static int vcn_v1_0_hw_init(void *handle)
+static int vcn_v1_0_hw_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	struct amdgpu_ring *ring = &adev->vcn.inst->ring_dec;
 	int i, r;
 
 	r = amdgpu_ring_test_helper(ring);
 	if (r)
-		goto done;
+		return r;
 
-	for (i = 0; i < adev->vcn.num_enc_rings; ++i) {
+	for (i = 0; i < adev->vcn.inst[0].num_enc_rings; ++i) {
 		ring = &adev->vcn.inst->ring_enc[i];
 		r = amdgpu_ring_test_helper(ring);
 		if (r)
-			goto done;
+			return r;
 	}
 
-	ring = &adev->jpeg.inst->ring_dec;
+	ring = adev->jpeg.inst->ring_dec;
 	r = amdgpu_ring_test_helper(ring);
-	if (r)
-		goto done;
-
-done:
-	if (!r)
-		DRM_INFO("VCN decode and encode initialized successfully(under %s).\n",
-			(adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)?"DPG Mode":"SPG Mode");
 
 	return r;
 }
@@ -216,20 +270,21 @@ done:
 /**
  * vcn_v1_0_hw_fini - stop the hardware block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Stop the VCN block, mark ring as not ready any more
  */
-static int vcn_v1_0_hw_fini(void *handle)
+static int vcn_v1_0_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
+	struct amdgpu_vcn_inst *vinst = adev->vcn.inst;
 
-	cancel_delayed_work_sync(&adev->vcn.idle_work);
+	cancel_delayed_work_sync(&vinst->idle_work);
 
 	if ((adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) ||
-		(adev->vcn.cur_state != AMD_PG_STATE_GATE &&
-		 RREG32_SOC15(VCN, 0, mmUVD_STATUS))) {
-		vcn_v1_0_set_powergating_state(adev, AMD_PG_STATE_GATE);
+	    (vinst->cur_state != AMD_PG_STATE_GATE &&
+	     RREG32_SOC15(VCN, 0, mmUVD_STATUS))) {
+		vinst->set_pg_state(vinst, AMD_PG_STATE_GATE);
 	}
 
 	return 0;
@@ -238,20 +293,27 @@ static int vcn_v1_0_hw_fini(void *handle)
 /**
  * vcn_v1_0_suspend - suspend VCN block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * HW fini and suspend VCN block
  */
-static int vcn_v1_0_suspend(void *handle)
+static int vcn_v1_0_suspend(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
+	bool idle_work_unexecuted;
 
-	r = vcn_v1_0_hw_fini(adev);
+	idle_work_unexecuted = cancel_delayed_work_sync(&adev->vcn.inst[0].idle_work);
+	if (idle_work_unexecuted) {
+		if (adev->pm.dpm_enabled)
+			amdgpu_dpm_enable_vcn(adev, false, 0);
+	}
+
+	r = vcn_v1_0_hw_fini(ip_block);
 	if (r)
 		return r;
 
-	r = amdgpu_vcn_suspend(adev);
+	r = amdgpu_vcn_suspend(adev, 0);
 
 	return r;
 }
@@ -259,20 +321,19 @@ static int vcn_v1_0_suspend(void *handle)
 /**
  * vcn_v1_0_resume - resume VCN block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Resume firmware and hw init VCN block
  */
-static int vcn_v1_0_resume(void *handle)
+static int vcn_v1_0_resume(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
-	r = amdgpu_vcn_resume(adev);
+	r = amdgpu_vcn_resume(ip_block->adev, 0);
 	if (r)
 		return r;
 
-	r = vcn_v1_0_hw_init(adev);
+	r = vcn_v1_0_hw_init(ip_block);
 
 	return r;
 }
@@ -280,13 +341,14 @@ static int vcn_v1_0_resume(void *handle)
 /**
  * vcn_v1_0_mc_resume_spg_mode - memory controller programming
  *
- * @adev: amdgpu_device pointer
+ * @vinst: VCN instance
  *
  * Let the VCN memory controller know it's offsets
  */
-static void vcn_v1_0_mc_resume_spg_mode(struct amdgpu_device *adev)
+static void vcn_v1_0_mc_resume_spg_mode(struct amdgpu_vcn_inst *vinst)
 {
-	uint32_t size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.fw->size + 4);
+	struct amdgpu_device *adev = vinst->adev;
+	uint32_t size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.inst[0].fw->size + 4);
 	uint32_t offset;
 
 	/* cache window 0: fw */
@@ -351,9 +413,10 @@ static void vcn_v1_0_mc_resume_spg_mode(struct amdgpu_device *adev)
 			adev->gfx.config.gb_addr_config);
 }
 
-static void vcn_v1_0_mc_resume_dpg_mode(struct amdgpu_device *adev)
+static void vcn_v1_0_mc_resume_dpg_mode(struct amdgpu_vcn_inst *vinst)
 {
-	uint32_t size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.fw->size + 4);
+	struct amdgpu_device *adev = vinst->adev;
+	uint32_t size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.inst[0].fw->size + 4);
 	uint32_t offset;
 
 	/* cache window 0: fw */
@@ -426,12 +489,13 @@ static void vcn_v1_0_mc_resume_dpg_mode(struct amdgpu_device *adev)
 /**
  * vcn_v1_0_disable_clock_gating - disable VCN clock gating
  *
- * @adev: amdgpu_device pointer
+ * @vinst: VCN instance
  *
  * Disable clock gating for VCN block
  */
-static void vcn_v1_0_disable_clock_gating(struct amdgpu_device *adev)
+static void vcn_v1_0_disable_clock_gating(struct amdgpu_vcn_inst *vinst)
 {
+	struct amdgpu_device *adev = vinst->adev;
 	uint32_t data;
 
 	/* JPEG disable CGC */
@@ -455,7 +519,7 @@ static void vcn_v1_0_disable_clock_gating(struct amdgpu_device *adev)
 	if (adev->cg_flags & AMD_CG_SUPPORT_VCN_MGCG)
 		data |= 1 << UVD_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
 	else
-		data &= ~ UVD_CGC_CTRL__DYN_CLOCK_MODE_MASK;
+		data &= ~UVD_CGC_CTRL__DYN_CLOCK_MODE_MASK;
 
 	data |= 1 << UVD_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT;
 	data |= 4 << UVD_CGC_CTRL__CLK_OFF_DELAY__SHIFT;
@@ -552,12 +616,13 @@ static void vcn_v1_0_disable_clock_gating(struct amdgpu_device *adev)
 /**
  * vcn_v1_0_enable_clock_gating - enable VCN clock gating
  *
- * @adev: amdgpu_device pointer
+ * @vinst: Pointer to the VCN instance structure
  *
  * Enable clock gating for VCN block
  */
-static void vcn_v1_0_enable_clock_gating(struct amdgpu_device *adev)
+static void vcn_v1_0_enable_clock_gating(struct amdgpu_vcn_inst *vinst)
 {
+	struct amdgpu_device *adev = vinst->adev;
 	uint32_t data = 0;
 
 	/* enable JPEG CGC */
@@ -621,8 +686,10 @@ static void vcn_v1_0_enable_clock_gating(struct amdgpu_device *adev)
 	WREG32_SOC15(VCN, 0, mmUVD_SUVD_CGC_CTRL, data);
 }
 
-static void vcn_v1_0_clock_gating_dpg_mode(struct amdgpu_device *adev, uint8_t sram_sel)
+static void vcn_v1_0_clock_gating_dpg_mode(struct amdgpu_vcn_inst *vinst,
+					   uint8_t sram_sel)
 {
+	struct amdgpu_device *adev = vinst->adev;
 	uint32_t reg_data = 0;
 
 	/* disable JPEG CGC */
@@ -675,8 +742,9 @@ static void vcn_v1_0_clock_gating_dpg_mode(struct amdgpu_device *adev, uint8_t s
 	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_SUVD_CGC_CTRL, 0, 0xFFFFFFFF, sram_sel);
 }
 
-static void vcn_1_0_disable_static_power_gating(struct amdgpu_device *adev)
+static void vcn_1_0_disable_static_power_gating(struct amdgpu_vcn_inst *vinst)
 {
+	struct amdgpu_device *adev = vinst->adev;
 	uint32_t data = 0;
 
 	if (adev->pg_flags & AMD_PG_SUPPORT_VCN) {
@@ -720,8 +788,9 @@ static void vcn_1_0_disable_static_power_gating(struct amdgpu_device *adev)
 	WREG32_SOC15(VCN, 0, mmUVD_POWER_STATUS, data);
 }
 
-static void vcn_1_0_enable_static_power_gating(struct amdgpu_device *adev)
+static void vcn_1_0_enable_static_power_gating(struct amdgpu_vcn_inst *vinst)
 {
+	struct amdgpu_device *adev = vinst->adev;
 	uint32_t data = 0;
 
 	if (adev->pg_flags & AMD_PG_SUPPORT_VCN) {
@@ -764,12 +833,13 @@ static void vcn_1_0_enable_static_power_gating(struct amdgpu_device *adev)
 /**
  * vcn_v1_0_start_spg_mode - start VCN block
  *
- * @adev: amdgpu_device pointer
+ * @vinst: VCN instance
  *
  * Setup and start the VCN block
  */
-static int vcn_v1_0_start_spg_mode(struct amdgpu_device *adev)
+static int vcn_v1_0_start_spg_mode(struct amdgpu_vcn_inst *vinst)
 {
+	struct amdgpu_device *adev = vinst->adev;
 	struct amdgpu_ring *ring = &adev->vcn.inst->ring_dec;
 	uint32_t rb_bufsz, tmp;
 	uint32_t lmi_swap_cntl;
@@ -778,13 +848,13 @@ static int vcn_v1_0_start_spg_mode(struct amdgpu_device *adev)
 	/* disable byte swapping */
 	lmi_swap_cntl = 0;
 
-	vcn_1_0_disable_static_power_gating(adev);
+	vcn_1_0_disable_static_power_gating(vinst);
 
 	tmp = RREG32_SOC15(UVD, 0, mmUVD_STATUS) | UVD_STATUS__UVD_BUSY;
 	WREG32_SOC15(UVD, 0, mmUVD_STATUS, tmp);
 
 	/* disable clock gating */
-	vcn_v1_0_disable_clock_gating(adev);
+	vcn_v1_0_disable_clock_gating(vinst);
 
 	/* disable interupt */
 	WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_MASTINT_EN), 0,
@@ -826,7 +896,7 @@ static int vcn_v1_0_start_spg_mode(struct amdgpu_device *adev)
 		(0x1 << UVD_MPC_SET_MUX__SET_1__SHIFT) |
 		(0x2 << UVD_MPC_SET_MUX__SET_2__SHIFT)));
 
-	vcn_v1_0_mc_resume_spg_mode(adev);
+	vcn_v1_0_mc_resume_spg_mode(vinst);
 
 	WREG32_SOC15(UVD, 0, mmUVD_REG_XX_MASK_1_0, 0x10);
 	WREG32_SOC15(UVD, 0, mmUVD_RBC_XX_IB_REG_CHECK_1_0,
@@ -939,11 +1009,17 @@ static int vcn_v1_0_start_spg_mode(struct amdgpu_device *adev)
 
 	jpeg_v1_0_start(adev, 0);
 
+	/* Keeping one read-back to ensure all register writes are done,
+	 * otherwise it may introduce race conditions.
+	 */
+	RREG32_SOC15(UVD, 0, mmUVD_STATUS);
+
 	return 0;
 }
 
-static int vcn_v1_0_start_dpg_mode(struct amdgpu_device *adev)
+static int vcn_v1_0_start_dpg_mode(struct amdgpu_vcn_inst *vinst)
 {
+	struct amdgpu_device *adev = vinst->adev;
 	struct amdgpu_ring *ring = &adev->vcn.inst->ring_dec;
 	uint32_t rb_bufsz, tmp;
 	uint32_t lmi_swap_cntl;
@@ -951,7 +1027,7 @@ static int vcn_v1_0_start_dpg_mode(struct amdgpu_device *adev)
 	/* disable byte swapping */
 	lmi_swap_cntl = 0;
 
-	vcn_1_0_enable_static_power_gating(adev);
+	vcn_1_0_enable_static_power_gating(vinst);
 
 	/* enable dynamic power gating mode */
 	tmp = RREG32_SOC15(UVD, 0, mmUVD_POWER_STATUS);
@@ -960,7 +1036,7 @@ static int vcn_v1_0_start_dpg_mode(struct amdgpu_device *adev)
 	WREG32_SOC15(UVD, 0, mmUVD_POWER_STATUS, tmp);
 
 	/* enable clock gating */
-	vcn_v1_0_clock_gating_dpg_mode(adev, 0);
+	vcn_v1_0_clock_gating_dpg_mode(vinst, 0);
 
 	/* enable VCPU clock */
 	tmp = (0xFF << UVD_VCPU_CNTL__PRB_TIMEOUT_VAL__SHIFT);
@@ -1009,7 +1085,7 @@ static int vcn_v1_0_start_dpg_mode(struct amdgpu_device *adev)
 		 (0x1 << UVD_MPC_SET_MUX__SET_1__SHIFT) |
 		 (0x2 << UVD_MPC_SET_MUX__SET_2__SHIFT)), 0xFFFFFFFF, 0);
 
-	vcn_v1_0_mc_resume_dpg_mode(adev);
+	vcn_v1_0_mc_resume_dpg_mode(vinst);
 
 	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_REG_XX_MASK, 0x10, 0xFFFFFFFF, 0);
 	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_RBC_XX_IB_REG_CHECK, 0x3, 0xFFFFFFFF, 0);
@@ -1026,7 +1102,7 @@ static int vcn_v1_0_start_dpg_mode(struct amdgpu_device *adev)
 	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_MASTINT_EN,
 			UVD_MASTINT_EN__VCPU_EN_MASK, UVD_MASTINT_EN__VCPU_EN_MASK, 0);
 
-	vcn_v1_0_clock_gating_dpg_mode(adev, 1);
+	vcn_v1_0_clock_gating_dpg_mode(vinst, 1);
 	/* setup mmUVD_LMI_CTRL */
 	WREG32_SOC15_DPG_MODE_1_0(UVD, 0, mmUVD_LMI_CTRL,
 		(8 << UVD_LMI_CTRL__WRITE_CLEAN_TIMER__SHIFT) |
@@ -1083,29 +1159,32 @@ static int vcn_v1_0_start_dpg_mode(struct amdgpu_device *adev)
 
 	jpeg_v1_0_start(adev, 1);
 
+	/* Keeping one read-back to ensure all register writes are done,
+	 * otherwise it may introduce race conditions.
+	 */
+	RREG32_SOC15(UVD, 0, mmUVD_STATUS);
+
 	return 0;
 }
 
-static int vcn_v1_0_start(struct amdgpu_device *adev)
+static int vcn_v1_0_start(struct amdgpu_vcn_inst *vinst)
 {
-	int r;
+	struct amdgpu_device *adev = vinst->adev;
 
-	if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)
-		r = vcn_v1_0_start_dpg_mode(adev);
-	else
-		r = vcn_v1_0_start_spg_mode(adev);
-	return r;
+	return (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) ?
+		vcn_v1_0_start_dpg_mode(vinst) : vcn_v1_0_start_spg_mode(vinst);
 }
 
 /**
  * vcn_v1_0_stop_spg_mode - stop VCN block
  *
- * @adev: amdgpu_device pointer
+ * @vinst: VCN instance
  *
  * stop the VCN block
  */
-static int vcn_v1_0_stop_spg_mode(struct amdgpu_device *adev)
+static int vcn_v1_0_stop_spg_mode(struct amdgpu_vcn_inst *vinst)
 {
+	struct amdgpu_device *adev = vinst->adev;
 	int tmp;
 
 	SOC15_WAIT_ON_RREG(UVD, 0, mmUVD_STATUS, UVD_STATUS__IDLE, 0x7);
@@ -1145,13 +1224,20 @@ static int vcn_v1_0_stop_spg_mode(struct amdgpu_device *adev)
 
 	WREG32_SOC15(UVD, 0, mmUVD_STATUS, 0);
 
-	vcn_v1_0_enable_clock_gating(adev);
-	vcn_1_0_enable_static_power_gating(adev);
+	vcn_v1_0_enable_clock_gating(vinst);
+	vcn_1_0_enable_static_power_gating(vinst);
+
+	/* Keeping one read-back to ensure all register writes are done,
+	 * otherwise it may introduce race conditions.
+	 */
+	RREG32_SOC15(UVD, 0, mmUVD_STATUS);
+
 	return 0;
 }
 
-static int vcn_v1_0_stop_dpg_mode(struct amdgpu_device *adev)
+static int vcn_v1_0_stop_dpg_mode(struct amdgpu_vcn_inst *vinst)
 {
+	struct amdgpu_device *adev = vinst->adev;
 	uint32_t tmp;
 
 	/* Wait for power status to be UVD_POWER_STATUS__UVD_POWER_STATUS_TILES_OFF */
@@ -1180,24 +1266,32 @@ static int vcn_v1_0_stop_dpg_mode(struct amdgpu_device *adev)
 	WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_POWER_STATUS), 0,
 			~UVD_POWER_STATUS__UVD_PG_MODE_MASK);
 
+	/* Keeping one read-back to ensure all register writes are done,
+	 * otherwise it may introduce race conditions.
+	 */
+	RREG32_SOC15(UVD, 0, mmUVD_STATUS);
+
 	return 0;
 }
 
-static int vcn_v1_0_stop(struct amdgpu_device *adev)
+static int vcn_v1_0_stop(struct amdgpu_vcn_inst *vinst)
 {
+	struct amdgpu_device *adev = vinst->adev;
 	int r;
 
 	if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)
-		r = vcn_v1_0_stop_dpg_mode(adev);
+		r = vcn_v1_0_stop_dpg_mode(vinst);
 	else
-		r = vcn_v1_0_stop_spg_mode(adev);
+		r = vcn_v1_0_stop_spg_mode(vinst);
 
 	return r;
 }
 
-static int vcn_v1_0_pause_dpg_mode(struct amdgpu_device *adev,
-				int inst_idx, struct dpg_pause_state *new_state)
+static int vcn_v1_0_pause_dpg_mode(struct amdgpu_vcn_inst *vinst,
+				   struct dpg_pause_state *new_state)
 {
+	struct amdgpu_device *adev = vinst->adev;
+	int inst_idx = vinst->inst;
 	int ret_code;
 	uint32_t reg_data = 0;
 	uint32_t reg_data2 = 0;
@@ -1244,7 +1338,6 @@ static int vcn_v1_0_pause_dpg_mode(struct amdgpu_device *adev,
 				WREG32_SOC15(UVD, 0, mmUVD_RB_RPTR2, lower_32_bits(ring->wptr));
 				WREG32_SOC15(UVD, 0, mmUVD_RB_WPTR2, lower_32_bits(ring->wptr));
 
-				ring = &adev->vcn.inst->ring_dec;
 				WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_WPTR,
 						   RREG32_SOC15(UVD, 0, mmUVD_SCRATCH2) & 0x7FFFFFFF);
 				SOC15_WAIT_ON_RREG(UVD, 0, mmUVD_POWER_STATUS,
@@ -1291,7 +1384,7 @@ static int vcn_v1_0_pause_dpg_mode(struct amdgpu_device *adev,
 							UVD_DPG_PAUSE__JPEG_PAUSE_DPG_ACK_MASK);
 
 				/* Restore */
-				ring = &adev->jpeg.inst->ring_dec;
+				ring = adev->jpeg.inst->ring_dec;
 				WREG32_SOC15(UVD, 0, mmUVD_LMI_JRBC_RB_VMID, 0);
 				WREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_CNTL,
 							UVD_JRBC_RB_CNTL__RB_NO_FETCH_MASK |
@@ -1305,7 +1398,6 @@ static int vcn_v1_0_pause_dpg_mode(struct amdgpu_device *adev,
 				WREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_CNTL,
 							UVD_JRBC_RB_CNTL__RB_RPTR_WR_EN_MASK);
 
-				ring = &adev->vcn.inst->ring_dec;
 				WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_WPTR,
 						   RREG32_SOC15(UVD, 0, mmUVD_SCRATCH2) & 0x7FFFFFFF);
 				SOC15_WAIT_ON_RREG(UVD, 0, mmUVD_POWER_STATUS,
@@ -1323,16 +1415,16 @@ static int vcn_v1_0_pause_dpg_mode(struct amdgpu_device *adev,
 	return 0;
 }
 
-static bool vcn_v1_0_is_idle(void *handle)
+static bool vcn_v1_0_is_idle(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	return (RREG32_SOC15(VCN, 0, mmUVD_STATUS) == UVD_STATUS__IDLE);
 }
 
-static int vcn_v1_0_wait_for_idle(void *handle)
+static int vcn_v1_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int ret;
 
 	ret = SOC15_WAIT_ON_RREG(VCN, 0, mmUVD_STATUS, UVD_STATUS__IDLE,
@@ -1341,20 +1433,21 @@ static int vcn_v1_0_wait_for_idle(void *handle)
 	return ret;
 }
 
-static int vcn_v1_0_set_clockgating_state(void *handle,
+static int vcn_v1_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_clockgating_state state)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
+	struct amdgpu_vcn_inst *vinst = adev->vcn.inst;
 	bool enable = (state == AMD_CG_STATE_GATE);
 
 	if (enable) {
 		/* wait for STATUS to clear */
-		if (!vcn_v1_0_is_idle(handle))
+		if (!vcn_v1_0_is_idle(ip_block))
 			return -EBUSY;
-		vcn_v1_0_enable_clock_gating(adev);
+		vcn_v1_0_enable_clock_gating(vinst);
 	} else {
 		/* disable HW gating and enable Sw gating */
-		vcn_v1_0_disable_clock_gating(adev);
+		vcn_v1_0_disable_clock_gating(vinst);
 	}
 	return 0;
 }
@@ -1537,7 +1630,7 @@ static void vcn_v1_0_dec_ring_emit_reg_wait(struct amdgpu_ring *ring,
 static void vcn_v1_0_dec_ring_emit_vm_flush(struct amdgpu_ring *ring,
 					    unsigned vmid, uint64_t pd_addr)
 {
-	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
+	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->vm_hub];
 	uint32_t data0, data1, mask;
 
 	pd_addr = amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pd_addr);
@@ -1682,7 +1775,7 @@ static void vcn_v1_0_enc_ring_emit_reg_wait(struct amdgpu_ring *ring,
 static void vcn_v1_0_enc_ring_emit_vm_flush(struct amdgpu_ring *ring,
 					    unsigned int vmid, uint64_t pd_addr)
 {
-	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
+	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->vm_hub];
 
 	pd_addr = amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pd_addr);
 
@@ -1746,8 +1839,8 @@ static void vcn_v1_0_dec_ring_insert_nop(struct amdgpu_ring *ring, uint32_t coun
 	}
 }
 
-static int vcn_v1_0_set_powergating_state(void *handle,
-					  enum amd_powergating_state state)
+static int vcn_v1_0_set_pg_state(struct amdgpu_vcn_inst *vinst,
+				 enum amd_powergating_state state)
 {
 	/* This doesn't actually powergate the VCN block.
 	 * That's done in the dpm code via the SMC.  This
@@ -1757,28 +1850,29 @@ static int vcn_v1_0_set_powergating_state(void *handle,
 	 * the smc and the hw blocks
 	 */
 	int ret;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
-	if(state == adev->vcn.cur_state)
+	if (state == vinst->cur_state)
 		return 0;
 
 	if (state == AMD_PG_STATE_GATE)
-		ret = vcn_v1_0_stop(adev);
+		ret = vcn_v1_0_stop(vinst);
 	else
-		ret = vcn_v1_0_start(adev);
+		ret = vcn_v1_0_start(vinst);
+
+	if (!ret)
+		vinst->cur_state = state;
 
-	if(!ret)
-		adev->vcn.cur_state = state;
 	return ret;
 }
 
 static void vcn_v1_0_idle_work_handler(struct work_struct *work)
 {
-	struct amdgpu_device *adev =
-		container_of(work, struct amdgpu_device, vcn.idle_work.work);
+	struct amdgpu_vcn_inst *vcn_inst =
+		container_of(work, struct amdgpu_vcn_inst, idle_work.work);
+	struct amdgpu_device *adev = vcn_inst->adev;
 	unsigned int fences = 0, i;
 
-	for (i = 0; i < adev->vcn.num_enc_rings; ++i)
+	for (i = 0; i < adev->vcn.inst[0].num_enc_rings; ++i)
 		fences += amdgpu_fence_count_emitted(&adev->vcn.inst->ring_enc[i]);
 
 	if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) {
@@ -1789,37 +1883,37 @@ static void vcn_v1_0_idle_work_handler(struct work_struct *work)
 		else
 			new_state.fw_based = VCN_DPG_STATE__UNPAUSE;
 
-		if (amdgpu_fence_count_emitted(&adev->jpeg.inst->ring_dec))
+		if (amdgpu_fence_count_emitted(adev->jpeg.inst->ring_dec))
 			new_state.jpeg = VCN_DPG_STATE__PAUSE;
 		else
 			new_state.jpeg = VCN_DPG_STATE__UNPAUSE;
 
-		adev->vcn.pause_dpg_mode(adev, 0, &new_state);
+		adev->vcn.inst->pause_dpg_mode(vcn_inst, &new_state);
 	}
 
-	fences += amdgpu_fence_count_emitted(&adev->jpeg.inst->ring_dec);
+	fences += amdgpu_fence_count_emitted(adev->jpeg.inst->ring_dec);
 	fences += amdgpu_fence_count_emitted(&adev->vcn.inst->ring_dec);
 
 	if (fences == 0) {
 		amdgpu_gfx_off_ctrl(adev, true);
 		if (adev->pm.dpm_enabled)
-			amdgpu_dpm_enable_uvd(adev, false);
+			amdgpu_dpm_enable_vcn(adev, false, 0);
 		else
 			amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_VCN,
 			       AMD_PG_STATE_GATE);
 	} else {
-		schedule_delayed_work(&adev->vcn.idle_work, VCN_IDLE_TIMEOUT);
+		schedule_delayed_work(&adev->vcn.inst[0].idle_work, VCN_IDLE_TIMEOUT);
 	}
 }
 
 static void vcn_v1_0_ring_begin_use(struct amdgpu_ring *ring)
 {
 	struct	amdgpu_device *adev = ring->adev;
-	bool set_clocks = !cancel_delayed_work_sync(&adev->vcn.idle_work);
+	bool set_clocks = !cancel_delayed_work_sync(&adev->vcn.inst[0].idle_work);
 
-	mutex_lock(&adev->vcn.vcn1_jpeg1_workaround);
+	mutex_lock(&adev->vcn.inst[0].vcn1_jpeg1_workaround);
 
-	if (amdgpu_fence_wait_empty(&ring->adev->jpeg.inst->ring_dec))
+	if (amdgpu_fence_wait_empty(ring->adev->jpeg.inst->ring_dec))
 		DRM_ERROR("VCN dec: jpeg dec ring may not be empty\n");
 
 	vcn_v1_0_set_pg_for_begin_use(ring, set_clocks);
@@ -1833,7 +1927,7 @@ void vcn_v1_0_set_pg_for_begin_use(struct amdgpu_ring *ring, bool set_clocks)
 	if (set_clocks) {
 		amdgpu_gfx_off_ctrl(adev, false);
 		if (adev->pm.dpm_enabled)
-			amdgpu_dpm_enable_uvd(adev, true);
+			amdgpu_dpm_enable_vcn(adev, true, 0);
 		else
 			amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_VCN,
 			       AMD_PG_STATE_UNGATE);
@@ -1843,7 +1937,7 @@ void vcn_v1_0_set_pg_for_begin_use(struct amdgpu_ring *ring, bool set_clocks)
 		struct dpg_pause_state new_state;
 		unsigned int fences = 0, i;
 
-		for (i = 0; i < adev->vcn.num_enc_rings; ++i)
+		for (i = 0; i < adev->vcn.inst[0].num_enc_rings; ++i)
 			fences += amdgpu_fence_count_emitted(&adev->vcn.inst->ring_enc[i]);
 
 		if (fences)
@@ -1851,7 +1945,7 @@ void vcn_v1_0_set_pg_for_begin_use(struct amdgpu_ring *ring, bool set_clocks)
 		else
 			new_state.fw_based = VCN_DPG_STATE__UNPAUSE;
 
-		if (amdgpu_fence_count_emitted(&adev->jpeg.inst->ring_dec))
+		if (amdgpu_fence_count_emitted(adev->jpeg.inst->ring_dec))
 			new_state.jpeg = VCN_DPG_STATE__PAUSE;
 		else
 			new_state.jpeg = VCN_DPG_STATE__UNPAUSE;
@@ -1861,20 +1955,79 @@ void vcn_v1_0_set_pg_for_begin_use(struct amdgpu_ring *ring, bool set_clocks)
 		else if (ring->funcs->type == AMDGPU_RING_TYPE_VCN_JPEG)
 			new_state.jpeg = VCN_DPG_STATE__PAUSE;
 
-		adev->vcn.pause_dpg_mode(adev, 0, &new_state);
+		adev->vcn.inst->pause_dpg_mode(adev->vcn.inst, &new_state);
 	}
 }
 
 void vcn_v1_0_ring_end_use(struct amdgpu_ring *ring)
 {
-	schedule_delayed_work(&ring->adev->vcn.idle_work, VCN_IDLE_TIMEOUT);
-	mutex_unlock(&ring->adev->vcn.vcn1_jpeg1_workaround);
+	schedule_delayed_work(&ring->adev->vcn.inst[0].idle_work, VCN_IDLE_TIMEOUT);
+	mutex_unlock(&ring->adev->vcn.inst[0].vcn1_jpeg1_workaround);
+}
+
+static void vcn_v1_0_print_ip_state(struct amdgpu_ip_block *ip_block, struct drm_printer *p)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, j;
+	uint32_t reg_count = ARRAY_SIZE(vcn_reg_list_1_0);
+	uint32_t inst_off, is_powered;
+
+	if (!adev->vcn.ip_dump)
+		return;
+
+	drm_printf(p, "num_instances:%d\n", adev->vcn.num_vcn_inst);
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+		if (adev->vcn.harvest_config & (1 << i)) {
+			drm_printf(p, "\nHarvested Instance:VCN%d Skipping dump\n", i);
+			continue;
+		}
+
+		inst_off = i * reg_count;
+		is_powered = (adev->vcn.ip_dump[inst_off] &
+				UVD_POWER_STATUS__UVD_POWER_STATUS_MASK) != 1;
+
+		if (is_powered) {
+			drm_printf(p, "\nActive Instance:VCN%d\n", i);
+			for (j = 0; j < reg_count; j++)
+				drm_printf(p, "%-50s \t 0x%08x\n", vcn_reg_list_1_0[j].reg_name,
+					   adev->vcn.ip_dump[inst_off + j]);
+		} else {
+			drm_printf(p, "\nInactive Instance:VCN%d\n", i);
+		}
+	}
+}
+
+static void vcn_v1_0_dump_ip_state(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, j;
+	bool is_powered;
+	uint32_t inst_off;
+	uint32_t reg_count = ARRAY_SIZE(vcn_reg_list_1_0);
+
+	if (!adev->vcn.ip_dump)
+		return;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+		if (adev->vcn.harvest_config & (1 << i))
+			continue;
+
+		inst_off = i * reg_count;
+		/* mmUVD_POWER_STATUS is always readable and is first element of the array */
+		adev->vcn.ip_dump[inst_off] = RREG32_SOC15(VCN, i, mmUVD_POWER_STATUS);
+		is_powered = (adev->vcn.ip_dump[inst_off] &
+				UVD_POWER_STATUS__UVD_POWER_STATUS_MASK) != 1;
+
+		if (is_powered)
+			for (j = 1; j < reg_count; j++)
+				adev->vcn.ip_dump[inst_off + j] =
+					RREG32(SOC15_REG_ENTRY_OFFSET_INST(vcn_reg_list_1_0[j], i));
+	}
 }
 
 static const struct amd_ip_funcs vcn_v1_0_ip_funcs = {
 	.name = "vcn_v1_0",
 	.early_init = vcn_v1_0_early_init,
-	.late_init = NULL,
 	.sw_init = vcn_v1_0_sw_init,
 	.sw_fini = vcn_v1_0_sw_fini,
 	.hw_init = vcn_v1_0_hw_init,
@@ -1883,23 +2036,91 @@ static const struct amd_ip_funcs vcn_v1_0_ip_funcs = {
 	.resume = vcn_v1_0_resume,
 	.is_idle = vcn_v1_0_is_idle,
 	.wait_for_idle = vcn_v1_0_wait_for_idle,
-	.check_soft_reset = NULL /* vcn_v1_0_check_soft_reset */,
-	.pre_soft_reset = NULL /* vcn_v1_0_pre_soft_reset */,
-	.soft_reset = NULL /* vcn_v1_0_soft_reset */,
-	.post_soft_reset = NULL /* vcn_v1_0_post_soft_reset */,
 	.set_clockgating_state = vcn_v1_0_set_clockgating_state,
-	.set_powergating_state = vcn_v1_0_set_powergating_state,
+	.set_powergating_state = vcn_set_powergating_state,
+	.dump_ip_state = vcn_v1_0_dump_ip_state,
+	.print_ip_state = vcn_v1_0_print_ip_state,
 };
 
+/*
+ * It is a hardware issue that VCN can't handle a GTT TMZ buffer on
+ * CHIP_RAVEN series ASIC. Move such a GTT TMZ buffer to VRAM domain
+ * before command submission as a workaround.
+ */
+static int vcn_v1_0_validate_bo(struct amdgpu_cs_parser *parser,
+				struct amdgpu_job *job,
+				uint64_t addr)
+{
+	struct ttm_operation_ctx ctx = { false, false };
+	struct amdgpu_fpriv *fpriv = parser->filp->driver_priv;
+	struct amdgpu_vm *vm = &fpriv->vm;
+	struct amdgpu_bo_va_mapping *mapping;
+	struct amdgpu_bo *bo;
+	int r;
+
+	addr &= AMDGPU_GMC_HOLE_MASK;
+	if (addr & 0x7) {
+		DRM_ERROR("VCN messages must be 8 byte aligned!\n");
+		return -EINVAL;
+	}
+
+	mapping = amdgpu_vm_bo_lookup_mapping(vm, addr/AMDGPU_GPU_PAGE_SIZE);
+	if (!mapping || !mapping->bo_va || !mapping->bo_va->base.bo)
+		return -EINVAL;
+
+	bo = mapping->bo_va->base.bo;
+	if (!(bo->flags & AMDGPU_GEM_CREATE_ENCRYPTED))
+		return 0;
+
+	amdgpu_bo_placement_from_domain(bo, AMDGPU_GEM_DOMAIN_VRAM);
+	r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
+	if (r) {
+		DRM_ERROR("Failed to validate the VCN message BO (%d)!\n", r);
+		return r;
+	}
+
+	return r;
+}
+
+static int vcn_v1_0_ring_patch_cs_in_place(struct amdgpu_cs_parser *p,
+					   struct amdgpu_job *job,
+					   struct amdgpu_ib *ib)
+{
+	uint32_t msg_lo = 0, msg_hi = 0;
+	int i, r;
+
+	if (!(ib->flags & AMDGPU_IB_FLAGS_SECURE))
+		return 0;
+
+	for (i = 0; i < ib->length_dw; i += 2) {
+		uint32_t reg = amdgpu_ib_get_value(ib, i);
+		uint32_t val = amdgpu_ib_get_value(ib, i + 1);
+
+		if (reg == PACKET0(p->adev->vcn.inst[0].internal.data0, 0)) {
+			msg_lo = val;
+		} else if (reg == PACKET0(p->adev->vcn.inst[0].internal.data1, 0)) {
+			msg_hi = val;
+		} else if (reg == PACKET0(p->adev->vcn.inst[0].internal.cmd, 0)) {
+			r = vcn_v1_0_validate_bo(p, job,
+						 ((u64)msg_hi) << 32 | msg_lo);
+			if (r)
+				return r;
+		}
+	}
+
+	return 0;
+}
+
 static const struct amdgpu_ring_funcs vcn_v1_0_dec_ring_vm_funcs = {
 	.type = AMDGPU_RING_TYPE_VCN_DEC,
 	.align_mask = 0xf,
 	.support_64bit_ptrs = false,
 	.no_user_fence = true,
-	.vmhub = AMDGPU_MMHUB_0,
+	.secure_submission_supported = true,
 	.get_rptr = vcn_v1_0_dec_ring_get_rptr,
 	.get_wptr = vcn_v1_0_dec_ring_get_wptr,
 	.set_wptr = vcn_v1_0_dec_ring_set_wptr,
+	.patch_cs_in_place = vcn_v1_0_ring_patch_cs_in_place,
 	.emit_frame_size =
 		6 + 6 + /* hdp invalidate / flush */
 		SOC15_FLUSH_GPU_TLB_NUM_WREG * 6 +
@@ -1930,7 +2151,6 @@ static const struct amdgpu_ring_funcs vcn_v1_0_enc_ring_vm_funcs = {
 	.nop = VCN_ENC_CMD_NO_OP,
 	.support_64bit_ptrs = false,
 	.no_user_fence = true,
-	.vmhub = AMDGPU_MMHUB_0,
 	.get_rptr = vcn_v1_0_enc_ring_get_rptr,
 	.get_wptr = vcn_v1_0_enc_ring_get_wptr,
 	.set_wptr = vcn_v1_0_enc_ring_set_wptr,
@@ -1959,17 +2179,14 @@ static const struct amdgpu_ring_funcs vcn_v1_0_enc_ring_vm_funcs = {
 static void vcn_v1_0_set_dec_ring_funcs(struct amdgpu_device *adev)
 {
 	adev->vcn.inst->ring_dec.funcs = &vcn_v1_0_dec_ring_vm_funcs;
-	DRM_INFO("VCN decode is enabled in VM mode\n");
 }
 
 static void vcn_v1_0_set_enc_ring_funcs(struct amdgpu_device *adev)
 {
 	int i;
 
-	for (i = 0; i < adev->vcn.num_enc_rings; ++i)
+	for (i = 0; i < adev->vcn.inst[0].num_enc_rings; ++i)
 		adev->vcn.inst->ring_enc[i].funcs = &vcn_v1_0_enc_ring_vm_funcs;
-
-	DRM_INFO("VCN encode is enabled in VM mode\n");
 }
 
 static const struct amdgpu_irq_src_funcs vcn_v1_0_irq_funcs = {
@@ -1979,12 +2196,11 @@ static const struct amdgpu_irq_src_funcs vcn_v1_0_irq_funcs = {
 
 static void vcn_v1_0_set_irq_funcs(struct amdgpu_device *adev)
 {
-	adev->vcn.inst->irq.num_types = adev->vcn.num_enc_rings + 2;
+	adev->vcn.inst->irq.num_types = adev->vcn.inst[0].num_enc_rings + 2;
 	adev->vcn.inst->irq.funcs = &vcn_v1_0_irq_funcs;
 }
 
-const struct amdgpu_ip_block_version vcn_v1_0_ip_block =
-{
+const struct amdgpu_ip_block_version vcn_v1_0_ip_block = {
 		.type = AMD_IP_BLOCK_TYPE_VCN,
 		.major = 1,
 		.minor = 0,
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c
index 313fc1b53999..8897dcc9c1a0 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c
@@ -37,6 +37,10 @@
 #include "vcn/vcn_2_0_0_sh_mask.h"
 #include "ivsrcid/vcn/irqsrcs_vcn_2_0.h"
 
+#define VCN_VID_SOC_ADDRESS_2_0					0x1fa00
+#define VCN1_VID_SOC_ADDRESS_3_0				0x48200
+#define VCN1_AON_SOC_ADDRESS_3_0				0x48000
+
 #define mmUVD_CONTEXT_ID_INTERNAL_OFFSET			0x1fd
 #define mmUVD_GPCOM_VCPU_CMD_INTERNAL_OFFSET			0x503
 #define mmUVD_GPCOM_VCPU_DATA0_INTERNAL_OFFSET			0x504
@@ -50,50 +54,90 @@
 #define mmUVD_LMI_RBC_IB_64BIT_BAR_LOW_INTERNAL_OFFSET		0x5a7
 #define mmUVD_RBC_IB_SIZE_INTERNAL_OFFSET			0x1e2
 
+static const struct amdgpu_hwip_reg_entry vcn_reg_list_2_0[] = {
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_POWER_STATUS),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_STATUS),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_CONTEXT_ID),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_CONTEXT_ID2),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_GPCOM_VCPU_DATA0),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_GPCOM_VCPU_DATA1),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_GPCOM_VCPU_CMD),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_BASE_HI),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_BASE_LO),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_BASE_HI2),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_BASE_LO2),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_BASE_HI3),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_BASE_LO3),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_BASE_HI4),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_BASE_LO4),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_RPTR),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_WPTR),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_RPTR2),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_WPTR2),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_RPTR3),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_WPTR3),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_RPTR4),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_WPTR4),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_SIZE),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_SIZE2),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_SIZE3),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_SIZE4),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_PGFSM_CONFIG),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_PGFSM_STATUS),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_DPG_LMA_CTL),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_DPG_LMA_DATA),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_DPG_LMA_MASK),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_DPG_PAUSE)
+};
+
 static void vcn_v2_0_set_dec_ring_funcs(struct amdgpu_device *adev);
 static void vcn_v2_0_set_enc_ring_funcs(struct amdgpu_device *adev);
 static void vcn_v2_0_set_irq_funcs(struct amdgpu_device *adev);
-static int vcn_v2_0_set_powergating_state(void *handle,
-				enum amd_powergating_state state);
-static int vcn_v2_0_pause_dpg_mode(struct amdgpu_device *adev,
-				int inst_idx, struct dpg_pause_state *new_state);
+static int vcn_v2_0_set_pg_state(struct amdgpu_vcn_inst *vinst,
+				 enum amd_powergating_state state);
+static int vcn_v2_0_pause_dpg_mode(struct amdgpu_vcn_inst *vinst,
+				   struct dpg_pause_state *new_state);
 static int vcn_v2_0_start_sriov(struct amdgpu_device *adev);
+static int vcn_v2_0_reset(struct amdgpu_vcn_inst *vinst);
+
 /**
- * vcn_v2_0_early_init - set function pointers
+ * vcn_v2_0_early_init - set function pointers and load microcode
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Set ring and irq function pointers
+ * Load microcode from filesystem
  */
-static int vcn_v2_0_early_init(void *handle)
+static int vcn_v2_0_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (amdgpu_sriov_vf(adev))
-		adev->vcn.num_enc_rings = 1;
+		adev->vcn.inst[0].num_enc_rings = 1;
 	else
-		adev->vcn.num_enc_rings = 2;
+		adev->vcn.inst[0].num_enc_rings = 2;
 
+	adev->vcn.inst->set_pg_state = vcn_v2_0_set_pg_state;
 	vcn_v2_0_set_dec_ring_funcs(adev);
 	vcn_v2_0_set_enc_ring_funcs(adev);
 	vcn_v2_0_set_irq_funcs(adev);
 
-	return 0;
+	return amdgpu_vcn_early_init(adev, 0);
 }
 
 /**
  * vcn_v2_0_sw_init - sw init for VCN block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Load firmware and sw initialization
  */
-static int vcn_v2_0_sw_init(void *handle)
+static int vcn_v2_0_sw_init(struct amdgpu_ip_block *ip_block)
 {
 	struct amdgpu_ring *ring;
 	int i, r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-	volatile struct amdgpu_fw_shared *fw_shared;
+	struct amdgpu_device *adev = ip_block->adev;
+	struct amdgpu_fw_shared *fw_shared;
 
 	/* VCN DEC TRAP */
 	r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_VCN,
@@ -103,7 +147,7 @@ static int vcn_v2_0_sw_init(void *handle)
 		return r;
 
 	/* VCN ENC TRAP */
-	for (i = 0; i < adev->vcn.num_enc_rings; ++i) {
+	for (i = 0; i < adev->vcn.inst[0].num_enc_rings; ++i) {
 		r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_VCN,
 				      i + VCN_2_0__SRCID__UVD_ENC_GENERAL_PURPOSE,
 				      &adev->vcn.inst->irq);
@@ -111,13 +155,13 @@ static int vcn_v2_0_sw_init(void *handle)
 			return r;
 	}
 
-	r = amdgpu_vcn_sw_init(adev);
+	r = amdgpu_vcn_sw_init(adev, 0);
 	if (r)
 		return r;
 
-	amdgpu_vcn_setup_ucode(adev);
+	amdgpu_vcn_setup_ucode(adev, 0);
 
-	r = amdgpu_vcn_resume(adev);
+	r = amdgpu_vcn_resume(adev, 0);
 	if (r)
 		return r;
 
@@ -125,6 +169,7 @@ static int vcn_v2_0_sw_init(void *handle)
 
 	ring->use_doorbell = true;
 	ring->doorbell_index = adev->doorbell_index.vcn.vcn_ring0_1 << 1;
+	ring->vm_hub = AMDGPU_MMHUB0(0);
 
 	sprintf(ring->name, "vcn_dec");
 	r = amdgpu_ring_init(adev, ring, 512, &adev->vcn.inst->irq, 0,
@@ -132,29 +177,30 @@ static int vcn_v2_0_sw_init(void *handle)
 	if (r)
 		return r;
 
-	adev->vcn.internal.context_id = mmUVD_CONTEXT_ID_INTERNAL_OFFSET;
-	adev->vcn.internal.ib_vmid = mmUVD_LMI_RBC_IB_VMID_INTERNAL_OFFSET;
-	adev->vcn.internal.ib_bar_low = mmUVD_LMI_RBC_IB_64BIT_BAR_LOW_INTERNAL_OFFSET;
-	adev->vcn.internal.ib_bar_high = mmUVD_LMI_RBC_IB_64BIT_BAR_HIGH_INTERNAL_OFFSET;
-	adev->vcn.internal.ib_size = mmUVD_RBC_IB_SIZE_INTERNAL_OFFSET;
-	adev->vcn.internal.gp_scratch8 = mmUVD_GP_SCRATCH8_INTERNAL_OFFSET;
+	adev->vcn.inst[0].internal.context_id = mmUVD_CONTEXT_ID_INTERNAL_OFFSET;
+	adev->vcn.inst[0].internal.ib_vmid = mmUVD_LMI_RBC_IB_VMID_INTERNAL_OFFSET;
+	adev->vcn.inst[0].internal.ib_bar_low = mmUVD_LMI_RBC_IB_64BIT_BAR_LOW_INTERNAL_OFFSET;
+	adev->vcn.inst[0].internal.ib_bar_high = mmUVD_LMI_RBC_IB_64BIT_BAR_HIGH_INTERNAL_OFFSET;
+	adev->vcn.inst[0].internal.ib_size = mmUVD_RBC_IB_SIZE_INTERNAL_OFFSET;
+	adev->vcn.inst[0].internal.gp_scratch8 = mmUVD_GP_SCRATCH8_INTERNAL_OFFSET;
 
-	adev->vcn.internal.scratch9 = mmUVD_SCRATCH9_INTERNAL_OFFSET;
+	adev->vcn.inst[0].internal.scratch9 = mmUVD_SCRATCH9_INTERNAL_OFFSET;
 	adev->vcn.inst->external.scratch9 = SOC15_REG_OFFSET(UVD, 0, mmUVD_SCRATCH9);
-	adev->vcn.internal.data0 = mmUVD_GPCOM_VCPU_DATA0_INTERNAL_OFFSET;
+	adev->vcn.inst[0].internal.data0 = mmUVD_GPCOM_VCPU_DATA0_INTERNAL_OFFSET;
 	adev->vcn.inst->external.data0 = SOC15_REG_OFFSET(UVD, 0, mmUVD_GPCOM_VCPU_DATA0);
-	adev->vcn.internal.data1 = mmUVD_GPCOM_VCPU_DATA1_INTERNAL_OFFSET;
+	adev->vcn.inst[0].internal.data1 = mmUVD_GPCOM_VCPU_DATA1_INTERNAL_OFFSET;
 	adev->vcn.inst->external.data1 = SOC15_REG_OFFSET(UVD, 0, mmUVD_GPCOM_VCPU_DATA1);
-	adev->vcn.internal.cmd = mmUVD_GPCOM_VCPU_CMD_INTERNAL_OFFSET;
+	adev->vcn.inst[0].internal.cmd = mmUVD_GPCOM_VCPU_CMD_INTERNAL_OFFSET;
 	adev->vcn.inst->external.cmd = SOC15_REG_OFFSET(UVD, 0, mmUVD_GPCOM_VCPU_CMD);
-	adev->vcn.internal.nop = mmUVD_NO_OP_INTERNAL_OFFSET;
+	adev->vcn.inst[0].internal.nop = mmUVD_NO_OP_INTERNAL_OFFSET;
 	adev->vcn.inst->external.nop = SOC15_REG_OFFSET(UVD, 0, mmUVD_NO_OP);
 
-	for (i = 0; i < adev->vcn.num_enc_rings; ++i) {
+	for (i = 0; i < adev->vcn.inst[0].num_enc_rings; ++i) {
 		enum amdgpu_ring_priority_level hw_prio = amdgpu_vcn_get_enc_ring_prio(i);
 
 		ring = &adev->vcn.inst->ring_enc[i];
 		ring->use_doorbell = true;
+		ring->vm_hub = AMDGPU_MMHUB0(0);
 		if (!amdgpu_sriov_vf(adev))
 			ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 2 + i;
 		else
@@ -166,29 +212,47 @@ static int vcn_v2_0_sw_init(void *handle)
 			return r;
 	}
 
-	adev->vcn.pause_dpg_mode = vcn_v2_0_pause_dpg_mode;
+	adev->vcn.inst[0].pause_dpg_mode = vcn_v2_0_pause_dpg_mode;
+	adev->vcn.inst[0].reset = vcn_v2_0_reset;
+
+	adev->vcn.supported_reset =
+		amdgpu_get_soft_full_reset_mask(&adev->vcn.inst[0].ring_enc[0]);
+	if (!amdgpu_sriov_vf(adev))
+		adev->vcn.supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE;
 
 	r = amdgpu_virt_alloc_mm_table(adev);
 	if (r)
 		return r;
 
-	fw_shared = adev->vcn.inst->fw_shared_cpu_addr;
+	fw_shared = adev->vcn.inst->fw_shared.cpu_addr;
 	fw_shared->present_flag_0 = cpu_to_le32(AMDGPU_VCN_MULTI_QUEUE_FLAG);
+
+	if (amdgpu_vcnfw_log)
+		amdgpu_vcn_fwlog_init(adev->vcn.inst);
+
+	r = amdgpu_vcn_reg_dump_init(adev, vcn_reg_list_2_0, ARRAY_SIZE(vcn_reg_list_2_0));
+	if (r)
+		return r;
+
+	r = amdgpu_vcn_sysfs_reset_mask_init(adev);
+	if (r)
+		return r;
+
 	return 0;
 }
 
 /**
  * vcn_v2_0_sw_fini - sw fini for VCN block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * VCN suspend and free up sw allocation
  */
-static int vcn_v2_0_sw_fini(void *handle)
+static int vcn_v2_0_sw_fini(struct amdgpu_ip_block *ip_block)
 {
 	int r, idx;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-	volatile struct amdgpu_fw_shared *fw_shared = adev->vcn.inst->fw_shared_cpu_addr;
+	struct amdgpu_device *adev = ip_block->adev;
+	struct amdgpu_fw_shared *fw_shared = adev->vcn.inst->fw_shared.cpu_addr;
 
 	if (drm_dev_enter(adev_to_drm(adev), &idx)) {
 		fw_shared->present_flag_0 = 0;
@@ -197,25 +261,27 @@ static int vcn_v2_0_sw_fini(void *handle)
 
 	amdgpu_virt_free_mm_table(adev);
 
-	r = amdgpu_vcn_suspend(adev);
+	r = amdgpu_vcn_suspend(adev, 0);
 	if (r)
 		return r;
 
-	r = amdgpu_vcn_sw_fini(adev);
+	amdgpu_vcn_sysfs_reset_mask_fini(adev);
 
-	return r;
+	amdgpu_vcn_sw_fini(adev, 0);
+
+	return 0;
 }
 
 /**
  * vcn_v2_0_hw_init - start and test VCN block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Initialize the hardware, boot up the VCPU and do some testing
  */
-static int vcn_v2_0_hw_init(void *handle)
+static int vcn_v2_0_hw_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	struct amdgpu_ring *ring = &adev->vcn.inst->ring_dec;
 	int i, r;
 
@@ -227,44 +293,40 @@ static int vcn_v2_0_hw_init(void *handle)
 
 	r = amdgpu_ring_test_helper(ring);
 	if (r)
-		goto done;
+		return r;
 
 	//Disable vcn decode for sriov
 	if (amdgpu_sriov_vf(adev))
 		ring->sched.ready = false;
 
-	for (i = 0; i < adev->vcn.num_enc_rings; ++i) {
+	for (i = 0; i < adev->vcn.inst[0].num_enc_rings; ++i) {
 		ring = &adev->vcn.inst->ring_enc[i];
 		r = amdgpu_ring_test_helper(ring);
 		if (r)
-			goto done;
+			return r;
 	}
 
-done:
-	if (!r)
-		DRM_INFO("VCN decode and encode initialized successfully(under %s).\n",
-			(adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)?"DPG Mode":"SPG Mode");
-
-	return r;
+	return 0;
 }
 
 /**
  * vcn_v2_0_hw_fini - stop the hardware block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Stop the VCN block, mark ring as not ready any more
  */
-static int vcn_v2_0_hw_fini(void *handle)
+static int vcn_v2_0_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
+	struct amdgpu_vcn_inst *vinst = adev->vcn.inst;
 
-	cancel_delayed_work_sync(&adev->vcn.idle_work);
+	cancel_delayed_work_sync(&vinst->idle_work);
 
 	if ((adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) ||
-	    (adev->vcn.cur_state != AMD_PG_STATE_GATE &&
-	      RREG32_SOC15(VCN, 0, mmUVD_STATUS)))
-		vcn_v2_0_set_powergating_state(adev, AMD_PG_STATE_GATE);
+	    (vinst->cur_state != AMD_PG_STATE_GATE &&
+	     RREG32_SOC15(VCN, 0, mmUVD_STATUS)))
+		vinst->set_pg_state(vinst, AMD_PG_STATE_GATE);
 
 	return 0;
 }
@@ -272,20 +334,19 @@ static int vcn_v2_0_hw_fini(void *handle)
 /**
  * vcn_v2_0_suspend - suspend VCN block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * HW fini and suspend VCN block
  */
-static int vcn_v2_0_suspend(void *handle)
+static int vcn_v2_0_suspend(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
-	r = vcn_v2_0_hw_fini(adev);
+	r = vcn_v2_0_hw_fini(ip_block);
 	if (r)
 		return r;
 
-	r = amdgpu_vcn_suspend(adev);
+	r = amdgpu_vcn_suspend(ip_block->adev, 0);
 
 	return r;
 }
@@ -293,20 +354,19 @@ static int vcn_v2_0_suspend(void *handle)
 /**
  * vcn_v2_0_resume - resume VCN block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Resume firmware and hw init VCN block
  */
-static int vcn_v2_0_resume(void *handle)
+static int vcn_v2_0_resume(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
-	r = amdgpu_vcn_resume(adev);
+	r = amdgpu_vcn_resume(ip_block->adev, 0);
 	if (r)
 		return r;
 
-	r = vcn_v2_0_hw_init(adev);
+	r = vcn_v2_0_hw_init(ip_block);
 
 	return r;
 }
@@ -314,13 +374,14 @@ static int vcn_v2_0_resume(void *handle)
 /**
  * vcn_v2_0_mc_resume - memory controller programming
  *
- * @adev: amdgpu_device pointer
+ * @vinst: Pointer to the VCN instance structure
  *
  * Let the VCN memory controller know it's offsets
  */
-static void vcn_v2_0_mc_resume(struct amdgpu_device *adev)
+static void vcn_v2_0_mc_resume(struct amdgpu_vcn_inst *vinst)
 {
-	uint32_t size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.fw->size + 4);
+	struct amdgpu_device *adev = vinst->adev;
+	uint32_t size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.inst[0].fw->size + 4);
 	uint32_t offset;
 
 	if (amdgpu_sriov_vf(adev))
@@ -364,9 +425,9 @@ static void vcn_v2_0_mc_resume(struct amdgpu_device *adev)
 
 	/* non-cache window */
 	WREG32_SOC15(UVD, 0, mmUVD_LMI_VCPU_NC0_64BIT_BAR_LOW,
-		lower_32_bits(adev->vcn.inst->fw_shared_gpu_addr));
+		lower_32_bits(adev->vcn.inst->fw_shared.gpu_addr));
 	WREG32_SOC15(UVD, 0, mmUVD_LMI_VCPU_NC0_64BIT_BAR_HIGH,
-		upper_32_bits(adev->vcn.inst->fw_shared_gpu_addr));
+		upper_32_bits(adev->vcn.inst->fw_shared.gpu_addr));
 	WREG32_SOC15(UVD, 0, mmUVD_VCPU_NONCACHE_OFFSET0, 0);
 	WREG32_SOC15(UVD, 0, mmUVD_VCPU_NONCACHE_SIZE0,
 		AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_fw_shared)));
@@ -374,9 +435,11 @@ static void vcn_v2_0_mc_resume(struct amdgpu_device *adev)
 	WREG32_SOC15(UVD, 0, mmUVD_GFX10_ADDR_CONFIG, adev->gfx.config.gb_addr_config);
 }
 
-static void vcn_v2_0_mc_resume_dpg_mode(struct amdgpu_device *adev, bool indirect)
+static void vcn_v2_0_mc_resume_dpg_mode(struct amdgpu_vcn_inst *vinst,
+					bool indirect)
 {
-	uint32_t size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.fw->size + 4);
+	struct amdgpu_device *adev = vinst->adev;
+	uint32_t size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.inst[0].fw->size + 4);
 	uint32_t offset;
 
 	/* cache window 0: fw */
@@ -455,10 +518,10 @@ static void vcn_v2_0_mc_resume_dpg_mode(struct amdgpu_device *adev, bool indirec
 	/* non-cache window */
 	WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 		UVD, 0, mmUVD_LMI_VCPU_NC0_64BIT_BAR_LOW),
-		lower_32_bits(adev->vcn.inst->fw_shared_gpu_addr), 0, indirect);
+		lower_32_bits(adev->vcn.inst->fw_shared.gpu_addr), 0, indirect);
 	WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 		UVD, 0, mmUVD_LMI_VCPU_NC0_64BIT_BAR_HIGH),
-		upper_32_bits(adev->vcn.inst->fw_shared_gpu_addr), 0, indirect);
+		upper_32_bits(adev->vcn.inst->fw_shared.gpu_addr), 0, indirect);
 	WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 		UVD, 0, mmUVD_VCPU_NONCACHE_OFFSET0), 0, 0, indirect);
 	WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
@@ -473,12 +536,13 @@ static void vcn_v2_0_mc_resume_dpg_mode(struct amdgpu_device *adev, bool indirec
 /**
  * vcn_v2_0_disable_clock_gating - disable VCN clock gating
  *
- * @adev: amdgpu_device pointer
+ * @vinst: VCN instance
  *
  * Disable clock gating for VCN block
  */
-static void vcn_v2_0_disable_clock_gating(struct amdgpu_device *adev)
+static void vcn_v2_0_disable_clock_gating(struct amdgpu_vcn_inst *vinst)
 {
+	struct amdgpu_device *adev = vinst->adev;
 	uint32_t data;
 
 	if (amdgpu_sriov_vf(adev))
@@ -582,9 +646,10 @@ static void vcn_v2_0_disable_clock_gating(struct amdgpu_device *adev)
 	WREG32_SOC15(VCN, 0, mmUVD_SUVD_CGC_CTRL, data);
 }
 
-static void vcn_v2_0_clock_gating_dpg_mode(struct amdgpu_device *adev,
+static void vcn_v2_0_clock_gating_dpg_mode(struct amdgpu_vcn_inst *vinst,
 		uint8_t sram_sel, uint8_t indirect)
 {
+	struct amdgpu_device *adev = vinst->adev;
 	uint32_t reg_data = 0;
 
 	/* enable sw clock gating control */
@@ -633,12 +698,13 @@ static void vcn_v2_0_clock_gating_dpg_mode(struct amdgpu_device *adev,
 /**
  * vcn_v2_0_enable_clock_gating - enable VCN clock gating
  *
- * @adev: amdgpu_device pointer
+ * @vinst: VCN instance
  *
  * Enable clock gating for VCN block
  */
-static void vcn_v2_0_enable_clock_gating(struct amdgpu_device *adev)
+static void vcn_v2_0_enable_clock_gating(struct amdgpu_vcn_inst *vinst)
 {
+	struct amdgpu_device *adev = vinst->adev;
 	uint32_t data = 0;
 
 	if (amdgpu_sriov_vf(adev))
@@ -691,8 +757,9 @@ static void vcn_v2_0_enable_clock_gating(struct amdgpu_device *adev)
 	WREG32_SOC15(VCN, 0, mmUVD_SUVD_CGC_CTRL, data);
 }
 
-static void vcn_v2_0_disable_static_power_gating(struct amdgpu_device *adev)
+static void vcn_v2_0_disable_static_power_gating(struct amdgpu_vcn_inst *vinst)
 {
+	struct amdgpu_device *adev = vinst->adev;
 	uint32_t data = 0;
 
 	if (amdgpu_sriov_vf(adev))
@@ -740,8 +807,9 @@ static void vcn_v2_0_disable_static_power_gating(struct amdgpu_device *adev)
 	WREG32_SOC15(VCN, 0, mmUVD_POWER_STATUS, data);
 }
 
-static void vcn_v2_0_enable_static_power_gating(struct amdgpu_device *adev)
+static void vcn_v2_0_enable_static_power_gating(struct amdgpu_vcn_inst *vinst)
 {
+	struct amdgpu_device *adev = vinst->adev;
 	uint32_t data = 0;
 
 	if (amdgpu_sriov_vf(adev))
@@ -782,13 +850,15 @@ static void vcn_v2_0_enable_static_power_gating(struct amdgpu_device *adev)
 	}
 }
 
-static int vcn_v2_0_start_dpg_mode(struct amdgpu_device *adev, bool indirect)
+static int vcn_v2_0_start_dpg_mode(struct amdgpu_vcn_inst *vinst, bool indirect)
 {
-	volatile struct amdgpu_fw_shared *fw_shared = adev->vcn.inst->fw_shared_cpu_addr;
+	struct amdgpu_device *adev = vinst->adev;
+	struct amdgpu_fw_shared *fw_shared = adev->vcn.inst->fw_shared.cpu_addr;
 	struct amdgpu_ring *ring = &adev->vcn.inst->ring_dec;
 	uint32_t rb_bufsz, tmp;
+	int ret;
 
-	vcn_v2_0_enable_static_power_gating(adev);
+	vcn_v2_0_enable_static_power_gating(vinst);
 
 	/* enable dynamic power gating mode */
 	tmp = RREG32_SOC15(UVD, 0, mmUVD_POWER_STATUS);
@@ -800,7 +870,7 @@ static int vcn_v2_0_start_dpg_mode(struct amdgpu_device *adev, bool indirect)
 		adev->vcn.inst->dpg_sram_curr_addr = (uint32_t *)adev->vcn.inst->dpg_sram_cpu_addr;
 
 	/* enable clock gating */
-	vcn_v2_0_clock_gating_dpg_mode(adev, 0, indirect);
+	vcn_v2_0_clock_gating_dpg_mode(vinst, 0, indirect);
 
 	/* enable VCPU clock */
 	tmp = (0xFF << UVD_VCPU_CNTL__PRB_TIMEOUT_VAL__SHIFT);
@@ -849,7 +919,7 @@ static int vcn_v2_0_start_dpg_mode(struct amdgpu_device *adev, bool indirect)
 		 (0x1 << UVD_MPC_SET_MUX__SET_1__SHIFT) |
 		 (0x2 << UVD_MPC_SET_MUX__SET_2__SHIFT)), 0, indirect);
 
-	vcn_v2_0_mc_resume_dpg_mode(adev, indirect);
+	vcn_v2_0_mc_resume_dpg_mode(vinst, indirect);
 
 	WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
 		UVD, 0, mmUVD_REG_XX_MASK), 0x10, 0, indirect);
@@ -870,10 +940,13 @@ static int vcn_v2_0_start_dpg_mode(struct amdgpu_device *adev, bool indirect)
 		UVD, 0, mmUVD_MASTINT_EN),
 		UVD_MASTINT_EN__VCPU_EN_MASK, 0, indirect);
 
-	if (indirect)
-		psp_update_vcn_sram(adev, 0, adev->vcn.inst->dpg_sram_gpu_addr,
-				    (uint32_t)((uintptr_t)adev->vcn.inst->dpg_sram_curr_addr -
-					       (uintptr_t)adev->vcn.inst->dpg_sram_cpu_addr));
+	if (indirect) {
+		ret = amdgpu_vcn_psp_update_sram(adev, 0, 0);
+		if (ret) {
+			dev_err(adev->dev, "vcn sram load failed %d\n", ret);
+			return ret;
+		}
+	}
 
 	/* force RBC into idle state */
 	rb_bufsz = order_base_2(ring->ring_size);
@@ -916,31 +989,38 @@ static int vcn_v2_0_start_dpg_mode(struct amdgpu_device *adev, bool indirect)
 	/* Unstall DPG */
 	WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_POWER_STATUS),
 		0, ~UVD_POWER_STATUS__STALL_DPG_POWER_UP_MASK);
+
+	/* Keeping one read-back to ensure all register writes are done,
+	 * otherwise it may introduce race conditions.
+	 */
+	RREG32_SOC15(UVD, 0, mmUVD_STATUS);
+
 	return 0;
 }
 
-static int vcn_v2_0_start(struct amdgpu_device *adev)
+static int vcn_v2_0_start(struct amdgpu_vcn_inst *vinst)
 {
-	volatile struct amdgpu_fw_shared *fw_shared = adev->vcn.inst->fw_shared_cpu_addr;
+	struct amdgpu_device *adev = vinst->adev;
+	struct amdgpu_fw_shared *fw_shared = adev->vcn.inst->fw_shared.cpu_addr;
 	struct amdgpu_ring *ring = &adev->vcn.inst->ring_dec;
 	uint32_t rb_bufsz, tmp;
 	uint32_t lmi_swap_cntl;
 	int i, j, r;
 
 	if (adev->pm.dpm_enabled)
-		amdgpu_dpm_enable_uvd(adev, true);
+		amdgpu_dpm_enable_vcn(adev, true, 0);
 
 	if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)
-		return vcn_v2_0_start_dpg_mode(adev, adev->vcn.indirect_sram);
+		return vcn_v2_0_start_dpg_mode(vinst, adev->vcn.inst->indirect_sram);
 
-	vcn_v2_0_disable_static_power_gating(adev);
+	vcn_v2_0_disable_static_power_gating(vinst);
 
 	/* set uvd status busy */
 	tmp = RREG32_SOC15(UVD, 0, mmUVD_STATUS) | UVD_STATUS__UVD_BUSY;
 	WREG32_SOC15(UVD, 0, mmUVD_STATUS, tmp);
 
 	/*SW clock gating */
-	vcn_v2_0_disable_clock_gating(adev);
+	vcn_v2_0_disable_clock_gating(vinst);
 
 	/* enable VCPU clock */
 	WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_VCPU_CNTL),
@@ -984,7 +1064,7 @@ static int vcn_v2_0_start(struct amdgpu_device *adev)
 		(0x1 << UVD_MPC_SET_MUX__SET_1__SHIFT) |
 		(0x2 << UVD_MPC_SET_MUX__SET_2__SHIFT)));
 
-	vcn_v2_0_mc_resume(adev);
+	vcn_v2_0_mc_resume(vinst);
 
 	/* release VCPU reset to boot */
 	WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_SOFT_RESET), 0,
@@ -1089,13 +1169,21 @@ static int vcn_v2_0_start(struct amdgpu_device *adev)
 	WREG32_SOC15(UVD, 0, mmUVD_RB_SIZE2, ring->ring_size / 4);
 	fw_shared->multi_queue.encode_lowlatency_queue_mode &= ~FW_QUEUE_RING_RESET;
 
+	/* Keeping one read-back to ensure all register writes are done,
+	 * otherwise it may introduce race conditions.
+	 */
+	RREG32_SOC15(UVD, 0, mmUVD_STATUS);
+
 	return 0;
 }
 
-static int vcn_v2_0_stop_dpg_mode(struct amdgpu_device *adev)
+static int vcn_v2_0_stop_dpg_mode(struct amdgpu_vcn_inst *vinst)
 {
+	struct amdgpu_device *adev = vinst->adev;
+	struct dpg_pause_state state = {.fw_based = VCN_DPG_STATE__UNPAUSE};
 	uint32_t tmp;
 
+	vcn_v2_0_pause_dpg_mode(vinst, &state);
 	/* Wait for power status to be 1 */
 	SOC15_WAIT_ON_RREG(UVD, 0, mmUVD_POWER_STATUS, 1,
 		UVD_POWER_STATUS__UVD_POWER_STATUS_MASK);
@@ -1117,16 +1205,22 @@ static int vcn_v2_0_stop_dpg_mode(struct amdgpu_device *adev)
 	WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_POWER_STATUS), 0,
 			~UVD_POWER_STATUS__UVD_PG_MODE_MASK);
 
+	/* Keeping one read-back to ensure all register writes are done,
+	 * otherwise it may introduce race conditions.
+	 */
+	RREG32_SOC15(UVD, 0, mmUVD_STATUS);
+
 	return 0;
 }
 
-static int vcn_v2_0_stop(struct amdgpu_device *adev)
+static int vcn_v2_0_stop(struct amdgpu_vcn_inst *vinst)
 {
+	struct amdgpu_device *adev = vinst->adev;
 	uint32_t tmp;
 	int r;
 
 	if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) {
-		r = vcn_v2_0_stop_dpg_mode(adev);
+		r = vcn_v2_0_stop_dpg_mode(vinst);
 		if (r)
 			return r;
 		goto power_off;
@@ -1178,19 +1272,26 @@ static int vcn_v2_0_stop(struct amdgpu_device *adev)
 	/* clear status */
 	WREG32_SOC15(VCN, 0, mmUVD_STATUS, 0);
 
-	vcn_v2_0_enable_clock_gating(adev);
-	vcn_v2_0_enable_static_power_gating(adev);
+	vcn_v2_0_enable_clock_gating(vinst);
+	vcn_v2_0_enable_static_power_gating(vinst);
+
+	/* Keeping one read-back to ensure all register writes are done,
+	 * otherwise it may introduce race conditions.
+	 */
+	RREG32_SOC15(VCN, 0, mmUVD_STATUS);
 
 power_off:
 	if (adev->pm.dpm_enabled)
-		amdgpu_dpm_enable_uvd(adev, false);
+		amdgpu_dpm_enable_vcn(adev, false, 0);
 
 	return 0;
 }
 
-static int vcn_v2_0_pause_dpg_mode(struct amdgpu_device *adev,
-				int inst_idx, struct dpg_pause_state *new_state)
+static int vcn_v2_0_pause_dpg_mode(struct amdgpu_vcn_inst *vinst,
+				   struct dpg_pause_state *new_state)
 {
+	struct amdgpu_device *adev = vinst->adev;
+	int inst_idx = vinst->inst;
 	struct amdgpu_ring *ring;
 	uint32_t reg_data = 0;
 	int ret_code;
@@ -1207,7 +1308,7 @@ static int vcn_v2_0_pause_dpg_mode(struct amdgpu_device *adev,
 				UVD_POWER_STATUS__UVD_POWER_STATUS_MASK);
 
 			if (!ret_code) {
-				volatile struct amdgpu_fw_shared *fw_shared = adev->vcn.inst->fw_shared_cpu_addr;
+				struct amdgpu_fw_shared *fw_shared = adev->vcn.inst->fw_shared.cpu_addr;
 				/* pause DPG */
 				reg_data |= UVD_DPG_PAUSE__NJ_PAUSE_DPG_REQ_MASK;
 				WREG32_SOC15(UVD, 0, mmUVD_DPG_PAUSE, reg_data);
@@ -1265,16 +1366,26 @@ static int vcn_v2_0_pause_dpg_mode(struct amdgpu_device *adev,
 	return 0;
 }
 
-static bool vcn_v2_0_is_idle(void *handle)
+static int vcn_v2_0_reset(struct amdgpu_vcn_inst *vinst)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	int r;
+
+	r = vcn_v2_0_stop(vinst);
+	if (r)
+		return r;
+	return vcn_v2_0_start(vinst);
+}
+
+static bool vcn_v2_0_is_idle(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
 
 	return (RREG32_SOC15(VCN, 0, mmUVD_STATUS) == UVD_STATUS__IDLE);
 }
 
-static int vcn_v2_0_wait_for_idle(void *handle)
+static int vcn_v2_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int ret;
 
 	ret = SOC15_WAIT_ON_RREG(VCN, 0, mmUVD_STATUS, UVD_STATUS__IDLE,
@@ -1283,10 +1394,10 @@ static int vcn_v2_0_wait_for_idle(void *handle)
 	return ret;
 }
 
-static int vcn_v2_0_set_clockgating_state(void *handle,
+static int vcn_v2_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_clockgating_state state)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	bool enable = (state == AMD_CG_STATE_GATE);
 
 	if (amdgpu_sriov_vf(adev))
@@ -1294,12 +1405,12 @@ static int vcn_v2_0_set_clockgating_state(void *handle,
 
 	if (enable) {
 		/* wait for STATUS to clear */
-		if (!vcn_v2_0_is_idle(handle))
+		if (!vcn_v2_0_is_idle(ip_block))
 			return -EBUSY;
-		vcn_v2_0_enable_clock_gating(adev);
+		vcn_v2_0_enable_clock_gating(&adev->vcn.inst[0]);
 	} else {
 		/* disable HW gating and enable Sw gating */
-		vcn_v2_0_disable_clock_gating(adev);
+		vcn_v2_0_disable_clock_gating(&adev->vcn.inst[0]);
 	}
 	return 0;
 }
@@ -1330,7 +1441,7 @@ static uint64_t vcn_v2_0_dec_ring_get_wptr(struct amdgpu_ring *ring)
 	struct amdgpu_device *adev = ring->adev;
 
 	if (ring->use_doorbell)
-		return adev->wb.wb[ring->wptr_offs];
+		return *ring->wptr_cpu_addr;
 	else
 		return RREG32_SOC15(UVD, 0, mmUVD_RBC_RB_WPTR);
 }
@@ -1351,7 +1462,7 @@ static void vcn_v2_0_dec_ring_set_wptr(struct amdgpu_ring *ring)
 			lower_32_bits(ring->wptr) | 0x80000000);
 
 	if (ring->use_doorbell) {
-		adev->wb.wb[ring->wptr_offs] = lower_32_bits(ring->wptr);
+		*ring->wptr_cpu_addr = lower_32_bits(ring->wptr);
 		WDOORBELL32(ring->doorbell_index, lower_32_bits(ring->wptr));
 	} else {
 		WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_WPTR, lower_32_bits(ring->wptr));
@@ -1369,9 +1480,9 @@ void vcn_v2_0_dec_ring_insert_start(struct amdgpu_ring *ring)
 {
 	struct amdgpu_device *adev = ring->adev;
 
-	amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.data0, 0));
+	amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.data0, 0));
 	amdgpu_ring_write(ring, 0);
-	amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.cmd, 0));
+	amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.cmd, 0));
 	amdgpu_ring_write(ring, VCN_DEC_KMD_CMD | (VCN_DEC_CMD_PACKET_START << 1));
 }
 
@@ -1386,7 +1497,7 @@ void vcn_v2_0_dec_ring_insert_end(struct amdgpu_ring *ring)
 {
 	struct amdgpu_device *adev = ring->adev;
 
-	amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.cmd, 0));
+	amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[0].internal.cmd, 0));
 	amdgpu_ring_write(ring, VCN_DEC_KMD_CMD | (VCN_DEC_CMD_PACKET_END << 1));
 }
 
@@ -1406,7 +1517,7 @@ void vcn_v2_0_dec_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count)
 	WARN_ON(ring->wptr % 2 || count % 2);
 
 	for (i = 0; i < count / 2; i++) {
-		amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.nop, 0));
+		amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.nop, 0));
 		amdgpu_ring_write(ring, 0);
 	}
 }
@@ -1427,25 +1538,25 @@ void vcn_v2_0_dec_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 seq,
 	struct amdgpu_device *adev = ring->adev;
 
 	WARN_ON(flags & AMDGPU_FENCE_FLAG_64BIT);
-	amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.context_id, 0));
+	amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.context_id, 0));
 	amdgpu_ring_write(ring, seq);
 
-	amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.data0, 0));
+	amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.data0, 0));
 	amdgpu_ring_write(ring, addr & 0xffffffff);
 
-	amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.data1, 0));
+	amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.data1, 0));
 	amdgpu_ring_write(ring, upper_32_bits(addr) & 0xff);
 
-	amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.cmd, 0));
+	amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.cmd, 0));
 	amdgpu_ring_write(ring, VCN_DEC_KMD_CMD | (VCN_DEC_CMD_FENCE << 1));
 
-	amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.data0, 0));
+	amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.data0, 0));
 	amdgpu_ring_write(ring, 0);
 
-	amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.data1, 0));
+	amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.data1, 0));
 	amdgpu_ring_write(ring, 0);
 
-	amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.cmd, 0));
+	amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.cmd, 0));
 
 	amdgpu_ring_write(ring, VCN_DEC_KMD_CMD | (VCN_DEC_CMD_TRAP << 1));
 }
@@ -1468,14 +1579,14 @@ void vcn_v2_0_dec_ring_emit_ib(struct amdgpu_ring *ring,
 	struct amdgpu_device *adev = ring->adev;
 	unsigned vmid = AMDGPU_JOB_GET_VMID(job);
 
-	amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.ib_vmid, 0));
+	amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.ib_vmid, 0));
 	amdgpu_ring_write(ring, vmid);
 
-	amdgpu_ring_write(ring,	PACKET0(adev->vcn.internal.ib_bar_low, 0));
+	amdgpu_ring_write(ring,	PACKET0(adev->vcn.inst[ring->me].internal.ib_bar_low, 0));
 	amdgpu_ring_write(ring, lower_32_bits(ib->gpu_addr));
-	amdgpu_ring_write(ring,	PACKET0(adev->vcn.internal.ib_bar_high, 0));
+	amdgpu_ring_write(ring,	PACKET0(adev->vcn.inst[ring->me].internal.ib_bar_high, 0));
 	amdgpu_ring_write(ring, upper_32_bits(ib->gpu_addr));
-	amdgpu_ring_write(ring,	PACKET0(adev->vcn.internal.ib_size, 0));
+	amdgpu_ring_write(ring,	PACKET0(adev->vcn.inst[ring->me].internal.ib_size, 0));
 	amdgpu_ring_write(ring, ib->length_dw);
 }
 
@@ -1484,16 +1595,16 @@ void vcn_v2_0_dec_ring_emit_reg_wait(struct amdgpu_ring *ring, uint32_t reg,
 {
 	struct amdgpu_device *adev = ring->adev;
 
-	amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.data0, 0));
+	amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.data0, 0));
 	amdgpu_ring_write(ring, reg << 2);
 
-	amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.data1, 0));
+	amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.data1, 0));
 	amdgpu_ring_write(ring, val);
 
-	amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.gp_scratch8, 0));
+	amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.gp_scratch8, 0));
 	amdgpu_ring_write(ring, mask);
 
-	amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.cmd, 0));
+	amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.cmd, 0));
 
 	amdgpu_ring_write(ring, VCN_DEC_KMD_CMD | (VCN_DEC_CMD_REG_READ_COND_WAIT << 1));
 }
@@ -1501,7 +1612,7 @@ void vcn_v2_0_dec_ring_emit_reg_wait(struct amdgpu_ring *ring, uint32_t reg,
 void vcn_v2_0_dec_ring_emit_vm_flush(struct amdgpu_ring *ring,
 				unsigned vmid, uint64_t pd_addr)
 {
-	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
+	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->vm_hub];
 	uint32_t data0, data1, mask;
 
 	pd_addr = amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pd_addr);
@@ -1518,13 +1629,13 @@ void vcn_v2_0_dec_ring_emit_wreg(struct amdgpu_ring *ring,
 {
 	struct amdgpu_device *adev = ring->adev;
 
-	amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.data0, 0));
+	amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.data0, 0));
 	amdgpu_ring_write(ring, reg << 2);
 
-	amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.data1, 0));
+	amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.data1, 0));
 	amdgpu_ring_write(ring, val);
 
-	amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.cmd, 0));
+	amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.cmd, 0));
 
 	amdgpu_ring_write(ring, VCN_DEC_KMD_CMD | (VCN_DEC_CMD_WRITE_REG << 1));
 }
@@ -1559,12 +1670,12 @@ static uint64_t vcn_v2_0_enc_ring_get_wptr(struct amdgpu_ring *ring)
 
 	if (ring == &adev->vcn.inst->ring_enc[0]) {
 		if (ring->use_doorbell)
-			return adev->wb.wb[ring->wptr_offs];
+			return *ring->wptr_cpu_addr;
 		else
 			return RREG32_SOC15(UVD, 0, mmUVD_RB_WPTR);
 	} else {
 		if (ring->use_doorbell)
-			return adev->wb.wb[ring->wptr_offs];
+			return *ring->wptr_cpu_addr;
 		else
 			return RREG32_SOC15(UVD, 0, mmUVD_RB_WPTR2);
 	}
@@ -1583,14 +1694,14 @@ static void vcn_v2_0_enc_ring_set_wptr(struct amdgpu_ring *ring)
 
 	if (ring == &adev->vcn.inst->ring_enc[0]) {
 		if (ring->use_doorbell) {
-			adev->wb.wb[ring->wptr_offs] = lower_32_bits(ring->wptr);
+			*ring->wptr_cpu_addr = lower_32_bits(ring->wptr);
 			WDOORBELL32(ring->doorbell_index, lower_32_bits(ring->wptr));
 		} else {
 			WREG32_SOC15(UVD, 0, mmUVD_RB_WPTR, lower_32_bits(ring->wptr));
 		}
 	} else {
 		if (ring->use_doorbell) {
-			adev->wb.wb[ring->wptr_offs] = lower_32_bits(ring->wptr);
+			*ring->wptr_cpu_addr = lower_32_bits(ring->wptr);
 			WDOORBELL32(ring->doorbell_index, lower_32_bits(ring->wptr));
 		} else {
 			WREG32_SOC15(UVD, 0, mmUVD_RB_WPTR2, lower_32_bits(ring->wptr));
@@ -1661,7 +1772,7 @@ void vcn_v2_0_enc_ring_emit_reg_wait(struct amdgpu_ring *ring, uint32_t reg,
 void vcn_v2_0_enc_ring_emit_vm_flush(struct amdgpu_ring *ring,
 				unsigned int vmid, uint64_t pd_addr)
 {
-	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
+	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->vm_hub];
 
 	pd_addr = amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pd_addr);
 
@@ -1725,9 +1836,9 @@ int vcn_v2_0_dec_ring_test_ring(struct amdgpu_ring *ring)
 	r = amdgpu_ring_alloc(ring, 4);
 	if (r)
 		return r;
-	amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.cmd, 0));
+	amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.cmd, 0));
 	amdgpu_ring_write(ring, VCN_DEC_KMD_CMD | (VCN_DEC_CMD_PACKET_START << 1));
-	amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.scratch9, 0));
+	amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.scratch9, 0));
 	amdgpu_ring_write(ring, 0xDEADBEEF);
 	amdgpu_ring_commit(ring);
 	for (i = 0; i < adev->usec_timeout; i++) {
@@ -1744,8 +1855,8 @@ int vcn_v2_0_dec_ring_test_ring(struct amdgpu_ring *ring)
 }
 
 
-static int vcn_v2_0_set_powergating_state(void *handle,
-					  enum amd_powergating_state state)
+static int vcn_v2_0_set_pg_state(struct amdgpu_vcn_inst *vinst,
+				 enum amd_powergating_state state)
 {
 	/* This doesn't actually powergate the VCN block.
 	 * That's done in the dpm code via the SMC.  This
@@ -1755,23 +1866,24 @@ static int vcn_v2_0_set_powergating_state(void *handle,
 	 * the smc and the hw blocks
 	 */
 	int ret;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = vinst->adev;
 
 	if (amdgpu_sriov_vf(adev)) {
-		adev->vcn.cur_state = AMD_PG_STATE_UNGATE;
+		vinst->cur_state = AMD_PG_STATE_UNGATE;
 		return 0;
 	}
 
-	if (state == adev->vcn.cur_state)
+	if (state == vinst->cur_state)
 		return 0;
 
 	if (state == AMD_PG_STATE_GATE)
-		ret = vcn_v2_0_stop(adev);
+		ret = vcn_v2_0_stop(vinst);
 	else
-		ret = vcn_v2_0_start(adev);
+		ret = vcn_v2_0_start(vinst);
 
 	if (!ret)
-		adev->vcn.cur_state = state;
+		vinst->cur_state = state;
+
 	return ret;
 }
 
@@ -1810,7 +1922,7 @@ static int vcn_v2_0_start_mmsch(struct amdgpu_device *adev,
 	adev->vcn.inst->ring_dec.wptr_old = 0;
 	vcn_v2_0_dec_ring_set_wptr(&adev->vcn.inst->ring_dec);
 
-	for (i = 0; i < adev->vcn.num_enc_rings; ++i) {
+	for (i = 0; i < adev->vcn.inst[0].num_enc_rings; ++i) {
 		adev->vcn.inst->ring_enc[i].wptr = 0;
 		adev->vcn.inst->ring_enc[i].wptr_old = 0;
 		vcn_v2_0_enc_ring_set_wptr(&adev->vcn.inst->ring_enc[i]);
@@ -1868,7 +1980,7 @@ static int vcn_v2_0_start_sriov(struct amdgpu_device *adev)
 
 		init_table += header->vcn_table_offset;
 
-		size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.fw->size + 4);
+		size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.inst[0].fw->size + 4);
 
 		MMSCH_V2_0_INSERT_DIRECT_RD_MOD_WT(
 			SOC15_REG_OFFSET(UVD, i, mmUVD_STATUS),
@@ -1936,7 +2048,7 @@ static int vcn_v2_0_start_sriov(struct amdgpu_device *adev)
 			SOC15_REG_OFFSET(UVD, i, mmUVD_VCPU_CACHE_SIZE2),
 			AMDGPU_VCN_CONTEXT_SIZE);
 
-		for (r = 0; r < adev->vcn.num_enc_rings; ++r) {
+		for (r = 0; r < adev->vcn.inst[0].num_enc_rings; ++r) {
 			ring = &adev->vcn.inst->ring_enc[r];
 			ring->wptr = 0;
 			MMSCH_V2_0_INSERT_DIRECT_WT(
@@ -1983,7 +2095,6 @@ static int vcn_v2_0_start_sriov(struct amdgpu_device *adev)
 static const struct amd_ip_funcs vcn_v2_0_ip_funcs = {
 	.name = "vcn_v2_0",
 	.early_init = vcn_v2_0_early_init,
-	.late_init = NULL,
 	.sw_init = vcn_v2_0_sw_init,
 	.sw_fini = vcn_v2_0_sw_fini,
 	.hw_init = vcn_v2_0_hw_init,
@@ -1992,18 +2103,16 @@ static const struct amd_ip_funcs vcn_v2_0_ip_funcs = {
 	.resume = vcn_v2_0_resume,
 	.is_idle = vcn_v2_0_is_idle,
 	.wait_for_idle = vcn_v2_0_wait_for_idle,
-	.check_soft_reset = NULL,
-	.pre_soft_reset = NULL,
-	.soft_reset = NULL,
-	.post_soft_reset = NULL,
 	.set_clockgating_state = vcn_v2_0_set_clockgating_state,
-	.set_powergating_state = vcn_v2_0_set_powergating_state,
+	.set_powergating_state = vcn_set_powergating_state,
+	.dump_ip_state = amdgpu_vcn_dump_ip_state,
+	.print_ip_state = amdgpu_vcn_print_ip_state,
 };
 
 static const struct amdgpu_ring_funcs vcn_v2_0_dec_ring_vm_funcs = {
 	.type = AMDGPU_RING_TYPE_VCN_DEC,
 	.align_mask = 0xf,
-	.vmhub = AMDGPU_MMHUB_0,
+	.secure_submission_supported = true,
 	.get_rptr = vcn_v2_0_dec_ring_get_rptr,
 	.get_wptr = vcn_v2_0_dec_ring_get_wptr,
 	.set_wptr = vcn_v2_0_dec_ring_set_wptr,
@@ -2028,13 +2137,13 @@ static const struct amdgpu_ring_funcs vcn_v2_0_dec_ring_vm_funcs = {
 	.emit_wreg = vcn_v2_0_dec_ring_emit_wreg,
 	.emit_reg_wait = vcn_v2_0_dec_ring_emit_reg_wait,
 	.emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper,
+	.reset = amdgpu_vcn_ring_reset,
 };
 
 static const struct amdgpu_ring_funcs vcn_v2_0_enc_ring_vm_funcs = {
 	.type = AMDGPU_RING_TYPE_VCN_ENC,
 	.align_mask = 0x3f,
 	.nop = VCN_ENC_CMD_NO_OP,
-	.vmhub = AMDGPU_MMHUB_0,
 	.get_rptr = vcn_v2_0_enc_ring_get_rptr,
 	.get_wptr = vcn_v2_0_enc_ring_get_wptr,
 	.set_wptr = vcn_v2_0_enc_ring_set_wptr,
@@ -2058,22 +2167,20 @@ static const struct amdgpu_ring_funcs vcn_v2_0_enc_ring_vm_funcs = {
 	.emit_wreg = vcn_v2_0_enc_ring_emit_wreg,
 	.emit_reg_wait = vcn_v2_0_enc_ring_emit_reg_wait,
 	.emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper,
+	.reset = amdgpu_vcn_ring_reset,
 };
 
 static void vcn_v2_0_set_dec_ring_funcs(struct amdgpu_device *adev)
 {
 	adev->vcn.inst->ring_dec.funcs = &vcn_v2_0_dec_ring_vm_funcs;
-	DRM_INFO("VCN decode is enabled in VM mode\n");
 }
 
 static void vcn_v2_0_set_enc_ring_funcs(struct amdgpu_device *adev)
 {
 	int i;
 
-	for (i = 0; i < adev->vcn.num_enc_rings; ++i)
+	for (i = 0; i < adev->vcn.inst[0].num_enc_rings; ++i)
 		adev->vcn.inst->ring_enc[i].funcs = &vcn_v2_0_enc_ring_vm_funcs;
-
-	DRM_INFO("VCN encode is enabled in VM mode\n");
 }
 
 static const struct amdgpu_irq_src_funcs vcn_v2_0_irq_funcs = {
@@ -2083,7 +2190,7 @@ static const struct amdgpu_irq_src_funcs vcn_v2_0_irq_funcs = {
 
 static void vcn_v2_0_set_irq_funcs(struct amdgpu_device *adev)
 {
-	adev->vcn.inst->irq.num_types = adev->vcn.num_enc_rings + 1;
+	adev->vcn.inst->irq.num_types = adev->vcn.inst[0].num_enc_rings + 1;
 	adev->vcn.inst->irq.funcs = &vcn_v2_0_irq_funcs;
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c b/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c
index 44fc4c218433..cebee453871c 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c
@@ -31,11 +31,16 @@
 #include "soc15d.h"
 #include "vcn_v2_0.h"
 #include "mmsch_v1_0.h"
+#include "vcn_v2_5.h"
 
 #include "vcn/vcn_2_5_offset.h"
 #include "vcn/vcn_2_5_sh_mask.h"
 #include "ivsrcid/vcn/irqsrcs_vcn_2_0.h"
 
+#define VCN_VID_SOC_ADDRESS_2_0					0x1fa00
+#define VCN1_VID_SOC_ADDRESS_3_0				0x48200
+#define VCN1_AON_SOC_ADDRESS_3_0				0x48000
+
 #define mmUVD_CONTEXT_ID_INTERNAL_OFFSET			0x27
 #define mmUVD_GPCOM_VCPU_CMD_INTERNAL_OFFSET			0x0f
 #define mmUVD_GPCOM_VCPU_DATA0_INTERNAL_OFFSET			0x10
@@ -51,35 +56,181 @@
 
 #define VCN25_MAX_HW_INSTANCES_ARCTURUS			2
 
+static const struct amdgpu_hwip_reg_entry vcn_reg_list_2_5[] = {
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_POWER_STATUS),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_POWER_STATUS),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_STATUS),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_CONTEXT_ID),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_CONTEXT_ID2),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_GPCOM_VCPU_DATA0),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_GPCOM_VCPU_DATA1),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_GPCOM_VCPU_CMD),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_BASE_HI),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_BASE_LO),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_BASE_HI2),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_BASE_LO2),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_BASE_HI3),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_BASE_LO3),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_BASE_HI4),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_BASE_LO4),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_RPTR),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_WPTR),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_RPTR2),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_WPTR2),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_RPTR3),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_WPTR3),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_RPTR4),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_WPTR4),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_SIZE),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_SIZE2),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_SIZE3),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_SIZE4),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_PGFSM_CONFIG),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_PGFSM_STATUS),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_DPG_LMA_CTL),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_DPG_LMA_DATA),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_DPG_LMA_MASK),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_DPG_PAUSE)
+};
+
 static void vcn_v2_5_set_dec_ring_funcs(struct amdgpu_device *adev);
 static void vcn_v2_5_set_enc_ring_funcs(struct amdgpu_device *adev);
 static void vcn_v2_5_set_irq_funcs(struct amdgpu_device *adev);
-static int vcn_v2_5_set_powergating_state(void *handle,
-				enum amd_powergating_state state);
-static int vcn_v2_5_pause_dpg_mode(struct amdgpu_device *adev,
-				int inst_idx, struct dpg_pause_state *new_state);
+static int vcn_v2_5_set_pg_state(struct amdgpu_vcn_inst *vinst,
+				 enum amd_powergating_state state);
+static int vcn_v2_5_pause_dpg_mode(struct amdgpu_vcn_inst *vinst,
+				   struct dpg_pause_state *new_state);
 static int vcn_v2_5_sriov_start(struct amdgpu_device *adev);
+static void vcn_v2_5_set_ras_funcs(struct amdgpu_device *adev);
+static int vcn_v2_5_reset(struct amdgpu_vcn_inst *vinst);
 
 static int amdgpu_ih_clientid_vcns[] = {
 	SOC15_IH_CLIENTID_VCN,
 	SOC15_IH_CLIENTID_VCN1
 };
 
+static void vcn_v2_5_idle_work_handler(struct work_struct *work)
+{
+	struct amdgpu_vcn_inst *vcn_inst =
+		container_of(work, struct amdgpu_vcn_inst, idle_work.work);
+	struct amdgpu_device *adev = vcn_inst->adev;
+	unsigned int fences = 0, fence[AMDGPU_MAX_VCN_INSTANCES] = {0};
+	unsigned int i, j;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		struct amdgpu_vcn_inst *v = &adev->vcn.inst[i];
+
+		if (adev->vcn.harvest_config & (1 << i))
+			continue;
+
+		for (j = 0; j < v->num_enc_rings; ++j)
+			fence[i] += amdgpu_fence_count_emitted(&v->ring_enc[j]);
+
+		/* Only set DPG pause for VCN3 or below, VCN4 and above will be handled by FW */
+		if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG &&
+		    !v->using_unified_queue) {
+			struct dpg_pause_state new_state;
+
+			if (fence[i] ||
+			    unlikely(atomic_read(&v->dpg_enc_submission_cnt)))
+				new_state.fw_based = VCN_DPG_STATE__PAUSE;
+			else
+				new_state.fw_based = VCN_DPG_STATE__UNPAUSE;
+
+			v->pause_dpg_mode(v, &new_state);
+		}
+
+		fence[i] += amdgpu_fence_count_emitted(&v->ring_dec);
+		fences += fence[i];
+
+	}
+
+	if (!fences && !atomic_read(&adev->vcn.inst[0].total_submission_cnt)) {
+		amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_VCN,
+						       AMD_PG_STATE_GATE);
+		amdgpu_vcn_put_profile(adev);
+	} else {
+		schedule_delayed_work(&adev->vcn.inst[0].idle_work, VCN_IDLE_TIMEOUT);
+	}
+}
+
+static void vcn_v2_5_ring_begin_use(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct amdgpu_vcn_inst *v = &adev->vcn.inst[ring->me];
+
+	atomic_inc(&adev->vcn.inst[0].total_submission_cnt);
+
+	cancel_delayed_work_sync(&adev->vcn.inst[0].idle_work);
+
+	/* We can safely return early here because we've cancelled the
+	 * the delayed work so there is no one else to set it to false
+	 * and we don't care if someone else sets it to true.
+	 */
+	mutex_lock(&adev->vcn.inst[0].vcn_pg_lock);
+	amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_VCN,
+					       AMD_PG_STATE_UNGATE);
+
+	/* Only set DPG pause for VCN3 or below, VCN4 and above will be handled by FW */
+	if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG &&
+	    !v->using_unified_queue) {
+		struct dpg_pause_state new_state;
+
+		if (ring->funcs->type == AMDGPU_RING_TYPE_VCN_ENC) {
+			atomic_inc(&v->dpg_enc_submission_cnt);
+			new_state.fw_based = VCN_DPG_STATE__PAUSE;
+		} else {
+			unsigned int fences = 0;
+			unsigned int i;
+
+			for (i = 0; i < v->num_enc_rings; ++i)
+				fences += amdgpu_fence_count_emitted(&v->ring_enc[i]);
+
+			if (fences || atomic_read(&v->dpg_enc_submission_cnt))
+				new_state.fw_based = VCN_DPG_STATE__PAUSE;
+			else
+				new_state.fw_based = VCN_DPG_STATE__UNPAUSE;
+		}
+		v->pause_dpg_mode(v, &new_state);
+	}
+	mutex_unlock(&adev->vcn.inst[0].vcn_pg_lock);
+	amdgpu_vcn_get_profile(adev);
+}
+
+static void vcn_v2_5_ring_end_use(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	/* Only set DPG pause for VCN3 or below, VCN4 and above will be handled by FW */
+	if (ring->adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG &&
+	    ring->funcs->type == AMDGPU_RING_TYPE_VCN_ENC &&
+	    !adev->vcn.inst[ring->me].using_unified_queue)
+		atomic_dec(&adev->vcn.inst[ring->me].dpg_enc_submission_cnt);
+
+	atomic_dec(&adev->vcn.inst[0].total_submission_cnt);
+
+	schedule_delayed_work(&adev->vcn.inst[0].idle_work,
+			      VCN_IDLE_TIMEOUT);
+}
+
 /**
- * vcn_v2_5_early_init - set function pointers
+ * vcn_v2_5_early_init - set function pointers and load microcode
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Set ring and irq function pointers
+ * Load microcode from filesystem
  */
-static int vcn_v2_5_early_init(void *handle)
+static int vcn_v2_5_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, r;
 
 	if (amdgpu_sriov_vf(adev)) {
 		adev->vcn.num_vcn_inst = 2;
 		adev->vcn.harvest_config = 0;
-		adev->vcn.num_enc_rings = 1;
+		for (i = 0; i < adev->vcn.num_vcn_inst; i++)
+			adev->vcn.inst[i].num_enc_rings = 1;
 	} else {
 		u32 harvest;
 		int i;
@@ -88,18 +239,26 @@ static int vcn_v2_5_early_init(void *handle)
 			harvest = RREG32_SOC15(VCN, i, mmCC_UVD_HARVESTING);
 			if (harvest & CC_UVD_HARVESTING__UVD_DISABLE_MASK)
 				adev->vcn.harvest_config |= 1 << i;
+			adev->vcn.inst[i].num_enc_rings = 2;
 		}
 		if (adev->vcn.harvest_config == (AMDGPU_VCN_HARVEST_VCN0 |
-					AMDGPU_VCN_HARVEST_VCN1))
+						 AMDGPU_VCN_HARVEST_VCN1))
 			/* both instances are harvested, disable the block */
 			return -ENOENT;
-
-		adev->vcn.num_enc_rings = 2;
 	}
 
 	vcn_v2_5_set_dec_ring_funcs(adev);
 	vcn_v2_5_set_enc_ring_funcs(adev);
 	vcn_v2_5_set_irq_funcs(adev);
+	vcn_v2_5_set_ras_funcs(adev);
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+		adev->vcn.inst[i].set_pg_state = vcn_v2_5_set_pg_state;
+
+		r = amdgpu_vcn_early_init(adev, i);
+		if (r)
+			return r;
+	}
 
 	return 0;
 }
@@ -107,17 +266,19 @@ static int vcn_v2_5_early_init(void *handle)
 /**
  * vcn_v2_5_sw_init - sw init for VCN block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Load firmware and sw initialization
  */
-static int vcn_v2_5_sw_init(void *handle)
+static int vcn_v2_5_sw_init(struct amdgpu_ip_block *ip_block)
 {
 	struct amdgpu_ring *ring;
 	int i, j, r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	for (j = 0; j < adev->vcn.num_vcn_inst; j++) {
+		struct amdgpu_fw_shared *fw_shared;
+
 		if (adev->vcn.harvest_config & (1 << j))
 			continue;
 		/* VCN DEC TRAP */
@@ -127,45 +288,48 @@ static int vcn_v2_5_sw_init(void *handle)
 			return r;
 
 		/* VCN ENC TRAP */
-		for (i = 0; i < adev->vcn.num_enc_rings; ++i) {
+		for (i = 0; i < adev->vcn.inst[j].num_enc_rings; ++i) {
 			r = amdgpu_irq_add_id(adev, amdgpu_ih_clientid_vcns[j],
 				i + VCN_2_0__SRCID__UVD_ENC_GENERAL_PURPOSE, &adev->vcn.inst[j].irq);
 			if (r)
 				return r;
 		}
-	}
 
-	r = amdgpu_vcn_sw_init(adev);
-	if (r)
-		return r;
+		/* VCN POISON TRAP */
+		r = amdgpu_irq_add_id(adev, amdgpu_ih_clientid_vcns[j],
+			VCN_2_6__SRCID_UVD_POISON, &adev->vcn.inst[j].ras_poison_irq);
+		if (r)
+			return r;
 
-	amdgpu_vcn_setup_ucode(adev);
+		r = amdgpu_vcn_sw_init(adev, j);
+		if (r)
+			return r;
 
-	r = amdgpu_vcn_resume(adev);
-	if (r)
-		return r;
+		/* Override the work func */
+		adev->vcn.inst[j].idle_work.work.func = vcn_v2_5_idle_work_handler;
 
-	for (j = 0; j < adev->vcn.num_vcn_inst; j++) {
-		volatile struct amdgpu_fw_shared *fw_shared;
+		amdgpu_vcn_setup_ucode(adev, j);
 
-		if (adev->vcn.harvest_config & (1 << j))
-			continue;
-		adev->vcn.internal.context_id = mmUVD_CONTEXT_ID_INTERNAL_OFFSET;
-		adev->vcn.internal.ib_vmid = mmUVD_LMI_RBC_IB_VMID_INTERNAL_OFFSET;
-		adev->vcn.internal.ib_bar_low = mmUVD_LMI_RBC_IB_64BIT_BAR_LOW_INTERNAL_OFFSET;
-		adev->vcn.internal.ib_bar_high = mmUVD_LMI_RBC_IB_64BIT_BAR_HIGH_INTERNAL_OFFSET;
-		adev->vcn.internal.ib_size = mmUVD_RBC_IB_SIZE_INTERNAL_OFFSET;
-		adev->vcn.internal.gp_scratch8 = mmUVD_GP_SCRATCH8_INTERNAL_OFFSET;
-
-		adev->vcn.internal.scratch9 = mmUVD_SCRATCH9_INTERNAL_OFFSET;
+		r = amdgpu_vcn_resume(adev, j);
+		if (r)
+			return r;
+
+		adev->vcn.inst[j].internal.context_id = mmUVD_CONTEXT_ID_INTERNAL_OFFSET;
+		adev->vcn.inst[j].internal.ib_vmid = mmUVD_LMI_RBC_IB_VMID_INTERNAL_OFFSET;
+		adev->vcn.inst[j].internal.ib_bar_low = mmUVD_LMI_RBC_IB_64BIT_BAR_LOW_INTERNAL_OFFSET;
+		adev->vcn.inst[j].internal.ib_bar_high = mmUVD_LMI_RBC_IB_64BIT_BAR_HIGH_INTERNAL_OFFSET;
+		adev->vcn.inst[j].internal.ib_size = mmUVD_RBC_IB_SIZE_INTERNAL_OFFSET;
+		adev->vcn.inst[j].internal.gp_scratch8 = mmUVD_GP_SCRATCH8_INTERNAL_OFFSET;
+
+		adev->vcn.inst[j].internal.scratch9 = mmUVD_SCRATCH9_INTERNAL_OFFSET;
 		adev->vcn.inst[j].external.scratch9 = SOC15_REG_OFFSET(VCN, j, mmUVD_SCRATCH9);
-		adev->vcn.internal.data0 = mmUVD_GPCOM_VCPU_DATA0_INTERNAL_OFFSET;
+		adev->vcn.inst[j].internal.data0 = mmUVD_GPCOM_VCPU_DATA0_INTERNAL_OFFSET;
 		adev->vcn.inst[j].external.data0 = SOC15_REG_OFFSET(VCN, j, mmUVD_GPCOM_VCPU_DATA0);
-		adev->vcn.internal.data1 = mmUVD_GPCOM_VCPU_DATA1_INTERNAL_OFFSET;
+		adev->vcn.inst[j].internal.data1 = mmUVD_GPCOM_VCPU_DATA1_INTERNAL_OFFSET;
 		adev->vcn.inst[j].external.data1 = SOC15_REG_OFFSET(VCN, j, mmUVD_GPCOM_VCPU_DATA1);
-		adev->vcn.internal.cmd = mmUVD_GPCOM_VCPU_CMD_INTERNAL_OFFSET;
+		adev->vcn.inst[j].internal.cmd = mmUVD_GPCOM_VCPU_CMD_INTERNAL_OFFSET;
 		adev->vcn.inst[j].external.cmd = SOC15_REG_OFFSET(VCN, j, mmUVD_GPCOM_VCPU_CMD);
-		adev->vcn.internal.nop = mmUVD_NO_OP_INTERNAL_OFFSET;
+		adev->vcn.inst[j].internal.nop = mmUVD_NO_OP_INTERNAL_OFFSET;
 		adev->vcn.inst[j].external.nop = SOC15_REG_OFFSET(VCN, j, mmUVD_NO_OP);
 
 		ring = &adev->vcn.inst[j].ring_dec;
@@ -173,13 +337,19 @@ static int vcn_v2_5_sw_init(void *handle)
 
 		ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) +
 				(amdgpu_sriov_vf(adev) ? 2*j : 8*j);
+
+		if (amdgpu_ip_version(adev, UVD_HWIP, 0) == IP_VERSION(2, 5, 0))
+			ring->vm_hub = AMDGPU_MMHUB1(0);
+		else
+			ring->vm_hub = AMDGPU_MMHUB0(0);
+
 		sprintf(ring->name, "vcn_dec_%d", j);
 		r = amdgpu_ring_init(adev, ring, 512, &adev->vcn.inst[j].irq,
 				     0, AMDGPU_RING_PRIO_DEFAULT, NULL);
 		if (r)
 			return r;
 
-		for (i = 0; i < adev->vcn.num_enc_rings; ++i) {
+		for (i = 0; i < adev->vcn.inst[j].num_enc_rings; ++i) {
 			enum amdgpu_ring_priority_level hw_prio = amdgpu_vcn_get_enc_ring_prio(i);
 
 			ring = &adev->vcn.inst[j].ring_enc[i];
@@ -188,6 +358,12 @@ static int vcn_v2_5_sw_init(void *handle)
 			ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) +
 					(amdgpu_sriov_vf(adev) ? (1 + i + 2*j) : (2 + i + 8*j));
 
+			if (amdgpu_ip_version(adev, UVD_HWIP, 0) ==
+			    IP_VERSION(2, 5, 0))
+				ring->vm_hub = AMDGPU_MMHUB1(0);
+			else
+				ring->vm_hub = AMDGPU_MMHUB0(0);
+
 			sprintf(ring->name, "vcn_enc_%d.%d", j, i);
 			r = amdgpu_ring_init(adev, ring, 512,
 					     &adev->vcn.inst[j].irq, 0,
@@ -196,18 +372,39 @@ static int vcn_v2_5_sw_init(void *handle)
 				return r;
 		}
 
-		fw_shared = adev->vcn.inst[j].fw_shared_cpu_addr;
+		fw_shared = adev->vcn.inst[j].fw_shared.cpu_addr;
 		fw_shared->present_flag_0 = cpu_to_le32(AMDGPU_VCN_MULTI_QUEUE_FLAG);
+
+		if (amdgpu_vcnfw_log)
+			amdgpu_vcn_fwlog_init(&adev->vcn.inst[i]);
+
+		if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)
+			adev->vcn.inst[j].pause_dpg_mode = vcn_v2_5_pause_dpg_mode;
+		adev->vcn.inst[j].reset = vcn_v2_5_reset;
 	}
 
+	adev->vcn.supported_reset =
+		amdgpu_get_soft_full_reset_mask(&adev->vcn.inst[0].ring_enc[0]);
+	if (!amdgpu_sriov_vf(adev))
+		adev->vcn.supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE;
+
 	if (amdgpu_sriov_vf(adev)) {
 		r = amdgpu_virt_alloc_mm_table(adev);
 		if (r)
 			return r;
 	}
 
-	if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)
-		adev->vcn.pause_dpg_mode = vcn_v2_5_pause_dpg_mode;
+	r = amdgpu_vcn_ras_sw_init(adev);
+	if (r)
+		return r;
+
+	r = amdgpu_vcn_reg_dump_init(adev, vcn_reg_list_2_5, ARRAY_SIZE(vcn_reg_list_2_5));
+	if (r)
+		return r;
+
+	r = amdgpu_vcn_sysfs_reset_mask_init(adev);
+	if (r)
+		return r;
 
 	return 0;
 }
@@ -215,21 +412,21 @@ static int vcn_v2_5_sw_init(void *handle)
 /**
  * vcn_v2_5_sw_fini - sw fini for VCN block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * VCN suspend and free up sw allocation
  */
-static int vcn_v2_5_sw_fini(void *handle)
+static int vcn_v2_5_sw_fini(struct amdgpu_ip_block *ip_block)
 {
 	int i, r, idx;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-	volatile struct amdgpu_fw_shared *fw_shared;
+	struct amdgpu_device *adev = ip_block->adev;
+	struct amdgpu_fw_shared *fw_shared;
 
 	if (drm_dev_enter(adev_to_drm(adev), &idx)) {
 		for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
 			if (adev->vcn.harvest_config & (1 << i))
 				continue;
-			fw_shared = adev->vcn.inst[i].fw_shared_cpu_addr;
+			fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr;
 			fw_shared->present_flag_0 = 0;
 		}
 		drm_dev_exit(idx);
@@ -239,25 +436,28 @@ static int vcn_v2_5_sw_fini(void *handle)
 	if (amdgpu_sriov_vf(adev))
 		amdgpu_virt_free_mm_table(adev);
 
-	r = amdgpu_vcn_suspend(adev);
-	if (r)
-		return r;
+	amdgpu_vcn_sysfs_reset_mask_fini(adev);
 
-	r = amdgpu_vcn_sw_fini(adev);
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+		r = amdgpu_vcn_suspend(adev, i);
+		if (r)
+			return r;
+		amdgpu_vcn_sw_fini(adev, i);
+	}
 
-	return r;
+	return 0;
 }
 
 /**
  * vcn_v2_5_hw_init - start and test VCN block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Initialize the hardware, boot up the VCPU and do some testing
  */
-static int vcn_v2_5_hw_init(void *handle)
+static int vcn_v2_5_hw_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	struct amdgpu_ring *ring;
 	int i, j, r = 0;
 
@@ -282,47 +482,47 @@ static int vcn_v2_5_hw_init(void *handle)
 
 			r = amdgpu_ring_test_helper(ring);
 			if (r)
-				goto done;
+				return r;
 
-			for (i = 0; i < adev->vcn.num_enc_rings; ++i) {
+			for (i = 0; i < adev->vcn.inst[j].num_enc_rings; ++i) {
 				ring = &adev->vcn.inst[j].ring_enc[i];
 				r = amdgpu_ring_test_helper(ring);
 				if (r)
-					goto done;
+					return r;
 			}
 		}
 	}
 
-done:
-	if (!r)
-		DRM_INFO("VCN decode and encode initialized successfully(under %s).\n",
-			(adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)?"DPG Mode":"SPG Mode");
-
 	return r;
 }
 
 /**
  * vcn_v2_5_hw_fini - stop the hardware block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Stop the VCN block, mark ring as not ready any more
  */
-static int vcn_v2_5_hw_fini(void *handle)
+static int vcn_v2_5_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int i;
 
-	cancel_delayed_work_sync(&adev->vcn.idle_work);
-
 	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		struct amdgpu_vcn_inst *vinst = &adev->vcn.inst[i];
+
 		if (adev->vcn.harvest_config & (1 << i))
 			continue;
 
+		cancel_delayed_work_sync(&vinst->idle_work);
+
 		if ((adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) ||
-		    (adev->vcn.cur_state != AMD_PG_STATE_GATE &&
+		    (vinst->cur_state != AMD_PG_STATE_GATE &&
 		     RREG32_SOC15(VCN, i, mmUVD_STATUS)))
-			vcn_v2_5_set_powergating_state(adev, AMD_PG_STATE_GATE);
+			vinst->set_pg_state(vinst, AMD_PG_STATE_GATE);
+
+		if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__VCN))
+			amdgpu_irq_put(adev, &vinst->ras_poison_irq, 0);
 	}
 
 	return 0;
@@ -331,41 +531,47 @@ static int vcn_v2_5_hw_fini(void *handle)
 /**
  * vcn_v2_5_suspend - suspend VCN block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * HW fini and suspend VCN block
  */
-static int vcn_v2_5_suspend(void *handle)
+static int vcn_v2_5_suspend(struct amdgpu_ip_block *ip_block)
 {
-	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
+	int r, i;
 
-	r = vcn_v2_5_hw_fini(adev);
+	r = vcn_v2_5_hw_fini(ip_block);
 	if (r)
 		return r;
 
-	r = amdgpu_vcn_suspend(adev);
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+		r = amdgpu_vcn_suspend(ip_block->adev, i);
+		if (r)
+			return r;
+	}
 
-	return r;
+	return 0;
 }
 
 /**
  * vcn_v2_5_resume - resume VCN block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Resume firmware and hw init VCN block
  */
-static int vcn_v2_5_resume(void *handle)
+static int vcn_v2_5_resume(struct amdgpu_ip_block *ip_block)
 {
-	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
+	int r, i;
 
-	r = amdgpu_vcn_resume(adev);
-	if (r)
-		return r;
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+		r = amdgpu_vcn_resume(ip_block->adev, i);
+		if (r)
+			return r;
+	}
 
-	r = vcn_v2_5_hw_init(adev);
+	r = vcn_v2_5_hw_init(ip_block);
 
 	return r;
 }
@@ -373,68 +579,72 @@ static int vcn_v2_5_resume(void *handle)
 /**
  * vcn_v2_5_mc_resume - memory controller programming
  *
- * @adev: amdgpu_device pointer
+ * @vinst: VCN instance
  *
  * Let the VCN memory controller know it's offsets
  */
-static void vcn_v2_5_mc_resume(struct amdgpu_device *adev)
+static void vcn_v2_5_mc_resume(struct amdgpu_vcn_inst *vinst)
 {
-	uint32_t size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.fw->size + 4);
+	struct amdgpu_device *adev = vinst->adev;
+	int i = vinst->inst;
+	uint32_t size;
 	uint32_t offset;
-	int i;
 
-	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
-		if (adev->vcn.harvest_config & (1 << i))
-			continue;
-		/* cache window 0: fw */
-		if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
-			WREG32_SOC15(VCN, i, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW,
-				(adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + i].tmr_mc_addr_lo));
-			WREG32_SOC15(VCN, i, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH,
-				(adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + i].tmr_mc_addr_hi));
-			WREG32_SOC15(VCN, i, mmUVD_VCPU_CACHE_OFFSET0, 0);
-			offset = 0;
-		} else {
-			WREG32_SOC15(VCN, i, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW,
-				lower_32_bits(adev->vcn.inst[i].gpu_addr));
-			WREG32_SOC15(VCN, i, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH,
-				upper_32_bits(adev->vcn.inst[i].gpu_addr));
-			offset = size;
-			WREG32_SOC15(VCN, i, mmUVD_VCPU_CACHE_OFFSET0,
-				AMDGPU_UVD_FIRMWARE_OFFSET >> 3);
-		}
-		WREG32_SOC15(VCN, i, mmUVD_VCPU_CACHE_SIZE0, size);
+	if (adev->vcn.harvest_config & (1 << i))
+		return;
 
-		/* cache window 1: stack */
-		WREG32_SOC15(VCN, i, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW,
-			lower_32_bits(adev->vcn.inst[i].gpu_addr + offset));
-		WREG32_SOC15(VCN, i, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH,
-			upper_32_bits(adev->vcn.inst[i].gpu_addr + offset));
-		WREG32_SOC15(VCN, i, mmUVD_VCPU_CACHE_OFFSET1, 0);
-		WREG32_SOC15(VCN, i, mmUVD_VCPU_CACHE_SIZE1, AMDGPU_VCN_STACK_SIZE);
-
-		/* cache window 2: context */
-		WREG32_SOC15(VCN, i, mmUVD_LMI_VCPU_CACHE2_64BIT_BAR_LOW,
-			lower_32_bits(adev->vcn.inst[i].gpu_addr + offset + AMDGPU_VCN_STACK_SIZE));
-		WREG32_SOC15(VCN, i, mmUVD_LMI_VCPU_CACHE2_64BIT_BAR_HIGH,
-			upper_32_bits(adev->vcn.inst[i].gpu_addr + offset + AMDGPU_VCN_STACK_SIZE));
-		WREG32_SOC15(VCN, i, mmUVD_VCPU_CACHE_OFFSET2, 0);
-		WREG32_SOC15(VCN, i, mmUVD_VCPU_CACHE_SIZE2, AMDGPU_VCN_CONTEXT_SIZE);
-
-		/* non-cache window */
-		WREG32_SOC15(VCN, i, mmUVD_LMI_VCPU_NC0_64BIT_BAR_LOW,
-			lower_32_bits(adev->vcn.inst[i].fw_shared_gpu_addr));
-		WREG32_SOC15(VCN, i, mmUVD_LMI_VCPU_NC0_64BIT_BAR_HIGH,
-			upper_32_bits(adev->vcn.inst[i].fw_shared_gpu_addr));
-		WREG32_SOC15(VCN, i, mmUVD_VCPU_NONCACHE_OFFSET0, 0);
-		WREG32_SOC15(VCN, i, mmUVD_VCPU_NONCACHE_SIZE0,
-			AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_fw_shared)));
+	size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.inst[i].fw->size + 4);
+	/* cache window 0: fw */
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
+		WREG32_SOC15(VCN, i, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW,
+			     (adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + i].tmr_mc_addr_lo));
+		WREG32_SOC15(VCN, i, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH,
+			     (adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + i].tmr_mc_addr_hi));
+		WREG32_SOC15(VCN, i, mmUVD_VCPU_CACHE_OFFSET0, 0);
+		offset = 0;
+	} else {
+		WREG32_SOC15(VCN, i, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW,
+			     lower_32_bits(adev->vcn.inst[i].gpu_addr));
+		WREG32_SOC15(VCN, i, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH,
+			     upper_32_bits(adev->vcn.inst[i].gpu_addr));
+		offset = size;
+		WREG32_SOC15(VCN, i, mmUVD_VCPU_CACHE_OFFSET0,
+			     AMDGPU_UVD_FIRMWARE_OFFSET >> 3);
 	}
+	WREG32_SOC15(VCN, i, mmUVD_VCPU_CACHE_SIZE0, size);
+
+	/* cache window 1: stack */
+	WREG32_SOC15(VCN, i, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW,
+		     lower_32_bits(adev->vcn.inst[i].gpu_addr + offset));
+	WREG32_SOC15(VCN, i, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH,
+		     upper_32_bits(adev->vcn.inst[i].gpu_addr + offset));
+	WREG32_SOC15(VCN, i, mmUVD_VCPU_CACHE_OFFSET1, 0);
+	WREG32_SOC15(VCN, i, mmUVD_VCPU_CACHE_SIZE1, AMDGPU_VCN_STACK_SIZE);
+
+	/* cache window 2: context */
+	WREG32_SOC15(VCN, i, mmUVD_LMI_VCPU_CACHE2_64BIT_BAR_LOW,
+		     lower_32_bits(adev->vcn.inst[i].gpu_addr + offset + AMDGPU_VCN_STACK_SIZE));
+	WREG32_SOC15(VCN, i, mmUVD_LMI_VCPU_CACHE2_64BIT_BAR_HIGH,
+		     upper_32_bits(adev->vcn.inst[i].gpu_addr + offset + AMDGPU_VCN_STACK_SIZE));
+	WREG32_SOC15(VCN, i, mmUVD_VCPU_CACHE_OFFSET2, 0);
+	WREG32_SOC15(VCN, i, mmUVD_VCPU_CACHE_SIZE2, AMDGPU_VCN_CONTEXT_SIZE);
+
+	/* non-cache window */
+	WREG32_SOC15(VCN, i, mmUVD_LMI_VCPU_NC0_64BIT_BAR_LOW,
+		     lower_32_bits(adev->vcn.inst[i].fw_shared.gpu_addr));
+	WREG32_SOC15(VCN, i, mmUVD_LMI_VCPU_NC0_64BIT_BAR_HIGH,
+		     upper_32_bits(adev->vcn.inst[i].fw_shared.gpu_addr));
+	WREG32_SOC15(VCN, i, mmUVD_VCPU_NONCACHE_OFFSET0, 0);
+	WREG32_SOC15(VCN, i, mmUVD_VCPU_NONCACHE_SIZE0,
+		     AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_fw_shared)));
 }
 
-static void vcn_v2_5_mc_resume_dpg_mode(struct amdgpu_device *adev, int inst_idx, bool indirect)
+static void vcn_v2_5_mc_resume_dpg_mode(struct amdgpu_vcn_inst *vinst,
+					bool indirect)
 {
-	uint32_t size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.fw->size + 4);
+	struct amdgpu_device *adev = vinst->adev;
+	int inst_idx = vinst->inst;
+	uint32_t size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.inst[inst_idx].fw->size + 4);
 	uint32_t offset;
 
 	/* cache window 0: fw */
@@ -513,10 +723,10 @@ static void vcn_v2_5_mc_resume_dpg_mode(struct amdgpu_device *adev, int inst_idx
 	/* non-cache window */
 	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, 0, mmUVD_LMI_VCPU_NC0_64BIT_BAR_LOW),
-		lower_32_bits(adev->vcn.inst[inst_idx].fw_shared_gpu_addr), 0, indirect);
+		lower_32_bits(adev->vcn.inst[inst_idx].fw_shared.gpu_addr), 0, indirect);
 	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, 0, mmUVD_LMI_VCPU_NC0_64BIT_BAR_HIGH),
-		upper_32_bits(adev->vcn.inst[inst_idx].fw_shared_gpu_addr), 0, indirect);
+		upper_32_bits(adev->vcn.inst[inst_idx].fw_shared.gpu_addr), 0, indirect);
 	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, 0, mmUVD_VCPU_NONCACHE_OFFSET0), 0, 0, indirect);
 	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
@@ -531,123 +741,124 @@ static void vcn_v2_5_mc_resume_dpg_mode(struct amdgpu_device *adev, int inst_idx
 /**
  * vcn_v2_5_disable_clock_gating - disable VCN clock gating
  *
- * @adev: amdgpu_device pointer
+ * @vinst: VCN instance
  *
  * Disable clock gating for VCN block
  */
-static void vcn_v2_5_disable_clock_gating(struct amdgpu_device *adev)
+static void vcn_v2_5_disable_clock_gating(struct amdgpu_vcn_inst *vinst)
 {
+	struct amdgpu_device *adev = vinst->adev;
+	int i = vinst->inst;
 	uint32_t data;
-	int i;
 
-	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
-		if (adev->vcn.harvest_config & (1 << i))
-			continue;
-		/* UVD disable CGC */
-		data = RREG32_SOC15(VCN, i, mmUVD_CGC_CTRL);
-		if (adev->cg_flags & AMD_CG_SUPPORT_VCN_MGCG)
-			data |= 1 << UVD_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
-		else
-			data &= ~UVD_CGC_CTRL__DYN_CLOCK_MODE_MASK;
-		data |= 1 << UVD_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT;
-		data |= 4 << UVD_CGC_CTRL__CLK_OFF_DELAY__SHIFT;
-		WREG32_SOC15(VCN, i, mmUVD_CGC_CTRL, data);
-
-		data = RREG32_SOC15(VCN, i, mmUVD_CGC_GATE);
-		data &= ~(UVD_CGC_GATE__SYS_MASK
-			| UVD_CGC_GATE__UDEC_MASK
-			| UVD_CGC_GATE__MPEG2_MASK
-			| UVD_CGC_GATE__REGS_MASK
-			| UVD_CGC_GATE__RBC_MASK
-			| UVD_CGC_GATE__LMI_MC_MASK
-			| UVD_CGC_GATE__LMI_UMC_MASK
-			| UVD_CGC_GATE__IDCT_MASK
-			| UVD_CGC_GATE__MPRD_MASK
-			| UVD_CGC_GATE__MPC_MASK
-			| UVD_CGC_GATE__LBSI_MASK
-			| UVD_CGC_GATE__LRBBM_MASK
-			| UVD_CGC_GATE__UDEC_RE_MASK
-			| UVD_CGC_GATE__UDEC_CM_MASK
-			| UVD_CGC_GATE__UDEC_IT_MASK
-			| UVD_CGC_GATE__UDEC_DB_MASK
-			| UVD_CGC_GATE__UDEC_MP_MASK
-			| UVD_CGC_GATE__WCB_MASK
-			| UVD_CGC_GATE__VCPU_MASK
-			| UVD_CGC_GATE__MMSCH_MASK);
-
-		WREG32_SOC15(VCN, i, mmUVD_CGC_GATE, data);
-
-		SOC15_WAIT_ON_RREG(VCN, i, mmUVD_CGC_GATE, 0,  0xFFFFFFFF);
-
-		data = RREG32_SOC15(VCN, i, mmUVD_CGC_CTRL);
-		data &= ~(UVD_CGC_CTRL__UDEC_RE_MODE_MASK
-			| UVD_CGC_CTRL__UDEC_CM_MODE_MASK
-			| UVD_CGC_CTRL__UDEC_IT_MODE_MASK
-			| UVD_CGC_CTRL__UDEC_DB_MODE_MASK
-			| UVD_CGC_CTRL__UDEC_MP_MODE_MASK
-			| UVD_CGC_CTRL__SYS_MODE_MASK
-			| UVD_CGC_CTRL__UDEC_MODE_MASK
-			| UVD_CGC_CTRL__MPEG2_MODE_MASK
-			| UVD_CGC_CTRL__REGS_MODE_MASK
-			| UVD_CGC_CTRL__RBC_MODE_MASK
-			| UVD_CGC_CTRL__LMI_MC_MODE_MASK
-			| UVD_CGC_CTRL__LMI_UMC_MODE_MASK
-			| UVD_CGC_CTRL__IDCT_MODE_MASK
-			| UVD_CGC_CTRL__MPRD_MODE_MASK
-			| UVD_CGC_CTRL__MPC_MODE_MASK
-			| UVD_CGC_CTRL__LBSI_MODE_MASK
-			| UVD_CGC_CTRL__LRBBM_MODE_MASK
-			| UVD_CGC_CTRL__WCB_MODE_MASK
-			| UVD_CGC_CTRL__VCPU_MODE_MASK
-			| UVD_CGC_CTRL__MMSCH_MODE_MASK);
-		WREG32_SOC15(VCN, i, mmUVD_CGC_CTRL, data);
-
-		/* turn on */
-		data = RREG32_SOC15(VCN, i, mmUVD_SUVD_CGC_GATE);
-		data |= (UVD_SUVD_CGC_GATE__SRE_MASK
-			| UVD_SUVD_CGC_GATE__SIT_MASK
-			| UVD_SUVD_CGC_GATE__SMP_MASK
-			| UVD_SUVD_CGC_GATE__SCM_MASK
-			| UVD_SUVD_CGC_GATE__SDB_MASK
-			| UVD_SUVD_CGC_GATE__SRE_H264_MASK
-			| UVD_SUVD_CGC_GATE__SRE_HEVC_MASK
-			| UVD_SUVD_CGC_GATE__SIT_H264_MASK
-			| UVD_SUVD_CGC_GATE__SIT_HEVC_MASK
-			| UVD_SUVD_CGC_GATE__SCM_H264_MASK
-			| UVD_SUVD_CGC_GATE__SCM_HEVC_MASK
-			| UVD_SUVD_CGC_GATE__SDB_H264_MASK
-			| UVD_SUVD_CGC_GATE__SDB_HEVC_MASK
-			| UVD_SUVD_CGC_GATE__SCLR_MASK
-			| UVD_SUVD_CGC_GATE__UVD_SC_MASK
-			| UVD_SUVD_CGC_GATE__ENT_MASK
-			| UVD_SUVD_CGC_GATE__SIT_HEVC_DEC_MASK
-			| UVD_SUVD_CGC_GATE__SIT_HEVC_ENC_MASK
-			| UVD_SUVD_CGC_GATE__SITE_MASK
-			| UVD_SUVD_CGC_GATE__SRE_VP9_MASK
-			| UVD_SUVD_CGC_GATE__SCM_VP9_MASK
-			| UVD_SUVD_CGC_GATE__SIT_VP9_DEC_MASK
-			| UVD_SUVD_CGC_GATE__SDB_VP9_MASK
-			| UVD_SUVD_CGC_GATE__IME_HEVC_MASK);
-		WREG32_SOC15(VCN, i, mmUVD_SUVD_CGC_GATE, data);
-
-		data = RREG32_SOC15(VCN, i, mmUVD_SUVD_CGC_CTRL);
-		data &= ~(UVD_SUVD_CGC_CTRL__SRE_MODE_MASK
-			| UVD_SUVD_CGC_CTRL__SIT_MODE_MASK
-			| UVD_SUVD_CGC_CTRL__SMP_MODE_MASK
-			| UVD_SUVD_CGC_CTRL__SCM_MODE_MASK
-			| UVD_SUVD_CGC_CTRL__SDB_MODE_MASK
-			| UVD_SUVD_CGC_CTRL__SCLR_MODE_MASK
-			| UVD_SUVD_CGC_CTRL__UVD_SC_MODE_MASK
-			| UVD_SUVD_CGC_CTRL__ENT_MODE_MASK
+	if (adev->vcn.harvest_config & (1 << i))
+		return;
+	/* UVD disable CGC */
+	data = RREG32_SOC15(VCN, i, mmUVD_CGC_CTRL);
+	if (adev->cg_flags & AMD_CG_SUPPORT_VCN_MGCG)
+		data |= 1 << UVD_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
+	else
+		data &= ~UVD_CGC_CTRL__DYN_CLOCK_MODE_MASK;
+	data |= 1 << UVD_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT;
+	data |= 4 << UVD_CGC_CTRL__CLK_OFF_DELAY__SHIFT;
+	WREG32_SOC15(VCN, i, mmUVD_CGC_CTRL, data);
+
+	data = RREG32_SOC15(VCN, i, mmUVD_CGC_GATE);
+	data &= ~(UVD_CGC_GATE__SYS_MASK
+		  | UVD_CGC_GATE__UDEC_MASK
+		  | UVD_CGC_GATE__MPEG2_MASK
+		  | UVD_CGC_GATE__REGS_MASK
+		  | UVD_CGC_GATE__RBC_MASK
+		  | UVD_CGC_GATE__LMI_MC_MASK
+		  | UVD_CGC_GATE__LMI_UMC_MASK
+		  | UVD_CGC_GATE__IDCT_MASK
+		  | UVD_CGC_GATE__MPRD_MASK
+		  | UVD_CGC_GATE__MPC_MASK
+		  | UVD_CGC_GATE__LBSI_MASK
+		  | UVD_CGC_GATE__LRBBM_MASK
+		  | UVD_CGC_GATE__UDEC_RE_MASK
+		  | UVD_CGC_GATE__UDEC_CM_MASK
+		  | UVD_CGC_GATE__UDEC_IT_MASK
+		  | UVD_CGC_GATE__UDEC_DB_MASK
+		  | UVD_CGC_GATE__UDEC_MP_MASK
+		  | UVD_CGC_GATE__WCB_MASK
+		  | UVD_CGC_GATE__VCPU_MASK
+		  | UVD_CGC_GATE__MMSCH_MASK);
+
+	WREG32_SOC15(VCN, i, mmUVD_CGC_GATE, data);
+
+	SOC15_WAIT_ON_RREG(VCN, i, mmUVD_CGC_GATE, 0,  0xFFFFFFFF);
+
+	data = RREG32_SOC15(VCN, i, mmUVD_CGC_CTRL);
+	data &= ~(UVD_CGC_CTRL__UDEC_RE_MODE_MASK
+		  | UVD_CGC_CTRL__UDEC_CM_MODE_MASK
+		  | UVD_CGC_CTRL__UDEC_IT_MODE_MASK
+		  | UVD_CGC_CTRL__UDEC_DB_MODE_MASK
+		  | UVD_CGC_CTRL__UDEC_MP_MODE_MASK
+		  | UVD_CGC_CTRL__SYS_MODE_MASK
+		  | UVD_CGC_CTRL__UDEC_MODE_MASK
+		  | UVD_CGC_CTRL__MPEG2_MODE_MASK
+		  | UVD_CGC_CTRL__REGS_MODE_MASK
+		  | UVD_CGC_CTRL__RBC_MODE_MASK
+		  | UVD_CGC_CTRL__LMI_MC_MODE_MASK
+		  | UVD_CGC_CTRL__LMI_UMC_MODE_MASK
+		  | UVD_CGC_CTRL__IDCT_MODE_MASK
+		  | UVD_CGC_CTRL__MPRD_MODE_MASK
+		  | UVD_CGC_CTRL__MPC_MODE_MASK
+		  | UVD_CGC_CTRL__LBSI_MODE_MASK
+		  | UVD_CGC_CTRL__LRBBM_MODE_MASK
+		  | UVD_CGC_CTRL__WCB_MODE_MASK
+		  | UVD_CGC_CTRL__VCPU_MODE_MASK
+		  | UVD_CGC_CTRL__MMSCH_MODE_MASK);
+	WREG32_SOC15(VCN, i, mmUVD_CGC_CTRL, data);
+
+	/* turn on */
+	data = RREG32_SOC15(VCN, i, mmUVD_SUVD_CGC_GATE);
+	data |= (UVD_SUVD_CGC_GATE__SRE_MASK
+		 | UVD_SUVD_CGC_GATE__SIT_MASK
+		 | UVD_SUVD_CGC_GATE__SMP_MASK
+		 | UVD_SUVD_CGC_GATE__SCM_MASK
+		 | UVD_SUVD_CGC_GATE__SDB_MASK
+		 | UVD_SUVD_CGC_GATE__SRE_H264_MASK
+		 | UVD_SUVD_CGC_GATE__SRE_HEVC_MASK
+		 | UVD_SUVD_CGC_GATE__SIT_H264_MASK
+		 | UVD_SUVD_CGC_GATE__SIT_HEVC_MASK
+		 | UVD_SUVD_CGC_GATE__SCM_H264_MASK
+		 | UVD_SUVD_CGC_GATE__SCM_HEVC_MASK
+		 | UVD_SUVD_CGC_GATE__SDB_H264_MASK
+		 | UVD_SUVD_CGC_GATE__SDB_HEVC_MASK
+		 | UVD_SUVD_CGC_GATE__SCLR_MASK
+		 | UVD_SUVD_CGC_GATE__UVD_SC_MASK
+		 | UVD_SUVD_CGC_GATE__ENT_MASK
+		 | UVD_SUVD_CGC_GATE__SIT_HEVC_DEC_MASK
+		 | UVD_SUVD_CGC_GATE__SIT_HEVC_ENC_MASK
+		 | UVD_SUVD_CGC_GATE__SITE_MASK
+		 | UVD_SUVD_CGC_GATE__SRE_VP9_MASK
+		 | UVD_SUVD_CGC_GATE__SCM_VP9_MASK
+		 | UVD_SUVD_CGC_GATE__SIT_VP9_DEC_MASK
+		 | UVD_SUVD_CGC_GATE__SDB_VP9_MASK
+		 | UVD_SUVD_CGC_GATE__IME_HEVC_MASK);
+	WREG32_SOC15(VCN, i, mmUVD_SUVD_CGC_GATE, data);
+
+	data = RREG32_SOC15(VCN, i, mmUVD_SUVD_CGC_CTRL);
+	data &= ~(UVD_SUVD_CGC_CTRL__SRE_MODE_MASK
+		  | UVD_SUVD_CGC_CTRL__SIT_MODE_MASK
+		  | UVD_SUVD_CGC_CTRL__SMP_MODE_MASK
+		  | UVD_SUVD_CGC_CTRL__SCM_MODE_MASK
+		  | UVD_SUVD_CGC_CTRL__SDB_MODE_MASK
+		  | UVD_SUVD_CGC_CTRL__SCLR_MODE_MASK
+		  | UVD_SUVD_CGC_CTRL__UVD_SC_MODE_MASK
+		  | UVD_SUVD_CGC_CTRL__ENT_MODE_MASK
 			| UVD_SUVD_CGC_CTRL__IME_MODE_MASK
-			| UVD_SUVD_CGC_CTRL__SITE_MODE_MASK);
-		WREG32_SOC15(VCN, i, mmUVD_SUVD_CGC_CTRL, data);
-	}
+		  | UVD_SUVD_CGC_CTRL__SITE_MODE_MASK);
+	WREG32_SOC15(VCN, i, mmUVD_SUVD_CGC_CTRL, data);
 }
 
-static void vcn_v2_5_clock_gating_dpg_mode(struct amdgpu_device *adev,
-		uint8_t sram_sel, int inst_idx, uint8_t indirect)
+static void vcn_v2_5_clock_gating_dpg_mode(struct amdgpu_vcn_inst *vinst,
+		uint8_t sram_sel, uint8_t indirect)
 {
+	struct amdgpu_device *adev = vinst->adev;
+	int inst_idx = vinst->inst;
 	uint32_t reg_data = 0;
 
 	/* enable sw clock gating control */
@@ -696,70 +907,101 @@ static void vcn_v2_5_clock_gating_dpg_mode(struct amdgpu_device *adev,
 /**
  * vcn_v2_5_enable_clock_gating - enable VCN clock gating
  *
- * @adev: amdgpu_device pointer
+ * @vinst: VCN instance
  *
  * Enable clock gating for VCN block
  */
-static void vcn_v2_5_enable_clock_gating(struct amdgpu_device *adev)
+static void vcn_v2_5_enable_clock_gating(struct amdgpu_vcn_inst *vinst)
 {
+	struct amdgpu_device *adev = vinst->adev;
+	int i = vinst->inst;
 	uint32_t data = 0;
-	int i;
 
-	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
-		if (adev->vcn.harvest_config & (1 << i))
-			continue;
-		/* enable UVD CGC */
-		data = RREG32_SOC15(VCN, i, mmUVD_CGC_CTRL);
-		if (adev->cg_flags & AMD_CG_SUPPORT_VCN_MGCG)
-			data |= 1 << UVD_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
-		else
-			data |= 0 << UVD_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
-		data |= 1 << UVD_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT;
-		data |= 4 << UVD_CGC_CTRL__CLK_OFF_DELAY__SHIFT;
-		WREG32_SOC15(VCN, i, mmUVD_CGC_CTRL, data);
-
-		data = RREG32_SOC15(VCN, i, mmUVD_CGC_CTRL);
-		data |= (UVD_CGC_CTRL__UDEC_RE_MODE_MASK
-			| UVD_CGC_CTRL__UDEC_CM_MODE_MASK
-			| UVD_CGC_CTRL__UDEC_IT_MODE_MASK
-			| UVD_CGC_CTRL__UDEC_DB_MODE_MASK
-			| UVD_CGC_CTRL__UDEC_MP_MODE_MASK
-			| UVD_CGC_CTRL__SYS_MODE_MASK
-			| UVD_CGC_CTRL__UDEC_MODE_MASK
-			| UVD_CGC_CTRL__MPEG2_MODE_MASK
-			| UVD_CGC_CTRL__REGS_MODE_MASK
-			| UVD_CGC_CTRL__RBC_MODE_MASK
-			| UVD_CGC_CTRL__LMI_MC_MODE_MASK
-			| UVD_CGC_CTRL__LMI_UMC_MODE_MASK
-			| UVD_CGC_CTRL__IDCT_MODE_MASK
-			| UVD_CGC_CTRL__MPRD_MODE_MASK
-			| UVD_CGC_CTRL__MPC_MODE_MASK
-			| UVD_CGC_CTRL__LBSI_MODE_MASK
-			| UVD_CGC_CTRL__LRBBM_MODE_MASK
-			| UVD_CGC_CTRL__WCB_MODE_MASK
-			| UVD_CGC_CTRL__VCPU_MODE_MASK);
-		WREG32_SOC15(VCN, i, mmUVD_CGC_CTRL, data);
-
-		data = RREG32_SOC15(VCN, i, mmUVD_SUVD_CGC_CTRL);
-		data |= (UVD_SUVD_CGC_CTRL__SRE_MODE_MASK
-			| UVD_SUVD_CGC_CTRL__SIT_MODE_MASK
-			| UVD_SUVD_CGC_CTRL__SMP_MODE_MASK
-			| UVD_SUVD_CGC_CTRL__SCM_MODE_MASK
-			| UVD_SUVD_CGC_CTRL__SDB_MODE_MASK
-			| UVD_SUVD_CGC_CTRL__SCLR_MODE_MASK
-			| UVD_SUVD_CGC_CTRL__UVD_SC_MODE_MASK
-			| UVD_SUVD_CGC_CTRL__ENT_MODE_MASK
-			| UVD_SUVD_CGC_CTRL__IME_MODE_MASK
-			| UVD_SUVD_CGC_CTRL__SITE_MODE_MASK);
-		WREG32_SOC15(VCN, i, mmUVD_SUVD_CGC_CTRL, data);
-	}
+	if (adev->vcn.harvest_config & (1 << i))
+		return;
+	/* enable UVD CGC */
+	data = RREG32_SOC15(VCN, i, mmUVD_CGC_CTRL);
+	if (adev->cg_flags & AMD_CG_SUPPORT_VCN_MGCG)
+		data |= 1 << UVD_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
+	else
+		data |= 0 << UVD_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
+	data |= 1 << UVD_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT;
+	data |= 4 << UVD_CGC_CTRL__CLK_OFF_DELAY__SHIFT;
+	WREG32_SOC15(VCN, i, mmUVD_CGC_CTRL, data);
+
+	data = RREG32_SOC15(VCN, i, mmUVD_CGC_CTRL);
+	data |= (UVD_CGC_CTRL__UDEC_RE_MODE_MASK
+		 | UVD_CGC_CTRL__UDEC_CM_MODE_MASK
+		 | UVD_CGC_CTRL__UDEC_IT_MODE_MASK
+		 | UVD_CGC_CTRL__UDEC_DB_MODE_MASK
+		 | UVD_CGC_CTRL__UDEC_MP_MODE_MASK
+		 | UVD_CGC_CTRL__SYS_MODE_MASK
+		 | UVD_CGC_CTRL__UDEC_MODE_MASK
+		 | UVD_CGC_CTRL__MPEG2_MODE_MASK
+		 | UVD_CGC_CTRL__REGS_MODE_MASK
+		 | UVD_CGC_CTRL__RBC_MODE_MASK
+		 | UVD_CGC_CTRL__LMI_MC_MODE_MASK
+		 | UVD_CGC_CTRL__LMI_UMC_MODE_MASK
+		 | UVD_CGC_CTRL__IDCT_MODE_MASK
+		 | UVD_CGC_CTRL__MPRD_MODE_MASK
+		 | UVD_CGC_CTRL__MPC_MODE_MASK
+		 | UVD_CGC_CTRL__LBSI_MODE_MASK
+		 | UVD_CGC_CTRL__LRBBM_MODE_MASK
+		 | UVD_CGC_CTRL__WCB_MODE_MASK
+		 | UVD_CGC_CTRL__VCPU_MODE_MASK);
+	WREG32_SOC15(VCN, i, mmUVD_CGC_CTRL, data);
+
+	data = RREG32_SOC15(VCN, i, mmUVD_SUVD_CGC_CTRL);
+	data |= (UVD_SUVD_CGC_CTRL__SRE_MODE_MASK
+		 | UVD_SUVD_CGC_CTRL__SIT_MODE_MASK
+		 | UVD_SUVD_CGC_CTRL__SMP_MODE_MASK
+		 | UVD_SUVD_CGC_CTRL__SCM_MODE_MASK
+		 | UVD_SUVD_CGC_CTRL__SDB_MODE_MASK
+		 | UVD_SUVD_CGC_CTRL__SCLR_MODE_MASK
+		 | UVD_SUVD_CGC_CTRL__UVD_SC_MODE_MASK
+		 | UVD_SUVD_CGC_CTRL__ENT_MODE_MASK
+		 | UVD_SUVD_CGC_CTRL__IME_MODE_MASK
+		 | UVD_SUVD_CGC_CTRL__SITE_MODE_MASK);
+	WREG32_SOC15(VCN, i, mmUVD_SUVD_CGC_CTRL, data);
+}
+
+static void vcn_v2_6_enable_ras(struct amdgpu_vcn_inst *vinst,
+				bool indirect)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int inst_idx = vinst->inst;
+	uint32_t tmp;
+
+	if (amdgpu_ip_version(adev, UVD_HWIP, 0) != IP_VERSION(2, 6, 0))
+		return;
+
+	tmp = VCN_RAS_CNTL__VCPU_VCODEC_REARM_MASK |
+	      VCN_RAS_CNTL__VCPU_VCODEC_IH_EN_MASK |
+	      VCN_RAS_CNTL__VCPU_VCODEC_PMI_EN_MASK |
+	      VCN_RAS_CNTL__VCPU_VCODEC_STALL_EN_MASK;
+	WREG32_SOC15_DPG_MODE(inst_idx,
+			      SOC15_DPG_MODE_OFFSET(VCN, 0, mmVCN_RAS_CNTL),
+			      tmp, 0, indirect);
+
+	tmp = UVD_VCPU_INT_EN__RASCNTL_VCPU_VCODEC_EN_MASK;
+	WREG32_SOC15_DPG_MODE(inst_idx,
+			      SOC15_DPG_MODE_OFFSET(VCN, 0, mmUVD_VCPU_INT_EN),
+			      tmp, 0, indirect);
+
+	tmp = UVD_SYS_INT_EN__RASCNTL_VCPU_VCODEC_EN_MASK;
+	WREG32_SOC15_DPG_MODE(inst_idx,
+			      SOC15_DPG_MODE_OFFSET(VCN, 0, mmUVD_SYS_INT_EN),
+			      tmp, 0, indirect);
 }
 
-static int vcn_v2_5_start_dpg_mode(struct amdgpu_device *adev, int inst_idx, bool indirect)
+static int vcn_v2_5_start_dpg_mode(struct amdgpu_vcn_inst *vinst, bool indirect)
 {
-	volatile struct amdgpu_fw_shared *fw_shared = adev->vcn.inst[inst_idx].fw_shared_cpu_addr;
+	struct amdgpu_device *adev = vinst->adev;
+	int inst_idx = vinst->inst;
+	struct amdgpu_fw_shared *fw_shared = adev->vcn.inst[inst_idx].fw_shared.cpu_addr;
 	struct amdgpu_ring *ring;
 	uint32_t rb_bufsz, tmp;
+	int ret;
 
 	/* disable register anti-hang mechanism */
 	WREG32_P(SOC15_REG_OFFSET(VCN, inst_idx, mmUVD_POWER_STATUS), 1,
@@ -774,7 +1016,7 @@ static int vcn_v2_5_start_dpg_mode(struct amdgpu_device *adev, int inst_idx, boo
 		adev->vcn.inst[inst_idx].dpg_sram_curr_addr = (uint32_t *)adev->vcn.inst[inst_idx].dpg_sram_cpu_addr;
 
 	/* enable clock gating */
-	vcn_v2_5_clock_gating_dpg_mode(adev, 0, inst_idx, indirect);
+	vcn_v2_5_clock_gating_dpg_mode(vinst, 0, indirect);
 
 	/* enable VCPU clock */
 	tmp = (0xFF << UVD_VCPU_CNTL__PRB_TIMEOUT_VAL__SHIFT);
@@ -823,7 +1065,7 @@ static int vcn_v2_5_start_dpg_mode(struct amdgpu_device *adev, int inst_idx, boo
 		 (0x1 << UVD_MPC_SET_MUX__SET_1__SHIFT) |
 		 (0x2 << UVD_MPC_SET_MUX__SET_2__SHIFT)), 0, indirect);
 
-	vcn_v2_5_mc_resume_dpg_mode(adev, inst_idx, indirect);
+	vcn_v2_5_mc_resume_dpg_mode(vinst, indirect);
 
 	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, 0, mmUVD_REG_XX_MASK), 0x10, 0, indirect);
@@ -834,6 +1076,8 @@ static int vcn_v2_5_start_dpg_mode(struct amdgpu_device *adev, int inst_idx, boo
 	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, 0, mmUVD_LMI_CTRL2), 0, 0, indirect);
 
+	vcn_v2_6_enable_ras(vinst, indirect);
+
 	/* unblock VCPU register access */
 	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, 0, mmUVD_RB_ARB_CTRL), 0, 0, indirect);
@@ -848,10 +1092,13 @@ static int vcn_v2_5_start_dpg_mode(struct amdgpu_device *adev, int inst_idx, boo
 		VCN, 0, mmUVD_MASTINT_EN),
 		UVD_MASTINT_EN__VCPU_EN_MASK, 0, indirect);
 
-	if (indirect)
-		psp_update_vcn_sram(adev, inst_idx, adev->vcn.inst[inst_idx].dpg_sram_gpu_addr,
-				    (uint32_t)((uintptr_t)adev->vcn.inst[inst_idx].dpg_sram_curr_addr -
-					       (uintptr_t)adev->vcn.inst[inst_idx].dpg_sram_cpu_addr));
+	if (indirect) {
+		ret = amdgpu_vcn_psp_update_sram(adev, inst_idx, 0);
+		if (ret) {
+			dev_err(adev->dev, "vcn sram load failed %d\n", ret);
+			return ret;
+		}
+	}
 
 	ring = &adev->vcn.inst[inst_idx].ring_dec;
 	/* force RBC into idle state */
@@ -896,198 +1143,200 @@ static int vcn_v2_5_start_dpg_mode(struct amdgpu_device *adev, int inst_idx, boo
 	WREG32_P(SOC15_REG_OFFSET(VCN, inst_idx, mmUVD_POWER_STATUS),
 		0, ~UVD_POWER_STATUS__STALL_DPG_POWER_UP_MASK);
 
+	/* Keeping one read-back to ensure all register writes are done,
+	 * otherwise it may introduce race conditions.
+	 */
+	RREG32_SOC15(VCN, inst_idx, mmUVD_STATUS);
+
 	return 0;
 }
 
-static int vcn_v2_5_start(struct amdgpu_device *adev)
+static int vcn_v2_5_start(struct amdgpu_vcn_inst *vinst)
 {
+	struct amdgpu_device *adev = vinst->adev;
+	int i = vinst->inst;
+	struct amdgpu_fw_shared *fw_shared =
+		adev->vcn.inst[i].fw_shared.cpu_addr;
 	struct amdgpu_ring *ring;
 	uint32_t rb_bufsz, tmp;
-	int i, j, k, r;
+	int j, k, r;
+
+	if (adev->vcn.harvest_config & (1 << i))
+		return 0;
 
 	if (adev->pm.dpm_enabled)
-		amdgpu_dpm_enable_uvd(adev, true);
+		amdgpu_dpm_enable_vcn(adev, true, i);
 
-	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
-		if (adev->vcn.harvest_config & (1 << i))
-			continue;
-		if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) {
-			r = vcn_v2_5_start_dpg_mode(adev, i, adev->vcn.indirect_sram);
-			continue;
-		}
+	if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)
+		return vcn_v2_5_start_dpg_mode(vinst, adev->vcn.inst[i].indirect_sram);
 
-		/* disable register anti-hang mechanism */
-		WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_POWER_STATUS), 0,
-			~UVD_POWER_STATUS__UVD_POWER_STATUS_MASK);
+	/* disable register anti-hang mechanism */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_POWER_STATUS), 0,
+		 ~UVD_POWER_STATUS__UVD_POWER_STATUS_MASK);
 
-		/* set uvd status busy */
-		tmp = RREG32_SOC15(VCN, i, mmUVD_STATUS) | UVD_STATUS__UVD_BUSY;
-		WREG32_SOC15(VCN, i, mmUVD_STATUS, tmp);
-	}
+	/* set uvd status busy */
+	tmp = RREG32_SOC15(VCN, i, mmUVD_STATUS) | UVD_STATUS__UVD_BUSY;
+	WREG32_SOC15(VCN, i, mmUVD_STATUS, tmp);
 
 	if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)
 		return 0;
 
-	/*SW clock gating */
-	vcn_v2_5_disable_clock_gating(adev);
+	/* SW clock gating */
+	vcn_v2_5_disable_clock_gating(vinst);
 
-	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
-		if (adev->vcn.harvest_config & (1 << i))
-			continue;
-		/* enable VCPU clock */
-		WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL),
-			UVD_VCPU_CNTL__CLK_EN_MASK, ~UVD_VCPU_CNTL__CLK_EN_MASK);
-
-		/* disable master interrupt */
-		WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_MASTINT_EN), 0,
-			~UVD_MASTINT_EN__VCPU_EN_MASK);
-
-		/* setup mmUVD_LMI_CTRL */
-		tmp = RREG32_SOC15(VCN, i, mmUVD_LMI_CTRL);
-		tmp &= ~0xff;
-		WREG32_SOC15(VCN, i, mmUVD_LMI_CTRL, tmp | 0x8|
-			UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK	|
-			UVD_LMI_CTRL__MASK_MC_URGENT_MASK |
-			UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK |
-			UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK);
-
-		/* setup mmUVD_MPC_CNTL */
-		tmp = RREG32_SOC15(VCN, i, mmUVD_MPC_CNTL);
-		tmp &= ~UVD_MPC_CNTL__REPLACEMENT_MODE_MASK;
-		tmp |= 0x2 << UVD_MPC_CNTL__REPLACEMENT_MODE__SHIFT;
-		WREG32_SOC15(VCN, i, mmUVD_MPC_CNTL, tmp);
-
-		/* setup UVD_MPC_SET_MUXA0 */
-		WREG32_SOC15(VCN, i, mmUVD_MPC_SET_MUXA0,
-			((0x1 << UVD_MPC_SET_MUXA0__VARA_1__SHIFT) |
-			(0x2 << UVD_MPC_SET_MUXA0__VARA_2__SHIFT) |
-			(0x3 << UVD_MPC_SET_MUXA0__VARA_3__SHIFT) |
-			(0x4 << UVD_MPC_SET_MUXA0__VARA_4__SHIFT)));
-
-		/* setup UVD_MPC_SET_MUXB0 */
-		WREG32_SOC15(VCN, i, mmUVD_MPC_SET_MUXB0,
-			((0x1 << UVD_MPC_SET_MUXB0__VARB_1__SHIFT) |
-			(0x2 << UVD_MPC_SET_MUXB0__VARB_2__SHIFT) |
-			(0x3 << UVD_MPC_SET_MUXB0__VARB_3__SHIFT) |
-			(0x4 << UVD_MPC_SET_MUXB0__VARB_4__SHIFT)));
-
-		/* setup mmUVD_MPC_SET_MUX */
-		WREG32_SOC15(VCN, i, mmUVD_MPC_SET_MUX,
-			((0x0 << UVD_MPC_SET_MUX__SET_0__SHIFT) |
-			(0x1 << UVD_MPC_SET_MUX__SET_1__SHIFT) |
-			(0x2 << UVD_MPC_SET_MUX__SET_2__SHIFT)));
-	}
+	/* enable VCPU clock */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL),
+		 UVD_VCPU_CNTL__CLK_EN_MASK, ~UVD_VCPU_CNTL__CLK_EN_MASK);
 
-	vcn_v2_5_mc_resume(adev);
+	/* disable master interrupt */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_MASTINT_EN), 0,
+		 ~UVD_MASTINT_EN__VCPU_EN_MASK);
 
-	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
-		volatile struct amdgpu_fw_shared *fw_shared = adev->vcn.inst[i].fw_shared_cpu_addr;
-		if (adev->vcn.harvest_config & (1 << i))
-			continue;
-		/* VCN global tiling registers */
-		WREG32_SOC15(VCN, i, mmUVD_GFX8_ADDR_CONFIG,
-			adev->gfx.config.gb_addr_config);
-		WREG32_SOC15(VCN, i, mmUVD_GFX8_ADDR_CONFIG,
-			adev->gfx.config.gb_addr_config);
+	/* setup mmUVD_LMI_CTRL */
+	tmp = RREG32_SOC15(VCN, i, mmUVD_LMI_CTRL);
+	tmp &= ~0xff;
+	WREG32_SOC15(VCN, i, mmUVD_LMI_CTRL, tmp | 0x8|
+		     UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK	|
+		     UVD_LMI_CTRL__MASK_MC_URGENT_MASK |
+		     UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK |
+		     UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK);
+
+	/* setup mmUVD_MPC_CNTL */
+	tmp = RREG32_SOC15(VCN, i, mmUVD_MPC_CNTL);
+	tmp &= ~UVD_MPC_CNTL__REPLACEMENT_MODE_MASK;
+	tmp |= 0x2 << UVD_MPC_CNTL__REPLACEMENT_MODE__SHIFT;
+	WREG32_SOC15(VCN, i, mmUVD_MPC_CNTL, tmp);
+
+	/* setup UVD_MPC_SET_MUXA0 */
+	WREG32_SOC15(VCN, i, mmUVD_MPC_SET_MUXA0,
+		     ((0x1 << UVD_MPC_SET_MUXA0__VARA_1__SHIFT) |
+		      (0x2 << UVD_MPC_SET_MUXA0__VARA_2__SHIFT) |
+		      (0x3 << UVD_MPC_SET_MUXA0__VARA_3__SHIFT) |
+		      (0x4 << UVD_MPC_SET_MUXA0__VARA_4__SHIFT)));
+
+	/* setup UVD_MPC_SET_MUXB0 */
+	WREG32_SOC15(VCN, i, mmUVD_MPC_SET_MUXB0,
+		     ((0x1 << UVD_MPC_SET_MUXB0__VARB_1__SHIFT) |
+		      (0x2 << UVD_MPC_SET_MUXB0__VARB_2__SHIFT) |
+		      (0x3 << UVD_MPC_SET_MUXB0__VARB_3__SHIFT) |
+		      (0x4 << UVD_MPC_SET_MUXB0__VARB_4__SHIFT)));
+
+	/* setup mmUVD_MPC_SET_MUX */
+	WREG32_SOC15(VCN, i, mmUVD_MPC_SET_MUX,
+		     ((0x0 << UVD_MPC_SET_MUX__SET_0__SHIFT) |
+		      (0x1 << UVD_MPC_SET_MUX__SET_1__SHIFT) |
+		      (0x2 << UVD_MPC_SET_MUX__SET_2__SHIFT)));
+
+	vcn_v2_5_mc_resume(vinst);
 
-		/* enable LMI MC and UMC channels */
-		WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_LMI_CTRL2), 0,
-			~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK);
+	/* VCN global tiling registers */
+	WREG32_SOC15(VCN, i, mmUVD_GFX8_ADDR_CONFIG,
+		     adev->gfx.config.gb_addr_config);
+	WREG32_SOC15(VCN, i, mmUVD_GFX8_ADDR_CONFIG,
+		     adev->gfx.config.gb_addr_config);
 
-		/* unblock VCPU register access */
-		WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_RB_ARB_CTRL), 0,
-			~UVD_RB_ARB_CTRL__VCPU_DIS_MASK);
+	/* enable LMI MC and UMC channels */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_LMI_CTRL2), 0,
+		 ~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK);
 
-		WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL), 0,
-			~UVD_VCPU_CNTL__BLK_RST_MASK);
-
-		for (k = 0; k < 10; ++k) {
-			uint32_t status;
-
-			for (j = 0; j < 100; ++j) {
-				status = RREG32_SOC15(VCN, i, mmUVD_STATUS);
-				if (status & 2)
-					break;
-				if (amdgpu_emu_mode == 1)
-					msleep(500);
-				else
-					mdelay(10);
-			}
-			r = 0;
+	/* unblock VCPU register access */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_RB_ARB_CTRL), 0,
+		 ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK);
+
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL), 0,
+		 ~UVD_VCPU_CNTL__BLK_RST_MASK);
+
+	for (k = 0; k < 10; ++k) {
+		uint32_t status;
+
+		for (j = 0; j < 100; ++j) {
+			status = RREG32_SOC15(VCN, i, mmUVD_STATUS);
 			if (status & 2)
 				break;
+			if (amdgpu_emu_mode == 1)
+				msleep(500);
+			else
+				mdelay(10);
+		}
+		r = 0;
+		if (status & 2)
+			break;
 
-			DRM_ERROR("VCN decode not responding, trying to reset the VCPU!!!\n");
-			WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL),
-				UVD_VCPU_CNTL__BLK_RST_MASK,
-				~UVD_VCPU_CNTL__BLK_RST_MASK);
-			mdelay(10);
-			WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL), 0,
-				~UVD_VCPU_CNTL__BLK_RST_MASK);
+		DRM_ERROR("VCN decode not responding, trying to reset the VCPU!!!\n");
+		WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL),
+			 UVD_VCPU_CNTL__BLK_RST_MASK,
+			 ~UVD_VCPU_CNTL__BLK_RST_MASK);
+		mdelay(10);
+		WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL), 0,
+			 ~UVD_VCPU_CNTL__BLK_RST_MASK);
 
-			mdelay(10);
-			r = -1;
-		}
+		mdelay(10);
+		r = -1;
+	}
 
-		if (r) {
-			DRM_ERROR("VCN decode not responding, giving up!!!\n");
-			return r;
-		}
+	if (r) {
+		DRM_ERROR("VCN decode not responding, giving up!!!\n");
+		return r;
+	}
 
-		/* enable master interrupt */
-		WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_MASTINT_EN),
-			UVD_MASTINT_EN__VCPU_EN_MASK,
-			~UVD_MASTINT_EN__VCPU_EN_MASK);
+	/* enable master interrupt */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_MASTINT_EN),
+		 UVD_MASTINT_EN__VCPU_EN_MASK,
+		 ~UVD_MASTINT_EN__VCPU_EN_MASK);
 
-		/* clear the busy bit of VCN_STATUS */
-		WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_STATUS), 0,
-			~(2 << UVD_STATUS__VCPU_REPORT__SHIFT));
+	/* clear the busy bit of VCN_STATUS */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_STATUS), 0,
+		 ~(2 << UVD_STATUS__VCPU_REPORT__SHIFT));
 
-		WREG32_SOC15(VCN, i, mmUVD_LMI_RBC_RB_VMID, 0);
+	WREG32_SOC15(VCN, i, mmUVD_LMI_RBC_RB_VMID, 0);
 
-		ring = &adev->vcn.inst[i].ring_dec;
-		/* force RBC into idle state */
-		rb_bufsz = order_base_2(ring->ring_size);
-		tmp = REG_SET_FIELD(0, UVD_RBC_RB_CNTL, RB_BUFSZ, rb_bufsz);
-		tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_BLKSZ, 1);
-		tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_NO_FETCH, 1);
-		tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_NO_UPDATE, 1);
-		tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_RPTR_WR_EN, 1);
-		WREG32_SOC15(VCN, i, mmUVD_RBC_RB_CNTL, tmp);
+	ring = &adev->vcn.inst[i].ring_dec;
+	/* force RBC into idle state */
+	rb_bufsz = order_base_2(ring->ring_size);
+	tmp = REG_SET_FIELD(0, UVD_RBC_RB_CNTL, RB_BUFSZ, rb_bufsz);
+	tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_BLKSZ, 1);
+	tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_NO_FETCH, 1);
+	tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_NO_UPDATE, 1);
+	tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_RPTR_WR_EN, 1);
+	WREG32_SOC15(VCN, i, mmUVD_RBC_RB_CNTL, tmp);
 
-		fw_shared->multi_queue.decode_queue_mode |= FW_QUEUE_RING_RESET;
-		/* program the RB_BASE for ring buffer */
-		WREG32_SOC15(VCN, i, mmUVD_LMI_RBC_RB_64BIT_BAR_LOW,
-			lower_32_bits(ring->gpu_addr));
-		WREG32_SOC15(VCN, i, mmUVD_LMI_RBC_RB_64BIT_BAR_HIGH,
-			upper_32_bits(ring->gpu_addr));
+	fw_shared->multi_queue.decode_queue_mode |= FW_QUEUE_RING_RESET;
+	/* program the RB_BASE for ring buffer */
+	WREG32_SOC15(VCN, i, mmUVD_LMI_RBC_RB_64BIT_BAR_LOW,
+		     lower_32_bits(ring->gpu_addr));
+	WREG32_SOC15(VCN, i, mmUVD_LMI_RBC_RB_64BIT_BAR_HIGH,
+		     upper_32_bits(ring->gpu_addr));
 
-		/* Initialize the ring buffer's read and write pointers */
-		WREG32_SOC15(VCN, i, mmUVD_RBC_RB_RPTR, 0);
+	/* Initialize the ring buffer's read and write pointers */
+	WREG32_SOC15(VCN, i, mmUVD_RBC_RB_RPTR, 0);
 
-		ring->wptr = RREG32_SOC15(VCN, i, mmUVD_RBC_RB_RPTR);
-		WREG32_SOC15(VCN, i, mmUVD_RBC_RB_WPTR,
-				lower_32_bits(ring->wptr));
-		fw_shared->multi_queue.decode_queue_mode &= ~FW_QUEUE_RING_RESET;
+	ring->wptr = RREG32_SOC15(VCN, i, mmUVD_RBC_RB_RPTR);
+	WREG32_SOC15(VCN, i, mmUVD_RBC_RB_WPTR,
+		     lower_32_bits(ring->wptr));
+	fw_shared->multi_queue.decode_queue_mode &= ~FW_QUEUE_RING_RESET;
 
-		fw_shared->multi_queue.encode_generalpurpose_queue_mode |= FW_QUEUE_RING_RESET;
-		ring = &adev->vcn.inst[i].ring_enc[0];
-		WREG32_SOC15(VCN, i, mmUVD_RB_RPTR, lower_32_bits(ring->wptr));
-		WREG32_SOC15(VCN, i, mmUVD_RB_WPTR, lower_32_bits(ring->wptr));
-		WREG32_SOC15(VCN, i, mmUVD_RB_BASE_LO, ring->gpu_addr);
-		WREG32_SOC15(VCN, i, mmUVD_RB_BASE_HI, upper_32_bits(ring->gpu_addr));
-		WREG32_SOC15(VCN, i, mmUVD_RB_SIZE, ring->ring_size / 4);
-		fw_shared->multi_queue.encode_generalpurpose_queue_mode &= ~FW_QUEUE_RING_RESET;
-
-		fw_shared->multi_queue.encode_lowlatency_queue_mode |= FW_QUEUE_RING_RESET;
-		ring = &adev->vcn.inst[i].ring_enc[1];
-		WREG32_SOC15(VCN, i, mmUVD_RB_RPTR2, lower_32_bits(ring->wptr));
-		WREG32_SOC15(VCN, i, mmUVD_RB_WPTR2, lower_32_bits(ring->wptr));
-		WREG32_SOC15(VCN, i, mmUVD_RB_BASE_LO2, ring->gpu_addr);
-		WREG32_SOC15(VCN, i, mmUVD_RB_BASE_HI2, upper_32_bits(ring->gpu_addr));
-		WREG32_SOC15(VCN, i, mmUVD_RB_SIZE2, ring->ring_size / 4);
-		fw_shared->multi_queue.encode_lowlatency_queue_mode &= ~FW_QUEUE_RING_RESET;
-	}
+	fw_shared->multi_queue.encode_generalpurpose_queue_mode |= FW_QUEUE_RING_RESET;
+	ring = &adev->vcn.inst[i].ring_enc[0];
+	WREG32_SOC15(VCN, i, mmUVD_RB_RPTR, lower_32_bits(ring->wptr));
+	WREG32_SOC15(VCN, i, mmUVD_RB_WPTR, lower_32_bits(ring->wptr));
+	WREG32_SOC15(VCN, i, mmUVD_RB_BASE_LO, ring->gpu_addr);
+	WREG32_SOC15(VCN, i, mmUVD_RB_BASE_HI, upper_32_bits(ring->gpu_addr));
+	WREG32_SOC15(VCN, i, mmUVD_RB_SIZE, ring->ring_size / 4);
+	fw_shared->multi_queue.encode_generalpurpose_queue_mode &= ~FW_QUEUE_RING_RESET;
+
+	fw_shared->multi_queue.encode_lowlatency_queue_mode |= FW_QUEUE_RING_RESET;
+	ring = &adev->vcn.inst[i].ring_enc[1];
+	WREG32_SOC15(VCN, i, mmUVD_RB_RPTR2, lower_32_bits(ring->wptr));
+	WREG32_SOC15(VCN, i, mmUVD_RB_WPTR2, lower_32_bits(ring->wptr));
+	WREG32_SOC15(VCN, i, mmUVD_RB_BASE_LO2, ring->gpu_addr);
+	WREG32_SOC15(VCN, i, mmUVD_RB_BASE_HI2, upper_32_bits(ring->gpu_addr));
+	WREG32_SOC15(VCN, i, mmUVD_RB_SIZE2, ring->ring_size / 4);
+	fw_shared->multi_queue.encode_lowlatency_queue_mode &= ~FW_QUEUE_RING_RESET;
+
+	/* Keeping one read-back to ensure all register writes are done,
+	 * otherwise it may introduce race conditions.
+	 */
+	RREG32_SOC15(VCN, i, mmUVD_STATUS);
 
 	return 0;
 }
@@ -1178,7 +1427,7 @@ static int vcn_v2_5_sriov_start(struct amdgpu_device *adev)
 			SOC15_REG_OFFSET(VCN, i, mmUVD_STATUS),
 			~UVD_STATUS__UVD_BUSY, UVD_STATUS__UVD_BUSY);
 
-		size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.fw->size + 4);
+		size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.inst[i].fw->size + 4);
 		/* mc resume*/
 		if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
 			MMSCH_V1_0_INSERT_DIRECT_WT(
@@ -1288,8 +1537,10 @@ static int vcn_v2_5_sriov_start(struct amdgpu_device *adev)
 	return vcn_v2_5_mmsch_start(adev, &adev->virt.mm_table);
 }
 
-static int vcn_v2_5_stop_dpg_mode(struct amdgpu_device *adev, int inst_idx)
+static int vcn_v2_5_stop_dpg_mode(struct amdgpu_vcn_inst *vinst)
 {
+	struct amdgpu_device *adev = vinst->adev;
+	int inst_idx = vinst->inst;
 	uint32_t tmp;
 
 	/* Wait for power status to be 1 */
@@ -1313,80 +1564,93 @@ static int vcn_v2_5_stop_dpg_mode(struct amdgpu_device *adev, int inst_idx)
 	WREG32_P(SOC15_REG_OFFSET(VCN, inst_idx, mmUVD_POWER_STATUS), 0,
 			~UVD_POWER_STATUS__UVD_PG_MODE_MASK);
 
+	/* Keeping one read-back to ensure all register writes are done,
+	 * otherwise it may introduce race conditions.
+	 */
+	RREG32_SOC15(VCN, inst_idx, mmUVD_STATUS);
+
 	return 0;
 }
 
-static int vcn_v2_5_stop(struct amdgpu_device *adev)
+static int vcn_v2_5_stop(struct amdgpu_vcn_inst *vinst)
 {
+	struct amdgpu_device *adev = vinst->adev;
+	int i = vinst->inst;
 	uint32_t tmp;
-	int i, r = 0;
+	int r;
 
-	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
-		if (adev->vcn.harvest_config & (1 << i))
-			continue;
-		if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) {
-			r = vcn_v2_5_stop_dpg_mode(adev, i);
-			continue;
-		}
+	if (adev->vcn.harvest_config & (1 << i))
+		return 0;
 
-		/* wait for vcn idle */
-		r = SOC15_WAIT_ON_RREG(VCN, i, mmUVD_STATUS, UVD_STATUS__IDLE, 0x7);
-		if (r)
-			return r;
+	if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) {
+		r = vcn_v2_5_stop_dpg_mode(vinst);
+		goto done;
+	}
 
-		tmp = UVD_LMI_STATUS__VCPU_LMI_WRITE_CLEAN_MASK |
-			UVD_LMI_STATUS__READ_CLEAN_MASK |
-			UVD_LMI_STATUS__WRITE_CLEAN_MASK |
-			UVD_LMI_STATUS__WRITE_CLEAN_RAW_MASK;
-		r = SOC15_WAIT_ON_RREG(VCN, i, mmUVD_LMI_STATUS, tmp, tmp);
-		if (r)
-			return r;
+	/* wait for vcn idle */
+	r = SOC15_WAIT_ON_RREG(VCN, i, mmUVD_STATUS, UVD_STATUS__IDLE, 0x7);
+	if (r)
+		goto done;
 
-		/* block LMI UMC channel */
-		tmp = RREG32_SOC15(VCN, i, mmUVD_LMI_CTRL2);
-		tmp |= UVD_LMI_CTRL2__STALL_ARB_UMC_MASK;
-		WREG32_SOC15(VCN, i, mmUVD_LMI_CTRL2, tmp);
+	tmp = UVD_LMI_STATUS__VCPU_LMI_WRITE_CLEAN_MASK |
+		UVD_LMI_STATUS__READ_CLEAN_MASK |
+		UVD_LMI_STATUS__WRITE_CLEAN_MASK |
+		UVD_LMI_STATUS__WRITE_CLEAN_RAW_MASK;
+	r = SOC15_WAIT_ON_RREG(VCN, i, mmUVD_LMI_STATUS, tmp, tmp);
+	if (r)
+		goto done;
 
-		tmp = UVD_LMI_STATUS__UMC_READ_CLEAN_RAW_MASK|
-			UVD_LMI_STATUS__UMC_WRITE_CLEAN_RAW_MASK;
-		r = SOC15_WAIT_ON_RREG(VCN, i, mmUVD_LMI_STATUS, tmp, tmp);
-		if (r)
-			return r;
+	/* block LMI UMC channel */
+	tmp = RREG32_SOC15(VCN, i, mmUVD_LMI_CTRL2);
+	tmp |= UVD_LMI_CTRL2__STALL_ARB_UMC_MASK;
+	WREG32_SOC15(VCN, i, mmUVD_LMI_CTRL2, tmp);
 
-		/* block VCPU register access */
-		WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_RB_ARB_CTRL),
-			UVD_RB_ARB_CTRL__VCPU_DIS_MASK,
-			~UVD_RB_ARB_CTRL__VCPU_DIS_MASK);
+	tmp = UVD_LMI_STATUS__UMC_READ_CLEAN_RAW_MASK|
+		UVD_LMI_STATUS__UMC_WRITE_CLEAN_RAW_MASK;
+	r = SOC15_WAIT_ON_RREG(VCN, i, mmUVD_LMI_STATUS, tmp, tmp);
+	if (r)
+		goto done;
 
-		/* reset VCPU */
-		WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL),
-			UVD_VCPU_CNTL__BLK_RST_MASK,
-			~UVD_VCPU_CNTL__BLK_RST_MASK);
+	/* block VCPU register access */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_RB_ARB_CTRL),
+		 UVD_RB_ARB_CTRL__VCPU_DIS_MASK,
+		 ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK);
 
-		/* disable VCPU clock */
-		WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL), 0,
-			~(UVD_VCPU_CNTL__CLK_EN_MASK));
+	/* reset VCPU */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL),
+		 UVD_VCPU_CNTL__BLK_RST_MASK,
+		 ~UVD_VCPU_CNTL__BLK_RST_MASK);
 
-		/* clear status */
-		WREG32_SOC15(VCN, i, mmUVD_STATUS, 0);
+	/* disable VCPU clock */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL), 0,
+		 ~(UVD_VCPU_CNTL__CLK_EN_MASK));
 
-		vcn_v2_5_enable_clock_gating(adev);
+	/* clear status */
+	WREG32_SOC15(VCN, i, mmUVD_STATUS, 0);
 
-		/* enable register anti-hang mechanism */
-		WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_POWER_STATUS),
-			UVD_POWER_STATUS__UVD_POWER_STATUS_MASK,
-			~UVD_POWER_STATUS__UVD_POWER_STATUS_MASK);
-	}
+	vcn_v2_5_enable_clock_gating(vinst);
 
+	/* enable register anti-hang mechanism */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_POWER_STATUS),
+		 UVD_POWER_STATUS__UVD_POWER_STATUS_MASK,
+		 ~UVD_POWER_STATUS__UVD_POWER_STATUS_MASK);
+
+	/* Keeping one read-back to ensure all register writes are done,
+	 * otherwise it may introduce race conditions.
+	 */
+	RREG32_SOC15(VCN, i, mmUVD_STATUS);
+done:
 	if (adev->pm.dpm_enabled)
-		amdgpu_dpm_enable_uvd(adev, false);
+		amdgpu_dpm_enable_vcn(adev, false, i);
 
-	return 0;
+	return r;
 }
 
-static int vcn_v2_5_pause_dpg_mode(struct amdgpu_device *adev,
-				int inst_idx, struct dpg_pause_state *new_state)
+static int vcn_v2_5_pause_dpg_mode(struct amdgpu_vcn_inst *vinst,
+				   struct dpg_pause_state *new_state)
 {
+	struct amdgpu_device *adev = vinst->adev;
+	int inst_idx = vinst->inst;
 	struct amdgpu_ring *ring;
 	uint32_t reg_data = 0;
 	int ret_code = 0;
@@ -1403,7 +1667,7 @@ static int vcn_v2_5_pause_dpg_mode(struct amdgpu_device *adev,
 				UVD_POWER_STATUS__UVD_POWER_STATUS_MASK);
 
 			if (!ret_code) {
-				volatile struct amdgpu_fw_shared *fw_shared = adev->vcn.inst[inst_idx].fw_shared_cpu_addr;
+				struct amdgpu_fw_shared *fw_shared = adev->vcn.inst[inst_idx].fw_shared.cpu_addr;
 
 				/* pause DPG */
 				reg_data |= UVD_DPG_PAUSE__NJ_PAUSE_DPG_REQ_MASK;
@@ -1485,7 +1749,7 @@ static uint64_t vcn_v2_5_dec_ring_get_wptr(struct amdgpu_ring *ring)
 	struct amdgpu_device *adev = ring->adev;
 
 	if (ring->use_doorbell)
-		return adev->wb.wb[ring->wptr_offs];
+		return *ring->wptr_cpu_addr;
 	else
 		return RREG32_SOC15(VCN, ring->me, mmUVD_RBC_RB_WPTR);
 }
@@ -1502,7 +1766,7 @@ static void vcn_v2_5_dec_ring_set_wptr(struct amdgpu_ring *ring)
 	struct amdgpu_device *adev = ring->adev;
 
 	if (ring->use_doorbell) {
-		adev->wb.wb[ring->wptr_offs] = lower_32_bits(ring->wptr);
+		*ring->wptr_cpu_addr = lower_32_bits(ring->wptr);
 		WDOORBELL32(ring->doorbell_index, lower_32_bits(ring->wptr));
 	} else {
 		WREG32_SOC15(VCN, ring->me, mmUVD_RBC_RB_WPTR, lower_32_bits(ring->wptr));
@@ -1512,7 +1776,7 @@ static void vcn_v2_5_dec_ring_set_wptr(struct amdgpu_ring *ring)
 static const struct amdgpu_ring_funcs vcn_v2_5_dec_ring_vm_funcs = {
 	.type = AMDGPU_RING_TYPE_VCN_DEC,
 	.align_mask = 0xf,
-	.vmhub = AMDGPU_MMHUB_1,
+	.secure_submission_supported = true,
 	.get_rptr = vcn_v2_5_dec_ring_get_rptr,
 	.get_wptr = vcn_v2_5_dec_ring_get_wptr,
 	.set_wptr = vcn_v2_5_dec_ring_set_wptr,
@@ -1532,41 +1796,12 @@ static const struct amdgpu_ring_funcs vcn_v2_5_dec_ring_vm_funcs = {
 	.insert_start = vcn_v2_0_dec_ring_insert_start,
 	.insert_end = vcn_v2_0_dec_ring_insert_end,
 	.pad_ib = amdgpu_ring_generic_pad_ib,
-	.begin_use = amdgpu_vcn_ring_begin_use,
-	.end_use = amdgpu_vcn_ring_end_use,
-	.emit_wreg = vcn_v2_0_dec_ring_emit_wreg,
-	.emit_reg_wait = vcn_v2_0_dec_ring_emit_reg_wait,
-	.emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper,
-};
-
-static const struct amdgpu_ring_funcs vcn_v2_6_dec_ring_vm_funcs = {
-	.type = AMDGPU_RING_TYPE_VCN_DEC,
-	.align_mask = 0xf,
-	.vmhub = AMDGPU_MMHUB_0,
-	.get_rptr = vcn_v2_5_dec_ring_get_rptr,
-	.get_wptr = vcn_v2_5_dec_ring_get_wptr,
-	.set_wptr = vcn_v2_5_dec_ring_set_wptr,
-	.emit_frame_size =
-		SOC15_FLUSH_GPU_TLB_NUM_WREG * 6 +
-		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 8 +
-		8 + /* vcn_v2_0_dec_ring_emit_vm_flush */
-		14 + 14 + /* vcn_v2_0_dec_ring_emit_fence x2 vm fence */
-		6,
-	.emit_ib_size = 8, /* vcn_v2_0_dec_ring_emit_ib */
-	.emit_ib = vcn_v2_0_dec_ring_emit_ib,
-	.emit_fence = vcn_v2_0_dec_ring_emit_fence,
-	.emit_vm_flush = vcn_v2_0_dec_ring_emit_vm_flush,
-	.test_ring = vcn_v2_0_dec_ring_test_ring,
-	.test_ib = amdgpu_vcn_dec_ring_test_ib,
-	.insert_nop = vcn_v2_0_dec_ring_insert_nop,
-	.insert_start = vcn_v2_0_dec_ring_insert_start,
-	.insert_end = vcn_v2_0_dec_ring_insert_end,
-	.pad_ib = amdgpu_ring_generic_pad_ib,
-	.begin_use = amdgpu_vcn_ring_begin_use,
-	.end_use = amdgpu_vcn_ring_end_use,
+	.begin_use = vcn_v2_5_ring_begin_use,
+	.end_use = vcn_v2_5_ring_end_use,
 	.emit_wreg = vcn_v2_0_dec_ring_emit_wreg,
 	.emit_reg_wait = vcn_v2_0_dec_ring_emit_reg_wait,
 	.emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper,
+	.reset = amdgpu_vcn_ring_reset,
 };
 
 /**
@@ -1599,12 +1834,12 @@ static uint64_t vcn_v2_5_enc_ring_get_wptr(struct amdgpu_ring *ring)
 
 	if (ring == &adev->vcn.inst[ring->me].ring_enc[0]) {
 		if (ring->use_doorbell)
-			return adev->wb.wb[ring->wptr_offs];
+			return *ring->wptr_cpu_addr;
 		else
 			return RREG32_SOC15(VCN, ring->me, mmUVD_RB_WPTR);
 	} else {
 		if (ring->use_doorbell)
-			return adev->wb.wb[ring->wptr_offs];
+			return *ring->wptr_cpu_addr;
 		else
 			return RREG32_SOC15(VCN, ring->me, mmUVD_RB_WPTR2);
 	}
@@ -1623,14 +1858,14 @@ static void vcn_v2_5_enc_ring_set_wptr(struct amdgpu_ring *ring)
 
 	if (ring == &adev->vcn.inst[ring->me].ring_enc[0]) {
 		if (ring->use_doorbell) {
-			adev->wb.wb[ring->wptr_offs] = lower_32_bits(ring->wptr);
+			*ring->wptr_cpu_addr = lower_32_bits(ring->wptr);
 			WDOORBELL32(ring->doorbell_index, lower_32_bits(ring->wptr));
 		} else {
 			WREG32_SOC15(VCN, ring->me, mmUVD_RB_WPTR, lower_32_bits(ring->wptr));
 		}
 	} else {
 		if (ring->use_doorbell) {
-			adev->wb.wb[ring->wptr_offs] = lower_32_bits(ring->wptr);
+			*ring->wptr_cpu_addr = lower_32_bits(ring->wptr);
 			WDOORBELL32(ring->doorbell_index, lower_32_bits(ring->wptr));
 		} else {
 			WREG32_SOC15(VCN, ring->me, mmUVD_RB_WPTR2, lower_32_bits(ring->wptr));
@@ -1642,7 +1877,6 @@ static const struct amdgpu_ring_funcs vcn_v2_5_enc_ring_vm_funcs = {
 	.type = AMDGPU_RING_TYPE_VCN_ENC,
 	.align_mask = 0x3f,
 	.nop = VCN_ENC_CMD_NO_OP,
-	.vmhub = AMDGPU_MMHUB_1,
 	.get_rptr = vcn_v2_5_enc_ring_get_rptr,
 	.get_wptr = vcn_v2_5_enc_ring_get_wptr,
 	.set_wptr = vcn_v2_5_enc_ring_set_wptr,
@@ -1661,41 +1895,12 @@ static const struct amdgpu_ring_funcs vcn_v2_5_enc_ring_vm_funcs = {
 	.insert_nop = amdgpu_ring_insert_nop,
 	.insert_end = vcn_v2_0_enc_ring_insert_end,
 	.pad_ib = amdgpu_ring_generic_pad_ib,
-	.begin_use = amdgpu_vcn_ring_begin_use,
-	.end_use = amdgpu_vcn_ring_end_use,
+	.begin_use = vcn_v2_5_ring_begin_use,
+	.end_use = vcn_v2_5_ring_end_use,
 	.emit_wreg = vcn_v2_0_enc_ring_emit_wreg,
 	.emit_reg_wait = vcn_v2_0_enc_ring_emit_reg_wait,
 	.emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper,
-};
-
-static const struct amdgpu_ring_funcs vcn_v2_6_enc_ring_vm_funcs = {
-        .type = AMDGPU_RING_TYPE_VCN_ENC,
-        .align_mask = 0x3f,
-        .nop = VCN_ENC_CMD_NO_OP,
-        .vmhub = AMDGPU_MMHUB_0,
-        .get_rptr = vcn_v2_5_enc_ring_get_rptr,
-        .get_wptr = vcn_v2_5_enc_ring_get_wptr,
-        .set_wptr = vcn_v2_5_enc_ring_set_wptr,
-        .emit_frame_size =
-                SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 +
-                SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 4 +
-                4 + /* vcn_v2_0_enc_ring_emit_vm_flush */
-                5 + 5 + /* vcn_v2_0_enc_ring_emit_fence x2 vm fence */
-                1, /* vcn_v2_0_enc_ring_insert_end */
-        .emit_ib_size = 5, /* vcn_v2_0_enc_ring_emit_ib */
-        .emit_ib = vcn_v2_0_enc_ring_emit_ib,
-        .emit_fence = vcn_v2_0_enc_ring_emit_fence,
-        .emit_vm_flush = vcn_v2_0_enc_ring_emit_vm_flush,
-        .test_ring = amdgpu_vcn_enc_ring_test_ring,
-        .test_ib = amdgpu_vcn_enc_ring_test_ib,
-        .insert_nop = amdgpu_ring_insert_nop,
-        .insert_end = vcn_v2_0_enc_ring_insert_end,
-        .pad_ib = amdgpu_ring_generic_pad_ib,
-        .begin_use = amdgpu_vcn_ring_begin_use,
-        .end_use = amdgpu_vcn_ring_end_use,
-        .emit_wreg = vcn_v2_0_enc_ring_emit_wreg,
-        .emit_reg_wait = vcn_v2_0_enc_ring_emit_reg_wait,
-        .emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper,
+	.reset = amdgpu_vcn_ring_reset,
 };
 
 static void vcn_v2_5_set_dec_ring_funcs(struct amdgpu_device *adev)
@@ -1705,12 +1910,8 @@ static void vcn_v2_5_set_dec_ring_funcs(struct amdgpu_device *adev)
 	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
 		if (adev->vcn.harvest_config & (1 << i))
 			continue;
-		if (adev->ip_versions[UVD_HWIP][0] == IP_VERSION(2, 5, 0))
-			adev->vcn.inst[i].ring_dec.funcs = &vcn_v2_5_dec_ring_vm_funcs;
-		else /* CHIP_ALDEBARAN */
-			adev->vcn.inst[i].ring_dec.funcs = &vcn_v2_6_dec_ring_vm_funcs;
+		adev->vcn.inst[i].ring_dec.funcs = &vcn_v2_5_dec_ring_vm_funcs;
 		adev->vcn.inst[i].ring_dec.me = i;
-		DRM_INFO("VCN(%d) decode is enabled in VM mode\n", i);
 	}
 }
 
@@ -1721,34 +1922,41 @@ static void vcn_v2_5_set_enc_ring_funcs(struct amdgpu_device *adev)
 	for (j = 0; j < adev->vcn.num_vcn_inst; ++j) {
 		if (adev->vcn.harvest_config & (1 << j))
 			continue;
-		for (i = 0; i < adev->vcn.num_enc_rings; ++i) {
-			if (adev->ip_versions[UVD_HWIP][0] == IP_VERSION(2, 5, 0))
-				adev->vcn.inst[j].ring_enc[i].funcs = &vcn_v2_5_enc_ring_vm_funcs;
-			else /* CHIP_ALDEBARAN */
-				adev->vcn.inst[j].ring_enc[i].funcs = &vcn_v2_6_enc_ring_vm_funcs;
+		for (i = 0; i < adev->vcn.inst[j].num_enc_rings; ++i) {
+			adev->vcn.inst[j].ring_enc[i].funcs = &vcn_v2_5_enc_ring_vm_funcs;
 			adev->vcn.inst[j].ring_enc[i].me = j;
 		}
-		DRM_INFO("VCN(%d) encode is enabled in VM mode\n", j);
 	}
 }
 
-static bool vcn_v2_5_is_idle(void *handle)
+static int vcn_v2_5_reset(struct amdgpu_vcn_inst *vinst)
+{
+	int r;
+
+	r = vcn_v2_5_stop(vinst);
+	if (r)
+		return r;
+	return vcn_v2_5_start(vinst);
+}
+
+static bool vcn_v2_5_is_idle(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int i, ret = 1;
 
 	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
 		if (adev->vcn.harvest_config & (1 << i))
 			continue;
+
 		ret &= (RREG32_SOC15(VCN, i, mmUVD_STATUS) == UVD_STATUS__IDLE);
 	}
 
 	return ret;
 }
 
-static int vcn_v2_5_wait_for_idle(void *handle)
+static int vcn_v2_5_wait_for_idle(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int i, ret = 0;
 
 	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
@@ -1763,45 +1971,50 @@ static int vcn_v2_5_wait_for_idle(void *handle)
 	return ret;
 }
 
-static int vcn_v2_5_set_clockgating_state(void *handle,
+static int vcn_v2_5_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_clockgating_state state)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	bool enable = (state == AMD_CG_STATE_GATE);
+	int i;
 
 	if (amdgpu_sriov_vf(adev))
 		return 0;
 
-	if (enable) {
-		if (!vcn_v2_5_is_idle(handle))
-			return -EBUSY;
-		vcn_v2_5_enable_clock_gating(adev);
-	} else {
-		vcn_v2_5_disable_clock_gating(adev);
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		struct amdgpu_vcn_inst *vinst = &adev->vcn.inst[i];
+
+		if (enable) {
+			if (!vcn_v2_5_is_idle(ip_block))
+				return -EBUSY;
+			vcn_v2_5_enable_clock_gating(vinst);
+		} else {
+			vcn_v2_5_disable_clock_gating(vinst);
+		}
 	}
 
 	return 0;
 }
 
-static int vcn_v2_5_set_powergating_state(void *handle,
-					  enum amd_powergating_state state)
+static int vcn_v2_5_set_pg_state(struct amdgpu_vcn_inst *vinst,
+				 enum amd_powergating_state state)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = vinst->adev;
 	int ret;
 
 	if (amdgpu_sriov_vf(adev))
 		return 0;
 
-	if(state == adev->vcn.cur_state)
+	if (state == vinst->cur_state)
 		return 0;
 
 	if (state == AMD_PG_STATE_GATE)
-		ret = vcn_v2_5_stop(adev);
+		ret = vcn_v2_5_stop(vinst);
 	else
-		ret = vcn_v2_5_start(adev);
+		ret = vcn_v2_5_start(vinst);
 
-	if(!ret)
-		adev->vcn.cur_state = state;
+	if (!ret)
+		vinst->cur_state = state;
 
 	return ret;
 }
@@ -1814,6 +2027,14 @@ static int vcn_v2_5_set_interrupt_state(struct amdgpu_device *adev,
 	return 0;
 }
 
+static int vcn_v2_6_set_ras_interrupt_state(struct amdgpu_device *adev,
+					struct amdgpu_irq_src *source,
+					unsigned int type,
+					enum amdgpu_interrupt_state state)
+{
+	return 0;
+}
+
 static int vcn_v2_5_process_interrupt(struct amdgpu_device *adev,
 				      struct amdgpu_irq_src *source,
 				      struct amdgpu_iv_entry *entry)
@@ -1858,6 +2079,11 @@ static const struct amdgpu_irq_src_funcs vcn_v2_5_irq_funcs = {
 	.process = vcn_v2_5_process_interrupt,
 };
 
+static const struct amdgpu_irq_src_funcs vcn_v2_6_ras_irq_funcs = {
+	.set = vcn_v2_6_set_ras_interrupt_state,
+	.process = amdgpu_vcn_process_poison_irq,
+};
+
 static void vcn_v2_5_set_irq_funcs(struct amdgpu_device *adev)
 {
 	int i;
@@ -1865,15 +2091,17 @@ static void vcn_v2_5_set_irq_funcs(struct amdgpu_device *adev)
 	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
 		if (adev->vcn.harvest_config & (1 << i))
 			continue;
-		adev->vcn.inst[i].irq.num_types = adev->vcn.num_enc_rings + 1;
+		adev->vcn.inst[i].irq.num_types = adev->vcn.inst[i].num_enc_rings + 1;
 		adev->vcn.inst[i].irq.funcs = &vcn_v2_5_irq_funcs;
+
+		adev->vcn.inst[i].ras_poison_irq.num_types = adev->vcn.inst[i].num_enc_rings + 1;
+		adev->vcn.inst[i].ras_poison_irq.funcs = &vcn_v2_6_ras_irq_funcs;
 	}
 }
 
 static const struct amd_ip_funcs vcn_v2_5_ip_funcs = {
 	.name = "vcn_v2_5",
 	.early_init = vcn_v2_5_early_init,
-	.late_init = NULL,
 	.sw_init = vcn_v2_5_sw_init,
 	.sw_fini = vcn_v2_5_sw_fini,
 	.hw_init = vcn_v2_5_hw_init,
@@ -1882,18 +2110,15 @@ static const struct amd_ip_funcs vcn_v2_5_ip_funcs = {
 	.resume = vcn_v2_5_resume,
 	.is_idle = vcn_v2_5_is_idle,
 	.wait_for_idle = vcn_v2_5_wait_for_idle,
-	.check_soft_reset = NULL,
-	.pre_soft_reset = NULL,
-	.soft_reset = NULL,
-	.post_soft_reset = NULL,
 	.set_clockgating_state = vcn_v2_5_set_clockgating_state,
-	.set_powergating_state = vcn_v2_5_set_powergating_state,
+	.set_powergating_state = vcn_set_powergating_state,
+	.dump_ip_state = amdgpu_vcn_dump_ip_state,
+	.print_ip_state = amdgpu_vcn_print_ip_state,
 };
 
 static const struct amd_ip_funcs vcn_v2_6_ip_funcs = {
         .name = "vcn_v2_6",
         .early_init = vcn_v2_5_early_init,
-        .late_init = NULL,
         .sw_init = vcn_v2_5_sw_init,
         .sw_fini = vcn_v2_5_sw_fini,
         .hw_init = vcn_v2_5_hw_init,
@@ -1902,12 +2127,10 @@ static const struct amd_ip_funcs vcn_v2_6_ip_funcs = {
         .resume = vcn_v2_5_resume,
         .is_idle = vcn_v2_5_is_idle,
         .wait_for_idle = vcn_v2_5_wait_for_idle,
-        .check_soft_reset = NULL,
-        .pre_soft_reset = NULL,
-        .soft_reset = NULL,
-        .post_soft_reset = NULL,
         .set_clockgating_state = vcn_v2_5_set_clockgating_state,
-        .set_powergating_state = vcn_v2_5_set_powergating_state,
+	.set_powergating_state = vcn_set_powergating_state,
+	.dump_ip_state = amdgpu_vcn_dump_ip_state,
+	.print_ip_state = amdgpu_vcn_print_ip_state,
 };
 
 const struct amdgpu_ip_block_version vcn_v2_5_ip_block =
@@ -1927,3 +2150,59 @@ const struct amdgpu_ip_block_version vcn_v2_6_ip_block =
 		.rev = 0,
 		.funcs = &vcn_v2_6_ip_funcs,
 };
+
+static uint32_t vcn_v2_6_query_poison_by_instance(struct amdgpu_device *adev,
+			uint32_t instance, uint32_t sub_block)
+{
+	uint32_t poison_stat = 0, reg_value = 0;
+
+	switch (sub_block) {
+	case AMDGPU_VCN_V2_6_VCPU_VCODEC:
+		reg_value = RREG32_SOC15(VCN, instance, mmUVD_RAS_VCPU_VCODEC_STATUS);
+		poison_stat = REG_GET_FIELD(reg_value, UVD_RAS_VCPU_VCODEC_STATUS, POISONED_PF);
+		break;
+	default:
+		break;
+	}
+
+	if (poison_stat)
+		dev_info(adev->dev, "Poison detected in VCN%d, sub_block%d\n",
+			instance, sub_block);
+
+	return poison_stat;
+}
+
+static bool vcn_v2_6_query_poison_status(struct amdgpu_device *adev)
+{
+	uint32_t inst, sub;
+	uint32_t poison_stat = 0;
+
+	for (inst = 0; inst < adev->vcn.num_vcn_inst; inst++)
+		for (sub = 0; sub < AMDGPU_VCN_V2_6_MAX_SUB_BLOCK; sub++)
+			poison_stat +=
+			vcn_v2_6_query_poison_by_instance(adev, inst, sub);
+
+	return !!poison_stat;
+}
+
+const struct amdgpu_ras_block_hw_ops vcn_v2_6_ras_hw_ops = {
+	.query_poison_status = vcn_v2_6_query_poison_status,
+};
+
+static struct amdgpu_vcn_ras vcn_v2_6_ras = {
+	.ras_block = {
+		.hw_ops = &vcn_v2_6_ras_hw_ops,
+		.ras_late_init = amdgpu_vcn_ras_late_init,
+	},
+};
+
+static void vcn_v2_5_set_ras_funcs(struct amdgpu_device *adev)
+{
+	switch (amdgpu_ip_version(adev, VCN_HWIP, 0)) {
+	case IP_VERSION(2, 6, 0):
+		adev->vcn.ras = &vcn_v2_6_ras;
+		break;
+	default:
+		break;
+	}
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.h b/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.h
index e72f799ed0fd..1c19af74e4fd 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.h
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.h
@@ -24,6 +24,12 @@
 #ifndef __VCN_V2_5_H__
 #define __VCN_V2_5_H__
 
+enum amdgpu_vcn_v2_6_sub_block {
+	AMDGPU_VCN_V2_6_VCPU_VCODEC = 0,
+
+	AMDGPU_VCN_V2_6_MAX_SUB_BLOCK,
+};
+
 extern const struct amdgpu_ip_block_version vcn_v2_5_ip_block;
 extern const struct amdgpu_ip_block_version vcn_v2_6_ip_block;
 
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
index da11ceba0698..d9cf8f0feeb3 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
@@ -25,10 +25,12 @@
 #include "amdgpu.h"
 #include "amdgpu_vcn.h"
 #include "amdgpu_pm.h"
+#include "amdgpu_cs.h"
 #include "soc15.h"
 #include "soc15d.h"
 #include "vcn_v2_0.h"
 #include "mmsch_v3_0.h"
+#include "vcn_sw_ring.h"
 
 #include "vcn/vcn_3_0_0_offset.h"
 #include "vcn/vcn_3_0_0_sh_mask.h"
@@ -36,6 +38,10 @@
 
 #include <drm/drm_drv.h>
 
+#define VCN_VID_SOC_ADDRESS_2_0					0x1fa00
+#define VCN1_VID_SOC_ADDRESS_3_0				0x48200
+#define VCN1_AON_SOC_ADDRESS_3_0				0x48000
+
 #define mmUVD_CONTEXT_ID_INTERNAL_OFFSET			0x27
 #define mmUVD_GPCOM_VCPU_CMD_INTERNAL_OFFSET			0x0f
 #define mmUVD_GPCOM_VCPU_DATA0_INTERNAL_OFFSET			0x10
@@ -55,6 +61,42 @@
 #define RDECODE_MSG_CREATE					0x00000000
 #define RDECODE_MESSAGE_CREATE					0x00000001
 
+static const struct amdgpu_hwip_reg_entry vcn_reg_list_3_0[] = {
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_POWER_STATUS),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_STATUS),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_CONTEXT_ID),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_CONTEXT_ID2),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_GPCOM_VCPU_DATA0),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_GPCOM_VCPU_DATA1),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_GPCOM_VCPU_CMD),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_BASE_HI),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_BASE_LO),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_BASE_HI2),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_BASE_LO2),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_BASE_HI3),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_BASE_LO3),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_BASE_HI4),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_BASE_LO4),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_RPTR),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_WPTR),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_RPTR2),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_WPTR2),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_RPTR3),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_WPTR3),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_RPTR4),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_WPTR4),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_SIZE),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_SIZE2),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_SIZE3),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_RB_SIZE4),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_PGFSM_CONFIG),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_PGFSM_STATUS),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_DPG_LMA_CTL),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_DPG_LMA_DATA),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_DPG_LMA_MASK),
+	SOC15_REG_ENTRY_STR(VCN, 0, mmUVD_DPG_PAUSE)
+};
+
 static int amdgpu_ih_clientid_vcns[] = {
 	SOC15_IH_CLIENTID_VCN,
 	SOC15_IH_CLIENTID_VCN1
@@ -64,29 +106,33 @@ static int vcn_v3_0_start_sriov(struct amdgpu_device *adev);
 static void vcn_v3_0_set_dec_ring_funcs(struct amdgpu_device *adev);
 static void vcn_v3_0_set_enc_ring_funcs(struct amdgpu_device *adev);
 static void vcn_v3_0_set_irq_funcs(struct amdgpu_device *adev);
-static int vcn_v3_0_set_powergating_state(void *handle,
-			enum amd_powergating_state state);
-static int vcn_v3_0_pause_dpg_mode(struct amdgpu_device *adev,
-			int inst_idx, struct dpg_pause_state *new_state);
+static int vcn_v3_0_set_pg_state(struct amdgpu_vcn_inst *vinst,
+				 enum amd_powergating_state state);
+static int vcn_v3_0_pause_dpg_mode(struct amdgpu_vcn_inst *vinst,
+				   struct dpg_pause_state *new_state);
+static int vcn_v3_0_reset(struct amdgpu_vcn_inst *vinst);
 
 static void vcn_v3_0_dec_ring_set_wptr(struct amdgpu_ring *ring);
 static void vcn_v3_0_enc_ring_set_wptr(struct amdgpu_ring *ring);
 
 /**
- * vcn_v3_0_early_init - set function pointers
+ * vcn_v3_0_early_init - set function pointers and load microcode
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Set ring and irq function pointers
+ * Load microcode from filesystem
  */
-static int vcn_v3_0_early_init(void *handle)
+static int vcn_v3_0_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, r;
 
 	if (amdgpu_sriov_vf(adev)) {
 		adev->vcn.num_vcn_inst = VCN_INSTANCES_SIENNA_CICHLID;
 		adev->vcn.harvest_config = 0;
-		adev->vcn.num_enc_rings = 1;
+		for (i = 0; i < adev->vcn.num_vcn_inst; i++)
+			adev->vcn.inst[i].num_enc_rings = 1;
 
 	} else {
 		if (adev->vcn.harvest_config == (AMDGPU_VCN_HARVEST_VCN0 |
@@ -94,42 +140,42 @@ static int vcn_v3_0_early_init(void *handle)
 			/* both instances are harvested, disable the block */
 			return -ENOENT;
 
-		if (adev->ip_versions[UVD_HWIP][0] == IP_VERSION(3, 0, 33))
-			adev->vcn.num_enc_rings = 0;
-		else
-			adev->vcn.num_enc_rings = 2;
+		for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+			if (amdgpu_ip_version(adev, UVD_HWIP, 0) ==
+			    IP_VERSION(3, 0, 33))
+				adev->vcn.inst[i].num_enc_rings = 0;
+			else
+				adev->vcn.inst[i].num_enc_rings = 2;
+		}
 	}
 
 	vcn_v3_0_set_dec_ring_funcs(adev);
 	vcn_v3_0_set_enc_ring_funcs(adev);
 	vcn_v3_0_set_irq_funcs(adev);
 
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+		adev->vcn.inst[i].set_pg_state = vcn_v3_0_set_pg_state;
+
+		r = amdgpu_vcn_early_init(adev, i);
+		if (r)
+			return r;
+	}
 	return 0;
 }
 
 /**
  * vcn_v3_0_sw_init - sw init for VCN block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Load firmware and sw initialization
  */
-static int vcn_v3_0_sw_init(void *handle)
+static int vcn_v3_0_sw_init(struct amdgpu_ip_block *ip_block)
 {
 	struct amdgpu_ring *ring;
 	int i, j, r;
 	int vcn_doorbell_index = 0;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	r = amdgpu_vcn_sw_init(adev);
-	if (r)
-		return r;
-
-	amdgpu_vcn_setup_ucode(adev);
-
-	r = amdgpu_vcn_resume(adev);
-	if (r)
-		return r;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	/*
 	 * Note: doorbell assignment is fixed for SRIOV multiple VCN engines
@@ -145,27 +191,37 @@ static int vcn_v3_0_sw_init(void *handle)
 	}
 
 	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
-		volatile struct amdgpu_fw_shared *fw_shared;
+		struct amdgpu_fw_shared *fw_shared;
 
 		if (adev->vcn.harvest_config & (1 << i))
 			continue;
 
-		adev->vcn.internal.context_id = mmUVD_CONTEXT_ID_INTERNAL_OFFSET;
-		adev->vcn.internal.ib_vmid = mmUVD_LMI_RBC_IB_VMID_INTERNAL_OFFSET;
-		adev->vcn.internal.ib_bar_low = mmUVD_LMI_RBC_IB_64BIT_BAR_LOW_INTERNAL_OFFSET;
-		adev->vcn.internal.ib_bar_high = mmUVD_LMI_RBC_IB_64BIT_BAR_HIGH_INTERNAL_OFFSET;
-		adev->vcn.internal.ib_size = mmUVD_RBC_IB_SIZE_INTERNAL_OFFSET;
-		adev->vcn.internal.gp_scratch8 = mmUVD_GP_SCRATCH8_INTERNAL_OFFSET;
+		r = amdgpu_vcn_sw_init(adev, i);
+		if (r)
+			return r;
+
+		amdgpu_vcn_setup_ucode(adev, i);
+
+		r = amdgpu_vcn_resume(adev, i);
+		if (r)
+			return r;
+
+		adev->vcn.inst[i].internal.context_id = mmUVD_CONTEXT_ID_INTERNAL_OFFSET;
+		adev->vcn.inst[i].internal.ib_vmid = mmUVD_LMI_RBC_IB_VMID_INTERNAL_OFFSET;
+		adev->vcn.inst[i].internal.ib_bar_low = mmUVD_LMI_RBC_IB_64BIT_BAR_LOW_INTERNAL_OFFSET;
+		adev->vcn.inst[i].internal.ib_bar_high = mmUVD_LMI_RBC_IB_64BIT_BAR_HIGH_INTERNAL_OFFSET;
+		adev->vcn.inst[i].internal.ib_size = mmUVD_RBC_IB_SIZE_INTERNAL_OFFSET;
+		adev->vcn.inst[i].internal.gp_scratch8 = mmUVD_GP_SCRATCH8_INTERNAL_OFFSET;
 
-		adev->vcn.internal.scratch9 = mmUVD_SCRATCH9_INTERNAL_OFFSET;
+		adev->vcn.inst[i].internal.scratch9 = mmUVD_SCRATCH9_INTERNAL_OFFSET;
 		adev->vcn.inst[i].external.scratch9 = SOC15_REG_OFFSET(VCN, i, mmUVD_SCRATCH9);
-		adev->vcn.internal.data0 = mmUVD_GPCOM_VCPU_DATA0_INTERNAL_OFFSET;
+		adev->vcn.inst[i].internal.data0 = mmUVD_GPCOM_VCPU_DATA0_INTERNAL_OFFSET;
 		adev->vcn.inst[i].external.data0 = SOC15_REG_OFFSET(VCN, i, mmUVD_GPCOM_VCPU_DATA0);
-		adev->vcn.internal.data1 = mmUVD_GPCOM_VCPU_DATA1_INTERNAL_OFFSET;
+		adev->vcn.inst[i].internal.data1 = mmUVD_GPCOM_VCPU_DATA1_INTERNAL_OFFSET;
 		adev->vcn.inst[i].external.data1 = SOC15_REG_OFFSET(VCN, i, mmUVD_GPCOM_VCPU_DATA1);
-		adev->vcn.internal.cmd = mmUVD_GPCOM_VCPU_CMD_INTERNAL_OFFSET;
+		adev->vcn.inst[i].internal.cmd = mmUVD_GPCOM_VCPU_CMD_INTERNAL_OFFSET;
 		adev->vcn.inst[i].external.cmd = SOC15_REG_OFFSET(VCN, i, mmUVD_GPCOM_VCPU_CMD);
-		adev->vcn.internal.nop = mmUVD_NO_OP_INTERNAL_OFFSET;
+		adev->vcn.inst[i].internal.nop = mmUVD_NO_OP_INTERNAL_OFFSET;
 		adev->vcn.inst[i].external.nop = SOC15_REG_OFFSET(VCN, i, mmUVD_NO_OP);
 
 		/* VCN DEC TRAP */
@@ -179,10 +235,11 @@ static int vcn_v3_0_sw_init(void *handle)
 		ring = &adev->vcn.inst[i].ring_dec;
 		ring->use_doorbell = true;
 		if (amdgpu_sriov_vf(adev)) {
-			ring->doorbell_index = vcn_doorbell_index + i * (adev->vcn.num_enc_rings + 1);
+			ring->doorbell_index = vcn_doorbell_index + i * (adev->vcn.inst[i].num_enc_rings + 1);
 		} else {
 			ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 8 * i;
 		}
+		ring->vm_hub = AMDGPU_MMHUB0(0);
 		sprintf(ring->name, "vcn_dec_%d", i);
 		r = amdgpu_ring_init(adev, ring, 512, &adev->vcn.inst[i].irq, 0,
 				     AMDGPU_RING_PRIO_DEFAULT,
@@ -190,7 +247,7 @@ static int vcn_v3_0_sw_init(void *handle)
 		if (r)
 			return r;
 
-		for (j = 0; j < adev->vcn.num_enc_rings; ++j) {
+		for (j = 0; j < adev->vcn.inst[i].num_enc_rings; ++j) {
 			enum amdgpu_ring_priority_level hw_prio = amdgpu_vcn_get_enc_ring_prio(j);
 
 			/* VCN ENC TRAP */
@@ -202,10 +259,11 @@ static int vcn_v3_0_sw_init(void *handle)
 			ring = &adev->vcn.inst[i].ring_enc[j];
 			ring->use_doorbell = true;
 			if (amdgpu_sriov_vf(adev)) {
-				ring->doorbell_index = vcn_doorbell_index + i * (adev->vcn.num_enc_rings + 1) + 1 + j;
+				ring->doorbell_index = vcn_doorbell_index + i * (adev->vcn.inst[i].num_enc_rings + 1) + 1 + j;
 			} else {
 				ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 2 + j + 8 * i;
 			}
+			ring->vm_hub = AMDGPU_MMHUB0(0);
 			sprintf(ring->name, "vcn_enc_%d.%d", i, j);
 			r = amdgpu_ring_init(adev, ring, 512, &adev->vcn.inst[i].irq, 0,
 					     hw_prio, &adev->vcn.inst[i].sched_score);
@@ -213,20 +271,44 @@ static int vcn_v3_0_sw_init(void *handle)
 				return r;
 		}
 
-		fw_shared = adev->vcn.inst[i].fw_shared_cpu_addr;
+		fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr;
 		fw_shared->present_flag_0 |= cpu_to_le32(AMDGPU_VCN_SW_RING_FLAG) |
 					     cpu_to_le32(AMDGPU_VCN_MULTI_QUEUE_FLAG) |
 					     cpu_to_le32(AMDGPU_VCN_FW_SHARED_FLAG_0_RB);
 		fw_shared->sw_ring.is_enabled = cpu_to_le32(DEC_SW_RING_ENABLED);
+		fw_shared->present_flag_0 |= AMDGPU_VCN_SMU_VERSION_INFO_FLAG;
+		if (amdgpu_ip_version(adev, UVD_HWIP, 0) == IP_VERSION(3, 1, 2))
+			fw_shared->smu_interface_info.smu_interface_type = 2;
+		else if (amdgpu_ip_version(adev, UVD_HWIP, 0) ==
+			 IP_VERSION(3, 1, 1))
+			fw_shared->smu_interface_info.smu_interface_type = 1;
+
+		if (amdgpu_vcnfw_log)
+			amdgpu_vcn_fwlog_init(&adev->vcn.inst[i]);
+
+		if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)
+			adev->vcn.inst[i].pause_dpg_mode = vcn_v3_0_pause_dpg_mode;
+		adev->vcn.inst[i].reset = vcn_v3_0_reset;
 	}
 
+	adev->vcn.supported_reset =
+		amdgpu_get_soft_full_reset_mask(&adev->vcn.inst[0].ring_enc[0]);
+	if (!amdgpu_sriov_vf(adev))
+		adev->vcn.supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE;
+
 	if (amdgpu_sriov_vf(adev)) {
 		r = amdgpu_virt_alloc_mm_table(adev);
 		if (r)
 			return r;
 	}
-	if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)
-		adev->vcn.pause_dpg_mode = vcn_v3_0_pause_dpg_mode;
+
+	r = amdgpu_vcn_reg_dump_init(adev, vcn_reg_list_3_0, ARRAY_SIZE(vcn_reg_list_3_0));
+	if (r)
+		return r;
+
+	r = amdgpu_vcn_sysfs_reset_mask_init(adev);
+	if (r)
+		return r;
 
 	return 0;
 }
@@ -234,22 +316,22 @@ static int vcn_v3_0_sw_init(void *handle)
 /**
  * vcn_v3_0_sw_fini - sw fini for VCN block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * VCN suspend and free up sw allocation
  */
-static int vcn_v3_0_sw_fini(void *handle)
+static int vcn_v3_0_sw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int i, r, idx;
 
 	if (drm_dev_enter(adev_to_drm(adev), &idx)) {
 		for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
-			volatile struct amdgpu_fw_shared *fw_shared;
+			struct amdgpu_fw_shared *fw_shared;
 
 			if (adev->vcn.harvest_config & (1 << i))
 				continue;
-			fw_shared = adev->vcn.inst[i].fw_shared_cpu_addr;
+			fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr;
 			fw_shared->present_flag_0 = 0;
 			fw_shared->sw_ring.is_enabled = false;
 		}
@@ -260,32 +342,36 @@ static int vcn_v3_0_sw_fini(void *handle)
 	if (amdgpu_sriov_vf(adev))
 		amdgpu_virt_free_mm_table(adev);
 
-	r = amdgpu_vcn_suspend(adev);
-	if (r)
-		return r;
+	amdgpu_vcn_sysfs_reset_mask_fini(adev);
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+		r = amdgpu_vcn_suspend(adev, i);
+		if (r)
+			return r;
 
-	r = amdgpu_vcn_sw_fini(adev);
+		amdgpu_vcn_sw_fini(adev, i);
+	}
 
-	return r;
+	return 0;
 }
 
 /**
  * vcn_v3_0_hw_init - start and test VCN block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Initialize the hardware, boot up the VCPU and do some testing
  */
-static int vcn_v3_0_hw_init(void *handle)
+static int vcn_v3_0_hw_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	struct amdgpu_ring *ring;
 	int i, j, r;
 
 	if (amdgpu_sriov_vf(adev)) {
 		r = vcn_v3_0_start_sriov(adev);
 		if (r)
-			goto done;
+			return r;
 
 		/* initialize VCN dec and enc ring buffers */
 		for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
@@ -295,6 +381,7 @@ static int vcn_v3_0_hw_init(void *handle)
 			ring = &adev->vcn.inst[i].ring_dec;
 			if (amdgpu_vcn_is_disabled_vcn(adev, VCN_DECODE_RING, i)) {
 				ring->sched.ready = false;
+				ring->no_scheduler = true;
 				dev_info(adev->dev, "ring %s is disabled by hypervisor\n", ring->name);
 			} else {
 				ring->wptr = 0;
@@ -303,10 +390,11 @@ static int vcn_v3_0_hw_init(void *handle)
 				ring->sched.ready = true;
 			}
 
-			for (j = 0; j < adev->vcn.num_enc_rings; ++j) {
+			for (j = 0; j < adev->vcn.inst[i].num_enc_rings; ++j) {
 				ring = &adev->vcn.inst[i].ring_enc[j];
 				if (amdgpu_vcn_is_disabled_vcn(adev, VCN_ENCODE_RING, i)) {
 					ring->sched.ready = false;
+					ring->no_scheduler = true;
 					dev_info(adev->dev, "ring %s is disabled by hypervisor\n", ring->name);
 				} else {
 					ring->wptr = 0;
@@ -328,48 +416,45 @@ static int vcn_v3_0_hw_init(void *handle)
 
 			r = amdgpu_ring_test_helper(ring);
 			if (r)
-				goto done;
+				return r;
 
-			for (j = 0; j < adev->vcn.num_enc_rings; ++j) {
+			for (j = 0; j < adev->vcn.inst[i].num_enc_rings; ++j) {
 				ring = &adev->vcn.inst[i].ring_enc[j];
 				r = amdgpu_ring_test_helper(ring);
 				if (r)
-					goto done;
+					return r;
 			}
 		}
 	}
 
-done:
-	if (!r)
-		DRM_INFO("VCN decode and encode initialized successfully(under %s).\n",
-			(adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)?"DPG Mode":"SPG Mode");
-
-	return r;
+	return 0;
 }
 
 /**
  * vcn_v3_0_hw_fini - stop the hardware block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Stop the VCN block, mark ring as not ready any more
  */
-static int vcn_v3_0_hw_fini(void *handle)
+static int vcn_v3_0_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int i;
 
-	cancel_delayed_work_sync(&adev->vcn.idle_work);
-
 	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		struct amdgpu_vcn_inst *vinst = &adev->vcn.inst[i];
+
 		if (adev->vcn.harvest_config & (1 << i))
 			continue;
 
+		cancel_delayed_work_sync(&vinst->idle_work);
+
 		if (!amdgpu_sriov_vf(adev)) {
 			if ((adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) ||
-					(adev->vcn.cur_state != AMD_PG_STATE_GATE &&
-					 RREG32_SOC15(VCN, i, mmUVD_STATUS))) {
-				vcn_v3_0_set_powergating_state(adev, AMD_PG_STATE_GATE);
+			    (vinst->cur_state != AMD_PG_STATE_GATE &&
+			     RREG32_SOC15(VCN, i, mmUVD_STATUS))) {
+				vinst->set_pg_state(vinst, AMD_PG_STATE_GATE);
 			}
 		}
 	}
@@ -380,41 +465,47 @@ static int vcn_v3_0_hw_fini(void *handle)
 /**
  * vcn_v3_0_suspend - suspend VCN block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * HW fini and suspend VCN block
  */
-static int vcn_v3_0_suspend(void *handle)
+static int vcn_v3_0_suspend(struct amdgpu_ip_block *ip_block)
 {
-	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
+	int r, i;
 
-	r = vcn_v3_0_hw_fini(adev);
+	r = vcn_v3_0_hw_fini(ip_block);
 	if (r)
 		return r;
 
-	r = amdgpu_vcn_suspend(adev);
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+		r = amdgpu_vcn_suspend(ip_block->adev, i);
+		if (r)
+			return r;
+	}
 
-	return r;
+	return 0;
 }
 
 /**
  * vcn_v3_0_resume - resume VCN block
  *
- * @handle: amdgpu_device pointer
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Resume firmware and hw init VCN block
  */
-static int vcn_v3_0_resume(void *handle)
+static int vcn_v3_0_resume(struct amdgpu_ip_block *ip_block)
 {
-	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
+	int r, i;
 
-	r = amdgpu_vcn_resume(adev);
-	if (r)
-		return r;
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+		r = amdgpu_vcn_resume(ip_block->adev, i);
+		if (r)
+			return r;
+	}
 
-	r = vcn_v3_0_hw_init(adev);
+	r = vcn_v3_0_hw_init(ip_block);
 
 	return r;
 }
@@ -422,14 +513,15 @@ static int vcn_v3_0_resume(void *handle)
 /**
  * vcn_v3_0_mc_resume - memory controller programming
  *
- * @adev: amdgpu_device pointer
- * @inst: instance number
+ * @vinst: VCN instance
  *
  * Let the VCN memory controller know it's offsets
  */
-static void vcn_v3_0_mc_resume(struct amdgpu_device *adev, int inst)
+static void vcn_v3_0_mc_resume(struct amdgpu_vcn_inst *vinst)
 {
-	uint32_t size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.fw->size + 4);
+	struct amdgpu_device *adev = vinst->adev;
+	int inst = vinst->inst;
+	uint32_t size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.inst[inst].fw->size + 4);
 	uint32_t offset;
 
 	/* cache window 0: fw */
@@ -469,17 +561,20 @@ static void vcn_v3_0_mc_resume(struct amdgpu_device *adev, int inst)
 
 	/* non-cache window */
 	WREG32_SOC15(VCN, inst, mmUVD_LMI_VCPU_NC0_64BIT_BAR_LOW,
-		lower_32_bits(adev->vcn.inst[inst].fw_shared_gpu_addr));
+		lower_32_bits(adev->vcn.inst[inst].fw_shared.gpu_addr));
 	WREG32_SOC15(VCN, inst, mmUVD_LMI_VCPU_NC0_64BIT_BAR_HIGH,
-		upper_32_bits(adev->vcn.inst[inst].fw_shared_gpu_addr));
+		upper_32_bits(adev->vcn.inst[inst].fw_shared.gpu_addr));
 	WREG32_SOC15(VCN, inst, mmUVD_VCPU_NONCACHE_OFFSET0, 0);
 	WREG32_SOC15(VCN, inst, mmUVD_VCPU_NONCACHE_SIZE0,
 		AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_fw_shared)));
 }
 
-static void vcn_v3_0_mc_resume_dpg_mode(struct amdgpu_device *adev, int inst_idx, bool indirect)
+static void vcn_v3_0_mc_resume_dpg_mode(struct amdgpu_vcn_inst *vinst,
+					bool indirect)
 {
-	uint32_t size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.fw->size + 4);
+	struct amdgpu_device *adev = vinst->adev;
+	int inst_idx = vinst->inst;
+	uint32_t size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.inst[inst_idx].fw->size + 4);
 	uint32_t offset;
 
 	/* cache window 0: fw */
@@ -558,10 +653,10 @@ static void vcn_v3_0_mc_resume_dpg_mode(struct amdgpu_device *adev, int inst_idx
 	/* non-cache window */
 	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 			VCN, inst_idx, mmUVD_LMI_VCPU_NC0_64BIT_BAR_LOW),
-			lower_32_bits(adev->vcn.inst[inst_idx].fw_shared_gpu_addr), 0, indirect);
+			lower_32_bits(adev->vcn.inst[inst_idx].fw_shared.gpu_addr), 0, indirect);
 	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 			VCN, inst_idx, mmUVD_LMI_VCPU_NC0_64BIT_BAR_HIGH),
-			upper_32_bits(adev->vcn.inst[inst_idx].fw_shared_gpu_addr), 0, indirect);
+			upper_32_bits(adev->vcn.inst[inst_idx].fw_shared.gpu_addr), 0, indirect);
 	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 			VCN, inst_idx, mmUVD_VCPU_NONCACHE_OFFSET0), 0, 0, indirect);
 	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
@@ -569,12 +664,14 @@ static void vcn_v3_0_mc_resume_dpg_mode(struct amdgpu_device *adev, int inst_idx
 			AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_fw_shared)), 0, indirect);
 
 	/* VCN global tiling registers */
-	WREG32_SOC15_DPG_MODE(0, SOC15_DPG_MODE_OFFSET(
-		UVD, 0, mmUVD_GFX10_ADDR_CONFIG), adev->gfx.config.gb_addr_config, 0, indirect);
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		UVD, inst_idx, mmUVD_GFX10_ADDR_CONFIG), adev->gfx.config.gb_addr_config, 0, indirect);
 }
 
-static void vcn_v3_0_disable_static_power_gating(struct amdgpu_device *adev, int inst)
+static void vcn_v3_0_disable_static_power_gating(struct amdgpu_vcn_inst *vinst)
 {
+	struct amdgpu_device *adev = vinst->adev;
+	int inst = vinst->inst;
 	uint32_t data = 0;
 
 	if (adev->pg_flags & AMD_PG_SUPPORT_VCN) {
@@ -624,8 +721,10 @@ static void vcn_v3_0_disable_static_power_gating(struct amdgpu_device *adev, int
 	WREG32_SOC15(VCN, inst, mmUVD_POWER_STATUS, data);
 }
 
-static void vcn_v3_0_enable_static_power_gating(struct amdgpu_device *adev, int inst)
+static void vcn_v3_0_enable_static_power_gating(struct amdgpu_vcn_inst *vinst)
 {
+	struct amdgpu_device *adev = vinst->adev;
+	int inst = vinst->inst;
 	uint32_t data;
 
 	if (adev->pg_flags & AMD_PG_SUPPORT_VCN) {
@@ -672,13 +771,14 @@ static void vcn_v3_0_enable_static_power_gating(struct amdgpu_device *adev, int
 /**
  * vcn_v3_0_disable_clock_gating - disable VCN clock gating
  *
- * @adev: amdgpu_device pointer
- * @inst: instance number
+ * @vinst: Pointer to the VCN instance structure
  *
  * Disable clock gating for VCN block
  */
-static void vcn_v3_0_disable_clock_gating(struct amdgpu_device *adev, int inst)
+static void vcn_v3_0_disable_clock_gating(struct amdgpu_vcn_inst *vinst)
 {
+	struct amdgpu_device *adev = vinst->adev;
+	int inst = vinst->inst;
 	uint32_t data;
 
 	/* VCN disable CGC */
@@ -805,9 +905,12 @@ static void vcn_v3_0_disable_clock_gating(struct amdgpu_device *adev, int inst)
 	WREG32_SOC15(VCN, inst, mmUVD_SUVD_CGC_CTRL, data);
 }
 
-static void vcn_v3_0_clock_gating_dpg_mode(struct amdgpu_device *adev,
-		uint8_t sram_sel, int inst_idx, uint8_t indirect)
+static void vcn_v3_0_clock_gating_dpg_mode(struct amdgpu_vcn_inst *vinst,
+					   uint8_t sram_sel,
+					   uint8_t indirect)
 {
+	struct amdgpu_device *adev = vinst->adev;
+	int inst_idx = vinst->inst;
 	uint32_t reg_data = 0;
 
 	/* enable sw clock gating control */
@@ -856,13 +959,14 @@ static void vcn_v3_0_clock_gating_dpg_mode(struct amdgpu_device *adev,
 /**
  * vcn_v3_0_enable_clock_gating - enable VCN clock gating
  *
- * @adev: amdgpu_device pointer
- * @inst: instance number
+ * @vinst: Pointer to the VCN instance structure
  *
  * Enable clock gating for VCN block
  */
-static void vcn_v3_0_enable_clock_gating(struct amdgpu_device *adev, int inst)
+static void vcn_v3_0_enable_clock_gating(struct amdgpu_vcn_inst *vinst)
 {
+	struct amdgpu_device *adev = vinst->adev;
+	int inst = vinst->inst;
 	uint32_t data;
 
 	/* enable VCN CGC */
@@ -921,11 +1025,14 @@ static void vcn_v3_0_enable_clock_gating(struct amdgpu_device *adev, int inst)
 	WREG32_SOC15(VCN, inst, mmUVD_SUVD_CGC_CTRL, data);
 }
 
-static int vcn_v3_0_start_dpg_mode(struct amdgpu_device *adev, int inst_idx, bool indirect)
+static int vcn_v3_0_start_dpg_mode(struct amdgpu_vcn_inst *vinst, bool indirect)
 {
-	volatile struct amdgpu_fw_shared *fw_shared = adev->vcn.inst[inst_idx].fw_shared_cpu_addr;
+	struct amdgpu_device *adev = vinst->adev;
+	int inst_idx = vinst->inst;
+	struct amdgpu_fw_shared *fw_shared = adev->vcn.inst[inst_idx].fw_shared.cpu_addr;
 	struct amdgpu_ring *ring;
 	uint32_t rb_bufsz, tmp;
+	int ret;
 
 	/* disable register anti-hang mechanism */
 	WREG32_P(SOC15_REG_OFFSET(VCN, inst_idx, mmUVD_POWER_STATUS), 1,
@@ -940,7 +1047,7 @@ static int vcn_v3_0_start_dpg_mode(struct amdgpu_device *adev, int inst_idx, boo
 		adev->vcn.inst[inst_idx].dpg_sram_curr_addr = (uint32_t *)adev->vcn.inst[inst_idx].dpg_sram_cpu_addr;
 
 	/* enable clock gating */
-	vcn_v3_0_clock_gating_dpg_mode(adev, 0, inst_idx, indirect);
+	vcn_v3_0_clock_gating_dpg_mode(vinst, 0, indirect);
 
 	/* enable VCPU clock */
 	tmp = (0xFF << UVD_VCPU_CNTL__PRB_TIMEOUT_VAL__SHIFT);
@@ -989,7 +1096,7 @@ static int vcn_v3_0_start_dpg_mode(struct amdgpu_device *adev, int inst_idx, boo
 		 (0x1 << UVD_MPC_SET_MUX__SET_1__SHIFT) |
 		 (0x2 << UVD_MPC_SET_MUX__SET_2__SHIFT)), 0, indirect);
 
-	vcn_v3_0_mc_resume_dpg_mode(adev, inst_idx, indirect);
+	vcn_v3_0_mc_resume_dpg_mode(vinst, indirect);
 
 	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, inst_idx, mmUVD_REG_XX_MASK), 0x10, 0, indirect);
@@ -1018,10 +1125,13 @@ static int vcn_v3_0_start_dpg_mode(struct amdgpu_device *adev, int inst_idx, boo
 	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
 		VCN, inst_idx, mmUVD_VCPU_CNTL), tmp, 0, indirect);
 
-	if (indirect)
-		psp_update_vcn_sram(adev, inst_idx, adev->vcn.inst[inst_idx].dpg_sram_gpu_addr,
-			(uint32_t)((uintptr_t)adev->vcn.inst[inst_idx].dpg_sram_curr_addr -
-				(uintptr_t)adev->vcn.inst[inst_idx].dpg_sram_cpu_addr));
+	if (indirect) {
+		ret = amdgpu_vcn_psp_update_sram(adev, inst_idx, 0);
+		if (ret) {
+			dev_err(adev->dev, "vcn sram load failed %d\n", ret);
+			return ret;
+		}
+	}
 
 	ring = &adev->vcn.inst[inst_idx].ring_dec;
 	/* force RBC into idle state */
@@ -1072,194 +1182,203 @@ static int vcn_v3_0_start_dpg_mode(struct amdgpu_device *adev, int inst_idx, boo
 	WREG32_P(SOC15_REG_OFFSET(VCN, inst_idx, mmUVD_POWER_STATUS),
 		0, ~UVD_POWER_STATUS__STALL_DPG_POWER_UP_MASK);
 
+	/* Keeping one read-back to ensure all register writes are done,
+	 * otherwise it may introduce race conditions.
+	 */
+	RREG32_SOC15(VCN, inst_idx, mmUVD_STATUS);
+
 	return 0;
 }
 
-static int vcn_v3_0_start(struct amdgpu_device *adev)
+static int vcn_v3_0_start(struct amdgpu_vcn_inst *vinst)
 {
-	volatile struct amdgpu_fw_shared *fw_shared;
+	struct amdgpu_device *adev = vinst->adev;
+	int i = vinst->inst;
+	struct amdgpu_fw_shared *fw_shared;
 	struct amdgpu_ring *ring;
 	uint32_t rb_bufsz, tmp;
-	int i, j, k, r;
+	int j, k, r;
+
+	if (adev->vcn.harvest_config & (1 << i))
+		return 0;
 
 	if (adev->pm.dpm_enabled)
-		amdgpu_dpm_enable_uvd(adev, true);
+		amdgpu_dpm_enable_vcn(adev, true, i);
 
-	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
-		if (adev->vcn.harvest_config & (1 << i))
-			continue;
+	if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)
+		return vcn_v3_0_start_dpg_mode(vinst, vinst->indirect_sram);
 
-		if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG){
-			r = vcn_v3_0_start_dpg_mode(adev, i, adev->vcn.indirect_sram);
-			continue;
-		}
+	/* disable VCN power gating */
+	vcn_v3_0_disable_static_power_gating(vinst);
+
+	/* set VCN status busy */
+	tmp = RREG32_SOC15(VCN, i, mmUVD_STATUS) | UVD_STATUS__UVD_BUSY;
+	WREG32_SOC15(VCN, i, mmUVD_STATUS, tmp);
 
-		/* disable VCN power gating */
-		vcn_v3_0_disable_static_power_gating(adev, i);
+	/* SW clock gating */
+	vcn_v3_0_disable_clock_gating(vinst);
 
-		/* set VCN status busy */
-		tmp = RREG32_SOC15(VCN, i, mmUVD_STATUS) | UVD_STATUS__UVD_BUSY;
-		WREG32_SOC15(VCN, i, mmUVD_STATUS, tmp);
+	/* enable VCPU clock */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL),
+		 UVD_VCPU_CNTL__CLK_EN_MASK, ~UVD_VCPU_CNTL__CLK_EN_MASK);
 
-		/*SW clock gating */
-		vcn_v3_0_disable_clock_gating(adev, i);
+	/* disable master interrupt */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_MASTINT_EN), 0,
+		 ~UVD_MASTINT_EN__VCPU_EN_MASK);
 
-		/* enable VCPU clock */
-		WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL),
-			UVD_VCPU_CNTL__CLK_EN_MASK, ~UVD_VCPU_CNTL__CLK_EN_MASK);
-
-		/* disable master interrupt */
-		WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_MASTINT_EN), 0,
-			~UVD_MASTINT_EN__VCPU_EN_MASK);
-
-		/* enable LMI MC and UMC channels */
-		WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_LMI_CTRL2), 0,
-			~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK);
-
-		tmp = RREG32_SOC15(VCN, i, mmUVD_SOFT_RESET);
-		tmp &= ~UVD_SOFT_RESET__LMI_SOFT_RESET_MASK;
-		tmp &= ~UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK;
-		WREG32_SOC15(VCN, i, mmUVD_SOFT_RESET, tmp);
-
-		/* setup mmUVD_LMI_CTRL */
-		tmp = RREG32_SOC15(VCN, i, mmUVD_LMI_CTRL);
-		WREG32_SOC15(VCN, i, mmUVD_LMI_CTRL, tmp |
-			UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK	|
-			UVD_LMI_CTRL__MASK_MC_URGENT_MASK |
-			UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK |
-			UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK);
-
-		/* setup mmUVD_MPC_CNTL */
-		tmp = RREG32_SOC15(VCN, i, mmUVD_MPC_CNTL);
-		tmp &= ~UVD_MPC_CNTL__REPLACEMENT_MODE_MASK;
-		tmp |= 0x2 << UVD_MPC_CNTL__REPLACEMENT_MODE__SHIFT;
-		WREG32_SOC15(VCN, i, mmUVD_MPC_CNTL, tmp);
-
-		/* setup UVD_MPC_SET_MUXA0 */
-		WREG32_SOC15(VCN, i, mmUVD_MPC_SET_MUXA0,
-			((0x1 << UVD_MPC_SET_MUXA0__VARA_1__SHIFT) |
-			(0x2 << UVD_MPC_SET_MUXA0__VARA_2__SHIFT) |
-			(0x3 << UVD_MPC_SET_MUXA0__VARA_3__SHIFT) |
-			(0x4 << UVD_MPC_SET_MUXA0__VARA_4__SHIFT)));
-
-		/* setup UVD_MPC_SET_MUXB0 */
-		WREG32_SOC15(VCN, i, mmUVD_MPC_SET_MUXB0,
-			((0x1 << UVD_MPC_SET_MUXB0__VARB_1__SHIFT) |
-			(0x2 << UVD_MPC_SET_MUXB0__VARB_2__SHIFT) |
-			(0x3 << UVD_MPC_SET_MUXB0__VARB_3__SHIFT) |
-			(0x4 << UVD_MPC_SET_MUXB0__VARB_4__SHIFT)));
-
-		/* setup mmUVD_MPC_SET_MUX */
-		WREG32_SOC15(VCN, i, mmUVD_MPC_SET_MUX,
-			((0x0 << UVD_MPC_SET_MUX__SET_0__SHIFT) |
-			(0x1 << UVD_MPC_SET_MUX__SET_1__SHIFT) |
-			(0x2 << UVD_MPC_SET_MUX__SET_2__SHIFT)));
-
-		vcn_v3_0_mc_resume(adev, i);
-
-		/* VCN global tiling registers */
-		WREG32_SOC15(VCN, i, mmUVD_GFX10_ADDR_CONFIG,
-			adev->gfx.config.gb_addr_config);
-
-		/* unblock VCPU register access */
-		WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_RB_ARB_CTRL), 0,
-			~UVD_RB_ARB_CTRL__VCPU_DIS_MASK);
-
-		/* release VCPU reset to boot */
-		WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL), 0,
-			~UVD_VCPU_CNTL__BLK_RST_MASK);
+	/* enable LMI MC and UMC channels */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_LMI_CTRL2), 0,
+		 ~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK);
 
-		for (j = 0; j < 10; ++j) {
-			uint32_t status;
+	tmp = RREG32_SOC15(VCN, i, mmUVD_SOFT_RESET);
+	tmp &= ~UVD_SOFT_RESET__LMI_SOFT_RESET_MASK;
+	tmp &= ~UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK;
+	WREG32_SOC15(VCN, i, mmUVD_SOFT_RESET, tmp);
 
-			for (k = 0; k < 100; ++k) {
-				status = RREG32_SOC15(VCN, i, mmUVD_STATUS);
-				if (status & 2)
-					break;
-				mdelay(10);
-			}
-			r = 0;
-			if (status & 2)
-				break;
+	/* setup mmUVD_LMI_CTRL */
+	tmp = RREG32_SOC15(VCN, i, mmUVD_LMI_CTRL);
+	WREG32_SOC15(VCN, i, mmUVD_LMI_CTRL, tmp |
+		     UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK	|
+		     UVD_LMI_CTRL__MASK_MC_URGENT_MASK |
+		     UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK |
+		     UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK);
+
+	/* setup mmUVD_MPC_CNTL */
+	tmp = RREG32_SOC15(VCN, i, mmUVD_MPC_CNTL);
+	tmp &= ~UVD_MPC_CNTL__REPLACEMENT_MODE_MASK;
+	tmp |= 0x2 << UVD_MPC_CNTL__REPLACEMENT_MODE__SHIFT;
+	WREG32_SOC15(VCN, i, mmUVD_MPC_CNTL, tmp);
+
+	/* setup UVD_MPC_SET_MUXA0 */
+	WREG32_SOC15(VCN, i, mmUVD_MPC_SET_MUXA0,
+		     ((0x1 << UVD_MPC_SET_MUXA0__VARA_1__SHIFT) |
+		      (0x2 << UVD_MPC_SET_MUXA0__VARA_2__SHIFT) |
+		      (0x3 << UVD_MPC_SET_MUXA0__VARA_3__SHIFT) |
+		      (0x4 << UVD_MPC_SET_MUXA0__VARA_4__SHIFT)));
+
+	/* setup UVD_MPC_SET_MUXB0 */
+	WREG32_SOC15(VCN, i, mmUVD_MPC_SET_MUXB0,
+		     ((0x1 << UVD_MPC_SET_MUXB0__VARB_1__SHIFT) |
+		      (0x2 << UVD_MPC_SET_MUXB0__VARB_2__SHIFT) |
+		      (0x3 << UVD_MPC_SET_MUXB0__VARB_3__SHIFT) |
+		      (0x4 << UVD_MPC_SET_MUXB0__VARB_4__SHIFT)));
+
+	/* setup mmUVD_MPC_SET_MUX */
+	WREG32_SOC15(VCN, i, mmUVD_MPC_SET_MUX,
+		     ((0x0 << UVD_MPC_SET_MUX__SET_0__SHIFT) |
+		      (0x1 << UVD_MPC_SET_MUX__SET_1__SHIFT) |
+		      (0x2 << UVD_MPC_SET_MUX__SET_2__SHIFT)));
+
+	vcn_v3_0_mc_resume(vinst);
 
-			DRM_ERROR("VCN[%d] decode not responding, trying to reset the VCPU!!!\n", i);
-			WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL),
-				UVD_VCPU_CNTL__BLK_RST_MASK,
-				~UVD_VCPU_CNTL__BLK_RST_MASK);
-			mdelay(10);
-			WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL), 0,
-				~UVD_VCPU_CNTL__BLK_RST_MASK);
+	/* VCN global tiling registers */
+	WREG32_SOC15(VCN, i, mmUVD_GFX10_ADDR_CONFIG,
+		     adev->gfx.config.gb_addr_config);
+
+	/* unblock VCPU register access */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_RB_ARB_CTRL), 0,
+		 ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK);
 
+	/* release VCPU reset to boot */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL), 0,
+		 ~UVD_VCPU_CNTL__BLK_RST_MASK);
+
+	for (j = 0; j < 10; ++j) {
+		uint32_t status;
+
+		for (k = 0; k < 100; ++k) {
+			status = RREG32_SOC15(VCN, i, mmUVD_STATUS);
+			if (status & 2)
+				break;
 			mdelay(10);
-			r = -1;
 		}
+		r = 0;
+		if (status & 2)
+			break;
 
-		if (r) {
-			DRM_ERROR("VCN[%d] decode not responding, giving up!!!\n", i);
-			return r;
-		}
+		DRM_ERROR("VCN[%d] decode not responding, trying to reset the VCPU!!!\n", i);
+		WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL),
+			 UVD_VCPU_CNTL__BLK_RST_MASK,
+			 ~UVD_VCPU_CNTL__BLK_RST_MASK);
+		mdelay(10);
+		WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL), 0,
+			 ~UVD_VCPU_CNTL__BLK_RST_MASK);
 
-		/* enable master interrupt */
-		WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_MASTINT_EN),
-			UVD_MASTINT_EN__VCPU_EN_MASK,
-			~UVD_MASTINT_EN__VCPU_EN_MASK);
+		mdelay(10);
+		r = -1;
+	}
 
-		/* clear the busy bit of VCN_STATUS */
-		WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_STATUS), 0,
-			~(2 << UVD_STATUS__VCPU_REPORT__SHIFT));
+	if (r) {
+		DRM_ERROR("VCN[%d] decode not responding, giving up!!!\n", i);
+		return r;
+	}
 
-		WREG32_SOC15(VCN, i, mmUVD_LMI_RBC_RB_VMID, 0);
+	/* enable master interrupt */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_MASTINT_EN),
+		 UVD_MASTINT_EN__VCPU_EN_MASK,
+		 ~UVD_MASTINT_EN__VCPU_EN_MASK);
 
-		ring = &adev->vcn.inst[i].ring_dec;
-		/* force RBC into idle state */
-		rb_bufsz = order_base_2(ring->ring_size);
-		tmp = REG_SET_FIELD(0, UVD_RBC_RB_CNTL, RB_BUFSZ, rb_bufsz);
-		tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_BLKSZ, 1);
-		tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_NO_FETCH, 1);
-		tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_NO_UPDATE, 1);
-		tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_RPTR_WR_EN, 1);
-		WREG32_SOC15(VCN, i, mmUVD_RBC_RB_CNTL, tmp);
+	/* clear the busy bit of VCN_STATUS */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_STATUS), 0,
+		 ~(2 << UVD_STATUS__VCPU_REPORT__SHIFT));
 
-		fw_shared = adev->vcn.inst[i].fw_shared_cpu_addr;
-		fw_shared->multi_queue.decode_queue_mode |= cpu_to_le32(FW_QUEUE_RING_RESET);
+	WREG32_SOC15(VCN, i, mmUVD_LMI_RBC_RB_VMID, 0);
 
-		/* programm the RB_BASE for ring buffer */
-		WREG32_SOC15(VCN, i, mmUVD_LMI_RBC_RB_64BIT_BAR_LOW,
-			lower_32_bits(ring->gpu_addr));
-		WREG32_SOC15(VCN, i, mmUVD_LMI_RBC_RB_64BIT_BAR_HIGH,
-			upper_32_bits(ring->gpu_addr));
+	ring = &adev->vcn.inst[i].ring_dec;
+	/* force RBC into idle state */
+	rb_bufsz = order_base_2(ring->ring_size);
+	tmp = REG_SET_FIELD(0, UVD_RBC_RB_CNTL, RB_BUFSZ, rb_bufsz);
+	tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_BLKSZ, 1);
+	tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_NO_FETCH, 1);
+	tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_NO_UPDATE, 1);
+	tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_RPTR_WR_EN, 1);
+	WREG32_SOC15(VCN, i, mmUVD_RBC_RB_CNTL, tmp);
 
-		/* Initialize the ring buffer's read and write pointers */
-		WREG32_SOC15(VCN, i, mmUVD_RBC_RB_RPTR, 0);
+	fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr;
+	fw_shared->multi_queue.decode_queue_mode |= cpu_to_le32(FW_QUEUE_RING_RESET);
 
-		WREG32_SOC15(VCN, i, mmUVD_SCRATCH2, 0);
-		ring->wptr = RREG32_SOC15(VCN, i, mmUVD_RBC_RB_RPTR);
-		WREG32_SOC15(VCN, i, mmUVD_RBC_RB_WPTR,
-			lower_32_bits(ring->wptr));
-		fw_shared->rb.wptr = lower_32_bits(ring->wptr);
-		fw_shared->multi_queue.decode_queue_mode &= cpu_to_le32(~FW_QUEUE_RING_RESET);
-
-		if (adev->ip_versions[UVD_HWIP][0] != IP_VERSION(3, 0, 33)) {
-			fw_shared->multi_queue.encode_generalpurpose_queue_mode |= cpu_to_le32(FW_QUEUE_RING_RESET);
-			ring = &adev->vcn.inst[i].ring_enc[0];
-			WREG32_SOC15(VCN, i, mmUVD_RB_RPTR, lower_32_bits(ring->wptr));
-			WREG32_SOC15(VCN, i, mmUVD_RB_WPTR, lower_32_bits(ring->wptr));
-			WREG32_SOC15(VCN, i, mmUVD_RB_BASE_LO, ring->gpu_addr);
-			WREG32_SOC15(VCN, i, mmUVD_RB_BASE_HI, upper_32_bits(ring->gpu_addr));
-			WREG32_SOC15(VCN, i, mmUVD_RB_SIZE, ring->ring_size / 4);
-			fw_shared->multi_queue.encode_generalpurpose_queue_mode &= cpu_to_le32(~FW_QUEUE_RING_RESET);
-
-			fw_shared->multi_queue.encode_lowlatency_queue_mode |= cpu_to_le32(FW_QUEUE_RING_RESET);
-			ring = &adev->vcn.inst[i].ring_enc[1];
-			WREG32_SOC15(VCN, i, mmUVD_RB_RPTR2, lower_32_bits(ring->wptr));
-			WREG32_SOC15(VCN, i, mmUVD_RB_WPTR2, lower_32_bits(ring->wptr));
-			WREG32_SOC15(VCN, i, mmUVD_RB_BASE_LO2, ring->gpu_addr);
-			WREG32_SOC15(VCN, i, mmUVD_RB_BASE_HI2, upper_32_bits(ring->gpu_addr));
-			WREG32_SOC15(VCN, i, mmUVD_RB_SIZE2, ring->ring_size / 4);
-			fw_shared->multi_queue.encode_lowlatency_queue_mode &= cpu_to_le32(~FW_QUEUE_RING_RESET);
-		}
+	/* programm the RB_BASE for ring buffer */
+	WREG32_SOC15(VCN, i, mmUVD_LMI_RBC_RB_64BIT_BAR_LOW,
+		     lower_32_bits(ring->gpu_addr));
+	WREG32_SOC15(VCN, i, mmUVD_LMI_RBC_RB_64BIT_BAR_HIGH,
+		     upper_32_bits(ring->gpu_addr));
+
+	/* Initialize the ring buffer's read and write pointers */
+	WREG32_SOC15(VCN, i, mmUVD_RBC_RB_RPTR, 0);
+
+	WREG32_SOC15(VCN, i, mmUVD_SCRATCH2, 0);
+	ring->wptr = RREG32_SOC15(VCN, i, mmUVD_RBC_RB_RPTR);
+	WREG32_SOC15(VCN, i, mmUVD_RBC_RB_WPTR,
+		     lower_32_bits(ring->wptr));
+	fw_shared->rb.wptr = lower_32_bits(ring->wptr);
+	fw_shared->multi_queue.decode_queue_mode &= cpu_to_le32(~FW_QUEUE_RING_RESET);
+
+	if (amdgpu_ip_version(adev, UVD_HWIP, 0) !=
+	    IP_VERSION(3, 0, 33)) {
+		fw_shared->multi_queue.encode_generalpurpose_queue_mode |= cpu_to_le32(FW_QUEUE_RING_RESET);
+		ring = &adev->vcn.inst[i].ring_enc[0];
+		WREG32_SOC15(VCN, i, mmUVD_RB_RPTR, lower_32_bits(ring->wptr));
+		WREG32_SOC15(VCN, i, mmUVD_RB_WPTR, lower_32_bits(ring->wptr));
+		WREG32_SOC15(VCN, i, mmUVD_RB_BASE_LO, ring->gpu_addr);
+		WREG32_SOC15(VCN, i, mmUVD_RB_BASE_HI, upper_32_bits(ring->gpu_addr));
+		WREG32_SOC15(VCN, i, mmUVD_RB_SIZE, ring->ring_size / 4);
+		fw_shared->multi_queue.encode_generalpurpose_queue_mode &= cpu_to_le32(~FW_QUEUE_RING_RESET);
+
+		fw_shared->multi_queue.encode_lowlatency_queue_mode |= cpu_to_le32(FW_QUEUE_RING_RESET);
+		ring = &adev->vcn.inst[i].ring_enc[1];
+		WREG32_SOC15(VCN, i, mmUVD_RB_RPTR2, lower_32_bits(ring->wptr));
+		WREG32_SOC15(VCN, i, mmUVD_RB_WPTR2, lower_32_bits(ring->wptr));
+		WREG32_SOC15(VCN, i, mmUVD_RB_BASE_LO2, ring->gpu_addr);
+		WREG32_SOC15(VCN, i, mmUVD_RB_BASE_HI2, upper_32_bits(ring->gpu_addr));
+		WREG32_SOC15(VCN, i, mmUVD_RB_SIZE2, ring->ring_size / 4);
+		fw_shared->multi_queue.encode_lowlatency_queue_mode &= cpu_to_le32(~FW_QUEUE_RING_RESET);
 	}
 
+	/* Keeping one read-back to ensure all register writes are done,
+	 * otherwise it may introduce race conditions.
+	 */
+	RREG32_SOC15(VCN, i, mmUVD_STATUS);
+
 	return 0;
 }
 
@@ -1295,7 +1414,7 @@ static int vcn_v3_0_start_sriov(struct amdgpu_device *adev)
 
 	header.version = MMSCH_VERSION;
 	header.total_size = sizeof(struct mmsch_v3_0_init_header) >> 2;
-	for (i = 0; i < AMDGPU_MAX_VCN_INSTANCES; i++) {
+	for (i = 0; i < MMSCH_V3_0_VCN_INSTANCES; i++) {
 		header.inst[i].init_status = 0;
 		header.inst[i].table_offset = 0;
 		header.inst[i].table_size = 0;
@@ -1313,7 +1432,7 @@ static int vcn_v3_0_start_sriov(struct amdgpu_device *adev)
 			mmUVD_STATUS),
 			~UVD_STATUS__UVD_BUSY, UVD_STATUS__UVD_BUSY);
 
-		cache_size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.fw->size + 4);
+		cache_size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.inst[i].fw->size + 4);
 
 		if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
 			MMSCH_V3_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, i,
@@ -1372,7 +1491,7 @@ static int vcn_v3_0_start_sriov(struct amdgpu_device *adev)
 			mmUVD_VCPU_CACHE_SIZE2),
 			AMDGPU_VCN_CONTEXT_SIZE);
 
-		for (j = 0; j < adev->vcn.num_enc_rings; ++j) {
+		for (j = 0; j < adev->vcn.inst[i].num_enc_rings; ++j) {
 			ring = &adev->vcn.inst[i].ring_enc[j];
 			ring->wptr = 0;
 			rb_addr = ring->gpu_addr;
@@ -1472,10 +1591,15 @@ static int vcn_v3_0_start_sriov(struct amdgpu_device *adev)
 	return 0;
 }
 
-static int vcn_v3_0_stop_dpg_mode(struct amdgpu_device *adev, int inst_idx)
+static int vcn_v3_0_stop_dpg_mode(struct amdgpu_vcn_inst *vinst)
 {
+	struct amdgpu_device *adev = vinst->adev;
+	int inst_idx = vinst->inst;
+	struct dpg_pause_state state = {.fw_based = VCN_DPG_STATE__UNPAUSE};
 	uint32_t tmp;
 
+	vcn_v3_0_pause_dpg_mode(vinst, &state);
+
 	/* Wait for power status to be 1 */
 	SOC15_WAIT_ON_RREG(VCN, inst_idx, mmUVD_POWER_STATUS, 1,
 		UVD_POWER_STATUS__UVD_POWER_STATUS_MASK);
@@ -1497,88 +1621,101 @@ static int vcn_v3_0_stop_dpg_mode(struct amdgpu_device *adev, int inst_idx)
 	WREG32_P(SOC15_REG_OFFSET(VCN, inst_idx, mmUVD_POWER_STATUS), 0,
 		~UVD_POWER_STATUS__UVD_PG_MODE_MASK);
 
+	/* Keeping one read-back to ensure all register writes are done,
+	 * otherwise it may introduce race conditions.
+	 */
+	RREG32_SOC15(VCN, inst_idx, mmUVD_STATUS);
+
 	return 0;
 }
 
-static int vcn_v3_0_stop(struct amdgpu_device *adev)
+static int vcn_v3_0_stop(struct amdgpu_vcn_inst *vinst)
 {
+	struct amdgpu_device *adev = vinst->adev;
+	int i = vinst->inst;
 	uint32_t tmp;
-	int i, r = 0;
+	int r = 0;
 
-	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
-		if (adev->vcn.harvest_config & (1 << i))
-			continue;
+	if (adev->vcn.harvest_config & (1 << i))
+		return 0;
 
-		if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) {
-			r = vcn_v3_0_stop_dpg_mode(adev, i);
-			continue;
-		}
+	if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) {
+		r = vcn_v3_0_stop_dpg_mode(vinst);
+		goto done;
+	}
 
-		/* wait for vcn idle */
-		r = SOC15_WAIT_ON_RREG(VCN, i, mmUVD_STATUS, UVD_STATUS__IDLE, 0x7);
-		if (r)
-			return r;
+	/* wait for vcn idle */
+	r = SOC15_WAIT_ON_RREG(VCN, i, mmUVD_STATUS, UVD_STATUS__IDLE, 0x7);
+	if (r)
+		goto done;
 
-		tmp = UVD_LMI_STATUS__VCPU_LMI_WRITE_CLEAN_MASK |
-			UVD_LMI_STATUS__READ_CLEAN_MASK |
-			UVD_LMI_STATUS__WRITE_CLEAN_MASK |
-			UVD_LMI_STATUS__WRITE_CLEAN_RAW_MASK;
-		r = SOC15_WAIT_ON_RREG(VCN, i, mmUVD_LMI_STATUS, tmp, tmp);
-		if (r)
-			return r;
+	tmp = UVD_LMI_STATUS__VCPU_LMI_WRITE_CLEAN_MASK |
+		UVD_LMI_STATUS__READ_CLEAN_MASK |
+		UVD_LMI_STATUS__WRITE_CLEAN_MASK |
+		UVD_LMI_STATUS__WRITE_CLEAN_RAW_MASK;
+	r = SOC15_WAIT_ON_RREG(VCN, i, mmUVD_LMI_STATUS, tmp, tmp);
+	if (r)
+		goto done;
+
+	/* disable LMI UMC channel */
+	tmp = RREG32_SOC15(VCN, i, mmUVD_LMI_CTRL2);
+	tmp |= UVD_LMI_CTRL2__STALL_ARB_UMC_MASK;
+	WREG32_SOC15(VCN, i, mmUVD_LMI_CTRL2, tmp);
+	tmp = UVD_LMI_STATUS__UMC_READ_CLEAN_RAW_MASK|
+		UVD_LMI_STATUS__UMC_WRITE_CLEAN_RAW_MASK;
+	r = SOC15_WAIT_ON_RREG(VCN, i, mmUVD_LMI_STATUS, tmp, tmp);
+	if (r)
+		goto done;
 
-		/* disable LMI UMC channel */
-		tmp = RREG32_SOC15(VCN, i, mmUVD_LMI_CTRL2);
-		tmp |= UVD_LMI_CTRL2__STALL_ARB_UMC_MASK;
-		WREG32_SOC15(VCN, i, mmUVD_LMI_CTRL2, tmp);
-		tmp = UVD_LMI_STATUS__UMC_READ_CLEAN_RAW_MASK|
-			UVD_LMI_STATUS__UMC_WRITE_CLEAN_RAW_MASK;
-		r = SOC15_WAIT_ON_RREG(VCN, i, mmUVD_LMI_STATUS, tmp, tmp);
-		if (r)
-			return r;
+	/* block VCPU register access */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_RB_ARB_CTRL),
+		 UVD_RB_ARB_CTRL__VCPU_DIS_MASK,
+		 ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK);
 
-		/* block VCPU register access */
-		WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_RB_ARB_CTRL),
-			UVD_RB_ARB_CTRL__VCPU_DIS_MASK,
-			~UVD_RB_ARB_CTRL__VCPU_DIS_MASK);
+	/* reset VCPU */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL),
+		 UVD_VCPU_CNTL__BLK_RST_MASK,
+		 ~UVD_VCPU_CNTL__BLK_RST_MASK);
 
-		/* reset VCPU */
-		WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL),
-			UVD_VCPU_CNTL__BLK_RST_MASK,
-			~UVD_VCPU_CNTL__BLK_RST_MASK);
+	/* disable VCPU clock */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL), 0,
+		 ~(UVD_VCPU_CNTL__CLK_EN_MASK));
 
-		/* disable VCPU clock */
-		WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL), 0,
-			~(UVD_VCPU_CNTL__CLK_EN_MASK));
+	/* apply soft reset */
+	tmp = RREG32_SOC15(VCN, i, mmUVD_SOFT_RESET);
+	tmp |= UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK;
+	WREG32_SOC15(VCN, i, mmUVD_SOFT_RESET, tmp);
+	tmp = RREG32_SOC15(VCN, i, mmUVD_SOFT_RESET);
+	tmp |= UVD_SOFT_RESET__LMI_SOFT_RESET_MASK;
+	WREG32_SOC15(VCN, i, mmUVD_SOFT_RESET, tmp);
 
-		/* apply soft reset */
-		tmp = RREG32_SOC15(VCN, i, mmUVD_SOFT_RESET);
-		tmp |= UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK;
-		WREG32_SOC15(VCN, i, mmUVD_SOFT_RESET, tmp);
-		tmp = RREG32_SOC15(VCN, i, mmUVD_SOFT_RESET);
-		tmp |= UVD_SOFT_RESET__LMI_SOFT_RESET_MASK;
-		WREG32_SOC15(VCN, i, mmUVD_SOFT_RESET, tmp);
+	/* clear status */
+	WREG32_SOC15(VCN, i, mmUVD_STATUS, 0);
 
-		/* clear status */
-		WREG32_SOC15(VCN, i, mmUVD_STATUS, 0);
+	/* apply HW clock gating */
+	vcn_v3_0_enable_clock_gating(vinst);
 
-		/* apply HW clock gating */
-		vcn_v3_0_enable_clock_gating(adev, i);
+	/* enable VCN power gating */
+	vcn_v3_0_enable_static_power_gating(vinst);
 
-		/* enable VCN power gating */
-		vcn_v3_0_enable_static_power_gating(adev, i);
-	}
+	/* Keeping one read-back to ensure all register writes are done,
+	 * otherwise it may introduce race conditions.
+	 */
+	RREG32_SOC15(VCN, i, mmUVD_STATUS);
 
+done:
 	if (adev->pm.dpm_enabled)
-		amdgpu_dpm_enable_uvd(adev, false);
+		amdgpu_dpm_enable_vcn(adev, false, i);
 
-	return 0;
+	return r;
 }
 
-static int vcn_v3_0_pause_dpg_mode(struct amdgpu_device *adev,
-		   int inst_idx, struct dpg_pause_state *new_state)
+static int vcn_v3_0_pause_dpg_mode(struct amdgpu_vcn_inst *vinst,
+				   struct dpg_pause_state *new_state)
 {
-	volatile struct amdgpu_fw_shared *fw_shared;
+	struct amdgpu_device *adev = vinst->adev;
+	int inst_idx = vinst->inst;
+	struct amdgpu_fw_shared *fw_shared;
 	struct amdgpu_ring *ring;
 	uint32_t reg_data = 0;
 	int ret_code;
@@ -1609,9 +1746,10 @@ static int vcn_v3_0_pause_dpg_mode(struct amdgpu_device *adev,
 					UVD_POWER_STATUS__STALL_DPG_POWER_UP_MASK,
 					~UVD_POWER_STATUS__STALL_DPG_POWER_UP_MASK);
 
-				if (adev->ip_versions[UVD_HWIP][0] != IP_VERSION(3, 0, 33)) {
+				if (amdgpu_ip_version(adev, UVD_HWIP, 0) !=
+				    IP_VERSION(3, 0, 33)) {
 					/* Restore */
-					fw_shared = adev->vcn.inst[inst_idx].fw_shared_cpu_addr;
+					fw_shared = adev->vcn.inst[inst_idx].fw_shared.cpu_addr;
 					fw_shared->multi_queue.encode_generalpurpose_queue_mode |= cpu_to_le32(FW_QUEUE_RING_RESET);
 					ring = &adev->vcn.inst[inst_idx].ring_enc[0];
 					ring->wptr = 0;
@@ -1681,7 +1819,7 @@ static uint64_t vcn_v3_0_dec_ring_get_wptr(struct amdgpu_ring *ring)
 	struct amdgpu_device *adev = ring->adev;
 
 	if (ring->use_doorbell)
-		return adev->wb.wb[ring->wptr_offs];
+		return *ring->wptr_cpu_addr;
 	else
 		return RREG32_SOC15(VCN, ring->me, mmUVD_RBC_RB_WPTR);
 }
@@ -1696,131 +1834,77 @@ static uint64_t vcn_v3_0_dec_ring_get_wptr(struct amdgpu_ring *ring)
 static void vcn_v3_0_dec_ring_set_wptr(struct amdgpu_ring *ring)
 {
 	struct amdgpu_device *adev = ring->adev;
-	volatile struct amdgpu_fw_shared *fw_shared;
+	struct amdgpu_fw_shared *fw_shared;
 
 	if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) {
 		/*whenever update RBC_RB_WPTR, we save the wptr in shared rb.wptr and scratch2 */
-		fw_shared = adev->vcn.inst[ring->me].fw_shared_cpu_addr;
+		fw_shared = adev->vcn.inst[ring->me].fw_shared.cpu_addr;
 		fw_shared->rb.wptr = lower_32_bits(ring->wptr);
 		WREG32_SOC15(VCN, ring->me, mmUVD_SCRATCH2,
 			lower_32_bits(ring->wptr));
 	}
 
 	if (ring->use_doorbell) {
-		adev->wb.wb[ring->wptr_offs] = lower_32_bits(ring->wptr);
+		*ring->wptr_cpu_addr = lower_32_bits(ring->wptr);
 		WDOORBELL32(ring->doorbell_index, lower_32_bits(ring->wptr));
 	} else {
 		WREG32_SOC15(VCN, ring->me, mmUVD_RBC_RB_WPTR, lower_32_bits(ring->wptr));
 	}
 }
 
-static void vcn_v3_0_dec_sw_ring_emit_fence(struct amdgpu_ring *ring, u64 addr,
-				u64 seq, uint32_t flags)
-{
-	WARN_ON(flags & AMDGPU_FENCE_FLAG_64BIT);
-
-	amdgpu_ring_write(ring, VCN_DEC_SW_CMD_FENCE);
-	amdgpu_ring_write(ring, addr);
-	amdgpu_ring_write(ring, upper_32_bits(addr));
-	amdgpu_ring_write(ring, seq);
-	amdgpu_ring_write(ring, VCN_DEC_SW_CMD_TRAP);
-}
-
-static void vcn_v3_0_dec_sw_ring_insert_end(struct amdgpu_ring *ring)
-{
-	amdgpu_ring_write(ring, VCN_DEC_SW_CMD_END);
-}
-
-static void vcn_v3_0_dec_sw_ring_emit_ib(struct amdgpu_ring *ring,
-			       struct amdgpu_job *job,
-			       struct amdgpu_ib *ib,
-			       uint32_t flags)
-{
-	uint32_t vmid = AMDGPU_JOB_GET_VMID(job);
-
-	amdgpu_ring_write(ring, VCN_DEC_SW_CMD_IB);
-	amdgpu_ring_write(ring, vmid);
-	amdgpu_ring_write(ring, lower_32_bits(ib->gpu_addr));
-	amdgpu_ring_write(ring, upper_32_bits(ib->gpu_addr));
-	amdgpu_ring_write(ring, ib->length_dw);
-}
-
-static void vcn_v3_0_dec_sw_ring_emit_reg_wait(struct amdgpu_ring *ring, uint32_t reg,
-				uint32_t val, uint32_t mask)
-{
-	amdgpu_ring_write(ring, VCN_DEC_SW_CMD_REG_WAIT);
-	amdgpu_ring_write(ring, reg << 2);
-	amdgpu_ring_write(ring, mask);
-	amdgpu_ring_write(ring, val);
-}
-
-static void vcn_v3_0_dec_sw_ring_emit_vm_flush(struct amdgpu_ring *ring,
-				uint32_t vmid, uint64_t pd_addr)
-{
-	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
-	uint32_t data0, data1, mask;
-
-	pd_addr = amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pd_addr);
-
-	/* wait for register write */
-	data0 = hub->ctx0_ptb_addr_lo32 + vmid * hub->ctx_addr_distance;
-	data1 = lower_32_bits(pd_addr);
-	mask = 0xffffffff;
-	vcn_v3_0_dec_sw_ring_emit_reg_wait(ring, data0, data1, mask);
-}
-
-static void vcn_v3_0_dec_sw_ring_emit_wreg(struct amdgpu_ring *ring, uint32_t reg, uint32_t val)
-{
-	amdgpu_ring_write(ring, VCN_DEC_SW_CMD_REG_WRITE);
-	amdgpu_ring_write(ring,	reg << 2);
-	amdgpu_ring_write(ring, val);
-}
-
 static const struct amdgpu_ring_funcs vcn_v3_0_dec_sw_ring_vm_funcs = {
 	.type = AMDGPU_RING_TYPE_VCN_DEC,
 	.align_mask = 0x3f,
 	.nop = VCN_DEC_SW_CMD_NO_OP,
-	.vmhub = AMDGPU_MMHUB_0,
+	.secure_submission_supported = true,
 	.get_rptr = vcn_v3_0_dec_ring_get_rptr,
 	.get_wptr = vcn_v3_0_dec_ring_get_wptr,
 	.set_wptr = vcn_v3_0_dec_ring_set_wptr,
 	.emit_frame_size =
 		SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 +
 		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 4 +
-		4 + /* vcn_v3_0_dec_sw_ring_emit_vm_flush */
-		5 + 5 + /* vcn_v3_0_dec_sw_ring_emit_fdec_swe x2 vm fdec_swe */
-		1, /* vcn_v3_0_dec_sw_ring_insert_end */
-	.emit_ib_size = 5, /* vcn_v3_0_dec_sw_ring_emit_ib */
-	.emit_ib = vcn_v3_0_dec_sw_ring_emit_ib,
-	.emit_fence = vcn_v3_0_dec_sw_ring_emit_fence,
-	.emit_vm_flush = vcn_v3_0_dec_sw_ring_emit_vm_flush,
+		VCN_SW_RING_EMIT_FRAME_SIZE,
+	.emit_ib_size = 5, /* vcn_dec_sw_ring_emit_ib */
+	.emit_ib = vcn_dec_sw_ring_emit_ib,
+	.emit_fence = vcn_dec_sw_ring_emit_fence,
+	.emit_vm_flush = vcn_dec_sw_ring_emit_vm_flush,
 	.test_ring = amdgpu_vcn_dec_sw_ring_test_ring,
 	.test_ib = NULL,//amdgpu_vcn_dec_sw_ring_test_ib,
 	.insert_nop = amdgpu_ring_insert_nop,
-	.insert_end = vcn_v3_0_dec_sw_ring_insert_end,
+	.insert_end = vcn_dec_sw_ring_insert_end,
 	.pad_ib = amdgpu_ring_generic_pad_ib,
 	.begin_use = amdgpu_vcn_ring_begin_use,
 	.end_use = amdgpu_vcn_ring_end_use,
-	.emit_wreg = vcn_v3_0_dec_sw_ring_emit_wreg,
-	.emit_reg_wait = vcn_v3_0_dec_sw_ring_emit_reg_wait,
+	.emit_wreg = vcn_dec_sw_ring_emit_wreg,
+	.emit_reg_wait = vcn_dec_sw_ring_emit_reg_wait,
 	.emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper,
 };
 
-static int vcn_v3_0_limit_sched(struct amdgpu_cs_parser *p)
+static int vcn_v3_0_limit_sched(struct amdgpu_cs_parser *p,
+				struct amdgpu_job *job)
 {
 	struct drm_gpu_scheduler **scheds;
+	struct dma_fence *fence;
 
-	/* The create msg must be in the first IB submitted */
-	if (atomic_read(&p->entity->fence_seq))
+	/* if VCN0 is harvested, we can't support AV1 */
+	if (p->adev->vcn.harvest_config & AMDGPU_VCN_HARVEST_VCN0)
 		return -EINVAL;
 
+	/* wait for all jobs to finish before switching to instance 0 */
+	fence = amdgpu_ctx_get_fence(p->ctx, job->base.entity, ~0ull);
+	if (fence) {
+		dma_fence_wait(fence, false);
+		dma_fence_put(fence);
+	}
+
 	scheds = p->adev->gpu_sched[AMDGPU_HW_IP_VCN_DEC]
 		[AMDGPU_RING_PRIO_DEFAULT].sched;
-	drm_sched_entity_modify_sched(p->entity, scheds, 1);
+	drm_sched_entity_modify_sched(job->base.entity, scheds, 1);
 	return 0;
 }
 
-static int vcn_v3_0_dec_msg(struct amdgpu_cs_parser *p, uint64_t addr)
+static int vcn_v3_0_dec_msg(struct amdgpu_cs_parser *p, struct amdgpu_job *job,
+			    uint64_t addr)
 {
 	struct ttm_operation_ctx ctx = { false, false };
 	struct amdgpu_bo_va_mapping *map;
@@ -1828,7 +1912,7 @@ static int vcn_v3_0_dec_msg(struct amdgpu_cs_parser *p, uint64_t addr)
 	struct amdgpu_bo *bo;
 	uint64_t start, end;
 	unsigned int i;
-	void * ptr;
+	void *ptr;
 	int r;
 
 	addr &= AMDGPU_GMC_HOLE_MASK;
@@ -1891,7 +1975,7 @@ static int vcn_v3_0_dec_msg(struct amdgpu_cs_parser *p, uint64_t addr)
 		if (create[0] == 0x7 || create[0] == 0x10 || create[0] == 0x11)
 			continue;
 
-		r = vcn_v3_0_limit_sched(p);
+		r = vcn_v3_0_limit_sched(p, job);
 		if (r)
 			goto out;
 	}
@@ -1902,10 +1986,10 @@ out:
 }
 
 static int vcn_v3_0_ring_patch_cs_in_place(struct amdgpu_cs_parser *p,
-					   uint32_t ib_idx)
+					   struct amdgpu_job *job,
+					   struct amdgpu_ib *ib)
 {
-	struct amdgpu_ring *ring = to_amdgpu_ring(p->entity->rq->sched);
-	struct amdgpu_ib *ib = &p->job->ibs[ib_idx];
+	struct amdgpu_ring *ring = amdgpu_job_ring(job);
 	uint32_t msg_lo = 0, msg_hi = 0;
 	unsigned i;
 	int r;
@@ -1915,16 +1999,17 @@ static int vcn_v3_0_ring_patch_cs_in_place(struct amdgpu_cs_parser *p,
 		return 0;
 
 	for (i = 0; i < ib->length_dw; i += 2) {
-		uint32_t reg = amdgpu_get_ib_value(p, ib_idx, i);
-		uint32_t val = amdgpu_get_ib_value(p, ib_idx, i + 1);
+		uint32_t reg = amdgpu_ib_get_value(ib, i);
+		uint32_t val = amdgpu_ib_get_value(ib, i + 1);
 
-		if (reg == PACKET0(p->adev->vcn.internal.data0, 0)) {
+		if (reg == PACKET0(p->adev->vcn.inst[ring->me].internal.data0, 0)) {
 			msg_lo = val;
-		} else if (reg == PACKET0(p->adev->vcn.internal.data1, 0)) {
+		} else if (reg == PACKET0(p->adev->vcn.inst[ring->me].internal.data1, 0)) {
 			msg_hi = val;
-		} else if (reg == PACKET0(p->adev->vcn.internal.cmd, 0) &&
+		} else if (reg == PACKET0(p->adev->vcn.inst[ring->me].internal.cmd, 0) &&
 			   val == 0) {
-			r = vcn_v3_0_dec_msg(p, ((u64)msg_hi) << 32 | msg_lo);
+			r = vcn_v3_0_dec_msg(p, job,
+					     ((u64)msg_hi) << 32 | msg_lo);
 			if (r)
 				return r;
 		}
@@ -1935,7 +2020,7 @@ static int vcn_v3_0_ring_patch_cs_in_place(struct amdgpu_cs_parser *p,
 static const struct amdgpu_ring_funcs vcn_v3_0_dec_ring_vm_funcs = {
 	.type = AMDGPU_RING_TYPE_VCN_DEC,
 	.align_mask = 0xf,
-	.vmhub = AMDGPU_MMHUB_0,
+	.secure_submission_supported = true,
 	.get_rptr = vcn_v3_0_dec_ring_get_rptr,
 	.get_wptr = vcn_v3_0_dec_ring_get_wptr,
 	.set_wptr = vcn_v3_0_dec_ring_set_wptr,
@@ -1961,6 +2046,7 @@ static const struct amdgpu_ring_funcs vcn_v3_0_dec_ring_vm_funcs = {
 	.emit_wreg = vcn_v2_0_dec_ring_emit_wreg,
 	.emit_reg_wait = vcn_v2_0_dec_ring_emit_reg_wait,
 	.emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper,
+	.reset = amdgpu_vcn_ring_reset,
 };
 
 /**
@@ -1993,12 +2079,12 @@ static uint64_t vcn_v3_0_enc_ring_get_wptr(struct amdgpu_ring *ring)
 
 	if (ring == &adev->vcn.inst[ring->me].ring_enc[0]) {
 		if (ring->use_doorbell)
-			return adev->wb.wb[ring->wptr_offs];
+			return *ring->wptr_cpu_addr;
 		else
 			return RREG32_SOC15(VCN, ring->me, mmUVD_RB_WPTR);
 	} else {
 		if (ring->use_doorbell)
-			return adev->wb.wb[ring->wptr_offs];
+			return *ring->wptr_cpu_addr;
 		else
 			return RREG32_SOC15(VCN, ring->me, mmUVD_RB_WPTR2);
 	}
@@ -2017,14 +2103,14 @@ static void vcn_v3_0_enc_ring_set_wptr(struct amdgpu_ring *ring)
 
 	if (ring == &adev->vcn.inst[ring->me].ring_enc[0]) {
 		if (ring->use_doorbell) {
-			adev->wb.wb[ring->wptr_offs] = lower_32_bits(ring->wptr);
+			*ring->wptr_cpu_addr = lower_32_bits(ring->wptr);
 			WDOORBELL32(ring->doorbell_index, lower_32_bits(ring->wptr));
 		} else {
 			WREG32_SOC15(VCN, ring->me, mmUVD_RB_WPTR, lower_32_bits(ring->wptr));
 		}
 	} else {
 		if (ring->use_doorbell) {
-			adev->wb.wb[ring->wptr_offs] = lower_32_bits(ring->wptr);
+			*ring->wptr_cpu_addr = lower_32_bits(ring->wptr);
 			WDOORBELL32(ring->doorbell_index, lower_32_bits(ring->wptr));
 		} else {
 			WREG32_SOC15(VCN, ring->me, mmUVD_RB_WPTR2, lower_32_bits(ring->wptr));
@@ -2036,7 +2122,6 @@ static const struct amdgpu_ring_funcs vcn_v3_0_enc_ring_vm_funcs = {
 	.type = AMDGPU_RING_TYPE_VCN_ENC,
 	.align_mask = 0x3f,
 	.nop = VCN_ENC_CMD_NO_OP,
-	.vmhub = AMDGPU_MMHUB_0,
 	.get_rptr = vcn_v3_0_enc_ring_get_rptr,
 	.get_wptr = vcn_v3_0_enc_ring_get_wptr,
 	.set_wptr = vcn_v3_0_enc_ring_set_wptr,
@@ -2060,6 +2145,7 @@ static const struct amdgpu_ring_funcs vcn_v3_0_enc_ring_vm_funcs = {
 	.emit_wreg = vcn_v2_0_enc_ring_emit_wreg,
 	.emit_reg_wait = vcn_v2_0_enc_ring_emit_reg_wait,
 	.emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper,
+	.reset = amdgpu_vcn_ring_reset,
 };
 
 static void vcn_v3_0_set_dec_ring_funcs(struct amdgpu_device *adev)
@@ -2075,8 +2161,6 @@ static void vcn_v3_0_set_dec_ring_funcs(struct amdgpu_device *adev)
 		else
 			adev->vcn.inst[i].ring_dec.funcs = &vcn_v3_0_dec_sw_ring_vm_funcs;
 		adev->vcn.inst[i].ring_dec.me = i;
-		DRM_INFO("VCN(%d) decode%s is enabled in VM mode\n", i,
-			  DEC_SW_RING_ENABLED?"(Software Ring)":"");
 	}
 }
 
@@ -2088,18 +2172,28 @@ static void vcn_v3_0_set_enc_ring_funcs(struct amdgpu_device *adev)
 		if (adev->vcn.harvest_config & (1 << i))
 			continue;
 
-		for (j = 0; j < adev->vcn.num_enc_rings; ++j) {
+		for (j = 0; j < adev->vcn.inst[i].num_enc_rings; ++j) {
 			adev->vcn.inst[i].ring_enc[j].funcs = &vcn_v3_0_enc_ring_vm_funcs;
 			adev->vcn.inst[i].ring_enc[j].me = i;
 		}
-		if (adev->vcn.num_enc_rings > 0)
-			DRM_INFO("VCN(%d) encode is enabled in VM mode\n", i);
 	}
 }
 
-static bool vcn_v3_0_is_idle(void *handle)
+static int vcn_v3_0_reset(struct amdgpu_vcn_inst *vinst)
+{
+	int r;
+
+	r = vcn_v3_0_stop(vinst);
+	if (r)
+		return r;
+	vcn_v3_0_enable_clock_gating(vinst);
+	vcn_v3_0_enable_static_power_gating(vinst);
+	return vcn_v3_0_start(vinst);
+}
+
+static bool vcn_v3_0_is_idle(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int i, ret = 1;
 
 	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
@@ -2112,9 +2206,9 @@ static bool vcn_v3_0_is_idle(void *handle)
 	return ret;
 }
 
-static int vcn_v3_0_wait_for_idle(void *handle)
+static int vcn_v3_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int i, ret = 0;
 
 	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
@@ -2130,54 +2224,55 @@ static int vcn_v3_0_wait_for_idle(void *handle)
 	return ret;
 }
 
-static int vcn_v3_0_set_clockgating_state(void *handle,
+static int vcn_v3_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_clockgating_state state)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-	bool enable = (state == AMD_CG_STATE_GATE) ? true : false;
+	struct amdgpu_device *adev = ip_block->adev;
+	bool enable = state == AMD_CG_STATE_GATE;
 	int i;
 
 	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		struct amdgpu_vcn_inst *vinst = &adev->vcn.inst[i];
 		if (adev->vcn.harvest_config & (1 << i))
 			continue;
 
 		if (enable) {
 			if (RREG32_SOC15(VCN, i, mmUVD_STATUS) != UVD_STATUS__IDLE)
 				return -EBUSY;
-			vcn_v3_0_enable_clock_gating(adev, i);
+			vcn_v3_0_enable_clock_gating(vinst);
 		} else {
-			vcn_v3_0_disable_clock_gating(adev, i);
+			vcn_v3_0_disable_clock_gating(vinst);
 		}
 	}
 
 	return 0;
 }
 
-static int vcn_v3_0_set_powergating_state(void *handle,
-					  enum amd_powergating_state state)
+static int vcn_v3_0_set_pg_state(struct amdgpu_vcn_inst *vinst,
+				 enum amd_powergating_state state)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-	int ret;
+	struct amdgpu_device *adev = vinst->adev;
+	int ret = 0;
 
 	/* for SRIOV, guest should not control VCN Power-gating
 	 * MMSCH FW should control Power-gating and clock-gating
 	 * guest should avoid touching CGC and PG
 	 */
 	if (amdgpu_sriov_vf(adev)) {
-		adev->vcn.cur_state = AMD_PG_STATE_UNGATE;
+		vinst->cur_state = AMD_PG_STATE_UNGATE;
 		return 0;
 	}
 
-	if(state == adev->vcn.cur_state)
+	if (state == vinst->cur_state)
 		return 0;
 
 	if (state == AMD_PG_STATE_GATE)
-		ret = vcn_v3_0_stop(adev);
+		ret = vcn_v3_0_stop(vinst);
 	else
-		ret = vcn_v3_0_start(adev);
+		ret = vcn_v3_0_start(vinst);
 
-	if(!ret)
-		adev->vcn.cur_state = state;
+	if (!ret)
+		vinst->cur_state = state;
 
 	return ret;
 }
@@ -2242,7 +2337,7 @@ static void vcn_v3_0_set_irq_funcs(struct amdgpu_device *adev)
 		if (adev->vcn.harvest_config & (1 << i))
 			continue;
 
-		adev->vcn.inst[i].irq.num_types = adev->vcn.num_enc_rings + 1;
+		adev->vcn.inst[i].irq.num_types = adev->vcn.inst[i].num_enc_rings + 1;
 		adev->vcn.inst[i].irq.funcs = &vcn_v3_0_irq_funcs;
 	}
 }
@@ -2250,7 +2345,6 @@ static void vcn_v3_0_set_irq_funcs(struct amdgpu_device *adev)
 static const struct amd_ip_funcs vcn_v3_0_ip_funcs = {
 	.name = "vcn_v3_0",
 	.early_init = vcn_v3_0_early_init,
-	.late_init = NULL,
 	.sw_init = vcn_v3_0_sw_init,
 	.sw_fini = vcn_v3_0_sw_fini,
 	.hw_init = vcn_v3_0_hw_init,
@@ -2259,16 +2353,13 @@ static const struct amd_ip_funcs vcn_v3_0_ip_funcs = {
 	.resume = vcn_v3_0_resume,
 	.is_idle = vcn_v3_0_is_idle,
 	.wait_for_idle = vcn_v3_0_wait_for_idle,
-	.check_soft_reset = NULL,
-	.pre_soft_reset = NULL,
-	.soft_reset = NULL,
-	.post_soft_reset = NULL,
 	.set_clockgating_state = vcn_v3_0_set_clockgating_state,
-	.set_powergating_state = vcn_v3_0_set_powergating_state,
+	.set_powergating_state = vcn_set_powergating_state,
+	.dump_ip_state = amdgpu_vcn_dump_ip_state,
+	.print_ip_state = amdgpu_vcn_print_ip_state,
 };
 
-const struct amdgpu_ip_block_version vcn_v3_0_ip_block =
-{
+const struct amdgpu_ip_block_version vcn_v3_0_ip_block = {
 	.type = AMD_IP_BLOCK_TYPE_VCN,
 	.major = 3,
 	.minor = 0,
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c
new file mode 100644
index 000000000000..3ae666522d57
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c
@@ -0,0 +1,2315 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/firmware.h>
+#include "amdgpu.h"
+#include "amdgpu_vcn.h"
+#include "amdgpu_pm.h"
+#include "amdgpu_cs.h"
+#include "soc15.h"
+#include "soc15d.h"
+#include "soc15_hw_ip.h"
+#include "vcn_v2_0.h"
+#include "mmsch_v4_0.h"
+#include "vcn_v4_0.h"
+
+#include "vcn/vcn_4_0_0_offset.h"
+#include "vcn/vcn_4_0_0_sh_mask.h"
+#include "ivsrcid/vcn/irqsrcs_vcn_4_0.h"
+
+#include <drm/drm_drv.h>
+
+#define mmUVD_DPG_LMA_CTL							regUVD_DPG_LMA_CTL
+#define mmUVD_DPG_LMA_CTL_BASE_IDX						regUVD_DPG_LMA_CTL_BASE_IDX
+#define mmUVD_DPG_LMA_DATA							regUVD_DPG_LMA_DATA
+#define mmUVD_DPG_LMA_DATA_BASE_IDX						regUVD_DPG_LMA_DATA_BASE_IDX
+
+#define VCN_VID_SOC_ADDRESS_2_0							0x1fb00
+#define VCN1_VID_SOC_ADDRESS_3_0						0x48300
+#define VCN1_AON_SOC_ADDRESS_3_0						0x48000
+
+#define VCN_HARVEST_MMSCH								0
+
+#define RDECODE_MSG_CREATE							0x00000000
+#define RDECODE_MESSAGE_CREATE							0x00000001
+
+static const struct amdgpu_hwip_reg_entry vcn_reg_list_4_0[] = {
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_POWER_STATUS),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_STATUS),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_CONTEXT_ID),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_CONTEXT_ID2),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_GPCOM_VCPU_DATA0),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_GPCOM_VCPU_DATA1),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_GPCOM_VCPU_CMD),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_BASE_HI),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_BASE_LO),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_BASE_HI2),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_BASE_LO2),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_BASE_HI3),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_BASE_LO3),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_BASE_HI4),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_BASE_LO4),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_RPTR),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_WPTR),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_RPTR2),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_WPTR2),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_RPTR3),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_WPTR3),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_RPTR4),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_WPTR4),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_SIZE),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_SIZE2),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_SIZE3),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_SIZE4),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_PGFSM_CONFIG),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_PGFSM_STATUS),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_DPG_LMA_CTL),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_DPG_LMA_DATA),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_DPG_LMA_MASK),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_DPG_PAUSE)
+};
+
+static int amdgpu_ih_clientid_vcns[] = {
+	SOC15_IH_CLIENTID_VCN,
+	SOC15_IH_CLIENTID_VCN1
+};
+
+static int vcn_v4_0_start_sriov(struct amdgpu_device *adev);
+static void vcn_v4_0_set_unified_ring_funcs(struct amdgpu_device *adev);
+static void vcn_v4_0_set_irq_funcs(struct amdgpu_device *adev);
+static int vcn_v4_0_set_pg_state(struct amdgpu_vcn_inst *vinst,
+				 enum amd_powergating_state state);
+static int vcn_v4_0_pause_dpg_mode(struct amdgpu_vcn_inst *vinst,
+				   struct dpg_pause_state *new_state);
+static void vcn_v4_0_unified_ring_set_wptr(struct amdgpu_ring *ring);
+static void vcn_v4_0_set_ras_funcs(struct amdgpu_device *adev);
+
+/**
+ * vcn_v4_0_early_init - set function pointers and load microcode
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Set ring and irq function pointers
+ * Load microcode from filesystem
+ */
+static int vcn_v4_0_early_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, r;
+
+	if (amdgpu_sriov_vf(adev)) {
+		adev->vcn.harvest_config = VCN_HARVEST_MMSCH;
+		for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+			if (amdgpu_vcn_is_disabled_vcn(adev, VCN_ENCODE_RING, i)) {
+				adev->vcn.harvest_config |= 1 << i;
+				dev_info(adev->dev, "VCN%d is disabled by hypervisor\n", i);
+			}
+		}
+	}
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i)
+		/* re-use enc ring as unified ring */
+		adev->vcn.inst[i].num_enc_rings = 1;
+
+	vcn_v4_0_set_unified_ring_funcs(adev);
+	vcn_v4_0_set_irq_funcs(adev);
+	vcn_v4_0_set_ras_funcs(adev);
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		adev->vcn.inst[i].set_pg_state = vcn_v4_0_set_pg_state;
+
+		r = amdgpu_vcn_early_init(adev, i);
+		if (r)
+			return r;
+	}
+
+	return 0;
+}
+
+static int vcn_v4_0_fw_shared_init(struct amdgpu_device *adev, int inst_idx)
+{
+	struct amdgpu_vcn4_fw_shared *fw_shared;
+
+	fw_shared = adev->vcn.inst[inst_idx].fw_shared.cpu_addr;
+	fw_shared->present_flag_0 = cpu_to_le32(AMDGPU_FW_SHARED_FLAG_0_UNIFIED_QUEUE);
+	fw_shared->sq.is_enabled = 1;
+
+	fw_shared->present_flag_0 |= cpu_to_le32(AMDGPU_VCN_SMU_DPM_INTERFACE_FLAG);
+	fw_shared->smu_dpm_interface.smu_interface_type = (adev->flags & AMD_IS_APU) ?
+		AMDGPU_VCN_SMU_DPM_INTERFACE_APU : AMDGPU_VCN_SMU_DPM_INTERFACE_DGPU;
+
+	if (amdgpu_ip_version(adev, VCN_HWIP, 0) ==
+	    IP_VERSION(4, 0, 2)) {
+		fw_shared->present_flag_0 |= AMDGPU_FW_SHARED_FLAG_0_DRM_KEY_INJECT;
+		fw_shared->drm_key_wa.method =
+			AMDGPU_DRM_KEY_INJECT_WORKAROUND_VCNFW_ASD_HANDSHAKING;
+	}
+
+	if (amdgpu_vcnfw_log)
+		amdgpu_vcn_fwlog_init(&adev->vcn.inst[inst_idx]);
+
+	return 0;
+}
+
+/**
+ * vcn_v4_0_sw_init - sw init for VCN block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Load firmware and sw initialization
+ */
+static int vcn_v4_0_sw_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_ring *ring;
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, r;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+		if (adev->vcn.harvest_config & (1 << i))
+			continue;
+
+		r = amdgpu_vcn_sw_init(adev, i);
+		if (r)
+			return r;
+
+		amdgpu_vcn_setup_ucode(adev, i);
+
+		r = amdgpu_vcn_resume(adev, i);
+		if (r)
+			return r;
+
+		/* Init instance 0 sched_score to 1, so it's scheduled after other instances */
+		if (i == 0)
+			atomic_set(&adev->vcn.inst[i].sched_score, 1);
+		else
+			atomic_set(&adev->vcn.inst[i].sched_score, 0);
+
+		/* VCN UNIFIED TRAP */
+		r = amdgpu_irq_add_id(adev, amdgpu_ih_clientid_vcns[i],
+				VCN_4_0__SRCID__UVD_ENC_GENERAL_PURPOSE, &adev->vcn.inst[i].irq);
+		if (r)
+			return r;
+
+		/* VCN POISON TRAP */
+		r = amdgpu_irq_add_id(adev, amdgpu_ih_clientid_vcns[i],
+				VCN_4_0__SRCID_UVD_POISON, &adev->vcn.inst[i].ras_poison_irq);
+		if (r)
+			return r;
+
+		ring = &adev->vcn.inst[i].ring_enc[0];
+		ring->use_doorbell = true;
+		if (amdgpu_sriov_vf(adev))
+			ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + i *
+				(adev->vcn.inst[i].num_enc_rings + 1) + 1;
+		else
+			ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 2 + 8 * i;
+		ring->vm_hub = AMDGPU_MMHUB0(0);
+		sprintf(ring->name, "vcn_unified_%d", i);
+
+		r = amdgpu_ring_init(adev, ring, 512, &adev->vcn.inst[i].irq, 0,
+						AMDGPU_RING_PRIO_0, &adev->vcn.inst[i].sched_score);
+		if (r)
+			return r;
+
+		vcn_v4_0_fw_shared_init(adev, i);
+
+		if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)
+			adev->vcn.inst[i].pause_dpg_mode = vcn_v4_0_pause_dpg_mode;
+	}
+
+	adev->vcn.supported_reset =
+		amdgpu_get_soft_full_reset_mask(&adev->vcn.inst[0].ring_enc[0]);
+	if (!amdgpu_sriov_vf(adev))
+		adev->vcn.supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE;
+
+	if (amdgpu_sriov_vf(adev)) {
+		r = amdgpu_virt_alloc_mm_table(adev);
+		if (r)
+			return r;
+	}
+
+
+	r = amdgpu_vcn_ras_sw_init(adev);
+	if (r)
+		return r;
+
+	r = amdgpu_vcn_reg_dump_init(adev, vcn_reg_list_4_0, ARRAY_SIZE(vcn_reg_list_4_0));
+	if (r)
+		return r;
+
+	r = amdgpu_vcn_sysfs_reset_mask_init(adev);
+	if (r)
+		return r;
+
+	return 0;
+}
+
+/**
+ * vcn_v4_0_sw_fini - sw fini for VCN block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * VCN suspend and free up sw allocation
+ */
+static int vcn_v4_0_sw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, r, idx;
+
+	if (drm_dev_enter(adev_to_drm(adev), &idx)) {
+		for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+			struct amdgpu_vcn4_fw_shared *fw_shared;
+
+			if (adev->vcn.harvest_config & (1 << i))
+				continue;
+
+			fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr;
+			fw_shared->present_flag_0 = 0;
+			fw_shared->sq.is_enabled = 0;
+		}
+
+		drm_dev_exit(idx);
+	}
+
+	if (amdgpu_sriov_vf(adev))
+		amdgpu_virt_free_mm_table(adev);
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+		r = amdgpu_vcn_suspend(adev, i);
+		if (r)
+			return r;
+	}
+
+	amdgpu_vcn_sysfs_reset_mask_fini(adev);
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++)
+		amdgpu_vcn_sw_fini(adev, i);
+
+	return 0;
+}
+
+/**
+ * vcn_v4_0_hw_init - start and test VCN block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Initialize the hardware, boot up the VCPU and do some testing
+ */
+static int vcn_v4_0_hw_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	struct amdgpu_ring *ring;
+	int i, r;
+
+	if (amdgpu_sriov_vf(adev)) {
+		r = vcn_v4_0_start_sriov(adev);
+		if (r)
+			return r;
+
+		for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+			if (adev->vcn.harvest_config & (1 << i))
+				continue;
+
+			ring = &adev->vcn.inst[i].ring_enc[0];
+			ring->wptr = 0;
+			ring->wptr_old = 0;
+			vcn_v4_0_unified_ring_set_wptr(ring);
+			ring->sched.ready = true;
+		}
+	} else {
+		for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+			if (adev->vcn.harvest_config & (1 << i))
+				continue;
+
+			ring = &adev->vcn.inst[i].ring_enc[0];
+
+			adev->nbio.funcs->vcn_doorbell_range(adev, ring->use_doorbell,
+					((adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 8 * i), i);
+
+			r = amdgpu_ring_test_helper(ring);
+			if (r)
+				return r;
+		}
+	}
+
+	return 0;
+}
+
+/**
+ * vcn_v4_0_hw_fini - stop the hardware block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Stop the VCN block, mark ring as not ready any more
+ */
+static int vcn_v4_0_hw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		struct amdgpu_vcn_inst *vinst = &adev->vcn.inst[i];
+
+		if (adev->vcn.harvest_config & (1 << i))
+			continue;
+
+		cancel_delayed_work_sync(&vinst->idle_work);
+
+		if (!amdgpu_sriov_vf(adev)) {
+			if ((adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) ||
+			    (vinst->cur_state != AMD_PG_STATE_GATE &&
+				 RREG32_SOC15(VCN, i, regUVD_STATUS))) {
+				vinst->set_pg_state(vinst, AMD_PG_STATE_GATE);
+			}
+		}
+		if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__VCN))
+			amdgpu_irq_put(adev, &vinst->ras_poison_irq, 0);
+	}
+
+	return 0;
+}
+
+/**
+ * vcn_v4_0_suspend - suspend VCN block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * HW fini and suspend VCN block
+ */
+static int vcn_v4_0_suspend(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int r, i;
+
+	r = vcn_v4_0_hw_fini(ip_block);
+	if (r)
+		return r;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+		r = amdgpu_vcn_suspend(ip_block->adev, i);
+		if (r)
+			return r;
+	}
+
+	return 0;
+}
+
+/**
+ * vcn_v4_0_resume - resume VCN block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Resume firmware and hw init VCN block
+ */
+static int vcn_v4_0_resume(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int r, i;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+		r = amdgpu_vcn_resume(ip_block->adev, i);
+		if (r)
+			return r;
+	}
+
+	r = vcn_v4_0_hw_init(ip_block);
+
+	return r;
+}
+
+/**
+ * vcn_v4_0_mc_resume - memory controller programming
+ *
+ * @vinst: VCN instance
+ *
+ * Let the VCN memory controller know it's offsets
+ */
+static void vcn_v4_0_mc_resume(struct amdgpu_vcn_inst *vinst)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int inst = vinst->inst;
+	uint32_t offset, size;
+	const struct common_firmware_header *hdr;
+
+	hdr = (const struct common_firmware_header *)adev->vcn.inst[inst].fw->data;
+	size = AMDGPU_GPU_PAGE_ALIGN(le32_to_cpu(hdr->ucode_size_bytes) + 8);
+
+	/* cache window 0: fw */
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
+		WREG32_SOC15(VCN, inst, regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW,
+			(adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + inst].tmr_mc_addr_lo));
+		WREG32_SOC15(VCN, inst, regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH,
+			(adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + inst].tmr_mc_addr_hi));
+		WREG32_SOC15(VCN, inst, regUVD_VCPU_CACHE_OFFSET0, 0);
+		offset = 0;
+	} else {
+		WREG32_SOC15(VCN, inst, regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW,
+			lower_32_bits(adev->vcn.inst[inst].gpu_addr));
+		WREG32_SOC15(VCN, inst, regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH,
+			upper_32_bits(adev->vcn.inst[inst].gpu_addr));
+		offset = size;
+                WREG32_SOC15(VCN, inst, regUVD_VCPU_CACHE_OFFSET0, AMDGPU_UVD_FIRMWARE_OFFSET >> 3);
+	}
+	WREG32_SOC15(VCN, inst, regUVD_VCPU_CACHE_SIZE0, size);
+
+	/* cache window 1: stack */
+	WREG32_SOC15(VCN, inst, regUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW,
+		lower_32_bits(adev->vcn.inst[inst].gpu_addr + offset));
+	WREG32_SOC15(VCN, inst, regUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH,
+		upper_32_bits(adev->vcn.inst[inst].gpu_addr + offset));
+	WREG32_SOC15(VCN, inst, regUVD_VCPU_CACHE_OFFSET1, 0);
+	WREG32_SOC15(VCN, inst, regUVD_VCPU_CACHE_SIZE1, AMDGPU_VCN_STACK_SIZE);
+
+	/* cache window 2: context */
+	WREG32_SOC15(VCN, inst, regUVD_LMI_VCPU_CACHE2_64BIT_BAR_LOW,
+		lower_32_bits(adev->vcn.inst[inst].gpu_addr + offset + AMDGPU_VCN_STACK_SIZE));
+	WREG32_SOC15(VCN, inst, regUVD_LMI_VCPU_CACHE2_64BIT_BAR_HIGH,
+		upper_32_bits(adev->vcn.inst[inst].gpu_addr + offset + AMDGPU_VCN_STACK_SIZE));
+	WREG32_SOC15(VCN, inst, regUVD_VCPU_CACHE_OFFSET2, 0);
+	WREG32_SOC15(VCN, inst, regUVD_VCPU_CACHE_SIZE2, AMDGPU_VCN_CONTEXT_SIZE);
+
+	/* non-cache window */
+	WREG32_SOC15(VCN, inst, regUVD_LMI_VCPU_NC0_64BIT_BAR_LOW,
+		lower_32_bits(adev->vcn.inst[inst].fw_shared.gpu_addr));
+	WREG32_SOC15(VCN, inst, regUVD_LMI_VCPU_NC0_64BIT_BAR_HIGH,
+		upper_32_bits(adev->vcn.inst[inst].fw_shared.gpu_addr));
+	WREG32_SOC15(VCN, inst, regUVD_VCPU_NONCACHE_OFFSET0, 0);
+	WREG32_SOC15(VCN, inst, regUVD_VCPU_NONCACHE_SIZE0,
+		AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_vcn4_fw_shared)));
+}
+
+/**
+ * vcn_v4_0_mc_resume_dpg_mode - memory controller programming for dpg mode
+ *
+ * @vinst: VCN instance
+ * @indirect: indirectly write sram
+ *
+ * Let the VCN memory controller know it's offsets with dpg mode
+ */
+static void vcn_v4_0_mc_resume_dpg_mode(struct amdgpu_vcn_inst *vinst,
+					bool indirect)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int inst_idx = vinst->inst;
+	uint32_t offset, size;
+	const struct common_firmware_header *hdr;
+	hdr = (const struct common_firmware_header *)adev->vcn.inst[inst_idx].fw->data;
+	size = AMDGPU_GPU_PAGE_ALIGN(le32_to_cpu(hdr->ucode_size_bytes) + 8);
+
+	/* cache window 0: fw */
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
+		if (!indirect) {
+			WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+				VCN, inst_idx, regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW),
+				(adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + inst_idx].tmr_mc_addr_lo), 0, indirect);
+			WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+				VCN, inst_idx, regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH),
+				(adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + inst_idx].tmr_mc_addr_hi), 0, indirect);
+			WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+				VCN, inst_idx, regUVD_VCPU_CACHE_OFFSET0), 0, 0, indirect);
+		} else {
+			WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+				VCN, inst_idx, regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW), 0, 0, indirect);
+			WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+				VCN, inst_idx, regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH), 0, 0, indirect);
+			WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+				VCN, inst_idx, regUVD_VCPU_CACHE_OFFSET0), 0, 0, indirect);
+		}
+		offset = 0;
+	} else {
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW),
+			lower_32_bits(adev->vcn.inst[inst_idx].gpu_addr), 0, indirect);
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH),
+			upper_32_bits(adev->vcn.inst[inst_idx].gpu_addr), 0, indirect);
+		offset = size;
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_VCPU_CACHE_OFFSET0),
+			AMDGPU_UVD_FIRMWARE_OFFSET >> 3, 0, indirect);
+	}
+
+	if (!indirect)
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_VCPU_CACHE_SIZE0), size, 0, indirect);
+	else
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_VCPU_CACHE_SIZE0), 0, 0, indirect);
+
+	/* cache window 1: stack */
+	if (!indirect) {
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW),
+			lower_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset), 0, indirect);
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH),
+			upper_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset), 0, indirect);
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_VCPU_CACHE_OFFSET1), 0, 0, indirect);
+	} else {
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW), 0, 0, indirect);
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH), 0, 0, indirect);
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_VCPU_CACHE_OFFSET1), 0, 0, indirect);
+	}
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_VCPU_CACHE_SIZE1), AMDGPU_VCN_STACK_SIZE, 0, indirect);
+
+	/* cache window 2: context */
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_LMI_VCPU_CACHE2_64BIT_BAR_LOW),
+			lower_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset + AMDGPU_VCN_STACK_SIZE), 0, indirect);
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_LMI_VCPU_CACHE2_64BIT_BAR_HIGH),
+			upper_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset + AMDGPU_VCN_STACK_SIZE), 0, indirect);
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_VCPU_CACHE_OFFSET2), 0, 0, indirect);
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_VCPU_CACHE_SIZE2), AMDGPU_VCN_CONTEXT_SIZE, 0, indirect);
+
+	/* non-cache window */
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_LMI_VCPU_NC0_64BIT_BAR_LOW),
+			lower_32_bits(adev->vcn.inst[inst_idx].fw_shared.gpu_addr), 0, indirect);
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_LMI_VCPU_NC0_64BIT_BAR_HIGH),
+			upper_32_bits(adev->vcn.inst[inst_idx].fw_shared.gpu_addr), 0, indirect);
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_VCPU_NONCACHE_OFFSET0), 0, 0, indirect);
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_VCPU_NONCACHE_SIZE0),
+			AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_vcn4_fw_shared)), 0, indirect);
+
+	/* VCN global tiling registers */
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_GFX10_ADDR_CONFIG),
+			adev->gfx.config.gb_addr_config, 0, indirect);
+}
+
+/**
+ * vcn_v4_0_disable_static_power_gating - disable VCN static power gating
+ *
+ * @vinst: VCN instance
+ *
+ * Disable static power gating for VCN block
+ */
+static void vcn_v4_0_disable_static_power_gating(struct amdgpu_vcn_inst *vinst)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int inst = vinst->inst;
+	uint32_t data = 0;
+
+	if (adev->pg_flags & AMD_PG_SUPPORT_VCN) {
+		data = (1 << UVD_PGFSM_CONFIG__UVDM_PWR_CONFIG__SHIFT
+			| 1 << UVD_PGFSM_CONFIG__UVDS_PWR_CONFIG__SHIFT
+			| 1 << UVD_PGFSM_CONFIG__UVDLM_PWR_CONFIG__SHIFT
+			| 2 << UVD_PGFSM_CONFIG__UVDF_PWR_CONFIG__SHIFT
+			| 2 << UVD_PGFSM_CONFIG__UVDTC_PWR_CONFIG__SHIFT
+			| 2 << UVD_PGFSM_CONFIG__UVDB_PWR_CONFIG__SHIFT
+			| 2 << UVD_PGFSM_CONFIG__UVDTA_PWR_CONFIG__SHIFT
+			| 2 << UVD_PGFSM_CONFIG__UVDTD_PWR_CONFIG__SHIFT
+			| 2 << UVD_PGFSM_CONFIG__UVDTE_PWR_CONFIG__SHIFT
+			| 2 << UVD_PGFSM_CONFIG__UVDE_PWR_CONFIG__SHIFT
+			| 2 << UVD_PGFSM_CONFIG__UVDAB_PWR_CONFIG__SHIFT
+			| 2 << UVD_PGFSM_CONFIG__UVDTB_PWR_CONFIG__SHIFT
+			| 2 << UVD_PGFSM_CONFIG__UVDNA_PWR_CONFIG__SHIFT
+			| 2 << UVD_PGFSM_CONFIG__UVDNB_PWR_CONFIG__SHIFT);
+
+		WREG32_SOC15(VCN, inst, regUVD_PGFSM_CONFIG, data);
+		SOC15_WAIT_ON_RREG(VCN, inst, regUVD_PGFSM_STATUS,
+			UVD_PGFSM_STATUS__UVDM_UVDU_UVDLM_PWR_ON_3_0, 0x3F3FFFFF);
+	} else {
+		uint32_t value;
+
+		value = (inst) ? 0x2200800 : 0;
+		data = (1 << UVD_PGFSM_CONFIG__UVDM_PWR_CONFIG__SHIFT
+			| 1 << UVD_PGFSM_CONFIG__UVDS_PWR_CONFIG__SHIFT
+			| 1 << UVD_PGFSM_CONFIG__UVDLM_PWR_CONFIG__SHIFT
+			| 1 << UVD_PGFSM_CONFIG__UVDF_PWR_CONFIG__SHIFT
+			| 1 << UVD_PGFSM_CONFIG__UVDTC_PWR_CONFIG__SHIFT
+			| 1 << UVD_PGFSM_CONFIG__UVDB_PWR_CONFIG__SHIFT
+			| 1 << UVD_PGFSM_CONFIG__UVDTA_PWR_CONFIG__SHIFT
+			| 1 << UVD_PGFSM_CONFIG__UVDTD_PWR_CONFIG__SHIFT
+			| 1 << UVD_PGFSM_CONFIG__UVDTE_PWR_CONFIG__SHIFT
+			| 1 << UVD_PGFSM_CONFIG__UVDE_PWR_CONFIG__SHIFT
+			| 1 << UVD_PGFSM_CONFIG__UVDAB_PWR_CONFIG__SHIFT
+			| 1 << UVD_PGFSM_CONFIG__UVDTB_PWR_CONFIG__SHIFT
+			| 1 << UVD_PGFSM_CONFIG__UVDNA_PWR_CONFIG__SHIFT
+			| 1 << UVD_PGFSM_CONFIG__UVDNB_PWR_CONFIG__SHIFT);
+
+                WREG32_SOC15(VCN, inst, regUVD_PGFSM_CONFIG, data);
+                SOC15_WAIT_ON_RREG(VCN, inst, regUVD_PGFSM_STATUS, value,  0x3F3FFFFF);
+        }
+
+        data = RREG32_SOC15(VCN, inst, regUVD_POWER_STATUS);
+        data &= ~0x103;
+        if (adev->pg_flags & AMD_PG_SUPPORT_VCN)
+                data |= UVD_PGFSM_CONFIG__UVDM_UVDU_PWR_ON |
+                        UVD_POWER_STATUS__UVD_PG_EN_MASK;
+
+        WREG32_SOC15(VCN, inst, regUVD_POWER_STATUS, data);
+
+        return;
+}
+
+/**
+ * vcn_v4_0_enable_static_power_gating - enable VCN static power gating
+ *
+ * @vinst: VCN instance
+ *
+ * Enable static power gating for VCN block
+ */
+static void vcn_v4_0_enable_static_power_gating(struct amdgpu_vcn_inst *vinst)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int inst = vinst->inst;
+	uint32_t data;
+
+	if (adev->pg_flags & AMD_PG_SUPPORT_VCN) {
+		/* Before power off, this indicator has to be turned on */
+		data = RREG32_SOC15(VCN, inst, regUVD_POWER_STATUS);
+		data &= ~UVD_POWER_STATUS__UVD_POWER_STATUS_MASK;
+		data |= UVD_POWER_STATUS__UVD_POWER_STATUS_TILES_OFF;
+		WREG32_SOC15(VCN, inst, regUVD_POWER_STATUS, data);
+
+		data = (2 << UVD_PGFSM_CONFIG__UVDM_PWR_CONFIG__SHIFT
+			| 2 << UVD_PGFSM_CONFIG__UVDS_PWR_CONFIG__SHIFT
+			| 2 << UVD_PGFSM_CONFIG__UVDF_PWR_CONFIG__SHIFT
+			| 2 << UVD_PGFSM_CONFIG__UVDTC_PWR_CONFIG__SHIFT
+			| 2 << UVD_PGFSM_CONFIG__UVDB_PWR_CONFIG__SHIFT
+			| 2 << UVD_PGFSM_CONFIG__UVDTA_PWR_CONFIG__SHIFT
+			| 2 << UVD_PGFSM_CONFIG__UVDLM_PWR_CONFIG__SHIFT
+			| 2 << UVD_PGFSM_CONFIG__UVDTD_PWR_CONFIG__SHIFT
+			| 2 << UVD_PGFSM_CONFIG__UVDTE_PWR_CONFIG__SHIFT
+			| 2 << UVD_PGFSM_CONFIG__UVDE_PWR_CONFIG__SHIFT
+			| 2 << UVD_PGFSM_CONFIG__UVDAB_PWR_CONFIG__SHIFT
+			| 2 << UVD_PGFSM_CONFIG__UVDTB_PWR_CONFIG__SHIFT
+			| 2 << UVD_PGFSM_CONFIG__UVDNA_PWR_CONFIG__SHIFT
+			| 2 << UVD_PGFSM_CONFIG__UVDNB_PWR_CONFIG__SHIFT);
+		WREG32_SOC15(VCN, inst, regUVD_PGFSM_CONFIG, data);
+
+		data = (2 << UVD_PGFSM_STATUS__UVDM_PWR_STATUS__SHIFT
+			| 2 << UVD_PGFSM_STATUS__UVDS_PWR_STATUS__SHIFT
+			| 2 << UVD_PGFSM_STATUS__UVDF_PWR_STATUS__SHIFT
+			| 2 << UVD_PGFSM_STATUS__UVDTC_PWR_STATUS__SHIFT
+			| 2 << UVD_PGFSM_STATUS__UVDB_PWR_STATUS__SHIFT
+			| 2 << UVD_PGFSM_STATUS__UVDTA_PWR_STATUS__SHIFT
+			| 2 << UVD_PGFSM_STATUS__UVDLM_PWR_STATUS__SHIFT
+			| 2 << UVD_PGFSM_STATUS__UVDTD_PWR_STATUS__SHIFT
+			| 2 << UVD_PGFSM_STATUS__UVDTE_PWR_STATUS__SHIFT
+			| 2 << UVD_PGFSM_STATUS__UVDE_PWR_STATUS__SHIFT
+			| 2 << UVD_PGFSM_STATUS__UVDAB_PWR_STATUS__SHIFT
+			| 2 << UVD_PGFSM_STATUS__UVDTB_PWR_STATUS__SHIFT
+			| 2 << UVD_PGFSM_STATUS__UVDNA_PWR_STATUS__SHIFT
+			| 2 << UVD_PGFSM_STATUS__UVDNB_PWR_STATUS__SHIFT);
+		SOC15_WAIT_ON_RREG(VCN, inst, regUVD_PGFSM_STATUS, data, 0x3F3FFFFF);
+	}
+
+        return;
+}
+
+/**
+ * vcn_v4_0_disable_clock_gating - disable VCN clock gating
+ *
+ * @vinst: VCN instance
+ *
+ * Disable clock gating for VCN block
+ */
+static void vcn_v4_0_disable_clock_gating(struct amdgpu_vcn_inst *vinst)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int inst = vinst->inst;
+	uint32_t data;
+
+	if (adev->cg_flags & AMD_CG_SUPPORT_VCN_MGCG)
+		return;
+
+	/* VCN disable CGC */
+	data = RREG32_SOC15(VCN, inst, regUVD_CGC_CTRL);
+	data &= ~UVD_CGC_CTRL__DYN_CLOCK_MODE_MASK;
+	data |= 1 << UVD_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT;
+	data |= 4 << UVD_CGC_CTRL__CLK_OFF_DELAY__SHIFT;
+	WREG32_SOC15(VCN, inst, regUVD_CGC_CTRL, data);
+
+	data = RREG32_SOC15(VCN, inst, regUVD_CGC_GATE);
+	data &= ~(UVD_CGC_GATE__SYS_MASK
+		| UVD_CGC_GATE__UDEC_MASK
+		| UVD_CGC_GATE__MPEG2_MASK
+		| UVD_CGC_GATE__REGS_MASK
+		| UVD_CGC_GATE__RBC_MASK
+		| UVD_CGC_GATE__LMI_MC_MASK
+		| UVD_CGC_GATE__LMI_UMC_MASK
+		| UVD_CGC_GATE__IDCT_MASK
+		| UVD_CGC_GATE__MPRD_MASK
+		| UVD_CGC_GATE__MPC_MASK
+		| UVD_CGC_GATE__LBSI_MASK
+		| UVD_CGC_GATE__LRBBM_MASK
+		| UVD_CGC_GATE__UDEC_RE_MASK
+		| UVD_CGC_GATE__UDEC_CM_MASK
+		| UVD_CGC_GATE__UDEC_IT_MASK
+		| UVD_CGC_GATE__UDEC_DB_MASK
+		| UVD_CGC_GATE__UDEC_MP_MASK
+		| UVD_CGC_GATE__WCB_MASK
+		| UVD_CGC_GATE__VCPU_MASK
+		| UVD_CGC_GATE__MMSCH_MASK);
+
+	WREG32_SOC15(VCN, inst, regUVD_CGC_GATE, data);
+	SOC15_WAIT_ON_RREG(VCN, inst, regUVD_CGC_GATE, 0,  0xFFFFFFFF);
+
+	data = RREG32_SOC15(VCN, inst, regUVD_CGC_CTRL);
+	data &= ~(UVD_CGC_CTRL__UDEC_RE_MODE_MASK
+		| UVD_CGC_CTRL__UDEC_CM_MODE_MASK
+		| UVD_CGC_CTRL__UDEC_IT_MODE_MASK
+		| UVD_CGC_CTRL__UDEC_DB_MODE_MASK
+		| UVD_CGC_CTRL__UDEC_MP_MODE_MASK
+		| UVD_CGC_CTRL__SYS_MODE_MASK
+		| UVD_CGC_CTRL__UDEC_MODE_MASK
+		| UVD_CGC_CTRL__MPEG2_MODE_MASK
+		| UVD_CGC_CTRL__REGS_MODE_MASK
+		| UVD_CGC_CTRL__RBC_MODE_MASK
+		| UVD_CGC_CTRL__LMI_MC_MODE_MASK
+		| UVD_CGC_CTRL__LMI_UMC_MODE_MASK
+		| UVD_CGC_CTRL__IDCT_MODE_MASK
+		| UVD_CGC_CTRL__MPRD_MODE_MASK
+		| UVD_CGC_CTRL__MPC_MODE_MASK
+		| UVD_CGC_CTRL__LBSI_MODE_MASK
+		| UVD_CGC_CTRL__LRBBM_MODE_MASK
+		| UVD_CGC_CTRL__WCB_MODE_MASK
+		| UVD_CGC_CTRL__VCPU_MODE_MASK
+		| UVD_CGC_CTRL__MMSCH_MODE_MASK);
+	WREG32_SOC15(VCN, inst, regUVD_CGC_CTRL, data);
+
+	data = RREG32_SOC15(VCN, inst, regUVD_SUVD_CGC_GATE);
+	data |= (UVD_SUVD_CGC_GATE__SRE_MASK
+		| UVD_SUVD_CGC_GATE__SIT_MASK
+		| UVD_SUVD_CGC_GATE__SMP_MASK
+		| UVD_SUVD_CGC_GATE__SCM_MASK
+		| UVD_SUVD_CGC_GATE__SDB_MASK
+		| UVD_SUVD_CGC_GATE__SRE_H264_MASK
+		| UVD_SUVD_CGC_GATE__SRE_HEVC_MASK
+		| UVD_SUVD_CGC_GATE__SIT_H264_MASK
+		| UVD_SUVD_CGC_GATE__SIT_HEVC_MASK
+		| UVD_SUVD_CGC_GATE__SCM_H264_MASK
+		| UVD_SUVD_CGC_GATE__SCM_HEVC_MASK
+		| UVD_SUVD_CGC_GATE__SDB_H264_MASK
+		| UVD_SUVD_CGC_GATE__SDB_HEVC_MASK
+		| UVD_SUVD_CGC_GATE__SCLR_MASK
+		| UVD_SUVD_CGC_GATE__UVD_SC_MASK
+		| UVD_SUVD_CGC_GATE__ENT_MASK
+		| UVD_SUVD_CGC_GATE__SIT_HEVC_DEC_MASK
+		| UVD_SUVD_CGC_GATE__SIT_HEVC_ENC_MASK
+		| UVD_SUVD_CGC_GATE__SITE_MASK
+		| UVD_SUVD_CGC_GATE__SRE_VP9_MASK
+		| UVD_SUVD_CGC_GATE__SCM_VP9_MASK
+		| UVD_SUVD_CGC_GATE__SIT_VP9_DEC_MASK
+		| UVD_SUVD_CGC_GATE__SDB_VP9_MASK
+		| UVD_SUVD_CGC_GATE__IME_HEVC_MASK);
+	WREG32_SOC15(VCN, inst, regUVD_SUVD_CGC_GATE, data);
+
+	data = RREG32_SOC15(VCN, inst, regUVD_SUVD_CGC_CTRL);
+	data &= ~(UVD_SUVD_CGC_CTRL__SRE_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SIT_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SMP_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SCM_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SDB_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SCLR_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__UVD_SC_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__ENT_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__IME_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SITE_MODE_MASK);
+	WREG32_SOC15(VCN, inst, regUVD_SUVD_CGC_CTRL, data);
+}
+
+/**
+ * vcn_v4_0_disable_clock_gating_dpg_mode - disable VCN clock gating dpg mode
+ *
+ * @vinst: VCN instance
+ * @sram_sel: sram select
+ * @indirect: indirectly write sram
+ *
+ * Disable clock gating for VCN block with dpg mode
+ */
+static void vcn_v4_0_disable_clock_gating_dpg_mode(struct amdgpu_vcn_inst *vinst,
+						   uint8_t sram_sel,
+						   uint8_t indirect)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int inst_idx = vinst->inst;
+	uint32_t reg_data = 0;
+
+	if (adev->cg_flags & AMD_CG_SUPPORT_VCN_MGCG)
+		return;
+
+	/* enable sw clock gating control */
+	reg_data = 0 << UVD_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
+	reg_data |= 1 << UVD_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT;
+	reg_data |= 4 << UVD_CGC_CTRL__CLK_OFF_DELAY__SHIFT;
+	reg_data &= ~(UVD_CGC_CTRL__UDEC_RE_MODE_MASK |
+		 UVD_CGC_CTRL__UDEC_CM_MODE_MASK |
+		 UVD_CGC_CTRL__UDEC_IT_MODE_MASK |
+		 UVD_CGC_CTRL__UDEC_DB_MODE_MASK |
+		 UVD_CGC_CTRL__UDEC_MP_MODE_MASK |
+		 UVD_CGC_CTRL__SYS_MODE_MASK |
+		 UVD_CGC_CTRL__UDEC_MODE_MASK |
+		 UVD_CGC_CTRL__MPEG2_MODE_MASK |
+		 UVD_CGC_CTRL__REGS_MODE_MASK |
+		 UVD_CGC_CTRL__RBC_MODE_MASK |
+		 UVD_CGC_CTRL__LMI_MC_MODE_MASK |
+		 UVD_CGC_CTRL__LMI_UMC_MODE_MASK |
+		 UVD_CGC_CTRL__IDCT_MODE_MASK |
+		 UVD_CGC_CTRL__MPRD_MODE_MASK |
+		 UVD_CGC_CTRL__MPC_MODE_MASK |
+		 UVD_CGC_CTRL__LBSI_MODE_MASK |
+		 UVD_CGC_CTRL__LRBBM_MODE_MASK |
+		 UVD_CGC_CTRL__WCB_MODE_MASK |
+		 UVD_CGC_CTRL__VCPU_MODE_MASK);
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_CGC_CTRL), reg_data, sram_sel, indirect);
+
+	/* turn off clock gating */
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_CGC_GATE), 0, sram_sel, indirect);
+
+	/* turn on SUVD clock gating */
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_SUVD_CGC_GATE), 1, sram_sel, indirect);
+
+	/* turn on sw mode in UVD_SUVD_CGC_CTRL */
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_SUVD_CGC_CTRL), 0, sram_sel, indirect);
+}
+
+/**
+ * vcn_v4_0_enable_clock_gating - enable VCN clock gating
+ *
+ * @vinst: VCN instance
+ *
+ * Enable clock gating for VCN block
+ */
+static void vcn_v4_0_enable_clock_gating(struct amdgpu_vcn_inst *vinst)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int inst = vinst->inst;
+	uint32_t data;
+
+	if (adev->cg_flags & AMD_CG_SUPPORT_VCN_MGCG)
+		return;
+
+	/* enable VCN CGC */
+	data = RREG32_SOC15(VCN, inst, regUVD_CGC_CTRL);
+	data |= 0 << UVD_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
+	data |= 1 << UVD_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT;
+	data |= 4 << UVD_CGC_CTRL__CLK_OFF_DELAY__SHIFT;
+	WREG32_SOC15(VCN, inst, regUVD_CGC_CTRL, data);
+
+	data = RREG32_SOC15(VCN, inst, regUVD_CGC_CTRL);
+	data |= (UVD_CGC_CTRL__UDEC_RE_MODE_MASK
+		| UVD_CGC_CTRL__UDEC_CM_MODE_MASK
+		| UVD_CGC_CTRL__UDEC_IT_MODE_MASK
+		| UVD_CGC_CTRL__UDEC_DB_MODE_MASK
+		| UVD_CGC_CTRL__UDEC_MP_MODE_MASK
+		| UVD_CGC_CTRL__SYS_MODE_MASK
+		| UVD_CGC_CTRL__UDEC_MODE_MASK
+		| UVD_CGC_CTRL__MPEG2_MODE_MASK
+		| UVD_CGC_CTRL__REGS_MODE_MASK
+		| UVD_CGC_CTRL__RBC_MODE_MASK
+		| UVD_CGC_CTRL__LMI_MC_MODE_MASK
+		| UVD_CGC_CTRL__LMI_UMC_MODE_MASK
+		| UVD_CGC_CTRL__IDCT_MODE_MASK
+		| UVD_CGC_CTRL__MPRD_MODE_MASK
+		| UVD_CGC_CTRL__MPC_MODE_MASK
+		| UVD_CGC_CTRL__LBSI_MODE_MASK
+		| UVD_CGC_CTRL__LRBBM_MODE_MASK
+		| UVD_CGC_CTRL__WCB_MODE_MASK
+		| UVD_CGC_CTRL__VCPU_MODE_MASK
+		| UVD_CGC_CTRL__MMSCH_MODE_MASK);
+	WREG32_SOC15(VCN, inst, regUVD_CGC_CTRL, data);
+
+	data = RREG32_SOC15(VCN, inst, regUVD_SUVD_CGC_CTRL);
+	data |= (UVD_SUVD_CGC_CTRL__SRE_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SIT_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SMP_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SCM_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SDB_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SCLR_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__UVD_SC_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__ENT_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__IME_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SITE_MODE_MASK);
+	WREG32_SOC15(VCN, inst, regUVD_SUVD_CGC_CTRL, data);
+}
+
+static void vcn_v4_0_enable_ras(struct amdgpu_vcn_inst *vinst,
+				bool indirect)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int inst_idx = vinst->inst;
+	uint32_t tmp;
+
+	if (!amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__VCN))
+		return;
+
+	tmp = VCN_RAS_CNTL__VCPU_VCODEC_REARM_MASK |
+	      VCN_RAS_CNTL__VCPU_VCODEC_IH_EN_MASK |
+	      VCN_RAS_CNTL__VCPU_VCODEC_PMI_EN_MASK |
+	      VCN_RAS_CNTL__VCPU_VCODEC_STALL_EN_MASK;
+	WREG32_SOC15_DPG_MODE(inst_idx,
+			      SOC15_DPG_MODE_OFFSET(VCN, 0, regVCN_RAS_CNTL),
+			      tmp, 0, indirect);
+
+	tmp = UVD_SYS_INT_EN__RASCNTL_VCPU_VCODEC_EN_MASK;
+	WREG32_SOC15_DPG_MODE(inst_idx,
+			      SOC15_DPG_MODE_OFFSET(VCN, 0, regUVD_SYS_INT_EN),
+			      tmp, 0, indirect);
+}
+
+/**
+ * vcn_v4_0_start_dpg_mode - VCN start with dpg mode
+ *
+ * @vinst: VCN instance
+ * @indirect: indirectly write sram
+ *
+ * Start VCN block with dpg mode
+ */
+static int vcn_v4_0_start_dpg_mode(struct amdgpu_vcn_inst *vinst, bool indirect)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int inst_idx = vinst->inst;
+	struct amdgpu_vcn4_fw_shared *fw_shared = adev->vcn.inst[inst_idx].fw_shared.cpu_addr;
+	struct amdgpu_ring *ring;
+	uint32_t tmp;
+	int ret;
+
+	/* disable register anti-hang mechanism */
+	WREG32_P(SOC15_REG_OFFSET(VCN, inst_idx, regUVD_POWER_STATUS), 1,
+		~UVD_POWER_STATUS__UVD_POWER_STATUS_MASK);
+	/* enable dynamic power gating mode */
+	tmp = RREG32_SOC15(VCN, inst_idx, regUVD_POWER_STATUS);
+	tmp |= UVD_POWER_STATUS__UVD_PG_MODE_MASK;
+	tmp |= UVD_POWER_STATUS__UVD_PG_EN_MASK;
+	WREG32_SOC15(VCN, inst_idx, regUVD_POWER_STATUS, tmp);
+
+	if (indirect)
+		adev->vcn.inst[inst_idx].dpg_sram_curr_addr = (uint32_t *)adev->vcn.inst[inst_idx].dpg_sram_cpu_addr;
+
+	/* enable clock gating */
+	vcn_v4_0_disable_clock_gating_dpg_mode(vinst, 0, indirect);
+
+	/* enable VCPU clock */
+	tmp = (0xFF << UVD_VCPU_CNTL__PRB_TIMEOUT_VAL__SHIFT);
+	tmp |= UVD_VCPU_CNTL__CLK_EN_MASK | UVD_VCPU_CNTL__BLK_RST_MASK;
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_VCPU_CNTL), tmp, 0, indirect);
+
+	/* disable master interupt */
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_MASTINT_EN), 0, 0, indirect);
+
+	/* setup regUVD_LMI_CTRL */
+	tmp = (UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK |
+		UVD_LMI_CTRL__REQ_MODE_MASK |
+		UVD_LMI_CTRL__CRC_RESET_MASK |
+		UVD_LMI_CTRL__MASK_MC_URGENT_MASK |
+		UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK |
+		UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK |
+		(8 << UVD_LMI_CTRL__WRITE_CLEAN_TIMER__SHIFT) |
+		0x00100000L);
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_LMI_CTRL), tmp, 0, indirect);
+
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_MPC_CNTL),
+		0x2 << UVD_MPC_CNTL__REPLACEMENT_MODE__SHIFT, 0, indirect);
+
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_MPC_SET_MUXA0),
+		((0x1 << UVD_MPC_SET_MUXA0__VARA_1__SHIFT) |
+		 (0x2 << UVD_MPC_SET_MUXA0__VARA_2__SHIFT) |
+		 (0x3 << UVD_MPC_SET_MUXA0__VARA_3__SHIFT) |
+		 (0x4 << UVD_MPC_SET_MUXA0__VARA_4__SHIFT)), 0, indirect);
+
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_MPC_SET_MUXB0),
+		 ((0x1 << UVD_MPC_SET_MUXB0__VARB_1__SHIFT) |
+		 (0x2 << UVD_MPC_SET_MUXB0__VARB_2__SHIFT) |
+		 (0x3 << UVD_MPC_SET_MUXB0__VARB_3__SHIFT) |
+		 (0x4 << UVD_MPC_SET_MUXB0__VARB_4__SHIFT)), 0, indirect);
+
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_MPC_SET_MUX),
+		((0x0 << UVD_MPC_SET_MUX__SET_0__SHIFT) |
+		 (0x1 << UVD_MPC_SET_MUX__SET_1__SHIFT) |
+		 (0x2 << UVD_MPC_SET_MUX__SET_2__SHIFT)), 0, indirect);
+
+	vcn_v4_0_mc_resume_dpg_mode(vinst, indirect);
+
+	tmp = (0xFF << UVD_VCPU_CNTL__PRB_TIMEOUT_VAL__SHIFT);
+	tmp |= UVD_VCPU_CNTL__CLK_EN_MASK;
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_VCPU_CNTL), tmp, 0, indirect);
+
+	/* enable LMI MC and UMC channels */
+	tmp = 0x1f << UVD_LMI_CTRL2__RE_OFLD_MIF_WR_REQ_NUM__SHIFT;
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_LMI_CTRL2), tmp, 0, indirect);
+
+	vcn_v4_0_enable_ras(vinst, indirect);
+
+	/* enable master interrupt */
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_MASTINT_EN),
+		UVD_MASTINT_EN__VCPU_EN_MASK, 0, indirect);
+
+
+	if (indirect) {
+		ret = amdgpu_vcn_psp_update_sram(adev, inst_idx, 0);
+		if (ret) {
+			dev_err(adev->dev, "vcn sram load failed %d\n", ret);
+			return ret;
+		}
+	}
+
+	ring = &adev->vcn.inst[inst_idx].ring_enc[0];
+
+	WREG32_SOC15(VCN, inst_idx, regUVD_RB_BASE_LO, ring->gpu_addr);
+	WREG32_SOC15(VCN, inst_idx, regUVD_RB_BASE_HI, upper_32_bits(ring->gpu_addr));
+	WREG32_SOC15(VCN, inst_idx, regUVD_RB_SIZE, ring->ring_size / 4);
+
+	tmp = RREG32_SOC15(VCN, inst_idx, regVCN_RB_ENABLE);
+	tmp &= ~(VCN_RB_ENABLE__RB1_EN_MASK);
+	WREG32_SOC15(VCN, inst_idx, regVCN_RB_ENABLE, tmp);
+	fw_shared->sq.queue_mode |= FW_QUEUE_RING_RESET;
+	WREG32_SOC15(VCN, inst_idx, regUVD_RB_RPTR, 0);
+	WREG32_SOC15(VCN, inst_idx, regUVD_RB_WPTR, 0);
+
+	tmp = RREG32_SOC15(VCN, inst_idx, regUVD_RB_RPTR);
+	WREG32_SOC15(VCN, inst_idx, regUVD_RB_WPTR, tmp);
+	ring->wptr = RREG32_SOC15(VCN, inst_idx, regUVD_RB_WPTR);
+
+	tmp = RREG32_SOC15(VCN, inst_idx, regVCN_RB_ENABLE);
+	tmp |= VCN_RB_ENABLE__RB1_EN_MASK;
+	WREG32_SOC15(VCN, inst_idx, regVCN_RB_ENABLE, tmp);
+	fw_shared->sq.queue_mode &= ~(FW_QUEUE_RING_RESET | FW_QUEUE_DPG_HOLD_OFF);
+
+	WREG32_SOC15(VCN, inst_idx, regVCN_RB1_DB_CTRL,
+			ring->doorbell_index << VCN_RB1_DB_CTRL__OFFSET__SHIFT |
+			VCN_RB1_DB_CTRL__EN_MASK);
+
+	/* Keeping one read-back to ensure all register writes are done,
+	 * otherwise it may introduce race conditions.
+	 */
+	RREG32_SOC15(VCN, inst_idx, regUVD_STATUS);
+
+	return 0;
+}
+
+
+/**
+ * vcn_v4_0_start - VCN start
+ *
+ * @vinst: VCN instance
+ *
+ * Start VCN block
+ */
+static int vcn_v4_0_start(struct amdgpu_vcn_inst *vinst)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int i = vinst->inst;
+	struct amdgpu_vcn4_fw_shared *fw_shared;
+	struct amdgpu_ring *ring;
+	uint32_t tmp;
+	int j, k, r;
+
+	if (adev->vcn.harvest_config & (1 << i))
+		return 0;
+
+	if (adev->pm.dpm_enabled)
+		amdgpu_dpm_enable_vcn(adev, true, i);
+
+	fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr;
+
+	if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)
+		return vcn_v4_0_start_dpg_mode(vinst, adev->vcn.inst[i].indirect_sram);
+
+	/* disable VCN power gating */
+	vcn_v4_0_disable_static_power_gating(vinst);
+
+	/* set VCN status busy */
+	tmp = RREG32_SOC15(VCN, i, regUVD_STATUS) | UVD_STATUS__UVD_BUSY;
+	WREG32_SOC15(VCN, i, regUVD_STATUS, tmp);
+
+	/*SW clock gating */
+	vcn_v4_0_disable_clock_gating(vinst);
+
+	/* enable VCPU clock */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL),
+		 UVD_VCPU_CNTL__CLK_EN_MASK, ~UVD_VCPU_CNTL__CLK_EN_MASK);
+
+	/* disable master interrupt */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_MASTINT_EN), 0,
+		 ~UVD_MASTINT_EN__VCPU_EN_MASK);
+
+	/* enable LMI MC and UMC channels */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_LMI_CTRL2), 0,
+		 ~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK);
+
+	tmp = RREG32_SOC15(VCN, i, regUVD_SOFT_RESET);
+	tmp &= ~UVD_SOFT_RESET__LMI_SOFT_RESET_MASK;
+	tmp &= ~UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK;
+	WREG32_SOC15(VCN, i, regUVD_SOFT_RESET, tmp);
+
+	/* setup regUVD_LMI_CTRL */
+	tmp = RREG32_SOC15(VCN, i, regUVD_LMI_CTRL);
+	WREG32_SOC15(VCN, i, regUVD_LMI_CTRL, tmp |
+		     UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK |
+		     UVD_LMI_CTRL__MASK_MC_URGENT_MASK |
+		     UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK |
+		     UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK);
+
+	/* setup regUVD_MPC_CNTL */
+	tmp = RREG32_SOC15(VCN, i, regUVD_MPC_CNTL);
+	tmp &= ~UVD_MPC_CNTL__REPLACEMENT_MODE_MASK;
+	tmp |= 0x2 << UVD_MPC_CNTL__REPLACEMENT_MODE__SHIFT;
+	WREG32_SOC15(VCN, i, regUVD_MPC_CNTL, tmp);
+
+	/* setup UVD_MPC_SET_MUXA0 */
+	WREG32_SOC15(VCN, i, regUVD_MPC_SET_MUXA0,
+		     ((0x1 << UVD_MPC_SET_MUXA0__VARA_1__SHIFT) |
+		      (0x2 << UVD_MPC_SET_MUXA0__VARA_2__SHIFT) |
+		      (0x3 << UVD_MPC_SET_MUXA0__VARA_3__SHIFT) |
+		      (0x4 << UVD_MPC_SET_MUXA0__VARA_4__SHIFT)));
+
+	/* setup UVD_MPC_SET_MUXB0 */
+	WREG32_SOC15(VCN, i, regUVD_MPC_SET_MUXB0,
+		     ((0x1 << UVD_MPC_SET_MUXB0__VARB_1__SHIFT) |
+		      (0x2 << UVD_MPC_SET_MUXB0__VARB_2__SHIFT) |
+		      (0x3 << UVD_MPC_SET_MUXB0__VARB_3__SHIFT) |
+		      (0x4 << UVD_MPC_SET_MUXB0__VARB_4__SHIFT)));
+
+	/* setup UVD_MPC_SET_MUX */
+	WREG32_SOC15(VCN, i, regUVD_MPC_SET_MUX,
+		     ((0x0 << UVD_MPC_SET_MUX__SET_0__SHIFT) |
+		      (0x1 << UVD_MPC_SET_MUX__SET_1__SHIFT) |
+		      (0x2 << UVD_MPC_SET_MUX__SET_2__SHIFT)));
+
+	vcn_v4_0_mc_resume(vinst);
+
+	/* VCN global tiling registers */
+	WREG32_SOC15(VCN, i, regUVD_GFX10_ADDR_CONFIG,
+		     adev->gfx.config.gb_addr_config);
+
+	/* unblock VCPU register access */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_RB_ARB_CTRL), 0,
+		 ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK);
+
+	/* release VCPU reset to boot */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), 0,
+		 ~UVD_VCPU_CNTL__BLK_RST_MASK);
+
+	for (j = 0; j < 10; ++j) {
+		uint32_t status;
+
+		for (k = 0; k < 100; ++k) {
+			status = RREG32_SOC15(VCN, i, regUVD_STATUS);
+			if (status & 2)
+				break;
+			mdelay(10);
+			if (amdgpu_emu_mode == 1)
+				msleep(1);
+		}
+
+		if (amdgpu_emu_mode == 1) {
+			r = -1;
+			if (status & 2) {
+				r = 0;
+				break;
+			}
+		} else {
+			r = 0;
+			if (status & 2)
+				break;
+
+			dev_err(adev->dev, "VCN[%d] is not responding, trying to reset the VCPU!!!\n", i);
+			WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL),
+				 UVD_VCPU_CNTL__BLK_RST_MASK,
+				 ~UVD_VCPU_CNTL__BLK_RST_MASK);
+			mdelay(10);
+			WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), 0,
+				 ~UVD_VCPU_CNTL__BLK_RST_MASK);
+
+			mdelay(10);
+			r = -1;
+		}
+	}
+
+	if (r) {
+		dev_err(adev->dev, "VCN[%d] is not responding, giving up!!!\n", i);
+		return r;
+	}
+
+	/* enable master interrupt */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_MASTINT_EN),
+		 UVD_MASTINT_EN__VCPU_EN_MASK,
+		 ~UVD_MASTINT_EN__VCPU_EN_MASK);
+
+	/* clear the busy bit of VCN_STATUS */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_STATUS), 0,
+		 ~(2 << UVD_STATUS__VCPU_REPORT__SHIFT));
+
+	ring = &adev->vcn.inst[i].ring_enc[0];
+	WREG32_SOC15(VCN, i, regVCN_RB1_DB_CTRL,
+		     ring->doorbell_index << VCN_RB1_DB_CTRL__OFFSET__SHIFT |
+		     VCN_RB1_DB_CTRL__EN_MASK);
+
+	WREG32_SOC15(VCN, i, regUVD_RB_BASE_LO, ring->gpu_addr);
+	WREG32_SOC15(VCN, i, regUVD_RB_BASE_HI, upper_32_bits(ring->gpu_addr));
+	WREG32_SOC15(VCN, i, regUVD_RB_SIZE, ring->ring_size / 4);
+
+	tmp = RREG32_SOC15(VCN, i, regVCN_RB_ENABLE);
+	tmp &= ~(VCN_RB_ENABLE__RB1_EN_MASK);
+	WREG32_SOC15(VCN, i, regVCN_RB_ENABLE, tmp);
+	fw_shared->sq.queue_mode |= FW_QUEUE_RING_RESET;
+	WREG32_SOC15(VCN, i, regUVD_RB_RPTR, 0);
+	WREG32_SOC15(VCN, i, regUVD_RB_WPTR, 0);
+
+	tmp = RREG32_SOC15(VCN, i, regUVD_RB_RPTR);
+	WREG32_SOC15(VCN, i, regUVD_RB_WPTR, tmp);
+	ring->wptr = RREG32_SOC15(VCN, i, regUVD_RB_WPTR);
+
+	tmp = RREG32_SOC15(VCN, i, regVCN_RB_ENABLE);
+	tmp |= VCN_RB_ENABLE__RB1_EN_MASK;
+	WREG32_SOC15(VCN, i, regVCN_RB_ENABLE, tmp);
+	fw_shared->sq.queue_mode &= ~(FW_QUEUE_RING_RESET | FW_QUEUE_DPG_HOLD_OFF);
+
+	/* Keeping one read-back to ensure all register writes are done,
+	 * otherwise it may introduce race conditions.
+	 */
+	RREG32_SOC15(VCN, i, regUVD_STATUS);
+
+	return 0;
+}
+
+static int vcn_v4_0_init_ring_metadata(struct amdgpu_device *adev, uint32_t vcn_inst, struct amdgpu_ring *ring_enc)
+{
+	struct amdgpu_vcn_rb_metadata *rb_metadata = NULL;
+	uint8_t *rb_ptr = (uint8_t *)ring_enc->ring;
+
+	rb_ptr += ring_enc->ring_size;
+	rb_metadata = (struct amdgpu_vcn_rb_metadata *)rb_ptr;
+
+	memset(rb_metadata, 0, sizeof(struct amdgpu_vcn_rb_metadata));
+	rb_metadata->size = sizeof(struct amdgpu_vcn_rb_metadata);
+	rb_metadata->present_flag_0 |= cpu_to_le32(AMDGPU_VCN_VF_RB_SETUP_FLAG);
+	rb_metadata->present_flag_0 |= cpu_to_le32(AMDGPU_VCN_VF_RB_DECOUPLE_FLAG);
+	rb_metadata->version = 1;
+	rb_metadata->ring_id = vcn_inst & 0xFF;
+
+	return 0;
+}
+
+static int vcn_v4_0_start_sriov(struct amdgpu_device *adev)
+{
+	int i;
+	struct amdgpu_ring *ring_enc;
+	uint64_t cache_addr;
+	uint64_t rb_enc_addr;
+	uint64_t ctx_addr;
+	uint32_t param, resp, expected;
+	uint32_t offset, cache_size;
+	uint32_t tmp, timeout;
+
+	struct amdgpu_mm_table *table = &adev->virt.mm_table;
+	uint32_t *table_loc;
+	uint32_t table_size;
+	uint32_t size, size_dw;
+	uint32_t init_status;
+	uint32_t enabled_vcn;
+
+	struct mmsch_v4_0_cmd_direct_write
+		direct_wt = { {0} };
+	struct mmsch_v4_0_cmd_direct_read_modify_write
+		direct_rd_mod_wt = { {0} };
+	struct mmsch_v4_0_cmd_end end = { {0} };
+	struct mmsch_v4_0_init_header header;
+
+	struct amdgpu_vcn4_fw_shared *fw_shared;
+	struct amdgpu_fw_shared_rb_setup *rb_setup;
+
+	direct_wt.cmd_header.command_type =
+		MMSCH_COMMAND__DIRECT_REG_WRITE;
+	direct_rd_mod_wt.cmd_header.command_type =
+		MMSCH_COMMAND__DIRECT_REG_READ_MODIFY_WRITE;
+	end.cmd_header.command_type =
+		MMSCH_COMMAND__END;
+
+	header.version = MMSCH_VERSION;
+	header.total_size = sizeof(struct mmsch_v4_0_init_header) >> 2;
+	for (i = 0; i < MMSCH_V4_0_VCN_INSTANCES; i++) {
+		header.inst[i].init_status = 0;
+		header.inst[i].table_offset = 0;
+		header.inst[i].table_size = 0;
+	}
+
+	table_loc = (uint32_t *)table->cpu_addr;
+	table_loc += header.total_size;
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+		if (adev->vcn.harvest_config & (1 << i))
+			continue;
+
+		// Must re/init fw_shared at beginning
+		vcn_v4_0_fw_shared_init(adev, i);
+
+		table_size = 0;
+
+		MMSCH_V4_0_INSERT_DIRECT_RD_MOD_WT(SOC15_REG_OFFSET(VCN, i,
+			regUVD_STATUS),
+			~UVD_STATUS__UVD_BUSY, UVD_STATUS__UVD_BUSY);
+
+		cache_size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.inst[i].fw->size + 4);
+
+		if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
+			MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, i,
+				regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW),
+				adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + i].tmr_mc_addr_lo);
+			MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, i,
+				regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH),
+				adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + i].tmr_mc_addr_hi);
+			offset = 0;
+			MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, i,
+				regUVD_VCPU_CACHE_OFFSET0),
+				0);
+		} else {
+			MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, i,
+				regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW),
+				lower_32_bits(adev->vcn.inst[i].gpu_addr));
+			MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, i,
+				regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH),
+				upper_32_bits(adev->vcn.inst[i].gpu_addr));
+			offset = cache_size;
+			MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, i,
+				regUVD_VCPU_CACHE_OFFSET0),
+				AMDGPU_UVD_FIRMWARE_OFFSET >> 3);
+		}
+
+		MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, i,
+			regUVD_VCPU_CACHE_SIZE0),
+			cache_size);
+
+		cache_addr = adev->vcn.inst[i].gpu_addr + offset;
+		MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, i,
+			regUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW),
+			lower_32_bits(cache_addr));
+		MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, i,
+			regUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH),
+			upper_32_bits(cache_addr));
+		MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, i,
+			regUVD_VCPU_CACHE_OFFSET1),
+			0);
+		MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, i,
+			regUVD_VCPU_CACHE_SIZE1),
+			AMDGPU_VCN_STACK_SIZE);
+
+		cache_addr = adev->vcn.inst[i].gpu_addr + offset +
+			AMDGPU_VCN_STACK_SIZE;
+		MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, i,
+			regUVD_LMI_VCPU_CACHE2_64BIT_BAR_LOW),
+			lower_32_bits(cache_addr));
+		MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, i,
+			regUVD_LMI_VCPU_CACHE2_64BIT_BAR_HIGH),
+			upper_32_bits(cache_addr));
+		MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, i,
+			regUVD_VCPU_CACHE_OFFSET2),
+			0);
+		MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, i,
+			regUVD_VCPU_CACHE_SIZE2),
+			AMDGPU_VCN_CONTEXT_SIZE);
+
+		fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr;
+		rb_setup = &fw_shared->rb_setup;
+
+		ring_enc = &adev->vcn.inst[i].ring_enc[0];
+		ring_enc->wptr = 0;
+		rb_enc_addr = ring_enc->gpu_addr;
+
+		rb_setup->is_rb_enabled_flags |= RB_ENABLED;
+		fw_shared->present_flag_0 |= cpu_to_le32(AMDGPU_VCN_VF_RB_SETUP_FLAG);
+
+		if (amdgpu_sriov_is_vcn_rb_decouple(adev)) {
+			vcn_v4_0_init_ring_metadata(adev, i, ring_enc);
+
+			memset((void *)&rb_setup->rb_info, 0, sizeof(struct amdgpu_vcn_rb_setup_info) * MAX_NUM_VCN_RB_SETUP);
+			if (!(adev->vcn.harvest_config & (1 << 0))) {
+				rb_setup->rb_info[0].rb_addr_lo = lower_32_bits(adev->vcn.inst[0].ring_enc[0].gpu_addr);
+				rb_setup->rb_info[0].rb_addr_hi = upper_32_bits(adev->vcn.inst[0].ring_enc[0].gpu_addr);
+				rb_setup->rb_info[0].rb_size = adev->vcn.inst[0].ring_enc[0].ring_size / 4;
+			}
+			if (!(adev->vcn.harvest_config & (1 << 1))) {
+				rb_setup->rb_info[2].rb_addr_lo = lower_32_bits(adev->vcn.inst[1].ring_enc[0].gpu_addr);
+				rb_setup->rb_info[2].rb_addr_hi = upper_32_bits(adev->vcn.inst[1].ring_enc[0].gpu_addr);
+				rb_setup->rb_info[2].rb_size = adev->vcn.inst[1].ring_enc[0].ring_size / 4;
+			}
+			fw_shared->decouple.is_enabled = 1;
+			fw_shared->present_flag_0 |= cpu_to_le32(AMDGPU_VCN_VF_RB_DECOUPLE_FLAG);
+		} else {
+			rb_setup->rb_addr_lo = lower_32_bits(rb_enc_addr);
+			rb_setup->rb_addr_hi = upper_32_bits(rb_enc_addr);
+			rb_setup->rb_size = ring_enc->ring_size / 4;
+		}
+
+		MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, i,
+			regUVD_LMI_VCPU_NC0_64BIT_BAR_LOW),
+			lower_32_bits(adev->vcn.inst[i].fw_shared.gpu_addr));
+		MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, i,
+			regUVD_LMI_VCPU_NC0_64BIT_BAR_HIGH),
+			upper_32_bits(adev->vcn.inst[i].fw_shared.gpu_addr));
+		MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, i,
+			regUVD_VCPU_NONCACHE_SIZE0),
+			AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_vcn4_fw_shared)));
+
+		/* add end packet */
+		MMSCH_V4_0_INSERT_END();
+
+		/* refine header */
+		header.inst[i].init_status = 0;
+		header.inst[i].table_offset = header.total_size;
+		header.inst[i].table_size = table_size;
+		header.total_size += table_size;
+	}
+
+	/* Update init table header in memory */
+	size = sizeof(struct mmsch_v4_0_init_header);
+	table_loc = (uint32_t *)table->cpu_addr;
+	memcpy((void *)table_loc, &header, size);
+
+	/* message MMSCH (in VCN[0]) to initialize this client
+	 * 1, write to mmsch_vf_ctx_addr_lo/hi register with GPU mc addr
+	 * of memory descriptor location
+	 */
+	ctx_addr = table->gpu_addr;
+	WREG32_SOC15(VCN, 0, regMMSCH_VF_CTX_ADDR_LO, lower_32_bits(ctx_addr));
+	WREG32_SOC15(VCN, 0, regMMSCH_VF_CTX_ADDR_HI, upper_32_bits(ctx_addr));
+
+	/* 2, update vmid of descriptor */
+	tmp = RREG32_SOC15(VCN, 0, regMMSCH_VF_VMID);
+	tmp &= ~MMSCH_VF_VMID__VF_CTX_VMID_MASK;
+	/* use domain0 for MM scheduler */
+	tmp |= (0 << MMSCH_VF_VMID__VF_CTX_VMID__SHIFT);
+	WREG32_SOC15(VCN, 0, regMMSCH_VF_VMID, tmp);
+
+	/* 3, notify mmsch about the size of this descriptor */
+	size = header.total_size;
+	WREG32_SOC15(VCN, 0, regMMSCH_VF_CTX_SIZE, size);
+
+	/* 4, set resp to zero */
+	WREG32_SOC15(VCN, 0, regMMSCH_VF_MAILBOX_RESP, 0);
+
+	/* 5, kick off the initialization and wait until
+	 * MMSCH_VF_MAILBOX_RESP becomes non-zero
+	 */
+	param = 0x00000001;
+	WREG32_SOC15(VCN, 0, regMMSCH_VF_MAILBOX_HOST, param);
+	tmp = 0;
+	timeout = 1000;
+	resp = 0;
+	expected = MMSCH_VF_MAILBOX_RESP__OK;
+	while (resp != expected) {
+		resp = RREG32_SOC15(VCN, 0, regMMSCH_VF_MAILBOX_RESP);
+		if (resp != 0)
+			break;
+
+		udelay(10);
+		tmp = tmp + 10;
+		if (tmp >= timeout) {
+			DRM_ERROR("failed to init MMSCH. TIME-OUT after %d usec"\
+				" waiting for regMMSCH_VF_MAILBOX_RESP "\
+				"(expected=0x%08x, readback=0x%08x)\n",
+				tmp, expected, resp);
+			return -EBUSY;
+		}
+	}
+	enabled_vcn = amdgpu_vcn_is_disabled_vcn(adev, VCN_DECODE_RING, 0) ? 1 : 0;
+	init_status = ((struct mmsch_v4_0_init_header *)(table_loc))->inst[enabled_vcn].init_status;
+	if (resp != expected && resp != MMSCH_VF_MAILBOX_RESP__INCOMPLETE
+	&& init_status != MMSCH_VF_ENGINE_STATUS__PASS)
+		DRM_ERROR("MMSCH init status is incorrect! readback=0x%08x, header init "\
+			"status for VCN%x: 0x%x\n", resp, enabled_vcn, init_status);
+
+	return 0;
+}
+
+/**
+ * vcn_v4_0_stop_dpg_mode - VCN stop with dpg mode
+ *
+ * @vinst: VCN instance
+ *
+ * Stop VCN block with dpg mode
+ */
+static void vcn_v4_0_stop_dpg_mode(struct amdgpu_vcn_inst *vinst)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int inst_idx = vinst->inst;
+	struct dpg_pause_state state = {.fw_based = VCN_DPG_STATE__UNPAUSE};
+	uint32_t tmp;
+
+	vcn_v4_0_pause_dpg_mode(vinst, &state);
+	/* Wait for power status to be 1 */
+	SOC15_WAIT_ON_RREG(VCN, inst_idx, regUVD_POWER_STATUS, 1,
+		UVD_POWER_STATUS__UVD_POWER_STATUS_MASK);
+
+	/* wait for read ptr to be equal to write ptr */
+	tmp = RREG32_SOC15(VCN, inst_idx, regUVD_RB_WPTR);
+	SOC15_WAIT_ON_RREG(VCN, inst_idx, regUVD_RB_RPTR, tmp, 0xFFFFFFFF);
+
+	SOC15_WAIT_ON_RREG(VCN, inst_idx, regUVD_POWER_STATUS, 1,
+		UVD_POWER_STATUS__UVD_POWER_STATUS_MASK);
+
+	/* disable dynamic power gating mode */
+	WREG32_P(SOC15_REG_OFFSET(VCN, inst_idx, regUVD_POWER_STATUS), 0,
+		~UVD_POWER_STATUS__UVD_PG_MODE_MASK);
+
+	/* Keeping one read-back to ensure all register writes are done,
+	 * otherwise it may introduce race conditions.
+	 */
+	RREG32_SOC15(VCN, inst_idx, regUVD_STATUS);
+}
+
+/**
+ * vcn_v4_0_stop - VCN stop
+ *
+ * @vinst: VCN instance
+ *
+ * Stop VCN block
+ */
+static int vcn_v4_0_stop(struct amdgpu_vcn_inst *vinst)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int i = vinst->inst;
+	struct amdgpu_vcn4_fw_shared *fw_shared;
+	uint32_t tmp;
+	int r = 0;
+
+	if (adev->vcn.harvest_config & (1 << i))
+		return 0;
+
+	fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr;
+	fw_shared->sq.queue_mode |= FW_QUEUE_DPG_HOLD_OFF;
+
+	if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) {
+		vcn_v4_0_stop_dpg_mode(vinst);
+		goto done;
+	}
+
+	/* wait for vcn idle */
+	r = SOC15_WAIT_ON_RREG(VCN, i, regUVD_STATUS, UVD_STATUS__IDLE, 0x7);
+	if (r)
+		goto done;
+
+	tmp = UVD_LMI_STATUS__VCPU_LMI_WRITE_CLEAN_MASK |
+		UVD_LMI_STATUS__READ_CLEAN_MASK |
+		UVD_LMI_STATUS__WRITE_CLEAN_MASK |
+		UVD_LMI_STATUS__WRITE_CLEAN_RAW_MASK;
+	r = SOC15_WAIT_ON_RREG(VCN, i, regUVD_LMI_STATUS, tmp, tmp);
+	if (r)
+		goto done;
+
+	/* disable LMI UMC channel */
+	tmp = RREG32_SOC15(VCN, i, regUVD_LMI_CTRL2);
+	tmp |= UVD_LMI_CTRL2__STALL_ARB_UMC_MASK;
+	WREG32_SOC15(VCN, i, regUVD_LMI_CTRL2, tmp);
+	tmp = UVD_LMI_STATUS__UMC_READ_CLEAN_RAW_MASK |
+		UVD_LMI_STATUS__UMC_WRITE_CLEAN_RAW_MASK;
+	r = SOC15_WAIT_ON_RREG(VCN, i, regUVD_LMI_STATUS, tmp, tmp);
+	if (r)
+		goto done;
+
+	/* block VCPU register access */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_RB_ARB_CTRL),
+		 UVD_RB_ARB_CTRL__VCPU_DIS_MASK,
+		 ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK);
+
+	/* reset VCPU */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL),
+		 UVD_VCPU_CNTL__BLK_RST_MASK,
+		 ~UVD_VCPU_CNTL__BLK_RST_MASK);
+
+	/* disable VCPU clock */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), 0,
+		 ~(UVD_VCPU_CNTL__CLK_EN_MASK));
+
+	/* apply soft reset */
+	tmp = RREG32_SOC15(VCN, i, regUVD_SOFT_RESET);
+	tmp |= UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK;
+	WREG32_SOC15(VCN, i, regUVD_SOFT_RESET, tmp);
+	tmp = RREG32_SOC15(VCN, i, regUVD_SOFT_RESET);
+	tmp |= UVD_SOFT_RESET__LMI_SOFT_RESET_MASK;
+	WREG32_SOC15(VCN, i, regUVD_SOFT_RESET, tmp);
+
+	/* clear status */
+	WREG32_SOC15(VCN, i, regUVD_STATUS, 0);
+
+	/* apply HW clock gating */
+	vcn_v4_0_enable_clock_gating(vinst);
+
+	/* enable VCN power gating */
+	vcn_v4_0_enable_static_power_gating(vinst);
+
+	/* Keeping one read-back to ensure all register writes are done,
+	 * otherwise it may introduce race conditions.
+	 */
+	RREG32_SOC15(VCN, i, regUVD_STATUS);
+
+done:
+	if (adev->pm.dpm_enabled)
+		amdgpu_dpm_enable_vcn(adev, false, i);
+
+	return 0;
+}
+
+/**
+ * vcn_v4_0_pause_dpg_mode - VCN pause with dpg mode
+ *
+ * @vinst: VCN instance
+ * @new_state: pause state
+ *
+ * Pause dpg mode for VCN block
+ */
+static int vcn_v4_0_pause_dpg_mode(struct amdgpu_vcn_inst *vinst,
+				   struct dpg_pause_state *new_state)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int inst_idx = vinst->inst;
+	uint32_t reg_data = 0;
+	int ret_code;
+
+	/* pause/unpause if state is changed */
+	if (adev->vcn.inst[inst_idx].pause_state.fw_based != new_state->fw_based) {
+		DRM_DEV_DEBUG(adev->dev, "dpg pause state changed %d -> %d",
+			adev->vcn.inst[inst_idx].pause_state.fw_based,	new_state->fw_based);
+		reg_data = RREG32_SOC15(VCN, inst_idx, regUVD_DPG_PAUSE) &
+			(~UVD_DPG_PAUSE__NJ_PAUSE_DPG_ACK_MASK);
+
+		if (new_state->fw_based == VCN_DPG_STATE__PAUSE) {
+			ret_code = SOC15_WAIT_ON_RREG(VCN, inst_idx, regUVD_POWER_STATUS, 0x1,
+				UVD_POWER_STATUS__UVD_POWER_STATUS_MASK);
+
+			if (!ret_code) {
+				/* pause DPG */
+				reg_data |= UVD_DPG_PAUSE__NJ_PAUSE_DPG_REQ_MASK;
+				WREG32_SOC15(VCN, inst_idx, regUVD_DPG_PAUSE, reg_data);
+
+				/* wait for ACK */
+				SOC15_WAIT_ON_RREG(VCN, inst_idx, regUVD_DPG_PAUSE,
+					UVD_DPG_PAUSE__NJ_PAUSE_DPG_ACK_MASK,
+					UVD_DPG_PAUSE__NJ_PAUSE_DPG_ACK_MASK);
+
+				SOC15_WAIT_ON_RREG(VCN, inst_idx, regUVD_POWER_STATUS,
+					UVD_PGFSM_CONFIG__UVDM_UVDU_PWR_ON, UVD_POWER_STATUS__UVD_POWER_STATUS_MASK);
+			}
+		} else {
+			/* unpause dpg, no need to wait */
+			reg_data &= ~UVD_DPG_PAUSE__NJ_PAUSE_DPG_REQ_MASK;
+			WREG32_SOC15(VCN, inst_idx, regUVD_DPG_PAUSE, reg_data);
+		}
+		adev->vcn.inst[inst_idx].pause_state.fw_based = new_state->fw_based;
+	}
+
+	return 0;
+}
+
+/**
+ * vcn_v4_0_unified_ring_get_rptr - get unified read pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Returns the current hardware unified read pointer
+ */
+static uint64_t vcn_v4_0_unified_ring_get_rptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	if (ring != &adev->vcn.inst[ring->me].ring_enc[0])
+		DRM_ERROR("wrong ring id is identified in %s", __func__);
+
+	return RREG32_SOC15(VCN, ring->me, regUVD_RB_RPTR);
+}
+
+/**
+ * vcn_v4_0_unified_ring_get_wptr - get unified write pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Returns the current hardware unified write pointer
+ */
+static uint64_t vcn_v4_0_unified_ring_get_wptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	if (ring != &adev->vcn.inst[ring->me].ring_enc[0])
+		DRM_ERROR("wrong ring id is identified in %s", __func__);
+
+	if (ring->use_doorbell)
+		return *ring->wptr_cpu_addr;
+	else
+		return RREG32_SOC15(VCN, ring->me, regUVD_RB_WPTR);
+}
+
+/**
+ * vcn_v4_0_unified_ring_set_wptr - set enc write pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Commits the enc write pointer to the hardware
+ */
+static void vcn_v4_0_unified_ring_set_wptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	if (ring != &adev->vcn.inst[ring->me].ring_enc[0])
+		DRM_ERROR("wrong ring id is identified in %s", __func__);
+
+	if (ring->use_doorbell) {
+		*ring->wptr_cpu_addr = lower_32_bits(ring->wptr);
+		WDOORBELL32(ring->doorbell_index, lower_32_bits(ring->wptr));
+	} else {
+		WREG32_SOC15(VCN, ring->me, regUVD_RB_WPTR, lower_32_bits(ring->wptr));
+	}
+}
+
+static int vcn_v4_0_limit_sched(struct amdgpu_cs_parser *p,
+				struct amdgpu_job *job)
+{
+	struct drm_gpu_scheduler **scheds;
+	struct dma_fence *fence;
+
+	/* if VCN0 is harvested, we can't support AV1 */
+	if (p->adev->vcn.harvest_config & AMDGPU_VCN_HARVEST_VCN0)
+		return -EINVAL;
+
+	/* wait for all jobs to finish before switching to instance 0 */
+	fence = amdgpu_ctx_get_fence(p->ctx, job->base.entity, ~0ull);
+	if (fence) {
+		dma_fence_wait(fence, false);
+		dma_fence_put(fence);
+	}
+
+	scheds = p->adev->gpu_sched[AMDGPU_HW_IP_VCN_ENC]
+		[AMDGPU_RING_PRIO_0].sched;
+	drm_sched_entity_modify_sched(job->base.entity, scheds, 1);
+	return 0;
+}
+
+static int vcn_v4_0_dec_msg(struct amdgpu_cs_parser *p, struct amdgpu_job *job,
+			    uint64_t addr)
+{
+	struct ttm_operation_ctx ctx = { false, false };
+	struct amdgpu_bo_va_mapping *map;
+	uint32_t *msg, num_buffers;
+	struct amdgpu_bo *bo;
+	uint64_t start, end;
+	unsigned int i;
+	void *ptr;
+	int r;
+
+	addr &= AMDGPU_GMC_HOLE_MASK;
+	r = amdgpu_cs_find_mapping(p, addr, &bo, &map);
+	if (r) {
+		DRM_ERROR("Can't find BO for addr 0x%08llx\n", addr);
+		return r;
+	}
+
+	start = map->start * AMDGPU_GPU_PAGE_SIZE;
+	end = (map->last + 1) * AMDGPU_GPU_PAGE_SIZE;
+	if (addr & 0x7) {
+		DRM_ERROR("VCN messages must be 8 byte aligned!\n");
+		return -EINVAL;
+	}
+
+	bo->flags |= AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED;
+	amdgpu_bo_placement_from_domain(bo, bo->allowed_domains);
+	r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
+	if (r) {
+		DRM_ERROR("Failed validating the VCN message BO (%d)!\n", r);
+		return r;
+	}
+
+	r = amdgpu_bo_kmap(bo, &ptr);
+	if (r) {
+		DRM_ERROR("Failed mapping the VCN message (%d)!\n", r);
+		return r;
+	}
+
+	msg = ptr + addr - start;
+
+	/* Check length */
+	if (msg[1] > end - addr) {
+		r = -EINVAL;
+		goto out;
+	}
+
+	if (msg[3] != RDECODE_MSG_CREATE)
+		goto out;
+
+	num_buffers = msg[2];
+	for (i = 0, msg = &msg[6]; i < num_buffers; ++i, msg += 4) {
+		uint32_t offset, size, *create;
+
+		if (msg[0] != RDECODE_MESSAGE_CREATE)
+			continue;
+
+		offset = msg[1];
+		size = msg[2];
+
+		if (offset + size > end) {
+			r = -EINVAL;
+			goto out;
+		}
+
+		create = ptr + addr + offset - start;
+
+		/* H264, HEVC and VP9 can run on any instance */
+		if (create[0] == 0x7 || create[0] == 0x10 || create[0] == 0x11)
+			continue;
+
+		r = vcn_v4_0_limit_sched(p, job);
+		if (r)
+			goto out;
+	}
+
+out:
+	amdgpu_bo_kunmap(bo);
+	return r;
+}
+
+#define RADEON_VCN_ENGINE_TYPE_ENCODE			(0x00000002)
+#define RADEON_VCN_ENGINE_TYPE_DECODE			(0x00000003)
+#define RADEON_VCN_ENGINE_INFO				(0x30000001)
+#define RENCODE_ENCODE_STANDARD_AV1			2
+#define RENCODE_IB_PARAM_SESSION_INIT			0x00000003
+
+/* return the offset in ib if id is found, -1 otherwise */
+static int vcn_v4_0_enc_find_ib_param(struct amdgpu_ib *ib, uint32_t id, int start)
+{
+	int i;
+
+	for (i = start; i < ib->length_dw && ib->ptr[i] >= 8; i += ib->ptr[i] / 4) {
+		if (ib->ptr[i + 1] == id)
+			return i;
+	}
+	return -1;
+}
+
+static int vcn_v4_0_ring_patch_cs_in_place(struct amdgpu_cs_parser *p,
+					   struct amdgpu_job *job,
+					   struct amdgpu_ib *ib)
+{
+	struct amdgpu_ring *ring = amdgpu_job_ring(job);
+	struct amdgpu_vcn_decode_buffer *decode_buffer;
+	uint64_t addr;
+	uint32_t val;
+	int idx = 0, sidx;
+
+	/* The first instance can decode anything */
+	if (!ring->me)
+		return 0;
+
+	while ((idx = vcn_v4_0_enc_find_ib_param(ib, RADEON_VCN_ENGINE_INFO, idx)) >= 0) {
+		val = amdgpu_ib_get_value(ib, idx + 2); /* RADEON_VCN_ENGINE_TYPE */
+		if (val == RADEON_VCN_ENGINE_TYPE_DECODE) {
+			decode_buffer = (struct amdgpu_vcn_decode_buffer *)&ib->ptr[idx + 6];
+
+			if (!(decode_buffer->valid_buf_flag & 0x1))
+				return 0;
+
+			addr = ((u64)decode_buffer->msg_buffer_address_hi) << 32 |
+				decode_buffer->msg_buffer_address_lo;
+			return vcn_v4_0_dec_msg(p, job, addr);
+		} else if (val == RADEON_VCN_ENGINE_TYPE_ENCODE) {
+			sidx = vcn_v4_0_enc_find_ib_param(ib, RENCODE_IB_PARAM_SESSION_INIT, idx);
+			if (sidx >= 0 && ib->ptr[sidx + 2] == RENCODE_ENCODE_STANDARD_AV1)
+				return vcn_v4_0_limit_sched(p, job);
+		}
+		idx += ib->ptr[idx] / 4;
+	}
+	return 0;
+}
+
+static int vcn_v4_0_ring_reset(struct amdgpu_ring *ring,
+			       unsigned int vmid,
+			       struct amdgpu_fence *timedout_fence)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct amdgpu_vcn_inst *vinst = &adev->vcn.inst[ring->me];
+	int r;
+
+	amdgpu_ring_reset_helper_begin(ring, timedout_fence);
+	r = vcn_v4_0_stop(vinst);
+	if (r)
+		return r;
+	r = vcn_v4_0_start(vinst);
+	if (r)
+		return r;
+	return amdgpu_ring_reset_helper_end(ring, timedout_fence);
+}
+
+static struct amdgpu_ring_funcs vcn_v4_0_unified_ring_vm_funcs = {
+	.type = AMDGPU_RING_TYPE_VCN_ENC,
+	.align_mask = 0x3f,
+	.nop = VCN_ENC_CMD_NO_OP,
+	.extra_bytes = sizeof(struct amdgpu_vcn_rb_metadata),
+	.get_rptr = vcn_v4_0_unified_ring_get_rptr,
+	.get_wptr = vcn_v4_0_unified_ring_get_wptr,
+	.set_wptr = vcn_v4_0_unified_ring_set_wptr,
+	.patch_cs_in_place = vcn_v4_0_ring_patch_cs_in_place,
+	.emit_frame_size =
+		SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 +
+		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 4 +
+		4 + /* vcn_v2_0_enc_ring_emit_vm_flush */
+		5 + 5 + /* vcn_v2_0_enc_ring_emit_fence x2 vm fence */
+		1, /* vcn_v2_0_enc_ring_insert_end */
+	.emit_ib_size = 5, /* vcn_v2_0_enc_ring_emit_ib */
+	.emit_ib = vcn_v2_0_enc_ring_emit_ib,
+	.emit_fence = vcn_v2_0_enc_ring_emit_fence,
+	.emit_vm_flush = vcn_v2_0_enc_ring_emit_vm_flush,
+	.test_ring = amdgpu_vcn_enc_ring_test_ring,
+	.test_ib = amdgpu_vcn_unified_ring_test_ib,
+	.insert_nop = amdgpu_ring_insert_nop,
+	.insert_end = vcn_v2_0_enc_ring_insert_end,
+	.pad_ib = amdgpu_ring_generic_pad_ib,
+	.begin_use = amdgpu_vcn_ring_begin_use,
+	.end_use = amdgpu_vcn_ring_end_use,
+	.emit_wreg = vcn_v2_0_enc_ring_emit_wreg,
+	.emit_reg_wait = vcn_v2_0_enc_ring_emit_reg_wait,
+	.emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper,
+	.reset = vcn_v4_0_ring_reset,
+};
+
+/**
+ * vcn_v4_0_set_unified_ring_funcs - set unified ring functions
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Set unified ring functions
+ */
+static void vcn_v4_0_set_unified_ring_funcs(struct amdgpu_device *adev)
+{
+	int i;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		if (adev->vcn.harvest_config & (1 << i))
+			continue;
+
+		if (amdgpu_ip_version(adev, VCN_HWIP, 0) == IP_VERSION(4, 0, 2))
+			vcn_v4_0_unified_ring_vm_funcs.secure_submission_supported = true;
+
+		adev->vcn.inst[i].ring_enc[0].funcs =
+		       (const struct amdgpu_ring_funcs *)&vcn_v4_0_unified_ring_vm_funcs;
+		adev->vcn.inst[i].ring_enc[0].me = i;
+	}
+}
+
+/**
+ * vcn_v4_0_is_idle - check VCN block is idle
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block structure
+ *
+ * Check whether VCN block is idle
+ */
+static bool vcn_v4_0_is_idle(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, ret = 1;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		if (adev->vcn.harvest_config & (1 << i))
+			continue;
+
+		ret &= (RREG32_SOC15(VCN, i, regUVD_STATUS) == UVD_STATUS__IDLE);
+	}
+
+	return ret;
+}
+
+/**
+ * vcn_v4_0_wait_for_idle - wait for VCN block idle
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Wait for VCN block idle
+ */
+static int vcn_v4_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, ret = 0;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		if (adev->vcn.harvest_config & (1 << i))
+			continue;
+
+		ret = SOC15_WAIT_ON_RREG(VCN, i, regUVD_STATUS, UVD_STATUS__IDLE,
+			UVD_STATUS__IDLE);
+		if (ret)
+			return ret;
+	}
+
+	return ret;
+}
+
+/**
+ * vcn_v4_0_set_clockgating_state - set VCN block clockgating state
+ *
+ * @ip_block: amdgpu_ip_block pointer
+ * @state: clock gating state
+ *
+ * Set VCN block clockgating state
+ */
+static int vcn_v4_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
+					  enum amd_clockgating_state state)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	bool enable = state == AMD_CG_STATE_GATE;
+	int i;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		struct amdgpu_vcn_inst *vinst = &adev->vcn.inst[i];
+
+		if (adev->vcn.harvest_config & (1 << i))
+			continue;
+
+		if (enable) {
+			if (RREG32_SOC15(VCN, i, regUVD_STATUS) != UVD_STATUS__IDLE)
+				return -EBUSY;
+			vcn_v4_0_enable_clock_gating(vinst);
+		} else {
+			vcn_v4_0_disable_clock_gating(vinst);
+		}
+	}
+
+	return 0;
+}
+
+static int vcn_v4_0_set_pg_state(struct amdgpu_vcn_inst *vinst,
+				 enum amd_powergating_state state)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int ret = 0;
+
+	/* for SRIOV, guest should not control VCN Power-gating
+	 * MMSCH FW should control Power-gating and clock-gating
+	 * guest should avoid touching CGC and PG
+	 */
+	if (amdgpu_sriov_vf(adev)) {
+		vinst->cur_state = AMD_PG_STATE_UNGATE;
+		return 0;
+	}
+
+	if (state == vinst->cur_state)
+		return 0;
+
+	if (state == AMD_PG_STATE_GATE)
+		ret = vcn_v4_0_stop(vinst);
+	else
+		ret = vcn_v4_0_start(vinst);
+
+	if (!ret)
+		vinst->cur_state = state;
+
+	return ret;
+}
+
+/**
+ * vcn_v4_0_set_ras_interrupt_state - set VCN block RAS interrupt state
+ *
+ * @adev: amdgpu_device pointer
+ * @source: interrupt sources
+ * @type: interrupt types
+ * @state: interrupt states
+ *
+ * Set VCN block RAS interrupt state
+ */
+static int vcn_v4_0_set_ras_interrupt_state(struct amdgpu_device *adev,
+	struct amdgpu_irq_src *source,
+	unsigned int type,
+	enum amdgpu_interrupt_state state)
+{
+	return 0;
+}
+
+/**
+ * vcn_v4_0_process_interrupt - process VCN block interrupt
+ *
+ * @adev: amdgpu_device pointer
+ * @source: interrupt sources
+ * @entry: interrupt entry from clients and sources
+ *
+ * Process VCN block interrupt
+ */
+static int vcn_v4_0_process_interrupt(struct amdgpu_device *adev, struct amdgpu_irq_src *source,
+      struct amdgpu_iv_entry *entry)
+{
+	uint32_t ip_instance;
+
+	if (amdgpu_sriov_is_vcn_rb_decouple(adev)) {
+		ip_instance = entry->ring_id;
+	} else {
+		switch (entry->client_id) {
+		case SOC15_IH_CLIENTID_VCN:
+			ip_instance = 0;
+			break;
+		case SOC15_IH_CLIENTID_VCN1:
+			ip_instance = 1;
+			break;
+		default:
+			DRM_ERROR("Unhandled client id: %d\n", entry->client_id);
+			return 0;
+		}
+	}
+
+	DRM_DEBUG("IH: VCN TRAP\n");
+
+	switch (entry->src_id) {
+	case VCN_4_0__SRCID__UVD_ENC_GENERAL_PURPOSE:
+		amdgpu_fence_process(&adev->vcn.inst[ip_instance].ring_enc[0]);
+		break;
+	default:
+		DRM_ERROR("Unhandled interrupt: %d %d\n",
+			  entry->src_id, entry->src_data[0]);
+		break;
+	}
+
+	return 0;
+}
+
+static const struct amdgpu_irq_src_funcs vcn_v4_0_irq_funcs = {
+	.process = vcn_v4_0_process_interrupt,
+};
+
+static const struct amdgpu_irq_src_funcs vcn_v4_0_ras_irq_funcs = {
+	.set = vcn_v4_0_set_ras_interrupt_state,
+	.process = amdgpu_vcn_process_poison_irq,
+};
+
+/**
+ * vcn_v4_0_set_irq_funcs - set VCN block interrupt irq functions
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Set VCN block interrupt irq functions
+ */
+static void vcn_v4_0_set_irq_funcs(struct amdgpu_device *adev)
+{
+	int i;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		if (adev->vcn.harvest_config & (1 << i))
+			continue;
+
+		adev->vcn.inst[i].irq.num_types = adev->vcn.inst[i].num_enc_rings + 1;
+		adev->vcn.inst[i].irq.funcs = &vcn_v4_0_irq_funcs;
+
+		adev->vcn.inst[i].ras_poison_irq.num_types = adev->vcn.inst[i].num_enc_rings + 1;
+		adev->vcn.inst[i].ras_poison_irq.funcs = &vcn_v4_0_ras_irq_funcs;
+	}
+}
+
+static const struct amd_ip_funcs vcn_v4_0_ip_funcs = {
+	.name = "vcn_v4_0",
+	.early_init = vcn_v4_0_early_init,
+	.sw_init = vcn_v4_0_sw_init,
+	.sw_fini = vcn_v4_0_sw_fini,
+	.hw_init = vcn_v4_0_hw_init,
+	.hw_fini = vcn_v4_0_hw_fini,
+	.suspend = vcn_v4_0_suspend,
+	.resume = vcn_v4_0_resume,
+	.is_idle = vcn_v4_0_is_idle,
+	.wait_for_idle = vcn_v4_0_wait_for_idle,
+	.set_clockgating_state = vcn_v4_0_set_clockgating_state,
+	.set_powergating_state = vcn_set_powergating_state,
+	.dump_ip_state = amdgpu_vcn_dump_ip_state,
+	.print_ip_state = amdgpu_vcn_print_ip_state,
+};
+
+const struct amdgpu_ip_block_version vcn_v4_0_ip_block = {
+	.type = AMD_IP_BLOCK_TYPE_VCN,
+	.major = 4,
+	.minor = 0,
+	.rev = 0,
+	.funcs = &vcn_v4_0_ip_funcs,
+};
+
+static uint32_t vcn_v4_0_query_poison_by_instance(struct amdgpu_device *adev,
+			uint32_t instance, uint32_t sub_block)
+{
+	uint32_t poison_stat = 0, reg_value = 0;
+
+	switch (sub_block) {
+	case AMDGPU_VCN_V4_0_VCPU_VCODEC:
+		reg_value = RREG32_SOC15(VCN, instance, regUVD_RAS_VCPU_VCODEC_STATUS);
+		poison_stat = REG_GET_FIELD(reg_value, UVD_RAS_VCPU_VCODEC_STATUS, POISONED_PF);
+		break;
+	default:
+		break;
+	}
+
+	if (poison_stat)
+		dev_info(adev->dev, "Poison detected in VCN%d, sub_block%d\n",
+			instance, sub_block);
+
+	return poison_stat;
+}
+
+static bool vcn_v4_0_query_ras_poison_status(struct amdgpu_device *adev)
+{
+	uint32_t inst, sub;
+	uint32_t poison_stat = 0;
+
+	for (inst = 0; inst < adev->vcn.num_vcn_inst; inst++)
+		for (sub = 0; sub < AMDGPU_VCN_V4_0_MAX_SUB_BLOCK; sub++)
+			poison_stat +=
+				vcn_v4_0_query_poison_by_instance(adev, inst, sub);
+
+	return !!poison_stat;
+}
+
+const struct amdgpu_ras_block_hw_ops vcn_v4_0_ras_hw_ops = {
+	.query_poison_status = vcn_v4_0_query_ras_poison_status,
+};
+
+static struct amdgpu_vcn_ras vcn_v4_0_ras = {
+	.ras_block = {
+		.hw_ops = &vcn_v4_0_ras_hw_ops,
+		.ras_late_init = amdgpu_vcn_ras_late_init,
+	},
+};
+
+static void vcn_v4_0_set_ras_funcs(struct amdgpu_device *adev)
+{
+	switch (amdgpu_ip_version(adev, VCN_HWIP, 0)) {
+	case IP_VERSION(4, 0, 0):
+		adev->vcn.ras = &vcn_v4_0_ras;
+		break;
+	default:
+		break;
+	}
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.h b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.h
new file mode 100644
index 000000000000..7d3d11f40f27
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.h
@@ -0,0 +1,35 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __VCN_V4_0_H__
+#define __VCN_V4_0_H__
+
+enum amdgpu_vcn_v4_0_sub_block {
+	AMDGPU_VCN_V4_0_VCPU_VCODEC = 0,
+
+	AMDGPU_VCN_V4_0_MAX_SUB_BLOCK,
+};
+
+extern const struct amdgpu_ip_block_version vcn_v4_0_ip_block;
+
+#endif /* __VCN_V4_0_H__ */
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c
new file mode 100644
index 000000000000..cb7123ec1a5d
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c
@@ -0,0 +1,2131 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/firmware.h>
+#include <drm/drm_drv.h>
+
+#include "amdgpu.h"
+#include "amdgpu_vcn.h"
+#include "amdgpu_pm.h"
+#include "soc15.h"
+#include "soc15d.h"
+#include "soc15_hw_ip.h"
+#include "vcn_v2_0.h"
+#include "vcn_v4_0_3.h"
+#include "mmsch_v4_0_3.h"
+
+#include "vcn/vcn_4_0_3_offset.h"
+#include "vcn/vcn_4_0_3_sh_mask.h"
+#include "ivsrcid/vcn/irqsrcs_vcn_4_0.h"
+
+#define mmUVD_DPG_LMA_CTL		regUVD_DPG_LMA_CTL
+#define mmUVD_DPG_LMA_CTL_BASE_IDX	regUVD_DPG_LMA_CTL_BASE_IDX
+#define mmUVD_DPG_LMA_DATA		regUVD_DPG_LMA_DATA
+#define mmUVD_DPG_LMA_DATA_BASE_IDX	regUVD_DPG_LMA_DATA_BASE_IDX
+
+#define VCN_VID_SOC_ADDRESS_2_0		0x1fb00
+#define VCN1_VID_SOC_ADDRESS_3_0	0x48300
+#define VCN1_AON_SOC_ADDRESS_3_0	0x48000
+
+static const struct amdgpu_hwip_reg_entry vcn_reg_list_4_0_3[] = {
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_POWER_STATUS),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_STATUS),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_CONTEXT_ID),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_CONTEXT_ID2),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_GPCOM_VCPU_DATA0),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_GPCOM_VCPU_DATA1),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_GPCOM_VCPU_CMD),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_BASE_HI),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_BASE_LO),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_BASE_HI2),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_BASE_LO2),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_BASE_HI3),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_BASE_LO3),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_BASE_HI4),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_BASE_LO4),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_RPTR),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_WPTR),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_RPTR2),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_WPTR2),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_RPTR3),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_WPTR3),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_RPTR4),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_WPTR4),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_SIZE),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_SIZE2),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_SIZE3),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_SIZE4),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_PGFSM_CONFIG),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_PGFSM_STATUS),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_DPG_LMA_CTL),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_DPG_LMA_DATA),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_DPG_LMA_MASK),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_DPG_PAUSE)
+};
+
+#define NORMALIZE_VCN_REG_OFFSET(offset) \
+		(offset & 0x1FFFF)
+
+static int vcn_v4_0_3_start_sriov(struct amdgpu_device *adev);
+static void vcn_v4_0_3_set_unified_ring_funcs(struct amdgpu_device *adev);
+static void vcn_v4_0_3_set_irq_funcs(struct amdgpu_device *adev);
+static int vcn_v4_0_3_set_pg_state(struct amdgpu_vcn_inst *vinst,
+				   enum amd_powergating_state state);
+static int vcn_v4_0_3_pause_dpg_mode(struct amdgpu_vcn_inst *vinst,
+				     struct dpg_pause_state *new_state);
+static void vcn_v4_0_3_unified_ring_set_wptr(struct amdgpu_ring *ring);
+static void vcn_v4_0_3_set_ras_funcs(struct amdgpu_device *adev);
+static void vcn_v4_0_3_enable_ras(struct amdgpu_device *adev,
+				  int inst_idx, bool indirect);
+
+static inline bool vcn_v4_0_3_normalizn_reqd(struct amdgpu_device *adev)
+{
+	return (adev->vcn.caps & AMDGPU_VCN_CAPS(RRMT_ENABLED)) == 0;
+}
+
+/**
+ * vcn_v4_0_3_early_init - set function pointers
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Set ring and irq function pointers
+ */
+static int vcn_v4_0_3_early_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, r;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i)
+		/* re-use enc ring as unified ring */
+		adev->vcn.inst[i].num_enc_rings = 1;
+
+	vcn_v4_0_3_set_unified_ring_funcs(adev);
+	vcn_v4_0_3_set_irq_funcs(adev);
+	vcn_v4_0_3_set_ras_funcs(adev);
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		adev->vcn.inst[i].set_pg_state = vcn_v4_0_3_set_pg_state;
+
+		r = amdgpu_vcn_early_init(adev, i);
+		if (r)
+			return r;
+	}
+
+	return 0;
+}
+
+static int vcn_v4_0_3_late_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	adev->vcn.supported_reset =
+		amdgpu_get_soft_full_reset_mask(&adev->vcn.inst[0].ring_enc[0]);
+
+	if (amdgpu_dpm_reset_vcn_is_supported(adev) && !amdgpu_sriov_vf(adev))
+		adev->vcn.supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE;
+
+	return 0;
+}
+
+static int vcn_v4_0_3_fw_shared_init(struct amdgpu_device *adev, int inst_idx)
+{
+	struct amdgpu_vcn4_fw_shared *fw_shared;
+
+	fw_shared = adev->vcn.inst[inst_idx].fw_shared.cpu_addr;
+	fw_shared->present_flag_0 = cpu_to_le32(AMDGPU_FW_SHARED_FLAG_0_UNIFIED_QUEUE);
+	fw_shared->sq.is_enabled = 1;
+
+	if (amdgpu_vcnfw_log)
+		amdgpu_vcn_fwlog_init(&adev->vcn.inst[inst_idx]);
+
+	return 0;
+}
+
+/**
+ * vcn_v4_0_3_sw_init - sw init for VCN block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Load firmware and sw initialization
+ */
+static int vcn_v4_0_3_sw_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	struct amdgpu_ring *ring;
+	int i, r, vcn_inst;
+
+	/* VCN DEC TRAP */
+	r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_VCN,
+		VCN_4_0__SRCID__UVD_ENC_GENERAL_PURPOSE, &adev->vcn.inst->irq);
+	if (r)
+		return r;
+
+	/* VCN POISON TRAP */
+	r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_VCN,
+		VCN_4_0__SRCID_UVD_POISON, &adev->vcn.inst->ras_poison_irq);
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+
+		r = amdgpu_vcn_sw_init(adev, i);
+		if (r)
+			return r;
+
+		amdgpu_vcn_setup_ucode(adev, i);
+
+		r = amdgpu_vcn_resume(adev, i);
+		if (r)
+			return r;
+
+		vcn_inst = GET_INST(VCN, i);
+
+		ring = &adev->vcn.inst[i].ring_enc[0];
+		ring->use_doorbell = true;
+
+		if (!amdgpu_sriov_vf(adev))
+			ring->doorbell_index =
+				(adev->doorbell_index.vcn.vcn_ring0_1 << 1) +
+				9 * vcn_inst;
+		else
+			ring->doorbell_index =
+				(adev->doorbell_index.vcn.vcn_ring0_1 << 1) +
+				32 * vcn_inst;
+
+		ring->vm_hub = AMDGPU_MMHUB0(adev->vcn.inst[i].aid_id);
+		sprintf(ring->name, "vcn_unified_%d", adev->vcn.inst[i].aid_id);
+
+		/* There are no per-instance irq source IDs on 4.0.3, the IH
+		 * packets use a separate field to differentiate instances.
+		 */
+		r = amdgpu_ring_init(adev, ring, 512, &adev->vcn.inst[0].irq, 0,
+				     AMDGPU_RING_PRIO_DEFAULT,
+				     &adev->vcn.inst[i].sched_score);
+		if (r)
+			return r;
+
+		vcn_v4_0_3_fw_shared_init(adev, i);
+
+		if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)
+			adev->vcn.inst[i].pause_dpg_mode = vcn_v4_0_3_pause_dpg_mode;
+	}
+
+	if (amdgpu_sriov_vf(adev)) {
+		r = amdgpu_virt_alloc_mm_table(adev);
+		if (r)
+			return r;
+	}
+
+	if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__VCN)) {
+		r = amdgpu_vcn_ras_sw_init(adev);
+		if (r) {
+			dev_err(adev->dev, "Failed to initialize vcn ras block!\n");
+			return r;
+		}
+	}
+
+	r = amdgpu_vcn_reg_dump_init(adev, vcn_reg_list_4_0_3, ARRAY_SIZE(vcn_reg_list_4_0_3));
+	if (r)
+		return r;
+
+	return amdgpu_vcn_sysfs_reset_mask_init(adev);
+}
+
+/**
+ * vcn_v4_0_3_sw_fini - sw fini for VCN block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * VCN suspend and free up sw allocation
+ */
+static int vcn_v4_0_3_sw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, r, idx;
+
+	if (drm_dev_enter(&adev->ddev, &idx)) {
+		for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+			struct amdgpu_vcn4_fw_shared *fw_shared;
+
+			fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr;
+			fw_shared->present_flag_0 = 0;
+			fw_shared->sq.is_enabled = cpu_to_le32(false);
+		}
+		drm_dev_exit(idx);
+	}
+
+	if (amdgpu_sriov_vf(adev))
+		amdgpu_virt_free_mm_table(adev);
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+		r = amdgpu_vcn_suspend(adev, i);
+		if (r)
+			return r;
+	}
+
+	amdgpu_vcn_sysfs_reset_mask_fini(adev);
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++)
+		amdgpu_vcn_sw_fini(adev, i);
+
+	return 0;
+}
+
+static int vcn_v4_0_3_hw_init_inst(struct amdgpu_vcn_inst *vinst)
+{
+	int vcn_inst;
+	struct amdgpu_device *adev = vinst->adev;
+	struct amdgpu_ring *ring;
+	int inst_idx = vinst->inst;
+
+	vcn_inst = GET_INST(VCN, inst_idx);
+	ring = &adev->vcn.inst[inst_idx].ring_enc[0];
+	if (ring->use_doorbell) {
+		adev->nbio.funcs->vcn_doorbell_range(adev, ring->use_doorbell,
+			(adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 9 * vcn_inst,
+			adev->vcn.inst[inst_idx].aid_id);
+
+		WREG32_SOC15(VCN, vcn_inst, regVCN_RB1_DB_CTRL,
+			ring->doorbell_index << VCN_RB1_DB_CTRL__OFFSET__SHIFT |
+			VCN_RB1_DB_CTRL__EN_MASK);
+
+		/* Read DB_CTRL to flush the write DB_CTRL command. */
+		RREG32_SOC15(VCN, vcn_inst, regVCN_RB1_DB_CTRL);
+	}
+
+	return 0;
+}
+
+/**
+ * vcn_v4_0_3_hw_init - start and test VCN block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Initialize the hardware, boot up the VCPU and do some testing
+ */
+static int vcn_v4_0_3_hw_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	struct amdgpu_ring *ring;
+	struct amdgpu_vcn_inst *vinst;
+	int i, r;
+
+	if (amdgpu_sriov_vf(adev)) {
+		r = vcn_v4_0_3_start_sriov(adev);
+		if (r)
+			return r;
+
+		for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+			ring = &adev->vcn.inst[i].ring_enc[0];
+			ring->wptr = 0;
+			ring->wptr_old = 0;
+			vcn_v4_0_3_unified_ring_set_wptr(ring);
+			ring->sched.ready = true;
+		}
+	} else {
+		/* This flag is not set for VF, assumed to be disabled always */
+		if (RREG32_SOC15(VCN, GET_INST(VCN, 0), regVCN_RRMT_CNTL) &
+		    0x100)
+			adev->vcn.caps |= AMDGPU_VCN_CAPS(RRMT_ENABLED);
+
+		for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+			struct amdgpu_vcn4_fw_shared *fw_shared;
+
+			ring = &adev->vcn.inst[i].ring_enc[0];
+			vinst = &adev->vcn.inst[i];
+			vcn_v4_0_3_hw_init_inst(vinst);
+
+			/* Re-init fw_shared when RAS fatal error occurred */
+			fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr;
+			if (!fw_shared->sq.is_enabled)
+				vcn_v4_0_3_fw_shared_init(adev, i);
+
+			r = amdgpu_ring_test_helper(ring);
+			if (r)
+				return r;
+		}
+	}
+
+	return r;
+}
+
+/**
+ * vcn_v4_0_3_hw_fini - stop the hardware block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Stop the VCN block, mark ring as not ready any more
+ */
+static int vcn_v4_0_3_hw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		struct amdgpu_vcn_inst *vinst = &adev->vcn.inst[i];
+
+		cancel_delayed_work_sync(&vinst->idle_work);
+
+		if (vinst->cur_state != AMD_PG_STATE_GATE)
+			vinst->set_pg_state(vinst, AMD_PG_STATE_GATE);
+	}
+
+	if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__VCN) && !amdgpu_sriov_vf(adev))
+		amdgpu_irq_put(adev, &adev->vcn.inst->ras_poison_irq, 0);
+
+	return 0;
+}
+
+/**
+ * vcn_v4_0_3_suspend - suspend VCN block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * HW fini and suspend VCN block
+ */
+static int vcn_v4_0_3_suspend(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int r, i;
+
+	r = vcn_v4_0_3_hw_fini(ip_block);
+	if (r)
+		return r;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+		r = amdgpu_vcn_suspend(adev, i);
+		if (r)
+			return r;
+	}
+
+	return 0;
+}
+
+/**
+ * vcn_v4_0_3_resume - resume VCN block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Resume firmware and hw init VCN block
+ */
+static int vcn_v4_0_3_resume(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int r, i;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+		r = amdgpu_vcn_resume(ip_block->adev, i);
+		if (r)
+			return r;
+	}
+
+	r = vcn_v4_0_3_hw_init(ip_block);
+
+	return r;
+}
+
+/**
+ * vcn_v4_0_3_mc_resume - memory controller programming
+ *
+ * @vinst: VCN instance
+ *
+ * Let the VCN memory controller know it's offsets
+ */
+static void vcn_v4_0_3_mc_resume(struct amdgpu_vcn_inst *vinst)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int inst_idx = vinst->inst;
+	uint32_t offset, size, vcn_inst;
+	const struct common_firmware_header *hdr;
+
+	hdr = (const struct common_firmware_header *)adev->vcn.inst[inst_idx].fw->data;
+	size = AMDGPU_GPU_PAGE_ALIGN(le32_to_cpu(hdr->ucode_size_bytes) + 8);
+
+	vcn_inst = GET_INST(VCN, inst_idx);
+	/* cache window 0: fw */
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
+		WREG32_SOC15(
+			VCN, vcn_inst, regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW,
+			(adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + inst_idx]
+				 .tmr_mc_addr_lo));
+		WREG32_SOC15(
+			VCN, vcn_inst, regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH,
+			(adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + inst_idx]
+				 .tmr_mc_addr_hi));
+		WREG32_SOC15(VCN, vcn_inst, regUVD_VCPU_CACHE_OFFSET0, 0);
+		offset = 0;
+	} else {
+		WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW,
+			     lower_32_bits(adev->vcn.inst[inst_idx].gpu_addr));
+		WREG32_SOC15(VCN, vcn_inst,
+			     regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH,
+			     upper_32_bits(adev->vcn.inst[inst_idx].gpu_addr));
+		offset = size;
+		WREG32_SOC15(VCN, vcn_inst, regUVD_VCPU_CACHE_OFFSET0,
+			     AMDGPU_UVD_FIRMWARE_OFFSET >> 3);
+	}
+	WREG32_SOC15(VCN, vcn_inst, regUVD_VCPU_CACHE_SIZE0, size);
+
+	/* cache window 1: stack */
+	WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW,
+		     lower_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset));
+	WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH,
+		     upper_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset));
+	WREG32_SOC15(VCN, vcn_inst, regUVD_VCPU_CACHE_OFFSET1, 0);
+	WREG32_SOC15(VCN, vcn_inst, regUVD_VCPU_CACHE_SIZE1,
+		     AMDGPU_VCN_STACK_SIZE);
+
+	/* cache window 2: context */
+	WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_VCPU_CACHE2_64BIT_BAR_LOW,
+		     lower_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset +
+				   AMDGPU_VCN_STACK_SIZE));
+	WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_VCPU_CACHE2_64BIT_BAR_HIGH,
+		     upper_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset +
+				   AMDGPU_VCN_STACK_SIZE));
+	WREG32_SOC15(VCN, vcn_inst, regUVD_VCPU_CACHE_OFFSET2, 0);
+	WREG32_SOC15(VCN, vcn_inst, regUVD_VCPU_CACHE_SIZE2,
+		     AMDGPU_VCN_CONTEXT_SIZE);
+
+	/* non-cache window */
+	WREG32_SOC15(
+		VCN, vcn_inst, regUVD_LMI_VCPU_NC0_64BIT_BAR_LOW,
+		lower_32_bits(adev->vcn.inst[inst_idx].fw_shared.gpu_addr));
+	WREG32_SOC15(
+		VCN, vcn_inst, regUVD_LMI_VCPU_NC0_64BIT_BAR_HIGH,
+		upper_32_bits(adev->vcn.inst[inst_idx].fw_shared.gpu_addr));
+	WREG32_SOC15(VCN, vcn_inst, regUVD_VCPU_NONCACHE_OFFSET0, 0);
+	WREG32_SOC15(
+		VCN, vcn_inst, regUVD_VCPU_NONCACHE_SIZE0,
+		AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_vcn4_fw_shared)));
+}
+
+/**
+ * vcn_v4_0_3_mc_resume_dpg_mode - memory controller programming for dpg mode
+ *
+ * @vinst: VCN instance
+ * @indirect: indirectly write sram
+ *
+ * Let the VCN memory controller know it's offsets with dpg mode
+ */
+static void vcn_v4_0_3_mc_resume_dpg_mode(struct amdgpu_vcn_inst *vinst,
+					  bool indirect)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int inst_idx = vinst->inst;
+	uint32_t offset, size;
+	const struct common_firmware_header *hdr;
+
+	hdr = (const struct common_firmware_header *)adev->vcn.inst[inst_idx].fw->data;
+	size = AMDGPU_GPU_PAGE_ALIGN(le32_to_cpu(hdr->ucode_size_bytes) + 8);
+
+	/* cache window 0: fw */
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
+		if (!indirect) {
+			WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+				VCN, 0, regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW),
+				(adev->firmware.ucode[AMDGPU_UCODE_ID_VCN +
+					inst_idx].tmr_mc_addr_lo), 0, indirect);
+			WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+				VCN, 0, regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH),
+				(adev->firmware.ucode[AMDGPU_UCODE_ID_VCN +
+					inst_idx].tmr_mc_addr_hi), 0, indirect);
+			WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+				VCN, 0, regUVD_VCPU_CACHE_OFFSET0), 0, 0, indirect);
+		} else {
+			WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+				VCN, 0, regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW), 0, 0, indirect);
+			WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+				VCN, 0, regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH), 0, 0, indirect);
+			WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+				VCN, 0, regUVD_VCPU_CACHE_OFFSET0), 0, 0, indirect);
+		}
+		offset = 0;
+	} else {
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, 0, regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW),
+			lower_32_bits(adev->vcn.inst[inst_idx].gpu_addr), 0, indirect);
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, 0, regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH),
+			upper_32_bits(adev->vcn.inst[inst_idx].gpu_addr), 0, indirect);
+		offset = size;
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, 0, regUVD_VCPU_CACHE_OFFSET0),
+			AMDGPU_UVD_FIRMWARE_OFFSET >> 3, 0, indirect);
+	}
+
+	if (!indirect)
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, 0, regUVD_VCPU_CACHE_SIZE0), size, 0, indirect);
+	else
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, 0, regUVD_VCPU_CACHE_SIZE0), 0, 0, indirect);
+
+	/* cache window 1: stack */
+	if (!indirect) {
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, 0, regUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW),
+			lower_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset), 0, indirect);
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, 0, regUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH),
+			upper_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset), 0, indirect);
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, 0, regUVD_VCPU_CACHE_OFFSET1), 0, 0, indirect);
+	} else {
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, 0, regUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW), 0, 0, indirect);
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, 0, regUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH), 0, 0, indirect);
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, 0, regUVD_VCPU_CACHE_OFFSET1), 0, 0, indirect);
+	}
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, 0, regUVD_VCPU_CACHE_SIZE1), AMDGPU_VCN_STACK_SIZE, 0, indirect);
+
+	/* cache window 2: context */
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, 0, regUVD_LMI_VCPU_CACHE2_64BIT_BAR_LOW),
+			lower_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset +
+				AMDGPU_VCN_STACK_SIZE), 0, indirect);
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, 0, regUVD_LMI_VCPU_CACHE2_64BIT_BAR_HIGH),
+			upper_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset +
+				AMDGPU_VCN_STACK_SIZE), 0, indirect);
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, 0, regUVD_VCPU_CACHE_OFFSET2), 0, 0, indirect);
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, 0, regUVD_VCPU_CACHE_SIZE2), AMDGPU_VCN_CONTEXT_SIZE, 0, indirect);
+
+	/* non-cache window */
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, 0, regUVD_LMI_VCPU_NC0_64BIT_BAR_LOW),
+			lower_32_bits(adev->vcn.inst[inst_idx].fw_shared.gpu_addr), 0, indirect);
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, 0, regUVD_LMI_VCPU_NC0_64BIT_BAR_HIGH),
+			upper_32_bits(adev->vcn.inst[inst_idx].fw_shared.gpu_addr), 0, indirect);
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, 0, regUVD_VCPU_NONCACHE_OFFSET0), 0, 0, indirect);
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, 0, regUVD_VCPU_NONCACHE_SIZE0),
+			AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_vcn4_fw_shared)), 0, indirect);
+
+	/* VCN global tiling registers */
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, 0, regUVD_GFX8_ADDR_CONFIG), adev->gfx.config.gb_addr_config, 0, indirect);
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, 0, regUVD_GFX10_ADDR_CONFIG), adev->gfx.config.gb_addr_config, 0, indirect);
+}
+
+/**
+ * vcn_v4_0_3_disable_clock_gating - disable VCN clock gating
+ *
+ * @vinst: VCN instance
+ *
+ * Disable clock gating for VCN block
+ */
+static void vcn_v4_0_3_disable_clock_gating(struct amdgpu_vcn_inst *vinst)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int inst_idx = vinst->inst;
+	uint32_t data;
+	int vcn_inst;
+
+	if (adev->cg_flags & AMD_CG_SUPPORT_VCN_MGCG)
+		return;
+
+	vcn_inst = GET_INST(VCN, inst_idx);
+
+	/* VCN disable CGC */
+	data = RREG32_SOC15(VCN, vcn_inst, regUVD_CGC_CTRL);
+	data &= ~UVD_CGC_CTRL__DYN_CLOCK_MODE_MASK;
+	data |= 1 << UVD_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT;
+	data |= 4 << UVD_CGC_CTRL__CLK_OFF_DELAY__SHIFT;
+	WREG32_SOC15(VCN, vcn_inst, regUVD_CGC_CTRL, data);
+
+	data = RREG32_SOC15(VCN, vcn_inst, regUVD_CGC_GATE);
+	data &= ~(UVD_CGC_GATE__SYS_MASK
+		| UVD_CGC_GATE__MPEG2_MASK
+		| UVD_CGC_GATE__REGS_MASK
+		| UVD_CGC_GATE__RBC_MASK
+		| UVD_CGC_GATE__LMI_MC_MASK
+		| UVD_CGC_GATE__LMI_UMC_MASK
+		| UVD_CGC_GATE__MPC_MASK
+		| UVD_CGC_GATE__LBSI_MASK
+		| UVD_CGC_GATE__LRBBM_MASK
+		| UVD_CGC_GATE__WCB_MASK
+		| UVD_CGC_GATE__VCPU_MASK
+		| UVD_CGC_GATE__MMSCH_MASK);
+
+	WREG32_SOC15(VCN, vcn_inst, regUVD_CGC_GATE, data);
+	SOC15_WAIT_ON_RREG(VCN, vcn_inst, regUVD_CGC_GATE, 0, 0xFFFFFFFF);
+
+	data = RREG32_SOC15(VCN, vcn_inst, regUVD_CGC_CTRL);
+	data &= ~(UVD_CGC_CTRL__SYS_MODE_MASK
+		| UVD_CGC_CTRL__MPEG2_MODE_MASK
+		| UVD_CGC_CTRL__REGS_MODE_MASK
+		| UVD_CGC_CTRL__RBC_MODE_MASK
+		| UVD_CGC_CTRL__LMI_MC_MODE_MASK
+		| UVD_CGC_CTRL__LMI_UMC_MODE_MASK
+		| UVD_CGC_CTRL__MPC_MODE_MASK
+		| UVD_CGC_CTRL__LBSI_MODE_MASK
+		| UVD_CGC_CTRL__LRBBM_MODE_MASK
+		| UVD_CGC_CTRL__WCB_MODE_MASK
+		| UVD_CGC_CTRL__VCPU_MODE_MASK
+		| UVD_CGC_CTRL__MMSCH_MODE_MASK);
+	WREG32_SOC15(VCN, vcn_inst, regUVD_CGC_CTRL, data);
+
+	data = RREG32_SOC15(VCN, vcn_inst, regUVD_SUVD_CGC_GATE);
+	data |= (UVD_SUVD_CGC_GATE__SRE_MASK
+		| UVD_SUVD_CGC_GATE__SIT_MASK
+		| UVD_SUVD_CGC_GATE__SMP_MASK
+		| UVD_SUVD_CGC_GATE__SCM_MASK
+		| UVD_SUVD_CGC_GATE__SDB_MASK
+		| UVD_SUVD_CGC_GATE__SRE_H264_MASK
+		| UVD_SUVD_CGC_GATE__SRE_HEVC_MASK
+		| UVD_SUVD_CGC_GATE__SIT_H264_MASK
+		| UVD_SUVD_CGC_GATE__SIT_HEVC_MASK
+		| UVD_SUVD_CGC_GATE__SCM_H264_MASK
+		| UVD_SUVD_CGC_GATE__SCM_HEVC_MASK
+		| UVD_SUVD_CGC_GATE__SDB_H264_MASK
+		| UVD_SUVD_CGC_GATE__SDB_HEVC_MASK
+		| UVD_SUVD_CGC_GATE__ENT_MASK
+		| UVD_SUVD_CGC_GATE__SIT_HEVC_DEC_MASK
+		| UVD_SUVD_CGC_GATE__SITE_MASK
+		| UVD_SUVD_CGC_GATE__SRE_VP9_MASK
+		| UVD_SUVD_CGC_GATE__SCM_VP9_MASK
+		| UVD_SUVD_CGC_GATE__SIT_VP9_DEC_MASK
+		| UVD_SUVD_CGC_GATE__SDB_VP9_MASK
+		| UVD_SUVD_CGC_GATE__IME_HEVC_MASK);
+	WREG32_SOC15(VCN, vcn_inst, regUVD_SUVD_CGC_GATE, data);
+
+	data = RREG32_SOC15(VCN, vcn_inst, regUVD_SUVD_CGC_CTRL);
+	data &= ~(UVD_SUVD_CGC_CTRL__SRE_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SIT_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SMP_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SCM_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SDB_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__ENT_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__IME_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SITE_MODE_MASK);
+	WREG32_SOC15(VCN, vcn_inst, regUVD_SUVD_CGC_CTRL, data);
+}
+
+/**
+ * vcn_v4_0_3_disable_clock_gating_dpg_mode - disable VCN clock gating dpg mode
+ *
+ * @vinst: VCN instance
+ * @sram_sel: sram select
+ * @indirect: indirectly write sram
+ *
+ * Disable clock gating for VCN block with dpg mode
+ */
+static void vcn_v4_0_3_disable_clock_gating_dpg_mode(struct amdgpu_vcn_inst *vinst,
+						     uint8_t sram_sel,
+						     uint8_t indirect)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int inst_idx = vinst->inst;
+	uint32_t reg_data = 0;
+
+	if (adev->cg_flags & AMD_CG_SUPPORT_VCN_MGCG)
+		return;
+
+	/* enable sw clock gating control */
+	reg_data = 0 << UVD_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
+	reg_data |= 1 << UVD_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT;
+	reg_data |= 4 << UVD_CGC_CTRL__CLK_OFF_DELAY__SHIFT;
+	reg_data &= ~(UVD_CGC_CTRL__SYS_MODE_MASK |
+		 UVD_CGC_CTRL__MPEG2_MODE_MASK |
+		 UVD_CGC_CTRL__REGS_MODE_MASK |
+		 UVD_CGC_CTRL__RBC_MODE_MASK |
+		 UVD_CGC_CTRL__LMI_MC_MODE_MASK |
+		 UVD_CGC_CTRL__LMI_UMC_MODE_MASK |
+		 UVD_CGC_CTRL__IDCT_MODE_MASK |
+		 UVD_CGC_CTRL__MPRD_MODE_MASK |
+		 UVD_CGC_CTRL__MPC_MODE_MASK |
+		 UVD_CGC_CTRL__LBSI_MODE_MASK |
+		 UVD_CGC_CTRL__LRBBM_MODE_MASK |
+		 UVD_CGC_CTRL__WCB_MODE_MASK |
+		 UVD_CGC_CTRL__VCPU_MODE_MASK);
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, 0, regUVD_CGC_CTRL), reg_data, sram_sel, indirect);
+
+	/* turn off clock gating */
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, 0, regUVD_CGC_GATE), 0, sram_sel, indirect);
+
+	/* turn on SUVD clock gating */
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, 0, regUVD_SUVD_CGC_GATE), 1, sram_sel, indirect);
+
+	/* turn on sw mode in UVD_SUVD_CGC_CTRL */
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, 0, regUVD_SUVD_CGC_CTRL), 0, sram_sel, indirect);
+}
+
+/**
+ * vcn_v4_0_3_enable_clock_gating - enable VCN clock gating
+ *
+ * @vinst: VCN instance
+ *
+ * Enable clock gating for VCN block
+ */
+static void vcn_v4_0_3_enable_clock_gating(struct amdgpu_vcn_inst *vinst)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int inst_idx = vinst->inst;
+	uint32_t data;
+	int vcn_inst;
+
+	if (adev->cg_flags & AMD_CG_SUPPORT_VCN_MGCG)
+		return;
+
+	vcn_inst = GET_INST(VCN, inst_idx);
+
+	/* enable VCN CGC */
+	data = RREG32_SOC15(VCN, vcn_inst, regUVD_CGC_CTRL);
+	data |= 0 << UVD_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
+	data |= 1 << UVD_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT;
+	data |= 4 << UVD_CGC_CTRL__CLK_OFF_DELAY__SHIFT;
+	WREG32_SOC15(VCN, vcn_inst, regUVD_CGC_CTRL, data);
+
+	data = RREG32_SOC15(VCN, vcn_inst, regUVD_CGC_CTRL);
+	data |= (UVD_CGC_CTRL__SYS_MODE_MASK
+		| UVD_CGC_CTRL__MPEG2_MODE_MASK
+		| UVD_CGC_CTRL__REGS_MODE_MASK
+		| UVD_CGC_CTRL__RBC_MODE_MASK
+		| UVD_CGC_CTRL__LMI_MC_MODE_MASK
+		| UVD_CGC_CTRL__LMI_UMC_MODE_MASK
+		| UVD_CGC_CTRL__MPC_MODE_MASK
+		| UVD_CGC_CTRL__LBSI_MODE_MASK
+		| UVD_CGC_CTRL__LRBBM_MODE_MASK
+		| UVD_CGC_CTRL__WCB_MODE_MASK
+		| UVD_CGC_CTRL__VCPU_MODE_MASK);
+	WREG32_SOC15(VCN, vcn_inst, regUVD_CGC_CTRL, data);
+
+	data = RREG32_SOC15(VCN, vcn_inst, regUVD_SUVD_CGC_CTRL);
+	data |= (UVD_SUVD_CGC_CTRL__SRE_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SIT_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SMP_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SCM_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SDB_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__ENT_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__IME_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SITE_MODE_MASK);
+	WREG32_SOC15(VCN, vcn_inst, regUVD_SUVD_CGC_CTRL, data);
+}
+
+/**
+ * vcn_v4_0_3_start_dpg_mode - VCN start with dpg mode
+ *
+ * @vinst: VCN instance
+ * @indirect: indirectly write sram
+ *
+ * Start VCN block with dpg mode
+ */
+static int vcn_v4_0_3_start_dpg_mode(struct amdgpu_vcn_inst *vinst,
+				     bool indirect)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int inst_idx = vinst->inst;
+	struct amdgpu_vcn4_fw_shared *fw_shared =
+						adev->vcn.inst[inst_idx].fw_shared.cpu_addr;
+	struct amdgpu_ring *ring;
+	int vcn_inst, ret;
+	uint32_t tmp;
+
+	vcn_inst = GET_INST(VCN, inst_idx);
+	/* disable register anti-hang mechanism */
+	WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_POWER_STATUS), 1,
+		 ~UVD_POWER_STATUS__UVD_POWER_STATUS_MASK);
+	/* enable dynamic power gating mode */
+	tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_POWER_STATUS);
+	tmp |= UVD_POWER_STATUS__UVD_PG_MODE_MASK;
+	tmp |= UVD_POWER_STATUS__UVD_PG_EN_MASK;
+	WREG32_SOC15(VCN, vcn_inst, regUVD_POWER_STATUS, tmp);
+
+	if (indirect) {
+		DRM_DEV_DEBUG(adev->dev, "VCN %d start: on AID %d",
+			inst_idx, adev->vcn.inst[inst_idx].aid_id);
+		adev->vcn.inst[inst_idx].dpg_sram_curr_addr =
+				(uint32_t *)adev->vcn.inst[inst_idx].dpg_sram_cpu_addr;
+		/* Use dummy register 0xDEADBEEF passing AID selection to PSP FW */
+		WREG32_SOC15_DPG_MODE(inst_idx, 0xDEADBEEF,
+			adev->vcn.inst[inst_idx].aid_id, 0, true);
+	}
+
+	/* enable clock gating */
+	vcn_v4_0_3_disable_clock_gating_dpg_mode(vinst, 0, indirect);
+
+	/* enable VCPU clock */
+	tmp = (0xFF << UVD_VCPU_CNTL__PRB_TIMEOUT_VAL__SHIFT);
+	tmp |= UVD_VCPU_CNTL__CLK_EN_MASK;
+	tmp |= UVD_VCPU_CNTL__BLK_RST_MASK;
+
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, 0, regUVD_VCPU_CNTL), tmp, 0, indirect);
+
+	/* disable master interrupt */
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, 0, regUVD_MASTINT_EN), 0, 0, indirect);
+
+	/* setup regUVD_LMI_CTRL */
+	tmp = (UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK |
+		UVD_LMI_CTRL__REQ_MODE_MASK |
+		UVD_LMI_CTRL__CRC_RESET_MASK |
+		UVD_LMI_CTRL__MASK_MC_URGENT_MASK |
+		UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK |
+		UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK |
+		(8 << UVD_LMI_CTRL__WRITE_CLEAN_TIMER__SHIFT) |
+		0x00100000L);
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, 0, regUVD_LMI_CTRL), tmp, 0, indirect);
+
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, 0, regUVD_MPC_CNTL),
+		0x2 << UVD_MPC_CNTL__REPLACEMENT_MODE__SHIFT, 0, indirect);
+
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, 0, regUVD_MPC_SET_MUXA0),
+		((0x1 << UVD_MPC_SET_MUXA0__VARA_1__SHIFT) |
+		 (0x2 << UVD_MPC_SET_MUXA0__VARA_2__SHIFT) |
+		 (0x3 << UVD_MPC_SET_MUXA0__VARA_3__SHIFT) |
+		 (0x4 << UVD_MPC_SET_MUXA0__VARA_4__SHIFT)), 0, indirect);
+
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, 0, regUVD_MPC_SET_MUXB0),
+		 ((0x1 << UVD_MPC_SET_MUXB0__VARB_1__SHIFT) |
+		 (0x2 << UVD_MPC_SET_MUXB0__VARB_2__SHIFT) |
+		 (0x3 << UVD_MPC_SET_MUXB0__VARB_3__SHIFT) |
+		 (0x4 << UVD_MPC_SET_MUXB0__VARB_4__SHIFT)), 0, indirect);
+
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, 0, regUVD_MPC_SET_MUX),
+		((0x0 << UVD_MPC_SET_MUX__SET_0__SHIFT) |
+		 (0x1 << UVD_MPC_SET_MUX__SET_1__SHIFT) |
+		 (0x2 << UVD_MPC_SET_MUX__SET_2__SHIFT)), 0, indirect);
+
+	vcn_v4_0_3_mc_resume_dpg_mode(vinst, indirect);
+
+	tmp = (0xFF << UVD_VCPU_CNTL__PRB_TIMEOUT_VAL__SHIFT);
+	tmp |= UVD_VCPU_CNTL__CLK_EN_MASK;
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, 0, regUVD_VCPU_CNTL), tmp, 0, indirect);
+
+	/* enable LMI MC and UMC channels */
+	tmp = 0x1f << UVD_LMI_CTRL2__RE_OFLD_MIF_WR_REQ_NUM__SHIFT;
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, 0, regUVD_LMI_CTRL2), tmp, 0, indirect);
+
+	vcn_v4_0_3_enable_ras(adev, inst_idx, indirect);
+
+	/* enable master interrupt */
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, 0, regUVD_MASTINT_EN),
+		UVD_MASTINT_EN__VCPU_EN_MASK, 0, indirect);
+
+	if (indirect) {
+		ret = amdgpu_vcn_psp_update_sram(adev, inst_idx, AMDGPU_UCODE_ID_VCN0_RAM);
+		if (ret) {
+			dev_err(adev->dev, "vcn sram load failed %d\n", ret);
+			return ret;
+		}
+	}
+
+	ring = &adev->vcn.inst[inst_idx].ring_enc[0];
+
+	/* program the RB_BASE for ring buffer */
+	WREG32_SOC15(VCN, vcn_inst, regUVD_RB_BASE_LO,
+		     lower_32_bits(ring->gpu_addr));
+	WREG32_SOC15(VCN, vcn_inst, regUVD_RB_BASE_HI,
+		     upper_32_bits(ring->gpu_addr));
+
+	WREG32_SOC15(VCN, vcn_inst, regUVD_RB_SIZE,
+		     ring->ring_size / sizeof(uint32_t));
+
+	/* resetting ring, fw should not check RB ring */
+	tmp = RREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE);
+	tmp &= ~(VCN_RB_ENABLE__RB_EN_MASK);
+	WREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE, tmp);
+	fw_shared->sq.queue_mode |= FW_QUEUE_RING_RESET;
+
+	/* Initialize the ring buffer's read and write pointers */
+	WREG32_SOC15(VCN, vcn_inst, regUVD_RB_RPTR, 0);
+	WREG32_SOC15(VCN, vcn_inst, regUVD_RB_WPTR, 0);
+	ring->wptr = RREG32_SOC15(VCN, vcn_inst, regUVD_RB_WPTR);
+
+	tmp = RREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE);
+	tmp |= VCN_RB_ENABLE__RB_EN_MASK;
+	WREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE, tmp);
+	fw_shared->sq.queue_mode &= ~(FW_QUEUE_RING_RESET | FW_QUEUE_DPG_HOLD_OFF);
+
+	/*resetting done, fw can check RB ring */
+	fw_shared->sq.queue_mode &= cpu_to_le32(~FW_QUEUE_RING_RESET);
+
+	/* Keeping one read-back to ensure all register writes are done,
+	 * otherwise it may introduce race conditions.
+	 */
+	RREG32_SOC15(VCN, vcn_inst, regUVD_STATUS);
+
+	return 0;
+}
+
+static int vcn_v4_0_3_start_sriov(struct amdgpu_device *adev)
+{
+	int i, vcn_inst;
+	struct amdgpu_ring *ring_enc;
+	uint64_t cache_addr;
+	uint64_t rb_enc_addr;
+	uint64_t ctx_addr;
+	uint32_t param, resp, expected;
+	uint32_t offset, cache_size;
+	uint32_t tmp, timeout;
+
+	struct amdgpu_mm_table *table = &adev->virt.mm_table;
+	uint32_t *table_loc;
+	uint32_t table_size;
+	uint32_t size, size_dw;
+	uint32_t init_status;
+	uint32_t enabled_vcn;
+
+	struct mmsch_v4_0_cmd_direct_write
+		direct_wt = { {0} };
+	struct mmsch_v4_0_cmd_direct_read_modify_write
+		direct_rd_mod_wt = { {0} };
+	struct mmsch_v4_0_cmd_end end = { {0} };
+	struct mmsch_v4_0_3_init_header header;
+
+	struct amdgpu_vcn4_fw_shared *fw_shared;
+	struct amdgpu_fw_shared_rb_setup *rb_setup;
+
+	direct_wt.cmd_header.command_type =
+		MMSCH_COMMAND__DIRECT_REG_WRITE;
+	direct_rd_mod_wt.cmd_header.command_type =
+		MMSCH_COMMAND__DIRECT_REG_READ_MODIFY_WRITE;
+	end.cmd_header.command_type = MMSCH_COMMAND__END;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+		vcn_inst = GET_INST(VCN, i);
+
+		vcn_v4_0_3_fw_shared_init(adev, vcn_inst);
+
+		memset(&header, 0, sizeof(struct mmsch_v4_0_3_init_header));
+		header.version = MMSCH_VERSION;
+		header.total_size = sizeof(struct mmsch_v4_0_3_init_header) >> 2;
+
+		table_loc = (uint32_t *)table->cpu_addr;
+		table_loc += header.total_size;
+
+		table_size = 0;
+
+		MMSCH_V4_0_INSERT_DIRECT_RD_MOD_WT(SOC15_REG_OFFSET(VCN, 0, regUVD_STATUS),
+			~UVD_STATUS__UVD_BUSY, UVD_STATUS__UVD_BUSY);
+
+		cache_size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.inst[i].fw->size + 4);
+
+		if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
+			MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0,
+				regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW),
+				adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + i].tmr_mc_addr_lo);
+
+			MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0,
+				regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH),
+				adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + i].tmr_mc_addr_hi);
+
+			offset = 0;
+			MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0,
+				regUVD_VCPU_CACHE_OFFSET0), 0);
+		} else {
+			MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0,
+				regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW),
+				lower_32_bits(adev->vcn.inst[i].gpu_addr));
+			MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0,
+				regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH),
+				upper_32_bits(adev->vcn.inst[i].gpu_addr));
+			offset = cache_size;
+			MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0,
+				regUVD_VCPU_CACHE_OFFSET0),
+				AMDGPU_UVD_FIRMWARE_OFFSET >> 3);
+		}
+
+		MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0,
+			regUVD_VCPU_CACHE_SIZE0),
+			cache_size);
+
+		cache_addr = adev->vcn.inst[vcn_inst].gpu_addr + offset;
+		MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0,
+			regUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW), lower_32_bits(cache_addr));
+		MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0,
+			regUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH), upper_32_bits(cache_addr));
+		MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0,
+			regUVD_VCPU_CACHE_OFFSET1), 0);
+		MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0,
+			regUVD_VCPU_CACHE_SIZE1), AMDGPU_VCN_STACK_SIZE);
+
+		cache_addr = adev->vcn.inst[vcn_inst].gpu_addr + offset +
+			AMDGPU_VCN_STACK_SIZE;
+
+		MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0,
+			regUVD_LMI_VCPU_CACHE2_64BIT_BAR_LOW), lower_32_bits(cache_addr));
+
+		MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0,
+			regUVD_LMI_VCPU_CACHE2_64BIT_BAR_HIGH), upper_32_bits(cache_addr));
+
+		MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0,
+			regUVD_VCPU_CACHE_OFFSET2), 0);
+
+		MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0,
+			regUVD_VCPU_CACHE_SIZE2), AMDGPU_VCN_CONTEXT_SIZE);
+
+		fw_shared = adev->vcn.inst[vcn_inst].fw_shared.cpu_addr;
+		rb_setup = &fw_shared->rb_setup;
+
+		ring_enc = &adev->vcn.inst[vcn_inst].ring_enc[0];
+		ring_enc->wptr = 0;
+		rb_enc_addr = ring_enc->gpu_addr;
+
+		rb_setup->is_rb_enabled_flags |= RB_ENABLED;
+		rb_setup->rb_addr_lo = lower_32_bits(rb_enc_addr);
+		rb_setup->rb_addr_hi = upper_32_bits(rb_enc_addr);
+		rb_setup->rb_size = ring_enc->ring_size / 4;
+		fw_shared->present_flag_0 |= cpu_to_le32(AMDGPU_VCN_VF_RB_SETUP_FLAG);
+
+		MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0,
+			regUVD_LMI_VCPU_NC0_64BIT_BAR_LOW),
+			lower_32_bits(adev->vcn.inst[vcn_inst].fw_shared.gpu_addr));
+		MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0,
+			regUVD_LMI_VCPU_NC0_64BIT_BAR_HIGH),
+			upper_32_bits(adev->vcn.inst[vcn_inst].fw_shared.gpu_addr));
+		MMSCH_V4_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0,
+			regUVD_VCPU_NONCACHE_SIZE0),
+			AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_vcn4_fw_shared)));
+		MMSCH_V4_0_INSERT_END();
+
+		header.vcn0.init_status = 0;
+		header.vcn0.table_offset = header.total_size;
+		header.vcn0.table_size = table_size;
+		header.total_size += table_size;
+
+		/* Send init table to mmsch */
+		size = sizeof(struct mmsch_v4_0_3_init_header);
+		table_loc = (uint32_t *)table->cpu_addr;
+		memcpy((void *)table_loc, &header, size);
+
+		ctx_addr = table->gpu_addr;
+		WREG32_SOC15(VCN, vcn_inst, regMMSCH_VF_CTX_ADDR_LO, lower_32_bits(ctx_addr));
+		WREG32_SOC15(VCN, vcn_inst, regMMSCH_VF_CTX_ADDR_HI, upper_32_bits(ctx_addr));
+
+		tmp = RREG32_SOC15(VCN, vcn_inst, regMMSCH_VF_VMID);
+		tmp &= ~MMSCH_VF_VMID__VF_CTX_VMID_MASK;
+		tmp |= (0 << MMSCH_VF_VMID__VF_CTX_VMID__SHIFT);
+		WREG32_SOC15(VCN, vcn_inst, regMMSCH_VF_VMID, tmp);
+
+		size = header.total_size;
+		WREG32_SOC15(VCN, vcn_inst, regMMSCH_VF_CTX_SIZE, size);
+
+		WREG32_SOC15(VCN, vcn_inst, regMMSCH_VF_MAILBOX_RESP, 0);
+
+		param = 0x00000001;
+		WREG32_SOC15(VCN, vcn_inst, regMMSCH_VF_MAILBOX_HOST, param);
+		tmp = 0;
+		timeout = 1000;
+		resp = 0;
+		expected = MMSCH_VF_MAILBOX_RESP__OK;
+		while (resp != expected) {
+			resp = RREG32_SOC15(VCN, vcn_inst, regMMSCH_VF_MAILBOX_RESP);
+			if (resp != 0)
+				break;
+
+			udelay(10);
+			tmp = tmp + 10;
+			if (tmp >= timeout) {
+				DRM_ERROR("failed to init MMSCH. TIME-OUT after %d usec"\
+					" waiting for regMMSCH_VF_MAILBOX_RESP "\
+					"(expected=0x%08x, readback=0x%08x)\n",
+					tmp, expected, resp);
+				return -EBUSY;
+			}
+		}
+
+		enabled_vcn = amdgpu_vcn_is_disabled_vcn(adev, VCN_DECODE_RING, 0) ? 1 : 0;
+		init_status = ((struct mmsch_v4_0_3_init_header *)(table_loc))->vcn0.init_status;
+		if (resp != expected && resp != MMSCH_VF_MAILBOX_RESP__INCOMPLETE
+					&& init_status != MMSCH_VF_ENGINE_STATUS__PASS) {
+			DRM_ERROR("MMSCH init status is incorrect! readback=0x%08x, header init "\
+				"status for VCN%x: 0x%x\n", resp, enabled_vcn, init_status);
+		}
+	}
+
+	return 0;
+}
+
+/**
+ * vcn_v4_0_3_start - VCN start
+ *
+ * @vinst: VCN instance
+ *
+ * Start VCN block
+ */
+static int vcn_v4_0_3_start(struct amdgpu_vcn_inst *vinst)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int i = vinst->inst;
+	struct amdgpu_vcn4_fw_shared *fw_shared;
+	struct amdgpu_ring *ring;
+	int j, k, r, vcn_inst;
+	uint32_t tmp;
+
+	if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)
+		return vcn_v4_0_3_start_dpg_mode(vinst, adev->vcn.inst[i].indirect_sram);
+
+	vcn_inst = GET_INST(VCN, i);
+	/* set VCN status busy */
+	tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_STATUS) |
+		UVD_STATUS__UVD_BUSY;
+	WREG32_SOC15(VCN, vcn_inst, regUVD_STATUS, tmp);
+
+	/* SW clock gating */
+	vcn_v4_0_3_disable_clock_gating(vinst);
+
+	/* enable VCPU clock */
+	WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_VCPU_CNTL),
+		 UVD_VCPU_CNTL__CLK_EN_MASK,
+		 ~UVD_VCPU_CNTL__CLK_EN_MASK);
+
+	/* disable master interrupt */
+	WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_MASTINT_EN), 0,
+		 ~UVD_MASTINT_EN__VCPU_EN_MASK);
+
+	/* enable LMI MC and UMC channels */
+	WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_LMI_CTRL2), 0,
+		 ~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK);
+
+	tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET);
+	tmp &= ~UVD_SOFT_RESET__LMI_SOFT_RESET_MASK;
+	tmp &= ~UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK;
+	WREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET, tmp);
+
+	/* setup regUVD_LMI_CTRL */
+	tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_LMI_CTRL);
+	WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_CTRL,
+		     tmp | UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK |
+		     UVD_LMI_CTRL__MASK_MC_URGENT_MASK |
+		     UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK |
+		     UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK);
+
+	/* setup regUVD_MPC_CNTL */
+	tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_MPC_CNTL);
+	tmp &= ~UVD_MPC_CNTL__REPLACEMENT_MODE_MASK;
+	tmp |= 0x2 << UVD_MPC_CNTL__REPLACEMENT_MODE__SHIFT;
+	WREG32_SOC15(VCN, vcn_inst, regUVD_MPC_CNTL, tmp);
+
+	/* setup UVD_MPC_SET_MUXA0 */
+	WREG32_SOC15(VCN, vcn_inst, regUVD_MPC_SET_MUXA0,
+		     ((0x1 << UVD_MPC_SET_MUXA0__VARA_1__SHIFT) |
+		      (0x2 << UVD_MPC_SET_MUXA0__VARA_2__SHIFT) |
+		      (0x3 << UVD_MPC_SET_MUXA0__VARA_3__SHIFT) |
+		      (0x4 << UVD_MPC_SET_MUXA0__VARA_4__SHIFT)));
+
+	/* setup UVD_MPC_SET_MUXB0 */
+	WREG32_SOC15(VCN, vcn_inst, regUVD_MPC_SET_MUXB0,
+		     ((0x1 << UVD_MPC_SET_MUXB0__VARB_1__SHIFT) |
+		      (0x2 << UVD_MPC_SET_MUXB0__VARB_2__SHIFT) |
+		      (0x3 << UVD_MPC_SET_MUXB0__VARB_3__SHIFT) |
+		      (0x4 << UVD_MPC_SET_MUXB0__VARB_4__SHIFT)));
+
+	/* setup UVD_MPC_SET_MUX */
+	WREG32_SOC15(VCN, vcn_inst, regUVD_MPC_SET_MUX,
+		     ((0x0 << UVD_MPC_SET_MUX__SET_0__SHIFT) |
+		      (0x1 << UVD_MPC_SET_MUX__SET_1__SHIFT) |
+		      (0x2 << UVD_MPC_SET_MUX__SET_2__SHIFT)));
+
+	vcn_v4_0_3_mc_resume(vinst);
+
+	/* VCN global tiling registers */
+	WREG32_SOC15(VCN, vcn_inst, regUVD_GFX8_ADDR_CONFIG,
+		     adev->gfx.config.gb_addr_config);
+	WREG32_SOC15(VCN, vcn_inst, regUVD_GFX10_ADDR_CONFIG,
+		     adev->gfx.config.gb_addr_config);
+
+	/* unblock VCPU register access */
+	WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_RB_ARB_CTRL), 0,
+		 ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK);
+
+	/* release VCPU reset to boot */
+	WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_VCPU_CNTL), 0,
+		 ~UVD_VCPU_CNTL__BLK_RST_MASK);
+
+	for (j = 0; j < 10; ++j) {
+		uint32_t status;
+
+		for (k = 0; k < 100; ++k) {
+			status = RREG32_SOC15(VCN, vcn_inst,
+					      regUVD_STATUS);
+			if (status & 2)
+				break;
+			mdelay(10);
+		}
+		r = 0;
+		if (status & 2)
+			break;
+
+		DRM_DEV_ERROR(adev->dev,
+			      "VCN decode not responding, trying to reset the VCPU!!!\n");
+		WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst,
+					  regUVD_VCPU_CNTL),
+			 UVD_VCPU_CNTL__BLK_RST_MASK,
+			 ~UVD_VCPU_CNTL__BLK_RST_MASK);
+		mdelay(10);
+		WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst,
+					  regUVD_VCPU_CNTL),
+			 0, ~UVD_VCPU_CNTL__BLK_RST_MASK);
+
+		mdelay(10);
+		r = -1;
+	}
+
+	if (r) {
+		DRM_DEV_ERROR(adev->dev, "VCN decode not responding, giving up!!!\n");
+		return r;
+	}
+
+	/* enable master interrupt */
+	WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_MASTINT_EN),
+		 UVD_MASTINT_EN__VCPU_EN_MASK,
+		 ~UVD_MASTINT_EN__VCPU_EN_MASK);
+
+	/* clear the busy bit of VCN_STATUS */
+	WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_STATUS), 0,
+		 ~(2 << UVD_STATUS__VCPU_REPORT__SHIFT));
+
+	ring = &adev->vcn.inst[i].ring_enc[0];
+	fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr;
+
+	/* program the RB_BASE for ring buffer */
+	WREG32_SOC15(VCN, vcn_inst, regUVD_RB_BASE_LO,
+		     lower_32_bits(ring->gpu_addr));
+	WREG32_SOC15(VCN, vcn_inst, regUVD_RB_BASE_HI,
+		     upper_32_bits(ring->gpu_addr));
+
+	WREG32_SOC15(VCN, vcn_inst, regUVD_RB_SIZE,
+		     ring->ring_size / sizeof(uint32_t));
+
+	/* resetting ring, fw should not check RB ring */
+	tmp = RREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE);
+	tmp &= ~(VCN_RB_ENABLE__RB_EN_MASK);
+	WREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE, tmp);
+
+	/* Initialize the ring buffer's read and write pointers */
+	WREG32_SOC15(VCN, vcn_inst, regUVD_RB_RPTR, 0);
+	WREG32_SOC15(VCN, vcn_inst, regUVD_RB_WPTR, 0);
+
+	tmp = RREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE);
+	tmp |= VCN_RB_ENABLE__RB_EN_MASK;
+	WREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE, tmp);
+
+	ring->wptr = RREG32_SOC15(VCN, vcn_inst, regUVD_RB_WPTR);
+	fw_shared->sq.queue_mode &=
+		cpu_to_le32(~(FW_QUEUE_RING_RESET | FW_QUEUE_DPG_HOLD_OFF));
+
+	return 0;
+}
+
+/**
+ * vcn_v4_0_3_stop_dpg_mode - VCN stop with dpg mode
+ *
+ * @vinst: VCN instance
+ *
+ * Stop VCN block with dpg mode
+ */
+static int vcn_v4_0_3_stop_dpg_mode(struct amdgpu_vcn_inst *vinst)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int inst_idx = vinst->inst;
+	uint32_t tmp;
+	int vcn_inst;
+
+	vcn_inst = GET_INST(VCN, inst_idx);
+
+	/* Wait for power status to be 1 */
+	SOC15_WAIT_ON_RREG(VCN, vcn_inst, regUVD_POWER_STATUS, 1,
+			   UVD_POWER_STATUS__UVD_POWER_STATUS_MASK);
+
+	/* wait for read ptr to be equal to write ptr */
+	tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_RB_WPTR);
+	SOC15_WAIT_ON_RREG(VCN, vcn_inst, regUVD_RB_RPTR, tmp, 0xFFFFFFFF);
+
+	SOC15_WAIT_ON_RREG(VCN, vcn_inst, regUVD_POWER_STATUS, 1,
+			   UVD_POWER_STATUS__UVD_POWER_STATUS_MASK);
+
+	/* disable dynamic power gating mode */
+	WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_POWER_STATUS), 0,
+		 ~UVD_POWER_STATUS__UVD_PG_MODE_MASK);
+
+	/* Keeping one read-back to ensure all register writes are done,
+	 * otherwise it may introduce race conditions.
+	 */
+	RREG32_SOC15(VCN, vcn_inst, regUVD_STATUS);
+
+	return 0;
+}
+
+/**
+ * vcn_v4_0_3_stop - VCN stop
+ *
+ * @vinst: VCN instance
+ *
+ * Stop VCN block
+ */
+static int vcn_v4_0_3_stop(struct amdgpu_vcn_inst *vinst)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int i = vinst->inst;
+	struct amdgpu_vcn4_fw_shared *fw_shared;
+	int r = 0, vcn_inst;
+	uint32_t tmp;
+
+	vcn_inst = GET_INST(VCN, i);
+
+	fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr;
+	fw_shared->sq.queue_mode |= FW_QUEUE_DPG_HOLD_OFF;
+
+	if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) {
+		vcn_v4_0_3_stop_dpg_mode(vinst);
+		goto Done;
+	}
+
+	/* wait for vcn idle */
+	r = SOC15_WAIT_ON_RREG(VCN, vcn_inst, regUVD_STATUS,
+			       UVD_STATUS__IDLE, 0x7);
+	if (r)
+		goto Done;
+
+	tmp = UVD_LMI_STATUS__VCPU_LMI_WRITE_CLEAN_MASK |
+		UVD_LMI_STATUS__READ_CLEAN_MASK |
+		UVD_LMI_STATUS__WRITE_CLEAN_MASK |
+		UVD_LMI_STATUS__WRITE_CLEAN_RAW_MASK;
+	r = SOC15_WAIT_ON_RREG(VCN, vcn_inst, regUVD_LMI_STATUS, tmp,
+			       tmp);
+	if (r)
+		goto Done;
+
+	/* stall UMC channel */
+	tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_LMI_CTRL2);
+	tmp |= UVD_LMI_CTRL2__STALL_ARB_UMC_MASK;
+	WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_CTRL2, tmp);
+	tmp = UVD_LMI_STATUS__UMC_READ_CLEAN_RAW_MASK |
+		UVD_LMI_STATUS__UMC_WRITE_CLEAN_RAW_MASK;
+	r = SOC15_WAIT_ON_RREG(VCN, vcn_inst, regUVD_LMI_STATUS, tmp,
+			       tmp);
+	if (r)
+		goto Done;
+
+	/* Unblock VCPU Register access */
+	WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_RB_ARB_CTRL),
+		 UVD_RB_ARB_CTRL__VCPU_DIS_MASK,
+		 ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK);
+
+	/* release VCPU reset to boot */
+	WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_VCPU_CNTL),
+		 UVD_VCPU_CNTL__BLK_RST_MASK,
+		 ~UVD_VCPU_CNTL__BLK_RST_MASK);
+
+	/* disable VCPU clock */
+	WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_VCPU_CNTL), 0,
+		 ~(UVD_VCPU_CNTL__CLK_EN_MASK));
+
+	/* reset LMI UMC/LMI/VCPU */
+	tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET);
+	tmp |= UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK;
+	WREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET, tmp);
+
+	tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET);
+	tmp |= UVD_SOFT_RESET__LMI_SOFT_RESET_MASK;
+	WREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET, tmp);
+
+	/* clear VCN status */
+	WREG32_SOC15(VCN, vcn_inst, regUVD_STATUS, 0);
+
+	/* apply HW clock gating */
+	vcn_v4_0_3_enable_clock_gating(vinst);
+
+	/* Keeping one read-back to ensure all register writes are done,
+	 * otherwise it may introduce race conditions.
+	 */
+	RREG32_SOC15(VCN, vcn_inst, regUVD_STATUS);
+
+Done:
+	return 0;
+}
+
+/**
+ * vcn_v4_0_3_pause_dpg_mode - VCN pause with dpg mode
+ *
+ * @vinst: VCN instance
+ * @new_state: pause state
+ *
+ * Pause dpg mode for VCN block
+ */
+static int vcn_v4_0_3_pause_dpg_mode(struct amdgpu_vcn_inst *vinst,
+				     struct dpg_pause_state *new_state)
+{
+
+	return 0;
+}
+
+/**
+ * vcn_v4_0_3_unified_ring_get_rptr - get unified read pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Returns the current hardware unified read pointer
+ */
+static uint64_t vcn_v4_0_3_unified_ring_get_rptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	if (ring != &adev->vcn.inst[ring->me].ring_enc[0])
+		DRM_ERROR("wrong ring id is identified in %s", __func__);
+
+	return RREG32_SOC15(VCN, GET_INST(VCN, ring->me), regUVD_RB_RPTR);
+}
+
+/**
+ * vcn_v4_0_3_unified_ring_get_wptr - get unified write pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Returns the current hardware unified write pointer
+ */
+static uint64_t vcn_v4_0_3_unified_ring_get_wptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	if (ring != &adev->vcn.inst[ring->me].ring_enc[0])
+		DRM_ERROR("wrong ring id is identified in %s", __func__);
+
+	if (ring->use_doorbell)
+		return *ring->wptr_cpu_addr;
+	else
+		return RREG32_SOC15(VCN, GET_INST(VCN, ring->me),
+				    regUVD_RB_WPTR);
+}
+
+void vcn_v4_0_3_enc_ring_emit_reg_wait(struct amdgpu_ring *ring, uint32_t reg,
+				       uint32_t val, uint32_t mask)
+{
+	/* Use normalized offsets when required */
+	if (vcn_v4_0_3_normalizn_reqd(ring->adev))
+		reg = NORMALIZE_VCN_REG_OFFSET(reg);
+
+	amdgpu_ring_write(ring, VCN_ENC_CMD_REG_WAIT);
+	amdgpu_ring_write(ring, reg << 2);
+	amdgpu_ring_write(ring, mask);
+	amdgpu_ring_write(ring, val);
+}
+
+void vcn_v4_0_3_enc_ring_emit_wreg(struct amdgpu_ring *ring, uint32_t reg,
+				   uint32_t val)
+{
+	/* Use normalized offsets when required */
+	if (vcn_v4_0_3_normalizn_reqd(ring->adev))
+		reg = NORMALIZE_VCN_REG_OFFSET(reg);
+
+	amdgpu_ring_write(ring, VCN_ENC_CMD_REG_WRITE);
+	amdgpu_ring_write(ring,	reg << 2);
+	amdgpu_ring_write(ring, val);
+}
+
+void vcn_v4_0_3_enc_ring_emit_vm_flush(struct amdgpu_ring *ring,
+				       unsigned int vmid, uint64_t pd_addr)
+{
+	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->vm_hub];
+
+	pd_addr = amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pd_addr);
+
+	/* wait for reg writes */
+	vcn_v4_0_3_enc_ring_emit_reg_wait(ring, hub->ctx0_ptb_addr_lo32 +
+					vmid * hub->ctx_addr_distance,
+					lower_32_bits(pd_addr), 0xffffffff);
+}
+
+void vcn_v4_0_3_ring_emit_hdp_flush(struct amdgpu_ring *ring)
+{
+	/* VCN engine access for HDP flush doesn't work when RRMT is enabled.
+	 * This is a workaround to avoid any HDP flush through VCN ring.
+	 */
+}
+
+/**
+ * vcn_v4_0_3_unified_ring_set_wptr - set enc write pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Commits the enc write pointer to the hardware
+ */
+static void vcn_v4_0_3_unified_ring_set_wptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	if (ring != &adev->vcn.inst[ring->me].ring_enc[0])
+		DRM_ERROR("wrong ring id is identified in %s", __func__);
+
+	if (ring->use_doorbell) {
+		*ring->wptr_cpu_addr = lower_32_bits(ring->wptr);
+		WDOORBELL32(ring->doorbell_index, lower_32_bits(ring->wptr));
+	} else {
+		WREG32_SOC15(VCN, GET_INST(VCN, ring->me), regUVD_RB_WPTR,
+			     lower_32_bits(ring->wptr));
+	}
+}
+
+static int vcn_v4_0_3_ring_reset(struct amdgpu_ring *ring,
+				 unsigned int vmid,
+				 struct amdgpu_fence *timedout_fence)
+{
+	int r = 0;
+	int vcn_inst;
+	struct amdgpu_device *adev = ring->adev;
+	struct amdgpu_vcn_inst *vinst = &adev->vcn.inst[ring->me];
+
+	amdgpu_ring_reset_helper_begin(ring, timedout_fence);
+
+	vcn_inst = GET_INST(VCN, ring->me);
+	r = amdgpu_dpm_reset_vcn(adev, 1 << vcn_inst);
+
+	if (r) {
+		DRM_DEV_ERROR(adev->dev, "VCN reset fail : %d\n", r);
+		return r;
+	}
+
+	/* This flag is not set for VF, assumed to be disabled always */
+	if (RREG32_SOC15(VCN, GET_INST(VCN, 0), regVCN_RRMT_CNTL) & 0x100)
+		adev->vcn.caps |= AMDGPU_VCN_CAPS(RRMT_ENABLED);
+	vcn_v4_0_3_hw_init_inst(vinst);
+	vcn_v4_0_3_start_dpg_mode(vinst, adev->vcn.inst[ring->me].indirect_sram);
+
+	return amdgpu_ring_reset_helper_end(ring, timedout_fence);
+}
+
+static const struct amdgpu_ring_funcs vcn_v4_0_3_unified_ring_vm_funcs = {
+	.type = AMDGPU_RING_TYPE_VCN_ENC,
+	.align_mask = 0x3f,
+	.nop = VCN_ENC_CMD_NO_OP,
+	.get_rptr = vcn_v4_0_3_unified_ring_get_rptr,
+	.get_wptr = vcn_v4_0_3_unified_ring_get_wptr,
+	.set_wptr = vcn_v4_0_3_unified_ring_set_wptr,
+	.emit_frame_size =
+		SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 +
+		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 4 +
+		4 + /* vcn_v2_0_enc_ring_emit_vm_flush */
+		5 + 5 + /* vcn_v2_0_enc_ring_emit_fence x2 vm fence */
+		1, /* vcn_v2_0_enc_ring_insert_end */
+	.emit_ib_size = 5, /* vcn_v2_0_enc_ring_emit_ib */
+	.emit_ib = vcn_v2_0_enc_ring_emit_ib,
+	.emit_fence = vcn_v2_0_enc_ring_emit_fence,
+	.emit_vm_flush = vcn_v4_0_3_enc_ring_emit_vm_flush,
+	.emit_hdp_flush = vcn_v4_0_3_ring_emit_hdp_flush,
+	.test_ring = amdgpu_vcn_enc_ring_test_ring,
+	.test_ib = amdgpu_vcn_unified_ring_test_ib,
+	.insert_nop = amdgpu_ring_insert_nop,
+	.insert_end = vcn_v2_0_enc_ring_insert_end,
+	.pad_ib = amdgpu_ring_generic_pad_ib,
+	.begin_use = amdgpu_vcn_ring_begin_use,
+	.end_use = amdgpu_vcn_ring_end_use,
+	.emit_wreg = vcn_v4_0_3_enc_ring_emit_wreg,
+	.emit_reg_wait = vcn_v4_0_3_enc_ring_emit_reg_wait,
+	.emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper,
+	.reset = vcn_v4_0_3_ring_reset,
+};
+
+/**
+ * vcn_v4_0_3_set_unified_ring_funcs - set unified ring functions
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Set unified ring functions
+ */
+static void vcn_v4_0_3_set_unified_ring_funcs(struct amdgpu_device *adev)
+{
+	int i, vcn_inst;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		adev->vcn.inst[i].ring_enc[0].funcs = &vcn_v4_0_3_unified_ring_vm_funcs;
+		adev->vcn.inst[i].ring_enc[0].me = i;
+		vcn_inst = GET_INST(VCN, i);
+		adev->vcn.inst[i].aid_id =
+			vcn_inst / adev->vcn.num_inst_per_aid;
+	}
+}
+
+/**
+ * vcn_v4_0_3_is_idle - check VCN block is idle
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block structure
+ *
+ * Check whether VCN block is idle
+ */
+static bool vcn_v4_0_3_is_idle(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, ret = 1;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		ret &= (RREG32_SOC15(VCN, GET_INST(VCN, i), regUVD_STATUS) ==
+			UVD_STATUS__IDLE);
+	}
+
+	return ret;
+}
+
+/**
+ * vcn_v4_0_3_wait_for_idle - wait for VCN block idle
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Wait for VCN block idle
+ */
+static int vcn_v4_0_3_wait_for_idle(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, ret = 0;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		ret = SOC15_WAIT_ON_RREG(VCN, GET_INST(VCN, i), regUVD_STATUS,
+					 UVD_STATUS__IDLE, UVD_STATUS__IDLE);
+		if (ret)
+			return ret;
+	}
+
+	return ret;
+}
+
+/* vcn_v4_0_3_set_clockgating_state - set VCN block clockgating state
+ *
+ * @ip_block: amdgpu_ip_block pointer
+ * @state: clock gating state
+ *
+ * Set VCN block clockgating state
+ */
+static int vcn_v4_0_3_set_clockgating_state(struct amdgpu_ip_block *ip_block,
+					  enum amd_clockgating_state state)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	bool enable = state == AMD_CG_STATE_GATE;
+	int i;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		struct amdgpu_vcn_inst *vinst = &adev->vcn.inst[i];
+
+		if (enable) {
+			if (RREG32_SOC15(VCN, GET_INST(VCN, i),
+					 regUVD_STATUS) != UVD_STATUS__IDLE)
+				return -EBUSY;
+			vcn_v4_0_3_enable_clock_gating(vinst);
+		} else {
+			vcn_v4_0_3_disable_clock_gating(vinst);
+		}
+	}
+	return 0;
+}
+
+static int vcn_v4_0_3_set_pg_state(struct amdgpu_vcn_inst *vinst,
+				   enum amd_powergating_state state)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int ret = 0;
+
+	/* for SRIOV, guest should not control VCN Power-gating
+	 * MMSCH FW should control Power-gating and clock-gating
+	 * guest should avoid touching CGC and PG
+	 */
+	if (amdgpu_sriov_vf(adev)) {
+		vinst->cur_state = AMD_PG_STATE_UNGATE;
+		return 0;
+	}
+
+	if (state == vinst->cur_state)
+		return 0;
+
+	if (state == AMD_PG_STATE_GATE)
+		ret = vcn_v4_0_3_stop(vinst);
+	else
+		ret = vcn_v4_0_3_start(vinst);
+
+	if (!ret)
+		vinst->cur_state = state;
+
+	return ret;
+}
+
+/**
+ * vcn_v4_0_3_set_interrupt_state - set VCN block interrupt state
+ *
+ * @adev: amdgpu_device pointer
+ * @source: interrupt sources
+ * @type: interrupt types
+ * @state: interrupt states
+ *
+ * Set VCN block interrupt state
+ */
+static int vcn_v4_0_3_set_interrupt_state(struct amdgpu_device *adev,
+					struct amdgpu_irq_src *source,
+					unsigned int type,
+					enum amdgpu_interrupt_state state)
+{
+	return 0;
+}
+
+/**
+ * vcn_v4_0_3_process_interrupt - process VCN block interrupt
+ *
+ * @adev: amdgpu_device pointer
+ * @source: interrupt sources
+ * @entry: interrupt entry from clients and sources
+ *
+ * Process VCN block interrupt
+ */
+static int vcn_v4_0_3_process_interrupt(struct amdgpu_device *adev,
+				      struct amdgpu_irq_src *source,
+				      struct amdgpu_iv_entry *entry)
+{
+	uint32_t i, inst;
+
+	i = node_id_to_phys_map[entry->node_id];
+
+	DRM_DEV_DEBUG(adev->dev, "IH: VCN TRAP\n");
+
+	for (inst = 0; inst < adev->vcn.num_vcn_inst; ++inst)
+		if (adev->vcn.inst[inst].aid_id == i)
+			break;
+
+	if (inst >= adev->vcn.num_vcn_inst) {
+		dev_WARN_ONCE(adev->dev, 1,
+			      "Interrupt received for unknown VCN instance %d",
+			      entry->node_id);
+		return 0;
+	}
+
+	switch (entry->src_id) {
+	case VCN_4_0__SRCID__UVD_ENC_GENERAL_PURPOSE:
+		amdgpu_fence_process(&adev->vcn.inst[inst].ring_enc[0]);
+		break;
+	default:
+		DRM_DEV_ERROR(adev->dev, "Unhandled interrupt: %d %d\n",
+			  entry->src_id, entry->src_data[0]);
+		break;
+	}
+
+	return 0;
+}
+
+static int vcn_v4_0_3_set_ras_interrupt_state(struct amdgpu_device *adev,
+					struct amdgpu_irq_src *source,
+					unsigned int type,
+					enum amdgpu_interrupt_state state)
+{
+	return 0;
+}
+
+static const struct amdgpu_irq_src_funcs vcn_v4_0_3_irq_funcs = {
+	.set = vcn_v4_0_3_set_interrupt_state,
+	.process = vcn_v4_0_3_process_interrupt,
+};
+
+static const struct amdgpu_irq_src_funcs vcn_v4_0_3_ras_irq_funcs = {
+	.set = vcn_v4_0_3_set_ras_interrupt_state,
+	.process = amdgpu_vcn_process_poison_irq,
+};
+
+/**
+ * vcn_v4_0_3_set_irq_funcs - set VCN block interrupt irq functions
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Set VCN block interrupt irq functions
+ */
+static void vcn_v4_0_3_set_irq_funcs(struct amdgpu_device *adev)
+{
+	int i;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		adev->vcn.inst->irq.num_types++;
+	}
+	adev->vcn.inst->irq.funcs = &vcn_v4_0_3_irq_funcs;
+
+	adev->vcn.inst->ras_poison_irq.num_types = 1;
+	adev->vcn.inst->ras_poison_irq.funcs = &vcn_v4_0_3_ras_irq_funcs;
+}
+
+static const struct amd_ip_funcs vcn_v4_0_3_ip_funcs = {
+	.name = "vcn_v4_0_3",
+	.early_init = vcn_v4_0_3_early_init,
+	.late_init = vcn_v4_0_3_late_init,
+	.sw_init = vcn_v4_0_3_sw_init,
+	.sw_fini = vcn_v4_0_3_sw_fini,
+	.hw_init = vcn_v4_0_3_hw_init,
+	.hw_fini = vcn_v4_0_3_hw_fini,
+	.suspend = vcn_v4_0_3_suspend,
+	.resume = vcn_v4_0_3_resume,
+	.is_idle = vcn_v4_0_3_is_idle,
+	.wait_for_idle = vcn_v4_0_3_wait_for_idle,
+	.set_clockgating_state = vcn_v4_0_3_set_clockgating_state,
+	.set_powergating_state = vcn_set_powergating_state,
+	.dump_ip_state = amdgpu_vcn_dump_ip_state,
+	.print_ip_state = amdgpu_vcn_print_ip_state,
+};
+
+const struct amdgpu_ip_block_version vcn_v4_0_3_ip_block = {
+	.type = AMD_IP_BLOCK_TYPE_VCN,
+	.major = 4,
+	.minor = 0,
+	.rev = 3,
+	.funcs = &vcn_v4_0_3_ip_funcs,
+};
+
+static const struct amdgpu_ras_err_status_reg_entry vcn_v4_0_3_ue_reg_list[] = {
+	{AMDGPU_RAS_REG_ENTRY(VCN, 0, regVCN_UE_ERR_STATUS_LO_VIDD, regVCN_UE_ERR_STATUS_HI_VIDD),
+	1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "VIDD"},
+	{AMDGPU_RAS_REG_ENTRY(VCN, 0, regVCN_UE_ERR_STATUS_LO_VIDV, regVCN_UE_ERR_STATUS_HI_VIDV),
+	1, (AMDGPU_RAS_ERR_INFO_VALID | AMDGPU_RAS_ERR_STATUS_VALID), "VIDV"},
+};
+
+static void vcn_v4_0_3_inst_query_ras_error_count(struct amdgpu_device *adev,
+						  uint32_t vcn_inst,
+						  void *ras_err_status)
+{
+	struct ras_err_data *err_data = (struct ras_err_data *)ras_err_status;
+
+	/* vcn v4_0_3 only support query uncorrectable errors */
+	amdgpu_ras_inst_query_ras_error_count(adev,
+			vcn_v4_0_3_ue_reg_list,
+			ARRAY_SIZE(vcn_v4_0_3_ue_reg_list),
+			NULL, 0, GET_INST(VCN, vcn_inst),
+			AMDGPU_RAS_ERROR__MULTI_UNCORRECTABLE,
+			&err_data->ue_count);
+}
+
+static void vcn_v4_0_3_query_ras_error_count(struct amdgpu_device *adev,
+					     void *ras_err_status)
+{
+	uint32_t i;
+
+	if (!amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__VCN)) {
+		dev_warn(adev->dev, "VCN RAS is not supported\n");
+		return;
+	}
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++)
+		vcn_v4_0_3_inst_query_ras_error_count(adev, i, ras_err_status);
+}
+
+static void vcn_v4_0_3_inst_reset_ras_error_count(struct amdgpu_device *adev,
+						  uint32_t vcn_inst)
+{
+	amdgpu_ras_inst_reset_ras_error_count(adev,
+					vcn_v4_0_3_ue_reg_list,
+					ARRAY_SIZE(vcn_v4_0_3_ue_reg_list),
+					GET_INST(VCN, vcn_inst));
+}
+
+static void vcn_v4_0_3_reset_ras_error_count(struct amdgpu_device *adev)
+{
+	uint32_t i;
+
+	if (!amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__VCN)) {
+		dev_warn(adev->dev, "VCN RAS is not supported\n");
+		return;
+	}
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++)
+		vcn_v4_0_3_inst_reset_ras_error_count(adev, i);
+}
+
+static uint32_t vcn_v4_0_3_query_poison_by_instance(struct amdgpu_device *adev,
+			uint32_t instance, uint32_t sub_block)
+{
+	uint32_t poison_stat = 0, reg_value = 0;
+
+	switch (sub_block) {
+	case AMDGPU_VCN_V4_0_3_VCPU_VCODEC:
+		reg_value = RREG32_SOC15(VCN, instance, regUVD_RAS_VCPU_VCODEC_STATUS);
+		poison_stat = REG_GET_FIELD(reg_value, UVD_RAS_VCPU_VCODEC_STATUS, POISONED_PF);
+		break;
+	default:
+		break;
+	}
+
+	if (poison_stat)
+		dev_info(adev->dev, "Poison detected in VCN%d, sub_block%d\n",
+			instance, sub_block);
+
+	return poison_stat;
+}
+
+static bool vcn_v4_0_3_query_poison_status(struct amdgpu_device *adev)
+{
+	uint32_t inst, sub;
+	uint32_t poison_stat = 0;
+
+	for (inst = 0; inst < adev->vcn.num_vcn_inst; inst++)
+		for (sub = 0; sub < AMDGPU_VCN_V4_0_3_MAX_SUB_BLOCK; sub++)
+			poison_stat +=
+			vcn_v4_0_3_query_poison_by_instance(adev, inst, sub);
+
+	return !!poison_stat;
+}
+
+static const struct amdgpu_ras_block_hw_ops vcn_v4_0_3_ras_hw_ops = {
+	.query_ras_error_count = vcn_v4_0_3_query_ras_error_count,
+	.reset_ras_error_count = vcn_v4_0_3_reset_ras_error_count,
+	.query_poison_status = vcn_v4_0_3_query_poison_status,
+};
+
+static int vcn_v4_0_3_aca_bank_parser(struct aca_handle *handle, struct aca_bank *bank,
+				      enum aca_smu_type type, void *data)
+{
+	struct aca_bank_info info;
+	u64 misc0;
+	int ret;
+
+	ret = aca_bank_info_decode(bank, &info);
+	if (ret)
+		return ret;
+
+	misc0 = bank->regs[ACA_REG_IDX_MISC0];
+	switch (type) {
+	case ACA_SMU_TYPE_UE:
+		bank->aca_err_type = ACA_ERROR_TYPE_UE;
+		ret = aca_error_cache_log_bank_error(handle, &info, ACA_ERROR_TYPE_UE,
+						     1ULL);
+		break;
+	case ACA_SMU_TYPE_CE:
+		bank->aca_err_type = ACA_ERROR_TYPE_CE;
+		ret = aca_error_cache_log_bank_error(handle, &info, bank->aca_err_type,
+						     ACA_REG__MISC0__ERRCNT(misc0));
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return ret;
+}
+
+/* reference to smu driver if header file */
+static int vcn_v4_0_3_err_codes[] = {
+	14, 15, /* VCN */
+};
+
+static bool vcn_v4_0_3_aca_bank_is_valid(struct aca_handle *handle, struct aca_bank *bank,
+					 enum aca_smu_type type, void *data)
+{
+	u32 instlo;
+
+	instlo = ACA_REG__IPID__INSTANCEIDLO(bank->regs[ACA_REG_IDX_IPID]);
+	instlo &= GENMASK(31, 1);
+
+	if (instlo != mmSMNAID_AID0_MCA_SMU)
+		return false;
+
+	if (aca_bank_check_error_codes(handle->adev, bank,
+				       vcn_v4_0_3_err_codes,
+				       ARRAY_SIZE(vcn_v4_0_3_err_codes)))
+		return false;
+
+	return true;
+}
+
+static const struct aca_bank_ops vcn_v4_0_3_aca_bank_ops = {
+	.aca_bank_parser = vcn_v4_0_3_aca_bank_parser,
+	.aca_bank_is_valid = vcn_v4_0_3_aca_bank_is_valid,
+};
+
+static const struct aca_info vcn_v4_0_3_aca_info = {
+	.hwip = ACA_HWIP_TYPE_SMU,
+	.mask = ACA_ERROR_UE_MASK,
+	.bank_ops = &vcn_v4_0_3_aca_bank_ops,
+};
+
+static int vcn_v4_0_3_ras_late_init(struct amdgpu_device *adev, struct ras_common_if *ras_block)
+{
+	int r;
+
+	r = amdgpu_ras_block_late_init(adev, ras_block);
+	if (r)
+		return r;
+
+	if (amdgpu_ras_is_supported(adev, ras_block->block) &&
+		adev->vcn.inst->ras_poison_irq.funcs) {
+		r = amdgpu_irq_get(adev, &adev->vcn.inst->ras_poison_irq, 0);
+		if (r)
+			goto late_fini;
+	}
+
+	r = amdgpu_ras_bind_aca(adev, AMDGPU_RAS_BLOCK__VCN,
+				&vcn_v4_0_3_aca_info, NULL);
+	if (r)
+		goto late_fini;
+
+	return 0;
+
+late_fini:
+	amdgpu_ras_block_late_fini(adev, ras_block);
+
+	return r;
+}
+
+static struct amdgpu_vcn_ras vcn_v4_0_3_ras = {
+	.ras_block = {
+		.hw_ops = &vcn_v4_0_3_ras_hw_ops,
+		.ras_late_init = vcn_v4_0_3_ras_late_init,
+	},
+};
+
+static void vcn_v4_0_3_set_ras_funcs(struct amdgpu_device *adev)
+{
+	adev->vcn.ras = &vcn_v4_0_3_ras;
+}
+
+static void vcn_v4_0_3_enable_ras(struct amdgpu_device *adev,
+				  int inst_idx, bool indirect)
+{
+	uint32_t tmp;
+
+	if (!amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__VCN))
+		return;
+
+	tmp = VCN_RAS_CNTL__VCPU_VCODEC_REARM_MASK |
+	      VCN_RAS_CNTL__VCPU_VCODEC_IH_EN_MASK |
+	      VCN_RAS_CNTL__VCPU_VCODEC_PMI_EN_MASK |
+	      VCN_RAS_CNTL__VCPU_VCODEC_STALL_EN_MASK;
+	WREG32_SOC15_DPG_MODE(inst_idx,
+			      SOC15_DPG_MODE_OFFSET(VCN, 0, regVCN_RAS_CNTL),
+			      tmp, 0, indirect);
+
+	tmp = UVD_VCPU_INT_EN2__RASCNTL_VCPU_VCODEC_EN_MASK;
+	WREG32_SOC15_DPG_MODE(inst_idx,
+			      SOC15_DPG_MODE_OFFSET(VCN, 0, regUVD_VCPU_INT_EN2),
+			      tmp, 0, indirect);
+
+	tmp = UVD_SYS_INT_EN__RASCNTL_VCPU_VCODEC_EN_MASK;
+	WREG32_SOC15_DPG_MODE(inst_idx,
+			      SOC15_DPG_MODE_OFFSET(VCN, 0, regUVD_SYS_INT_EN),
+			      tmp, 0, indirect);
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.h b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.h
new file mode 100644
index 000000000000..aeab89853a92
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.h
@@ -0,0 +1,44 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __VCN_V4_0_3_H__
+#define __VCN_V4_0_3_H__
+
+enum amdgpu_vcn_v4_0_3_sub_block {
+	AMDGPU_VCN_V4_0_3_VCPU_VCODEC = 0,
+
+	AMDGPU_VCN_V4_0_3_MAX_SUB_BLOCK,
+};
+
+extern const struct amdgpu_ip_block_version vcn_v4_0_3_ip_block;
+
+void vcn_v4_0_3_enc_ring_emit_reg_wait(struct amdgpu_ring *ring, uint32_t reg,
+				       uint32_t val, uint32_t mask);
+
+void vcn_v4_0_3_enc_ring_emit_wreg(struct amdgpu_ring *ring, uint32_t reg,
+				   uint32_t val);
+void vcn_v4_0_3_enc_ring_emit_vm_flush(struct amdgpu_ring *ring,
+				       unsigned int vmid, uint64_t pd_addr);
+void vcn_v4_0_3_ring_emit_hdp_flush(struct amdgpu_ring *ring);
+
+#endif /* __VCN_V4_0_3_H__ */
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c
new file mode 100644
index 000000000000..b107ee80e472
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c
@@ -0,0 +1,1723 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/firmware.h>
+#include "amdgpu.h"
+#include "amdgpu_vcn.h"
+#include "amdgpu_pm.h"
+#include "amdgpu_cs.h"
+#include "soc15.h"
+#include "soc15d.h"
+#include "soc15_hw_ip.h"
+#include "vcn_v2_0.h"
+#include "mmsch_v4_0.h"
+#include "vcn_v4_0_5.h"
+
+#include "vcn/vcn_4_0_5_offset.h"
+#include "vcn/vcn_4_0_5_sh_mask.h"
+#include "ivsrcid/vcn/irqsrcs_vcn_4_0.h"
+
+#include <drm/drm_drv.h>
+
+#define mmUVD_DPG_LMA_CTL							regUVD_DPG_LMA_CTL
+#define mmUVD_DPG_LMA_CTL_BASE_IDX					regUVD_DPG_LMA_CTL_BASE_IDX
+#define mmUVD_DPG_LMA_DATA							regUVD_DPG_LMA_DATA
+#define mmUVD_DPG_LMA_DATA_BASE_IDX					regUVD_DPG_LMA_DATA_BASE_IDX
+
+#define VCN_VID_SOC_ADDRESS_2_0						0x1fb00
+#define VCN1_VID_SOC_ADDRESS_3_0					(0x48300 + 0x38000)
+#define VCN1_AON_SOC_ADDRESS_3_0					(0x48000 + 0x38000)
+
+#define VCN_HARVEST_MMSCH							0
+
+#define RDECODE_MSG_CREATE							0x00000000
+#define RDECODE_MESSAGE_CREATE						0x00000001
+
+static const struct amdgpu_hwip_reg_entry vcn_reg_list_4_0_5[] = {
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_POWER_STATUS),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_STATUS),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_CONTEXT_ID),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_CONTEXT_ID2),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_GPCOM_VCPU_DATA0),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_GPCOM_VCPU_DATA1),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_GPCOM_VCPU_CMD),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_BASE_HI),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_BASE_LO),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_BASE_HI2),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_BASE_LO2),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_BASE_HI3),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_BASE_LO3),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_BASE_HI4),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_BASE_LO4),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_RPTR),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_WPTR),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_RPTR2),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_WPTR2),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_RPTR3),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_WPTR3),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_RPTR4),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_WPTR4),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_SIZE),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_SIZE2),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_SIZE3),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_SIZE4),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_PGFSM_CONFIG),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_PGFSM_STATUS),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_DPG_LMA_CTL),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_DPG_LMA_DATA),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_DPG_LMA_MASK),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_DPG_PAUSE)
+};
+
+static int amdgpu_ih_clientid_vcns[] = {
+	SOC15_IH_CLIENTID_VCN,
+	SOC15_IH_CLIENTID_VCN1
+};
+
+static void vcn_v4_0_5_set_unified_ring_funcs(struct amdgpu_device *adev);
+static void vcn_v4_0_5_set_irq_funcs(struct amdgpu_device *adev);
+static int vcn_v4_0_5_set_pg_state(struct amdgpu_vcn_inst *vinst,
+				   enum amd_powergating_state state);
+static int vcn_v4_0_5_pause_dpg_mode(struct amdgpu_vcn_inst *vinst,
+				     struct dpg_pause_state *new_state);
+static void vcn_v4_0_5_unified_ring_set_wptr(struct amdgpu_ring *ring);
+
+/**
+ * vcn_v4_0_5_early_init - set function pointers and load microcode
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Set ring and irq function pointers
+ * Load microcode from filesystem
+ */
+static int vcn_v4_0_5_early_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, r;
+
+	if (amdgpu_ip_version(adev, UVD_HWIP, 0) == IP_VERSION(4, 0, 6))
+		adev->vcn.per_inst_fw = true;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i)
+		/* re-use enc ring as unified ring */
+		adev->vcn.inst[i].num_enc_rings = 1;
+	vcn_v4_0_5_set_unified_ring_funcs(adev);
+	vcn_v4_0_5_set_irq_funcs(adev);
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		adev->vcn.inst[i].set_pg_state = vcn_v4_0_5_set_pg_state;
+
+		r = amdgpu_vcn_early_init(adev, i);
+		if (r)
+			return r;
+	}
+
+	return 0;
+}
+
+/**
+ * vcn_v4_0_5_sw_init - sw init for VCN block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Load firmware and sw initialization
+ */
+static int vcn_v4_0_5_sw_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_ring *ring;
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, r;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+		struct amdgpu_vcn4_fw_shared *fw_shared;
+
+		if (adev->vcn.harvest_config & (1 << i))
+			continue;
+
+		r = amdgpu_vcn_sw_init(adev, i);
+		if (r)
+			return r;
+
+		amdgpu_vcn_setup_ucode(adev, i);
+
+		r = amdgpu_vcn_resume(adev, i);
+		if (r)
+			return r;
+
+		atomic_set(&adev->vcn.inst[i].sched_score, 0);
+
+		/* VCN UNIFIED TRAP */
+		r = amdgpu_irq_add_id(adev, amdgpu_ih_clientid_vcns[i],
+				VCN_4_0__SRCID__UVD_ENC_GENERAL_PURPOSE, &adev->vcn.inst[i].irq);
+		if (r)
+			return r;
+
+		/* VCN POISON TRAP */
+		r = amdgpu_irq_add_id(adev, amdgpu_ih_clientid_vcns[i],
+				VCN_4_0__SRCID_UVD_POISON, &adev->vcn.inst[i].irq);
+		if (r)
+			return r;
+
+		ring = &adev->vcn.inst[i].ring_enc[0];
+		ring->use_doorbell = true;
+		if (amdgpu_sriov_vf(adev))
+			ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) +
+						i * (adev->vcn.inst[i].num_enc_rings + 1) + 1;
+		else
+			ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) +
+						2 + 8 * i;
+		ring->vm_hub = AMDGPU_MMHUB0(0);
+		sprintf(ring->name, "vcn_unified_%d", i);
+
+		r = amdgpu_ring_init(adev, ring, 512, &adev->vcn.inst[i].irq, 0,
+				AMDGPU_RING_PRIO_0, &adev->vcn.inst[i].sched_score);
+		if (r)
+			return r;
+
+		fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr;
+		fw_shared->present_flag_0 = cpu_to_le32(AMDGPU_FW_SHARED_FLAG_0_UNIFIED_QUEUE);
+		fw_shared->sq.is_enabled = 1;
+
+		fw_shared->present_flag_0 |= cpu_to_le32(AMDGPU_VCN_SMU_DPM_INTERFACE_FLAG);
+		fw_shared->smu_dpm_interface.smu_interface_type = (adev->flags & AMD_IS_APU) ?
+			AMDGPU_VCN_SMU_DPM_INTERFACE_APU : AMDGPU_VCN_SMU_DPM_INTERFACE_DGPU;
+
+		if (amdgpu_sriov_vf(adev))
+			fw_shared->present_flag_0 |= cpu_to_le32(AMDGPU_VCN_VF_RB_SETUP_FLAG);
+
+		fw_shared->present_flag_0 |= AMDGPU_FW_SHARED_FLAG_0_DRM_KEY_INJECT;
+		fw_shared->drm_key_wa.method =
+			AMDGPU_DRM_KEY_INJECT_WORKAROUND_VCNFW_ASD_HANDSHAKING;
+
+		if (amdgpu_vcnfw_log)
+			amdgpu_vcn_fwlog_init(&adev->vcn.inst[i]);
+
+		if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)
+			adev->vcn.inst[i].pause_dpg_mode = vcn_v4_0_5_pause_dpg_mode;
+	}
+
+	adev->vcn.supported_reset = amdgpu_get_soft_full_reset_mask(&adev->vcn.inst[0].ring_enc[0]);
+	if (!amdgpu_sriov_vf(adev))
+		adev->vcn.supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE;
+
+	r = amdgpu_vcn_sysfs_reset_mask_init(adev);
+	if (r)
+		return r;
+
+	if (amdgpu_sriov_vf(adev)) {
+		r = amdgpu_virt_alloc_mm_table(adev);
+		if (r)
+			return r;
+	}
+
+	r = amdgpu_vcn_reg_dump_init(adev, vcn_reg_list_4_0_5, ARRAY_SIZE(vcn_reg_list_4_0_5));
+
+	return r;
+}
+
+/**
+ * vcn_v4_0_5_sw_fini - sw fini for VCN block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * VCN suspend and free up sw allocation
+ */
+static int vcn_v4_0_5_sw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, r, idx;
+
+	if (drm_dev_enter(adev_to_drm(adev), &idx)) {
+		for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+			struct amdgpu_vcn4_fw_shared *fw_shared;
+
+			if (adev->vcn.harvest_config & (1 << i))
+				continue;
+
+			fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr;
+			fw_shared->present_flag_0 = 0;
+			fw_shared->sq.is_enabled = 0;
+		}
+
+		drm_dev_exit(idx);
+	}
+
+	if (amdgpu_sriov_vf(adev))
+		amdgpu_virt_free_mm_table(adev);
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+		r = amdgpu_vcn_suspend(adev, i);
+		if (r)
+			return r;
+
+		amdgpu_vcn_sw_fini(adev, i);
+	}
+
+	return 0;
+}
+
+/**
+ * vcn_v4_0_5_hw_init - start and test VCN block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Initialize the hardware, boot up the VCPU and do some testing
+ */
+static int vcn_v4_0_5_hw_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	struct amdgpu_ring *ring;
+	int i, r;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		if (adev->vcn.harvest_config & (1 << i))
+			continue;
+
+		ring = &adev->vcn.inst[i].ring_enc[0];
+
+		adev->nbio.funcs->vcn_doorbell_range(adev, ring->use_doorbell,
+				((adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 8 * i), i);
+
+		r = amdgpu_ring_test_helper(ring);
+		if (r)
+			return r;
+	}
+
+	return 0;
+}
+
+/**
+ * vcn_v4_0_5_hw_fini - stop the hardware block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Stop the VCN block, mark ring as not ready any more
+ */
+static int vcn_v4_0_5_hw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		struct amdgpu_vcn_inst *vinst = &adev->vcn.inst[i];
+
+		if (adev->vcn.harvest_config & (1 << i))
+			continue;
+
+		cancel_delayed_work_sync(&vinst->idle_work);
+
+		if (!amdgpu_sriov_vf(adev)) {
+			if ((adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) ||
+			    (vinst->cur_state != AMD_PG_STATE_GATE &&
+			     RREG32_SOC15(VCN, i, regUVD_STATUS))) {
+				vinst->set_pg_state(vinst, AMD_PG_STATE_GATE);
+			}
+		}
+	}
+
+	return 0;
+}
+
+/**
+ * vcn_v4_0_5_suspend - suspend VCN block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * HW fini and suspend VCN block
+ */
+static int vcn_v4_0_5_suspend(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int r, i;
+
+	r = vcn_v4_0_5_hw_fini(ip_block);
+	if (r)
+		return r;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+		r = amdgpu_vcn_suspend(ip_block->adev, i);
+		if (r)
+			return r;
+	}
+
+	return r;
+}
+
+/**
+ * vcn_v4_0_5_resume - resume VCN block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Resume firmware and hw init VCN block
+ */
+static int vcn_v4_0_5_resume(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int r, i;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+		r = amdgpu_vcn_resume(ip_block->adev, i);
+		if (r)
+			return r;
+	}
+
+	r = vcn_v4_0_5_hw_init(ip_block);
+
+	return r;
+}
+
+/**
+ * vcn_v4_0_5_mc_resume - memory controller programming
+ *
+ * @vinst: VCN instance
+ *
+ * Let the VCN memory controller know it's offsets
+ */
+static void vcn_v4_0_5_mc_resume(struct amdgpu_vcn_inst *vinst)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int inst = vinst->inst;
+	uint32_t offset, size;
+	const struct common_firmware_header *hdr;
+
+	hdr = (const struct common_firmware_header *)adev->vcn.inst[inst].fw->data;
+	size = AMDGPU_GPU_PAGE_ALIGN(le32_to_cpu(hdr->ucode_size_bytes) + 8);
+
+	/* cache window 0: fw */
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
+		WREG32_SOC15(VCN, inst, regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW,
+			(adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + inst].tmr_mc_addr_lo));
+		WREG32_SOC15(VCN, inst, regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH,
+			(adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + inst].tmr_mc_addr_hi));
+		WREG32_SOC15(VCN, inst, regUVD_VCPU_CACHE_OFFSET0, 0);
+		offset = 0;
+	} else {
+		WREG32_SOC15(VCN, inst, regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW,
+			lower_32_bits(adev->vcn.inst[inst].gpu_addr));
+		WREG32_SOC15(VCN, inst, regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH,
+			upper_32_bits(adev->vcn.inst[inst].gpu_addr));
+		offset = size;
+		WREG32_SOC15(VCN, inst, regUVD_VCPU_CACHE_OFFSET0, AMDGPU_UVD_FIRMWARE_OFFSET >> 3);
+	}
+	WREG32_SOC15(VCN, inst, regUVD_VCPU_CACHE_SIZE0, size);
+
+	/* cache window 1: stack */
+	WREG32_SOC15(VCN, inst, regUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW,
+		lower_32_bits(adev->vcn.inst[inst].gpu_addr + offset));
+	WREG32_SOC15(VCN, inst, regUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH,
+		upper_32_bits(adev->vcn.inst[inst].gpu_addr + offset));
+	WREG32_SOC15(VCN, inst, regUVD_VCPU_CACHE_OFFSET1, 0);
+	WREG32_SOC15(VCN, inst, regUVD_VCPU_CACHE_SIZE1, AMDGPU_VCN_STACK_SIZE);
+
+	/* cache window 2: context */
+	WREG32_SOC15(VCN, inst, regUVD_LMI_VCPU_CACHE2_64BIT_BAR_LOW,
+		lower_32_bits(adev->vcn.inst[inst].gpu_addr + offset + AMDGPU_VCN_STACK_SIZE));
+	WREG32_SOC15(VCN, inst, regUVD_LMI_VCPU_CACHE2_64BIT_BAR_HIGH,
+		upper_32_bits(adev->vcn.inst[inst].gpu_addr + offset + AMDGPU_VCN_STACK_SIZE));
+	WREG32_SOC15(VCN, inst, regUVD_VCPU_CACHE_OFFSET2, 0);
+	WREG32_SOC15(VCN, inst, regUVD_VCPU_CACHE_SIZE2, AMDGPU_VCN_CONTEXT_SIZE);
+
+	/* non-cache window */
+	WREG32_SOC15(VCN, inst, regUVD_LMI_VCPU_NC0_64BIT_BAR_LOW,
+		lower_32_bits(adev->vcn.inst[inst].fw_shared.gpu_addr));
+	WREG32_SOC15(VCN, inst, regUVD_LMI_VCPU_NC0_64BIT_BAR_HIGH,
+		upper_32_bits(adev->vcn.inst[inst].fw_shared.gpu_addr));
+	WREG32_SOC15(VCN, inst, regUVD_VCPU_NONCACHE_OFFSET0, 0);
+	WREG32_SOC15(VCN, inst, regUVD_VCPU_NONCACHE_SIZE0,
+		AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_vcn4_fw_shared)));
+}
+
+/**
+ * vcn_v4_0_5_mc_resume_dpg_mode - memory controller programming for dpg mode
+ *
+ * @vinst: VCN instance
+ * @indirect: indirectly write sram
+ *
+ * Let the VCN memory controller know it's offsets with dpg mode
+ */
+static void vcn_v4_0_5_mc_resume_dpg_mode(struct amdgpu_vcn_inst *vinst,
+					  bool indirect)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int inst_idx = vinst->inst;
+	uint32_t offset, size;
+	const struct common_firmware_header *hdr;
+
+	hdr = (const struct common_firmware_header *)adev->vcn.inst[inst_idx].fw->data;
+	size = AMDGPU_GPU_PAGE_ALIGN(le32_to_cpu(hdr->ucode_size_bytes) + 8);
+
+	/* cache window 0: fw */
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
+		if (!indirect) {
+			WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW),
+			(adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + inst_idx].tmr_mc_addr_lo),
+			0, indirect);
+			WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH),
+			(adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + inst_idx].tmr_mc_addr_hi),
+			0, indirect);
+			WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+				VCN, inst_idx, regUVD_VCPU_CACHE_OFFSET0), 0, 0, indirect);
+		} else {
+			WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW), 0, 0, indirect);
+			WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH), 0, 0, indirect);
+			WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+				VCN, inst_idx, regUVD_VCPU_CACHE_OFFSET0), 0, 0, indirect);
+		}
+		offset = 0;
+	} else {
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW),
+			lower_32_bits(adev->vcn.inst[inst_idx].gpu_addr), 0, indirect);
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH),
+			upper_32_bits(adev->vcn.inst[inst_idx].gpu_addr), 0, indirect);
+		offset = size;
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_VCPU_CACHE_OFFSET0),
+			AMDGPU_UVD_FIRMWARE_OFFSET >> 3, 0, indirect);
+	}
+
+	if (!indirect)
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_VCPU_CACHE_SIZE0), size, 0, indirect);
+	else
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_VCPU_CACHE_SIZE0), 0, 0, indirect);
+
+	/* cache window 1: stack */
+	if (!indirect) {
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW),
+			lower_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset), 0, indirect);
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH),
+			upper_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset), 0, indirect);
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_VCPU_CACHE_OFFSET1), 0, 0, indirect);
+	} else {
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW), 0, 0, indirect);
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH), 0, 0, indirect);
+		WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_VCPU_CACHE_OFFSET1), 0, 0, indirect);
+	}
+
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_VCPU_CACHE_SIZE1), AMDGPU_VCN_STACK_SIZE, 0, indirect);
+
+	/* cache window 2: context */
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_LMI_VCPU_CACHE2_64BIT_BAR_LOW),
+		lower_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset + AMDGPU_VCN_STACK_SIZE),
+		0, indirect);
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_LMI_VCPU_CACHE2_64BIT_BAR_HIGH),
+		upper_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset + AMDGPU_VCN_STACK_SIZE),
+		0, indirect);
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_VCPU_CACHE_OFFSET2), 0, 0, indirect);
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_VCPU_CACHE_SIZE2), AMDGPU_VCN_CONTEXT_SIZE, 0, indirect);
+
+	/* non-cache window */
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_LMI_VCPU_NC0_64BIT_BAR_LOW),
+		lower_32_bits(adev->vcn.inst[inst_idx].fw_shared.gpu_addr), 0, indirect);
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_LMI_VCPU_NC0_64BIT_BAR_HIGH),
+		upper_32_bits(adev->vcn.inst[inst_idx].fw_shared.gpu_addr), 0, indirect);
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_VCPU_NONCACHE_OFFSET0), 0, 0, indirect);
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_VCPU_NONCACHE_SIZE0),
+		AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_vcn4_fw_shared)), 0, indirect);
+
+	/* VCN global tiling registers */
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_GFX10_ADDR_CONFIG),
+		adev->gfx.config.gb_addr_config, 0, indirect);
+}
+
+/**
+ * vcn_v4_0_5_disable_static_power_gating - disable VCN static power gating
+ *
+ * @vinst: VCN instance
+ *
+ * Disable static power gating for VCN block
+ */
+static void vcn_v4_0_5_disable_static_power_gating(struct amdgpu_vcn_inst *vinst)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int inst = vinst->inst;
+	uint32_t data = 0;
+
+	if (adev->pg_flags & AMD_PG_SUPPORT_VCN) {
+		WREG32_SOC15(VCN, inst, regUVD_IPX_DLDO_CONFIG,
+					1 << UVD_IPX_DLDO_CONFIG__ONO2_PWR_CONFIG__SHIFT);
+		SOC15_WAIT_ON_RREG(VCN, inst, regUVD_IPX_DLDO_STATUS, 0,
+					UVD_IPX_DLDO_STATUS__ONO2_PWR_STATUS_MASK);
+		WREG32_SOC15(VCN, inst, regUVD_IPX_DLDO_CONFIG,
+					2 << UVD_IPX_DLDO_CONFIG__ONO3_PWR_CONFIG__SHIFT);
+		SOC15_WAIT_ON_RREG(VCN, inst, regUVD_IPX_DLDO_STATUS,
+					1 << UVD_IPX_DLDO_STATUS__ONO3_PWR_STATUS__SHIFT,
+					UVD_IPX_DLDO_STATUS__ONO3_PWR_STATUS_MASK);
+		WREG32_SOC15(VCN, inst, regUVD_IPX_DLDO_CONFIG,
+					2 << UVD_IPX_DLDO_CONFIG__ONO4_PWR_CONFIG__SHIFT);
+		SOC15_WAIT_ON_RREG(VCN, inst, regUVD_IPX_DLDO_STATUS,
+					1 << UVD_IPX_DLDO_STATUS__ONO4_PWR_STATUS__SHIFT,
+					UVD_IPX_DLDO_STATUS__ONO4_PWR_STATUS_MASK);
+		WREG32_SOC15(VCN, inst, regUVD_IPX_DLDO_CONFIG,
+					2 << UVD_IPX_DLDO_CONFIG__ONO5_PWR_CONFIG__SHIFT);
+		SOC15_WAIT_ON_RREG(VCN, inst, regUVD_IPX_DLDO_STATUS,
+					1 << UVD_IPX_DLDO_STATUS__ONO5_PWR_STATUS__SHIFT,
+					UVD_IPX_DLDO_STATUS__ONO5_PWR_STATUS_MASK);
+	} else {
+		WREG32_SOC15(VCN, inst, regUVD_IPX_DLDO_CONFIG,
+			1 << UVD_IPX_DLDO_CONFIG__ONO2_PWR_CONFIG__SHIFT);
+		SOC15_WAIT_ON_RREG(VCN, inst, regUVD_IPX_DLDO_STATUS,
+			0, UVD_IPX_DLDO_STATUS__ONO2_PWR_STATUS_MASK);
+		WREG32_SOC15(VCN, inst, regUVD_IPX_DLDO_CONFIG,
+			1 << UVD_IPX_DLDO_CONFIG__ONO3_PWR_CONFIG__SHIFT);
+		SOC15_WAIT_ON_RREG(VCN, inst, regUVD_IPX_DLDO_STATUS,
+			0, UVD_IPX_DLDO_STATUS__ONO3_PWR_STATUS_MASK);
+		WREG32_SOC15(VCN, inst, regUVD_IPX_DLDO_CONFIG,
+			1 << UVD_IPX_DLDO_CONFIG__ONO4_PWR_CONFIG__SHIFT);
+		SOC15_WAIT_ON_RREG(VCN, inst, regUVD_IPX_DLDO_STATUS,
+			0, UVD_IPX_DLDO_STATUS__ONO4_PWR_STATUS_MASK);
+		WREG32_SOC15(VCN, inst, regUVD_IPX_DLDO_CONFIG,
+			1 << UVD_IPX_DLDO_CONFIG__ONO5_PWR_CONFIG__SHIFT);
+		SOC15_WAIT_ON_RREG(VCN, inst, regUVD_IPX_DLDO_STATUS,
+			0, UVD_IPX_DLDO_STATUS__ONO5_PWR_STATUS_MASK);
+	}
+
+	data = RREG32_SOC15(VCN, inst, regUVD_POWER_STATUS);
+	data &= ~0x103;
+	if (adev->pg_flags & AMD_PG_SUPPORT_VCN)
+		data |= UVD_PGFSM_CONFIG__UVDM_UVDU_PWR_ON |
+			UVD_POWER_STATUS__UVD_PG_EN_MASK;
+	WREG32_SOC15(VCN, inst, regUVD_POWER_STATUS, data);
+}
+
+/**
+ * vcn_v4_0_5_enable_static_power_gating - enable VCN static power gating
+ *
+ * @vinst: VCN instance
+ *
+ * Enable static power gating for VCN block
+ */
+static void vcn_v4_0_5_enable_static_power_gating(struct amdgpu_vcn_inst *vinst)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int inst = vinst->inst;
+	uint32_t data;
+
+	if (adev->pg_flags & AMD_PG_SUPPORT_VCN) {
+		/* Before power off, this indicator has to be turned on */
+		data = RREG32_SOC15(VCN, inst, regUVD_POWER_STATUS);
+		data &= ~UVD_POWER_STATUS__UVD_POWER_STATUS_MASK;
+		data |= UVD_POWER_STATUS__UVD_POWER_STATUS_TILES_OFF;
+		WREG32_SOC15(VCN, inst, regUVD_POWER_STATUS, data);
+
+		WREG32_SOC15(VCN, inst, regUVD_IPX_DLDO_CONFIG,
+			2 << UVD_IPX_DLDO_CONFIG__ONO5_PWR_CONFIG__SHIFT);
+		SOC15_WAIT_ON_RREG(VCN, inst, regUVD_IPX_DLDO_STATUS,
+			1 << UVD_IPX_DLDO_STATUS__ONO5_PWR_STATUS__SHIFT,
+			UVD_IPX_DLDO_STATUS__ONO5_PWR_STATUS_MASK);
+		WREG32_SOC15(VCN, inst, regUVD_IPX_DLDO_CONFIG,
+			2 << UVD_IPX_DLDO_CONFIG__ONO4_PWR_CONFIG__SHIFT);
+		SOC15_WAIT_ON_RREG(VCN, inst, regUVD_IPX_DLDO_STATUS,
+			1 << UVD_IPX_DLDO_STATUS__ONO4_PWR_STATUS__SHIFT,
+			UVD_IPX_DLDO_STATUS__ONO4_PWR_STATUS_MASK);
+		WREG32_SOC15(VCN, inst, regUVD_IPX_DLDO_CONFIG,
+			2 << UVD_IPX_DLDO_CONFIG__ONO3_PWR_CONFIG__SHIFT);
+		SOC15_WAIT_ON_RREG(VCN, inst, regUVD_IPX_DLDO_STATUS,
+			1 << UVD_IPX_DLDO_STATUS__ONO3_PWR_STATUS__SHIFT,
+			UVD_IPX_DLDO_STATUS__ONO3_PWR_STATUS_MASK);
+		WREG32_SOC15(VCN, inst, regUVD_IPX_DLDO_CONFIG,
+			2 << UVD_IPX_DLDO_CONFIG__ONO2_PWR_CONFIG__SHIFT);
+		SOC15_WAIT_ON_RREG(VCN, inst, regUVD_IPX_DLDO_STATUS,
+			1 << UVD_IPX_DLDO_STATUS__ONO2_PWR_STATUS__SHIFT,
+			UVD_IPX_DLDO_STATUS__ONO2_PWR_STATUS_MASK);
+	}
+}
+
+/**
+ * vcn_v4_0_5_disable_clock_gating - disable VCN clock gating
+ *
+ * @vinst: VCN instance
+ *
+ * Disable clock gating for VCN block
+ */
+static void vcn_v4_0_5_disable_clock_gating(struct amdgpu_vcn_inst *vinst)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int inst = vinst->inst;
+	uint32_t data;
+
+	if (adev->cg_flags & AMD_CG_SUPPORT_VCN_MGCG)
+		return;
+
+	/* VCN disable CGC */
+	data = RREG32_SOC15(VCN, inst, regUVD_CGC_CTRL);
+	data &= ~UVD_CGC_CTRL__DYN_CLOCK_MODE_MASK;
+	data |= 1 << UVD_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT;
+	data |= 4 << UVD_CGC_CTRL__CLK_OFF_DELAY__SHIFT;
+	WREG32_SOC15(VCN, inst, regUVD_CGC_CTRL, data);
+
+	data = RREG32_SOC15(VCN, inst, regUVD_CGC_GATE);
+	data &= ~(UVD_CGC_GATE__SYS_MASK
+		| UVD_CGC_GATE__UDEC_MASK
+		| UVD_CGC_GATE__MPEG2_MASK
+		| UVD_CGC_GATE__REGS_MASK
+		| UVD_CGC_GATE__RBC_MASK
+		| UVD_CGC_GATE__LMI_MC_MASK
+		| UVD_CGC_GATE__LMI_UMC_MASK
+		| UVD_CGC_GATE__IDCT_MASK
+		| UVD_CGC_GATE__MPRD_MASK
+		| UVD_CGC_GATE__MPC_MASK
+		| UVD_CGC_GATE__LBSI_MASK
+		| UVD_CGC_GATE__LRBBM_MASK
+		| UVD_CGC_GATE__UDEC_RE_MASK
+		| UVD_CGC_GATE__UDEC_CM_MASK
+		| UVD_CGC_GATE__UDEC_IT_MASK
+		| UVD_CGC_GATE__UDEC_DB_MASK
+		| UVD_CGC_GATE__UDEC_MP_MASK
+		| UVD_CGC_GATE__WCB_MASK
+		| UVD_CGC_GATE__VCPU_MASK
+		| UVD_CGC_GATE__MMSCH_MASK);
+
+	WREG32_SOC15(VCN, inst, regUVD_CGC_GATE, data);
+	SOC15_WAIT_ON_RREG(VCN, inst, regUVD_CGC_GATE, 0,  0xFFFFFFFF);
+
+	data = RREG32_SOC15(VCN, inst, regUVD_CGC_CTRL);
+	data &= ~(UVD_CGC_CTRL__UDEC_RE_MODE_MASK
+		| UVD_CGC_CTRL__UDEC_CM_MODE_MASK
+		| UVD_CGC_CTRL__UDEC_IT_MODE_MASK
+		| UVD_CGC_CTRL__UDEC_DB_MODE_MASK
+		| UVD_CGC_CTRL__UDEC_MP_MODE_MASK
+		| UVD_CGC_CTRL__SYS_MODE_MASK
+		| UVD_CGC_CTRL__UDEC_MODE_MASK
+		| UVD_CGC_CTRL__MPEG2_MODE_MASK
+		| UVD_CGC_CTRL__REGS_MODE_MASK
+		| UVD_CGC_CTRL__RBC_MODE_MASK
+		| UVD_CGC_CTRL__LMI_MC_MODE_MASK
+		| UVD_CGC_CTRL__LMI_UMC_MODE_MASK
+		| UVD_CGC_CTRL__IDCT_MODE_MASK
+		| UVD_CGC_CTRL__MPRD_MODE_MASK
+		| UVD_CGC_CTRL__MPC_MODE_MASK
+		| UVD_CGC_CTRL__LBSI_MODE_MASK
+		| UVD_CGC_CTRL__LRBBM_MODE_MASK
+		| UVD_CGC_CTRL__WCB_MODE_MASK
+		| UVD_CGC_CTRL__VCPU_MODE_MASK
+		| UVD_CGC_CTRL__MMSCH_MODE_MASK);
+	WREG32_SOC15(VCN, inst, regUVD_CGC_CTRL, data);
+
+	data = RREG32_SOC15(VCN, inst, regUVD_SUVD_CGC_GATE);
+	data |= (UVD_SUVD_CGC_GATE__SRE_MASK
+		| UVD_SUVD_CGC_GATE__SIT_MASK
+		| UVD_SUVD_CGC_GATE__SMP_MASK
+		| UVD_SUVD_CGC_GATE__SCM_MASK
+		| UVD_SUVD_CGC_GATE__SDB_MASK
+		| UVD_SUVD_CGC_GATE__SRE_H264_MASK
+		| UVD_SUVD_CGC_GATE__SRE_HEVC_MASK
+		| UVD_SUVD_CGC_GATE__SIT_H264_MASK
+		| UVD_SUVD_CGC_GATE__SIT_HEVC_MASK
+		| UVD_SUVD_CGC_GATE__SCM_H264_MASK
+		| UVD_SUVD_CGC_GATE__SCM_HEVC_MASK
+		| UVD_SUVD_CGC_GATE__SDB_H264_MASK
+		| UVD_SUVD_CGC_GATE__SDB_HEVC_MASK
+		| UVD_SUVD_CGC_GATE__SCLR_MASK
+		| UVD_SUVD_CGC_GATE__UVD_SC_MASK
+		| UVD_SUVD_CGC_GATE__ENT_MASK
+		| UVD_SUVD_CGC_GATE__SIT_HEVC_DEC_MASK
+		| UVD_SUVD_CGC_GATE__SIT_HEVC_ENC_MASK
+		| UVD_SUVD_CGC_GATE__SITE_MASK
+		| UVD_SUVD_CGC_GATE__SRE_VP9_MASK
+		| UVD_SUVD_CGC_GATE__SCM_VP9_MASK
+		| UVD_SUVD_CGC_GATE__SIT_VP9_DEC_MASK
+		| UVD_SUVD_CGC_GATE__SDB_VP9_MASK
+		| UVD_SUVD_CGC_GATE__IME_HEVC_MASK);
+	WREG32_SOC15(VCN, inst, regUVD_SUVD_CGC_GATE, data);
+
+	data = RREG32_SOC15(VCN, inst, regUVD_SUVD_CGC_CTRL);
+	data &= ~(UVD_SUVD_CGC_CTRL__SRE_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SIT_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SMP_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SCM_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SDB_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SCLR_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__UVD_SC_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__ENT_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__IME_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SITE_MODE_MASK);
+	WREG32_SOC15(VCN, inst, regUVD_SUVD_CGC_CTRL, data);
+}
+
+/**
+ * vcn_v4_0_5_disable_clock_gating_dpg_mode - disable VCN clock gating dpg mode
+ *
+ * @vinst: VCN instance
+ * @sram_sel: sram select
+ * @indirect: indirectly write sram
+ *
+ * Disable clock gating for VCN block with dpg mode
+ */
+static void vcn_v4_0_5_disable_clock_gating_dpg_mode(struct amdgpu_vcn_inst *vinst,
+						     uint8_t sram_sel,
+						     uint8_t indirect)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int inst_idx = vinst->inst;
+	uint32_t reg_data = 0;
+
+	if (adev->cg_flags & AMD_CG_SUPPORT_VCN_MGCG)
+		return;
+
+	/* enable sw clock gating control */
+	reg_data = 0 << UVD_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
+	reg_data |= 1 << UVD_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT;
+	reg_data |= 4 << UVD_CGC_CTRL__CLK_OFF_DELAY__SHIFT;
+	reg_data &= ~(UVD_CGC_CTRL__UDEC_RE_MODE_MASK |
+		 UVD_CGC_CTRL__UDEC_CM_MODE_MASK |
+		 UVD_CGC_CTRL__UDEC_IT_MODE_MASK |
+		 UVD_CGC_CTRL__UDEC_DB_MODE_MASK |
+		 UVD_CGC_CTRL__UDEC_MP_MODE_MASK |
+		 UVD_CGC_CTRL__SYS_MODE_MASK |
+		 UVD_CGC_CTRL__UDEC_MODE_MASK |
+		 UVD_CGC_CTRL__MPEG2_MODE_MASK |
+		 UVD_CGC_CTRL__REGS_MODE_MASK |
+		 UVD_CGC_CTRL__RBC_MODE_MASK |
+		 UVD_CGC_CTRL__LMI_MC_MODE_MASK |
+		 UVD_CGC_CTRL__LMI_UMC_MODE_MASK |
+		 UVD_CGC_CTRL__IDCT_MODE_MASK |
+		 UVD_CGC_CTRL__MPRD_MODE_MASK |
+		 UVD_CGC_CTRL__MPC_MODE_MASK |
+		 UVD_CGC_CTRL__LBSI_MODE_MASK |
+		 UVD_CGC_CTRL__LRBBM_MODE_MASK |
+		 UVD_CGC_CTRL__WCB_MODE_MASK |
+		 UVD_CGC_CTRL__VCPU_MODE_MASK);
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_CGC_CTRL), reg_data, sram_sel, indirect);
+
+	/* turn off clock gating */
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_CGC_GATE), 0, sram_sel, indirect);
+
+	/* turn on SUVD clock gating */
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_SUVD_CGC_GATE), 1, sram_sel, indirect);
+
+	/* turn on sw mode in UVD_SUVD_CGC_CTRL */
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_SUVD_CGC_CTRL), 0, sram_sel, indirect);
+}
+
+/**
+ * vcn_v4_0_5_enable_clock_gating - enable VCN clock gating
+ *
+ * @vinst: VCN instance
+ *
+ * Enable clock gating for VCN block
+ */
+static void vcn_v4_0_5_enable_clock_gating(struct amdgpu_vcn_inst *vinst)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int inst = vinst->inst;
+	uint32_t data;
+
+	if (adev->cg_flags & AMD_CG_SUPPORT_VCN_MGCG)
+		return;
+
+	/* enable VCN CGC */
+	data = RREG32_SOC15(VCN, inst, regUVD_CGC_CTRL);
+	data |= 0 << UVD_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
+	data |= 1 << UVD_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT;
+	data |= 4 << UVD_CGC_CTRL__CLK_OFF_DELAY__SHIFT;
+	WREG32_SOC15(VCN, inst, regUVD_CGC_CTRL, data);
+
+	data = RREG32_SOC15(VCN, inst, regUVD_CGC_CTRL);
+	data |= (UVD_CGC_CTRL__UDEC_RE_MODE_MASK
+		| UVD_CGC_CTRL__UDEC_CM_MODE_MASK
+		| UVD_CGC_CTRL__UDEC_IT_MODE_MASK
+		| UVD_CGC_CTRL__UDEC_DB_MODE_MASK
+		| UVD_CGC_CTRL__UDEC_MP_MODE_MASK
+		| UVD_CGC_CTRL__SYS_MODE_MASK
+		| UVD_CGC_CTRL__UDEC_MODE_MASK
+		| UVD_CGC_CTRL__MPEG2_MODE_MASK
+		| UVD_CGC_CTRL__REGS_MODE_MASK
+		| UVD_CGC_CTRL__RBC_MODE_MASK
+		| UVD_CGC_CTRL__LMI_MC_MODE_MASK
+		| UVD_CGC_CTRL__LMI_UMC_MODE_MASK
+		| UVD_CGC_CTRL__IDCT_MODE_MASK
+		| UVD_CGC_CTRL__MPRD_MODE_MASK
+		| UVD_CGC_CTRL__MPC_MODE_MASK
+		| UVD_CGC_CTRL__LBSI_MODE_MASK
+		| UVD_CGC_CTRL__LRBBM_MODE_MASK
+		| UVD_CGC_CTRL__WCB_MODE_MASK
+		| UVD_CGC_CTRL__VCPU_MODE_MASK
+		| UVD_CGC_CTRL__MMSCH_MODE_MASK);
+	WREG32_SOC15(VCN, inst, regUVD_CGC_CTRL, data);
+
+	data = RREG32_SOC15(VCN, inst, regUVD_SUVD_CGC_CTRL);
+	data |= (UVD_SUVD_CGC_CTRL__SRE_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SIT_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SMP_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SCM_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SDB_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SCLR_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__UVD_SC_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__ENT_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__IME_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SITE_MODE_MASK);
+	WREG32_SOC15(VCN, inst, regUVD_SUVD_CGC_CTRL, data);
+}
+
+/**
+ * vcn_v4_0_5_start_dpg_mode - VCN start with dpg mode
+ *
+ * @vinst: VCN instance
+ * @indirect: indirectly write sram
+ *
+ * Start VCN block with dpg mode
+ */
+static int vcn_v4_0_5_start_dpg_mode(struct amdgpu_vcn_inst *vinst,
+				     bool indirect)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int inst_idx = vinst->inst;
+	struct amdgpu_vcn4_fw_shared *fw_shared = adev->vcn.inst[inst_idx].fw_shared.cpu_addr;
+	struct amdgpu_ring *ring;
+	uint32_t tmp;
+	int ret;
+
+	/* disable register anti-hang mechanism */
+	WREG32_P(SOC15_REG_OFFSET(VCN, inst_idx, regUVD_POWER_STATUS), 1,
+		~UVD_POWER_STATUS__UVD_POWER_STATUS_MASK);
+	/* enable dynamic power gating mode */
+	tmp = RREG32_SOC15(VCN, inst_idx, regUVD_POWER_STATUS);
+	tmp |= UVD_POWER_STATUS__UVD_PG_MODE_MASK;
+	tmp |= UVD_POWER_STATUS__UVD_PG_EN_MASK;
+	WREG32_SOC15(VCN, inst_idx, regUVD_POWER_STATUS, tmp);
+
+	if (indirect)
+		adev->vcn.inst[inst_idx].dpg_sram_curr_addr =
+					(uint32_t *)adev->vcn.inst[inst_idx].dpg_sram_cpu_addr;
+
+	/* enable clock gating */
+	vcn_v4_0_5_disable_clock_gating_dpg_mode(vinst, 0, indirect);
+
+	/* enable VCPU clock */
+	tmp = (0xFF << UVD_VCPU_CNTL__PRB_TIMEOUT_VAL__SHIFT);
+	tmp |= UVD_VCPU_CNTL__CLK_EN_MASK | UVD_VCPU_CNTL__BLK_RST_MASK;
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_VCPU_CNTL), tmp, 0, indirect);
+
+	/* disable master interrupt */
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_MASTINT_EN), 0, 0, indirect);
+
+	/* setup regUVD_LMI_CTRL */
+	tmp = (UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK |
+		UVD_LMI_CTRL__REQ_MODE_MASK |
+		UVD_LMI_CTRL__CRC_RESET_MASK |
+		UVD_LMI_CTRL__MASK_MC_URGENT_MASK |
+		UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK |
+		UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK |
+		(8 << UVD_LMI_CTRL__WRITE_CLEAN_TIMER__SHIFT) |
+		0x00100000L);
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_LMI_CTRL), tmp, 0, indirect);
+
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_MPC_CNTL),
+		0x2 << UVD_MPC_CNTL__REPLACEMENT_MODE__SHIFT, 0, indirect);
+
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_MPC_SET_MUXA0),
+		((0x1 << UVD_MPC_SET_MUXA0__VARA_1__SHIFT) |
+		 (0x2 << UVD_MPC_SET_MUXA0__VARA_2__SHIFT) |
+		 (0x3 << UVD_MPC_SET_MUXA0__VARA_3__SHIFT) |
+		 (0x4 << UVD_MPC_SET_MUXA0__VARA_4__SHIFT)), 0, indirect);
+
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_MPC_SET_MUXB0),
+		 ((0x1 << UVD_MPC_SET_MUXB0__VARB_1__SHIFT) |
+		 (0x2 << UVD_MPC_SET_MUXB0__VARB_2__SHIFT) |
+		 (0x3 << UVD_MPC_SET_MUXB0__VARB_3__SHIFT) |
+		 (0x4 << UVD_MPC_SET_MUXB0__VARB_4__SHIFT)), 0, indirect);
+
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_MPC_SET_MUX),
+		((0x0 << UVD_MPC_SET_MUX__SET_0__SHIFT) |
+		 (0x1 << UVD_MPC_SET_MUX__SET_1__SHIFT) |
+		 (0x2 << UVD_MPC_SET_MUX__SET_2__SHIFT)), 0, indirect);
+
+	vcn_v4_0_5_mc_resume_dpg_mode(vinst, indirect);
+
+	tmp = (0xFF << UVD_VCPU_CNTL__PRB_TIMEOUT_VAL__SHIFT);
+	tmp |= UVD_VCPU_CNTL__CLK_EN_MASK;
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_VCPU_CNTL), tmp, 0, indirect);
+
+	/* enable LMI MC and UMC channels */
+	tmp = 0x1f << UVD_LMI_CTRL2__RE_OFLD_MIF_WR_REQ_NUM__SHIFT;
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_LMI_CTRL2), tmp, 0, indirect);
+
+	/* enable master interrupt */
+	WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_MASTINT_EN),
+		UVD_MASTINT_EN__VCPU_EN_MASK, 0, indirect);
+
+	if (indirect) {
+		ret = amdgpu_vcn_psp_update_sram(adev, inst_idx, 0);
+		if (ret) {
+			dev_err(adev->dev, "vcn sram load failed %d\n", ret);
+			return ret;
+		}
+	}
+
+	ring = &adev->vcn.inst[inst_idx].ring_enc[0];
+
+	WREG32_SOC15(VCN, inst_idx, regUVD_RB_BASE_LO, ring->gpu_addr);
+	WREG32_SOC15(VCN, inst_idx, regUVD_RB_BASE_HI, upper_32_bits(ring->gpu_addr));
+	WREG32_SOC15(VCN, inst_idx, regUVD_RB_SIZE, ring->ring_size / 4);
+
+	tmp = RREG32_SOC15(VCN, inst_idx, regVCN_RB_ENABLE);
+	tmp &= ~(VCN_RB_ENABLE__RB1_EN_MASK);
+	WREG32_SOC15(VCN, inst_idx, regVCN_RB_ENABLE, tmp);
+	fw_shared->sq.queue_mode |= FW_QUEUE_RING_RESET;
+	WREG32_SOC15(VCN, inst_idx, regUVD_RB_RPTR, 0);
+	WREG32_SOC15(VCN, inst_idx, regUVD_RB_WPTR, 0);
+
+	tmp = RREG32_SOC15(VCN, inst_idx, regUVD_RB_RPTR);
+	WREG32_SOC15(VCN, inst_idx, regUVD_RB_WPTR, tmp);
+	ring->wptr = RREG32_SOC15(VCN, inst_idx, regUVD_RB_WPTR);
+
+	tmp = RREG32_SOC15(VCN, inst_idx, regVCN_RB_ENABLE);
+	tmp |= VCN_RB_ENABLE__RB1_EN_MASK;
+	WREG32_SOC15(VCN, inst_idx, regVCN_RB_ENABLE, tmp);
+	fw_shared->sq.queue_mode &= ~(FW_QUEUE_RING_RESET | FW_QUEUE_DPG_HOLD_OFF);
+
+	WREG32_SOC15(VCN, inst_idx, regVCN_RB1_DB_CTRL,
+			ring->doorbell_index << VCN_RB1_DB_CTRL__OFFSET__SHIFT |
+			VCN_RB1_DB_CTRL__EN_MASK);
+
+	/* Keeping one read-back to ensure all register writes are done, otherwise
+	 * it may introduce race conditions */
+	RREG32_SOC15(VCN, inst_idx, regVCN_RB1_DB_CTRL);
+
+	return 0;
+}
+
+
+/**
+ * vcn_v4_0_5_start - VCN start
+ *
+ * @vinst: VCN instance
+ *
+ * Start VCN block
+ */
+static int vcn_v4_0_5_start(struct amdgpu_vcn_inst *vinst)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int i = vinst->inst;
+	struct amdgpu_vcn4_fw_shared *fw_shared;
+	struct amdgpu_ring *ring;
+	uint32_t tmp;
+	int j, k, r;
+
+	if (adev->vcn.harvest_config & (1 << i))
+		return 0;
+
+	if (adev->pm.dpm_enabled)
+		amdgpu_dpm_enable_vcn(adev, true, i);
+
+	fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr;
+
+	if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)
+		return vcn_v4_0_5_start_dpg_mode(vinst, adev->vcn.inst[i].indirect_sram);
+
+	/* disable VCN power gating */
+	vcn_v4_0_5_disable_static_power_gating(vinst);
+
+	/* set VCN status busy */
+	tmp = RREG32_SOC15(VCN, i, regUVD_STATUS) | UVD_STATUS__UVD_BUSY;
+	WREG32_SOC15(VCN, i, regUVD_STATUS, tmp);
+
+	/* SW clock gating */
+	vcn_v4_0_5_disable_clock_gating(vinst);
+
+	/* enable VCPU clock */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL),
+		 UVD_VCPU_CNTL__CLK_EN_MASK, ~UVD_VCPU_CNTL__CLK_EN_MASK);
+
+	/* disable master interrupt */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_MASTINT_EN), 0,
+		 ~UVD_MASTINT_EN__VCPU_EN_MASK);
+
+	/* enable LMI MC and UMC channels */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_LMI_CTRL2), 0,
+		 ~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK);
+
+	tmp = RREG32_SOC15(VCN, i, regUVD_SOFT_RESET);
+	tmp &= ~UVD_SOFT_RESET__LMI_SOFT_RESET_MASK;
+	tmp &= ~UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK;
+	WREG32_SOC15(VCN, i, regUVD_SOFT_RESET, tmp);
+
+	/* setup regUVD_LMI_CTRL */
+	tmp = RREG32_SOC15(VCN, i, regUVD_LMI_CTRL);
+	WREG32_SOC15(VCN, i, regUVD_LMI_CTRL, tmp |
+		     UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK |
+		     UVD_LMI_CTRL__MASK_MC_URGENT_MASK |
+		     UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK |
+		     UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK);
+
+	/* setup regUVD_MPC_CNTL */
+	tmp = RREG32_SOC15(VCN, i, regUVD_MPC_CNTL);
+	tmp &= ~UVD_MPC_CNTL__REPLACEMENT_MODE_MASK;
+	tmp |= 0x2 << UVD_MPC_CNTL__REPLACEMENT_MODE__SHIFT;
+	WREG32_SOC15(VCN, i, regUVD_MPC_CNTL, tmp);
+
+	/* setup UVD_MPC_SET_MUXA0 */
+	WREG32_SOC15(VCN, i, regUVD_MPC_SET_MUXA0,
+		     ((0x1 << UVD_MPC_SET_MUXA0__VARA_1__SHIFT) |
+		      (0x2 << UVD_MPC_SET_MUXA0__VARA_2__SHIFT) |
+		      (0x3 << UVD_MPC_SET_MUXA0__VARA_3__SHIFT) |
+		      (0x4 << UVD_MPC_SET_MUXA0__VARA_4__SHIFT)));
+
+	/* setup UVD_MPC_SET_MUXB0 */
+	WREG32_SOC15(VCN, i, regUVD_MPC_SET_MUXB0,
+		     ((0x1 << UVD_MPC_SET_MUXB0__VARB_1__SHIFT) |
+		      (0x2 << UVD_MPC_SET_MUXB0__VARB_2__SHIFT) |
+		      (0x3 << UVD_MPC_SET_MUXB0__VARB_3__SHIFT) |
+		      (0x4 << UVD_MPC_SET_MUXB0__VARB_4__SHIFT)));
+
+	/* setup UVD_MPC_SET_MUX */
+	WREG32_SOC15(VCN, i, regUVD_MPC_SET_MUX,
+		     ((0x0 << UVD_MPC_SET_MUX__SET_0__SHIFT) |
+		      (0x1 << UVD_MPC_SET_MUX__SET_1__SHIFT) |
+		      (0x2 << UVD_MPC_SET_MUX__SET_2__SHIFT)));
+
+	vcn_v4_0_5_mc_resume(vinst);
+
+	/* VCN global tiling registers */
+	WREG32_SOC15(VCN, i, regUVD_GFX10_ADDR_CONFIG,
+		     adev->gfx.config.gb_addr_config);
+
+	/* unblock VCPU register access */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_RB_ARB_CTRL), 0,
+		 ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK);
+
+	/* release VCPU reset to boot */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), 0,
+		 ~UVD_VCPU_CNTL__BLK_RST_MASK);
+
+	for (j = 0; j < 10; ++j) {
+		uint32_t status;
+
+		for (k = 0; k < 100; ++k) {
+			status = RREG32_SOC15(VCN, i, regUVD_STATUS);
+			if (status & 2)
+				break;
+			mdelay(10);
+			if (amdgpu_emu_mode == 1)
+				msleep(1);
+		}
+
+		if (amdgpu_emu_mode == 1) {
+			r = -1;
+			if (status & 2) {
+				r = 0;
+				break;
+			}
+		} else {
+			r = 0;
+			if (status & 2)
+				break;
+
+			dev_err(adev->dev,
+				"VCN[%d] is not responding, trying to reset VCPU!!!\n", i);
+			WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL),
+				 UVD_VCPU_CNTL__BLK_RST_MASK,
+				 ~UVD_VCPU_CNTL__BLK_RST_MASK);
+			mdelay(10);
+			WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), 0,
+				 ~UVD_VCPU_CNTL__BLK_RST_MASK);
+
+			mdelay(10);
+			r = -1;
+		}
+	}
+
+	if (r) {
+		dev_err(adev->dev, "VCN[%d] is not responding, giving up!!!\n", i);
+		return r;
+	}
+
+	/* enable master interrupt */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_MASTINT_EN),
+		 UVD_MASTINT_EN__VCPU_EN_MASK,
+		 ~UVD_MASTINT_EN__VCPU_EN_MASK);
+
+	/* clear the busy bit of VCN_STATUS */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_STATUS), 0,
+		 ~(2 << UVD_STATUS__VCPU_REPORT__SHIFT));
+
+	ring = &adev->vcn.inst[i].ring_enc[0];
+	WREG32_SOC15(VCN, i, regVCN_RB1_DB_CTRL,
+		     ring->doorbell_index << VCN_RB1_DB_CTRL__OFFSET__SHIFT |
+		     VCN_RB1_DB_CTRL__EN_MASK);
+
+	WREG32_SOC15(VCN, i, regUVD_RB_BASE_LO, ring->gpu_addr);
+	WREG32_SOC15(VCN, i, regUVD_RB_BASE_HI, upper_32_bits(ring->gpu_addr));
+	WREG32_SOC15(VCN, i, regUVD_RB_SIZE, ring->ring_size / 4);
+
+	tmp = RREG32_SOC15(VCN, i, regVCN_RB_ENABLE);
+	tmp &= ~(VCN_RB_ENABLE__RB1_EN_MASK);
+	WREG32_SOC15(VCN, i, regVCN_RB_ENABLE, tmp);
+	fw_shared->sq.queue_mode |= FW_QUEUE_RING_RESET;
+	WREG32_SOC15(VCN, i, regUVD_RB_RPTR, 0);
+	WREG32_SOC15(VCN, i, regUVD_RB_WPTR, 0);
+
+	tmp = RREG32_SOC15(VCN, i, regUVD_RB_RPTR);
+	WREG32_SOC15(VCN, i, regUVD_RB_WPTR, tmp);
+	ring->wptr = RREG32_SOC15(VCN, i, regUVD_RB_WPTR);
+
+	tmp = RREG32_SOC15(VCN, i, regVCN_RB_ENABLE);
+	tmp |= VCN_RB_ENABLE__RB1_EN_MASK;
+	WREG32_SOC15(VCN, i, regVCN_RB_ENABLE, tmp);
+	fw_shared->sq.queue_mode &= ~(FW_QUEUE_RING_RESET | FW_QUEUE_DPG_HOLD_OFF);
+
+	/* Keeping one read-back to ensure all register writes are done, otherwise
+	 * it may introduce race conditions */
+	RREG32_SOC15(VCN, i, regVCN_RB_ENABLE);
+
+	return 0;
+}
+
+/**
+ * vcn_v4_0_5_stop_dpg_mode - VCN stop with dpg mode
+ *
+ * @vinst: VCN instance
+ *
+ * Stop VCN block with dpg mode
+ */
+static void vcn_v4_0_5_stop_dpg_mode(struct amdgpu_vcn_inst *vinst)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int inst_idx = vinst->inst;
+	uint32_t tmp;
+
+	/* Wait for power status to be 1 */
+	SOC15_WAIT_ON_RREG(VCN, inst_idx, regUVD_POWER_STATUS, 1,
+		UVD_POWER_STATUS__UVD_POWER_STATUS_MASK);
+
+	/* wait for read ptr to be equal to write ptr */
+	tmp = RREG32_SOC15(VCN, inst_idx, regUVD_RB_WPTR);
+	SOC15_WAIT_ON_RREG(VCN, inst_idx, regUVD_RB_RPTR, tmp, 0xFFFFFFFF);
+
+	SOC15_WAIT_ON_RREG(VCN, inst_idx, regUVD_POWER_STATUS, 1,
+		UVD_POWER_STATUS__UVD_POWER_STATUS_MASK);
+
+	/* disable dynamic power gating mode */
+	WREG32_P(SOC15_REG_OFFSET(VCN, inst_idx, regUVD_POWER_STATUS), 0,
+		~UVD_POWER_STATUS__UVD_PG_MODE_MASK);
+
+	/* Keeping one read-back to ensure all register writes are done,
+	 * otherwise it may introduce race conditions.
+	 */
+	RREG32_SOC15(VCN, inst_idx, regUVD_STATUS);
+}
+
+/**
+ * vcn_v4_0_5_stop - VCN stop
+ *
+ * @vinst: VCN instance
+ *
+ * Stop VCN block
+ */
+static int vcn_v4_0_5_stop(struct amdgpu_vcn_inst *vinst)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int i = vinst->inst;
+	struct amdgpu_vcn4_fw_shared *fw_shared;
+	uint32_t tmp;
+	int r = 0;
+
+	if (adev->vcn.harvest_config & (1 << i))
+		return 0;
+
+	fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr;
+	fw_shared->sq.queue_mode |= FW_QUEUE_DPG_HOLD_OFF;
+
+	if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) {
+		vcn_v4_0_5_stop_dpg_mode(vinst);
+		r = 0;
+		goto done;
+	}
+
+	/* wait for vcn idle */
+	r = SOC15_WAIT_ON_RREG(VCN, i, regUVD_STATUS, UVD_STATUS__IDLE, 0x7);
+	if (r)
+		goto done;
+
+	tmp = UVD_LMI_STATUS__VCPU_LMI_WRITE_CLEAN_MASK |
+		UVD_LMI_STATUS__READ_CLEAN_MASK |
+		UVD_LMI_STATUS__WRITE_CLEAN_MASK |
+		UVD_LMI_STATUS__WRITE_CLEAN_RAW_MASK;
+	r = SOC15_WAIT_ON_RREG(VCN, i, regUVD_LMI_STATUS, tmp, tmp);
+	if (r)
+		goto done;
+
+	/* disable LMI UMC channel */
+	tmp = RREG32_SOC15(VCN, i, regUVD_LMI_CTRL2);
+	tmp |= UVD_LMI_CTRL2__STALL_ARB_UMC_MASK;
+	WREG32_SOC15(VCN, i, regUVD_LMI_CTRL2, tmp);
+	tmp = UVD_LMI_STATUS__UMC_READ_CLEAN_RAW_MASK |
+		UVD_LMI_STATUS__UMC_WRITE_CLEAN_RAW_MASK;
+	r = SOC15_WAIT_ON_RREG(VCN, i, regUVD_LMI_STATUS, tmp, tmp);
+	if (r)
+		goto done;
+
+	/* block VCPU register access */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_RB_ARB_CTRL),
+		 UVD_RB_ARB_CTRL__VCPU_DIS_MASK,
+		 ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK);
+
+	/* reset VCPU */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL),
+		 UVD_VCPU_CNTL__BLK_RST_MASK,
+		 ~UVD_VCPU_CNTL__BLK_RST_MASK);
+
+	/* disable VCPU clock */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), 0,
+		 ~(UVD_VCPU_CNTL__CLK_EN_MASK));
+
+	/* apply soft reset */
+	tmp = RREG32_SOC15(VCN, i, regUVD_SOFT_RESET);
+	tmp |= UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK;
+	WREG32_SOC15(VCN, i, regUVD_SOFT_RESET, tmp);
+	tmp = RREG32_SOC15(VCN, i, regUVD_SOFT_RESET);
+	tmp |= UVD_SOFT_RESET__LMI_SOFT_RESET_MASK;
+	WREG32_SOC15(VCN, i, regUVD_SOFT_RESET, tmp);
+
+	/* clear status */
+	WREG32_SOC15(VCN, i, regUVD_STATUS, 0);
+
+	/* apply HW clock gating */
+	vcn_v4_0_5_enable_clock_gating(vinst);
+
+	/* enable VCN power gating */
+	vcn_v4_0_5_enable_static_power_gating(vinst);
+
+	/* Keeping one read-back to ensure all register writes are done,
+	 * otherwise it may introduce race conditions.
+	 */
+	RREG32_SOC15(VCN, i, regUVD_STATUS);
+
+done:
+	if (adev->pm.dpm_enabled)
+		amdgpu_dpm_enable_vcn(adev, false, i);
+
+	return r;
+}
+
+/**
+ * vcn_v4_0_5_pause_dpg_mode - VCN pause with dpg mode
+ *
+ * @vinst: VCN instance
+ * @new_state: pause state
+ *
+ * Pause dpg mode for VCN block
+ */
+static int vcn_v4_0_5_pause_dpg_mode(struct amdgpu_vcn_inst *vinst,
+				     struct dpg_pause_state *new_state)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int inst_idx = vinst->inst;
+	uint32_t reg_data = 0;
+	int ret_code;
+
+	/* pause/unpause if state is changed */
+	if (adev->vcn.inst[inst_idx].pause_state.fw_based != new_state->fw_based) {
+		DRM_DEV_DEBUG(adev->dev, "dpg pause state changed %d -> %d",
+			adev->vcn.inst[inst_idx].pause_state.fw_based,	new_state->fw_based);
+		reg_data = RREG32_SOC15(VCN, inst_idx, regUVD_DPG_PAUSE) &
+			(~UVD_DPG_PAUSE__NJ_PAUSE_DPG_ACK_MASK);
+
+		if (new_state->fw_based == VCN_DPG_STATE__PAUSE) {
+			ret_code = SOC15_WAIT_ON_RREG(VCN, inst_idx, regUVD_POWER_STATUS, 0x1,
+				UVD_POWER_STATUS__UVD_POWER_STATUS_MASK);
+
+			if (!ret_code) {
+				/* pause DPG */
+				reg_data |= UVD_DPG_PAUSE__NJ_PAUSE_DPG_REQ_MASK;
+				WREG32_SOC15(VCN, inst_idx, regUVD_DPG_PAUSE, reg_data);
+
+				/* wait for ACK */
+				SOC15_WAIT_ON_RREG(VCN, inst_idx, regUVD_DPG_PAUSE,
+					UVD_DPG_PAUSE__NJ_PAUSE_DPG_ACK_MASK,
+					UVD_DPG_PAUSE__NJ_PAUSE_DPG_ACK_MASK);
+
+				SOC15_WAIT_ON_RREG(VCN, inst_idx, regUVD_POWER_STATUS,
+					UVD_PGFSM_CONFIG__UVDM_UVDU_PWR_ON,
+					UVD_POWER_STATUS__UVD_POWER_STATUS_MASK);
+			}
+		} else {
+			/* unpause dpg, no need to wait */
+			reg_data &= ~UVD_DPG_PAUSE__NJ_PAUSE_DPG_REQ_MASK;
+			WREG32_SOC15(VCN, inst_idx, regUVD_DPG_PAUSE, reg_data);
+		}
+		adev->vcn.inst[inst_idx].pause_state.fw_based = new_state->fw_based;
+	}
+
+	return 0;
+}
+
+/**
+ * vcn_v4_0_5_unified_ring_get_rptr - get unified read pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Returns the current hardware unified read pointer
+ */
+static uint64_t vcn_v4_0_5_unified_ring_get_rptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	if (ring != &adev->vcn.inst[ring->me].ring_enc[0])
+		DRM_ERROR("wrong ring id is identified in %s", __func__);
+
+	return RREG32_SOC15(VCN, ring->me, regUVD_RB_RPTR);
+}
+
+/**
+ * vcn_v4_0_5_unified_ring_get_wptr - get unified write pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Returns the current hardware unified write pointer
+ */
+static uint64_t vcn_v4_0_5_unified_ring_get_wptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	if (ring != &adev->vcn.inst[ring->me].ring_enc[0])
+		DRM_ERROR("wrong ring id is identified in %s", __func__);
+
+	if (ring->use_doorbell)
+		return *ring->wptr_cpu_addr;
+	else
+		return RREG32_SOC15(VCN, ring->me, regUVD_RB_WPTR);
+}
+
+/**
+ * vcn_v4_0_5_unified_ring_set_wptr - set enc write pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Commits the enc write pointer to the hardware
+ */
+static void vcn_v4_0_5_unified_ring_set_wptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	if (ring != &adev->vcn.inst[ring->me].ring_enc[0])
+		DRM_ERROR("wrong ring id is identified in %s", __func__);
+
+	if (ring->use_doorbell) {
+		*ring->wptr_cpu_addr = lower_32_bits(ring->wptr);
+		WDOORBELL32(ring->doorbell_index, lower_32_bits(ring->wptr));
+	} else {
+		WREG32_SOC15(VCN, ring->me, regUVD_RB_WPTR, lower_32_bits(ring->wptr));
+	}
+}
+
+static int vcn_v4_0_5_ring_reset(struct amdgpu_ring *ring,
+				 unsigned int vmid,
+				 struct amdgpu_fence *timedout_fence)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct amdgpu_vcn_inst *vinst = &adev->vcn.inst[ring->me];
+	int r;
+
+	amdgpu_ring_reset_helper_begin(ring, timedout_fence);
+	r = vcn_v4_0_5_stop(vinst);
+	if (r)
+		return r;
+	r = vcn_v4_0_5_start(vinst);
+	if (r)
+		return r;
+	return amdgpu_ring_reset_helper_end(ring, timedout_fence);
+}
+
+static struct amdgpu_ring_funcs vcn_v4_0_5_unified_ring_vm_funcs = {
+	.type = AMDGPU_RING_TYPE_VCN_ENC,
+	.align_mask = 0x3f,
+	.nop = VCN_ENC_CMD_NO_OP,
+	.get_rptr = vcn_v4_0_5_unified_ring_get_rptr,
+	.get_wptr = vcn_v4_0_5_unified_ring_get_wptr,
+	.set_wptr = vcn_v4_0_5_unified_ring_set_wptr,
+	.emit_frame_size =
+		SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 +
+		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 4 +
+		4 + /* vcn_v2_0_enc_ring_emit_vm_flush */
+		5 + 5 + /* vcn_v2_0_enc_ring_emit_fence x2 vm fence */
+		1, /* vcn_v2_0_enc_ring_insert_end */
+	.emit_ib_size = 5, /* vcn_v2_0_enc_ring_emit_ib */
+	.emit_ib = vcn_v2_0_enc_ring_emit_ib,
+	.emit_fence = vcn_v2_0_enc_ring_emit_fence,
+	.emit_vm_flush = vcn_v2_0_enc_ring_emit_vm_flush,
+	.test_ring = amdgpu_vcn_enc_ring_test_ring,
+	.test_ib = amdgpu_vcn_unified_ring_test_ib,
+	.insert_nop = amdgpu_ring_insert_nop,
+	.insert_end = vcn_v2_0_enc_ring_insert_end,
+	.pad_ib = amdgpu_ring_generic_pad_ib,
+	.begin_use = amdgpu_vcn_ring_begin_use,
+	.end_use = amdgpu_vcn_ring_end_use,
+	.emit_wreg = vcn_v2_0_enc_ring_emit_wreg,
+	.emit_reg_wait = vcn_v2_0_enc_ring_emit_reg_wait,
+	.emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper,
+	.reset = vcn_v4_0_5_ring_reset,
+};
+
+/**
+ * vcn_v4_0_5_set_unified_ring_funcs - set unified ring functions
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Set unified ring functions
+ */
+static void vcn_v4_0_5_set_unified_ring_funcs(struct amdgpu_device *adev)
+{
+	int i;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		if (adev->vcn.harvest_config & (1 << i))
+			continue;
+
+		if (amdgpu_ip_version(adev, VCN_HWIP, 0) == IP_VERSION(4, 0, 5))
+			vcn_v4_0_5_unified_ring_vm_funcs.secure_submission_supported = true;
+
+		adev->vcn.inst[i].ring_enc[0].funcs = &vcn_v4_0_5_unified_ring_vm_funcs;
+		adev->vcn.inst[i].ring_enc[0].me = i;
+	}
+}
+
+/**
+ * vcn_v4_0_5_is_idle - check VCN block is idle
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block structure
+ *
+ * Check whether VCN block is idle
+ */
+static bool vcn_v4_0_5_is_idle(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, ret = 1;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		if (adev->vcn.harvest_config & (1 << i))
+			continue;
+
+		ret &= (RREG32_SOC15(VCN, i, regUVD_STATUS) == UVD_STATUS__IDLE);
+	}
+
+	return ret;
+}
+
+/**
+ * vcn_v4_0_5_wait_for_idle - wait for VCN block idle
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Wait for VCN block idle
+ */
+static int vcn_v4_0_5_wait_for_idle(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, ret = 0;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		if (adev->vcn.harvest_config & (1 << i))
+			continue;
+
+		ret = SOC15_WAIT_ON_RREG(VCN, i, regUVD_STATUS, UVD_STATUS__IDLE,
+			UVD_STATUS__IDLE);
+		if (ret)
+			return ret;
+	}
+
+	return ret;
+}
+
+/**
+ * vcn_v4_0_5_set_clockgating_state - set VCN block clockgating state
+ *
+ * @ip_block: amdgpu_ip_block pointer
+ * @state: clock gating state
+ *
+ * Set VCN block clockgating state
+ */
+static int vcn_v4_0_5_set_clockgating_state(struct amdgpu_ip_block *ip_block,
+					  enum amd_clockgating_state state)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	bool enable = state == AMD_CG_STATE_GATE;
+	int i;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		struct amdgpu_vcn_inst *vinst = &adev->vcn.inst[i];
+
+		if (adev->vcn.harvest_config & (1 << i))
+			continue;
+
+		if (enable) {
+			if (RREG32_SOC15(VCN, i, regUVD_STATUS) != UVD_STATUS__IDLE)
+				return -EBUSY;
+			vcn_v4_0_5_enable_clock_gating(vinst);
+		} else {
+			vcn_v4_0_5_disable_clock_gating(vinst);
+		}
+	}
+
+	return 0;
+}
+
+static int vcn_v4_0_5_set_pg_state(struct amdgpu_vcn_inst *vinst,
+				   enum amd_powergating_state state)
+{
+	int ret = 0;
+
+	if (state == vinst->cur_state)
+		return 0;
+
+	if (state == AMD_PG_STATE_GATE)
+		ret = vcn_v4_0_5_stop(vinst);
+	else
+		ret = vcn_v4_0_5_start(vinst);
+
+	if (!ret)
+		vinst->cur_state = state;
+
+	return ret;
+}
+
+/**
+ * vcn_v4_0_5_process_interrupt - process VCN block interrupt
+ *
+ * @adev: amdgpu_device pointer
+ * @source: interrupt sources
+ * @entry: interrupt entry from clients and sources
+ *
+ * Process VCN block interrupt
+ */
+static int vcn_v4_0_5_process_interrupt(struct amdgpu_device *adev, struct amdgpu_irq_src *source,
+		struct amdgpu_iv_entry *entry)
+{
+	uint32_t ip_instance;
+
+	switch (entry->client_id) {
+	case SOC15_IH_CLIENTID_VCN:
+		ip_instance = 0;
+		break;
+	case SOC15_IH_CLIENTID_VCN1:
+		ip_instance = 1;
+		break;
+	default:
+		DRM_ERROR("Unhandled client id: %d\n", entry->client_id);
+		return 0;
+	}
+
+	DRM_DEBUG("IH: VCN TRAP\n");
+
+	switch (entry->src_id) {
+	case VCN_4_0__SRCID__UVD_ENC_GENERAL_PURPOSE:
+		amdgpu_fence_process(&adev->vcn.inst[ip_instance].ring_enc[0]);
+		break;
+	case VCN_4_0__SRCID_UVD_POISON:
+		amdgpu_vcn_process_poison_irq(adev, source, entry);
+		break;
+	default:
+		DRM_ERROR("Unhandled interrupt: %d %d\n",
+			  entry->src_id, entry->src_data[0]);
+		break;
+	}
+
+	return 0;
+}
+
+static const struct amdgpu_irq_src_funcs vcn_v4_0_5_irq_funcs = {
+	.process = vcn_v4_0_5_process_interrupt,
+};
+
+/**
+ * vcn_v4_0_5_set_irq_funcs - set VCN block interrupt irq functions
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Set VCN block interrupt irq functions
+ */
+static void vcn_v4_0_5_set_irq_funcs(struct amdgpu_device *adev)
+{
+	int i;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		if (adev->vcn.harvest_config & (1 << i))
+			continue;
+
+		adev->vcn.inst[i].irq.num_types = adev->vcn.inst[i].num_enc_rings + 1;
+		adev->vcn.inst[i].irq.funcs = &vcn_v4_0_5_irq_funcs;
+	}
+}
+
+static const struct amd_ip_funcs vcn_v4_0_5_ip_funcs = {
+	.name = "vcn_v4_0_5",
+	.early_init = vcn_v4_0_5_early_init,
+	.sw_init = vcn_v4_0_5_sw_init,
+	.sw_fini = vcn_v4_0_5_sw_fini,
+	.hw_init = vcn_v4_0_5_hw_init,
+	.hw_fini = vcn_v4_0_5_hw_fini,
+	.suspend = vcn_v4_0_5_suspend,
+	.resume = vcn_v4_0_5_resume,
+	.is_idle = vcn_v4_0_5_is_idle,
+	.wait_for_idle = vcn_v4_0_5_wait_for_idle,
+	.set_clockgating_state = vcn_v4_0_5_set_clockgating_state,
+	.set_powergating_state = vcn_set_powergating_state,
+	.dump_ip_state = amdgpu_vcn_dump_ip_state,
+	.print_ip_state = amdgpu_vcn_print_ip_state,
+};
+
+const struct amdgpu_ip_block_version vcn_v4_0_5_ip_block = {
+	.type = AMD_IP_BLOCK_TYPE_VCN,
+	.major = 4,
+	.minor = 0,
+	.rev = 5,
+	.funcs = &vcn_v4_0_5_ip_funcs,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.h b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.h
new file mode 100644
index 000000000000..ff9b3d6f6a47
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.h
@@ -0,0 +1,35 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __VCN_V4_0_5_H__
+#define __VCN_V4_0_5_H__
+
+enum amdgpu_vcn_v4_0_5_sub_block {
+	AMDGPU_VCN_V4_0_5_VCPU_VCODEC = 0,
+
+	AMDGPU_VCN_V4_0_5_MAX_SUB_BLOCK,
+};
+
+extern const struct amdgpu_ip_block_version vcn_v4_0_5_ip_block;
+
+#endif /* __VCN_V4_0_5_H__ */
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c
new file mode 100644
index 000000000000..0202df5db1e1
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c
@@ -0,0 +1,1442 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/firmware.h>
+#include "amdgpu.h"
+#include "amdgpu_vcn.h"
+#include "amdgpu_pm.h"
+#include "soc15.h"
+#include "soc15d.h"
+#include "soc15_hw_ip.h"
+#include "vcn_v2_0.h"
+
+#include "vcn/vcn_5_0_0_offset.h"
+#include "vcn/vcn_5_0_0_sh_mask.h"
+#include "ivsrcid/vcn/irqsrcs_vcn_5_0.h"
+#include "vcn_v5_0_0.h"
+
+#include <drm/drm_drv.h>
+
+static const struct amdgpu_hwip_reg_entry vcn_reg_list_5_0[] = {
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_POWER_STATUS),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_STATUS),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_CONTEXT_ID),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_CONTEXT_ID2),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_GPCOM_VCPU_DATA0),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_GPCOM_VCPU_DATA1),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_GPCOM_VCPU_CMD),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_BASE_HI),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_BASE_LO),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_BASE_HI2),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_BASE_LO2),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_BASE_HI3),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_BASE_LO3),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_BASE_HI4),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_BASE_LO4),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_RPTR),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_WPTR),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_RPTR2),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_WPTR2),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_RPTR3),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_WPTR3),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_RPTR4),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_WPTR4),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_SIZE),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_SIZE2),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_SIZE3),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_SIZE4),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_DPG_LMA_CTL),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_DPG_LMA_DATA),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_DPG_LMA_MASK),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_DPG_PAUSE)
+};
+
+static int amdgpu_ih_clientid_vcns[] = {
+	SOC15_IH_CLIENTID_VCN,
+	SOC15_IH_CLIENTID_VCN1
+};
+
+static void vcn_v5_0_0_set_unified_ring_funcs(struct amdgpu_device *adev);
+static void vcn_v5_0_0_set_irq_funcs(struct amdgpu_device *adev);
+static int vcn_v5_0_0_set_pg_state(struct amdgpu_vcn_inst *vinst,
+				   enum amd_powergating_state state);
+static int vcn_v5_0_0_pause_dpg_mode(struct amdgpu_vcn_inst *vinst,
+				     struct dpg_pause_state *new_state);
+static void vcn_v5_0_0_unified_ring_set_wptr(struct amdgpu_ring *ring);
+
+/**
+ * vcn_v5_0_0_early_init - set function pointers and load microcode
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Set ring and irq function pointers
+ * Load microcode from filesystem
+ */
+static int vcn_v5_0_0_early_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, r;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i)
+		/* re-use enc ring as unified ring */
+		adev->vcn.inst[i].num_enc_rings = 1;
+
+	vcn_v5_0_0_set_unified_ring_funcs(adev);
+	vcn_v5_0_0_set_irq_funcs(adev);
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		adev->vcn.inst[i].set_pg_state = vcn_v5_0_0_set_pg_state;
+
+		r = amdgpu_vcn_early_init(adev, i);
+		if (r)
+			return r;
+	}
+
+	return 0;
+}
+
+/**
+ * vcn_v5_0_0_sw_init - sw init for VCN block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Load firmware and sw initialization
+ */
+static int vcn_v5_0_0_sw_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_ring *ring;
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, r;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+		struct amdgpu_vcn5_fw_shared *fw_shared;
+
+		if (adev->vcn.harvest_config & (1 << i))
+			continue;
+
+		r = amdgpu_vcn_sw_init(adev, i);
+		if (r)
+			return r;
+
+		amdgpu_vcn_setup_ucode(adev, i);
+
+		r = amdgpu_vcn_resume(adev, i);
+		if (r)
+			return r;
+
+		atomic_set(&adev->vcn.inst[i].sched_score, 0);
+
+		/* VCN UNIFIED TRAP */
+		r = amdgpu_irq_add_id(adev, amdgpu_ih_clientid_vcns[i],
+				VCN_5_0__SRCID__UVD_ENC_GENERAL_PURPOSE, &adev->vcn.inst[i].irq);
+		if (r)
+			return r;
+
+		/* VCN POISON TRAP */
+		r = amdgpu_irq_add_id(adev, amdgpu_ih_clientid_vcns[i],
+				VCN_5_0__SRCID_UVD_POISON, &adev->vcn.inst[i].irq);
+		if (r)
+			return r;
+
+		ring = &adev->vcn.inst[i].ring_enc[0];
+		ring->use_doorbell = true;
+		ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 2 + 8 * i;
+
+		ring->vm_hub = AMDGPU_MMHUB0(0);
+		sprintf(ring->name, "vcn_unified_%d", i);
+
+		r = amdgpu_ring_init(adev, ring, 512, &adev->vcn.inst[i].irq, 0,
+						AMDGPU_RING_PRIO_0, &adev->vcn.inst[i].sched_score);
+		if (r)
+			return r;
+
+		fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr;
+		fw_shared->present_flag_0 = cpu_to_le32(AMDGPU_FW_SHARED_FLAG_0_UNIFIED_QUEUE);
+		fw_shared->sq.is_enabled = 1;
+
+		if (amdgpu_vcnfw_log)
+			amdgpu_vcn_fwlog_init(&adev->vcn.inst[i]);
+
+		if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)
+			adev->vcn.inst[i].pause_dpg_mode = vcn_v5_0_0_pause_dpg_mode;
+	}
+
+	adev->vcn.supported_reset =
+		amdgpu_get_soft_full_reset_mask(&adev->vcn.inst[0].ring_enc[0]);
+	if (!amdgpu_sriov_vf(adev))
+		adev->vcn.supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE;
+
+	r = amdgpu_vcn_reg_dump_init(adev, vcn_reg_list_5_0, ARRAY_SIZE(vcn_reg_list_5_0));
+	if (r)
+		return r;
+
+	r = amdgpu_vcn_sysfs_reset_mask_init(adev);
+	if (r)
+		return r;
+
+	return 0;
+}
+
+/**
+ * vcn_v5_0_0_sw_fini - sw fini for VCN block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * VCN suspend and free up sw allocation
+ */
+static int vcn_v5_0_0_sw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, r, idx;
+
+	if (drm_dev_enter(adev_to_drm(adev), &idx)) {
+		for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+			struct amdgpu_vcn5_fw_shared *fw_shared;
+
+			if (adev->vcn.harvest_config & (1 << i))
+				continue;
+
+			fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr;
+			fw_shared->present_flag_0 = 0;
+			fw_shared->sq.is_enabled = 0;
+		}
+
+		drm_dev_exit(idx);
+	}
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+		r = amdgpu_vcn_suspend(adev, i);
+		if (r)
+			return r;
+	}
+
+	amdgpu_vcn_sysfs_reset_mask_fini(adev);
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++)
+		amdgpu_vcn_sw_fini(adev, i);
+
+	return 0;
+}
+
+/**
+ * vcn_v5_0_0_hw_init - start and test VCN block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Initialize the hardware, boot up the VCPU and do some testing
+ */
+static int vcn_v5_0_0_hw_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	struct amdgpu_ring *ring;
+	int i, r;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		if (adev->vcn.harvest_config & (1 << i))
+			continue;
+
+		ring = &adev->vcn.inst[i].ring_enc[0];
+
+		adev->nbio.funcs->vcn_doorbell_range(adev, ring->use_doorbell,
+			((adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 8 * i), i);
+
+		r = amdgpu_ring_test_helper(ring);
+		if (r)
+			return r;
+	}
+
+	return 0;
+}
+
+/**
+ * vcn_v5_0_0_hw_fini - stop the hardware block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Stop the VCN block, mark ring as not ready any more
+ */
+static int vcn_v5_0_0_hw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		struct amdgpu_vcn_inst *vinst = &adev->vcn.inst[i];
+
+		if (adev->vcn.harvest_config & (1 << i))
+			continue;
+
+		cancel_delayed_work_sync(&vinst->idle_work);
+
+		if (!amdgpu_sriov_vf(adev)) {
+			if ((adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) ||
+			    (vinst->cur_state != AMD_PG_STATE_GATE &&
+			     RREG32_SOC15(VCN, i, regUVD_STATUS))) {
+				vinst->set_pg_state(vinst, AMD_PG_STATE_GATE);
+			}
+		}
+	}
+
+	return 0;
+}
+
+/**
+ * vcn_v5_0_0_suspend - suspend VCN block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * HW fini and suspend VCN block
+ */
+static int vcn_v5_0_0_suspend(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int r, i;
+
+	r = vcn_v5_0_0_hw_fini(ip_block);
+	if (r)
+		return r;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+		r = amdgpu_vcn_suspend(ip_block->adev, i);
+		if (r)
+			return r;
+	}
+
+	return r;
+}
+
+/**
+ * vcn_v5_0_0_resume - resume VCN block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Resume firmware and hw init VCN block
+ */
+static int vcn_v5_0_0_resume(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int r, i;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+		r = amdgpu_vcn_resume(ip_block->adev, i);
+		if (r)
+			return r;
+	}
+
+	r = vcn_v5_0_0_hw_init(ip_block);
+
+	return r;
+}
+
+/**
+ * vcn_v5_0_0_mc_resume - memory controller programming
+ *
+ * @vinst: VCN instance
+ *
+ * Let the VCN memory controller know it's offsets
+ */
+static void vcn_v5_0_0_mc_resume(struct amdgpu_vcn_inst *vinst)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int inst = vinst->inst;
+	uint32_t offset, size;
+	const struct common_firmware_header *hdr;
+
+	hdr = (const struct common_firmware_header *)adev->vcn.inst[inst].fw->data;
+	size = AMDGPU_GPU_PAGE_ALIGN(le32_to_cpu(hdr->ucode_size_bytes) + 8);
+
+	/* cache window 0: fw */
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
+		WREG32_SOC15(VCN, inst, regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW,
+			(adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + inst].tmr_mc_addr_lo));
+		WREG32_SOC15(VCN, inst, regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH,
+			(adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + inst].tmr_mc_addr_hi));
+		WREG32_SOC15(VCN, inst, regUVD_VCPU_CACHE_OFFSET0, 0);
+		offset = 0;
+	} else {
+		WREG32_SOC15(VCN, inst, regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW,
+			lower_32_bits(adev->vcn.inst[inst].gpu_addr));
+		WREG32_SOC15(VCN, inst, regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH,
+			upper_32_bits(adev->vcn.inst[inst].gpu_addr));
+		offset = size;
+		WREG32_SOC15(VCN, inst, regUVD_VCPU_CACHE_OFFSET0, AMDGPU_UVD_FIRMWARE_OFFSET >> 3);
+	}
+	WREG32_SOC15(VCN, inst, regUVD_VCPU_CACHE_SIZE0, size);
+
+	/* cache window 1: stack */
+	WREG32_SOC15(VCN, inst, regUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW,
+		lower_32_bits(adev->vcn.inst[inst].gpu_addr + offset));
+	WREG32_SOC15(VCN, inst, regUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH,
+		upper_32_bits(adev->vcn.inst[inst].gpu_addr + offset));
+	WREG32_SOC15(VCN, inst, regUVD_VCPU_CACHE_OFFSET1, 0);
+	WREG32_SOC15(VCN, inst, regUVD_VCPU_CACHE_SIZE1, AMDGPU_VCN_STACK_SIZE);
+
+	/* cache window 2: context */
+	WREG32_SOC15(VCN, inst, regUVD_LMI_VCPU_CACHE2_64BIT_BAR_LOW,
+		lower_32_bits(adev->vcn.inst[inst].gpu_addr + offset + AMDGPU_VCN_STACK_SIZE));
+	WREG32_SOC15(VCN, inst, regUVD_LMI_VCPU_CACHE2_64BIT_BAR_HIGH,
+		upper_32_bits(adev->vcn.inst[inst].gpu_addr + offset + AMDGPU_VCN_STACK_SIZE));
+	WREG32_SOC15(VCN, inst, regUVD_VCPU_CACHE_OFFSET2, 0);
+	WREG32_SOC15(VCN, inst, regUVD_VCPU_CACHE_SIZE2, AMDGPU_VCN_CONTEXT_SIZE);
+
+	/* non-cache window */
+	WREG32_SOC15(VCN, inst, regUVD_LMI_VCPU_NC0_64BIT_BAR_LOW,
+		lower_32_bits(adev->vcn.inst[inst].fw_shared.gpu_addr));
+	WREG32_SOC15(VCN, inst, regUVD_LMI_VCPU_NC0_64BIT_BAR_HIGH,
+		upper_32_bits(adev->vcn.inst[inst].fw_shared.gpu_addr));
+	WREG32_SOC15(VCN, inst, regUVD_VCPU_NONCACHE_OFFSET0, 0);
+	WREG32_SOC15(VCN, inst, regUVD_VCPU_NONCACHE_SIZE0,
+		AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_vcn5_fw_shared)));
+}
+
+/**
+ * vcn_v5_0_0_mc_resume_dpg_mode - memory controller programming for dpg mode
+ *
+ * @vinst: VCN instance
+ * @indirect: indirectly write sram
+ *
+ * Let the VCN memory controller know it's offsets with dpg mode
+ */
+static void vcn_v5_0_0_mc_resume_dpg_mode(struct amdgpu_vcn_inst *vinst,
+					  bool indirect)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int inst_idx = vinst->inst;
+	uint32_t offset, size;
+	const struct common_firmware_header *hdr;
+
+	hdr = (const struct common_firmware_header *)adev->vcn.inst[inst_idx].fw->data;
+	size = AMDGPU_GPU_PAGE_ALIGN(le32_to_cpu(hdr->ucode_size_bytes) + 8);
+
+	/* cache window 0: fw */
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
+		if (!indirect) {
+			WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+				VCN, inst_idx, regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW),
+				(adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + inst_idx].tmr_mc_addr_lo), 0, indirect);
+			WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+				VCN, inst_idx, regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH),
+				(adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + inst_idx].tmr_mc_addr_hi), 0, indirect);
+			WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+				VCN, inst_idx, regUVD_VCPU_CACHE_OFFSET0), 0, 0, indirect);
+		} else {
+			WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+				VCN, inst_idx, regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW), 0, 0, indirect);
+			WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+				VCN, inst_idx, regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH), 0, 0, indirect);
+			WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+				VCN, inst_idx, regUVD_VCPU_CACHE_OFFSET0), 0, 0, indirect);
+		}
+		offset = 0;
+	} else {
+		WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW),
+			lower_32_bits(adev->vcn.inst[inst_idx].gpu_addr), 0, indirect);
+		WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH),
+			upper_32_bits(adev->vcn.inst[inst_idx].gpu_addr), 0, indirect);
+		offset = size;
+		WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_VCPU_CACHE_OFFSET0),
+			AMDGPU_UVD_FIRMWARE_OFFSET >> 3, 0, indirect);
+	}
+
+	if (!indirect)
+		WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_VCPU_CACHE_SIZE0), size, 0, indirect);
+	else
+		WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_VCPU_CACHE_SIZE0), 0, 0, indirect);
+
+	/* cache window 1: stack */
+	if (!indirect) {
+		WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW),
+			lower_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset), 0, indirect);
+		WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH),
+			upper_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset), 0, indirect);
+		WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_VCPU_CACHE_OFFSET1), 0, 0, indirect);
+	} else {
+		WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW), 0, 0, indirect);
+		WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH), 0, 0, indirect);
+		WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_VCPU_CACHE_OFFSET1), 0, 0, indirect);
+	}
+		WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+			VCN, inst_idx, regUVD_VCPU_CACHE_SIZE1), AMDGPU_VCN_STACK_SIZE, 0, indirect);
+
+	/* cache window 2: context */
+	WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_LMI_VCPU_CACHE2_64BIT_BAR_LOW),
+		lower_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset + AMDGPU_VCN_STACK_SIZE), 0, indirect);
+	WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_LMI_VCPU_CACHE2_64BIT_BAR_HIGH),
+		upper_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset + AMDGPU_VCN_STACK_SIZE), 0, indirect);
+	WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_VCPU_CACHE_OFFSET2), 0, 0, indirect);
+	WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_VCPU_CACHE_SIZE2), AMDGPU_VCN_CONTEXT_SIZE, 0, indirect);
+
+	/* non-cache window */
+	WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_LMI_VCPU_NC0_64BIT_BAR_LOW),
+		lower_32_bits(adev->vcn.inst[inst_idx].fw_shared.gpu_addr), 0, indirect);
+	WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_LMI_VCPU_NC0_64BIT_BAR_HIGH),
+		upper_32_bits(adev->vcn.inst[inst_idx].fw_shared.gpu_addr), 0, indirect);
+	WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_VCPU_NONCACHE_OFFSET0), 0, 0, indirect);
+	WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_VCPU_NONCACHE_SIZE0),
+		AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_vcn5_fw_shared)), 0, indirect);
+
+	/* VCN global tiling registers */
+	WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_GFX10_ADDR_CONFIG),
+		adev->gfx.config.gb_addr_config, 0, indirect);
+
+	return;
+}
+
+/**
+ * vcn_v5_0_0_disable_static_power_gating - disable VCN static power gating
+ *
+ * @vinst: VCN instance
+ *
+ * Disable static power gating for VCN block
+ */
+static void vcn_v5_0_0_disable_static_power_gating(struct amdgpu_vcn_inst *vinst)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int inst = vinst->inst;
+	uint32_t data = 0;
+
+	if (adev->pg_flags & AMD_PG_SUPPORT_VCN) {
+		data = 1 << UVD_IPX_DLDO_CONFIG__ONO2_PWR_CONFIG__SHIFT;
+		WREG32_SOC15(VCN, inst, regUVD_IPX_DLDO_CONFIG, data);
+		SOC15_WAIT_ON_RREG(VCN, inst, regUVD_IPX_DLDO_STATUS, 0,
+				UVD_IPX_DLDO_STATUS__ONO2_PWR_STATUS_MASK);
+
+		data = 2 << UVD_IPX_DLDO_CONFIG__ONO3_PWR_CONFIG__SHIFT;
+		WREG32_SOC15(VCN, inst, regUVD_IPX_DLDO_CONFIG, data);
+		SOC15_WAIT_ON_RREG(VCN, inst, regUVD_IPX_DLDO_STATUS,
+				1 << UVD_IPX_DLDO_STATUS__ONO3_PWR_STATUS__SHIFT,
+				UVD_IPX_DLDO_STATUS__ONO3_PWR_STATUS_MASK);
+
+		data = 2 << UVD_IPX_DLDO_CONFIG__ONO4_PWR_CONFIG__SHIFT;
+		WREG32_SOC15(VCN, inst, regUVD_IPX_DLDO_CONFIG, data);
+		SOC15_WAIT_ON_RREG(VCN, inst, regUVD_IPX_DLDO_STATUS,
+				1 << UVD_IPX_DLDO_STATUS__ONO4_PWR_STATUS__SHIFT,
+				UVD_IPX_DLDO_STATUS__ONO4_PWR_STATUS_MASK);
+
+		data = 2 << UVD_IPX_DLDO_CONFIG__ONO5_PWR_CONFIG__SHIFT;
+		WREG32_SOC15(VCN, inst, regUVD_IPX_DLDO_CONFIG, data);
+		SOC15_WAIT_ON_RREG(VCN, inst, regUVD_IPX_DLDO_STATUS,
+				1 << UVD_IPX_DLDO_STATUS__ONO5_PWR_STATUS__SHIFT,
+				UVD_IPX_DLDO_STATUS__ONO5_PWR_STATUS_MASK);
+	} else {
+		data = 1 << UVD_IPX_DLDO_CONFIG__ONO2_PWR_CONFIG__SHIFT;
+		WREG32_SOC15(VCN, inst, regUVD_IPX_DLDO_CONFIG, data);
+		SOC15_WAIT_ON_RREG(VCN, inst, regUVD_IPX_DLDO_STATUS, 0,
+				UVD_IPX_DLDO_STATUS__ONO2_PWR_STATUS_MASK);
+
+		data = 1 << UVD_IPX_DLDO_CONFIG__ONO3_PWR_CONFIG__SHIFT;
+		WREG32_SOC15(VCN, inst, regUVD_IPX_DLDO_CONFIG, data);
+		SOC15_WAIT_ON_RREG(VCN, inst, regUVD_IPX_DLDO_STATUS, 0,
+				UVD_IPX_DLDO_STATUS__ONO3_PWR_STATUS_MASK);
+
+		data = 1 << UVD_IPX_DLDO_CONFIG__ONO4_PWR_CONFIG__SHIFT;
+		WREG32_SOC15(VCN, inst, regUVD_IPX_DLDO_CONFIG, data);
+		SOC15_WAIT_ON_RREG(VCN, inst, regUVD_IPX_DLDO_STATUS, 0,
+				UVD_IPX_DLDO_STATUS__ONO4_PWR_STATUS_MASK);
+
+		data = 1 << UVD_IPX_DLDO_CONFIG__ONO5_PWR_CONFIG__SHIFT;
+		WREG32_SOC15(VCN, inst, regUVD_IPX_DLDO_CONFIG, data);
+		SOC15_WAIT_ON_RREG(VCN, inst, regUVD_IPX_DLDO_STATUS, 0,
+				UVD_IPX_DLDO_STATUS__ONO5_PWR_STATUS_MASK);
+	}
+
+	data = RREG32_SOC15(VCN, inst, regUVD_POWER_STATUS);
+	data &= ~0x103;
+	if (adev->pg_flags & AMD_PG_SUPPORT_VCN)
+		data |= UVD_PGFSM_CONFIG__UVDM_UVDU_PWR_ON |
+			UVD_POWER_STATUS__UVD_PG_EN_MASK;
+
+	WREG32_SOC15(VCN, inst, regUVD_POWER_STATUS, data);
+	return;
+}
+
+/**
+ * vcn_v5_0_0_enable_static_power_gating - enable VCN static power gating
+ *
+ * @vinst: VCN instance
+ *
+ * Enable static power gating for VCN block
+ */
+static void vcn_v5_0_0_enable_static_power_gating(struct amdgpu_vcn_inst *vinst)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int inst = vinst->inst;
+	uint32_t data;
+
+	if (adev->pg_flags & AMD_PG_SUPPORT_VCN) {
+		/* Before power off, this indicator has to be turned on */
+		data = RREG32_SOC15(VCN, inst, regUVD_POWER_STATUS);
+		data &= ~UVD_POWER_STATUS__UVD_POWER_STATUS_MASK;
+		data |= UVD_POWER_STATUS__UVD_POWER_STATUS_TILES_OFF;
+		WREG32_SOC15(VCN, inst, regUVD_POWER_STATUS, data);
+
+		data = 2 << UVD_IPX_DLDO_CONFIG__ONO5_PWR_CONFIG__SHIFT;
+		WREG32_SOC15(VCN, inst, regUVD_IPX_DLDO_CONFIG, data);
+		SOC15_WAIT_ON_RREG(VCN, inst, regUVD_IPX_DLDO_STATUS,
+				1 << UVD_IPX_DLDO_STATUS__ONO5_PWR_STATUS__SHIFT,
+				UVD_IPX_DLDO_STATUS__ONO5_PWR_STATUS_MASK);
+
+		data = 2 << UVD_IPX_DLDO_CONFIG__ONO4_PWR_CONFIG__SHIFT;
+		WREG32_SOC15(VCN, inst, regUVD_IPX_DLDO_CONFIG, data);
+		SOC15_WAIT_ON_RREG(VCN, inst, regUVD_IPX_DLDO_STATUS,
+				1 << UVD_IPX_DLDO_STATUS__ONO4_PWR_STATUS__SHIFT,
+				UVD_IPX_DLDO_STATUS__ONO4_PWR_STATUS_MASK);
+
+		data = 2 << UVD_IPX_DLDO_CONFIG__ONO3_PWR_CONFIG__SHIFT;
+		WREG32_SOC15(VCN, inst, regUVD_IPX_DLDO_CONFIG, data);
+		SOC15_WAIT_ON_RREG(VCN, inst, regUVD_IPX_DLDO_STATUS,
+				1 << UVD_IPX_DLDO_STATUS__ONO3_PWR_STATUS__SHIFT,
+				UVD_IPX_DLDO_STATUS__ONO3_PWR_STATUS_MASK);
+
+		data = 2 << UVD_IPX_DLDO_CONFIG__ONO2_PWR_CONFIG__SHIFT;
+		WREG32_SOC15(VCN, inst, regUVD_IPX_DLDO_CONFIG, data);
+		SOC15_WAIT_ON_RREG(VCN, inst, regUVD_IPX_DLDO_STATUS,
+				1 << UVD_IPX_DLDO_STATUS__ONO2_PWR_STATUS__SHIFT,
+				UVD_IPX_DLDO_STATUS__ONO2_PWR_STATUS_MASK);
+	}
+	return;
+}
+
+/**
+ * vcn_v5_0_0_disable_clock_gating - disable VCN clock gating
+ *
+ * @vinst: VCN instance
+ *
+ * Disable clock gating for VCN block
+ */
+static void vcn_v5_0_0_disable_clock_gating(struct amdgpu_vcn_inst *vinst)
+{
+	return;
+}
+
+#if 0
+/**
+ * vcn_v5_0_0_disable_clock_gating_dpg_mode - disable VCN clock gating dpg mode
+ *
+ * @vinst: VCN instance
+ * @sram_sel: sram select
+ * @indirect: indirectly write sram
+ *
+ * Disable clock gating for VCN block with dpg mode
+ */
+static void vcn_v5_0_0_disable_clock_gating_dpg_mode(struct amdgpu_vcn_inst *vinst,
+						     uint8_t sram_sel,
+						     uint8_t indirect)
+{
+	return;
+}
+#endif
+
+/**
+ * vcn_v5_0_0_enable_clock_gating - enable VCN clock gating
+ *
+ * @vinst: VCN instance
+ *
+ * Enable clock gating for VCN block
+ */
+static void vcn_v5_0_0_enable_clock_gating(struct amdgpu_vcn_inst *vinst)
+{
+	return;
+}
+
+/**
+ * vcn_v5_0_0_start_dpg_mode - VCN start with dpg mode
+ *
+ * @vinst: VCN instance
+ * @indirect: indirectly write sram
+ *
+ * Start VCN block with dpg mode
+ */
+static int vcn_v5_0_0_start_dpg_mode(struct amdgpu_vcn_inst *vinst,
+				     bool indirect)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int inst_idx = vinst->inst;
+	struct amdgpu_vcn5_fw_shared *fw_shared = adev->vcn.inst[inst_idx].fw_shared.cpu_addr;
+	struct amdgpu_ring *ring;
+	uint32_t tmp;
+	int ret;
+
+	/* disable register anti-hang mechanism */
+	WREG32_P(SOC15_REG_OFFSET(VCN, inst_idx, regUVD_POWER_STATUS), 1,
+		~UVD_POWER_STATUS__UVD_POWER_STATUS_MASK);
+
+	/* enable dynamic power gating mode */
+	tmp = RREG32_SOC15(VCN, inst_idx, regUVD_POWER_STATUS);
+	tmp |= UVD_POWER_STATUS__UVD_PG_MODE_MASK;
+	tmp |= UVD_POWER_STATUS__UVD_PG_EN_MASK;
+	WREG32_SOC15(VCN, inst_idx, regUVD_POWER_STATUS, tmp);
+
+	if (indirect)
+		adev->vcn.inst[inst_idx].dpg_sram_curr_addr = (uint32_t *)adev->vcn.inst[inst_idx].dpg_sram_cpu_addr;
+
+	/* enable VCPU clock */
+	tmp = (0xFF << UVD_VCPU_CNTL__PRB_TIMEOUT_VAL__SHIFT);
+	tmp |= UVD_VCPU_CNTL__CLK_EN_MASK | UVD_VCPU_CNTL__BLK_RST_MASK;
+	WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_VCPU_CNTL), tmp, 0, indirect);
+
+	/* disable master interrupt */
+	WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_MASTINT_EN), 0, 0, indirect);
+
+	/* setup regUVD_LMI_CTRL */
+	tmp = (UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK |
+		UVD_LMI_CTRL__REQ_MODE_MASK |
+		UVD_LMI_CTRL__CRC_RESET_MASK |
+		UVD_LMI_CTRL__MASK_MC_URGENT_MASK |
+		UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK |
+		UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK |
+		(8 << UVD_LMI_CTRL__WRITE_CLEAN_TIMER__SHIFT) |
+		0x00100000L);
+	WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_LMI_CTRL), tmp, 0, indirect);
+
+	vcn_v5_0_0_mc_resume_dpg_mode(vinst, indirect);
+
+	tmp = (0xFF << UVD_VCPU_CNTL__PRB_TIMEOUT_VAL__SHIFT);
+	tmp |= UVD_VCPU_CNTL__CLK_EN_MASK;
+	WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_VCPU_CNTL), tmp, 0, indirect);
+
+	/* enable LMI MC and UMC channels */
+	tmp = 0x1f << UVD_LMI_CTRL2__RE_OFLD_MIF_WR_REQ_NUM__SHIFT;
+	WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_LMI_CTRL2), tmp, 0, indirect);
+
+	/* enable master interrupt */
+	WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+		VCN, inst_idx, regUVD_MASTINT_EN),
+		UVD_MASTINT_EN__VCPU_EN_MASK, 0, indirect);
+
+	if (indirect) {
+		ret = amdgpu_vcn_psp_update_sram(adev, inst_idx, 0);
+		if (ret) {
+			dev_err(adev->dev, "%s: vcn sram load failed %d\n", __func__, ret);
+			return ret;
+		}
+	}
+
+	ring = &adev->vcn.inst[inst_idx].ring_enc[0];
+
+	WREG32_SOC15(VCN, inst_idx, regUVD_RB_BASE_LO, ring->gpu_addr);
+	WREG32_SOC15(VCN, inst_idx, regUVD_RB_BASE_HI, upper_32_bits(ring->gpu_addr));
+	WREG32_SOC15(VCN, inst_idx, regUVD_RB_SIZE, ring->ring_size / 4);
+
+	tmp = RREG32_SOC15(VCN, inst_idx, regVCN_RB_ENABLE);
+	tmp &= ~(VCN_RB_ENABLE__RB1_EN_MASK);
+	WREG32_SOC15(VCN, inst_idx, regVCN_RB_ENABLE, tmp);
+	fw_shared->sq.queue_mode |= FW_QUEUE_RING_RESET;
+	WREG32_SOC15(VCN, inst_idx, regUVD_RB_RPTR, 0);
+	WREG32_SOC15(VCN, inst_idx, regUVD_RB_WPTR, 0);
+
+	tmp = RREG32_SOC15(VCN, inst_idx, regUVD_RB_RPTR);
+	WREG32_SOC15(VCN, inst_idx, regUVD_RB_WPTR, tmp);
+	ring->wptr = RREG32_SOC15(VCN, inst_idx, regUVD_RB_WPTR);
+
+	tmp = RREG32_SOC15(VCN, inst_idx, regVCN_RB_ENABLE);
+	tmp |= VCN_RB_ENABLE__RB1_EN_MASK;
+	WREG32_SOC15(VCN, inst_idx, regVCN_RB_ENABLE, tmp);
+	fw_shared->sq.queue_mode &= ~(FW_QUEUE_RING_RESET | FW_QUEUE_DPG_HOLD_OFF);
+
+	WREG32_SOC15(VCN, inst_idx, regVCN_RB1_DB_CTRL,
+		ring->doorbell_index << VCN_RB1_DB_CTRL__OFFSET__SHIFT |
+		VCN_RB1_DB_CTRL__EN_MASK);
+
+	/* Keeping one read-back to ensure all register writes are done,
+	 * otherwise it may introduce race conditions.
+	 */
+	RREG32_SOC15(VCN, inst_idx, regUVD_STATUS);
+
+	return 0;
+}
+
+/**
+ * vcn_v5_0_0_start - VCN start
+ *
+ * @vinst: VCN instance
+ *
+ * Start VCN block
+ */
+static int vcn_v5_0_0_start(struct amdgpu_vcn_inst *vinst)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int i = vinst->inst;
+	struct amdgpu_vcn5_fw_shared *fw_shared;
+	struct amdgpu_ring *ring;
+	uint32_t tmp;
+	int j, k, r;
+
+	if (adev->vcn.harvest_config & (1 << i))
+		return 0;
+
+	if (adev->pm.dpm_enabled)
+		amdgpu_dpm_enable_vcn(adev, true, i);
+
+	fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr;
+
+	if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)
+		return vcn_v5_0_0_start_dpg_mode(vinst, adev->vcn.inst[i].indirect_sram);
+
+	/* disable VCN power gating */
+	vcn_v5_0_0_disable_static_power_gating(vinst);
+
+	/* set VCN status busy */
+	tmp = RREG32_SOC15(VCN, i, regUVD_STATUS) | UVD_STATUS__UVD_BUSY;
+	WREG32_SOC15(VCN, i, regUVD_STATUS, tmp);
+
+	/* enable VCPU clock */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL),
+		 UVD_VCPU_CNTL__CLK_EN_MASK, ~UVD_VCPU_CNTL__CLK_EN_MASK);
+
+	/* disable master interrupt */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_MASTINT_EN), 0,
+		 ~UVD_MASTINT_EN__VCPU_EN_MASK);
+
+	/* enable LMI MC and UMC channels */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_LMI_CTRL2), 0,
+		 ~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK);
+
+	tmp = RREG32_SOC15(VCN, i, regUVD_SOFT_RESET);
+	tmp &= ~UVD_SOFT_RESET__LMI_SOFT_RESET_MASK;
+	tmp &= ~UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK;
+	WREG32_SOC15(VCN, i, regUVD_SOFT_RESET, tmp);
+
+	/* setup regUVD_LMI_CTRL */
+	tmp = RREG32_SOC15(VCN, i, regUVD_LMI_CTRL);
+	WREG32_SOC15(VCN, i, regUVD_LMI_CTRL, tmp |
+		     UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK |
+		     UVD_LMI_CTRL__MASK_MC_URGENT_MASK |
+		     UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK |
+		     UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK);
+
+	vcn_v5_0_0_mc_resume(vinst);
+
+	/* VCN global tiling registers */
+	WREG32_SOC15(VCN, i, regUVD_GFX10_ADDR_CONFIG,
+		     adev->gfx.config.gb_addr_config);
+
+	/* unblock VCPU register access */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_RB_ARB_CTRL), 0,
+		 ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK);
+
+	/* release VCPU reset to boot */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), 0,
+		 ~UVD_VCPU_CNTL__BLK_RST_MASK);
+
+	for (j = 0; j < 10; ++j) {
+		uint32_t status;
+
+		for (k = 0; k < 100; ++k) {
+			status = RREG32_SOC15(VCN, i, regUVD_STATUS);
+			if (status & 2)
+				break;
+			mdelay(10);
+			if (amdgpu_emu_mode == 1)
+				msleep(1);
+		}
+
+		if (amdgpu_emu_mode == 1) {
+			r = -1;
+			if (status & 2) {
+				r = 0;
+				break;
+			}
+		} else {
+			r = 0;
+			if (status & 2)
+				break;
+
+			dev_err(adev->dev,
+				"VCN[%d] is not responding, trying to reset the VCPU!!!\n", i);
+			WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL),
+				 UVD_VCPU_CNTL__BLK_RST_MASK,
+				 ~UVD_VCPU_CNTL__BLK_RST_MASK);
+			mdelay(10);
+			WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), 0,
+				 ~UVD_VCPU_CNTL__BLK_RST_MASK);
+
+			mdelay(10);
+			r = -1;
+		}
+	}
+
+	if (r) {
+		dev_err(adev->dev, "VCN[%d] is not responding, giving up!!!\n", i);
+		return r;
+	}
+
+	/* enable master interrupt */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_MASTINT_EN),
+		 UVD_MASTINT_EN__VCPU_EN_MASK,
+		 ~UVD_MASTINT_EN__VCPU_EN_MASK);
+
+	/* clear the busy bit of VCN_STATUS */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_STATUS), 0,
+		 ~(2 << UVD_STATUS__VCPU_REPORT__SHIFT));
+
+	ring = &adev->vcn.inst[i].ring_enc[0];
+	WREG32_SOC15(VCN, i, regVCN_RB1_DB_CTRL,
+		     ring->doorbell_index << VCN_RB1_DB_CTRL__OFFSET__SHIFT |
+		     VCN_RB1_DB_CTRL__EN_MASK);
+
+	WREG32_SOC15(VCN, i, regUVD_RB_BASE_LO, ring->gpu_addr);
+	WREG32_SOC15(VCN, i, regUVD_RB_BASE_HI, upper_32_bits(ring->gpu_addr));
+	WREG32_SOC15(VCN, i, regUVD_RB_SIZE, ring->ring_size / 4);
+
+	tmp = RREG32_SOC15(VCN, i, regVCN_RB_ENABLE);
+	tmp &= ~(VCN_RB_ENABLE__RB1_EN_MASK);
+	WREG32_SOC15(VCN, i, regVCN_RB_ENABLE, tmp);
+	fw_shared->sq.queue_mode |= FW_QUEUE_RING_RESET;
+	WREG32_SOC15(VCN, i, regUVD_RB_RPTR, 0);
+	WREG32_SOC15(VCN, i, regUVD_RB_WPTR, 0);
+
+	tmp = RREG32_SOC15(VCN, i, regUVD_RB_RPTR);
+	WREG32_SOC15(VCN, i, regUVD_RB_WPTR, tmp);
+	ring->wptr = RREG32_SOC15(VCN, i, regUVD_RB_WPTR);
+
+	tmp = RREG32_SOC15(VCN, i, regVCN_RB_ENABLE);
+	tmp |= VCN_RB_ENABLE__RB1_EN_MASK;
+	WREG32_SOC15(VCN, i, regVCN_RB_ENABLE, tmp);
+	fw_shared->sq.queue_mode &= ~(FW_QUEUE_RING_RESET | FW_QUEUE_DPG_HOLD_OFF);
+
+	/* Keeping one read-back to ensure all register writes are done,
+	 * otherwise it may introduce race conditions.
+	 */
+	RREG32_SOC15(VCN, i, regUVD_STATUS);
+
+	return 0;
+}
+
+/**
+ * vcn_v5_0_0_stop_dpg_mode - VCN stop with dpg mode
+ *
+ * @vinst: VCN instance
+ *
+ * Stop VCN block with dpg mode
+ */
+static void vcn_v5_0_0_stop_dpg_mode(struct amdgpu_vcn_inst *vinst)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int inst_idx = vinst->inst;
+	struct dpg_pause_state state = {.fw_based = VCN_DPG_STATE__UNPAUSE};
+	uint32_t tmp;
+
+	vcn_v5_0_0_pause_dpg_mode(vinst, &state);
+
+	/* Wait for power status to be 1 */
+	SOC15_WAIT_ON_RREG(VCN, inst_idx, regUVD_POWER_STATUS, 1,
+		UVD_POWER_STATUS__UVD_POWER_STATUS_MASK);
+
+	/* wait for read ptr to be equal to write ptr */
+	tmp = RREG32_SOC15(VCN, inst_idx, regUVD_RB_WPTR);
+	SOC15_WAIT_ON_RREG(VCN, inst_idx, regUVD_RB_RPTR, tmp, 0xFFFFFFFF);
+
+	/* disable dynamic power gating mode */
+	WREG32_P(SOC15_REG_OFFSET(VCN, inst_idx, regUVD_POWER_STATUS), 0,
+		~UVD_POWER_STATUS__UVD_PG_MODE_MASK);
+
+	/* Keeping one read-back to ensure all register writes are done,
+	 * otherwise it may introduce race conditions.
+	 */
+	RREG32_SOC15(VCN, inst_idx, regUVD_STATUS);
+
+	return;
+}
+
+/**
+ * vcn_v5_0_0_stop - VCN stop
+ *
+ * @vinst: VCN instance
+ *
+ * Stop VCN block
+ */
+static int vcn_v5_0_0_stop(struct amdgpu_vcn_inst *vinst)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int i = vinst->inst;
+	struct amdgpu_vcn5_fw_shared *fw_shared;
+	uint32_t tmp;
+	int r = 0;
+
+	if (adev->vcn.harvest_config & (1 << i))
+		return 0;
+
+	fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr;
+	fw_shared->sq.queue_mode |= FW_QUEUE_DPG_HOLD_OFF;
+
+	if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) {
+		vcn_v5_0_0_stop_dpg_mode(vinst);
+		r = 0;
+		goto done;
+	}
+
+	/* wait for vcn idle */
+	r = SOC15_WAIT_ON_RREG(VCN, i, regUVD_STATUS, UVD_STATUS__IDLE, 0x7);
+	if (r)
+		goto done;
+
+	tmp = UVD_LMI_STATUS__VCPU_LMI_WRITE_CLEAN_MASK |
+		UVD_LMI_STATUS__READ_CLEAN_MASK |
+		UVD_LMI_STATUS__WRITE_CLEAN_MASK |
+		UVD_LMI_STATUS__WRITE_CLEAN_RAW_MASK;
+	r = SOC15_WAIT_ON_RREG(VCN, i, regUVD_LMI_STATUS, tmp, tmp);
+	if (r)
+		goto done;
+
+	/* disable LMI UMC channel */
+	tmp = RREG32_SOC15(VCN, i, regUVD_LMI_CTRL2);
+	tmp |= UVD_LMI_CTRL2__STALL_ARB_UMC_MASK;
+	WREG32_SOC15(VCN, i, regUVD_LMI_CTRL2, tmp);
+	tmp = UVD_LMI_STATUS__UMC_READ_CLEAN_RAW_MASK |
+		UVD_LMI_STATUS__UMC_WRITE_CLEAN_RAW_MASK;
+	r = SOC15_WAIT_ON_RREG(VCN, i, regUVD_LMI_STATUS, tmp, tmp);
+	if (r)
+		goto done;
+
+	/* block VCPU register access */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_RB_ARB_CTRL),
+		 UVD_RB_ARB_CTRL__VCPU_DIS_MASK,
+		 ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK);
+
+	/* reset VCPU */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL),
+		 UVD_VCPU_CNTL__BLK_RST_MASK,
+		 ~UVD_VCPU_CNTL__BLK_RST_MASK);
+
+	/* disable VCPU clock */
+	WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), 0,
+		 ~(UVD_VCPU_CNTL__CLK_EN_MASK));
+
+	/* apply soft reset */
+	tmp = RREG32_SOC15(VCN, i, regUVD_SOFT_RESET);
+	tmp |= UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK;
+	WREG32_SOC15(VCN, i, regUVD_SOFT_RESET, tmp);
+	tmp = RREG32_SOC15(VCN, i, regUVD_SOFT_RESET);
+	tmp |= UVD_SOFT_RESET__LMI_SOFT_RESET_MASK;
+	WREG32_SOC15(VCN, i, regUVD_SOFT_RESET, tmp);
+
+	/* clear status */
+	WREG32_SOC15(VCN, i, regUVD_STATUS, 0);
+
+	/* enable VCN power gating */
+	vcn_v5_0_0_enable_static_power_gating(vinst);
+
+	/* Keeping one read-back to ensure all register writes are done,
+	 * otherwise it may introduce race conditions.
+	 */
+	RREG32_SOC15(VCN, i, regUVD_STATUS);
+
+done:
+	if (adev->pm.dpm_enabled)
+		amdgpu_dpm_enable_vcn(adev, false, i);
+
+	return r;
+}
+
+/**
+ * vcn_v5_0_0_pause_dpg_mode - VCN pause with dpg mode
+ *
+ * @vinst: VCN instance
+ * @new_state: pause state
+ *
+ * Pause dpg mode for VCN block
+ */
+static int vcn_v5_0_0_pause_dpg_mode(struct amdgpu_vcn_inst *vinst,
+				     struct dpg_pause_state *new_state)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int inst_idx = vinst->inst;
+	uint32_t reg_data = 0;
+	int ret_code;
+
+	/* pause/unpause if state is changed */
+	if (adev->vcn.inst[inst_idx].pause_state.fw_based != new_state->fw_based) {
+		DRM_DEV_DEBUG(adev->dev, "dpg pause state changed %d -> %d",
+			adev->vcn.inst[inst_idx].pause_state.fw_based,  new_state->fw_based);
+		reg_data = RREG32_SOC15(VCN, inst_idx, regUVD_DPG_PAUSE) &
+			(~UVD_DPG_PAUSE__NJ_PAUSE_DPG_ACK_MASK);
+
+		if (new_state->fw_based == VCN_DPG_STATE__PAUSE) {
+			ret_code = SOC15_WAIT_ON_RREG(VCN, inst_idx, regUVD_POWER_STATUS, 0x1,
+					UVD_POWER_STATUS__UVD_POWER_STATUS_MASK);
+
+			if (!ret_code) {
+				/* pause DPG */
+				reg_data |= UVD_DPG_PAUSE__NJ_PAUSE_DPG_REQ_MASK;
+				WREG32_SOC15(VCN, inst_idx, regUVD_DPG_PAUSE, reg_data);
+
+				/* wait for ACK */
+				SOC15_WAIT_ON_RREG(VCN, inst_idx, regUVD_DPG_PAUSE,
+					UVD_DPG_PAUSE__NJ_PAUSE_DPG_ACK_MASK,
+					UVD_DPG_PAUSE__NJ_PAUSE_DPG_ACK_MASK);
+			}
+		} else {
+			/* unpause dpg, no need to wait */
+			reg_data &= ~UVD_DPG_PAUSE__NJ_PAUSE_DPG_REQ_MASK;
+			WREG32_SOC15(VCN, inst_idx, regUVD_DPG_PAUSE, reg_data);
+		}
+		adev->vcn.inst[inst_idx].pause_state.fw_based = new_state->fw_based;
+	}
+
+	return 0;
+}
+
+/**
+ * vcn_v5_0_0_unified_ring_get_rptr - get unified read pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Returns the current hardware unified read pointer
+ */
+static uint64_t vcn_v5_0_0_unified_ring_get_rptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	if (ring != &adev->vcn.inst[ring->me].ring_enc[0])
+		DRM_ERROR("wrong ring id is identified in %s", __func__);
+
+	return RREG32_SOC15(VCN, ring->me, regUVD_RB_RPTR);
+}
+
+/**
+ * vcn_v5_0_0_unified_ring_get_wptr - get unified write pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Returns the current hardware unified write pointer
+ */
+static uint64_t vcn_v5_0_0_unified_ring_get_wptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	if (ring != &adev->vcn.inst[ring->me].ring_enc[0])
+		DRM_ERROR("wrong ring id is identified in %s", __func__);
+
+	if (ring->use_doorbell)
+		return *ring->wptr_cpu_addr;
+	else
+		return RREG32_SOC15(VCN, ring->me, regUVD_RB_WPTR);
+}
+
+/**
+ * vcn_v5_0_0_unified_ring_set_wptr - set enc write pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Commits the enc write pointer to the hardware
+ */
+static void vcn_v5_0_0_unified_ring_set_wptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	if (ring != &adev->vcn.inst[ring->me].ring_enc[0])
+		DRM_ERROR("wrong ring id is identified in %s", __func__);
+
+	if (ring->use_doorbell) {
+		*ring->wptr_cpu_addr = lower_32_bits(ring->wptr);
+		WDOORBELL32(ring->doorbell_index, lower_32_bits(ring->wptr));
+	} else {
+		WREG32_SOC15(VCN, ring->me, regUVD_RB_WPTR, lower_32_bits(ring->wptr));
+	}
+}
+
+static int vcn_v5_0_0_ring_reset(struct amdgpu_ring *ring,
+				 unsigned int vmid,
+				 struct amdgpu_fence *timedout_fence)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct amdgpu_vcn_inst *vinst = &adev->vcn.inst[ring->me];
+	int r;
+
+	amdgpu_ring_reset_helper_begin(ring, timedout_fence);
+	r = vcn_v5_0_0_stop(vinst);
+	if (r)
+		return r;
+	r = vcn_v5_0_0_start(vinst);
+	if (r)
+		return r;
+	return amdgpu_ring_reset_helper_end(ring, timedout_fence);
+}
+
+static const struct amdgpu_ring_funcs vcn_v5_0_0_unified_ring_vm_funcs = {
+	.type = AMDGPU_RING_TYPE_VCN_ENC,
+	.align_mask = 0x3f,
+	.nop = VCN_ENC_CMD_NO_OP,
+	.get_rptr = vcn_v5_0_0_unified_ring_get_rptr,
+	.get_wptr = vcn_v5_0_0_unified_ring_get_wptr,
+	.set_wptr = vcn_v5_0_0_unified_ring_set_wptr,
+	.emit_frame_size =
+		SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 +
+		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 4 +
+		4 + /* vcn_v2_0_enc_ring_emit_vm_flush */
+		5 + 5 + /* vcn_v2_0_enc_ring_emit_fence x2 vm fence */
+		1, /* vcn_v2_0_enc_ring_insert_end */
+	.emit_ib_size = 5, /* vcn_v2_0_enc_ring_emit_ib */
+	.emit_ib = vcn_v2_0_enc_ring_emit_ib,
+	.emit_fence = vcn_v2_0_enc_ring_emit_fence,
+	.emit_vm_flush = vcn_v2_0_enc_ring_emit_vm_flush,
+	.test_ring = amdgpu_vcn_enc_ring_test_ring,
+	.test_ib = amdgpu_vcn_unified_ring_test_ib,
+	.insert_nop = amdgpu_ring_insert_nop,
+	.insert_end = vcn_v2_0_enc_ring_insert_end,
+	.pad_ib = amdgpu_ring_generic_pad_ib,
+	.begin_use = amdgpu_vcn_ring_begin_use,
+	.end_use = amdgpu_vcn_ring_end_use,
+	.emit_wreg = vcn_v2_0_enc_ring_emit_wreg,
+	.emit_reg_wait = vcn_v2_0_enc_ring_emit_reg_wait,
+	.emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper,
+	.reset = vcn_v5_0_0_ring_reset,
+};
+
+/**
+ * vcn_v5_0_0_set_unified_ring_funcs - set unified ring functions
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Set unified ring functions
+ */
+static void vcn_v5_0_0_set_unified_ring_funcs(struct amdgpu_device *adev)
+{
+	int i;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		if (adev->vcn.harvest_config & (1 << i))
+			continue;
+
+		adev->vcn.inst[i].ring_enc[0].funcs = &vcn_v5_0_0_unified_ring_vm_funcs;
+		adev->vcn.inst[i].ring_enc[0].me = i;
+	}
+}
+
+/**
+ * vcn_v5_0_0_is_idle - check VCN block is idle
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block structure
+ *
+ * Check whether VCN block is idle
+ */
+static bool vcn_v5_0_0_is_idle(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, ret = 1;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		if (adev->vcn.harvest_config & (1 << i))
+			continue;
+
+		ret &= (RREG32_SOC15(VCN, i, regUVD_STATUS) == UVD_STATUS__IDLE);
+	}
+
+	return ret;
+}
+
+/**
+ * vcn_v5_0_0_wait_for_idle - wait for VCN block idle
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Wait for VCN block idle
+ */
+static int vcn_v5_0_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, ret = 0;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		if (adev->vcn.harvest_config & (1 << i))
+			continue;
+
+		ret = SOC15_WAIT_ON_RREG(VCN, i, regUVD_STATUS, UVD_STATUS__IDLE,
+			UVD_STATUS__IDLE);
+		if (ret)
+			return ret;
+	}
+
+	return ret;
+}
+
+/**
+ * vcn_v5_0_0_set_clockgating_state - set VCN block clockgating state
+ *
+ * @ip_block: amdgpu_ip_block pointer
+ * @state: clock gating state
+ *
+ * Set VCN block clockgating state
+ */
+static int vcn_v5_0_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
+					  enum amd_clockgating_state state)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	bool enable = state == AMD_CG_STATE_GATE;
+	int i;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		struct amdgpu_vcn_inst *vinst = &adev->vcn.inst[i];
+
+		if (adev->vcn.harvest_config & (1 << i))
+			continue;
+
+		if (enable) {
+			if (RREG32_SOC15(VCN, i, regUVD_STATUS) != UVD_STATUS__IDLE)
+				return -EBUSY;
+			vcn_v5_0_0_enable_clock_gating(vinst);
+		} else {
+			vcn_v5_0_0_disable_clock_gating(vinst);
+		}
+	}
+
+	return 0;
+}
+
+static int vcn_v5_0_0_set_pg_state(struct amdgpu_vcn_inst *vinst,
+				   enum amd_powergating_state state)
+{
+	int ret = 0;
+
+	if (state == vinst->cur_state)
+		return 0;
+
+	if (state == AMD_PG_STATE_GATE)
+		ret = vcn_v5_0_0_stop(vinst);
+	else
+		ret = vcn_v5_0_0_start(vinst);
+
+	if (!ret)
+		vinst->cur_state = state;
+
+	return ret;
+}
+
+/**
+ * vcn_v5_0_0_process_interrupt - process VCN block interrupt
+ *
+ * @adev: amdgpu_device pointer
+ * @source: interrupt sources
+ * @entry: interrupt entry from clients and sources
+ *
+ * Process VCN block interrupt
+ */
+static int vcn_v5_0_0_process_interrupt(struct amdgpu_device *adev, struct amdgpu_irq_src *source,
+	struct amdgpu_iv_entry *entry)
+{
+	uint32_t ip_instance;
+
+	switch (entry->client_id) {
+	case SOC15_IH_CLIENTID_VCN:
+		ip_instance = 0;
+		break;
+	case SOC15_IH_CLIENTID_VCN1:
+		ip_instance = 1;
+		break;
+	default:
+		DRM_ERROR("Unhandled client id: %d\n", entry->client_id);
+		return 0;
+	}
+
+	DRM_DEBUG("IH: VCN TRAP\n");
+
+	switch (entry->src_id) {
+	case VCN_5_0__SRCID__UVD_ENC_GENERAL_PURPOSE:
+		amdgpu_fence_process(&adev->vcn.inst[ip_instance].ring_enc[0]);
+		break;
+	case VCN_5_0__SRCID_UVD_POISON:
+		amdgpu_vcn_process_poison_irq(adev, source, entry);
+		break;
+	default:
+		DRM_ERROR("Unhandled interrupt: %d %d\n",
+			  entry->src_id, entry->src_data[0]);
+		break;
+	}
+
+	return 0;
+}
+
+static const struct amdgpu_irq_src_funcs vcn_v5_0_0_irq_funcs = {
+	.process = vcn_v5_0_0_process_interrupt,
+};
+
+/**
+ * vcn_v5_0_0_set_irq_funcs - set VCN block interrupt irq functions
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Set VCN block interrupt irq functions
+ */
+static void vcn_v5_0_0_set_irq_funcs(struct amdgpu_device *adev)
+{
+	int i;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		if (adev->vcn.harvest_config & (1 << i))
+			continue;
+
+		adev->vcn.inst[i].irq.num_types = adev->vcn.inst[i].num_enc_rings + 1;
+		adev->vcn.inst[i].irq.funcs = &vcn_v5_0_0_irq_funcs;
+	}
+}
+
+static const struct amd_ip_funcs vcn_v5_0_0_ip_funcs = {
+	.name = "vcn_v5_0_0",
+	.early_init = vcn_v5_0_0_early_init,
+	.sw_init = vcn_v5_0_0_sw_init,
+	.sw_fini = vcn_v5_0_0_sw_fini,
+	.hw_init = vcn_v5_0_0_hw_init,
+	.hw_fini = vcn_v5_0_0_hw_fini,
+	.suspend = vcn_v5_0_0_suspend,
+	.resume = vcn_v5_0_0_resume,
+	.is_idle = vcn_v5_0_0_is_idle,
+	.wait_for_idle = vcn_v5_0_0_wait_for_idle,
+	.set_clockgating_state = vcn_v5_0_0_set_clockgating_state,
+	.set_powergating_state = vcn_set_powergating_state,
+	.dump_ip_state = amdgpu_vcn_dump_ip_state,
+	.print_ip_state = amdgpu_vcn_print_ip_state,
+};
+
+const struct amdgpu_ip_block_version vcn_v5_0_0_ip_block = {
+	.type = AMD_IP_BLOCK_TYPE_VCN,
+	.major = 5,
+	.minor = 0,
+	.rev = 0,
+	.funcs = &vcn_v5_0_0_ip_funcs,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.h b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.h
new file mode 100644
index 000000000000..51bbccd4360f
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.h
@@ -0,0 +1,37 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __VCN_V5_0_0_H__
+#define __VCN_V5_0_0_H__
+
+#define VCN_VID_SOC_ADDRESS                        0x1FC00
+#define VCN_AON_SOC_ADDRESS                        0x1F800
+#define VCN1_VID_SOC_ADDRESS                       0x48300
+#define VCN1_AON_SOC_ADDRESS                       0x48000
+
+#define VCN_VID_IP_ADDRESS                         0x0
+#define VCN_AON_IP_ADDRESS                         0x30000
+
+extern const struct amdgpu_ip_block_version vcn_v5_0_0_ip_block;
+
+#endif /* __VCN_V5_0_0_H__ */
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c
new file mode 100644
index 000000000000..8bd457dea4cf
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c
@@ -0,0 +1,1729 @@
+/*
+ * Copyright 2024 Advanced Micro Devices, Inc. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/firmware.h>
+#include "amdgpu.h"
+#include "amdgpu_vcn.h"
+#include "amdgpu_pm.h"
+#include "soc15.h"
+#include "soc15d.h"
+#include "soc15_hw_ip.h"
+#include "vcn_v2_0.h"
+#include "vcn_v4_0_3.h"
+#include "mmsch_v5_0.h"
+
+#include "vcn/vcn_5_0_0_offset.h"
+#include "vcn/vcn_5_0_0_sh_mask.h"
+#include "ivsrcid/vcn/irqsrcs_vcn_5_0.h"
+#include "vcn_v5_0_0.h"
+#include "vcn_v5_0_1.h"
+
+#include <drm/drm_drv.h>
+
+static const struct amdgpu_hwip_reg_entry vcn_reg_list_5_0_1[] = {
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_POWER_STATUS),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_STATUS),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_CONTEXT_ID),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_CONTEXT_ID2),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_GPCOM_VCPU_DATA0),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_GPCOM_VCPU_DATA1),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_GPCOM_VCPU_CMD),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_BASE_HI),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_BASE_LO),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_BASE_HI2),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_BASE_LO2),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_BASE_HI3),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_BASE_LO3),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_BASE_HI4),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_BASE_LO4),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_RPTR),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_WPTR),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_RPTR2),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_WPTR2),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_RPTR3),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_WPTR3),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_RPTR4),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_WPTR4),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_SIZE),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_SIZE2),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_SIZE3),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_RB_SIZE4),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_DPG_LMA_CTL),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_DPG_LMA_DATA),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_DPG_LMA_MASK),
+	SOC15_REG_ENTRY_STR(VCN, 0, regUVD_DPG_PAUSE)
+};
+
+static int vcn_v5_0_1_start_sriov(struct amdgpu_device *adev);
+static void vcn_v5_0_1_set_unified_ring_funcs(struct amdgpu_device *adev);
+static void vcn_v5_0_1_set_irq_funcs(struct amdgpu_device *adev);
+static int vcn_v5_0_1_set_pg_state(struct amdgpu_vcn_inst *vinst,
+				   enum amd_powergating_state state);
+static void vcn_v5_0_1_unified_ring_set_wptr(struct amdgpu_ring *ring);
+static void vcn_v5_0_1_set_ras_funcs(struct amdgpu_device *adev);
+/**
+ * vcn_v5_0_1_early_init - set function pointers and load microcode
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Set ring and irq function pointers
+ * Load microcode from filesystem
+ */
+static int vcn_v5_0_1_early_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, r;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i)
+		/* re-use enc ring as unified ring */
+		adev->vcn.inst[i].num_enc_rings = 1;
+
+	vcn_v5_0_1_set_unified_ring_funcs(adev);
+	vcn_v5_0_1_set_irq_funcs(adev);
+	vcn_v5_0_1_set_ras_funcs(adev);
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		adev->vcn.inst[i].set_pg_state = vcn_v5_0_1_set_pg_state;
+
+		r = amdgpu_vcn_early_init(adev, i);
+		if (r)
+			return r;
+	}
+
+	return 0;
+}
+
+static int vcn_v5_0_1_late_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+
+	adev->vcn.supported_reset =
+		amdgpu_get_soft_full_reset_mask(&adev->vcn.inst[0].ring_enc[0]);
+
+	switch (amdgpu_ip_version(adev, MP0_HWIP, 0)) {
+	case IP_VERSION(13, 0, 12):
+		if ((adev->psp.sos.fw_version >= 0x00450025) &&
+			amdgpu_dpm_reset_vcn_is_supported(adev) &&
+			!amdgpu_sriov_vf(adev))
+			adev->vcn.supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE;
+		break;
+	default:
+		break;
+	}
+
+	return 0;
+}
+
+static void vcn_v5_0_1_fw_shared_init(struct amdgpu_device *adev, int inst_idx)
+{
+	struct amdgpu_vcn5_fw_shared *fw_shared;
+
+	fw_shared = adev->vcn.inst[inst_idx].fw_shared.cpu_addr;
+
+	if (fw_shared->sq.is_enabled)
+		return;
+	fw_shared->present_flag_0 =
+		cpu_to_le32(AMDGPU_FW_SHARED_FLAG_0_UNIFIED_QUEUE);
+	fw_shared->sq.is_enabled = 1;
+
+	if (amdgpu_vcnfw_log)
+		amdgpu_vcn_fwlog_init(&adev->vcn.inst[inst_idx]);
+}
+
+/**
+ * vcn_v5_0_1_sw_init - sw init for VCN block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Load firmware and sw initialization
+ */
+static int vcn_v5_0_1_sw_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	struct amdgpu_ring *ring;
+	int i, r, vcn_inst;
+
+	/* VCN UNIFIED TRAP */
+	r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_VCN,
+		VCN_5_0__SRCID__UVD_ENC_GENERAL_PURPOSE, &adev->vcn.inst->irq);
+	if (r)
+		return r;
+
+	/* VCN POISON TRAP */
+	r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_VCN,
+		VCN_5_0__SRCID_UVD_POISON, &adev->vcn.inst->ras_poison_irq);
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+		vcn_inst = GET_INST(VCN, i);
+
+		r = amdgpu_vcn_sw_init(adev, i);
+		if (r)
+			return r;
+
+		amdgpu_vcn_setup_ucode(adev, i);
+
+		r = amdgpu_vcn_resume(adev, i);
+		if (r)
+			return r;
+
+		ring = &adev->vcn.inst[i].ring_enc[0];
+		ring->use_doorbell = true;
+		if (!amdgpu_sriov_vf(adev))
+			ring->doorbell_index =
+				(adev->doorbell_index.vcn.vcn_ring0_1 << 1) +
+				11 * vcn_inst;
+		else
+			ring->doorbell_index =
+				(adev->doorbell_index.vcn.vcn_ring0_1 << 1) +
+				32 * vcn_inst;
+
+		ring->vm_hub = AMDGPU_MMHUB0(adev->vcn.inst[i].aid_id);
+		sprintf(ring->name, "vcn_unified_%d", adev->vcn.inst[i].aid_id);
+
+		r = amdgpu_ring_init(adev, ring, 512, &adev->vcn.inst[i].irq, 0,
+					AMDGPU_RING_PRIO_DEFAULT, &adev->vcn.inst[i].sched_score);
+		if (r)
+			return r;
+
+		vcn_v5_0_1_fw_shared_init(adev, i);
+	}
+
+	if (amdgpu_sriov_vf(adev)) {
+		r = amdgpu_virt_alloc_mm_table(adev);
+		if (r)
+			return r;
+	}
+
+	if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__VCN)) {
+		r = amdgpu_vcn_ras_sw_init(adev);
+		if (r) {
+			dev_err(adev->dev, "Failed to initialize vcn ras block!\n");
+			return r;
+		}
+	}
+
+	r = amdgpu_vcn_reg_dump_init(adev, vcn_reg_list_5_0_1, ARRAY_SIZE(vcn_reg_list_5_0_1));
+	if (r)
+		return r;
+
+	return amdgpu_vcn_sysfs_reset_mask_init(adev);
+}
+
+/**
+ * vcn_v5_0_1_sw_fini - sw fini for VCN block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * VCN suspend and free up sw allocation
+ */
+static int vcn_v5_0_1_sw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, r, idx;
+
+	if (drm_dev_enter(adev_to_drm(adev), &idx)) {
+		for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+			struct amdgpu_vcn5_fw_shared *fw_shared;
+
+			fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr;
+			fw_shared->present_flag_0 = 0;
+			fw_shared->sq.is_enabled = 0;
+		}
+
+		drm_dev_exit(idx);
+	}
+
+	if (amdgpu_sriov_vf(adev))
+		amdgpu_virt_free_mm_table(adev);
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+		r = amdgpu_vcn_suspend(adev, i);
+		if (r)
+			return r;
+	}
+
+	amdgpu_vcn_sysfs_reset_mask_fini(adev);
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++)
+		amdgpu_vcn_sw_fini(adev, i);
+
+	return 0;
+}
+
+static int vcn_v5_0_1_hw_init_inst(struct amdgpu_device *adev, int i)
+{
+	struct amdgpu_ring *ring;
+	int vcn_inst;
+
+	vcn_inst = GET_INST(VCN, i);
+	ring = &adev->vcn.inst[i].ring_enc[0];
+
+	if (ring->use_doorbell)
+		adev->nbio.funcs->vcn_doorbell_range(adev, ring->use_doorbell,
+			((adev->doorbell_index.vcn.vcn_ring0_1 << 1) +
+			11 * vcn_inst),
+			adev->vcn.inst[i].aid_id);
+
+	return 0;
+}
+
+/**
+ * vcn_v5_0_1_hw_init - start and test VCN block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Initialize the hardware, boot up the VCPU and do some testing
+ */
+static int vcn_v5_0_1_hw_init(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	struct amdgpu_ring *ring;
+	int i, r;
+
+	if (amdgpu_sriov_vf(adev)) {
+		r = vcn_v5_0_1_start_sriov(adev);
+		if (r)
+			return r;
+
+		for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+			ring = &adev->vcn.inst[i].ring_enc[0];
+			ring->wptr = 0;
+			ring->wptr_old = 0;
+			vcn_v5_0_1_unified_ring_set_wptr(ring);
+			ring->sched.ready = true;
+		}
+	} else {
+		if (RREG32_SOC15(VCN, GET_INST(VCN, 0), regVCN_RRMT_CNTL) & 0x100)
+			adev->vcn.caps |= AMDGPU_VCN_CAPS(RRMT_ENABLED);
+		for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+			ring = &adev->vcn.inst[i].ring_enc[0];
+			vcn_v5_0_1_hw_init_inst(adev, i);
+
+			/* Re-init fw_shared, if required */
+			vcn_v5_0_1_fw_shared_init(adev, i);
+
+			r = amdgpu_ring_test_helper(ring);
+			if (r)
+				return r;
+		}
+	}
+
+	return 0;
+}
+
+/**
+ * vcn_v5_0_1_hw_fini - stop the hardware block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Stop the VCN block, mark ring as not ready any more
+ */
+static int vcn_v5_0_1_hw_fini(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		struct amdgpu_vcn_inst *vinst = &adev->vcn.inst[i];
+
+		cancel_delayed_work_sync(&adev->vcn.inst[i].idle_work);
+		if (vinst->cur_state != AMD_PG_STATE_GATE)
+			vinst->set_pg_state(vinst, AMD_PG_STATE_GATE);
+	}
+
+	if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__VCN) && !amdgpu_sriov_vf(adev))
+		amdgpu_irq_put(adev, &adev->vcn.inst->ras_poison_irq, 0);
+
+	return 0;
+}
+
+/**
+ * vcn_v5_0_1_suspend - suspend VCN block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * HW fini and suspend VCN block
+ */
+static int vcn_v5_0_1_suspend(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int r, i;
+
+	r = vcn_v5_0_1_hw_fini(ip_block);
+	if (r)
+		return r;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+		r = amdgpu_vcn_suspend(ip_block->adev, i);
+		if (r)
+			return r;
+	}
+
+	return r;
+}
+
+/**
+ * vcn_v5_0_1_resume - resume VCN block
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Resume firmware and hw init VCN block
+ */
+static int vcn_v5_0_1_resume(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int r, i;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+		struct amdgpu_vcn_inst *vinst = &adev->vcn.inst[i];
+
+		if (amdgpu_in_reset(adev))
+			vinst->cur_state = AMD_PG_STATE_GATE;
+
+		r = amdgpu_vcn_resume(ip_block->adev, i);
+		if (r)
+			return r;
+	}
+
+	r = vcn_v5_0_1_hw_init(ip_block);
+
+	return r;
+}
+
+/**
+ * vcn_v5_0_1_mc_resume - memory controller programming
+ *
+ * @vinst: VCN instance
+ *
+ * Let the VCN memory controller know it's offsets
+ */
+static void vcn_v5_0_1_mc_resume(struct amdgpu_vcn_inst *vinst)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int inst = vinst->inst;
+	uint32_t offset, size, vcn_inst;
+	const struct common_firmware_header *hdr;
+
+	hdr = (const struct common_firmware_header *)adev->vcn.inst[inst].fw->data;
+	size = AMDGPU_GPU_PAGE_ALIGN(le32_to_cpu(hdr->ucode_size_bytes) + 8);
+
+	vcn_inst = GET_INST(VCN, inst);
+	/* cache window 0: fw */
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
+		WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW,
+			(adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + inst].tmr_mc_addr_lo));
+		WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH,
+			(adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + inst].tmr_mc_addr_hi));
+		WREG32_SOC15(VCN, vcn_inst, regUVD_VCPU_CACHE_OFFSET0, 0);
+		offset = 0;
+	} else {
+		WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW,
+			lower_32_bits(adev->vcn.inst[inst].gpu_addr));
+		WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH,
+			upper_32_bits(adev->vcn.inst[inst].gpu_addr));
+		offset = size;
+		WREG32_SOC15(VCN, vcn_inst, regUVD_VCPU_CACHE_OFFSET0,
+				AMDGPU_UVD_FIRMWARE_OFFSET >> 3);
+	}
+	WREG32_SOC15(VCN, vcn_inst, regUVD_VCPU_CACHE_SIZE0, size);
+
+	/* cache window 1: stack */
+	WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW,
+		lower_32_bits(adev->vcn.inst[inst].gpu_addr + offset));
+	WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH,
+		upper_32_bits(adev->vcn.inst[inst].gpu_addr + offset));
+	WREG32_SOC15(VCN, vcn_inst, regUVD_VCPU_CACHE_OFFSET1, 0);
+	WREG32_SOC15(VCN, vcn_inst, regUVD_VCPU_CACHE_SIZE1, AMDGPU_VCN_STACK_SIZE);
+
+	/* cache window 2: context */
+	WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_VCPU_CACHE2_64BIT_BAR_LOW,
+		lower_32_bits(adev->vcn.inst[inst].gpu_addr + offset + AMDGPU_VCN_STACK_SIZE));
+	WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_VCPU_CACHE2_64BIT_BAR_HIGH,
+		upper_32_bits(adev->vcn.inst[inst].gpu_addr + offset + AMDGPU_VCN_STACK_SIZE));
+	WREG32_SOC15(VCN, vcn_inst, regUVD_VCPU_CACHE_OFFSET2, 0);
+	WREG32_SOC15(VCN, vcn_inst, regUVD_VCPU_CACHE_SIZE2, AMDGPU_VCN_CONTEXT_SIZE);
+
+	/* non-cache window */
+	WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_VCPU_NC0_64BIT_BAR_LOW,
+		lower_32_bits(adev->vcn.inst[inst].fw_shared.gpu_addr));
+	WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_VCPU_NC0_64BIT_BAR_HIGH,
+		upper_32_bits(adev->vcn.inst[inst].fw_shared.gpu_addr));
+	WREG32_SOC15(VCN, vcn_inst, regUVD_VCPU_NONCACHE_OFFSET0, 0);
+	WREG32_SOC15(VCN, vcn_inst, regUVD_VCPU_NONCACHE_SIZE0,
+		AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_vcn5_fw_shared)));
+}
+
+/**
+ * vcn_v5_0_1_mc_resume_dpg_mode - memory controller programming for dpg mode
+ *
+ * @vinst: VCN instance
+ * @indirect: indirectly write sram
+ *
+ * Let the VCN memory controller know it's offsets with dpg mode
+ */
+static void vcn_v5_0_1_mc_resume_dpg_mode(struct amdgpu_vcn_inst *vinst,
+					  bool indirect)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int inst_idx = vinst->inst;
+	uint32_t offset, size;
+	const struct common_firmware_header *hdr;
+
+	hdr = (const struct common_firmware_header *)adev->vcn.inst[inst_idx].fw->data;
+	size = AMDGPU_GPU_PAGE_ALIGN(le32_to_cpu(hdr->ucode_size_bytes) + 8);
+
+	/* cache window 0: fw */
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
+		if (!indirect) {
+			WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+				VCN, 0, regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW),
+				(adev->firmware.ucode[AMDGPU_UCODE_ID_VCN +
+				 inst_idx].tmr_mc_addr_lo), 0, indirect);
+			WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+				VCN, 0, regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH),
+				(adev->firmware.ucode[AMDGPU_UCODE_ID_VCN +
+				 inst_idx].tmr_mc_addr_hi), 0, indirect);
+			WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+				VCN, 0, regUVD_VCPU_CACHE_OFFSET0), 0, 0, indirect);
+		} else {
+			WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+				VCN, 0, regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW), 0, 0, indirect);
+			WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+				VCN, 0, regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH), 0, 0, indirect);
+			WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+				VCN, 0, regUVD_VCPU_CACHE_OFFSET0), 0, 0, indirect);
+		}
+		offset = 0;
+	} else {
+		WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+			VCN, 0, regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW),
+			lower_32_bits(adev->vcn.inst[inst_idx].gpu_addr), 0, indirect);
+		WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+			VCN, 0, regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH),
+			upper_32_bits(adev->vcn.inst[inst_idx].gpu_addr), 0, indirect);
+		offset = size;
+		WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+			VCN, 0, regUVD_VCPU_CACHE_OFFSET0),
+			AMDGPU_UVD_FIRMWARE_OFFSET >> 3, 0, indirect);
+	}
+
+	if (!indirect)
+		WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+			VCN, 0, regUVD_VCPU_CACHE_SIZE0), size, 0, indirect);
+	else
+		WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+			VCN, 0, regUVD_VCPU_CACHE_SIZE0), 0, 0, indirect);
+
+	/* cache window 1: stack */
+	if (!indirect) {
+		WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+			VCN, 0, regUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW),
+			lower_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset), 0, indirect);
+		WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+			VCN, 0, regUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH),
+			upper_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset), 0, indirect);
+		WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+			VCN, 0, regUVD_VCPU_CACHE_OFFSET1), 0, 0, indirect);
+	} else {
+		WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+			VCN, 0, regUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW), 0, 0, indirect);
+		WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+			VCN, 0, regUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH), 0, 0, indirect);
+		WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+			VCN, 0, regUVD_VCPU_CACHE_OFFSET1), 0, 0, indirect);
+	}
+	WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+			VCN, 0, regUVD_VCPU_CACHE_SIZE1), AMDGPU_VCN_STACK_SIZE, 0, indirect);
+
+	/* cache window 2: context */
+	WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+		VCN, 0, regUVD_LMI_VCPU_CACHE2_64BIT_BAR_LOW),
+		lower_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset +
+			AMDGPU_VCN_STACK_SIZE), 0, indirect);
+	WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+		VCN, 0, regUVD_LMI_VCPU_CACHE2_64BIT_BAR_HIGH),
+		upper_32_bits(adev->vcn.inst[inst_idx].gpu_addr + offset +
+			AMDGPU_VCN_STACK_SIZE), 0, indirect);
+	WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+		VCN, 0, regUVD_VCPU_CACHE_OFFSET2), 0, 0, indirect);
+	WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+		VCN, 0, regUVD_VCPU_CACHE_SIZE2), AMDGPU_VCN_CONTEXT_SIZE, 0, indirect);
+
+	/* non-cache window */
+	WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+		VCN, 0, regUVD_LMI_VCPU_NC0_64BIT_BAR_LOW),
+		lower_32_bits(adev->vcn.inst[inst_idx].fw_shared.gpu_addr), 0, indirect);
+	WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+		VCN, 0, regUVD_LMI_VCPU_NC0_64BIT_BAR_HIGH),
+		upper_32_bits(adev->vcn.inst[inst_idx].fw_shared.gpu_addr), 0, indirect);
+	WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+		VCN, 0, regUVD_VCPU_NONCACHE_OFFSET0), 0, 0, indirect);
+	WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+		VCN, 0, regUVD_VCPU_NONCACHE_SIZE0),
+		AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_vcn5_fw_shared)), 0, indirect);
+
+	/* VCN global tiling registers */
+	WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+		VCN, 0, regUVD_GFX10_ADDR_CONFIG), adev->gfx.config.gb_addr_config, 0, indirect);
+}
+
+/**
+ * vcn_v5_0_1_disable_clock_gating - disable VCN clock gating
+ *
+ * @vinst: VCN instance
+ *
+ * Disable clock gating for VCN block
+ */
+static void vcn_v5_0_1_disable_clock_gating(struct amdgpu_vcn_inst *vinst)
+{
+}
+
+/**
+ * vcn_v5_0_1_enable_clock_gating - enable VCN clock gating
+ *
+ * @vinst: VCN instance
+ *
+ * Enable clock gating for VCN block
+ */
+static void vcn_v5_0_1_enable_clock_gating(struct amdgpu_vcn_inst *vinst)
+{
+}
+
+/**
+ * vcn_v5_0_1_pause_dpg_mode - VCN pause with dpg mode
+ *
+ * @vinst: VCN instance
+ * @new_state: pause state
+ *
+ * Pause dpg mode for VCN block
+ */
+static int vcn_v5_0_1_pause_dpg_mode(struct amdgpu_vcn_inst *vinst,
+				     struct dpg_pause_state *new_state)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	uint32_t reg_data = 0;
+	int vcn_inst;
+
+	vcn_inst = GET_INST(VCN, vinst->inst);
+
+	/* pause/unpause if state is changed */
+	if (vinst->pause_state.fw_based != new_state->fw_based) {
+		DRM_DEV_DEBUG(adev->dev, "dpg pause state changed %d -> %d %s\n",
+			vinst->pause_state.fw_based, new_state->fw_based,
+			new_state->fw_based ? "VCN_DPG_STATE__PAUSE" : "VCN_DPG_STATE__UNPAUSE");
+		reg_data = RREG32_SOC15(VCN, vcn_inst, regUVD_DPG_PAUSE) &
+			(~UVD_DPG_PAUSE__NJ_PAUSE_DPG_ACK_MASK);
+
+		if (new_state->fw_based == VCN_DPG_STATE__PAUSE) {
+			/* pause DPG */
+			reg_data |= UVD_DPG_PAUSE__NJ_PAUSE_DPG_REQ_MASK;
+			WREG32_SOC15(VCN, vcn_inst, regUVD_DPG_PAUSE, reg_data);
+
+			/* wait for ACK */
+			SOC15_WAIT_ON_RREG(VCN, vcn_inst, regUVD_DPG_PAUSE,
+					UVD_DPG_PAUSE__NJ_PAUSE_DPG_ACK_MASK,
+					UVD_DPG_PAUSE__NJ_PAUSE_DPG_ACK_MASK);
+		} else {
+			/* unpause DPG, no need to wait */
+			reg_data &= ~UVD_DPG_PAUSE__NJ_PAUSE_DPG_REQ_MASK;
+			WREG32_SOC15(VCN, vcn_inst, regUVD_DPG_PAUSE, reg_data);
+		}
+		vinst->pause_state.fw_based = new_state->fw_based;
+	}
+
+	return 0;
+}
+
+
+/**
+ * vcn_v5_0_1_start_dpg_mode - VCN start with dpg mode
+ *
+ * @vinst: VCN instance
+ * @indirect: indirectly write sram
+ *
+ * Start VCN block with dpg mode
+ */
+static int vcn_v5_0_1_start_dpg_mode(struct amdgpu_vcn_inst *vinst,
+				     bool indirect)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int inst_idx = vinst->inst;
+	struct amdgpu_vcn5_fw_shared *fw_shared =
+		adev->vcn.inst[inst_idx].fw_shared.cpu_addr;
+	struct amdgpu_ring *ring;
+	struct dpg_pause_state state = {.fw_based = VCN_DPG_STATE__PAUSE};
+	int vcn_inst, ret;
+	uint32_t tmp;
+
+	vcn_inst = GET_INST(VCN, inst_idx);
+
+	/* disable register anti-hang mechanism */
+	WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_POWER_STATUS), 1,
+		~UVD_POWER_STATUS__UVD_POWER_STATUS_MASK);
+
+	/* enable dynamic power gating mode */
+	tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_POWER_STATUS);
+	tmp |= UVD_POWER_STATUS__UVD_PG_MODE_MASK;
+	WREG32_SOC15(VCN, vcn_inst, regUVD_POWER_STATUS, tmp);
+
+	if (indirect) {
+		adev->vcn.inst[inst_idx].dpg_sram_curr_addr =
+			(uint32_t *)adev->vcn.inst[inst_idx].dpg_sram_cpu_addr;
+		/* Use dummy register 0xDEADBEEF passing AID selection to PSP FW */
+		WREG32_SOC24_DPG_MODE(inst_idx, 0xDEADBEEF,
+				adev->vcn.inst[inst_idx].aid_id, 0, true);
+	}
+
+	/* enable VCPU clock */
+	tmp = (0xFF << UVD_VCPU_CNTL__PRB_TIMEOUT_VAL__SHIFT);
+	tmp |= UVD_VCPU_CNTL__CLK_EN_MASK | UVD_VCPU_CNTL__BLK_RST_MASK;
+	WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+		VCN, 0, regUVD_VCPU_CNTL), tmp, 0, indirect);
+
+	/* disable master interrupt */
+	WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+		VCN, 0, regUVD_MASTINT_EN), 0, 0, indirect);
+
+	/* setup regUVD_LMI_CTRL */
+	tmp = (UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK |
+		UVD_LMI_CTRL__REQ_MODE_MASK |
+		UVD_LMI_CTRL__CRC_RESET_MASK |
+		UVD_LMI_CTRL__MASK_MC_URGENT_MASK |
+		UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK |
+		UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK |
+		(8 << UVD_LMI_CTRL__WRITE_CLEAN_TIMER__SHIFT) |
+		0x00100000L);
+	WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+		VCN, 0, regUVD_LMI_CTRL), tmp, 0, indirect);
+
+	vcn_v5_0_1_mc_resume_dpg_mode(vinst, indirect);
+
+	tmp = (0xFF << UVD_VCPU_CNTL__PRB_TIMEOUT_VAL__SHIFT);
+	tmp |= UVD_VCPU_CNTL__CLK_EN_MASK;
+	WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+		VCN, 0, regUVD_VCPU_CNTL), tmp, 0, indirect);
+
+	/* enable LMI MC and UMC channels */
+	tmp = 0x1f << UVD_LMI_CTRL2__RE_OFLD_MIF_WR_REQ_NUM__SHIFT;
+	WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+		VCN, 0, regUVD_LMI_CTRL2), tmp, 0, indirect);
+
+	/* enable master interrupt */
+	WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
+		VCN, 0, regUVD_MASTINT_EN),
+		UVD_MASTINT_EN__VCPU_EN_MASK, 0, indirect);
+
+	if (indirect) {
+		ret = amdgpu_vcn_psp_update_sram(adev, inst_idx, AMDGPU_UCODE_ID_VCN0_RAM);
+		if (ret) {
+			dev_err(adev->dev, "vcn sram load failed %d\n", ret);
+			return ret;
+		}
+	}
+
+	/* resetting ring, fw should not check RB ring */
+	fw_shared->sq.queue_mode |= FW_QUEUE_RING_RESET;
+
+	/* Pause dpg */
+	vcn_v5_0_1_pause_dpg_mode(vinst, &state);
+
+	ring = &adev->vcn.inst[inst_idx].ring_enc[0];
+
+	WREG32_SOC15(VCN, vcn_inst, regUVD_RB_BASE_LO, lower_32_bits(ring->gpu_addr));
+	WREG32_SOC15(VCN, vcn_inst, regUVD_RB_BASE_HI, upper_32_bits(ring->gpu_addr));
+	WREG32_SOC15(VCN, vcn_inst, regUVD_RB_SIZE, ring->ring_size / sizeof(uint32_t));
+
+	tmp = RREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE);
+	tmp &= ~(VCN_RB_ENABLE__RB1_EN_MASK);
+	WREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE, tmp);
+
+	WREG32_SOC15(VCN, vcn_inst, regUVD_RB_RPTR, 0);
+	WREG32_SOC15(VCN, vcn_inst, regUVD_RB_WPTR, 0);
+
+	tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_RB_RPTR);
+	WREG32_SOC15(VCN, vcn_inst, regUVD_RB_WPTR, tmp);
+	ring->wptr = RREG32_SOC15(VCN, vcn_inst, regUVD_RB_WPTR);
+
+	tmp = RREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE);
+	tmp |= VCN_RB_ENABLE__RB1_EN_MASK;
+	WREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE, tmp);
+	/* resetting done, fw can check RB ring */
+	fw_shared->sq.queue_mode &= ~(FW_QUEUE_RING_RESET | FW_QUEUE_DPG_HOLD_OFF);
+
+	WREG32_SOC15(VCN, vcn_inst, regVCN_RB1_DB_CTRL,
+		ring->doorbell_index << VCN_RB1_DB_CTRL__OFFSET__SHIFT |
+		VCN_RB1_DB_CTRL__EN_MASK);
+	/* Read DB_CTRL to flush the write DB_CTRL command. */
+	RREG32_SOC15(VCN, vcn_inst, regVCN_RB1_DB_CTRL);
+
+	return 0;
+}
+
+static int vcn_v5_0_1_start_sriov(struct amdgpu_device *adev)
+{
+	int i, vcn_inst;
+	struct amdgpu_ring *ring_enc;
+	uint64_t cache_addr;
+	uint64_t rb_enc_addr;
+	uint64_t ctx_addr;
+	uint32_t param, resp, expected;
+	uint32_t offset, cache_size;
+	uint32_t tmp, timeout;
+
+	struct amdgpu_mm_table *table = &adev->virt.mm_table;
+	uint32_t *table_loc;
+	uint32_t table_size;
+	uint32_t size, size_dw;
+	uint32_t init_status;
+	uint32_t enabled_vcn;
+
+	struct mmsch_v5_0_cmd_direct_write
+		direct_wt = { {0} };
+	struct mmsch_v5_0_cmd_direct_read_modify_write
+		direct_rd_mod_wt = { {0} };
+	struct mmsch_v5_0_cmd_end end = { {0} };
+	struct mmsch_v5_0_init_header header;
+
+	struct amdgpu_vcn5_fw_shared *fw_shared;
+	struct amdgpu_fw_shared_rb_setup *rb_setup;
+
+	direct_wt.cmd_header.command_type =
+		MMSCH_COMMAND__DIRECT_REG_WRITE;
+	direct_rd_mod_wt.cmd_header.command_type =
+		MMSCH_COMMAND__DIRECT_REG_READ_MODIFY_WRITE;
+	end.cmd_header.command_type = MMSCH_COMMAND__END;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+		vcn_inst = GET_INST(VCN, i);
+
+		vcn_v5_0_1_fw_shared_init(adev, vcn_inst);
+
+		memset(&header, 0, sizeof(struct mmsch_v5_0_init_header));
+		header.version = MMSCH_VERSION;
+		header.total_size = sizeof(struct mmsch_v5_0_init_header) >> 2;
+
+		table_loc = (uint32_t *)table->cpu_addr;
+		table_loc += header.total_size;
+
+		table_size = 0;
+
+		MMSCH_V5_0_INSERT_DIRECT_RD_MOD_WT(SOC15_REG_OFFSET(VCN, 0, regUVD_STATUS),
+			~UVD_STATUS__UVD_BUSY, UVD_STATUS__UVD_BUSY);
+
+		cache_size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.inst[i].fw->size + 4);
+
+		if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
+			MMSCH_V5_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0,
+				regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW),
+				adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + i].tmr_mc_addr_lo);
+
+			MMSCH_V5_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0,
+				regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH),
+				adev->firmware.ucode[AMDGPU_UCODE_ID_VCN + i].tmr_mc_addr_hi);
+
+			offset = 0;
+			MMSCH_V5_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0,
+				regUVD_VCPU_CACHE_OFFSET0), 0);
+		} else {
+			MMSCH_V5_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0,
+				regUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW),
+				lower_32_bits(adev->vcn.inst[i].gpu_addr));
+			MMSCH_V5_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0,
+				regUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH),
+				upper_32_bits(adev->vcn.inst[i].gpu_addr));
+			offset = cache_size;
+			MMSCH_V5_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0,
+				regUVD_VCPU_CACHE_OFFSET0),
+				AMDGPU_UVD_FIRMWARE_OFFSET >> 3);
+		}
+
+		MMSCH_V5_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0,
+			regUVD_VCPU_CACHE_SIZE0),
+			cache_size);
+
+		cache_addr = adev->vcn.inst[vcn_inst].gpu_addr + offset;
+		MMSCH_V5_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0,
+			regUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW), lower_32_bits(cache_addr));
+		MMSCH_V5_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0,
+			regUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH), upper_32_bits(cache_addr));
+		MMSCH_V5_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0,
+			regUVD_VCPU_CACHE_OFFSET1), 0);
+		MMSCH_V5_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0,
+			regUVD_VCPU_CACHE_SIZE1), AMDGPU_VCN_STACK_SIZE);
+
+		cache_addr = adev->vcn.inst[vcn_inst].gpu_addr + offset +
+			AMDGPU_VCN_STACK_SIZE;
+
+		MMSCH_V5_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0,
+			regUVD_LMI_VCPU_CACHE2_64BIT_BAR_LOW), lower_32_bits(cache_addr));
+
+		MMSCH_V5_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0,
+			regUVD_LMI_VCPU_CACHE2_64BIT_BAR_HIGH), upper_32_bits(cache_addr));
+
+		MMSCH_V5_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0,
+			regUVD_VCPU_CACHE_OFFSET2), 0);
+
+		MMSCH_V5_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0,
+			regUVD_VCPU_CACHE_SIZE2), AMDGPU_VCN_CONTEXT_SIZE);
+
+		fw_shared = adev->vcn.inst[vcn_inst].fw_shared.cpu_addr;
+		rb_setup = &fw_shared->rb_setup;
+
+		ring_enc = &adev->vcn.inst[vcn_inst].ring_enc[0];
+		ring_enc->wptr = 0;
+		rb_enc_addr = ring_enc->gpu_addr;
+
+		rb_setup->is_rb_enabled_flags |= RB_ENABLED;
+		rb_setup->rb_addr_lo = lower_32_bits(rb_enc_addr);
+		rb_setup->rb_addr_hi = upper_32_bits(rb_enc_addr);
+		rb_setup->rb_size = ring_enc->ring_size / 4;
+		fw_shared->present_flag_0 |= cpu_to_le32(AMDGPU_VCN_VF_RB_SETUP_FLAG);
+
+		MMSCH_V5_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0,
+			regUVD_LMI_VCPU_NC0_64BIT_BAR_LOW),
+			lower_32_bits(adev->vcn.inst[vcn_inst].fw_shared.gpu_addr));
+		MMSCH_V5_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0,
+			regUVD_LMI_VCPU_NC0_64BIT_BAR_HIGH),
+			upper_32_bits(adev->vcn.inst[vcn_inst].fw_shared.gpu_addr));
+		MMSCH_V5_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCN, 0,
+			regUVD_VCPU_NONCACHE_SIZE0),
+			AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_vcn4_fw_shared)));
+		MMSCH_V5_0_INSERT_END();
+
+		header.vcn0.init_status = 0;
+		header.vcn0.table_offset = header.total_size;
+		header.vcn0.table_size = table_size;
+		header.total_size += table_size;
+
+		/* Send init table to mmsch */
+		size = sizeof(struct mmsch_v5_0_init_header);
+		table_loc = (uint32_t *)table->cpu_addr;
+		memcpy((void *)table_loc, &header, size);
+
+		ctx_addr = table->gpu_addr;
+		WREG32_SOC15(VCN, vcn_inst, regMMSCH_VF_CTX_ADDR_LO, lower_32_bits(ctx_addr));
+		WREG32_SOC15(VCN, vcn_inst, regMMSCH_VF_CTX_ADDR_HI, upper_32_bits(ctx_addr));
+
+		tmp = RREG32_SOC15(VCN, vcn_inst, regMMSCH_VF_VMID);
+		tmp &= ~MMSCH_VF_VMID__VF_CTX_VMID_MASK;
+		tmp |= (0 << MMSCH_VF_VMID__VF_CTX_VMID__SHIFT);
+		WREG32_SOC15(VCN, vcn_inst, regMMSCH_VF_VMID, tmp);
+
+		size = header.total_size;
+		WREG32_SOC15(VCN, vcn_inst, regMMSCH_VF_CTX_SIZE, size);
+
+		WREG32_SOC15(VCN, vcn_inst, regMMSCH_VF_MAILBOX_RESP, 0);
+
+		param = 0x00000001;
+		WREG32_SOC15(VCN, vcn_inst, regMMSCH_VF_MAILBOX_HOST, param);
+		tmp = 0;
+		timeout = 1000;
+		resp = 0;
+		expected = MMSCH_VF_MAILBOX_RESP__OK;
+		while (resp != expected) {
+			resp = RREG32_SOC15(VCN, vcn_inst, regMMSCH_VF_MAILBOX_RESP);
+			if (resp != 0)
+				break;
+
+			udelay(10);
+			tmp = tmp + 10;
+			if (tmp >= timeout) {
+				DRM_ERROR("failed to init MMSCH. TIME-OUT after %d usec"\
+					" waiting for regMMSCH_VF_MAILBOX_RESP "\
+					"(expected=0x%08x, readback=0x%08x)\n",
+					tmp, expected, resp);
+				return -EBUSY;
+			}
+		}
+
+		enabled_vcn = amdgpu_vcn_is_disabled_vcn(adev, VCN_DECODE_RING, 0) ? 1 : 0;
+		init_status = ((struct mmsch_v5_0_init_header *)(table_loc))->vcn0.init_status;
+		if (resp != expected && resp != MMSCH_VF_MAILBOX_RESP__INCOMPLETE
+					&& init_status != MMSCH_VF_ENGINE_STATUS__PASS) {
+			DRM_ERROR("MMSCH init status is incorrect! readback=0x%08x, header init "\
+				"status for VCN%x: 0x%x\n", resp, enabled_vcn, init_status);
+		}
+	}
+
+	return 0;
+}
+
+/**
+ * vcn_v5_0_1_start - VCN start
+ *
+ * @vinst: VCN instance
+ *
+ * Start VCN block
+ */
+static int vcn_v5_0_1_start(struct amdgpu_vcn_inst *vinst)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int i = vinst->inst;
+	struct amdgpu_vcn5_fw_shared *fw_shared;
+	struct amdgpu_ring *ring;
+	uint32_t tmp;
+	int j, k, r, vcn_inst;
+
+	fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr;
+
+	if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)
+		return vcn_v5_0_1_start_dpg_mode(vinst, adev->vcn.inst[i].indirect_sram);
+
+	vcn_inst = GET_INST(VCN, i);
+
+	/* set VCN status busy */
+	tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_STATUS) | UVD_STATUS__UVD_BUSY;
+	WREG32_SOC15(VCN, vcn_inst, regUVD_STATUS, tmp);
+
+	/* enable VCPU clock */
+	WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_VCPU_CNTL),
+		 UVD_VCPU_CNTL__CLK_EN_MASK, ~UVD_VCPU_CNTL__CLK_EN_MASK);
+
+	/* disable master interrupt */
+	WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_MASTINT_EN), 0,
+		 ~UVD_MASTINT_EN__VCPU_EN_MASK);
+
+	/* enable LMI MC and UMC channels */
+	WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_LMI_CTRL2), 0,
+		 ~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK);
+
+	tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET);
+	tmp &= ~UVD_SOFT_RESET__LMI_SOFT_RESET_MASK;
+	tmp &= ~UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK;
+	WREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET, tmp);
+
+	/* setup regUVD_LMI_CTRL */
+	tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_LMI_CTRL);
+	WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_CTRL, tmp |
+		     UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK |
+		     UVD_LMI_CTRL__MASK_MC_URGENT_MASK |
+		     UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK |
+		     UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK);
+
+	vcn_v5_0_1_mc_resume(vinst);
+
+	/* VCN global tiling registers */
+	WREG32_SOC15(VCN, vcn_inst, regUVD_GFX10_ADDR_CONFIG,
+		     adev->gfx.config.gb_addr_config);
+
+	/* unblock VCPU register access */
+	WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_RB_ARB_CTRL), 0,
+		 ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK);
+
+	/* release VCPU reset to boot */
+	WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_VCPU_CNTL), 0,
+		 ~UVD_VCPU_CNTL__BLK_RST_MASK);
+
+	for (j = 0; j < 10; ++j) {
+		uint32_t status;
+
+		for (k = 0; k < 100; ++k) {
+			status = RREG32_SOC15(VCN, vcn_inst, regUVD_STATUS);
+			if (status & 2)
+				break;
+			mdelay(100);
+			if (amdgpu_emu_mode == 1)
+				msleep(20);
+		}
+
+		if (amdgpu_emu_mode == 1) {
+			r = -1;
+			if (status & 2) {
+				r = 0;
+				break;
+			}
+		} else {
+			r = 0;
+			if (status & 2)
+				break;
+
+			dev_err(adev->dev,
+				"VCN[%d] is not responding, trying to reset the VCPU!!!\n", i);
+			WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_VCPU_CNTL),
+				 UVD_VCPU_CNTL__BLK_RST_MASK,
+				 ~UVD_VCPU_CNTL__BLK_RST_MASK);
+			mdelay(10);
+			WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_VCPU_CNTL), 0,
+				 ~UVD_VCPU_CNTL__BLK_RST_MASK);
+
+			mdelay(10);
+			r = -1;
+		}
+	}
+
+	if (r) {
+		dev_err(adev->dev, "VCN[%d] is not responding, giving up!!!\n", i);
+		return r;
+	}
+
+	/* enable master interrupt */
+	WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_MASTINT_EN),
+		 UVD_MASTINT_EN__VCPU_EN_MASK,
+		 ~UVD_MASTINT_EN__VCPU_EN_MASK);
+
+	/* clear the busy bit of VCN_STATUS */
+	WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_STATUS), 0,
+		 ~(2 << UVD_STATUS__VCPU_REPORT__SHIFT));
+
+	ring = &adev->vcn.inst[i].ring_enc[0];
+
+	WREG32_SOC15(VCN, vcn_inst, regVCN_RB1_DB_CTRL,
+		     ring->doorbell_index << VCN_RB1_DB_CTRL__OFFSET__SHIFT |
+		     VCN_RB1_DB_CTRL__EN_MASK);
+
+	/* Read DB_CTRL to flush the write DB_CTRL command. */
+	RREG32_SOC15(VCN, vcn_inst, regVCN_RB1_DB_CTRL);
+
+	WREG32_SOC15(VCN, vcn_inst, regUVD_RB_BASE_LO, ring->gpu_addr);
+	WREG32_SOC15(VCN, vcn_inst, regUVD_RB_BASE_HI, upper_32_bits(ring->gpu_addr));
+	WREG32_SOC15(VCN, vcn_inst, regUVD_RB_SIZE, ring->ring_size / 4);
+
+	tmp = RREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE);
+	tmp &= ~(VCN_RB_ENABLE__RB1_EN_MASK);
+	WREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE, tmp);
+	fw_shared->sq.queue_mode |= FW_QUEUE_RING_RESET;
+	WREG32_SOC15(VCN, vcn_inst, regUVD_RB_RPTR, 0);
+	WREG32_SOC15(VCN, vcn_inst, regUVD_RB_WPTR, 0);
+
+	tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_RB_RPTR);
+	WREG32_SOC15(VCN, vcn_inst, regUVD_RB_WPTR, tmp);
+	ring->wptr = RREG32_SOC15(VCN, vcn_inst, regUVD_RB_WPTR);
+
+	tmp = RREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE);
+	tmp |= VCN_RB_ENABLE__RB1_EN_MASK;
+	WREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE, tmp);
+	fw_shared->sq.queue_mode &= ~(FW_QUEUE_RING_RESET | FW_QUEUE_DPG_HOLD_OFF);
+
+	/* Keeping one read-back to ensure all register writes are done,
+	 * otherwise it may introduce race conditions.
+	 */
+	RREG32_SOC15(VCN, vcn_inst, regUVD_STATUS);
+
+	return 0;
+}
+
+/**
+ * vcn_v5_0_1_stop_dpg_mode - VCN stop with dpg mode
+ *
+ * @vinst: VCN instance
+ *
+ * Stop VCN block with dpg mode
+ */
+static void vcn_v5_0_1_stop_dpg_mode(struct amdgpu_vcn_inst *vinst)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int inst_idx = vinst->inst;
+	uint32_t tmp;
+	int vcn_inst;
+	struct dpg_pause_state state = {.fw_based = VCN_DPG_STATE__UNPAUSE};
+
+	vcn_inst = GET_INST(VCN, inst_idx);
+
+	/* Unpause dpg */
+	vcn_v5_0_1_pause_dpg_mode(vinst, &state);
+
+	/* Wait for power status to be 1 */
+	SOC15_WAIT_ON_RREG(VCN, vcn_inst, regUVD_POWER_STATUS, 1,
+		UVD_POWER_STATUS__UVD_POWER_STATUS_MASK);
+
+	/* wait for read ptr to be equal to write ptr */
+	tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_RB_WPTR);
+	SOC15_WAIT_ON_RREG(VCN, vcn_inst, regUVD_RB_RPTR, tmp, 0xFFFFFFFF);
+
+	/* disable dynamic power gating mode */
+	WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_POWER_STATUS), 0,
+		~UVD_POWER_STATUS__UVD_PG_MODE_MASK);
+
+	/* Keeping one read-back to ensure all register writes are done,
+	 * otherwise it may introduce race conditions.
+	 */
+	RREG32_SOC15(VCN, vcn_inst, regUVD_STATUS);
+}
+
+/**
+ * vcn_v5_0_1_stop - VCN stop
+ *
+ * @vinst: VCN instance
+ *
+ * Stop VCN block
+ */
+static int vcn_v5_0_1_stop(struct amdgpu_vcn_inst *vinst)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int i = vinst->inst;
+	struct amdgpu_vcn5_fw_shared *fw_shared;
+	uint32_t tmp;
+	int r = 0, vcn_inst;
+
+	vcn_inst = GET_INST(VCN, i);
+
+	fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr;
+	fw_shared->sq.queue_mode |= FW_QUEUE_DPG_HOLD_OFF;
+
+	if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) {
+		vcn_v5_0_1_stop_dpg_mode(vinst);
+		return 0;
+	}
+
+	/* wait for vcn idle */
+	r = SOC15_WAIT_ON_RREG(VCN, vcn_inst, regUVD_STATUS, UVD_STATUS__IDLE, 0x7);
+	if (r)
+		return r;
+
+	tmp = UVD_LMI_STATUS__VCPU_LMI_WRITE_CLEAN_MASK |
+		UVD_LMI_STATUS__READ_CLEAN_MASK |
+		UVD_LMI_STATUS__WRITE_CLEAN_MASK |
+		UVD_LMI_STATUS__WRITE_CLEAN_RAW_MASK;
+	r = SOC15_WAIT_ON_RREG(VCN, vcn_inst, regUVD_LMI_STATUS, tmp, tmp);
+	if (r)
+		return r;
+
+	/* disable LMI UMC channel */
+	tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_LMI_CTRL2);
+	tmp |= UVD_LMI_CTRL2__STALL_ARB_UMC_MASK;
+	WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_CTRL2, tmp);
+	tmp = UVD_LMI_STATUS__UMC_READ_CLEAN_RAW_MASK |
+		UVD_LMI_STATUS__UMC_WRITE_CLEAN_RAW_MASK;
+	r = SOC15_WAIT_ON_RREG(VCN, vcn_inst, regUVD_LMI_STATUS, tmp, tmp);
+	if (r)
+		return r;
+
+	/* block VCPU register access */
+	WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_RB_ARB_CTRL),
+		 UVD_RB_ARB_CTRL__VCPU_DIS_MASK,
+		 ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK);
+
+	/* reset VCPU */
+	WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_VCPU_CNTL),
+		 UVD_VCPU_CNTL__BLK_RST_MASK,
+		 ~UVD_VCPU_CNTL__BLK_RST_MASK);
+
+	/* disable VCPU clock */
+	WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_VCPU_CNTL), 0,
+		 ~(UVD_VCPU_CNTL__CLK_EN_MASK));
+
+	/* apply soft reset */
+	tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET);
+	tmp |= UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK;
+	WREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET, tmp);
+	tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET);
+	tmp |= UVD_SOFT_RESET__LMI_SOFT_RESET_MASK;
+	WREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET, tmp);
+
+	/* clear status */
+	WREG32_SOC15(VCN, vcn_inst, regUVD_STATUS, 0);
+
+	/* Keeping one read-back to ensure all register writes are done,
+	 * otherwise it may introduce race conditions.
+	 */
+	RREG32_SOC15(VCN, vcn_inst, regUVD_STATUS);
+
+	return 0;
+}
+
+/**
+ * vcn_v5_0_1_unified_ring_get_rptr - get unified read pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Returns the current hardware unified read pointer
+ */
+static uint64_t vcn_v5_0_1_unified_ring_get_rptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	if (ring != &adev->vcn.inst[ring->me].ring_enc[0])
+		DRM_ERROR("wrong ring id is identified in %s", __func__);
+
+	return RREG32_SOC15(VCN, GET_INST(VCN, ring->me), regUVD_RB_RPTR);
+}
+
+/**
+ * vcn_v5_0_1_unified_ring_get_wptr - get unified write pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Returns the current hardware unified write pointer
+ */
+static uint64_t vcn_v5_0_1_unified_ring_get_wptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	if (ring != &adev->vcn.inst[ring->me].ring_enc[0])
+		DRM_ERROR("wrong ring id is identified in %s", __func__);
+
+	if (ring->use_doorbell)
+		return *ring->wptr_cpu_addr;
+	else
+		return RREG32_SOC15(VCN, GET_INST(VCN, ring->me), regUVD_RB_WPTR);
+}
+
+/**
+ * vcn_v5_0_1_unified_ring_set_wptr - set enc write pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Commits the enc write pointer to the hardware
+ */
+static void vcn_v5_0_1_unified_ring_set_wptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	if (ring != &adev->vcn.inst[ring->me].ring_enc[0])
+		DRM_ERROR("wrong ring id is identified in %s", __func__);
+
+	if (ring->use_doorbell) {
+		*ring->wptr_cpu_addr = lower_32_bits(ring->wptr);
+		WDOORBELL32(ring->doorbell_index, lower_32_bits(ring->wptr));
+	} else {
+		WREG32_SOC15(VCN, GET_INST(VCN, ring->me), regUVD_RB_WPTR,
+				lower_32_bits(ring->wptr));
+	}
+}
+
+static int vcn_v5_0_1_ring_reset(struct amdgpu_ring *ring,
+				 unsigned int vmid,
+				 struct amdgpu_fence *timedout_fence)
+{
+	int r = 0;
+	int vcn_inst;
+	struct amdgpu_device *adev = ring->adev;
+	struct amdgpu_vcn_inst *vinst = &adev->vcn.inst[ring->me];
+
+	amdgpu_ring_reset_helper_begin(ring, timedout_fence);
+
+	vcn_inst = GET_INST(VCN, ring->me);
+	r = amdgpu_dpm_reset_vcn(adev, 1 << vcn_inst);
+
+	if (r) {
+		DRM_DEV_ERROR(adev->dev, "VCN reset fail : %d\n", r);
+		return r;
+	}
+
+	vcn_v5_0_1_hw_init_inst(adev, ring->me);
+	vcn_v5_0_1_start_dpg_mode(vinst, vinst->indirect_sram);
+
+	return amdgpu_ring_reset_helper_end(ring, timedout_fence);
+}
+
+static const struct amdgpu_ring_funcs vcn_v5_0_1_unified_ring_vm_funcs = {
+	.type = AMDGPU_RING_TYPE_VCN_ENC,
+	.align_mask = 0x3f,
+	.nop = VCN_ENC_CMD_NO_OP,
+	.get_rptr = vcn_v5_0_1_unified_ring_get_rptr,
+	.get_wptr = vcn_v5_0_1_unified_ring_get_wptr,
+	.set_wptr = vcn_v5_0_1_unified_ring_set_wptr,
+	.emit_frame_size = SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 +
+			   SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 4 +
+			   4 + /* vcn_v2_0_enc_ring_emit_vm_flush */
+			   5 +
+			   5 + /* vcn_v2_0_enc_ring_emit_fence x2 vm fence */
+			   1, /* vcn_v2_0_enc_ring_insert_end */
+	.emit_ib_size = 5, /* vcn_v2_0_enc_ring_emit_ib */
+	.emit_ib = vcn_v2_0_enc_ring_emit_ib,
+	.emit_fence = vcn_v2_0_enc_ring_emit_fence,
+	.emit_vm_flush = vcn_v4_0_3_enc_ring_emit_vm_flush,
+	.emit_hdp_flush = vcn_v4_0_3_ring_emit_hdp_flush,
+	.test_ring = amdgpu_vcn_enc_ring_test_ring,
+	.test_ib = amdgpu_vcn_unified_ring_test_ib,
+	.insert_nop = amdgpu_ring_insert_nop,
+	.insert_end = vcn_v2_0_enc_ring_insert_end,
+	.pad_ib = amdgpu_ring_generic_pad_ib,
+	.begin_use = amdgpu_vcn_ring_begin_use,
+	.end_use = amdgpu_vcn_ring_end_use,
+	.emit_wreg = vcn_v4_0_3_enc_ring_emit_wreg,
+	.emit_reg_wait = vcn_v4_0_3_enc_ring_emit_reg_wait,
+	.emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper,
+	.reset = vcn_v5_0_1_ring_reset,
+};
+
+/**
+ * vcn_v5_0_1_set_unified_ring_funcs - set unified ring functions
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Set unified ring functions
+ */
+static void vcn_v5_0_1_set_unified_ring_funcs(struct amdgpu_device *adev)
+{
+	int i, vcn_inst;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		adev->vcn.inst[i].ring_enc[0].funcs = &vcn_v5_0_1_unified_ring_vm_funcs;
+		adev->vcn.inst[i].ring_enc[0].me = i;
+		vcn_inst = GET_INST(VCN, i);
+		adev->vcn.inst[i].aid_id = vcn_inst / adev->vcn.num_inst_per_aid;
+	}
+}
+
+/**
+ * vcn_v5_0_1_is_idle - check VCN block is idle
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block structure
+ *
+ * Check whether VCN block is idle
+ */
+static bool vcn_v5_0_1_is_idle(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, ret = 1;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i)
+		ret &= (RREG32_SOC15(VCN, GET_INST(VCN, i), regUVD_STATUS) == UVD_STATUS__IDLE);
+
+	return ret;
+}
+
+/**
+ * vcn_v5_0_1_wait_for_idle - wait for VCN block idle
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ *
+ * Wait for VCN block idle
+ */
+static int vcn_v5_0_1_wait_for_idle(struct amdgpu_ip_block *ip_block)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	int i, ret = 0;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		ret = SOC15_WAIT_ON_RREG(VCN, GET_INST(VCN, i), regUVD_STATUS, UVD_STATUS__IDLE,
+			UVD_STATUS__IDLE);
+		if (ret)
+			return ret;
+	}
+
+	return ret;
+}
+
+/**
+ * vcn_v5_0_1_set_clockgating_state - set VCN block clockgating state
+ *
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
+ * @state: clock gating state
+ *
+ * Set VCN block clockgating state
+ */
+static int vcn_v5_0_1_set_clockgating_state(struct amdgpu_ip_block *ip_block,
+					    enum amd_clockgating_state state)
+{
+	struct amdgpu_device *adev = ip_block->adev;
+	bool enable = state == AMD_CG_STATE_GATE;
+	int i;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		struct amdgpu_vcn_inst *vinst = &adev->vcn.inst[i];
+
+		if (enable) {
+			if (RREG32_SOC15(VCN, GET_INST(VCN, i), regUVD_STATUS) != UVD_STATUS__IDLE)
+				return -EBUSY;
+			vcn_v5_0_1_enable_clock_gating(vinst);
+		} else {
+			vcn_v5_0_1_disable_clock_gating(vinst);
+		}
+	}
+
+	return 0;
+}
+
+static int vcn_v5_0_1_set_pg_state(struct amdgpu_vcn_inst *vinst,
+				   enum amd_powergating_state state)
+{
+	struct amdgpu_device *adev = vinst->adev;
+	int ret = 0;
+
+	/* for SRIOV, guest should not control VCN Power-gating
+	 * MMSCH FW should control Power-gating and clock-gating
+	 * guest should avoid touching CGC and PG
+	 */
+	if (amdgpu_sriov_vf(adev)) {
+		vinst->cur_state = AMD_PG_STATE_UNGATE;
+		return 0;
+	}
+
+	if (state == vinst->cur_state)
+		return 0;
+
+	if (state == AMD_PG_STATE_GATE)
+		ret = vcn_v5_0_1_stop(vinst);
+	else
+		ret = vcn_v5_0_1_start(vinst);
+
+	if (!ret)
+		vinst->cur_state = state;
+
+	return ret;
+}
+
+/**
+ * vcn_v5_0_1_process_interrupt - process VCN block interrupt
+ *
+ * @adev: amdgpu_device pointer
+ * @source: interrupt sources
+ * @entry: interrupt entry from clients and sources
+ *
+ * Process VCN block interrupt
+ */
+static int vcn_v5_0_1_process_interrupt(struct amdgpu_device *adev, struct amdgpu_irq_src *source,
+	struct amdgpu_iv_entry *entry)
+{
+	uint32_t i, inst;
+
+	i = node_id_to_phys_map[entry->node_id];
+
+	DRM_DEV_DEBUG(adev->dev, "IH: VCN TRAP\n");
+
+	for (inst = 0; inst < adev->vcn.num_vcn_inst; ++inst)
+		if (adev->vcn.inst[inst].aid_id == i)
+			break;
+	if (inst >= adev->vcn.num_vcn_inst) {
+		dev_WARN_ONCE(adev->dev, 1,
+				"Interrupt received for unknown VCN instance %d",
+				entry->node_id);
+		return 0;
+	}
+
+	switch (entry->src_id) {
+	case VCN_5_0__SRCID__UVD_ENC_GENERAL_PURPOSE:
+		amdgpu_fence_process(&adev->vcn.inst[inst].ring_enc[0]);
+		break;
+	default:
+		DRM_DEV_ERROR(adev->dev, "Unhandled interrupt: %d %d\n",
+			  entry->src_id, entry->src_data[0]);
+		break;
+	}
+
+	return 0;
+}
+
+static int vcn_v5_0_1_set_ras_interrupt_state(struct amdgpu_device *adev,
+					struct amdgpu_irq_src *source,
+					unsigned int type,
+					enum amdgpu_interrupt_state state)
+{
+	return 0;
+}
+
+static const struct amdgpu_irq_src_funcs vcn_v5_0_1_irq_funcs = {
+	.process = vcn_v5_0_1_process_interrupt,
+};
+
+static const struct amdgpu_irq_src_funcs vcn_v5_0_1_ras_irq_funcs = {
+	.set = vcn_v5_0_1_set_ras_interrupt_state,
+	.process = amdgpu_vcn_process_poison_irq,
+};
+
+
+/**
+ * vcn_v5_0_1_set_irq_funcs - set VCN block interrupt irq functions
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Set VCN block interrupt irq functions
+ */
+static void vcn_v5_0_1_set_irq_funcs(struct amdgpu_device *adev)
+{
+	int i;
+
+	for (i = 0; i < adev->vcn.num_vcn_inst; ++i)
+		adev->vcn.inst->irq.num_types++;
+
+	adev->vcn.inst->irq.funcs = &vcn_v5_0_1_irq_funcs;
+
+	adev->vcn.inst->ras_poison_irq.num_types = 1;
+	adev->vcn.inst->ras_poison_irq.funcs = &vcn_v5_0_1_ras_irq_funcs;
+
+}
+
+static const struct amd_ip_funcs vcn_v5_0_1_ip_funcs = {
+	.name = "vcn_v5_0_1",
+	.early_init = vcn_v5_0_1_early_init,
+	.late_init = vcn_v5_0_1_late_init,
+	.sw_init = vcn_v5_0_1_sw_init,
+	.sw_fini = vcn_v5_0_1_sw_fini,
+	.hw_init = vcn_v5_0_1_hw_init,
+	.hw_fini = vcn_v5_0_1_hw_fini,
+	.suspend = vcn_v5_0_1_suspend,
+	.resume = vcn_v5_0_1_resume,
+	.is_idle = vcn_v5_0_1_is_idle,
+	.wait_for_idle = vcn_v5_0_1_wait_for_idle,
+	.check_soft_reset = NULL,
+	.pre_soft_reset = NULL,
+	.soft_reset = NULL,
+	.post_soft_reset = NULL,
+	.set_clockgating_state = vcn_v5_0_1_set_clockgating_state,
+	.set_powergating_state = vcn_set_powergating_state,
+	.dump_ip_state = amdgpu_vcn_dump_ip_state,
+	.print_ip_state = amdgpu_vcn_print_ip_state,
+};
+
+const struct amdgpu_ip_block_version vcn_v5_0_1_ip_block = {
+	.type = AMD_IP_BLOCK_TYPE_VCN,
+	.major = 5,
+	.minor = 0,
+	.rev = 1,
+	.funcs = &vcn_v5_0_1_ip_funcs,
+};
+
+static uint32_t vcn_v5_0_1_query_poison_by_instance(struct amdgpu_device *adev,
+			uint32_t instance, uint32_t sub_block)
+{
+	uint32_t poison_stat = 0, reg_value = 0;
+
+	switch (sub_block) {
+	case AMDGPU_VCN_V5_0_1_VCPU_VCODEC:
+		reg_value = RREG32_SOC15(VCN, instance, regUVD_RAS_VCPU_VCODEC_STATUS);
+		poison_stat = REG_GET_FIELD(reg_value, UVD_RAS_VCPU_VCODEC_STATUS, POISONED_PF);
+		break;
+	default:
+		break;
+	}
+
+	if (poison_stat)
+		dev_info(adev->dev, "Poison detected in VCN%d, sub_block%d\n",
+			instance, sub_block);
+
+	return poison_stat;
+}
+
+static bool vcn_v5_0_1_query_poison_status(struct amdgpu_device *adev)
+{
+	uint32_t inst, sub;
+	uint32_t poison_stat = 0;
+
+	for (inst = 0; inst < adev->vcn.num_vcn_inst; inst++)
+		for (sub = 0; sub < AMDGPU_VCN_V5_0_1_MAX_SUB_BLOCK; sub++)
+			poison_stat +=
+			vcn_v5_0_1_query_poison_by_instance(adev, inst, sub);
+
+	return !!poison_stat;
+}
+
+static const struct amdgpu_ras_block_hw_ops vcn_v5_0_1_ras_hw_ops = {
+	.query_poison_status = vcn_v5_0_1_query_poison_status,
+};
+
+static int vcn_v5_0_1_aca_bank_parser(struct aca_handle *handle, struct aca_bank *bank,
+				      enum aca_smu_type type, void *data)
+{
+	struct aca_bank_info info;
+	u64 misc0;
+	int ret;
+
+	ret = aca_bank_info_decode(bank, &info);
+	if (ret)
+		return ret;
+
+	misc0 = bank->regs[ACA_REG_IDX_MISC0];
+	switch (type) {
+	case ACA_SMU_TYPE_UE:
+		bank->aca_err_type = ACA_ERROR_TYPE_UE;
+		ret = aca_error_cache_log_bank_error(handle, &info, ACA_ERROR_TYPE_UE,
+						     1ULL);
+		break;
+	case ACA_SMU_TYPE_CE:
+		bank->aca_err_type = ACA_ERROR_TYPE_CE;
+		ret = aca_error_cache_log_bank_error(handle, &info, bank->aca_err_type,
+						     ACA_REG__MISC0__ERRCNT(misc0));
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return ret;
+}
+
+/* reference to smu driver if header file */
+static int vcn_v5_0_1_err_codes[] = {
+	14, 15, 47, /* VCN [D|V|S] */
+};
+
+static bool vcn_v5_0_1_aca_bank_is_valid(struct aca_handle *handle, struct aca_bank *bank,
+					 enum aca_smu_type type, void *data)
+{
+	u32 instlo;
+
+	instlo = ACA_REG__IPID__INSTANCEIDLO(bank->regs[ACA_REG_IDX_IPID]);
+	instlo &= GENMASK(31, 1);
+
+	if (instlo != mmSMNAID_AID0_MCA_SMU)
+		return false;
+
+	if (aca_bank_check_error_codes(handle->adev, bank,
+				       vcn_v5_0_1_err_codes,
+				       ARRAY_SIZE(vcn_v5_0_1_err_codes)))
+		return false;
+
+	return true;
+}
+
+static const struct aca_bank_ops vcn_v5_0_1_aca_bank_ops = {
+	.aca_bank_parser = vcn_v5_0_1_aca_bank_parser,
+	.aca_bank_is_valid = vcn_v5_0_1_aca_bank_is_valid,
+};
+
+static const struct aca_info vcn_v5_0_1_aca_info = {
+	.hwip = ACA_HWIP_TYPE_SMU,
+	.mask = ACA_ERROR_UE_MASK,
+	.bank_ops = &vcn_v5_0_1_aca_bank_ops,
+};
+
+static int vcn_v5_0_1_ras_late_init(struct amdgpu_device *adev, struct ras_common_if *ras_block)
+{
+	int r;
+
+	r = amdgpu_ras_block_late_init(adev, ras_block);
+	if (r)
+		return r;
+
+	r = amdgpu_ras_bind_aca(adev, AMDGPU_RAS_BLOCK__VCN,
+				&vcn_v5_0_1_aca_info, NULL);
+	if (r)
+		goto late_fini;
+
+	if (amdgpu_ras_is_supported(adev, ras_block->block) &&
+		adev->vcn.inst->ras_poison_irq.funcs) {
+		r = amdgpu_irq_get(adev, &adev->vcn.inst->ras_poison_irq, 0);
+		if (r)
+			goto late_fini;
+	}
+
+	return 0;
+
+late_fini:
+	amdgpu_ras_block_late_fini(adev, ras_block);
+
+	return r;
+}
+
+static struct amdgpu_vcn_ras vcn_v5_0_1_ras = {
+	.ras_block = {
+		.hw_ops = &vcn_v5_0_1_ras_hw_ops,
+		.ras_late_init = vcn_v5_0_1_ras_late_init,
+	},
+};
+
+static void vcn_v5_0_1_set_ras_funcs(struct amdgpu_device *adev)
+{
+	adev->vcn.ras = &vcn_v5_0_1_ras;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.h b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.h
new file mode 100644
index 000000000000..b72e4da68317
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.h
@@ -0,0 +1,39 @@
+/*
+ * Copyright 2024 Advanced Micro Devices, Inc. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __VCN_v5_0_1_H__
+#define __VCN_v5_0_1_H__
+
+#define regVCN_RRMT_CNTL                          0x0940
+#define regVCN_RRMT_CNTL_BASE_IDX                 1
+
+
+enum amdgpu_vcn_v5_0_1_sub_block {
+	AMDGPU_VCN_V5_0_1_VCPU_VCODEC = 0,
+
+	AMDGPU_VCN_V5_0_1_MAX_SUB_BLOCK,
+};
+
+extern const struct amdgpu_ip_block_version vcn_v5_0_1_ip_block;
+
+#endif /* __VCN_v5_0_1_H__ */
diff --git a/drivers/gpu/drm/amd/amdgpu/vega10_ih.c b/drivers/gpu/drm/amd/amdgpu/vega10_ih.c
index a9ca6988009e..eb16916c6473 100644
--- a/drivers/gpu/drm/amd/amdgpu/vega10_ih.c
+++ b/drivers/gpu/drm/amd/amdgpu/vega10_ih.c
@@ -289,6 +289,10 @@ static int vega10_ih_irq_init(struct amdgpu_device *adev)
 		}
 	}
 
+	if (!amdgpu_sriov_vf(adev))
+		adev->nbio.funcs->ih_doorbell_range(adev, adev->irq.ih.use_doorbell,
+						    adev->irq.ih.doorbell_index);
+
 	pci_set_master(adev->pdev);
 
 	/* enable interrupts */
@@ -334,9 +338,11 @@ static u32 vega10_ih_get_wptr(struct amdgpu_device *adev,
 	u32 wptr, tmp;
 	struct amdgpu_ih_regs *ih_regs;
 
-	if (ih == &adev->irq.ih) {
+	if (ih == &adev->irq.ih || ih == &adev->irq.ih_soft) {
 		/* Only ring0 supports writeback. On other rings fall back
 		 * to register-based code with overflow checking below.
+		 * ih_soft ring doesn't have any backing hardware registers,
+		 * update wptr and return.
 		 */
 		wptr = le32_to_cpu(*ih->wptr_cpu);
 
@@ -358,15 +364,20 @@ static u32 vega10_ih_get_wptr(struct amdgpu_device *adev,
 	 * this should allow us to catchup.
 	 */
 	tmp = (wptr + 32) & ih->ptr_mask;
-	dev_warn(adev->dev, "IH ring buffer overflow "
-		 "(0x%08X, 0x%08X, 0x%08X)\n",
-		 wptr, ih->rptr, tmp);
+	dev_warn_ratelimited(adev->dev, "%s ring buffer overflow (0x%08X, 0x%08X, 0x%08X)\n",
+			     amdgpu_ih_ring_name(adev, ih), wptr, ih->rptr, tmp);
 	ih->rptr = tmp;
 
 	tmp = RREG32_NO_KIQ(ih_regs->ih_rb_cntl);
 	tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 1);
 	WREG32_NO_KIQ(ih_regs->ih_rb_cntl, tmp);
 
+	/* Unset the CLEAR_OVERFLOW bit immediately so new overflows
+	 * can be detected.
+	 */
+	tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 0);
+	WREG32_NO_KIQ(ih_regs->ih_rb_cntl, tmp);
+
 out:
 	return (wptr & ih->ptr_mask);
 }
@@ -409,6 +420,9 @@ static void vega10_ih_set_rptr(struct amdgpu_device *adev,
 {
 	struct amdgpu_ih_regs *ih_regs;
 
+	if (ih == &adev->irq.ih_soft)
+		return;
+
 	if (ih->use_doorbell) {
 		/* XXX check if swapping is necessary on BE */
 		*ih->rptr_cpu = ih->rptr;
@@ -457,18 +471,18 @@ static void vega10_ih_set_self_irq_funcs(struct amdgpu_device *adev)
 	adev->irq.self_irq.funcs = &vega10_ih_self_irq_funcs;
 }
 
-static int vega10_ih_early_init(void *handle)
+static int vega10_ih_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	vega10_ih_set_interrupt_funcs(adev);
 	vega10_ih_set_self_irq_funcs(adev);
 	return 0;
 }
 
-static int vega10_ih_sw_init(void *handle)
+static int vega10_ih_sw_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int r;
 
 	r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_IH, 0,
@@ -476,7 +490,7 @@ static int vega10_ih_sw_init(void *handle)
 	if (r)
 		return r;
 
-	r = amdgpu_ih_ring_init(adev, &adev->irq.ih, 256 * 1024, true);
+	r = amdgpu_ih_ring_init(adev, &adev->irq.ih, IH_RING_SIZE, true);
 	if (r)
 		return r;
 
@@ -501,7 +515,7 @@ static int vega10_ih_sw_init(void *handle)
 	/* initialize ih control registers offset */
 	vega10_ih_init_register_offset(adev);
 
-	r = amdgpu_ih_ring_init(adev, &adev->irq.ih_soft, PAGE_SIZE, true);
+	r = amdgpu_ih_ring_init(adev, &adev->irq.ih_soft, IH_SW_RING_SIZE, true);
 	if (r)
 		return r;
 
@@ -510,63 +524,50 @@ static int vega10_ih_sw_init(void *handle)
 	return r;
 }
 
-static int vega10_ih_sw_fini(void *handle)
+static int vega10_ih_sw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	amdgpu_irq_fini_sw(adev);
 
 	return 0;
 }
 
-static int vega10_ih_hw_init(void *handle)
+static int vega10_ih_hw_init(struct amdgpu_ip_block *ip_block)
 {
-	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	r = vega10_ih_irq_init(adev);
-	if (r)
-		return r;
-
-	return 0;
+	return vega10_ih_irq_init(ip_block->adev);
 }
 
-static int vega10_ih_hw_fini(void *handle)
+static int vega10_ih_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	vega10_ih_irq_disable(adev);
+	vega10_ih_irq_disable(ip_block->adev);
 
 	return 0;
 }
 
-static int vega10_ih_suspend(void *handle)
+static int vega10_ih_suspend(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	return vega10_ih_hw_fini(adev);
+	return vega10_ih_hw_fini(ip_block);
 }
 
-static int vega10_ih_resume(void *handle)
+static int vega10_ih_resume(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	return vega10_ih_hw_init(adev);
+	return vega10_ih_hw_init(ip_block);
 }
 
-static bool vega10_ih_is_idle(void *handle)
+static bool vega10_ih_is_idle(struct amdgpu_ip_block *ip_block)
 {
 	/* todo */
 	return true;
 }
 
-static int vega10_ih_wait_for_idle(void *handle)
+static int vega10_ih_wait_for_idle(struct amdgpu_ip_block *ip_block)
 {
 	/* todo */
 	return -ETIMEDOUT;
 }
 
-static int vega10_ih_soft_reset(void *handle)
+static int vega10_ih_soft_reset(struct amdgpu_ip_block *ip_block)
 {
 	/* todo */
 
@@ -603,10 +604,10 @@ static void vega10_ih_update_clockgating_state(struct amdgpu_device *adev,
 	}
 }
 
-static int vega10_ih_set_clockgating_state(void *handle,
+static int vega10_ih_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_clockgating_state state)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	vega10_ih_update_clockgating_state(adev,
 				state == AMD_CG_STATE_GATE);
@@ -614,7 +615,7 @@ static int vega10_ih_set_clockgating_state(void *handle,
 
 }
 
-static int vega10_ih_set_powergating_state(void *handle,
+static int vega10_ih_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_powergating_state state)
 {
 	return 0;
@@ -623,7 +624,6 @@ static int vega10_ih_set_powergating_state(void *handle,
 const struct amd_ip_funcs vega10_ih_ip_funcs = {
 	.name = "vega10_ih",
 	.early_init = vega10_ih_early_init,
-	.late_init = NULL,
 	.sw_init = vega10_ih_sw_init,
 	.sw_fini = vega10_ih_sw_fini,
 	.hw_init = vega10_ih_hw_init,
@@ -640,6 +640,7 @@ const struct amd_ip_funcs vega10_ih_ip_funcs = {
 static const struct amdgpu_ih_funcs vega10_ih_funcs = {
 	.get_wptr = vega10_ih_get_wptr,
 	.decode_iv = amdgpu_ih_decode_iv_helper,
+	.decode_iv_ts = amdgpu_ih_decode_iv_ts_helper,
 	.set_rptr = vega10_ih_set_rptr
 };
 
diff --git a/drivers/gpu/drm/amd/amdgpu/vega20_ih.c b/drivers/gpu/drm/amd/amdgpu/vega20_ih.c
index f51dfc38ac65..85846fd08ce4 100644
--- a/drivers/gpu/drm/amd/amdgpu/vega20_ih.c
+++ b/drivers/gpu/drm/amd/amdgpu/vega20_ih.c
@@ -38,6 +38,11 @@
 #define mmIH_CHICKEN_ALDEBARAN			0x18d
 #define mmIH_CHICKEN_ALDEBARAN_BASE_IDX		0
 
+#define mmIH_RETRY_INT_CAM_CNTL_ALDEBARAN		0x00ea
+#define mmIH_RETRY_INT_CAM_CNTL_ALDEBARAN_BASE_IDX	0
+#define IH_RETRY_INT_CAM_CNTL_ALDEBARAN__ENABLE__SHIFT	0x10
+#define IH_RETRY_INT_CAM_CNTL_ALDEBARAN__ENABLE_MASK	0x00010000L
+
 static void vega20_ih_set_interrupt_funcs(struct amdgpu_device *adev);
 
 /**
@@ -109,6 +114,33 @@ static int vega20_ih_toggle_ring_interrupts(struct amdgpu_device *adev,
 	tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, RB_ENABLE, (enable ? 1 : 0));
 	tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, RB_GPU_TS_ENABLE, 1);
 
+	if (enable) {
+		/* Unset the CLEAR_OVERFLOW bit to make sure the next step
+		 * is switching the bit from 0 to 1
+		 */
+		tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 0);
+		if (amdgpu_sriov_vf(adev) && amdgpu_sriov_reg_indirect_ih(adev)) {
+			if (psp_reg_program(&adev->psp, ih_regs->psp_reg_id, tmp))
+				return -ETIMEDOUT;
+		} else {
+			WREG32_NO_KIQ(ih_regs->ih_rb_cntl, tmp);
+		}
+
+		/* Clear RB_OVERFLOW bit */
+		tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 1);
+		if (amdgpu_sriov_vf(adev) && amdgpu_sriov_reg_indirect_ih(adev)) {
+			if (psp_reg_program(&adev->psp, ih_regs->psp_reg_id, tmp))
+				return -ETIMEDOUT;
+		} else {
+			WREG32_NO_KIQ(ih_regs->ih_rb_cntl, tmp);
+		}
+
+		/* Unset the CLEAR_OVERFLOW bit immediately so new overflows
+		 * can be detected.
+		 */
+		tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 0);
+	}
+
 	/* enable_intr field is only valid in ring0 */
 	if (ih == &adev->irq.ih)
 		tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, ENABLE_INTR, (enable ? 1 : 0));
@@ -251,36 +283,14 @@ static int vega20_ih_enable_ring(struct amdgpu_device *adev,
 	return 0;
 }
 
-/**
- * vega20_ih_reroute_ih - reroute VMC/UTCL2 ih to an ih ring
- *
- * @adev: amdgpu_device pointer
- *
- * Reroute VMC and UMC interrupts on primary ih ring to
- * ih ring 1 so they won't lose when bunches of page faults
- * interrupts overwhelms the interrupt handler(VEGA20)
- */
-static void vega20_ih_reroute_ih(struct amdgpu_device *adev)
+static uint32_t vega20_setup_retry_doorbell(u32 doorbell_index)
 {
-	uint32_t tmp;
+	u32 val = 0;
 
-	/* vega20 ih reroute will go through psp this
-	 * function is used for newer asics starting arcturus
-	 */
-	if (adev->asic_type >= CHIP_ARCTURUS) {
-		/* Reroute to IH ring 1 for VMC */
-		WREG32_SOC15(OSSSYS, 0, mmIH_CLIENT_CFG_INDEX, 0x12);
-		tmp = RREG32_SOC15(OSSSYS, 0, mmIH_CLIENT_CFG_DATA);
-		tmp = REG_SET_FIELD(tmp, IH_CLIENT_CFG_DATA, CLIENT_TYPE, 1);
-		tmp = REG_SET_FIELD(tmp, IH_CLIENT_CFG_DATA, RING_ID, 1);
-		WREG32_SOC15(OSSSYS, 0, mmIH_CLIENT_CFG_DATA, tmp);
-
-		/* Reroute IH ring 1 for UTCL2 */
-		WREG32_SOC15(OSSSYS, 0, mmIH_CLIENT_CFG_INDEX, 0x1B);
-		tmp = RREG32_SOC15(OSSSYS, 0, mmIH_CLIENT_CFG_DATA);
-		tmp = REG_SET_FIELD(tmp, IH_CLIENT_CFG_DATA, RING_ID, 1);
-		WREG32_SOC15(OSSSYS, 0, mmIH_CLIENT_CFG_DATA, tmp);
-	}
+	val = REG_SET_FIELD(val, IH_DOORBELL_RPTR, OFFSET, doorbell_index);
+	val = REG_SET_FIELD(val, IH_DOORBELL_RPTR, ENABLE, 1);
+
+	return val;
 }
 
 /**
@@ -308,40 +318,64 @@ static int vega20_ih_irq_init(struct amdgpu_device *adev)
 
 	adev->nbio.funcs->ih_control(adev);
 
-	if (adev->asic_type == CHIP_ARCTURUS &&
-	    adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT) {
-		ih_chicken = RREG32_SOC15(OSSSYS, 0, mmIH_CHICKEN);
-		if (adev->irq.ih.use_bus_addr) {
-			ih_chicken = REG_SET_FIELD(ih_chicken, IH_CHICKEN,
-						   MC_SPACE_GPA_ENABLE, 1);
+	if (!amdgpu_sriov_vf(adev)) {
+		if ((amdgpu_ip_version(adev, OSSSYS_HWIP, 0) == IP_VERSION(4, 2, 1)) &&
+		    adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT) {
+			ih_chicken = RREG32_SOC15(OSSSYS, 0, mmIH_CHICKEN);
+			if (adev->irq.ih.use_bus_addr) {
+				ih_chicken = REG_SET_FIELD(ih_chicken, IH_CHICKEN,
+							   MC_SPACE_GPA_ENABLE, 1);
+			}
+			WREG32_SOC15(OSSSYS, 0, mmIH_CHICKEN, ih_chicken);
 		}
-		WREG32_SOC15(OSSSYS, 0, mmIH_CHICKEN, ih_chicken);
-	}
 
-	/* psp firmware won't program IH_CHICKEN for aldebaran
-	 * driver needs to program it properly according to
-	 * MC_SPACE type in IH_RB_CNTL */
-	if (adev->asic_type == CHIP_ALDEBARAN) {
-		ih_chicken = RREG32_SOC15(OSSSYS, 0, mmIH_CHICKEN_ALDEBARAN);
-		if (adev->irq.ih.use_bus_addr) {
-			ih_chicken = REG_SET_FIELD(ih_chicken, IH_CHICKEN,
-						   MC_SPACE_GPA_ENABLE, 1);
+		/* psp firmware won't program IH_CHICKEN for aldebaran
+		 * driver needs to program it properly according to
+		 * MC_SPACE type in IH_RB_CNTL */
+		if ((amdgpu_ip_version(adev, OSSSYS_HWIP, 0) == IP_VERSION(4, 4, 0)) ||
+		    (amdgpu_ip_version(adev, OSSSYS_HWIP, 0) == IP_VERSION(4, 4, 2)) ||
+		    (amdgpu_ip_version(adev, OSSSYS_HWIP, 0) == IP_VERSION(4, 4, 5))) {
+			ih_chicken = RREG32_SOC15(OSSSYS, 0, mmIH_CHICKEN_ALDEBARAN);
+			if (adev->irq.ih.use_bus_addr) {
+				ih_chicken = REG_SET_FIELD(ih_chicken, IH_CHICKEN,
+							   MC_SPACE_GPA_ENABLE, 1);
+			}
+			WREG32_SOC15(OSSSYS, 0, mmIH_CHICKEN_ALDEBARAN, ih_chicken);
 		}
-		WREG32_SOC15(OSSSYS, 0, mmIH_CHICKEN_ALDEBARAN, ih_chicken);
 	}
 
 	for (i = 0; i < ARRAY_SIZE(ih); i++) {
 		if (ih[i]->ring_size) {
-			if (i == 1)
-				vega20_ih_reroute_ih(adev);
 			ret = vega20_ih_enable_ring(adev, ih[i]);
 			if (ret)
 				return ret;
 		}
+		ih[i]->overflow = false;
 	}
 
+	if (!amdgpu_sriov_vf(adev))
+		adev->nbio.funcs->ih_doorbell_range(adev, adev->irq.ih.use_doorbell,
+						    adev->irq.ih.doorbell_index);
+
 	pci_set_master(adev->pdev);
 
+	/* Allocate the doorbell for IH Retry CAM */
+	adev->irq.retry_cam_doorbell_index = (adev->doorbell_index.ih + 3) << 1;
+	WREG32_SOC15(OSSSYS, 0, mmIH_DOORBELL_RETRY_CAM,
+		vega20_setup_retry_doorbell(adev->irq.retry_cam_doorbell_index));
+
+	/* Enable IH Retry CAM */
+	if (amdgpu_ip_version(adev, OSSSYS_HWIP, 0) == IP_VERSION(4, 4, 0) ||
+	    amdgpu_ip_version(adev, OSSSYS_HWIP, 0) == IP_VERSION(4, 4, 2) ||
+	    amdgpu_ip_version(adev, OSSSYS_HWIP, 0) == IP_VERSION(4, 4, 4) ||
+	    amdgpu_ip_version(adev, OSSSYS_HWIP, 0) == IP_VERSION(4, 4, 5))
+		WREG32_FIELD15(OSSSYS, 0, IH_RETRY_INT_CAM_CNTL_ALDEBARAN,
+			       ENABLE, 1);
+	else
+		WREG32_FIELD15(OSSSYS, 0, IH_RETRY_INT_CAM_CNTL, ENABLE, 1);
+
+	adev->irq.retry_cam_enabled = true;
+
 	/* enable interrupts */
 	ret = vega20_ih_toggle_interrupts(adev, true);
 	if (ret)
@@ -385,9 +419,11 @@ static u32 vega20_ih_get_wptr(struct amdgpu_device *adev,
 	u32 wptr, tmp;
 	struct amdgpu_ih_regs *ih_regs;
 
-	if (ih == &adev->irq.ih) {
+	if (ih == &adev->irq.ih || ih == &adev->irq.ih_soft) {
 		/* Only ring0 supports writeback. On other rings fall back
 		 * to register-based code with overflow checking below.
+		 * ih_soft ring doesn't have any backing hardware registers,
+		 * update wptr and return.
 		 */
 		wptr = le32_to_cpu(*ih->wptr_cpu);
 
@@ -402,22 +438,30 @@ static u32 vega20_ih_get_wptr(struct amdgpu_device *adev,
 	if (!REG_GET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW))
 		goto out;
 
-	wptr = REG_SET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW, 0);
+	if (!amdgpu_sriov_vf(adev))
+		wptr = REG_SET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW, 0);
+	else
+		ih->overflow = true;
 
 	/* When a ring buffer overflow happen start parsing interrupt
 	 * from the last not overwritten vector (wptr + 32). Hopefully
 	 * this should allow us to catchup.
 	 */
 	tmp = (wptr + 32) & ih->ptr_mask;
-	dev_warn(adev->dev, "IH ring buffer overflow "
-		 "(0x%08X, 0x%08X, 0x%08X)\n",
-		 wptr, ih->rptr, tmp);
+	dev_warn_ratelimited(adev->dev, "%s ring buffer overflow (0x%08X, 0x%08X, 0x%08X)\n",
+			     amdgpu_ih_ring_name(adev, ih), wptr, ih->rptr, tmp);
 	ih->rptr = tmp;
 
 	tmp = RREG32_NO_KIQ(ih_regs->ih_rb_cntl);
 	tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 1);
 	WREG32_NO_KIQ(ih_regs->ih_rb_cntl, tmp);
 
+	/* Unset the CLEAR_OVERFLOW bit immediately so new overflows
+	 * can be detected.
+	 */
+	tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 0);
+	WREG32_NO_KIQ(ih_regs->ih_rb_cntl, tmp);
+
 out:
 	return (wptr & ih->ptr_mask);
 }
@@ -461,6 +505,9 @@ static void vega20_ih_set_rptr(struct amdgpu_device *adev,
 {
 	struct amdgpu_ih_regs *ih_regs;
 
+	if (ih == &adev->irq.ih_soft)
+		return;
+
 	if (ih->use_doorbell) {
 		/* XXX check if swapping is necessary on BE */
 		*ih->rptr_cpu = ih->rptr;
@@ -494,7 +541,8 @@ static int vega20_ih_self_irq(struct amdgpu_device *adev,
 	case 2:
 		schedule_work(&adev->irq.ih2_work);
 		break;
-	default: break;
+	default:
+		break;
 	}
 	return 0;
 }
@@ -509,18 +557,19 @@ static void vega20_ih_set_self_irq_funcs(struct amdgpu_device *adev)
 	adev->irq.self_irq.funcs = &vega20_ih_self_irq_funcs;
 }
 
-static int vega20_ih_early_init(void *handle)
+static int vega20_ih_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	vega20_ih_set_interrupt_funcs(adev);
 	vega20_ih_set_self_irq_funcs(adev);
 	return 0;
 }
 
-static int vega20_ih_sw_init(void *handle)
+static int vega20_ih_sw_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
+	bool use_bus_addr = true;
 	int r;
 
 	r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_IH, 0,
@@ -528,31 +577,38 @@ static int vega20_ih_sw_init(void *handle)
 	if (r)
 		return r;
 
-	r = amdgpu_ih_ring_init(adev, &adev->irq.ih, 256 * 1024, true);
+	if ((adev->flags & AMD_IS_APU) &&
+	    (amdgpu_ip_version(adev, OSSSYS_HWIP, 0) == IP_VERSION(4, 4, 2)))
+		use_bus_addr = false;
+
+	r = amdgpu_ih_ring_init(adev, &adev->irq.ih, IH_RING_SIZE, use_bus_addr);
 	if (r)
 		return r;
 
 	adev->irq.ih.use_doorbell = true;
 	adev->irq.ih.doorbell_index = adev->doorbell_index.ih << 1;
 
-	r = amdgpu_ih_ring_init(adev, &adev->irq.ih1, PAGE_SIZE, true);
+	r = amdgpu_ih_ring_init(adev, &adev->irq.ih1, PAGE_SIZE, use_bus_addr);
 	if (r)
 		return r;
 
 	adev->irq.ih1.use_doorbell = true;
 	adev->irq.ih1.doorbell_index = (adev->doorbell_index.ih + 1) << 1;
 
-	r = amdgpu_ih_ring_init(adev, &adev->irq.ih2, PAGE_SIZE, true);
-	if (r)
-		return r;
+	if (amdgpu_ip_version(adev, OSSSYS_HWIP, 0) != IP_VERSION(4, 4, 2) &&
+	    amdgpu_ip_version(adev, OSSSYS_HWIP, 0) != IP_VERSION(4, 4, 5)) {
+		r = amdgpu_ih_ring_init(adev, &adev->irq.ih2, PAGE_SIZE, true);
+		if (r)
+			return r;
 
-	adev->irq.ih2.use_doorbell = true;
-	adev->irq.ih2.doorbell_index = (adev->doorbell_index.ih + 2) << 1;
+		adev->irq.ih2.use_doorbell = true;
+		adev->irq.ih2.doorbell_index = (adev->doorbell_index.ih + 2) << 1;
+	}
 
 	/* initialize ih control registers offset */
 	vega20_ih_init_register_offset(adev);
 
-	r = amdgpu_ih_ring_init(adev, &adev->irq.ih_soft, PAGE_SIZE, true);
+	r = amdgpu_ih_ring_init(adev, &adev->irq.ih_soft, IH_SW_RING_SIZE, use_bus_addr);
 	if (r)
 		return r;
 
@@ -561,19 +617,19 @@ static int vega20_ih_sw_init(void *handle)
 	return r;
 }
 
-static int vega20_ih_sw_fini(void *handle)
+static int vega20_ih_sw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	amdgpu_irq_fini_sw(adev);
 
 	return 0;
 }
 
-static int vega20_ih_hw_init(void *handle)
+static int vega20_ih_hw_init(struct amdgpu_ip_block *ip_block)
 {
 	int r;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	r = vega20_ih_irq_init(adev);
 	if (r)
@@ -582,42 +638,36 @@ static int vega20_ih_hw_init(void *handle)
 	return 0;
 }
 
-static int vega20_ih_hw_fini(void *handle)
+static int vega20_ih_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	vega20_ih_irq_disable(adev);
+	vega20_ih_irq_disable(ip_block->adev);
 
 	return 0;
 }
 
-static int vega20_ih_suspend(void *handle)
+static int vega20_ih_suspend(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	return vega20_ih_hw_fini(adev);
+	return vega20_ih_hw_fini(ip_block);
 }
 
-static int vega20_ih_resume(void *handle)
+static int vega20_ih_resume(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	return vega20_ih_hw_init(adev);
+	return vega20_ih_hw_init(ip_block);
 }
 
-static bool vega20_ih_is_idle(void *handle)
+static bool vega20_ih_is_idle(struct amdgpu_ip_block *ip_block)
 {
 	/* todo */
 	return true;
 }
 
-static int vega20_ih_wait_for_idle(void *handle)
+static int vega20_ih_wait_for_idle(struct amdgpu_ip_block *ip_block)
 {
 	/* todo */
 	return -ETIMEDOUT;
 }
 
-static int vega20_ih_soft_reset(void *handle)
+static int vega20_ih_soft_reset(struct amdgpu_ip_block *ip_block)
 {
 	/* todo */
 
@@ -651,10 +701,10 @@ static void vega20_ih_update_clockgating_state(struct amdgpu_device *adev,
 	}
 }
 
-static int vega20_ih_set_clockgating_state(void *handle,
+static int vega20_ih_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_clockgating_state state)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	vega20_ih_update_clockgating_state(adev,
 				state == AMD_CG_STATE_GATE);
@@ -662,7 +712,7 @@ static int vega20_ih_set_clockgating_state(void *handle,
 
 }
 
-static int vega20_ih_set_powergating_state(void *handle,
+static int vega20_ih_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					  enum amd_powergating_state state)
 {
 	return 0;
@@ -671,7 +721,6 @@ static int vega20_ih_set_powergating_state(void *handle,
 const struct amd_ip_funcs vega20_ih_ip_funcs = {
 	.name = "vega20_ih",
 	.early_init = vega20_ih_early_init,
-	.late_init = NULL,
 	.sw_init = vega20_ih_sw_init,
 	.sw_fini = vega20_ih_sw_fini,
 	.hw_init = vega20_ih_hw_init,
@@ -688,6 +737,7 @@ const struct amd_ip_funcs vega20_ih_ip_funcs = {
 static const struct amdgpu_ih_funcs vega20_ih_funcs = {
 	.get_wptr = vega20_ih_get_wptr,
 	.decode_iv = amdgpu_ih_decode_iv_helper,
+	.decode_iv_ts = amdgpu_ih_decode_iv_ts_helper,
 	.set_rptr = vega20_ih_set_rptr
 };
 
@@ -696,8 +746,7 @@ static void vega20_ih_set_interrupt_funcs(struct amdgpu_device *adev)
 	adev->irq.ih_funcs = &vega20_ih_funcs;
 }
 
-const struct amdgpu_ip_block_version vega20_ih_ip_block =
-{
+const struct amdgpu_ip_block_version vega20_ih_ip_block = {
 	.type = AMD_IP_BLOCK_TYPE_IH,
 	.major = 4,
 	.minor = 2,
diff --git a/drivers/gpu/drm/amd/amdgpu/vi.c b/drivers/gpu/drm/amd/amdgpu/vi.c
index fe9a7cc8d9eb..a611a7345125 100644
--- a/drivers/gpu/drm/amd/amdgpu/vi.c
+++ b/drivers/gpu/drm/amd/amdgpu/vi.c
@@ -67,7 +67,6 @@
 #include "sdma_v2_4.h"
 #include "sdma_v3_0.h"
 #include "dce_v10_0.h"
-#include "dce_v11_0.h"
 #include "iceland_ih.h"
 #include "tonga_ih.h"
 #include "cz_ih.h"
@@ -136,15 +135,15 @@ static const struct amdgpu_video_codec_info polaris_video_codecs_encode_array[]
 	{
 		.codec_type = AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC,
 		.max_width = 4096,
-		.max_height = 2304,
-		.max_pixels_per_frame = 4096 * 2304,
+		.max_height = 4096,
+		.max_pixels_per_frame = 4096 * 4096,
 		.max_level = 0,
 	},
 	{
 		.codec_type = AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC,
 		.max_width = 4096,
-		.max_height = 2304,
-		.max_pixels_per_frame = 4096 * 2304,
+		.max_height = 4096,
+		.max_pixels_per_frame = 4096 * 4096,
 		.max_level = 0,
 	},
 };
@@ -167,16 +166,16 @@ static const struct amdgpu_video_codec_info tonga_video_codecs_decode_array[] =
 {
 	{
 		.codec_type = AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG2,
-		.max_width = 4096,
-		.max_height = 4096,
-		.max_pixels_per_frame = 4096 * 4096,
+		.max_width = 1920,
+		.max_height = 1088,
+		.max_pixels_per_frame = 1920 * 1088,
 		.max_level = 3,
 	},
 	{
 		.codec_type = AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4,
-		.max_width = 4096,
-		.max_height = 4096,
-		.max_pixels_per_frame = 4096 * 4096,
+		.max_width = 1920,
+		.max_height = 1088,
+		.max_pixels_per_frame = 1920 * 1088,
 		.max_level = 5,
 	},
 	{
@@ -188,9 +187,9 @@ static const struct amdgpu_video_codec_info tonga_video_codecs_decode_array[] =
 	},
 	{
 		.codec_type = AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VC1,
-		.max_width = 4096,
-		.max_height = 4096,
-		.max_pixels_per_frame = 4096 * 4096,
+		.max_width = 1920,
+		.max_height = 1088,
+		.max_pixels_per_frame = 1920 * 1088,
 		.max_level = 4,
 	},
 };
@@ -206,16 +205,16 @@ static const struct amdgpu_video_codec_info cz_video_codecs_decode_array[] =
 {
 	{
 		.codec_type = AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG2,
-		.max_width = 4096,
-		.max_height = 4096,
-		.max_pixels_per_frame = 4096 * 4096,
+		.max_width = 1920,
+		.max_height = 1088,
+		.max_pixels_per_frame = 1920 * 1088,
 		.max_level = 3,
 	},
 	{
 		.codec_type = AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4,
-		.max_width = 4096,
-		.max_height = 4096,
-		.max_pixels_per_frame = 4096 * 4096,
+		.max_width = 1920,
+		.max_height = 1088,
+		.max_pixels_per_frame = 1920 * 1088,
 		.max_level = 5,
 	},
 	{
@@ -227,9 +226,9 @@ static const struct amdgpu_video_codec_info cz_video_codecs_decode_array[] =
 	},
 	{
 		.codec_type = AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VC1,
-		.max_width = 4096,
-		.max_height = 4096,
-		.max_pixels_per_frame = 4096 * 4096,
+		.max_width = 1920,
+		.max_height = 1088,
+		.max_pixels_per_frame = 1920 * 1088,
 		.max_level = 4,
 	},
 	{
@@ -542,8 +541,15 @@ static u32 vi_get_xclk(struct amdgpu_device *adev)
 	u32 reference_clock = adev->clock.spll.reference_freq;
 	u32 tmp;
 
-	if (adev->flags & AMD_IS_APU)
-		return reference_clock;
+	if (adev->flags & AMD_IS_APU) {
+		switch (adev->asic_type) {
+		case CHIP_STONEY:
+			/* vbios says 48Mhz, but the actual freq is 100Mhz */
+			return 10000;
+		default:
+			return reference_clock;
+		}
+	}
 
 	tmp = RREG32_SMC(ixCG_CLKPIN_CNTL_2);
 	if (REG_GET_FIELD(tmp, CG_CLKPIN_CNTL_2, MUX_TCLK_TO_XCLK))
@@ -580,11 +586,6 @@ void vi_srbm_select(struct amdgpu_device *adev,
 	WREG32(mmSRBM_GFX_CNTL, srbm_gfx_cntl);
 }
 
-static void vi_vga_set_state(struct amdgpu_device *adev, bool state)
-{
-	/* todo */
-}
-
 static bool vi_read_disabled_bios(struct amdgpu_device *adev)
 {
 	u32 bus_cntl;
@@ -762,12 +763,12 @@ static uint32_t vi_get_register_value(struct amdgpu_device *adev,
 
 		mutex_lock(&adev->grbm_idx_mutex);
 		if (se_num != 0xffffffff || sh_num != 0xffffffff)
-			amdgpu_gfx_select_se_sh(adev, se_num, sh_num, 0xffffffff);
+			amdgpu_gfx_select_se_sh(adev, se_num, sh_num, 0xffffffff, 0);
 
 		val = RREG32(reg_offset);
 
 		if (se_num != 0xffffffff || sh_num != 0xffffffff)
-			amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+			amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff, 0);
 		mutex_unlock(&adev->grbm_idx_mutex);
 		return val;
 	} else {
@@ -895,7 +896,7 @@ static int vi_asic_pci_config_reset(struct amdgpu_device *adev)
 	return r;
 }
 
-static bool vi_asic_supports_baco(struct amdgpu_device *adev)
+static int vi_asic_supports_baco(struct amdgpu_device *adev)
 {
 	switch (adev->asic_type) {
 	case CHIP_FIJI:
@@ -906,14 +907,14 @@ static bool vi_asic_supports_baco(struct amdgpu_device *adev)
 	case CHIP_TOPAZ:
 		return amdgpu_dpm_is_baco_supported(adev);
 	default:
-		return false;
+		return 0;
 	}
 }
 
 static enum amd_reset_method
 vi_asic_reset_method(struct amdgpu_device *adev)
 {
-	bool baco_reset;
+	int baco_reset;
 
 	if (amdgpu_reset_method == AMD_RESET_METHOD_LEGACY ||
 	    amdgpu_reset_method == AMD_RESET_METHOD_BACO)
@@ -933,7 +934,7 @@ vi_asic_reset_method(struct amdgpu_device *adev)
 		baco_reset = amdgpu_dpm_is_baco_supported(adev);
 		break;
 	default:
-		baco_reset = false;
+		baco_reset = 0;
 		break;
 	}
 
@@ -956,6 +957,10 @@ static int vi_asic_reset(struct amdgpu_device *adev)
 {
 	int r;
 
+	/* APUs don't have full asic reset */
+	if (adev->flags & AMD_IS_APU)
+		return 0;
+
 	if (vi_asic_reset_method(adev) == AMD_RESET_METHOD_BACO) {
 		dev_info(adev->dev, "BACO reset\n");
 		r = amdgpu_dpm_baco_reset(adev);
@@ -1097,24 +1102,6 @@ static int vi_set_vce_clocks(struct amdgpu_device *adev, u32 evclk, u32 ecclk)
 	return 0;
 }
 
-static void vi_pcie_gen3_enable(struct amdgpu_device *adev)
-{
-	if (pci_is_root_bus(adev->pdev->bus))
-		return;
-
-	if (amdgpu_pcie_gen2 == 0)
-		return;
-
-	if (adev->flags & AMD_IS_APU)
-		return;
-
-	if (!(adev->pm.pcie_gen_mask & (CAIL_PCIE_LINK_SPEED_SUPPORT_GEN2 |
-					CAIL_PCIE_LINK_SPEED_SUPPORT_GEN3)))
-		return;
-
-	/* todo */
-}
-
 static void vi_enable_aspm(struct amdgpu_device *adev)
 {
 	u32 data, orig;
@@ -1136,11 +1123,10 @@ static void vi_program_aspm(struct amdgpu_device *adev)
 	bool bL1SS = false;
 	bool bClkReqSupport = true;
 
-	if (!amdgpu_aspm)
+	if (!amdgpu_device_should_use_aspm(adev))
 		return;
 
-	if (adev->flags & AMD_IS_APU ||
-	    adev->asic_type < CHIP_POLARIS10)
+	if (adev->asic_type < CHIP_POLARIS10)
 		return;
 
 	orig = data = RREG32_PCIE(ixPCIE_LC_CNTL);
@@ -1449,7 +1435,6 @@ static const struct amdgpu_asic_funcs vi_asic_funcs =
 	.read_register = &vi_read_register,
 	.reset = &vi_asic_reset,
 	.reset_method = &vi_asic_reset_method,
-	.set_vga_state = &vi_vga_set_state,
 	.get_xclk = &vi_get_xclk,
 	.set_uvd_clocks = &vi_set_uvd_clocks,
 	.set_vce_clocks = &vi_set_vce_clocks,
@@ -1469,9 +1454,9 @@ static const struct amdgpu_asic_funcs vi_asic_funcs =
 #define CZ_REV_BRISTOL(rev)	 \
 	((rev >= 0xC8 && rev <= 0xCE) || (rev >= 0xE1 && rev <= 0xE6))
 
-static int vi_common_early_init(void *handle)
+static int vi_common_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (adev->flags & AMD_IS_APU) {
 		adev->smc_rreg = &cz_smc_rreg;
@@ -1693,9 +1678,9 @@ static int vi_common_early_init(void *handle)
 	return 0;
 }
 
-static int vi_common_late_init(void *handle)
+static int vi_common_late_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (amdgpu_sriov_vf(adev))
 		xgpu_vi_mailbox_get_irq(adev);
@@ -1703,9 +1688,9 @@ static int vi_common_late_init(void *handle)
 	return 0;
 }
 
-static int vi_common_sw_init(void *handle)
+static int vi_common_sw_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (amdgpu_sriov_vf(adev))
 		xgpu_vi_mailbox_add_irq_id(adev);
@@ -1713,19 +1698,12 @@ static int vi_common_sw_init(void *handle)
 	return 0;
 }
 
-static int vi_common_sw_fini(void *handle)
-{
-	return 0;
-}
-
-static int vi_common_hw_init(void *handle)
+static int vi_common_hw_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	/* move the golden regs per IP block */
 	vi_init_golden_registers(adev);
-	/* enable pcie gen2/3 link */
-	vi_pcie_gen3_enable(adev);
 	/* enable aspm */
 	vi_program_aspm(adev);
 	/* enable the doorbell aperture */
@@ -1734,9 +1712,9 @@ static int vi_common_hw_init(void *handle)
 	return 0;
 }
 
-static int vi_common_hw_fini(void *handle)
+static int vi_common_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	/* enable the doorbell aperture */
 	vi_enable_doorbell_aperture(adev, false);
@@ -1747,35 +1725,21 @@ static int vi_common_hw_fini(void *handle)
 	return 0;
 }
 
-static int vi_common_suspend(void *handle)
+static int vi_common_suspend(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	return vi_common_hw_fini(adev);
+	return vi_common_hw_fini(ip_block);
 }
 
-static int vi_common_resume(void *handle)
+static int vi_common_resume(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	return vi_common_hw_init(adev);
+	return vi_common_hw_init(ip_block);
 }
 
-static bool vi_common_is_idle(void *handle)
+static bool vi_common_is_idle(struct amdgpu_ip_block *ip_block)
 {
 	return true;
 }
 
-static int vi_common_wait_for_idle(void *handle)
-{
-	return 0;
-}
-
-static int vi_common_soft_reset(void *handle)
-{
-	return 0;
-}
-
 static void vi_update_bif_medium_grain_light_sleep(struct amdgpu_device *adev,
 						   bool enable)
 {
@@ -1980,10 +1944,10 @@ static int vi_common_set_clockgating_state_by_smu(void *handle,
 	return 0;
 }
 
-static int vi_common_set_clockgating_state(void *handle,
+static int vi_common_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 					   enum amd_clockgating_state state)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	if (amdgpu_sriov_vf(adev))
 		return 0;
@@ -2023,15 +1987,15 @@ static int vi_common_set_clockgating_state(void *handle,
 	return 0;
 }
 
-static int vi_common_set_powergating_state(void *handle,
+static int vi_common_set_powergating_state(struct amdgpu_ip_block *ip_block,
 					    enum amd_powergating_state state)
 {
 	return 0;
 }
 
-static void vi_common_get_clockgating_state(void *handle, u32 *flags)
+static void vi_common_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int data;
 
 	if (amdgpu_sriov_vf(adev))
@@ -2063,14 +2027,11 @@ static const struct amd_ip_funcs vi_common_ip_funcs = {
 	.early_init = vi_common_early_init,
 	.late_init = vi_common_late_init,
 	.sw_init = vi_common_sw_init,
-	.sw_fini = vi_common_sw_fini,
 	.hw_init = vi_common_hw_init,
 	.hw_fini = vi_common_hw_fini,
 	.suspend = vi_common_suspend,
 	.resume = vi_common_resume,
 	.is_idle = vi_common_is_idle,
-	.wait_for_idle = vi_common_wait_for_idle,
-	.soft_reset = vi_common_soft_reset,
 	.set_clockgating_state = vi_common_set_clockgating_state,
 	.set_powergating_state = vi_common_set_powergating_state,
 	.get_clockgating_state = vi_common_get_clockgating_state,
@@ -2092,6 +2053,8 @@ void vi_set_virt_ops(struct amdgpu_device *adev)
 
 int vi_set_ip_blocks(struct amdgpu_device *adev)
 {
+	amdgpu_device_set_sriov_virtual_display(adev);
+
 	switch (adev->asic_type) {
 	case CHIP_TOPAZ:
 		/* topaz has no DCE, UVD, VCE */
@@ -2111,7 +2074,7 @@ int vi_set_ip_blocks(struct amdgpu_device *adev)
 		amdgpu_device_ip_block_add(adev, &gfx_v8_0_ip_block);
 		amdgpu_device_ip_block_add(adev, &sdma_v3_0_ip_block);
 		amdgpu_device_ip_block_add(adev, &pp_smu_ip_block);
-		if (adev->enable_virtual_display || amdgpu_sriov_vf(adev))
+		if (adev->enable_virtual_display)
 			amdgpu_device_ip_block_add(adev, &amdgpu_vkms_ip_block);
 #if defined(CONFIG_DRM_AMD_DC)
 		else if (amdgpu_device_has_dc_support(adev))
@@ -2131,7 +2094,7 @@ int vi_set_ip_blocks(struct amdgpu_device *adev)
 		amdgpu_device_ip_block_add(adev, &gfx_v8_0_ip_block);
 		amdgpu_device_ip_block_add(adev, &sdma_v3_0_ip_block);
 		amdgpu_device_ip_block_add(adev, &pp_smu_ip_block);
-		if (adev->enable_virtual_display || amdgpu_sriov_vf(adev))
+		if (adev->enable_virtual_display)
 			amdgpu_device_ip_block_add(adev, &amdgpu_vkms_ip_block);
 #if defined(CONFIG_DRM_AMD_DC)
 		else if (amdgpu_device_has_dc_support(adev))
@@ -2160,8 +2123,6 @@ int vi_set_ip_blocks(struct amdgpu_device *adev)
 		else if (amdgpu_device_has_dc_support(adev))
 			amdgpu_device_ip_block_add(adev, &dm_ip_block);
 #endif
-		else
-			amdgpu_device_ip_block_add(adev, &dce_v11_2_ip_block);
 		amdgpu_device_ip_block_add(adev, &uvd_v6_3_ip_block);
 		amdgpu_device_ip_block_add(adev, &vce_v3_4_ip_block);
 		break;
@@ -2178,8 +2139,6 @@ int vi_set_ip_blocks(struct amdgpu_device *adev)
 		else if (amdgpu_device_has_dc_support(adev))
 			amdgpu_device_ip_block_add(adev, &dm_ip_block);
 #endif
-		else
-			amdgpu_device_ip_block_add(adev, &dce_v11_0_ip_block);
 		amdgpu_device_ip_block_add(adev, &uvd_v6_0_ip_block);
 		amdgpu_device_ip_block_add(adev, &vce_v3_1_ip_block);
 #if defined(CONFIG_DRM_AMD_ACP)
@@ -2199,8 +2158,6 @@ int vi_set_ip_blocks(struct amdgpu_device *adev)
 		else if (amdgpu_device_has_dc_support(adev))
 			amdgpu_device_ip_block_add(adev, &dm_ip_block);
 #endif
-		else
-			amdgpu_device_ip_block_add(adev, &dce_v11_0_ip_block);
 		amdgpu_device_ip_block_add(adev, &uvd_v6_2_ip_block);
 		amdgpu_device_ip_block_add(adev, &vce_v3_4_ip_block);
 #if defined(CONFIG_DRM_AMD_ACP)
diff --git a/drivers/gpu/drm/amd/amdgpu/vid.h b/drivers/gpu/drm/amd/amdgpu/vid.h
index 80ce42aacc0c..b61f6b838ec2 100644
--- a/drivers/gpu/drm/amd/amdgpu/vid.h
+++ b/drivers/gpu/drm/amd/amdgpu/vid.h
@@ -246,6 +246,7 @@
 		 * 1 - Stream
 		 * 2 - Bypass
 		 */
+#define		EOP_EXEC				(1 << 28) /* For Trailing Fence */
 #define		DATA_SEL(x)                             ((x) << 29)
 		/* 0 - discard
 		 * 1 - send low 32bit data
diff --git a/drivers/gpu/drm/amd/amdgpu/vpe_6_1_fw_if.h b/drivers/gpu/drm/amd/amdgpu/vpe_6_1_fw_if.h
new file mode 100644
index 000000000000..47534dbbd137
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/vpe_6_1_fw_if.h
@@ -0,0 +1,218 @@
+/* Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __VPE_6_1_FW_IF_H_
+#define __VPE_6_1_FW_IF_H_
+
+/****************
+ * VPE OP Codes
+ ****************/
+enum VPE_CMD_OPCODE {
+    VPE_CMD_OPCODE_NOP          = 0x0,
+    VPE_CMD_OPCODE_VPE_DESC     = 0x1,
+    VPE_CMD_OPCODE_PLANE_CFG    = 0x2,
+    VPE_CMD_OPCODE_VPEP_CFG     = 0x3,
+    VPE_CMD_OPCODE_INDIRECT     = 0x4,
+    VPE_CMD_OPCODE_FENCE        = 0x5,
+    VPE_CMD_OPCODE_TRAP         = 0x6,
+    VPE_CMD_OPCODE_REG_WRITE    = 0x7,
+    VPE_CMD_OPCODE_POLL_REGMEM  = 0x8,
+    VPE_CMD_OPCODE_COND_EXE     = 0x9,
+    VPE_CMD_OPCODE_ATOMIC       = 0xA,
+    VPE_CMD_OPCODE_PRED_EXE     = 0xB,
+    VPE_CMD_OPCODE_COLLAB_SYNC  = 0xC,
+    VPE_CMD_OPCODE_TIMESTAMP    = 0xD
+};
+
+/** Generic Command Header
+ * Generic Commands include:
+ *  Noop, Fence, Trap,
+ *  RegisterWrite, PollRegisterWriteMemory,
+ *  SetLocalTimestamp, GetLocalTimestamp
+ *  GetGlobalGPUTimestamp */
+#define VPE_HEADER_SUB_OPCODE__SHIFT    8
+#define VPE_HEADER_SUB_OPCODE_MASK      0x0000FF00
+#define VPE_HEADER_OPCODE__SHIFT        0
+#define VPE_HEADER_OPCODE_MASK          0x000000FF
+
+#define VPE_CMD_HEADER(op, subop) \
+    (((subop << VPE_HEADER_SUB_OPCODE__SHIFT) & VPE_HEADER_SUB_OPCODE_MASK) | \
+     ((op << VPE_HEADER_OPCODE__SHIFT) & VPE_HEADER_OPCODE_MASK))
+
+
+ /***************************
+  * VPE NOP
+  ***************************/
+#define VPE_CMD_NOP_HEADER_COUNT__SHIFT    16
+#define VPE_CMD_NOP_HEADER_COUNT_MASK      0x00003FFF
+
+#define VPE_CMD_NOP_HEADER_COUNT(count) \
+     (((count) & VPE_CMD_NOP_HEADER_COUNT_MASK) << VPE_CMD_NOP_HEADER_COUNT__SHIFT)
+
+ /***************************
+  * VPE Descriptor
+  ***************************/
+#define VPE_DESC_CD__SHIFT          16
+#define VPE_DESC_CD_MASK            0x000F0000
+
+#define VPE_DESC_CMD_HEADER(cd) \
+    (VPE_CMD_HEADER(VPE_CMD_OPCODE_VPE_DESC, 0) | \
+     (((cd) << VPE_DESC_CD__SHIFT) & VPE_DESC_CD_MASK))
+
+ /***************************
+  * VPE Plane Config
+  ***************************/
+enum VPE_PLANE_CFG_SUBOP {
+    VPE_PLANE_CFG_SUBOP_1_TO_1 = 0x0,
+    VPE_PLANE_CFG_SUBOP_2_TO_1 = 0x1,
+    VPE_PLANE_CFG_SUBOP_2_TO_2 = 0x2
+};
+
+#define VPE_PLANE_CFG_ONE_PLANE     0
+#define VPE_PLANE_CFG_TWO_PLANES    1
+
+#define VPE_PLANE_CFG_NPS0__SHIFT   16
+#define VPE_PLANE_CFG_NPS0_MASK     0x00030000
+
+#define VPE_PLANE_CFG_NPD0__SHIFT   18
+#define VPE_PLANE_CFG_NPD0_MASK     0x000C0000
+
+#define VPE_PLANE_CFG_NPS1__SHIFT   20
+#define VPE_PLANE_CFG_NPS1_MASK     0x00300000
+
+#define VPE_PLANE_CFG_NPD1__SHIFT   22
+#define VPE_PLANE_CFG_NPD1_MASK     0x00C00000
+
+#define VPE_PLANE_CFG_TMZ__SHIFT    16
+#define VPE_PLANE_CFG_TMZ_MASK      0x00010000
+
+#define VPE_PLANE_CFG_SWIZZLE_MODE__SHIFT   3
+#define VPE_PLANE_CFG_SWIZZLE_MODE_MASK     0x000000F8
+
+#define VPE_PLANE_CFG_ROTATION__SHIFT       0
+#define VPE_PLANE_CFG_ROTATION_MASK         0x00000003
+
+#define VPE_PLANE_ADDR_LO__SHIFT        0
+#define VPE_PLANE_ADDR_LO_MASK          0xFFFFFF00
+
+#define VPE_PLANE_CFG_PITCH__SHIFT      0
+#define VPE_PLANE_CFG_PITCH_MASK        0x00003FFF
+
+#define VPE_PLANE_CFG_VIEWPORT_Y__SHIFT 16
+#define VPE_PLANE_CFG_VIEWPORT_Y_MASK   0x3FFF0000
+#define VPE_PLANE_CFG_VIEWPORT_X__SHIFT 0
+#define VPE_PLANE_CFG_VIEWPORT_X_MASK   0x00003FFF
+
+
+#define VPE_PLANE_CFG_VIEWPORT_HEIGHT__SHIFT 16
+#define VPE_PLANE_CFG_VIEWPORT_HEIGHT_MASK   0x1FFF0000
+#define VPE_PLANE_CFG_VIEWPORT_ELEMENT_SIZE__SHIFT  13
+#define VPE_PLANE_CFG_VIEWPORT_ELEMENT_SIZE_MASK    0x0000E000
+#define VPE_PLANE_CFG_VIEWPORT_WIDTH__SHIFT 0
+#define VPE_PLANE_CFG_VIEWPORT_WIDTH_MASK   0x00001FFF
+
+enum VPE_PLANE_CFG_ELEMENT_SIZE {
+    VPE_PLANE_CFG_ELEMENT_SIZE_8BPE     = 0,
+    VPE_PLANE_CFG_ELEMENT_SIZE_16BPE    = 1,
+    VPE_PLANE_CFG_ELEMENT_SIZE_32BPE    = 2,
+    VPE_PLANE_CFG_ELEMENT_SIZE_64BPE    = 3
+};
+
+#define VPE_PLANE_CFG_CMD_HEADER(subop, nps0, npd0, nps1, npd1) \
+    (VPE_CMD_HEADER(VPE_CMD_OPCODE_PLANE_CFG, subop) | \
+     (((nps0) << VPE_PLANE_CFG_NPS0__SHIFT) & VPE_PLANE_CFG_NPS0_MASK) | \
+     (((npd0) << VPE_PLANE_CFG_NPD0__SHIFT) & VPE_PLANE_CFG_NPD0_MASK) | \
+     (((nps1) << VPE_PLANE_CFG_NPS1__SHIFT) & VPE_PLANE_CFG_NPS1_MASK) | \
+     (((npd0) << VPE_PLANE_CFG_NPD1__SHIFT) & VPE_PLANE_CFG_NPD1_MASK))
+
+
+/************************
+ * VPEP Config
+ ************************/
+enum VPE_VPEP_CFG_SUBOP {
+    VPE_VPEP_CFG_SUBOP_DIR_CFG = 0x0,
+    VPE_VPEP_CFG_SUBOP_IND_CFG = 0x1
+};
+
+
+// Direct Config Command Header
+#define VPE_DIR_CFG_HEADER_ARRAY_SIZE__SHIFT    16
+#define VPE_DIR_CFG_HEADER_ARRAY_SIZE_MASK      0xFFFF0000
+
+#define VPE_DIR_CFG_CMD_HEADER(subop, arr_sz) \
+    (VPE_CMD_HEADER(VPE_CMD_OPCODE_VPEP_CFG, subop) | \
+     (((arr_sz) << VPE_DIR_CFG_HEADER_ARRAY_SIZE__SHIFT) & VPE_DIR_CFG_HEADER_ARRAY_SIZE_MASK))
+
+
+#define VPE_DIR_CFG_PKT_REGISTER_OFFSET__SHIFT  2
+#define VPE_DIR_CFG_PKT_REGISTER_OFFSET_MASK    0x000FFFFC
+
+#define VPE_DIR_CFG_PKT_DATA_SIZE__SHIFT        20
+#define VPE_DIR_CFG_PKT_DATA_SIZE_MASK          0xFFF00000
+
+
+// InDirect Config Command Header
+#define VPE_IND_CFG_HEADER_NUM_DST__SHIFT   28
+#define VPE_IND_CFG_HEADER_NUM_DST_MASK     0xF0000000
+
+#define VPE_IND_CFG_CMD_HEADER(subop, num_dst) \
+    (VPE_CMD_HEADER(VPE_CMD_OPCODE_VPEP_CFG, subop) | \
+     (((num_dst) << VPE_IND_CFG_HEADER_NUM_DST__SHIFT) & VPE_IND_CFG_HEADER_NUM_DST_MASK))
+
+// Indirect Buffer Command Header
+#define VPE_CMD_INDIRECT_HEADER_VMID__SHIFT   16
+#define VPE_CMD_INDIRECT_HEADER_VMID_MASK     0x0000000F
+#define VPE_CMD_INDIRECT_HEADER_VMID(vmid) \
+     (((vmid) & VPE_CMD_INDIRECT_HEADER_VMID_MASK) << VPE_CMD_INDIRECT_HEADER_VMID__SHIFT)
+
+
+/**************************
+ * Poll Reg/Mem Sub-OpCode
+ **************************/
+enum VPE_POLL_REGMEM_SUBOP {
+    VPE_POLL_REGMEM_SUBOP_REGMEM = 0x0,
+    VPE_POLL_REGMEM_SUBOP_REGMEM_WRITE = 0x1
+};
+
+#define VPE_CMD_POLL_REGMEM_HEADER_FUNC__SHIFT   28
+#define VPE_CMD_POLL_REGMEM_HEADER_FUNC_MASK     0x00000007
+#define VPE_CMD_POLL_REGMEM_HEADER_FUNC(func) \
+     (((func) & VPE_CMD_POLL_REGMEM_HEADER_FUNC_MASK) << VPE_CMD_POLL_REGMEM_HEADER_FUNC__SHIFT)
+
+#define VPE_CMD_POLL_REGMEM_HEADER_MEM__SHIFT   31
+#define VPE_CMD_POLL_REGMEM_HEADER_MEM_MASK     0x00000001
+#define VPE_CMD_POLL_REGMEM_HEADER_MEM(mem) \
+     (((mem) & VPE_CMD_POLL_REGMEM_HEADER_MEM_MASK) << VPE_CMD_POLL_REGMEM_HEADER_MEM__SHIFT)
+
+#define VPE_CMD_POLL_REGMEM_DW5_INTERVAL__SHIFT   0
+#define VPE_CMD_POLL_REGMEM_DW5_INTERVAL_MASK     0x0000FFFF
+#define VPE_CMD_POLL_REGMEM_DW5_INTERVAL(interval) \
+     (((interval) & VPE_CMD_POLL_REGMEM_DW5_INTERVAL_MASK) << VPE_CMD_POLL_REGMEM_DW5_INTERVAL__SHIFT)
+
+#define VPE_CMD_POLL_REGMEM_DW5_RETRY_COUNT__SHIFT   16
+#define VPE_CMD_POLL_REGMEM_DW5_RETRY_COUNT_MASK     0x00000FFF
+#define VPE_CMD_POLL_REGMEM_DW5_RETRY_COUNT(count) \
+     (((count) & VPE_CMD_POLL_REGMEM_DW5_RETRY_COUNT_MASK) << VPE_CMD_POLL_REGMEM_DW5_RETRY_COUNT__SHIFT)
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/vpe_v6_1.c b/drivers/gpu/drm/amd/amdgpu/vpe_v6_1.c
new file mode 100644
index 000000000000..45876883bbf3
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/vpe_v6_1.c
@@ -0,0 +1,398 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#include <linux/firmware.h>
+#include <drm/drm_drv.h>
+
+#include "amdgpu.h"
+#include "amdgpu_ucode.h"
+#include "amdgpu_vpe.h"
+#include "vpe_v6_1.h"
+#include "soc15_common.h"
+#include "ivsrcid/vpe/irqsrcs_vpe_6_1.h"
+#include "vpe/vpe_6_1_0_offset.h"
+#include "vpe/vpe_6_1_0_sh_mask.h"
+
+MODULE_FIRMWARE("amdgpu/vpe_6_1_0.bin");
+MODULE_FIRMWARE("amdgpu/vpe_6_1_1.bin");
+MODULE_FIRMWARE("amdgpu/vpe_6_1_3.bin");
+
+#define VPE_THREAD1_UCODE_OFFSET	0x8000
+
+#define regVPEC_COLLABORATE_CNTL                                                0x0013
+#define regVPEC_COLLABORATE_CNTL_BASE_IDX                                       0
+#define VPEC_COLLABORATE_CNTL__COLLABORATE_MODE_EN__SHIFT                       0x0
+#define VPEC_COLLABORATE_CNTL__COLLABORATE_MODE_EN_MASK                         0x00000001L
+
+#define regVPEC_COLLABORATE_CFG                                                 0x0014
+#define regVPEC_COLLABORATE_CFG_BASE_IDX                                        0
+#define VPEC_COLLABORATE_CFG__MASTER_ID__SHIFT                                  0x0
+#define VPEC_COLLABORATE_CFG__MASTER_EN__SHIFT                                  0x3
+#define VPEC_COLLABORATE_CFG__SLAVE0_ID__SHIFT                                  0x4
+#define VPEC_COLLABORATE_CFG__SLAVE0_EN__SHIFT                                  0x7
+#define VPEC_COLLABORATE_CFG__MASTER_ID_MASK                                    0x00000007L
+#define VPEC_COLLABORATE_CFG__MASTER_EN_MASK                                    0x00000008L
+#define VPEC_COLLABORATE_CFG__SLAVE0_ID_MASK                                    0x00000070L
+#define VPEC_COLLABORATE_CFG__SLAVE0_EN_MASK                                    0x00000080L
+
+#define regVPEC_CNTL_6_1_1                                                      0x0016
+#define regVPEC_CNTL_6_1_1_BASE_IDX                                             0
+#define regVPEC_QUEUE_RESET_REQ_6_1_1                                           0x002c
+#define regVPEC_QUEUE_RESET_REQ_6_1_1_BASE_IDX                                  0
+#define regVPEC_PUB_DUMMY2_6_1_1                                                0x004c
+#define regVPEC_PUB_DUMMY2_6_1_1_BASE_IDX                                       0
+
+static uint32_t vpe_v6_1_get_reg_offset(struct amdgpu_vpe *vpe, uint32_t inst, uint32_t offset)
+{
+	uint32_t base;
+
+	base = vpe->ring.adev->reg_offset[VPE_HWIP][inst][0];
+
+	return base + offset;
+}
+
+static void vpe_v6_1_halt(struct amdgpu_vpe *vpe, bool halt)
+{
+	struct amdgpu_device *adev = vpe->ring.adev;
+	uint32_t i, f32_cntl;
+
+	for (i = 0; i < vpe->num_instances; i++) {
+		f32_cntl = RREG32(vpe_get_reg_offset(vpe, i, regVPEC_F32_CNTL));
+		f32_cntl = REG_SET_FIELD(f32_cntl, VPEC_F32_CNTL, HALT, halt ? 1 : 0);
+		f32_cntl = REG_SET_FIELD(f32_cntl, VPEC_F32_CNTL, TH1_RESET, halt ? 1 : 0);
+		WREG32(vpe_get_reg_offset(vpe, i, regVPEC_F32_CNTL), f32_cntl);
+	}
+}
+
+static int vpe_v6_1_irq_init(struct amdgpu_vpe *vpe)
+{
+	struct amdgpu_device *adev = container_of(vpe, struct amdgpu_device, vpe);
+	int ret;
+
+	ret = amdgpu_irq_add_id(adev, SOC21_IH_CLIENTID_VPE,
+				VPE_6_1_SRCID__VPE_TRAP,
+				&adev->vpe.trap_irq);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+static void vpe_v6_1_set_collaborate_mode(struct amdgpu_vpe *vpe, bool enable)
+{
+	struct amdgpu_device *adev = vpe->ring.adev;
+	uint32_t vpe_colla_cntl, vpe_colla_cfg, i;
+
+	if (!vpe->collaborate_mode)
+		return;
+
+	for (i = 0; i < vpe->num_instances; i++) {
+		vpe_colla_cntl = RREG32(vpe_get_reg_offset(vpe, i, regVPEC_COLLABORATE_CNTL));
+		vpe_colla_cntl = REG_SET_FIELD(vpe_colla_cntl, VPEC_COLLABORATE_CNTL,
+					       COLLABORATE_MODE_EN, enable ? 1 : 0);
+		WREG32(vpe_get_reg_offset(vpe, i, regVPEC_COLLABORATE_CNTL), vpe_colla_cntl);
+
+		vpe_colla_cfg = RREG32(vpe_get_reg_offset(vpe, i, regVPEC_COLLABORATE_CFG));
+		vpe_colla_cfg = REG_SET_FIELD(vpe_colla_cfg, VPEC_COLLABORATE_CFG, MASTER_ID, 0);
+		vpe_colla_cfg = REG_SET_FIELD(vpe_colla_cfg, VPEC_COLLABORATE_CFG, MASTER_EN, enable ? 1 : 0);
+		vpe_colla_cfg = REG_SET_FIELD(vpe_colla_cfg, VPEC_COLLABORATE_CFG, SLAVE0_ID, 1);
+		vpe_colla_cfg = REG_SET_FIELD(vpe_colla_cfg, VPEC_COLLABORATE_CFG, SLAVE0_EN, enable ? 1 : 0);
+		WREG32(vpe_get_reg_offset(vpe, i, regVPEC_COLLABORATE_CFG), vpe_colla_cfg);
+	}
+}
+
+static int vpe_v6_1_load_microcode(struct amdgpu_vpe *vpe)
+{
+	struct amdgpu_device *adev = vpe->ring.adev;
+	const struct vpe_firmware_header_v1_0 *vpe_hdr;
+	const __le32 *data;
+	uint32_t ucode_offset[2], ucode_size[2];
+	uint32_t i, j, size_dw;
+	uint32_t ret;
+
+	/* disable UMSCH_INT_ENABLE */
+	for (j = 0; j < vpe->num_instances; j++) {
+
+		if (amdgpu_ip_version(adev, VPE_HWIP, 0) == IP_VERSION(6, 1, 1))
+			ret = RREG32(vpe_get_reg_offset(vpe, j, regVPEC_CNTL_6_1_1));
+		else
+			ret = RREG32(vpe_get_reg_offset(vpe, j, regVPEC_CNTL));
+
+		ret = REG_SET_FIELD(ret, VPEC_CNTL, UMSCH_INT_ENABLE, 0);
+
+		if (amdgpu_ip_version(adev, VPE_HWIP, 0) == IP_VERSION(6, 1, 1))
+			WREG32(vpe_get_reg_offset(vpe, j, regVPEC_CNTL_6_1_1), ret);
+		else
+			WREG32(vpe_get_reg_offset(vpe, j, regVPEC_CNTL), ret);
+	}
+
+	/* setup collaborate mode */
+	vpe_v6_1_set_collaborate_mode(vpe, true);
+	/* setup DPM */
+	if (amdgpu_vpe_configure_dpm(vpe))
+		dev_warn(adev->dev, "VPE failed to enable DPM\n");
+
+	/*
+	 * For VPE 6.1.1, still only need to add master's offset, and psp will apply it to slave as well.
+	 * Here use instance 0 as master.
+	 */
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
+		uint32_t f32_offset, f32_cntl;
+
+		f32_offset = vpe_get_reg_offset(vpe, 0, regVPEC_F32_CNTL);
+		f32_cntl = RREG32(f32_offset);
+		f32_cntl = REG_SET_FIELD(f32_cntl, VPEC_F32_CNTL, HALT, 0);
+		f32_cntl = REG_SET_FIELD(f32_cntl, VPEC_F32_CNTL, TH1_RESET, 0);
+
+		adev->vpe.cmdbuf_cpu_addr[0] = f32_offset;
+		adev->vpe.cmdbuf_cpu_addr[1] = f32_cntl;
+
+		return amdgpu_vpe_psp_update_sram(adev);
+	}
+
+	vpe_hdr = (const struct vpe_firmware_header_v1_0 *)adev->vpe.fw->data;
+
+	/* Thread 0(command thread) ucode offset/size */
+	ucode_offset[0] = le32_to_cpu(vpe_hdr->header.ucode_array_offset_bytes);
+	ucode_size[0] = le32_to_cpu(vpe_hdr->ctx_ucode_size_bytes);
+	/* Thread 1(control thread) ucode offset/size */
+	ucode_offset[1] = le32_to_cpu(vpe_hdr->ctl_ucode_offset);
+	ucode_size[1] = le32_to_cpu(vpe_hdr->ctl_ucode_size_bytes);
+
+	vpe_v6_1_halt(vpe, true);
+
+	for (j = 0; j < vpe->num_instances; j++) {
+		for (i = 0; i < 2; i++) {
+			if (i > 0)
+				WREG32(vpe_get_reg_offset(vpe, j, regVPEC_UCODE_ADDR), VPE_THREAD1_UCODE_OFFSET);
+			else
+				WREG32(vpe_get_reg_offset(vpe, j, regVPEC_UCODE_ADDR), 0);
+
+			data = (const __le32 *)(adev->vpe.fw->data + ucode_offset[i]);
+			size_dw = ucode_size[i] / sizeof(__le32);
+
+			while (size_dw--) {
+				if (amdgpu_emu_mode && size_dw % 500 == 0)
+					msleep(1);
+				WREG32(vpe_get_reg_offset(vpe, j, regVPEC_UCODE_DATA), le32_to_cpup(data++));
+			}
+		}
+	}
+
+	vpe_v6_1_halt(vpe, false);
+
+	return 0;
+}
+
+static int vpe_v6_1_ring_start(struct amdgpu_vpe *vpe)
+{
+	struct amdgpu_ring *ring = &vpe->ring;
+	struct amdgpu_device *adev = ring->adev;
+	uint32_t doorbell, doorbell_offset;
+	uint32_t rb_bufsz, rb_cntl;
+	uint32_t ib_cntl, i;
+	int ret;
+
+	for (i = 0; i < vpe->num_instances; i++) {
+		/* Set ring buffer size in dwords */
+		rb_bufsz = order_base_2(ring->ring_size / 4);
+		rb_cntl = RREG32(vpe_get_reg_offset(vpe, i, regVPEC_QUEUE0_RB_CNTL));
+		rb_cntl = REG_SET_FIELD(rb_cntl, VPEC_QUEUE0_RB_CNTL, RB_SIZE, rb_bufsz);
+		rb_cntl = REG_SET_FIELD(rb_cntl, VPEC_QUEUE0_RB_CNTL, RB_PRIV, 1);
+		rb_cntl = REG_SET_FIELD(rb_cntl, VPEC_QUEUE0_RB_CNTL, RB_VMID, 0);
+		WREG32(vpe_get_reg_offset(vpe, i, regVPEC_QUEUE0_RB_CNTL), rb_cntl);
+
+		/* Initialize the ring buffer's read and write pointers */
+		WREG32(vpe_get_reg_offset(vpe, i, regVPEC_QUEUE0_RB_RPTR), 0);
+		WREG32(vpe_get_reg_offset(vpe, i, regVPEC_QUEUE0_RB_RPTR_HI), 0);
+		WREG32(vpe_get_reg_offset(vpe, i, regVPEC_QUEUE0_RB_WPTR), 0);
+		WREG32(vpe_get_reg_offset(vpe, i, regVPEC_QUEUE0_RB_WPTR_HI), 0);
+
+		/* set the wb address whether it's enabled or not */
+		WREG32(vpe_get_reg_offset(vpe, i, regVPEC_QUEUE0_RB_RPTR_ADDR_LO),
+			lower_32_bits(ring->rptr_gpu_addr) & 0xFFFFFFFC);
+		WREG32(vpe_get_reg_offset(vpe, i, regVPEC_QUEUE0_RB_RPTR_ADDR_HI),
+			upper_32_bits(ring->rptr_gpu_addr) & 0xFFFFFFFF);
+
+		rb_cntl = REG_SET_FIELD(rb_cntl, VPEC_QUEUE0_RB_CNTL, RPTR_WRITEBACK_ENABLE, 1);
+
+		WREG32(vpe_get_reg_offset(vpe, i, regVPEC_QUEUE0_RB_BASE), ring->gpu_addr >> 8);
+		WREG32(vpe_get_reg_offset(vpe, i, regVPEC_QUEUE0_RB_BASE_HI), ring->gpu_addr >> 40);
+
+		ring->wptr = 0;
+
+		/* before programing wptr to a less value, need set minor_ptr_update first */
+		WREG32(vpe_get_reg_offset(vpe, i, regVPEC_QUEUE0_MINOR_PTR_UPDATE), 1);
+		WREG32(vpe_get_reg_offset(vpe, i, regVPEC_QUEUE0_RB_WPTR), lower_32_bits(ring->wptr) << 2);
+		WREG32(vpe_get_reg_offset(vpe, i, regVPEC_QUEUE0_RB_WPTR_HI), upper_32_bits(ring->wptr) << 2);
+		/* set minor_ptr_update to 0 after wptr programed */
+		WREG32(vpe_get_reg_offset(vpe, i, regVPEC_QUEUE0_MINOR_PTR_UPDATE), 0);
+
+		doorbell_offset = RREG32(vpe_get_reg_offset(vpe, i, regVPEC_QUEUE0_DOORBELL_OFFSET));
+		doorbell_offset = REG_SET_FIELD(doorbell_offset, VPEC_QUEUE0_DOORBELL_OFFSET, OFFSET, ring->doorbell_index + i*4);
+		WREG32(vpe_get_reg_offset(vpe, i, regVPEC_QUEUE0_DOORBELL_OFFSET), doorbell_offset);
+
+		doorbell = RREG32(vpe_get_reg_offset(vpe, i, regVPEC_QUEUE0_DOORBELL));
+		doorbell = REG_SET_FIELD(doorbell, VPEC_QUEUE0_DOORBELL, ENABLE, ring->use_doorbell ? 1 : 0);
+		WREG32(vpe_get_reg_offset(vpe, i, regVPEC_QUEUE0_DOORBELL), doorbell);
+
+		adev->nbio.funcs->vpe_doorbell_range(adev, i, ring->use_doorbell, ring->doorbell_index + i*4, 4);
+
+		rb_cntl = REG_SET_FIELD(rb_cntl, VPEC_QUEUE0_RB_CNTL, RPTR_WRITEBACK_ENABLE, 1);
+		rb_cntl = REG_SET_FIELD(rb_cntl, VPEC_QUEUE0_RB_CNTL, RB_ENABLE, 1);
+		WREG32(vpe_get_reg_offset(vpe, i, regVPEC_QUEUE0_RB_CNTL), rb_cntl);
+
+		ib_cntl = RREG32(vpe_get_reg_offset(vpe, i, regVPEC_QUEUE0_IB_CNTL));
+		ib_cntl = REG_SET_FIELD(ib_cntl, VPEC_QUEUE0_IB_CNTL, IB_ENABLE, 1);
+		WREG32(vpe_get_reg_offset(vpe, i, regVPEC_QUEUE0_IB_CNTL), ib_cntl);
+	}
+
+	ret = amdgpu_ring_test_helper(ring);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+static int vpe_v_6_1_ring_stop(struct amdgpu_vpe *vpe)
+{
+	struct amdgpu_device *adev = vpe->ring.adev;
+	uint32_t queue_reset, i;
+	int ret;
+
+	for (i = 0; i < vpe->num_instances; i++) {
+		if (amdgpu_ip_version(adev, VPE_HWIP, 0) == IP_VERSION(6, 1, 1))
+			queue_reset = RREG32(vpe_get_reg_offset(vpe, i, regVPEC_QUEUE_RESET_REQ_6_1_1));
+		else
+			queue_reset = RREG32(vpe_get_reg_offset(vpe, i, regVPEC_QUEUE_RESET_REQ));
+
+		queue_reset = REG_SET_FIELD(queue_reset, VPEC_QUEUE_RESET_REQ, QUEUE0_RESET, 1);
+
+		if (amdgpu_ip_version(adev, VPE_HWIP, 0) == IP_VERSION(6, 1, 1)) {
+			WREG32(vpe_get_reg_offset(vpe, i, regVPEC_QUEUE_RESET_REQ_6_1_1), queue_reset);
+			ret = SOC15_WAIT_ON_RREG(VPE, i, regVPEC_QUEUE_RESET_REQ_6_1_1, 0,
+						 VPEC_QUEUE_RESET_REQ__QUEUE0_RESET_MASK);
+		} else {
+			WREG32(vpe_get_reg_offset(vpe, i, regVPEC_QUEUE_RESET_REQ), queue_reset);
+			ret = SOC15_WAIT_ON_RREG(VPE, i, regVPEC_QUEUE_RESET_REQ, 0,
+						 VPEC_QUEUE_RESET_REQ__QUEUE0_RESET_MASK);
+		}
+
+		if (ret)
+			dev_err(adev->dev, "VPE queue reset failed\n");
+	}
+
+	vpe->ring.sched.ready = false;
+
+	return ret;
+}
+
+static int vpe_v6_1_set_trap_irq_state(struct amdgpu_device *adev,
+				       struct amdgpu_irq_src *source,
+				       unsigned int type,
+				       enum amdgpu_interrupt_state state)
+{
+	struct amdgpu_vpe *vpe = &adev->vpe;
+	uint32_t vpe_cntl;
+
+	if (amdgpu_ip_version(adev, VPE_HWIP, 0) == IP_VERSION(6, 1, 1))
+		vpe_cntl = RREG32(vpe_get_reg_offset(vpe, 0, regVPEC_CNTL_6_1_1));
+	else
+		vpe_cntl = RREG32(vpe_get_reg_offset(vpe, 0, regVPEC_CNTL));
+
+	vpe_cntl = REG_SET_FIELD(vpe_cntl, VPEC_CNTL, TRAP_ENABLE,
+				 state == AMDGPU_IRQ_STATE_ENABLE ? 1 : 0);
+
+	if (amdgpu_ip_version(adev, VPE_HWIP, 0) == IP_VERSION(6, 1, 1))
+		WREG32(vpe_get_reg_offset(vpe, 0, regVPEC_CNTL_6_1_1), vpe_cntl);
+	else
+		WREG32(vpe_get_reg_offset(vpe, 0, regVPEC_CNTL), vpe_cntl);
+
+	return 0;
+}
+
+static int vpe_v6_1_process_trap_irq(struct amdgpu_device *adev,
+				     struct amdgpu_irq_src *source,
+				     struct amdgpu_iv_entry *entry)
+{
+
+	dev_dbg(adev->dev, "IH: VPE trap\n");
+
+	switch (entry->client_id) {
+	case SOC21_IH_CLIENTID_VPE:
+		amdgpu_fence_process(&adev->vpe.ring);
+		break;
+	default:
+		break;
+	}
+
+	return 0;
+}
+
+static int vpe_v6_1_set_regs(struct amdgpu_vpe *vpe)
+{
+	struct amdgpu_device *adev = container_of(vpe, struct amdgpu_device, vpe);
+
+	vpe->regs.queue0_rb_rptr_lo = regVPEC_QUEUE0_RB_RPTR;
+	vpe->regs.queue0_rb_rptr_hi = regVPEC_QUEUE0_RB_RPTR_HI;
+	vpe->regs.queue0_rb_wptr_lo = regVPEC_QUEUE0_RB_WPTR;
+	vpe->regs.queue0_rb_wptr_hi = regVPEC_QUEUE0_RB_WPTR_HI;
+	vpe->regs.queue0_preempt = regVPEC_QUEUE0_PREEMPT;
+
+	if (amdgpu_ip_version(adev, VPE_HWIP, 0) == IP_VERSION(6, 1, 1))
+		vpe->regs.dpm_enable = regVPEC_PUB_DUMMY2_6_1_1;
+	else
+		vpe->regs.dpm_enable = regVPEC_PUB_DUMMY2;
+
+	vpe->regs.dpm_pratio = regVPEC_QUEUE6_DUMMY4;
+	vpe->regs.dpm_request_interval = regVPEC_QUEUE5_DUMMY3;
+	vpe->regs.dpm_decision_threshold = regVPEC_QUEUE5_DUMMY4;
+	vpe->regs.dpm_busy_clamp_threshold = regVPEC_QUEUE7_DUMMY2;
+	vpe->regs.dpm_idle_clamp_threshold = regVPEC_QUEUE7_DUMMY3;
+	vpe->regs.dpm_request_lv = regVPEC_QUEUE7_DUMMY1;
+	vpe->regs.context_indicator = regVPEC_QUEUE6_DUMMY3;
+
+	return 0;
+}
+
+static const struct vpe_funcs vpe_v6_1_funcs = {
+	.get_reg_offset = vpe_v6_1_get_reg_offset,
+	.set_regs = vpe_v6_1_set_regs,
+	.irq_init = vpe_v6_1_irq_init,
+	.init_microcode = amdgpu_vpe_init_microcode,
+	.load_microcode = vpe_v6_1_load_microcode,
+	.ring_init = amdgpu_vpe_ring_init,
+	.ring_start = vpe_v6_1_ring_start,
+	.ring_stop = vpe_v_6_1_ring_stop,
+	.ring_fini = amdgpu_vpe_ring_fini,
+};
+
+static const struct amdgpu_irq_src_funcs vpe_v6_1_trap_irq_funcs = {
+	.set = vpe_v6_1_set_trap_irq_state,
+	.process = vpe_v6_1_process_trap_irq,
+};
+
+void vpe_v6_1_set_funcs(struct amdgpu_vpe *vpe)
+{
+	vpe->funcs = &vpe_v6_1_funcs;
+	vpe->trap_irq.funcs = &vpe_v6_1_trap_irq_funcs;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/vpe_v6_1.h b/drivers/gpu/drm/amd/amdgpu/vpe_v6_1.h
new file mode 100644
index 000000000000..a9bea7905a77
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/vpe_v6_1.h
@@ -0,0 +1,29 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+#ifndef __VPE_V6_1_H__
+#define __VPE_V6_1_H__
+
+#include "amdgpu_vpe.h"
+
+void vpe_v6_1_set_funcs(struct amdgpu_vpe *vpe);
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdkfd/Kconfig b/drivers/gpu/drm/amd/amdkfd/Kconfig
index 8cc0a76ddf9f..16e12c9913f9 100644
--- a/drivers/gpu/drm/amd/amdkfd/Kconfig
+++ b/drivers/gpu/drm/amd/amdkfd/Kconfig
@@ -5,8 +5,7 @@
 
 config HSA_AMD
 	bool "HSA kernel driver for AMD GPU devices"
-	depends on DRM_AMDGPU && (X86_64 || ARM64 || PPC64)
-	imply AMD_IOMMU_V2 if X86_64
+	depends on DRM_AMDGPU && (X86_64 || ARM64 || PPC64 || (RISCV && 64BIT) || (LOONGARCH && 64BIT))
 	select HMM_MIRROR
 	select MMU_NOTIFIER
 	select DRM_AMDGPU_USERPTR
@@ -25,3 +24,17 @@ config HSA_AMD_SVM
 	  preemptions and one based on page faults. To enable page fault
 	  based memory management on most GFXv9 GPUs, set the module
 	  parameter amdgpu.noretry=0.
+
+config HSA_AMD_P2P
+	bool "HSA kernel driver support for peer-to-peer for AMD GPU devices"
+	depends on HSA_AMD && PCI_P2PDMA && DMABUF_MOVE_NOTIFY
+	help
+	  Enable peer-to-peer (P2P) communication between AMD GPUs over
+	  the PCIe bus. This can improve performance of multi-GPU compute
+	  applications and libraries by enabling GPUs to access data directly
+	  in peer GPUs' memory without intermediate copies in system memory.
+
+	  This P2P feature is only enabled on compatible chipsets, and between
+	  GPUs with large memory BARs that expose the entire VRAM in PCIe bus
+	  address space within the physical address limits of the GPUs.
+
diff --git a/drivers/gpu/drm/amd/amdkfd/Makefile b/drivers/gpu/drm/amd/amdkfd/Makefile
index c4f3aff11072..0ce08113c9f0 100644
--- a/drivers/gpu/drm/amd/amdkfd/Makefile
+++ b/drivers/gpu/drm/amd/amdkfd/Makefile
@@ -27,7 +27,6 @@ AMDKFD_FILES	:= $(AMDKFD_PATH)/kfd_module.o \
 		$(AMDKFD_PATH)/kfd_device.o \
 		$(AMDKFD_PATH)/kfd_chardev.o \
 		$(AMDKFD_PATH)/kfd_topology.o \
-		$(AMDKFD_PATH)/kfd_pasid.o \
 		$(AMDKFD_PATH)/kfd_doorbell.o \
 		$(AMDKFD_PATH)/kfd_flat_memory.o \
 		$(AMDKFD_PATH)/kfd_process.o \
@@ -37,6 +36,8 @@ AMDKFD_FILES	:= $(AMDKFD_PATH)/kfd_module.o \
 		$(AMDKFD_PATH)/kfd_mqd_manager_vi.o \
 		$(AMDKFD_PATH)/kfd_mqd_manager_v9.o \
 		$(AMDKFD_PATH)/kfd_mqd_manager_v10.o \
+		$(AMDKFD_PATH)/kfd_mqd_manager_v11.o \
+		$(AMDKFD_PATH)/kfd_mqd_manager_v12.o \
 		$(AMDKFD_PATH)/kfd_kernel_queue.o \
 		$(AMDKFD_PATH)/kfd_packet_manager.o \
 		$(AMDKFD_PATH)/kfd_packet_manager_vi.o \
@@ -47,18 +48,17 @@ AMDKFD_FILES	:= $(AMDKFD_PATH)/kfd_module.o \
 		$(AMDKFD_PATH)/kfd_device_queue_manager_vi.o \
 		$(AMDKFD_PATH)/kfd_device_queue_manager_v9.o \
 		$(AMDKFD_PATH)/kfd_device_queue_manager_v10.o \
+		$(AMDKFD_PATH)/kfd_device_queue_manager_v11.o \
+		$(AMDKFD_PATH)/kfd_device_queue_manager_v12.o \
 		$(AMDKFD_PATH)/kfd_interrupt.o \
 		$(AMDKFD_PATH)/kfd_events.o \
 		$(AMDKFD_PATH)/cik_event_interrupt.o \
 		$(AMDKFD_PATH)/kfd_int_process_v9.o \
-		$(AMDKFD_PATH)/kfd_dbgdev.o \
-		$(AMDKFD_PATH)/kfd_dbgmgr.o \
+		$(AMDKFD_PATH)/kfd_int_process_v10.o \
+		$(AMDKFD_PATH)/kfd_int_process_v11.o \
 		$(AMDKFD_PATH)/kfd_smi_events.o \
-		$(AMDKFD_PATH)/kfd_crat.o
-
-ifneq ($(CONFIG_AMD_IOMMU_V2),)
-AMDKFD_FILES += $(AMDKFD_PATH)/kfd_iommu.o
-endif
+		$(AMDKFD_PATH)/kfd_crat.o \
+		$(AMDKFD_PATH)/kfd_debug.o
 
 ifneq ($(CONFIG_DEBUG_FS),)
 AMDKFD_FILES += $(AMDKFD_PATH)/kfd_debugfs.o
diff --git a/drivers/gpu/drm/amd/amdkfd/cik_event_interrupt.c b/drivers/gpu/drm/amd/amdkfd/cik_event_interrupt.c
index f6233019f042..73acbe0b7c21 100644
--- a/drivers/gpu/drm/amd/amdkfd/cik_event_interrupt.c
+++ b/drivers/gpu/drm/amd/amdkfd/cik_event_interrupt.c
@@ -26,7 +26,7 @@
 #include "amdgpu_amdkfd.h"
 #include "kfd_smi_events.h"
 
-static bool cik_event_interrupt_isr(struct kfd_dev *dev,
+static bool cik_event_interrupt_isr(struct kfd_node *dev,
 					const uint32_t *ih_ring_entry,
 					uint32_t *patched_ihre,
 					bool *patched_flag)
@@ -43,15 +43,15 @@ static bool cik_event_interrupt_isr(struct kfd_dev *dev,
 	 */
 	if ((ihre->source_id == CIK_INTSRC_GFX_PAGE_INV_FAULT ||
 		ihre->source_id == CIK_INTSRC_GFX_MEM_PROT_FAULT) &&
-		dev->device_info->asic_family == CHIP_HAWAII) {
+		dev->adev->asic_type == CHIP_HAWAII) {
 		struct cik_ih_ring_entry *tmp_ihre =
 			(struct cik_ih_ring_entry *)patched_ihre;
 
 		*patched_flag = true;
 		*tmp_ihre = *ihre;
 
-		vmid = f2g->read_vmid_from_vmfault_reg(dev->kgd);
-		ret = f2g->get_atc_vmid_pasid_mapping_info(dev->kgd, vmid, &pasid);
+		vmid = f2g->read_vmid_from_vmfault_reg(dev->adev);
+		ret = f2g->get_atc_vmid_pasid_mapping_info(dev->adev, vmid, &pasid);
 
 		tmp_ihre->ring_id &= 0x000000ff;
 		tmp_ihre->ring_id |= vmid << 8;
@@ -85,13 +85,12 @@ static bool cik_event_interrupt_isr(struct kfd_dev *dev,
 		!amdgpu_no_queue_eviction_on_vm_fault);
 }
 
-static void cik_event_interrupt_wq(struct kfd_dev *dev,
+static void cik_event_interrupt_wq(struct kfd_node *dev,
 					const uint32_t *ih_ring_entry)
 {
 	const struct cik_ih_ring_entry *ihre =
 			(const struct cik_ih_ring_entry *)ih_ring_entry;
 	uint32_t context_id = ihre->data & 0xfffffff;
-	unsigned int vmid  = (ihre->ring_id & 0x0000ff00) >> 8;
 	u32 pasid = (ihre->ring_id & 0xffff0000) >> 16;
 
 	if (pasid == 0)
@@ -107,20 +106,26 @@ static void cik_event_interrupt_wq(struct kfd_dev *dev,
 		kfd_signal_hw_exception_event(pasid);
 	else if (ihre->source_id == CIK_INTSRC_GFX_PAGE_INV_FAULT ||
 		ihre->source_id == CIK_INTSRC_GFX_MEM_PROT_FAULT) {
+		struct kfd_process_device *pdd = NULL;
 		struct kfd_vm_fault_info info;
+		struct kfd_process *p;
 
 		kfd_smi_event_update_vmfault(dev, pasid);
-		kfd_process_vm_fault(dev->dqm, pasid);
+		p = kfd_lookup_process_by_pasid(pasid, &pdd);
+		if (!pdd)
+			return;
+
+		kfd_evict_process_device(pdd);
 
 		memset(&info, 0, sizeof(info));
-		amdgpu_amdkfd_gpuvm_get_vm_fault_info(dev->kgd, &info);
-		if (!info.page_addr && !info.status)
+		amdgpu_amdkfd_gpuvm_get_vm_fault_info(dev->adev, &info);
+		if (!info.page_addr && !info.status) {
+			kfd_unref_process(p);
 			return;
+		}
 
-		if (info.vmid == vmid)
-			kfd_signal_vm_fault_event(dev, pasid, &info);
-		else
-			kfd_signal_vm_fault_event(dev, pasid, NULL);
+		kfd_signal_vm_fault_event(pdd, &info, NULL);
+		kfd_unref_process(p);
 	}
 }
 
diff --git a/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler.h b/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler.h
index 475f89700c74..0320163b6e74 100644
--- a/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler.h
+++ b/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler.h
@@ -166,7 +166,7 @@ static const uint32_t cwsr_trap_gfx8_hex[] = {
 	0x807c847c, 0x806eff6e,
 	0x00000400, 0xbf0a757c,
 	0xbf85ffef, 0xbf9c0000,
-	0xbf8200cd, 0xbef8007e,
+	0xbf8200ce, 0xbef8007e,
 	0x8679ff7f, 0x0000ffff,
 	0x8779ff79, 0x00040000,
 	0xbefa0080, 0xbefb00ff,
@@ -212,176 +212,185 @@ static const uint32_t cwsr_trap_gfx8_hex[] = {
 	0x761e0000, 0xe0524100,
 	0x761e0100, 0xe0524200,
 	0x761e0200, 0xe0524300,
-	0x761e0300, 0xb8f22a05,
-	0x80728172, 0x8e728a72,
-	0xb8f61605, 0x80768176,
-	0x8e768676, 0x80727672,
-	0x80f2c072, 0xb8f31605,
-	0x80738173, 0x8e738473,
-	0x8e7a8273, 0xbefa00ff,
-	0x01000000, 0xbefc0073,
-	0xc031003c, 0x00000072,
-	0x80f2c072, 0xbf8c007f,
-	0x80fc907c, 0xbe802d00,
-	0xbe822d02, 0xbe842d04,
-	0xbe862d06, 0xbe882d08,
-	0xbe8a2d0a, 0xbe8c2d0c,
-	0xbe8e2d0e, 0xbf06807c,
-	0xbf84fff1, 0xb8f22a05,
-	0x80728172, 0x8e728a72,
-	0xb8f61605, 0x80768176,
-	0x8e768676, 0x80727672,
-	0xbefa0084, 0xbefa00ff,
-	0x01000000, 0xc0211cfc,
+	0x761e0300, 0xbf8c0f70,
+	0xb8f22a05, 0x80728172,
+	0x8e728a72, 0xb8f61605,
+	0x80768176, 0x8e768676,
+	0x80727672, 0x80f2c072,
+	0xb8f31605, 0x80738173,
+	0x8e738473, 0x8e7a8273,
+	0xbefa00ff, 0x01000000,
+	0xbefc0073, 0xc031003c,
+	0x00000072, 0x80f2c072,
+	0xbf8c007f, 0x80fc907c,
+	0xbe802d00, 0xbe822d02,
+	0xbe842d04, 0xbe862d06,
+	0xbe882d08, 0xbe8a2d0a,
+	0xbe8c2d0c, 0xbe8e2d0e,
+	0xbf06807c, 0xbf84fff1,
+	0xb8f22a05, 0x80728172,
+	0x8e728a72, 0xb8f61605,
+	0x80768176, 0x8e768676,
+	0x80727672, 0xbefa0084,
+	0xbefa00ff, 0x01000000,
+	0xc0211cfc, 0x00000072,
+	0x80728472, 0xc0211c3c,
 	0x00000072, 0x80728472,
-	0xc0211c3c, 0x00000072,
-	0x80728472, 0xc0211c7c,
+	0xc0211c7c, 0x00000072,
+	0x80728472, 0xc0211bbc,
 	0x00000072, 0x80728472,
-	0xc0211bbc, 0x00000072,
-	0x80728472, 0xc0211bfc,
+	0xc0211bfc, 0x00000072,
+	0x80728472, 0xc0211d3c,
 	0x00000072, 0x80728472,
-	0xc0211d3c, 0x00000072,
-	0x80728472, 0xc0211d7c,
+	0xc0211d7c, 0x00000072,
+	0x80728472, 0xc0211a3c,
 	0x00000072, 0x80728472,
-	0xc0211a3c, 0x00000072,
-	0x80728472, 0xc0211a7c,
+	0xc0211a7c, 0x00000072,
+	0x80728472, 0xc0211dfc,
 	0x00000072, 0x80728472,
-	0xc0211dfc, 0x00000072,
-	0x80728472, 0xc0211b3c,
+	0xc0211b3c, 0x00000072,
+	0x80728472, 0xc0211b7c,
 	0x00000072, 0x80728472,
-	0xc0211b7c, 0x00000072,
-	0x80728472, 0xbf8c007f,
-	0xbefc0073, 0xbefe006e,
-	0xbeff006f, 0x867375ff,
-	0x000003ff, 0xb9734803,
-	0x867375ff, 0xfffff800,
-	0x8f738b73, 0xb973a2c3,
-	0xb977f801, 0x8673ff71,
-	0xf0000000, 0x8f739c73,
-	0x8e739073, 0xbef60080,
-	0x87767376, 0x8673ff71,
-	0x08000000, 0x8f739b73,
-	0x8e738f73, 0x87767376,
-	0x8673ff74, 0x00800000,
-	0x8f739773, 0xb976f807,
-	0x8671ff71, 0x0000ffff,
-	0x86fe7e7e, 0x86ea6a6a,
-	0x8f768374, 0xb976e0c2,
-	0xbf800002, 0xb9740002,
-	0xbf8a0000, 0x95807370,
-	0xbf810000, 0x00000000,
+	0xbf8c007f, 0xbefc0073,
+	0xbefe006e, 0xbeff006f,
+	0x867375ff, 0x000003ff,
+	0xb9734803, 0x867375ff,
+	0xfffff800, 0x8f738b73,
+	0xb973a2c3, 0xb977f801,
+	0x8673ff71, 0xf0000000,
+	0x8f739c73, 0x8e739073,
+	0xbef60080, 0x87767376,
+	0x8673ff71, 0x08000000,
+	0x8f739b73, 0x8e738f73,
+	0x87767376, 0x8673ff74,
+	0x00800000, 0x8f739773,
+	0xb976f807, 0x8671ff71,
+	0x0000ffff, 0x86fe7e7e,
+	0x86ea6a6a, 0x8f768374,
+	0xb976e0c2, 0xbf800002,
+	0xb9740002, 0xbf8a0000,
+	0x95807370, 0xbf810000,
 };
 
 
 static const uint32_t cwsr_trap_gfx9_hex[] = {
-	0xbf820001, 0xbf820248,
-	0xb8f8f802, 0x89788678,
-	0xb8eef801, 0x866eff6e,
-	0x00000800, 0xbf840003,
+	0xbf820001, 0xbf820259,
+	0xb8f8f802, 0x8978ff78,
+	0x00020006, 0xb8fbf803,
 	0x866eff78, 0x00002000,
-	0xbf840016, 0xb8fbf803,
+	0xbf840009, 0x866eff6d,
+	0x00ff0000, 0xbf85001e,
 	0x866eff7b, 0x00000400,
-	0xbf85003b, 0x866eff7b,
-	0x00000800, 0xbf850003,
-	0x866eff7b, 0x00000100,
-	0xbf84000c, 0x866eff78,
-	0x00002000, 0xbf840005,
-	0xbf8e0010, 0xb8eef803,
-	0x866eff6e, 0x00000400,
-	0xbf84fffb, 0x8778ff78,
-	0x00002000, 0x80ec886c,
-	0x82ed806d, 0xb8eef807,
-	0x866fff6e, 0x001f8000,
-	0x8e6f8b6f, 0x8977ff77,
-	0xfc000000, 0x87776f77,
-	0x896eff6e, 0x001f8000,
-	0xb96ef807, 0xb8faf812,
+	0xbf850055, 0xbf8e0010,
+	0xb8fbf803, 0xbf82fffa,
+	0x866eff7b, 0x03c00900,
+	0xbf850015, 0x866eff7b,
+	0x000071ff, 0xbf840008,
+	0x866fff7b, 0x00007080,
+	0xbf840001, 0xbeee1a87,
+	0xb8eff801, 0x8e6e8c6e,
+	0x866e6f6e, 0xbf85000a,
+	0x866eff6d, 0x00ff0000,
+	0xbf850007, 0xb8eef801,
+	0x866eff6e, 0x00000800,
+	0xbf850003, 0x866eff7b,
+	0x00000400, 0xbf85003a,
+	0xb8faf807, 0x867aff7a,
+	0x001f8000, 0x8e7a8b7a,
+	0x8977ff77, 0xfc000000,
+	0x87777a77, 0xba7ff807,
+	0x00000000, 0xb8faf812,
 	0xb8fbf813, 0x8efa887a,
-	0xc0071bbd, 0x00000000,
-	0xbf8cc07f, 0xc0071ebd,
-	0x00000008, 0xbf8cc07f,
-	0x86ee6e6e, 0xbf840001,
-	0xbe801d6e, 0xb8fbf803,
-	0x867bff7b, 0x000001ff,
+	0xbf0d8f7b, 0xbf840002,
+	0x877bff7b, 0xffff0000,
+	0xc0031bbd, 0x00000010,
+	0xbf8cc07f, 0x8e6e976e,
+	0x8977ff77, 0x00800000,
+	0x87776e77, 0xc0071bbd,
+	0x00000000, 0xbf8cc07f,
+	0xc0071ebd, 0x00000008,
+	0xbf8cc07f, 0x86ee6e6e,
+	0xbf840001, 0xbe801d6e,
+	0x866eff6d, 0x01ff0000,
+	0xbf850005, 0x8778ff78,
+	0x00002000, 0x80ec886c,
+	0x82ed806d, 0xbf820005,
+	0x866eff6d, 0x01000000,
 	0xbf850002, 0x806c846c,
 	0x826d806d, 0x866dff6d,
-	0x0000ffff, 0x8f6e8b77,
-	0x866eff6e, 0x001f8000,
-	0xb96ef807, 0x86fe7e7e,
+	0x0000ffff, 0x8f7a8b77,
+	0x867aff7a, 0x001f8000,
+	0xb97af807, 0x86fe7e7e,
 	0x86ea6a6a, 0x8f6e8378,
 	0xb96ee0c2, 0xbf800002,
 	0xb9780002, 0xbe801f6c,
 	0x866dff6d, 0x0000ffff,
 	0xbefa0080, 0xb97a0283,
-	0xb8fa2407, 0x8e7a9b7a,
-	0x876d7a6d, 0xb8fa03c7,
-	0x8e7a9a7a, 0x876d7a6d,
 	0xb8faf807, 0x867aff7a,
-	0x00007fff, 0xb97af807,
-	0xbeee007e, 0xbeef007f,
-	0xbefe0180, 0xbf900004,
-	0x877a8478, 0xb97af802,
-	0xbf8e0002, 0xbf88fffe,
-	0xb8fa2a05, 0x807a817a,
-	0x8e7a8a7a, 0xb8fb1605,
-	0x807b817b, 0x8e7b867b,
-	0x807a7b7a, 0x807a7e7a,
-	0x827b807f, 0x867bff7b,
-	0x0000ffff, 0xc04b1c3d,
-	0x00000050, 0xbf8cc07f,
-	0xc04b1d3d, 0x00000060,
-	0xbf8cc07f, 0xc0431e7d,
-	0x00000074, 0xbf8cc07f,
-	0xbef4007e, 0x8675ff7f,
-	0x0000ffff, 0x8775ff75,
-	0x00040000, 0xbef60080,
-	0xbef700ff, 0x00807fac,
-	0x867aff7f, 0x08000000,
-	0x8f7a837a, 0x87777a77,
-	0x867aff7f, 0x70000000,
-	0x8f7a817a, 0x87777a77,
-	0xbef1007c, 0xbef00080,
-	0xb8f02a05, 0x80708170,
-	0x8e708a70, 0xb8fa1605,
-	0x807a817a, 0x8e7a867a,
-	0x80707a70, 0xbef60084,
-	0xbef600ff, 0x01000000,
-	0xbefe007c, 0xbefc0070,
-	0xc0611c7a, 0x0000007c,
-	0xbf8cc07f, 0x80708470,
-	0xbefc007e, 0xbefe007c,
-	0xbefc0070, 0xc0611b3a,
+	0x001f8000, 0x8e7a8b7a,
+	0x8977ff77, 0xfc000000,
+	0x87777a77, 0xba7ff807,
+	0x00000000, 0xbeee007e,
+	0xbeef007f, 0xbefe0180,
+	0xbf900004, 0x877a8478,
+	0xb97af802, 0xbf8e0002,
+	0xbf88fffe, 0xb8fa2a05,
+	0x807a817a, 0x8e7a8a7a,
+	0xb8fb1605, 0x807b817b,
+	0x8e7b867b, 0x807a7b7a,
+	0x807a7e7a, 0x827b807f,
+	0x867bff7b, 0x0000ffff,
+	0xc04b1c3d, 0x00000050,
+	0xbf8cc07f, 0xc04b1d3d,
+	0x00000060, 0xbf8cc07f,
+	0xc0431e7d, 0x00000074,
+	0xbf8cc07f, 0xbef4007e,
+	0x8675ff7f, 0x0000ffff,
+	0x8775ff75, 0x00040000,
+	0xbef60080, 0xbef700ff,
+	0x00807fac, 0xbef1007c,
+	0xbef00080, 0xb8f02a05,
+	0x80708170, 0x8e708a70,
+	0xb8fa1605, 0x807a817a,
+	0x8e7a867a, 0x80707a70,
+	0xbef60084, 0xbef600ff,
+	0x01000000, 0xbefe007c,
+	0xbefc0070, 0xc0611c7a,
 	0x0000007c, 0xbf8cc07f,
 	0x80708470, 0xbefc007e,
 	0xbefe007c, 0xbefc0070,
-	0xc0611b7a, 0x0000007c,
+	0xc0611b3a, 0x0000007c,
 	0xbf8cc07f, 0x80708470,
 	0xbefc007e, 0xbefe007c,
-	0xbefc0070, 0xc0611bba,
+	0xbefc0070, 0xc0611b7a,
 	0x0000007c, 0xbf8cc07f,
 	0x80708470, 0xbefc007e,
 	0xbefe007c, 0xbefc0070,
-	0xc0611bfa, 0x0000007c,
+	0xc0611bba, 0x0000007c,
 	0xbf8cc07f, 0x80708470,
 	0xbefc007e, 0xbefe007c,
-	0xbefc0070, 0xc0611e3a,
-	0x0000007c, 0xbf8cc07f,
-	0x80708470, 0xbefc007e,
-	0xb8fbf803, 0xbefe007c,
-	0xbefc0070, 0xc0611efa,
+	0xbefc0070, 0xc0611bfa,
 	0x0000007c, 0xbf8cc07f,
 	0x80708470, 0xbefc007e,
 	0xbefe007c, 0xbefc0070,
-	0xc0611a3a, 0x0000007c,
+	0xc0611e3a, 0x0000007c,
+	0xbf8cc07f, 0x80708470,
+	0xbefc007e, 0xb8fbf803,
+	0xbefe007c, 0xbefc0070,
+	0xc0611efa, 0x0000007c,
 	0xbf8cc07f, 0x80708470,
 	0xbefc007e, 0xbefe007c,
-	0xbefc0070, 0xc0611a7a,
-	0x0000007c, 0xbf8cc07f,
-	0x80708470, 0xbefc007e,
-	0xb8f1f801, 0xbefe007c,
-	0xbefc0070, 0xc0611c7a,
+	0xbefc0070, 0xc0611a3a,
 	0x0000007c, 0xbf8cc07f,
 	0x80708470, 0xbefc007e,
+	0xbefe007c, 0xbefc0070,
+	0xc0611a7a, 0x0000007c,
+	0xbf8cc07f, 0x80708470,
+	0xbefc007e, 0xb8f1f801,
+	0xbefe007c, 0xbefc0070,
+	0xc0611c7a, 0x0000007c,
+	0xbf8cc07f, 0x80708470,
+	0xbefc007e, 0xbf108080,
 	0x867aff7f, 0x04000000,
 	0xbeef0080, 0x876f6f7a,
 	0xb8f02a05, 0x80708170,
@@ -509,97 +518,92 @@ static const uint32_t cwsr_trap_gfx9_hex[] = {
 	0x00000400, 0xbefe00c1,
 	0xbeff00c1, 0xb8fb2a05,
 	0x807b817b, 0x8e7b827b,
-	0x8e76887b, 0xbef600ff,
-	0x01000000, 0xbefc0084,
-	0xbf0a7b7c, 0xbf84006d,
-	0xbf11017c, 0x807bff7b,
-	0x00001000, 0x867aff78,
-	0x00400000, 0xbf850003,
-	0xb8faf803, 0x897a7aff,
-	0x10000000, 0xbf850051,
-	0xbe840080, 0xd2890000,
-	0x00000900, 0x80048104,
-	0xd2890001, 0x00000900,
-	0x80048104, 0xd2890002,
-	0x00000900, 0x80048104,
-	0xd2890003, 0x00000900,
-	0x80048104, 0xc069003a,
-	0x00000070, 0xbf8cc07f,
-	0x80709070, 0xbf06c004,
-	0xbf84ffee, 0xbe840080,
-	0xd2890000, 0x00000901,
+	0xbef600ff, 0x01000000,
+	0xbefc0084, 0xbf0a7b7c,
+	0xbf84006d, 0xbf11017c,
+	0x807bff7b, 0x00001000,
+	0x867aff78, 0x00400000,
+	0xbf850003, 0xb8faf803,
+	0x897a7aff, 0x10000000,
+	0xbf850051, 0xbe840080,
+	0xd2890000, 0x00000900,
 	0x80048104, 0xd2890001,
-	0x00000901, 0x80048104,
-	0xd2890002, 0x00000901,
+	0x00000900, 0x80048104,
+	0xd2890002, 0x00000900,
 	0x80048104, 0xd2890003,
-	0x00000901, 0x80048104,
+	0x00000900, 0x80048104,
 	0xc069003a, 0x00000070,
 	0xbf8cc07f, 0x80709070,
 	0xbf06c004, 0xbf84ffee,
 	0xbe840080, 0xd2890000,
-	0x00000902, 0x80048104,
-	0xd2890001, 0x00000902,
+	0x00000901, 0x80048104,
+	0xd2890001, 0x00000901,
 	0x80048104, 0xd2890002,
-	0x00000902, 0x80048104,
-	0xd2890003, 0x00000902,
+	0x00000901, 0x80048104,
+	0xd2890003, 0x00000901,
 	0x80048104, 0xc069003a,
 	0x00000070, 0xbf8cc07f,
 	0x80709070, 0xbf06c004,
 	0xbf84ffee, 0xbe840080,
-	0xd2890000, 0x00000903,
+	0xd2890000, 0x00000902,
 	0x80048104, 0xd2890001,
-	0x00000903, 0x80048104,
-	0xd2890002, 0x00000903,
+	0x00000902, 0x80048104,
+	0xd2890002, 0x00000902,
 	0x80048104, 0xd2890003,
-	0x00000903, 0x80048104,
+	0x00000902, 0x80048104,
 	0xc069003a, 0x00000070,
 	0xbf8cc07f, 0x80709070,
 	0xbf06c004, 0xbf84ffee,
-	0x807c847c, 0xbf0a7b7c,
-	0xbf85ffb1, 0xbf9c0000,
-	0xbf820012, 0x7e000300,
-	0x7e020301, 0x7e040302,
-	0x7e060303, 0xe0724000,
-	0x701d0000, 0xe0724100,
-	0x701d0100, 0xe0724200,
-	0x701d0200, 0xe0724300,
-	0x701d0300, 0x807c847c,
-	0x8070ff70, 0x00000400,
-	0xbf0a7b7c, 0xbf85ffef,
-	0xbf9c0000, 0xbf8200da,
-	0xbef4007e, 0x8675ff7f,
-	0x0000ffff, 0x8775ff75,
-	0x00040000, 0xbef60080,
-	0xbef700ff, 0x00807fac,
-	0x866eff7f, 0x08000000,
-	0x8f6e836e, 0x87776e77,
-	0x866eff7f, 0x70000000,
-	0x8f6e816e, 0x87776e77,
-	0x866eff7f, 0x04000000,
-	0xbf84001e, 0xbefe00c1,
-	0xbeff00c1, 0xb8ef4306,
-	0x866fc16f, 0xbf840019,
-	0x8e6f866f, 0x8e6f826f,
-	0xbef6006f, 0xb8f82a05,
-	0x80788178, 0x8e788a78,
-	0xb8ee1605, 0x806e816e,
-	0x8e6e866e, 0x80786e78,
-	0x8078ff78, 0x00000080,
-	0xbef600ff, 0x01000000,
-	0xbefc0080, 0xe0510000,
-	0x781d0000, 0xe0510100,
-	0x781d0000, 0x807cff7c,
-	0x00000200, 0x8078ff78,
-	0x00000200, 0xbf0a6f7c,
-	0xbf85fff6, 0xbef80080,
+	0xbe840080, 0xd2890000,
+	0x00000903, 0x80048104,
+	0xd2890001, 0x00000903,
+	0x80048104, 0xd2890002,
+	0x00000903, 0x80048104,
+	0xd2890003, 0x00000903,
+	0x80048104, 0xc069003a,
+	0x00000070, 0xbf8cc07f,
+	0x80709070, 0xbf06c004,
+	0xbf84ffee, 0x807c847c,
+	0xbf0a7b7c, 0xbf85ffb1,
+	0xbf9c0000, 0xbf820012,
+	0x7e000300, 0x7e020301,
+	0x7e040302, 0x7e060303,
+	0xe0724000, 0x701d0000,
+	0xe0724100, 0x701d0100,
+	0xe0724200, 0x701d0200,
+	0xe0724300, 0x701d0300,
+	0x807c847c, 0x8070ff70,
+	0x00000400, 0xbf0a7b7c,
+	0xbf85ffef, 0xbf9c0000,
+	0xbf8200c7, 0xbef4007e,
+	0x8675ff7f, 0x0000ffff,
+	0x8775ff75, 0x00040000,
+	0xbef60080, 0xbef700ff,
+	0x00807fac, 0x866eff7f,
+	0x04000000, 0xbf84001e,
+	0xbefe00c1, 0xbeff00c1,
+	0xb8ef4306, 0x866fc16f,
+	0xbf840019, 0x8e6f866f,
+	0x8e6f826f, 0xbef6006f,
+	0xb8f82a05, 0x80788178,
+	0x8e788a78, 0xb8ee1605,
+	0x806e816e, 0x8e6e866e,
+	0x80786e78, 0x8078ff78,
+	0x00000080, 0xbef600ff,
+	0x01000000, 0xbefc0080,
+	0xe0510000, 0x781d0000,
+	0xe0510100, 0x781d0000,
+	0x807cff7c, 0x00000200,
+	0x8078ff78, 0x00000200,
+	0xbf0a6f7c, 0xbf85fff6,
 	0xbefe00c1, 0xbeff00c1,
-	0xb8ef2a05, 0x806f816f,
-	0x8e6f826f, 0x8e76886f,
 	0xbef600ff, 0x01000000,
+	0xb8ef2a05, 0x806f816f,
+	0x8e6f826f, 0x806fff6f,
+	0x00008000, 0xbef80080,
 	0xbeee0078, 0x8078ff78,
 	0x00000400, 0xbefc0084,
-	0xbf11087c, 0x806fff6f,
-	0x00008000, 0xe0524000,
+	0xbf11087c, 0xe0524000,
 	0x781d0000, 0xe0524100,
 	0x781d0100, 0xe0524200,
 	0x781d0200, 0xe0524300,
@@ -613,157 +617,248 @@ static const uint32_t cwsr_trap_gfx9_hex[] = {
 	0xe0524100, 0x6e1d0100,
 	0xe0524200, 0x6e1d0200,
 	0xe0524300, 0x6e1d0300,
+	0xbf8c0f70, 0xb8f82a05,
+	0x80788178, 0x8e788a78,
+	0xb8ee1605, 0x806e816e,
+	0x8e6e866e, 0x80786e78,
+	0x80f8c078, 0xb8ef1605,
+	0x806f816f, 0x8e6f846f,
+	0x8e76826f, 0xbef600ff,
+	0x01000000, 0xbefc006f,
+	0xc031003a, 0x00000078,
+	0x80f8c078, 0xbf8cc07f,
+	0x80fc907c, 0xbf800000,
+	0xbe802d00, 0xbe822d02,
+	0xbe842d04, 0xbe862d06,
+	0xbe882d08, 0xbe8a2d0a,
+	0xbe8c2d0c, 0xbe8e2d0e,
+	0xbf06807c, 0xbf84fff0,
 	0xb8f82a05, 0x80788178,
 	0x8e788a78, 0xb8ee1605,
 	0x806e816e, 0x8e6e866e,
-	0x80786e78, 0x80f8c078,
-	0xb8ef1605, 0x806f816f,
-	0x8e6f846f, 0x8e76826f,
+	0x80786e78, 0xbef60084,
 	0xbef600ff, 0x01000000,
-	0xbefc006f, 0xc031003a,
-	0x00000078, 0x80f8c078,
-	0xbf8cc07f, 0x80fc907c,
-	0xbf800000, 0xbe802d00,
-	0xbe822d02, 0xbe842d04,
-	0xbe862d06, 0xbe882d08,
-	0xbe8a2d0a, 0xbe8c2d0c,
-	0xbe8e2d0e, 0xbf06807c,
-	0xbf84fff0, 0xb8f82a05,
-	0x80788178, 0x8e788a78,
-	0xb8ee1605, 0x806e816e,
-	0x8e6e866e, 0x80786e78,
-	0xbef60084, 0xbef600ff,
-	0x01000000, 0xc0211bfa,
+	0xc0211bfa, 0x00000078,
+	0x80788478, 0xc0211b3a,
 	0x00000078, 0x80788478,
-	0xc0211b3a, 0x00000078,
-	0x80788478, 0xc0211b7a,
+	0xc0211b7a, 0x00000078,
+	0x80788478, 0xc0211c3a,
 	0x00000078, 0x80788478,
-	0xc0211c3a, 0x00000078,
-	0x80788478, 0xc0211c7a,
+	0xc0211c7a, 0x00000078,
+	0x80788478, 0xc0211eba,
 	0x00000078, 0x80788478,
-	0xc0211eba, 0x00000078,
-	0x80788478, 0xc0211efa,
+	0xc0211efa, 0x00000078,
+	0x80788478, 0xc0211a3a,
 	0x00000078, 0x80788478,
-	0xc0211a3a, 0x00000078,
-	0x80788478, 0xc0211a7a,
+	0xc0211a7a, 0x00000078,
+	0x80788478, 0xc0211cfa,
 	0x00000078, 0x80788478,
-	0xc0211cfa, 0x00000078,
-	0x80788478, 0xbf8cc07f,
-	0xbefc006f, 0xbefe0070,
-	0xbeff0071, 0x866f7bff,
-	0x000003ff, 0xb96f4803,
-	0x866f7bff, 0xfffff800,
-	0x8f6f8b6f, 0xb96fa2c3,
-	0xb973f801, 0xb8ee2a05,
-	0x806e816e, 0x8e6e8a6e,
-	0xb8ef1605, 0x806f816f,
-	0x8e6f866f, 0x806e6f6e,
-	0x806e746e, 0x826f8075,
-	0x866fff6f, 0x0000ffff,
-	0xc00b1c37, 0x00000050,
-	0xc00b1d37, 0x00000060,
-	0xc0031e77, 0x00000074,
-	0xbf8cc07f, 0x866fff6d,
-	0xf8000000, 0x8f6f9b6f,
-	0x8e6f906f, 0xbeee0080,
-	0x876e6f6e, 0x866fff6d,
-	0x04000000, 0x8f6f9a6f,
-	0x8e6f8f6f, 0x876e6f6e,
-	0x866fff7a, 0x00800000,
-	0x8f6f976f, 0xb96ef807,
+	0xbf8cc07f, 0xbefc006f,
+	0xbefe0070, 0xbeff0071,
+	0x866f7bff, 0x000003ff,
+	0xb96f4803, 0x866f7bff,
+	0xfffff800, 0x8f6f8b6f,
+	0xb96fa2c3, 0xb973f801,
+	0xb8ee2a05, 0x806e816e,
+	0x8e6e8a6e, 0xb8ef1605,
+	0x806f816f, 0x8e6f866f,
+	0x806e6f6e, 0x806e746e,
+	0x826f8075, 0x866fff6f,
+	0x0000ffff, 0xc00b1c37,
+	0x00000050, 0xc00b1d37,
+	0x00000060, 0xc0031e77,
+	0x00000074, 0xbf8cc07f,
+	0x8f6e8b77, 0x866eff6e,
+	0x001f8000, 0xb96ef807,
 	0x866dff6d, 0x0000ffff,
 	0x86fe7e7e, 0x86ea6a6a,
 	0x8f6e837a, 0xb96ee0c2,
 	0xbf800002, 0xb97a0002,
-	0xbf8a0000, 0x95806f6c,
-	0xbf810000, 0x00000000,
+	0xbf8a0000, 0xbe801f6c,
+	0xbf9b0000, 0x00000000,
 };
 
 static const uint32_t cwsr_trap_nv1x_hex[] = {
-	0xbf820001, 0xbf8201cd,
+	0xbf820001, 0xbf820393,
 	0xb0804004, 0xb978f802,
-	0x8a788678, 0xb96ef801,
-	0x876eff6e, 0x00000800,
-	0xbf840003, 0x876eff78,
+	0x8a78ff78, 0x00020006,
+	0xb97bf803, 0x876eff78,
 	0x00002000, 0xbf840009,
-	0xb97bf803, 0x876eff7b,
-	0x00000400, 0xbf850033,
-	0x876eff7b, 0x00000100,
-	0xbf840002, 0x8878ff78,
-	0x00002000, 0x8a77ff77,
-	0xff000000, 0xb96ef807,
-	0x876fff6e, 0x02000000,
-	0x8f6f866f, 0x88776f77,
-	0x876fff6e, 0x003f8000,
-	0x8f6f896f, 0x88776f77,
-	0x8a6eff6e, 0x023f8000,
-	0xb9eef807, 0xb97af812,
+	0x876eff6d, 0x00ff0000,
+	0xbf85001e, 0x876eff7b,
+	0x00000400, 0xbf85005b,
+	0xbf8e0010, 0xb97bf803,
+	0xbf82fffa, 0x876eff7b,
+	0x00000900, 0xbf850015,
+	0x876eff7b, 0x000071ff,
+	0xbf840008, 0x876fff7b,
+	0x00007080, 0xbf840001,
+	0xbeee1d87, 0xb96ff801,
+	0x8f6e8c6e, 0x876e6f6e,
+	0xbf85000a, 0x876eff6d,
+	0x00ff0000, 0xbf850007,
+	0xb96ef801, 0x876eff6e,
+	0x00000800, 0xbf850003,
+	0x876eff7b, 0x00000400,
+	0xbf850040, 0x8a77ff77,
+	0xff000000, 0xb97af807,
+	0x877bff7a, 0x02000000,
+	0x8f7b867b, 0x88777b77,
+	0x877bff7a, 0x003f8000,
+	0x8f7b897b, 0x88777b77,
+	0x8a7aff7a, 0x023f8000,
+	0xb9faf807, 0xb97af812,
 	0xb97bf813, 0x8ffa887a,
-	0xf4051bbd, 0xfa000000,
-	0xbf8cc07f, 0xf4051ebd,
-	0xfa000008, 0xbf8cc07f,
-	0x87ee6e6e, 0xbf840001,
-	0xbe80206e, 0xb97bf803,
-	0x877bff7b, 0x000001ff,
-	0xbf850002, 0x806c846c,
+	0xbf0d8f7b, 0xbf840002,
+	0x887bff7b, 0xffff0000,
+	0xf4011bbd, 0xfa000010,
+	0xbf8c0000, 0x8f6e976e,
+	0x8a77ff77, 0x00800000,
+	0x88776e77, 0xf4051bbd,
+	0xfa000000, 0xbf8c0000,
+	0xf4051ebd, 0xfa000008,
+	0xbf8c0000, 0x87ee6e6e,
+	0xbf840001, 0xbe80206e,
+	0x876eff6d, 0x00ff0000,
+	0xbf850008, 0x876eff6d,
+	0x01000000, 0xbf850007,
+	0x8878ff78, 0x00002000,
+	0x80ec886c, 0x82ed806d,
+	0xbf820002, 0x806c846c,
 	0x826d806d, 0x876dff6d,
-	0x0000ffff, 0x906e8977,
-	0x876fff6e, 0x003f8000,
-	0x906e8677, 0x876eff6e,
-	0x02000000, 0x886e6f6e,
-	0xb9eef807, 0x87fe7e7e,
+	0x0000ffff, 0x907a8977,
+	0x877bff7a, 0x003f8000,
+	0x907a8677, 0x877aff7a,
+	0x02000000, 0x887a7b7a,
+	0xb9faf807, 0x87fe7e7e,
 	0x87ea6a6a, 0xb9f8f802,
 	0xbe80226c, 0x876dff6d,
 	0x0000ffff, 0xbefa0380,
-	0xb9fa0283, 0xb97a2c07,
-	0x8f7a9a7a, 0x886d7a6d,
-	0xb97a03c7, 0x8f7a997a,
-	0x886d7a6d, 0xb97a0647,
-	0x8f7a987a, 0x886d7a6d,
-	0xb97af807, 0x877aff7a,
-	0x00007fff, 0xb9faf807,
-	0xbeee037e, 0xbeef037f,
-	0xbefe0480, 0xbf900004,
-	0xbf8e0002, 0xbf88fffe,
-	0xb97b02dc, 0x8f7b997b,
-	0x887b7b7f, 0xb97a2a05,
+	0xb9fa0283, 0x8a77ff77,
+	0xff000000, 0xb97af807,
+	0x877bff7a, 0x02000000,
+	0x8f7b867b, 0x88777b77,
+	0x877bff7a, 0x003f8000,
+	0x8f7b897b, 0x88777b77,
+	0x8a7aff7a, 0x023f8000,
+	0xb9faf807, 0xbeee037e,
+	0xbeef037f, 0xbefe0480,
+	0xbf900004, 0xbf8e0002,
+	0xbf88fffe, 0x877aff7f,
+	0x04000000, 0x8f7a857a,
+	0x886d7a6d, 0xb97b02dc,
+	0x8f7b997b, 0xb97a3a05,
 	0x807a817a, 0xbf0d997b,
 	0xbf850002, 0x8f7a897a,
 	0xbf820001, 0x8f7a8a7a,
-	0x877bff7f, 0x0000ffff,
-	0x807aff7a, 0x00000200,
-	0x807a7e7a, 0x827b807b,
-	0xf4491c3d, 0xfa000050,
-	0xf4491d3d, 0xfa000060,
-	0xf4411e7d, 0xfa000074,
-	0xbef4037e, 0x8775ff7f,
-	0x0000ffff, 0x8875ff75,
-	0x00040000, 0xbef60380,
-	0xbef703ff, 0x10807fac,
-	0x877aff7f, 0x08000000,
-	0x907a837a, 0x88777a77,
-	0x877aff7f, 0x70000000,
-	0x907a817a, 0x88777a77,
-	0xbef1037c, 0xbef00380,
-	0xb97302dc, 0x8f739973,
-	0x8873737f, 0xb97bf816,
+	0xb97b1e06, 0x8f7b8a7b,
+	0x807a7b7a, 0x877bff7f,
+	0x0000ffff, 0x807aff7a,
+	0x00000200, 0x807a7e7a,
+	0x827b807b, 0xf4491c3d,
+	0xfa000050, 0xf4491d3d,
+	0xfa000060, 0xf4411e7d,
+	0xfa000074, 0xbef4037e,
+	0x8775ff7f, 0x0000ffff,
+	0x8875ff75, 0x00040000,
+	0xbef60380, 0xbef703ff,
+	0x10807fac, 0xbef1037c,
+	0xbef00380, 0xb97302dc,
+	0x8f739973, 0xb97bf816,
 	0xba80f816, 0x00000000,
 	0xbefe03c1, 0x907c9973,
 	0x877c817c, 0xbf06817c,
 	0xbf850002, 0xbeff0380,
 	0xbf820002, 0xbeff03c1,
-	0xbf82000b, 0xbef603ff,
-	0x01000000, 0xe0704000,
-	0x705d0000, 0xe0704080,
-	0x705d0100, 0xe0704100,
-	0x705d0200, 0xe0704180,
-	0x705d0300, 0xbf82000a,
-	0xbef603ff, 0x01000000,
+	0xbf820058, 0xbef603ff,
+	0x01000000, 0xb97af803,
+	0x8a7a7aff, 0x10000000,
+	0xbf850049, 0xbe840380,
+	0xd7600000, 0x00000900,
+	0x80048104, 0xd7600001,
+	0x00000900, 0x80048104,
+	0xd7600002, 0x00000900,
+	0x80048104, 0xd7600003,
+	0x00000900, 0x80048104,
+	0xf469003a, 0xe0000000,
+	0x80709070, 0xbf06a004,
+	0xbf84ffef, 0xbe840380,
+	0xd7600000, 0x00000901,
+	0x80048104, 0xd7600001,
+	0x00000901, 0x80048104,
+	0xd7600002, 0x00000901,
+	0x80048104, 0xd7600003,
+	0x00000901, 0x80048104,
+	0xf469003a, 0xe0000000,
+	0x80709070, 0xbf06a004,
+	0xbf84ffef, 0xbe840380,
+	0xd7600000, 0x00000902,
+	0x80048104, 0xd7600001,
+	0x00000902, 0x80048104,
+	0xd7600002, 0x00000902,
+	0x80048104, 0xd7600003,
+	0x00000902, 0x80048104,
+	0xf469003a, 0xe0000000,
+	0x80709070, 0xbf06a004,
+	0xbf84ffef, 0xbe840380,
+	0xd7600000, 0x00000903,
+	0x80048104, 0xd7600001,
+	0x00000903, 0x80048104,
+	0xd7600002, 0x00000903,
+	0x80048104, 0xd7600003,
+	0x00000903, 0x80048104,
+	0xf469003a, 0xe0000000,
+	0x80709070, 0xbf06a004,
+	0xbf84ffef, 0xbf820060,
+	0xe0704000, 0x705d0000,
+	0xe0704080, 0x705d0100,
+	0xe0704100, 0x705d0200,
+	0xe0704180, 0x705d0300,
+	0xbf820057, 0xbef603ff,
+	0x01000000, 0xb97af803,
+	0x8a7a7aff, 0x10000000,
+	0xbf850049, 0xbe840380,
+	0xd7600000, 0x00000900,
+	0x80048104, 0xd7600001,
+	0x00000900, 0x80048104,
+	0xd7600002, 0x00000900,
+	0x80048104, 0xd7600003,
+	0x00000900, 0x80048104,
+	0xf469003a, 0xe0000000,
+	0x80709070, 0xbf06c004,
+	0xbf84ffef, 0xbe840380,
+	0xd7600000, 0x00000901,
+	0x80048104, 0xd7600001,
+	0x00000901, 0x80048104,
+	0xd7600002, 0x00000901,
+	0x80048104, 0xd7600003,
+	0x00000901, 0x80048104,
+	0xf469003a, 0xe0000000,
+	0x80709070, 0xbf06c004,
+	0xbf84ffef, 0xbe840380,
+	0xd7600000, 0x00000902,
+	0x80048104, 0xd7600001,
+	0x00000902, 0x80048104,
+	0xd7600002, 0x00000902,
+	0x80048104, 0xd7600003,
+	0x00000902, 0x80048104,
+	0xf469003a, 0xe0000000,
+	0x80709070, 0xbf06c004,
+	0xbf84ffef, 0xbe840380,
+	0xd7600000, 0x00000903,
+	0x80048104, 0xd7600001,
+	0x00000903, 0x80048104,
+	0xd7600002, 0x00000903,
+	0x80048104, 0xd7600003,
+	0x00000903, 0x80048104,
+	0xf469003a, 0xe0000000,
+	0x80709070, 0xbf06c004,
+	0xbf84ffef, 0xbf820008,
 	0xe0704000, 0x705d0000,
 	0xe0704100, 0x705d0100,
 	0xe0704200, 0x705d0200,
 	0xe0704300, 0x705d0300,
-	0xb9702a05, 0x80708170,
+	0xb9703a05, 0x80708170,
 	0xbf0d9973, 0xbf850002,
 	0x8f708970, 0xbf820001,
 	0x8f708a70, 0xb97a1e06,
@@ -776,8 +871,9 @@ static const uint32_t cwsr_trap_nv1x_hex[] = {
 	0xbefe037c, 0xbefc0370,
 	0xf4611b3a, 0xf8000000,
 	0x80708470, 0xbefc037e,
+	0x8a7aff6d, 0x80000000,
 	0xbefe037c, 0xbefc0370,
-	0xf4611b7a, 0xf8000000,
+	0xf4611eba, 0xf8000000,
 	0x80708470, 0xbefc037e,
 	0xbefe037c, 0xbefc0370,
 	0xf4611bba, 0xf8000000,
@@ -805,7 +901,7 @@ static const uint32_t cwsr_trap_nv1x_hex[] = {
 	0xbefe037c, 0xbefc0370,
 	0xf4611c7a, 0xf8000000,
 	0x80708470, 0xbefc037e,
-	0xb9702a05, 0x80708170,
+	0xb9703a05, 0x80708170,
 	0xbf0d9973, 0xbf850002,
 	0x8f708970, 0xbf820001,
 	0x8f708a70, 0xb97a1e06,
@@ -837,99 +933,224 @@ static const uint32_t cwsr_trap_nv1x_hex[] = {
 	0xbf850002, 0xbeff0380,
 	0xbf820001, 0xbeff03c1,
 	0xb97b4306, 0x877bc17b,
-	0xbf840044, 0xbf8a0000,
-	0x877aff73, 0x04000000,
-	0xbf840040, 0x8f7b867b,
-	0x8f7b827b, 0xbef6037b,
-	0xb9702a05, 0x80708170,
-	0xbf0d9973, 0xbf850002,
-	0x8f708970, 0xbf820001,
-	0x8f708a70, 0xb97a1e06,
-	0x8f7a8a7a, 0x80707a70,
-	0x8070ff70, 0x00000200,
-	0x8070ff70, 0x00000080,
-	0xbef603ff, 0x01000000,
-	0xd7650000, 0x000100c1,
-	0xd7660000, 0x000200c1,
-	0x16000084, 0x907c9973,
-	0x877c817c, 0xbf06817c,
-	0xbefc0380, 0xbf850012,
-	0xbe8303ff, 0x00000080,
-	0xbf800000, 0xbf800000,
-	0xbf800000, 0xd8d80000,
+	0xbf840085, 0xbf8a0000,
+	0x877aff6d, 0x80000000,
+	0xbf840081, 0x8f7b887b,
+	0xbef6037b, 0xb9703a05,
+	0x80708170, 0xbf0d9973,
+	0xbf850002, 0x8f708970,
+	0xbf820001, 0x8f708a70,
+	0xb97a1e06, 0x8f7a8a7a,
+	0x80707a70, 0x8070ff70,
+	0x00000200, 0x8070ff70,
+	0x00000080, 0xbef603ff,
+	0x01000000, 0xd7650000,
+	0x000100c1, 0xd7660000,
+	0x000200c1, 0x16000084,
+	0x907c9973, 0x877c817c,
+	0xbf06817c, 0xbefc0380,
+	0xbf850033, 0xb97af803,
+	0x8a7a7aff, 0x10000000,
+	0xbf85001d, 0xd8d80000,
 	0x01000000, 0xbf8c0000,
-	0xe0704000, 0x705d0100,
-	0x807c037c, 0x80700370,
+	0xbe840380, 0xd7600000,
+	0x00000901, 0x80048104,
+	0xd7600001, 0x00000901,
+	0x80048104, 0xd7600002,
+	0x00000901, 0x80048104,
+	0xd7600003, 0x00000901,
+	0x80048104, 0xf469003a,
+	0xe0000000, 0x80709070,
+	0xbf06a004, 0xbf84ffef,
+	0x807cff7c, 0x00000080,
 	0xd5250000, 0x0001ff00,
 	0x00000080, 0xbf0a7b7c,
-	0xbf85fff4, 0xbf820011,
-	0xbe8303ff, 0x00000100,
+	0xbf85ffe4, 0xbf820044,
+	0xbe8303ff, 0x00000080,
 	0xbf800000, 0xbf800000,
 	0xbf800000, 0xd8d80000,
 	0x01000000, 0xbf8c0000,
 	0xe0704000, 0x705d0100,
 	0x807c037c, 0x80700370,
 	0xd5250000, 0x0001ff00,
-	0x00000100, 0xbf0a7b7c,
-	0xbf85fff4, 0xbefe03c1,
-	0x907c9973, 0x877c817c,
-	0xbf06817c, 0xbf850004,
-	0xbef003ff, 0x00000200,
-	0xbeff0380, 0xbf820003,
-	0xbef003ff, 0x00000400,
-	0xbeff03c1, 0xb97b2a05,
-	0x807b817b, 0x8f7b827b,
-	0x907c9973, 0x877c817c,
-	0xbf06817c, 0xbf850017,
+	0x00000080, 0xbf0a7b7c,
+	0xbf85fff4, 0xbf820032,
+	0xb97af803, 0x8a7a7aff,
+	0x10000000, 0xbf85001d,
+	0xd8d80000, 0x01000000,
+	0xbf8c0000, 0xbe840380,
+	0xd7600000, 0x00000901,
+	0x80048104, 0xd7600001,
+	0x00000901, 0x80048104,
+	0xd7600002, 0x00000901,
+	0x80048104, 0xd7600003,
+	0x00000901, 0x80048104,
+	0xf469003a, 0xe0000000,
+	0x80709070, 0xbf06c004,
+	0xbf84ffef, 0x807cff7c,
+	0x00000100, 0xd5250000,
+	0x0001ff00, 0x00000100,
+	0xbf0a7b7c, 0xbf85ffe4,
+	0xbf820011, 0xbe8303ff,
+	0x00000100, 0xbf800000,
+	0xbf800000, 0xbf800000,
+	0xd8d80000, 0x01000000,
+	0xbf8c0000, 0xe0704000,
+	0x705d0100, 0x807c037c,
+	0x80700370, 0xd5250000,
+	0x0001ff00, 0x00000100,
+	0xbf0a7b7c, 0xbf85fff4,
+	0xbefe03c1, 0x907c9973,
+	0x877c817c, 0xbf06817c,
+	0xbf850004, 0xbef003ff,
+	0x00000200, 0xbeff0380,
+	0xbf820003, 0xbef003ff,
+	0x00000400, 0xbeff03c1,
+	0xb97b3a05, 0x807b817b,
+	0x8f7b827b, 0x907c9973,
+	0x877c817c, 0xbf06817c,
+	0xbf85006b, 0xbef603ff,
+	0x01000000, 0xbefc0384,
+	0xbf0a7b7c, 0xbf8400fa,
+	0xb97af803, 0x8a7a7aff,
+	0x10000000, 0xbf850050,
+	0x7e008700, 0x7e028701,
+	0x7e048702, 0x7e068703,
+	0xbe840380, 0xd7600000,
+	0x00000900, 0x80048104,
+	0xd7600001, 0x00000900,
+	0x80048104, 0xd7600002,
+	0x00000900, 0x80048104,
+	0xd7600003, 0x00000900,
+	0x80048104, 0xf469003a,
+	0xe0000000, 0x80709070,
+	0xbf06a004, 0xbf84ffef,
+	0xbe840380, 0xd7600000,
+	0x00000901, 0x80048104,
+	0xd7600001, 0x00000901,
+	0x80048104, 0xd7600002,
+	0x00000901, 0x80048104,
+	0xd7600003, 0x00000901,
+	0x80048104, 0xf469003a,
+	0xe0000000, 0x80709070,
+	0xbf06a004, 0xbf84ffef,
+	0xbe840380, 0xd7600000,
+	0x00000902, 0x80048104,
+	0xd7600001, 0x00000902,
+	0x80048104, 0xd7600002,
+	0x00000902, 0x80048104,
+	0xd7600003, 0x00000902,
+	0x80048104, 0xf469003a,
+	0xe0000000, 0x80709070,
+	0xbf06a004, 0xbf84ffef,
+	0xbe840380, 0xd7600000,
+	0x00000903, 0x80048104,
+	0xd7600001, 0x00000903,
+	0x80048104, 0xd7600002,
+	0x00000903, 0x80048104,
+	0xd7600003, 0x00000903,
+	0x80048104, 0xf469003a,
+	0xe0000000, 0x80709070,
+	0xbf06a004, 0xbf84ffef,
+	0x807c847c, 0xbf0a7b7c,
+	0xbf85ffb1, 0xbf8200a6,
+	0x7e008700, 0x7e028701,
+	0x7e048702, 0x7e068703,
+	0xe0704000, 0x705d0000,
+	0xe0704080, 0x705d0100,
+	0xe0704100, 0x705d0200,
+	0xe0704180, 0x705d0300,
+	0x807c847c, 0x8070ff70,
+	0x00000200, 0xbf0a7b7c,
+	0xbf85ffef, 0xbf820094,
 	0xbef603ff, 0x01000000,
 	0xbefc0384, 0xbf0a7b7c,
-	0xbf840037, 0x7e008700,
+	0xbf840065, 0xb97af803,
+	0x8a7a7aff, 0x10000000,
+	0xbf850050, 0x7e008700,
+	0x7e028701, 0x7e048702,
+	0x7e068703, 0xbe840380,
+	0xd7600000, 0x00000900,
+	0x80048104, 0xd7600001,
+	0x00000900, 0x80048104,
+	0xd7600002, 0x00000900,
+	0x80048104, 0xd7600003,
+	0x00000900, 0x80048104,
+	0xf469003a, 0xe0000000,
+	0x80709070, 0xbf06c004,
+	0xbf84ffef, 0xbe840380,
+	0xd7600000, 0x00000901,
+	0x80048104, 0xd7600001,
+	0x00000901, 0x80048104,
+	0xd7600002, 0x00000901,
+	0x80048104, 0xd7600003,
+	0x00000901, 0x80048104,
+	0xf469003a, 0xe0000000,
+	0x80709070, 0xbf06c004,
+	0xbf84ffef, 0xbe840380,
+	0xd7600000, 0x00000902,
+	0x80048104, 0xd7600001,
+	0x00000902, 0x80048104,
+	0xd7600002, 0x00000902,
+	0x80048104, 0xd7600003,
+	0x00000902, 0x80048104,
+	0xf469003a, 0xe0000000,
+	0x80709070, 0xbf06c004,
+	0xbf84ffef, 0xbe840380,
+	0xd7600000, 0x00000903,
+	0x80048104, 0xd7600001,
+	0x00000903, 0x80048104,
+	0xd7600002, 0x00000903,
+	0x80048104, 0xd7600003,
+	0x00000903, 0x80048104,
+	0xf469003a, 0xe0000000,
+	0x80709070, 0xbf06c004,
+	0xbf84ffef, 0x807c847c,
+	0xbf0a7b7c, 0xbf85ffb1,
+	0xbf82003b, 0x7e008700,
 	0x7e028701, 0x7e048702,
 	0x7e068703, 0xe0704000,
-	0x705d0000, 0xe0704080,
-	0x705d0100, 0xe0704100,
-	0x705d0200, 0xe0704180,
+	0x705d0000, 0xe0704100,
+	0x705d0100, 0xe0704200,
+	0x705d0200, 0xe0704300,
 	0x705d0300, 0x807c847c,
-	0x8070ff70, 0x00000200,
+	0x8070ff70, 0x00000400,
 	0xbf0a7b7c, 0xbf85ffef,
-	0xbf820025, 0xbef603ff,
-	0x01000000, 0xbefc0384,
-	0xbf0a7b7c, 0xbf840020,
-	0x7e008700, 0x7e028701,
-	0x7e048702, 0x7e068703,
-	0xe0704000, 0x705d0000,
-	0xe0704100, 0x705d0100,
-	0xe0704200, 0x705d0200,
-	0xe0704300, 0x705d0300,
-	0x807c847c, 0x8070ff70,
-	0x00000400, 0xbf0a7b7c,
-	0xbf85ffef, 0xb97b1e06,
-	0x877bc17b, 0xbf84000c,
-	0x8f7b837b, 0x807b7c7b,
-	0xbefe03c1, 0xbeff0380,
+	0xb97b1e06, 0x877bc17b,
+	0xbf840027, 0x8f7b837b,
+	0x807b7c7b, 0xbefe03c1,
+	0xbeff0380, 0xb97af803,
+	0x8a7a7aff, 0x10000000,
+	0xbf850017, 0x7e008700,
+	0xbe840380, 0xd7600000,
+	0x00000900, 0x80048104,
+	0xd7600001, 0x00000900,
+	0x80048104, 0xd7600002,
+	0x00000900, 0x80048104,
+	0xd7600003, 0x00000900,
+	0x80048104, 0xf469003a,
+	0xe0000000, 0x80709070,
+	0xbf06c004, 0xbf84ffef,
+	0x807c817c, 0xbf0a7b7c,
+	0xbf85ffea, 0xbf820008,
 	0x7e008700, 0xe0704000,
 	0x705d0000, 0x807c817c,
 	0x8070ff70, 0x00000080,
 	0xbf0a7b7c, 0xbf85fff8,
-	0xbf820151, 0xbef4037e,
+	0xbf82013f, 0xbef4037e,
 	0x8775ff7f, 0x0000ffff,
 	0x8875ff75, 0x00040000,
 	0xbef60380, 0xbef703ff,
-	0x10807fac, 0x876eff7f,
-	0x08000000, 0x906e836e,
-	0x88776e77, 0x876eff7f,
-	0x70000000, 0x906e816e,
-	0x88776e77, 0xb97202dc,
-	0x8f729972, 0x8872727f,
-	0x876eff7f, 0x04000000,
-	0xbf840034, 0xbefe03c1,
-	0x907c9972, 0x877c817c,
-	0xbf06817c, 0xbf850002,
-	0xbeff0380, 0xbf820001,
-	0xbeff03c1, 0xb96f4306,
-	0x876fc16f, 0xbf840029,
-	0x8f6f866f, 0x8f6f826f,
-	0xbef6036f, 0xb9782a05,
+	0x10807fac, 0xb97202dc,
+	0x8f729972, 0x876eff7f,
+	0x04000000, 0xbf840033,
+	0xbefe03c1, 0x907c9972,
+	0x877c817c, 0xbf06817c,
+	0xbf850002, 0xbeff0380,
+	0xbf820001, 0xbeff03c1,
+	0xb96f4306, 0x876fc16f,
+	0xbf840028, 0x8f6f886f,
+	0xbef6036f, 0xb9783a05,
 	0x80788178, 0xbf0d9972,
 	0xbf850002, 0x8f788978,
 	0xbf820001, 0x8f788a78,
@@ -953,17 +1174,18 @@ static const uint32_t cwsr_trap_nv1x_hex[] = {
 	0x877c817c, 0xbf06817c,
 	0xbf850002, 0xbeff0380,
 	0xbf820001, 0xbeff03c1,
-	0xb96f2a05, 0x806f816f,
+	0xb96f3a05, 0x806f816f,
 	0x8f6f826f, 0x907c9972,
 	0x877c817c, 0xbf06817c,
-	0xbf850021, 0xbef603ff,
+	0xbf850024, 0xbef603ff,
 	0x01000000, 0xbeee0378,
 	0x8078ff78, 0x00000200,
-	0xbefc0384, 0xe0304000,
+	0xbefc0384, 0xbf0a6f7c,
+	0xbf840050, 0xe0304000,
 	0x785d0000, 0xe0304080,
 	0x785d0100, 0xe0304100,
 	0x785d0200, 0xe0304180,
-	0x785d0300, 0xbf8c3f70,
+	0x785d0300, 0xbf8c0000,
 	0x7e008500, 0x7e028501,
 	0x7e048502, 0x7e068503,
 	0x807c847c, 0x8078ff78,
@@ -972,229 +1194,233 @@ static const uint32_t cwsr_trap_nv1x_hex[] = {
 	0x6e5d0000, 0xe0304080,
 	0x6e5d0100, 0xe0304100,
 	0x6e5d0200, 0xe0304180,
-	0x6e5d0300, 0xbf820032,
-	0xbef603ff, 0x01000000,
-	0xbeee0378, 0x8078ff78,
-	0x00000400, 0xbefc0384,
-	0xe0304000, 0x785d0000,
-	0xe0304100, 0x785d0100,
-	0xe0304200, 0x785d0200,
-	0xe0304300, 0x785d0300,
-	0xbf8c3f70, 0x7e008500,
-	0x7e028501, 0x7e048502,
-	0x7e068503, 0x807c847c,
+	0x6e5d0300, 0xbf8c0000,
+	0xbf820034, 0xbef603ff,
+	0x01000000, 0xbeee0378,
 	0x8078ff78, 0x00000400,
-	0xbf0a6f7c, 0xbf85ffee,
-	0xb96f1e06, 0x876fc16f,
-	0xbf84000e, 0x8f6f836f,
-	0x806f7c6f, 0xbefe03c1,
-	0xbeff0380, 0xe0304000,
-	0x785d0000, 0xbf8c3f70,
-	0x7e008500, 0x807c817c,
-	0x8078ff78, 0x00000080,
-	0xbf0a6f7c, 0xbf85fff7,
-	0xbeff03c1, 0xe0304000,
-	0x6e5d0000, 0xe0304100,
-	0x6e5d0100, 0xe0304200,
-	0x6e5d0200, 0xe0304300,
-	0x6e5d0300, 0xbf8c3f70,
-	0xb9782a05, 0x80788178,
+	0xbefc0384, 0xbf0a6f7c,
+	0xbf840012, 0xe0304000,
+	0x785d0000, 0xe0304100,
+	0x785d0100, 0xe0304200,
+	0x785d0200, 0xe0304300,
+	0x785d0300, 0xbf8c0000,
+	0x7e008500, 0x7e028501,
+	0x7e048502, 0x7e068503,
+	0x807c847c, 0x8078ff78,
+	0x00000400, 0xbf0a6f7c,
+	0xbf85ffee, 0xb96f1e06,
+	0x876fc16f, 0xbf84000e,
+	0x8f6f836f, 0x806f7c6f,
+	0xbefe03c1, 0xbeff0380,
+	0xe0304000, 0x785d0000,
+	0xbf8c0000, 0x7e008500,
+	0x807c817c, 0x8078ff78,
+	0x00000080, 0xbf0a6f7c,
+	0xbf85fff7, 0xbeff03c1,
+	0xe0304000, 0x6e5d0000,
+	0xe0304100, 0x6e5d0100,
+	0xe0304200, 0x6e5d0200,
+	0xe0304300, 0x6e5d0300,
+	0xbf8c0000, 0xb9783a05,
+	0x80788178, 0xbf0d9972,
+	0xbf850002, 0x8f788978,
+	0xbf820001, 0x8f788a78,
+	0xb96e1e06, 0x8f6e8a6e,
+	0x80786e78, 0x8078ff78,
+	0x00000200, 0x80f8ff78,
+	0x00000050, 0xbef603ff,
+	0x01000000, 0xbefc03ff,
+	0x0000006c, 0x80f89078,
+	0xf429003a, 0xf0000000,
+	0xbf8c0000, 0x80fc847c,
+	0xbf800000, 0xbe803100,
+	0xbe823102, 0x80f8a078,
+	0xf42d003a, 0xf0000000,
+	0xbf8c0000, 0x80fc887c,
+	0xbf800000, 0xbe803100,
+	0xbe823102, 0xbe843104,
+	0xbe863106, 0x80f8c078,
+	0xf431003a, 0xf0000000,
+	0xbf8c0000, 0x80fc907c,
+	0xbf800000, 0xbe803100,
+	0xbe823102, 0xbe843104,
+	0xbe863106, 0xbe883108,
+	0xbe8a310a, 0xbe8c310c,
+	0xbe8e310e, 0xbf06807c,
+	0xbf84fff0, 0xba80f801,
+	0x00000000, 0xbf8a0000,
+	0xb9783a05, 0x80788178,
 	0xbf0d9972, 0xbf850002,
 	0x8f788978, 0xbf820001,
 	0x8f788a78, 0xb96e1e06,
 	0x8f6e8a6e, 0x80786e78,
 	0x8078ff78, 0x00000200,
-	0x80f8ff78, 0x00000050,
 	0xbef603ff, 0x01000000,
-	0xbefc03ff, 0x0000006c,
-	0x80f89078, 0xf429003a,
-	0xf0000000, 0xbf8cc07f,
-	0x80fc847c, 0xbf800000,
-	0xbe803100, 0xbe823102,
-	0x80f8a078, 0xf42d003a,
-	0xf0000000, 0xbf8cc07f,
-	0x80fc887c, 0xbf800000,
-	0xbe803100, 0xbe823102,
-	0xbe843104, 0xbe863106,
-	0x80f8c078, 0xf431003a,
-	0xf0000000, 0xbf8cc07f,
-	0x80fc907c, 0xbf800000,
-	0xbe803100, 0xbe823102,
-	0xbe843104, 0xbe863106,
-	0xbe883108, 0xbe8a310a,
-	0xbe8c310c, 0xbe8e310e,
-	0xbf06807c, 0xbf84fff0,
-	0xba80f801, 0x00000000,
-	0xbf8a0000, 0xb9782a05,
-	0x80788178, 0xbf0d9972,
-	0xbf850002, 0x8f788978,
-	0xbf820001, 0x8f788a78,
-	0xb96e1e06, 0x8f6e8a6e,
-	0x80786e78, 0x8078ff78,
-	0x00000200, 0xbef603ff,
-	0x01000000, 0xf4211bfa,
+	0xf4211bfa, 0xf0000000,
+	0x80788478, 0xf4211b3a,
 	0xf0000000, 0x80788478,
-	0xf4211b3a, 0xf0000000,
-	0x80788478, 0xf4211b7a,
+	0xf4211b7a, 0xf0000000,
+	0x80788478, 0xf4211c3a,
 	0xf0000000, 0x80788478,
-	0xf4211c3a, 0xf0000000,
-	0x80788478, 0xf4211c7a,
+	0xf4211c7a, 0xf0000000,
+	0x80788478, 0xf4211eba,
 	0xf0000000, 0x80788478,
-	0xf4211eba, 0xf0000000,
-	0x80788478, 0xf4211efa,
+	0xf4211efa, 0xf0000000,
+	0x80788478, 0xf4211e7a,
 	0xf0000000, 0x80788478,
-	0xf4211e7a, 0xf0000000,
-	0x80788478, 0xf4211cfa,
+	0xf4211cfa, 0xf0000000,
+	0x80788478, 0xf4211bba,
 	0xf0000000, 0x80788478,
+	0xbf8c0000, 0xb9eef814,
 	0xf4211bba, 0xf0000000,
-	0x80788478, 0xbf8cc07f,
-	0xb9eef814, 0xf4211bba,
-	0xf0000000, 0x80788478,
-	0xbf8cc07f, 0xb9eef815,
-	0xbefc036f, 0xbefe0370,
-	0xbeff0371, 0x876f7bff,
-	0x000003ff, 0xb9ef4803,
-	0xb9f9f816, 0x876f7bff,
-	0xfffff800, 0x906f8b6f,
-	0xb9efa2c3, 0xb9f3f801,
-	0xb96e2a05, 0x806e816e,
-	0xbf0d9972, 0xbf850002,
-	0x8f6e896e, 0xbf820001,
-	0x8f6e8a6e, 0x806eff6e,
+	0x80788478, 0xbf8c0000,
+	0xb9eef815, 0xbefc036f,
+	0xbefe0370, 0xbeff0371,
+	0xb9f9f816, 0xb9fb4803,
+	0x907b8b7b, 0xb9fba2c3,
+	0xb9f3f801, 0xb96e3a05,
+	0x806e816e, 0xbf0d9972,
+	0xbf850002, 0x8f6e896e,
+	0xbf820001, 0x8f6e8a6e,
+	0xb96f1e06, 0x8f6f8a6f,
+	0x806e6f6e, 0x806eff6e,
 	0x00000200, 0x806e746e,
 	0x826f8075, 0x876fff6f,
 	0x0000ffff, 0xf4091c37,
 	0xfa000050, 0xf4091d37,
 	0xfa000060, 0xf4011e77,
-	0xfa000074, 0xbf8cc07f,
-	0x876fff6d, 0xfc000000,
-	0x906f9a6f, 0x8f6f906f,
-	0xbeee0380, 0x886e6f6e,
-	0x876fff6d, 0x02000000,
-	0x906f996f, 0x8f6f8f6f,
-	0x886e6f6e, 0x876fff6d,
-	0x01000000, 0x906f986f,
-	0x8f6f996f, 0x886e6f6e,
-	0x876fff7a, 0x00800000,
-	0x906f976f, 0xb9eef807,
+	0xfa000074, 0xbf8c0000,
+	0x906e8977, 0x876fff6e,
+	0x003f8000, 0x906e8677,
+	0x876eff6e, 0x02000000,
+	0x886e6f6e, 0xb9eef807,
 	0x876dff6d, 0x0000ffff,
 	0x87fe7e7e, 0x87ea6a6a,
 	0xb9faf802, 0xbe80226c,
-	0xbf810000, 0xbf9f0000,
+	0xbf9b0000, 0xbf9f0000,
 	0xbf9f0000, 0xbf9f0000,
 	0xbf9f0000, 0xbf9f0000,
 };
 
 static const uint32_t cwsr_trap_arcturus_hex[] = {
-	0xbf820001, 0xbf8202c4,
-	0xb8f8f802, 0x89788678,
-	0xb8eef801, 0x866eff6e,
-	0x00000800, 0xbf840003,
+	0xbf820001, 0xbf8202d5,
+	0xb8f8f802, 0x8978ff78,
+	0x00020006, 0xb8fbf803,
 	0x866eff78, 0x00002000,
-	0xbf840016, 0xb8fbf803,
+	0xbf840009, 0x866eff6d,
+	0x00ff0000, 0xbf85001e,
 	0x866eff7b, 0x00000400,
-	0xbf85003b, 0x866eff7b,
-	0x00000800, 0xbf850003,
-	0x866eff7b, 0x00000100,
-	0xbf84000c, 0x866eff78,
-	0x00002000, 0xbf840005,
-	0xbf8e0010, 0xb8eef803,
-	0x866eff6e, 0x00000400,
-	0xbf84fffb, 0x8778ff78,
-	0x00002000, 0x80ec886c,
-	0x82ed806d, 0xb8eef807,
-	0x866fff6e, 0x001f8000,
-	0x8e6f8b6f, 0x8977ff77,
-	0xfc000000, 0x87776f77,
-	0x896eff6e, 0x001f8000,
-	0xb96ef807, 0xb8faf812,
+	0xbf850055, 0xbf8e0010,
+	0xb8fbf803, 0xbf82fffa,
+	0x866eff7b, 0x03c00900,
+	0xbf850015, 0x866eff7b,
+	0x000071ff, 0xbf840008,
+	0x866fff7b, 0x00007080,
+	0xbf840001, 0xbeee1a87,
+	0xb8eff801, 0x8e6e8c6e,
+	0x866e6f6e, 0xbf85000a,
+	0x866eff6d, 0x00ff0000,
+	0xbf850007, 0xb8eef801,
+	0x866eff6e, 0x00000800,
+	0xbf850003, 0x866eff7b,
+	0x00000400, 0xbf85003a,
+	0xb8faf807, 0x867aff7a,
+	0x001f8000, 0x8e7a8b7a,
+	0x8977ff77, 0xfc000000,
+	0x87777a77, 0xba7ff807,
+	0x00000000, 0xb8faf812,
 	0xb8fbf813, 0x8efa887a,
-	0xc0071bbd, 0x00000000,
-	0xbf8cc07f, 0xc0071ebd,
-	0x00000008, 0xbf8cc07f,
-	0x86ee6e6e, 0xbf840001,
-	0xbe801d6e, 0xb8fbf803,
-	0x867bff7b, 0x000001ff,
+	0xbf0d8f7b, 0xbf840002,
+	0x877bff7b, 0xffff0000,
+	0xc0031bbd, 0x00000010,
+	0xbf8cc07f, 0x8e6e976e,
+	0x8977ff77, 0x00800000,
+	0x87776e77, 0xc0071bbd,
+	0x00000000, 0xbf8cc07f,
+	0xc0071ebd, 0x00000008,
+	0xbf8cc07f, 0x86ee6e6e,
+	0xbf840001, 0xbe801d6e,
+	0x866eff6d, 0x01ff0000,
+	0xbf850005, 0x8778ff78,
+	0x00002000, 0x80ec886c,
+	0x82ed806d, 0xbf820005,
+	0x866eff6d, 0x01000000,
 	0xbf850002, 0x806c846c,
 	0x826d806d, 0x866dff6d,
-	0x0000ffff, 0x8f6e8b77,
-	0x866eff6e, 0x001f8000,
-	0xb96ef807, 0x86fe7e7e,
+	0x0000ffff, 0x8f7a8b77,
+	0x867aff7a, 0x001f8000,
+	0xb97af807, 0x86fe7e7e,
 	0x86ea6a6a, 0x8f6e8378,
 	0xb96ee0c2, 0xbf800002,
 	0xb9780002, 0xbe801f6c,
 	0x866dff6d, 0x0000ffff,
 	0xbefa0080, 0xb97a0283,
-	0xb8fa2407, 0x8e7a9b7a,
-	0x876d7a6d, 0xb8fa03c7,
-	0x8e7a9a7a, 0x876d7a6d,
 	0xb8faf807, 0x867aff7a,
-	0x00007fff, 0xb97af807,
-	0xbeee007e, 0xbeef007f,
-	0xbefe0180, 0xbf900004,
-	0x877a8478, 0xb97af802,
-	0xbf8e0002, 0xbf88fffe,
-	0xb8fa2a05, 0x807a817a,
-	0x8e7a8a7a, 0x8e7a817a,
-	0xb8fb1605, 0x807b817b,
-	0x8e7b867b, 0x807a7b7a,
-	0x807a7e7a, 0x827b807f,
-	0x867bff7b, 0x0000ffff,
-	0xc04b1c3d, 0x00000050,
-	0xbf8cc07f, 0xc04b1d3d,
-	0x00000060, 0xbf8cc07f,
-	0xc0431e7d, 0x00000074,
-	0xbf8cc07f, 0xbef4007e,
-	0x8675ff7f, 0x0000ffff,
-	0x8775ff75, 0x00040000,
-	0xbef60080, 0xbef700ff,
-	0x00807fac, 0x867aff7f,
-	0x08000000, 0x8f7a837a,
-	0x87777a77, 0x867aff7f,
-	0x70000000, 0x8f7a817a,
-	0x87777a77, 0xbef1007c,
-	0xbef00080, 0xb8f02a05,
-	0x80708170, 0x8e708a70,
-	0x8e708170, 0xb8fa1605,
-	0x807a817a, 0x8e7a867a,
-	0x80707a70, 0xbef60084,
-	0xbef600ff, 0x01000000,
-	0xbefe007c, 0xbefc0070,
-	0xc0611c7a, 0x0000007c,
-	0xbf8cc07f, 0x80708470,
-	0xbefc007e, 0xbefe007c,
-	0xbefc0070, 0xc0611b3a,
+	0x001f8000, 0x8e7a8b7a,
+	0x8977ff77, 0xfc000000,
+	0x87777a77, 0xba7ff807,
+	0x00000000, 0xbeee007e,
+	0xbeef007f, 0xbefe0180,
+	0xbf900004, 0x877a8478,
+	0xb97af802, 0xbf8e0002,
+	0xbf88fffe, 0xb8fa2a05,
+	0x807a817a, 0x8e7a8a7a,
+	0x8e7a817a, 0xb8fb1605,
+	0x807b817b, 0x8e7b867b,
+	0x807a7b7a, 0x807a7e7a,
+	0x827b807f, 0x867bff7b,
+	0x0000ffff, 0xc04b1c3d,
+	0x00000050, 0xbf8cc07f,
+	0xc04b1d3d, 0x00000060,
+	0xbf8cc07f, 0xc0431e7d,
+	0x00000074, 0xbf8cc07f,
+	0xbef4007e, 0x8675ff7f,
+	0x0000ffff, 0x8775ff75,
+	0x00040000, 0xbef60080,
+	0xbef700ff, 0x00807fac,
+	0xbef1007c, 0xbef00080,
+	0xb8f02a05, 0x80708170,
+	0x8e708a70, 0x8e708170,
+	0xb8fa1605, 0x807a817a,
+	0x8e7a867a, 0x80707a70,
+	0xbef60084, 0xbef600ff,
+	0x01000000, 0xbefe007c,
+	0xbefc0070, 0xc0611c7a,
 	0x0000007c, 0xbf8cc07f,
 	0x80708470, 0xbefc007e,
 	0xbefe007c, 0xbefc0070,
-	0xc0611b7a, 0x0000007c,
+	0xc0611b3a, 0x0000007c,
 	0xbf8cc07f, 0x80708470,
 	0xbefc007e, 0xbefe007c,
-	0xbefc0070, 0xc0611bba,
+	0xbefc0070, 0xc0611b7a,
 	0x0000007c, 0xbf8cc07f,
 	0x80708470, 0xbefc007e,
 	0xbefe007c, 0xbefc0070,
-	0xc0611bfa, 0x0000007c,
+	0xc0611bba, 0x0000007c,
 	0xbf8cc07f, 0x80708470,
 	0xbefc007e, 0xbefe007c,
-	0xbefc0070, 0xc0611e3a,
-	0x0000007c, 0xbf8cc07f,
-	0x80708470, 0xbefc007e,
-	0xb8fbf803, 0xbefe007c,
-	0xbefc0070, 0xc0611efa,
+	0xbefc0070, 0xc0611bfa,
 	0x0000007c, 0xbf8cc07f,
 	0x80708470, 0xbefc007e,
 	0xbefe007c, 0xbefc0070,
-	0xc0611a3a, 0x0000007c,
+	0xc0611e3a, 0x0000007c,
+	0xbf8cc07f, 0x80708470,
+	0xbefc007e, 0xb8fbf803,
+	0xbefe007c, 0xbefc0070,
+	0xc0611efa, 0x0000007c,
 	0xbf8cc07f, 0x80708470,
 	0xbefc007e, 0xbefe007c,
-	0xbefc0070, 0xc0611a7a,
-	0x0000007c, 0xbf8cc07f,
-	0x80708470, 0xbefc007e,
-	0xb8f1f801, 0xbefe007c,
-	0xbefc0070, 0xc0611c7a,
+	0xbefc0070, 0xc0611a3a,
 	0x0000007c, 0xbf8cc07f,
 	0x80708470, 0xbefc007e,
+	0xbefe007c, 0xbefc0070,
+	0xc0611a7a, 0x0000007c,
+	0xbf8cc07f, 0x80708470,
+	0xbefc007e, 0xb8f1f801,
+	0xbefe007c, 0xbefc0070,
+	0xc0611c7a, 0x0000007c,
+	0xbf8cc07f, 0x80708470,
+	0xbefc007e, 0xbf108080,
 	0x867aff7f, 0x04000000,
 	0xbeef0080, 0x876f6f7a,
 	0xb8f02a05, 0x80708170,
@@ -1323,368 +1549,359 @@ static const uint32_t cwsr_trap_arcturus_hex[] = {
 	0x00000400, 0xbefe00c1,
 	0xbeff00c1, 0xb8fb2a05,
 	0x807b817b, 0x8e7b827b,
-	0x8e76887b, 0xbef600ff,
-	0x01000000, 0xbefc0084,
-	0xbf0a7b7c, 0xbf84006d,
-	0xbf11017c, 0x807bff7b,
-	0x00001000, 0x867aff78,
-	0x00400000, 0xbf850003,
-	0xb8faf803, 0x897a7aff,
-	0x10000000, 0xbf850051,
-	0xbe840080, 0xd2890000,
-	0x00000900, 0x80048104,
-	0xd2890001, 0x00000900,
-	0x80048104, 0xd2890002,
-	0x00000900, 0x80048104,
-	0xd2890003, 0x00000900,
-	0x80048104, 0xc069003a,
-	0x00000070, 0xbf8cc07f,
-	0x80709070, 0xbf06c004,
-	0xbf84ffee, 0xbe840080,
-	0xd2890000, 0x00000901,
+	0xbef600ff, 0x01000000,
+	0xbefc0084, 0xbf0a7b7c,
+	0xbf84006d, 0xbf11017c,
+	0x807bff7b, 0x00001000,
+	0x867aff78, 0x00400000,
+	0xbf850003, 0xb8faf803,
+	0x897a7aff, 0x10000000,
+	0xbf850051, 0xbe840080,
+	0xd2890000, 0x00000900,
 	0x80048104, 0xd2890001,
-	0x00000901, 0x80048104,
-	0xd2890002, 0x00000901,
+	0x00000900, 0x80048104,
+	0xd2890002, 0x00000900,
 	0x80048104, 0xd2890003,
-	0x00000901, 0x80048104,
+	0x00000900, 0x80048104,
 	0xc069003a, 0x00000070,
 	0xbf8cc07f, 0x80709070,
 	0xbf06c004, 0xbf84ffee,
 	0xbe840080, 0xd2890000,
-	0x00000902, 0x80048104,
-	0xd2890001, 0x00000902,
+	0x00000901, 0x80048104,
+	0xd2890001, 0x00000901,
 	0x80048104, 0xd2890002,
-	0x00000902, 0x80048104,
-	0xd2890003, 0x00000902,
+	0x00000901, 0x80048104,
+	0xd2890003, 0x00000901,
 	0x80048104, 0xc069003a,
 	0x00000070, 0xbf8cc07f,
 	0x80709070, 0xbf06c004,
 	0xbf84ffee, 0xbe840080,
-	0xd2890000, 0x00000903,
+	0xd2890000, 0x00000902,
 	0x80048104, 0xd2890001,
-	0x00000903, 0x80048104,
-	0xd2890002, 0x00000903,
+	0x00000902, 0x80048104,
+	0xd2890002, 0x00000902,
 	0x80048104, 0xd2890003,
-	0x00000903, 0x80048104,
+	0x00000902, 0x80048104,
 	0xc069003a, 0x00000070,
 	0xbf8cc07f, 0x80709070,
 	0xbf06c004, 0xbf84ffee,
-	0x807c847c, 0xbf0a7b7c,
-	0xbf85ffb1, 0xbf9c0000,
-	0xbf820012, 0x7e000300,
-	0x7e020301, 0x7e040302,
-	0x7e060303, 0xe0724000,
-	0x701d0000, 0xe0724100,
-	0x701d0100, 0xe0724200,
-	0x701d0200, 0xe0724300,
-	0x701d0300, 0x807c847c,
-	0x8070ff70, 0x00000400,
-	0xbf0a7b7c, 0xbf85ffef,
-	0xbf9c0000, 0xbefc0080,
-	0xbf11017c, 0x867aff78,
-	0x00400000, 0xbf850003,
-	0xb8faf803, 0x897a7aff,
-	0x10000000, 0xbf850059,
-	0xd3d84000, 0x18000100,
-	0xd3d84001, 0x18000101,
-	0xd3d84002, 0x18000102,
-	0xd3d84003, 0x18000103,
 	0xbe840080, 0xd2890000,
-	0x00000900, 0x80048104,
-	0xd2890001, 0x00000900,
+	0x00000903, 0x80048104,
+	0xd2890001, 0x00000903,
 	0x80048104, 0xd2890002,
-	0x00000900, 0x80048104,
-	0xd2890003, 0x00000900,
+	0x00000903, 0x80048104,
+	0xd2890003, 0x00000903,
 	0x80048104, 0xc069003a,
 	0x00000070, 0xbf8cc07f,
 	0x80709070, 0xbf06c004,
-	0xbf84ffee, 0xbe840080,
-	0xd2890000, 0x00000901,
+	0xbf84ffee, 0x807c847c,
+	0xbf0a7b7c, 0xbf85ffb1,
+	0xbf9c0000, 0xbf820012,
+	0x7e000300, 0x7e020301,
+	0x7e040302, 0x7e060303,
+	0xe0724000, 0x701d0000,
+	0xe0724100, 0x701d0100,
+	0xe0724200, 0x701d0200,
+	0xe0724300, 0x701d0300,
+	0x807c847c, 0x8070ff70,
+	0x00000400, 0xbf0a7b7c,
+	0xbf85ffef, 0xbf9c0000,
+	0xbefc0080, 0xbf11017c,
+	0x867aff78, 0x00400000,
+	0xbf850003, 0xb8faf803,
+	0x897a7aff, 0x10000000,
+	0xbf850059, 0xd3d84000,
+	0x18000100, 0xd3d84001,
+	0x18000101, 0xd3d84002,
+	0x18000102, 0xd3d84003,
+	0x18000103, 0xbe840080,
+	0xd2890000, 0x00000900,
 	0x80048104, 0xd2890001,
-	0x00000901, 0x80048104,
-	0xd2890002, 0x00000901,
+	0x00000900, 0x80048104,
+	0xd2890002, 0x00000900,
 	0x80048104, 0xd2890003,
-	0x00000901, 0x80048104,
+	0x00000900, 0x80048104,
 	0xc069003a, 0x00000070,
 	0xbf8cc07f, 0x80709070,
 	0xbf06c004, 0xbf84ffee,
 	0xbe840080, 0xd2890000,
-	0x00000902, 0x80048104,
-	0xd2890001, 0x00000902,
+	0x00000901, 0x80048104,
+	0xd2890001, 0x00000901,
 	0x80048104, 0xd2890002,
-	0x00000902, 0x80048104,
-	0xd2890003, 0x00000902,
+	0x00000901, 0x80048104,
+	0xd2890003, 0x00000901,
 	0x80048104, 0xc069003a,
 	0x00000070, 0xbf8cc07f,
 	0x80709070, 0xbf06c004,
 	0xbf84ffee, 0xbe840080,
-	0xd2890000, 0x00000903,
+	0xd2890000, 0x00000902,
 	0x80048104, 0xd2890001,
-	0x00000903, 0x80048104,
-	0xd2890002, 0x00000903,
+	0x00000902, 0x80048104,
+	0xd2890002, 0x00000902,
 	0x80048104, 0xd2890003,
-	0x00000903, 0x80048104,
+	0x00000902, 0x80048104,
 	0xc069003a, 0x00000070,
 	0xbf8cc07f, 0x80709070,
 	0xbf06c004, 0xbf84ffee,
-	0x807c847c, 0xbf0a7b7c,
-	0xbf85ffa9, 0xbf9c0000,
-	0xbf820016, 0xd3d84000,
-	0x18000100, 0xd3d84001,
-	0x18000101, 0xd3d84002,
-	0x18000102, 0xd3d84003,
-	0x18000103, 0xe0724000,
-	0x701d0000, 0xe0724100,
-	0x701d0100, 0xe0724200,
-	0x701d0200, 0xe0724300,
-	0x701d0300, 0x807c847c,
-	0x8070ff70, 0x00000400,
-	0xbf0a7b7c, 0xbf85ffeb,
-	0xbf9c0000, 0xbf820106,
-	0xbef4007e, 0x8675ff7f,
-	0x0000ffff, 0x8775ff75,
-	0x00040000, 0xbef60080,
-	0xbef700ff, 0x00807fac,
-	0x866eff7f, 0x08000000,
-	0x8f6e836e, 0x87776e77,
-	0x866eff7f, 0x70000000,
-	0x8f6e816e, 0x87776e77,
-	0x866eff7f, 0x04000000,
-	0xbf84001f, 0xbefe00c1,
-	0xbeff00c1, 0xb8ef4306,
-	0x866fc16f, 0xbf84001a,
-	0x8e6f866f, 0x8e6f826f,
-	0xbef6006f, 0xb8f82a05,
-	0x80788178, 0x8e788a78,
-	0x8e788178, 0xb8ee1605,
-	0x806e816e, 0x8e6e866e,
-	0x80786e78, 0x8078ff78,
-	0x00000080, 0xbef600ff,
-	0x01000000, 0xbefc0080,
-	0xe0510000, 0x781d0000,
-	0xe0510100, 0x781d0000,
-	0x807cff7c, 0x00000200,
-	0x8078ff78, 0x00000200,
-	0xbf0a6f7c, 0xbf85fff6,
-	0xbef80080, 0xbefe00c1,
-	0xbeff00c1, 0xb8ef2a05,
-	0x806f816f, 0x8e6f826f,
-	0x8e76886f, 0xbef90076,
+	0xbe840080, 0xd2890000,
+	0x00000903, 0x80048104,
+	0xd2890001, 0x00000903,
+	0x80048104, 0xd2890002,
+	0x00000903, 0x80048104,
+	0xd2890003, 0x00000903,
+	0x80048104, 0xc069003a,
+	0x00000070, 0xbf8cc07f,
+	0x80709070, 0xbf06c004,
+	0xbf84ffee, 0x807c847c,
+	0xbf0a7b7c, 0xbf85ffa9,
+	0xbf9c0000, 0xbf820016,
+	0xd3d84000, 0x18000100,
+	0xd3d84001, 0x18000101,
+	0xd3d84002, 0x18000102,
+	0xd3d84003, 0x18000103,
+	0xe0724000, 0x701d0000,
+	0xe0724100, 0x701d0100,
+	0xe0724200, 0x701d0200,
+	0xe0724300, 0x701d0300,
+	0x807c847c, 0x8070ff70,
+	0x00000400, 0xbf0a7b7c,
+	0xbf85ffeb, 0xbf9c0000,
+	0xbf8200e3, 0xbef4007e,
+	0x8675ff7f, 0x0000ffff,
+	0x8775ff75, 0x00040000,
+	0xbef60080, 0xbef700ff,
+	0x00807fac, 0x866eff7f,
+	0x04000000, 0xbf84001f,
+	0xbefe00c1, 0xbeff00c1,
+	0xb8ef4306, 0x866fc16f,
+	0xbf84001a, 0x8e6f866f,
+	0x8e6f826f, 0xbef6006f,
+	0xb8f82a05, 0x80788178,
+	0x8e788a78, 0x8e788178,
+	0xb8ee1605, 0x806e816e,
+	0x8e6e866e, 0x80786e78,
+	0x8078ff78, 0x00000080,
 	0xbef600ff, 0x01000000,
-	0xbeee0078, 0x8078ff78,
-	0x00000400, 0xbef30079,
-	0x8079ff79, 0x00000400,
-	0xbefc0084, 0xbf11087c,
+	0xbefc0080, 0xe0510000,
+	0x781d0000, 0xe0510100,
+	0x781d0000, 0x807cff7c,
+	0x00000200, 0x8078ff78,
+	0x00000200, 0xbf0a6f7c,
+	0xbf85fff6, 0xbefe00c1,
+	0xbeff00c1, 0xbef600ff,
+	0x01000000, 0xb8ef2a05,
+	0x806f816f, 0x8e6f826f,
 	0x806fff6f, 0x00008000,
-	0xe0524000, 0x791d0000,
-	0xe0524100, 0x791d0100,
-	0xe0524200, 0x791d0200,
-	0xe0524300, 0x791d0300,
-	0x8079ff79, 0x00000400,
-	0xbf8c0f70, 0xd3d94000,
-	0x18000100, 0xd3d94001,
-	0x18000101, 0xd3d94002,
-	0x18000102, 0xd3d94003,
-	0x18000103, 0xe0524000,
-	0x781d0000, 0xe0524100,
-	0x781d0100, 0xe0524200,
-	0x781d0200, 0xe0524300,
-	0x781d0300, 0xbf8c0f70,
-	0x7e000300, 0x7e020301,
-	0x7e040302, 0x7e060303,
-	0x807c847c, 0x8078ff78,
-	0x00000400, 0xbf0a6f7c,
-	0xbf85ffdb, 0xbf9c0000,
-	0xe0524000, 0x731d0000,
-	0xe0524100, 0x731d0100,
-	0xe0524200, 0x731d0200,
-	0xe0524300, 0x731d0300,
+	0xbef80080, 0xbeee0078,
+	0x8078ff78, 0x00000400,
+	0xbefc0084, 0xbf11087c,
+	0xe0524000, 0x781d0000,
+	0xe0524100, 0x781d0100,
+	0xe0524200, 0x781d0200,
+	0xe0524300, 0x781d0300,
+	0xbf8c0f70, 0x7e000300,
+	0x7e020301, 0x7e040302,
+	0x7e060303, 0x807c847c,
+	0x8078ff78, 0x00000400,
+	0xbf0a6f7c, 0xbf85ffee,
+	0xbefc0080, 0xbf11087c,
+	0xe0524000, 0x781d0000,
+	0xe0524100, 0x781d0100,
+	0xe0524200, 0x781d0200,
+	0xe0524300, 0x781d0300,
 	0xbf8c0f70, 0xd3d94000,
 	0x18000100, 0xd3d94001,
 	0x18000101, 0xd3d94002,
 	0x18000102, 0xd3d94003,
-	0x18000103, 0xe0524000,
+	0x18000103, 0x807c847c,
+	0x8078ff78, 0x00000400,
+	0xbf0a6f7c, 0xbf85ffea,
+	0xbf9c0000, 0xe0524000,
 	0x6e1d0000, 0xe0524100,
 	0x6e1d0100, 0xe0524200,
 	0x6e1d0200, 0xe0524300,
-	0x6e1d0300, 0xb8f82a05,
-	0x80788178, 0x8e788a78,
-	0x8e788178, 0xb8ee1605,
-	0x806e816e, 0x8e6e866e,
-	0x80786e78, 0x80f8c078,
-	0xb8ef1605, 0x806f816f,
-	0x8e6f846f, 0x8e76826f,
-	0xbef600ff, 0x01000000,
-	0xbefc006f, 0xc031003a,
-	0x00000078, 0x80f8c078,
-	0xbf8cc07f, 0x80fc907c,
-	0xbf800000, 0xbe802d00,
-	0xbe822d02, 0xbe842d04,
-	0xbe862d06, 0xbe882d08,
-	0xbe8a2d0a, 0xbe8c2d0c,
-	0xbe8e2d0e, 0xbf06807c,
-	0xbf84fff0, 0xb8f82a05,
-	0x80788178, 0x8e788a78,
-	0x8e788178, 0xb8ee1605,
-	0x806e816e, 0x8e6e866e,
-	0x80786e78, 0xbef60084,
-	0xbef600ff, 0x01000000,
-	0xc0211bfa, 0x00000078,
-	0x80788478, 0xc0211b3a,
+	0x6e1d0300, 0xbf8c0f70,
+	0xb8f82a05, 0x80788178,
+	0x8e788a78, 0x8e788178,
+	0xb8ee1605, 0x806e816e,
+	0x8e6e866e, 0x80786e78,
+	0x80f8c078, 0xb8ef1605,
+	0x806f816f, 0x8e6f846f,
+	0x8e76826f, 0xbef600ff,
+	0x01000000, 0xbefc006f,
+	0xc031003a, 0x00000078,
+	0x80f8c078, 0xbf8cc07f,
+	0x80fc907c, 0xbf800000,
+	0xbe802d00, 0xbe822d02,
+	0xbe842d04, 0xbe862d06,
+	0xbe882d08, 0xbe8a2d0a,
+	0xbe8c2d0c, 0xbe8e2d0e,
+	0xbf06807c, 0xbf84fff0,
+	0xb8f82a05, 0x80788178,
+	0x8e788a78, 0x8e788178,
+	0xb8ee1605, 0x806e816e,
+	0x8e6e866e, 0x80786e78,
+	0xbef60084, 0xbef600ff,
+	0x01000000, 0xc0211bfa,
 	0x00000078, 0x80788478,
-	0xc0211b7a, 0x00000078,
-	0x80788478, 0xc0211c3a,
+	0xc0211b3a, 0x00000078,
+	0x80788478, 0xc0211b7a,
 	0x00000078, 0x80788478,
-	0xc0211c7a, 0x00000078,
-	0x80788478, 0xc0211eba,
+	0xc0211c3a, 0x00000078,
+	0x80788478, 0xc0211c7a,
 	0x00000078, 0x80788478,
-	0xc0211efa, 0x00000078,
-	0x80788478, 0xc0211a3a,
+	0xc0211eba, 0x00000078,
+	0x80788478, 0xc0211efa,
 	0x00000078, 0x80788478,
-	0xc0211a7a, 0x00000078,
-	0x80788478, 0xc0211cfa,
+	0xc0211a3a, 0x00000078,
+	0x80788478, 0xc0211a7a,
 	0x00000078, 0x80788478,
-	0xbf8cc07f, 0xbefc006f,
-	0xbefe0070, 0xbeff0071,
-	0x866f7bff, 0x000003ff,
-	0xb96f4803, 0x866f7bff,
-	0xfffff800, 0x8f6f8b6f,
-	0xb96fa2c3, 0xb973f801,
-	0xb8ee2a05, 0x806e816e,
-	0x8e6e8a6e, 0x8e6e816e,
-	0xb8ef1605, 0x806f816f,
-	0x8e6f866f, 0x806e6f6e,
-	0x806e746e, 0x826f8075,
-	0x866fff6f, 0x0000ffff,
-	0xc00b1c37, 0x00000050,
-	0xc00b1d37, 0x00000060,
-	0xc0031e77, 0x00000074,
-	0xbf8cc07f, 0x866fff6d,
-	0xf8000000, 0x8f6f9b6f,
-	0x8e6f906f, 0xbeee0080,
-	0x876e6f6e, 0x866fff6d,
-	0x04000000, 0x8f6f9a6f,
-	0x8e6f8f6f, 0x876e6f6e,
-	0x866fff7a, 0x00800000,
-	0x8f6f976f, 0xb96ef807,
+	0xc0211cfa, 0x00000078,
+	0x80788478, 0xbf8cc07f,
+	0xbefc006f, 0xbefe0070,
+	0xbeff0071, 0x866f7bff,
+	0x000003ff, 0xb96f4803,
+	0x866f7bff, 0xfffff800,
+	0x8f6f8b6f, 0xb96fa2c3,
+	0xb973f801, 0xb8ee2a05,
+	0x806e816e, 0x8e6e8a6e,
+	0x8e6e816e, 0xb8ef1605,
+	0x806f816f, 0x8e6f866f,
+	0x806e6f6e, 0x806e746e,
+	0x826f8075, 0x866fff6f,
+	0x0000ffff, 0xc00b1c37,
+	0x00000050, 0xc00b1d37,
+	0x00000060, 0xc0031e77,
+	0x00000074, 0xbf8cc07f,
+	0x8f6e8b77, 0x866eff6e,
+	0x001f8000, 0xb96ef807,
 	0x866dff6d, 0x0000ffff,
 	0x86fe7e7e, 0x86ea6a6a,
 	0x8f6e837a, 0xb96ee0c2,
 	0xbf800002, 0xb97a0002,
-	0xbf8a0000, 0x95806f6c,
-	0xbf810000, 0x00000000,
+	0xbf8a0000, 0xbe801f6c,
+	0xbf9b0000, 0x00000000,
 };
 
 static const uint32_t cwsr_trap_aldebaran_hex[] = {
-	0xbf820001, 0xbf8202ce,
-	0xb8f8f802, 0x89788678,
-	0xb8eef801, 0x866eff6e,
-	0x00000800, 0xbf840003,
+	0xbf820001, 0xbf8202e0,
+	0xb8f8f802, 0x8978ff78,
+	0x00020006, 0xb8fbf803,
 	0x866eff78, 0x00002000,
-	0xbf840016, 0xb8fbf803,
+	0xbf840009, 0x866eff6d,
+	0x00ff0000, 0xbf85001e,
 	0x866eff7b, 0x00000400,
-	0xbf85003b, 0x866eff7b,
-	0x00000800, 0xbf850003,
-	0x866eff7b, 0x00000100,
-	0xbf84000c, 0x866eff78,
-	0x00002000, 0xbf840005,
-	0xbf8e0010, 0xb8eef803,
-	0x866eff6e, 0x00000400,
-	0xbf84fffb, 0x8778ff78,
-	0x00002000, 0x80ec886c,
-	0x82ed806d, 0xb8eef807,
-	0x866fff6e, 0x001f8000,
-	0x8e6f8b6f, 0x8977ff77,
-	0xfc000000, 0x87776f77,
-	0x896eff6e, 0x001f8000,
-	0xb96ef807, 0xb8faf812,
+	0xbf850055, 0xbf8e0010,
+	0xb8fbf803, 0xbf82fffa,
+	0x866eff7b, 0x03c00900,
+	0xbf850015, 0x866eff7b,
+	0x000071ff, 0xbf840008,
+	0x866fff7b, 0x00007080,
+	0xbf840001, 0xbeee1a87,
+	0xb8eff801, 0x8e6e8c6e,
+	0x866e6f6e, 0xbf85000a,
+	0x866eff6d, 0x00ff0000,
+	0xbf850007, 0xb8eef801,
+	0x866eff6e, 0x00000800,
+	0xbf850003, 0x866eff7b,
+	0x00000400, 0xbf85003a,
+	0xb8faf807, 0x867aff7a,
+	0x001f8000, 0x8e7a8b7a,
+	0x8977ff77, 0xfc000000,
+	0x87777a77, 0xba7ff807,
+	0x00000000, 0xb8faf812,
 	0xb8fbf813, 0x8efa887a,
-	0xc0071bbd, 0x00000000,
-	0xbf8cc07f, 0xc0071ebd,
-	0x00000008, 0xbf8cc07f,
-	0x86ee6e6e, 0xbf840001,
-	0xbe801d6e, 0xb8fbf803,
-	0x867bff7b, 0x000001ff,
+	0xbf0d8f7b, 0xbf840002,
+	0x877bff7b, 0xffff0000,
+	0xc0031bbd, 0x00000010,
+	0xbf8cc07f, 0x8e6e976e,
+	0x8977ff77, 0x00800000,
+	0x87776e77, 0xc0071bbd,
+	0x00000000, 0xbf8cc07f,
+	0xc0071ebd, 0x00000008,
+	0xbf8cc07f, 0x86ee6e6e,
+	0xbf840001, 0xbe801d6e,
+	0x866eff6d, 0x01ff0000,
+	0xbf850005, 0x8778ff78,
+	0x00002000, 0x80ec886c,
+	0x82ed806d, 0xbf820005,
+	0x866eff6d, 0x01000000,
 	0xbf850002, 0x806c846c,
 	0x826d806d, 0x866dff6d,
-	0x0000ffff, 0x8f6e8b77,
-	0x866eff6e, 0x001f8000,
-	0xb96ef807, 0x86fe7e7e,
+	0x0000ffff, 0x8f7a8b77,
+	0x867aff7a, 0x001f8000,
+	0xb97af807, 0x86fe7e7e,
 	0x86ea6a6a, 0x8f6e8378,
 	0xb96ee0c2, 0xbf800002,
 	0xb9780002, 0xbe801f6c,
 	0x866dff6d, 0x0000ffff,
 	0xbefa0080, 0xb97a0283,
-	0xb8fa2407, 0x8e7a9b7a,
-	0x876d7a6d, 0xb8fa03c7,
-	0x8e7a9a7a, 0x876d7a6d,
 	0xb8faf807, 0x867aff7a,
-	0x00007fff, 0xb97af807,
-	0xbeee007e, 0xbeef007f,
-	0xbefe0180, 0xbf900004,
-	0x877a8478, 0xb97af802,
-	0xbf8e0002, 0xbf88fffe,
-	0xb8fa2985, 0x807a817a,
-	0x8e7a8a7a, 0x8e7a817a,
-	0xb8fb1605, 0x807b817b,
-	0x8e7b867b, 0x807a7b7a,
-	0x807a7e7a, 0x827b807f,
-	0x867bff7b, 0x0000ffff,
-	0xc04b1c3d, 0x00000050,
-	0xbf8cc07f, 0xc04b1d3d,
-	0x00000060, 0xbf8cc07f,
-	0xc0431e7d, 0x00000074,
-	0xbf8cc07f, 0xbef4007e,
-	0x8675ff7f, 0x0000ffff,
-	0x8775ff75, 0x00040000,
-	0xbef60080, 0xbef700ff,
-	0x00807fac, 0x867aff7f,
-	0x08000000, 0x8f7a837a,
-	0x87777a77, 0x867aff7f,
-	0x70000000, 0x8f7a817a,
-	0x87777a77, 0xbef1007c,
-	0xbef00080, 0xb8f02985,
-	0x80708170, 0x8e708a70,
-	0x8e708170, 0xb8fa1605,
-	0x807a817a, 0x8e7a867a,
-	0x80707a70, 0xbef60084,
-	0xbef600ff, 0x01000000,
-	0xbefe007c, 0xbefc0070,
-	0xc0611c7a, 0x0000007c,
-	0xbf8cc07f, 0x80708470,
-	0xbefc007e, 0xbefe007c,
-	0xbefc0070, 0xc0611b3a,
+	0x001f8000, 0x8e7a8b7a,
+	0x8977ff77, 0xfc000000,
+	0x87777a77, 0xba7ff807,
+	0x00000000, 0xbeee007e,
+	0xbeef007f, 0xbefe0180,
+	0xbf900004, 0x877a8478,
+	0xb97af802, 0xbf8e0002,
+	0xbf88fffe, 0xb8fa2985,
+	0x807a817a, 0x8e7a8a7a,
+	0x8e7a817a, 0xb8fb1605,
+	0x807b817b, 0x8e7b867b,
+	0x807a7b7a, 0x807a7e7a,
+	0x827b807f, 0x867bff7b,
+	0x0000ffff, 0xc04b1c3d,
+	0x00000050, 0xbf8cc07f,
+	0xc04b1d3d, 0x00000060,
+	0xbf8cc07f, 0xc0431e7d,
+	0x00000074, 0xbf8cc07f,
+	0xbef4007e, 0x8675ff7f,
+	0x0000ffff, 0x8775ff75,
+	0x00040000, 0xbef60080,
+	0xbef700ff, 0x00807fac,
+	0xbef1007c, 0xbef00080,
+	0xb8f02985, 0x80708170,
+	0x8e708a70, 0x8e708170,
+	0xb8fa1605, 0x807a817a,
+	0x8e7a867a, 0x80707a70,
+	0xbef60084, 0xbef600ff,
+	0x01000000, 0xbefe007c,
+	0xbefc0070, 0xc0611c7a,
 	0x0000007c, 0xbf8cc07f,
 	0x80708470, 0xbefc007e,
 	0xbefe007c, 0xbefc0070,
-	0xc0611b7a, 0x0000007c,
+	0xc0611b3a, 0x0000007c,
 	0xbf8cc07f, 0x80708470,
 	0xbefc007e, 0xbefe007c,
-	0xbefc0070, 0xc0611bba,
+	0xbefc0070, 0xc0611b7a,
 	0x0000007c, 0xbf8cc07f,
 	0x80708470, 0xbefc007e,
 	0xbefe007c, 0xbefc0070,
-	0xc0611bfa, 0x0000007c,
+	0xc0611bba, 0x0000007c,
 	0xbf8cc07f, 0x80708470,
 	0xbefc007e, 0xbefe007c,
-	0xbefc0070, 0xc0611e3a,
-	0x0000007c, 0xbf8cc07f,
-	0x80708470, 0xbefc007e,
-	0xb8fbf803, 0xbefe007c,
-	0xbefc0070, 0xc0611efa,
+	0xbefc0070, 0xc0611bfa,
 	0x0000007c, 0xbf8cc07f,
 	0x80708470, 0xbefc007e,
 	0xbefe007c, 0xbefc0070,
-	0xc0611a3a, 0x0000007c,
+	0xc0611e3a, 0x0000007c,
+	0xbf8cc07f, 0x80708470,
+	0xbefc007e, 0xb8fbf803,
+	0xbefe007c, 0xbefc0070,
+	0xc0611efa, 0x0000007c,
 	0xbf8cc07f, 0x80708470,
 	0xbefc007e, 0xbefe007c,
-	0xbefc0070, 0xc0611a7a,
-	0x0000007c, 0xbf8cc07f,
-	0x80708470, 0xbefc007e,
-	0xb8f1f801, 0xbefe007c,
-	0xbefc0070, 0xc0611c7a,
+	0xbefc0070, 0xc0611a3a,
 	0x0000007c, 0xbf8cc07f,
 	0x80708470, 0xbefc007e,
+	0xbefe007c, 0xbefc0070,
+	0xc0611a7a, 0x0000007c,
+	0xbf8cc07f, 0x80708470,
+	0xbefc007e, 0xb8f1f801,
+	0xbefe007c, 0xbefc0070,
+	0xc0611c7a, 0x0000007c,
+	0xbf8cc07f, 0x80708470,
+	0xbefc007e, 0xbf108080,
 	0x867aff7f, 0x04000000,
 	0xbeef0080, 0x876f6f7a,
 	0xb8f02985, 0x80708170,
@@ -1935,15 +2152,11 @@ static const uint32_t cwsr_trap_aldebaran_hex[] = {
 	0x701d0300, 0x807c847c,
 	0x8070ff70, 0x00000400,
 	0xbf0a7b7c, 0xbf85ffeb,
-	0xbf9c0000, 0xbf820101,
+	0xbf9c0000, 0xbf8200ee,
 	0xbef4007e, 0x8675ff7f,
 	0x0000ffff, 0x8775ff75,
 	0x00040000, 0xbef60080,
 	0xbef700ff, 0x00807fac,
-	0x866eff7f, 0x08000000,
-	0x8f6e836e, 0x87776e77,
-	0x866eff7f, 0x70000000,
-	0x8f6e816e, 0x87776e77,
 	0x866eff7f, 0x04000000,
 	0xbf84001f, 0xbefe00c1,
 	0xbeff00c1, 0xb8ef4306,
@@ -2052,41 +2265,56 @@ static const uint32_t cwsr_trap_aldebaran_hex[] = {
 	0x00000050, 0xc00b1d37,
 	0x00000060, 0xc0031e77,
 	0x00000074, 0xbf8cc07f,
-	0x866fff6d, 0xf8000000,
-	0x8f6f9b6f, 0x8e6f906f,
-	0xbeee0080, 0x876e6f6e,
-	0x866fff6d, 0x04000000,
-	0x8f6f9a6f, 0x8e6f8f6f,
-	0x876e6f6e, 0x866fff7a,
-	0x00800000, 0x8f6f976f,
-	0xb96ef807, 0x866dff6d,
-	0x0000ffff, 0x86fe7e7e,
-	0x86ea6a6a, 0x8f6e837a,
-	0xb96ee0c2, 0xbf800002,
-	0xb97a0002, 0xbf8a0000,
-	0x95806f6c, 0xbf810000,
+	0x8f6e8b77, 0x866eff6e,
+	0x001f8000, 0xb96ef807,
+	0x866dff6d, 0x0000ffff,
+	0x86fe7e7e, 0x86ea6a6a,
+	0x8f6e837a, 0xb96ee0c2,
+	0xbf800002, 0xb97a0002,
+	0xbf8a0000, 0xbe801f6c,
+	0xbf9b0000, 0x00000000,
 };
 
 static const uint32_t cwsr_trap_gfx10_hex[] = {
-	0xbf820001, 0xbf8201cf,
+	0xbf820001, 0xbf820220,
 	0xb0804004, 0xb978f802,
-	0x8a788678, 0xb96ef801,
-	0x876eff6e, 0x00000800,
-	0xbf840003, 0x876eff78,
+	0x8a78ff78, 0x00020006,
+	0xb97bf803, 0x876eff78,
 	0x00002000, 0xbf840009,
-	0xb97bf803, 0x876eff7b,
-	0x00000400, 0xbf85001d,
-	0x876eff7b, 0x00000100,
-	0xbf840002, 0x8878ff78,
-	0x00002000, 0xb97af812,
+	0x876eff6d, 0x00ff0000,
+	0xbf85001e, 0x876eff7b,
+	0x00000400, 0xbf850045,
+	0xbf8e0010, 0xb97bf803,
+	0xbf82fffa, 0x876eff7b,
+	0x00000900, 0xbf850015,
+	0x876eff7b, 0x000071ff,
+	0xbf840008, 0x876fff7b,
+	0x00007080, 0xbf840001,
+	0xbeee1d87, 0xb96ff801,
+	0x8f6e8c6e, 0x876e6f6e,
+	0xbf85000a, 0x876eff6d,
+	0x00ff0000, 0xbf850007,
+	0xb96ef801, 0x876eff6e,
+	0x00000800, 0xbf850003,
+	0x876eff7b, 0x00000400,
+	0xbf85002a, 0xb97af812,
 	0xb97bf813, 0x8ffa887a,
-	0xf4051bbd, 0xfa000000,
-	0xbf8cc07f, 0xf4051ebd,
-	0xfa000008, 0xbf8cc07f,
-	0x87ee6e6e, 0xbf840001,
-	0xbe80206e, 0xb97bf803,
-	0x877bff7b, 0x000001ff,
-	0xbf850002, 0x806c846c,
+	0xbf0d8f7b, 0xbf840002,
+	0x887bff7b, 0xffff0000,
+	0xf4011bbd, 0xfa000010,
+	0xbf8c0000, 0x8f6e976e,
+	0x8a77ff77, 0x00800000,
+	0x88776e77, 0xf4051bbd,
+	0xfa000000, 0xbf8c0000,
+	0xf4051ebd, 0xfa000008,
+	0xbf8c0000, 0x87ee6e6e,
+	0xbf840001, 0xbe80206e,
+	0x876eff6d, 0x00ff0000,
+	0xbf850008, 0x876eff6d,
+	0x01000000, 0xbf850007,
+	0x8878ff78, 0x00002000,
+	0x80ec886c, 0x82ed806d,
+	0xbf820002, 0x806c846c,
 	0x826d806d, 0x876dff6d,
 	0x0000ffff, 0x87fe7e7e,
 	0x87ea6a6a, 0xb9f8f802,
@@ -2094,70 +2322,91 @@ static const uint32_t cwsr_trap_gfx10_hex[] = {
 	0x0000ffff, 0xbefa0380,
 	0xb9fa0283, 0xbeee037e,
 	0xbeef037f, 0xbefe0480,
-	0xbf900004, 0xbf8cc07f,
-	0xb97b02dc, 0x8f7b997b,
-	0x887b7b7f, 0xb97a2a05,
+	0xbf900004, 0xbf8c0000,
+	0x877aff7f, 0x04000000,
+	0x8f7a857a, 0x886d7a6d,
+	0x7e008200, 0xbefa037e,
+	0x877bff7f, 0x0000ffff,
+	0xbefe03c1, 0xbeff03c1,
+	0xdc5f8000, 0x007a0000,
+	0x7e000280, 0xbefe037a,
+	0xbeff037b, 0xb97b02dc,
+	0x8f7b997b, 0xb97a3a05,
 	0x807a817a, 0xbf0d997b,
 	0xbf850002, 0x8f7a897a,
 	0xbf820001, 0x8f7a8a7a,
-	0x877bff7f, 0x0000ffff,
-	0x807aff7a, 0x00000200,
-	0x807a7e7a, 0x827b807b,
+	0xb97b1e06, 0x8f7b8a7b,
+	0x807a7b7a, 0x877bff7f,
+	0x0000ffff, 0x807aff7a,
+	0x00000200, 0x807a7e7a,
+	0x827b807b, 0xd7610000,
+	0x00010870, 0xd7610000,
+	0x00010a71, 0xd7610000,
+	0x00010c72, 0xd7610000,
+	0x00010e73, 0xd7610000,
+	0x00011074, 0xd7610000,
+	0x00011275, 0xd7610000,
+	0x00011476, 0xd7610000,
+	0x00011677, 0xd7610000,
+	0x00011a79, 0xd7610000,
+	0x00011c7e, 0xd7610000,
+	0x00011e7f, 0xbefe03ff,
+	0x00003fff, 0xbeff0380,
+	0xdc5f8040, 0x007a0000,
+	0xd760007a, 0x00011d00,
+	0xd760007b, 0x00011f00,
+	0xbefe037a, 0xbeff037b,
 	0xbef4037e, 0x8775ff7f,
 	0x0000ffff, 0x8875ff75,
 	0x00040000, 0xbef60380,
 	0xbef703ff, 0x10807fac,
-	0x877aff7f, 0x08000000,
-	0x907a837a, 0x88777a77,
-	0x877aff7f, 0x70000000,
-	0x907a817a, 0x88777a77,
 	0xbef1037c, 0xbef00380,
 	0xb97302dc, 0x8f739973,
-	0x8873737f, 0xbefe03c1,
-	0x907c9973, 0x877c817c,
-	0xbf06817c, 0xbf850002,
-	0xbeff0380, 0xbf820002,
-	0xbeff03c1, 0xbf82000b,
+	0xbefe03c1, 0x907c9973,
+	0x877c817c, 0xbf06817c,
+	0xbf850002, 0xbeff0380,
+	0xbf820002, 0xbeff03c1,
+	0xbf820009, 0xbef603ff,
+	0x01000000, 0xe0704080,
+	0x705d0100, 0xe0704100,
+	0x705d0200, 0xe0704180,
+	0x705d0300, 0xbf820008,
 	0xbef603ff, 0x01000000,
-	0xe0704000, 0x705d0000,
-	0xe0704080, 0x705d0100,
-	0xe0704100, 0x705d0200,
-	0xe0704180, 0x705d0300,
-	0xbf82000a, 0xbef603ff,
-	0x01000000, 0xe0704000,
-	0x705d0000, 0xe0704100,
-	0x705d0100, 0xe0704200,
-	0x705d0200, 0xe0704300,
-	0x705d0300, 0xb9702a05,
-	0x80708170, 0xbf0d9973,
-	0xbf850002, 0x8f708970,
-	0xbf820001, 0x8f708a70,
-	0xb97a1e06, 0x8f7a8a7a,
-	0x80707a70, 0x8070ff70,
-	0x00000200, 0xbef603ff,
-	0x01000000, 0x7e000280,
-	0x7e020280, 0x7e040280,
-	0xbefc0380, 0xd7610002,
-	0x0000f871, 0x807c817c,
-	0xd7610002, 0x0000f86c,
-	0x807c817c, 0xd7610002,
-	0x0000f86d, 0x807c817c,
-	0xd7610002, 0x0000f86e,
+	0xe0704100, 0x705d0100,
+	0xe0704200, 0x705d0200,
+	0xe0704300, 0x705d0300,
+	0xb9703a05, 0x80708170,
+	0xbf0d9973, 0xbf850002,
+	0x8f708970, 0xbf820001,
+	0x8f708a70, 0xb97a1e06,
+	0x8f7a8a7a, 0x80707a70,
+	0x8070ff70, 0x00000200,
+	0xbef603ff, 0x01000000,
+	0x7e000280, 0x7e020280,
+	0x7e040280, 0xbefc0380,
+	0xd7610002, 0x0000f871,
 	0x807c817c, 0xd7610002,
-	0x0000f86f, 0x807c817c,
-	0xd7610002, 0x0000f878,
-	0x807c817c, 0xb97af803,
+	0x0000f86c, 0x807c817c,
+	0x8a7aff6d, 0x80000000,
 	0xd7610002, 0x0000f87a,
 	0x807c817c, 0xd7610002,
-	0x0000f87b, 0x807c817c,
-	0xb971f801, 0xd7610002,
-	0x0000f871, 0x807c817c,
-	0xb971f814, 0xd7610002,
-	0x0000f871, 0x807c817c,
-	0xb971f815, 0xd7610002,
-	0x0000f871, 0x807c817c,
-	0xbeff0380, 0xe0704000,
-	0x705d0200, 0xb9702a05,
+	0x0000f86e, 0x807c817c,
+	0xd7610002, 0x0000f86f,
+	0x807c817c, 0xd7610002,
+	0x0000f878, 0x807c817c,
+	0xb97af803, 0xd7610002,
+	0x0000f87a, 0x807c817c,
+	0xd7610002, 0x0000f87b,
+	0x807c817c, 0xb971f801,
+	0xd7610002, 0x0000f871,
+	0x807c817c, 0xb971f814,
+	0xd7610002, 0x0000f871,
+	0x807c817c, 0xb971f815,
+	0xd7610002, 0x0000f871,
+	0x807c817c, 0xbefe03ff,
+	0x0000ffff, 0xbeff0380,
+	0xe0704000, 0x705d0200,
+	0xbefe03c1, 0xb9703a05,
 	0x80708170, 0xbf0d9973,
 	0xbf850002, 0x8f708970,
 	0xbf820001, 0x8f708a70,
@@ -2226,150 +2475,147 @@ static const uint32_t cwsr_trap_gfx10_hex[] = {
 	0xbf850002, 0xbeff0380,
 	0xbf820001, 0xbeff03c1,
 	0xb97b4306, 0x877bc17b,
-	0xbf840044, 0xbf8a0000,
-	0x877aff73, 0x04000000,
-	0xbf840040, 0x8f7b867b,
-	0x8f7b827b, 0xbef6037b,
-	0xb9702a05, 0x80708170,
-	0xbf0d9973, 0xbf850002,
-	0x8f708970, 0xbf820001,
-	0x8f708a70, 0xb97a1e06,
-	0x8f7a8a7a, 0x80707a70,
-	0x8070ff70, 0x00000200,
-	0x8070ff70, 0x00000080,
-	0xbef603ff, 0x01000000,
-	0xd7650000, 0x000100c1,
-	0xd7660000, 0x000200c1,
-	0x16000084, 0x907c9973,
-	0x877c817c, 0xbf06817c,
-	0xbefc0380, 0xbf850012,
-	0xbe8303ff, 0x00000080,
+	0xbf840043, 0xbf8a0000,
+	0x877aff6d, 0x80000000,
+	0xbf84003f, 0x8f7b887b,
+	0xbef6037b, 0xb9703a05,
+	0x80708170, 0xbf0d9973,
+	0xbf850002, 0x8f708970,
+	0xbf820001, 0x8f708a70,
+	0xb97a1e06, 0x8f7a8a7a,
+	0x80707a70, 0x8070ff70,
+	0x00000200, 0x8070ff70,
+	0x00000080, 0xbef603ff,
+	0x01000000, 0xd7650000,
+	0x000100c1, 0xd7660000,
+	0x000200c1, 0x16000084,
+	0x907c9973, 0x877c817c,
+	0xbf06817c, 0xbefc0380,
+	0xbf850012, 0xbe8303ff,
+	0x00000080, 0xbf800000,
 	0xbf800000, 0xbf800000,
-	0xbf800000, 0xd8d80000,
-	0x01000000, 0xbf8c0000,
-	0xe0704000, 0x705d0100,
-	0x807c037c, 0x80700370,
-	0xd5250000, 0x0001ff00,
-	0x00000080, 0xbf0a7b7c,
-	0xbf85fff4, 0xbf820011,
-	0xbe8303ff, 0x00000100,
+	0xd8d80000, 0x01000000,
+	0xbf8c0000, 0xe0704000,
+	0x705d0100, 0x807c037c,
+	0x80700370, 0xd5250000,
+	0x0001ff00, 0x00000080,
+	0xbf0a7b7c, 0xbf85fff4,
+	0xbf820011, 0xbe8303ff,
+	0x00000100, 0xbf800000,
 	0xbf800000, 0xbf800000,
-	0xbf800000, 0xd8d80000,
-	0x01000000, 0xbf8c0000,
-	0xe0704000, 0x705d0100,
-	0x807c037c, 0x80700370,
-	0xd5250000, 0x0001ff00,
-	0x00000100, 0xbf0a7b7c,
-	0xbf85fff4, 0xbefe03c1,
-	0x907c9973, 0x877c817c,
-	0xbf06817c, 0xbf850004,
-	0xbef003ff, 0x00000200,
-	0xbeff0380, 0xbf820003,
-	0xbef003ff, 0x00000400,
-	0xbeff03c1, 0xb97b2a05,
-	0x807b817b, 0x8f7b827b,
-	0x907c9973, 0x877c817c,
-	0xbf06817c, 0xbf850017,
+	0xd8d80000, 0x01000000,
+	0xbf8c0000, 0xe0704000,
+	0x705d0100, 0x807c037c,
+	0x80700370, 0xd5250000,
+	0x0001ff00, 0x00000100,
+	0xbf0a7b7c, 0xbf85fff4,
+	0xbefe03c1, 0x907c9973,
+	0x877c817c, 0xbf06817c,
+	0xbf850004, 0xbef003ff,
+	0x00000200, 0xbeff0380,
+	0xbf820003, 0xbef003ff,
+	0x00000400, 0xbeff03c1,
+	0xb97b3a05, 0x807b817b,
+	0x8f7b827b, 0x907c9973,
+	0x877c817c, 0xbf06817c,
+	0xbf850017, 0xbef603ff,
+	0x01000000, 0xbefc0384,
+	0xbf0a7b7c, 0xbf840037,
+	0x7e008700, 0x7e028701,
+	0x7e048702, 0x7e068703,
+	0xe0704000, 0x705d0000,
+	0xe0704080, 0x705d0100,
+	0xe0704100, 0x705d0200,
+	0xe0704180, 0x705d0300,
+	0x807c847c, 0x8070ff70,
+	0x00000200, 0xbf0a7b7c,
+	0xbf85ffef, 0xbf820025,
 	0xbef603ff, 0x01000000,
 	0xbefc0384, 0xbf0a7b7c,
-	0xbf840037, 0x7e008700,
+	0xbf840011, 0x7e008700,
 	0x7e028701, 0x7e048702,
 	0x7e068703, 0xe0704000,
-	0x705d0000, 0xe0704080,
-	0x705d0100, 0xe0704100,
-	0x705d0200, 0xe0704180,
+	0x705d0000, 0xe0704100,
+	0x705d0100, 0xe0704200,
+	0x705d0200, 0xe0704300,
 	0x705d0300, 0x807c847c,
-	0x8070ff70, 0x00000200,
+	0x8070ff70, 0x00000400,
 	0xbf0a7b7c, 0xbf85ffef,
-	0xbf820025, 0xbef603ff,
-	0x01000000, 0xbefc0384,
-	0xbf0a7b7c, 0xbf840020,
-	0x7e008700, 0x7e028701,
-	0x7e048702, 0x7e068703,
+	0xb97b1e06, 0x877bc17b,
+	0xbf84000c, 0x8f7b837b,
+	0x807b7c7b, 0xbefe03c1,
+	0xbeff0380, 0x7e008700,
 	0xe0704000, 0x705d0000,
-	0xe0704100, 0x705d0100,
-	0xe0704200, 0x705d0200,
-	0xe0704300, 0x705d0300,
-	0x807c847c, 0x8070ff70,
-	0x00000400, 0xbf0a7b7c,
-	0xbf85ffef, 0xb97b1e06,
-	0x877bc17b, 0xbf84000c,
-	0x8f7b837b, 0x807b7c7b,
-	0xbefe03c1, 0xbeff0380,
-	0x7e008700, 0xe0704000,
-	0x705d0000, 0x807c817c,
-	0x8070ff70, 0x00000080,
-	0xbf0a7b7c, 0xbf85fff8,
-	0xbf82013c, 0xbef4037e,
-	0x8775ff7f, 0x0000ffff,
-	0x8875ff75, 0x00040000,
-	0xbef60380, 0xbef703ff,
-	0x10807fac, 0x876eff7f,
-	0x08000000, 0x906e836e,
-	0x88776e77, 0x876eff7f,
-	0x70000000, 0x906e816e,
-	0x88776e77, 0xb97202dc,
-	0x8f729972, 0x8872727f,
+	0x807c817c, 0x8070ff70,
+	0x00000080, 0xbf0a7b7c,
+	0xbf85fff8, 0xbf820136,
+	0xbef4037e, 0x8775ff7f,
+	0x0000ffff, 0x8875ff75,
+	0x00040000, 0xbef60380,
+	0xbef703ff, 0x10807fac,
+	0xb97202dc, 0x8f729972,
 	0x876eff7f, 0x04000000,
-	0xbf840034, 0xbefe03c1,
+	0xbf840033, 0xbefe03c1,
 	0x907c9972, 0x877c817c,
 	0xbf06817c, 0xbf850002,
 	0xbeff0380, 0xbf820001,
 	0xbeff03c1, 0xb96f4306,
-	0x876fc16f, 0xbf840029,
-	0x8f6f866f, 0x8f6f826f,
-	0xbef6036f, 0xb9782a05,
-	0x80788178, 0xbf0d9972,
-	0xbf850002, 0x8f788978,
-	0xbf820001, 0x8f788a78,
-	0xb96e1e06, 0x8f6e8a6e,
-	0x80786e78, 0x8078ff78,
-	0x00000200, 0x8078ff78,
-	0x00000080, 0xbef603ff,
-	0x01000000, 0x907c9972,
-	0x877c817c, 0xbf06817c,
-	0xbefc0380, 0xbf850009,
-	0xe0310000, 0x781d0000,
-	0x807cff7c, 0x00000080,
+	0x876fc16f, 0xbf840028,
+	0x8f6f886f, 0xbef6036f,
+	0xb9783a05, 0x80788178,
+	0xbf0d9972, 0xbf850002,
+	0x8f788978, 0xbf820001,
+	0x8f788a78, 0xb96e1e06,
+	0x8f6e8a6e, 0x80786e78,
+	0x8078ff78, 0x00000200,
 	0x8078ff78, 0x00000080,
-	0xbf0a6f7c, 0xbf85fff8,
-	0xbf820008, 0xe0310000,
+	0xbef603ff, 0x01000000,
+	0x907c9972, 0x877c817c,
+	0xbf06817c, 0xbefc0380,
+	0xbf850009, 0xe0310000,
 	0x781d0000, 0x807cff7c,
-	0x00000100, 0x8078ff78,
-	0x00000100, 0xbf0a6f7c,
-	0xbf85fff8, 0xbef80380,
-	0xbefe03c1, 0x907c9972,
-	0x877c817c, 0xbf06817c,
-	0xbf850002, 0xbeff0380,
-	0xbf820001, 0xbeff03c1,
-	0xb96f2a05, 0x806f816f,
-	0x8f6f826f, 0x907c9972,
-	0x877c817c, 0xbf06817c,
-	0xbf850021, 0xbef603ff,
-	0x01000000, 0xbeee0378,
+	0x00000080, 0x8078ff78,
+	0x00000080, 0xbf0a6f7c,
+	0xbf85fff8, 0xbf820008,
+	0xe0310000, 0x781d0000,
+	0x807cff7c, 0x00000100,
+	0x8078ff78, 0x00000100,
+	0xbf0a6f7c, 0xbf85fff8,
+	0xbef80380, 0xbefe03c1,
+	0x907c9972, 0x877c817c,
+	0xbf06817c, 0xbf850002,
+	0xbeff0380, 0xbf820001,
+	0xbeff03c1, 0xb96f3a05,
+	0x806f816f, 0x8f6f826f,
+	0x907c9972, 0x877c817c,
+	0xbf06817c, 0xbf850024,
+	0xbef603ff, 0x01000000,
+	0xbeee0378, 0x8078ff78,
+	0x00000200, 0xbefc0384,
+	0xbf0a6f7c, 0xbf840050,
+	0xe0304000, 0x785d0000,
+	0xe0304080, 0x785d0100,
+	0xe0304100, 0x785d0200,
+	0xe0304180, 0x785d0300,
+	0xbf8c0000, 0x7e008500,
+	0x7e028501, 0x7e048502,
+	0x7e068503, 0x807c847c,
 	0x8078ff78, 0x00000200,
-	0xbefc0384, 0xe0304000,
-	0x785d0000, 0xe0304080,
-	0x785d0100, 0xe0304100,
-	0x785d0200, 0xe0304180,
-	0x785d0300, 0xbf8c3f70,
-	0x7e008500, 0x7e028501,
-	0x7e048502, 0x7e068503,
-	0x807c847c, 0x8078ff78,
-	0x00000200, 0xbf0a6f7c,
-	0xbf85ffee, 0xe0304000,
-	0x6e5d0000, 0xe0304080,
-	0x6e5d0100, 0xe0304100,
-	0x6e5d0200, 0xe0304180,
-	0x6e5d0300, 0xbf820032,
+	0xbf0a6f7c, 0xbf85ffee,
+	0xe0304000, 0x6e5d0000,
+	0xe0304080, 0x6e5d0100,
+	0xe0304100, 0x6e5d0200,
+	0xe0304180, 0x6e5d0300,
+	0xbf8c0000, 0xbf820034,
 	0xbef603ff, 0x01000000,
 	0xbeee0378, 0x8078ff78,
 	0x00000400, 0xbefc0384,
+	0xbf0a6f7c, 0xbf840012,
 	0xe0304000, 0x785d0000,
 	0xe0304100, 0x785d0100,
 	0xe0304200, 0x785d0200,
 	0xe0304300, 0x785d0300,
-	0xbf8c3f70, 0x7e008500,
+	0xbf8c0000, 0x7e008500,
 	0x7e028501, 0x7e048502,
 	0x7e068503, 0x807c847c,
 	0x8078ff78, 0x00000400,
@@ -2378,7 +2624,7 @@ static const uint32_t cwsr_trap_gfx10_hex[] = {
 	0xbf84000e, 0x8f6f836f,
 	0x806f7c6f, 0xbefe03c1,
 	0xbeff0380, 0xe0304000,
-	0x785d0000, 0xbf8c3f70,
+	0x785d0000, 0xbf8c0000,
 	0x7e008500, 0x807c817c,
 	0x8078ff78, 0x00000080,
 	0xbf0a6f7c, 0xbf85fff7,
@@ -2386,8 +2632,8 @@ static const uint32_t cwsr_trap_gfx10_hex[] = {
 	0x6e5d0000, 0xe0304100,
 	0x6e5d0100, 0xe0304200,
 	0x6e5d0200, 0xe0304300,
-	0x6e5d0300, 0xbf8c3f70,
-	0xb9782a05, 0x80788178,
+	0x6e5d0300, 0xbf8c0000,
+	0xb9783a05, 0x80788178,
 	0xbf0d9972, 0xbf850002,
 	0x8f788978, 0xbf820001,
 	0x8f788a78, 0xb96e1e06,
@@ -2397,16 +2643,16 @@ static const uint32_t cwsr_trap_gfx10_hex[] = {
 	0xbef603ff, 0x01000000,
 	0xbefc03ff, 0x0000006c,
 	0x80f89078, 0xf429003a,
-	0xf0000000, 0xbf8cc07f,
+	0xf0000000, 0xbf8c0000,
 	0x80fc847c, 0xbf800000,
 	0xbe803100, 0xbe823102,
 	0x80f8a078, 0xf42d003a,
-	0xf0000000, 0xbf8cc07f,
+	0xf0000000, 0xbf8c0000,
 	0x80fc887c, 0xbf800000,
 	0xbe803100, 0xbe823102,
 	0xbe843104, 0xbe863106,
 	0x80f8c078, 0xf431003a,
-	0xf0000000, 0xbf8cc07f,
+	0xf0000000, 0xbf8c0000,
 	0x80fc907c, 0xbf800000,
 	0xbe803100, 0xbe823102,
 	0xbe843104, 0xbe863106,
@@ -2414,7 +2660,7 @@ static const uint32_t cwsr_trap_gfx10_hex[] = {
 	0xbe8c310c, 0xbe8e310e,
 	0xbf06807c, 0xbf84fff0,
 	0xba80f801, 0x00000000,
-	0xbf8a0000, 0xb9782a05,
+	0xbf8a0000, 0xb9783a05,
 	0x80788178, 0xbf0d9972,
 	0xbf850002, 0x8f788978,
 	0xbf820001, 0x8f788a78,
@@ -2436,30 +2682,1975 @@ static const uint32_t cwsr_trap_gfx10_hex[] = {
 	0x80788478, 0xf4211cfa,
 	0xf0000000, 0x80788478,
 	0xf4211bba, 0xf0000000,
-	0x80788478, 0xbf8cc07f,
+	0x80788478, 0xbf8c0000,
 	0xb9eef814, 0xf4211bba,
 	0xf0000000, 0x80788478,
-	0xbf8cc07f, 0xb9eef815,
+	0xbf8c0000, 0xb9eef815,
 	0xbefc036f, 0xbefe0370,
-	0xbeff0371, 0x876f7bff,
-	0x000003ff, 0xb9ef4803,
-	0x876f7bff, 0xfffff800,
-	0x906f8b6f, 0xb9efa2c3,
-	0xb9f3f801, 0xb96e2a05,
+	0xbeff0371, 0xb9fb4803,
+	0x907b8b7b, 0xb9fba2c3,
+	0xb9f3f801, 0xb96e3a05,
 	0x806e816e, 0xbf0d9972,
 	0xbf850002, 0x8f6e896e,
 	0xbf820001, 0x8f6e8a6e,
+	0xb96f1e06, 0x8f6f8a6f,
+	0x806e6f6e, 0x806eff6e,
+	0x00000200, 0x806e746e,
+	0x826f8075, 0x876fff6f,
+	0x0000ffff, 0xf4091c37,
+	0xfa000050, 0xf4091d37,
+	0xfa000060, 0xf4011e77,
+	0xfa000074, 0xbf8c0000,
+	0x876dff6d, 0x0000ffff,
+	0x87fe7e7e, 0x87ea6a6a,
+	0xb9faf802, 0xbe80226c,
+	0xbf9b0000, 0xbf9f0000,
+	0xbf9f0000, 0xbf9f0000,
+	0xbf9f0000, 0xbf9f0000,
+};
+
+static const uint32_t cwsr_trap_gfx11_hex[] = {
+	0xbfa00001, 0xbfa00227,
+	0xb0804006, 0xb8f8f802,
+	0x9178ff78, 0x00020006,
+	0xb8fbf803, 0xbf0d9e6d,
+	0xbfa10001, 0xbfbd0000,
+	0xbf0d9f6d, 0xbfa20006,
+	0x8b6eff78, 0x00002000,
+	0xbfa10009, 0x8b6eff6d,
+	0x00ff0000, 0xbfa2001e,
+	0x8b6eff7b, 0x00000400,
+	0xbfa20045, 0xbf830010,
+	0xb8fbf803, 0xbfa0fffa,
+	0x8b6eff7b, 0x00160900,
+	0xbfa20015, 0x8b6eff7b,
+	0x000071ff, 0xbfa10008,
+	0x8b6fff7b, 0x00007080,
+	0xbfa10001, 0xbeee1287,
+	0xb8eff801, 0x846e8c6e,
+	0x8b6e6f6e, 0xbfa2000a,
+	0x8b6eff6d, 0x00ff0000,
+	0xbfa20007, 0xb8eef801,
+	0x8b6eff6e, 0x00000800,
+	0xbfa20003, 0x8b6eff7b,
+	0x00000400, 0xbfa2002a,
+	0xbefa4d82, 0xbf890000,
+	0x84fa887a, 0xbf0d8f7b,
+	0xbfa10002, 0x8c7bff7b,
+	0xffff0000, 0xf4005bbd,
+	0xf8000010, 0xbf890000,
+	0x846e976e, 0x9177ff77,
+	0x00800000, 0x8c776e77,
+	0xf4045bbd, 0xf8000000,
+	0xbf890000, 0xf4045ebd,
+	0xf8000008, 0xbf890000,
+	0x8bee6e6e, 0xbfa10001,
+	0xbe80486e, 0x8b6eff6d,
+	0x00ff0000, 0xbfa20008,
+	0x8b6eff6d, 0x01000000,
+	0xbfa20007, 0x8c78ff78,
+	0x00002000, 0x80ec886c,
+	0x82ed806d, 0xbfa00002,
+	0x806c846c, 0x826d806d,
+	0x8b6dff6d, 0x0000ffff,
+	0x8bfe7e7e, 0x8bea6a6a,
+	0xb978f802, 0xbe804a6c,
+	0xbf0d9878, 0xbfa10001,
+	0xbfb00000, 0x8b6dff6d,
+	0x0000ffff, 0xbefa0080,
+	0xb97a0283, 0xbeee007e,
+	0xbeef007f, 0xbefe0180,
+	0xbefe4d84, 0xbf890000,
+	0x8b7aff7f, 0x04000000,
+	0x847a857a, 0x8c6d7a6d,
+	0xbefa007e, 0x8b7bff7f,
+	0x0000ffff, 0xbefe00c1,
+	0xbeff00c1, 0xdca6c000,
+	0x007a0000, 0x7e000280,
+	0xbefe007a, 0xbeff007b,
+	0xb8fb02dc, 0x847b997b,
+	0xb8fa3b05, 0x807a817a,
+	0xbf0d997b, 0xbfa20002,
+	0x847a897a, 0xbfa00001,
+	0x847a8a7a, 0xb8fb1e06,
+	0x847b8a7b, 0x807a7b7a,
+	0x8b7bff7f, 0x0000ffff,
+	0x807aff7a, 0x00000200,
+	0x807a7e7a, 0x827b807b,
+	0xd7610000, 0x00010870,
+	0xd7610000, 0x00010a71,
+	0xd7610000, 0x00010c72,
+	0xd7610000, 0x00010e73,
+	0xd7610000, 0x00011074,
+	0xd7610000, 0x00011275,
+	0xd7610000, 0x00011476,
+	0xd7610000, 0x00011677,
+	0xd7610000, 0x00011a79,
+	0xd7610000, 0x00011c7e,
+	0xd7610000, 0x00011e7f,
+	0xbefe00ff, 0x00003fff,
+	0xbeff0080, 0xdca6c040,
+	0x007a0000, 0xd760007a,
+	0x00011d00, 0xd760007b,
+	0x00011f00, 0xbefe007a,
+	0xbeff007b, 0xbef4007e,
+	0x8b75ff7f, 0x0000ffff,
+	0x8c75ff75, 0x00040000,
+	0xbef60080, 0xbef700ff,
+	0x10807fac, 0xbef1007d,
+	0xbef00080, 0xb8f302dc,
+	0x84739973, 0xbefe00c1,
+	0x857d9973, 0x8b7d817d,
+	0xbf06817d, 0xbfa20002,
+	0xbeff0080, 0xbfa00002,
+	0xbeff00c1, 0xbfa00009,
+	0xbef600ff, 0x01000000,
+	0xe0685080, 0x701d0100,
+	0xe0685100, 0x701d0200,
+	0xe0685180, 0x701d0300,
+	0xbfa00008, 0xbef600ff,
+	0x01000000, 0xe0685100,
+	0x701d0100, 0xe0685200,
+	0x701d0200, 0xe0685300,
+	0x701d0300, 0xb8f03b05,
+	0x80708170, 0xbf0d9973,
+	0xbfa20002, 0x84708970,
+	0xbfa00001, 0x84708a70,
+	0xb8fa1e06, 0x847a8a7a,
+	0x80707a70, 0x8070ff70,
+	0x00000200, 0xbef600ff,
+	0x01000000, 0x7e000280,
+	0x7e020280, 0x7e040280,
+	0xbefd0080, 0xd7610002,
+	0x0000fa71, 0x807d817d,
+	0xd7610002, 0x0000fa6c,
+	0x807d817d, 0x917aff6d,
+	0x80000000, 0xd7610002,
+	0x0000fa7a, 0x807d817d,
+	0xd7610002, 0x0000fa6e,
+	0x807d817d, 0xd7610002,
+	0x0000fa6f, 0x807d817d,
+	0xd7610002, 0x0000fa78,
+	0x807d817d, 0xb8faf803,
+	0xd7610002, 0x0000fa7a,
+	0x807d817d, 0xd7610002,
+	0x0000fa7b, 0x807d817d,
+	0xb8f1f801, 0xd7610002,
+	0x0000fa71, 0x807d817d,
+	0xb8f1f814, 0xd7610002,
+	0x0000fa71, 0x807d817d,
+	0xb8f1f815, 0xd7610002,
+	0x0000fa71, 0x807d817d,
+	0xbefe00ff, 0x0000ffff,
+	0xbeff0080, 0xe0685000,
+	0x701d0200, 0xbefe00c1,
+	0xb8f03b05, 0x80708170,
+	0xbf0d9973, 0xbfa20002,
+	0x84708970, 0xbfa00001,
+	0x84708a70, 0xb8fa1e06,
+	0x847a8a7a, 0x80707a70,
+	0xbef600ff, 0x01000000,
+	0xbef90080, 0xbefd0080,
+	0xbf800000, 0xbe804100,
+	0xbe824102, 0xbe844104,
+	0xbe864106, 0xbe884108,
+	0xbe8a410a, 0xbe8c410c,
+	0xbe8e410e, 0xd7610002,
+	0x0000f200, 0x80798179,
+	0xd7610002, 0x0000f201,
+	0x80798179, 0xd7610002,
+	0x0000f202, 0x80798179,
+	0xd7610002, 0x0000f203,
+	0x80798179, 0xd7610002,
+	0x0000f204, 0x80798179,
+	0xd7610002, 0x0000f205,
+	0x80798179, 0xd7610002,
+	0x0000f206, 0x80798179,
+	0xd7610002, 0x0000f207,
+	0x80798179, 0xd7610002,
+	0x0000f208, 0x80798179,
+	0xd7610002, 0x0000f209,
+	0x80798179, 0xd7610002,
+	0x0000f20a, 0x80798179,
+	0xd7610002, 0x0000f20b,
+	0x80798179, 0xd7610002,
+	0x0000f20c, 0x80798179,
+	0xd7610002, 0x0000f20d,
+	0x80798179, 0xd7610002,
+	0x0000f20e, 0x80798179,
+	0xd7610002, 0x0000f20f,
+	0x80798179, 0xbf06a079,
+	0xbfa10006, 0xe0685000,
+	0x701d0200, 0x8070ff70,
+	0x00000080, 0xbef90080,
+	0x7e040280, 0x807d907d,
+	0xbf0aff7d, 0x00000060,
+	0xbfa2ffbc, 0xbe804100,
+	0xbe824102, 0xbe844104,
+	0xbe864106, 0xbe884108,
+	0xbe8a410a, 0xd7610002,
+	0x0000f200, 0x80798179,
+	0xd7610002, 0x0000f201,
+	0x80798179, 0xd7610002,
+	0x0000f202, 0x80798179,
+	0xd7610002, 0x0000f203,
+	0x80798179, 0xd7610002,
+	0x0000f204, 0x80798179,
+	0xd7610002, 0x0000f205,
+	0x80798179, 0xd7610002,
+	0x0000f206, 0x80798179,
+	0xd7610002, 0x0000f207,
+	0x80798179, 0xd7610002,
+	0x0000f208, 0x80798179,
+	0xd7610002, 0x0000f209,
+	0x80798179, 0xd7610002,
+	0x0000f20a, 0x80798179,
+	0xd7610002, 0x0000f20b,
+	0x80798179, 0xe0685000,
+	0x701d0200, 0xbefe00c1,
+	0x857d9973, 0x8b7d817d,
+	0xbf06817d, 0xbfa20002,
+	0xbeff0080, 0xbfa00001,
+	0xbeff00c1, 0xb8fb4306,
+	0x8b7bc17b, 0xbfa10043,
+	0xbfbd0000, 0x8b7aff6d,
+	0x80000000, 0xbfa1003f,
+	0x847b887b, 0xbef6007b,
+	0xb8f03b05, 0x80708170,
+	0xbf0d9973, 0xbfa20002,
+	0x84708970, 0xbfa00001,
+	0x84708a70, 0xb8fa1e06,
+	0x847a8a7a, 0x80707a70,
+	0x8070ff70, 0x00000200,
+	0x8070ff70, 0x00000080,
+	0xbef600ff, 0x01000000,
+	0xd71f0000, 0x000100c1,
+	0xd7200000, 0x000200c1,
+	0x16000084, 0x857d9973,
+	0x8b7d817d, 0xbf06817d,
+	0xbefd0080, 0xbfa20012,
+	0xbe8300ff, 0x00000080,
+	0xbf800000, 0xbf800000,
+	0xbf800000, 0xd8d80000,
+	0x01000000, 0xbf890000,
+	0xe0685000, 0x701d0100,
+	0x807d037d, 0x80700370,
+	0xd5250000, 0x0001ff00,
+	0x00000080, 0xbf0a7b7d,
+	0xbfa2fff4, 0xbfa00011,
+	0xbe8300ff, 0x00000100,
+	0xbf800000, 0xbf800000,
+	0xbf800000, 0xd8d80000,
+	0x01000000, 0xbf890000,
+	0xe0685000, 0x701d0100,
+	0x807d037d, 0x80700370,
+	0xd5250000, 0x0001ff00,
+	0x00000100, 0xbf0a7b7d,
+	0xbfa2fff4, 0xbefe00c1,
+	0x857d9973, 0x8b7d817d,
+	0xbf06817d, 0xbfa20004,
+	0xbef000ff, 0x00000200,
+	0xbeff0080, 0xbfa00003,
+	0xbef000ff, 0x00000400,
+	0xbeff00c1, 0xb8fb3b05,
+	0x807b817b, 0x847b827b,
+	0x857d9973, 0x8b7d817d,
+	0xbf06817d, 0xbfa20017,
+	0xbef600ff, 0x01000000,
+	0xbefd0084, 0xbf0a7b7d,
+	0xbfa10037, 0x7e008700,
+	0x7e028701, 0x7e048702,
+	0x7e068703, 0xe0685000,
+	0x701d0000, 0xe0685080,
+	0x701d0100, 0xe0685100,
+	0x701d0200, 0xe0685180,
+	0x701d0300, 0x807d847d,
+	0x8070ff70, 0x00000200,
+	0xbf0a7b7d, 0xbfa2ffef,
+	0xbfa00025, 0xbef600ff,
+	0x01000000, 0xbefd0084,
+	0xbf0a7b7d, 0xbfa10011,
+	0x7e008700, 0x7e028701,
+	0x7e048702, 0x7e068703,
+	0xe0685000, 0x701d0000,
+	0xe0685100, 0x701d0100,
+	0xe0685200, 0x701d0200,
+	0xe0685300, 0x701d0300,
+	0x807d847d, 0x8070ff70,
+	0x00000400, 0xbf0a7b7d,
+	0xbfa2ffef, 0xb8fb1e06,
+	0x8b7bc17b, 0xbfa1000c,
+	0x847b837b, 0x807b7d7b,
+	0xbefe00c1, 0xbeff0080,
+	0x7e008700, 0xe0685000,
+	0x701d0000, 0x807d817d,
+	0x8070ff70, 0x00000080,
+	0xbf0a7b7d, 0xbfa2fff8,
+	0xbfa00143, 0xbef4007e,
+	0x8b75ff7f, 0x0000ffff,
+	0x8c75ff75, 0x00040000,
+	0xbef60080, 0xbef700ff,
+	0x10807fac, 0xb8f202dc,
+	0x84729972, 0x8b6eff7f,
+	0x04000000, 0xbfa10039,
+	0xbefe00c1, 0x857d9972,
+	0x8b7d817d, 0xbf06817d,
+	0xbfa20002, 0xbeff0080,
+	0xbfa00001, 0xbeff00c1,
+	0xb8ef4306, 0x8b6fc16f,
+	0xbfa1002e, 0x846f886f,
+	0xbef6006f, 0xb8f83b05,
+	0x80788178, 0xbf0d9972,
+	0xbfa20002, 0x84788978,
+	0xbfa00001, 0x84788a78,
+	0xb8ee1e06, 0x846e8a6e,
+	0x80786e78, 0x8078ff78,
+	0x00000200, 0x8078ff78,
+	0x00000080, 0xbef600ff,
+	0x01000000, 0x857d9972,
+	0x8b7d817d, 0xbf06817d,
+	0xbefd0080, 0xbfa2000c,
+	0xe0500000, 0x781d0000,
+	0xbf890000, 0xdac00000,
+	0x00000000, 0x807dff7d,
+	0x00000080, 0x8078ff78,
+	0x00000080, 0xbf0a6f7d,
+	0xbfa2fff5, 0xbfa0000b,
+	0xe0500000, 0x781d0000,
+	0xbf890000, 0xdac00000,
+	0x00000000, 0x807dff7d,
+	0x00000100, 0x8078ff78,
+	0x00000100, 0xbf0a6f7d,
+	0xbfa2fff5, 0xbef80080,
+	0xbefe00c1, 0x857d9972,
+	0x8b7d817d, 0xbf06817d,
+	0xbfa20002, 0xbeff0080,
+	0xbfa00001, 0xbeff00c1,
+	0xb8ef3b05, 0x806f816f,
+	0x846f826f, 0x857d9972,
+	0x8b7d817d, 0xbf06817d,
+	0xbfa20024, 0xbef600ff,
+	0x01000000, 0xbeee0078,
+	0x8078ff78, 0x00000200,
+	0xbefd0084, 0xbf0a6f7d,
+	0xbfa10050, 0xe0505000,
+	0x781d0000, 0xe0505080,
+	0x781d0100, 0xe0505100,
+	0x781d0200, 0xe0505180,
+	0x781d0300, 0xbf890000,
+	0x7e008500, 0x7e028501,
+	0x7e048502, 0x7e068503,
+	0x807d847d, 0x8078ff78,
+	0x00000200, 0xbf0a6f7d,
+	0xbfa2ffee, 0xe0505000,
+	0x6e1d0000, 0xe0505080,
+	0x6e1d0100, 0xe0505100,
+	0x6e1d0200, 0xe0505180,
+	0x6e1d0300, 0xbf890000,
+	0xbfa00034, 0xbef600ff,
+	0x01000000, 0xbeee0078,
+	0x8078ff78, 0x00000400,
+	0xbefd0084, 0xbf0a6f7d,
+	0xbfa10012, 0xe0505000,
+	0x781d0000, 0xe0505100,
+	0x781d0100, 0xe0505200,
+	0x781d0200, 0xe0505300,
+	0x781d0300, 0xbf890000,
+	0x7e008500, 0x7e028501,
+	0x7e048502, 0x7e068503,
+	0x807d847d, 0x8078ff78,
+	0x00000400, 0xbf0a6f7d,
+	0xbfa2ffee, 0xb8ef1e06,
+	0x8b6fc16f, 0xbfa1000e,
+	0x846f836f, 0x806f7d6f,
+	0xbefe00c1, 0xbeff0080,
+	0xe0505000, 0x781d0000,
+	0xbf890000, 0x7e008500,
+	0x807d817d, 0x8078ff78,
+	0x00000080, 0xbf0a6f7d,
+	0xbfa2fff7, 0xbeff00c1,
+	0xe0505000, 0x6e1d0000,
+	0xe0505100, 0x6e1d0100,
+	0xe0505200, 0x6e1d0200,
+	0xe0505300, 0x6e1d0300,
+	0xbf890000, 0xb8f83b05,
+	0x80788178, 0xbf0d9972,
+	0xbfa20002, 0x84788978,
+	0xbfa00001, 0x84788a78,
+	0xb8ee1e06, 0x846e8a6e,
+	0x80786e78, 0x8078ff78,
+	0x00000200, 0x80f8ff78,
+	0x00000050, 0xbef600ff,
+	0x01000000, 0xbefd00ff,
+	0x0000006c, 0x80f89078,
+	0xf428403a, 0xf0000000,
+	0xbf890000, 0x80fd847d,
+	0xbf800000, 0xbe804300,
+	0xbe824302, 0x80f8a078,
+	0xf42c403a, 0xf0000000,
+	0xbf890000, 0x80fd887d,
+	0xbf800000, 0xbe804300,
+	0xbe824302, 0xbe844304,
+	0xbe864306, 0x80f8c078,
+	0xf430403a, 0xf0000000,
+	0xbf890000, 0x80fd907d,
+	0xbf800000, 0xbe804300,
+	0xbe824302, 0xbe844304,
+	0xbe864306, 0xbe884308,
+	0xbe8a430a, 0xbe8c430c,
+	0xbe8e430e, 0xbf06807d,
+	0xbfa1fff0, 0xb980f801,
+	0x00000000, 0xbfbd0000,
+	0xb8f83b05, 0x80788178,
+	0xbf0d9972, 0xbfa20002,
+	0x84788978, 0xbfa00001,
+	0x84788a78, 0xb8ee1e06,
+	0x846e8a6e, 0x80786e78,
+	0x8078ff78, 0x00000200,
+	0xbef600ff, 0x01000000,
+	0xf4205bfa, 0xf0000000,
+	0x80788478, 0xf4205b3a,
+	0xf0000000, 0x80788478,
+	0xf4205b7a, 0xf0000000,
+	0x80788478, 0xf4205c3a,
+	0xf0000000, 0x80788478,
+	0xf4205c7a, 0xf0000000,
+	0x80788478, 0xf4205eba,
+	0xf0000000, 0x80788478,
+	0xf4205efa, 0xf0000000,
+	0x80788478, 0xf4205e7a,
+	0xf0000000, 0x80788478,
+	0xf4205cfa, 0xf0000000,
+	0x80788478, 0xf4205bba,
+	0xf0000000, 0x80788478,
+	0xbf890000, 0xb96ef814,
+	0xf4205bba, 0xf0000000,
+	0x80788478, 0xbf890000,
+	0xb96ef815, 0xbefd006f,
+	0xbefe0070, 0xbeff0071,
+	0xb97b4803, 0x857b8b7b,
+	0xb97b22c3, 0x857b867b,
+	0xb97b7443, 0xb973f801,
+	0xb8ee3b05, 0x806e816e,
+	0xbf0d9972, 0xbfa20002,
+	0x846e896e, 0xbfa00001,
+	0x846e8a6e, 0xb8ef1e06,
+	0x846f8a6f, 0x806e6f6e,
 	0x806eff6e, 0x00000200,
 	0x806e746e, 0x826f8075,
-	0x876fff6f, 0x0000ffff,
-	0xf4091c37, 0xfa000050,
-	0xf4091d37, 0xfa000060,
-	0xf4011e77, 0xfa000074,
-	0xbf8cc07f, 0x876dff6d,
-	0x0000ffff, 0x87fe7e7e,
-	0x87ea6a6a, 0xb9faf802,
-	0xbe80226c, 0xbf810000,
+	0x8b6fff6f, 0x0000ffff,
+	0xf4085c37, 0xf8000050,
+	0xf4085d37, 0xf8000060,
+	0xf4005e77, 0xf8000074,
+	0xbf890000, 0x8b6dff6d,
+	0x0000ffff, 0x8bfe7e7e,
+	0x8bea6a6a, 0xb8eef802,
+	0xbf0d866e, 0xbfa20002,
+	0xb97af802, 0xbe80486c,
+	0xb97af802, 0xbe804a6c,
+	0xbfb10000, 0xbf9f0000,
+	0xbf9f0000, 0xbf9f0000,
+	0xbf9f0000, 0xbf9f0000,
+};
+
+static const uint32_t cwsr_trap_gfx9_4_3_hex[] = {
+	0xbf820001, 0xbf8202dc,
+	0xb8f8f802, 0x8978ff78,
+	0x00020006, 0xb8fbf803,
+	0x866eff78, 0x00002000,
+	0xbf840009, 0x866eff6d,
+	0x00ff0000, 0xbf85001a,
+	0x866eff7b, 0x00000400,
+	0xbf850051, 0xbf8e0010,
+	0xb8fbf803, 0xbf82fffa,
+	0x866eff7b, 0x03c00900,
+	0xbf850011, 0x866eff7b,
+	0x000071ff, 0xbf840008,
+	0x866fff7b, 0x00007080,
+	0xbf840001, 0xbeee1a87,
+	0xb8eff801, 0x8e6e8c6e,
+	0x866e6f6e, 0xbf850006,
+	0x866eff6d, 0x00ff0000,
+	0xbf850003, 0x866eff7b,
+	0x00000400, 0xbf85003a,
+	0xb8faf807, 0x867aff7a,
+	0x001f8000, 0x8e7a8b7a,
+	0x8979ff79, 0xfc000000,
+	0x87797a79, 0xba7ff807,
+	0x00000000, 0xb8faf812,
+	0xb8fbf813, 0x8efa887a,
+	0xbf0d8f7b, 0xbf840002,
+	0x877bff7b, 0xffff0000,
+	0xc0031bbd, 0x00000010,
+	0xbf8cc07f, 0x8e6e976e,
+	0x8979ff79, 0x00800000,
+	0x87796e79, 0xc0071bbd,
+	0x00000000, 0xbf8cc07f,
+	0xc0071ebd, 0x00000008,
+	0xbf8cc07f, 0x86ee6e6e,
+	0xbf840001, 0xbe801d6e,
+	0x866eff6d, 0x01ff0000,
+	0xbf850005, 0x8778ff78,
+	0x00002000, 0x80ec886c,
+	0x82ed806d, 0xbf820005,
+	0x866eff6d, 0x01000000,
+	0xbf850002, 0x806c846c,
+	0x826d806d, 0x866dff6d,
+	0x0000ffff, 0x8f7a8b79,
+	0x867aff7a, 0x001f8000,
+	0xb97af807, 0x86fe7e7e,
+	0x86ea6a6a, 0x8f6e8378,
+	0xb96ee0c2, 0xbf800002,
+	0xb9780002, 0xbe801f6c,
+	0x866dff6d, 0x0000ffff,
+	0xbefa0080, 0xb97a0283,
+	0xb8faf807, 0x867aff7a,
+	0x001f8000, 0x8e7a8b7a,
+	0x8979ff79, 0xfc000000,
+	0x87797a79, 0xba7ff807,
+	0x00000000, 0xbeee007e,
+	0xbeef007f, 0xbefe0180,
+	0xbf900004, 0x877a8478,
+	0xb97af802, 0xbf8e0002,
+	0xbf88fffe, 0xb8fa2985,
+	0x807a817a, 0x8e7a8a7a,
+	0x8e7a817a, 0xb8fb1605,
+	0x807b817b, 0x8e7b867b,
+	0x807a7b7a, 0x807a7e7a,
+	0x827b807f, 0x867bff7b,
+	0x0000ffff, 0xc04b1c3d,
+	0x00000050, 0xbf8cc07f,
+	0xc04b1d3d, 0x00000060,
+	0xbf8cc07f, 0xc0431e7d,
+	0x00000074, 0xbf8cc07f,
+	0xbef4007e, 0x8675ff7f,
+	0x0000ffff, 0x8775ff75,
+	0x00040000, 0xbef60080,
+	0xbef700ff, 0x00807fac,
+	0xbef1007c, 0xbef00080,
+	0xb8f02985, 0x80708170,
+	0x8e708a70, 0x8e708170,
+	0xb8fa1605, 0x807a817a,
+	0x8e7a867a, 0x80707a70,
+	0xbef60084, 0xbef600ff,
+	0x01000000, 0xbefe007c,
+	0xbefc0070, 0xc0611c7a,
+	0x0000007c, 0xbf8cc07f,
+	0x80708470, 0xbefc007e,
+	0xbefe007c, 0xbefc0070,
+	0xc0611b3a, 0x0000007c,
+	0xbf8cc07f, 0x80708470,
+	0xbefc007e, 0xbefe007c,
+	0xbefc0070, 0xc0611b7a,
+	0x0000007c, 0xbf8cc07f,
+	0x80708470, 0xbefc007e,
+	0xbefe007c, 0xbefc0070,
+	0xc0611bba, 0x0000007c,
+	0xbf8cc07f, 0x80708470,
+	0xbefc007e, 0xbefe007c,
+	0xbefc0070, 0xc0611bfa,
+	0x0000007c, 0xbf8cc07f,
+	0x80708470, 0xbefc007e,
+	0xbefe007c, 0xbefc0070,
+	0xc0611e3a, 0x0000007c,
+	0xbf8cc07f, 0x80708470,
+	0xbefc007e, 0xb8fbf803,
+	0xbefe007c, 0xbefc0070,
+	0xc0611efa, 0x0000007c,
+	0xbf8cc07f, 0x80708470,
+	0xbefc007e, 0xbefe007c,
+	0xbefc0070, 0xc0611a3a,
+	0x0000007c, 0xbf8cc07f,
+	0x80708470, 0xbefc007e,
+	0xbefe007c, 0xbefc0070,
+	0xc0611a7a, 0x0000007c,
+	0xbf8cc07f, 0x80708470,
+	0xbefc007e, 0xb8f1f801,
+	0xbefe007c, 0xbefc0070,
+	0xc0611c7a, 0x0000007c,
+	0xbf8cc07f, 0x80708470,
+	0xbefc007e, 0xbf108080,
+	0x867aff7f, 0x04000000,
+	0xbeef0080, 0x876f6f7a,
+	0xb8f02985, 0x80708170,
+	0x8e708a70, 0x8e708170,
+	0xb8fb1605, 0x807b817b,
+	0x8e7b847b, 0x8e76827b,
+	0xbef600ff, 0x01000000,
+	0xbef20174, 0x80747074,
+	0x82758075, 0xbefc0080,
+	0xbf800000, 0xbe802b00,
+	0xbe822b02, 0xbe842b04,
+	0xbe862b06, 0xbe882b08,
+	0xbe8a2b0a, 0xbe8c2b0c,
+	0xbe8e2b0e, 0xc06b003a,
+	0x00000000, 0xbf8cc07f,
+	0xc06b013a, 0x00000010,
+	0xbf8cc07f, 0xc06b023a,
+	0x00000020, 0xbf8cc07f,
+	0xc06b033a, 0x00000030,
+	0xbf8cc07f, 0x8074c074,
+	0x82758075, 0x807c907c,
+	0xbf0a7b7c, 0xbf85ffe7,
+	0xbef40172, 0xbef00080,
+	0xbefe00c1, 0xbeff00c1,
+	0xbee80080, 0xbee90080,
+	0xbef600ff, 0x01000000,
+	0x867aff78, 0x00400000,
+	0xbf850003, 0xb8faf803,
+	0x897a7aff, 0x10000000,
+	0xbf85004d, 0xbe840080,
+	0xd2890000, 0x00000900,
+	0x80048104, 0xd2890001,
+	0x00000900, 0x80048104,
+	0xd2890002, 0x00000900,
+	0x80048104, 0xd2890003,
+	0x00000900, 0x80048104,
+	0xc069003a, 0x00000070,
+	0xbf8cc07f, 0x80709070,
+	0xbf06c004, 0xbf84ffee,
+	0xbe840080, 0xd2890000,
+	0x00000901, 0x80048104,
+	0xd2890001, 0x00000901,
+	0x80048104, 0xd2890002,
+	0x00000901, 0x80048104,
+	0xd2890003, 0x00000901,
+	0x80048104, 0xc069003a,
+	0x00000070, 0xbf8cc07f,
+	0x80709070, 0xbf06c004,
+	0xbf84ffee, 0xbe840080,
+	0xd2890000, 0x00000902,
+	0x80048104, 0xd2890001,
+	0x00000902, 0x80048104,
+	0xd2890002, 0x00000902,
+	0x80048104, 0xd2890003,
+	0x00000902, 0x80048104,
+	0xc069003a, 0x00000070,
+	0xbf8cc07f, 0x80709070,
+	0xbf06c004, 0xbf84ffee,
+	0xbe840080, 0xd2890000,
+	0x00000903, 0x80048104,
+	0xd2890001, 0x00000903,
+	0x80048104, 0xd2890002,
+	0x00000903, 0x80048104,
+	0xd2890003, 0x00000903,
+	0x80048104, 0xc069003a,
+	0x00000070, 0xbf8cc07f,
+	0x80709070, 0xbf06c004,
+	0xbf84ffee, 0xbf820008,
+	0xe0724000, 0x701d0000,
+	0xe0724100, 0x701d0100,
+	0xe0724200, 0x701d0200,
+	0xe0724300, 0x701d0300,
+	0xbefe00c1, 0xbeff00c1,
+	0xb8fb4306, 0x867bc17b,
+	0xbf840064, 0xbf8a0000,
+	0x867aff6f, 0x04000000,
+	0xbf840060, 0x8e7b867b,
+	0x8e7b827b, 0xbef6007b,
+	0xb8f02985, 0x80708170,
+	0x8e708a70, 0x8e708170,
+	0xb8fa1605, 0x807a817a,
+	0x8e7a867a, 0x80707a70,
+	0x8070ff70, 0x00000080,
+	0xbef600ff, 0x01000000,
+	0xbefc0080, 0xd28c0002,
+	0x000100c1, 0xd28d0003,
+	0x000204c1, 0x867aff78,
+	0x00400000, 0xbf850003,
+	0xb8faf803, 0x897a7aff,
+	0x10000000, 0xbf850030,
+	0x24040682, 0xd86e4000,
+	0x00000002, 0xbf8cc07f,
+	0xbe840080, 0xd2890000,
+	0x00000900, 0x80048104,
+	0xd2890001, 0x00000900,
+	0x80048104, 0xd2890002,
+	0x00000900, 0x80048104,
+	0xd2890003, 0x00000900,
+	0x80048104, 0xc069003a,
+	0x00000070, 0xbf8cc07f,
+	0x80709070, 0xbf06c004,
+	0xbf84ffee, 0xbe840080,
+	0xd2890000, 0x00000901,
+	0x80048104, 0xd2890001,
+	0x00000901, 0x80048104,
+	0xd2890002, 0x00000901,
+	0x80048104, 0xd2890003,
+	0x00000901, 0x80048104,
+	0xc069003a, 0x00000070,
+	0xbf8cc07f, 0x80709070,
+	0xbf06c004, 0xbf84ffee,
+	0x680404ff, 0x00000200,
+	0xd0c9006a, 0x0000f702,
+	0xbf87ffd2, 0xbf820015,
+	0xd1060002, 0x00011103,
+	0x7e0602ff, 0x00000200,
+	0xbefc00ff, 0x00010000,
+	0xbe800077, 0x8677ff77,
+	0xff7fffff, 0x8777ff77,
+	0x00058000, 0xd8ec0000,
+	0x00000002, 0xbf8cc07f,
+	0xe0765000, 0x701d0002,
+	0x68040702, 0xd0c9006a,
+	0x0000f702, 0xbf87fff7,
+	0xbef70000, 0xbef000ff,
+	0x00000400, 0xbefe00c1,
+	0xbeff00c1, 0xb8fb2b05,
+	0x807b817b, 0x8e7b827b,
+	0xbef600ff, 0x01000000,
+	0xbefc0084, 0xbf0a7b7c,
+	0xbf84006d, 0xbf11017c,
+	0x807bff7b, 0x00001000,
+	0x867aff78, 0x00400000,
+	0xbf850003, 0xb8faf803,
+	0x897a7aff, 0x10000000,
+	0xbf850051, 0xbe840080,
+	0xd2890000, 0x00000900,
+	0x80048104, 0xd2890001,
+	0x00000900, 0x80048104,
+	0xd2890002, 0x00000900,
+	0x80048104, 0xd2890003,
+	0x00000900, 0x80048104,
+	0xc069003a, 0x00000070,
+	0xbf8cc07f, 0x80709070,
+	0xbf06c004, 0xbf84ffee,
+	0xbe840080, 0xd2890000,
+	0x00000901, 0x80048104,
+	0xd2890001, 0x00000901,
+	0x80048104, 0xd2890002,
+	0x00000901, 0x80048104,
+	0xd2890003, 0x00000901,
+	0x80048104, 0xc069003a,
+	0x00000070, 0xbf8cc07f,
+	0x80709070, 0xbf06c004,
+	0xbf84ffee, 0xbe840080,
+	0xd2890000, 0x00000902,
+	0x80048104, 0xd2890001,
+	0x00000902, 0x80048104,
+	0xd2890002, 0x00000902,
+	0x80048104, 0xd2890003,
+	0x00000902, 0x80048104,
+	0xc069003a, 0x00000070,
+	0xbf8cc07f, 0x80709070,
+	0xbf06c004, 0xbf84ffee,
+	0xbe840080, 0xd2890000,
+	0x00000903, 0x80048104,
+	0xd2890001, 0x00000903,
+	0x80048104, 0xd2890002,
+	0x00000903, 0x80048104,
+	0xd2890003, 0x00000903,
+	0x80048104, 0xc069003a,
+	0x00000070, 0xbf8cc07f,
+	0x80709070, 0xbf06c004,
+	0xbf84ffee, 0x807c847c,
+	0xbf0a7b7c, 0xbf85ffb1,
+	0xbf9c0000, 0xbf820012,
+	0x7e000300, 0x7e020301,
+	0x7e040302, 0x7e060303,
+	0xe0724000, 0x701d0000,
+	0xe0724100, 0x701d0100,
+	0xe0724200, 0x701d0200,
+	0xe0724300, 0x701d0300,
+	0x807c847c, 0x8070ff70,
+	0x00000400, 0xbf0a7b7c,
+	0xbf85ffef, 0xbf9c0000,
+	0xb8fb2985, 0x807b817b,
+	0x8e7b837b, 0xb8fa2b05,
+	0x807a817a, 0x8e7a827a,
+	0x80fb7a7b, 0x867b7b7b,
+	0xbf84007a, 0x807bff7b,
+	0x00001000, 0xbefc0080,
+	0xbf11017c, 0x867aff78,
+	0x00400000, 0xbf850003,
+	0xb8faf803, 0x897a7aff,
+	0x10000000, 0xbf850059,
+	0xd3d84000, 0x18000100,
+	0xd3d84001, 0x18000101,
+	0xd3d84002, 0x18000102,
+	0xd3d84003, 0x18000103,
+	0xbe840080, 0xd2890000,
+	0x00000900, 0x80048104,
+	0xd2890001, 0x00000900,
+	0x80048104, 0xd2890002,
+	0x00000900, 0x80048104,
+	0xd2890003, 0x00000900,
+	0x80048104, 0xc069003a,
+	0x00000070, 0xbf8cc07f,
+	0x80709070, 0xbf06c004,
+	0xbf84ffee, 0xbe840080,
+	0xd2890000, 0x00000901,
+	0x80048104, 0xd2890001,
+	0x00000901, 0x80048104,
+	0xd2890002, 0x00000901,
+	0x80048104, 0xd2890003,
+	0x00000901, 0x80048104,
+	0xc069003a, 0x00000070,
+	0xbf8cc07f, 0x80709070,
+	0xbf06c004, 0xbf84ffee,
+	0xbe840080, 0xd2890000,
+	0x00000902, 0x80048104,
+	0xd2890001, 0x00000902,
+	0x80048104, 0xd2890002,
+	0x00000902, 0x80048104,
+	0xd2890003, 0x00000902,
+	0x80048104, 0xc069003a,
+	0x00000070, 0xbf8cc07f,
+	0x80709070, 0xbf06c004,
+	0xbf84ffee, 0xbe840080,
+	0xd2890000, 0x00000903,
+	0x80048104, 0xd2890001,
+	0x00000903, 0x80048104,
+	0xd2890002, 0x00000903,
+	0x80048104, 0xd2890003,
+	0x00000903, 0x80048104,
+	0xc069003a, 0x00000070,
+	0xbf8cc07f, 0x80709070,
+	0xbf06c004, 0xbf84ffee,
+	0x807c847c, 0xbf0a7b7c,
+	0xbf85ffa9, 0xbf9c0000,
+	0xbf820016, 0xd3d84000,
+	0x18000100, 0xd3d84001,
+	0x18000101, 0xd3d84002,
+	0x18000102, 0xd3d84003,
+	0x18000103, 0xe0724000,
+	0x701d0000, 0xe0724100,
+	0x701d0100, 0xe0724200,
+	0x701d0200, 0xe0724300,
+	0x701d0300, 0x807c847c,
+	0x8070ff70, 0x00000400,
+	0xbf0a7b7c, 0xbf85ffeb,
+	0xbf9c0000, 0xbf8200ee,
+	0xbef4007e, 0x8675ff7f,
+	0x0000ffff, 0x8775ff75,
+	0x00040000, 0xbef60080,
+	0xbef700ff, 0x00807fac,
+	0x866eff7f, 0x04000000,
+	0xbf84001f, 0xbefe00c1,
+	0xbeff00c1, 0xb8ef4306,
+	0x866fc16f, 0xbf84001a,
+	0x8e6f866f, 0x8e6f826f,
+	0xbef6006f, 0xb8f82985,
+	0x80788178, 0x8e788a78,
+	0x8e788178, 0xb8ee1605,
+	0x806e816e, 0x8e6e866e,
+	0x80786e78, 0x8078ff78,
+	0x00000080, 0xbef600ff,
+	0x01000000, 0xbefc0080,
+	0xe0510000, 0x781d0000,
+	0xe0510100, 0x781d0000,
+	0x807cff7c, 0x00000200,
+	0x8078ff78, 0x00000200,
+	0xbf0a6f7c, 0xbf85fff6,
+	0xbefe00c1, 0xbeff00c1,
+	0xbef600ff, 0x01000000,
+	0xb8ef2b05, 0x806f816f,
+	0x8e6f826f, 0x806fff6f,
+	0x00008000, 0xbef80080,
+	0xbeee0078, 0x8078ff78,
+	0x00000400, 0xbefc0084,
+	0xbf11087c, 0xe0524000,
+	0x781d0000, 0xe0524100,
+	0x781d0100, 0xe0524200,
+	0x781d0200, 0xe0524300,
+	0x781d0300, 0xbf8c0f70,
+	0x7e000300, 0x7e020301,
+	0x7e040302, 0x7e060303,
+	0x807c847c, 0x8078ff78,
+	0x00000400, 0xbf0a6f7c,
+	0xbf85ffee, 0xb8ef2985,
+	0x806f816f, 0x8e6f836f,
+	0xb8f92b05, 0x80798179,
+	0x8e798279, 0x80ef796f,
+	0x866f6f6f, 0xbf84001a,
+	0x806fff6f, 0x00008000,
+	0xbefc0080, 0xbf11087c,
+	0xe0524000, 0x781d0000,
+	0xe0524100, 0x781d0100,
+	0xe0524200, 0x781d0200,
+	0xe0524300, 0x781d0300,
+	0xbf8c0f70, 0xd3d94000,
+	0x18000100, 0xd3d94001,
+	0x18000101, 0xd3d94002,
+	0x18000102, 0xd3d94003,
+	0x18000103, 0x807c847c,
+	0x8078ff78, 0x00000400,
+	0xbf0a6f7c, 0xbf85ffea,
+	0xbf9c0000, 0xe0524000,
+	0x6e1d0000, 0xe0524100,
+	0x6e1d0100, 0xe0524200,
+	0x6e1d0200, 0xe0524300,
+	0x6e1d0300, 0xbf8c0f70,
+	0xb8f82985, 0x80788178,
+	0x8e788a78, 0x8e788178,
+	0xb8ee1605, 0x806e816e,
+	0x8e6e866e, 0x80786e78,
+	0x80f8c078, 0xb8ef1605,
+	0x806f816f, 0x8e6f846f,
+	0x8e76826f, 0xbef600ff,
+	0x01000000, 0xbefc006f,
+	0xc031003a, 0x00000078,
+	0x80f8c078, 0xbf8cc07f,
+	0x80fc907c, 0xbf800000,
+	0xbe802d00, 0xbe822d02,
+	0xbe842d04, 0xbe862d06,
+	0xbe882d08, 0xbe8a2d0a,
+	0xbe8c2d0c, 0xbe8e2d0e,
+	0xbf06807c, 0xbf84fff0,
+	0xb8f82985, 0x80788178,
+	0x8e788a78, 0x8e788178,
+	0xb8ee1605, 0x806e816e,
+	0x8e6e866e, 0x80786e78,
+	0xbef60084, 0xbef600ff,
+	0x01000000, 0xc0211bfa,
+	0x00000078, 0x80788478,
+	0xc0211b3a, 0x00000078,
+	0x80788478, 0xc0211b7a,
+	0x00000078, 0x80788478,
+	0xc0211c3a, 0x00000078,
+	0x80788478, 0xc0211c7a,
+	0x00000078, 0x80788478,
+	0xc0211eba, 0x00000078,
+	0x80788478, 0xc0211efa,
+	0x00000078, 0x80788478,
+	0xc0211a3a, 0x00000078,
+	0x80788478, 0xc0211a7a,
+	0x00000078, 0x80788478,
+	0xc0211cfa, 0x00000078,
+	0x80788478, 0xbf8cc07f,
+	0xbefc006f, 0xbefe0070,
+	0xbeff0071, 0x866f7bff,
+	0x000003ff, 0xb96f4803,
+	0x866f7bff, 0xfffff800,
+	0x8f6f8b6f, 0xb96fa2c3,
+	0xb973f801, 0xb8ee2985,
+	0x806e816e, 0x8e6e8a6e,
+	0x8e6e816e, 0xb8ef1605,
+	0x806f816f, 0x8e6f866f,
+	0x806e6f6e, 0x806e746e,
+	0x826f8075, 0x866fff6f,
+	0x0000ffff, 0xc00b1c37,
+	0x00000050, 0xc00b1d37,
+	0x00000060, 0xc0031e77,
+	0x00000074, 0xbf8cc07f,
+	0x8f6e8b79, 0x866eff6e,
+	0x001f8000, 0xb96ef807,
+	0x866dff6d, 0x0000ffff,
+	0x86fe7e7e, 0x86ea6a6a,
+	0x8f6e837a, 0xb96ee0c2,
+	0xbf800002, 0xb97a0002,
+	0xbf8a0000, 0xbe801f6c,
+	0xbf9b0000, 0x00000000,
+};
+
+static const uint32_t cwsr_trap_gfx12_hex[] = {
+	0xbfa00001, 0xbfa002a2,
+	0xb0804009, 0xb8f8f804,
+	0x9178ff78, 0x00008c00,
+	0xb8fbf811, 0x8b6eff78,
+	0x00004000, 0xbfa10008,
+	0x8b6eff7b, 0x00000080,
+	0xbfa20018, 0x8b6ea07b,
+	0xbfa20042, 0xbf830010,
+	0xb8fbf811, 0xbfa0fffb,
+	0x8b6eff7b, 0x00000bd0,
+	0xbfa20010, 0xb8eef812,
+	0x8b6f8f7b, 0xbfa10002,
+	0x8c6eff6e, 0x00000080,
+	0xb8eff813, 0x8b6e6e6f,
+	0xbfa20008, 0x8b6eff6d,
+	0xf0000000, 0xbfa20005,
+	0x8b6fff6f, 0x00000200,
+	0xbfa20002, 0x8b6ea07b,
+	0xbfa2002c, 0xbefa4d82,
+	0xbf8a0000, 0x84fa887a,
+	0xbf0d8f7b, 0xbfa10002,
+	0x8c7bff7b, 0xffff0000,
+	0xf4601bbd, 0xf8000010,
+	0xbf8a0000, 0x846e976e,
+	0x9177ff77, 0x00800000,
+	0x8c776e77, 0xf4603bbd,
+	0xf8000000, 0xbf8a0000,
+	0xf4603ebd, 0xf8000008,
+	0xbf8a0000, 0x8bee6e6e,
+	0xbfa10001, 0xbe80486e,
+	0x8b6eff6d, 0xf0000000,
+	0xbfa20009, 0xb8eef811,
+	0x8b6eff6e, 0x00000080,
+	0xbfa20007, 0x8c78ff78,
+	0x00004000, 0x80ec886c,
+	0x82ed806d, 0xbfa00002,
+	0x806c846c, 0x826d806d,
+	0x8b6dff6d, 0x0000ffff,
+	0x8bfe7e7e, 0x8bea6a6a,
+	0x85788978, 0xb9783244,
+	0xbe804a6c, 0xb8faf802,
+	0xbf0d987a, 0xbfa10001,
+	0xbfb00000, 0x8b6dff6d,
+	0x0000ffff, 0xbefa0080,
+	0xb97a0151, 0xbeee007e,
+	0xbeef007f, 0xbefe0180,
+	0xbefe4d84, 0xbf8a0000,
+	0x8b7aff7f, 0x04000000,
+	0x847a857a, 0x8c6d7a6d,
+	0xbefa007e, 0x8b7bff7f,
+	0x0000ffff, 0xbefe00c1,
+	0xbeff00c1, 0xee0a407a,
+	0x000c0000, 0x00000000,
+	0x7e000280, 0xbefe007a,
+	0xbeff007b, 0xb8fb0742,
+	0x847b997b, 0xb8fa3b05,
+	0x807a817a, 0xbf0d997b,
+	0xbfa20002, 0x847a897a,
+	0xbfa00001, 0x847a8a7a,
+	0xb8fb1e06, 0x847b8a7b,
+	0x807a7b7a, 0x8b7bff7f,
+	0x0000ffff, 0x807aff7a,
+	0x00000200, 0x807a7e7a,
+	0x827b807b, 0xd7610000,
+	0x00010870, 0xd7610000,
+	0x00010a71, 0xd7610000,
+	0x00010c72, 0xd7610000,
+	0x00010e73, 0xd7610000,
+	0x00011074, 0xd7610000,
+	0x00011275, 0xd7610000,
+	0x00011476, 0xd7610000,
+	0x00011677, 0xd7610000,
+	0x00011a79, 0xd7610000,
+	0x00011c7e, 0xd7610000,
+	0x00011e7f, 0xd8500000,
+	0x00000000, 0xd8500000,
+	0x00000000, 0xd8500000,
+	0x00000000, 0xd8500000,
+	0x00000000, 0xd8500000,
+	0x00000000, 0xd8500000,
+	0x00000000, 0xd8500000,
+	0x00000000, 0xd8500000,
+	0x00000000, 0xbefe00ff,
+	0x00003fff, 0xbeff0080,
+	0xee0a407a, 0x000c0000,
+	0x00004000, 0xd760007a,
+	0x00011d00, 0xd760007b,
+	0x00011f00, 0xbefe007a,
+	0xbeff007b, 0xbef4007e,
+	0x8b75ff7f, 0x0000ffff,
+	0x8c75ff75, 0x00040000,
+	0xbef60080, 0xbef700ff,
+	0x10807fac, 0xbef1007d,
+	0xbef00080, 0xb8f30742,
+	0x84739973, 0xbefe00c1,
+	0x857d9973, 0x8b7d817d,
+	0xbf06817d, 0xbfa20002,
+	0xbeff0080, 0xbfa00002,
+	0xbeff00c1, 0xbfa0000c,
+	0xbef600ff, 0x01000000,
+	0xc4068070, 0x008ce801,
+	0x00008000, 0xc4068070,
+	0x008ce802, 0x00010000,
+	0xc4068070, 0x008ce803,
+	0x00018000, 0xbfa0000b,
+	0xbef600ff, 0x01000000,
+	0xc4068070, 0x008ce801,
+	0x00010000, 0xc4068070,
+	0x008ce802, 0x00020000,
+	0xc4068070, 0x008ce803,
+	0x00030000, 0xb8f03b05,
+	0x80708170, 0xbf0d9973,
+	0xbfa20002, 0x84708970,
+	0xbfa00001, 0x84708a70,
+	0xb8fa1e06, 0x847a8a7a,
+	0x80707a70, 0x8070ff70,
+	0x00000200, 0xbef600ff,
+	0x01000000, 0x7e000280,
+	0x7e020280, 0x7e040280,
+	0xbe804ec2, 0xbf94fffe,
+	0xb8faf804, 0x8b7a847a,
+	0x91788478, 0x8c787a78,
+	0x917aff6d, 0x80000000,
+	0xd7610002, 0x00010071,
+	0xd7610002, 0x0001026c,
+	0xd7610002, 0x0001047a,
+	0xd7610002, 0x0001066e,
+	0xd7610002, 0x0001086f,
+	0xd7610002, 0x00010a78,
+	0xd7610002, 0x00010e7b,
+	0xd8500000, 0x00000000,
+	0xd8500000, 0x00000000,
+	0xd8500000, 0x00000000,
+	0xd8500000, 0x00000000,
+	0xd8500000, 0x00000000,
+	0xd8500000, 0x00000000,
+	0xd8500000, 0x00000000,
+	0xd8500000, 0x00000000,
+	0xb8faf811, 0xd7610002,
+	0x00010c7a, 0xb8faf801,
+	0xd7610002, 0x0001107a,
+	0xb8faf814, 0xd7610002,
+	0x0001127a, 0xb8faf815,
+	0xd7610002, 0x0001147a,
+	0xb8faf812, 0xd7610002,
+	0x0001167a, 0xb8faf813,
+	0xd7610002, 0x0001187a,
+	0xb8faf802, 0xd7610002,
+	0x00011a7a, 0xbefa50c1,
+	0xbfc70000, 0xd7610002,
+	0x00011c7a, 0xd8500000,
+	0x00000000, 0xd8500000,
+	0x00000000, 0xd8500000,
+	0x00000000, 0xd8500000,
+	0x00000000, 0xd8500000,
+	0x00000000, 0xd8500000,
+	0x00000000, 0xd8500000,
+	0x00000000, 0xd8500000,
+	0x00000000, 0xbefe00ff,
+	0x0000ffff, 0xbeff0080,
+	0xc4068070, 0x008ce802,
+	0x00000000, 0xbefe00c1,
+	0xb8f03b05, 0x80708170,
+	0xbf0d9973, 0xbfa20002,
+	0x84708970, 0xbfa00001,
+	0x84708a70, 0xb8fa1e06,
+	0x847a8a7a, 0x80707a70,
+	0xbef600ff, 0x01000000,
+	0xbef90080, 0xbefd0080,
+	0xbf800000, 0xbe804100,
+	0xbe824102, 0xbe844104,
+	0xbe864106, 0xbe884108,
+	0xbe8a410a, 0xbe8c410c,
+	0xbe8e410e, 0xbf068079,
+	0xbfa10032, 0xd7610002,
+	0x00010000, 0xd7610002,
+	0x00010201, 0xd7610002,
+	0x00010402, 0xd7610002,
+	0x00010603, 0xd7610002,
+	0x00010804, 0xd7610002,
+	0x00010a05, 0xd7610002,
+	0x00010c06, 0xd7610002,
+	0x00010e07, 0xd7610002,
+	0x00011008, 0xd7610002,
+	0x00011209, 0xd7610002,
+	0x0001140a, 0xd7610002,
+	0x0001160b, 0xd7610002,
+	0x0001180c, 0xd7610002,
+	0x00011a0d, 0xd7610002,
+	0x00011c0e, 0xd7610002,
+	0x00011e0f, 0xd8500000,
+	0x00000000, 0xd8500000,
+	0x00000000, 0xd8500000,
+	0x00000000, 0xd8500000,
+	0x00000000, 0xd8500000,
+	0x00000000, 0xd8500000,
+	0x00000000, 0xd8500000,
+	0x00000000, 0xd8500000,
+	0x00000000, 0x80799079,
+	0xbfa00038, 0xd7610002,
+	0x00012000, 0xd7610002,
+	0x00012201, 0xd7610002,
+	0x00012402, 0xd7610002,
+	0x00012603, 0xd7610002,
+	0x00012804, 0xd7610002,
+	0x00012a05, 0xd7610002,
+	0x00012c06, 0xd7610002,
+	0x00012e07, 0xd7610002,
+	0x00013008, 0xd7610002,
+	0x00013209, 0xd7610002,
+	0x0001340a, 0xd7610002,
+	0x0001360b, 0xd7610002,
+	0x0001380c, 0xd7610002,
+	0x00013a0d, 0xd7610002,
+	0x00013c0e, 0xd7610002,
+	0x00013e0f, 0xd8500000,
+	0x00000000, 0xd8500000,
+	0x00000000, 0xd8500000,
+	0x00000000, 0xd8500000,
+	0x00000000, 0xd8500000,
+	0x00000000, 0xd8500000,
+	0x00000000, 0xd8500000,
+	0x00000000, 0xd8500000,
+	0x00000000, 0x80799079,
+	0xc4068070, 0x008ce802,
+	0x00000000, 0x8070ff70,
+	0x00000080, 0xbef90080,
+	0x7e040280, 0x807d907d,
+	0xbf0aff7d, 0x00000060,
+	0xbfa2ff88, 0xbe804100,
+	0xbe824102, 0xbe844104,
+	0xbe864106, 0xbe884108,
+	0xbe8a410a, 0xd7610002,
+	0x00010000, 0xd7610002,
+	0x00010201, 0xd7610002,
+	0x00010402, 0xd7610002,
+	0x00010603, 0xd7610002,
+	0x00010804, 0xd7610002,
+	0x00010a05, 0xd7610002,
+	0x00010c06, 0xd7610002,
+	0x00010e07, 0xd7610002,
+	0x00011008, 0xd7610002,
+	0x00011209, 0xd7610002,
+	0x0001140a, 0xd7610002,
+	0x0001160b, 0xd8500000,
+	0x00000000, 0xd8500000,
+	0x00000000, 0xd8500000,
+	0x00000000, 0xd8500000,
+	0x00000000, 0xd8500000,
+	0x00000000, 0xd8500000,
+	0x00000000, 0xd8500000,
+	0x00000000, 0xd8500000,
+	0x00000000, 0xc4068070,
+	0x008ce802, 0x00000000,
+	0xbefe00c1, 0x857d9973,
+	0x8b7d817d, 0xbf06817d,
+	0xbfa20002, 0xbeff0080,
+	0xbfa00001, 0xbeff00c1,
+	0xb8fb4306, 0x8b7bc17b,
+	0xbfa10044, 0x8b7aff6d,
+	0x80000000, 0xbfa10041,
+	0x847b897b, 0xbef6007b,
+	0xb8f03b05, 0x80708170,
+	0xbf0d9973, 0xbfa20002,
+	0x84708970, 0xbfa00001,
+	0x84708a70, 0xb8fa1e06,
+	0x847a8a7a, 0x80707a70,
+	0x8070ff70, 0x00000200,
+	0x8070ff70, 0x00000080,
+	0xbef600ff, 0x01000000,
+	0xd71f0000, 0x000100c1,
+	0xd7200000, 0x000200c1,
+	0x16000084, 0x857d9973,
+	0x8b7d817d, 0xbf06817d,
+	0xbefd0080, 0xbfa20013,
+	0xbe8300ff, 0x00000080,
+	0xbf800000, 0xbf800000,
+	0xbf800000, 0xd8d80000,
+	0x01000000, 0xbf8a0000,
+	0xc4068070, 0x008ce801,
+	0x00000000, 0x807d037d,
+	0x80700370, 0xd5250000,
+	0x0001ff00, 0x00000080,
+	0xbf0a7b7d, 0xbfa2fff3,
+	0xbfa00012, 0xbe8300ff,
+	0x00000100, 0xbf800000,
+	0xbf800000, 0xbf800000,
+	0xd8d80000, 0x01000000,
+	0xbf8a0000, 0xc4068070,
+	0x008ce801, 0x00000000,
+	0x807d037d, 0x80700370,
+	0xd5250000, 0x0001ff00,
+	0x00000100, 0xbf0a7b7d,
+	0xbfa2fff3, 0xbefe00c1,
+	0x857d9973, 0x8b7d817d,
+	0xbf06817d, 0xbfa20004,
+	0xbef000ff, 0x00000200,
+	0xbeff0080, 0xbfa00003,
+	0xbef000ff, 0x00000400,
+	0xbeff00c1, 0xb8fb3b05,
+	0x807b817b, 0x847b827b,
+	0x857d9973, 0x8b7d817d,
+	0xbf06817d, 0xbfa2001b,
+	0xbef600ff, 0x01000000,
+	0xbefd0084, 0xbf0a7b7d,
+	0xbfa10040, 0x7e008700,
+	0x7e028701, 0x7e048702,
+	0x7e068703, 0xc4068070,
+	0x008ce800, 0x00000000,
+	0xc4068070, 0x008ce801,
+	0x00008000, 0xc4068070,
+	0x008ce802, 0x00010000,
+	0xc4068070, 0x008ce803,
+	0x00018000, 0x807d847d,
+	0x8070ff70, 0x00000200,
+	0xbf0a7b7d, 0xbfa2ffeb,
+	0xbfa0002a, 0xbef600ff,
+	0x01000000, 0xbefd0084,
+	0xbf0a7b7d, 0xbfa10015,
+	0x7e008700, 0x7e028701,
+	0x7e048702, 0x7e068703,
+	0xc4068070, 0x008ce800,
+	0x00000000, 0xc4068070,
+	0x008ce801, 0x00010000,
+	0xc4068070, 0x008ce802,
+	0x00020000, 0xc4068070,
+	0x008ce803, 0x00030000,
+	0x807d847d, 0x8070ff70,
+	0x00000400, 0xbf0a7b7d,
+	0xbfa2ffeb, 0xb8fb1e06,
+	0x8b7bc17b, 0xbfa1000d,
+	0x847b837b, 0x807b7d7b,
+	0xbefe00c1, 0xbeff0080,
+	0x7e008700, 0xc4068070,
+	0x008ce800, 0x00000000,
+	0x807d817d, 0x8070ff70,
+	0x00000080, 0xbf0a7b7d,
+	0xbfa2fff7, 0xbfa0016e,
+	0xbef4007e, 0x8b75ff7f,
+	0x0000ffff, 0x8c75ff75,
+	0x00040000, 0xbef60080,
+	0xbef700ff, 0x10807fac,
+	0xbef1007f, 0xb8f20742,
+	0x84729972, 0x8b6eff7f,
+	0x04000000, 0xbfa1003b,
+	0xbefe00c1, 0x857d9972,
+	0x8b7d817d, 0xbf06817d,
+	0xbfa20002, 0xbeff0080,
+	0xbfa00001, 0xbeff00c1,
+	0xb8ef4306, 0x8b6fc16f,
+	0xbfa10030, 0x846f896f,
+	0xbef6006f, 0xb8f83b05,
+	0x80788178, 0xbf0d9972,
+	0xbfa20002, 0x84788978,
+	0xbfa00001, 0x84788a78,
+	0xb8ee1e06, 0x846e8a6e,
+	0x80786e78, 0x8078ff78,
+	0x00000200, 0x8078ff78,
+	0x00000080, 0xbef600ff,
+	0x01000000, 0x857d9972,
+	0x8b7d817d, 0xbf06817d,
+	0xbefd0080, 0xbfa2000d,
+	0xc4050078, 0x0080e800,
+	0x00000000, 0xbf8a0000,
+	0xdac00000, 0x00000000,
+	0x807dff7d, 0x00000080,
+	0x8078ff78, 0x00000080,
+	0xbf0a6f7d, 0xbfa2fff4,
+	0xbfa0000c, 0xc4050078,
+	0x0080e800, 0x00000000,
+	0xbf8a0000, 0xdac00000,
+	0x00000000, 0x807dff7d,
+	0x00000100, 0x8078ff78,
+	0x00000100, 0xbf0a6f7d,
+	0xbfa2fff4, 0xbef80080,
+	0xbefe00c1, 0x857d9972,
+	0x8b7d817d, 0xbf06817d,
+	0xbfa20002, 0xbeff0080,
+	0xbfa00001, 0xbeff00c1,
+	0xb8ef3b05, 0x806f816f,
+	0x846f826f, 0x857d9972,
+	0x8b7d817d, 0xbf06817d,
+	0xbfa2002c, 0xbef600ff,
+	0x01000000, 0xbeee0078,
+	0x8078ff78, 0x00000200,
+	0xbefd0084, 0xbf0a6f7d,
+	0xbfa10061, 0xc4050078,
+	0x008ce800, 0x00000000,
+	0xc4050078, 0x008ce801,
+	0x00008000, 0xc4050078,
+	0x008ce802, 0x00010000,
+	0xc4050078, 0x008ce803,
+	0x00018000, 0xbf8a0000,
+	0x7e008500, 0x7e028501,
+	0x7e048502, 0x7e068503,
+	0x807d847d, 0x8078ff78,
+	0x00000200, 0xbf0a6f7d,
+	0xbfa2ffea, 0xc405006e,
+	0x008ce800, 0x00000000,
+	0xc405006e, 0x008ce801,
+	0x00008000, 0xc405006e,
+	0x008ce802, 0x00010000,
+	0xc405006e, 0x008ce803,
+	0x00018000, 0xbf8a0000,
+	0xbfa0003d, 0xbef600ff,
+	0x01000000, 0xbeee0078,
+	0x8078ff78, 0x00000400,
+	0xbefd0084, 0xbf0a6f7d,
+	0xbfa10016, 0xc4050078,
+	0x008ce800, 0x00000000,
+	0xc4050078, 0x008ce801,
+	0x00010000, 0xc4050078,
+	0x008ce802, 0x00020000,
+	0xc4050078, 0x008ce803,
+	0x00030000, 0xbf8a0000,
+	0x7e008500, 0x7e028501,
+	0x7e048502, 0x7e068503,
+	0x807d847d, 0x8078ff78,
+	0x00000400, 0xbf0a6f7d,
+	0xbfa2ffea, 0xb8ef1e06,
+	0x8b6fc16f, 0xbfa1000f,
+	0x846f836f, 0x806f7d6f,
+	0xbefe00c1, 0xbeff0080,
+	0xc4050078, 0x008ce800,
+	0x00000000, 0xbf8a0000,
+	0x7e008500, 0x807d817d,
+	0x8078ff78, 0x00000080,
+	0xbf0a6f7d, 0xbfa2fff6,
+	0xbeff00c1, 0xc405006e,
+	0x008ce800, 0x00000000,
+	0xc405006e, 0x008ce801,
+	0x00010000, 0xc405006e,
+	0x008ce802, 0x00020000,
+	0xc405006e, 0x008ce803,
+	0x00030000, 0xbf8a0000,
+	0xb8f83b05, 0x80788178,
+	0xbf0d9972, 0xbfa20002,
+	0x84788978, 0xbfa00001,
+	0x84788a78, 0xb8ee1e06,
+	0x846e8a6e, 0x80786e78,
+	0x8078ff78, 0x00000200,
+	0x80f8ff78, 0x00000050,
+	0xbef600ff, 0x01000000,
+	0xbefd00ff, 0x0000006c,
+	0x80f89078, 0xf462403a,
+	0xf0000000, 0xbf8a0000,
+	0x80fd847d, 0xbf800000,
+	0xbe804300, 0xbe824302,
+	0x80f8a078, 0xf462603a,
+	0xf0000000, 0xbf8a0000,
+	0x80fd887d, 0xbf800000,
+	0xbe804300, 0xbe824302,
+	0xbe844304, 0xbe864306,
+	0x80f8c078, 0xf462803a,
+	0xf0000000, 0xbf8a0000,
+	0x80fd907d, 0xbf800000,
+	0xbe804300, 0xbe824302,
+	0xbe844304, 0xbe864306,
+	0xbe884308, 0xbe8a430a,
+	0xbe8c430c, 0xbe8e430e,
+	0xbf06807d, 0xbfa1fff0,
+	0xb980f801, 0x00000000,
+	0xb8f83b05, 0x80788178,
+	0xbf0d9972, 0xbfa20002,
+	0x84788978, 0xbfa00001,
+	0x84788a78, 0xb8ee1e06,
+	0x846e8a6e, 0x80786e78,
+	0x8078ff78, 0x00000200,
+	0xbef600ff, 0x01000000,
+	0xbeff0071, 0xf4621bfa,
+	0xf0000000, 0x80788478,
+	0xf4621b3a, 0xf0000000,
+	0x80788478, 0xf4621b7a,
+	0xf0000000, 0x80788478,
+	0xf4621c3a, 0xf0000000,
+	0x80788478, 0xf4621c7a,
+	0xf0000000, 0x80788478,
+	0xf4621eba, 0xf0000000,
+	0x80788478, 0xf4621efa,
+	0xf0000000, 0x80788478,
+	0xf4621e7a, 0xf0000000,
+	0x80788478, 0xf4621cfa,
+	0xf0000000, 0x80788478,
+	0xf4621bba, 0xf0000000,
+	0x80788478, 0xbf8a0000,
+	0xb96ef814, 0xf4621bba,
+	0xf0000000, 0x80788478,
+	0xbf8a0000, 0xb96ef815,
+	0xf4621bba, 0xf0000000,
+	0x80788478, 0xbf8a0000,
+	0xb96ef812, 0xf4621bba,
+	0xf0000000, 0x80788478,
+	0xbf8a0000, 0xb96ef813,
+	0x8b6eff7f, 0x04000000,
+	0xbfa1000d, 0x80788478,
+	0xf4621bba, 0xf0000000,
+	0x80788478, 0xbf8a0000,
+	0xbf0d806e, 0xbfa10006,
+	0x856e906e, 0x8b6e6e6e,
+	0xbfa10003, 0xbe804ec1,
+	0x816ec16e, 0xbfa0fffb,
+	0xbefd006f, 0xbefe0070,
+	0xbeff0071, 0xb97b2011,
+	0x857b867b, 0xb97b0191,
+	0x857b827b, 0xb97bba11,
+	0xb973f801, 0xb8ee3b05,
+	0x806e816e, 0xbf0d9972,
+	0xbfa20002, 0x846e896e,
+	0xbfa00001, 0x846e8a6e,
+	0xb8ef1e06, 0x846f8a6f,
+	0x806e6f6e, 0x806eff6e,
+	0x00000200, 0x806e746e,
+	0x826f8075, 0x8b6fff6f,
+	0x0000ffff, 0xf4605c37,
+	0xf8000050, 0xf4605d37,
+	0xf8000060, 0xf4601e77,
+	0xf8000074, 0xbf8a0000,
+	0x8b6dff6d, 0x0000ffff,
+	0x8bfe7e7e, 0x8bea6a6a,
+	0xb97af804, 0xbe804ec2,
+	0xbf94fffe, 0xbe804a6c,
+	0xbe804ec2, 0xbf94fffe,
+	0xbfb10000, 0xbf9f0000,
 	0xbf9f0000, 0xbf9f0000,
 	0xbf9f0000, 0xbf9f0000,
-	0xbf9f0000, 0x00000000,
+};
+
+static const uint32_t cwsr_trap_gfx9_5_0_hex[] = {
+	0xbf820001, 0xbf8202ca,
+	0xb8f8f802, 0x8978ff78,
+	0x00020006, 0xb8fbf803,
+	0x866eff78, 0x00002000,
+	0xbf840009, 0x866eff6d,
+	0x00ff0000, 0xbf85001a,
+	0x866eff7b, 0x00000400,
+	0xbf850051, 0xbf8e0010,
+	0xb8fbf803, 0xbf82fffa,
+	0x866eff7b, 0x03c00900,
+	0xbf850011, 0x866eff7b,
+	0x000071ff, 0xbf840008,
+	0x866fff7b, 0x00007080,
+	0xbf840001, 0xbeee1a87,
+	0xb8eff801, 0x8e6e8c6e,
+	0x866e6f6e, 0xbf850006,
+	0x866eff6d, 0x00ff0000,
+	0xbf850003, 0x866eff7b,
+	0x00000400, 0xbf85003a,
+	0xb8faf807, 0x867aff7a,
+	0x001f8000, 0x8e7a8b7a,
+	0x8979ff79, 0xfc000000,
+	0x87797a79, 0xba7ff807,
+	0x00000000, 0xb8faf812,
+	0xb8fbf813, 0x8efa887a,
+	0xbf0d8f7b, 0xbf840002,
+	0x877bff7b, 0xffff0000,
+	0xc0031bbd, 0x00000010,
+	0xbf8cc07f, 0x8e6e976e,
+	0x8979ff79, 0x00800000,
+	0x87796e79, 0xc0071bbd,
+	0x00000000, 0xbf8cc07f,
+	0xc0071ebd, 0x00000008,
+	0xbf8cc07f, 0x86ee6e6e,
+	0xbf840001, 0xbe801d6e,
+	0x866eff6d, 0x01ff0000,
+	0xbf850005, 0x8778ff78,
+	0x00002000, 0x80ec886c,
+	0x82ed806d, 0xbf820005,
+	0x866eff6d, 0x01000000,
+	0xbf850002, 0x806c846c,
+	0x826d806d, 0x866dff6d,
+	0x0000ffff, 0x8f7a8b79,
+	0x867aff7a, 0x001f8000,
+	0xb97af807, 0x86fe7e7e,
+	0x86ea6a6a, 0x8f6e8378,
+	0xb96ee0c2, 0xbf800002,
+	0xb9780002, 0xbe801f6c,
+	0x866dff6d, 0x0000ffff,
+	0xbefa0080, 0xb97a0283,
+	0xb8faf807, 0x867aff7a,
+	0x001f8000, 0x8e7a8b7a,
+	0x8979ff79, 0xfc000000,
+	0x87797a79, 0xba7ff807,
+	0x00000000, 0xbeee007e,
+	0xbeef007f, 0xbefe0180,
+	0xbf900004, 0x877a8478,
+	0xb97af802, 0xbf8e0002,
+	0xbf88fffe, 0xb8fa2985,
+	0x807a817a, 0x8e7a8a7a,
+	0x8e7a817a, 0xb8fb1605,
+	0x807b817b, 0x8e7b867b,
+	0x807a7b7a, 0x807a7e7a,
+	0x827b807f, 0x867bff7b,
+	0x0000ffff, 0xc04b1c3d,
+	0x00000050, 0xbf8cc07f,
+	0xc04b1d3d, 0x00000060,
+	0xbf8cc07f, 0xc0431e7d,
+	0x00000074, 0xbf8cc07f,
+	0xbef4007e, 0x8675ff7f,
+	0x0000ffff, 0x8775ff75,
+	0x00040000, 0xbef60080,
+	0xbef700ff, 0x00807fac,
+	0xbef1007c, 0xbef00080,
+	0xb8f02985, 0x80708170,
+	0x8e708a70, 0x8e708170,
+	0xb8fa1605, 0x807a817a,
+	0x8e7a867a, 0x80707a70,
+	0xbef60084, 0xbef600ff,
+	0x01000000, 0xbefe007c,
+	0xbefc0070, 0xc0611c7a,
+	0x0000007c, 0xbf8cc07f,
+	0x80708470, 0xbefc007e,
+	0xbefe007c, 0xbefc0070,
+	0xc0611b3a, 0x0000007c,
+	0xbf8cc07f, 0x80708470,
+	0xbefc007e, 0xbefe007c,
+	0xbefc0070, 0xc0611b7a,
+	0x0000007c, 0xbf8cc07f,
+	0x80708470, 0xbefc007e,
+	0xbefe007c, 0xbefc0070,
+	0xc0611bba, 0x0000007c,
+	0xbf8cc07f, 0x80708470,
+	0xbefc007e, 0xbefe007c,
+	0xbefc0070, 0xc0611bfa,
+	0x0000007c, 0xbf8cc07f,
+	0x80708470, 0xbefc007e,
+	0xbefe007c, 0xbefc0070,
+	0xc0611e3a, 0x0000007c,
+	0xbf8cc07f, 0x80708470,
+	0xbefc007e, 0xb8fbf803,
+	0xbefe007c, 0xbefc0070,
+	0xc0611efa, 0x0000007c,
+	0xbf8cc07f, 0x80708470,
+	0xbefc007e, 0xbefe007c,
+	0xbefc0070, 0xc0611a3a,
+	0x0000007c, 0xbf8cc07f,
+	0x80708470, 0xbefc007e,
+	0xbefe007c, 0xbefc0070,
+	0xc0611a7a, 0x0000007c,
+	0xbf8cc07f, 0x80708470,
+	0xbefc007e, 0xb8f1f801,
+	0xbefe007c, 0xbefc0070,
+	0xc0611c7a, 0x0000007c,
+	0xbf8cc07f, 0x80708470,
+	0xbefc007e, 0xbf108080,
+	0x867aff7f, 0x04000000,
+	0xbeef0080, 0x876f6f7a,
+	0xb8f02985, 0x80708170,
+	0x8e708a70, 0x8e708170,
+	0xb8fb1605, 0x807b817b,
+	0x8e7b847b, 0x8e76827b,
+	0xbef600ff, 0x01000000,
+	0xbef20174, 0x80747074,
+	0x82758075, 0xbefc0080,
+	0xbf800000, 0xbe802b00,
+	0xbe822b02, 0xbe842b04,
+	0xbe862b06, 0xbe882b08,
+	0xbe8a2b0a, 0xbe8c2b0c,
+	0xbe8e2b0e, 0xc06b003a,
+	0x00000000, 0xbf8cc07f,
+	0xc06b013a, 0x00000010,
+	0xbf8cc07f, 0xc06b023a,
+	0x00000020, 0xbf8cc07f,
+	0xc06b033a, 0x00000030,
+	0xbf8cc07f, 0x8074c074,
+	0x82758075, 0x807c907c,
+	0xbf0a7b7c, 0xbf85ffe7,
+	0xbef40172, 0xbef00080,
+	0xbefe00c1, 0xbeff00c1,
+	0xbee80080, 0xbee90080,
+	0xbef600ff, 0x01000000,
+	0x867aff78, 0x00400000,
+	0xbf850003, 0xb8faf803,
+	0x897a7aff, 0x10000000,
+	0xbf85004d, 0xbe840080,
+	0xd2890000, 0x00000900,
+	0x80048104, 0xd2890001,
+	0x00000900, 0x80048104,
+	0xd2890002, 0x00000900,
+	0x80048104, 0xd2890003,
+	0x00000900, 0x80048104,
+	0xc069003a, 0x00000070,
+	0xbf8cc07f, 0x80709070,
+	0xbf06c004, 0xbf84ffee,
+	0xbe840080, 0xd2890000,
+	0x00000901, 0x80048104,
+	0xd2890001, 0x00000901,
+	0x80048104, 0xd2890002,
+	0x00000901, 0x80048104,
+	0xd2890003, 0x00000901,
+	0x80048104, 0xc069003a,
+	0x00000070, 0xbf8cc07f,
+	0x80709070, 0xbf06c004,
+	0xbf84ffee, 0xbe840080,
+	0xd2890000, 0x00000902,
+	0x80048104, 0xd2890001,
+	0x00000902, 0x80048104,
+	0xd2890002, 0x00000902,
+	0x80048104, 0xd2890003,
+	0x00000902, 0x80048104,
+	0xc069003a, 0x00000070,
+	0xbf8cc07f, 0x80709070,
+	0xbf06c004, 0xbf84ffee,
+	0xbe840080, 0xd2890000,
+	0x00000903, 0x80048104,
+	0xd2890001, 0x00000903,
+	0x80048104, 0xd2890002,
+	0x00000903, 0x80048104,
+	0xd2890003, 0x00000903,
+	0x80048104, 0xc069003a,
+	0x00000070, 0xbf8cc07f,
+	0x80709070, 0xbf06c004,
+	0xbf84ffee, 0xbf820008,
+	0xe0724000, 0x701d0000,
+	0xe0724100, 0x701d0100,
+	0xe0724200, 0x701d0200,
+	0xe0724300, 0x701d0300,
+	0xbefe00c1, 0xbeff00c1,
+	0xb8fb5306, 0x867bc17b,
+	0xbf840052, 0xbf8a0000,
+	0x867aff6f, 0x04000000,
+	0xbf84004e, 0x8e7b867b,
+	0x8e7b827b, 0xbef6007b,
+	0xb8f02985, 0x80708170,
+	0x8e708a70, 0x8e708170,
+	0xb8fa1605, 0x807a817a,
+	0x8e7a867a, 0x80707a70,
+	0x8070ff70, 0x00000080,
+	0xbef600ff, 0x01000000,
+	0xbefc0080, 0xd28c0002,
+	0x000100c1, 0xd28d0003,
+	0x000204c1, 0x867aff78,
+	0x00400000, 0xbf850003,
+	0xb8faf803, 0x897a7aff,
+	0x10000000, 0xbf85001d,
+	0x24040682, 0xd86c0000,
+	0x00000002, 0xbf8cc07f,
+	0xbe840080, 0xd2890000,
+	0x00000900, 0x80048104,
+	0xd2890001, 0x00000900,
+	0x80048104, 0xd2890002,
+	0x00000900, 0x80048104,
+	0xd2890003, 0x00000900,
+	0x80048104, 0xc069003a,
+	0x00000070, 0xbf8cc07f,
+	0x80709070, 0xbf06c004,
+	0xbf84ffee, 0x680404ff,
+	0x00000100, 0xd0c9006a,
+	0x0000f702, 0xbf87ffe5,
+	0xbf820016, 0xd1060002,
+	0x00011103, 0x7e0602ff,
+	0x00000200, 0xbefc00ff,
+	0x00010000, 0xbe800077,
+	0x8677ff77, 0xff7fffff,
+	0x8777ff77, 0x00058000,
+	0xd8ec0000, 0x00000002,
+	0xbf8cc07f, 0xe0765000,
+	0x701d0002, 0x68040702,
+	0xd0c9006a, 0x0000f702,
+	0xbefe016a, 0xbf87fff6,
+	0xbef70000, 0xbef000ff,
+	0x00000400, 0xbefe00c1,
+	0xbeff00c1, 0xb8fb2b05,
+	0x807b817b, 0x8e7b827b,
+	0xbef600ff, 0x01000000,
+	0xbefc0084, 0xbf0a7b7c,
+	0xbf84006d, 0xbf11017c,
+	0x807bff7b, 0x00001000,
+	0x867aff78, 0x00400000,
+	0xbf850003, 0xb8faf803,
+	0x897a7aff, 0x10000000,
+	0xbf850051, 0xbe840080,
+	0xd2890000, 0x00000900,
+	0x80048104, 0xd2890001,
+	0x00000900, 0x80048104,
+	0xd2890002, 0x00000900,
+	0x80048104, 0xd2890003,
+	0x00000900, 0x80048104,
+	0xc069003a, 0x00000070,
+	0xbf8cc07f, 0x80709070,
+	0xbf06c004, 0xbf84ffee,
+	0xbe840080, 0xd2890000,
+	0x00000901, 0x80048104,
+	0xd2890001, 0x00000901,
+	0x80048104, 0xd2890002,
+	0x00000901, 0x80048104,
+	0xd2890003, 0x00000901,
+	0x80048104, 0xc069003a,
+	0x00000070, 0xbf8cc07f,
+	0x80709070, 0xbf06c004,
+	0xbf84ffee, 0xbe840080,
+	0xd2890000, 0x00000902,
+	0x80048104, 0xd2890001,
+	0x00000902, 0x80048104,
+	0xd2890002, 0x00000902,
+	0x80048104, 0xd2890003,
+	0x00000902, 0x80048104,
+	0xc069003a, 0x00000070,
+	0xbf8cc07f, 0x80709070,
+	0xbf06c004, 0xbf84ffee,
+	0xbe840080, 0xd2890000,
+	0x00000903, 0x80048104,
+	0xd2890001, 0x00000903,
+	0x80048104, 0xd2890002,
+	0x00000903, 0x80048104,
+	0xd2890003, 0x00000903,
+	0x80048104, 0xc069003a,
+	0x00000070, 0xbf8cc07f,
+	0x80709070, 0xbf06c004,
+	0xbf84ffee, 0x807c847c,
+	0xbf0a7b7c, 0xbf85ffb1,
+	0xbf9c0000, 0xbf820012,
+	0x7e000300, 0x7e020301,
+	0x7e040302, 0x7e060303,
+	0xe0724000, 0x701d0000,
+	0xe0724100, 0x701d0100,
+	0xe0724200, 0x701d0200,
+	0xe0724300, 0x701d0300,
+	0x807c847c, 0x8070ff70,
+	0x00000400, 0xbf0a7b7c,
+	0xbf85ffef, 0xbf9c0000,
+	0xb8fb2985, 0x807b817b,
+	0x8e7b837b, 0xb8fa2b05,
+	0x807a817a, 0x8e7a827a,
+	0x80fb7a7b, 0x867b7b7b,
+	0xbf84007a, 0x807bff7b,
+	0x00001000, 0xbefc0080,
+	0xbf11017c, 0x867aff78,
+	0x00400000, 0xbf850003,
+	0xb8faf803, 0x897a7aff,
+	0x10000000, 0xbf850059,
+	0xd3d84000, 0x18000100,
+	0xd3d84001, 0x18000101,
+	0xd3d84002, 0x18000102,
+	0xd3d84003, 0x18000103,
+	0xbe840080, 0xd2890000,
+	0x00000900, 0x80048104,
+	0xd2890001, 0x00000900,
+	0x80048104, 0xd2890002,
+	0x00000900, 0x80048104,
+	0xd2890003, 0x00000900,
+	0x80048104, 0xc069003a,
+	0x00000070, 0xbf8cc07f,
+	0x80709070, 0xbf06c004,
+	0xbf84ffee, 0xbe840080,
+	0xd2890000, 0x00000901,
+	0x80048104, 0xd2890001,
+	0x00000901, 0x80048104,
+	0xd2890002, 0x00000901,
+	0x80048104, 0xd2890003,
+	0x00000901, 0x80048104,
+	0xc069003a, 0x00000070,
+	0xbf8cc07f, 0x80709070,
+	0xbf06c004, 0xbf84ffee,
+	0xbe840080, 0xd2890000,
+	0x00000902, 0x80048104,
+	0xd2890001, 0x00000902,
+	0x80048104, 0xd2890002,
+	0x00000902, 0x80048104,
+	0xd2890003, 0x00000902,
+	0x80048104, 0xc069003a,
+	0x00000070, 0xbf8cc07f,
+	0x80709070, 0xbf06c004,
+	0xbf84ffee, 0xbe840080,
+	0xd2890000, 0x00000903,
+	0x80048104, 0xd2890001,
+	0x00000903, 0x80048104,
+	0xd2890002, 0x00000903,
+	0x80048104, 0xd2890003,
+	0x00000903, 0x80048104,
+	0xc069003a, 0x00000070,
+	0xbf8cc07f, 0x80709070,
+	0xbf06c004, 0xbf84ffee,
+	0x807c847c, 0xbf0a7b7c,
+	0xbf85ffa9, 0xbf9c0000,
+	0xbf820016, 0xd3d84000,
+	0x18000100, 0xd3d84001,
+	0x18000101, 0xd3d84002,
+	0x18000102, 0xd3d84003,
+	0x18000103, 0xe0724000,
+	0x701d0000, 0xe0724100,
+	0x701d0100, 0xe0724200,
+	0x701d0200, 0xe0724300,
+	0x701d0300, 0x807c847c,
+	0x8070ff70, 0x00000400,
+	0xbf0a7b7c, 0xbf85ffeb,
+	0xbf9c0000, 0xbf8200f4,
+	0xbef4007e, 0x8675ff7f,
+	0x0000ffff, 0x8775ff75,
+	0x00040000, 0xbef60080,
+	0xbef700ff, 0x00807fac,
+	0x866eff7f, 0x04000000,
+	0xbf840025, 0xbefe00c1,
+	0xbeff00c1, 0xb8ef5306,
+	0x866fc16f, 0xbf840020,
+	0x8e6f866f, 0x8e6f826f,
+	0xbef6006f, 0xb8f82985,
+	0x80788178, 0x8e788a78,
+	0x8e788178, 0xb8ee1605,
+	0x806e816e, 0x8e6e866e,
+	0x80786e78, 0x8078ff78,
+	0x00000080, 0xbef600ff,
+	0x01000000, 0xbefc0080,
+	0xe0510000, 0x781d0000,
+	0xe0510100, 0x781d0000,
+	0xe0510200, 0x781d0000,
+	0xe0510300, 0x781d0000,
+	0xe0510400, 0x781d0000,
+	0x807cff7c, 0x00000500,
+	0x8078ff78, 0x00000500,
+	0xbf0a6f7c, 0xbf85fff0,
+	0xbefe00c1, 0xbeff00c1,
+	0xbef600ff, 0x01000000,
+	0xb8ef2b05, 0x806f816f,
+	0x8e6f826f, 0x806fff6f,
+	0x00008000, 0xbef80080,
+	0xbeee0078, 0x8078ff78,
+	0x00000400, 0xbefc0084,
+	0xbf11087c, 0xe0524000,
+	0x781d0000, 0xe0524100,
+	0x781d0100, 0xe0524200,
+	0x781d0200, 0xe0524300,
+	0x781d0300, 0xbf8c0f70,
+	0x7e000300, 0x7e020301,
+	0x7e040302, 0x7e060303,
+	0x807c847c, 0x8078ff78,
+	0x00000400, 0xbf0a6f7c,
+	0xbf85ffee, 0xb8ef2985,
+	0x806f816f, 0x8e6f836f,
+	0xb8f92b05, 0x80798179,
+	0x8e798279, 0x80ef796f,
+	0x866f6f6f, 0xbf84001a,
+	0x806fff6f, 0x00008000,
+	0xbefc0080, 0xbf11087c,
+	0xe0524000, 0x781d0000,
+	0xe0524100, 0x781d0100,
+	0xe0524200, 0x781d0200,
+	0xe0524300, 0x781d0300,
+	0xbf8c0f70, 0xd3d94000,
+	0x18000100, 0xd3d94001,
+	0x18000101, 0xd3d94002,
+	0x18000102, 0xd3d94003,
+	0x18000103, 0x807c847c,
+	0x8078ff78, 0x00000400,
+	0xbf0a6f7c, 0xbf85ffea,
+	0xbf9c0000, 0xe0524000,
+	0x6e1d0000, 0xe0524100,
+	0x6e1d0100, 0xe0524200,
+	0x6e1d0200, 0xe0524300,
+	0x6e1d0300, 0xbf8c0f70,
+	0xb8f82985, 0x80788178,
+	0x8e788a78, 0x8e788178,
+	0xb8ee1605, 0x806e816e,
+	0x8e6e866e, 0x80786e78,
+	0x80f8c078, 0xb8ef1605,
+	0x806f816f, 0x8e6f846f,
+	0x8e76826f, 0xbef600ff,
+	0x01000000, 0xbefc006f,
+	0xc031003a, 0x00000078,
+	0x80f8c078, 0xbf8cc07f,
+	0x80fc907c, 0xbf800000,
+	0xbe802d00, 0xbe822d02,
+	0xbe842d04, 0xbe862d06,
+	0xbe882d08, 0xbe8a2d0a,
+	0xbe8c2d0c, 0xbe8e2d0e,
+	0xbf06807c, 0xbf84fff0,
+	0xb8f82985, 0x80788178,
+	0x8e788a78, 0x8e788178,
+	0xb8ee1605, 0x806e816e,
+	0x8e6e866e, 0x80786e78,
+	0xbef60084, 0xbef600ff,
+	0x01000000, 0xc0211bfa,
+	0x00000078, 0x80788478,
+	0xc0211b3a, 0x00000078,
+	0x80788478, 0xc0211b7a,
+	0x00000078, 0x80788478,
+	0xc0211c3a, 0x00000078,
+	0x80788478, 0xc0211c7a,
+	0x00000078, 0x80788478,
+	0xc0211eba, 0x00000078,
+	0x80788478, 0xc0211efa,
+	0x00000078, 0x80788478,
+	0xc0211a3a, 0x00000078,
+	0x80788478, 0xc0211a7a,
+	0x00000078, 0x80788478,
+	0xc0211cfa, 0x00000078,
+	0x80788478, 0xbf8cc07f,
+	0xbefc006f, 0xbefe0070,
+	0xbeff0071, 0x866f7bff,
+	0x000003ff, 0xb96f4803,
+	0x866f7bff, 0xfffff800,
+	0x8f6f8b6f, 0xb96fa2c3,
+	0xb973f801, 0xb8ee2985,
+	0x806e816e, 0x8e6e8a6e,
+	0x8e6e816e, 0xb8ef1605,
+	0x806f816f, 0x8e6f866f,
+	0x806e6f6e, 0x806e746e,
+	0x826f8075, 0x866fff6f,
+	0x0000ffff, 0xc00b1c37,
+	0x00000050, 0xc00b1d37,
+	0x00000060, 0xc0031e77,
+	0x00000074, 0xbf8cc07f,
+	0x8f6e8b79, 0x866eff6e,
+	0x001f8000, 0xb96ef807,
+	0x866dff6d, 0x0000ffff,
+	0x86fe7e7e, 0x86ea6a6a,
+	0x8f6e837a, 0xb96ee0c2,
+	0xbf800002, 0xb97a0002,
+	0xbf8a0000, 0xbe801f6c,
+	0xbf9b0000, 0x00000000,
 };
diff --git a/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler_gfx10.asm b/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler_gfx10.asm
index 5081f91190b8..96fbb16ceb21 100644
--- a/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler_gfx10.asm
+++ b/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler_gfx10.asm
@@ -23,86 +23,129 @@
 /* To compile this assembly code:
  *
  * Navi1x:
- *   cpp -DASIC_TARGET_NAVI1X=1 cwsr_trap_handler_gfx10.asm -P -o nv1x.sp3
- *   sp3-nv1x nv1x.sp3 -hex nv1x.hex
+ *   cpp -DASIC_FAMILY=CHIP_NAVI10 cwsr_trap_handler_gfx10.asm -P -o nv1x.sp3
+ *   sp3 nv1x.sp3 -hex nv1x.hex
+ *
+ * gfx10:
+ *   cpp -DASIC_FAMILY=CHIP_SIENNA_CICHLID cwsr_trap_handler_gfx10.asm -P -o gfx10.sp3
+ *   sp3 gfx10.sp3 -hex gfx10.hex
+ *
+ * gfx11:
+ *   cpp -DASIC_FAMILY=CHIP_PLUM_BONITO cwsr_trap_handler_gfx10.asm -P -o gfx11.sp3
+ *   sp3 gfx11.sp3 -hex gfx11.hex
  *
- * Others:
- *   cpp -DASIC_TARGET_NAVI1X=0 cwsr_trap_handler_gfx10.asm -P -o gfx10.sp3
- *   sp3-gfx10 gfx10.sp3 -hex gfx10.hex
  */
 
-#define NO_SQC_STORE !ASIC_TARGET_NAVI1X
+#define CHIP_NAVI10 26
+#define CHIP_SIENNA_CICHLID 30
+#define CHIP_PLUM_BONITO 36
+
+#define NO_SQC_STORE (ASIC_FAMILY >= CHIP_SIENNA_CICHLID)
+#define HAVE_XNACK (ASIC_FAMILY < CHIP_SIENNA_CICHLID)
+#define HAVE_SENDMSG_RTN (ASIC_FAMILY >= CHIP_PLUM_BONITO)
+#define HAVE_BUFFER_LDS_LOAD (ASIC_FAMILY < CHIP_PLUM_BONITO)
+#define SW_SA_TRAP (ASIC_FAMILY == CHIP_PLUM_BONITO)
+#define SAVE_AFTER_XNACK_ERROR (HAVE_XNACK && !NO_SQC_STORE) // workaround for TCP store failure after XNACK error when ALLOW_REPLAY=0, for debugger
+#define SINGLE_STEP_MISSED_WORKAROUND 1	//workaround for lost MODE.DEBUG_EN exception when SAVECTX raised
 
-var SINGLE_STEP_MISSED_WORKAROUND		= 1	//workaround for lost MODE.DEBUG_EN exception when SAVECTX raised
+#define S_COHERENCE glc:1
+#define V_COHERENCE slc:1 glc:1
+#define S_WAITCNT_0 s_waitcnt 0
 
-var SQ_WAVE_STATUS_INST_ATC_SHIFT		= 23
-var SQ_WAVE_STATUS_INST_ATC_MASK		= 0x00800000
 var SQ_WAVE_STATUS_SPI_PRIO_MASK		= 0x00000006
 var SQ_WAVE_STATUS_HALT_MASK			= 0x2000
-
+var SQ_WAVE_STATUS_ECC_ERR_MASK			= 0x20000
+var SQ_WAVE_STATUS_TRAP_EN_SHIFT		= 6
+var SQ_WAVE_IB_STS2_WAVE64_SHIFT		= 11
+var SQ_WAVE_IB_STS2_WAVE64_SIZE			= 1
+var SQ_WAVE_LDS_ALLOC_GRANULARITY		= 8
+var S_STATUS_HWREG				= HW_REG_STATUS
+var S_STATUS_ALWAYS_CLEAR_MASK			= SQ_WAVE_STATUS_SPI_PRIO_MASK|SQ_WAVE_STATUS_ECC_ERR_MASK
+var S_STATUS_HALT_MASK				= SQ_WAVE_STATUS_HALT_MASK
+var S_SAVE_PC_HI_TRAP_ID_MASK			= 0x00FF0000
+var S_SAVE_PC_HI_HT_MASK			= 0x01000000
+
+var SQ_WAVE_STATUS_NO_VGPRS_SHIFT		= 24
 var SQ_WAVE_LDS_ALLOC_LDS_SIZE_SHIFT		= 12
 var SQ_WAVE_LDS_ALLOC_LDS_SIZE_SIZE		= 9
-var SQ_WAVE_GPR_ALLOC_VGPR_SIZE_SHIFT		= 8
-var SQ_WAVE_GPR_ALLOC_VGPR_SIZE_SIZE		= 6
-var SQ_WAVE_GPR_ALLOC_SGPR_SIZE_SHIFT		= 24
-var SQ_WAVE_GPR_ALLOC_SGPR_SIZE_SIZE		= 4
+var SQ_WAVE_GPR_ALLOC_VGPR_SIZE_SIZE		= 8
 var SQ_WAVE_LDS_ALLOC_VGPR_SHARED_SIZE_SHIFT	= 24
 var SQ_WAVE_LDS_ALLOC_VGPR_SHARED_SIZE_SIZE	= 4
-var SQ_WAVE_IB_STS2_WAVE64_SHIFT		= 11
-var SQ_WAVE_IB_STS2_WAVE64_SIZE			= 1
+
+#if ASIC_FAMILY < CHIP_PLUM_BONITO
+var SQ_WAVE_GPR_ALLOC_VGPR_SIZE_SHIFT		= 8
+#else
+var SQ_WAVE_GPR_ALLOC_VGPR_SIZE_SHIFT		= 12
+#endif
 
 var SQ_WAVE_TRAPSTS_SAVECTX_MASK		= 0x400
-var SQ_WAVE_TRAPSTS_EXCE_MASK			= 0x1FF
+var SQ_WAVE_TRAPSTS_EXCP_MASK			= 0x1FF
 var SQ_WAVE_TRAPSTS_SAVECTX_SHIFT		= 10
+var SQ_WAVE_TRAPSTS_ADDR_WATCH_MASK		= 0x80
+var SQ_WAVE_TRAPSTS_ADDR_WATCH_SHIFT		= 7
 var SQ_WAVE_TRAPSTS_MEM_VIOL_MASK		= 0x100
 var SQ_WAVE_TRAPSTS_MEM_VIOL_SHIFT		= 8
-var SQ_WAVE_TRAPSTS_PRE_SAVECTX_MASK		= 0x3FF
-var SQ_WAVE_TRAPSTS_PRE_SAVECTX_SHIFT		= 0x0
-var SQ_WAVE_TRAPSTS_PRE_SAVECTX_SIZE		= 10
-var SQ_WAVE_TRAPSTS_POST_SAVECTX_MASK		= 0xFFFFF800
-var SQ_WAVE_TRAPSTS_POST_SAVECTX_SHIFT		= 11
-var SQ_WAVE_TRAPSTS_POST_SAVECTX_SIZE		= 21
 var SQ_WAVE_TRAPSTS_ILLEGAL_INST_MASK		= 0x800
+var SQ_WAVE_TRAPSTS_ILLEGAL_INST_SHIFT		= 11
+var SQ_WAVE_TRAPSTS_EXCP_HI_MASK		= 0x7000
+#if ASIC_FAMILY >= CHIP_PLUM_BONITO
+var SQ_WAVE_TRAPSTS_HOST_TRAP_SHIFT		= 16
+var SQ_WAVE_TRAPSTS_WAVE_START_MASK		= 0x20000
+var SQ_WAVE_TRAPSTS_WAVE_START_SHIFT		= 17
+var SQ_WAVE_TRAPSTS_WAVE_END_MASK		= 0x40000
+var SQ_WAVE_TRAPSTS_TRAP_AFTER_INST_MASK	= 0x100000
+#endif
+var SQ_WAVE_TRAPSTS_XNACK_ERROR_MASK		= 0x10000000
+
+var SQ_WAVE_MODE_EXCP_EN_SHIFT			= 12
+var SQ_WAVE_MODE_EXCP_EN_ADDR_WATCH_SHIFT	= 19
 
-var SQ_WAVE_IB_STS_RCNT_SHIFT			= 16
 var SQ_WAVE_IB_STS_FIRST_REPLAY_SHIFT		= 15
 var SQ_WAVE_IB_STS_REPLAY_W64H_SHIFT		= 25
-var SQ_WAVE_IB_STS_REPLAY_W64H_SIZE		= 1
 var SQ_WAVE_IB_STS_REPLAY_W64H_MASK		= 0x02000000
-var SQ_WAVE_IB_STS_FIRST_REPLAY_SIZE		= 1
-var SQ_WAVE_IB_STS_RCNT_SIZE			= 6
 var SQ_WAVE_IB_STS_RCNT_FIRST_REPLAY_MASK	= 0x003F8000
-var SQ_WAVE_IB_STS_RCNT_FIRST_REPLAY_MASK_NEG	= 0x00007FFF
 
 var SQ_WAVE_MODE_DEBUG_EN_MASK			= 0x800
 
-var SQ_BUF_RSRC_WORD1_ATC_SHIFT			= 24
-var SQ_BUF_RSRC_WORD3_MTYPE_SHIFT		= 27
+var S_TRAPSTS_RESTORE_PART_1_SIZE		= SQ_WAVE_TRAPSTS_SAVECTX_SHIFT
+var S_TRAPSTS_RESTORE_PART_2_SHIFT		= SQ_WAVE_TRAPSTS_ILLEGAL_INST_SHIFT
+
+#if ASIC_FAMILY < CHIP_PLUM_BONITO
+var S_TRAPSTS_NON_MASKABLE_EXCP_MASK		= SQ_WAVE_TRAPSTS_MEM_VIOL_MASK|SQ_WAVE_TRAPSTS_ILLEGAL_INST_MASK
+var S_TRAPSTS_RESTORE_PART_2_SIZE		= 32 - S_TRAPSTS_RESTORE_PART_2_SHIFT
+var S_TRAPSTS_RESTORE_PART_3_SHIFT		= 0
+var S_TRAPSTS_RESTORE_PART_3_SIZE		= 0
+#else
+var S_TRAPSTS_NON_MASKABLE_EXCP_MASK		= SQ_WAVE_TRAPSTS_MEM_VIOL_MASK		|\
+						  SQ_WAVE_TRAPSTS_ILLEGAL_INST_MASK	|\
+						  SQ_WAVE_TRAPSTS_WAVE_START_MASK	|\
+						  SQ_WAVE_TRAPSTS_WAVE_END_MASK		|\
+						  SQ_WAVE_TRAPSTS_TRAP_AFTER_INST_MASK
+var S_TRAPSTS_RESTORE_PART_2_SIZE		= SQ_WAVE_TRAPSTS_HOST_TRAP_SHIFT - SQ_WAVE_TRAPSTS_ILLEGAL_INST_SHIFT
+var S_TRAPSTS_RESTORE_PART_3_SHIFT		= SQ_WAVE_TRAPSTS_WAVE_START_SHIFT
+var S_TRAPSTS_RESTORE_PART_3_SIZE		= 32 - S_TRAPSTS_RESTORE_PART_3_SHIFT
+#endif
+var S_TRAPSTS_HWREG				= HW_REG_TRAPSTS
+var S_TRAPSTS_SAVE_CONTEXT_MASK			= SQ_WAVE_TRAPSTS_SAVECTX_MASK
+var S_TRAPSTS_SAVE_CONTEXT_SHIFT		= SQ_WAVE_TRAPSTS_SAVECTX_SHIFT
 
 // bits [31:24] unused by SPI debug data
 var TTMP11_SAVE_REPLAY_W64H_SHIFT		= 31
 var TTMP11_SAVE_REPLAY_W64H_MASK		= 0x80000000
 var TTMP11_SAVE_RCNT_FIRST_REPLAY_SHIFT		= 24
 var TTMP11_SAVE_RCNT_FIRST_REPLAY_MASK		= 0x7F000000
+var TTMP11_DEBUG_TRAP_ENABLED_SHIFT		= 23
+var TTMP11_DEBUG_TRAP_ENABLED_MASK		= 0x800000
 
 // SQ_SEL_X/Y/Z/W, BUF_NUM_FORMAT_FLOAT, (0 for MUBUF stride[17:14]
 // when ADD_TID_ENABLE and BUF_DATA_FORMAT_32 for MTBUF), ADD_TID_ENABLE
 var S_SAVE_BUF_RSRC_WORD1_STRIDE		= 0x00040000
 var S_SAVE_BUF_RSRC_WORD3_MISC			= 0x10807FAC
-
-var S_SAVE_SPI_INIT_ATC_MASK			= 0x08000000
-var S_SAVE_SPI_INIT_ATC_SHIFT			= 27
-var S_SAVE_SPI_INIT_MTYPE_MASK			= 0x70000000
-var S_SAVE_SPI_INIT_MTYPE_SHIFT			= 28
 var S_SAVE_SPI_INIT_FIRST_WAVE_MASK		= 0x04000000
 var S_SAVE_SPI_INIT_FIRST_WAVE_SHIFT		= 26
 
-var S_SAVE_PC_HI_RCNT_SHIFT			= 26
-var S_SAVE_PC_HI_RCNT_MASK			= 0xFC000000
-var S_SAVE_PC_HI_FIRST_REPLAY_SHIFT		= 25
-var S_SAVE_PC_HI_FIRST_REPLAY_MASK		= 0x02000000
-var S_SAVE_PC_HI_REPLAY_W64H_SHIFT		= 24
-var S_SAVE_PC_HI_REPLAY_W64H_MASK		= 0x01000000
+var S_SAVE_PC_HI_FIRST_WAVE_MASK		= 0x80000000
+var S_SAVE_PC_HI_FIRST_WAVE_SHIFT		= 31
 
 var s_sgpr_save_num				= 108
 
@@ -130,19 +173,10 @@ var s_save_ttmps_hi				= s_save_trapsts
 var S_RESTORE_BUF_RSRC_WORD1_STRIDE		= S_SAVE_BUF_RSRC_WORD1_STRIDE
 var S_RESTORE_BUF_RSRC_WORD3_MISC		= S_SAVE_BUF_RSRC_WORD3_MISC
 
-var S_RESTORE_SPI_INIT_ATC_MASK			= 0x08000000
-var S_RESTORE_SPI_INIT_ATC_SHIFT		= 27
-var S_RESTORE_SPI_INIT_MTYPE_MASK		= 0x70000000
-var S_RESTORE_SPI_INIT_MTYPE_SHIFT		= 28
 var S_RESTORE_SPI_INIT_FIRST_WAVE_MASK		= 0x04000000
 var S_RESTORE_SPI_INIT_FIRST_WAVE_SHIFT		= 26
 var S_WAVE_SIZE					= 25
 
-var S_RESTORE_PC_HI_RCNT_SHIFT			= S_SAVE_PC_HI_RCNT_SHIFT
-var S_RESTORE_PC_HI_RCNT_MASK			= S_SAVE_PC_HI_RCNT_MASK
-var S_RESTORE_PC_HI_FIRST_REPLAY_SHIFT		= S_SAVE_PC_HI_FIRST_REPLAY_SHIFT
-var S_RESTORE_PC_HI_FIRST_REPLAY_MASK		= S_SAVE_PC_HI_FIRST_REPLAY_MASK
-
 var s_restore_spi_init_lo			= exec_lo
 var s_restore_spi_init_hi			= exec_hi
 var s_restore_mem_offset			= ttmp12
@@ -166,6 +200,7 @@ var s_restore_buf_rsrc3				= ttmp11
 var s_restore_size				= ttmp6
 var s_restore_ttmps_lo				= s_restore_tmp
 var s_restore_ttmps_hi				= s_restore_alloc_size
+var s_restore_spi_init_hi_save			= s_restore_exec_hi
 
 shader main
 	asic(DEFAULT)
@@ -178,113 +213,176 @@ L_JUMP_TO_RESTORE:
 	s_branch	L_RESTORE
 
 L_SKIP_RESTORE:
-	s_getreg_b32	s_save_status, hwreg(HW_REG_STATUS)			//save STATUS since we will change SCC
-	s_andn2_b32	s_save_status, s_save_status, SQ_WAVE_STATUS_SPI_PRIO_MASK
+	s_getreg_b32	s_save_status, hwreg(S_STATUS_HWREG)			//save STATUS since we will change SCC
 
-if SINGLE_STEP_MISSED_WORKAROUND
-	// No single step exceptions if MODE.DEBUG_EN=0.
-	s_getreg_b32    ttmp2, hwreg(HW_REG_MODE)
-	s_and_b32       ttmp2, ttmp2, SQ_WAVE_MODE_DEBUG_EN_MASK
-	s_cbranch_scc0  L_NO_SINGLE_STEP_WORKAROUND
+	// Clear SPI_PRIO: do not save with elevated priority.
+	// Clear ECC_ERR: prevents SQC store and triggers FATAL_HALT if setreg'd.
+	s_andn2_b32	s_save_status, s_save_status, S_STATUS_ALWAYS_CLEAR_MASK
 
-	// Second-level trap already handled exception if STATUS.HALT=1.
-	s_and_b32       ttmp2, s_save_status, SQ_WAVE_STATUS_HALT_MASK
+	s_getreg_b32	s_save_trapsts, hwreg(S_TRAPSTS_HWREG)
 
-	// Prioritize single step exception over context save.
-	// Second-level trap will halt wave and RFE, re-entering for SAVECTX.
-	s_cbranch_scc0  L_FETCH_2ND_TRAP
+#if SW_SA_TRAP
+	// If ttmp1[30] is set then issue s_barrier to unblock dependent waves.
+	s_bitcmp1_b32	s_save_pc_hi, 30
+	s_cbranch_scc0	L_TRAP_NO_BARRIER
+	s_barrier
 
-L_NO_SINGLE_STEP_WORKAROUND:
-end
+L_TRAP_NO_BARRIER:
+	// If ttmp1[31] is set then trap may occur early.
+	// Spin wait until SAVECTX exception is raised.
+	s_bitcmp1_b32	s_save_pc_hi, 31
+	s_cbranch_scc1  L_CHECK_SAVE
+#endif
 
+	s_and_b32       ttmp2, s_save_status, S_STATUS_HALT_MASK
+	s_cbranch_scc0	L_NOT_HALTED
 
-	s_getreg_b32	s_save_trapsts, hwreg(HW_REG_TRAPSTS)
-	s_and_b32	ttmp2, s_save_trapsts, SQ_WAVE_TRAPSTS_SAVECTX_MASK	//check whether this is for save
+L_HALTED:
+	// Host trap may occur while wave is halted.
+	s_and_b32	ttmp2, s_save_pc_hi, S_SAVE_PC_HI_TRAP_ID_MASK
+	s_cbranch_scc1	L_FETCH_2ND_TRAP
+
+L_CHECK_SAVE:
+	s_and_b32	ttmp2, s_save_trapsts, S_TRAPSTS_SAVE_CONTEXT_MASK
 	s_cbranch_scc1	L_SAVE
 
-	// If STATUS.MEM_VIOL is asserted then halt the wave to prevent
-	// the exception raising again and blocking context save.
-	s_and_b32	ttmp2, s_save_trapsts, SQ_WAVE_TRAPSTS_MEM_VIOL_MASK
-	s_cbranch_scc0	L_FETCH_2ND_TRAP
-	s_or_b32	s_save_status, s_save_status, SQ_WAVE_STATUS_HALT_MASK
+	// Wave is halted but neither host trap nor SAVECTX is raised.
+	// Caused by instruction fetch memory violation.
+	// Spin wait until context saved to prevent interrupt storm.
+	s_sleep		0x10
+	s_getreg_b32	s_save_trapsts, hwreg(S_TRAPSTS_HWREG)
+	s_branch	L_CHECK_SAVE
+
+L_NOT_HALTED:
+	// Let second-level handle non-SAVECTX exception or trap.
+	// Any concurrent SAVECTX will be handled upon re-entry once halted.
+
+	// Check non-maskable exceptions. memory_violation, illegal_instruction
+	// and xnack_error exceptions always cause the wave to enter the trap
+	// handler.
+	s_and_b32	ttmp2, s_save_trapsts, S_TRAPSTS_NON_MASKABLE_EXCP_MASK
+	s_cbranch_scc1	L_FETCH_2ND_TRAP
+
+	// Check for maskable exceptions in trapsts.excp and trapsts.excp_hi.
+	// Maskable exceptions only cause the wave to enter the trap handler if
+	// their respective bit in mode.excp_en is set.
+	s_and_b32	ttmp2, s_save_trapsts, SQ_WAVE_TRAPSTS_EXCP_MASK|SQ_WAVE_TRAPSTS_EXCP_HI_MASK
+	s_cbranch_scc0	L_CHECK_TRAP_ID
+
+	s_and_b32	ttmp3, s_save_trapsts, SQ_WAVE_TRAPSTS_ADDR_WATCH_MASK|SQ_WAVE_TRAPSTS_EXCP_HI_MASK
+	s_cbranch_scc0	L_NOT_ADDR_WATCH
+	s_bitset1_b32	ttmp2, SQ_WAVE_TRAPSTS_ADDR_WATCH_SHIFT // Check all addr_watch[123] exceptions against excp_en.addr_watch
+
+L_NOT_ADDR_WATCH:
+	s_getreg_b32	ttmp3, hwreg(HW_REG_MODE)
+	s_lshl_b32	ttmp2, ttmp2, SQ_WAVE_MODE_EXCP_EN_SHIFT
+	s_and_b32	ttmp2, ttmp2, ttmp3
+	s_cbranch_scc1	L_FETCH_2ND_TRAP
+
+L_CHECK_TRAP_ID:
+	// Check trap_id != 0
+	s_and_b32	ttmp2, s_save_pc_hi, S_SAVE_PC_HI_TRAP_ID_MASK
+	s_cbranch_scc1	L_FETCH_2ND_TRAP
+
+#if SINGLE_STEP_MISSED_WORKAROUND
+	// Prioritize single step exception over context save.
+	// Second-level trap will halt wave and RFE, re-entering for SAVECTX.
+	s_getreg_b32	ttmp2, hwreg(HW_REG_MODE)
+	s_and_b32	ttmp2, ttmp2, SQ_WAVE_MODE_DEBUG_EN_MASK
+	s_cbranch_scc1	L_FETCH_2ND_TRAP
+#endif
 
-L_FETCH_2ND_TRAP:
+	s_and_b32	ttmp2, s_save_trapsts, S_TRAPSTS_SAVE_CONTEXT_MASK
+	s_cbranch_scc1	L_SAVE
 
-#if ASIC_TARGET_NAVI1X
-	// Preserve and clear scalar XNACK state before issuing scalar loads.
-	// Save IB_STS.REPLAY_W64H[25], RCNT[21:16], FIRST_REPLAY[15] into
-	// unused space ttmp11[31:24].
-	s_andn2_b32	ttmp11, ttmp11, (TTMP11_SAVE_REPLAY_W64H_MASK | TTMP11_SAVE_RCNT_FIRST_REPLAY_MASK)
-	s_getreg_b32	ttmp2, hwreg(HW_REG_IB_STS)
-	s_and_b32	ttmp3, ttmp2, SQ_WAVE_IB_STS_REPLAY_W64H_MASK
-	s_lshl_b32	ttmp3, ttmp3, (TTMP11_SAVE_REPLAY_W64H_SHIFT - SQ_WAVE_IB_STS_REPLAY_W64H_SHIFT)
-	s_or_b32	ttmp11, ttmp11, ttmp3
-	s_and_b32	ttmp3, ttmp2, SQ_WAVE_IB_STS_RCNT_FIRST_REPLAY_MASK
-	s_lshl_b32	ttmp3, ttmp3, (TTMP11_SAVE_RCNT_FIRST_REPLAY_SHIFT - SQ_WAVE_IB_STS_FIRST_REPLAY_SHIFT)
-	s_or_b32	ttmp11, ttmp11, ttmp3
-	s_andn2_b32	ttmp2, ttmp2, (SQ_WAVE_IB_STS_REPLAY_W64H_MASK | SQ_WAVE_IB_STS_RCNT_FIRST_REPLAY_MASK)
-	s_setreg_b32	hwreg(HW_REG_IB_STS), ttmp2
+L_FETCH_2ND_TRAP:
+#if HAVE_XNACK
+	save_and_clear_ib_sts(ttmp14, ttmp15)
 #endif
 
 	// Read second-level TBA/TMA from first-level TMA and jump if available.
 	// ttmp[2:5] and ttmp12 can be used (others hold SPI-initialized debug data)
 	// ttmp12 holds SQ_WAVE_STATUS
+#if HAVE_SENDMSG_RTN
+	s_sendmsg_rtn_b64       [ttmp14, ttmp15], sendmsg(MSG_RTN_GET_TMA)
+	S_WAITCNT_0
+#else
 	s_getreg_b32	ttmp14, hwreg(HW_REG_SHADER_TMA_LO)
 	s_getreg_b32	ttmp15, hwreg(HW_REG_SHADER_TMA_HI)
+#endif
 	s_lshl_b64	[ttmp14, ttmp15], [ttmp14, ttmp15], 0x8
-	s_load_dwordx2	[ttmp2, ttmp3], [ttmp14, ttmp15], 0x0 glc:1		// second-level TBA
-	s_waitcnt	lgkmcnt(0)
-	s_load_dwordx2	[ttmp14, ttmp15], [ttmp14, ttmp15], 0x8 glc:1		// second-level TMA
-	s_waitcnt	lgkmcnt(0)
+
+	s_bitcmp1_b32	ttmp15, 0xF
+	s_cbranch_scc0	L_NO_SIGN_EXTEND_TMA
+	s_or_b32	ttmp15, ttmp15, 0xFFFF0000
+L_NO_SIGN_EXTEND_TMA:
+
+	s_load_dword    ttmp2, [ttmp14, ttmp15], 0x10 S_COHERENCE		// debug trap enabled flag
+	S_WAITCNT_0
+	s_lshl_b32      ttmp2, ttmp2, TTMP11_DEBUG_TRAP_ENABLED_SHIFT
+	s_andn2_b32     ttmp11, ttmp11, TTMP11_DEBUG_TRAP_ENABLED_MASK
+	s_or_b32        ttmp11, ttmp11, ttmp2
+
+	s_load_dwordx2	[ttmp2, ttmp3], [ttmp14, ttmp15], 0x0 S_COHERENCE	// second-level TBA
+	S_WAITCNT_0
+	s_load_dwordx2	[ttmp14, ttmp15], [ttmp14, ttmp15], 0x8 S_COHERENCE	// second-level TMA
+	S_WAITCNT_0
+
 	s_and_b64	[ttmp2, ttmp3], [ttmp2, ttmp3], [ttmp2, ttmp3]
 	s_cbranch_scc0	L_NO_NEXT_TRAP						// second-level trap handler not been set
 	s_setpc_b64	[ttmp2, ttmp3]						// jump to second-level trap handler
 
 L_NO_NEXT_TRAP:
-	s_getreg_b32	s_save_trapsts, hwreg(HW_REG_TRAPSTS)
-	s_and_b32	s_save_trapsts, s_save_trapsts, SQ_WAVE_TRAPSTS_EXCE_MASK
-	s_cbranch_scc1	L_EXCP_CASE						// Exception, jump back to the shader program directly.
-	s_add_u32	ttmp0, ttmp0, 4						// S_TRAP case, add 4 to ttmp0
-	s_addc_u32	ttmp1, ttmp1, 0
-L_EXCP_CASE:
+	// If not caused by trap then halt wave to prevent re-entry.
+	s_and_b32	ttmp2, s_save_pc_hi, S_SAVE_PC_HI_TRAP_ID_MASK
+	s_cbranch_scc1	L_TRAP_CASE
+
+	// Host trap will not cause trap re-entry.
+	s_and_b32	ttmp2, s_save_pc_hi, S_SAVE_PC_HI_HT_MASK
+	s_cbranch_scc1	L_EXIT_TRAP
+	s_or_b32	s_save_status, s_save_status, S_STATUS_HALT_MASK
+
+	// If the PC points to S_ENDPGM then context save will fail if STATUS.HALT is set.
+	// Rewind the PC to prevent this from occurring.
+	s_sub_u32	ttmp0, ttmp0, 0x8
+	s_subb_u32	ttmp1, ttmp1, 0x0
+
+	s_branch	L_EXIT_TRAP
+
+L_TRAP_CASE:
+	// Advance past trap instruction to prevent re-entry.
+	s_add_u32	ttmp0, ttmp0, 0x4
+	s_addc_u32	ttmp1, ttmp1, 0x0
+
+L_EXIT_TRAP:
 	s_and_b32	ttmp1, ttmp1, 0xFFFF
 
-#if ASIC_TARGET_NAVI1X
-	// Restore SQ_WAVE_IB_STS.
-	s_lshr_b32	ttmp2, ttmp11, (TTMP11_SAVE_RCNT_FIRST_REPLAY_SHIFT - SQ_WAVE_IB_STS_FIRST_REPLAY_SHIFT)
-	s_and_b32	ttmp3, ttmp2, SQ_WAVE_IB_STS_RCNT_FIRST_REPLAY_MASK
-	s_lshr_b32	ttmp2, ttmp11, (TTMP11_SAVE_REPLAY_W64H_SHIFT - SQ_WAVE_IB_STS_REPLAY_W64H_SHIFT)
-	s_and_b32	ttmp2, ttmp2, SQ_WAVE_IB_STS_REPLAY_W64H_MASK
-	s_or_b32	ttmp2, ttmp2, ttmp3
-	s_setreg_b32	hwreg(HW_REG_IB_STS), ttmp2
+#if HAVE_XNACK
+	restore_ib_sts(ttmp14, ttmp15)
 #endif
 
 	// Restore SQ_WAVE_STATUS.
 	s_and_b64	exec, exec, exec					// Restore STATUS.EXECZ, not writable by s_setreg_b32
 	s_and_b64	vcc, vcc, vcc						// Restore STATUS.VCCZ, not writable by s_setreg_b32
-	s_setreg_b32	hwreg(HW_REG_STATUS), s_save_status
 
+	s_setreg_b32	hwreg(S_STATUS_HWREG), s_save_status
 	s_rfe_b64	[ttmp0, ttmp1]
 
 L_SAVE:
+	// If VGPRs have been deallocated then terminate the wavefront.
+	// It has no remaining program to run and cannot save without VGPRs.
+#if ASIC_FAMILY == CHIP_PLUM_BONITO
+	s_bitcmp1_b32	s_save_status, SQ_WAVE_STATUS_NO_VGPRS_SHIFT
+	s_cbranch_scc0	L_HAVE_VGPRS
+	s_endpgm
+L_HAVE_VGPRS:
+#endif
 	s_and_b32	s_save_pc_hi, s_save_pc_hi, 0x0000ffff			//pc[47:32]
 	s_mov_b32	s_save_tmp, 0
-	s_setreg_b32	hwreg(HW_REG_TRAPSTS, SQ_WAVE_TRAPSTS_SAVECTX_SHIFT, 1), s_save_tmp	//clear saveCtx bit
+	s_setreg_b32	hwreg(S_TRAPSTS_HWREG, S_TRAPSTS_SAVE_CONTEXT_SHIFT, 1), s_save_tmp	//clear saveCtx bit
 
-#if ASIC_TARGET_NAVI1X
-	s_getreg_b32	s_save_tmp, hwreg(HW_REG_IB_STS, SQ_WAVE_IB_STS_RCNT_SHIFT, SQ_WAVE_IB_STS_RCNT_SIZE)
-	s_lshl_b32	s_save_tmp, s_save_tmp, S_SAVE_PC_HI_RCNT_SHIFT
-	s_or_b32	s_save_pc_hi, s_save_pc_hi, s_save_tmp
-	s_getreg_b32	s_save_tmp, hwreg(HW_REG_IB_STS, SQ_WAVE_IB_STS_FIRST_REPLAY_SHIFT, SQ_WAVE_IB_STS_FIRST_REPLAY_SIZE)
-	s_lshl_b32	s_save_tmp, s_save_tmp, S_SAVE_PC_HI_FIRST_REPLAY_SHIFT
-	s_or_b32	s_save_pc_hi, s_save_pc_hi, s_save_tmp
-	s_getreg_b32	s_save_tmp, hwreg(HW_REG_IB_STS, SQ_WAVE_IB_STS_REPLAY_W64H_SHIFT, SQ_WAVE_IB_STS_REPLAY_W64H_SIZE)
-	s_lshl_b32	s_save_tmp, s_save_tmp, S_SAVE_PC_HI_REPLAY_W64H_SHIFT
-	s_or_b32	s_save_pc_hi, s_save_pc_hi, s_save_tmp
-	s_getreg_b32	s_save_tmp, hwreg(HW_REG_IB_STS)			//clear RCNT and FIRST_REPLAY and REPLAY_W64H in IB_STS
-	s_and_b32	s_save_tmp, s_save_tmp, SQ_WAVE_IB_STS_RCNT_FIRST_REPLAY_MASK_NEG
-
-	s_setreg_b32	hwreg(HW_REG_IB_STS), s_save_tmp
+#if HAVE_XNACK
+	save_and_clear_ib_sts(s_save_tmp, s_save_trapsts)
 #endif
 
 	/* inform SPI the readiness and wait for SPI's go signal */
@@ -292,9 +390,13 @@ L_SAVE:
 	s_mov_b32	s_save_exec_hi, exec_hi
 	s_mov_b64	exec, 0x0						//clear EXEC to get ready to receive
 
+#if HAVE_SENDMSG_RTN
+	s_sendmsg_rtn_b64       [exec_lo, exec_hi], sendmsg(MSG_RTN_SAVE_WAVE)
+#else
 	s_sendmsg	sendmsg(MSG_SAVEWAVE)					//send SPI a message and wait for SPI's write to EXEC
+#endif
 
-#if ASIC_TARGET_NAVI1X
+#if ASIC_FAMILY < CHIP_SIENNA_CICHLID
 L_SLEEP:
 	// sleep 1 (64clk) is not enough for 8 waves per SIMD, which will cause
 	// SQ hang, since the 7,8th wave could not get arbit to exec inst, while
@@ -302,22 +404,69 @@ L_SLEEP:
 	s_sleep		0x2
 	s_cbranch_execz	L_SLEEP
 #else
-	s_waitcnt	lgkmcnt(0)
+	S_WAITCNT_0
+#endif
+
+	// Save first_wave flag so we can clear high bits of save address.
+	s_and_b32	s_save_tmp, s_save_spi_init_hi, S_SAVE_SPI_INIT_FIRST_WAVE_MASK
+	s_lshl_b32	s_save_tmp, s_save_tmp, (S_SAVE_PC_HI_FIRST_WAVE_SHIFT - S_SAVE_SPI_INIT_FIRST_WAVE_SHIFT)
+	s_or_b32	s_save_pc_hi, s_save_pc_hi, s_save_tmp
+
+#if NO_SQC_STORE
+#if ASIC_FAMILY <= CHIP_SIENNA_CICHLID
+	// gfx10: If there was a VALU exception, the exception state must be
+	// cleared before executing the VALU instructions below.
+	v_clrexcp
+#endif
+
+	// Trap temporaries must be saved via VGPR but all VGPRs are in use.
+	// There is no ttmp space to hold the resource constant for VGPR save.
+	// Save v0 by itself since it requires only two SGPRs.
+	s_mov_b32	s_save_ttmps_lo, exec_lo
+	s_and_b32	s_save_ttmps_hi, exec_hi, 0xFFFF
+	s_mov_b32	exec_lo, 0xFFFFFFFF
+	s_mov_b32	exec_hi, 0xFFFFFFFF
+	global_store_dword_addtid	v0, [s_save_ttmps_lo, s_save_ttmps_hi] V_COHERENCE
+	v_mov_b32	v0, 0x0
+	s_mov_b32	exec_lo, s_save_ttmps_lo
+	s_mov_b32	exec_hi, s_save_ttmps_hi
 #endif
 
 	// Save trap temporaries 4-11, 13 initialized by SPI debug dispatch logic
-	// ttmp SR memory offset : size(VGPR)+size(SGPR)+0x40
-	get_wave_size(s_save_ttmps_hi)
+	// ttmp SR memory offset : size(VGPR)+size(SVGPR)+size(SGPR)+0x40
+	get_wave_size2(s_save_ttmps_hi)
 	get_vgpr_size_bytes(s_save_ttmps_lo, s_save_ttmps_hi)
+	get_svgpr_size_bytes(s_save_ttmps_hi)
+	s_add_u32	s_save_ttmps_lo, s_save_ttmps_lo, s_save_ttmps_hi
 	s_and_b32	s_save_ttmps_hi, s_save_spi_init_hi, 0xFFFF
 	s_add_u32	s_save_ttmps_lo, s_save_ttmps_lo, get_sgpr_size_bytes()
 	s_add_u32	s_save_ttmps_lo, s_save_ttmps_lo, s_save_spi_init_lo
 	s_addc_u32	s_save_ttmps_hi, s_save_ttmps_hi, 0x0
 
-#if ASIC_TARGET_NAVI1X
-	s_store_dwordx4	[ttmp4, ttmp5, ttmp6, ttmp7], [s_save_ttmps_lo, s_save_ttmps_hi], 0x50 glc:1
-	s_store_dwordx4	[ttmp8, ttmp9, ttmp10, ttmp11], [s_save_ttmps_lo, s_save_ttmps_hi], 0x60 glc:1
-	s_store_dword   ttmp13, [s_save_ttmps_lo, s_save_ttmps_hi], 0x74 glc:1
+#if NO_SQC_STORE
+	v_writelane_b32	v0, ttmp4, 0x4
+	v_writelane_b32	v0, ttmp5, 0x5
+	v_writelane_b32	v0, ttmp6, 0x6
+	v_writelane_b32	v0, ttmp7, 0x7
+	v_writelane_b32	v0, ttmp8, 0x8
+	v_writelane_b32	v0, ttmp9, 0x9
+	v_writelane_b32	v0, ttmp10, 0xA
+	v_writelane_b32	v0, ttmp11, 0xB
+	v_writelane_b32	v0, ttmp13, 0xD
+	v_writelane_b32	v0, exec_lo, 0xE
+	v_writelane_b32	v0, exec_hi, 0xF
+
+	s_mov_b32	exec_lo, 0x3FFF
+	s_mov_b32	exec_hi, 0x0
+	global_store_dword_addtid	v0, [s_save_ttmps_lo, s_save_ttmps_hi] inst_offset:0x40 V_COHERENCE
+	v_readlane_b32	ttmp14, v0, 0xE
+	v_readlane_b32	ttmp15, v0, 0xF
+	s_mov_b32	exec_lo, ttmp14
+	s_mov_b32	exec_hi, ttmp15
+#else
+	s_store_dwordx4	[ttmp4, ttmp5, ttmp6, ttmp7], [s_save_ttmps_lo, s_save_ttmps_hi], 0x50 S_COHERENCE
+	s_store_dwordx4	[ttmp8, ttmp9, ttmp10, ttmp11], [s_save_ttmps_lo, s_save_ttmps_hi], 0x60 S_COHERENCE
+	s_store_dword   ttmp13, [s_save_ttmps_lo, s_save_ttmps_hi], 0x74 S_COHERENCE
 #endif
 
 	/* setup Resource Contants */
@@ -326,20 +475,14 @@ L_SLEEP:
 	s_or_b32	s_save_buf_rsrc1, s_save_buf_rsrc1, S_SAVE_BUF_RSRC_WORD1_STRIDE
 	s_mov_b32	s_save_buf_rsrc2, 0					//NUM_RECORDS initial value = 0 (in bytes) although not neccessarily inited
 	s_mov_b32	s_save_buf_rsrc3, S_SAVE_BUF_RSRC_WORD3_MISC
-	s_and_b32	s_save_tmp, s_save_spi_init_hi, S_SAVE_SPI_INIT_ATC_MASK
-	s_lshr_b32	s_save_tmp, s_save_tmp, (S_SAVE_SPI_INIT_ATC_SHIFT-SQ_BUF_RSRC_WORD1_ATC_SHIFT)
-	s_or_b32	s_save_buf_rsrc3, s_save_buf_rsrc3, s_save_tmp		//or ATC
-	s_and_b32	s_save_tmp, s_save_spi_init_hi, S_SAVE_SPI_INIT_MTYPE_MASK
-	s_lshr_b32	s_save_tmp, s_save_tmp, (S_SAVE_SPI_INIT_MTYPE_SHIFT-SQ_BUF_RSRC_WORD3_MTYPE_SHIFT)
-	s_or_b32	s_save_buf_rsrc3, s_save_buf_rsrc3, s_save_tmp		//or MTYPE
 
 	s_mov_b32	s_save_m0, m0
 
 	/* global mem offset */
 	s_mov_b32	s_save_mem_offset, 0x0
-	get_wave_size(s_wave_size)
+	get_wave_size2(s_wave_size)
 
-#if ASIC_TARGET_NAVI1X
+#if HAVE_XNACK
 	// Save and clear vector XNACK state late to free up SGPRs.
 	s_getreg_b32	s_save_xnack_mask, hwreg(HW_REG_SHADER_XNACK_MASK)
 	s_setreg_imm32_b32	hwreg(HW_REG_SHADER_XNACK_MASK), 0x0
@@ -361,10 +504,22 @@ L_SAVE_4VGPR_WAVE32:
 
 	// VGPR Allocated in 4-GPR granularity
 
-	buffer_store_dword	v0, v0, s_save_buf_rsrc0, s_save_mem_offset slc:1 glc:1
-	buffer_store_dword	v1, v0, s_save_buf_rsrc0, s_save_mem_offset slc:1 glc:1 offset:128
-	buffer_store_dword	v2, v0, s_save_buf_rsrc0, s_save_mem_offset slc:1 glc:1 offset:128*2
-	buffer_store_dword	v3, v0, s_save_buf_rsrc0, s_save_mem_offset slc:1 glc:1 offset:128*3
+#if SAVE_AFTER_XNACK_ERROR
+	check_if_tcp_store_ok()
+	s_cbranch_scc1 L_SAVE_FIRST_VGPRS32_WITH_TCP
+
+	write_vgprs_to_mem_with_sqc_w32(v0, 4, s_save_buf_rsrc0, s_save_mem_offset)
+	s_branch L_SAVE_HWREG
+
+L_SAVE_FIRST_VGPRS32_WITH_TCP:
+#endif
+
+#if !NO_SQC_STORE
+	buffer_store_dword	v0, v0, s_save_buf_rsrc0, s_save_mem_offset V_COHERENCE
+#endif
+	buffer_store_dword	v1, v0, s_save_buf_rsrc0, s_save_mem_offset V_COHERENCE offset:128
+	buffer_store_dword	v2, v0, s_save_buf_rsrc0, s_save_mem_offset V_COHERENCE offset:128*2
+	buffer_store_dword	v3, v0, s_save_buf_rsrc0, s_save_mem_offset V_COHERENCE offset:128*3
 	s_branch	L_SAVE_HWREG
 
 L_SAVE_4VGPR_WAVE64:
@@ -372,10 +527,22 @@ L_SAVE_4VGPR_WAVE64:
 
 	// VGPR Allocated in 4-GPR granularity
 
-	buffer_store_dword	v0, v0, s_save_buf_rsrc0, s_save_mem_offset slc:1 glc:1
-	buffer_store_dword	v1, v0, s_save_buf_rsrc0, s_save_mem_offset slc:1 glc:1 offset:256
-	buffer_store_dword	v2, v0, s_save_buf_rsrc0, s_save_mem_offset slc:1 glc:1 offset:256*2
-	buffer_store_dword	v3, v0, s_save_buf_rsrc0, s_save_mem_offset slc:1 glc:1 offset:256*3
+#if  SAVE_AFTER_XNACK_ERROR
+	check_if_tcp_store_ok()
+	s_cbranch_scc1 L_SAVE_FIRST_VGPRS64_WITH_TCP
+
+	write_vgprs_to_mem_with_sqc_w64(v0, 4, s_save_buf_rsrc0, s_save_mem_offset)
+	s_branch L_SAVE_HWREG
+
+L_SAVE_FIRST_VGPRS64_WITH_TCP:
+#endif
+
+#if !NO_SQC_STORE
+	buffer_store_dword	v0, v0, s_save_buf_rsrc0, s_save_mem_offset V_COHERENCE
+#endif
+	buffer_store_dword	v1, v0, s_save_buf_rsrc0, s_save_mem_offset V_COHERENCE offset:256
+	buffer_store_dword	v2, v0, s_save_buf_rsrc0, s_save_mem_offset V_COHERENCE offset:256*2
+	buffer_store_dword	v3, v0, s_save_buf_rsrc0, s_save_mem_offset V_COHERENCE offset:256*3
 
 	/* save HW registers */
 
@@ -397,12 +564,13 @@ L_SAVE_HWREG:
 
 	write_hwreg_to_mem(s_save_m0, s_save_buf_rsrc0, s_save_mem_offset)
 	write_hwreg_to_mem(s_save_pc_lo, s_save_buf_rsrc0, s_save_mem_offset)
-	write_hwreg_to_mem(s_save_pc_hi, s_save_buf_rsrc0, s_save_mem_offset)
+	s_andn2_b32	s_save_tmp, s_save_pc_hi, S_SAVE_PC_HI_FIRST_WAVE_MASK
+	write_hwreg_to_mem(s_save_tmp, s_save_buf_rsrc0, s_save_mem_offset)
 	write_hwreg_to_mem(s_save_exec_lo, s_save_buf_rsrc0, s_save_mem_offset)
 	write_hwreg_to_mem(s_save_exec_hi, s_save_buf_rsrc0, s_save_mem_offset)
 	write_hwreg_to_mem(s_save_status, s_save_buf_rsrc0, s_save_mem_offset)
 
-	s_getreg_b32	s_save_tmp, hwreg(HW_REG_TRAPSTS)
+	s_getreg_b32	s_save_tmp, hwreg(S_TRAPSTS_HWREG)
 	write_hwreg_to_mem(s_save_tmp, s_save_buf_rsrc0, s_save_mem_offset)
 
 	// Not used on Sienna_Cichlid but keep layout same for debugger.
@@ -418,9 +586,13 @@ L_SAVE_HWREG:
 	write_hwreg_to_mem(s_save_m0, s_save_buf_rsrc0, s_save_mem_offset)
 
 #if NO_SQC_STORE
-	// Write HWREG/SGPRs with 32 VGPR lanes, wave32 is common case.
+	// Write HWREGs with 16 VGPR lanes. TTMPs occupy space after this.
+	s_mov_b32       exec_lo, 0xFFFF
 	s_mov_b32	exec_hi, 0x0
-	buffer_store_dword	v2, v0, s_save_buf_rsrc0, s_save_mem_offset slc:1 glc:1
+	buffer_store_dword	v2, v0, s_save_buf_rsrc0, s_save_mem_offset V_COHERENCE
+
+	// Write SGPRs with 32 VGPR lanes. This works in wave32 and wave64 mode.
+	s_mov_b32       exec_lo, 0xFFFFFFFF
 #endif
 
 	/* save SGPRs */
@@ -460,7 +632,7 @@ L_SAVE_SGPR_LOOP:
 	s_cmp_eq_u32	ttmp13, 0x20						//have 32 VGPR lanes filled?
 	s_cbranch_scc0	L_SAVE_SGPR_SKIP_TCP_STORE
 
-	buffer_store_dword	v2, v0, s_save_buf_rsrc0, s_save_mem_offset slc:1 glc:1
+	buffer_store_dword	v2, v0, s_save_buf_rsrc0, s_save_mem_offset V_COHERENCE
 	s_add_u32	s_save_mem_offset, s_save_mem_offset, 0x80
 	s_mov_b32	ttmp13, 0x0
 	v_mov_b32	v2, 0x0
@@ -481,7 +653,7 @@ L_SAVE_SGPR_SKIP_TCP_STORE:
 	write_12sgpr_to_mem(s0, s_save_buf_rsrc0, s_save_mem_offset)
 
 #if NO_SQC_STORE
-	buffer_store_dword	v2, v0, s_save_buf_rsrc0, s_save_mem_offset slc:1 glc:1
+	buffer_store_dword	v2, v0, s_save_buf_rsrc0, s_save_mem_offset V_COHERENCE
 #else
 	// restore s_save_buf_rsrc0,1
 	s_mov_b32	s_save_buf_rsrc0, s_save_xnack_mask
@@ -506,13 +678,12 @@ L_SAVE_LDS_NORMAL:
 	s_cbranch_scc0	L_SAVE_LDS_DONE						//no lds used? jump to L_SAVE_DONE
 
 	s_barrier								//LDS is used? wait for other waves in the same TG
-	s_and_b32	s_save_tmp, s_wave_size, S_SAVE_SPI_INIT_FIRST_WAVE_MASK
+	s_and_b32	s_save_tmp, s_save_pc_hi, S_SAVE_PC_HI_FIRST_WAVE_MASK
 	s_cbranch_scc0	L_SAVE_LDS_DONE
 
 	// first wave do LDS save;
 
-	s_lshl_b32	s_save_alloc_size, s_save_alloc_size, 6			//LDS size in dwords = lds_size * 64dw
-	s_lshl_b32	s_save_alloc_size, s_save_alloc_size, 2			//LDS size in bytes
+	s_lshl_b32	s_save_alloc_size, s_save_alloc_size, SQ_WAVE_LDS_ALLOC_GRANULARITY
 	s_mov_b32	s_save_buf_rsrc2, s_save_alloc_size			//NUM_RECORDS in bytes
 
 	// LDS at offset: size(VGPR)+size(SVGPR)+SIZE(SGPR)+SIZE(HWREG)
@@ -537,16 +708,36 @@ L_SAVE_LDS_NORMAL:
 	s_cbranch_scc1	L_SAVE_LDS_W64
 
 L_SAVE_LDS_W32:
+#if SAVE_AFTER_XNACK_ERROR
+	check_if_tcp_store_ok()
+	s_cbranch_scc1 L_SAVE_LDS_WITH_TCP_W32
+
+L_SAVE_LDS_LOOP_SQC_W32:
+	ds_read_b32	v1, v0
+	S_WAITCNT_0
+
+	write_vgprs_to_mem_with_sqc_w32(v1, 1, s_save_buf_rsrc0, s_save_mem_offset)
+
+	s_add_u32	m0, m0, 128						//every buffer_store_lds does 128 bytes
+	v_add_nc_u32	v0, v0, 128						//mem offset increased by 128 bytes
+	s_cmp_lt_u32	m0, s_save_alloc_size					//scc=(m0 < s_save_alloc_size) ? 1 : 0
+	s_cbranch_scc1	L_SAVE_LDS_LOOP_SQC_W32					//LDS save is complete?
+
+	s_branch	L_SAVE_LDS_DONE
+
+L_SAVE_LDS_WITH_TCP_W32:
+#endif
+
 	s_mov_b32	s3, 128
 	s_nop		0
 	s_nop		0
 	s_nop		0
 L_SAVE_LDS_LOOP_W32:
 	ds_read_b32	v1, v0
-	s_waitcnt	0
-	buffer_store_dword	v1, v0, s_save_buf_rsrc0, s_save_mem_offset slc:1 glc:1
+	S_WAITCNT_0
+	buffer_store_dword	v1, v0, s_save_buf_rsrc0, s_save_mem_offset V_COHERENCE
 
-	s_add_u32	m0, m0, s3						//every buffer_store_lds does 256 bytes
+	s_add_u32	m0, m0, s3						//every buffer_store_lds does 128 bytes
 	s_add_u32	s_save_mem_offset, s_save_mem_offset, s3
 	v_add_nc_u32	v0, v0, 128						//mem offset increased by 128 bytes
 	s_cmp_lt_u32	m0, s_save_alloc_size					//scc=(m0 < s_save_alloc_size) ? 1 : 0
@@ -555,14 +746,34 @@ L_SAVE_LDS_LOOP_W32:
 	s_branch	L_SAVE_LDS_DONE
 
 L_SAVE_LDS_W64:
+#if  SAVE_AFTER_XNACK_ERROR
+	check_if_tcp_store_ok()
+	s_cbranch_scc1 L_SAVE_LDS_WITH_TCP_W64
+
+L_SAVE_LDS_LOOP_SQC_W64:
+	ds_read_b32	v1, v0
+	S_WAITCNT_0
+
+	write_vgprs_to_mem_with_sqc_w64(v1, 1, s_save_buf_rsrc0, s_save_mem_offset)
+
+	s_add_u32	m0, m0, 256						//every buffer_store_lds does 256 bytes
+	v_add_nc_u32	v0, v0, 256						//mem offset increased by 256 bytes
+	s_cmp_lt_u32	m0, s_save_alloc_size					//scc=(m0 < s_save_alloc_size) ? 1 : 0
+	s_cbranch_scc1	L_SAVE_LDS_LOOP_SQC_W64					//LDS save is complete?
+
+	s_branch	L_SAVE_LDS_DONE
+
+L_SAVE_LDS_WITH_TCP_W64:
+#endif
+
 	s_mov_b32	s3, 256
 	s_nop		0
 	s_nop		0
 	s_nop		0
 L_SAVE_LDS_LOOP_W64:
 	ds_read_b32	v1, v0
-	s_waitcnt	0
-	buffer_store_dword	v1, v0, s_save_buf_rsrc0, s_save_mem_offset slc:1 glc:1
+	S_WAITCNT_0
+	buffer_store_dword	v1, v0, s_save_buf_rsrc0, s_save_mem_offset V_COHERENCE
 
 	s_add_u32	m0, m0, s3						//every buffer_store_lds does 256 bytes
 	s_add_u32	s_save_mem_offset, s_save_mem_offset, s3
@@ -604,16 +815,35 @@ L_SAVE_VGPR_NORMAL:
 	s_cmp_lt_u32	m0, s_save_alloc_size
 	s_cbranch_scc0	L_SAVE_VGPR_END
 
+#if  SAVE_AFTER_XNACK_ERROR
+	check_if_tcp_store_ok()
+	s_cbranch_scc1 L_SAVE_VGPR_W32_LOOP
+
+L_SAVE_VGPR_LOOP_SQC_W32:
+	v_movrels_b32	v0, v0							//v0 = v[0+m0]
+	v_movrels_b32	v1, v1							//v1 = v[1+m0]
+	v_movrels_b32	v2, v2							//v2 = v[2+m0]
+	v_movrels_b32	v3, v3							//v3 = v[3+m0]
+
+	write_vgprs_to_mem_with_sqc_w32(v0, 4, s_save_buf_rsrc0, s_save_mem_offset)
+
+	s_add_u32 m0, m0, 4
+	s_cmp_lt_u32 m0, s_save_alloc_size
+	s_cbranch_scc1 L_SAVE_VGPR_LOOP_SQC_W32
+
+	s_branch L_SAVE_VGPR_END
+#endif
+
 L_SAVE_VGPR_W32_LOOP:
 	v_movrels_b32	v0, v0							//v0 = v[0+m0]
 	v_movrels_b32	v1, v1							//v1 = v[1+m0]
 	v_movrels_b32	v2, v2							//v2 = v[2+m0]
 	v_movrels_b32	v3, v3							//v3 = v[3+m0]
 
-	buffer_store_dword	v0, v0, s_save_buf_rsrc0, s_save_mem_offset slc:1 glc:1
-	buffer_store_dword	v1, v0, s_save_buf_rsrc0, s_save_mem_offset slc:1 glc:1 offset:128
-	buffer_store_dword	v2, v0, s_save_buf_rsrc0, s_save_mem_offset slc:1 glc:1 offset:128*2
-	buffer_store_dword	v3, v0, s_save_buf_rsrc0, s_save_mem_offset slc:1 glc:1 offset:128*3
+	buffer_store_dword	v0, v0, s_save_buf_rsrc0, s_save_mem_offset V_COHERENCE
+	buffer_store_dword	v1, v0, s_save_buf_rsrc0, s_save_mem_offset V_COHERENCE offset:128
+	buffer_store_dword	v2, v0, s_save_buf_rsrc0, s_save_mem_offset V_COHERENCE offset:128*2
+	buffer_store_dword	v3, v0, s_save_buf_rsrc0, s_save_mem_offset V_COHERENCE offset:128*3
 
 	s_add_u32	m0, m0, 4						//next vgpr index
 	s_add_u32	s_save_mem_offset, s_save_mem_offset, 128*4		//every buffer_store_dword does 128 bytes
@@ -628,7 +858,26 @@ L_SAVE_VGPR_WAVE64:
 	// VGPR store using dw burst
 	s_mov_b32	m0, 0x4							//VGPR initial index value =4
 	s_cmp_lt_u32	m0, s_save_alloc_size
-	s_cbranch_scc0	L_SAVE_VGPR_END
+	s_cbranch_scc0	L_SAVE_SHARED_VGPR
+
+#if  SAVE_AFTER_XNACK_ERROR
+	check_if_tcp_store_ok()
+	s_cbranch_scc1 L_SAVE_VGPR_W64_LOOP
+
+L_SAVE_VGPR_LOOP_SQC_W64:
+	v_movrels_b32	v0, v0							//v0 = v[0+m0]
+	v_movrels_b32	v1, v1							//v1 = v[1+m0]
+	v_movrels_b32	v2, v2							//v2 = v[2+m0]
+	v_movrels_b32	v3, v3							//v3 = v[3+m0]
+
+	write_vgprs_to_mem_with_sqc_w64(v0, 4, s_save_buf_rsrc0, s_save_mem_offset)
+
+	s_add_u32 m0, m0, 4
+	s_cmp_lt_u32 m0, s_save_alloc_size
+	s_cbranch_scc1 L_SAVE_VGPR_LOOP_SQC_W64
+
+	s_branch L_SAVE_VGPR_END
+#endif
 
 L_SAVE_VGPR_W64_LOOP:
 	v_movrels_b32	v0, v0							//v0 = v[0+m0]
@@ -636,16 +885,17 @@ L_SAVE_VGPR_W64_LOOP:
 	v_movrels_b32	v2, v2							//v2 = v[2+m0]
 	v_movrels_b32	v3, v3							//v3 = v[3+m0]
 
-	buffer_store_dword	v0, v0, s_save_buf_rsrc0, s_save_mem_offset slc:1 glc:1
-	buffer_store_dword	v1, v0, s_save_buf_rsrc0, s_save_mem_offset slc:1 glc:1 offset:256
-	buffer_store_dword	v2, v0, s_save_buf_rsrc0, s_save_mem_offset slc:1 glc:1 offset:256*2
-	buffer_store_dword	v3, v0, s_save_buf_rsrc0, s_save_mem_offset slc:1 glc:1 offset:256*3
+	buffer_store_dword	v0, v0, s_save_buf_rsrc0, s_save_mem_offset V_COHERENCE
+	buffer_store_dword	v1, v0, s_save_buf_rsrc0, s_save_mem_offset V_COHERENCE offset:256
+	buffer_store_dword	v2, v0, s_save_buf_rsrc0, s_save_mem_offset V_COHERENCE offset:256*2
+	buffer_store_dword	v3, v0, s_save_buf_rsrc0, s_save_mem_offset V_COHERENCE offset:256*3
 
 	s_add_u32	m0, m0, 4						//next vgpr index
 	s_add_u32	s_save_mem_offset, s_save_mem_offset, 256*4		//every buffer_store_dword does 256 bytes
 	s_cmp_lt_u32	m0, s_save_alloc_size					//scc = (m0 < s_save_alloc_size) ? 1 : 0
 	s_cbranch_scc1	L_SAVE_VGPR_W64_LOOP					//VGPR save is complete?
 
+L_SAVE_SHARED_VGPR:
 	//Below part will be the save shared vgpr part (new for gfx10)
 	s_getreg_b32	s_save_alloc_size, hwreg(HW_REG_LDS_ALLOC,SQ_WAVE_LDS_ALLOC_VGPR_SHARED_SIZE_SHIFT,SQ_WAVE_LDS_ALLOC_VGPR_SHARED_SIZE_SIZE)
 	s_and_b32	s_save_alloc_size, s_save_alloc_size, 0xFFFFFFFF	//shared_vgpr_size is zero?
@@ -656,9 +906,26 @@ L_SAVE_VGPR_W64_LOOP:
 	s_add_u32	s_save_alloc_size, s_save_alloc_size, m0
 	s_mov_b32	exec_lo, 0xFFFFFFFF
 	s_mov_b32	exec_hi, 0x00000000
+
+#if  SAVE_AFTER_XNACK_ERROR
+	check_if_tcp_store_ok()
+	s_cbranch_scc1 L_SAVE_SHARED_VGPR_WAVE64_LOOP
+
+L_SAVE_SHARED_VGPR_WAVE64_LOOP_SQC:
+	v_movrels_b32	v0, v0
+
+	write_vgprs_to_mem_with_sqc_w64(v0, 1, s_save_buf_rsrc0, s_save_mem_offset)
+
+	s_add_u32 m0, m0, 1
+	s_cmp_lt_u32 m0, s_save_alloc_size
+	s_cbranch_scc1 L_SAVE_SHARED_VGPR_WAVE64_LOOP_SQC
+
+	s_branch L_SAVE_VGPR_END
+#endif
+
 L_SAVE_SHARED_VGPR_WAVE64_LOOP:
 	v_movrels_b32	v0, v0							//v0 = v[0+m0]
-	buffer_store_dword	v0, v0, s_save_buf_rsrc0, s_save_mem_offset slc:1 glc:1
+	buffer_store_dword	v0, v0, s_save_buf_rsrc0, s_save_mem_offset V_COHERENCE
 	s_add_u32	m0, m0, 1						//next vgpr index
 	s_add_u32	s_save_mem_offset, s_save_mem_offset, 128
 	s_cmp_lt_u32	m0, s_save_alloc_size					//scc = (m0 < s_save_alloc_size) ? 1 : 0
@@ -674,14 +941,9 @@ L_RESTORE:
 	s_or_b32	s_restore_buf_rsrc1, s_restore_buf_rsrc1, S_RESTORE_BUF_RSRC_WORD1_STRIDE
 	s_mov_b32	s_restore_buf_rsrc2, 0					//NUM_RECORDS initial value = 0 (in bytes)
 	s_mov_b32	s_restore_buf_rsrc3, S_RESTORE_BUF_RSRC_WORD3_MISC
-	s_and_b32	s_restore_tmp, s_restore_spi_init_hi, S_RESTORE_SPI_INIT_ATC_MASK
-	s_lshr_b32	s_restore_tmp, s_restore_tmp, (S_RESTORE_SPI_INIT_ATC_SHIFT-SQ_BUF_RSRC_WORD1_ATC_SHIFT)
-	s_or_b32	s_restore_buf_rsrc3, s_restore_buf_rsrc3, s_restore_tmp	//or ATC
-	s_and_b32	s_restore_tmp, s_restore_spi_init_hi, S_RESTORE_SPI_INIT_MTYPE_MASK
-	s_lshr_b32	s_restore_tmp, s_restore_tmp, (S_RESTORE_SPI_INIT_MTYPE_SHIFT-SQ_BUF_RSRC_WORD3_MTYPE_SHIFT)
-	s_or_b32	s_restore_buf_rsrc3, s_restore_buf_rsrc3, s_restore_tmp	//or MTYPE
+
 	//determine it is wave32 or wave64
-	get_wave_size(s_restore_size)
+	get_wave_size2(s_restore_size)
 
 	s_and_b32	s_restore_tmp, s_restore_spi_init_hi, S_RESTORE_SPI_INIT_FIRST_WAVE_MASK
 	s_cbranch_scc0	L_RESTORE_VGPR
@@ -701,8 +963,7 @@ L_RESTORE_LDS_NORMAL:
 	s_getreg_b32	s_restore_alloc_size, hwreg(HW_REG_LDS_ALLOC,SQ_WAVE_LDS_ALLOC_LDS_SIZE_SHIFT,SQ_WAVE_LDS_ALLOC_LDS_SIZE_SIZE)
 	s_and_b32	s_restore_alloc_size, s_restore_alloc_size, 0xFFFFFFFF	//lds_size is zero?
 	s_cbranch_scc0	L_RESTORE_VGPR						//no lds used? jump to L_RESTORE_VGPR
-	s_lshl_b32	s_restore_alloc_size, s_restore_alloc_size, 6		//LDS size in dwords = lds_size * 64dw
-	s_lshl_b32	s_restore_alloc_size, s_restore_alloc_size, 2		//LDS size in bytes
+	s_lshl_b32	s_restore_alloc_size, s_restore_alloc_size, SQ_WAVE_LDS_ALLOC_GRANULARITY
 	s_mov_b32	s_restore_buf_rsrc2, s_restore_alloc_size		//NUM_RECORDS in bytes
 
 	// LDS at offset: size(VGPR)+size(SVGPR)+SIZE(SGPR)+SIZE(HWREG)
@@ -722,7 +983,13 @@ L_RESTORE_LDS_NORMAL:
 	s_cbranch_scc1	L_RESTORE_LDS_LOOP_W64
 
 L_RESTORE_LDS_LOOP_W32:
+#if HAVE_BUFFER_LDS_LOAD
 	buffer_load_dword	v0, v0, s_restore_buf_rsrc0, s_restore_mem_offset lds:1	// first 64DW
+#else
+	buffer_load_dword       v0, v0, s_restore_buf_rsrc0, s_restore_mem_offset
+	S_WAITCNT_0
+	ds_store_addtid_b32     v0
+#endif
 	s_add_u32	m0, m0, 128						// 128 DW
 	s_add_u32	s_restore_mem_offset, s_restore_mem_offset, 128		//mem offset increased by 128DW
 	s_cmp_lt_u32	m0, s_restore_alloc_size				//scc=(m0 < s_restore_alloc_size) ? 1 : 0
@@ -730,7 +997,13 @@ L_RESTORE_LDS_LOOP_W32:
 	s_branch	L_RESTORE_VGPR
 
 L_RESTORE_LDS_LOOP_W64:
+#if HAVE_BUFFER_LDS_LOAD
 	buffer_load_dword	v0, v0, s_restore_buf_rsrc0, s_restore_mem_offset lds:1	// first 64DW
+#else
+	buffer_load_dword       v0, v0, s_restore_buf_rsrc0, s_restore_mem_offset
+	S_WAITCNT_0
+	ds_store_addtid_b32     v0
+#endif
 	s_add_u32	m0, m0, 256						// 256 DW
 	s_add_u32	s_restore_mem_offset, s_restore_mem_offset, 256		//mem offset increased by 256DW
 	s_cmp_lt_u32	m0, s_restore_alloc_size				//scc=(m0 < s_restore_alloc_size) ? 1 : 0
@@ -765,13 +1038,15 @@ L_RESTORE_VGPR_NORMAL:
 	s_mov_b32	s_restore_mem_offset_save, s_restore_mem_offset		// restore start with v1, v0 will be the last
 	s_add_u32	s_restore_mem_offset, s_restore_mem_offset, 128*4
 	s_mov_b32	m0, 4							//VGPR initial index value = 4
+	s_cmp_lt_u32	m0, s_restore_alloc_size
+	s_cbranch_scc0	L_RESTORE_SGPR
 
 L_RESTORE_VGPR_WAVE32_LOOP:
-	buffer_load_dword	v0, v0, s_restore_buf_rsrc0, s_restore_mem_offset slc:1 glc:1
-	buffer_load_dword	v1, v0, s_restore_buf_rsrc0, s_restore_mem_offset slc:1 glc:1 offset:128
-	buffer_load_dword	v2, v0, s_restore_buf_rsrc0, s_restore_mem_offset slc:1 glc:1 offset:128*2
-	buffer_load_dword	v3, v0, s_restore_buf_rsrc0, s_restore_mem_offset slc:1 glc:1 offset:128*3
-	s_waitcnt	vmcnt(0)
+	buffer_load_dword	v0, v0, s_restore_buf_rsrc0, s_restore_mem_offset V_COHERENCE
+	buffer_load_dword	v1, v0, s_restore_buf_rsrc0, s_restore_mem_offset V_COHERENCE offset:128
+	buffer_load_dword	v2, v0, s_restore_buf_rsrc0, s_restore_mem_offset V_COHERENCE offset:128*2
+	buffer_load_dword	v3, v0, s_restore_buf_rsrc0, s_restore_mem_offset V_COHERENCE offset:128*3
+	S_WAITCNT_0
 	v_movreld_b32	v0, v0							//v[0+m0] = v0
 	v_movreld_b32	v1, v1
 	v_movreld_b32	v2, v2
@@ -782,10 +1057,11 @@ L_RESTORE_VGPR_WAVE32_LOOP:
 	s_cbranch_scc1	L_RESTORE_VGPR_WAVE32_LOOP				//VGPR restore (except v0) is complete?
 
 	/* VGPR restore on v0 */
-	buffer_load_dword	v0, v0, s_restore_buf_rsrc0, s_restore_mem_offset_save slc:1 glc:1
-	buffer_load_dword	v1, v0, s_restore_buf_rsrc0, s_restore_mem_offset_save slc:1 glc:1 offset:128
-	buffer_load_dword	v2, v0, s_restore_buf_rsrc0, s_restore_mem_offset_save slc:1 glc:1 offset:128*2
-	buffer_load_dword	v3, v0, s_restore_buf_rsrc0, s_restore_mem_offset_save slc:1 glc:1 offset:128*3
+	buffer_load_dword	v0, v0, s_restore_buf_rsrc0, s_restore_mem_offset_save V_COHERENCE
+	buffer_load_dword	v1, v0, s_restore_buf_rsrc0, s_restore_mem_offset_save V_COHERENCE offset:128
+	buffer_load_dword	v2, v0, s_restore_buf_rsrc0, s_restore_mem_offset_save V_COHERENCE offset:128*2
+	buffer_load_dword	v3, v0, s_restore_buf_rsrc0, s_restore_mem_offset_save V_COHERENCE offset:128*3
+	S_WAITCNT_0
 
 	s_branch	L_RESTORE_SGPR
 
@@ -796,13 +1072,15 @@ L_RESTORE_VGPR_WAVE64:
 	s_mov_b32	s_restore_mem_offset_save, s_restore_mem_offset		// restore start with v4, v0 will be the last
 	s_add_u32	s_restore_mem_offset, s_restore_mem_offset, 256*4
 	s_mov_b32	m0, 4							//VGPR initial index value = 4
+	s_cmp_lt_u32	m0, s_restore_alloc_size
+	s_cbranch_scc0	L_RESTORE_SHARED_VGPR
 
 L_RESTORE_VGPR_WAVE64_LOOP:
-	buffer_load_dword	v0, v0, s_restore_buf_rsrc0, s_restore_mem_offset slc:1 glc:1
-	buffer_load_dword	v1, v0, s_restore_buf_rsrc0, s_restore_mem_offset slc:1 glc:1 offset:256
-	buffer_load_dword	v2, v0, s_restore_buf_rsrc0, s_restore_mem_offset slc:1 glc:1 offset:256*2
-	buffer_load_dword	v3, v0, s_restore_buf_rsrc0, s_restore_mem_offset slc:1 glc:1 offset:256*3
-	s_waitcnt	vmcnt(0)
+	buffer_load_dword	v0, v0, s_restore_buf_rsrc0, s_restore_mem_offset V_COHERENCE
+	buffer_load_dword	v1, v0, s_restore_buf_rsrc0, s_restore_mem_offset V_COHERENCE offset:256
+	buffer_load_dword	v2, v0, s_restore_buf_rsrc0, s_restore_mem_offset V_COHERENCE offset:256*2
+	buffer_load_dword	v3, v0, s_restore_buf_rsrc0, s_restore_mem_offset V_COHERENCE offset:256*3
+	S_WAITCNT_0
 	v_movreld_b32	v0, v0							//v[0+m0] = v0
 	v_movreld_b32	v1, v1
 	v_movreld_b32	v2, v2
@@ -812,6 +1090,7 @@ L_RESTORE_VGPR_WAVE64_LOOP:
 	s_cmp_lt_u32	m0, s_restore_alloc_size				//scc = (m0 < s_restore_alloc_size) ? 1 : 0
 	s_cbranch_scc1	L_RESTORE_VGPR_WAVE64_LOOP				//VGPR restore (except v0) is complete?
 
+L_RESTORE_SHARED_VGPR:
 	//Below part will be the restore shared vgpr part (new for gfx10)
 	s_getreg_b32	s_restore_alloc_size, hwreg(HW_REG_LDS_ALLOC,SQ_WAVE_LDS_ALLOC_VGPR_SHARED_SIZE_SHIFT,SQ_WAVE_LDS_ALLOC_VGPR_SHARED_SIZE_SIZE)	//shared_vgpr_size
 	s_and_b32	s_restore_alloc_size, s_restore_alloc_size, 0xFFFFFFFF	//shared_vgpr_size is zero?
@@ -823,8 +1102,8 @@ L_RESTORE_VGPR_WAVE64_LOOP:
 	s_mov_b32	exec_lo, 0xFFFFFFFF
 	s_mov_b32	exec_hi, 0x00000000
 L_RESTORE_SHARED_VGPR_WAVE64_LOOP:
-	buffer_load_dword	v0, v0, s_restore_buf_rsrc0, s_restore_mem_offset slc:1 glc:1
-	s_waitcnt	vmcnt(0)
+	buffer_load_dword	v0, v0, s_restore_buf_rsrc0, s_restore_mem_offset V_COHERENCE
+	S_WAITCNT_0
 	v_movreld_b32	v0, v0							//v[0+m0] = v0
 	s_add_u32	m0, m0, 1						//next vgpr index
 	s_add_u32	s_restore_mem_offset, s_restore_mem_offset, 128
@@ -835,11 +1114,11 @@ L_RESTORE_SHARED_VGPR_WAVE64_LOOP:
 
 	/* VGPR restore on v0 */
 L_RESTORE_V0:
-	buffer_load_dword	v0, v0, s_restore_buf_rsrc0, s_restore_mem_offset_save slc:1 glc:1
-	buffer_load_dword	v1, v0, s_restore_buf_rsrc0, s_restore_mem_offset_save slc:1 glc:1 offset:256
-	buffer_load_dword	v2, v0, s_restore_buf_rsrc0, s_restore_mem_offset_save slc:1 glc:1 offset:256*2
-	buffer_load_dword	v3, v0, s_restore_buf_rsrc0, s_restore_mem_offset_save slc:1 glc:1 offset:256*3
-	s_waitcnt	vmcnt(0)
+	buffer_load_dword	v0, v0, s_restore_buf_rsrc0, s_restore_mem_offset_save V_COHERENCE
+	buffer_load_dword	v1, v0, s_restore_buf_rsrc0, s_restore_mem_offset_save V_COHERENCE offset:256
+	buffer_load_dword	v2, v0, s_restore_buf_rsrc0, s_restore_mem_offset_save V_COHERENCE offset:256*2
+	buffer_load_dword	v3, v0, s_restore_buf_rsrc0, s_restore_mem_offset_save V_COHERENCE offset:256*3
+	S_WAITCNT_0
 
 	/* restore SGPRs */
 	//will be 2+8+16*6
@@ -856,7 +1135,7 @@ L_RESTORE_SGPR:
 	s_mov_b32	m0, s_sgpr_save_num
 
 	read_4sgpr_from_mem(s0, s_restore_buf_rsrc0, s_restore_mem_offset)
-	s_waitcnt	lgkmcnt(0)
+	S_WAITCNT_0
 
 	s_sub_u32	m0, m0, 4						// Restore from S[0] to S[104]
 	s_nop		0							// hazard SALU M0=> S_MOVREL
@@ -865,7 +1144,7 @@ L_RESTORE_SGPR:
 	s_movreld_b64	s2, s2
 
 	read_8sgpr_from_mem(s0, s_restore_buf_rsrc0, s_restore_mem_offset)
-	s_waitcnt	lgkmcnt(0)
+	S_WAITCNT_0
 
 	s_sub_u32	m0, m0, 8						// Restore from S[0] to S[96]
 	s_nop		0							// hazard SALU M0=> S_MOVREL
@@ -877,7 +1156,7 @@ L_RESTORE_SGPR:
 
  L_RESTORE_SGPR_LOOP:
 	read_16sgpr_from_mem(s0, s_restore_buf_rsrc0, s_restore_mem_offset)
-	s_waitcnt	lgkmcnt(0)
+	S_WAITCNT_0
 
 	s_sub_u32	m0, m0, 16						// Restore from S[n] to S[0]
 	s_nop		0							// hazard SALU M0=> S_MOVREL
@@ -919,12 +1198,12 @@ L_RESTORE_HWREG:
 	read_hwreg_from_mem(s_restore_xnack_mask, s_restore_buf_rsrc0, s_restore_mem_offset)
 	read_hwreg_from_mem(s_restore_mode, s_restore_buf_rsrc0, s_restore_mem_offset)
 	read_hwreg_from_mem(s_restore_flat_scratch, s_restore_buf_rsrc0, s_restore_mem_offset)
-	s_waitcnt	lgkmcnt(0)
+	S_WAITCNT_0
 
 	s_setreg_b32	hwreg(HW_REG_SHADER_FLAT_SCRATCH_LO), s_restore_flat_scratch
 
 	read_hwreg_from_mem(s_restore_flat_scratch, s_restore_buf_rsrc0, s_restore_mem_offset)
-	s_waitcnt	lgkmcnt(0)						//from now on, it is safe to restore STATUS and IB_STS
+	S_WAITCNT_0
 
 	s_setreg_b32	hwreg(HW_REG_SHADER_FLAT_SCRATCH_HI), s_restore_flat_scratch
 
@@ -932,59 +1211,63 @@ L_RESTORE_HWREG:
 	s_mov_b32	exec_lo, s_restore_exec_lo
 	s_mov_b32	exec_hi, s_restore_exec_hi
 
-	s_and_b32	s_restore_m0, SQ_WAVE_TRAPSTS_PRE_SAVECTX_MASK, s_restore_trapsts
-	s_setreg_b32	hwreg(HW_REG_TRAPSTS, SQ_WAVE_TRAPSTS_PRE_SAVECTX_SHIFT, SQ_WAVE_TRAPSTS_PRE_SAVECTX_SIZE), s_restore_m0
-
-#if ASIC_TARGET_NAVI1X
+#if HAVE_XNACK
 	s_setreg_b32	hwreg(HW_REG_SHADER_XNACK_MASK), s_restore_xnack_mask
 #endif
 
-	s_and_b32	s_restore_m0, SQ_WAVE_TRAPSTS_POST_SAVECTX_MASK, s_restore_trapsts
-	s_lshr_b32	s_restore_m0, s_restore_m0, SQ_WAVE_TRAPSTS_POST_SAVECTX_SHIFT
-	s_setreg_b32	hwreg(HW_REG_TRAPSTS, SQ_WAVE_TRAPSTS_POST_SAVECTX_SHIFT, SQ_WAVE_TRAPSTS_POST_SAVECTX_SIZE), s_restore_m0
+	// {TRAPSTS/EXCP_FLAG_PRIV}.SAVE_CONTEXT and HOST_TRAP may have changed.
+	// Only restore the other fields to avoid clobbering them.
+	s_setreg_b32	hwreg(S_TRAPSTS_HWREG, 0, S_TRAPSTS_RESTORE_PART_1_SIZE), s_restore_trapsts
+	s_lshr_b32	s_restore_trapsts, s_restore_trapsts, S_TRAPSTS_RESTORE_PART_2_SHIFT
+	s_setreg_b32	hwreg(S_TRAPSTS_HWREG, S_TRAPSTS_RESTORE_PART_2_SHIFT, S_TRAPSTS_RESTORE_PART_2_SIZE), s_restore_trapsts
+
+if S_TRAPSTS_RESTORE_PART_3_SIZE > 0
+	s_lshr_b32	s_restore_trapsts, s_restore_trapsts, S_TRAPSTS_RESTORE_PART_3_SHIFT - S_TRAPSTS_RESTORE_PART_2_SHIFT
+	s_setreg_b32	hwreg(S_TRAPSTS_HWREG, S_TRAPSTS_RESTORE_PART_3_SHIFT, S_TRAPSTS_RESTORE_PART_3_SIZE), s_restore_trapsts
+end
+
 	s_setreg_b32	hwreg(HW_REG_MODE), s_restore_mode
 
 	// Restore trap temporaries 4-11, 13 initialized by SPI debug dispatch logic
-	// ttmp SR memory offset : size(VGPR)+size(SGPR)+0x40
+	// ttmp SR memory offset : size(VGPR)+size(SVGPR)+size(SGPR)+0x40
 	get_vgpr_size_bytes(s_restore_ttmps_lo, s_restore_size)
+	get_svgpr_size_bytes(s_restore_ttmps_hi)
+	s_add_u32	s_restore_ttmps_lo, s_restore_ttmps_lo, s_restore_ttmps_hi
 	s_add_u32	s_restore_ttmps_lo, s_restore_ttmps_lo, get_sgpr_size_bytes()
 	s_add_u32	s_restore_ttmps_lo, s_restore_ttmps_lo, s_restore_buf_rsrc0
 	s_addc_u32	s_restore_ttmps_hi, s_restore_buf_rsrc1, 0x0
 	s_and_b32	s_restore_ttmps_hi, s_restore_ttmps_hi, 0xFFFF
-	s_load_dwordx4	[ttmp4, ttmp5, ttmp6, ttmp7], [s_restore_ttmps_lo, s_restore_ttmps_hi], 0x50 glc:1
-	s_load_dwordx4	[ttmp8, ttmp9, ttmp10, ttmp11], [s_restore_ttmps_lo, s_restore_ttmps_hi], 0x60 glc:1
-	s_load_dword	ttmp13, [s_restore_ttmps_lo, s_restore_ttmps_hi], 0x74 glc:1
-	s_waitcnt	lgkmcnt(0)
-
-#if ASIC_TARGET_NAVI1X
-	s_and_b32	s_restore_m0, s_restore_pc_hi, S_SAVE_PC_HI_RCNT_MASK
-	s_lshr_b32	s_restore_m0, s_restore_m0, S_SAVE_PC_HI_RCNT_SHIFT
-	s_lshl_b32	s_restore_m0, s_restore_m0, SQ_WAVE_IB_STS_RCNT_SHIFT
-	s_mov_b32	s_restore_tmp, 0x0
-	s_or_b32	s_restore_tmp, s_restore_tmp, s_restore_m0
-	s_and_b32	s_restore_m0, s_restore_pc_hi, S_SAVE_PC_HI_FIRST_REPLAY_MASK
-	s_lshr_b32	s_restore_m0, s_restore_m0, S_SAVE_PC_HI_FIRST_REPLAY_SHIFT
-	s_lshl_b32	s_restore_m0, s_restore_m0, SQ_WAVE_IB_STS_FIRST_REPLAY_SHIFT
-	s_or_b32	s_restore_tmp, s_restore_tmp, s_restore_m0
-	s_and_b32	s_restore_m0, s_restore_pc_hi, S_SAVE_PC_HI_REPLAY_W64H_MASK
-	s_lshr_b32	s_restore_m0, s_restore_m0, S_SAVE_PC_HI_REPLAY_W64H_SHIFT
-	s_lshl_b32	s_restore_m0, s_restore_m0, SQ_WAVE_IB_STS_REPLAY_W64H_SHIFT
-	s_or_b32	s_restore_tmp, s_restore_tmp, s_restore_m0
-
-	s_and_b32	s_restore_m0, s_restore_status, SQ_WAVE_STATUS_INST_ATC_MASK
-	s_lshr_b32	s_restore_m0, s_restore_m0, SQ_WAVE_STATUS_INST_ATC_SHIFT
-	s_setreg_b32 	hwreg(HW_REG_IB_STS), s_restore_tmp
+	s_load_dwordx4	[ttmp4, ttmp5, ttmp6, ttmp7], [s_restore_ttmps_lo, s_restore_ttmps_hi], 0x50 S_COHERENCE
+	s_load_dwordx4	[ttmp8, ttmp9, ttmp10, ttmp11], [s_restore_ttmps_lo, s_restore_ttmps_hi], 0x60 S_COHERENCE
+	s_load_dword	ttmp13, [s_restore_ttmps_lo, s_restore_ttmps_hi], 0x74 S_COHERENCE
+	S_WAITCNT_0
+
+#if HAVE_XNACK
+	restore_ib_sts(s_restore_tmp, s_restore_m0)
 #endif
 
 	s_and_b32	s_restore_pc_hi, s_restore_pc_hi, 0x0000ffff		//pc[47:32] //Do it here in order not to affect STATUS
 	s_and_b64	exec, exec, exec					// Restore STATUS.EXECZ, not writable by s_setreg_b32
 	s_and_b64	vcc, vcc, vcc						// Restore STATUS.VCCZ, not writable by s_setreg_b32
+
+#if SW_SA_TRAP
+	// If traps are enabled then return to the shader with PRIV=0.
+	// Otherwise retain PRIV=1 for subsequent context save requests.
+	s_getreg_b32	s_restore_tmp, hwreg(HW_REG_STATUS)
+	s_bitcmp1_b32	s_restore_tmp, SQ_WAVE_STATUS_TRAP_EN_SHIFT
+	s_cbranch_scc1	L_RETURN_WITHOUT_PRIV
+
 	s_setreg_b32	hwreg(HW_REG_STATUS), s_restore_status			// SCC is included, which is changed by previous salu
+	s_setpc_b64	[s_restore_pc_lo, s_restore_pc_hi]
+L_RETURN_WITHOUT_PRIV:
+#endif
+
+	s_setreg_b32	hwreg(S_STATUS_HWREG), s_restore_status			// SCC is included, which is changed by previous salu
 
 	s_rfe_b64	s_restore_pc_lo						//Return to the main shader program and resume execution
 
 L_END_PGM:
-	s_endpgm
+	s_endpgm_saved
 end
 
 function write_hwreg_to_mem(s, s_rsrc, s_mem_offset)
@@ -995,7 +1278,7 @@ function write_hwreg_to_mem(s, s_rsrc, s_mem_offset)
 #else
 	s_mov_b32	exec_lo, m0
 	s_mov_b32	m0, s_mem_offset
-	s_buffer_store_dword	s, s_rsrc, m0 glc:1
+	s_buffer_store_dword	s, s_rsrc, m0 S_COHERENCE
 	s_add_u32	s_mem_offset, s_mem_offset, 4
 	s_mov_b32	m0, exec_lo
 #endif
@@ -1010,10 +1293,10 @@ function write_16sgpr_to_mem(s, s_rsrc, s_mem_offset)
 		s_add_u32	ttmp13, ttmp13, 0x1
 	end
 #else
-	s_buffer_store_dwordx4	s[0], s_rsrc, 0 glc:1
-	s_buffer_store_dwordx4	s[4], s_rsrc, 16 glc:1
-	s_buffer_store_dwordx4	s[8], s_rsrc, 32 glc:1
-	s_buffer_store_dwordx4	s[12], s_rsrc, 48 glc:1
+	s_buffer_store_dwordx4	s[0], s_rsrc, 0 S_COHERENCE
+	s_buffer_store_dwordx4	s[4], s_rsrc, 16 S_COHERENCE
+	s_buffer_store_dwordx4	s[8], s_rsrc, 32 S_COHERENCE
+	s_buffer_store_dwordx4	s[12], s_rsrc, 48 S_COHERENCE
 	s_add_u32	s_rsrc[0], s_rsrc[0], 4*16
 	s_addc_u32	s_rsrc[1], s_rsrc[1], 0x0
 #endif
@@ -1027,40 +1310,72 @@ function write_12sgpr_to_mem(s, s_rsrc, s_mem_offset)
 		s_add_u32	ttmp13, ttmp13, 0x1
 	end
 #else
-	s_buffer_store_dwordx4	s[0], s_rsrc, 0 glc:1
-	s_buffer_store_dwordx4	s[4], s_rsrc, 16 glc:1
-	s_buffer_store_dwordx4	s[8], s_rsrc, 32 glc:1
+	s_buffer_store_dwordx4	s[0], s_rsrc, 0 S_COHERENCE
+	s_buffer_store_dwordx4	s[4], s_rsrc, 16 S_COHERENCE
+	s_buffer_store_dwordx4	s[8], s_rsrc, 32 S_COHERENCE
 	s_add_u32	s_rsrc[0], s_rsrc[0], 4*12
 	s_addc_u32	s_rsrc[1], s_rsrc[1], 0x0
 #endif
 end
 
 function read_hwreg_from_mem(s, s_rsrc, s_mem_offset)
-	s_buffer_load_dword	s, s_rsrc, s_mem_offset glc:1
+	s_buffer_load_dword	s, s_rsrc, s_mem_offset S_COHERENCE
 	s_add_u32	s_mem_offset, s_mem_offset, 4
 end
 
 function read_16sgpr_from_mem(s, s_rsrc, s_mem_offset)
 	s_sub_u32	s_mem_offset, s_mem_offset, 4*16
-	s_buffer_load_dwordx16	s, s_rsrc, s_mem_offset glc:1
+	s_buffer_load_dwordx16	s, s_rsrc, s_mem_offset S_COHERENCE
 end
 
 function read_8sgpr_from_mem(s, s_rsrc, s_mem_offset)
 	s_sub_u32	s_mem_offset, s_mem_offset, 4*8
-	s_buffer_load_dwordx8	s, s_rsrc, s_mem_offset glc:1
+	s_buffer_load_dwordx8	s, s_rsrc, s_mem_offset S_COHERENCE
 end
 
 function read_4sgpr_from_mem(s, s_rsrc, s_mem_offset)
 	s_sub_u32	s_mem_offset, s_mem_offset, 4*4
-	s_buffer_load_dwordx4	s, s_rsrc, s_mem_offset glc:1
+	s_buffer_load_dwordx4	s, s_rsrc, s_mem_offset S_COHERENCE
 end
 
+#if SAVE_AFTER_XNACK_ERROR
+function check_if_tcp_store_ok
+	// If TRAPSTS.XNACK_ERROR=1 then TCP stores will fail.
+	s_getreg_b32 s_save_tmp, hwreg(HW_REG_TRAPSTS)
+	s_andn2_b32 s_save_tmp, SQ_WAVE_TRAPSTS_XNACK_ERROR_MASK, s_save_tmp
 
-function get_lds_size_bytes(s_lds_size_byte)
-	s_getreg_b32	s_lds_size_byte, hwreg(HW_REG_LDS_ALLOC, SQ_WAVE_LDS_ALLOC_LDS_SIZE_SHIFT, SQ_WAVE_LDS_ALLOC_LDS_SIZE_SIZE)
-	s_lshl_b32	s_lds_size_byte, s_lds_size_byte, 8			//LDS size in dwords = lds_size * 64 *4Bytes // granularity 64DW
+L_TCP_STORE_CHECK_DONE:
 end
 
+function write_vgpr_to_mem_with_sqc(vgpr, n_lanes, s_rsrc, s_mem_offset)
+	s_mov_b32 s4, 0
+
+L_WRITE_VGPR_LANE_LOOP:
+	for var lane = 0; lane < 4; ++lane
+		v_readlane_b32 s[lane], vgpr, s4
+		s_add_u32 s4, s4, 1
+	end
+
+	s_buffer_store_dwordx4 s[0:3], s_rsrc, s_mem_offset glc:1
+
+	s_add_u32 s_mem_offset, s_mem_offset, 0x10
+	s_cmp_eq_u32 s4, n_lanes
+	s_cbranch_scc0 L_WRITE_VGPR_LANE_LOOP
+end
+
+function write_vgprs_to_mem_with_sqc_w32(vgpr0, n_vgprs, s_rsrc, s_mem_offset)
+	for var vgpr = 0; vgpr < n_vgprs; ++vgpr
+		write_vgpr_to_mem_with_sqc(vgpr0[vgpr], 32, s_rsrc, s_mem_offset)
+	end
+end
+
+function write_vgprs_to_mem_with_sqc_w64(vgpr0, n_vgprs, s_rsrc, s_mem_offset)
+	for var vgpr = 0; vgpr < n_vgprs; ++vgpr
+		write_vgpr_to_mem_with_sqc(vgpr0[vgpr], 64, s_rsrc, s_mem_offset)
+	end
+end
+#endif
+
 function get_vgpr_size_bytes(s_vgpr_size_byte, s_size)
 	s_getreg_b32	s_vgpr_size_byte, hwreg(HW_REG_GPR_ALLOC,SQ_WAVE_GPR_ALLOC_VGPR_SIZE_SHIFT,SQ_WAVE_GPR_ALLOC_VGPR_SIZE_SIZE)
 	s_add_u32	s_vgpr_size_byte, s_vgpr_size_byte, 1
@@ -1086,8 +1401,34 @@ function get_hwreg_size_bytes
 	return 128
 end
 
-function get_wave_size(s_reg)
+function get_wave_size2(s_reg)
 	s_getreg_b32	s_reg, hwreg(HW_REG_IB_STS2,SQ_WAVE_IB_STS2_WAVE64_SHIFT,SQ_WAVE_IB_STS2_WAVE64_SIZE)
 	s_lshl_b32	s_reg, s_reg, S_WAVE_SIZE
-	s_or_b32	s_reg, s_save_spi_init_hi, s_reg			//share with exec_hi, it's at bit25
 end
+
+#if HAVE_XNACK
+function save_and_clear_ib_sts(tmp1, tmp2)
+	// Preserve and clear scalar XNACK state before issuing scalar loads.
+	// Save IB_STS.REPLAY_W64H[25], RCNT[21:16], FIRST_REPLAY[15] into
+	// unused space ttmp11[31:24].
+	s_andn2_b32	ttmp11, ttmp11, (TTMP11_SAVE_REPLAY_W64H_MASK | TTMP11_SAVE_RCNT_FIRST_REPLAY_MASK)
+	s_getreg_b32	tmp1, hwreg(HW_REG_IB_STS)
+	s_and_b32	tmp2, tmp1, SQ_WAVE_IB_STS_REPLAY_W64H_MASK
+	s_lshl_b32	tmp2, tmp2, (TTMP11_SAVE_REPLAY_W64H_SHIFT - SQ_WAVE_IB_STS_REPLAY_W64H_SHIFT)
+	s_or_b32	ttmp11, ttmp11, tmp2
+	s_and_b32	tmp2, tmp1, SQ_WAVE_IB_STS_RCNT_FIRST_REPLAY_MASK
+	s_lshl_b32	tmp2, tmp2, (TTMP11_SAVE_RCNT_FIRST_REPLAY_SHIFT - SQ_WAVE_IB_STS_FIRST_REPLAY_SHIFT)
+	s_or_b32	ttmp11, ttmp11, tmp2
+	s_andn2_b32	tmp1, tmp1, (SQ_WAVE_IB_STS_REPLAY_W64H_MASK | SQ_WAVE_IB_STS_RCNT_FIRST_REPLAY_MASK)
+	s_setreg_b32	hwreg(HW_REG_IB_STS), tmp1
+end
+
+function restore_ib_sts(tmp1, tmp2)
+	s_lshr_b32	tmp1, ttmp11, (TTMP11_SAVE_RCNT_FIRST_REPLAY_SHIFT - SQ_WAVE_IB_STS_FIRST_REPLAY_SHIFT)
+	s_and_b32	tmp2, tmp1, SQ_WAVE_IB_STS_RCNT_FIRST_REPLAY_MASK
+	s_lshr_b32	tmp1, ttmp11, (TTMP11_SAVE_REPLAY_W64H_SHIFT - SQ_WAVE_IB_STS_REPLAY_W64H_SHIFT)
+	s_and_b32	tmp1, tmp1, SQ_WAVE_IB_STS_REPLAY_W64H_MASK
+	s_or_b32	tmp1, tmp1, tmp2
+	s_setreg_b32	hwreg(HW_REG_IB_STS), tmp1
+end
+#endif
diff --git a/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler_gfx12.asm b/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler_gfx12.asm
new file mode 100644
index 000000000000..5a1a1b1f897f
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler_gfx12.asm
@@ -0,0 +1,1136 @@
+/*
+ * Copyright 2018 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+/* To compile this assembly code:
+ *
+ * gfx12:
+ *   cpp -DASIC_FAMILY=CHIP_GFX12 cwsr_trap_handler_gfx12.asm -P -o gfx12.sp3
+ *   sp3 gfx12.sp3 -hex gfx12.hex
+ */
+
+#define CHIP_GFX12 37
+
+#define SINGLE_STEP_MISSED_WORKAROUND 1	//workaround for lost TRAP_AFTER_INST exception when SAVECTX raised
+#define HAVE_VALU_SGPR_HAZARD (ASIC_FAMILY == CHIP_GFX12)
+
+var SQ_WAVE_STATE_PRIV_BARRIER_COMPLETE_MASK	= 0x4
+var SQ_WAVE_STATE_PRIV_SCC_SHIFT		= 9
+var SQ_WAVE_STATE_PRIV_SYS_PRIO_MASK		= 0xC00
+var SQ_WAVE_STATE_PRIV_HALT_MASK		= 0x4000
+var SQ_WAVE_STATE_PRIV_POISON_ERR_MASK		= 0x8000
+var SQ_WAVE_STATE_PRIV_POISON_ERR_SHIFT		= 15
+var SQ_WAVE_STATUS_WAVE64_SHIFT			= 29
+var SQ_WAVE_STATUS_WAVE64_SIZE			= 1
+var SQ_WAVE_STATUS_NO_VGPRS_SHIFT		= 24
+var SQ_WAVE_STATE_PRIV_ALWAYS_CLEAR_MASK	= SQ_WAVE_STATE_PRIV_SYS_PRIO_MASK|SQ_WAVE_STATE_PRIV_POISON_ERR_MASK
+var S_SAVE_PC_HI_TRAP_ID_MASK			= 0xF0000000
+
+var SQ_WAVE_LDS_ALLOC_LDS_SIZE_SHIFT		= 12
+var SQ_WAVE_LDS_ALLOC_LDS_SIZE_SIZE		= 9
+var SQ_WAVE_GPR_ALLOC_VGPR_SIZE_SIZE		= 8
+var SQ_WAVE_GPR_ALLOC_VGPR_SIZE_SHIFT		= 12
+var SQ_WAVE_LDS_ALLOC_VGPR_SHARED_SIZE_SHIFT	= 24
+var SQ_WAVE_LDS_ALLOC_VGPR_SHARED_SIZE_SIZE	= 4
+var SQ_WAVE_LDS_ALLOC_GRANULARITY		= 9
+
+var SQ_WAVE_EXCP_FLAG_PRIV_ADDR_WATCH_MASK	= 0xF
+var SQ_WAVE_EXCP_FLAG_PRIV_MEM_VIOL_MASK	= 0x10
+var SQ_WAVE_EXCP_FLAG_PRIV_SAVE_CONTEXT_SHIFT	= 5
+var SQ_WAVE_EXCP_FLAG_PRIV_SAVE_CONTEXT_MASK	= 0x20
+var SQ_WAVE_EXCP_FLAG_PRIV_ILLEGAL_INST_MASK	= 0x40
+var SQ_WAVE_EXCP_FLAG_PRIV_ILLEGAL_INST_SHIFT	= 6
+var SQ_WAVE_EXCP_FLAG_PRIV_HOST_TRAP_MASK	= 0x80
+var SQ_WAVE_EXCP_FLAG_PRIV_HOST_TRAP_SHIFT	= 7
+var SQ_WAVE_EXCP_FLAG_PRIV_WAVE_START_MASK	= 0x100
+var SQ_WAVE_EXCP_FLAG_PRIV_WAVE_START_SHIFT	= 8
+var SQ_WAVE_EXCP_FLAG_PRIV_WAVE_END_MASK	= 0x200
+var SQ_WAVE_EXCP_FLAG_PRIV_TRAP_AFTER_INST_MASK	= 0x800
+var SQ_WAVE_TRAP_CTRL_ADDR_WATCH_MASK		= 0x80
+var SQ_WAVE_TRAP_CTRL_TRAP_AFTER_INST_MASK	= 0x200
+
+var SQ_WAVE_EXCP_FLAG_PRIV_NON_MASKABLE_EXCP_MASK= SQ_WAVE_EXCP_FLAG_PRIV_MEM_VIOL_MASK		|\
+						  SQ_WAVE_EXCP_FLAG_PRIV_ILLEGAL_INST_MASK	|\
+						  SQ_WAVE_EXCP_FLAG_PRIV_HOST_TRAP_MASK		|\
+						  SQ_WAVE_EXCP_FLAG_PRIV_WAVE_START_MASK	|\
+						  SQ_WAVE_EXCP_FLAG_PRIV_WAVE_END_MASK		|\
+						  SQ_WAVE_EXCP_FLAG_PRIV_TRAP_AFTER_INST_MASK
+var SQ_WAVE_EXCP_FLAG_PRIV_RESTORE_PART_1_SIZE	= SQ_WAVE_EXCP_FLAG_PRIV_SAVE_CONTEXT_SHIFT
+var SQ_WAVE_EXCP_FLAG_PRIV_RESTORE_PART_2_SHIFT	= SQ_WAVE_EXCP_FLAG_PRIV_ILLEGAL_INST_SHIFT
+var SQ_WAVE_EXCP_FLAG_PRIV_RESTORE_PART_2_SIZE	= SQ_WAVE_EXCP_FLAG_PRIV_HOST_TRAP_SHIFT - SQ_WAVE_EXCP_FLAG_PRIV_ILLEGAL_INST_SHIFT
+var SQ_WAVE_EXCP_FLAG_PRIV_RESTORE_PART_3_SHIFT	= SQ_WAVE_EXCP_FLAG_PRIV_WAVE_START_SHIFT
+var SQ_WAVE_EXCP_FLAG_PRIV_RESTORE_PART_3_SIZE	= 32 - SQ_WAVE_EXCP_FLAG_PRIV_RESTORE_PART_3_SHIFT
+var BARRIER_STATE_SIGNAL_OFFSET			= 16
+var BARRIER_STATE_VALID_OFFSET			= 0
+
+var TTMP11_DEBUG_TRAP_ENABLED_SHIFT		= 23
+var TTMP11_DEBUG_TRAP_ENABLED_MASK		= 0x800000
+
+// SQ_SEL_X/Y/Z/W, BUF_NUM_FORMAT_FLOAT, (0 for MUBUF stride[17:14]
+// when ADD_TID_ENABLE and BUF_DATA_FORMAT_32 for MTBUF), ADD_TID_ENABLE
+var S_SAVE_BUF_RSRC_WORD1_STRIDE		= 0x00040000
+var S_SAVE_BUF_RSRC_WORD3_MISC			= 0x10807FAC
+var S_SAVE_SPI_INIT_FIRST_WAVE_MASK		= 0x04000000
+var S_SAVE_SPI_INIT_FIRST_WAVE_SHIFT		= 26
+
+var S_SAVE_PC_HI_FIRST_WAVE_MASK		= 0x80000000
+var S_SAVE_PC_HI_FIRST_WAVE_SHIFT		= 31
+
+var s_sgpr_save_num				= 108
+
+var s_save_spi_init_lo				= exec_lo
+var s_save_spi_init_hi				= exec_hi
+var s_save_pc_lo				= ttmp0
+var s_save_pc_hi				= ttmp1
+var s_save_exec_lo				= ttmp2
+var s_save_exec_hi				= ttmp3
+var s_save_state_priv				= ttmp12
+var s_save_excp_flag_priv			= ttmp15
+var s_save_xnack_mask				= s_save_excp_flag_priv
+var s_wave_size					= ttmp7
+var s_save_buf_rsrc0				= ttmp8
+var s_save_buf_rsrc1				= ttmp9
+var s_save_buf_rsrc2				= ttmp10
+var s_save_buf_rsrc3				= ttmp11
+var s_save_mem_offset				= ttmp4
+var s_save_alloc_size				= s_save_excp_flag_priv
+var s_save_tmp					= ttmp14
+var s_save_m0					= ttmp5
+var s_save_ttmps_lo				= s_save_tmp
+var s_save_ttmps_hi				= s_save_excp_flag_priv
+
+var S_RESTORE_BUF_RSRC_WORD1_STRIDE		= S_SAVE_BUF_RSRC_WORD1_STRIDE
+var S_RESTORE_BUF_RSRC_WORD3_MISC		= S_SAVE_BUF_RSRC_WORD3_MISC
+
+var S_RESTORE_SPI_INIT_FIRST_WAVE_MASK		= 0x04000000
+var S_RESTORE_SPI_INIT_FIRST_WAVE_SHIFT		= 26
+var S_WAVE_SIZE					= 25
+
+var s_restore_spi_init_lo			= exec_lo
+var s_restore_spi_init_hi			= exec_hi
+var s_restore_mem_offset			= ttmp12
+var s_restore_alloc_size			= ttmp3
+var s_restore_tmp				= ttmp2
+var s_restore_mem_offset_save			= s_restore_tmp
+var s_restore_m0				= s_restore_alloc_size
+var s_restore_mode				= ttmp7
+var s_restore_flat_scratch			= s_restore_tmp
+var s_restore_pc_lo				= ttmp0
+var s_restore_pc_hi				= ttmp1
+var s_restore_exec_lo				= ttmp4
+var s_restore_exec_hi				= ttmp5
+var s_restore_state_priv			= ttmp14
+var s_restore_excp_flag_priv			= ttmp15
+var s_restore_xnack_mask			= ttmp13
+var s_restore_buf_rsrc0				= ttmp8
+var s_restore_buf_rsrc1				= ttmp9
+var s_restore_buf_rsrc2				= ttmp10
+var s_restore_buf_rsrc3				= ttmp11
+var s_restore_size				= ttmp6
+var s_restore_ttmps_lo				= s_restore_tmp
+var s_restore_ttmps_hi				= s_restore_alloc_size
+var s_restore_spi_init_hi_save			= s_restore_exec_hi
+
+shader main
+	asic(DEFAULT)
+	type(CS)
+	wave_size(32)
+
+	s_branch	L_SKIP_RESTORE						//NOT restore. might be a regular trap or save
+
+L_JUMP_TO_RESTORE:
+	s_branch	L_RESTORE
+
+L_SKIP_RESTORE:
+	s_getreg_b32	s_save_state_priv, hwreg(HW_REG_WAVE_STATE_PRIV)	//save STATUS since we will change SCC
+
+	// Clear SPI_PRIO: do not save with elevated priority.
+	// Clear ECC_ERR: prevents SQC store and triggers FATAL_HALT if setreg'd.
+	s_andn2_b32	s_save_state_priv, s_save_state_priv, SQ_WAVE_STATE_PRIV_ALWAYS_CLEAR_MASK
+
+	s_getreg_b32	s_save_excp_flag_priv, hwreg(HW_REG_WAVE_EXCP_FLAG_PRIV)
+
+	s_and_b32       ttmp2, s_save_state_priv, SQ_WAVE_STATE_PRIV_HALT_MASK
+	s_cbranch_scc0	L_NOT_HALTED
+
+L_HALTED:
+	// Host trap may occur while wave is halted.
+	s_and_b32	ttmp2, s_save_excp_flag_priv, SQ_WAVE_EXCP_FLAG_PRIV_HOST_TRAP_MASK
+	s_cbranch_scc1	L_FETCH_2ND_TRAP
+
+L_CHECK_SAVE:
+	s_and_b32	ttmp2, s_save_excp_flag_priv, SQ_WAVE_EXCP_FLAG_PRIV_SAVE_CONTEXT_MASK
+	s_cbranch_scc1	L_SAVE
+
+	// Wave is halted but neither host trap nor SAVECTX is raised.
+	// Caused by instruction fetch memory violation.
+	// Spin wait until context saved to prevent interrupt storm.
+	s_sleep		0x10
+	s_getreg_b32	s_save_excp_flag_priv, hwreg(HW_REG_WAVE_EXCP_FLAG_PRIV)
+	s_branch	L_CHECK_SAVE
+
+L_NOT_HALTED:
+	// Let second-level handle non-SAVECTX exception or trap.
+	// Any concurrent SAVECTX will be handled upon re-entry once halted.
+
+	// Check non-maskable exceptions. memory_violation, illegal_instruction
+	// and xnack_error exceptions always cause the wave to enter the trap
+	// handler.
+	s_and_b32	ttmp2, s_save_excp_flag_priv, SQ_WAVE_EXCP_FLAG_PRIV_NON_MASKABLE_EXCP_MASK
+	s_cbranch_scc1	L_FETCH_2ND_TRAP
+
+	// Check for maskable exceptions in trapsts.excp and trapsts.excp_hi.
+	// Maskable exceptions only cause the wave to enter the trap handler if
+	// their respective bit in mode.excp_en is set.
+	s_getreg_b32	ttmp2, hwreg(HW_REG_WAVE_EXCP_FLAG_USER)
+	s_and_b32	ttmp3, s_save_excp_flag_priv, SQ_WAVE_EXCP_FLAG_PRIV_ADDR_WATCH_MASK
+	s_cbranch_scc0	L_NOT_ADDR_WATCH
+	s_or_b32	ttmp2, ttmp2, SQ_WAVE_TRAP_CTRL_ADDR_WATCH_MASK
+
+L_NOT_ADDR_WATCH:
+	s_getreg_b32	ttmp3, hwreg(HW_REG_WAVE_TRAP_CTRL)
+	s_and_b32	ttmp2, ttmp3, ttmp2
+	s_cbranch_scc1	L_FETCH_2ND_TRAP
+
+L_CHECK_TRAP_ID:
+	// Check trap_id != 0
+	s_and_b32	ttmp2, s_save_pc_hi, S_SAVE_PC_HI_TRAP_ID_MASK
+	s_cbranch_scc1	L_FETCH_2ND_TRAP
+
+#if SINGLE_STEP_MISSED_WORKAROUND
+	// Prioritize single step exception over context save.
+	// Second-level trap will halt wave and RFE, re-entering for SAVECTX.
+	// WAVE_TRAP_CTRL is already in ttmp3.
+	s_and_b32	ttmp3, ttmp3, SQ_WAVE_TRAP_CTRL_TRAP_AFTER_INST_MASK
+	s_cbranch_scc1	L_FETCH_2ND_TRAP
+#endif
+
+	s_and_b32	ttmp2, s_save_excp_flag_priv, SQ_WAVE_EXCP_FLAG_PRIV_SAVE_CONTEXT_MASK
+	s_cbranch_scc1	L_SAVE
+
+L_FETCH_2ND_TRAP:
+	// Read second-level TBA/TMA from first-level TMA and jump if available.
+	// ttmp[2:5] and ttmp12 can be used (others hold SPI-initialized debug data)
+	// ttmp12 holds SQ_WAVE_STATUS
+	s_sendmsg_rtn_b64       [ttmp14, ttmp15], sendmsg(MSG_RTN_GET_TMA)
+	s_wait_idle
+	s_lshl_b64	[ttmp14, ttmp15], [ttmp14, ttmp15], 0x8
+
+	s_bitcmp1_b32	ttmp15, 0xF
+	s_cbranch_scc0	L_NO_SIGN_EXTEND_TMA
+	s_or_b32	ttmp15, ttmp15, 0xFFFF0000
+L_NO_SIGN_EXTEND_TMA:
+
+	s_load_dword    ttmp2, [ttmp14, ttmp15], 0x10 scope:SCOPE_SYS		// debug trap enabled flag
+	s_wait_idle
+	s_lshl_b32      ttmp2, ttmp2, TTMP11_DEBUG_TRAP_ENABLED_SHIFT
+	s_andn2_b32     ttmp11, ttmp11, TTMP11_DEBUG_TRAP_ENABLED_MASK
+	s_or_b32        ttmp11, ttmp11, ttmp2
+
+	s_load_dwordx2	[ttmp2, ttmp3], [ttmp14, ttmp15], 0x0 scope:SCOPE_SYS	// second-level TBA
+	s_wait_idle
+	s_load_dwordx2	[ttmp14, ttmp15], [ttmp14, ttmp15], 0x8 scope:SCOPE_SYS	// second-level TMA
+	s_wait_idle
+
+	s_and_b64	[ttmp2, ttmp3], [ttmp2, ttmp3], [ttmp2, ttmp3]
+	s_cbranch_scc0	L_NO_NEXT_TRAP						// second-level trap handler not been set
+	s_setpc_b64	[ttmp2, ttmp3]						// jump to second-level trap handler
+
+L_NO_NEXT_TRAP:
+	// If not caused by trap then halt wave to prevent re-entry.
+	s_and_b32	ttmp2, s_save_pc_hi, S_SAVE_PC_HI_TRAP_ID_MASK
+	s_cbranch_scc1	L_TRAP_CASE
+
+	// Host trap will not cause trap re-entry.
+	s_getreg_b32	ttmp2, hwreg(HW_REG_WAVE_EXCP_FLAG_PRIV)
+	s_and_b32	ttmp2, ttmp2, SQ_WAVE_EXCP_FLAG_PRIV_HOST_TRAP_MASK
+	s_cbranch_scc1	L_EXIT_TRAP
+	s_or_b32	s_save_state_priv, s_save_state_priv, SQ_WAVE_STATE_PRIV_HALT_MASK
+
+	// If the PC points to S_ENDPGM then context save will fail if STATE_PRIV.HALT is set.
+	// Rewind the PC to prevent this from occurring.
+	s_sub_u32	ttmp0, ttmp0, 0x8
+	s_subb_u32	ttmp1, ttmp1, 0x0
+
+	s_branch	L_EXIT_TRAP
+
+L_TRAP_CASE:
+	// Advance past trap instruction to prevent re-entry.
+	s_add_u32	ttmp0, ttmp0, 0x4
+	s_addc_u32	ttmp1, ttmp1, 0x0
+
+L_EXIT_TRAP:
+	s_and_b32	ttmp1, ttmp1, 0xFFFF
+
+	// Restore SQ_WAVE_STATUS.
+	s_and_b64	exec, exec, exec					// Restore STATUS.EXECZ, not writable by s_setreg_b32
+	s_and_b64	vcc, vcc, vcc						// Restore STATUS.VCCZ, not writable by s_setreg_b32
+
+	// STATE_PRIV.BARRIER_COMPLETE may have changed since we read it.
+	// Only restore fields which the trap handler changes.
+	s_lshr_b32	s_save_state_priv, s_save_state_priv, SQ_WAVE_STATE_PRIV_SCC_SHIFT
+	s_setreg_b32	hwreg(HW_REG_WAVE_STATE_PRIV, SQ_WAVE_STATE_PRIV_SCC_SHIFT, \
+		SQ_WAVE_STATE_PRIV_POISON_ERR_SHIFT - SQ_WAVE_STATE_PRIV_SCC_SHIFT + 1), s_save_state_priv
+
+	s_rfe_b64	[ttmp0, ttmp1]
+
+L_SAVE:
+	// If VGPRs have been deallocated then terminate the wavefront.
+	// It has no remaining program to run and cannot save without VGPRs.
+	s_getreg_b32	s_save_tmp, hwreg(HW_REG_WAVE_STATUS)
+	s_bitcmp1_b32	s_save_tmp, SQ_WAVE_STATUS_NO_VGPRS_SHIFT
+	s_cbranch_scc0	L_HAVE_VGPRS
+	s_endpgm
+L_HAVE_VGPRS:
+
+	s_and_b32	s_save_pc_hi, s_save_pc_hi, 0x0000ffff			//pc[47:32]
+	s_mov_b32	s_save_tmp, 0
+	s_setreg_b32	hwreg(HW_REG_WAVE_EXCP_FLAG_PRIV, SQ_WAVE_EXCP_FLAG_PRIV_SAVE_CONTEXT_SHIFT, 1), s_save_tmp	//clear saveCtx bit
+
+	/* inform SPI the readiness and wait for SPI's go signal */
+	s_mov_b32	s_save_exec_lo, exec_lo					//save EXEC and use EXEC for the go signal from SPI
+	s_mov_b32	s_save_exec_hi, exec_hi
+	s_mov_b64	exec, 0x0						//clear EXEC to get ready to receive
+
+	s_sendmsg_rtn_b64       [exec_lo, exec_hi], sendmsg(MSG_RTN_SAVE_WAVE)
+	s_wait_idle
+
+	// Save first_wave flag so we can clear high bits of save address.
+	s_and_b32	s_save_tmp, s_save_spi_init_hi, S_SAVE_SPI_INIT_FIRST_WAVE_MASK
+	s_lshl_b32	s_save_tmp, s_save_tmp, (S_SAVE_PC_HI_FIRST_WAVE_SHIFT - S_SAVE_SPI_INIT_FIRST_WAVE_SHIFT)
+	s_or_b32	s_save_pc_hi, s_save_pc_hi, s_save_tmp
+
+	// Trap temporaries must be saved via VGPR but all VGPRs are in use.
+	// There is no ttmp space to hold the resource constant for VGPR save.
+	// Save v0 by itself since it requires only two SGPRs.
+	s_mov_b32	s_save_ttmps_lo, exec_lo
+	s_and_b32	s_save_ttmps_hi, exec_hi, 0xFFFF
+	s_mov_b32	exec_lo, 0xFFFFFFFF
+	s_mov_b32	exec_hi, 0xFFFFFFFF
+	global_store_dword_addtid	v0, [s_save_ttmps_lo, s_save_ttmps_hi] scope:SCOPE_SYS
+	v_mov_b32	v0, 0x0
+	s_mov_b32	exec_lo, s_save_ttmps_lo
+	s_mov_b32	exec_hi, s_save_ttmps_hi
+
+	// Save trap temporaries 4-11, 13 initialized by SPI debug dispatch logic
+	// ttmp SR memory offset : size(VGPR)+size(SVGPR)+size(SGPR)+0x40
+	get_wave_size2(s_save_ttmps_hi)
+	get_vgpr_size_bytes(s_save_ttmps_lo, s_save_ttmps_hi)
+	get_svgpr_size_bytes(s_save_ttmps_hi)
+	s_add_u32	s_save_ttmps_lo, s_save_ttmps_lo, s_save_ttmps_hi
+	s_and_b32	s_save_ttmps_hi, s_save_spi_init_hi, 0xFFFF
+	s_add_u32	s_save_ttmps_lo, s_save_ttmps_lo, get_sgpr_size_bytes()
+	s_add_u32	s_save_ttmps_lo, s_save_ttmps_lo, s_save_spi_init_lo
+	s_addc_u32	s_save_ttmps_hi, s_save_ttmps_hi, 0x0
+
+	v_writelane_b32	v0, ttmp4, 0x4
+	v_writelane_b32	v0, ttmp5, 0x5
+	v_writelane_b32	v0, ttmp6, 0x6
+	v_writelane_b32	v0, ttmp7, 0x7
+	v_writelane_b32	v0, ttmp8, 0x8
+	v_writelane_b32	v0, ttmp9, 0x9
+	v_writelane_b32	v0, ttmp10, 0xA
+	v_writelane_b32	v0, ttmp11, 0xB
+	v_writelane_b32	v0, ttmp13, 0xD
+	v_writelane_b32	v0, exec_lo, 0xE
+	v_writelane_b32	v0, exec_hi, 0xF
+	valu_sgpr_hazard()
+
+	s_mov_b32	exec_lo, 0x3FFF
+	s_mov_b32	exec_hi, 0x0
+	global_store_dword_addtid	v0, [s_save_ttmps_lo, s_save_ttmps_hi] offset:0x40 scope:SCOPE_SYS
+	v_readlane_b32	ttmp14, v0, 0xE
+	v_readlane_b32	ttmp15, v0, 0xF
+	s_mov_b32	exec_lo, ttmp14
+	s_mov_b32	exec_hi, ttmp15
+
+	/* setup Resource Contants */
+	s_mov_b32	s_save_buf_rsrc0, s_save_spi_init_lo			//base_addr_lo
+	s_and_b32	s_save_buf_rsrc1, s_save_spi_init_hi, 0x0000FFFF	//base_addr_hi
+	s_or_b32	s_save_buf_rsrc1, s_save_buf_rsrc1, S_SAVE_BUF_RSRC_WORD1_STRIDE
+	s_mov_b32	s_save_buf_rsrc2, 0					//NUM_RECORDS initial value = 0 (in bytes) although not neccessarily inited
+	s_mov_b32	s_save_buf_rsrc3, S_SAVE_BUF_RSRC_WORD3_MISC
+
+	s_mov_b32	s_save_m0, m0
+
+	/* global mem offset */
+	s_mov_b32	s_save_mem_offset, 0x0
+	get_wave_size2(s_wave_size)
+
+	/* save first 4 VGPRs, needed for SGPR save */
+	s_mov_b32	exec_lo, 0xFFFFFFFF					//need every thread from now on
+	s_lshr_b32	m0, s_wave_size, S_WAVE_SIZE
+	s_and_b32	m0, m0, 1
+	s_cmp_eq_u32	m0, 1
+	s_cbranch_scc1	L_ENABLE_SAVE_4VGPR_EXEC_HI
+	s_mov_b32	exec_hi, 0x00000000
+	s_branch	L_SAVE_4VGPR_WAVE32
+L_ENABLE_SAVE_4VGPR_EXEC_HI:
+	s_mov_b32	exec_hi, 0xFFFFFFFF
+	s_branch	L_SAVE_4VGPR_WAVE64
+L_SAVE_4VGPR_WAVE32:
+	s_mov_b32	s_save_buf_rsrc2, 0x1000000				//NUM_RECORDS in bytes
+
+	// VGPR Allocated in 4-GPR granularity
+
+	buffer_store_dword	v1, v0, s_save_buf_rsrc0, s_save_mem_offset scope:SCOPE_SYS offset:128
+	buffer_store_dword	v2, v0, s_save_buf_rsrc0, s_save_mem_offset scope:SCOPE_SYS offset:128*2
+	buffer_store_dword	v3, v0, s_save_buf_rsrc0, s_save_mem_offset scope:SCOPE_SYS offset:128*3
+	s_branch	L_SAVE_HWREG
+
+L_SAVE_4VGPR_WAVE64:
+	s_mov_b32	s_save_buf_rsrc2, 0x1000000				//NUM_RECORDS in bytes
+
+	// VGPR Allocated in 4-GPR granularity
+
+	buffer_store_dword	v1, v0, s_save_buf_rsrc0, s_save_mem_offset scope:SCOPE_SYS offset:256
+	buffer_store_dword	v2, v0, s_save_buf_rsrc0, s_save_mem_offset scope:SCOPE_SYS offset:256*2
+	buffer_store_dword	v3, v0, s_save_buf_rsrc0, s_save_mem_offset scope:SCOPE_SYS offset:256*3
+
+	/* save HW registers */
+
+L_SAVE_HWREG:
+	// HWREG SR memory offset : size(VGPR)+size(SVGPR)+size(SGPR)
+	get_vgpr_size_bytes(s_save_mem_offset, s_wave_size)
+	get_svgpr_size_bytes(s_save_tmp)
+	s_add_u32	s_save_mem_offset, s_save_mem_offset, s_save_tmp
+	s_add_u32	s_save_mem_offset, s_save_mem_offset, get_sgpr_size_bytes()
+
+	s_mov_b32	s_save_buf_rsrc2, 0x1000000				//NUM_RECORDS in bytes
+
+	v_mov_b32	v0, 0x0							//Offset[31:0] from buffer resource
+	v_mov_b32	v1, 0x0							//Offset[63:32] from buffer resource
+	v_mov_b32	v2, 0x0							//Set of SGPRs for TCP store
+
+	// Ensure no further changes to barrier or LDS state.
+	// STATE_PRIV.BARRIER_COMPLETE may change up to this point.
+	s_barrier_signal	-2
+	s_barrier_wait	-2
+
+	// Re-read final state of BARRIER_COMPLETE field for save.
+	s_getreg_b32	s_save_tmp, hwreg(HW_REG_WAVE_STATE_PRIV)
+	s_and_b32	s_save_tmp, s_save_tmp, SQ_WAVE_STATE_PRIV_BARRIER_COMPLETE_MASK
+	s_andn2_b32	s_save_state_priv, s_save_state_priv, SQ_WAVE_STATE_PRIV_BARRIER_COMPLETE_MASK
+	s_or_b32	s_save_state_priv, s_save_state_priv, s_save_tmp
+
+	s_andn2_b32	s_save_tmp, s_save_pc_hi, S_SAVE_PC_HI_FIRST_WAVE_MASK
+	v_writelane_b32	v2, s_save_m0, 0x0
+	v_writelane_b32	v2, s_save_pc_lo, 0x1
+	v_writelane_b32	v2, s_save_tmp, 0x2
+	v_writelane_b32	v2, s_save_exec_lo, 0x3
+	v_writelane_b32	v2, s_save_exec_hi, 0x4
+	v_writelane_b32	v2, s_save_state_priv, 0x5
+	v_writelane_b32	v2, s_save_xnack_mask, 0x7
+	valu_sgpr_hazard()
+
+	s_getreg_b32	s_save_tmp, hwreg(HW_REG_WAVE_EXCP_FLAG_PRIV)
+	v_writelane_b32	v2, s_save_tmp, 0x6
+
+	s_getreg_b32	s_save_tmp, hwreg(HW_REG_WAVE_MODE)
+	v_writelane_b32	v2, s_save_tmp, 0x8
+
+	s_getreg_b32	s_save_tmp, hwreg(HW_REG_WAVE_SCRATCH_BASE_LO)
+	v_writelane_b32	v2, s_save_tmp, 0x9
+
+	s_getreg_b32	s_save_tmp, hwreg(HW_REG_WAVE_SCRATCH_BASE_HI)
+	v_writelane_b32	v2, s_save_tmp, 0xA
+
+	s_getreg_b32	s_save_tmp, hwreg(HW_REG_WAVE_EXCP_FLAG_USER)
+	v_writelane_b32	v2, s_save_tmp, 0xB
+
+	s_getreg_b32	s_save_tmp, hwreg(HW_REG_WAVE_TRAP_CTRL)
+	v_writelane_b32	v2, s_save_tmp, 0xC
+
+	s_getreg_b32	s_save_tmp, hwreg(HW_REG_WAVE_STATUS)
+	v_writelane_b32	v2, s_save_tmp, 0xD
+
+	s_get_barrier_state s_save_tmp, -1
+	s_wait_kmcnt (0)
+	v_writelane_b32	v2, s_save_tmp, 0xE
+	valu_sgpr_hazard()
+
+	// Write HWREGs with 16 VGPR lanes. TTMPs occupy space after this.
+	s_mov_b32       exec_lo, 0xFFFF
+	s_mov_b32	exec_hi, 0x0
+	buffer_store_dword	v2, v0, s_save_buf_rsrc0, s_save_mem_offset scope:SCOPE_SYS
+
+	// Write SGPRs with 32 VGPR lanes. This works in wave32 and wave64 mode.
+	s_mov_b32       exec_lo, 0xFFFFFFFF
+
+	/* save SGPRs */
+	// Save SGPR before LDS save, then the s0 to s4 can be used during LDS save...
+
+	// SGPR SR memory offset : size(VGPR)+size(SVGPR)
+	get_vgpr_size_bytes(s_save_mem_offset, s_wave_size)
+	get_svgpr_size_bytes(s_save_tmp)
+	s_add_u32	s_save_mem_offset, s_save_mem_offset, s_save_tmp
+	s_mov_b32	s_save_buf_rsrc2, 0x1000000				//NUM_RECORDS in bytes
+
+	s_mov_b32	ttmp13, 0x0						//next VGPR lane to copy SGPR into
+
+	s_mov_b32	m0, 0x0							//SGPR initial index value =0
+	s_nop		0x0							//Manually inserted wait states
+L_SAVE_SGPR_LOOP:
+	// SGPR is allocated in 16 SGPR granularity
+	s_movrels_b64	s0, s0							//s0 = s[0+m0], s1 = s[1+m0]
+	s_movrels_b64	s2, s2							//s2 = s[2+m0], s3 = s[3+m0]
+	s_movrels_b64	s4, s4							//s4 = s[4+m0], s5 = s[5+m0]
+	s_movrels_b64	s6, s6							//s6 = s[6+m0], s7 = s[7+m0]
+	s_movrels_b64	s8, s8							//s8 = s[8+m0], s9 = s[9+m0]
+	s_movrels_b64	s10, s10						//s10 = s[10+m0], s11 = s[11+m0]
+	s_movrels_b64	s12, s12						//s12 = s[12+m0], s13 = s[13+m0]
+	s_movrels_b64	s14, s14						//s14 = s[14+m0], s15 = s[15+m0]
+
+	s_cmp_eq_u32	ttmp13, 0x0
+	s_cbranch_scc0	L_WRITE_V2_SECOND_HALF
+	write_16sgpr_to_v2(s0, 0x0)
+	s_branch	L_SAVE_SGPR_SKIP_TCP_STORE
+L_WRITE_V2_SECOND_HALF:
+	write_16sgpr_to_v2(s0, 0x10)
+
+	buffer_store_dword	v2, v0, s_save_buf_rsrc0, s_save_mem_offset scope:SCOPE_SYS
+	s_add_u32	s_save_mem_offset, s_save_mem_offset, 0x80
+	s_mov_b32	ttmp13, 0x0
+	v_mov_b32	v2, 0x0
+L_SAVE_SGPR_SKIP_TCP_STORE:
+
+	s_add_u32	m0, m0, 16						//next sgpr index
+	s_cmp_lt_u32	m0, 96							//scc = (m0 < first 96 SGPR) ? 1 : 0
+	s_cbranch_scc1	L_SAVE_SGPR_LOOP					//first 96 SGPR save is complete?
+
+	//save the rest 12 SGPR
+	s_movrels_b64	s0, s0							//s0 = s[0+m0], s1 = s[1+m0]
+	s_movrels_b64	s2, s2							//s2 = s[2+m0], s3 = s[3+m0]
+	s_movrels_b64	s4, s4							//s4 = s[4+m0], s5 = s[5+m0]
+	s_movrels_b64	s6, s6							//s6 = s[6+m0], s7 = s[7+m0]
+	s_movrels_b64	s8, s8							//s8 = s[8+m0], s9 = s[9+m0]
+	s_movrels_b64	s10, s10						//s10 = s[10+m0], s11 = s[11+m0]
+	write_12sgpr_to_v2(s0)
+
+	buffer_store_dword	v2, v0, s_save_buf_rsrc0, s_save_mem_offset scope:SCOPE_SYS
+
+	/* save LDS */
+
+L_SAVE_LDS:
+	// Change EXEC to all threads...
+	s_mov_b32	exec_lo, 0xFFFFFFFF					//need every thread from now on
+	s_lshr_b32	m0, s_wave_size, S_WAVE_SIZE
+	s_and_b32	m0, m0, 1
+	s_cmp_eq_u32	m0, 1
+	s_cbranch_scc1	L_ENABLE_SAVE_LDS_EXEC_HI
+	s_mov_b32	exec_hi, 0x00000000
+	s_branch	L_SAVE_LDS_NORMAL
+L_ENABLE_SAVE_LDS_EXEC_HI:
+	s_mov_b32	exec_hi, 0xFFFFFFFF
+L_SAVE_LDS_NORMAL:
+	s_getreg_b32	s_save_alloc_size, hwreg(HW_REG_WAVE_LDS_ALLOC,SQ_WAVE_LDS_ALLOC_LDS_SIZE_SHIFT,SQ_WAVE_LDS_ALLOC_LDS_SIZE_SIZE)
+	s_and_b32	s_save_alloc_size, s_save_alloc_size, 0xFFFFFFFF	//lds_size is zero?
+	s_cbranch_scc0	L_SAVE_LDS_DONE						//no lds used? jump to L_SAVE_DONE
+
+	s_and_b32	s_save_tmp, s_save_pc_hi, S_SAVE_PC_HI_FIRST_WAVE_MASK
+	s_cbranch_scc0	L_SAVE_LDS_DONE
+
+	// first wave do LDS save;
+
+	s_lshl_b32	s_save_alloc_size, s_save_alloc_size, SQ_WAVE_LDS_ALLOC_GRANULARITY
+	s_mov_b32	s_save_buf_rsrc2, s_save_alloc_size			//NUM_RECORDS in bytes
+
+	// LDS at offset: size(VGPR)+size(SVGPR)+SIZE(SGPR)+SIZE(HWREG)
+	//
+	get_vgpr_size_bytes(s_save_mem_offset, s_wave_size)
+	get_svgpr_size_bytes(s_save_tmp)
+	s_add_u32	s_save_mem_offset, s_save_mem_offset, s_save_tmp
+	s_add_u32	s_save_mem_offset, s_save_mem_offset, get_sgpr_size_bytes()
+	s_add_u32	s_save_mem_offset, s_save_mem_offset, get_hwreg_size_bytes()
+
+	s_mov_b32	s_save_buf_rsrc2, 0x1000000				//NUM_RECORDS in bytes
+
+	//load 0~63*4(byte address) to vgpr v0
+	v_mbcnt_lo_u32_b32	v0, -1, 0
+	v_mbcnt_hi_u32_b32	v0, -1, v0
+	v_mul_u32_u24	v0, 4, v0
+
+	s_lshr_b32	m0, s_wave_size, S_WAVE_SIZE
+	s_and_b32	m0, m0, 1
+	s_cmp_eq_u32	m0, 1
+	s_mov_b32	m0, 0x0
+	s_cbranch_scc1	L_SAVE_LDS_W64
+
+L_SAVE_LDS_W32:
+	s_mov_b32	s3, 128
+	s_nop		0
+	s_nop		0
+	s_nop		0
+L_SAVE_LDS_LOOP_W32:
+	ds_read_b32	v1, v0
+	s_wait_idle
+	buffer_store_dword	v1, v0, s_save_buf_rsrc0, s_save_mem_offset scope:SCOPE_SYS
+
+	s_add_u32	m0, m0, s3						//every buffer_store_lds does 128 bytes
+	s_add_u32	s_save_mem_offset, s_save_mem_offset, s3
+	v_add_nc_u32	v0, v0, 128						//mem offset increased by 128 bytes
+	s_cmp_lt_u32	m0, s_save_alloc_size					//scc=(m0 < s_save_alloc_size) ? 1 : 0
+	s_cbranch_scc1	L_SAVE_LDS_LOOP_W32					//LDS save is complete?
+
+	s_branch	L_SAVE_LDS_DONE
+
+L_SAVE_LDS_W64:
+	s_mov_b32	s3, 256
+	s_nop		0
+	s_nop		0
+	s_nop		0
+L_SAVE_LDS_LOOP_W64:
+	ds_read_b32	v1, v0
+	s_wait_idle
+	buffer_store_dword	v1, v0, s_save_buf_rsrc0, s_save_mem_offset scope:SCOPE_SYS
+
+	s_add_u32	m0, m0, s3						//every buffer_store_lds does 256 bytes
+	s_add_u32	s_save_mem_offset, s_save_mem_offset, s3
+	v_add_nc_u32	v0, v0, 256						//mem offset increased by 256 bytes
+	s_cmp_lt_u32	m0, s_save_alloc_size					//scc=(m0 < s_save_alloc_size) ? 1 : 0
+	s_cbranch_scc1	L_SAVE_LDS_LOOP_W64					//LDS save is complete?
+
+L_SAVE_LDS_DONE:
+	/* save VGPRs  - set the Rest VGPRs */
+L_SAVE_VGPR:
+	// VGPR SR memory offset: 0
+	s_mov_b32	exec_lo, 0xFFFFFFFF					//need every thread from now on
+	s_lshr_b32	m0, s_wave_size, S_WAVE_SIZE
+	s_and_b32	m0, m0, 1
+	s_cmp_eq_u32	m0, 1
+	s_cbranch_scc1	L_ENABLE_SAVE_VGPR_EXEC_HI
+	s_mov_b32	s_save_mem_offset, (0+128*4)				// for the rest VGPRs
+	s_mov_b32	exec_hi, 0x00000000
+	s_branch	L_SAVE_VGPR_NORMAL
+L_ENABLE_SAVE_VGPR_EXEC_HI:
+	s_mov_b32	s_save_mem_offset, (0+256*4)				// for the rest VGPRs
+	s_mov_b32	exec_hi, 0xFFFFFFFF
+L_SAVE_VGPR_NORMAL:
+	s_getreg_b32	s_save_alloc_size, hwreg(HW_REG_WAVE_GPR_ALLOC,SQ_WAVE_GPR_ALLOC_VGPR_SIZE_SHIFT,SQ_WAVE_GPR_ALLOC_VGPR_SIZE_SIZE)
+	s_add_u32	s_save_alloc_size, s_save_alloc_size, 1
+	s_lshl_b32	s_save_alloc_size, s_save_alloc_size, 2			//Number of VGPRs = (vgpr_size + 1) * 4    (non-zero value)
+	//determine it is wave32 or wave64
+	s_lshr_b32	m0, s_wave_size, S_WAVE_SIZE
+	s_and_b32	m0, m0, 1
+	s_cmp_eq_u32	m0, 1
+	s_cbranch_scc1	L_SAVE_VGPR_WAVE64
+
+	s_mov_b32	s_save_buf_rsrc2, 0x1000000				//NUM_RECORDS in bytes
+
+	// VGPR Allocated in 4-GPR granularity
+
+	// VGPR store using dw burst
+	s_mov_b32	m0, 0x4							//VGPR initial index value =4
+	s_cmp_lt_u32	m0, s_save_alloc_size
+	s_cbranch_scc0	L_SAVE_VGPR_END
+
+L_SAVE_VGPR_W32_LOOP:
+	v_movrels_b32	v0, v0							//v0 = v[0+m0]
+	v_movrels_b32	v1, v1							//v1 = v[1+m0]
+	v_movrels_b32	v2, v2							//v2 = v[2+m0]
+	v_movrels_b32	v3, v3							//v3 = v[3+m0]
+
+	buffer_store_dword	v0, v0, s_save_buf_rsrc0, s_save_mem_offset scope:SCOPE_SYS
+	buffer_store_dword	v1, v0, s_save_buf_rsrc0, s_save_mem_offset scope:SCOPE_SYS offset:128
+	buffer_store_dword	v2, v0, s_save_buf_rsrc0, s_save_mem_offset scope:SCOPE_SYS offset:128*2
+	buffer_store_dword	v3, v0, s_save_buf_rsrc0, s_save_mem_offset scope:SCOPE_SYS offset:128*3
+
+	s_add_u32	m0, m0, 4						//next vgpr index
+	s_add_u32	s_save_mem_offset, s_save_mem_offset, 128*4		//every buffer_store_dword does 128 bytes
+	s_cmp_lt_u32	m0, s_save_alloc_size					//scc = (m0 < s_save_alloc_size) ? 1 : 0
+	s_cbranch_scc1	L_SAVE_VGPR_W32_LOOP					//VGPR save is complete?
+
+	s_branch	L_SAVE_VGPR_END
+
+L_SAVE_VGPR_WAVE64:
+	s_mov_b32	s_save_buf_rsrc2, 0x1000000				//NUM_RECORDS in bytes
+
+	// VGPR store using dw burst
+	s_mov_b32	m0, 0x4							//VGPR initial index value =4
+	s_cmp_lt_u32	m0, s_save_alloc_size
+	s_cbranch_scc0	L_SAVE_SHARED_VGPR
+
+L_SAVE_VGPR_W64_LOOP:
+	v_movrels_b32	v0, v0							//v0 = v[0+m0]
+	v_movrels_b32	v1, v1							//v1 = v[1+m0]
+	v_movrels_b32	v2, v2							//v2 = v[2+m0]
+	v_movrels_b32	v3, v3							//v3 = v[3+m0]
+
+	buffer_store_dword	v0, v0, s_save_buf_rsrc0, s_save_mem_offset scope:SCOPE_SYS
+	buffer_store_dword	v1, v0, s_save_buf_rsrc0, s_save_mem_offset scope:SCOPE_SYS offset:256
+	buffer_store_dword	v2, v0, s_save_buf_rsrc0, s_save_mem_offset scope:SCOPE_SYS offset:256*2
+	buffer_store_dword	v3, v0, s_save_buf_rsrc0, s_save_mem_offset scope:SCOPE_SYS offset:256*3
+
+	s_add_u32	m0, m0, 4						//next vgpr index
+	s_add_u32	s_save_mem_offset, s_save_mem_offset, 256*4		//every buffer_store_dword does 256 bytes
+	s_cmp_lt_u32	m0, s_save_alloc_size					//scc = (m0 < s_save_alloc_size) ? 1 : 0
+	s_cbranch_scc1	L_SAVE_VGPR_W64_LOOP					//VGPR save is complete?
+
+L_SAVE_SHARED_VGPR:
+	s_getreg_b32	s_save_alloc_size, hwreg(HW_REG_WAVE_LDS_ALLOC,SQ_WAVE_LDS_ALLOC_VGPR_SHARED_SIZE_SHIFT,SQ_WAVE_LDS_ALLOC_VGPR_SHARED_SIZE_SIZE)
+	s_and_b32	s_save_alloc_size, s_save_alloc_size, 0xFFFFFFFF	//shared_vgpr_size is zero?
+	s_cbranch_scc0	L_SAVE_VGPR_END						//no shared_vgpr used? jump to L_SAVE_LDS
+	s_lshl_b32	s_save_alloc_size, s_save_alloc_size, 3			//Number of SHARED_VGPRs = shared_vgpr_size * 8    (non-zero value)
+	//m0 now has the value of normal vgpr count, just add the m0 with shared_vgpr count to get the total count.
+	//save shared_vgpr will start from the index of m0
+	s_add_u32	s_save_alloc_size, s_save_alloc_size, m0
+	s_mov_b32	exec_lo, 0xFFFFFFFF
+	s_mov_b32	exec_hi, 0x00000000
+
+L_SAVE_SHARED_VGPR_WAVE64_LOOP:
+	v_movrels_b32	v0, v0							//v0 = v[0+m0]
+	buffer_store_dword	v0, v0, s_save_buf_rsrc0, s_save_mem_offset scope:SCOPE_SYS
+	s_add_u32	m0, m0, 1						//next vgpr index
+	s_add_u32	s_save_mem_offset, s_save_mem_offset, 128
+	s_cmp_lt_u32	m0, s_save_alloc_size					//scc = (m0 < s_save_alloc_size) ? 1 : 0
+	s_cbranch_scc1	L_SAVE_SHARED_VGPR_WAVE64_LOOP				//SHARED_VGPR save is complete?
+
+L_SAVE_VGPR_END:
+	s_branch	L_END_PGM
+
+L_RESTORE:
+	/* Setup Resource Contants */
+	s_mov_b32	s_restore_buf_rsrc0, s_restore_spi_init_lo		//base_addr_lo
+	s_and_b32	s_restore_buf_rsrc1, s_restore_spi_init_hi, 0x0000FFFF	//base_addr_hi
+	s_or_b32	s_restore_buf_rsrc1, s_restore_buf_rsrc1, S_RESTORE_BUF_RSRC_WORD1_STRIDE
+	s_mov_b32	s_restore_buf_rsrc2, 0					//NUM_RECORDS initial value = 0 (in bytes)
+	s_mov_b32	s_restore_buf_rsrc3, S_RESTORE_BUF_RSRC_WORD3_MISC
+
+	// Save s_restore_spi_init_hi for later use.
+	s_mov_b32 s_restore_spi_init_hi_save, s_restore_spi_init_hi
+
+	//determine it is wave32 or wave64
+	get_wave_size2(s_restore_size)
+
+	s_and_b32	s_restore_tmp, s_restore_spi_init_hi, S_RESTORE_SPI_INIT_FIRST_WAVE_MASK
+	s_cbranch_scc0	L_RESTORE_VGPR
+
+	/* restore LDS */
+L_RESTORE_LDS:
+	s_mov_b32	exec_lo, 0xFFFFFFFF					//need every thread from now on
+	s_lshr_b32	m0, s_restore_size, S_WAVE_SIZE
+	s_and_b32	m0, m0, 1
+	s_cmp_eq_u32	m0, 1
+	s_cbranch_scc1	L_ENABLE_RESTORE_LDS_EXEC_HI
+	s_mov_b32	exec_hi, 0x00000000
+	s_branch	L_RESTORE_LDS_NORMAL
+L_ENABLE_RESTORE_LDS_EXEC_HI:
+	s_mov_b32	exec_hi, 0xFFFFFFFF
+L_RESTORE_LDS_NORMAL:
+	s_getreg_b32	s_restore_alloc_size, hwreg(HW_REG_WAVE_LDS_ALLOC,SQ_WAVE_LDS_ALLOC_LDS_SIZE_SHIFT,SQ_WAVE_LDS_ALLOC_LDS_SIZE_SIZE)
+	s_and_b32	s_restore_alloc_size, s_restore_alloc_size, 0xFFFFFFFF	//lds_size is zero?
+	s_cbranch_scc0	L_RESTORE_VGPR						//no lds used? jump to L_RESTORE_VGPR
+	s_lshl_b32	s_restore_alloc_size, s_restore_alloc_size, SQ_WAVE_LDS_ALLOC_GRANULARITY
+	s_mov_b32	s_restore_buf_rsrc2, s_restore_alloc_size		//NUM_RECORDS in bytes
+
+	// LDS at offset: size(VGPR)+size(SVGPR)+SIZE(SGPR)+SIZE(HWREG)
+	//
+	get_vgpr_size_bytes(s_restore_mem_offset, s_restore_size)
+	get_svgpr_size_bytes(s_restore_tmp)
+	s_add_u32	s_restore_mem_offset, s_restore_mem_offset, s_restore_tmp
+	s_add_u32	s_restore_mem_offset, s_restore_mem_offset, get_sgpr_size_bytes()
+	s_add_u32	s_restore_mem_offset, s_restore_mem_offset, get_hwreg_size_bytes()
+
+	s_mov_b32	s_restore_buf_rsrc2, 0x1000000				//NUM_RECORDS in bytes
+
+	s_lshr_b32	m0, s_restore_size, S_WAVE_SIZE
+	s_and_b32	m0, m0, 1
+	s_cmp_eq_u32	m0, 1
+	s_mov_b32	m0, 0x0
+	s_cbranch_scc1	L_RESTORE_LDS_LOOP_W64
+
+L_RESTORE_LDS_LOOP_W32:
+	buffer_load_dword       v0, v0, s_restore_buf_rsrc0, s_restore_mem_offset
+	s_wait_idle
+	ds_store_addtid_b32     v0
+	s_add_u32	m0, m0, 128						// 128 DW
+	s_add_u32	s_restore_mem_offset, s_restore_mem_offset, 128		//mem offset increased by 128DW
+	s_cmp_lt_u32	m0, s_restore_alloc_size				//scc=(m0 < s_restore_alloc_size) ? 1 : 0
+	s_cbranch_scc1	L_RESTORE_LDS_LOOP_W32					//LDS restore is complete?
+	s_branch	L_RESTORE_VGPR
+
+L_RESTORE_LDS_LOOP_W64:
+	buffer_load_dword       v0, v0, s_restore_buf_rsrc0, s_restore_mem_offset
+	s_wait_idle
+	ds_store_addtid_b32     v0
+	s_add_u32	m0, m0, 256						// 256 DW
+	s_add_u32	s_restore_mem_offset, s_restore_mem_offset, 256		//mem offset increased by 256DW
+	s_cmp_lt_u32	m0, s_restore_alloc_size				//scc=(m0 < s_restore_alloc_size) ? 1 : 0
+	s_cbranch_scc1	L_RESTORE_LDS_LOOP_W64					//LDS restore is complete?
+
+	/* restore VGPRs */
+L_RESTORE_VGPR:
+	// VGPR SR memory offset : 0
+	s_mov_b32	s_restore_mem_offset, 0x0
+	s_mov_b32	exec_lo, 0xFFFFFFFF					//need every thread from now on
+	s_lshr_b32	m0, s_restore_size, S_WAVE_SIZE
+	s_and_b32	m0, m0, 1
+	s_cmp_eq_u32	m0, 1
+	s_cbranch_scc1	L_ENABLE_RESTORE_VGPR_EXEC_HI
+	s_mov_b32	exec_hi, 0x00000000
+	s_branch	L_RESTORE_VGPR_NORMAL
+L_ENABLE_RESTORE_VGPR_EXEC_HI:
+	s_mov_b32	exec_hi, 0xFFFFFFFF
+L_RESTORE_VGPR_NORMAL:
+	s_getreg_b32	s_restore_alloc_size, hwreg(HW_REG_WAVE_GPR_ALLOC,SQ_WAVE_GPR_ALLOC_VGPR_SIZE_SHIFT,SQ_WAVE_GPR_ALLOC_VGPR_SIZE_SIZE)
+	s_add_u32	s_restore_alloc_size, s_restore_alloc_size, 1
+	s_lshl_b32	s_restore_alloc_size, s_restore_alloc_size, 2		//Number of VGPRs = (vgpr_size + 1) * 4    (non-zero value)
+	//determine it is wave32 or wave64
+	s_lshr_b32	m0, s_restore_size, S_WAVE_SIZE
+	s_and_b32	m0, m0, 1
+	s_cmp_eq_u32	m0, 1
+	s_cbranch_scc1	L_RESTORE_VGPR_WAVE64
+
+	s_mov_b32	s_restore_buf_rsrc2, 0x1000000				//NUM_RECORDS in bytes
+
+	// VGPR load using dw burst
+	s_mov_b32	s_restore_mem_offset_save, s_restore_mem_offset		// restore start with v1, v0 will be the last
+	s_add_u32	s_restore_mem_offset, s_restore_mem_offset, 128*4
+	s_mov_b32	m0, 4							//VGPR initial index value = 4
+	s_cmp_lt_u32	m0, s_restore_alloc_size
+	s_cbranch_scc0	L_RESTORE_SGPR
+
+L_RESTORE_VGPR_WAVE32_LOOP:
+	buffer_load_dword	v0, v0, s_restore_buf_rsrc0, s_restore_mem_offset scope:SCOPE_SYS
+	buffer_load_dword	v1, v0, s_restore_buf_rsrc0, s_restore_mem_offset scope:SCOPE_SYS offset:128
+	buffer_load_dword	v2, v0, s_restore_buf_rsrc0, s_restore_mem_offset scope:SCOPE_SYS offset:128*2
+	buffer_load_dword	v3, v0, s_restore_buf_rsrc0, s_restore_mem_offset scope:SCOPE_SYS offset:128*3
+	s_wait_idle
+	v_movreld_b32	v0, v0							//v[0+m0] = v0
+	v_movreld_b32	v1, v1
+	v_movreld_b32	v2, v2
+	v_movreld_b32	v3, v3
+	s_add_u32	m0, m0, 4						//next vgpr index
+	s_add_u32	s_restore_mem_offset, s_restore_mem_offset, 128*4	//every buffer_load_dword does 128 bytes
+	s_cmp_lt_u32	m0, s_restore_alloc_size				//scc = (m0 < s_restore_alloc_size) ? 1 : 0
+	s_cbranch_scc1	L_RESTORE_VGPR_WAVE32_LOOP				//VGPR restore (except v0) is complete?
+
+	/* VGPR restore on v0 */
+	buffer_load_dword	v0, v0, s_restore_buf_rsrc0, s_restore_mem_offset_save scope:SCOPE_SYS
+	buffer_load_dword	v1, v0, s_restore_buf_rsrc0, s_restore_mem_offset_save scope:SCOPE_SYS offset:128
+	buffer_load_dword	v2, v0, s_restore_buf_rsrc0, s_restore_mem_offset_save scope:SCOPE_SYS offset:128*2
+	buffer_load_dword	v3, v0, s_restore_buf_rsrc0, s_restore_mem_offset_save scope:SCOPE_SYS offset:128*3
+	s_wait_idle
+
+	s_branch	L_RESTORE_SGPR
+
+L_RESTORE_VGPR_WAVE64:
+	s_mov_b32	s_restore_buf_rsrc2, 0x1000000				//NUM_RECORDS in bytes
+
+	// VGPR load using dw burst
+	s_mov_b32	s_restore_mem_offset_save, s_restore_mem_offset		// restore start with v4, v0 will be the last
+	s_add_u32	s_restore_mem_offset, s_restore_mem_offset, 256*4
+	s_mov_b32	m0, 4							//VGPR initial index value = 4
+	s_cmp_lt_u32	m0, s_restore_alloc_size
+	s_cbranch_scc0	L_RESTORE_SHARED_VGPR
+
+L_RESTORE_VGPR_WAVE64_LOOP:
+	buffer_load_dword	v0, v0, s_restore_buf_rsrc0, s_restore_mem_offset scope:SCOPE_SYS
+	buffer_load_dword	v1, v0, s_restore_buf_rsrc0, s_restore_mem_offset scope:SCOPE_SYS offset:256
+	buffer_load_dword	v2, v0, s_restore_buf_rsrc0, s_restore_mem_offset scope:SCOPE_SYS offset:256*2
+	buffer_load_dword	v3, v0, s_restore_buf_rsrc0, s_restore_mem_offset scope:SCOPE_SYS offset:256*3
+	s_wait_idle
+	v_movreld_b32	v0, v0							//v[0+m0] = v0
+	v_movreld_b32	v1, v1
+	v_movreld_b32	v2, v2
+	v_movreld_b32	v3, v3
+	s_add_u32	m0, m0, 4						//next vgpr index
+	s_add_u32	s_restore_mem_offset, s_restore_mem_offset, 256*4	//every buffer_load_dword does 256 bytes
+	s_cmp_lt_u32	m0, s_restore_alloc_size				//scc = (m0 < s_restore_alloc_size) ? 1 : 0
+	s_cbranch_scc1	L_RESTORE_VGPR_WAVE64_LOOP				//VGPR restore (except v0) is complete?
+
+L_RESTORE_SHARED_VGPR:
+	s_getreg_b32	s_restore_alloc_size, hwreg(HW_REG_WAVE_LDS_ALLOC,SQ_WAVE_LDS_ALLOC_VGPR_SHARED_SIZE_SHIFT,SQ_WAVE_LDS_ALLOC_VGPR_SHARED_SIZE_SIZE)	//shared_vgpr_size
+	s_and_b32	s_restore_alloc_size, s_restore_alloc_size, 0xFFFFFFFF	//shared_vgpr_size is zero?
+	s_cbranch_scc0	L_RESTORE_V0						//no shared_vgpr used?
+	s_lshl_b32	s_restore_alloc_size, s_restore_alloc_size, 3		//Number of SHARED_VGPRs = shared_vgpr_size * 8    (non-zero value)
+	//m0 now has the value of normal vgpr count, just add the m0 with shared_vgpr count to get the total count.
+	//restore shared_vgpr will start from the index of m0
+	s_add_u32	s_restore_alloc_size, s_restore_alloc_size, m0
+	s_mov_b32	exec_lo, 0xFFFFFFFF
+	s_mov_b32	exec_hi, 0x00000000
+L_RESTORE_SHARED_VGPR_WAVE64_LOOP:
+	buffer_load_dword	v0, v0, s_restore_buf_rsrc0, s_restore_mem_offset scope:SCOPE_SYS
+	s_wait_idle
+	v_movreld_b32	v0, v0							//v[0+m0] = v0
+	s_add_u32	m0, m0, 1						//next vgpr index
+	s_add_u32	s_restore_mem_offset, s_restore_mem_offset, 128
+	s_cmp_lt_u32	m0, s_restore_alloc_size				//scc = (m0 < s_restore_alloc_size) ? 1 : 0
+	s_cbranch_scc1	L_RESTORE_SHARED_VGPR_WAVE64_LOOP			//VGPR restore (except v0) is complete?
+
+	s_mov_b32	exec_hi, 0xFFFFFFFF					//restore back exec_hi before restoring V0!!
+
+	/* VGPR restore on v0 */
+L_RESTORE_V0:
+	buffer_load_dword	v0, v0, s_restore_buf_rsrc0, s_restore_mem_offset_save scope:SCOPE_SYS
+	buffer_load_dword	v1, v0, s_restore_buf_rsrc0, s_restore_mem_offset_save scope:SCOPE_SYS offset:256
+	buffer_load_dword	v2, v0, s_restore_buf_rsrc0, s_restore_mem_offset_save scope:SCOPE_SYS offset:256*2
+	buffer_load_dword	v3, v0, s_restore_buf_rsrc0, s_restore_mem_offset_save scope:SCOPE_SYS offset:256*3
+	s_wait_idle
+
+	/* restore SGPRs */
+	//will be 2+8+16*6
+	// SGPR SR memory offset : size(VGPR)+size(SVGPR)
+L_RESTORE_SGPR:
+	get_vgpr_size_bytes(s_restore_mem_offset, s_restore_size)
+	get_svgpr_size_bytes(s_restore_tmp)
+	s_add_u32	s_restore_mem_offset, s_restore_mem_offset, s_restore_tmp
+	s_add_u32	s_restore_mem_offset, s_restore_mem_offset, get_sgpr_size_bytes()
+	s_sub_u32	s_restore_mem_offset, s_restore_mem_offset, 20*4	//s108~s127 is not saved
+
+	s_mov_b32	s_restore_buf_rsrc2, 0x1000000				//NUM_RECORDS in bytes
+
+	s_mov_b32	m0, s_sgpr_save_num
+
+	read_4sgpr_from_mem(s0, s_restore_buf_rsrc0, s_restore_mem_offset)
+	s_wait_idle
+
+	s_sub_u32	m0, m0, 4						// Restore from S[0] to S[104]
+	s_nop		0							// hazard SALU M0=> S_MOVREL
+
+	s_movreld_b64	s0, s0							//s[0+m0] = s0
+	s_movreld_b64	s2, s2
+
+	read_8sgpr_from_mem(s0, s_restore_buf_rsrc0, s_restore_mem_offset)
+	s_wait_idle
+
+	s_sub_u32	m0, m0, 8						// Restore from S[0] to S[96]
+	s_nop		0							// hazard SALU M0=> S_MOVREL
+
+	s_movreld_b64	s0, s0							//s[0+m0] = s0
+	s_movreld_b64	s2, s2
+	s_movreld_b64	s4, s4
+	s_movreld_b64	s6, s6
+
+ L_RESTORE_SGPR_LOOP:
+	read_16sgpr_from_mem(s0, s_restore_buf_rsrc0, s_restore_mem_offset)
+	s_wait_idle
+
+	s_sub_u32	m0, m0, 16						// Restore from S[n] to S[0]
+	s_nop		0							// hazard SALU M0=> S_MOVREL
+
+	s_movreld_b64	s0, s0							//s[0+m0] = s0
+	s_movreld_b64	s2, s2
+	s_movreld_b64	s4, s4
+	s_movreld_b64	s6, s6
+	s_movreld_b64	s8, s8
+	s_movreld_b64	s10, s10
+	s_movreld_b64	s12, s12
+	s_movreld_b64	s14, s14
+
+	s_cmp_eq_u32	m0, 0							//scc = (m0 < s_sgpr_save_num) ? 1 : 0
+	s_cbranch_scc0	L_RESTORE_SGPR_LOOP
+
+	// s_barrier with STATE_PRIV.TRAP_AFTER_INST=1, STATUS.PRIV=1 incorrectly asserts debug exception.
+	// Clear DEBUG_EN before and restore MODE after the barrier.
+	s_setreg_imm32_b32	hwreg(HW_REG_WAVE_MODE), 0
+
+	/* restore HW registers */
+L_RESTORE_HWREG:
+	// HWREG SR memory offset : size(VGPR)+size(SVGPR)+size(SGPR)
+	get_vgpr_size_bytes(s_restore_mem_offset, s_restore_size)
+	get_svgpr_size_bytes(s_restore_tmp)
+	s_add_u32	s_restore_mem_offset, s_restore_mem_offset, s_restore_tmp
+	s_add_u32	s_restore_mem_offset, s_restore_mem_offset, get_sgpr_size_bytes()
+
+	s_mov_b32	s_restore_buf_rsrc2, 0x1000000				//NUM_RECORDS in bytes
+
+	// Restore s_restore_spi_init_hi before the saved value gets clobbered.
+	s_mov_b32 s_restore_spi_init_hi, s_restore_spi_init_hi_save
+
+	read_hwreg_from_mem(s_restore_m0, s_restore_buf_rsrc0, s_restore_mem_offset)
+	read_hwreg_from_mem(s_restore_pc_lo, s_restore_buf_rsrc0, s_restore_mem_offset)
+	read_hwreg_from_mem(s_restore_pc_hi, s_restore_buf_rsrc0, s_restore_mem_offset)
+	read_hwreg_from_mem(s_restore_exec_lo, s_restore_buf_rsrc0, s_restore_mem_offset)
+	read_hwreg_from_mem(s_restore_exec_hi, s_restore_buf_rsrc0, s_restore_mem_offset)
+	read_hwreg_from_mem(s_restore_state_priv, s_restore_buf_rsrc0, s_restore_mem_offset)
+	read_hwreg_from_mem(s_restore_excp_flag_priv, s_restore_buf_rsrc0, s_restore_mem_offset)
+	read_hwreg_from_mem(s_restore_xnack_mask, s_restore_buf_rsrc0, s_restore_mem_offset)
+	read_hwreg_from_mem(s_restore_mode, s_restore_buf_rsrc0, s_restore_mem_offset)
+	read_hwreg_from_mem(s_restore_flat_scratch, s_restore_buf_rsrc0, s_restore_mem_offset)
+	s_wait_idle
+
+	s_setreg_b32	hwreg(HW_REG_WAVE_SCRATCH_BASE_LO), s_restore_flat_scratch
+
+	read_hwreg_from_mem(s_restore_flat_scratch, s_restore_buf_rsrc0, s_restore_mem_offset)
+	s_wait_idle
+
+	s_setreg_b32	hwreg(HW_REG_WAVE_SCRATCH_BASE_HI), s_restore_flat_scratch
+
+	read_hwreg_from_mem(s_restore_tmp, s_restore_buf_rsrc0, s_restore_mem_offset)
+	s_wait_idle
+	s_setreg_b32	hwreg(HW_REG_WAVE_EXCP_FLAG_USER), s_restore_tmp
+
+	read_hwreg_from_mem(s_restore_tmp, s_restore_buf_rsrc0, s_restore_mem_offset)
+	s_wait_idle
+	s_setreg_b32	hwreg(HW_REG_WAVE_TRAP_CTRL), s_restore_tmp
+
+	// Only the first wave needs to restore the workgroup barrier.
+	s_and_b32	s_restore_tmp, s_restore_spi_init_hi, S_RESTORE_SPI_INIT_FIRST_WAVE_MASK
+	s_cbranch_scc0	L_SKIP_BARRIER_RESTORE
+
+	// Skip over WAVE_STATUS, since there is no state to restore from it
+	s_add_u32	s_restore_mem_offset, s_restore_mem_offset, 4
+
+	read_hwreg_from_mem(s_restore_tmp, s_restore_buf_rsrc0, s_restore_mem_offset)
+	s_wait_idle
+
+	s_bitcmp1_b32	s_restore_tmp, BARRIER_STATE_VALID_OFFSET
+	s_cbranch_scc0	L_SKIP_BARRIER_RESTORE
+
+	// extract the saved signal count from s_restore_tmp
+	s_lshr_b32	s_restore_tmp, s_restore_tmp, BARRIER_STATE_SIGNAL_OFFSET
+
+	// We need to call s_barrier_signal repeatedly to restore the signal
+	// count of the work group barrier.  The member count is already
+	// initialized with the number of waves in the work group.
+L_BARRIER_RESTORE_LOOP:
+	s_and_b32	s_restore_tmp, s_restore_tmp, s_restore_tmp
+	s_cbranch_scc0	L_SKIP_BARRIER_RESTORE
+	s_barrier_signal	-1
+	s_add_i32	s_restore_tmp, s_restore_tmp, -1
+	s_branch	L_BARRIER_RESTORE_LOOP
+
+L_SKIP_BARRIER_RESTORE:
+
+	s_mov_b32	m0, s_restore_m0
+	s_mov_b32	exec_lo, s_restore_exec_lo
+	s_mov_b32	exec_hi, s_restore_exec_hi
+
+	// EXCP_FLAG_PRIV.SAVE_CONTEXT and HOST_TRAP may have changed.
+	// Only restore the other fields to avoid clobbering them.
+	s_setreg_b32	hwreg(HW_REG_WAVE_EXCP_FLAG_PRIV, 0, SQ_WAVE_EXCP_FLAG_PRIV_RESTORE_PART_1_SIZE), s_restore_excp_flag_priv
+	s_lshr_b32	s_restore_excp_flag_priv, s_restore_excp_flag_priv, SQ_WAVE_EXCP_FLAG_PRIV_RESTORE_PART_2_SHIFT
+	s_setreg_b32	hwreg(HW_REG_WAVE_EXCP_FLAG_PRIV, SQ_WAVE_EXCP_FLAG_PRIV_RESTORE_PART_2_SHIFT, SQ_WAVE_EXCP_FLAG_PRIV_RESTORE_PART_2_SIZE), s_restore_excp_flag_priv
+	s_lshr_b32	s_restore_excp_flag_priv, s_restore_excp_flag_priv, SQ_WAVE_EXCP_FLAG_PRIV_RESTORE_PART_3_SHIFT - SQ_WAVE_EXCP_FLAG_PRIV_RESTORE_PART_2_SHIFT
+	s_setreg_b32	hwreg(HW_REG_WAVE_EXCP_FLAG_PRIV, SQ_WAVE_EXCP_FLAG_PRIV_RESTORE_PART_3_SHIFT, SQ_WAVE_EXCP_FLAG_PRIV_RESTORE_PART_3_SIZE), s_restore_excp_flag_priv
+
+	s_setreg_b32	hwreg(HW_REG_WAVE_MODE), s_restore_mode
+
+	// Restore trap temporaries 4-11, 13 initialized by SPI debug dispatch logic
+	// ttmp SR memory offset : size(VGPR)+size(SVGPR)+size(SGPR)+0x40
+	get_vgpr_size_bytes(s_restore_ttmps_lo, s_restore_size)
+	get_svgpr_size_bytes(s_restore_ttmps_hi)
+	s_add_u32	s_restore_ttmps_lo, s_restore_ttmps_lo, s_restore_ttmps_hi
+	s_add_u32	s_restore_ttmps_lo, s_restore_ttmps_lo, get_sgpr_size_bytes()
+	s_add_u32	s_restore_ttmps_lo, s_restore_ttmps_lo, s_restore_buf_rsrc0
+	s_addc_u32	s_restore_ttmps_hi, s_restore_buf_rsrc1, 0x0
+	s_and_b32	s_restore_ttmps_hi, s_restore_ttmps_hi, 0xFFFF
+	s_load_dwordx4	[ttmp4, ttmp5, ttmp6, ttmp7], [s_restore_ttmps_lo, s_restore_ttmps_hi], 0x50 scope:SCOPE_SYS
+	s_load_dwordx4	[ttmp8, ttmp9, ttmp10, ttmp11], [s_restore_ttmps_lo, s_restore_ttmps_hi], 0x60 scope:SCOPE_SYS
+	s_load_dword	ttmp13, [s_restore_ttmps_lo, s_restore_ttmps_hi], 0x74 scope:SCOPE_SYS
+	s_wait_idle
+
+	s_and_b32	s_restore_pc_hi, s_restore_pc_hi, 0x0000ffff		//pc[47:32] //Do it here in order not to affect STATUS
+	s_and_b64	exec, exec, exec					// Restore STATUS.EXECZ, not writable by s_setreg_b32
+	s_and_b64	vcc, vcc, vcc						// Restore STATUS.VCCZ, not writable by s_setreg_b32
+
+	s_setreg_b32	hwreg(HW_REG_WAVE_STATE_PRIV), s_restore_state_priv	// SCC is included, which is changed by previous salu
+
+	// Make barrier and LDS state visible to all waves in the group.
+	// STATE_PRIV.BARRIER_COMPLETE may change after this point.
+	s_barrier_signal	-2
+	s_barrier_wait	-2
+
+	s_rfe_b64	s_restore_pc_lo						//Return to the main shader program and resume execution
+
+L_END_PGM:
+	// Make sure that no wave of the workgroup can exit the trap handler
+	// before the workgroup barrier state is saved.
+	s_barrier_signal	-2
+	s_barrier_wait	-2
+	s_endpgm_saved
+end
+
+function write_16sgpr_to_v2(s, lane_offset)
+	// Copy into VGPR for later TCP store.
+	for var sgpr_idx = 0; sgpr_idx < 16; sgpr_idx ++
+		v_writelane_b32	v2, s[sgpr_idx], sgpr_idx + lane_offset
+	end
+	valu_sgpr_hazard()
+	s_add_u32	ttmp13, ttmp13, 0x10
+end
+
+function write_12sgpr_to_v2(s)
+	// Copy into VGPR for later TCP store.
+	for var sgpr_idx = 0; sgpr_idx < 12; sgpr_idx ++
+		v_writelane_b32	v2, s[sgpr_idx], sgpr_idx
+	end
+	valu_sgpr_hazard()
+end
+
+function read_hwreg_from_mem(s, s_rsrc, s_mem_offset)
+	s_buffer_load_dword	s, s_rsrc, s_mem_offset scope:SCOPE_SYS
+	s_add_u32	s_mem_offset, s_mem_offset, 4
+end
+
+function read_16sgpr_from_mem(s, s_rsrc, s_mem_offset)
+	s_sub_u32	s_mem_offset, s_mem_offset, 4*16
+	s_buffer_load_dwordx16	s, s_rsrc, s_mem_offset scope:SCOPE_SYS
+end
+
+function read_8sgpr_from_mem(s, s_rsrc, s_mem_offset)
+	s_sub_u32	s_mem_offset, s_mem_offset, 4*8
+	s_buffer_load_dwordx8	s, s_rsrc, s_mem_offset scope:SCOPE_SYS
+end
+
+function read_4sgpr_from_mem(s, s_rsrc, s_mem_offset)
+	s_sub_u32	s_mem_offset, s_mem_offset, 4*4
+	s_buffer_load_dwordx4	s, s_rsrc, s_mem_offset scope:SCOPE_SYS
+end
+
+function get_vgpr_size_bytes(s_vgpr_size_byte, s_size)
+	s_getreg_b32	s_vgpr_size_byte, hwreg(HW_REG_WAVE_GPR_ALLOC,SQ_WAVE_GPR_ALLOC_VGPR_SIZE_SHIFT,SQ_WAVE_GPR_ALLOC_VGPR_SIZE_SIZE)
+	s_add_u32	s_vgpr_size_byte, s_vgpr_size_byte, 1
+	s_bitcmp1_b32	s_size, S_WAVE_SIZE
+	s_cbranch_scc1	L_ENABLE_SHIFT_W64
+	s_lshl_b32	s_vgpr_size_byte, s_vgpr_size_byte, (2+7)		//Number of VGPRs = (vgpr_size + 1) * 4 * 32 * 4   (non-zero value)
+	s_branch	L_SHIFT_DONE
+L_ENABLE_SHIFT_W64:
+	s_lshl_b32	s_vgpr_size_byte, s_vgpr_size_byte, (2+8)		//Number of VGPRs = (vgpr_size + 1) * 4 * 64 * 4   (non-zero value)
+L_SHIFT_DONE:
+end
+
+function get_svgpr_size_bytes(s_svgpr_size_byte)
+	s_getreg_b32	s_svgpr_size_byte, hwreg(HW_REG_WAVE_LDS_ALLOC,SQ_WAVE_LDS_ALLOC_VGPR_SHARED_SIZE_SHIFT,SQ_WAVE_LDS_ALLOC_VGPR_SHARED_SIZE_SIZE)
+	s_lshl_b32	s_svgpr_size_byte, s_svgpr_size_byte, (3+7)
+end
+
+function get_sgpr_size_bytes
+	return 512
+end
+
+function get_hwreg_size_bytes
+	return 128
+end
+
+function get_wave_size2(s_reg)
+	s_getreg_b32	s_reg, hwreg(HW_REG_WAVE_STATUS,SQ_WAVE_STATUS_WAVE64_SHIFT,SQ_WAVE_STATUS_WAVE64_SIZE)
+	s_lshl_b32	s_reg, s_reg, S_WAVE_SIZE
+end
+
+function valu_sgpr_hazard
+#if HAVE_VALU_SGPR_HAZARD
+	for var rep = 0; rep < 8; rep ++
+		ds_nop
+	end
+#endif
+end
diff --git a/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler_gfx9.asm b/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler_gfx9.asm
index eed78a04e7c7..6869e07a2fff 100644
--- a/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler_gfx9.asm
+++ b/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler_gfx9.asm
@@ -33,21 +33,35 @@
  * aldebaran:
  *   cpp -DASIC_FAMILY=CHIP_ALDEBARAN cwsr_trap_handler_gfx9.asm -P -o aldebaran.sp3
  *   sp3 aldebaran.sp3 -hex aldebaran.hex
+ *
+ * gc_9_4_3:
+ *   cpp -DASIC_FAMILY=GC_9_4_3 cwsr_trap_handler_gfx9.asm -P -o gc_9_4_3.sp3
+ *   sp3 gc_9_4_3.sp3 -hex gc_9_4_3.hex
+ *
+ * gc_9_5_0:
+ *   cpp -DASIC_FAMILY=GC_9_5_0 cwsr_trap_handler_gfx9.asm -P -o gc_9_5_0.sp3
+ *   sp3 gc_9_5_0.sp3 -hex gc_9_5_0.hex
  */
 
 #define CHIP_VEGAM 18
 #define CHIP_ARCTURUS 23
 #define CHIP_ALDEBARAN 25
+#define CHIP_GC_9_4_3 26
+#define CHIP_GC_9_5_0 27
 
 var ACK_SQC_STORE		    =	1		    //workaround for suspected SQC store bug causing incorrect stores under concurrency
 var SAVE_AFTER_XNACK_ERROR	    =	1		    //workaround for TCP store failure after XNACK error when ALLOW_REPLAY=0, for debugger
-var SINGLE_STEP_MISSED_WORKAROUND   =	1		    //workaround for lost MODE.DEBUG_EN exception when SAVECTX raised
+var SINGLE_STEP_MISSED_WORKAROUND   =	(ASIC_FAMILY <= CHIP_ALDEBARAN)	//workaround for lost MODE.DEBUG_EN exception when SAVECTX raised
+
+#if ASIC_FAMILY < CHIP_GC_9_4_3
+#define VMEM_MODIFIERS slc:1 glc:1
+#else
+#define VMEM_MODIFIERS sc0:1 nt:1
+#endif
 
 /**************************************************************************/
 /*			variables					  */
 /**************************************************************************/
-var SQ_WAVE_STATUS_INST_ATC_SHIFT  = 23
-var SQ_WAVE_STATUS_INST_ATC_MASK   = 0x00800000
 var SQ_WAVE_STATUS_SPI_PRIO_SHIFT  = 1
 var SQ_WAVE_STATUS_SPI_PRIO_MASK   = 0x00000006
 var SQ_WAVE_STATUS_HALT_MASK       = 0x2000
@@ -56,9 +70,16 @@ var SQ_WAVE_STATUS_PRE_SPI_PRIO_SIZE    = 1
 var SQ_WAVE_STATUS_POST_SPI_PRIO_SHIFT  = 3
 var SQ_WAVE_STATUS_POST_SPI_PRIO_SIZE   = 29
 var SQ_WAVE_STATUS_ALLOW_REPLAY_MASK    = 0x400000
+var SQ_WAVE_STATUS_ECC_ERR_MASK         = 0x20000
 
 var SQ_WAVE_LDS_ALLOC_LDS_SIZE_SHIFT	= 12
+#if ASIC_FAMILY >= CHIP_GC_9_5_0
+var SQ_WAVE_LDS_ALLOC_LDS_SIZE_SIZE	= 11
+var LDS_RESTORE_GRANULARITY_BYTES	= 1280
+#else
 var SQ_WAVE_LDS_ALLOC_LDS_SIZE_SIZE	= 9
+var LDS_RESTORE_GRANULARITY_BYTES	= 512
+#endif
 var SQ_WAVE_GPR_ALLOC_VGPR_SIZE_SIZE	= 6
 var SQ_WAVE_GPR_ALLOC_SGPR_SIZE_SIZE	= 3			//FIXME	 sq.blk still has 4 bits at this time while SQ programming guide has 3 bits
 var SQ_WAVE_GPR_ALLOC_SGPR_SIZE_SHIFT	= 24
@@ -72,10 +93,16 @@ var SQ_WAVE_GPR_ALLOC_VGPR_SIZE_SHIFT	= 8
 #endif
 
 var SQ_WAVE_TRAPSTS_SAVECTX_MASK    =	0x400
-var SQ_WAVE_TRAPSTS_EXCE_MASK	    =	0x1FF			// Exception mask
+var SQ_WAVE_TRAPSTS_EXCP_MASK	    =	0x1FF
 var SQ_WAVE_TRAPSTS_SAVECTX_SHIFT   =	10
+var SQ_WAVE_TRAPSTS_ADDR_WATCH_MASK =	0x80
+var SQ_WAVE_TRAPSTS_ADDR_WATCH_SHIFT =	7
 var SQ_WAVE_TRAPSTS_MEM_VIOL_MASK   =	0x100
 var SQ_WAVE_TRAPSTS_MEM_VIOL_SHIFT  =	8
+var SQ_WAVE_TRAPSTS_HOST_TRAP_MASK  =	0x400000
+var SQ_WAVE_TRAPSTS_WAVE_BEGIN_MASK =	0x800000
+var SQ_WAVE_TRAPSTS_WAVE_END_MASK   =	0x1000000
+var SQ_WAVE_TRAPSTS_TRAP_AFTER_INST_MASK =  0x2000000
 var SQ_WAVE_TRAPSTS_PRE_SAVECTX_MASK	=   0x3FF
 var SQ_WAVE_TRAPSTS_PRE_SAVECTX_SHIFT	=   0x0
 var SQ_WAVE_TRAPSTS_PRE_SAVECTX_SIZE	=   10
@@ -83,37 +110,30 @@ var SQ_WAVE_TRAPSTS_POST_SAVECTX_MASK	=   0xFFFFF800
 var SQ_WAVE_TRAPSTS_POST_SAVECTX_SHIFT	=   11
 var SQ_WAVE_TRAPSTS_POST_SAVECTX_SIZE	=   21
 var SQ_WAVE_TRAPSTS_ILLEGAL_INST_MASK	=   0x800
+var SQ_WAVE_TRAPSTS_EXCP_HI_MASK	=   0x7000
 var SQ_WAVE_TRAPSTS_XNACK_ERROR_MASK	=   0x10000000
 
-var SQ_WAVE_IB_STS_RCNT_SHIFT		=   16			//FIXME
+var SQ_WAVE_MODE_EXCP_EN_SHIFT		=   12
+var SQ_WAVE_MODE_EXCP_EN_ADDR_WATCH_SHIFT	= 19
+
 var SQ_WAVE_IB_STS_FIRST_REPLAY_SHIFT	=   15			//FIXME
 var SQ_WAVE_IB_STS_RCNT_FIRST_REPLAY_MASK	= 0x1F8000
-var SQ_WAVE_IB_STS_RCNT_FIRST_REPLAY_MASK_NEG	= 0x00007FFF	//FIXME
 
 var SQ_WAVE_MODE_DEBUG_EN_MASK		=   0x800
 
-var SQ_BUF_RSRC_WORD1_ATC_SHIFT	    =	24
-var SQ_BUF_RSRC_WORD3_MTYPE_SHIFT   =	27
-
-var TTMP11_SAVE_RCNT_FIRST_REPLAY_SHIFT	=   26			// bits [31:26] unused by SPI debug data
-var TTMP11_SAVE_RCNT_FIRST_REPLAY_MASK	=   0xFC000000
+var TTMP_SAVE_RCNT_FIRST_REPLAY_SHIFT	=   26			// bits [31:26] unused by SPI debug data
+var TTMP_SAVE_RCNT_FIRST_REPLAY_MASK	=   0xFC000000
+var TTMP_DEBUG_TRAP_ENABLED_SHIFT	=   23
+var TTMP_DEBUG_TRAP_ENABLED_MASK	=   0x800000
 
 /*	Save	    */
 var S_SAVE_BUF_RSRC_WORD1_STRIDE	=   0x00040000		//stride is 4 bytes
 var S_SAVE_BUF_RSRC_WORD3_MISC		=   0x00807FAC		//SQ_SEL_X/Y/Z/W, BUF_NUM_FORMAT_FLOAT, (0 for MUBUF stride[17:14] when ADD_TID_ENABLE and BUF_DATA_FORMAT_32 for MTBUF), ADD_TID_ENABLE
-
-var S_SAVE_SPI_INIT_ATC_MASK		=   0x08000000		//bit[27]: ATC bit
-var S_SAVE_SPI_INIT_ATC_SHIFT		=   27
-var S_SAVE_SPI_INIT_MTYPE_MASK		=   0x70000000		//bit[30:28]: Mtype
-var S_SAVE_SPI_INIT_MTYPE_SHIFT		=   28
+var S_SAVE_PC_HI_TRAP_ID_MASK		=   0x00FF0000
+var S_SAVE_PC_HI_HT_MASK		=   0x01000000
 var S_SAVE_SPI_INIT_FIRST_WAVE_MASK	=   0x04000000		//bit[26]: FirstWaveInTG
 var S_SAVE_SPI_INIT_FIRST_WAVE_SHIFT	=   26
 
-var S_SAVE_PC_HI_RCNT_SHIFT		=   27			//FIXME	 check with Brian to ensure all fields other than PC[47:0] can be used
-var S_SAVE_PC_HI_RCNT_MASK		=   0xF8000000		//FIXME
-var S_SAVE_PC_HI_FIRST_REPLAY_SHIFT	=   26			//FIXME
-var S_SAVE_PC_HI_FIRST_REPLAY_MASK	=   0x04000000		//FIXME
-
 var s_save_spi_init_lo		    =	exec_lo
 var s_save_spi_init_hi		    =	exec_hi
 
@@ -135,23 +155,19 @@ var s_save_alloc_size	    =	s_save_trapsts		//conflict
 var s_save_m0		    =	ttmp5
 var s_save_ttmps_lo	    =	s_save_tmp		//no conflict
 var s_save_ttmps_hi	    =	s_save_trapsts		//no conflict
+#if ASIC_FAMILY >= CHIP_GC_9_4_3
+var s_save_ib_sts       =	ttmp13
+#else
+var s_save_ib_sts       =	ttmp11
+#endif
 
 /*	Restore	    */
 var S_RESTORE_BUF_RSRC_WORD1_STRIDE	    =	S_SAVE_BUF_RSRC_WORD1_STRIDE
 var S_RESTORE_BUF_RSRC_WORD3_MISC	    =	S_SAVE_BUF_RSRC_WORD3_MISC
 
-var S_RESTORE_SPI_INIT_ATC_MASK		    =	0x08000000	    //bit[27]: ATC bit
-var S_RESTORE_SPI_INIT_ATC_SHIFT	    =	27
-var S_RESTORE_SPI_INIT_MTYPE_MASK	    =	0x70000000	    //bit[30:28]: Mtype
-var S_RESTORE_SPI_INIT_MTYPE_SHIFT	    =	28
 var S_RESTORE_SPI_INIT_FIRST_WAVE_MASK	    =	0x04000000	    //bit[26]: FirstWaveInTG
 var S_RESTORE_SPI_INIT_FIRST_WAVE_SHIFT	    =	26
 
-var S_RESTORE_PC_HI_RCNT_SHIFT		    =	S_SAVE_PC_HI_RCNT_SHIFT
-var S_RESTORE_PC_HI_RCNT_MASK		    =	S_SAVE_PC_HI_RCNT_MASK
-var S_RESTORE_PC_HI_FIRST_REPLAY_SHIFT	    =	S_SAVE_PC_HI_FIRST_REPLAY_SHIFT
-var S_RESTORE_PC_HI_FIRST_REPLAY_MASK	    =	S_SAVE_PC_HI_FIRST_REPLAY_MASK
-
 var s_restore_spi_init_lo		    =	exec_lo
 var s_restore_spi_init_hi		    =	exec_hi
 
@@ -199,71 +215,83 @@ L_JUMP_TO_RESTORE:
 L_SKIP_RESTORE:
 
     s_getreg_b32    s_save_status, hwreg(HW_REG_STATUS)				    //save STATUS since we will change SCC
-    s_andn2_b32	    s_save_status, s_save_status, SQ_WAVE_STATUS_SPI_PRIO_MASK	    //check whether this is for save
 
-if SINGLE_STEP_MISSED_WORKAROUND
-    // No single step exceptions if MODE.DEBUG_EN=0.
-    s_getreg_b32    ttmp2, hwreg(HW_REG_MODE)
-    s_and_b32       ttmp2, ttmp2, SQ_WAVE_MODE_DEBUG_EN_MASK
-    s_cbranch_scc0  L_NO_SINGLE_STEP_WORKAROUND
+    // Clear SPI_PRIO: do not save with elevated priority.
+    // Clear ECC_ERR: prevents SQC store and triggers FATAL_HALT if setreg'd.
+    s_andn2_b32     s_save_status, s_save_status, SQ_WAVE_STATUS_SPI_PRIO_MASK|SQ_WAVE_STATUS_ECC_ERR_MASK
 
-    // Second-level trap already handled exception if STATUS.HALT=1.
-    s_and_b32       ttmp2, s_save_status, SQ_WAVE_STATUS_HALT_MASK
+    s_getreg_b32    s_save_trapsts, hwreg(HW_REG_TRAPSTS)
 
-    // Prioritize single step exception over context save.
-    // Second-level trap will halt wave and RFE, re-entering for SAVECTX.
-    s_cbranch_scc0  L_FETCH_2ND_TRAP
+    s_and_b32       ttmp2, s_save_status, SQ_WAVE_STATUS_HALT_MASK
+    s_cbranch_scc0  L_NOT_HALTED
 
-L_NO_SINGLE_STEP_WORKAROUND:
-end
+L_HALTED:
+    // Host trap may occur while wave is halted.
+    s_and_b32       ttmp2, s_save_pc_hi, S_SAVE_PC_HI_TRAP_ID_MASK
+    s_cbranch_scc1  L_FETCH_2ND_TRAP
 
-    s_getreg_b32    s_save_trapsts, hwreg(HW_REG_TRAPSTS)
+L_CHECK_SAVE:
     s_and_b32       ttmp2, s_save_trapsts, SQ_WAVE_TRAPSTS_SAVECTX_MASK    //check whether this is for save
     s_cbranch_scc1  L_SAVE					//this is the operation for save
 
-    // *********    Handle non-CWSR traps	*******************
-
-    // Illegal instruction is a non-maskable exception which blocks context save.
-    // Halt the wavefront and return from the trap.
-    s_and_b32       ttmp2, s_save_trapsts, SQ_WAVE_TRAPSTS_ILLEGAL_INST_MASK
-    s_cbranch_scc1  L_HALT_WAVE
-
-    // If STATUS.MEM_VIOL is asserted then we cannot fetch from the TMA.
-    // Instead, halt the wavefront and return from the trap.
-    s_and_b32       ttmp2, s_save_trapsts, SQ_WAVE_TRAPSTS_MEM_VIOL_MASK
-    s_cbranch_scc0  L_FETCH_2ND_TRAP
-
-L_HALT_WAVE:
-    // If STATUS.HALT is set then this fault must come from SQC instruction fetch.
-    // We cannot prevent further faults. Spin wait until context saved.
-    s_and_b32       ttmp2, s_save_status, SQ_WAVE_STATUS_HALT_MASK
-    s_cbranch_scc0  L_NOT_ALREADY_HALTED
-
-L_WAIT_CTX_SAVE:
+    // Wave is halted but neither host trap nor SAVECTX is raised.
+    // Caused by instruction fetch memory violation.
+    // Spin wait until context saved to prevent interrupt storm.
     s_sleep         0x10
-    s_getreg_b32    ttmp2, hwreg(HW_REG_TRAPSTS)
-    s_and_b32       ttmp2, ttmp2, SQ_WAVE_TRAPSTS_SAVECTX_MASK
-    s_cbranch_scc0  L_WAIT_CTX_SAVE
+    s_getreg_b32    s_save_trapsts, hwreg(HW_REG_TRAPSTS)
+    s_branch        L_CHECK_SAVE
+
+L_NOT_HALTED:
+    // Let second-level handle non-SAVECTX exception or trap.
+    // Any concurrent SAVECTX will be handled upon re-entry once halted.
+
+    // Check non-maskable exceptions. memory_violation, illegal_instruction
+    // and debugger (host trap, wave start/end, trap after instruction)
+    // exceptions always cause the wave to enter the trap handler.
+    s_and_b32       ttmp2, s_save_trapsts,      \
+        SQ_WAVE_TRAPSTS_MEM_VIOL_MASK         | \
+        SQ_WAVE_TRAPSTS_ILLEGAL_INST_MASK     | \
+        SQ_WAVE_TRAPSTS_HOST_TRAP_MASK        | \
+        SQ_WAVE_TRAPSTS_WAVE_BEGIN_MASK       | \
+        SQ_WAVE_TRAPSTS_WAVE_END_MASK         | \
+        SQ_WAVE_TRAPSTS_TRAP_AFTER_INST_MASK
+    s_cbranch_scc1  L_FETCH_2ND_TRAP
+
+    // Check for maskable exceptions in trapsts.excp and trapsts.excp_hi.
+    // Maskable exceptions only cause the wave to enter the trap handler if
+    // their respective bit in mode.excp_en is set.
+    s_and_b32       ttmp2, s_save_trapsts, SQ_WAVE_TRAPSTS_EXCP_MASK|SQ_WAVE_TRAPSTS_EXCP_HI_MASK
+    s_cbranch_scc0  L_CHECK_TRAP_ID
+
+    s_and_b32       ttmp3, s_save_trapsts, SQ_WAVE_TRAPSTS_ADDR_WATCH_MASK|SQ_WAVE_TRAPSTS_EXCP_HI_MASK
+    s_cbranch_scc0  L_NOT_ADDR_WATCH
+    s_bitset1_b32   ttmp2, SQ_WAVE_TRAPSTS_ADDR_WATCH_SHIFT // Check all addr_watch[123] exceptions against excp_en.addr_watch
+
+L_NOT_ADDR_WATCH:
+    s_getreg_b32    ttmp3, hwreg(HW_REG_MODE)
+    s_lshl_b32      ttmp2, ttmp2, SQ_WAVE_MODE_EXCP_EN_SHIFT
+    s_and_b32       ttmp2, ttmp2, ttmp3
+    s_cbranch_scc1  L_FETCH_2ND_TRAP
+
+L_CHECK_TRAP_ID:
+    // Check trap_id != 0
+    s_and_b32       ttmp2, s_save_pc_hi, S_SAVE_PC_HI_TRAP_ID_MASK
+    s_cbranch_scc1  L_FETCH_2ND_TRAP
 
-L_NOT_ALREADY_HALTED:
-    s_or_b32        s_save_status, s_save_status, SQ_WAVE_STATUS_HALT_MASK
+if SINGLE_STEP_MISSED_WORKAROUND
+    // Prioritize single step exception over context save.
+    // Second-level trap will halt wave and RFE, re-entering for SAVECTX.
+    s_getreg_b32    ttmp2, hwreg(HW_REG_MODE)
+    s_and_b32       ttmp2, ttmp2, SQ_WAVE_MODE_DEBUG_EN_MASK
+    s_cbranch_scc1  L_FETCH_2ND_TRAP
+end
 
-    // If the PC points to S_ENDPGM then context save will fail if STATUS.HALT is set.
-    // Rewind the PC to prevent this from occurring. The debugger compensates for this.
-    s_sub_u32       ttmp0, ttmp0, 0x8
-    s_subb_u32      ttmp1, ttmp1, 0x0
+    s_and_b32       ttmp2, s_save_trapsts, SQ_WAVE_TRAPSTS_SAVECTX_MASK
+    s_cbranch_scc1  L_SAVE
 
 L_FETCH_2ND_TRAP:
     // Preserve and clear scalar XNACK state before issuing scalar reads.
-    // Save IB_STS.FIRST_REPLAY[15] and IB_STS.RCNT[20:16] into unused space ttmp11[31:26].
-    s_getreg_b32    ttmp2, hwreg(HW_REG_IB_STS)
-    s_and_b32       ttmp3, ttmp2, SQ_WAVE_IB_STS_RCNT_FIRST_REPLAY_MASK
-    s_lshl_b32      ttmp3, ttmp3, (TTMP11_SAVE_RCNT_FIRST_REPLAY_SHIFT - SQ_WAVE_IB_STS_FIRST_REPLAY_SHIFT)
-    s_andn2_b32     ttmp11, ttmp11, TTMP11_SAVE_RCNT_FIRST_REPLAY_MASK
-    s_or_b32        ttmp11, ttmp11, ttmp3
-
-    s_andn2_b32     ttmp2, ttmp2, SQ_WAVE_IB_STS_RCNT_FIRST_REPLAY_MASK
-    s_setreg_b32    hwreg(HW_REG_IB_STS), ttmp2
+    save_and_clear_ib_sts(ttmp14)
 
     // Read second-level TBA/TMA from first-level TMA and jump if available.
     // ttmp[2:5] and ttmp12 can be used (others hold SPI-initialized debug data)
@@ -271,27 +299,53 @@ L_FETCH_2ND_TRAP:
     s_getreg_b32    ttmp14, hwreg(HW_REG_SQ_SHADER_TMA_LO)
     s_getreg_b32    ttmp15, hwreg(HW_REG_SQ_SHADER_TMA_HI)
     s_lshl_b64      [ttmp14, ttmp15], [ttmp14, ttmp15], 0x8
+
+    s_bitcmp1_b32   ttmp15, 0xF
+    s_cbranch_scc0  L_NO_SIGN_EXTEND_TMA
+    s_or_b32        ttmp15, ttmp15, 0xFFFF0000
+L_NO_SIGN_EXTEND_TMA:
+
+    s_load_dword    ttmp2, [ttmp14, ttmp15], 0x10 glc:1 // debug trap enabled flag
+    s_waitcnt       lgkmcnt(0)
+    s_lshl_b32      ttmp2, ttmp2, TTMP_DEBUG_TRAP_ENABLED_SHIFT
+    s_andn2_b32     s_save_ib_sts, s_save_ib_sts, TTMP_DEBUG_TRAP_ENABLED_MASK
+    s_or_b32        s_save_ib_sts, s_save_ib_sts, ttmp2
+
     s_load_dwordx2  [ttmp2, ttmp3], [ttmp14, ttmp15], 0x0 glc:1 // second-level TBA
     s_waitcnt       lgkmcnt(0)
     s_load_dwordx2  [ttmp14, ttmp15], [ttmp14, ttmp15], 0x8 glc:1 // second-level TMA
     s_waitcnt       lgkmcnt(0)
+
     s_and_b64       [ttmp2, ttmp3], [ttmp2, ttmp3], [ttmp2, ttmp3]
     s_cbranch_scc0  L_NO_NEXT_TRAP // second-level trap handler not been set
     s_setpc_b64     [ttmp2, ttmp3] // jump to second-level trap handler
 
 L_NO_NEXT_TRAP:
-    s_getreg_b32    s_save_trapsts, hwreg(HW_REG_TRAPSTS)
-    s_and_b32	    s_save_trapsts, s_save_trapsts, SQ_WAVE_TRAPSTS_EXCE_MASK // Check whether it is an exception
-    s_cbranch_scc1  L_EXCP_CASE	  // Exception, jump back to the shader program directly.
-    s_add_u32	    ttmp0, ttmp0, 4   // S_TRAP case, add 4 to ttmp0
-    s_addc_u32	ttmp1, ttmp1, 0
-L_EXCP_CASE:
+    // If not caused by trap then halt wave to prevent re-entry.
+    s_and_b32       ttmp2, s_save_pc_hi, (S_SAVE_PC_HI_TRAP_ID_MASK|S_SAVE_PC_HI_HT_MASK)
+    s_cbranch_scc1  L_TRAP_CASE
+    s_or_b32        s_save_status, s_save_status, SQ_WAVE_STATUS_HALT_MASK
+
+    // If the PC points to S_ENDPGM then context save will fail if STATUS.HALT is set.
+    // Rewind the PC to prevent this from occurring.
+    s_sub_u32       ttmp0, ttmp0, 0x8
+    s_subb_u32      ttmp1, ttmp1, 0x0
+
+    s_branch        L_EXIT_TRAP
+
+L_TRAP_CASE:
+    // Host trap will not cause trap re-entry.
+    s_and_b32       ttmp2, s_save_pc_hi, S_SAVE_PC_HI_HT_MASK
+    s_cbranch_scc1  L_EXIT_TRAP
+
+    // Advance past trap instruction to prevent re-entry.
+    s_add_u32       ttmp0, ttmp0, 0x4
+    s_addc_u32      ttmp1, ttmp1, 0x0
+
+L_EXIT_TRAP:
     s_and_b32	ttmp1, ttmp1, 0xFFFF
 
-    // Restore SQ_WAVE_IB_STS.
-    s_lshr_b32      ttmp2, ttmp11, (TTMP11_SAVE_RCNT_FIRST_REPLAY_SHIFT - SQ_WAVE_IB_STS_FIRST_REPLAY_SHIFT)
-    s_and_b32       ttmp2, ttmp2, SQ_WAVE_IB_STS_RCNT_FIRST_REPLAY_MASK
-    s_setreg_b32    hwreg(HW_REG_IB_STS), ttmp2
+    restore_ib_sts(ttmp14)
 
     // Restore SQ_WAVE_STATUS.
     s_and_b64       exec, exec, exec // Restore STATUS.EXECZ, not writable by s_setreg_b32
@@ -312,16 +366,7 @@ L_SAVE:
     s_mov_b32	    s_save_tmp, 0							    //clear saveCtx bit
     s_setreg_b32    hwreg(HW_REG_TRAPSTS, SQ_WAVE_TRAPSTS_SAVECTX_SHIFT, 1), s_save_tmp	    //clear saveCtx bit
 
-    s_getreg_b32    s_save_tmp, hwreg(HW_REG_IB_STS, SQ_WAVE_IB_STS_RCNT_SHIFT, SQ_WAVE_IB_STS_RCNT_SIZE)		    //save RCNT
-    s_lshl_b32	    s_save_tmp, s_save_tmp, S_SAVE_PC_HI_RCNT_SHIFT
-    s_or_b32	    s_save_pc_hi, s_save_pc_hi, s_save_tmp
-    s_getreg_b32    s_save_tmp, hwreg(HW_REG_IB_STS, SQ_WAVE_IB_STS_FIRST_REPLAY_SHIFT, SQ_WAVE_IB_STS_FIRST_REPLAY_SIZE)   //save FIRST_REPLAY
-    s_lshl_b32	    s_save_tmp, s_save_tmp, S_SAVE_PC_HI_FIRST_REPLAY_SHIFT
-    s_or_b32	    s_save_pc_hi, s_save_pc_hi, s_save_tmp
-    s_getreg_b32    s_save_tmp, hwreg(HW_REG_IB_STS)					    //clear RCNT and FIRST_REPLAY in IB_STS
-    s_and_b32	    s_save_tmp, s_save_tmp, SQ_WAVE_IB_STS_RCNT_FIRST_REPLAY_MASK_NEG
-
-    s_setreg_b32    hwreg(HW_REG_IB_STS), s_save_tmp
+    save_and_clear_ib_sts(s_save_tmp)
 
     /*	    inform SPI the readiness and wait for SPI's go signal */
     s_mov_b32	    s_save_exec_lo, exec_lo						    //save EXEC and use EXEC for the go signal from SPI
@@ -360,12 +405,6 @@ L_SAVE:
     s_or_b32	    s_save_buf_rsrc1,	s_save_buf_rsrc1,  S_SAVE_BUF_RSRC_WORD1_STRIDE
     s_mov_b32	    s_save_buf_rsrc2,	0									//NUM_RECORDS initial value = 0 (in bytes) although not neccessarily inited
     s_mov_b32	    s_save_buf_rsrc3,	S_SAVE_BUF_RSRC_WORD3_MISC
-    s_and_b32	    s_save_tmp,		s_save_spi_init_hi, S_SAVE_SPI_INIT_ATC_MASK
-    s_lshr_b32	    s_save_tmp,		s_save_tmp, (S_SAVE_SPI_INIT_ATC_SHIFT-SQ_BUF_RSRC_WORD1_ATC_SHIFT)	    //get ATC bit into position
-    s_or_b32	    s_save_buf_rsrc3,	s_save_buf_rsrc3,  s_save_tmp						//or ATC
-    s_and_b32	    s_save_tmp,		s_save_spi_init_hi, S_SAVE_SPI_INIT_MTYPE_MASK
-    s_lshr_b32	    s_save_tmp,		s_save_tmp, (S_SAVE_SPI_INIT_MTYPE_SHIFT-SQ_BUF_RSRC_WORD3_MTYPE_SHIFT)	    //get MTYPE bits into position
-    s_or_b32	    s_save_buf_rsrc3,	s_save_buf_rsrc3,  s_save_tmp						//or MTYPE
 
     //FIXME  right now s_save_m0/s_save_mem_offset use tma_lo/tma_hi  (might need to save them before using them?)
     s_mov_b32	    s_save_m0,		m0								    //save M0
@@ -408,7 +447,9 @@ L_SAVE:
     s_getreg_b32    s_save_m0, hwreg(HW_REG_MODE)						    //MODE
     write_hwreg_to_mem(s_save_m0, s_save_buf_rsrc0, s_save_mem_offset)
 
-
+    // Clear VSKIP state now that MODE.VSKIP has been saved.
+    // If user shader set it then vector instructions would be skipped.
+    s_setvskip	0,0
 
     /*	    the first wave in the threadgroup	 */
     s_and_b32	    s_save_tmp, s_save_spi_init_hi, S_SAVE_SPI_INIT_FIRST_WAVE_MASK	// extract fisrt wave bit
@@ -535,12 +576,21 @@ if SAVE_AFTER_XNACK_ERROR
 
 	v_lshlrev_b32 v2, 2, v3
 L_SAVE_LDS_LOOP_SQC:
+#if ASIC_FAMILY < CHIP_GC_9_5_0
 	ds_read2_b32 v[0:1], v2 offset0:0 offset1:0x40
 	s_waitcnt lgkmcnt(0)
-
 	write_vgprs_to_mem_with_sqc(v0, 2, s_save_buf_rsrc0, s_save_mem_offset)
 
 	v_add_u32 v2, 0x200, v2
+#else
+	// gfx950 needs to save in multiple of 256 bytes.
+	ds_read_b32 v0, v2
+	s_waitcnt lgkmcnt(0)
+	write_vgprs_to_mem_with_sqc(v0, 1, s_save_buf_rsrc0, s_save_mem_offset)
+
+	v_add_u32 v2, 0x100, v2
+#endif
+
 	v_cmp_lt_u32 vcc[0:1], v2, s_save_alloc_size
 	s_cbranch_vccnz L_SAVE_LDS_LOOP_SQC
 
@@ -559,11 +609,14 @@ end
 L_SAVE_LDS_LOOP_VECTOR:
       ds_read_b64 v[0:1], v2	//x =LDS[a], byte address
       s_waitcnt lgkmcnt(0)
-      buffer_store_dwordx2  v[0:1], v2, s_save_buf_rsrc0, s_save_mem_offset offen:1  glc:1  slc:1
+      buffer_store_dwordx2  v[0:1], v2, s_save_buf_rsrc0, s_save_mem_offset VMEM_MODIFIERS offen:1
 //	s_waitcnt vmcnt(0)
 //	v_add_u32 v2, vcc[0:1], v2, v3
       v_add_u32 v2, v2, v3
       v_cmp_lt_u32 vcc[0:1], v2, s_save_alloc_size
+#if ASIC_FAMILY >= CHIP_GC_9_5_0
+      s_mov_b64 exec, vcc
+#endif
       s_cbranch_vccnz L_SAVE_LDS_LOOP_VECTOR
 
       // restore rsrc3
@@ -690,12 +743,6 @@ L_RESTORE:
     s_or_b32	    s_restore_buf_rsrc1,    s_restore_buf_rsrc1,  S_RESTORE_BUF_RSRC_WORD1_STRIDE
     s_mov_b32	    s_restore_buf_rsrc2,    0										    //NUM_RECORDS initial value = 0 (in bytes)
     s_mov_b32	    s_restore_buf_rsrc3,    S_RESTORE_BUF_RSRC_WORD3_MISC
-    s_and_b32	    s_restore_tmp,	    s_restore_spi_init_hi, S_RESTORE_SPI_INIT_ATC_MASK
-    s_lshr_b32	    s_restore_tmp,	    s_restore_tmp, (S_RESTORE_SPI_INIT_ATC_SHIFT-SQ_BUF_RSRC_WORD1_ATC_SHIFT)	    //get ATC bit into position
-    s_or_b32	    s_restore_buf_rsrc3,    s_restore_buf_rsrc3,  s_restore_tmp						    //or ATC
-    s_and_b32	    s_restore_tmp,	    s_restore_spi_init_hi, S_RESTORE_SPI_INIT_MTYPE_MASK
-    s_lshr_b32	    s_restore_tmp,	    s_restore_tmp, (S_RESTORE_SPI_INIT_MTYPE_SHIFT-SQ_BUF_RSRC_WORD3_MTYPE_SHIFT)   //get MTYPE bits into position
-    s_or_b32	    s_restore_buf_rsrc3,    s_restore_buf_rsrc3,  s_restore_tmp						    //or MTYPE
 
     /*	    global mem offset		*/
 //  s_mov_b32	    s_restore_mem_offset, 0x0				    //mem offset initial value = 0
@@ -732,8 +779,13 @@ L_RESTORE:
   L_RESTORE_LDS_LOOP:
 	buffer_load_dword   v0, v0, s_restore_buf_rsrc0, s_restore_mem_offset lds:1		       // first 64DW
 	buffer_load_dword   v0, v0, s_restore_buf_rsrc0, s_restore_mem_offset lds:1 offset:256	       // second 64DW
-    s_add_u32	    m0, m0, 256*2						// 128 DW
-    s_add_u32	    s_restore_mem_offset, s_restore_mem_offset, 256*2		//mem offset increased by 128DW
+#if ASIC_FAMILY >= CHIP_GC_9_5_0
+	buffer_load_dword   v0, v0, s_restore_buf_rsrc0, s_restore_mem_offset lds:1 offset:512	// third 64DW
+	buffer_load_dword   v0, v0, s_restore_buf_rsrc0, s_restore_mem_offset lds:1 offset:768	// forth 64DW
+	buffer_load_dword   v0, v0, s_restore_buf_rsrc0, s_restore_mem_offset lds:1 offset:1024	// fifth 64DW
+#endif
+    s_add_u32	    m0, m0, LDS_RESTORE_GRANULARITY_BYTES					// 128/320 DW
+    s_add_u32	    s_restore_mem_offset, s_restore_mem_offset, LDS_RESTORE_GRANULARITY_BYTES	//mem offset increased by 128/320 DW
     s_cmp_lt_u32    m0, s_restore_alloc_size					//scc=(m0 < s_restore_alloc_size) ? 1 : 0
     s_cbranch_scc1  L_RESTORE_LDS_LOOP							    //LDS restore is complete?
 
@@ -889,19 +941,7 @@ L_RESTORE:
     s_load_dword    ttmp13, [s_restore_ttmps_lo, s_restore_ttmps_hi], 0x74 glc:1
     s_waitcnt	    lgkmcnt(0)
 
-    //reuse s_restore_m0 as a temp register
-    s_and_b32	    s_restore_m0, s_restore_pc_hi, S_SAVE_PC_HI_RCNT_MASK
-    s_lshr_b32	    s_restore_m0, s_restore_m0, S_SAVE_PC_HI_RCNT_SHIFT
-    s_lshl_b32	    s_restore_m0, s_restore_m0, SQ_WAVE_IB_STS_RCNT_SHIFT
-    s_mov_b32	    s_restore_tmp, 0x0										    //IB_STS is zero
-    s_or_b32	    s_restore_tmp, s_restore_tmp, s_restore_m0
-    s_and_b32	    s_restore_m0, s_restore_pc_hi, S_SAVE_PC_HI_FIRST_REPLAY_MASK
-    s_lshr_b32	    s_restore_m0, s_restore_m0, S_SAVE_PC_HI_FIRST_REPLAY_SHIFT
-    s_lshl_b32	    s_restore_m0, s_restore_m0, SQ_WAVE_IB_STS_FIRST_REPLAY_SHIFT
-    s_or_b32	    s_restore_tmp, s_restore_tmp, s_restore_m0
-    s_and_b32	    s_restore_m0, s_restore_status, SQ_WAVE_STATUS_INST_ATC_MASK
-    s_lshr_b32	    s_restore_m0, s_restore_m0, SQ_WAVE_STATUS_INST_ATC_SHIFT
-    s_setreg_b32    hwreg(HW_REG_IB_STS),   s_restore_tmp
+    restore_ib_sts(s_restore_tmp)
 
     s_and_b32 s_restore_pc_hi, s_restore_pc_hi, 0x0000ffff	//pc[47:32]	   //Do it here in order not to affect STATUS
     s_and_b64	 exec, exec, exec  // Restore STATUS.EXECZ, not writable by s_setreg_b32
@@ -910,15 +950,14 @@ L_RESTORE:
 
     s_barrier							//barrier to ensure the readiness of LDS before access attempts from any other wave in the same TG //FIXME not performance-optimal at this time
 
-//  s_rfe_b64 s_restore_pc_lo					//Return to the main shader program and resume execution
-    s_rfe_restore_b64  s_restore_pc_lo, s_restore_m0		// s_restore_m0[0] is used to set STATUS.inst_atc
+    s_rfe_b64 s_restore_pc_lo					//Return to the main shader program and resume execution
 
 
 /**************************************************************************/
 /*			the END						  */
 /**************************************************************************/
 L_END_PGM:
-    s_endpgm
+    s_endpgm_saved
 
 end
 
@@ -976,17 +1015,17 @@ L_TCP_STORE_CHECK_DONE:
 end
 
 function write_4vgprs_to_mem(s_rsrc, s_mem_offset)
-	buffer_store_dword v0, v0, s_rsrc, s_mem_offset slc:1 glc:1
-	buffer_store_dword v1, v0, s_rsrc, s_mem_offset slc:1 glc:1  offset:256
-	buffer_store_dword v2, v0, s_rsrc, s_mem_offset slc:1 glc:1  offset:256*2
-	buffer_store_dword v3, v0, s_rsrc, s_mem_offset slc:1 glc:1  offset:256*3
+	buffer_store_dword v0, v0, s_rsrc, s_mem_offset VMEM_MODIFIERS
+	buffer_store_dword v1, v0, s_rsrc, s_mem_offset VMEM_MODIFIERS offset:256
+	buffer_store_dword v2, v0, s_rsrc, s_mem_offset VMEM_MODIFIERS offset:256*2
+	buffer_store_dword v3, v0, s_rsrc, s_mem_offset VMEM_MODIFIERS offset:256*3
 end
 
 function read_4vgprs_from_mem(s_rsrc, s_mem_offset)
-	buffer_load_dword v0, v0, s_rsrc, s_mem_offset slc:1 glc:1
-	buffer_load_dword v1, v0, s_rsrc, s_mem_offset slc:1 glc:1 offset:256
-	buffer_load_dword v2, v0, s_rsrc, s_mem_offset slc:1 glc:1 offset:256*2
-	buffer_load_dword v3, v0, s_rsrc, s_mem_offset slc:1 glc:1 offset:256*3
+	buffer_load_dword v0, v0, s_rsrc, s_mem_offset VMEM_MODIFIERS
+	buffer_load_dword v1, v0, s_rsrc, s_mem_offset VMEM_MODIFIERS offset:256
+	buffer_load_dword v2, v0, s_rsrc, s_mem_offset VMEM_MODIFIERS offset:256*2
+	buffer_load_dword v3, v0, s_rsrc, s_mem_offset VMEM_MODIFIERS offset:256*3
 	s_waitcnt vmcnt(0)
 end
 
@@ -1078,3 +1117,19 @@ function set_status_without_spi_prio(status, tmp)
     s_nop           0x2 // avoid S_SETREG => S_SETREG hazard
     s_setreg_b32    hwreg(HW_REG_STATUS, SQ_WAVE_STATUS_PRE_SPI_PRIO_SHIFT, SQ_WAVE_STATUS_PRE_SPI_PRIO_SIZE), status
 end
+
+function save_and_clear_ib_sts(tmp)
+    // Save IB_STS.FIRST_REPLAY[15] and IB_STS.RCNT[20:16] into unused space s_save_ib_sts[31:26].
+    s_getreg_b32    tmp, hwreg(HW_REG_IB_STS)
+    s_and_b32       tmp, tmp, SQ_WAVE_IB_STS_RCNT_FIRST_REPLAY_MASK
+    s_lshl_b32      tmp, tmp, (TTMP_SAVE_RCNT_FIRST_REPLAY_SHIFT - SQ_WAVE_IB_STS_FIRST_REPLAY_SHIFT)
+    s_andn2_b32     s_save_ib_sts, s_save_ib_sts, TTMP_SAVE_RCNT_FIRST_REPLAY_MASK
+    s_or_b32        s_save_ib_sts, s_save_ib_sts, tmp
+    s_setreg_imm32_b32 hwreg(HW_REG_IB_STS), 0x0
+end
+
+function restore_ib_sts(tmp)
+    s_lshr_b32      tmp, s_save_ib_sts, (TTMP_SAVE_RCNT_FIRST_REPLAY_SHIFT - SQ_WAVE_IB_STS_FIRST_REPLAY_SHIFT)
+    s_and_b32       tmp, tmp, SQ_WAVE_IB_STS_RCNT_FIRST_REPLAY_MASK
+    s_setreg_b32    hwreg(HW_REG_IB_STS), tmp
+end
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
index 24ebd61395d8..22925df6a791 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
@@ -1,5 +1,6 @@
+// SPDX-License-Identifier: GPL-2.0 OR MIT
 /*
- * Copyright 2014 Advanced Micro Devices, Inc.
+ * Copyright 2014-2022 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -21,7 +22,6 @@
  */
 
 #include <linux/device.h>
-#include <linux/export.h>
 #include <linux/err.h>
 #include <linux/fs.h>
 #include <linux/file.h>
@@ -33,14 +33,16 @@
 #include <linux/time.h>
 #include <linux/mm.h>
 #include <linux/mman.h>
+#include <linux/ptrace.h>
 #include <linux/dma-buf.h>
-#include <asm/processor.h>
+#include <linux/processor.h>
 #include "kfd_priv.h"
 #include "kfd_device_queue_manager.h"
-#include "kfd_dbgmgr.h"
 #include "kfd_svm.h"
 #include "amdgpu_amdkfd.h"
 #include "kfd_smi_events.h"
+#include "amdgpu_dma_buf.h"
+#include "kfd_debug.h"
 
 static long kfd_ioctl(struct file *, unsigned int, unsigned long);
 static int kfd_open(struct inode *, struct file *);
@@ -59,8 +61,29 @@ static const struct file_operations kfd_fops = {
 };
 
 static int kfd_char_dev_major = -1;
-static struct class *kfd_class;
 struct device *kfd_device;
+static const struct class kfd_class = {
+	.name = kfd_dev_name,
+};
+
+static inline struct kfd_process_device *kfd_lock_pdd_by_id(struct kfd_process *p, __u32 gpu_id)
+{
+	struct kfd_process_device *pdd;
+
+	mutex_lock(&p->mutex);
+	pdd = kfd_process_device_data_by_id(p, gpu_id);
+
+	if (pdd)
+		return pdd;
+
+	mutex_unlock(&p->mutex);
+	return NULL;
+}
+
+static inline void kfd_unlock_pdd(struct kfd_process_device *pdd)
+{
+	mutex_unlock(&pdd->process->mutex);
+}
 
 int kfd_chardev_init(void)
 {
@@ -71,14 +94,13 @@ int kfd_chardev_init(void)
 	if (err < 0)
 		goto err_register_chrdev;
 
-	kfd_class = class_create(THIS_MODULE, kfd_dev_name);
-	err = PTR_ERR(kfd_class);
-	if (IS_ERR(kfd_class))
+	err = class_register(&kfd_class);
+	if (err)
 		goto err_class_create;
 
-	kfd_device = device_create(kfd_class, NULL,
-					MKDEV(kfd_char_dev_major, 0),
-					NULL, kfd_dev_name);
+	kfd_device = device_create(&kfd_class, NULL,
+				   MKDEV(kfd_char_dev_major, 0),
+				   NULL, kfd_dev_name);
 	err = PTR_ERR(kfd_device);
 	if (IS_ERR(kfd_device))
 		goto err_device_create;
@@ -86,7 +108,7 @@ int kfd_chardev_init(void)
 	return 0;
 
 err_device_create:
-	class_destroy(kfd_class);
+	class_unregister(&kfd_class);
 err_class_create:
 	unregister_chrdev(kfd_char_dev_major, kfd_dev_name);
 err_register_chrdev:
@@ -95,17 +117,12 @@ err_register_chrdev:
 
 void kfd_chardev_exit(void)
 {
-	device_destroy(kfd_class, MKDEV(kfd_char_dev_major, 0));
-	class_destroy(kfd_class);
+	device_destroy(&kfd_class, MKDEV(kfd_char_dev_major, 0));
+	class_unregister(&kfd_class);
 	unregister_chrdev(kfd_char_dev_major, kfd_dev_name);
 	kfd_device = NULL;
 }
 
-struct device *kfd_chardev(void)
-{
-	return kfd_device;
-}
-
 
 static int kfd_open(struct inode *inode, struct file *filep)
 {
@@ -125,22 +142,20 @@ static int kfd_open(struct inode *inode, struct file *filep)
 		return -EPERM;
 	}
 
-	process = kfd_create_process(filep);
+	process = kfd_create_process(current);
 	if (IS_ERR(process))
 		return PTR_ERR(process);
 
-	if (kfd_is_locked()) {
-		dev_dbg(kfd_device, "kfd is locked!\n"
-				"process %d unreferenced", process->pasid);
+	if (kfd_process_init_cwsr_apu(process, filep)) {
 		kfd_unref_process(process);
-		return -EAGAIN;
+		return -EFAULT;
 	}
 
 	/* filep now owns the reference returned by kfd_create_process */
 	filep->private_data = process;
 
-	dev_dbg(kfd_device, "process %d opened, compat mode (32 bit) - %d\n",
-		process->pasid, process->is_32bit_user_mode);
+	dev_dbg(kfd_device, "process pid %d opened kfd node, compat mode (32 bit) - %d\n",
+		process->lead_thread->pid, process->is_32bit_user_mode);
 
 	return 0;
 }
@@ -169,7 +184,12 @@ static int kfd_ioctl_get_version(struct file *filep, struct kfd_process *p,
 static int set_queue_properties_from_user(struct queue_properties *q_properties,
 				struct kfd_ioctl_create_queue_args *args)
 {
-	if (args->queue_percentage > KFD_MAX_QUEUE_PERCENTAGE) {
+	/*
+	 * Repurpose queue percentage to accommodate new features:
+	 * bit 0-7: queue percentage
+	 * bit 8-15: pm4_target_xcc
+	 */
+	if ((args->queue_percentage & 0xFF) > KFD_MAX_QUEUE_PERCENTAGE) {
 		pr_err("Queue percentage must be between 0 to KFD_MAX_QUEUE_PERCENTAGE\n");
 		return -EINVAL;
 	}
@@ -191,6 +211,11 @@ static int set_queue_properties_from_user(struct queue_properties *q_properties,
 		return -EINVAL;
 	}
 
+	if (args->ring_size < KFD_MIN_QUEUE_RING_SIZE) {
+		args->ring_size = KFD_MIN_QUEUE_RING_SIZE;
+		pr_debug("Size lower. clamped to KFD_MIN_QUEUE_RING_SIZE");
+	}
+
 	if (!access_ok((const void __user *) args->read_pointer_address,
 			sizeof(uint32_t))) {
 		pr_err("Can't access read pointer\n");
@@ -219,18 +244,21 @@ static int set_queue_properties_from_user(struct queue_properties *q_properties,
 
 	q_properties->is_interop = false;
 	q_properties->is_gws = false;
-	q_properties->queue_percent = args->queue_percentage;
+	q_properties->queue_percent = args->queue_percentage & 0xFF;
+	/* bit 8-15 are repurposed to be PM4 target XCC */
+	q_properties->pm4_target_xcc = (args->queue_percentage >> 8) & 0xFF;
 	q_properties->priority = args->queue_priority;
 	q_properties->queue_address = args->ring_base_address;
 	q_properties->queue_size = args->ring_size;
-	q_properties->read_ptr = (uint32_t *) args->read_pointer_address;
-	q_properties->write_ptr = (uint32_t *) args->write_pointer_address;
+	q_properties->read_ptr = (void __user *)args->read_pointer_address;
+	q_properties->write_ptr = (void __user *)args->write_pointer_address;
 	q_properties->eop_ring_buffer_address = args->eop_buffer_address;
 	q_properties->eop_ring_buffer_size = args->eop_buffer_size;
 	q_properties->ctx_save_restore_area_address =
 			args->ctx_save_restore_address;
 	q_properties->ctx_save_restore_area_size = args->ctx_save_restore_size;
 	q_properties->ctl_stack_size = args->ctl_stack_size;
+	q_properties->sdma_engine_id = args->sdma_engine_id;
 	if (args->queue_type == KFD_IOC_QUEUE_TYPE_COMPUTE ||
 		args->queue_type == KFD_IOC_QUEUE_TYPE_COMPUTE_AQL)
 		q_properties->type = KFD_QUEUE_TYPE_COMPUTE;
@@ -238,6 +266,8 @@ static int set_queue_properties_from_user(struct queue_properties *q_properties,
 		q_properties->type = KFD_QUEUE_TYPE_SDMA;
 	else if (args->queue_type == KFD_IOC_QUEUE_TYPE_SDMA_XGMI)
 		q_properties->type = KFD_QUEUE_TYPE_SDMA_XGMI;
+	else if (args->queue_type == KFD_IOC_QUEUE_TYPE_SDMA_BY_ENG_ID)
+		q_properties->type = KFD_QUEUE_TYPE_SDMA_BY_ENG_ID;
 	else
 		return -ENOTSUPP;
 
@@ -276,7 +306,7 @@ static int kfd_ioctl_create_queue(struct file *filep, struct kfd_process *p,
 					void *data)
 {
 	struct kfd_ioctl_create_queue_args *args = data;
-	struct kfd_dev *dev;
+	struct kfd_node *dev;
 	int err = 0;
 	unsigned int queue_id;
 	struct kfd_process_device *pdd;
@@ -292,26 +322,55 @@ static int kfd_ioctl_create_queue(struct file *filep, struct kfd_process *p,
 		return err;
 
 	pr_debug("Looking for gpu id 0x%x\n", args->gpu_id);
-	dev = kfd_device_by_id(args->gpu_id);
-	if (!dev) {
-		pr_debug("Could not find gpu id 0x%x\n", args->gpu_id);
-		return -EINVAL;
-	}
 
 	mutex_lock(&p->mutex);
 
+	pdd = kfd_process_device_data_by_id(p, args->gpu_id);
+	if (!pdd) {
+		pr_debug("Could not find gpu id 0x%x\n", args->gpu_id);
+		err = -EINVAL;
+		goto err_pdd;
+	}
+	dev = pdd->dev;
+
 	pdd = kfd_bind_process_to_device(dev, p);
 	if (IS_ERR(pdd)) {
 		err = -ESRCH;
 		goto err_bind_process;
 	}
 
-	pr_debug("Creating queue for PASID 0x%x on gpu 0x%x\n",
-			p->pasid,
+	if (q_properties.type == KFD_QUEUE_TYPE_SDMA_BY_ENG_ID) {
+		int max_sdma_eng_id = kfd_get_num_sdma_engines(dev) +
+				      kfd_get_num_xgmi_sdma_engines(dev) - 1;
+
+		if (q_properties.sdma_engine_id > max_sdma_eng_id) {
+			err = -EINVAL;
+			pr_err("sdma_engine_id %i exceeds maximum id of %i\n",
+			       q_properties.sdma_engine_id, max_sdma_eng_id);
+			goto err_sdma_engine_id;
+		}
+	}
+
+	if (!pdd->qpd.proc_doorbells) {
+		err = kfd_alloc_process_doorbells(dev->kfd, pdd);
+		if (err) {
+			pr_debug("failed to allocate process doorbells\n");
+			goto err_bind_process;
+		}
+	}
+
+	err = kfd_queue_acquire_buffers(pdd, &q_properties);
+	if (err) {
+		pr_debug("failed to acquire user queue buffers\n");
+		goto err_acquire_queue_buf;
+	}
+
+	pr_debug("Creating queue for process pid %d on gpu 0x%x\n",
+			p->lead_thread->pid,
 			dev->id);
 
-	err = pqm_create_queue(&p->pqm, dev, filep, &q_properties, &queue_id,
-			&doorbell_offset_in_process);
+	err = pqm_create_queue(&p->pqm, dev, &q_properties, &queue_id,
+			NULL, NULL, NULL, &doorbell_offset_in_process);
 	if (err != 0)
 		goto err_create_queue;
 
@@ -321,7 +380,7 @@ static int kfd_ioctl_create_queue(struct file *filep, struct kfd_process *p,
 	/* Return gpu_id as doorbell offset for mmap usage */
 	args->doorbell_offset = KFD_MMAP_TYPE_DOORBELL;
 	args->doorbell_offset |= KFD_MMAP_GPU_ID(args->gpu_id);
-	if (KFD_IS_SOC15(dev->device_info->asic_family))
+	if (KFD_IS_SOC15(dev))
 		/* On SOC15 ASICs, include the doorbell offset within the
 		 * process doorbell frame, which is 2 pages.
 		 */
@@ -340,10 +399,16 @@ static int kfd_ioctl_create_queue(struct file *filep, struct kfd_process *p,
 	pr_debug("Write ptr address   == 0x%016llX\n",
 			args->write_pointer_address);
 
+	kfd_dbg_ev_raise(KFD_EC_MASK(EC_QUEUE_NEW), p, dev, queue_id, false, NULL, 0);
 	return 0;
 
 err_create_queue:
+	kfd_queue_unref_bo_vas(pdd, &q_properties);
+	kfd_queue_release_buffers(pdd, &q_properties);
+err_acquire_queue_buf:
+err_sdma_engine_id:
 err_bind_process:
+err_pdd:
 	mutex_unlock(&p->mutex);
 	return err;
 }
@@ -354,9 +419,9 @@ static int kfd_ioctl_destroy_queue(struct file *filp, struct kfd_process *p,
 	int retval;
 	struct kfd_ioctl_destroy_queue_args *args = data;
 
-	pr_debug("Destroying queue id %d for pasid 0x%x\n",
+	pr_debug("Destroying queue id %d for process pid %d\n",
 				args->queue_id,
-				p->pasid);
+				p->lead_thread->pid);
 
 	mutex_lock(&p->mutex);
 
@@ -373,7 +438,12 @@ static int kfd_ioctl_update_queue(struct file *filp, struct kfd_process *p,
 	struct kfd_ioctl_update_queue_args *args = data;
 	struct queue_properties properties;
 
-	if (args->queue_percentage > KFD_MAX_QUEUE_PERCENTAGE) {
+	/*
+	 * Repurpose queue percentage to accommodate new features:
+	 * bit 0-7: queue percentage
+	 * bit 8-15: pm4_target_xcc
+	 */
+	if ((args->queue_percentage & 0xFF) > KFD_MAX_QUEUE_PERCENTAGE) {
 		pr_err("Queue percentage must be between 0 to KFD_MAX_QUEUE_PERCENTAGE\n");
 		return -EINVAL;
 	}
@@ -395,13 +465,20 @@ static int kfd_ioctl_update_queue(struct file *filp, struct kfd_process *p,
 		return -EINVAL;
 	}
 
+	if (args->ring_size < KFD_MIN_QUEUE_RING_SIZE) {
+		args->ring_size = KFD_MIN_QUEUE_RING_SIZE;
+		pr_debug("Size lower. clamped to KFD_MIN_QUEUE_RING_SIZE");
+	}
+
 	properties.queue_address = args->ring_base_address;
 	properties.queue_size = args->ring_size;
-	properties.queue_percent = args->queue_percentage;
+	properties.queue_percent = args->queue_percentage & 0xFF;
+	/* bit 8-15 are repurposed to be PM4 target XCC */
+	properties.pm4_target_xcc = (args->queue_percentage >> 8) & 0xFF;
 	properties.priority = args->queue_priority;
 
-	pr_debug("Updating queue id %d for pasid 0x%x\n",
-			args->queue_id, p->pasid);
+	pr_debug("Updating queue id %d for process pid %d\n",
+			args->queue_id, p->lead_thread->pid);
 
 	mutex_lock(&p->mutex);
 
@@ -444,26 +521,18 @@ static int kfd_ioctl_set_cu_mask(struct file *filp, struct kfd_process *p,
 		cu_mask_size = sizeof(uint32_t) * (max_num_cus/32);
 	}
 
-	minfo.cu_mask.ptr = kzalloc(cu_mask_size, GFP_KERNEL);
-	if (!minfo.cu_mask.ptr)
-		return -ENOMEM;
-
-	retval = copy_from_user(minfo.cu_mask.ptr, cu_mask_ptr, cu_mask_size);
-	if (retval) {
+	minfo.cu_mask.ptr = memdup_user(cu_mask_ptr, cu_mask_size);
+	if (IS_ERR(minfo.cu_mask.ptr)) {
 		pr_debug("Could not copy CU mask from userspace");
-		retval = -EFAULT;
-		goto out;
+		return PTR_ERR(minfo.cu_mask.ptr);
 	}
 
-	minfo.update_flag = UPDATE_FLAG_CU_MASK;
-
 	mutex_lock(&p->mutex);
 
 	retval = pqm_update_mqd(&p->pqm, args->queue_id, &minfo);
 
 	mutex_unlock(&p->mutex);
 
-out:
 	kfree(minfo.cu_mask.ptr);
 	return retval;
 }
@@ -490,7 +559,6 @@ static int kfd_ioctl_set_memory_policy(struct file *filep,
 					struct kfd_process *p, void *data)
 {
 	struct kfd_ioctl_set_memory_policy_args *args = data;
-	struct kfd_dev *dev;
 	int err = 0;
 	struct kfd_process_device *pdd;
 	enum cache_policy default_policy, alternate_policy;
@@ -505,13 +573,15 @@ static int kfd_ioctl_set_memory_policy(struct file *filep,
 		return -EINVAL;
 	}
 
-	dev = kfd_device_by_id(args->gpu_id);
-	if (!dev)
-		return -EINVAL;
-
 	mutex_lock(&p->mutex);
+	pdd = kfd_process_device_data_by_id(p, args->gpu_id);
+	if (!pdd) {
+		pr_debug("Could not find gpu id 0x%x\n", args->gpu_id);
+		err = -EINVAL;
+		goto err_pdd;
+	}
 
-	pdd = kfd_bind_process_to_device(dev, p);
+	pdd = kfd_bind_process_to_device(pdd->dev, p);
 	if (IS_ERR(pdd)) {
 		err = -ESRCH;
 		goto out;
@@ -524,15 +594,17 @@ static int kfd_ioctl_set_memory_policy(struct file *filep,
 		(args->alternate_policy == KFD_IOC_CACHE_POLICY_COHERENT)
 		   ? cache_policy_coherent : cache_policy_noncoherent;
 
-	if (!dev->dqm->ops.set_cache_memory_policy(dev->dqm,
+	if (!pdd->dev->dqm->ops.set_cache_memory_policy(pdd->dev->dqm,
 				&pdd->qpd,
 				default_policy,
 				alternate_policy,
 				(void __user *)args->alternate_aperture_base,
-				args->alternate_aperture_size))
+				args->alternate_aperture_size,
+				args->misc_process_flag))
 		err = -EINVAL;
 
 out:
+err_pdd:
 	mutex_unlock(&p->mutex);
 
 	return err;
@@ -542,17 +614,18 @@ static int kfd_ioctl_set_trap_handler(struct file *filep,
 					struct kfd_process *p, void *data)
 {
 	struct kfd_ioctl_set_trap_handler_args *args = data;
-	struct kfd_dev *dev;
 	int err = 0;
 	struct kfd_process_device *pdd;
 
-	dev = kfd_device_by_id(args->gpu_id);
-	if (!dev)
-		return -EINVAL;
-
 	mutex_lock(&p->mutex);
 
-	pdd = kfd_bind_process_to_device(dev, p);
+	pdd = kfd_process_device_data_by_id(p, args->gpu_id);
+	if (!pdd) {
+		err = -EINVAL;
+		goto err_pdd;
+	}
+
+	pdd = kfd_bind_process_to_device(pdd->dev, p);
 	if (IS_ERR(pdd)) {
 		err = -ESRCH;
 		goto out;
@@ -561,6 +634,7 @@ static int kfd_ioctl_set_trap_handler(struct file *filep,
 	kfd_process_set_trap_handler(&pdd->qpd, args->tba_addr, args->tma_addr);
 
 out:
+err_pdd:
 	mutex_unlock(&p->mutex);
 
 	return err;
@@ -569,289 +643,40 @@ out:
 static int kfd_ioctl_dbg_register(struct file *filep,
 				struct kfd_process *p, void *data)
 {
-	struct kfd_ioctl_dbg_register_args *args = data;
-	struct kfd_dev *dev;
-	struct kfd_dbgmgr *dbgmgr_ptr;
-	struct kfd_process_device *pdd;
-	bool create_ok;
-	long status = 0;
-
-	dev = kfd_device_by_id(args->gpu_id);
-	if (!dev)
-		return -EINVAL;
-
-	if (dev->device_info->asic_family == CHIP_CARRIZO) {
-		pr_debug("kfd_ioctl_dbg_register not supported on CZ\n");
-		return -EINVAL;
-	}
-
-	mutex_lock(&p->mutex);
-	mutex_lock(kfd_get_dbgmgr_mutex());
-
-	/*
-	 * make sure that we have pdd, if this the first queue created for
-	 * this process
-	 */
-	pdd = kfd_bind_process_to_device(dev, p);
-	if (IS_ERR(pdd)) {
-		status = PTR_ERR(pdd);
-		goto out;
-	}
-
-	if (!dev->dbgmgr) {
-		/* In case of a legal call, we have no dbgmgr yet */
-		create_ok = kfd_dbgmgr_create(&dbgmgr_ptr, dev);
-		if (create_ok) {
-			status = kfd_dbgmgr_register(dbgmgr_ptr, p);
-			if (status != 0)
-				kfd_dbgmgr_destroy(dbgmgr_ptr);
-			else
-				dev->dbgmgr = dbgmgr_ptr;
-		}
-	} else {
-		pr_debug("debugger already registered\n");
-		status = -EINVAL;
-	}
-
-out:
-	mutex_unlock(kfd_get_dbgmgr_mutex());
-	mutex_unlock(&p->mutex);
-
-	return status;
+	return -EPERM;
 }
 
 static int kfd_ioctl_dbg_unregister(struct file *filep,
 				struct kfd_process *p, void *data)
 {
-	struct kfd_ioctl_dbg_unregister_args *args = data;
-	struct kfd_dev *dev;
-	long status;
-
-	dev = kfd_device_by_id(args->gpu_id);
-	if (!dev || !dev->dbgmgr)
-		return -EINVAL;
-
-	if (dev->device_info->asic_family == CHIP_CARRIZO) {
-		pr_debug("kfd_ioctl_dbg_unregister not supported on CZ\n");
-		return -EINVAL;
-	}
-
-	mutex_lock(kfd_get_dbgmgr_mutex());
-
-	status = kfd_dbgmgr_unregister(dev->dbgmgr, p);
-	if (!status) {
-		kfd_dbgmgr_destroy(dev->dbgmgr);
-		dev->dbgmgr = NULL;
-	}
-
-	mutex_unlock(kfd_get_dbgmgr_mutex());
-
-	return status;
+	return -EPERM;
 }
 
-/*
- * Parse and generate variable size data structure for address watch.
- * Total size of the buffer and # watch points is limited in order
- * to prevent kernel abuse. (no bearing to the much smaller HW limitation
- * which is enforced by dbgdev module)
- * please also note that the watch address itself are not "copied from user",
- * since it be set into the HW in user mode values.
- *
- */
 static int kfd_ioctl_dbg_address_watch(struct file *filep,
 					struct kfd_process *p, void *data)
 {
-	struct kfd_ioctl_dbg_address_watch_args *args = data;
-	struct kfd_dev *dev;
-	struct dbg_address_watch_info aw_info;
-	unsigned char *args_buff;
-	long status;
-	void __user *cmd_from_user;
-	uint64_t watch_mask_value = 0;
-	unsigned int args_idx = 0;
-
-	memset((void *) &aw_info, 0, sizeof(struct dbg_address_watch_info));
-
-	dev = kfd_device_by_id(args->gpu_id);
-	if (!dev)
-		return -EINVAL;
-
-	if (dev->device_info->asic_family == CHIP_CARRIZO) {
-		pr_debug("kfd_ioctl_dbg_wave_control not supported on CZ\n");
-		return -EINVAL;
-	}
-
-	cmd_from_user = (void __user *) args->content_ptr;
-
-	/* Validate arguments */
-
-	if ((args->buf_size_in_bytes > MAX_ALLOWED_AW_BUFF_SIZE) ||
-		(args->buf_size_in_bytes <= sizeof(*args) + sizeof(int) * 2) ||
-		(cmd_from_user == NULL))
-		return -EINVAL;
-
-	/* this is the actual buffer to work with */
-	args_buff = memdup_user(cmd_from_user,
-				args->buf_size_in_bytes - sizeof(*args));
-	if (IS_ERR(args_buff))
-		return PTR_ERR(args_buff);
-
-	aw_info.process = p;
-
-	aw_info.num_watch_points = *((uint32_t *)(&args_buff[args_idx]));
-	args_idx += sizeof(aw_info.num_watch_points);
-
-	aw_info.watch_mode = (enum HSA_DBG_WATCH_MODE *) &args_buff[args_idx];
-	args_idx += sizeof(enum HSA_DBG_WATCH_MODE) * aw_info.num_watch_points;
-
-	/*
-	 * set watch address base pointer to point on the array base
-	 * within args_buff
-	 */
-	aw_info.watch_address = (uint64_t *) &args_buff[args_idx];
-
-	/* skip over the addresses buffer */
-	args_idx += sizeof(aw_info.watch_address) * aw_info.num_watch_points;
-
-	if (args_idx >= args->buf_size_in_bytes - sizeof(*args)) {
-		status = -EINVAL;
-		goto out;
-	}
-
-	watch_mask_value = (uint64_t) args_buff[args_idx];
-
-	if (watch_mask_value > 0) {
-		/*
-		 * There is an array of masks.
-		 * set watch mask base pointer to point on the array base
-		 * within args_buff
-		 */
-		aw_info.watch_mask = (uint64_t *) &args_buff[args_idx];
-
-		/* skip over the masks buffer */
-		args_idx += sizeof(aw_info.watch_mask) *
-				aw_info.num_watch_points;
-	} else {
-		/* just the NULL mask, set to NULL and skip over it */
-		aw_info.watch_mask = NULL;
-		args_idx += sizeof(aw_info.watch_mask);
-	}
-
-	if (args_idx >= args->buf_size_in_bytes - sizeof(args)) {
-		status = -EINVAL;
-		goto out;
-	}
-
-	/* Currently HSA Event is not supported for DBG */
-	aw_info.watch_event = NULL;
-
-	mutex_lock(kfd_get_dbgmgr_mutex());
-
-	status = kfd_dbgmgr_address_watch(dev->dbgmgr, &aw_info);
-
-	mutex_unlock(kfd_get_dbgmgr_mutex());
-
-out:
-	kfree(args_buff);
-
-	return status;
+	return -EPERM;
 }
 
 /* Parse and generate fixed size data structure for wave control */
 static int kfd_ioctl_dbg_wave_control(struct file *filep,
 					struct kfd_process *p, void *data)
 {
-	struct kfd_ioctl_dbg_wave_control_args *args = data;
-	struct kfd_dev *dev;
-	struct dbg_wave_control_info wac_info;
-	unsigned char *args_buff;
-	uint32_t computed_buff_size;
-	long status;
-	void __user *cmd_from_user;
-	unsigned int args_idx = 0;
-
-	memset((void *) &wac_info, 0, sizeof(struct dbg_wave_control_info));
-
-	/* we use compact form, independent of the packing attribute value */
-	computed_buff_size = sizeof(*args) +
-				sizeof(wac_info.mode) +
-				sizeof(wac_info.operand) +
-				sizeof(wac_info.dbgWave_msg.DbgWaveMsg) +
-				sizeof(wac_info.dbgWave_msg.MemoryVA) +
-				sizeof(wac_info.trapId);
-
-	dev = kfd_device_by_id(args->gpu_id);
-	if (!dev)
-		return -EINVAL;
-
-	if (dev->device_info->asic_family == CHIP_CARRIZO) {
-		pr_debug("kfd_ioctl_dbg_wave_control not supported on CZ\n");
-		return -EINVAL;
-	}
-
-	/* input size must match the computed "compact" size */
-	if (args->buf_size_in_bytes != computed_buff_size) {
-		pr_debug("size mismatch, computed : actual %u : %u\n",
-				args->buf_size_in_bytes, computed_buff_size);
-		return -EINVAL;
-	}
-
-	cmd_from_user = (void __user *) args->content_ptr;
-
-	if (cmd_from_user == NULL)
-		return -EINVAL;
-
-	/* copy the entire buffer from user */
-
-	args_buff = memdup_user(cmd_from_user,
-				args->buf_size_in_bytes - sizeof(*args));
-	if (IS_ERR(args_buff))
-		return PTR_ERR(args_buff);
-
-	/* move ptr to the start of the "pay-load" area */
-	wac_info.process = p;
-
-	wac_info.operand = *((enum HSA_DBG_WAVEOP *)(&args_buff[args_idx]));
-	args_idx += sizeof(wac_info.operand);
-
-	wac_info.mode = *((enum HSA_DBG_WAVEMODE *)(&args_buff[args_idx]));
-	args_idx += sizeof(wac_info.mode);
-
-	wac_info.trapId = *((uint32_t *)(&args_buff[args_idx]));
-	args_idx += sizeof(wac_info.trapId);
-
-	wac_info.dbgWave_msg.DbgWaveMsg.WaveMsgInfoGen2.Value =
-					*((uint32_t *)(&args_buff[args_idx]));
-	wac_info.dbgWave_msg.MemoryVA = NULL;
-
-	mutex_lock(kfd_get_dbgmgr_mutex());
-
-	pr_debug("Calling dbg manager process %p, operand %u, mode %u, trapId %u, message %u\n",
-			wac_info.process, wac_info.operand,
-			wac_info.mode, wac_info.trapId,
-			wac_info.dbgWave_msg.DbgWaveMsg.WaveMsgInfoGen2.Value);
-
-	status = kfd_dbgmgr_wave_control(dev->dbgmgr, &wac_info);
-
-	pr_debug("Returned status of dbg manager is %ld\n", status);
-
-	mutex_unlock(kfd_get_dbgmgr_mutex());
-
-	kfree(args_buff);
-
-	return status;
+	return -EPERM;
 }
 
 static int kfd_ioctl_get_clock_counters(struct file *filep,
 				struct kfd_process *p, void *data)
 {
 	struct kfd_ioctl_get_clock_counters_args *args = data;
-	struct kfd_dev *dev;
+	struct kfd_process_device *pdd;
 
-	dev = kfd_device_by_id(args->gpu_id);
-	if (dev)
+	mutex_lock(&p->mutex);
+	pdd = kfd_process_device_data_by_id(p, args->gpu_id);
+	mutex_unlock(&p->mutex);
+	if (pdd)
 		/* Reading GPU clock counter from KGD */
-		args->gpu_clock_counter = amdgpu_amdkfd_get_gpu_clock_counter(dev->kgd);
+		args->gpu_clock_counter = amdgpu_amdkfd_get_gpu_clock_counter(pdd->dev->adev);
 	else
 		/* Node without GPU resource */
 		args->gpu_clock_counter = 0;
@@ -874,7 +699,7 @@ static int kfd_ioctl_get_process_apertures(struct file *filp,
 	struct kfd_process_device_apertures *pAperture;
 	int i;
 
-	dev_dbg(kfd_device, "get apertures for PASID 0x%x", p->pasid);
+	dev_dbg(kfd_device, "get apertures for process pid %d", p->lead_thread->pid);
 
 	args->num_of_nodes = 0;
 
@@ -926,7 +751,8 @@ static int kfd_ioctl_get_process_apertures_new(struct file *filp,
 	int ret;
 	int i;
 
-	dev_dbg(kfd_device, "get apertures for PASID 0x%x", p->pasid);
+	dev_dbg(kfd_device, "get apertures for process pid %d",
+			p->lead_thread->pid);
 
 	if (args->num_of_nodes == 0) {
 		/* Return number of nodes, so that user space can alloacate
@@ -941,8 +767,8 @@ static int kfd_ioctl_get_process_apertures_new(struct file *filp,
 	 * nodes, but not more than args->num_of_nodes as that is
 	 * the amount of memory allocated by user
 	 */
-	pa = kzalloc((sizeof(struct kfd_process_device_apertures) *
-				args->num_of_nodes), GFP_KERNEL);
+	pa = kcalloc(args->num_of_nodes, sizeof(struct kfd_process_device_apertures),
+		     GFP_KERNEL);
 	if (!pa)
 		return -ENOMEM;
 
@@ -1007,57 +833,11 @@ static int kfd_ioctl_create_event(struct file *filp, struct kfd_process *p,
 	 * through the event_page_offset field.
 	 */
 	if (args->event_page_offset) {
-		struct kfd_dev *kfd;
-		struct kfd_process_device *pdd;
-		void *mem, *kern_addr;
-		uint64_t size;
-
-		kfd = kfd_device_by_id(GET_GPU_ID(args->event_page_offset));
-		if (!kfd) {
-			pr_err("Getting device by id failed in %s\n", __func__);
-			return -EINVAL;
-		}
-
 		mutex_lock(&p->mutex);
-
-		if (p->signal_page) {
-			pr_err("Event page is already set\n");
-			err = -EINVAL;
-			goto out_unlock;
-		}
-
-		pdd = kfd_bind_process_to_device(kfd, p);
-		if (IS_ERR(pdd)) {
-			err = PTR_ERR(pdd);
-			goto out_unlock;
-		}
-
-		mem = kfd_process_device_translate_handle(pdd,
-				GET_IDR_HANDLE(args->event_page_offset));
-		if (!mem) {
-			pr_err("Can't find BO, offset is 0x%llx\n",
-			       args->event_page_offset);
-			err = -EINVAL;
-			goto out_unlock;
-		}
-
-		err = amdgpu_amdkfd_gpuvm_map_gtt_bo_to_kernel(kfd->kgd,
-						mem, &kern_addr, &size);
-		if (err) {
-			pr_err("Failed to map event page to kernel\n");
-			goto out_unlock;
-		}
-
-		err = kfd_event_page_set(p, kern_addr, size);
-		if (err) {
-			pr_err("Failed to set event page\n");
-			amdgpu_amdkfd_gpuvm_unmap_gtt_bo_from_kernel(kfd->kgd, mem);
-			goto out_unlock;
-		}
-
-		p->signal_handle = args->event_page_offset;
-
+		err = kfd_kmap_event_page(p, args->event_page_offset);
 		mutex_unlock(&p->mutex);
+		if (err)
+			return err;
 	}
 
 	err = kfd_event_create(filp, p, args->event_type,
@@ -1066,10 +846,7 @@ static int kfd_ioctl_create_event(struct file *filp, struct kfd_process *p,
 				&args->event_page_offset,
 				&args->event_slot_index);
 
-	return err;
-
-out_unlock:
-	mutex_unlock(&p->mutex);
+	pr_debug("Created event (id:0x%08x) (%s)\n", args->event_id, __func__);
 	return err;
 }
 
@@ -1101,28 +878,27 @@ static int kfd_ioctl_wait_events(struct file *filp, struct kfd_process *p,
 				void *data)
 {
 	struct kfd_ioctl_wait_events_args *args = data;
-	int err;
 
-	err = kfd_wait_on_events(p, args->num_events,
+	return kfd_wait_on_events(p, args->num_events,
 			(void __user *)args->events_ptr,
 			(args->wait_for_all != 0),
-			args->timeout, &args->wait_result);
-
-	return err;
+			&args->timeout, &args->wait_result);
 }
 static int kfd_ioctl_set_scratch_backing_va(struct file *filep,
 					struct kfd_process *p, void *data)
 {
 	struct kfd_ioctl_set_scratch_backing_va_args *args = data;
 	struct kfd_process_device *pdd;
-	struct kfd_dev *dev;
+	struct kfd_node *dev;
 	long err;
 
-	dev = kfd_device_by_id(args->gpu_id);
-	if (!dev)
-		return -EINVAL;
-
 	mutex_lock(&p->mutex);
+	pdd = kfd_process_device_data_by_id(p, args->gpu_id);
+	if (!pdd) {
+		err = -EINVAL;
+		goto err_pdd;
+	}
+	dev = pdd->dev;
 
 	pdd = kfd_bind_process_to_device(dev, p);
 	if (IS_ERR(pdd)) {
@@ -1137,11 +913,12 @@ static int kfd_ioctl_set_scratch_backing_va(struct file *filep,
 	if (dev->dqm->sched_policy == KFD_SCHED_POLICY_NO_HWS &&
 	    pdd->qpd.vmid != 0 && dev->kfd2kgd->set_scratch_backing_va)
 		dev->kfd2kgd->set_scratch_backing_va(
-			dev->kgd, args->va_addr, pdd->qpd.vmid);
+			dev->adev, args->va_addr, pdd->qpd.vmid);
 
 	return 0;
 
 bind_process_to_device_fail:
+err_pdd:
 	mutex_unlock(&p->mutex);
 	return err;
 }
@@ -1150,15 +927,17 @@ static int kfd_ioctl_get_tile_config(struct file *filep,
 		struct kfd_process *p, void *data)
 {
 	struct kfd_ioctl_get_tile_config_args *args = data;
-	struct kfd_dev *dev;
+	struct kfd_process_device *pdd;
 	struct tile_config config;
 	int err = 0;
 
-	dev = kfd_device_by_id(args->gpu_id);
-	if (!dev)
+	mutex_lock(&p->mutex);
+	pdd = kfd_process_device_data_by_id(p, args->gpu_id);
+	mutex_unlock(&p->mutex);
+	if (!pdd)
 		return -EINVAL;
 
-	amdgpu_amdkfd_get_tile_config(dev->kgd, &config);
+	amdgpu_amdkfd_get_tile_config(pdd->dev->adev, &config);
 
 	args->gb_addr_config = config.gb_addr_config;
 	args->num_banks = config.num_banks;
@@ -1193,71 +972,83 @@ static int kfd_ioctl_acquire_vm(struct file *filep, struct kfd_process *p,
 {
 	struct kfd_ioctl_acquire_vm_args *args = data;
 	struct kfd_process_device *pdd;
-	struct kfd_dev *dev;
 	struct file *drm_file;
 	int ret;
 
-	dev = kfd_device_by_id(args->gpu_id);
-	if (!dev)
-		return -EINVAL;
-
 	drm_file = fget(args->drm_fd);
 	if (!drm_file)
 		return -EINVAL;
 
 	mutex_lock(&p->mutex);
-
-	pdd = kfd_get_process_device_data(dev, p);
+	pdd = kfd_process_device_data_by_id(p, args->gpu_id);
 	if (!pdd) {
 		ret = -EINVAL;
-		goto err_unlock;
+		goto err_pdd;
 	}
 
 	if (pdd->drm_file) {
 		ret = pdd->drm_file == drm_file ? 0 : -EBUSY;
-		goto err_unlock;
+		goto err_drm_file;
 	}
 
 	ret = kfd_process_device_init_vm(pdd, drm_file);
 	if (ret)
 		goto err_unlock;
+
 	/* On success, the PDD keeps the drm_file reference */
 	mutex_unlock(&p->mutex);
 
 	return 0;
 
 err_unlock:
+err_pdd:
+err_drm_file:
 	mutex_unlock(&p->mutex);
 	fput(drm_file);
 	return ret;
 }
 
-bool kfd_dev_is_large_bar(struct kfd_dev *dev)
+bool kfd_dev_is_large_bar(struct kfd_node *dev)
 {
-	struct kfd_local_mem_info mem_info;
-
-	if (debug_largebar) {
+	if (dev->kfd->adev->debug_largebar) {
 		pr_debug("Simulate large-bar allocation on non large-bar machine\n");
 		return true;
 	}
 
-	if (dev->use_iommu_v2)
-		return false;
+	if (dev->local_mem_info.local_mem_size_private == 0 &&
+	    dev->local_mem_info.local_mem_size_public > 0)
+		return true;
 
-	amdgpu_amdkfd_get_local_mem_info(dev->kgd, &mem_info);
-	if (mem_info.local_mem_size_private == 0 &&
-			mem_info.local_mem_size_public > 0)
+	if (dev->local_mem_info.local_mem_size_public == 0 &&
+	    dev->kfd->adev->gmc.is_app_apu) {
+		pr_debug("APP APU, Consider like a large bar system\n");
 		return true;
+	}
+
 	return false;
 }
 
+static int kfd_ioctl_get_available_memory(struct file *filep,
+					  struct kfd_process *p, void *data)
+{
+	struct kfd_ioctl_get_available_memory_args *args = data;
+	struct kfd_process_device *pdd = kfd_lock_pdd_by_id(p, args->gpu_id);
+
+	if (!pdd)
+		return -EINVAL;
+	args->available = amdgpu_amdkfd_get_available_memory(pdd->dev->adev,
+							pdd->dev->node_id);
+	kfd_unlock_pdd(pdd);
+	return 0;
+}
+
 static int kfd_ioctl_alloc_memory_of_gpu(struct file *filep,
 					struct kfd_process *p, void *data)
 {
 	struct kfd_ioctl_alloc_memory_of_gpu_args *args = data;
 	struct kfd_process_device *pdd;
 	void *mem;
-	struct kfd_dev *dev;
+	struct kfd_node *dev;
 	int idr_handle;
 	long err;
 	uint64_t offset = args->mmap_offset;
@@ -1273,7 +1064,12 @@ static int kfd_ioctl_alloc_memory_of_gpu(struct file *filep,
 	svm_range_list_lock_and_flush_work(&p->svms, current->mm);
 	mutex_lock(&p->svms.lock);
 	mmap_write_unlock(current->mm);
-	if (interval_tree_iter_first(&p->svms.objects,
+
+	/* Skip a special case that allocates VRAM without VA,
+	 * VA will be invalid of 0.
+	 */
+	if (!(!args->va_addr && (flags & KFD_IOC_ALLOC_MEM_FLAGS_VRAM)) &&
+	    interval_tree_iter_first(&p->svms.objects,
 				     args->va_addr >> PAGE_SHIFT,
 				     (args->va_addr + args->size - 1) >> PAGE_SHIFT)) {
 		pr_err("Address: 0x%llx already allocated by SVM\n",
@@ -1281,21 +1077,39 @@ static int kfd_ioctl_alloc_memory_of_gpu(struct file *filep,
 		mutex_unlock(&p->svms.lock);
 		return -EADDRINUSE;
 	}
+
+	/* When register user buffer check if it has been registered by svm by
+	 * buffer cpu virtual address.
+	 */
+	if ((flags & KFD_IOC_ALLOC_MEM_FLAGS_USERPTR) &&
+	    interval_tree_iter_first(&p->svms.objects,
+				     args->mmap_offset >> PAGE_SHIFT,
+				     (args->mmap_offset  + args->size - 1) >> PAGE_SHIFT)) {
+		pr_err("User Buffer Address: 0x%llx already allocated by SVM\n",
+			args->mmap_offset);
+		mutex_unlock(&p->svms.lock);
+		return -EADDRINUSE;
+	}
+
 	mutex_unlock(&p->svms.lock);
 #endif
-	dev = kfd_device_by_id(args->gpu_id);
-	if (!dev)
-		return -EINVAL;
+	mutex_lock(&p->mutex);
+	pdd = kfd_process_device_data_by_id(p, args->gpu_id);
+	if (!pdd) {
+		err = -EINVAL;
+		goto err_pdd;
+	}
+
+	dev = pdd->dev;
 
 	if ((flags & KFD_IOC_ALLOC_MEM_FLAGS_PUBLIC) &&
 		(flags & KFD_IOC_ALLOC_MEM_FLAGS_VRAM) &&
 		!kfd_dev_is_large_bar(dev)) {
 		pr_err("Alloc host visible vram on small bar is not allowed\n");
-		return -EINVAL;
+		err = -EINVAL;
+		goto err_large_bar;
 	}
 
-	mutex_lock(&p->mutex);
-
 	pdd = kfd_bind_process_to_device(dev, p);
 	if (IS_ERR(pdd)) {
 		err = PTR_ERR(pdd);
@@ -1303,27 +1117,31 @@ static int kfd_ioctl_alloc_memory_of_gpu(struct file *filep,
 	}
 
 	if (flags & KFD_IOC_ALLOC_MEM_FLAGS_DOORBELL) {
-		if (args->size != kfd_doorbell_process_slice(dev)) {
+		if (args->size != kfd_doorbell_process_slice(dev->kfd)) {
 			err = -EINVAL;
 			goto err_unlock;
 		}
 		offset = kfd_get_process_doorbells(pdd);
+		if (!offset) {
+			err = -ENOMEM;
+			goto err_unlock;
+		}
 	} else if (flags & KFD_IOC_ALLOC_MEM_FLAGS_MMIO_REMAP) {
 		if (args->size != PAGE_SIZE) {
 			err = -EINVAL;
 			goto err_unlock;
 		}
-		offset = amdgpu_amdkfd_get_mmio_remap_phys_addr(dev->kgd);
-		if (!offset) {
+		offset = dev->adev->rmmio_remap.bus_addr;
+		if (!offset || (PAGE_SIZE > 4096)) {
 			err = -ENOMEM;
 			goto err_unlock;
 		}
 	}
 
 	err = amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu(
-		dev->kgd, args->va_addr, args->size,
+		dev->adev, args->va_addr, args->size,
 		pdd->drm_priv, (struct kgd_mem **) &mem, &offset,
-		flags);
+		flags, false);
 
 	if (err)
 		goto err_unlock;
@@ -1335,8 +1153,13 @@ static int kfd_ioctl_alloc_memory_of_gpu(struct file *filep,
 	}
 
 	/* Update the VRAM usage count */
-	if (flags & KFD_IOC_ALLOC_MEM_FLAGS_VRAM)
-		WRITE_ONCE(pdd->vram_usage, pdd->vram_usage + args->size);
+	if (flags & KFD_IOC_ALLOC_MEM_FLAGS_VRAM) {
+		uint64_t size = args->size;
+
+		if (flags & KFD_IOC_ALLOC_MEM_FLAGS_AQL_QUEUE_MEM)
+			size >>= 1;
+		atomic64_add(PAGE_ALIGN(size), &pdd->vram_usage);
+	}
 
 	mutex_unlock(&p->mutex);
 
@@ -1353,9 +1176,11 @@ static int kfd_ioctl_alloc_memory_of_gpu(struct file *filep,
 	return 0;
 
 err_free:
-	amdgpu_amdkfd_gpuvm_free_memory_of_gpu(dev->kgd, (struct kgd_mem *)mem,
+	amdgpu_amdkfd_gpuvm_free_memory_of_gpu(dev->adev, (struct kgd_mem *)mem,
 					       pdd->drm_priv, NULL);
 err_unlock:
+err_pdd:
+err_large_bar:
 	mutex_unlock(&p->mutex);
 	return err;
 }
@@ -1366,14 +1191,9 @@ static int kfd_ioctl_free_memory_of_gpu(struct file *filep,
 	struct kfd_ioctl_free_memory_of_gpu_args *args = data;
 	struct kfd_process_device *pdd;
 	void *mem;
-	struct kfd_dev *dev;
 	int ret;
 	uint64_t size = 0;
 
-	dev = kfd_device_by_id(GET_GPU_ID(args->handle));
-	if (!dev)
-		return -EINVAL;
-
 	mutex_lock(&p->mutex);
 	/*
 	 * Safeguard to prevent user space from freeing signal BO.
@@ -1385,11 +1205,11 @@ static int kfd_ioctl_free_memory_of_gpu(struct file *filep,
 		goto err_unlock;
 	}
 
-	pdd = kfd_get_process_device_data(dev, p);
+	pdd = kfd_process_device_data_by_id(p, GET_GPU_ID(args->handle));
 	if (!pdd) {
 		pr_err("Process device data doesn't exist\n");
 		ret = -EINVAL;
-		goto err_unlock;
+		goto err_pdd;
 	}
 
 	mem = kfd_process_device_translate_handle(
@@ -1399,7 +1219,7 @@ static int kfd_ioctl_free_memory_of_gpu(struct file *filep,
 		goto err_unlock;
 	}
 
-	ret = amdgpu_amdkfd_gpuvm_free_memory_of_gpu(dev->kgd,
+	ret = amdgpu_amdkfd_gpuvm_free_memory_of_gpu(pdd->dev->adev,
 				(struct kgd_mem *)mem, pdd->drm_priv, &size);
 
 	/* If freeing the buffer failed, leave the handle in place for
@@ -1409,9 +1229,10 @@ static int kfd_ioctl_free_memory_of_gpu(struct file *filep,
 		kfd_process_device_remove_obj_handle(
 			pdd, GET_IDR_HANDLE(args->handle));
 
-	WRITE_ONCE(pdd->vram_usage, pdd->vram_usage - size);
+	atomic64_sub(size, &pdd->vram_usage);
 
 err_unlock:
+err_pdd:
 	mutex_unlock(&p->mutex);
 	return ret;
 }
@@ -1422,15 +1243,10 @@ static int kfd_ioctl_map_memory_to_gpu(struct file *filep,
 	struct kfd_ioctl_map_memory_to_gpu_args *args = data;
 	struct kfd_process_device *pdd, *peer_pdd;
 	void *mem;
-	struct kfd_dev *dev, *peer;
+	struct kfd_node *dev;
 	long err = 0;
 	int i;
 	uint32_t *devices_arr = NULL;
-	bool table_freed = false;
-
-	dev = kfd_device_by_id(GET_GPU_ID(args->handle));
-	if (!dev)
-		return -EINVAL;
 
 	if (!args->n_devices) {
 		pr_debug("Device IDs array empty\n");
@@ -1455,6 +1271,12 @@ static int kfd_ioctl_map_memory_to_gpu(struct file *filep,
 	}
 
 	mutex_lock(&p->mutex);
+	pdd = kfd_process_device_data_by_id(p, GET_GPU_ID(args->handle));
+	if (!pdd) {
+		err = -EINVAL;
+		goto get_process_device_data_failed;
+	}
+	dev = pdd->dev;
 
 	pdd = kfd_bind_process_to_device(dev, p);
 	if (IS_ERR(pdd)) {
@@ -1470,60 +1292,64 @@ static int kfd_ioctl_map_memory_to_gpu(struct file *filep,
 	}
 
 	for (i = args->n_success; i < args->n_devices; i++) {
-		peer = kfd_device_by_id(devices_arr[i]);
-		if (!peer) {
+		peer_pdd = kfd_process_device_data_by_id(p, devices_arr[i]);
+		if (!peer_pdd) {
 			pr_debug("Getting device by id failed for 0x%x\n",
 				 devices_arr[i]);
 			err = -EINVAL;
 			goto get_mem_obj_from_handle_failed;
 		}
 
-		peer_pdd = kfd_bind_process_to_device(peer, p);
+		peer_pdd = kfd_bind_process_to_device(peer_pdd->dev, p);
 		if (IS_ERR(peer_pdd)) {
 			err = PTR_ERR(peer_pdd);
 			goto get_mem_obj_from_handle_failed;
 		}
+
 		err = amdgpu_amdkfd_gpuvm_map_memory_to_gpu(
-			peer->kgd, (struct kgd_mem *)mem,
-			peer_pdd->drm_priv, &table_freed);
+			peer_pdd->dev->adev, (struct kgd_mem *)mem,
+			peer_pdd->drm_priv);
 		if (err) {
-			pr_err("Failed to map to gpu %d/%d\n",
-			       i, args->n_devices);
+			struct pci_dev *pdev = peer_pdd->dev->adev->pdev;
+
+			dev_err(dev->adev->dev,
+			       "Failed to map peer:%04x:%02x:%02x.%d mem_domain:%d\n",
+			       pci_domain_nr(pdev->bus),
+			       pdev->bus->number,
+			       PCI_SLOT(pdev->devfn),
+			       PCI_FUNC(pdev->devfn),
+			       ((struct kgd_mem *)mem)->domain);
 			goto map_memory_to_gpu_failed;
 		}
 		args->n_success = i+1;
 	}
 
-	mutex_unlock(&p->mutex);
-
-	err = amdgpu_amdkfd_gpuvm_sync_memory(dev->kgd, (struct kgd_mem *) mem, true);
+	err = amdgpu_amdkfd_gpuvm_sync_memory(dev->adev, (struct kgd_mem *) mem, true);
 	if (err) {
 		pr_debug("Sync memory failed, wait interrupted by user signal\n");
 		goto sync_memory_failed;
 	}
 
+	mutex_unlock(&p->mutex);
+
 	/* Flush TLBs after waiting for the page table updates to complete */
-	if (table_freed) {
-		for (i = 0; i < args->n_devices; i++) {
-			peer = kfd_device_by_id(devices_arr[i]);
-			if (WARN_ON_ONCE(!peer))
-				continue;
-			peer_pdd = kfd_get_process_device_data(peer, p);
-			if (WARN_ON_ONCE(!peer_pdd))
-				continue;
-			kfd_flush_tlb(peer_pdd, TLB_FLUSH_LEGACY);
-		}
+	for (i = 0; i < args->n_devices; i++) {
+		peer_pdd = kfd_process_device_data_by_id(p, devices_arr[i]);
+		if (WARN_ON_ONCE(!peer_pdd))
+			continue;
+		kfd_flush_tlb(peer_pdd, TLB_FLUSH_LEGACY);
 	}
 	kfree(devices_arr);
 
 	return err;
 
+get_process_device_data_failed:
 bind_process_to_device_failed:
 get_mem_obj_from_handle_failed:
 map_memory_to_gpu_failed:
+sync_memory_failed:
 	mutex_unlock(&p->mutex);
 copy_from_user_failed:
-sync_memory_failed:
 	kfree(devices_arr);
 
 	return err;
@@ -1535,13 +1361,9 @@ static int kfd_ioctl_unmap_memory_from_gpu(struct file *filep,
 	struct kfd_ioctl_unmap_memory_from_gpu_args *args = data;
 	struct kfd_process_device *pdd, *peer_pdd;
 	void *mem;
-	struct kfd_dev *dev, *peer;
 	long err = 0;
 	uint32_t *devices_arr = NULL, i;
-
-	dev = kfd_device_by_id(GET_GPU_ID(args->handle));
-	if (!dev)
-		return -EINVAL;
+	bool flush_tlb;
 
 	if (!args->n_devices) {
 		pr_debug("Device IDs array empty\n");
@@ -1566,8 +1388,7 @@ static int kfd_ioctl_unmap_memory_from_gpu(struct file *filep,
 	}
 
 	mutex_lock(&p->mutex);
-
-	pdd = kfd_get_process_device_data(dev, p);
+	pdd = kfd_process_device_data_by_id(p, GET_GPU_ID(args->handle));
 	if (!pdd) {
 		err = -EINVAL;
 		goto bind_process_to_device_failed;
@@ -1581,47 +1402,46 @@ static int kfd_ioctl_unmap_memory_from_gpu(struct file *filep,
 	}
 
 	for (i = args->n_success; i < args->n_devices; i++) {
-		peer = kfd_device_by_id(devices_arr[i]);
-		if (!peer) {
-			err = -EINVAL;
-			goto get_mem_obj_from_handle_failed;
-		}
-
-		peer_pdd = kfd_get_process_device_data(peer, p);
+		peer_pdd = kfd_process_device_data_by_id(p, devices_arr[i]);
 		if (!peer_pdd) {
-			err = -ENODEV;
+			err = -EINVAL;
 			goto get_mem_obj_from_handle_failed;
 		}
 		err = amdgpu_amdkfd_gpuvm_unmap_memory_from_gpu(
-			peer->kgd, (struct kgd_mem *)mem, peer_pdd->drm_priv);
+			peer_pdd->dev->adev, (struct kgd_mem *)mem, peer_pdd->drm_priv);
 		if (err) {
-			pr_err("Failed to unmap from gpu %d/%d\n",
-			       i, args->n_devices);
+			pr_debug("Failed to unmap from gpu %d/%d\n", i, args->n_devices);
 			goto unmap_memory_from_gpu_failed;
 		}
 		args->n_success = i+1;
 	}
-	mutex_unlock(&p->mutex);
 
-	if (dev->device_info->asic_family == CHIP_ALDEBARAN) {
-		err = amdgpu_amdkfd_gpuvm_sync_memory(dev->kgd,
+	flush_tlb = kfd_flush_tlb_after_unmap(pdd->dev->kfd);
+	if (flush_tlb) {
+		err = amdgpu_amdkfd_gpuvm_sync_memory(pdd->dev->adev,
 				(struct kgd_mem *) mem, true);
 		if (err) {
 			pr_debug("Sync memory failed, wait interrupted by user signal\n");
 			goto sync_memory_failed;
 		}
+	}
 
-		/* Flush TLBs after waiting for the page table updates to complete */
-		for (i = 0; i < args->n_devices; i++) {
-			peer = kfd_device_by_id(devices_arr[i]);
-			if (WARN_ON_ONCE(!peer))
-				continue;
-			peer_pdd = kfd_get_process_device_data(peer, p);
-			if (WARN_ON_ONCE(!peer_pdd))
-				continue;
+	/* Flush TLBs after waiting for the page table updates to complete */
+	for (i = 0; i < args->n_devices; i++) {
+		peer_pdd = kfd_process_device_data_by_id(p, devices_arr[i]);
+		if (WARN_ON_ONCE(!peer_pdd))
+			continue;
+		if (flush_tlb)
 			kfd_flush_tlb(peer_pdd, TLB_FLUSH_HEAVYWEIGHT);
-		}
+
+		/* Remove dma mapping after tlb flush to avoid IO_PAGE_FAULT */
+		err = amdgpu_amdkfd_gpuvm_dmaunmap_mem(mem, peer_pdd->drm_priv);
+		if (err)
+			goto sync_memory_failed;
 	}
+
+	mutex_unlock(&p->mutex);
+
 	kfree(devices_arr);
 
 	return 0;
@@ -1629,9 +1449,9 @@ static int kfd_ioctl_unmap_memory_from_gpu(struct file *filep,
 bind_process_to_device_failed:
 get_mem_obj_from_handle_failed:
 unmap_memory_from_gpu_failed:
+sync_memory_failed:
 	mutex_unlock(&p->mutex);
 copy_from_user_failed:
-sync_memory_failed:
 	kfree(devices_arr);
 	return err;
 }
@@ -1642,7 +1462,7 @@ static int kfd_ioctl_alloc_queue_gws(struct file *filep,
 	int retval;
 	struct kfd_ioctl_alloc_queue_gws_args *args = data;
 	struct queue *q;
-	struct kfd_dev *dev;
+	struct kfd_node *dev;
 
 	mutex_lock(&p->mutex);
 	q = pqm_get_user_queue(&p->pqm, args->queue_id);
@@ -1664,6 +1484,12 @@ static int kfd_ioctl_alloc_queue_gws(struct file *filep,
 		goto out_unlock;
 	}
 
+	if (p->debug_trap_enabled && (!kfd_dbg_has_gws_support(dev) ||
+				      kfd_dbg_has_cwsr_workaround(dev))) {
+		retval = -EBUSY;
+		goto out_unlock;
+	}
+
 	retval = pqm_set_gws(&p->pqm, args->queue_id, args->num_gws ? dev->gws : NULL);
 	mutex_unlock(&p->mutex);
 
@@ -1679,16 +1505,17 @@ static int kfd_ioctl_get_dmabuf_info(struct file *filep,
 		struct kfd_process *p, void *data)
 {
 	struct kfd_ioctl_get_dmabuf_info_args *args = data;
-	struct kfd_dev *dev = NULL;
-	struct kgd_dev *dma_buf_kgd;
+	struct kfd_node *dev = NULL;
+	struct amdgpu_device *dmabuf_adev;
 	void *metadata_buffer = NULL;
 	uint32_t flags;
+	int8_t xcp_id;
 	unsigned int i;
 	int r;
 
 	/* Find a KFD GPU device that supports the get_dmabuf_info query */
 	for (i = 0; kfd_topology_enum_kfd_devices(i, &dev) == 0; i++)
-		if (dev)
+		if (dev && !kfd_devcgroup_check_permission(dev))
 			break;
 	if (!dev)
 		return -EINVAL;
@@ -1700,20 +1527,17 @@ static int kfd_ioctl_get_dmabuf_info(struct file *filep,
 	}
 
 	/* Get dmabuf info from KGD */
-	r = amdgpu_amdkfd_get_dmabuf_info(dev->kgd, args->dmabuf_fd,
-					  &dma_buf_kgd, &args->size,
+	r = amdgpu_amdkfd_get_dmabuf_info(dev->adev, args->dmabuf_fd,
+					  &dmabuf_adev, &args->size,
 					  metadata_buffer, args->metadata_size,
-					  &args->metadata_size, &flags);
+					  &args->metadata_size, &flags, &xcp_id);
 	if (r)
 		goto exit;
 
-	/* Reverse-lookup gpu_id from kgd pointer */
-	dev = kfd_device_by_kgd(dma_buf_kgd);
-	if (!dev) {
-		r = -EINVAL;
-		goto exit;
-	}
-	args->gpu_id = dev->id;
+	if (xcp_id >= 0)
+		args->gpu_id = dmabuf_adev->kfd.dev->nodes[xcp_id]->id;
+	else
+		args->gpu_id = dev->id;
 	args->flags = flags;
 
 	/* Copy metadata buffer to user mode */
@@ -1735,33 +1559,28 @@ static int kfd_ioctl_import_dmabuf(struct file *filep,
 {
 	struct kfd_ioctl_import_dmabuf_args *args = data;
 	struct kfd_process_device *pdd;
-	struct dma_buf *dmabuf;
-	struct kfd_dev *dev;
 	int idr_handle;
 	uint64_t size;
 	void *mem;
 	int r;
 
-	dev = kfd_device_by_id(args->gpu_id);
-	if (!dev)
-		return -EINVAL;
-
-	dmabuf = dma_buf_get(args->dmabuf_fd);
-	if (IS_ERR(dmabuf))
-		return PTR_ERR(dmabuf);
-
 	mutex_lock(&p->mutex);
+	pdd = kfd_process_device_data_by_id(p, args->gpu_id);
+	if (!pdd) {
+		r = -EINVAL;
+		goto err_unlock;
+	}
 
-	pdd = kfd_bind_process_to_device(dev, p);
+	pdd = kfd_bind_process_to_device(pdd->dev, p);
 	if (IS_ERR(pdd)) {
 		r = PTR_ERR(pdd);
 		goto err_unlock;
 	}
 
-	r = amdgpu_amdkfd_gpuvm_import_dmabuf(dev->kgd, dmabuf,
-					      args->va_addr, pdd->drm_priv,
-					      (struct kgd_mem **)&mem, &size,
-					      NULL);
+	r = amdgpu_amdkfd_gpuvm_import_dmabuf_fd(pdd->dev->adev, args->dmabuf_fd,
+						 args->va_addr, pdd->drm_priv,
+						 (struct kgd_mem **)&mem, &size,
+						 NULL);
 	if (r)
 		goto err_unlock;
 
@@ -1772,35 +1591,90 @@ static int kfd_ioctl_import_dmabuf(struct file *filep,
 	}
 
 	mutex_unlock(&p->mutex);
-	dma_buf_put(dmabuf);
 
 	args->handle = MAKE_HANDLE(args->gpu_id, idr_handle);
 
 	return 0;
 
 err_free:
-	amdgpu_amdkfd_gpuvm_free_memory_of_gpu(dev->kgd, (struct kgd_mem *)mem,
+	amdgpu_amdkfd_gpuvm_free_memory_of_gpu(pdd->dev->adev, (struct kgd_mem *)mem,
 					       pdd->drm_priv, NULL);
 err_unlock:
 	mutex_unlock(&p->mutex);
-	dma_buf_put(dmabuf);
 	return r;
 }
 
+static int kfd_ioctl_export_dmabuf(struct file *filep,
+				   struct kfd_process *p, void *data)
+{
+	struct kfd_ioctl_export_dmabuf_args *args = data;
+	struct kfd_process_device *pdd;
+	struct dma_buf *dmabuf;
+	struct kfd_node *dev;
+	void *mem;
+	int ret = 0;
+
+	dev = kfd_device_by_id(GET_GPU_ID(args->handle));
+	if (!dev)
+		return -EINVAL;
+
+	mutex_lock(&p->mutex);
+
+	pdd = kfd_get_process_device_data(dev, p);
+	if (!pdd) {
+		ret = -EINVAL;
+		goto err_unlock;
+	}
+
+	mem = kfd_process_device_translate_handle(pdd,
+						GET_IDR_HANDLE(args->handle));
+	if (!mem) {
+		ret = -EINVAL;
+		goto err_unlock;
+	}
+
+	ret = amdgpu_amdkfd_gpuvm_export_dmabuf(mem, &dmabuf);
+	mutex_unlock(&p->mutex);
+	if (ret)
+		goto err_out;
+
+	ret = dma_buf_fd(dmabuf, args->flags);
+	if (ret < 0) {
+		dma_buf_put(dmabuf);
+		goto err_out;
+	}
+	/* dma_buf_fd assigns the reference count to the fd, no need to
+	 * put the reference here.
+	 */
+	args->dmabuf_fd = ret;
+
+	return 0;
+
+err_unlock:
+	mutex_unlock(&p->mutex);
+err_out:
+	return ret;
+}
+
 /* Handle requests for watching SMI events */
 static int kfd_ioctl_smi_events(struct file *filep,
 				struct kfd_process *p, void *data)
 {
 	struct kfd_ioctl_smi_events_args *args = data;
-	struct kfd_dev *dev;
+	struct kfd_process_device *pdd;
 
-	dev = kfd_device_by_id(args->gpuid);
-	if (!dev)
+	mutex_lock(&p->mutex);
+
+	pdd = kfd_process_device_data_by_id(p, args->gpuid);
+	mutex_unlock(&p->mutex);
+	if (!pdd)
 		return -EINVAL;
 
-	return kfd_smi_event_open(dev, &args->anon_fd);
+	return kfd_smi_event_open(pdd->dev, &args->anon_fd);
 }
 
+#if IS_ENABLED(CONFIG_HSA_AMD_SVM)
+
 static int kfd_ioctl_set_xnack_mode(struct file *filep,
 				    struct kfd_process *p, void *data)
 {
@@ -1811,22 +1685,29 @@ static int kfd_ioctl_set_xnack_mode(struct file *filep,
 	if (args->xnack_enabled >= 0) {
 		if (!list_empty(&p->pqm.queues)) {
 			pr_debug("Process has user queues running\n");
-			mutex_unlock(&p->mutex);
-			return -EBUSY;
+			r = -EBUSY;
+			goto out_unlock;
 		}
-		if (args->xnack_enabled && !kfd_process_xnack_mode(p, true))
+
+		if (p->xnack_enabled == args->xnack_enabled)
+			goto out_unlock;
+
+		if (args->xnack_enabled && !kfd_process_xnack_mode(p, true)) {
 			r = -EPERM;
-		else
-			p->xnack_enabled = args->xnack_enabled;
+			goto out_unlock;
+		}
+
+		r = svm_range_switch_xnack_reserve_mem(p, args->xnack_enabled);
 	} else {
 		args->xnack_enabled = p->xnack_enabled;
 	}
+
+out_unlock:
 	mutex_unlock(&p->mutex);
 
 	return r;
 }
 
-#if IS_ENABLED(CONFIG_HSA_AMD_SVM)
 static int kfd_ioctl_svm(struct file *filep, struct kfd_process *p, void *data)
 {
 	struct kfd_ioctl_svm_args *args = data;
@@ -1840,22 +1721,1406 @@ static int kfd_ioctl_svm(struct file *filep, struct kfd_process *p, void *data)
 	if (!args->start_addr || !args->size)
 		return -EINVAL;
 
-	mutex_lock(&p->mutex);
-
 	r = svm_ioctl(p, args->op, args->start_addr, args->size, args->nattr,
 		      args->attrs);
 
-	mutex_unlock(&p->mutex);
-
 	return r;
 }
 #else
+static int kfd_ioctl_set_xnack_mode(struct file *filep,
+				    struct kfd_process *p, void *data)
+{
+	return -EPERM;
+}
 static int kfd_ioctl_svm(struct file *filep, struct kfd_process *p, void *data)
 {
 	return -EPERM;
 }
 #endif
 
+static int criu_checkpoint_process(struct kfd_process *p,
+			     uint8_t __user *user_priv_data,
+			     uint64_t *priv_offset)
+{
+	struct kfd_criu_process_priv_data process_priv;
+	int ret;
+
+	memset(&process_priv, 0, sizeof(process_priv));
+
+	process_priv.version = KFD_CRIU_PRIV_VERSION;
+	/* For CR, we don't consider negative xnack mode which is used for
+	 * querying without changing it, here 0 simply means disabled and 1
+	 * means enabled so retry for finding a valid PTE.
+	 */
+	process_priv.xnack_mode = p->xnack_enabled ? 1 : 0;
+
+	ret = copy_to_user(user_priv_data + *priv_offset,
+				&process_priv, sizeof(process_priv));
+
+	if (ret) {
+		pr_err("Failed to copy process information to user\n");
+		ret = -EFAULT;
+	}
+
+	*priv_offset += sizeof(process_priv);
+	return ret;
+}
+
+static int criu_checkpoint_devices(struct kfd_process *p,
+			     uint32_t num_devices,
+			     uint8_t __user *user_addr,
+			     uint8_t __user *user_priv_data,
+			     uint64_t *priv_offset)
+{
+	struct kfd_criu_device_priv_data *device_priv = NULL;
+	struct kfd_criu_device_bucket *device_buckets = NULL;
+	int ret = 0, i;
+
+	device_buckets = kvzalloc(num_devices * sizeof(*device_buckets), GFP_KERNEL);
+	if (!device_buckets) {
+		ret = -ENOMEM;
+		goto exit;
+	}
+
+	device_priv = kvzalloc(num_devices * sizeof(*device_priv), GFP_KERNEL);
+	if (!device_priv) {
+		ret = -ENOMEM;
+		goto exit;
+	}
+
+	for (i = 0; i < num_devices; i++) {
+		struct kfd_process_device *pdd = p->pdds[i];
+
+		device_buckets[i].user_gpu_id = pdd->user_gpu_id;
+		device_buckets[i].actual_gpu_id = pdd->dev->id;
+
+		/*
+		 * priv_data does not contain useful information for now and is reserved for
+		 * future use, so we do not set its contents.
+		 */
+	}
+
+	ret = copy_to_user(user_addr, device_buckets, num_devices * sizeof(*device_buckets));
+	if (ret) {
+		pr_err("Failed to copy device information to user\n");
+		ret = -EFAULT;
+		goto exit;
+	}
+
+	ret = copy_to_user(user_priv_data + *priv_offset,
+			   device_priv,
+			   num_devices * sizeof(*device_priv));
+	if (ret) {
+		pr_err("Failed to copy device information to user\n");
+		ret = -EFAULT;
+	}
+	*priv_offset += num_devices * sizeof(*device_priv);
+
+exit:
+	kvfree(device_buckets);
+	kvfree(device_priv);
+	return ret;
+}
+
+static uint32_t get_process_num_bos(struct kfd_process *p)
+{
+	uint32_t num_of_bos = 0;
+	int i;
+
+	/* Run over all PDDs of the process */
+	for (i = 0; i < p->n_pdds; i++) {
+		struct kfd_process_device *pdd = p->pdds[i];
+		void *mem;
+		int id;
+
+		idr_for_each_entry(&pdd->alloc_idr, mem, id) {
+			struct kgd_mem *kgd_mem = (struct kgd_mem *)mem;
+
+			if (!kgd_mem->va || kgd_mem->va > pdd->gpuvm_base)
+				num_of_bos++;
+		}
+	}
+	return num_of_bos;
+}
+
+static int criu_get_prime_handle(struct kgd_mem *mem,
+				 int flags, u32 *shared_fd,
+				 struct file **file)
+{
+	struct dma_buf *dmabuf;
+	int ret;
+
+	ret = amdgpu_amdkfd_gpuvm_export_dmabuf(mem, &dmabuf);
+	if (ret) {
+		pr_err("dmabuf export failed for the BO\n");
+		return ret;
+	}
+
+	ret = get_unused_fd_flags(flags);
+	if (ret < 0) {
+		pr_err("dmabuf create fd failed, ret:%d\n", ret);
+		goto out_free_dmabuf;
+	}
+
+	*shared_fd = ret;
+	*file = dmabuf->file;
+	return 0;
+
+out_free_dmabuf:
+	dma_buf_put(dmabuf);
+	return ret;
+}
+
+static void commit_files(struct file **files,
+			 struct kfd_criu_bo_bucket *bo_buckets,
+			 unsigned int count,
+			 int err)
+{
+	while (count--) {
+		struct file *file = files[count];
+
+		if (!file)
+			continue;
+		if (err) {
+			fput(file);
+			put_unused_fd(bo_buckets[count].dmabuf_fd);
+		} else {
+			fd_install(bo_buckets[count].dmabuf_fd, file);
+		}
+	}
+}
+
+static int criu_checkpoint_bos(struct kfd_process *p,
+			       uint32_t num_bos,
+			       uint8_t __user *user_bos,
+			       uint8_t __user *user_priv_data,
+			       uint64_t *priv_offset)
+{
+	struct kfd_criu_bo_bucket *bo_buckets;
+	struct kfd_criu_bo_priv_data *bo_privs;
+	struct file **files = NULL;
+	int ret = 0, pdd_index, bo_index = 0, id;
+	void *mem;
+
+	bo_buckets = kvzalloc(num_bos * sizeof(*bo_buckets), GFP_KERNEL);
+	if (!bo_buckets)
+		return -ENOMEM;
+
+	bo_privs = kvzalloc(num_bos * sizeof(*bo_privs), GFP_KERNEL);
+	if (!bo_privs) {
+		ret = -ENOMEM;
+		goto exit;
+	}
+
+	files = kvzalloc(num_bos * sizeof(struct file *), GFP_KERNEL);
+	if (!files) {
+		ret = -ENOMEM;
+		goto exit;
+	}
+
+	for (pdd_index = 0; pdd_index < p->n_pdds; pdd_index++) {
+		struct kfd_process_device *pdd = p->pdds[pdd_index];
+		struct amdgpu_bo *dumper_bo;
+		struct kgd_mem *kgd_mem;
+
+		idr_for_each_entry(&pdd->alloc_idr, mem, id) {
+			struct kfd_criu_bo_bucket *bo_bucket;
+			struct kfd_criu_bo_priv_data *bo_priv;
+			int i, dev_idx = 0;
+
+			kgd_mem = (struct kgd_mem *)mem;
+			dumper_bo = kgd_mem->bo;
+
+			/* Skip checkpointing BOs that are used for Trap handler
+			 * code and state. Currently, these BOs have a VA that
+			 * is less GPUVM Base
+			 */
+			if (kgd_mem->va && kgd_mem->va <= pdd->gpuvm_base)
+				continue;
+
+			bo_bucket = &bo_buckets[bo_index];
+			bo_priv = &bo_privs[bo_index];
+
+			bo_bucket->gpu_id = pdd->user_gpu_id;
+			bo_bucket->addr = (uint64_t)kgd_mem->va;
+			bo_bucket->size = amdgpu_bo_size(dumper_bo);
+			bo_bucket->alloc_flags = (uint32_t)kgd_mem->alloc_flags;
+			bo_priv->idr_handle = id;
+
+			if (bo_bucket->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_USERPTR) {
+				ret = amdgpu_ttm_tt_get_userptr(&dumper_bo->tbo,
+								&bo_priv->user_addr);
+				if (ret) {
+					pr_err("Failed to obtain user address for user-pointer bo\n");
+					goto exit;
+				}
+			}
+			if (bo_bucket->alloc_flags
+			    & (KFD_IOC_ALLOC_MEM_FLAGS_VRAM | KFD_IOC_ALLOC_MEM_FLAGS_GTT)) {
+				ret = criu_get_prime_handle(kgd_mem,
+						bo_bucket->alloc_flags &
+						KFD_IOC_ALLOC_MEM_FLAGS_WRITABLE ? DRM_RDWR : 0,
+						&bo_bucket->dmabuf_fd, &files[bo_index]);
+				if (ret)
+					goto exit;
+			} else {
+				bo_bucket->dmabuf_fd = KFD_INVALID_FD;
+			}
+
+			if (bo_bucket->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_DOORBELL)
+				bo_bucket->offset = KFD_MMAP_TYPE_DOORBELL |
+					KFD_MMAP_GPU_ID(pdd->dev->id);
+			else if (bo_bucket->alloc_flags &
+				KFD_IOC_ALLOC_MEM_FLAGS_MMIO_REMAP)
+				bo_bucket->offset = KFD_MMAP_TYPE_MMIO |
+					KFD_MMAP_GPU_ID(pdd->dev->id);
+			else
+				bo_bucket->offset = amdgpu_bo_mmap_offset(dumper_bo);
+
+			for (i = 0; i < p->n_pdds; i++) {
+				if (amdgpu_amdkfd_bo_mapped_to_dev(p->pdds[i]->drm_priv, kgd_mem))
+					bo_priv->mapped_gpuids[dev_idx++] = p->pdds[i]->user_gpu_id;
+			}
+
+			pr_debug("bo_size = 0x%llx, bo_addr = 0x%llx bo_offset = 0x%llx\n"
+					"gpu_id = 0x%x alloc_flags = 0x%x idr_handle = 0x%x",
+					bo_bucket->size,
+					bo_bucket->addr,
+					bo_bucket->offset,
+					bo_bucket->gpu_id,
+					bo_bucket->alloc_flags,
+					bo_priv->idr_handle);
+			bo_index++;
+		}
+	}
+
+	ret = copy_to_user(user_bos, bo_buckets, num_bos * sizeof(*bo_buckets));
+	if (ret) {
+		pr_err("Failed to copy BO information to user\n");
+		ret = -EFAULT;
+		goto exit;
+	}
+
+	ret = copy_to_user(user_priv_data + *priv_offset, bo_privs, num_bos * sizeof(*bo_privs));
+	if (ret) {
+		pr_err("Failed to copy BO priv information to user\n");
+		ret = -EFAULT;
+		goto exit;
+	}
+
+	*priv_offset += num_bos * sizeof(*bo_privs);
+
+exit:
+	commit_files(files, bo_buckets, bo_index, ret);
+	kvfree(files);
+	kvfree(bo_buckets);
+	kvfree(bo_privs);
+	return ret;
+}
+
+static int criu_get_process_object_info(struct kfd_process *p,
+					uint32_t *num_devices,
+					uint32_t *num_bos,
+					uint32_t *num_objects,
+					uint64_t *objs_priv_size)
+{
+	uint64_t queues_priv_data_size, svm_priv_data_size, priv_size;
+	uint32_t num_queues, num_events, num_svm_ranges;
+	int ret;
+
+	*num_devices = p->n_pdds;
+	*num_bos = get_process_num_bos(p);
+
+	ret = kfd_process_get_queue_info(p, &num_queues, &queues_priv_data_size);
+	if (ret)
+		return ret;
+
+	num_events = kfd_get_num_events(p);
+
+	svm_range_get_info(p, &num_svm_ranges, &svm_priv_data_size);
+
+	*num_objects = num_queues + num_events + num_svm_ranges;
+
+	if (objs_priv_size) {
+		priv_size = sizeof(struct kfd_criu_process_priv_data);
+		priv_size += *num_devices * sizeof(struct kfd_criu_device_priv_data);
+		priv_size += *num_bos * sizeof(struct kfd_criu_bo_priv_data);
+		priv_size += queues_priv_data_size;
+		priv_size += num_events * sizeof(struct kfd_criu_event_priv_data);
+		priv_size += svm_priv_data_size;
+		*objs_priv_size = priv_size;
+	}
+	return 0;
+}
+
+static int criu_checkpoint(struct file *filep,
+			   struct kfd_process *p,
+			   struct kfd_ioctl_criu_args *args)
+{
+	int ret;
+	uint32_t num_devices, num_bos, num_objects;
+	uint64_t priv_size, priv_offset = 0, bo_priv_offset;
+
+	if (!args->devices || !args->bos || !args->priv_data)
+		return -EINVAL;
+
+	mutex_lock(&p->mutex);
+
+	if (!p->n_pdds) {
+		pr_err("No pdd for given process\n");
+		ret = -ENODEV;
+		goto exit_unlock;
+	}
+
+	/* Confirm all process queues are evicted */
+	if (!p->queues_paused) {
+		pr_err("Cannot dump process when queues are not in evicted state\n");
+		/* CRIU plugin did not call op PROCESS_INFO before checkpointing */
+		ret = -EINVAL;
+		goto exit_unlock;
+	}
+
+	ret = criu_get_process_object_info(p, &num_devices, &num_bos, &num_objects, &priv_size);
+	if (ret)
+		goto exit_unlock;
+
+	if (num_devices != args->num_devices ||
+	    num_bos != args->num_bos ||
+	    num_objects != args->num_objects ||
+	    priv_size != args->priv_data_size) {
+
+		ret = -EINVAL;
+		goto exit_unlock;
+	}
+
+	/* each function will store private data inside priv_data and adjust priv_offset */
+	ret = criu_checkpoint_process(p, (uint8_t __user *)args->priv_data, &priv_offset);
+	if (ret)
+		goto exit_unlock;
+
+	ret = criu_checkpoint_devices(p, num_devices, (uint8_t __user *)args->devices,
+				(uint8_t __user *)args->priv_data, &priv_offset);
+	if (ret)
+		goto exit_unlock;
+
+	/* Leave room for BOs in the private data. They need to be restored
+	 * before events, but we checkpoint them last to simplify the error
+	 * handling.
+	 */
+	bo_priv_offset = priv_offset;
+	priv_offset += num_bos * sizeof(struct kfd_criu_bo_priv_data);
+
+	if (num_objects) {
+		ret = kfd_criu_checkpoint_queues(p, (uint8_t __user *)args->priv_data,
+						 &priv_offset);
+		if (ret)
+			goto exit_unlock;
+
+		ret = kfd_criu_checkpoint_events(p, (uint8_t __user *)args->priv_data,
+						 &priv_offset);
+		if (ret)
+			goto exit_unlock;
+
+		ret = kfd_criu_checkpoint_svm(p, (uint8_t __user *)args->priv_data, &priv_offset);
+		if (ret)
+			goto exit_unlock;
+	}
+
+	/* This must be the last thing in this function that can fail.
+	 * Otherwise we leak dmabuf file descriptors.
+	 */
+	ret = criu_checkpoint_bos(p, num_bos, (uint8_t __user *)args->bos,
+			   (uint8_t __user *)args->priv_data, &bo_priv_offset);
+
+exit_unlock:
+	mutex_unlock(&p->mutex);
+	if (ret)
+		pr_err("Failed to dump CRIU ret:%d\n", ret);
+	else
+		pr_debug("CRIU dump ret:%d\n", ret);
+
+	return ret;
+}
+
+static int criu_restore_process(struct kfd_process *p,
+				struct kfd_ioctl_criu_args *args,
+				uint64_t *priv_offset,
+				uint64_t max_priv_data_size)
+{
+	int ret = 0;
+	struct kfd_criu_process_priv_data process_priv;
+
+	if (*priv_offset + sizeof(process_priv) > max_priv_data_size)
+		return -EINVAL;
+
+	ret = copy_from_user(&process_priv,
+				(void __user *)(args->priv_data + *priv_offset),
+				sizeof(process_priv));
+	if (ret) {
+		pr_err("Failed to copy process private information from user\n");
+		ret = -EFAULT;
+		goto exit;
+	}
+	*priv_offset += sizeof(process_priv);
+
+	if (process_priv.version != KFD_CRIU_PRIV_VERSION) {
+		pr_err("Invalid CRIU API version (checkpointed:%d current:%d)\n",
+			process_priv.version, KFD_CRIU_PRIV_VERSION);
+		return -EINVAL;
+	}
+
+	pr_debug("Setting XNACK mode\n");
+	if (process_priv.xnack_mode && !kfd_process_xnack_mode(p, true)) {
+		pr_err("xnack mode cannot be set\n");
+		ret = -EPERM;
+		goto exit;
+	} else {
+		pr_debug("set xnack mode: %d\n", process_priv.xnack_mode);
+		p->xnack_enabled = process_priv.xnack_mode;
+	}
+
+exit:
+	return ret;
+}
+
+static int criu_restore_devices(struct kfd_process *p,
+				struct kfd_ioctl_criu_args *args,
+				uint64_t *priv_offset,
+				uint64_t max_priv_data_size)
+{
+	struct kfd_criu_device_bucket *device_buckets;
+	struct kfd_criu_device_priv_data *device_privs;
+	int ret = 0;
+	uint32_t i;
+
+	if (args->num_devices != p->n_pdds)
+		return -EINVAL;
+
+	if (*priv_offset + (args->num_devices * sizeof(*device_privs)) > max_priv_data_size)
+		return -EINVAL;
+
+	device_buckets = kmalloc_array(args->num_devices, sizeof(*device_buckets), GFP_KERNEL);
+	if (!device_buckets)
+		return -ENOMEM;
+
+	ret = copy_from_user(device_buckets, (void __user *)args->devices,
+				args->num_devices * sizeof(*device_buckets));
+	if (ret) {
+		pr_err("Failed to copy devices buckets from user\n");
+		ret = -EFAULT;
+		goto exit;
+	}
+
+	for (i = 0; i < args->num_devices; i++) {
+		struct kfd_node *dev;
+		struct kfd_process_device *pdd;
+		struct file *drm_file;
+
+		/* device private data is not currently used */
+
+		if (!device_buckets[i].user_gpu_id) {
+			pr_err("Invalid user gpu_id\n");
+			ret = -EINVAL;
+			goto exit;
+		}
+
+		dev = kfd_device_by_id(device_buckets[i].actual_gpu_id);
+		if (!dev) {
+			pr_err("Failed to find device with gpu_id = %x\n",
+				device_buckets[i].actual_gpu_id);
+			ret = -EINVAL;
+			goto exit;
+		}
+
+		pdd = kfd_get_process_device_data(dev, p);
+		if (!pdd) {
+			pr_err("Failed to get pdd for gpu_id = %x\n",
+					device_buckets[i].actual_gpu_id);
+			ret = -EINVAL;
+			goto exit;
+		}
+		pdd->user_gpu_id = device_buckets[i].user_gpu_id;
+
+		drm_file = fget(device_buckets[i].drm_fd);
+		if (!drm_file) {
+			pr_err("Invalid render node file descriptor sent from plugin (%d)\n",
+				device_buckets[i].drm_fd);
+			ret = -EINVAL;
+			goto exit;
+		}
+
+		if (pdd->drm_file) {
+			ret = -EINVAL;
+			goto exit;
+		}
+
+		/* create the vm using render nodes for kfd pdd */
+		if (kfd_process_device_init_vm(pdd, drm_file)) {
+			pr_err("could not init vm for given pdd\n");
+			/* On success, the PDD keeps the drm_file reference */
+			fput(drm_file);
+			ret = -EINVAL;
+			goto exit;
+		}
+		/*
+		 * pdd now already has the vm bound to render node so below api won't create a new
+		 * exclusive kfd mapping but use existing one with renderDXXX but is still needed
+		 * for iommu v2 binding  and runtime pm.
+		 */
+		pdd = kfd_bind_process_to_device(dev, p);
+		if (IS_ERR(pdd)) {
+			ret = PTR_ERR(pdd);
+			goto exit;
+		}
+
+		if (!pdd->qpd.proc_doorbells) {
+			ret = kfd_alloc_process_doorbells(dev->kfd, pdd);
+			if (ret)
+				goto exit;
+		}
+	}
+
+	/*
+	 * We are not copying device private data from user as we are not using the data for now,
+	 * but we still adjust for its private data.
+	 */
+	*priv_offset += args->num_devices * sizeof(*device_privs);
+
+exit:
+	kfree(device_buckets);
+	return ret;
+}
+
+static int criu_restore_memory_of_gpu(struct kfd_process_device *pdd,
+				      struct kfd_criu_bo_bucket *bo_bucket,
+				      struct kfd_criu_bo_priv_data *bo_priv,
+				      struct kgd_mem **kgd_mem)
+{
+	int idr_handle;
+	int ret;
+	const bool criu_resume = true;
+	u64 offset;
+
+	if (bo_bucket->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_DOORBELL) {
+		if (bo_bucket->size !=
+				kfd_doorbell_process_slice(pdd->dev->kfd))
+			return -EINVAL;
+
+		offset = kfd_get_process_doorbells(pdd);
+		if (!offset)
+			return -ENOMEM;
+	} else if (bo_bucket->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_MMIO_REMAP) {
+		/* MMIO BOs need remapped bus address */
+		if (bo_bucket->size != PAGE_SIZE) {
+			pr_err("Invalid page size\n");
+			return -EINVAL;
+		}
+		offset = pdd->dev->adev->rmmio_remap.bus_addr;
+		if (!offset || (PAGE_SIZE > 4096)) {
+			pr_err("amdgpu_amdkfd_get_mmio_remap_phys_addr failed\n");
+			return -ENOMEM;
+		}
+	} else if (bo_bucket->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_USERPTR) {
+		offset = bo_priv->user_addr;
+	}
+	/* Create the BO */
+	ret = amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu(pdd->dev->adev, bo_bucket->addr,
+						      bo_bucket->size, pdd->drm_priv, kgd_mem,
+						      &offset, bo_bucket->alloc_flags, criu_resume);
+	if (ret) {
+		pr_err("Could not create the BO\n");
+		return ret;
+	}
+	pr_debug("New BO created: size:0x%llx addr:0x%llx offset:0x%llx\n",
+		 bo_bucket->size, bo_bucket->addr, offset);
+
+	/* Restore previous IDR handle */
+	pr_debug("Restoring old IDR handle for the BO");
+	idr_handle = idr_alloc(&pdd->alloc_idr, *kgd_mem, bo_priv->idr_handle,
+			       bo_priv->idr_handle + 1, GFP_KERNEL);
+
+	if (idr_handle < 0) {
+		pr_err("Could not allocate idr\n");
+		amdgpu_amdkfd_gpuvm_free_memory_of_gpu(pdd->dev->adev, *kgd_mem, pdd->drm_priv,
+						       NULL);
+		return -ENOMEM;
+	}
+
+	if (bo_bucket->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_DOORBELL)
+		bo_bucket->restored_offset = KFD_MMAP_TYPE_DOORBELL | KFD_MMAP_GPU_ID(pdd->dev->id);
+	if (bo_bucket->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_MMIO_REMAP) {
+		bo_bucket->restored_offset = KFD_MMAP_TYPE_MMIO | KFD_MMAP_GPU_ID(pdd->dev->id);
+	} else if (bo_bucket->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_GTT) {
+		bo_bucket->restored_offset = offset;
+	} else if (bo_bucket->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_VRAM) {
+		bo_bucket->restored_offset = offset;
+		/* Update the VRAM usage count */
+		atomic64_add(bo_bucket->size, &pdd->vram_usage);
+	}
+	return 0;
+}
+
+static int criu_restore_bo(struct kfd_process *p,
+			   struct kfd_criu_bo_bucket *bo_bucket,
+			   struct kfd_criu_bo_priv_data *bo_priv,
+			   struct file **file)
+{
+	struct kfd_process_device *pdd;
+	struct kgd_mem *kgd_mem;
+	int ret;
+	int j;
+
+	pr_debug("Restoring BO size:0x%llx addr:0x%llx gpu_id:0x%x flags:0x%x idr_handle:0x%x\n",
+		 bo_bucket->size, bo_bucket->addr, bo_bucket->gpu_id, bo_bucket->alloc_flags,
+		 bo_priv->idr_handle);
+
+	pdd = kfd_process_device_data_by_id(p, bo_bucket->gpu_id);
+	if (!pdd) {
+		pr_err("Failed to get pdd\n");
+		return -ENODEV;
+	}
+
+	ret = criu_restore_memory_of_gpu(pdd, bo_bucket, bo_priv, &kgd_mem);
+	if (ret)
+		return ret;
+
+	/* now map these BOs to GPU/s */
+	for (j = 0; j < p->n_pdds; j++) {
+		struct kfd_node *peer;
+		struct kfd_process_device *peer_pdd;
+
+		if (!bo_priv->mapped_gpuids[j])
+			break;
+
+		peer_pdd = kfd_process_device_data_by_id(p, bo_priv->mapped_gpuids[j]);
+		if (!peer_pdd)
+			return -EINVAL;
+
+		peer = peer_pdd->dev;
+
+		peer_pdd = kfd_bind_process_to_device(peer, p);
+		if (IS_ERR(peer_pdd))
+			return PTR_ERR(peer_pdd);
+
+		ret = amdgpu_amdkfd_gpuvm_map_memory_to_gpu(peer->adev, kgd_mem,
+							    peer_pdd->drm_priv);
+		if (ret) {
+			pr_err("Failed to map to gpu %d/%d\n", j, p->n_pdds);
+			return ret;
+		}
+	}
+
+	pr_debug("map memory was successful for the BO\n");
+	/* create the dmabuf object and export the bo */
+	if (bo_bucket->alloc_flags
+	    & (KFD_IOC_ALLOC_MEM_FLAGS_VRAM | KFD_IOC_ALLOC_MEM_FLAGS_GTT)) {
+		ret = criu_get_prime_handle(kgd_mem, DRM_RDWR,
+					    &bo_bucket->dmabuf_fd, file);
+		if (ret)
+			return ret;
+	} else {
+		bo_bucket->dmabuf_fd = KFD_INVALID_FD;
+	}
+
+	return 0;
+}
+
+static int criu_restore_bos(struct kfd_process *p,
+			    struct kfd_ioctl_criu_args *args,
+			    uint64_t *priv_offset,
+			    uint64_t max_priv_data_size)
+{
+	struct kfd_criu_bo_bucket *bo_buckets = NULL;
+	struct kfd_criu_bo_priv_data *bo_privs = NULL;
+	struct file **files = NULL;
+	int ret = 0;
+	uint32_t i = 0;
+
+	if (*priv_offset + (args->num_bos * sizeof(*bo_privs)) > max_priv_data_size)
+		return -EINVAL;
+
+	/* Prevent MMU notifications until stage-4 IOCTL (CRIU_RESUME) is received */
+	amdgpu_amdkfd_block_mmu_notifications(p->kgd_process_info);
+
+	bo_buckets = kvmalloc_array(args->num_bos, sizeof(*bo_buckets), GFP_KERNEL);
+	if (!bo_buckets)
+		return -ENOMEM;
+
+	files = kvzalloc(args->num_bos * sizeof(struct file *), GFP_KERNEL);
+	if (!files) {
+		ret = -ENOMEM;
+		goto exit;
+	}
+
+	ret = copy_from_user(bo_buckets, (void __user *)args->bos,
+			     args->num_bos * sizeof(*bo_buckets));
+	if (ret) {
+		pr_err("Failed to copy BOs information from user\n");
+		ret = -EFAULT;
+		goto exit;
+	}
+
+	bo_privs = kvmalloc_array(args->num_bos, sizeof(*bo_privs), GFP_KERNEL);
+	if (!bo_privs) {
+		ret = -ENOMEM;
+		goto exit;
+	}
+
+	ret = copy_from_user(bo_privs, (void __user *)args->priv_data + *priv_offset,
+			     args->num_bos * sizeof(*bo_privs));
+	if (ret) {
+		pr_err("Failed to copy BOs information from user\n");
+		ret = -EFAULT;
+		goto exit;
+	}
+	*priv_offset += args->num_bos * sizeof(*bo_privs);
+
+	/* Create and map new BOs */
+	for (; i < args->num_bos; i++) {
+		ret = criu_restore_bo(p, &bo_buckets[i], &bo_privs[i], &files[i]);
+		if (ret) {
+			pr_debug("Failed to restore BO[%d] ret%d\n", i, ret);
+			goto exit;
+		}
+	} /* done */
+
+	/* Copy only the buckets back so user can read bo_buckets[N].restored_offset */
+	ret = copy_to_user((void __user *)args->bos,
+				bo_buckets,
+				(args->num_bos * sizeof(*bo_buckets)));
+	if (ret)
+		ret = -EFAULT;
+
+exit:
+	commit_files(files, bo_buckets, i, ret);
+	kvfree(files);
+	kvfree(bo_buckets);
+	kvfree(bo_privs);
+	return ret;
+}
+
+static int criu_restore_objects(struct file *filep,
+				struct kfd_process *p,
+				struct kfd_ioctl_criu_args *args,
+				uint64_t *priv_offset,
+				uint64_t max_priv_data_size)
+{
+	int ret = 0;
+	uint32_t i;
+
+	BUILD_BUG_ON(offsetof(struct kfd_criu_queue_priv_data, object_type));
+	BUILD_BUG_ON(offsetof(struct kfd_criu_event_priv_data, object_type));
+	BUILD_BUG_ON(offsetof(struct kfd_criu_svm_range_priv_data, object_type));
+
+	for (i = 0; i < args->num_objects; i++) {
+		uint32_t object_type;
+
+		if (*priv_offset + sizeof(object_type) > max_priv_data_size) {
+			pr_err("Invalid private data size\n");
+			return -EINVAL;
+		}
+
+		ret = get_user(object_type, (uint32_t __user *)(args->priv_data + *priv_offset));
+		if (ret) {
+			pr_err("Failed to copy private information from user\n");
+			goto exit;
+		}
+
+		switch (object_type) {
+		case KFD_CRIU_OBJECT_TYPE_QUEUE:
+			ret = kfd_criu_restore_queue(p, (uint8_t __user *)args->priv_data,
+						     priv_offset, max_priv_data_size);
+			if (ret)
+				goto exit;
+			break;
+		case KFD_CRIU_OBJECT_TYPE_EVENT:
+			ret = kfd_criu_restore_event(filep, p, (uint8_t __user *)args->priv_data,
+						     priv_offset, max_priv_data_size);
+			if (ret)
+				goto exit;
+			break;
+		case KFD_CRIU_OBJECT_TYPE_SVM_RANGE:
+			ret = kfd_criu_restore_svm(p, (uint8_t __user *)args->priv_data,
+						     priv_offset, max_priv_data_size);
+			if (ret)
+				goto exit;
+			break;
+		default:
+			pr_err("Invalid object type:%u at index:%d\n", object_type, i);
+			ret = -EINVAL;
+			goto exit;
+		}
+	}
+exit:
+	return ret;
+}
+
+static int criu_restore(struct file *filep,
+			struct kfd_process *p,
+			struct kfd_ioctl_criu_args *args)
+{
+	uint64_t priv_offset = 0;
+	int ret = 0;
+
+	pr_debug("CRIU restore (num_devices:%u num_bos:%u num_objects:%u priv_data_size:%llu)\n",
+		 args->num_devices, args->num_bos, args->num_objects, args->priv_data_size);
+
+	if ((args->num_bos > 0 && !args->bos) || !args->devices || !args->priv_data ||
+	    !args->priv_data_size || !args->num_devices)
+		return -EINVAL;
+
+	mutex_lock(&p->mutex);
+
+	/*
+	 * Set the process to evicted state to avoid running any new queues before all the memory
+	 * mappings are ready.
+	 */
+	ret = kfd_process_evict_queues(p, KFD_QUEUE_EVICTION_CRIU_RESTORE);
+	if (ret)
+		goto exit_unlock;
+
+	/* Each function will adjust priv_offset based on how many bytes they consumed */
+	ret = criu_restore_process(p, args, &priv_offset, args->priv_data_size);
+	if (ret)
+		goto exit_unlock;
+
+	ret = criu_restore_devices(p, args, &priv_offset, args->priv_data_size);
+	if (ret)
+		goto exit_unlock;
+
+	ret = criu_restore_bos(p, args, &priv_offset, args->priv_data_size);
+	if (ret)
+		goto exit_unlock;
+
+	ret = criu_restore_objects(filep, p, args, &priv_offset, args->priv_data_size);
+	if (ret)
+		goto exit_unlock;
+
+	if (priv_offset != args->priv_data_size) {
+		pr_err("Invalid private data size\n");
+		ret = -EINVAL;
+	}
+
+exit_unlock:
+	mutex_unlock(&p->mutex);
+	if (ret)
+		pr_err("Failed to restore CRIU ret:%d\n", ret);
+	else
+		pr_debug("CRIU restore successful\n");
+
+	return ret;
+}
+
+static int criu_unpause(struct file *filep,
+			struct kfd_process *p,
+			struct kfd_ioctl_criu_args *args)
+{
+	int ret;
+
+	mutex_lock(&p->mutex);
+
+	if (!p->queues_paused) {
+		mutex_unlock(&p->mutex);
+		return -EINVAL;
+	}
+
+	ret = kfd_process_restore_queues(p);
+	if (ret)
+		pr_err("Failed to unpause queues ret:%d\n", ret);
+	else
+		p->queues_paused = false;
+
+	mutex_unlock(&p->mutex);
+
+	return ret;
+}
+
+static int criu_resume(struct file *filep,
+			struct kfd_process *p,
+			struct kfd_ioctl_criu_args *args)
+{
+	struct kfd_process *target = NULL;
+	struct pid *pid = NULL;
+	int ret = 0;
+
+	pr_debug("Inside %s, target pid for criu restore: %d\n", __func__,
+		 args->pid);
+
+	pid = find_get_pid(args->pid);
+	if (!pid) {
+		pr_err("Cannot find pid info for %i\n", args->pid);
+		return -ESRCH;
+	}
+
+	pr_debug("calling kfd_lookup_process_by_pid\n");
+	target = kfd_lookup_process_by_pid(pid);
+
+	put_pid(pid);
+
+	if (!target) {
+		pr_debug("Cannot find process info for %i\n", args->pid);
+		return -ESRCH;
+	}
+
+	mutex_lock(&target->mutex);
+	ret = kfd_criu_resume_svm(target);
+	if (ret) {
+		pr_err("kfd_criu_resume_svm failed for %i\n", args->pid);
+		goto exit;
+	}
+
+	ret =  amdgpu_amdkfd_criu_resume(target->kgd_process_info);
+	if (ret)
+		pr_err("amdgpu_amdkfd_criu_resume failed for %i\n", args->pid);
+
+exit:
+	mutex_unlock(&target->mutex);
+
+	kfd_unref_process(target);
+	return ret;
+}
+
+static int criu_process_info(struct file *filep,
+				struct kfd_process *p,
+				struct kfd_ioctl_criu_args *args)
+{
+	int ret = 0;
+
+	mutex_lock(&p->mutex);
+
+	if (!p->n_pdds) {
+		pr_err("No pdd for given process\n");
+		ret = -ENODEV;
+		goto err_unlock;
+	}
+
+	ret = kfd_process_evict_queues(p, KFD_QUEUE_EVICTION_CRIU_CHECKPOINT);
+	if (ret)
+		goto err_unlock;
+
+	p->queues_paused = true;
+
+	args->pid = task_pid_nr_ns(p->lead_thread,
+					task_active_pid_ns(p->lead_thread));
+
+	ret = criu_get_process_object_info(p, &args->num_devices, &args->num_bos,
+					   &args->num_objects, &args->priv_data_size);
+	if (ret)
+		goto err_unlock;
+
+	dev_dbg(kfd_device, "Num of devices:%u bos:%u objects:%u priv_data_size:%lld\n",
+				args->num_devices, args->num_bos, args->num_objects,
+				args->priv_data_size);
+
+err_unlock:
+	if (ret) {
+		kfd_process_restore_queues(p);
+		p->queues_paused = false;
+	}
+	mutex_unlock(&p->mutex);
+	return ret;
+}
+
+static int kfd_ioctl_criu(struct file *filep, struct kfd_process *p, void *data)
+{
+	struct kfd_ioctl_criu_args *args = data;
+	int ret;
+
+	dev_dbg(kfd_device, "CRIU operation: %d\n", args->op);
+	switch (args->op) {
+	case KFD_CRIU_OP_PROCESS_INFO:
+		ret = criu_process_info(filep, p, args);
+		break;
+	case KFD_CRIU_OP_CHECKPOINT:
+		ret = criu_checkpoint(filep, p, args);
+		break;
+	case KFD_CRIU_OP_UNPAUSE:
+		ret = criu_unpause(filep, p, args);
+		break;
+	case KFD_CRIU_OP_RESTORE:
+		ret = criu_restore(filep, p, args);
+		break;
+	case KFD_CRIU_OP_RESUME:
+		ret = criu_resume(filep, p, args);
+		break;
+	default:
+		dev_dbg(kfd_device, "Unsupported CRIU operation:%d\n", args->op);
+		ret = -EINVAL;
+		break;
+	}
+
+	if (ret)
+		dev_dbg(kfd_device, "CRIU operation:%d err:%d\n", args->op, ret);
+
+	return ret;
+}
+
+static int runtime_enable(struct kfd_process *p, uint64_t r_debug,
+			bool enable_ttmp_setup)
+{
+	int i = 0, ret = 0;
+
+	if (p->is_runtime_retry)
+		goto retry;
+
+	if (p->runtime_info.runtime_state != DEBUG_RUNTIME_STATE_DISABLED)
+		return -EBUSY;
+
+	for (i = 0; i < p->n_pdds; i++) {
+		struct kfd_process_device *pdd = p->pdds[i];
+
+		if (pdd->qpd.queue_count)
+			return -EEXIST;
+
+		/*
+		 * Setup TTMPs by default.
+		 * Note that this call must remain here for MES ADD QUEUE to
+		 * skip_process_ctx_clear unconditionally as the first call to
+		 * SET_SHADER_DEBUGGER clears any stale process context data
+		 * saved in MES.
+		 */
+		if (pdd->dev->kfd->shared_resources.enable_mes)
+			kfd_dbg_set_mes_debug_mode(pdd, !kfd_dbg_has_cwsr_workaround(pdd->dev));
+	}
+
+	p->runtime_info.runtime_state = DEBUG_RUNTIME_STATE_ENABLED;
+	p->runtime_info.r_debug = r_debug;
+	p->runtime_info.ttmp_setup = enable_ttmp_setup;
+
+	if (p->runtime_info.ttmp_setup) {
+		for (i = 0; i < p->n_pdds; i++) {
+			struct kfd_process_device *pdd = p->pdds[i];
+
+			if (!kfd_dbg_is_rlc_restore_supported(pdd->dev)) {
+				amdgpu_gfx_off_ctrl(pdd->dev->adev, false);
+				pdd->dev->kfd2kgd->enable_debug_trap(
+						pdd->dev->adev,
+						true,
+						pdd->dev->vm_info.last_vmid_kfd);
+			} else if (kfd_dbg_is_per_vmid_supported(pdd->dev)) {
+				pdd->spi_dbg_override = pdd->dev->kfd2kgd->enable_debug_trap(
+						pdd->dev->adev,
+						false,
+						0);
+			}
+		}
+	}
+
+retry:
+	if (p->debug_trap_enabled) {
+		if (!p->is_runtime_retry) {
+			kfd_dbg_trap_activate(p);
+			kfd_dbg_ev_raise(KFD_EC_MASK(EC_PROCESS_RUNTIME),
+					p, NULL, 0, false, NULL, 0);
+		}
+
+		mutex_unlock(&p->mutex);
+		ret = down_interruptible(&p->runtime_enable_sema);
+		mutex_lock(&p->mutex);
+
+		p->is_runtime_retry = !!ret;
+	}
+
+	return ret;
+}
+
+static int runtime_disable(struct kfd_process *p)
+{
+	int i = 0, ret = 0;
+	bool was_enabled = p->runtime_info.runtime_state == DEBUG_RUNTIME_STATE_ENABLED;
+
+	p->runtime_info.runtime_state = DEBUG_RUNTIME_STATE_DISABLED;
+	p->runtime_info.r_debug = 0;
+
+	if (p->debug_trap_enabled) {
+		if (was_enabled)
+			kfd_dbg_trap_deactivate(p, false, 0);
+
+		if (!p->is_runtime_retry)
+			kfd_dbg_ev_raise(KFD_EC_MASK(EC_PROCESS_RUNTIME),
+					p, NULL, 0, false, NULL, 0);
+
+		mutex_unlock(&p->mutex);
+		ret = down_interruptible(&p->runtime_enable_sema);
+		mutex_lock(&p->mutex);
+
+		p->is_runtime_retry = !!ret;
+		if (ret)
+			return ret;
+	}
+
+	if (was_enabled && p->runtime_info.ttmp_setup) {
+		for (i = 0; i < p->n_pdds; i++) {
+			struct kfd_process_device *pdd = p->pdds[i];
+
+			if (!kfd_dbg_is_rlc_restore_supported(pdd->dev))
+				amdgpu_gfx_off_ctrl(pdd->dev->adev, true);
+		}
+	}
+
+	p->runtime_info.ttmp_setup = false;
+
+	/* disable ttmp setup */
+	for (i = 0; i < p->n_pdds; i++) {
+		struct kfd_process_device *pdd = p->pdds[i];
+		int last_err = 0;
+
+		if (kfd_dbg_is_per_vmid_supported(pdd->dev)) {
+			pdd->spi_dbg_override =
+					pdd->dev->kfd2kgd->disable_debug_trap(
+					pdd->dev->adev,
+					false,
+					pdd->dev->vm_info.last_vmid_kfd);
+
+			if (!pdd->dev->kfd->shared_resources.enable_mes)
+				last_err = debug_refresh_runlist(pdd->dev->dqm);
+			else
+				last_err = kfd_dbg_set_mes_debug_mode(pdd,
+							   !kfd_dbg_has_cwsr_workaround(pdd->dev));
+
+			if (last_err)
+				ret = last_err;
+		}
+	}
+
+	return ret;
+}
+
+static int kfd_ioctl_runtime_enable(struct file *filep, struct kfd_process *p, void *data)
+{
+	struct kfd_ioctl_runtime_enable_args *args = data;
+	int r;
+
+	mutex_lock(&p->mutex);
+
+	if (args->mode_mask & KFD_RUNTIME_ENABLE_MODE_ENABLE_MASK)
+		r = runtime_enable(p, args->r_debug,
+				!!(args->mode_mask & KFD_RUNTIME_ENABLE_MODE_TTMP_SAVE_MASK));
+	else
+		r = runtime_disable(p);
+
+	mutex_unlock(&p->mutex);
+
+	return r;
+}
+
+static int kfd_ioctl_set_debug_trap(struct file *filep, struct kfd_process *p, void *data)
+{
+	struct kfd_ioctl_dbg_trap_args *args = data;
+	struct task_struct *thread = NULL;
+	struct mm_struct *mm = NULL;
+	struct pid *pid = NULL;
+	struct kfd_process *target = NULL;
+	struct kfd_process_device *pdd = NULL;
+	int r = 0;
+
+	if (sched_policy == KFD_SCHED_POLICY_NO_HWS) {
+		pr_err("Debugging does not support sched_policy %i", sched_policy);
+		return -EINVAL;
+	}
+
+	pid = find_get_pid(args->pid);
+	if (!pid) {
+		pr_debug("Cannot find pid info for %i\n", args->pid);
+		r = -ESRCH;
+		goto out;
+	}
+
+	thread = get_pid_task(pid, PIDTYPE_PID);
+	if (!thread) {
+		r = -ESRCH;
+		goto out;
+	}
+
+	mm = get_task_mm(thread);
+	if (!mm) {
+		r = -ESRCH;
+		goto out;
+	}
+
+	if (args->op == KFD_IOC_DBG_TRAP_ENABLE) {
+		bool create_process;
+
+		rcu_read_lock();
+		create_process = thread && thread != current && ptrace_parent(thread) == current;
+		rcu_read_unlock();
+
+		target = create_process ? kfd_create_process(thread) :
+					kfd_lookup_process_by_pid(pid);
+	} else {
+		target = kfd_lookup_process_by_pid(pid);
+	}
+
+	if (IS_ERR_OR_NULL(target)) {
+		pr_debug("Cannot find process PID %i to debug\n", args->pid);
+		r = target ? PTR_ERR(target) : -ESRCH;
+		target = NULL;
+		goto out;
+	}
+
+	/* Check if target is still PTRACED. */
+	rcu_read_lock();
+	if (target != p && args->op != KFD_IOC_DBG_TRAP_DISABLE
+				&& ptrace_parent(target->lead_thread) != current) {
+		pr_err("PID %i is not PTRACED and cannot be debugged\n", args->pid);
+		r = -EPERM;
+	}
+	rcu_read_unlock();
+
+	if (r)
+		goto out;
+
+	mutex_lock(&target->mutex);
+
+	if (args->op != KFD_IOC_DBG_TRAP_ENABLE && !target->debug_trap_enabled) {
+		pr_err("PID %i not debug enabled for op %i\n", args->pid, args->op);
+		r = -EINVAL;
+		goto unlock_out;
+	}
+
+	if (target->runtime_info.runtime_state != DEBUG_RUNTIME_STATE_ENABLED &&
+			(args->op == KFD_IOC_DBG_TRAP_SET_WAVE_LAUNCH_OVERRIDE ||
+			 args->op == KFD_IOC_DBG_TRAP_SET_WAVE_LAUNCH_MODE ||
+			 args->op == KFD_IOC_DBG_TRAP_SUSPEND_QUEUES ||
+			 args->op == KFD_IOC_DBG_TRAP_RESUME_QUEUES ||
+			 args->op == KFD_IOC_DBG_TRAP_SET_NODE_ADDRESS_WATCH ||
+			 args->op == KFD_IOC_DBG_TRAP_CLEAR_NODE_ADDRESS_WATCH ||
+			 args->op == KFD_IOC_DBG_TRAP_SET_FLAGS)) {
+		r = -EPERM;
+		goto unlock_out;
+	}
+
+	if (args->op == KFD_IOC_DBG_TRAP_SET_NODE_ADDRESS_WATCH ||
+	    args->op == KFD_IOC_DBG_TRAP_CLEAR_NODE_ADDRESS_WATCH) {
+		int user_gpu_id = kfd_process_get_user_gpu_id(target,
+				args->op == KFD_IOC_DBG_TRAP_SET_NODE_ADDRESS_WATCH ?
+					args->set_node_address_watch.gpu_id :
+					args->clear_node_address_watch.gpu_id);
+
+		pdd = kfd_process_device_data_by_id(target, user_gpu_id);
+		if (user_gpu_id == -EINVAL || !pdd) {
+			r = -ENODEV;
+			goto unlock_out;
+		}
+	}
+
+	switch (args->op) {
+	case KFD_IOC_DBG_TRAP_ENABLE:
+		if (target != p)
+			target->debugger_process = p;
+
+		r = kfd_dbg_trap_enable(target,
+					args->enable.dbg_fd,
+					(void __user *)args->enable.rinfo_ptr,
+					&args->enable.rinfo_size);
+		if (!r)
+			target->exception_enable_mask = args->enable.exception_mask;
+
+		break;
+	case KFD_IOC_DBG_TRAP_DISABLE:
+		r = kfd_dbg_trap_disable(target);
+		break;
+	case KFD_IOC_DBG_TRAP_SEND_RUNTIME_EVENT:
+		r = kfd_dbg_send_exception_to_runtime(target,
+				args->send_runtime_event.gpu_id,
+				args->send_runtime_event.queue_id,
+				args->send_runtime_event.exception_mask);
+		break;
+	case KFD_IOC_DBG_TRAP_SET_EXCEPTIONS_ENABLED:
+		kfd_dbg_set_enabled_debug_exception_mask(target,
+				args->set_exceptions_enabled.exception_mask);
+		break;
+	case KFD_IOC_DBG_TRAP_SET_WAVE_LAUNCH_OVERRIDE:
+		r = kfd_dbg_trap_set_wave_launch_override(target,
+				args->launch_override.override_mode,
+				args->launch_override.enable_mask,
+				args->launch_override.support_request_mask,
+				&args->launch_override.enable_mask,
+				&args->launch_override.support_request_mask);
+		break;
+	case KFD_IOC_DBG_TRAP_SET_WAVE_LAUNCH_MODE:
+		r = kfd_dbg_trap_set_wave_launch_mode(target,
+				args->launch_mode.launch_mode);
+		break;
+	case KFD_IOC_DBG_TRAP_SUSPEND_QUEUES:
+		r = suspend_queues(target,
+				args->suspend_queues.num_queues,
+				args->suspend_queues.grace_period,
+				args->suspend_queues.exception_mask,
+				(uint32_t *)args->suspend_queues.queue_array_ptr);
+
+		break;
+	case KFD_IOC_DBG_TRAP_RESUME_QUEUES:
+		r = resume_queues(target, args->resume_queues.num_queues,
+				(uint32_t *)args->resume_queues.queue_array_ptr);
+		break;
+	case KFD_IOC_DBG_TRAP_SET_NODE_ADDRESS_WATCH:
+		r = kfd_dbg_trap_set_dev_address_watch(pdd,
+				args->set_node_address_watch.address,
+				args->set_node_address_watch.mask,
+				&args->set_node_address_watch.id,
+				args->set_node_address_watch.mode);
+		break;
+	case KFD_IOC_DBG_TRAP_CLEAR_NODE_ADDRESS_WATCH:
+		r = kfd_dbg_trap_clear_dev_address_watch(pdd,
+				args->clear_node_address_watch.id);
+		break;
+	case KFD_IOC_DBG_TRAP_SET_FLAGS:
+		r = kfd_dbg_trap_set_flags(target, &args->set_flags.flags);
+		break;
+	case KFD_IOC_DBG_TRAP_QUERY_DEBUG_EVENT:
+		r = kfd_dbg_ev_query_debug_event(target,
+				&args->query_debug_event.queue_id,
+				&args->query_debug_event.gpu_id,
+				args->query_debug_event.exception_mask,
+				&args->query_debug_event.exception_mask);
+		break;
+	case KFD_IOC_DBG_TRAP_QUERY_EXCEPTION_INFO:
+		r = kfd_dbg_trap_query_exception_info(target,
+				args->query_exception_info.source_id,
+				args->query_exception_info.exception_code,
+				args->query_exception_info.clear_exception,
+				(void __user *)args->query_exception_info.info_ptr,
+				&args->query_exception_info.info_size);
+		break;
+	case KFD_IOC_DBG_TRAP_GET_QUEUE_SNAPSHOT:
+		r = pqm_get_queue_snapshot(&target->pqm,
+				args->queue_snapshot.exception_mask,
+				(void __user *)args->queue_snapshot.snapshot_buf_ptr,
+				&args->queue_snapshot.num_queues,
+				&args->queue_snapshot.entry_size);
+		break;
+	case KFD_IOC_DBG_TRAP_GET_DEVICE_SNAPSHOT:
+		r = kfd_dbg_trap_device_snapshot(target,
+				args->device_snapshot.exception_mask,
+				(void __user *)args->device_snapshot.snapshot_buf_ptr,
+				&args->device_snapshot.num_devices,
+				&args->device_snapshot.entry_size);
+		break;
+	default:
+		pr_err("Invalid option: %i\n", args->op);
+		r = -EINVAL;
+	}
+
+unlock_out:
+	mutex_unlock(&target->mutex);
+
+out:
+	if (thread)
+		put_task_struct(thread);
+
+	if (mm)
+		mmput(mm);
+
+	if (pid)
+		put_pid(pid);
+
+	if (target)
+		kfd_unref_process(target);
+
+	return r;
+}
+
 #define AMDKFD_IOCTL_DEF(ioctl, _func, _flags) \
 	[_IOC_NR(ioctl)] = {.cmd = ioctl, .func = _func, .flags = _flags, \
 			    .cmd_drv = 0, .name = #ioctl}
@@ -1898,16 +3163,16 @@ static const struct amdkfd_ioctl_desc amdkfd_ioctls[] = {
 	AMDKFD_IOCTL_DEF(AMDKFD_IOC_WAIT_EVENTS,
 			kfd_ioctl_wait_events, 0),
 
-	AMDKFD_IOCTL_DEF(AMDKFD_IOC_DBG_REGISTER,
+	AMDKFD_IOCTL_DEF(AMDKFD_IOC_DBG_REGISTER_DEPRECATED,
 			kfd_ioctl_dbg_register, 0),
 
-	AMDKFD_IOCTL_DEF(AMDKFD_IOC_DBG_UNREGISTER,
+	AMDKFD_IOCTL_DEF(AMDKFD_IOC_DBG_UNREGISTER_DEPRECATED,
 			kfd_ioctl_dbg_unregister, 0),
 
-	AMDKFD_IOCTL_DEF(AMDKFD_IOC_DBG_ADDRESS_WATCH,
+	AMDKFD_IOCTL_DEF(AMDKFD_IOC_DBG_ADDRESS_WATCH_DEPRECATED,
 			kfd_ioctl_dbg_address_watch, 0),
 
-	AMDKFD_IOCTL_DEF(AMDKFD_IOC_DBG_WAVE_CONTROL,
+	AMDKFD_IOCTL_DEF(AMDKFD_IOC_DBG_WAVE_CONTROL_DEPRECATED,
 			kfd_ioctl_dbg_wave_control, 0),
 
 	AMDKFD_IOCTL_DEF(AMDKFD_IOC_SET_SCRATCH_BACKING_VA,
@@ -1959,6 +3224,21 @@ static const struct amdkfd_ioctl_desc amdkfd_ioctls[] = {
 
 	AMDKFD_IOCTL_DEF(AMDKFD_IOC_SET_XNACK_MODE,
 			kfd_ioctl_set_xnack_mode, 0),
+
+	AMDKFD_IOCTL_DEF(AMDKFD_IOC_CRIU_OP,
+			kfd_ioctl_criu, KFD_IOC_FLAG_CHECKPOINT_RESTORE),
+
+	AMDKFD_IOCTL_DEF(AMDKFD_IOC_AVAILABLE_MEMORY,
+			kfd_ioctl_get_available_memory, 0),
+
+	AMDKFD_IOCTL_DEF(AMDKFD_IOC_EXPORT_DMABUF,
+				kfd_ioctl_export_dmabuf, 0),
+
+	AMDKFD_IOCTL_DEF(AMDKFD_IOC_RUNTIME_ENABLE,
+			kfd_ioctl_runtime_enable, 0),
+
+	AMDKFD_IOCTL_DEF(AMDKFD_IOC_DBG_TRAP,
+			kfd_ioctl_set_debug_trap, 0),
 };
 
 #define AMDKFD_CORE_IOCTL_COUNT	ARRAY_SIZE(amdkfd_ioctls)
@@ -1973,9 +3253,12 @@ static long kfd_ioctl(struct file *filep, unsigned int cmd, unsigned long arg)
 	char *kdata = NULL;
 	unsigned int usize, asize;
 	int retcode = -EINVAL;
+	bool ptrace_attached = false;
 
-	if (nr >= AMDKFD_CORE_IOCTL_COUNT)
+	if (nr >= AMDKFD_CORE_IOCTL_COUNT) {
+		retcode = -ENOTTY;
 		goto err_i1;
+	}
 
 	if ((nr >= AMDKFD_COMMAND_START) && (nr < AMDKFD_COMMAND_END)) {
 		u32 amdkfd_size;
@@ -1988,8 +3271,10 @@ static long kfd_ioctl(struct file *filep, unsigned int cmd, unsigned long arg)
 			asize = amdkfd_size;
 
 		cmd = ioctl->cmd;
-	} else
+	} else {
+		retcode = -ENOTTY;
 		goto err_i1;
+	}
 
 	dev_dbg(kfd_device, "ioctl cmd 0x%x (#0x%x), arg 0x%lx\n", cmd, nr, arg);
 
@@ -1998,7 +3283,15 @@ static long kfd_ioctl(struct file *filep, unsigned int cmd, unsigned long arg)
 	 * processes need to create their own KFD device context.
 	 */
 	process = filep->private_data;
-	if (process->lead_thread != current->group_leader) {
+
+	rcu_read_lock();
+	if ((ioctl->flags & KFD_IOC_FLAG_CHECKPOINT_RESTORE) &&
+	    ptrace_parent(process->lead_thread) == current)
+		ptrace_attached = true;
+	rcu_read_unlock();
+
+	if (process->lead_thread != current->group_leader
+	    && !ptrace_attached) {
 		dev_dbg(kfd_device, "Using KFD FD in wrong process\n");
 		retcode = -EBADF;
 		goto err_i1;
@@ -2013,6 +3306,19 @@ static long kfd_ioctl(struct file *filep, unsigned int cmd, unsigned long arg)
 		goto err_i1;
 	}
 
+	/*
+	 * Versions of docker shipped in Ubuntu 18.xx and 20.xx do not support
+	 * CAP_CHECKPOINT_RESTORE, so we also allow access if CAP_SYS_ADMIN as CAP_SYS_ADMIN is a
+	 * more priviledged access.
+	 */
+	if (unlikely(ioctl->flags & KFD_IOC_FLAG_CHECKPOINT_RESTORE)) {
+		if (!capable(CAP_CHECKPOINT_RESTORE) &&
+						!capable(CAP_SYS_ADMIN)) {
+			retcode = -EACCES;
+			goto err_i1;
+		}
+	}
+
 	if (cmd & (IOC_IN | IOC_OUT)) {
 		if (asize <= sizeof(stack_kdata)) {
 			kdata = stack_kdata;
@@ -2057,43 +3363,44 @@ err_i1:
 	return retcode;
 }
 
-static int kfd_mmio_mmap(struct kfd_dev *dev, struct kfd_process *process,
+static int kfd_mmio_mmap(struct kfd_node *dev, struct kfd_process *process,
 		      struct vm_area_struct *vma)
 {
 	phys_addr_t address;
-	int ret;
 
 	if (vma->vm_end - vma->vm_start != PAGE_SIZE)
 		return -EINVAL;
 
-	address = amdgpu_amdkfd_get_mmio_remap_phys_addr(dev->kgd);
+	if (PAGE_SIZE > 4096)
+		return -EINVAL;
+
+	address = dev->adev->rmmio_remap.bus_addr;
 
-	vma->vm_flags |= VM_IO | VM_DONTCOPY | VM_DONTEXPAND | VM_NORESERVE |
-				VM_DONTDUMP | VM_PFNMAP;
+	vm_flags_set(vma, VM_IO | VM_DONTCOPY | VM_DONTEXPAND | VM_NORESERVE |
+				VM_DONTDUMP | VM_PFNMAP);
 
 	vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot);
 
-	pr_debug("pasid 0x%x mapping mmio page\n"
+	pr_debug("process pid %d mapping mmio page\n"
 		 "     target user address == 0x%08llX\n"
 		 "     physical address    == 0x%08llX\n"
 		 "     vm_flags            == 0x%04lX\n"
 		 "     size                == 0x%04lX\n",
-		 process->pasid, (unsigned long long) vma->vm_start,
+		 process->lead_thread->pid, (unsigned long long) vma->vm_start,
 		 address, vma->vm_flags, PAGE_SIZE);
 
-	ret = io_remap_pfn_range(vma,
+	return io_remap_pfn_range(vma,
 				vma->vm_start,
 				address >> PAGE_SHIFT,
 				PAGE_SIZE,
 				vma->vm_page_prot);
-	return ret;
 }
 
 
 static int kfd_mmap(struct file *filp, struct vm_area_struct *vma)
 {
 	struct kfd_process *process;
-	struct kfd_dev *dev = NULL;
+	struct kfd_node *dev = NULL;
 	unsigned long mmap_offset;
 	unsigned int gpu_id;
 
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_crat.c b/drivers/gpu/drm/amd/amdkfd/kfd_crat.c
index cfedfb1e8596..4a7180b46b71 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_crat.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_crat.c
@@ -1,5 +1,6 @@
+// SPDX-License-Identifier: GPL-2.0 OR MIT
 /*
- * Copyright 2015-2017 Advanced Micro Devices, Inc.
+ * Copyright 2015-2022 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -25,9 +26,9 @@
 #include "kfd_crat.h"
 #include "kfd_priv.h"
 #include "kfd_topology.h"
-#include "kfd_iommu.h"
 #include "amdgpu.h"
 #include "amdgpu_amdkfd.h"
+#include "amdgpu_xgmi.h"
 
 /* GPU Processor ID base for dGPUs for which VCRAT needs to be created.
  * GPU processor ID are expressed with Bit[31]=1.
@@ -49,22 +50,13 @@ static inline unsigned int get_and_inc_gpu_processor_id(
 	return current_id;
 }
 
-/* Static table to describe GPU Cache information */
-struct kfd_gpu_cache_info {
-	uint32_t	cache_size;
-	uint32_t	cache_level;
-	uint32_t	flags;
-	/* Indicates how many Compute Units share this cache
-	 * within a SA. Value = 1 indicates the cache is not shared
-	 */
-	uint32_t	num_cu_shared;
-};
 
 static struct kfd_gpu_cache_info kaveri_cache_info[] = {
 	{
 		/* TCP L1 Cache per CU */
 		.cache_size = 16,
 		.cache_level = 1,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -74,6 +66,7 @@ static struct kfd_gpu_cache_info kaveri_cache_info[] = {
 		/* Scalar L1 Instruction Cache (in SQC module) per bank */
 		.cache_size = 16,
 		.cache_level = 1,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_INST_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -83,6 +76,7 @@ static struct kfd_gpu_cache_info kaveri_cache_info[] = {
 		/* Scalar L1 Data Cache (in SQC module) per bank */
 		.cache_size = 8,
 		.cache_level = 1,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -98,6 +92,7 @@ static struct kfd_gpu_cache_info carrizo_cache_info[] = {
 		/* TCP L1 Cache per CU */
 		.cache_size = 16,
 		.cache_level = 1,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -105,8 +100,9 @@ static struct kfd_gpu_cache_info carrizo_cache_info[] = {
 	},
 	{
 		/* Scalar L1 Instruction Cache (in SQC module) per bank */
-		.cache_size = 8,
+		.cache_size = 32,
 		.cache_level = 1,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_INST_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -114,8 +110,9 @@ static struct kfd_gpu_cache_info carrizo_cache_info[] = {
 	},
 	{
 		/* Scalar L1 Data Cache (in SQC module) per bank. */
-		.cache_size = 4,
+		.cache_size = 16,
 		.cache_level = 1,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -145,6 +142,7 @@ static struct kfd_gpu_cache_info vega10_cache_info[] = {
 		/* TCP L1 Cache per CU */
 		.cache_size = 16,
 		.cache_level = 1,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -154,6 +152,7 @@ static struct kfd_gpu_cache_info vega10_cache_info[] = {
 		/* Scalar L1 Instruction Cache per SQC */
 		.cache_size = 32,
 		.cache_level = 1,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_INST_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -163,6 +162,7 @@ static struct kfd_gpu_cache_info vega10_cache_info[] = {
 		/* Scalar L1 Data Cache per SQC */
 		.cache_size = 16,
 		.cache_level = 1,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -172,6 +172,7 @@ static struct kfd_gpu_cache_info vega10_cache_info[] = {
 		/* L2 Data Cache per GPU (Total Tex Cache) */
 		.cache_size = 4096,
 		.cache_level = 2,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -184,6 +185,7 @@ static struct kfd_gpu_cache_info raven_cache_info[] = {
 		/* TCP L1 Cache per CU */
 		.cache_size = 16,
 		.cache_level = 1,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -193,6 +195,7 @@ static struct kfd_gpu_cache_info raven_cache_info[] = {
 		/* Scalar L1 Instruction Cache per SQC */
 		.cache_size = 32,
 		.cache_level = 1,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_INST_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -202,6 +205,7 @@ static struct kfd_gpu_cache_info raven_cache_info[] = {
 		/* Scalar L1 Data Cache per SQC */
 		.cache_size = 16,
 		.cache_level = 1,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -211,6 +215,7 @@ static struct kfd_gpu_cache_info raven_cache_info[] = {
 		/* L2 Data Cache per GPU (Total Tex Cache) */
 		.cache_size = 1024,
 		.cache_level = 2,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -223,6 +228,7 @@ static struct kfd_gpu_cache_info renoir_cache_info[] = {
 		/* TCP L1 Cache per CU */
 		.cache_size = 16,
 		.cache_level = 1,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -232,6 +238,7 @@ static struct kfd_gpu_cache_info renoir_cache_info[] = {
 		/* Scalar L1 Instruction Cache per SQC */
 		.cache_size = 32,
 		.cache_level = 1,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_INST_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -241,6 +248,7 @@ static struct kfd_gpu_cache_info renoir_cache_info[] = {
 		/* Scalar L1 Data Cache per SQC */
 		.cache_size = 16,
 		.cache_level = 1,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -250,6 +258,7 @@ static struct kfd_gpu_cache_info renoir_cache_info[] = {
 		/* L2 Data Cache per GPU (Total Tex Cache) */
 		.cache_size = 1024,
 		.cache_level = 2,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -262,6 +271,7 @@ static struct kfd_gpu_cache_info vega12_cache_info[] = {
 		/* TCP L1 Cache per CU */
 		.cache_size = 16,
 		.cache_level = 1,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -271,6 +281,7 @@ static struct kfd_gpu_cache_info vega12_cache_info[] = {
 		/* Scalar L1 Instruction Cache per SQC */
 		.cache_size = 32,
 		.cache_level = 1,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_INST_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -280,6 +291,7 @@ static struct kfd_gpu_cache_info vega12_cache_info[] = {
 		/* Scalar L1 Data Cache per SQC */
 		.cache_size = 16,
 		.cache_level = 1,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -289,6 +301,7 @@ static struct kfd_gpu_cache_info vega12_cache_info[] = {
 		/* L2 Data Cache per GPU (Total Tex Cache) */
 		.cache_size = 2048,
 		.cache_level = 2,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -301,6 +314,7 @@ static struct kfd_gpu_cache_info vega20_cache_info[] = {
 		/* TCP L1 Cache per CU */
 		.cache_size = 16,
 		.cache_level = 1,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -310,6 +324,7 @@ static struct kfd_gpu_cache_info vega20_cache_info[] = {
 		/* Scalar L1 Instruction Cache per SQC */
 		.cache_size = 32,
 		.cache_level = 1,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_INST_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -319,6 +334,7 @@ static struct kfd_gpu_cache_info vega20_cache_info[] = {
 		/* Scalar L1 Data Cache per SQC */
 		.cache_size = 16,
 		.cache_level = 1,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -328,6 +344,7 @@ static struct kfd_gpu_cache_info vega20_cache_info[] = {
 		/* L2 Data Cache per GPU (Total Tex Cache) */
 		.cache_size = 8192,
 		.cache_level = 2,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -340,6 +357,7 @@ static struct kfd_gpu_cache_info aldebaran_cache_info[] = {
 		/* TCP L1 Cache per CU */
 		.cache_size = 16,
 		.cache_level = 1,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -349,6 +367,7 @@ static struct kfd_gpu_cache_info aldebaran_cache_info[] = {
 		/* Scalar L1 Instruction Cache per SQC */
 		.cache_size = 32,
 		.cache_level = 1,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_INST_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -358,6 +377,7 @@ static struct kfd_gpu_cache_info aldebaran_cache_info[] = {
 		/* Scalar L1 Data Cache per SQC */
 		.cache_size = 16,
 		.cache_level = 1,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -367,6 +387,7 @@ static struct kfd_gpu_cache_info aldebaran_cache_info[] = {
 		/* L2 Data Cache per GPU (Total Tex Cache) */
 		.cache_size = 8192,
 		.cache_level = 2,
+		.cache_line_size = 128,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -379,6 +400,7 @@ static struct kfd_gpu_cache_info navi10_cache_info[] = {
 		/* TCP L1 Cache per CU */
 		.cache_size = 16,
 		.cache_level = 1,
+		.cache_line_size = 128,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -388,6 +410,7 @@ static struct kfd_gpu_cache_info navi10_cache_info[] = {
 		/* Scalar L1 Instruction Cache per SQC */
 		.cache_size = 32,
 		.cache_level = 1,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_INST_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -397,6 +420,7 @@ static struct kfd_gpu_cache_info navi10_cache_info[] = {
 		/* Scalar L1 Data Cache per SQC */
 		.cache_size = 16,
 		.cache_level = 1,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -406,6 +430,7 @@ static struct kfd_gpu_cache_info navi10_cache_info[] = {
 		/* GL1 Data Cache per SA */
 		.cache_size = 128,
 		.cache_level = 1,
+		.cache_line_size = 128,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -415,6 +440,7 @@ static struct kfd_gpu_cache_info navi10_cache_info[] = {
 		/* L2 Data Cache per GPU (Total Tex Cache) */
 		.cache_size = 4096,
 		.cache_level = 2,
+		.cache_line_size = 128,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -427,6 +453,7 @@ static struct kfd_gpu_cache_info vangogh_cache_info[] = {
 		/* TCP L1 Cache per CU */
 		.cache_size = 16,
 		.cache_level = 1,
+		.cache_line_size = 128,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -436,6 +463,7 @@ static struct kfd_gpu_cache_info vangogh_cache_info[] = {
 		/* Scalar L1 Instruction Cache per SQC */
 		.cache_size = 32,
 		.cache_level = 1,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_INST_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -445,6 +473,7 @@ static struct kfd_gpu_cache_info vangogh_cache_info[] = {
 		/* Scalar L1 Data Cache per SQC */
 		.cache_size = 16,
 		.cache_level = 1,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -454,6 +483,7 @@ static struct kfd_gpu_cache_info vangogh_cache_info[] = {
 		/* GL1 Data Cache per SA */
 		.cache_size = 128,
 		.cache_level = 1,
+		.cache_line_size = 128,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -463,6 +493,7 @@ static struct kfd_gpu_cache_info vangogh_cache_info[] = {
 		/* L2 Data Cache per GPU (Total Tex Cache) */
 		.cache_size = 1024,
 		.cache_level = 2,
+		.cache_line_size = 128,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -475,6 +506,7 @@ static struct kfd_gpu_cache_info navi14_cache_info[] = {
 		/* TCP L1 Cache per CU */
 		.cache_size = 16,
 		.cache_level = 1,
+		.cache_line_size = 128,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -484,6 +516,7 @@ static struct kfd_gpu_cache_info navi14_cache_info[] = {
 		/* Scalar L1 Instruction Cache per SQC */
 		.cache_size = 32,
 		.cache_level = 1,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_INST_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -493,6 +526,7 @@ static struct kfd_gpu_cache_info navi14_cache_info[] = {
 		/* Scalar L1 Data Cache per SQC */
 		.cache_size = 16,
 		.cache_level = 1,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -502,6 +536,7 @@ static struct kfd_gpu_cache_info navi14_cache_info[] = {
 		/* GL1 Data Cache per SA */
 		.cache_size = 128,
 		.cache_level = 1,
+		.cache_line_size = 128,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -511,6 +546,7 @@ static struct kfd_gpu_cache_info navi14_cache_info[] = {
 		/* L2 Data Cache per GPU (Total Tex Cache) */
 		.cache_size = 2048,
 		.cache_level = 2,
+		.cache_line_size = 128,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -523,6 +559,7 @@ static struct kfd_gpu_cache_info sienna_cichlid_cache_info[] = {
 		/* TCP L1 Cache per CU */
 		.cache_size = 16,
 		.cache_level = 1,
+		.cache_line_size = 128,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -532,6 +569,7 @@ static struct kfd_gpu_cache_info sienna_cichlid_cache_info[] = {
 		/* Scalar L1 Instruction Cache per SQC */
 		.cache_size = 32,
 		.cache_level = 1,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_INST_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -541,6 +579,7 @@ static struct kfd_gpu_cache_info sienna_cichlid_cache_info[] = {
 		/* Scalar L1 Data Cache per SQC */
 		.cache_size = 16,
 		.cache_level = 1,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -550,6 +589,7 @@ static struct kfd_gpu_cache_info sienna_cichlid_cache_info[] = {
 		/* GL1 Data Cache per SA */
 		.cache_size = 128,
 		.cache_level = 1,
+		.cache_line_size = 128,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -559,6 +599,7 @@ static struct kfd_gpu_cache_info sienna_cichlid_cache_info[] = {
 		/* L2 Data Cache per GPU (Total Tex Cache) */
 		.cache_size = 4096,
 		.cache_level = 2,
+		.cache_line_size = 128,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -568,6 +609,7 @@ static struct kfd_gpu_cache_info sienna_cichlid_cache_info[] = {
 		/* L3 Data Cache per GPU */
 		.cache_size = 128*1024,
 		.cache_level = 3,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -580,6 +622,7 @@ static struct kfd_gpu_cache_info navy_flounder_cache_info[] = {
 		/* TCP L1 Cache per CU */
 		.cache_size = 16,
 		.cache_level = 1,
+		.cache_line_size = 128,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -589,6 +632,7 @@ static struct kfd_gpu_cache_info navy_flounder_cache_info[] = {
 		/* Scalar L1 Instruction Cache per SQC */
 		.cache_size = 32,
 		.cache_level = 1,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_INST_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -598,6 +642,7 @@ static struct kfd_gpu_cache_info navy_flounder_cache_info[] = {
 		/* Scalar L1 Data Cache per SQC */
 		.cache_size = 16,
 		.cache_level = 1,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -607,6 +652,7 @@ static struct kfd_gpu_cache_info navy_flounder_cache_info[] = {
 		/* GL1 Data Cache per SA */
 		.cache_size = 128,
 		.cache_level = 1,
+		.cache_line_size = 128,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -616,6 +662,7 @@ static struct kfd_gpu_cache_info navy_flounder_cache_info[] = {
 		/* L2 Data Cache per GPU (Total Tex Cache) */
 		.cache_size = 3072,
 		.cache_level = 2,
+		.cache_line_size = 128,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -625,6 +672,7 @@ static struct kfd_gpu_cache_info navy_flounder_cache_info[] = {
 		/* L3 Data Cache per GPU */
 		.cache_size = 96*1024,
 		.cache_level = 3,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -637,6 +685,7 @@ static struct kfd_gpu_cache_info dimgrey_cavefish_cache_info[] = {
 		/* TCP L1 Cache per CU */
 		.cache_size = 16,
 		.cache_level = 1,
+		.cache_line_size = 128,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -646,6 +695,7 @@ static struct kfd_gpu_cache_info dimgrey_cavefish_cache_info[] = {
 		/* Scalar L1 Instruction Cache per SQC */
 		.cache_size = 32,
 		.cache_level = 1,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_INST_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -655,6 +705,7 @@ static struct kfd_gpu_cache_info dimgrey_cavefish_cache_info[] = {
 		/* Scalar L1 Data Cache per SQC */
 		.cache_size = 16,
 		.cache_level = 1,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -664,6 +715,7 @@ static struct kfd_gpu_cache_info dimgrey_cavefish_cache_info[] = {
 		/* GL1 Data Cache per SA */
 		.cache_size = 128,
 		.cache_level = 1,
+		.cache_line_size = 128,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -673,6 +725,7 @@ static struct kfd_gpu_cache_info dimgrey_cavefish_cache_info[] = {
 		/* L2 Data Cache per GPU (Total Tex Cache) */
 		.cache_size = 2048,
 		.cache_level = 2,
+		.cache_line_size = 128,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -682,6 +735,7 @@ static struct kfd_gpu_cache_info dimgrey_cavefish_cache_info[] = {
 		/* L3 Data Cache per GPU */
 		.cache_size = 32*1024,
 		.cache_level = 3,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -694,6 +748,7 @@ static struct kfd_gpu_cache_info beige_goby_cache_info[] = {
 		/* TCP L1 Cache per CU */
 		.cache_size = 16,
 		.cache_level = 1,
+		.cache_line_size = 128,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -703,6 +758,7 @@ static struct kfd_gpu_cache_info beige_goby_cache_info[] = {
 		/* Scalar L1 Instruction Cache per SQC */
 		.cache_size = 32,
 		.cache_level = 1,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_INST_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -712,6 +768,7 @@ static struct kfd_gpu_cache_info beige_goby_cache_info[] = {
 		/* Scalar L1 Data Cache per SQC */
 		.cache_size = 16,
 		.cache_level = 1,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -721,6 +778,7 @@ static struct kfd_gpu_cache_info beige_goby_cache_info[] = {
 		/* GL1 Data Cache per SA */
 		.cache_size = 128,
 		.cache_level = 1,
+		.cache_line_size = 128,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -730,6 +788,7 @@ static struct kfd_gpu_cache_info beige_goby_cache_info[] = {
 		/* L2 Data Cache per GPU (Total Tex Cache) */
 		.cache_size = 1024,
 		.cache_level = 2,
+		.cache_line_size = 128,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -739,6 +798,7 @@ static struct kfd_gpu_cache_info beige_goby_cache_info[] = {
 		/* L3 Data Cache per GPU */
 		.cache_size = 16*1024,
 		.cache_level = 3,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -751,6 +811,7 @@ static struct kfd_gpu_cache_info yellow_carp_cache_info[] = {
 		/* TCP L1 Cache per CU */
 		.cache_size = 16,
 		.cache_level = 1,
+		.cache_line_size = 128,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -760,6 +821,7 @@ static struct kfd_gpu_cache_info yellow_carp_cache_info[] = {
 		/* Scalar L1 Instruction Cache per SQC */
 		.cache_size = 32,
 		.cache_level = 1,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_INST_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -769,6 +831,7 @@ static struct kfd_gpu_cache_info yellow_carp_cache_info[] = {
 		/* Scalar L1 Data Cache per SQC */
 		.cache_size = 16,
 		.cache_level = 1,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -778,6 +841,7 @@ static struct kfd_gpu_cache_info yellow_carp_cache_info[] = {
 		/* GL1 Data Cache per SA */
 		.cache_size = 128,
 		.cache_level = 1,
+		.cache_line_size = 128,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -787,6 +851,166 @@ static struct kfd_gpu_cache_info yellow_carp_cache_info[] = {
 		/* L2 Data Cache per GPU (Total Tex Cache) */
 		.cache_size = 2048,
 		.cache_level = 2,
+		.cache_line_size = 128,
+		.flags = (CRAT_CACHE_FLAGS_ENABLED |
+				CRAT_CACHE_FLAGS_DATA_CACHE |
+				CRAT_CACHE_FLAGS_SIMD_CACHE),
+		.num_cu_shared = 6,
+	},
+};
+
+static struct kfd_gpu_cache_info gfx1037_cache_info[] = {
+	{
+		/* TCP L1 Cache per CU */
+		.cache_size = 16,
+		.cache_level = 1,
+		.cache_line_size = 128,
+		.flags = (CRAT_CACHE_FLAGS_ENABLED |
+				CRAT_CACHE_FLAGS_DATA_CACHE |
+				CRAT_CACHE_FLAGS_SIMD_CACHE),
+		.num_cu_shared = 1,
+	},
+	{
+		/* Scalar L1 Instruction Cache per SQC */
+		.cache_size = 32,
+		.cache_level = 1,
+		.cache_line_size = 64,
+		.flags = (CRAT_CACHE_FLAGS_ENABLED |
+				CRAT_CACHE_FLAGS_INST_CACHE |
+				CRAT_CACHE_FLAGS_SIMD_CACHE),
+		.num_cu_shared = 2,
+	},
+	{
+		/* Scalar L1 Data Cache per SQC */
+		.cache_size = 16,
+		.cache_level = 1,
+		.cache_line_size = 64,
+		.flags = (CRAT_CACHE_FLAGS_ENABLED |
+				CRAT_CACHE_FLAGS_DATA_CACHE |
+				CRAT_CACHE_FLAGS_SIMD_CACHE),
+		.num_cu_shared = 2,
+	},
+	{
+		/* GL1 Data Cache per SA */
+		.cache_size = 128,
+		.cache_level = 1,
+		.cache_line_size = 128,
+		.flags = (CRAT_CACHE_FLAGS_ENABLED |
+				CRAT_CACHE_FLAGS_DATA_CACHE |
+				CRAT_CACHE_FLAGS_SIMD_CACHE),
+		.num_cu_shared = 2,
+	},
+	{
+		/* L2 Data Cache per GPU (Total Tex Cache) */
+		.cache_size = 256,
+		.cache_level = 2,
+		.cache_line_size = 128,
+		.flags = (CRAT_CACHE_FLAGS_ENABLED |
+				CRAT_CACHE_FLAGS_DATA_CACHE |
+				CRAT_CACHE_FLAGS_SIMD_CACHE),
+		.num_cu_shared = 2,
+	},
+};
+
+static struct kfd_gpu_cache_info gc_10_3_6_cache_info[] = {
+	{
+		/* TCP L1 Cache per CU */
+		.cache_size = 16,
+		.cache_level = 1,
+		.cache_line_size = 128,
+		.flags = (CRAT_CACHE_FLAGS_ENABLED |
+			  CRAT_CACHE_FLAGS_DATA_CACHE |
+			  CRAT_CACHE_FLAGS_SIMD_CACHE),
+		.num_cu_shared = 1,
+	},
+	{
+		/* Scalar L1 Instruction Cache per SQC */
+		.cache_size = 32,
+		.cache_level = 1,
+		.cache_line_size = 64,
+		.flags = (CRAT_CACHE_FLAGS_ENABLED |
+			  CRAT_CACHE_FLAGS_INST_CACHE |
+			  CRAT_CACHE_FLAGS_SIMD_CACHE),
+		.num_cu_shared = 2,
+	},
+	{
+		/* Scalar L1 Data Cache per SQC */
+		.cache_size = 16,
+		.cache_level = 1,
+		.cache_line_size = 64,
+		.flags = (CRAT_CACHE_FLAGS_ENABLED |
+			  CRAT_CACHE_FLAGS_DATA_CACHE |
+			  CRAT_CACHE_FLAGS_SIMD_CACHE),
+		.num_cu_shared = 2,
+	},
+	{
+		/* GL1 Data Cache per SA */
+		.cache_size = 128,
+		.cache_level = 1,
+		.cache_line_size = 128,
+		.flags = (CRAT_CACHE_FLAGS_ENABLED |
+			  CRAT_CACHE_FLAGS_DATA_CACHE |
+			  CRAT_CACHE_FLAGS_SIMD_CACHE),
+		.num_cu_shared = 2,
+	},
+	{
+		/* L2 Data Cache per GPU (Total Tex Cache) */
+		.cache_size = 256,
+		.cache_level = 2,
+		.cache_line_size = 128,
+		.flags = (CRAT_CACHE_FLAGS_ENABLED |
+			  CRAT_CACHE_FLAGS_DATA_CACHE |
+			  CRAT_CACHE_FLAGS_SIMD_CACHE),
+		.num_cu_shared = 2,
+	},
+};
+
+static struct kfd_gpu_cache_info dummy_cache_info[] = {
+	{
+		/* TCP L1 Cache per CU */
+		.cache_size = 16,
+		.cache_level = 1,
+		.cache_line_size = 64,
+		.flags = (CRAT_CACHE_FLAGS_ENABLED |
+				CRAT_CACHE_FLAGS_DATA_CACHE |
+				CRAT_CACHE_FLAGS_SIMD_CACHE),
+		.num_cu_shared = 1,
+	},
+	{
+		/* Scalar L1 Instruction Cache per SQC */
+		.cache_size = 32,
+		.cache_level = 1,
+		.cache_line_size = 64,
+		.flags = (CRAT_CACHE_FLAGS_ENABLED |
+				CRAT_CACHE_FLAGS_INST_CACHE |
+				CRAT_CACHE_FLAGS_SIMD_CACHE),
+		.num_cu_shared = 2,
+	},
+	{
+		/* Scalar L1 Data Cache per SQC */
+		.cache_size = 16,
+		.cache_level = 1,
+		.cache_line_size = 64,
+		.flags = (CRAT_CACHE_FLAGS_ENABLED |
+				CRAT_CACHE_FLAGS_DATA_CACHE |
+				CRAT_CACHE_FLAGS_SIMD_CACHE),
+		.num_cu_shared = 2,
+	},
+	{
+		/* GL1 Data Cache per SA */
+		.cache_size = 128,
+		.cache_level = 1,
+		.cache_line_size = 64,
+		.flags = (CRAT_CACHE_FLAGS_ENABLED |
+				CRAT_CACHE_FLAGS_DATA_CACHE |
+				CRAT_CACHE_FLAGS_SIMD_CACHE),
+		.num_cu_shared = 6,
+	},
+	{
+		/* L2 Data Cache per GPU (Total Tex Cache) */
+		.cache_size = 2048,
+		.cache_level = 2,
+		.cache_line_size = 64,
 		.flags = (CRAT_CACHE_FLAGS_ENABLED |
 				CRAT_CACHE_FLAGS_DATA_CACHE |
 				CRAT_CACHE_FLAGS_SIMD_CACHE),
@@ -974,8 +1198,12 @@ static int kfd_parse_subtype_cache(struct crat_subtype_cache *cache,
 			props->cachelines_per_tag = cache->lines_per_tag;
 			props->cache_assoc = cache->associativity;
 			props->cache_latency = cache->cache_latency;
+
 			memcpy(props->sibling_map, cache->sibling_map,
-					sizeof(props->sibling_map));
+					CRAT_SIBLINGMAP_SIZE);
+
+			/* set the sibling_map_size as 32 for CRAT from ACPI */
+			props->sibling_map_size = CRAT_SIBLINGMAP_SIZE;
 
 			if (cache->flags & CRAT_CACHE_FLAGS_DATA_CACHE)
 				props->cache_type |= HSA_CACHE_TYPE_DATA;
@@ -986,7 +1214,6 @@ static int kfd_parse_subtype_cache(struct crat_subtype_cache *cache,
 			if (cache->flags & CRAT_CACHE_FLAGS_SIMD_CACHE)
 				props->cache_type |= HSA_CACHE_TYPE_HSACU;
 
-			dev->cache_count++;
 			dev->node_props.caches_count++;
 			list_add_tail(&props->list, &dev->cache_props);
 
@@ -1028,7 +1255,7 @@ static int kfd_parse_subtype_iolink(struct crat_subtype_iolink *iolink,
 			if (props->iolink_type == CRAT_IOLINK_TYPE_PCIEXPRESS)
 				props->weight = 20;
 			else if (props->iolink_type == CRAT_IOLINK_TYPE_XGMI)
-				props->weight = 15 * iolink->num_hops_xgmi;
+				props->weight = iolink->weight_xgmi;
 			else
 				props->weight = node_distance(id_from, id_to);
 
@@ -1039,7 +1266,6 @@ static int kfd_parse_subtype_iolink(struct crat_subtype_iolink *iolink,
 			props->rec_transfer_size =
 					iolink->recommended_transfer_size;
 
-			dev->io_link_count++;
 			dev->node_props.io_links_count++;
 			list_add_tail(&props->list, &dev->io_link_props);
 			break;
@@ -1055,15 +1281,17 @@ static int kfd_parse_subtype_iolink(struct crat_subtype_iolink *iolink,
 	 * table, add corresponded reversed direction link now.
 	 */
 	if (props && (iolink->flags & CRAT_IOLINK_FLAGS_BI_DIRECTIONAL)) {
-		to_dev = kfd_topology_device_by_proximity_domain(id_to);
+		to_dev = kfd_topology_device_by_proximity_domain_no_lock(id_to);
 		if (!to_dev)
 			return -ENODEV;
 		/* same everything but the other direction */
 		props2 = kmemdup(props, sizeof(*props2), GFP_KERNEL);
+		if (!props2)
+			return -ENOMEM;
+
 		props2->node_from = id_to;
 		props2->node_to = id_from;
 		props2->kobj = NULL;
-		to_dev->io_link_count++;
 		to_dev->node_props.io_links_count++;
 		list_add_tail(&props2->list, &to_dev->io_link_props);
 	}
@@ -1193,394 +1421,310 @@ err:
 	return ret;
 }
 
-/* Helper function. See kfd_fill_gpu_cache_info for parameter description */
-static int fill_in_l1_pcache(struct crat_subtype_cache *pcache,
-				struct kfd_gpu_cache_info *pcache_info,
-				struct kfd_cu_info *cu_info,
-				int mem_available,
-				int cu_bitmask,
-				int cache_type, unsigned int cu_processor_id,
-				int cu_block)
-{
-	unsigned int cu_sibling_map_mask;
-	int first_active_cu;
-
-	/* First check if enough memory is available */
-	if (sizeof(struct crat_subtype_cache) > mem_available)
-		return -ENOMEM;
-
-	cu_sibling_map_mask = cu_bitmask;
-	cu_sibling_map_mask >>= cu_block;
-	cu_sibling_map_mask &=
-		((1 << pcache_info[cache_type].num_cu_shared) - 1);
-	first_active_cu = ffs(cu_sibling_map_mask);
 
-	/* CU could be inactive. In case of shared cache find the first active
-	 * CU. and incase of non-shared cache check if the CU is inactive. If
-	 * inactive active skip it
-	 */
-	if (first_active_cu) {
-		memset(pcache, 0, sizeof(struct crat_subtype_cache));
-		pcache->type = CRAT_SUBTYPE_CACHE_AFFINITY;
-		pcache->length = sizeof(struct crat_subtype_cache);
-		pcache->flags = pcache_info[cache_type].flags;
-		pcache->processor_id_low = cu_processor_id
-					 + (first_active_cu - 1);
-		pcache->cache_level = pcache_info[cache_type].cache_level;
-		pcache->cache_size = pcache_info[cache_type].cache_size;
-
-		/* Sibling map is w.r.t processor_id_low, so shift out
-		 * inactive CU
-		 */
-		cu_sibling_map_mask =
-			cu_sibling_map_mask >> (first_active_cu - 1);
-
-		pcache->sibling_map[0] = (uint8_t)(cu_sibling_map_mask & 0xFF);
-		pcache->sibling_map[1] =
-				(uint8_t)((cu_sibling_map_mask >> 8) & 0xFF);
-		pcache->sibling_map[2] =
-				(uint8_t)((cu_sibling_map_mask >> 16) & 0xFF);
-		pcache->sibling_map[3] =
-				(uint8_t)((cu_sibling_map_mask >> 24) & 0xFF);
-		return 0;
+static int kfd_fill_gpu_cache_info_from_gfx_config(struct kfd_dev *kdev,
+						   bool cache_line_size_missing,
+						   struct kfd_gpu_cache_info *pcache_info)
+{
+	struct amdgpu_device *adev = kdev->adev;
+	int i = 0;
+
+	/* TCP L1 Cache per CU */
+	if (adev->gfx.config.gc_tcp_l1_size) {
+		pcache_info[i].cache_size = adev->gfx.config.gc_tcp_l1_size;
+		pcache_info[i].cache_level = 1;
+		pcache_info[i].flags = (CRAT_CACHE_FLAGS_ENABLED |
+					CRAT_CACHE_FLAGS_DATA_CACHE |
+					CRAT_CACHE_FLAGS_SIMD_CACHE);
+		pcache_info[i].num_cu_shared = adev->gfx.config.gc_num_tcp_per_wpg / 2;
+		pcache_info[i].cache_line_size = adev->gfx.config.gc_tcp_cache_line_size;
+		if (cache_line_size_missing && !pcache_info[i].cache_line_size)
+			pcache_info[i].cache_line_size = 128;
+		i++;
+	}
+	/* Scalar L1 Instruction Cache per SQC */
+	if (adev->gfx.config.gc_l1_instruction_cache_size_per_sqc) {
+		pcache_info[i].cache_size =
+			adev->gfx.config.gc_l1_instruction_cache_size_per_sqc;
+		pcache_info[i].cache_level = 1;
+		pcache_info[i].flags = (CRAT_CACHE_FLAGS_ENABLED |
+					CRAT_CACHE_FLAGS_INST_CACHE |
+					CRAT_CACHE_FLAGS_SIMD_CACHE);
+		pcache_info[i].num_cu_shared = adev->gfx.config.gc_num_sqc_per_wgp * 2;
+		pcache_info[i].cache_line_size = adev->gfx.config.gc_instruction_cache_line_size;
+		if (cache_line_size_missing && !pcache_info[i].cache_line_size)
+			pcache_info[i].cache_line_size = 128;
+		i++;
 	}
-	return 1;
+	/* Scalar L1 Data Cache per SQC */
+	if (adev->gfx.config.gc_l1_data_cache_size_per_sqc) {
+		pcache_info[i].cache_size = adev->gfx.config.gc_l1_data_cache_size_per_sqc;
+		pcache_info[i].cache_level = 1;
+		pcache_info[i].flags = (CRAT_CACHE_FLAGS_ENABLED |
+					CRAT_CACHE_FLAGS_DATA_CACHE |
+					CRAT_CACHE_FLAGS_SIMD_CACHE);
+		pcache_info[i].num_cu_shared = adev->gfx.config.gc_num_sqc_per_wgp * 2;
+		pcache_info[i].cache_line_size = adev->gfx.config.gc_scalar_data_cache_line_size;
+		if (cache_line_size_missing && !pcache_info[i].cache_line_size)
+			pcache_info[i].cache_line_size = 64;
+		i++;
+	}
+	/* GL1 Data Cache per SA */
+	if (adev->gfx.config.gc_gl1c_per_sa &&
+	    adev->gfx.config.gc_gl1c_size_per_instance) {
+		pcache_info[i].cache_size = adev->gfx.config.gc_gl1c_per_sa *
+			adev->gfx.config.gc_gl1c_size_per_instance;
+		pcache_info[i].cache_level = 1;
+		pcache_info[i].flags = (CRAT_CACHE_FLAGS_ENABLED |
+					CRAT_CACHE_FLAGS_DATA_CACHE |
+					CRAT_CACHE_FLAGS_SIMD_CACHE);
+		pcache_info[i].num_cu_shared = adev->gfx.config.max_cu_per_sh;
+		if (cache_line_size_missing)
+			pcache_info[i].cache_line_size = 128;
+		i++;
+	}
+	/* L2 Data Cache per GPU (Total Tex Cache) */
+	if (adev->gfx.config.gc_gl2c_per_gpu) {
+		pcache_info[i].cache_size = adev->gfx.config.gc_gl2c_per_gpu;
+		pcache_info[i].cache_level = 2;
+		pcache_info[i].flags = (CRAT_CACHE_FLAGS_ENABLED |
+					CRAT_CACHE_FLAGS_DATA_CACHE |
+					CRAT_CACHE_FLAGS_SIMD_CACHE);
+		pcache_info[i].num_cu_shared = adev->gfx.config.max_cu_per_sh;
+		pcache_info[i].cache_line_size = adev->gfx.config.gc_tcc_cache_line_size;
+		if (cache_line_size_missing && !pcache_info[i].cache_line_size)
+			pcache_info[i].cache_line_size = 128;
+		i++;
+	}
+	/* L3 Data Cache per GPU */
+	if (adev->gmc.mall_size) {
+		pcache_info[i].cache_size = adev->gmc.mall_size / 1024;
+		pcache_info[i].cache_level = 3;
+		pcache_info[i].flags = (CRAT_CACHE_FLAGS_ENABLED |
+					CRAT_CACHE_FLAGS_DATA_CACHE |
+					CRAT_CACHE_FLAGS_SIMD_CACHE);
+		pcache_info[i].num_cu_shared = adev->gfx.config.max_cu_per_sh;
+		pcache_info[i].cache_line_size = 64;
+		i++;
+	}
+	return i;
 }
 
-/* Helper function. See kfd_fill_gpu_cache_info for parameter description */
-static int fill_in_l2_l3_pcache(struct crat_subtype_cache *pcache,
-				struct kfd_gpu_cache_info *pcache_info,
-				struct kfd_cu_info *cu_info,
-				int mem_available,
-				int cache_type, unsigned int cu_processor_id)
+static int kfd_fill_gpu_cache_info_from_gfx_config_v2(struct kfd_dev *kdev,
+						   struct kfd_gpu_cache_info *pcache_info)
 {
-	unsigned int cu_sibling_map_mask;
-	int first_active_cu;
-	int i, j, k;
-
-	/* First check if enough memory is available */
-	if (sizeof(struct crat_subtype_cache) > mem_available)
-		return -ENOMEM;
-
-	cu_sibling_map_mask = cu_info->cu_bitmap[0][0];
-	cu_sibling_map_mask &=
-		((1 << pcache_info[cache_type].num_cu_shared) - 1);
-	first_active_cu = ffs(cu_sibling_map_mask);
-
-	/* CU could be inactive. In case of shared cache find the first active
-	 * CU. and incase of non-shared cache check if the CU is inactive. If
-	 * inactive active skip it
-	 */
-	if (first_active_cu) {
-		memset(pcache, 0, sizeof(struct crat_subtype_cache));
-		pcache->type = CRAT_SUBTYPE_CACHE_AFFINITY;
-		pcache->length = sizeof(struct crat_subtype_cache);
-		pcache->flags = pcache_info[cache_type].flags;
-		pcache->processor_id_low = cu_processor_id
-					 + (first_active_cu - 1);
-		pcache->cache_level = pcache_info[cache_type].cache_level;
-		pcache->cache_size = pcache_info[cache_type].cache_size;
-
-		/* Sibling map is w.r.t processor_id_low, so shift out
-		 * inactive CU
-		 */
-		cu_sibling_map_mask =
-			cu_sibling_map_mask >> (first_active_cu - 1);
-		k = 0;
-		for (i = 0; i < cu_info->num_shader_engines; i++) {
-			for (j = 0; j < cu_info->num_shader_arrays_per_engine;
-				j++) {
-				pcache->sibling_map[k] =
-				 (uint8_t)(cu_sibling_map_mask & 0xFF);
-				pcache->sibling_map[k+1] =
-				 (uint8_t)((cu_sibling_map_mask >> 8) & 0xFF);
-				pcache->sibling_map[k+2] =
-				 (uint8_t)((cu_sibling_map_mask >> 16) & 0xFF);
-				pcache->sibling_map[k+3] =
-				 (uint8_t)((cu_sibling_map_mask >> 24) & 0xFF);
-				k += 4;
-				cu_sibling_map_mask =
-					cu_info->cu_bitmap[i % 4][j + i / 4];
-				cu_sibling_map_mask &= (
-				 (1 << pcache_info[cache_type].num_cu_shared)
-				 - 1);
-			}
-		}
-		return 0;
+	struct amdgpu_device *adev = kdev->adev;
+	int i = 0;
+
+	/* TCP L1 Cache per CU */
+	if (adev->gfx.config.gc_tcp_size_per_cu) {
+		pcache_info[i].cache_size = adev->gfx.config.gc_tcp_size_per_cu;
+		pcache_info[i].cache_level = 1;
+		/* Cacheline size not available in IP discovery for gc943,gc944 */
+		pcache_info[i].cache_line_size = 128;
+		pcache_info[i].flags = (CRAT_CACHE_FLAGS_ENABLED |
+					CRAT_CACHE_FLAGS_DATA_CACHE |
+					CRAT_CACHE_FLAGS_SIMD_CACHE);
+		pcache_info[i].num_cu_shared = 1;
+		i++;
+	}
+	/* Scalar L1 Instruction Cache per SQC */
+	if (adev->gfx.config.gc_l1_instruction_cache_size_per_sqc) {
+		pcache_info[i].cache_size =
+			adev->gfx.config.gc_l1_instruction_cache_size_per_sqc;
+		pcache_info[i].cache_level = 1;
+		pcache_info[i].cache_line_size = 64;
+		pcache_info[i].flags = (CRAT_CACHE_FLAGS_ENABLED |
+					CRAT_CACHE_FLAGS_INST_CACHE |
+					CRAT_CACHE_FLAGS_SIMD_CACHE);
+		pcache_info[i].num_cu_shared = adev->gfx.config.gc_num_cu_per_sqc;
+		i++;
+	}
+	/* Scalar L1 Data Cache per SQC */
+	if (adev->gfx.config.gc_l1_data_cache_size_per_sqc) {
+		pcache_info[i].cache_size = adev->gfx.config.gc_l1_data_cache_size_per_sqc;
+		pcache_info[i].cache_level = 1;
+		pcache_info[i].cache_line_size = 64;
+		pcache_info[i].flags = (CRAT_CACHE_FLAGS_ENABLED |
+					CRAT_CACHE_FLAGS_DATA_CACHE |
+					CRAT_CACHE_FLAGS_SIMD_CACHE);
+		pcache_info[i].num_cu_shared = adev->gfx.config.gc_num_cu_per_sqc;
+		i++;
 	}
-	return 1;
+	/* L2 Data Cache per GPU (Total Tex Cache) */
+	if (adev->gfx.config.gc_tcc_size) {
+		pcache_info[i].cache_size = adev->gfx.config.gc_tcc_size;
+		pcache_info[i].cache_level = 2;
+		pcache_info[i].cache_line_size = 128;
+		pcache_info[i].flags = (CRAT_CACHE_FLAGS_ENABLED |
+					CRAT_CACHE_FLAGS_DATA_CACHE |
+					CRAT_CACHE_FLAGS_SIMD_CACHE);
+		pcache_info[i].num_cu_shared = adev->gfx.config.max_cu_per_sh;
+		i++;
+	}
+	/* L3 Data Cache per GPU */
+	if (adev->gmc.mall_size) {
+		pcache_info[i].cache_size = adev->gmc.mall_size / 1024;
+		pcache_info[i].cache_level = 3;
+		pcache_info[i].cache_line_size = 64;
+		pcache_info[i].flags = (CRAT_CACHE_FLAGS_ENABLED |
+					CRAT_CACHE_FLAGS_DATA_CACHE |
+					CRAT_CACHE_FLAGS_SIMD_CACHE);
+		pcache_info[i].num_cu_shared = adev->gfx.config.max_cu_per_sh;
+		i++;
+	}
+	return i;
 }
 
-/* kfd_fill_gpu_cache_info - Fill GPU cache info using kfd_gpu_cache_info
- * tables
- *
- *	@kdev - [IN] GPU device
- *	@gpu_processor_id - [IN] GPU processor ID to which these caches
- *			    associate
- *	@available_size - [IN] Amount of memory available in pcache
- *	@cu_info - [IN] Compute Unit info obtained from KGD
- *	@pcache - [OUT] memory into which cache data is to be filled in.
- *	@size_filled - [OUT] amount of data used up in pcache.
- *	@num_of_entries - [OUT] number of caches added
- */
-static int kfd_fill_gpu_cache_info(struct kfd_dev *kdev,
-			int gpu_processor_id,
-			int available_size,
-			struct kfd_cu_info *cu_info,
-			struct crat_subtype_cache *pcache,
-			int *size_filled,
-			int *num_of_entries)
+int kfd_get_gpu_cache_info(struct kfd_node *kdev, struct kfd_gpu_cache_info **pcache_info)
 {
-	struct kfd_gpu_cache_info *pcache_info;
 	int num_of_cache_types = 0;
-	int i, j, k;
-	int ct = 0;
-	int mem_available = available_size;
-	unsigned int cu_processor_id;
-	int ret;
-	unsigned int num_cu_shared;
-
-	switch (kdev->device_info->asic_family) {
+	bool cache_line_size_missing = false;
+
+	switch (kdev->adev->asic_type) {
 	case CHIP_KAVERI:
-		pcache_info = kaveri_cache_info;
+		*pcache_info = kaveri_cache_info;
 		num_of_cache_types = ARRAY_SIZE(kaveri_cache_info);
 		break;
 	case CHIP_HAWAII:
-		pcache_info = hawaii_cache_info;
+		*pcache_info = hawaii_cache_info;
 		num_of_cache_types = ARRAY_SIZE(hawaii_cache_info);
 		break;
 	case CHIP_CARRIZO:
-		pcache_info = carrizo_cache_info;
+		*pcache_info = carrizo_cache_info;
 		num_of_cache_types = ARRAY_SIZE(carrizo_cache_info);
 		break;
 	case CHIP_TONGA:
-		pcache_info = tonga_cache_info;
+		*pcache_info = tonga_cache_info;
 		num_of_cache_types = ARRAY_SIZE(tonga_cache_info);
 		break;
 	case CHIP_FIJI:
-		pcache_info = fiji_cache_info;
+		*pcache_info = fiji_cache_info;
 		num_of_cache_types = ARRAY_SIZE(fiji_cache_info);
 		break;
 	case CHIP_POLARIS10:
-		pcache_info = polaris10_cache_info;
+		*pcache_info = polaris10_cache_info;
 		num_of_cache_types = ARRAY_SIZE(polaris10_cache_info);
 		break;
 	case CHIP_POLARIS11:
-		pcache_info = polaris11_cache_info;
+		*pcache_info = polaris11_cache_info;
 		num_of_cache_types = ARRAY_SIZE(polaris11_cache_info);
 		break;
 	case CHIP_POLARIS12:
-		pcache_info = polaris12_cache_info;
+		*pcache_info = polaris12_cache_info;
 		num_of_cache_types = ARRAY_SIZE(polaris12_cache_info);
 		break;
 	case CHIP_VEGAM:
-		pcache_info = vegam_cache_info;
+		*pcache_info = vegam_cache_info;
 		num_of_cache_types = ARRAY_SIZE(vegam_cache_info);
 		break;
-	case CHIP_VEGA10:
-		pcache_info = vega10_cache_info;
-		num_of_cache_types = ARRAY_SIZE(vega10_cache_info);
-		break;
-	case CHIP_VEGA12:
-		pcache_info = vega12_cache_info;
-		num_of_cache_types = ARRAY_SIZE(vega12_cache_info);
-		break;
-	case CHIP_VEGA20:
-	case CHIP_ARCTURUS:
-		pcache_info = vega20_cache_info;
-		num_of_cache_types = ARRAY_SIZE(vega20_cache_info);
-		break;
-	case CHIP_ALDEBARAN:
-		pcache_info = aldebaran_cache_info;
-		num_of_cache_types = ARRAY_SIZE(aldebaran_cache_info);
-		break;
-	case CHIP_RAVEN:
-		pcache_info = raven_cache_info;
-		num_of_cache_types = ARRAY_SIZE(raven_cache_info);
-		break;
-	case CHIP_RENOIR:
-		pcache_info = renoir_cache_info;
-		num_of_cache_types = ARRAY_SIZE(renoir_cache_info);
-		break;
-	case CHIP_NAVI10:
-	case CHIP_NAVI12:
-	case CHIP_CYAN_SKILLFISH:
-		pcache_info = navi10_cache_info;
-		num_of_cache_types = ARRAY_SIZE(navi10_cache_info);
-		break;
-	case CHIP_NAVI14:
-		pcache_info = navi14_cache_info;
-		num_of_cache_types = ARRAY_SIZE(navi14_cache_info);
-		break;
-	case CHIP_SIENNA_CICHLID:
-		pcache_info = sienna_cichlid_cache_info;
-		num_of_cache_types = ARRAY_SIZE(sienna_cichlid_cache_info);
-		break;
-	case CHIP_NAVY_FLOUNDER:
-		pcache_info = navy_flounder_cache_info;
-		num_of_cache_types = ARRAY_SIZE(navy_flounder_cache_info);
-		break;
-	case CHIP_DIMGREY_CAVEFISH:
-		pcache_info = dimgrey_cavefish_cache_info;
-		num_of_cache_types = ARRAY_SIZE(dimgrey_cavefish_cache_info);
-		break;
-	case CHIP_VANGOGH:
-		pcache_info = vangogh_cache_info;
-		num_of_cache_types = ARRAY_SIZE(vangogh_cache_info);
-		break;
-	case CHIP_BEIGE_GOBY:
-		pcache_info = beige_goby_cache_info;
-		num_of_cache_types = ARRAY_SIZE(beige_goby_cache_info);
-		break;
-	case CHIP_YELLOW_CARP:
-		pcache_info = yellow_carp_cache_info;
-		num_of_cache_types = ARRAY_SIZE(yellow_carp_cache_info);
-		break;
 	default:
-		return -EINVAL;
-	}
-
-	*size_filled = 0;
-	*num_of_entries = 0;
-
-	/* For each type of cache listed in the kfd_gpu_cache_info table,
-	 * go through all available Compute Units.
-	 * The [i,j,k] loop will
-	 *		if kfd_gpu_cache_info.num_cu_shared = 1
-	 *			will parse through all available CU
-	 *		If (kfd_gpu_cache_info.num_cu_shared != 1)
-	 *			then it will consider only one CU from
-	 *			the shared unit
-	 */
-
-	for (ct = 0; ct < num_of_cache_types; ct++) {
-	  cu_processor_id = gpu_processor_id;
-	  if (pcache_info[ct].cache_level == 1) {
-	    for (i = 0; i < cu_info->num_shader_engines; i++) {
-	      for (j = 0; j < cu_info->num_shader_arrays_per_engine; j++) {
-	        for (k = 0; k < cu_info->num_cu_per_sh;
-		  k += pcache_info[ct].num_cu_shared) {
-		  ret = fill_in_l1_pcache(pcache,
-					pcache_info,
-					cu_info,
-					mem_available,
-					cu_info->cu_bitmap[i % 4][j + i / 4],
-					ct,
-					cu_processor_id,
-					k);
-
-		  if (ret < 0)
+		switch (KFD_GC_VERSION(kdev)) {
+		case IP_VERSION(9, 0, 1):
+			*pcache_info = vega10_cache_info;
+			num_of_cache_types = ARRAY_SIZE(vega10_cache_info);
+			break;
+		case IP_VERSION(9, 2, 1):
+			*pcache_info = vega12_cache_info;
+			num_of_cache_types = ARRAY_SIZE(vega12_cache_info);
+			break;
+		case IP_VERSION(9, 4, 0):
+		case IP_VERSION(9, 4, 1):
+			*pcache_info = vega20_cache_info;
+			num_of_cache_types = ARRAY_SIZE(vega20_cache_info);
+			break;
+		case IP_VERSION(9, 4, 2):
+			*pcache_info = aldebaran_cache_info;
+			num_of_cache_types = ARRAY_SIZE(aldebaran_cache_info);
+			break;
+		case IP_VERSION(9, 4, 3):
+		case IP_VERSION(9, 4, 4):
+		case IP_VERSION(9, 5, 0):
+			num_of_cache_types =
+				kfd_fill_gpu_cache_info_from_gfx_config_v2(kdev->kfd,
+									*pcache_info);
+			break;
+		case IP_VERSION(9, 1, 0):
+		case IP_VERSION(9, 2, 2):
+			*pcache_info = raven_cache_info;
+			num_of_cache_types = ARRAY_SIZE(raven_cache_info);
+			break;
+		case IP_VERSION(9, 3, 0):
+			*pcache_info = renoir_cache_info;
+			num_of_cache_types = ARRAY_SIZE(renoir_cache_info);
+			break;
+		case IP_VERSION(10, 1, 10):
+		case IP_VERSION(10, 1, 2):
+		case IP_VERSION(10, 1, 3):
+		case IP_VERSION(10, 1, 4):
+			*pcache_info = navi10_cache_info;
+			num_of_cache_types = ARRAY_SIZE(navi10_cache_info);
+			break;
+		case IP_VERSION(10, 1, 1):
+			*pcache_info = navi14_cache_info;
+			num_of_cache_types = ARRAY_SIZE(navi14_cache_info);
+			break;
+		case IP_VERSION(10, 3, 0):
+			*pcache_info = sienna_cichlid_cache_info;
+			num_of_cache_types = ARRAY_SIZE(sienna_cichlid_cache_info);
+			break;
+		case IP_VERSION(10, 3, 2):
+			*pcache_info = navy_flounder_cache_info;
+			num_of_cache_types = ARRAY_SIZE(navy_flounder_cache_info);
+			break;
+		case IP_VERSION(10, 3, 4):
+			*pcache_info = dimgrey_cavefish_cache_info;
+			num_of_cache_types = ARRAY_SIZE(dimgrey_cavefish_cache_info);
+			break;
+		case IP_VERSION(10, 3, 1):
+			*pcache_info = vangogh_cache_info;
+			num_of_cache_types = ARRAY_SIZE(vangogh_cache_info);
+			break;
+		case IP_VERSION(10, 3, 5):
+			*pcache_info = beige_goby_cache_info;
+			num_of_cache_types = ARRAY_SIZE(beige_goby_cache_info);
+			break;
+		case IP_VERSION(10, 3, 3):
+			*pcache_info = yellow_carp_cache_info;
+			num_of_cache_types = ARRAY_SIZE(yellow_carp_cache_info);
+			break;
+		case IP_VERSION(10, 3, 6):
+			*pcache_info = gc_10_3_6_cache_info;
+			num_of_cache_types = ARRAY_SIZE(gc_10_3_6_cache_info);
+			break;
+		case IP_VERSION(10, 3, 7):
+			*pcache_info = gfx1037_cache_info;
+			num_of_cache_types = ARRAY_SIZE(gfx1037_cache_info);
+			break;
+		case IP_VERSION(11, 0, 0):
+		case IP_VERSION(11, 0, 1):
+		case IP_VERSION(11, 0, 2):
+		case IP_VERSION(11, 0, 3):
+		case IP_VERSION(11, 0, 4):
+		case IP_VERSION(11, 5, 0):
+		case IP_VERSION(11, 5, 1):
+		case IP_VERSION(11, 5, 2):
+		case IP_VERSION(11, 5, 3):
+			/* Cacheline size not available in IP discovery for gc11.
+			 * kfd_fill_gpu_cache_info_from_gfx_config to hard code it
+			 */
+			cache_line_size_missing = true;
+			fallthrough;
+		case IP_VERSION(12, 0, 0):
+		case IP_VERSION(12, 0, 1):
+			num_of_cache_types =
+				kfd_fill_gpu_cache_info_from_gfx_config(kdev->kfd,
+									cache_line_size_missing,
+									*pcache_info);
+			break;
+		default:
+			*pcache_info = dummy_cache_info;
+			num_of_cache_types = ARRAY_SIZE(dummy_cache_info);
+			pr_warn("dummy cache info is used temporarily and real cache info need update later.\n");
 			break;
-
-		  if (!ret) {
-				pcache++;
-				(*num_of_entries)++;
-				mem_available -= sizeof(*pcache);
-				(*size_filled) += sizeof(*pcache);
-		  }
-
-		  /* Move to next CU block */
-		  num_cu_shared = ((k + pcache_info[ct].num_cu_shared) <=
-					cu_info->num_cu_per_sh) ?
-					pcache_info[ct].num_cu_shared :
-					(cu_info->num_cu_per_sh - k);
-		  cu_processor_id += num_cu_shared;
 		}
-	      }
-	    }
-	  } else {
-			ret = fill_in_l2_l3_pcache(pcache,
-				pcache_info,
-				cu_info,
-				mem_available,
-				ct,
-				cu_processor_id);
-
-			if (ret < 0)
-				break;
-
-			if (!ret) {
-				pcache++;
-				(*num_of_entries)++;
-				mem_available -= sizeof(*pcache);
-				(*size_filled) += sizeof(*pcache);
-			}
-	  }
-	}
-
-	pr_debug("Added [%d] GPU cache entries\n", *num_of_entries);
-
-	return 0;
-}
-
-static bool kfd_ignore_crat(void)
-{
-	bool ret;
-
-	if (ignore_crat)
-		return true;
-
-#ifndef KFD_SUPPORT_IOMMU_V2
-	ret = true;
-#else
-	ret = false;
-#endif
-
-	return ret;
-}
-
-/*
- * kfd_create_crat_image_acpi - Allocates memory for CRAT image and
- * copies CRAT from ACPI (if available).
- * NOTE: Call kfd_destroy_crat_image to free CRAT image memory
- *
- *	@crat_image: CRAT read from ACPI. If no CRAT in ACPI then
- *		     crat_image will be NULL
- *	@size: [OUT] size of crat_image
- *
- *	Return 0 if successful else return error code
- */
-int kfd_create_crat_image_acpi(void **crat_image, size_t *size)
-{
-	struct acpi_table_header *crat_table;
-	acpi_status status;
-	void *pcrat_image;
-	int rc = 0;
-
-	if (!crat_image)
-		return -EINVAL;
-
-	*crat_image = NULL;
-
-	if (kfd_ignore_crat()) {
-		pr_info("CRAT table disabled by module option\n");
-		return -ENODATA;
-	}
-
-	/* Fetch the CRAT table from ACPI */
-	status = acpi_get_table(CRAT_SIGNATURE, 0, &crat_table);
-	if (status == AE_NOT_FOUND) {
-		pr_warn("CRAT table not found\n");
-		return -ENODATA;
-	} else if (ACPI_FAILURE(status)) {
-		const char *err = acpi_format_exception(status);
-
-		pr_err("CRAT table error: %s\n", err);
-		return -EINVAL;
-	}
-
-	pcrat_image = kvmalloc(crat_table->length, GFP_KERNEL);
-	if (!pcrat_image) {
-		rc = -ENOMEM;
-		goto out;
 	}
-
-	memcpy(pcrat_image, crat_table, crat_table->length);
-	*crat_image = pcrat_image;
-	*size = crat_table->length;
-out:
-	acpi_put_table(crat_table);
-	return rc;
+	return num_of_cache_types;
 }
 
 /* Memory required to create Virtual CRAT.
@@ -1835,7 +1979,7 @@ static int kfd_create_vcrat_image_cpu(void *pcrat_image, size_t *size)
 }
 
 static int kfd_fill_gpu_memory_affinity(int *avail_size,
-		struct kfd_dev *kdev, uint8_t type, uint64_t size,
+		struct kfd_node *kdev, uint8_t type, uint64_t size,
 		struct crat_subtype_memory *sub_type_hdr,
 		uint32_t proximity_domain,
 		const struct kfd_local_mem_info *local_mem_info)
@@ -1864,13 +2008,13 @@ static int kfd_fill_gpu_memory_affinity(int *avail_size,
 }
 
 #ifdef CONFIG_ACPI_NUMA
-static void kfd_find_numa_node_in_srat(struct kfd_dev *kdev)
+static void kfd_find_numa_node_in_srat(struct kfd_node *kdev)
 {
 	struct acpi_table_header *table_header = NULL;
 	struct acpi_subtable_header *sub_header = NULL;
 	unsigned long table_end, subtable_len;
-	u32 pci_id = pci_domain_nr(kdev->pdev->bus) << 16 |
-			pci_dev_id(kdev->pdev);
+	u32 pci_id = pci_domain_nr(kdev->adev->pdev->bus) << 16 |
+			pci_dev_id(kdev->adev->pdev);
 	u32 bdf;
 	acpi_status status;
 	struct acpi_srat_cpu_affinity *cpu;
@@ -1945,10 +2089,13 @@ static void kfd_find_numa_node_in_srat(struct kfd_dev *kdev)
 		numa_node = 0;
 
 	if (numa_node != NUMA_NO_NODE)
-		set_dev_node(&kdev->pdev->dev, numa_node);
+		set_dev_node(&kdev->adev->pdev->dev, numa_node);
 }
 #endif
 
+#define KFD_CRAT_INTRA_SOCKET_WEIGHT	13
+#define KFD_CRAT_XGMI_WEIGHT		15
+
 /* kfd_fill_gpu_direct_io_link - Fill in direct io link from GPU
  * to its NUMA node
  *	@avail_size: Available size in the memory
@@ -1959,12 +2106,10 @@ static void kfd_find_numa_node_in_srat(struct kfd_dev *kdev)
  *	Return 0 if successful else return -ve value
  */
 static int kfd_fill_gpu_direct_io_link_to_cpu(int *avail_size,
-			struct kfd_dev *kdev,
+			struct kfd_node *kdev,
 			struct crat_subtype_iolink *sub_type_hdr,
 			uint32_t proximity_domain)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)kdev->kgd;
-
 	*avail_size -= sizeof(struct crat_subtype_iolink);
 	if (*avail_size < 0)
 		return -ENOMEM;
@@ -1981,7 +2126,13 @@ static int kfd_fill_gpu_direct_io_link_to_cpu(int *avail_size,
 	/* Fill in IOLINK subtype.
 	 * TODO: Fill-in other fields of iolink subtype
 	 */
-	if (adev->gmc.xgmi.connected_to_cpu) {
+	if (kdev->adev->gmc.xgmi.connected_to_cpu ||
+	    (KFD_GC_VERSION(kdev) == IP_VERSION(9, 4, 3) &&
+	     kdev->adev->smuio.funcs->get_pkg_type(kdev->adev) ==
+	     AMDGPU_PKG_TYPE_APU)) {
+		bool ext_cpu = KFD_GC_VERSION(kdev) != IP_VERSION(9, 4, 3);
+		int mem_bw = 819200, weight = ext_cpu ? KFD_CRAT_XGMI_WEIGHT :
+							KFD_CRAT_INTRA_SOCKET_WEIGHT;
 		/*
 		 * with host gpu xgmi link, host can access gpu memory whether
 		 * or not pcie bar type is large, so always create bidirectional
@@ -1989,33 +2140,37 @@ static int kfd_fill_gpu_direct_io_link_to_cpu(int *avail_size,
 		 */
 		sub_type_hdr->flags |= CRAT_IOLINK_FLAGS_BI_DIRECTIONAL;
 		sub_type_hdr->io_interface_type = CRAT_IOLINK_TYPE_XGMI;
-		sub_type_hdr->num_hops_xgmi = 1;
-		if (adev->asic_type == CHIP_ALDEBARAN) {
-			sub_type_hdr->minimum_bandwidth_mbs =
-					amdgpu_amdkfd_get_xgmi_bandwidth_mbytes(
-							kdev->kgd, NULL, true);
-			sub_type_hdr->maximum_bandwidth_mbs =
-					sub_type_hdr->minimum_bandwidth_mbs;
+		sub_type_hdr->weight_xgmi = weight;
+		if (ext_cpu) {
+			amdgpu_xgmi_get_bandwidth(kdev->adev, NULL,
+						  AMDGPU_XGMI_BW_MODE_PER_LINK,
+						  AMDGPU_XGMI_BW_UNIT_MBYTES,
+						  &sub_type_hdr->minimum_bandwidth_mbs,
+						  &sub_type_hdr->maximum_bandwidth_mbs);
+		} else {
+			sub_type_hdr->minimum_bandwidth_mbs = mem_bw;
+			sub_type_hdr->maximum_bandwidth_mbs = mem_bw;
 		}
 	} else {
 		sub_type_hdr->io_interface_type = CRAT_IOLINK_TYPE_PCIEXPRESS;
 		sub_type_hdr->minimum_bandwidth_mbs =
-				amdgpu_amdkfd_get_pcie_bandwidth_mbytes(kdev->kgd, true);
+				amdgpu_amdkfd_get_pcie_bandwidth_mbytes(kdev->adev, true);
 		sub_type_hdr->maximum_bandwidth_mbs =
-				amdgpu_amdkfd_get_pcie_bandwidth_mbytes(kdev->kgd, false);
+				amdgpu_amdkfd_get_pcie_bandwidth_mbytes(kdev->adev, false);
 	}
 
 	sub_type_hdr->proximity_domain_from = proximity_domain;
 
 #ifdef CONFIG_ACPI_NUMA
-	if (kdev->pdev->dev.numa_node == NUMA_NO_NODE)
+	if (kdev->adev->pdev->dev.numa_node == NUMA_NO_NODE &&
+	    num_possible_nodes() > 1)
 		kfd_find_numa_node_in_srat(kdev);
 #endif
 #ifdef CONFIG_NUMA
-	if (kdev->pdev->dev.numa_node == NUMA_NO_NODE)
+	if (kdev->adev->pdev->dev.numa_node == NUMA_NO_NODE)
 		sub_type_hdr->proximity_domain_to = 0;
 	else
-		sub_type_hdr->proximity_domain_to = kdev->pdev->dev.numa_node;
+		sub_type_hdr->proximity_domain_to = kdev->adev->pdev->dev.numa_node;
 #else
 	sub_type_hdr->proximity_domain_to = 0;
 #endif
@@ -2023,12 +2178,14 @@ static int kfd_fill_gpu_direct_io_link_to_cpu(int *avail_size,
 }
 
 static int kfd_fill_gpu_xgmi_link_to_gpu(int *avail_size,
-			struct kfd_dev *kdev,
-			struct kfd_dev *peer_kdev,
+			struct kfd_node *kdev,
+			struct kfd_node *peer_kdev,
 			struct crat_subtype_iolink *sub_type_hdr,
 			uint32_t proximity_domain_from,
 			uint32_t proximity_domain_to)
 {
+	bool use_ta_info = kdev->kfd->num_nodes == 1;
+
 	*avail_size -= sizeof(struct crat_subtype_iolink);
 	if (*avail_size < 0)
 		return -ENOMEM;
@@ -2043,12 +2200,25 @@ static int kfd_fill_gpu_xgmi_link_to_gpu(int *avail_size,
 	sub_type_hdr->io_interface_type = CRAT_IOLINK_TYPE_XGMI;
 	sub_type_hdr->proximity_domain_from = proximity_domain_from;
 	sub_type_hdr->proximity_domain_to = proximity_domain_to;
-	sub_type_hdr->num_hops_xgmi =
-		amdgpu_amdkfd_get_xgmi_hops_count(kdev->kgd, peer_kdev->kgd);
-	sub_type_hdr->maximum_bandwidth_mbs =
-		amdgpu_amdkfd_get_xgmi_bandwidth_mbytes(kdev->kgd, peer_kdev->kgd, false);
-	sub_type_hdr->minimum_bandwidth_mbs = sub_type_hdr->maximum_bandwidth_mbs ?
-		amdgpu_amdkfd_get_xgmi_bandwidth_mbytes(kdev->kgd, NULL, true) : 0;
+
+	if (use_ta_info) {
+		sub_type_hdr->weight_xgmi = KFD_CRAT_XGMI_WEIGHT *
+			amdgpu_xgmi_get_hops_count(kdev->adev, peer_kdev->adev);
+		amdgpu_xgmi_get_bandwidth(kdev->adev, peer_kdev->adev,
+					  AMDGPU_XGMI_BW_MODE_PER_PEER,
+					  AMDGPU_XGMI_BW_UNIT_MBYTES,
+					  &sub_type_hdr->minimum_bandwidth_mbs,
+					  &sub_type_hdr->maximum_bandwidth_mbs);
+	} else {
+		bool is_single_hop = kdev->kfd == peer_kdev->kfd;
+		int weight = is_single_hop ? KFD_CRAT_INTRA_SOCKET_WEIGHT :
+			(2 * KFD_CRAT_INTRA_SOCKET_WEIGHT) + KFD_CRAT_XGMI_WEIGHT;
+		int mem_bw = 819200;
+
+		sub_type_hdr->weight_xgmi = weight;
+		sub_type_hdr->maximum_bandwidth_mbs = is_single_hop ? mem_bw : 0;
+		sub_type_hdr->minimum_bandwidth_mbs = is_single_hop ? mem_bw : 0;
+	}
 
 	return 0;
 }
@@ -2060,19 +2230,18 @@ static int kfd_fill_gpu_xgmi_link_to_gpu(int *avail_size,
  *		[OUT] actual size of data filled in crat_image
  */
 static int kfd_create_vcrat_image_gpu(void *pcrat_image,
-				      size_t *size, struct kfd_dev *kdev,
+				      size_t *size, struct kfd_node *kdev,
 				      uint32_t proximity_domain)
 {
 	struct crat_header *crat_table = (struct crat_header *)pcrat_image;
+	struct amdgpu_gfx_config *gfx_info = &kdev->adev->gfx.config;
+	struct amdgpu_cu_info *cu_info = &kdev->adev->gfx.cu_info;
 	struct crat_subtype_generic *sub_type_hdr;
 	struct kfd_local_mem_info local_mem_info;
 	struct kfd_topology_device *peer_dev;
 	struct crat_subtype_computeunit *cu;
-	struct kfd_cu_info cu_info;
 	int avail_size = *size;
 	uint32_t total_num_of_cu;
-	int num_of_cache_entries = 0;
-	int cache_mem_filled = 0;
 	uint32_t nid = 0;
 	int ret = 0;
 
@@ -2083,9 +2252,6 @@ static int kfd_create_vcrat_image_gpu(void *pcrat_image,
 	 * Modify length and total_entries as subunits are added.
 	 */
 	avail_size -= sizeof(struct crat_header);
-	if (avail_size < 0)
-		return -ENOMEM;
-
 	memset(crat_table, 0, sizeof(struct crat_header));
 
 	memcpy(&crat_table->signature, CRAT_SIGNATURE,
@@ -2099,9 +2265,6 @@ static int kfd_create_vcrat_image_gpu(void *pcrat_image,
 	 * First fill in the sub type header and then sub type data
 	 */
 	avail_size -= sizeof(struct crat_subtype_computeunit);
-	if (avail_size < 0)
-		return -ENOMEM;
-
 	sub_type_hdr = (struct crat_subtype_generic *)(crat_table + 1);
 	memset(sub_type_hdr, 0, sizeof(struct crat_subtype_computeunit));
 
@@ -2114,29 +2277,23 @@ static int kfd_create_vcrat_image_gpu(void *pcrat_image,
 	cu->flags |= CRAT_CU_FLAGS_GPU_PRESENT;
 	cu->proximity_domain = proximity_domain;
 
-	amdgpu_amdkfd_get_cu_info(kdev->kgd, &cu_info);
-	cu->num_simd_per_cu = cu_info.simd_per_cu;
-	cu->num_simd_cores = cu_info.simd_per_cu * cu_info.cu_active_number;
-	cu->max_waves_simd = cu_info.max_waves_per_simd;
+	cu->num_simd_per_cu = cu_info->simd_per_cu;
+	cu->num_simd_cores = cu_info->simd_per_cu *
+			(cu_info->number / kdev->kfd->num_nodes);
+	cu->max_waves_simd = cu_info->max_waves_per_simd;
 
-	cu->wave_front_size = cu_info.wave_front_size;
-	cu->array_count = cu_info.num_shader_arrays_per_engine *
-		cu_info.num_shader_engines;
-	total_num_of_cu = (cu->array_count * cu_info.num_cu_per_sh);
+	cu->wave_front_size = cu_info->wave_front_size;
+	cu->array_count = gfx_info->max_sh_per_se *
+		gfx_info->max_shader_engines;
+	total_num_of_cu = (cu->array_count * gfx_info->max_cu_per_sh);
 	cu->processor_id_low = get_and_inc_gpu_processor_id(total_num_of_cu);
-	cu->num_cu_per_array = cu_info.num_cu_per_sh;
-	cu->max_slots_scatch_cu = cu_info.max_scratch_slots_per_cu;
-	cu->num_banks = cu_info.num_shader_engines;
-	cu->lds_size_in_kb = cu_info.lds_size;
+	cu->num_cu_per_array = gfx_info->max_cu_per_sh;
+	cu->max_slots_scatch_cu = cu_info->max_scratch_slots_per_cu;
+	cu->num_banks = gfx_info->max_shader_engines;
+	cu->lds_size_in_kb = cu_info->lds_size;
 
 	cu->hsa_capability = 0;
 
-	/* Check if this node supports IOMMU. During parsing this flag will
-	 * translate to HSA_CAP_ATS_PRESENT
-	 */
-	if (!kfd_iommu_check_device(kdev))
-		cu->hsa_capability |= CRAT_CU_FLAGS_IOMMU_PRESENT;
-
 	crat_table->length += sub_type_hdr->length;
 	crat_table->total_entries++;
 
@@ -2145,11 +2302,11 @@ static int kfd_create_vcrat_image_gpu(void *pcrat_image,
 	 * report the total FB size (public+private) as a single
 	 * private heap.
 	 */
-	amdgpu_amdkfd_get_local_mem_info(kdev->kgd, &local_mem_info);
+	local_mem_info = kdev->local_mem_info;
 	sub_type_hdr = (typeof(sub_type_hdr))((char *)sub_type_hdr +
 			sub_type_hdr->length);
 
-	if (debug_largebar)
+	if (kdev->adev->debug_largebar)
 		local_mem_info.local_mem_size_private = 0;
 
 	if (local_mem_info.local_mem_size_private == 0)
@@ -2173,31 +2330,12 @@ static int kfd_create_vcrat_image_gpu(void *pcrat_image,
 	crat_table->length += sizeof(struct crat_subtype_memory);
 	crat_table->total_entries++;
 
-	/* TODO: Fill in cache information. This information is NOT readily
-	 * available in KGD
-	 */
-	sub_type_hdr = (typeof(sub_type_hdr))((char *)sub_type_hdr +
-		sub_type_hdr->length);
-	ret = kfd_fill_gpu_cache_info(kdev, cu->processor_id_low,
-				avail_size,
-				&cu_info,
-				(struct crat_subtype_cache *)sub_type_hdr,
-				&cache_mem_filled,
-				&num_of_cache_entries);
-
-	if (ret < 0)
-		return ret;
-
-	crat_table->length += cache_mem_filled;
-	crat_table->total_entries += num_of_cache_entries;
-	avail_size -= cache_mem_filled;
-
 	/* Fill in Subtype: IO_LINKS
 	 *  Only direct links are added here which is Link from GPU to
-	 *  to its NUMA node. Indirect links are added by userspace.
+	 *  its NUMA node. Indirect links are added by userspace.
 	 */
 	sub_type_hdr = (typeof(sub_type_hdr))((char *)sub_type_hdr +
-		cache_mem_filled);
+		sub_type_hdr->length);
 	ret = kfd_fill_gpu_direct_io_link_to_cpu(&avail_size, kdev,
 		(struct crat_subtype_iolink *)sub_type_hdr, proximity_domain);
 
@@ -2216,12 +2354,14 @@ static int kfd_create_vcrat_image_gpu(void *pcrat_image,
 	 * (from other GPU to this GPU) will be added
 	 * in kfd_parse_subtype_iolink.
 	 */
-	if (kdev->hive_id) {
+	if (kdev->kfd->hive_id) {
 		for (nid = 0; nid < proximity_domain; ++nid) {
-			peer_dev = kfd_topology_device_by_proximity_domain(nid);
+			peer_dev = kfd_topology_device_by_proximity_domain_no_lock(nid);
 			if (!peer_dev->gpu)
 				continue;
-			if (peer_dev->gpu->hive_id != kdev->hive_id)
+			if (peer_dev->gpu->kfd->hive_id != kdev->kfd->hive_id)
+				continue;
+			if (!amdgpu_xgmi_get_is_sharing_enabled(kdev->adev, peer_dev->gpu->adev))
 				continue;
 			sub_type_hdr = (typeof(sub_type_hdr))(
 				(char *)sub_type_hdr +
@@ -2255,12 +2395,12 @@ static int kfd_create_vcrat_image_gpu(void *pcrat_image,
  *		(COMPUTE_UNIT_CPU | COMPUTE_UNIT_GPU) - Create VCRAT for APU
  *			-- this option is not currently implemented.
  *			The assumption is that all AMD APUs will have CRAT
- *	@kdev: Valid kfd_device required if flags contain COMPUTE_UNIT_GPU
+ *	@kdev: Valid kfd_node required if flags contain COMPUTE_UNIT_GPU
  *
  *	Return 0 if successful else return -ve value
  */
 int kfd_create_crat_image_virtual(void **crat_image, size_t *size,
-				  int flags, struct kfd_dev *kdev,
+				  int flags, struct kfd_node *kdev,
 				  uint32_t proximity_domain)
 {
 	void *pcrat_image = NULL;
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_crat.h b/drivers/gpu/drm/amd/amdkfd/kfd_crat.h
index d54ceebd346b..a8ca7ecb6d27 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_crat.h
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_crat.h
@@ -1,5 +1,6 @@
+/* SPDX-License-Identifier: GPL-2.0 OR MIT */
 /*
- * Copyright 2014 Advanced Micro Devices, Inc.
+ * Copyright 2014-2022 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -28,11 +29,10 @@
 #pragma pack(1)
 
 /*
- * 4CC signature values for the CRAT and CDIT ACPI tables
+ * 4CC signature value for the CRAT ACPI table
  */
 
 #define CRAT_SIGNATURE	"CRAT"
-#define CDIT_SIGNATURE	"CDIT"
 
 /*
  * Component Resource Association Table (CRAT)
@@ -42,8 +42,6 @@
 #define CRAT_OEMTABLEID_LENGTH	8
 #define CRAT_RESERVED_LENGTH	6
 
-#define CRAT_OEMID_64BIT_MASK ((1ULL << (CRAT_OEMID_LENGTH * 8)) - 1)
-
 /* Compute Unit flags */
 #define COMPUTE_UNIT_CPU	(1 << 0)  /* Create Virtual CRAT for CPU */
 #define COMPUTE_UNIT_GPU	(1 << 1)  /* Create Virtual CRAT for GPU */
@@ -79,6 +77,10 @@ struct crat_header {
 #define CRAT_SUBTYPE_IOLINK_AFFINITY		5
 #define CRAT_SUBTYPE_MAX			6
 
+/*
+ * Do not change the value of CRAT_SIBLINGMAP_SIZE from 32
+ * as it breaks the ABI.
+ */
 #define CRAT_SIBLINGMAP_SIZE	32
 
 /*
@@ -232,7 +234,7 @@ struct crat_subtype_ccompute {
 #define CRAT_IOLINK_FLAGS_NO_ATOMICS_32_BIT	(1 << 2)
 #define CRAT_IOLINK_FLAGS_NO_ATOMICS_64_BIT	(1 << 3)
 #define CRAT_IOLINK_FLAGS_NO_PEER_TO_PEER_DMA	(1 << 4)
-#define CRAT_IOLINK_FLAGS_BI_DIRECTIONAL 	(1 << 31)
+#define CRAT_IOLINK_FLAGS_BI_DIRECTIONAL	(1 << 31)
 #define CRAT_IOLINK_FLAGS_RESERVED_MASK		0x7fffffe0
 
 /*
@@ -275,7 +277,7 @@ struct crat_subtype_iolink {
 	uint32_t	maximum_bandwidth_mbs;
 	uint32_t	recommended_transfer_size;
 	uint8_t		reserved2[CRAT_IOLINK_RESERVED_LENGTH - 1];
-	uint8_t		num_hops_xgmi;
+	uint8_t		weight_xgmi;
 };
 
 /*
@@ -291,37 +293,28 @@ struct crat_subtype_generic {
 	uint32_t	flags;
 };
 
-/*
- * Component Locality Distance Information Table (CDIT)
- */
-#define CDIT_OEMID_LENGTH	6
-#define CDIT_OEMTABLEID_LENGTH	8
-
-struct cdit_header {
-	uint32_t	signature;
-	uint32_t	length;
-	uint8_t		revision;
-	uint8_t		checksum;
-	uint8_t		oem_id[CDIT_OEMID_LENGTH];
-	uint8_t		oem_table_id[CDIT_OEMTABLEID_LENGTH];
-	uint32_t	oem_revision;
-	uint32_t	creator_id;
-	uint32_t	creator_revision;
-	uint32_t	total_entries;
-	uint16_t	num_domains;
-	uint8_t		entry[1];
-};
-
 #pragma pack()
 
-struct kfd_dev;
+struct kfd_node;
+
+/* Static table to describe GPU Cache information */
+struct kfd_gpu_cache_info {
+	uint32_t	cache_size;
+	uint32_t	cache_level;
+	uint32_t	cache_line_size;
+	uint32_t	flags;
+	/* Indicates how many Compute Units share this cache
+	 * within a SA. Value = 1 indicates the cache is not shared
+	 */
+	uint32_t	num_cu_shared;
+};
+int kfd_get_gpu_cache_info(struct kfd_node *kdev, struct kfd_gpu_cache_info **pcache_info);
 
-int kfd_create_crat_image_acpi(void **crat_image, size_t *size);
 void kfd_destroy_crat_image(void *crat_image);
 int kfd_parse_crat_table(void *crat_image, struct list_head *device_list,
 			 uint32_t proximity_domain);
 int kfd_create_crat_image_virtual(void **crat_image, size_t *size,
-				  int flags, struct kfd_dev *kdev,
+				  int flags, struct kfd_node *kdev,
 				  uint32_t proximity_domain);
 
 #endif /* KFD_CRAT_H_INCLUDED */
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_dbgdev.c b/drivers/gpu/drm/amd/amdkfd/kfd_dbgdev.c
deleted file mode 100644
index 159add0f5aaa..000000000000
--- a/drivers/gpu/drm/amd/amdkfd/kfd_dbgdev.c
+++ /dev/null
@@ -1,845 +0,0 @@
-/*
- * Copyright 2014 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- */
-
-#include <linux/types.h>
-#include <linux/kernel.h>
-#include <linux/log2.h>
-#include <linux/sched.h>
-#include <linux/slab.h>
-#include <linux/mutex.h>
-#include <linux/device.h>
-
-#include "kfd_pm4_headers.h"
-#include "kfd_pm4_headers_diq.h"
-#include "kfd_kernel_queue.h"
-#include "kfd_priv.h"
-#include "kfd_pm4_opcodes.h"
-#include "cik_regs.h"
-#include "kfd_dbgmgr.h"
-#include "kfd_dbgdev.h"
-#include "kfd_device_queue_manager.h"
-
-static void dbgdev_address_watch_disable_nodiq(struct kfd_dev *dev)
-{
-	dev->kfd2kgd->address_watch_disable(dev->kgd);
-}
-
-static int dbgdev_diq_submit_ib(struct kfd_dbgdev *dbgdev,
-				u32 pasid, uint64_t vmid0_address,
-				uint32_t *packet_buff, size_t size_in_bytes)
-{
-	struct pm4__release_mem *rm_packet;
-	struct pm4__indirect_buffer_pasid *ib_packet;
-	struct kfd_mem_obj *mem_obj;
-	size_t pq_packets_size_in_bytes;
-	union ULARGE_INTEGER *largep;
-	union ULARGE_INTEGER addr;
-	struct kernel_queue *kq;
-	uint64_t *rm_state;
-	unsigned int *ib_packet_buff;
-	int status;
-
-	if (WARN_ON(!size_in_bytes))
-		return -EINVAL;
-
-	kq = dbgdev->kq;
-
-	pq_packets_size_in_bytes = sizeof(struct pm4__release_mem) +
-				sizeof(struct pm4__indirect_buffer_pasid);
-
-	/*
-	 * We acquire a buffer from DIQ
-	 * The receive packet buff will be sitting on the Indirect Buffer
-	 * and in the PQ we put the IB packet + sync packet(s).
-	 */
-	status = kq_acquire_packet_buffer(kq,
-				pq_packets_size_in_bytes / sizeof(uint32_t),
-				&ib_packet_buff);
-	if (status) {
-		pr_err("kq_acquire_packet_buffer failed\n");
-		return status;
-	}
-
-	memset(ib_packet_buff, 0, pq_packets_size_in_bytes);
-
-	ib_packet = (struct pm4__indirect_buffer_pasid *) (ib_packet_buff);
-
-	ib_packet->header.count = 3;
-	ib_packet->header.opcode = IT_INDIRECT_BUFFER_PASID;
-	ib_packet->header.type = PM4_TYPE_3;
-
-	largep = (union ULARGE_INTEGER *) &vmid0_address;
-
-	ib_packet->bitfields2.ib_base_lo = largep->u.low_part >> 2;
-	ib_packet->bitfields3.ib_base_hi = largep->u.high_part;
-
-	ib_packet->control = (1 << 23) | (1 << 31) |
-			((size_in_bytes / 4) & 0xfffff);
-
-	ib_packet->bitfields5.pasid = pasid;
-
-	/*
-	 * for now we use release mem for GPU-CPU synchronization
-	 * Consider WaitRegMem + WriteData as a better alternative
-	 * we get a GART allocations ( gpu/cpu mapping),
-	 * for the sync variable, and wait until:
-	 * (a) Sync with HW
-	 * (b) Sync var is written by CP to mem.
-	 */
-	rm_packet = (struct pm4__release_mem *) (ib_packet_buff +
-			(sizeof(struct pm4__indirect_buffer_pasid) /
-					sizeof(unsigned int)));
-
-	status = kfd_gtt_sa_allocate(dbgdev->dev, sizeof(uint64_t),
-					&mem_obj);
-
-	if (status) {
-		pr_err("Failed to allocate GART memory\n");
-		kq_rollback_packet(kq);
-		return status;
-	}
-
-	rm_state = (uint64_t *) mem_obj->cpu_ptr;
-
-	*rm_state = QUEUESTATE__ACTIVE_COMPLETION_PENDING;
-
-	rm_packet->header.opcode = IT_RELEASE_MEM;
-	rm_packet->header.type = PM4_TYPE_3;
-	rm_packet->header.count = sizeof(struct pm4__release_mem) / 4 - 2;
-
-	rm_packet->bitfields2.event_type = CACHE_FLUSH_AND_INV_TS_EVENT;
-	rm_packet->bitfields2.event_index =
-				event_index___release_mem__end_of_pipe;
-
-	rm_packet->bitfields2.cache_policy = cache_policy___release_mem__lru;
-	rm_packet->bitfields2.atc = 0;
-	rm_packet->bitfields2.tc_wb_action_ena = 1;
-
-	addr.quad_part = mem_obj->gpu_addr;
-
-	rm_packet->bitfields4.address_lo_32b = addr.u.low_part >> 2;
-	rm_packet->address_hi = addr.u.high_part;
-
-	rm_packet->bitfields3.data_sel =
-				data_sel___release_mem__send_64_bit_data;
-
-	rm_packet->bitfields3.int_sel =
-			int_sel___release_mem__send_data_after_write_confirm;
-
-	rm_packet->bitfields3.dst_sel =
-			dst_sel___release_mem__memory_controller;
-
-	rm_packet->data_lo = QUEUESTATE__ACTIVE;
-
-	kq_submit_packet(kq);
-
-	/* Wait till CP writes sync code: */
-	status = amdkfd_fence_wait_timeout(
-			rm_state,
-			QUEUESTATE__ACTIVE, 1500);
-
-	kfd_gtt_sa_free(dbgdev->dev, mem_obj);
-
-	return status;
-}
-
-static int dbgdev_register_nodiq(struct kfd_dbgdev *dbgdev)
-{
-	/*
-	 * no action is needed in this case,
-	 * just make sure diq will not be used
-	 */
-
-	dbgdev->kq = NULL;
-
-	return 0;
-}
-
-static int dbgdev_register_diq(struct kfd_dbgdev *dbgdev)
-{
-	struct queue_properties properties;
-	unsigned int qid;
-	struct kernel_queue *kq = NULL;
-	int status;
-
-	properties.type = KFD_QUEUE_TYPE_DIQ;
-
-	status = pqm_create_queue(dbgdev->pqm, dbgdev->dev, NULL,
-				&properties, &qid, NULL);
-
-	if (status) {
-		pr_err("Failed to create DIQ\n");
-		return status;
-	}
-
-	pr_debug("DIQ Created with queue id: %d\n", qid);
-
-	kq = pqm_get_kernel_queue(dbgdev->pqm, qid);
-
-	if (!kq) {
-		pr_err("Error getting DIQ\n");
-		pqm_destroy_queue(dbgdev->pqm, qid);
-		return -EFAULT;
-	}
-
-	dbgdev->kq = kq;
-
-	return status;
-}
-
-static int dbgdev_unregister_nodiq(struct kfd_dbgdev *dbgdev)
-{
-	/* disable watch address */
-	dbgdev_address_watch_disable_nodiq(dbgdev->dev);
-	return 0;
-}
-
-static int dbgdev_unregister_diq(struct kfd_dbgdev *dbgdev)
-{
-	/* todo - disable address watch */
-	int status;
-
-	status = pqm_destroy_queue(dbgdev->pqm,
-			dbgdev->kq->queue->properties.queue_id);
-	dbgdev->kq = NULL;
-
-	return status;
-}
-
-static void dbgdev_address_watch_set_registers(
-			const struct dbg_address_watch_info *adw_info,
-			union TCP_WATCH_ADDR_H_BITS *addrHi,
-			union TCP_WATCH_ADDR_L_BITS *addrLo,
-			union TCP_WATCH_CNTL_BITS *cntl,
-			unsigned int index, unsigned int vmid)
-{
-	union ULARGE_INTEGER addr;
-
-	addr.quad_part = 0;
-	addrHi->u32All = 0;
-	addrLo->u32All = 0;
-	cntl->u32All = 0;
-
-	if (adw_info->watch_mask)
-		cntl->bitfields.mask =
-			(uint32_t) (adw_info->watch_mask[index] &
-					ADDRESS_WATCH_REG_CNTL_DEFAULT_MASK);
-	else
-		cntl->bitfields.mask = ADDRESS_WATCH_REG_CNTL_DEFAULT_MASK;
-
-	addr.quad_part = (unsigned long long) adw_info->watch_address[index];
-
-	addrHi->bitfields.addr = addr.u.high_part &
-					ADDRESS_WATCH_REG_ADDHIGH_MASK;
-	addrLo->bitfields.addr =
-			(addr.u.low_part >> ADDRESS_WATCH_REG_ADDLOW_SHIFT);
-
-	cntl->bitfields.mode = adw_info->watch_mode[index];
-	cntl->bitfields.vmid = (uint32_t) vmid;
-	/* for now assume it is an ATC address */
-	cntl->u32All |= ADDRESS_WATCH_REG_CNTL_ATC_BIT;
-
-	pr_debug("\t\t%20s %08x\n", "set reg mask :", cntl->bitfields.mask);
-	pr_debug("\t\t%20s %08x\n", "set reg add high :",
-			addrHi->bitfields.addr);
-	pr_debug("\t\t%20s %08x\n", "set reg add low :",
-			addrLo->bitfields.addr);
-}
-
-static int dbgdev_address_watch_nodiq(struct kfd_dbgdev *dbgdev,
-				      struct dbg_address_watch_info *adw_info)
-{
-	union TCP_WATCH_ADDR_H_BITS addrHi;
-	union TCP_WATCH_ADDR_L_BITS addrLo;
-	union TCP_WATCH_CNTL_BITS cntl;
-	struct kfd_process_device *pdd;
-	unsigned int i;
-
-	/* taking the vmid for that process on the safe way using pdd */
-	pdd = kfd_get_process_device_data(dbgdev->dev,
-					adw_info->process);
-	if (!pdd) {
-		pr_err("Failed to get pdd for wave control no DIQ\n");
-		return -EFAULT;
-	}
-
-	addrHi.u32All = 0;
-	addrLo.u32All = 0;
-	cntl.u32All = 0;
-
-	if ((adw_info->num_watch_points > MAX_WATCH_ADDRESSES) ||
-			(adw_info->num_watch_points == 0)) {
-		pr_err("num_watch_points is invalid\n");
-		return -EINVAL;
-	}
-
-	if (!adw_info->watch_mode || !adw_info->watch_address) {
-		pr_err("adw_info fields are not valid\n");
-		return -EINVAL;
-	}
-
-	for (i = 0; i < adw_info->num_watch_points; i++) {
-		dbgdev_address_watch_set_registers(adw_info, &addrHi, &addrLo,
-						&cntl, i, pdd->qpd.vmid);
-
-		pr_debug("\t\t%30s\n", "* * * * * * * * * * * * * * * * * *");
-		pr_debug("\t\t%20s %08x\n", "register index :", i);
-		pr_debug("\t\t%20s %08x\n", "vmid is :", pdd->qpd.vmid);
-		pr_debug("\t\t%20s %08x\n", "Address Low is :",
-				addrLo.bitfields.addr);
-		pr_debug("\t\t%20s %08x\n", "Address high is :",
-				addrHi.bitfields.addr);
-		pr_debug("\t\t%20s %08x\n", "Address high is :",
-				addrHi.bitfields.addr);
-		pr_debug("\t\t%20s %08x\n", "Control Mask is :",
-				cntl.bitfields.mask);
-		pr_debug("\t\t%20s %08x\n", "Control Mode is :",
-				cntl.bitfields.mode);
-		pr_debug("\t\t%20s %08x\n", "Control Vmid is :",
-				cntl.bitfields.vmid);
-		pr_debug("\t\t%20s %08x\n", "Control atc  is :",
-				cntl.bitfields.atc);
-		pr_debug("\t\t%30s\n", "* * * * * * * * * * * * * * * * * *");
-
-		pdd->dev->kfd2kgd->address_watch_execute(
-						dbgdev->dev->kgd,
-						i,
-						cntl.u32All,
-						addrHi.u32All,
-						addrLo.u32All);
-	}
-
-	return 0;
-}
-
-static int dbgdev_address_watch_diq(struct kfd_dbgdev *dbgdev,
-				    struct dbg_address_watch_info *adw_info)
-{
-	struct pm4__set_config_reg *packets_vec;
-	union TCP_WATCH_ADDR_H_BITS addrHi;
-	union TCP_WATCH_ADDR_L_BITS addrLo;
-	union TCP_WATCH_CNTL_BITS cntl;
-	struct kfd_mem_obj *mem_obj;
-	unsigned int aw_reg_add_dword;
-	uint32_t *packet_buff_uint;
-	unsigned int i;
-	int status;
-	size_t ib_size = sizeof(struct pm4__set_config_reg) * 4;
-	/* we do not control the vmid in DIQ mode, just a place holder */
-	unsigned int vmid = 0;
-
-	addrHi.u32All = 0;
-	addrLo.u32All = 0;
-	cntl.u32All = 0;
-
-	if ((adw_info->num_watch_points > MAX_WATCH_ADDRESSES) ||
-			(adw_info->num_watch_points == 0)) {
-		pr_err("num_watch_points is invalid\n");
-		return -EINVAL;
-	}
-
-	if (!adw_info->watch_mode || !adw_info->watch_address) {
-		pr_err("adw_info fields are not valid\n");
-		return -EINVAL;
-	}
-
-	status = kfd_gtt_sa_allocate(dbgdev->dev, ib_size, &mem_obj);
-
-	if (status) {
-		pr_err("Failed to allocate GART memory\n");
-		return status;
-	}
-
-	packet_buff_uint = mem_obj->cpu_ptr;
-
-	memset(packet_buff_uint, 0, ib_size);
-
-	packets_vec = (struct pm4__set_config_reg *) (packet_buff_uint);
-
-	packets_vec[0].header.count = 1;
-	packets_vec[0].header.opcode = IT_SET_CONFIG_REG;
-	packets_vec[0].header.type = PM4_TYPE_3;
-	packets_vec[0].bitfields2.vmid_shift = ADDRESS_WATCH_CNTL_OFFSET;
-	packets_vec[0].bitfields2.insert_vmid = 1;
-	packets_vec[1].ordinal1 = packets_vec[0].ordinal1;
-	packets_vec[1].bitfields2.insert_vmid = 0;
-	packets_vec[2].ordinal1 = packets_vec[0].ordinal1;
-	packets_vec[2].bitfields2.insert_vmid = 0;
-	packets_vec[3].ordinal1 = packets_vec[0].ordinal1;
-	packets_vec[3].bitfields2.vmid_shift = ADDRESS_WATCH_CNTL_OFFSET;
-	packets_vec[3].bitfields2.insert_vmid = 1;
-
-	for (i = 0; i < adw_info->num_watch_points; i++) {
-		dbgdev_address_watch_set_registers(adw_info,
-						&addrHi,
-						&addrLo,
-						&cntl,
-						i,
-						vmid);
-
-		pr_debug("\t\t%30s\n", "* * * * * * * * * * * * * * * * * *");
-		pr_debug("\t\t%20s %08x\n", "register index :", i);
-		pr_debug("\t\t%20s %08x\n", "vmid is :", vmid);
-		pr_debug("\t\t%20s %p\n", "Add ptr is :",
-				adw_info->watch_address);
-		pr_debug("\t\t%20s %08llx\n", "Add     is :",
-				adw_info->watch_address[i]);
-		pr_debug("\t\t%20s %08x\n", "Address Low is :",
-				addrLo.bitfields.addr);
-		pr_debug("\t\t%20s %08x\n", "Address high is :",
-				addrHi.bitfields.addr);
-		pr_debug("\t\t%20s %08x\n", "Control Mask is :",
-				cntl.bitfields.mask);
-		pr_debug("\t\t%20s %08x\n", "Control Mode is :",
-				cntl.bitfields.mode);
-		pr_debug("\t\t%20s %08x\n", "Control Vmid is :",
-				cntl.bitfields.vmid);
-		pr_debug("\t\t%20s %08x\n", "Control atc  is :",
-				cntl.bitfields.atc);
-		pr_debug("\t\t%30s\n", "* * * * * * * * * * * * * * * * * *");
-
-		aw_reg_add_dword =
-				dbgdev->dev->kfd2kgd->address_watch_get_offset(
-					dbgdev->dev->kgd,
-					i,
-					ADDRESS_WATCH_REG_CNTL);
-
-		packets_vec[0].bitfields2.reg_offset =
-					aw_reg_add_dword - AMD_CONFIG_REG_BASE;
-
-		packets_vec[0].reg_data[0] = cntl.u32All;
-
-		aw_reg_add_dword =
-				dbgdev->dev->kfd2kgd->address_watch_get_offset(
-					dbgdev->dev->kgd,
-					i,
-					ADDRESS_WATCH_REG_ADDR_HI);
-
-		packets_vec[1].bitfields2.reg_offset =
-					aw_reg_add_dword - AMD_CONFIG_REG_BASE;
-		packets_vec[1].reg_data[0] = addrHi.u32All;
-
-		aw_reg_add_dword =
-				dbgdev->dev->kfd2kgd->address_watch_get_offset(
-					dbgdev->dev->kgd,
-					i,
-					ADDRESS_WATCH_REG_ADDR_LO);
-
-		packets_vec[2].bitfields2.reg_offset =
-				aw_reg_add_dword - AMD_CONFIG_REG_BASE;
-		packets_vec[2].reg_data[0] = addrLo.u32All;
-
-		/* enable watch flag if address is not zero*/
-		if (adw_info->watch_address[i] > 0)
-			cntl.bitfields.valid = 1;
-		else
-			cntl.bitfields.valid = 0;
-
-		aw_reg_add_dword =
-				dbgdev->dev->kfd2kgd->address_watch_get_offset(
-					dbgdev->dev->kgd,
-					i,
-					ADDRESS_WATCH_REG_CNTL);
-
-		packets_vec[3].bitfields2.reg_offset =
-					aw_reg_add_dword - AMD_CONFIG_REG_BASE;
-		packets_vec[3].reg_data[0] = cntl.u32All;
-
-		status = dbgdev_diq_submit_ib(
-					dbgdev,
-					adw_info->process->pasid,
-					mem_obj->gpu_addr,
-					packet_buff_uint,
-					ib_size);
-
-		if (status) {
-			pr_err("Failed to submit IB to DIQ\n");
-			break;
-		}
-	}
-
-	kfd_gtt_sa_free(dbgdev->dev, mem_obj);
-	return status;
-}
-
-static int dbgdev_wave_control_set_registers(
-				struct dbg_wave_control_info *wac_info,
-				union SQ_CMD_BITS *in_reg_sq_cmd,
-				union GRBM_GFX_INDEX_BITS *in_reg_gfx_index)
-{
-	int status = 0;
-	union SQ_CMD_BITS reg_sq_cmd;
-	union GRBM_GFX_INDEX_BITS reg_gfx_index;
-	struct HsaDbgWaveMsgAMDGen2 *pMsg;
-
-	reg_sq_cmd.u32All = 0;
-	reg_gfx_index.u32All = 0;
-	pMsg = &wac_info->dbgWave_msg.DbgWaveMsg.WaveMsgInfoGen2;
-
-	switch (wac_info->mode) {
-	/* Send command to single wave */
-	case HSA_DBG_WAVEMODE_SINGLE:
-		/*
-		 * Limit access to the process waves only,
-		 * by setting vmid check
-		 */
-		reg_sq_cmd.bits.check_vmid = 1;
-		reg_sq_cmd.bits.simd_id = pMsg->ui32.SIMD;
-		reg_sq_cmd.bits.wave_id = pMsg->ui32.WaveId;
-		reg_sq_cmd.bits.mode = SQ_IND_CMD_MODE_SINGLE;
-
-		reg_gfx_index.bits.sh_index = pMsg->ui32.ShaderArray;
-		reg_gfx_index.bits.se_index = pMsg->ui32.ShaderEngine;
-		reg_gfx_index.bits.instance_index = pMsg->ui32.HSACU;
-
-		break;
-
-	/* Send command to all waves with matching VMID */
-	case HSA_DBG_WAVEMODE_BROADCAST_PROCESS:
-
-		reg_gfx_index.bits.sh_broadcast_writes = 1;
-		reg_gfx_index.bits.se_broadcast_writes = 1;
-		reg_gfx_index.bits.instance_broadcast_writes = 1;
-
-		reg_sq_cmd.bits.mode = SQ_IND_CMD_MODE_BROADCAST;
-
-		break;
-
-	/* Send command to all CU waves with matching VMID */
-	case HSA_DBG_WAVEMODE_BROADCAST_PROCESS_CU:
-
-		reg_sq_cmd.bits.check_vmid = 1;
-		reg_sq_cmd.bits.mode = SQ_IND_CMD_MODE_BROADCAST;
-
-		reg_gfx_index.bits.sh_index = pMsg->ui32.ShaderArray;
-		reg_gfx_index.bits.se_index = pMsg->ui32.ShaderEngine;
-		reg_gfx_index.bits.instance_index = pMsg->ui32.HSACU;
-
-		break;
-
-	default:
-		return -EINVAL;
-	}
-
-	switch (wac_info->operand) {
-	case HSA_DBG_WAVEOP_HALT:
-		reg_sq_cmd.bits.cmd = SQ_IND_CMD_CMD_HALT;
-		break;
-
-	case HSA_DBG_WAVEOP_RESUME:
-		reg_sq_cmd.bits.cmd = SQ_IND_CMD_CMD_RESUME;
-		break;
-
-	case HSA_DBG_WAVEOP_KILL:
-		reg_sq_cmd.bits.cmd = SQ_IND_CMD_CMD_KILL;
-		break;
-
-	case HSA_DBG_WAVEOP_DEBUG:
-		reg_sq_cmd.bits.cmd = SQ_IND_CMD_CMD_DEBUG;
-		break;
-
-	case HSA_DBG_WAVEOP_TRAP:
-		if (wac_info->trapId < MAX_TRAPID) {
-			reg_sq_cmd.bits.cmd = SQ_IND_CMD_CMD_TRAP;
-			reg_sq_cmd.bits.trap_id = wac_info->trapId;
-		} else {
-			status = -EINVAL;
-		}
-		break;
-
-	default:
-		status = -EINVAL;
-		break;
-	}
-
-	if (status == 0) {
-		*in_reg_sq_cmd = reg_sq_cmd;
-		*in_reg_gfx_index = reg_gfx_index;
-	}
-
-	return status;
-}
-
-static int dbgdev_wave_control_diq(struct kfd_dbgdev *dbgdev,
-					struct dbg_wave_control_info *wac_info)
-{
-
-	int status;
-	union SQ_CMD_BITS reg_sq_cmd;
-	union GRBM_GFX_INDEX_BITS reg_gfx_index;
-	struct kfd_mem_obj *mem_obj;
-	uint32_t *packet_buff_uint;
-	struct pm4__set_config_reg *packets_vec;
-	size_t ib_size = sizeof(struct pm4__set_config_reg) * 3;
-
-	reg_sq_cmd.u32All = 0;
-
-	status = dbgdev_wave_control_set_registers(wac_info, &reg_sq_cmd,
-							&reg_gfx_index);
-	if (status) {
-		pr_err("Failed to set wave control registers\n");
-		return status;
-	}
-
-	/* we do not control the VMID in DIQ, so reset it to a known value */
-	reg_sq_cmd.bits.vm_id = 0;
-
-	pr_debug("\t\t %30s\n", "* * * * * * * * * * * * * * * * * *");
-
-	pr_debug("\t\t mode      is: %u\n", wac_info->mode);
-	pr_debug("\t\t operand   is: %u\n", wac_info->operand);
-	pr_debug("\t\t trap id   is: %u\n", wac_info->trapId);
-	pr_debug("\t\t msg value is: %u\n",
-			wac_info->dbgWave_msg.DbgWaveMsg.WaveMsgInfoGen2.Value);
-	pr_debug("\t\t vmid      is: N/A\n");
-
-	pr_debug("\t\t chk_vmid  is : %u\n", reg_sq_cmd.bitfields.check_vmid);
-	pr_debug("\t\t command   is : %u\n", reg_sq_cmd.bitfields.cmd);
-	pr_debug("\t\t queue id  is : %u\n", reg_sq_cmd.bitfields.queue_id);
-	pr_debug("\t\t simd id   is : %u\n", reg_sq_cmd.bitfields.simd_id);
-	pr_debug("\t\t mode      is : %u\n", reg_sq_cmd.bitfields.mode);
-	pr_debug("\t\t vm_id     is : %u\n", reg_sq_cmd.bitfields.vm_id);
-	pr_debug("\t\t wave_id   is : %u\n", reg_sq_cmd.bitfields.wave_id);
-
-	pr_debug("\t\t ibw       is : %u\n",
-			reg_gfx_index.bitfields.instance_broadcast_writes);
-	pr_debug("\t\t ii        is : %u\n",
-			reg_gfx_index.bitfields.instance_index);
-	pr_debug("\t\t sebw      is : %u\n",
-			reg_gfx_index.bitfields.se_broadcast_writes);
-	pr_debug("\t\t se_ind    is : %u\n", reg_gfx_index.bitfields.se_index);
-	pr_debug("\t\t sh_ind    is : %u\n", reg_gfx_index.bitfields.sh_index);
-	pr_debug("\t\t sbw       is : %u\n",
-			reg_gfx_index.bitfields.sh_broadcast_writes);
-
-	pr_debug("\t\t %30s\n", "* * * * * * * * * * * * * * * * * *");
-
-	status = kfd_gtt_sa_allocate(dbgdev->dev, ib_size, &mem_obj);
-
-	if (status != 0) {
-		pr_err("Failed to allocate GART memory\n");
-		return status;
-	}
-
-	packet_buff_uint = mem_obj->cpu_ptr;
-
-	memset(packet_buff_uint, 0, ib_size);
-
-	packets_vec =  (struct pm4__set_config_reg *) packet_buff_uint;
-	packets_vec[0].header.count = 1;
-	packets_vec[0].header.opcode = IT_SET_UCONFIG_REG;
-	packets_vec[0].header.type = PM4_TYPE_3;
-	packets_vec[0].bitfields2.reg_offset =
-			GRBM_GFX_INDEX / 4 - USERCONFIG_REG_BASE;
-
-	packets_vec[0].bitfields2.insert_vmid = 0;
-	packets_vec[0].reg_data[0] = reg_gfx_index.u32All;
-
-	packets_vec[1].header.count = 1;
-	packets_vec[1].header.opcode = IT_SET_CONFIG_REG;
-	packets_vec[1].header.type = PM4_TYPE_3;
-	packets_vec[1].bitfields2.reg_offset = SQ_CMD / 4 - AMD_CONFIG_REG_BASE;
-
-	packets_vec[1].bitfields2.vmid_shift = SQ_CMD_VMID_OFFSET;
-	packets_vec[1].bitfields2.insert_vmid = 1;
-	packets_vec[1].reg_data[0] = reg_sq_cmd.u32All;
-
-	/* Restore the GRBM_GFX_INDEX register */
-
-	reg_gfx_index.u32All = 0;
-	reg_gfx_index.bits.sh_broadcast_writes = 1;
-	reg_gfx_index.bits.instance_broadcast_writes = 1;
-	reg_gfx_index.bits.se_broadcast_writes = 1;
-
-
-	packets_vec[2].ordinal1 = packets_vec[0].ordinal1;
-	packets_vec[2].bitfields2.reg_offset =
-				GRBM_GFX_INDEX / 4 - USERCONFIG_REG_BASE;
-
-	packets_vec[2].bitfields2.insert_vmid = 0;
-	packets_vec[2].reg_data[0] = reg_gfx_index.u32All;
-
-	status = dbgdev_diq_submit_ib(
-			dbgdev,
-			wac_info->process->pasid,
-			mem_obj->gpu_addr,
-			packet_buff_uint,
-			ib_size);
-
-	if (status)
-		pr_err("Failed to submit IB to DIQ\n");
-
-	kfd_gtt_sa_free(dbgdev->dev, mem_obj);
-
-	return status;
-}
-
-static int dbgdev_wave_control_nodiq(struct kfd_dbgdev *dbgdev,
-					struct dbg_wave_control_info *wac_info)
-{
-	int status;
-	union SQ_CMD_BITS reg_sq_cmd;
-	union GRBM_GFX_INDEX_BITS reg_gfx_index;
-	struct kfd_process_device *pdd;
-
-	reg_sq_cmd.u32All = 0;
-
-	/* taking the VMID for that process on the safe way using PDD */
-	pdd = kfd_get_process_device_data(dbgdev->dev, wac_info->process);
-
-	if (!pdd) {
-		pr_err("Failed to get pdd for wave control no DIQ\n");
-		return -EFAULT;
-	}
-	status = dbgdev_wave_control_set_registers(wac_info, &reg_sq_cmd,
-							&reg_gfx_index);
-	if (status) {
-		pr_err("Failed to set wave control registers\n");
-		return status;
-	}
-
-	/* for non DIQ we need to patch the VMID: */
-
-	reg_sq_cmd.bits.vm_id = pdd->qpd.vmid;
-
-	pr_debug("\t\t %30s\n", "* * * * * * * * * * * * * * * * * *");
-
-	pr_debug("\t\t mode      is: %u\n", wac_info->mode);
-	pr_debug("\t\t operand   is: %u\n", wac_info->operand);
-	pr_debug("\t\t trap id   is: %u\n", wac_info->trapId);
-	pr_debug("\t\t msg value is: %u\n",
-			wac_info->dbgWave_msg.DbgWaveMsg.WaveMsgInfoGen2.Value);
-	pr_debug("\t\t vmid      is: %u\n", pdd->qpd.vmid);
-
-	pr_debug("\t\t chk_vmid  is : %u\n", reg_sq_cmd.bitfields.check_vmid);
-	pr_debug("\t\t command   is : %u\n", reg_sq_cmd.bitfields.cmd);
-	pr_debug("\t\t queue id  is : %u\n", reg_sq_cmd.bitfields.queue_id);
-	pr_debug("\t\t simd id   is : %u\n", reg_sq_cmd.bitfields.simd_id);
-	pr_debug("\t\t mode      is : %u\n", reg_sq_cmd.bitfields.mode);
-	pr_debug("\t\t vm_id     is : %u\n", reg_sq_cmd.bitfields.vm_id);
-	pr_debug("\t\t wave_id   is : %u\n", reg_sq_cmd.bitfields.wave_id);
-
-	pr_debug("\t\t ibw       is : %u\n",
-			reg_gfx_index.bitfields.instance_broadcast_writes);
-	pr_debug("\t\t ii        is : %u\n",
-			reg_gfx_index.bitfields.instance_index);
-	pr_debug("\t\t sebw      is : %u\n",
-			reg_gfx_index.bitfields.se_broadcast_writes);
-	pr_debug("\t\t se_ind    is : %u\n", reg_gfx_index.bitfields.se_index);
-	pr_debug("\t\t sh_ind    is : %u\n", reg_gfx_index.bitfields.sh_index);
-	pr_debug("\t\t sbw       is : %u\n",
-			reg_gfx_index.bitfields.sh_broadcast_writes);
-
-	pr_debug("\t\t %30s\n", "* * * * * * * * * * * * * * * * * *");
-
-	return dbgdev->dev->kfd2kgd->wave_control_execute(dbgdev->dev->kgd,
-							reg_gfx_index.u32All,
-							reg_sq_cmd.u32All);
-}
-
-int dbgdev_wave_reset_wavefronts(struct kfd_dev *dev, struct kfd_process *p)
-{
-	int status = 0;
-	unsigned int vmid;
-	uint16_t queried_pasid;
-	union SQ_CMD_BITS reg_sq_cmd;
-	union GRBM_GFX_INDEX_BITS reg_gfx_index;
-	struct kfd_process_device *pdd;
-	struct dbg_wave_control_info wac_info;
-	int first_vmid_to_scan = dev->vm_info.first_vmid_kfd;
-	int last_vmid_to_scan = dev->vm_info.last_vmid_kfd;
-
-	reg_sq_cmd.u32All = 0;
-	status = 0;
-
-	wac_info.mode = HSA_DBG_WAVEMODE_BROADCAST_PROCESS;
-	wac_info.operand = HSA_DBG_WAVEOP_KILL;
-
-	pr_debug("Killing all process wavefronts\n");
-
-	/* Scan all registers in the range ATC_VMID8_PASID_MAPPING ..
-	 * ATC_VMID15_PASID_MAPPING
-	 * to check which VMID the current process is mapped to.
-	 */
-
-	for (vmid = first_vmid_to_scan; vmid <= last_vmid_to_scan; vmid++) {
-		status = dev->kfd2kgd->get_atc_vmid_pasid_mapping_info
-				(dev->kgd, vmid, &queried_pasid);
-
-		if (status && queried_pasid == p->pasid) {
-			pr_debug("Killing wave fronts of vmid %d and pasid 0x%x\n",
-					vmid, p->pasid);
-			break;
-		}
-	}
-
-	if (vmid > last_vmid_to_scan) {
-		pr_err("Didn't find vmid for pasid 0x%x\n", p->pasid);
-		return -EFAULT;
-	}
-
-	/* taking the VMID for that process on the safe way using PDD */
-	pdd = kfd_get_process_device_data(dev, p);
-	if (!pdd)
-		return -EFAULT;
-
-	status = dbgdev_wave_control_set_registers(&wac_info, &reg_sq_cmd,
-			&reg_gfx_index);
-	if (status != 0)
-		return -EINVAL;
-
-	/* for non DIQ we need to patch the VMID: */
-	reg_sq_cmd.bits.vm_id = vmid;
-
-	dev->kfd2kgd->wave_control_execute(dev->kgd,
-					reg_gfx_index.u32All,
-					reg_sq_cmd.u32All);
-
-	return 0;
-}
-
-void kfd_dbgdev_init(struct kfd_dbgdev *pdbgdev, struct kfd_dev *pdev,
-			enum DBGDEV_TYPE type)
-{
-	pdbgdev->dev = pdev;
-	pdbgdev->kq = NULL;
-	pdbgdev->type = type;
-	pdbgdev->pqm = NULL;
-
-	switch (type) {
-	case DBGDEV_TYPE_NODIQ:
-		pdbgdev->dbgdev_register = dbgdev_register_nodiq;
-		pdbgdev->dbgdev_unregister = dbgdev_unregister_nodiq;
-		pdbgdev->dbgdev_wave_control = dbgdev_wave_control_nodiq;
-		pdbgdev->dbgdev_address_watch = dbgdev_address_watch_nodiq;
-		break;
-	case DBGDEV_TYPE_DIQ:
-	default:
-		pdbgdev->dbgdev_register = dbgdev_register_diq;
-		pdbgdev->dbgdev_unregister = dbgdev_unregister_diq;
-		pdbgdev->dbgdev_wave_control =  dbgdev_wave_control_diq;
-		pdbgdev->dbgdev_address_watch = dbgdev_address_watch_diq;
-		break;
-	}
-
-}
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_dbgdev.h b/drivers/gpu/drm/amd/amdkfd/kfd_dbgdev.h
deleted file mode 100644
index 0619c777b47e..000000000000
--- a/drivers/gpu/drm/amd/amdkfd/kfd_dbgdev.h
+++ /dev/null
@@ -1,230 +0,0 @@
-/*
- * Copyright 2014 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- */
-
-#ifndef KFD_DBGDEV_H_
-#define KFD_DBGDEV_H_
-
-enum {
-	SQ_CMD_VMID_OFFSET = 28,
-	ADDRESS_WATCH_CNTL_OFFSET = 24
-};
-
-enum {
-	PRIV_QUEUE_SYNC_TIME_MS = 200
-};
-
-/* CONTEXT reg space definition */
-enum {
-	CONTEXT_REG_BASE = 0xA000,
-	CONTEXT_REG_END = 0xA400,
-	CONTEXT_REG_SIZE = CONTEXT_REG_END - CONTEXT_REG_BASE
-};
-
-/* USER CONFIG reg space definition */
-enum {
-	USERCONFIG_REG_BASE = 0xC000,
-	USERCONFIG_REG_END = 0x10000,
-	USERCONFIG_REG_SIZE = USERCONFIG_REG_END - USERCONFIG_REG_BASE
-};
-
-/* CONFIG reg space definition */
-enum {
-	AMD_CONFIG_REG_BASE = 0x2000,	/* in dwords */
-	AMD_CONFIG_REG_END = 0x2B00,
-	AMD_CONFIG_REG_SIZE = AMD_CONFIG_REG_END - AMD_CONFIG_REG_BASE
-};
-
-/* SH reg space definition */
-enum {
-	SH_REG_BASE = 0x2C00,
-	SH_REG_END = 0x3000,
-	SH_REG_SIZE = SH_REG_END - SH_REG_BASE
-};
-
-/* SQ_CMD definitions */
-#define SQ_CMD						0x8DEC
-
-enum SQ_IND_CMD_CMD {
-	SQ_IND_CMD_CMD_NULL = 0x00000000,
-	SQ_IND_CMD_CMD_HALT = 0x00000001,
-	SQ_IND_CMD_CMD_RESUME = 0x00000002,
-	SQ_IND_CMD_CMD_KILL = 0x00000003,
-	SQ_IND_CMD_CMD_DEBUG = 0x00000004,
-	SQ_IND_CMD_CMD_TRAP = 0x00000005,
-};
-
-enum SQ_IND_CMD_MODE {
-	SQ_IND_CMD_MODE_SINGLE = 0x00000000,
-	SQ_IND_CMD_MODE_BROADCAST = 0x00000001,
-	SQ_IND_CMD_MODE_BROADCAST_QUEUE = 0x00000002,
-	SQ_IND_CMD_MODE_BROADCAST_PIPE = 0x00000003,
-	SQ_IND_CMD_MODE_BROADCAST_ME = 0x00000004,
-};
-
-union SQ_IND_INDEX_BITS {
-	struct {
-		uint32_t wave_id:4;
-		uint32_t simd_id:2;
-		uint32_t thread_id:6;
-		 uint32_t:1;
-		uint32_t force_read:1;
-		uint32_t read_timeout:1;
-		uint32_t unindexed:1;
-		uint32_t index:16;
-
-	} bitfields, bits;
-	uint32_t u32All;
-	signed int i32All;
-	float f32All;
-};
-
-union SQ_IND_CMD_BITS {
-	struct {
-		uint32_t data:32;
-	} bitfields, bits;
-	uint32_t u32All;
-	signed int i32All;
-	float f32All;
-};
-
-union SQ_CMD_BITS {
-	struct {
-		uint32_t cmd:3;
-		 uint32_t:1;
-		uint32_t mode:3;
-		uint32_t check_vmid:1;
-		uint32_t trap_id:3;
-		 uint32_t:5;
-		uint32_t wave_id:4;
-		uint32_t simd_id:2;
-		 uint32_t:2;
-		uint32_t queue_id:3;
-		 uint32_t:1;
-		uint32_t vm_id:4;
-	} bitfields, bits;
-	uint32_t u32All;
-	signed int i32All;
-	float f32All;
-};
-
-union SQ_IND_DATA_BITS {
-	struct {
-		uint32_t data:32;
-	} bitfields, bits;
-	uint32_t u32All;
-	signed int i32All;
-	float f32All;
-};
-
-union GRBM_GFX_INDEX_BITS {
-	struct {
-		uint32_t instance_index:8;
-		uint32_t sh_index:8;
-		uint32_t se_index:8;
-		 uint32_t:5;
-		uint32_t sh_broadcast_writes:1;
-		uint32_t instance_broadcast_writes:1;
-		uint32_t se_broadcast_writes:1;
-	} bitfields, bits;
-	uint32_t u32All;
-	signed int i32All;
-	float f32All;
-};
-
-union TCP_WATCH_ADDR_H_BITS {
-	struct {
-		uint32_t addr:16;
-		 uint32_t:16;
-
-	} bitfields, bits;
-	uint32_t u32All;
-	signed int i32All;
-	float f32All;
-};
-
-union TCP_WATCH_ADDR_L_BITS {
-	struct {
-		uint32_t:6;
-		uint32_t addr:26;
-	} bitfields, bits;
-	uint32_t u32All;
-	signed int i32All;
-	float f32All;
-};
-
-enum {
-	QUEUESTATE__INVALID = 0, /* so by default we'll get invalid state */
-	QUEUESTATE__ACTIVE_COMPLETION_PENDING,
-	QUEUESTATE__ACTIVE
-};
-
-union ULARGE_INTEGER {
-	struct {
-		uint32_t low_part;
-		uint32_t high_part;
-	} u;
-	unsigned long long quad_part;
-};
-
-
-#define KFD_CIK_VMID_START_OFFSET (8)
-#define KFD_CIK_VMID_END_OFFSET (KFD_CIK_VMID_START_OFFSET + (8))
-
-
-void kfd_dbgdev_init(struct kfd_dbgdev *pdbgdev, struct kfd_dev *pdev,
-			enum DBGDEV_TYPE type);
-
-union TCP_WATCH_CNTL_BITS {
-	struct {
-		uint32_t mask:24;
-		uint32_t vmid:4;
-		uint32_t atc:1;
-		uint32_t mode:2;
-		uint32_t valid:1;
-	} bitfields, bits;
-	uint32_t u32All;
-	signed int i32All;
-	float f32All;
-};
-
-enum {
-	ADDRESS_WATCH_REG_CNTL_ATC_BIT = 0x10000000UL,
-	ADDRESS_WATCH_REG_CNTL_DEFAULT_MASK = 0x00FFFFFF,
-	ADDRESS_WATCH_REG_ADDLOW_MASK_EXTENSION = 0x03000000,
-	/* extend the mask to 26 bits in order to match the low address field */
-	ADDRESS_WATCH_REG_ADDLOW_SHIFT = 6,
-	ADDRESS_WATCH_REG_ADDHIGH_MASK = 0xFFFF
-};
-
-enum {
-	MAX_TRAPID = 8,		/* 3 bits in the bitfield. */
-	MAX_WATCH_ADDRESSES = 4
-};
-
-enum {
-	ADDRESS_WATCH_REG_ADDR_HI = 0,
-	ADDRESS_WATCH_REG_ADDR_LO,
-	ADDRESS_WATCH_REG_CNTL,
-	ADDRESS_WATCH_REG_MAX
-};
-
-#endif	/* KFD_DBGDEV_H_ */
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_dbgmgr.c b/drivers/gpu/drm/amd/amdkfd/kfd_dbgmgr.c
deleted file mode 100644
index 9bfa50633654..000000000000
--- a/drivers/gpu/drm/amd/amdkfd/kfd_dbgmgr.c
+++ /dev/null
@@ -1,158 +0,0 @@
-/*
- * Copyright 2014 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- */
-#include <linux/types.h>
-#include <linux/kernel.h>
-#include <linux/log2.h>
-#include <linux/sched.h>
-#include <linux/slab.h>
-#include <linux/device.h>
-
-#include "kfd_priv.h"
-#include "cik_regs.h"
-#include "kfd_pm4_headers.h"
-#include "kfd_pm4_headers_diq.h"
-#include "kfd_dbgmgr.h"
-#include "kfd_dbgdev.h"
-#include "kfd_device_queue_manager.h"
-
-static DEFINE_MUTEX(kfd_dbgmgr_mutex);
-
-struct mutex *kfd_get_dbgmgr_mutex(void)
-{
-	return &kfd_dbgmgr_mutex;
-}
-
-
-static void kfd_dbgmgr_uninitialize(struct kfd_dbgmgr *pmgr)
-{
-	kfree(pmgr->dbgdev);
-
-	pmgr->dbgdev = NULL;
-	pmgr->pasid = 0;
-	pmgr->dev = NULL;
-}
-
-void kfd_dbgmgr_destroy(struct kfd_dbgmgr *pmgr)
-{
-	if (pmgr) {
-		kfd_dbgmgr_uninitialize(pmgr);
-		kfree(pmgr);
-	}
-}
-
-bool kfd_dbgmgr_create(struct kfd_dbgmgr **ppmgr, struct kfd_dev *pdev)
-{
-	enum DBGDEV_TYPE type = DBGDEV_TYPE_DIQ;
-	struct kfd_dbgmgr *new_buff;
-
-	if (WARN_ON(!pdev->init_complete))
-		return false;
-
-	new_buff = kfd_alloc_struct(new_buff);
-	if (!new_buff) {
-		pr_err("Failed to allocate dbgmgr instance\n");
-		return false;
-	}
-
-	new_buff->pasid = 0;
-	new_buff->dev = pdev;
-	new_buff->dbgdev = kfd_alloc_struct(new_buff->dbgdev);
-	if (!new_buff->dbgdev) {
-		pr_err("Failed to allocate dbgdev instance\n");
-		kfree(new_buff);
-		return false;
-	}
-
-	/* get actual type of DBGDevice cpsch or not */
-	if (pdev->dqm->sched_policy == KFD_SCHED_POLICY_NO_HWS)
-		type = DBGDEV_TYPE_NODIQ;
-
-	kfd_dbgdev_init(new_buff->dbgdev, pdev, type);
-	*ppmgr = new_buff;
-
-	return true;
-}
-
-long kfd_dbgmgr_register(struct kfd_dbgmgr *pmgr, struct kfd_process *p)
-{
-	if (pmgr->pasid != 0) {
-		pr_debug("H/W debugger is already active using pasid 0x%x\n",
-				pmgr->pasid);
-		return -EBUSY;
-	}
-
-	/* remember pasid */
-	pmgr->pasid = p->pasid;
-
-	/* provide the pqm for diq generation */
-	pmgr->dbgdev->pqm = &p->pqm;
-
-	/* activate the actual registering */
-	pmgr->dbgdev->dbgdev_register(pmgr->dbgdev);
-
-	return 0;
-}
-
-long kfd_dbgmgr_unregister(struct kfd_dbgmgr *pmgr, struct kfd_process *p)
-{
-	/* Is the requests coming from the already registered process? */
-	if (pmgr->pasid != p->pasid) {
-		pr_debug("H/W debugger is not registered by calling pasid 0x%x\n",
-				p->pasid);
-		return -EINVAL;
-	}
-
-	pmgr->dbgdev->dbgdev_unregister(pmgr->dbgdev);
-
-	pmgr->pasid = 0;
-
-	return 0;
-}
-
-long kfd_dbgmgr_wave_control(struct kfd_dbgmgr *pmgr,
-				struct dbg_wave_control_info *wac_info)
-{
-	/* Is the requests coming from the already registered process? */
-	if (pmgr->pasid != wac_info->process->pasid) {
-		pr_debug("H/W debugger support was not registered for requester pasid 0x%x\n",
-				wac_info->process->pasid);
-		return -EINVAL;
-	}
-
-	return (long) pmgr->dbgdev->dbgdev_wave_control(pmgr->dbgdev, wac_info);
-}
-
-long kfd_dbgmgr_address_watch(struct kfd_dbgmgr *pmgr,
-				struct dbg_address_watch_info *adw_info)
-{
-	/* Is the requests coming from the already registered process? */
-	if (pmgr->pasid != adw_info->process->pasid) {
-		pr_debug("H/W debugger support was not registered for requester pasid 0x%x\n",
-				adw_info->process->pasid);
-		return -EINVAL;
-	}
-
-	return (long) pmgr->dbgdev->dbgdev_address_watch(pmgr->dbgdev,
-							adw_info);
-}
-
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_dbgmgr.h b/drivers/gpu/drm/amd/amdkfd/kfd_dbgmgr.h
deleted file mode 100644
index f9c6df1fdc5c..000000000000
--- a/drivers/gpu/drm/amd/amdkfd/kfd_dbgmgr.h
+++ /dev/null
@@ -1,293 +0,0 @@
-/*
- * Copyright 2014 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- */
-
-#ifndef KFD_DBGMGR_H_
-#define KFD_DBGMGR_H_
-
-#include "kfd_priv.h"
-
-/* must align with hsakmttypes definition */
-#pragma pack(push, 4)
-
-enum HSA_DBG_WAVEOP {
-	HSA_DBG_WAVEOP_HALT = 1,   /* Halts a wavefront */
-	HSA_DBG_WAVEOP_RESUME = 2, /* Resumes a wavefront */
-	HSA_DBG_WAVEOP_KILL = 3,   /* Kills a wavefront */
-	HSA_DBG_WAVEOP_DEBUG = 4,  /* Causes wavefront to enter dbg mode */
-	HSA_DBG_WAVEOP_TRAP = 5,   /* Causes wavefront to take a trap */
-	HSA_DBG_NUM_WAVEOP = 5,
-	HSA_DBG_MAX_WAVEOP = 0xFFFFFFFF
-};
-
-enum HSA_DBG_WAVEMODE {
-	/* send command to a single wave */
-	HSA_DBG_WAVEMODE_SINGLE = 0,
-	/*
-	 * Broadcast to all wavefronts of all processes is not
-	 * supported for HSA user mode
-	 */
-
-	/* send to waves within current process */
-	HSA_DBG_WAVEMODE_BROADCAST_PROCESS = 2,
-	/* send to waves within current process on CU  */
-	HSA_DBG_WAVEMODE_BROADCAST_PROCESS_CU = 3,
-	HSA_DBG_NUM_WAVEMODE = 3,
-	HSA_DBG_MAX_WAVEMODE = 0xFFFFFFFF
-};
-
-enum HSA_DBG_WAVEMSG_TYPE {
-	HSA_DBG_WAVEMSG_AUTO = 0,
-	HSA_DBG_WAVEMSG_USER = 1,
-	HSA_DBG_WAVEMSG_ERROR = 2,
-	HSA_DBG_NUM_WAVEMSG,
-	HSA_DBG_MAX_WAVEMSG = 0xFFFFFFFF
-};
-
-enum HSA_DBG_WATCH_MODE {
-	HSA_DBG_WATCH_READ = 0,		/* Read operations only */
-	HSA_DBG_WATCH_NONREAD = 1,	/* Write or Atomic operations only */
-	HSA_DBG_WATCH_ATOMIC = 2,	/* Atomic Operations only */
-	HSA_DBG_WATCH_ALL = 3,		/* Read, Write or Atomic operations */
-	HSA_DBG_WATCH_NUM,
-	HSA_DBG_WATCH_SIZE = 0xFFFFFFFF
-};
-
-/* This structure is hardware specific and may change in the future */
-struct HsaDbgWaveMsgAMDGen2 {
-	union {
-		struct ui32 {
-			uint32_t UserData:8;	/* user data */
-			uint32_t ShaderArray:1;	/* Shader array */
-			uint32_t Priv:1;	/* Privileged */
-			uint32_t Reserved0:4;	/* Reserved, should be 0 */
-			uint32_t WaveId:4;	/* wave id */
-			uint32_t SIMD:2;	/* SIMD id */
-			uint32_t HSACU:4;	/* Compute unit */
-			uint32_t ShaderEngine:2;/* Shader engine */
-			uint32_t MessageType:2;	/* see HSA_DBG_WAVEMSG_TYPE */
-			uint32_t Reserved1:4;	/* Reserved, should be 0 */
-		} ui32;
-		uint32_t Value;
-	};
-	uint32_t Reserved2;
-};
-
-union HsaDbgWaveMessageAMD {
-	struct HsaDbgWaveMsgAMDGen2 WaveMsgInfoGen2;
-	/* for future HsaDbgWaveMsgAMDGen3; */
-};
-
-struct HsaDbgWaveMessage {
-	void *MemoryVA;		/* ptr to associated host-accessible data */
-	union HsaDbgWaveMessageAMD DbgWaveMsg;
-};
-
-/*
- * TODO: This definitions to be MOVED to kfd_event, once it is implemented.
- *
- * HSA sync primitive, Event and HW Exception notification API definitions.
- * The API functions allow the runtime to define a so-called sync-primitive,
- * a SW object combining a user-mode provided "syncvar" and a scheduler event
- * that can be signaled through a defined GPU interrupt. A syncvar is
- * a process virtual memory location of a certain size that can be accessed
- * by CPU and GPU shader code within the process to set and query the content
- * within that memory. The definition of the content is determined by the HSA
- * runtime and potentially GPU shader code interfacing with the HSA runtime.
- * The syncvar values may be commonly written through an PM4 WRITE_DATA packet
- * in the user mode instruction stream. The OS scheduler event is typically
- * associated and signaled by an interrupt issued by the GPU, but other HSA
- * system interrupt conditions from other HW (e.g. IOMMUv2) may be surfaced
- * by the KFD by this mechanism, too.
- */
-
-/* these are the new definitions for events */
-enum HSA_EVENTTYPE {
-	HSA_EVENTTYPE_SIGNAL = 0,	/* user-mode generated GPU signal */
-	HSA_EVENTTYPE_NODECHANGE = 1,	/* HSA node change (attach/detach) */
-	HSA_EVENTTYPE_DEVICESTATECHANGE = 2,	/* HSA device state change
-						 * (start/stop)
-						 */
-	HSA_EVENTTYPE_HW_EXCEPTION = 3,	/* GPU shader exception event */
-	HSA_EVENTTYPE_SYSTEM_EVENT = 4,	/* GPU SYSCALL with parameter info */
-	HSA_EVENTTYPE_DEBUG_EVENT = 5,	/* GPU signal for debugging */
-	HSA_EVENTTYPE_PROFILE_EVENT = 6,/* GPU signal for profiling */
-	HSA_EVENTTYPE_QUEUE_EVENT = 7,	/* GPU signal queue idle state
-					 * (EOP pm4)
-					 */
-	/* ...  */
-	HSA_EVENTTYPE_MAXID,
-	HSA_EVENTTYPE_TYPE_SIZE = 0xFFFFFFFF
-};
-
-/* Sub-definitions for various event types: Syncvar */
-struct HsaSyncVar {
-	union SyncVar {
-		void *UserData;	/* pointer to user mode data */
-		uint64_t UserDataPtrValue; /* 64bit compatibility of value */
-	} SyncVar;
-	uint64_t SyncVarSize;
-};
-
-/* Sub-definitions for various event types: NodeChange */
-
-enum HSA_EVENTTYPE_NODECHANGE_FLAGS {
-	HSA_EVENTTYPE_NODECHANGE_ADD = 0,
-	HSA_EVENTTYPE_NODECHANGE_REMOVE = 1,
-	HSA_EVENTTYPE_NODECHANGE_SIZE = 0xFFFFFFFF
-};
-
-struct HsaNodeChange {
-	/* HSA node added/removed on the platform */
-	enum HSA_EVENTTYPE_NODECHANGE_FLAGS Flags;
-};
-
-/* Sub-definitions for various event types: DeviceStateChange */
-enum HSA_EVENTTYPE_DEVICESTATECHANGE_FLAGS {
-	/* device started (and available) */
-	HSA_EVENTTYPE_DEVICESTATUSCHANGE_START = 0,
-	/* device stopped (i.e. unavailable) */
-	HSA_EVENTTYPE_DEVICESTATUSCHANGE_STOP = 1,
-	HSA_EVENTTYPE_DEVICESTATUSCHANGE_SIZE = 0xFFFFFFFF
-};
-
-enum HSA_DEVICE {
-	HSA_DEVICE_CPU = 0,
-	HSA_DEVICE_GPU = 1,
-	MAX_HSA_DEVICE = 2
-};
-
-struct HsaDeviceStateChange {
-	uint32_t NodeId;	/* F-NUMA node that contains the device */
-	enum HSA_DEVICE Device;	/* device type: GPU or CPU */
-	enum HSA_EVENTTYPE_DEVICESTATECHANGE_FLAGS Flags; /* event flags */
-};
-
-struct HsaEventData {
-	enum HSA_EVENTTYPE EventType; /* event type */
-	union EventData {
-		/*
-		 * return data associated with HSA_EVENTTYPE_SIGNAL
-		 * and other events
-		 */
-		struct HsaSyncVar SyncVar;
-
-		/* data associated with HSA_EVENTTYPE_NODE_CHANGE */
-		struct HsaNodeChange NodeChangeState;
-
-		/* data associated with HSA_EVENTTYPE_DEVICE_STATE_CHANGE */
-		struct HsaDeviceStateChange DeviceState;
-	} EventData;
-
-	/* the following data entries are internal to the KFD & thunk itself */
-
-	/* internal thunk store for Event data (OsEventHandle) */
-	uint64_t HWData1;
-	/* internal thunk store for Event data (HWAddress) */
-	uint64_t HWData2;
-	/* internal thunk store for Event data (HWData) */
-	uint32_t HWData3;
-};
-
-struct HsaEventDescriptor {
-	/* event type to allocate */
-	enum HSA_EVENTTYPE EventType;
-	/* H-NUMA node containing GPU device that is event source */
-	uint32_t NodeId;
-	/* pointer to user mode syncvar data, syncvar->UserDataPtrValue
-	 * may be NULL
-	 */
-	struct HsaSyncVar SyncVar;
-};
-
-struct HsaEvent {
-	uint32_t EventId;
-	struct HsaEventData EventData;
-};
-
-#pragma pack(pop)
-
-enum DBGDEV_TYPE {
-	DBGDEV_TYPE_ILLEGAL = 0,
-	DBGDEV_TYPE_NODIQ = 1,
-	DBGDEV_TYPE_DIQ = 2,
-	DBGDEV_TYPE_TEST = 3
-};
-
-struct dbg_address_watch_info {
-	struct kfd_process *process;
-	enum HSA_DBG_WATCH_MODE *watch_mode;
-	uint64_t *watch_address;
-	uint64_t *watch_mask;
-	struct HsaEvent *watch_event;
-	uint32_t num_watch_points;
-};
-
-struct dbg_wave_control_info {
-	struct kfd_process *process;
-	uint32_t trapId;
-	enum HSA_DBG_WAVEOP operand;
-	enum HSA_DBG_WAVEMODE mode;
-	struct HsaDbgWaveMessage dbgWave_msg;
-};
-
-struct kfd_dbgdev {
-
-	/* The device that owns this data. */
-	struct kfd_dev *dev;
-
-	/* kernel queue for DIQ */
-	struct kernel_queue *kq;
-
-	/* a pointer to the pqm of the calling process */
-	struct process_queue_manager *pqm;
-
-	/* type of debug device ( DIQ, non DIQ, etc. ) */
-	enum DBGDEV_TYPE type;
-
-	/* virtualized function pointers to device dbg */
-	int (*dbgdev_register)(struct kfd_dbgdev *dbgdev);
-	int (*dbgdev_unregister)(struct kfd_dbgdev *dbgdev);
-	int (*dbgdev_address_watch)(struct kfd_dbgdev *dbgdev,
-				struct dbg_address_watch_info *adw_info);
-	int (*dbgdev_wave_control)(struct kfd_dbgdev *dbgdev,
-				struct dbg_wave_control_info *wac_info);
-
-};
-
-struct kfd_dbgmgr {
-	u32 pasid;
-	struct kfd_dev *dev;
-	struct kfd_dbgdev *dbgdev;
-};
-
-/* prototypes for debug manager functions */
-struct mutex *kfd_get_dbgmgr_mutex(void);
-void kfd_dbgmgr_destroy(struct kfd_dbgmgr *pmgr);
-bool kfd_dbgmgr_create(struct kfd_dbgmgr **ppmgr, struct kfd_dev *pdev);
-long kfd_dbgmgr_register(struct kfd_dbgmgr *pmgr, struct kfd_process *p);
-long kfd_dbgmgr_unregister(struct kfd_dbgmgr *pmgr, struct kfd_process *p);
-long kfd_dbgmgr_wave_control(struct kfd_dbgmgr *pmgr,
-				struct dbg_wave_control_info *wac_info);
-long kfd_dbgmgr_address_watch(struct kfd_dbgmgr *pmgr,
-			struct dbg_address_watch_info *adw_info);
-#endif /* KFD_DBGMGR_H_ */
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_debug.c b/drivers/gpu/drm/amd/amdkfd/kfd_debug.c
new file mode 100644
index 000000000000..ba99e0f258ae
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_debug.c
@@ -0,0 +1,1152 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#include "kfd_debug.h"
+#include "kfd_device_queue_manager.h"
+#include "kfd_topology.h"
+#include <linux/file.h>
+#include <uapi/linux/kfd_ioctl.h>
+#include <uapi/linux/kfd_sysfs.h>
+
+#define MAX_WATCH_ADDRESSES	4
+
+int kfd_dbg_ev_query_debug_event(struct kfd_process *process,
+		      unsigned int *queue_id,
+		      unsigned int *gpu_id,
+		      uint64_t exception_clear_mask,
+		      uint64_t *event_status)
+{
+	struct process_queue_manager *pqm;
+	struct process_queue_node *pqn;
+	int i;
+
+	if (!(process && process->debug_trap_enabled))
+		return -ENODATA;
+
+	mutex_lock(&process->event_mutex);
+	*event_status = 0;
+	*queue_id = 0;
+	*gpu_id = 0;
+
+	/* find and report queue events */
+	pqm = &process->pqm;
+	list_for_each_entry(pqn, &pqm->queues, process_queue_list) {
+		uint64_t tmp = process->exception_enable_mask;
+
+		if (!pqn->q)
+			continue;
+
+		tmp &= pqn->q->properties.exception_status;
+
+		if (!tmp)
+			continue;
+
+		*event_status = pqn->q->properties.exception_status;
+		*queue_id = pqn->q->properties.queue_id;
+		*gpu_id = pqn->q->device->id;
+		pqn->q->properties.exception_status &= ~exception_clear_mask;
+		goto out;
+	}
+
+	/* find and report device events */
+	for (i = 0; i < process->n_pdds; i++) {
+		struct kfd_process_device *pdd = process->pdds[i];
+		uint64_t tmp = process->exception_enable_mask
+						& pdd->exception_status;
+
+		if (!tmp)
+			continue;
+
+		*event_status = pdd->exception_status;
+		*gpu_id = pdd->dev->id;
+		pdd->exception_status &= ~exception_clear_mask;
+		goto out;
+	}
+
+	/* report process events */
+	if (process->exception_enable_mask & process->exception_status) {
+		*event_status = process->exception_status;
+		process->exception_status &= ~exception_clear_mask;
+	}
+
+out:
+	mutex_unlock(&process->event_mutex);
+	return *event_status ? 0 : -EAGAIN;
+}
+
+void debug_event_write_work_handler(struct work_struct *work)
+{
+	struct kfd_process *process;
+
+	static const char write_data = '.';
+	loff_t pos = 0;
+
+	process = container_of(work,
+			struct kfd_process,
+			debug_event_workarea);
+
+	if (process->debug_trap_enabled && process->dbg_ev_file)
+		kernel_write(process->dbg_ev_file, &write_data, 1, &pos);
+}
+
+/* update process/device/queue exception status, write to descriptor
+ * only if exception_status is enabled.
+ */
+bool kfd_dbg_ev_raise(uint64_t event_mask,
+			struct kfd_process *process, struct kfd_node *dev,
+			unsigned int source_id, bool use_worker,
+			void *exception_data, size_t exception_data_size)
+{
+	struct process_queue_manager *pqm;
+	struct process_queue_node *pqn;
+	int i;
+	static const char write_data = '.';
+	loff_t pos = 0;
+	bool is_subscribed = true;
+
+	if (!(process && process->debug_trap_enabled))
+		return false;
+
+	mutex_lock(&process->event_mutex);
+
+	if (event_mask & KFD_EC_MASK_DEVICE) {
+		for (i = 0; i < process->n_pdds; i++) {
+			struct kfd_process_device *pdd = process->pdds[i];
+
+			if (pdd->dev != dev)
+				continue;
+
+			pdd->exception_status |= event_mask & KFD_EC_MASK_DEVICE;
+
+			if (event_mask & KFD_EC_MASK(EC_DEVICE_MEMORY_VIOLATION)) {
+				if (!pdd->vm_fault_exc_data) {
+					pdd->vm_fault_exc_data = kmemdup(
+							exception_data,
+							exception_data_size,
+							GFP_KERNEL);
+					if (!pdd->vm_fault_exc_data)
+						pr_debug("Failed to allocate exception data memory");
+				} else {
+					pr_debug("Debugger exception data not saved\n");
+					print_hex_dump_bytes("exception data: ",
+							DUMP_PREFIX_OFFSET,
+							exception_data,
+							exception_data_size);
+				}
+			}
+			break;
+		}
+	} else if (event_mask & KFD_EC_MASK_PROCESS) {
+		process->exception_status |= event_mask & KFD_EC_MASK_PROCESS;
+	} else {
+		pqm = &process->pqm;
+		list_for_each_entry(pqn, &pqm->queues,
+				process_queue_list) {
+			int target_id;
+
+			if (!pqn->q)
+				continue;
+
+			target_id = event_mask & KFD_EC_MASK(EC_QUEUE_NEW) ?
+					pqn->q->properties.queue_id :
+							pqn->q->doorbell_id;
+
+			if (pqn->q->device != dev || target_id != source_id)
+				continue;
+
+			pqn->q->properties.exception_status |= event_mask;
+			break;
+		}
+	}
+
+	if (process->exception_enable_mask & event_mask) {
+		if (use_worker)
+			schedule_work(&process->debug_event_workarea);
+		else
+			kernel_write(process->dbg_ev_file,
+					&write_data,
+					1,
+					&pos);
+	} else {
+		is_subscribed = false;
+	}
+
+	mutex_unlock(&process->event_mutex);
+
+	return is_subscribed;
+}
+
+/* set pending event queue entry from ring entry  */
+bool kfd_set_dbg_ev_from_interrupt(struct kfd_node *dev,
+				   unsigned int pasid,
+				   uint32_t doorbell_id,
+				   uint64_t trap_mask,
+				   void *exception_data,
+				   size_t exception_data_size)
+{
+	struct kfd_process *p;
+	struct kfd_process_device *pdd = NULL;
+	bool signaled_to_debugger_or_runtime = false;
+
+	p = kfd_lookup_process_by_pasid(pasid, &pdd);
+
+	if (!pdd)
+		return false;
+
+	if (!kfd_dbg_ev_raise(trap_mask, p, dev, doorbell_id, true,
+			      exception_data, exception_data_size)) {
+		struct process_queue_manager *pqm;
+		struct process_queue_node *pqn;
+
+		if (!!(trap_mask & KFD_EC_MASK_QUEUE) &&
+		       p->runtime_info.runtime_state == DEBUG_RUNTIME_STATE_ENABLED) {
+			mutex_lock(&p->mutex);
+
+			pqm = &p->pqm;
+			list_for_each_entry(pqn, &pqm->queues,
+							process_queue_list) {
+
+				if (!(pqn->q && pqn->q->device == dev &&
+				      pqn->q->doorbell_id == doorbell_id))
+					continue;
+
+				kfd_send_exception_to_runtime(p, pqn->q->properties.queue_id,
+							      trap_mask);
+
+				signaled_to_debugger_or_runtime = true;
+
+				break;
+			}
+
+			mutex_unlock(&p->mutex);
+		} else if (trap_mask & KFD_EC_MASK(EC_DEVICE_MEMORY_VIOLATION)) {
+			kfd_evict_process_device(pdd);
+			kfd_signal_vm_fault_event(pdd, NULL, exception_data);
+
+			signaled_to_debugger_or_runtime = true;
+		}
+	} else {
+		signaled_to_debugger_or_runtime = true;
+	}
+
+	kfd_unref_process(p);
+
+	return signaled_to_debugger_or_runtime;
+}
+
+int kfd_dbg_send_exception_to_runtime(struct kfd_process *p,
+					unsigned int dev_id,
+					unsigned int queue_id,
+					uint64_t error_reason)
+{
+	if (error_reason & KFD_EC_MASK(EC_DEVICE_MEMORY_VIOLATION)) {
+		struct kfd_process_device *pdd = NULL;
+		struct kfd_hsa_memory_exception_data *data;
+		int i;
+
+		for (i = 0; i < p->n_pdds; i++) {
+			if (p->pdds[i]->dev->id == dev_id) {
+				pdd = p->pdds[i];
+				break;
+			}
+		}
+
+		if (!pdd)
+			return -ENODEV;
+
+		data = (struct kfd_hsa_memory_exception_data *)
+						pdd->vm_fault_exc_data;
+
+		kfd_evict_process_device(pdd);
+		kfd_signal_vm_fault_event(pdd, NULL, data);
+		error_reason &= ~KFD_EC_MASK(EC_DEVICE_MEMORY_VIOLATION);
+	}
+
+	if (error_reason & (KFD_EC_MASK(EC_PROCESS_RUNTIME))) {
+		/*
+		 * block should only happen after the debugger receives runtime
+		 * enable notice.
+		 */
+		up(&p->runtime_enable_sema);
+		error_reason &= ~KFD_EC_MASK(EC_PROCESS_RUNTIME);
+	}
+
+	if (error_reason)
+		return kfd_send_exception_to_runtime(p, queue_id, error_reason);
+
+	return 0;
+}
+
+static int kfd_dbg_set_queue_workaround(struct queue *q, bool enable)
+{
+	struct mqd_update_info minfo = {0};
+	int err;
+
+	if (!q)
+		return 0;
+
+	if (!kfd_dbg_has_cwsr_workaround(q->device))
+		return 0;
+
+	if (enable && q->properties.is_user_cu_masked)
+		return -EBUSY;
+
+	minfo.update_flag = enable ? UPDATE_FLAG_DBG_WA_ENABLE : UPDATE_FLAG_DBG_WA_DISABLE;
+
+	q->properties.is_dbg_wa = enable;
+	err = q->device->dqm->ops.update_queue(q->device->dqm, q, &minfo);
+	if (err)
+		q->properties.is_dbg_wa = false;
+
+	return err;
+}
+
+static int kfd_dbg_set_workaround(struct kfd_process *target, bool enable)
+{
+	struct process_queue_manager *pqm = &target->pqm;
+	struct process_queue_node *pqn;
+	int r = 0;
+
+	list_for_each_entry(pqn, &pqm->queues, process_queue_list) {
+		r = kfd_dbg_set_queue_workaround(pqn->q, enable);
+		if (enable && r)
+			goto unwind;
+	}
+
+	return 0;
+
+unwind:
+	list_for_each_entry(pqn, &pqm->queues, process_queue_list)
+		kfd_dbg_set_queue_workaround(pqn->q, false);
+
+	if (enable)
+		target->runtime_info.runtime_state = r == -EBUSY ?
+				DEBUG_RUNTIME_STATE_ENABLED_BUSY :
+				DEBUG_RUNTIME_STATE_ENABLED_ERROR;
+
+	return r;
+}
+
+int kfd_dbg_set_mes_debug_mode(struct kfd_process_device *pdd, bool sq_trap_en)
+{
+	uint32_t spi_dbg_cntl = pdd->spi_dbg_override | pdd->spi_dbg_launch_mode;
+	uint32_t flags = pdd->process->dbg_flags;
+	struct amdgpu_device *adev = pdd->dev->adev;
+	int r;
+
+	if (!kfd_dbg_is_per_vmid_supported(pdd->dev))
+		return 0;
+
+	if (!pdd->proc_ctx_cpu_ptr) {
+		r = amdgpu_amdkfd_alloc_gtt_mem(adev,
+			AMDGPU_MES_PROC_CTX_SIZE,
+			&pdd->proc_ctx_bo,
+			&pdd->proc_ctx_gpu_addr,
+			&pdd->proc_ctx_cpu_ptr,
+			false);
+		if (r) {
+			dev_err(adev->dev,
+			"failed to allocate process context bo\n");
+			return r;
+		}
+		memset(pdd->proc_ctx_cpu_ptr, 0, AMDGPU_MES_PROC_CTX_SIZE);
+	}
+
+	return amdgpu_mes_set_shader_debugger(pdd->dev->adev, pdd->proc_ctx_gpu_addr, spi_dbg_cntl,
+						pdd->watch_points, flags, sq_trap_en);
+}
+
+#define KFD_DEBUGGER_INVALID_WATCH_POINT_ID -1
+static int kfd_dbg_get_dev_watch_id(struct kfd_process_device *pdd, int *watch_id)
+{
+	int i;
+
+	*watch_id = KFD_DEBUGGER_INVALID_WATCH_POINT_ID;
+
+	spin_lock(&pdd->dev->watch_points_lock);
+
+	for (i = 0; i < MAX_WATCH_ADDRESSES; i++) {
+		/* device watchpoint in use so skip */
+		if ((pdd->dev->alloc_watch_ids >> i) & 0x1)
+			continue;
+
+		pdd->alloc_watch_ids |= 0x1 << i;
+		pdd->dev->alloc_watch_ids |= 0x1 << i;
+		*watch_id = i;
+		spin_unlock(&pdd->dev->watch_points_lock);
+		return 0;
+	}
+
+	spin_unlock(&pdd->dev->watch_points_lock);
+
+	return -ENOMEM;
+}
+
+static void kfd_dbg_clear_dev_watch_id(struct kfd_process_device *pdd, int watch_id)
+{
+	spin_lock(&pdd->dev->watch_points_lock);
+
+	/* process owns device watch point so safe to clear */
+	if ((pdd->alloc_watch_ids >> watch_id) & 0x1) {
+		pdd->alloc_watch_ids &= ~(0x1 << watch_id);
+		pdd->dev->alloc_watch_ids &= ~(0x1 << watch_id);
+	}
+
+	spin_unlock(&pdd->dev->watch_points_lock);
+}
+
+static bool kfd_dbg_owns_dev_watch_id(struct kfd_process_device *pdd, int watch_id)
+{
+	bool owns_watch_id = false;
+
+	spin_lock(&pdd->dev->watch_points_lock);
+	owns_watch_id = watch_id < MAX_WATCH_ADDRESSES &&
+			((pdd->alloc_watch_ids >> watch_id) & 0x1);
+
+	spin_unlock(&pdd->dev->watch_points_lock);
+
+	return owns_watch_id;
+}
+
+int kfd_dbg_trap_clear_dev_address_watch(struct kfd_process_device *pdd,
+					uint32_t watch_id)
+{
+	int r;
+
+	if (!kfd_dbg_owns_dev_watch_id(pdd, watch_id))
+		return -EINVAL;
+
+	if (!pdd->dev->kfd->shared_resources.enable_mes) {
+		r = debug_lock_and_unmap(pdd->dev->dqm);
+		if (r)
+			return r;
+	}
+
+	amdgpu_gfx_off_ctrl(pdd->dev->adev, false);
+	pdd->watch_points[watch_id] = pdd->dev->kfd2kgd->clear_address_watch(
+							pdd->dev->adev,
+							watch_id);
+	amdgpu_gfx_off_ctrl(pdd->dev->adev, true);
+
+	if (!pdd->dev->kfd->shared_resources.enable_mes)
+		r = debug_map_and_unlock(pdd->dev->dqm);
+	else
+		r = kfd_dbg_set_mes_debug_mode(pdd, true);
+
+	kfd_dbg_clear_dev_watch_id(pdd, watch_id);
+
+	return r;
+}
+
+int kfd_dbg_trap_set_dev_address_watch(struct kfd_process_device *pdd,
+					uint64_t watch_address,
+					uint32_t watch_address_mask,
+					uint32_t *watch_id,
+					uint32_t watch_mode)
+{
+	int xcc_id, r = kfd_dbg_get_dev_watch_id(pdd, watch_id);
+	uint32_t xcc_mask = pdd->dev->xcc_mask;
+
+	if (r)
+		return r;
+
+	if (!pdd->dev->kfd->shared_resources.enable_mes) {
+		r = debug_lock_and_unmap(pdd->dev->dqm);
+		if (r) {
+			kfd_dbg_clear_dev_watch_id(pdd, *watch_id);
+			return r;
+		}
+	}
+
+	amdgpu_gfx_off_ctrl(pdd->dev->adev, false);
+	for_each_inst(xcc_id, xcc_mask)
+		pdd->watch_points[*watch_id] = pdd->dev->kfd2kgd->set_address_watch(
+				pdd->dev->adev,
+				watch_address,
+				watch_address_mask,
+				*watch_id,
+				watch_mode,
+				pdd->dev->vm_info.last_vmid_kfd,
+				xcc_id);
+	amdgpu_gfx_off_ctrl(pdd->dev->adev, true);
+
+	if (!pdd->dev->kfd->shared_resources.enable_mes)
+		r = debug_map_and_unlock(pdd->dev->dqm);
+	else
+		r = kfd_dbg_set_mes_debug_mode(pdd, true);
+
+	/* HWS is broken so no point in HW rollback but release the watchpoint anyways */
+	if (r)
+		kfd_dbg_clear_dev_watch_id(pdd, *watch_id);
+
+	return 0;
+}
+
+static void kfd_dbg_clear_process_address_watch(struct kfd_process *target)
+{
+	int i, j;
+
+	for (i = 0; i < target->n_pdds; i++)
+		for (j = 0; j < MAX_WATCH_ADDRESSES; j++)
+			kfd_dbg_trap_clear_dev_address_watch(target->pdds[i], j);
+}
+
+int kfd_dbg_trap_set_flags(struct kfd_process *target, uint32_t *flags)
+{
+	uint32_t prev_flags = target->dbg_flags;
+	int i, r = 0, rewind_count = 0;
+
+	for (i = 0; i < target->n_pdds; i++) {
+		struct kfd_topology_device *topo_dev =
+				kfd_topology_device_by_id(target->pdds[i]->dev->id);
+		uint32_t caps = topo_dev->node_props.capability;
+
+		if (!(caps & HSA_CAP_TRAP_DEBUG_PRECISE_MEMORY_OPERATIONS_SUPPORTED) &&
+			(*flags & KFD_DBG_TRAP_FLAG_SINGLE_MEM_OP)) {
+			*flags = prev_flags;
+			return -EACCES;
+		}
+
+		if (!(caps & HSA_CAP_TRAP_DEBUG_PRECISE_ALU_OPERATIONS_SUPPORTED) &&
+		    (*flags & KFD_DBG_TRAP_FLAG_SINGLE_ALU_OP)) {
+			*flags = prev_flags;
+			return -EACCES;
+		}
+	}
+
+	target->dbg_flags = *flags;
+	*flags = prev_flags;
+	for (i = 0; i < target->n_pdds; i++) {
+		struct kfd_process_device *pdd = target->pdds[i];
+
+		if (!kfd_dbg_is_per_vmid_supported(pdd->dev))
+			continue;
+
+		if (!pdd->dev->kfd->shared_resources.enable_mes)
+			r = debug_refresh_runlist(pdd->dev->dqm);
+		else
+			r = kfd_dbg_set_mes_debug_mode(pdd, true);
+
+		if (r) {
+			target->dbg_flags = prev_flags;
+			break;
+		}
+
+		rewind_count++;
+	}
+
+	/* Rewind flags */
+	if (r) {
+		target->dbg_flags = prev_flags;
+
+		for (i = 0; i < rewind_count; i++) {
+			struct kfd_process_device *pdd = target->pdds[i];
+
+			if (!kfd_dbg_is_per_vmid_supported(pdd->dev))
+				continue;
+
+			if (!pdd->dev->kfd->shared_resources.enable_mes)
+				debug_refresh_runlist(pdd->dev->dqm);
+			else
+				kfd_dbg_set_mes_debug_mode(pdd, true);
+		}
+	}
+
+	return r;
+}
+
+/* kfd_dbg_trap_deactivate:
+ *	target: target process
+ *	unwind: If this is unwinding a failed kfd_dbg_trap_enable()
+ *	unwind_count:
+ *		If unwind == true, how far down the pdd list we need
+ *				to unwind
+ *		else: ignored
+ */
+void kfd_dbg_trap_deactivate(struct kfd_process *target, bool unwind, int unwind_count)
+{
+	int i;
+
+	if (!unwind) {
+		uint32_t flags = 0;
+		int resume_count = resume_queues(target, 0, NULL);
+
+		if (resume_count)
+			pr_debug("Resumed %d queues\n", resume_count);
+
+		cancel_work_sync(&target->debug_event_workarea);
+		kfd_dbg_clear_process_address_watch(target);
+		kfd_dbg_trap_set_wave_launch_mode(target, 0);
+
+		kfd_dbg_trap_set_flags(target, &flags);
+	}
+
+	for (i = 0; i < target->n_pdds; i++) {
+		struct kfd_process_device *pdd = target->pdds[i];
+
+		/* If this is an unwind, and we have unwound the required
+		 * enable calls on the pdd list, we need to stop now
+		 * otherwise we may mess up another debugger session.
+		 */
+		if (unwind && i == unwind_count)
+			break;
+
+		kfd_process_set_trap_debug_flag(&pdd->qpd, false);
+
+		/* GFX off is already disabled by debug activate if not RLC restore supported. */
+		if (kfd_dbg_is_rlc_restore_supported(pdd->dev))
+			amdgpu_gfx_off_ctrl(pdd->dev->adev, false);
+		pdd->spi_dbg_override =
+				pdd->dev->kfd2kgd->disable_debug_trap(
+				pdd->dev->adev,
+				target->runtime_info.ttmp_setup,
+				pdd->dev->vm_info.last_vmid_kfd);
+		amdgpu_gfx_off_ctrl(pdd->dev->adev, true);
+
+		if (!kfd_dbg_is_per_vmid_supported(pdd->dev) &&
+				release_debug_trap_vmid(pdd->dev->dqm, &pdd->qpd))
+			pr_err("Failed to release debug vmid on [%i]\n", pdd->dev->id);
+
+		if (!pdd->dev->kfd->shared_resources.enable_mes)
+			debug_refresh_runlist(pdd->dev->dqm);
+		else
+			kfd_dbg_set_mes_debug_mode(pdd, !kfd_dbg_has_cwsr_workaround(pdd->dev));
+	}
+
+	kfd_dbg_set_workaround(target, false);
+}
+
+static void kfd_dbg_clean_exception_status(struct kfd_process *target)
+{
+	struct process_queue_manager *pqm;
+	struct process_queue_node *pqn;
+	int i;
+
+	for (i = 0; i < target->n_pdds; i++) {
+		struct kfd_process_device *pdd = target->pdds[i];
+
+		kfd_process_drain_interrupts(pdd);
+
+		pdd->exception_status = 0;
+	}
+
+	pqm = &target->pqm;
+	list_for_each_entry(pqn, &pqm->queues, process_queue_list) {
+		if (!pqn->q)
+			continue;
+
+		pqn->q->properties.exception_status = 0;
+	}
+
+	target->exception_status = 0;
+}
+
+int kfd_dbg_trap_disable(struct kfd_process *target)
+{
+	if (!target->debug_trap_enabled)
+		return 0;
+
+	/*
+	 * Defer deactivation to runtime if runtime not enabled otherwise reset
+	 * attached running target runtime state to enable for re-attach.
+	 */
+	if (target->runtime_info.runtime_state == DEBUG_RUNTIME_STATE_ENABLED)
+		kfd_dbg_trap_deactivate(target, false, 0);
+	else if (target->runtime_info.runtime_state != DEBUG_RUNTIME_STATE_DISABLED)
+		target->runtime_info.runtime_state = DEBUG_RUNTIME_STATE_ENABLED;
+
+	cancel_work_sync(&target->debug_event_workarea);
+	fput(target->dbg_ev_file);
+	target->dbg_ev_file = NULL;
+
+	if (target->debugger_process) {
+		atomic_dec(&target->debugger_process->debugged_process_count);
+		target->debugger_process = NULL;
+	}
+
+	target->debug_trap_enabled = false;
+	kfd_dbg_clean_exception_status(target);
+	kfd_unref_process(target);
+
+	return 0;
+}
+
+int kfd_dbg_trap_activate(struct kfd_process *target)
+{
+	int i, r = 0;
+
+	r = kfd_dbg_set_workaround(target, true);
+	if (r)
+		return r;
+
+	for (i = 0; i < target->n_pdds; i++) {
+		struct kfd_process_device *pdd = target->pdds[i];
+
+		if (!kfd_dbg_is_per_vmid_supported(pdd->dev)) {
+			r = reserve_debug_trap_vmid(pdd->dev->dqm, &pdd->qpd);
+
+			if (r) {
+				target->runtime_info.runtime_state = (r == -EBUSY) ?
+							DEBUG_RUNTIME_STATE_ENABLED_BUSY :
+							DEBUG_RUNTIME_STATE_ENABLED_ERROR;
+
+				goto unwind_err;
+			}
+		}
+
+		/* Disable GFX OFF to prevent garbage read/writes to debug registers.
+		 * If RLC restore of debug registers is not supported and runtime enable
+		 * hasn't done so already on ttmp setup request, restore the trap config registers.
+		 *
+		 * If RLC restore of debug registers is not supported, keep gfx off disabled for
+		 * the debug session.
+		 */
+		amdgpu_gfx_off_ctrl(pdd->dev->adev, false);
+		if (!(kfd_dbg_is_rlc_restore_supported(pdd->dev) ||
+						target->runtime_info.ttmp_setup))
+			pdd->dev->kfd2kgd->enable_debug_trap(pdd->dev->adev, true,
+								pdd->dev->vm_info.last_vmid_kfd);
+
+		pdd->spi_dbg_override = pdd->dev->kfd2kgd->enable_debug_trap(
+					pdd->dev->adev,
+					false,
+					pdd->dev->vm_info.last_vmid_kfd);
+
+		if (kfd_dbg_is_rlc_restore_supported(pdd->dev))
+			amdgpu_gfx_off_ctrl(pdd->dev->adev, true);
+
+		/*
+		 * Setting the debug flag in the trap handler requires that the TMA has been
+		 * allocated, which occurs during CWSR initialization.
+		 * In the event that CWSR has not been initialized at this point, setting the
+		 * flag will be called again during CWSR initialization if the target process
+		 * is still debug enabled.
+		 */
+		kfd_process_set_trap_debug_flag(&pdd->qpd, true);
+
+		if (!pdd->dev->kfd->shared_resources.enable_mes)
+			r = debug_refresh_runlist(pdd->dev->dqm);
+		else
+			r = kfd_dbg_set_mes_debug_mode(pdd, true);
+
+		if (r) {
+			target->runtime_info.runtime_state =
+					DEBUG_RUNTIME_STATE_ENABLED_ERROR;
+			goto unwind_err;
+		}
+	}
+
+	return 0;
+
+unwind_err:
+	/* Enabling debug failed, we need to disable on
+	 * all GPUs so the enable is all or nothing.
+	 */
+	kfd_dbg_trap_deactivate(target, true, i);
+	return r;
+}
+
+int kfd_dbg_trap_enable(struct kfd_process *target, uint32_t fd,
+			void __user *runtime_info, uint32_t *runtime_size)
+{
+	struct file *f;
+	uint32_t copy_size;
+	int i, r = 0;
+
+	if (target->debug_trap_enabled)
+		return -EALREADY;
+
+	/* Enable pre-checks */
+	for (i = 0; i < target->n_pdds; i++) {
+		struct kfd_process_device *pdd = target->pdds[i];
+
+		if (!KFD_IS_SOC15(pdd->dev))
+			return -ENODEV;
+
+		if (pdd->qpd.num_gws && (!kfd_dbg_has_gws_support(pdd->dev) ||
+					 kfd_dbg_has_cwsr_workaround(pdd->dev)))
+			return -EBUSY;
+	}
+
+	copy_size = min((size_t)(*runtime_size), sizeof(target->runtime_info));
+
+	f = fget(fd);
+	if (!f) {
+		pr_err("Failed to get file for (%i)\n", fd);
+		return -EBADF;
+	}
+
+	target->dbg_ev_file = f;
+
+	/* defer activation to runtime if not runtime enabled */
+	if (target->runtime_info.runtime_state == DEBUG_RUNTIME_STATE_ENABLED)
+		kfd_dbg_trap_activate(target);
+
+	/* We already hold the process reference but hold another one for the
+	 * debug session.
+	 */
+	kref_get(&target->ref);
+	target->debug_trap_enabled = true;
+
+	if (target->debugger_process)
+		atomic_inc(&target->debugger_process->debugged_process_count);
+
+	if (copy_to_user(runtime_info, (void *)&target->runtime_info, copy_size)) {
+		kfd_dbg_trap_deactivate(target, false, 0);
+		r = -EFAULT;
+	}
+
+	*runtime_size = sizeof(target->runtime_info);
+
+	return r;
+}
+
+static int kfd_dbg_validate_trap_override_request(struct kfd_process *p,
+						uint32_t trap_override,
+						uint32_t trap_mask_request,
+						uint32_t *trap_mask_supported)
+{
+	int i = 0;
+
+	*trap_mask_supported = 0xffffffff;
+
+	for (i = 0; i < p->n_pdds; i++) {
+		struct kfd_process_device *pdd = p->pdds[i];
+		int err = pdd->dev->kfd2kgd->validate_trap_override_request(
+								pdd->dev->adev,
+								trap_override,
+								trap_mask_supported);
+
+		if (err)
+			return err;
+	}
+
+	if (trap_mask_request & ~*trap_mask_supported)
+		return -EACCES;
+
+	return 0;
+}
+
+int kfd_dbg_trap_set_wave_launch_override(struct kfd_process *target,
+					uint32_t trap_override,
+					uint32_t trap_mask_bits,
+					uint32_t trap_mask_request,
+					uint32_t *trap_mask_prev,
+					uint32_t *trap_mask_supported)
+{
+	int r = 0, i;
+
+	r = kfd_dbg_validate_trap_override_request(target,
+						trap_override,
+						trap_mask_request,
+						trap_mask_supported);
+
+	if (r)
+		return r;
+
+	for (i = 0; i < target->n_pdds; i++) {
+		struct kfd_process_device *pdd = target->pdds[i];
+
+		amdgpu_gfx_off_ctrl(pdd->dev->adev, false);
+		pdd->spi_dbg_override = pdd->dev->kfd2kgd->set_wave_launch_trap_override(
+				pdd->dev->adev,
+				pdd->dev->vm_info.last_vmid_kfd,
+				trap_override,
+				trap_mask_bits,
+				trap_mask_request,
+				trap_mask_prev,
+				pdd->spi_dbg_override);
+		amdgpu_gfx_off_ctrl(pdd->dev->adev, true);
+
+		if (!pdd->dev->kfd->shared_resources.enable_mes)
+			r = debug_refresh_runlist(pdd->dev->dqm);
+		else
+			r = kfd_dbg_set_mes_debug_mode(pdd, true);
+
+		if (r)
+			break;
+	}
+
+	return r;
+}
+
+int kfd_dbg_trap_set_wave_launch_mode(struct kfd_process *target,
+					uint8_t wave_launch_mode)
+{
+	int r = 0, i;
+
+	if (wave_launch_mode != KFD_DBG_TRAP_WAVE_LAUNCH_MODE_NORMAL &&
+			wave_launch_mode != KFD_DBG_TRAP_WAVE_LAUNCH_MODE_HALT &&
+			wave_launch_mode != KFD_DBG_TRAP_WAVE_LAUNCH_MODE_DEBUG)
+		return -EINVAL;
+
+	for (i = 0; i < target->n_pdds; i++) {
+		struct kfd_process_device *pdd = target->pdds[i];
+
+		amdgpu_gfx_off_ctrl(pdd->dev->adev, false);
+		pdd->spi_dbg_launch_mode = pdd->dev->kfd2kgd->set_wave_launch_mode(
+				pdd->dev->adev,
+				wave_launch_mode,
+				pdd->dev->vm_info.last_vmid_kfd);
+		amdgpu_gfx_off_ctrl(pdd->dev->adev, true);
+
+		if (!pdd->dev->kfd->shared_resources.enable_mes)
+			r = debug_refresh_runlist(pdd->dev->dqm);
+		else
+			r = kfd_dbg_set_mes_debug_mode(pdd, true);
+
+		if (r)
+			break;
+	}
+
+	return r;
+}
+
+int kfd_dbg_trap_query_exception_info(struct kfd_process *target,
+		uint32_t source_id,
+		uint32_t exception_code,
+		bool clear_exception,
+		void __user *info,
+		uint32_t *info_size)
+{
+	bool found = false;
+	int r = 0;
+	uint32_t copy_size, actual_info_size = 0;
+	uint64_t *exception_status_ptr = NULL;
+
+	if (!target)
+		return -EINVAL;
+
+	if (!info || !info_size)
+		return -EINVAL;
+
+	mutex_lock(&target->event_mutex);
+
+	if (KFD_DBG_EC_TYPE_IS_QUEUE(exception_code)) {
+		/* Per queue exceptions */
+		struct queue *queue = NULL;
+		int i;
+
+		for (i = 0; i < target->n_pdds; i++) {
+			struct kfd_process_device *pdd = target->pdds[i];
+			struct qcm_process_device *qpd = &pdd->qpd;
+
+			list_for_each_entry(queue, &qpd->queues_list, list) {
+				if (!found && queue->properties.queue_id == source_id) {
+					found = true;
+					break;
+				}
+			}
+			if (found)
+				break;
+		}
+
+		if (!found) {
+			r = -EINVAL;
+			goto out;
+		}
+
+		if (!(queue->properties.exception_status & KFD_EC_MASK(exception_code))) {
+			r = -ENODATA;
+			goto out;
+		}
+		exception_status_ptr = &queue->properties.exception_status;
+	} else if (KFD_DBG_EC_TYPE_IS_DEVICE(exception_code)) {
+		/* Per device exceptions */
+		struct kfd_process_device *pdd = NULL;
+		int i;
+
+		for (i = 0; i < target->n_pdds; i++) {
+			pdd = target->pdds[i];
+			if (pdd->dev->id == source_id) {
+				found = true;
+				break;
+			}
+		}
+
+		if (!found) {
+			r = -EINVAL;
+			goto out;
+		}
+
+		if (!(pdd->exception_status & KFD_EC_MASK(exception_code))) {
+			r = -ENODATA;
+			goto out;
+		}
+
+		if (exception_code == EC_DEVICE_MEMORY_VIOLATION) {
+			copy_size = min((size_t)(*info_size), pdd->vm_fault_exc_data_size);
+
+			if (copy_to_user(info, pdd->vm_fault_exc_data, copy_size)) {
+				r = -EFAULT;
+				goto out;
+			}
+			actual_info_size = pdd->vm_fault_exc_data_size;
+			if (clear_exception) {
+				kfree(pdd->vm_fault_exc_data);
+				pdd->vm_fault_exc_data = NULL;
+				pdd->vm_fault_exc_data_size = 0;
+			}
+		}
+		exception_status_ptr = &pdd->exception_status;
+	} else if (KFD_DBG_EC_TYPE_IS_PROCESS(exception_code)) {
+		/* Per process exceptions */
+		if (!(target->exception_status & KFD_EC_MASK(exception_code))) {
+			r = -ENODATA;
+			goto out;
+		}
+
+		if (exception_code == EC_PROCESS_RUNTIME) {
+			copy_size = min((size_t)(*info_size), sizeof(target->runtime_info));
+
+			if (copy_to_user(info, (void *)&target->runtime_info, copy_size)) {
+				r = -EFAULT;
+				goto out;
+			}
+
+			actual_info_size = sizeof(target->runtime_info);
+		}
+
+		exception_status_ptr = &target->exception_status;
+	} else {
+		pr_debug("Bad exception type [%i]\n", exception_code);
+		r = -EINVAL;
+		goto out;
+	}
+
+	*info_size = actual_info_size;
+	if (clear_exception)
+		*exception_status_ptr &= ~KFD_EC_MASK(exception_code);
+out:
+	mutex_unlock(&target->event_mutex);
+	return r;
+}
+
+int kfd_dbg_trap_device_snapshot(struct kfd_process *target,
+		uint64_t exception_clear_mask,
+		void __user *user_info,
+		uint32_t *number_of_device_infos,
+		uint32_t *entry_size)
+{
+	struct kfd_dbg_device_info_entry device_info;
+	uint32_t tmp_entry_size, tmp_num_devices;
+	int i, r = 0;
+
+	if (!(target && user_info && number_of_device_infos && entry_size))
+		return -EINVAL;
+
+	tmp_entry_size = *entry_size;
+
+	tmp_num_devices = min_t(size_t, *number_of_device_infos, target->n_pdds);
+	*number_of_device_infos = target->n_pdds;
+	*entry_size = min_t(size_t, *entry_size, sizeof(device_info));
+
+	if (!tmp_num_devices)
+		return 0;
+
+	memset(&device_info, 0, sizeof(device_info));
+
+	mutex_lock(&target->event_mutex);
+
+	/* Run over all pdd of the process */
+	for (i = 0; i < tmp_num_devices; i++) {
+		struct kfd_process_device *pdd = target->pdds[i];
+		struct kfd_topology_device *topo_dev = kfd_topology_device_by_id(pdd->dev->id);
+
+		device_info.gpu_id = pdd->dev->id;
+		device_info.exception_status = pdd->exception_status;
+		device_info.lds_base = pdd->lds_base;
+		device_info.lds_limit = pdd->lds_limit;
+		device_info.scratch_base = pdd->scratch_base;
+		device_info.scratch_limit = pdd->scratch_limit;
+		device_info.gpuvm_base = pdd->gpuvm_base;
+		device_info.gpuvm_limit = pdd->gpuvm_limit;
+		device_info.location_id = topo_dev->node_props.location_id;
+		device_info.vendor_id = topo_dev->node_props.vendor_id;
+		device_info.device_id = topo_dev->node_props.device_id;
+		device_info.revision_id = pdd->dev->adev->pdev->revision;
+		device_info.subsystem_vendor_id = pdd->dev->adev->pdev->subsystem_vendor;
+		device_info.subsystem_device_id = pdd->dev->adev->pdev->subsystem_device;
+		device_info.fw_version = pdd->dev->kfd->mec_fw_version;
+		device_info.gfx_target_version =
+			topo_dev->node_props.gfx_target_version;
+		device_info.simd_count = topo_dev->node_props.simd_count;
+		device_info.max_waves_per_simd =
+			topo_dev->node_props.max_waves_per_simd;
+		device_info.array_count = topo_dev->node_props.array_count;
+		device_info.simd_arrays_per_engine =
+			topo_dev->node_props.simd_arrays_per_engine;
+		device_info.num_xcc = NUM_XCC(pdd->dev->xcc_mask);
+		device_info.capability = topo_dev->node_props.capability;
+		device_info.debug_prop = topo_dev->node_props.debug_prop;
+
+		if (exception_clear_mask)
+			pdd->exception_status &= ~exception_clear_mask;
+
+		if (copy_to_user(user_info, &device_info, *entry_size)) {
+			r = -EFAULT;
+			break;
+		}
+
+		user_info += tmp_entry_size;
+	}
+
+	mutex_unlock(&target->event_mutex);
+
+	return r;
+}
+
+void kfd_dbg_set_enabled_debug_exception_mask(struct kfd_process *target,
+					uint64_t exception_set_mask)
+{
+	uint64_t found_mask = 0;
+	struct process_queue_manager *pqm;
+	struct process_queue_node *pqn;
+	static const char write_data = '.';
+	loff_t pos = 0;
+	int i;
+
+	mutex_lock(&target->event_mutex);
+
+	found_mask |= target->exception_status;
+
+	pqm = &target->pqm;
+	list_for_each_entry(pqn, &pqm->queues, process_queue_list) {
+		if (!pqn->q)
+			continue;
+
+		found_mask |= pqn->q->properties.exception_status;
+	}
+
+	for (i = 0; i < target->n_pdds; i++) {
+		struct kfd_process_device *pdd = target->pdds[i];
+
+		found_mask |= pdd->exception_status;
+	}
+
+	if (exception_set_mask & found_mask)
+		kernel_write(target->dbg_ev_file, &write_data, 1, &pos);
+
+	target->exception_enable_mask = exception_set_mask;
+
+	mutex_unlock(&target->event_mutex);
+}
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_debug.h b/drivers/gpu/drm/amd/amdkfd/kfd_debug.h
new file mode 100644
index 000000000000..27aa1a5b120f
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_debug.h
@@ -0,0 +1,142 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#ifndef KFD_DEBUG_EVENTS_H_INCLUDED
+#define KFD_DEBUG_EVENTS_H_INCLUDED
+
+#include "kfd_priv.h"
+
+void kfd_dbg_trap_deactivate(struct kfd_process *target, bool unwind, int unwind_count);
+int kfd_dbg_trap_activate(struct kfd_process *target);
+int kfd_dbg_ev_query_debug_event(struct kfd_process *process,
+			unsigned int *queue_id,
+			unsigned int *gpu_id,
+			uint64_t exception_clear_mask,
+			uint64_t *event_status);
+bool kfd_set_dbg_ev_from_interrupt(struct kfd_node *dev,
+				   unsigned int pasid,
+				   uint32_t doorbell_id,
+				   uint64_t trap_mask,
+				   void *exception_data,
+				   size_t exception_data_size);
+bool kfd_dbg_ev_raise(uint64_t event_mask,
+			struct kfd_process *process, struct kfd_node *dev,
+			unsigned int source_id, bool use_worker,
+			void *exception_data,
+			size_t exception_data_size);
+int kfd_dbg_trap_disable(struct kfd_process *target);
+int kfd_dbg_trap_enable(struct kfd_process *target, uint32_t fd,
+			void __user *runtime_info,
+			uint32_t *runtime_info_size);
+int kfd_dbg_trap_set_wave_launch_override(struct kfd_process *target,
+					uint32_t trap_override,
+					uint32_t trap_mask_bits,
+					uint32_t trap_mask_request,
+					uint32_t *trap_mask_prev,
+					uint32_t *trap_mask_supported);
+int kfd_dbg_trap_set_wave_launch_mode(struct kfd_process *target,
+					uint8_t wave_launch_mode);
+int kfd_dbg_trap_clear_dev_address_watch(struct kfd_process_device *pdd,
+					uint32_t watch_id);
+int kfd_dbg_trap_set_dev_address_watch(struct kfd_process_device *pdd,
+					uint64_t watch_address,
+					uint32_t watch_address_mask,
+					uint32_t *watch_id,
+					uint32_t watch_mode);
+int kfd_dbg_trap_set_flags(struct kfd_process *target, uint32_t *flags);
+int kfd_dbg_trap_query_exception_info(struct kfd_process *target,
+		uint32_t source_id,
+		uint32_t exception_code,
+		bool clear_exception,
+		void __user *info,
+		uint32_t *info_size);
+int kfd_dbg_send_exception_to_runtime(struct kfd_process *p,
+					unsigned int dev_id,
+					unsigned int queue_id,
+					uint64_t error_reason);
+
+static inline bool kfd_dbg_is_per_vmid_supported(struct kfd_node *dev)
+{
+	return (KFD_GC_VERSION(dev) == IP_VERSION(9, 4, 2) ||
+		KFD_GC_VERSION(dev) == IP_VERSION(9, 4, 3) ||
+		KFD_GC_VERSION(dev) == IP_VERSION(9, 4, 4) ||
+		KFD_GC_VERSION(dev) == IP_VERSION(9, 5, 0) ||
+		KFD_GC_VERSION(dev) >= IP_VERSION(11, 0, 0));
+}
+
+void debug_event_write_work_handler(struct work_struct *work);
+int kfd_dbg_trap_device_snapshot(struct kfd_process *target,
+		uint64_t exception_clear_mask,
+		void __user *user_info,
+		uint32_t *number_of_device_infos,
+		uint32_t *entry_size);
+
+void kfd_dbg_set_enabled_debug_exception_mask(struct kfd_process *target,
+					uint64_t exception_set_mask);
+/*
+ * If GFX off is enabled, chips that do not support RLC restore for the debug
+ * registers will disable GFX off temporarily for the entire debug session.
+ * See disable_on_trap_action_entry and enable_on_trap_action_exit for details.
+ */
+static inline bool kfd_dbg_is_rlc_restore_supported(struct kfd_node *dev)
+{
+	return !(KFD_GC_VERSION(dev) == IP_VERSION(10, 1, 10) ||
+		 KFD_GC_VERSION(dev) == IP_VERSION(10, 1, 1));
+}
+
+static inline bool kfd_dbg_has_cwsr_workaround(struct kfd_node *dev)
+{
+	return KFD_GC_VERSION(dev) >= IP_VERSION(11, 0, 0) &&
+	       KFD_GC_VERSION(dev) <= IP_VERSION(11, 0, 3);
+}
+
+static inline bool kfd_dbg_has_gws_support(struct kfd_node *dev)
+{
+	if ((KFD_GC_VERSION(dev) == IP_VERSION(9, 0, 1)
+			&& dev->kfd->mec2_fw_version < 0x81b6) ||
+		(KFD_GC_VERSION(dev) >= IP_VERSION(9, 1, 0)
+			&& KFD_GC_VERSION(dev) <= IP_VERSION(9, 2, 2)
+			&& dev->kfd->mec2_fw_version < 0x1b6) ||
+		(KFD_GC_VERSION(dev) == IP_VERSION(9, 4, 0)
+			&& dev->kfd->mec2_fw_version < 0x1b6) ||
+		(KFD_GC_VERSION(dev) == IP_VERSION(9, 4, 1)
+			&& dev->kfd->mec2_fw_version < 0x30) ||
+		(KFD_GC_VERSION(dev) >= IP_VERSION(11, 0, 0) &&
+			KFD_GC_VERSION(dev) < IP_VERSION(12, 0, 0)))
+		return false;
+
+	/* Assume debugging and cooperative launch supported otherwise. */
+	return true;
+}
+
+int kfd_dbg_set_mes_debug_mode(struct kfd_process_device *pdd, bool sq_trap_en);
+
+static inline bool kfd_dbg_has_ttmps_always_setup(struct kfd_node *dev)
+{
+	return (KFD_GC_VERSION(dev) < IP_VERSION(11, 0, 0) &&
+			KFD_GC_VERSION(dev) != IP_VERSION(9, 4, 2)) ||
+	       (KFD_GC_VERSION(dev) >= IP_VERSION(11, 0, 0) &&
+			KFD_GC_VERSION(dev) < IP_VERSION(12, 0, 0) &&
+			(dev->adev->mes.sched_version & AMDGPU_MES_VERSION_MASK) >= 70) ||
+	       (KFD_GC_VERSION(dev) >= IP_VERSION(12, 0, 0));
+}
+#endif
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_debugfs.c b/drivers/gpu/drm/amd/amdkfd/kfd_debugfs.c
index 673d5e34f213..9bde2c64540f 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_debugfs.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_debugfs.c
@@ -1,5 +1,6 @@
+// SPDX-License-Identifier: GPL-2.0 OR MIT
 /*
- * Copyright 2016-2017 Advanced Micro Devices, Inc.
+ * Copyright 2016-2022 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -26,6 +27,16 @@
 #include "kfd_priv.h"
 
 static struct dentry *debugfs_root;
+static struct dentry *debugfs_proc;
+static struct list_head procs;
+
+struct debugfs_proc_entry {
+	struct list_head list;
+	struct dentry *proc_dentry;
+	pid_t pid;
+};
+
+#define MAX_DEBUGFS_FILENAME_LEN 32
 
 static int kfd_debugfs_open(struct inode *inode, struct file *file)
 {
@@ -35,14 +46,14 @@ static int kfd_debugfs_open(struct inode *inode, struct file *file)
 }
 static int kfd_debugfs_hang_hws_read(struct seq_file *m, void *data)
 {
-	seq_printf(m, "echo gpu_id > hang_hws\n");
+	seq_puts(m, "echo gpu_id > hang_hws\n");
 	return 0;
 }
 
 static ssize_t kfd_debugfs_hang_hws_write(struct file *file,
 	const char __user *user_buf, size_t size, loff_t *ppos)
 {
-	struct kfd_dev *dev;
+	struct kfd_node *dev;
 	char tmp[16];
 	uint32_t gpu_id;
 	int ret = -EINVAL;
@@ -91,6 +102,8 @@ static const struct file_operations kfd_debugfs_hang_hws_fops = {
 void kfd_debugfs_init(void)
 {
 	debugfs_root = debugfs_create_dir("kfd", NULL);
+	debugfs_proc = debugfs_create_dir("proc", debugfs_root);
+	INIT_LIST_HEAD(&procs);
 
 	debugfs_create_file("mqds", S_IFREG | 0444, debugfs_root,
 			    kfd_debugfs_mqds_by_process, &kfd_debugfs_fops);
@@ -100,9 +113,75 @@ void kfd_debugfs_init(void)
 			    kfd_debugfs_rls_by_device, &kfd_debugfs_fops);
 	debugfs_create_file("hang_hws", S_IFREG | 0200, debugfs_root,
 			    kfd_debugfs_hang_hws_read, &kfd_debugfs_hang_hws_fops);
+	debugfs_create_file("mem_limit", S_IFREG | 0200, debugfs_root,
+			    kfd_debugfs_kfd_mem_limits, &kfd_debugfs_fops);
 }
 
 void kfd_debugfs_fini(void)
 {
+	debugfs_remove_recursive(debugfs_proc);
 	debugfs_remove_recursive(debugfs_root);
 }
+
+static ssize_t kfd_debugfs_pasid_read(struct file *file, char __user *buf,
+				      size_t count, loff_t *ppos)
+{
+	struct kfd_process_device *pdd = file_inode(file)->i_private;
+	char tmp[32];
+	int len;
+
+	len = snprintf(tmp, sizeof(tmp), "%u\n", pdd->pasid);
+
+	return simple_read_from_buffer(buf, count, ppos, tmp, len);
+}
+
+static const struct file_operations kfd_debugfs_pasid_fops = {
+	.owner = THIS_MODULE,
+	.read = kfd_debugfs_pasid_read,
+};
+
+void kfd_debugfs_add_process(struct kfd_process *p)
+{
+	int i;
+	char name[MAX_DEBUGFS_FILENAME_LEN];
+	struct debugfs_proc_entry *entry;
+
+	entry = kzalloc(sizeof(*entry), GFP_KERNEL);
+	if (!entry)
+		return;
+
+	list_add(&entry->list, &procs);
+	entry->pid = p->lead_thread->pid;
+	snprintf(name, MAX_DEBUGFS_FILENAME_LEN, "%d",
+		 (int)entry->pid);
+	entry->proc_dentry = debugfs_create_dir(name, debugfs_proc);
+
+	/* Create debugfs files for each GPU:
+	 * - proc/<pid>/pasid_<gpuid>
+	 */
+	for (i = 0; i < p->n_pdds; i++) {
+		struct kfd_process_device *pdd = p->pdds[i];
+
+		snprintf(name, MAX_DEBUGFS_FILENAME_LEN, "pasid_%u",
+			 pdd->dev->id);
+		debugfs_create_file((const char *)name, S_IFREG | 0444,
+				    entry->proc_dentry, pdd,
+				    &kfd_debugfs_pasid_fops);
+	}
+}
+
+void kfd_debugfs_remove_process(struct kfd_process *p)
+{
+	struct debugfs_proc_entry *entry, *next;
+
+	mutex_lock(&kfd_processes_mutex);
+	list_for_each_entry_safe(entry, next, &procs, list) {
+		if (entry->pid != p->lead_thread->pid)
+			continue;
+
+		debugfs_remove_recursive(entry->proc_dentry);
+		list_del(&entry->list);
+		kfree(entry);
+	}
+	mutex_unlock(&kfd_processes_mutex);
+}
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device.c b/drivers/gpu/drm/amd/amdkfd/kfd_device.c
index 3b119db16003..e9cfb80bd436 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_device.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device.c
@@ -1,5 +1,6 @@
+// SPDX-License-Identifier: GPL-2.0 OR MIT
 /*
- * Copyright 2014 Advanced Micro Devices, Inc.
+ * Copyright 2014-2022 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -28,11 +29,12 @@
 #include "kfd_pm4_headers_vi.h"
 #include "kfd_pm4_headers_aldebaran.h"
 #include "cwsr_trap_handler.h"
-#include "kfd_iommu.h"
 #include "amdgpu_amdkfd.h"
 #include "kfd_smi_events.h"
+#include "kfd_svm.h"
 #include "kfd_migrate.h"
 #include "amdgpu.h"
+#include "amdgpu_xcp.h"
 
 #define MQD_SIZE_ALIGNED 768
 
@@ -41,7 +43,7 @@
  * once locked, kfd driver will stop any further GPU execution.
  * create process (open) will return -EAGAIN.
  */
-static atomic_t kfd_locked = ATOMIC_INIT(0);
+static int kfd_locked;
 
 #ifdef CONFIG_DRM_AMDGPU_CIK
 extern const struct kfd2kgd_calls gfx_v7_kfd2kgd;
@@ -50,773 +52,433 @@ extern const struct kfd2kgd_calls gfx_v8_kfd2kgd;
 extern const struct kfd2kgd_calls gfx_v9_kfd2kgd;
 extern const struct kfd2kgd_calls arcturus_kfd2kgd;
 extern const struct kfd2kgd_calls aldebaran_kfd2kgd;
+extern const struct kfd2kgd_calls gc_9_4_3_kfd2kgd;
 extern const struct kfd2kgd_calls gfx_v10_kfd2kgd;
 extern const struct kfd2kgd_calls gfx_v10_3_kfd2kgd;
-
-#ifdef KFD_SUPPORT_IOMMU_V2
-static const struct kfd_device_info kaveri_device_info = {
-	.asic_family = CHIP_KAVERI,
-	.asic_name = "kaveri",
-	.gfx_target_version = 70000,
-	.max_pasid_bits = 16,
-	/* max num of queues for KV.TODO should be a dynamic value */
-	.max_no_of_hqd	= 24,
-	.doorbell_size  = 4,
-	.ih_ring_entry_size = 4 * sizeof(uint32_t),
-	.event_interrupt_class = &event_interrupt_class_cik,
-	.num_of_watch_points = 4,
-	.mqd_size_aligned = MQD_SIZE_ALIGNED,
-	.supports_cwsr = false,
-	.needs_iommu_device = true,
-	.needs_pci_atomics = false,
-	.num_sdma_engines = 2,
-	.num_xgmi_sdma_engines = 0,
-	.num_sdma_queues_per_engine = 2,
-};
-
-static const struct kfd_device_info carrizo_device_info = {
-	.asic_family = CHIP_CARRIZO,
-	.asic_name = "carrizo",
-	.gfx_target_version = 80001,
-	.max_pasid_bits = 16,
-	/* max num of queues for CZ.TODO should be a dynamic value */
-	.max_no_of_hqd	= 24,
-	.doorbell_size  = 4,
-	.ih_ring_entry_size = 4 * sizeof(uint32_t),
-	.event_interrupt_class = &event_interrupt_class_cik,
-	.num_of_watch_points = 4,
-	.mqd_size_aligned = MQD_SIZE_ALIGNED,
-	.supports_cwsr = true,
-	.needs_iommu_device = true,
-	.needs_pci_atomics = false,
-	.num_sdma_engines = 2,
-	.num_xgmi_sdma_engines = 0,
-	.num_sdma_queues_per_engine = 2,
-};
-
-static const struct kfd_device_info raven_device_info = {
-	.asic_family = CHIP_RAVEN,
-	.asic_name = "raven",
-	.gfx_target_version = 90002,
-	.max_pasid_bits = 16,
-	.max_no_of_hqd  = 24,
-	.doorbell_size  = 8,
-	.ih_ring_entry_size = 8 * sizeof(uint32_t),
-	.event_interrupt_class = &event_interrupt_class_v9,
-	.num_of_watch_points = 4,
-	.mqd_size_aligned = MQD_SIZE_ALIGNED,
-	.supports_cwsr = true,
-	.needs_iommu_device = true,
-	.needs_pci_atomics = true,
-	.num_sdma_engines = 1,
-	.num_xgmi_sdma_engines = 0,
-	.num_sdma_queues_per_engine = 2,
-};
-#endif
-
-#ifdef CONFIG_DRM_AMDGPU_CIK
-static const struct kfd_device_info hawaii_device_info = {
-	.asic_family = CHIP_HAWAII,
-	.asic_name = "hawaii",
-	.gfx_target_version = 70001,
-	.max_pasid_bits = 16,
-	/* max num of queues for KV.TODO should be a dynamic value */
-	.max_no_of_hqd	= 24,
-	.doorbell_size  = 4,
-	.ih_ring_entry_size = 4 * sizeof(uint32_t),
-	.event_interrupt_class = &event_interrupt_class_cik,
-	.num_of_watch_points = 4,
-	.mqd_size_aligned = MQD_SIZE_ALIGNED,
-	.supports_cwsr = false,
-	.needs_iommu_device = false,
-	.needs_pci_atomics = false,
-	.num_sdma_engines = 2,
-	.num_xgmi_sdma_engines = 0,
-	.num_sdma_queues_per_engine = 2,
-};
-#endif
-
-static const struct kfd_device_info tonga_device_info = {
-	.asic_family = CHIP_TONGA,
-	.asic_name = "tonga",
-	.gfx_target_version = 80002,
-	.max_pasid_bits = 16,
-	.max_no_of_hqd  = 24,
-	.doorbell_size  = 4,
-	.ih_ring_entry_size = 4 * sizeof(uint32_t),
-	.event_interrupt_class = &event_interrupt_class_cik,
-	.num_of_watch_points = 4,
-	.mqd_size_aligned = MQD_SIZE_ALIGNED,
-	.supports_cwsr = false,
-	.needs_iommu_device = false,
-	.needs_pci_atomics = true,
-	.num_sdma_engines = 2,
-	.num_xgmi_sdma_engines = 0,
-	.num_sdma_queues_per_engine = 2,
-};
-
-static const struct kfd_device_info fiji_device_info = {
-	.asic_family = CHIP_FIJI,
-	.asic_name = "fiji",
-	.gfx_target_version = 80003,
-	.max_pasid_bits = 16,
-	.max_no_of_hqd  = 24,
-	.doorbell_size  = 4,
-	.ih_ring_entry_size = 4 * sizeof(uint32_t),
-	.event_interrupt_class = &event_interrupt_class_cik,
-	.num_of_watch_points = 4,
-	.mqd_size_aligned = MQD_SIZE_ALIGNED,
-	.supports_cwsr = true,
-	.needs_iommu_device = false,
-	.needs_pci_atomics = true,
-	.num_sdma_engines = 2,
-	.num_xgmi_sdma_engines = 0,
-	.num_sdma_queues_per_engine = 2,
-};
-
-static const struct kfd_device_info fiji_vf_device_info = {
-	.asic_family = CHIP_FIJI,
-	.asic_name = "fiji",
-	.gfx_target_version = 80003,
-	.max_pasid_bits = 16,
-	.max_no_of_hqd  = 24,
-	.doorbell_size  = 4,
-	.ih_ring_entry_size = 4 * sizeof(uint32_t),
-	.event_interrupt_class = &event_interrupt_class_cik,
-	.num_of_watch_points = 4,
-	.mqd_size_aligned = MQD_SIZE_ALIGNED,
-	.supports_cwsr = true,
-	.needs_iommu_device = false,
-	.needs_pci_atomics = false,
-	.num_sdma_engines = 2,
-	.num_xgmi_sdma_engines = 0,
-	.num_sdma_queues_per_engine = 2,
-};
-
-
-static const struct kfd_device_info polaris10_device_info = {
-	.asic_family = CHIP_POLARIS10,
-	.asic_name = "polaris10",
-	.gfx_target_version = 80003,
-	.max_pasid_bits = 16,
-	.max_no_of_hqd  = 24,
-	.doorbell_size  = 4,
-	.ih_ring_entry_size = 4 * sizeof(uint32_t),
-	.event_interrupt_class = &event_interrupt_class_cik,
-	.num_of_watch_points = 4,
-	.mqd_size_aligned = MQD_SIZE_ALIGNED,
-	.supports_cwsr = true,
-	.needs_iommu_device = false,
-	.needs_pci_atomics = true,
-	.num_sdma_engines = 2,
-	.num_xgmi_sdma_engines = 0,
-	.num_sdma_queues_per_engine = 2,
-};
-
-static const struct kfd_device_info polaris10_vf_device_info = {
-	.asic_family = CHIP_POLARIS10,
-	.asic_name = "polaris10",
-	.gfx_target_version = 80003,
-	.max_pasid_bits = 16,
-	.max_no_of_hqd  = 24,
-	.doorbell_size  = 4,
-	.ih_ring_entry_size = 4 * sizeof(uint32_t),
-	.event_interrupt_class = &event_interrupt_class_cik,
-	.num_of_watch_points = 4,
-	.mqd_size_aligned = MQD_SIZE_ALIGNED,
-	.supports_cwsr = true,
-	.needs_iommu_device = false,
-	.needs_pci_atomics = false,
-	.num_sdma_engines = 2,
-	.num_xgmi_sdma_engines = 0,
-	.num_sdma_queues_per_engine = 2,
-};
-
-static const struct kfd_device_info polaris11_device_info = {
-	.asic_family = CHIP_POLARIS11,
-	.asic_name = "polaris11",
-	.gfx_target_version = 80003,
-	.max_pasid_bits = 16,
-	.max_no_of_hqd  = 24,
-	.doorbell_size  = 4,
-	.ih_ring_entry_size = 4 * sizeof(uint32_t),
-	.event_interrupt_class = &event_interrupt_class_cik,
-	.num_of_watch_points = 4,
-	.mqd_size_aligned = MQD_SIZE_ALIGNED,
-	.supports_cwsr = true,
-	.needs_iommu_device = false,
-	.needs_pci_atomics = true,
-	.num_sdma_engines = 2,
-	.num_xgmi_sdma_engines = 0,
-	.num_sdma_queues_per_engine = 2,
-};
-
-static const struct kfd_device_info polaris12_device_info = {
-	.asic_family = CHIP_POLARIS12,
-	.asic_name = "polaris12",
-	.gfx_target_version = 80003,
-	.max_pasid_bits = 16,
-	.max_no_of_hqd  = 24,
-	.doorbell_size  = 4,
-	.ih_ring_entry_size = 4 * sizeof(uint32_t),
-	.event_interrupt_class = &event_interrupt_class_cik,
-	.num_of_watch_points = 4,
-	.mqd_size_aligned = MQD_SIZE_ALIGNED,
-	.supports_cwsr = true,
-	.needs_iommu_device = false,
-	.needs_pci_atomics = true,
-	.num_sdma_engines = 2,
-	.num_xgmi_sdma_engines = 0,
-	.num_sdma_queues_per_engine = 2,
-};
-
-static const struct kfd_device_info vegam_device_info = {
-	.asic_family = CHIP_VEGAM,
-	.asic_name = "vegam",
-	.gfx_target_version = 80003,
-	.max_pasid_bits = 16,
-	.max_no_of_hqd  = 24,
-	.doorbell_size  = 4,
-	.ih_ring_entry_size = 4 * sizeof(uint32_t),
-	.event_interrupt_class = &event_interrupt_class_cik,
-	.num_of_watch_points = 4,
-	.mqd_size_aligned = MQD_SIZE_ALIGNED,
-	.supports_cwsr = true,
-	.needs_iommu_device = false,
-	.needs_pci_atomics = true,
-	.num_sdma_engines = 2,
-	.num_xgmi_sdma_engines = 0,
-	.num_sdma_queues_per_engine = 2,
-};
-
-static const struct kfd_device_info vega10_device_info = {
-	.asic_family = CHIP_VEGA10,
-	.asic_name = "vega10",
-	.gfx_target_version = 90000,
-	.max_pasid_bits = 16,
-	.max_no_of_hqd  = 24,
-	.doorbell_size  = 8,
-	.ih_ring_entry_size = 8 * sizeof(uint32_t),
-	.event_interrupt_class = &event_interrupt_class_v9,
-	.num_of_watch_points = 4,
-	.mqd_size_aligned = MQD_SIZE_ALIGNED,
-	.supports_cwsr = true,
-	.needs_iommu_device = false,
-	.needs_pci_atomics = false,
-	.num_sdma_engines = 2,
-	.num_xgmi_sdma_engines = 0,
-	.num_sdma_queues_per_engine = 2,
-};
-
-static const struct kfd_device_info vega10_vf_device_info = {
-	.asic_family = CHIP_VEGA10,
-	.asic_name = "vega10",
-	.gfx_target_version = 90000,
-	.max_pasid_bits = 16,
-	.max_no_of_hqd  = 24,
-	.doorbell_size  = 8,
-	.ih_ring_entry_size = 8 * sizeof(uint32_t),
-	.event_interrupt_class = &event_interrupt_class_v9,
-	.num_of_watch_points = 4,
-	.mqd_size_aligned = MQD_SIZE_ALIGNED,
-	.supports_cwsr = true,
-	.needs_iommu_device = false,
-	.needs_pci_atomics = false,
-	.num_sdma_engines = 2,
-	.num_xgmi_sdma_engines = 0,
-	.num_sdma_queues_per_engine = 2,
-};
-
-static const struct kfd_device_info vega12_device_info = {
-	.asic_family = CHIP_VEGA12,
-	.asic_name = "vega12",
-	.gfx_target_version = 90004,
-	.max_pasid_bits = 16,
-	.max_no_of_hqd  = 24,
-	.doorbell_size  = 8,
-	.ih_ring_entry_size = 8 * sizeof(uint32_t),
-	.event_interrupt_class = &event_interrupt_class_v9,
-	.num_of_watch_points = 4,
-	.mqd_size_aligned = MQD_SIZE_ALIGNED,
-	.supports_cwsr = true,
-	.needs_iommu_device = false,
-	.needs_pci_atomics = false,
-	.num_sdma_engines = 2,
-	.num_xgmi_sdma_engines = 0,
-	.num_sdma_queues_per_engine = 2,
-};
-
-static const struct kfd_device_info vega20_device_info = {
-	.asic_family = CHIP_VEGA20,
-	.asic_name = "vega20",
-	.gfx_target_version = 90006,
-	.max_pasid_bits = 16,
-	.max_no_of_hqd	= 24,
-	.doorbell_size	= 8,
-	.ih_ring_entry_size = 8 * sizeof(uint32_t),
-	.event_interrupt_class = &event_interrupt_class_v9,
-	.num_of_watch_points = 4,
-	.mqd_size_aligned = MQD_SIZE_ALIGNED,
-	.supports_cwsr = true,
-	.needs_iommu_device = false,
-	.needs_pci_atomics = false,
-	.num_sdma_engines = 2,
-	.num_xgmi_sdma_engines = 0,
-	.num_sdma_queues_per_engine = 8,
-};
-
-static const struct kfd_device_info arcturus_device_info = {
-	.asic_family = CHIP_ARCTURUS,
-	.asic_name = "arcturus",
-	.gfx_target_version = 90008,
-	.max_pasid_bits = 16,
-	.max_no_of_hqd	= 24,
-	.doorbell_size	= 8,
-	.ih_ring_entry_size = 8 * sizeof(uint32_t),
-	.event_interrupt_class = &event_interrupt_class_v9,
-	.num_of_watch_points = 4,
-	.mqd_size_aligned = MQD_SIZE_ALIGNED,
-	.supports_cwsr = true,
-	.needs_iommu_device = false,
-	.needs_pci_atomics = false,
-	.num_sdma_engines = 2,
-	.num_xgmi_sdma_engines = 6,
-	.num_sdma_queues_per_engine = 8,
-};
-
-static const struct kfd_device_info aldebaran_device_info = {
-	.asic_family = CHIP_ALDEBARAN,
-	.asic_name = "aldebaran",
-	.gfx_target_version = 90010,
-	.max_pasid_bits = 16,
-	.max_no_of_hqd	= 24,
-	.doorbell_size	= 8,
-	.ih_ring_entry_size = 8 * sizeof(uint32_t),
-	.event_interrupt_class = &event_interrupt_class_v9,
-	.num_of_watch_points = 4,
-	.mqd_size_aligned = MQD_SIZE_ALIGNED,
-	.supports_cwsr = true,
-	.needs_iommu_device = false,
-	.needs_pci_atomics = false,
-	.num_sdma_engines = 2,
-	.num_xgmi_sdma_engines = 3,
-	.num_sdma_queues_per_engine = 8,
-};
-
-static const struct kfd_device_info renoir_device_info = {
-	.asic_family = CHIP_RENOIR,
-	.asic_name = "renoir",
-	.gfx_target_version = 90012,
-	.max_pasid_bits = 16,
-	.max_no_of_hqd  = 24,
-	.doorbell_size  = 8,
-	.ih_ring_entry_size = 8 * sizeof(uint32_t),
-	.event_interrupt_class = &event_interrupt_class_v9,
-	.num_of_watch_points = 4,
-	.mqd_size_aligned = MQD_SIZE_ALIGNED,
-	.supports_cwsr = true,
-	.needs_iommu_device = false,
-	.needs_pci_atomics = false,
-	.num_sdma_engines = 1,
-	.num_xgmi_sdma_engines = 0,
-	.num_sdma_queues_per_engine = 2,
-};
-
-static const struct kfd_device_info navi10_device_info = {
-	.asic_family = CHIP_NAVI10,
-	.asic_name = "navi10",
-	.gfx_target_version = 100100,
-	.max_pasid_bits = 16,
-	.max_no_of_hqd  = 24,
-	.doorbell_size  = 8,
-	.ih_ring_entry_size = 8 * sizeof(uint32_t),
-	.event_interrupt_class = &event_interrupt_class_v9,
-	.num_of_watch_points = 4,
-	.mqd_size_aligned = MQD_SIZE_ALIGNED,
-	.needs_iommu_device = false,
-	.supports_cwsr = true,
-	.needs_pci_atomics = true,
-	.no_atomic_fw_version = 145,
-	.num_sdma_engines = 2,
-	.num_xgmi_sdma_engines = 0,
-	.num_sdma_queues_per_engine = 8,
-};
-
-static const struct kfd_device_info navi12_device_info = {
-	.asic_family = CHIP_NAVI12,
-	.asic_name = "navi12",
-	.gfx_target_version = 100101,
-	.max_pasid_bits = 16,
-	.max_no_of_hqd  = 24,
-	.doorbell_size  = 8,
-	.ih_ring_entry_size = 8 * sizeof(uint32_t),
-	.event_interrupt_class = &event_interrupt_class_v9,
-	.num_of_watch_points = 4,
-	.mqd_size_aligned = MQD_SIZE_ALIGNED,
-	.needs_iommu_device = false,
-	.supports_cwsr = true,
-	.needs_pci_atomics = true,
-	.no_atomic_fw_version = 145,
-	.num_sdma_engines = 2,
-	.num_xgmi_sdma_engines = 0,
-	.num_sdma_queues_per_engine = 8,
-};
-
-static const struct kfd_device_info navi14_device_info = {
-	.asic_family = CHIP_NAVI14,
-	.asic_name = "navi14",
-	.gfx_target_version = 100102,
-	.max_pasid_bits = 16,
-	.max_no_of_hqd  = 24,
-	.doorbell_size  = 8,
-	.ih_ring_entry_size = 8 * sizeof(uint32_t),
-	.event_interrupt_class = &event_interrupt_class_v9,
-	.num_of_watch_points = 4,
-	.mqd_size_aligned = MQD_SIZE_ALIGNED,
-	.needs_iommu_device = false,
-	.supports_cwsr = true,
-	.needs_pci_atomics = true,
-	.no_atomic_fw_version = 145,
-	.num_sdma_engines = 2,
-	.num_xgmi_sdma_engines = 0,
-	.num_sdma_queues_per_engine = 8,
-};
-
-static const struct kfd_device_info sienna_cichlid_device_info = {
-	.asic_family = CHIP_SIENNA_CICHLID,
-	.asic_name = "sienna_cichlid",
-	.gfx_target_version = 100300,
-	.max_pasid_bits = 16,
-	.max_no_of_hqd  = 24,
-	.doorbell_size  = 8,
-	.ih_ring_entry_size = 8 * sizeof(uint32_t),
-	.event_interrupt_class = &event_interrupt_class_v9,
-	.num_of_watch_points = 4,
-	.mqd_size_aligned = MQD_SIZE_ALIGNED,
-	.needs_iommu_device = false,
-	.supports_cwsr = true,
-	.needs_pci_atomics = true,
-	.no_atomic_fw_version = 92,
-	.num_sdma_engines = 4,
-	.num_xgmi_sdma_engines = 0,
-	.num_sdma_queues_per_engine = 8,
-};
-
-static const struct kfd_device_info navy_flounder_device_info = {
-	.asic_family = CHIP_NAVY_FLOUNDER,
-	.asic_name = "navy_flounder",
-	.gfx_target_version = 100301,
-	.max_pasid_bits = 16,
-	.max_no_of_hqd  = 24,
-	.doorbell_size  = 8,
-	.ih_ring_entry_size = 8 * sizeof(uint32_t),
-	.event_interrupt_class = &event_interrupt_class_v9,
-	.num_of_watch_points = 4,
-	.mqd_size_aligned = MQD_SIZE_ALIGNED,
-	.needs_iommu_device = false,
-	.supports_cwsr = true,
-	.needs_pci_atomics = true,
-	.no_atomic_fw_version = 92,
-	.num_sdma_engines = 2,
-	.num_xgmi_sdma_engines = 0,
-	.num_sdma_queues_per_engine = 8,
-};
-
-static const struct kfd_device_info vangogh_device_info = {
-	.asic_family = CHIP_VANGOGH,
-	.asic_name = "vangogh",
-	.gfx_target_version = 100303,
-	.max_pasid_bits = 16,
-	.max_no_of_hqd  = 24,
-	.doorbell_size  = 8,
-	.ih_ring_entry_size = 8 * sizeof(uint32_t),
-	.event_interrupt_class = &event_interrupt_class_v9,
-	.num_of_watch_points = 4,
-	.mqd_size_aligned = MQD_SIZE_ALIGNED,
-	.needs_iommu_device = false,
-	.supports_cwsr = true,
-	.needs_pci_atomics = true,
-	.no_atomic_fw_version = 92,
-	.num_sdma_engines = 1,
-	.num_xgmi_sdma_engines = 0,
-	.num_sdma_queues_per_engine = 2,
-};
-
-static const struct kfd_device_info dimgrey_cavefish_device_info = {
-	.asic_family = CHIP_DIMGREY_CAVEFISH,
-	.asic_name = "dimgrey_cavefish",
-	.gfx_target_version = 100302,
-	.max_pasid_bits = 16,
-	.max_no_of_hqd  = 24,
-	.doorbell_size  = 8,
-	.ih_ring_entry_size = 8 * sizeof(uint32_t),
-	.event_interrupt_class = &event_interrupt_class_v9,
-	.num_of_watch_points = 4,
-	.mqd_size_aligned = MQD_SIZE_ALIGNED,
-	.needs_iommu_device = false,
-	.supports_cwsr = true,
-	.needs_pci_atomics = true,
-	.no_atomic_fw_version = 92,
-	.num_sdma_engines = 2,
-	.num_xgmi_sdma_engines = 0,
-	.num_sdma_queues_per_engine = 8,
-};
-
-static const struct kfd_device_info beige_goby_device_info = {
-	.asic_family = CHIP_BEIGE_GOBY,
-	.asic_name = "beige_goby",
-	.gfx_target_version = 100304,
-	.max_pasid_bits = 16,
-	.max_no_of_hqd  = 24,
-	.doorbell_size  = 8,
-	.ih_ring_entry_size = 8 * sizeof(uint32_t),
-	.event_interrupt_class = &event_interrupt_class_v9,
-	.num_of_watch_points = 4,
-	.mqd_size_aligned = MQD_SIZE_ALIGNED,
-	.needs_iommu_device = false,
-	.supports_cwsr = true,
-	.needs_pci_atomics = true,
-	.no_atomic_fw_version = 92,
-	.num_sdma_engines = 1,
-	.num_xgmi_sdma_engines = 0,
-	.num_sdma_queues_per_engine = 8,
-};
-
-static const struct kfd_device_info yellow_carp_device_info = {
-	.asic_family = CHIP_YELLOW_CARP,
-	.asic_name = "yellow_carp",
-	.gfx_target_version = 100305,
-	.max_pasid_bits = 16,
-	.max_no_of_hqd  = 24,
-	.doorbell_size  = 8,
-	.ih_ring_entry_size = 8 * sizeof(uint32_t),
-	.event_interrupt_class = &event_interrupt_class_v9,
-	.num_of_watch_points = 4,
-	.mqd_size_aligned = MQD_SIZE_ALIGNED,
-	.needs_iommu_device = false,
-	.supports_cwsr = true,
-	.needs_pci_atomics = true,
-	.no_atomic_fw_version = 92,
-	.num_sdma_engines = 1,
-	.num_xgmi_sdma_engines = 0,
-	.num_sdma_queues_per_engine = 2,
-};
-
-static const struct kfd_device_info cyan_skillfish_device_info = {
-	.asic_family = CHIP_CYAN_SKILLFISH,
-	.asic_name = "cyan_skillfish",
-	.gfx_target_version = 100103,
-	.max_pasid_bits = 16,
-	.max_no_of_hqd  = 24,
-	.doorbell_size  = 8,
-	.ih_ring_entry_size = 8 * sizeof(uint32_t),
-	.event_interrupt_class = &event_interrupt_class_v9,
-	.num_of_watch_points = 4,
-	.mqd_size_aligned = MQD_SIZE_ALIGNED,
-	.needs_iommu_device = false,
-	.supports_cwsr = true,
-	.needs_pci_atomics = true,
-	.num_sdma_engines = 2,
-	.num_xgmi_sdma_engines = 0,
-	.num_sdma_queues_per_engine = 8,
-};
+extern const struct kfd2kgd_calls gfx_v11_kfd2kgd;
+extern const struct kfd2kgd_calls gfx_v12_kfd2kgd;
 
 static int kfd_gtt_sa_init(struct kfd_dev *kfd, unsigned int buf_size,
 				unsigned int chunk_size);
 static void kfd_gtt_sa_fini(struct kfd_dev *kfd);
 
-static int kfd_resume(struct kfd_dev *kfd);
+static int kfd_resume(struct kfd_node *kfd);
 
-struct kfd_dev *kgd2kfd_probe(struct kgd_dev *kgd, bool vf)
+static void kfd_device_info_set_sdma_info(struct kfd_dev *kfd)
 {
-	struct kfd_dev *kfd;
-	const struct kfd_device_info *device_info;
-	const struct kfd2kgd_calls *f2g;
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
-	struct pci_dev *pdev = adev->pdev;
+	uint32_t sdma_version = amdgpu_ip_version(kfd->adev, SDMA0_HWIP, 0);
+
+	switch (sdma_version) {
+	case IP_VERSION(4, 0, 0):/* VEGA10 */
+	case IP_VERSION(4, 0, 1):/* VEGA12 */
+	case IP_VERSION(4, 1, 0):/* RAVEN */
+	case IP_VERSION(4, 1, 1):/* RAVEN */
+	case IP_VERSION(4, 1, 2):/* RENOIR */
+	case IP_VERSION(5, 2, 1):/* VANGOGH */
+	case IP_VERSION(5, 2, 3):/* YELLOW_CARP */
+	case IP_VERSION(5, 2, 6):/* GC 10.3.6 */
+	case IP_VERSION(5, 2, 7):/* GC 10.3.7 */
+		kfd->device_info.num_sdma_queues_per_engine = 2;
+		break;
+	case IP_VERSION(4, 2, 0):/* VEGA20 */
+	case IP_VERSION(4, 2, 2):/* ARCTURUS */
+	case IP_VERSION(4, 4, 0):/* ALDEBARAN */
+	case IP_VERSION(4, 4, 2):
+	case IP_VERSION(4, 4, 5):
+	case IP_VERSION(4, 4, 4):
+	case IP_VERSION(5, 0, 0):/* NAVI10 */
+	case IP_VERSION(5, 0, 1):/* CYAN_SKILLFISH */
+	case IP_VERSION(5, 0, 2):/* NAVI14 */
+	case IP_VERSION(5, 0, 5):/* NAVI12 */
+	case IP_VERSION(5, 2, 0):/* SIENNA_CICHLID */
+	case IP_VERSION(5, 2, 2):/* NAVY_FLOUNDER */
+	case IP_VERSION(5, 2, 4):/* DIMGREY_CAVEFISH */
+	case IP_VERSION(5, 2, 5):/* BEIGE_GOBY */
+	case IP_VERSION(6, 0, 0):
+	case IP_VERSION(6, 0, 1):
+	case IP_VERSION(6, 0, 2):
+	case IP_VERSION(6, 0, 3):
+	case IP_VERSION(6, 1, 0):
+	case IP_VERSION(6, 1, 1):
+	case IP_VERSION(6, 1, 2):
+	case IP_VERSION(6, 1, 3):
+	case IP_VERSION(7, 0, 0):
+	case IP_VERSION(7, 0, 1):
+		kfd->device_info.num_sdma_queues_per_engine = 8;
+		break;
+	default:
+		dev_warn(kfd_device,
+			"Default sdma queue per engine(8) is set due to mismatch of sdma ip block(SDMA_HWIP:0x%x).\n",
+			sdma_version);
+		kfd->device_info.num_sdma_queues_per_engine = 8;
+	}
+
+	bitmap_zero(kfd->device_info.reserved_sdma_queues_bitmap, KFD_MAX_SDMA_QUEUES);
+
+	switch (sdma_version) {
+	case IP_VERSION(6, 0, 0):
+	case IP_VERSION(6, 0, 1):
+	case IP_VERSION(6, 0, 2):
+	case IP_VERSION(6, 0, 3):
+	case IP_VERSION(6, 1, 0):
+	case IP_VERSION(6, 1, 1):
+	case IP_VERSION(6, 1, 2):
+	case IP_VERSION(6, 1, 3):
+	case IP_VERSION(7, 0, 0):
+	case IP_VERSION(7, 0, 1):
+		/* Reserve 1 for paging and 1 for gfx */
+		kfd->device_info.num_reserved_sdma_queues_per_engine = 2;
+		/* BIT(0)=engine-0 queue-0; BIT(1)=engine-1 queue-0; BIT(2)=engine-0 queue-1; ... */
+		bitmap_set(kfd->device_info.reserved_sdma_queues_bitmap, 0,
+			   kfd->adev->sdma.num_instances *
+			   kfd->device_info.num_reserved_sdma_queues_per_engine);
+		break;
+	default:
+		break;
+	}
+}
+
+static void kfd_device_info_set_event_interrupt_class(struct kfd_dev *kfd)
+{
+	uint32_t gc_version = KFD_GC_VERSION(kfd);
+
+	switch (gc_version) {
+	case IP_VERSION(9, 0, 1): /* VEGA10 */
+	case IP_VERSION(9, 1, 0): /* RAVEN */
+	case IP_VERSION(9, 2, 1): /* VEGA12 */
+	case IP_VERSION(9, 2, 2): /* RAVEN */
+	case IP_VERSION(9, 3, 0): /* RENOIR */
+	case IP_VERSION(9, 4, 0): /* VEGA20 */
+	case IP_VERSION(9, 4, 1): /* ARCTURUS */
+	case IP_VERSION(9, 4, 2): /* ALDEBARAN */
+		kfd->device_info.event_interrupt_class = &event_interrupt_class_v9;
+		break;
+	case IP_VERSION(9, 4, 3): /* GC 9.4.3 */
+	case IP_VERSION(9, 4, 4): /* GC 9.4.4 */
+	case IP_VERSION(9, 5, 0): /* GC 9.5.0 */
+		kfd->device_info.event_interrupt_class =
+						&event_interrupt_class_v9_4_3;
+		break;
+	case IP_VERSION(10, 3, 1): /* VANGOGH */
+	case IP_VERSION(10, 3, 3): /* YELLOW_CARP */
+	case IP_VERSION(10, 3, 6): /* GC 10.3.6 */
+	case IP_VERSION(10, 3, 7): /* GC 10.3.7 */
+	case IP_VERSION(10, 1, 3): /* CYAN_SKILLFISH */
+	case IP_VERSION(10, 1, 4):
+	case IP_VERSION(10, 1, 10): /* NAVI10 */
+	case IP_VERSION(10, 1, 2): /* NAVI12 */
+	case IP_VERSION(10, 1, 1): /* NAVI14 */
+	case IP_VERSION(10, 3, 0): /* SIENNA_CICHLID */
+	case IP_VERSION(10, 3, 2): /* NAVY_FLOUNDER */
+	case IP_VERSION(10, 3, 4): /* DIMGREY_CAVEFISH */
+	case IP_VERSION(10, 3, 5): /* BEIGE_GOBY */
+		kfd->device_info.event_interrupt_class = &event_interrupt_class_v10;
+		break;
+	case IP_VERSION(11, 0, 0):
+	case IP_VERSION(11, 0, 1):
+	case IP_VERSION(11, 0, 2):
+	case IP_VERSION(11, 0, 3):
+	case IP_VERSION(11, 0, 4):
+	case IP_VERSION(11, 5, 0):
+	case IP_VERSION(11, 5, 1):
+	case IP_VERSION(11, 5, 2):
+	case IP_VERSION(11, 5, 3):
+		kfd->device_info.event_interrupt_class = &event_interrupt_class_v11;
+		break;
+	case IP_VERSION(12, 0, 0):
+	case IP_VERSION(12, 0, 1):
+		/* GFX12_TODO: Change to v12 version. */
+		kfd->device_info.event_interrupt_class = &event_interrupt_class_v11;
+		break;
+	default:
+		dev_warn(kfd_device, "v9 event interrupt handler is set due to "
+			"mismatch of gc ip block(GC_HWIP:0x%x).\n", gc_version);
+		kfd->device_info.event_interrupt_class = &event_interrupt_class_v9;
+	}
+}
+
+static void kfd_device_info_init(struct kfd_dev *kfd,
+				 bool vf, uint32_t gfx_target_version)
+{
+	uint32_t gc_version = KFD_GC_VERSION(kfd);
+	uint32_t asic_type = kfd->adev->asic_type;
+
+	kfd->device_info.max_pasid_bits = 16;
+	kfd->device_info.max_no_of_hqd = 24;
+	kfd->device_info.num_of_watch_points = 4;
+	kfd->device_info.mqd_size_aligned = MQD_SIZE_ALIGNED;
+	kfd->device_info.gfx_target_version = gfx_target_version;
+
+	if (KFD_IS_SOC15(kfd)) {
+		kfd->device_info.doorbell_size = 8;
+		kfd->device_info.ih_ring_entry_size = 8 * sizeof(uint32_t);
+		kfd->device_info.supports_cwsr = true;
+
+		kfd_device_info_set_sdma_info(kfd);
+
+		kfd_device_info_set_event_interrupt_class(kfd);
+
+		if (gc_version < IP_VERSION(11, 0, 0)) {
+			/* Navi2x+, Navi1x+ */
+			if (gc_version == IP_VERSION(10, 3, 6))
+				kfd->device_info.no_atomic_fw_version = 14;
+			else if (gc_version == IP_VERSION(10, 3, 7))
+				kfd->device_info.no_atomic_fw_version = 3;
+			else if (gc_version >= IP_VERSION(10, 3, 0))
+				kfd->device_info.no_atomic_fw_version = 92;
+			else if (gc_version >= IP_VERSION(10, 1, 1))
+				kfd->device_info.no_atomic_fw_version = 145;
+
+			/* Navi1x+ */
+			if (gc_version >= IP_VERSION(10, 1, 1))
+				kfd->device_info.needs_pci_atomics = true;
+		} else if (gc_version < IP_VERSION(12, 0, 0)) {
+			/*
+			 * PCIe atomics support acknowledgment in GFX11 RS64 CPFW requires
+			 * MEC version >= 509. Prior RS64 CPFW versions (and all F32) require
+			 * PCIe atomics support.
+			 */
+			kfd->device_info.needs_pci_atomics = true;
+			kfd->device_info.no_atomic_fw_version = kfd->adev->gfx.rs64_enable ? 509 : 0;
+		} else if (gc_version < IP_VERSION(13, 0, 0)) {
+			kfd->device_info.needs_pci_atomics = true;
+			kfd->device_info.no_atomic_fw_version = 2090;
+		} else {
+			kfd->device_info.needs_pci_atomics = true;
+		}
+	} else {
+		kfd->device_info.doorbell_size = 4;
+		kfd->device_info.ih_ring_entry_size = 4 * sizeof(uint32_t);
+		kfd->device_info.event_interrupt_class = &event_interrupt_class_cik;
+		kfd->device_info.num_sdma_queues_per_engine = 2;
+
+		if (asic_type != CHIP_KAVERI &&
+		    asic_type != CHIP_HAWAII &&
+		    asic_type != CHIP_TONGA)
+			kfd->device_info.supports_cwsr = true;
+
+		if (asic_type != CHIP_HAWAII && !vf)
+			kfd->device_info.needs_pci_atomics = true;
+	}
+}
+
+struct kfd_dev *kgd2kfd_probe(struct amdgpu_device *adev, bool vf)
+{
+	struct kfd_dev *kfd = NULL;
+	const struct kfd2kgd_calls *f2g = NULL;
+	uint32_t gfx_target_version = 0;
 
 	switch (adev->asic_type) {
-#ifdef KFD_SUPPORT_IOMMU_V2
 #ifdef CONFIG_DRM_AMDGPU_CIK
 	case CHIP_KAVERI:
-		if (vf)
-			device_info = NULL;
-		else
-			device_info = &kaveri_device_info;
-		f2g = &gfx_v7_kfd2kgd;
+		gfx_target_version = 70000;
+		if (!vf)
+			f2g = &gfx_v7_kfd2kgd;
 		break;
 #endif
 	case CHIP_CARRIZO:
-		if (vf)
-			device_info = NULL;
-		else
-			device_info = &carrizo_device_info;
-		f2g = &gfx_v8_kfd2kgd;
+		gfx_target_version = 80001;
+		if (!vf)
+			f2g = &gfx_v8_kfd2kgd;
 		break;
-#endif
 #ifdef CONFIG_DRM_AMDGPU_CIK
 	case CHIP_HAWAII:
-		if (vf)
-			device_info = NULL;
-		else
-			device_info = &hawaii_device_info;
-		f2g = &gfx_v7_kfd2kgd;
+		gfx_target_version = 70001;
+		if (!amdgpu_exp_hw_support)
+			pr_info(
+	"KFD support on Hawaii is experimental. See modparam exp_hw_support\n"
+				);
+		else if (!vf)
+			f2g = &gfx_v7_kfd2kgd;
 		break;
 #endif
 	case CHIP_TONGA:
-		if (vf)
-			device_info = NULL;
-		else
-			device_info = &tonga_device_info;
-		f2g = &gfx_v8_kfd2kgd;
+		gfx_target_version = 80002;
+		if (!vf)
+			f2g = &gfx_v8_kfd2kgd;
 		break;
 	case CHIP_FIJI:
-		if (vf)
-			device_info = &fiji_vf_device_info;
-		else
-			device_info = &fiji_device_info;
-		f2g = &gfx_v8_kfd2kgd;
-		break;
 	case CHIP_POLARIS10:
-		if (vf)
-			device_info = &polaris10_vf_device_info;
-		else
-			device_info = &polaris10_device_info;
+		gfx_target_version = 80003;
 		f2g = &gfx_v8_kfd2kgd;
 		break;
 	case CHIP_POLARIS11:
-		if (vf)
-			device_info = NULL;
-		else
-			device_info = &polaris11_device_info;
-		f2g = &gfx_v8_kfd2kgd;
-		break;
 	case CHIP_POLARIS12:
-		if (vf)
-			device_info = NULL;
-		else
-			device_info = &polaris12_device_info;
-		f2g = &gfx_v8_kfd2kgd;
-		break;
 	case CHIP_VEGAM:
-		if (vf)
-			device_info = NULL;
-		else
-			device_info = &vegam_device_info;
-		f2g = &gfx_v8_kfd2kgd;
+		gfx_target_version = 80003;
+		if (!vf)
+			f2g = &gfx_v8_kfd2kgd;
 		break;
 	default:
-		switch (adev->ip_versions[GC_HWIP][0]) {
+		switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+		/* Vega 10 */
 		case IP_VERSION(9, 0, 1):
-			if (vf)
-				device_info = &vega10_vf_device_info;
-			else
-				device_info = &vega10_device_info;
+			gfx_target_version = 90000;
 			f2g = &gfx_v9_kfd2kgd;
 			break;
-#ifdef KFD_SUPPORT_IOMMU_V2
+		/* Raven */
 		case IP_VERSION(9, 1, 0):
 		case IP_VERSION(9, 2, 2):
-			if (vf)
-				device_info = NULL;
-			else
-				device_info = &raven_device_info;
-			f2g = &gfx_v9_kfd2kgd;
+			gfx_target_version = 90002;
+			if (!vf)
+				f2g = &gfx_v9_kfd2kgd;
 			break;
-#endif
+		/* Vega12 */
 		case IP_VERSION(9, 2, 1):
-			if (vf)
-				device_info = NULL;
-			else
-				device_info = &vega12_device_info;
-			f2g = &gfx_v9_kfd2kgd;
+			gfx_target_version = 90004;
+			if (!vf)
+				f2g = &gfx_v9_kfd2kgd;
 			break;
+		/* Renoir */
 		case IP_VERSION(9, 3, 0):
-			if (vf)
-				device_info = NULL;
-			else
-				device_info = &renoir_device_info;
-			f2g = &gfx_v9_kfd2kgd;
+			gfx_target_version = 90012;
+			if (!vf)
+				f2g = &gfx_v9_kfd2kgd;
 			break;
+		/* Vega20 */
 		case IP_VERSION(9, 4, 0):
-			if (vf)
-				device_info = NULL;
-			else
-				device_info = &vega20_device_info;
-			f2g = &gfx_v9_kfd2kgd;
+			gfx_target_version = 90006;
+			if (!vf)
+				f2g = &gfx_v9_kfd2kgd;
 			break;
+		/* Arcturus */
 		case IP_VERSION(9, 4, 1):
-			device_info = &arcturus_device_info;
+			gfx_target_version = 90008;
 			f2g = &arcturus_kfd2kgd;
 			break;
+		/* Aldebaran */
 		case IP_VERSION(9, 4, 2):
-			device_info = &aldebaran_device_info;
+			gfx_target_version = 90010;
 			f2g = &aldebaran_kfd2kgd;
 			break;
+		case IP_VERSION(9, 4, 3):
+		case IP_VERSION(9, 4, 4):
+			gfx_target_version = 90402;
+			f2g = &gc_9_4_3_kfd2kgd;
+			break;
+		case IP_VERSION(9, 5, 0):
+			gfx_target_version = 90500;
+			f2g = &gc_9_4_3_kfd2kgd;
+			break;
+		/* Navi10 */
 		case IP_VERSION(10, 1, 10):
-			if (vf)
-				device_info = NULL;
-			else
-				device_info = &navi10_device_info;
-			f2g = &gfx_v10_kfd2kgd;
+			gfx_target_version = 100100;
+			if (!vf)
+				f2g = &gfx_v10_kfd2kgd;
 			break;
+		/* Navi12 */
 		case IP_VERSION(10, 1, 2):
-			device_info = &navi12_device_info;
+			gfx_target_version = 100101;
 			f2g = &gfx_v10_kfd2kgd;
 			break;
+		/* Navi14 */
 		case IP_VERSION(10, 1, 1):
-			if (vf)
-				device_info = NULL;
-			else
-				device_info = &navi14_device_info;
-			f2g = &gfx_v10_kfd2kgd;
+			gfx_target_version = 100102;
+			if (!vf)
+				f2g = &gfx_v10_kfd2kgd;
 			break;
+		/* Cyan Skillfish */
 		case IP_VERSION(10, 1, 3):
-			if (vf)
-				device_info = NULL;
-			else
-				device_info = &cyan_skillfish_device_info;
-			f2g = &gfx_v10_kfd2kgd;
+		case IP_VERSION(10, 1, 4):
+			gfx_target_version = 100103;
+			if (!vf)
+				f2g = &gfx_v10_kfd2kgd;
 			break;
+		/* Sienna Cichlid */
 		case IP_VERSION(10, 3, 0):
-			device_info = &sienna_cichlid_device_info;
+			gfx_target_version = 100300;
 			f2g = &gfx_v10_3_kfd2kgd;
 			break;
+		/* Navy Flounder */
 		case IP_VERSION(10, 3, 2):
-			device_info = &navy_flounder_device_info;
+			gfx_target_version = 100301;
 			f2g = &gfx_v10_3_kfd2kgd;
 			break;
+		/* Van Gogh */
 		case IP_VERSION(10, 3, 1):
-			if (vf)
-				device_info = NULL;
-			else
-				device_info = &vangogh_device_info;
-			f2g = &gfx_v10_3_kfd2kgd;
+			gfx_target_version = 100303;
+			if (!vf)
+				f2g = &gfx_v10_3_kfd2kgd;
 			break;
+		/* Dimgrey Cavefish */
 		case IP_VERSION(10, 3, 4):
-			device_info = &dimgrey_cavefish_device_info;
+			gfx_target_version = 100302;
 			f2g = &gfx_v10_3_kfd2kgd;
 			break;
+		/* Beige Goby */
 		case IP_VERSION(10, 3, 5):
-			device_info = &beige_goby_device_info;
+			gfx_target_version = 100304;
 			f2g = &gfx_v10_3_kfd2kgd;
 			break;
+		/* Yellow Carp */
 		case IP_VERSION(10, 3, 3):
-			if (vf)
-				device_info = NULL;
-			else
-				device_info = &yellow_carp_device_info;
-			f2g = &gfx_v10_3_kfd2kgd;
+			gfx_target_version = 100305;
+			if (!vf)
+				f2g = &gfx_v10_3_kfd2kgd;
+			break;
+		case IP_VERSION(10, 3, 6):
+		case IP_VERSION(10, 3, 7):
+			gfx_target_version = 100306;
+			if (!vf)
+				f2g = &gfx_v10_3_kfd2kgd;
+			break;
+		case IP_VERSION(11, 0, 0):
+			gfx_target_version = 110000;
+			f2g = &gfx_v11_kfd2kgd;
+			break;
+		case IP_VERSION(11, 0, 1):
+		case IP_VERSION(11, 0, 4):
+			gfx_target_version = 110003;
+			f2g = &gfx_v11_kfd2kgd;
+			break;
+		case IP_VERSION(11, 0, 2):
+			gfx_target_version = 110002;
+			f2g = &gfx_v11_kfd2kgd;
+			break;
+		case IP_VERSION(11, 0, 3):
+			/* Note: Compiler version is 11.0.1 while HW version is 11.0.3 */
+			gfx_target_version = 110001;
+			f2g = &gfx_v11_kfd2kgd;
+			break;
+		case IP_VERSION(11, 5, 0):
+			gfx_target_version = 110500;
+			f2g = &gfx_v11_kfd2kgd;
+			break;
+		case IP_VERSION(11, 5, 1):
+			gfx_target_version = 110501;
+			f2g = &gfx_v11_kfd2kgd;
+			break;
+		case IP_VERSION(11, 5, 2):
+			gfx_target_version = 110502;
+			f2g = &gfx_v11_kfd2kgd;
+			break;
+		case IP_VERSION(11, 5, 3):
+			gfx_target_version = 110503;
+			f2g = &gfx_v11_kfd2kgd;
+			break;
+		case IP_VERSION(12, 0, 0):
+			gfx_target_version = 120000;
+			f2g = &gfx_v12_kfd2kgd;
+			break;
+		case IP_VERSION(12, 0, 1):
+			gfx_target_version = 120001;
+			f2g = &gfx_v12_kfd2kgd;
 			break;
 		default:
-			return NULL;
+			break;
 		}
 		break;
 	}
 
-	if (!device_info || !f2g) {
-		dev_err(kfd_device, "%s %s not supported in kfd\n",
-			amdgpu_asic_name[adev->asic_type], vf ? "VF" : "");
+	if (!f2g) {
+		if (amdgpu_ip_version(adev, GC_HWIP, 0))
+			dev_info(kfd_device,
+				"GC IP %06x %s not supported in kfd\n",
+				amdgpu_ip_version(adev, GC_HWIP, 0),
+				vf ? "VF" : "");
+		else
+			dev_info(kfd_device, "%s %s not supported in kfd\n",
+				amdgpu_asic_name[adev->asic_type], vf ? "VF" : "");
 		return NULL;
 	}
 
@@ -824,145 +486,315 @@ struct kfd_dev *kgd2kfd_probe(struct kgd_dev *kgd, bool vf)
 	if (!kfd)
 		return NULL;
 
-	kfd->kgd = kgd;
-	kfd->device_info = device_info;
-	kfd->pdev = pdev;
+	kfd->adev = adev;
+	kfd_device_info_init(kfd, vf, gfx_target_version);
 	kfd->init_complete = false;
 	kfd->kfd2kgd = f2g;
 	atomic_set(&kfd->compute_profile, 0);
 
 	mutex_init(&kfd->doorbell_mutex);
-	memset(&kfd->doorbell_available_index, 0,
-		sizeof(kfd->doorbell_available_index));
-
-	atomic_set(&kfd->sram_ecc_flag, 0);
 
 	ida_init(&kfd->doorbell_ida);
+	atomic_set(&kfd->kfd_processes_count, 0);
 
 	return kfd;
 }
 
 static void kfd_cwsr_init(struct kfd_dev *kfd)
 {
-	if (cwsr_enable && kfd->device_info->supports_cwsr) {
-		if (kfd->device_info->asic_family < CHIP_VEGA10) {
-			BUILD_BUG_ON(sizeof(cwsr_trap_gfx8_hex) > PAGE_SIZE);
+	if (cwsr_enable && kfd->device_info.supports_cwsr) {
+		if (KFD_GC_VERSION(kfd) < IP_VERSION(9, 0, 1)) {
+			BUILD_BUG_ON(sizeof(cwsr_trap_gfx8_hex)
+					     > KFD_CWSR_TMA_OFFSET);
 			kfd->cwsr_isa = cwsr_trap_gfx8_hex;
 			kfd->cwsr_isa_size = sizeof(cwsr_trap_gfx8_hex);
-		} else if (kfd->device_info->asic_family == CHIP_ARCTURUS) {
-			BUILD_BUG_ON(sizeof(cwsr_trap_arcturus_hex) > PAGE_SIZE);
+		} else if (KFD_GC_VERSION(kfd) == IP_VERSION(9, 4, 1)) {
+			BUILD_BUG_ON(sizeof(cwsr_trap_arcturus_hex)
+					     > KFD_CWSR_TMA_OFFSET);
 			kfd->cwsr_isa = cwsr_trap_arcturus_hex;
 			kfd->cwsr_isa_size = sizeof(cwsr_trap_arcturus_hex);
-		} else if (kfd->device_info->asic_family == CHIP_ALDEBARAN) {
-			BUILD_BUG_ON(sizeof(cwsr_trap_aldebaran_hex) > PAGE_SIZE);
+		} else if (KFD_GC_VERSION(kfd) == IP_VERSION(9, 4, 2)) {
+			BUILD_BUG_ON(sizeof(cwsr_trap_aldebaran_hex)
+					     > KFD_CWSR_TMA_OFFSET);
 			kfd->cwsr_isa = cwsr_trap_aldebaran_hex;
 			kfd->cwsr_isa_size = sizeof(cwsr_trap_aldebaran_hex);
-		} else if (kfd->device_info->asic_family < CHIP_NAVI10) {
-			BUILD_BUG_ON(sizeof(cwsr_trap_gfx9_hex) > PAGE_SIZE);
+		} else if (KFD_GC_VERSION(kfd) == IP_VERSION(9, 4, 3) ||
+			   KFD_GC_VERSION(kfd) == IP_VERSION(9, 4, 4)) {
+			BUILD_BUG_ON(sizeof(cwsr_trap_gfx9_4_3_hex)
+					     > KFD_CWSR_TMA_OFFSET);
+			kfd->cwsr_isa = cwsr_trap_gfx9_4_3_hex;
+			kfd->cwsr_isa_size = sizeof(cwsr_trap_gfx9_4_3_hex);
+		} else if (KFD_GC_VERSION(kfd) == IP_VERSION(9, 5, 0)) {
+			BUILD_BUG_ON(sizeof(cwsr_trap_gfx9_5_0_hex) > PAGE_SIZE);
+			kfd->cwsr_isa = cwsr_trap_gfx9_5_0_hex;
+			kfd->cwsr_isa_size = sizeof(cwsr_trap_gfx9_5_0_hex);
+		} else if (KFD_GC_VERSION(kfd) < IP_VERSION(10, 1, 1)) {
+			BUILD_BUG_ON(sizeof(cwsr_trap_gfx9_hex)
+					     > KFD_CWSR_TMA_OFFSET);
 			kfd->cwsr_isa = cwsr_trap_gfx9_hex;
 			kfd->cwsr_isa_size = sizeof(cwsr_trap_gfx9_hex);
-		} else if (kfd->device_info->asic_family < CHIP_SIENNA_CICHLID) {
-			BUILD_BUG_ON(sizeof(cwsr_trap_nv1x_hex) > PAGE_SIZE);
+		} else if (KFD_GC_VERSION(kfd) < IP_VERSION(10, 3, 0)) {
+			BUILD_BUG_ON(sizeof(cwsr_trap_nv1x_hex)
+					     > KFD_CWSR_TMA_OFFSET);
 			kfd->cwsr_isa = cwsr_trap_nv1x_hex;
 			kfd->cwsr_isa_size = sizeof(cwsr_trap_nv1x_hex);
-		} else {
-			BUILD_BUG_ON(sizeof(cwsr_trap_gfx10_hex) > PAGE_SIZE);
+		} else if (KFD_GC_VERSION(kfd) < IP_VERSION(11, 0, 0)) {
+			BUILD_BUG_ON(sizeof(cwsr_trap_gfx10_hex)
+					     > KFD_CWSR_TMA_OFFSET);
 			kfd->cwsr_isa = cwsr_trap_gfx10_hex;
 			kfd->cwsr_isa_size = sizeof(cwsr_trap_gfx10_hex);
+		} else if (KFD_GC_VERSION(kfd) < IP_VERSION(12, 0, 0)) {
+			/* The gfx11 cwsr trap handler must fit inside a single
+			   page. */
+			BUILD_BUG_ON(sizeof(cwsr_trap_gfx11_hex) > PAGE_SIZE);
+			kfd->cwsr_isa = cwsr_trap_gfx11_hex;
+			kfd->cwsr_isa_size = sizeof(cwsr_trap_gfx11_hex);
+		} else {
+			BUILD_BUG_ON(sizeof(cwsr_trap_gfx12_hex)
+					     > KFD_CWSR_TMA_OFFSET);
+			kfd->cwsr_isa = cwsr_trap_gfx12_hex;
+			kfd->cwsr_isa_size = sizeof(cwsr_trap_gfx12_hex);
 		}
 
 		kfd->cwsr_enabled = true;
 	}
 }
 
-static int kfd_gws_init(struct kfd_dev *kfd)
+static int kfd_gws_init(struct kfd_node *node)
 {
 	int ret = 0;
+	struct kfd_dev *kfd = node->kfd;
+	uint32_t mes_rev = node->adev->mes.sched_version & AMDGPU_MES_VERSION_MASK;
 
-	if (kfd->dqm->sched_policy == KFD_SCHED_POLICY_NO_HWS)
+	if (node->dqm->sched_policy == KFD_SCHED_POLICY_NO_HWS)
 		return 0;
 
-	if (hws_gws_support
-		|| (kfd->device_info->asic_family == CHIP_VEGA10
-			&& kfd->mec2_fw_version >= 0x81b3)
-		|| (kfd->device_info->asic_family >= CHIP_VEGA12
-			&& kfd->device_info->asic_family <= CHIP_RAVEN
-			&& kfd->mec2_fw_version >= 0x1b3)
-		|| (kfd->device_info->asic_family == CHIP_ARCTURUS
-			&& kfd->mec2_fw_version >= 0x30)
-		|| (kfd->device_info->asic_family == CHIP_ALDEBARAN
-			&& kfd->mec2_fw_version >= 0x28))
-		ret = amdgpu_amdkfd_alloc_gws(kfd->kgd,
-				amdgpu_amdkfd_get_num_gws(kfd->kgd), &kfd->gws);
+	if (hws_gws_support || (KFD_IS_SOC15(node) &&
+		((KFD_GC_VERSION(node) == IP_VERSION(9, 0, 1)
+			&& kfd->mec2_fw_version >= 0x81b3) ||
+		(KFD_GC_VERSION(node) <= IP_VERSION(9, 4, 0)
+			&& kfd->mec2_fw_version >= 0x1b3)  ||
+		(KFD_GC_VERSION(node) == IP_VERSION(9, 4, 1)
+			&& kfd->mec2_fw_version >= 0x30)   ||
+		(KFD_GC_VERSION(node) == IP_VERSION(9, 4, 2)
+			&& kfd->mec2_fw_version >= 0x28) ||
+		(KFD_GC_VERSION(node) == IP_VERSION(9, 4, 3) ||
+		 KFD_GC_VERSION(node) == IP_VERSION(9, 4, 4)) ||
+		(KFD_GC_VERSION(node) == IP_VERSION(9, 5, 0)) ||
+		(KFD_GC_VERSION(node) >= IP_VERSION(10, 3, 0)
+			&& KFD_GC_VERSION(node) < IP_VERSION(11, 0, 0)
+			&& kfd->mec2_fw_version >= 0x6b) ||
+		(KFD_GC_VERSION(node) >= IP_VERSION(11, 0, 0)
+			&& KFD_GC_VERSION(node) < IP_VERSION(12, 0, 0)
+			&& mes_rev >= 68) ||
+		(KFD_GC_VERSION(node) >= IP_VERSION(12, 0, 0))))) {
+		if (KFD_GC_VERSION(node) >= IP_VERSION(12, 0, 0))
+			node->adev->gds.gws_size = 64;
+		ret = amdgpu_amdkfd_alloc_gws(node->adev,
+				node->adev->gds.gws_size, &node->gws);
+	}
 
 	return ret;
 }
 
-static void kfd_smi_init(struct kfd_dev *dev) {
+static void kfd_smi_init(struct kfd_node *dev)
+{
 	INIT_LIST_HEAD(&dev->smi_clients);
 	spin_lock_init(&dev->smi_lock);
 }
 
+static int kfd_init_node(struct kfd_node *node)
+{
+	int err = -1;
+
+	if (kfd_interrupt_init(node)) {
+		dev_err(kfd_device, "Error initializing interrupts\n");
+		goto kfd_interrupt_error;
+	}
+
+	node->dqm = device_queue_manager_init(node);
+	if (!node->dqm) {
+		dev_err(kfd_device, "Error initializing queue manager\n");
+		goto device_queue_manager_error;
+	}
+
+	if (kfd_gws_init(node)) {
+		dev_err(kfd_device, "Could not allocate %d gws\n",
+			node->adev->gds.gws_size);
+		goto gws_error;
+	}
+
+	if (kfd_resume(node))
+		goto kfd_resume_error;
+
+	if (kfd_topology_add_device(node)) {
+		dev_err(kfd_device, "Error adding device to topology\n");
+		goto kfd_topology_add_device_error;
+	}
+
+	kfd_smi_init(node);
+
+	return 0;
+
+kfd_topology_add_device_error:
+kfd_resume_error:
+gws_error:
+	device_queue_manager_uninit(node->dqm);
+device_queue_manager_error:
+	kfd_interrupt_exit(node);
+kfd_interrupt_error:
+	if (node->gws)
+		amdgpu_amdkfd_free_gws(node->adev, node->gws);
+
+	/* Cleanup the node memory here */
+	kfree(node);
+	return err;
+}
+
+static void kfd_cleanup_nodes(struct kfd_dev *kfd, unsigned int num_nodes)
+{
+	struct kfd_node *knode;
+	unsigned int i;
+
+	/*
+	 * flush_work ensures that there are no outstanding
+	 * work-queue items that will access interrupt_ring. New work items
+	 * can't be created because we stopped interrupt handling above.
+	 */
+	flush_workqueue(kfd->ih_wq);
+	destroy_workqueue(kfd->ih_wq);
+
+	for (i = 0; i < num_nodes; i++) {
+		knode = kfd->nodes[i];
+		device_queue_manager_uninit(knode->dqm);
+		kfd_interrupt_exit(knode);
+		kfd_topology_remove_device(knode);
+		if (knode->gws)
+			amdgpu_amdkfd_free_gws(knode->adev, knode->gws);
+		kfree(knode);
+		kfd->nodes[i] = NULL;
+	}
+}
+
+static void kfd_setup_interrupt_bitmap(struct kfd_node *node,
+				       unsigned int kfd_node_idx)
+{
+	struct amdgpu_device *adev = node->adev;
+	uint32_t xcc_mask = node->xcc_mask;
+	uint32_t xcc, mapped_xcc;
+	/*
+	 * Interrupt bitmap is setup for processing interrupts from
+	 * different XCDs and AIDs.
+	 * Interrupt bitmap is defined as follows:
+	 * 1. Bits 0-15 - correspond to the NodeId field.
+	 *    Each bit corresponds to NodeId number. For example, if
+	 *    a KFD node has interrupt bitmap set to 0x7, then this
+	 *    KFD node will process interrupts with NodeId = 0, 1 and 2
+	 *    in the IH cookie.
+	 * 2. Bits 16-31 - unused.
+	 *
+	 * Please note that the kfd_node_idx argument passed to this
+	 * function is not related to NodeId field received in the
+	 * IH cookie.
+	 *
+	 * In CPX mode, a KFD node will process an interrupt if:
+	 * - the Node Id matches the corresponding bit set in
+	 *   Bits 0-15.
+	 * - AND VMID reported in the interrupt lies within the
+	 *   VMID range of the node.
+	 */
+	for_each_inst(xcc, xcc_mask) {
+		mapped_xcc = GET_INST(GC, xcc);
+		node->interrupt_bitmap |= (mapped_xcc % 2 ? 5 : 3) << (4 * (mapped_xcc / 2));
+	}
+	dev_info(kfd_device, "Node: %d, interrupt_bitmap: %x\n", kfd_node_idx,
+							node->interrupt_bitmap);
+}
+
 bool kgd2kfd_device_init(struct kfd_dev *kfd,
-			 struct drm_device *ddev,
 			 const struct kgd2kfd_shared_resources *gpu_resources)
 {
-	unsigned int size, map_process_packet_size;
-
-	kfd->ddev = ddev;
-	kfd->mec_fw_version = amdgpu_amdkfd_get_fw_version(kfd->kgd,
+	unsigned int size, map_process_packet_size, i;
+	struct kfd_node *node;
+	uint32_t first_vmid_kfd, last_vmid_kfd, vmid_num_kfd;
+	unsigned int max_proc_per_quantum;
+	int partition_mode;
+	int xcp_idx;
+
+	kfd->mec_fw_version = amdgpu_amdkfd_get_fw_version(kfd->adev,
 			KGD_ENGINE_MEC1);
-	kfd->mec2_fw_version = amdgpu_amdkfd_get_fw_version(kfd->kgd,
+	kfd->mec2_fw_version = amdgpu_amdkfd_get_fw_version(kfd->adev,
 			KGD_ENGINE_MEC2);
-	kfd->sdma_fw_version = amdgpu_amdkfd_get_fw_version(kfd->kgd,
+	kfd->sdma_fw_version = amdgpu_amdkfd_get_fw_version(kfd->adev,
 			KGD_ENGINE_SDMA1);
 	kfd->shared_resources = *gpu_resources;
 
-	kfd->vm_info.first_vmid_kfd = ffs(gpu_resources->compute_vmid_bitmap)-1;
-	kfd->vm_info.last_vmid_kfd = fls(gpu_resources->compute_vmid_bitmap)-1;
-	kfd->vm_info.vmid_num_kfd = kfd->vm_info.last_vmid_kfd
-			- kfd->vm_info.first_vmid_kfd + 1;
+	kfd->num_nodes = amdgpu_xcp_get_num_xcp(kfd->adev->xcp_mgr);
+
+	if (kfd->num_nodes == 0) {
+		dev_err(kfd_device,
+			"KFD num nodes cannot be 0, num_xcc_in_node: %d\n",
+			kfd->adev->gfx.num_xcc_per_xcp);
+		goto out;
+	}
 
 	/* Allow BIF to recode atomics to PCIe 3.0 AtomicOps.
 	 * 32 and 64-bit requests are possible and must be
 	 * supported.
 	 */
-	kfd->pci_atomic_requested = amdgpu_amdkfd_have_atomics_support(kfd->kgd);
+	kfd->pci_atomic_requested = amdgpu_amdkfd_have_atomics_support(kfd->adev);
 	if (!kfd->pci_atomic_requested &&
-	    kfd->device_info->needs_pci_atomics &&
-	    (!kfd->device_info->no_atomic_fw_version ||
-	     kfd->mec_fw_version < kfd->device_info->no_atomic_fw_version)) {
+	    kfd->device_info.needs_pci_atomics &&
+	    (!kfd->device_info.no_atomic_fw_version ||
+	     kfd->mec_fw_version < kfd->device_info.no_atomic_fw_version)) {
 		dev_info(kfd_device,
 			 "skipped device %x:%x, PCI rejects atomics %d<%d\n",
-			 kfd->pdev->vendor, kfd->pdev->device,
+			 kfd->adev->pdev->vendor, kfd->adev->pdev->device,
 			 kfd->mec_fw_version,
-			 kfd->device_info->no_atomic_fw_version);
+			 kfd->device_info.no_atomic_fw_version);
 		return false;
 	}
 
+	first_vmid_kfd = ffs(gpu_resources->compute_vmid_bitmap)-1;
+	last_vmid_kfd = fls(gpu_resources->compute_vmid_bitmap)-1;
+	vmid_num_kfd = last_vmid_kfd - first_vmid_kfd + 1;
+
+	/* For multi-partition capable GPUs, we need special handling for VMIDs
+	 * depending on partition mode.
+	 * In CPX mode, the VMID range needs to be shared between XCDs.
+	 * Additionally, there are 13 VMIDs (3-15) available for KFD. To
+	 * divide them equally, we change starting VMID to 4 and not use
+	 * VMID 3.
+	 * If the VMID range changes for multi-partition capable GPUs, then
+	 * this code MUST be revisited.
+	 */
+	if (kfd->adev->xcp_mgr) {
+		partition_mode = amdgpu_xcp_query_partition_mode(kfd->adev->xcp_mgr,
+								 AMDGPU_XCP_FL_LOCKED);
+		if (partition_mode == AMDGPU_CPX_PARTITION_MODE &&
+		    kfd->num_nodes != 1) {
+			vmid_num_kfd /= 2;
+			first_vmid_kfd = last_vmid_kfd + 1 - vmid_num_kfd*2;
+		}
+	}
+
 	/* Verify module parameters regarding mapped process number*/
-	if ((hws_max_conc_proc < 0)
-			|| (hws_max_conc_proc > kfd->vm_info.vmid_num_kfd)) {
-		dev_err(kfd_device,
-			"hws_max_conc_proc %d must be between 0 and %d, use %d instead\n",
-			hws_max_conc_proc, kfd->vm_info.vmid_num_kfd,
-			kfd->vm_info.vmid_num_kfd);
-		kfd->max_proc_per_quantum = kfd->vm_info.vmid_num_kfd;
-	} else
-		kfd->max_proc_per_quantum = hws_max_conc_proc;
+	if (hws_max_conc_proc >= 0)
+		max_proc_per_quantum = min((u32)hws_max_conc_proc, vmid_num_kfd);
+	else
+		max_proc_per_quantum = vmid_num_kfd;
 
 	/* calculate max size of mqds needed for queues */
 	size = max_num_of_queues_per_device *
-			kfd->device_info->mqd_size_aligned;
+			kfd->device_info.mqd_size_aligned;
 
 	/*
 	 * calculate max size of runlist packet.
 	 * There can be only 2 packets at once
 	 */
-	map_process_packet_size =
-			kfd->device_info->asic_family == CHIP_ALDEBARAN ?
+	map_process_packet_size = KFD_GC_VERSION(kfd) == IP_VERSION(9, 4, 2) ?
 				sizeof(struct pm4_mes_map_process_aldebaran) :
-					sizeof(struct pm4_mes_map_process);
+				sizeof(struct pm4_mes_map_process);
 	size += (KFD_MAX_NUM_OF_PROCESSES * map_process_packet_size +
 		max_num_of_queues_per_device * sizeof(struct pm4_mes_map_queues)
 		+ sizeof(struct pm4_mes_runlist)) * 2;
@@ -974,7 +806,7 @@ bool kgd2kfd_device_init(struct kfd_dev *kfd,
 	size += 512 * 1024;
 
 	if (amdgpu_amdkfd_alloc_gtt_mem(
-			kfd->kgd, size, &kfd->gtt_mem,
+			kfd->adev, size, &kfd->gtt_mem,
 			&kfd->gtt_start_gpu_addr, &kfd->gtt_start_cpu_ptr,
 			false)) {
 		dev_err(kfd_device, "Could not allocate %d bytes\n", size);
@@ -995,86 +827,118 @@ bool kgd2kfd_device_init(struct kfd_dev *kfd,
 		goto kfd_doorbell_error;
 	}
 
-	kfd->hive_id = amdgpu_amdkfd_get_hive_id(kfd->kgd);
-
-	kfd->noretry = amdgpu_amdkfd_get_noretry(kfd->kgd);
+	if (amdgpu_use_xgmi_p2p)
+		kfd->hive_id = kfd->adev->gmc.xgmi.hive_id;
 
-	if (kfd_interrupt_init(kfd)) {
-		dev_err(kfd_device, "Error initializing interrupts\n");
-		goto kfd_interrupt_error;
-	}
+	/*
+	 * For multi-partition capable GPUs, the KFD abstracts all partitions
+	 * within a socket as xGMI connected in the topology so assign a unique
+	 * hive id per device based on the pci device location if device is in
+	 * PCIe mode.
+	 */
+	if (!kfd->hive_id && kfd->num_nodes > 1)
+		kfd->hive_id = pci_dev_id(kfd->adev->pdev);
 
-	kfd->dqm = device_queue_manager_init(kfd);
-	if (!kfd->dqm) {
-		dev_err(kfd_device, "Error initializing queue manager\n");
-		goto device_queue_manager_error;
-	}
+	kfd->noretry = kfd->adev->gmc.noretry;
 
-	/* If supported on this device, allocate global GWS that is shared
-	 * by all KFD processes
-	 */
-	if (kfd_gws_init(kfd)) {
-		dev_err(kfd_device, "Could not allocate %d gws\n",
-			amdgpu_amdkfd_get_num_gws(kfd->kgd));
-		goto gws_error;
-	}
+	kfd_cwsr_init(kfd);
 
-	/* If CRAT is broken, won't set iommu enabled */
-	kfd_double_confirm_iommu_support(kfd);
+	dev_info(kfd_device, "Total number of KFD nodes to be created: %d\n",
+				kfd->num_nodes);
+
+	/* Allocate the KFD nodes */
+	for (i = 0, xcp_idx = 0; i < kfd->num_nodes; i++) {
+		node = kzalloc(sizeof(struct kfd_node), GFP_KERNEL);
+		if (!node)
+			goto node_alloc_error;
+
+		node->node_id = i;
+		node->adev = kfd->adev;
+		node->kfd = kfd;
+		node->kfd2kgd = kfd->kfd2kgd;
+		node->vm_info.vmid_num_kfd = vmid_num_kfd;
+		node->xcp = amdgpu_get_next_xcp(kfd->adev->xcp_mgr, &xcp_idx);
+		/* TODO : Check if error handling is needed */
+		if (node->xcp) {
+			amdgpu_xcp_get_inst_details(node->xcp, AMDGPU_XCP_GFX,
+						    &node->xcc_mask);
+			++xcp_idx;
+		} else {
+			node->xcc_mask =
+				(1U << NUM_XCC(kfd->adev->gfx.xcc_mask)) - 1;
+		}
 
-	if (kfd_iommu_device_init(kfd)) {
-		kfd->use_iommu_v2 = false;
-		dev_err(kfd_device, "Error initializing iommuv2\n");
-		goto device_iommu_error;
-	}
+		if (node->xcp) {
+			dev_info(kfd_device, "KFD node %d partition %d size %lldM\n",
+				node->node_id, node->xcp->mem_id,
+				KFD_XCP_MEMORY_SIZE(node->adev, node->node_id) >> 20);
+		}
 
-	kfd_cwsr_init(kfd);
+		if (partition_mode == AMDGPU_CPX_PARTITION_MODE &&
+		    kfd->num_nodes != 1) {
+			/* For multi-partition capable GPUs and CPX mode, first
+			 * XCD gets VMID range 4-9 and second XCD gets VMID
+			 * range 10-15.
+			 */
+
+			node->vm_info.first_vmid_kfd = (i%2 == 0) ?
+						first_vmid_kfd :
+						first_vmid_kfd+vmid_num_kfd;
+			node->vm_info.last_vmid_kfd = (i%2 == 0) ?
+						last_vmid_kfd-vmid_num_kfd :
+						last_vmid_kfd;
+			node->compute_vmid_bitmap =
+				((0x1 << (node->vm_info.last_vmid_kfd + 1)) - 1) -
+				((0x1 << (node->vm_info.first_vmid_kfd)) - 1);
+		} else {
+			node->vm_info.first_vmid_kfd = first_vmid_kfd;
+			node->vm_info.last_vmid_kfd = last_vmid_kfd;
+			node->compute_vmid_bitmap =
+				gpu_resources->compute_vmid_bitmap;
+		}
+		node->max_proc_per_quantum = max_proc_per_quantum;
+		atomic_set(&node->sram_ecc_flag, 0);
 
-	svm_migrate_init((struct amdgpu_device *)kfd->kgd);
+		amdgpu_amdkfd_get_local_mem_info(kfd->adev,
+					&node->local_mem_info, node->xcp);
 
-	if(kgd2kfd_resume_iommu(kfd))
-		goto device_iommu_error;
+		if (kfd->adev->xcp_mgr)
+			kfd_setup_interrupt_bitmap(node, i);
 
-	if (kfd_resume(kfd))
-		goto kfd_resume_error;
+		/* Initialize the KFD node */
+		if (kfd_init_node(node)) {
+			dev_err(kfd_device, "Error initializing KFD node\n");
+			goto node_init_error;
+		}
 
-	kfd->dbgmgr = NULL;
+		spin_lock_init(&node->watch_points_lock);
 
-	if (kfd_topology_add_device(kfd)) {
-		dev_err(kfd_device, "Error adding device to topology\n");
-		goto kfd_topology_add_device_error;
+		kfd->nodes[i] = node;
 	}
 
-	kfd_smi_init(kfd);
+	svm_range_set_max_pages(kfd->adev);
 
 	kfd->init_complete = true;
-	dev_info(kfd_device, "added device %x:%x\n", kfd->pdev->vendor,
-		 kfd->pdev->device);
+	dev_info(kfd_device, "added device %x:%x\n", kfd->adev->pdev->vendor,
+		 kfd->adev->pdev->device);
 
 	pr_debug("Starting kfd with the following scheduling policy %d\n",
-		kfd->dqm->sched_policy);
+		node->dqm->sched_policy);
 
 	goto out;
 
-kfd_topology_add_device_error:
-kfd_resume_error:
-device_iommu_error:
-gws_error:
-	device_queue_manager_uninit(kfd->dqm);
-device_queue_manager_error:
-	kfd_interrupt_exit(kfd);
-kfd_interrupt_error:
+node_init_error:
+node_alloc_error:
+	kfd_cleanup_nodes(kfd, i);
 	kfd_doorbell_fini(kfd);
 kfd_doorbell_error:
 	kfd_gtt_sa_fini(kfd);
 kfd_gtt_sa_init_error:
-	amdgpu_amdkfd_free_gtt_mem(kfd->kgd, kfd->gtt_mem);
+	amdgpu_amdkfd_free_gtt_mem(kfd->adev, &kfd->gtt_mem);
 alloc_gtt_mem_failure:
-	if (kfd->gws)
-		amdgpu_amdkfd_free_gws(kfd->kgd, kfd->gws);
 	dev_err(kfd_device,
 		"device %x:%x NOT added due to errors\n",
-		kfd->pdev->vendor, kfd->pdev->device);
+		kfd->adev->pdev->vendor, kfd->adev->pdev->device);
 out:
 	return kfd->init_complete;
 }
@@ -1082,32 +946,37 @@ out:
 void kgd2kfd_device_exit(struct kfd_dev *kfd)
 {
 	if (kfd->init_complete) {
-		device_queue_manager_uninit(kfd->dqm);
-		kfd_interrupt_exit(kfd);
-		kfd_topology_remove_device(kfd);
+		/* Cleanup KFD nodes */
+		kfd_cleanup_nodes(kfd, kfd->num_nodes);
+		/* Cleanup common/shared resources */
 		kfd_doorbell_fini(kfd);
 		ida_destroy(&kfd->doorbell_ida);
 		kfd_gtt_sa_fini(kfd);
-		amdgpu_amdkfd_free_gtt_mem(kfd->kgd, kfd->gtt_mem);
-		if (kfd->gws)
-			amdgpu_amdkfd_free_gws(kfd->kgd, kfd->gws);
+		amdgpu_amdkfd_free_gtt_mem(kfd->adev, &kfd->gtt_mem);
 	}
 
 	kfree(kfd);
 }
 
-int kgd2kfd_pre_reset(struct kfd_dev *kfd)
+int kgd2kfd_pre_reset(struct kfd_dev *kfd,
+		      struct amdgpu_reset_context *reset_context)
 {
+	struct kfd_node *node;
+	int i;
+
 	if (!kfd->init_complete)
 		return 0;
 
-	kfd_smi_event_update_gpu_reset(kfd, false);
+	for (i = 0; i < kfd->num_nodes; i++) {
+		node = kfd->nodes[i];
+		kfd_smi_event_update_gpu_reset(node, false, reset_context);
+	}
 
-	kfd->dqm->ops.pre_reset(kfd->dqm);
+	kgd2kfd_suspend(kfd, true);
 
-	kgd2kfd_suspend(kfd, false);
+	for (i = 0; i < kfd->num_nodes; i++)
+		kfd_signal_reset_event(kfd->nodes[i]);
 
-	kfd_signal_reset_event(kfd);
 	return 0;
 }
 
@@ -1120,133 +989,177 @@ int kgd2kfd_pre_reset(struct kfd_dev *kfd)
 int kgd2kfd_post_reset(struct kfd_dev *kfd)
 {
 	int ret;
+	struct kfd_node *node;
+	int i;
 
 	if (!kfd->init_complete)
 		return 0;
 
-	ret = kfd_resume(kfd);
-	if (ret)
-		return ret;
-	atomic_dec(&kfd_locked);
+	for (i = 0; i < kfd->num_nodes; i++) {
+		ret = kfd_resume(kfd->nodes[i]);
+		if (ret)
+			return ret;
+	}
 
-	atomic_set(&kfd->sram_ecc_flag, 0);
+	mutex_lock(&kfd_processes_mutex);
+	--kfd_locked;
+	mutex_unlock(&kfd_processes_mutex);
 
-	kfd_smi_event_update_gpu_reset(kfd, true);
+	for (i = 0; i < kfd->num_nodes; i++) {
+		node = kfd->nodes[i];
+		atomic_set(&node->sram_ecc_flag, 0);
+		kfd_smi_event_update_gpu_reset(node, true, NULL);
+	}
 
 	return 0;
 }
 
-bool kfd_is_locked(void)
+bool kfd_is_locked(struct kfd_dev *kfd)
 {
-	return  (atomic_read(&kfd_locked) > 0);
+	uint8_t id  = 0;
+	struct kfd_node *dev;
+
+	lockdep_assert_held(&kfd_processes_mutex);
+
+	/* check reset/suspend lock */
+	if (kfd_locked > 0)
+		return true;
+
+	if (kfd)
+		return kfd->kfd_dev_lock > 0;
+
+	/* check lock on all cgroup accessible devices */
+	while (kfd_topology_enum_kfd_devices(id++, &dev) == 0) {
+		if (!dev || kfd_devcgroup_check_permission(dev))
+			continue;
+
+		if (dev->kfd->kfd_dev_lock > 0)
+			return true;
+	}
+
+	return false;
 }
 
-void kgd2kfd_suspend(struct kfd_dev *kfd, bool run_pm)
+void kgd2kfd_suspend(struct kfd_dev *kfd, bool suspend_proc)
 {
+	struct kfd_node *node;
+	int i;
+
 	if (!kfd->init_complete)
 		return;
 
-	/* for runtime suspend, skip locking kfd */
-	if (!run_pm) {
-		/* For first KFD device suspend all the KFD processes */
-		if (atomic_inc_return(&kfd_locked) == 1)
-			kfd_suspend_all_processes();
-	}
+	if (suspend_proc)
+		kgd2kfd_suspend_process(kfd);
 
-	kfd->dqm->ops.stop(kfd->dqm);
-	kfd_iommu_suspend(kfd);
+	for (i = 0; i < kfd->num_nodes; i++) {
+		node = kfd->nodes[i];
+		node->dqm->ops.stop(node->dqm);
+	}
 }
 
-int kgd2kfd_resume(struct kfd_dev *kfd, bool run_pm)
+int kgd2kfd_resume(struct kfd_dev *kfd, bool resume_proc)
 {
-	int ret, count;
+	int ret, i;
 
 	if (!kfd->init_complete)
 		return 0;
 
-	ret = kfd_resume(kfd);
-	if (ret)
-		return ret;
-
-	/* for runtime resume, skip unlocking kfd */
-	if (!run_pm) {
-		count = atomic_dec_return(&kfd_locked);
-		WARN_ONCE(count < 0, "KFD suspend / resume ref. error");
-		if (count == 0)
-			ret = kfd_resume_all_processes();
+	for (i = 0; i < kfd->num_nodes; i++) {
+		ret = kfd_resume(kfd->nodes[i]);
+		if (ret)
+			return ret;
 	}
 
+	if (resume_proc)
+		ret = kgd2kfd_resume_process(kfd);
+
 	return ret;
 }
 
-int kgd2kfd_resume_iommu(struct kfd_dev *kfd)
+void kgd2kfd_suspend_process(struct kfd_dev *kfd)
 {
-	int err = 0;
+	if (!kfd->init_complete)
+		return;
 
-	err = kfd_iommu_resume(kfd);
-	if (err)
-		dev_err(kfd_device,
-			"Failed to resume IOMMU for device %x:%x\n",
-			kfd->pdev->vendor, kfd->pdev->device);
-	return err;
+	mutex_lock(&kfd_processes_mutex);
+	/* For first KFD device suspend all the KFD processes */
+	if (++kfd_locked == 1)
+		kfd_suspend_all_processes();
+	mutex_unlock(&kfd_processes_mutex);
 }
 
-static int kfd_resume(struct kfd_dev *kfd)
+int kgd2kfd_resume_process(struct kfd_dev *kfd)
 {
-	int err = 0;
+	int ret = 0;
 
-	err = kfd->dqm->ops.start(kfd->dqm);
-	if (err)
-		dev_err(kfd_device,
-			"Error starting queue manager for device %x:%x\n",
-			kfd->pdev->vendor, kfd->pdev->device);
+	if (!kfd->init_complete)
+		return 0;
 
-	return err;
+	mutex_lock(&kfd_processes_mutex);
+	if (--kfd_locked == 0)
+		ret = kfd_resume_all_processes();
+	WARN_ONCE(kfd_locked < 0, "KFD suspend / resume ref. error");
+	mutex_unlock(&kfd_processes_mutex);
+
+	return ret;
 }
 
-static inline void kfd_queue_work(struct workqueue_struct *wq,
-				  struct work_struct *work)
+static int kfd_resume(struct kfd_node *node)
 {
-	int cpu, new_cpu;
+	int err = 0;
 
-	cpu = new_cpu = smp_processor_id();
-	do {
-		new_cpu = cpumask_next(new_cpu, cpu_online_mask) % nr_cpu_ids;
-		if (cpu_to_node(new_cpu) == numa_node_id())
-			break;
-	} while (cpu != new_cpu);
+	err = node->dqm->ops.start(node->dqm);
+	if (err)
+		dev_err(kfd_device,
+			"Error starting queue manager for device %x:%x\n",
+			node->adev->pdev->vendor, node->adev->pdev->device);
 
-	queue_work_on(new_cpu, wq, work);
+	return err;
 }
 
 /* This is called directly from KGD at ISR. */
 void kgd2kfd_interrupt(struct kfd_dev *kfd, const void *ih_ring_entry)
 {
-	uint32_t patched_ihre[KFD_MAX_RING_ENTRY_SIZE];
+	uint32_t patched_ihre[KFD_MAX_RING_ENTRY_SIZE], i;
 	bool is_patched = false;
 	unsigned long flags;
+	struct kfd_node *node;
 
 	if (!kfd->init_complete)
 		return;
 
-	if (kfd->device_info->ih_ring_entry_size > sizeof(patched_ihre)) {
+	if (kfd->device_info.ih_ring_entry_size > sizeof(patched_ihre)) {
 		dev_err_once(kfd_device, "Ring entry too small\n");
 		return;
 	}
 
-	spin_lock_irqsave(&kfd->interrupt_lock, flags);
-
-	if (kfd->interrupts_active
-	    && interrupt_is_wanted(kfd, ih_ring_entry,
-				   patched_ihre, &is_patched)
-	    && enqueue_ih_ring_entry(kfd,
-				     is_patched ? patched_ihre : ih_ring_entry))
-		kfd_queue_work(kfd->ih_wq, &kfd->interrupt_work);
+	for (i = 0; i < kfd->num_nodes; i++) {
+		/* Race if another thread in b/w
+		 * kfd_cleanup_nodes and kfree(kfd),
+		 * when kfd->nodes[i] = NULL
+		 */
+		if (kfd->nodes[i])
+			node = kfd->nodes[i];
+		else
+			return;
+
+		spin_lock_irqsave(&node->interrupt_lock, flags);
+
+		if (node->interrupts_active
+		    && interrupt_is_wanted(node, ih_ring_entry,
+			    	patched_ihre, &is_patched)
+		    && enqueue_ih_ring_entry(node,
+			    	is_patched ? patched_ihre : ih_ring_entry)) {
+			queue_work(node->kfd->ih_wq, &node->interrupt_work);
+			spin_unlock_irqrestore(&node->interrupt_lock, flags);
+			return;
+		}
+		spin_unlock_irqrestore(&node->interrupt_lock, flags);
+	}
 
-	spin_unlock_irqrestore(&kfd->interrupt_lock, flags);
 }
 
-int kgd2kfd_quiesce_mm(struct mm_struct *mm)
+int kgd2kfd_quiesce_mm(struct mm_struct *mm, uint32_t trigger)
 {
 	struct kfd_process *p;
 	int r;
@@ -1260,7 +1173,7 @@ int kgd2kfd_quiesce_mm(struct mm_struct *mm)
 		return -ESRCH;
 
 	WARN(debug_evictions, "Evicting pid %d", p->lead_thread->pid);
-	r = kfd_process_evict_queues(p);
+	r = kfd_process_evict_queues(p, trigger);
 
 	kfd_unref_process(p);
 	return r;
@@ -1338,8 +1251,6 @@ out:
 static int kfd_gtt_sa_init(struct kfd_dev *kfd, unsigned int buf_size,
 				unsigned int chunk_size)
 {
-	unsigned int num_of_longs;
-
 	if (WARN_ON(buf_size < chunk_size))
 		return -EINVAL;
 	if (WARN_ON(buf_size == 0))
@@ -1350,11 +1261,8 @@ static int kfd_gtt_sa_init(struct kfd_dev *kfd, unsigned int buf_size,
 	kfd->gtt_sa_chunk_size = chunk_size;
 	kfd->gtt_sa_num_of_chunks = buf_size / chunk_size;
 
-	num_of_longs = (kfd->gtt_sa_num_of_chunks + BITS_PER_LONG - 1) /
-		BITS_PER_LONG;
-
-	kfd->gtt_sa_bitmap = kcalloc(num_of_longs, sizeof(long), GFP_KERNEL);
-
+	kfd->gtt_sa_bitmap = bitmap_zalloc(kfd->gtt_sa_num_of_chunks,
+					   GFP_KERNEL);
 	if (!kfd->gtt_sa_bitmap)
 		return -ENOMEM;
 
@@ -1364,13 +1272,12 @@ static int kfd_gtt_sa_init(struct kfd_dev *kfd, unsigned int buf_size,
 	mutex_init(&kfd->gtt_sa_lock);
 
 	return 0;
-
 }
 
 static void kfd_gtt_sa_fini(struct kfd_dev *kfd)
 {
 	mutex_destroy(&kfd->gtt_sa_lock);
-	kfree(kfd->gtt_sa_bitmap);
+	bitmap_free(kfd->gtt_sa_bitmap);
 }
 
 static inline uint64_t kfd_gtt_sa_calc_gpu_addr(uint64_t start_addr,
@@ -1387,10 +1294,11 @@ static inline uint32_t *kfd_gtt_sa_calc_cpu_addr(void *start_addr,
 	return (uint32_t *) ((uint64_t) start_addr + bit_num * chunk_size);
 }
 
-int kfd_gtt_sa_allocate(struct kfd_dev *kfd, unsigned int size,
+int kfd_gtt_sa_allocate(struct kfd_node *node, unsigned int size,
 			struct kfd_mem_obj **mem_obj)
 {
 	unsigned int found, start_search, cur_size;
+	struct kfd_dev *kfd = node->kfd;
 
 	if (size == 0)
 		return -EINVAL;
@@ -1438,7 +1346,7 @@ kfd_gtt_restart_search:
 	/* If we need only one chunk, mark it as allocated and get out */
 	if (size <= kfd->gtt_sa_chunk_size) {
 		pr_debug("Single bit\n");
-		set_bit(found, kfd->gtt_sa_bitmap);
+		__set_bit(found, kfd->gtt_sa_bitmap);
 		goto kfd_gtt_out;
 	}
 
@@ -1476,10 +1384,8 @@ kfd_gtt_restart_search:
 		(*mem_obj)->range_start, (*mem_obj)->range_end);
 
 	/* Mark the chunks as allocated */
-	for (found = (*mem_obj)->range_start;
-		found <= (*mem_obj)->range_end;
-		found++)
-		set_bit(found, kfd->gtt_sa_bitmap);
+	bitmap_set(kfd->gtt_sa_bitmap, (*mem_obj)->range_start,
+		   (*mem_obj)->range_end - (*mem_obj)->range_start + 1);
 
 kfd_gtt_out:
 	mutex_unlock(&kfd->gtt_sa_lock);
@@ -1492,9 +1398,9 @@ kfd_gtt_no_free_chunk:
 	return -ENOMEM;
 }
 
-int kfd_gtt_sa_free(struct kfd_dev *kfd, struct kfd_mem_obj *mem_obj)
+int kfd_gtt_sa_free(struct kfd_node *node, struct kfd_mem_obj *mem_obj)
 {
-	unsigned int bit;
+	struct kfd_dev *kfd = node->kfd;
 
 	/* Act like kfree when trying to free a NULL object */
 	if (!mem_obj)
@@ -1506,10 +1412,8 @@ int kfd_gtt_sa_free(struct kfd_dev *kfd, struct kfd_mem_obj *mem_obj)
 	mutex_lock(&kfd->gtt_sa_lock);
 
 	/* Mark the chunks as free */
-	for (bit = mem_obj->range_start;
-		bit <= mem_obj->range_end;
-		bit++)
-		clear_bit(bit, kfd->gtt_sa_bitmap);
+	bitmap_clear(kfd->gtt_sa_bitmap, mem_obj->range_start,
+		     mem_obj->range_end - mem_obj->range_start + 1);
 
 	mutex_unlock(&kfd->gtt_sa_lock);
 
@@ -1519,29 +1423,287 @@ int kfd_gtt_sa_free(struct kfd_dev *kfd, struct kfd_mem_obj *mem_obj)
 
 void kgd2kfd_set_sram_ecc_flag(struct kfd_dev *kfd)
 {
+	/*
+	 * TODO: Currently update SRAM ECC flag for first node.
+	 * This needs to be updated later when we can
+	 * identify SRAM ECC error on other nodes also.
+	 */
 	if (kfd)
-		atomic_inc(&kfd->sram_ecc_flag);
+		atomic_inc(&kfd->nodes[0]->sram_ecc_flag);
 }
 
-void kfd_inc_compute_active(struct kfd_dev *kfd)
+void kfd_inc_compute_active(struct kfd_node *node)
 {
-	if (atomic_inc_return(&kfd->compute_profile) == 1)
-		amdgpu_amdkfd_set_compute_idle(kfd->kgd, false);
+	if (atomic_inc_return(&node->kfd->compute_profile) == 1)
+		amdgpu_amdkfd_set_compute_idle(node->adev, false);
 }
 
-void kfd_dec_compute_active(struct kfd_dev *kfd)
+void kfd_dec_compute_active(struct kfd_node *node)
 {
-	int count = atomic_dec_return(&kfd->compute_profile);
+	int count = atomic_dec_return(&node->kfd->compute_profile);
 
 	if (count == 0)
-		amdgpu_amdkfd_set_compute_idle(kfd->kgd, true);
+		amdgpu_amdkfd_set_compute_idle(node->adev, true);
 	WARN_ONCE(count < 0, "Compute profile ref. count error");
 }
 
+static bool kfd_compute_active(struct kfd_node *node)
+{
+	if (atomic_read(&node->kfd->compute_profile))
+		return true;
+	return false;
+}
+
 void kgd2kfd_smi_event_throttle(struct kfd_dev *kfd, uint64_t throttle_bitmask)
 {
+	/*
+	 * TODO: For now, raise the throttling event only on first node.
+	 * This will need to change after we are able to determine
+	 * which node raised the throttling event.
+	 */
 	if (kfd && kfd->init_complete)
-		kfd_smi_event_update_thermal_throttling(kfd, throttle_bitmask);
+		kfd_smi_event_update_thermal_throttling(kfd->nodes[0],
+							throttle_bitmask);
+}
+
+/* kfd_get_num_sdma_engines returns the number of PCIe optimized SDMA and
+ * kfd_get_num_xgmi_sdma_engines returns the number of XGMI SDMA.
+ * When the device has more than two engines, we reserve two for PCIe to enable
+ * full-duplex and the rest are used as XGMI.
+ */
+unsigned int kfd_get_num_sdma_engines(struct kfd_node *node)
+{
+	/* If XGMI is not supported, all SDMA engines are PCIe */
+	if (!node->adev->gmc.xgmi.supported)
+		return node->adev->sdma.num_instances/(int)node->kfd->num_nodes;
+
+	return min(node->adev->sdma.num_instances/(int)node->kfd->num_nodes, 2);
+}
+
+unsigned int kfd_get_num_xgmi_sdma_engines(struct kfd_node *node)
+{
+	/* After reserved for PCIe, the rest of engines are XGMI */
+	return node->adev->sdma.num_instances/(int)node->kfd->num_nodes -
+		kfd_get_num_sdma_engines(node);
+}
+
+int kgd2kfd_check_and_lock_kfd(struct kfd_dev *kfd)
+{
+	struct kfd_process *p;
+	int r = 0, temp, idx;
+
+	mutex_lock(&kfd_processes_mutex);
+
+	/* kfd_processes_count is per kfd_dev, return -EBUSY without
+	 * further check
+	 */
+	if (!!atomic_read(&kfd->kfd_processes_count)) {
+		pr_debug("process_wq_release not finished\n");
+		r = -EBUSY;
+		goto out;
+	}
+
+	if (hash_empty(kfd_processes_table) && !kfd_is_locked(kfd))
+		goto out;
+
+	/* fail under system reset/resume or kfd device is partition switching. */
+	if (kfd_is_locked(kfd)) {
+		r = -EBUSY;
+		goto out;
+	}
+
+	/*
+	 * ensure all running processes are cgroup excluded from device before mode switch.
+	 * i.e. no pdd was created on the process socket.
+	 */
+	idx = srcu_read_lock(&kfd_processes_srcu);
+	hash_for_each_rcu(kfd_processes_table, temp, p, kfd_processes) {
+		int i;
+
+		for (i = 0; i < p->n_pdds; i++) {
+			if (p->pdds[i]->dev->kfd != kfd)
+				continue;
+
+			r = -EBUSY;
+			goto proc_check_unlock;
+		}
+	}
+
+proc_check_unlock:
+	srcu_read_unlock(&kfd_processes_srcu, idx);
+out:
+	if (!r)
+		++kfd->kfd_dev_lock;
+	mutex_unlock(&kfd_processes_mutex);
+
+	return r;
+}
+
+void kgd2kfd_unlock_kfd(struct kfd_dev *kfd)
+{
+	mutex_lock(&kfd_processes_mutex);
+	--kfd->kfd_dev_lock;
+	mutex_unlock(&kfd_processes_mutex);
+}
+
+int kgd2kfd_start_sched(struct kfd_dev *kfd, uint32_t node_id)
+{
+	struct kfd_node *node;
+	int ret;
+
+	if (!kfd->init_complete)
+		return 0;
+
+	if (node_id >= kfd->num_nodes) {
+		dev_warn(kfd->adev->dev, "Invalid node ID: %u exceeds %u\n",
+			 node_id, kfd->num_nodes - 1);
+		return -EINVAL;
+	}
+	node = kfd->nodes[node_id];
+
+	ret = node->dqm->ops.unhalt(node->dqm);
+	if (ret)
+		dev_err(kfd_device, "Error in starting scheduler\n");
+
+	return ret;
+}
+
+int kgd2kfd_start_sched_all_nodes(struct kfd_dev *kfd)
+{
+	struct kfd_node *node;
+	int i, r;
+
+	if (!kfd->init_complete)
+		return 0;
+
+	for (i = 0; i < kfd->num_nodes; i++) {
+		node = kfd->nodes[i];
+		r = node->dqm->ops.unhalt(node->dqm);
+		if (r) {
+			dev_err(kfd_device, "Error in starting scheduler\n");
+			return r;
+		}
+	}
+	return 0;
+}
+
+int kgd2kfd_stop_sched(struct kfd_dev *kfd, uint32_t node_id)
+{
+	struct kfd_node *node;
+
+	if (!kfd->init_complete)
+		return 0;
+
+	if (node_id >= kfd->num_nodes) {
+		dev_warn(kfd->adev->dev, "Invalid node ID: %u exceeds %u\n",
+			 node_id, kfd->num_nodes - 1);
+		return -EINVAL;
+	}
+
+	node = kfd->nodes[node_id];
+	return node->dqm->ops.halt(node->dqm);
+}
+
+int kgd2kfd_stop_sched_all_nodes(struct kfd_dev *kfd)
+{
+	struct kfd_node *node;
+	int i, r;
+
+	if (!kfd->init_complete)
+		return 0;
+
+	for (i = 0; i < kfd->num_nodes; i++) {
+		node = kfd->nodes[i];
+		r = node->dqm->ops.halt(node->dqm);
+		if (r)
+			return r;
+	}
+	return 0;
+}
+
+bool kgd2kfd_compute_active(struct kfd_dev *kfd, uint32_t node_id)
+{
+	struct kfd_node *node;
+
+	if (!kfd->init_complete)
+		return false;
+
+	if (node_id >= kfd->num_nodes) {
+		dev_warn(kfd->adev->dev, "Invalid node ID: %u exceeds %u\n",
+			 node_id, kfd->num_nodes - 1);
+		return false;
+	}
+
+	node = kfd->nodes[node_id];
+
+	return kfd_compute_active(node);
+}
+
+/**
+ * kgd2kfd_vmfault_fast_path() - KFD vm page fault interrupt handling fast path for gmc v9
+ * @adev: amdgpu device
+ * @entry: vm fault interrupt vector
+ * @retry_fault: if this is retry fault
+ *
+ * retry fault -
+ *    with CAM enabled, adev primary ring
+ *                           |  gmc_v9_0_process_interrupt()
+ *                      adev soft_ring
+ *                           |  gmc_v9_0_process_interrupt() worker failed to recover page fault
+ *                      KFD node ih_fifo
+ *                           |  KFD interrupt_wq worker
+ *                      kfd_signal_vm_fault_event
+ *
+ *    without CAM,      adev primary ring1
+ *                           |  gmc_v9_0_process_interrupt worker failed to recvoer page fault
+ *                      KFD node ih_fifo
+ *                           |  KFD interrupt_wq worker
+ *                      kfd_signal_vm_fault_event
+ *
+ * no-retry fault -
+ *                      adev primary ring
+ *                           |  gmc_v9_0_process_interrupt()
+ *                      KFD node ih_fifo
+ *                           |  KFD interrupt_wq worker
+ *                      kfd_signal_vm_fault_event
+ *
+ * fast path - After kfd_signal_vm_fault_event, gmc_v9_0_process_interrupt drop the page fault
+ *            of same process, don't copy interrupt to KFD node ih_fifo.
+ *            With gdb debugger enabled, need convert the retry fault to no-retry fault for
+ *            debugger, cannot use the fast path.
+ *
+ * Return:
+ *   true - use the fast path to handle this fault
+ *   false - use normal path to handle it
+ */
+bool kgd2kfd_vmfault_fast_path(struct amdgpu_device *adev, struct amdgpu_iv_entry *entry,
+			       bool retry_fault)
+{
+	struct kfd_process *p;
+	u32 cam_index;
+
+	if (entry->ih == &adev->irq.ih_soft || entry->ih == &adev->irq.ih1) {
+		p = kfd_lookup_process_by_pasid(entry->pasid, NULL);
+		if (!p)
+			return true;
+
+		if (p->gpu_page_fault && !p->debug_trap_enabled) {
+			if (retry_fault && adev->irq.retry_cam_enabled) {
+				cam_index = entry->src_data[2] & 0x3ff;
+				WDOORBELL32(adev->irq.retry_cam_doorbell_index, cam_index);
+			}
+
+			kfd_unref_process(p);
+			return true;
+		}
+
+		/*
+		 * This is the first page fault, set flag and then signal user space
+		 */
+		p->gpu_page_fault = true;
+		kfd_unref_process(p);
+	}
+	return false;
 }
 
 #if defined(CONFIG_DEBUG_FS)
@@ -1549,13 +1711,18 @@ void kgd2kfd_smi_event_throttle(struct kfd_dev *kfd, uint64_t throttle_bitmask)
 /* This function will send a package to HIQ to hang the HWS
  * which will trigger a GPU reset and bring the HWS back to normal state
  */
-int kfd_debugfs_hang_hws(struct kfd_dev *dev)
+int kfd_debugfs_hang_hws(struct kfd_node *dev)
 {
 	if (dev->dqm->sched_policy != KFD_SCHED_POLICY_HWS) {
 		pr_err("HWS is not enabled");
 		return -EINVAL;
 	}
 
+	if (dev->kfd->shared_resources.enable_mes) {
+		dev_err(dev->adev->dev, "Inducing MES hang is not supported\n");
+		return -EINVAL;
+	}
+
 	return dqm_debugfs_hang_hws(dev->dqm);
 }
 
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c
index 93e33dd84dd4..d7a2e7178ea9 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c
@@ -1,5 +1,6 @@
+// SPDX-License-Identifier: GPL-2.0 OR MIT
 /*
- * Copyright 2014 Advanced Micro Devices, Inc.
+ * Copyright 2014-2022 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -34,20 +35,29 @@
 #include "cik_regs.h"
 #include "kfd_kernel_queue.h"
 #include "amdgpu_amdkfd.h"
+#include "amdgpu_reset.h"
+#include "amdgpu_sdma.h"
+#include "mes_v11_api_def.h"
+#include "kfd_debug.h"
 
 /* Size of the per-pipe EOP queue */
 #define CIK_HPD_EOP_BYTES_LOG2 11
 #define CIK_HPD_EOP_BYTES (1U << CIK_HPD_EOP_BYTES_LOG2)
+/* See unmap_queues_cpsch() */
+#define USE_DEFAULT_GRACE_PERIOD 0xffffffff
 
 static int set_pasid_vmid_mapping(struct device_queue_manager *dqm,
 				  u32 pasid, unsigned int vmid);
 
 static int execute_queues_cpsch(struct device_queue_manager *dqm,
 				enum kfd_unmap_queues_filter filter,
-				uint32_t filter_param);
+				uint32_t filter_param,
+				uint32_t grace_period);
 static int unmap_queues_cpsch(struct device_queue_manager *dqm,
 				enum kfd_unmap_queues_filter filter,
-				uint32_t filter_param);
+				uint32_t filter_param,
+				uint32_t grace_period,
+				bool reset);
 
 static int map_queues_cpsch(struct device_queue_manager *dqm);
 
@@ -58,8 +68,9 @@ static inline void deallocate_hqd(struct device_queue_manager *dqm,
 				struct queue *q);
 static int allocate_hqd(struct device_queue_manager *dqm, struct queue *q);
 static int allocate_sdma_queue(struct device_queue_manager *dqm,
-				struct queue *q);
-static void kfd_process_hw_exception(struct work_struct *work);
+				struct queue *q, const uint32_t *restore_sdma_id);
+
+static int reset_queues_on_hws_hang(struct device_queue_manager *dqm, bool is_sdma);
 
 static inline
 enum KFD_MQD_TYPE get_mqd_type_from_queue_type(enum kfd_queue_type type)
@@ -72,95 +83,365 @@ enum KFD_MQD_TYPE get_mqd_type_from_queue_type(enum kfd_queue_type type)
 static bool is_pipe_enabled(struct device_queue_manager *dqm, int mec, int pipe)
 {
 	int i;
-	int pipe_offset = (mec * dqm->dev->shared_resources.num_pipe_per_mec
-		+ pipe) * dqm->dev->shared_resources.num_queue_per_pipe;
+	int pipe_offset = (mec * dqm->dev->kfd->shared_resources.num_pipe_per_mec
+		+ pipe) * dqm->dev->kfd->shared_resources.num_queue_per_pipe;
 
 	/* queue is available for KFD usage if bit is 1 */
-	for (i = 0; i <  dqm->dev->shared_resources.num_queue_per_pipe; ++i)
+	for (i = 0; i <  dqm->dev->kfd->shared_resources.num_queue_per_pipe; ++i)
 		if (test_bit(pipe_offset + i,
-			      dqm->dev->shared_resources.cp_queue_bitmap))
+			      dqm->dev->kfd->shared_resources.cp_queue_bitmap))
 			return true;
 	return false;
 }
 
 unsigned int get_cp_queues_num(struct device_queue_manager *dqm)
 {
-	return bitmap_weight(dqm->dev->shared_resources.cp_queue_bitmap,
-				KGD_MAX_QUEUES);
+	return bitmap_weight(dqm->dev->kfd->shared_resources.cp_queue_bitmap,
+				AMDGPU_MAX_QUEUES);
 }
 
 unsigned int get_queues_per_pipe(struct device_queue_manager *dqm)
 {
-	return dqm->dev->shared_resources.num_queue_per_pipe;
+	return dqm->dev->kfd->shared_resources.num_queue_per_pipe;
 }
 
 unsigned int get_pipes_per_mec(struct device_queue_manager *dqm)
 {
-	return dqm->dev->shared_resources.num_pipe_per_mec;
+	return dqm->dev->kfd->shared_resources.num_pipe_per_mec;
 }
 
-static unsigned int get_num_sdma_engines(struct device_queue_manager *dqm)
+static unsigned int get_num_all_sdma_engines(struct device_queue_manager *dqm)
 {
-	return dqm->dev->device_info->num_sdma_engines;
+	return kfd_get_num_sdma_engines(dqm->dev) +
+		kfd_get_num_xgmi_sdma_engines(dqm->dev);
 }
 
-static unsigned int get_num_xgmi_sdma_engines(struct device_queue_manager *dqm)
+unsigned int get_num_sdma_queues(struct device_queue_manager *dqm)
 {
-	return dqm->dev->device_info->num_xgmi_sdma_engines;
+	return kfd_get_num_sdma_engines(dqm->dev) *
+		dqm->dev->kfd->device_info.num_sdma_queues_per_engine;
 }
 
-static unsigned int get_num_all_sdma_engines(struct device_queue_manager *dqm)
+unsigned int get_num_xgmi_sdma_queues(struct device_queue_manager *dqm)
 {
-	return get_num_sdma_engines(dqm) + get_num_xgmi_sdma_engines(dqm);
+	return kfd_get_num_xgmi_sdma_engines(dqm->dev) *
+		dqm->dev->kfd->device_info.num_sdma_queues_per_engine;
 }
 
-unsigned int get_num_sdma_queues(struct device_queue_manager *dqm)
+static void init_sdma_bitmaps(struct device_queue_manager *dqm)
 {
-	return dqm->dev->device_info->num_sdma_engines
-			* dqm->dev->device_info->num_sdma_queues_per_engine;
-}
+	bitmap_zero(dqm->sdma_bitmap, KFD_MAX_SDMA_QUEUES);
+	bitmap_set(dqm->sdma_bitmap, 0, get_num_sdma_queues(dqm));
 
-unsigned int get_num_xgmi_sdma_queues(struct device_queue_manager *dqm)
-{
-	return dqm->dev->device_info->num_xgmi_sdma_engines
-			* dqm->dev->device_info->num_sdma_queues_per_engine;
+	bitmap_zero(dqm->xgmi_sdma_bitmap, KFD_MAX_SDMA_QUEUES);
+	bitmap_set(dqm->xgmi_sdma_bitmap, 0, get_num_xgmi_sdma_queues(dqm));
+
+	/* Mask out the reserved queues */
+	bitmap_andnot(dqm->sdma_bitmap, dqm->sdma_bitmap,
+		      dqm->dev->kfd->device_info.reserved_sdma_queues_bitmap,
+		      KFD_MAX_SDMA_QUEUES);
 }
 
 void program_sh_mem_settings(struct device_queue_manager *dqm,
 					struct qcm_process_device *qpd)
 {
-	return dqm->dev->kfd2kgd->program_sh_mem_settings(
-						dqm->dev->kgd, qpd->vmid,
-						qpd->sh_mem_config,
-						qpd->sh_mem_ape1_base,
-						qpd->sh_mem_ape1_limit,
-						qpd->sh_mem_bases);
+	uint32_t xcc_mask = dqm->dev->xcc_mask;
+	int xcc_id;
+
+	for_each_inst(xcc_id, xcc_mask)
+		dqm->dev->kfd2kgd->program_sh_mem_settings(
+			dqm->dev->adev, qpd->vmid, qpd->sh_mem_config,
+			qpd->sh_mem_ape1_base, qpd->sh_mem_ape1_limit,
+			qpd->sh_mem_bases, xcc_id);
+}
+
+static void kfd_hws_hang(struct device_queue_manager *dqm)
+{
+	struct device_process_node *cur;
+	struct qcm_process_device *qpd;
+	struct queue *q;
+
+	/* Mark all device queues as reset. */
+	list_for_each_entry(cur, &dqm->queues, list) {
+		qpd = cur->qpd;
+		list_for_each_entry(q, &qpd->queues_list, list) {
+			struct kfd_process_device *pdd = qpd_to_pdd(qpd);
+
+			pdd->has_reset_queue = true;
+		}
+	}
+
+	/*
+	 * Issue a GPU reset if HWS is unresponsive
+	 */
+	amdgpu_amdkfd_gpu_reset(dqm->dev->adev);
+}
+
+static int convert_to_mes_queue_type(int queue_type)
+{
+	int mes_queue_type;
+
+	switch (queue_type) {
+	case KFD_QUEUE_TYPE_COMPUTE:
+		mes_queue_type = MES_QUEUE_TYPE_COMPUTE;
+		break;
+	case KFD_QUEUE_TYPE_SDMA:
+		mes_queue_type = MES_QUEUE_TYPE_SDMA;
+		break;
+	default:
+		WARN(1, "Invalid queue type %d", queue_type);
+		mes_queue_type = -EINVAL;
+		break;
+	}
+
+	return mes_queue_type;
+}
+
+static int add_queue_mes(struct device_queue_manager *dqm, struct queue *q,
+			 struct qcm_process_device *qpd)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)dqm->dev->adev;
+	struct kfd_process_device *pdd = qpd_to_pdd(qpd);
+	struct mes_add_queue_input queue_input;
+	int r, queue_type;
+	uint64_t wptr_addr_off;
+
+	if (!dqm->sched_running || dqm->sched_halt)
+		return 0;
+	if (!down_read_trylock(&adev->reset_domain->sem))
+		return -EIO;
+
+	memset(&queue_input, 0x0, sizeof(struct mes_add_queue_input));
+	queue_input.process_id = pdd->pasid;
+	queue_input.page_table_base_addr =  qpd->page_table_base;
+	queue_input.process_va_start = 0;
+	queue_input.process_va_end = adev->vm_manager.max_pfn - 1;
+	/* MES unit for quantum is 100ns */
+	queue_input.process_quantum = KFD_MES_PROCESS_QUANTUM;  /* Equivalent to 10ms. */
+	queue_input.process_context_addr = pdd->proc_ctx_gpu_addr;
+	queue_input.gang_quantum = KFD_MES_GANG_QUANTUM; /* Equivalent to 1ms */
+	queue_input.gang_context_addr = q->gang_ctx_gpu_addr;
+	queue_input.inprocess_gang_priority = q->properties.priority;
+	queue_input.gang_global_priority_level =
+					AMDGPU_MES_PRIORITY_LEVEL_NORMAL;
+	queue_input.doorbell_offset = q->properties.doorbell_off;
+	queue_input.mqd_addr = q->gart_mqd_addr;
+	queue_input.wptr_addr = (uint64_t)q->properties.write_ptr;
+
+	wptr_addr_off = (uint64_t)q->properties.write_ptr & (PAGE_SIZE - 1);
+	queue_input.wptr_mc_addr = amdgpu_bo_gpu_offset(q->properties.wptr_bo) + wptr_addr_off;
+
+	queue_input.is_kfd_process = 1;
+	queue_input.is_aql_queue = (q->properties.format == KFD_QUEUE_FORMAT_AQL);
+	queue_input.queue_size = q->properties.queue_size >> 2;
+
+	queue_input.paging = false;
+	queue_input.tba_addr = qpd->tba_addr;
+	queue_input.tma_addr = qpd->tma_addr;
+	queue_input.trap_en = !kfd_dbg_has_cwsr_workaround(q->device);
+	queue_input.skip_process_ctx_clear =
+		qpd->pqm->process->runtime_info.runtime_state == DEBUG_RUNTIME_STATE_ENABLED &&
+						(qpd->pqm->process->debug_trap_enabled ||
+						 kfd_dbg_has_ttmps_always_setup(q->device));
+
+	queue_type = convert_to_mes_queue_type(q->properties.type);
+	if (queue_type < 0) {
+		dev_err(adev->dev, "Queue type not supported with MES, queue:%d\n",
+			q->properties.type);
+		up_read(&adev->reset_domain->sem);
+		return -EINVAL;
+	}
+	queue_input.queue_type = (uint32_t)queue_type;
+
+	queue_input.exclusively_scheduled = q->properties.is_gws;
+
+	amdgpu_mes_lock(&adev->mes);
+	r = adev->mes.funcs->add_hw_queue(&adev->mes, &queue_input);
+	amdgpu_mes_unlock(&adev->mes);
+	up_read(&adev->reset_domain->sem);
+	if (r) {
+		dev_err(adev->dev, "failed to add hardware queue to MES, doorbell=0x%x\n",
+			q->properties.doorbell_off);
+		dev_err(adev->dev, "MES might be in unrecoverable state, issue a GPU reset\n");
+		kfd_hws_hang(dqm);
+	}
+
+	return r;
+}
+
+static int remove_queue_mes(struct device_queue_manager *dqm, struct queue *q,
+			struct qcm_process_device *qpd)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)dqm->dev->adev;
+	int r;
+	struct mes_remove_queue_input queue_input;
+
+	if (!dqm->sched_running || dqm->sched_halt)
+		return 0;
+	if (!down_read_trylock(&adev->reset_domain->sem))
+		return -EIO;
+
+	memset(&queue_input, 0x0, sizeof(struct mes_remove_queue_input));
+	queue_input.doorbell_offset = q->properties.doorbell_off;
+	queue_input.gang_context_addr = q->gang_ctx_gpu_addr;
+
+	amdgpu_mes_lock(&adev->mes);
+	r = adev->mes.funcs->remove_hw_queue(&adev->mes, &queue_input);
+	amdgpu_mes_unlock(&adev->mes);
+	up_read(&adev->reset_domain->sem);
+
+	if (r) {
+		dev_err(adev->dev, "failed to remove hardware queue from MES, doorbell=0x%x\n",
+			q->properties.doorbell_off);
+		dev_err(adev->dev, "MES might be in unrecoverable state, issue a GPU reset\n");
+		kfd_hws_hang(dqm);
+	}
+
+	return r;
+}
+
+static int remove_all_kfd_queues_mes(struct device_queue_manager *dqm)
+{
+	struct device_process_node *cur;
+	struct device *dev = dqm->dev->adev->dev;
+	struct qcm_process_device *qpd;
+	struct queue *q;
+	int retval = 0;
+
+	list_for_each_entry(cur, &dqm->queues, list) {
+		qpd = cur->qpd;
+		list_for_each_entry(q, &qpd->queues_list, list) {
+			if (q->properties.is_active) {
+				retval = remove_queue_mes(dqm, q, qpd);
+				if (retval) {
+					dev_err(dev, "%s: Failed to remove queue %d for dev %d",
+						__func__,
+						q->properties.queue_id,
+						dqm->dev->id);
+					return retval;
+				}
+			}
+		}
+	}
+
+	return retval;
+}
+
+static int add_all_kfd_queues_mes(struct device_queue_manager *dqm)
+{
+	struct device_process_node *cur;
+	struct device *dev = dqm->dev->adev->dev;
+	struct qcm_process_device *qpd;
+	struct queue *q;
+	int retval = 0;
+
+	list_for_each_entry(cur, &dqm->queues, list) {
+		qpd = cur->qpd;
+		list_for_each_entry(q, &qpd->queues_list, list) {
+			if (!q->properties.is_active)
+				continue;
+			retval = add_queue_mes(dqm, q, qpd);
+			if (retval) {
+				dev_err(dev, "%s: Failed to add queue %d for dev %d",
+					__func__,
+					q->properties.queue_id,
+					dqm->dev->id);
+				return retval;
+			}
+		}
+	}
+
+	return retval;
+}
+
+static int suspend_all_queues_mes(struct device_queue_manager *dqm)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)dqm->dev->adev;
+	int r = 0;
+
+	if (!down_read_trylock(&adev->reset_domain->sem))
+		return -EIO;
+
+	r = amdgpu_mes_suspend(adev);
+	up_read(&adev->reset_domain->sem);
+
+	if (r) {
+		dev_err(adev->dev, "failed to suspend gangs from MES\n");
+		dev_err(adev->dev, "MES might be in unrecoverable state, issue a GPU reset\n");
+		kfd_hws_hang(dqm);
+	}
+
+	return r;
+}
+
+static int resume_all_queues_mes(struct device_queue_manager *dqm)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)dqm->dev->adev;
+	int r = 0;
+
+	if (!down_read_trylock(&adev->reset_domain->sem))
+		return -EIO;
+
+	r = amdgpu_mes_resume(adev);
+	up_read(&adev->reset_domain->sem);
+
+	if (r) {
+		dev_err(adev->dev, "failed to resume gangs from MES\n");
+		dev_err(adev->dev, "MES might be in unrecoverable state, issue a GPU reset\n");
+		kfd_hws_hang(dqm);
+	}
+
+	return r;
 }
 
 static void increment_queue_count(struct device_queue_manager *dqm,
-			enum kfd_queue_type type)
+				  struct qcm_process_device *qpd,
+				  struct queue *q)
 {
 	dqm->active_queue_count++;
-	if (type == KFD_QUEUE_TYPE_COMPUTE || type == KFD_QUEUE_TYPE_DIQ)
+	if (q->properties.type == KFD_QUEUE_TYPE_COMPUTE ||
+	    q->properties.type == KFD_QUEUE_TYPE_DIQ)
 		dqm->active_cp_queue_count++;
+
+	if (q->properties.is_gws) {
+		dqm->gws_queue_count++;
+		qpd->mapped_gws_queue = true;
+	}
 }
 
 static void decrement_queue_count(struct device_queue_manager *dqm,
-			enum kfd_queue_type type)
+				  struct qcm_process_device *qpd,
+				  struct queue *q)
 {
 	dqm->active_queue_count--;
-	if (type == KFD_QUEUE_TYPE_COMPUTE || type == KFD_QUEUE_TYPE_DIQ)
+	if (q->properties.type == KFD_QUEUE_TYPE_COMPUTE ||
+	    q->properties.type == KFD_QUEUE_TYPE_DIQ)
 		dqm->active_cp_queue_count--;
+
+	if (q->properties.is_gws) {
+		dqm->gws_queue_count--;
+		qpd->mapped_gws_queue = false;
+	}
 }
 
-static int allocate_doorbell(struct qcm_process_device *qpd, struct queue *q)
+/*
+ * Allocate a doorbell ID to this queue.
+ * If doorbell_id is passed in, make sure requested ID is valid then allocate it.
+ */
+static int allocate_doorbell(struct qcm_process_device *qpd,
+			     struct queue *q,
+			     uint32_t const *restore_id)
 {
-	struct kfd_dev *dev = qpd->dqm->dev;
+	struct kfd_node *dev = qpd->dqm->dev;
 
-	if (!KFD_IS_SOC15(dev->device_info->asic_family)) {
+	if (!KFD_IS_SOC15(dev)) {
 		/* On pre-SOC15 chips we need to use the queue ID to
 		 * preserve the user mode ABI.
 		 */
+
+		if (restore_id && *restore_id != q->properties.queue_id)
+			return -EINVAL;
+
 		q->doorbell_id = q->properties.queue_id;
 	} else if (q->properties.type == KFD_QUEUE_TYPE_SDMA ||
 			q->properties.type == KFD_QUEUE_TYPE_SDMA_XGMI) {
@@ -169,30 +450,52 @@ static int allocate_doorbell(struct qcm_process_device *qpd, struct queue *q)
 		 * The doobell index distance between RLC (2*i) and (2*i+1)
 		 * for a SDMA engine is 512.
 		 */
-		uint32_t *idx_offset =
-				dev->shared_resources.sdma_doorbell_idx;
 
-		q->doorbell_id = idx_offset[q->properties.sdma_engine_id]
-			+ (q->properties.sdma_queue_id & 1)
-			* KFD_QUEUE_DOORBELL_MIRROR_OFFSET
-			+ (q->properties.sdma_queue_id >> 1);
+		uint32_t *idx_offset = dev->kfd->shared_resources.sdma_doorbell_idx;
+
+		/*
+		 * q->properties.sdma_engine_id corresponds to the virtual
+		 * sdma engine number. However, for doorbell allocation,
+		 * we need the physical sdma engine id in order to get the
+		 * correct doorbell offset.
+		 */
+		uint32_t valid_id = idx_offset[qpd->dqm->dev->node_id *
+					       get_num_all_sdma_engines(qpd->dqm) +
+					       q->properties.sdma_engine_id]
+						+ (q->properties.sdma_queue_id & 1)
+						* KFD_QUEUE_DOORBELL_MIRROR_OFFSET
+						+ (q->properties.sdma_queue_id >> 1);
+
+		if (restore_id && *restore_id != valid_id)
+			return -EINVAL;
+		q->doorbell_id = valid_id;
 	} else {
-		/* For CP queues on SOC15 reserve a free doorbell ID */
-		unsigned int found;
-
-		found = find_first_zero_bit(qpd->doorbell_bitmap,
-					    KFD_MAX_NUM_OF_QUEUES_PER_PROCESS);
-		if (found >= KFD_MAX_NUM_OF_QUEUES_PER_PROCESS) {
-			pr_debug("No doorbells available");
-			return -EBUSY;
+		/* For CP queues on SOC15 */
+		if (restore_id) {
+			/* make sure that ID is free  */
+			if (__test_and_set_bit(*restore_id, qpd->doorbell_bitmap))
+				return -EINVAL;
+
+			q->doorbell_id = *restore_id;
+		} else {
+			/* or reserve a free doorbell ID */
+			unsigned int found;
+
+			found = find_first_zero_bit(qpd->doorbell_bitmap,
+						    KFD_MAX_NUM_OF_QUEUES_PER_PROCESS);
+			if (found >= KFD_MAX_NUM_OF_QUEUES_PER_PROCESS) {
+				pr_debug("No doorbells available");
+				return -EBUSY;
+			}
+			set_bit(found, qpd->doorbell_bitmap);
+			q->doorbell_id = found;
 		}
-		set_bit(found, qpd->doorbell_bitmap);
-		q->doorbell_id = found;
 	}
 
-	q->properties.doorbell_off =
-		kfd_get_doorbell_dw_offset_in_bar(dev, qpd_to_pdd(qpd),
-					  q->doorbell_id);
+	q->properties.doorbell_off = amdgpu_doorbell_index_on_bar(dev->adev,
+								  qpd->proc_doorbells,
+								  q->doorbell_id,
+								  dev->kfd->device_info.doorbell_size);
 	return 0;
 }
 
@@ -200,9 +503,9 @@ static void deallocate_doorbell(struct qcm_process_device *qpd,
 				struct queue *q)
 {
 	unsigned int old;
-	struct kfd_dev *dev = qpd->dqm->dev;
+	struct kfd_node *dev = qpd->dqm->dev;
 
-	if (!KFD_IS_SOC15(dev->device_info->asic_family) ||
+	if (!KFD_IS_SOC15(dev) ||
 	    q->properties.type == KFD_QUEUE_TYPE_SDMA ||
 	    q->properties.type == KFD_QUEUE_TYPE_SDMA_XGMI)
 		return;
@@ -214,16 +517,22 @@ static void deallocate_doorbell(struct qcm_process_device *qpd,
 static void program_trap_handler_settings(struct device_queue_manager *dqm,
 				struct qcm_process_device *qpd)
 {
+	uint32_t xcc_mask = dqm->dev->xcc_mask;
+	int xcc_id;
+
 	if (dqm->dev->kfd2kgd->program_trap_handler_settings)
-		dqm->dev->kfd2kgd->program_trap_handler_settings(
-						dqm->dev->kgd, qpd->vmid,
-						qpd->tba_addr, qpd->tma_addr);
+		for_each_inst(xcc_id, xcc_mask)
+			dqm->dev->kfd2kgd->program_trap_handler_settings(
+				dqm->dev->adev, qpd->vmid, qpd->tba_addr,
+				qpd->tma_addr, xcc_id);
 }
 
 static int allocate_vmid(struct device_queue_manager *dqm,
 			struct qcm_process_device *qpd,
 			struct queue *q)
 {
+	struct kfd_process_device *pdd = qpd_to_pdd(qpd);
+	struct device *dev = dqm->dev->adev->dev;
 	int allocated_vmid = -1, i;
 
 	for (i = dqm->dev->vm_info.first_vmid_kfd;
@@ -235,42 +544,41 @@ static int allocate_vmid(struct device_queue_manager *dqm,
 	}
 
 	if (allocated_vmid < 0) {
-		pr_err("no more vmid to allocate\n");
+		dev_err(dev, "no more vmid to allocate\n");
 		return -ENOSPC;
 	}
 
 	pr_debug("vmid allocated: %d\n", allocated_vmid);
 
-	dqm->vmid_pasid[allocated_vmid] = q->process->pasid;
+	dqm->vmid_pasid[allocated_vmid] = pdd->pasid;
 
-	set_pasid_vmid_mapping(dqm, q->process->pasid, allocated_vmid);
+	set_pasid_vmid_mapping(dqm, pdd->pasid, allocated_vmid);
 
 	qpd->vmid = allocated_vmid;
 	q->properties.vmid = allocated_vmid;
 
 	program_sh_mem_settings(dqm, qpd);
 
-	if (dqm->dev->device_info->asic_family >= CHIP_VEGA10 &&
-	    dqm->dev->cwsr_enabled)
+	if (KFD_IS_SOC15(dqm->dev) && dqm->dev->kfd->cwsr_enabled)
 		program_trap_handler_settings(dqm, qpd);
 
 	/* qpd->page_table_base is set earlier when register_process()
 	 * is called, i.e. when the first queue is created.
 	 */
-	dqm->dev->kfd2kgd->set_vm_context_page_table_base(dqm->dev->kgd,
+	dqm->dev->kfd2kgd->set_vm_context_page_table_base(dqm->dev->adev,
 			qpd->vmid,
 			qpd->page_table_base);
 	/* invalidate the VM context after pasid and vmid mapping is set up */
 	kfd_flush_tlb(qpd_to_pdd(qpd), TLB_FLUSH_LEGACY);
 
 	if (dqm->dev->kfd2kgd->set_scratch_backing_va)
-		dqm->dev->kfd2kgd->set_scratch_backing_va(dqm->dev->kgd,
+		dqm->dev->kfd2kgd->set_scratch_backing_va(dqm->dev->adev,
 				qpd->sh_hidden_private_base, qpd->vmid);
 
 	return 0;
 }
 
-static int flush_texture_cache_nocpsch(struct kfd_dev *kdev,
+static int flush_texture_cache_nocpsch(struct kfd_node *kdev,
 				struct qcm_process_device *qpd)
 {
 	const struct packet_manager_funcs *pmf = qpd->dqm->packet_mgr.pmf;
@@ -283,7 +591,7 @@ static int flush_texture_cache_nocpsch(struct kfd_dev *kdev,
 	if (ret)
 		return ret;
 
-	return amdgpu_amdkfd_submit_ib(kdev->kgd, KGD_ENGINE_MEC1, qpd->vmid,
+	return amdgpu_amdkfd_submit_ib(kdev->adev, KGD_ENGINE_MEC1, qpd->vmid,
 				qpd->ib_base, (uint32_t *)qpd->ib_kaddr,
 				pmf->release_mem_size / sizeof(uint32_t));
 }
@@ -292,10 +600,12 @@ static void deallocate_vmid(struct device_queue_manager *dqm,
 				struct qcm_process_device *qpd,
 				struct queue *q)
 {
+	struct device *dev = dqm->dev->adev->dev;
+
 	/* On GFX v7, CP doesn't flush TC at dequeue */
-	if (q->device->device_info->asic_family == CHIP_HAWAII)
+	if (q->device->adev->asic_type == CHIP_HAWAII)
 		if (flush_texture_cache_nocpsch(q->device, qpd))
-			pr_err("Failed to flush TC\n");
+			dev_err(dev, "Failed to flush TC\n");
 
 	kfd_flush_tlb(qpd_to_pdd(qpd), TLB_FLUSH_LEGACY);
 
@@ -309,7 +619,9 @@ static void deallocate_vmid(struct device_queue_manager *dqm,
 
 static int create_queue_nocpsch(struct device_queue_manager *dqm,
 				struct queue *q,
-				struct qcm_process_device *qpd)
+				struct qcm_process_device *qpd,
+				const struct kfd_criu_queue_priv_data *qd,
+				const void *restore_mqd, const void *restore_ctl_stack)
 {
 	struct mqd_manager *mqd_mgr;
 	int retval;
@@ -349,13 +661,13 @@ static int create_queue_nocpsch(struct device_queue_manager *dqm,
 			q->pipe, q->queue);
 	} else if (q->properties.type == KFD_QUEUE_TYPE_SDMA ||
 		q->properties.type == KFD_QUEUE_TYPE_SDMA_XGMI) {
-		retval = allocate_sdma_queue(dqm, q);
+		retval = allocate_sdma_queue(dqm, q, qd ? &qd->sdma_id : NULL);
 		if (retval)
 			goto deallocate_vmid;
 		dqm->asic_ops.init_sdma_vm(dqm, q, qpd);
 	}
 
-	retval = allocate_doorbell(qpd, q);
+	retval = allocate_doorbell(qpd, q, qd ? &qd->doorbell_id : NULL);
 	if (retval)
 		goto out_deallocate_hqd;
 
@@ -368,8 +680,15 @@ static int create_queue_nocpsch(struct device_queue_manager *dqm,
 		retval = -ENOMEM;
 		goto out_deallocate_doorbell;
 	}
-	mqd_mgr->init_mqd(mqd_mgr, &q->mqd, q->mqd_mem_obj,
-				&q->gart_mqd_addr, &q->properties);
+
+	if (qd)
+		mqd_mgr->restore_mqd(mqd_mgr, &q->mqd, q->mqd_mem_obj, &q->gart_mqd_addr,
+				     &q->properties, restore_mqd, restore_ctl_stack,
+				     qd->ctl_stack_size);
+	else
+		mqd_mgr->init_mqd(mqd_mgr, &q->mqd, q->mqd_mem_obj,
+					&q->gart_mqd_addr, &q->properties);
+
 	if (q->properties.is_active) {
 		if (!dqm->sched_running) {
 			WARN_ONCE(1, "Load non-HWS mqd while stopped\n");
@@ -390,7 +709,7 @@ add_queue_to_list:
 	list_add(&q->list, &qpd->queues_list);
 	qpd->queue_count++;
 	if (q->properties.is_active)
-		increment_queue_count(dqm, q->properties.type);
+		increment_queue_count(dqm, qpd, q);
 
 	/*
 	 * Unconditionally increment this counter, regardless of the queue's
@@ -459,6 +778,74 @@ static inline void deallocate_hqd(struct device_queue_manager *dqm,
 	dqm->allocated_queues[q->pipe] |= (1 << q->queue);
 }
 
+#define SQ_IND_CMD_CMD_KILL		0x00000003
+#define SQ_IND_CMD_MODE_BROADCAST	0x00000001
+
+static int dbgdev_wave_reset_wavefronts(struct kfd_node *dev, struct kfd_process *p)
+{
+	int status = 0;
+	unsigned int vmid;
+	uint16_t queried_pasid;
+	union SQ_CMD_BITS reg_sq_cmd;
+	union GRBM_GFX_INDEX_BITS reg_gfx_index;
+	struct kfd_process_device *pdd;
+	int first_vmid_to_scan = dev->vm_info.first_vmid_kfd;
+	int last_vmid_to_scan = dev->vm_info.last_vmid_kfd;
+	uint32_t xcc_mask = dev->xcc_mask;
+	int xcc_id;
+
+	reg_sq_cmd.u32All = 0;
+	reg_gfx_index.u32All = 0;
+
+	pr_debug("Killing all process wavefronts\n");
+
+	if (!dev->kfd2kgd->get_atc_vmid_pasid_mapping_info) {
+		dev_err(dev->adev->dev, "no vmid pasid mapping supported\n");
+		return -EOPNOTSUPP;
+	}
+
+	/* taking the VMID for that process on the safe way using PDD */
+	pdd = kfd_get_process_device_data(dev, p);
+	if (!pdd)
+		return -EFAULT;
+
+	/* Scan all registers in the range ATC_VMID8_PASID_MAPPING ..
+	 * ATC_VMID15_PASID_MAPPING
+	 * to check which VMID the current process is mapped to.
+	 */
+
+	for (vmid = first_vmid_to_scan; vmid <= last_vmid_to_scan; vmid++) {
+		status = dev->kfd2kgd->get_atc_vmid_pasid_mapping_info
+				(dev->adev, vmid, &queried_pasid);
+
+		if (status && queried_pasid == pdd->pasid) {
+			pr_debug("Killing wave fronts of vmid %d and process pid %d\n",
+					vmid, p->lead_thread->pid);
+			break;
+		}
+	}
+
+	if (vmid > last_vmid_to_scan) {
+		dev_err(dev->adev->dev, "Didn't find vmid for process pid %d\n",
+				p->lead_thread->pid);
+		return -EFAULT;
+	}
+
+	reg_gfx_index.bits.sh_broadcast_writes = 1;
+	reg_gfx_index.bits.se_broadcast_writes = 1;
+	reg_gfx_index.bits.instance_broadcast_writes = 1;
+	reg_sq_cmd.bits.mode = SQ_IND_CMD_MODE_BROADCAST;
+	reg_sq_cmd.bits.cmd = SQ_IND_CMD_CMD_KILL;
+	reg_sq_cmd.bits.vm_id = vmid;
+
+	for_each_inst(xcc_id, xcc_mask)
+		dev->kfd2kgd->wave_control_execute(
+			dev->adev, reg_gfx_index.u32All,
+			reg_sq_cmd.u32All, xcc_id);
+
+	return 0;
+}
+
 /* Access to DQM has to be locked before calling destroy_queue_nocpsch_locked
  * to avoid asynchronized access
  */
@@ -515,13 +902,8 @@ static int destroy_queue_nocpsch_locked(struct device_queue_manager *dqm,
 		deallocate_vmid(dqm, qpd, q);
 	}
 	qpd->queue_count--;
-	if (q->properties.is_active) {
-		decrement_queue_count(dqm, q->properties.type);
-		if (q->properties.is_gws) {
-			dqm->gws_queue_count--;
-			qpd->mapped_gws_queue = false;
-		}
-	}
+	if (q->properties.is_active)
+		decrement_queue_count(dqm, qpd, q);
 
 	return retval;
 }
@@ -532,6 +914,7 @@ static int destroy_queue_nocpsch(struct device_queue_manager *dqm,
 {
 	int retval;
 	uint64_t sdma_val = 0;
+	struct device *dev = dqm->dev->adev->dev;
 	struct kfd_process_device *pdd = qpd_to_pdd(qpd);
 	struct mqd_manager *mqd_mgr =
 		dqm->mqd_mgrs[get_mqd_type_from_queue_type(q->properties.type)];
@@ -542,7 +925,7 @@ static int destroy_queue_nocpsch(struct device_queue_manager *dqm,
 		retval = read_sdma_queue_counter((uint64_t __user *)q->properties.read_ptr,
 							&sdma_val);
 		if (retval)
-			pr_err("Failed to read SDMA queue counter for queue: %d\n",
+			dev_err(dev, "Failed to read SDMA queue counter for queue: %d\n",
 				q->properties.queue_id);
 	}
 
@@ -561,6 +944,7 @@ static int update_queue(struct device_queue_manager *dqm, struct queue *q,
 			struct mqd_update_info *minfo)
 {
 	int retval = 0;
+	struct device *dev = dqm->dev->adev->dev;
 	struct mqd_manager *mqd_mgr;
 	struct kfd_process_device *pdd;
 	bool prev_active = false;
@@ -579,10 +963,20 @@ static int update_queue(struct device_queue_manager *dqm, struct queue *q,
 
 	/* Make sure the queue is unmapped before updating the MQD */
 	if (dqm->sched_policy != KFD_SCHED_POLICY_NO_HWS) {
-		retval = unmap_queues_cpsch(dqm,
-				KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0);
+		if (!dqm->dev->kfd->shared_resources.enable_mes)
+			retval = unmap_queues_cpsch(dqm,
+						    KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0, USE_DEFAULT_GRACE_PERIOD, false);
+		else if (prev_active)
+			retval = remove_queue_mes(dqm, q, &pdd->qpd);
+
+		/* queue is reset so inaccessable  */
+		if (pdd->has_reset_queue) {
+			retval = -EACCES;
+			goto out_unlock;
+		}
+
 		if (retval) {
-			pr_err("unmap queue failed\n");
+			dev_err(dev, "unmap queue failed\n");
 			goto out_unlock;
 		}
 	} else if (prev_active &&
@@ -596,12 +990,12 @@ static int update_queue(struct device_queue_manager *dqm, struct queue *q,
 		}
 
 		retval = mqd_mgr->destroy_mqd(mqd_mgr, q->mqd,
-				(dqm->dev->cwsr_enabled?
-				 KFD_PREEMPT_TYPE_WAVEFRONT_SAVE:
-				KFD_PREEMPT_TYPE_WAVEFRONT_DRAIN),
+				(dqm->dev->kfd->cwsr_enabled ?
+				 KFD_PREEMPT_TYPE_WAVEFRONT_SAVE :
+				 KFD_PREEMPT_TYPE_WAVEFRONT_DRAIN),
 				KFD_UNMAP_LATENCY_MS, q->pipe, q->queue);
 		if (retval) {
-			pr_err("destroy mqd failed\n");
+			dev_err(dev, "destroy mqd failed\n");
 			goto out_unlock;
 		}
 	}
@@ -614,12 +1008,11 @@ static int update_queue(struct device_queue_manager *dqm, struct queue *q,
 	 * dqm->active_queue_count to determine whether a new runlist must be
 	 * uploaded.
 	 */
-	if (q->properties.is_active && !prev_active)
-		increment_queue_count(dqm, q->properties.type);
-	else if (!q->properties.is_active && prev_active)
-		decrement_queue_count(dqm, q->properties.type);
-
-	if (q->gws && !q->properties.is_gws) {
+	if (q->properties.is_active && !prev_active) {
+		increment_queue_count(dqm, &pdd->qpd, q);
+	} else if (!q->properties.is_active && prev_active) {
+		decrement_queue_count(dqm, &pdd->qpd, q);
+	} else if (q->gws && !q->properties.is_gws) {
 		if (q->properties.is_active) {
 			dqm->gws_queue_count++;
 			pdd->qpd.mapped_gws_queue = true;
@@ -633,9 +1026,12 @@ static int update_queue(struct device_queue_manager *dqm, struct queue *q,
 		q->properties.is_gws = false;
 	}
 
-	if (dqm->sched_policy != KFD_SCHED_POLICY_NO_HWS)
-		retval = map_queues_cpsch(dqm);
-	else if (q->properties.is_active &&
+	if (dqm->sched_policy != KFD_SCHED_POLICY_NO_HWS) {
+		if (!dqm->dev->kfd->shared_resources.enable_mes)
+			retval = map_queues_cpsch(dqm);
+		else if (q->properties.is_active)
+			retval = add_queue_mes(dqm, q, &pdd->qpd);
+	} else if (q->properties.is_active &&
 		 (q->properties.type == KFD_QUEUE_TYPE_COMPUTE ||
 		  q->properties.type == KFD_QUEUE_TYPE_SDMA ||
 		  q->properties.type == KFD_QUEUE_TYPE_SDMA_XGMI)) {
@@ -653,6 +1049,92 @@ out_unlock:
 	return retval;
 }
 
+/* suspend_single_queue does not lock the dqm like the
+ * evict_process_queues_cpsch or evict_process_queues_nocpsch. You should
+ * lock the dqm before calling, and unlock after calling.
+ *
+ * The reason we don't lock the dqm is because this function may be
+ * called on multiple queues in a loop, so rather than locking/unlocking
+ * multiple times, we will just keep the dqm locked for all of the calls.
+ */
+static int suspend_single_queue(struct device_queue_manager *dqm,
+				      struct kfd_process_device *pdd,
+				      struct queue *q)
+{
+	bool is_new;
+
+	if (q->properties.is_suspended)
+		return 0;
+
+	pr_debug("Suspending process pid %d queue [%i]\n",
+			pdd->process->lead_thread->pid,
+			q->properties.queue_id);
+
+	is_new = q->properties.exception_status & KFD_EC_MASK(EC_QUEUE_NEW);
+
+	if (is_new || q->properties.is_being_destroyed) {
+		pr_debug("Suspend: skip %s queue id %i\n",
+				is_new ? "new" : "destroyed",
+				q->properties.queue_id);
+		return -EBUSY;
+	}
+
+	q->properties.is_suspended = true;
+	if (q->properties.is_active) {
+		if (dqm->dev->kfd->shared_resources.enable_mes) {
+			int r = remove_queue_mes(dqm, q, &pdd->qpd);
+
+			if (r)
+				return r;
+		}
+
+		decrement_queue_count(dqm, &pdd->qpd, q);
+		q->properties.is_active = false;
+	}
+
+	return 0;
+}
+
+/* resume_single_queue does not lock the dqm like the functions
+ * restore_process_queues_cpsch or restore_process_queues_nocpsch. You should
+ * lock the dqm before calling, and unlock after calling.
+ *
+ * The reason we don't lock the dqm is because this function may be
+ * called on multiple queues in a loop, so rather than locking/unlocking
+ * multiple times, we will just keep the dqm locked for all of the calls.
+ */
+static int resume_single_queue(struct device_queue_manager *dqm,
+				      struct qcm_process_device *qpd,
+				      struct queue *q)
+{
+	struct kfd_process_device *pdd;
+
+	if (!q->properties.is_suspended)
+		return 0;
+
+	pdd = qpd_to_pdd(qpd);
+
+	pr_debug("Restoring from suspend process pid %d queue [%i]\n",
+			    pdd->process->lead_thread->pid,
+			    q->properties.queue_id);
+
+	q->properties.is_suspended = false;
+
+	if (QUEUE_IS_ACTIVE(q->properties)) {
+		if (dqm->dev->kfd->shared_resources.enable_mes) {
+			int r = add_queue_mes(dqm, q, &pdd->qpd);
+
+			if (r)
+				return r;
+		}
+
+		q->properties.is_active = true;
+		increment_queue_count(dqm, qpd, q);
+	}
+
+	return 0;
+}
+
 static int evict_process_queues_nocpsch(struct device_queue_manager *dqm,
 					struct qcm_process_device *qpd)
 {
@@ -666,8 +1148,8 @@ static int evict_process_queues_nocpsch(struct device_queue_manager *dqm,
 		goto out;
 
 	pdd = qpd_to_pdd(qpd);
-	pr_debug_ratelimited("Evicting PASID 0x%x queues\n",
-			    pdd->process->pasid);
+	pr_debug_ratelimited("Evicting process pid %d queues\n",
+			    pdd->process->lead_thread->pid);
 
 	pdd->last_evict_timestamp = get_jiffies_64();
 	/* Mark all queues as evicted. Deactivate all active queues on
@@ -681,19 +1163,15 @@ static int evict_process_queues_nocpsch(struct device_queue_manager *dqm,
 		mqd_mgr = dqm->mqd_mgrs[get_mqd_type_from_queue_type(
 				q->properties.type)];
 		q->properties.is_active = false;
-		decrement_queue_count(dqm, q->properties.type);
-		if (q->properties.is_gws) {
-			dqm->gws_queue_count--;
-			qpd->mapped_gws_queue = false;
-		}
+		decrement_queue_count(dqm, qpd, q);
 
 		if (WARN_ONCE(!dqm->sched_running, "Evict when stopped\n"))
 			continue;
 
 		retval = mqd_mgr->destroy_mqd(mqd_mgr, q->mqd,
-				(dqm->dev->cwsr_enabled?
-				 KFD_PREEMPT_TYPE_WAVEFRONT_SAVE:
-				KFD_PREEMPT_TYPE_WAVEFRONT_DRAIN),
+				(dqm->dev->kfd->cwsr_enabled ?
+				 KFD_PREEMPT_TYPE_WAVEFRONT_SAVE :
+				 KFD_PREEMPT_TYPE_WAVEFRONT_DRAIN),
 				KFD_UNMAP_LATENCY_MS, q->pipe, q->queue);
 		if (retval && !ret)
 			/* Return the first error, but keep going to
@@ -711,6 +1189,7 @@ static int evict_process_queues_cpsch(struct device_queue_manager *dqm,
 				      struct qcm_process_device *qpd)
 {
 	struct queue *q;
+	struct device *dev = dqm->dev->adev->dev;
 	struct kfd_process_device *pdd;
 	int retval = 0;
 
@@ -719,8 +1198,25 @@ static int evict_process_queues_cpsch(struct device_queue_manager *dqm,
 		goto out;
 
 	pdd = qpd_to_pdd(qpd);
-	pr_debug_ratelimited("Evicting PASID 0x%x queues\n",
-			    pdd->process->pasid);
+
+	/* The debugger creates processes that temporarily have not acquired
+	 * all VMs for all devices and has no VMs itself.
+	 * Skip queue eviction on process eviction.
+	 */
+	if (!pdd->drm_priv)
+		goto out;
+
+	pr_debug_ratelimited("Evicting process pid %d queues\n",
+			    pdd->process->lead_thread->pid);
+
+	if (dqm->dev->kfd->shared_resources.enable_mes) {
+		pdd->last_evict_timestamp = get_jiffies_64();
+		retval = suspend_all_queues_mes(dqm);
+		if (retval) {
+			dev_err(dev, "Suspending all queues failed");
+			goto out;
+		}
+	}
 
 	/* Mark all queues as evicted. Deactivate all active queues on
 	 * the qpd.
@@ -731,13 +1227,30 @@ static int evict_process_queues_cpsch(struct device_queue_manager *dqm,
 			continue;
 
 		q->properties.is_active = false;
-		decrement_queue_count(dqm, q->properties.type);
+		decrement_queue_count(dqm, qpd, q);
+
+		if (dqm->dev->kfd->shared_resources.enable_mes) {
+			retval = remove_queue_mes(dqm, q, qpd);
+			if (retval) {
+				dev_err(dev, "Failed to evict queue %d\n",
+					q->properties.queue_id);
+				goto out;
+			}
+		}
+	}
+
+	if (!dqm->dev->kfd->shared_resources.enable_mes) {
+		pdd->last_evict_timestamp = get_jiffies_64();
+		retval = execute_queues_cpsch(dqm,
+					      qpd->is_debug ?
+					      KFD_UNMAP_QUEUES_FILTER_ALL_QUEUES :
+					      KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0,
+					      USE_DEFAULT_GRACE_PERIOD);
+	} else {
+		retval = resume_all_queues_mes(dqm);
+		if (retval)
+			dev_err(dev, "Resuming all queues failed");
 	}
-	pdd->last_evict_timestamp = get_jiffies_64();
-	retval = execute_queues_cpsch(dqm,
-				qpd->is_debug ?
-				KFD_UNMAP_QUEUES_FILTER_ALL_QUEUES :
-				KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0);
 
 out:
 	dqm_unlock(dqm);
@@ -767,8 +1280,8 @@ static int restore_process_queues_nocpsch(struct device_queue_manager *dqm,
 		goto out;
 	}
 
-	pr_debug_ratelimited("Restoring PASID 0x%x queues\n",
-			    pdd->process->pasid);
+	pr_debug_ratelimited("Restoring process pid %d queues\n",
+			    pdd->process->lead_thread->pid);
 
 	/* Update PD Base in QPD */
 	qpd->page_table_base = pd_base;
@@ -776,7 +1289,7 @@ static int restore_process_queues_nocpsch(struct device_queue_manager *dqm,
 
 	if (!list_empty(&qpd->queues_list)) {
 		dqm->dev->kfd2kgd->set_vm_context_page_table_base(
-				dqm->dev->kgd,
+				dqm->dev->adev,
 				qpd->vmid,
 				qpd->page_table_base);
 		kfd_flush_tlb(pdd, TLB_FLUSH_LEGACY);
@@ -802,11 +1315,7 @@ static int restore_process_queues_nocpsch(struct device_queue_manager *dqm,
 		mqd_mgr = dqm->mqd_mgrs[get_mqd_type_from_queue_type(
 				q->properties.type)];
 		q->properties.is_active = true;
-		increment_queue_count(dqm, q->properties.type);
-		if (q->properties.is_gws) {
-			dqm->gws_queue_count++;
-			qpd->mapped_gws_queue = true;
-		}
+		increment_queue_count(dqm, qpd, q);
 
 		if (WARN_ONCE(!dqm->sched_running, "Restore when stopped\n"))
 			continue;
@@ -833,14 +1342,12 @@ static int restore_process_queues_cpsch(struct device_queue_manager *dqm,
 					struct qcm_process_device *qpd)
 {
 	struct queue *q;
+	struct device *dev = dqm->dev->adev->dev;
 	struct kfd_process_device *pdd;
-	uint64_t pd_base;
 	uint64_t eviction_duration;
 	int retval = 0;
 
 	pdd = qpd_to_pdd(qpd);
-	/* Retrieve PD base */
-	pd_base = amdgpu_amdkfd_gpuvm_get_process_page_dir(pdd->drm_priv);
 
 	dqm_lock(dqm);
 	if (WARN_ON_ONCE(!qpd->evicted)) /* already restored, do nothing */
@@ -850,12 +1357,19 @@ static int restore_process_queues_cpsch(struct device_queue_manager *dqm,
 		goto out;
 	}
 
-	pr_debug_ratelimited("Restoring PASID 0x%x queues\n",
-			    pdd->process->pasid);
+	/* The debugger creates processes that temporarily have not acquired
+	 * all VMs for all devices and has no VMs itself.
+	 * Skip queue restore on process restore.
+	 */
+	if (!pdd->drm_priv)
+		goto vm_not_acquired;
+
+	pr_debug_ratelimited("Restoring process pid %d queues\n",
+			    pdd->process->lead_thread->pid);
 
 	/* Update PD Base in QPD */
-	qpd->page_table_base = pd_base;
-	pr_debug("Updated PD address to 0x%llx\n", pd_base);
+	qpd->page_table_base = amdgpu_amdkfd_gpuvm_get_process_page_dir(pdd->drm_priv);
+	pr_debug("Updated PD address to 0x%llx\n", qpd->page_table_base);
 
 	/* activate all active queues on the qpd */
 	list_for_each_entry(q, &qpd->queues_list, list) {
@@ -864,13 +1378,24 @@ static int restore_process_queues_cpsch(struct device_queue_manager *dqm,
 			continue;
 
 		q->properties.is_active = true;
-		increment_queue_count(dqm, q->properties.type);
+		increment_queue_count(dqm, &pdd->qpd, q);
+
+		if (dqm->dev->kfd->shared_resources.enable_mes) {
+			retval = add_queue_mes(dqm, q, qpd);
+			if (retval) {
+				dev_err(dev, "Failed to restore queue %d\n",
+					q->properties.queue_id);
+				goto out;
+			}
+		}
 	}
-	retval = execute_queues_cpsch(dqm,
-				KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0);
-	qpd->evicted = 0;
+	if (!dqm->dev->kfd->shared_resources.enable_mes)
+		retval = execute_queues_cpsch(dqm,
+					      KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0, USE_DEFAULT_GRACE_PERIOD);
 	eviction_duration = get_jiffies_64() - pdd->last_evict_timestamp;
 	atomic64_add(eviction_duration, &pdd->evict_duration_counter);
+vm_not_acquired:
+	qpd->evicted = 0;
 out:
 	dqm_unlock(dqm);
 	return retval;
@@ -953,17 +1478,32 @@ static int
 set_pasid_vmid_mapping(struct device_queue_manager *dqm, u32 pasid,
 			unsigned int vmid)
 {
-	return dqm->dev->kfd2kgd->set_pasid_vmid_mapping(
-						dqm->dev->kgd, pasid, vmid);
+	uint32_t xcc_mask = dqm->dev->xcc_mask;
+	int xcc_id, ret;
+
+	for_each_inst(xcc_id, xcc_mask) {
+		ret = dqm->dev->kfd2kgd->set_pasid_vmid_mapping(
+			dqm->dev->adev, pasid, vmid, xcc_id);
+		if (ret)
+			break;
+	}
+
+	return ret;
 }
 
 static void init_interrupts(struct device_queue_manager *dqm)
 {
-	unsigned int i;
+	uint32_t xcc_mask = dqm->dev->xcc_mask;
+	unsigned int i, xcc_id;
 
-	for (i = 0 ; i < get_pipes_per_mec(dqm) ; i++)
-		if (is_pipe_enabled(dqm, 0, i))
-			dqm->dev->kfd2kgd->init_interrupts(dqm->dev->kgd, i);
+	for_each_inst(xcc_id, xcc_mask) {
+		for (i = 0 ; i < get_pipes_per_mec(dqm) ; i++) {
+			if (is_pipe_enabled(dqm, 0, i)) {
+				dqm->dev->kfd2kgd->init_interrupts(
+					dqm->dev->adev, i, xcc_id);
+			}
+		}
+	}
 }
 
 static int initialize_nocpsch(struct device_queue_manager *dqm)
@@ -988,14 +1528,13 @@ static int initialize_nocpsch(struct device_queue_manager *dqm)
 
 		for (queue = 0; queue < get_queues_per_pipe(dqm); queue++)
 			if (test_bit(pipe_offset + queue,
-				     dqm->dev->shared_resources.cp_queue_bitmap))
+				     dqm->dev->kfd->shared_resources.cp_queue_bitmap))
 				dqm->allocated_queues[pipe] |= 1 << queue;
 	}
 
 	memset(dqm->vmid_pasid, 0, sizeof(dqm->vmid_pasid));
 
-	dqm->sdma_bitmap = ~0ULL >> (64 - get_num_sdma_queues(dqm));
-	dqm->xgmi_sdma_bitmap = ~0ULL >> (64 - get_num_xgmi_sdma_queues(dqm));
+	init_sdma_bitmaps(dqm);
 
 	return 0;
 }
@@ -1014,68 +1553,134 @@ static void uninitialize(struct device_queue_manager *dqm)
 
 static int start_nocpsch(struct device_queue_manager *dqm)
 {
+	int r = 0;
+
 	pr_info("SW scheduler is used");
 	init_interrupts(dqm);
-	
-	if (dqm->dev->device_info->asic_family == CHIP_HAWAII)
-		return pm_init(&dqm->packet_mgr, dqm);
-	dqm->sched_running = true;
 
-	return 0;
+	if (dqm->dev->adev->asic_type == CHIP_HAWAII)
+		r = pm_init(&dqm->packet_mgr, dqm);
+	if (!r)
+		dqm->sched_running = true;
+
+	return r;
 }
 
 static int stop_nocpsch(struct device_queue_manager *dqm)
 {
-	if (dqm->dev->device_info->asic_family == CHIP_HAWAII)
-		pm_uninit(&dqm->packet_mgr, false);
+	dqm_lock(dqm);
+	if (!dqm->sched_running) {
+		dqm_unlock(dqm);
+		return 0;
+	}
+
+	if (dqm->dev->adev->asic_type == CHIP_HAWAII)
+		pm_uninit(&dqm->packet_mgr);
 	dqm->sched_running = false;
+	dqm_unlock(dqm);
 
 	return 0;
 }
 
-static void pre_reset(struct device_queue_manager *dqm)
-{
-	dqm_lock(dqm);
-	dqm->is_resetting = true;
-	dqm_unlock(dqm);
-}
-
 static int allocate_sdma_queue(struct device_queue_manager *dqm,
-				struct queue *q)
+				struct queue *q, const uint32_t *restore_sdma_id)
 {
+	struct device *dev = dqm->dev->adev->dev;
 	int bit;
 
 	if (q->properties.type == KFD_QUEUE_TYPE_SDMA) {
-		if (dqm->sdma_bitmap == 0) {
-			pr_err("No more SDMA queue to allocate\n");
+		if (bitmap_empty(dqm->sdma_bitmap, get_num_sdma_queues(dqm))) {
+			dev_warn(dev, "No more SDMA queue to allocate (%d total queues)\n",
+				 get_num_sdma_queues(dqm));
 			return -ENOMEM;
 		}
 
-		bit = __ffs64(dqm->sdma_bitmap);
-		dqm->sdma_bitmap &= ~(1ULL << bit);
-		q->sdma_id = bit;
-		q->properties.sdma_engine_id = q->sdma_id %
-				get_num_sdma_engines(dqm);
+		if (restore_sdma_id) {
+			/* Re-use existing sdma_id */
+			if (!test_bit(*restore_sdma_id, dqm->sdma_bitmap)) {
+				dev_err(dev, "SDMA queue already in use\n");
+				return -EBUSY;
+			}
+			clear_bit(*restore_sdma_id, dqm->sdma_bitmap);
+			q->sdma_id = *restore_sdma_id;
+		} else {
+			/* Find first available sdma_id */
+			bit = find_first_bit(dqm->sdma_bitmap,
+					     get_num_sdma_queues(dqm));
+			clear_bit(bit, dqm->sdma_bitmap);
+			q->sdma_id = bit;
+		}
+
+		q->properties.sdma_engine_id =
+			q->sdma_id % kfd_get_num_sdma_engines(dqm->dev);
 		q->properties.sdma_queue_id = q->sdma_id /
-				get_num_sdma_engines(dqm);
+				kfd_get_num_sdma_engines(dqm->dev);
 	} else if (q->properties.type == KFD_QUEUE_TYPE_SDMA_XGMI) {
-		if (dqm->xgmi_sdma_bitmap == 0) {
-			pr_err("No more XGMI SDMA queue to allocate\n");
+		if (bitmap_empty(dqm->xgmi_sdma_bitmap, get_num_xgmi_sdma_queues(dqm))) {
+			dev_warn(dev, "No more XGMI SDMA queue to allocate (%d total queues)\n",
+				 get_num_xgmi_sdma_queues(dqm));
 			return -ENOMEM;
 		}
-		bit = __ffs64(dqm->xgmi_sdma_bitmap);
-		dqm->xgmi_sdma_bitmap &= ~(1ULL << bit);
-		q->sdma_id = bit;
+		if (restore_sdma_id) {
+			/* Re-use existing sdma_id */
+			if (!test_bit(*restore_sdma_id, dqm->xgmi_sdma_bitmap)) {
+				dev_err(dev, "SDMA queue already in use\n");
+				return -EBUSY;
+			}
+			clear_bit(*restore_sdma_id, dqm->xgmi_sdma_bitmap);
+			q->sdma_id = *restore_sdma_id;
+		} else {
+			bit = find_first_bit(dqm->xgmi_sdma_bitmap,
+					     get_num_xgmi_sdma_queues(dqm));
+			clear_bit(bit, dqm->xgmi_sdma_bitmap);
+			q->sdma_id = bit;
+		}
 		/* sdma_engine_id is sdma id including
 		 * both PCIe-optimized SDMAs and XGMI-
 		 * optimized SDMAs. The calculation below
 		 * assumes the first N engines are always
 		 * PCIe-optimized ones
 		 */
-		q->properties.sdma_engine_id = get_num_sdma_engines(dqm) +
-				q->sdma_id % get_num_xgmi_sdma_engines(dqm);
+		q->properties.sdma_engine_id =
+			kfd_get_num_sdma_engines(dqm->dev) +
+			q->sdma_id % kfd_get_num_xgmi_sdma_engines(dqm->dev);
 		q->properties.sdma_queue_id = q->sdma_id /
-				get_num_xgmi_sdma_engines(dqm);
+			kfd_get_num_xgmi_sdma_engines(dqm->dev);
+	} else if (q->properties.type == KFD_QUEUE_TYPE_SDMA_BY_ENG_ID) {
+		int i, num_queues, num_engines, eng_offset = 0, start_engine;
+		bool free_bit_found = false, is_xgmi = false;
+
+		if (q->properties.sdma_engine_id < kfd_get_num_sdma_engines(dqm->dev)) {
+			num_queues = get_num_sdma_queues(dqm);
+			num_engines = kfd_get_num_sdma_engines(dqm->dev);
+			q->properties.type = KFD_QUEUE_TYPE_SDMA;
+		} else {
+			num_queues = get_num_xgmi_sdma_queues(dqm);
+			num_engines = kfd_get_num_xgmi_sdma_engines(dqm->dev);
+			eng_offset = kfd_get_num_sdma_engines(dqm->dev);
+			q->properties.type = KFD_QUEUE_TYPE_SDMA_XGMI;
+			is_xgmi = true;
+		}
+
+		/* Scan available bit based on target engine ID. */
+		start_engine = q->properties.sdma_engine_id - eng_offset;
+		for (i = start_engine; i < num_queues; i += num_engines) {
+
+			if (!test_bit(i, is_xgmi ? dqm->xgmi_sdma_bitmap : dqm->sdma_bitmap))
+				continue;
+
+			clear_bit(i, is_xgmi ? dqm->xgmi_sdma_bitmap : dqm->sdma_bitmap);
+			q->sdma_id = i;
+			q->properties.sdma_queue_id = q->sdma_id / num_engines;
+			free_bit_found = true;
+			break;
+		}
+
+		if (!free_bit_found) {
+			dev_warn(dev, "No more SDMA queue to allocate for target ID %i (%d total queues)\n",
+				 q->properties.sdma_engine_id, num_queues);
+			return -ENOMEM;
+		}
 	}
 
 	pr_debug("SDMA engine id: %d\n", q->properties.sdma_engine_id);
@@ -1090,11 +1695,11 @@ static void deallocate_sdma_queue(struct device_queue_manager *dqm,
 	if (q->properties.type == KFD_QUEUE_TYPE_SDMA) {
 		if (q->sdma_id >= get_num_sdma_queues(dqm))
 			return;
-		dqm->sdma_bitmap |= (1ULL << q->sdma_id);
+		set_bit(q->sdma_id, dqm->sdma_bitmap);
 	} else if (q->properties.type == KFD_QUEUE_TYPE_SDMA_XGMI) {
 		if (q->sdma_id >= get_num_xgmi_sdma_queues(dqm))
 			return;
-		dqm->xgmi_sdma_bitmap |= (1ULL << q->sdma_id);
+		set_bit(q->sdma_id, dqm->xgmi_sdma_bitmap);
 	}
 }
 
@@ -1106,15 +1711,16 @@ static int set_sched_resources(struct device_queue_manager *dqm)
 {
 	int i, mec;
 	struct scheduling_resources res;
+	struct device *dev = dqm->dev->adev->dev;
 
-	res.vmid_mask = dqm->dev->shared_resources.compute_vmid_bitmap;
+	res.vmid_mask = dqm->dev->compute_vmid_bitmap;
 
 	res.queue_mask = 0;
-	for (i = 0; i < KGD_MAX_QUEUES; ++i) {
-		mec = (i / dqm->dev->shared_resources.num_queue_per_pipe)
-			/ dqm->dev->shared_resources.num_pipe_per_mec;
+	for (i = 0; i < AMDGPU_MAX_QUEUES; ++i) {
+		mec = (i / dqm->dev->kfd->shared_resources.num_queue_per_pipe)
+			/ dqm->dev->kfd->shared_resources.num_pipe_per_mec;
 
-		if (!test_bit(i, dqm->dev->shared_resources.cp_queue_bitmap))
+		if (!test_bit(i, dqm->dev->kfd->shared_resources.cp_queue_bitmap))
 			continue;
 
 		/* only acquire queues from the first MEC */
@@ -1126,13 +1732,13 @@ static int set_sched_resources(struct device_queue_manager *dqm)
 		 * definition of res.queue_mask needs updating
 		 */
 		if (WARN_ON(i >= (sizeof(res.queue_mask)*8))) {
-			pr_err("Invalid queue enabled by amdgpu: %d\n", i);
+			dev_err(dev, "Invalid queue enabled by amdgpu: %d\n", i);
 			break;
 		}
 
 		res.queue_mask |= 1ull
 			<< amdgpu_queue_mask_bit_to_set_resource_bit(
-				(struct amdgpu_device *)dqm->dev->kgd, i);
+				dqm->dev->adev, i);
 	}
 	res.gws_mask = ~0ull;
 	res.oac_mask = res.gds_heap_base = res.gds_heap_size = 0;
@@ -1147,9 +1753,6 @@ static int set_sched_resources(struct device_queue_manager *dqm)
 
 static int initialize_cpsch(struct device_queue_manager *dqm)
 {
-	uint64_t num_sdma_queues;
-	uint64_t num_xgmi_sdma_queues;
-
 	pr_debug("num of pipes: %d\n", get_pipes_per_mec(dqm));
 
 	mutex_init(&dqm->lock_hidden);
@@ -1158,39 +1761,89 @@ static int initialize_cpsch(struct device_queue_manager *dqm)
 	dqm->active_cp_queue_count = 0;
 	dqm->gws_queue_count = 0;
 	dqm->active_runlist = false;
+	dqm->trap_debug_vmid = 0;
 
-	num_sdma_queues = get_num_sdma_queues(dqm);
-	if (num_sdma_queues >= BITS_PER_TYPE(dqm->sdma_bitmap))
-		dqm->sdma_bitmap = ULLONG_MAX;
-	else
-		dqm->sdma_bitmap = (BIT_ULL(num_sdma_queues) - 1);
+	init_sdma_bitmaps(dqm);
 
-	num_xgmi_sdma_queues = get_num_xgmi_sdma_queues(dqm);
-	if (num_xgmi_sdma_queues >= BITS_PER_TYPE(dqm->xgmi_sdma_bitmap))
-		dqm->xgmi_sdma_bitmap = ULLONG_MAX;
+	update_dqm_wait_times(dqm);
+	return 0;
+}
+
+/* halt_cpsch:
+ * Unmap queues so the schedule doesn't continue remaining jobs in the queue.
+ * Then set dqm->sched_halt so queues don't map to runlist until unhalt_cpsch
+ * is called.
+ */
+static int halt_cpsch(struct device_queue_manager *dqm)
+{
+	int ret = 0;
+
+	dqm_lock(dqm);
+	if (!dqm->sched_running) {
+		dqm_unlock(dqm);
+		return 0;
+	}
+
+	WARN_ONCE(dqm->sched_halt, "Scheduling is already on halt\n");
+
+	if (!dqm->is_hws_hang) {
+		if (!dqm->dev->kfd->shared_resources.enable_mes)
+			ret = unmap_queues_cpsch(dqm,
+						 KFD_UNMAP_QUEUES_FILTER_ALL_QUEUES, 0,
+				USE_DEFAULT_GRACE_PERIOD, false);
+		else
+			ret = remove_all_kfd_queues_mes(dqm);
+	}
+	dqm->sched_halt = true;
+	dqm_unlock(dqm);
+
+	return ret;
+}
+
+/* unhalt_cpsch
+ * Unset dqm->sched_halt and map queues back to runlist
+ */
+static int unhalt_cpsch(struct device_queue_manager *dqm)
+{
+	int ret = 0;
+
+	dqm_lock(dqm);
+	if (!dqm->sched_running || !dqm->sched_halt) {
+		WARN_ONCE(!dqm->sched_halt, "Scheduling is not on halt.\n");
+		dqm_unlock(dqm);
+		return 0;
+	}
+	dqm->sched_halt = false;
+	if (!dqm->dev->kfd->shared_resources.enable_mes)
+		ret = execute_queues_cpsch(dqm,
+					   KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES,
+			0, USE_DEFAULT_GRACE_PERIOD);
 	else
-		dqm->xgmi_sdma_bitmap = (BIT_ULL(num_xgmi_sdma_queues) - 1);
+		ret = add_all_kfd_queues_mes(dqm);
 
-	INIT_WORK(&dqm->hw_exception_work, kfd_process_hw_exception);
+	dqm_unlock(dqm);
 
-	return 0;
+	return ret;
 }
 
 static int start_cpsch(struct device_queue_manager *dqm)
 {
-	int retval;
+	struct device *dev = dqm->dev->adev->dev;
+	int retval, num_hw_queue_slots;
 
 	retval = 0;
 
 	dqm_lock(dqm);
-	retval = pm_init(&dqm->packet_mgr, dqm);
-	if (retval)
-		goto fail_packet_manager_init;
 
-	retval = set_sched_resources(dqm);
-	if (retval)
-		goto fail_set_sched_resources;
+	if (!dqm->dev->kfd->shared_resources.enable_mes) {
+		retval = pm_init(&dqm->packet_mgr, dqm);
+		if (retval)
+			goto fail_packet_manager_init;
 
+		retval = set_sched_resources(dqm);
+		if (retval)
+			goto fail_set_sched_resources;
+	}
 	pr_debug("Allocating fence memory\n");
 
 	/* allocate fence memory on the gart */
@@ -1206,16 +1859,37 @@ static int start_cpsch(struct device_queue_manager *dqm)
 	init_interrupts(dqm);
 
 	/* clear hang status when driver try to start the hw scheduler */
-	dqm->is_hws_hang = false;
-	dqm->is_resetting = false;
 	dqm->sched_running = true;
-	execute_queues_cpsch(dqm, KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0);
+
+	if (!dqm->dev->kfd->shared_resources.enable_mes) {
+		if (pm_config_dequeue_wait_counts(&dqm->packet_mgr,
+				KFD_DEQUEUE_WAIT_INIT, 0 /* unused */))
+			dev_err(dev, "Setting optimized dequeue wait failed. Using default values\n");
+		execute_queues_cpsch(dqm, KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0, USE_DEFAULT_GRACE_PERIOD);
+	}
+
+	/* setup per-queue reset detection buffer  */
+	num_hw_queue_slots =  dqm->dev->kfd->shared_resources.num_queue_per_pipe *
+			      dqm->dev->kfd->shared_resources.num_pipe_per_mec *
+			      NUM_XCC(dqm->dev->xcc_mask);
+
+	dqm->detect_hang_info_size = num_hw_queue_slots * sizeof(struct dqm_detect_hang_info);
+	dqm->detect_hang_info = kzalloc(dqm->detect_hang_info_size, GFP_KERNEL);
+
+	if (!dqm->detect_hang_info) {
+		retval = -ENOMEM;
+		goto fail_detect_hang_buffer;
+	}
+
 	dqm_unlock(dqm);
 
 	return 0;
+fail_detect_hang_buffer:
+	kfd_gtt_sa_free(dqm->dev, dqm->fence_mem);
 fail_allocate_vidmem:
 fail_set_sched_resources:
-	pm_uninit(&dqm->packet_mgr, false);
+	if (!dqm->dev->kfd->shared_resources.enable_mes)
+		pm_uninit(&dqm->packet_mgr);
 fail_packet_manager_init:
 	dqm_unlock(dqm);
 	return retval;
@@ -1223,7 +1897,7 @@ fail_packet_manager_init:
 
 static int stop_cpsch(struct device_queue_manager *dqm)
 {
-	bool hanging;
+	int ret = 0;
 
 	dqm_lock(dqm);
 	if (!dqm->sched_running) {
@@ -1231,18 +1905,25 @@ static int stop_cpsch(struct device_queue_manager *dqm)
 		return 0;
 	}
 
-	if (!dqm->is_hws_hang)
-		unmap_queues_cpsch(dqm, KFD_UNMAP_QUEUES_FILTER_ALL_QUEUES, 0);
-	hanging = dqm->is_hws_hang || dqm->is_resetting;
+	if (!dqm->dev->kfd->shared_resources.enable_mes)
+		ret = unmap_queues_cpsch(dqm, KFD_UNMAP_QUEUES_FILTER_ALL_QUEUES,
+								0, USE_DEFAULT_GRACE_PERIOD, false);
+	else
+		ret = remove_all_kfd_queues_mes(dqm);
+
 	dqm->sched_running = false;
 
-	pm_release_ib(&dqm->packet_mgr);
+	if (!dqm->dev->kfd->shared_resources.enable_mes)
+		pm_release_ib(&dqm->packet_mgr);
 
 	kfd_gtt_sa_free(dqm->dev, dqm->fence_mem);
-	pm_uninit(&dqm->packet_mgr, hanging);
+	if (!dqm->dev->kfd->shared_resources.enable_mes)
+		pm_uninit(&dqm->packet_mgr);
+	kfree(dqm->detect_hang_info);
+	dqm->detect_hang_info = NULL;
 	dqm_unlock(dqm);
 
-	return 0;
+	return ret;
 }
 
 static int create_kernel_queue_cpsch(struct device_queue_manager *dqm,
@@ -1266,9 +1947,10 @@ static int create_kernel_queue_cpsch(struct device_queue_manager *dqm,
 			dqm->total_queue_count);
 
 	list_add(&kq->list, &qpd->priv_queue_list);
-	increment_queue_count(dqm, kq->queue->properties.type);
+	increment_queue_count(dqm, qpd, kq->queue);
 	qpd->is_debug = true;
-	execute_queues_cpsch(dqm, KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0);
+	execute_queues_cpsch(dqm, KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0,
+			USE_DEFAULT_GRACE_PERIOD);
 	dqm_unlock(dqm);
 
 	return 0;
@@ -1280,9 +1962,10 @@ static void destroy_kernel_queue_cpsch(struct device_queue_manager *dqm,
 {
 	dqm_lock(dqm);
 	list_del(&kq->list);
-	decrement_queue_count(dqm, kq->queue->properties.type);
+	decrement_queue_count(dqm, qpd, kq->queue);
 	qpd->is_debug = false;
-	execute_queues_cpsch(dqm, KFD_UNMAP_QUEUES_FILTER_ALL_QUEUES, 0);
+	execute_queues_cpsch(dqm, KFD_UNMAP_QUEUES_FILTER_ALL_QUEUES, 0,
+			USE_DEFAULT_GRACE_PERIOD);
 	/*
 	 * Unconditionally decrement this counter, regardless of the queue's
 	 * type.
@@ -1294,7 +1977,9 @@ static void destroy_kernel_queue_cpsch(struct device_queue_manager *dqm,
 }
 
 static int create_queue_cpsch(struct device_queue_manager *dqm, struct queue *q,
-			struct qcm_process_device *qpd)
+			struct qcm_process_device *qpd,
+			const struct kfd_criu_queue_priv_data *qd,
+			const void *restore_mqd, const void *restore_ctl_stack)
 {
 	int retval;
 	struct mqd_manager *mqd_mgr;
@@ -1307,15 +1992,16 @@ static int create_queue_cpsch(struct device_queue_manager *dqm, struct queue *q,
 	}
 
 	if (q->properties.type == KFD_QUEUE_TYPE_SDMA ||
-		q->properties.type == KFD_QUEUE_TYPE_SDMA_XGMI) {
+		q->properties.type == KFD_QUEUE_TYPE_SDMA_XGMI ||
+		q->properties.type == KFD_QUEUE_TYPE_SDMA_BY_ENG_ID) {
 		dqm_lock(dqm);
-		retval = allocate_sdma_queue(dqm, q);
+		retval = allocate_sdma_queue(dqm, q, qd ? &qd->sdma_id : NULL);
 		dqm_unlock(dqm);
 		if (retval)
 			goto out;
 	}
 
-	retval = allocate_doorbell(qpd, q);
+	retval = allocate_doorbell(qpd, q, qd ? &qd->doorbell_id : NULL);
 	if (retval)
 		goto out_deallocate_sdma_queue;
 
@@ -1340,17 +2026,30 @@ static int create_queue_cpsch(struct device_queue_manager *dqm, struct queue *q,
 	 * updates the is_evicted flag but is a no-op otherwise.
 	 */
 	q->properties.is_evicted = !!qpd->evicted;
-	mqd_mgr->init_mqd(mqd_mgr, &q->mqd, q->mqd_mem_obj,
-				&q->gart_mqd_addr, &q->properties);
+	q->properties.is_dbg_wa = qpd->pqm->process->debug_trap_enabled &&
+				  kfd_dbg_has_cwsr_workaround(q->device);
+
+	if (qd)
+		mqd_mgr->restore_mqd(mqd_mgr, &q->mqd, q->mqd_mem_obj, &q->gart_mqd_addr,
+				     &q->properties, restore_mqd, restore_ctl_stack,
+				     qd->ctl_stack_size);
+	else
+		mqd_mgr->init_mqd(mqd_mgr, &q->mqd, q->mqd_mem_obj,
+					&q->gart_mqd_addr, &q->properties);
 
 	list_add(&q->list, &qpd->queues_list);
 	qpd->queue_count++;
 
 	if (q->properties.is_active) {
-		increment_queue_count(dqm, q->properties.type);
+		increment_queue_count(dqm, qpd, q);
 
-		execute_queues_cpsch(dqm,
-				KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0);
+		if (!dqm->dev->kfd->shared_resources.enable_mes)
+			retval = execute_queues_cpsch(dqm,
+					KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0, USE_DEFAULT_GRACE_PERIOD);
+		else
+			retval = add_queue_mes(dqm, q, qpd);
+		if (retval)
+			goto cleanup_queue;
 	}
 
 	/*
@@ -1365,6 +2064,13 @@ static int create_queue_cpsch(struct device_queue_manager *dqm, struct queue *q,
 	dqm_unlock(dqm);
 	return retval;
 
+cleanup_queue:
+	qpd->queue_count--;
+	list_del(&q->list);
+	if (q->properties.is_active)
+		decrement_queue_count(dqm, qpd, q);
+	mqd_mgr->free_mqd(mqd_mgr, q->mqd, q->mqd_mem_obj);
+	dqm_unlock(dqm);
 out_deallocate_doorbell:
 	deallocate_doorbell(qpd, q);
 out_deallocate_sdma_queue:
@@ -1378,15 +2084,22 @@ out:
 	return retval;
 }
 
-int amdkfd_fence_wait_timeout(uint64_t *fence_addr,
-				uint64_t fence_value,
-				unsigned int timeout_ms)
+int amdkfd_fence_wait_timeout(struct device_queue_manager *dqm,
+			      uint64_t fence_value,
+			      unsigned int timeout_ms)
 {
 	unsigned long end_jiffies = msecs_to_jiffies(timeout_ms) + jiffies;
+	struct device *dev = dqm->dev->adev->dev;
+	uint64_t *fence_addr = dqm->fence_addr;
 
 	while (*fence_addr != fence_value) {
+		/* Fatal err detected, this response won't come */
+		if (amdgpu_amdkfd_is_fed(dqm->dev->adev) ||
+		    amdgpu_in_reset(dqm->dev->adev))
+			return -EIO;
+
 		if (time_after(jiffies, end_jiffies)) {
-			pr_err("qcm fence wait loop timeout expired\n");
+			dev_err(dev, "qcm fence wait loop timeout expired\n");
 			/* In HWS case, this is used to halt the driver thread
 			 * in order not to mess up CP states before doing
 			 * scandumps for FW debugging.
@@ -1405,9 +2118,10 @@ int amdkfd_fence_wait_timeout(uint64_t *fence_addr,
 /* dqm->lock mutex has to be locked before calling this function */
 static int map_queues_cpsch(struct device_queue_manager *dqm)
 {
+	struct device *dev = dqm->dev->adev->dev;
 	int retval;
 
-	if (!dqm->sched_running)
+	if (!dqm->sched_running || dqm->sched_halt)
 		return 0;
 	if (dqm->active_queue_count <= 0 || dqm->processes_count <= 0)
 		return 0;
@@ -1417,7 +2131,7 @@ static int map_queues_cpsch(struct device_queue_manager *dqm)
 	retval = pm_send_runlist(&dqm->packet_mgr, &dqm->queues);
 	pr_debug("%s sent runlist\n", __func__);
 	if (retval) {
-		pr_err("failed to execute runlist\n");
+		dev_err(dev, "failed to execute runlist\n");
 		return retval;
 	}
 	dqm->active_runlist = true;
@@ -1425,42 +2139,288 @@ static int map_queues_cpsch(struct device_queue_manager *dqm)
 	return retval;
 }
 
-/* dqm->lock mutex has to be locked before calling this function */
+static void set_queue_as_reset(struct device_queue_manager *dqm, struct queue *q,
+			       struct qcm_process_device *qpd)
+{
+	struct kfd_process_device *pdd = qpd_to_pdd(qpd);
+
+	dev_err(dqm->dev->adev->dev, "queue id 0x%0x at pasid %d is reset\n",
+		q->properties.queue_id, pdd->process->lead_thread->pid);
+
+	pdd->has_reset_queue = true;
+	if (q->properties.is_active) {
+		q->properties.is_active = false;
+		decrement_queue_count(dqm, qpd, q);
+	}
+}
+
+static int detect_queue_hang(struct device_queue_manager *dqm)
+{
+	int i;
+
+	/* detect should be used only in dqm locked queue reset */
+	if (WARN_ON(dqm->detect_hang_count > 0))
+		return 0;
+
+	memset(dqm->detect_hang_info, 0, dqm->detect_hang_info_size);
+
+	for (i = 0; i < AMDGPU_MAX_QUEUES; ++i) {
+		uint32_t mec, pipe, queue;
+		int xcc_id;
+
+		mec = (i / dqm->dev->kfd->shared_resources.num_queue_per_pipe)
+			/ dqm->dev->kfd->shared_resources.num_pipe_per_mec;
+
+		if (mec || !test_bit(i, dqm->dev->kfd->shared_resources.cp_queue_bitmap))
+			continue;
+
+		amdgpu_queue_mask_bit_to_mec_queue(dqm->dev->adev, i, &mec, &pipe, &queue);
+
+		for_each_inst(xcc_id, dqm->dev->xcc_mask) {
+			uint64_t queue_addr = dqm->dev->kfd2kgd->hqd_get_pq_addr(
+						dqm->dev->adev, pipe, queue, xcc_id);
+			struct dqm_detect_hang_info hang_info;
+
+			if (!queue_addr)
+				continue;
+
+			hang_info.pipe_id = pipe;
+			hang_info.queue_id = queue;
+			hang_info.xcc_id = xcc_id;
+			hang_info.queue_address = queue_addr;
+
+			dqm->detect_hang_info[dqm->detect_hang_count] = hang_info;
+			dqm->detect_hang_count++;
+		}
+	}
+
+	return dqm->detect_hang_count;
+}
+
+static struct queue *find_queue_by_address(struct device_queue_manager *dqm, uint64_t queue_address)
+{
+	struct device_process_node *cur;
+	struct qcm_process_device *qpd;
+	struct queue *q;
+
+	list_for_each_entry(cur, &dqm->queues, list) {
+		qpd = cur->qpd;
+		list_for_each_entry(q, &qpd->queues_list, list) {
+			if (queue_address == q->properties.queue_address)
+				return q;
+		}
+	}
+
+	return NULL;
+}
+
+static int reset_hung_queues(struct device_queue_manager *dqm)
+{
+	int r = 0, reset_count = 0, i;
+
+	if (!dqm->detect_hang_info || dqm->is_hws_hang)
+		return -EIO;
+
+	/* assume dqm locked. */
+	if (!detect_queue_hang(dqm))
+		return -ENOTRECOVERABLE;
+
+	for (i = 0; i < dqm->detect_hang_count; i++) {
+		struct dqm_detect_hang_info hang_info = dqm->detect_hang_info[i];
+		struct queue *q = find_queue_by_address(dqm, hang_info.queue_address);
+		struct kfd_process_device *pdd;
+		uint64_t queue_addr = 0;
+
+		if (!q) {
+			r = -ENOTRECOVERABLE;
+			goto reset_fail;
+		}
+
+		pdd = kfd_get_process_device_data(dqm->dev, q->process);
+		if (!pdd) {
+			r = -ENOTRECOVERABLE;
+			goto reset_fail;
+		}
+
+		queue_addr = dqm->dev->kfd2kgd->hqd_reset(dqm->dev->adev,
+				hang_info.pipe_id, hang_info.queue_id, hang_info.xcc_id,
+				KFD_UNMAP_LATENCY_MS);
+
+		/* either reset failed or we reset an unexpected queue. */
+		if (queue_addr != q->properties.queue_address) {
+			r = -ENOTRECOVERABLE;
+			goto reset_fail;
+		}
+
+		set_queue_as_reset(dqm, q, &pdd->qpd);
+		reset_count++;
+	}
+
+	if (reset_count == dqm->detect_hang_count)
+		kfd_signal_reset_event(dqm->dev);
+	else
+		r = -ENOTRECOVERABLE;
+
+reset_fail:
+	dqm->detect_hang_count = 0;
+
+	return r;
+}
+
+static bool sdma_has_hang(struct device_queue_manager *dqm)
+{
+	int engine_start = dqm->dev->node_id * get_num_all_sdma_engines(dqm);
+	int engine_end = engine_start + get_num_all_sdma_engines(dqm);
+	int num_queues_per_eng =  dqm->dev->kfd->device_info.num_sdma_queues_per_engine;
+	int i, j;
+
+	for (i = engine_start; i < engine_end; i++) {
+		for (j = 0; j < num_queues_per_eng; j++) {
+			if (!dqm->dev->kfd2kgd->hqd_sdma_get_doorbell(dqm->dev->adev, i, j))
+				continue;
+
+			return true;
+		}
+	}
+
+	return false;
+}
+
+static bool set_sdma_queue_as_reset(struct device_queue_manager *dqm,
+				    uint32_t doorbell_off)
+{
+	struct device_process_node *cur;
+	struct qcm_process_device *qpd;
+	struct queue *q;
+
+	list_for_each_entry(cur, &dqm->queues, list) {
+		qpd = cur->qpd;
+		list_for_each_entry(q, &qpd->queues_list, list) {
+			if ((q->properties.type == KFD_QUEUE_TYPE_SDMA ||
+			     q->properties.type == KFD_QUEUE_TYPE_SDMA_XGMI) &&
+			     q->properties.doorbell_off == doorbell_off) {
+				set_queue_as_reset(dqm, q, qpd);
+				return true;
+			}
+		}
+	}
+
+	return false;
+}
+
+static int reset_hung_queues_sdma(struct device_queue_manager *dqm)
+{
+	int engine_start = dqm->dev->node_id * get_num_all_sdma_engines(dqm);
+	int engine_end = engine_start + get_num_all_sdma_engines(dqm);
+	int num_queues_per_eng =  dqm->dev->kfd->device_info.num_sdma_queues_per_engine;
+	int r = 0, i, j;
+
+	if (dqm->is_hws_hang)
+		return -EIO;
+
+	/* Scan for hung HW queues and reset engine. */
+	dqm->detect_hang_count = 0;
+	for (i = engine_start; i < engine_end; i++) {
+		for (j = 0; j < num_queues_per_eng; j++) {
+			uint32_t doorbell_off =
+				dqm->dev->kfd2kgd->hqd_sdma_get_doorbell(dqm->dev->adev, i, j);
+
+			if (!doorbell_off)
+				continue;
+
+			/* Reset engine and check. */
+			if (amdgpu_sdma_reset_engine(dqm->dev->adev, i, false) ||
+			    dqm->dev->kfd2kgd->hqd_sdma_get_doorbell(dqm->dev->adev, i, j) ||
+			    !set_sdma_queue_as_reset(dqm, doorbell_off)) {
+				r = -ENOTRECOVERABLE;
+				goto reset_fail;
+			}
+
+			/* Should only expect one queue active per engine */
+			dqm->detect_hang_count++;
+			break;
+		}
+	}
+
+	/* Signal process reset */
+	if (dqm->detect_hang_count)
+		kfd_signal_reset_event(dqm->dev);
+	else
+		r = -ENOTRECOVERABLE;
+
+reset_fail:
+	dqm->detect_hang_count = 0;
+
+	return r;
+}
+
+static int reset_queues_on_hws_hang(struct device_queue_manager *dqm, bool is_sdma)
+{
+	struct amdgpu_device *adev = dqm->dev->adev;
+
+	while (halt_if_hws_hang)
+		schedule();
+
+	if (adev->debug_disable_gpu_ring_reset) {
+		dev_info_once(adev->dev,
+			      "%s queue hung, but ring reset disabled",
+			      is_sdma ? "sdma" : "compute");
+
+		return -EPERM;
+	}
+	if (!amdgpu_gpu_recovery)
+		return -ENOTRECOVERABLE;
+
+	return is_sdma ? reset_hung_queues_sdma(dqm) : reset_hung_queues(dqm);
+}
+
+/* dqm->lock mutex has to be locked before calling this function
+ *
+ * @grace_period: If USE_DEFAULT_GRACE_PERIOD then default wait time
+ *   for context switch latency. Lower values are used by debugger
+ *   since context switching are triggered at high frequency.
+ *   This is configured by setting CP_IQ_WAIT_TIME2.SCH_WAVE
+ *
+ */
 static int unmap_queues_cpsch(struct device_queue_manager *dqm,
 				enum kfd_unmap_queues_filter filter,
-				uint32_t filter_param)
+				uint32_t filter_param,
+				uint32_t grace_period,
+				bool reset)
 {
-	int retval = 0;
+	struct device *dev = dqm->dev->adev->dev;
 	struct mqd_manager *mqd_mgr;
+	int retval;
 
 	if (!dqm->sched_running)
 		return 0;
-	if (dqm->is_hws_hang || dqm->is_resetting)
-		return -EIO;
 	if (!dqm->active_runlist)
-		return retval;
+		return 0;
+	if (!down_read_trylock(&dqm->dev->adev->reset_domain->sem))
+		return -EIO;
+
+	if (grace_period != USE_DEFAULT_GRACE_PERIOD) {
+		retval = pm_config_dequeue_wait_counts(&dqm->packet_mgr,
+				KFD_DEQUEUE_WAIT_SET_SCH_WAVE, grace_period);
+		if (retval)
+			goto out;
+	}
 
-	retval = pm_send_unmap_queue(&dqm->packet_mgr, KFD_QUEUE_TYPE_COMPUTE,
-			filter, filter_param, false, 0);
+	retval = pm_send_unmap_queue(&dqm->packet_mgr, filter, filter_param, reset);
 	if (retval)
-		return retval;
+		goto out;
 
 	*dqm->fence_addr = KFD_FENCE_INIT;
+	mb();
 	pm_send_query_status(&dqm->packet_mgr, dqm->fence_gpu_addr,
 				KFD_FENCE_COMPLETED);
 	/* should be timed out */
-	retval = amdkfd_fence_wait_timeout(dqm->fence_addr, KFD_FENCE_COMPLETED,
-				queue_preemption_timeout_ms);
+	retval = amdkfd_fence_wait_timeout(dqm, KFD_FENCE_COMPLETED,
+					   queue_preemption_timeout_ms);
 	if (retval) {
-		pr_err("The cp might be in an unrecoverable state due to an unsuccessful queues preemption\n");
-		dqm->is_hws_hang = true;
-		/* It's possible we're detecting a HWS hang in the
-		 * middle of a GPU reset. No need to schedule another
-		 * reset in this case.
-		 */
-		if (!dqm->is_resetting)
-			schedule_work(&dqm->hw_exception_work);
-		return retval;
+		dev_err(dev, "The cp might be in an unrecoverable state due to an unsuccessful queues preemption\n");
+		kfd_hws_hang(dqm);
+		goto out;
 	}
 
 	/* In the current MEC firmware implementation, if compute queue
@@ -1472,33 +2432,91 @@ static int unmap_queues_cpsch(struct device_queue_manager *dqm,
 	 * check those fields
 	 */
 	mqd_mgr = dqm->mqd_mgrs[KFD_MQD_TYPE_HIQ];
-	if (mqd_mgr->read_doorbell_id(dqm->packet_mgr.priv_queue->queue->mqd)) {
-		pr_err("HIQ MQD's queue_doorbell_id0 is not 0, Queue preemption time out\n");
-		while (halt_if_hws_hang)
-			schedule();
-		return -ETIME;
+	if (mqd_mgr->check_preemption_failed(mqd_mgr, dqm->packet_mgr.priv_queue->queue->mqd) &&
+	    reset_queues_on_hws_hang(dqm, false))
+		goto reset_fail;
+
+	/* Check for SDMA hang and attempt SDMA reset */
+	if (sdma_has_hang(dqm) && reset_queues_on_hws_hang(dqm, true))
+		goto reset_fail;
+
+	/* We need to reset the grace period value for this device */
+	if (grace_period != USE_DEFAULT_GRACE_PERIOD) {
+		if (pm_config_dequeue_wait_counts(&dqm->packet_mgr,
+				KFD_DEQUEUE_WAIT_RESET, 0 /* unused */))
+			dev_err(dev, "Failed to reset grace period\n");
 	}
 
 	pm_release_ib(&dqm->packet_mgr);
 	dqm->active_runlist = false;
+out:
+	up_read(&dqm->dev->adev->reset_domain->sem);
+	return retval;
 
+reset_fail:
+	dqm->is_hws_hang = true;
+	kfd_hws_hang(dqm);
+	up_read(&dqm->dev->adev->reset_domain->sem);
+	return -ETIME;
+}
+
+/* only for compute queue */
+static int reset_queues_cpsch(struct device_queue_manager *dqm, uint16_t pasid)
+{
+	int retval;
+
+	dqm_lock(dqm);
+
+	retval = unmap_queues_cpsch(dqm, KFD_UNMAP_QUEUES_FILTER_BY_PASID,
+			pasid, USE_DEFAULT_GRACE_PERIOD, true);
+
+	dqm_unlock(dqm);
 	return retval;
 }
 
 /* dqm->lock mutex has to be locked before calling this function */
 static int execute_queues_cpsch(struct device_queue_manager *dqm,
 				enum kfd_unmap_queues_filter filter,
-				uint32_t filter_param)
+				uint32_t filter_param,
+				uint32_t grace_period)
 {
 	int retval;
 
-	if (dqm->is_hws_hang)
+	if (!down_read_trylock(&dqm->dev->adev->reset_domain->sem))
 		return -EIO;
-	retval = unmap_queues_cpsch(dqm, filter, filter_param);
-	if (retval)
-		return retval;
+	retval = unmap_queues_cpsch(dqm, filter, filter_param, grace_period, false);
+	if (!retval)
+		retval = map_queues_cpsch(dqm);
+	up_read(&dqm->dev->adev->reset_domain->sem);
+	return retval;
+}
+
+static int wait_on_destroy_queue(struct device_queue_manager *dqm,
+				 struct queue *q)
+{
+	struct kfd_process_device *pdd = kfd_get_process_device_data(q->device,
+								q->process);
+	int ret = 0;
+
+	if (WARN_ON(!pdd))
+		return ret;
+
+	if (pdd->qpd.is_debug)
+		return ret;
 
-	return map_queues_cpsch(dqm);
+	q->properties.is_being_destroyed = true;
+
+	if (pdd->process->debug_trap_enabled && q->properties.is_suspended) {
+		dqm_unlock(dqm);
+		mutex_unlock(&q->process->mutex);
+		ret = wait_event_interruptible(dqm->destroy_wait,
+						!q->properties.is_suspended);
+
+		mutex_lock(&q->process->mutex);
+		dqm_lock(dqm);
+	}
+
+	return ret;
 }
 
 static int destroy_queue_cpsch(struct device_queue_manager *dqm,
@@ -1509,6 +2527,7 @@ static int destroy_queue_cpsch(struct device_queue_manager *dqm,
 	struct mqd_manager *mqd_mgr;
 	uint64_t sdma_val = 0;
 	struct kfd_process_device *pdd = qpd_to_pdd(qpd);
+	struct device *dev = dqm->dev->adev->dev;
 
 	/* Get the SDMA queue stats */
 	if ((q->properties.type == KFD_QUEUE_TYPE_SDMA) ||
@@ -1516,15 +2535,20 @@ static int destroy_queue_cpsch(struct device_queue_manager *dqm,
 		retval = read_sdma_queue_counter((uint64_t __user *)q->properties.read_ptr,
 							&sdma_val);
 		if (retval)
-			pr_err("Failed to read SDMA queue counter for queue: %d\n",
+			dev_err(dev, "Failed to read SDMA queue counter for queue: %d\n",
 				q->properties.queue_id);
 	}
 
-	retval = 0;
-
 	/* remove queue from list to prevent rescheduling after preemption */
 	dqm_lock(dqm);
 
+	retval = wait_on_destroy_queue(dqm, q);
+
+	if (retval) {
+		dqm_unlock(dqm);
+		return retval;
+	}
+
 	if (qpd->is_debug) {
 		/*
 		 * error, currently we do not allow to destroy a queue
@@ -1546,19 +2570,21 @@ static int destroy_queue_cpsch(struct device_queue_manager *dqm,
 		pdd->sdma_past_activity_counter += sdma_val;
 	}
 
-	list_del(&q->list);
-	qpd->queue_count--;
 	if (q->properties.is_active) {
-		decrement_queue_count(dqm, q->properties.type);
-		retval = execute_queues_cpsch(dqm,
-				KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0);
-		if (retval == -ETIME)
-			qpd->reset_wavefronts = true;
-		if (q->properties.is_gws) {
-			dqm->gws_queue_count--;
-			qpd->mapped_gws_queue = false;
+		decrement_queue_count(dqm, qpd, q);
+		q->properties.is_active = false;
+		if (!dqm->dev->kfd->shared_resources.enable_mes) {
+			retval = execute_queues_cpsch(dqm,
+						      KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0,
+						      USE_DEFAULT_GRACE_PERIOD);
+			if (retval == -ETIME)
+				qpd->reset_wavefronts = true;
+		} else {
+			retval = remove_queue_mes(dqm, q, qpd);
 		}
 	}
+	list_del(&q->list);
+	qpd->queue_count--;
 
 	/*
 	 * Unconditionally decrement this counter, regardless of the queue's
@@ -1570,7 +2596,14 @@ static int destroy_queue_cpsch(struct device_queue_manager *dqm,
 
 	dqm_unlock(dqm);
 
-	/* Do free_mqd after dqm_unlock(dqm) to avoid circular locking */
+	/*
+	 * Do free_mqd and raise delete event after dqm_unlock(dqm) to avoid
+	 * circular locking
+	 */
+	kfd_dbg_ev_raise(KFD_EC_MASK(EC_DEVICE_QUEUE_DELETE),
+				qpd->pqm->process, q->device,
+				-1, false, NULL, 0);
+
 	mqd_mgr->free_mqd(mqd_mgr, q->mqd, q->mqd_mem_obj);
 
 	return retval;
@@ -1581,20 +2614,13 @@ failed_try_destroy_debugged_queue:
 	return retval;
 }
 
-/*
- * Low bits must be 0000/FFFF as required by HW, high bits must be 0 to
- * stay in user mode.
- */
-#define APE1_FIXED_BITS_MASK 0xFFFF80000000FFFFULL
-/* APE1 limit is inclusive and 64K aligned. */
-#define APE1_LIMIT_ALIGNMENT 0xFFFF
-
 static bool set_cache_memory_policy(struct device_queue_manager *dqm,
 				   struct qcm_process_device *qpd,
 				   enum cache_policy default_policy,
 				   enum cache_policy alternate_policy,
 				   void __user *alternate_aperture_base,
-				   uint64_t alternate_aperture_size)
+				   uint64_t alternate_aperture_size,
+				   u32 misc_process_properties)
 {
 	bool retval = true;
 
@@ -1603,41 +2629,17 @@ static bool set_cache_memory_policy(struct device_queue_manager *dqm,
 
 	dqm_lock(dqm);
 
-	if (alternate_aperture_size == 0) {
-		/* base > limit disables APE1 */
-		qpd->sh_mem_ape1_base = 1;
-		qpd->sh_mem_ape1_limit = 0;
-	} else {
-		/*
-		 * In FSA64, APE1_Base[63:0] = { 16{SH_MEM_APE1_BASE[31]},
-		 *			SH_MEM_APE1_BASE[31:0], 0x0000 }
-		 * APE1_Limit[63:0] = { 16{SH_MEM_APE1_LIMIT[31]},
-		 *			SH_MEM_APE1_LIMIT[31:0], 0xFFFF }
-		 * Verify that the base and size parameters can be
-		 * represented in this format and convert them.
-		 * Additionally restrict APE1 to user-mode addresses.
-		 */
-
-		uint64_t base = (uintptr_t)alternate_aperture_base;
-		uint64_t limit = base + alternate_aperture_size - 1;
-
-		if (limit <= base || (base & APE1_FIXED_BITS_MASK) != 0 ||
-		   (limit & APE1_FIXED_BITS_MASK) != APE1_LIMIT_ALIGNMENT) {
-			retval = false;
-			goto out;
-		}
-
-		qpd->sh_mem_ape1_base = base >> 16;
-		qpd->sh_mem_ape1_limit = limit >> 16;
-	}
-
 	retval = dqm->asic_ops.set_cache_memory_policy(
 			dqm,
 			qpd,
 			default_policy,
 			alternate_policy,
 			alternate_aperture_base,
-			alternate_aperture_size);
+			alternate_aperture_size,
+			misc_process_properties);
+
+	if (retval)
+		goto out;
 
 	if ((dqm->sched_policy == KFD_SCHED_POLICY_NO_HWS) && (qpd->vmid != 0))
 		program_sh_mem_settings(dqm, qpd);
@@ -1712,7 +2714,7 @@ static int get_wave_state(struct device_queue_manager *dqm,
 	mqd_mgr = dqm->mqd_mgrs[KFD_MQD_TYPE_CP];
 
 	if (q->properties.type != KFD_QUEUE_TYPE_COMPUTE ||
-	    q->properties.is_active || !q->device->cwsr_enabled ||
+	    q->properties.is_active || !q->device->kfd->cwsr_enabled ||
 	    !mqd_mgr->get_wave_state) {
 		dqm_unlock(dqm);
 		return -EINVAL;
@@ -1725,8 +2727,58 @@ static int get_wave_state(struct device_queue_manager *dqm,
 	 * and the queue should be protected against destruction by the process
 	 * lock.
 	 */
-	return mqd_mgr->get_wave_state(mqd_mgr, q->mqd, ctl_stack,
-			ctl_stack_used_size, save_area_used_size);
+	return mqd_mgr->get_wave_state(mqd_mgr, q->mqd, &q->properties,
+			ctl_stack, ctl_stack_used_size, save_area_used_size);
+}
+
+static void get_queue_checkpoint_info(struct device_queue_manager *dqm,
+			const struct queue *q,
+			u32 *mqd_size,
+			u32 *ctl_stack_size)
+{
+	struct mqd_manager *mqd_mgr;
+	enum KFD_MQD_TYPE mqd_type =
+			get_mqd_type_from_queue_type(q->properties.type);
+
+	dqm_lock(dqm);
+	mqd_mgr = dqm->mqd_mgrs[mqd_type];
+	*mqd_size = mqd_mgr->mqd_size * NUM_XCC(mqd_mgr->dev->xcc_mask);
+	*ctl_stack_size = 0;
+
+	if (q->properties.type == KFD_QUEUE_TYPE_COMPUTE && mqd_mgr->get_checkpoint_info)
+		mqd_mgr->get_checkpoint_info(mqd_mgr, q->mqd, ctl_stack_size);
+
+	dqm_unlock(dqm);
+}
+
+static int checkpoint_mqd(struct device_queue_manager *dqm,
+			  const struct queue *q,
+			  void *mqd,
+			  void *ctl_stack)
+{
+	struct mqd_manager *mqd_mgr;
+	int r = 0;
+	enum KFD_MQD_TYPE mqd_type =
+			get_mqd_type_from_queue_type(q->properties.type);
+
+	dqm_lock(dqm);
+
+	if (q->properties.is_active || !q->device->kfd->cwsr_enabled) {
+		r = -EINVAL;
+		goto dqm_unlock;
+	}
+
+	mqd_mgr = dqm->mqd_mgrs[mqd_type];
+	if (!mqd_mgr->checkpoint_mqd) {
+		r = -EOPNOTSUPP;
+		goto dqm_unlock;
+	}
+
+	mqd_mgr->checkpoint_mqd(mqd_mgr, q->mqd, mqd, ctl_stack);
+
+dqm_unlock:
+	dqm_unlock(dqm);
+	return r;
 }
 
 static int process_termination_cpsch(struct device_queue_manager *dqm,
@@ -1734,6 +2786,7 @@ static int process_termination_cpsch(struct device_queue_manager *dqm,
 {
 	int retval;
 	struct queue *q;
+	struct device *dev = dqm->dev->adev->dev;
 	struct kernel_queue *kq, *kq_next;
 	struct mqd_manager *mqd_mgr;
 	struct device_process_node *cur, *next_dpn;
@@ -1748,7 +2801,7 @@ static int process_termination_cpsch(struct device_queue_manager *dqm,
 	/* Clean all kernel queues */
 	list_for_each_entry_safe(kq, kq_next, &qpd->priv_queue_list, list) {
 		list_del(&kq->list);
-		decrement_queue_count(dqm, kq->queue->properties.type);
+		decrement_queue_count(dqm, qpd, kq->queue);
 		qpd->is_debug = false;
 		dqm->total_queue_count--;
 		filter = KFD_UNMAP_QUEUES_FILTER_ALL_QUEUES;
@@ -1762,10 +2815,13 @@ static int process_termination_cpsch(struct device_queue_manager *dqm,
 			deallocate_sdma_queue(dqm, q);
 
 		if (q->properties.is_active) {
-			decrement_queue_count(dqm, q->properties.type);
-			if (q->properties.is_gws) {
-				dqm->gws_queue_count--;
-				qpd->mapped_gws_queue = false;
+			decrement_queue_count(dqm, qpd, q);
+
+			if (dqm->dev->kfd->shared_resources.enable_mes) {
+				retval = remove_queue_mes(dqm, q, qpd);
+				if (retval)
+					dev_err(dev, "Failed to remove queue %d\n",
+						q->properties.queue_id);
 			}
 		}
 
@@ -1783,11 +2839,15 @@ static int process_termination_cpsch(struct device_queue_manager *dqm,
 		}
 	}
 
-	retval = execute_queues_cpsch(dqm, filter, 0);
-	if ((!dqm->is_hws_hang) && (retval || qpd->reset_wavefronts)) {
+	if (!dqm->dev->kfd->shared_resources.enable_mes)
+		retval = execute_queues_cpsch(dqm, filter, 0, USE_DEFAULT_GRACE_PERIOD);
+
+	if ((retval || qpd->reset_wavefronts) &&
+	    down_read_trylock(&dqm->dev->adev->reset_domain->sem)) {
 		pr_warn("Resetting wave fronts (cpsch) on dev %p\n", dqm->dev);
 		dbgdev_wave_reset_wavefronts(dqm->dev, qpd->pqm->process);
 		qpd->reset_wavefronts = false;
+		up_read(&dqm->dev->adev->reset_domain->sem);
 	}
 
 	/* Lastly, free mqd resources.
@@ -1817,12 +2877,13 @@ static int process_termination_cpsch(struct device_queue_manager *dqm,
 static int init_mqd_managers(struct device_queue_manager *dqm)
 {
 	int i, j;
+	struct device *dev = dqm->dev->adev->dev;
 	struct mqd_manager *mqd_mgr;
 
 	for (i = 0; i < KFD_MQD_TYPE_MAX; i++) {
 		mqd_mgr = dqm->asic_ops.mqd_manager_init(i, dqm->dev);
 		if (!mqd_mgr) {
-			pr_err("mqd manager [%d] initialization failed\n", i);
+			dev_err(dev, "mqd manager [%d] initialization failed\n", i);
 			goto out_free;
 		}
 		dqm->mqd_mgrs[i] = mqd_mgr;
@@ -1843,21 +2904,22 @@ out_free:
 static int allocate_hiq_sdma_mqd(struct device_queue_manager *dqm)
 {
 	int retval;
-	struct kfd_dev *dev = dqm->dev;
+	struct kfd_node *dev = dqm->dev;
 	struct kfd_mem_obj *mem_obj = &dqm->hiq_sdma_mqd;
 	uint32_t size = dqm->mqd_mgrs[KFD_MQD_TYPE_SDMA]->mqd_size *
 		get_num_all_sdma_engines(dqm) *
-		dev->device_info->num_sdma_queues_per_engine +
-		dqm->mqd_mgrs[KFD_MQD_TYPE_HIQ]->mqd_size;
+		dev->kfd->device_info.num_sdma_queues_per_engine +
+		(dqm->mqd_mgrs[KFD_MQD_TYPE_HIQ]->mqd_size *
+		NUM_XCC(dqm->dev->xcc_mask));
 
-	retval = amdgpu_amdkfd_alloc_gtt_mem(dev->kgd, size,
+	retval = amdgpu_amdkfd_alloc_gtt_mem(dev->adev, size,
 		&(mem_obj->gtt_mem), &(mem_obj->gpu_addr),
 		(void *)&(mem_obj->cpu_ptr), false);
 
 	return retval;
 }
 
-struct device_queue_manager *device_queue_manager_init(struct kfd_dev *dev)
+struct device_queue_manager *device_queue_manager_init(struct kfd_node *dev)
 {
 	struct device_queue_manager *dqm;
 
@@ -1867,7 +2929,7 @@ struct device_queue_manager *device_queue_manager_init(struct kfd_dev *dev)
 	if (!dqm)
 		return NULL;
 
-	switch (dev->device_info->asic_family) {
+	switch (dev->adev->asic_type) {
 	/* HWS is not available on Hawaii. */
 	case CHIP_HAWAII:
 	/* HWS depends on CWSR for timely dequeue. CWSR is not
@@ -1892,7 +2954,8 @@ struct device_queue_manager *device_queue_manager_init(struct kfd_dev *dev)
 		dqm->ops.initialize = initialize_cpsch;
 		dqm->ops.start = start_cpsch;
 		dqm->ops.stop = stop_cpsch;
-		dqm->ops.pre_reset = pre_reset;
+		dqm->ops.halt = halt_cpsch;
+		dqm->ops.unhalt = unhalt_cpsch;
 		dqm->ops.destroy_queue = destroy_queue_cpsch;
 		dqm->ops.update_queue = update_queue;
 		dqm->ops.register_process = register_process;
@@ -1905,12 +2968,14 @@ struct device_queue_manager *device_queue_manager_init(struct kfd_dev *dev)
 		dqm->ops.evict_process_queues = evict_process_queues_cpsch;
 		dqm->ops.restore_process_queues = restore_process_queues_cpsch;
 		dqm->ops.get_wave_state = get_wave_state;
+		dqm->ops.reset_queues = reset_queues_cpsch;
+		dqm->ops.get_queue_checkpoint_info = get_queue_checkpoint_info;
+		dqm->ops.checkpoint_mqd = checkpoint_mqd;
 		break;
 	case KFD_SCHED_POLICY_NO_HWS:
 		/* initialize dqm for no cp scheduling */
 		dqm->ops.start = start_nocpsch;
 		dqm->ops.stop = stop_nocpsch;
-		dqm->ops.pre_reset = pre_reset;
 		dqm->ops.create_queue = create_queue_nocpsch;
 		dqm->ops.destroy_queue = destroy_queue_nocpsch;
 		dqm->ops.update_queue = update_queue;
@@ -1924,116 +2989,662 @@ struct device_queue_manager *device_queue_manager_init(struct kfd_dev *dev)
 		dqm->ops.restore_process_queues =
 			restore_process_queues_nocpsch;
 		dqm->ops.get_wave_state = get_wave_state;
+		dqm->ops.get_queue_checkpoint_info = get_queue_checkpoint_info;
+		dqm->ops.checkpoint_mqd = checkpoint_mqd;
 		break;
 	default:
-		pr_err("Invalid scheduling policy %d\n", dqm->sched_policy);
+		dev_err(dev->adev->dev, "Invalid scheduling policy %d\n", dqm->sched_policy);
 		goto out_free;
 	}
 
-	switch (dev->device_info->asic_family) {
-	case CHIP_CARRIZO:
-		device_queue_manager_init_vi(&dqm->asic_ops);
-		break;
-
+	switch (dev->adev->asic_type) {
 	case CHIP_KAVERI:
-		device_queue_manager_init_cik(&dqm->asic_ops);
-		break;
-
 	case CHIP_HAWAII:
-		device_queue_manager_init_cik_hawaii(&dqm->asic_ops);
+		device_queue_manager_init_cik(&dqm->asic_ops);
 		break;
 
+	case CHIP_CARRIZO:
 	case CHIP_TONGA:
 	case CHIP_FIJI:
 	case CHIP_POLARIS10:
 	case CHIP_POLARIS11:
 	case CHIP_POLARIS12:
 	case CHIP_VEGAM:
-		device_queue_manager_init_vi_tonga(&dqm->asic_ops);
+		device_queue_manager_init_vi(&dqm->asic_ops);
 		break;
 
-	case CHIP_VEGA10:
-	case CHIP_VEGA12:
-	case CHIP_VEGA20:
-	case CHIP_RAVEN:
-	case CHIP_RENOIR:
-	case CHIP_ARCTURUS:
-	case CHIP_ALDEBARAN:
-		device_queue_manager_init_v9(&dqm->asic_ops);
-		break;
-	case CHIP_NAVI10:
-	case CHIP_NAVI12:
-	case CHIP_NAVI14:
-	case CHIP_SIENNA_CICHLID:
-	case CHIP_NAVY_FLOUNDER:
-	case CHIP_VANGOGH:
-	case CHIP_DIMGREY_CAVEFISH:
-	case CHIP_BEIGE_GOBY:
-	case CHIP_YELLOW_CARP:
-	case CHIP_CYAN_SKILLFISH:
-		device_queue_manager_init_v10_navi10(&dqm->asic_ops);
-		break;
 	default:
-		WARN(1, "Unexpected ASIC family %u",
-		     dev->device_info->asic_family);
-		goto out_free;
+		if (KFD_GC_VERSION(dev) >= IP_VERSION(12, 0, 0))
+			device_queue_manager_init_v12(&dqm->asic_ops);
+		else if (KFD_GC_VERSION(dev) >= IP_VERSION(11, 0, 0))
+			device_queue_manager_init_v11(&dqm->asic_ops);
+		else if (KFD_GC_VERSION(dev) >= IP_VERSION(10, 1, 1))
+			device_queue_manager_init_v10(&dqm->asic_ops);
+		else if (KFD_GC_VERSION(dev) >= IP_VERSION(9, 0, 1))
+			device_queue_manager_init_v9(&dqm->asic_ops);
+		else {
+			WARN(1, "Unexpected ASIC family %u",
+			     dev->adev->asic_type);
+			goto out_free;
+		}
 	}
 
 	if (init_mqd_managers(dqm))
 		goto out_free;
 
-	if (allocate_hiq_sdma_mqd(dqm)) {
-		pr_err("Failed to allocate hiq sdma mqd trunk buffer\n");
+	if (!dev->kfd->shared_resources.enable_mes && allocate_hiq_sdma_mqd(dqm)) {
+		dev_err(dev->adev->dev, "Failed to allocate hiq sdma mqd trunk buffer\n");
 		goto out_free;
 	}
 
-	if (!dqm->ops.initialize(dqm))
+	if (!dqm->ops.initialize(dqm)) {
+		init_waitqueue_head(&dqm->destroy_wait);
 		return dqm;
+	}
 
 out_free:
 	kfree(dqm);
 	return NULL;
 }
 
-static void deallocate_hiq_sdma_mqd(struct kfd_dev *dev,
+static void deallocate_hiq_sdma_mqd(struct kfd_node *dev,
 				    struct kfd_mem_obj *mqd)
 {
 	WARN(!mqd, "No hiq sdma mqd trunk to free");
 
-	amdgpu_amdkfd_free_gtt_mem(dev->kgd, mqd->gtt_mem);
+	amdgpu_amdkfd_free_gtt_mem(dev->adev, &mqd->gtt_mem);
 }
 
 void device_queue_manager_uninit(struct device_queue_manager *dqm)
 {
+	dqm->ops.stop(dqm);
 	dqm->ops.uninitialize(dqm);
-	deallocate_hiq_sdma_mqd(dqm->dev, &dqm->hiq_sdma_mqd);
+	if (!dqm->dev->kfd->shared_resources.enable_mes)
+		deallocate_hiq_sdma_mqd(dqm->dev, &dqm->hiq_sdma_mqd);
 	kfree(dqm);
 }
 
-int kfd_process_vm_fault(struct device_queue_manager *dqm, u32 pasid)
+int kfd_dqm_suspend_bad_queue_mes(struct kfd_node *knode, u32 pasid, u32 doorbell_id)
 {
-	struct kfd_process_device *pdd;
-	struct kfd_process *p = kfd_lookup_process_by_pasid(pasid);
+	struct kfd_process_device *pdd = NULL;
+	struct kfd_process *p = kfd_lookup_process_by_pasid(pasid, &pdd);
+	struct device_queue_manager *dqm = knode->dqm;
+	struct device *dev = dqm->dev->adev->dev;
+	struct qcm_process_device *qpd;
+	struct queue *q = NULL;
 	int ret = 0;
 
-	if (!p)
+	if (!pdd)
 		return -EINVAL;
-	WARN(debug_evictions, "Evicting pid %d", p->lead_thread->pid);
-	pdd = kfd_get_process_device_data(dqm->dev, p);
-	if (pdd)
-		ret = dqm->ops.evict_process_queues(dqm, &pdd->qpd);
-	kfd_unref_process(p);
 
+	dqm_lock(dqm);
+
+	if (pdd) {
+		qpd = &pdd->qpd;
+
+		list_for_each_entry(q, &qpd->queues_list, list) {
+			if (q->doorbell_id == doorbell_id && q->properties.is_active) {
+				ret = suspend_all_queues_mes(dqm);
+				if (ret) {
+					dev_err(dev, "Suspending all queues failed");
+					goto out;
+				}
+
+				q->properties.is_evicted = true;
+				q->properties.is_active = false;
+				decrement_queue_count(dqm, qpd, q);
+
+				ret = remove_queue_mes(dqm, q, qpd);
+				if (ret) {
+					dev_err(dev, "Removing bad queue failed");
+					goto out;
+				}
+
+				ret = resume_all_queues_mes(dqm);
+				if (ret)
+					dev_err(dev, "Resuming all queues failed");
+
+				break;
+			}
+		}
+	}
+
+out:
+	dqm_unlock(dqm);
+	kfd_unref_process(p);
 	return ret;
 }
 
-static void kfd_process_hw_exception(struct work_struct *work)
+int kfd_evict_process_device(struct kfd_process_device *pdd)
 {
-	struct device_queue_manager *dqm = container_of(work,
-			struct device_queue_manager, hw_exception_work);
-	amdgpu_amdkfd_gpu_reset(dqm->dev->kgd);
+	struct device_queue_manager *dqm;
+	struct kfd_process *p;
+
+	p = pdd->process;
+	dqm = pdd->dev->dqm;
+
+	WARN(debug_evictions, "Evicting pid %d", p->lead_thread->pid);
+
+	return dqm->ops.evict_process_queues(dqm, &pdd->qpd);
 }
 
+int reserve_debug_trap_vmid(struct device_queue_manager *dqm,
+				struct qcm_process_device *qpd)
+{
+	int r;
+	struct device *dev = dqm->dev->adev->dev;
+	int updated_vmid_mask;
+
+	if (dqm->sched_policy == KFD_SCHED_POLICY_NO_HWS) {
+		dev_err(dev, "Unsupported on sched_policy: %i\n", dqm->sched_policy);
+		return -EINVAL;
+	}
+
+	dqm_lock(dqm);
+
+	if (dqm->trap_debug_vmid != 0) {
+		dev_err(dev, "Trap debug id already reserved\n");
+		r = -EBUSY;
+		goto out_unlock;
+	}
+
+	r = unmap_queues_cpsch(dqm, KFD_UNMAP_QUEUES_FILTER_ALL_QUEUES, 0,
+			USE_DEFAULT_GRACE_PERIOD, false);
+	if (r)
+		goto out_unlock;
+
+	updated_vmid_mask = dqm->dev->kfd->shared_resources.compute_vmid_bitmap;
+	updated_vmid_mask &= ~(1 << dqm->dev->vm_info.last_vmid_kfd);
+
+	dqm->dev->kfd->shared_resources.compute_vmid_bitmap = updated_vmid_mask;
+	dqm->trap_debug_vmid = dqm->dev->vm_info.last_vmid_kfd;
+	r = set_sched_resources(dqm);
+	if (r)
+		goto out_unlock;
+
+	r = map_queues_cpsch(dqm);
+	if (r)
+		goto out_unlock;
+
+	pr_debug("Reserved VMID for trap debug: %i\n", dqm->trap_debug_vmid);
+
+out_unlock:
+	dqm_unlock(dqm);
+	return r;
+}
+
+/*
+ * Releases vmid for the trap debugger
+ */
+int release_debug_trap_vmid(struct device_queue_manager *dqm,
+			struct qcm_process_device *qpd)
+{
+	struct device *dev = dqm->dev->adev->dev;
+	int r;
+	int updated_vmid_mask;
+	uint32_t trap_debug_vmid;
+
+	if (dqm->sched_policy == KFD_SCHED_POLICY_NO_HWS) {
+		dev_err(dev, "Unsupported on sched_policy: %i\n", dqm->sched_policy);
+		return -EINVAL;
+	}
+
+	dqm_lock(dqm);
+	trap_debug_vmid = dqm->trap_debug_vmid;
+	if (dqm->trap_debug_vmid == 0) {
+		dev_err(dev, "Trap debug id is not reserved\n");
+		r = -EINVAL;
+		goto out_unlock;
+	}
+
+	r = unmap_queues_cpsch(dqm, KFD_UNMAP_QUEUES_FILTER_ALL_QUEUES, 0,
+			USE_DEFAULT_GRACE_PERIOD, false);
+	if (r)
+		goto out_unlock;
+
+	updated_vmid_mask = dqm->dev->kfd->shared_resources.compute_vmid_bitmap;
+	updated_vmid_mask |= (1 << dqm->dev->vm_info.last_vmid_kfd);
+
+	dqm->dev->kfd->shared_resources.compute_vmid_bitmap = updated_vmid_mask;
+	dqm->trap_debug_vmid = 0;
+	r = set_sched_resources(dqm);
+	if (r)
+		goto out_unlock;
+
+	r = map_queues_cpsch(dqm);
+	if (r)
+		goto out_unlock;
+
+	pr_debug("Released VMID for trap debug: %i\n", trap_debug_vmid);
+
+out_unlock:
+	dqm_unlock(dqm);
+	return r;
+}
+
+#define QUEUE_NOT_FOUND		-1
+/* invalidate queue operation in array */
+static void q_array_invalidate(uint32_t num_queues, uint32_t *queue_ids)
+{
+	int i;
+
+	for (i = 0; i < num_queues; i++)
+		queue_ids[i] |= KFD_DBG_QUEUE_INVALID_MASK;
+}
+
+/* find queue index in array */
+static int q_array_get_index(unsigned int queue_id,
+		uint32_t num_queues,
+		uint32_t *queue_ids)
+{
+	int i;
+
+	for (i = 0; i < num_queues; i++)
+		if (queue_id == (queue_ids[i] & ~KFD_DBG_QUEUE_INVALID_MASK))
+			return i;
+
+	return QUEUE_NOT_FOUND;
+}
+
+struct copy_context_work_handler_workarea {
+	struct work_struct copy_context_work;
+	struct kfd_process *p;
+};
+
+static void copy_context_work_handler(struct work_struct *work)
+{
+	struct copy_context_work_handler_workarea *workarea;
+	struct mqd_manager *mqd_mgr;
+	struct queue *q;
+	struct mm_struct *mm;
+	struct kfd_process *p;
+	uint32_t tmp_ctl_stack_used_size, tmp_save_area_used_size;
+	int i;
+
+	workarea = container_of(work,
+			struct copy_context_work_handler_workarea,
+			copy_context_work);
+
+	p = workarea->p;
+	mm = get_task_mm(p->lead_thread);
+
+	if (!mm)
+		return;
+
+	kthread_use_mm(mm);
+	for (i = 0; i < p->n_pdds; i++) {
+		struct kfd_process_device *pdd = p->pdds[i];
+		struct device_queue_manager *dqm = pdd->dev->dqm;
+		struct qcm_process_device *qpd = &pdd->qpd;
+
+		list_for_each_entry(q, &qpd->queues_list, list) {
+			if (q->properties.type != KFD_QUEUE_TYPE_COMPUTE)
+				continue;
+
+			mqd_mgr = dqm->mqd_mgrs[KFD_MQD_TYPE_CP];
+
+			/* We ignore the return value from get_wave_state
+			 * because
+			 * i) right now, it always returns 0, and
+			 * ii) if we hit an error, we would continue to the
+			 *      next queue anyway.
+			 */
+			mqd_mgr->get_wave_state(mqd_mgr,
+					q->mqd,
+					&q->properties,
+					(void __user *)	q->properties.ctx_save_restore_area_address,
+					&tmp_ctl_stack_used_size,
+					&tmp_save_area_used_size);
+		}
+	}
+	kthread_unuse_mm(mm);
+	mmput(mm);
+}
+
+static uint32_t *get_queue_ids(uint32_t num_queues, uint32_t *usr_queue_id_array)
+{
+	size_t array_size = num_queues * sizeof(uint32_t);
+
+	if (!usr_queue_id_array)
+		return NULL;
+
+	return memdup_user(usr_queue_id_array, array_size);
+}
+
+int resume_queues(struct kfd_process *p,
+		uint32_t num_queues,
+		uint32_t *usr_queue_id_array)
+{
+	uint32_t *queue_ids = NULL;
+	int total_resumed = 0;
+	int i;
+
+	if (usr_queue_id_array) {
+		queue_ids = get_queue_ids(num_queues, usr_queue_id_array);
+
+		if (IS_ERR(queue_ids))
+			return PTR_ERR(queue_ids);
+
+		/* mask all queues as invalid.  unmask per successful request */
+		q_array_invalidate(num_queues, queue_ids);
+	}
+
+	for (i = 0; i < p->n_pdds; i++) {
+		struct kfd_process_device *pdd = p->pdds[i];
+		struct device_queue_manager *dqm = pdd->dev->dqm;
+		struct device *dev = dqm->dev->adev->dev;
+		struct qcm_process_device *qpd = &pdd->qpd;
+		struct queue *q;
+		int r, per_device_resumed = 0;
+
+		dqm_lock(dqm);
+
+		/* unmask queues that resume or already resumed as valid */
+		list_for_each_entry(q, &qpd->queues_list, list) {
+			int q_idx = QUEUE_NOT_FOUND;
+
+			if (queue_ids)
+				q_idx = q_array_get_index(
+						q->properties.queue_id,
+						num_queues,
+						queue_ids);
+
+			if (!queue_ids || q_idx != QUEUE_NOT_FOUND) {
+				int err = resume_single_queue(dqm, &pdd->qpd, q);
+
+				if (queue_ids) {
+					if (!err) {
+						queue_ids[q_idx] &=
+							~KFD_DBG_QUEUE_INVALID_MASK;
+					} else {
+						queue_ids[q_idx] |=
+							KFD_DBG_QUEUE_ERROR_MASK;
+						break;
+					}
+				}
+
+				if (dqm->dev->kfd->shared_resources.enable_mes) {
+					wake_up_all(&dqm->destroy_wait);
+					if (!err)
+						total_resumed++;
+				} else {
+					per_device_resumed++;
+				}
+			}
+		}
+
+		if (!per_device_resumed) {
+			dqm_unlock(dqm);
+			continue;
+		}
+
+		r = execute_queues_cpsch(dqm,
+					KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES,
+					0,
+					USE_DEFAULT_GRACE_PERIOD);
+		if (r) {
+			dev_err(dev, "Failed to resume process queues\n");
+			if (queue_ids) {
+				list_for_each_entry(q, &qpd->queues_list, list) {
+					int q_idx = q_array_get_index(
+							q->properties.queue_id,
+							num_queues,
+							queue_ids);
+
+					/* mask queue as error on resume fail */
+					if (q_idx != QUEUE_NOT_FOUND)
+						queue_ids[q_idx] |=
+							KFD_DBG_QUEUE_ERROR_MASK;
+				}
+			}
+		} else {
+			wake_up_all(&dqm->destroy_wait);
+			total_resumed += per_device_resumed;
+		}
+
+		dqm_unlock(dqm);
+	}
+
+	if (queue_ids) {
+		if (copy_to_user((void __user *)usr_queue_id_array, queue_ids,
+				num_queues * sizeof(uint32_t)))
+			pr_err("copy_to_user failed on queue resume\n");
+
+		kfree(queue_ids);
+	}
+
+	return total_resumed;
+}
+
+int suspend_queues(struct kfd_process *p,
+			uint32_t num_queues,
+			uint32_t grace_period,
+			uint64_t exception_clear_mask,
+			uint32_t *usr_queue_id_array)
+{
+	uint32_t *queue_ids = get_queue_ids(num_queues, usr_queue_id_array);
+	int total_suspended = 0;
+	int i;
+
+	if (IS_ERR(queue_ids))
+		return PTR_ERR(queue_ids);
+
+	/* mask all queues as invalid.  umask on successful request */
+	q_array_invalidate(num_queues, queue_ids);
+
+	for (i = 0; i < p->n_pdds; i++) {
+		struct kfd_process_device *pdd = p->pdds[i];
+		struct device_queue_manager *dqm = pdd->dev->dqm;
+		struct device *dev = dqm->dev->adev->dev;
+		struct qcm_process_device *qpd = &pdd->qpd;
+		struct queue *q;
+		int r, per_device_suspended = 0;
+
+		mutex_lock(&p->event_mutex);
+		dqm_lock(dqm);
+
+		/* unmask queues that suspend or already suspended */
+		list_for_each_entry(q, &qpd->queues_list, list) {
+			int q_idx = q_array_get_index(q->properties.queue_id,
+							num_queues,
+							queue_ids);
+
+			if (q_idx != QUEUE_NOT_FOUND) {
+				int err = suspend_single_queue(dqm, pdd, q);
+				bool is_mes = dqm->dev->kfd->shared_resources.enable_mes;
+
+				if (!err) {
+					queue_ids[q_idx] &= ~KFD_DBG_QUEUE_INVALID_MASK;
+					if (exception_clear_mask && is_mes)
+						q->properties.exception_status &=
+							~exception_clear_mask;
+
+					if (is_mes)
+						total_suspended++;
+					else
+						per_device_suspended++;
+				} else if (err != -EBUSY) {
+					r = err;
+					queue_ids[q_idx] |= KFD_DBG_QUEUE_ERROR_MASK;
+					break;
+				}
+			}
+		}
+
+		if (!per_device_suspended) {
+			dqm_unlock(dqm);
+			mutex_unlock(&p->event_mutex);
+			if (total_suspended)
+				amdgpu_amdkfd_debug_mem_fence(dqm->dev->adev);
+			continue;
+		}
+
+		r = execute_queues_cpsch(dqm,
+			KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0,
+			grace_period);
+
+		if (r)
+			dev_err(dev, "Failed to suspend process queues.\n");
+		else
+			total_suspended += per_device_suspended;
+
+		list_for_each_entry(q, &qpd->queues_list, list) {
+			int q_idx = q_array_get_index(q->properties.queue_id,
+						num_queues, queue_ids);
+
+			if (q_idx == QUEUE_NOT_FOUND)
+				continue;
+
+			/* mask queue as error on suspend fail */
+			if (r)
+				queue_ids[q_idx] |= KFD_DBG_QUEUE_ERROR_MASK;
+			else if (exception_clear_mask)
+				q->properties.exception_status &=
+							~exception_clear_mask;
+		}
+
+		dqm_unlock(dqm);
+		mutex_unlock(&p->event_mutex);
+		amdgpu_device_flush_hdp(dqm->dev->adev, NULL);
+	}
+
+	if (total_suspended) {
+		struct copy_context_work_handler_workarea copy_context_worker;
+
+		INIT_WORK_ONSTACK(
+				&copy_context_worker.copy_context_work,
+				copy_context_work_handler);
+
+		copy_context_worker.p = p;
+
+		schedule_work(&copy_context_worker.copy_context_work);
+
+
+		flush_work(&copy_context_worker.copy_context_work);
+		destroy_work_on_stack(&copy_context_worker.copy_context_work);
+	}
+
+	if (copy_to_user((void __user *)usr_queue_id_array, queue_ids,
+			num_queues * sizeof(uint32_t)))
+		pr_err("copy_to_user failed on queue suspend\n");
+
+	kfree(queue_ids);
+
+	return total_suspended;
+}
+
+static uint32_t set_queue_type_for_user(struct queue_properties *q_props)
+{
+	switch (q_props->type) {
+	case KFD_QUEUE_TYPE_COMPUTE:
+		return q_props->format == KFD_QUEUE_FORMAT_PM4
+					? KFD_IOC_QUEUE_TYPE_COMPUTE
+					: KFD_IOC_QUEUE_TYPE_COMPUTE_AQL;
+	case KFD_QUEUE_TYPE_SDMA:
+		return KFD_IOC_QUEUE_TYPE_SDMA;
+	case KFD_QUEUE_TYPE_SDMA_XGMI:
+		return KFD_IOC_QUEUE_TYPE_SDMA_XGMI;
+	default:
+		WARN_ONCE(true, "queue type not recognized!");
+		return 0xffffffff;
+	};
+}
+
+void set_queue_snapshot_entry(struct queue *q,
+			      uint64_t exception_clear_mask,
+			      struct kfd_queue_snapshot_entry *qss_entry)
+{
+	qss_entry->ring_base_address = q->properties.queue_address;
+	qss_entry->write_pointer_address = (uint64_t)q->properties.write_ptr;
+	qss_entry->read_pointer_address = (uint64_t)q->properties.read_ptr;
+	qss_entry->ctx_save_restore_address =
+				q->properties.ctx_save_restore_area_address;
+	qss_entry->ctx_save_restore_area_size =
+				q->properties.ctx_save_restore_area_size;
+	qss_entry->exception_status = q->properties.exception_status;
+	qss_entry->queue_id = q->properties.queue_id;
+	qss_entry->gpu_id = q->device->id;
+	qss_entry->ring_size = (uint32_t)q->properties.queue_size;
+	qss_entry->queue_type = set_queue_type_for_user(&q->properties);
+	q->properties.exception_status &= ~exception_clear_mask;
+}
+
+int debug_lock_and_unmap(struct device_queue_manager *dqm)
+{
+	struct device *dev = dqm->dev->adev->dev;
+	int r;
+
+	if (dqm->sched_policy == KFD_SCHED_POLICY_NO_HWS) {
+		dev_err(dev, "Unsupported on sched_policy: %i\n", dqm->sched_policy);
+		return -EINVAL;
+	}
+
+	if (!kfd_dbg_is_per_vmid_supported(dqm->dev))
+		return 0;
+
+	dqm_lock(dqm);
+
+	r = unmap_queues_cpsch(dqm, KFD_UNMAP_QUEUES_FILTER_ALL_QUEUES, 0, 0, false);
+	if (r)
+		dqm_unlock(dqm);
+
+	return r;
+}
+
+int debug_map_and_unlock(struct device_queue_manager *dqm)
+{
+	struct device *dev = dqm->dev->adev->dev;
+	int r;
+
+	if (dqm->sched_policy == KFD_SCHED_POLICY_NO_HWS) {
+		dev_err(dev, "Unsupported on sched_policy: %i\n", dqm->sched_policy);
+		return -EINVAL;
+	}
+
+	if (!kfd_dbg_is_per_vmid_supported(dqm->dev))
+		return 0;
+
+	r = map_queues_cpsch(dqm);
+
+	dqm_unlock(dqm);
+
+	return r;
+}
+
+int debug_refresh_runlist(struct device_queue_manager *dqm)
+{
+	int r = debug_lock_and_unmap(dqm);
+
+	if (r)
+		return r;
+
+	return debug_map_and_unlock(dqm);
+}
+
+bool kfd_dqm_is_queue_in_process(struct device_queue_manager *dqm,
+				 struct qcm_process_device *qpd,
+				 int doorbell_off, u32 *queue_format)
+{
+	struct queue *q;
+	bool r = false;
+
+	if (!queue_format)
+		return r;
+
+	dqm_lock(dqm);
+
+	list_for_each_entry(q, &qpd->queues_list, list) {
+		if (q->properties.doorbell_off == doorbell_off) {
+			*queue_format = q->properties.format;
+			r = true;
+			goto out;
+		}
+	}
+
+out:
+	dqm_unlock(dqm);
+	return r;
+}
 #if defined(CONFIG_DEBUG_FS)
 
 static void seq_reg_dump(struct seq_file *m,
@@ -2060,56 +3671,69 @@ static void seq_reg_dump(struct seq_file *m,
 int dqm_debugfs_hqds(struct seq_file *m, void *data)
 {
 	struct device_queue_manager *dqm = data;
+	uint32_t xcc_mask = dqm->dev->xcc_mask;
 	uint32_t (*dump)[2], n_regs;
 	int pipe, queue;
-	int r = 0;
+	int r = 0, xcc_id;
+	uint32_t sdma_engine_start;
 
 	if (!dqm->sched_running) {
-		seq_printf(m, " Device is stopped\n");
-
+		seq_puts(m, " Device is stopped\n");
 		return 0;
 	}
 
-	r = dqm->dev->kfd2kgd->hqd_dump(dqm->dev->kgd,
-					KFD_CIK_HIQ_PIPE, KFD_CIK_HIQ_QUEUE,
-					&dump, &n_regs);
-	if (!r) {
-		seq_printf(m, "  HIQ on MEC %d Pipe %d Queue %d\n",
-			   KFD_CIK_HIQ_PIPE/get_pipes_per_mec(dqm)+1,
-			   KFD_CIK_HIQ_PIPE%get_pipes_per_mec(dqm),
-			   KFD_CIK_HIQ_QUEUE);
-		seq_reg_dump(m, dump, n_regs);
+	for_each_inst(xcc_id, xcc_mask) {
+		r = dqm->dev->kfd2kgd->hqd_dump(dqm->dev->adev,
+						KFD_CIK_HIQ_PIPE,
+						KFD_CIK_HIQ_QUEUE, &dump,
+						&n_regs, xcc_id);
+		if (!r) {
+			seq_printf(
+				m,
+				"   Inst %d, HIQ on MEC %d Pipe %d Queue %d\n",
+				xcc_id,
+				KFD_CIK_HIQ_PIPE / get_pipes_per_mec(dqm) + 1,
+				KFD_CIK_HIQ_PIPE % get_pipes_per_mec(dqm),
+				KFD_CIK_HIQ_QUEUE);
+			seq_reg_dump(m, dump, n_regs);
 
-		kfree(dump);
-	}
+			kfree(dump);
+		}
 
-	for (pipe = 0; pipe < get_pipes_per_mec(dqm); pipe++) {
-		int pipe_offset = pipe * get_queues_per_pipe(dqm);
+		for (pipe = 0; pipe < get_pipes_per_mec(dqm); pipe++) {
+			int pipe_offset = pipe * get_queues_per_pipe(dqm);
 
-		for (queue = 0; queue < get_queues_per_pipe(dqm); queue++) {
-			if (!test_bit(pipe_offset + queue,
-				      dqm->dev->shared_resources.cp_queue_bitmap))
-				continue;
+			for (queue = 0; queue < get_queues_per_pipe(dqm); queue++) {
+				if (!test_bit(pipe_offset + queue,
+				      dqm->dev->kfd->shared_resources.cp_queue_bitmap))
+					continue;
 
-			r = dqm->dev->kfd2kgd->hqd_dump(
-				dqm->dev->kgd, pipe, queue, &dump, &n_regs);
-			if (r)
-				break;
+				r = dqm->dev->kfd2kgd->hqd_dump(dqm->dev->adev,
+								pipe, queue,
+								&dump, &n_regs,
+								xcc_id);
+				if (r)
+					break;
 
-			seq_printf(m, "  CP Pipe %d, Queue %d\n",
-				  pipe, queue);
-			seq_reg_dump(m, dump, n_regs);
+				seq_printf(m,
+					   " Inst %d,  CP Pipe %d, Queue %d\n",
+					   xcc_id, pipe, queue);
+				seq_reg_dump(m, dump, n_regs);
 
-			kfree(dump);
+				kfree(dump);
+			}
 		}
 	}
 
-	for (pipe = 0; pipe < get_num_all_sdma_engines(dqm); pipe++) {
+	sdma_engine_start = dqm->dev->node_id * get_num_all_sdma_engines(dqm);
+	for (pipe = sdma_engine_start;
+	     pipe < (sdma_engine_start + get_num_all_sdma_engines(dqm));
+	     pipe++) {
 		for (queue = 0;
-		     queue < dqm->dev->device_info->num_sdma_queues_per_engine;
+		     queue < dqm->dev->kfd->device_info.num_sdma_queues_per_engine;
 		     queue++) {
 			r = dqm->dev->kfd2kgd->hqd_sdma_dump(
-				dqm->dev->kgd, pipe, queue, &dump, &n_regs);
+				dqm->dev->adev, pipe, queue, &dump, &n_regs);
 			if (r)
 				break;
 
@@ -2135,7 +3759,8 @@ int dqm_debugfs_hang_hws(struct device_queue_manager *dqm)
 		return r;
 	}
 	dqm->active_runlist = true;
-	r = execute_queues_cpsch(dqm, KFD_UNMAP_QUEUES_FILTER_ALL_QUEUES, 0);
+	r = execute_queues_cpsch(dqm, KFD_UNMAP_QUEUES_FILTER_ALL_QUEUES,
+				0, USE_DEFAULT_GRACE_PERIOD);
 	dqm_unlock(dqm);
 
 	return r;
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.h b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.h
index 499fc0ea387f..74a61b5b2f0b 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.h
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.h
@@ -1,5 +1,6 @@
+/* SPDX-License-Identifier: GPL-2.0 OR MIT */
 /*
- * Copyright 2014 Advanced Micro Devices, Inc.
+ * Copyright 2014-2022 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -34,11 +35,49 @@
 
 #define VMID_NUM 16
 
+#define KFD_MES_PROCESS_QUANTUM		100000
+#define KFD_MES_GANG_QUANTUM		10000
+
 struct device_process_node {
 	struct qcm_process_device *qpd;
 	struct list_head list;
 };
 
+union SQ_CMD_BITS {
+	struct {
+		uint32_t cmd:3;
+		uint32_t:1;
+		uint32_t mode:3;
+		uint32_t check_vmid:1;
+		uint32_t trap_id:3;
+		uint32_t:5;
+		uint32_t wave_id:4;
+		uint32_t simd_id:2;
+		uint32_t:2;
+		uint32_t queue_id:3;
+		uint32_t:1;
+		uint32_t vm_id:4;
+	} bitfields, bits;
+	uint32_t u32All;
+	signed int i32All;
+	float f32All;
+};
+
+union GRBM_GFX_INDEX_BITS {
+	struct {
+		uint32_t instance_index:8;
+		uint32_t sh_index:8;
+		uint32_t se_index:8;
+		uint32_t:5;
+		uint32_t sh_broadcast_writes:1;
+		uint32_t instance_broadcast_writes:1;
+		uint32_t se_broadcast_writes:1;
+	} bitfields, bits;
+	uint32_t u32All;
+	signed int i32All;
+	float f32All;
+};
+
 /**
  * struct device_queue_manager_ops
  *
@@ -56,7 +95,7 @@ struct device_process_node {
  *
  * @initialize: Initializes the pipelines and memory module for that device.
  *
- * @start: Initializes the resources/modules the the device needs for queues
+ * @start: Initializes the resources/modules the device needs for queues
  * execution. This function is called on device initialization and after the
  * system woke up after suspension.
  *
@@ -66,6 +105,12 @@ struct device_process_node {
  * @uninitialize: Destroys all the device queue manager resources allocated in
  * initialize routine.
  *
+ * @halt: This routine unmaps queues from runlist and set halt status to true
+ * so no more queues will be mapped to runlist until unhalt.
+ *
+ * @unhalt: This routine unset halt status to flase and maps queues back to
+ * runlist.
+ *
  * @create_kernel_queue: Creates kernel queue. Used for debug queue.
  *
  * @destroy_kernel_queue: Destroys kernel queue. Used for debug queue.
@@ -77,16 +122,24 @@ struct device_process_node {
  *
  * @evict_process_queues: Evict all active queues of a process
  *
- * @restore_process_queues: Restore all evicted queues queues of a process
+ * @restore_process_queues: Restore all evicted queues of a process
  *
  * @get_wave_state: Retrieves context save state and optionally copies the
  * control stack, if kept in the MQD, to the given userspace address.
+ *
+ * @reset_queues: reset queues which consume RAS poison
+ * @get_queue_checkpoint_info: Retrieves queue size information for CRIU checkpoint.
+ *
+ * @checkpoint_mqd: checkpoint queue MQD contents for CRIU.
  */
 
 struct device_queue_manager_ops {
 	int	(*create_queue)(struct device_queue_manager *dqm,
 				struct queue *q,
-				struct qcm_process_device *qpd);
+				struct qcm_process_device *qpd,
+				const struct kfd_criu_queue_priv_data *qd,
+				const void *restore_mqd,
+				const void *restore_ctl_stack);
 
 	int	(*destroy_queue)(struct device_queue_manager *dqm,
 				struct qcm_process_device *qpd,
@@ -104,8 +157,9 @@ struct device_queue_manager_ops {
 	int	(*initialize)(struct device_queue_manager *dqm);
 	int	(*start)(struct device_queue_manager *dqm);
 	int	(*stop)(struct device_queue_manager *dqm);
-	void	(*pre_reset)(struct device_queue_manager *dqm);
 	void	(*uninitialize)(struct device_queue_manager *dqm);
+	int     (*halt)(struct device_queue_manager *dqm);
+	int     (*unhalt)(struct device_queue_manager *dqm);
 	int	(*create_kernel_queue)(struct device_queue_manager *dqm,
 					struct kernel_queue *kq,
 					struct qcm_process_device *qpd);
@@ -119,7 +173,8 @@ struct device_queue_manager_ops {
 					   enum cache_policy default_policy,
 					   enum cache_policy alternate_policy,
 					   void __user *alternate_aperture_base,
-					   uint64_t alternate_aperture_size);
+					   uint64_t alternate_aperture_size,
+					   u32 misc_process_properties);
 
 	int (*process_termination)(struct device_queue_manager *dqm,
 			struct qcm_process_device *qpd);
@@ -134,6 +189,17 @@ struct device_queue_manager_ops {
 				  void __user *ctl_stack,
 				  u32 *ctl_stack_used_size,
 				  u32 *save_area_used_size);
+
+	int (*reset_queues)(struct device_queue_manager *dqm,
+					uint16_t pasid);
+	void	(*get_queue_checkpoint_info)(struct device_queue_manager *dqm,
+				  const struct queue *q, u32 *mqd_size,
+				  u32 *ctl_stack_size);
+
+	int	(*checkpoint_mqd)(struct device_queue_manager *dqm,
+				  const struct queue *q,
+				  void *mqd,
+				  void *ctl_stack);
 };
 
 struct device_queue_manager_asic_ops {
@@ -144,12 +210,20 @@ struct device_queue_manager_asic_ops {
 					   enum cache_policy default_policy,
 					   enum cache_policy alternate_policy,
 					   void __user *alternate_aperture_base,
-					   uint64_t alternate_aperture_size);
+					   uint64_t alternate_aperture_size,
+					   u32 misc_process_properties);
 	void	(*init_sdma_vm)(struct device_queue_manager *dqm,
 				struct queue *q,
 				struct qcm_process_device *qpd);
 	struct mqd_manager *	(*mqd_manager_init)(enum KFD_MQD_TYPE type,
-				 struct kfd_dev *dev);
+				 struct kfd_node *dev);
+};
+
+struct dqm_detect_hang_info {
+	int pipe_id;
+	int queue_id;
+	int xcc_id;
+	uint64_t queue_address;
 };
 
 /**
@@ -170,7 +244,7 @@ struct device_queue_manager {
 
 	struct mqd_manager	*mqd_mgrs[KFD_MQD_TYPE_MAX];
 	struct packet_manager	packet_mgr;
-	struct kfd_dev		*dev;
+	struct kfd_node		*dev;
 	struct mutex		lock_hidden; /* use dqm_lock/unlock(dqm) */
 	struct list_head	queues;
 	unsigned int		saved_flags;
@@ -181,8 +255,8 @@ struct device_queue_manager {
 	unsigned int		total_queue_count;
 	unsigned int		next_pipe_to_allocate;
 	unsigned int		*allocated_queues;
-	uint64_t		sdma_bitmap;
-	uint64_t		xgmi_sdma_bitmap;
+	DECLARE_BITMAP(sdma_bitmap, KFD_MAX_SDMA_QUEUES);
+	DECLARE_BITMAP(xgmi_sdma_bitmap, KFD_MAX_SDMA_QUEUES);
 	/* the pasid mapping for each kfd vmid */
 	uint16_t		vmid_pasid[VMID_NUM];
 	uint64_t		pipelines_addr;
@@ -191,26 +265,39 @@ struct device_queue_manager {
 	struct kfd_mem_obj	*fence_mem;
 	bool			active_runlist;
 	int			sched_policy;
+	uint32_t		trap_debug_vmid;
 
 	/* hw exception  */
 	bool			is_hws_hang;
 	bool			is_resetting;
-	struct work_struct	hw_exception_work;
 	struct kfd_mem_obj	hiq_sdma_mqd;
 	bool			sched_running;
+	bool			sched_halt;
+
+	/* used for GFX 9.4.3 only */
+	uint32_t		current_logical_xcc_start;
+
+	uint32_t		wait_times;
+
+	wait_queue_head_t	destroy_wait;
+
+	/* for per-queue reset support */
+	struct dqm_detect_hang_info *detect_hang_info;
+	size_t detect_hang_info_size;
+	int detect_hang_count;
 };
 
 void device_queue_manager_init_cik(
 		struct device_queue_manager_asic_ops *asic_ops);
-void device_queue_manager_init_cik_hawaii(
-		struct device_queue_manager_asic_ops *asic_ops);
 void device_queue_manager_init_vi(
 		struct device_queue_manager_asic_ops *asic_ops);
-void device_queue_manager_init_vi_tonga(
-		struct device_queue_manager_asic_ops *asic_ops);
 void device_queue_manager_init_v9(
 		struct device_queue_manager_asic_ops *asic_ops);
-void device_queue_manager_init_v10_navi10(
+void device_queue_manager_init_v10(
+		struct device_queue_manager_asic_ops *asic_ops);
+void device_queue_manager_init_v11(
+		struct device_queue_manager_asic_ops *asic_ops);
+void device_queue_manager_init_v12(
 		struct device_queue_manager_asic_ops *asic_ops);
 void program_sh_mem_settings(struct device_queue_manager *dqm,
 					struct qcm_process_device *qpd);
@@ -219,6 +306,27 @@ unsigned int get_queues_per_pipe(struct device_queue_manager *dqm);
 unsigned int get_pipes_per_mec(struct device_queue_manager *dqm);
 unsigned int get_num_sdma_queues(struct device_queue_manager *dqm);
 unsigned int get_num_xgmi_sdma_queues(struct device_queue_manager *dqm);
+int reserve_debug_trap_vmid(struct device_queue_manager *dqm,
+			struct qcm_process_device *qpd);
+int release_debug_trap_vmid(struct device_queue_manager *dqm,
+			struct qcm_process_device *qpd);
+int suspend_queues(struct kfd_process *p,
+			uint32_t num_queues,
+			uint32_t grace_period,
+			uint64_t exception_clear_mask,
+			uint32_t *usr_queue_id_array);
+int resume_queues(struct kfd_process *p,
+		uint32_t num_queues,
+		uint32_t *usr_queue_id_array);
+void set_queue_snapshot_entry(struct queue *q,
+			      uint64_t exception_clear_mask,
+			      struct kfd_queue_snapshot_entry *qss_entry);
+int debug_lock_and_unmap(struct device_queue_manager *dqm);
+int debug_map_and_unlock(struct device_queue_manager *dqm);
+int debug_refresh_runlist(struct device_queue_manager *dqm);
+bool kfd_dqm_is_queue_in_process(struct device_queue_manager *dqm,
+				 struct qcm_process_device *qpd,
+				 int doorbell_off, u32 *queue_format);
 
 static inline unsigned int get_sh_mem_bases_32(struct kfd_process_device *pdd)
 {
@@ -248,9 +356,17 @@ static inline void dqm_unlock(struct device_queue_manager *dqm)
 
 static inline int read_sdma_queue_counter(uint64_t __user *q_rptr, uint64_t *val)
 {
-        /*
-         * SDMA activity counter is stored at queue's RPTR + 0x8 location.
-         */
+	/* SDMA activity counter is stored at queue's RPTR + 0x8 location. */
 	return get_user(*val, q_rptr + 1);
 }
+
+static inline void update_dqm_wait_times(struct device_queue_manager *dqm)
+{
+	if (dqm->dev->kfd2kgd->get_iq_wait_times)
+		dqm->dev->kfd2kgd->get_iq_wait_times(dqm->dev->adev,
+					&dqm->wait_times,
+					ffs(dqm->dev->xcc_mask) - 1);
+}
+
+
 #endif /* KFD_DEVICE_QUEUE_MANAGER_H_ */
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_cik.c b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_cik.c
index 0d26506798cf..0508ef5a41d7 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_cik.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_cik.c
@@ -1,5 +1,6 @@
+// SPDX-License-Identifier: GPL-2.0 OR MIT
 /*
- * Copyright 2014 Advanced Micro Devices, Inc.
+ * Copyright 2014-2022 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -26,24 +27,29 @@
 #include "oss/oss_2_4_sh_mask.h"
 #include "gca/gfx_7_2_sh_mask.h"
 
+/*
+ * Low bits must be 0000/FFFF as required by HW, high bits must be 0 to
+ * stay in user mode.
+ */
+#define APE1_FIXED_BITS_MASK 0xFFFF80000000FFFFULL
+/* APE1 limit is inclusive and 64K aligned. */
+#define APE1_LIMIT_ALIGNMENT 0xFFFF
+
 static bool set_cache_memory_policy_cik(struct device_queue_manager *dqm,
 				   struct qcm_process_device *qpd,
 				   enum cache_policy default_policy,
 				   enum cache_policy alternate_policy,
 				   void __user *alternate_aperture_base,
-				   uint64_t alternate_aperture_size);
+				   uint64_t alternate_aperture_size,
+				   u32 misc_process_properties);
 static int update_qpd_cik(struct device_queue_manager *dqm,
-					struct qcm_process_device *qpd);
-static int update_qpd_cik_hawaii(struct device_queue_manager *dqm,
-					struct qcm_process_device *qpd);
-static void init_sdma_vm(struct device_queue_manager *dqm, struct queue *q,
-				struct qcm_process_device *qpd);
-static void init_sdma_vm_hawaii(struct device_queue_manager *dqm,
-				struct queue *q,
-				struct qcm_process_device *qpd);
+			  struct qcm_process_device *qpd);
+static void init_sdma_vm(struct device_queue_manager *dqm,
+			 struct queue *q,
+			 struct qcm_process_device *qpd);
 
 void device_queue_manager_init_cik(
-		struct device_queue_manager_asic_ops *asic_ops)
+	struct device_queue_manager_asic_ops *asic_ops)
 {
 	asic_ops->set_cache_memory_policy = set_cache_memory_policy_cik;
 	asic_ops->update_qpd = update_qpd_cik;
@@ -51,15 +57,6 @@ void device_queue_manager_init_cik(
 	asic_ops->mqd_manager_init = mqd_manager_init_cik;
 }
 
-void device_queue_manager_init_cik_hawaii(
-		struct device_queue_manager_asic_ops *asic_ops)
-{
-	asic_ops->set_cache_memory_policy = set_cache_memory_policy_cik;
-	asic_ops->update_qpd = update_qpd_cik_hawaii;
-	asic_ops->init_sdma_vm = init_sdma_vm_hawaii;
-	asic_ops->mqd_manager_init = mqd_manager_init_cik_hawaii;
-}
-
 static uint32_t compute_sh_mem_bases_64bit(unsigned int top_address_nybble)
 {
 	/* In 64-bit mode, we can only control the top 3 bits of the LDS,
@@ -92,10 +89,41 @@ static bool set_cache_memory_policy_cik(struct device_queue_manager *dqm,
 				   enum cache_policy default_policy,
 				   enum cache_policy alternate_policy,
 				   void __user *alternate_aperture_base,
-				   uint64_t alternate_aperture_size)
+				   uint64_t alternate_aperture_size,
+				   u32 misc_process_properties)
 {
 	uint32_t default_mtype;
 	uint32_t ape1_mtype;
+	unsigned int temp;
+	bool retval = true;
+
+	if (alternate_aperture_size == 0) {
+		/* base > limit disables APE1 */
+		qpd->sh_mem_ape1_base = 1;
+		qpd->sh_mem_ape1_limit = 0;
+	} else {
+		/*
+		 * In FSA64, APE1_Base[63:0] = { 16{SH_MEM_APE1_BASE[31]},
+		 *			SH_MEM_APE1_BASE[31:0], 0x0000 }
+		 * APE1_Limit[63:0] = { 16{SH_MEM_APE1_LIMIT[31]},
+		 *			SH_MEM_APE1_LIMIT[31:0], 0xFFFF }
+		 * Verify that the base and size parameters can be
+		 * represented in this format and convert them.
+		 * Additionally restrict APE1 to user-mode addresses.
+		 */
+
+		uint64_t base = (uintptr_t)alternate_aperture_base;
+		uint64_t limit = base + alternate_aperture_size - 1;
+
+		if (limit <= base || (base & APE1_FIXED_BITS_MASK) != 0 ||
+		   (limit & APE1_FIXED_BITS_MASK) != APE1_LIMIT_ALIGNMENT) {
+			retval = false;
+			goto out;
+		}
+
+		qpd->sh_mem_ape1_base = base >> 16;
+		qpd->sh_mem_ape1_limit = limit >> 16;
+	}
 
 	default_mtype = (default_policy == cache_policy_coherent) ?
 			MTYPE_NONCACHED :
@@ -109,93 +137,28 @@ static bool set_cache_memory_policy_cik(struct device_queue_manager *dqm,
 			| ALIGNMENT_MODE(SH_MEM_ALIGNMENT_MODE_UNALIGNED)
 			| DEFAULT_MTYPE(default_mtype)
 			| APE1_MTYPE(ape1_mtype);
-
-	return true;
-}
-
-static int update_qpd_cik(struct device_queue_manager *dqm,
-		struct qcm_process_device *qpd)
-{
-	struct kfd_process_device *pdd;
-	unsigned int temp;
-
-	pdd = qpd_to_pdd(qpd);
-
-	/* check if sh_mem_config register already configured */
-	if (qpd->sh_mem_config == 0) {
-		qpd->sh_mem_config =
-			ALIGNMENT_MODE(SH_MEM_ALIGNMENT_MODE_UNALIGNED) |
-			DEFAULT_MTYPE(MTYPE_NONCACHED) |
-			APE1_MTYPE(MTYPE_NONCACHED);
-		qpd->sh_mem_ape1_limit = 0;
-		qpd->sh_mem_ape1_base = 0;
-	}
-
-	if (qpd->pqm->process->is_32bit_user_mode) {
-		temp = get_sh_mem_bases_32(pdd);
-		qpd->sh_mem_bases = SHARED_BASE(temp);
-		qpd->sh_mem_config |= PTR32;
-	} else {
-		temp = get_sh_mem_bases_nybble_64(pdd);
-		qpd->sh_mem_bases = compute_sh_mem_bases_64bit(temp);
-		qpd->sh_mem_config |= 1  << SH_MEM_CONFIG__PRIVATE_ATC__SHIFT;
-	}
-
-	pr_debug("is32bit process: %d sh_mem_bases nybble: 0x%X and register 0x%X\n",
-		qpd->pqm->process->is_32bit_user_mode, temp, qpd->sh_mem_bases);
-
-	return 0;
-}
-
-static int update_qpd_cik_hawaii(struct device_queue_manager *dqm,
-		struct qcm_process_device *qpd)
-{
-	struct kfd_process_device *pdd;
-	unsigned int temp;
-
-	pdd = qpd_to_pdd(qpd);
-
-	/* check if sh_mem_config register already configured */
-	if (qpd->sh_mem_config == 0) {
-		qpd->sh_mem_config =
-			ALIGNMENT_MODE(SH_MEM_ALIGNMENT_MODE_UNALIGNED) |
-			DEFAULT_MTYPE(MTYPE_NONCACHED) |
-			APE1_MTYPE(MTYPE_NONCACHED);
-		qpd->sh_mem_ape1_limit = 0;
-		qpd->sh_mem_ape1_base = 0;
-	}
-
 	/* On dGPU we're always in GPUVM64 addressing mode with 64-bit
 	 * aperture addresses.
 	 */
-	temp = get_sh_mem_bases_nybble_64(pdd);
+	temp = get_sh_mem_bases_nybble_64(qpd_to_pdd(qpd));
 	qpd->sh_mem_bases = compute_sh_mem_bases_64bit(temp);
 
 	pr_debug("is32bit process: %d sh_mem_bases nybble: 0x%X and register 0x%X\n",
 		qpd->pqm->process->is_32bit_user_mode, temp, qpd->sh_mem_bases);
 
-	return 0;
+out:
+	return retval;
 }
 
-static void init_sdma_vm(struct device_queue_manager *dqm, struct queue *q,
-				struct qcm_process_device *qpd)
+static int update_qpd_cik(struct device_queue_manager *dqm,
+			  struct qcm_process_device *qpd)
 {
-	uint32_t value = (1 << SDMA0_RLC0_VIRTUAL_ADDR__ATC__SHIFT);
-
-	if (q->process->is_32bit_user_mode)
-		value |= (1 << SDMA0_RLC0_VIRTUAL_ADDR__PTR32__SHIFT) |
-				get_sh_mem_bases_32(qpd_to_pdd(qpd));
-	else
-		value |= ((get_sh_mem_bases_nybble_64(qpd_to_pdd(qpd))) <<
-				SDMA0_RLC0_VIRTUAL_ADDR__SHARED_BASE__SHIFT) &
-				SDMA0_RLC0_VIRTUAL_ADDR__SHARED_BASE_MASK;
-
-	q->properties.sdma_vm_addr = value;
+	return 0;
 }
 
-static void init_sdma_vm_hawaii(struct device_queue_manager *dqm,
-				struct queue *q,
-				struct qcm_process_device *qpd)
+static void init_sdma_vm(struct device_queue_manager *dqm,
+			 struct queue *q,
+			 struct qcm_process_device *qpd)
 {
 	/* On dGPU we're always in GPUVM64 addressing mode with 64-bit
 	 * aperture addresses.
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v10.c b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v10.c
index ad0593342333..ba6e3d747ccd 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v10.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v10.c
@@ -1,5 +1,6 @@
+// SPDX-License-Identifier: GPL-2.0 OR MIT
 /*
- * Copyright 2018 Advanced Micro Devices, Inc.
+ * Copyright 2018-2022 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -30,10 +31,18 @@ static int update_qpd_v10(struct device_queue_manager *dqm,
 			 struct qcm_process_device *qpd);
 static void init_sdma_vm_v10(struct device_queue_manager *dqm, struct queue *q,
 			    struct qcm_process_device *qpd);
+static bool set_cache_memory_policy_v10(struct device_queue_manager *dqm,
+				   struct qcm_process_device *qpd,
+				   enum cache_policy default_policy,
+				   enum cache_policy alternate_policy,
+				   void __user *alternate_aperture_base,
+				   uint64_t alternate_aperture_size,
+				   u32 misc_process_properties);
 
-void device_queue_manager_init_v10_navi10(
+void device_queue_manager_init_v10(
 	struct device_queue_manager_asic_ops *asic_ops)
 {
+	asic_ops->set_cache_memory_policy = set_cache_memory_policy_v10;
 	asic_ops->update_qpd = update_qpd_v10;
 	asic_ops->init_sdma_vm = init_sdma_vm_v10;
 	asic_ops->mqd_manager_init = mqd_manager_init_v10;
@@ -48,36 +57,28 @@ static uint32_t compute_sh_mem_bases_64bit(struct kfd_process_device *pdd)
 		private_base;
 }
 
-static int update_qpd_v10(struct device_queue_manager *dqm,
-			 struct qcm_process_device *qpd)
+static bool set_cache_memory_policy_v10(struct device_queue_manager *dqm,
+				   struct qcm_process_device *qpd,
+				   enum cache_policy default_policy,
+				   enum cache_policy alternate_policy,
+				   void __user *alternate_aperture_base,
+				   uint64_t alternate_aperture_size,
+				   u32 misc_process_properties)
 {
-	struct kfd_process_device *pdd;
-
-	pdd = qpd_to_pdd(qpd);
-
-	/* check if sh_mem_config register already configured */
-	if (qpd->sh_mem_config == 0) {
-		qpd->sh_mem_config =
-			(SH_MEM_ALIGNMENT_MODE_UNALIGNED <<
-				SH_MEM_CONFIG__ALIGNMENT_MODE__SHIFT) |
-			(3 << SH_MEM_CONFIG__INITIAL_INST_PREFETCH__SHIFT);
-#if 0
-		/* TODO:
-		 *    This shouldn't be an issue with Navi10.  Verify.
-		 */
-		if (vega10_noretry)
-			qpd->sh_mem_config |=
-				1 << SH_MEM_CONFIG__RETRY_DISABLE__SHIFT;
-#endif
-
-		qpd->sh_mem_ape1_limit = 0;
-		qpd->sh_mem_ape1_base = 0;
-	}
-
-	qpd->sh_mem_bases = compute_sh_mem_bases_64bit(pdd);
+	qpd->sh_mem_config = (SH_MEM_ALIGNMENT_MODE_UNALIGNED <<
+			      SH_MEM_CONFIG__ALIGNMENT_MODE__SHIFT) |
+			      (3 << SH_MEM_CONFIG__INITIAL_INST_PREFETCH__SHIFT);
+	qpd->sh_mem_ape1_limit = 0;
+	qpd->sh_mem_ape1_base = 0;
+	qpd->sh_mem_bases = compute_sh_mem_bases_64bit(qpd_to_pdd(qpd));
 
 	pr_debug("sh_mem_bases 0x%X\n", qpd->sh_mem_bases);
+	return true;
+}
 
+static int update_qpd_v10(struct device_queue_manager *dqm,
+			 struct qcm_process_device *qpd)
+{
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v11.c b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v11.c
new file mode 100644
index 000000000000..8b447d04558f
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v11.c
@@ -0,0 +1,90 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "kfd_device_queue_manager.h"
+#include "gc/gc_11_0_0_offset.h"
+#include "gc/gc_11_0_0_sh_mask.h"
+#include "soc21_enum.h"
+
+static int update_qpd_v11(struct device_queue_manager *dqm,
+			 struct qcm_process_device *qpd);
+static void init_sdma_vm_v11(struct device_queue_manager *dqm, struct queue *q,
+			    struct qcm_process_device *qpd);
+static bool set_cache_memory_policy_v11(struct device_queue_manager *dqm,
+				   struct qcm_process_device *qpd,
+				   enum cache_policy default_policy,
+				   enum cache_policy alternate_policy,
+				   void __user *alternate_aperture_base,
+				   uint64_t alternate_aperture_size,
+				   u32 misc_process_properties);
+
+void device_queue_manager_init_v11(
+	struct device_queue_manager_asic_ops *asic_ops)
+{
+	asic_ops->set_cache_memory_policy = set_cache_memory_policy_v11;
+	asic_ops->update_qpd = update_qpd_v11;
+	asic_ops->init_sdma_vm = init_sdma_vm_v11;
+	asic_ops->mqd_manager_init = mqd_manager_init_v11;
+}
+
+static uint32_t compute_sh_mem_bases_64bit(struct kfd_process_device *pdd)
+{
+	uint32_t shared_base = pdd->lds_base >> 48;
+	uint32_t private_base = pdd->scratch_base >> 48;
+
+	return (shared_base << SH_MEM_BASES__SHARED_BASE__SHIFT) |
+		private_base;
+}
+
+static bool set_cache_memory_policy_v11(struct device_queue_manager *dqm,
+				   struct qcm_process_device *qpd,
+				   enum cache_policy default_policy,
+				   enum cache_policy alternate_policy,
+				   void __user *alternate_aperture_base,
+				   uint64_t alternate_aperture_size,
+				   u32 misc_process_properties)
+{
+	qpd->sh_mem_config = (SH_MEM_ALIGNMENT_MODE_UNALIGNED <<
+			      SH_MEM_CONFIG__ALIGNMENT_MODE__SHIFT) |
+			      (3 << SH_MEM_CONFIG__INITIAL_INST_PREFETCH__SHIFT);
+
+	qpd->sh_mem_ape1_limit = 0;
+	qpd->sh_mem_ape1_base = 0;
+	qpd->sh_mem_bases = compute_sh_mem_bases_64bit(qpd_to_pdd(qpd));
+
+	pr_debug("sh_mem_bases 0x%X\n", qpd->sh_mem_bases);
+	return true;
+}
+
+static int update_qpd_v11(struct device_queue_manager *dqm,
+			 struct qcm_process_device *qpd)
+{
+	return 0;
+}
+
+static void init_sdma_vm_v11(struct device_queue_manager *dqm, struct queue *q,
+			    struct qcm_process_device *qpd)
+{
+	/* Not needed on SDMAv4 onwards any more */
+	q->properties.sdma_vm_addr = 0;
+}
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v12.c b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v12.c
new file mode 100644
index 000000000000..3550da3a46f9
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v12.c
@@ -0,0 +1,90 @@
+// SPDX-License-Identifier: GPL-2.0 OR MIT
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "kfd_device_queue_manager.h"
+#include "gc/gc_12_0_0_sh_mask.h"
+#include "soc24_enum.h"
+
+static int update_qpd_v12(struct device_queue_manager *dqm,
+			 struct qcm_process_device *qpd);
+static void init_sdma_vm_v12(struct device_queue_manager *dqm, struct queue *q,
+			    struct qcm_process_device *qpd);
+static bool set_cache_memory_policy_v12(struct device_queue_manager *dqm,
+				   struct qcm_process_device *qpd,
+				   enum cache_policy default_policy,
+				   enum cache_policy alternate_policy,
+				   void __user *alternate_aperture_base,
+				   uint64_t alternate_aperture_size,
+				   u32 misc_process_properties);
+
+void device_queue_manager_init_v12(
+	struct device_queue_manager_asic_ops *asic_ops)
+{
+	asic_ops->set_cache_memory_policy = set_cache_memory_policy_v12;
+	asic_ops->update_qpd = update_qpd_v12;
+	asic_ops->init_sdma_vm = init_sdma_vm_v12;
+	asic_ops->mqd_manager_init = mqd_manager_init_v12;
+}
+
+static uint32_t compute_sh_mem_bases_64bit(struct kfd_process_device *pdd)
+{
+	uint32_t shared_base = pdd->lds_base >> 48;
+	uint32_t private_base = pdd->scratch_base >> 48;
+
+	return (shared_base << SH_MEM_BASES__SHARED_BASE__SHIFT) |
+		private_base;
+}
+
+static bool set_cache_memory_policy_v12(struct device_queue_manager *dqm,
+				   struct qcm_process_device *qpd,
+				   enum cache_policy default_policy,
+				   enum cache_policy alternate_policy,
+				   void __user *alternate_aperture_base,
+				   uint64_t alternate_aperture_size,
+				   u32 misc_process_properties)
+{
+	qpd->sh_mem_config = (SH_MEM_ALIGNMENT_MODE_UNALIGNED <<
+			      SH_MEM_CONFIG__ALIGNMENT_MODE__SHIFT) |
+			      (3 << SH_MEM_CONFIG__INITIAL_INST_PREFETCH__SHIFT);
+
+	qpd->sh_mem_ape1_limit = 0;
+	qpd->sh_mem_ape1_base = 0;
+	qpd->sh_mem_bases = compute_sh_mem_bases_64bit(qpd_to_pdd(qpd));
+
+	pr_debug("sh_mem_bases 0x%X\n", qpd->sh_mem_bases);
+	return true;
+}
+
+static int update_qpd_v12(struct device_queue_manager *dqm,
+			 struct qcm_process_device *qpd)
+{
+	return 0;
+}
+
+static void init_sdma_vm_v12(struct device_queue_manager *dqm, struct queue *q,
+			    struct qcm_process_device *qpd)
+{
+	/* Not needed on SDMAv4 onwards any more */
+	q->properties.sdma_vm_addr = 0;
+}
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v9.c b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v9.c
index b5c3d13643f1..9fcc8c6e57b7 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v9.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v9.c
@@ -1,5 +1,6 @@
+// SPDX-License-Identifier: GPL-2.0 OR MIT
 /*
- * Copyright 2016-2018 Advanced Micro Devices, Inc.
+ * Copyright 2016-2022 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -23,18 +24,24 @@
 
 #include "kfd_device_queue_manager.h"
 #include "vega10_enum.h"
-#include "gc/gc_9_0_offset.h"
-#include "gc/gc_9_0_sh_mask.h"
-#include "sdma0/sdma0_4_0_sh_mask.h"
+#include "gc/gc_9_4_3_sh_mask.h"
 
 static int update_qpd_v9(struct device_queue_manager *dqm,
 			 struct qcm_process_device *qpd);
 static void init_sdma_vm_v9(struct device_queue_manager *dqm, struct queue *q,
 			    struct qcm_process_device *qpd);
+static bool set_cache_memory_policy_v9(struct device_queue_manager *dqm,
+				   struct qcm_process_device *qpd,
+				   enum cache_policy default_policy,
+				   enum cache_policy alternate_policy,
+				   void __user *alternate_aperture_base,
+				   uint64_t alternate_aperture_size,
+				   u32 misc_process_properties);
 
 void device_queue_manager_init_v9(
 	struct device_queue_manager_asic_ops *asic_ops)
 {
+	asic_ops->set_cache_memory_policy = set_cache_memory_policy_v9;
 	asic_ops->update_qpd = update_qpd_v9;
 	asic_ops->init_sdma_vm = init_sdma_vm_v9;
 	asic_ops->mqd_manager_init = mqd_manager_init_v9;
@@ -49,39 +56,73 @@ static uint32_t compute_sh_mem_bases_64bit(struct kfd_process_device *pdd)
 		private_base;
 }
 
+static bool set_cache_memory_policy_v9(struct device_queue_manager *dqm,
+				   struct qcm_process_device *qpd,
+				   enum cache_policy default_policy,
+				   enum cache_policy alternate_policy,
+				   void __user *alternate_aperture_base,
+				   uint64_t alternate_aperture_size,
+				   u32 misc_process_properties)
+{
+	qpd->sh_mem_config = SH_MEM_ALIGNMENT_MODE_UNALIGNED <<
+				SH_MEM_CONFIG__ALIGNMENT_MODE__SHIFT;
+
+	if (dqm->dev->kfd->noretry)
+		qpd->sh_mem_config |= 1 << SH_MEM_CONFIG__RETRY_DISABLE__SHIFT;
+
+	if (KFD_GC_VERSION(dqm->dev->kfd) == IP_VERSION(9, 4, 3) ||
+		KFD_GC_VERSION(dqm->dev->kfd) == IP_VERSION(9, 4, 4))
+		qpd->sh_mem_config |= (1 << SH_MEM_CONFIG__F8_MODE__SHIFT);
+
+	if (KFD_GC_VERSION(dqm->dev->kfd) == IP_VERSION(9, 5, 0)) {
+		if (misc_process_properties & KFD_PROC_FLAG_MFMA_HIGH_PRECISION)
+			qpd->sh_mem_config |= 1 << SH_MEM_CONFIG__PRECISION_MODE__SHIFT;
+	}
+
+	qpd->sh_mem_ape1_limit = 0;
+	qpd->sh_mem_ape1_base = 0;
+	qpd->sh_mem_bases = compute_sh_mem_bases_64bit(qpd_to_pdd(qpd));
+
+	pr_debug("sh_mem_bases 0x%X sh_mem_config 0x%X\n", qpd->sh_mem_bases,
+		 qpd->sh_mem_config);
+	return true;
+}
+
 static int update_qpd_v9(struct device_queue_manager *dqm,
 			 struct qcm_process_device *qpd)
 {
-	struct kfd_process_device *pdd;
+	struct kfd_process_device *pdd = qpd_to_pdd(qpd);
 
 	pdd = qpd_to_pdd(qpd);
 
 	/* check if sh_mem_config register already configured */
 	if (qpd->sh_mem_config == 0) {
-		qpd->sh_mem_config =
-				SH_MEM_ALIGNMENT_MODE_UNALIGNED <<
+		qpd->sh_mem_config = SH_MEM_ALIGNMENT_MODE_UNALIGNED <<
 					SH_MEM_CONFIG__ALIGNMENT_MODE__SHIFT;
 
-		if (dqm->dev->device_info->asic_family == CHIP_ALDEBARAN) {
-			/* Aldebaran can safely support different XNACK modes
-			 * per process
-			 */
-			if (!pdd->process->xnack_enabled)
-				qpd->sh_mem_config |=
-					1 << SH_MEM_CONFIG__RETRY_DISABLE__SHIFT;
-		} else if (dqm->dev->noretry &&
-			   !dqm->dev->use_iommu_v2) {
+		if (dqm->dev->kfd->noretry)
+			qpd->sh_mem_config |= 1 << SH_MEM_CONFIG__RETRY_DISABLE__SHIFT;
+
+		if (KFD_GC_VERSION(dqm->dev->kfd) == IP_VERSION(9, 4, 3) ||
+		    KFD_GC_VERSION(dqm->dev->kfd) == IP_VERSION(9, 4, 4))
 			qpd->sh_mem_config |=
-				1 << SH_MEM_CONFIG__RETRY_DISABLE__SHIFT;
-		}
+				(1 << SH_MEM_CONFIG__F8_MODE__SHIFT);
 
 		qpd->sh_mem_ape1_limit = 0;
 		qpd->sh_mem_ape1_base = 0;
 	}
 
+	if (KFD_SUPPORT_XNACK_PER_PROCESS(dqm->dev)) {
+		if (!pdd->process->xnack_enabled)
+			qpd->sh_mem_config |= 1 << SH_MEM_CONFIG__RETRY_DISABLE__SHIFT;
+		else
+			qpd->sh_mem_config &= ~(1 << SH_MEM_CONFIG__RETRY_DISABLE__SHIFT);
+	}
+
 	qpd->sh_mem_bases = compute_sh_mem_bases_64bit(pdd);
 
-	pr_debug("sh_mem_bases 0x%X\n", qpd->sh_mem_bases);
+	pr_debug("sh_mem_bases 0x%X sh_mem_config 0x%X\n", qpd->sh_mem_bases,
+		 qpd->sh_mem_config);
 
 	return 0;
 }
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_vi.c b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_vi.c
index 3a7cb2f88366..dad83356e976 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_vi.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_vi.c
@@ -1,5 +1,6 @@
+// SPDX-License-Identifier: GPL-2.0 OR MIT
 /*
- * Copyright 2014 Advanced Micro Devices, Inc.
+ * Copyright 2014-2022 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -26,30 +27,29 @@
 #include "gca/gfx_8_0_sh_mask.h"
 #include "oss/oss_3_0_sh_mask.h"
 
+/*
+ * Low bits must be 0000/FFFF as required by HW, high bits must be 0 to
+ * stay in user mode.
+ */
+#define APE1_FIXED_BITS_MASK 0xFFFF80000000FFFFULL
+/* APE1 limit is inclusive and 64K aligned. */
+#define APE1_LIMIT_ALIGNMENT 0xFFFF
+
 static bool set_cache_memory_policy_vi(struct device_queue_manager *dqm,
-				   struct qcm_process_device *qpd,
-				   enum cache_policy default_policy,
-				   enum cache_policy alternate_policy,
-				   void __user *alternate_aperture_base,
-				   uint64_t alternate_aperture_size);
-static bool set_cache_memory_policy_vi_tonga(struct device_queue_manager *dqm,
-			struct qcm_process_device *qpd,
-			enum cache_policy default_policy,
-			enum cache_policy alternate_policy,
-			void __user *alternate_aperture_base,
-			uint64_t alternate_aperture_size);
+				       struct qcm_process_device *qpd,
+				       enum cache_policy default_policy,
+				       enum cache_policy alternate_policy,
+				       void __user *alternate_aperture_base,
+				       uint64_t alternate_aperture_size,
+				       u32 misc_process_properties);
 static int update_qpd_vi(struct device_queue_manager *dqm,
-					struct qcm_process_device *qpd);
-static int update_qpd_vi_tonga(struct device_queue_manager *dqm,
-			struct qcm_process_device *qpd);
-static void init_sdma_vm(struct device_queue_manager *dqm, struct queue *q,
-				struct qcm_process_device *qpd);
-static void init_sdma_vm_tonga(struct device_queue_manager *dqm,
-			struct queue *q,
-			struct qcm_process_device *qpd);
+			 struct qcm_process_device *qpd);
+static void init_sdma_vm(struct device_queue_manager *dqm,
+			 struct queue *q,
+			 struct qcm_process_device *qpd);
 
 void device_queue_manager_init_vi(
-		struct device_queue_manager_asic_ops *asic_ops)
+	struct device_queue_manager_asic_ops *asic_ops)
 {
 	asic_ops->set_cache_memory_policy = set_cache_memory_policy_vi;
 	asic_ops->update_qpd = update_qpd_vi;
@@ -57,15 +57,6 @@ void device_queue_manager_init_vi(
 	asic_ops->mqd_manager_init = mqd_manager_init_vi;
 }
 
-void device_queue_manager_init_vi_tonga(
-		struct device_queue_manager_asic_ops *asic_ops)
-{
-	asic_ops->set_cache_memory_policy = set_cache_memory_policy_vi_tonga;
-	asic_ops->update_qpd = update_qpd_vi_tonga;
-	asic_ops->init_sdma_vm = init_sdma_vm_tonga;
-	asic_ops->mqd_manager_init = mqd_manager_init_vi_tonga;
-}
-
 static uint32_t compute_sh_mem_bases_64bit(unsigned int top_address_nybble)
 {
 	/* In 64-bit mode, we can only control the top 3 bits of the LDS,
@@ -95,43 +86,45 @@ static uint32_t compute_sh_mem_bases_64bit(unsigned int top_address_nybble)
 }
 
 static bool set_cache_memory_policy_vi(struct device_queue_manager *dqm,
-				   struct qcm_process_device *qpd,
-				   enum cache_policy default_policy,
-				   enum cache_policy alternate_policy,
-				   void __user *alternate_aperture_base,
-				   uint64_t alternate_aperture_size)
-{
-	uint32_t default_mtype;
-	uint32_t ape1_mtype;
-
-	default_mtype = (default_policy == cache_policy_coherent) ?
-			MTYPE_CC :
-			MTYPE_NC;
-
-	ape1_mtype = (alternate_policy == cache_policy_coherent) ?
-			MTYPE_CC :
-			MTYPE_NC;
-
-	qpd->sh_mem_config = (qpd->sh_mem_config &
-			SH_MEM_CONFIG__ADDRESS_MODE_MASK) |
-		SH_MEM_ALIGNMENT_MODE_UNALIGNED <<
-				SH_MEM_CONFIG__ALIGNMENT_MODE__SHIFT |
-		default_mtype << SH_MEM_CONFIG__DEFAULT_MTYPE__SHIFT |
-		ape1_mtype << SH_MEM_CONFIG__APE1_MTYPE__SHIFT |
-		SH_MEM_CONFIG__PRIVATE_ATC_MASK;
-
-	return true;
-}
-
-static bool set_cache_memory_policy_vi_tonga(struct device_queue_manager *dqm,
 		struct qcm_process_device *qpd,
 		enum cache_policy default_policy,
 		enum cache_policy alternate_policy,
 		void __user *alternate_aperture_base,
-		uint64_t alternate_aperture_size)
+		uint64_t alternate_aperture_size,
+		u32 misc_process_properties)
 {
 	uint32_t default_mtype;
 	uint32_t ape1_mtype;
+	unsigned int temp;
+	bool retval = true;
+
+	if (alternate_aperture_size == 0) {
+		/* base > limit disables APE1 */
+		qpd->sh_mem_ape1_base = 1;
+		qpd->sh_mem_ape1_limit = 0;
+	} else {
+		/*
+		 * In FSA64, APE1_Base[63:0] = { 16{SH_MEM_APE1_BASE[31]},
+		 *			SH_MEM_APE1_BASE[31:0], 0x0000 }
+		 * APE1_Limit[63:0] = { 16{SH_MEM_APE1_LIMIT[31]},
+		 *			SH_MEM_APE1_LIMIT[31:0], 0xFFFF }
+		 * Verify that the base and size parameters can be
+		 * represented in this format and convert them.
+		 * Additionally restrict APE1 to user-mode addresses.
+		 */
+
+		uint64_t base = (uintptr_t)alternate_aperture_base;
+		uint64_t limit = base + alternate_aperture_size - 1;
+
+		if (limit <= base || (base & APE1_FIXED_BITS_MASK) != 0 ||
+		   (limit & APE1_FIXED_BITS_MASK) != APE1_LIMIT_ALIGNMENT) {
+			retval = false;
+			goto out;
+		}
+
+		qpd->sh_mem_ape1_base = base >> 16;
+		qpd->sh_mem_ape1_limit = limit >> 16;
+	}
 
 	default_mtype = (default_policy == cache_policy_coherent) ?
 			MTYPE_UC :
@@ -147,103 +140,27 @@ static bool set_cache_memory_policy_vi_tonga(struct device_queue_manager *dqm,
 			default_mtype << SH_MEM_CONFIG__DEFAULT_MTYPE__SHIFT |
 			ape1_mtype << SH_MEM_CONFIG__APE1_MTYPE__SHIFT;
 
-	return true;
-}
-
-static int update_qpd_vi(struct device_queue_manager *dqm,
-					struct qcm_process_device *qpd)
-{
-	struct kfd_process_device *pdd;
-	unsigned int temp;
-
-	pdd = qpd_to_pdd(qpd);
-
-	/* check if sh_mem_config register already configured */
-	if (qpd->sh_mem_config == 0) {
-		qpd->sh_mem_config =
-			SH_MEM_ALIGNMENT_MODE_UNALIGNED <<
-				SH_MEM_CONFIG__ALIGNMENT_MODE__SHIFT |
-			MTYPE_CC << SH_MEM_CONFIG__DEFAULT_MTYPE__SHIFT |
-			MTYPE_CC << SH_MEM_CONFIG__APE1_MTYPE__SHIFT |
-			SH_MEM_CONFIG__PRIVATE_ATC_MASK;
-
-		qpd->sh_mem_ape1_limit = 0;
-		qpd->sh_mem_ape1_base = 0;
-	}
-
-	if (qpd->pqm->process->is_32bit_user_mode) {
-		temp = get_sh_mem_bases_32(pdd);
-		qpd->sh_mem_bases = temp << SH_MEM_BASES__SHARED_BASE__SHIFT;
-		qpd->sh_mem_config |= SH_MEM_ADDRESS_MODE_HSA32 <<
-					SH_MEM_CONFIG__ADDRESS_MODE__SHIFT;
-	} else {
-		temp = get_sh_mem_bases_nybble_64(pdd);
-		qpd->sh_mem_bases = compute_sh_mem_bases_64bit(temp);
-		qpd->sh_mem_config |= SH_MEM_ADDRESS_MODE_HSA64 <<
-			SH_MEM_CONFIG__ADDRESS_MODE__SHIFT;
-		qpd->sh_mem_config |= 1  <<
-			SH_MEM_CONFIG__PRIVATE_ATC__SHIFT;
-	}
-
-	pr_debug("is32bit process: %d sh_mem_bases nybble: 0x%X and register 0x%X\n",
-		qpd->pqm->process->is_32bit_user_mode, temp, qpd->sh_mem_bases);
-
-	return 0;
-}
-
-static int update_qpd_vi_tonga(struct device_queue_manager *dqm,
-			struct qcm_process_device *qpd)
-{
-	struct kfd_process_device *pdd;
-	unsigned int temp;
-
-	pdd = qpd_to_pdd(qpd);
-
-	/* check if sh_mem_config register already configured */
-	if (qpd->sh_mem_config == 0) {
-		qpd->sh_mem_config =
-				SH_MEM_ALIGNMENT_MODE_UNALIGNED <<
-					SH_MEM_CONFIG__ALIGNMENT_MODE__SHIFT |
-				MTYPE_UC <<
-					SH_MEM_CONFIG__DEFAULT_MTYPE__SHIFT |
-				MTYPE_UC <<
-					SH_MEM_CONFIG__APE1_MTYPE__SHIFT;
-
-		qpd->sh_mem_ape1_limit = 0;
-		qpd->sh_mem_ape1_base = 0;
-	}
-
 	/* On dGPU we're always in GPUVM64 addressing mode with 64-bit
 	 * aperture addresses.
 	 */
-	temp = get_sh_mem_bases_nybble_64(pdd);
+	temp = get_sh_mem_bases_nybble_64(qpd_to_pdd(qpd));
 	qpd->sh_mem_bases = compute_sh_mem_bases_64bit(temp);
 
 	pr_debug("sh_mem_bases nybble: 0x%X and register 0x%X\n",
 		temp, qpd->sh_mem_bases);
-
-	return 0;
+out:
+	return retval;
 }
 
-static void init_sdma_vm(struct device_queue_manager *dqm, struct queue *q,
-				struct qcm_process_device *qpd)
+static int update_qpd_vi(struct device_queue_manager *dqm,
+			 struct qcm_process_device *qpd)
 {
-	uint32_t value = (1 << SDMA0_RLC0_VIRTUAL_ADDR__ATC__SHIFT);
-
-	if (q->process->is_32bit_user_mode)
-		value |= (1 << SDMA0_RLC0_VIRTUAL_ADDR__PTR32__SHIFT) |
-				get_sh_mem_bases_32(qpd_to_pdd(qpd));
-	else
-		value |= ((get_sh_mem_bases_nybble_64(qpd_to_pdd(qpd))) <<
-				SDMA0_RLC0_VIRTUAL_ADDR__SHARED_BASE__SHIFT) &
-				SDMA0_RLC0_VIRTUAL_ADDR__SHARED_BASE_MASK;
-
-	q->properties.sdma_vm_addr = value;
+	return 0;
 }
 
-static void init_sdma_vm_tonga(struct device_queue_manager *dqm,
-			struct queue *q,
-			struct qcm_process_device *qpd)
+static void init_sdma_vm(struct device_queue_manager *dqm,
+			 struct queue *q,
+			 struct qcm_process_device *qpd)
 {
 	/* On dGPU we're always in GPUVM64 addressing mode with 64-bit
 	 * aperture addresses.
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_doorbell.c b/drivers/gpu/drm/amd/amdkfd/kfd_doorbell.c
index 768d153acff4..05c74887fd6f 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_doorbell.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_doorbell.c
@@ -1,5 +1,6 @@
+// SPDX-License-Identifier: GPL-2.0 OR MIT
 /*
- * Copyright 2014 Advanced Micro Devices, Inc.
+ * Copyright 2014-2022 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -48,82 +49,61 @@
 /* # of doorbell bytes allocated for each process. */
 size_t kfd_doorbell_process_slice(struct kfd_dev *kfd)
 {
-	return roundup(kfd->device_info->doorbell_size *
-			KFD_MAX_NUM_OF_QUEUES_PER_PROCESS,
-			PAGE_SIZE);
+	if (!kfd->shared_resources.enable_mes)
+		return roundup(kfd->device_info.doorbell_size *
+				KFD_MAX_NUM_OF_QUEUES_PER_PROCESS,
+				PAGE_SIZE);
+	else
+		return amdgpu_mes_doorbell_process_slice(
+					(struct amdgpu_device *)kfd->adev);
 }
 
 /* Doorbell calculations for device init. */
 int kfd_doorbell_init(struct kfd_dev *kfd)
 {
-	size_t doorbell_start_offset;
-	size_t doorbell_aperture_size;
-	size_t doorbell_process_limit;
+	int size = PAGE_SIZE;
+	int r;
 
 	/*
-	 * We start with calculations in bytes because the input data might
-	 * only be byte-aligned.
-	 * Only after we have done the rounding can we assume any alignment.
+	 * Todo: KFD kernel level operations need only one doorbell for
+	 * ring test/HWS. So instead of reserving a whole page here for
+	 * kernel, reserve and consume a doorbell from existing KGD kernel
+	 * doorbell page.
 	 */
 
-	doorbell_start_offset =
-			roundup(kfd->shared_resources.doorbell_start_offset,
-					kfd_doorbell_process_slice(kfd));
-
-	doorbell_aperture_size =
-			rounddown(kfd->shared_resources.doorbell_aperture_size,
-					kfd_doorbell_process_slice(kfd));
-
-	if (doorbell_aperture_size > doorbell_start_offset)
-		doorbell_process_limit =
-			(doorbell_aperture_size - doorbell_start_offset) /
-						kfd_doorbell_process_slice(kfd);
-	else
-		return -ENOSPC;
-
-	if (!kfd->max_doorbell_slices ||
-	    doorbell_process_limit < kfd->max_doorbell_slices)
-		kfd->max_doorbell_slices = doorbell_process_limit;
-
-	kfd->doorbell_base = kfd->shared_resources.doorbell_physical_address +
-				doorbell_start_offset;
-
-	kfd->doorbell_base_dw_offset = doorbell_start_offset / sizeof(u32);
-
-	kfd->doorbell_kernel_ptr = ioremap(kfd->doorbell_base,
-					   kfd_doorbell_process_slice(kfd));
-
-	if (!kfd->doorbell_kernel_ptr)
+	/* Bitmap to dynamically allocate doorbells from kernel page */
+	kfd->doorbell_bitmap = bitmap_zalloc(size / sizeof(u32), GFP_KERNEL);
+	if (!kfd->doorbell_bitmap) {
+		DRM_ERROR("Failed to allocate kernel doorbell bitmap\n");
 		return -ENOMEM;
+	}
 
-	pr_debug("Doorbell initialization:\n");
-	pr_debug("doorbell base           == 0x%08lX\n",
-			(uintptr_t)kfd->doorbell_base);
-
-	pr_debug("doorbell_base_dw_offset      == 0x%08lX\n",
-			kfd->doorbell_base_dw_offset);
-
-	pr_debug("doorbell_process_limit  == 0x%08lX\n",
-			doorbell_process_limit);
-
-	pr_debug("doorbell_kernel_offset  == 0x%08lX\n",
-			(uintptr_t)kfd->doorbell_base);
-
-	pr_debug("doorbell aperture size  == 0x%08lX\n",
-			kfd->shared_resources.doorbell_aperture_size);
-
-	pr_debug("doorbell kernel address == %p\n", kfd->doorbell_kernel_ptr);
+	/* Alloc a doorbell page for KFD kernel usages */
+	r = amdgpu_bo_create_kernel(kfd->adev,
+				    size,
+				    PAGE_SIZE,
+				    AMDGPU_GEM_DOMAIN_DOORBELL,
+				    &kfd->doorbells,
+				    NULL,
+				    (void **)&kfd->doorbell_kernel_ptr);
+	if (r) {
+		pr_err("failed to allocate kernel doorbells\n");
+		bitmap_free(kfd->doorbell_bitmap);
+		return r;
+	}
 
+	pr_debug("Doorbell kernel address == %p\n", kfd->doorbell_kernel_ptr);
 	return 0;
 }
 
 void kfd_doorbell_fini(struct kfd_dev *kfd)
 {
-	if (kfd->doorbell_kernel_ptr)
-		iounmap(kfd->doorbell_kernel_ptr);
+	bitmap_free(kfd->doorbell_bitmap);
+	amdgpu_bo_free_kernel(&kfd->doorbells, NULL,
+			     (void **)&kfd->doorbell_kernel_ptr);
 }
 
-int kfd_doorbell_mmap(struct kfd_dev *dev, struct kfd_process *process,
+int kfd_doorbell_mmap(struct kfd_node *dev, struct kfd_process *process,
 		      struct vm_area_struct *vma)
 {
 	phys_addr_t address;
@@ -133,7 +113,7 @@ int kfd_doorbell_mmap(struct kfd_dev *dev, struct kfd_process *process,
 	 * For simplicitly we only allow mapping of the entire doorbell
 	 * allocation of a single device & process.
 	 */
-	if (vma->vm_end - vma->vm_start != kfd_doorbell_process_slice(dev))
+	if (vma->vm_end - vma->vm_start != kfd_doorbell_process_slice(dev->kfd))
 		return -EINVAL;
 
 	pdd = kfd_get_process_device_data(dev, process);
@@ -142,8 +122,10 @@ int kfd_doorbell_mmap(struct kfd_dev *dev, struct kfd_process *process,
 
 	/* Calculate physical address of doorbell */
 	address = kfd_get_process_doorbells(pdd);
-	vma->vm_flags |= VM_IO | VM_DONTCOPY | VM_DONTEXPAND | VM_NORESERVE |
-				VM_DONTDUMP | VM_PFNMAP;
+	if (!address)
+		return -ENOMEM;
+	vm_flags_set(vma, VM_IO | VM_DONTCOPY | VM_DONTEXPAND | VM_NORESERVE |
+				VM_DONTDUMP | VM_PFNMAP);
 
 	vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot);
 
@@ -153,13 +135,13 @@ int kfd_doorbell_mmap(struct kfd_dev *dev, struct kfd_process *process,
 		 "     vm_flags            == 0x%04lX\n"
 		 "     size                == 0x%04lX\n",
 		 (unsigned long long) vma->vm_start, address, vma->vm_flags,
-		 kfd_doorbell_process_slice(dev));
+		 kfd_doorbell_process_slice(dev->kfd));
 
 
 	return io_remap_pfn_range(vma,
 				vma->vm_start,
 				address >> PAGE_SHIFT,
-				kfd_doorbell_process_slice(dev),
+				kfd_doorbell_process_slice(dev->kfd),
 				vma->vm_page_prot);
 }
 
@@ -171,22 +153,19 @@ void __iomem *kfd_get_kernel_doorbell(struct kfd_dev *kfd,
 	u32 inx;
 
 	mutex_lock(&kfd->doorbell_mutex);
-	inx = find_first_zero_bit(kfd->doorbell_available_index,
-					KFD_MAX_NUM_OF_QUEUES_PER_PROCESS);
+	inx = find_first_zero_bit(kfd->doorbell_bitmap, PAGE_SIZE / sizeof(u32));
 
-	__set_bit(inx, kfd->doorbell_available_index);
+	__set_bit(inx, kfd->doorbell_bitmap);
 	mutex_unlock(&kfd->doorbell_mutex);
 
 	if (inx >= KFD_MAX_NUM_OF_QUEUES_PER_PROCESS)
 		return NULL;
 
-	inx *= kfd->device_info->doorbell_size / sizeof(u32);
-
-	/*
-	 * Calculating the kernel doorbell offset using the first
-	 * doorbell page.
-	 */
-	*doorbell_off = kfd->doorbell_base_dw_offset + inx;
+	*doorbell_off = amdgpu_doorbell_index_on_bar(kfd->adev,
+						     kfd->doorbells,
+						     inx,
+						     kfd->device_info.doorbell_size);
+	inx *= 2;
 
 	pr_debug("Get kernel queue doorbell\n"
 			"     doorbell offset   == 0x%08X\n"
@@ -200,11 +179,11 @@ void kfd_release_kernel_doorbell(struct kfd_dev *kfd, u32 __iomem *db_addr)
 {
 	unsigned int inx;
 
-	inx = (unsigned int)(db_addr - kfd->doorbell_kernel_ptr)
-		* sizeof(u32) / kfd->device_info->doorbell_size;
+	inx = (unsigned int)(db_addr - kfd->doorbell_kernel_ptr);
+	inx /= 2;
 
 	mutex_lock(&kfd->doorbell_mutex);
-	__clear_bit(inx, kfd->doorbell_available_index);
+	__clear_bit(inx, kfd->doorbell_bitmap);
 	mutex_unlock(&kfd->doorbell_mutex);
 }
 
@@ -226,50 +205,99 @@ void write_kernel_doorbell64(void __iomem *db, u64 value)
 	}
 }
 
-unsigned int kfd_get_doorbell_dw_offset_in_bar(struct kfd_dev *kfd,
-					struct kfd_process_device *pdd,
-					unsigned int doorbell_id)
+static int init_doorbell_bitmap(struct qcm_process_device *qpd,
+				struct kfd_dev *dev)
 {
-	/*
-	 * doorbell_base_dw_offset accounts for doorbells taken by KGD.
-	 * index * kfd_doorbell_process_slice/sizeof(u32) adjusts to
-	 * the process's doorbells. The offset returned is in dword
-	 * units regardless of the ASIC-dependent doorbell size.
-	 */
-	return kfd->doorbell_base_dw_offset +
-		pdd->doorbell_index
-		* kfd_doorbell_process_slice(kfd) / sizeof(u32) +
-		doorbell_id * kfd->device_info->doorbell_size / sizeof(u32);
+	unsigned int i;
+	int range_start = dev->shared_resources.non_cp_doorbells_start;
+	int range_end = dev->shared_resources.non_cp_doorbells_end;
+
+	if (!KFD_IS_SOC15(dev))
+		return 0;
+
+	/* Mask out doorbells reserved for SDMA, IH, and VCN on SOC15. */
+	pr_debug("reserved doorbell 0x%03x - 0x%03x\n", range_start, range_end);
+	pr_debug("reserved doorbell 0x%03x - 0x%03x\n",
+			range_start + KFD_QUEUE_DOORBELL_MIRROR_OFFSET,
+			range_end + KFD_QUEUE_DOORBELL_MIRROR_OFFSET);
+
+	for (i = 0; i < KFD_MAX_NUM_OF_QUEUES_PER_PROCESS / 2; i++) {
+		if (i >= range_start && i <= range_end) {
+			__set_bit(i, qpd->doorbell_bitmap);
+			__set_bit(i + KFD_QUEUE_DOORBELL_MIRROR_OFFSET,
+				  qpd->doorbell_bitmap);
+		}
+	}
+
+	return 0;
 }
 
-uint64_t kfd_get_number_elems(struct kfd_dev *kfd)
+phys_addr_t kfd_get_process_doorbells(struct kfd_process_device *pdd)
 {
-	uint64_t num_of_elems = (kfd->shared_resources.doorbell_aperture_size -
-				kfd->shared_resources.doorbell_start_offset) /
-					kfd_doorbell_process_slice(kfd) + 1;
+	struct amdgpu_device *adev = pdd->dev->adev;
+	uint32_t first_db_index;
 
-	return num_of_elems;
+	if (!pdd->qpd.proc_doorbells) {
+		if (kfd_alloc_process_doorbells(pdd->dev->kfd, pdd))
+			/* phys_addr_t 0 is error */
+			return 0;
+	}
 
+	first_db_index = amdgpu_doorbell_index_on_bar(adev,
+						      pdd->qpd.proc_doorbells,
+						      0,
+						      pdd->dev->kfd->device_info.doorbell_size);
+	return adev->doorbell.base + first_db_index * sizeof(uint32_t);
 }
 
-phys_addr_t kfd_get_process_doorbells(struct kfd_process_device *pdd)
+int kfd_alloc_process_doorbells(struct kfd_dev *kfd, struct kfd_process_device *pdd)
 {
-	return pdd->dev->doorbell_base +
-		pdd->doorbell_index * kfd_doorbell_process_slice(pdd->dev);
-}
+	int r;
+	struct qcm_process_device *qpd = &pdd->qpd;
+
+	/* Allocate bitmap for dynamic doorbell allocation */
+	qpd->doorbell_bitmap = bitmap_zalloc(KFD_MAX_NUM_OF_QUEUES_PER_PROCESS,
+					     GFP_KERNEL);
+	if (!qpd->doorbell_bitmap) {
+		DRM_ERROR("Failed to allocate process doorbell bitmap\n");
+		return -ENOMEM;
+	}
 
-int kfd_alloc_process_doorbells(struct kfd_dev *kfd, unsigned int *doorbell_index)
-{
-	int r = ida_simple_get(&kfd->doorbell_ida, 1, kfd->max_doorbell_slices,
-				GFP_KERNEL);
-	if (r > 0)
-		*doorbell_index = r;
+	r = init_doorbell_bitmap(&pdd->qpd, kfd);
+	if (r) {
+		DRM_ERROR("Failed to initialize process doorbells\n");
+		r = -ENOMEM;
+		goto err;
+	}
+
+	/* Allocate doorbells for this process */
+	r = amdgpu_bo_create_kernel(kfd->adev,
+				    kfd_doorbell_process_slice(kfd),
+				    PAGE_SIZE,
+				    AMDGPU_GEM_DOMAIN_DOORBELL,
+				    &qpd->proc_doorbells,
+				    NULL,
+				    NULL);
+	if (r) {
+		DRM_ERROR("Failed to allocate process doorbells\n");
+		goto err;
+	}
+	return 0;
 
+err:
+	bitmap_free(qpd->doorbell_bitmap);
+	qpd->doorbell_bitmap = NULL;
 	return r;
 }
 
-void kfd_free_process_doorbells(struct kfd_dev *kfd, unsigned int doorbell_index)
+void kfd_free_process_doorbells(struct kfd_dev *kfd, struct kfd_process_device *pdd)
 {
-	if (doorbell_index)
-		ida_simple_remove(&kfd->doorbell_ida, doorbell_index);
+	struct qcm_process_device *qpd = &pdd->qpd;
+
+	if (qpd->doorbell_bitmap) {
+		bitmap_free(qpd->doorbell_bitmap);
+		qpd->doorbell_bitmap = NULL;
+	}
+
+	amdgpu_bo_free_kernel(&qpd->proc_doorbells, NULL, NULL);
 }
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_events.c b/drivers/gpu/drm/amd/amdkfd/kfd_events.c
index 3eea4edee355..5a190dd6be4e 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_events.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_events.c
@@ -1,5 +1,6 @@
+// SPDX-License-Identifier: GPL-2.0 OR MIT
 /*
- * Copyright 2014 Advanced Micro Devices, Inc.
+ * Copyright 2014-2022 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -30,7 +31,7 @@
 #include <linux/memory.h>
 #include "kfd_priv.h"
 #include "kfd_events.h"
-#include "kfd_iommu.h"
+#include "kfd_device_queue_manager.h"
 #include <linux/device.h>
 
 /*
@@ -40,6 +41,7 @@ struct kfd_event_waiter {
 	wait_queue_entry_t wait;
 	struct kfd_event *event; /* Event to wait for */
 	bool activated;		 /* Becomes true when event is signaled */
+	bool event_age_enabled;  /* set to true when last_event_age is non-zero */
 };
 
 /*
@@ -55,7 +57,6 @@ struct kfd_signal_page {
 	bool need_to_free_pages;
 };
 
-
 static uint64_t *page_slots(struct kfd_signal_page *page)
 {
 	return page->kernel_address;
@@ -92,7 +93,8 @@ fail_alloc_signal_store:
 }
 
 static int allocate_event_notification_slot(struct kfd_process *p,
-					    struct kfd_event *ev)
+					    struct kfd_event *ev,
+					    const int *restore_id)
 {
 	int id;
 
@@ -104,14 +106,19 @@ static int allocate_event_notification_slot(struct kfd_process *p,
 		p->signal_mapped_size = 256*8;
 	}
 
-	/*
-	 * Compatibility with old user mode: Only use signal slots
-	 * user mode has mapped, may be less than
-	 * KFD_SIGNAL_EVENT_LIMIT. This also allows future increase
-	 * of the event limit without breaking user mode.
-	 */
-	id = idr_alloc(&p->event_idr, ev, 0, p->signal_mapped_size / 8,
-		       GFP_KERNEL);
+	if (restore_id) {
+		id = idr_alloc(&p->event_idr, ev, *restore_id, *restore_id + 1,
+				GFP_KERNEL);
+	} else {
+		/*
+		 * Compatibility with old user mode: Only use signal slots
+		 * user mode has mapped, may be less than
+		 * KFD_SIGNAL_EVENT_LIMIT. This also allows future increase
+		 * of the event limit without breaking user mode.
+		 */
+		id = idr_alloc(&p->event_idr, ev, 0, p->signal_mapped_size / 8,
+				GFP_KERNEL);
+	}
 	if (id < 0)
 		return id;
 
@@ -122,8 +129,8 @@ static int allocate_event_notification_slot(struct kfd_process *p,
 }
 
 /*
- * Assumes that p->event_mutex is held and of course that p is not going
- * away (current or locked).
+ * Assumes that p->event_mutex or rcu_readlock is held and of course that p is
+ * not going away.
  */
 static struct kfd_event *lookup_event_by_id(struct kfd_process *p, uint32_t id)
 {
@@ -178,9 +185,8 @@ static struct kfd_event *lookup_signaled_event_by_partial_id(
 	return ev;
 }
 
-static int create_signal_event(struct file *devkfd,
-				struct kfd_process *p,
-				struct kfd_event *ev)
+static int create_signal_event(struct file *devkfd, struct kfd_process *p,
+				struct kfd_event *ev, const int *restore_id)
 {
 	int ret;
 
@@ -193,7 +199,7 @@ static int create_signal_event(struct file *devkfd,
 		return -ENOSPC;
 	}
 
-	ret = allocate_event_notification_slot(p, ev);
+	ret = allocate_event_notification_slot(p, ev, restore_id);
 	if (ret) {
 		pr_warn("Signal event wasn't created because out of kernel memory\n");
 		return ret;
@@ -209,16 +215,22 @@ static int create_signal_event(struct file *devkfd,
 	return 0;
 }
 
-static int create_other_event(struct kfd_process *p, struct kfd_event *ev)
+static int create_other_event(struct kfd_process *p, struct kfd_event *ev, const int *restore_id)
 {
-	/* Cast KFD_LAST_NONSIGNAL_EVENT to uint32_t. This allows an
-	 * intentional integer overflow to -1 without a compiler
-	 * warning. idr_alloc treats a negative value as "maximum
-	 * signed integer".
-	 */
-	int id = idr_alloc(&p->event_idr, ev, KFD_FIRST_NONSIGNAL_EVENT_ID,
-			   (uint32_t)KFD_LAST_NONSIGNAL_EVENT_ID + 1,
-			   GFP_KERNEL);
+	int id;
+
+	if (restore_id)
+		id = idr_alloc(&p->event_idr, ev, *restore_id, *restore_id + 1,
+			GFP_KERNEL);
+	else
+		/* Cast KFD_LAST_NONSIGNAL_EVENT to uint32_t. This allows an
+		 * intentional integer overflow to -1 without a compiler
+		 * warning. idr_alloc treats a negative value as "maximum
+		 * signed integer".
+		 */
+		id = idr_alloc(&p->event_idr, ev, KFD_FIRST_NONSIGNAL_EVENT_ID,
+				(uint32_t)KFD_LAST_NONSIGNAL_EVENT_ID + 1,
+				GFP_KERNEL);
 
 	if (id < 0)
 		return id;
@@ -227,12 +239,24 @@ static int create_other_event(struct kfd_process *p, struct kfd_event *ev)
 	return 0;
 }
 
-void kfd_event_init_process(struct kfd_process *p)
+int kfd_event_init_process(struct kfd_process *p)
 {
+	int id;
+
 	mutex_init(&p->event_mutex);
 	idr_init(&p->event_idr);
 	p->signal_page = NULL;
-	p->signal_event_count = 0;
+	p->signal_event_count = 1;
+	/* Allocate event ID 0. It is used for a fast path to ignore bogus events
+	 * that are sent by the CP without a context ID
+	 */
+	id = idr_alloc(&p->event_idr, NULL, 0, 1, GFP_KERNEL);
+	if (id < 0) {
+		idr_destroy(&p->event_idr);
+		mutex_destroy(&p->event_mutex);
+		return id;
+	}
+	return 0;
 }
 
 static void destroy_event(struct kfd_process *p, struct kfd_event *ev)
@@ -240,16 +264,18 @@ static void destroy_event(struct kfd_process *p, struct kfd_event *ev)
 	struct kfd_event_waiter *waiter;
 
 	/* Wake up pending waiters. They will return failure */
+	spin_lock(&ev->lock);
 	list_for_each_entry(waiter, &ev->wq.head, wait.entry)
-		waiter->event = NULL;
+		WRITE_ONCE(waiter->event, NULL);
 	wake_up_all(&ev->wq);
+	spin_unlock(&ev->lock);
 
 	if (ev->type == KFD_EVENT_TYPE_SIGNAL ||
 	    ev->type == KFD_EVENT_TYPE_DEBUG)
 		p->signal_event_count--;
 
 	idr_remove(&p->event_idr, ev->event_id);
-	kfree(ev);
+	kfree_rcu(ev, rcu);
 }
 
 static void destroy_events(struct kfd_process *p)
@@ -258,8 +284,10 @@ static void destroy_events(struct kfd_process *p)
 	uint32_t id;
 
 	idr_for_each_entry(&p->event_idr, ev, id)
-		destroy_event(p, ev);
+		if (ev)
+			destroy_event(p, ev);
 	idr_destroy(&p->event_idr);
+	mutex_destroy(&p->event_mutex);
 }
 
 /*
@@ -295,8 +323,8 @@ static bool event_can_be_cpu_signaled(const struct kfd_event *ev)
 	return ev->type == KFD_EVENT_TYPE_SIGNAL;
 }
 
-int kfd_event_page_set(struct kfd_process *p, void *kernel_address,
-		       uint64_t size)
+static int kfd_event_page_set(struct kfd_process *p, void *kernel_address,
+		       uint64_t size, uint64_t user_handle)
 {
 	struct kfd_signal_page *page;
 
@@ -315,10 +343,56 @@ int kfd_event_page_set(struct kfd_process *p, void *kernel_address,
 
 	p->signal_page = page;
 	p->signal_mapped_size = size;
-
+	p->signal_handle = user_handle;
 	return 0;
 }
 
+int kfd_kmap_event_page(struct kfd_process *p, uint64_t event_page_offset)
+{
+	struct kfd_node *kfd;
+	struct kfd_process_device *pdd;
+	void *mem, *kern_addr;
+	uint64_t size;
+	int err = 0;
+
+	if (p->signal_page) {
+		pr_err("Event page is already set\n");
+		return -EINVAL;
+	}
+
+	pdd = kfd_process_device_data_by_id(p, GET_GPU_ID(event_page_offset));
+	if (!pdd) {
+		pr_err("Getting device by id failed in %s\n", __func__);
+		return -EINVAL;
+	}
+	kfd = pdd->dev;
+
+	pdd = kfd_bind_process_to_device(kfd, p);
+	if (IS_ERR(pdd))
+		return PTR_ERR(pdd);
+
+	mem = kfd_process_device_translate_handle(pdd,
+			GET_IDR_HANDLE(event_page_offset));
+	if (!mem) {
+		pr_err("Can't find BO, offset is 0x%llx\n", event_page_offset);
+		return -EINVAL;
+	}
+
+	err = amdgpu_amdkfd_gpuvm_map_gtt_bo_to_kernel(mem, &kern_addr, &size);
+	if (err) {
+		pr_err("Failed to map event page to kernel\n");
+		return err;
+	}
+
+	err = kfd_event_page_set(p, kern_addr, size, event_page_offset);
+	if (err) {
+		pr_err("Failed to set event page\n");
+		amdgpu_amdkfd_gpuvm_unmap_gtt_bo_from_kernel(mem);
+		return err;
+	}
+	return err;
+}
+
 int kfd_event_create(struct file *devkfd, struct kfd_process *p,
 		     uint32_t event_type, bool auto_reset, uint32_t node_id,
 		     uint32_t *event_id, uint32_t *event_trigger_data,
@@ -334,6 +408,7 @@ int kfd_event_create(struct file *devkfd, struct kfd_process *p,
 	ev->auto_reset = auto_reset;
 	ev->signaled = false;
 
+	spin_lock_init(&ev->lock);
 	init_waitqueue_head(&ev->wq);
 
 	*event_page_offset = 0;
@@ -343,20 +418,21 @@ int kfd_event_create(struct file *devkfd, struct kfd_process *p,
 	switch (event_type) {
 	case KFD_EVENT_TYPE_SIGNAL:
 	case KFD_EVENT_TYPE_DEBUG:
-		ret = create_signal_event(devkfd, p, ev);
+		ret = create_signal_event(devkfd, p, ev, NULL);
 		if (!ret) {
 			*event_page_offset = KFD_MMAP_TYPE_EVENTS;
 			*event_slot_index = ev->event_id;
 		}
 		break;
 	default:
-		ret = create_other_event(p, ev);
+		ret = create_other_event(p, ev, NULL);
 		break;
 	}
 
 	if (!ret) {
 		*event_id = ev->event_id;
 		*event_trigger_data = ev->event_id;
+		ev->event_age = 1;
 	} else {
 		kfree(ev);
 	}
@@ -366,6 +442,166 @@ int kfd_event_create(struct file *devkfd, struct kfd_process *p,
 	return ret;
 }
 
+int kfd_criu_restore_event(struct file *devkfd,
+			   struct kfd_process *p,
+			   uint8_t __user *user_priv_ptr,
+			   uint64_t *priv_data_offset,
+			   uint64_t max_priv_data_size)
+{
+	struct kfd_criu_event_priv_data *ev_priv;
+	struct kfd_event *ev = NULL;
+	int ret = 0;
+
+	ev_priv = kmalloc(sizeof(*ev_priv), GFP_KERNEL);
+	if (!ev_priv)
+		return -ENOMEM;
+
+	ev = kzalloc(sizeof(*ev), GFP_KERNEL);
+	if (!ev) {
+		ret = -ENOMEM;
+		goto exit;
+	}
+
+	if (*priv_data_offset + sizeof(*ev_priv) > max_priv_data_size) {
+		ret = -EINVAL;
+		goto exit;
+	}
+
+	ret = copy_from_user(ev_priv, user_priv_ptr + *priv_data_offset, sizeof(*ev_priv));
+	if (ret) {
+		ret = -EFAULT;
+		goto exit;
+	}
+	*priv_data_offset += sizeof(*ev_priv);
+
+	if (ev_priv->user_handle) {
+		ret = kfd_kmap_event_page(p, ev_priv->user_handle);
+		if (ret)
+			goto exit;
+	}
+
+	ev->type = ev_priv->type;
+	ev->auto_reset = ev_priv->auto_reset;
+	ev->signaled = ev_priv->signaled;
+
+	spin_lock_init(&ev->lock);
+	init_waitqueue_head(&ev->wq);
+
+	mutex_lock(&p->event_mutex);
+	switch (ev->type) {
+	case KFD_EVENT_TYPE_SIGNAL:
+	case KFD_EVENT_TYPE_DEBUG:
+		ret = create_signal_event(devkfd, p, ev, &ev_priv->event_id);
+		break;
+	case KFD_EVENT_TYPE_MEMORY:
+		memcpy(&ev->memory_exception_data,
+			&ev_priv->memory_exception_data,
+			sizeof(struct kfd_hsa_memory_exception_data));
+
+		ret = create_other_event(p, ev, &ev_priv->event_id);
+		break;
+	case KFD_EVENT_TYPE_HW_EXCEPTION:
+		memcpy(&ev->hw_exception_data,
+			&ev_priv->hw_exception_data,
+			sizeof(struct kfd_hsa_hw_exception_data));
+
+		ret = create_other_event(p, ev, &ev_priv->event_id);
+		break;
+	}
+	mutex_unlock(&p->event_mutex);
+
+exit:
+	if (ret)
+		kfree(ev);
+
+	kfree(ev_priv);
+
+	return ret;
+}
+
+int kfd_criu_checkpoint_events(struct kfd_process *p,
+			 uint8_t __user *user_priv_data,
+			 uint64_t *priv_data_offset)
+{
+	struct kfd_criu_event_priv_data *ev_privs;
+	int i = 0;
+	int ret =  0;
+	struct kfd_event *ev;
+	uint32_t ev_id;
+
+	uint32_t num_events = kfd_get_num_events(p);
+
+	if (!num_events)
+		return 0;
+
+	ev_privs = kvzalloc(num_events * sizeof(*ev_privs), GFP_KERNEL);
+	if (!ev_privs)
+		return -ENOMEM;
+
+
+	idr_for_each_entry(&p->event_idr, ev, ev_id) {
+		struct kfd_criu_event_priv_data *ev_priv;
+
+		/*
+		 * Currently, all events have same size of private_data, but the current ioctl's
+		 * and CRIU plugin supports private_data of variable sizes
+		 */
+		ev_priv = &ev_privs[i];
+
+		ev_priv->object_type = KFD_CRIU_OBJECT_TYPE_EVENT;
+
+		/* We store the user_handle with the first event */
+		if (i == 0 && p->signal_page)
+			ev_priv->user_handle = p->signal_handle;
+
+		ev_priv->event_id = ev->event_id;
+		ev_priv->auto_reset = ev->auto_reset;
+		ev_priv->type = ev->type;
+		ev_priv->signaled = ev->signaled;
+
+		if (ev_priv->type == KFD_EVENT_TYPE_MEMORY)
+			memcpy(&ev_priv->memory_exception_data,
+				&ev->memory_exception_data,
+				sizeof(struct kfd_hsa_memory_exception_data));
+		else if (ev_priv->type == KFD_EVENT_TYPE_HW_EXCEPTION)
+			memcpy(&ev_priv->hw_exception_data,
+				&ev->hw_exception_data,
+				sizeof(struct kfd_hsa_hw_exception_data));
+
+		pr_debug("Checkpointed event[%d] id = 0x%08x auto_reset = %x type = %x signaled = %x\n",
+			  i,
+			  ev_priv->event_id,
+			  ev_priv->auto_reset,
+			  ev_priv->type,
+			  ev_priv->signaled);
+		i++;
+	}
+
+	ret = copy_to_user(user_priv_data + *priv_data_offset,
+			   ev_privs, num_events * sizeof(*ev_privs));
+	if (ret) {
+		pr_err("Failed to copy events priv to user\n");
+		ret = -EFAULT;
+	}
+
+	*priv_data_offset += num_events * sizeof(*ev_privs);
+
+	kvfree(ev_privs);
+	return ret;
+}
+
+int kfd_get_num_events(struct kfd_process *p)
+{
+	struct kfd_event *ev;
+	uint32_t id;
+	u32 num_events = 0;
+
+	idr_for_each_entry(&p->event_idr, ev, id)
+		num_events++;
+
+	return num_events;
+}
+
 /* Assumes that p is current. */
 int kfd_event_destroy(struct kfd_process *p, uint32_t event_id)
 {
@@ -391,13 +627,18 @@ static void set_event(struct kfd_event *ev)
 
 	/* Auto reset if the list is non-empty and we're waking
 	 * someone. waitqueue_active is safe here because we're
-	 * protected by the p->event_mutex, which is also held when
+	 * protected by the ev->lock, which is also held when
 	 * updating the wait queues in kfd_wait_on_events.
 	 */
 	ev->signaled = !ev->auto_reset || !waitqueue_active(&ev->wq);
+	if (!(++ev->event_age)) {
+		/* Never wrap back to reserved/default event age 0/1 */
+		ev->event_age = 2;
+		WARN_ONCE(1, "event_age wrap back!");
+	}
 
 	list_for_each_entry(waiter, &ev->wq.head, wait.entry)
-		waiter->activated = true;
+		WRITE_ONCE(waiter->activated, true);
 
 	wake_up_all(&ev->wq);
 }
@@ -408,16 +649,23 @@ int kfd_set_event(struct kfd_process *p, uint32_t event_id)
 	int ret = 0;
 	struct kfd_event *ev;
 
-	mutex_lock(&p->event_mutex);
+	rcu_read_lock();
 
 	ev = lookup_event_by_id(p, event_id);
+	if (!ev) {
+		ret = -EINVAL;
+		goto unlock_rcu;
+	}
+	spin_lock(&ev->lock);
 
-	if (ev && event_can_be_cpu_signaled(ev))
+	if (event_can_be_cpu_signaled(ev))
 		set_event(ev);
 	else
 		ret = -EINVAL;
 
-	mutex_unlock(&p->event_mutex);
+	spin_unlock(&ev->lock);
+unlock_rcu:
+	rcu_read_unlock();
 	return ret;
 }
 
@@ -432,23 +680,30 @@ int kfd_reset_event(struct kfd_process *p, uint32_t event_id)
 	int ret = 0;
 	struct kfd_event *ev;
 
-	mutex_lock(&p->event_mutex);
+	rcu_read_lock();
 
 	ev = lookup_event_by_id(p, event_id);
+	if (!ev) {
+		ret = -EINVAL;
+		goto unlock_rcu;
+	}
+	spin_lock(&ev->lock);
 
-	if (ev && event_can_be_cpu_signaled(ev))
+	if (event_can_be_cpu_signaled(ev))
 		reset_event(ev);
 	else
 		ret = -EINVAL;
 
-	mutex_unlock(&p->event_mutex);
+	spin_unlock(&ev->lock);
+unlock_rcu:
+	rcu_read_unlock();
 	return ret;
 
 }
 
 static void acknowledge_signal(struct kfd_process *p, struct kfd_event *ev)
 {
-	page_slots(p->signal_page)[ev->event_id] = UNSIGNALED_EVENT_SLOT;
+	WRITE_ONCE(page_slots(p->signal_page)[ev->event_id], UNSIGNALED_EVENT_SLOT);
 }
 
 static void set_event_from_interrupt(struct kfd_process *p,
@@ -456,7 +711,9 @@ static void set_event_from_interrupt(struct kfd_process *p,
 {
 	if (ev && event_can_be_gpu_signaled(ev)) {
 		acknowledge_signal(p, ev);
+		spin_lock(&ev->lock);
 		set_event(ev);
+		spin_unlock(&ev->lock);
 	}
 }
 
@@ -470,12 +727,12 @@ void kfd_signal_event_interrupt(u32 pasid, uint32_t partial_id,
 	 * to process context, kfd_process could attempt to exit while we are
 	 * running so the lookup function increments the process ref count.
 	 */
-	struct kfd_process *p = kfd_lookup_process_by_pasid(pasid);
+	struct kfd_process *p = kfd_lookup_process_by_pasid(pasid, NULL);
 
 	if (!p)
 		return; /* Presumably process exited. */
 
-	mutex_lock(&p->event_mutex);
+	rcu_read_lock();
 
 	if (valid_id_bits)
 		ev = lookup_signaled_event_by_partial_id(p, partial_id,
@@ -503,7 +760,7 @@ void kfd_signal_event_interrupt(u32 pasid, uint32_t partial_id,
 				if (id >= KFD_SIGNAL_EVENT_LIMIT)
 					break;
 
-				if (slots[id] != UNSIGNALED_EVENT_SLOT)
+				if (READ_ONCE(slots[id]) != UNSIGNALED_EVENT_SLOT)
 					set_event_from_interrupt(p, ev);
 			}
 		} else {
@@ -511,15 +768,15 @@ void kfd_signal_event_interrupt(u32 pasid, uint32_t partial_id,
 			 * iterate over the signal slots and lookup
 			 * only signaled events from the IDR.
 			 */
-			for (id = 0; id < KFD_SIGNAL_EVENT_LIMIT; id++)
-				if (slots[id] != UNSIGNALED_EVENT_SLOT) {
+			for (id = 1; id < KFD_SIGNAL_EVENT_LIMIT; id++)
+				if (READ_ONCE(slots[id]) != UNSIGNALED_EVENT_SLOT) {
 					ev = lookup_event_by_id(p, id);
 					set_event_from_interrupt(p, ev);
 				}
 		}
 	}
 
-	mutex_unlock(&p->event_mutex);
+	rcu_read_unlock();
 	kfd_unref_process(p);
 }
 
@@ -528,43 +785,44 @@ static struct kfd_event_waiter *alloc_event_waiters(uint32_t num_events)
 	struct kfd_event_waiter *event_waiters;
 	uint32_t i;
 
-	event_waiters = kmalloc_array(num_events,
-					sizeof(struct kfd_event_waiter),
-					GFP_KERNEL);
+	event_waiters = kcalloc(num_events, sizeof(struct kfd_event_waiter),
+				GFP_KERNEL);
+	if (!event_waiters)
+		return NULL;
 
-	for (i = 0; (event_waiters) && (i < num_events) ; i++) {
+	for (i = 0; i < num_events; i++)
 		init_wait(&event_waiters[i].wait);
-		event_waiters[i].activated = false;
-	}
 
 	return event_waiters;
 }
 
-static int init_event_waiter_get_status(struct kfd_process *p,
+static int init_event_waiter(struct kfd_process *p,
 		struct kfd_event_waiter *waiter,
-		uint32_t event_id)
+		struct kfd_event_data *event_data)
 {
-	struct kfd_event *ev = lookup_event_by_id(p, event_id);
+	struct kfd_event *ev = lookup_event_by_id(p, event_data->event_id);
 
 	if (!ev)
 		return -EINVAL;
 
+	spin_lock(&ev->lock);
 	waiter->event = ev;
 	waiter->activated = ev->signaled;
 	ev->signaled = ev->signaled && !ev->auto_reset;
 
-	return 0;
-}
-
-static void init_event_waiter_add_to_waitlist(struct kfd_event_waiter *waiter)
-{
-	struct kfd_event *ev = waiter->event;
+	/* last_event_age = 0 reserved for backward compatible */
+	if (waiter->event->type == KFD_EVENT_TYPE_SIGNAL &&
+		event_data->signal_event_data.last_event_age) {
+		waiter->event_age_enabled = true;
+		if (ev->event_age != event_data->signal_event_data.last_event_age)
+			waiter->activated = true;
+	}
 
-	/* Only add to the wait list if we actually need to
-	 * wait on this event.
-	 */
 	if (!waiter->activated)
 		add_wait_queue(&ev->wq, &waiter->wait);
+	spin_unlock(&ev->lock);
+
+	return 0;
 }
 
 /* test_event_condition - Test condition of events being waited for
@@ -584,10 +842,10 @@ static uint32_t test_event_condition(bool all, uint32_t num_events,
 	uint32_t activated_count = 0;
 
 	for (i = 0; i < num_events; i++) {
-		if (!event_waiters[i].event)
+		if (!READ_ONCE(event_waiters[i].event))
 			return KFD_IOC_WAIT_RESULT_FAIL;
 
-		if (event_waiters[i].activated) {
+		if (READ_ONCE(event_waiters[i].activated)) {
 			if (!all)
 				return KFD_IOC_WAIT_RESULT_COMPLETE;
 
@@ -607,30 +865,40 @@ static int copy_signaled_event_data(uint32_t num_events,
 		struct kfd_event_waiter *event_waiters,
 		struct kfd_event_data __user *data)
 {
-	struct kfd_hsa_memory_exception_data *src;
-	struct kfd_hsa_memory_exception_data __user *dst;
+	void *src;
+	void __user *dst;
 	struct kfd_event_waiter *waiter;
 	struct kfd_event *event;
-	uint32_t i;
+	uint32_t i, size = 0;
 
 	for (i = 0; i < num_events; i++) {
 		waiter = &event_waiters[i];
 		event = waiter->event;
-		if (waiter->activated && event->type == KFD_EVENT_TYPE_MEMORY) {
-			dst = &data[i].memory_exception_data;
-			src = &event->memory_exception_data;
-			if (copy_to_user(dst, src,
-				sizeof(struct kfd_hsa_memory_exception_data)))
+		if (!event)
+			return -EINVAL; /* event was destroyed */
+		if (waiter->activated) {
+			if (event->type == KFD_EVENT_TYPE_MEMORY) {
+				dst = &data[i].memory_exception_data;
+				src = &event->memory_exception_data;
+				size = sizeof(struct kfd_hsa_memory_exception_data);
+			} else if (event->type == KFD_EVENT_TYPE_HW_EXCEPTION) {
+				dst = &data[i].memory_exception_data;
+				src = &event->hw_exception_data;
+				size = sizeof(struct kfd_hsa_hw_exception_data);
+			} else if (event->type == KFD_EVENT_TYPE_SIGNAL &&
+				waiter->event_age_enabled) {
+				dst = &data[i].signal_event_data.last_event_age;
+				src = &event->event_age;
+				size = sizeof(u64);
+			}
+			if (size && copy_to_user(dst, src, size))
 				return -EFAULT;
 		}
 	}
 
 	return 0;
-
 }
 
-
-
 static long user_timeout_to_jiffies(uint32_t user_timeout_ms)
 {
 	if (user_timeout_ms == KFD_EVENT_TIMEOUT_IMMEDIATE)
@@ -649,21 +917,28 @@ static long user_timeout_to_jiffies(uint32_t user_timeout_ms)
 	return msecs_to_jiffies(user_timeout_ms) + 1;
 }
 
-static void free_waiters(uint32_t num_events, struct kfd_event_waiter *waiters)
+static void free_waiters(uint32_t num_events, struct kfd_event_waiter *waiters,
+			 bool undo_auto_reset)
 {
 	uint32_t i;
 
 	for (i = 0; i < num_events; i++)
-		if (waiters[i].event)
+		if (waiters[i].event) {
+			spin_lock(&waiters[i].event->lock);
 			remove_wait_queue(&waiters[i].event->wq,
 					  &waiters[i].wait);
+			if (undo_auto_reset && waiters[i].activated &&
+			    waiters[i].event && waiters[i].event->auto_reset)
+				set_event(waiters[i].event);
+			spin_unlock(&waiters[i].event->lock);
+		}
 
 	kfree(waiters);
 }
 
 int kfd_wait_on_events(struct kfd_process *p,
 		       uint32_t num_events, void __user *data,
-		       bool all, uint32_t user_timeout_ms,
+		       bool all, uint32_t *user_timeout_ms,
 		       uint32_t *wait_result)
 {
 	struct kfd_event_data __user *events =
@@ -672,7 +947,7 @@ int kfd_wait_on_events(struct kfd_process *p,
 	int ret = 0;
 
 	struct kfd_event_waiter *event_waiters = NULL;
-	long timeout = user_timeout_to_jiffies(user_timeout_ms);
+	long timeout = user_timeout_to_jiffies(*user_timeout_ms);
 
 	event_waiters = alloc_event_waiters(num_events);
 	if (!event_waiters) {
@@ -680,6 +955,9 @@ int kfd_wait_on_events(struct kfd_process *p,
 		goto out;
 	}
 
+	/* Use p->event_mutex here to protect against concurrent creation and
+	 * destruction of events while we initialize event_waiters.
+	 */
 	mutex_lock(&p->event_mutex);
 
 	for (i = 0; i < num_events; i++) {
@@ -691,8 +969,7 @@ int kfd_wait_on_events(struct kfd_process *p,
 			goto out_unlock;
 		}
 
-		ret = init_event_waiter_get_status(p, &event_waiters[i],
-				event_data.event_id);
+		ret = init_event_waiter(p, &event_waiters[i], &event_data);
 		if (ret)
 			goto out_unlock;
 	}
@@ -710,10 +987,6 @@ int kfd_wait_on_events(struct kfd_process *p,
 		goto out_unlock;
 	}
 
-	/* Add to wait lists if we need to wait. */
-	for (i = 0; i < num_events; i++)
-		init_event_waiter_add_to_waitlist(&event_waiters[i]);
-
 	mutex_unlock(&p->event_mutex);
 
 	while (true) {
@@ -723,15 +996,11 @@ int kfd_wait_on_events(struct kfd_process *p,
 		}
 
 		if (signal_pending(current)) {
-			/*
-			 * This is wrong when a nonzero, non-infinite timeout
-			 * is specified. We need to use
-			 * ERESTARTSYS_RESTARTBLOCK, but struct restart_block
-			 * contains a union with data for each user and it's
-			 * in generic kernel code that I don't want to
-			 * touch yet.
-			 */
 			ret = -ERESTARTSYS;
+			if (*user_timeout_ms != KFD_EVENT_TIMEOUT_IMMEDIATE &&
+			    *user_timeout_ms != KFD_EVENT_TIMEOUT_INFINITE)
+				*user_timeout_ms = jiffies_to_msecs(
+					max(0l, timeout-1));
 			break;
 		}
 
@@ -758,16 +1027,21 @@ int kfd_wait_on_events(struct kfd_process *p,
 	}
 	__set_current_state(TASK_RUNNING);
 
+	mutex_lock(&p->event_mutex);
 	/* copy_signaled_event_data may sleep. So this has to happen
 	 * after the task state is set back to RUNNING.
+	 *
+	 * The event may also have been destroyed after signaling. So
+	 * copy_signaled_event_data also must confirm that the event
+	 * still exists. Therefore this must be under the p->event_mutex
+	 * which is also held when events are destroyed.
 	 */
 	if (!ret && *wait_result == KFD_IOC_WAIT_RESULT_COMPLETE)
 		ret = copy_signaled_event_data(num_events,
 					       event_waiters, events);
 
-	mutex_lock(&p->event_mutex);
 out_unlock:
-	free_waiters(num_events, event_waiters);
+	free_waiters(num_events, event_waiters, ret == -ERESTARTSYS);
 	mutex_unlock(&p->event_mutex);
 out:
 	if (ret)
@@ -801,8 +1075,8 @@ int kfd_event_mmap(struct kfd_process *p, struct vm_area_struct *vma)
 	pfn = __pa(page->kernel_address);
 	pfn >>= PAGE_SHIFT;
 
-	vma->vm_flags |= VM_IO | VM_DONTCOPY | VM_DONTEXPAND | VM_NORESERVE
-		       | VM_DONTDUMP | VM_PFNMAP;
+	vm_flags_set(vma, VM_IO | VM_DONTCOPY | VM_DONTEXPAND | VM_NORESERVE
+		       | VM_DONTDUMP | VM_PFNMAP);
 
 	pr_debug("Mapping signal page\n");
 	pr_debug("     start user address  == 0x%08lx\n", vma->vm_start);
@@ -824,8 +1098,7 @@ int kfd_event_mmap(struct kfd_process *p, struct vm_area_struct *vma)
 }
 
 /*
- * Assumes that p->event_mutex is held and of course
- * that p is not going away (current or locked).
+ * Assumes that p is not going away.
  */
 static void lookup_events_by_type_and_signal(struct kfd_process *p,
 		int type, void *event_data)
@@ -837,6 +1110,8 @@ static void lookup_events_by_type_and_signal(struct kfd_process *p,
 
 	ev_data = (struct kfd_hsa_memory_exception_data *) event_data;
 
+	rcu_read_lock();
+
 	id = KFD_FIRST_NONSIGNAL_EVENT_ID;
 	idr_for_each_entry_continue(&p->event_idr, ev, id)
 		if (ev->type == type) {
@@ -844,15 +1119,17 @@ static void lookup_events_by_type_and_signal(struct kfd_process *p,
 			dev_dbg(kfd_device,
 					"Event found: id %X type %d",
 					ev->event_id, ev->type);
+			spin_lock(&ev->lock);
 			set_event(ev);
 			if (ev->type == KFD_EVENT_TYPE_MEMORY && ev_data)
 				ev->memory_exception_data = *ev_data;
+			spin_unlock(&ev->lock);
 		}
 
 	if (type == KFD_EVENT_TYPE_MEMORY) {
 		dev_warn(kfd_device,
-			"Sending SIGSEGV to process %d (pasid 0x%x)",
-				p->lead_thread->pid, p->pasid);
+			"Sending SIGSEGV to process pid %d",
+				p->lead_thread->pid);
 		send_sig(SIGSEGV, p->lead_thread, 0);
 	}
 
@@ -860,95 +1137,18 @@ static void lookup_events_by_type_and_signal(struct kfd_process *p,
 	if (send_signal) {
 		if (send_sigterm) {
 			dev_warn(kfd_device,
-				"Sending SIGTERM to process %d (pasid 0x%x)",
-					p->lead_thread->pid, p->pasid);
+				"Sending SIGTERM to process pid %d",
+					p->lead_thread->pid);
 			send_sig(SIGTERM, p->lead_thread, 0);
 		} else {
 			dev_err(kfd_device,
-				"Process %d (pasid 0x%x) got unhandled exception",
-				p->lead_thread->pid, p->pasid);
+				"Process pid %d got unhandled exception",
+				p->lead_thread->pid);
 		}
 	}
-}
-
-#ifdef KFD_SUPPORT_IOMMU_V2
-void kfd_signal_iommu_event(struct kfd_dev *dev, u32 pasid,
-		unsigned long address, bool is_write_requested,
-		bool is_execute_requested)
-{
-	struct kfd_hsa_memory_exception_data memory_exception_data;
-	struct vm_area_struct *vma;
-
-	/*
-	 * Because we are called from arbitrary context (workqueue) as opposed
-	 * to process context, kfd_process could attempt to exit while we are
-	 * running so the lookup function increments the process ref count.
-	 */
-	struct kfd_process *p = kfd_lookup_process_by_pasid(pasid);
-	struct mm_struct *mm;
-
-	if (!p)
-		return; /* Presumably process exited. */
-
-	/* Take a safe reference to the mm_struct, which may otherwise
-	 * disappear even while the kfd_process is still referenced.
-	 */
-	mm = get_task_mm(p->lead_thread);
-	if (!mm) {
-		kfd_unref_process(p);
-		return; /* Process is exiting */
-	}
-
-	memset(&memory_exception_data, 0, sizeof(memory_exception_data));
-
-	mmap_read_lock(mm);
-	vma = find_vma(mm, address);
-
-	memory_exception_data.gpu_id = dev->id;
-	memory_exception_data.va = address;
-	/* Set failure reason */
-	memory_exception_data.failure.NotPresent = 1;
-	memory_exception_data.failure.NoExecute = 0;
-	memory_exception_data.failure.ReadOnly = 0;
-	if (vma && address >= vma->vm_start) {
-		memory_exception_data.failure.NotPresent = 0;
-
-		if (is_write_requested && !(vma->vm_flags & VM_WRITE))
-			memory_exception_data.failure.ReadOnly = 1;
-		else
-			memory_exception_data.failure.ReadOnly = 0;
-
-		if (is_execute_requested && !(vma->vm_flags & VM_EXEC))
-			memory_exception_data.failure.NoExecute = 1;
-		else
-			memory_exception_data.failure.NoExecute = 0;
-	}
-
-	mmap_read_unlock(mm);
-	mmput(mm);
 
-	pr_debug("notpresent %d, noexecute %d, readonly %d\n",
-			memory_exception_data.failure.NotPresent,
-			memory_exception_data.failure.NoExecute,
-			memory_exception_data.failure.ReadOnly);
-
-	/* Workaround on Raven to not kill the process when memory is freed
-	 * before IOMMU is able to finish processing all the excessive PPRs
-	 */
-	if (dev->device_info->asic_family != CHIP_RAVEN &&
-	    dev->device_info->asic_family != CHIP_RENOIR) {
-		mutex_lock(&p->event_mutex);
-
-		/* Lookup events by type and signal them */
-		lookup_events_by_type_and_signal(p, KFD_EVENT_TYPE_MEMORY,
-				&memory_exception_data);
-
-		mutex_unlock(&p->event_mutex);
-	}
-
-	kfd_unref_process(p);
+	rcu_read_unlock();
 }
-#endif /* KFD_SUPPORT_IOMMU_V2 */
 
 void kfd_signal_hw_exception_event(u32 pasid)
 {
@@ -957,58 +1157,87 @@ void kfd_signal_hw_exception_event(u32 pasid)
 	 * to process context, kfd_process could attempt to exit while we are
 	 * running so the lookup function increments the process ref count.
 	 */
-	struct kfd_process *p = kfd_lookup_process_by_pasid(pasid);
+	struct kfd_process *p = kfd_lookup_process_by_pasid(pasid, NULL);
 
 	if (!p)
 		return; /* Presumably process exited. */
 
-	mutex_lock(&p->event_mutex);
-
-	/* Lookup events by type and signal them */
 	lookup_events_by_type_and_signal(p, KFD_EVENT_TYPE_HW_EXCEPTION, NULL);
-
-	mutex_unlock(&p->event_mutex);
 	kfd_unref_process(p);
 }
 
-void kfd_signal_vm_fault_event(struct kfd_dev *dev, u32 pasid,
-				struct kfd_vm_fault_info *info)
+void kfd_signal_vm_fault_event_with_userptr(struct kfd_process *p, uint64_t gpu_va)
+{
+	struct kfd_process_device *pdd;
+	struct kfd_hsa_memory_exception_data exception_data;
+	int i;
+
+	memset(&exception_data, 0, sizeof(exception_data));
+	exception_data.va = gpu_va;
+	exception_data.failure.NotPresent = 1;
+
+	// Send VM seg fault to all kfd process device
+	for (i = 0; i < p->n_pdds; i++) {
+		pdd = p->pdds[i];
+		exception_data.gpu_id = pdd->user_gpu_id;
+		kfd_evict_process_device(pdd);
+		kfd_signal_vm_fault_event(pdd, NULL, &exception_data);
+	}
+}
+
+void kfd_signal_vm_fault_event(struct kfd_process_device *pdd,
+				struct kfd_vm_fault_info *info,
+				struct kfd_hsa_memory_exception_data *data)
 {
 	struct kfd_event *ev;
 	uint32_t id;
-	struct kfd_process *p = kfd_lookup_process_by_pasid(pasid);
+	struct kfd_process *p = pdd->process;
 	struct kfd_hsa_memory_exception_data memory_exception_data;
+	int user_gpu_id;
 
-	if (!p)
-		return; /* Presumably process exited. */
-	memset(&memory_exception_data, 0, sizeof(memory_exception_data));
-	memory_exception_data.gpu_id = dev->id;
-	memory_exception_data.failure.imprecise = true;
-	/* Set failure reason */
-	if (info) {
-		memory_exception_data.va = (info->page_addr) << PAGE_SHIFT;
-		memory_exception_data.failure.NotPresent =
-			info->prot_valid ? 1 : 0;
-		memory_exception_data.failure.NoExecute =
-			info->prot_exec ? 1 : 0;
-		memory_exception_data.failure.ReadOnly =
-			info->prot_write ? 1 : 0;
-		memory_exception_data.failure.imprecise = 0;
+	user_gpu_id = kfd_process_get_user_gpu_id(p, pdd->dev->id);
+	if (unlikely(user_gpu_id == -EINVAL)) {
+		WARN_ONCE(1, "Could not get user_gpu_id from dev->id:%x\n",
+			  pdd->dev->id);
+		return;
 	}
-	mutex_lock(&p->event_mutex);
+
+	/* SoC15 chips and onwards will pass in data from now on. */
+	if (!data) {
+		memset(&memory_exception_data, 0, sizeof(memory_exception_data));
+		memory_exception_data.gpu_id = user_gpu_id;
+		memory_exception_data.failure.imprecise = true;
+
+		/* Set failure reason */
+		if (info) {
+			memory_exception_data.va = (info->page_addr) <<
+								PAGE_SHIFT;
+			memory_exception_data.failure.NotPresent =
+				info->prot_valid ? 1 : 0;
+			memory_exception_data.failure.NoExecute =
+				info->prot_exec ? 1 : 0;
+			memory_exception_data.failure.ReadOnly =
+				info->prot_write ? 1 : 0;
+			memory_exception_data.failure.imprecise = 0;
+		}
+	}
+
+	rcu_read_lock();
 
 	id = KFD_FIRST_NONSIGNAL_EVENT_ID;
 	idr_for_each_entry_continue(&p->event_idr, ev, id)
 		if (ev->type == KFD_EVENT_TYPE_MEMORY) {
-			ev->memory_exception_data = memory_exception_data;
+			spin_lock(&ev->lock);
+			ev->memory_exception_data = data ? *data :
+							memory_exception_data;
 			set_event(ev);
+			spin_unlock(&ev->lock);
 		}
 
-	mutex_unlock(&p->event_mutex);
-	kfd_unref_process(p);
+	rcu_read_unlock();
 }
 
-void kfd_signal_reset_event(struct kfd_dev *dev)
+void kfd_signal_reset_event(struct kfd_node *dev)
 {
 	struct kfd_hsa_hw_exception_data hw_exception_data;
 	struct kfd_hsa_memory_exception_data memory_exception_data;
@@ -1022,69 +1251,129 @@ void kfd_signal_reset_event(struct kfd_dev *dev)
 
 	/* Whole gpu reset caused by GPU hang and memory is lost */
 	memset(&hw_exception_data, 0, sizeof(hw_exception_data));
-	hw_exception_data.gpu_id = dev->id;
 	hw_exception_data.memory_lost = 1;
 	hw_exception_data.reset_cause = reset_cause;
 
 	memset(&memory_exception_data, 0, sizeof(memory_exception_data));
 	memory_exception_data.ErrorType = KFD_MEM_ERR_SRAM_ECC;
-	memory_exception_data.gpu_id = dev->id;
 	memory_exception_data.failure.imprecise = true;
 
 	idx = srcu_read_lock(&kfd_processes_srcu);
 	hash_for_each_rcu(kfd_processes_table, temp, p, kfd_processes) {
-		mutex_lock(&p->event_mutex);
+		int user_gpu_id = kfd_process_get_user_gpu_id(p, dev->id);
+		struct kfd_process_device *pdd = kfd_get_process_device_data(dev, p);
+
+		if (unlikely(user_gpu_id == -EINVAL)) {
+			WARN_ONCE(1, "Could not get user_gpu_id from dev->id:%x\n", dev->id);
+			continue;
+		}
+
+		if (unlikely(!pdd)) {
+			WARN_ONCE(1, "Could not get device data from process pid:%d\n",
+				  p->lead_thread->pid);
+			continue;
+		}
+
+		if (dev->dqm->detect_hang_count && !pdd->has_reset_queue)
+			continue;
+
+		if (dev->dqm->detect_hang_count) {
+			struct amdgpu_task_info *ti;
+			struct amdgpu_fpriv *drv_priv;
+
+			if (unlikely(amdgpu_file_to_fpriv(pdd->drm_file, &drv_priv))) {
+				WARN_ONCE(1, "Could not get vm for device %x from pid:%d\n",
+					  dev->id, p->lead_thread->pid);
+				continue;
+			}
+
+			ti = amdgpu_vm_get_task_info_vm(&drv_priv->vm);
+			if (ti) {
+				dev_err(dev->adev->dev,
+					"Queues reset on process %s tid %d thread %s pid %d\n",
+					ti->process_name, ti->tgid, ti->task.comm, ti->task.pid);
+				amdgpu_vm_put_task_info(ti);
+			}
+		}
+
+		rcu_read_lock();
+
 		id = KFD_FIRST_NONSIGNAL_EVENT_ID;
 		idr_for_each_entry_continue(&p->event_idr, ev, id) {
 			if (ev->type == KFD_EVENT_TYPE_HW_EXCEPTION) {
+				spin_lock(&ev->lock);
 				ev->hw_exception_data = hw_exception_data;
+				ev->hw_exception_data.gpu_id = user_gpu_id;
 				set_event(ev);
+				spin_unlock(&ev->lock);
 			}
 			if (ev->type == KFD_EVENT_TYPE_MEMORY &&
 			    reset_cause == KFD_HW_EXCEPTION_ECC) {
+				spin_lock(&ev->lock);
 				ev->memory_exception_data = memory_exception_data;
+				ev->memory_exception_data.gpu_id = user_gpu_id;
 				set_event(ev);
+				spin_unlock(&ev->lock);
 			}
 		}
-		mutex_unlock(&p->event_mutex);
+
+		rcu_read_unlock();
 	}
 	srcu_read_unlock(&kfd_processes_srcu, idx);
 }
 
-void kfd_signal_poison_consumed_event(struct kfd_dev *dev, u32 pasid)
+void kfd_signal_poison_consumed_event(struct kfd_node *dev, u32 pasid)
 {
-	struct kfd_process *p = kfd_lookup_process_by_pasid(pasid);
+	struct kfd_process *p = kfd_lookup_process_by_pasid(pasid, NULL);
 	struct kfd_hsa_memory_exception_data memory_exception_data;
 	struct kfd_hsa_hw_exception_data hw_exception_data;
 	struct kfd_event *ev;
 	uint32_t id = KFD_FIRST_NONSIGNAL_EVENT_ID;
+	int user_gpu_id;
 
-	if (!p)
+	if (!p) {
+		dev_warn(dev->adev->dev, "Not find process with pasid:%d\n", pasid);
 		return; /* Presumably process exited. */
+	}
+
+	user_gpu_id = kfd_process_get_user_gpu_id(p, dev->id);
+	if (unlikely(user_gpu_id == -EINVAL)) {
+		WARN_ONCE(1, "Could not get user_gpu_id from dev->id:%x\n", dev->id);
+		kfd_unref_process(p);
+		return;
+	}
 
 	memset(&hw_exception_data, 0, sizeof(hw_exception_data));
-	hw_exception_data.gpu_id = dev->id;
+	hw_exception_data.gpu_id = user_gpu_id;
 	hw_exception_data.memory_lost = 1;
 	hw_exception_data.reset_cause = KFD_HW_EXCEPTION_ECC;
 
 	memset(&memory_exception_data, 0, sizeof(memory_exception_data));
 	memory_exception_data.ErrorType = KFD_MEM_ERR_POISON_CONSUMED;
-	memory_exception_data.gpu_id = dev->id;
+	memory_exception_data.gpu_id = user_gpu_id;
 	memory_exception_data.failure.imprecise = true;
 
-	mutex_lock(&p->event_mutex);
+	rcu_read_lock();
+
 	idr_for_each_entry_continue(&p->event_idr, ev, id) {
 		if (ev->type == KFD_EVENT_TYPE_HW_EXCEPTION) {
+			spin_lock(&ev->lock);
 			ev->hw_exception_data = hw_exception_data;
 			set_event(ev);
+			spin_unlock(&ev->lock);
 		}
 
 		if (ev->type == KFD_EVENT_TYPE_MEMORY) {
+			spin_lock(&ev->lock);
 			ev->memory_exception_data = memory_exception_data;
 			set_event(ev);
+			spin_unlock(&ev->lock);
 		}
 	}
-	mutex_unlock(&p->event_mutex);
+
+	dev_warn(dev->adev->dev, "Send SIGBUS to process %s(pasid:%d)\n",
+		p->lead_thread->comm, pasid);
+	rcu_read_unlock();
 
 	/* user application will handle SIGBUS signal */
 	send_sig(SIGBUS, p->lead_thread, 0);
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_events.h b/drivers/gpu/drm/amd/amdkfd/kfd_events.h
index c8fe5dbdad55..52ccfd397c2b 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_events.h
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_events.h
@@ -1,5 +1,6 @@
+/* SPDX-License-Identifier: GPL-2.0 OR MIT */
 /*
- * Copyright 2014 Advanced Micro Devices, Inc.
+ * Copyright 2014-2022 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -52,12 +53,14 @@ struct signal_page;
 
 struct kfd_event {
 	u32 event_id;
+	u64 event_age;
 
 	bool signaled;
 	bool auto_reset;
 
 	int type;
 
+	spinlock_t lock;
 	wait_queue_head_t wq; /* List of event waiters. */
 
 	/* Only for signal events. */
@@ -68,6 +71,8 @@ struct kfd_event {
 		struct kfd_hsa_memory_exception_data memory_exception_data;
 		struct kfd_hsa_hw_exception_data hw_exception_data;
 	};
+
+	struct rcu_head rcu; /* for asynchronous kfree_rcu */
 };
 
 #define KFD_EVENT_TIMEOUT_IMMEDIATE 0
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c b/drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c
index d1388896f9c1..1d170dc50df3 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c
@@ -1,5 +1,6 @@
+// SPDX-License-Identifier: GPL-2.0 OR MIT
 /*
- * Copyright 2014 Advanced Micro Devices, Inc.
+ * Copyright 2014-2022 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -22,7 +23,6 @@
  */
 
 #include <linux/device.h>
-#include <linux/export.h>
 #include <linux/err.h>
 #include <linux/fs.h>
 #include <linux/sched.h>
@@ -34,7 +34,8 @@
 #include "kfd_priv.h"
 #include <linux/mm.h>
 #include <linux/mman.h>
-#include <asm/processor.h>
+#include <linux/processor.h>
+#include "amdgpu_vm.h"
 
 /*
  * The primary memory I/O features being added for revisions of gfxip
@@ -321,22 +322,19 @@ static void kfd_init_apertures_vi(struct kfd_process_device *pdd, uint8_t id)
 	pdd->lds_base = MAKE_LDS_APP_BASE_VI();
 	pdd->lds_limit = MAKE_LDS_APP_LIMIT(pdd->lds_base);
 
-	if (!pdd->dev->use_iommu_v2) {
-		/* dGPUs: SVM aperture starting at 0
-		 * with small reserved space for kernel.
-		 * Set them to CANONICAL addresses.
-		 */
-		pdd->gpuvm_base = SVM_USER_BASE;
-		pdd->gpuvm_limit =
-			pdd->dev->shared_resources.gpuvm_size - 1;
-	} else {
-		/* set them to non CANONICAL addresses, and no SVM is
-		 * allocated.
-		 */
-		pdd->gpuvm_base = MAKE_GPUVM_APP_BASE_VI(id + 1);
-		pdd->gpuvm_limit = MAKE_GPUVM_APP_LIMIT(pdd->gpuvm_base,
-				pdd->dev->shared_resources.gpuvm_size);
-	}
+	/* dGPUs: SVM aperture starting at 0
+	 * with small reserved space for kernel.
+	 * Set them to CANONICAL addresses.
+	 */
+	pdd->gpuvm_base = max(SVM_USER_BASE, AMDGPU_VA_RESERVED_BOTTOM);
+	pdd->gpuvm_limit =
+		pdd->dev->kfd->shared_resources.gpuvm_size - 1;
+
+	/* dGPUs: the reserved space for kernel
+	 * before SVM
+	 */
+	pdd->qpd.cwsr_base = SVM_CWSR_BASE;
+	pdd->qpd.ib_base = SVM_IB_BASE;
 
 	pdd->scratch_base = MAKE_SCRATCH_APP_BASE_VI();
 	pdd->scratch_limit = MAKE_SCRATCH_APP_LIMIT(pdd->scratch_base);
@@ -347,24 +345,24 @@ static void kfd_init_apertures_v9(struct kfd_process_device *pdd, uint8_t id)
 	pdd->lds_base = MAKE_LDS_APP_BASE_V9();
 	pdd->lds_limit = MAKE_LDS_APP_LIMIT(pdd->lds_base);
 
-	/* Raven needs SVM to support graphic handle, etc. Leave the small
-	 * reserved space before SVM on Raven as well, even though we don't
-	 * have to.
-	 * Set gpuvm_base and gpuvm_limit to CANONICAL addresses so that they
-	 * are used in Thunk to reserve SVM.
-	 */
-	pdd->gpuvm_base = SVM_USER_BASE;
+	pdd->gpuvm_base = AMDGPU_VA_RESERVED_BOTTOM;
 	pdd->gpuvm_limit =
-		pdd->dev->shared_resources.gpuvm_size - 1;
+		pdd->dev->kfd->shared_resources.gpuvm_size - 1;
 
 	pdd->scratch_base = MAKE_SCRATCH_APP_BASE_V9();
 	pdd->scratch_limit = MAKE_SCRATCH_APP_LIMIT(pdd->scratch_base);
+
+	/*
+	 * Place TBA/TMA on opposite side of VM hole to prevent
+	 * stray faults from triggering SVM on these pages.
+	 */
+	pdd->qpd.cwsr_base = AMDGPU_VA_RESERVED_TRAP_START(pdd->dev->adev);
 }
 
 int kfd_init_apertures(struct kfd_process *process)
 {
 	uint8_t id  = 0;
-	struct kfd_dev *dev;
+	struct kfd_node *dev;
 	struct kfd_process_device *pdd;
 
 	/*Iterating over all devices*/
@@ -381,7 +379,8 @@ int kfd_init_apertures(struct kfd_process *process)
 
 		pdd = kfd_create_process_device_data(dev, process);
 		if (!pdd) {
-			pr_err("Failed to create process device data\n");
+			dev_err(dev->adev->dev,
+				"Failed to create process device data\n");
 			return -ENOMEM;
 		}
 		/*
@@ -394,7 +393,7 @@ int kfd_init_apertures(struct kfd_process *process)
 			pdd->gpuvm_base = pdd->gpuvm_limit = 0;
 			pdd->scratch_base = pdd->scratch_limit = 0;
 		} else {
-			switch (dev->device_info->asic_family) {
+			switch (dev->adev->asic_type) {
 			case CHIP_KAVERI:
 			case CHIP_HAWAII:
 			case CHIP_CARRIZO:
@@ -406,37 +405,14 @@ int kfd_init_apertures(struct kfd_process *process)
 			case CHIP_VEGAM:
 				kfd_init_apertures_vi(pdd, id);
 				break;
-			case CHIP_VEGA10:
-			case CHIP_VEGA12:
-			case CHIP_VEGA20:
-			case CHIP_RAVEN:
-			case CHIP_RENOIR:
-			case CHIP_ARCTURUS:
-			case CHIP_ALDEBARAN:
-			case CHIP_NAVI10:
-			case CHIP_NAVI12:
-			case CHIP_NAVI14:
-			case CHIP_SIENNA_CICHLID:
-			case CHIP_NAVY_FLOUNDER:
-			case CHIP_VANGOGH:
-			case CHIP_DIMGREY_CAVEFISH:
-			case CHIP_BEIGE_GOBY:
-			case CHIP_YELLOW_CARP:
-			case CHIP_CYAN_SKILLFISH:
-				kfd_init_apertures_v9(pdd, id);
-				break;
 			default:
-				WARN(1, "Unexpected ASIC family %u",
-				     dev->device_info->asic_family);
-				return -EINVAL;
-			}
-
-			if (!dev->use_iommu_v2) {
-				/* dGPUs: the reserved space for kernel
-				 * before SVM
-				 */
-				pdd->qpd.cwsr_base = SVM_CWSR_BASE;
-				pdd->qpd.ib_base = SVM_IB_BASE;
+				if (KFD_GC_VERSION(dev) >= IP_VERSION(9, 0, 1))
+					kfd_init_apertures_v9(pdd, id);
+				else {
+					WARN(1, "Unexpected ASIC family %u",
+					     dev->adev->asic_type);
+					return -EINVAL;
+				}
 			}
 		}
 
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v10.c b/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v10.c
new file mode 100644
index 000000000000..3e1ad8974797
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v10.c
@@ -0,0 +1,386 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#include "kfd_events.h"
+#include "kfd_debug.h"
+#include "soc15_int.h"
+#include "kfd_device_queue_manager.h"
+
+/*
+ * GFX10 SQ Interrupts
+ *
+ * There are 3 encoding types of interrupts sourced from SQ sent as a 44-bit
+ * packet to the Interrupt Handler:
+ * Auto - Generated by the SQG (various cmd overflows, timestamps etc)
+ * Wave - Generated by S_SENDMSG through a shader program
+ * Error - HW generated errors (Illegal instructions, Memviols, EDC etc)
+ *
+ * The 44-bit packet is mapped as {context_id1[7:0],context_id0[31:0]} plus
+ * 4-bits for VMID (SOC15_VMID_FROM_IH_ENTRY) as such:
+ *
+ * - context_id1[7:6]
+ * Encoding type (0 = Auto, 1 = Wave, 2 = Error)
+ *
+ * - context_id0[24]
+ * PRIV bit indicates that Wave S_SEND or error occurred within trap
+ *
+ * - context_id0[22:0]
+ * 23-bit data with the following layout per encoding type:
+ * Auto - only context_id0[8:0] is used, which reports various interrupts
+ * generated by SQG.  The rest is 0.
+ * Wave - user data sent from m0 via S_SENDMSG
+ * Error - Error type (context_id0[22:19]), Error Details (rest of bits)
+ *
+ * The other context_id bits show coordinates (SE/SH/CU/SIMD/WGP) for wave
+ * S_SENDMSG and Errors.  These are 0 for Auto.
+ */
+
+enum SQ_INTERRUPT_WORD_ENCODING {
+	SQ_INTERRUPT_WORD_ENCODING_AUTO = 0x0,
+	SQ_INTERRUPT_WORD_ENCODING_INST,
+	SQ_INTERRUPT_WORD_ENCODING_ERROR,
+};
+
+enum SQ_INTERRUPT_ERROR_TYPE {
+	SQ_INTERRUPT_ERROR_TYPE_EDC_FUE = 0x0,
+	SQ_INTERRUPT_ERROR_TYPE_ILLEGAL_INST,
+	SQ_INTERRUPT_ERROR_TYPE_MEMVIOL,
+	SQ_INTERRUPT_ERROR_TYPE_EDC_FED,
+};
+
+/* SQ_INTERRUPT_WORD_AUTO_CTXID */
+#define SQ_INTERRUPT_WORD_AUTO_CTXID0__THREAD_TRACE__SHIFT 0
+#define SQ_INTERRUPT_WORD_AUTO_CTXID0__WLT__SHIFT 1
+#define SQ_INTERRUPT_WORD_AUTO_CTXID0__THREAD_TRACE_BUF0_FULL__SHIFT 2
+#define SQ_INTERRUPT_WORD_AUTO_CTXID0__THREAD_TRACE_BUF1_FULL__SHIFT 3
+#define SQ_INTERRUPT_WORD_AUTO_CTXID0__THREAD_TRACE_UTC_ERROR__SHIFT 7
+#define SQ_INTERRUPT_WORD_AUTO_CTXID1__SE_ID__SHIFT 4
+#define SQ_INTERRUPT_WORD_AUTO_CTXID1__ENCODING__SHIFT 6
+
+#define SQ_INTERRUPT_WORD_AUTO_CTXID0__THREAD_TRACE_MASK 0x00000001
+#define SQ_INTERRUPT_WORD_AUTO_CTXID0__WLT_MASK 0x00000002
+#define SQ_INTERRUPT_WORD_AUTO_CTXID0__THREAD_TRACE_BUF0_FULL_MASK 0x00000004
+#define SQ_INTERRUPT_WORD_AUTO_CTXID0__THREAD_TRACE_BUF1_FULL_MASK 0x00000008
+#define SQ_INTERRUPT_WORD_AUTO_CTXID0__THREAD_TRACE_UTC_ERROR_MASK 0x00000080
+#define SQ_INTERRUPT_WORD_AUTO_CTXID1__SE_ID_MASK 0x030
+#define SQ_INTERRUPT_WORD_AUTO_CTXID1__ENCODING_MASK 0x0c0
+
+/* SQ_INTERRUPT_WORD_WAVE_CTXID */
+#define SQ_INTERRUPT_WORD_WAVE_CTXID0__DATA__SHIFT 0
+#define SQ_INTERRUPT_WORD_WAVE_CTXID0__SA_ID__SHIFT 23
+#define SQ_INTERRUPT_WORD_WAVE_CTXID0__PRIV__SHIFT 24
+#define SQ_INTERRUPT_WORD_WAVE_CTXID0__WAVE_ID__SHIFT 25
+#define SQ_INTERRUPT_WORD_WAVE_CTXID0__SIMD_ID__SHIFT 30
+#define SQ_INTERRUPT_WORD_WAVE_CTXID1__WGP_ID__SHIFT 0
+#define SQ_INTERRUPT_WORD_WAVE_CTXID1__SE_ID__SHIFT 4
+#define SQ_INTERRUPT_WORD_WAVE_CTXID1__ENCODING__SHIFT 6
+
+#define SQ_INTERRUPT_WORD_WAVE_CTXID0__DATA_MASK 0x000007fffff
+#define SQ_INTERRUPT_WORD_WAVE_CTXID0__SA_ID_MASK 0x0000800000
+#define SQ_INTERRUPT_WORD_WAVE_CTXID0__PRIV_MASK 0x00001000000
+#define SQ_INTERRUPT_WORD_WAVE_CTXID0__WAVE_ID_MASK 0x0003e000000
+#define SQ_INTERRUPT_WORD_WAVE_CTXID0__SIMD_ID_MASK 0x000c0000000
+#define SQ_INTERRUPT_WORD_WAVE_CTXID1__WGP_ID_MASK 0x00f
+#define SQ_INTERRUPT_WORD_WAVE_CTXID1__SE_ID_MASK 0x030
+#define SQ_INTERRUPT_WORD_WAVE_CTXID1__ENCODING_MASK 0x0c0
+
+#define KFD_CTXID0__ERR_TYPE_MASK 0x780000
+#define KFD_CTXID0__ERR_TYPE__SHIFT 19
+
+/* GFX10 SQ interrupt ENC type bit (context_id1[7:6]) for wave s_sendmsg */
+#define KFD_CONTEXT_ID1_ENC_TYPE_WAVE_MASK	0x40
+/* GFX10 SQ interrupt PRIV bit (context_id0[24]) for s_sendmsg inside trap */
+#define KFD_CONTEXT_ID0_PRIV_MASK		0x1000000
+/*
+ * The debugger will send user data(m0) with PRIV=1 to indicate it requires
+ * notification from the KFD with the following queue id (DOORBELL_ID) and
+ * trap code (TRAP_CODE).
+ */
+#define KFD_CONTEXT_ID0_DEBUG_DOORBELL_MASK	0x0003ff
+#define KFD_CONTEXT_ID0_DEBUG_TRAP_CODE_SHIFT	10
+#define KFD_CONTEXT_ID0_DEBUG_TRAP_CODE_MASK	0x07fc00
+#define KFD_DEBUG_DOORBELL_ID(ctxid0)	((ctxid0) &	\
+				KFD_CONTEXT_ID0_DEBUG_DOORBELL_MASK)
+#define KFD_DEBUG_TRAP_CODE(ctxid0)	(((ctxid0) &	\
+				KFD_CONTEXT_ID0_DEBUG_TRAP_CODE_MASK)	\
+				>> KFD_CONTEXT_ID0_DEBUG_TRAP_CODE_SHIFT)
+#define KFD_DEBUG_CP_BAD_OP_ECODE_MASK		0x3fffc00
+#define KFD_DEBUG_CP_BAD_OP_ECODE_SHIFT		10
+#define KFD_DEBUG_CP_BAD_OP_ECODE(ctxid0) (((ctxid0) &			\
+				KFD_DEBUG_CP_BAD_OP_ECODE_MASK)		\
+				>> KFD_DEBUG_CP_BAD_OP_ECODE_SHIFT)
+
+static bool event_interrupt_isr_v10(struct kfd_node *dev,
+					const uint32_t *ih_ring_entry,
+					uint32_t *patched_ihre,
+					bool *patched_flag)
+{
+	uint16_t source_id, client_id, pasid, vmid;
+	const uint32_t *data = ih_ring_entry;
+
+	source_id = SOC15_SOURCE_ID_FROM_IH_ENTRY(ih_ring_entry);
+	client_id = SOC15_CLIENT_ID_FROM_IH_ENTRY(ih_ring_entry);
+
+	/* Only handle interrupts from KFD VMIDs */
+	vmid = SOC15_VMID_FROM_IH_ENTRY(ih_ring_entry);
+	if (!KFD_IRQ_IS_FENCE(client_id, source_id) &&
+	   (vmid < dev->vm_info.first_vmid_kfd ||
+	    vmid > dev->vm_info.last_vmid_kfd))
+		return false;
+
+	pasid = SOC15_PASID_FROM_IH_ENTRY(ih_ring_entry);
+
+	/* Only handle clients we care about */
+	if (client_id != SOC15_IH_CLIENTID_GRBM_CP &&
+	    client_id != SOC15_IH_CLIENTID_SDMA0 &&
+	    client_id != SOC15_IH_CLIENTID_SDMA1 &&
+	    client_id != SOC15_IH_CLIENTID_SDMA2 &&
+	    client_id != SOC15_IH_CLIENTID_SDMA3 &&
+	    client_id != SOC15_IH_CLIENTID_SDMA4 &&
+	    client_id != SOC15_IH_CLIENTID_SDMA5 &&
+	    client_id != SOC15_IH_CLIENTID_SDMA6 &&
+	    client_id != SOC15_IH_CLIENTID_SDMA7 &&
+	    client_id != SOC15_IH_CLIENTID_VMC &&
+	    client_id != SOC15_IH_CLIENTID_VMC1 &&
+	    client_id != SOC15_IH_CLIENTID_UTCL2 &&
+	    client_id != SOC15_IH_CLIENTID_SE0SH &&
+	    client_id != SOC15_IH_CLIENTID_SE1SH &&
+	    client_id != SOC15_IH_CLIENTID_SE2SH &&
+	    client_id != SOC15_IH_CLIENTID_SE3SH)
+		return false;
+
+	dev_dbg(dev->adev->dev,
+		"client id 0x%x, source id %d, vmid %d, pasid 0x%x. raw data:\n",
+		client_id, source_id, vmid, pasid);
+	dev_dbg(dev->adev->dev, "%8X, %8X, %8X, %8X, %8X, %8X, %8X, %8X.\n",
+		data[0], data[1], data[2], data[3], data[4], data[5], data[6],
+		data[7]);
+
+	if (pasid == 0)
+		return 0;
+
+	/* Interrupt types we care about: various signals and faults.
+	 * They will be forwarded to a work queue (see below).
+	 */
+	return source_id == SOC15_INTSRC_CP_END_OF_PIPE ||
+		source_id == SOC15_INTSRC_SDMA_TRAP ||
+		source_id == SOC15_INTSRC_SQ_INTERRUPT_MSG ||
+		source_id == SOC15_INTSRC_CP_BAD_OPCODE ||
+		client_id == SOC15_IH_CLIENTID_VMC ||
+		client_id == SOC15_IH_CLIENTID_VMC1 ||
+		client_id == SOC15_IH_CLIENTID_UTCL2 ||
+		KFD_IRQ_IS_FENCE(client_id, source_id);
+}
+
+static void event_interrupt_wq_v10(struct kfd_node *dev,
+					const uint32_t *ih_ring_entry)
+{
+	uint16_t source_id, client_id, pasid, vmid;
+	uint32_t context_id0, context_id1;
+	uint32_t encoding, sq_intr_err_type;
+
+	source_id = SOC15_SOURCE_ID_FROM_IH_ENTRY(ih_ring_entry);
+	client_id = SOC15_CLIENT_ID_FROM_IH_ENTRY(ih_ring_entry);
+	pasid = SOC15_PASID_FROM_IH_ENTRY(ih_ring_entry);
+	vmid = SOC15_VMID_FROM_IH_ENTRY(ih_ring_entry);
+	context_id0 = SOC15_CONTEXT_ID0_FROM_IH_ENTRY(ih_ring_entry);
+	context_id1 = SOC15_CONTEXT_ID1_FROM_IH_ENTRY(ih_ring_entry);
+
+	if (client_id == SOC15_IH_CLIENTID_GRBM_CP ||
+	    client_id == SOC15_IH_CLIENTID_SE0SH ||
+	    client_id == SOC15_IH_CLIENTID_SE1SH ||
+	    client_id == SOC15_IH_CLIENTID_SE2SH ||
+	    client_id == SOC15_IH_CLIENTID_SE3SH) {
+		if (source_id == SOC15_INTSRC_CP_END_OF_PIPE)
+			kfd_signal_event_interrupt(pasid, context_id0, 32);
+		else if (source_id == SOC15_INTSRC_SQ_INTERRUPT_MSG) {
+			encoding = REG_GET_FIELD(context_id1,
+						SQ_INTERRUPT_WORD_WAVE_CTXID1, ENCODING);
+			switch (encoding) {
+			case SQ_INTERRUPT_WORD_ENCODING_AUTO:
+				dev_dbg_ratelimited(
+					dev->adev->dev,
+					"sq_intr: auto, se %d, ttrace %d, wlt %d, ttrac_buf0_full %d, ttrac_buf1_full %d, ttrace_utc_err %d\n",
+					REG_GET_FIELD(
+						context_id1,
+						SQ_INTERRUPT_WORD_AUTO_CTXID1,
+						SE_ID),
+					REG_GET_FIELD(
+						context_id0,
+						SQ_INTERRUPT_WORD_AUTO_CTXID0,
+						THREAD_TRACE),
+					REG_GET_FIELD(
+						context_id0,
+						SQ_INTERRUPT_WORD_AUTO_CTXID0,
+						WLT),
+					REG_GET_FIELD(
+						context_id0,
+						SQ_INTERRUPT_WORD_AUTO_CTXID0,
+						THREAD_TRACE_BUF0_FULL),
+					REG_GET_FIELD(
+						context_id0,
+						SQ_INTERRUPT_WORD_AUTO_CTXID0,
+						THREAD_TRACE_BUF1_FULL),
+					REG_GET_FIELD(
+						context_id0,
+						SQ_INTERRUPT_WORD_AUTO_CTXID0,
+						THREAD_TRACE_UTC_ERROR));
+				break;
+			case SQ_INTERRUPT_WORD_ENCODING_INST:
+				dev_dbg_ratelimited(
+					dev->adev->dev,
+					"sq_intr: inst, se %d, data 0x%x, sa %d, priv %d, wave_id %d, simd_id %d, wgp_id %d\n",
+					REG_GET_FIELD(
+						context_id1,
+						SQ_INTERRUPT_WORD_WAVE_CTXID1,
+						SE_ID),
+					REG_GET_FIELD(
+						context_id0,
+						SQ_INTERRUPT_WORD_WAVE_CTXID0,
+						DATA),
+					REG_GET_FIELD(
+						context_id0,
+						SQ_INTERRUPT_WORD_WAVE_CTXID0,
+						SA_ID),
+					REG_GET_FIELD(
+						context_id0,
+						SQ_INTERRUPT_WORD_WAVE_CTXID0,
+						PRIV),
+					REG_GET_FIELD(
+						context_id0,
+						SQ_INTERRUPT_WORD_WAVE_CTXID0,
+						WAVE_ID),
+					REG_GET_FIELD(
+						context_id0,
+						SQ_INTERRUPT_WORD_WAVE_CTXID0,
+						SIMD_ID),
+					REG_GET_FIELD(
+						context_id1,
+						SQ_INTERRUPT_WORD_WAVE_CTXID1,
+						WGP_ID));
+				if (context_id0 & SQ_INTERRUPT_WORD_WAVE_CTXID0__PRIV_MASK) {
+					if (kfd_set_dbg_ev_from_interrupt(dev, pasid,
+							KFD_DEBUG_DOORBELL_ID(context_id0),
+							KFD_DEBUG_TRAP_CODE(context_id0),
+							NULL, 0))
+						return;
+				}
+				break;
+			case SQ_INTERRUPT_WORD_ENCODING_ERROR:
+				sq_intr_err_type = REG_GET_FIELD(context_id0, KFD_CTXID0,
+								ERR_TYPE);
+				dev_warn_ratelimited(
+					dev->adev->dev,
+					"sq_intr: error, se %d, data 0x%x, sa %d, priv %d, wave_id %d, simd_id %d, wgp_id %d, err_type %d\n",
+					REG_GET_FIELD(
+						context_id1,
+						SQ_INTERRUPT_WORD_WAVE_CTXID1,
+						SE_ID),
+					REG_GET_FIELD(
+						context_id0,
+						SQ_INTERRUPT_WORD_WAVE_CTXID0,
+						DATA),
+					REG_GET_FIELD(
+						context_id0,
+						SQ_INTERRUPT_WORD_WAVE_CTXID0,
+						SA_ID),
+					REG_GET_FIELD(
+						context_id0,
+						SQ_INTERRUPT_WORD_WAVE_CTXID0,
+						PRIV),
+					REG_GET_FIELD(
+						context_id0,
+						SQ_INTERRUPT_WORD_WAVE_CTXID0,
+						WAVE_ID),
+					REG_GET_FIELD(
+						context_id0,
+						SQ_INTERRUPT_WORD_WAVE_CTXID0,
+						SIMD_ID),
+					REG_GET_FIELD(
+						context_id1,
+						SQ_INTERRUPT_WORD_WAVE_CTXID1,
+						WGP_ID),
+					sq_intr_err_type);
+				break;
+			default:
+				break;
+			}
+			kfd_signal_event_interrupt(pasid, context_id0 & 0x7fffff, 23);
+		} else if (source_id == SOC15_INTSRC_CP_BAD_OPCODE &&
+			   KFD_DBG_EC_TYPE_IS_PACKET(KFD_DEBUG_CP_BAD_OP_ECODE(context_id0))) {
+			kfd_set_dbg_ev_from_interrupt(dev, pasid,
+				KFD_DEBUG_DOORBELL_ID(context_id0),
+				KFD_EC_MASK(KFD_DEBUG_CP_BAD_OP_ECODE(context_id0)),
+				NULL,
+				0);
+		}
+	} else if (client_id == SOC15_IH_CLIENTID_SDMA0 ||
+		   client_id == SOC15_IH_CLIENTID_SDMA1 ||
+		   client_id == SOC15_IH_CLIENTID_SDMA2 ||
+		   client_id == SOC15_IH_CLIENTID_SDMA3 ||
+		   (client_id == SOC15_IH_CLIENTID_SDMA3_Sienna_Cichlid &&
+		    KFD_GC_VERSION(dev) == IP_VERSION(10, 3, 0)) ||
+		   client_id == SOC15_IH_CLIENTID_SDMA4 ||
+		   client_id == SOC15_IH_CLIENTID_SDMA5 ||
+		   client_id == SOC15_IH_CLIENTID_SDMA6 ||
+		   client_id == SOC15_IH_CLIENTID_SDMA7) {
+		if (source_id == SOC15_INTSRC_SDMA_TRAP) {
+			kfd_signal_event_interrupt(pasid, context_id0 & 0xfffffff, 28);
+		}
+	} else if (client_id == SOC15_IH_CLIENTID_VMC ||
+		   client_id == SOC15_IH_CLIENTID_VMC1 ||
+		   client_id == SOC15_IH_CLIENTID_UTCL2) {
+		struct kfd_vm_fault_info info = {0};
+		uint16_t ring_id = SOC15_RING_ID_FROM_IH_ENTRY(ih_ring_entry);
+		struct kfd_hsa_memory_exception_data exception_data;
+
+		info.vmid = vmid;
+		info.mc_id = client_id;
+		info.page_addr = ih_ring_entry[4] |
+			(uint64_t)(ih_ring_entry[5] & 0xf) << 32;
+		info.prot_valid = ring_id & 0x08;
+		info.prot_read  = ring_id & 0x10;
+		info.prot_write = ring_id & 0x20;
+
+		memset(&exception_data, 0, sizeof(exception_data));
+		exception_data.gpu_id = dev->id;
+		exception_data.va = (info.page_addr) << PAGE_SHIFT;
+		exception_data.failure.NotPresent = info.prot_valid ? 1 : 0;
+		exception_data.failure.NoExecute = info.prot_exec ? 1 : 0;
+		exception_data.failure.ReadOnly = info.prot_write ? 1 : 0;
+		exception_data.failure.imprecise = 0;
+
+		kfd_set_dbg_ev_from_interrupt(dev,
+						pasid,
+						-1,
+						KFD_EC_MASK(EC_DEVICE_MEMORY_VIOLATION),
+						&exception_data,
+						sizeof(exception_data));
+	} else if (KFD_IRQ_IS_FENCE(client_id, source_id)) {
+		kfd_process_close_interrupt_drain(pasid);
+	}
+}
+
+const struct kfd_event_interrupt_class event_interrupt_class_v10 = {
+	.interrupt_isr = event_interrupt_isr_v10,
+	.interrupt_wq = event_interrupt_wq_v10,
+};
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v11.c b/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v11.c
new file mode 100644
index 000000000000..2788a52714d1
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v11.c
@@ -0,0 +1,418 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#include "kfd_priv.h"
+#include "kfd_events.h"
+#include "soc15_int.h"
+#include "kfd_device_queue_manager.h"
+#include "ivsrcid/vmc/irqsrcs_vmc_1_0.h"
+#include "kfd_smi_events.h"
+#include "kfd_debug.h"
+
+/*
+ * GFX11 SQ Interrupts
+ *
+ * There are 3 encoding types of interrupts sourced from SQ sent as a 44-bit
+ * packet to the Interrupt Handler:
+ * Auto - Generated by the SQG (various cmd overflows, timestamps etc)
+ * Wave - Generated by S_SENDMSG through a shader program
+ * Error - HW generated errors (Illegal instructions, Memviols, EDC etc)
+ *
+ * The 44-bit packet is mapped as {context_id1[7:0],context_id0[31:0]} plus
+ * 4-bits for VMID (SOC15_VMID_FROM_IH_ENTRY) as such:
+ *
+ * - context_id1[7:6]
+ * Encoding type (0 = Auto, 1 = Wave, 2 = Error)
+ *
+ * - context_id0[26]
+ * PRIV bit indicates that Wave S_SEND or error occurred within trap
+ *
+ * - context_id0[24:0]
+ * 25-bit data with the following layout per encoding type:
+ * Auto - only context_id0[8:0] is used, which reports various interrupts
+ * generated by SQG.  The rest is 0.
+ * Wave - user data sent from m0 via S_SENDMSG (context_id0[23:0])
+ * Error - Error Type (context_id0[24:21]), Error Details (context_id0[20:0])
+ *
+ * The other context_id bits show coordinates (SE/SH/CU/SIMD/WGP) for wave
+ * S_SENDMSG and Errors.  These are 0 for Auto.
+ */
+
+enum SQ_INTERRUPT_WORD_ENCODING {
+	SQ_INTERRUPT_WORD_ENCODING_AUTO = 0x0,
+	SQ_INTERRUPT_WORD_ENCODING_INST,
+	SQ_INTERRUPT_WORD_ENCODING_ERROR,
+};
+
+enum SQ_INTERRUPT_ERROR_TYPE {
+	SQ_INTERRUPT_ERROR_TYPE_EDC_FUE = 0x0,
+	SQ_INTERRUPT_ERROR_TYPE_ILLEGAL_INST,
+	SQ_INTERRUPT_ERROR_TYPE_MEMVIOL,
+	SQ_INTERRUPT_ERROR_TYPE_EDC_FED,
+};
+
+/* SQ_INTERRUPT_WORD_AUTO_CTXID */
+#define SQ_INTERRUPT_WORD_AUTO_CTXID0__THREAD_TRACE__SHIFT		0
+#define SQ_INTERRUPT_WORD_AUTO_CTXID0__WLT__SHIFT			1
+#define SQ_INTERRUPT_WORD_AUTO_CTXID0__THREAD_TRACE_BUF_FULL__SHIFT	2
+#define SQ_INTERRUPT_WORD_AUTO_CTXID0__REG_TIMESTAMP__SHIFT		3
+#define SQ_INTERRUPT_WORD_AUTO_CTXID0__CMD_TIMESTAMP__SHIFT		4
+#define SQ_INTERRUPT_WORD_AUTO_CTXID0__HOST_CMD_OVERFLOW__SHIFT		5
+#define SQ_INTERRUPT_WORD_AUTO_CTXID0__HOST_REG_OVERFLOW__SHIFT		6
+#define SQ_INTERRUPT_WORD_AUTO_CTXID0__IMMED_OVERFLOW__SHIFT		7
+#define SQ_INTERRUPT_WORD_AUTO_CTXID0__THREAD_TRACE_UTC_ERROR__SHIFT	8
+#define SQ_INTERRUPT_WORD_AUTO_CTXID1__ENCODING__SHIFT			6
+
+#define SQ_INTERRUPT_WORD_AUTO_CTXID0__THREAD_TRACE_MASK		0x00000001
+#define SQ_INTERRUPT_WORD_AUTO_CTXID0__WLT_MASK				0x00000002
+#define SQ_INTERRUPT_WORD_AUTO_CTXID0__THREAD_TRACE_BUF_FULL_MASK	0x00000004
+#define SQ_INTERRUPT_WORD_AUTO_CTXID0__REG_TIMESTAMP_MASK		0x00000008
+#define SQ_INTERRUPT_WORD_AUTO_CTXID0__CMD_TIMESTAMP_MASK		0x00000010
+#define SQ_INTERRUPT_WORD_AUTO_CTXID0__HOST_CMD_OVERFLOW_MASK		0x00000020
+#define SQ_INTERRUPT_WORD_AUTO_CTXID0__HOST_REG_OVERFLOW_MASK		0x00000040
+#define SQ_INTERRUPT_WORD_AUTO_CTXID0__IMMED_OVERFLOW_MASK		0x00000080
+#define SQ_INTERRUPT_WORD_AUTO_CTXID0__THREAD_TRACE_UTC_ERROR_MASK	0x00000100
+#define SQ_INTERRUPT_WORD_AUTO_CTXID1__ENCODING_MASK			0x000000c0
+
+/* SQ_INTERRUPT_WORD_WAVE_CTXID */
+#define SQ_INTERRUPT_WORD_WAVE_CTXID0__DATA__SHIFT	0
+#define SQ_INTERRUPT_WORD_WAVE_CTXID0__SH_ID__SHIFT	25
+#define SQ_INTERRUPT_WORD_WAVE_CTXID0__PRIV__SHIFT	26
+#define SQ_INTERRUPT_WORD_WAVE_CTXID0__WAVE_ID__SHIFT	27
+#define SQ_INTERRUPT_WORD_WAVE_CTXID1__SIMD_ID__SHIFT	0
+#define SQ_INTERRUPT_WORD_WAVE_CTXID1__WGP_ID__SHIFT	2
+#define SQ_INTERRUPT_WORD_WAVE_CTXID1__ENCODING__SHIFT	6
+
+#define SQ_INTERRUPT_WORD_WAVE_CTXID0__DATA_MASK	0x00ffffff /* [23:0] */
+#define SQ_INTERRUPT_WORD_WAVE_CTXID0__SH_ID_MASK	0x02000000 /* [25] */
+#define SQ_INTERRUPT_WORD_WAVE_CTXID0__PRIV_MASK	0x04000000 /* [26] */
+#define SQ_INTERRUPT_WORD_WAVE_CTXID0__WAVE_ID_MASK	0xf8000000 /* [31:27] */
+#define SQ_INTERRUPT_WORD_WAVE_CTXID1__SIMD_ID_MASK	0x00000003 /* [33:32] */
+#define SQ_INTERRUPT_WORD_WAVE_CTXID1__WGP_ID_MASK	0x0000003c /* [37:34] */
+#define SQ_INTERRUPT_WORD_WAVE_CTXID1__ENCODING_MASK	0x000000c0 /* [39:38] */
+
+/* SQ_INTERRUPT_WORD_ERROR_CTXID */
+#define SQ_INTERRUPT_WORD_ERROR_CTXID0__DETAIL__SHIFT	0
+#define SQ_INTERRUPT_WORD_ERROR_CTXID0__TYPE__SHIFT	21
+#define SQ_INTERRUPT_WORD_ERROR_CTXID0__SH_ID__SHIFT	25
+#define SQ_INTERRUPT_WORD_ERROR_CTXID0__PRIV__SHIFT	26
+#define SQ_INTERRUPT_WORD_ERROR_CTXID0__WAVE_ID__SHIFT	27
+#define SQ_INTERRUPT_WORD_ERROR_CTXID1__SIMD_ID__SHIFT	0
+#define SQ_INTERRUPT_WORD_ERROR_CTXID1__WGP_ID__SHIFT	2
+#define SQ_INTERRUPT_WORD_ERROR_CTXID1__ENCODING__SHIFT	6
+
+#define SQ_INTERRUPT_WORD_ERROR_CTXID0__DETAIL_MASK	0x001fffff /* [20:0] */
+#define SQ_INTERRUPT_WORD_ERROR_CTXID0__TYPE_MASK	0x01e00000 /* [24:21] */
+#define SQ_INTERRUPT_WORD_ERROR_CTXID0__SH_ID_MASK	0x02000000 /* [25] */
+#define SQ_INTERRUPT_WORD_ERROR_CTXID0__PRIV_MASK	0x04000000 /* [26] */
+#define SQ_INTERRUPT_WORD_ERROR_CTXID0__WAVE_ID_MASK	0xf8000000 /* [31:27] */
+#define SQ_INTERRUPT_WORD_ERROR_CTXID1__SIMD_ID_MASK	0x00000003 /* [33:32] */
+#define SQ_INTERRUPT_WORD_ERROR_CTXID1__WGP_ID_MASK	0x0000003c /* [37:34] */
+#define SQ_INTERRUPT_WORD_ERROR_CTXID1__ENCODING_MASK	0x000000c0 /* [39:38] */
+
+/*
+ * The debugger will send user data(m0) with PRIV=1 to indicate it requires
+ * notification from the KFD with the following queue id (DOORBELL_ID) and
+ * trap code (TRAP_CODE).
+ */
+#define KFD_CTXID0_TRAP_CODE_SHIFT	10
+#define KFD_CTXID0_TRAP_CODE_MASK	0xfffc00
+#define KFD_CTXID0_CP_BAD_OP_ECODE_MASK	0x3ffffff
+#define KFD_CTXID0_DOORBELL_ID_MASK	0x0003ff
+
+#define KFD_CTXID0_TRAP_CODE(ctxid0)		(((ctxid0) &  \
+				KFD_CTXID0_TRAP_CODE_MASK) >> \
+				KFD_CTXID0_TRAP_CODE_SHIFT)
+#define KFD_CTXID0_CP_BAD_OP_ECODE(ctxid0)	(((ctxid0) &        \
+				KFD_CTXID0_CP_BAD_OP_ECODE_MASK) >> \
+				KFD_CTXID0_TRAP_CODE_SHIFT)
+#define KFD_CTXID0_DOORBELL_ID(ctxid0)		((ctxid0) & \
+				KFD_CTXID0_DOORBELL_ID_MASK)
+
+static void print_sq_intr_info_auto(struct kfd_node *dev, uint32_t context_id0,
+				    uint32_t context_id1)
+{
+	dev_dbg_ratelimited(
+		dev->adev->dev,
+		"sq_intr: auto, ttrace %d, wlt %d, ttrace_buf_full %d, reg_tms %d, cmd_tms %d, host_cmd_ovf %d, host_reg_ovf %d, immed_ovf %d, ttrace_utc_err %d\n",
+		REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID0,
+			      THREAD_TRACE),
+		REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID0, WLT),
+		REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID0,
+			      THREAD_TRACE_BUF_FULL),
+		REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID0,
+			      REG_TIMESTAMP),
+		REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID0,
+			      CMD_TIMESTAMP),
+		REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID0,
+			      HOST_CMD_OVERFLOW),
+		REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID0,
+			      HOST_REG_OVERFLOW),
+		REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID0,
+			      IMMED_OVERFLOW),
+		REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID0,
+			      THREAD_TRACE_UTC_ERROR));
+}
+
+static void print_sq_intr_info_inst(struct kfd_node *dev, uint32_t context_id0,
+				    uint32_t context_id1)
+{
+	dev_dbg_ratelimited(
+		dev->adev->dev,
+		"sq_intr: inst, data 0x%08x, sh %d, priv %d, wave_id %d, simd_id %d, wgp_id %d\n",
+		REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID0, DATA),
+		REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID0,
+			      SH_ID),
+		REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID0, PRIV),
+		REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID0,
+			      WAVE_ID),
+		REG_GET_FIELD(context_id1, SQ_INTERRUPT_WORD_WAVE_CTXID1,
+			      SIMD_ID),
+		REG_GET_FIELD(context_id1, SQ_INTERRUPT_WORD_WAVE_CTXID1,
+			      WGP_ID));
+}
+
+static void print_sq_intr_info_error(struct kfd_node *dev, uint32_t context_id0,
+				     uint32_t context_id1)
+{
+	dev_warn_ratelimited(
+		dev->adev->dev,
+		"sq_intr: error, detail 0x%08x, type %d, sh %d, priv %d, wave_id %d, simd_id %d, wgp_id %d\n",
+		REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_ERROR_CTXID0,
+			      DETAIL),
+		REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_ERROR_CTXID0,
+			      TYPE),
+		REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_ERROR_CTXID0,
+			      SH_ID),
+		REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_ERROR_CTXID0,
+			      PRIV),
+		REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_ERROR_CTXID0,
+			      WAVE_ID),
+		REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_ERROR_CTXID1,
+			      SIMD_ID),
+		REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_ERROR_CTXID1,
+			      WGP_ID));
+}
+
+static void event_interrupt_poison_consumption_v11(struct kfd_node *dev,
+				uint16_t pasid, uint16_t source_id)
+{
+	enum amdgpu_ras_block block = 0;
+	int ret = -EINVAL;
+	uint32_t reset = 0;
+	struct kfd_process *p = kfd_lookup_process_by_pasid(pasid, NULL);
+
+	if (!p)
+		return;
+
+	/* all queues of a process will be unmapped in one time */
+	if (atomic_read(&p->poison)) {
+		kfd_unref_process(p);
+		return;
+	}
+
+	atomic_set(&p->poison, 1);
+	kfd_unref_process(p);
+
+	switch (source_id) {
+	case SOC15_INTSRC_SQ_INTERRUPT_MSG:
+		if (dev->dqm->ops.reset_queues)
+			ret = dev->dqm->ops.reset_queues(dev->dqm, pasid);
+		block = AMDGPU_RAS_BLOCK__GFX;
+		if (ret)
+			reset = AMDGPU_RAS_GPU_RESET_MODE2_RESET;
+		break;
+	case SOC21_INTSRC_SDMA_ECC:
+	default:
+		block = AMDGPU_RAS_BLOCK__GFX;
+		reset = AMDGPU_RAS_GPU_RESET_MODE2_RESET;
+		break;
+	}
+
+	kfd_signal_poison_consumed_event(dev, pasid);
+
+	/* resetting queue passes, do page retirement without gpu reset
+	   resetting queue fails, fallback to gpu reset solution */
+	amdgpu_amdkfd_ras_poison_consumption_handler(dev->adev, block, reset);
+}
+
+static bool event_interrupt_isr_v11(struct kfd_node *dev,
+					const uint32_t *ih_ring_entry,
+					uint32_t *patched_ihre,
+					bool *patched_flag)
+{
+	uint16_t source_id, client_id, pasid, vmid;
+	const uint32_t *data = ih_ring_entry;
+	uint32_t context_id0;
+
+	source_id = SOC15_SOURCE_ID_FROM_IH_ENTRY(ih_ring_entry);
+	client_id = SOC15_CLIENT_ID_FROM_IH_ENTRY(ih_ring_entry);
+	/* Only handle interrupts from KFD VMIDs */
+	vmid = SOC15_VMID_FROM_IH_ENTRY(ih_ring_entry);
+	if (!KFD_IRQ_IS_FENCE(client_id, source_id) &&
+	    (vmid < dev->vm_info.first_vmid_kfd ||
+	    vmid > dev->vm_info.last_vmid_kfd))
+		return false;
+
+	pasid = SOC15_PASID_FROM_IH_ENTRY(ih_ring_entry);
+	context_id0 = SOC15_CONTEXT_ID0_FROM_IH_ENTRY(ih_ring_entry);
+
+	if ((source_id == SOC15_INTSRC_CP_END_OF_PIPE) &&
+	    (context_id0 & AMDGPU_FENCE_MES_QUEUE_FLAG))
+		return false;
+
+	dev_dbg(dev->adev->dev,
+		"client id 0x%x, source id %d, vmid %d, pasid 0x%x. raw data:\n",
+		client_id, source_id, vmid, pasid);
+	dev_dbg(dev->adev->dev, "%8X, %8X, %8X, %8X, %8X, %8X, %8X, %8X.\n",
+		data[0], data[1], data[2], data[3], data[4], data[5], data[6],
+		data[7]);
+
+	if (pasid == 0)
+		return false;
+
+	/* Interrupt types we care about: various signals and faults.
+	 * They will be forwarded to a work queue (see below).
+	 */
+	return source_id == SOC15_INTSRC_CP_END_OF_PIPE ||
+		source_id == SOC15_INTSRC_SQ_INTERRUPT_MSG ||
+		source_id == SOC15_INTSRC_CP_BAD_OPCODE ||
+		source_id == SOC21_INTSRC_SDMA_TRAP ||
+		KFD_IRQ_IS_FENCE(client_id, source_id) ||
+		(((client_id == SOC21_IH_CLIENTID_VMC) ||
+		 ((client_id == SOC21_IH_CLIENTID_GFX) &&
+		  (source_id == UTCL2_1_0__SRCID__FAULT))) &&
+		  !amdgpu_no_queue_eviction_on_vm_fault);
+}
+
+static void event_interrupt_wq_v11(struct kfd_node *dev,
+					const uint32_t *ih_ring_entry)
+{
+	uint16_t source_id, client_id, ring_id, pasid, vmid;
+	uint32_t context_id0, context_id1;
+	uint8_t sq_int_enc, sq_int_priv, sq_int_errtype;
+	struct kfd_vm_fault_info info = {0};
+	struct kfd_hsa_memory_exception_data exception_data;
+
+	source_id = SOC15_SOURCE_ID_FROM_IH_ENTRY(ih_ring_entry);
+	client_id = SOC15_CLIENT_ID_FROM_IH_ENTRY(ih_ring_entry);
+	ring_id = SOC15_RING_ID_FROM_IH_ENTRY(ih_ring_entry);
+	pasid = SOC15_PASID_FROM_IH_ENTRY(ih_ring_entry);
+	vmid = SOC15_VMID_FROM_IH_ENTRY(ih_ring_entry);
+	context_id0 = SOC15_CONTEXT_ID0_FROM_IH_ENTRY(ih_ring_entry);
+	context_id1 = SOC15_CONTEXT_ID1_FROM_IH_ENTRY(ih_ring_entry);
+
+	/* VMC, UTCL2 */
+	if (client_id == SOC21_IH_CLIENTID_VMC ||
+	     ((client_id == SOC21_IH_CLIENTID_GFX) &&
+	     (source_id == UTCL2_1_0__SRCID__FAULT))) {
+
+		info.vmid = vmid;
+		info.mc_id = client_id;
+		info.page_addr = ih_ring_entry[4] |
+			(uint64_t)(ih_ring_entry[5] & 0xf) << 32;
+		info.prot_valid = ring_id & 0x08;
+		info.prot_read  = ring_id & 0x10;
+		info.prot_write = ring_id & 0x20;
+
+		memset(&exception_data, 0, sizeof(exception_data));
+		exception_data.gpu_id = dev->id;
+		exception_data.va = (info.page_addr) << PAGE_SHIFT;
+		exception_data.failure.NotPresent = info.prot_valid ? 1 : 0;
+		exception_data.failure.NoExecute = info.prot_exec ? 1 : 0;
+		exception_data.failure.ReadOnly = info.prot_write ? 1 : 0;
+		exception_data.failure.imprecise = 0;
+
+		kfd_set_dbg_ev_from_interrupt(dev, pasid, -1,
+					      KFD_EC_MASK(EC_DEVICE_MEMORY_VIOLATION),
+					      &exception_data, sizeof(exception_data));
+		kfd_smi_event_update_vmfault(dev, pasid);
+
+	/* GRBM, SDMA, SE, PMM */
+	} else if (client_id == SOC21_IH_CLIENTID_GRBM_CP ||
+		   client_id == SOC21_IH_CLIENTID_GFX) {
+
+		/* CP */
+		if (source_id == SOC15_INTSRC_CP_END_OF_PIPE)
+			kfd_signal_event_interrupt(pasid, context_id0, 32);
+		else if (source_id == SOC15_INTSRC_CP_BAD_OPCODE &&
+			 KFD_DBG_EC_TYPE_IS_PACKET(KFD_CTXID0_CP_BAD_OP_ECODE(context_id0))) {
+			u32 doorbell_id = KFD_CTXID0_DOORBELL_ID(context_id0);
+
+			kfd_set_dbg_ev_from_interrupt(dev, pasid, doorbell_id,
+				KFD_EC_MASK(KFD_CTXID0_CP_BAD_OP_ECODE(context_id0)),
+				NULL, 0);
+			kfd_dqm_suspend_bad_queue_mes(dev, pasid, doorbell_id);
+		}
+
+		/* SDMA */
+		else if (source_id == SOC21_INTSRC_SDMA_TRAP)
+			kfd_signal_event_interrupt(pasid, context_id0 & 0xfffffff, 28);
+		else if (source_id == SOC21_INTSRC_SDMA_ECC) {
+			event_interrupt_poison_consumption_v11(dev, pasid, source_id);
+			return;
+		}
+
+		/* SQ */
+		else if (source_id == SOC15_INTSRC_SQ_INTERRUPT_MSG) {
+			sq_int_enc = REG_GET_FIELD(context_id1,
+					SQ_INTERRUPT_WORD_WAVE_CTXID1, ENCODING);
+			switch (sq_int_enc) {
+			case SQ_INTERRUPT_WORD_ENCODING_AUTO:
+				print_sq_intr_info_auto(dev, context_id0, context_id1);
+				break;
+			case SQ_INTERRUPT_WORD_ENCODING_INST:
+				print_sq_intr_info_inst(dev, context_id0, context_id1);
+				sq_int_priv = REG_GET_FIELD(context_id0,
+						SQ_INTERRUPT_WORD_WAVE_CTXID0, PRIV);
+				if (sq_int_priv && (kfd_set_dbg_ev_from_interrupt(dev, pasid,
+						KFD_CTXID0_DOORBELL_ID(context_id0),
+						KFD_CTXID0_TRAP_CODE(context_id0),
+						NULL, 0)))
+					return;
+				break;
+			case SQ_INTERRUPT_WORD_ENCODING_ERROR:
+				print_sq_intr_info_error(dev, context_id0, context_id1);
+				sq_int_errtype = REG_GET_FIELD(context_id0,
+						SQ_INTERRUPT_WORD_ERROR_CTXID0, TYPE);
+				if (sq_int_errtype != SQ_INTERRUPT_ERROR_TYPE_ILLEGAL_INST &&
+				    sq_int_errtype != SQ_INTERRUPT_ERROR_TYPE_MEMVIOL) {
+					event_interrupt_poison_consumption_v11(
+							dev, pasid, source_id);
+					return;
+				}
+				break;
+			default:
+				break;
+			}
+			kfd_signal_event_interrupt(pasid, context_id0 & 0xffffff, 24);
+		}
+
+	} else if (KFD_IRQ_IS_FENCE(client_id, source_id)) {
+		kfd_process_close_interrupt_drain(pasid);
+	}
+}
+
+const struct kfd_event_interrupt_class event_interrupt_class_v11 = {
+	.interrupt_isr = event_interrupt_isr_v11,
+	.interrupt_wq = event_interrupt_wq_v11,
+};
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v9.c b/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v9.c
index 543e7ea75593..d76fb61869c7 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v9.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v9.c
@@ -1,5 +1,6 @@
+// SPDX-License-Identifier: GPL-2.0 OR MIT
 /*
- * Copyright 2016-2018 Advanced Micro Devices, Inc.
+ * Copyright 2016-2022 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -22,9 +23,41 @@
 
 #include "kfd_priv.h"
 #include "kfd_events.h"
+#include "kfd_debug.h"
 #include "soc15_int.h"
 #include "kfd_device_queue_manager.h"
 #include "kfd_smi_events.h"
+#include "amdgpu_ras.h"
+#include "amdgpu_ras_mgr.h"
+
+/*
+ * GFX9 SQ Interrupts
+ *
+ * There are 3 encoding types of interrupts sourced from SQ sent as a 44-bit
+ * packet to the Interrupt Handler:
+ * Auto - Generated by the SQG (various cmd overflows, timestamps etc)
+ * Wave - Generated by S_SENDMSG through a shader program
+ * Error - HW generated errors (Illegal instructions, Memviols, EDC etc)
+ *
+ * The 44-bit packet is mapped as {context_id1[7:0],context_id0[31:0]} plus
+ * 4-bits for VMID (SOC15_VMID_FROM_IH_ENTRY) as such:
+ *
+ * - context_id0[27:26]
+ * Encoding type (0 = Auto, 1 = Wave, 2 = Error)
+ *
+ * - context_id0[13]
+ * PRIV bit indicates that Wave S_SEND or error occurred within trap
+ *
+ * - {context_id1[7:0],context_id0[31:28],context_id0[11:0]}
+ * 24-bit data with the following layout per encoding type:
+ * Auto - only context_id0[8:0] is used, which reports various interrupts
+ * generated by SQG.  The rest is 0.
+ * Wave - user data sent from m0 via S_SENDMSG
+ * Error - Error type (context_id1[7:4]), Error Details (rest of bits)
+ *
+ * The other context_id bits show coordinates (SE/SH/CU/SIMD/WAVE) for wave
+ * S_SENDMSG and Errors.  These are 0 for Auto.
+ */
 
 enum SQ_INTERRUPT_WORD_ENCODING {
 	SQ_INTERRUPT_WORD_ENCODING_AUTO = 0x0,
@@ -83,13 +116,152 @@ enum SQ_INTERRUPT_ERROR_TYPE {
 #define SQ_INTERRUPT_WORD_WAVE_CTXID__SE_ID_MASK 0x03000000
 #define SQ_INTERRUPT_WORD_WAVE_CTXID__ENCODING_MASK 0x0c000000
 
+/* GFX9 SQ interrupt 24-bit data from context_id<0,1> */
 #define KFD_CONTEXT_ID_GET_SQ_INT_DATA(ctx0, ctx1)                             \
 	((ctx0 & 0xfff) | ((ctx0 >> 16) & 0xf000) | ((ctx1 << 16) & 0xff0000))
 
 #define KFD_SQ_INT_DATA__ERR_TYPE_MASK 0xF00000
 #define KFD_SQ_INT_DATA__ERR_TYPE__SHIFT 20
 
-static bool event_interrupt_isr_v9(struct kfd_dev *dev,
+/*
+ * The debugger will send user data(m0) with PRIV=1 to indicate it requires
+ * notification from the KFD with the following queue id (DOORBELL_ID) and
+ * trap code (TRAP_CODE).
+ */
+#define KFD_INT_DATA_DEBUG_DOORBELL_MASK	0x0003ff
+#define KFD_INT_DATA_DEBUG_TRAP_CODE_SHIFT	10
+#define KFD_INT_DATA_DEBUG_TRAP_CODE_MASK	0x07fc00
+#define KFD_DEBUG_DOORBELL_ID(sq_int_data)	((sq_int_data) &	\
+				KFD_INT_DATA_DEBUG_DOORBELL_MASK)
+#define KFD_DEBUG_TRAP_CODE(sq_int_data)	(((sq_int_data) &	\
+				KFD_INT_DATA_DEBUG_TRAP_CODE_MASK)	\
+				>> KFD_INT_DATA_DEBUG_TRAP_CODE_SHIFT)
+#define KFD_DEBUG_CP_BAD_OP_ECODE_MASK		0x3fffc00
+#define KFD_DEBUG_CP_BAD_OP_ECODE_SHIFT		10
+#define KFD_DEBUG_CP_BAD_OP_ECODE(ctxid0)	(((ctxid0) &		\
+				KFD_DEBUG_CP_BAD_OP_ECODE_MASK)		\
+				>> KFD_DEBUG_CP_BAD_OP_ECODE_SHIFT)
+
+static void event_interrupt_poison_consumption_v9(struct kfd_node *dev,
+				uint16_t pasid, uint16_t client_id)
+{
+	enum amdgpu_ras_block block = 0;
+	uint32_t reset = 0;
+	struct kfd_process *p = kfd_lookup_process_by_pasid(pasid, NULL);
+	enum ras_event_type type = RAS_EVENT_TYPE_POISON_CONSUMPTION;
+	u64 event_id;
+	int old_poison, ret;
+
+	if (!p)
+		return;
+
+	/* all queues of a process will be unmapped in one time */
+	old_poison = atomic_cmpxchg(&p->poison, 0, 1);
+	kfd_unref_process(p);
+	if (old_poison)
+		return;
+
+	switch (client_id) {
+	case SOC15_IH_CLIENTID_SE0SH:
+	case SOC15_IH_CLIENTID_SE1SH:
+	case SOC15_IH_CLIENTID_SE2SH:
+	case SOC15_IH_CLIENTID_SE3SH:
+	case SOC15_IH_CLIENTID_UTCL2:
+		block = AMDGPU_RAS_BLOCK__GFX;
+		if (amdgpu_ip_version(dev->adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3)) {
+			/* driver mode-2 for gfx poison is only supported by
+			 * pmfw 0x00557300 and onwards */
+			if (dev->adev->pm.fw_version < 0x00557300)
+				reset = AMDGPU_RAS_GPU_RESET_MODE1_RESET;
+			else
+				reset = AMDGPU_RAS_GPU_RESET_MODE2_RESET;
+		} else if (amdgpu_ip_version(dev->adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4)) {
+			/* driver mode-2 for gfx poison is only supported by
+			 * pmfw 0x05550C00 and onwards */
+			if (dev->adev->pm.fw_version < 0x05550C00)
+				reset = AMDGPU_RAS_GPU_RESET_MODE1_RESET;
+			else
+				reset = AMDGPU_RAS_GPU_RESET_MODE2_RESET;
+		} else {
+			reset = AMDGPU_RAS_GPU_RESET_MODE2_RESET;
+		}
+		amdgpu_ras_set_err_poison(dev->adev, AMDGPU_RAS_BLOCK__GFX);
+		break;
+	case SOC15_IH_CLIENTID_VMC:
+	case SOC15_IH_CLIENTID_VMC1:
+		block = AMDGPU_RAS_BLOCK__MMHUB;
+		reset = AMDGPU_RAS_GPU_RESET_MODE1_RESET;
+		break;
+	case SOC15_IH_CLIENTID_SDMA0:
+	case SOC15_IH_CLIENTID_SDMA1:
+	case SOC15_IH_CLIENTID_SDMA2:
+	case SOC15_IH_CLIENTID_SDMA3:
+	case SOC15_IH_CLIENTID_SDMA4:
+		block = AMDGPU_RAS_BLOCK__SDMA;
+		if (amdgpu_ip_version(dev->adev, SDMA0_HWIP, 0) == IP_VERSION(4, 4, 2)) {
+			/* driver mode-2 for gfx poison is only supported by
+			 * pmfw 0x00557300 and onwards */
+			if (dev->adev->pm.fw_version < 0x00557300)
+				reset = AMDGPU_RAS_GPU_RESET_MODE1_RESET;
+			else
+				reset = AMDGPU_RAS_GPU_RESET_MODE2_RESET;
+		} else if (amdgpu_ip_version(dev->adev, SDMA0_HWIP, 0) == IP_VERSION(4, 4, 5)) {
+			/* driver mode-2 for gfx poison is only supported by
+			 * pmfw 0x05550C00 and onwards */
+			if (dev->adev->pm.fw_version < 0x05550C00)
+				reset = AMDGPU_RAS_GPU_RESET_MODE1_RESET;
+			else
+				reset = AMDGPU_RAS_GPU_RESET_MODE2_RESET;
+		} else {
+			reset = AMDGPU_RAS_GPU_RESET_MODE2_RESET;
+		}
+		amdgpu_ras_set_err_poison(dev->adev, AMDGPU_RAS_BLOCK__SDMA);
+		break;
+	default:
+		dev_warn(dev->adev->dev,
+			 "client %d does not support poison consumption\n", client_id);
+		return;
+	}
+
+	ret = amdgpu_ras_mark_ras_event(dev->adev, type);
+	if (ret)
+		return;
+
+	kfd_signal_poison_consumed_event(dev, pasid);
+
+	if (amdgpu_uniras_enabled(dev->adev))
+		event_id = amdgpu_ras_mgr_gen_ras_event_seqno(dev->adev,
+					RAS_SEQNO_TYPE_POISON_CONSUMPTION);
+	else
+		event_id = amdgpu_ras_acquire_event_id(dev->adev, type);
+
+	RAS_EVENT_LOG(dev->adev, event_id,
+		      "poison is consumed by client %d, kick off gpu reset flow\n", client_id);
+
+	amdgpu_amdkfd_ras_pasid_poison_consumption_handler(dev->adev,
+		block, pasid, NULL, NULL, reset);
+}
+
+static bool context_id_expected(struct kfd_dev *dev)
+{
+	switch (KFD_GC_VERSION(dev)) {
+	case IP_VERSION(9, 0, 1):
+		return dev->mec_fw_version >= 0x817a;
+	case IP_VERSION(9, 1, 0):
+	case IP_VERSION(9, 2, 1):
+	case IP_VERSION(9, 2, 2):
+	case IP_VERSION(9, 3, 0):
+	case IP_VERSION(9, 4, 0):
+		return dev->mec_fw_version >= 0x17a;
+	default:
+		/* Other GFXv9 and later GPUs always sent valid context IDs
+		 * on legitimate events
+		 */
+		return KFD_GC_VERSION(dev) >= IP_VERSION(9, 4, 1);
+	}
+}
+
+static bool event_interrupt_isr_v9(struct kfd_node *dev,
 					const uint32_t *ih_ring_entry,
 					uint32_t *patched_ihre,
 					bool *patched_flag)
@@ -97,14 +269,16 @@ static bool event_interrupt_isr_v9(struct kfd_dev *dev,
 	uint16_t source_id, client_id, pasid, vmid;
 	const uint32_t *data = ih_ring_entry;
 
+	source_id = SOC15_SOURCE_ID_FROM_IH_ENTRY(ih_ring_entry);
+	client_id = SOC15_CLIENT_ID_FROM_IH_ENTRY(ih_ring_entry);
+
 	/* Only handle interrupts from KFD VMIDs */
 	vmid = SOC15_VMID_FROM_IH_ENTRY(ih_ring_entry);
-	if (vmid < dev->vm_info.first_vmid_kfd ||
-	    vmid > dev->vm_info.last_vmid_kfd)
+	if (!KFD_IRQ_IS_FENCE(client_id, source_id) &&
+	   (vmid < dev->vm_info.first_vmid_kfd ||
+	    vmid > dev->vm_info.last_vmid_kfd))
 		return false;
 
-	source_id = SOC15_SOURCE_ID_FROM_IH_ENTRY(ih_ring_entry);
-	client_id = SOC15_CLIENT_ID_FROM_IH_ENTRY(ih_ring_entry);
 	pasid = SOC15_PASID_FROM_IH_ENTRY(ih_ring_entry);
 
 	/* Only handle clients we care about */
@@ -123,7 +297,8 @@ static bool event_interrupt_isr_v9(struct kfd_dev *dev,
 	    client_id != SOC15_IH_CLIENTID_SE0SH &&
 	    client_id != SOC15_IH_CLIENTID_SE1SH &&
 	    client_id != SOC15_IH_CLIENTID_SE2SH &&
-	    client_id != SOC15_IH_CLIENTID_SE3SH)
+	    client_id != SOC15_IH_CLIENTID_SE3SH &&
+	    !KFD_IRQ_IS_FENCE(client_id, source_id))
 		return false;
 
 	/* This is a known issue for gfx9. Under non HWS, pasid is not set
@@ -135,7 +310,7 @@ static bool event_interrupt_isr_v9(struct kfd_dev *dev,
 
 		*patched_flag = true;
 		memcpy(patched_ihre, ih_ring_entry,
-				dev->device_info->ih_ring_entry_size);
+				dev->kfd->device_info.ih_ring_entry_size);
 
 		pasid = dev->dqm->vmid_pasid[vmid];
 
@@ -144,30 +319,47 @@ static bool event_interrupt_isr_v9(struct kfd_dev *dev,
 					& ~pasid_mask) | pasid);
 	}
 
-	pr_debug("client id 0x%x, source id %d, vmid %d, pasid 0x%x. raw data:\n",
-		 client_id, source_id, vmid, pasid);
-	pr_debug("%8X, %8X, %8X, %8X, %8X, %8X, %8X, %8X.\n",
-		 data[0], data[1], data[2], data[3],
-		 data[4], data[5], data[6], data[7]);
+	dev_dbg(dev->adev->dev,
+		"client id 0x%x, source id %d, vmid %d, pasid 0x%x. raw data:\n",
+		client_id, source_id, vmid, pasid);
+	dev_dbg(dev->adev->dev, "%8X, %8X, %8X, %8X, %8X, %8X, %8X, %8X.\n",
+		data[0], data[1], data[2], data[3], data[4], data[5], data[6],
+		data[7]);
 
 	/* If there is no valid PASID, it's likely a bug */
 	if (WARN_ONCE(pasid == 0, "Bug: No PASID in KFD interrupt"))
 		return false;
 
+	/* Workaround CP firmware sending bogus signals with 0 context_id.
+	 * Those can be safely ignored on hardware and firmware versions that
+	 * include a valid context_id on legitimate signals. This avoids the
+	 * slow path in kfd_signal_event_interrupt that scans all event slots
+	 * for signaled events.
+	 */
+	if (source_id == SOC15_INTSRC_CP_END_OF_PIPE) {
+		uint32_t context_id =
+			SOC15_CONTEXT_ID0_FROM_IH_ENTRY(ih_ring_entry);
+
+		if (context_id == 0 && context_id_expected(dev->kfd))
+			return false;
+	}
+
 	/* Interrupt types we care about: various signals and faults.
 	 * They will be forwarded to a work queue (see below).
 	 */
 	return source_id == SOC15_INTSRC_CP_END_OF_PIPE ||
 		source_id == SOC15_INTSRC_SDMA_TRAP ||
+		source_id == SOC15_INTSRC_SDMA_ECC ||
 		source_id == SOC15_INTSRC_SQ_INTERRUPT_MSG ||
 		source_id == SOC15_INTSRC_CP_BAD_OPCODE ||
+		KFD_IRQ_IS_FENCE(client_id, source_id) ||
 		((client_id == SOC15_IH_CLIENTID_VMC ||
 		client_id == SOC15_IH_CLIENTID_VMC1 ||
 		client_id == SOC15_IH_CLIENTID_UTCL2) &&
 		!amdgpu_no_queue_eviction_on_vm_fault);
 }
 
-static void event_interrupt_wq_v9(struct kfd_dev *dev,
+static void event_interrupt_wq_v9(struct kfd_node *dev,
 					const uint32_t *ih_ring_entry)
 {
 	uint16_t source_id, client_id, pasid, vmid;
@@ -193,54 +385,142 @@ static void event_interrupt_wq_v9(struct kfd_dev *dev,
 			encoding = REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID, ENCODING);
 			switch (encoding) {
 			case SQ_INTERRUPT_WORD_ENCODING_AUTO:
-				pr_debug(
+				dev_dbg_ratelimited(
+					dev->adev->dev,
 					"sq_intr: auto, se %d, ttrace %d, wlt %d, ttrac_buf_full %d, reg_tms %d, cmd_tms %d, host_cmd_ovf %d, host_reg_ovf %d, immed_ovf %d, ttrace_utc_err %d\n",
-					REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID, SE_ID),
-					REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID, THREAD_TRACE),
-					REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID, WLT),
-					REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID, THREAD_TRACE_BUF_FULL),
-					REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID, REG_TIMESTAMP),
-					REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID, CMD_TIMESTAMP),
-					REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID, HOST_CMD_OVERFLOW),
-					REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID, HOST_REG_OVERFLOW),
-					REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID, IMMED_OVERFLOW),
-					REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_AUTO_CTXID, THREAD_TRACE_UTC_ERROR));
+					REG_GET_FIELD(
+						context_id0,
+						SQ_INTERRUPT_WORD_AUTO_CTXID,
+						SE_ID),
+					REG_GET_FIELD(
+						context_id0,
+						SQ_INTERRUPT_WORD_AUTO_CTXID,
+						THREAD_TRACE),
+					REG_GET_FIELD(
+						context_id0,
+						SQ_INTERRUPT_WORD_AUTO_CTXID,
+						WLT),
+					REG_GET_FIELD(
+						context_id0,
+						SQ_INTERRUPT_WORD_AUTO_CTXID,
+						THREAD_TRACE_BUF_FULL),
+					REG_GET_FIELD(
+						context_id0,
+						SQ_INTERRUPT_WORD_AUTO_CTXID,
+						REG_TIMESTAMP),
+					REG_GET_FIELD(
+						context_id0,
+						SQ_INTERRUPT_WORD_AUTO_CTXID,
+						CMD_TIMESTAMP),
+					REG_GET_FIELD(
+						context_id0,
+						SQ_INTERRUPT_WORD_AUTO_CTXID,
+						HOST_CMD_OVERFLOW),
+					REG_GET_FIELD(
+						context_id0,
+						SQ_INTERRUPT_WORD_AUTO_CTXID,
+						HOST_REG_OVERFLOW),
+					REG_GET_FIELD(
+						context_id0,
+						SQ_INTERRUPT_WORD_AUTO_CTXID,
+						IMMED_OVERFLOW),
+					REG_GET_FIELD(
+						context_id0,
+						SQ_INTERRUPT_WORD_AUTO_CTXID,
+						THREAD_TRACE_UTC_ERROR));
 				break;
 			case SQ_INTERRUPT_WORD_ENCODING_INST:
-				pr_debug("sq_intr: inst, se %d, data 0x%x, sh %d, priv %d, wave_id %d, simd_id %d, cu_id %d, intr_data 0x%x\n",
-					REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID, SE_ID),
-					REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID, DATA),
-					REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID, SH_ID),
-					REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID, PRIV),
-					REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID, WAVE_ID),
-					REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID, SIMD_ID),
-					REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID, CU_ID),
+				dev_dbg_ratelimited(
+					dev->adev->dev,
+					"sq_intr: inst, se %d, data 0x%x, sh %d, priv %d, wave_id %d, simd_id %d, cu_id %d, intr_data 0x%x\n",
+					REG_GET_FIELD(
+						context_id0,
+						SQ_INTERRUPT_WORD_WAVE_CTXID,
+						SE_ID),
+					REG_GET_FIELD(
+						context_id0,
+						SQ_INTERRUPT_WORD_WAVE_CTXID,
+						DATA),
+					REG_GET_FIELD(
+						context_id0,
+						SQ_INTERRUPT_WORD_WAVE_CTXID,
+						SH_ID),
+					REG_GET_FIELD(
+						context_id0,
+						SQ_INTERRUPT_WORD_WAVE_CTXID,
+						PRIV),
+					REG_GET_FIELD(
+						context_id0,
+						SQ_INTERRUPT_WORD_WAVE_CTXID,
+						WAVE_ID),
+					REG_GET_FIELD(
+						context_id0,
+						SQ_INTERRUPT_WORD_WAVE_CTXID,
+						SIMD_ID),
+					REG_GET_FIELD(
+						context_id0,
+						SQ_INTERRUPT_WORD_WAVE_CTXID,
+						CU_ID),
 					sq_int_data);
+				if (context_id0 & SQ_INTERRUPT_WORD_WAVE_CTXID__PRIV_MASK) {
+					if (kfd_set_dbg_ev_from_interrupt(dev, pasid,
+							KFD_DEBUG_DOORBELL_ID(sq_int_data),
+							KFD_DEBUG_TRAP_CODE(sq_int_data),
+							NULL, 0))
+						return;
+				}
 				break;
 			case SQ_INTERRUPT_WORD_ENCODING_ERROR:
 				sq_intr_err = REG_GET_FIELD(sq_int_data, KFD_SQ_INT_DATA, ERR_TYPE);
-				pr_warn("sq_intr: error, se %d, data 0x%x, sh %d, priv %d, wave_id %d, simd_id %d, cu_id %d, err_type %d\n",
-					REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID, SE_ID),
-					REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID, DATA),
-					REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID, SH_ID),
-					REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID, PRIV),
-					REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID, WAVE_ID),
-					REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID, SIMD_ID),
-					REG_GET_FIELD(context_id0, SQ_INTERRUPT_WORD_WAVE_CTXID, CU_ID),
+				dev_warn_ratelimited(
+					dev->adev->dev,
+					"sq_intr: error, se %d, data 0x%x, sh %d, priv %d, wave_id %d, simd_id %d, cu_id %d, err_type %d\n",
+					REG_GET_FIELD(
+						context_id0,
+						SQ_INTERRUPT_WORD_WAVE_CTXID,
+						SE_ID),
+					REG_GET_FIELD(
+						context_id0,
+						SQ_INTERRUPT_WORD_WAVE_CTXID,
+						DATA),
+					REG_GET_FIELD(
+						context_id0,
+						SQ_INTERRUPT_WORD_WAVE_CTXID,
+						SH_ID),
+					REG_GET_FIELD(
+						context_id0,
+						SQ_INTERRUPT_WORD_WAVE_CTXID,
+						PRIV),
+					REG_GET_FIELD(
+						context_id0,
+						SQ_INTERRUPT_WORD_WAVE_CTXID,
+						WAVE_ID),
+					REG_GET_FIELD(
+						context_id0,
+						SQ_INTERRUPT_WORD_WAVE_CTXID,
+						SIMD_ID),
+					REG_GET_FIELD(
+						context_id0,
+						SQ_INTERRUPT_WORD_WAVE_CTXID,
+						CU_ID),
 					sq_intr_err);
 				if (sq_intr_err != SQ_INTERRUPT_ERROR_TYPE_ILLEGAL_INST &&
 					sq_intr_err != SQ_INTERRUPT_ERROR_TYPE_MEMVIOL) {
-					kfd_signal_poison_consumed_event(dev, pasid);
-					amdgpu_amdkfd_ras_poison_consumption_handler(dev->kgd);
+					event_interrupt_poison_consumption_v9(dev, pasid, client_id);
 					return;
 				}
 				break;
 			default:
 				break;
 			}
-			kfd_signal_event_interrupt(pasid, context_id0 & 0xffffff, 24);
-		} else if (source_id == SOC15_INTSRC_CP_BAD_OPCODE)
-			kfd_signal_hw_exception_event(pasid);
+			kfd_signal_event_interrupt(pasid, sq_int_data, 24);
+		} else if (source_id == SOC15_INTSRC_CP_BAD_OPCODE &&
+			   KFD_DBG_EC_TYPE_IS_PACKET(KFD_DEBUG_CP_BAD_OP_ECODE(context_id0))) {
+			kfd_set_dbg_ev_from_interrupt(dev, pasid,
+				KFD_DEBUG_DOORBELL_ID(context_id0),
+				KFD_EC_MASK(KFD_DEBUG_CP_BAD_OP_ECODE(context_id0)),
+				NULL, 0);
+		}
 	} else if (client_id == SOC15_IH_CLIENTID_SDMA0 ||
 		   client_id == SOC15_IH_CLIENTID_SDMA1 ||
 		   client_id == SOC15_IH_CLIENTID_SDMA2 ||
@@ -252,8 +532,7 @@ static void event_interrupt_wq_v9(struct kfd_dev *dev,
 		if (source_id == SOC15_INTSRC_SDMA_TRAP) {
 			kfd_signal_event_interrupt(pasid, context_id0 & 0xfffffff, 28);
 		} else if (source_id == SOC15_INTSRC_SDMA_ECC) {
-			kfd_signal_poison_consumed_event(dev, pasid);
-			amdgpu_amdkfd_ras_poison_consumption_handler(dev->kgd);
+			event_interrupt_poison_consumption_v9(dev, pasid, client_id);
 			return;
 		}
 	} else if (client_id == SOC15_IH_CLIENTID_VMC ||
@@ -261,6 +540,12 @@ static void event_interrupt_wq_v9(struct kfd_dev *dev,
 		   client_id == SOC15_IH_CLIENTID_UTCL2) {
 		struct kfd_vm_fault_info info = {0};
 		uint16_t ring_id = SOC15_RING_ID_FROM_IH_ENTRY(ih_ring_entry);
+		struct kfd_hsa_memory_exception_data exception_data;
+
+		if (source_id == SOC15_INTSRC_VMC_UTCL2_POISON) {
+			event_interrupt_poison_consumption_v9(dev, pasid, client_id);
+			return;
+		}
 
 		info.vmid = vmid;
 		info.mc_id = client_id;
@@ -270,13 +555,56 @@ static void event_interrupt_wq_v9(struct kfd_dev *dev,
 		info.prot_read  = ring_id & 0x10;
 		info.prot_write = ring_id & 0x20;
 
+		memset(&exception_data, 0, sizeof(exception_data));
+		exception_data.gpu_id = dev->id;
+		exception_data.va = (info.page_addr) << PAGE_SHIFT;
+		exception_data.failure.NotPresent = info.prot_valid ? 1 : 0;
+		exception_data.failure.NoExecute = info.prot_exec ? 1 : 0;
+		exception_data.failure.ReadOnly = info.prot_write ? 1 : 0;
+		exception_data.failure.imprecise = 0;
+
+		kfd_set_dbg_ev_from_interrupt(dev,
+						pasid,
+						-1,
+						KFD_EC_MASK(EC_DEVICE_MEMORY_VIOLATION),
+						&exception_data,
+						sizeof(exception_data));
 		kfd_smi_event_update_vmfault(dev, pasid);
-		kfd_process_vm_fault(dev->dqm, pasid);
-		kfd_signal_vm_fault_event(dev, pasid, &info);
+	} else if (KFD_IRQ_IS_FENCE(client_id, source_id)) {
+		kfd_process_close_interrupt_drain(pasid);
 	}
 }
 
+static bool event_interrupt_isr_v9_4_3(struct kfd_node *node,
+				const uint32_t *ih_ring_entry,
+				uint32_t *patched_ihre,
+				bool *patched_flag)
+{
+	uint16_t node_id, vmid;
+
+	/*
+	 * For GFX 9.4.3, process the interrupt if:
+	 * - NodeID field in IH entry matches the corresponding bit
+	 *   set in interrupt_bitmap Bits 0-15.
+	 *   OR
+	 * - If partition mode is CPX and interrupt came from
+	 *   Node_id 0,4,8,12, then check if the Bit (16 + client id)
+	 *   is set in interrupt bitmap Bits 16-31.
+	 */
+	node_id = SOC15_NODEID_FROM_IH_ENTRY(ih_ring_entry);
+	vmid = SOC15_VMID_FROM_IH_ENTRY(ih_ring_entry);
+	if (kfd_irq_is_from_node(node, node_id, vmid))
+		return event_interrupt_isr_v9(node, ih_ring_entry,
+					patched_ihre, patched_flag);
+	return false;
+}
+
 const struct kfd_event_interrupt_class event_interrupt_class_v9 = {
 	.interrupt_isr = event_interrupt_isr_v9,
 	.interrupt_wq = event_interrupt_wq_v9,
 };
+
+const struct kfd_event_interrupt_class event_interrupt_class_v9_4_3 = {
+	.interrupt_isr = event_interrupt_isr_v9_4_3,
+	.interrupt_wq = event_interrupt_wq_v9,
+};
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_interrupt.c b/drivers/gpu/drm/amd/amdkfd/kfd_interrupt.c
index bc47f6a44456..783c2f5a04e4 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_interrupt.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_interrupt.c
@@ -1,5 +1,6 @@
+// SPDX-License-Identifier: GPL-2.0 OR MIT
 /*
- * Copyright 2014 Advanced Micro Devices, Inc.
+ * Copyright 2014-2022 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -45,33 +46,36 @@
 #include <linux/kfifo.h>
 #include "kfd_priv.h"
 
-#define KFD_IH_NUM_ENTRIES 8192
+#define KFD_IH_NUM_ENTRIES 16384
 
 static void interrupt_wq(struct work_struct *);
 
-int kfd_interrupt_init(struct kfd_dev *kfd)
+int kfd_interrupt_init(struct kfd_node *node)
 {
 	int r;
 
-	r = kfifo_alloc(&kfd->ih_fifo,
-		KFD_IH_NUM_ENTRIES * kfd->device_info->ih_ring_entry_size,
+	r = kfifo_alloc(&node->ih_fifo,
+		KFD_IH_NUM_ENTRIES * node->kfd->device_info.ih_ring_entry_size,
 		GFP_KERNEL);
 	if (r) {
-		dev_err(kfd_chardev(), "Failed to allocate IH fifo\n");
+		dev_err(node->adev->dev, "Failed to allocate IH fifo\n");
 		return r;
 	}
 
-	kfd->ih_wq = alloc_workqueue("KFD IH", WQ_HIGHPRI, 1);
-	if (unlikely(!kfd->ih_wq)) {
-		kfifo_free(&kfd->ih_fifo);
-		dev_err(kfd_chardev(), "Failed to allocate KFD IH workqueue\n");
-		return -ENOMEM;
+	if (!node->kfd->ih_wq) {
+		node->kfd->ih_wq = alloc_workqueue("KFD IH", WQ_HIGHPRI | WQ_UNBOUND,
+						   node->kfd->num_nodes);
+		if (unlikely(!node->kfd->ih_wq)) {
+			kfifo_free(&node->ih_fifo);
+			dev_err(node->adev->dev, "Failed to allocate KFD IH workqueue\n");
+			return -ENOMEM;
+		}
 	}
-	spin_lock_init(&kfd->interrupt_lock);
+	spin_lock_init(&node->interrupt_lock);
 
-	INIT_WORK(&kfd->interrupt_work, interrupt_wq);
+	INIT_WORK(&node->interrupt_work, interrupt_wq);
 
-	kfd->interrupts_active = true;
+	node->interrupts_active = true;
 
 	/*
 	 * After this function returns, the interrupt will be enabled. This
@@ -83,7 +87,7 @@ int kfd_interrupt_init(struct kfd_dev *kfd)
 	return 0;
 }
 
-void kfd_interrupt_exit(struct kfd_dev *kfd)
+void kfd_interrupt_exit(struct kfd_node *node)
 {
 	/*
 	 * Stop the interrupt handler from writing to the ring and scheduling
@@ -92,78 +96,72 @@ void kfd_interrupt_exit(struct kfd_dev *kfd)
 	 */
 	unsigned long flags;
 
-	spin_lock_irqsave(&kfd->interrupt_lock, flags);
-	kfd->interrupts_active = false;
-	spin_unlock_irqrestore(&kfd->interrupt_lock, flags);
-
-	/*
-	 * flush_work ensures that there are no outstanding
-	 * work-queue items that will access interrupt_ring. New work items
-	 * can't be created because we stopped interrupt handling above.
-	 */
-	flush_workqueue(kfd->ih_wq);
-
-	kfifo_free(&kfd->ih_fifo);
+	spin_lock_irqsave(&node->interrupt_lock, flags);
+	node->interrupts_active = false;
+	spin_unlock_irqrestore(&node->interrupt_lock, flags);
+	kfifo_free(&node->ih_fifo);
 }
 
 /*
  * Assumption: single reader/writer. This function is not re-entrant
  */
-bool enqueue_ih_ring_entry(struct kfd_dev *kfd,	const void *ih_ring_entry)
+bool enqueue_ih_ring_entry(struct kfd_node *node, const void *ih_ring_entry)
 {
-	int count;
-
-	count = kfifo_in(&kfd->ih_fifo, ih_ring_entry,
-				kfd->device_info->ih_ring_entry_size);
-	if (count != kfd->device_info->ih_ring_entry_size) {
-		dev_err_ratelimited(kfd_chardev(),
-			"Interrupt ring overflow, dropping interrupt %d\n",
-			count);
+	if (kfifo_is_full(&node->ih_fifo)) {
+		dev_warn_ratelimited(node->adev->dev, "KFD node %d ih_fifo overflow\n",
+				     node->node_id);
 		return false;
 	}
 
+	kfifo_in(&node->ih_fifo, ih_ring_entry, node->kfd->device_info.ih_ring_entry_size);
 	return true;
 }
 
 /*
  * Assumption: single reader/writer. This function is not re-entrant
  */
-static bool dequeue_ih_ring_entry(struct kfd_dev *kfd, void *ih_ring_entry)
+static bool dequeue_ih_ring_entry(struct kfd_node *node, u32 **ih_ring_entry)
 {
 	int count;
 
-	count = kfifo_out(&kfd->ih_fifo, ih_ring_entry,
-				kfd->device_info->ih_ring_entry_size);
-
-	WARN_ON(count && count != kfd->device_info->ih_ring_entry_size);
+	if (kfifo_is_empty(&node->ih_fifo))
+		return false;
 
-	return count == kfd->device_info->ih_ring_entry_size;
+	count = kfifo_out_linear_ptr(&node->ih_fifo, ih_ring_entry,
+				     node->kfd->device_info.ih_ring_entry_size);
+	WARN_ON(count != node->kfd->device_info.ih_ring_entry_size);
+	return count == node->kfd->device_info.ih_ring_entry_size;
 }
 
 static void interrupt_wq(struct work_struct *work)
 {
-	struct kfd_dev *dev = container_of(work, struct kfd_dev,
-						interrupt_work);
-	uint32_t ih_ring_entry[KFD_MAX_RING_ENTRY_SIZE];
+	struct kfd_node *dev = container_of(work, struct kfd_node, interrupt_work);
+	uint32_t *ih_ring_entry;
+	unsigned long start_jiffies = jiffies;
 
-	if (dev->device_info->ih_ring_entry_size > sizeof(ih_ring_entry)) {
-		dev_err_once(kfd_chardev(), "Ring entry too small\n");
-		return;
-	}
-
-	while (dequeue_ih_ring_entry(dev, ih_ring_entry))
-		dev->device_info->event_interrupt_class->interrupt_wq(dev,
+	while (dequeue_ih_ring_entry(dev, &ih_ring_entry)) {
+		dev->kfd->device_info.event_interrupt_class->interrupt_wq(dev,
 								ih_ring_entry);
+		kfifo_skip_count(&dev->ih_fifo, dev->kfd->device_info.ih_ring_entry_size);
+
+		if (time_is_before_jiffies(start_jiffies + HZ)) {
+			/* If we spent more than a second processing signals,
+			 * reschedule the worker to avoid soft-lockup warnings
+			 */
+			queue_work(dev->kfd->ih_wq, &dev->interrupt_work);
+			break;
+		}
+	}
 }
 
-bool interrupt_is_wanted(struct kfd_dev *dev,
+bool interrupt_is_wanted(struct kfd_node *dev,
 			const uint32_t *ih_ring_entry,
 			uint32_t *patched_ihre, bool *flag)
 {
 	/* integer and bitwise OR so there is no boolean short-circuiting */
 	unsigned int wanted = 0;
 
-	wanted |= dev->device_info->event_interrupt_class->interrupt_isr(dev,
+	wanted |= dev->kfd->device_info.event_interrupt_class->interrupt_isr(dev,
 					 ih_ring_entry, patched_ihre, flag);
 
 	return wanted != 0;
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_iommu.c b/drivers/gpu/drm/amd/amdkfd/kfd_iommu.c
deleted file mode 100644
index 73f2257acc23..000000000000
--- a/drivers/gpu/drm/amd/amdkfd/kfd_iommu.c
+++ /dev/null
@@ -1,359 +0,0 @@
-/*
- * Copyright 2018 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- */
-
-#include <linux/kconfig.h>
-
-#if IS_REACHABLE(CONFIG_AMD_IOMMU_V2)
-
-#include <linux/printk.h>
-#include <linux/device.h>
-#include <linux/slab.h>
-#include <linux/pci.h>
-#include <linux/amd-iommu.h>
-#include "kfd_priv.h"
-#include "kfd_dbgmgr.h"
-#include "kfd_topology.h"
-#include "kfd_iommu.h"
-
-static const u32 required_iommu_flags = AMD_IOMMU_DEVICE_FLAG_ATS_SUP |
-					AMD_IOMMU_DEVICE_FLAG_PRI_SUP |
-					AMD_IOMMU_DEVICE_FLAG_PASID_SUP;
-
-/** kfd_iommu_check_device - Check whether IOMMU is available for device
- */
-int kfd_iommu_check_device(struct kfd_dev *kfd)
-{
-	struct amd_iommu_device_info iommu_info;
-	int err;
-
-	if (!kfd->use_iommu_v2)
-		return -ENODEV;
-
-	iommu_info.flags = 0;
-	err = amd_iommu_device_info(kfd->pdev, &iommu_info);
-	if (err)
-		return err;
-
-	if ((iommu_info.flags & required_iommu_flags) != required_iommu_flags)
-		return -ENODEV;
-
-	return 0;
-}
-
-/** kfd_iommu_device_init - Initialize IOMMU for device
- */
-int kfd_iommu_device_init(struct kfd_dev *kfd)
-{
-	struct amd_iommu_device_info iommu_info;
-	unsigned int pasid_limit;
-	int err;
-
-	if (!kfd->use_iommu_v2)
-		return 0;
-
-	iommu_info.flags = 0;
-	err = amd_iommu_device_info(kfd->pdev, &iommu_info);
-	if (err < 0) {
-		dev_err(kfd_device,
-			"error getting iommu info. is the iommu enabled?\n");
-		return -ENODEV;
-	}
-
-	if ((iommu_info.flags & required_iommu_flags) != required_iommu_flags) {
-		dev_err(kfd_device,
-			"error required iommu flags ats %i, pri %i, pasid %i\n",
-		       (iommu_info.flags & AMD_IOMMU_DEVICE_FLAG_ATS_SUP) != 0,
-		       (iommu_info.flags & AMD_IOMMU_DEVICE_FLAG_PRI_SUP) != 0,
-		       (iommu_info.flags & AMD_IOMMU_DEVICE_FLAG_PASID_SUP)
-									!= 0);
-		return -ENODEV;
-	}
-
-	pasid_limit = min_t(unsigned int,
-			(unsigned int)(1 << kfd->device_info->max_pasid_bits),
-			iommu_info.max_pasids);
-
-	if (!kfd_set_pasid_limit(pasid_limit)) {
-		dev_err(kfd_device, "error setting pasid limit\n");
-		return -EBUSY;
-	}
-
-	return 0;
-}
-
-/** kfd_iommu_bind_process_to_device - Have the IOMMU bind a process
- *
- * Binds the given process to the given device using its PASID. This
- * enables IOMMUv2 address translation for the process on the device.
- *
- * This function assumes that the process mutex is held.
- */
-int kfd_iommu_bind_process_to_device(struct kfd_process_device *pdd)
-{
-	struct kfd_dev *dev = pdd->dev;
-	struct kfd_process *p = pdd->process;
-	int err;
-
-	if (!dev->use_iommu_v2 || pdd->bound == PDD_BOUND)
-		return 0;
-
-	if (unlikely(pdd->bound == PDD_BOUND_SUSPENDED)) {
-		pr_err("Binding PDD_BOUND_SUSPENDED pdd is unexpected!\n");
-		return -EINVAL;
-	}
-
-	err = amd_iommu_bind_pasid(dev->pdev, p->pasid, p->lead_thread);
-	if (!err)
-		pdd->bound = PDD_BOUND;
-
-	return err;
-}
-
-/** kfd_iommu_unbind_process - Unbind process from all devices
- *
- * This removes all IOMMU device bindings of the process. To be used
- * before process termination.
- */
-void kfd_iommu_unbind_process(struct kfd_process *p)
-{
-	int i;
-
-	for (i = 0; i < p->n_pdds; i++)
-		if (p->pdds[i]->bound == PDD_BOUND)
-			amd_iommu_unbind_pasid(p->pdds[i]->dev->pdev, p->pasid);
-}
-
-/* Callback for process shutdown invoked by the IOMMU driver */
-static void iommu_pasid_shutdown_callback(struct pci_dev *pdev, u32 pasid)
-{
-	struct kfd_dev *dev = kfd_device_by_pci_dev(pdev);
-	struct kfd_process *p;
-	struct kfd_process_device *pdd;
-
-	if (!dev)
-		return;
-
-	/*
-	 * Look for the process that matches the pasid. If there is no such
-	 * process, we either released it in amdkfd's own notifier, or there
-	 * is a bug. Unfortunately, there is no way to tell...
-	 */
-	p = kfd_lookup_process_by_pasid(pasid);
-	if (!p)
-		return;
-
-	pr_debug("Unbinding process 0x%x from IOMMU\n", pasid);
-
-	mutex_lock(kfd_get_dbgmgr_mutex());
-
-	if (dev->dbgmgr && dev->dbgmgr->pasid == p->pasid) {
-		if (!kfd_dbgmgr_unregister(dev->dbgmgr, p)) {
-			kfd_dbgmgr_destroy(dev->dbgmgr);
-			dev->dbgmgr = NULL;
-		}
-	}
-
-	mutex_unlock(kfd_get_dbgmgr_mutex());
-
-	mutex_lock(&p->mutex);
-
-	pdd = kfd_get_process_device_data(dev, p);
-	if (pdd)
-		/* For GPU relying on IOMMU, we need to dequeue here
-		 * when PASID is still bound.
-		 */
-		kfd_process_dequeue_from_device(pdd);
-
-	mutex_unlock(&p->mutex);
-
-	kfd_unref_process(p);
-}
-
-/* This function called by IOMMU driver on PPR failure */
-static int iommu_invalid_ppr_cb(struct pci_dev *pdev, u32 pasid,
-				unsigned long address, u16 flags)
-{
-	struct kfd_dev *dev;
-
-	dev_warn_ratelimited(kfd_device,
-			"Invalid PPR device %x:%x.%x pasid 0x%x address 0x%lX flags 0x%X",
-			pdev->bus->number,
-			PCI_SLOT(pdev->devfn),
-			PCI_FUNC(pdev->devfn),
-			pasid,
-			address,
-			flags);
-
-	dev = kfd_device_by_pci_dev(pdev);
-	if (!WARN_ON(!dev))
-		kfd_signal_iommu_event(dev, pasid, address,
-			flags & PPR_FAULT_WRITE, flags & PPR_FAULT_EXEC);
-
-	return AMD_IOMMU_INV_PRI_RSP_INVALID;
-}
-
-/*
- * Bind processes do the device that have been temporarily unbound
- * (PDD_BOUND_SUSPENDED) in kfd_unbind_processes_from_device.
- */
-static int kfd_bind_processes_to_device(struct kfd_dev *kfd)
-{
-	struct kfd_process_device *pdd;
-	struct kfd_process *p;
-	unsigned int temp;
-	int err = 0;
-
-	int idx = srcu_read_lock(&kfd_processes_srcu);
-
-	hash_for_each_rcu(kfd_processes_table, temp, p, kfd_processes) {
-		mutex_lock(&p->mutex);
-		pdd = kfd_get_process_device_data(kfd, p);
-
-		if (WARN_ON(!pdd) || pdd->bound != PDD_BOUND_SUSPENDED) {
-			mutex_unlock(&p->mutex);
-			continue;
-		}
-
-		err = amd_iommu_bind_pasid(kfd->pdev, p->pasid,
-				p->lead_thread);
-		if (err < 0) {
-			pr_err("Unexpected pasid 0x%x binding failure\n",
-					p->pasid);
-			mutex_unlock(&p->mutex);
-			break;
-		}
-
-		pdd->bound = PDD_BOUND;
-		mutex_unlock(&p->mutex);
-	}
-
-	srcu_read_unlock(&kfd_processes_srcu, idx);
-
-	return err;
-}
-
-/*
- * Mark currently bound processes as PDD_BOUND_SUSPENDED. These
- * processes will be restored to PDD_BOUND state in
- * kfd_bind_processes_to_device.
- */
-static void kfd_unbind_processes_from_device(struct kfd_dev *kfd)
-{
-	struct kfd_process_device *pdd;
-	struct kfd_process *p;
-	unsigned int temp;
-
-	int idx = srcu_read_lock(&kfd_processes_srcu);
-
-	hash_for_each_rcu(kfd_processes_table, temp, p, kfd_processes) {
-		mutex_lock(&p->mutex);
-		pdd = kfd_get_process_device_data(kfd, p);
-
-		if (WARN_ON(!pdd)) {
-			mutex_unlock(&p->mutex);
-			continue;
-		}
-
-		if (pdd->bound == PDD_BOUND)
-			pdd->bound = PDD_BOUND_SUSPENDED;
-		mutex_unlock(&p->mutex);
-	}
-
-	srcu_read_unlock(&kfd_processes_srcu, idx);
-}
-
-/** kfd_iommu_suspend - Prepare IOMMU for suspend
- *
- * This unbinds processes from the device and disables the IOMMU for
- * the device.
- */
-void kfd_iommu_suspend(struct kfd_dev *kfd)
-{
-	if (!kfd->use_iommu_v2)
-		return;
-
-	kfd_unbind_processes_from_device(kfd);
-
-	amd_iommu_set_invalidate_ctx_cb(kfd->pdev, NULL);
-	amd_iommu_set_invalid_ppr_cb(kfd->pdev, NULL);
-	amd_iommu_free_device(kfd->pdev);
-}
-
-/** kfd_iommu_resume - Restore IOMMU after resume
- *
- * This reinitializes the IOMMU for the device and re-binds previously
- * suspended processes to the device.
- */
-int kfd_iommu_resume(struct kfd_dev *kfd)
-{
-	unsigned int pasid_limit;
-	int err;
-
-	if (!kfd->use_iommu_v2)
-		return 0;
-
-	pasid_limit = kfd_get_pasid_limit();
-
-	err = amd_iommu_init_device(kfd->pdev, pasid_limit);
-	if (err)
-		return -ENXIO;
-
-	amd_iommu_set_invalidate_ctx_cb(kfd->pdev,
-					iommu_pasid_shutdown_callback);
-	amd_iommu_set_invalid_ppr_cb(kfd->pdev,
-				     iommu_invalid_ppr_cb);
-
-	err = kfd_bind_processes_to_device(kfd);
-	if (err) {
-		amd_iommu_set_invalidate_ctx_cb(kfd->pdev, NULL);
-		amd_iommu_set_invalid_ppr_cb(kfd->pdev, NULL);
-		amd_iommu_free_device(kfd->pdev);
-		return err;
-	}
-
-	return 0;
-}
-
-/** kfd_iommu_add_perf_counters - Add IOMMU performance counters to topology
- */
-int kfd_iommu_add_perf_counters(struct kfd_topology_device *kdev)
-{
-	struct kfd_perf_properties *props;
-
-	if (!(kdev->node_props.capability & HSA_CAP_ATS_PRESENT))
-		return 0;
-
-	if (!amd_iommu_pc_supported())
-		return 0;
-
-	props = kfd_alloc_struct(props);
-	if (!props)
-		return -ENOMEM;
-	strcpy(props->block_name, "iommu");
-	props->max_concurrent = amd_iommu_pc_get_max_banks(0) *
-		amd_iommu_pc_get_max_counters(0); /* assume one iommu */
-	list_add_tail(&props->list, &kdev->perf_props);
-
-	return 0;
-}
-
-#endif
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_iommu.h b/drivers/gpu/drm/amd/amdkfd/kfd_iommu.h
deleted file mode 100644
index afd420b01a0c..000000000000
--- a/drivers/gpu/drm/amd/amdkfd/kfd_iommu.h
+++ /dev/null
@@ -1,83 +0,0 @@
-/*
- * Copyright 2018 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- */
-
-#ifndef __KFD_IOMMU_H__
-#define __KFD_IOMMU_H__
-
-#include <linux/kconfig.h>
-
-#if IS_REACHABLE(CONFIG_AMD_IOMMU_V2)
-
-#define KFD_SUPPORT_IOMMU_V2
-
-int kfd_iommu_check_device(struct kfd_dev *kfd);
-int kfd_iommu_device_init(struct kfd_dev *kfd);
-
-int kfd_iommu_bind_process_to_device(struct kfd_process_device *pdd);
-void kfd_iommu_unbind_process(struct kfd_process *p);
-
-void kfd_iommu_suspend(struct kfd_dev *kfd);
-int kfd_iommu_resume(struct kfd_dev *kfd);
-
-int kfd_iommu_add_perf_counters(struct kfd_topology_device *kdev);
-
-#else
-
-static inline int kfd_iommu_check_device(struct kfd_dev *kfd)
-{
-	return -ENODEV;
-}
-static inline int kfd_iommu_device_init(struct kfd_dev *kfd)
-{
-#if IS_MODULE(CONFIG_AMD_IOMMU_V2)
-	WARN_ONCE(1, "iommu_v2 module is not usable by built-in KFD");
-#endif
-	return 0;
-}
-
-static inline int kfd_iommu_bind_process_to_device(
-	struct kfd_process_device *pdd)
-{
-	return 0;
-}
-static inline void kfd_iommu_unbind_process(struct kfd_process *p)
-{
-	/* empty */
-}
-
-static inline void kfd_iommu_suspend(struct kfd_dev *kfd)
-{
-	/* empty */
-}
-static inline int kfd_iommu_resume(struct kfd_dev *kfd)
-{
-	return 0;
-}
-
-static inline int kfd_iommu_add_perf_counters(struct kfd_topology_device *kdev)
-{
-	return 0;
-}
-
-#endif /* IS_REACHABLE(CONFIG_AMD_IOMMU_V2) */
-
-#endif /* __KFD_IOMMU_H__ */
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_kernel_queue.c b/drivers/gpu/drm/amd/amdkfd/kfd_kernel_queue.c
index 64b4ac339904..fb3129883a4c 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_kernel_queue.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_kernel_queue.c
@@ -1,5 +1,6 @@
+// SPDX-License-Identifier: GPL-2.0 OR MIT
 /*
- * Copyright 2014 Advanced Micro Devices, Inc.
+ * Copyright 2014-2022 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -31,24 +32,21 @@
 #include "kfd_device_queue_manager.h"
 #include "kfd_pm4_headers.h"
 #include "kfd_pm4_opcodes.h"
+#include "amdgpu_reset.h"
 
 #define PM4_COUNT_ZERO (((1 << 15) - 1) << 16)
 
 /* Initialize a kernel queue, including allocations of GART memory
  * needed for the queue.
  */
-static bool kq_initialize(struct kernel_queue *kq, struct kfd_dev *dev,
+static bool kq_initialize(struct kernel_queue *kq, struct kfd_node *dev,
 		enum kfd_queue_type type, unsigned int queue_size)
 {
 	struct queue_properties prop;
 	int retval;
 	union PM4_MES_TYPE_3_HEADER nop;
 
-	if (WARN_ON(type != KFD_QUEUE_TYPE_DIQ && type != KFD_QUEUE_TYPE_HIQ))
-		return false;
-
-	pr_debug("Initializing queue type %d size %d\n", KFD_QUEUE_TYPE_HIQ,
-			queue_size);
+	pr_debug("Initializing queue type %d size %d\n", type, queue_size);
 
 	memset(&prop, 0, sizeof(prop));
 	memset(&nop, 0, sizeof(nop));
@@ -67,23 +65,25 @@ static bool kq_initialize(struct kernel_queue *kq, struct kfd_dev *dev,
 		kq->mqd_mgr = dev->dqm->mqd_mgrs[KFD_MQD_TYPE_HIQ];
 		break;
 	default:
-		pr_err("Invalid queue type %d\n", type);
+		WARN(1, "Invalid queue type %d\n", type);
+		dev_err(dev->adev->dev, "Invalid queue type %d\n", type);
 		return false;
 	}
 
 	if (!kq->mqd_mgr)
 		return false;
 
-	prop.doorbell_ptr = kfd_get_kernel_doorbell(dev, &prop.doorbell_off);
+	prop.doorbell_ptr = kfd_get_kernel_doorbell(dev->kfd, &prop.doorbell_off);
 
 	if (!prop.doorbell_ptr) {
-		pr_err("Failed to initialize doorbell");
+		dev_err(dev->adev->dev, "Failed to initialize doorbell");
 		goto err_get_kernel_doorbell;
 	}
 
 	retval = kfd_gtt_sa_allocate(dev, queue_size, &kq->pq);
 	if (retval != 0) {
-		pr_err("Failed to init pq queues size %d\n", queue_size);
+		dev_err(dev->adev->dev, "Failed to init pq queues size %d\n",
+			queue_size);
 		goto err_pq_allocate_vidmem;
 	}
 
@@ -91,7 +91,7 @@ static bool kq_initialize(struct kernel_queue *kq, struct kfd_dev *dev,
 	kq->pq_gpu_addr = kq->pq->gpu_addr;
 
 	/* For CIK family asics, kq->eop_mem is not needed */
-	if (dev->device_info->asic_family > CHIP_MULLINS) {
+	if (dev->adev->asic_type > CHIP_MULLINS) {
 		retval = kfd_gtt_sa_allocate(dev, PAGE_SIZE, &kq->eop_mem);
 		if (retval != 0)
 			goto err_eop_allocate_vidmem;
@@ -111,7 +111,7 @@ static bool kq_initialize(struct kernel_queue *kq, struct kfd_dev *dev,
 	kq->rptr_kernel = kq->rptr_mem->cpu_ptr;
 	kq->rptr_gpu_addr = kq->rptr_mem->gpu_addr;
 
-	retval = kfd_gtt_sa_allocate(dev, dev->device_info->doorbell_size,
+	retval = kfd_gtt_sa_allocate(dev, dev->kfd->device_info.doorbell_size,
 					&kq->wptr_mem);
 
 	if (retval != 0)
@@ -122,7 +122,7 @@ static bool kq_initialize(struct kernel_queue *kq, struct kfd_dev *dev,
 
 	memset(kq->pq_kernel_addr, 0, queue_size);
 	memset(kq->rptr_kernel, 0, sizeof(*kq->rptr_kernel));
-	memset(kq->wptr_kernel, 0, sizeof(*kq->wptr_kernel));
+	memset(kq->wptr_kernel, 0, dev->kfd->device_info.doorbell_size);
 
 	prop.queue_size = queue_size;
 	prop.is_interop = false;
@@ -188,22 +188,24 @@ err_rptr_allocate_vidmem:
 err_eop_allocate_vidmem:
 	kfd_gtt_sa_free(dev, kq->pq);
 err_pq_allocate_vidmem:
-	kfd_release_kernel_doorbell(dev, prop.doorbell_ptr);
+	kfd_release_kernel_doorbell(dev->kfd, prop.doorbell_ptr);
 err_get_kernel_doorbell:
 	return false;
 
 }
 
 /* Uninitialize a kernel queue and free all its memory usages. */
-static void kq_uninitialize(struct kernel_queue *kq, bool hanging)
+static void kq_uninitialize(struct kernel_queue *kq)
 {
-	if (kq->queue->properties.type == KFD_QUEUE_TYPE_HIQ && !hanging)
+	if (kq->queue->properties.type == KFD_QUEUE_TYPE_HIQ && down_read_trylock(&kq->dev->adev->reset_domain->sem)) {
 		kq->mqd_mgr->destroy_mqd(kq->mqd_mgr,
 					kq->queue->mqd,
 					KFD_PREEMPT_TYPE_WAVEFRONT_RESET,
 					KFD_UNMAP_LATENCY_MS,
 					kq->queue->pipe,
 					kq->queue->queue);
+		up_read(&kq->dev->adev->reset_domain->sem);
+	}
 	else if (kq->queue->properties.type == KFD_QUEUE_TYPE_DIQ)
 		kfd_gtt_sa_free(kq->dev, kq->fence_mem_obj);
 
@@ -219,7 +221,7 @@ static void kq_uninitialize(struct kernel_queue *kq, bool hanging)
 	kfd_gtt_sa_free(kq->dev, kq->eop_mem);
 
 	kfd_gtt_sa_free(kq->dev, kq->pq);
-	kfd_release_kernel_doorbell(kq->dev,
+	kfd_release_kernel_doorbell(kq->dev->kfd,
 					kq->queue->properties.doorbell_ptr);
 	uninit_queue(kq->queue);
 }
@@ -285,7 +287,7 @@ err_no_space:
 	return -ENOMEM;
 }
 
-void kq_submit_packet(struct kernel_queue *kq)
+int kq_submit_packet(struct kernel_queue *kq)
 {
 #ifdef DEBUG
 	int i;
@@ -297,20 +299,31 @@ void kq_submit_packet(struct kernel_queue *kq)
 	}
 	pr_debug("\n");
 #endif
-	if (kq->dev->device_info->doorbell_size == 8) {
+	/* Fatal err detected, packet submission won't go through */
+	if (amdgpu_amdkfd_is_fed(kq->dev->adev))
+		return -EIO;
+
+	/* Make sure ring buffer is updated before wptr updated */
+	mb();
+
+	if (kq->dev->kfd->device_info.doorbell_size == 8) {
 		*kq->wptr64_kernel = kq->pending_wptr64;
+		mb(); /* Make sure wptr updated before ring doorbell */
 		write_kernel_doorbell64(kq->queue->properties.doorbell_ptr,
 					kq->pending_wptr64);
 	} else {
 		*kq->wptr_kernel = kq->pending_wptr;
+		mb(); /* Make sure wptr updated before ring doorbell */
 		write_kernel_doorbell(kq->queue->properties.doorbell_ptr,
 					kq->pending_wptr);
 	}
+
+	return 0;
 }
 
 void kq_rollback_packet(struct kernel_queue *kq)
 {
-	if (kq->dev->device_info->doorbell_size == 8) {
+	if (kq->dev->kfd->device_info.doorbell_size == 8) {
 		kq->pending_wptr64 = *kq->wptr64_kernel;
 		kq->pending_wptr = *kq->wptr_kernel %
 			(kq->queue->properties.queue_size / 4);
@@ -319,7 +332,7 @@ void kq_rollback_packet(struct kernel_queue *kq)
 	}
 }
 
-struct kernel_queue *kernel_queue_init(struct kfd_dev *dev,
+struct kernel_queue *kernel_queue_init(struct kfd_node *dev,
 					enum kfd_queue_type type)
 {
 	struct kernel_queue *kq;
@@ -331,45 +344,45 @@ struct kernel_queue *kernel_queue_init(struct kfd_dev *dev,
 	if (kq_initialize(kq, dev, type, KFD_KERNEL_QUEUE_SIZE))
 		return kq;
 
-	pr_err("Failed to init kernel queue\n");
+	dev_err(dev->adev->dev, "Failed to init kernel queue\n");
 
 	kfree(kq);
 	return NULL;
 }
 
-void kernel_queue_uninit(struct kernel_queue *kq, bool hanging)
+void kernel_queue_uninit(struct kernel_queue *kq)
 {
-	kq_uninitialize(kq, hanging);
+	kq_uninitialize(kq);
 	kfree(kq);
 }
 
 /* FIXME: Can this test be removed? */
-static __attribute__((unused)) void test_kq(struct kfd_dev *dev)
+static __attribute__((unused)) void test_kq(struct kfd_node *dev)
 {
 	struct kernel_queue *kq;
 	uint32_t *buffer, i;
 	int retval;
 
-	pr_err("Starting kernel queue test\n");
+	dev_err(dev->adev->dev, "Starting kernel queue test\n");
 
 	kq = kernel_queue_init(dev, KFD_QUEUE_TYPE_HIQ);
 	if (unlikely(!kq)) {
-		pr_err("  Failed to initialize HIQ\n");
-		pr_err("Kernel queue test failed\n");
+		dev_err(dev->adev->dev, "  Failed to initialize HIQ\n");
+		dev_err(dev->adev->dev, "Kernel queue test failed\n");
 		return;
 	}
 
 	retval = kq_acquire_packet_buffer(kq, 5, &buffer);
 	if (unlikely(retval != 0)) {
-		pr_err("  Failed to acquire packet buffer\n");
-		pr_err("Kernel queue test failed\n");
+		dev_err(dev->adev->dev, "  Failed to acquire packet buffer\n");
+		dev_err(dev->adev->dev, "Kernel queue test failed\n");
 		return;
 	}
 	for (i = 0; i < 5; i++)
 		buffer[i] = kq->nop_packet;
 	kq_submit_packet(kq);
 
-	pr_err("Ending kernel queue test\n");
+	dev_err(dev->adev->dev, "Ending kernel queue test\n");
 }
 
 
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_kernel_queue.h b/drivers/gpu/drm/amd/amdkfd/kfd_kernel_queue.h
index f4cfe9f1871c..e24ee50acdf0 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_kernel_queue.h
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_kernel_queue.h
@@ -1,5 +1,6 @@
+/* SPDX-License-Identifier: GPL-2.0 OR MIT */
 /*
- * Copyright 2014 Advanced Micro Devices, Inc.
+ * Copyright 2014-2022 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -46,13 +47,13 @@
 int kq_acquire_packet_buffer(struct kernel_queue *kq,
 				size_t packet_size_in_dwords,
 				unsigned int **buffer_ptr);
-void kq_submit_packet(struct kernel_queue *kq);
+int kq_submit_packet(struct kernel_queue *kq);
 void kq_rollback_packet(struct kernel_queue *kq);
 
 
 struct kernel_queue {
 	/* data */
-	struct kfd_dev		*dev;
+	struct kfd_node		*dev;
 	struct mqd_manager	*mqd_mgr;
 	struct queue		*queue;
 	uint64_t		pending_wptr64;
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_migrate.c b/drivers/gpu/drm/amd/amdkfd/kfd_migrate.c
index 9b9c2b9bf2ef..af53e796ea1b 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_migrate.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_migrate.c
@@ -21,39 +21,39 @@
  * OTHER DEALINGS IN THE SOFTWARE.
  */
 #include <linux/types.h>
-#include <linux/hmm.h>
 #include <linux/dma-direction.h>
 #include <linux/dma-mapping.h>
+#include <linux/migrate.h>
 #include "amdgpu_sync.h"
 #include "amdgpu_object.h"
 #include "amdgpu_vm.h"
-#include "amdgpu_mn.h"
 #include "amdgpu_res_cursor.h"
 #include "kfd_priv.h"
 #include "kfd_svm.h"
 #include "kfd_migrate.h"
+#include "kfd_smi_events.h"
 
 #ifdef dev_fmt
 #undef dev_fmt
 #endif
-#define dev_fmt(fmt) "kfd_migrate: %s: " fmt, __func__
+#define dev_fmt(fmt) "kfd_migrate: " fmt
 
-static uint64_t
-svm_migrate_direct_mapping_addr(struct amdgpu_device *adev, uint64_t addr)
+static u64
+svm_migrate_direct_mapping_addr(struct amdgpu_device *adev, u64 addr)
 {
 	return addr + amdgpu_ttm_domain_start(adev, TTM_PL_VRAM);
 }
 
 static int
-svm_migrate_gart_map(struct amdgpu_ring *ring, uint64_t npages,
-		     dma_addr_t *addr, uint64_t *gart_addr, uint64_t flags)
+svm_migrate_gart_map(struct amdgpu_ring *ring, u64 npages,
+		     dma_addr_t *addr, u64 *gart_addr, u64 flags)
 {
 	struct amdgpu_device *adev = ring->adev;
 	struct amdgpu_job *job;
 	unsigned int num_dw, num_bytes;
 	struct dma_fence *fence;
-	uint64_t src_addr, dst_addr;
-	uint64_t pte_flags;
+	u64 src_addr, dst_addr;
+	u64 pte_flags;
 	void *cpu_addr;
 	int r;
 
@@ -63,8 +63,12 @@ svm_migrate_gart_map(struct amdgpu_ring *ring, uint64_t npages,
 	num_dw = ALIGN(adev->mman.buffer_funcs->copy_num_dw, 8);
 	num_bytes = npages * 8;
 
-	r = amdgpu_job_alloc_with_ib(adev, num_dw * 4 + num_bytes,
-				     AMDGPU_IB_POOL_DELAYED, &job);
+	r = amdgpu_job_alloc_with_ib(adev, &adev->mman.high_pr,
+				     AMDGPU_FENCE_OWNER_UNDEFINED,
+				     num_dw * 4 + num_bytes,
+				     AMDGPU_IB_POOL_DELAYED,
+				     &job,
+				     AMDGPU_KERNEL_JOB_ID_KFD_GART_MAP);
 	if (r)
 		return r;
 
@@ -73,7 +77,7 @@ svm_migrate_gart_map(struct amdgpu_ring *ring, uint64_t npages,
 
 	dst_addr = amdgpu_bo_gpu_offset(adev->gart.bo);
 	amdgpu_emit_copy_buffer(adev, &job->ibs[0], src_addr,
-				dst_addr, num_bytes, false);
+				dst_addr, num_bytes, 0);
 
 	amdgpu_ring_pad_ib(ring, &job->ibs[0]);
 	WARN_ON(job->ibs[0].length_dw > num_dw);
@@ -86,30 +90,19 @@ svm_migrate_gart_map(struct amdgpu_ring *ring, uint64_t npages,
 
 	cpu_addr = &job->ibs[0].ptr[num_dw];
 
-	r = amdgpu_gart_map(adev, 0, npages, addr, pte_flags, cpu_addr);
-	if (r)
-		goto error_free;
-
-	r = amdgpu_job_submit(job, &adev->mman.entity,
-			      AMDGPU_FENCE_OWNER_UNDEFINED, &fence);
-	if (r)
-		goto error_free;
-
+	amdgpu_gart_map(adev, 0, npages, addr, pte_flags, cpu_addr);
+	fence = amdgpu_job_submit(job);
 	dma_fence_put(fence);
 
 	return r;
-
-error_free:
-	amdgpu_job_free(job);
-	return r;
 }
 
 /**
  * svm_migrate_copy_memory_gart - sdma copy data between ram and vram
  *
  * @adev: amdgpu device the sdma ring running
- * @src: source page address array
- * @dst: destination page address array
+ * @sys: system DMA pointer to be copied
+ * @vram: vram destination DMA pointer
  * @npages: number of pages to copy
  * @direction: enum MIGRATION_COPY_DIR
  * @mfence: output, sdma fence to signal after sdma is done
@@ -129,15 +122,15 @@ error_free:
 
 static int
 svm_migrate_copy_memory_gart(struct amdgpu_device *adev, dma_addr_t *sys,
-			     uint64_t *vram, uint64_t npages,
+			     u64 *vram, u64 npages,
 			     enum MIGRATION_COPY_DIR direction,
 			     struct dma_fence **mfence)
 {
-	const uint64_t GTT_MAX_PAGES = AMDGPU_GTT_MAX_TRANSFER_SIZE;
+	const u64 GTT_MAX_PAGES = AMDGPU_GTT_MAX_TRANSFER_SIZE;
 	struct amdgpu_ring *ring = adev->mman.buffer_funcs_ring;
-	uint64_t gart_s, gart_d;
+	u64 gart_s, gart_d;
 	struct dma_fence *next;
-	uint64_t size;
+	u64 size;
 	int r;
 
 	mutex_lock(&adev->mman.gtt_window_lock);
@@ -160,7 +153,7 @@ svm_migrate_copy_memory_gart(struct amdgpu_device *adev, dma_addr_t *sys,
 		}
 
 		r = amdgpu_copy_buffer(ring, gart_s, gart_d, size * PAGE_SIZE,
-				       NULL, &next, false, true, false);
+				       NULL, &next, false, true, 0);
 		if (r) {
 			dev_err(adev->dev, "fail %d to copy memory\n", r);
 			goto out_unlock;
@@ -213,7 +206,7 @@ svm_migrate_copy_done(struct amdgpu_device *adev, struct dma_fence *mfence)
 unsigned long
 svm_migrate_addr_to_pfn(struct amdgpu_device *adev, unsigned long addr)
 {
-	return (addr + adev->kfd.dev->pgmap.range.start) >> PAGE_SHIFT;
+	return (addr + adev->kfd.pgmap.range.start) >> PAGE_SHIFT;
 }
 
 static void
@@ -224,8 +217,7 @@ svm_migrate_get_vram_page(struct svm_range *prange, unsigned long pfn)
 	page = pfn_to_page(pfn);
 	svm_range_bo_ref(prange->svm_bo);
 	page->zone_device_data = prange->svm_bo;
-	get_page(page);
-	lock_page(page);
+	zone_device_page_init(page, 0);
 }
 
 static void
@@ -244,7 +236,7 @@ svm_migrate_addr(struct amdgpu_device *adev, struct page *page)
 	unsigned long addr;
 
 	addr = page_to_pfn(page) << PAGE_SHIFT;
-	return (addr - adev->kfd.dev->pgmap.range.start);
+	return (addr - adev->kfd.pgmap.range.start);
 }
 
 static struct page *
@@ -270,71 +262,58 @@ static void svm_migrate_put_sys_page(unsigned long addr)
 
 static unsigned long svm_migrate_successful_pages(struct migrate_vma *migrate)
 {
-	unsigned long cpages = 0;
+	unsigned long mpages = 0;
 	unsigned long i;
 
 	for (i = 0; i < migrate->npages; i++) {
-		if (migrate->src[i] & MIGRATE_PFN_VALID &&
+		if (migrate->dst[i] & MIGRATE_PFN_VALID &&
 		    migrate->src[i] & MIGRATE_PFN_MIGRATE)
-			cpages++;
+			mpages++;
 	}
-	return cpages;
-}
-
-static unsigned long svm_migrate_unsuccessful_pages(struct migrate_vma *migrate)
-{
-	unsigned long upages = 0;
-	unsigned long i;
-
-	for (i = 0; i < migrate->npages; i++) {
-		if (migrate->src[i] & MIGRATE_PFN_VALID &&
-		    !(migrate->src[i] & MIGRATE_PFN_MIGRATE))
-			upages++;
-	}
-	return upages;
+	return mpages;
 }
 
 static int
-svm_migrate_copy_to_vram(struct amdgpu_device *adev, struct svm_range *prange,
+svm_migrate_copy_to_vram(struct kfd_node *node, struct svm_range *prange,
 			 struct migrate_vma *migrate, struct dma_fence **mfence,
-			 dma_addr_t *scratch)
+			 dma_addr_t *scratch, u64 ttm_res_offset)
 {
-	uint64_t npages = migrate->cpages;
+	u64 npages = migrate->npages;
+	struct amdgpu_device *adev = node->adev;
 	struct device *dev = adev->dev;
 	struct amdgpu_res_cursor cursor;
+	u64 mpages = 0;
 	dma_addr_t *src;
-	uint64_t *dst;
-	uint64_t i, j;
+	u64 *dst;
+	u64 i, j;
 	int r;
 
-	pr_debug("svms 0x%p [0x%lx 0x%lx]\n", prange->svms, prange->start,
-		 prange->last);
+	pr_debug("svms 0x%p [0x%lx 0x%lx 0x%llx]\n", prange->svms, prange->start,
+		 prange->last, ttm_res_offset);
 
 	src = scratch;
-	dst = (uint64_t *)(scratch + npages);
+	dst = (u64 *)(scratch + npages);
 
-	r = svm_range_vram_node_new(adev, prange, true);
-	if (r) {
-		dev_err(adev->dev, "fail %d to alloc vram\n", r);
-		goto out;
-	}
-
-	amdgpu_res_first(prange->ttm_res, prange->offset << PAGE_SHIFT,
+	amdgpu_res_first(prange->ttm_res, ttm_res_offset,
 			 npages << PAGE_SHIFT, &cursor);
-	for (i = j = 0; i < npages; i++) {
+	for (i = j = 0; (i < npages) && (mpages < migrate->cpages); i++) {
 		struct page *spage;
 
-		spage = migrate_pfn_to_page(migrate->src[i]);
-		if (spage && !is_zone_device_page(spage)) {
+		if (migrate->src[i] & MIGRATE_PFN_MIGRATE) {
 			dst[i] = cursor.start + (j << PAGE_SHIFT);
 			migrate->dst[i] = svm_migrate_addr_to_pfn(adev, dst[i]);
 			svm_migrate_get_vram_page(prange, migrate->dst[i]);
 			migrate->dst[i] = migrate_pfn(migrate->dst[i]);
+			mpages++;
+		}
+		spage = migrate_pfn_to_page(migrate->src[i]);
+		if (spage && !is_zone_device_page(spage)) {
 			src[i] = dma_map_page(dev, spage, 0, PAGE_SIZE,
-					      DMA_TO_DEVICE);
+					      DMA_BIDIRECTIONAL);
 			r = dma_mapping_error(dev, src[i]);
 			if (r) {
-				dev_err(adev->dev, "fail %d dma_map_page\n", r);
+				dev_err(dev, "%s: fail %d dma_map_page\n",
+					__func__, r);
 				goto out_free_vram_pages;
 			}
 		} else {
@@ -346,7 +325,7 @@ svm_migrate_copy_to_vram(struct amdgpu_device *adev, struct svm_range *prange,
 						mfence);
 				if (r)
 					goto out_free_vram_pages;
-				amdgpu_res_next(&cursor, j << PAGE_SHIFT);
+				amdgpu_res_next(&cursor, (j + 1) << PAGE_SHIFT);
 				j = 0;
 			} else {
 				amdgpu_res_next(&cursor, PAGE_SIZE);
@@ -365,7 +344,7 @@ svm_migrate_copy_to_vram(struct amdgpu_device *adev, struct svm_range *prange,
 			if (r)
 				goto out_free_vram_pages;
 			amdgpu_res_next(&cursor, (j + 1) * PAGE_SIZE);
-			j= 0;
+			j = 0;
 		} else {
 			j++;
 		}
@@ -377,9 +356,12 @@ svm_migrate_copy_to_vram(struct amdgpu_device *adev, struct svm_range *prange,
 out_free_vram_pages:
 	if (r) {
 		pr_debug("failed %d to copy memory to vram\n", r);
-		while (i--) {
+		for (i = 0; i < npages && mpages; i++) {
+			if (!dst[i])
+				continue;
 			svm_migrate_put_vram_page(adev, dst[i]);
 			migrate->dst[i] = 0;
+			mpages--;
 		}
 	}
 
@@ -397,22 +379,24 @@ out_free_vram_pages:
 		migrate->dst[i + 3] = 0;
 	}
 #endif
-out:
+
 	return r;
 }
 
 static long
-svm_migrate_vma_to_vram(struct amdgpu_device *adev, struct svm_range *prange,
-			struct vm_area_struct *vma, uint64_t start,
-			uint64_t end)
+svm_migrate_vma_to_vram(struct kfd_node *node, struct svm_range *prange,
+			struct vm_area_struct *vma, u64 start,
+			u64 end, uint32_t trigger, u64 ttm_res_offset)
 {
-	uint64_t npages = (end - start) >> PAGE_SHIFT;
+	struct kfd_process *p = container_of(prange->svms, struct kfd_process, svms);
+	u64 npages = (end - start) >> PAGE_SHIFT;
+	struct amdgpu_device *adev = node->adev;
 	struct kfd_process_device *pdd;
 	struct dma_fence *mfence = NULL;
-	struct migrate_vma migrate;
+	struct migrate_vma migrate = { 0 };
 	unsigned long cpages = 0;
+	unsigned long mpages = 0;
 	dma_addr_t *scratch;
-	size_t size;
 	void *buf;
 	int r = -ENOMEM;
 
@@ -423,9 +407,9 @@ svm_migrate_vma_to_vram(struct amdgpu_device *adev, struct svm_range *prange,
 	migrate.flags = MIGRATE_VMA_SELECT_SYSTEM;
 	migrate.pgmap_owner = SVM_ADEV_PGMAP_OWNER(adev);
 
-	size = 2 * sizeof(*migrate.src) + sizeof(uint64_t) + sizeof(dma_addr_t);
-	size *= npages;
-	buf = kvmalloc(size, GFP_KERNEL | __GFP_ZERO);
+	buf = kvcalloc(npages,
+		       2 * sizeof(*migrate.src) + sizeof(u64) + sizeof(dma_addr_t),
+		       GFP_KERNEL);
 	if (!buf)
 		goto out;
 
@@ -433,10 +417,15 @@ svm_migrate_vma_to_vram(struct amdgpu_device *adev, struct svm_range *prange,
 	migrate.dst = migrate.src + npages;
 	scratch = (dma_addr_t *)(migrate.dst + npages);
 
+	kfd_smi_event_migration_start(node, p->lead_thread->pid,
+				      start >> PAGE_SHIFT, end >> PAGE_SHIFT,
+				      0, node->id, prange->prefetch_loc,
+				      prange->preferred_loc, trigger);
+
 	r = migrate_vma_setup(&migrate);
 	if (r) {
-		dev_err(adev->dev, "vma setup fail %d range [0x%lx 0x%lx]\n", r,
-			prange->start, prange->last);
+		dev_err(adev->dev, "%s: vma setup fail %d range [0x%lx 0x%lx]\n",
+			__func__, r, prange->start, prange->last);
 		goto out_free;
 	}
 
@@ -447,32 +436,35 @@ svm_migrate_vma_to_vram(struct amdgpu_device *adev, struct svm_range *prange,
 		goto out_free;
 	}
 	if (cpages != npages)
-		pr_debug("partial migration, 0x%lx/0x%llx pages migrated\n",
+		pr_debug("partial migration, 0x%lx/0x%llx pages collected\n",
 			 cpages, npages);
 	else
-		pr_debug("0x%lx pages migrated\n", cpages);
+		pr_debug("0x%lx pages collected\n", cpages);
 
-	r = svm_migrate_copy_to_vram(adev, prange, &migrate, &mfence, scratch);
+	r = svm_migrate_copy_to_vram(node, prange, &migrate, &mfence, scratch, ttm_res_offset);
 	migrate_vma_pages(&migrate);
 
-	pr_debug("successful/cpages/npages 0x%lx/0x%lx/0x%lx\n",
-		svm_migrate_successful_pages(&migrate), cpages, migrate.npages);
-
 	svm_migrate_copy_done(adev, mfence);
 	migrate_vma_finalize(&migrate);
 
-	svm_range_dma_unmap(adev->dev, scratch, 0, npages);
-	svm_range_free_dma_mappings(prange);
+	mpages = svm_migrate_successful_pages(&migrate);
+	pr_debug("migrated/collected/requested 0x%lx/0x%lx/0x%lx\n",
+		 mpages, cpages, migrate.npages);
+
+	svm_range_dma_unmap_dev(adev->dev, scratch, 0, npages);
 
 out_free:
 	kvfree(buf);
+	kfd_smi_event_migration_end(node, p->lead_thread->pid,
+				    start >> PAGE_SHIFT, end >> PAGE_SHIFT,
+				    0, node->id, trigger, r);
 out:
-	if (!r && cpages) {
-		pdd = svm_range_get_pdd_by_adev(prange, adev);
+	if (!r && mpages) {
+		pdd = svm_range_get_pdd_by_node(prange, node);
 		if (pdd)
-			WRITE_ONCE(pdd->page_in, pdd->page_in + cpages);
+			WRITE_ONCE(pdd->page_in, pdd->page_in + mpages);
 
-		return cpages;
+		return mpages;
 	}
 	return r;
 }
@@ -481,7 +473,10 @@ out:
  * svm_migrate_ram_to_vram - migrate svm range from system to device
  * @prange: range structure
  * @best_loc: the device to migrate to
+ * @start_mgr: start page to migrate
+ * @last_mgr: last page to migrate
  * @mm: the process mm structure
+ * @trigger: reason of migration
  *
  * Context: Process context, caller hold mmap read lock, svms lock, prange lock
  *
@@ -490,88 +485,118 @@ out:
  */
 static int
 svm_migrate_ram_to_vram(struct svm_range *prange, uint32_t best_loc,
-			struct mm_struct *mm)
+			unsigned long start_mgr, unsigned long last_mgr,
+			struct mm_struct *mm, uint32_t trigger)
 {
 	unsigned long addr, start, end;
 	struct vm_area_struct *vma;
-	struct amdgpu_device *adev;
-	unsigned long cpages = 0;
+	u64 ttm_res_offset;
+	struct kfd_node *node;
+	unsigned long mpages = 0;
 	long r = 0;
 
-	if (prange->actual_loc == best_loc) {
-		pr_debug("svms 0x%p [0x%lx 0x%lx] already on best_loc 0x%x\n",
-			 prange->svms, prange->start, prange->last, best_loc);
-		return 0;
+	if (start_mgr < prange->start || last_mgr > prange->last) {
+		pr_debug("range [0x%lx 0x%lx] out prange [0x%lx 0x%lx]\n",
+			 start_mgr, last_mgr, prange->start, prange->last);
+		return -EFAULT;
 	}
 
-	adev = svm_range_get_adev_by_id(prange, best_loc);
-	if (!adev) {
-		pr_debug("failed to get device by id 0x%x\n", best_loc);
+	node = svm_range_get_node_by_id(prange, best_loc);
+	if (!node) {
+		pr_debug("failed to get kfd node by id 0x%x\n", best_loc);
 		return -ENODEV;
 	}
 
-	pr_debug("svms 0x%p [0x%lx 0x%lx] to gpu 0x%x\n", prange->svms,
-		 prange->start, prange->last, best_loc);
+	pr_debug("svms 0x%p [0x%lx 0x%lx] in [0x%lx 0x%lx] to gpu 0x%x\n",
+		prange->svms, start_mgr, last_mgr, prange->start, prange->last,
+		best_loc);
 
-	/* FIXME: workaround for page locking bug with invalid pages */
-	svm_range_prefault(prange, mm, SVM_ADEV_PGMAP_OWNER(adev));
+	start = start_mgr << PAGE_SHIFT;
+	end = (last_mgr + 1) << PAGE_SHIFT;
 
-	start = prange->start << PAGE_SHIFT;
-	end = (prange->last + 1) << PAGE_SHIFT;
+	r = amdgpu_amdkfd_reserve_mem_limit(node->adev,
+					prange->npages * PAGE_SIZE,
+					KFD_IOC_ALLOC_MEM_FLAGS_VRAM,
+					node->xcp ? node->xcp->id : 0);
+	if (r) {
+		dev_dbg(node->adev->dev, "failed to reserve VRAM, r: %ld\n", r);
+		return -ENOSPC;
+	}
+
+	r = svm_range_vram_node_new(node, prange, true);
+	if (r) {
+		dev_dbg(node->adev->dev, "fail %ld to alloc vram\n", r);
+		goto out;
+	}
+	ttm_res_offset = (start_mgr - prange->start + prange->offset) << PAGE_SHIFT;
 
 	for (addr = start; addr < end;) {
 		unsigned long next;
 
-		vma = find_vma(mm, addr);
-		if (!vma || addr < vma->vm_start)
+		vma = vma_lookup(mm, addr);
+		if (!vma)
 			break;
 
 		next = min(vma->vm_end, end);
-		r = svm_migrate_vma_to_vram(adev, prange, vma, addr, next);
+		r = svm_migrate_vma_to_vram(node, prange, vma, addr, next, trigger, ttm_res_offset);
 		if (r < 0) {
 			pr_debug("failed %ld to migrate\n", r);
 			break;
 		} else {
-			cpages += r;
+			mpages += r;
 		}
+		ttm_res_offset += next - addr;
 		addr = next;
 	}
 
-	if (cpages)
+	if (mpages) {
 		prange->actual_loc = best_loc;
+		prange->vram_pages += mpages;
+	} else if (!prange->actual_loc) {
+		/* if no page migrated and all pages from prange are at
+		 * sys ram drop svm_bo got from svm_range_vram_node_new
+		 */
+		svm_range_vram_node_free(prange);
+	}
 
+out:
+	amdgpu_amdkfd_unreserve_mem_limit(node->adev,
+					prange->npages * PAGE_SIZE,
+					KFD_IOC_ALLOC_MEM_FLAGS_VRAM,
+					node->xcp ? node->xcp->id : 0);
 	return r < 0 ? r : 0;
 }
 
-static void svm_migrate_page_free(struct page *page)
+static void svm_migrate_folio_free(struct folio *folio)
 {
+	struct page *page = &folio->page;
 	struct svm_range_bo *svm_bo = page->zone_device_data;
 
 	if (svm_bo) {
 		pr_debug_ratelimited("ref: %d\n", kref_read(&svm_bo->kref));
-		svm_range_bo_unref(svm_bo);
+		svm_range_bo_unref_async(svm_bo);
 	}
 }
 
 static int
 svm_migrate_copy_to_ram(struct amdgpu_device *adev, struct svm_range *prange,
 			struct migrate_vma *migrate, struct dma_fence **mfence,
-			dma_addr_t *scratch, uint64_t npages)
+			dma_addr_t *scratch, u64 npages)
 {
 	struct device *dev = adev->dev;
-	uint64_t *src;
+	u64 *src;
 	dma_addr_t *dst;
 	struct page *dpage;
-	uint64_t i = 0, j;
-	uint64_t addr;
+	u64 i = 0, j;
+	u64 addr;
 	int r = 0;
 
 	pr_debug("svms 0x%p [0x%lx 0x%lx]\n", prange->svms, prange->start,
 		 prange->last);
 
-	addr = prange->start << PAGE_SHIFT;
+	addr = migrate->start;
 
-	src = (uint64_t *)(scratch + npages);
+	src = (u64 *)(scratch + npages);
 	dst = scratch;
 
 	for (i = 0, j = 0; i < npages; i++, addr += PAGE_SIZE) {
@@ -593,7 +618,7 @@ svm_migrate_copy_to_ram(struct amdgpu_device *adev, struct svm_range *prange,
 			continue;
 		}
 		src[i] = svm_migrate_addr(adev, spage);
-		if (i > 0 && src[i] != src[i - 1] + PAGE_SIZE) {
+		if (j > 0 && src[i] != src[i - 1] + PAGE_SIZE) {
 			r = svm_migrate_copy_memory_gart(adev, dst + i - j,
 							 src + i - j, j,
 							 FROM_VRAM_TO_RAM,
@@ -611,10 +636,10 @@ svm_migrate_copy_to_ram(struct amdgpu_device *adev, struct svm_range *prange,
 			goto out_oom;
 		}
 
-		dst[i] = dma_map_page(dev, dpage, 0, PAGE_SIZE, DMA_FROM_DEVICE);
+		dst[i] = dma_map_page(dev, dpage, 0, PAGE_SIZE, DMA_BIDIRECTIONAL);
 		r = dma_mapping_error(dev, dst[i]);
 		if (r) {
-			dev_err(adev->dev, "fail %d dma_map_page\n", r);
+			dev_err(adev->dev, "%s: fail %d dma_map_page\n", __func__, r);
 			goto out_oom;
 		}
 
@@ -640,18 +665,37 @@ out_oom:
 	return r;
 }
 
+/**
+ * svm_migrate_vma_to_ram - migrate range inside one vma from device to system
+ *
+ * @prange: svm range structure
+ * @vma: vm_area_struct that range [start, end] belongs to
+ * @start: range start virtual address in pages
+ * @end: range end virtual address in pages
+ * @node: kfd node device to migrate from
+ * @trigger: reason of migration
+ * @fault_page: is from vmf->page, svm_migrate_to_ram(), this is CPU page fault callback
+ *
+ * Context: Process context, caller hold mmap read lock, prange->migrate_mutex
+ *
+ * Return:
+ *   negative values - indicate error
+ *   positive values or zero - number of pages got migrated
+ */
 static long
-svm_migrate_vma_to_ram(struct amdgpu_device *adev, struct svm_range *prange,
-		       struct vm_area_struct *vma, uint64_t start, uint64_t end)
+svm_migrate_vma_to_ram(struct kfd_node *node, struct svm_range *prange,
+		       struct vm_area_struct *vma, u64 start, u64 end,
+		       uint32_t trigger, struct page *fault_page)
 {
-	uint64_t npages = (end - start) >> PAGE_SHIFT;
-	unsigned long upages = npages;
+	struct kfd_process *p = container_of(prange->svms, struct kfd_process, svms);
+	u64 npages = (end - start) >> PAGE_SHIFT;
 	unsigned long cpages = 0;
+	unsigned long mpages = 0;
+	struct amdgpu_device *adev = node->adev;
 	struct kfd_process_device *pdd;
 	struct dma_fence *mfence = NULL;
-	struct migrate_vma migrate;
+	struct migrate_vma migrate = { 0 };
 	dma_addr_t *scratch;
-	size_t size;
 	void *buf;
 	int r = -ENOMEM;
 
@@ -659,23 +703,32 @@ svm_migrate_vma_to_ram(struct amdgpu_device *adev, struct svm_range *prange,
 	migrate.vma = vma;
 	migrate.start = start;
 	migrate.end = end;
-	migrate.flags = MIGRATE_VMA_SELECT_DEVICE_PRIVATE;
 	migrate.pgmap_owner = SVM_ADEV_PGMAP_OWNER(adev);
+	if (adev->gmc.xgmi.connected_to_cpu)
+		migrate.flags = MIGRATE_VMA_SELECT_DEVICE_COHERENT;
+	else
+		migrate.flags = MIGRATE_VMA_SELECT_DEVICE_PRIVATE;
 
-	size = 2 * sizeof(*migrate.src) + sizeof(uint64_t) + sizeof(dma_addr_t);
-	size *= npages;
-	buf = kvmalloc(size, GFP_KERNEL | __GFP_ZERO);
+	buf = kvcalloc(npages,
+		       2 * sizeof(*migrate.src) + sizeof(u64) + sizeof(dma_addr_t),
+		       GFP_KERNEL);
 	if (!buf)
 		goto out;
 
 	migrate.src = buf;
 	migrate.dst = migrate.src + npages;
+	migrate.fault_page = fault_page;
 	scratch = (dma_addr_t *)(migrate.dst + npages);
 
+	kfd_smi_event_migration_start(node, p->lead_thread->pid,
+				      start >> PAGE_SHIFT, end >> PAGE_SHIFT,
+				      node->id, 0, prange->prefetch_loc,
+				      prange->preferred_loc, trigger);
+
 	r = migrate_vma_setup(&migrate);
 	if (r) {
-		dev_err(adev->dev, "vma setup fail %d range [0x%lx 0x%lx]\n", r,
-			prange->start, prange->last);
+		dev_err(adev->dev, "%s: vma setup fail %d range [0x%lx 0x%lx]\n",
+			__func__, r, prange->start, prange->last);
 		goto out_free;
 	}
 
@@ -683,101 +736,128 @@ svm_migrate_vma_to_ram(struct amdgpu_device *adev, struct svm_range *prange,
 	if (!cpages) {
 		pr_debug("failed collect migrate device pages [0x%lx 0x%lx]\n",
 			 prange->start, prange->last);
-		upages = svm_migrate_unsuccessful_pages(&migrate);
 		goto out_free;
 	}
 	if (cpages != npages)
-		pr_debug("partial migration, 0x%lx/0x%llx pages migrated\n",
+		pr_debug("partial migration, 0x%lx/0x%llx pages collected\n",
 			 cpages, npages);
 	else
-		pr_debug("0x%lx pages migrated\n", cpages);
+		pr_debug("0x%lx pages collected\n", cpages);
 
 	r = svm_migrate_copy_to_ram(adev, prange, &migrate, &mfence,
 				    scratch, npages);
 	migrate_vma_pages(&migrate);
 
-	upages = svm_migrate_unsuccessful_pages(&migrate);
-	pr_debug("unsuccessful/cpages/npages 0x%lx/0x%lx/0x%lx\n",
-		 upages, cpages, migrate.npages);
+	mpages = svm_migrate_successful_pages(&migrate);
+	pr_debug("migrated/collected/requested 0x%lx/0x%lx/0x%lx\n",
+		 mpages, cpages, migrate.npages);
 
 	svm_migrate_copy_done(adev, mfence);
 	migrate_vma_finalize(&migrate);
-	svm_range_dma_unmap(adev->dev, scratch, 0, npages);
+
+	svm_range_dma_unmap_dev(adev->dev, scratch, 0, npages);
 
 out_free:
 	kvfree(buf);
+	kfd_smi_event_migration_end(node, p->lead_thread->pid,
+				    start >> PAGE_SHIFT, end >> PAGE_SHIFT,
+				    node->id, 0, trigger, r);
 out:
-	if (!r && cpages) {
-		pdd = svm_range_get_pdd_by_adev(prange, adev);
+	if (!r && mpages) {
+		pdd = svm_range_get_pdd_by_node(prange, node);
 		if (pdd)
-			WRITE_ONCE(pdd->page_out, pdd->page_out + cpages);
-
-		return upages;
+			WRITE_ONCE(pdd->page_out, pdd->page_out + mpages);
 	}
-	return r ? r : upages;
+
+	return r ? r : mpages;
 }
 
 /**
  * svm_migrate_vram_to_ram - migrate svm range from device to system
  * @prange: range structure
  * @mm: process mm, use current->mm if NULL
+ * @start_mgr: start page need be migrated to sys ram
+ * @last_mgr: last page need be migrated to sys ram
+ * @trigger: reason of migration
+ * @fault_page: is from vmf->page, svm_migrate_to_ram(), this is CPU page fault callback
  *
- * Context: Process context, caller hold mmap read lock, svms lock, prange lock
+ * Context: Process context, caller hold mmap read lock, prange->migrate_mutex
  *
  * Return:
  * 0 - OK, otherwise error code
  */
-int svm_migrate_vram_to_ram(struct svm_range *prange, struct mm_struct *mm)
+int svm_migrate_vram_to_ram(struct svm_range *prange, struct mm_struct *mm,
+			    unsigned long start_mgr, unsigned long last_mgr,
+			    uint32_t trigger, struct page *fault_page)
 {
-	struct amdgpu_device *adev;
+	struct kfd_node *node;
 	struct vm_area_struct *vma;
 	unsigned long addr;
 	unsigned long start;
 	unsigned long end;
-	unsigned long upages = 0;
+	unsigned long mpages = 0;
 	long r = 0;
 
+	/* this pragne has no any vram page to migrate to sys ram */
 	if (!prange->actual_loc) {
 		pr_debug("[0x%lx 0x%lx] already migrated to ram\n",
 			 prange->start, prange->last);
 		return 0;
 	}
 
-	adev = svm_range_get_adev_by_id(prange, prange->actual_loc);
-	if (!adev) {
-		pr_debug("failed to get device by id 0x%x\n",
-			 prange->actual_loc);
-		return -ENODEV;
+	if (start_mgr < prange->start || last_mgr > prange->last) {
+		pr_debug("range [0x%lx 0x%lx] out prange [0x%lx 0x%lx]\n",
+			 start_mgr, last_mgr, prange->start, prange->last);
+		return -EFAULT;
 	}
 
+	node = svm_range_get_node_by_id(prange, prange->actual_loc);
+	if (!node) {
+		pr_debug("failed to get kfd node by id 0x%x\n", prange->actual_loc);
+		return -ENODEV;
+	}
 	pr_debug("svms 0x%p prange 0x%p [0x%lx 0x%lx] from gpu 0x%x to ram\n",
-		 prange->svms, prange, prange->start, prange->last,
+		 prange->svms, prange, start_mgr, last_mgr,
 		 prange->actual_loc);
 
-	start = prange->start << PAGE_SHIFT;
-	end = (prange->last + 1) << PAGE_SHIFT;
+	start = start_mgr << PAGE_SHIFT;
+	end = (last_mgr + 1) << PAGE_SHIFT;
 
 	for (addr = start; addr < end;) {
 		unsigned long next;
 
-		vma = find_vma(mm, addr);
-		if (!vma || addr < vma->vm_start)
+		vma = vma_lookup(mm, addr);
+		if (!vma) {
+			pr_debug("failed to find vma for prange %p\n", prange);
+			r = -EFAULT;
 			break;
+		}
 
 		next = min(vma->vm_end, end);
-		r = svm_migrate_vma_to_ram(adev, prange, vma, addr, next);
+		r = svm_migrate_vma_to_ram(node, prange, vma, addr, next, trigger,
+			fault_page);
 		if (r < 0) {
-			pr_debug("failed %ld to migrate\n", r);
+			pr_debug("failed %ld to migrate prange %p\n", r, prange);
 			break;
 		} else {
-			upages += r;
+			mpages += r;
 		}
 		addr = next;
 	}
 
-	if (!upages) {
-		svm_range_vram_node_free(prange);
-		prange->actual_loc = 0;
+	if (r >= 0) {
+		WARN_ONCE(prange->vram_pages < mpages,
+			  "Recorded vram pages(0x%llx) should not be less than migration pages(0x%lx).",
+			  prange->vram_pages, mpages);
+		prange->vram_pages -= mpages;
+
+		/* prange does not have vram page set its actual_loc to system
+		 * and drop its svm_bo ref
+		 */
+		if (prange->vram_pages == 0 && prange->ttm_res) {
+			prange->actual_loc = 0;
+			svm_range_vram_node_free(prange);
+		}
 	}
 
 	return r < 0 ? r : 0;
@@ -787,16 +867,23 @@ int svm_migrate_vram_to_ram(struct svm_range *prange, struct mm_struct *mm)
  * svm_migrate_vram_to_vram - migrate svm range from device to device
  * @prange: range structure
  * @best_loc: the device to migrate to
+ * @start: start page need be migrated to sys ram
+ * @last: last page need be migrated to sys ram
  * @mm: process mm, use current->mm if NULL
+ * @trigger: reason of migration
  *
  * Context: Process context, caller hold mmap read lock, svms lock, prange lock
  *
+ * migrate all vram pages in prange to sys ram, then migrate
+ * [start, last] pages from sys ram to gpu node best_loc.
+ *
  * Return:
  * 0 - OK, otherwise error code
  */
 static int
 svm_migrate_vram_to_vram(struct svm_range *prange, uint32_t best_loc,
-			 struct mm_struct *mm)
+			unsigned long start, unsigned long last,
+			struct mm_struct *mm, uint32_t trigger)
 {
 	int r, retries = 3;
 
@@ -808,7 +895,8 @@ svm_migrate_vram_to_vram(struct svm_range *prange, uint32_t best_loc,
 	pr_debug("from gpu 0x%x to gpu 0x%x\n", prange->actual_loc, best_loc);
 
 	do {
-		r = svm_migrate_vram_to_ram(prange, mm);
+		r = svm_migrate_vram_to_ram(prange, mm, prange->start, prange->last,
+					    trigger, NULL);
 		if (r)
 			return r;
 	} while (prange->actual_loc && --retries);
@@ -816,17 +904,21 @@ svm_migrate_vram_to_vram(struct svm_range *prange, uint32_t best_loc,
 	if (prange->actual_loc)
 		return -EDEADLK;
 
-	return svm_migrate_ram_to_vram(prange, best_loc, mm);
+	return svm_migrate_ram_to_vram(prange, best_loc, start, last, mm, trigger);
 }
 
 int
 svm_migrate_to_vram(struct svm_range *prange, uint32_t best_loc,
-		    struct mm_struct *mm)
+		    unsigned long start, unsigned long last,
+		    struct mm_struct *mm, uint32_t trigger)
 {
-	if  (!prange->actual_loc)
-		return svm_migrate_ram_to_vram(prange, best_loc, mm);
+	if  (!prange->actual_loc || prange->actual_loc == best_loc)
+		return svm_migrate_ram_to_vram(prange, best_loc, start, last,
+					       mm, trigger);
+
 	else
-		return svm_migrate_vram_to_vram(prange, best_loc, mm);
+		return svm_migrate_vram_to_vram(prange, best_loc, start, last,
+						mm, trigger);
 
 }
 
@@ -842,105 +934,103 @@ svm_migrate_to_vram(struct svm_range *prange, uint32_t best_loc,
  */
 static vm_fault_t svm_migrate_to_ram(struct vm_fault *vmf)
 {
+	unsigned long start, last, size;
 	unsigned long addr = vmf->address;
-	struct vm_area_struct *vma;
-	enum svm_work_list_ops op;
-	struct svm_range *parent;
+	struct svm_range_bo *svm_bo;
 	struct svm_range *prange;
 	struct kfd_process *p;
 	struct mm_struct *mm;
 	int r = 0;
 
-	vma = vmf->vma;
-	mm = vma->vm_mm;
+	svm_bo = vmf->page->zone_device_data;
+	if (!svm_bo) {
+		pr_debug("failed get device page at addr 0x%lx\n", addr);
+		return VM_FAULT_SIGBUS;
+	}
+	if (!mmget_not_zero(svm_bo->eviction_fence->mm)) {
+		pr_debug("addr 0x%lx of process mm is destroyed\n", addr);
+		return VM_FAULT_SIGBUS;
+	}
 
-	p = kfd_lookup_process_by_mm(vma->vm_mm);
+	mm = svm_bo->eviction_fence->mm;
+	if (mm != vmf->vma->vm_mm)
+		pr_debug("addr 0x%lx is COW mapping in child process\n", addr);
+
+	p = kfd_lookup_process_by_mm(mm);
 	if (!p) {
 		pr_debug("failed find process at fault address 0x%lx\n", addr);
-		return VM_FAULT_SIGBUS;
+		r = VM_FAULT_SIGBUS;
+		goto out_mmput;
 	}
 	if (READ_ONCE(p->svms.faulting_task) == current) {
 		pr_debug("skipping ram migration\n");
-		kfd_unref_process(p);
-		return 0;
+		r = 0;
+		goto out_unref_process;
 	}
-	addr >>= PAGE_SHIFT;
+
 	pr_debug("CPU page fault svms 0x%p address 0x%lx\n", &p->svms, addr);
+	addr >>= PAGE_SHIFT;
 
 	mutex_lock(&p->svms.lock);
 
-	prange = svm_range_from_addr(&p->svms, addr, &parent);
+	prange = svm_range_from_addr(&p->svms, addr, NULL);
 	if (!prange) {
-		pr_debug("cannot find svm range at 0x%lx\n", addr);
+		pr_debug("failed get range svms 0x%p addr 0x%lx\n", &p->svms, addr);
 		r = -EFAULT;
-		goto out;
+		goto out_unlock_svms;
 	}
 
-	mutex_lock(&parent->migrate_mutex);
-	if (prange != parent)
-		mutex_lock_nested(&prange->migrate_mutex, 1);
+	mutex_lock(&prange->migrate_mutex);
 
 	if (!prange->actual_loc)
 		goto out_unlock_prange;
 
-	svm_range_lock(parent);
-	if (prange != parent)
-		mutex_lock_nested(&prange->lock, 1);
-	r = svm_range_split_by_granularity(p, mm, addr, parent, prange);
-	if (prange != parent)
-		mutex_unlock(&prange->lock);
-	svm_range_unlock(parent);
-	if (r) {
-		pr_debug("failed %d to split range by granularity\n", r);
-		goto out_unlock_prange;
-	}
+	/* Align migration range start and size to granularity size */
+	size = 1UL << prange->granularity;
+	start = max(ALIGN_DOWN(addr, size), prange->start);
+	last = min(ALIGN(addr + 1, size) - 1, prange->last);
 
-	r = svm_migrate_vram_to_ram(prange, mm);
+	r = svm_migrate_vram_to_ram(prange, vmf->vma->vm_mm, start, last,
+				    KFD_MIGRATE_TRIGGER_PAGEFAULT_CPU, vmf->page);
 	if (r)
-		pr_debug("failed %d migrate 0x%p [0x%lx 0x%lx] to ram\n", r,
-			 prange, prange->start, prange->last);
-
-	/* xnack on, update mapping on GPUs with ACCESS_IN_PLACE */
-	if (p->xnack_enabled && parent == prange)
-		op = SVM_OP_UPDATE_RANGE_NOTIFIER_AND_MAP;
-	else
-		op = SVM_OP_UPDATE_RANGE_NOTIFIER;
-	svm_range_add_list_work(&p->svms, parent, mm, op);
-	schedule_deferred_list_work(&p->svms);
+		pr_debug("failed %d migrate svms 0x%p range 0x%p [0x%lx 0x%lx]\n",
+			r, prange->svms, prange, start, last);
 
 out_unlock_prange:
-	if (prange != parent)
-		mutex_unlock(&prange->migrate_mutex);
-	mutex_unlock(&parent->migrate_mutex);
-out:
+	mutex_unlock(&prange->migrate_mutex);
+out_unlock_svms:
 	mutex_unlock(&p->svms.lock);
-	kfd_unref_process(p);
-
+out_unref_process:
 	pr_debug("CPU fault svms 0x%p address 0x%lx done\n", &p->svms, addr);
-
+	kfd_unref_process(p);
+out_mmput:
+	mmput(mm);
 	return r ? VM_FAULT_SIGBUS : 0;
 }
 
 static const struct dev_pagemap_ops svm_migrate_pgmap_ops = {
-	.page_free		= svm_migrate_page_free,
+	.folio_free		= svm_migrate_folio_free,
 	.migrate_to_ram		= svm_migrate_to_ram,
 };
 
 /* Each VRAM page uses sizeof(struct page) on system memory */
 #define SVM_HMM_PAGE_STRUCT_SIZE(size) ((size)/PAGE_SIZE * sizeof(struct page))
 
-int svm_migrate_init(struct amdgpu_device *adev)
+int kgd2kfd_init_zone_device(struct amdgpu_device *adev)
 {
-	struct kfd_dev *kfddev = adev->kfd.dev;
+	struct amdgpu_kfd_dev *kfddev = &adev->kfd;
 	struct dev_pagemap *pgmap;
-	struct resource *res;
+	struct resource *res = NULL;
 	unsigned long size;
 	void *r;
 
-	/* Page migration works on Vega10 or newer */
-	if (kfddev->device_info->asic_family < CHIP_VEGA10)
+	/* Page migration works on gfx9 or newer */
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) < IP_VERSION(9, 0, 1))
 		return -EINVAL;
 
+	if (adev->apu_prefer_gtt)
+		return 0;
+
 	pgmap = &kfddev->pgmap;
 	memset(pgmap, 0, sizeof(*pgmap));
 
@@ -948,28 +1038,33 @@ int svm_migrate_init(struct amdgpu_device *adev)
 	 * should remove reserved size
 	 */
 	size = ALIGN(adev->gmc.real_vram_size, 2ULL << 20);
-	res = devm_request_free_mem_region(adev->dev, &iomem_resource, size);
-	if (IS_ERR(res))
-		return -ENOMEM;
+	if (adev->gmc.xgmi.connected_to_cpu) {
+		pgmap->range.start = adev->gmc.aper_base;
+		pgmap->range.end = adev->gmc.aper_base + adev->gmc.aper_size - 1;
+		pgmap->type = MEMORY_DEVICE_COHERENT;
+	} else {
+		res = devm_request_free_mem_region(adev->dev, &iomem_resource, size);
+		if (IS_ERR(res))
+			return PTR_ERR(res);
+		pgmap->range.start = res->start;
+		pgmap->range.end = res->end;
+		pgmap->type = MEMORY_DEVICE_PRIVATE;
+	}
 
-	pgmap->type = MEMORY_DEVICE_PRIVATE;
 	pgmap->nr_range = 1;
-	pgmap->range.start = res->start;
-	pgmap->range.end = res->end;
 	pgmap->ops = &svm_migrate_pgmap_ops;
 	pgmap->owner = SVM_ADEV_PGMAP_OWNER(adev);
-	pgmap->flags = MIGRATE_VMA_SELECT_DEVICE_PRIVATE;
-
+	pgmap->flags = 0;
 	/* Device manager releases device-specific resources, memory region and
 	 * pgmap when driver disconnects from device.
 	 */
 	r = devm_memremap_pages(adev->dev, pgmap);
 	if (IS_ERR(r)) {
 		pr_err("failed to register HMM device memory\n");
-
+		if (pgmap->type == MEMORY_DEVICE_PRIVATE)
+			devm_release_mem_region(adev->dev, res->start, resource_size(res));
 		/* Disable SVM support capability */
 		pgmap->type = 0;
-		devm_release_mem_region(adev->dev, res->start, resource_size(res));
 		return PTR_ERR(r);
 	}
 
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_migrate.h b/drivers/gpu/drm/amd/amdkfd/kfd_migrate.h
index 2f5b3394c9ed..2b7fd442d29c 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_migrate.h
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_migrate.h
@@ -31,7 +31,6 @@
 #include <linux/list.h>
 #include <linux/mutex.h>
 #include <linux/sched/mm.h>
-#include <linux/hmm.h>
 #include "kfd_priv.h"
 #include "kfd_svm.h"
 
@@ -41,19 +40,15 @@ enum MIGRATION_COPY_DIR {
 };
 
 int svm_migrate_to_vram(struct svm_range *prange,  uint32_t best_loc,
-			struct mm_struct *mm);
-int svm_migrate_vram_to_ram(struct svm_range *prange, struct mm_struct *mm);
-unsigned long
-svm_migrate_addr_to_pfn(struct amdgpu_device *adev, unsigned long addr);
-
-int svm_migrate_init(struct amdgpu_device *adev);
+			unsigned long start, unsigned long last,
+			struct mm_struct *mm, uint32_t trigger);
 
-#else
+int svm_migrate_vram_to_ram(struct svm_range *prange, struct mm_struct *mm,
+			    unsigned long start, unsigned long last,
+			    uint32_t trigger, struct page *fault_page);
 
-static inline int svm_migrate_init(struct amdgpu_device *adev)
-{
-	return 0;
-}
+unsigned long
+svm_migrate_addr_to_pfn(struct amdgpu_device *adev, unsigned long addr);
 
 #endif /* IS_ENABLED(CONFIG_HSA_AMD_SVM) */
 
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_module.c b/drivers/gpu/drm/amd/amdkfd/kfd_module.c
index 5e90fe642192..33aa23450b3f 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_module.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_module.c
@@ -1,5 +1,6 @@
+// SPDX-License-Identifier: GPL-2.0 OR MIT
 /*
- * Copyright 2014 Advanced Micro Devices, Inc.
+ * Copyright 2014-2022 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -76,8 +77,9 @@ err_ioctl:
 
 static void kfd_exit(void)
 {
-	kfd_debugfs_fini();
+	kfd_cleanup_processes();
 	kfd_process_destroy_wq();
+	kfd_debugfs_fini();
 	kfd_procfs_shutdown();
 	kfd_topology_shutdown();
 	kfd_chardev_exit();
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager.c b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager.c
index c021519af810..d9ae854b6908 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager.c
@@ -1,5 +1,6 @@
+// SPDX-License-Identifier: GPL-2.0 OR MIT
 /*
- * Copyright 2014 Advanced Micro Devices, Inc.
+ * Copyright 2014-2022 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -45,9 +46,9 @@ int pipe_priority_map[] = {
 	KFD_PIPE_PRIORITY_CS_HIGH
 };
 
-struct kfd_mem_obj *allocate_hiq_mqd(struct kfd_dev *dev, struct queue_properties *q)
+struct kfd_mem_obj *allocate_hiq_mqd(struct kfd_node *dev, struct queue_properties *q)
 {
-	struct kfd_mem_obj *mqd_mem_obj = NULL;
+	struct kfd_mem_obj *mqd_mem_obj;
 
 	mqd_mem_obj = kzalloc(sizeof(struct kfd_mem_obj), GFP_KERNEL);
 	if (!mqd_mem_obj)
@@ -60,10 +61,10 @@ struct kfd_mem_obj *allocate_hiq_mqd(struct kfd_dev *dev, struct queue_propertie
 	return mqd_mem_obj;
 }
 
-struct kfd_mem_obj *allocate_sdma_mqd(struct kfd_dev *dev,
+struct kfd_mem_obj *allocate_sdma_mqd(struct kfd_node *dev,
 					struct queue_properties *q)
 {
-	struct kfd_mem_obj *mqd_mem_obj = NULL;
+	struct kfd_mem_obj *mqd_mem_obj;
 	uint64_t offset;
 
 	mqd_mem_obj = kzalloc(sizeof(struct kfd_mem_obj), GFP_KERNEL);
@@ -71,11 +72,12 @@ struct kfd_mem_obj *allocate_sdma_mqd(struct kfd_dev *dev,
 		return NULL;
 
 	offset = (q->sdma_engine_id *
-		dev->device_info->num_sdma_queues_per_engine +
+		dev->kfd->device_info.num_sdma_queues_per_engine +
 		q->sdma_queue_id) *
 		dev->dqm->mqd_mgrs[KFD_MQD_TYPE_SDMA]->mqd_size;
 
-	offset += dev->dqm->mqd_mgrs[KFD_MQD_TYPE_HIQ]->mqd_size;
+	offset += dev->dqm->mqd_mgrs[KFD_MQD_TYPE_HIQ]->mqd_size *
+		  NUM_XCC(dev->xcc_mask);
 
 	mqd_mem_obj->gtt_mem = (void *)((uint64_t)dev->dqm->hiq_sdma_mqd.gtt_mem
 				+ offset);
@@ -95,29 +97,42 @@ void free_mqd_hiq_sdma(struct mqd_manager *mm, void *mqd,
 
 void mqd_symmetrically_map_cu_mask(struct mqd_manager *mm,
 		const uint32_t *cu_mask, uint32_t cu_mask_count,
-		uint32_t *se_mask)
+		uint32_t *se_mask, uint32_t inst)
 {
-	struct kfd_cu_info cu_info;
+	struct amdgpu_cu_info *cu_info = &mm->dev->adev->gfx.cu_info;
+	struct amdgpu_gfx_config *gfx_info = &mm->dev->adev->gfx.config;
 	uint32_t cu_per_sh[KFD_MAX_NUM_SE][KFD_MAX_NUM_SH_PER_SE] = {0};
-	int i, se, sh, cu;
-	amdgpu_amdkfd_get_cu_info(mm->dev->kgd, &cu_info);
+	bool wgp_mode_req = KFD_GC_VERSION(mm->dev) >= IP_VERSION(10, 0, 0);
+	uint32_t en_mask = wgp_mode_req ? 0x3 : 0x1;
+	int i, se, sh, cu, cu_bitmap_sh_mul, cu_inc = wgp_mode_req ? 2 : 1;
+	uint32_t cu_active_per_node;
+	int inc = cu_inc * NUM_XCC(mm->dev->xcc_mask);
+	int xcc_inst = inst + ffs(mm->dev->xcc_mask) - 1;
 
-	if (cu_mask_count > cu_info.cu_active_number)
-		cu_mask_count = cu_info.cu_active_number;
+	cu_active_per_node = cu_info->number / mm->dev->kfd->num_nodes;
+	if (cu_mask_count > cu_active_per_node)
+		cu_mask_count = cu_active_per_node;
 
 	/* Exceeding these bounds corrupts the stack and indicates a coding error.
 	 * Returning with no CU's enabled will hang the queue, which should be
 	 * attention grabbing.
 	 */
-	if (cu_info.num_shader_engines > KFD_MAX_NUM_SE) {
-		pr_err("Exceeded KFD_MAX_NUM_SE, chip reports %d\n", cu_info.num_shader_engines);
+	if (gfx_info->max_shader_engines > KFD_MAX_NUM_SE) {
+		dev_err(mm->dev->adev->dev,
+			"Exceeded KFD_MAX_NUM_SE, chip reports %d\n",
+			gfx_info->max_shader_engines);
 		return;
 	}
-	if (cu_info.num_shader_arrays_per_engine > KFD_MAX_NUM_SH_PER_SE) {
-		pr_err("Exceeded KFD_MAX_NUM_SH, chip reports %d\n",
-			cu_info.num_shader_arrays_per_engine * cu_info.num_shader_engines);
+	if (gfx_info->max_sh_per_se > KFD_MAX_NUM_SH_PER_SE) {
+		dev_err(mm->dev->adev->dev,
+			"Exceeded KFD_MAX_NUM_SH, chip reports %d\n",
+			gfx_info->max_sh_per_se * gfx_info->max_shader_engines);
 		return;
 	}
+
+	cu_bitmap_sh_mul = (KFD_GC_VERSION(mm->dev) >= IP_VERSION(11, 0, 0) &&
+			    KFD_GC_VERSION(mm->dev) < IP_VERSION(13, 0, 0)) ? 2 : 1;
+
 	/* Count active CUs per SH.
 	 *
 	 * Some CUs in an SH may be disabled.	HW expects disabled CUs to be
@@ -127,10 +142,13 @@ void mqd_symmetrically_map_cu_mask(struct mqd_manager *mm,
 	 * Each half of se_mask must be filled only on bits 0-cu_per_sh[se][sh]-1.
 	 *
 	 * See note on Arcturus cu_bitmap layout in gfx_v9_0_get_cu_info.
+	 * See note on GFX11 cu_bitmap layout in gfx_v11_0_get_cu_info.
 	 */
-	for (se = 0; se < cu_info.num_shader_engines; se++)
-		for (sh = 0; sh < cu_info.num_shader_arrays_per_engine; sh++)
-			cu_per_sh[se][sh] = hweight32(cu_info.cu_bitmap[se % 4][sh + (se / 4)]);
+	for (se = 0; se < gfx_info->max_shader_engines; se++)
+		for (sh = 0; sh < gfx_info->max_sh_per_se; sh++)
+			cu_per_sh[se][sh] = hweight32(
+				cu_info->bitmap[xcc_inst][se % 4][sh + (se / 4) *
+				cu_bitmap_sh_mul]);
 
 	/* Symmetrically map cu_mask to all SEs & SHs:
 	 * se_mask programs up to 2 SH in the upper and lower 16 bits.
@@ -153,23 +171,142 @@ void mqd_symmetrically_map_cu_mask(struct mqd_manager *mm,
 	 * cu_mask[0] bit8 -> se_mask[0] bit1 (SE0,SH0,CU1)
 	 * ...
 	 *
+	 * For GFX 9.4.3, the following code only looks at a
+	 * subset of the cu_mask corresponding to the inst parameter.
+	 * If we have n XCCs under one GPU node
+	 * cu_mask[0] bit0 -> XCC0 se_mask[0] bit0 (XCC0,SE0,SH0,CU0)
+	 * cu_mask[0] bit1 -> XCC1 se_mask[0] bit0 (XCC1,SE0,SH0,CU0)
+	 * ..
+	 * cu_mask[0] bitn -> XCCn se_mask[0] bit0 (XCCn,SE0,SH0,CU0)
+	 * cu_mask[0] bit n+1 -> XCC0 se_mask[1] bit0 (XCC0,SE1,SH0,CU0)
+	 *
+	 * For example, if there are 6 XCCs under 1 KFD node, this code
+	 * running for each inst, will look at the bits as:
+	 * inst, inst + 6, inst + 12...
+	 *
 	 * First ensure all CUs are disabled, then enable user specified CUs.
 	 */
-	for (i = 0; i < cu_info.num_shader_engines; i++)
+	for (i = 0; i < gfx_info->max_shader_engines; i++)
 		se_mask[i] = 0;
 
-	i = 0;
-	for (cu = 0; cu < 16; cu++) {
-		for (sh = 0; sh < cu_info.num_shader_arrays_per_engine; sh++) {
-			for (se = 0; se < cu_info.num_shader_engines; se++) {
+	i = inst;
+	for (cu = 0; cu < 16; cu += cu_inc) {
+		for (sh = 0; sh < gfx_info->max_sh_per_se; sh++) {
+			for (se = 0; se < gfx_info->max_shader_engines; se++) {
 				if (cu_per_sh[se][sh] > cu) {
-					if (cu_mask[i / 32] & (1 << (i % 32)))
-						se_mask[se] |= 1 << (cu + sh * 16);
-					i++;
-					if (i == cu_mask_count)
+					if (cu_mask[i / 32] & (en_mask << (i % 32)))
+						se_mask[se] |= en_mask << (cu + sh * 16);
+					i += inc;
+					if (i >= cu_mask_count)
 						return;
 				}
 			}
 		}
 	}
 }
+
+int kfd_hiq_load_mqd_kiq(struct mqd_manager *mm, void *mqd,
+		     uint32_t pipe_id, uint32_t queue_id,
+		     struct queue_properties *p, struct mm_struct *mms)
+{
+	return mm->dev->kfd2kgd->hiq_mqd_load(mm->dev->adev, mqd, pipe_id,
+					      queue_id, p->doorbell_off, 0);
+}
+
+int kfd_destroy_mqd_cp(struct mqd_manager *mm, void *mqd,
+		enum kfd_preempt_type type, unsigned int timeout,
+		uint32_t pipe_id, uint32_t queue_id)
+{
+	return mm->dev->kfd2kgd->hqd_destroy(mm->dev->adev, mqd, type, timeout,
+						pipe_id, queue_id, 0);
+}
+
+void kfd_free_mqd_cp(struct mqd_manager *mm, void *mqd,
+	      struct kfd_mem_obj *mqd_mem_obj)
+{
+	if (mqd_mem_obj->gtt_mem) {
+		amdgpu_amdkfd_free_gtt_mem(mm->dev->adev, &mqd_mem_obj->gtt_mem);
+		kfree(mqd_mem_obj);
+	} else {
+		kfd_gtt_sa_free(mm->dev, mqd_mem_obj);
+	}
+}
+
+bool kfd_is_occupied_cp(struct mqd_manager *mm, void *mqd,
+		 uint64_t queue_address, uint32_t pipe_id,
+		 uint32_t queue_id)
+{
+	return mm->dev->kfd2kgd->hqd_is_occupied(mm->dev->adev, queue_address,
+						pipe_id, queue_id, 0);
+}
+
+int kfd_load_mqd_sdma(struct mqd_manager *mm, void *mqd,
+		  uint32_t pipe_id, uint32_t queue_id,
+		  struct queue_properties *p, struct mm_struct *mms)
+{
+	return mm->dev->kfd2kgd->hqd_sdma_load(mm->dev->adev, mqd,
+						(uint32_t __user *)p->write_ptr,
+						mms);
+}
+
+/*
+ * preempt type here is ignored because there is only one way
+ * to preempt sdma queue
+ */
+int kfd_destroy_mqd_sdma(struct mqd_manager *mm, void *mqd,
+		     enum kfd_preempt_type type,
+		     unsigned int timeout, uint32_t pipe_id,
+		     uint32_t queue_id)
+{
+	return mm->dev->kfd2kgd->hqd_sdma_destroy(mm->dev->adev, mqd, timeout);
+}
+
+bool kfd_is_occupied_sdma(struct mqd_manager *mm, void *mqd,
+		      uint64_t queue_address, uint32_t pipe_id,
+		      uint32_t queue_id)
+{
+	return mm->dev->kfd2kgd->hqd_sdma_is_occupied(mm->dev->adev, mqd);
+}
+
+uint64_t kfd_hiq_mqd_stride(struct kfd_node *dev)
+{
+	return dev->dqm->mqd_mgrs[KFD_MQD_TYPE_HIQ]->mqd_size;
+}
+
+void kfd_get_hiq_xcc_mqd(struct kfd_node *dev, struct kfd_mem_obj *mqd_mem_obj,
+		     uint32_t virtual_xcc_id)
+{
+	uint64_t offset;
+
+	offset = kfd_hiq_mqd_stride(dev) * virtual_xcc_id;
+
+	mqd_mem_obj->gtt_mem = (virtual_xcc_id == 0) ?
+			dev->dqm->hiq_sdma_mqd.gtt_mem : NULL;
+	mqd_mem_obj->gpu_addr = dev->dqm->hiq_sdma_mqd.gpu_addr + offset;
+	mqd_mem_obj->cpu_ptr = (uint32_t *)((uintptr_t)
+				dev->dqm->hiq_sdma_mqd.cpu_ptr + offset);
+}
+
+uint64_t kfd_mqd_stride(struct mqd_manager *mm,
+			struct queue_properties *q)
+{
+	return mm->mqd_size;
+}
+
+bool kfd_check_hiq_mqd_doorbell_id(struct kfd_node *node, uint32_t doorbell_id,
+				   uint32_t inst)
+{
+	if (doorbell_id) {
+		struct device *dev = node->adev->dev;
+
+		if (node->adev->xcp_mgr && node->adev->xcp_mgr->num_xcps > 0)
+			dev_err(dev, "XCC %d: Queue preemption failed for queue with doorbell_id: %x\n",
+							inst, doorbell_id);
+		else
+			dev_err(dev, "Queue preemption failed for queue with doorbell_id: %x\n",
+							doorbell_id);
+		return true;
+	}
+
+	return false;
+}
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager.h b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager.h
index 965e17c5dbb4..17cc1f25c8d0 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager.h
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager.h
@@ -1,5 +1,6 @@
+/* SPDX-License-Identifier: GPL-2.0 OR MIT */
 /*
- * Copyright 2014 Advanced Micro Devices, Inc.
+ * Copyright 2014-2022 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -67,7 +68,7 @@
  */
 extern int pipe_priority_map[];
 struct mqd_manager {
-	struct kfd_mem_obj*	(*allocate_mqd)(struct kfd_dev *kfd,
+	struct kfd_mem_obj*	(*allocate_mqd)(struct kfd_node *kfd,
 		struct queue_properties *q);
 
 	void	(*init_mqd)(struct mqd_manager *mm, void **mqd,
@@ -96,30 +97,107 @@ struct mqd_manager {
 				uint32_t queue_id);
 
 	int	(*get_wave_state)(struct mqd_manager *mm, void *mqd,
+				  struct queue_properties *q,
 				  void __user *ctl_stack,
 				  u32 *ctl_stack_used_size,
 				  u32 *save_area_used_size);
 
+	void	(*get_checkpoint_info)(struct mqd_manager *mm, void *mqd, uint32_t *ctl_stack_size);
+
+	void	(*checkpoint_mqd)(struct mqd_manager *mm,
+				  void *mqd,
+				  void *mqd_dst,
+				  void *ctl_stack_dst);
+
+	void	(*restore_mqd)(struct mqd_manager *mm, void **mqd,
+				struct kfd_mem_obj *mqd_mem_obj, uint64_t *gart_addr,
+				struct queue_properties *p,
+				const void *mqd_src,
+				const void *ctl_stack_src,
+				const u32 ctl_stack_size);
+
 #if defined(CONFIG_DEBUG_FS)
 	int	(*debugfs_show_mqd)(struct seq_file *m, void *data);
 #endif
-	uint32_t (*read_doorbell_id)(void *mqd);
+	bool (*check_preemption_failed)(struct mqd_manager *mm, void *mqd);
+	uint64_t (*mqd_stride)(struct mqd_manager *mm,
+				struct queue_properties *p);
 
 	struct mutex	mqd_mutex;
-	struct kfd_dev	*dev;
+	struct kfd_node	*dev;
 	uint32_t mqd_size;
 };
 
-struct kfd_mem_obj *allocate_hiq_mqd(struct kfd_dev *dev,
+struct mqd_user_context_save_area_header {
+	/* Byte offset from start of user context
+	 * save area to the last saved top (lowest
+	 * address) of control stack data. Must be
+	 * 4 byte aligned.
+	 */
+	uint32_t control_stack_offset;
+
+	/* Byte size of the last saved control stack
+	 * data. Must be 4 byte aligned.
+	 */
+	uint32_t control_stack_size;
+
+	/* Byte offset from start of user context save
+	 * area to the last saved base (lowest address)
+	 * of wave state data. Must be 4 byte aligned.
+	 */
+	uint32_t wave_state_offset;
+
+	/* Byte size of the last saved wave state data.
+	 * Must be 4 byte aligned.
+	 */
+	uint32_t wave_state_size;
+};
+
+struct kfd_mem_obj *allocate_hiq_mqd(struct kfd_node *dev,
 				struct queue_properties *q);
 
-struct kfd_mem_obj *allocate_sdma_mqd(struct kfd_dev *dev,
+struct kfd_mem_obj *allocate_sdma_mqd(struct kfd_node *dev,
 					struct queue_properties *q);
 void free_mqd_hiq_sdma(struct mqd_manager *mm, void *mqd,
 				struct kfd_mem_obj *mqd_mem_obj);
 
 void mqd_symmetrically_map_cu_mask(struct mqd_manager *mm,
 		const uint32_t *cu_mask, uint32_t cu_mask_count,
-		uint32_t *se_mask);
+		uint32_t *se_mask, uint32_t inst);
+
+int kfd_hiq_load_mqd_kiq(struct mqd_manager *mm, void *mqd,
+		uint32_t pipe_id, uint32_t queue_id,
+		struct queue_properties *p, struct mm_struct *mms);
+
+int kfd_destroy_mqd_cp(struct mqd_manager *mm, void *mqd,
+		enum kfd_preempt_type type, unsigned int timeout,
+		uint32_t pipe_id, uint32_t queue_id);
 
+void kfd_free_mqd_cp(struct mqd_manager *mm, void *mqd,
+		struct kfd_mem_obj *mqd_mem_obj);
+
+bool kfd_is_occupied_cp(struct mqd_manager *mm, void *mqd,
+		 uint64_t queue_address, uint32_t pipe_id,
+		 uint32_t queue_id);
+
+int kfd_load_mqd_sdma(struct mqd_manager *mm, void *mqd,
+		uint32_t pipe_id, uint32_t queue_id,
+		struct queue_properties *p, struct mm_struct *mms);
+
+int kfd_destroy_mqd_sdma(struct mqd_manager *mm, void *mqd,
+		enum kfd_preempt_type type, unsigned int timeout,
+		uint32_t pipe_id, uint32_t queue_id);
+
+bool kfd_is_occupied_sdma(struct mqd_manager *mm, void *mqd,
+		uint64_t queue_address, uint32_t pipe_id,
+		uint32_t queue_id);
+
+void kfd_get_hiq_xcc_mqd(struct kfd_node *dev,
+		struct kfd_mem_obj *mqd_mem_obj, uint32_t virtual_xcc_id);
+
+uint64_t kfd_hiq_mqd_stride(struct kfd_node *dev);
+uint64_t kfd_mqd_stride(struct mqd_manager *mm,
+			struct queue_properties *q);
+bool kfd_check_hiq_mqd_doorbell_id(struct kfd_node *node, uint32_t doorbell_id,
+				   uint32_t inst);
 #endif /* KFD_MQD_MANAGER_H_ */
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_cik.c b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_cik.c
index 8128f4d312f1..05f3ac2eaef9 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_cik.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_cik.c
@@ -1,5 +1,6 @@
+// SPDX-License-Identifier: GPL-2.0 OR MIT
 /*
- * Copyright 2014 Advanced Micro Devices, Inc.
+ * Copyright 2014-2022 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -47,12 +48,11 @@ static void update_cu_mask(struct mqd_manager *mm, void *mqd,
 	struct cik_mqd *m;
 	uint32_t se_mask[4] = {0}; /* 4 is the max # of SEs */
 
-	if (!minfo || (minfo->update_flag != UPDATE_FLAG_CU_MASK) ||
-	    !minfo->cu_mask.ptr)
+	if (!minfo || !minfo->cu_mask.ptr)
 		return;
 
 	mqd_symmetrically_map_cu_mask(mm,
-		minfo->cu_mask.ptr, minfo->cu_mask.count, se_mask);
+		minfo->cu_mask.ptr, minfo->cu_mask.count, se_mask, 0);
 
 	m = get_mqd(mqd);
 	m->compute_static_thread_mgmt_se0 = se_mask[0];
@@ -73,7 +73,7 @@ static void set_priority(struct cik_mqd *m, struct queue_properties *q)
 	m->cp_hqd_queue_priority = q->priority;
 }
 
-static struct kfd_mem_obj *allocate_mqd(struct kfd_dev *kfd,
+static struct kfd_mem_obj *allocate_mqd(struct kfd_node *kfd,
 					struct queue_properties *q)
 {
 	struct kfd_mem_obj *mqd_mem_obj;
@@ -156,13 +156,6 @@ static void init_mqd_sdma(struct mqd_manager *mm, void **mqd,
 	mm->update_mqd(mm, m, q, NULL);
 }
 
-static void free_mqd(struct mqd_manager *mm, void *mqd,
-			struct kfd_mem_obj *mqd_mem_obj)
-{
-	kfd_gtt_sa_free(mm->dev, mqd_mem_obj);
-}
-
-
 static int load_mqd(struct mqd_manager *mm, void *mqd, uint32_t pipe_id,
 		    uint32_t queue_id, struct queue_properties *p,
 		    struct mm_struct *mms)
@@ -171,18 +164,9 @@ static int load_mqd(struct mqd_manager *mm, void *mqd, uint32_t pipe_id,
 	uint32_t wptr_shift = (p->format == KFD_QUEUE_FORMAT_AQL ? 4 : 0);
 	uint32_t wptr_mask = (uint32_t)((p->queue_size / 4) - 1);
 
-	return mm->dev->kfd2kgd->hqd_load(mm->dev->kgd, mqd, pipe_id, queue_id,
+	return mm->dev->kfd2kgd->hqd_load(mm->dev->adev, mqd, pipe_id, queue_id,
 					  (uint32_t __user *)p->write_ptr,
-					  wptr_shift, wptr_mask, mms);
-}
-
-static int load_mqd_sdma(struct mqd_manager *mm, void *mqd,
-			 uint32_t pipe_id, uint32_t queue_id,
-			 struct queue_properties *p, struct mm_struct *mms)
-{
-	return mm->dev->kfd2kgd->hqd_sdma_load(mm->dev->kgd, mqd,
-					       (uint32_t __user *)p->write_ptr,
-					       mms);
+					  wptr_shift, wptr_mask, mms, 0);
 }
 
 static void __update_mqd(struct mqd_manager *mm, void *mqd,
@@ -222,23 +206,16 @@ static void __update_mqd(struct mqd_manager *mm, void *mqd,
 	q->is_active = QUEUE_IS_ACTIVE(*q);
 }
 
-static void update_mqd(struct mqd_manager *mm, void *mqd,
-			struct queue_properties *q,
-			struct mqd_update_info *minfo)
-{
-	__update_mqd(mm, mqd, q, minfo, 1);
-}
-
-static uint32_t read_doorbell_id(void *mqd)
+static bool check_preemption_failed(struct mqd_manager *mm, void *mqd)
 {
 	struct cik_mqd *m = (struct cik_mqd *)mqd;
 
-	return m->queue_doorbell_id0;
+	return kfd_check_hiq_mqd_doorbell_id(mm->dev, m->queue_doorbell_id0, 0);
 }
 
-static void update_mqd_hawaii(struct mqd_manager *mm, void *mqd,
-			struct queue_properties *q,
-			struct mqd_update_info *minfo)
+static void update_mqd(struct mqd_manager *mm, void *mqd,
+		       struct queue_properties *q,
+		       struct mqd_update_info *minfo)
 {
 	__update_mqd(mm, mqd, q, minfo, 0);
 }
@@ -271,42 +248,75 @@ static void update_mqd_sdma(struct mqd_manager *mm, void *mqd,
 	q->is_active = QUEUE_IS_ACTIVE(*q);
 }
 
-static int destroy_mqd(struct mqd_manager *mm, void *mqd,
-			enum kfd_preempt_type type,
-			unsigned int timeout, uint32_t pipe_id,
-			uint32_t queue_id)
+static void checkpoint_mqd(struct mqd_manager *mm, void *mqd, void *mqd_dst, void *ctl_stack_dst)
 {
-	return mm->dev->kfd2kgd->hqd_destroy(mm->dev->kgd, mqd, type, timeout,
-					pipe_id, queue_id);
+	struct cik_mqd *m;
+
+	m = get_mqd(mqd);
+
+	memcpy(mqd_dst, m, sizeof(struct cik_mqd));
 }
 
-/*
- * preempt type here is ignored because there is only one way
- * to preempt sdma queue
- */
-static int destroy_mqd_sdma(struct mqd_manager *mm, void *mqd,
-				enum kfd_preempt_type type,
-				unsigned int timeout, uint32_t pipe_id,
-				uint32_t queue_id)
+static void restore_mqd(struct mqd_manager *mm, void **mqd,
+			struct kfd_mem_obj *mqd_mem_obj, uint64_t *gart_addr,
+			struct queue_properties *qp,
+			const void *mqd_src,
+			const void *ctl_stack_src, const u32 ctl_stack_size)
 {
-	return mm->dev->kfd2kgd->hqd_sdma_destroy(mm->dev->kgd, mqd, timeout);
+	uint64_t addr;
+	struct cik_mqd *m;
+
+	m = (struct cik_mqd *) mqd_mem_obj->cpu_ptr;
+	addr = mqd_mem_obj->gpu_addr;
+
+	memcpy(m, mqd_src, sizeof(*m));
+
+	*mqd = m;
+	if (gart_addr)
+		*gart_addr = addr;
+
+	m->cp_hqd_pq_doorbell_control = DOORBELL_OFFSET(qp->doorbell_off);
+
+	pr_debug("cp_hqd_pq_doorbell_control 0x%x\n",
+			m->cp_hqd_pq_doorbell_control);
+
+	qp->is_active = 0;
 }
 
-static bool is_occupied(struct mqd_manager *mm, void *mqd,
-			uint64_t queue_address,	uint32_t pipe_id,
-			uint32_t queue_id)
+static void checkpoint_mqd_sdma(struct mqd_manager *mm,
+				void *mqd,
+				void *mqd_dst,
+				void *ctl_stack_dst)
 {
+	struct cik_sdma_rlc_registers *m;
 
-	return mm->dev->kfd2kgd->hqd_is_occupied(mm->dev->kgd, queue_address,
-					pipe_id, queue_id);
+	m = get_sdma_mqd(mqd);
 
+	memcpy(mqd_dst, m, sizeof(struct cik_sdma_rlc_registers));
 }
 
-static bool is_occupied_sdma(struct mqd_manager *mm, void *mqd,
-			uint64_t queue_address,	uint32_t pipe_id,
-			uint32_t queue_id)
+static void restore_mqd_sdma(struct mqd_manager *mm, void **mqd,
+				struct kfd_mem_obj *mqd_mem_obj, uint64_t *gart_addr,
+				struct queue_properties *qp,
+				const void *mqd_src,
+				const void *ctl_stack_src, const u32 ctl_stack_size)
 {
-	return mm->dev->kfd2kgd->hqd_sdma_is_occupied(mm->dev->kgd, mqd);
+	uint64_t addr;
+	struct cik_sdma_rlc_registers *m;
+
+	m = (struct cik_sdma_rlc_registers *) mqd_mem_obj->cpu_ptr;
+	addr = mqd_mem_obj->gpu_addr;
+
+	memcpy(m, mqd_src, sizeof(*m));
+
+	m->sdma_rlc_doorbell =
+		qp->doorbell_off << SDMA0_RLC0_DOORBELL__OFFSET__SHIFT;
+
+	*mqd = m;
+	if (gart_addr)
+		*gart_addr = addr;
+
+	qp->is_active = 0;
 }
 
 /*
@@ -370,9 +380,8 @@ static int debugfs_show_mqd_sdma(struct seq_file *m, void *data)
 
 #endif
 
-
 struct mqd_manager *mqd_manager_init_cik(enum KFD_MQD_TYPE type,
-		struct kfd_dev *dev)
+		struct kfd_node *dev)
 {
 	struct mqd_manager *mqd;
 
@@ -389,11 +398,13 @@ struct mqd_manager *mqd_manager_init_cik(enum KFD_MQD_TYPE type,
 	case KFD_MQD_TYPE_CP:
 		mqd->allocate_mqd = allocate_mqd;
 		mqd->init_mqd = init_mqd;
-		mqd->free_mqd = free_mqd;
+		mqd->free_mqd = kfd_free_mqd_cp;
 		mqd->load_mqd = load_mqd;
 		mqd->update_mqd = update_mqd;
-		mqd->destroy_mqd = destroy_mqd;
-		mqd->is_occupied = is_occupied;
+		mqd->destroy_mqd = kfd_destroy_mqd_cp;
+		mqd->is_occupied = kfd_is_occupied_cp;
+		mqd->checkpoint_mqd = checkpoint_mqd;
+		mqd->restore_mqd = restore_mqd;
 		mqd->mqd_size = sizeof(struct cik_mqd);
 #if defined(CONFIG_DEBUG_FS)
 		mqd->debugfs_show_mqd = debugfs_show_mqd;
@@ -405,23 +416,25 @@ struct mqd_manager *mqd_manager_init_cik(enum KFD_MQD_TYPE type,
 		mqd->free_mqd = free_mqd_hiq_sdma;
 		mqd->load_mqd = load_mqd;
 		mqd->update_mqd = update_mqd_hiq;
-		mqd->destroy_mqd = destroy_mqd;
-		mqd->is_occupied = is_occupied;
+		mqd->destroy_mqd = kfd_destroy_mqd_cp;
+		mqd->is_occupied = kfd_is_occupied_cp;
 		mqd->mqd_size = sizeof(struct cik_mqd);
+		mqd->mqd_stride = kfd_mqd_stride;
 #if defined(CONFIG_DEBUG_FS)
 		mqd->debugfs_show_mqd = debugfs_show_mqd;
 #endif
-		mqd->read_doorbell_id = read_doorbell_id;
+		mqd->check_preemption_failed = check_preemption_failed;
 		break;
 	case KFD_MQD_TYPE_DIQ:
 		mqd->allocate_mqd = allocate_mqd;
 		mqd->init_mqd = init_mqd_hiq;
-		mqd->free_mqd = free_mqd;
+		mqd->free_mqd = kfd_free_mqd_cp;
 		mqd->load_mqd = load_mqd;
 		mqd->update_mqd = update_mqd_hiq;
-		mqd->destroy_mqd = destroy_mqd;
-		mqd->is_occupied = is_occupied;
+		mqd->destroy_mqd = kfd_destroy_mqd_cp;
+		mqd->is_occupied = kfd_is_occupied_cp;
 		mqd->mqd_size = sizeof(struct cik_mqd);
+		mqd->mqd_stride = kfd_mqd_stride;
 #if defined(CONFIG_DEBUG_FS)
 		mqd->debugfs_show_mqd = debugfs_show_mqd;
 #endif
@@ -430,11 +443,14 @@ struct mqd_manager *mqd_manager_init_cik(enum KFD_MQD_TYPE type,
 		mqd->allocate_mqd = allocate_sdma_mqd;
 		mqd->init_mqd = init_mqd_sdma;
 		mqd->free_mqd = free_mqd_hiq_sdma;
-		mqd->load_mqd = load_mqd_sdma;
+		mqd->load_mqd = kfd_load_mqd_sdma;
 		mqd->update_mqd = update_mqd_sdma;
-		mqd->destroy_mqd = destroy_mqd_sdma;
-		mqd->is_occupied = is_occupied_sdma;
+		mqd->destroy_mqd = kfd_destroy_mqd_sdma;
+		mqd->is_occupied = kfd_is_occupied_sdma;
+		mqd->checkpoint_mqd = checkpoint_mqd_sdma;
+		mqd->restore_mqd = restore_mqd_sdma;
 		mqd->mqd_size = sizeof(struct cik_sdma_rlc_registers);
+		mqd->mqd_stride = kfd_mqd_stride;
 #if defined(CONFIG_DEBUG_FS)
 		mqd->debugfs_show_mqd = debugfs_show_mqd_sdma;
 #endif
@@ -446,16 +462,3 @@ struct mqd_manager *mqd_manager_init_cik(enum KFD_MQD_TYPE type,
 
 	return mqd;
 }
-
-struct mqd_manager *mqd_manager_init_cik_hawaii(enum KFD_MQD_TYPE type,
-			struct kfd_dev *dev)
-{
-	struct mqd_manager *mqd;
-
-	mqd = mqd_manager_init_cik(type, dev);
-	if (!mqd)
-		return NULL;
-	if (type == KFD_MQD_TYPE_CP)
-		mqd->update_mqd = update_mqd_hawaii;
-	return mqd;
-}
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v10.c b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v10.c
index 270160fc401b..1695dd78ede8 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v10.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v10.c
@@ -1,5 +1,6 @@
+// SPDX-License-Identifier: GPL-2.0 OR MIT
 /*
- * Copyright 2018 Advanced Micro Devices, Inc.
+ * Copyright 2018-2022 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -47,12 +48,11 @@ static void update_cu_mask(struct mqd_manager *mm, void *mqd,
 	struct v10_compute_mqd *m;
 	uint32_t se_mask[4] = {0}; /* 4 is the max # of SEs */
 
-	if (!minfo || (minfo->update_flag != UPDATE_FLAG_CU_MASK) ||
-	    !minfo->cu_mask.ptr)
+	if (!minfo || !minfo->cu_mask.ptr)
 		return;
 
 	mqd_symmetrically_map_cu_mask(mm,
-		minfo->cu_mask.ptr, minfo->cu_mask.count, se_mask);
+		minfo->cu_mask.ptr, minfo->cu_mask.count, se_mask, 0);
 
 	m = get_mqd(mqd);
 	m->compute_static_thread_mgmt_se0 = se_mask[0];
@@ -73,7 +73,7 @@ static void set_priority(struct v10_compute_mqd *m, struct queue_properties *q)
 	m->cp_hqd_queue_priority = q->priority;
 }
 
-static struct kfd_mem_obj *allocate_mqd(struct kfd_dev *kfd,
+static struct kfd_mem_obj *allocate_mqd(struct kfd_node *kfd,
 		struct queue_properties *q)
 {
 	struct kfd_mem_obj *mqd_mem_obj;
@@ -107,6 +107,8 @@ static void init_mqd(struct mqd_manager *mm, void **mqd,
 	m->cp_hqd_persistent_state = CP_HQD_PERSISTENT_STATE__PRELOAD_REQ_MASK |
 			0x53 << CP_HQD_PERSISTENT_STATE__PRELOAD_SIZE__SHIFT;
 
+	m->cp_hqd_pq_control = 5 << CP_HQD_PQ_CONTROL__RPTR_BLOCK_SIZE__SHIFT;
+	m->cp_hqd_pq_control |= CP_HQD_PQ_CONTROL__UNORD_DISPATCH_MASK;
 	m->cp_mqd_control = 1 << CP_MQD_CONTROL__PRIV_STATE__SHIFT;
 
 	m->cp_mqd_base_addr_lo        = lower_32_bits(addr);
@@ -116,12 +118,17 @@ static void init_mqd(struct mqd_manager *mm, void **mqd,
 			1 << CP_HQD_QUANTUM__QUANTUM_SCALE__SHIFT |
 			1 << CP_HQD_QUANTUM__QUANTUM_DURATION__SHIFT;
 
+	/* Set cp_hqd_hq_scheduler0 bit 14 to 1 to have the CP set up the
+	 * DISPATCH_PTR.  This is required for the kfd debugger
+	 */
+	m->cp_hqd_hq_scheduler0 = 1 << 14;
+
 	if (q->format == KFD_QUEUE_FORMAT_AQL) {
 		m->cp_hqd_aql_control =
 			1 << CP_HQD_AQL_CONTROL__CONTROL0__SHIFT;
 	}
 
-	if (mm->dev->cwsr_enabled) {
+	if (mm->dev->kfd->cwsr_enabled) {
 		m->cp_hqd_persistent_state |=
 			(1 << CP_HQD_PERSISTENT_STATE__QSWITCH_MODE__SHIFT);
 		m->cp_hqd_ctx_save_base_addr_lo =
@@ -148,20 +155,12 @@ static int load_mqd(struct mqd_manager *mm, void *mqd,
 	/* AQL write pointer counts in 64B packets, PM4/CP counts in dwords. */
 	uint32_t wptr_shift = (p->format == KFD_QUEUE_FORMAT_AQL ? 4 : 0);
 
-	r = mm->dev->kfd2kgd->hqd_load(mm->dev->kgd, mqd, pipe_id, queue_id,
+	r = mm->dev->kfd2kgd->hqd_load(mm->dev->adev, mqd, pipe_id, queue_id,
 					  (uint32_t __user *)p->write_ptr,
-					  wptr_shift, 0, mms);
+					  wptr_shift, 0, mms, 0);
 	return r;
 }
 
-static int hiq_load_mqd_kiq(struct mqd_manager *mm, void *mqd,
-			    uint32_t pipe_id, uint32_t queue_id,
-			    struct queue_properties *p, struct mm_struct *mms)
-{
-	return mm->dev->kfd2kgd->hiq_mqd_load(mm->dev->kgd, mqd, pipe_id,
-					      queue_id, p->doorbell_off);
-}
-
 static void update_mqd(struct mqd_manager *mm, void *mqd,
 			struct queue_properties *q,
 			struct mqd_update_info *minfo)
@@ -170,9 +169,10 @@ static void update_mqd(struct mqd_manager *mm, void *mqd,
 
 	m = get_mqd(mqd);
 
-	m->cp_hqd_pq_control = 5 << CP_HQD_PQ_CONTROL__RPTR_BLOCK_SIZE__SHIFT;
+	m->cp_hqd_pq_control &= ~CP_HQD_PQ_CONTROL__QUEUE_SIZE_MASK;
 	m->cp_hqd_pq_control |=
 			ffs(q->queue_size / sizeof(unsigned int)) - 1 - 1;
+
 	pr_debug("cp_hqd_pq_control 0x%x\n", m->cp_hqd_pq_control);
 
 	m->cp_hqd_pq_base_lo = lower_32_bits((uint64_t)q->queue_address >> 8);
@@ -213,11 +213,11 @@ static void update_mqd(struct mqd_manager *mm, void *mqd,
 		/* GC 10 removed WPP_CLAMP from PQ Control */
 		m->cp_hqd_pq_control |= CP_HQD_PQ_CONTROL__NO_UPDATE_RPTR_MASK |
 				2 << CP_HQD_PQ_CONTROL__SLOT_BASED_WPTR__SHIFT |
-				1 << CP_HQD_PQ_CONTROL__QUEUE_FULL_EN__SHIFT ;
+				1 << CP_HQD_PQ_CONTROL__QUEUE_FULL_EN__SHIFT;
 		m->cp_hqd_pq_doorbell_control |=
 			1 << CP_HQD_PQ_DOORBELL_CONTROL__DOORBELL_BIF_DROP__SHIFT;
 	}
-	if (mm->dev->cwsr_enabled)
+	if (mm->dev->kfd->cwsr_enabled)
 		m->cp_hqd_ctx_save_control = 0;
 
 	update_cu_mask(mm, mqd, minfo);
@@ -226,44 +226,21 @@ static void update_mqd(struct mqd_manager *mm, void *mqd,
 	q->is_active = QUEUE_IS_ACTIVE(*q);
 }
 
-static uint32_t read_doorbell_id(void *mqd)
+static bool check_preemption_failed(struct mqd_manager *mm, void *mqd)
 {
 	struct v10_compute_mqd *m = (struct v10_compute_mqd *)mqd;
 
-	return m->queue_doorbell_id0;
-}
-
-static int destroy_mqd(struct mqd_manager *mm, void *mqd,
-		       enum kfd_preempt_type type,
-		       unsigned int timeout, uint32_t pipe_id,
-		       uint32_t queue_id)
-{
-	return mm->dev->kfd2kgd->hqd_destroy
-		(mm->dev->kgd, mqd, type, timeout,
-		 pipe_id, queue_id);
-}
-
-static void free_mqd(struct mqd_manager *mm, void *mqd,
-			struct kfd_mem_obj *mqd_mem_obj)
-{
-	kfd_gtt_sa_free(mm->dev, mqd_mem_obj);
-}
-
-static bool is_occupied(struct mqd_manager *mm, void *mqd,
-			uint64_t queue_address,	uint32_t pipe_id,
-			uint32_t queue_id)
-{
-	return mm->dev->kfd2kgd->hqd_is_occupied(
-		mm->dev->kgd, queue_address,
-		pipe_id, queue_id);
+	return kfd_check_hiq_mqd_doorbell_id(mm->dev, m->queue_doorbell_id0, 0);
 }
 
 static int get_wave_state(struct mqd_manager *mm, void *mqd,
+			  struct queue_properties *q,
 			  void __user *ctl_stack,
 			  u32 *ctl_stack_used_size,
 			  u32 *save_area_used_size)
 {
 	struct v10_compute_mqd *m;
+	struct kfd_context_save_area_header header;
 
 	m = get_mqd(mqd);
 
@@ -282,9 +259,54 @@ static int get_wave_state(struct mqd_manager *mm, void *mqd,
 	 * accessible to user mode
 	 */
 
+	header.wave_state.control_stack_size = *ctl_stack_used_size;
+	header.wave_state.wave_state_size = *save_area_used_size;
+
+	header.wave_state.wave_state_offset = m->cp_hqd_wg_state_offset;
+	header.wave_state.control_stack_offset = m->cp_hqd_cntl_stack_offset;
+
+	if (copy_to_user(ctl_stack, &header, sizeof(header.wave_state)))
+		return -EFAULT;
+
 	return 0;
 }
 
+static void checkpoint_mqd(struct mqd_manager *mm, void *mqd, void *mqd_dst, void *ctl_stack_dst)
+{
+	struct v10_compute_mqd *m;
+
+	m = get_mqd(mqd);
+
+	memcpy(mqd_dst, m, sizeof(struct v10_compute_mqd));
+}
+
+static void restore_mqd(struct mqd_manager *mm, void **mqd,
+			struct kfd_mem_obj *mqd_mem_obj, uint64_t *gart_addr,
+			struct queue_properties *qp,
+			const void *mqd_src,
+			const void *ctl_stack_src, const u32 ctl_stack_size)
+{
+	uint64_t addr;
+	struct v10_compute_mqd *m;
+
+	m = (struct v10_compute_mqd *) mqd_mem_obj->cpu_ptr;
+	addr = mqd_mem_obj->gpu_addr;
+
+	memcpy(m, mqd_src, sizeof(*m));
+
+	*mqd = m;
+	if (gart_addr)
+		*gart_addr = addr;
+
+	m->cp_hqd_pq_doorbell_control =
+		qp->doorbell_off <<
+			CP_HQD_PQ_DOORBELL_CONTROL__DOORBELL_OFFSET__SHIFT;
+	pr_debug("cp_hqd_pq_doorbell_control 0x%x\n",
+			m->cp_hqd_pq_doorbell_control);
+
+	qp->is_active = 0;
+}
+
 static void init_mqd_hiq(struct mqd_manager *mm, void **mqd,
 			struct kfd_mem_obj *mqd_mem_obj, uint64_t *gart_addr,
 			struct queue_properties *q)
@@ -299,6 +321,26 @@ static void init_mqd_hiq(struct mqd_manager *mm, void **mqd,
 			1 << CP_HQD_PQ_CONTROL__KMD_QUEUE__SHIFT;
 }
 
+static int destroy_hiq_mqd(struct mqd_manager *mm, void *mqd,
+			enum kfd_preempt_type type, unsigned int timeout,
+			uint32_t pipe_id, uint32_t queue_id)
+{
+	int err;
+	struct v10_compute_mqd *m;
+	u32 doorbell_off;
+
+	m = get_mqd(mqd);
+
+	doorbell_off = m->cp_hqd_pq_doorbell_control >>
+			CP_HQD_PQ_DOORBELL_CONTROL__DOORBELL_OFFSET__SHIFT;
+
+	err = amdgpu_amdkfd_unmap_hiq(mm->dev->adev, doorbell_off, 0);
+	if (err)
+		pr_debug("Destroy HIQ MQD failed: %d\n", err);
+
+	return err;
+}
+
 static void init_mqd_sdma(struct mqd_manager *mm, void **mqd,
 		struct kfd_mem_obj *mqd_mem_obj, uint64_t *gart_addr,
 		struct queue_properties *q)
@@ -316,15 +358,6 @@ static void init_mqd_sdma(struct mqd_manager *mm, void **mqd,
 	mm->update_mqd(mm, m, q, NULL);
 }
 
-static int load_mqd_sdma(struct mqd_manager *mm, void *mqd,
-		uint32_t pipe_id, uint32_t queue_id,
-		struct queue_properties *p, struct mm_struct *mms)
-{
-	return mm->dev->kfd2kgd->hqd_sdma_load(mm->dev->kgd, mqd,
-					       (uint32_t __user *)p->write_ptr,
-					       mms);
-}
-
 #define SDMA_RLC_DUMMY_DEFAULT 0xf
 
 static void update_mqd_sdma(struct mqd_manager *mm, void *mqd,
@@ -354,23 +387,41 @@ static void update_mqd_sdma(struct mqd_manager *mm, void *mqd,
 	q->is_active = QUEUE_IS_ACTIVE(*q);
 }
 
-/*
- *  * preempt type here is ignored because there is only one way
- *  * to preempt sdma queue
- */
-static int destroy_mqd_sdma(struct mqd_manager *mm, void *mqd,
-		enum kfd_preempt_type type,
-		unsigned int timeout, uint32_t pipe_id,
-		uint32_t queue_id)
+static void checkpoint_mqd_sdma(struct mqd_manager *mm,
+				void *mqd,
+				void *mqd_dst,
+				void *ctl_stack_dst)
 {
-	return mm->dev->kfd2kgd->hqd_sdma_destroy(mm->dev->kgd, mqd, timeout);
+	struct v10_sdma_mqd *m;
+
+	m = get_sdma_mqd(mqd);
+
+	memcpy(mqd_dst, m, sizeof(struct v10_sdma_mqd));
 }
 
-static bool is_occupied_sdma(struct mqd_manager *mm, void *mqd,
-		uint64_t queue_address, uint32_t pipe_id,
-		uint32_t queue_id)
+static void restore_mqd_sdma(struct mqd_manager *mm, void **mqd,
+			     struct kfd_mem_obj *mqd_mem_obj, uint64_t *gart_addr,
+			     struct queue_properties *qp,
+			     const void *mqd_src,
+			     const void *ctl_stack_src,
+			     const u32 ctl_stack_size)
 {
-	return mm->dev->kfd2kgd->hqd_sdma_is_occupied(mm->dev->kgd, mqd);
+	uint64_t addr;
+	struct v10_sdma_mqd *m;
+
+	m = (struct v10_sdma_mqd *) mqd_mem_obj->cpu_ptr;
+	addr = mqd_mem_obj->gpu_addr;
+
+	memcpy(m, mqd_src, sizeof(*m));
+
+	m->sdmax_rlcx_doorbell_offset =
+		qp->doorbell_off << SDMA0_RLC0_DOORBELL_OFFSET__OFFSET__SHIFT;
+
+	*mqd = m;
+	if (gart_addr)
+		*gart_addr = addr;
+
+	qp->is_active = 0;
 }
 
 #if defined(CONFIG_DEBUG_FS)
@@ -392,7 +443,7 @@ static int debugfs_show_mqd_sdma(struct seq_file *m, void *data)
 #endif
 
 struct mqd_manager *mqd_manager_init_v10(enum KFD_MQD_TYPE type,
-		struct kfd_dev *dev)
+		struct kfd_node *dev)
 {
 	struct mqd_manager *mqd;
 
@@ -410,13 +461,16 @@ struct mqd_manager *mqd_manager_init_v10(enum KFD_MQD_TYPE type,
 		pr_debug("%s@%i\n", __func__, __LINE__);
 		mqd->allocate_mqd = allocate_mqd;
 		mqd->init_mqd = init_mqd;
-		mqd->free_mqd = free_mqd;
+		mqd->free_mqd = kfd_free_mqd_cp;
 		mqd->load_mqd = load_mqd;
 		mqd->update_mqd = update_mqd;
-		mqd->destroy_mqd = destroy_mqd;
-		mqd->is_occupied = is_occupied;
+		mqd->destroy_mqd = kfd_destroy_mqd_cp;
+		mqd->is_occupied = kfd_is_occupied_cp;
 		mqd->mqd_size = sizeof(struct v10_compute_mqd);
 		mqd->get_wave_state = get_wave_state;
+		mqd->checkpoint_mqd = checkpoint_mqd;
+		mqd->restore_mqd = restore_mqd;
+		mqd->mqd_stride = kfd_mqd_stride;
 #if defined(CONFIG_DEBUG_FS)
 		mqd->debugfs_show_mqd = debugfs_show_mqd;
 #endif
@@ -427,25 +481,26 @@ struct mqd_manager *mqd_manager_init_v10(enum KFD_MQD_TYPE type,
 		mqd->allocate_mqd = allocate_hiq_mqd;
 		mqd->init_mqd = init_mqd_hiq;
 		mqd->free_mqd = free_mqd_hiq_sdma;
-		mqd->load_mqd = hiq_load_mqd_kiq;
+		mqd->load_mqd = kfd_hiq_load_mqd_kiq;
 		mqd->update_mqd = update_mqd;
-		mqd->destroy_mqd = destroy_mqd;
-		mqd->is_occupied = is_occupied;
+		mqd->destroy_mqd = destroy_hiq_mqd;
+		mqd->is_occupied = kfd_is_occupied_cp;
 		mqd->mqd_size = sizeof(struct v10_compute_mqd);
+		mqd->mqd_stride = kfd_mqd_stride;
 #if defined(CONFIG_DEBUG_FS)
 		mqd->debugfs_show_mqd = debugfs_show_mqd;
 #endif
-		mqd->read_doorbell_id = read_doorbell_id;
+		mqd->check_preemption_failed = check_preemption_failed;
 		pr_debug("%s@%i\n", __func__, __LINE__);
 		break;
 	case KFD_MQD_TYPE_DIQ:
 		mqd->allocate_mqd = allocate_mqd;
 		mqd->init_mqd = init_mqd_hiq;
-		mqd->free_mqd = free_mqd;
+		mqd->free_mqd = kfd_free_mqd_cp;
 		mqd->load_mqd = load_mqd;
 		mqd->update_mqd = update_mqd;
-		mqd->destroy_mqd = destroy_mqd;
-		mqd->is_occupied = is_occupied;
+		mqd->destroy_mqd = kfd_destroy_mqd_cp;
+		mqd->is_occupied = kfd_is_occupied_cp;
 		mqd->mqd_size = sizeof(struct v10_compute_mqd);
 #if defined(CONFIG_DEBUG_FS)
 		mqd->debugfs_show_mqd = debugfs_show_mqd;
@@ -456,11 +511,14 @@ struct mqd_manager *mqd_manager_init_v10(enum KFD_MQD_TYPE type,
 		mqd->allocate_mqd = allocate_sdma_mqd;
 		mqd->init_mqd = init_mqd_sdma;
 		mqd->free_mqd = free_mqd_hiq_sdma;
-		mqd->load_mqd = load_mqd_sdma;
+		mqd->load_mqd = kfd_load_mqd_sdma;
 		mqd->update_mqd = update_mqd_sdma;
-		mqd->destroy_mqd = destroy_mqd_sdma;
-		mqd->is_occupied = is_occupied_sdma;
+		mqd->destroy_mqd = kfd_destroy_mqd_sdma;
+		mqd->is_occupied = kfd_is_occupied_sdma;
+		mqd->checkpoint_mqd = checkpoint_mqd_sdma;
+		mqd->restore_mqd = restore_mqd_sdma;
 		mqd->mqd_size = sizeof(struct v10_sdma_mqd);
+		mqd->mqd_stride = kfd_mqd_stride;
 #if defined(CONFIG_DEBUG_FS)
 		mqd->debugfs_show_mqd = debugfs_show_mqd_sdma;
 #endif
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v11.c b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v11.c
new file mode 100644
index 000000000000..3c0ae28c5923
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v11.c
@@ -0,0 +1,569 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/printk.h>
+#include <linux/slab.h>
+#include <linux/uaccess.h>
+#include "kfd_priv.h"
+#include "kfd_mqd_manager.h"
+#include "v11_structs.h"
+#include "gc/gc_11_0_0_offset.h"
+#include "gc/gc_11_0_0_sh_mask.h"
+#include "amdgpu_amdkfd.h"
+
+static inline struct v11_compute_mqd *get_mqd(void *mqd)
+{
+	return (struct v11_compute_mqd *)mqd;
+}
+
+static inline struct v11_sdma_mqd *get_sdma_mqd(void *mqd)
+{
+	return (struct v11_sdma_mqd *)mqd;
+}
+
+static void update_cu_mask(struct mqd_manager *mm, void *mqd,
+			   struct mqd_update_info *minfo)
+{
+	struct v11_compute_mqd *m;
+	uint32_t se_mask[KFD_MAX_NUM_SE] = {0};
+	bool has_wa_flag = minfo && (minfo->update_flag & (UPDATE_FLAG_DBG_WA_ENABLE |
+			UPDATE_FLAG_DBG_WA_DISABLE));
+
+	if (!minfo || !(has_wa_flag || minfo->cu_mask.ptr))
+		return;
+
+	m = get_mqd(mqd);
+
+	if (has_wa_flag) {
+		uint32_t wa_mask =
+			(minfo->update_flag & UPDATE_FLAG_DBG_WA_ENABLE) ? 0xffff : 0xffffffff;
+
+		m->compute_static_thread_mgmt_se0 = wa_mask;
+		m->compute_static_thread_mgmt_se1 = wa_mask;
+		m->compute_static_thread_mgmt_se2 = wa_mask;
+		m->compute_static_thread_mgmt_se3 = wa_mask;
+		m->compute_static_thread_mgmt_se4 = wa_mask;
+		m->compute_static_thread_mgmt_se5 = wa_mask;
+		m->compute_static_thread_mgmt_se6 = wa_mask;
+		m->compute_static_thread_mgmt_se7 = wa_mask;
+
+		return;
+	}
+
+	mqd_symmetrically_map_cu_mask(mm,
+		minfo->cu_mask.ptr, minfo->cu_mask.count, se_mask, 0);
+
+	m->compute_static_thread_mgmt_se0 = se_mask[0];
+	m->compute_static_thread_mgmt_se1 = se_mask[1];
+	m->compute_static_thread_mgmt_se2 = se_mask[2];
+	m->compute_static_thread_mgmt_se3 = se_mask[3];
+	m->compute_static_thread_mgmt_se4 = se_mask[4];
+	m->compute_static_thread_mgmt_se5 = se_mask[5];
+	m->compute_static_thread_mgmt_se6 = se_mask[6];
+	m->compute_static_thread_mgmt_se7 = se_mask[7];
+
+	pr_debug("update cu mask to %#x %#x %#x %#x %#x %#x %#x %#x\n",
+		m->compute_static_thread_mgmt_se0,
+		m->compute_static_thread_mgmt_se1,
+		m->compute_static_thread_mgmt_se2,
+		m->compute_static_thread_mgmt_se3,
+		m->compute_static_thread_mgmt_se4,
+		m->compute_static_thread_mgmt_se5,
+		m->compute_static_thread_mgmt_se6,
+		m->compute_static_thread_mgmt_se7);
+}
+
+static void set_priority(struct v11_compute_mqd *m, struct queue_properties *q)
+{
+	m->cp_hqd_pipe_priority = pipe_priority_map[q->priority];
+	m->cp_hqd_queue_priority = q->priority;
+}
+
+static struct kfd_mem_obj *allocate_mqd(struct kfd_node *node,
+		struct queue_properties *q)
+{
+	struct kfd_mem_obj *mqd_mem_obj;
+	int size;
+
+	/*
+	 * MES write to areas beyond MQD size. So allocate
+	 * 1 PAGE_SIZE memory for MQD is MES is enabled.
+	 */
+	if (node->kfd->shared_resources.enable_mes)
+		size = PAGE_SIZE;
+	else
+		size = sizeof(struct v11_compute_mqd);
+
+	if (kfd_gtt_sa_allocate(node, size, &mqd_mem_obj))
+		return NULL;
+
+	return mqd_mem_obj;
+}
+
+static void init_mqd(struct mqd_manager *mm, void **mqd,
+			struct kfd_mem_obj *mqd_mem_obj, uint64_t *gart_addr,
+			struct queue_properties *q)
+{
+	uint64_t addr;
+	struct v11_compute_mqd *m;
+	int size;
+	uint32_t wa_mask = q->is_dbg_wa ? 0xffff : 0xffffffff;
+
+	m = (struct v11_compute_mqd *) mqd_mem_obj->cpu_ptr;
+	addr = mqd_mem_obj->gpu_addr;
+
+	if (mm->dev->kfd->shared_resources.enable_mes)
+		size = PAGE_SIZE;
+	else
+		size = sizeof(struct v11_compute_mqd);
+
+	memset(m, 0, size);
+
+	m->header = 0xC0310800;
+	m->compute_pipelinestat_enable = 1;
+
+	m->compute_static_thread_mgmt_se0 = wa_mask;
+	m->compute_static_thread_mgmt_se1 = wa_mask;
+	m->compute_static_thread_mgmt_se2 = wa_mask;
+	m->compute_static_thread_mgmt_se3 = wa_mask;
+	m->compute_static_thread_mgmt_se4 = wa_mask;
+	m->compute_static_thread_mgmt_se5 = wa_mask;
+	m->compute_static_thread_mgmt_se6 = wa_mask;
+	m->compute_static_thread_mgmt_se7 = wa_mask;
+
+	m->cp_hqd_persistent_state = CP_HQD_PERSISTENT_STATE__PRELOAD_REQ_MASK |
+			0x55 << CP_HQD_PERSISTENT_STATE__PRELOAD_SIZE__SHIFT;
+
+	m->cp_hqd_pq_control = 5 << CP_HQD_PQ_CONTROL__RPTR_BLOCK_SIZE__SHIFT;
+	m->cp_hqd_pq_control |= CP_HQD_PQ_CONTROL__UNORD_DISPATCH_MASK;
+	m->cp_mqd_control = 1 << CP_MQD_CONTROL__PRIV_STATE__SHIFT;
+
+	m->cp_mqd_base_addr_lo        = lower_32_bits(addr);
+	m->cp_mqd_base_addr_hi        = upper_32_bits(addr);
+
+	m->cp_hqd_quantum = 1 << CP_HQD_QUANTUM__QUANTUM_EN__SHIFT |
+			1 << CP_HQD_QUANTUM__QUANTUM_SCALE__SHIFT |
+			1 << CP_HQD_QUANTUM__QUANTUM_DURATION__SHIFT;
+
+	/* Set cp_hqd_hq_scheduler0 bit 14 to 1 to have the CP set up the
+	 * DISPATCH_PTR.  This is required for the kfd debugger
+	 */
+	m->cp_hqd_hq_status0 = 1 << 14;
+
+	/*
+	 * GFX11 RS64 CPFW version >= 509 supports PCIe atomics support
+	 * acknowledgment.
+	 */
+	if (amdgpu_amdkfd_have_atomics_support(mm->dev->adev))
+		m->cp_hqd_hq_status0 |= 1 << 29;
+
+	if (q->format == KFD_QUEUE_FORMAT_AQL) {
+		m->cp_hqd_aql_control =
+			1 << CP_HQD_AQL_CONTROL__CONTROL0__SHIFT;
+	}
+
+	if (mm->dev->kfd->cwsr_enabled) {
+		m->cp_hqd_persistent_state |=
+			(1 << CP_HQD_PERSISTENT_STATE__QSWITCH_MODE__SHIFT);
+		m->cp_hqd_ctx_save_base_addr_lo =
+			lower_32_bits(q->ctx_save_restore_area_address);
+		m->cp_hqd_ctx_save_base_addr_hi =
+			upper_32_bits(q->ctx_save_restore_area_address);
+		m->cp_hqd_ctx_save_size = q->ctx_save_restore_area_size;
+		m->cp_hqd_cntl_stack_size = q->ctl_stack_size;
+		m->cp_hqd_cntl_stack_offset = q->ctl_stack_size;
+		m->cp_hqd_wg_state_offset = q->ctl_stack_size;
+	}
+
+	*mqd = m;
+	if (gart_addr)
+		*gart_addr = addr;
+	mm->update_mqd(mm, m, q, NULL);
+}
+
+static int load_mqd(struct mqd_manager *mm, void *mqd,
+			uint32_t pipe_id, uint32_t queue_id,
+			struct queue_properties *p, struct mm_struct *mms)
+{
+	int r = 0;
+	/* AQL write pointer counts in 64B packets, PM4/CP counts in dwords. */
+	uint32_t wptr_shift = (p->format == KFD_QUEUE_FORMAT_AQL ? 4 : 0);
+
+	r = mm->dev->kfd2kgd->hqd_load(mm->dev->adev, mqd, pipe_id, queue_id,
+					  (uint32_t __user *)p->write_ptr,
+					  wptr_shift, 0, mms, 0);
+	return r;
+}
+
+static void update_mqd(struct mqd_manager *mm, void *mqd,
+		       struct queue_properties *q,
+		       struct mqd_update_info *minfo)
+{
+	struct v11_compute_mqd *m;
+
+	m = get_mqd(mqd);
+
+	m->cp_hqd_pq_control &= ~CP_HQD_PQ_CONTROL__QUEUE_SIZE_MASK;
+	m->cp_hqd_pq_control |=
+			ffs(q->queue_size / sizeof(unsigned int)) - 1 - 1;
+	pr_debug("cp_hqd_pq_control 0x%x\n", m->cp_hqd_pq_control);
+
+	m->cp_hqd_pq_base_lo = lower_32_bits((uint64_t)q->queue_address >> 8);
+	m->cp_hqd_pq_base_hi = upper_32_bits((uint64_t)q->queue_address >> 8);
+
+	m->cp_hqd_pq_rptr_report_addr_lo = lower_32_bits((uint64_t)q->read_ptr);
+	m->cp_hqd_pq_rptr_report_addr_hi = upper_32_bits((uint64_t)q->read_ptr);
+	m->cp_hqd_pq_wptr_poll_addr_lo = lower_32_bits((uint64_t)q->write_ptr);
+	m->cp_hqd_pq_wptr_poll_addr_hi = upper_32_bits((uint64_t)q->write_ptr);
+
+	m->cp_hqd_pq_doorbell_control =
+		q->doorbell_off <<
+			CP_HQD_PQ_DOORBELL_CONTROL__DOORBELL_OFFSET__SHIFT;
+	pr_debug("cp_hqd_pq_doorbell_control 0x%x\n",
+			m->cp_hqd_pq_doorbell_control);
+
+	m->cp_hqd_ib_control = 3 << CP_HQD_IB_CONTROL__MIN_IB_AVAIL_SIZE__SHIFT;
+
+	/*
+	 * HW does not clamp this field correctly. Maximum EOP queue size
+	 * is constrained by per-SE EOP done signal count, which is 8-bit.
+	 * Limit is 0xFF EOP entries (= 0x7F8 dwords). CP will not submit
+	 * more than (EOP entry count - 1) so a queue size of 0x800 dwords
+	 * is safe, giving a maximum field value of 0xA.
+	 */
+	m->cp_hqd_eop_control = min(0xA,
+		ffs(q->eop_ring_buffer_size / sizeof(unsigned int)) - 1 - 1);
+	m->cp_hqd_eop_base_addr_lo =
+			lower_32_bits(q->eop_ring_buffer_address >> 8);
+	m->cp_hqd_eop_base_addr_hi =
+			upper_32_bits(q->eop_ring_buffer_address >> 8);
+
+	m->cp_hqd_iq_timer = 0;
+
+	m->cp_hqd_vmid = q->vmid;
+
+	if (q->format == KFD_QUEUE_FORMAT_AQL) {
+		/* GC 10 removed WPP_CLAMP from PQ Control */
+		m->cp_hqd_pq_control |= CP_HQD_PQ_CONTROL__NO_UPDATE_RPTR_MASK |
+				2 << CP_HQD_PQ_CONTROL__SLOT_BASED_WPTR__SHIFT |
+				1 << CP_HQD_PQ_CONTROL__QUEUE_FULL_EN__SHIFT ;
+		m->cp_hqd_pq_doorbell_control |=
+			1 << CP_HQD_PQ_DOORBELL_CONTROL__DOORBELL_BIF_DROP__SHIFT;
+	}
+	if (mm->dev->kfd->cwsr_enabled)
+		m->cp_hqd_ctx_save_control = 0;
+
+	update_cu_mask(mm, mqd, minfo);
+	set_priority(m, q);
+
+	q->is_active = QUEUE_IS_ACTIVE(*q);
+}
+
+static bool check_preemption_failed(struct mqd_manager *mm, void *mqd)
+{
+	struct v11_compute_mqd *m = (struct v11_compute_mqd *)mqd;
+
+	return kfd_check_hiq_mqd_doorbell_id(mm->dev, m->queue_doorbell_id0, 0);
+}
+
+static int get_wave_state(struct mqd_manager *mm, void *mqd,
+			  struct queue_properties *q,
+			  void __user *ctl_stack,
+			  u32 *ctl_stack_used_size,
+			  u32 *save_area_used_size)
+{
+	struct v11_compute_mqd *m;
+	struct kfd_context_save_area_header header;
+
+	m = get_mqd(mqd);
+
+	/* Control stack is written backwards, while workgroup context data
+	 * is written forwards. Both starts from m->cp_hqd_cntl_stack_size.
+	 * Current position is at m->cp_hqd_cntl_stack_offset and
+	 * m->cp_hqd_wg_state_offset, respectively.
+	 */
+	*ctl_stack_used_size = m->cp_hqd_cntl_stack_size -
+		m->cp_hqd_cntl_stack_offset;
+	*save_area_used_size = m->cp_hqd_wg_state_offset -
+		m->cp_hqd_cntl_stack_size;
+
+	/* Control stack is not copied to user mode for GFXv11 because
+	 * it's part of the context save area that is already
+	 * accessible to user mode
+	 */
+	header.wave_state.control_stack_size = *ctl_stack_used_size;
+	header.wave_state.wave_state_size = *save_area_used_size;
+
+	header.wave_state.wave_state_offset = m->cp_hqd_wg_state_offset;
+	header.wave_state.control_stack_offset = m->cp_hqd_cntl_stack_offset;
+
+	if (copy_to_user(ctl_stack, &header, sizeof(header.wave_state)))
+		return -EFAULT;
+
+	return 0;
+}
+
+static void checkpoint_mqd(struct mqd_manager *mm, void *mqd, void *mqd_dst, void *ctl_stack_dst)
+{
+	struct v11_compute_mqd *m;
+
+	m = get_mqd(mqd);
+
+	memcpy(mqd_dst, m, sizeof(struct v11_compute_mqd));
+}
+
+static void restore_mqd(struct mqd_manager *mm, void **mqd,
+			struct kfd_mem_obj *mqd_mem_obj, uint64_t *gart_addr,
+			struct queue_properties *qp,
+			const void *mqd_src,
+			const void *ctl_stack_src, const u32 ctl_stack_size)
+{
+	uint64_t addr;
+	struct v11_compute_mqd *m;
+
+	m = (struct v11_compute_mqd *) mqd_mem_obj->cpu_ptr;
+	addr = mqd_mem_obj->gpu_addr;
+
+	memcpy(m, mqd_src, sizeof(*m));
+
+	*mqd = m;
+	if (gart_addr)
+		*gart_addr = addr;
+
+	m->cp_hqd_pq_doorbell_control =
+		qp->doorbell_off <<
+			CP_HQD_PQ_DOORBELL_CONTROL__DOORBELL_OFFSET__SHIFT;
+	pr_debug("cp_hqd_pq_doorbell_control 0x%x\n",
+			m->cp_hqd_pq_doorbell_control);
+
+	qp->is_active = 0;
+}
+
+
+static void init_mqd_hiq(struct mqd_manager *mm, void **mqd,
+			struct kfd_mem_obj *mqd_mem_obj, uint64_t *gart_addr,
+			struct queue_properties *q)
+{
+	struct v11_compute_mqd *m;
+
+	init_mqd(mm, mqd, mqd_mem_obj, gart_addr, q);
+
+	m = get_mqd(*mqd);
+
+	m->cp_hqd_pq_control |= 1 << CP_HQD_PQ_CONTROL__PRIV_STATE__SHIFT |
+			1 << CP_HQD_PQ_CONTROL__KMD_QUEUE__SHIFT;
+}
+
+static int destroy_hiq_mqd(struct mqd_manager *mm, void *mqd,
+			enum kfd_preempt_type type, unsigned int timeout,
+			uint32_t pipe_id, uint32_t queue_id)
+{
+	int err;
+	struct v11_compute_mqd *m;
+	u32 doorbell_off;
+
+	m = get_mqd(mqd);
+
+	doorbell_off = m->cp_hqd_pq_doorbell_control >>
+			CP_HQD_PQ_DOORBELL_CONTROL__DOORBELL_OFFSET__SHIFT;
+
+	err = amdgpu_amdkfd_unmap_hiq(mm->dev->adev, doorbell_off, 0);
+	if (err)
+		pr_debug("Destroy HIQ MQD failed: %d\n", err);
+
+	return err;
+}
+
+static void init_mqd_sdma(struct mqd_manager *mm, void **mqd,
+		struct kfd_mem_obj *mqd_mem_obj, uint64_t *gart_addr,
+		struct queue_properties *q)
+{
+	struct v11_sdma_mqd *m;
+	int size;
+
+	m = (struct v11_sdma_mqd *) mqd_mem_obj->cpu_ptr;
+
+	if (mm->dev->kfd->shared_resources.enable_mes)
+		size = PAGE_SIZE;
+	else
+		size = sizeof(struct v11_sdma_mqd);
+
+	memset(m, 0, size);
+	*mqd = m;
+	if (gart_addr)
+		*gart_addr = mqd_mem_obj->gpu_addr;
+
+	mm->update_mqd(mm, m, q, NULL);
+}
+
+#define SDMA_RLC_DUMMY_DEFAULT 0xf
+
+static void update_mqd_sdma(struct mqd_manager *mm, void *mqd,
+		struct queue_properties *q,
+		struct mqd_update_info *minfo)
+{
+	struct v11_sdma_mqd *m;
+
+	m = get_sdma_mqd(mqd);
+	m->sdmax_rlcx_rb_cntl = (ffs(q->queue_size / sizeof(unsigned int)) - 1)
+		<< SDMA0_QUEUE0_RB_CNTL__RB_SIZE__SHIFT |
+		q->vmid << SDMA0_QUEUE0_RB_CNTL__RB_VMID__SHIFT |
+		1 << SDMA0_QUEUE0_RB_CNTL__RPTR_WRITEBACK_ENABLE__SHIFT |
+		6 << SDMA0_QUEUE0_RB_CNTL__RPTR_WRITEBACK_TIMER__SHIFT |
+		1 << SDMA0_QUEUE0_RB_CNTL__F32_WPTR_POLL_ENABLE__SHIFT;
+
+	m->sdmax_rlcx_rb_base = lower_32_bits(q->queue_address >> 8);
+	m->sdmax_rlcx_rb_base_hi = upper_32_bits(q->queue_address >> 8);
+	m->sdmax_rlcx_rb_rptr_addr_lo = lower_32_bits((uint64_t)q->read_ptr);
+	m->sdmax_rlcx_rb_rptr_addr_hi = upper_32_bits((uint64_t)q->read_ptr);
+	m->sdmax_rlcx_rb_wptr_poll_addr_lo = lower_32_bits((uint64_t)q->write_ptr);
+	m->sdmax_rlcx_rb_wptr_poll_addr_hi = upper_32_bits((uint64_t)q->write_ptr);
+	m->sdmax_rlcx_doorbell_offset =
+		q->doorbell_off << SDMA0_QUEUE0_DOORBELL_OFFSET__OFFSET__SHIFT;
+
+	m->sdmax_rlcx_sched_cntl = (amdgpu_sdma_phase_quantum
+		<< SDMA0_QUEUE0_SCHEDULE_CNTL__CONTEXT_QUANTUM__SHIFT)
+		 & SDMA0_QUEUE0_SCHEDULE_CNTL__CONTEXT_QUANTUM_MASK;
+
+	m->sdma_engine_id = q->sdma_engine_id;
+	m->sdma_queue_id = q->sdma_queue_id;
+	m->sdmax_rlcx_dummy_reg = SDMA_RLC_DUMMY_DEFAULT;
+
+	q->is_active = QUEUE_IS_ACTIVE(*q);
+}
+
+#if defined(CONFIG_DEBUG_FS)
+
+static int debugfs_show_mqd(struct seq_file *m, void *data)
+{
+	seq_hex_dump(m, "    ", DUMP_PREFIX_OFFSET, 32, 4,
+		     data, sizeof(struct v11_compute_mqd), false);
+	return 0;
+}
+
+static int debugfs_show_mqd_sdma(struct seq_file *m, void *data)
+{
+	seq_hex_dump(m, "    ", DUMP_PREFIX_OFFSET, 32, 4,
+		     data, sizeof(struct v11_sdma_mqd), false);
+	return 0;
+}
+
+#endif
+
+struct mqd_manager *mqd_manager_init_v11(enum KFD_MQD_TYPE type,
+		struct kfd_node *dev)
+{
+	struct mqd_manager *mqd;
+
+	if (WARN_ON(type >= KFD_MQD_TYPE_MAX))
+		return NULL;
+
+	mqd = kzalloc(sizeof(*mqd), GFP_KERNEL);
+	if (!mqd)
+		return NULL;
+
+	mqd->dev = dev;
+
+	switch (type) {
+	case KFD_MQD_TYPE_CP:
+		pr_debug("%s@%i\n", __func__, __LINE__);
+		mqd->allocate_mqd = allocate_mqd;
+		mqd->init_mqd = init_mqd;
+		mqd->free_mqd = kfd_free_mqd_cp;
+		mqd->load_mqd = load_mqd;
+		mqd->update_mqd = update_mqd;
+		mqd->destroy_mqd = kfd_destroy_mqd_cp;
+		mqd->is_occupied = kfd_is_occupied_cp;
+		mqd->mqd_size = sizeof(struct v11_compute_mqd);
+		mqd->get_wave_state = get_wave_state;
+		mqd->mqd_stride = kfd_mqd_stride;
+		mqd->checkpoint_mqd = checkpoint_mqd;
+		mqd->restore_mqd = restore_mqd;
+#if defined(CONFIG_DEBUG_FS)
+		mqd->debugfs_show_mqd = debugfs_show_mqd;
+#endif
+		pr_debug("%s@%i\n", __func__, __LINE__);
+		break;
+	case KFD_MQD_TYPE_HIQ:
+		pr_debug("%s@%i\n", __func__, __LINE__);
+		mqd->allocate_mqd = allocate_hiq_mqd;
+		mqd->init_mqd = init_mqd_hiq;
+		mqd->free_mqd = free_mqd_hiq_sdma;
+		mqd->load_mqd = kfd_hiq_load_mqd_kiq;
+		mqd->update_mqd = update_mqd;
+		mqd->destroy_mqd = destroy_hiq_mqd;
+		mqd->is_occupied = kfd_is_occupied_cp;
+		mqd->mqd_size = sizeof(struct v11_compute_mqd);
+		mqd->mqd_stride = kfd_mqd_stride;
+#if defined(CONFIG_DEBUG_FS)
+		mqd->debugfs_show_mqd = debugfs_show_mqd;
+#endif
+		mqd->check_preemption_failed = check_preemption_failed;
+		pr_debug("%s@%i\n", __func__, __LINE__);
+		break;
+	case KFD_MQD_TYPE_DIQ:
+		mqd->allocate_mqd = allocate_mqd;
+		mqd->init_mqd = init_mqd_hiq;
+		mqd->free_mqd = kfd_free_mqd_cp;
+		mqd->load_mqd = load_mqd;
+		mqd->update_mqd = update_mqd;
+		mqd->destroy_mqd = kfd_destroy_mqd_cp;
+		mqd->is_occupied = kfd_is_occupied_cp;
+		mqd->mqd_size = sizeof(struct v11_compute_mqd);
+#if defined(CONFIG_DEBUG_FS)
+		mqd->debugfs_show_mqd = debugfs_show_mqd;
+#endif
+		break;
+	case KFD_MQD_TYPE_SDMA:
+		pr_debug("%s@%i\n", __func__, __LINE__);
+		mqd->allocate_mqd = allocate_sdma_mqd;
+		mqd->init_mqd = init_mqd_sdma;
+		mqd->free_mqd = free_mqd_hiq_sdma;
+		mqd->load_mqd = kfd_load_mqd_sdma;
+		mqd->update_mqd = update_mqd_sdma;
+		mqd->destroy_mqd = kfd_destroy_mqd_sdma;
+		mqd->is_occupied = kfd_is_occupied_sdma;
+		mqd->checkpoint_mqd = checkpoint_mqd;
+		mqd->restore_mqd = restore_mqd;
+		mqd->mqd_size = sizeof(struct v11_sdma_mqd);
+		mqd->mqd_stride = kfd_mqd_stride;
+#if defined(CONFIG_DEBUG_FS)
+		mqd->debugfs_show_mqd = debugfs_show_mqd_sdma;
+#endif
+		/*
+		 * To allocate SDMA MQDs by generic functions
+		 * when MES is enabled.
+		 */
+		if (dev->kfd->shared_resources.enable_mes) {
+			mqd->allocate_mqd = allocate_mqd;
+			mqd->free_mqd = kfd_free_mqd_cp;
+		}
+		pr_debug("%s@%i\n", __func__, __LINE__);
+		break;
+	default:
+		kfree(mqd);
+		return NULL;
+	}
+
+	return mqd;
+}
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v12.c b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v12.c
new file mode 100644
index 000000000000..565858b9044d
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v12.c
@@ -0,0 +1,459 @@
+// SPDX-License-Identifier: GPL-2.0 OR MIT
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/printk.h>
+#include <linux/slab.h>
+#include <linux/uaccess.h>
+#include "kfd_priv.h"
+#include "kfd_mqd_manager.h"
+#include "v12_structs.h"
+#include "gc/gc_12_0_0_sh_mask.h"
+#include "amdgpu_amdkfd.h"
+
+static inline struct v12_compute_mqd *get_mqd(void *mqd)
+{
+	return (struct v12_compute_mqd *)mqd;
+}
+
+static inline struct v12_sdma_mqd *get_sdma_mqd(void *mqd)
+{
+	return (struct v12_sdma_mqd *)mqd;
+}
+
+static void update_cu_mask(struct mqd_manager *mm, void *mqd,
+			   struct mqd_update_info *minfo)
+{
+	struct v12_compute_mqd *m;
+	uint32_t se_mask[KFD_MAX_NUM_SE] = {0};
+
+	if (!minfo || !minfo->cu_mask.ptr)
+		return;
+
+	mqd_symmetrically_map_cu_mask(mm,
+		minfo->cu_mask.ptr, minfo->cu_mask.count, se_mask, 0);
+
+	m = get_mqd(mqd);
+	m->compute_static_thread_mgmt_se0 = se_mask[0];
+	m->compute_static_thread_mgmt_se1 = se_mask[1];
+	m->compute_static_thread_mgmt_se2 = se_mask[2];
+	m->compute_static_thread_mgmt_se3 = se_mask[3];
+	m->compute_static_thread_mgmt_se4 = se_mask[4];
+	m->compute_static_thread_mgmt_se5 = se_mask[5];
+	m->compute_static_thread_mgmt_se6 = se_mask[6];
+	m->compute_static_thread_mgmt_se7 = se_mask[7];
+
+	pr_debug("update cu mask to %#x %#x %#x %#x %#x %#x %#x %#x\n",
+		m->compute_static_thread_mgmt_se0,
+		m->compute_static_thread_mgmt_se1,
+		m->compute_static_thread_mgmt_se2,
+		m->compute_static_thread_mgmt_se3,
+		m->compute_static_thread_mgmt_se4,
+		m->compute_static_thread_mgmt_se5,
+		m->compute_static_thread_mgmt_se6,
+		m->compute_static_thread_mgmt_se7);
+}
+
+static void set_priority(struct v12_compute_mqd *m, struct queue_properties *q)
+{
+	m->cp_hqd_pipe_priority = pipe_priority_map[q->priority];
+	m->cp_hqd_queue_priority = q->priority;
+}
+
+static struct kfd_mem_obj *allocate_mqd(struct kfd_node *node,
+		struct queue_properties *q)
+{
+	struct kfd_mem_obj *mqd_mem_obj;
+
+	/*
+	 * Allocate one PAGE_SIZE memory for MQD as MES writes to areas beyond
+	 * struct MQD size.
+	 */
+	if (kfd_gtt_sa_allocate(node, PAGE_SIZE, &mqd_mem_obj))
+		return NULL;
+
+	return mqd_mem_obj;
+}
+
+static void init_mqd(struct mqd_manager *mm, void **mqd,
+			struct kfd_mem_obj *mqd_mem_obj, uint64_t *gart_addr,
+			struct queue_properties *q)
+{
+	uint64_t addr;
+	struct v12_compute_mqd *m;
+
+	m = (struct v12_compute_mqd *) mqd_mem_obj->cpu_ptr;
+	addr = mqd_mem_obj->gpu_addr;
+
+	memset(m, 0, PAGE_SIZE);
+
+	m->header = 0xC0310800;
+	m->compute_pipelinestat_enable = 1;
+	m->compute_static_thread_mgmt_se0 = 0xFFFFFFFF;
+	m->compute_static_thread_mgmt_se1 = 0xFFFFFFFF;
+	m->compute_static_thread_mgmt_se2 = 0xFFFFFFFF;
+	m->compute_static_thread_mgmt_se3 = 0xFFFFFFFF;
+	m->compute_static_thread_mgmt_se4 = 0xFFFFFFFF;
+	m->compute_static_thread_mgmt_se5 = 0xFFFFFFFF;
+	m->compute_static_thread_mgmt_se6 = 0xFFFFFFFF;
+	m->compute_static_thread_mgmt_se7 = 0xFFFFFFFF;
+
+	m->cp_hqd_persistent_state = CP_HQD_PERSISTENT_STATE__PRELOAD_REQ_MASK |
+			0x55 << CP_HQD_PERSISTENT_STATE__PRELOAD_SIZE__SHIFT;
+
+	m->cp_hqd_pq_control = 5 << CP_HQD_PQ_CONTROL__RPTR_BLOCK_SIZE__SHIFT;
+	m->cp_hqd_pq_control |= CP_HQD_PQ_CONTROL__UNORD_DISPATCH_MASK;
+	m->cp_mqd_control = 1 << CP_MQD_CONTROL__PRIV_STATE__SHIFT;
+
+	m->cp_mqd_base_addr_lo        = lower_32_bits(addr);
+	m->cp_mqd_base_addr_hi        = upper_32_bits(addr);
+
+	m->cp_hqd_quantum = 1 << CP_HQD_QUANTUM__QUANTUM_EN__SHIFT |
+			1 << CP_HQD_QUANTUM__QUANTUM_SCALE__SHIFT |
+			1 << CP_HQD_QUANTUM__QUANTUM_DURATION__SHIFT;
+
+	/* Set cp_hqd_hq_status0.c_queue_debug_en to 1 to have the CP set up the
+	 * DISPATCH_PTR.  This is required for the kfd debugger
+	 */
+	m->cp_hqd_hq_status0 = 1 << 14;
+
+	if (amdgpu_amdkfd_have_atomics_support(mm->dev->adev))
+		m->cp_hqd_hq_status0 |= 1 << 29;
+
+	if (q->format == KFD_QUEUE_FORMAT_AQL) {
+		m->cp_hqd_aql_control =
+			1 << CP_HQD_AQL_CONTROL__CONTROL0__SHIFT;
+	}
+
+	if (mm->dev->kfd->cwsr_enabled) {
+		m->cp_hqd_persistent_state |=
+			(1 << CP_HQD_PERSISTENT_STATE__QSWITCH_MODE__SHIFT);
+		m->cp_hqd_ctx_save_base_addr_lo =
+			lower_32_bits(q->ctx_save_restore_area_address);
+		m->cp_hqd_ctx_save_base_addr_hi =
+			upper_32_bits(q->ctx_save_restore_area_address);
+		m->cp_hqd_ctx_save_size = q->ctx_save_restore_area_size;
+		m->cp_hqd_cntl_stack_size = q->ctl_stack_size;
+		m->cp_hqd_cntl_stack_offset = q->ctl_stack_size;
+		m->cp_hqd_wg_state_offset = q->ctl_stack_size;
+	}
+
+	*mqd = m;
+	if (gart_addr)
+		*gart_addr = addr;
+	mm->update_mqd(mm, m, q, NULL);
+}
+
+static int load_mqd(struct mqd_manager *mm, void *mqd,
+			uint32_t pipe_id, uint32_t queue_id,
+			struct queue_properties *p, struct mm_struct *mms)
+{
+	int r = 0;
+	/* AQL write pointer counts in 64B packets, PM4/CP counts in dwords. */
+	uint32_t wptr_shift = (p->format == KFD_QUEUE_FORMAT_AQL ? 4 : 0);
+
+	r = mm->dev->kfd2kgd->hqd_load(mm->dev->adev, mqd, pipe_id, queue_id,
+					  (uint32_t __user *)p->write_ptr,
+					  wptr_shift, 0, mms, 0);
+	return r;
+}
+
+static void update_mqd(struct mqd_manager *mm, void *mqd,
+		       struct queue_properties *q,
+		       struct mqd_update_info *minfo)
+{
+	struct v12_compute_mqd *m;
+
+	m = get_mqd(mqd);
+
+	m->cp_hqd_pq_control &= ~CP_HQD_PQ_CONTROL__QUEUE_SIZE_MASK;
+	m->cp_hqd_pq_control |=
+			ffs(q->queue_size / sizeof(unsigned int)) - 1 - 1;
+	pr_debug("cp_hqd_pq_control 0x%x\n", m->cp_hqd_pq_control);
+
+	m->cp_hqd_pq_base_lo = lower_32_bits((uint64_t)q->queue_address >> 8);
+	m->cp_hqd_pq_base_hi = upper_32_bits((uint64_t)q->queue_address >> 8);
+
+	m->cp_hqd_pq_rptr_report_addr_lo = lower_32_bits((uint64_t)q->read_ptr);
+	m->cp_hqd_pq_rptr_report_addr_hi = upper_32_bits((uint64_t)q->read_ptr);
+	m->cp_hqd_pq_wptr_poll_addr_lo = lower_32_bits((uint64_t)q->write_ptr);
+	m->cp_hqd_pq_wptr_poll_addr_hi = upper_32_bits((uint64_t)q->write_ptr);
+
+	m->cp_hqd_pq_doorbell_control =
+		q->doorbell_off <<
+			CP_HQD_PQ_DOORBELL_CONTROL__DOORBELL_OFFSET__SHIFT;
+	pr_debug("cp_hqd_pq_doorbell_control 0x%x\n",
+			m->cp_hqd_pq_doorbell_control);
+
+	m->cp_hqd_ib_control = 3 << CP_HQD_IB_CONTROL__MIN_IB_AVAIL_SIZE__SHIFT;
+
+	/*
+	 * HW does not clamp this field correctly. Maximum EOP queue size
+	 * is constrained by per-SE EOP done signal count, which is 8-bit.
+	 * Limit is 0xFF EOP entries (= 0x7F8 dwords). CP will not submit
+	 * more than (EOP entry count - 1) so a queue size of 0x800 dwords
+	 * is safe, giving a maximum field value of 0xA.
+	 */
+	m->cp_hqd_eop_control = min(0xA,
+		ffs(q->eop_ring_buffer_size / sizeof(unsigned int)) - 1 - 1);
+	m->cp_hqd_eop_base_addr_lo =
+			lower_32_bits(q->eop_ring_buffer_address >> 8);
+	m->cp_hqd_eop_base_addr_hi =
+			upper_32_bits(q->eop_ring_buffer_address >> 8);
+
+	m->cp_hqd_iq_timer = 0;
+
+	m->cp_hqd_vmid = q->vmid;
+
+	if (q->format == KFD_QUEUE_FORMAT_AQL) {
+		/* GC 10 removed WPP_CLAMP from PQ Control */
+		m->cp_hqd_pq_control |= CP_HQD_PQ_CONTROL__NO_UPDATE_RPTR_MASK |
+				2 << CP_HQD_PQ_CONTROL__SLOT_BASED_WPTR__SHIFT |
+				1 << CP_HQD_PQ_CONTROL__QUEUE_FULL_EN__SHIFT;
+		m->cp_hqd_pq_doorbell_control |=
+			1 << CP_HQD_PQ_DOORBELL_CONTROL__DOORBELL_BIF_DROP__SHIFT;
+	}
+	if (mm->dev->kfd->cwsr_enabled)
+		m->cp_hqd_ctx_save_control = 0;
+
+	update_cu_mask(mm, mqd, minfo);
+	set_priority(m, q);
+
+	q->is_active = QUEUE_IS_ACTIVE(*q);
+}
+
+static bool check_preemption_failed(struct mqd_manager *mm, void *mqd)
+{
+	struct v12_compute_mqd *m = (struct v12_compute_mqd *)mqd;
+
+	return kfd_check_hiq_mqd_doorbell_id(mm->dev, m->queue_doorbell_id0, 0);
+}
+
+static int get_wave_state(struct mqd_manager *mm, void *mqd,
+			  struct queue_properties *q,
+			  void __user *ctl_stack,
+			  u32 *ctl_stack_used_size,
+			  u32 *save_area_used_size)
+{
+	struct v12_compute_mqd *m;
+	struct mqd_user_context_save_area_header header;
+
+	m = get_mqd(mqd);
+
+	/* Control stack is written backwards, while workgroup context data
+	 * is written forwards. Both starts from m->cp_hqd_cntl_stack_size.
+	 * Current position is at m->cp_hqd_cntl_stack_offset and
+	 * m->cp_hqd_wg_state_offset, respectively.
+	 */
+	*ctl_stack_used_size = m->cp_hqd_cntl_stack_size -
+		m->cp_hqd_cntl_stack_offset;
+	*save_area_used_size = m->cp_hqd_wg_state_offset -
+		m->cp_hqd_cntl_stack_size;
+
+	/* Control stack is not copied to user mode for GFXv12 because
+	 * it's part of the context save area that is already
+	 * accessible to user mode
+	 */
+	header.control_stack_size = *ctl_stack_used_size;
+	header.wave_state_size = *save_area_used_size;
+
+	header.wave_state_offset = m->cp_hqd_wg_state_offset;
+	header.control_stack_offset = m->cp_hqd_cntl_stack_offset;
+
+	if (copy_to_user(ctl_stack, &header, sizeof(header)))
+		return -EFAULT;
+
+	return 0;
+}
+
+static void init_mqd_hiq(struct mqd_manager *mm, void **mqd,
+			struct kfd_mem_obj *mqd_mem_obj, uint64_t *gart_addr,
+			struct queue_properties *q)
+{
+	struct v12_compute_mqd *m;
+
+	init_mqd(mm, mqd, mqd_mem_obj, gart_addr, q);
+
+	m = get_mqd(*mqd);
+
+	m->cp_hqd_pq_control |= 1 << CP_HQD_PQ_CONTROL__PRIV_STATE__SHIFT |
+			1 << CP_HQD_PQ_CONTROL__KMD_QUEUE__SHIFT;
+}
+
+static void init_mqd_sdma(struct mqd_manager *mm, void **mqd,
+		struct kfd_mem_obj *mqd_mem_obj, uint64_t *gart_addr,
+		struct queue_properties *q)
+{
+	struct v12_sdma_mqd *m;
+
+	m = (struct v12_sdma_mqd *) mqd_mem_obj->cpu_ptr;
+
+	memset(m, 0, sizeof(struct v12_sdma_mqd));
+
+	*mqd = m;
+	if (gart_addr)
+		*gart_addr = mqd_mem_obj->gpu_addr;
+
+	mm->update_mqd(mm, m, q, NULL);
+}
+
+#define SDMA_RLC_DUMMY_DEFAULT 0xf
+
+static void update_mqd_sdma(struct mqd_manager *mm, void *mqd,
+		struct queue_properties *q,
+		struct mqd_update_info *minfo)
+{
+	struct v12_sdma_mqd *m;
+
+	m = get_sdma_mqd(mqd);
+	m->sdmax_rlcx_rb_cntl = (ffs(q->queue_size / sizeof(unsigned int)) - 1)
+		<< SDMA0_QUEUE0_RB_CNTL__RB_SIZE__SHIFT |
+		q->vmid << SDMA0_QUEUE0_RB_CNTL__RB_VMID__SHIFT |
+		1 << SDMA0_QUEUE0_RB_CNTL__RPTR_WRITEBACK_ENABLE__SHIFT |
+		6 << SDMA0_QUEUE0_RB_CNTL__RPTR_WRITEBACK_TIMER__SHIFT |
+		1 << SDMA0_QUEUE0_RB_CNTL__MCU_WPTR_POLL_ENABLE__SHIFT;
+
+	m->sdmax_rlcx_rb_base = lower_32_bits(q->queue_address >> 8);
+	m->sdmax_rlcx_rb_base_hi = upper_32_bits(q->queue_address >> 8);
+	m->sdmax_rlcx_rb_rptr_addr_lo = lower_32_bits((uint64_t)q->read_ptr);
+	m->sdmax_rlcx_rb_rptr_addr_hi = upper_32_bits((uint64_t)q->read_ptr);
+	m->sdmax_rlcx_rb_wptr_poll_addr_lo = lower_32_bits((uint64_t)q->write_ptr);
+	m->sdmax_rlcx_rb_wptr_poll_addr_hi = upper_32_bits((uint64_t)q->write_ptr);
+	m->sdmax_rlcx_doorbell_offset =
+		q->doorbell_off << SDMA0_QUEUE0_DOORBELL_OFFSET__OFFSET__SHIFT;
+
+	m->sdmax_rlcx_sched_cntl = (amdgpu_sdma_phase_quantum
+		<< SDMA0_QUEUE0_SCHEDULE_CNTL__CONTEXT_QUANTUM__SHIFT)
+		 & SDMA0_QUEUE0_SCHEDULE_CNTL__CONTEXT_QUANTUM_MASK;
+
+	m->sdma_engine_id = q->sdma_engine_id;
+	m->sdma_queue_id = q->sdma_queue_id;
+
+	m->sdmax_rlcx_dummy_reg = SDMA_RLC_DUMMY_DEFAULT;
+
+	q->is_active = QUEUE_IS_ACTIVE(*q);
+}
+
+#if defined(CONFIG_DEBUG_FS)
+
+static int debugfs_show_mqd(struct seq_file *m, void *data)
+{
+	seq_hex_dump(m, "    ", DUMP_PREFIX_OFFSET, 32, 4,
+		     data, sizeof(struct v12_compute_mqd), false);
+	return 0;
+}
+
+static int debugfs_show_mqd_sdma(struct seq_file *m, void *data)
+{
+	seq_hex_dump(m, "    ", DUMP_PREFIX_OFFSET, 32, 4,
+		     data, sizeof(struct v12_sdma_mqd), false);
+	return 0;
+}
+
+#endif
+
+struct mqd_manager *mqd_manager_init_v12(enum KFD_MQD_TYPE type,
+		struct kfd_node *dev)
+{
+	struct mqd_manager *mqd;
+
+	if (WARN_ON(type >= KFD_MQD_TYPE_MAX))
+		return NULL;
+
+	mqd = kzalloc(sizeof(*mqd), GFP_KERNEL);
+	if (!mqd)
+		return NULL;
+
+	mqd->dev = dev;
+
+	switch (type) {
+	case KFD_MQD_TYPE_CP:
+		pr_debug("%s@%i\n", __func__, __LINE__);
+		mqd->allocate_mqd = allocate_mqd;
+		mqd->init_mqd = init_mqd;
+		mqd->free_mqd = kfd_free_mqd_cp;
+		mqd->load_mqd = load_mqd;
+		mqd->update_mqd = update_mqd;
+		mqd->destroy_mqd = kfd_destroy_mqd_cp;
+		mqd->is_occupied = kfd_is_occupied_cp;
+		mqd->mqd_size = sizeof(struct v12_compute_mqd);
+		mqd->get_wave_state = get_wave_state;
+		mqd->mqd_stride = kfd_mqd_stride;
+#if defined(CONFIG_DEBUG_FS)
+		mqd->debugfs_show_mqd = debugfs_show_mqd;
+#endif
+		pr_debug("%s@%i\n", __func__, __LINE__);
+		break;
+	case KFD_MQD_TYPE_HIQ:
+		pr_debug("%s@%i\n", __func__, __LINE__);
+		mqd->allocate_mqd = allocate_hiq_mqd;
+		mqd->init_mqd = init_mqd_hiq;
+		mqd->free_mqd = free_mqd_hiq_sdma;
+		mqd->load_mqd = kfd_hiq_load_mqd_kiq;
+		mqd->update_mqd = update_mqd;
+		mqd->destroy_mqd = kfd_destroy_mqd_cp;
+		mqd->is_occupied = kfd_is_occupied_cp;
+		mqd->mqd_size = sizeof(struct v12_compute_mqd);
+		mqd->mqd_stride = kfd_mqd_stride;
+#if defined(CONFIG_DEBUG_FS)
+		mqd->debugfs_show_mqd = debugfs_show_mqd;
+#endif
+		mqd->check_preemption_failed = check_preemption_failed;
+		pr_debug("%s@%i\n", __func__, __LINE__);
+		break;
+	case KFD_MQD_TYPE_DIQ:
+		mqd->allocate_mqd = allocate_mqd;
+		mqd->init_mqd = init_mqd_hiq;
+		mqd->free_mqd = kfd_free_mqd_cp;
+		mqd->load_mqd = load_mqd;
+		mqd->update_mqd = update_mqd;
+		mqd->destroy_mqd = kfd_destroy_mqd_cp;
+		mqd->is_occupied = kfd_is_occupied_cp;
+		mqd->mqd_size = sizeof(struct v12_compute_mqd);
+#if defined(CONFIG_DEBUG_FS)
+		mqd->debugfs_show_mqd = debugfs_show_mqd;
+#endif
+		break;
+	case KFD_MQD_TYPE_SDMA:
+		pr_debug("%s@%i\n", __func__, __LINE__);
+		mqd->allocate_mqd = allocate_mqd;
+		mqd->init_mqd = init_mqd_sdma;
+		mqd->free_mqd = kfd_free_mqd_cp;
+		mqd->load_mqd = kfd_load_mqd_sdma;
+		mqd->update_mqd = update_mqd_sdma;
+		mqd->destroy_mqd = kfd_destroy_mqd_sdma;
+		mqd->is_occupied = kfd_is_occupied_sdma;
+		mqd->mqd_size = sizeof(struct v12_sdma_mqd);
+		mqd->mqd_stride = kfd_mqd_stride;
+#if defined(CONFIG_DEBUG_FS)
+		mqd->debugfs_show_mqd = debugfs_show_mqd_sdma;
+#endif
+		pr_debug("%s@%i\n", __func__, __LINE__);
+		break;
+	default:
+		kfree(mqd);
+		return NULL;
+	}
+
+	return mqd;
+}
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v9.c b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v9.c
index 4e5932f54b5a..f2dee320fada 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v9.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v9.c
@@ -1,5 +1,6 @@
+// SPDX-License-Identifier: GPL-2.0 OR MIT
 /*
- * Copyright 2016-2018 Advanced Micro Devices, Inc.
+ * Copyright 2016-2022 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -31,6 +32,22 @@
 #include "gc/gc_9_0_sh_mask.h"
 #include "sdma0/sdma0_4_0_sh_mask.h"
 #include "amdgpu_amdkfd.h"
+#include "kfd_device_queue_manager.h"
+
+static void update_mqd(struct mqd_manager *mm, void *mqd,
+		       struct queue_properties *q,
+		       struct mqd_update_info *minfo);
+
+static uint64_t mqd_stride_v9(struct mqd_manager *mm,
+				struct queue_properties *q)
+{
+	if (mm->dev->kfd->cwsr_enabled &&
+	    q->type == KFD_QUEUE_TYPE_COMPUTE)
+		return ALIGN(q->ctl_stack_size, PAGE_SIZE) +
+			ALIGN(sizeof(struct v9_mqd), PAGE_SIZE);
+
+	return mm->mqd_size;
+}
 
 static inline struct v9_mqd *get_mqd(void *mqd)
 {
@@ -43,37 +60,47 @@ static inline struct v9_sdma_mqd *get_sdma_mqd(void *mqd)
 }
 
 static void update_cu_mask(struct mqd_manager *mm, void *mqd,
-			struct mqd_update_info *minfo)
+			struct mqd_update_info *minfo, uint32_t inst)
 {
 	struct v9_mqd *m;
 	uint32_t se_mask[KFD_MAX_NUM_SE] = {0};
 
-	if (!minfo || (minfo->update_flag != UPDATE_FLAG_CU_MASK) ||
-	    !minfo->cu_mask.ptr)
+	if (!minfo || !minfo->cu_mask.ptr)
 		return;
 
 	mqd_symmetrically_map_cu_mask(mm,
-		minfo->cu_mask.ptr, minfo->cu_mask.count, se_mask);
+		minfo->cu_mask.ptr, minfo->cu_mask.count, se_mask, inst);
 
 	m = get_mqd(mqd);
+
 	m->compute_static_thread_mgmt_se0 = se_mask[0];
 	m->compute_static_thread_mgmt_se1 = se_mask[1];
 	m->compute_static_thread_mgmt_se2 = se_mask[2];
 	m->compute_static_thread_mgmt_se3 = se_mask[3];
-	m->compute_static_thread_mgmt_se4 = se_mask[4];
-	m->compute_static_thread_mgmt_se5 = se_mask[5];
-	m->compute_static_thread_mgmt_se6 = se_mask[6];
-	m->compute_static_thread_mgmt_se7 = se_mask[7];
-
-	pr_debug("update cu mask to %#x %#x %#x %#x %#x %#x %#x %#x\n",
-		m->compute_static_thread_mgmt_se0,
-		m->compute_static_thread_mgmt_se1,
-		m->compute_static_thread_mgmt_se2,
-		m->compute_static_thread_mgmt_se3,
-		m->compute_static_thread_mgmt_se4,
-		m->compute_static_thread_mgmt_se5,
-		m->compute_static_thread_mgmt_se6,
-		m->compute_static_thread_mgmt_se7);
+	if (KFD_GC_VERSION(mm->dev) != IP_VERSION(9, 4, 3) &&
+	    KFD_GC_VERSION(mm->dev) != IP_VERSION(9, 4, 4) &&
+	    KFD_GC_VERSION(mm->dev) != IP_VERSION(9, 5, 0)) {
+		m->compute_static_thread_mgmt_se4 = se_mask[4];
+		m->compute_static_thread_mgmt_se5 = se_mask[5];
+		m->compute_static_thread_mgmt_se6 = se_mask[6];
+		m->compute_static_thread_mgmt_se7 = se_mask[7];
+
+		pr_debug("update cu mask to %#x %#x %#x %#x %#x %#x %#x %#x\n",
+			m->compute_static_thread_mgmt_se0,
+			m->compute_static_thread_mgmt_se1,
+			m->compute_static_thread_mgmt_se2,
+			m->compute_static_thread_mgmt_se3,
+			m->compute_static_thread_mgmt_se4,
+			m->compute_static_thread_mgmt_se5,
+			m->compute_static_thread_mgmt_se6,
+			m->compute_static_thread_mgmt_se7);
+	} else {
+		pr_debug("inst: %u, update cu mask to %#x %#x %#x %#x\n",
+			inst, m->compute_static_thread_mgmt_se0,
+			m->compute_static_thread_mgmt_se1,
+			m->compute_static_thread_mgmt_se2,
+			m->compute_static_thread_mgmt_se3);
+	}
 }
 
 static void set_priority(struct v9_mqd *m, struct queue_properties *q)
@@ -82,7 +109,7 @@ static void set_priority(struct v9_mqd *m, struct queue_properties *q)
 	m->cp_hqd_queue_priority = q->priority;
 }
 
-static struct kfd_mem_obj *allocate_mqd(struct kfd_dev *kfd,
+static struct kfd_mem_obj *allocate_mqd(struct kfd_node *node,
 		struct queue_properties *q)
 {
 	int retval;
@@ -104,28 +131,30 @@ static struct kfd_mem_obj *allocate_mqd(struct kfd_dev *kfd,
 	 * pass a special bo flag AMDGPU_GEM_CREATE_CP_MQD_GFX9 to instruct
 	 * amdgpu memory functions to do so.
 	 */
-	if (kfd->cwsr_enabled && (q->type == KFD_QUEUE_TYPE_COMPUTE)) {
+	if (node->kfd->cwsr_enabled && (q->type == KFD_QUEUE_TYPE_COMPUTE)) {
 		mqd_mem_obj = kzalloc(sizeof(struct kfd_mem_obj), GFP_KERNEL);
 		if (!mqd_mem_obj)
 			return NULL;
-		retval = amdgpu_amdkfd_alloc_gtt_mem(kfd->kgd,
-			ALIGN(q->ctl_stack_size, PAGE_SIZE) +
-				ALIGN(sizeof(struct v9_mqd), PAGE_SIZE),
+		retval = amdgpu_amdkfd_alloc_gtt_mem(node->adev,
+			(ALIGN(q->ctl_stack_size, PAGE_SIZE) +
+			ALIGN(sizeof(struct v9_mqd), PAGE_SIZE)) *
+			NUM_XCC(node->xcc_mask),
 			&(mqd_mem_obj->gtt_mem),
 			&(mqd_mem_obj->gpu_addr),
 			(void *)&(mqd_mem_obj->cpu_ptr), true);
+
+		if (retval) {
+			kfree(mqd_mem_obj);
+			return NULL;
+		}
 	} else {
-		retval = kfd_gtt_sa_allocate(kfd, sizeof(struct v9_mqd),
+		retval = kfd_gtt_sa_allocate(node, sizeof(struct v9_mqd),
 				&mqd_mem_obj);
-	}
-
-	if (retval) {
-		kfree(mqd_mem_obj);
-		return NULL;
+		if (retval)
+			return NULL;
 	}
 
 	return mqd_mem_obj;
-
 }
 
 static void init_mqd(struct mqd_manager *mm, void **mqd,
@@ -154,6 +183,9 @@ static void init_mqd(struct mqd_manager *mm, void **mqd,
 	m->cp_hqd_persistent_state = CP_HQD_PERSISTENT_STATE__PRELOAD_REQ_MASK |
 			0x53 << CP_HQD_PERSISTENT_STATE__PRELOAD_SIZE__SHIFT;
 
+	m->cp_hqd_pq_control = 5 << CP_HQD_PQ_CONTROL__RPTR_BLOCK_SIZE__SHIFT;
+	m->cp_hqd_pq_control |= CP_HQD_PQ_CONTROL__UNORD_DISPATCH_MASK;
+
 	m->cp_mqd_control = 1 << CP_MQD_CONTROL__PRIV_STATE__SHIFT;
 
 	m->cp_mqd_base_addr_lo        = lower_32_bits(addr);
@@ -163,17 +195,21 @@ static void init_mqd(struct mqd_manager *mm, void **mqd,
 			1 << CP_HQD_QUANTUM__QUANTUM_SCALE__SHIFT |
 			1 << CP_HQD_QUANTUM__QUANTUM_DURATION__SHIFT;
 
-	if (q->format == KFD_QUEUE_FORMAT_AQL) {
+	/* Set cp_hqd_hq_scheduler0 bit 14 to 1 to have the CP set up the
+	 * DISPATCH_PTR.  This is required for the kfd debugger
+	 */
+	m->cp_hqd_hq_status0 = 1 << 14;
+
+	if (q->format == KFD_QUEUE_FORMAT_AQL)
 		m->cp_hqd_aql_control =
 			1 << CP_HQD_AQL_CONTROL__CONTROL0__SHIFT;
-	}
 
 	if (q->tba_addr) {
 		m->compute_pgm_rsrc2 |=
 			(1 << COMPUTE_PGM_RSRC2__TRAP_PRESENT__SHIFT);
 	}
 
-	if (mm->dev->cwsr_enabled && q->ctx_save_restore_area_address) {
+	if (mm->dev->kfd->cwsr_enabled && q->ctx_save_restore_area_address) {
 		m->cp_hqd_persistent_state |=
 			(1 << CP_HQD_PERSISTENT_STATE__QSWITCH_MODE__SHIFT);
 		m->cp_hqd_ctx_save_base_addr_lo =
@@ -189,7 +225,7 @@ static void init_mqd(struct mqd_manager *mm, void **mqd,
 	*mqd = m;
 	if (gart_addr)
 		*gart_addr = addr;
-	mm->update_mqd(mm, m, q, NULL);
+	update_mqd(mm, m, q, NULL);
 }
 
 static int load_mqd(struct mqd_manager *mm, void *mqd,
@@ -199,17 +235,9 @@ static int load_mqd(struct mqd_manager *mm, void *mqd,
 	/* AQL write pointer counts in 64B packets, PM4/CP counts in dwords. */
 	uint32_t wptr_shift = (p->format == KFD_QUEUE_FORMAT_AQL ? 4 : 0);
 
-	return mm->dev->kfd2kgd->hqd_load(mm->dev->kgd, mqd, pipe_id, queue_id,
+	return mm->dev->kfd2kgd->hqd_load(mm->dev->adev, mqd, pipe_id, queue_id,
 					  (uint32_t __user *)p->write_ptr,
-					  wptr_shift, 0, mms);
-}
-
-static int hiq_load_mqd_kiq(struct mqd_manager *mm, void *mqd,
-			    uint32_t pipe_id, uint32_t queue_id,
-			    struct queue_properties *p, struct mm_struct *mms)
-{
-	return mm->dev->kfd2kgd->hiq_mqd_load(mm->dev->kgd, mqd, pipe_id,
-					      queue_id, p->doorbell_off);
+					  wptr_shift, 0, mms, 0);
 }
 
 static void update_mqd(struct mqd_manager *mm, void *mqd,
@@ -220,7 +248,7 @@ static void update_mqd(struct mqd_manager *mm, void *mqd,
 
 	m = get_mqd(mqd);
 
-	m->cp_hqd_pq_control = 5 << CP_HQD_PQ_CONTROL__RPTR_BLOCK_SIZE__SHIFT;
+	m->cp_hqd_pq_control &= ~CP_HQD_PQ_CONTROL__QUEUE_SIZE_MASK;
 	m->cp_hqd_pq_control |= order_base_2(q->queue_size / 4) - 1;
 	pr_debug("cp_hqd_pq_control 0x%x\n", m->cp_hqd_pq_control);
 
@@ -248,9 +276,14 @@ static void update_mqd(struct mqd_manager *mm, void *mqd,
 	 * Limit is 0xFF EOP entries (= 0x7F8 dwords). CP will not submit
 	 * more than (EOP entry count - 1) so a queue size of 0x800 dwords
 	 * is safe, giving a maximum field value of 0xA.
+	 *
+	 * Also, do calculation only if EOP is used (size > 0), otherwise
+	 * the order_base_2 calculation provides incorrect result.
+	 *
 	 */
-	m->cp_hqd_eop_control = min(0xA,
-		order_base_2(q->eop_ring_buffer_size / 4) - 1);
+	m->cp_hqd_eop_control = q->eop_ring_buffer_size ?
+		min(0xA, order_base_2(q->eop_ring_buffer_size / 4) - 1) : 0;
+
 	m->cp_hqd_eop_base_addr_lo =
 			lower_32_bits(q->eop_ring_buffer_address >> 8);
 	m->cp_hqd_eop_base_addr_hi =
@@ -268,61 +301,46 @@ static void update_mqd(struct mqd_manager *mm, void *mqd,
 		m->cp_hqd_pq_doorbell_control |= 1 <<
 			CP_HQD_PQ_DOORBELL_CONTROL__DOORBELL_BIF_DROP__SHIFT;
 	}
-	if (mm->dev->cwsr_enabled && q->ctx_save_restore_area_address)
+	if (mm->dev->kfd->cwsr_enabled && q->ctx_save_restore_area_address)
 		m->cp_hqd_ctx_save_control = 0;
 
-	update_cu_mask(mm, mqd, minfo);
+	if (KFD_GC_VERSION(mm->dev) != IP_VERSION(9, 4, 3) &&
+	    KFD_GC_VERSION(mm->dev) != IP_VERSION(9, 4, 4) &&
+	    KFD_GC_VERSION(mm->dev) != IP_VERSION(9, 5, 0))
+		update_cu_mask(mm, mqd, minfo, 0);
 	set_priority(m, q);
 
+	if (minfo && KFD_GC_VERSION(mm->dev) >= IP_VERSION(9, 4, 2)) {
+		if (minfo->update_flag & UPDATE_FLAG_IS_GWS)
+			m->compute_resource_limits |=
+				COMPUTE_RESOURCE_LIMITS__FORCE_SIMD_DIST_MASK;
+		else
+			m->compute_resource_limits &=
+				~COMPUTE_RESOURCE_LIMITS__FORCE_SIMD_DIST_MASK;
+	}
+
 	q->is_active = QUEUE_IS_ACTIVE(*q);
 }
 
 
-static uint32_t read_doorbell_id(void *mqd)
+static bool check_preemption_failed(struct mqd_manager *mm, void *mqd)
 {
 	struct v9_mqd *m = (struct v9_mqd *)mqd;
+	uint32_t doorbell_id = m->queue_doorbell_id0;
 
-	return m->queue_doorbell_id0;
-}
-
-static int destroy_mqd(struct mqd_manager *mm, void *mqd,
-			enum kfd_preempt_type type,
-			unsigned int timeout, uint32_t pipe_id,
-			uint32_t queue_id)
-{
-	return mm->dev->kfd2kgd->hqd_destroy
-		(mm->dev->kgd, mqd, type, timeout,
-		pipe_id, queue_id);
-}
-
-static void free_mqd(struct mqd_manager *mm, void *mqd,
-			struct kfd_mem_obj *mqd_mem_obj)
-{
-	struct kfd_dev *kfd = mm->dev;
-
-	if (mqd_mem_obj->gtt_mem) {
-		amdgpu_amdkfd_free_gtt_mem(kfd->kgd, mqd_mem_obj->gtt_mem);
-		kfree(mqd_mem_obj);
-	} else {
-		kfd_gtt_sa_free(mm->dev, mqd_mem_obj);
-	}
-}
+	m->queue_doorbell_id0 = 0;
 
-static bool is_occupied(struct mqd_manager *mm, void *mqd,
-			uint64_t queue_address,	uint32_t pipe_id,
-			uint32_t queue_id)
-{
-	return mm->dev->kfd2kgd->hqd_is_occupied(
-		mm->dev->kgd, queue_address,
-		pipe_id, queue_id);
+	return kfd_check_hiq_mqd_doorbell_id(mm->dev, doorbell_id, 0);
 }
 
 static int get_wave_state(struct mqd_manager *mm, void *mqd,
+			  struct queue_properties *q,
 			  void __user *ctl_stack,
 			  u32 *ctl_stack_used_size,
 			  u32 *save_area_used_size)
 {
 	struct v9_mqd *m;
+	struct kfd_context_save_area_header header;
 
 	/* Control stack is located one page after MQD. */
 	void *mqd_ctl_stack = (void *)((uintptr_t)mqd + PAGE_SIZE);
@@ -334,12 +352,92 @@ static int get_wave_state(struct mqd_manager *mm, void *mqd,
 	*save_area_used_size = m->cp_hqd_wg_state_offset -
 		m->cp_hqd_cntl_stack_size;
 
-	if (copy_to_user(ctl_stack, mqd_ctl_stack, m->cp_hqd_cntl_stack_size))
+	header.wave_state.control_stack_size = *ctl_stack_used_size;
+	header.wave_state.wave_state_size = *save_area_used_size;
+
+	header.wave_state.wave_state_offset = m->cp_hqd_wg_state_offset;
+	header.wave_state.control_stack_offset = m->cp_hqd_cntl_stack_offset;
+
+	if (copy_to_user(ctl_stack, &header, sizeof(header.wave_state)))
+		return -EFAULT;
+
+	if (copy_to_user(ctl_stack + m->cp_hqd_cntl_stack_offset,
+				mqd_ctl_stack + m->cp_hqd_cntl_stack_offset,
+				*ctl_stack_used_size))
 		return -EFAULT;
 
 	return 0;
 }
 
+static void get_checkpoint_info(struct mqd_manager *mm, void *mqd, u32 *ctl_stack_size)
+{
+	struct v9_mqd *m = get_mqd(mqd);
+
+	*ctl_stack_size = m->cp_hqd_cntl_stack_size * NUM_XCC(mm->dev->xcc_mask);
+}
+
+static void checkpoint_mqd(struct mqd_manager *mm, void *mqd, void *mqd_dst, void *ctl_stack_dst)
+{
+	struct v9_mqd *m;
+	/* Control stack is located one page after MQD. */
+	void *ctl_stack = (void *)((uintptr_t)mqd + PAGE_SIZE);
+
+	m = get_mqd(mqd);
+
+	memcpy(mqd_dst, m, sizeof(struct v9_mqd));
+	memcpy(ctl_stack_dst, ctl_stack, m->cp_hqd_cntl_stack_size);
+}
+
+static void checkpoint_mqd_v9_4_3(struct mqd_manager *mm,
+								  void *mqd,
+								  void *mqd_dst,
+								  void *ctl_stack_dst)
+{
+	struct v9_mqd *m;
+	int xcc;
+	uint64_t size = get_mqd(mqd)->cp_mqd_stride_size;
+
+	for (xcc = 0; xcc < NUM_XCC(mm->dev->xcc_mask); xcc++) {
+		m = get_mqd(mqd + size * xcc);
+
+		checkpoint_mqd(mm, m,
+				(uint8_t *)mqd_dst + sizeof(*m) * xcc,
+				(uint8_t *)ctl_stack_dst + m->cp_hqd_cntl_stack_size * xcc);
+	}
+}
+
+static void restore_mqd(struct mqd_manager *mm, void **mqd,
+			struct kfd_mem_obj *mqd_mem_obj, uint64_t *gart_addr,
+			struct queue_properties *qp,
+			const void *mqd_src,
+			const void *ctl_stack_src, u32 ctl_stack_size)
+{
+	uint64_t addr;
+	struct v9_mqd *m;
+	void *ctl_stack;
+
+	m = (struct v9_mqd *) mqd_mem_obj->cpu_ptr;
+	addr = mqd_mem_obj->gpu_addr;
+
+	memcpy(m, mqd_src, sizeof(*m));
+
+	*mqd = m;
+	if (gart_addr)
+		*gart_addr = addr;
+
+	/* Control stack is located one page after MQD. */
+	ctl_stack = (void *)((uintptr_t)*mqd + PAGE_SIZE);
+	memcpy(ctl_stack, ctl_stack_src, ctl_stack_size);
+
+	m->cp_hqd_pq_doorbell_control =
+		qp->doorbell_off <<
+			CP_HQD_PQ_DOORBELL_CONTROL__DOORBELL_OFFSET__SHIFT;
+	pr_debug("cp_hqd_pq_doorbell_control 0x%x\n",
+				m->cp_hqd_pq_doorbell_control);
+
+	qp->is_active = 0;
+}
+
 static void init_mqd_hiq(struct mqd_manager *mm, void **mqd,
 			struct kfd_mem_obj *mqd_mem_obj, uint64_t *gart_addr,
 			struct queue_properties *q)
@@ -354,6 +452,25 @@ static void init_mqd_hiq(struct mqd_manager *mm, void **mqd,
 			1 << CP_HQD_PQ_CONTROL__KMD_QUEUE__SHIFT;
 }
 
+static int destroy_hiq_mqd(struct mqd_manager *mm, void *mqd,
+			enum kfd_preempt_type type, unsigned int timeout,
+			uint32_t pipe_id, uint32_t queue_id)
+{
+	int err;
+	struct v9_mqd *m;
+	u32 doorbell_off;
+
+	m = get_mqd(mqd);
+
+	doorbell_off = m->cp_hqd_pq_doorbell_control >>
+			CP_HQD_PQ_DOORBELL_CONTROL__DOORBELL_OFFSET__SHIFT;
+	err = amdgpu_amdkfd_unmap_hiq(mm->dev->adev, doorbell_off, 0);
+	if (err)
+		pr_debug("Destroy HIQ MQD failed: %d\n", err);
+
+	return err;
+}
+
 static void init_mqd_sdma(struct mqd_manager *mm, void **mqd,
 		struct kfd_mem_obj *mqd_mem_obj, uint64_t *gart_addr,
 		struct queue_properties *q)
@@ -371,15 +488,6 @@ static void init_mqd_sdma(struct mqd_manager *mm, void **mqd,
 	mm->update_mqd(mm, m, q, NULL);
 }
 
-static int load_mqd_sdma(struct mqd_manager *mm, void *mqd,
-		uint32_t pipe_id, uint32_t queue_id,
-		struct queue_properties *p, struct mm_struct *mms)
-{
-	return mm->dev->kfd2kgd->hqd_sdma_load(mm->dev->kgd, mqd,
-					       (uint32_t __user *)p->write_ptr,
-					       mms);
-}
-
 #define SDMA_RLC_DUMMY_DEFAULT 0xf
 
 static void update_mqd_sdma(struct mqd_manager *mm, void *mqd,
@@ -405,27 +513,399 @@ static void update_mqd_sdma(struct mqd_manager *mm, void *mqd,
 	m->sdma_engine_id = q->sdma_engine_id;
 	m->sdma_queue_id = q->sdma_queue_id;
 	m->sdmax_rlcx_dummy_reg = SDMA_RLC_DUMMY_DEFAULT;
+	/* Allow context switch so we don't cross-process starve with a massive
+	 * command buffer of long-running SDMA commands
+	 */
+	m->sdmax_rlcx_ib_cntl |= SDMA0_GFX_IB_CNTL__SWITCH_INSIDE_IB_MASK;
 
 	q->is_active = QUEUE_IS_ACTIVE(*q);
 }
 
-/*
- *  * preempt type here is ignored because there is only one way
- *  * to preempt sdma queue
- */
-static int destroy_mqd_sdma(struct mqd_manager *mm, void *mqd,
-		enum kfd_preempt_type type,
-		unsigned int timeout, uint32_t pipe_id,
-		uint32_t queue_id)
+static void checkpoint_mqd_sdma(struct mqd_manager *mm,
+				void *mqd,
+				void *mqd_dst,
+				void *ctl_stack_dst)
 {
-	return mm->dev->kfd2kgd->hqd_sdma_destroy(mm->dev->kgd, mqd, timeout);
+	struct v9_sdma_mqd *m;
+
+	m = get_sdma_mqd(mqd);
+
+	memcpy(mqd_dst, m, sizeof(struct v9_sdma_mqd));
+}
+
+static void restore_mqd_sdma(struct mqd_manager *mm, void **mqd,
+			     struct kfd_mem_obj *mqd_mem_obj, uint64_t *gart_addr,
+			     struct queue_properties *qp,
+			     const void *mqd_src,
+			     const void *ctl_stack_src, const u32 ctl_stack_size)
+{
+	uint64_t addr;
+	struct v9_sdma_mqd *m;
+
+	m = (struct v9_sdma_mqd *) mqd_mem_obj->cpu_ptr;
+	addr = mqd_mem_obj->gpu_addr;
+
+	memcpy(m, mqd_src, sizeof(*m));
+
+	m->sdmax_rlcx_doorbell_offset =
+		qp->doorbell_off << SDMA0_RLC0_DOORBELL_OFFSET__OFFSET__SHIFT;
+
+	*mqd = m;
+	if (gart_addr)
+		*gart_addr = addr;
+
+	qp->is_active = 0;
+}
+
+static void init_mqd_hiq_v9_4_3(struct mqd_manager *mm, void **mqd,
+			struct kfd_mem_obj *mqd_mem_obj, uint64_t *gart_addr,
+			struct queue_properties *q)
+{
+	struct v9_mqd *m;
+	int xcc = 0;
+	struct kfd_mem_obj xcc_mqd_mem_obj;
+	uint64_t xcc_gart_addr = 0;
+
+	memset(&xcc_mqd_mem_obj, 0x0, sizeof(struct kfd_mem_obj));
+
+	for (xcc = 0; xcc < NUM_XCC(mm->dev->xcc_mask); xcc++) {
+		kfd_get_hiq_xcc_mqd(mm->dev, &xcc_mqd_mem_obj, xcc);
+
+		init_mqd(mm, (void **)&m, &xcc_mqd_mem_obj, &xcc_gart_addr, q);
+
+		m->cp_hqd_pq_control |= CP_HQD_PQ_CONTROL__NO_UPDATE_RPTR_MASK |
+					1 << CP_HQD_PQ_CONTROL__PRIV_STATE__SHIFT |
+					1 << CP_HQD_PQ_CONTROL__KMD_QUEUE__SHIFT;
+		if (amdgpu_sriov_multi_vf_mode(mm->dev->adev))
+			m->cp_hqd_pq_doorbell_control |= 1 <<
+				CP_HQD_PQ_DOORBELL_CONTROL__DOORBELL_MODE__SHIFT;
+		m->cp_mqd_stride_size = kfd_hiq_mqd_stride(mm->dev);
+		if (xcc == 0) {
+			/* Set no_update_rptr = 0 in Master XCC */
+			m->cp_hqd_pq_control &= ~CP_HQD_PQ_CONTROL__NO_UPDATE_RPTR_MASK;
+
+			/* Set the MQD pointer and gart address to XCC0 MQD */
+			*mqd = m;
+			*gart_addr = xcc_gart_addr;
+		}
+	}
+}
+
+static int hiq_load_mqd_kiq_v9_4_3(struct mqd_manager *mm, void *mqd,
+			uint32_t pipe_id, uint32_t queue_id,
+			struct queue_properties *p, struct mm_struct *mms)
+{
+	uint32_t xcc_mask = mm->dev->xcc_mask;
+	int xcc_id, err, inst = 0;
+	void *xcc_mqd;
+	uint64_t hiq_mqd_size = kfd_hiq_mqd_stride(mm->dev);
+
+	for_each_inst(xcc_id, xcc_mask) {
+		xcc_mqd = mqd + hiq_mqd_size * inst;
+		err = mm->dev->kfd2kgd->hiq_mqd_load(mm->dev->adev, xcc_mqd,
+						     pipe_id, queue_id,
+						     p->doorbell_off, xcc_id);
+		if (err) {
+			pr_debug("Failed to load HIQ MQD for XCC: %d\n", inst);
+			break;
+		}
+		++inst;
+	}
+
+	return err;
+}
+
+static int destroy_hiq_mqd_v9_4_3(struct mqd_manager *mm, void *mqd,
+			enum kfd_preempt_type type, unsigned int timeout,
+			uint32_t pipe_id, uint32_t queue_id)
+{
+	uint32_t xcc_mask = mm->dev->xcc_mask;
+	int xcc_id, err, inst = 0;
+	uint64_t hiq_mqd_size = kfd_hiq_mqd_stride(mm->dev);
+	struct v9_mqd *m;
+	u32 doorbell_off;
+
+	for_each_inst(xcc_id, xcc_mask) {
+		m = get_mqd(mqd + hiq_mqd_size * inst);
+
+		doorbell_off = m->cp_hqd_pq_doorbell_control >>
+				CP_HQD_PQ_DOORBELL_CONTROL__DOORBELL_OFFSET__SHIFT;
+
+		err = amdgpu_amdkfd_unmap_hiq(mm->dev->adev, doorbell_off, xcc_id);
+		if (err) {
+			pr_debug("Destroy HIQ MQD failed for xcc: %d\n", inst);
+			break;
+		}
+		++inst;
+	}
+
+	return err;
 }
 
-static bool is_occupied_sdma(struct mqd_manager *mm, void *mqd,
-		uint64_t queue_address, uint32_t pipe_id,
-		uint32_t queue_id)
+static bool check_preemption_failed_v9_4_3(struct mqd_manager *mm, void *mqd)
 {
-	return mm->dev->kfd2kgd->hqd_sdma_is_occupied(mm->dev->kgd, mqd);
+	uint64_t hiq_mqd_size = kfd_hiq_mqd_stride(mm->dev);
+	uint32_t xcc_mask = mm->dev->xcc_mask;
+	int inst = 0, xcc_id;
+	struct v9_mqd *m;
+	bool ret = false;
+
+	for_each_inst(xcc_id, xcc_mask) {
+		m = get_mqd(mqd + hiq_mqd_size * inst);
+		ret |= kfd_check_hiq_mqd_doorbell_id(mm->dev,
+					m->queue_doorbell_id0, inst);
+		m->queue_doorbell_id0 = 0;
+		++inst;
+	}
+
+	return ret;
+}
+
+static void get_xcc_mqd(struct kfd_mem_obj *mqd_mem_obj,
+			       struct kfd_mem_obj *xcc_mqd_mem_obj,
+			       uint64_t offset)
+{
+	xcc_mqd_mem_obj->gtt_mem = (offset == 0) ?
+					mqd_mem_obj->gtt_mem : NULL;
+	xcc_mqd_mem_obj->gpu_addr = mqd_mem_obj->gpu_addr + offset;
+	xcc_mqd_mem_obj->cpu_ptr = (uint32_t *)((uintptr_t)mqd_mem_obj->cpu_ptr
+						+ offset);
+}
+
+static void init_mqd_v9_4_3(struct mqd_manager *mm, void **mqd,
+			struct kfd_mem_obj *mqd_mem_obj, uint64_t *gart_addr,
+			struct queue_properties *q)
+{
+	struct v9_mqd *m;
+	int xcc = 0;
+	struct kfd_mem_obj xcc_mqd_mem_obj;
+	uint64_t xcc_gart_addr = 0;
+	uint64_t xcc_ctx_save_restore_area_address;
+	uint64_t offset = mm->mqd_stride(mm, q);
+	uint32_t local_xcc_start = mm->dev->dqm->current_logical_xcc_start++;
+
+	memset(&xcc_mqd_mem_obj, 0x0, sizeof(struct kfd_mem_obj));
+	for (xcc = 0; xcc < NUM_XCC(mm->dev->xcc_mask); xcc++) {
+		get_xcc_mqd(mqd_mem_obj, &xcc_mqd_mem_obj, offset*xcc);
+
+		init_mqd(mm, (void **)&m, &xcc_mqd_mem_obj, &xcc_gart_addr, q);
+		if (amdgpu_sriov_multi_vf_mode(mm->dev->adev))
+				m->cp_hqd_pq_doorbell_control |= 1 <<
+					CP_HQD_PQ_DOORBELL_CONTROL__DOORBELL_MODE__SHIFT;
+		m->cp_mqd_stride_size = offset;
+
+		/*
+		 * Update the CWSR address for each XCC if CWSR is enabled
+		 * and CWSR area is allocated in thunk
+		 */
+		if (mm->dev->kfd->cwsr_enabled &&
+		    q->ctx_save_restore_area_address) {
+			xcc_ctx_save_restore_area_address =
+				q->ctx_save_restore_area_address +
+				(xcc * q->ctx_save_restore_area_size);
+
+			m->cp_hqd_ctx_save_base_addr_lo =
+				lower_32_bits(xcc_ctx_save_restore_area_address);
+			m->cp_hqd_ctx_save_base_addr_hi =
+				upper_32_bits(xcc_ctx_save_restore_area_address);
+		}
+
+		if (q->format == KFD_QUEUE_FORMAT_AQL) {
+			m->compute_tg_chunk_size = 1;
+			m->compute_current_logic_xcc_id =
+					(local_xcc_start + xcc) %
+					NUM_XCC(mm->dev->xcc_mask);
+
+			switch (xcc) {
+			case 0:
+				/* Master XCC */
+				m->cp_hqd_pq_control &=
+					~CP_HQD_PQ_CONTROL__NO_UPDATE_RPTR_MASK;
+				break;
+			default:
+				break;
+			}
+		} else {
+			/* PM4 Queue */
+			m->compute_current_logic_xcc_id = 0;
+			m->compute_tg_chunk_size = 0;
+			m->pm4_target_xcc_in_xcp = q->pm4_target_xcc;
+		}
+
+		if (xcc == 0) {
+			/* Set the MQD pointer and gart address to XCC0 MQD */
+			*mqd = m;
+			*gart_addr = xcc_gart_addr;
+		}
+	}
+}
+
+static void update_mqd_v9_4_3(struct mqd_manager *mm, void *mqd,
+		      struct queue_properties *q, struct mqd_update_info *minfo)
+{
+	struct v9_mqd *m;
+	int xcc = 0;
+	uint64_t size = mm->mqd_stride(mm, q);
+
+	for (xcc = 0; xcc < NUM_XCC(mm->dev->xcc_mask); xcc++) {
+		m = get_mqd(mqd + size * xcc);
+		update_mqd(mm, m, q, minfo);
+
+		if (amdgpu_sriov_multi_vf_mode(mm->dev->adev))
+				m->cp_hqd_pq_doorbell_control |= 1 <<
+					CP_HQD_PQ_DOORBELL_CONTROL__DOORBELL_MODE__SHIFT;
+		update_cu_mask(mm, m, minfo, xcc);
+
+		if (q->format == KFD_QUEUE_FORMAT_AQL) {
+			switch (xcc) {
+			case 0:
+				/* Master XCC */
+				m->cp_hqd_pq_control &=
+					~CP_HQD_PQ_CONTROL__NO_UPDATE_RPTR_MASK;
+				break;
+			default:
+				break;
+			}
+			m->compute_tg_chunk_size = 1;
+		} else {
+			/* PM4 Queue */
+			m->compute_current_logic_xcc_id = 0;
+			m->compute_tg_chunk_size = 0;
+			m->pm4_target_xcc_in_xcp = q->pm4_target_xcc;
+		}
+	}
+}
+
+static void restore_mqd_v9_4_3(struct mqd_manager *mm, void **mqd,
+			struct kfd_mem_obj *mqd_mem_obj, uint64_t *gart_addr,
+			struct queue_properties *qp,
+			const void *mqd_src,
+			const void *ctl_stack_src, u32 ctl_stack_size)
+{
+	struct kfd_mem_obj xcc_mqd_mem_obj;
+	u32 mqd_ctl_stack_size;
+	struct v9_mqd *m;
+	u32 num_xcc;
+	int xcc;
+
+	uint64_t offset = mm->mqd_stride(mm, qp);
+
+	mm->dev->dqm->current_logical_xcc_start++;
+
+	num_xcc = NUM_XCC(mm->dev->xcc_mask);
+	mqd_ctl_stack_size = ctl_stack_size / num_xcc;
+
+	memset(&xcc_mqd_mem_obj, 0x0, sizeof(struct kfd_mem_obj));
+
+	/* Set the MQD pointer and gart address to XCC0 MQD */
+	*mqd = mqd_mem_obj->cpu_ptr;
+	if (gart_addr)
+		*gart_addr = mqd_mem_obj->gpu_addr;
+
+	for (xcc = 0; xcc < num_xcc; xcc++) {
+		get_xcc_mqd(mqd_mem_obj, &xcc_mqd_mem_obj, offset * xcc);
+		restore_mqd(mm, (void **)&m,
+					&xcc_mqd_mem_obj,
+					NULL,
+					qp,
+					(uint8_t *)mqd_src + xcc * sizeof(*m),
+					(uint8_t *)ctl_stack_src + xcc *  mqd_ctl_stack_size,
+					mqd_ctl_stack_size);
+	}
+}
+static int destroy_mqd_v9_4_3(struct mqd_manager *mm, void *mqd,
+		   enum kfd_preempt_type type, unsigned int timeout,
+		   uint32_t pipe_id, uint32_t queue_id)
+{
+	uint32_t xcc_mask = mm->dev->xcc_mask;
+	int xcc_id, err, inst = 0;
+	void *xcc_mqd;
+	struct v9_mqd *m;
+	uint64_t mqd_offset;
+
+	m = get_mqd(mqd);
+	mqd_offset = m->cp_mqd_stride_size;
+
+	for_each_inst(xcc_id, xcc_mask) {
+		xcc_mqd = mqd + mqd_offset * inst;
+		err = mm->dev->kfd2kgd->hqd_destroy(mm->dev->adev, xcc_mqd,
+						    type, timeout, pipe_id,
+						    queue_id, xcc_id);
+		if (err) {
+			pr_debug("Destroy MQD failed for xcc: %d\n", inst);
+			break;
+		}
+		++inst;
+	}
+
+	return err;
+}
+
+static int load_mqd_v9_4_3(struct mqd_manager *mm, void *mqd,
+			uint32_t pipe_id, uint32_t queue_id,
+			struct queue_properties *p, struct mm_struct *mms)
+{
+	/* AQL write pointer counts in 64B packets, PM4/CP counts in dwords. */
+	uint32_t wptr_shift = (p->format == KFD_QUEUE_FORMAT_AQL ? 4 : 0);
+	uint32_t xcc_mask = mm->dev->xcc_mask;
+	int xcc_id, err, inst = 0;
+	void *xcc_mqd;
+	uint64_t mqd_stride_size = mm->mqd_stride(mm, p);
+
+	for_each_inst(xcc_id, xcc_mask) {
+		xcc_mqd = mqd + mqd_stride_size * inst;
+		err = mm->dev->kfd2kgd->hqd_load(
+			mm->dev->adev, xcc_mqd, pipe_id, queue_id,
+			(uint32_t __user *)p->write_ptr, wptr_shift, 0, mms,
+			xcc_id);
+		if (err) {
+			pr_debug("Load MQD failed for xcc: %d\n", inst);
+			break;
+		}
+		++inst;
+	}
+
+	return err;
+}
+
+static int get_wave_state_v9_4_3(struct mqd_manager *mm, void *mqd,
+				 struct queue_properties *q,
+				 void __user *ctl_stack,
+				 u32 *ctl_stack_used_size,
+				 u32 *save_area_used_size)
+{
+	int xcc, err = 0;
+	void *xcc_mqd;
+	void __user *xcc_ctl_stack;
+	uint64_t mqd_stride_size = mm->mqd_stride(mm, q);
+	u32 tmp_ctl_stack_used_size = 0, tmp_save_area_used_size = 0;
+
+	for (xcc = 0; xcc < NUM_XCC(mm->dev->xcc_mask); xcc++) {
+		xcc_mqd = mqd + mqd_stride_size * xcc;
+		xcc_ctl_stack = (void __user *)((uintptr_t)ctl_stack +
+					q->ctx_save_restore_area_size * xcc);
+
+		err = get_wave_state(mm, xcc_mqd, q, xcc_ctl_stack,
+				     &tmp_ctl_stack_used_size,
+				     &tmp_save_area_used_size);
+		if (err)
+			break;
+
+		/*
+		 * Set the ctl_stack_used_size and save_area_used_size to
+		 * ctl_stack_used_size and save_area_used_size of XCC 0 when
+		 * passing the info the user-space.
+		 * For multi XCC, user-space would have to look at the header
+		 * info of each Control stack area to determine the control
+		 * stack size and save area used.
+		 */
+		if (xcc == 0) {
+			*ctl_stack_used_size = tmp_ctl_stack_used_size;
+			*save_area_used_size = tmp_save_area_used_size;
+		}
+	}
+
+	return err;
 }
 
 #if defined(CONFIG_DEBUG_FS)
@@ -447,7 +927,7 @@ static int debugfs_show_mqd_sdma(struct seq_file *m, void *data)
 #endif
 
 struct mqd_manager *mqd_manager_init_v9(enum KFD_MQD_TYPE type,
-		struct kfd_dev *dev)
+		struct kfd_node *dev)
 {
 	struct mqd_manager *mqd;
 
@@ -463,40 +943,67 @@ struct mqd_manager *mqd_manager_init_v9(enum KFD_MQD_TYPE type,
 	switch (type) {
 	case KFD_MQD_TYPE_CP:
 		mqd->allocate_mqd = allocate_mqd;
-		mqd->init_mqd = init_mqd;
-		mqd->free_mqd = free_mqd;
-		mqd->load_mqd = load_mqd;
-		mqd->update_mqd = update_mqd;
-		mqd->destroy_mqd = destroy_mqd;
-		mqd->is_occupied = is_occupied;
-		mqd->get_wave_state = get_wave_state;
+		mqd->free_mqd = kfd_free_mqd_cp;
+		mqd->is_occupied = kfd_is_occupied_cp;
+		mqd->get_checkpoint_info = get_checkpoint_info;
 		mqd->mqd_size = sizeof(struct v9_mqd);
+		mqd->mqd_stride = mqd_stride_v9;
 #if defined(CONFIG_DEBUG_FS)
 		mqd->debugfs_show_mqd = debugfs_show_mqd;
 #endif
+		if (KFD_GC_VERSION(dev) == IP_VERSION(9, 4, 3) ||
+		    KFD_GC_VERSION(dev) == IP_VERSION(9, 4, 4) ||
+		    KFD_GC_VERSION(dev) == IP_VERSION(9, 5, 0)) {
+			mqd->init_mqd = init_mqd_v9_4_3;
+			mqd->load_mqd = load_mqd_v9_4_3;
+			mqd->update_mqd = update_mqd_v9_4_3;
+			mqd->destroy_mqd = destroy_mqd_v9_4_3;
+			mqd->get_wave_state = get_wave_state_v9_4_3;
+			mqd->checkpoint_mqd = checkpoint_mqd_v9_4_3;
+			mqd->restore_mqd = restore_mqd_v9_4_3;
+		} else {
+			mqd->init_mqd = init_mqd;
+			mqd->load_mqd = load_mqd;
+			mqd->update_mqd = update_mqd;
+			mqd->destroy_mqd = kfd_destroy_mqd_cp;
+			mqd->get_wave_state = get_wave_state;
+			mqd->checkpoint_mqd = checkpoint_mqd;
+			mqd->restore_mqd = restore_mqd;
+		}
 		break;
 	case KFD_MQD_TYPE_HIQ:
 		mqd->allocate_mqd = allocate_hiq_mqd;
-		mqd->init_mqd = init_mqd_hiq;
 		mqd->free_mqd = free_mqd_hiq_sdma;
-		mqd->load_mqd = hiq_load_mqd_kiq;
 		mqd->update_mqd = update_mqd;
-		mqd->destroy_mqd = destroy_mqd;
-		mqd->is_occupied = is_occupied;
+		mqd->is_occupied = kfd_is_occupied_cp;
 		mqd->mqd_size = sizeof(struct v9_mqd);
+		mqd->mqd_stride = kfd_mqd_stride;
 #if defined(CONFIG_DEBUG_FS)
 		mqd->debugfs_show_mqd = debugfs_show_mqd;
 #endif
-		mqd->read_doorbell_id = read_doorbell_id;
+		mqd->check_preemption_failed = check_preemption_failed;
+		if (KFD_GC_VERSION(dev) == IP_VERSION(9, 4, 3) ||
+		    KFD_GC_VERSION(dev) == IP_VERSION(9, 4, 4) ||
+		    KFD_GC_VERSION(dev) == IP_VERSION(9, 5, 0)) {
+			mqd->init_mqd = init_mqd_hiq_v9_4_3;
+			mqd->load_mqd = hiq_load_mqd_kiq_v9_4_3;
+			mqd->destroy_mqd = destroy_hiq_mqd_v9_4_3;
+			mqd->check_preemption_failed = check_preemption_failed_v9_4_3;
+		} else {
+			mqd->init_mqd = init_mqd_hiq;
+			mqd->load_mqd = kfd_hiq_load_mqd_kiq;
+			mqd->destroy_mqd = destroy_hiq_mqd;
+			mqd->check_preemption_failed = check_preemption_failed;
+		}
 		break;
 	case KFD_MQD_TYPE_DIQ:
 		mqd->allocate_mqd = allocate_mqd;
 		mqd->init_mqd = init_mqd_hiq;
-		mqd->free_mqd = free_mqd;
+		mqd->free_mqd = kfd_free_mqd_cp;
 		mqd->load_mqd = load_mqd;
 		mqd->update_mqd = update_mqd;
-		mqd->destroy_mqd = destroy_mqd;
-		mqd->is_occupied = is_occupied;
+		mqd->destroy_mqd = kfd_destroy_mqd_cp;
+		mqd->is_occupied = kfd_is_occupied_cp;
 		mqd->mqd_size = sizeof(struct v9_mqd);
 #if defined(CONFIG_DEBUG_FS)
 		mqd->debugfs_show_mqd = debugfs_show_mqd;
@@ -506,11 +1013,14 @@ struct mqd_manager *mqd_manager_init_v9(enum KFD_MQD_TYPE type,
 		mqd->allocate_mqd = allocate_sdma_mqd;
 		mqd->init_mqd = init_mqd_sdma;
 		mqd->free_mqd = free_mqd_hiq_sdma;
-		mqd->load_mqd = load_mqd_sdma;
+		mqd->load_mqd = kfd_load_mqd_sdma;
 		mqd->update_mqd = update_mqd_sdma;
-		mqd->destroy_mqd = destroy_mqd_sdma;
-		mqd->is_occupied = is_occupied_sdma;
+		mqd->destroy_mqd = kfd_destroy_mqd_sdma;
+		mqd->is_occupied = kfd_is_occupied_sdma;
+		mqd->checkpoint_mqd = checkpoint_mqd_sdma;
+		mqd->restore_mqd = restore_mqd_sdma;
 		mqd->mqd_size = sizeof(struct v9_sdma_mqd);
+		mqd->mqd_stride = kfd_mqd_stride;
 #if defined(CONFIG_DEBUG_FS)
 		mqd->debugfs_show_mqd = debugfs_show_mqd_sdma;
 #endif
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_vi.c b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_vi.c
index cd9220eb8a7a..c1fafc502515 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_vi.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_vi.c
@@ -1,5 +1,6 @@
+// SPDX-License-Identifier: GPL-2.0 OR MIT
 /*
- * Copyright 2014 Advanced Micro Devices, Inc.
+ * Copyright 2014-2022 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -50,12 +51,11 @@ static void update_cu_mask(struct mqd_manager *mm, void *mqd,
 	struct vi_mqd *m;
 	uint32_t se_mask[4] = {0}; /* 4 is the max # of SEs */
 
-	if (!minfo || (minfo->update_flag != UPDATE_FLAG_CU_MASK) ||
-	    !minfo->cu_mask.ptr)
+	if (!minfo || !minfo->cu_mask.ptr)
 		return;
 
 	mqd_symmetrically_map_cu_mask(mm,
-		minfo->cu_mask.ptr, minfo->cu_mask.count, se_mask);
+		minfo->cu_mask.ptr, minfo->cu_mask.count, se_mask, 0);
 
 	m = get_mqd(mqd);
 	m->compute_static_thread_mgmt_se0 = se_mask[0];
@@ -76,7 +76,7 @@ static void set_priority(struct vi_mqd *m, struct queue_properties *q)
 	m->cp_hqd_queue_priority = q->priority;
 }
 
-static struct kfd_mem_obj *allocate_mqd(struct kfd_dev *kfd,
+static struct kfd_mem_obj *allocate_mqd(struct kfd_node *kfd,
 					struct queue_properties *q)
 {
 	struct kfd_mem_obj *mqd_mem_obj;
@@ -135,7 +135,7 @@ static void init_mqd(struct mqd_manager *mm, void **mqd,
 			(1 << COMPUTE_PGM_RSRC2__TRAP_PRESENT__SHIFT);
 	}
 
-	if (mm->dev->cwsr_enabled && q->ctx_save_restore_area_address) {
+	if (mm->dev->kfd->cwsr_enabled && q->ctx_save_restore_area_address) {
 		m->cp_hqd_persistent_state |=
 			(1 << CP_HQD_PERSISTENT_STATE__QSWITCH_MODE__SHIFT);
 		m->cp_hqd_ctx_save_base_addr_lo =
@@ -162,9 +162,9 @@ static int load_mqd(struct mqd_manager *mm, void *mqd,
 	uint32_t wptr_shift = (p->format == KFD_QUEUE_FORMAT_AQL ? 4 : 0);
 	uint32_t wptr_mask = (uint32_t)((p->queue_size / 4) - 1);
 
-	return mm->dev->kfd2kgd->hqd_load(mm->dev->kgd, mqd, pipe_id, queue_id,
+	return mm->dev->kfd2kgd->hqd_load(mm->dev->adev, mqd, pipe_id, queue_id,
 					  (uint32_t __user *)p->write_ptr,
-					  wptr_shift, wptr_mask, mms);
+					  wptr_shift, wptr_mask, mms, 0);
 }
 
 static void __update_mqd(struct mqd_manager *mm, void *mqd,
@@ -226,7 +226,7 @@ static void __update_mqd(struct mqd_manager *mm, void *mqd,
 				2 << CP_HQD_PQ_CONTROL__SLOT_BASED_WPTR__SHIFT;
 	}
 
-	if (mm->dev->cwsr_enabled && q->ctx_save_restore_area_address)
+	if (mm->dev->kfd->cwsr_enabled && q->ctx_save_restore_area_address)
 		m->cp_hqd_ctx_save_control =
 			atc_bit << CP_HQD_CTX_SAVE_CONTROL__ATC__SHIFT |
 			mtype << CP_HQD_CTX_SAVE_CONTROL__MTYPE__SHIFT;
@@ -237,54 +237,22 @@ static void __update_mqd(struct mqd_manager *mm, void *mqd,
 	q->is_active = QUEUE_IS_ACTIVE(*q);
 }
 
-
-static void update_mqd(struct mqd_manager *mm, void *mqd,
-			struct queue_properties *q,
-			struct mqd_update_info *minfo)
-{
-	__update_mqd(mm, mqd, q, minfo, MTYPE_CC, 1);
-}
-
-static uint32_t read_doorbell_id(void *mqd)
+static bool check_preemption_failed(struct mqd_manager *mm, void *mqd)
 {
 	struct vi_mqd *m = (struct vi_mqd *)mqd;
 
-	return m->queue_doorbell_id0;
+	return kfd_check_hiq_mqd_doorbell_id(mm->dev, m->queue_doorbell_id0, 0);
 }
 
-static void update_mqd_tonga(struct mqd_manager *mm, void *mqd,
-			struct queue_properties *q,
-			struct mqd_update_info *minfo)
+static void update_mqd(struct mqd_manager *mm, void *mqd,
+		       struct queue_properties *q,
+		       struct mqd_update_info *minfo)
 {
 	__update_mqd(mm, mqd, q, minfo, MTYPE_UC, 0);
 }
 
-static int destroy_mqd(struct mqd_manager *mm, void *mqd,
-			enum kfd_preempt_type type,
-			unsigned int timeout, uint32_t pipe_id,
-			uint32_t queue_id)
-{
-	return mm->dev->kfd2kgd->hqd_destroy
-		(mm->dev->kgd, mqd, type, timeout,
-		pipe_id, queue_id);
-}
-
-static void free_mqd(struct mqd_manager *mm, void *mqd,
-			struct kfd_mem_obj *mqd_mem_obj)
-{
-	kfd_gtt_sa_free(mm->dev, mqd_mem_obj);
-}
-
-static bool is_occupied(struct mqd_manager *mm, void *mqd,
-			uint64_t queue_address,	uint32_t pipe_id,
-			uint32_t queue_id)
-{
-	return mm->dev->kfd2kgd->hqd_is_occupied(
-		mm->dev->kgd, queue_address,
-		pipe_id, queue_id);
-}
-
 static int get_wave_state(struct mqd_manager *mm, void *mqd,
+			  struct queue_properties *q,
 			  void __user *ctl_stack,
 			  u32 *ctl_stack_used_size,
 			  u32 *save_area_used_size)
@@ -306,11 +274,54 @@ static int get_wave_state(struct mqd_manager *mm, void *mqd,
 	return 0;
 }
 
+static void get_checkpoint_info(struct mqd_manager *mm, void *mqd, u32 *ctl_stack_size)
+{
+	/* Control stack is stored in user mode */
+	*ctl_stack_size = 0;
+}
+
+static void checkpoint_mqd(struct mqd_manager *mm, void *mqd, void *mqd_dst, void *ctl_stack_dst)
+{
+	struct vi_mqd *m;
+
+	m = get_mqd(mqd);
+
+	memcpy(mqd_dst, m, sizeof(struct vi_mqd));
+}
+
+static void restore_mqd(struct mqd_manager *mm, void **mqd,
+			struct kfd_mem_obj *mqd_mem_obj, uint64_t *gart_addr,
+			struct queue_properties *qp,
+			const void *mqd_src,
+			const void *ctl_stack_src, const u32 ctl_stack_size)
+{
+	uint64_t addr;
+	struct vi_mqd *m;
+
+	m = (struct vi_mqd *) mqd_mem_obj->cpu_ptr;
+	addr = mqd_mem_obj->gpu_addr;
+
+	memcpy(m, mqd_src, sizeof(*m));
+
+	*mqd = m;
+	if (gart_addr)
+		*gart_addr = addr;
+
+	m->cp_hqd_pq_doorbell_control =
+		qp->doorbell_off <<
+			CP_HQD_PQ_DOORBELL_CONTROL__DOORBELL_OFFSET__SHIFT;
+	pr_debug("cp_hqd_pq_doorbell_control 0x%x\n",
+			m->cp_hqd_pq_doorbell_control);
+
+	qp->is_active = 0;
+}
+
 static void init_mqd_hiq(struct mqd_manager *mm, void **mqd,
 			struct kfd_mem_obj *mqd_mem_obj, uint64_t *gart_addr,
 			struct queue_properties *q)
 {
 	struct vi_mqd *m;
+
 	init_mqd(mm, mqd, mqd_mem_obj, gart_addr, q);
 
 	m = get_mqd(*mqd);
@@ -343,15 +354,6 @@ static void init_mqd_sdma(struct mqd_manager *mm, void **mqd,
 	mm->update_mqd(mm, m, q, NULL);
 }
 
-static int load_mqd_sdma(struct mqd_manager *mm, void *mqd,
-		uint32_t pipe_id, uint32_t queue_id,
-		struct queue_properties *p, struct mm_struct *mms)
-{
-	return mm->dev->kfd2kgd->hqd_sdma_load(mm->dev->kgd, mqd,
-					       (uint32_t __user *)p->write_ptr,
-					       mms);
-}
-
 static void update_mqd_sdma(struct mqd_manager *mm, void *mqd,
 			struct queue_properties *q,
 			struct mqd_update_info *minfo)
@@ -380,27 +382,45 @@ static void update_mqd_sdma(struct mqd_manager *mm, void *mqd,
 	q->is_active = QUEUE_IS_ACTIVE(*q);
 }
 
-/*
- *  * preempt type here is ignored because there is only one way
- *  * to preempt sdma queue
- */
-static int destroy_mqd_sdma(struct mqd_manager *mm, void *mqd,
-		enum kfd_preempt_type type,
-		unsigned int timeout, uint32_t pipe_id,
-		uint32_t queue_id)
+static void checkpoint_mqd_sdma(struct mqd_manager *mm,
+				void *mqd,
+				void *mqd_dst,
+				void *ctl_stack_dst)
 {
-	return mm->dev->kfd2kgd->hqd_sdma_destroy(mm->dev->kgd, mqd, timeout);
+	struct vi_sdma_mqd *m;
+
+	m = get_sdma_mqd(mqd);
+
+	memcpy(mqd_dst, m, sizeof(struct vi_sdma_mqd));
 }
 
-static bool is_occupied_sdma(struct mqd_manager *mm, void *mqd,
-		uint64_t queue_address, uint32_t pipe_id,
-		uint32_t queue_id)
+static void restore_mqd_sdma(struct mqd_manager *mm, void **mqd,
+			     struct kfd_mem_obj *mqd_mem_obj, uint64_t *gart_addr,
+			     struct queue_properties *qp,
+			     const void *mqd_src,
+			     const void *ctl_stack_src, const u32 ctl_stack_size)
 {
-	return mm->dev->kfd2kgd->hqd_sdma_is_occupied(mm->dev->kgd, mqd);
+	uint64_t addr;
+	struct vi_sdma_mqd *m;
+
+	m = (struct vi_sdma_mqd *) mqd_mem_obj->cpu_ptr;
+	addr = mqd_mem_obj->gpu_addr;
+
+	memcpy(m, mqd_src, sizeof(*m));
+
+	m->sdmax_rlcx_doorbell =
+		qp->doorbell_off << SDMA0_RLC0_DOORBELL__OFFSET__SHIFT;
+
+	*mqd = m;
+	if (gart_addr)
+		*gart_addr = addr;
+
+	qp->is_active = 0;
 }
 
 #if defined(CONFIG_DEBUG_FS)
 
+
 static int debugfs_show_mqd(struct seq_file *m, void *data)
 {
 	seq_hex_dump(m, "    ", DUMP_PREFIX_OFFSET, 32, 4,
@@ -418,7 +438,7 @@ static int debugfs_show_mqd_sdma(struct seq_file *m, void *data)
 #endif
 
 struct mqd_manager *mqd_manager_init_vi(enum KFD_MQD_TYPE type,
-		struct kfd_dev *dev)
+		struct kfd_node *dev)
 {
 	struct mqd_manager *mqd;
 
@@ -435,12 +455,15 @@ struct mqd_manager *mqd_manager_init_vi(enum KFD_MQD_TYPE type,
 	case KFD_MQD_TYPE_CP:
 		mqd->allocate_mqd = allocate_mqd;
 		mqd->init_mqd = init_mqd;
-		mqd->free_mqd = free_mqd;
+		mqd->free_mqd = kfd_free_mqd_cp;
 		mqd->load_mqd = load_mqd;
 		mqd->update_mqd = update_mqd;
-		mqd->destroy_mqd = destroy_mqd;
-		mqd->is_occupied = is_occupied;
+		mqd->destroy_mqd = kfd_destroy_mqd_cp;
+		mqd->is_occupied = kfd_is_occupied_cp;
 		mqd->get_wave_state = get_wave_state;
+		mqd->get_checkpoint_info = get_checkpoint_info;
+		mqd->checkpoint_mqd = checkpoint_mqd;
+		mqd->restore_mqd = restore_mqd;
 		mqd->mqd_size = sizeof(struct vi_mqd);
 #if defined(CONFIG_DEBUG_FS)
 		mqd->debugfs_show_mqd = debugfs_show_mqd;
@@ -452,23 +475,25 @@ struct mqd_manager *mqd_manager_init_vi(enum KFD_MQD_TYPE type,
 		mqd->free_mqd = free_mqd_hiq_sdma;
 		mqd->load_mqd = load_mqd;
 		mqd->update_mqd = update_mqd_hiq;
-		mqd->destroy_mqd = destroy_mqd;
-		mqd->is_occupied = is_occupied;
+		mqd->destroy_mqd = kfd_destroy_mqd_cp;
+		mqd->is_occupied = kfd_is_occupied_cp;
 		mqd->mqd_size = sizeof(struct vi_mqd);
+		mqd->mqd_stride = kfd_mqd_stride;
 #if defined(CONFIG_DEBUG_FS)
 		mqd->debugfs_show_mqd = debugfs_show_mqd;
 #endif
-		mqd->read_doorbell_id = read_doorbell_id;
+		mqd->check_preemption_failed = check_preemption_failed;
 		break;
 	case KFD_MQD_TYPE_DIQ:
 		mqd->allocate_mqd = allocate_mqd;
 		mqd->init_mqd = init_mqd_hiq;
-		mqd->free_mqd = free_mqd;
+		mqd->free_mqd = kfd_free_mqd_cp;
 		mqd->load_mqd = load_mqd;
 		mqd->update_mqd = update_mqd_hiq;
-		mqd->destroy_mqd = destroy_mqd;
-		mqd->is_occupied = is_occupied;
+		mqd->destroy_mqd = kfd_destroy_mqd_cp;
+		mqd->is_occupied = kfd_is_occupied_cp;
 		mqd->mqd_size = sizeof(struct vi_mqd);
+		mqd->mqd_stride = kfd_mqd_stride;
 #if defined(CONFIG_DEBUG_FS)
 		mqd->debugfs_show_mqd = debugfs_show_mqd;
 #endif
@@ -477,11 +502,14 @@ struct mqd_manager *mqd_manager_init_vi(enum KFD_MQD_TYPE type,
 		mqd->allocate_mqd = allocate_sdma_mqd;
 		mqd->init_mqd = init_mqd_sdma;
 		mqd->free_mqd = free_mqd_hiq_sdma;
-		mqd->load_mqd = load_mqd_sdma;
+		mqd->load_mqd = kfd_load_mqd_sdma;
 		mqd->update_mqd = update_mqd_sdma;
-		mqd->destroy_mqd = destroy_mqd_sdma;
-		mqd->is_occupied = is_occupied_sdma;
+		mqd->destroy_mqd = kfd_destroy_mqd_sdma;
+		mqd->is_occupied = kfd_is_occupied_sdma;
+		mqd->checkpoint_mqd = checkpoint_mqd_sdma;
+		mqd->restore_mqd = restore_mqd_sdma;
 		mqd->mqd_size = sizeof(struct vi_sdma_mqd);
+		mqd->mqd_stride = kfd_mqd_stride;
 #if defined(CONFIG_DEBUG_FS)
 		mqd->debugfs_show_mqd = debugfs_show_mqd_sdma;
 #endif
@@ -493,16 +521,3 @@ struct mqd_manager *mqd_manager_init_vi(enum KFD_MQD_TYPE type,
 
 	return mqd;
 }
-
-struct mqd_manager *mqd_manager_init_vi_tonga(enum KFD_MQD_TYPE type,
-			struct kfd_dev *dev)
-{
-	struct mqd_manager *mqd;
-
-	mqd = mqd_manager_init_vi(type, dev);
-	if (!mqd)
-		return NULL;
-	if (type == KFD_MQD_TYPE_CP)
-		mqd->update_mqd = update_mqd_tonga;
-	return mqd;
-}
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager.c b/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager.c
index e547f1f8c49f..b1a6eb349bb3 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager.c
@@ -1,5 +1,6 @@
+// SPDX-License-Identifier: GPL-2.0 OR MIT
 /*
- * Copyright 2014 Advanced Micro Devices, Inc.
+ * Copyright 2014-2022 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -27,6 +28,11 @@
 #include "kfd_kernel_queue.h"
 #include "kfd_priv.h"
 
+#define OVER_SUBSCRIPTION_PROCESS_COUNT (1 << 0)
+#define OVER_SUBSCRIPTION_COMPUTE_QUEUE_COUNT (1 << 1)
+#define OVER_SUBSCRIPTION_GWS_QUEUE_COUNT (1 << 2)
+#define OVER_SUBSCRIPTION_XNACK_CONFLICT (1 << 3)
+
 static inline void inc_wptr(unsigned int *wptr, unsigned int increment_bytes,
 				unsigned int buffer_size_bytes)
 {
@@ -39,12 +45,14 @@ static inline void inc_wptr(unsigned int *wptr, unsigned int increment_bytes,
 
 static void pm_calc_rlib_size(struct packet_manager *pm,
 				unsigned int *rlib_size,
-				bool *over_subscription)
+				int *over_subscription,
+				int xnack_conflict)
 {
 	unsigned int process_count, queue_count, compute_queue_count, gws_queue_count;
 	unsigned int map_queue_size;
 	unsigned int max_proc_per_quantum = 1;
-	struct kfd_dev *dev = pm->dqm->dev;
+	struct kfd_node *node = pm->dqm->dev;
+	struct device *dev = node->adev->dev;
 
 	process_count = pm->dqm->processes_count;
 	queue_count = pm->dqm->active_queue_count;
@@ -56,17 +64,22 @@ static void pm_calc_rlib_size(struct packet_manager *pm,
 	 * hws_max_conc_proc has been done in
 	 * kgd2kfd_device_init().
 	 */
-	*over_subscription = false;
+	*over_subscription = 0;
 
-	if (dev->max_proc_per_quantum > 1)
-		max_proc_per_quantum = dev->max_proc_per_quantum;
+	if (node->max_proc_per_quantum > 1)
+		max_proc_per_quantum = node->max_proc_per_quantum;
 
-	if ((process_count > max_proc_per_quantum) ||
-	    compute_queue_count > get_cp_queues_num(pm->dqm) ||
-	    gws_queue_count > 1) {
-		*over_subscription = true;
-		pr_debug("Over subscribed runlist\n");
-	}
+	if (process_count > max_proc_per_quantum)
+		*over_subscription |= OVER_SUBSCRIPTION_PROCESS_COUNT;
+	if (compute_queue_count > get_cp_queues_num(pm->dqm))
+		*over_subscription |= OVER_SUBSCRIPTION_COMPUTE_QUEUE_COUNT;
+	if (gws_queue_count > 1)
+		*over_subscription |= OVER_SUBSCRIPTION_GWS_QUEUE_COUNT;
+	if (xnack_conflict && (node->adev->gmc.xnack_flags & AMDGPU_GMC_XNACK_FLAG_CHAIN))
+		*over_subscription |= OVER_SUBSCRIPTION_XNACK_CONFLICT;
+
+	if (*over_subscription)
+		dev_dbg(dev, "Over subscribed runlist\n");
 
 	map_queue_size = pm->pmf->map_queues_size;
 	/* calculate run list ib allocation size */
@@ -80,29 +93,32 @@ static void pm_calc_rlib_size(struct packet_manager *pm,
 	if (*over_subscription)
 		*rlib_size += pm->pmf->runlist_size;
 
-	pr_debug("runlist ib size %d\n", *rlib_size);
+	dev_dbg(dev, "runlist ib size %d\n", *rlib_size);
 }
 
 static int pm_allocate_runlist_ib(struct packet_manager *pm,
 				unsigned int **rl_buffer,
 				uint64_t *rl_gpu_buffer,
 				unsigned int *rl_buffer_size,
-				bool *is_over_subscription)
+				int *is_over_subscription,
+				int xnack_conflict)
 {
+	struct kfd_node *node = pm->dqm->dev;
+	struct device *dev = node->adev->dev;
 	int retval;
 
 	if (WARN_ON(pm->allocated))
 		return -EINVAL;
 
-	pm_calc_rlib_size(pm, rl_buffer_size, is_over_subscription);
+	pm_calc_rlib_size(pm, rl_buffer_size, is_over_subscription,
+				xnack_conflict);
 
 	mutex_lock(&pm->lock);
 
-	retval = kfd_gtt_sa_allocate(pm->dqm->dev, *rl_buffer_size,
-					&pm->ib_buffer_obj);
+	retval = kfd_gtt_sa_allocate(node, *rl_buffer_size, &pm->ib_buffer_obj);
 
 	if (retval) {
-		pr_err("Failed to allocate runlist IB\n");
+		dev_err(dev, "Failed to allocate runlist IB\n");
 		goto out;
 	}
 
@@ -124,32 +140,54 @@ static int pm_create_runlist_ib(struct packet_manager *pm,
 {
 	unsigned int alloc_size_bytes;
 	unsigned int *rl_buffer, rl_wptr, i;
+	struct kfd_node *node = pm->dqm->dev;
+	struct device *dev = node->adev->dev;
 	int retval, processes_mapped;
 	struct device_process_node *cur;
 	struct qcm_process_device *qpd;
 	struct queue *q;
 	struct kernel_queue *kq;
-	bool is_over_subscription;
+	int is_over_subscription;
+	int xnack_enabled = -1;
+	bool xnack_conflict = 0;
 
 	rl_wptr = retval = processes_mapped = 0;
 
+	/* Check if processes set different xnack modes */
+	list_for_each_entry(cur, queues, list) {
+		qpd = cur->qpd;
+		if (xnack_enabled < 0)
+			/* First process */
+			xnack_enabled = qpd->pqm->process->xnack_enabled;
+		else if (qpd->pqm->process->xnack_enabled != xnack_enabled) {
+			/* Found a process with a different xnack mode */
+			xnack_conflict = 1;
+			break;
+		}
+	}
+
 	retval = pm_allocate_runlist_ib(pm, &rl_buffer, rl_gpu_addr,
-				&alloc_size_bytes, &is_over_subscription);
+				&alloc_size_bytes, &is_over_subscription,
+				xnack_conflict);
 	if (retval)
 		return retval;
 
 	*rl_size_bytes = alloc_size_bytes;
 	pm->ib_size_bytes = alloc_size_bytes;
 
-	pr_debug("Building runlist ib process count: %d queues count %d\n",
+	dev_dbg(dev, "Building runlist ib process count: %d queues count %d\n",
 		pm->dqm->processes_count, pm->dqm->active_queue_count);
 
+build_runlist_ib:
 	/* build the run list ib packet */
 	list_for_each_entry(cur, queues, list) {
 		qpd = cur->qpd;
+		/* group processes with the same xnack mode together */
+		if (qpd->pqm->process->xnack_enabled != xnack_enabled)
+			continue;
 		/* build map process packet */
 		if (processes_mapped >= pm->dqm->processes_count) {
-			pr_debug("Not enough space left in runlist IB\n");
+			dev_dbg(dev, "Not enough space left in runlist IB\n");
 			pm_release_ib(pm);
 			return -ENOMEM;
 		}
@@ -166,7 +204,8 @@ static int pm_create_runlist_ib(struct packet_manager *pm,
 			if (!kq->queue->properties.is_active)
 				continue;
 
-			pr_debug("static_queue, mapping kernel q %d, is debug status %d\n",
+			dev_dbg(dev,
+				"static_queue, mapping kernel q %d, is debug status %d\n",
 				kq->queue->queue, qpd->is_debug);
 
 			retval = pm->pmf->map_queues(pm,
@@ -185,7 +224,8 @@ static int pm_create_runlist_ib(struct packet_manager *pm,
 			if (!q->properties.is_active)
 				continue;
 
-			pr_debug("static_queue, mapping user queue %d, is debug status %d\n",
+			dev_dbg(dev,
+				"static_queue, mapping user queue %d, is debug status %d\n",
 				q->queue, qpd->is_debug);
 
 			retval = pm->pmf->map_queues(pm,
@@ -201,18 +241,33 @@ static int pm_create_runlist_ib(struct packet_manager *pm,
 				alloc_size_bytes);
 		}
 	}
+	if (xnack_conflict) {
+		/* pick up processes with the other xnack mode */
+		xnack_enabled = !xnack_enabled;
+		xnack_conflict = 0;
+		goto build_runlist_ib;
+	}
 
-	pr_debug("Finished map process and queues to runlist\n");
+	dev_dbg(dev, "Finished map process and queues to runlist\n");
 
 	if (is_over_subscription) {
 		if (!pm->is_over_subscription)
-			pr_warn("Runlist is getting oversubscribed. Expect reduced ROCm performance.\n");
+			dev_warn(dev, "Runlist is getting oversubscribed due to%s%s%s%s. Expect reduced ROCm performance.\n",
+				is_over_subscription & OVER_SUBSCRIPTION_PROCESS_COUNT ?
+				" too many processes" : "",
+				is_over_subscription & OVER_SUBSCRIPTION_COMPUTE_QUEUE_COUNT ?
+				" too many queues" : "",
+				is_over_subscription & OVER_SUBSCRIPTION_GWS_QUEUE_COUNT ?
+				" multiple processes using cooperative launch" : "",
+				is_over_subscription & OVER_SUBSCRIPTION_XNACK_CONFLICT ?
+				" xnack on/off processes mixed on gfx9" : "");
+
 		retval = pm->pmf->runlist(pm, &rl_buffer[rl_wptr],
 					*rl_gpu_addr,
 					alloc_size_bytes / sizeof(uint32_t),
 					true);
 	}
-	pm->is_over_subscription = is_over_subscription;
+	pm->is_over_subscription = !!is_over_subscription;
 
 	for (i = 0; i < alloc_size_bytes / sizeof(uint32_t); i++)
 		pr_debug("0x%2X ", rl_buffer[i]);
@@ -223,7 +278,7 @@ static int pm_create_runlist_ib(struct packet_manager *pm,
 
 int pm_init(struct packet_manager *pm, struct device_queue_manager *dqm)
 {
-	switch (dqm->dev->device_info->asic_family) {
+	switch (dqm->dev->adev->asic_type) {
 	case CHIP_KAVERI:
 	case CHIP_HAWAII:
 		/* PM4 packet structures on CIK are the same as on VI */
@@ -236,31 +291,19 @@ int pm_init(struct packet_manager *pm, struct device_queue_manager *dqm)
 	case CHIP_VEGAM:
 		pm->pmf = &kfd_vi_pm_funcs;
 		break;
-	case CHIP_VEGA10:
-	case CHIP_VEGA12:
-	case CHIP_VEGA20:
-	case CHIP_RAVEN:
-	case CHIP_RENOIR:
-	case CHIP_ARCTURUS:
-	case CHIP_NAVI10:
-	case CHIP_NAVI12:
-	case CHIP_NAVI14:
-	case CHIP_SIENNA_CICHLID:
-	case CHIP_NAVY_FLOUNDER:
-	case CHIP_VANGOGH:
-	case CHIP_DIMGREY_CAVEFISH:
-	case CHIP_BEIGE_GOBY:
-	case CHIP_YELLOW_CARP:
-	case CHIP_CYAN_SKILLFISH:
-		pm->pmf = &kfd_v9_pm_funcs;
-		break;
-	case CHIP_ALDEBARAN:
-		pm->pmf = &kfd_aldebaran_pm_funcs;
-		break;
 	default:
-		WARN(1, "Unexpected ASIC family %u",
-		     dqm->dev->device_info->asic_family);
-		return -EINVAL;
+		if (KFD_GC_VERSION(dqm->dev) == IP_VERSION(9, 4, 2) ||
+		    KFD_GC_VERSION(dqm->dev) == IP_VERSION(9, 4, 3) ||
+		    KFD_GC_VERSION(dqm->dev) == IP_VERSION(9, 4, 4) ||
+		    KFD_GC_VERSION(dqm->dev) == IP_VERSION(9, 5, 0))
+			pm->pmf = &kfd_aldebaran_pm_funcs;
+		else if (KFD_GC_VERSION(dqm->dev) >= IP_VERSION(9, 0, 1))
+			pm->pmf = &kfd_v9_pm_funcs;
+		else {
+			WARN(1, "Unexpected ASIC family %u",
+			     dqm->dev->adev->asic_type);
+			return -EINVAL;
+		}
 	}
 
 	pm->dqm = dqm;
@@ -275,16 +318,18 @@ int pm_init(struct packet_manager *pm, struct device_queue_manager *dqm)
 	return 0;
 }
 
-void pm_uninit(struct packet_manager *pm, bool hanging)
+void pm_uninit(struct packet_manager *pm)
 {
 	mutex_destroy(&pm->lock);
-	kernel_queue_uninit(pm->priv_queue, hanging);
+	kernel_queue_uninit(pm->priv_queue);
 	pm->priv_queue = NULL;
 }
 
 int pm_send_set_resources(struct packet_manager *pm,
 				struct scheduling_resources *res)
 {
+	struct kfd_node *node = pm->dqm->dev;
+	struct device *dev = node->adev->dev;
 	uint32_t *buffer, size;
 	int retval = 0;
 
@@ -294,14 +339,14 @@ int pm_send_set_resources(struct packet_manager *pm,
 					size / sizeof(uint32_t),
 					(unsigned int **)&buffer);
 	if (!buffer) {
-		pr_err("Failed to allocate buffer on kernel queue\n");
+		dev_err(dev, "Failed to allocate buffer on kernel queue\n");
 		retval = -ENOMEM;
 		goto out;
 	}
 
 	retval = pm->pmf->set_resources(pm, buffer, res);
 	if (!retval)
-		kq_submit_packet(pm->priv_queue);
+		retval = kq_submit_packet(pm->priv_queue);
 	else
 		kq_rollback_packet(pm->priv_queue);
 
@@ -338,7 +383,7 @@ int pm_send_runlist(struct packet_manager *pm, struct list_head *dqm_queues)
 	if (retval)
 		goto fail_create_runlist;
 
-	kq_submit_packet(pm->priv_queue);
+	retval = kq_submit_packet(pm->priv_queue);
 
 	mutex_unlock(&pm->lock);
 
@@ -356,6 +401,8 @@ fail_create_runlist_ib:
 int pm_send_query_status(struct packet_manager *pm, uint64_t fence_address,
 			uint64_t fence_value)
 {
+	struct kfd_node *node = pm->dqm->dev;
+	struct device *dev = node->adev->dev;
 	uint32_t *buffer, size;
 	int retval = 0;
 
@@ -367,14 +414,14 @@ int pm_send_query_status(struct packet_manager *pm, uint64_t fence_address,
 	kq_acquire_packet_buffer(pm->priv_queue,
 			size / sizeof(uint32_t), (unsigned int **)&buffer);
 	if (!buffer) {
-		pr_err("Failed to allocate buffer on kernel queue\n");
+		dev_err(dev, "Failed to allocate buffer on kernel queue\n");
 		retval = -ENOMEM;
 		goto out;
 	}
 
 	retval = pm->pmf->query_status(pm, buffer, fence_address, fence_value);
 	if (!retval)
-		kq_submit_packet(pm->priv_queue);
+		retval = kq_submit_packet(pm->priv_queue);
 	else
 		kq_rollback_packet(pm->priv_queue);
 
@@ -383,11 +430,71 @@ out:
 	return retval;
 }
 
-int pm_send_unmap_queue(struct packet_manager *pm, enum kfd_queue_type type,
+/* pm_config_dequeue_wait_counts: Configure dequeue timer Wait Counts
+ *  by writing to CP_IQ_WAIT_TIME2 registers.
+ *
+ *  @cmd: See emum kfd_config_dequeue_wait_counts_cmd definition
+ *  @value: Depends on the cmd. This parameter is unused for
+ *    KFD_DEQUEUE_WAIT_INIT and KFD_DEQUEUE_WAIT_RESET. For
+ *    KFD_DEQUEUE_WAIT_SET_SCH_WAVE it holds value to be set
+ *
+ */
+int pm_config_dequeue_wait_counts(struct packet_manager *pm,
+		enum kfd_config_dequeue_wait_counts_cmd cmd,
+		uint32_t value)
+{
+	struct kfd_node *node = pm->dqm->dev;
+	struct device *dev = node->adev->dev;
+	int retval = 0;
+	uint32_t *buffer, size;
+
+	if (!pm->pmf->config_dequeue_wait_counts ||
+	    !pm->pmf->config_dequeue_wait_counts_size)
+		return 0;
+
+	if (cmd == KFD_DEQUEUE_WAIT_INIT && (KFD_GC_VERSION(pm->dqm->dev) < IP_VERSION(9, 4, 1) ||
+	   KFD_GC_VERSION(pm->dqm->dev) >= IP_VERSION(10, 0, 0)))
+		return 0;
+
+	size = pm->pmf->config_dequeue_wait_counts_size;
+
+	mutex_lock(&pm->lock);
+
+	if (size) {
+		kq_acquire_packet_buffer(pm->priv_queue,
+			size / sizeof(uint32_t),
+			(unsigned int **)&buffer);
+
+		if (!buffer) {
+			dev_err(dev,
+				"Failed to allocate buffer on kernel queue\n");
+			retval = -ENOMEM;
+			goto out;
+		}
+
+		retval = pm->pmf->config_dequeue_wait_counts(pm, buffer,
+							     cmd, value);
+		if (!retval) {
+			retval = kq_submit_packet(pm->priv_queue);
+
+			/* If default value is modified, cache that in dqm->wait_times */
+			if (!retval && cmd == KFD_DEQUEUE_WAIT_INIT)
+				update_dqm_wait_times(pm->dqm);
+		} else {
+			kq_rollback_packet(pm->priv_queue);
+		}
+	}
+out:
+	mutex_unlock(&pm->lock);
+	return retval;
+}
+
+int pm_send_unmap_queue(struct packet_manager *pm,
 			enum kfd_unmap_queues_filter filter,
-			uint32_t filter_param, bool reset,
-			unsigned int sdma_engine)
+			uint32_t filter_param, bool reset)
 {
+	struct kfd_node *node = pm->dqm->dev;
+	struct device *dev = node->adev->dev;
 	uint32_t *buffer, size;
 	int retval = 0;
 
@@ -396,15 +503,14 @@ int pm_send_unmap_queue(struct packet_manager *pm, enum kfd_queue_type type,
 	kq_acquire_packet_buffer(pm->priv_queue,
 			size / sizeof(uint32_t), (unsigned int **)&buffer);
 	if (!buffer) {
-		pr_err("Failed to allocate buffer on kernel queue\n");
+		dev_err(dev, "Failed to allocate buffer on kernel queue\n");
 		retval = -ENOMEM;
 		goto out;
 	}
 
-	retval = pm->pmf->unmap_queues(pm, buffer, type, filter, filter_param,
-				       reset, sdma_engine);
+	retval = pm->pmf->unmap_queues(pm, buffer, filter, filter_param, reset);
 	if (!retval)
-		kq_submit_packet(pm->priv_queue);
+		retval = kq_submit_packet(pm->priv_queue);
 	else
 		kq_rollback_packet(pm->priv_queue);
 
@@ -446,6 +552,8 @@ out:
 
 int pm_debugfs_hang_hws(struct packet_manager *pm)
 {
+	struct kfd_node *node = pm->dqm->dev;
+	struct device *dev = node->adev->dev;
 	uint32_t *buffer, size;
 	int r = 0;
 
@@ -457,16 +565,16 @@ int pm_debugfs_hang_hws(struct packet_manager *pm)
 	kq_acquire_packet_buffer(pm->priv_queue,
 			size / sizeof(uint32_t), (unsigned int **)&buffer);
 	if (!buffer) {
-		pr_err("Failed to allocate buffer on kernel queue\n");
+		dev_err(dev, "Failed to allocate buffer on kernel queue\n");
 		r = -ENOMEM;
 		goto out;
 	}
 	memset(buffer, 0x55, size);
 	kq_submit_packet(pm->priv_queue);
 
-	pr_info("Submitting %x %x %x %x %x %x %x to HIQ to hang the HWS.",
-		buffer[0], buffer[1], buffer[2], buffer[3],
-		buffer[4], buffer[5], buffer[6]);
+	dev_info(dev, "Submitting %x %x %x %x %x %x %x to HIQ to hang the HWS.",
+		 buffer[0], buffer[1], buffer[2], buffer[3], buffer[4],
+		 buffer[5], buffer[6]);
 out:
 	mutex_unlock(&pm->lock);
 	return r;
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager_v9.c b/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager_v9.c
index 7ea3f671b325..505036968a77 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager_v9.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager_v9.c
@@ -1,5 +1,6 @@
+// SPDX-License-Identifier: GPL-2.0 OR MIT
 /*
- * Copyright 2016-2018 Advanced Micro Devices, Inc.
+ * Copyright 2016-2022 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -33,14 +34,20 @@ static int pm_map_process_v9(struct packet_manager *pm,
 {
 	struct pm4_mes_map_process *packet;
 	uint64_t vm_page_table_base_addr = qpd->page_table_base;
+	struct kfd_node *kfd = pm->dqm->dev;
+	struct kfd_process_device *pdd =
+			container_of(qpd, struct kfd_process_device, qpd);
+	struct amdgpu_device *adev = kfd->adev;
 
 	packet = (struct pm4_mes_map_process *)buffer;
 	memset(buffer, 0, sizeof(struct pm4_mes_map_process));
 	packet->header.u32All = pm_build_pm4_header(IT_MAP_PROCESS,
 					sizeof(struct pm4_mes_map_process));
+	if (adev->enforce_isolation[kfd->node_id] == AMDGPU_ENFORCE_ISOLATION_ENABLE)
+		packet->bitfields2.exec_cleaner_shader = 1;
 	packet->bitfields2.diq_enable = (qpd->is_debug) ? 1 : 0;
 	packet->bitfields2.process_quantum = 10;
-	packet->bitfields2.pasid = qpd->pqm->process->pasid;
+	packet->bitfields2.pasid = pdd->pasid;
 	packet->bitfields14.gds_size = qpd->gds_size & 0x3F;
 	packet->bitfields14.gds_size_hi = (qpd->gds_size >> 6) & 0xF;
 	packet->bitfields14.num_gws = (qpd->mapped_gws_queue) ? qpd->num_gws : 0;
@@ -48,6 +55,12 @@ static int pm_map_process_v9(struct packet_manager *pm,
 	packet->bitfields14.sdma_enable = 1;
 	packet->bitfields14.num_queues = (qpd->is_debug) ? 0 : qpd->queue_count;
 
+	if (kfd->dqm->trap_debug_vmid && pdd->process->debug_trap_enabled &&
+			pdd->process->runtime_info.runtime_state == DEBUG_RUNTIME_STATE_ENABLED) {
+		packet->bitfields2.debug_vmid = kfd->dqm->trap_debug_vmid;
+		packet->bitfields2.new_debug = 1;
+	}
+
 	packet->sh_mem_config = qpd->sh_mem_config;
 	packet->sh_mem_bases = qpd->sh_mem_bases;
 	if (qpd->tba_addr) {
@@ -78,25 +91,45 @@ static int pm_map_process_aldebaran(struct packet_manager *pm,
 {
 	struct pm4_mes_map_process_aldebaran *packet;
 	uint64_t vm_page_table_base_addr = qpd->page_table_base;
+	struct kfd_dev *kfd = pm->dqm->dev->kfd;
+	struct kfd_node *knode = pm->dqm->dev;
+	struct kfd_process_device *pdd =
+			container_of(qpd, struct kfd_process_device, qpd);
+	int i;
+	struct amdgpu_device *adev = kfd->adev;
 
 	packet = (struct pm4_mes_map_process_aldebaran *)buffer;
 	memset(buffer, 0, sizeof(struct pm4_mes_map_process_aldebaran));
 	packet->header.u32All = pm_build_pm4_header(IT_MAP_PROCESS,
 			sizeof(struct pm4_mes_map_process_aldebaran));
+	if (adev->enforce_isolation[knode->node_id] ==
+	    AMDGPU_ENFORCE_ISOLATION_ENABLE)
+		packet->bitfields2.exec_cleaner_shader = 1;
 	packet->bitfields2.diq_enable = (qpd->is_debug) ? 1 : 0;
 	packet->bitfields2.process_quantum = 10;
-	packet->bitfields2.pasid = qpd->pqm->process->pasid;
+	packet->bitfields2.pasid = pdd->pasid;
 	packet->bitfields14.gds_size = qpd->gds_size & 0x3F;
 	packet->bitfields14.gds_size_hi = (qpd->gds_size >> 6) & 0xF;
 	packet->bitfields14.num_gws = (qpd->mapped_gws_queue) ? qpd->num_gws : 0;
 	packet->bitfields14.num_oac = qpd->num_oac;
 	packet->bitfields14.sdma_enable = 1;
 	packet->bitfields14.num_queues = (qpd->is_debug) ? 0 : qpd->queue_count;
+	packet->spi_gdbg_per_vmid_cntl = pdd->spi_dbg_override |
+						pdd->spi_dbg_launch_mode;
+
+	if (pdd->process->debug_trap_enabled) {
+		for (i = 0; i < kfd->device_info.num_of_watch_points; i++)
+			packet->tcp_watch_cntl[i] = pdd->watch_points[i];
+
+		packet->bitfields2.single_memops =
+				!!(pdd->process->dbg_flags & KFD_DBG_TRAP_FLAG_SINGLE_MEM_OP);
+	}
 
 	packet->sh_mem_config = qpd->sh_mem_config;
 	packet->sh_mem_bases = qpd->sh_mem_bases;
 	if (qpd->tba_addr) {
 		packet->sq_shader_tba_lo = lower_32_bits(qpd->tba_addr >> 8);
+		packet->sq_shader_tba_hi = upper_32_bits(qpd->tba_addr >> 8);
 		packet->sq_shader_tma_lo = lower_32_bits(qpd->tma_addr >> 8);
 		packet->sq_shader_tma_hi = upper_32_bits(qpd->tma_addr >> 8);
 	}
@@ -118,19 +151,24 @@ static int pm_runlist_v9(struct packet_manager *pm, uint32_t *buffer,
 	struct pm4_mes_runlist *packet;
 
 	int concurrent_proc_cnt = 0;
-	struct kfd_dev *kfd = pm->dqm->dev;
+	struct kfd_node *kfd = pm->dqm->dev;
+	struct amdgpu_device *adev = kfd->adev;
 
 	/* Determine the number of processes to map together to HW:
 	 * it can not exceed the number of VMIDs available to the
 	 * scheduler, and it is determined by the smaller of the number
 	 * of processes in the runlist and kfd module parameter
 	 * hws_max_conc_proc.
+	 * However, if enforce_isolation is set (toggle LDS/VGPRs/SGPRs
+	 * cleaner between process switch), enable single-process mode
+	 * in HWS.
 	 * Note: the arbitration between the number of VMIDs and
 	 * hws_max_conc_proc has been done in
 	 * kgd2kfd_device_init().
 	 */
-	concurrent_proc_cnt = min(pm->dqm->processes_count,
-			kfd->max_proc_per_quantum);
+	concurrent_proc_cnt = (adev->enforce_isolation[kfd->node_id] ==
+			       AMDGPU_ENFORCE_ISOLATION_ENABLE) ?
+		1 : min(pm->dqm->processes_count, kfd->max_proc_per_quantum);
 
 	packet = (struct pm4_mes_runlist *)buffer;
 
@@ -165,6 +203,8 @@ static int pm_set_resources_v9(struct packet_manager *pm, uint32_t *buffer,
 			queue_type__mes_set_resources__hsa_interface_queue_hiq;
 	packet->bitfields2.vmid_mask = res->vmid_mask;
 	packet->bitfields2.unmap_latency = KFD_UNMAP_LATENCY_MS / 100;
+	if (pm->dqm->dev->adev->gmc.xnack_flags & AMDGPU_GMC_XNACK_FLAG_CHAIN)
+		packet->bitfields2.enb_xnack_retry_disable_check = 1;
 	packet->bitfields7.oac_mask = res->oac_mask;
 	packet->bitfields8.gds_heap_base = res->gds_heap_base;
 	packet->bitfields8.gds_heap_size = res->gds_heap_size;
@@ -178,11 +218,16 @@ static int pm_set_resources_v9(struct packet_manager *pm, uint32_t *buffer,
 	return 0;
 }
 
+static inline bool pm_use_ext_eng(struct kfd_dev *dev)
+{
+	return amdgpu_ip_version(dev->adev, SDMA0_HWIP, 0) >=
+	       IP_VERSION(5, 2, 0);
+}
+
 static int pm_map_queues_v9(struct packet_manager *pm, uint32_t *buffer,
 		struct queue *q, bool is_static)
 {
 	struct pm4_mes_map_queues *packet;
-	bool use_static = is_static;
 
 	packet = (struct pm4_mes_map_queues *)buffer;
 	memset(buffer, 0, sizeof(struct pm4_mes_map_queues));
@@ -195,7 +240,7 @@ static int pm_map_queues_v9(struct packet_manager *pm, uint32_t *buffer,
 
 	packet->bitfields2.engine_sel =
 		engine_sel__mes_map_queues__compute_vi;
-	packet->bitfields2.gws_control_queue = q->gws ? 1 : 0;
+	packet->bitfields2.gws_control_queue = q->properties.is_gws ? 1 : 0;
 	packet->bitfields2.extended_engine_sel =
 		extended_engine_sel__mes_map_queues__legacy_engine_sel;
 	packet->bitfields2.queue_type =
@@ -203,7 +248,7 @@ static int pm_map_queues_v9(struct packet_manager *pm, uint32_t *buffer,
 
 	switch (q->properties.type) {
 	case KFD_QUEUE_TYPE_COMPUTE:
-		if (use_static)
+		if (is_static)
 			packet->bitfields2.queue_type =
 		queue_type__mes_map_queues__normal_latency_static_queue_vi;
 		break;
@@ -213,14 +258,24 @@ static int pm_map_queues_v9(struct packet_manager *pm, uint32_t *buffer,
 		break;
 	case KFD_QUEUE_TYPE_SDMA:
 	case KFD_QUEUE_TYPE_SDMA_XGMI:
-		use_static = false; /* no static queues under SDMA */
-		if (q->properties.sdma_engine_id < 2)
+		if (q->properties.sdma_engine_id < 2 &&
+		    !pm_use_ext_eng(q->device->kfd))
 			packet->bitfields2.engine_sel = q->properties.sdma_engine_id +
 				engine_sel__mes_map_queues__sdma0_vi;
 		else {
-			packet->bitfields2.extended_engine_sel =
-				extended_engine_sel__mes_map_queues__sdma0_to_7_sel;
-			packet->bitfields2.engine_sel = q->properties.sdma_engine_id;
+			/*
+			 * For GFX9.4.3, SDMA engine id can be greater than 8.
+			 * For such cases, set extended_engine_sel to 2 and
+			 * ensure engine_sel lies between 0-7.
+			 */
+			if (q->properties.sdma_engine_id >= 8)
+				packet->bitfields2.extended_engine_sel =
+					extended_engine_sel__mes_map_queues__sdma8_to_15_sel;
+			else
+				packet->bitfields2.extended_engine_sel =
+					extended_engine_sel__mes_map_queues__sdma0_to_7_sel;
+
+			packet->bitfields2.engine_sel = q->properties.sdma_engine_id % 8;
 		}
 		break;
 	default:
@@ -245,11 +300,100 @@ static int pm_map_queues_v9(struct packet_manager *pm, uint32_t *buffer,
 	return 0;
 }
 
+static inline void pm_build_dequeue_wait_counts_packet_info(struct packet_manager *pm,
+			uint32_t sch_value, uint32_t que_sleep, uint32_t *reg_offset,
+			uint32_t *reg_data)
+{
+	pm->dqm->dev->kfd2kgd->build_dequeue_wait_counts_packet_info(
+		pm->dqm->dev->adev,
+		pm->dqm->wait_times,
+		sch_value,
+		que_sleep,
+		reg_offset,
+		reg_data);
+}
+
+/* pm_config_dequeue_wait_counts_v9: Builds WRITE_DATA packet with
+ *    register/value for configuring dequeue wait counts
+ *
+ * @return: -ve for failure and 0 for success and buffer is
+ *  filled in with packet
+ *
+ **/
+static int pm_config_dequeue_wait_counts_v9(struct packet_manager *pm,
+		uint32_t *buffer,
+		enum kfd_config_dequeue_wait_counts_cmd cmd,
+		uint32_t value)
+{
+	struct pm4_mec_write_data_mmio *packet;
+	uint32_t reg_offset = 0;
+	uint32_t reg_data = 0;
+
+	switch (cmd) {
+	case KFD_DEQUEUE_WAIT_INIT: {
+		uint32_t sch_wave = 0, que_sleep = 1;
+
+		/* For all gfx9 ASICs > gfx941,
+		 * Reduce CP_IQ_WAIT_TIME2.QUE_SLEEP to 0x1 from default 0x40.
+		 * On a 1GHz machine this is roughly 1 microsecond, which is
+		 * about how long it takes to load data out of memory during
+		 * queue connect
+		 * QUE_SLEEP: Wait Count for Dequeue Retry.
+		 *
+		 * Set CWSR grace period to 1x1000 cycle for GFX9.4.3 APU
+		 */
+		if (KFD_GC_VERSION(pm->dqm->dev) < IP_VERSION(9, 4, 1) ||
+		    KFD_GC_VERSION(pm->dqm->dev) >= IP_VERSION(10, 0, 0))
+			return -EPERM;
+
+		if (amdgpu_emu_mode == 0 && pm->dqm->dev->adev->gmc.is_app_apu &&
+		    (KFD_GC_VERSION(pm->dqm->dev) == IP_VERSION(9, 4, 3)))
+			sch_wave = 1;
+
+		pm_build_dequeue_wait_counts_packet_info(pm, sch_wave, que_sleep,
+			&reg_offset, &reg_data);
+
+		break;
+	}
+	case KFD_DEQUEUE_WAIT_RESET:
+		/* reg_data would be set to dqm->wait_times */
+		pm_build_dequeue_wait_counts_packet_info(pm, 0, 0, &reg_offset, &reg_data);
+		break;
+
+	case KFD_DEQUEUE_WAIT_SET_SCH_WAVE:
+		/* The CP cannot handle value 0 and it will result in
+		 * an infinite grace period being set so set to 1 to prevent this. Also
+		 * avoid debugger API breakage as it sets 0 and expects a low value.
+		 */
+		if (!value)
+			value = 1;
+		pm_build_dequeue_wait_counts_packet_info(pm, value, 0, &reg_offset, &reg_data);
+		break;
+	default:
+		pr_err("Invalid dequeue wait cmd\n");
+		return -EINVAL;
+	}
+
+	packet = (struct pm4_mec_write_data_mmio *)buffer;
+	memset(buffer, 0, sizeof(struct pm4_mec_write_data_mmio));
+
+	packet->header.u32All = pm_build_pm4_header(IT_WRITE_DATA,
+					sizeof(struct pm4_mec_write_data_mmio));
+
+	packet->bitfields2.dst_sel  = dst_sel___write_data__mem_mapped_register;
+	packet->bitfields2.addr_incr =
+			addr_incr___write_data__do_not_increment_address;
+
+	packet->bitfields3.dst_mmreg_addr = reg_offset;
+
+	packet->data = reg_data;
+
+	return 0;
+}
+
 static int pm_unmap_queues_v9(struct packet_manager *pm, uint32_t *buffer,
-			enum kfd_queue_type type,
 			enum kfd_unmap_queues_filter filter,
-			uint32_t filter_param, bool reset,
-			unsigned int sdma_engine)
+			uint32_t filter_param, bool reset)
 {
 	struct pm4_mes_unmap_queues *packet;
 
@@ -258,31 +402,14 @@ static int pm_unmap_queues_v9(struct packet_manager *pm, uint32_t *buffer,
 
 	packet->header.u32All = pm_build_pm4_header(IT_UNMAP_QUEUES,
 					sizeof(struct pm4_mes_unmap_queues));
-	switch (type) {
-	case KFD_QUEUE_TYPE_COMPUTE:
-	case KFD_QUEUE_TYPE_DIQ:
-		packet->bitfields2.extended_engine_sel =
-			extended_engine_sel__mes_unmap_queues__legacy_engine_sel;
-		packet->bitfields2.engine_sel =
-			engine_sel__mes_unmap_queues__compute;
-		break;
-	case KFD_QUEUE_TYPE_SDMA:
-	case KFD_QUEUE_TYPE_SDMA_XGMI:
-		if (sdma_engine < 2) {
-			packet->bitfields2.extended_engine_sel =
-				extended_engine_sel__mes_unmap_queues__legacy_engine_sel;
-			packet->bitfields2.engine_sel =
-				engine_sel__mes_unmap_queues__sdma0 + sdma_engine;
-		} else {
-			packet->bitfields2.extended_engine_sel =
-				extended_engine_sel__mes_unmap_queues__sdma0_to_7_sel;
-			packet->bitfields2.engine_sel = sdma_engine;
-		}
-		break;
-	default:
-		WARN(1, "queue type %d", type);
-		return -EINVAL;
-	}
+
+	packet->bitfields2.extended_engine_sel =
+				pm_use_ext_eng(pm->dqm->dev->kfd) ?
+		extended_engine_sel__mes_unmap_queues__sdma0_to_7_sel :
+		extended_engine_sel__mes_unmap_queues__legacy_engine_sel;
+
+	packet->bitfields2.engine_sel =
+		engine_sel__mes_unmap_queues__compute;
 
 	if (reset)
 		packet->bitfields2.action =
@@ -292,12 +419,6 @@ static int pm_unmap_queues_v9(struct packet_manager *pm, uint32_t *buffer,
 			action__mes_unmap_queues__preempt_queues;
 
 	switch (filter) {
-	case KFD_UNMAP_QUEUES_FILTER_SINGLE_QUEUE:
-		packet->bitfields2.queue_sel =
-			queue_sel__mes_unmap_queues__perform_request_on_specified_queues;
-		packet->bitfields2.num_queues = 1;
-		packet->bitfields3b.doorbell_offset0 = filter_param;
-		break;
 	case KFD_UNMAP_QUEUES_FILTER_BY_PASID:
 		packet->bitfields2.queue_sel =
 			queue_sel__mes_unmap_queues__perform_request_on_pasid_queues;
@@ -353,6 +474,7 @@ const struct packet_manager_funcs kfd_v9_pm_funcs = {
 	.set_resources		= pm_set_resources_v9,
 	.map_queues		= pm_map_queues_v9,
 	.unmap_queues		= pm_unmap_queues_v9,
+	.config_dequeue_wait_counts = pm_config_dequeue_wait_counts_v9,
 	.query_status		= pm_query_status_v9,
 	.release_mem		= NULL,
 	.map_process_size	= sizeof(struct pm4_mes_map_process),
@@ -360,6 +482,7 @@ const struct packet_manager_funcs kfd_v9_pm_funcs = {
 	.set_resources_size	= sizeof(struct pm4_mes_set_resources),
 	.map_queues_size	= sizeof(struct pm4_mes_map_queues),
 	.unmap_queues_size	= sizeof(struct pm4_mes_unmap_queues),
+	.config_dequeue_wait_counts_size  = sizeof(struct pm4_mec_write_data_mmio),
 	.query_status_size	= sizeof(struct pm4_mes_query_status),
 	.release_mem_size	= 0,
 };
@@ -370,6 +493,7 @@ const struct packet_manager_funcs kfd_aldebaran_pm_funcs = {
 	.set_resources		= pm_set_resources_v9,
 	.map_queues		= pm_map_queues_v9,
 	.unmap_queues		= pm_unmap_queues_v9,
+	.config_dequeue_wait_counts = pm_config_dequeue_wait_counts_v9,
 	.query_status		= pm_query_status_v9,
 	.release_mem		= NULL,
 	.map_process_size	= sizeof(struct pm4_mes_map_process_aldebaran),
@@ -377,6 +501,7 @@ const struct packet_manager_funcs kfd_aldebaran_pm_funcs = {
 	.set_resources_size	= sizeof(struct pm4_mes_set_resources),
 	.map_queues_size	= sizeof(struct pm4_mes_map_queues),
 	.unmap_queues_size	= sizeof(struct pm4_mes_unmap_queues),
+	.config_dequeue_wait_counts_size  = sizeof(struct pm4_mec_write_data_mmio),
 	.query_status_size	= sizeof(struct pm4_mes_query_status),
 	.release_mem_size	= 0,
 };
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager_vi.c b/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager_vi.c
index 08442e7d9944..a1de5d7e173a 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager_vi.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager_vi.c
@@ -1,5 +1,6 @@
+// SPDX-License-Identifier: GPL-2.0 OR MIT
 /*
- * Copyright 2014 Advanced Micro Devices, Inc.
+ * Copyright 2014-2022 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -41,6 +42,7 @@ unsigned int pm_build_pm4_header(unsigned int opcode, size_t packet_size)
 static int pm_map_process_vi(struct packet_manager *pm, uint32_t *buffer,
 				struct qcm_process_device *qpd)
 {
+	struct kfd_process_device *pdd = qpd_to_pdd(qpd);
 	struct pm4_mes_map_process *packet;
 
 	packet = (struct pm4_mes_map_process *)buffer;
@@ -51,7 +53,7 @@ static int pm_map_process_vi(struct packet_manager *pm, uint32_t *buffer,
 					sizeof(struct pm4_mes_map_process));
 	packet->bitfields2.diq_enable = (qpd->is_debug) ? 1 : 0;
 	packet->bitfields2.process_quantum = 10;
-	packet->bitfields2.pasid = qpd->pqm->process->pasid;
+	packet->bitfields2.pasid = pdd->pasid;
 	packet->bitfields3.page_table_base = qpd->page_table_base;
 	packet->bitfields10.gds_size = qpd->gds_size;
 	packet->bitfields10.num_gws = qpd->num_gws;
@@ -76,7 +78,7 @@ static int pm_runlist_vi(struct packet_manager *pm, uint32_t *buffer,
 {
 	struct pm4_mes_runlist *packet;
 	int concurrent_proc_cnt = 0;
-	struct kfd_dev *kfd = pm->dqm->dev;
+	struct kfd_node *kfd = pm->dqm->dev;
 
 	if (WARN_ON(!ib))
 		return -EFAULT;
@@ -110,8 +112,8 @@ static int pm_runlist_vi(struct packet_manager *pm, uint32_t *buffer,
 	return 0;
 }
 
-int pm_set_resources_vi(struct packet_manager *pm, uint32_t *buffer,
-				struct scheduling_resources *res)
+static int pm_set_resources_vi(struct packet_manager *pm, uint32_t *buffer,
+			       struct scheduling_resources *res)
 {
 	struct pm4_mes_set_resources *packet;
 
@@ -197,10 +199,8 @@ static int pm_map_queues_vi(struct packet_manager *pm, uint32_t *buffer,
 }
 
 static int pm_unmap_queues_vi(struct packet_manager *pm, uint32_t *buffer,
-			enum kfd_queue_type type,
 			enum kfd_unmap_queues_filter filter,
-			uint32_t filter_param, bool reset,
-			unsigned int sdma_engine)
+			uint32_t filter_param, bool reset)
 {
 	struct pm4_mes_unmap_queues *packet;
 
@@ -209,21 +209,9 @@ static int pm_unmap_queues_vi(struct packet_manager *pm, uint32_t *buffer,
 
 	packet->header.u32All = pm_build_pm4_header(IT_UNMAP_QUEUES,
 					sizeof(struct pm4_mes_unmap_queues));
-	switch (type) {
-	case KFD_QUEUE_TYPE_COMPUTE:
-	case KFD_QUEUE_TYPE_DIQ:
-		packet->bitfields2.engine_sel =
+
+	packet->bitfields2.engine_sel =
 			engine_sel__mes_unmap_queues__compute;
-		break;
-	case KFD_QUEUE_TYPE_SDMA:
-	case KFD_QUEUE_TYPE_SDMA_XGMI:
-		packet->bitfields2.engine_sel =
-			engine_sel__mes_unmap_queues__sdma0 + sdma_engine;
-		break;
-	default:
-		WARN(1, "queue type %d", type);
-		return -EINVAL;
-	}
 
 	if (reset)
 		packet->bitfields2.action =
@@ -233,12 +221,6 @@ static int pm_unmap_queues_vi(struct packet_manager *pm, uint32_t *buffer,
 			action__mes_unmap_queues__preempt_queues;
 
 	switch (filter) {
-	case KFD_UNMAP_QUEUES_FILTER_SINGLE_QUEUE:
-		packet->bitfields2.queue_sel =
-			queue_sel__mes_unmap_queues__perform_request_on_specified_queues;
-		packet->bitfields2.num_queues = 1;
-		packet->bitfields3b.doorbell_offset0 = filter_param;
-		break;
 	case KFD_UNMAP_QUEUES_FILTER_BY_PASID:
 		packet->bitfields2.queue_sel =
 			queue_sel__mes_unmap_queues__perform_request_on_pasid_queues;
@@ -322,6 +304,7 @@ const struct packet_manager_funcs kfd_vi_pm_funcs = {
 	.set_resources		= pm_set_resources_vi,
 	.map_queues		= pm_map_queues_vi,
 	.unmap_queues		= pm_unmap_queues_vi,
+	.config_dequeue_wait_counts	= NULL,
 	.query_status		= pm_query_status_vi,
 	.release_mem		= pm_release_mem_vi,
 	.map_process_size	= sizeof(struct pm4_mes_map_process),
@@ -329,6 +312,7 @@ const struct packet_manager_funcs kfd_vi_pm_funcs = {
 	.set_resources_size	= sizeof(struct pm4_mes_set_resources),
 	.map_queues_size	= sizeof(struct pm4_mes_map_queues),
 	.unmap_queues_size	= sizeof(struct pm4_mes_unmap_queues),
+	.config_dequeue_wait_counts_size	= 0,
 	.query_status_size	= sizeof(struct pm4_mes_query_status),
 	.release_mem_size	= sizeof(struct pm4_mec_release_mem)
 };
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_pasid.c b/drivers/gpu/drm/amd/amdkfd/kfd_pasid.c
deleted file mode 100644
index af5816f51e55..000000000000
--- a/drivers/gpu/drm/amd/amdkfd/kfd_pasid.c
+++ /dev/null
@@ -1,69 +0,0 @@
-/*
- * Copyright 2014 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- */
-
-#include <linux/types.h>
-#include "kfd_priv.h"
-#include "amdgpu_ids.h"
-
-static unsigned int pasid_bits = 16;
-static bool pasids_allocated; /* = false */
-
-bool kfd_set_pasid_limit(unsigned int new_limit)
-{
-	if (new_limit < 2)
-		return false;
-
-	if (new_limit < (1U << pasid_bits)) {
-		if (pasids_allocated)
-			/* We've already allocated user PASIDs, too late to
-			 * change the limit
-			 */
-			return false;
-
-		while (new_limit < (1U << pasid_bits))
-			pasid_bits--;
-	}
-
-	return true;
-}
-
-unsigned int kfd_get_pasid_limit(void)
-{
-	return 1U << pasid_bits;
-}
-
-u32 kfd_pasid_alloc(void)
-{
-	int r = amdgpu_pasid_alloc(pasid_bits);
-
-	if (r > 0) {
-		pasids_allocated = true;
-		return r;
-	}
-
-	return 0;
-}
-
-void kfd_pasid_free(u32 pasid)
-{
-	amdgpu_pasid_free(pasid);
-}
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_pm4_headers.h b/drivers/gpu/drm/amd/amdkfd/kfd_pm4_headers.h
index e50f73d25de6..7274edfd3f38 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_pm4_headers.h
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_pm4_headers.h
@@ -1,5 +1,6 @@
+/* SPDX-License-Identifier: GPL-2.0 OR MIT */
 /*
- * Copyright 2014 Advanced Micro Devices, Inc.
+ * Copyright 2014-2022 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_pm4_headers_ai.h b/drivers/gpu/drm/amd/amdkfd/kfd_pm4_headers_ai.h
index 4d7add843746..e356a207d03c 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_pm4_headers_ai.h
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_pm4_headers_ai.h
@@ -1,5 +1,6 @@
+/* SPDX-License-Identifier: GPL-2.0 OR MIT */
 /*
- * Copyright 2016 Advanced Micro Devices, Inc.
+ * Copyright 2016-2022 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -62,7 +63,8 @@ struct pm4_mes_set_resources {
 		struct {
 			uint32_t vmid_mask:16;
 			uint32_t unmap_latency:8;
-			uint32_t reserved1:5;
+			uint32_t reserved1:4;
+			uint32_t enb_xnack_retry_disable_check:1;
 			enum mes_set_resources_queue_type_enum queue_type:3;
 		} bitfields2;
 		uint32_t ordinal2;
@@ -144,8 +146,12 @@ struct pm4_mes_map_process {
 
 	union {
 		struct {
-			uint32_t pasid:16;
-			uint32_t reserved1:8;
+			uint32_t pasid:16;		/* 0 - 15  */
+			uint32_t reserved1:1;		/* 16      */
+			uint32_t exec_cleaner_shader:1;	/* 17      */
+			uint32_t debug_vmid:4;
+			uint32_t new_debug:1;
+			uint32_t reserved2:1;
 			uint32_t diq_enable:1;
 			uint32_t process_quantum:7;
 		} bitfields2;
@@ -262,7 +268,8 @@ enum mes_map_queues_engine_sel_enum {
 
 enum mes_map_queues_extended_engine_sel_enum {
 	extended_engine_sel__mes_map_queues__legacy_engine_sel = 0,
-	extended_engine_sel__mes_map_queues__sdma0_to_7_sel = 1
+	extended_engine_sel__mes_map_queues__sdma0_to_7_sel  = 1,
+	extended_engine_sel__mes_map_queues__sdma8_to_15_sel = 2
 };
 
 struct pm4_mes_map_queues {
@@ -582,6 +589,71 @@ struct pm4_mec_release_mem {
 
 #endif
 
+#ifndef PM4_MEC_WRITE_DATA_DEFINED
+#define PM4_MEC_WRITE_DATA_DEFINED
+
+enum WRITE_DATA_dst_sel_enum {
+	dst_sel___write_data__mem_mapped_register = 0,
+	dst_sel___write_data__tc_l2 = 2,
+	dst_sel___write_data__gds = 3,
+	dst_sel___write_data__memory = 5,
+	dst_sel___write_data__memory_mapped_adc_persistent_state = 6,
+};
+
+enum WRITE_DATA_addr_incr_enum {
+	addr_incr___write_data__increment_address = 0,
+	addr_incr___write_data__do_not_increment_address = 1
+};
+
+enum WRITE_DATA_wr_confirm_enum {
+	wr_confirm___write_data__do_not_wait_for_write_confirmation = 0,
+	wr_confirm___write_data__wait_for_write_confirmation = 1
+};
+
+enum WRITE_DATA_cache_policy_enum {
+	cache_policy___write_data__lru = 0,
+	cache_policy___write_data__stream = 1
+};
+
+
+struct pm4_mec_write_data_mmio {
+	union {
+		union PM4_MES_TYPE_3_HEADER header;     /*header */
+		unsigned int ordinal1;
+	};
+
+	union {
+		struct {
+			unsigned int reserved1:8;
+			unsigned int dst_sel:4;
+			unsigned int reserved2:4;
+			unsigned int addr_incr:1;
+			unsigned int reserved3:2;
+			unsigned int resume_vf:1;
+			unsigned int wr_confirm:1;
+			unsigned int reserved4:4;
+			unsigned int cache_policy:2;
+			unsigned int reserved5:5;
+		} bitfields2;
+		unsigned int ordinal2;
+	};
+
+	union {
+		struct {
+			unsigned int dst_mmreg_addr:18;
+			unsigned int reserved6:14;
+		} bitfields3;
+		unsigned int ordinal3;
+	};
+
+	uint32_t reserved7;
+
+	uint32_t data;
+
+};
+
+#endif
+
 enum {
 	CACHE_FLUSH_AND_INV_TS_EVENT = 0x00000014
 };
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_pm4_headers_aldebaran.h b/drivers/gpu/drm/amd/amdkfd/kfd_pm4_headers_aldebaran.h
index f795ec815e2a..e0ed62c4ade0 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_pm4_headers_aldebaran.h
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_pm4_headers_aldebaran.h
@@ -1,5 +1,6 @@
+/* SPDX-License-Identifier: GPL-2.0 OR MIT */
 /*
- * Copyright 2020 Advanced Micro Devices, Inc.
+ * Copyright 2020-2022 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -36,7 +37,7 @@ struct pm4_mes_map_process_aldebaran {
 		struct {
 			uint32_t pasid:16;	    /* 0 - 15  */
 			uint32_t single_memops:1;   /* 16      */
-			uint32_t reserved1:1;	    /* 17      */
+			uint32_t exec_cleaner_shader:1;	    /* 17      */
 			uint32_t debug_vmid:4;	    /* 18 - 21 */
 			uint32_t new_debug:1;	    /* 22      */
 			uint32_t tmz:1;		    /* 23      */
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_pm4_headers_diq.h b/drivers/gpu/drm/amd/amdkfd/kfd_pm4_headers_diq.h
deleted file mode 100644
index a0ff34878163..000000000000
--- a/drivers/gpu/drm/amd/amdkfd/kfd_pm4_headers_diq.h
+++ /dev/null
@@ -1,290 +0,0 @@
-/*
- * Copyright 2014 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- */
-
-#ifndef KFD_PM4_HEADERS_DIQ_H_
-#define KFD_PM4_HEADERS_DIQ_H_
-
-/*--------------------_INDIRECT_BUFFER-------------------- */
-
-#ifndef _PM4__INDIRECT_BUFFER_DEFINED
-#define _PM4__INDIRECT_BUFFER_DEFINED
-enum _INDIRECT_BUFFER_cache_policy_enum {
-	cache_policy___indirect_buffer__lru = 0,
-	cache_policy___indirect_buffer__stream = 1,
-	cache_policy___indirect_buffer__bypass = 2
-};
-
-enum {
-	IT_INDIRECT_BUFFER_PASID = 0x5C
-};
-
-struct pm4__indirect_buffer_pasid {
-	union {
-		union PM4_MES_TYPE_3_HEADER header;	/* header */
-		unsigned int ordinal1;
-	};
-
-	union {
-		struct {
-			unsigned int reserved1:2;
-			unsigned int ib_base_lo:30;
-		} bitfields2;
-		unsigned int ordinal2;
-	};
-
-	union {
-		struct {
-			unsigned int ib_base_hi:16;
-			unsigned int reserved2:16;
-		} bitfields3;
-		unsigned int ordinal3;
-	};
-
-	union {
-		unsigned int control;
-		unsigned int ordinal4;
-	};
-
-	union {
-		struct {
-			unsigned int pasid:10;
-			unsigned int reserved4:22;
-		} bitfields5;
-		unsigned int ordinal5;
-	};
-
-};
-
-#endif
-
-/*--------------------_RELEASE_MEM-------------------- */
-
-#ifndef _PM4__RELEASE_MEM_DEFINED
-#define _PM4__RELEASE_MEM_DEFINED
-enum _RELEASE_MEM_event_index_enum {
-	event_index___release_mem__end_of_pipe = 5,
-	event_index___release_mem__shader_done = 6
-};
-
-enum _RELEASE_MEM_cache_policy_enum {
-	cache_policy___release_mem__lru = 0,
-	cache_policy___release_mem__stream = 1,
-	cache_policy___release_mem__bypass = 2
-};
-
-enum _RELEASE_MEM_dst_sel_enum {
-	dst_sel___release_mem__memory_controller = 0,
-	dst_sel___release_mem__tc_l2 = 1,
-	dst_sel___release_mem__queue_write_pointer_register = 2,
-	dst_sel___release_mem__queue_write_pointer_poll_mask_bit = 3
-};
-
-enum _RELEASE_MEM_int_sel_enum {
-	int_sel___release_mem__none = 0,
-	int_sel___release_mem__send_interrupt_only = 1,
-	int_sel___release_mem__send_interrupt_after_write_confirm = 2,
-	int_sel___release_mem__send_data_after_write_confirm = 3
-};
-
-enum _RELEASE_MEM_data_sel_enum {
-	data_sel___release_mem__none = 0,
-	data_sel___release_mem__send_32_bit_low = 1,
-	data_sel___release_mem__send_64_bit_data = 2,
-	data_sel___release_mem__send_gpu_clock_counter = 3,
-	data_sel___release_mem__send_cp_perfcounter_hi_lo = 4,
-	data_sel___release_mem__store_gds_data_to_memory = 5
-};
-
-struct pm4__release_mem {
-	union {
-		union PM4_MES_TYPE_3_HEADER header;	/*header */
-		unsigned int ordinal1;
-	};
-
-	union {
-		struct {
-			unsigned int event_type:6;
-			unsigned int reserved1:2;
-			enum _RELEASE_MEM_event_index_enum event_index:4;
-			unsigned int tcl1_vol_action_ena:1;
-			unsigned int tc_vol_action_ena:1;
-			unsigned int reserved2:1;
-			unsigned int tc_wb_action_ena:1;
-			unsigned int tcl1_action_ena:1;
-			unsigned int tc_action_ena:1;
-			unsigned int reserved3:6;
-			unsigned int atc:1;
-			enum _RELEASE_MEM_cache_policy_enum cache_policy:2;
-			unsigned int reserved4:5;
-		} bitfields2;
-		unsigned int ordinal2;
-	};
-
-	union {
-		struct {
-			unsigned int reserved5:16;
-			enum _RELEASE_MEM_dst_sel_enum dst_sel:2;
-			unsigned int reserved6:6;
-			enum _RELEASE_MEM_int_sel_enum int_sel:3;
-			unsigned int reserved7:2;
-			enum _RELEASE_MEM_data_sel_enum data_sel:3;
-		} bitfields3;
-		unsigned int ordinal3;
-	};
-
-	union {
-		struct {
-			unsigned int reserved8:2;
-			unsigned int address_lo_32b:30;
-		} bitfields4;
-		struct {
-			unsigned int reserved9:3;
-			unsigned int address_lo_64b:29;
-		} bitfields5;
-		unsigned int ordinal4;
-	};
-
-	unsigned int address_hi;
-
-	unsigned int data_lo;
-
-	unsigned int data_hi;
-
-};
-#endif
-
-
-/*--------------------_SET_CONFIG_REG-------------------- */
-
-#ifndef _PM4__SET_CONFIG_REG_DEFINED
-#define _PM4__SET_CONFIG_REG_DEFINED
-
-struct pm4__set_config_reg {
-	union {
-		union PM4_MES_TYPE_3_HEADER header;	/*header */
-		unsigned int ordinal1;
-	};
-
-	union {
-		struct {
-			unsigned int reg_offset:16;
-			unsigned int reserved1:7;
-			unsigned int vmid_shift:5;
-			unsigned int insert_vmid:1;
-			unsigned int reserved2:3;
-		} bitfields2;
-		unsigned int ordinal2;
-	};
-
-	unsigned int reg_data[1];	/*1..N of these fields */
-
-};
-#endif
-
-/*--------------------_WAIT_REG_MEM-------------------- */
-
-#ifndef _PM4__WAIT_REG_MEM_DEFINED
-#define _PM4__WAIT_REG_MEM_DEFINED
-enum _WAIT_REG_MEM_function_enum {
-	function___wait_reg_mem__always_pass = 0,
-	function___wait_reg_mem__less_than_ref_value = 1,
-	function___wait_reg_mem__less_than_equal_to_the_ref_value = 2,
-	function___wait_reg_mem__equal_to_the_reference_value = 3,
-	function___wait_reg_mem__not_equal_reference_value = 4,
-	function___wait_reg_mem__greater_than_or_equal_reference_value = 5,
-	function___wait_reg_mem__greater_than_reference_value = 6,
-	function___wait_reg_mem__reserved = 7
-};
-
-enum _WAIT_REG_MEM_mem_space_enum {
-	mem_space___wait_reg_mem__register_space = 0,
-	mem_space___wait_reg_mem__memory_space = 1
-};
-
-enum _WAIT_REG_MEM_operation_enum {
-	operation___wait_reg_mem__wait_reg_mem = 0,
-	operation___wait_reg_mem__wr_wait_wr_reg = 1
-};
-
-struct pm4__wait_reg_mem {
-	union {
-		union PM4_MES_TYPE_3_HEADER header;	/*header */
-		unsigned int ordinal1;
-	};
-
-	union {
-		struct {
-			enum _WAIT_REG_MEM_function_enum function:3;
-			unsigned int reserved1:1;
-			enum _WAIT_REG_MEM_mem_space_enum mem_space:2;
-			enum _WAIT_REG_MEM_operation_enum operation:2;
-			unsigned int reserved2:24;
-		} bitfields2;
-		unsigned int ordinal2;
-	};
-
-	union {
-		struct {
-			unsigned int reserved3:2;
-			unsigned int memory_poll_addr_lo:30;
-		} bitfields3;
-		struct {
-			unsigned int register_poll_addr:16;
-			unsigned int reserved4:16;
-		} bitfields4;
-		struct {
-			unsigned int register_write_addr:16;
-			unsigned int reserved5:16;
-		} bitfields5;
-		unsigned int ordinal3;
-	};
-
-	union {
-		struct {
-			unsigned int poll_address_hi:16;
-			unsigned int reserved6:16;
-		} bitfields6;
-		struct {
-			unsigned int register_write_addr:16;
-			unsigned int reserved7:16;
-		} bitfields7;
-		unsigned int ordinal4;
-	};
-
-	unsigned int reference;
-
-	unsigned int mask;
-
-	union {
-		struct {
-			unsigned int poll_interval:16;
-			unsigned int reserved8:16;
-		} bitfields8;
-		unsigned int ordinal7;
-	};
-
-};
-#endif
-
-
-#endif /* KFD_PM4_HEADERS_DIQ_H_ */
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_pm4_headers_vi.h b/drivers/gpu/drm/amd/amdkfd/kfd_pm4_headers_vi.h
index 5466cfe1c3cc..8147395c083b 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_pm4_headers_vi.h
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_pm4_headers_vi.h
@@ -1,5 +1,6 @@
+/* SPDX-License-Identifier: GPL-2.0 OR MIT */
 /*
- * Copyright 2014 Advanced Micro Devices, Inc.
+ * Copyright 2014-2022 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -101,8 +102,8 @@ struct pm4_mes_set_resources {
 
 struct pm4_mes_runlist {
 	union {
-	    union PM4_MES_TYPE_3_HEADER   header;            /* header */
-	    uint32_t            ordinal1;
+		union PM4_MES_TYPE_3_HEADER   header;            /* header */
+		uint32_t            ordinal1;
 	};
 
 	union {
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_pm4_opcodes.h b/drivers/gpu/drm/amd/amdkfd/kfd_pm4_opcodes.h
index b72fa3b8c2d4..5bfd0f9cbe23 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_pm4_opcodes.h
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_pm4_opcodes.h
@@ -1,5 +1,6 @@
+/* SPDX-License-Identifier: GPL-2.0 OR MIT */
 /*
- * Copyright 2014 Advanced Micro Devices, Inc.
+ * Copyright 2014-2022 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_priv.h b/drivers/gpu/drm/amd/amdkfd/kfd_priv.h
index 8fd48d0ed240..70ef051511bb 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_priv.h
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_priv.h
@@ -1,5 +1,6 @@
+/* SPDX-License-Identifier: GPL-2.0 OR MIT */
 /*
- * Copyright 2014 Advanced Micro Devices, Inc.
+ * Copyright 2014-2022 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -25,12 +26,13 @@
 
 #include <linux/hashtable.h>
 #include <linux/mmu_notifier.h>
+#include <linux/memremap.h>
 #include <linux/mutex.h>
 #include <linux/types.h>
 #include <linux/atomic.h>
 #include <linux/workqueue.h>
 #include <linux/spinlock.h>
-#include <linux/kfd_ioctl.h>
+#include <uapi/linux/kfd_ioctl.h>
 #include <linux/idr.h>
 #include <linux/kfifo.h>
 #include <linux/seq_file.h>
@@ -97,11 +99,11 @@
 /*
  * Size of the per-process TBA+TMA buffer: 2 pages
  *
- * The first page is the TBA used for the CWSR ISA code. The second
- * page is used as TMA for user-mode trap handler setup in daisy-chain mode.
+ * The first chunk is the TBA used for the CWSR ISA code. The second
+ * chunk is used as TMA for user-mode trap handler setup in daisy-chain mode.
  */
 #define KFD_CWSR_TBA_TMA_SIZE (PAGE_SIZE * 2)
-#define KFD_CWSR_TMA_OFFSET PAGE_SIZE
+#define KFD_CWSR_TMA_OFFSET (PAGE_SIZE + 2048)
 
 #define KFD_MAX_NUM_OF_QUEUES_PER_DEVICE		\
 	(KFD_MAX_NUM_OF_PROCESSES *			\
@@ -109,7 +111,16 @@
 
 #define KFD_KERNEL_QUEUE_SIZE 2048
 
-#define KFD_UNMAP_LATENCY_MS	(4000)
+/*  KFD_UNMAP_LATENCY_MS is the timeout CP waiting for SDMA preemption. One XCC
+ *  can be associated to 2 SDMA engines. queue_preemption_timeout_ms is the time
+ *  driver waiting for CP returning the UNMAP_QUEUE fence. Thus the math is
+ *  queue_preemption_timeout_ms = sdma_preemption_time * 2 + cp workload
+ *  The format here makes CP workload 10% of total timeout
+ */
+#define KFD_UNMAP_LATENCY_MS	\
+	((queue_preemption_timeout_ms - queue_preemption_timeout_ms / 10) >> 1)
+
+#define KFD_MAX_SDMA_QUEUES	128
 
 /*
  * 512 = 0x200
@@ -121,7 +132,26 @@
  */
 #define KFD_QUEUE_DOORBELL_MIRROR_OFFSET 512
 
-
+/**
+ * enum kfd_ioctl_flags - KFD ioctl flags
+ * Various flags that can be set in &amdkfd_ioctl_desc.flags to control how
+ * userspace can use a given ioctl.
+ */
+enum kfd_ioctl_flags {
+	/*
+	 * @KFD_IOC_FLAG_CHECKPOINT_RESTORE:
+	 * Certain KFD ioctls such as AMDKFD_IOC_CRIU_OP can potentially
+	 * perform privileged operations and load arbitrary data into MQDs and
+	 * eventually HQD registers when the queue is mapped by HWS. In order to
+	 * prevent this we should perform additional security checks.
+	 *
+	 * This is equivalent to callers with the CHECKPOINT_RESTORE capability.
+	 *
+	 * Note: Since earlier versions of docker do not support CHECKPOINT_RESTORE,
+	 * we also allow ioctls with SYS_ADMIN capability.
+	 */
+	KFD_IOC_FLAG_CHECKPOINT_RESTORE = BIT(0),
+};
 /*
  * Kernel module parameter to specify maximum number of supported queues per
  * device
@@ -152,12 +182,6 @@ extern int send_sigterm;
  */
 extern int debug_largebar;
 
-/*
- * Ignore CRAT table during KFD initialization, can be used to work around
- * broken CRAT tables on some AMD systems
- */
-extern int ignore_crat;
-
 /* Set sh_mem_config.retry_disable on GFX v9 */
 extern int amdgpu_noretry;
 
@@ -178,24 +202,32 @@ extern int amdgpu_no_queue_eviction_on_vm_fault;
 /* Enable eviction debug messages */
 extern bool debug_evictions;
 
+extern struct mutex kfd_processes_mutex;
+
 enum cache_policy {
 	cache_policy_coherent,
 	cache_policy_noncoherent
 };
 
-#define KFD_IS_SOC15(chip) ((chip) >= CHIP_VEGA10)
+#define KFD_GC_VERSION(dev) (amdgpu_ip_version((dev)->adev, GC_HWIP, 0))
+#define KFD_IS_SOC15(dev)   ((KFD_GC_VERSION(dev)) >= (IP_VERSION(9, 0, 1)))
+#define KFD_SUPPORT_XNACK_PER_PROCESS(dev)\
+	((KFD_GC_VERSION(dev) == IP_VERSION(9, 4, 2)) ||	\
+	 (KFD_GC_VERSION(dev) == IP_VERSION(9, 4, 3)) ||	\
+	 (KFD_GC_VERSION(dev) == IP_VERSION(9, 4, 4)) ||	\
+	 (KFD_GC_VERSION(dev) == IP_VERSION(9, 5, 0)))
+
+struct kfd_node;
 
 struct kfd_event_interrupt_class {
-	bool (*interrupt_isr)(struct kfd_dev *dev,
+	bool (*interrupt_isr)(struct kfd_node *dev,
 			const uint32_t *ih_ring_entry, uint32_t *patched_ihre,
 			bool *patched_flag);
-	void (*interrupt_wq)(struct kfd_dev *dev,
+	void (*interrupt_wq)(struct kfd_node *dev,
 			const uint32_t *ih_ring_entry);
 };
 
 struct kfd_device_info {
-	enum amd_asic_type asic_family;
-	const char *asic_name;
 	uint32_t gfx_target_version;
 	const struct kfd_event_interrupt_class *event_interrupt_class;
 	unsigned int max_pasid_bits;
@@ -205,14 +237,16 @@ struct kfd_device_info {
 	uint8_t num_of_watch_points;
 	uint16_t mqd_size_aligned;
 	bool supports_cwsr;
-	bool needs_iommu_device;
 	bool needs_pci_atomics;
 	uint32_t no_atomic_fw_version;
-	unsigned int num_sdma_engines;
-	unsigned int num_xgmi_sdma_engines;
 	unsigned int num_sdma_queues_per_engine;
+	unsigned int num_reserved_sdma_queues_per_engine;
+	DECLARE_BITMAP(reserved_sdma_queues_bitmap, KFD_MAX_SDMA_QUEUES);
 };
 
+unsigned int kfd_get_num_sdma_engines(struct kfd_node *kdev);
+unsigned int kfd_get_num_xgmi_sdma_engines(struct kfd_node *kdev);
+
 struct kfd_mem_obj {
 	uint32_t range_start;
 	uint32_t range_end;
@@ -227,35 +261,80 @@ struct kfd_vmid_info {
 	uint32_t vmid_num_kfd;
 };
 
-struct kfd_dev {
-	struct kgd_dev *kgd;
-
-	const struct kfd_device_info *device_info;
-	struct pci_dev *pdev;
-	struct drm_device *ddev;
+#define MAX_KFD_NODES	8
 
-	unsigned int id;		/* topology stub index */
+struct kfd_dev;
 
-	phys_addr_t doorbell_base;	/* Start of actual doorbells used by
-					 * KFD. It is aligned for mapping
-					 * into user mode
-					 */
-	size_t doorbell_base_dw_offset;	/* Offset from the start of the PCI
-					 * doorbell BAR to the first KFD
-					 * doorbell in dwords. GFX reserves
-					 * the segment before this offset.
+struct kfd_node {
+	unsigned int node_id;
+	struct amdgpu_device *adev;     /* Duplicated here along with keeping
+					 * a copy in kfd_dev to save a hop
 					 */
+	const struct kfd2kgd_calls *kfd2kgd; /* Duplicated here along with
+					      * keeping a copy in kfd_dev to
+					      * save a hop
+					      */
+	struct kfd_vmid_info vm_info;
+	unsigned int id;                /* topology stub index */
+	uint32_t xcc_mask; /* Instance mask of XCCs present */
+	struct amdgpu_xcp *xcp;
+
+	/* Interrupts */
+	struct kfifo ih_fifo;
+	struct work_struct interrupt_work;
+	spinlock_t interrupt_lock;
+
+	/*
+	 * Interrupts of interest to KFD are copied
+	 * from the HW ring into a SW ring.
+	 */
+	bool interrupts_active;
+	uint32_t interrupt_bitmap; /* Only used for GFX 9.4.3 */
+
+	/* QCM Device instance */
+	struct device_queue_manager *dqm;
+
+	/* Global GWS resource shared between processes */
+	void *gws;
+
+	/* Clients watching SMI events */
+	struct list_head smi_clients;
+	spinlock_t smi_lock;
+	uint32_t reset_seq_num;
+
+	/* SRAM ECC flag */
+	atomic_t sram_ecc_flag;
+
+	/*spm process id */
+	unsigned int spm_pasid;
+
+	/* Maximum process number mapped to HW scheduler */
+	unsigned int max_proc_per_quantum;
+
+	unsigned int compute_vmid_bitmap;
+
+	struct kfd_local_mem_info local_mem_info;
+
+	struct kfd_dev *kfd;
+
+	/* Track per device allocated watch points */
+	uint32_t alloc_watch_ids;
+	spinlock_t watch_points_lock;
+};
+
+struct kfd_dev {
+	struct amdgpu_device *adev;
+
+	struct kfd_device_info device_info;
+
 	u32 __iomem *doorbell_kernel_ptr; /* This is a pointer for a doorbells
 					   * page used by kernel queue
 					   */
 
 	struct kgd2kfd_shared_resources shared_resources;
-	struct kfd_vmid_info vm_info;
 
 	const struct kfd2kgd_calls *kfd2kgd;
 	struct mutex doorbell_mutex;
-	DECLARE_BITMAP(doorbell_available_index,
-			KFD_MAX_NUM_OF_QUEUES_PER_PROCESS);
 
 	void *gtt_mem;
 	uint64_t gtt_start_gpu_addr;
@@ -265,33 +344,13 @@ struct kfd_dev {
 	unsigned int gtt_sa_chunk_size;
 	unsigned int gtt_sa_num_of_chunks;
 
-	/* Interrupts */
-	struct kfifo ih_fifo;
-	struct workqueue_struct *ih_wq;
-	struct work_struct interrupt_work;
-	spinlock_t interrupt_lock;
-
-	/* QCM Device instance */
-	struct device_queue_manager *dqm;
-
 	bool init_complete;
-	/*
-	 * Interrupts of interest to KFD are copied
-	 * from the HW ring into a SW ring.
-	 */
-	bool interrupts_active;
-
-	/* Debug manager */
-	struct kfd_dbgmgr *dbgmgr;
 
 	/* Firmware versions */
 	uint16_t mec_fw_version;
 	uint16_t mec2_fw_version;
 	uint16_t sdma_fw_version;
 
-	/* Maximum process number mapped to HW scheduler */
-	unsigned int max_proc_per_quantum;
-
 	/* CWSR */
 	bool cwsr_enabled;
 	const void *cwsr_isa;
@@ -302,31 +361,29 @@ struct kfd_dev {
 
 	bool pci_atomic_requested;
 
-	/* Use IOMMU v2 flag */
-	bool use_iommu_v2;
-
-	/* SRAM ECC flag */
-	atomic_t sram_ecc_flag;
-
 	/* Compute Profile ref. count */
 	atomic_t compute_profile;
 
-	/* Global GWS resource shared between processes */
-	void *gws;
-
-	/* Clients watching SMI events */
-	struct list_head smi_clients;
-	spinlock_t smi_lock;
-
-	uint32_t reset_seq_num;
-
 	struct ida doorbell_ida;
 	unsigned int max_doorbell_slices;
 
 	int noretry;
 
-	/* HMM page migration MEMORY_DEVICE_PRIVATE mapping */
-	struct dev_pagemap pgmap;
+	struct kfd_node *nodes[MAX_KFD_NODES];
+	unsigned int num_nodes;
+
+	struct workqueue_struct *ih_wq;
+
+	/* Kernel doorbells for KFD device */
+	struct amdgpu_bo *doorbells;
+
+	/* bitmap for dynamic doorbell allocation from doorbell object */
+	unsigned long *doorbell_bitmap;
+
+	/* for dynamic partitioning */
+	int kfd_dev_lock;
+
+	atomic_t kfd_processes_count;
 };
 
 enum kfd_mempool {
@@ -338,25 +395,24 @@ enum kfd_mempool {
 /* Character device interface */
 int kfd_chardev_init(void);
 void kfd_chardev_exit(void);
-struct device *kfd_chardev(void);
 
 /**
  * enum kfd_unmap_queues_filter - Enum for queue filters.
  *
- * @KFD_UNMAP_QUEUES_FILTER_SINGLE_QUEUE: Preempts single queue.
- *
  * @KFD_UNMAP_QUEUES_FILTER_ALL_QUEUES: Preempts all queues in the
  *						running queues list.
  *
+ * @KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES: Preempts all non-static queues
+ *						in the run list.
+ *
  * @KFD_UNMAP_QUEUES_FILTER_BY_PASID: Preempts queues that belongs to
  *						specific process.
  *
  */
 enum kfd_unmap_queues_filter {
-	KFD_UNMAP_QUEUES_FILTER_SINGLE_QUEUE,
-	KFD_UNMAP_QUEUES_FILTER_ALL_QUEUES,
-	KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES,
-	KFD_UNMAP_QUEUES_FILTER_BY_PASID
+	KFD_UNMAP_QUEUES_FILTER_ALL_QUEUES = 1,
+	KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES = 2,
+	KFD_UNMAP_QUEUES_FILTER_BY_PASID = 3
 };
 
 /**
@@ -371,13 +427,16 @@ enum kfd_unmap_queues_filter {
  * @KFD_QUEUE_TYPE_DIQ: DIQ queue type.
  *
  * @KFD_QUEUE_TYPE_SDMA_XGMI: Special SDMA queue for XGMI interface.
+ *
+ * @KFD_QUEUE_TYPE_SDMA_BY_ENG_ID:  SDMA user mode queue with target SDMA engine ID.
  */
 enum kfd_queue_type  {
 	KFD_QUEUE_TYPE_COMPUTE,
 	KFD_QUEUE_TYPE_SDMA,
 	KFD_QUEUE_TYPE_HIQ,
 	KFD_QUEUE_TYPE_DIQ,
-	KFD_QUEUE_TYPE_SDMA_XGMI
+	KFD_QUEUE_TYPE_SDMA_XGMI,
+	KFD_QUEUE_TYPE_SDMA_BY_ENG_ID
 };
 
 enum kfd_queue_format {
@@ -442,6 +501,7 @@ enum KFD_QUEUE_PRIORITY {
  * it's user mode or kernel mode queue.
  *
  */
+
 struct queue_properties {
 	enum kfd_queue_type type;
 	enum kfd_queue_format format;
@@ -450,14 +510,19 @@ struct queue_properties {
 	uint64_t  queue_size;
 	uint32_t priority;
 	uint32_t queue_percent;
-	uint32_t *read_ptr;
-	uint32_t *write_ptr;
+	void __user *read_ptr;
+	void __user *write_ptr;
 	void __iomem *doorbell_ptr;
 	uint32_t doorbell_off;
 	bool is_interop;
 	bool is_evicted;
+	bool is_suspended;
+	bool is_being_destroyed;
 	bool is_active;
 	bool is_gws;
+	uint32_t pm4_target_xcc;
+	bool is_dbg_wa;
+	bool is_user_cu_masked;
 	/* Not relevant for user mode queues in cp scheduling */
 	unsigned int vmid;
 	/* Relevant only for sdma queues*/
@@ -472,15 +537,25 @@ struct queue_properties {
 	uint32_t ctl_stack_size;
 	uint64_t tba_addr;
 	uint64_t tma_addr;
+	uint64_t exception_status;
+
+	struct amdgpu_bo *wptr_bo;
+	struct amdgpu_bo *rptr_bo;
+	struct amdgpu_bo *ring_bo;
+	struct amdgpu_bo *eop_buf_bo;
+	struct amdgpu_bo *cwsr_bo;
 };
 
 #define QUEUE_IS_ACTIVE(q) ((q).queue_size > 0 &&	\
 			    (q).queue_address != 0 &&	\
 			    (q).queue_percent > 0 &&	\
-			    !(q).is_evicted)
+			    !(q).is_evicted &&		\
+			    !(q).is_suspended)
 
 enum mqd_update_flag {
-	UPDATE_FLAG_CU_MASK = 0,
+	UPDATE_FLAG_DBG_WA_ENABLE = 1,
+	UPDATE_FLAG_DBG_WA_DISABLE = 2,
+	UPDATE_FLAG_IS_GWS = 4, /* quirk for gfx9 IP */
 };
 
 struct mqd_update_info {
@@ -541,11 +616,17 @@ struct queue {
 	unsigned int doorbell_id;
 
 	struct kfd_process	*process;
-	struct kfd_dev		*device;
+	struct kfd_node		*device;
 	void *gws;
 
 	/* procfs */
 	struct kobject kobj;
+
+	void *gang_ctx_bo;
+	uint64_t gang_ctx_gpu_addr;
+	void *gang_ctx_cpu_ptr;
+
+	struct amdgpu_bo *wptr_bo_gart;
 };
 
 enum KFD_MQD_TYPE {
@@ -630,7 +711,10 @@ struct qcm_process_device {
 	uint64_t ib_base;
 	void *ib_kaddr;
 
-	/* doorbell resources per process per device */
+	/* doorbells for kfd process */
+	struct amdgpu_bo *proc_doorbells;
+
+	/* bitmap for dynamic doorbell allocation from the bo */
 	unsigned long *doorbell_bitmap;
 };
 
@@ -669,7 +753,7 @@ enum kfd_pdd_bound {
 /* Data that is per-process-per device. */
 struct kfd_process_device {
 	/* The device that owns this data. */
-	struct kfd_dev *dev;
+	struct kfd_node *dev;
 
 	/* The process that owns this kfd_process_device. */
 	struct kfd_process *process;
@@ -704,7 +788,7 @@ struct kfd_process_device {
 	enum kfd_pdd_bound bound;
 
 	/* VRAM usage */
-	uint64_t vram_usage;
+	atomic64_t vram_usage;
 	struct attribute attr_vram;
 	char vram_filename[MAX_SYSFS_FILENAME_LEN];
 
@@ -719,7 +803,6 @@ struct kfd_process_device {
 	struct attribute attr_evict;
 
 	struct kobject *kobj_stats;
-	unsigned int doorbell_index;
 
 	/*
 	 * @cu_occupancy: Reports occupancy of Compute Units (CU) of a process
@@ -754,6 +837,33 @@ struct kfd_process_device {
 	uint64_t faults;
 	uint64_t page_in;
 	uint64_t page_out;
+
+	/* Exception code status*/
+	uint64_t exception_status;
+	void *vm_fault_exc_data;
+	size_t vm_fault_exc_data_size;
+
+	/* Tracks debug per-vmid request settings */
+	uint32_t spi_dbg_override;
+	uint32_t spi_dbg_launch_mode;
+	uint32_t watch_points[4];
+	uint32_t alloc_watch_ids;
+
+	/*
+	 * If this process has been checkpointed before, then the user
+	 * application will use the original gpu_id on the
+	 * checkpointed node to refer to this device.
+	 */
+	uint32_t user_gpu_id;
+
+	void *proc_ctx_bo;
+	uint64_t proc_ctx_gpu_addr;
+	void *proc_ctx_cpu_ptr;
+
+	/* Tracks queue reset status */
+	bool has_reset_queue;
+
+	u32 pasid;
 };
 
 #define qpd_to_pdd(x) container_of(x, struct kfd_process_device, qpd)
@@ -764,12 +874,21 @@ struct svm_range_list {
 	struct list_head		list;
 	struct work_struct		deferred_list_work;
 	struct list_head		deferred_range_list;
+	struct list_head                criu_svm_metadata_list;
 	spinlock_t			deferred_list_lock;
 	atomic_t			evicted_ranges;
 	atomic_t			drain_pagefaults;
 	struct delayed_work		restore_work;
 	DECLARE_BITMAP(bitmap_supported, MAX_GPU_INSTANCE);
-	struct task_struct 		*faulting_task;
+	struct task_struct		*faulting_task;
+	/* check point ts decides if page fault recovery need be dropped */
+	uint64_t			checkpoint_ts[MAX_GPU_INSTANCE];
+
+	/* Default granularity to use in buffer migration
+	 * and restoration of backing memory while handling
+	 * recoverable page faults
+	 */
+	uint8_t default_granularity;
 };
 
 /* Process data */
@@ -804,8 +923,6 @@ struct kfd_process {
 	/* We want to receive a notification when the mm_struct is destroyed */
 	struct mmu_notifier mmu_notifier;
 
-	u32 pasid;
-
 	/*
 	 * Array of kfd_process_device pointers,
 	 * one for each device the process is using.
@@ -835,7 +952,7 @@ struct kfd_process {
 	 * fence will be triggered during eviction and new one will be created
 	 * during restore
 	 */
-	struct dma_fence *ef;
+	struct dma_fence __rcu *ef;
 
 	/* Work items for evicting and restoring BOs */
 	struct delayed_work eviction_work;
@@ -847,15 +964,60 @@ struct kfd_process {
 	 */
 	unsigned long last_restore_timestamp;
 
+	/* Indicates device process is debug attached with reserved vmid. */
+	bool debug_trap_enabled;
+
+	/* per-process-per device debug event fd file */
+	struct file *dbg_ev_file;
+
+	/* If the process is a kfd debugger, we need to know so we can clean
+	 * up at exit time.  If a process enables debugging on itself, it does
+	 * its own clean-up, so we don't set the flag here.  We track this by
+	 * counting the number of processes this process is debugging.
+	 */
+	atomic_t debugged_process_count;
+
+	/* If the process is a debugged, this is the debugger process */
+	struct kfd_process *debugger_process;
+
 	/* Kobj for our procfs */
 	struct kobject *kobj;
 	struct kobject *kobj_queues;
 	struct attribute attr_pasid;
 
+	/* Keep track cwsr init */
+	bool has_cwsr;
+
+	/* Exception code enable mask and status */
+	uint64_t exception_enable_mask;
+	uint64_t exception_status;
+
+	/* Used to drain stale interrupts */
+	wait_queue_head_t wait_irq_drain;
+	bool irq_drain_is_open;
+
 	/* shared virtual memory registered by this process */
 	struct svm_range_list svms;
 
 	bool xnack_enabled;
+
+	/* Work area for debugger event writer worker. */
+	struct work_struct debug_event_workarea;
+
+	/* Tracks debug per-vmid request for debug flags */
+	u32 dbg_flags;
+
+	atomic_t poison;
+	/* Queues are in paused stated because we are in the process of doing a CRIU checkpoint */
+	bool queues_paused;
+
+	/* Tracks runtime enable status */
+	struct semaphore runtime_enable_sema;
+	bool is_runtime_retry;
+	struct kfd_runtime_info runtime_info;
+
+	/* if gpu page fault sent to KFD */
+	bool gpu_page_fault;
 };
 
 #define KFD_PROCESS_TABLE_SIZE 5 /* bits: 32 entries */
@@ -881,19 +1043,20 @@ struct amdkfd_ioctl_desc {
 	unsigned int cmd_drv;
 	const char *name;
 };
-bool kfd_dev_is_large_bar(struct kfd_dev *dev);
+bool kfd_dev_is_large_bar(struct kfd_node *dev);
 
 int kfd_process_create_wq(void);
 void kfd_process_destroy_wq(void);
-struct kfd_process *kfd_create_process(struct file *filep);
-struct kfd_process *kfd_get_process(const struct task_struct *);
-struct kfd_process *kfd_lookup_process_by_pasid(u32 pasid);
+void kfd_cleanup_processes(void);
+struct kfd_process *kfd_create_process(struct task_struct *thread);
+struct kfd_process *kfd_get_process(const struct task_struct *task);
+struct kfd_process *kfd_lookup_process_by_pasid(u32 pasid,
+						 struct kfd_process_device **pdd);
 struct kfd_process *kfd_lookup_process_by_mm(const struct mm_struct *mm);
 
 int kfd_process_gpuidx_from_gpuid(struct kfd_process *p, uint32_t gpu_id);
-int kfd_process_gpuid_from_kgd(struct kfd_process *p,
-			       struct amdgpu_device *adev, uint32_t *gpuid,
-			       uint32_t *gpuidx);
+int kfd_process_gpuid_from_node(struct kfd_process *p, struct kfd_node *node,
+				uint32_t *gpuid, uint32_t *gpuidx);
 static inline int kfd_process_gpuid_from_gpuidx(struct kfd_process *p,
 				uint32_t gpuidx, uint32_t *gpuid) {
 	return gpuidx < p->n_pdds ? p->pdds[gpuidx]->dev->id : -EINVAL;
@@ -904,23 +1067,28 @@ static inline struct kfd_process_device *kfd_process_device_from_gpuidx(
 }
 
 void kfd_unref_process(struct kfd_process *p);
-int kfd_process_evict_queues(struct kfd_process *p);
+int kfd_process_evict_queues(struct kfd_process *p, uint32_t trigger);
 int kfd_process_restore_queues(struct kfd_process *p);
 void kfd_suspend_all_processes(void);
 int kfd_resume_all_processes(void);
 
+struct kfd_process_device *kfd_process_device_data_by_id(struct kfd_process *process,
+							 uint32_t gpu_id);
+
+int kfd_process_get_user_gpu_id(struct kfd_process *p, uint32_t actual_gpu_id);
+
 int kfd_process_device_init_vm(struct kfd_process_device *pdd,
 			       struct file *drm_file);
-struct kfd_process_device *kfd_bind_process_to_device(struct kfd_dev *dev,
+struct kfd_process_device *kfd_bind_process_to_device(struct kfd_node *dev,
 						struct kfd_process *p);
-struct kfd_process_device *kfd_get_process_device_data(struct kfd_dev *dev,
+struct kfd_process_device *kfd_get_process_device_data(struct kfd_node *dev,
 							struct kfd_process *p);
-struct kfd_process_device *kfd_create_process_device_data(struct kfd_dev *dev,
+struct kfd_process_device *kfd_create_process_device_data(struct kfd_node *dev,
 							struct kfd_process *p);
 
 bool kfd_process_xnack_mode(struct kfd_process *p, bool supported);
 
-int kfd_reserved_mem_mmap(struct kfd_dev *dev, struct kfd_process *process,
+int kfd_reserved_mem_mmap(struct kfd_node *dev, struct kfd_process *process,
 			  struct vm_area_struct *vma);
 
 /* KFD process API for creating and translating handles */
@@ -930,12 +1098,11 @@ void *kfd_process_device_translate_handle(struct kfd_process_device *p,
 					int handle);
 void kfd_process_device_remove_obj_handle(struct kfd_process_device *pdd,
 					int handle);
+struct kfd_process *kfd_lookup_process_by_pid(struct pid *pid);
 
 /* PASIDs */
 int kfd_pasid_init(void);
 void kfd_pasid_exit(void);
-bool kfd_set_pasid_limit(unsigned int new_limit);
-unsigned int kfd_get_pasid_limit(void);
 u32 kfd_pasid_alloc(void);
 void kfd_pasid_free(u32 pasid);
 
@@ -943,7 +1110,7 @@ void kfd_pasid_free(u32 pasid);
 size_t kfd_doorbell_process_slice(struct kfd_dev *kfd);
 int kfd_doorbell_init(struct kfd_dev *kfd);
 void kfd_doorbell_fini(struct kfd_dev *kfd);
-int kfd_doorbell_mmap(struct kfd_dev *dev, struct kfd_process *process,
+int kfd_doorbell_mmap(struct kfd_node *dev, struct kfd_process *process,
 		      struct vm_area_struct *vma);
 void __iomem *kfd_get_kernel_doorbell(struct kfd_dev *kfd,
 					unsigned int *doorbell_off);
@@ -956,15 +1123,15 @@ unsigned int kfd_get_doorbell_dw_offset_in_bar(struct kfd_dev *kfd,
 					unsigned int doorbell_id);
 phys_addr_t kfd_get_process_doorbells(struct kfd_process_device *pdd);
 int kfd_alloc_process_doorbells(struct kfd_dev *kfd,
-				unsigned int *doorbell_index);
+				struct kfd_process_device *pdd);
 void kfd_free_process_doorbells(struct kfd_dev *kfd,
-				unsigned int doorbell_index);
+				struct kfd_process_device *pdd);
 /* GTT Sub-Allocator */
 
-int kfd_gtt_sa_allocate(struct kfd_dev *kfd, unsigned int size,
+int kfd_gtt_sa_allocate(struct kfd_node *node, unsigned int size,
 			struct kfd_mem_obj **mem_obj);
 
-int kfd_gtt_sa_free(struct kfd_dev *kfd, struct kfd_mem_obj *mem_obj);
+int kfd_gtt_sa_free(struct kfd_node *node, struct kfd_mem_obj *mem_obj);
 
 extern struct device *kfd_device;
 
@@ -977,25 +1144,53 @@ void kfd_procfs_del_queue(struct queue *q);
 /* Topology */
 int kfd_topology_init(void);
 void kfd_topology_shutdown(void);
-int kfd_topology_add_device(struct kfd_dev *gpu);
-int kfd_topology_remove_device(struct kfd_dev *gpu);
+int kfd_topology_add_device(struct kfd_node *gpu);
+int kfd_topology_remove_device(struct kfd_node *gpu);
 struct kfd_topology_device *kfd_topology_device_by_proximity_domain(
 						uint32_t proximity_domain);
+struct kfd_topology_device *kfd_topology_device_by_proximity_domain_no_lock(
+						uint32_t proximity_domain);
 struct kfd_topology_device *kfd_topology_device_by_id(uint32_t gpu_id);
-struct kfd_dev *kfd_device_by_id(uint32_t gpu_id);
-struct kfd_dev *kfd_device_by_pci_dev(const struct pci_dev *pdev);
-struct kfd_dev *kfd_device_by_kgd(const struct kgd_dev *kgd);
-int kfd_topology_enum_kfd_devices(uint8_t idx, struct kfd_dev **kdev);
+struct kfd_node *kfd_device_by_id(uint32_t gpu_id);
+static inline bool kfd_irq_is_from_node(struct kfd_node *node, uint32_t node_id,
+					uint32_t vmid)
+{
+	return (node->interrupt_bitmap & (1 << node_id)) != 0 &&
+	       (node->compute_vmid_bitmap & (1 << vmid)) != 0;
+}
+static inline struct kfd_node *kfd_node_by_irq_ids(struct amdgpu_device *adev,
+					uint32_t node_id, uint32_t vmid) {
+	struct kfd_dev *dev = adev->kfd.dev;
+	uint32_t i;
+
+	if (KFD_GC_VERSION(dev) != IP_VERSION(9, 4, 3) &&
+	    KFD_GC_VERSION(dev) != IP_VERSION(9, 4, 4) &&
+	    KFD_GC_VERSION(dev) != IP_VERSION(9, 5, 0))
+		return dev->nodes[0];
+
+	for (i = 0; i < dev->num_nodes; i++)
+		if (kfd_irq_is_from_node(dev->nodes[i], node_id, vmid))
+			return dev->nodes[i];
+
+	return NULL;
+}
+int kfd_topology_enum_kfd_devices(uint8_t idx, struct kfd_node **kdev);
 int kfd_numa_node_to_apic_id(int numa_node_id);
-void kfd_double_confirm_iommu_support(struct kfd_dev *gpu);
 
 /* Interrupts */
-int kfd_interrupt_init(struct kfd_dev *dev);
-void kfd_interrupt_exit(struct kfd_dev *dev);
-bool enqueue_ih_ring_entry(struct kfd_dev *kfd,	const void *ih_ring_entry);
-bool interrupt_is_wanted(struct kfd_dev *dev,
+#define	KFD_IRQ_FENCE_CLIENTID	0xff
+#define	KFD_IRQ_FENCE_SOURCEID	0xff
+#define	KFD_IRQ_IS_FENCE(client, source)				\
+				((client) == KFD_IRQ_FENCE_CLIENTID &&	\
+				(source) == KFD_IRQ_FENCE_SOURCEID)
+int kfd_interrupt_init(struct kfd_node *dev);
+void kfd_interrupt_exit(struct kfd_node *dev);
+bool enqueue_ih_ring_entry(struct kfd_node *kfd, const void *ih_ring_entry);
+bool interrupt_is_wanted(struct kfd_node *dev,
 				const uint32_t *ih_ring_entry,
 				uint32_t *patched_ihre, bool *flag);
+int kfd_process_drain_interrupts(struct kfd_process_device *pdd);
+void kfd_process_close_interrupt_drain(unsigned int pasid);
 
 /* amdkfd Apertures */
 int kfd_init_apertures(struct kfd_process *process);
@@ -1003,31 +1198,156 @@ int kfd_init_apertures(struct kfd_process *process);
 void kfd_process_set_trap_handler(struct qcm_process_device *qpd,
 				  uint64_t tba_addr,
 				  uint64_t tma_addr);
+void kfd_process_set_trap_debug_flag(struct qcm_process_device *qpd,
+				     bool enabled);
+
+/* CWSR initialization */
+int kfd_process_init_cwsr_apu(struct kfd_process *process, struct file *filep);
+
+/* CRIU */
+/*
+ * Need to increment KFD_CRIU_PRIV_VERSION each time a change is made to any of the CRIU private
+ * structures:
+ * kfd_criu_process_priv_data
+ * kfd_criu_device_priv_data
+ * kfd_criu_bo_priv_data
+ * kfd_criu_queue_priv_data
+ * kfd_criu_event_priv_data
+ * kfd_criu_svm_range_priv_data
+ */
+
+#define KFD_CRIU_PRIV_VERSION 1
+
+struct kfd_criu_process_priv_data {
+	uint32_t version;
+	uint32_t xnack_mode;
+};
+
+struct kfd_criu_device_priv_data {
+	/* For future use */
+	uint64_t reserved;
+};
+
+struct kfd_criu_bo_priv_data {
+	uint64_t user_addr;
+	uint32_t idr_handle;
+	uint32_t mapped_gpuids[MAX_GPU_INSTANCE];
+};
+
+/*
+ * The first 4 bytes of kfd_criu_queue_priv_data, kfd_criu_event_priv_data,
+ * kfd_criu_svm_range_priv_data is the object type
+ */
+enum kfd_criu_object_type {
+	KFD_CRIU_OBJECT_TYPE_QUEUE,
+	KFD_CRIU_OBJECT_TYPE_EVENT,
+	KFD_CRIU_OBJECT_TYPE_SVM_RANGE,
+};
+
+struct kfd_criu_svm_range_priv_data {
+	uint32_t object_type;
+	uint64_t start_addr;
+	uint64_t size;
+	/* Variable length array of attributes */
+	struct kfd_ioctl_svm_attribute attrs[];
+};
+
+struct kfd_criu_queue_priv_data {
+	uint32_t object_type;
+	uint64_t q_address;
+	uint64_t q_size;
+	uint64_t read_ptr_addr;
+	uint64_t write_ptr_addr;
+	uint64_t doorbell_off;
+	uint64_t eop_ring_buffer_address;
+	uint64_t ctx_save_restore_area_address;
+	uint32_t gpu_id;
+	uint32_t type;
+	uint32_t format;
+	uint32_t q_id;
+	uint32_t priority;
+	uint32_t q_percent;
+	uint32_t doorbell_id;
+	uint32_t gws;
+	uint32_t sdma_id;
+	uint32_t eop_ring_buffer_size;
+	uint32_t ctx_save_restore_area_size;
+	uint32_t ctl_stack_size;
+	uint32_t mqd_size;
+};
+
+struct kfd_criu_event_priv_data {
+	uint32_t object_type;
+	uint64_t user_handle;
+	uint32_t event_id;
+	uint32_t auto_reset;
+	uint32_t type;
+	uint32_t signaled;
+
+	union {
+		struct kfd_hsa_memory_exception_data memory_exception_data;
+		struct kfd_hsa_hw_exception_data hw_exception_data;
+	};
+};
+
+int kfd_process_get_queue_info(struct kfd_process *p,
+			       uint32_t *num_queues,
+			       uint64_t *priv_data_sizes);
+
+int kfd_criu_checkpoint_queues(struct kfd_process *p,
+			 uint8_t __user *user_priv_data,
+			 uint64_t *priv_data_offset);
+
+int kfd_criu_restore_queue(struct kfd_process *p,
+			   uint8_t __user *user_priv_data,
+			   uint64_t *priv_data_offset,
+			   uint64_t max_priv_data_size);
+
+int kfd_criu_checkpoint_events(struct kfd_process *p,
+			 uint8_t __user *user_priv_data,
+			 uint64_t *priv_data_offset);
+
+int kfd_criu_restore_event(struct file *devkfd,
+			   struct kfd_process *p,
+			   uint8_t __user *user_priv_data,
+			   uint64_t *priv_data_offset,
+			   uint64_t max_priv_data_size);
+/* CRIU - End */
 
 /* Queue Context Management */
 int init_queue(struct queue **q, const struct queue_properties *properties);
 void uninit_queue(struct queue *q);
 void print_queue_properties(struct queue_properties *q);
 void print_queue(struct queue *q);
+int kfd_queue_buffer_get(struct amdgpu_vm *vm, void __user *addr, struct amdgpu_bo **pbo,
+			 u64 expected_size);
+void kfd_queue_buffer_put(struct amdgpu_bo **bo);
+int kfd_queue_acquire_buffers(struct kfd_process_device *pdd, struct queue_properties *properties);
+int kfd_queue_release_buffers(struct kfd_process_device *pdd, struct queue_properties *properties);
+void kfd_queue_unref_bo_va(struct amdgpu_vm *vm, struct amdgpu_bo **bo);
+int kfd_queue_unref_bo_vas(struct kfd_process_device *pdd,
+			   struct queue_properties *properties);
+void kfd_queue_ctx_save_restore_size(struct kfd_topology_device *dev);
 
 struct mqd_manager *mqd_manager_init_cik(enum KFD_MQD_TYPE type,
-		struct kfd_dev *dev);
-struct mqd_manager *mqd_manager_init_cik_hawaii(enum KFD_MQD_TYPE type,
-		struct kfd_dev *dev);
+		struct kfd_node *dev);
 struct mqd_manager *mqd_manager_init_vi(enum KFD_MQD_TYPE type,
-		struct kfd_dev *dev);
-struct mqd_manager *mqd_manager_init_vi_tonga(enum KFD_MQD_TYPE type,
-		struct kfd_dev *dev);
+		struct kfd_node *dev);
 struct mqd_manager *mqd_manager_init_v9(enum KFD_MQD_TYPE type,
-		struct kfd_dev *dev);
+		struct kfd_node *dev);
 struct mqd_manager *mqd_manager_init_v10(enum KFD_MQD_TYPE type,
-		struct kfd_dev *dev);
-struct device_queue_manager *device_queue_manager_init(struct kfd_dev *dev);
+		struct kfd_node *dev);
+struct mqd_manager *mqd_manager_init_v11(enum KFD_MQD_TYPE type,
+		struct kfd_node *dev);
+struct mqd_manager *mqd_manager_init_v12(enum KFD_MQD_TYPE type,
+		struct kfd_node *dev);
+struct device_queue_manager *device_queue_manager_init(struct kfd_node *dev);
 void device_queue_manager_uninit(struct device_queue_manager *dqm);
-struct kernel_queue *kernel_queue_init(struct kfd_dev *dev,
+struct kernel_queue *kernel_queue_init(struct kfd_node *dev,
 					enum kfd_queue_type type);
-void kernel_queue_uninit(struct kernel_queue *kq, bool hanging);
-int kfd_process_vm_fault(struct device_queue_manager *dqm, u32 pasid);
+void kernel_queue_uninit(struct kernel_queue *kq);
+int kfd_evict_process_device(struct kfd_process_device *pdd);
+int kfd_dqm_suspend_bad_queue_mes(struct kfd_node *knode, u32 pasid, u32 doorbell_id);
 
 /* Process Queue Manager */
 struct process_queue_node {
@@ -1041,10 +1361,12 @@ void kfd_process_dequeue_from_all_devices(struct kfd_process *p);
 int pqm_init(struct process_queue_manager *pqm, struct kfd_process *p);
 void pqm_uninit(struct process_queue_manager *pqm);
 int pqm_create_queue(struct process_queue_manager *pqm,
-			    struct kfd_dev *dev,
-			    struct file *f,
+			    struct kfd_node *dev,
 			    struct queue_properties *properties,
 			    unsigned int *qid,
+			    const struct kfd_criu_queue_priv_data *q_data,
+			    const void *restore_mqd,
+			    const void *restore_ctl_stack,
 			    uint32_t *p_doorbell_offset_in_process);
 int pqm_destroy_queue(struct process_queue_manager *pqm, unsigned int qid);
 int pqm_update_queue_properties(struct process_queue_manager *pqm, unsigned int qid,
@@ -1053,8 +1375,6 @@ int pqm_update_mqd(struct process_queue_manager *pqm, unsigned int qid,
 			struct mqd_update_info *minfo);
 int pqm_set_gws(struct process_queue_manager *pqm, unsigned int qid,
 			void *gws);
-struct kernel_queue *pqm_get_kernel_queue(struct process_queue_manager *pqm,
-						unsigned int qid);
 struct queue *pqm_get_user_queue(struct process_queue_manager *pqm,
 						unsigned int qid);
 int pqm_get_wave_state(struct process_queue_manager *pqm,
@@ -1062,16 +1382,43 @@ int pqm_get_wave_state(struct process_queue_manager *pqm,
 		       void __user *ctl_stack,
 		       u32 *ctl_stack_used_size,
 		       u32 *save_area_used_size);
+int pqm_get_queue_snapshot(struct process_queue_manager *pqm,
+			   uint64_t exception_clear_mask,
+			   void __user *buf,
+			   int *num_qss_entries,
+			   uint32_t *entry_size);
 
-int amdkfd_fence_wait_timeout(uint64_t *fence_addr,
+int amdkfd_fence_wait_timeout(struct device_queue_manager *dqm,
 			      uint64_t fence_value,
 			      unsigned int timeout_ms);
 
+int pqm_get_queue_checkpoint_info(struct process_queue_manager *pqm,
+				  unsigned int qid,
+				  u32 *mqd_size,
+				  u32 *ctl_stack_size);
 /* Packet Manager */
 
 #define KFD_FENCE_COMPLETED (100)
 #define KFD_FENCE_INIT   (10)
 
+/**
+ * enum kfd_config_dequeue_wait_counts_cmd - Command for configuring
+ *  dequeue wait counts.
+ *
+ * @KFD_DEQUEUE_WAIT_INIT: Set optimized dequeue wait counts for a
+ *	certain ASICs. For these ASICs, this is default value used by RESET
+ * @KFD_DEQUEUE_WAIT_RESET: Reset dequeue wait counts to the optimized value
+ *	for certain ASICs. For others set it to default hardware reset value
+ * @KFD_DEQUEUE_WAIT_SET_SCH_WAVE: Set context switch latency wait
+ *
+ */
+enum kfd_config_dequeue_wait_counts_cmd {
+	KFD_DEQUEUE_WAIT_INIT = 1,
+	KFD_DEQUEUE_WAIT_RESET = 2,
+	KFD_DEQUEUE_WAIT_SET_SCH_WAVE = 3
+};
+
+
 struct packet_manager {
 	struct device_queue_manager *dqm;
 	struct kernel_queue *priv_queue;
@@ -1095,10 +1442,10 @@ struct packet_manager_funcs {
 	int (*map_queues)(struct packet_manager *pm, uint32_t *buffer,
 			struct queue *q, bool is_static);
 	int (*unmap_queues)(struct packet_manager *pm, uint32_t *buffer,
-			enum kfd_queue_type type,
 			enum kfd_unmap_queues_filter mode,
-			uint32_t filter_param, bool reset,
-			unsigned int sdma_engine);
+			uint32_t filter_param, bool reset);
+	int (*config_dequeue_wait_counts)(struct packet_manager *pm, uint32_t *buffer,
+			enum kfd_config_dequeue_wait_counts_cmd cmd, uint32_t value);
 	int (*query_status)(struct packet_manager *pm, uint32_t *buffer,
 			uint64_t fence_address,	uint64_t fence_value);
 	int (*release_mem)(uint64_t gpu_addr, uint32_t *buffer);
@@ -1109,6 +1456,7 @@ struct packet_manager_funcs {
 	int set_resources_size;
 	int map_queues_size;
 	int unmap_queues_size;
+	int config_dequeue_wait_counts_size;
 	int query_status_size;
 	int release_mem_size;
 };
@@ -1118,20 +1466,23 @@ extern const struct packet_manager_funcs kfd_v9_pm_funcs;
 extern const struct packet_manager_funcs kfd_aldebaran_pm_funcs;
 
 int pm_init(struct packet_manager *pm, struct device_queue_manager *dqm);
-void pm_uninit(struct packet_manager *pm, bool hanging);
+void pm_uninit(struct packet_manager *pm);
 int pm_send_set_resources(struct packet_manager *pm,
 				struct scheduling_resources *res);
 int pm_send_runlist(struct packet_manager *pm, struct list_head *dqm_queues);
 int pm_send_query_status(struct packet_manager *pm, uint64_t fence_address,
 				uint64_t fence_value);
 
-int pm_send_unmap_queue(struct packet_manager *pm, enum kfd_queue_type type,
+int pm_send_unmap_queue(struct packet_manager *pm,
 			enum kfd_unmap_queues_filter mode,
-			uint32_t filter_param, bool reset,
-			unsigned int sdma_engine);
+			uint32_t filter_param, bool reset);
 
 void pm_release_ib(struct packet_manager *pm);
 
+int pm_config_dequeue_wait_counts(struct packet_manager *pm,
+			enum kfd_config_dequeue_wait_counts_cmd cmd,
+			uint32_t wait_counts_config);
+
 /* Following PM funcs can be shared among VI and AI */
 unsigned int pm_build_pm4_header(unsigned int opcode, size_t packet_size);
 
@@ -1140,55 +1491,80 @@ uint64_t kfd_get_number_elems(struct kfd_dev *kfd);
 /* Events */
 extern const struct kfd_event_interrupt_class event_interrupt_class_cik;
 extern const struct kfd_event_interrupt_class event_interrupt_class_v9;
+extern const struct kfd_event_interrupt_class event_interrupt_class_v9_4_3;
+extern const struct kfd_event_interrupt_class event_interrupt_class_v10;
+extern const struct kfd_event_interrupt_class event_interrupt_class_v11;
 
 extern const struct kfd_device_global_init_class device_global_init_class_cik;
 
-void kfd_event_init_process(struct kfd_process *p);
+int kfd_event_init_process(struct kfd_process *p);
 void kfd_event_free_process(struct kfd_process *p);
 int kfd_event_mmap(struct kfd_process *process, struct vm_area_struct *vma);
 int kfd_wait_on_events(struct kfd_process *p,
 		       uint32_t num_events, void __user *data,
-		       bool all, uint32_t user_timeout_ms,
+		       bool all, uint32_t *user_timeout_ms,
 		       uint32_t *wait_result);
 void kfd_signal_event_interrupt(u32 pasid, uint32_t partial_id,
 				uint32_t valid_id_bits);
-void kfd_signal_iommu_event(struct kfd_dev *dev,
-			    u32 pasid, unsigned long address,
-			    bool is_write_requested, bool is_execute_requested);
 void kfd_signal_hw_exception_event(u32 pasid);
 int kfd_set_event(struct kfd_process *p, uint32_t event_id);
 int kfd_reset_event(struct kfd_process *p, uint32_t event_id);
-int kfd_event_page_set(struct kfd_process *p, void *kernel_address,
-		       uint64_t size);
+int kfd_kmap_event_page(struct kfd_process *p, uint64_t event_page_offset);
+
 int kfd_event_create(struct file *devkfd, struct kfd_process *p,
 		     uint32_t event_type, bool auto_reset, uint32_t node_id,
 		     uint32_t *event_id, uint32_t *event_trigger_data,
 		     uint64_t *event_page_offset, uint32_t *event_slot_index);
+
+int kfd_get_num_events(struct kfd_process *p);
 int kfd_event_destroy(struct kfd_process *p, uint32_t event_id);
 
-void kfd_signal_vm_fault_event(struct kfd_dev *dev, u32 pasid,
-				struct kfd_vm_fault_info *info);
+void kfd_signal_vm_fault_event_with_userptr(struct kfd_process *p, uint64_t gpu_va);
 
-void kfd_signal_reset_event(struct kfd_dev *dev);
+void kfd_signal_vm_fault_event(struct kfd_process_device *pdd,
+				struct kfd_vm_fault_info *info,
+				struct kfd_hsa_memory_exception_data *data);
 
-void kfd_signal_poison_consumed_event(struct kfd_dev *dev, u32 pasid);
+void kfd_signal_reset_event(struct kfd_node *dev);
 
-void kfd_flush_tlb(struct kfd_process_device *pdd, enum TLB_FLUSH_TYPE type);
+void kfd_signal_poison_consumed_event(struct kfd_node *dev, u32 pasid);
 
-int dbgdev_wave_reset_wavefronts(struct kfd_dev *dev, struct kfd_process *p);
+static inline void kfd_flush_tlb(struct kfd_process_device *pdd,
+				 enum TLB_FLUSH_TYPE type)
+{
+	struct amdgpu_device *adev = pdd->dev->adev;
+	struct amdgpu_vm *vm = drm_priv_to_vm(pdd->drm_priv);
+
+	amdgpu_vm_flush_compute_tlb(adev, vm, type, pdd->dev->xcc_mask);
+}
+
+static inline bool kfd_flush_tlb_after_unmap(struct kfd_dev *dev)
+{
+	return KFD_GC_VERSION(dev) >= IP_VERSION(9, 4, 2) ||
+	       (KFD_GC_VERSION(dev) == IP_VERSION(9, 4, 1) && dev->sdma_fw_version >= 18) ||
+	       KFD_GC_VERSION(dev) == IP_VERSION(9, 4, 0);
+}
 
-bool kfd_is_locked(void);
+int kfd_send_exception_to_runtime(struct kfd_process *p,
+				unsigned int queue_id,
+				uint64_t error_reason);
+bool kfd_is_locked(struct kfd_dev *kfd);
 
 /* Compute profile */
-void kfd_inc_compute_active(struct kfd_dev *dev);
-void kfd_dec_compute_active(struct kfd_dev *dev);
+void kfd_inc_compute_active(struct kfd_node *dev);
+void kfd_dec_compute_active(struct kfd_node *dev);
 
 /* Cgroup Support */
 /* Check with device cgroup if @kfd device is accessible */
-static inline int kfd_devcgroup_check_permission(struct kfd_dev *kfd)
+static inline int kfd_devcgroup_check_permission(struct kfd_node *node)
 {
 #if defined(CONFIG_CGROUP_DEVICE) || defined(CONFIG_CGROUP_BPF)
-	struct drm_device *ddev = kfd->ddev;
+	struct drm_device *ddev;
+
+	if (node->xcp)
+		ddev = node->xcp->ddev;
+	else
+		ddev = adev_to_drm(node->adev);
 
 	return devcgroup_check_permission(DEVCG_DEV_CHAR, DRM_MAJOR,
 					  ddev->render->index,
@@ -1198,6 +1574,11 @@ static inline int kfd_devcgroup_check_permission(struct kfd_dev *kfd)
 #endif
 }
 
+static inline bool kfd_is_first_node(struct kfd_node *node)
+{
+	return (node == node->kfd->nodes[0]);
+}
+
 /* Debugfs */
 #if defined(CONFIG_DEBUG_FS)
 
@@ -1210,14 +1591,19 @@ int dqm_debugfs_hqds(struct seq_file *m, void *data);
 int kfd_debugfs_rls_by_device(struct seq_file *m, void *data);
 int pm_debugfs_runlist(struct seq_file *m, void *data);
 
-int kfd_debugfs_hang_hws(struct kfd_dev *dev);
+int kfd_debugfs_hang_hws(struct kfd_node *dev);
 int pm_debugfs_hang_hws(struct packet_manager *pm);
 int dqm_debugfs_hang_hws(struct device_queue_manager *dqm);
 
+void kfd_debugfs_add_process(struct kfd_process *p);
+void kfd_debugfs_remove_process(struct kfd_process *p);
+
 #else
 
 static inline void kfd_debugfs_init(void) {}
 static inline void kfd_debugfs_fini(void) {}
+static inline void kfd_debugfs_add_process(struct kfd_process *p) {}
+static inline void kfd_debugfs_remove_process(struct kfd_process *p) {}
 
 #endif
 
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_process.c b/drivers/gpu/drm/amd/amdkfd/kfd_process.c
index b993011cfa64..a085faac9fe1 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_process.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_process.c
@@ -1,5 +1,6 @@
+// SPDX-License-Identifier: GPL-2.0 OR MIT
 /*
- * Copyright 2014 Advanced Micro Devices, Inc.
+ * Copyright 2014-2022 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -27,7 +28,6 @@
 #include <linux/sched/task.h>
 #include <linux/mmu_context.h>
 #include <linux/slab.h>
-#include <linux/amd-iommu.h>
 #include <linux/notifier.h>
 #include <linux/compat.h>
 #include <linux/mman.h>
@@ -35,21 +35,22 @@
 #include <linux/pm_runtime.h>
 #include "amdgpu_amdkfd.h"
 #include "amdgpu.h"
+#include "amdgpu_reset.h"
 
 struct mm_struct;
 
 #include "kfd_priv.h"
 #include "kfd_device_queue_manager.h"
-#include "kfd_dbgmgr.h"
-#include "kfd_iommu.h"
 #include "kfd_svm.h"
+#include "kfd_smi_events.h"
+#include "kfd_debug.h"
 
 /*
  * List of struct kfd_process (field kfd_process).
  * Unique/indexed by mm_struct*
  */
 DEFINE_HASHTABLE(kfd_processes_table, KFD_PROCESS_TABLE_SIZE);
-static DEFINE_MUTEX(kfd_processes_mutex);
+DEFINE_MUTEX(kfd_processes_mutex);
 
 DEFINE_SRCU(kfd_processes_srcu);
 
@@ -64,10 +65,10 @@ static struct workqueue_struct *kfd_process_wq;
  */
 static struct workqueue_struct *kfd_restore_wq;
 
-static struct kfd_process *find_process(const struct task_struct *thread);
+static struct kfd_process *find_process(const struct task_struct *thread,
+					bool ref);
 static void kfd_process_ref_release(struct kref *ref);
 static struct kfd_process *create_process(const struct task_struct *thread);
-static int kfd_process_init_cwsr_apu(struct kfd_process *p, struct file *filep);
 
 static void evict_process_worker(struct work_struct *work);
 static void restore_process_worker(struct work_struct *work);
@@ -251,14 +252,13 @@ cleanup:
 }
 
 /**
- * @kfd_get_cu_occupancy - Collect number of waves in-flight on this device
+ * kfd_get_cu_occupancy - Collect number of waves in-flight on this device
  * by current process. Translates acquired wave count into number of compute units
  * that are occupied.
  *
- * @atr: Handle of attribute that allows reporting of wave count. The attribute
+ * @attr: Handle of attribute that allows reporting of wave count. The attribute
  * handle encapsulates GPU device it is associated with, thereby allowing collection
  * of waves in flight, etc
- *
  * @buffer: Handle of user provided buffer updated with wave count
  *
  * Return: Number of bytes written to user buffer or an error value
@@ -268,9 +268,12 @@ static int kfd_get_cu_occupancy(struct attribute *attr, char *buffer)
 	int cu_cnt;
 	int wave_cnt;
 	int max_waves_per_cu;
-	struct kfd_dev *dev = NULL;
+	struct kfd_node *dev = NULL;
 	struct kfd_process *proc = NULL;
 	struct kfd_process_device *pdd = NULL;
+	int i;
+	struct kfd_cu_occupancy *cu_occupancy;
+	u32 queue_format;
 
 	pdd = container_of(attr, struct kfd_process_device, attr_cu_occupancy);
 	dev = pdd->dev;
@@ -280,41 +283,64 @@ static int kfd_get_cu_occupancy(struct attribute *attr, char *buffer)
 	cu_cnt = 0;
 	proc = pdd->process;
 	if (pdd->qpd.queue_count == 0) {
-		pr_debug("Gpu-Id: %d has no active queues for process %d\n",
-			 dev->id, proc->pasid);
+		pr_debug("Gpu-Id: %d has no active queues for process pid %d\n",
+			 dev->id, (int)proc->lead_thread->pid);
 		return snprintf(buffer, PAGE_SIZE, "%d\n", cu_cnt);
 	}
 
 	/* Collect wave count from device if it supports */
 	wave_cnt = 0;
 	max_waves_per_cu = 0;
-	dev->kfd2kgd->get_cu_occupancy(dev->kgd, proc->pasid, &wave_cnt,
-			&max_waves_per_cu);
+
+	cu_occupancy = kcalloc(AMDGPU_MAX_QUEUES, sizeof(*cu_occupancy), GFP_KERNEL);
+	if (!cu_occupancy)
+		return -ENOMEM;
+
+	/*
+	 * For GFX 9.4.3, fetch the CU occupancy from the first XCC in the partition.
+	 * For AQL queues, because of cooperative dispatch we multiply the wave count
+	 * by number of XCCs in the partition to get the total wave counts across all
+	 * XCCs in the partition.
+	 * For PM4 queues, there is no cooperative dispatch so wave_cnt stay as it is.
+	 */
+	dev->kfd2kgd->get_cu_occupancy(dev->adev, cu_occupancy,
+			&max_waves_per_cu, ffs(dev->xcc_mask) - 1);
+
+	for (i = 0; i < AMDGPU_MAX_QUEUES; i++) {
+		if (cu_occupancy[i].wave_cnt != 0 &&
+		    kfd_dqm_is_queue_in_process(dev->dqm, &pdd->qpd,
+						cu_occupancy[i].doorbell_off,
+						&queue_format)) {
+			if (unlikely(queue_format == KFD_QUEUE_FORMAT_PM4))
+				wave_cnt += cu_occupancy[i].wave_cnt;
+			else
+				wave_cnt += (NUM_XCC(dev->xcc_mask) *
+						cu_occupancy[i].wave_cnt);
+		}
+	}
 
 	/* Translate wave count to number of compute units */
 	cu_cnt = (wave_cnt + (max_waves_per_cu - 1)) / max_waves_per_cu;
+	kfree(cu_occupancy);
 	return snprintf(buffer, PAGE_SIZE, "%d\n", cu_cnt);
 }
 
 static ssize_t kfd_procfs_show(struct kobject *kobj, struct attribute *attr,
 			       char *buffer)
 {
-	if (strcmp(attr->name, "pasid") == 0) {
-		struct kfd_process *p = container_of(attr, struct kfd_process,
-						     attr_pasid);
-
-		return snprintf(buffer, PAGE_SIZE, "%d\n", p->pasid);
-	} else if (strncmp(attr->name, "vram_", 5) == 0) {
+	if (strcmp(attr->name, "pasid") == 0)
+		return snprintf(buffer, PAGE_SIZE, "%d\n", 0);
+	else if (strncmp(attr->name, "vram_", 5) == 0) {
 		struct kfd_process_device *pdd = container_of(attr, struct kfd_process_device,
 							      attr_vram);
-		return snprintf(buffer, PAGE_SIZE, "%llu\n", READ_ONCE(pdd->vram_usage));
+		return snprintf(buffer, PAGE_SIZE, "%llu\n", atomic64_read(&pdd->vram_usage));
 	} else if (strncmp(attr->name, "sdma_", 5) == 0) {
 		struct kfd_process_device *pdd = container_of(attr, struct kfd_process_device,
 							      attr_sdma);
 		struct kfd_sdma_activity_handler_workarea sdma_activity_work_handler;
 
-		INIT_WORK(&sdma_activity_work_handler.sdma_activity_work,
-					kfd_sdma_activity_worker);
+		INIT_WORK_ONSTACK(&sdma_activity_work_handler.sdma_activity_work,
+				  kfd_sdma_activity_worker);
 
 		sdma_activity_work_handler.pdd = pdd;
 		sdma_activity_work_handler.sdma_activity_counter = 0;
@@ -322,6 +348,7 @@ static ssize_t kfd_procfs_show(struct kobject *kobj, struct attribute *attr,
 		schedule_work(&sdma_activity_work_handler.sdma_activity_work);
 
 		flush_work(&sdma_activity_work_handler.sdma_activity_work);
+		destroy_work_on_stack(&sdma_activity_work_handler.sdma_activity_work);
 
 		return snprintf(buffer, PAGE_SIZE, "%llu\n",
 				(sdma_activity_work_handler.sdma_activity_counter)/
@@ -343,7 +370,7 @@ static const struct sysfs_ops kfd_procfs_ops = {
 	.show = kfd_procfs_show,
 };
 
-static struct kobj_type procfs_type = {
+static const struct kobj_type procfs_type = {
 	.release = kfd_procfs_kobj_release,
 	.sysfs_ops = &kfd_procfs_ops,
 };
@@ -462,21 +489,22 @@ static struct attribute *procfs_queue_attrs[] = {
 	&attr_queue_gpuid,
 	NULL
 };
+ATTRIBUTE_GROUPS(procfs_queue);
 
 static const struct sysfs_ops procfs_queue_ops = {
 	.show = kfd_procfs_queue_show,
 };
 
-static struct kobj_type procfs_queue_type = {
+static const struct kobj_type procfs_queue_type = {
 	.sysfs_ops = &procfs_queue_ops,
-	.default_attrs = procfs_queue_attrs,
+	.default_groups = procfs_queue_groups,
 };
 
 static const struct sysfs_ops procfs_stats_ops = {
 	.show = kfd_procfs_stats_show,
 };
 
-static struct kobj_type procfs_stats_type = {
+static const struct kobj_type procfs_stats_type = {
 	.sysfs_ops = &procfs_stats_ops,
 	.release = kfd_procfs_kobj_release,
 };
@@ -485,7 +513,7 @@ static const struct sysfs_ops sysfs_counters_ops = {
 	.show = kfd_sysfs_counters_show,
 };
 
-static struct kobj_type sysfs_counters_type = {
+static const struct kobj_type sysfs_counters_type = {
 	.sysfs_ops = &sysfs_counters_ops,
 	.release = kfd_procfs_kobj_release,
 };
@@ -664,7 +692,8 @@ int kfd_process_create_wq(void)
 	if (!kfd_process_wq)
 		kfd_process_wq = alloc_workqueue("kfd_process_wq", 0, 0);
 	if (!kfd_restore_wq)
-		kfd_restore_wq = alloc_ordered_workqueue("kfd_restore_wq", 0);
+		kfd_restore_wq = alloc_ordered_workqueue("kfd_restore_wq",
+							 WQ_FREEZABLE);
 
 	if (!kfd_process_wq || !kfd_restore_wq) {
 		kfd_process_destroy_wq();
@@ -687,17 +716,17 @@ void kfd_process_destroy_wq(void)
 }
 
 static void kfd_process_free_gpuvm(struct kgd_mem *mem,
-			struct kfd_process_device *pdd, void *kptr)
+			struct kfd_process_device *pdd, void **kptr)
 {
-	struct kfd_dev *dev = pdd->dev;
+	struct kfd_node *dev = pdd->dev;
 
-	if (kptr) {
-		amdgpu_amdkfd_gpuvm_unmap_gtt_bo_from_kernel(dev->kgd, mem);
-		kptr = NULL;
+	if (kptr && *kptr) {
+		amdgpu_amdkfd_gpuvm_unmap_gtt_bo_from_kernel(mem);
+		*kptr = NULL;
 	}
 
-	amdgpu_amdkfd_gpuvm_unmap_memory_from_gpu(dev->kgd, mem, pdd->drm_priv);
-	amdgpu_amdkfd_gpuvm_free_memory_of_gpu(dev->kgd, mem, pdd->drm_priv,
+	amdgpu_amdkfd_gpuvm_unmap_memory_from_gpu(dev->adev, mem, pdd->drm_priv);
+	amdgpu_amdkfd_gpuvm_free_memory_of_gpu(dev->adev, mem, pdd->drm_priv,
 					       NULL);
 }
 
@@ -711,27 +740,28 @@ static int kfd_process_alloc_gpuvm(struct kfd_process_device *pdd,
 				   uint64_t gpu_va, uint32_t size,
 				   uint32_t flags, struct kgd_mem **mem, void **kptr)
 {
-	struct kfd_dev *kdev = pdd->dev;
+	struct kfd_node *kdev = pdd->dev;
 	int err;
 
-	err = amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu(kdev->kgd, gpu_va, size,
-						 pdd->drm_priv, mem, NULL, flags);
+	err = amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu(kdev->adev, gpu_va, size,
+						 pdd->drm_priv, mem, NULL,
+						 flags, false);
 	if (err)
 		goto err_alloc_mem;
 
-	err = amdgpu_amdkfd_gpuvm_map_memory_to_gpu(kdev->kgd, *mem,
-			pdd->drm_priv, NULL);
+	err = amdgpu_amdkfd_gpuvm_map_memory_to_gpu(kdev->adev, *mem,
+			pdd->drm_priv);
 	if (err)
 		goto err_map_mem;
 
-	err = amdgpu_amdkfd_gpuvm_sync_memory(kdev->kgd, *mem, true);
+	err = amdgpu_amdkfd_gpuvm_sync_memory(kdev->adev, *mem, true);
 	if (err) {
 		pr_debug("Sync memory failed, wait interrupted by user signal\n");
 		goto sync_memory_failed;
 	}
 
 	if (kptr) {
-		err = amdgpu_amdkfd_gpuvm_map_gtt_bo_to_kernel(kdev->kgd,
+		err = amdgpu_amdkfd_gpuvm_map_gtt_bo_to_kernel(
 				(struct kgd_mem *)*mem, kptr, NULL);
 		if (err) {
 			pr_debug("Map GTT BO to kernel failed\n");
@@ -742,10 +772,10 @@ static int kfd_process_alloc_gpuvm(struct kfd_process_device *pdd,
 	return err;
 
 sync_memory_failed:
-	amdgpu_amdkfd_gpuvm_unmap_memory_from_gpu(kdev->kgd, *mem, pdd->drm_priv);
+	amdgpu_amdkfd_gpuvm_unmap_memory_from_gpu(kdev->adev, *mem, pdd->drm_priv);
 
 err_map_mem:
-	amdgpu_amdkfd_gpuvm_free_memory_of_gpu(kdev->kgd, *mem, pdd->drm_priv,
+	amdgpu_amdkfd_gpuvm_free_memory_of_gpu(kdev->adev, *mem, pdd->drm_priv,
 					       NULL);
 err_alloc_mem:
 	*mem = NULL;
@@ -792,21 +822,30 @@ static void kfd_process_device_destroy_ib_mem(struct kfd_process_device *pdd)
 	if (!qpd->ib_kaddr || !qpd->ib_base)
 		return;
 
-	kfd_process_free_gpuvm(qpd->ib_mem, pdd, qpd->ib_kaddr);
+	kfd_process_free_gpuvm(qpd->ib_mem, pdd, &qpd->ib_kaddr);
 }
 
-struct kfd_process *kfd_create_process(struct file *filep)
+struct kfd_process *kfd_create_process(struct task_struct *thread)
 {
 	struct kfd_process *process;
-	struct task_struct *thread = current;
 	int ret;
 
-	if (!thread->mm)
+	if (!(thread->mm && mmget_not_zero(thread->mm)))
 		return ERR_PTR(-EINVAL);
 
 	/* Only the pthreads threading model is supported. */
-	if (thread->group_leader->mm != thread->mm)
+	if (thread->group_leader->mm != thread->mm) {
+		mmput(thread->mm);
 		return ERR_PTR(-EINVAL);
+	}
+
+	/* If the process just called exec(3), it is possible that the
+	 * cleanup of the kfd_process (following the release of the mm
+	 * of the old process image) is still in the cleanup work queue.
+	 * Make sure to drain any job before trying to recreate any
+	 * resource for this process.
+	 */
+	flush_workqueue(kfd_process_wq);
 
 	/*
 	 * take kfd processes mutex before starting of process creation
@@ -815,8 +854,16 @@ struct kfd_process *kfd_create_process(struct file *filep)
 	 */
 	mutex_lock(&kfd_processes_mutex);
 
-	/* A prior open of /dev/kfd could have already created the process. */
-	process = find_process(thread);
+	if (kfd_is_locked(NULL)) {
+		pr_debug("KFD is locked! Cannot create process");
+		process = ERR_PTR(-EINVAL);
+		goto out;
+	}
+
+	/* A prior open of /dev/kfd could have already created the process.
+	 * find_process will increase process kref in this case
+	 */
+	process = find_process(thread, true);
 	if (process) {
 		pr_debug("Process already found\n");
 	} else {
@@ -824,10 +871,6 @@ struct kfd_process *kfd_create_process(struct file *filep)
 		if (IS_ERR(process))
 			goto out;
 
-		ret = kfd_process_init_cwsr_apu(process, filep);
-		if (ret)
-			goto out_destroy;
-
 		if (!procfs.kobj)
 			goto out;
 
@@ -856,21 +899,16 @@ struct kfd_process *kfd_create_process(struct file *filep)
 		kfd_procfs_add_sysfs_stats(process);
 		kfd_procfs_add_sysfs_files(process);
 		kfd_procfs_add_sysfs_counters(process);
+
+		kfd_debugfs_add_process(process);
+
+		init_waitqueue_head(&process->wait_irq_drain);
 	}
 out:
-	if (!IS_ERR(process))
-		kref_get(&process->ref);
 	mutex_unlock(&kfd_processes_mutex);
+	mmput(thread->mm);
 
 	return process;
-
-out_destroy:
-	hash_del_rcu(&process->kfd_processes);
-	mutex_unlock(&kfd_processes_mutex);
-	synchronize_srcu(&kfd_processes_srcu);
-	/* kfd_process_free_notifier will trigger the cleanup */
-	mmu_notifier_put(&process->mmu_notifier);
-	return ERR_PTR(ret);
 }
 
 struct kfd_process *kfd_get_process(const struct task_struct *thread)
@@ -884,7 +922,7 @@ struct kfd_process *kfd_get_process(const struct task_struct *thread)
 	if (thread->group_leader->mm != thread->mm)
 		return ERR_PTR(-EINVAL);
 
-	process = find_process(thread);
+	process = find_process(thread, false);
 	if (!process)
 		return ERR_PTR(-EINVAL);
 
@@ -903,13 +941,16 @@ static struct kfd_process *find_process_by_mm(const struct mm_struct *mm)
 	return NULL;
 }
 
-static struct kfd_process *find_process(const struct task_struct *thread)
+static struct kfd_process *find_process(const struct task_struct *thread,
+					bool ref)
 {
 	struct kfd_process *p;
 	int idx;
 
 	idx = srcu_read_lock(&kfd_processes_srcu);
 	p = find_process_by_mm(thread->mm);
+	if (p && ref)
+		kref_get(&p->ref);
 	srcu_read_unlock(&kfd_processes_srcu, idx);
 
 	return p;
@@ -920,6 +961,26 @@ void kfd_unref_process(struct kfd_process *p)
 	kref_put(&p->ref, kfd_process_ref_release);
 }
 
+/* This increments the process->ref counter. */
+struct kfd_process *kfd_lookup_process_by_pid(struct pid *pid)
+{
+	struct task_struct *task = NULL;
+	struct kfd_process *p    = NULL;
+
+	if (!pid) {
+		task = current;
+		get_task_struct(task);
+	} else {
+		task = get_pid_task(pid, PIDTYPE_PID);
+	}
+
+	if (task) {
+		p = find_process(task, true);
+		put_task_struct(task);
+	}
+
+	return p;
+}
 
 static void kfd_process_device_free_bos(struct kfd_process_device *pdd)
 {
@@ -940,10 +1001,10 @@ static void kfd_process_device_free_bos(struct kfd_process_device *pdd)
 			if (!peer_pdd->drm_priv)
 				continue;
 			amdgpu_amdkfd_gpuvm_unmap_memory_from_gpu(
-				peer_pdd->dev->kgd, mem, peer_pdd->drm_priv);
+				peer_pdd->dev->adev, mem, peer_pdd->drm_priv);
 		}
 
-		amdgpu_amdkfd_gpuvm_free_memory_of_gpu(pdd->dev->kgd, mem,
+		amdgpu_amdkfd_gpuvm_free_memory_of_gpu(pdd->dev->adev, mem,
 						       pdd->drm_priv, NULL);
 		kfd_process_device_remove_obj_handle(pdd, id);
 	}
@@ -956,7 +1017,7 @@ static void kfd_process_device_free_bos(struct kfd_process_device *pdd)
 static void kfd_process_kunmap_signal_bo(struct kfd_process *p)
 {
 	struct kfd_process_device *pdd;
-	struct kfd_dev *kdev;
+	struct kfd_node *kdev;
 	void *mem;
 
 	kdev = kfd_device_by_id(GET_GPU_ID(p->signal_handle));
@@ -974,7 +1035,7 @@ static void kfd_process_kunmap_signal_bo(struct kfd_process *p)
 	if (!mem)
 		goto out;
 
-	amdgpu_amdkfd_gpuvm_unmap_gtt_bo_from_kernel(kdev->kgd, mem);
+	amdgpu_amdkfd_gpuvm_unmap_gtt_bo_from_kernel(mem);
 
 out:
 	mutex_unlock(&p->mutex);
@@ -995,37 +1056,39 @@ static void kfd_process_destroy_pdds(struct kfd_process *p)
 	for (i = 0; i < p->n_pdds; i++) {
 		struct kfd_process_device *pdd = p->pdds[i];
 
-		pr_debug("Releasing pdd (topology id %d) for process (pasid 0x%x)\n",
-				pdd->dev->id, p->pasid);
+		kfd_smi_event_process(pdd, false);
 
+		pr_debug("Releasing pdd (topology id %d, for pid %d)\n",
+			pdd->dev->id, p->lead_thread->pid);
 		kfd_process_device_destroy_cwsr_dgpu(pdd);
 		kfd_process_device_destroy_ib_mem(pdd);
 
-		if (pdd->drm_file) {
-			amdgpu_amdkfd_gpuvm_release_process_vm(
-					pdd->dev->kgd, pdd->drm_priv);
+		if (pdd->drm_file)
 			fput(pdd->drm_file);
-		}
 
 		if (pdd->qpd.cwsr_kaddr && !pdd->qpd.cwsr_base)
 			free_pages((unsigned long)pdd->qpd.cwsr_kaddr,
 				get_order(KFD_CWSR_TBA_TMA_SIZE));
 
-		kfree(pdd->qpd.doorbell_bitmap);
 		idr_destroy(&pdd->alloc_idr);
 
-		kfd_free_process_doorbells(pdd->dev, pdd->doorbell_index);
+		kfd_free_process_doorbells(pdd->dev->kfd, pdd);
 
+		if (pdd->dev->kfd->shared_resources.enable_mes &&
+			pdd->proc_ctx_cpu_ptr)
+			amdgpu_amdkfd_free_gtt_mem(pdd->dev->adev,
+						   &pdd->proc_ctx_bo);
 		/*
 		 * before destroying pdd, make sure to report availability
 		 * for auto suspend
 		 */
 		if (pdd->runtime_inuse) {
-			pm_runtime_mark_last_busy(pdd->dev->ddev->dev);
-			pm_runtime_put_autosuspend(pdd->dev->ddev->dev);
+			pm_runtime_put_autosuspend(adev_to_drm(pdd->dev->adev)->dev);
 			pdd->runtime_inuse = false;
 		}
 
+		atomic_dec(&pdd->dev->kfd->kfd_processes_count);
+
 		kfree(pdd);
 		p->pdds[i] = NULL;
 	}
@@ -1076,6 +1139,17 @@ static void kfd_process_remove_sysfs(struct kfd_process *p)
 	p->kobj = NULL;
 }
 
+/*
+ * If any GPU is ongoing reset, wait for reset complete.
+ */
+static void kfd_process_wait_gpu_reset_complete(struct kfd_process *p)
+{
+	int i;
+
+	for (i = 0; i < p->n_pdds; i++)
+		flush_workqueue(p->pdds[i]->dev->adev->reset_domain->wq);
+}
+
 /* No process locking is needed in this function, because the process
  * is not findable any more. We must assume that no other thread is
  * using it any more, otherwise we couldn't safely free the process
@@ -1085,20 +1159,34 @@ static void kfd_process_wq_release(struct work_struct *work)
 {
 	struct kfd_process *p = container_of(work, struct kfd_process,
 					     release_work);
+	struct dma_fence *ef;
+
+	/*
+	 * If GPU in reset, user queues may still running, wait for reset complete.
+	 */
+	kfd_process_wait_gpu_reset_complete(p);
+
+	/* Signal the eviction fence after user mode queues are
+	 * destroyed. This allows any BOs to be freed without
+	 * triggering pointless evictions or waiting for fences.
+	 */
+	synchronize_rcu();
+	ef = rcu_access_pointer(p->ef);
+	if (ef)
+		dma_fence_signal(ef);
 
 	kfd_process_remove_sysfs(p);
-	kfd_iommu_unbind_process(p);
+	kfd_debugfs_remove_process(p);
 
 	kfd_process_kunmap_signal_bo(p);
 	kfd_process_free_outstanding_kfd_bos(p);
 	svm_range_list_fini(p);
 
 	kfd_process_destroy_pdds(p);
-	dma_fence_put(p->ef);
+	dma_fence_put(ef);
 
 	kfd_event_free_process(p);
 
-	kfd_pasid_free(p->pasid);
 	mutex_destroy(&p->mutex);
 
 	put_task_struct(p->lead_thread);
@@ -1116,10 +1204,8 @@ static void kfd_process_ref_release(struct kref *ref)
 
 static struct mmu_notifier *kfd_process_alloc_notifier(struct mm_struct *mm)
 {
-	int idx = srcu_read_lock(&kfd_processes_srcu);
-	struct kfd_process *p = find_process_by_mm(mm);
-
-	srcu_read_unlock(&kfd_processes_srcu, idx);
+	/* This increments p->ref counter if kfd process p exists */
+	struct kfd_process *p = kfd_lookup_process_by_mm(mm);
 
 	return p ? &p->mmu_notifier : ERR_PTR(-ESRCH);
 }
@@ -1129,11 +1215,58 @@ static void kfd_process_free_notifier(struct mmu_notifier *mn)
 	kfd_unref_process(container_of(mn, struct kfd_process, mmu_notifier));
 }
 
+static void kfd_process_notifier_release_internal(struct kfd_process *p)
+{
+	int i;
+
+	cancel_delayed_work_sync(&p->eviction_work);
+	cancel_delayed_work_sync(&p->restore_work);
+
+	/*
+	 * Dequeue and destroy user queues, it is not safe for GPU to access
+	 * system memory after mmu release notifier callback returns because
+	 * exit_mmap free process memory afterwards.
+	 */
+	kfd_process_dequeue_from_all_devices(p);
+	pqm_uninit(&p->pqm);
+
+	for (i = 0; i < p->n_pdds; i++) {
+		struct kfd_process_device *pdd = p->pdds[i];
+
+		/* re-enable GFX OFF since runtime enable with ttmp setup disabled it. */
+		if (!kfd_dbg_is_rlc_restore_supported(pdd->dev) && p->runtime_info.ttmp_setup)
+			amdgpu_gfx_off_ctrl(pdd->dev->adev, true);
+	}
+
+	/* Indicate to other users that MM is no longer valid */
+	p->mm = NULL;
+	kfd_dbg_trap_disable(p);
+
+	if (atomic_read(&p->debugged_process_count) > 0) {
+		struct kfd_process *target;
+		unsigned int temp;
+		int idx = srcu_read_lock(&kfd_processes_srcu);
+
+		hash_for_each_rcu(kfd_processes_table, temp, target, kfd_processes) {
+			if (target->debugger_process && target->debugger_process == p) {
+				mutex_lock_nested(&target->mutex, 1);
+				kfd_dbg_trap_disable(target);
+				mutex_unlock(&target->mutex);
+				if (atomic_read(&p->debugged_process_count) == 0)
+					break;
+			}
+		}
+
+		srcu_read_unlock(&kfd_processes_srcu, idx);
+	}
+
+	mmu_notifier_put(&p->mmu_notifier);
+}
+
 static void kfd_process_notifier_release(struct mmu_notifier *mn,
 					struct mm_struct *mm)
 {
 	struct kfd_process *p;
-	int i;
 
 	/*
 	 * The kfd_process structure can not be free because the
@@ -1144,65 +1277,80 @@ static void kfd_process_notifier_release(struct mmu_notifier *mn,
 		return;
 
 	mutex_lock(&kfd_processes_mutex);
+	/*
+	 * Do early return if table is empty.
+	 *
+	 * This could potentially happen if this function is called concurrently
+	 * by mmu_notifier and by kfd_cleanup_pocesses.
+	 *
+	 */
+	if (hash_empty(kfd_processes_table)) {
+		mutex_unlock(&kfd_processes_mutex);
+		return;
+	}
 	hash_del_rcu(&p->kfd_processes);
 	mutex_unlock(&kfd_processes_mutex);
 	synchronize_srcu(&kfd_processes_srcu);
 
-	cancel_delayed_work_sync(&p->eviction_work);
-	cancel_delayed_work_sync(&p->restore_work);
-	cancel_delayed_work_sync(&p->svms.restore_work);
+	kfd_process_notifier_release_internal(p);
+}
 
-	mutex_lock(&p->mutex);
+static const struct mmu_notifier_ops kfd_process_mmu_notifier_ops = {
+	.release = kfd_process_notifier_release,
+	.alloc_notifier = kfd_process_alloc_notifier,
+	.free_notifier = kfd_process_free_notifier,
+};
 
-	/* Iterate over all process device data structures and if the
-	 * pdd is in debug mode, we should first force unregistration,
-	 * then we will be able to destroy the queues
-	 */
-	for (i = 0; i < p->n_pdds; i++) {
-		struct kfd_dev *dev = p->pdds[i]->dev;
+/*
+ * This code handles the case when driver is being unloaded before all
+ * mm_struct are released.  We need to safely free the kfd_process and
+ * avoid race conditions with mmu_notifier that might try to free them.
+ *
+ */
+void kfd_cleanup_processes(void)
+{
+	struct kfd_process *p;
+	struct hlist_node *p_temp;
+	unsigned int temp;
+	HLIST_HEAD(cleanup_list);
 
-		mutex_lock(kfd_get_dbgmgr_mutex());
-		if (dev && dev->dbgmgr && dev->dbgmgr->pasid == p->pasid) {
-			if (!kfd_dbgmgr_unregister(dev->dbgmgr, p)) {
-				kfd_dbgmgr_destroy(dev->dbgmgr);
-				dev->dbgmgr = NULL;
-			}
-		}
-		mutex_unlock(kfd_get_dbgmgr_mutex());
+	/*
+	 * Move all remaining kfd_process from the process table to a
+	 * temp list for processing.   Once done, callback from mmu_notifier
+	 * release will not see the kfd_process in the table and do early return,
+	 * avoiding double free issues.
+	 */
+	mutex_lock(&kfd_processes_mutex);
+	hash_for_each_safe(kfd_processes_table, temp, p_temp, p, kfd_processes) {
+		hash_del_rcu(&p->kfd_processes);
+		synchronize_srcu(&kfd_processes_srcu);
+		hlist_add_head(&p->kfd_processes, &cleanup_list);
 	}
+	mutex_unlock(&kfd_processes_mutex);
 
-	kfd_process_dequeue_from_all_devices(p);
-	pqm_uninit(&p->pqm);
+	hlist_for_each_entry_safe(p, p_temp, &cleanup_list, kfd_processes)
+		kfd_process_notifier_release_internal(p);
 
-	/* Indicate to other users that MM is no longer valid */
-	p->mm = NULL;
-	/* Signal the eviction fence after user mode queues are
-	 * destroyed. This allows any BOs to be freed without
-	 * triggering pointless evictions or waiting for fences.
+	/*
+	 * Ensures that all outstanding free_notifier get called, triggering
+	 * the release of the kfd_process struct.
 	 */
-	dma_fence_signal(p->ef);
-
-	mutex_unlock(&p->mutex);
-
-	mmu_notifier_put(&p->mmu_notifier);
+	mmu_notifier_synchronize();
 }
 
-static const struct mmu_notifier_ops kfd_process_mmu_notifier_ops = {
-	.release = kfd_process_notifier_release,
-	.alloc_notifier = kfd_process_alloc_notifier,
-	.free_notifier = kfd_process_free_notifier,
-};
-
-static int kfd_process_init_cwsr_apu(struct kfd_process *p, struct file *filep)
+int kfd_process_init_cwsr_apu(struct kfd_process *p, struct file *filep)
 {
 	unsigned long  offset;
 	int i;
 
+	if (p->has_cwsr)
+		return 0;
+
 	for (i = 0; i < p->n_pdds; i++) {
-		struct kfd_dev *dev = p->pdds[i]->dev;
+		struct kfd_node *dev = p->pdds[i]->dev;
 		struct qcm_process_device *qpd = &p->pdds[i]->qpd;
 
-		if (!dev->cwsr_enabled || qpd->cwsr_kaddr || qpd->cwsr_base)
+		if (!dev->kfd->cwsr_enabled || qpd->cwsr_kaddr || qpd->cwsr_base)
 			continue;
 
 		offset = KFD_MMAP_TYPE_RESERVED_MEM | KFD_MMAP_GPU_ID(dev->id);
@@ -1213,25 +1361,30 @@ static int kfd_process_init_cwsr_apu(struct kfd_process *p, struct file *filep)
 		if (IS_ERR_VALUE(qpd->tba_addr)) {
 			int err = qpd->tba_addr;
 
-			pr_err("Failure to set tba address. error %d.\n", err);
+			dev_err(dev->adev->dev,
+				"Failure to set tba address. error %d.\n", err);
 			qpd->tba_addr = 0;
 			qpd->cwsr_kaddr = NULL;
 			return err;
 		}
 
-		memcpy(qpd->cwsr_kaddr, dev->cwsr_isa, dev->cwsr_isa_size);
+		memcpy(qpd->cwsr_kaddr, dev->kfd->cwsr_isa, dev->kfd->cwsr_isa_size);
+
+		kfd_process_set_trap_debug_flag(qpd, p->debug_trap_enabled);
 
 		qpd->tma_addr = qpd->tba_addr + KFD_CWSR_TMA_OFFSET;
 		pr_debug("set tba :0x%llx, tma:0x%llx, cwsr_kaddr:%p for pqm.\n",
 			qpd->tba_addr, qpd->tma_addr, qpd->cwsr_kaddr);
 	}
 
+	p->has_cwsr = true;
+
 	return 0;
 }
 
 static int kfd_process_device_init_cwsr_dgpu(struct kfd_process_device *pdd)
 {
-	struct kfd_dev *dev = pdd->dev;
+	struct kfd_node *dev = pdd->dev;
 	struct qcm_process_device *qpd = &pdd->qpd;
 	uint32_t flags = KFD_IOC_ALLOC_MEM_FLAGS_GTT
 			| KFD_IOC_ALLOC_MEM_FLAGS_NO_SUBSTITUTE
@@ -1240,7 +1393,7 @@ static int kfd_process_device_init_cwsr_dgpu(struct kfd_process_device *pdd)
 	void *kaddr;
 	int ret;
 
-	if (!dev->cwsr_enabled || qpd->cwsr_kaddr || !qpd->cwsr_base)
+	if (!dev->kfd->cwsr_enabled || qpd->cwsr_kaddr || !qpd->cwsr_base)
 		return 0;
 
 	/* cwsr_base is only set for dGPU */
@@ -1253,7 +1406,10 @@ static int kfd_process_device_init_cwsr_dgpu(struct kfd_process_device *pdd)
 	qpd->cwsr_kaddr = kaddr;
 	qpd->tba_addr = qpd->cwsr_base;
 
-	memcpy(qpd->cwsr_kaddr, dev->cwsr_isa, dev->cwsr_isa_size);
+	memcpy(qpd->cwsr_kaddr, dev->kfd->cwsr_isa, dev->kfd->cwsr_isa_size);
+
+	kfd_process_set_trap_debug_flag(&pdd->qpd,
+					pdd->process->debug_trap_enabled);
 
 	qpd->tma_addr = qpd->tba_addr + KFD_CWSR_TMA_OFFSET;
 	pr_debug("set tba :0x%llx, tma:0x%llx, cwsr_kaddr:%p for pqm.\n",
@@ -1264,13 +1420,13 @@ static int kfd_process_device_init_cwsr_dgpu(struct kfd_process_device *pdd)
 
 static void kfd_process_device_destroy_cwsr_dgpu(struct kfd_process_device *pdd)
 {
-	struct kfd_dev *dev = pdd->dev;
+	struct kfd_node *dev = pdd->dev;
 	struct qcm_process_device *qpd = &pdd->qpd;
 
-	if (!dev->cwsr_enabled || !qpd->cwsr_kaddr || !qpd->cwsr_base)
+	if (!dev->kfd->cwsr_enabled || !qpd->cwsr_kaddr || !qpd->cwsr_base)
 		return;
 
-	kfd_process_free_gpuvm(qpd->cwsr_mem, pdd, qpd->cwsr_kaddr);
+	kfd_process_free_gpuvm(qpd->cwsr_mem, pdd, &qpd->cwsr_kaddr);
 }
 
 void kfd_process_set_trap_handler(struct qcm_process_device *qpd,
@@ -1311,37 +1467,51 @@ bool kfd_process_xnack_mode(struct kfd_process *p, bool supported)
 	 * support retry.
 	 */
 	for (i = 0; i < p->n_pdds; i++) {
-		struct kfd_dev *dev = p->pdds[i]->dev;
+		struct kfd_node *dev = p->pdds[i]->dev;
 
 		/* Only consider GFXv9 and higher GPUs. Older GPUs don't
 		 * support the SVM APIs and don't need to be considered
 		 * for the XNACK mode selection.
 		 */
-		if (dev->device_info->asic_family < CHIP_VEGA10)
+		if (!KFD_IS_SOC15(dev))
 			continue;
 		/* Aldebaran can always support XNACK because it can support
 		 * per-process XNACK mode selection. But let the dev->noretry
 		 * setting still influence the default XNACK mode.
 		 */
-		if (supported &&
-		    dev->device_info->asic_family == CHIP_ALDEBARAN)
+		if (supported && KFD_SUPPORT_XNACK_PER_PROCESS(dev)) {
+			if (!amdgpu_sriov_xnack_support(dev->kfd->adev)) {
+				pr_debug("SRIOV platform xnack not supported\n");
+				return false;
+			}
 			continue;
+		}
 
 		/* GFXv10 and later GPUs do not support shader preemption
 		 * during page faults. This can lead to poor QoS for queue
 		 * management and memory-manager-related preemptions or
 		 * even deadlocks.
 		 */
-		if (dev->device_info->asic_family >= CHIP_NAVI10)
+		if (KFD_GC_VERSION(dev) >= IP_VERSION(10, 1, 1))
 			return false;
 
-		if (dev->noretry)
+		if (dev->kfd->noretry)
 			return false;
 	}
 
 	return true;
 }
 
+void kfd_process_set_trap_debug_flag(struct qcm_process_device *qpd,
+				     bool enabled)
+{
+	if (qpd->cwsr_kaddr) {
+		uint64_t *tma =
+			(uint64_t *)(qpd->cwsr_kaddr + KFD_CWSR_TMA_OFFSET);
+		tma[2] = enabled;
+	}
+}
+
 /*
  * On return the kfd_process is fully operational and will be freed when the
  * mm is released
@@ -1361,15 +1531,19 @@ static struct kfd_process *create_process(const struct task_struct *thread)
 	process->mm = thread->mm;
 	process->lead_thread = thread->group_leader;
 	process->n_pdds = 0;
+	process->queues_paused = false;
 	INIT_DELAYED_WORK(&process->eviction_work, evict_process_worker);
 	INIT_DELAYED_WORK(&process->restore_work, restore_process_worker);
 	process->last_restore_timestamp = get_jiffies_64();
-	kfd_event_init_process(process);
+	err = kfd_event_init_process(process);
+	if (err)
+		goto err_event_init;
 	process->is_32bit_user_mode = in_compat_syscall();
-
-	process->pasid = kfd_pasid_alloc();
-	if (process->pasid == 0)
-		goto err_alloc_pasid;
+	process->debug_trap_enabled = false;
+	process->debugger_process = NULL;
+	process->exception_enable_mask = 0;
+	atomic_set(&process->debugged_process_count, 0);
+	sema_init(&process->runtime_enable_sema, 0);
 
 	err = pqm_init(&process->pqm, process);
 	if (err != 0)
@@ -1391,6 +1565,11 @@ static struct kfd_process *create_process(const struct task_struct *thread)
 	hash_add_rcu(kfd_processes_table, &process->kfd_processes,
 			(uintptr_t)process->mm);
 
+	/* Avoid free_notifier to start kfd_process_wq_release if
+	 * mmu_notifier_get failed because of pending signal.
+	 */
+	kref_get(&process->ref);
+
 	/* MMU notifier registration must be the last call that can fail
 	 * because after this point we cannot unwind the process creation.
 	 * After this point, mmu_notifier_put will trigger the cleanup by
@@ -1403,8 +1582,11 @@ static struct kfd_process *create_process(const struct task_struct *thread)
 	}
 	BUG_ON(mn != &process->mmu_notifier);
 
+	kfd_unref_process(process);
 	get_task_struct(process->lead_thread);
 
+	INIT_WORK(&process->debug_event_workarea, debug_event_write_work_handler);
+
 	return process;
 
 err_register_notifier:
@@ -1416,48 +1598,15 @@ err_init_svm_range_list:
 err_init_apertures:
 	pqm_uninit(&process->pqm);
 err_process_pqm_init:
-	kfd_pasid_free(process->pasid);
-err_alloc_pasid:
+	kfd_event_free_process(process);
+err_event_init:
 	mutex_destroy(&process->mutex);
 	kfree(process);
 err_alloc_process:
 	return ERR_PTR(err);
 }
 
-static int init_doorbell_bitmap(struct qcm_process_device *qpd,
-			struct kfd_dev *dev)
-{
-	unsigned int i;
-	int range_start = dev->shared_resources.non_cp_doorbells_start;
-	int range_end = dev->shared_resources.non_cp_doorbells_end;
-
-	if (!KFD_IS_SOC15(dev->device_info->asic_family))
-		return 0;
-
-	qpd->doorbell_bitmap =
-		kzalloc(DIV_ROUND_UP(KFD_MAX_NUM_OF_QUEUES_PER_PROCESS,
-				     BITS_PER_BYTE), GFP_KERNEL);
-	if (!qpd->doorbell_bitmap)
-		return -ENOMEM;
-
-	/* Mask out doorbells reserved for SDMA, IH, and VCN on SOC15. */
-	pr_debug("reserved doorbell 0x%03x - 0x%03x\n", range_start, range_end);
-	pr_debug("reserved doorbell 0x%03x - 0x%03x\n",
-			range_start + KFD_QUEUE_DOORBELL_MIRROR_OFFSET,
-			range_end + KFD_QUEUE_DOORBELL_MIRROR_OFFSET);
-
-	for (i = 0; i < KFD_MAX_NUM_OF_QUEUES_PER_PROCESS / 2; i++) {
-		if (i >= range_start && i <= range_end) {
-			set_bit(i, qpd->doorbell_bitmap);
-			set_bit(i + KFD_QUEUE_DOORBELL_MIRROR_OFFSET,
-				qpd->doorbell_bitmap);
-		}
-	}
-
-	return 0;
-}
-
-struct kfd_process_device *kfd_get_process_device_data(struct kfd_dev *dev,
+struct kfd_process_device *kfd_get_process_device_data(struct kfd_node *dev,
 							struct kfd_process *p)
 {
 	int i;
@@ -1469,7 +1618,7 @@ struct kfd_process_device *kfd_get_process_device_data(struct kfd_dev *dev,
 	return NULL;
 }
 
-struct kfd_process_device *kfd_create_process_device_data(struct kfd_dev *dev,
+struct kfd_process_device *kfd_create_process_device_data(struct kfd_node *dev,
 							struct kfd_process *p)
 {
 	struct kfd_process_device *pdd = NULL;
@@ -1480,16 +1629,6 @@ struct kfd_process_device *kfd_create_process_device_data(struct kfd_dev *dev,
 	if (!pdd)
 		return NULL;
 
-	if (kfd_alloc_process_doorbells(dev, &pdd->doorbell_index) < 0) {
-		pr_err("Failed to alloc doorbell for pdd\n");
-		goto err_free_pdd;
-	}
-
-	if (init_doorbell_bitmap(&pdd->qpd, dev)) {
-		pr_err("Failed to init doorbell for process\n");
-		goto err_free_pdd;
-	}
-
 	pdd->dev = dev;
 	INIT_LIST_HEAD(&pdd->qpd.queues_list);
 	INIT_LIST_HEAD(&pdd->qpd.priv_queue_list);
@@ -1501,19 +1640,24 @@ struct kfd_process_device *kfd_create_process_device_data(struct kfd_dev *dev,
 	pdd->bound = PDD_UNBOUND;
 	pdd->already_dequeued = false;
 	pdd->runtime_inuse = false;
-	pdd->vram_usage = 0;
+	atomic64_set(&pdd->vram_usage, 0);
 	pdd->sdma_past_activity_counter = 0;
+	pdd->user_gpu_id = dev->id;
 	atomic64_set(&pdd->evict_duration_counter, 0);
+
 	p->pdds[p->n_pdds++] = pdd;
+	if (kfd_dbg_is_per_vmid_supported(pdd->dev))
+		pdd->spi_dbg_override = pdd->dev->kfd2kgd->disable_debug_trap(
+							pdd->dev->adev,
+							false,
+							0);
 
 	/* Init idr used for memory handle translation */
 	idr_init(&pdd->alloc_idr);
 
-	return pdd;
+	atomic_inc(&dev->kfd->kfd_processes_count);
 
-err_free_pdd:
-	kfree(pdd);
-	return NULL;
+	return pdd;
 }
 
 /**
@@ -1533,8 +1677,11 @@ err_free_pdd:
 int kfd_process_device_init_vm(struct kfd_process_device *pdd,
 			       struct file *drm_file)
 {
+	struct amdgpu_fpriv *drv_priv;
+	struct amdgpu_vm *avm;
 	struct kfd_process *p;
-	struct kfd_dev *dev;
+	struct dma_fence *ef;
+	struct kfd_node *dev;
 	int ret;
 
 	if (!drm_file)
@@ -1543,16 +1690,25 @@ int kfd_process_device_init_vm(struct kfd_process_device *pdd,
 	if (pdd->drm_priv)
 		return -EBUSY;
 
+	ret = amdgpu_file_to_fpriv(drm_file, &drv_priv);
+	if (ret)
+		return ret;
+	avm = &drv_priv->vm;
+
 	p = pdd->process;
 	dev = pdd->dev;
 
-	ret = amdgpu_amdkfd_gpuvm_acquire_process_vm(
-		dev->kgd, drm_file, p->pasid,
-		&p->kgd_process_info, &p->ef);
+	ret = amdgpu_amdkfd_gpuvm_acquire_process_vm(dev->adev, avm,
+						     &p->kgd_process_info,
+						     p->ef ? NULL : &ef);
 	if (ret) {
-		pr_err("Failed to create process VM object\n");
+		dev_err(dev->adev->dev, "Failed to create process VM object\n");
 		return ret;
 	}
+
+	if (!p->ef)
+		RCU_INIT_POINTER(p->ef, ef);
+
 	pdd->drm_priv = drm_file->private_data;
 
 	ret = kfd_process_device_reserve_ib_mem(pdd);
@@ -1562,14 +1718,27 @@ int kfd_process_device_init_vm(struct kfd_process_device *pdd,
 	if (ret)
 		goto err_init_cwsr;
 
+	if (unlikely(!avm->pasid)) {
+		dev_warn(pdd->dev->adev->dev, "WARN: vm %p has no pasid associated",
+				 avm);
+		ret = -EINVAL;
+		goto err_get_pasid;
+	}
+
+	pdd->pasid = avm->pasid;
 	pdd->drm_file = drm_file;
 
+	kfd_smi_event_process(pdd, true);
+
 	return 0;
 
+err_get_pasid:
+	kfd_process_device_destroy_cwsr_dgpu(pdd);
 err_init_cwsr:
+	kfd_process_device_destroy_ib_mem(pdd);
 err_reserve_ib_mem:
-	kfd_process_device_free_bos(pdd);
 	pdd->drm_priv = NULL;
+	amdgpu_amdkfd_gpuvm_destroy_cb(dev->adev, avm);
 
 	return ret;
 }
@@ -1581,7 +1750,7 @@ err_reserve_ib_mem:
  *
  * Assumes that the process lock is held.
  */
-struct kfd_process_device *kfd_bind_process_to_device(struct kfd_dev *dev,
+struct kfd_process_device *kfd_bind_process_to_device(struct kfd_node *dev,
 							struct kfd_process *p)
 {
 	struct kfd_process_device *pdd;
@@ -1589,7 +1758,7 @@ struct kfd_process_device *kfd_bind_process_to_device(struct kfd_dev *dev,
 
 	pdd = kfd_get_process_device_data(dev, p);
 	if (!pdd) {
-		pr_err("Process device data doesn't exist\n");
+		dev_err(dev->adev->dev, "Process device data doesn't exist\n");
 		return ERR_PTR(-ENOMEM);
 	}
 
@@ -1602,17 +1771,13 @@ struct kfd_process_device *kfd_bind_process_to_device(struct kfd_dev *dev,
 	 * pdd is destroyed.
 	 */
 	if (!pdd->runtime_inuse) {
-		err = pm_runtime_get_sync(dev->ddev->dev);
+		err = pm_runtime_get_sync(adev_to_drm(dev->adev)->dev);
 		if (err < 0) {
-			pm_runtime_put_autosuspend(dev->ddev->dev);
+			pm_runtime_put_autosuspend(adev_to_drm(dev->adev)->dev);
 			return ERR_PTR(err);
 		}
 	}
 
-	err = kfd_iommu_bind_process_to_device(pdd);
-	if (err)
-		goto out;
-
 	/*
 	 * make sure that runtime_usage counter is incremented just once
 	 * per pdd
@@ -1620,15 +1785,6 @@ struct kfd_process_device *kfd_bind_process_to_device(struct kfd_dev *dev,
 	pdd->runtime_inuse = true;
 
 	return pdd;
-
-out:
-	/* balance runpm reference count and exit with error */
-	if (!pdd->runtime_inuse) {
-		pm_runtime_mark_last_busy(dev->ddev->dev);
-		pm_runtime_put_autosuspend(dev->ddev->dev);
-	}
-
-	return ERR_PTR(err);
 }
 
 /* Create specific handle mapped to mem from process local memory idr
@@ -1662,25 +1818,50 @@ void kfd_process_device_remove_obj_handle(struct kfd_process_device *pdd,
 		idr_remove(&pdd->alloc_idr, handle);
 }
 
-/* This increments the process->ref counter. */
-struct kfd_process *kfd_lookup_process_by_pasid(u32 pasid)
+static struct kfd_process_device *kfd_lookup_process_device_by_pasid(u32 pasid)
 {
-	struct kfd_process *p, *ret_p = NULL;
+	struct kfd_process_device *ret_p = NULL;
+	struct kfd_process *p;
 	unsigned int temp;
-
-	int idx = srcu_read_lock(&kfd_processes_srcu);
+	int i;
 
 	hash_for_each_rcu(kfd_processes_table, temp, p, kfd_processes) {
-		if (p->pasid == pasid) {
-			kref_get(&p->ref);
-			ret_p = p;
-			break;
+		for (i = 0; i < p->n_pdds; i++) {
+			if (p->pdds[i]->pasid == pasid) {
+				ret_p = p->pdds[i];
+				break;
+			}
 		}
+		if (ret_p)
+			break;
+	}
+	return ret_p;
+}
+
+/* This increments the process->ref counter. */
+struct kfd_process *kfd_lookup_process_by_pasid(u32 pasid,
+						struct kfd_process_device **pdd)
+{
+	struct kfd_process_device *ret_p;
+
+	int idx = srcu_read_lock(&kfd_processes_srcu);
+
+	ret_p = kfd_lookup_process_device_by_pasid(pasid);
+	if (ret_p) {
+		if (pdd)
+			*pdd = ret_p;
+		kref_get(&ret_p->process->ref);
+
+		srcu_read_unlock(&kfd_processes_srcu, idx);
+		return ret_p->process;
 	}
 
 	srcu_read_unlock(&kfd_processes_srcu, idx);
 
-	return ret_p;
+	if (pdd)
+		*pdd = NULL;
+
+	return NULL;
 }
 
 /* This increments the process->ref counter. */
@@ -1704,7 +1885,7 @@ struct kfd_process *kfd_lookup_process_by_mm(const struct mm_struct *mm)
  * Eviction is reference-counted per process-device. This means multiple
  * evictions from different sources can be nested safely.
  */
-int kfd_process_evict_queues(struct kfd_process *p)
+int kfd_process_evict_queues(struct kfd_process *p, uint32_t trigger)
 {
 	int r = 0;
 	int i;
@@ -1712,6 +1893,10 @@ int kfd_process_evict_queues(struct kfd_process *p)
 
 	for (i = 0; i < p->n_pdds; i++) {
 		struct kfd_process_device *pdd = p->pdds[i];
+		struct device *dev = pdd->dev->adev->dev;
+
+		kfd_smi_event_queue_eviction(pdd->dev, p->lead_thread->pid,
+					     trigger);
 
 		r = pdd->dev->dqm->ops.evict_process_queues(pdd->dev->dqm,
 							    &pdd->qpd);
@@ -1720,10 +1905,12 @@ int kfd_process_evict_queues(struct kfd_process *p)
 		 * them been add back since they actually not be saved right now.
 		 */
 		if (r && r != -EIO) {
-			pr_err("Failed to evict process queues\n");
+			dev_err(dev, "Failed to evict process queues\n");
 			goto fail;
 		}
 		n_evicted++;
+
+		pdd->dev->dqm->is_hws_hang = false;
 	}
 
 	return r;
@@ -1737,9 +1924,13 @@ fail:
 
 		if (n_evicted == 0)
 			break;
+
+		kfd_smi_event_queue_restore(pdd->dev, p->lead_thread->pid);
+
 		if (pdd->dev->dqm->ops.restore_process_queues(pdd->dev->dqm,
 							      &pdd->qpd))
-			pr_err("Failed to restore queues\n");
+			dev_err(pdd->dev->adev->dev,
+				"Failed to restore queues\n");
 
 		n_evicted--;
 	}
@@ -1755,11 +1946,14 @@ int kfd_process_restore_queues(struct kfd_process *p)
 
 	for (i = 0; i < p->n_pdds; i++) {
 		struct kfd_process_device *pdd = p->pdds[i];
+		struct device *dev = pdd->dev->adev->dev;
+
+		kfd_smi_event_queue_restore(pdd->dev, p->lead_thread->pid);
 
 		r = pdd->dev->dqm->ops.restore_process_queues(pdd->dev->dqm,
 							      &pdd->qpd);
 		if (r) {
-			pr_err("Failed to restore process queues\n");
+			dev_err(dev, "Failed to restore process queues\n");
 			if (!ret)
 				ret = r;
 		}
@@ -1773,27 +1967,43 @@ int kfd_process_gpuidx_from_gpuid(struct kfd_process *p, uint32_t gpu_id)
 	int i;
 
 	for (i = 0; i < p->n_pdds; i++)
-		if (p->pdds[i] && gpu_id == p->pdds[i]->dev->id)
+		if (p->pdds[i] && gpu_id == p->pdds[i]->user_gpu_id)
 			return i;
 	return -EINVAL;
 }
 
 int
-kfd_process_gpuid_from_kgd(struct kfd_process *p, struct amdgpu_device *adev,
-			   uint32_t *gpuid, uint32_t *gpuidx)
+kfd_process_gpuid_from_node(struct kfd_process *p, struct kfd_node *node,
+			    uint32_t *gpuid, uint32_t *gpuidx)
 {
-	struct kgd_dev *kgd = (struct kgd_dev *)adev;
 	int i;
 
 	for (i = 0; i < p->n_pdds; i++)
-		if (p->pdds[i] && p->pdds[i]->dev->kgd == kgd) {
-			*gpuid = p->pdds[i]->dev->id;
+		if (p->pdds[i] && p->pdds[i]->dev == node) {
+			*gpuid = p->pdds[i]->user_gpu_id;
 			*gpuidx = i;
 			return 0;
 		}
 	return -EINVAL;
 }
 
+static int signal_eviction_fence(struct kfd_process *p)
+{
+	struct dma_fence *ef;
+	int ret;
+
+	rcu_read_lock();
+	ef = dma_fence_get_rcu_safe(&p->ef);
+	rcu_read_unlock();
+	if (!ef)
+		return -EINVAL;
+
+	ret = dma_fence_signal(ef);
+	dma_fence_put(ef);
+
+	return ret;
+}
+
 static void evict_process_worker(struct work_struct *work)
 {
 	int ret;
@@ -1806,29 +2016,45 @@ static void evict_process_worker(struct work_struct *work)
 	 * lifetime of this thread, kfd_process p will be valid
 	 */
 	p = container_of(dwork, struct kfd_process, eviction_work);
-	WARN_ONCE(p->last_eviction_seqno != p->ef->seqno,
-		  "Eviction fence mismatch\n");
-
-	/* Narrow window of overlap between restore and evict work
-	 * item is possible. Once amdgpu_amdkfd_gpuvm_restore_process_bos
-	 * unreserves KFD BOs, it is possible to evicted again. But
-	 * restore has few more steps of finish. So lets wait for any
-	 * previous restore work to complete
-	 */
-	flush_delayed_work(&p->restore_work);
 
-	pr_debug("Started evicting pasid 0x%x\n", p->pasid);
-	ret = kfd_process_evict_queues(p);
+	pr_debug("Started evicting process pid %d\n", p->lead_thread->pid);
+	ret = kfd_process_evict_queues(p, KFD_QUEUE_EVICTION_TRIGGER_TTM);
 	if (!ret) {
-		dma_fence_signal(p->ef);
-		dma_fence_put(p->ef);
-		p->ef = NULL;
-		queue_delayed_work(kfd_restore_wq, &p->restore_work,
-				msecs_to_jiffies(PROCESS_RESTORE_TIME_MS));
+		/* If another thread already signaled the eviction fence,
+		 * they are responsible stopping the queues and scheduling
+		 * the restore work.
+		 */
+		if (signal_eviction_fence(p) ||
+		    mod_delayed_work(kfd_restore_wq, &p->restore_work,
+				     msecs_to_jiffies(PROCESS_RESTORE_TIME_MS)))
+			kfd_process_restore_queues(p);
 
-		pr_debug("Finished evicting pasid 0x%x\n", p->pasid);
+		pr_debug("Finished evicting process pid %d\n", p->lead_thread->pid);
 	} else
-		pr_err("Failed to evict queues of pasid 0x%x\n", p->pasid);
+		pr_err("Failed to evict queues of process pid %d\n", p->lead_thread->pid);
+}
+
+static int restore_process_helper(struct kfd_process *p)
+{
+	int ret = 0;
+
+	/* VMs may not have been acquired yet during debugging. */
+	if (p->kgd_process_info) {
+		ret = amdgpu_amdkfd_gpuvm_restore_process_bos(
+			p->kgd_process_info, &p->ef);
+		if (ret)
+			return ret;
+	}
+
+	ret = kfd_process_restore_queues(p);
+	if (!ret)
+		pr_debug("Finished restoring process pid %d\n",
+			p->lead_thread->pid);
+	else
+		pr_err("Failed to restore queues of process pid %d\n",
+		      p->lead_thread->pid);
+
+	return ret;
 }
 
 static void restore_process_worker(struct work_struct *work)
@@ -1843,7 +2069,7 @@ static void restore_process_worker(struct work_struct *work)
 	 * lifetime of this thread, kfd_process p will be valid
 	 */
 	p = container_of(dwork, struct kfd_process, restore_work);
-	pr_debug("Started restoring pasid 0x%x\n", p->pasid);
+	pr_debug("Started restoring process pasid %d\n", (int)p->lead_thread->pid);
 
 	/* Setting last_restore_timestamp before successful restoration.
 	 * Otherwise this would have to be set by KGD (restore_process_bos)
@@ -1856,22 +2082,15 @@ static void restore_process_worker(struct work_struct *work)
 	 */
 
 	p->last_restore_timestamp = get_jiffies_64();
-	ret = amdgpu_amdkfd_gpuvm_restore_process_bos(p->kgd_process_info,
-						     &p->ef);
+
+	ret = restore_process_helper(p);
 	if (ret) {
-		pr_debug("Failed to restore BOs of pasid 0x%x, retry after %d ms\n",
-			 p->pasid, PROCESS_BACK_OFF_TIME_MS);
-		ret = queue_delayed_work(kfd_restore_wq, &p->restore_work,
-				msecs_to_jiffies(PROCESS_BACK_OFF_TIME_MS));
-		WARN(!ret, "reschedule restore work failed\n");
-		return;
+		pr_debug("Failed to restore BOs of process pid %d, retry after %d ms\n",
+			 p->lead_thread->pid, PROCESS_BACK_OFF_TIME_MS);
+		if (mod_delayed_work(kfd_restore_wq, &p->restore_work,
+				     msecs_to_jiffies(PROCESS_RESTORE_TIME_MS)))
+			kfd_process_restore_queues(p);
 	}
-
-	ret = kfd_process_restore_queues(p);
-	if (!ret)
-		pr_debug("Finished restoring pasid 0x%x\n", p->pasid);
-	else
-		pr_err("Failed to restore queues of pasid 0x%x\n", p->pasid);
 }
 
 void kfd_suspend_all_processes(void)
@@ -1882,14 +2101,9 @@ void kfd_suspend_all_processes(void)
 
 	WARN(debug_evictions, "Evicting all processes");
 	hash_for_each_rcu(kfd_processes_table, temp, p, kfd_processes) {
-		cancel_delayed_work_sync(&p->eviction_work);
-		cancel_delayed_work_sync(&p->restore_work);
-
-		if (kfd_process_evict_queues(p))
-			pr_err("Failed to suspend process 0x%x\n", p->pasid);
-		dma_fence_signal(p->ef);
-		dma_fence_put(p->ef);
-		p->ef = NULL;
+		if (kfd_process_evict_queues(p, KFD_QUEUE_EVICTION_TRIGGER_SUSPEND))
+			pr_err("Failed to suspend process pid %d\n", p->lead_thread->pid);
+		signal_eviction_fence(p);
 	}
 	srcu_read_unlock(&kfd_processes_srcu, idx);
 }
@@ -1901,9 +2115,9 @@ int kfd_resume_all_processes(void)
 	int ret = 0, idx = srcu_read_lock(&kfd_processes_srcu);
 
 	hash_for_each_rcu(kfd_processes_table, temp, p, kfd_processes) {
-		if (!queue_delayed_work(kfd_restore_wq, &p->restore_work, 0)) {
-			pr_err("Restore process %d failed during resume\n",
-			       p->pasid);
+		if (restore_process_helper(p)) {
+			pr_err("Restore process pid %d failed during resume\n",
+			      p->lead_thread->pid);
 			ret = -EFAULT;
 		}
 	}
@@ -1911,14 +2125,14 @@ int kfd_resume_all_processes(void)
 	return ret;
 }
 
-int kfd_reserved_mem_mmap(struct kfd_dev *dev, struct kfd_process *process,
+int kfd_reserved_mem_mmap(struct kfd_node *dev, struct kfd_process *process,
 			  struct vm_area_struct *vma)
 {
 	struct kfd_process_device *pdd;
 	struct qcm_process_device *qpd;
 
 	if ((vma->vm_end - vma->vm_start) != KFD_CWSR_TBA_TMA_SIZE) {
-		pr_err("Incorrect CWSR mapping size.\n");
+		dev_err(dev->adev->dev, "Incorrect CWSR mapping size.\n");
 		return -EINVAL;
 	}
 
@@ -1930,33 +2144,173 @@ int kfd_reserved_mem_mmap(struct kfd_dev *dev, struct kfd_process *process,
 	qpd->cwsr_kaddr = (void *)__get_free_pages(GFP_KERNEL | __GFP_ZERO,
 					get_order(KFD_CWSR_TBA_TMA_SIZE));
 	if (!qpd->cwsr_kaddr) {
-		pr_err("Error allocating per process CWSR buffer.\n");
+		dev_err(dev->adev->dev,
+			"Error allocating per process CWSR buffer.\n");
 		return -ENOMEM;
 	}
 
-	vma->vm_flags |= VM_IO | VM_DONTCOPY | VM_DONTEXPAND
-		| VM_NORESERVE | VM_DONTDUMP | VM_PFNMAP;
+	vm_flags_set(vma, VM_IO | VM_DONTCOPY | VM_DONTEXPAND
+		| VM_NORESERVE | VM_DONTDUMP | VM_PFNMAP);
 	/* Mapping pages to user process */
 	return remap_pfn_range(vma, vma->vm_start,
 			       PFN_DOWN(__pa(qpd->cwsr_kaddr)),
 			       KFD_CWSR_TBA_TMA_SIZE, vma->vm_page_prot);
 }
 
-void kfd_flush_tlb(struct kfd_process_device *pdd, enum TLB_FLUSH_TYPE type)
+/* assumes caller holds process lock. */
+int kfd_process_drain_interrupts(struct kfd_process_device *pdd)
 {
-	struct kfd_dev *dev = pdd->dev;
+	uint32_t irq_drain_fence[8];
+	uint8_t node_id = 0;
+	int r = 0;
 
-	if (dev->dqm->sched_policy == KFD_SCHED_POLICY_NO_HWS) {
-		/* Nothing to flush until a VMID is assigned, which
-		 * only happens when the first queue is created.
-		 */
-		if (pdd->qpd.vmid)
-			amdgpu_amdkfd_flush_gpu_tlb_vmid(dev->kgd,
-							pdd->qpd.vmid);
-	} else {
-		amdgpu_amdkfd_flush_gpu_tlb_pasid(dev->kgd,
-					pdd->process->pasid, type);
+	if (!KFD_IS_SOC15(pdd->dev))
+		return 0;
+
+	pdd->process->irq_drain_is_open = true;
+
+	memset(irq_drain_fence, 0, sizeof(irq_drain_fence));
+	irq_drain_fence[0] = (KFD_IRQ_FENCE_SOURCEID << 8) |
+							KFD_IRQ_FENCE_CLIENTID;
+	irq_drain_fence[3] = pdd->pasid;
+
+	/*
+	 * For GFX 9.4.3/9.5.0, send the NodeId also in IH cookie DW[3]
+	 */
+	if (KFD_GC_VERSION(pdd->dev->kfd) == IP_VERSION(9, 4, 3) ||
+	    KFD_GC_VERSION(pdd->dev->kfd) == IP_VERSION(9, 4, 4) ||
+	    KFD_GC_VERSION(pdd->dev->kfd) == IP_VERSION(9, 5, 0)) {
+		node_id = ffs(pdd->dev->interrupt_bitmap) - 1;
+		irq_drain_fence[3] |= node_id << 16;
+	}
+
+	/* ensure stale irqs scheduled KFD interrupts and send drain fence. */
+	if (amdgpu_amdkfd_send_close_event_drain_irq(pdd->dev->adev,
+						     irq_drain_fence)) {
+		pdd->process->irq_drain_is_open = false;
+		return 0;
+	}
+
+	r = wait_event_interruptible(pdd->process->wait_irq_drain,
+				     !READ_ONCE(pdd->process->irq_drain_is_open));
+	if (r)
+		pdd->process->irq_drain_is_open = false;
+
+	return r;
+}
+
+void kfd_process_close_interrupt_drain(unsigned int pasid)
+{
+	struct kfd_process *p;
+
+	p = kfd_lookup_process_by_pasid(pasid, NULL);
+
+	if (!p)
+		return;
+
+	WRITE_ONCE(p->irq_drain_is_open, false);
+	wake_up_all(&p->wait_irq_drain);
+	kfd_unref_process(p);
+}
+
+struct send_exception_work_handler_workarea {
+	struct work_struct work;
+	struct kfd_process *p;
+	unsigned int queue_id;
+	uint64_t error_reason;
+};
+
+static void send_exception_work_handler(struct work_struct *work)
+{
+	struct send_exception_work_handler_workarea *workarea;
+	struct kfd_process *p;
+	struct queue *q;
+	struct mm_struct *mm;
+	struct kfd_context_save_area_header __user *csa_header;
+	uint64_t __user *err_payload_ptr;
+	uint64_t cur_err;
+	uint32_t ev_id;
+
+	workarea = container_of(work,
+				struct send_exception_work_handler_workarea,
+				work);
+	p = workarea->p;
+
+	mm = get_task_mm(p->lead_thread);
+
+	if (!mm)
+		return;
+
+	kthread_use_mm(mm);
+
+	q = pqm_get_user_queue(&p->pqm, workarea->queue_id);
+
+	if (!q)
+		goto out;
+
+	csa_header = (void __user *)q->properties.ctx_save_restore_area_address;
+
+	get_user(err_payload_ptr, (uint64_t __user **)&csa_header->err_payload_addr);
+	get_user(cur_err, err_payload_ptr);
+	cur_err |= workarea->error_reason;
+	put_user(cur_err, err_payload_ptr);
+	get_user(ev_id, &csa_header->err_event_id);
+
+	kfd_set_event(p, ev_id);
+
+out:
+	kthread_unuse_mm(mm);
+	mmput(mm);
+}
+
+int kfd_send_exception_to_runtime(struct kfd_process *p,
+			unsigned int queue_id,
+			uint64_t error_reason)
+{
+	struct send_exception_work_handler_workarea worker;
+
+	INIT_WORK_ONSTACK(&worker.work, send_exception_work_handler);
+
+	worker.p = p;
+	worker.queue_id = queue_id;
+	worker.error_reason = error_reason;
+
+	schedule_work(&worker.work);
+	flush_work(&worker.work);
+	destroy_work_on_stack(&worker.work);
+
+	return 0;
+}
+
+struct kfd_process_device *kfd_process_device_data_by_id(struct kfd_process *p, uint32_t gpu_id)
+{
+	int i;
+
+	if (gpu_id) {
+		for (i = 0; i < p->n_pdds; i++) {
+			struct kfd_process_device *pdd = p->pdds[i];
+
+			if (pdd->user_gpu_id == gpu_id)
+				return pdd;
+		}
+	}
+	return NULL;
+}
+
+int kfd_process_get_user_gpu_id(struct kfd_process *p, uint32_t actual_gpu_id)
+{
+	int i;
+
+	if (!actual_gpu_id)
+		return 0;
+
+	for (i = 0; i < p->n_pdds; i++) {
+		struct kfd_process_device *pdd = p->pdds[i];
+
+		if (pdd->dev->id == actual_gpu_id)
+			return pdd->user_gpu_id;
 	}
+	return -EINVAL;
 }
 
 #if defined(CONFIG_DEBUG_FS)
@@ -1970,8 +2324,8 @@ int kfd_debugfs_mqds_by_process(struct seq_file *m, void *data)
 	int idx = srcu_read_lock(&kfd_processes_srcu);
 
 	hash_for_each_rcu(kfd_processes_table, temp, p, kfd_processes) {
-		seq_printf(m, "Process %d PASID 0x%x:\n",
-			   p->lead_thread->tgid, p->pasid);
+		seq_printf(m, "Process %d PASID %d:\n",
+			   p->lead_thread->tgid, p->lead_thread->pid);
 
 		mutex_lock(&p->mutex);
 		r = pqm_debugfs_mqds(m, &p->pqm);
@@ -1987,4 +2341,3 @@ int kfd_debugfs_mqds_by_process(struct seq_file *m, void *data)
 }
 
 #endif
-
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c b/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c
index 3627e7ac161b..7fbb5c274ccc 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c
@@ -1,5 +1,6 @@
+// SPDX-License-Identifier: GPL-2.0 OR MIT
 /*
- * Copyright 2014 Advanced Micro Devices, Inc.
+ * Copyright 2014-2022 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -27,6 +28,7 @@
 #include "kfd_priv.h"
 #include "kfd_kernel_queue.h"
 #include "amdgpu_amdkfd.h"
+#include "amdgpu_reset.h"
 
 static inline struct process_queue_node *get_queue_by_qid(
 			struct process_queue_manager *pqm, unsigned int qid)
@@ -42,6 +44,20 @@ static inline struct process_queue_node *get_queue_by_qid(
 	return NULL;
 }
 
+static int assign_queue_slot_by_qid(struct process_queue_manager *pqm,
+				    unsigned int qid)
+{
+	if (qid >= KFD_MAX_NUM_OF_QUEUES_PER_PROCESS)
+		return -EINVAL;
+
+	if (__test_and_set_bit(qid, pqm->queue_slot_bitmap)) {
+		pr_err("Cannot create new queue because requested qid(%u) is in use\n", qid);
+		return -ENOSPC;
+	}
+
+	return 0;
+}
+
 static int find_available_queue_slot(struct process_queue_manager *pqm,
 					unsigned int *qid)
 {
@@ -53,8 +69,8 @@ static int find_available_queue_slot(struct process_queue_manager *pqm,
 	pr_debug("The new slot id %lu\n", found);
 
 	if (found >= KFD_MAX_NUM_OF_QUEUES_PER_PROCESS) {
-		pr_info("Cannot open more queues for process with pasid 0x%x\n",
-				pqm->process->pasid);
+		pr_info("Cannot open more queues for process with pid %d\n",
+			pqm->process->lead_thread->pid);
 		return -ENOMEM;
 	}
 
@@ -66,19 +82,29 @@ static int find_available_queue_slot(struct process_queue_manager *pqm,
 
 void kfd_process_dequeue_from_device(struct kfd_process_device *pdd)
 {
-	struct kfd_dev *dev = pdd->dev;
+	struct kfd_node *dev = pdd->dev;
 
 	if (pdd->already_dequeued)
 		return;
-
+	/* The MES context flush needs to filter out the case which the
+	 * KFD process is created without setting up the MES context and
+	 * queue for creating a compute queue.
+	 */
 	dev->dqm->ops.process_termination(dev->dqm, &pdd->qpd);
+	if (dev->kfd->shared_resources.enable_mes && !!pdd->proc_ctx_gpu_addr &&
+	    down_read_trylock(&dev->adev->reset_domain->sem)) {
+		amdgpu_mes_flush_shader_debugger(dev->adev,
+						 pdd->proc_ctx_gpu_addr);
+		up_read(&dev->adev->reset_domain->sem);
+	}
 	pdd->already_dequeued = true;
 }
 
 int pqm_set_gws(struct process_queue_manager *pqm, unsigned int qid,
 			void *gws)
 {
-	struct kfd_dev *dev = NULL;
+	struct mqd_update_info minfo = {0};
+	struct kfd_node *dev = NULL;
 	struct process_queue_node *pqn;
 	struct kfd_process_device *pdd;
 	struct kgd_mem *mem = NULL;
@@ -108,20 +134,34 @@ int pqm_set_gws(struct process_queue_manager *pqm, unsigned int qid,
 	if (!gws && pdd->qpd.num_gws == 0)
 		return -EINVAL;
 
-	if (gws)
-		ret = amdgpu_amdkfd_add_gws_to_process(pdd->process->kgd_process_info,
-			gws, &mem);
-	else
-		ret = amdgpu_amdkfd_remove_gws_from_process(pdd->process->kgd_process_info,
-			pqn->q->gws);
-	if (unlikely(ret))
-		return ret;
+	if ((KFD_GC_VERSION(dev) != IP_VERSION(9, 4, 3) &&
+	     KFD_GC_VERSION(dev) != IP_VERSION(9, 4, 4) &&
+	     KFD_GC_VERSION(dev) != IP_VERSION(9, 5, 0)) &&
+	    !dev->kfd->shared_resources.enable_mes) {
+		if (gws)
+			ret = amdgpu_amdkfd_add_gws_to_process(pdd->process->kgd_process_info,
+				gws, &mem);
+		else
+			ret = amdgpu_amdkfd_remove_gws_from_process(pdd->process->kgd_process_info,
+				pqn->q->gws);
+		if (unlikely(ret))
+			return ret;
+		pqn->q->gws = mem;
+	} else {
+		/*
+		 * Intentionally set GWS to a non-NULL value
+		 * for devices that do not use GWS for global wave
+		 * synchronization but require the formality
+		 * of setting GWS for cooperative groups.
+		 */
+		pqn->q->gws = gws ? ERR_PTR(-ENOMEM) : NULL;
+	}
 
-	pqn->q->gws = mem;
-	pdd->qpd.num_gws = gws ? amdgpu_amdkfd_get_num_gws(dev->kgd) : 0;
+	pdd->qpd.num_gws = gws ? dev->adev->gds.gws_size : 0;
+	minfo.update_flag = gws ? UPDATE_FLAG_IS_GWS : 0;
 
 	return pqn->q->device->dqm->ops.update_queue(pqn->q->device->dqm,
-							pqn->q, NULL);
+							pqn->q, &minfo);
 }
 
 void kfd_process_dequeue_from_all_devices(struct kfd_process *p)
@@ -135,9 +175,8 @@ void kfd_process_dequeue_from_all_devices(struct kfd_process *p)
 int pqm_init(struct process_queue_manager *pqm, struct kfd_process *p)
 {
 	INIT_LIST_HEAD(&pqm->queues);
-	pqm->queue_slot_bitmap =
-			kzalloc(DIV_ROUND_UP(KFD_MAX_NUM_OF_QUEUES_PER_PROCESS,
-					BITS_PER_BYTE), GFP_KERNEL);
+	pqm->queue_slot_bitmap = bitmap_zalloc(KFD_MAX_NUM_OF_QUEUES_PER_PROCESS,
+					       GFP_KERNEL);
 	if (!pqm->queue_slot_bitmap)
 		return -ENOMEM;
 	pqm->process = p;
@@ -145,33 +184,73 @@ int pqm_init(struct process_queue_manager *pqm, struct kfd_process *p)
 	return 0;
 }
 
+static void pqm_clean_queue_resource(struct process_queue_manager *pqm,
+				     struct process_queue_node *pqn)
+{
+	struct kfd_node *dev;
+	struct kfd_process_device *pdd;
+
+	dev = pqn->q->device;
+
+	pdd = kfd_get_process_device_data(dev, pqm->process);
+	if (!pdd) {
+		pr_err("Process device data doesn't exist\n");
+		return;
+	}
+
+	if (pqn->q->gws) {
+		if (KFD_GC_VERSION(pqn->q->device) != IP_VERSION(9, 4, 3) &&
+		    KFD_GC_VERSION(pqn->q->device) != IP_VERSION(9, 4, 4) &&
+		    KFD_GC_VERSION(pqn->q->device) != IP_VERSION(9, 5, 0) &&
+		    !dev->kfd->shared_resources.enable_mes)
+			amdgpu_amdkfd_remove_gws_from_process(
+				pqm->process->kgd_process_info, pqn->q->gws);
+		pdd->qpd.num_gws = 0;
+	}
+
+	if (dev->kfd->shared_resources.enable_mes) {
+		amdgpu_amdkfd_free_gtt_mem(dev->adev, &pqn->q->gang_ctx_bo);
+		amdgpu_amdkfd_free_gtt_mem(dev->adev, (void **)&pqn->q->wptr_bo_gart);
+	}
+}
+
 void pqm_uninit(struct process_queue_manager *pqm)
 {
 	struct process_queue_node *pqn, *next;
 
 	list_for_each_entry_safe(pqn, next, &pqm->queues, process_queue_list) {
-		if (pqn->q && pqn->q->gws)
-			amdgpu_amdkfd_remove_gws_from_process(pqm->process->kgd_process_info,
-				pqn->q->gws);
+		if (pqn->q) {
+			struct kfd_process_device *pdd = kfd_get_process_device_data(pqn->q->device,
+										     pqm->process);
+			if (pdd) {
+				kfd_queue_unref_bo_vas(pdd, &pqn->q->properties);
+				kfd_queue_release_buffers(pdd, &pqn->q->properties);
+			} else {
+				WARN_ON(!pdd);
+			}
+			pqm_clean_queue_resource(pqm, pqn);
+		}
+
 		kfd_procfs_del_queue(pqn->q);
 		uninit_queue(pqn->q);
 		list_del(&pqn->process_queue_list);
 		kfree(pqn);
 	}
 
-	kfree(pqm->queue_slot_bitmap);
+	bitmap_free(pqm->queue_slot_bitmap);
 	pqm->queue_slot_bitmap = NULL;
 }
 
 static int init_user_queue(struct process_queue_manager *pqm,
-				struct kfd_dev *dev, struct queue **q,
+				struct kfd_node *dev, struct queue **q,
 				struct queue_properties *q_properties,
-				struct file *f, unsigned int qid)
+				unsigned int qid)
 {
 	int retval;
 
 	/* Doorbell initialized in user space*/
 	q_properties->doorbell_ptr = NULL;
+	q_properties->exception_status = KFD_EC_MASK(EC_QUEUE_NEW);
 
 	/* let DQM handle it*/
 	q_properties->vmid = 0;
@@ -184,16 +263,54 @@ static int init_user_queue(struct process_queue_manager *pqm,
 	(*q)->device = dev;
 	(*q)->process = pqm->process;
 
+	if (dev->kfd->shared_resources.enable_mes) {
+		retval = amdgpu_amdkfd_alloc_gtt_mem(dev->adev,
+						AMDGPU_MES_GANG_CTX_SIZE,
+						&(*q)->gang_ctx_bo,
+						&(*q)->gang_ctx_gpu_addr,
+						&(*q)->gang_ctx_cpu_ptr,
+						false);
+		if (retval) {
+			pr_err("failed to allocate gang context bo\n");
+			goto cleanup;
+		}
+		memset((*q)->gang_ctx_cpu_ptr, 0, AMDGPU_MES_GANG_CTX_SIZE);
+
+		/* Starting with GFX11, wptr BOs must be mapped to GART for MES to determine work
+		 * on unmapped queues for usermode queue oversubscription (no aggregated doorbell)
+		 */
+		if (dev->adev != amdgpu_ttm_adev(q_properties->wptr_bo->tbo.bdev)) {
+			pr_err("Queue memory allocated to wrong device\n");
+			retval = -EINVAL;
+			goto free_gang_ctx_bo;
+		}
+
+		retval = amdgpu_amdkfd_map_gtt_bo_to_gart(q_properties->wptr_bo,
+							  &(*q)->wptr_bo_gart);
+		if (retval) {
+			pr_err("Failed to map wptr bo to GART\n");
+			goto free_gang_ctx_bo;
+		}
+	}
+
 	pr_debug("PQM After init queue");
+	return 0;
 
+free_gang_ctx_bo:
+	amdgpu_amdkfd_free_gtt_mem(dev->adev, &(*q)->gang_ctx_bo);
+cleanup:
+	uninit_queue(*q);
+	*q = NULL;
 	return retval;
 }
 
 int pqm_create_queue(struct process_queue_manager *pqm,
-			    struct kfd_dev *dev,
-			    struct file *f,
+			    struct kfd_node *dev,
 			    struct queue_properties *properties,
 			    unsigned int *qid,
+			    const struct kfd_criu_queue_priv_data *q_data,
+			    const void *restore_mqd,
+			    const void *restore_ctl_stack,
 			    uint32_t *p_doorbell_offset_in_process)
 {
 	int retval;
@@ -204,6 +321,15 @@ int pqm_create_queue(struct process_queue_manager *pqm,
 	enum kfd_queue_type type = properties->type;
 	unsigned int max_queues = 127; /* HWS limit */
 
+	/*
+	 * On GFX 9.4.3/9.5.0, increase the number of queues that
+	 * can be created to 255. No HWS limit on GFX 9.4.3/9.5.0.
+	 */
+	if (KFD_GC_VERSION(dev) == IP_VERSION(9, 4, 3) ||
+	    KFD_GC_VERSION(dev) == IP_VERSION(9, 4, 4) ||
+	    KFD_GC_VERSION(dev) == IP_VERSION(9, 5, 0))
+		max_queues = 255;
+
 	q = NULL;
 	kq = NULL;
 
@@ -220,19 +346,40 @@ int pqm_create_queue(struct process_queue_manager *pqm,
 	 * Hence we also check the type as well
 	 */
 	if ((pdd->qpd.is_debug) || (type == KFD_QUEUE_TYPE_DIQ))
-		max_queues = dev->device_info->max_no_of_hqd/2;
+		max_queues = dev->kfd->device_info.max_no_of_hqd/2;
 
 	if (pdd->qpd.queue_count >= max_queues)
 		return -ENOSPC;
 
-	retval = find_available_queue_slot(pqm, qid);
+	if (q_data) {
+		retval = assign_queue_slot_by_qid(pqm, q_data->q_id);
+		*qid = q_data->q_id;
+	} else
+		retval = find_available_queue_slot(pqm, qid);
+
 	if (retval != 0)
 		return retval;
 
+	/* Register process if this is the first queue */
 	if (list_empty(&pdd->qpd.queues_list) &&
 	    list_empty(&pdd->qpd.priv_queue_list))
 		dev->dqm->ops.register_process(dev->dqm, &pdd->qpd);
 
+	/* Allocate proc_ctx_bo only if MES is enabled and this is the first queue */
+	if (!pdd->proc_ctx_cpu_ptr && dev->kfd->shared_resources.enable_mes) {
+		retval = amdgpu_amdkfd_alloc_gtt_mem(dev->adev,
+						     AMDGPU_MES_PROC_CTX_SIZE,
+						     &pdd->proc_ctx_bo,
+						     &pdd->proc_ctx_gpu_addr,
+						     &pdd->proc_ctx_cpu_ptr,
+						     false);
+		if (retval) {
+			dev_err(dev->adev->dev, "failed to allocate process context bo\n");
+			return retval;
+		}
+		memset(pdd->proc_ctx_cpu_ptr, 0, AMDGPU_MES_PROC_CTX_SIZE);
+	}
+
 	pqn = kzalloc(sizeof(*pqn), GFP_KERNEL);
 	if (!pqn) {
 		retval = -ENOMEM;
@@ -242,18 +389,20 @@ int pqm_create_queue(struct process_queue_manager *pqm,
 	switch (type) {
 	case KFD_QUEUE_TYPE_SDMA:
 	case KFD_QUEUE_TYPE_SDMA_XGMI:
+	case KFD_QUEUE_TYPE_SDMA_BY_ENG_ID:
 		/* SDMA queues are always allocated statically no matter
 		 * which scheduler mode is used. We also do not need to
 		 * check whether a SDMA queue can be allocated here, because
 		 * allocate_sdma_queue() in create_queue() has the
 		 * corresponding check logic.
 		 */
-		retval = init_user_queue(pqm, dev, &q, properties, f, *qid);
+		retval = init_user_queue(pqm, dev, &q, properties, *qid);
 		if (retval != 0)
 			goto err_create_queue;
 		pqn->q = q;
 		pqn->kq = NULL;
-		retval = dev->dqm->ops.create_queue(dev->dqm, q, &pdd->qpd);
+		retval = dev->dqm->ops.create_queue(dev->dqm, q, &pdd->qpd, q_data,
+						    restore_mqd, restore_ctl_stack);
 		print_queue(q);
 		break;
 
@@ -268,12 +417,13 @@ int pqm_create_queue(struct process_queue_manager *pqm,
 			goto err_create_queue;
 		}
 
-		retval = init_user_queue(pqm, dev, &q, properties, f, *qid);
+		retval = init_user_queue(pqm, dev, &q, properties, *qid);
 		if (retval != 0)
 			goto err_create_queue;
 		pqn->q = q;
 		pqn->kq = NULL;
-		retval = dev->dqm->ops.create_queue(dev->dqm, q, &pdd->qpd);
+		retval = dev->dqm->ops.create_queue(dev->dqm, q, &pdd->qpd, q_data,
+						    restore_mqd, restore_ctl_stack);
 		print_queue(q);
 		break;
 	case KFD_QUEUE_TYPE_DIQ:
@@ -285,6 +435,10 @@ int pqm_create_queue(struct process_queue_manager *pqm,
 		kq->queue->properties.queue_id = *qid;
 		pqn->kq = kq;
 		pqn->q = NULL;
+		retval = kfd_process_drain_interrupts(pdd);
+		if (retval)
+			break;
+
 		retval = dev->dqm->ops.create_kernel_queue(dev->dqm,
 							kq, &pdd->qpd);
 		break;
@@ -294,22 +448,33 @@ int pqm_create_queue(struct process_queue_manager *pqm,
 	}
 
 	if (retval != 0) {
-		pr_err("Pasid 0x%x DQM create queue type %d failed. ret %d\n",
-			pqm->process->pasid, type, retval);
+		if ((type == KFD_QUEUE_TYPE_SDMA ||
+		    type == KFD_QUEUE_TYPE_SDMA_XGMI ||
+		    type == KFD_QUEUE_TYPE_SDMA_BY_ENG_ID) &&
+		    retval == -ENOMEM)
+			pr_warn("process pid %d DQM create queue type %d failed. ret %d\n",
+				pqm->process->lead_thread->pid, type, retval);
+		else
+			pr_err("process pid %d DQM create queue type %d failed. ret %d\n",
+				pqm->process->lead_thread->pid, type, retval);
 		goto err_create_queue;
 	}
 
-	if (q && p_doorbell_offset_in_process)
+	if (q && p_doorbell_offset_in_process) {
 		/* Return the doorbell offset within the doorbell page
 		 * to the caller so it can be passed up to user mode
 		 * (in bytes).
-		 * There are always 1024 doorbells per process, so in case
-		 * of 8-byte doorbells, there are two doorbell pages per
-		 * process.
+		 * relative doorbell index = Absolute doorbell index -
+		 * absolute index of first doorbell in the page.
 		 */
-		*p_doorbell_offset_in_process =
-			(q->properties.doorbell_off * sizeof(uint32_t)) &
-			(kfd_doorbell_process_slice(dev) - 1);
+		uint32_t first_db_index = amdgpu_doorbell_index_on_bar(pdd->dev->adev,
+								       pdd->qpd.proc_doorbells,
+								       0,
+								       pdd->dev->kfd->device_info.doorbell_size);
+
+		*p_doorbell_offset_in_process = (q->properties.doorbell_off
+						- first_db_index) * sizeof(uint32_t);
+	}
 
 	pr_debug("PQM After DQM create queue\n");
 
@@ -326,7 +491,7 @@ int pqm_create_queue(struct process_queue_manager *pqm,
 err_create_queue:
 	uninit_queue(q);
 	if (kq)
-		kernel_queue_uninit(kq, false);
+		kernel_queue_uninit(kq);
 	kfree(pqn);
 err_allocate_pqn:
 	/* check if queues list is empty unregister process from device */
@@ -342,7 +507,7 @@ int pqm_destroy_queue(struct process_queue_manager *pqm, unsigned int qid)
 	struct process_queue_node *pqn;
 	struct kfd_process_device *pdd;
 	struct device_queue_manager *dqm;
-	struct kfd_dev *dev;
+	struct kfd_node *dev;
 	int retval;
 
 	dqm = NULL;
@@ -373,27 +538,26 @@ int pqm_destroy_queue(struct process_queue_manager *pqm, unsigned int qid)
 		/* destroy kernel queue (DIQ) */
 		dqm = pqn->kq->dev->dqm;
 		dqm->ops.destroy_kernel_queue(dqm, pqn->kq, &pdd->qpd);
-		kernel_queue_uninit(pqn->kq, false);
+		kernel_queue_uninit(pqn->kq);
 	}
 
 	if (pqn->q) {
-		kfd_procfs_del_queue(pqn->q);
+		retval = kfd_queue_unref_bo_vas(pdd, &pqn->q->properties);
+		if (retval)
+			goto err_destroy_queue;
+
 		dqm = pqn->q->device->dqm;
 		retval = dqm->ops.destroy_queue(dqm, &pdd->qpd, pqn->q);
 		if (retval) {
 			pr_err("Pasid 0x%x destroy queue %d failed, ret %d\n",
-				pqm->process->pasid,
+				pdd->pasid,
 				pqn->q->properties.queue_id, retval);
-			if (retval != -ETIME)
+			if (retval != -ETIME && retval != -EIO)
 				goto err_destroy_queue;
 		}
-
-		if (pqn->q->gws) {
-			amdgpu_amdkfd_remove_gws_from_process(pqm->process->kgd_process_info,
-				pqn->q->gws);
-			pdd->qpd.num_gws = 0;
-		}
-
+		kfd_procfs_del_queue(pqn->q);
+		kfd_queue_release_buffers(pdd, &pqn->q->properties);
+		pqm_clean_queue_resource(pqm, pqn);
 		uninit_queue(pqn->q);
 	}
 
@@ -416,15 +580,47 @@ int pqm_update_queue_properties(struct process_queue_manager *pqm,
 	struct process_queue_node *pqn;
 
 	pqn = get_queue_by_qid(pqm, qid);
-	if (!pqn) {
+	if (!pqn || !pqn->q) {
 		pr_debug("No queue %d exists for update operation\n", qid);
 		return -EFAULT;
 	}
 
+	/*
+	 * Update with NULL ring address is used to disable the queue
+	 */
+	if (p->queue_address && p->queue_size) {
+		struct kfd_process_device *pdd;
+		struct amdgpu_vm *vm;
+		struct queue *q = pqn->q;
+		int err;
+
+		pdd = kfd_get_process_device_data(q->device, q->process);
+		if (!pdd)
+			return -ENODEV;
+		vm = drm_priv_to_vm(pdd->drm_priv);
+		err = amdgpu_bo_reserve(vm->root.bo, false);
+		if (err)
+			return err;
+
+		if (kfd_queue_buffer_get(vm, (void *)p->queue_address, &p->ring_bo,
+					 p->queue_size)) {
+			pr_debug("ring buf 0x%llx size 0x%llx not mapped on GPU\n",
+				 p->queue_address, p->queue_size);
+			return -EFAULT;
+		}
+
+		kfd_queue_unref_bo_va(vm, &pqn->q->properties.ring_bo);
+		kfd_queue_buffer_put(&pqn->q->properties.ring_bo);
+		amdgpu_bo_unreserve(vm->root.bo);
+
+		pqn->q->properties.ring_bo = p->ring_bo;
+	}
+
 	pqn->q->properties.queue_address = p->queue_address;
 	pqn->q->properties.queue_size = p->queue_size;
 	pqn->q->properties.queue_percent = p->queue_percent;
 	pqn->q->properties.priority = p->priority;
+	pqn->q->properties.pm4_target_xcc = p->pm4_target_xcc;
 
 	retval = pqn->q->device->dqm->ops.update_queue(pqn->q->device->dqm,
 							pqn->q, NULL);
@@ -446,25 +642,34 @@ int pqm_update_mqd(struct process_queue_manager *pqm,
 		return -EFAULT;
 	}
 
+	/* CUs are masked for debugger requirements so deny user mask  */
+	if (pqn->q->properties.is_dbg_wa && minfo && minfo->cu_mask.ptr)
+		return -EBUSY;
+
+	/* ASICs that have WGPs must enforce pairwise enabled mask checks. */
+	if (minfo && minfo->cu_mask.ptr &&
+			KFD_GC_VERSION(pqn->q->device) >= IP_VERSION(10, 0, 0)) {
+		int i;
+
+		for (i = 0; i < minfo->cu_mask.count; i += 2) {
+			uint32_t cu_pair = (minfo->cu_mask.ptr[i / 32] >> (i % 32)) & 0x3;
+
+			if (cu_pair && cu_pair != 0x3) {
+				pr_debug("CUs must be adjacent pairwise enabled.\n");
+				return -EINVAL;
+			}
+		}
+	}
+
 	retval = pqn->q->device->dqm->ops.update_queue(pqn->q->device->dqm,
 							pqn->q, minfo);
 	if (retval != 0)
 		return retval;
 
-	return 0;
-}
-
-struct kernel_queue *pqm_get_kernel_queue(
-					struct process_queue_manager *pqm,
-					unsigned int qid)
-{
-	struct process_queue_node *pqn;
-
-	pqn = get_queue_by_qid(pqm, qid);
-	if (pqn && pqn->kq)
-		return pqn->kq;
+	if (minfo && minfo->cu_mask.ptr)
+		pqn->q->properties.is_user_cu_masked = true;
 
-	return NULL;
+	return 0;
 }
 
 struct queue *pqm_get_user_queue(struct process_queue_manager *pqm,
@@ -498,6 +703,403 @@ int pqm_get_wave_state(struct process_queue_manager *pqm,
 						       save_area_used_size);
 }
 
+int pqm_get_queue_snapshot(struct process_queue_manager *pqm,
+			   uint64_t exception_clear_mask,
+			   void __user *buf,
+			   int *num_qss_entries,
+			   uint32_t *entry_size)
+{
+	struct process_queue_node *pqn;
+	struct kfd_queue_snapshot_entry src;
+	uint32_t tmp_entry_size = *entry_size, tmp_qss_entries = *num_qss_entries;
+	int r = 0;
+
+	*num_qss_entries = 0;
+	if (!(*entry_size))
+		return -EINVAL;
+
+	*entry_size = min_t(size_t, *entry_size, sizeof(struct kfd_queue_snapshot_entry));
+	mutex_lock(&pqm->process->event_mutex);
+
+	memset(&src, 0, sizeof(src));
+
+	list_for_each_entry(pqn, &pqm->queues, process_queue_list) {
+		if (!pqn->q)
+			continue;
+
+		if (*num_qss_entries < tmp_qss_entries) {
+			set_queue_snapshot_entry(pqn->q, exception_clear_mask, &src);
+
+			if (copy_to_user(buf, &src, *entry_size)) {
+				r = -EFAULT;
+				break;
+			}
+			buf += tmp_entry_size;
+		}
+		*num_qss_entries += 1;
+	}
+
+	mutex_unlock(&pqm->process->event_mutex);
+	return r;
+}
+
+static int get_queue_data_sizes(struct kfd_process_device *pdd,
+				struct queue *q,
+				uint32_t *mqd_size,
+				uint32_t *ctl_stack_size)
+{
+	int ret;
+
+	ret = pqm_get_queue_checkpoint_info(&pdd->process->pqm,
+					    q->properties.queue_id,
+					    mqd_size,
+					    ctl_stack_size);
+	if (ret)
+		pr_err("Failed to get queue dump info (%d)\n", ret);
+
+	return ret;
+}
+
+int kfd_process_get_queue_info(struct kfd_process *p,
+			       uint32_t *num_queues,
+			       uint64_t *priv_data_sizes)
+{
+	uint32_t extra_data_sizes = 0;
+	struct queue *q;
+	int i;
+	int ret;
+
+	*num_queues = 0;
+
+	/* Run over all PDDs of the process */
+	for (i = 0; i < p->n_pdds; i++) {
+		struct kfd_process_device *pdd = p->pdds[i];
+
+		list_for_each_entry(q, &pdd->qpd.queues_list, list) {
+			if (q->properties.type == KFD_QUEUE_TYPE_COMPUTE ||
+				q->properties.type == KFD_QUEUE_TYPE_SDMA ||
+				q->properties.type == KFD_QUEUE_TYPE_SDMA_XGMI) {
+				uint32_t mqd_size, ctl_stack_size;
+
+				*num_queues = *num_queues + 1;
+
+				ret = get_queue_data_sizes(pdd, q, &mqd_size, &ctl_stack_size);
+				if (ret)
+					return ret;
+
+				extra_data_sizes += mqd_size + ctl_stack_size;
+			} else {
+				pr_err("Unsupported queue type (%d)\n", q->properties.type);
+				return -EOPNOTSUPP;
+			}
+		}
+	}
+	*priv_data_sizes = extra_data_sizes +
+				(*num_queues * sizeof(struct kfd_criu_queue_priv_data));
+
+	return 0;
+}
+
+static int pqm_checkpoint_mqd(struct process_queue_manager *pqm,
+			      unsigned int qid,
+			      void *mqd,
+			      void *ctl_stack)
+{
+	struct process_queue_node *pqn;
+
+	pqn = get_queue_by_qid(pqm, qid);
+	if (!pqn) {
+		pr_debug("amdkfd: No queue %d exists for operation\n", qid);
+		return -EFAULT;
+	}
+
+	if (!pqn->q->device->dqm->ops.checkpoint_mqd) {
+		pr_err("amdkfd: queue dumping not supported on this device\n");
+		return -EOPNOTSUPP;
+	}
+
+	return pqn->q->device->dqm->ops.checkpoint_mqd(pqn->q->device->dqm,
+						       pqn->q, mqd, ctl_stack);
+}
+
+static int criu_checkpoint_queue(struct kfd_process_device *pdd,
+			   struct queue *q,
+			   struct kfd_criu_queue_priv_data *q_data)
+{
+	uint8_t *mqd, *ctl_stack;
+	int ret;
+
+	mqd = (void *)(q_data + 1);
+	ctl_stack = mqd + q_data->mqd_size;
+
+	q_data->gpu_id = pdd->user_gpu_id;
+	q_data->type = q->properties.type;
+	q_data->format = q->properties.format;
+	q_data->q_id =  q->properties.queue_id;
+	q_data->q_address = q->properties.queue_address;
+	q_data->q_size = q->properties.queue_size;
+	q_data->priority = q->properties.priority;
+	q_data->q_percent = q->properties.queue_percent;
+	q_data->read_ptr_addr = (uint64_t)q->properties.read_ptr;
+	q_data->write_ptr_addr = (uint64_t)q->properties.write_ptr;
+	q_data->doorbell_id = q->doorbell_id;
+
+	q_data->sdma_id = q->sdma_id;
+
+	q_data->eop_ring_buffer_address =
+		q->properties.eop_ring_buffer_address;
+
+	q_data->eop_ring_buffer_size = q->properties.eop_ring_buffer_size;
+
+	q_data->ctx_save_restore_area_address =
+		q->properties.ctx_save_restore_area_address;
+
+	q_data->ctx_save_restore_area_size =
+		q->properties.ctx_save_restore_area_size;
+
+	q_data->gws = !!q->gws;
+
+	ret = pqm_checkpoint_mqd(&pdd->process->pqm, q->properties.queue_id, mqd, ctl_stack);
+	if (ret) {
+		pr_err("Failed checkpoint queue_mqd (%d)\n", ret);
+		return ret;
+	}
+
+	pr_debug("Dumping Queue: gpu_id:%x queue_id:%u\n", q_data->gpu_id, q_data->q_id);
+	return ret;
+}
+
+static int criu_checkpoint_queues_device(struct kfd_process_device *pdd,
+				   uint8_t __user *user_priv,
+				   unsigned int *q_index,
+				   uint64_t *queues_priv_data_offset)
+{
+	unsigned int q_private_data_size = 0;
+	uint8_t *q_private_data = NULL; /* Local buffer to store individual queue private data */
+	struct queue *q;
+	int ret = 0;
+
+	list_for_each_entry(q, &pdd->qpd.queues_list, list) {
+		struct kfd_criu_queue_priv_data *q_data;
+		uint64_t q_data_size;
+		uint32_t mqd_size;
+		uint32_t ctl_stack_size;
+
+		if (q->properties.type != KFD_QUEUE_TYPE_COMPUTE &&
+			q->properties.type != KFD_QUEUE_TYPE_SDMA &&
+			q->properties.type != KFD_QUEUE_TYPE_SDMA_XGMI) {
+
+			pr_err("Unsupported queue type (%d)\n", q->properties.type);
+			ret = -EOPNOTSUPP;
+			break;
+		}
+
+		ret = get_queue_data_sizes(pdd, q, &mqd_size, &ctl_stack_size);
+		if (ret)
+			break;
+
+		q_data_size = sizeof(*q_data) + mqd_size + ctl_stack_size;
+
+		/* Increase local buffer space if needed */
+		if (q_private_data_size < q_data_size) {
+			kfree(q_private_data);
+
+			q_private_data = kzalloc(q_data_size, GFP_KERNEL);
+			if (!q_private_data) {
+				ret = -ENOMEM;
+				break;
+			}
+			q_private_data_size = q_data_size;
+		}
+
+		q_data = (struct kfd_criu_queue_priv_data *)q_private_data;
+
+		/*
+		 * data stored in this order:
+		 * priv_data, mqd[xcc0], mqd[xcc1],..., ctl_stack[xcc0], ctl_stack[xcc1]...
+		 */
+		q_data->mqd_size = mqd_size;
+		q_data->ctl_stack_size = ctl_stack_size;
+
+		ret = criu_checkpoint_queue(pdd, q, q_data);
+		if (ret)
+			break;
+
+		q_data->object_type = KFD_CRIU_OBJECT_TYPE_QUEUE;
+
+		ret = copy_to_user(user_priv + *queues_priv_data_offset,
+				q_data, q_data_size);
+		if (ret) {
+			ret = -EFAULT;
+			break;
+		}
+		*queues_priv_data_offset += q_data_size;
+		*q_index = *q_index + 1;
+	}
+
+	kfree(q_private_data);
+
+	return ret;
+}
+
+int kfd_criu_checkpoint_queues(struct kfd_process *p,
+			 uint8_t __user *user_priv_data,
+			 uint64_t *priv_data_offset)
+{
+	int ret = 0, pdd_index, q_index = 0;
+
+	for (pdd_index = 0; pdd_index < p->n_pdds; pdd_index++) {
+		struct kfd_process_device *pdd = p->pdds[pdd_index];
+
+		/*
+		 * criu_checkpoint_queues_device will copy data to user and update q_index and
+		 * queues_priv_data_offset
+		 */
+		ret = criu_checkpoint_queues_device(pdd, user_priv_data, &q_index,
+					      priv_data_offset);
+
+		if (ret)
+			break;
+	}
+
+	return ret;
+}
+
+static void set_queue_properties_from_criu(struct queue_properties *qp,
+					  struct kfd_criu_queue_priv_data *q_data, uint32_t num_xcc)
+{
+	qp->is_interop = false;
+	qp->queue_percent = q_data->q_percent;
+	qp->priority = q_data->priority;
+	qp->queue_address = q_data->q_address;
+	qp->queue_size = q_data->q_size;
+	qp->read_ptr = (uint32_t *) q_data->read_ptr_addr;
+	qp->write_ptr = (uint32_t *) q_data->write_ptr_addr;
+	qp->eop_ring_buffer_address = q_data->eop_ring_buffer_address;
+	qp->eop_ring_buffer_size = q_data->eop_ring_buffer_size;
+	qp->ctx_save_restore_area_address = q_data->ctx_save_restore_area_address;
+	qp->ctx_save_restore_area_size = q_data->ctx_save_restore_area_size;
+	if (q_data->type == KFD_QUEUE_TYPE_COMPUTE)
+		qp->ctl_stack_size = q_data->ctl_stack_size / num_xcc;
+	else
+		qp->ctl_stack_size = q_data->ctl_stack_size;
+
+	qp->type = q_data->type;
+	qp->format = q_data->format;
+}
+
+int kfd_criu_restore_queue(struct kfd_process *p,
+			   uint8_t __user *user_priv_ptr,
+			   uint64_t *priv_data_offset,
+			   uint64_t max_priv_data_size)
+{
+	uint8_t *mqd, *ctl_stack, *q_extra_data = NULL;
+	struct kfd_criu_queue_priv_data *q_data;
+	struct kfd_process_device *pdd;
+	uint64_t q_extra_data_size;
+	struct queue_properties qp;
+	unsigned int queue_id;
+	int ret = 0;
+
+	if (*priv_data_offset + sizeof(*q_data) > max_priv_data_size)
+		return -EINVAL;
+
+	q_data = kmalloc(sizeof(*q_data), GFP_KERNEL);
+	if (!q_data)
+		return -ENOMEM;
+
+	ret = copy_from_user(q_data, user_priv_ptr + *priv_data_offset, sizeof(*q_data));
+	if (ret) {
+		ret = -EFAULT;
+		goto exit;
+	}
+
+	*priv_data_offset += sizeof(*q_data);
+	q_extra_data_size = (uint64_t)q_data->ctl_stack_size + q_data->mqd_size;
+
+	if (*priv_data_offset + q_extra_data_size > max_priv_data_size) {
+		ret = -EINVAL;
+		goto exit;
+	}
+
+	q_extra_data = kmalloc(q_extra_data_size, GFP_KERNEL);
+	if (!q_extra_data) {
+		ret = -ENOMEM;
+		goto exit;
+	}
+
+	ret = copy_from_user(q_extra_data, user_priv_ptr + *priv_data_offset, q_extra_data_size);
+	if (ret) {
+		ret = -EFAULT;
+		goto exit;
+	}
+
+	*priv_data_offset += q_extra_data_size;
+
+	pdd = kfd_process_device_data_by_id(p, q_data->gpu_id);
+	if (!pdd) {
+		pr_err("Failed to get pdd\n");
+		ret = -EINVAL;
+		goto exit;
+	}
+
+	/*
+	 * data stored in this order:
+	 * mqd[xcc0], mqd[xcc1],..., ctl_stack[xcc0], ctl_stack[xcc1]...
+	 */
+	mqd = q_extra_data;
+	ctl_stack = mqd + q_data->mqd_size;
+
+	memset(&qp, 0, sizeof(qp));
+	set_queue_properties_from_criu(&qp, q_data, NUM_XCC(pdd->dev->adev->gfx.xcc_mask));
+
+	print_queue_properties(&qp);
+
+	ret = pqm_create_queue(&p->pqm, pdd->dev, &qp, &queue_id, q_data, mqd, ctl_stack, NULL);
+	if (ret) {
+		pr_err("Failed to create new queue err:%d\n", ret);
+		goto exit;
+	}
+
+	if (q_data->gws)
+		ret = pqm_set_gws(&p->pqm, q_data->q_id, pdd->dev->gws);
+
+exit:
+	if (ret)
+		pr_err("Failed to restore queue (%d)\n", ret);
+	else
+		pr_debug("Queue id %d was restored successfully\n", queue_id);
+
+	kfree(q_data);
+	kfree(q_extra_data);
+
+	return ret;
+}
+
+int pqm_get_queue_checkpoint_info(struct process_queue_manager *pqm,
+				  unsigned int qid,
+				  uint32_t *mqd_size,
+				  uint32_t *ctl_stack_size)
+{
+	struct process_queue_node *pqn;
+
+	pqn = get_queue_by_qid(pqm, qid);
+	if (!pqn) {
+		pr_debug("amdkfd: No queue %d exists for operation\n", qid);
+		return -EFAULT;
+	}
+
+	if (!pqn->q->device->dqm->ops.get_queue_checkpoint_info) {
+		pr_err("amdkfd: queue dumping not supported on this device\n");
+		return -EOPNOTSUPP;
+	}
+
+	pqn->q->device->dqm->ops.get_queue_checkpoint_info(pqn->q->device->dqm,
+						       pqn->q, mqd_size,
+						       ctl_stack_size);
+	return 0;
+}
+
 #if defined(CONFIG_DEBUG_FS)
 
 int pqm_debugfs_mqds(struct seq_file *m, void *data)
@@ -507,7 +1109,9 @@ int pqm_debugfs_mqds(struct seq_file *m, void *data)
 	struct queue *q;
 	enum KFD_MQD_TYPE mqd_type;
 	struct mqd_manager *mqd_mgr;
-	int r = 0;
+	int r = 0, xcc, num_xccs = 1;
+	void *mqd;
+	uint64_t size = 0;
 
 	list_for_each_entry(pqn, &pqm->queues, process_queue_list) {
 		if (pqn->q) {
@@ -523,6 +1127,7 @@ int pqm_debugfs_mqds(struct seq_file *m, void *data)
 				seq_printf(m, "  Compute queue on device %x\n",
 					   q->device->id);
 				mqd_type = KFD_MQD_TYPE_CP;
+				num_xccs = NUM_XCC(q->device->xcc_mask);
 				break;
 			default:
 				seq_printf(m,
@@ -531,6 +1136,8 @@ int pqm_debugfs_mqds(struct seq_file *m, void *data)
 				continue;
 			}
 			mqd_mgr = q->device->dqm->mqd_mgrs[mqd_type];
+			size = mqd_mgr->mqd_stride(mqd_mgr,
+							&q->properties);
 		} else if (pqn->kq) {
 			q = pqn->kq->queue;
 			mqd_mgr = pqn->kq->mqd_mgr;
@@ -552,9 +1159,12 @@ int pqm_debugfs_mqds(struct seq_file *m, void *data)
 			continue;
 		}
 
-		r = mqd_mgr->debugfs_show_mqd(m, q->mqd);
-		if (r != 0)
-			break;
+		for (xcc = 0; xcc < num_xccs; xcc++) {
+			mqd = q->mqd + size * xcc;
+			r = mqd_mgr->debugfs_show_mqd(m, mqd);
+			if (r != 0)
+				break;
+		}
 	}
 
 	return r;
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_queue.c b/drivers/gpu/drm/amd/amdkfd/kfd_queue.c
index 6dcd621e5b71..f1e7583650c4 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_queue.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_queue.c
@@ -1,5 +1,6 @@
+// SPDX-License-Identifier: GPL-2.0 OR MIT
 /*
- * Copyright 2014 Advanced Micro Devices, Inc.
+ * Copyright 2014-2022 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -23,6 +24,8 @@
 
 #include <linux/slab.h>
 #include "kfd_priv.h"
+#include "kfd_topology.h"
+#include "kfd_svm.h"
 
 void print_queue_properties(struct queue_properties *q)
 {
@@ -81,3 +84,386 @@ void uninit_queue(struct queue *q)
 {
 	kfree(q);
 }
+
+#if IS_ENABLED(CONFIG_HSA_AMD_SVM)
+
+static int kfd_queue_buffer_svm_get(struct kfd_process_device *pdd, u64 addr, u64 size)
+{
+	struct kfd_process *p = pdd->process;
+	struct list_head update_list;
+	struct svm_range *prange;
+	int ret = -EINVAL;
+
+	INIT_LIST_HEAD(&update_list);
+	addr >>= PAGE_SHIFT;
+	size >>= PAGE_SHIFT;
+
+	mutex_lock(&p->svms.lock);
+
+	/*
+	 * range may split to multiple svm pranges aligned to granularity boundaery.
+	 */
+	while (size) {
+		uint32_t gpuid, gpuidx;
+		int r;
+
+		prange = svm_range_from_addr(&p->svms, addr, NULL);
+		if (!prange)
+			break;
+
+		if (!prange->mapped_to_gpu)
+			break;
+
+		r = kfd_process_gpuid_from_node(p, pdd->dev, &gpuid, &gpuidx);
+		if (r < 0)
+			break;
+		if (!test_bit(gpuidx, prange->bitmap_access) &&
+		    !test_bit(gpuidx, prange->bitmap_aip))
+			break;
+
+		if (!(prange->flags & KFD_IOCTL_SVM_FLAG_GPU_ALWAYS_MAPPED))
+			break;
+
+		list_add(&prange->update_list, &update_list);
+
+		if (prange->last - prange->start + 1 >= size) {
+			size = 0;
+			break;
+		}
+
+		size -= prange->last - prange->start + 1;
+		addr += prange->last - prange->start + 1;
+	}
+	if (size) {
+		pr_debug("[0x%llx 0x%llx] not registered\n", addr, addr + size - 1);
+		goto out_unlock;
+	}
+
+	list_for_each_entry(prange, &update_list, update_list)
+		atomic_inc(&prange->queue_refcount);
+	ret = 0;
+
+out_unlock:
+	mutex_unlock(&p->svms.lock);
+	return ret;
+}
+
+static void kfd_queue_buffer_svm_put(struct kfd_process_device *pdd, u64 addr, u64 size)
+{
+	struct kfd_process *p = pdd->process;
+	struct svm_range *prange, *pchild;
+	struct interval_tree_node *node;
+	unsigned long last;
+
+	addr >>= PAGE_SHIFT;
+	last = addr + (size >> PAGE_SHIFT) - 1;
+
+	mutex_lock(&p->svms.lock);
+
+	node = interval_tree_iter_first(&p->svms.objects, addr, last);
+	while (node) {
+		struct interval_tree_node *next_node;
+		unsigned long next_start;
+
+		prange = container_of(node, struct svm_range, it_node);
+		next_node = interval_tree_iter_next(node, addr, last);
+		next_start = min(node->last, last) + 1;
+
+		if (atomic_add_unless(&prange->queue_refcount, -1, 0)) {
+			list_for_each_entry(pchild, &prange->child_list, child_list)
+				atomic_add_unless(&pchild->queue_refcount, -1, 0);
+		}
+
+		node = next_node;
+		addr = next_start;
+	}
+
+	mutex_unlock(&p->svms.lock);
+}
+#else
+
+static int kfd_queue_buffer_svm_get(struct kfd_process_device *pdd, u64 addr, u64 size)
+{
+	return -EINVAL;
+}
+
+static void kfd_queue_buffer_svm_put(struct kfd_process_device *pdd, u64 addr, u64 size)
+{
+}
+
+#endif
+
+int kfd_queue_buffer_get(struct amdgpu_vm *vm, void __user *addr, struct amdgpu_bo **pbo,
+			 u64 expected_size)
+{
+	struct amdgpu_bo_va_mapping *mapping;
+	u64 user_addr;
+	u64 size;
+
+	user_addr = (u64)addr >> AMDGPU_GPU_PAGE_SHIFT;
+	size = expected_size >> AMDGPU_GPU_PAGE_SHIFT;
+
+	mapping = amdgpu_vm_bo_lookup_mapping(vm, user_addr);
+	if (!mapping)
+		goto out_err;
+
+	if (user_addr != mapping->start ||
+	    (size != 0 && user_addr + size - 1 != mapping->last)) {
+		pr_debug("expected size 0x%llx not equal to mapping addr 0x%llx size 0x%llx\n",
+			expected_size, mapping->start << AMDGPU_GPU_PAGE_SHIFT,
+			(mapping->last - mapping->start + 1) << AMDGPU_GPU_PAGE_SHIFT);
+		goto out_err;
+	}
+
+	*pbo = amdgpu_bo_ref(mapping->bo_va->base.bo);
+	mapping->bo_va->queue_refcount++;
+	return 0;
+
+out_err:
+	*pbo = NULL;
+	return -EINVAL;
+}
+
+/* FIXME: remove this function, just call amdgpu_bo_unref directly */
+void kfd_queue_buffer_put(struct amdgpu_bo **bo)
+{
+	amdgpu_bo_unref(bo);
+}
+
+int kfd_queue_acquire_buffers(struct kfd_process_device *pdd, struct queue_properties *properties)
+{
+	struct kfd_topology_device *topo_dev;
+	u64 expected_queue_size;
+	struct amdgpu_vm *vm;
+	u32 total_cwsr_size;
+	int err;
+
+	topo_dev = kfd_topology_device_by_id(pdd->dev->id);
+	if (!topo_dev)
+		return -EINVAL;
+
+	/* AQL queues on GFX7 and GFX8 appear twice their actual size */
+	if (properties->type == KFD_QUEUE_TYPE_COMPUTE &&
+	    properties->format == KFD_QUEUE_FORMAT_AQL &&
+	    topo_dev->node_props.gfx_target_version >= 70000 &&
+	    topo_dev->node_props.gfx_target_version < 90000)
+		expected_queue_size = properties->queue_size / 2;
+	else
+		expected_queue_size = properties->queue_size;
+
+	vm = drm_priv_to_vm(pdd->drm_priv);
+	err = amdgpu_bo_reserve(vm->root.bo, false);
+	if (err)
+		return err;
+
+	err = kfd_queue_buffer_get(vm, properties->write_ptr, &properties->wptr_bo, PAGE_SIZE);
+	if (err)
+		goto out_err_unreserve;
+
+	err = kfd_queue_buffer_get(vm, properties->read_ptr, &properties->rptr_bo, PAGE_SIZE);
+	if (err)
+		goto out_err_unreserve;
+
+	err = kfd_queue_buffer_get(vm, (void *)properties->queue_address,
+				   &properties->ring_bo, expected_queue_size);
+	if (err)
+		goto out_err_unreserve;
+
+	/* only compute queue requires EOP buffer and CWSR area */
+	if (properties->type != KFD_QUEUE_TYPE_COMPUTE)
+		goto out_unreserve;
+
+	/* EOP buffer is not required for all ASICs */
+	if (properties->eop_ring_buffer_address) {
+		if (properties->eop_ring_buffer_size != topo_dev->node_props.eop_buffer_size) {
+			pr_debug("queue eop bo size 0x%x not equal to node eop buf size 0x%x\n",
+				properties->eop_ring_buffer_size,
+				topo_dev->node_props.eop_buffer_size);
+			err = -EINVAL;
+			goto out_err_unreserve;
+		}
+		err = kfd_queue_buffer_get(vm, (void *)properties->eop_ring_buffer_address,
+					   &properties->eop_buf_bo,
+					   properties->eop_ring_buffer_size);
+		if (err)
+			goto out_err_unreserve;
+	}
+
+	if (properties->ctl_stack_size != topo_dev->node_props.ctl_stack_size) {
+		pr_debug("queue ctl stack size 0x%x not equal to node ctl stack size 0x%x\n",
+			properties->ctl_stack_size,
+			topo_dev->node_props.ctl_stack_size);
+		err = -EINVAL;
+		goto out_err_unreserve;
+	}
+
+	if (properties->ctx_save_restore_area_size < topo_dev->node_props.cwsr_size) {
+		pr_debug("queue cwsr size 0x%x not sufficient for node cwsr size 0x%x\n",
+			properties->ctx_save_restore_area_size,
+			topo_dev->node_props.cwsr_size);
+		err = -EINVAL;
+		goto out_err_unreserve;
+	}
+
+	total_cwsr_size = (properties->ctx_save_restore_area_size +
+			   topo_dev->node_props.debug_memory_size) * NUM_XCC(pdd->dev->xcc_mask);
+	total_cwsr_size = ALIGN(total_cwsr_size, PAGE_SIZE);
+
+	err = kfd_queue_buffer_get(vm, (void *)properties->ctx_save_restore_area_address,
+				   &properties->cwsr_bo, total_cwsr_size);
+	if (!err)
+		goto out_unreserve;
+
+	amdgpu_bo_unreserve(vm->root.bo);
+
+	err = kfd_queue_buffer_svm_get(pdd, properties->ctx_save_restore_area_address,
+				       total_cwsr_size);
+	if (err)
+		goto out_err_release;
+
+	return 0;
+
+out_unreserve:
+	amdgpu_bo_unreserve(vm->root.bo);
+	return 0;
+
+out_err_unreserve:
+	amdgpu_bo_unreserve(vm->root.bo);
+out_err_release:
+	/* FIXME: make a _locked version of this that can be called before
+	 * dropping the VM reservation.
+	 */
+	kfd_queue_unref_bo_vas(pdd, properties);
+	kfd_queue_release_buffers(pdd, properties);
+	return err;
+}
+
+int kfd_queue_release_buffers(struct kfd_process_device *pdd, struct queue_properties *properties)
+{
+	struct kfd_topology_device *topo_dev;
+	u32 total_cwsr_size;
+
+	kfd_queue_buffer_put(&properties->wptr_bo);
+	kfd_queue_buffer_put(&properties->rptr_bo);
+	kfd_queue_buffer_put(&properties->ring_bo);
+	kfd_queue_buffer_put(&properties->eop_buf_bo);
+	kfd_queue_buffer_put(&properties->cwsr_bo);
+
+	topo_dev = kfd_topology_device_by_id(pdd->dev->id);
+	if (!topo_dev)
+		return -EINVAL;
+	total_cwsr_size = (properties->ctx_save_restore_area_size +
+			   topo_dev->node_props.debug_memory_size) * NUM_XCC(pdd->dev->xcc_mask);
+	total_cwsr_size = ALIGN(total_cwsr_size, PAGE_SIZE);
+
+	kfd_queue_buffer_svm_put(pdd, properties->ctx_save_restore_area_address, total_cwsr_size);
+	return 0;
+}
+
+void kfd_queue_unref_bo_va(struct amdgpu_vm *vm, struct amdgpu_bo **bo)
+{
+	if (*bo) {
+		struct amdgpu_bo_va *bo_va;
+
+		bo_va = amdgpu_vm_bo_find(vm, *bo);
+		if (bo_va && bo_va->queue_refcount)
+			bo_va->queue_refcount--;
+	}
+}
+
+int kfd_queue_unref_bo_vas(struct kfd_process_device *pdd,
+			   struct queue_properties *properties)
+{
+	struct amdgpu_vm *vm;
+	int err;
+
+	vm = drm_priv_to_vm(pdd->drm_priv);
+	err = amdgpu_bo_reserve(vm->root.bo, false);
+	if (err)
+		return err;
+
+	kfd_queue_unref_bo_va(vm, &properties->wptr_bo);
+	kfd_queue_unref_bo_va(vm, &properties->rptr_bo);
+	kfd_queue_unref_bo_va(vm, &properties->ring_bo);
+	kfd_queue_unref_bo_va(vm, &properties->eop_buf_bo);
+	kfd_queue_unref_bo_va(vm, &properties->cwsr_bo);
+
+	amdgpu_bo_unreserve(vm->root.bo);
+	return 0;
+}
+
+#define SGPR_SIZE_PER_CU	0x4000
+#define LDS_SIZE_PER_CU		0x10000
+#define HWREG_SIZE_PER_CU	0x1000
+#define DEBUGGER_BYTES_ALIGN	64
+#define DEBUGGER_BYTES_PER_WAVE	32
+
+static u32 kfd_get_vgpr_size_per_cu(u32 gfxv)
+{
+	u32 vgpr_size = 0x40000;
+
+	if (gfxv == 90402 ||			/* GFX_VERSION_AQUA_VANJARAM */
+	    gfxv == 90010 ||			/* GFX_VERSION_ALDEBARAN */
+	    gfxv == 90008 ||			/* GFX_VERSION_ARCTURUS */
+	    gfxv == 90500)
+		vgpr_size = 0x80000;
+	else if (gfxv == 110000 ||		/* GFX_VERSION_PLUM_BONITO */
+		 gfxv == 110001 ||		/* GFX_VERSION_WHEAT_NAS */
+		 gfxv == 120000 ||		/* GFX_VERSION_GFX1200 */
+		 gfxv == 120001)		/* GFX_VERSION_GFX1201 */
+		vgpr_size = 0x60000;
+
+	return vgpr_size;
+}
+
+#define WG_CONTEXT_DATA_SIZE_PER_CU(gfxv, props)	\
+	(kfd_get_vgpr_size_per_cu(gfxv) + SGPR_SIZE_PER_CU +\
+	 (((gfxv) == 90500) ? (props->lds_size_in_kb << 10) : LDS_SIZE_PER_CU) +\
+	 HWREG_SIZE_PER_CU)
+
+#define CNTL_STACK_BYTES_PER_WAVE(gfxv)	\
+	((gfxv) >= 100100 ? 12 : 8)	/* GFX_VERSION_NAVI10*/
+
+#define SIZEOF_HSA_USER_CONTEXT_SAVE_AREA_HEADER 40
+
+void kfd_queue_ctx_save_restore_size(struct kfd_topology_device *dev)
+{
+	struct kfd_node_properties *props = &dev->node_props;
+	u32 gfxv = props->gfx_target_version;
+	u32 ctl_stack_size;
+	u32 wg_data_size;
+	u32 wave_num;
+	u32 cu_num;
+
+	if (gfxv < 80001)	/* GFX_VERSION_CARRIZO */
+		return;
+
+	cu_num = props->simd_count / props->simd_per_cu / NUM_XCC(dev->gpu->xcc_mask);
+	wave_num = (gfxv < 100100) ?	/* GFX_VERSION_NAVI10 */
+		    min(cu_num * 40, props->array_count / props->simd_arrays_per_engine * 512)
+		    : cu_num * 32;
+
+	wg_data_size = ALIGN(cu_num * WG_CONTEXT_DATA_SIZE_PER_CU(gfxv, props), PAGE_SIZE);
+	ctl_stack_size = wave_num * CNTL_STACK_BYTES_PER_WAVE(gfxv) + 8;
+	ctl_stack_size = ALIGN(SIZEOF_HSA_USER_CONTEXT_SAVE_AREA_HEADER + ctl_stack_size,
+			       PAGE_SIZE);
+
+	if ((gfxv / 10000 * 10000) == 100000) {
+		/* HW design limits control stack size to 0x7000.
+		 * This is insufficient for theoretical PM4 cases
+		 * but sufficient for AQL, limited by SPI events.
+		 */
+		ctl_stack_size = min(ctl_stack_size, 0x7000);
+	}
+
+	props->ctl_stack_size = ctl_stack_size;
+	props->debug_memory_size = ALIGN(wave_num * DEBUGGER_BYTES_PER_WAVE, DEBUGGER_BYTES_ALIGN);
+	props->cwsr_size = ctl_stack_size + wg_data_size;
+
+	if (gfxv == 80002)	/* GFX_VERSION_TONGA */
+		props->eop_buffer_size = 0x8000;
+	else if (gfxv == 90402)	/* GFX_VERSION_AQUA_VANJARAM */
+		props->eop_buffer_size = 4096;
+	else if (gfxv >= 80000)
+		props->eop_buffer_size = 4096;
+}
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_smi_events.c b/drivers/gpu/drm/amd/amdkfd/kfd_smi_events.c
index ed4bc5f844ce..a499449fcb06 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_smi_events.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_smi_events.c
@@ -1,5 +1,6 @@
+// SPDX-License-Identifier: GPL-2.0 OR MIT
 /*
- * Copyright 2020 Advanced Micro Devices, Inc.
+ * Copyright 2020-2022 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -28,6 +29,7 @@
 #include "amdgpu_vm.h"
 #include "kfd_priv.h"
 #include "kfd_smi_events.h"
+#include "amdgpu_reset.h"
 
 struct kfd_smi_client {
 	struct list_head list;
@@ -35,11 +37,14 @@ struct kfd_smi_client {
 	wait_queue_head_t wait_queue;
 	/* events enabled */
 	uint64_t events;
-	struct kfd_dev *dev;
+	struct kfd_node *dev;
 	spinlock_t lock;
+	struct rcu_head rcu;
+	pid_t pid;
+	bool suser;
 };
 
-#define MAX_KFIFO_SIZE	1024
+#define KFD_MAX_KFIFO_SIZE	8192
 
 static __poll_t kfd_smi_ev_poll(struct file *, struct poll_table_struct *);
 static ssize_t kfd_smi_ev_read(struct file *, char __user *, size_t, loff_t *);
@@ -81,7 +86,8 @@ static ssize_t kfd_smi_ev_read(struct file *filep, char __user *user,
 	struct kfd_smi_client *client = filep->private_data;
 	unsigned char *buf;
 
-	buf = kmalloc_array(MAX_KFIFO_SIZE, sizeof(*buf), GFP_KERNEL);
+	size = min_t(size_t, size, KFD_MAX_KFIFO_SIZE);
+	buf = kmalloc(size, GFP_KERNEL);
 	if (!buf)
 		return -ENOMEM;
 
@@ -95,7 +101,7 @@ static ssize_t kfd_smi_ev_read(struct file *filep, char __user *user,
 		ret = -EAGAIN;
 		goto ret_err;
 	}
-	to_copy = min3(size, sizeof(buf), to_copy);
+	to_copy = min(size, to_copy);
 	ret = kfifo_out(&client->fifo, buf, to_copy);
 	spin_unlock(&client->lock);
 	if (ret <= 0) {
@@ -133,32 +139,47 @@ static ssize_t kfd_smi_ev_write(struct file *filep, const char __user *user,
 	return sizeof(events);
 }
 
+static void kfd_smi_ev_client_free(struct rcu_head *p)
+{
+	struct kfd_smi_client *ev = container_of(p, struct kfd_smi_client, rcu);
+
+	kfifo_free(&ev->fifo);
+	kfree(ev);
+}
+
 static int kfd_smi_ev_release(struct inode *inode, struct file *filep)
 {
 	struct kfd_smi_client *client = filep->private_data;
-	struct kfd_dev *dev = client->dev;
+	struct kfd_node *dev = client->dev;
 
 	spin_lock(&dev->smi_lock);
 	list_del_rcu(&client->list);
 	spin_unlock(&dev->smi_lock);
 
-	synchronize_rcu();
-	kfifo_free(&client->fifo);
-	kfree(client);
-
+	call_rcu(&client->rcu, kfd_smi_ev_client_free);
 	return 0;
 }
 
-static void add_event_to_kfifo(struct kfd_dev *dev, unsigned int smi_event,
-			      char *event_msg, int len)
+static bool kfd_smi_ev_enabled(pid_t pid, struct kfd_smi_client *client,
+			       unsigned int event)
+{
+	uint64_t events = READ_ONCE(client->events);
+
+	if (pid && client->pid != pid && !client->suser)
+		return false;
+
+	return events & KFD_SMI_EVENT_MASK_FROM_INDEX(event);
+}
+
+static void add_event_to_kfifo(pid_t pid, struct kfd_node *dev,
+			       unsigned int smi_event, char *event_msg, int len)
 {
 	struct kfd_smi_client *client;
 
 	rcu_read_lock();
 
 	list_for_each_entry_rcu(client, &dev->smi_clients, list) {
-		if (!(READ_ONCE(client->events) &
-				KFD_SMI_EVENT_MASK_FROM_INDEX(smi_event)))
+		if (!kfd_smi_ev_enabled(pid, client, smi_event))
 			continue;
 		spin_lock(&client->lock);
 		if (kfifo_avail(&client->fifo) >= len) {
@@ -174,22 +195,31 @@ static void add_event_to_kfifo(struct kfd_dev *dev, unsigned int smi_event,
 	rcu_read_unlock();
 }
 
-void kfd_smi_event_update_gpu_reset(struct kfd_dev *dev, bool post_reset)
+__printf(4, 5)
+static void kfd_smi_event_add(pid_t pid, struct kfd_node *dev,
+			      unsigned int event, char *fmt, ...)
 {
-	/*
-	 * GpuReset msg = Reset seq number (incremented for
-	 * every reset message sent before GPU reset).
-	 * 1 byte event + 1 byte space + 8 bytes seq num +
-	 * 1 byte \n + 1 byte \0 = 12
-	 */
-	char fifo_in[12];
+	char fifo_in[KFD_SMI_EVENT_MSG_SIZE];
 	int len;
-	unsigned int event;
+	va_list args;
 
 	if (list_empty(&dev->smi_clients))
 		return;
 
-	memset(fifo_in, 0x0, sizeof(fifo_in));
+	len = snprintf(fifo_in, sizeof(fifo_in), "%x ", event);
+
+	va_start(args, fmt);
+	len += vsnprintf(fifo_in + len, sizeof(fifo_in) - len, fmt, args);
+	va_end(args);
+
+	add_event_to_kfifo(pid, dev, event, fifo_in, len);
+}
+
+void kfd_smi_event_update_gpu_reset(struct kfd_node *dev, bool post_reset,
+				    struct amdgpu_reset_context *reset_context)
+{
+	unsigned int event;
+	char reset_cause[64];
 
 	if (post_reset) {
 		event = KFD_SMI_EVENT_GPU_POST_RESET;
@@ -198,63 +228,144 @@ void kfd_smi_event_update_gpu_reset(struct kfd_dev *dev, bool post_reset)
 		++(dev->reset_seq_num);
 	}
 
-	len = snprintf(fifo_in, sizeof(fifo_in), "%x %x\n", event,
-						dev->reset_seq_num);
+	memset(reset_cause, 0, sizeof(reset_cause));
+
+	if (reset_context)
+		amdgpu_reset_get_desc(reset_context, reset_cause,
+				      sizeof(reset_cause));
 
-	add_event_to_kfifo(dev, event, fifo_in, len);
+	kfd_smi_event_add(0, dev, event, KFD_EVENT_FMT_UPDATE_GPU_RESET(
+			  dev->reset_seq_num, reset_cause));
 }
 
-void kfd_smi_event_update_thermal_throttling(struct kfd_dev *dev,
+void kfd_smi_event_update_thermal_throttling(struct kfd_node *dev,
 					     uint64_t throttle_bitmask)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)dev->kgd;
-	/*
-	 * ThermalThrottle msg = throttle_bitmask(8):
-	 * 			 thermal_interrupt_count(16):
-	 * 1 byte event + 1 byte space + 16 byte throttle_bitmask +
-	 * 1 byte : + 16 byte thermal_interupt_counter + 1 byte \n +
-	 * 1 byte \0 = 37
-	 */
-	char fifo_in[37];
-	int len;
+	kfd_smi_event_add(0, dev, KFD_SMI_EVENT_THERMAL_THROTTLE, KFD_EVENT_FMT_THERMAL_THROTTLING(
+			  throttle_bitmask,
+			  amdgpu_dpm_get_thermal_throttling_counter(dev->adev)));
+}
 
-	if (list_empty(&dev->smi_clients))
-		return;
+void kfd_smi_event_update_vmfault(struct kfd_node *dev, uint16_t pasid)
+{
+	struct amdgpu_task_info *task_info;
+
+	task_info = amdgpu_vm_get_task_info_pasid(dev->adev, pasid);
+	if (task_info) {
+		/* Report VM faults from user applications, not retry from kernel */
+		if (task_info->task.pid)
+			kfd_smi_event_add(0, dev, KFD_SMI_EVENT_VMFAULT, KFD_EVENT_FMT_VMFAULT(
+					  task_info->task.pid, task_info->task.comm));
+		amdgpu_vm_put_task_info(task_info);
+	}
+}
 
-	len = snprintf(fifo_in, sizeof(fifo_in), "%x %llx:%llx\n",
-		       KFD_SMI_EVENT_THERMAL_THROTTLE, throttle_bitmask,
-		       atomic64_read(&adev->smu.throttle_int_counter));
+void kfd_smi_event_page_fault_start(struct kfd_node *node, pid_t pid,
+				    unsigned long address, bool write_fault,
+				    ktime_t ts)
+{
+	kfd_smi_event_add(pid, node, KFD_SMI_EVENT_PAGE_FAULT_START,
+			  KFD_EVENT_FMT_PAGEFAULT_START(ktime_to_ns(ts), pid,
+			  address, node->id, write_fault ? 'W' : 'R'));
+}
 
-	add_event_to_kfifo(dev, KFD_SMI_EVENT_THERMAL_THROTTLE,	fifo_in, len);
+void kfd_smi_event_page_fault_end(struct kfd_node *node, pid_t pid,
+				  unsigned long address, bool migration)
+{
+	kfd_smi_event_add(pid, node, KFD_SMI_EVENT_PAGE_FAULT_END,
+			  KFD_EVENT_FMT_PAGEFAULT_END(ktime_get_boottime_ns(),
+			  pid, address, node->id, migration ? 'M' : 'U'));
 }
 
-void kfd_smi_event_update_vmfault(struct kfd_dev *dev, uint16_t pasid)
+void kfd_smi_event_migration_start(struct kfd_node *node, pid_t pid,
+				   unsigned long start, unsigned long end,
+				   uint32_t from, uint32_t to,
+				   uint32_t prefetch_loc, uint32_t preferred_loc,
+				   uint32_t trigger)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)dev->kgd;
-	struct amdgpu_task_info task_info;
-	/* VmFault msg = (hex)uint32_pid(8) + :(1) + task name(16) = 25 */
-	/* 1 byte event + 1 byte space + 25 bytes msg + 1 byte \n +
-	 * 1 byte \0 = 29
-	 */
-	char fifo_in[29];
-	int len;
+	kfd_smi_event_add(pid, node, KFD_SMI_EVENT_MIGRATE_START,
+			  KFD_EVENT_FMT_MIGRATE_START(
+			  ktime_get_boottime_ns(), pid, start, end - start,
+			  from, to, prefetch_loc, preferred_loc, trigger));
+}
 
-	if (list_empty(&dev->smi_clients))
+void kfd_smi_event_migration_end(struct kfd_node *node, pid_t pid,
+				 unsigned long start, unsigned long end,
+				 uint32_t from, uint32_t to, uint32_t trigger,
+				 int error_code)
+{
+	kfd_smi_event_add(pid, node, KFD_SMI_EVENT_MIGRATE_END,
+			  KFD_EVENT_FMT_MIGRATE_END(
+			  ktime_get_boottime_ns(), pid, start, end - start,
+			  from, to, trigger, error_code));
+}
+
+void kfd_smi_event_queue_eviction(struct kfd_node *node, pid_t pid,
+				  uint32_t trigger)
+{
+	kfd_smi_event_add(pid, node, KFD_SMI_EVENT_QUEUE_EVICTION,
+			  KFD_EVENT_FMT_QUEUE_EVICTION(ktime_get_boottime_ns(), pid,
+			  node->id, trigger));
+}
+
+void kfd_smi_event_queue_restore(struct kfd_node *node, pid_t pid)
+{
+	kfd_smi_event_add(pid, node, KFD_SMI_EVENT_QUEUE_RESTORE,
+			  KFD_EVENT_FMT_QUEUE_RESTORE(ktime_get_boottime_ns(), pid,
+			  node->id, 0));
+}
+
+void kfd_smi_event_queue_restore_rescheduled(struct mm_struct *mm)
+{
+	struct kfd_process *p;
+	int i;
+
+	p = kfd_lookup_process_by_mm(mm);
+	if (!p)
 		return;
 
-	memset(&task_info, 0, sizeof(struct amdgpu_task_info));
-	amdgpu_vm_get_task_info(adev, pasid, &task_info);
-	/* Report VM faults from user applications, not retry from kernel */
-	if (!task_info.pid)
+	for (i = 0; i < p->n_pdds; i++) {
+		struct kfd_process_device *pdd = p->pdds[i];
+
+		kfd_smi_event_add(p->lead_thread->pid, pdd->dev,
+				  KFD_SMI_EVENT_QUEUE_RESTORE,
+				  KFD_EVENT_FMT_QUEUE_RESTORE(ktime_get_boottime_ns(),
+				  p->lead_thread->pid, pdd->dev->id, 'R'));
+	}
+	kfd_unref_process(p);
+}
+
+void kfd_smi_event_unmap_from_gpu(struct kfd_node *node, pid_t pid,
+				  unsigned long address, unsigned long last,
+				  uint32_t trigger)
+{
+	kfd_smi_event_add(pid, node, KFD_SMI_EVENT_UNMAP_FROM_GPU,
+			  KFD_EVENT_FMT_UNMAP_FROM_GPU(ktime_get_boottime_ns(),
+			  pid, address, last - address + 1, node->id, trigger));
+}
+
+void kfd_smi_event_process(struct kfd_process_device *pdd, bool start)
+{
+	struct amdgpu_task_info *task_info;
+	struct amdgpu_vm *avm;
+
+	if (!pdd->drm_priv)
 		return;
 
-	len = snprintf(fifo_in, sizeof(fifo_in), "%x %x:%s\n", KFD_SMI_EVENT_VMFAULT,
-		task_info.pid, task_info.task_name);
+	avm = drm_priv_to_vm(pdd->drm_priv);
+	task_info = amdgpu_vm_get_task_info_vm(avm);
 
-	add_event_to_kfifo(dev, KFD_SMI_EVENT_VMFAULT, fifo_in, len);
+	if (task_info) {
+		kfd_smi_event_add(0, pdd->dev,
+				  start ? KFD_SMI_EVENT_PROCESS_START :
+				  KFD_SMI_EVENT_PROCESS_END,
+				  KFD_EVENT_FMT_PROCESS(task_info->task.pid,
+				  task_info->task.comm));
+		amdgpu_vm_put_task_info(task_info);
+	}
 }
 
-int kfd_smi_event_open(struct kfd_dev *dev, uint32_t *fd)
+int kfd_smi_event_open(struct kfd_node *dev, uint32_t *fd)
 {
 	struct kfd_smi_client *client;
 	int ret;
@@ -264,29 +375,37 @@ int kfd_smi_event_open(struct kfd_dev *dev, uint32_t *fd)
 		return -ENOMEM;
 	INIT_LIST_HEAD(&client->list);
 
-	ret = kfifo_alloc(&client->fifo, MAX_KFIFO_SIZE, GFP_KERNEL);
+	ret = kfifo_alloc(&client->fifo, KFD_MAX_KFIFO_SIZE, GFP_KERNEL);
 	if (ret) {
 		kfree(client);
 		return ret;
 	}
 
-	ret = anon_inode_getfd(kfd_smi_name, &kfd_smi_ev_fops, (void *)client,
-			       O_RDWR);
-	if (ret < 0) {
-		kfifo_free(&client->fifo);
-		kfree(client);
-		return ret;
-	}
-	*fd = ret;
-
 	init_waitqueue_head(&client->wait_queue);
 	spin_lock_init(&client->lock);
 	client->events = 0;
 	client->dev = dev;
+	client->pid = current->tgid;
+	client->suser = capable(CAP_SYS_ADMIN);
 
 	spin_lock(&dev->smi_lock);
 	list_add_rcu(&client->list, &dev->smi_clients);
 	spin_unlock(&dev->smi_lock);
 
+	ret = anon_inode_getfd(kfd_smi_name, &kfd_smi_ev_fops, (void *)client,
+			       O_RDWR);
+	if (ret < 0) {
+		spin_lock(&dev->smi_lock);
+		list_del_rcu(&client->list);
+		spin_unlock(&dev->smi_lock);
+
+		synchronize_rcu();
+
+		kfifo_free(&client->fifo);
+		kfree(client);
+		return ret;
+	}
+	*fd = ret;
+
 	return 0;
 }
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_smi_events.h b/drivers/gpu/drm/amd/amdkfd/kfd_smi_events.h
index bffd0c32b060..bb4d72b57387 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_smi_events.h
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_smi_events.h
@@ -1,5 +1,6 @@
+/* SPDX-License-Identifier: GPL-2.0 OR MIT */
 /*
- * Copyright 2020 Advanced Micro Devices, Inc.
+ * Copyright 2020-2022 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -23,10 +24,34 @@
 #ifndef KFD_SMI_EVENTS_H_INCLUDED
 #define KFD_SMI_EVENTS_H_INCLUDED
 
-int kfd_smi_event_open(struct kfd_dev *dev, uint32_t *fd);
-void kfd_smi_event_update_vmfault(struct kfd_dev *dev, uint16_t pasid);
-void kfd_smi_event_update_thermal_throttling(struct kfd_dev *dev,
-					     uint64_t throttle_bitmask);
-void kfd_smi_event_update_gpu_reset(struct kfd_dev *dev, bool post_reset);
+struct amdgpu_reset_context;
 
+int kfd_smi_event_open(struct kfd_node *dev, uint32_t *fd);
+void kfd_smi_event_update_vmfault(struct kfd_node *dev, uint16_t pasid);
+void kfd_smi_event_update_thermal_throttling(struct kfd_node *dev,
+					     uint64_t throttle_bitmask);
+void kfd_smi_event_update_gpu_reset(struct kfd_node *dev, bool post_reset,
+				    struct amdgpu_reset_context *reset_context);
+void kfd_smi_event_page_fault_start(struct kfd_node *node, pid_t pid,
+				    unsigned long address, bool write_fault,
+				    ktime_t ts);
+void kfd_smi_event_page_fault_end(struct kfd_node *node, pid_t pid,
+				  unsigned long address, bool migration);
+void kfd_smi_event_migration_start(struct kfd_node *node, pid_t pid,
+			     unsigned long start, unsigned long end,
+			     uint32_t from, uint32_t to,
+			     uint32_t prefetch_loc, uint32_t preferred_loc,
+			     uint32_t trigger);
+void kfd_smi_event_migration_end(struct kfd_node *node, pid_t pid,
+			     unsigned long start, unsigned long end,
+			     uint32_t from, uint32_t to, uint32_t trigger,
+			     int error_code);
+void kfd_smi_event_queue_eviction(struct kfd_node *node, pid_t pid,
+				  uint32_t trigger);
+void kfd_smi_event_queue_restore(struct kfd_node *node, pid_t pid);
+void kfd_smi_event_queue_restore_rescheduled(struct mm_struct *mm);
+void kfd_smi_event_unmap_from_gpu(struct kfd_node *node, pid_t pid,
+				  unsigned long address, unsigned long last,
+				  uint32_t trigger);
+void kfd_smi_event_process(struct kfd_process_device *pdd, bool start);
 #endif
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
index 3cb4681c5f53..97c2270f278f 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
@@ -23,15 +23,20 @@
 
 #include <linux/types.h>
 #include <linux/sched/task.h>
+#include <linux/dynamic_debug.h>
+#include <drm/ttm/ttm_tt.h>
+#include <drm/drm_exec.h>
+
 #include "amdgpu_sync.h"
 #include "amdgpu_object.h"
 #include "amdgpu_vm.h"
-#include "amdgpu_mn.h"
+#include "amdgpu_hmm.h"
 #include "amdgpu.h"
 #include "amdgpu_xgmi.h"
 #include "kfd_priv.h"
 #include "kfd_svm.h"
 #include "kfd_migrate.h"
+#include "kfd_smi_events.h"
 
 #ifdef dev_fmt
 #undef dev_fmt
@@ -43,7 +48,25 @@
 /* Long enough to ensure no retry fault comes after svm range is restored and
  * page table is updated.
  */
-#define AMDGPU_SVM_RANGE_RETRY_FAULT_PENDING	2000
+#define AMDGPU_SVM_RANGE_RETRY_FAULT_PENDING	(2UL * NSEC_PER_MSEC)
+#if IS_ENABLED(CONFIG_DYNAMIC_DEBUG)
+#define dynamic_svm_range_dump(svms) \
+	_dynamic_func_call_no_desc("svm_range_dump", svm_range_debug_dump, svms)
+#else
+#define dynamic_svm_range_dump(svms) \
+	do { if (0) svm_range_debug_dump(svms); } while (0)
+#endif
+
+/* Giant svm range split into smaller ranges based on this, it is decided using
+ * minimum of all dGPU/APU 1/32 VRAM size, between 2MB to 1GB and alignment to
+ * power of 2MB.
+ */
+static uint64_t max_svm_range_pages;
+
+struct criu_svm_metadata {
+	struct list_head list;
+	struct kfd_criu_svm_range_priv_data data;
+};
 
 static void svm_range_evict_svm_bo_worker(struct work_struct *work);
 static bool
@@ -107,7 +130,7 @@ static void svm_range_add_to_svms(struct svm_range *prange)
 	pr_debug("svms 0x%p prange 0x%p [0x%lx 0x%lx]\n", prange->svms,
 		 prange, prange->start, prange->last);
 
-	list_add_tail(&prange->list, &prange->svms->list);
+	list_move_tail(&prange->list, &prange->svms->list);
 	prange->it_node.start = prange->start;
 	prange->it_node.last = prange->last;
 	interval_tree_insert(&prange->it_node, &prange->svms->objects);
@@ -144,8 +167,7 @@ svm_range_dma_map_dev(struct amdgpu_device *adev, struct svm_range *prange,
 	int i, r;
 
 	if (!addr) {
-		addr = kvmalloc_array(prange->npages, sizeof(*addr),
-				      GFP_KERNEL | __GFP_ZERO);
+		addr = kvcalloc(prange->npages, sizeof(*addr), GFP_KERNEL);
 		if (!addr)
 			return -ENOMEM;
 		prange->dma_addr[gpuidx] = addr;
@@ -158,12 +180,11 @@ svm_range_dma_map_dev(struct amdgpu_device *adev, struct svm_range *prange,
 
 		page = hmm_pfn_to_page(hmm_pfns[i]);
 		if (is_zone_device_page(page)) {
-			struct amdgpu_device *bo_adev =
-					amdgpu_ttm_adev(prange->svm_bo->bo->tbo.bdev);
+			struct amdgpu_device *bo_adev = prange->svm_bo->node->adev;
 
 			addr[i] = (hmm_pfns[i] << PAGE_SHIFT) +
 				   bo_adev->vm_manager.vram_base_offset -
-				   bo_adev->kfd.dev->pgmap.range.start;
+				   bo_adev->kfd.pgmap.range.start;
 			addr[i] |= SVM_RANGE_VRAM_DOMAIN;
 			pr_debug_ratelimited("vram address: 0x%llx\n", addr[i]);
 			continue;
@@ -177,6 +198,7 @@ svm_range_dma_map_dev(struct amdgpu_device *adev, struct svm_range *prange,
 		pr_debug_ratelimited("dma mapping 0x%llx for page addr 0x%lx\n",
 				     addr[i] >> PAGE_SHIFT, page_to_pfn(page));
 	}
+
 	return 0;
 }
 
@@ -193,7 +215,6 @@ svm_range_dma_map(struct svm_range *prange, unsigned long *bitmap,
 
 	for_each_set_bit(gpuidx, bitmap, MAX_GPU_INSTANCE) {
 		struct kfd_process_device *pdd;
-		struct amdgpu_device *adev;
 
 		pr_debug("mapping to gpu idx 0x%x\n", gpuidx);
 		pdd = kfd_process_device_from_gpuidx(p, gpuidx);
@@ -201,9 +222,8 @@ svm_range_dma_map(struct svm_range *prange, unsigned long *bitmap,
 			pr_debug("failed to find device idx %d\n", gpuidx);
 			return -EINVAL;
 		}
-		adev = (struct amdgpu_device *)pdd->dev->kgd;
 
-		r = svm_range_dma_map_dev(adev, prange, offset, npages,
+		r = svm_range_dma_map_dev(pdd->dev->adev, prange, offset, npages,
 					  hmm_pfns, gpuidx);
 		if (r)
 			break;
@@ -212,7 +232,7 @@ svm_range_dma_map(struct svm_range *prange, unsigned long *bitmap,
 	return r;
 }
 
-void svm_range_dma_unmap(struct device *dev, dma_addr_t *dma_addr,
+void svm_range_dma_unmap_dev(struct device *dev, dma_addr_t *dma_addr,
 			 unsigned long offset, unsigned long npages)
 {
 	enum dma_data_direction dir = DMA_BIDIRECTIONAL;
@@ -230,7 +250,7 @@ void svm_range_dma_unmap(struct device *dev, dma_addr_t *dma_addr,
 	}
 }
 
-void svm_range_free_dma_mappings(struct svm_range *prange)
+void svm_range_dma_unmap(struct svm_range *prange)
 {
 	struct kfd_process_device *pdd;
 	dma_addr_t *dma_addr;
@@ -250,39 +270,59 @@ void svm_range_free_dma_mappings(struct svm_range *prange)
 			pr_debug("failed to find device idx %d\n", gpuidx);
 			continue;
 		}
-		dev = &pdd->dev->pdev->dev;
-		svm_range_dma_unmap(dev, dma_addr, 0, prange->npages);
-		kvfree(dma_addr);
-		prange->dma_addr[gpuidx] = NULL;
+		dev = &pdd->dev->adev->pdev->dev;
+
+		svm_range_dma_unmap_dev(dev, dma_addr, 0, prange->npages);
 	}
 }
 
-static void svm_range_free(struct svm_range *prange)
+static void svm_range_free(struct svm_range *prange, bool do_unmap)
 {
+	uint64_t size = (prange->last - prange->start + 1) << PAGE_SHIFT;
+	struct kfd_process *p = container_of(prange->svms, struct kfd_process, svms);
+	uint32_t gpuidx;
+
 	pr_debug("svms 0x%p prange 0x%p [0x%lx 0x%lx]\n", prange->svms, prange,
 		 prange->start, prange->last);
 
 	svm_range_vram_node_free(prange);
-	svm_range_free_dma_mappings(prange);
+	if (do_unmap)
+		svm_range_dma_unmap(prange);
+
+	if (do_unmap && !p->xnack_enabled) {
+		pr_debug("unreserve prange 0x%p size: 0x%llx\n", prange, size);
+		amdgpu_amdkfd_unreserve_mem_limit(NULL, size,
+					KFD_IOC_ALLOC_MEM_FLAGS_USERPTR, 0);
+	}
+
+	/* free dma_addr array for each gpu */
+	for (gpuidx = 0; gpuidx < MAX_GPU_INSTANCE; gpuidx++) {
+		if (prange->dma_addr[gpuidx]) {
+			kvfree(prange->dma_addr[gpuidx]);
+			prange->dma_addr[gpuidx] = NULL;
+		}
+	}
+
 	mutex_destroy(&prange->lock);
 	mutex_destroy(&prange->migrate_mutex);
 	kfree(prange);
 }
 
 static void
-svm_range_set_default_attributes(int32_t *location, int32_t *prefetch_loc,
-				 uint8_t *granularity, uint32_t *flags)
+svm_range_set_default_attributes(struct svm_range_list *svms, int32_t *location,
+				 int32_t *prefetch_loc, uint8_t *granularity,
+				 uint32_t *flags)
 {
 	*location = KFD_IOCTL_SVM_LOCATION_UNDEFINED;
 	*prefetch_loc = KFD_IOCTL_SVM_LOCATION_UNDEFINED;
-	*granularity = 9;
+	*granularity = svms->default_granularity;
 	*flags =
 		KFD_IOCTL_SVM_FLAG_HOST_ACCESS | KFD_IOCTL_SVM_FLAG_COHERENT;
 }
 
 static struct
 svm_range *svm_range_new(struct svm_range_list *svms, uint64_t start,
-			 uint64_t last)
+			 uint64_t last, bool update_mem_usage)
 {
 	uint64_t size = last - start + 1;
 	struct svm_range *prange;
@@ -291,28 +331,35 @@ svm_range *svm_range_new(struct svm_range_list *svms, uint64_t start,
 	prange = kzalloc(sizeof(*prange), GFP_KERNEL);
 	if (!prange)
 		return NULL;
+
+	p = container_of(svms, struct kfd_process, svms);
+	if (!p->xnack_enabled && update_mem_usage &&
+	    amdgpu_amdkfd_reserve_mem_limit(NULL, size << PAGE_SHIFT,
+				    KFD_IOC_ALLOC_MEM_FLAGS_USERPTR, 0)) {
+		pr_info("SVM mapping failed, exceeds resident system memory limit\n");
+		kfree(prange);
+		return NULL;
+	}
 	prange->npages = size;
 	prange->svms = svms;
 	prange->start = start;
 	prange->last = last;
 	INIT_LIST_HEAD(&prange->list);
 	INIT_LIST_HEAD(&prange->update_list);
-	INIT_LIST_HEAD(&prange->remove_list);
-	INIT_LIST_HEAD(&prange->insert_list);
 	INIT_LIST_HEAD(&prange->svm_bo_list);
 	INIT_LIST_HEAD(&prange->deferred_list);
 	INIT_LIST_HEAD(&prange->child_list);
 	atomic_set(&prange->invalid, 0);
 	prange->validate_timestamp = 0;
+	prange->vram_pages = 0;
 	mutex_init(&prange->migrate_mutex);
 	mutex_init(&prange->lock);
 
-	p = container_of(svms, struct kfd_process, svms);
 	if (p->xnack_enabled)
 		bitmap_copy(prange->bitmap_access, svms->bitmap_supported,
 			    MAX_GPU_INSTANCE);
 
-	svm_range_set_default_attributes(&prange->preferred_loc,
+	svm_range_set_default_attributes(svms, &prange->preferred_loc,
 					 &prange->prefetch_loc,
 					 &prange->granularity, &prange->flags);
 
@@ -334,6 +381,8 @@ static void svm_range_bo_release(struct kref *kref)
 	struct svm_range_bo *svm_bo;
 
 	svm_bo = container_of(kref, struct svm_range_bo, kref);
+	pr_debug("svm_bo 0x%p\n", svm_bo);
+
 	spin_lock(&svm_bo->list_lock);
 	while (!list_empty(&svm_bo->range_list)) {
 		struct svm_range *prange =
@@ -349,37 +398,74 @@ static void svm_range_bo_release(struct kref *kref)
 			 prange->start, prange->last);
 		mutex_lock(&prange->lock);
 		prange->svm_bo = NULL;
+		/* prange should not hold vram page now */
+		WARN_ONCE(prange->actual_loc, "prange should not hold vram page");
 		mutex_unlock(&prange->lock);
 
 		spin_lock(&svm_bo->list_lock);
 	}
 	spin_unlock(&svm_bo->list_lock);
-	if (!dma_fence_is_signaled(&svm_bo->eviction_fence->base)) {
-		/* We're not in the eviction worker.
-		 * Signal the fence and synchronize with any
-		 * pending eviction work.
+
+	if (mmget_not_zero(svm_bo->eviction_fence->mm)) {
+		struct kfd_process_device *pdd;
+		struct kfd_process *p;
+		struct mm_struct *mm;
+
+		mm = svm_bo->eviction_fence->mm;
+		/*
+		 * The forked child process takes svm_bo device pages ref, svm_bo could be
+		 * released after parent process is gone.
 		 */
-		dma_fence_signal(&svm_bo->eviction_fence->base);
-		cancel_work_sync(&svm_bo->eviction_work);
+		p = kfd_lookup_process_by_mm(mm);
+		if (p) {
+			pdd = kfd_get_process_device_data(svm_bo->node, p);
+			if (pdd)
+				atomic64_sub(amdgpu_bo_size(svm_bo->bo), &pdd->vram_usage);
+			kfd_unref_process(p);
+		}
+		mmput(mm);
 	}
+
+	if (!dma_fence_is_signaled(&svm_bo->eviction_fence->base))
+		/* We're not in the eviction worker. Signal the fence. */
+		dma_fence_signal(&svm_bo->eviction_fence->base);
 	dma_fence_put(&svm_bo->eviction_fence->base);
 	amdgpu_bo_unref(&svm_bo->bo);
 	kfree(svm_bo);
 }
 
-void svm_range_bo_unref(struct svm_range_bo *svm_bo)
+static void svm_range_bo_wq_release(struct work_struct *work)
 {
-	if (!svm_bo)
-		return;
+	struct svm_range_bo *svm_bo;
 
-	kref_put(&svm_bo->kref, svm_range_bo_release);
+	svm_bo = container_of(work, struct svm_range_bo, release_work);
+	svm_range_bo_release(&svm_bo->kref);
 }
 
-static bool
-svm_range_validate_svm_bo(struct amdgpu_device *adev, struct svm_range *prange)
+static void svm_range_bo_release_async(struct kref *kref)
 {
-	struct amdgpu_device *bo_adev;
+	struct svm_range_bo *svm_bo;
 
+	svm_bo = container_of(kref, struct svm_range_bo, kref);
+	pr_debug("svm_bo 0x%p\n", svm_bo);
+	INIT_WORK(&svm_bo->release_work, svm_range_bo_wq_release);
+	schedule_work(&svm_bo->release_work);
+}
+
+void svm_range_bo_unref_async(struct svm_range_bo *svm_bo)
+{
+	kref_put(&svm_bo->kref, svm_range_bo_release_async);
+}
+
+static void svm_range_bo_unref(struct svm_range_bo *svm_bo)
+{
+	if (svm_bo)
+		kref_put(&svm_bo->kref, svm_range_bo_release);
+}
+
+static bool
+svm_range_validate_svm_bo(struct kfd_node *node, struct svm_range *prange)
+{
 	mutex_lock(&prange->lock);
 	if (!prange->svm_bo) {
 		mutex_unlock(&prange->lock);
@@ -392,12 +478,11 @@ svm_range_validate_svm_bo(struct amdgpu_device *adev, struct svm_range *prange)
 	}
 	if (svm_bo_ref_unless_zero(prange->svm_bo)) {
 		/*
-		 * Migrate from GPU to GPU, remove range from source bo_adev
-		 * svm_bo range list, and return false to allocate svm_bo from
-		 * destination adev.
+		 * Migrate from GPU to GPU, remove range from source svm_bo->node
+		 * range list, and return false to allocate svm_bo from destination
+		 * node.
 		 */
-		bo_adev = amdgpu_ttm_adev(prange->svm_bo->bo->tbo.bdev);
-		if (bo_adev != adev) {
+		if (prange->svm_bo->node != node) {
 			mutex_unlock(&prange->lock);
 
 			spin_lock(&prange->svm_bo->list_lock);
@@ -440,11 +525,11 @@ svm_range_validate_svm_bo(struct amdgpu_device *adev, struct svm_range *prange)
 
 	/* We need a new svm_bo. Spin-loop to wait for concurrent
 	 * svm_range_bo_release to finish removing this range from
-	 * its range list. After this, it is safe to reuse the
-	 * svm_bo pointer and svm_bo_list head.
+	 * its range list and set prange->svm_bo to null. After this,
+	 * it is safe to reuse the svm_bo pointer and svm_bo_list head.
 	 */
-	while (!list_empty_careful(&prange->svm_bo_list))
-		;
+	while (!list_empty_careful(&prange->svm_bo_list) || prange->svm_bo)
+		cond_resched();
 
 	return false;
 }
@@ -465,9 +550,10 @@ static struct svm_range_bo *svm_range_bo_new(void)
 }
 
 int
-svm_range_vram_node_new(struct amdgpu_device *adev, struct svm_range *prange,
+svm_range_vram_node_new(struct kfd_node *node, struct svm_range *prange,
 			bool clear)
 {
+	struct kfd_process_device *pdd;
 	struct amdgpu_bo_param bp;
 	struct svm_range_bo *svm_bo;
 	struct amdgpu_bo_user *ubo;
@@ -477,10 +563,11 @@ svm_range_vram_node_new(struct amdgpu_device *adev, struct svm_range *prange,
 	int r;
 
 	p = container_of(prange->svms, struct kfd_process, svms);
-	pr_debug("pasid: %x svms 0x%p [0x%lx 0x%lx]\n", p->pasid, prange->svms,
+	pr_debug("process pid: %d svms 0x%p [0x%lx 0x%lx]\n",
+		 p->lead_thread->pid, prange->svms,
 		 prange->start, prange->last);
 
-	if (svm_range_validate_svm_bo(adev, prange))
+	if (svm_range_validate_svm_bo(node, prange))
 		return 0;
 
 	svm_bo = svm_range_bo_new();
@@ -494,7 +581,7 @@ svm_range_vram_node_new(struct amdgpu_device *adev, struct svm_range *prange,
 		kfree(svm_bo);
 		return -ESRCH;
 	}
-	svm_bo->svms = prange->svms;
+	svm_bo->node = node;
 	svm_bo->eviction_fence =
 		amdgpu_amdkfd_fence_create(dma_fence_context_alloc(1),
 					   mm,
@@ -508,23 +595,39 @@ svm_range_vram_node_new(struct amdgpu_device *adev, struct svm_range *prange,
 	bp.domain = AMDGPU_GEM_DOMAIN_VRAM;
 	bp.flags = AMDGPU_GEM_CREATE_NO_CPU_ACCESS;
 	bp.flags |= clear ? AMDGPU_GEM_CREATE_VRAM_CLEARED : 0;
-	bp.flags |= AMDGPU_AMDKFD_CREATE_SVM_BO;
+	bp.flags |= AMDGPU_GEM_CREATE_DISCARDABLE;
 	bp.type = ttm_bo_type_device;
 	bp.resv = NULL;
+	if (node->xcp)
+		bp.xcp_id_plus1 = node->xcp->id + 1;
 
-	r = amdgpu_bo_create_user(adev, &bp, &ubo);
+	r = amdgpu_bo_create_user(node->adev, &bp, &ubo);
 	if (r) {
 		pr_debug("failed %d to create bo\n", r);
 		goto create_bo_failed;
 	}
 	bo = &ubo->bo;
+
+	pr_debug("alloc bo at offset 0x%lx size 0x%lx on partition %d\n",
+		 bo->tbo.resource->start << PAGE_SHIFT, bp.size,
+		 bp.xcp_id_plus1 - 1);
+
 	r = amdgpu_bo_reserve(bo, true);
 	if (r) {
 		pr_debug("failed %d to reserve bo\n", r);
 		goto reserve_bo_failed;
 	}
 
-	r = dma_resv_reserve_shared(bo->tbo.base.resv, 1);
+	if (clear) {
+		r = amdgpu_bo_sync_wait(bo, AMDGPU_FENCE_OWNER_KFD, false);
+		if (r) {
+			pr_debug("failed %d to sync bo\n", r);
+			amdgpu_bo_unreserve(bo);
+			goto reserve_bo_failed;
+		}
+	}
+
+	r = dma_resv_reserve_fences(bo->tbo.base.resv, 1);
 	if (r) {
 		pr_debug("failed %d to reserve bo\n", r);
 		amdgpu_bo_unreserve(bo);
@@ -543,6 +646,10 @@ svm_range_vram_node_new(struct amdgpu_device *adev, struct svm_range *prange,
 	list_add(&prange->svm_bo_list, &svm_bo->range_list);
 	spin_unlock(&svm_bo->list_lock);
 
+	pdd = svm_range_get_pdd_by_node(prange, node);
+	if (pdd)
+		atomic64_add(amdgpu_bo_size(bo), &pdd->vram_usage);
+
 	return 0;
 
 reserve_bo_failed:
@@ -557,49 +664,41 @@ create_bo_failed:
 
 void svm_range_vram_node_free(struct svm_range *prange)
 {
-	svm_range_bo_unref(prange->svm_bo);
-	prange->ttm_res = NULL;
+	/* serialize prange->svm_bo unref */
+	mutex_lock(&prange->lock);
+	/* prange->svm_bo has not been unref */
+	if (prange->ttm_res) {
+		prange->ttm_res = NULL;
+		mutex_unlock(&prange->lock);
+		svm_range_bo_unref(prange->svm_bo);
+	} else
+		mutex_unlock(&prange->lock);
 }
 
-struct amdgpu_device *
-svm_range_get_adev_by_id(struct svm_range *prange, uint32_t gpu_id)
+struct kfd_node *
+svm_range_get_node_by_id(struct svm_range *prange, uint32_t gpu_id)
 {
-	struct kfd_process_device *pdd;
 	struct kfd_process *p;
-	int32_t gpu_idx;
+	struct kfd_process_device *pdd;
 
 	p = container_of(prange->svms, struct kfd_process, svms);
-
-	gpu_idx = kfd_process_gpuidx_from_gpuid(p, gpu_id);
-	if (gpu_idx < 0) {
-		pr_debug("failed to get device by id 0x%x\n", gpu_id);
-		return NULL;
-	}
-	pdd = kfd_process_device_from_gpuidx(p, gpu_idx);
+	pdd = kfd_process_device_data_by_id(p, gpu_id);
 	if (!pdd) {
-		pr_debug("failed to get device by idx 0x%x\n", gpu_idx);
+		pr_debug("failed to get kfd process device by id 0x%x\n", gpu_id);
 		return NULL;
 	}
 
-	return (struct amdgpu_device *)pdd->dev->kgd;
+	return pdd->dev;
 }
 
 struct kfd_process_device *
-svm_range_get_pdd_by_adev(struct svm_range *prange, struct amdgpu_device *adev)
+svm_range_get_pdd_by_node(struct svm_range *prange, struct kfd_node *node)
 {
 	struct kfd_process *p;
-	int32_t gpu_idx, gpuid;
-	int r;
 
 	p = container_of(prange->svms, struct kfd_process, svms);
 
-	r = kfd_process_gpuid_from_kgd(p, adev, &gpuid, &gpu_idx);
-	if (r) {
-		pr_debug("failed to get device id by adev %p\n", adev);
-		return NULL;
-	}
-
-	return kfd_process_device_from_gpuidx(p, gpu_idx);
+	return kfd_get_process_device_data(node, p);
 }
 
 static int svm_range_bo_validate(void *param, struct amdgpu_bo *bo)
@@ -662,7 +761,8 @@ svm_range_check_attr(struct kfd_process *p,
 
 static void
 svm_range_apply_attrs(struct kfd_process *p, struct svm_range *prange,
-		      uint32_t nattr, struct kfd_ioctl_svm_attribute *attrs)
+		      uint32_t nattr, struct kfd_ioctl_svm_attribute *attrs,
+		      bool *update_mapping)
 {
 	uint32_t i;
 	int gpuidx;
@@ -678,6 +778,9 @@ svm_range_apply_attrs(struct kfd_process *p, struct svm_range *prange,
 		case KFD_IOCTL_SVM_ATTR_ACCESS:
 		case KFD_IOCTL_SVM_ATTR_ACCESS_IN_PLACE:
 		case KFD_IOCTL_SVM_ATTR_NO_ACCESS:
+			if (!p->xnack_enabled)
+				*update_mapping = true;
+
 			gpuidx = kfd_process_gpuidx_from_gpuid(p,
 							       attrs[i].value);
 			if (attrs[i].type == KFD_IOCTL_SVM_ATTR_NO_ACCESS) {
@@ -692,18 +795,75 @@ svm_range_apply_attrs(struct kfd_process *p, struct svm_range *prange,
 			}
 			break;
 		case KFD_IOCTL_SVM_ATTR_SET_FLAGS:
+			*update_mapping = true;
 			prange->flags |= attrs[i].value;
 			break;
 		case KFD_IOCTL_SVM_ATTR_CLR_FLAGS:
+			*update_mapping = true;
 			prange->flags &= ~attrs[i].value;
 			break;
 		case KFD_IOCTL_SVM_ATTR_GRANULARITY:
-			prange->granularity = attrs[i].value;
+			prange->granularity = min_t(uint32_t, attrs[i].value, 0x3F);
+			break;
+		default:
+			WARN_ONCE(1, "svm_range_check_attrs wasn't called?");
+		}
+	}
+}
+
+static bool
+svm_range_is_same_attrs(struct kfd_process *p, struct svm_range *prange,
+			uint32_t nattr, struct kfd_ioctl_svm_attribute *attrs)
+{
+	uint32_t i;
+	int gpuidx;
+
+	for (i = 0; i < nattr; i++) {
+		switch (attrs[i].type) {
+		case KFD_IOCTL_SVM_ATTR_PREFERRED_LOC:
+			if (prange->preferred_loc != attrs[i].value)
+				return false;
+			break;
+		case KFD_IOCTL_SVM_ATTR_PREFETCH_LOC:
+			/* Prefetch should always trigger a migration even
+			 * if the value of the attribute didn't change.
+			 */
+			return false;
+		case KFD_IOCTL_SVM_ATTR_ACCESS:
+		case KFD_IOCTL_SVM_ATTR_ACCESS_IN_PLACE:
+		case KFD_IOCTL_SVM_ATTR_NO_ACCESS:
+			gpuidx = kfd_process_gpuidx_from_gpuid(p,
+							       attrs[i].value);
+			if (attrs[i].type == KFD_IOCTL_SVM_ATTR_NO_ACCESS) {
+				if (test_bit(gpuidx, prange->bitmap_access) ||
+				    test_bit(gpuidx, prange->bitmap_aip))
+					return false;
+			} else if (attrs[i].type == KFD_IOCTL_SVM_ATTR_ACCESS) {
+				if (!test_bit(gpuidx, prange->bitmap_access))
+					return false;
+			} else {
+				if (!test_bit(gpuidx, prange->bitmap_aip))
+					return false;
+			}
+			break;
+		case KFD_IOCTL_SVM_ATTR_SET_FLAGS:
+			if ((prange->flags & attrs[i].value) != attrs[i].value)
+				return false;
+			break;
+		case KFD_IOCTL_SVM_ATTR_CLR_FLAGS:
+			if ((prange->flags & attrs[i].value) != 0)
+				return false;
+			break;
+		case KFD_IOCTL_SVM_ATTR_GRANULARITY:
+			if (prange->granularity != attrs[i].value)
+				return false;
 			break;
 		default:
 			WARN_ONCE(1, "svm_range_check_attrs wasn't called?");
 		}
 	}
+
+	return true;
 }
 
 /**
@@ -743,18 +903,56 @@ static void svm_range_debug_dump(struct svm_range_list *svms)
 	}
 }
 
-static bool
-svm_range_is_same_attrs(struct svm_range *old, struct svm_range *new)
+static void *
+svm_range_copy_array(void *psrc, size_t size, uint64_t num_elements,
+		     uint64_t offset, uint64_t *vram_pages)
 {
-	return (old->prefetch_loc == new->prefetch_loc &&
-		old->flags == new->flags &&
-		old->granularity == new->granularity);
+	unsigned char *src = (unsigned char *)psrc + offset;
+	unsigned char *dst;
+	uint64_t i;
+
+	dst = kvmalloc_array(num_elements, size, GFP_KERNEL);
+	if (!dst)
+		return NULL;
+
+	if (!vram_pages) {
+		memcpy(dst, src, num_elements * size);
+		return (void *)dst;
+	}
+
+	*vram_pages = 0;
+	for (i = 0; i < num_elements; i++) {
+		dma_addr_t *temp;
+		temp = (dma_addr_t *)dst + i;
+		*temp = *((dma_addr_t *)src + i);
+		if (*temp&SVM_RANGE_VRAM_DOMAIN)
+			(*vram_pages)++;
+	}
+
+	return (void *)dst;
+}
+
+static int
+svm_range_copy_dma_addrs(struct svm_range *dst, struct svm_range *src)
+{
+	int i;
+
+	for (i = 0; i < MAX_GPU_INSTANCE; i++) {
+		if (!src->dma_addr[i])
+			continue;
+		dst->dma_addr[i] = svm_range_copy_array(src->dma_addr[i],
+					sizeof(*src->dma_addr[i]), src->npages, 0, NULL);
+		if (!dst->dma_addr[i])
+			return -ENOMEM;
+	}
+
+	return 0;
 }
 
 static int
 svm_range_split_array(void *ppnew, void *ppold, size_t size,
 		      uint64_t old_start, uint64_t old_n,
-		      uint64_t new_start, uint64_t new_n)
+		      uint64_t new_start, uint64_t new_n, uint64_t *new_vram_pages)
 {
 	unsigned char *new, *old, *pold;
 	uint64_t d;
@@ -765,22 +963,17 @@ svm_range_split_array(void *ppnew, void *ppold, size_t size,
 	if (!pold)
 		return 0;
 
-	new = kvmalloc_array(new_n, size, GFP_KERNEL);
+	d = (new_start - old_start) * size;
+	/* get dma addr array for new range and calculte its vram page number */
+	new = svm_range_copy_array(pold, size, new_n, d, new_vram_pages);
 	if (!new)
 		return -ENOMEM;
-
-	d = (new_start - old_start) * size;
-	memcpy(new, pold + d, new_n * size);
-
-	old = kvmalloc_array(old_n, size, GFP_KERNEL);
+	d = (new_start == old_start) ? new_n * size : 0;
+	old = svm_range_copy_array(pold, size, old_n, d, NULL);
 	if (!old) {
 		kvfree(new);
 		return -ENOMEM;
 	}
-
-	d = (new_start == old_start) ? new_n * size : 0;
-	memcpy(old, pold + d, old_n * size);
-
 	kvfree(pold);
 	*(void **)ppold = old;
 	*(void **)ppnew = new;
@@ -798,10 +991,13 @@ svm_range_split_pages(struct svm_range *new, struct svm_range *old,
 	for (i = 0; i < MAX_GPU_INSTANCE; i++) {
 		r = svm_range_split_array(&new->dma_addr[i], &old->dma_addr[i],
 					  sizeof(*old->dma_addr[i]), old->start,
-					  npages, new->start, new->npages);
+					  npages, new->start, new->npages,
+					  old->actual_loc ? &new->vram_pages : NULL);
 		if (r)
 			return r;
 	}
+	if (old->actual_loc)
+		old->vram_pages -= new->vram_pages;
 
 	return 0;
 }
@@ -880,8 +1076,10 @@ svm_range_split_adjust(struct svm_range *new, struct svm_range *old,
 	new->prefetch_loc = old->prefetch_loc;
 	new->actual_loc = old->actual_loc;
 	new->granularity = old->granularity;
+	new->mapped_to_gpu = old->mapped_to_gpu;
 	bitmap_copy(new->bitmap_access, old->bitmap_access, MAX_GPU_INSTANCE);
 	bitmap_copy(new->bitmap_aip, old->bitmap_aip, MAX_GPU_INSTANCE);
+	atomic_set(&new->queue_refcount, atomic_read(&old->queue_refcount));
 
 	return 0;
 }
@@ -925,9 +1123,9 @@ svm_range_split(struct svm_range *prange, uint64_t start, uint64_t last,
 
 	svms = prange->svms;
 	if (old_start == start)
-		*new = svm_range_new(svms, last + 1, old_last);
+		*new = svm_range_new(svms, last + 1, old_last, false);
 	else
-		*new = svm_range_new(svms, old_start, start - 1);
+		*new = svm_range_new(svms, old_start, start - 1, false);
 	if (!*new)
 		return -ENOMEM;
 
@@ -935,7 +1133,7 @@ svm_range_split(struct svm_range *prange, uint64_t start, uint64_t last,
 	if (r) {
 		pr_debug("failed %d split [0x%llx 0x%llx] to [0x%llx 0x%llx]\n",
 			 r, old_start, old_last, start, last);
-		svm_range_free(*new);
+		svm_range_free(*new, false);
 		*new = NULL;
 	}
 
@@ -943,126 +1141,80 @@ svm_range_split(struct svm_range *prange, uint64_t start, uint64_t last,
 }
 
 static int
-svm_range_split_tail(struct svm_range *prange, struct svm_range *new,
-		     uint64_t new_last, struct list_head *insert_list)
+svm_range_split_tail(struct svm_range *prange, uint64_t new_last,
+		     struct list_head *insert_list, struct list_head *remap_list)
 {
-	struct svm_range *tail;
+	struct svm_range *tail = NULL;
 	int r = svm_range_split(prange, prange->start, new_last, &tail);
 
-	if (!r)
-		list_add(&tail->insert_list, insert_list);
+	if (!r) {
+		list_add(&tail->list, insert_list);
+		if (!IS_ALIGNED(new_last + 1, 1UL << prange->granularity))
+			list_add(&tail->update_list, remap_list);
+	}
 	return r;
 }
 
 static int
-svm_range_split_head(struct svm_range *prange, struct svm_range *new,
-		     uint64_t new_start, struct list_head *insert_list)
+svm_range_split_head(struct svm_range *prange, uint64_t new_start,
+		     struct list_head *insert_list, struct list_head *remap_list)
 {
-	struct svm_range *head;
+	struct svm_range *head = NULL;
 	int r = svm_range_split(prange, new_start, prange->last, &head);
 
-	if (!r)
-		list_add(&head->insert_list, insert_list);
+	if (!r) {
+		list_add(&head->list, insert_list);
+		if (!IS_ALIGNED(new_start, 1UL << prange->granularity))
+			list_add(&head->update_list, remap_list);
+	}
 	return r;
 }
 
 static void
-svm_range_add_child(struct svm_range *prange, struct mm_struct *mm,
-		    struct svm_range *pchild, enum svm_work_list_ops op)
+svm_range_add_child(struct svm_range *prange, struct svm_range *pchild, enum svm_work_list_ops op)
 {
 	pr_debug("add child 0x%p [0x%lx 0x%lx] to prange 0x%p child list %d\n",
 		 pchild, pchild->start, pchild->last, prange, op);
 
-	pchild->work_item.mm = mm;
+	pchild->work_item.mm = NULL;
 	pchild->work_item.op = op;
 	list_add_tail(&pchild->child_list, &prange->child_list);
 }
 
-/**
- * svm_range_split_by_granularity - collect ranges within granularity boundary
- *
- * @p: the process with svms list
- * @mm: mm structure
- * @addr: the vm fault address in pages, to split the prange
- * @parent: parent range if prange is from child list
- * @prange: prange to split
- *
- * Trims @prange to be a single aligned block of prange->granularity if
- * possible. The head and tail are added to the child_list in @parent.
- *
- * Context: caller must hold mmap_read_lock and prange->lock
- *
- * Return:
- * 0 - OK, otherwise error code
- */
-int
-svm_range_split_by_granularity(struct kfd_process *p, struct mm_struct *mm,
-			       unsigned long addr, struct svm_range *parent,
-			       struct svm_range *prange)
+static bool
+svm_nodes_in_same_hive(struct kfd_node *node_a, struct kfd_node *node_b)
 {
-	struct svm_range *head, *tail;
-	unsigned long start, last, size;
-	int r;
-
-	/* Align splited range start and size to granularity size, then a single
-	 * PTE will be used for whole range, this reduces the number of PTE
-	 * updated and the L1 TLB space used for translation.
-	 */
-	size = 1UL << prange->granularity;
-	start = ALIGN_DOWN(addr, size);
-	last = ALIGN(addr + 1, size) - 1;
-
-	pr_debug("svms 0x%p split [0x%lx 0x%lx] to [0x%lx 0x%lx] size 0x%lx\n",
-		 prange->svms, prange->start, prange->last, start, last, size);
-
-	if (start > prange->start) {
-		r = svm_range_split(prange, start, prange->last, &head);
-		if (r)
-			return r;
-		svm_range_add_child(parent, mm, head, SVM_OP_ADD_RANGE);
-	}
-
-	if (last < prange->last) {
-		r = svm_range_split(prange, prange->start, last, &tail);
-		if (r)
-			return r;
-		svm_range_add_child(parent, mm, tail, SVM_OP_ADD_RANGE);
-	}
-
-	/* xnack on, update mapping on GPUs with ACCESS_IN_PLACE */
-	if (p->xnack_enabled && prange->work_item.op == SVM_OP_ADD_RANGE) {
-		prange->work_item.op = SVM_OP_ADD_RANGE_AND_MAP;
-		pr_debug("change prange 0x%p [0x%lx 0x%lx] op %d\n",
-			 prange, prange->start, prange->last,
-			 SVM_OP_ADD_RANGE_AND_MAP);
-	}
-	return 0;
+	return (node_a->adev == node_b->adev ||
+		amdgpu_xgmi_same_hive(node_a->adev, node_b->adev));
 }
 
 static uint64_t
-svm_range_get_pte_flags(struct amdgpu_device *adev, struct svm_range *prange,
-			int domain)
+svm_range_get_pte_flags(struct kfd_node *node, struct amdgpu_vm *vm,
+			struct svm_range *prange, int domain)
 {
-	struct amdgpu_device *bo_adev;
+	struct kfd_node *bo_node;
 	uint32_t flags = prange->flags;
 	uint32_t mapping_flags = 0;
+	uint32_t gc_ip_version = KFD_GC_VERSION(node);
 	uint64_t pte_flags;
 	bool snoop = (domain != SVM_RANGE_VRAM_DOMAIN);
-	bool coherent = flags & KFD_IOCTL_SVM_FLAG_COHERENT;
+	bool coherent = flags & (KFD_IOCTL_SVM_FLAG_COHERENT | KFD_IOCTL_SVM_FLAG_EXT_COHERENT);
+	bool ext_coherent = flags & KFD_IOCTL_SVM_FLAG_EXT_COHERENT;
+	unsigned int mtype_local;
 
 	if (domain == SVM_RANGE_VRAM_DOMAIN)
-		bo_adev = amdgpu_ttm_adev(prange->svm_bo->bo->tbo.bdev);
+		bo_node = prange->svm_bo->node;
 
-	switch (adev->asic_type) {
-	case CHIP_ARCTURUS:
+	switch (gc_ip_version) {
+	case IP_VERSION(9, 4, 1):
 		if (domain == SVM_RANGE_VRAM_DOMAIN) {
-			if (bo_adev == adev) {
+			if (bo_node == node) {
 				mapping_flags |= coherent ?
 					AMDGPU_VM_MTYPE_CC : AMDGPU_VM_MTYPE_RW;
 			} else {
 				mapping_flags |= coherent ?
 					AMDGPU_VM_MTYPE_UC : AMDGPU_VM_MTYPE_NC;
-				if (amdgpu_xgmi_same_hive(adev, bo_adev))
+				if (svm_nodes_in_same_hive(node, bo_node))
 					snoop = true;
 			}
 		} else {
@@ -1070,17 +1222,17 @@ svm_range_get_pte_flags(struct amdgpu_device *adev, struct svm_range *prange,
 				AMDGPU_VM_MTYPE_UC : AMDGPU_VM_MTYPE_NC;
 		}
 		break;
-	case CHIP_ALDEBARAN:
+	case IP_VERSION(9, 4, 2):
 		if (domain == SVM_RANGE_VRAM_DOMAIN) {
-			if (bo_adev == adev) {
+			if (bo_node == node) {
 				mapping_flags |= coherent ?
 					AMDGPU_VM_MTYPE_CC : AMDGPU_VM_MTYPE_RW;
-				if (adev->gmc.xgmi.connected_to_cpu)
+				if (node->adev->gmc.xgmi.connected_to_cpu)
 					snoop = true;
 			} else {
 				mapping_flags |= coherent ?
 					AMDGPU_VM_MTYPE_UC : AMDGPU_VM_MTYPE_NC;
-				if (amdgpu_xgmi_same_hive(adev, bo_adev))
+				if (svm_nodes_in_same_hive(node, bo_node))
 					snoop = true;
 			}
 		} else {
@@ -1088,23 +1240,71 @@ svm_range_get_pte_flags(struct amdgpu_device *adev, struct svm_range *prange,
 				AMDGPU_VM_MTYPE_UC : AMDGPU_VM_MTYPE_NC;
 		}
 		break;
+	case IP_VERSION(9, 4, 3):
+	case IP_VERSION(9, 4, 4):
+	case IP_VERSION(9, 5, 0):
+		if (ext_coherent)
+			mtype_local = AMDGPU_VM_MTYPE_CC;
+		else
+			mtype_local = amdgpu_mtype_local == 1 ? AMDGPU_VM_MTYPE_NC :
+				amdgpu_mtype_local == 2 ? AMDGPU_VM_MTYPE_CC : AMDGPU_VM_MTYPE_RW;
+		snoop = true;
+		if (domain == SVM_RANGE_VRAM_DOMAIN) {
+			/* local HBM region close to partition */
+			if (bo_node->adev == node->adev &&
+			    (!bo_node->xcp || !node->xcp || bo_node->xcp->mem_id == node->xcp->mem_id))
+				mapping_flags |= mtype_local;
+			/* local HBM region far from partition or remote XGMI GPU
+			 * with regular system scope coherence
+			 */
+			else if (svm_nodes_in_same_hive(bo_node, node) && !ext_coherent)
+				mapping_flags |= AMDGPU_VM_MTYPE_NC;
+			/* PCIe P2P on GPUs pre-9.5.0 */
+			else if (gc_ip_version < IP_VERSION(9, 5, 0) &&
+				 !svm_nodes_in_same_hive(bo_node, node))
+				mapping_flags |= AMDGPU_VM_MTYPE_UC;
+			/* Other remote memory */
+			else
+				mapping_flags |= ext_coherent ? AMDGPU_VM_MTYPE_UC : AMDGPU_VM_MTYPE_NC;
+		/* system memory accessed by the APU */
+		} else if (node->adev->flags & AMD_IS_APU) {
+			/* On NUMA systems, locality is determined per-page
+			 * in amdgpu_gmc_override_vm_pte_flags
+			 */
+			if (num_possible_nodes() <= 1)
+				mapping_flags |= mtype_local;
+			else
+				mapping_flags |= ext_coherent ? AMDGPU_VM_MTYPE_UC : AMDGPU_VM_MTYPE_NC;
+		/* system memory accessed by the dGPU */
+		} else {
+			if (gc_ip_version < IP_VERSION(9, 5, 0) || ext_coherent)
+				mapping_flags |= AMDGPU_VM_MTYPE_UC;
+			else
+				mapping_flags |= AMDGPU_VM_MTYPE_NC;
+		}
+		break;
+	case IP_VERSION(12, 0, 0):
+	case IP_VERSION(12, 0, 1):
+		mapping_flags |= AMDGPU_VM_MTYPE_NC;
+		break;
 	default:
 		mapping_flags |= coherent ?
 			AMDGPU_VM_MTYPE_UC : AMDGPU_VM_MTYPE_NC;
 	}
 
-	mapping_flags |= AMDGPU_VM_PAGE_READABLE | AMDGPU_VM_PAGE_WRITEABLE;
-
-	if (flags & KFD_IOCTL_SVM_FLAG_GPU_RO)
-		mapping_flags &= ~AMDGPU_VM_PAGE_WRITEABLE;
 	if (flags & KFD_IOCTL_SVM_FLAG_GPU_EXEC)
 		mapping_flags |= AMDGPU_VM_PAGE_EXECUTABLE;
 
 	pte_flags = AMDGPU_PTE_VALID;
 	pte_flags |= (domain == SVM_RANGE_VRAM_DOMAIN) ? 0 : AMDGPU_PTE_SYSTEM;
 	pte_flags |= snoop ? AMDGPU_PTE_SNOOPED : 0;
+	if (gc_ip_version >= IP_VERSION(12, 0, 0))
+		pte_flags |= AMDGPU_PTE_IS_PTE;
 
-	pte_flags |= amdgpu_gem_va_map_flags(adev, mapping_flags);
+	amdgpu_gmc_get_vm_pte(node->adev, vm, NULL, mapping_flags, &pte_flags);
+	pte_flags |= AMDGPU_PTE_READABLE;
+	if (!(flags & KFD_IOCTL_SVM_FLAG_GPU_RO))
+		pte_flags |= AMDGPU_PTE_WRITEABLE;
 	return pte_flags;
 }
 
@@ -1117,23 +1317,33 @@ svm_range_unmap_from_gpu(struct amdgpu_device *adev, struct amdgpu_vm *vm,
 
 	pr_debug("[0x%llx 0x%llx]\n", start, last);
 
-	return amdgpu_vm_bo_update_mapping(adev, adev, vm, false, true, NULL,
-					   start, last, init_pte_value, 0,
-					   NULL, NULL, fence, NULL);
+	return amdgpu_vm_update_range(adev, vm, false, true, true, false, NULL, start,
+				      last, init_pte_value, 0, 0, NULL, NULL,
+				      fence);
 }
 
 static int
 svm_range_unmap_from_gpus(struct svm_range *prange, unsigned long start,
-			  unsigned long last)
+			  unsigned long last, uint32_t trigger)
 {
 	DECLARE_BITMAP(bitmap, MAX_GPU_INSTANCE);
 	struct kfd_process_device *pdd;
 	struct dma_fence *fence = NULL;
-	struct amdgpu_device *adev;
 	struct kfd_process *p;
 	uint32_t gpuidx;
 	int r = 0;
 
+	if (!prange->mapped_to_gpu) {
+		pr_debug("prange 0x%p [0x%lx 0x%lx] not mapped to GPU\n",
+			 prange, prange->start, prange->last);
+		return 0;
+	}
+
+	if (prange->start == start && prange->last == last) {
+		pr_debug("unmap svms 0x%p prange 0x%p\n", prange->svms, prange);
+		prange->mapped_to_gpu = false;
+	}
+
 	bitmap_or(bitmap, prange->bitmap_access, prange->bitmap_aip,
 		  MAX_GPU_INSTANCE);
 	p = container_of(prange->svms, struct kfd_process, svms);
@@ -1145,9 +1355,12 @@ svm_range_unmap_from_gpus(struct svm_range *prange, unsigned long start,
 			pr_debug("failed to find device idx %d\n", gpuidx);
 			return -EINVAL;
 		}
-		adev = (struct amdgpu_device *)pdd->dev->kgd;
 
-		r = svm_range_unmap_from_gpu(adev, drm_priv_to_vm(pdd->drm_priv),
+		kfd_smi_event_unmap_from_gpu(pdd->dev, p->lead_thread->pid,
+					     start, last, trigger);
+
+		r = svm_range_unmap_from_gpu(pdd->dev->adev,
+					     drm_priv_to_vm(pdd->drm_priv),
 					     start, last, &fence);
 		if (r)
 			break;
@@ -1159,21 +1372,20 @@ svm_range_unmap_from_gpus(struct svm_range *prange, unsigned long start,
 			if (r)
 				break;
 		}
-		amdgpu_amdkfd_flush_gpu_tlb_pasid((struct kgd_dev *)adev,
-					p->pasid, TLB_FLUSH_HEAVYWEIGHT);
+		kfd_flush_tlb(pdd, TLB_FLUSH_HEAVYWEIGHT);
 	}
 
 	return r;
 }
 
 static int
-svm_range_map_to_gpu(struct amdgpu_device *adev, struct amdgpu_vm *vm,
-		     struct svm_range *prange, unsigned long offset,
-		     unsigned long npages, bool readonly, dma_addr_t *dma_addr,
-		     struct amdgpu_device *bo_adev, struct dma_fence **fence)
+svm_range_map_to_gpu(struct kfd_process_device *pdd, struct svm_range *prange,
+		     unsigned long offset, unsigned long npages, bool readonly,
+		     dma_addr_t *dma_addr, struct amdgpu_device *bo_adev,
+		     struct dma_fence **fence, bool flush_tlb)
 {
-	struct amdgpu_bo_va bo_va;
-	bool table_freed = false;
+	struct amdgpu_device *adev = pdd->dev->adev;
+	struct amdgpu_vm *vm = drm_priv_to_vm(pdd->drm_priv);
 	uint64_t pte_flags;
 	unsigned long last_start;
 	int last_domain;
@@ -1185,9 +1397,6 @@ svm_range_map_to_gpu(struct amdgpu_device *adev, struct amdgpu_vm *vm,
 	pr_debug("svms 0x%p [0x%lx 0x%lx] readonly %d\n", prange->svms,
 		 last_start, last_start + npages - 1, readonly);
 
-	if (prange->svm_bo && prange->ttm_res)
-		bo_va.is_xgmi = amdgpu_xgmi_same_hive(adev, bo_adev);
-
 	for (i = offset; i < offset + npages; i++) {
 		last_domain = dma_addr[i] & SVM_RANGE_VRAM_DOMAIN;
 		dma_addr[i] &= ~SVM_RANGE_VRAM_DOMAIN;
@@ -1202,7 +1411,7 @@ svm_range_map_to_gpu(struct amdgpu_device *adev, struct amdgpu_vm *vm,
 		pr_debug("Mapping range [0x%lx 0x%llx] on domain: %s\n",
 			 last_start, prange->start + i, last_domain ? "GPU" : "CPU");
 
-		pte_flags = svm_range_get_pte_flags(adev, prange, last_domain);
+		pte_flags = svm_range_get_pte_flags(pdd->dev, vm, prange, last_domain);
 		if (readonly)
 			pte_flags &= ~AMDGPU_PTE_WRITEABLE;
 
@@ -1211,13 +1420,16 @@ svm_range_map_to_gpu(struct amdgpu_device *adev, struct amdgpu_vm *vm,
 			 (last_domain == SVM_RANGE_VRAM_DOMAIN) ? 1 : 0,
 			 pte_flags);
 
-		r = amdgpu_vm_bo_update_mapping(adev, bo_adev, vm, false, false,
-						NULL, last_start,
-						prange->start + i, pte_flags,
-						last_start - prange->start,
-						NULL, dma_addr,
-						&vm->last_update,
-						&table_freed);
+		/* For dGPU mode, we use same vm_manager to allocate VRAM for
+		 * different memory partition based on fpfn/lpfn, we should use
+		 * same vm_manager.vram_base_offset regardless memory partition.
+		 */
+		r = amdgpu_vm_update_range(adev, vm, false, false, flush_tlb, true,
+					   NULL, last_start, prange->start + i,
+					   pte_flags,
+					   (last_start - prange->start) << PAGE_SHIFT,
+					   bo_adev ? bo_adev->vm_manager.vram_base_offset : 0,
+					   NULL, dma_addr, &vm->last_update);
 
 		for (j = last_start - prange->start; j <= i; j++)
 			dma_addr[j] |= last_domain;
@@ -1239,13 +1451,6 @@ svm_range_map_to_gpu(struct amdgpu_device *adev, struct amdgpu_vm *vm,
 	if (fence)
 		*fence = dma_fence_get(vm->last_update);
 
-	if (table_freed) {
-		struct kfd_process *p;
-
-		p = container_of(prange->svms, struct kfd_process, svms);
-		amdgpu_amdkfd_flush_gpu_tlb_pasid((struct kgd_dev *)adev,
-						p->pasid, TLB_FLUSH_LEGACY);
-	}
 out:
 	return r;
 }
@@ -1253,20 +1458,17 @@ out:
 static int
 svm_range_map_to_gpus(struct svm_range *prange, unsigned long offset,
 		      unsigned long npages, bool readonly,
-		      unsigned long *bitmap, bool wait)
+		      unsigned long *bitmap, bool wait, bool flush_tlb)
 {
 	struct kfd_process_device *pdd;
-	struct amdgpu_device *bo_adev;
-	struct amdgpu_device *adev;
+	struct amdgpu_device *bo_adev = NULL;
 	struct kfd_process *p;
 	struct dma_fence *fence = NULL;
 	uint32_t gpuidx;
 	int r = 0;
 
 	if (prange->svm_bo && prange->ttm_res)
-		bo_adev = amdgpu_ttm_adev(prange->svm_bo->bo->tbo.bdev);
-	else
-		bo_adev = NULL;
+		bo_adev = prange->svm_bo->node->adev;
 
 	p = container_of(prange->svms, struct kfd_process, svms);
 	for_each_set_bit(gpuidx, bitmap, MAX_GPU_INSTANCE) {
@@ -1276,22 +1478,21 @@ svm_range_map_to_gpus(struct svm_range *prange, unsigned long offset,
 			pr_debug("failed to find device idx %d\n", gpuidx);
 			return -EINVAL;
 		}
-		adev = (struct amdgpu_device *)pdd->dev->kgd;
 
 		pdd = kfd_bind_process_to_device(pdd->dev, p);
 		if (IS_ERR(pdd))
 			return -EINVAL;
 
-		if (bo_adev && adev != bo_adev &&
-		    !amdgpu_xgmi_same_hive(adev, bo_adev)) {
+		if (bo_adev && pdd->dev->adev != bo_adev &&
+		    !amdgpu_xgmi_same_hive(pdd->dev->adev, bo_adev)) {
 			pr_debug("cannot map to device idx %d\n", gpuidx);
 			continue;
 		}
 
-		r = svm_range_map_to_gpu(adev, drm_priv_to_vm(pdd->drm_priv),
-					 prange, offset, npages, readonly,
+		r = svm_range_map_to_gpu(pdd, prange, offset, npages, readonly,
 					 prange->dma_addr[gpuidx],
-					 bo_adev, wait ? &fence : NULL);
+					 bo_adev, wait ? &fence : NULL,
+					 flush_tlb);
 		if (r)
 			break;
 
@@ -1304,6 +1505,8 @@ svm_range_map_to_gpus(struct svm_range *prange, unsigned long offset,
 				break;
 			}
 		}
+
+		kfd_flush_tlb(pdd, TLB_FLUSH_LEGACY);
 	}
 
 	return r;
@@ -1313,40 +1516,35 @@ struct svm_validate_context {
 	struct kfd_process *process;
 	struct svm_range *prange;
 	bool intr;
-	unsigned long bitmap[MAX_GPU_INSTANCE];
-	struct ttm_validate_buffer tv[MAX_GPU_INSTANCE];
-	struct list_head validate_list;
-	struct ww_acquire_ctx ticket;
+	DECLARE_BITMAP(bitmap, MAX_GPU_INSTANCE);
+	struct drm_exec exec;
 };
 
-static int svm_range_reserve_bos(struct svm_validate_context *ctx)
+static int svm_range_reserve_bos(struct svm_validate_context *ctx, bool intr)
 {
 	struct kfd_process_device *pdd;
-	struct amdgpu_device *adev;
 	struct amdgpu_vm *vm;
 	uint32_t gpuidx;
 	int r;
 
-	INIT_LIST_HEAD(&ctx->validate_list);
-	for_each_set_bit(gpuidx, ctx->bitmap, MAX_GPU_INSTANCE) {
-		pdd = kfd_process_device_from_gpuidx(ctx->process, gpuidx);
-		if (!pdd) {
-			pr_debug("failed to find device idx %d\n", gpuidx);
-			return -EINVAL;
-		}
-		adev = (struct amdgpu_device *)pdd->dev->kgd;
-		vm = drm_priv_to_vm(pdd->drm_priv);
-
-		ctx->tv[gpuidx].bo = &vm->root.bo->tbo;
-		ctx->tv[gpuidx].num_shared = 4;
-		list_add(&ctx->tv[gpuidx].head, &ctx->validate_list);
-	}
+	drm_exec_init(&ctx->exec, intr ? DRM_EXEC_INTERRUPTIBLE_WAIT: 0, 0);
+	drm_exec_until_all_locked(&ctx->exec) {
+		for_each_set_bit(gpuidx, ctx->bitmap, MAX_GPU_INSTANCE) {
+			pdd = kfd_process_device_from_gpuidx(ctx->process, gpuidx);
+			if (!pdd) {
+				pr_debug("failed to find device idx %d\n", gpuidx);
+				r = -EINVAL;
+				goto unreserve_out;
+			}
+			vm = drm_priv_to_vm(pdd->drm_priv);
 
-	r = ttm_eu_reserve_buffers(&ctx->ticket, &ctx->validate_list,
-				   ctx->intr, NULL);
-	if (r) {
-		pr_debug("failed %d to reserve bo\n", r);
-		return r;
+			r = amdgpu_vm_lock_pd(vm, &ctx->exec, 2);
+			drm_exec_retry_on_contention(&ctx->exec);
+			if (unlikely(r)) {
+				pr_debug("failed %d to reserve bo\n", r);
+				goto unreserve_out;
+			}
+		}
 	}
 
 	for_each_set_bit(gpuidx, ctx->bitmap, MAX_GPU_INSTANCE) {
@@ -1356,10 +1554,10 @@ static int svm_range_reserve_bos(struct svm_validate_context *ctx)
 			r = -EINVAL;
 			goto unreserve_out;
 		}
-		adev = (struct amdgpu_device *)pdd->dev->kgd;
 
-		r = amdgpu_vm_validate_pt_bos(adev, drm_priv_to_vm(pdd->drm_priv),
-					      svm_range_bo_validate, NULL);
+		r = amdgpu_vm_validate(pdd->dev->adev,
+				       drm_priv_to_vm(pdd->drm_priv), NULL,
+				       svm_range_bo_validate, NULL);
 		if (r) {
 			pr_debug("failed %d validate pt bos\n", r);
 			goto unreserve_out;
@@ -1369,24 +1567,24 @@ static int svm_range_reserve_bos(struct svm_validate_context *ctx)
 	return 0;
 
 unreserve_out:
-	ttm_eu_backoff_reservation(&ctx->ticket, &ctx->validate_list);
+	drm_exec_fini(&ctx->exec);
 	return r;
 }
 
 static void svm_range_unreserve_bos(struct svm_validate_context *ctx)
 {
-	ttm_eu_backoff_reservation(&ctx->ticket, &ctx->validate_list);
+	drm_exec_fini(&ctx->exec);
 }
 
 static void *kfd_svm_page_owner(struct kfd_process *p, int32_t gpuidx)
 {
 	struct kfd_process_device *pdd;
-	struct amdgpu_device *adev;
 
 	pdd = kfd_process_device_from_gpuidx(p, gpuidx);
-	adev = (struct amdgpu_device *)pdd->dev->kgd;
+	if (!pdd)
+		return NULL;
 
-	return SVM_ADEV_PGMAP_OWNER(adev);
+	return SVM_ADEV_PGMAP_OWNER(pdd->dev->adev);
 }
 
 /*
@@ -1414,135 +1612,196 @@ static void *kfd_svm_page_owner(struct kfd_process *p, int32_t gpuidx)
  * 5. Release page table (and SVM BO) reservation
  */
 static int svm_range_validate_and_map(struct mm_struct *mm,
-				      struct svm_range *prange,
-				      int32_t gpuidx, bool intr, bool wait)
+				      unsigned long map_start, unsigned long map_last,
+				      struct svm_range *prange, int32_t gpuidx,
+				      bool intr, bool wait, bool flush_tlb)
 {
-	struct svm_validate_context ctx;
+	struct svm_validate_context *ctx;
 	unsigned long start, end, addr;
 	struct kfd_process *p;
 	void *owner;
 	int32_t idx;
 	int r = 0;
 
-	ctx.process = container_of(prange->svms, struct kfd_process, svms);
-	ctx.prange = prange;
-	ctx.intr = intr;
+	ctx = kzalloc(sizeof(struct svm_validate_context), GFP_KERNEL);
+	if (!ctx)
+		return -ENOMEM;
+	ctx->process = container_of(prange->svms, struct kfd_process, svms);
+	ctx->prange = prange;
+	ctx->intr = intr;
 
 	if (gpuidx < MAX_GPU_INSTANCE) {
-		bitmap_zero(ctx.bitmap, MAX_GPU_INSTANCE);
-		bitmap_set(ctx.bitmap, gpuidx, 1);
-	} else if (ctx.process->xnack_enabled) {
-		bitmap_copy(ctx.bitmap, prange->bitmap_aip, MAX_GPU_INSTANCE);
+		bitmap_zero(ctx->bitmap, MAX_GPU_INSTANCE);
+		bitmap_set(ctx->bitmap, gpuidx, 1);
+	} else if (ctx->process->xnack_enabled) {
+		bitmap_copy(ctx->bitmap, prange->bitmap_aip, MAX_GPU_INSTANCE);
 
 		/* If prefetch range to GPU, or GPU retry fault migrate range to
 		 * GPU, which has ACCESS attribute to the range, create mapping
 		 * on that GPU.
 		 */
 		if (prange->actual_loc) {
-			gpuidx = kfd_process_gpuidx_from_gpuid(ctx.process,
+			gpuidx = kfd_process_gpuidx_from_gpuid(ctx->process,
 							prange->actual_loc);
 			if (gpuidx < 0) {
 				WARN_ONCE(1, "failed get device by id 0x%x\n",
 					 prange->actual_loc);
-				return -EINVAL;
+				r = -EINVAL;
+				goto free_ctx;
 			}
 			if (test_bit(gpuidx, prange->bitmap_access))
-				bitmap_set(ctx.bitmap, gpuidx, 1);
+				bitmap_set(ctx->bitmap, gpuidx, 1);
+		}
+
+		/*
+		 * If prange is already mapped or with always mapped flag,
+		 * update mapping on GPUs with ACCESS attribute
+		 */
+		if (bitmap_empty(ctx->bitmap, MAX_GPU_INSTANCE)) {
+			if (prange->mapped_to_gpu ||
+			    prange->flags & KFD_IOCTL_SVM_FLAG_GPU_ALWAYS_MAPPED)
+				bitmap_copy(ctx->bitmap, prange->bitmap_access, MAX_GPU_INSTANCE);
 		}
 	} else {
-		bitmap_or(ctx.bitmap, prange->bitmap_access,
+		bitmap_or(ctx->bitmap, prange->bitmap_access,
 			  prange->bitmap_aip, MAX_GPU_INSTANCE);
 	}
 
-	if (bitmap_empty(ctx.bitmap, MAX_GPU_INSTANCE))
-		return 0;
+	if (bitmap_empty(ctx->bitmap, MAX_GPU_INSTANCE)) {
+		r = 0;
+		goto free_ctx;
+	}
 
 	if (prange->actual_loc && !prange->ttm_res) {
 		/* This should never happen. actual_loc gets set by
 		 * svm_migrate_ram_to_vram after allocating a BO.
 		 */
 		WARN_ONCE(1, "VRAM BO missing during validation\n");
-		return -EINVAL;
+		r = -EINVAL;
+		goto free_ctx;
 	}
 
-	svm_range_reserve_bos(&ctx);
+	r = svm_range_reserve_bos(ctx, intr);
+	if (r)
+		goto free_ctx;
 
 	p = container_of(prange->svms, struct kfd_process, svms);
-	owner = kfd_svm_page_owner(p, find_first_bit(ctx.bitmap,
+	owner = kfd_svm_page_owner(p, find_first_bit(ctx->bitmap,
 						MAX_GPU_INSTANCE));
-	for_each_set_bit(idx, ctx.bitmap, MAX_GPU_INSTANCE) {
+	for_each_set_bit(idx, ctx->bitmap, MAX_GPU_INSTANCE) {
 		if (kfd_svm_page_owner(p, idx) != owner) {
 			owner = NULL;
 			break;
 		}
 	}
 
-	start = prange->start << PAGE_SHIFT;
-	end = (prange->last + 1) << PAGE_SHIFT;
-	for (addr = start; addr < end && !r; ) {
-		struct hmm_range *hmm_range;
+	start = map_start << PAGE_SHIFT;
+	end = (map_last + 1) << PAGE_SHIFT;
+	for (addr = start; !r && addr < end; ) {
+		struct amdgpu_hmm_range *range = NULL;
+		unsigned long map_start_vma;
+		unsigned long map_last_vma;
 		struct vm_area_struct *vma;
-		unsigned long next;
+		unsigned long next = 0;
 		unsigned long offset;
 		unsigned long npages;
 		bool readonly;
 
-		vma = find_vma(mm, addr);
-		if (!vma || addr < vma->vm_start) {
-			r = -EFAULT;
-			goto unreserve_out;
-		}
-		readonly = !(vma->vm_flags & VM_WRITE);
+		vma = vma_lookup(mm, addr);
+		if (vma) {
+			readonly = !(vma->vm_flags & VM_WRITE);
 
-		next = min(vma->vm_end, end);
-		npages = (next - addr) >> PAGE_SHIFT;
-		WRITE_ONCE(p->svms.faulting_task, current);
-		r = amdgpu_hmm_range_get_pages(&prange->notifier, mm, NULL,
-					       addr, npages, &hmm_range,
-					       readonly, true, owner);
-		WRITE_ONCE(p->svms.faulting_task, NULL);
-		if (r) {
-			pr_debug("failed %d to get svm range pages\n", r);
-			goto unreserve_out;
+			next = min(vma->vm_end, end);
+			npages = (next - addr) >> PAGE_SHIFT;
+			/* HMM requires at least READ permissions. If provided with PROT_NONE,
+			 * unmap the memory. If it's not already mapped, this is a no-op
+			 * If PROT_WRITE is provided without READ, warn first then unmap
+			 */
+			if (!(vma->vm_flags & VM_READ)) {
+				unsigned long e, s;
+
+				svm_range_lock(prange);
+				if (vma->vm_flags & VM_WRITE)
+					pr_debug("VM_WRITE without VM_READ is not supported");
+				s = max(start, prange->start);
+				e = min(end, prange->last);
+				if (e >= s)
+					r = svm_range_unmap_from_gpus(prange, s, e,
+						       KFD_SVM_UNMAP_TRIGGER_UNMAP_FROM_CPU);
+				svm_range_unlock(prange);
+				/* If unmap returns non-zero, we'll bail on the next for loop
+				 * iteration, so just leave r and continue
+				 */
+				addr = next;
+				continue;
+			}
+
+			WRITE_ONCE(p->svms.faulting_task, current);
+			range = amdgpu_hmm_range_alloc(NULL);
+			if (likely(range))
+				r = amdgpu_hmm_range_get_pages(&prange->notifier, addr, npages,
+							       readonly, owner, range);
+			else
+				r = -ENOMEM;
+			WRITE_ONCE(p->svms.faulting_task, NULL);
+			if (r)
+				pr_debug("failed %d to get svm range pages\n", r);
+		} else {
+			r = -EFAULT;
 		}
 
-		offset = (addr - start) >> PAGE_SHIFT;
-		r = svm_range_dma_map(prange, ctx.bitmap, offset, npages,
-				      hmm_range->hmm_pfns);
-		if (r) {
-			pr_debug("failed %d to dma map range\n", r);
-			goto unreserve_out;
+		if (!r) {
+			offset = (addr >> PAGE_SHIFT) - prange->start;
+			r = svm_range_dma_map(prange, ctx->bitmap, offset, npages,
+					      range->hmm_range.hmm_pfns);
+			if (r)
+				pr_debug("failed %d to dma map range\n", r);
 		}
 
 		svm_range_lock(prange);
-		if (amdgpu_hmm_range_get_pages_done(hmm_range)) {
+
+		/* Free backing memory of hmm_range if it was initialized
+		 * Override return value to TRY AGAIN only if prior returns
+		 * were successful
+		 */
+		if (range && !amdgpu_hmm_range_valid(range) && !r) {
 			pr_debug("hmm update the range, need validate again\n");
 			r = -EAGAIN;
-			goto unlock_out;
 		}
-		if (!list_empty(&prange->child_list)) {
+
+		/* Free the hmm range */
+		amdgpu_hmm_range_free(range);
+
+		if (!r && !list_empty(&prange->child_list)) {
 			pr_debug("range split by unmap in parallel, validate again\n");
 			r = -EAGAIN;
-			goto unlock_out;
 		}
 
-		r = svm_range_map_to_gpus(prange, offset, npages, readonly,
-					  ctx.bitmap, wait);
+		if (!r) {
+			map_start_vma = max(map_start, prange->start + offset);
+			map_last_vma = min(map_last, prange->start + offset + npages - 1);
+			if (map_start_vma <= map_last_vma) {
+				offset = map_start_vma - prange->start;
+				npages = map_last_vma - map_start_vma + 1;
+				r = svm_range_map_to_gpus(prange, offset, npages, readonly,
+							  ctx->bitmap, wait, flush_tlb);
+			}
+		}
+
+		if (!r && next == end)
+			prange->mapped_to_gpu = true;
 
-unlock_out:
 		svm_range_unlock(prange);
 
 		addr = next;
 	}
 
-	if (addr == end)
-		prange->validated_once = true;
-
-unreserve_out:
-	svm_range_unreserve_bos(&ctx);
-
+	svm_range_unreserve_bos(ctx);
 	if (!r)
-		prange->validate_timestamp = ktime_to_us(ktime_get());
+		prange->validate_timestamp = ktime_get_boottime();
+
+free_ctx:
+	kfree(ctx);
 
 	return r;
 }
@@ -1574,6 +1833,7 @@ retry_flush_work:
 static void svm_range_restore_work(struct work_struct *work)
 {
 	struct delayed_work *dwork = to_delayed_work(work);
+	struct amdkfd_process_info *process_info;
 	struct svm_range_list *svms;
 	struct svm_range *prange;
 	struct kfd_process *p;
@@ -1589,14 +1849,17 @@ static void svm_range_restore_work(struct work_struct *work)
 
 	pr_debug("restore svm ranges\n");
 
-	/* kfd_process_notifier_release destroys this worker thread. So during
-	 * the lifetime of this thread, kfd_process and mm will be valid.
-	 */
 	p = container_of(svms, struct kfd_process, svms);
-	mm = p->mm;
-	if (!mm)
+	process_info = p->kgd_process_info;
+
+	/* Keep mm reference when svm_range_validate_and_map ranges */
+	mm = get_task_mm(p->lead_thread);
+	if (!mm) {
+		pr_debug("svms 0x%p process mm gone\n", svms);
 		return;
+	}
 
+	mutex_lock(&process_info->lock);
 	svm_range_list_lock_and_flush_work(svms, mm);
 	mutex_lock(&svms->lock);
 
@@ -1616,8 +1879,8 @@ static void svm_range_restore_work(struct work_struct *work)
 		 */
 		mutex_lock(&prange->migrate_mutex);
 
-		r = svm_range_validate_and_map(mm, prange, MAX_GPU_INSTANCE,
-					       false, true);
+		r = svm_range_validate_and_map(mm, prange->start, prange->last, prange,
+					       MAX_GPU_INSTANCE, false, true, false);
 		if (r)
 			pr_debug("failed %d to map 0x%lx to gpus\n", r,
 				 prange->start);
@@ -1649,17 +1912,26 @@ static void svm_range_restore_work(struct work_struct *work)
 out_reschedule:
 	mutex_unlock(&svms->lock);
 	mmap_write_unlock(mm);
+	mutex_unlock(&process_info->lock);
 
 	/* If validation failed, reschedule another attempt */
 	if (evicted_ranges) {
 		pr_debug("reschedule to restore svm range\n");
-		schedule_delayed_work(&svms->restore_work,
+		queue_delayed_work(system_freezable_wq, &svms->restore_work,
 			msecs_to_jiffies(AMDGPU_SVM_RANGE_RESTORE_DELAY_MS));
+
+		kfd_smi_event_queue_restore_rescheduled(mm);
 	}
+	mmput(mm);
 }
 
 /**
  * svm_range_evict - evict svm range
+ * @prange: svm range structure
+ * @mm: current process mm_struct
+ * @start: starting process queue number
+ * @last: last process queue number
+ * @event: mmu notifier event when range is evicted or migrated
  *
  * Stop all queues of the process to ensure GPU doesn't access the memory, then
  * return to let CPU evict the buffer and proceed CPU pagetable update.
@@ -1671,7 +1943,8 @@ out_reschedule:
  */
 static int
 svm_range_evict(struct svm_range *prange, struct mm_struct *mm,
-		unsigned long start, unsigned long last)
+		unsigned long start, unsigned long last,
+		enum mmu_notifier_event event)
 {
 	struct svm_range_list *svms = prange->svms;
 	struct svm_range *pchild;
@@ -1683,10 +1956,15 @@ svm_range_evict(struct svm_range *prange, struct mm_struct *mm,
 	pr_debug("invalidate svms 0x%p prange [0x%lx 0x%lx] [0x%lx 0x%lx]\n",
 		 svms, prange->start, prange->last, start, last);
 
-	if (!p->xnack_enabled) {
+	if (!p->xnack_enabled ||
+	    (prange->flags & KFD_IOCTL_SVM_FLAG_GPU_ALWAYS_MAPPED)) {
 		int evicted_ranges;
+		bool mapped = prange->mapped_to_gpu;
 
 		list_for_each_entry(pchild, &prange->child_list, child_list) {
+			if (!pchild->mapped_to_gpu)
+				continue;
+			mapped = true;
 			mutex_lock_nested(&pchild->lock, 1);
 			if (pchild->start <= last && pchild->last >= start) {
 				pr_debug("increment pchild invalid [0x%lx 0x%lx]\n",
@@ -1696,6 +1974,9 @@ svm_range_evict(struct svm_range *prange, struct mm_struct *mm,
 			mutex_unlock(&pchild->lock);
 		}
 
+		if (!mapped)
+			return r;
+
 		if (prange->start <= last && prange->last >= start)
 			atomic_inc(&prange->invalid);
 
@@ -1707,15 +1988,21 @@ svm_range_evict(struct svm_range *prange, struct mm_struct *mm,
 			 prange->svms, prange->start, prange->last);
 
 		/* First eviction, stop the queues */
-		r = kgd2kfd_quiesce_mm(mm);
+		r = kgd2kfd_quiesce_mm(mm, KFD_QUEUE_EVICTION_TRIGGER_SVM);
 		if (r)
 			pr_debug("failed to quiesce KFD\n");
 
 		pr_debug("schedule to restore svm %p ranges\n", svms);
-		schedule_delayed_work(&svms->restore_work,
+		queue_delayed_work(system_freezable_wq, &svms->restore_work,
 			msecs_to_jiffies(AMDGPU_SVM_RANGE_RESTORE_DELAY_MS));
 	} else {
 		unsigned long s, l;
+		uint32_t trigger;
+
+		if (event == MMU_NOTIFY_MIGRATE)
+			trigger = KFD_SVM_UNMAP_TRIGGER_MMU_NOTIFY_MIGRATE;
+		else
+			trigger = KFD_SVM_UNMAP_TRIGGER_MMU_NOTIFY;
 
 		pr_debug("invalidate unmap svms 0x%p [0x%lx 0x%lx] from GPUs\n",
 			 prange->svms, start, last);
@@ -1724,13 +2011,13 @@ svm_range_evict(struct svm_range *prange, struct mm_struct *mm,
 			s = max(start, pchild->start);
 			l = min(last, pchild->last);
 			if (l >= s)
-				svm_range_unmap_from_gpus(pchild, s, l);
+				svm_range_unmap_from_gpus(pchild, s, l, trigger);
 			mutex_unlock(&pchild->lock);
 		}
 		s = max(start, prange->start);
 		l = min(last, prange->last);
 		if (l >= s)
-			svm_range_unmap_from_gpus(prange, s, l);
+			svm_range_unmap_from_gpus(prange, s, l, trigger);
 	}
 
 	return r;
@@ -1740,10 +2027,13 @@ static struct svm_range *svm_range_clone(struct svm_range *old)
 {
 	struct svm_range *new;
 
-	new = svm_range_new(old->svms, old->start, old->last);
+	new = svm_range_new(old->svms, old->start, old->last, false);
 	if (!new)
 		return NULL;
-
+	if (svm_range_copy_dma_addrs(new, old)) {
+		svm_range_free(new, false);
+		return NULL;
+	}
 	if (old->svm_bo) {
 		new->ttm_res = old->ttm_res;
 		new->offset = old->offset;
@@ -1757,92 +2047,159 @@ static struct svm_range *svm_range_clone(struct svm_range *old)
 	new->prefetch_loc = old->prefetch_loc;
 	new->actual_loc = old->actual_loc;
 	new->granularity = old->granularity;
+	new->mapped_to_gpu = old->mapped_to_gpu;
+	new->vram_pages = old->vram_pages;
 	bitmap_copy(new->bitmap_access, old->bitmap_access, MAX_GPU_INSTANCE);
 	bitmap_copy(new->bitmap_aip, old->bitmap_aip, MAX_GPU_INSTANCE);
+	atomic_set(&new->queue_refcount, atomic_read(&old->queue_refcount));
 
 	return new;
 }
 
+void svm_range_set_max_pages(struct amdgpu_device *adev)
+{
+	uint64_t max_pages;
+	uint64_t pages, _pages;
+	uint64_t min_pages = 0;
+	int i, id;
+
+	for (i = 0; i < adev->kfd.dev->num_nodes; i++) {
+		if (adev->kfd.dev->nodes[i]->xcp)
+			id = adev->kfd.dev->nodes[i]->xcp->id;
+		else
+			id = -1;
+		pages = KFD_XCP_MEMORY_SIZE(adev, id) >> 17;
+		pages = clamp(pages, 1ULL << 9, 1ULL << 18);
+		pages = rounddown_pow_of_two(pages);
+		min_pages = min_not_zero(min_pages, pages);
+	}
+
+	do {
+		max_pages = READ_ONCE(max_svm_range_pages);
+		_pages = min_not_zero(max_pages, min_pages);
+	} while (cmpxchg(&max_svm_range_pages, max_pages, _pages) != max_pages);
+}
+
+static int
+svm_range_split_new(struct svm_range_list *svms, uint64_t start, uint64_t last,
+		    uint64_t max_pages, struct list_head *insert_list,
+		    struct list_head *update_list)
+{
+	struct svm_range *prange;
+	uint64_t l;
+
+	pr_debug("max_svm_range_pages 0x%llx adding [0x%llx 0x%llx]\n",
+		 max_pages, start, last);
+
+	while (last >= start) {
+		l = min(last, ALIGN_DOWN(start + max_pages, max_pages) - 1);
+
+		prange = svm_range_new(svms, start, l, true);
+		if (!prange)
+			return -ENOMEM;
+		list_add(&prange->list, insert_list);
+		list_add(&prange->update_list, update_list);
+
+		start = l + 1;
+	}
+	return 0;
+}
+
 /**
- * svm_range_handle_overlap - split overlap ranges
- * @svms: svm range list header
- * @new: range added with this attributes
- * @start: range added start address, in pages
- * @last: range last address, in pages
- * @update_list: output, the ranges attributes are updated. For set_attr, this
- *               will do validation and map to GPUs. For unmap, this will be
- *               removed and unmap from GPUs
- * @insert_list: output, the ranges will be inserted into svms, attributes are
- *               not changes. For set_attr, this will add into svms.
- * @remove_list:output, the ranges will be removed from svms
- * @left: the remaining range after overlap, For set_attr, this will be added
- *        as new range.
+ * svm_range_add - add svm range and handle overlap
+ * @p: the range add to this process svms
+ * @start: page size aligned
+ * @size: page size aligned
+ * @nattr: number of attributes
+ * @attrs: array of attributes
+ * @update_list: output, the ranges need validate and update GPU mapping
+ * @insert_list: output, the ranges need insert to svms
+ * @remove_list: output, the ranges are replaced and need remove from svms
+ * @remap_list: output, remap unaligned svm ranges
  *
- * Total have 5 overlap cases.
+ * Check if the virtual address range has overlap with any existing ranges,
+ * split partly overlapping ranges and add new ranges in the gaps. All changes
+ * should be applied to the range_list and interval tree transactionally. If
+ * any range split or allocation fails, the entire update fails. Therefore any
+ * existing overlapping svm_ranges are cloned and the original svm_ranges left
+ * unchanged.
  *
- * This function handles overlap of an address interval with existing
- * struct svm_ranges for applying new attributes. This may require
- * splitting existing struct svm_ranges. All changes should be applied to
- * the range_list and interval tree transactionally. If any split operation
- * fails, the entire update fails. Therefore the existing overlapping
- * svm_ranges are cloned and the original svm_ranges left unchanged. If the
- * transaction succeeds, the modified clones are added and the originals
- * freed. Otherwise the clones are removed and the old svm_ranges remain.
+ * If the transaction succeeds, the caller can update and insert clones and
+ * new ranges, then free the originals.
  *
- * Context: The caller must hold svms->lock
+ * Otherwise the caller can free the clones and new ranges, while the old
+ * svm_ranges remain unchanged.
+ *
+ * Context: Process context, caller must hold svms->lock
+ *
+ * Return:
+ * 0 - OK, otherwise error code
  */
 static int
-svm_range_handle_overlap(struct svm_range_list *svms, struct svm_range *new,
-			 unsigned long start, unsigned long last,
-			 struct list_head *update_list,
-			 struct list_head *insert_list,
-			 struct list_head *remove_list,
-			 unsigned long *left)
+svm_range_add(struct kfd_process *p, uint64_t start, uint64_t size,
+	      uint32_t nattr, struct kfd_ioctl_svm_attribute *attrs,
+	      struct list_head *update_list, struct list_head *insert_list,
+	      struct list_head *remove_list, struct list_head *remap_list)
 {
+	unsigned long last = start + size - 1UL;
+	struct svm_range_list *svms = &p->svms;
 	struct interval_tree_node *node;
 	struct svm_range *prange;
 	struct svm_range *tmp;
+	struct list_head new_list;
 	int r = 0;
 
+	pr_debug("svms 0x%p [0x%llx 0x%lx]\n", &p->svms, start, last);
+
 	INIT_LIST_HEAD(update_list);
 	INIT_LIST_HEAD(insert_list);
 	INIT_LIST_HEAD(remove_list);
+	INIT_LIST_HEAD(&new_list);
+	INIT_LIST_HEAD(remap_list);
 
 	node = interval_tree_iter_first(&svms->objects, start, last);
 	while (node) {
 		struct interval_tree_node *next;
-		struct svm_range *old;
 		unsigned long next_start;
 
 		pr_debug("found overlap node [0x%lx 0x%lx]\n", node->start,
 			 node->last);
 
-		old = container_of(node, struct svm_range, it_node);
+		prange = container_of(node, struct svm_range, it_node);
 		next = interval_tree_iter_next(node, start, last);
 		next_start = min(node->last, last) + 1;
 
-		if (node->start < start || node->last > last) {
-			/* node intersects the updated range, clone+split it */
+		if (svm_range_is_same_attrs(p, prange, nattr, attrs) &&
+		    prange->mapped_to_gpu) {
+			/* nothing to do */
+		} else if (node->start < start || node->last > last) {
+			/* node intersects the update range and its attributes
+			 * will change. Clone and split it, apply updates only
+			 * to the overlapping part
+			 */
+			struct svm_range *old = prange;
+
 			prange = svm_range_clone(old);
 			if (!prange) {
 				r = -ENOMEM;
 				goto out;
 			}
 
-			list_add(&old->remove_list, remove_list);
-			list_add(&prange->insert_list, insert_list);
+			list_add(&old->update_list, remove_list);
+			list_add(&prange->list, insert_list);
+			list_add(&prange->update_list, update_list);
 
 			if (node->start < start) {
 				pr_debug("change old range start\n");
-				r = svm_range_split_head(prange, new, start,
-							 insert_list);
+				r = svm_range_split_head(prange, start,
+							 insert_list, remap_list);
 				if (r)
 					goto out;
 			}
 			if (node->last > last) {
 				pr_debug("change old range last\n");
-				r = svm_range_split_tail(prange, new, last,
-							 insert_list);
+				r = svm_range_split_tail(prange, last,
+							 insert_list, remap_list);
 				if (r)
 					goto out;
 			}
@@ -1850,36 +2207,37 @@ svm_range_handle_overlap(struct svm_range_list *svms, struct svm_range *new,
 			/* The node is contained within start..last,
 			 * just update it
 			 */
-			prange = old;
-		}
-
-		if (!svm_range_is_same_attrs(prange, new))
 			list_add(&prange->update_list, update_list);
+		}
 
 		/* insert a new node if needed */
 		if (node->start > start) {
-			prange = svm_range_new(prange->svms, start,
-					       node->start - 1);
-			if (!prange) {
-				r = -ENOMEM;
+			r = svm_range_split_new(svms, start, node->start - 1,
+						READ_ONCE(max_svm_range_pages),
+						&new_list, update_list);
+			if (r)
 				goto out;
-			}
-
-			list_add(&prange->insert_list, insert_list);
-			list_add(&prange->update_list, update_list);
 		}
 
 		node = next;
 		start = next_start;
 	}
 
-	if (left && start <= last)
-		*left = last - start + 1;
+	/* add a final range at the end if needed */
+	if (start <= last)
+		r = svm_range_split_new(svms, start, last,
+					READ_ONCE(max_svm_range_pages),
+					&new_list, update_list);
 
 out:
-	if (r)
-		list_for_each_entry_safe(prange, tmp, insert_list, insert_list)
-			svm_range_free(prange);
+	if (r) {
+		list_for_each_entry_safe(prange, tmp, insert_list, list)
+			svm_range_free(prange, false);
+		list_for_each_entry_safe(prange, tmp, &new_list, list)
+			svm_range_free(prange, true);
+	} else {
+		list_splice(&new_list, insert_list);
+	}
 
 	return r;
 }
@@ -1913,10 +2271,9 @@ svm_range_update_notifier_and_interval_tree(struct mm_struct *mm,
 }
 
 static void
-svm_range_handle_list_op(struct svm_range_list *svms, struct svm_range *prange)
+svm_range_handle_list_op(struct svm_range_list *svms, struct svm_range *prange,
+			 struct mm_struct *mm)
 {
-	struct mm_struct *mm = prange->work_item.mm;
-
 	switch (prange->work_item.op) {
 	case SVM_OP_NULL:
 		pr_debug("NULL OP 0x%p prange 0x%p [0x%lx 0x%lx]\n",
@@ -1927,7 +2284,7 @@ svm_range_handle_list_op(struct svm_range_list *svms, struct svm_range *prange)
 			 svms, prange, prange->start, prange->last);
 		svm_range_unlink(prange);
 		svm_range_remove_notifier(prange);
-		svm_range_free(prange);
+		svm_range_free(prange, true);
 		break;
 	case SVM_OP_UPDATE_RANGE_NOTIFIER:
 		pr_debug("update notifier 0x%p prange 0x%p [0x%lx 0x%lx]\n",
@@ -1962,31 +2319,30 @@ svm_range_handle_list_op(struct svm_range_list *svms, struct svm_range *prange)
 static void svm_range_drain_retry_fault(struct svm_range_list *svms)
 {
 	struct kfd_process_device *pdd;
-	struct amdgpu_device *adev;
 	struct kfd_process *p;
-	int drain;
 	uint32_t i;
 
 	p = container_of(svms, struct kfd_process, svms);
 
-restart:
-	drain = atomic_read(&svms->drain_pagefaults);
-	if (!drain)
-		return;
-
 	for_each_set_bit(i, svms->bitmap_supported, p->n_pdds) {
 		pdd = p->pdds[i];
 		if (!pdd)
 			continue;
 
 		pr_debug("drain retry fault gpu %d svms %p\n", i, svms);
-		adev = (struct amdgpu_device *)pdd->dev->kgd;
 
-		amdgpu_ih_wait_on_checkpoint_process(adev, &adev->irq.ih1);
+		amdgpu_ih_wait_on_checkpoint_process_ts(pdd->dev->adev,
+				pdd->dev->adev->irq.retry_cam_enabled ?
+				&pdd->dev->adev->irq.ih :
+				&pdd->dev->adev->irq.ih1);
+
+		if (pdd->dev->adev->irq.retry_cam_enabled)
+			amdgpu_ih_wait_on_checkpoint_process_ts(pdd->dev->adev,
+				&pdd->dev->adev->irq.ih_soft);
+
+
 		pr_debug("drain retry fault gpu %d svms 0x%p done\n", i, svms);
 	}
-	if (atomic_cmpxchg(&svms->drain_pagefaults, drain, 0) != drain)
-		goto restart;
 }
 
 static void svm_range_deferred_list_work(struct work_struct *work)
@@ -1994,40 +2350,35 @@ static void svm_range_deferred_list_work(struct work_struct *work)
 	struct svm_range_list *svms;
 	struct svm_range *prange;
 	struct mm_struct *mm;
-	struct kfd_process *p;
 
 	svms = container_of(work, struct svm_range_list, deferred_list_work);
 	pr_debug("enter svms 0x%p\n", svms);
 
-	p = container_of(svms, struct kfd_process, svms);
-	/* Avoid mm is gone when inserting mmu notifier */
-	mm = get_task_mm(p->lead_thread);
-	if (!mm) {
-		pr_debug("svms 0x%p process mm gone\n", svms);
-		return;
-	}
-retry:
-	mmap_write_lock(mm);
-
-	/* Checking for the need to drain retry faults must be inside
-	 * mmap write lock to serialize with munmap notifiers.
-	 */
-	if (unlikely(atomic_read(&svms->drain_pagefaults))) {
-		mmap_write_unlock(mm);
-		svm_range_drain_retry_fault(svms);
-		goto retry;
-	}
-
 	spin_lock(&svms->deferred_list_lock);
 	while (!list_empty(&svms->deferred_range_list)) {
 		prange = list_first_entry(&svms->deferred_range_list,
 					  struct svm_range, deferred_list);
-		list_del_init(&prange->deferred_list);
 		spin_unlock(&svms->deferred_list_lock);
 
 		pr_debug("prange 0x%p [0x%lx 0x%lx] op %d\n", prange,
 			 prange->start, prange->last, prange->work_item.op);
 
+		mm = prange->work_item.mm;
+
+		mmap_write_lock(mm);
+
+		/* Remove from deferred_list must be inside mmap write lock, for
+		 * two race cases:
+		 * 1. unmap_from_cpu may change work_item.op and add the range
+		 *    to deferred_list again, cause use after free bug.
+		 * 2. svm_range_list_lock_and_flush_work may hold mmap write
+		 *    lock and continue because deferred_list is empty, but
+		 *    deferred_list work is actually waiting for mmap lock.
+		 */
+		spin_lock(&svms->deferred_list_lock);
+		list_del_init(&prange->deferred_list);
+		spin_unlock(&svms->deferred_list_lock);
+
 		mutex_lock(&svms->lock);
 		mutex_lock(&prange->migrate_mutex);
 		while (!list_empty(&prange->child_list)) {
@@ -2038,19 +2389,22 @@ retry:
 			pr_debug("child prange 0x%p op %d\n", pchild,
 				 pchild->work_item.op);
 			list_del_init(&pchild->child_list);
-			svm_range_handle_list_op(svms, pchild);
+			svm_range_handle_list_op(svms, pchild, mm);
 		}
 		mutex_unlock(&prange->migrate_mutex);
 
-		svm_range_handle_list_op(svms, prange);
+		svm_range_handle_list_op(svms, prange, mm);
 		mutex_unlock(&svms->lock);
+		mmap_write_unlock(mm);
+
+		/* Pairs with mmget in svm_range_add_list_work. If dropping the
+		 * last mm refcount, schedule release work to avoid circular locking
+		 */
+		mmput_async(mm);
 
 		spin_lock(&svms->deferred_list_lock);
 	}
 	spin_unlock(&svms->deferred_list_lock);
-
-	mmap_write_unlock(mm);
-	mmput(mm);
 	pr_debug("exit svms 0x%p\n", svms);
 }
 
@@ -2067,12 +2421,17 @@ svm_range_add_list_work(struct svm_range_list *svms, struct svm_range *prange,
 		    prange->work_item.op != SVM_OP_UNMAP_RANGE)
 			prange->work_item.op = op;
 	} else {
-		prange->work_item.op = op;
-		prange->work_item.mm = mm;
-		list_add_tail(&prange->deferred_list,
-			      &prange->svms->deferred_range_list);
-		pr_debug("add prange 0x%p [0x%lx 0x%lx] to work list op %d\n",
-			 prange, prange->start, prange->last, op);
+		/* Pairs with mmput in deferred_list_work.
+		 * If process is exiting and mm is gone, don't update mmu notifier.
+		 */
+		if (mmget_not_zero(mm)) {
+			prange->work_item.mm = mm;
+			prange->work_item.op = op;
+			list_add_tail(&prange->deferred_list,
+				      &prange->svms->deferred_range_list);
+			pr_debug("add prange 0x%p [0x%lx 0x%lx] to work list op %d\n",
+				 prange, prange->start, prange->last, op);
+		}
 	}
 	spin_unlock(&svms->deferred_list_lock);
 }
@@ -2086,8 +2445,7 @@ void schedule_deferred_list_work(struct svm_range_list *svms)
 }
 
 static void
-svm_range_unmap_split(struct mm_struct *mm, struct svm_range *parent,
-		      struct svm_range *prange, unsigned long start,
+svm_range_unmap_split(struct svm_range *parent, struct svm_range *prange, unsigned long start,
 		      unsigned long last)
 {
 	struct svm_range *head;
@@ -2108,12 +2466,12 @@ svm_range_unmap_split(struct mm_struct *mm, struct svm_range *parent,
 		svm_range_split(tail, last + 1, tail->last, &head);
 
 	if (head != prange && tail != prange) {
-		svm_range_add_child(parent, mm, head, SVM_OP_UNMAP_RANGE);
-		svm_range_add_child(parent, mm, tail, SVM_OP_ADD_RANGE);
+		svm_range_add_child(parent, head, SVM_OP_UNMAP_RANGE);
+		svm_range_add_child(parent, tail, SVM_OP_ADD_RANGE);
 	} else if (tail != prange) {
-		svm_range_add_child(parent, mm, tail, SVM_OP_UNMAP_RANGE);
+		svm_range_add_child(parent, tail, SVM_OP_UNMAP_RANGE);
 	} else if (head != prange) {
-		svm_range_add_child(parent, mm, head, SVM_OP_UNMAP_RANGE);
+		svm_range_add_child(parent, head, SVM_OP_UNMAP_RANGE);
 	} else if (parent != prange) {
 		prange->work_item.op = SVM_OP_UNMAP_RANGE;
 	}
@@ -2123,11 +2481,23 @@ static void
 svm_range_unmap_from_cpu(struct mm_struct *mm, struct svm_range *prange,
 			 unsigned long start, unsigned long last)
 {
+	uint32_t trigger = KFD_SVM_UNMAP_TRIGGER_UNMAP_FROM_CPU;
 	struct svm_range_list *svms;
 	struct svm_range *pchild;
 	struct kfd_process *p;
 	unsigned long s, l;
 	bool unmap_parent;
+	uint32_t i;
+
+	if (atomic_read(&prange->queue_refcount)) {
+		int r;
+
+		pr_warn("Freeing queue vital buffer 0x%lx, queue evicted\n",
+			prange->start << PAGE_SHIFT);
+		r = kgd2kfd_quiesce_mm(mm, KFD_QUEUE_EVICTION_TRIGGER_SVM);
+		if (r)
+			pr_debug("failed %d to quiesce KFD queues\n", r);
+	}
 
 	p = kfd_lookup_process_by_mm(mm);
 	if (!p)
@@ -2137,11 +2507,38 @@ svm_range_unmap_from_cpu(struct mm_struct *mm, struct svm_range *prange,
 	pr_debug("svms 0x%p prange 0x%p [0x%lx 0x%lx] [0x%lx 0x%lx]\n", svms,
 		 prange, prange->start, prange->last, start, last);
 
-	/* Make sure pending page faults are drained in the deferred worker
-	 * before the range is freed to avoid straggler interrupts on
-	 * unmapped memory causing "phantom faults".
+	/* calculate time stamps that are used to decide which page faults need be
+	 * dropped or handled before unmap pages from gpu vm
 	 */
-	atomic_inc(&svms->drain_pagefaults);
+	for_each_set_bit(i, svms->bitmap_supported, p->n_pdds) {
+		struct kfd_process_device *pdd;
+		struct amdgpu_device *adev;
+		struct amdgpu_ih_ring *ih;
+		uint32_t checkpoint_wptr;
+
+		pdd = p->pdds[i];
+		if (!pdd)
+			continue;
+
+		adev = pdd->dev->adev;
+
+		/* Check and drain ih1 ring if cam not available */
+		if (adev->irq.ih1.ring_size) {
+			ih = &adev->irq.ih1;
+			checkpoint_wptr = amdgpu_ih_get_wptr(adev, ih);
+			if (ih->rptr != checkpoint_wptr) {
+				svms->checkpoint_ts[i] =
+					amdgpu_ih_decode_iv_ts(adev, ih, checkpoint_wptr, -1);
+				continue;
+			}
+		}
+
+		/* check if dev->irq.ih_soft is not empty */
+		ih = &adev->irq.ih_soft;
+		checkpoint_wptr = amdgpu_ih_get_wptr(adev, ih);
+		if (ih->rptr != checkpoint_wptr)
+			svms->checkpoint_ts[i] = amdgpu_ih_decode_iv_ts(adev, ih, checkpoint_wptr, -1);
+	}
 
 	unmap_parent = start <= prange->start && last >= prange->last;
 
@@ -2150,15 +2547,15 @@ svm_range_unmap_from_cpu(struct mm_struct *mm, struct svm_range *prange,
 		s = max(start, pchild->start);
 		l = min(last, pchild->last);
 		if (l >= s)
-			svm_range_unmap_from_gpus(pchild, s, l);
-		svm_range_unmap_split(mm, prange, pchild, start, last);
+			svm_range_unmap_from_gpus(pchild, s, l, trigger);
+		svm_range_unmap_split(prange, pchild, start, last);
 		mutex_unlock(&pchild->lock);
 	}
 	s = max(start, prange->start);
 	l = min(last, prange->last);
 	if (l >= s)
-		svm_range_unmap_from_gpus(prange, s, l);
-	svm_range_unmap_split(mm, prange, prange, start, last);
+		svm_range_unmap_from_gpus(prange, s, l, trigger);
+	svm_range_unmap_split(prange, prange, start, last);
 
 	if (unmap_parent)
 		svm_range_add_list_work(svms, prange, mm, SVM_OP_UNMAP_RANGE);
@@ -2172,6 +2569,9 @@ svm_range_unmap_from_cpu(struct mm_struct *mm, struct svm_range *prange,
 
 /**
  * svm_range_cpu_invalidate_pagetables - interval notifier callback
+ * @mni: mmu_interval_notifier struct
+ * @range: mmu_notifier_range struct
+ * @cur_seq: value to pass to mmu_interval_set_seq()
  *
  * If event is MMU_NOTIFY_UNMAP, this is from CPU unmap range, otherwise, it
  * is from migration, or CPU page invalidation callback.
@@ -2201,8 +2601,8 @@ svm_range_cpu_invalidate_pagetables(struct mmu_interval_notifier *mni,
 
 	start = mni->interval_tree.start;
 	last = mni->interval_tree.last;
-	start = (start > range->start ? start : range->start) >> PAGE_SHIFT;
-	last = (last < (range->end - 1) ? last : range->end - 1) >> PAGE_SHIFT;
+	start = max(start, range->start) >> PAGE_SHIFT;
+	last = min(last, range->end - 1) >> PAGE_SHIFT;
 	pr_debug("[0x%lx 0x%lx] range[0x%lx 0x%lx] notifier[0x%lx 0x%lx] %d\n",
 		 start, last, range->start >> PAGE_SHIFT,
 		 (range->end - 1) >> PAGE_SHIFT,
@@ -2219,7 +2619,7 @@ svm_range_cpu_invalidate_pagetables(struct mmu_interval_notifier *mni,
 		svm_range_unmap_from_cpu(mni->mm, prange, start, last);
 		break;
 	default:
-		svm_range_evict(prange, mni->mm, start, last);
+		svm_range_evict(prange, mni->mm, start, last, range->event);
 		break;
 	}
 
@@ -2294,29 +2694,31 @@ svm_range_from_addr(struct svm_range_list *svms, unsigned long addr,
  */
 static int32_t
 svm_range_best_restore_location(struct svm_range *prange,
-				struct amdgpu_device *adev,
+				struct kfd_node *node,
 				int32_t *gpuidx)
 {
-	struct amdgpu_device *bo_adev, *preferred_adev;
+	struct kfd_node *bo_node, *preferred_node;
 	struct kfd_process *p;
 	uint32_t gpuid;
 	int r;
 
 	p = container_of(prange->svms, struct kfd_process, svms);
 
-	r = kfd_process_gpuid_from_kgd(p, adev, &gpuid, gpuidx);
+	r = kfd_process_gpuid_from_node(p, node, &gpuid, gpuidx);
 	if (r < 0) {
 		pr_debug("failed to get gpuid from kgd\n");
 		return -1;
 	}
 
+	if (node->adev->apu_prefer_gtt)
+		return 0;
+
 	if (prange->preferred_loc == gpuid ||
 	    prange->preferred_loc == KFD_IOCTL_SVM_LOCATION_SYSMEM) {
 		return prange->preferred_loc;
 	} else if (prange->preferred_loc != KFD_IOCTL_SVM_LOCATION_UNDEFINED) {
-		preferred_adev = svm_range_get_adev_by_id(prange,
-							prange->preferred_loc);
-		if (amdgpu_xgmi_same_hive(adev, preferred_adev))
+		preferred_node = svm_range_get_node_by_id(prange, prange->preferred_loc);
+		if (preferred_node && svm_nodes_in_same_hive(node, preferred_node))
 			return prange->preferred_loc;
 		/* fall through */
 	}
@@ -2328,8 +2730,8 @@ svm_range_best_restore_location(struct svm_range *prange,
 		if (!prange->actual_loc)
 			return 0;
 
-		bo_adev = svm_range_get_adev_by_id(prange, prange->actual_loc);
-		if (amdgpu_xgmi_same_hive(adev, bo_adev))
+		bo_node = svm_range_get_node_by_id(prange, prange->actual_loc);
+		if (bo_node && svm_nodes_in_same_hive(node, bo_node))
 			return prange->actual_loc;
 		else
 			return 0;
@@ -2345,38 +2747,36 @@ svm_range_get_range_boundaries(struct kfd_process *p, int64_t addr,
 {
 	struct vm_area_struct *vma;
 	struct interval_tree_node *node;
+	struct rb_node *rb_node;
 	unsigned long start_limit, end_limit;
 
-	vma = find_vma(p->mm, addr << PAGE_SHIFT);
-	if (!vma || (addr << PAGE_SHIFT) < vma->vm_start) {
+	vma = vma_lookup(p->mm, addr << PAGE_SHIFT);
+	if (!vma) {
 		pr_debug("VMA does not exist in address [0x%llx]\n", addr);
 		return -EFAULT;
 	}
 
-	*is_heap_stack = (vma->vm_start <= vma->vm_mm->brk &&
-			  vma->vm_end >= vma->vm_mm->start_brk) ||
-			 (vma->vm_start <= vma->vm_mm->start_stack &&
-			  vma->vm_end >= vma->vm_mm->start_stack);
+	*is_heap_stack = vma_is_initial_heap(vma) || vma_is_initial_stack(vma);
 
 	start_limit = max(vma->vm_start >> PAGE_SHIFT,
-		      (unsigned long)ALIGN_DOWN(addr, 2UL << 8));
+		      (unsigned long)ALIGN_DOWN(addr, 1UL << p->svms.default_granularity));
 	end_limit = min(vma->vm_end >> PAGE_SHIFT,
-		    (unsigned long)ALIGN(addr + 1, 2UL << 8));
+		    (unsigned long)ALIGN(addr + 1, 1UL << p->svms.default_granularity));
+
 	/* First range that starts after the fault address */
 	node = interval_tree_iter_first(&p->svms.objects, addr + 1, ULONG_MAX);
 	if (node) {
 		end_limit = min(end_limit, node->start);
 		/* Last range that ends before the fault address */
-		node = container_of(rb_prev(&node->rb),
-				    struct interval_tree_node, rb);
+		rb_node = rb_prev(&node->rb);
 	} else {
 		/* Last range must end before addr because
 		 * there was no range after addr
 		 */
-		node = container_of(rb_last(&p->svms.objects.rb_root),
-				    struct interval_tree_node, rb);
+		rb_node = rb_last(&p->svms.objects.rb_root);
 	}
-	if (node) {
+	if (rb_node) {
+		node = container_of(rb_node, struct interval_tree_node, rb);
 		if (node->last >= addr) {
 			WARN(1, "Overlap with prev node and page fault addr\n");
 			return -EFAULT;
@@ -2446,7 +2846,7 @@ svm_range_check_vm_userptr(struct kfd_process *p, uint64_t start, uint64_t last,
 }
 
 static struct
-svm_range *svm_range_create_unregistered_range(struct amdgpu_device *adev,
+svm_range *svm_range_create_unregistered_range(struct kfd_node *node,
 						struct kfd_process *p,
 						struct mm_struct *mm,
 						int64_t addr)
@@ -2476,14 +2876,14 @@ svm_range *svm_range_create_unregistered_range(struct amdgpu_device *adev,
 		last = addr;
 	}
 
-	prange = svm_range_new(&p->svms, start, last);
+	prange = svm_range_new(&p->svms, start, last, true);
 	if (!prange) {
 		pr_debug("Failed to create prange in address [0x%llx]\n", addr);
 		return NULL;
 	}
-	if (kfd_process_gpuid_from_kgd(p, adev, &gpuid, &gpuidx)) {
+	if (kfd_process_gpuid_from_node(p, node, &gpuid, &gpuidx)) {
 		pr_debug("failed to get gpuid from kgd\n");
-		svm_range_free(prange);
+		svm_range_free(prange, true);
 		return NULL;
 	}
 
@@ -2535,7 +2935,7 @@ static bool svm_range_skip_recover(struct svm_range *prange)
 }
 
 static void
-svm_range_count_fault(struct amdgpu_device *adev, struct kfd_process *p,
+svm_range_count_fault(struct kfd_node *node, struct kfd_process *p,
 		      int32_t gpuidx)
 {
 	struct kfd_process_device *pdd;
@@ -2548,7 +2948,7 @@ svm_range_count_fault(struct amdgpu_device *adev, struct kfd_process *p,
 		uint32_t gpuid;
 		int r;
 
-		r = kfd_process_gpuid_from_kgd(p, adev, &gpuid, &gpuidx);
+		r = kfd_process_gpuid_from_node(p, node, &gpuid, &gpuidx);
 		if (r < 0)
 			return;
 	}
@@ -2576,44 +2976,63 @@ svm_fault_allowed(struct vm_area_struct *vma, bool write_fault)
 
 int
 svm_range_restore_pages(struct amdgpu_device *adev, unsigned int pasid,
-			uint64_t addr, bool write_fault)
+			uint32_t vmid, uint32_t node_id,
+			uint64_t addr, uint64_t ts, bool write_fault)
 {
+	unsigned long start, last, size;
 	struct mm_struct *mm = NULL;
 	struct svm_range_list *svms;
 	struct svm_range *prange;
 	struct kfd_process *p;
-	uint64_t timestamp;
+	ktime_t timestamp = ktime_get_boottime();
+	struct kfd_node *node;
 	int32_t best_loc;
-	int32_t gpuidx = MAX_GPU_INSTANCE;
+	int32_t gpuid, gpuidx = MAX_GPU_INSTANCE;
 	bool write_locked = false;
 	struct vm_area_struct *vma;
+	bool migration = false;
 	int r = 0;
 
-	if (!KFD_IS_SVM_API_SUPPORTED(adev->kfd.dev)) {
+	if (!KFD_IS_SVM_API_SUPPORTED(adev)) {
 		pr_debug("device does not support SVM\n");
 		return -EFAULT;
 	}
 
-	p = kfd_lookup_process_by_pasid(pasid);
+	p = kfd_lookup_process_by_pasid(pasid, NULL);
 	if (!p) {
 		pr_debug("kfd process not founded pasid 0x%x\n", pasid);
 		return 0;
 	}
-	if (!p->xnack_enabled) {
-		pr_debug("XNACK not enabled for pasid 0x%x\n", pasid);
-		r = -EFAULT;
-		goto out;
-	}
 	svms = &p->svms;
 
 	pr_debug("restoring svms 0x%p fault address 0x%llx\n", svms, addr);
 
 	if (atomic_read(&svms->drain_pagefaults)) {
-		pr_debug("draining retry fault, drop fault 0x%llx\n", addr);
+		pr_debug("page fault handling disabled, drop fault 0x%llx\n", addr);
 		r = 0;
 		goto out;
 	}
 
+	node = kfd_node_by_irq_ids(adev, node_id, vmid);
+	if (!node) {
+		pr_debug("kfd node does not exist node_id: %d, vmid: %d\n", node_id,
+			 vmid);
+		r = -EFAULT;
+		goto out;
+	}
+
+	if (kfd_process_gpuid_from_node(p, node, &gpuid, &gpuidx)) {
+		pr_debug("failed to get gpuid/gpuidex for node_id: %d\n", node_id);
+		r = -EFAULT;
+		goto out;
+	}
+
+	if (!p->xnack_enabled) {
+		pr_debug("XNACK not enabled for pasid 0x%x\n", pasid);
+		r = -EFAULT;
+		goto out;
+	}
+
 	/* p->lead_thread is available as kfd_process_wq_release flush the work
 	 * before releasing task ref.
 	 */
@@ -2627,6 +3046,23 @@ svm_range_restore_pages(struct amdgpu_device *adev, unsigned int pasid,
 	mmap_read_lock(mm);
 retry_write_locked:
 	mutex_lock(&svms->lock);
+
+	/* check if this page fault time stamp is before svms->checkpoint_ts */
+	if (svms->checkpoint_ts[gpuidx] != 0) {
+		if (amdgpu_ih_ts_after_or_equal(ts,  svms->checkpoint_ts[gpuidx])) {
+			pr_debug("draining retry fault, drop fault 0x%llx\n", addr);
+			if (write_locked)
+				mmap_write_downgrade(mm);
+			r = -EAGAIN;
+			goto out_unlock_svms;
+		} else {
+			/* ts is after svms->checkpoint_ts now, reset svms->checkpoint_ts
+			 * to zero to avoid following ts wrap around give wrong comparing
+			 */
+			svms->checkpoint_ts[gpuidx] = 0;
+		}
+	}
+
 	prange = svm_range_from_addr(svms, addr, NULL);
 	if (!prange) {
 		pr_debug("failed to find prange svms 0x%p address [0x%llx]\n",
@@ -2642,7 +3078,7 @@ retry_write_locked:
 			write_locked = true;
 			goto retry_write_locked;
 		}
-		prange = svm_range_create_unregistered_range(adev, p, mm, addr);
+		prange = svm_range_create_unregistered_range(node, p, mm, addr);
 		if (!prange) {
 			pr_debug("failed to create unregistered range svms 0x%p address [0x%llx]\n",
 				 svms, addr);
@@ -2657,14 +3093,14 @@ retry_write_locked:
 	mutex_lock(&prange->migrate_mutex);
 
 	if (svm_range_skip_recover(prange)) {
-		amdgpu_gmc_filter_faults_remove(adev, addr, pasid);
+		amdgpu_gmc_filter_faults_remove(node->adev, addr, pasid);
 		r = 0;
 		goto out_unlock_range;
 	}
 
-	timestamp = ktime_to_us(ktime_get()) - prange->validate_timestamp;
 	/* skip duplicate vm fault on different pages of same range */
-	if (timestamp < AMDGPU_SVM_RANGE_RETRY_FAULT_PENDING) {
+	if (ktime_before(timestamp, ktime_add_ns(prange->validate_timestamp,
+				AMDGPU_SVM_RANGE_RETRY_FAULT_PENDING))) {
 		pr_debug("svms 0x%p [0x%lx %lx] already restored\n",
 			 svms, prange->start, prange->last);
 		r = 0;
@@ -2674,8 +3110,8 @@ retry_write_locked:
 	/* __do_munmap removed VMA, return success as we are handling stale
 	 * retry fault.
 	 */
-	vma = find_vma(mm, addr << PAGE_SHIFT);
-	if (!vma || (addr << PAGE_SHIFT) < vma->vm_start) {
+	vma = vma_lookup(mm, addr << PAGE_SHIFT);
+	if (!vma) {
 		pr_debug("address 0x%llx VMA is removed\n", addr);
 		r = 0;
 		goto out_unlock_range;
@@ -2688,7 +3124,7 @@ retry_write_locked:
 		goto out_unlock_range;
 	}
 
-	best_loc = svm_range_best_restore_location(prange, adev, &gpuidx);
+	best_loc = svm_range_best_restore_location(prange, node, &gpuidx);
 	if (best_loc == -1) {
 		pr_debug("svms %p failed get best restore loc [0x%lx 0x%lx]\n",
 			 svms, prange->start, prange->last);
@@ -2700,34 +3136,51 @@ retry_write_locked:
 		 svms, prange->start, prange->last, best_loc,
 		 prange->actual_loc);
 
-	if (prange->actual_loc != best_loc) {
+	kfd_smi_event_page_fault_start(node, p->lead_thread->pid, addr,
+				       write_fault, timestamp);
+
+	/* Align migration range start and size to granularity size */
+	size = 1UL << prange->granularity;
+	start = max_t(unsigned long, ALIGN_DOWN(addr, size), prange->start);
+	last = min_t(unsigned long, ALIGN(addr + 1, size) - 1, prange->last);
+	if (prange->actual_loc != 0 || best_loc != 0) {
 		if (best_loc) {
-			r = svm_migrate_to_vram(prange, best_loc, mm);
+			r = svm_migrate_to_vram(prange, best_loc, start, last,
+					mm, KFD_MIGRATE_TRIGGER_PAGEFAULT_GPU);
 			if (r) {
 				pr_debug("svm_migrate_to_vram failed (%d) at %llx, falling back to system memory\n",
 					 r, addr);
 				/* Fallback to system memory if migration to
 				 * VRAM failed
 				 */
-				if (prange->actual_loc)
-					r = svm_migrate_vram_to_ram(prange, mm);
+				if (prange->actual_loc && prange->actual_loc != best_loc)
+					r = svm_migrate_vram_to_ram(prange, mm, start, last,
+						KFD_MIGRATE_TRIGGER_PAGEFAULT_GPU, NULL);
 				else
 					r = 0;
 			}
 		} else {
-			r = svm_migrate_vram_to_ram(prange, mm);
+			r = svm_migrate_vram_to_ram(prange, mm, start, last,
+					KFD_MIGRATE_TRIGGER_PAGEFAULT_GPU, NULL);
 		}
 		if (r) {
 			pr_debug("failed %d to migrate svms %p [0x%lx 0x%lx]\n",
-				 r, svms, prange->start, prange->last);
-			goto out_unlock_range;
+				 r, svms, start, last);
+			goto out_migrate_fail;
+		} else {
+			migration = true;
 		}
 	}
 
-	r = svm_range_validate_and_map(mm, prange, gpuidx, false, false);
+	r = svm_range_validate_and_map(mm, start, last, prange, gpuidx, false,
+				       false, false);
 	if (r)
 		pr_debug("failed %d to map svms 0x%p [0x%lx 0x%lx] to gpus\n",
-			 r, svms, prange->start, prange->last);
+			 r, svms, start, last);
+
+out_migrate_fail:
+	kfd_smi_event_page_fault_end(node, p->lead_thread->pid, addr,
+				     migration);
 
 out_unlock_range:
 	mutex_unlock(&prange->migrate_mutex);
@@ -2735,7 +3188,8 @@ out_unlock_svms:
 	mutex_unlock(&svms->lock);
 	mmap_read_unlock(mm);
 
-	svm_range_count_fault(adev, p, gpuidx);
+	if (r != -EAGAIN)
+		svm_range_count_fault(node, p, gpuidx);
 
 	mmput(mm);
 out:
@@ -2743,18 +3197,79 @@ out:
 
 	if (r == -EAGAIN) {
 		pr_debug("recover vm fault later\n");
-		amdgpu_gmc_filter_faults_remove(adev, addr, pasid);
+		amdgpu_gmc_filter_faults_remove(node->adev, addr, pasid);
 		r = 0;
 	}
 	return r;
 }
 
+int
+svm_range_switch_xnack_reserve_mem(struct kfd_process *p, bool xnack_enabled)
+{
+	struct svm_range *prange, *pchild;
+	uint64_t reserved_size = 0;
+	uint64_t size;
+	int r = 0;
+
+	pr_debug("switching xnack from %d to %d\n", p->xnack_enabled, xnack_enabled);
+
+	mutex_lock(&p->svms.lock);
+
+	list_for_each_entry(prange, &p->svms.list, list) {
+		svm_range_lock(prange);
+		list_for_each_entry(pchild, &prange->child_list, child_list) {
+			size = (pchild->last - pchild->start + 1) << PAGE_SHIFT;
+			if (xnack_enabled) {
+				amdgpu_amdkfd_unreserve_mem_limit(NULL, size,
+					KFD_IOC_ALLOC_MEM_FLAGS_USERPTR, 0);
+			} else {
+				r = amdgpu_amdkfd_reserve_mem_limit(NULL, size,
+					KFD_IOC_ALLOC_MEM_FLAGS_USERPTR, 0);
+				if (r)
+					goto out_unlock;
+				reserved_size += size;
+			}
+		}
+
+		size = (prange->last - prange->start + 1) << PAGE_SHIFT;
+		if (xnack_enabled) {
+			amdgpu_amdkfd_unreserve_mem_limit(NULL, size,
+					KFD_IOC_ALLOC_MEM_FLAGS_USERPTR, 0);
+		} else {
+			r = amdgpu_amdkfd_reserve_mem_limit(NULL, size,
+					KFD_IOC_ALLOC_MEM_FLAGS_USERPTR, 0);
+			if (r)
+				goto out_unlock;
+			reserved_size += size;
+		}
+out_unlock:
+		svm_range_unlock(prange);
+		if (r)
+			break;
+	}
+
+	if (r)
+		amdgpu_amdkfd_unreserve_mem_limit(NULL, reserved_size,
+					KFD_IOC_ALLOC_MEM_FLAGS_USERPTR, 0);
+	else
+		/* Change xnack mode must be inside svms lock, to avoid race with
+		 * svm_range_deferred_list_work unreserve memory in parallel.
+		 */
+		p->xnack_enabled = xnack_enabled;
+
+	mutex_unlock(&p->svms.lock);
+	return r;
+}
+
 void svm_range_list_fini(struct kfd_process *p)
 {
 	struct svm_range *prange;
 	struct svm_range *next;
 
-	pr_debug("pasid 0x%x svms 0x%p\n", p->pasid, &p->svms);
+	pr_debug("process pid %d svms 0x%p\n", p->lead_thread->pid,
+		 &p->svms);
+
+	cancel_delayed_work_sync(&p->svms.restore_work);
 
 	/* Ensure list work is finished before process is destroyed */
 	flush_work(&p->svms.deferred_list_work);
@@ -2762,20 +3277,21 @@ void svm_range_list_fini(struct kfd_process *p)
 	/*
 	 * Ensure no retry fault comes in afterwards, as page fault handler will
 	 * not find kfd process and take mm lock to recover fault.
+	 * stop kfd page fault handing, then wait pending page faults got drained
 	 */
-	atomic_inc(&p->svms.drain_pagefaults);
+	atomic_set(&p->svms.drain_pagefaults, 1);
 	svm_range_drain_retry_fault(&p->svms);
 
-
 	list_for_each_entry_safe(prange, next, &p->svms.list, list) {
 		svm_range_unlink(prange);
 		svm_range_remove_notifier(prange);
-		svm_range_free(prange);
+		svm_range_free(prange, true);
 	}
 
 	mutex_destroy(&p->svms.lock);
 
-	pr_debug("pasid 0x%x svms 0x%p done\n", p->pasid, &p->svms);
+	pr_debug("process pid %d svms 0x%p done\n",
+		p->lead_thread->pid, &p->svms);
 }
 
 int svm_range_list_init(struct kfd_process *p)
@@ -2791,12 +3307,19 @@ int svm_range_list_init(struct kfd_process *p)
 	INIT_DELAYED_WORK(&svms->restore_work, svm_range_restore_work);
 	INIT_WORK(&svms->deferred_list_work, svm_range_deferred_list_work);
 	INIT_LIST_HEAD(&svms->deferred_range_list);
+	INIT_LIST_HEAD(&svms->criu_svm_metadata_list);
 	spin_lock_init(&svms->deferred_list_lock);
 
 	for (i = 0; i < p->n_pdds; i++)
-		if (KFD_IS_SVM_API_SUPPORTED(p->pdds[i]->dev))
+		if (KFD_IS_SVM_API_SUPPORTED(p->pdds[i]->dev->adev))
 			bitmap_set(svms->bitmap_supported, i, 1);
 
+	 /* Value of default granularity cannot exceed 0x1B, the
+	  * number of pages supported by a 4-level paging table
+	  */
+	svms->default_granularity = min_t(u8, amdgpu_svm_default_granularity, 0x1B);
+	pr_debug("Default SVM Granularity to use: %d\n", svms->default_granularity);
+
 	return 0;
 }
 
@@ -2883,9 +3406,8 @@ svm_range_is_valid(struct kfd_process *p, uint64_t start, uint64_t size)
 	start <<= PAGE_SHIFT;
 	end = start + (size << PAGE_SHIFT);
 	do {
-		vma = find_vma(p->mm, start);
-		if (!vma || start < vma->vm_start ||
-		    (vma->vm_flags & device_vma))
+		vma = vma_lookup(p->mm, start);
+		if (!vma || (vma->vm_flags & device_vma))
 			return -EFAULT;
 		start = min(end, vma->vm_end);
 	} while (start < end);
@@ -2895,59 +3417,6 @@ svm_range_is_valid(struct kfd_process *p, uint64_t start, uint64_t size)
 }
 
 /**
- * svm_range_add - add svm range and handle overlap
- * @p: the range add to this process svms
- * @start: page size aligned
- * @size: page size aligned
- * @nattr: number of attributes
- * @attrs: array of attributes
- * @update_list: output, the ranges need validate and update GPU mapping
- * @insert_list: output, the ranges need insert to svms
- * @remove_list: output, the ranges are replaced and need remove from svms
- *
- * Check if the virtual address range has overlap with the registered ranges,
- * split the overlapped range, copy and adjust pages address and vram nodes in
- * old and new ranges.
- *
- * Context: Process context, caller must hold svms->lock
- *
- * Return:
- * 0 - OK, otherwise error code
- */
-static int
-svm_range_add(struct kfd_process *p, uint64_t start, uint64_t size,
-	      uint32_t nattr, struct kfd_ioctl_svm_attribute *attrs,
-	      struct list_head *update_list, struct list_head *insert_list,
-	      struct list_head *remove_list)
-{
-	uint64_t last = start + size - 1UL;
-	struct svm_range_list *svms;
-	struct svm_range new = {0};
-	struct svm_range *prange;
-	unsigned long left = 0;
-	int r = 0;
-
-	pr_debug("svms 0x%p [0x%llx 0x%llx]\n", &p->svms, start, last);
-
-	svm_range_apply_attrs(p, &new, nattr, attrs);
-
-	svms = &p->svms;
-
-	r = svm_range_handle_overlap(svms, &new, start, last, update_list,
-				     insert_list, remove_list, &left);
-	if (r)
-		return r;
-
-	if (left) {
-		prange = svm_range_new(svms, last - left + 1, last);
-		list_add(&prange->insert_list, insert_list);
-		list_add(&prange->update_list, update_list);
-	}
-
-	return 0;
-}
-
-/**
  * svm_range_best_prefetch_location - decide the best prefetch location
  * @prange: svm range structure
  *
@@ -2979,8 +3448,7 @@ svm_range_best_prefetch_location(struct svm_range *prange)
 	DECLARE_BITMAP(bitmap, MAX_GPU_INSTANCE);
 	uint32_t best_loc = prange->prefetch_loc;
 	struct kfd_process_device *pdd;
-	struct amdgpu_device *bo_adev;
-	struct amdgpu_device *adev;
+	struct kfd_node *bo_node;
 	struct kfd_process *p;
 	uint32_t gpuidx;
 
@@ -2989,9 +3457,14 @@ svm_range_best_prefetch_location(struct svm_range *prange)
 	if (!best_loc || best_loc == KFD_IOCTL_SVM_LOCATION_UNDEFINED)
 		goto out;
 
-	bo_adev = svm_range_get_adev_by_id(prange, best_loc);
-	if (!bo_adev) {
-		WARN_ONCE(1, "failed to get device by id 0x%x\n", best_loc);
+	bo_node = svm_range_get_node_by_id(prange, best_loc);
+	if (!bo_node) {
+		WARN_ONCE(1, "failed to get valid kfd node at id%x\n", best_loc);
+		best_loc = 0;
+		goto out;
+	}
+
+	if (bo_node->adev->apu_prefer_gtt) {
 		best_loc = 0;
 		goto out;
 	}
@@ -3008,12 +3481,11 @@ svm_range_best_prefetch_location(struct svm_range *prange)
 			pr_debug("failed to get device by idx 0x%x\n", gpuidx);
 			continue;
 		}
-		adev = (struct amdgpu_device *)pdd->dev->kgd;
 
-		if (adev == bo_adev)
+		if (pdd->dev->adev == bo_node->adev)
 			continue;
 
-		if (!amdgpu_xgmi_same_hive(adev, bo_adev)) {
+		if (!svm_nodes_in_same_hive(pdd->dev, bo_node)) {
 			best_loc = 0;
 			break;
 		}
@@ -3027,28 +3499,6 @@ out:
 	return best_loc;
 }
 
-/* FIXME: This is a workaround for page locking bug when some pages are
- * invalid during migration to VRAM
- */
-void svm_range_prefault(struct svm_range *prange, struct mm_struct *mm,
-			void *owner)
-{
-	struct hmm_range *hmm_range;
-	int r;
-
-	if (prange->validated_once)
-		return;
-
-	r = amdgpu_hmm_range_get_pages(&prange->notifier, mm, NULL,
-				       prange->start << PAGE_SHIFT,
-				       prange->npages, &hmm_range,
-				       false, true, owner);
-	if (!r) {
-		amdgpu_hmm_range_get_pages_done(hmm_range);
-		prange->validated_once = true;
-	}
-}
-
 /* svm_range_trigger_migration - start page migration if prefetch loc changed
  * @mm: current process mm_struct
  * @prange: svm range structure
@@ -3083,31 +3533,39 @@ svm_range_trigger_migration(struct mm_struct *mm, struct svm_range *prange,
 	*migrated = false;
 	best_loc = svm_range_best_prefetch_location(prange);
 
-	if (best_loc == KFD_IOCTL_SVM_LOCATION_UNDEFINED ||
-	    best_loc == prange->actual_loc)
+	/* when best_loc is a gpu node and same as prange->actual_loc
+	 * we still need do migration as prange->actual_loc !=0 does
+	 * not mean all pages in prange are vram. hmm migrate will pick
+	 * up right pages during migration.
+	 */
+	if ((best_loc == KFD_IOCTL_SVM_LOCATION_UNDEFINED) ||
+	    (best_loc == 0 && prange->actual_loc == 0))
 		return 0;
 
 	if (!best_loc) {
-		r = svm_migrate_vram_to_ram(prange, mm);
+		r = svm_migrate_vram_to_ram(prange, mm, prange->start, prange->last,
+					KFD_MIGRATE_TRIGGER_PREFETCH, NULL);
 		*migrated = !r;
 		return r;
 	}
 
-	r = svm_migrate_to_vram(prange, best_loc, mm);
+	r = svm_migrate_to_vram(prange, best_loc, prange->start, prange->last,
+				mm, KFD_MIGRATE_TRIGGER_PREFETCH);
 	*migrated = !r;
 
-	return r;
+	return 0;
 }
 
 int svm_range_schedule_evict_svm_bo(struct amdgpu_amdkfd_fence *fence)
 {
-	if (!fence)
-		return -EINVAL;
-
-	if (dma_fence_is_signaled(&fence->base))
-		return 0;
-
-	if (fence->svm_bo) {
+	/* Dereferencing fence->svm_bo is safe here because the fence hasn't
+	 * signaled yet and we're under the protection of the fence->lock.
+	 * After the fence is signaled in svm_range_bo_release, we cannot get
+	 * here any more.
+	 *
+	 * Reference is dropped in svm_range_evict_svm_bo_worker.
+	 */
+	if (svm_bo_ref_unless_zero(fence->svm_bo)) {
 		WRITE_ONCE(fence->svm_bo->evicting, 1);
 		schedule_work(&fence->svm_bo->eviction_work);
 	}
@@ -3118,24 +3576,21 @@ int svm_range_schedule_evict_svm_bo(struct amdgpu_amdkfd_fence *fence)
 static void svm_range_evict_svm_bo_worker(struct work_struct *work)
 {
 	struct svm_range_bo *svm_bo;
-	struct kfd_process *p;
 	struct mm_struct *mm;
+	int r = 0;
 
 	svm_bo = container_of(work, struct svm_range_bo, eviction_work);
-	if (!svm_bo_ref_unless_zero(svm_bo))
-		return; /* svm_bo was freed while eviction was pending */
 
-	/* svm_range_bo_release destroys this worker thread. So during
-	 * the lifetime of this thread, kfd_process and mm will be valid.
-	 */
-	p = container_of(svm_bo->svms, struct kfd_process, svms);
-	mm = p->mm;
-	if (!mm)
+	if (mmget_not_zero(svm_bo->eviction_fence->mm)) {
+		mm = svm_bo->eviction_fence->mm;
+	} else {
+		svm_range_bo_unref(svm_bo);
 		return;
+	}
 
 	mmap_read_lock(mm);
 	spin_lock(&svm_bo->list_lock);
-	while (!list_empty(&svm_bo->range_list)) {
+	while (!list_empty(&svm_bo->range_list) && !r) {
 		struct svm_range *prange =
 				list_first_entry(&svm_bo->range_list,
 						struct svm_range, svm_bo_list);
@@ -3149,45 +3604,58 @@ static void svm_range_evict_svm_bo_worker(struct work_struct *work)
 
 		mutex_lock(&prange->migrate_mutex);
 		do {
-			svm_migrate_vram_to_ram(prange,
-						svm_bo->eviction_fence->mm);
-		} while (prange->actual_loc && --retries);
-		WARN(prange->actual_loc, "Migration failed during eviction");
+			/* migrate all vram pages in this prange to sys ram
+			 * after that prange->actual_loc should be zero
+			 */
+			r = svm_migrate_vram_to_ram(prange, mm,
+					prange->start, prange->last,
+					KFD_MIGRATE_TRIGGER_TTM_EVICTION, NULL);
+		} while (!r && prange->actual_loc && --retries);
 
-		mutex_lock(&prange->lock);
-		prange->svm_bo = NULL;
-		mutex_unlock(&prange->lock);
+		if (!r && prange->actual_loc)
+			pr_info_once("Migration failed during eviction");
 
+		if (!prange->actual_loc) {
+			mutex_lock(&prange->lock);
+			prange->svm_bo = NULL;
+			mutex_unlock(&prange->lock);
+		}
 		mutex_unlock(&prange->migrate_mutex);
 
 		spin_lock(&svm_bo->list_lock);
 	}
 	spin_unlock(&svm_bo->list_lock);
 	mmap_read_unlock(mm);
+	mmput(mm);
 
 	dma_fence_signal(&svm_bo->eviction_fence->base);
+
 	/* This is the last reference to svm_bo, after svm_range_vram_node_free
 	 * has been called in svm_migrate_vram_to_ram
 	 */
-	WARN_ONCE(kref_read(&svm_bo->kref) != 1, "This was not the last reference\n");
+	WARN_ONCE(!r && kref_read(&svm_bo->kref) != 1, "This was not the last reference\n");
 	svm_range_bo_unref(svm_bo);
 }
 
 static int
-svm_range_set_attr(struct kfd_process *p, uint64_t start, uint64_t size,
-		   uint32_t nattr, struct kfd_ioctl_svm_attribute *attrs)
+svm_range_set_attr(struct kfd_process *p, struct mm_struct *mm,
+		   uint64_t start, uint64_t size, uint32_t nattr,
+		   struct kfd_ioctl_svm_attribute *attrs)
 {
-	struct mm_struct *mm = current->mm;
+	struct amdkfd_process_info *process_info = p->kgd_process_info;
 	struct list_head update_list;
 	struct list_head insert_list;
 	struct list_head remove_list;
+	struct list_head remap_list;
 	struct svm_range_list *svms;
 	struct svm_range *prange;
 	struct svm_range *next;
-	int r = 0;
+	bool update_mapping = false;
+	bool flush_tlb;
+	int r, ret = 0;
 
-	pr_debug("pasid 0x%x svms 0x%p [0x%llx 0x%llx] pages 0x%llx\n",
-		 p->pasid, &p->svms, start, start + size - 1, size);
+	pr_debug("process pid %d svms 0x%p [0x%llx 0x%llx] pages 0x%llx\n",
+		 p->lead_thread->pid, &p->svms, start, start + size - 1, size);
 
 	r = svm_range_check_attr(p, nattr, attrs);
 	if (r)
@@ -3195,6 +3663,8 @@ svm_range_set_attr(struct kfd_process *p, uint64_t start, uint64_t size,
 
 	svms = &p->svms;
 
+	mutex_lock(&process_info->lock);
+
 	svm_range_list_lock_and_flush_work(svms, mm);
 
 	r = svm_range_is_valid(p, start, size);
@@ -3208,29 +3678,30 @@ svm_range_set_attr(struct kfd_process *p, uint64_t start, uint64_t size,
 
 	/* Add new range and split existing ranges as needed */
 	r = svm_range_add(p, start, size, nattr, attrs, &update_list,
-			  &insert_list, &remove_list);
+			  &insert_list, &remove_list, &remap_list);
 	if (r) {
 		mutex_unlock(&svms->lock);
 		mmap_write_unlock(mm);
 		goto out;
 	}
 	/* Apply changes as a transaction */
-	list_for_each_entry_safe(prange, next, &insert_list, insert_list) {
+	list_for_each_entry_safe(prange, next, &insert_list, list) {
 		svm_range_add_to_svms(prange);
 		svm_range_add_notifier_locked(mm, prange);
 	}
 	list_for_each_entry(prange, &update_list, update_list) {
-		svm_range_apply_attrs(p, prange, nattr, attrs);
+		svm_range_apply_attrs(p, prange, nattr, attrs, &update_mapping);
 		/* TODO: unmap ranges from GPU that lost access */
 	}
-	list_for_each_entry_safe(prange, next, &remove_list,
-				remove_list) {
+	update_mapping |= !p->xnack_enabled && !list_empty(&remap_list);
+
+	list_for_each_entry_safe(prange, next, &remove_list, update_list) {
 		pr_debug("unlink old 0x%p prange 0x%p [0x%lx 0x%lx]\n",
 			 prange->svms, prange, prange->start,
 			 prange->last);
 		svm_range_unlink(prange);
 		svm_range_remove_notifier(prange);
-		svm_range_free(prange);
+		svm_range_free(prange, false);
 	}
 
 	mmap_write_downgrade(mm);
@@ -3248,37 +3719,62 @@ svm_range_set_attr(struct kfd_process *p, uint64_t start, uint64_t size,
 		if (r)
 			goto out_unlock_range;
 
-		if (migrated && !p->xnack_enabled) {
+		if (migrated && (!p->xnack_enabled ||
+		    (prange->flags & KFD_IOCTL_SVM_FLAG_GPU_ALWAYS_MAPPED)) &&
+		    prange->mapped_to_gpu) {
 			pr_debug("restore_work will update mappings of GPUs\n");
 			mutex_unlock(&prange->migrate_mutex);
 			continue;
 		}
 
-		r = svm_range_validate_and_map(mm, prange, MAX_GPU_INSTANCE,
-					       true, true);
+		if (!migrated && !update_mapping) {
+			mutex_unlock(&prange->migrate_mutex);
+			continue;
+		}
+
+		flush_tlb = !migrated && update_mapping && prange->mapped_to_gpu;
+
+		r = svm_range_validate_and_map(mm, prange->start, prange->last, prange,
+					       MAX_GPU_INSTANCE, true, true, flush_tlb);
 		if (r)
 			pr_debug("failed %d to map svm range\n", r);
 
 out_unlock_range:
 		mutex_unlock(&prange->migrate_mutex);
 		if (r)
-			break;
+			ret = r;
 	}
 
-	svm_range_debug_dump(svms);
+	list_for_each_entry(prange, &remap_list, update_list) {
+		pr_debug("Remapping prange 0x%p [0x%lx 0x%lx]\n",
+			 prange, prange->start, prange->last);
+		mutex_lock(&prange->migrate_mutex);
+		r = svm_range_validate_and_map(mm,  prange->start, prange->last, prange,
+					       MAX_GPU_INSTANCE, true, true, prange->mapped_to_gpu);
+		if (r)
+			pr_debug("failed %d on remap svm range\n", r);
+		mutex_unlock(&prange->migrate_mutex);
+		if (r)
+			ret = r;
+	}
+
+	dynamic_svm_range_dump(svms);
 
 	mutex_unlock(&svms->lock);
 	mmap_read_unlock(mm);
 out:
-	pr_debug("pasid 0x%x svms 0x%p [0x%llx 0x%llx] done, r=%d\n", p->pasid,
-		 &p->svms, start, start + size - 1, r);
+	mutex_unlock(&process_info->lock);
 
-	return r;
+	pr_debug("process pid %d svms 0x%p [0x%llx 0x%llx] done, r=%d\n",
+		 p->lead_thread->pid, &p->svms, start, start + size - 1, r);
+
+	return ret ? ret : r;
 }
 
 static int
-svm_range_get_attr(struct kfd_process *p, uint64_t start, uint64_t size,
-		   uint32_t nattr, struct kfd_ioctl_svm_attribute *attrs)
+svm_range_get_attr(struct kfd_process *p, struct mm_struct *mm,
+		   uint64_t start, uint64_t size, uint32_t nattr,
+		   struct kfd_ioctl_svm_attribute *attrs)
 {
 	DECLARE_BITMAP(bitmap_access, MAX_GPU_INSTANCE);
 	DECLARE_BITMAP(bitmap_aip, MAX_GPU_INSTANCE);
@@ -3288,7 +3784,6 @@ svm_range_get_attr(struct kfd_process *p, uint64_t start, uint64_t size,
 	bool get_accessible = false;
 	bool get_flags = false;
 	uint64_t last = start + size - 1UL;
-	struct mm_struct *mm = current->mm;
 	uint8_t granularity = 0xff;
 	struct interval_tree_node *node;
 	struct svm_range_list *svms;
@@ -3354,7 +3849,7 @@ svm_range_get_attr(struct kfd_process *p, uint64_t start, uint64_t size,
 	node = interval_tree_iter_first(&svms->objects, start, last);
 	if (!node) {
 		pr_debug("range attrs not found return default values\n");
-		svm_range_set_default_attributes(&location, &prefetch_loc,
+		svm_range_set_default_attributes(svms, &location, &prefetch_loc,
 						 &granularity, &flags_and);
 		flags_or = flags_and;
 		if (p->xnack_enabled)
@@ -3453,10 +3948,316 @@ fill_values:
 	return 0;
 }
 
+int kfd_criu_resume_svm(struct kfd_process *p)
+{
+	struct kfd_ioctl_svm_attribute *set_attr_new, *set_attr = NULL;
+	int nattr_common = 4, nattr_accessibility = 1;
+	struct criu_svm_metadata *criu_svm_md = NULL;
+	struct svm_range_list *svms = &p->svms;
+	struct criu_svm_metadata *next = NULL;
+	uint32_t set_flags = 0xffffffff;
+	int i, j, num_attrs, ret = 0;
+	uint64_t set_attr_size;
+	struct mm_struct *mm;
+
+	if (list_empty(&svms->criu_svm_metadata_list)) {
+		pr_debug("No SVM data from CRIU restore stage 2\n");
+		return ret;
+	}
+
+	mm = get_task_mm(p->lead_thread);
+	if (!mm) {
+		pr_err("failed to get mm for the target process\n");
+		return -ESRCH;
+	}
+
+	num_attrs = nattr_common + (nattr_accessibility * p->n_pdds);
+
+	i = j = 0;
+	list_for_each_entry(criu_svm_md, &svms->criu_svm_metadata_list, list) {
+		pr_debug("criu_svm_md[%d]\n\tstart: 0x%llx size: 0x%llx (npages)\n",
+			 i, criu_svm_md->data.start_addr, criu_svm_md->data.size);
+
+		for (j = 0; j < num_attrs; j++) {
+			pr_debug("\ncriu_svm_md[%d]->attrs[%d].type : 0x%x\ncriu_svm_md[%d]->attrs[%d].value : 0x%x\n",
+				 i, j, criu_svm_md->data.attrs[j].type,
+				 i, j, criu_svm_md->data.attrs[j].value);
+			switch (criu_svm_md->data.attrs[j].type) {
+			/* During Checkpoint operation, the query for
+			 * KFD_IOCTL_SVM_ATTR_PREFETCH_LOC attribute might
+			 * return KFD_IOCTL_SVM_LOCATION_UNDEFINED if they were
+			 * not used by the range which was checkpointed. Care
+			 * must be taken to not restore with an invalid value
+			 * otherwise the gpuidx value will be invalid and
+			 * set_attr would eventually fail so just replace those
+			 * with another dummy attribute such as
+			 * KFD_IOCTL_SVM_ATTR_SET_FLAGS.
+			 */
+			case KFD_IOCTL_SVM_ATTR_PREFETCH_LOC:
+				if (criu_svm_md->data.attrs[j].value ==
+				    KFD_IOCTL_SVM_LOCATION_UNDEFINED) {
+					criu_svm_md->data.attrs[j].type =
+						KFD_IOCTL_SVM_ATTR_SET_FLAGS;
+					criu_svm_md->data.attrs[j].value = 0;
+				}
+				break;
+			case KFD_IOCTL_SVM_ATTR_SET_FLAGS:
+				set_flags = criu_svm_md->data.attrs[j].value;
+				break;
+			default:
+				break;
+			}
+		}
+
+		/* CLR_FLAGS is not available via get_attr during checkpoint but
+		 * it needs to be inserted before restoring the ranges so
+		 * allocate extra space for it before calling set_attr
+		 */
+		set_attr_size = sizeof(struct kfd_ioctl_svm_attribute) *
+						(num_attrs + 1);
+		set_attr_new = krealloc(set_attr, set_attr_size,
+					    GFP_KERNEL);
+		if (!set_attr_new) {
+			ret = -ENOMEM;
+			goto exit;
+		}
+		set_attr = set_attr_new;
+
+		memcpy(set_attr, criu_svm_md->data.attrs, num_attrs *
+					sizeof(struct kfd_ioctl_svm_attribute));
+		set_attr[num_attrs].type = KFD_IOCTL_SVM_ATTR_CLR_FLAGS;
+		set_attr[num_attrs].value = ~set_flags;
+
+		ret = svm_range_set_attr(p, mm, criu_svm_md->data.start_addr,
+					 criu_svm_md->data.size, num_attrs + 1,
+					 set_attr);
+		if (ret) {
+			pr_err("CRIU: failed to set range attributes\n");
+			goto exit;
+		}
+
+		i++;
+	}
+exit:
+	kfree(set_attr);
+	list_for_each_entry_safe(criu_svm_md, next, &svms->criu_svm_metadata_list, list) {
+		pr_debug("freeing criu_svm_md[]\n\tstart: 0x%llx\n",
+						criu_svm_md->data.start_addr);
+		kfree(criu_svm_md);
+	}
+
+	mmput(mm);
+	return ret;
+
+}
+
+int kfd_criu_restore_svm(struct kfd_process *p,
+			 uint8_t __user *user_priv_ptr,
+			 uint64_t *priv_data_offset,
+			 uint64_t max_priv_data_size)
+{
+	uint64_t svm_priv_data_size, svm_object_md_size, svm_attrs_size;
+	int nattr_common = 4, nattr_accessibility = 1;
+	struct criu_svm_metadata *criu_svm_md = NULL;
+	struct svm_range_list *svms = &p->svms;
+	uint32_t num_devices;
+	int ret = 0;
+
+	num_devices = p->n_pdds;
+	/* Handle one SVM range object at a time, also the number of gpus are
+	 * assumed to be same on the restore node, checking must be done while
+	 * evaluating the topology earlier
+	 */
+
+	svm_attrs_size = sizeof(struct kfd_ioctl_svm_attribute) *
+		(nattr_common + nattr_accessibility * num_devices);
+	svm_object_md_size = sizeof(struct criu_svm_metadata) + svm_attrs_size;
+
+	svm_priv_data_size = sizeof(struct kfd_criu_svm_range_priv_data) +
+								svm_attrs_size;
+
+	criu_svm_md = kzalloc(svm_object_md_size, GFP_KERNEL);
+	if (!criu_svm_md) {
+		pr_err("failed to allocate memory to store svm metadata\n");
+		return -ENOMEM;
+	}
+	if (*priv_data_offset + svm_priv_data_size > max_priv_data_size) {
+		ret = -EINVAL;
+		goto exit;
+	}
+
+	ret = copy_from_user(&criu_svm_md->data, user_priv_ptr + *priv_data_offset,
+			     svm_priv_data_size);
+	if (ret) {
+		ret = -EFAULT;
+		goto exit;
+	}
+	*priv_data_offset += svm_priv_data_size;
+
+	list_add_tail(&criu_svm_md->list, &svms->criu_svm_metadata_list);
+
+	return 0;
+
+
+exit:
+	kfree(criu_svm_md);
+	return ret;
+}
+
+void svm_range_get_info(struct kfd_process *p, uint32_t *num_svm_ranges,
+			uint64_t *svm_priv_data_size)
+{
+	uint64_t total_size, accessibility_size, common_attr_size;
+	int nattr_common = 4, nattr_accessibility = 1;
+	int num_devices = p->n_pdds;
+	struct svm_range_list *svms;
+	struct svm_range *prange;
+	uint32_t count = 0;
+
+	*svm_priv_data_size = 0;
+
+	svms = &p->svms;
+
+	mutex_lock(&svms->lock);
+	list_for_each_entry(prange, &svms->list, list) {
+		pr_debug("prange: 0x%p start: 0x%lx\t npages: 0x%llx\t end: 0x%llx\n",
+			 prange, prange->start, prange->npages,
+			 prange->start + prange->npages - 1);
+		count++;
+	}
+	mutex_unlock(&svms->lock);
+
+	*num_svm_ranges = count;
+	/* Only the accessbility attributes need to be queried for all the gpus
+	 * individually, remaining ones are spanned across the entire process
+	 * regardless of the various gpu nodes. Of the remaining attributes,
+	 * KFD_IOCTL_SVM_ATTR_CLR_FLAGS need not be saved.
+	 *
+	 * KFD_IOCTL_SVM_ATTR_PREFERRED_LOC
+	 * KFD_IOCTL_SVM_ATTR_PREFETCH_LOC
+	 * KFD_IOCTL_SVM_ATTR_SET_FLAGS
+	 * KFD_IOCTL_SVM_ATTR_GRANULARITY
+	 *
+	 * ** ACCESSBILITY ATTRIBUTES **
+	 * (Considered as one, type is altered during query, value is gpuid)
+	 * KFD_IOCTL_SVM_ATTR_ACCESS
+	 * KFD_IOCTL_SVM_ATTR_ACCESS_IN_PLACE
+	 * KFD_IOCTL_SVM_ATTR_NO_ACCESS
+	 */
+	if (*num_svm_ranges > 0) {
+		common_attr_size = sizeof(struct kfd_ioctl_svm_attribute) *
+			nattr_common;
+		accessibility_size = sizeof(struct kfd_ioctl_svm_attribute) *
+			nattr_accessibility * num_devices;
+
+		total_size = sizeof(struct kfd_criu_svm_range_priv_data) +
+			common_attr_size + accessibility_size;
+
+		*svm_priv_data_size = *num_svm_ranges * total_size;
+	}
+
+	pr_debug("num_svm_ranges %u total_priv_size %llu\n", *num_svm_ranges,
+		 *svm_priv_data_size);
+}
+
+int kfd_criu_checkpoint_svm(struct kfd_process *p,
+			    uint8_t __user *user_priv_data,
+			    uint64_t *priv_data_offset)
+{
+	struct kfd_criu_svm_range_priv_data *svm_priv = NULL;
+	struct kfd_ioctl_svm_attribute *query_attr = NULL;
+	uint64_t svm_priv_data_size, query_attr_size = 0;
+	int index, nattr_common = 4, ret = 0;
+	struct svm_range_list *svms;
+	int num_devices = p->n_pdds;
+	struct svm_range *prange;
+	struct mm_struct *mm;
+
+	svms = &p->svms;
+
+	mm = get_task_mm(p->lead_thread);
+	if (!mm) {
+		pr_err("failed to get mm for the target process\n");
+		return -ESRCH;
+	}
+
+	query_attr_size = sizeof(struct kfd_ioctl_svm_attribute) *
+				(nattr_common + num_devices);
+
+	query_attr = kzalloc(query_attr_size, GFP_KERNEL);
+	if (!query_attr) {
+		ret = -ENOMEM;
+		goto exit;
+	}
+
+	query_attr[0].type = KFD_IOCTL_SVM_ATTR_PREFERRED_LOC;
+	query_attr[1].type = KFD_IOCTL_SVM_ATTR_PREFETCH_LOC;
+	query_attr[2].type = KFD_IOCTL_SVM_ATTR_SET_FLAGS;
+	query_attr[3].type = KFD_IOCTL_SVM_ATTR_GRANULARITY;
+
+	for (index = 0; index < num_devices; index++) {
+		struct kfd_process_device *pdd = p->pdds[index];
+
+		query_attr[index + nattr_common].type =
+			KFD_IOCTL_SVM_ATTR_ACCESS;
+		query_attr[index + nattr_common].value = pdd->user_gpu_id;
+	}
+
+	svm_priv_data_size = sizeof(*svm_priv) + query_attr_size;
+
+	svm_priv = kzalloc(svm_priv_data_size, GFP_KERNEL);
+	if (!svm_priv) {
+		ret = -ENOMEM;
+		goto exit_query;
+	}
+
+	index = 0;
+	list_for_each_entry(prange, &svms->list, list) {
+
+		svm_priv->object_type = KFD_CRIU_OBJECT_TYPE_SVM_RANGE;
+		svm_priv->start_addr = prange->start;
+		svm_priv->size = prange->npages;
+		memcpy(&svm_priv->attrs, query_attr, query_attr_size);
+		pr_debug("CRIU: prange: 0x%p start: 0x%lx\t npages: 0x%llx end: 0x%llx\t size: 0x%llx\n",
+			 prange, prange->start, prange->npages,
+			 prange->start + prange->npages - 1,
+			 prange->npages * PAGE_SIZE);
+
+		ret = svm_range_get_attr(p, mm, svm_priv->start_addr,
+					 svm_priv->size,
+					 (nattr_common + num_devices),
+					 svm_priv->attrs);
+		if (ret) {
+			pr_err("CRIU: failed to obtain range attributes\n");
+			goto exit_priv;
+		}
+
+		if (copy_to_user(user_priv_data + *priv_data_offset, svm_priv,
+				 svm_priv_data_size)) {
+			pr_err("Failed to copy svm priv to user\n");
+			ret = -EFAULT;
+			goto exit_priv;
+		}
+
+		*priv_data_offset += svm_priv_data_size;
+
+	}
+
+
+exit_priv:
+	kfree(svm_priv);
+exit_query:
+	kfree(query_attr);
+exit:
+	mmput(mm);
+	return ret;
+}
+
 int
 svm_ioctl(struct kfd_process *p, enum kfd_ioctl_svm_op op, uint64_t start,
 	  uint64_t size, uint32_t nattrs, struct kfd_ioctl_svm_attribute *attrs)
 {
+	struct mm_struct *mm = current->mm;
 	int r;
 
 	start >>= PAGE_SHIFT;
@@ -3464,13 +4265,13 @@ svm_ioctl(struct kfd_process *p, enum kfd_ioctl_svm_op op, uint64_t start,
 
 	switch (op) {
 	case KFD_IOCTL_SVM_OP_SET_ATTR:
-		r = svm_range_set_attr(p, start, size, nattrs, attrs);
+		r = svm_range_set_attr(p, mm, start, size, nattrs, attrs);
 		break;
 	case KFD_IOCTL_SVM_OP_GET_ATTR:
-		r = svm_range_get_attr(p, start, size, nattrs, attrs);
+		r = svm_range_get_attr(p, mm, start, size, nattrs, attrs);
 		break;
 	default:
-		r = EINVAL;
+		r = -EINVAL;
 		break;
 	}
 
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_svm.h b/drivers/gpu/drm/amd/amdkfd/kfd_svm.h
index 6dc91c33e80f..a63dfc95b602 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_svm.h
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_svm.h
@@ -31,7 +31,6 @@
 #include <linux/list.h>
 #include <linux/mutex.h>
 #include <linux/sched/mm.h>
-#include <linux/hmm.h>
 #include "amdgpu.h"
 #include "kfd_priv.h"
 
@@ -46,8 +45,9 @@ struct svm_range_bo {
 	spinlock_t			list_lock;
 	struct amdgpu_amdkfd_fence	*eviction_fence;
 	struct work_struct		eviction_work;
-	struct svm_range_list		*svms;
 	uint32_t			evicting;
+	struct work_struct		release_work;
+	struct kfd_node			*node;
 };
 
 enum svm_work_list_ops {
@@ -75,10 +75,9 @@ struct svm_work_list_item {
  *              aligned, page size is (last - start + 1)
  * @list:       link list node, used to scan all ranges of svms
  * @update_list:link list node used to add to update_list
- * @remove_list:link list node used to add to remove list
- * @insert_list:link list node used to add to insert list
  * @mapping:    bo_va mapping structure to create and update GPU page table
  * @npages:     number of pages
+ * @vram_pages: vram pages number in this svm_range
  * @dma_addr:   dma mapping address on each GPU for system memory physical page
  * @ttm_res:    vram ttm resource map
  * @offset:     range start offset within mm_nodes
@@ -89,7 +88,9 @@ struct svm_work_list_item {
  * @flags:      flags defined as KFD_IOCTL_SVM_FLAG_*
  * @perferred_loc: perferred location, 0 for CPU, or GPU id
  * @perfetch_loc: last prefetch location, 0 for CPU, or GPU id
- * @actual_loc: the actual location, 0 for CPU, or GPU id
+ * @actual_loc: this svm_range location. 0: all pages are from sys ram;
+ *              GPU id: this svm_range may include vram pages from GPU with
+ *              id actual_loc.
  * @granularity:migration granularity, log2 num pages
  * @invalid:    not 0 means cpu page table is invalidated
  * @validate_timestamp: system timestamp when range is validated
@@ -112,9 +113,8 @@ struct svm_range {
 	struct interval_tree_node	it_node;
 	struct list_head		list;
 	struct list_head		update_list;
-	struct list_head		remove_list;
-	struct list_head		insert_list;
 	uint64_t			npages;
+	uint64_t			vram_pages;
 	dma_addr_t			*dma_addr[MAX_GPU_INSTANCE];
 	struct ttm_resource		*ttm_res;
 	uint64_t			offset;
@@ -128,14 +128,15 @@ struct svm_range {
 	uint32_t			actual_loc;
 	uint8_t				granularity;
 	atomic_t			invalid;
-	uint64_t			validate_timestamp;
+	ktime_t				validate_timestamp;
 	struct mmu_interval_notifier	notifier;
 	struct svm_work_list_item	work_item;
 	struct list_head		deferred_list;
 	struct list_head		child_list;
 	DECLARE_BITMAP(bitmap_access, MAX_GPU_INSTANCE);
 	DECLARE_BITMAP(bitmap_aip, MAX_GPU_INSTANCE);
-	bool				validated_once;
+	bool				mapped_to_gpu;
+	atomic_t			queue_refcount;
 };
 
 static inline void svm_range_lock(struct svm_range *prange)
@@ -166,36 +167,47 @@ int svm_ioctl(struct kfd_process *p, enum kfd_ioctl_svm_op op, uint64_t start,
 struct svm_range *svm_range_from_addr(struct svm_range_list *svms,
 				      unsigned long addr,
 				      struct svm_range **parent);
-struct amdgpu_device *svm_range_get_adev_by_id(struct svm_range *prange,
-					       uint32_t id);
-int svm_range_vram_node_new(struct amdgpu_device *adev,
-			    struct svm_range *prange, bool clear);
+struct kfd_node *svm_range_get_node_by_id(struct svm_range *prange,
+					  uint32_t gpu_id);
+int svm_range_vram_node_new(struct kfd_node *node, struct svm_range *prange,
+			    bool clear);
 void svm_range_vram_node_free(struct svm_range *prange);
-int svm_range_split_by_granularity(struct kfd_process *p, struct mm_struct *mm,
-			       unsigned long addr, struct svm_range *parent,
-			       struct svm_range *prange);
-int svm_range_restore_pages(struct amdgpu_device *adev,
-			    unsigned int pasid, uint64_t addr, bool write_fault);
+int svm_range_restore_pages(struct amdgpu_device *adev, unsigned int pasid,
+			    uint32_t vmid, uint32_t node_id, uint64_t addr, uint64_t ts,
+			    bool write_fault);
 int svm_range_schedule_evict_svm_bo(struct amdgpu_amdkfd_fence *fence);
 void svm_range_add_list_work(struct svm_range_list *svms,
 			     struct svm_range *prange, struct mm_struct *mm,
 			     enum svm_work_list_ops op);
 void schedule_deferred_list_work(struct svm_range_list *svms);
-void svm_range_dma_unmap(struct device *dev, dma_addr_t *dma_addr,
+void svm_range_dma_unmap_dev(struct device *dev, dma_addr_t *dma_addr,
 			 unsigned long offset, unsigned long npages);
-void svm_range_free_dma_mappings(struct svm_range *prange);
-void svm_range_prefault(struct svm_range *prange, struct mm_struct *mm,
-			void *owner);
+void svm_range_dma_unmap(struct svm_range *prange);
+void svm_range_get_info(struct kfd_process *p, uint32_t *num_svm_ranges,
+			uint64_t *svm_priv_data_size);
+int kfd_criu_checkpoint_svm(struct kfd_process *p,
+			    uint8_t __user *user_priv_data,
+			    uint64_t *priv_offset);
+int kfd_criu_restore_svm(struct kfd_process *p,
+			 uint8_t __user *user_priv_ptr,
+			 uint64_t *priv_data_offset,
+			 uint64_t max_priv_data_size);
+int kfd_criu_resume_svm(struct kfd_process *p);
 struct kfd_process_device *
-svm_range_get_pdd_by_adev(struct svm_range *prange, struct amdgpu_device *adev);
+svm_range_get_pdd_by_node(struct svm_range *prange, struct kfd_node *node);
 void svm_range_list_lock_and_flush_work(struct svm_range_list *svms, struct mm_struct *mm);
 
 /* SVM API and HMM page migration work together, device memory type
  * is initialized to not 0 when page migration register device memory.
  */
-#define KFD_IS_SVM_API_SUPPORTED(dev) ((dev)->pgmap.type != 0)
+#define KFD_IS_SVM_API_SUPPORTED(adev) ((adev)->kfd.pgmap.type != 0 ||\
+					((adev)->apu_prefer_gtt))
+
+void svm_range_bo_unref_async(struct svm_range_bo *svm_bo);
+
+void svm_range_set_max_pages(struct amdgpu_device *adev);
+int svm_range_switch_xnack_reserve_mem(struct kfd_process *p, bool xnack_enabled);
 
-void svm_range_bo_unref(struct svm_range_bo *svm_bo);
 #else
 
 struct kfd_process;
@@ -210,8 +222,9 @@ static inline void svm_range_list_fini(struct kfd_process *p)
 }
 
 static inline int svm_range_restore_pages(struct amdgpu_device *adev,
-					  unsigned int pasid, uint64_t addr,
-					  bool write_fault)
+					  unsigned int pasid,
+					  uint32_t client_id, uint32_t node_id,
+					  uint64_t addr, uint64_t ts, bool write_fault)
 {
 	return -EFAULT;
 }
@@ -223,6 +236,38 @@ static inline int svm_range_schedule_evict_svm_bo(
 	return -EINVAL;
 }
 
+static inline void svm_range_get_info(struct kfd_process *p,
+				      uint32_t *num_svm_ranges,
+				      uint64_t *svm_priv_data_size)
+{
+	*num_svm_ranges = 0;
+	*svm_priv_data_size = 0;
+}
+
+static inline int kfd_criu_checkpoint_svm(struct kfd_process *p,
+					  uint8_t __user *user_priv_data,
+					  uint64_t *priv_offset)
+{
+	return 0;
+}
+
+static inline int kfd_criu_restore_svm(struct kfd_process *p,
+				       uint8_t __user *user_priv_ptr,
+				       uint64_t *priv_data_offset,
+				       uint64_t max_priv_data_size)
+{
+	return -EINVAL;
+}
+
+static inline int kfd_criu_resume_svm(struct kfd_process *p)
+{
+	return 0;
+}
+
+static inline void svm_range_set_max_pages(struct amdgpu_device *adev)
+{
+}
+
 #define KFD_IS_SVM_API_SUPPORTED(dev) false
 
 #endif /* IS_ENABLED(CONFIG_HSA_AMD_SVM) */
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_topology.c b/drivers/gpu/drm/amd/amdkfd/kfd_topology.c
index dd593ad0614a..811636af14ea 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_topology.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_topology.c
@@ -1,5 +1,6 @@
+// SPDX-License-Identifier: GPL-2.0 OR MIT
 /*
- * Copyright 2014 Advanced Micro Devices, Inc.
+ * Copyright 2014-2022 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -30,37 +31,49 @@
 #include <linux/log2.h>
 #include <linux/dmi.h>
 #include <linux/atomic.h>
+#include <linux/crc16.h>
 
 #include "kfd_priv.h"
 #include "kfd_crat.h"
 #include "kfd_topology.h"
 #include "kfd_device_queue_manager.h"
-#include "kfd_iommu.h"
 #include "kfd_svm.h"
+#include "kfd_debug.h"
 #include "amdgpu_amdkfd.h"
 #include "amdgpu_ras.h"
+#include "amdgpu.h"
 
 /* topology_device_list - Master list of all topology devices */
 static struct list_head topology_device_list;
 static struct kfd_system_properties sys_props;
 
 static DECLARE_RWSEM(topology_lock);
-static atomic_t topology_crat_proximity_domain;
+static uint32_t topology_crat_proximity_domain;
 
-struct kfd_topology_device *kfd_topology_device_by_proximity_domain(
+struct kfd_topology_device *kfd_topology_device_by_proximity_domain_no_lock(
 						uint32_t proximity_domain)
 {
 	struct kfd_topology_device *top_dev;
 	struct kfd_topology_device *device = NULL;
 
-	down_read(&topology_lock);
-
 	list_for_each_entry(top_dev, &topology_device_list, list)
 		if (top_dev->proximity_domain == proximity_domain) {
 			device = top_dev;
 			break;
 		}
 
+	return device;
+}
+
+struct kfd_topology_device *kfd_topology_device_by_proximity_domain(
+						uint32_t proximity_domain)
+{
+	struct kfd_topology_device *device = NULL;
+
+	down_read(&topology_lock);
+
+	device = kfd_topology_device_by_proximity_domain_no_lock(
+							proximity_domain);
 	up_read(&topology_lock);
 
 	return device;
@@ -84,7 +97,7 @@ struct kfd_topology_device *kfd_topology_device_by_id(uint32_t gpu_id)
 	return ret;
 }
 
-struct kfd_dev *kfd_device_by_id(uint32_t gpu_id)
+struct kfd_node *kfd_device_by_id(uint32_t gpu_id)
 {
 	struct kfd_topology_device *top_dev;
 
@@ -95,48 +108,13 @@ struct kfd_dev *kfd_device_by_id(uint32_t gpu_id)
 	return top_dev->gpu;
 }
 
-struct kfd_dev *kfd_device_by_pci_dev(const struct pci_dev *pdev)
-{
-	struct kfd_topology_device *top_dev;
-	struct kfd_dev *device = NULL;
-
-	down_read(&topology_lock);
-
-	list_for_each_entry(top_dev, &topology_device_list, list)
-		if (top_dev->gpu && top_dev->gpu->pdev == pdev) {
-			device = top_dev->gpu;
-			break;
-		}
-
-	up_read(&topology_lock);
-
-	return device;
-}
-
-struct kfd_dev *kfd_device_by_kgd(const struct kgd_dev *kgd)
-{
-	struct kfd_topology_device *top_dev;
-	struct kfd_dev *device = NULL;
-
-	down_read(&topology_lock);
-
-	list_for_each_entry(top_dev, &topology_device_list, list)
-		if (top_dev->gpu && top_dev->gpu->kgd == kgd) {
-			device = top_dev->gpu;
-			break;
-		}
-
-	up_read(&topology_lock);
-
-	return device;
-}
-
 /* Called with write topology_lock acquired */
 static void kfd_release_topology_device(struct kfd_topology_device *dev)
 {
 	struct kfd_mem_properties *mem;
 	struct kfd_cache_properties *cache;
 	struct kfd_iolink_properties *iolink;
+	struct kfd_iolink_properties *p2plink;
 	struct kfd_perf_properties *perf;
 
 	list_del(&dev->list);
@@ -162,6 +140,13 @@ static void kfd_release_topology_device(struct kfd_topology_device *dev)
 		kfree(iolink);
 	}
 
+	while (dev->p2p_link_props.next != &dev->p2p_link_props) {
+		p2plink = container_of(dev->p2p_link_props.next,
+				struct kfd_iolink_properties, list);
+		list_del(&p2plink->list);
+		kfree(p2plink);
+	}
+
 	while (dev->perf_props.next != &dev->perf_props) {
 		perf = container_of(dev->perf_props.next,
 				struct kfd_perf_properties, list);
@@ -203,6 +188,7 @@ struct kfd_topology_device *kfd_create_topology_device(
 	INIT_LIST_HEAD(&dev->mem_props);
 	INIT_LIST_HEAD(&dev->cache_props);
 	INIT_LIST_HEAD(&dev->io_link_props);
+	INIT_LIST_HEAD(&dev->p2p_link_props);
 	INIT_LIST_HEAD(&dev->perf_props);
 
 	list_add_tail(&dev->list, device_list);
@@ -257,7 +243,7 @@ static const struct sysfs_ops sysprops_ops = {
 	.show = sysprops_show,
 };
 
-static struct kobj_type sysprops_type = {
+static const struct kobj_type sysprops_type = {
 	.release = kfd_topology_kobj_release,
 	.sysfs_ops = &sysprops_ops,
 };
@@ -288,6 +274,8 @@ static ssize_t iolink_show(struct kobject *kobj, struct attribute *attr,
 			      iolink->max_bandwidth);
 	sysfs_show_32bit_prop(buffer, offs, "recommended_transfer_size",
 			      iolink->rec_transfer_size);
+	sysfs_show_32bit_prop(buffer, offs, "recommended_sdma_engine_id_mask",
+			      iolink->rec_sdma_eng_id_mask);
 	sysfs_show_32bit_prop(buffer, offs, "flags", iolink->flags);
 
 	return offs;
@@ -297,7 +285,7 @@ static const struct sysfs_ops iolink_ops = {
 	.show = iolink_show,
 };
 
-static struct kobj_type iolink_type = {
+static const struct kobj_type iolink_type = {
 	.release = kfd_topology_kobj_release,
 	.sysfs_ops = &iolink_ops,
 };
@@ -329,7 +317,7 @@ static const struct sysfs_ops mem_ops = {
 	.show = mem_show,
 };
 
-static struct kobj_type mem_type = {
+static const struct kobj_type mem_type = {
 	.release = kfd_topology_kobj_release,
 	.sysfs_ops = &mem_ops,
 };
@@ -343,7 +331,6 @@ static ssize_t kfd_cache_show(struct kobject *kobj, struct attribute *attr,
 
 	/* Making sure that the buffer is an empty string */
 	buffer[0] = 0;
-
 	cache = container_of(attr, struct kfd_cache_properties, attr);
 	if (cache->gpu && kfd_devcgroup_check_permission(cache->gpu))
 		return -EPERM;
@@ -358,12 +345,13 @@ static ssize_t kfd_cache_show(struct kobject *kobj, struct attribute *attr,
 	sysfs_show_32bit_prop(buffer, offs, "association", cache->cache_assoc);
 	sysfs_show_32bit_prop(buffer, offs, "latency", cache->cache_latency);
 	sysfs_show_32bit_prop(buffer, offs, "type", cache->cache_type);
+
 	offs += snprintf(buffer+offs, PAGE_SIZE-offs, "sibling_map ");
-	for (i = 0; i < CRAT_SIBLINGMAP_SIZE; i++)
+	for (i = 0; i < cache->sibling_map_size; i++)
 		for (j = 0; j < sizeof(cache->sibling_map[0])*8; j++)
 			/* Check each bit */
 			offs += snprintf(buffer+offs, PAGE_SIZE-offs, "%d,",
-					 (cache->sibling_map[i] >> j) & 1);
+						(cache->sibling_map[i] >> j) & 1);
 
 	/* Replace the last "," with end of line */
 	buffer[offs-1] = '\n';
@@ -374,7 +362,7 @@ static const struct sysfs_ops cache_ops = {
 	.show = kfd_cache_show,
 };
 
-static struct kobj_type cache_type = {
+static const struct kobj_type cache_type = {
 	.release = kfd_topology_kobj_release,
 	.sysfs_ops = &cache_ops,
 };
@@ -454,6 +442,8 @@ static ssize_t node_show(struct kobject *kobj, struct attribute *attr,
 			      dev->node_props.caches_count);
 	sysfs_show_32bit_prop(buffer, offs, "io_links_count",
 			      dev->node_props.io_links_count);
+	sysfs_show_32bit_prop(buffer, offs, "p2p_links_count",
+			      dev->node_props.p2p_links_count);
 	sysfs_show_32bit_prop(buffer, offs, "cpu_core_id_base",
 			      dev->node_props.cpu_core_id_base);
 	sysfs_show_32bit_prop(buffer, offs, "simd_id_base",
@@ -469,7 +459,8 @@ static ssize_t node_show(struct kobject *kobj, struct attribute *attr,
 	sysfs_show_32bit_prop(buffer, offs, "wave_front_size",
 			      dev->node_props.wave_front_size);
 	sysfs_show_32bit_prop(buffer, offs, "array_count",
-			      dev->node_props.array_count);
+			      dev->gpu ? (dev->node_props.array_count *
+					  NUM_XCC(dev->gpu->xcc_mask)) : 0);
 	sysfs_show_32bit_prop(buffer, offs, "simd_arrays_per_engine",
 			      dev->node_props.simd_arrays_per_engine);
 	sysfs_show_32bit_prop(buffer, offs, "cu_per_simd_array",
@@ -503,7 +494,7 @@ static ssize_t node_show(struct kobject *kobj, struct attribute *attr,
 
 	if (dev->gpu) {
 		log_max_watch_addr =
-			__ilog2_u32(dev->gpu->device_info->num_of_watch_points);
+			__ilog2_u32(dev->gpu->kfd->device_info.num_of_watch_points);
 
 		if (log_max_watch_addr) {
 			dev->node_props.capability |=
@@ -515,24 +506,37 @@ static ssize_t node_show(struct kobject *kobj, struct attribute *attr,
 				HSA_CAP_WATCH_POINTS_TOTALBITS_MASK);
 		}
 
-		if (dev->gpu->device_info->asic_family == CHIP_TONGA)
+		if (dev->gpu->adev->asic_type == CHIP_TONGA)
 			dev->node_props.capability |=
 					HSA_CAP_AQL_QUEUE_DOUBLE_MAP;
 
+		if (KFD_GC_VERSION(dev->gpu) < IP_VERSION(10, 0, 0) &&
+			(dev->gpu->adev->sdma.supported_reset & AMDGPU_RESET_TYPE_PER_QUEUE))
+				dev->node_props.capability2 |= HSA_CAP2_PER_SDMA_QUEUE_RESET_SUPPORTED;
+
 		sysfs_show_32bit_prop(buffer, offs, "max_engine_clk_fcompute",
 			dev->node_props.max_engine_clk_fcompute);
 
 		sysfs_show_64bit_prop(buffer, offs, "local_mem_size", 0ULL);
 
 		sysfs_show_32bit_prop(buffer, offs, "fw_version",
-				      dev->gpu->mec_fw_version);
+				      dev->gpu->kfd->mec_fw_version);
 		sysfs_show_32bit_prop(buffer, offs, "capability",
 				      dev->node_props.capability);
+		sysfs_show_32bit_prop(buffer, offs, "capability2",
+				      dev->node_props.capability2);
+		sysfs_show_64bit_prop(buffer, offs, "debug_prop",
+				      dev->node_props.debug_prop);
 		sysfs_show_32bit_prop(buffer, offs, "sdma_fw_version",
-				      dev->gpu->sdma_fw_version);
+				      dev->gpu->kfd->sdma_fw_version);
 		sysfs_show_64bit_prop(buffer, offs, "unique_id",
-				      amdgpu_amdkfd_get_unique_id(dev->gpu->kgd));
-
+				      dev->gpu->xcp &&
+				      (dev->gpu->xcp->xcp_mgr->mode !=
+				       AMDGPU_SPX_PARTITION_MODE) ?
+				      dev->gpu->xcp->unique_id :
+				      dev->gpu->adev->unique_id);
+		sysfs_show_32bit_prop(buffer, offs, "num_xcc",
+				      NUM_XCC(dev->gpu->xcc_mask));
 	}
 
 	return sysfs_show_32bit_prop(buffer, offs, "max_engine_clk_ccompute",
@@ -543,7 +547,7 @@ static const struct sysfs_ops node_ops = {
 	.show = node_show,
 };
 
-static struct kobj_type node_type = {
+static const struct kobj_type node_type = {
 	.release = kfd_topology_kobj_release,
 	.sysfs_ops = &node_ops,
 };
@@ -557,6 +561,7 @@ static void kfd_remove_sysfs_file(struct kobject *kobj, struct attribute *attr)
 
 static void kfd_remove_sysfs_node_entry(struct kfd_topology_device *dev)
 {
+	struct kfd_iolink_properties *p2plink;
 	struct kfd_iolink_properties *iolink;
 	struct kfd_cache_properties *cache;
 	struct kfd_mem_properties *mem;
@@ -574,6 +579,18 @@ static void kfd_remove_sysfs_node_entry(struct kfd_topology_device *dev)
 		dev->kobj_iolink = NULL;
 	}
 
+	if (dev->kobj_p2plink) {
+		list_for_each_entry(p2plink, &dev->p2p_link_props, list)
+			if (p2plink->kobj) {
+				kfd_remove_sysfs_file(p2plink->kobj,
+							&p2plink->attr);
+				p2plink->kobj = NULL;
+			}
+		kobject_del(dev->kobj_p2plink);
+		kobject_put(dev->kobj_p2plink);
+		dev->kobj_p2plink = NULL;
+	}
+
 	if (dev->kobj_cache) {
 		list_for_each_entry(cache, &dev->cache_props, list)
 			if (cache->kobj) {
@@ -620,6 +637,7 @@ static void kfd_remove_sysfs_node_entry(struct kfd_topology_device *dev)
 static int kfd_build_sysfs_node_entry(struct kfd_topology_device *dev,
 		uint32_t id)
 {
+	struct kfd_iolink_properties *p2plink;
 	struct kfd_iolink_properties *iolink;
 	struct kfd_cache_properties *cache;
 	struct kfd_mem_properties *mem;
@@ -657,6 +675,10 @@ static int kfd_build_sysfs_node_entry(struct kfd_topology_device *dev,
 	if (!dev->kobj_iolink)
 		return -ENOMEM;
 
+	dev->kobj_p2plink = kobject_create_and_add("p2p_links", dev->kobj_node);
+	if (!dev->kobj_p2plink)
+		return -ENOMEM;
+
 	dev->kobj_perf = kobject_create_and_add("perf", dev->kobj_node);
 	if (!dev->kobj_perf)
 		return -ENOMEM;
@@ -746,6 +768,27 @@ static int kfd_build_sysfs_node_entry(struct kfd_topology_device *dev,
 		i++;
 	}
 
+	i = 0;
+	list_for_each_entry(p2plink, &dev->p2p_link_props, list) {
+		p2plink->kobj = kzalloc(sizeof(struct kobject), GFP_KERNEL);
+		if (!p2plink->kobj)
+			return -ENOMEM;
+		ret = kobject_init_and_add(p2plink->kobj, &iolink_type,
+				dev->kobj_p2plink, "%d", i);
+		if (ret < 0) {
+			kobject_put(p2plink->kobj);
+			return ret;
+		}
+
+		p2plink->attr.name = "properties";
+		p2plink->attr.mode = KFD_SYSFS_FILE_MODE;
+		sysfs_attr_init(&p2plink->attr);
+		ret = sysfs_create_file(p2plink->kobj, &p2plink->attr);
+		if (ret < 0)
+			return ret;
+		i++;
+	}
+
 	/* All hardware blocks have the same number of attributes. */
 	num_attrs = ARRAY_SIZE(perf_attr_iommu);
 	list_for_each_entry(perf, &dev->perf_props, list) {
@@ -910,44 +953,31 @@ static void kfd_update_system_properties(void)
 	dev = list_last_entry(&topology_device_list,
 			struct kfd_topology_device, list);
 	if (dev) {
-		sys_props.platform_id =
-			(*((uint64_t *)dev->oem_id)) & CRAT_OEMID_64BIT_MASK;
+		sys_props.platform_id = dev->oem_id64;
 		sys_props.platform_oem = *((uint64_t *)dev->oem_table_id);
 		sys_props.platform_rev = dev->oem_revision;
 	}
 	up_read(&topology_lock);
 }
 
-static void find_system_memory(const struct dmi_header *dm,
-	void *private)
+static void find_system_memory(const struct dmi_header *dm, void *private)
 {
+	struct dmi_mem_device *memdev = container_of(dm, struct dmi_mem_device, header);
 	struct kfd_mem_properties *mem;
-	u16 mem_width, mem_clock;
 	struct kfd_topology_device *kdev =
 		(struct kfd_topology_device *)private;
-	const u8 *dmi_data = (const u8 *)(dm + 1);
-
-	if (dm->type == DMI_ENTRY_MEM_DEVICE && dm->length >= 0x15) {
-		mem_width = (u16)(*(const u16 *)(dmi_data + 0x6));
-		mem_clock = (u16)(*(const u16 *)(dmi_data + 0x11));
-		list_for_each_entry(mem, &kdev->mem_props, list) {
-			if (mem_width != 0xFFFF && mem_width != 0)
-				mem->width = mem_width;
-			if (mem_clock != 0)
-				mem->mem_clk_max = mem_clock;
-		}
-	}
-}
 
-/*
- * Performance counters information is not part of CRAT but we would like to
- * put them in the sysfs under topology directory for Thunk to get the data.
- * This function is called before updating the sysfs.
- */
-static int kfd_add_perf_to_topology(struct kfd_topology_device *kdev)
-{
-	/* These are the only counters supported so far */
-	return kfd_iommu_add_perf_counters(kdev);
+	if (memdev->header.type != DMI_ENTRY_MEM_DEVICE)
+		return;
+	if (memdev->header.length < sizeof(struct dmi_mem_device))
+		return;
+
+	list_for_each_entry(mem, &kdev->mem_props, list) {
+		if (memdev->total_width != 0xFFFF && memdev->total_width != 0)
+			mem->width = memdev->total_width;
+		if (memdev->speed != 0)
+			mem->mem_clk_max = memdev->speed;
+	}
 }
 
 /* kfd_add_non_crat_information - Add information that is not currently
@@ -964,25 +994,6 @@ static void kfd_add_non_crat_information(struct kfd_topology_device *kdev)
 	/* TODO: For GPU node, rearrange code from kfd_topology_add_device */
 }
 
-/* kfd_is_acpi_crat_invalid - CRAT from ACPI is valid only for AMD APU devices.
- *	Ignore CRAT for all other devices. AMD APU is identified if both CPU
- *	and GPU cores are present.
- * @device_list - topology device list created by parsing ACPI CRAT table.
- * @return - TRUE if invalid, FALSE is valid.
- */
-static bool kfd_is_acpi_crat_invalid(struct list_head *device_list)
-{
-	struct kfd_topology_device *dev;
-
-	list_for_each_entry(dev, device_list, list) {
-		if (dev->node_props.cpu_cores_count &&
-			dev->node_props.simd_count)
-			return false;
-	}
-	pr_info("Ignoring ACPI CRAT on non-APU system\n");
-	return true;
-}
-
 int kfd_topology_init(void)
 {
 	void *crat_image = NULL;
@@ -1013,53 +1024,30 @@ int kfd_topology_init(void)
 	 */
 	proximity_domain = 0;
 
-	/*
-	 * Get the CRAT image from the ACPI. If ACPI doesn't have one
-	 * or if ACPI CRAT is invalid create a virtual CRAT.
-	 * NOTE: The current implementation expects all AMD APUs to have
-	 *	CRAT. If no CRAT is available, it is assumed to be a CPU
-	 */
-	ret = kfd_create_crat_image_acpi(&crat_image, &image_size);
-	if (!ret) {
-		ret = kfd_parse_crat_table(crat_image,
-					   &temp_topology_device_list,
-					   proximity_domain);
-		if (ret ||
-		    kfd_is_acpi_crat_invalid(&temp_topology_device_list)) {
-			kfd_release_topology_device_list(
-				&temp_topology_device_list);
-			kfd_destroy_crat_image(crat_image);
-			crat_image = NULL;
-		}
+	ret = kfd_create_crat_image_virtual(&crat_image, &image_size,
+					    COMPUTE_UNIT_CPU, NULL,
+					    proximity_domain);
+	cpu_only_node = 1;
+	if (ret) {
+		pr_err("Error creating VCRAT table for CPU\n");
+		return ret;
 	}
 
-	if (!crat_image) {
-		ret = kfd_create_crat_image_virtual(&crat_image, &image_size,
-						    COMPUTE_UNIT_CPU, NULL,
-						    proximity_domain);
-		cpu_only_node = 1;
-		if (ret) {
-			pr_err("Error creating VCRAT table for CPU\n");
-			return ret;
-		}
-
-		ret = kfd_parse_crat_table(crat_image,
-					   &temp_topology_device_list,
-					   proximity_domain);
-		if (ret) {
-			pr_err("Error parsing VCRAT table for CPU\n");
-			goto err;
-		}
+	ret = kfd_parse_crat_table(crat_image,
+				   &temp_topology_device_list,
+				   proximity_domain);
+	if (ret) {
+		pr_err("Error parsing VCRAT table for CPU\n");
+		goto err;
 	}
 
 	kdev = list_first_entry(&temp_topology_device_list,
 				struct kfd_topology_device, list);
-	kfd_add_perf_to_topology(kdev);
 
 	down_write(&topology_lock);
 	kfd_topology_update_device_list(&temp_topology_device_list,
 					&topology_device_list);
-	atomic_set(&topology_crat_proximity_domain, sys_props.num_devices-1);
+	topology_crat_proximity_domain = sys_props.num_devices-1;
 	ret = kfd_topology_update_sysfs();
 	up_write(&topology_lock);
 
@@ -1095,56 +1083,79 @@ void kfd_topology_shutdown(void)
 	up_write(&topology_lock);
 }
 
-static uint32_t kfd_generate_gpu_id(struct kfd_dev *gpu)
+static uint32_t kfd_generate_gpu_id(struct kfd_node *gpu)
 {
-	uint32_t hashout;
-	uint32_t buf[7];
+	uint32_t gpu_id;
+	uint32_t buf[8];
 	uint64_t local_mem_size;
-	int i;
-	struct kfd_local_mem_info local_mem_info;
+	struct kfd_topology_device *dev;
+	bool is_unique;
+	uint8_t *crc_buf;
 
 	if (!gpu)
 		return 0;
 
-	amdgpu_amdkfd_get_local_mem_info(gpu->kgd, &local_mem_info);
-
-	local_mem_size = local_mem_info.local_mem_size_private +
-			local_mem_info.local_mem_size_public;
-
-	buf[0] = gpu->pdev->devfn;
-	buf[1] = gpu->pdev->subsystem_vendor |
-		(gpu->pdev->subsystem_device << 16);
-	buf[2] = pci_domain_nr(gpu->pdev->bus);
-	buf[3] = gpu->pdev->device;
-	buf[4] = gpu->pdev->bus->number;
+	crc_buf = (uint8_t *)&buf;
+	local_mem_size = gpu->local_mem_info.local_mem_size_private +
+			gpu->local_mem_info.local_mem_size_public;
+	buf[0] = gpu->adev->pdev->devfn;
+	buf[1] = gpu->adev->pdev->subsystem_vendor |
+		(gpu->adev->pdev->subsystem_device << 16);
+	buf[2] = pci_domain_nr(gpu->adev->pdev->bus);
+	buf[3] = gpu->adev->pdev->device;
+	buf[4] = gpu->adev->pdev->bus->number;
 	buf[5] = lower_32_bits(local_mem_size);
 	buf[6] = upper_32_bits(local_mem_size);
+	buf[7] = (ffs(gpu->xcc_mask) - 1) | (NUM_XCC(gpu->xcc_mask) << 16);
+
+	gpu_id = crc16(0, crc_buf, sizeof(buf)) &
+		 ((1 << KFD_GPU_ID_HASH_WIDTH) - 1);
+
+	/* There is a very small possibility when generating a
+	 * 16 (KFD_GPU_ID_HASH_WIDTH) bit value from 8 word buffer
+	 * that the value could be 0 or non-unique. So, check if
+	 * it is unique and non-zero. If not unique increment till
+	 * unique one is found. In case of overflow, restart from 1
+	 */
 
-	for (i = 0, hashout = 0; i < 7; i++)
-		hashout ^= hash_32(buf[i], KFD_GPU_ID_HASH_WIDTH);
+	down_read(&topology_lock);
+	do {
+		is_unique = true;
+		if (!gpu_id)
+			gpu_id = 1;
+		list_for_each_entry(dev, &topology_device_list, list) {
+			if (dev->gpu && dev->gpu_id == gpu_id) {
+				is_unique = false;
+				break;
+			}
+		}
+		if (unlikely(!is_unique))
+			gpu_id = (gpu_id + 1) &
+				  ((1 << KFD_GPU_ID_HASH_WIDTH) - 1);
+	} while (!is_unique);
+	up_read(&topology_lock);
 
-	return hashout;
+	return gpu_id;
 }
 /* kfd_assign_gpu - Attach @gpu to the correct kfd topology device. If
  *		the GPU device is not already present in the topology device
  *		list then return NULL. This means a new topology device has to
  *		be created for this GPU.
  */
-static struct kfd_topology_device *kfd_assign_gpu(struct kfd_dev *gpu)
+static struct kfd_topology_device *kfd_assign_gpu(struct kfd_node *gpu)
 {
 	struct kfd_topology_device *dev;
 	struct kfd_topology_device *out_dev = NULL;
 	struct kfd_mem_properties *mem;
 	struct kfd_cache_properties *cache;
 	struct kfd_iolink_properties *iolink;
+	struct kfd_iolink_properties *p2plink;
 
-	down_write(&topology_lock);
 	list_for_each_entry(dev, &topology_device_list, list) {
 		/* Discrete GPUs need their own topology device list
 		 * entries. Don't assign them to CPU/APU nodes.
 		 */
-		if (!gpu->use_iommu_v2 &&
-		    dev->node_props.cpu_cores_count)
+		if (dev->node_props.cpu_cores_count)
 			continue;
 
 		if (!dev->gpu && (dev->node_props.simd_count > 0)) {
@@ -1157,10 +1168,11 @@ static struct kfd_topology_device *kfd_assign_gpu(struct kfd_dev *gpu)
 				cache->gpu = dev->gpu;
 			list_for_each_entry(iolink, &dev->io_link_props, list)
 				iolink->gpu = dev->gpu;
+			list_for_each_entry(p2plink, &dev->p2p_link_props, list)
+				p2plink->gpu = dev->gpu;
 			break;
 		}
 	}
-	up_write(&topology_lock);
 	return out_dev;
 }
 
@@ -1189,7 +1201,8 @@ static void kfd_fill_mem_clk_max_info(struct kfd_topology_device *dev)
 	 * for APUs - If CRAT from ACPI reports more than one bank, then
 	 *	all the banks will report the same mem_clk_max information
 	 */
-	amdgpu_amdkfd_get_local_mem_info(dev->gpu->kgd, &local_mem_info);
+	amdgpu_amdkfd_get_local_mem_info(dev->gpu->adev, &local_mem_info,
+					 dev->gpu->xcp);
 
 	list_for_each_entry(mem, &dev->mem_props, list)
 		mem->mem_clk_max = local_mem_info.mem_clk_max;
@@ -1207,7 +1220,7 @@ static void kfd_set_iolink_no_atomics(struct kfd_topology_device *dev,
 	if (target_gpu_dev) {
 		uint32_t cap;
 
-		pcie_capability_read_dword(target_gpu_dev->gpu->pdev,
+		pcie_capability_read_dword(target_gpu_dev->gpu->adev->pdev,
 				PCI_EXP_DEVCAP2, &cap);
 
 		if (!(cap & (PCI_EXP_DEVCAP2_ATOMIC_COMP32 |
@@ -1216,9 +1229,8 @@ static void kfd_set_iolink_no_atomics(struct kfd_topology_device *dev,
 				CRAT_IOLINK_FLAGS_NO_ATOMICS_64_BIT;
 	/* set gpu (dev) flags. */
 	} else {
-		if (!dev->gpu->pci_atomic_requested ||
-				dev->gpu->device_info->asic_family ==
-							CHIP_HAWAII)
+		if (!dev->gpu->kfd->pci_atomic_requested ||
+				dev->gpu->adev->asic_type == CHIP_HAWAII)
 			link->flags |= CRAT_IOLINK_FLAGS_NO_ATOMICS_32_BIT |
 				CRAT_IOLINK_FLAGS_NO_ATOMICS_64_BIT;
 	}
@@ -1239,13 +1251,68 @@ static void kfd_set_iolink_non_coherent(struct kfd_topology_device *to_dev,
 		 */
 		if (inbound_link->iolink_type == CRAT_IOLINK_TYPE_PCIEXPRESS ||
 		    (inbound_link->iolink_type == CRAT_IOLINK_TYPE_XGMI &&
-		    to_dev->gpu->device_info->asic_family == CHIP_VEGA20)) {
+		    KFD_GC_VERSION(to_dev->gpu) == IP_VERSION(9, 4, 0))) {
 			outbound_link->flags |= CRAT_IOLINK_FLAGS_NON_COHERENT;
 			inbound_link->flags |= CRAT_IOLINK_FLAGS_NON_COHERENT;
 		}
 	}
 }
 
+#define REC_SDMA_NUM_GPU	8
+static const int rec_sdma_eng_map[REC_SDMA_NUM_GPU][REC_SDMA_NUM_GPU] = {
+							{ -1, 14, 12, 2, 4, 8, 10, 6 },
+							{ 14, -1, 2, 10, 8, 4, 6, 12 },
+							{ 10, 2, -1, 12, 14, 6, 4, 8 },
+							{ 2, 12, 10, -1, 6, 14, 8, 4 },
+							{ 4, 8, 14, 6, -1, 10, 12, 2 },
+							{ 8, 4, 6, 14, 12, -1, 2, 10 },
+							{ 10, 6, 4, 8, 12, 2, -1, 14 },
+							{ 6, 12, 8, 4, 2, 10, 14, -1 }};
+
+static void kfd_set_recommended_sdma_engines(struct kfd_topology_device *to_dev,
+					     struct kfd_iolink_properties *outbound_link,
+					     struct kfd_iolink_properties *inbound_link)
+{
+	struct kfd_node *gpu = outbound_link->gpu;
+	struct amdgpu_device *adev = gpu->adev;
+	unsigned int num_xgmi_nodes = adev->gmc.xgmi.num_physical_nodes;
+	unsigned int num_xgmi_sdma_engines = kfd_get_num_xgmi_sdma_engines(gpu);
+	unsigned int num_sdma_engines = kfd_get_num_sdma_engines(gpu);
+	uint32_t sdma_eng_id_mask = (1 << num_sdma_engines) - 1;
+	uint32_t xgmi_sdma_eng_id_mask =
+			((1 << num_xgmi_sdma_engines) - 1) << num_sdma_engines;
+
+	bool support_rec_eng = !amdgpu_sriov_vf(adev) && to_dev->gpu &&
+		adev->aid_mask && num_xgmi_nodes && gpu->kfd->num_nodes == 1 &&
+		num_xgmi_sdma_engines >= 6 && (!(adev->flags & AMD_IS_APU) &&
+		num_xgmi_nodes == 8);
+
+	if (support_rec_eng) {
+		int src_socket_id = adev->gmc.xgmi.physical_node_id;
+		int dst_socket_id = to_dev->gpu->adev->gmc.xgmi.physical_node_id;
+		unsigned int reshift = num_xgmi_sdma_engines == 6 ? 1 : 0;
+
+		outbound_link->rec_sdma_eng_id_mask =
+			1 << (rec_sdma_eng_map[src_socket_id][dst_socket_id] >> reshift);
+		inbound_link->rec_sdma_eng_id_mask =
+			1 << (rec_sdma_eng_map[dst_socket_id][src_socket_id] >> reshift);
+
+		/* If recommended engine is out of range, need to reset the mask */
+		if (outbound_link->rec_sdma_eng_id_mask & sdma_eng_id_mask)
+			outbound_link->rec_sdma_eng_id_mask = xgmi_sdma_eng_id_mask;
+		if (inbound_link->rec_sdma_eng_id_mask & sdma_eng_id_mask)
+			inbound_link->rec_sdma_eng_id_mask = xgmi_sdma_eng_id_mask;
+
+	} else {
+		uint32_t engine_mask = (outbound_link->iolink_type == CRAT_IOLINK_TYPE_XGMI &&
+				num_xgmi_sdma_engines && to_dev->gpu) ? xgmi_sdma_eng_id_mask :
+				sdma_eng_id_mask;
+
+		outbound_link->rec_sdma_eng_id_mask = engine_mask;
+		inbound_link->rec_sdma_eng_id_mask = engine_mask;
+	}
+}
+
 static void kfd_fill_iolink_non_crat_info(struct kfd_topology_device *dev)
 {
 	struct kfd_iolink_properties *link, *inbound_link;
@@ -1264,6 +1331,18 @@ static void kfd_fill_iolink_non_crat_info(struct kfd_topology_device *dev)
 		if (!peer_dev)
 			continue;
 
+		/* Include the CPU peer in GPU hive if connected over xGMI. */
+		if (!peer_dev->gpu &&
+		    link->iolink_type == CRAT_IOLINK_TYPE_XGMI) {
+			/*
+			 * If the GPU is not part of a GPU hive, use its pci
+			 * device location as the hive ID to bind with the CPU.
+			 */
+			if (!dev->node_props.hive_id)
+				dev->node_props.hive_id = pci_dev_id(dev->gpu->adev->pdev);
+			peer_dev->node_props.hive_id = dev->node_props.hive_id;
+		}
+
 		list_for_each_entry(inbound_link, &peer_dev->io_link_props,
 									list) {
 			if (inbound_link->node_to != link->node_from)
@@ -1272,46 +1351,700 @@ static void kfd_fill_iolink_non_crat_info(struct kfd_topology_device *dev)
 			inbound_link->flags = CRAT_IOLINK_FLAGS_ENABLED;
 			kfd_set_iolink_no_atomics(peer_dev, dev, inbound_link);
 			kfd_set_iolink_non_coherent(peer_dev, link, inbound_link);
+			kfd_set_recommended_sdma_engines(peer_dev, link, inbound_link);
+		}
+	}
+
+	/* Create indirect links so apply flags setting to all */
+	list_for_each_entry(link, &dev->p2p_link_props, list) {
+		link->flags = CRAT_IOLINK_FLAGS_ENABLED;
+		kfd_set_iolink_no_atomics(dev, NULL, link);
+		peer_dev = kfd_topology_device_by_proximity_domain(
+				link->node_to);
+
+		if (!peer_dev)
+			continue;
+
+		list_for_each_entry(inbound_link, &peer_dev->p2p_link_props,
+									list) {
+			if (inbound_link->node_to != link->node_from)
+				continue;
+
+			inbound_link->flags = CRAT_IOLINK_FLAGS_ENABLED;
+			kfd_set_iolink_no_atomics(peer_dev, dev, inbound_link);
+			kfd_set_iolink_non_coherent(peer_dev, link, inbound_link);
 		}
 	}
 }
 
-int kfd_topology_add_device(struct kfd_dev *gpu)
+static int kfd_build_p2p_node_entry(struct kfd_topology_device *dev,
+				struct kfd_iolink_properties *p2plink)
+{
+	int ret;
+
+	p2plink->kobj = kzalloc(sizeof(struct kobject), GFP_KERNEL);
+	if (!p2plink->kobj)
+		return -ENOMEM;
+
+	ret = kobject_init_and_add(p2plink->kobj, &iolink_type,
+			dev->kobj_p2plink, "%d", dev->node_props.p2p_links_count - 1);
+	if (ret < 0) {
+		kobject_put(p2plink->kobj);
+		return ret;
+	}
+
+	p2plink->attr.name = "properties";
+	p2plink->attr.mode = KFD_SYSFS_FILE_MODE;
+	sysfs_attr_init(&p2plink->attr);
+	ret = sysfs_create_file(p2plink->kobj, &p2plink->attr);
+	if (ret < 0)
+		return ret;
+
+	return 0;
+}
+
+static int kfd_create_indirect_link_prop(struct kfd_topology_device *kdev, int gpu_node)
+{
+	struct kfd_iolink_properties *gpu_link, *tmp_link, *cpu_link;
+	struct kfd_iolink_properties *props = NULL, *props2 = NULL;
+	struct kfd_topology_device *cpu_dev;
+	int ret = 0;
+	int i, num_cpu;
+
+	num_cpu = 0;
+	list_for_each_entry(cpu_dev, &topology_device_list, list) {
+		if (cpu_dev->gpu)
+			break;
+		num_cpu++;
+	}
+
+	if (list_empty(&kdev->io_link_props))
+		return -ENODATA;
+
+	gpu_link = list_first_entry(&kdev->io_link_props,
+				    struct kfd_iolink_properties, list);
+
+	for (i = 0; i < num_cpu; i++) {
+		/* CPU <--> GPU */
+		if (gpu_link->node_to == i)
+			continue;
+
+		/* find CPU <-->  CPU links */
+		cpu_link = NULL;
+		cpu_dev = kfd_topology_device_by_proximity_domain(i);
+		if (cpu_dev) {
+			list_for_each_entry(tmp_link,
+					&cpu_dev->io_link_props, list) {
+				if (tmp_link->node_to == gpu_link->node_to) {
+					cpu_link = tmp_link;
+					break;
+				}
+			}
+		}
+
+		if (!cpu_link)
+			return -ENOMEM;
+
+		/* CPU <--> CPU <--> GPU, GPU node*/
+		props = kfd_alloc_struct(props);
+		if (!props)
+			return -ENOMEM;
+
+		memcpy(props, gpu_link, sizeof(struct kfd_iolink_properties));
+		props->weight = gpu_link->weight + cpu_link->weight;
+		props->min_latency = gpu_link->min_latency + cpu_link->min_latency;
+		props->max_latency = gpu_link->max_latency + cpu_link->max_latency;
+		props->min_bandwidth = min(gpu_link->min_bandwidth, cpu_link->min_bandwidth);
+		props->max_bandwidth = min(gpu_link->max_bandwidth, cpu_link->max_bandwidth);
+
+		props->node_from = gpu_node;
+		props->node_to = i;
+		kdev->node_props.p2p_links_count++;
+		list_add_tail(&props->list, &kdev->p2p_link_props);
+		ret = kfd_build_p2p_node_entry(kdev, props);
+		if (ret < 0)
+			return ret;
+
+		/* for small Bar, no CPU --> GPU in-direct links */
+		if (kfd_dev_is_large_bar(kdev->gpu)) {
+			/* CPU <--> CPU <--> GPU, CPU node*/
+			props2 = kfd_alloc_struct(props2);
+			if (!props2)
+				return -ENOMEM;
+
+			memcpy(props2, props, sizeof(struct kfd_iolink_properties));
+			props2->node_from = i;
+			props2->node_to = gpu_node;
+			props2->kobj = NULL;
+			cpu_dev->node_props.p2p_links_count++;
+			list_add_tail(&props2->list, &cpu_dev->p2p_link_props);
+			ret = kfd_build_p2p_node_entry(cpu_dev, props2);
+			if (ret < 0)
+				return ret;
+		}
+	}
+	return ret;
+}
+
+#if defined(CONFIG_HSA_AMD_P2P)
+static int kfd_add_peer_prop(struct kfd_topology_device *kdev,
+		struct kfd_topology_device *peer, int from, int to)
+{
+	struct kfd_iolink_properties *props = NULL;
+	struct kfd_iolink_properties *iolink1, *iolink2, *iolink3;
+	struct kfd_topology_device *cpu_dev;
+	int ret = 0;
+
+	if (!amdgpu_device_is_peer_accessible(
+				kdev->gpu->adev,
+				peer->gpu->adev))
+		return ret;
+
+	if (list_empty(&kdev->io_link_props))
+		return -ENODATA;
+
+	iolink1 = list_first_entry(&kdev->io_link_props,
+				   struct kfd_iolink_properties, list);
+
+	if (list_empty(&peer->io_link_props))
+		return -ENODATA;
+
+	iolink2 = list_first_entry(&peer->io_link_props,
+				   struct kfd_iolink_properties, list);
+
+	props = kfd_alloc_struct(props);
+	if (!props)
+		return -ENOMEM;
+
+	memcpy(props, iolink1, sizeof(struct kfd_iolink_properties));
+
+	props->weight = iolink1->weight + iolink2->weight;
+	props->min_latency = iolink1->min_latency + iolink2->min_latency;
+	props->max_latency = iolink1->max_latency + iolink2->max_latency;
+	props->min_bandwidth = min(iolink1->min_bandwidth, iolink2->min_bandwidth);
+	props->max_bandwidth = min(iolink2->max_bandwidth, iolink2->max_bandwidth);
+
+	if (iolink1->node_to != iolink2->node_to) {
+		/* CPU->CPU  link*/
+		cpu_dev = kfd_topology_device_by_proximity_domain(iolink1->node_to);
+		if (cpu_dev) {
+			list_for_each_entry(iolink3, &cpu_dev->io_link_props, list) {
+				if (iolink3->node_to != iolink2->node_to)
+					continue;
+
+				props->weight += iolink3->weight;
+				props->min_latency += iolink3->min_latency;
+				props->max_latency += iolink3->max_latency;
+				props->min_bandwidth = min(props->min_bandwidth,
+							   iolink3->min_bandwidth);
+				props->max_bandwidth = min(props->max_bandwidth,
+							   iolink3->max_bandwidth);
+				break;
+			}
+		} else {
+			WARN(1, "CPU node not found");
+		}
+	}
+
+	props->node_from = from;
+	props->node_to = to;
+	peer->node_props.p2p_links_count++;
+	list_add_tail(&props->list, &peer->p2p_link_props);
+	ret = kfd_build_p2p_node_entry(peer, props);
+
+	return ret;
+}
+#endif
+
+static int kfd_dev_create_p2p_links(void)
 {
-	uint32_t gpu_id;
 	struct kfd_topology_device *dev;
-	struct kfd_cu_info cu_info;
-	int res = 0;
+	struct kfd_topology_device *new_dev;
+#if defined(CONFIG_HSA_AMD_P2P)
+	uint32_t i;
+#endif
+	uint32_t k;
+	int ret = 0;
+
+	k = 0;
+	list_for_each_entry(dev, &topology_device_list, list)
+		k++;
+	if (k < 2)
+		return 0;
+
+	new_dev = list_last_entry(&topology_device_list, struct kfd_topology_device, list);
+	if (WARN_ON(!new_dev->gpu))
+		return 0;
+
+	k--;
+
+	/* create in-direct links */
+	ret = kfd_create_indirect_link_prop(new_dev, k);
+	if (ret < 0)
+		goto out;
+
+	/* create p2p links */
+#if defined(CONFIG_HSA_AMD_P2P)
+	i = 0;
+	list_for_each_entry(dev, &topology_device_list, list) {
+		if (dev == new_dev)
+			break;
+		if (!dev->gpu || !dev->gpu->adev ||
+		    (dev->gpu->kfd->hive_id &&
+		     dev->gpu->kfd->hive_id == new_dev->gpu->kfd->hive_id &&
+		     amdgpu_xgmi_get_is_sharing_enabled(dev->gpu->adev, new_dev->gpu->adev)))
+			goto next;
+
+		/* check if node(s) is/are peer accessible in one direction or bi-direction */
+		ret = kfd_add_peer_prop(new_dev, dev, i, k);
+		if (ret < 0)
+			goto out;
+
+		ret = kfd_add_peer_prop(dev, new_dev, k, i);
+		if (ret < 0)
+			goto out;
+next:
+		i++;
+	}
+#endif
+
+out:
+	return ret;
+}
+
+/* Helper function. See kfd_fill_gpu_cache_info for parameter description */
+static int fill_in_l1_pcache(struct kfd_cache_properties **props_ext,
+				struct kfd_gpu_cache_info *pcache_info,
+				int cu_bitmask,
+				int cache_type, unsigned int cu_processor_id,
+				int cu_block)
+{
+	unsigned int cu_sibling_map_mask;
+	int first_active_cu;
+	struct kfd_cache_properties *pcache = NULL;
+
+	cu_sibling_map_mask = cu_bitmask;
+	cu_sibling_map_mask >>= cu_block;
+	cu_sibling_map_mask &= ((1 << pcache_info[cache_type].num_cu_shared) - 1);
+	first_active_cu = ffs(cu_sibling_map_mask);
+
+	/* CU could be inactive. In case of shared cache find the first active
+	 * CU. and incase of non-shared cache check if the CU is inactive. If
+	 * inactive active skip it
+	 */
+	if (first_active_cu) {
+		pcache = kfd_alloc_struct(pcache);
+		if (!pcache)
+			return -ENOMEM;
+
+		memset(pcache, 0, sizeof(struct kfd_cache_properties));
+		pcache->processor_id_low = cu_processor_id + (first_active_cu - 1);
+		pcache->cache_level = pcache_info[cache_type].cache_level;
+		pcache->cache_size = pcache_info[cache_type].cache_size;
+		pcache->cacheline_size = pcache_info[cache_type].cache_line_size;
+
+		if (pcache_info[cache_type].flags & CRAT_CACHE_FLAGS_DATA_CACHE)
+			pcache->cache_type |= HSA_CACHE_TYPE_DATA;
+		if (pcache_info[cache_type].flags & CRAT_CACHE_FLAGS_INST_CACHE)
+			pcache->cache_type |= HSA_CACHE_TYPE_INSTRUCTION;
+		if (pcache_info[cache_type].flags & CRAT_CACHE_FLAGS_CPU_CACHE)
+			pcache->cache_type |= HSA_CACHE_TYPE_CPU;
+		if (pcache_info[cache_type].flags & CRAT_CACHE_FLAGS_SIMD_CACHE)
+			pcache->cache_type |= HSA_CACHE_TYPE_HSACU;
+
+		/* Sibling map is w.r.t processor_id_low, so shift out
+		 * inactive CU
+		 */
+		cu_sibling_map_mask =
+			cu_sibling_map_mask >> (first_active_cu - 1);
+
+		pcache->sibling_map[0] = (uint8_t)(cu_sibling_map_mask & 0xFF);
+		pcache->sibling_map[1] =
+				(uint8_t)((cu_sibling_map_mask >> 8) & 0xFF);
+		pcache->sibling_map[2] =
+				(uint8_t)((cu_sibling_map_mask >> 16) & 0xFF);
+		pcache->sibling_map[3] =
+				(uint8_t)((cu_sibling_map_mask >> 24) & 0xFF);
+
+		pcache->sibling_map_size = 4;
+		*props_ext = pcache;
+
+		return 0;
+	}
+	return 1;
+}
+
+/* Helper function. See kfd_fill_gpu_cache_info for parameter description */
+static int fill_in_l2_l3_pcache(struct kfd_cache_properties **props_ext,
+				struct kfd_gpu_cache_info *pcache_info,
+				struct amdgpu_cu_info *cu_info,
+				struct amdgpu_gfx_config *gfx_info,
+				int cache_type, unsigned int cu_processor_id,
+				struct kfd_node *knode)
+{
+	unsigned int cu_sibling_map_mask = 0;
+	int first_active_cu;
+	int i, j, k, xcc, start, end;
+	int num_xcc = NUM_XCC(knode->xcc_mask);
+	struct kfd_cache_properties *pcache = NULL;
+	enum amdgpu_memory_partition mode;
+	struct amdgpu_device *adev = knode->adev;
+	bool found = false;
+
+	start = ffs(knode->xcc_mask) - 1;
+	end = start + num_xcc;
+
+	/* To find the bitmap in the first active cu in the first
+	 * xcc, it is based on the assumption that evrey xcc must
+	 * have at least one active cu.
+	 */
+	for (i = 0; i < gfx_info->max_shader_engines && !found; i++) {
+		for (j = 0; j < gfx_info->max_sh_per_se && !found; j++) {
+			if (cu_info->bitmap[start][i % 4][j % 4]) {
+				cu_sibling_map_mask =
+					cu_info->bitmap[start][i % 4][j % 4];
+				found = true;
+			}
+		}
+	}
+
+	cu_sibling_map_mask &=
+		((1 << pcache_info[cache_type].num_cu_shared) - 1);
+	first_active_cu = ffs(cu_sibling_map_mask);
+
+	/* CU could be inactive. In case of shared cache find the first active
+	 * CU. and incase of non-shared cache check if the CU is inactive. If
+	 * inactive active skip it
+	 */
+	if (first_active_cu) {
+		pcache = kfd_alloc_struct(pcache);
+		if (!pcache)
+			return -ENOMEM;
+
+		memset(pcache, 0, sizeof(struct kfd_cache_properties));
+		pcache->processor_id_low = cu_processor_id
+					+ (first_active_cu - 1);
+		pcache->cache_level = pcache_info[cache_type].cache_level;
+		pcache->cacheline_size = pcache_info[cache_type].cache_line_size;
+
+		if (KFD_GC_VERSION(knode) == IP_VERSION(9, 4, 3) ||
+		    KFD_GC_VERSION(knode) == IP_VERSION(9, 4, 4) ||
+		    KFD_GC_VERSION(knode) == IP_VERSION(9, 5, 0))
+			mode = adev->gmc.gmc_funcs->query_mem_partition_mode(adev);
+		else
+			mode = UNKNOWN_MEMORY_PARTITION_MODE;
+
+		pcache->cache_size = pcache_info[cache_type].cache_size;
+		/* Partition mode only affects L3 cache size */
+		if (mode && pcache->cache_level == 3)
+			pcache->cache_size /= mode;
+
+		if (pcache_info[cache_type].flags & CRAT_CACHE_FLAGS_DATA_CACHE)
+			pcache->cache_type |= HSA_CACHE_TYPE_DATA;
+		if (pcache_info[cache_type].flags & CRAT_CACHE_FLAGS_INST_CACHE)
+			pcache->cache_type |= HSA_CACHE_TYPE_INSTRUCTION;
+		if (pcache_info[cache_type].flags & CRAT_CACHE_FLAGS_CPU_CACHE)
+			pcache->cache_type |= HSA_CACHE_TYPE_CPU;
+		if (pcache_info[cache_type].flags & CRAT_CACHE_FLAGS_SIMD_CACHE)
+			pcache->cache_type |= HSA_CACHE_TYPE_HSACU;
+
+		/* Sibling map is w.r.t processor_id_low, so shift out
+		 * inactive CU
+		 */
+		cu_sibling_map_mask = cu_sibling_map_mask >> (first_active_cu - 1);
+		k = 0;
+
+		for (xcc = start; xcc < end; xcc++) {
+			for (i = 0; i < gfx_info->max_shader_engines; i++) {
+				for (j = 0; j < gfx_info->max_sh_per_se; j++) {
+					pcache->sibling_map[k] = (uint8_t)(cu_sibling_map_mask & 0xFF);
+					pcache->sibling_map[k+1] = (uint8_t)((cu_sibling_map_mask >> 8) & 0xFF);
+					pcache->sibling_map[k+2] = (uint8_t)((cu_sibling_map_mask >> 16) & 0xFF);
+					pcache->sibling_map[k+3] = (uint8_t)((cu_sibling_map_mask >> 24) & 0xFF);
+					k += 4;
+
+					cu_sibling_map_mask = cu_info->bitmap[xcc][i % 4][j + i / 4];
+					cu_sibling_map_mask &= ((1 << pcache_info[cache_type].num_cu_shared) - 1);
+				}
+			}
+		}
+		pcache->sibling_map_size = k;
+		*props_ext = pcache;
+		return 0;
+	}
+	return 1;
+}
+
+#define KFD_MAX_CACHE_TYPES 6
+
+/* kfd_fill_cache_non_crat_info - Fill GPU cache info using kfd_gpu_cache_info
+ * tables
+ */
+static void kfd_fill_cache_non_crat_info(struct kfd_topology_device *dev, struct kfd_node *kdev)
+{
+	struct kfd_gpu_cache_info *pcache_info = NULL;
+	int i, j, k, xcc, start, end;
+	int ct = 0;
+	unsigned int cu_processor_id;
+	int ret;
+	unsigned int num_cu_shared;
+	struct amdgpu_cu_info *cu_info = &kdev->adev->gfx.cu_info;
+	struct amdgpu_gfx_config *gfx_info = &kdev->adev->gfx.config;
+	int gpu_processor_id;
+	struct kfd_cache_properties *props_ext = NULL;
+	int num_of_entries = 0;
+	int num_of_cache_types = 0;
+	struct kfd_gpu_cache_info cache_info[KFD_MAX_CACHE_TYPES];
+
+
+	gpu_processor_id = dev->node_props.simd_id_base;
+
+	memset(cache_info, 0, sizeof(cache_info));
+	pcache_info = cache_info;
+	num_of_cache_types = kfd_get_gpu_cache_info(kdev, &pcache_info);
+	if (!num_of_cache_types) {
+		pr_warn("no cache info found\n");
+		return;
+	}
+
+	/* For each type of cache listed in the kfd_gpu_cache_info table,
+	 * go through all available Compute Units.
+	 * The [i,j,k] loop will
+	 *		if kfd_gpu_cache_info.num_cu_shared = 1
+	 *			will parse through all available CU
+	 *		If (kfd_gpu_cache_info.num_cu_shared != 1)
+	 *			then it will consider only one CU from
+	 *			the shared unit
+	 */
+	start = ffs(kdev->xcc_mask) - 1;
+	end = start + NUM_XCC(kdev->xcc_mask);
+
+	for (ct = 0; ct < num_of_cache_types; ct++) {
+		cu_processor_id = gpu_processor_id;
+		if (pcache_info[ct].cache_level == 1) {
+			for (xcc = start; xcc < end; xcc++) {
+				for (i = 0; i < gfx_info->max_shader_engines; i++) {
+					for (j = 0; j < gfx_info->max_sh_per_se; j++) {
+						for (k = 0; k < gfx_info->max_cu_per_sh; k += pcache_info[ct].num_cu_shared) {
+
+							ret = fill_in_l1_pcache(&props_ext, pcache_info,
+										cu_info->bitmap[xcc][i % 4][j + i / 4], ct,
+										cu_processor_id, k);
+
+							if (ret < 0)
+								break;
+
+							if (!ret) {
+								num_of_entries++;
+								list_add_tail(&props_ext->list, &dev->cache_props);
+							}
+
+							/* Move to next CU block */
+							num_cu_shared = ((k + pcache_info[ct].num_cu_shared) <=
+								gfx_info->max_cu_per_sh) ?
+								pcache_info[ct].num_cu_shared :
+								(gfx_info->max_cu_per_sh - k);
+							cu_processor_id += num_cu_shared;
+						}
+					}
+				}
+			}
+		} else {
+			ret = fill_in_l2_l3_pcache(&props_ext, pcache_info,
+						   cu_info, gfx_info, ct, cu_processor_id, kdev);
+
+			if (ret < 0)
+				break;
+
+			if (!ret) {
+				num_of_entries++;
+				list_add_tail(&props_ext->list, &dev->cache_props);
+			}
+		}
+	}
+	dev->node_props.caches_count += num_of_entries;
+	pr_debug("Added [%d] GPU cache entries\n", num_of_entries);
+}
+
+static int kfd_topology_add_device_locked(struct kfd_node *gpu,
+					  struct kfd_topology_device **dev)
+{
+	int proximity_domain = ++topology_crat_proximity_domain;
 	struct list_head temp_topology_device_list;
 	void *crat_image = NULL;
 	size_t image_size = 0;
-	int proximity_domain;
-	struct amdgpu_device *adev;
+	int res;
+
+	res = kfd_create_crat_image_virtual(&crat_image, &image_size,
+					    COMPUTE_UNIT_GPU, gpu,
+					    proximity_domain);
+	if (res) {
+		dev_err(gpu->adev->dev, "Error creating VCRAT\n");
+		topology_crat_proximity_domain--;
+		goto err;
+	}
 
 	INIT_LIST_HEAD(&temp_topology_device_list);
 
-	gpu_id = kfd_generate_gpu_id(gpu);
+	res = kfd_parse_crat_table(crat_image,
+				   &temp_topology_device_list,
+				   proximity_domain);
+	if (res) {
+		dev_err(gpu->adev->dev, "Error parsing VCRAT\n");
+		topology_crat_proximity_domain--;
+		goto err;
+	}
 
-	pr_debug("Adding new GPU (ID: 0x%x) to topology\n", gpu_id);
+	kfd_topology_update_device_list(&temp_topology_device_list,
+					&topology_device_list);
 
-	proximity_domain = atomic_inc_return(&topology_crat_proximity_domain);
+	*dev = kfd_assign_gpu(gpu);
+	if (WARN_ON(!*dev)) {
+		res = -ENODEV;
+		goto err;
+	}
 
-	adev = (struct amdgpu_device *)(gpu->kgd);
+	/* Fill the cache affinity information here for the GPUs
+	 * using VCRAT
+	 */
+	kfd_fill_cache_non_crat_info(*dev, gpu);
 
-	/* Include the CPU in xGMI hive if xGMI connected by assigning it the hive ID. */
-	if (gpu->hive_id && adev->gmc.xgmi.connected_to_cpu) {
-		struct kfd_topology_device *top_dev;
+	/* Update the SYSFS tree, since we added another topology
+	 * device
+	 */
+	res = kfd_topology_update_sysfs();
+	if (!res)
+		sys_props.generation_count++;
+	else
+		dev_err(gpu->adev->dev, "Failed to update GPU to sysfs topology. res=%d\n",
+			res);
 
-		down_read(&topology_lock);
+err:
+	kfd_destroy_crat_image(crat_image);
+	return res;
+}
 
-		list_for_each_entry(top_dev, &topology_device_list, list) {
-			if (top_dev->gpu)
-				break;
+static void kfd_topology_set_dbg_firmware_support(struct kfd_topology_device *dev)
+{
+	bool firmware_supported = true;
+
+	if (KFD_GC_VERSION(dev->gpu) >= IP_VERSION(11, 0, 0) &&
+			KFD_GC_VERSION(dev->gpu) < IP_VERSION(12, 0, 0)) {
+		uint32_t mes_api_rev = (dev->gpu->adev->mes.sched_version &
+						AMDGPU_MES_API_VERSION_MASK) >>
+						AMDGPU_MES_API_VERSION_SHIFT;
+		uint32_t mes_rev = dev->gpu->adev->mes.sched_version &
+						AMDGPU_MES_VERSION_MASK;
+
+		firmware_supported = (mes_api_rev >= 14) && (mes_rev >= 64);
+		goto out;
+	}
 
-			top_dev->node_props.hive_id = gpu->hive_id;
-		}
+	/*
+	 * Note: Any unlisted devices here are assumed to support exception handling.
+	 * Add additional checks here as needed.
+	 */
+	switch (KFD_GC_VERSION(dev->gpu)) {
+	case IP_VERSION(9, 0, 1):
+		firmware_supported = dev->gpu->kfd->mec_fw_version >= 459 + 32768;
+		break;
+	case IP_VERSION(9, 1, 0):
+	case IP_VERSION(9, 2, 1):
+	case IP_VERSION(9, 2, 2):
+	case IP_VERSION(9, 3, 0):
+	case IP_VERSION(9, 4, 0):
+		firmware_supported = dev->gpu->kfd->mec_fw_version >= 459;
+		break;
+	case IP_VERSION(9, 4, 1):
+		firmware_supported = dev->gpu->kfd->mec_fw_version >= 60;
+		break;
+	case IP_VERSION(9, 4, 2):
+		firmware_supported = dev->gpu->kfd->mec_fw_version >= 51;
+		break;
+	case IP_VERSION(10, 1, 10):
+	case IP_VERSION(10, 1, 2):
+	case IP_VERSION(10, 1, 1):
+		firmware_supported = dev->gpu->kfd->mec_fw_version >= 144;
+		break;
+	case IP_VERSION(10, 3, 0):
+	case IP_VERSION(10, 3, 2):
+	case IP_VERSION(10, 3, 1):
+	case IP_VERSION(10, 3, 4):
+	case IP_VERSION(10, 3, 5):
+		firmware_supported = dev->gpu->kfd->mec_fw_version >= 89;
+		break;
+	case IP_VERSION(10, 1, 3):
+	case IP_VERSION(10, 3, 3):
+		firmware_supported = false;
+		break;
+	default:
+		break;
+	}
 
-		up_read(&topology_lock);
+out:
+	if (firmware_supported)
+		dev->node_props.capability |= HSA_CAP_TRAP_DEBUG_FIRMWARE_SUPPORTED;
+}
+
+static void kfd_topology_set_capabilities(struct kfd_topology_device *dev)
+{
+	dev->node_props.capability |= ((HSA_CAP_DOORBELL_TYPE_2_0 <<
+				HSA_CAP_DOORBELL_TYPE_TOTALBITS_SHIFT) &
+				HSA_CAP_DOORBELL_TYPE_TOTALBITS_MASK);
+
+	dev->node_props.capability |= HSA_CAP_TRAP_DEBUG_SUPPORT |
+			HSA_CAP_TRAP_DEBUG_WAVE_LAUNCH_TRAP_OVERRIDE_SUPPORTED |
+			HSA_CAP_TRAP_DEBUG_WAVE_LAUNCH_MODE_SUPPORTED;
+
+	if (kfd_dbg_has_ttmps_always_setup(dev->gpu))
+		dev->node_props.debug_prop |= HSA_DBG_DISPATCH_INFO_ALWAYS_VALID;
+
+	if (KFD_GC_VERSION(dev->gpu) < IP_VERSION(10, 0, 0)) {
+		if (KFD_GC_VERSION(dev->gpu) == IP_VERSION(9, 4, 3) ||
+		    KFD_GC_VERSION(dev->gpu) == IP_VERSION(9, 4, 4))
+			dev->node_props.debug_prop |=
+				HSA_DBG_WATCH_ADDR_MASK_LO_BIT_GFX9_4_3 |
+				HSA_DBG_WATCH_ADDR_MASK_HI_BIT_GFX9_4_3;
+		else
+			dev->node_props.debug_prop |=
+				HSA_DBG_WATCH_ADDR_MASK_LO_BIT_GFX9 |
+				HSA_DBG_WATCH_ADDR_MASK_HI_BIT;
+
+		if (KFD_GC_VERSION(dev->gpu) >= IP_VERSION(9, 4, 2))
+			dev->node_props.capability |=
+				HSA_CAP_TRAP_DEBUG_PRECISE_MEMORY_OPERATIONS_SUPPORTED;
+
+		if (!amdgpu_sriov_vf(dev->gpu->adev))
+			dev->node_props.capability |= HSA_CAP_PER_QUEUE_RESET_SUPPORTED;
+
+	} else {
+		dev->node_props.debug_prop |= HSA_DBG_WATCH_ADDR_MASK_LO_BIT_GFX10 |
+					HSA_DBG_WATCH_ADDR_MASK_HI_BIT;
+
+		if (KFD_GC_VERSION(dev->gpu) >= IP_VERSION(12, 0, 0))
+			dev->node_props.capability |=
+				HSA_CAP_TRAP_DEBUG_PRECISE_ALU_OPERATIONS_SUPPORTED;
+	}
+
+	kfd_topology_set_dbg_firmware_support(dev);
+}
+
+int kfd_topology_add_device(struct kfd_node *gpu)
+{
+	uint32_t gpu_id;
+	struct kfd_topology_device *dev;
+	int res = 0;
+	int i;
+	const char *asic_name = amdgpu_asic_name[gpu->adev->asic_type];
+	struct amdgpu_gfx_config *gfx_info = &gpu->adev->gfx.config;
+	struct amdgpu_cu_info *cu_info = &gpu->adev->gfx.cu_info;
+
+	if (gpu->xcp && !gpu->xcp->ddev) {
+		dev_warn(gpu->adev->dev,
+			 "Won't add GPU to topology since it has no drm node assigned.");
+		return 0;
+	} else {
+		dev_dbg(gpu->adev->dev, "Adding new GPU to topology\n");
 	}
 
 	/* Check to see if this gpu device exists in the topology_device_list.
@@ -1320,50 +2053,20 @@ int kfd_topology_add_device(struct kfd_dev *gpu)
 	 * CRAT to create a new topology device. Once created assign the gpu to
 	 * that topology device
 	 */
+	down_write(&topology_lock);
 	dev = kfd_assign_gpu(gpu);
-	if (!dev) {
-		res = kfd_create_crat_image_virtual(&crat_image, &image_size,
-						    COMPUTE_UNIT_GPU, gpu,
-						    proximity_domain);
-		if (res) {
-			pr_err("Error creating VCRAT for GPU (ID: 0x%x)\n",
-			       gpu_id);
-			return res;
-		}
-		res = kfd_parse_crat_table(crat_image,
-					   &temp_topology_device_list,
-					   proximity_domain);
-		if (res) {
-			pr_err("Error parsing VCRAT for GPU (ID: 0x%x)\n",
-			       gpu_id);
-			goto err;
-		}
-
-		down_write(&topology_lock);
-		kfd_topology_update_device_list(&temp_topology_device_list,
-			&topology_device_list);
-
-		/* Update the SYSFS tree, since we added another topology
-		 * device
-		 */
-		res = kfd_topology_update_sysfs();
-		up_write(&topology_lock);
-
-		if (!res)
-			sys_props.generation_count++;
-		else
-			pr_err("Failed to update GPU (ID: 0x%x) to sysfs topology. res=%d\n",
-						gpu_id, res);
-		dev = kfd_assign_gpu(gpu);
-		if (WARN_ON(!dev)) {
-			res = -ENODEV;
-			goto err;
-		}
-	}
+	if (!dev)
+		res = kfd_topology_add_device_locked(gpu, &dev);
+	up_write(&topology_lock);
+	if (res)
+		return res;
 
+	gpu_id = kfd_generate_gpu_id(gpu);
 	dev->gpu_id = gpu_id;
 	gpu->id = gpu_id;
 
+	kfd_dev_create_p2p_links();
+
 	/* TODO: Move the following lines to function
 	 *	kfd_add_non_crat_information
 	 */
@@ -1371,46 +2074,56 @@ int kfd_topology_add_device(struct kfd_dev *gpu)
 	/* Fill-in additional information that is not available in CRAT but
 	 * needed for the topology
 	 */
-
-	amdgpu_amdkfd_get_cu_info(dev->gpu->kgd, &cu_info);
-
-	strncpy(dev->node_props.name, gpu->device_info->asic_name,
-			KFD_TOPOLOGY_PUBLIC_NAME_SIZE);
+	for (i = 0; i < KFD_TOPOLOGY_PUBLIC_NAME_SIZE-1; i++) {
+		dev->node_props.name[i] = __tolower(asic_name[i]);
+		if (asic_name[i] == '\0')
+			break;
+	}
+	dev->node_props.name[i] = '\0';
 
 	dev->node_props.simd_arrays_per_engine =
-		cu_info.num_shader_arrays_per_engine;
+		gfx_info->max_sh_per_se;
 
-	dev->node_props.gfx_target_version = gpu->device_info->gfx_target_version;
-	dev->node_props.vendor_id = gpu->pdev->vendor;
-	dev->node_props.device_id = gpu->pdev->device;
+	dev->node_props.gfx_target_version =
+				gpu->kfd->device_info.gfx_target_version;
+	dev->node_props.vendor_id = gpu->adev->pdev->vendor;
+	dev->node_props.device_id = gpu->adev->pdev->device;
 	dev->node_props.capability |=
-		((amdgpu_amdkfd_get_asic_rev_id(dev->gpu->kgd) <<
-			HSA_CAP_ASIC_REVISION_SHIFT) &
+		((dev->gpu->adev->rev_id << HSA_CAP_ASIC_REVISION_SHIFT) &
 			HSA_CAP_ASIC_REVISION_MASK);
-	dev->node_props.location_id = pci_dev_id(gpu->pdev);
-	dev->node_props.domain = pci_domain_nr(gpu->pdev->bus);
+
+	dev->node_props.location_id = pci_dev_id(gpu->adev->pdev);
+	if (gpu->kfd->num_nodes > 1)
+		dev->node_props.location_id |= dev->gpu->node_id;
+
+	dev->node_props.domain = pci_domain_nr(gpu->adev->pdev->bus);
 	dev->node_props.max_engine_clk_fcompute =
-		amdgpu_amdkfd_get_max_engine_clock_in_mhz(dev->gpu->kgd);
+		amdgpu_amdkfd_get_max_engine_clock_in_mhz(dev->gpu->adev);
 	dev->node_props.max_engine_clk_ccompute =
 		cpufreq_quick_get_max(0) / 1000;
-	dev->node_props.drm_render_minor =
-		gpu->shared_resources.drm_render_minor;
 
-	dev->node_props.hive_id = gpu->hive_id;
-	dev->node_props.num_sdma_engines = gpu->device_info->num_sdma_engines;
+	if (gpu->xcp)
+		dev->node_props.drm_render_minor = gpu->xcp->ddev->render->index;
+	else
+		dev->node_props.drm_render_minor =
+				gpu->kfd->shared_resources.drm_render_minor;
+
+	dev->node_props.hive_id = gpu->kfd->hive_id;
+	dev->node_props.num_sdma_engines = kfd_get_num_sdma_engines(gpu);
 	dev->node_props.num_sdma_xgmi_engines =
-				gpu->device_info->num_xgmi_sdma_engines;
+					kfd_get_num_xgmi_sdma_engines(gpu);
 	dev->node_props.num_sdma_queues_per_engine =
-				gpu->device_info->num_sdma_queues_per_engine;
+				gpu->kfd->device_info.num_sdma_queues_per_engine -
+				gpu->kfd->device_info.num_reserved_sdma_queues_per_engine;
 	dev->node_props.num_gws = (dev->gpu->gws &&
 		dev->gpu->dqm->sched_policy != KFD_SCHED_POLICY_NO_HWS) ?
-		amdgpu_amdkfd_get_num_gws(dev->gpu->kgd) : 0;
+		dev->gpu->adev->gds.gws_size : 0;
 	dev->node_props.num_cp_queues = get_cp_queues_num(dev->gpu->dqm);
 
 	kfd_fill_mem_clk_max_info(dev);
 	kfd_fill_iolink_non_crat_info(dev);
 
-	switch (dev->gpu->device_info->asic_family) {
+	switch (dev->gpu->adev->asic_type) {
 	case CHIP_KAVERI:
 	case CHIP_HAWAII:
 	case CHIP_TONGA:
@@ -1429,94 +2142,146 @@ int kfd_topology_add_device(struct kfd_dev *gpu)
 			HSA_CAP_DOORBELL_TYPE_TOTALBITS_SHIFT) &
 			HSA_CAP_DOORBELL_TYPE_TOTALBITS_MASK);
 		break;
-	case CHIP_VEGA10:
-	case CHIP_VEGA12:
-	case CHIP_VEGA20:
-	case CHIP_RAVEN:
-	case CHIP_RENOIR:
-	case CHIP_ARCTURUS:
-	case CHIP_ALDEBARAN:
-	case CHIP_NAVI10:
-	case CHIP_NAVI12:
-	case CHIP_NAVI14:
-	case CHIP_SIENNA_CICHLID:
-	case CHIP_NAVY_FLOUNDER:
-	case CHIP_VANGOGH:
-	case CHIP_DIMGREY_CAVEFISH:
-	case CHIP_BEIGE_GOBY:
-	case CHIP_YELLOW_CARP:
-	case CHIP_CYAN_SKILLFISH:
-		dev->node_props.capability |= ((HSA_CAP_DOORBELL_TYPE_2_0 <<
-			HSA_CAP_DOORBELL_TYPE_TOTALBITS_SHIFT) &
-			HSA_CAP_DOORBELL_TYPE_TOTALBITS_MASK);
-		break;
 	default:
-		WARN(1, "Unexpected ASIC family %u",
-		     dev->gpu->device_info->asic_family);
+		if (KFD_GC_VERSION(dev->gpu) < IP_VERSION(9, 0, 1))
+			WARN(1, "Unexpected ASIC family %u",
+			     dev->gpu->adev->asic_type);
+		else
+			kfd_topology_set_capabilities(dev);
 	}
 
 	/*
-	* Overwrite ATS capability according to needs_iommu_device to fix
-	* potential missing corresponding bit in CRAT of BIOS.
-	*/
-	if (dev->gpu->use_iommu_v2)
-		dev->node_props.capability |= HSA_CAP_ATS_PRESENT;
-	else
-		dev->node_props.capability &= ~HSA_CAP_ATS_PRESENT;
+	 * Overwrite ATS capability according to needs_iommu_device to fix
+	 * potential missing corresponding bit in CRAT of BIOS.
+	 */
+	dev->node_props.capability &= ~HSA_CAP_ATS_PRESENT;
 
 	/* Fix errors in CZ CRAT.
 	 * simd_count: Carrizo CRAT reports wrong simd_count, probably
 	 *		because it doesn't consider masked out CUs
 	 * max_waves_per_simd: Carrizo reports wrong max_waves_per_simd
 	 */
-	if (dev->gpu->device_info->asic_family == CHIP_CARRIZO) {
+	if (dev->gpu->adev->asic_type == CHIP_CARRIZO) {
 		dev->node_props.simd_count =
-			cu_info.simd_per_cu * cu_info.cu_active_number;
+			cu_info->simd_per_cu * cu_info->number;
 		dev->node_props.max_waves_per_simd = 10;
 	}
 
 	/* kfd only concerns sram ecc on GFX and HBM ecc on UMC */
 	dev->node_props.capability |=
-		((adev->ras_enabled & BIT(AMDGPU_RAS_BLOCK__GFX)) != 0) ?
+		((dev->gpu->adev->ras_enabled & BIT(AMDGPU_RAS_BLOCK__GFX)) != 0) ?
 		HSA_CAP_SRAM_EDCSUPPORTED : 0;
-	dev->node_props.capability |= ((adev->ras_enabled & BIT(AMDGPU_RAS_BLOCK__UMC)) != 0) ?
+	dev->node_props.capability |=
+		((dev->gpu->adev->ras_enabled & BIT(AMDGPU_RAS_BLOCK__UMC)) != 0) ?
 		HSA_CAP_MEM_EDCSUPPORTED : 0;
 
-	if (adev->asic_type != CHIP_VEGA10)
-		dev->node_props.capability |= (adev->ras_enabled != 0) ?
+	if (KFD_GC_VERSION(dev->gpu) != IP_VERSION(9, 0, 1))
+		dev->node_props.capability |= (dev->gpu->adev->ras_enabled != 0) ?
 			HSA_CAP_RASEVENTNOTIFY : 0;
 
-	if (KFD_IS_SVM_API_SUPPORTED(adev->kfd.dev))
+	if (KFD_IS_SVM_API_SUPPORTED(dev->gpu->adev))
 		dev->node_props.capability |= HSA_CAP_SVMAPI_SUPPORTED;
 
+	if (dev->gpu->adev->gmc.is_app_apu ||
+		dev->gpu->adev->gmc.xgmi.connected_to_cpu)
+		dev->node_props.capability |= HSA_CAP_FLAGS_COHERENTHOSTACCESS;
+
+	kfd_queue_ctx_save_restore_size(dev);
+
 	kfd_debug_print_topology();
 
-	if (!res)
-		kfd_notify_gpu_change(gpu_id, 1);
-err:
-	kfd_destroy_crat_image(crat_image);
-	return res;
+	kfd_notify_gpu_change(gpu_id, 1);
+
+	return 0;
 }
 
-int kfd_topology_remove_device(struct kfd_dev *gpu)
+/**
+ * kfd_topology_update_io_links() - Update IO links after device removal.
+ * @proximity_domain: Proximity domain value of the dev being removed.
+ *
+ * The topology list currently is arranged in increasing order of
+ * proximity domain.
+ *
+ * Two things need to be done when a device is removed:
+ * 1. All the IO links to this device need to be removed.
+ * 2. All nodes after the current device node need to move
+ *    up once this device node is removed from the topology
+ *    list. As a result, the proximity domain values for
+ *    all nodes after the node being deleted reduce by 1.
+ *    This would also cause the proximity domain values for
+ *    io links to be updated based on new proximity domain
+ *    values.
+ *
+ * Context: The caller must hold write topology_lock.
+ */
+static void kfd_topology_update_io_links(int proximity_domain)
+{
+	struct kfd_topology_device *dev;
+	struct kfd_iolink_properties *iolink, *p2plink, *tmp;
+
+	list_for_each_entry(dev, &topology_device_list, list) {
+		if (dev->proximity_domain > proximity_domain)
+			dev->proximity_domain--;
+
+		list_for_each_entry_safe(iolink, tmp, &dev->io_link_props, list) {
+			/*
+			 * If there is an io link to the dev being deleted
+			 * then remove that IO link also.
+			 */
+			if (iolink->node_to == proximity_domain) {
+				list_del(&iolink->list);
+				dev->node_props.io_links_count--;
+			} else {
+				if (iolink->node_from > proximity_domain)
+					iolink->node_from--;
+				if (iolink->node_to > proximity_domain)
+					iolink->node_to--;
+			}
+		}
+
+		list_for_each_entry_safe(p2plink, tmp, &dev->p2p_link_props, list) {
+			/*
+			 * If there is a p2p link to the dev being deleted
+			 * then remove that p2p link also.
+			 */
+			if (p2plink->node_to == proximity_domain) {
+				list_del(&p2plink->list);
+				dev->node_props.p2p_links_count--;
+			} else {
+				if (p2plink->node_from > proximity_domain)
+					p2plink->node_from--;
+				if (p2plink->node_to > proximity_domain)
+					p2plink->node_to--;
+			}
+		}
+	}
+}
+
+int kfd_topology_remove_device(struct kfd_node *gpu)
 {
 	struct kfd_topology_device *dev, *tmp;
 	uint32_t gpu_id;
 	int res = -ENODEV;
+	int i = 0;
 
 	down_write(&topology_lock);
 
-	list_for_each_entry_safe(dev, tmp, &topology_device_list, list)
+	list_for_each_entry_safe(dev, tmp, &topology_device_list, list) {
 		if (dev->gpu == gpu) {
 			gpu_id = dev->gpu_id;
 			kfd_remove_sysfs_node_entry(dev);
 			kfd_release_topology_device(dev);
 			sys_props.num_devices--;
+			kfd_topology_update_io_links(i);
+			topology_crat_proximity_domain = sys_props.num_devices-1;
+			sys_props.generation_count++;
 			res = 0;
 			if (kfd_topology_update_sysfs() < 0)
 				kfd_topology_release_sysfs();
 			break;
 		}
+		i++;
+	}
 
 	up_write(&topology_lock);
 
@@ -1532,7 +2297,7 @@ int kfd_topology_remove_device(struct kfd_dev *gpu)
  * Return -	0: On success (@kdev will be NULL for non GPU nodes)
  *		-1: If end of list
  */
-int kfd_topology_enum_kfd_devices(uint8_t idx, struct kfd_dev **kdev)
+int kfd_topology_enum_kfd_devices(uint8_t idx, struct kfd_node **kdev)
 {
 
 	struct kfd_topology_device *top_dev;
@@ -1567,7 +2332,7 @@ static int kfd_cpumask_to_apic_id(const struct cpumask *cpumask)
 	if (first_cpu_of_numa_node >= nr_cpu_ids)
 		return -1;
 #ifdef CONFIG_X86_64
-	return cpu_data(first_cpu_of_numa_node).apicid;
+	return cpu_data(first_cpu_of_numa_node).topo.apicid;
 #else
 	return first_cpu_of_numa_node;
 #endif
@@ -1586,29 +2351,6 @@ int kfd_numa_node_to_apic_id(int numa_node_id)
 	return kfd_cpumask_to_apic_id(cpumask_of_node(numa_node_id));
 }
 
-void kfd_double_confirm_iommu_support(struct kfd_dev *gpu)
-{
-	struct kfd_topology_device *dev;
-
-	gpu->use_iommu_v2 = false;
-
-	if (!gpu->device_info->needs_iommu_device)
-		return;
-
-	down_read(&topology_lock);
-
-	/* Only use IOMMUv2 if there is an APU topology node with no GPU
-	 * assigned yet. This GPU will be assigned to it.
-	 */
-	list_for_each_entry(dev, &topology_device_list, list)
-		if (dev->node_props.cpu_cores_count &&
-		    dev->node_props.simd_count &&
-		    !dev->gpu)
-			gpu->use_iommu_v2 = true;
-
-	up_read(&topology_lock);
-}
-
 #if defined(CONFIG_DEBUG_FS)
 
 int kfd_debugfs_hqds_by_device(struct seq_file *m, void *data)
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_topology.h b/drivers/gpu/drm/amd/amdkfd/kfd_topology.h
index a8db017c9b8e..3de8ec0043bb 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_topology.h
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_topology.h
@@ -1,5 +1,6 @@
+/* SPDX-License-Identifier: GPL-2.0 OR MIT */
 /*
- * Copyright 2014 Advanced Micro Devices, Inc.
+ * Copyright 2014-2022 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -23,39 +24,21 @@
 #ifndef __KFD_TOPOLOGY_H__
 #define __KFD_TOPOLOGY_H__
 
+#include <linux/dmi.h>
 #include <linux/types.h>
 #include <linux/list.h>
+#include <linux/kfd_sysfs.h>
 #include "kfd_crat.h"
 
 #define KFD_TOPOLOGY_PUBLIC_NAME_SIZE 32
 
-#define HSA_CAP_HOT_PLUGGABLE			0x00000001
-#define HSA_CAP_ATS_PRESENT			0x00000002
-#define HSA_CAP_SHARED_WITH_GRAPHICS		0x00000004
-#define HSA_CAP_QUEUE_SIZE_POW2			0x00000008
-#define HSA_CAP_QUEUE_SIZE_32BIT		0x00000010
-#define HSA_CAP_QUEUE_IDLE_EVENT		0x00000020
-#define HSA_CAP_VA_LIMIT			0x00000040
-#define HSA_CAP_WATCH_POINTS_SUPPORTED		0x00000080
-#define HSA_CAP_WATCH_POINTS_TOTALBITS_MASK	0x00000f00
-#define HSA_CAP_WATCH_POINTS_TOTALBITS_SHIFT	8
-#define HSA_CAP_DOORBELL_TYPE_TOTALBITS_MASK	0x00003000
-#define HSA_CAP_DOORBELL_TYPE_TOTALBITS_SHIFT	12
-
-#define HSA_CAP_DOORBELL_TYPE_PRE_1_0		0x0
-#define HSA_CAP_DOORBELL_TYPE_1_0		0x1
-#define HSA_CAP_DOORBELL_TYPE_2_0		0x2
-#define HSA_CAP_AQL_QUEUE_DOUBLE_MAP		0x00004000
-
-#define HSA_CAP_RESERVED_WAS_SRAM_EDCSUPPORTED	0x00080000 /* Old buggy user mode depends on this being 0 */
-#define HSA_CAP_MEM_EDCSUPPORTED		0x00100000
-#define HSA_CAP_RASEVENTNOTIFY			0x00200000
-#define HSA_CAP_ASIC_REVISION_MASK		0x03c00000
-#define HSA_CAP_ASIC_REVISION_SHIFT		22
-#define HSA_CAP_SRAM_EDCSUPPORTED		0x04000000
-#define HSA_CAP_SVMAPI_SUPPORTED		0x08000000
-#define HSA_CAP_FLAGS_COHERENTHOSTACCESS	0x10000000
-#define HSA_CAP_RESERVED			0xe00f8000
+#define HSA_DBG_WATCH_ADDR_MASK_LO_BIT_GFX9	6
+#define HSA_DBG_WATCH_ADDR_MASK_LO_BIT_GFX9_4_3 7
+#define HSA_DBG_WATCH_ADDR_MASK_LO_BIT_GFX10	7
+#define HSA_DBG_WATCH_ADDR_MASK_HI_BIT  \
+			(29 << HSA_DBG_WATCH_ADDR_MASK_HI_BIT_SHIFT)
+#define HSA_DBG_WATCH_ADDR_MASK_HI_BIT_GFX9_4_3 \
+			(30 << HSA_DBG_WATCH_ADDR_MASK_HI_BIT_SHIFT)
 
 struct kfd_node_properties {
 	uint64_t hive_id;
@@ -64,9 +47,12 @@ struct kfd_node_properties {
 	uint32_t mem_banks_count;
 	uint32_t caches_count;
 	uint32_t io_links_count;
+	uint32_t p2p_links_count;
 	uint32_t cpu_core_id_base;
 	uint32_t simd_id_base;
 	uint32_t capability;
+	uint32_t capability2;
+	uint64_t debug_prop;
 	uint32_t max_waves_per_simd;
 	uint32_t lds_size_in_kb;
 	uint32_t gds_size_in_kb;
@@ -90,20 +76,13 @@ struct kfd_node_properties {
 	uint32_t num_sdma_xgmi_engines;
 	uint32_t num_sdma_queues_per_engine;
 	uint32_t num_cp_queues;
+	uint32_t cwsr_size;
+	uint32_t ctl_stack_size;
+	uint32_t eop_buffer_size;
+	uint32_t debug_memory_size;
 	char name[KFD_TOPOLOGY_PUBLIC_NAME_SIZE];
 };
 
-#define HSA_MEM_HEAP_TYPE_SYSTEM	0
-#define HSA_MEM_HEAP_TYPE_FB_PUBLIC	1
-#define HSA_MEM_HEAP_TYPE_FB_PRIVATE	2
-#define HSA_MEM_HEAP_TYPE_GPU_GDS	3
-#define HSA_MEM_HEAP_TYPE_GPU_LDS	4
-#define HSA_MEM_HEAP_TYPE_GPU_SCRATCH	5
-
-#define HSA_MEM_FLAGS_HOT_PLUGGABLE		0x00000001
-#define HSA_MEM_FLAGS_NON_VOLATILE		0x00000002
-#define HSA_MEM_FLAGS_RESERVED			0xfffffffc
-
 struct kfd_mem_properties {
 	struct list_head	list;
 	uint32_t		heap_type;
@@ -111,16 +90,12 @@ struct kfd_mem_properties {
 	uint32_t		flags;
 	uint32_t		width;
 	uint32_t		mem_clk_max;
-	struct kfd_dev		*gpu;
+	struct kfd_node		*gpu;
 	struct kobject		*kobj;
 	struct attribute	attr;
 };
 
-#define HSA_CACHE_TYPE_DATA		0x00000001
-#define HSA_CACHE_TYPE_INSTRUCTION	0x00000002
-#define HSA_CACHE_TYPE_CPU		0x00000004
-#define HSA_CACHE_TYPE_HSACU		0x00000008
-#define HSA_CACHE_TYPE_RESERVED		0xfffffff0
+#define CACHE_SIBLINGMAP_SIZE 128
 
 struct kfd_cache_properties {
 	struct list_head	list;
@@ -132,10 +107,11 @@ struct kfd_cache_properties {
 	uint32_t		cache_assoc;
 	uint32_t		cache_latency;
 	uint32_t		cache_type;
-	uint8_t			sibling_map[CRAT_SIBLINGMAP_SIZE];
-	struct kfd_dev		*gpu;
+	uint8_t			sibling_map[CACHE_SIBLINGMAP_SIZE];
+	struct kfd_node		*gpu;
 	struct kobject		*kobj;
 	struct attribute	attr;
+	uint32_t		sibling_map_size;
 };
 
 struct kfd_iolink_properties {
@@ -151,8 +127,9 @@ struct kfd_iolink_properties {
 	uint32_t		min_bandwidth;
 	uint32_t		max_bandwidth;
 	uint32_t		rec_transfer_size;
+	uint32_t		rec_sdma_eng_id_mask;
 	uint32_t		flags;
-	struct kfd_dev		*gpu;
+	struct kfd_node		*gpu;
 	struct kobject		*kobj;
 	struct attribute	attr;
 };
@@ -170,21 +147,24 @@ struct kfd_topology_device {
 	uint32_t			proximity_domain;
 	struct kfd_node_properties	node_props;
 	struct list_head		mem_props;
-	uint32_t			cache_count;
 	struct list_head		cache_props;
-	uint32_t			io_link_count;
 	struct list_head		io_link_props;
+	struct list_head		p2p_link_props;
 	struct list_head		perf_props;
-	struct kfd_dev			*gpu;
+	struct kfd_node			*gpu;
 	struct kobject			*kobj_node;
 	struct kobject			*kobj_mem;
 	struct kobject			*kobj_cache;
 	struct kobject			*kobj_iolink;
+	struct kobject			*kobj_p2plink;
 	struct kobject			*kobj_perf;
 	struct attribute		attr_gpuid;
 	struct attribute		attr_name;
 	struct attribute		attr_props;
-	uint8_t				oem_id[CRAT_OEMID_LENGTH];
+	union {
+		uint8_t				oem_id[CRAT_OEMID_LENGTH];
+		uint64_t			oem_id64;
+	};
 	uint8_t				oem_table_id[CRAT_OEMTABLEID_LENGTH];
 	uint32_t			oem_revision;
 };
@@ -201,6 +181,22 @@ struct kfd_system_properties {
 	struct attribute	attr_props;
 };
 
+struct dmi_mem_device {
+	struct dmi_header header;
+	u16 physical_handle;
+	u16 error_handle;
+	u16 total_width;
+	u16 data_width;
+	u16 size;
+	u8 form_factor;
+	u8 device_set;
+	u8 device_locator;
+	u8 bank_locator;
+	u8 memory_type;
+	u16 type_detail;
+	u16 speed;
+} __packed;
+
 struct kfd_topology_device *kfd_create_topology_device(
 		struct list_head *device_list);
 void kfd_release_topology_device_list(struct list_head *device_list);
diff --git a/drivers/gpu/drm/amd/amdkfd/soc15_int.h b/drivers/gpu/drm/amd/amdkfd/soc15_int.h
index daf3c44547d3..e5c0205f2618 100644
--- a/drivers/gpu/drm/amd/amdkfd/soc15_int.h
+++ b/drivers/gpu/drm/amd/amdkfd/soc15_int.h
@@ -29,9 +29,11 @@
 #define SOC15_INTSRC_CP_BAD_OPCODE	183
 #define SOC15_INTSRC_SQ_INTERRUPT_MSG	239
 #define SOC15_INTSRC_VMC_FAULT		0
+#define SOC15_INTSRC_VMC_UTCL2_POISON	1
 #define SOC15_INTSRC_SDMA_TRAP		224
 #define SOC15_INTSRC_SDMA_ECC		220
-
+#define SOC21_INTSRC_SDMA_TRAP		49
+#define SOC21_INTSRC_SDMA_ECC		62
 
 #define SOC15_CLIENT_ID_FROM_IH_ENTRY(entry) (le32_to_cpu(entry[0]) & 0xff)
 #define SOC15_SOURCE_ID_FROM_IH_ENTRY(entry) (le32_to_cpu(entry[0]) >> 8 & 0xff)
@@ -39,6 +41,7 @@
 #define SOC15_VMID_FROM_IH_ENTRY(entry) (le32_to_cpu(entry[0]) >> 24 & 0xf)
 #define SOC15_VMID_TYPE_FROM_IH_ENTRY(entry) (le32_to_cpu(entry[0]) >> 31 & 0x1)
 #define SOC15_PASID_FROM_IH_ENTRY(entry) (le32_to_cpu(entry[3]) & 0xffff)
+#define SOC15_NODEID_FROM_IH_ENTRY(entry) (le32_to_cpu(entry[3]) >> 16 & 0xff)
 #define SOC15_CONTEXT_ID0_FROM_IH_ENTRY(entry) (le32_to_cpu(entry[4]))
 #define SOC15_CONTEXT_ID1_FROM_IH_ENTRY(entry) (le32_to_cpu(entry[5]))
 #define SOC15_CONTEXT_ID2_FROM_IH_ENTRY(entry) (le32_to_cpu(entry[6]))
diff --git a/drivers/gpu/drm/amd/amdxcp/Makefile b/drivers/gpu/drm/amd/amdxcp/Makefile
new file mode 100644
index 000000000000..870501a4bb8c
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdxcp/Makefile
@@ -0,0 +1,25 @@
+#
+# Copyright 2023 Advanced Micro Devices, Inc.
+#
+# Permission is hereby granted, free of charge, to any person obtaining a
+# copy of this software and associated documentation files (the "Software"),
+# to deal in the Software without restriction, including without limitation
+# the rights to use, copy, modify, merge, publish, distribute, sublicense,
+# and/or sell copies of the Software, and to permit persons to whom the
+# Software is furnished to do so, subject to the following conditions:
+#
+# The above copyright notice and this permission notice shall be included in
+# all copies or substantial portions of the Software.
+#
+# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+# THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+# OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+# ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+# OTHER DEALINGS IN THE SOFTWARE.
+#
+
+amdxcp-y := amdgpu_xcp_drv.o
+
+obj-$(CONFIG_DRM_AMDGPU) += amdxcp.o
diff --git a/drivers/gpu/drm/amd/amdxcp/amdgpu_xcp_drv.c b/drivers/gpu/drm/amd/amdxcp/amdgpu_xcp_drv.c
new file mode 100644
index 000000000000..44009aa8216e
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdxcp/amdgpu_xcp_drv.c
@@ -0,0 +1,152 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/export.h>
+#include <linux/init.h>
+#include <linux/module.h>
+#include <linux/platform_device.h>
+
+#include <drm/drm_drv.h>
+
+#include "amdgpu_xcp_drv.h"
+
+#define MAX_XCP_PLATFORM_DEVICE 64
+
+struct xcp_device {
+	struct drm_device drm;
+	struct platform_device *pdev;
+};
+
+static const struct drm_driver amdgpu_xcp_driver = {
+	.driver_features = DRIVER_GEM | DRIVER_RENDER,
+	.name = "amdgpu_xcp_drv",
+	.major = 1,
+	.minor = 0,
+};
+
+static int8_t pdev_num;
+static struct xcp_device *xcp_dev[MAX_XCP_PLATFORM_DEVICE];
+static DEFINE_MUTEX(xcp_mutex);
+
+int amdgpu_xcp_drm_dev_alloc(struct drm_device **ddev)
+{
+	struct platform_device *pdev;
+	struct xcp_device *pxcp_dev;
+	char dev_name[20];
+	int ret, i;
+
+	guard(mutex)(&xcp_mutex);
+
+	if (pdev_num >= MAX_XCP_PLATFORM_DEVICE)
+		return -ENODEV;
+
+	for (i = 0; i < MAX_XCP_PLATFORM_DEVICE; i++) {
+		if (!xcp_dev[i])
+			break;
+	}
+
+	if (i >= MAX_XCP_PLATFORM_DEVICE)
+		return -ENODEV;
+
+	snprintf(dev_name, sizeof(dev_name), "amdgpu_xcp_%d", i);
+	pdev = platform_device_register_simple(dev_name, -1, NULL, 0);
+	if (IS_ERR(pdev))
+		return PTR_ERR(pdev);
+
+	if (!devres_open_group(&pdev->dev, NULL, GFP_KERNEL)) {
+		ret = -ENOMEM;
+		goto out_unregister;
+	}
+
+	pxcp_dev = devm_drm_dev_alloc(&pdev->dev, &amdgpu_xcp_driver, struct xcp_device, drm);
+	if (IS_ERR(pxcp_dev)) {
+		ret = PTR_ERR(pxcp_dev);
+		goto out_devres;
+	}
+
+	xcp_dev[i] = pxcp_dev;
+	xcp_dev[i]->pdev = pdev;
+	*ddev = &pxcp_dev->drm;
+	pdev_num++;
+
+	return 0;
+
+out_devres:
+	devres_release_group(&pdev->dev, NULL);
+out_unregister:
+	platform_device_unregister(pdev);
+
+	return ret;
+}
+EXPORT_SYMBOL(amdgpu_xcp_drm_dev_alloc);
+
+static void free_xcp_dev(int8_t index)
+{
+	if ((index < MAX_XCP_PLATFORM_DEVICE) && (xcp_dev[index])) {
+		struct platform_device *pdev = xcp_dev[index]->pdev;
+
+		devres_release_group(&pdev->dev, NULL);
+		platform_device_unregister(pdev);
+
+		xcp_dev[index] = NULL;
+		pdev_num--;
+	}
+}
+
+void amdgpu_xcp_drm_dev_free(struct drm_device *ddev)
+{
+	int8_t i;
+
+	guard(mutex)(&xcp_mutex);
+
+	for (i = 0; i < MAX_XCP_PLATFORM_DEVICE; i++) {
+		if ((xcp_dev[i]) && (&xcp_dev[i]->drm == ddev)) {
+			free_xcp_dev(i);
+			break;
+		}
+	}
+}
+EXPORT_SYMBOL(amdgpu_xcp_drm_dev_free);
+
+void amdgpu_xcp_drv_release(void)
+{
+	int8_t i;
+
+	guard(mutex)(&xcp_mutex);
+
+	for (i = 0; pdev_num && i < MAX_XCP_PLATFORM_DEVICE; i++) {
+		free_xcp_dev(i);
+	}
+}
+EXPORT_SYMBOL(amdgpu_xcp_drv_release);
+
+static void __exit amdgpu_xcp_drv_exit(void)
+{
+	amdgpu_xcp_drv_release();
+}
+
+module_exit(amdgpu_xcp_drv_exit);
+
+MODULE_AUTHOR("AMD linux driver team");
+MODULE_DESCRIPTION("AMD XCP PLATFORM DEVICES");
+MODULE_LICENSE("GPL and additional rights");
diff --git a/drivers/gpu/drm/amd/amdxcp/amdgpu_xcp_drv.h b/drivers/gpu/drm/amd/amdxcp/amdgpu_xcp_drv.h
new file mode 100644
index 000000000000..580a1602c8e3
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdxcp/amdgpu_xcp_drv.h
@@ -0,0 +1,30 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef _AMDGPU_XCP_DRV_H_
+#define _AMDGPU_XCP_DRV_H_
+
+int amdgpu_xcp_drm_dev_alloc(struct drm_device **ddev);
+void amdgpu_xcp_drm_dev_free(struct drm_device *ddev);
+void amdgpu_xcp_drv_release(void);
+#endif /* _AMDGPU_XCP_DRV_H_ */
diff --git a/drivers/gpu/drm/amd/display/Kconfig b/drivers/gpu/drm/amd/display/Kconfig
index 127667e549c1..abd3b6564373 100644
--- a/drivers/gpu/drm/amd/display/Kconfig
+++ b/drivers/gpu/drm/amd/display/Kconfig
@@ -1,33 +1,39 @@
 # SPDX-License-Identifier: MIT
+# Copyright © 2019-2024 Advanced Micro Devices, Inc. All rights reserved.
+
 menu "Display Engine Configuration"
 	depends on DRM && DRM_AMDGPU
 
 config DRM_AMD_DC
 	bool "AMD DC - Enable new display engine"
 	default y
+	depends on BROKEN || !CC_IS_CLANG || ARM64 || LOONGARCH || RISCV || SPARC64 || X86_64
+	select CEC_CORE
+	select CEC_NOTIFIER
 	select SND_HDA_COMPONENT if SND_HDA_CORE
-	select DRM_AMD_DC_DCN if (X86 || PPC64) && !(KCOV_INSTRUMENT_ALL && KCOV_ENABLE_COMPARISONS)
+	# !CC_IS_CLANG: https://github.com/ClangBuiltLinux/linux/issues/1752
+	select DRM_AMD_DC_FP if ARCH_HAS_KERNEL_FPU_SUPPORT && !(CC_IS_CLANG && (ARM64 || LOONGARCH || RISCV))
 	help
 	  Choose this option if you want to use the new display engine
 	  support for AMDGPU. This adds required support for Vega and
 	  Raven ASICs.
 
-config DRM_AMD_DC_DCN
-	def_bool n
-	help
-	  Raven, Navi, and newer family support for display engine
+	  calculate_bandwidth() is presently broken on all !(X86_64 || SPARC64 ||
+	  ARM64 || LOONGARCH || RISCV) architectures built with Clang (all released
+	  versions), whereby the stack frame gets blown up to well over 5k.  This
+	  would cause an immediate kernel panic on most architectures.  We'll revert
+	  this when the following bug report has been resolved:
+	  https://github.com/llvm/llvm-project/issues/41896.
 
-config DRM_AMD_DC_HDCP
-	bool "Enable HDCP support in DC"
-	depends on DRM_AMD_DC
+config DRM_AMD_DC_FP
+	def_bool n
 	help
-	  Choose this option if you want to support HDCP authentication.
+	  Floating point support, required for DCN-based SoCs
 
 config DRM_AMD_DC_SI
 	bool "AMD DC support for Southern Islands ASICs"
 	depends on DRM_AMDGPU_SI
 	depends on DRM_AMD_DC
-	default n
 	help
 	  Choose this option to enable new AMD DC support for SI asics
 	  by default. This includes Tahiti, Pitcairn, Cape Verde, Oland.
@@ -41,17 +47,13 @@ config DEBUG_KERNEL_DC
 	  Choose this option if you want to hit kdgb_break in assert.
 
 config DRM_AMD_SECURE_DISPLAY
-        bool "Enable secure display support"
-        default n
-        depends on DEBUG_FS
-        depends on DRM_AMD_DC_DCN
-        help
-            Choose this option if you want to
-            support secure display
-
-            This option enables the calculation
-            of crc of specific region via debugfs.
-            Cooperate with specific DMCU FW.
+	bool "Enable secure display support"
+	depends on DEBUG_FS
+	depends on DRM_AMD_DC_FP
+	help
+	  Choose this option if you want to support secure display
 
+	  This option enables the calculation of crc of specific region via
+	  debugfs. Cooperate with specific DMCU FW.
 
 endmenu
diff --git a/drivers/gpu/drm/amd/display/Makefile b/drivers/gpu/drm/amd/display/Makefile
index 2633de77de5e..0084a8d55254 100644
--- a/drivers/gpu/drm/amd/display/Makefile
+++ b/drivers/gpu/drm/amd/display/Makefile
@@ -29,6 +29,22 @@ AMDDALPATH = $(RELATIVE_AMD_DISPLAY_PATH)
 subdir-ccflags-y += -I$(FULL_AMD_DISPLAY_PATH)/dc/inc/
 subdir-ccflags-y += -I$(FULL_AMD_DISPLAY_PATH)/dc/inc/hw
 subdir-ccflags-y += -I$(FULL_AMD_DISPLAY_PATH)/dc/clk_mgr
+subdir-ccflags-y += -I$(FULL_AMD_DISPLAY_PATH)/dc/hwss
+subdir-ccflags-y += -I$(FULL_AMD_DISPLAY_PATH)/dc/resource
+subdir-ccflags-y += -I$(FULL_AMD_DISPLAY_PATH)/dc/dsc
+subdir-ccflags-y += -I$(FULL_AMD_DISPLAY_PATH)/dc/optc
+subdir-ccflags-y += -I$(FULL_AMD_DISPLAY_PATH)/dc/dpp
+subdir-ccflags-y += -I$(FULL_AMD_DISPLAY_PATH)/dc/hubbub
+subdir-ccflags-y += -I$(FULL_AMD_DISPLAY_PATH)/dc/dccg
+subdir-ccflags-y += -I$(FULL_AMD_DISPLAY_PATH)/dc/hubp
+subdir-ccflags-y += -I$(FULL_AMD_DISPLAY_PATH)/dc/dio
+subdir-ccflags-y += -I$(FULL_AMD_DISPLAY_PATH)/dc/dwb
+subdir-ccflags-y += -I$(FULL_AMD_DISPLAY_PATH)/dc/hpo
+subdir-ccflags-y += -I$(FULL_AMD_DISPLAY_PATH)/dc/mmhubbub
+subdir-ccflags-y += -I$(FULL_AMD_DISPLAY_PATH)/dc/mpc
+subdir-ccflags-y += -I$(FULL_AMD_DISPLAY_PATH)/dc/opp
+subdir-ccflags-y += -I$(FULL_AMD_DISPLAY_PATH)/dc/pg
+subdir-ccflags-y += -I$(FULL_AMD_DISPLAY_PATH)/dc/soc_and_ip_translator
 subdir-ccflags-y += -I$(FULL_AMD_DISPLAY_PATH)/modules/inc
 subdir-ccflags-y += -I$(FULL_AMD_DISPLAY_PATH)/modules/freesync
 subdir-ccflags-y += -I$(FULL_AMD_DISPLAY_PATH)/modules/color
@@ -36,18 +52,14 @@ subdir-ccflags-y += -I$(FULL_AMD_DISPLAY_PATH)/modules/info_packet
 subdir-ccflags-y += -I$(FULL_AMD_DISPLAY_PATH)/modules/power
 subdir-ccflags-y += -I$(FULL_AMD_DISPLAY_PATH)/dmub/inc
 
-ifdef CONFIG_DRM_AMD_DC_HDCP
 subdir-ccflags-y += -I$(FULL_AMD_DISPLAY_PATH)/modules/hdcp
-endif
 
 #TODO: remove when Timing Sync feature is complete
 subdir-ccflags-y += -DBUILD_FEATURE_TIMING_SYNC=0
 
 DAL_LIBS = amdgpu_dm dc	modules/freesync modules/color modules/info_packet modules/power dmub/src
 
-ifdef CONFIG_DRM_AMD_DC_HDCP
 DAL_LIBS += modules/hdcp
-endif
 
 AMD_DAL = $(addsuffix /Makefile, $(addprefix $(FULL_AMD_DISPLAY_PATH)/,$(DAL_LIBS)))
 
diff --git a/drivers/gpu/drm/amd/display/TODO b/drivers/gpu/drm/amd/display/TODO
deleted file mode 100644
index a8a6c106e8c7..000000000000
--- a/drivers/gpu/drm/amd/display/TODO
+++ /dev/null
@@ -1,110 +0,0 @@
-===============================================================================
-TODOs
-===============================================================================
-
-1. Base this on drm-next - WIP
-
-
-2. Cleanup commit history
-
-
-3. WIP - Drop page flip helper and use DRM's version
-
-
-4. DONE - Flatten all DC objects
-    * dc_stream/core_stream/stream should just be dc_stream
-    * Same for other DC objects
-
-    "Is there any major reason to keep all those abstractions?
-
-    Could you collapse everything into struct dc_stream?
-
-    I haven't looked recently but I didn't get the impression there was a
-    lot of design around what was public/protected, more whatever needed
-    to be used by someone else was in public."
-    ~ Dave Airlie
-
-
-5. DONE - Rename DC objects to align more with DRM
-    * dc_surface -> dc_plane_state
-    * dc_stream -> dc_stream_state
-
-
-6. DONE - Per-plane and per-stream validation
-
-
-7. WIP - Per-plane and per-stream commit
-
-
-8. WIP - Split pipe_ctx into plane and stream resource structs
-
-
-9. Attach plane and stream reources to state object instead of validate_context
-
-
-10. Remove dc_edid_caps and drm_helpers_parse_edid_caps
-    * Use drm_display_info instead
-    * Remove DC's edid quirks and rely on DRM's quirks (add quirks if needed)
-
-    "Making sure you use the sink-specific helper libraries and kernel
-    subsystems, since there's really no good reason to have 2nd
-    implementation of those in the kernel. Looks likes that's done for mst
-    and edid parsing. There's still a bit a midlayer feeling to the edid
-    parsing side (e.g. dc_edid_caps and dm_helpers_parse_edid_caps, I
-    think it'd be much better if you convert that over to reading stuff
-    from drm_display_info and if needed, push stuff into the core). Also,
-    I can't come up with a good reason why DC needs all this (except to
-    reimplement half of our edid quirk table, which really isn't a good
-    idea). Might be good if you put this onto the list of things to fix
-    long-term, but imo not a blocker. Definitely make sure new stuff
-    doesn't slip in (i.e. if you start adding edid quirks to DC instead of
-    the drm core, refactoring to use the core edid stuff was pointless)."
-    ~ Daniel Vetter
-
-
-11. Remove dc/i2caux. This folder can be somewhat misleading. It's basically an
-overy complicated HW programming function for sendind and receiving i2c/aux
-commands. We can greatly simplify that and move it into dc/dceXYZ like other
-HW blocks.
-
-12. drm_modeset_lock in MST should no longer be needed in recent kernels
-    * Adopt appropriate locking scheme
-
-13. get_modes and best_encoder callbacks look a bit funny. Can probably rip out
-a few indirections, and consider removing entirely and using the
-drm_atomic_helper_best_encoder default behaviour.
-
-14. core/dc_debug.c, consider switching to the atomic state debug helpers and
-moving all your driver state printing into the various atomic_print_state
-callbacks. There's also plans to expose this stuff in a standard way across all
-drivers, to make debugging userspace compositors easier across different hw.
-
-15. Move DP/HDMI dual mode adaptors to drm_dp_dual_mode_helper.c. See
-dal_ddc_service_i2c_query_dp_dual_mode_adaptor.
-
-16. Move to core SCDC helpers (I think those are new since initial DC review).
-
-17. There's still a pretty massive layer cake around dp aux and DPCD handling,
-with like 3 levels of abstraction and using your own structures instead of the
-stuff in drm_dp_helper.h. drm_dp_helper.h isn't really great and already has 2
-incompatible styles, just means more reasons not to add a third (or well third
-one gets to do the cleanup refactor).
-
-18. There's a pile of sink handling code, both for DP and HDMI where I didn't
-immediately recognize the standard. I think long term it'd be best for the drm
-subsystem if we try to move as much of that into helpers/core as possible, and
-share it with drivers. But that's a very long term goal, and by far not just an
-issue with DC - other drivers, especially around DP sink handling, are equally
-guilty.
-
-19. DONE - The DC logger is still a rather sore thing, but I know that the
-DRM_DEBUG stuff just isn't up to the challenges either. We need to figure out
-something that integrates better with DRM and linux debug printing, while not
-being useless with filtering output. dynamic debug printing might be an option.
-
-20. Use kernel i2c device to program HDMI retimer. Some boards have an HDMI
-retimer that we need to program to pass PHY compliance. Currently that's
-bypassing the i2c device and goes directly to HW. This should be changed.
-
-21. Remove vector.c from dc/basics. It's used in DDC code which can probably
-be simplified enough to no longer need a vector implementation.
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/Makefile b/drivers/gpu/drm/amd/display/amdgpu_dm/Makefile
index 718e123a3230..8e949fe77312 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/Makefile
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/Makefile
@@ -25,19 +25,28 @@
 
 
 
-AMDGPUDM = amdgpu_dm.o amdgpu_dm_irq.o amdgpu_dm_mst_types.o amdgpu_dm_color.o
-
-ifdef CONFIG_DRM_AMD_DC_DCN
-AMDGPUDM += dc_fpu.o
-endif
-
 ifneq ($(CONFIG_DRM_AMD_DC),)
-AMDGPUDM += amdgpu_dm_services.o amdgpu_dm_helpers.o amdgpu_dm_pp_smu.o amdgpu_dm_psr.o
+AMDGPUDM = \
+	amdgpu_dm.o \
+	amdgpu_dm_plane.o \
+	amdgpu_dm_crtc.o \
+	amdgpu_dm_irq.o \
+	amdgpu_dm_mst_types.o \
+	amdgpu_dm_color.o \
+	amdgpu_dm_services.o \
+	amdgpu_dm_helpers.o \
+	amdgpu_dm_pp_smu.o \
+	amdgpu_dm_psr.o \
+	amdgpu_dm_replay.o \
+	amdgpu_dm_quirks.o \
+	amdgpu_dm_wb.o \
+	amdgpu_dm_colorop.o
+
+ifdef CONFIG_DRM_AMD_DC_FP
+AMDGPUDM += dc_fpu.o
 endif
 
-ifdef CONFIG_DRM_AMD_DC_HDCP
 AMDGPUDM += amdgpu_dm_hdcp.o
-endif
 
 ifneq ($(CONFIG_DEBUG_FS),)
 AMDGPUDM += amdgpu_dm_crc.o amdgpu_dm_debugfs.o
@@ -48,3 +57,4 @@ subdir-ccflags-y += -I$(FULL_AMD_DISPLAY_PATH)/dc
 AMDGPU_DM = $(addprefix $(AMDDALPATH)/amdgpu_dm/,$(AMDGPUDM))
 
 AMD_DISPLAY_FILES += $(AMDGPU_DM)
+endif
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
index e727f1dd2a9a..740711ac1037 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
@@ -1,3 +1,4 @@
+// SPDX-License-Identifier: MIT
 /*
  * Copyright 2015 Advanced Micro Devices, Inc.
  *
@@ -28,7 +29,6 @@
 
 #include "dm_services_types.h"
 #include "dc.h"
-#include "dc_link_dp.h"
 #include "link_enc_cfg.h"
 #include "dc/inc/core_types.h"
 #include "dal_asic_id.h"
@@ -38,19 +38,23 @@
 #include "dc/dc_dmub_srv.h"
 #include "dc/dc_edid_parser.h"
 #include "dc/dc_stat.h"
+#include "dc/dc_state.h"
 #include "amdgpu_dm_trace.h"
+#include "link/protocols/link_dpcd.h"
+#include "link_service_types.h"
+#include "link/protocols/link_dp_capability.h"
+#include "link/protocols/link_ddc.h"
 
-#include "vid.h"
 #include "amdgpu.h"
 #include "amdgpu_display.h"
 #include "amdgpu_ucode.h"
 #include "atom.h"
 #include "amdgpu_dm.h"
-#ifdef CONFIG_DRM_AMD_DC_HDCP
+#include "amdgpu_dm_plane.h"
+#include "amdgpu_dm_crtc.h"
 #include "amdgpu_dm_hdcp.h"
-#include <drm/drm_hdcp.h>
-#endif
-#include "amdgpu_pm.h"
+#include <drm/display/drm_hdcp_helper.h>
+#include "amdgpu_dm_wb.h"
 #include "amdgpu_atombios.h"
 
 #include "amd_shared.h"
@@ -61,42 +65,46 @@
 #include "amdgpu_dm_debugfs.h"
 #endif
 #include "amdgpu_dm_psr.h"
+#include "amdgpu_dm_replay.h"
 
 #include "ivsrcid/ivsrcid_vislands30.h"
 
-#include "i2caux_interface.h"
+#include <linux/backlight.h>
 #include <linux/module.h>
 #include <linux/moduleparam.h>
 #include <linux/types.h>
 #include <linux/pm_runtime.h>
 #include <linux/pci.h>
+#include <linux/power_supply.h>
 #include <linux/firmware.h>
 #include <linux/component.h>
+#include <linux/sort.h>
 
+#include <drm/drm_privacy_screen_consumer.h>
+#include <drm/display/drm_dp_mst_helper.h>
+#include <drm/display/drm_hdmi_helper.h>
 #include <drm/drm_atomic.h>
 #include <drm/drm_atomic_uapi.h>
 #include <drm/drm_atomic_helper.h>
-#include <drm/drm_dp_mst_helper.h>
-#include <drm/drm_fb_helper.h>
+#include <drm/drm_blend.h>
+#include <drm/drm_fixed.h>
 #include <drm/drm_fourcc.h>
 #include <drm/drm_edid.h>
+#include <drm/drm_eld.h>
+#include <drm/drm_utils.h>
 #include <drm/drm_vblank.h>
 #include <drm/drm_audio_component.h>
+#include <drm/drm_gem_atomic_helper.h>
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-#include "ivsrcid/dcn/irqsrcs_dcn_1_0.h"
-
-#include "dcn/dcn_1_0_offset.h"
-#include "dcn/dcn_1_0_sh_mask.h"
-#include "soc15_hw_ip.h"
-#include "vega10_ip_offset.h"
+#include <media/cec-notifier.h>
+#include <acpi/video.h>
 
-#include "soc15_common.h"
-#endif
+#include "ivsrcid/dcn/irqsrcs_dcn_1_0.h"
 
 #include "modules/inc/mod_freesync.h"
 #include "modules/power/power_helpers.h"
-#include "modules/inc/mod_info_packet.h"
+
+static_assert(AMDGPU_DMUB_NOTIFICATION_MAX == DMUB_NOTIFICATION_MAX, "AMDGPU_DMUB_NOTIFICATION_MAX mismatch");
 
 #define FIRMWARE_RENOIR_DMUB "amdgpu/renoir_dmcub.bin"
 MODULE_FIRMWARE(FIRMWARE_RENOIR_DMUB);
@@ -114,6 +122,17 @@ MODULE_FIRMWARE(FIRMWARE_DIMGREY_CAVEFISH_DMUB);
 MODULE_FIRMWARE(FIRMWARE_BEIGE_GOBY_DMUB);
 #define FIRMWARE_YELLOW_CARP_DMUB "amdgpu/yellow_carp_dmcub.bin"
 MODULE_FIRMWARE(FIRMWARE_YELLOW_CARP_DMUB);
+#define FIRMWARE_DCN_314_DMUB "amdgpu/dcn_3_1_4_dmcub.bin"
+MODULE_FIRMWARE(FIRMWARE_DCN_314_DMUB);
+#define FIRMWARE_DCN_315_DMUB "amdgpu/dcn_3_1_5_dmcub.bin"
+MODULE_FIRMWARE(FIRMWARE_DCN_315_DMUB);
+#define FIRMWARE_DCN316_DMUB "amdgpu/dcn_3_1_6_dmcub.bin"
+MODULE_FIRMWARE(FIRMWARE_DCN316_DMUB);
+
+#define FIRMWARE_DCN_V3_2_0_DMCUB "amdgpu/dcn_3_2_0_dmcub.bin"
+MODULE_FIRMWARE(FIRMWARE_DCN_V3_2_0_DMCUB);
+#define FIRMWARE_DCN_V3_2_1_DMCUB "amdgpu/dcn_3_2_1_dmcub.bin"
+MODULE_FIRMWARE(FIRMWARE_DCN_V3_2_1_DMCUB);
 
 #define FIRMWARE_RAVEN_DMCU		"amdgpu/raven_dmcu.bin"
 MODULE_FIRMWARE(FIRMWARE_RAVEN_DMCU);
@@ -121,6 +140,18 @@ MODULE_FIRMWARE(FIRMWARE_RAVEN_DMCU);
 #define FIRMWARE_NAVI12_DMCU            "amdgpu/navi12_dmcu.bin"
 MODULE_FIRMWARE(FIRMWARE_NAVI12_DMCU);
 
+#define FIRMWARE_DCN_35_DMUB "amdgpu/dcn_3_5_dmcub.bin"
+MODULE_FIRMWARE(FIRMWARE_DCN_35_DMUB);
+
+#define FIRMWARE_DCN_351_DMUB "amdgpu/dcn_3_5_1_dmcub.bin"
+MODULE_FIRMWARE(FIRMWARE_DCN_351_DMUB);
+
+#define FIRMWARE_DCN_36_DMUB "amdgpu/dcn_3_6_dmcub.bin"
+MODULE_FIRMWARE(FIRMWARE_DCN_36_DMUB);
+
+#define FIRMWARE_DCN_401_DMUB "amdgpu/dcn_4_0_1_dmcub.bin"
+MODULE_FIRMWARE(FIRMWARE_DCN_401_DMUB);
+
 /* Number of bytes in PSP header for firmware. */
 #define PSP_HEADER_BYTES 0x100
 
@@ -141,6 +172,9 @@ MODULE_FIRMWARE(FIRMWARE_NAVI12_DMCU);
 static int amdgpu_dm_init(struct amdgpu_device *adev);
 static void amdgpu_dm_fini(struct amdgpu_device *adev);
 static bool is_freesync_video_mode(const struct drm_display_mode *mode, struct amdgpu_dm_connector *aconnector);
+static void reset_freesync_config_for_crtc(struct dm_crtc_state *new_crtc_state);
+static struct amdgpu_i2c_adapter *
+create_i2c(struct ddc_service *ddc_service, bool oem);
 
 static enum drm_mode_subconnector get_subconnector_type(struct dc_link *link)
 {
@@ -189,16 +223,9 @@ static int amdgpu_dm_initialize_drm_device(struct amdgpu_device *adev);
 /* removes and deallocates the drm structures, created by the above function */
 static void amdgpu_dm_destroy_drm_device(struct amdgpu_display_manager *dm);
 
-static int amdgpu_dm_plane_init(struct amdgpu_display_manager *dm,
-				struct drm_plane *plane,
-				unsigned long possible_crtcs,
-				const struct dc_plane_cap *plane_cap);
-static int amdgpu_dm_crtc_init(struct amdgpu_display_manager *dm,
-			       struct drm_plane *plane,
-			       uint32_t link_index);
 static int amdgpu_dm_connector_init(struct amdgpu_display_manager *dm,
 				    struct amdgpu_dm_connector *amdgpu_dm_connector,
-				    uint32_t link_index,
+				    u32 link_index,
 				    struct amdgpu_encoder *amdgpu_encoder);
 static int amdgpu_dm_encoder_init(struct drm_device *dev,
 				  struct amdgpu_encoder *aencoder,
@@ -206,20 +233,19 @@ static int amdgpu_dm_encoder_init(struct drm_device *dev,
 
 static int amdgpu_dm_connector_get_modes(struct drm_connector *connector);
 
+static int amdgpu_dm_atomic_setup_commit(struct drm_atomic_state *state);
 static void amdgpu_dm_atomic_commit_tail(struct drm_atomic_state *state);
 
 static int amdgpu_dm_atomic_check(struct drm_device *dev,
 				  struct drm_atomic_state *state);
 
-static void handle_cursor_update(struct drm_plane *plane,
-				 struct drm_plane_state *old_plane_state);
-
-static const struct drm_format_info *
-amd_get_format_info(const struct drm_mode_fb_cmd2 *cmd);
-
 static void handle_hpd_irq_helper(struct amdgpu_dm_connector *aconnector);
 static void handle_hpd_rx_irq(void *param);
 
+static void amdgpu_dm_backlight_set_level(struct amdgpu_display_manager *dm,
+					 int bl_idx,
+					 u32 user_brightness);
+
 static bool
 is_timing_unchanged_for_freesync(struct drm_crtc_state *old_crtc_state,
 				 struct drm_crtc_state *new_crtc_state);
@@ -238,72 +264,77 @@ is_timing_unchanged_for_freesync(struct drm_crtc_state *old_crtc_state,
  */
 static u32 dm_vblank_get_counter(struct amdgpu_device *adev, int crtc)
 {
+	struct amdgpu_crtc *acrtc = NULL;
+
 	if (crtc >= adev->mode_info.num_crtc)
 		return 0;
-	else {
-		struct amdgpu_crtc *acrtc = adev->mode_info.crtcs[crtc];
 
-		if (acrtc->dm_irq_params.stream == NULL) {
-			DRM_ERROR("dc_stream_state is NULL for crtc '%d'!\n",
-				  crtc);
-			return 0;
-		}
+	acrtc = adev->mode_info.crtcs[crtc];
 
-		return dc_stream_get_vblank_counter(acrtc->dm_irq_params.stream);
+	if (!acrtc->dm_irq_params.stream) {
+		drm_err(adev_to_drm(adev), "dc_stream_state is NULL for crtc '%d'!\n",
+			  crtc);
+		return 0;
 	}
+
+	return dc_stream_get_vblank_counter(acrtc->dm_irq_params.stream);
 }
 
 static int dm_crtc_get_scanoutpos(struct amdgpu_device *adev, int crtc,
 				  u32 *vbl, u32 *position)
 {
-	uint32_t v_blank_start, v_blank_end, h_position, v_position;
+	u32 v_blank_start = 0, v_blank_end = 0, h_position = 0, v_position = 0;
+	struct amdgpu_crtc *acrtc = NULL;
+	struct dc *dc = adev->dm.dc;
 
 	if ((crtc < 0) || (crtc >= adev->mode_info.num_crtc))
 		return -EINVAL;
-	else {
-		struct amdgpu_crtc *acrtc = adev->mode_info.crtcs[crtc];
-
-		if (acrtc->dm_irq_params.stream ==  NULL) {
-			DRM_ERROR("dc_stream_state is NULL for crtc '%d'!\n",
-				  crtc);
-			return 0;
-		}
 
-		/*
-		 * TODO rework base driver to use values directly.
-		 * for now parse it back into reg-format
-		 */
-		dc_stream_get_scanoutpos(acrtc->dm_irq_params.stream,
-					 &v_blank_start,
-					 &v_blank_end,
-					 &h_position,
-					 &v_position);
+	acrtc = adev->mode_info.crtcs[crtc];
 
-		*position = v_position | (h_position << 16);
-		*vbl = v_blank_start | (v_blank_end << 16);
+	if (!acrtc->dm_irq_params.stream) {
+		drm_err(adev_to_drm(adev), "dc_stream_state is NULL for crtc '%d'!\n",
+			  crtc);
+		return 0;
 	}
 
+	if (dc && dc->caps.ips_support && dc->idle_optimizations_allowed)
+		dc_allow_idle_optimizations(dc, false);
+
+	/*
+	 * TODO rework base driver to use values directly.
+	 * for now parse it back into reg-format
+	 */
+	dc_stream_get_scanoutpos(acrtc->dm_irq_params.stream,
+				 &v_blank_start,
+				 &v_blank_end,
+				 &h_position,
+				 &v_position);
+
+	*position = v_position | (h_position << 16);
+	*vbl = v_blank_start | (v_blank_end << 16);
+
 	return 0;
 }
 
-static bool dm_is_idle(void *handle)
+static bool dm_is_idle(struct amdgpu_ip_block *ip_block)
 {
 	/* XXX todo */
 	return true;
 }
 
-static int dm_wait_for_idle(void *handle)
+static int dm_wait_for_idle(struct amdgpu_ip_block *ip_block)
 {
 	/* XXX todo */
 	return 0;
 }
 
-static bool dm_check_soft_reset(void *handle)
+static bool dm_check_soft_reset(struct amdgpu_ip_block *ip_block)
 {
 	return false;
 }
 
-static int dm_soft_reset(void *handle)
+static int dm_soft_reset(struct amdgpu_ip_block *ip_block)
 {
 	/* XXX todo */
 	return 0;
@@ -330,31 +361,72 @@ get_crtc_by_otg_inst(struct amdgpu_device *adev,
 	return NULL;
 }
 
-static inline bool amdgpu_dm_vrr_active_irq(struct amdgpu_crtc *acrtc)
-{
-	return acrtc->dm_irq_params.freesync_config.state ==
-		       VRR_STATE_ACTIVE_VARIABLE ||
-	       acrtc->dm_irq_params.freesync_config.state ==
-		       VRR_STATE_ACTIVE_FIXED;
-}
-
-static inline bool amdgpu_dm_vrr_active(struct dm_crtc_state *dm_state)
-{
-	return dm_state->freesync_config.state == VRR_STATE_ACTIVE_VARIABLE ||
-	       dm_state->freesync_config.state == VRR_STATE_ACTIVE_FIXED;
-}
-
 static inline bool is_dc_timing_adjust_needed(struct dm_crtc_state *old_state,
 					      struct dm_crtc_state *new_state)
 {
+	if (new_state->stream->adjust.timing_adjust_pending)
+		return true;
 	if (new_state->freesync_config.state ==  VRR_STATE_ACTIVE_FIXED)
 		return true;
-	else if (amdgpu_dm_vrr_active(old_state) != amdgpu_dm_vrr_active(new_state))
+	else if (amdgpu_dm_crtc_vrr_active(old_state) != amdgpu_dm_crtc_vrr_active(new_state))
 		return true;
 	else
 		return false;
 }
 
+/*
+ * DC will program planes with their z-order determined by their ordering
+ * in the dc_surface_updates array. This comparator is used to sort them
+ * by descending zpos.
+ */
+static int dm_plane_layer_index_cmp(const void *a, const void *b)
+{
+	const struct dc_surface_update *sa = (struct dc_surface_update *)a;
+	const struct dc_surface_update *sb = (struct dc_surface_update *)b;
+
+	/* Sort by descending dc_plane layer_index (i.e. normalized_zpos) */
+	return sb->surface->layer_index - sa->surface->layer_index;
+}
+
+/**
+ * update_planes_and_stream_adapter() - Send planes to be updated in DC
+ *
+ * DC has a generic way to update planes and stream via
+ * dc_update_planes_and_stream function; however, DM might need some
+ * adjustments and preparation before calling it. This function is a wrapper
+ * for the dc_update_planes_and_stream that does any required configuration
+ * before passing control to DC.
+ *
+ * @dc: Display Core control structure
+ * @update_type: specify whether it is FULL/MEDIUM/FAST update
+ * @planes_count: planes count to update
+ * @stream: stream state
+ * @stream_update: stream update
+ * @array_of_surface_update: dc surface update pointer
+ *
+ */
+static inline bool update_planes_and_stream_adapter(struct dc *dc,
+						    int update_type,
+						    int planes_count,
+						    struct dc_stream_state *stream,
+						    struct dc_stream_update *stream_update,
+						    struct dc_surface_update *array_of_surface_update)
+{
+	sort(array_of_surface_update, planes_count,
+	     sizeof(*array_of_surface_update), dm_plane_layer_index_cmp, NULL);
+
+	/*
+	 * Previous frame finished and HW is ready for optimization.
+	 */
+	dc_post_update_surfaces_to_stream(dc);
+
+	return dc_update_planes_and_stream(dc,
+					   array_of_surface_update,
+					   planes_count,
+					   stream,
+					   stream_update);
+}
+
 /**
  * dm_pflip_high_irq() - Handle pageflip interrupt
  * @interrupt_params: ignored
@@ -367,9 +439,10 @@ static void dm_pflip_high_irq(void *interrupt_params)
 	struct amdgpu_crtc *amdgpu_crtc;
 	struct common_irq_params *irq_params = interrupt_params;
 	struct amdgpu_device *adev = irq_params->adev;
+	struct drm_device *dev = adev_to_drm(adev);
 	unsigned long flags;
 	struct drm_pending_vblank_event *e;
-	uint32_t vpos, hpos, v_blank_start, v_blank_end;
+	u32 vpos, hpos, v_blank_start, v_blank_end;
 	bool vrr_active;
 
 	amdgpu_crtc = get_crtc_by_otg_inst(adev, irq_params->irq_src - IRQ_TYPE_PFLIP);
@@ -377,18 +450,17 @@ static void dm_pflip_high_irq(void *interrupt_params)
 	/* IRQ could occur when in initial stage */
 	/* TODO work and BO cleanup */
 	if (amdgpu_crtc == NULL) {
-		DC_LOG_PFLIP("CRTC is null, returning.\n");
+		drm_dbg_state(dev, "CRTC is null, returning.\n");
 		return;
 	}
 
 	spin_lock_irqsave(&adev_to_drm(adev)->event_lock, flags);
 
-	if (amdgpu_crtc->pflip_status != AMDGPU_FLIP_SUBMITTED){
-		DC_LOG_PFLIP("amdgpu_crtc->pflip_status = %d !=AMDGPU_FLIP_SUBMITTED(%d) on crtc:%d[%p] \n",
-						 amdgpu_crtc->pflip_status,
-						 AMDGPU_FLIP_SUBMITTED,
-						 amdgpu_crtc->crtc_id,
-						 amdgpu_crtc);
+	if (amdgpu_crtc->pflip_status != AMDGPU_FLIP_SUBMITTED) {
+		drm_dbg_state(dev,
+			      "amdgpu_crtc->pflip_status = %d != AMDGPU_FLIP_SUBMITTED(%d) on crtc:%d[%p]\n",
+			      amdgpu_crtc->pflip_status, AMDGPU_FLIP_SUBMITTED,
+			      amdgpu_crtc->crtc_id, amdgpu_crtc);
 		spin_unlock_irqrestore(&adev_to_drm(adev)->event_lock, flags);
 		return;
 	}
@@ -399,7 +471,7 @@ static void dm_pflip_high_irq(void *interrupt_params)
 
 	WARN_ON(!e);
 
-	vrr_active = amdgpu_dm_vrr_active_irq(amdgpu_crtc);
+	vrr_active = amdgpu_dm_crtc_vrr_active_irq(amdgpu_crtc);
 
 	/* Fixed refresh rate, or VRR scanout position outside front-porch? */
 	if (!vrr_active ||
@@ -454,9 +526,53 @@ static void dm_pflip_high_irq(void *interrupt_params)
 	amdgpu_crtc->pflip_status = AMDGPU_FLIP_NONE;
 	spin_unlock_irqrestore(&adev_to_drm(adev)->event_lock, flags);
 
-	DC_LOG_PFLIP("crtc:%d[%p], pflip_stat:AMDGPU_FLIP_NONE, vrr[%d]-fp %d\n",
-		     amdgpu_crtc->crtc_id, amdgpu_crtc,
-		     vrr_active, (int) !e);
+	drm_dbg_state(dev,
+		      "crtc:%d[%p], pflip_stat:AMDGPU_FLIP_NONE, vrr[%d]-fp %d\n",
+		      amdgpu_crtc->crtc_id, amdgpu_crtc, vrr_active, (int)!e);
+}
+
+static void dm_handle_vmin_vmax_update(struct work_struct *offload_work)
+{
+	struct vupdate_offload_work *work = container_of(offload_work, struct vupdate_offload_work, work);
+	struct amdgpu_device *adev = work->adev;
+	struct dc_stream_state *stream = work->stream;
+	struct dc_crtc_timing_adjust *adjust = work->adjust;
+
+	mutex_lock(&adev->dm.dc_lock);
+	dc_stream_adjust_vmin_vmax(adev->dm.dc, stream, adjust);
+	mutex_unlock(&adev->dm.dc_lock);
+
+	dc_stream_release(stream);
+	kfree(work->adjust);
+	kfree(work);
+}
+
+static void schedule_dc_vmin_vmax(struct amdgpu_device *adev,
+	struct dc_stream_state *stream,
+	struct dc_crtc_timing_adjust *adjust)
+{
+	struct vupdate_offload_work *offload_work = kzalloc(sizeof(*offload_work), GFP_NOWAIT);
+	if (!offload_work) {
+		drm_dbg_driver(adev_to_drm(adev), "Failed to allocate vupdate_offload_work\n");
+		return;
+	}
+
+	struct dc_crtc_timing_adjust *adjust_copy = kzalloc(sizeof(*adjust_copy), GFP_NOWAIT);
+	if (!adjust_copy) {
+		drm_dbg_driver(adev_to_drm(adev), "Failed to allocate adjust_copy\n");
+		kfree(offload_work);
+		return;
+	}
+
+	dc_stream_retain(stream);
+	memcpy(adjust_copy, adjust, sizeof(*adjust_copy));
+
+	INIT_WORK(&offload_work->work, dm_handle_vmin_vmax_update);
+	offload_work->adev = adev;
+	offload_work->stream = stream;
+	offload_work->adjust = adjust_copy;
+
+	queue_work(system_wq, &offload_work->work);
 }
 
 static void dm_vupdate_high_irq(void *interrupt_params)
@@ -473,9 +589,9 @@ static void dm_vupdate_high_irq(void *interrupt_params)
 	acrtc = get_crtc_by_otg_inst(adev, irq_params->irq_src - IRQ_TYPE_VUPDATE);
 
 	if (acrtc) {
-		vrr_active = amdgpu_dm_vrr_active_irq(acrtc);
+		vrr_active = amdgpu_dm_crtc_vrr_active_irq(acrtc);
 		drm_dev = acrtc->base.dev;
-		vblank = &drm_dev->vblank[acrtc->base.index];
+		vblank = drm_crtc_vblank_crtc(&acrtc->base);
 		previous_timestamp = atomic64_read(&irq_params->previous_timestamp);
 		frame_duration_ns = vblank->time - previous_timestamp;
 
@@ -486,9 +602,9 @@ static void dm_vupdate_high_irq(void *interrupt_params)
 			atomic64_set(&irq_params->previous_timestamp, vblank->time);
 		}
 
-		DC_LOG_VBLANK("crtc:%d, vupdate-vrr:%d\n",
-			      acrtc->crtc_id,
-			      vrr_active);
+		drm_dbg_vbl(drm_dev,
+			    "crtc:%d, vupdate-vrr:%d\n", acrtc->crtc_id,
+			    vrr_active);
 
 		/* Core vblank handling is done here after end of front-porch in
 		 * vrr mode, as vblank timestamping will give valid results
@@ -496,22 +612,27 @@ static void dm_vupdate_high_irq(void *interrupt_params)
 		 * page-flip completion events that have been queued to us
 		 * if a pageflip happened inside front-porch.
 		 */
-		if (vrr_active) {
-			drm_crtc_handle_vblank(&acrtc->base);
+		if (vrr_active && acrtc->dm_irq_params.stream) {
+			bool replay_en = acrtc->dm_irq_params.stream->link->replay_settings.replay_feature_enabled;
+			bool psr_en = acrtc->dm_irq_params.stream->link->psr_settings.psr_feature_enabled;
+			bool fs_active_var_en = acrtc->dm_irq_params.freesync_config.state
+				== VRR_STATE_ACTIVE_VARIABLE;
+
+			amdgpu_dm_crtc_handle_vblank(acrtc);
 
 			/* BTR processing for pre-DCE12 ASICs */
-			if (acrtc->dm_irq_params.stream &&
-			    adev->family < AMDGPU_FAMILY_AI) {
+			if (adev->family < AMDGPU_FAMILY_AI) {
 				spin_lock_irqsave(&adev_to_drm(adev)->event_lock, flags);
 				mod_freesync_handle_v_update(
 				    adev->dm.freesync_module,
 				    acrtc->dm_irq_params.stream,
 				    &acrtc->dm_irq_params.vrr_params);
 
-				dc_stream_adjust_vmin_vmax(
-				    adev->dm.dc,
-				    acrtc->dm_irq_params.stream,
-				    &acrtc->dm_irq_params.vrr_params.adjust);
+				if (fs_active_var_en || (!fs_active_var_en && !replay_en && !psr_en)) {
+					schedule_dc_vmin_vmax(adev,
+						acrtc->dm_irq_params.stream,
+						&acrtc->dm_irq_params.vrr_params.adjust);
+				}
 				spin_unlock_irqrestore(&adev_to_drm(adev)->event_lock, flags);
 			}
 		}
@@ -529,6 +650,7 @@ static void dm_crtc_high_irq(void *interrupt_params)
 {
 	struct common_irq_params *irq_params = interrupt_params;
 	struct amdgpu_device *adev = irq_params->adev;
+	struct drm_writeback_job *job;
 	struct amdgpu_crtc *acrtc;
 	unsigned long flags;
 	int vrr_active;
@@ -537,10 +659,39 @@ static void dm_crtc_high_irq(void *interrupt_params)
 	if (!acrtc)
 		return;
 
-	vrr_active = amdgpu_dm_vrr_active_irq(acrtc);
+	if (acrtc->wb_conn) {
+		spin_lock_irqsave(&acrtc->wb_conn->job_lock, flags);
+
+		if (acrtc->wb_pending) {
+			job = list_first_entry_or_null(&acrtc->wb_conn->job_queue,
+						       struct drm_writeback_job,
+						       list_entry);
+			acrtc->wb_pending = false;
+			spin_unlock_irqrestore(&acrtc->wb_conn->job_lock, flags);
+
+			if (job) {
+				unsigned int v_total, refresh_hz;
+				struct dc_stream_state *stream = acrtc->dm_irq_params.stream;
+
+				v_total = stream->adjust.v_total_max ?
+					  stream->adjust.v_total_max : stream->timing.v_total;
+				refresh_hz = div_u64((uint64_t) stream->timing.pix_clk_100hz *
+					     100LL, (v_total * stream->timing.h_total));
+				mdelay(1000 / refresh_hz);
+
+				drm_writeback_signal_completion(acrtc->wb_conn, 0);
+				dc_stream_fc_disable_writeback(adev->dm.dc,
+							       acrtc->dm_irq_params.stream, 0);
+			}
+		} else
+			spin_unlock_irqrestore(&acrtc->wb_conn->job_lock, flags);
+	}
+
+	vrr_active = amdgpu_dm_crtc_vrr_active_irq(acrtc);
 
-	DC_LOG_VBLANK("crtc:%d, vupdate-vrr:%d, planes:%d\n", acrtc->crtc_id,
-		      vrr_active, acrtc->dm_irq_params.active_planes);
+	drm_dbg_vbl(adev_to_drm(adev),
+		    "crtc:%d, vupdate-vrr:%d, planes:%d\n", acrtc->crtc_id,
+		    vrr_active, acrtc->dm_irq_params.active_planes);
 
 	/**
 	 * Core vblank handling at start of front-porch is only possible
@@ -549,7 +700,7 @@ static void dm_crtc_high_irq(void *interrupt_params)
 	 * to dm_vupdate_high_irq after end of front-porch.
 	 */
 	if (!vrr_active)
-		drm_crtc_handle_vblank(&acrtc->base);
+		amdgpu_dm_crtc_handle_vblank(acrtc);
 
 	/**
 	 * Following stuff must happen at start of vblank, for crc
@@ -564,15 +715,20 @@ static void dm_crtc_high_irq(void *interrupt_params)
 	spin_lock_irqsave(&adev_to_drm(adev)->event_lock, flags);
 
 	if (acrtc->dm_irq_params.stream &&
-	    acrtc->dm_irq_params.vrr_params.supported &&
-	    acrtc->dm_irq_params.freesync_config.state ==
-		    VRR_STATE_ACTIVE_VARIABLE) {
+		acrtc->dm_irq_params.vrr_params.supported) {
+		bool replay_en = acrtc->dm_irq_params.stream->link->replay_settings.replay_feature_enabled;
+		bool psr_en = acrtc->dm_irq_params.stream->link->psr_settings.psr_feature_enabled;
+		bool fs_active_var_en = acrtc->dm_irq_params.freesync_config.state == VRR_STATE_ACTIVE_VARIABLE;
+
 		mod_freesync_handle_v_update(adev->dm.freesync_module,
 					     acrtc->dm_irq_params.stream,
 					     &acrtc->dm_irq_params.vrr_params);
 
-		dc_stream_adjust_vmin_vmax(adev->dm.dc, acrtc->dm_irq_params.stream,
-					   &acrtc->dm_irq_params.vrr_params.adjust);
+		/* update vmin_vmax only if freesync is enabled, or only if PSR and REPLAY are disabled */
+		if (fs_active_var_en || (!fs_active_var_en && !replay_en && !psr_en)) {
+			schedule_dc_vmin_vmax(adev, acrtc->dm_irq_params.stream,
+					&acrtc->dm_irq_params.vrr_params.adjust);
+		}
 	}
 
 	/*
@@ -599,7 +755,6 @@ static void dm_crtc_high_irq(void *interrupt_params)
 	spin_unlock_irqrestore(&adev_to_drm(adev)->event_lock, flags);
 }
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 #if defined(CONFIG_DRM_AMD_SECURE_DISPLAY)
 /**
  * dm_dcn_vertical_interrupt0_high_irq() - Handles OTG Vertical interrupt0 for
@@ -624,7 +779,7 @@ static void dm_dcn_vertical_interrupt0_high_irq(void *interrupt_params)
 #endif /* CONFIG_DRM_AMD_SECURE_DISPLAY */
 
 /**
- * dmub_aux_setconfig_reply_callback - Callback for AUX or SET_CONFIG command.
+ * dmub_aux_setconfig_callback - Callback for AUX or SET_CONFIG command.
  * @adev: amdgpu_device pointer
  * @notify: dmub notification structure
  *
@@ -632,7 +787,8 @@ static void dm_dcn_vertical_interrupt0_high_irq(void *interrupt_params)
  * Copies dmub notification to DM which is to be read by AUX command.
  * issuing thread and also signals the event to wake up the thread.
  */
-void dmub_aux_setconfig_callback(struct amdgpu_device *adev, struct dmub_notification *notify)
+static void dmub_aux_setconfig_callback(struct amdgpu_device *adev,
+					struct dmub_notification *notify)
 {
 	if (adev->dm.dmub_notify)
 		memcpy(adev->dm.dmub_notify, notify, sizeof(struct dmub_notification));
@@ -640,6 +796,29 @@ void dmub_aux_setconfig_callback(struct amdgpu_device *adev, struct dmub_notific
 		complete(&adev->dm.dmub_aux_transfer_done);
 }
 
+static void dmub_aux_fused_io_callback(struct amdgpu_device *adev,
+					struct dmub_notification *notify)
+{
+	if (!adev || !notify) {
+		ASSERT(false);
+		return;
+	}
+
+	const struct dmub_cmd_fused_request *req = &notify->fused_request;
+	const uint8_t ddc_line = req->u.aux.ddc_line;
+
+	if (ddc_line >= ARRAY_SIZE(adev->dm.fused_io)) {
+		ASSERT(false);
+		return;
+	}
+
+	struct fused_io_sync *sync = &adev->dm.fused_io[ddc_line];
+
+	static_assert(sizeof(*req) <= sizeof(sync->reply_data), "Size mismatch");
+	memcpy(sync->reply_data, req, sizeof(*req));
+	complete(&sync->replied);
+}
+
 /**
  * dmub_hpd_callback - DMUB HPD interrupt processing callback.
  * @adev: amdgpu_device pointer
@@ -648,37 +827,56 @@ void dmub_aux_setconfig_callback(struct amdgpu_device *adev, struct dmub_notific
  * Dmub Hpd interrupt processing callback. Gets displayindex through the
  * ink index and calls helper to do the processing.
  */
-void dmub_hpd_callback(struct amdgpu_device *adev, struct dmub_notification *notify)
+static void dmub_hpd_callback(struct amdgpu_device *adev,
+			      struct dmub_notification *notify)
 {
 	struct amdgpu_dm_connector *aconnector;
 	struct amdgpu_dm_connector *hpd_aconnector = NULL;
 	struct drm_connector *connector;
 	struct drm_connector_list_iter iter;
 	struct dc_link *link;
-	uint8_t link_index = 0;
-	struct drm_device *dev = adev->dm.ddev;
+	u8 link_index = 0;
+	struct drm_device *dev;
 
 	if (adev == NULL)
 		return;
 
 	if (notify == NULL) {
-		DRM_ERROR("DMUB HPD callback notification was NULL");
+		drm_err(adev_to_drm(adev), "DMUB HPD callback notification was NULL");
 		return;
 	}
 
 	if (notify->link_index > adev->dm.dc->link_count) {
-		DRM_ERROR("DMUB HPD index (%u)is abnormal", notify->link_index);
+		drm_err(adev_to_drm(adev), "DMUB HPD index (%u)is abnormal", notify->link_index);
+		return;
+	}
+
+	/* Skip DMUB HPD IRQ in suspend/resume. We will probe them later. */
+	if (notify->type == DMUB_NOTIFICATION_HPD && adev->in_suspend) {
+		drm_info(adev_to_drm(adev), "Skip DMUB HPD IRQ callback in suspend/resume\n");
 		return;
 	}
 
 	link_index = notify->link_index;
 	link = adev->dm.dc->links[link_index];
+	dev = adev->dm.ddev;
 
 	drm_connector_list_iter_begin(dev, &iter);
 	drm_for_each_connector_iter(connector, &iter) {
+
+		if (connector->connector_type == DRM_MODE_CONNECTOR_WRITEBACK)
+			continue;
+
 		aconnector = to_amdgpu_dm_connector(connector);
 		if (link && aconnector->dc_link == link) {
-			DRM_INFO("DMUB HPD callback: link_index=%u\n", link_index);
+			if (notify->type == DMUB_NOTIFICATION_HPD)
+				drm_info(adev_to_drm(adev), "DMUB HPD IRQ callback: link_index=%u\n", link_index);
+			else if (notify->type == DMUB_NOTIFICATION_HPD_IRQ)
+				drm_info(adev_to_drm(adev), "DMUB HPD RX IRQ callback: link_index=%u\n", link_index);
+			else
+				drm_warn(adev_to_drm(adev), "DMUB Unknown HPD callback type %d, link_index=%u\n",
+						notify->type, link_index);
+
 			hpd_aconnector = aconnector;
 			break;
 		}
@@ -686,14 +884,31 @@ void dmub_hpd_callback(struct amdgpu_device *adev, struct dmub_notification *not
 	drm_connector_list_iter_end(&iter);
 
 	if (hpd_aconnector) {
-		if (notify->type == DMUB_NOTIFICATION_HPD)
+		if (notify->type == DMUB_NOTIFICATION_HPD) {
+			if (hpd_aconnector->dc_link->hpd_status == (notify->hpd_status == DP_HPD_PLUG))
+				drm_warn(adev_to_drm(adev), "DMUB reported hpd status unchanged. link_index=%u\n", link_index);
 			handle_hpd_irq_helper(hpd_aconnector);
-		else if (notify->type == DMUB_NOTIFICATION_HPD_IRQ)
+		} else if (notify->type == DMUB_NOTIFICATION_HPD_IRQ) {
 			handle_hpd_rx_irq(hpd_aconnector);
+		}
 	}
 }
 
 /**
+ * dmub_hpd_sense_callback - DMUB HPD sense processing callback.
+ * @adev: amdgpu_device pointer
+ * @notify: dmub notification structure
+ *
+ * HPD sense changes can occur during low power states and need to be
+ * notified from firmware to driver.
+ */
+static void dmub_hpd_sense_callback(struct amdgpu_device *adev,
+			      struct dmub_notification *notify)
+{
+	drm_dbg_driver(adev_to_drm(adev), "DMUB HPD SENSE callback.\n");
+}
+
+/**
  * register_dmub_notify_callback - Sets callback for DMUB notify
  * @adev: amdgpu_device pointer
  * @type: Type of dmub notification
@@ -705,8 +920,10 @@ void dmub_hpd_callback(struct amdgpu_device *adev, struct dmub_notification *not
  * to dmub interrupt handling thread
  * Return: true if successfully registered, false if there is existing registration
  */
-bool register_dmub_notify_callback(struct amdgpu_device *adev, enum dmub_notification_type type,
-dmub_notify_interrupt_callback_t callback, bool dmub_int_thread_offload)
+static bool register_dmub_notify_callback(struct amdgpu_device *adev,
+					  enum dmub_notification_type type,
+					  dmub_notify_interrupt_callback_t callback,
+					  bool dmub_int_thread_offload)
 {
 	if (callback != NULL && type < ARRAY_SIZE(adev->dm.dmub_thread_offload)) {
 		adev->dm.dmub_callback[type] = callback;
@@ -724,7 +941,7 @@ static void dm_handle_hpd_work(struct work_struct *work)
 	dmub_hpd_wrk = container_of(work, struct dmub_hpd_work, handle_hpd_work);
 
 	if (!dmub_hpd_wrk->dmub_notify) {
-		DRM_ERROR("dmub_hpd_wrk dmub_notify is NULL");
+		drm_err(adev_to_drm(dmub_hpd_wrk->adev), "dmub_hpd_wrk dmub_notify is NULL");
 		return;
 	}
 
@@ -738,6 +955,30 @@ static void dm_handle_hpd_work(struct work_struct *work)
 
 }
 
+static const char *dmub_notification_type_str(enum dmub_notification_type e)
+{
+	switch (e) {
+	case DMUB_NOTIFICATION_NO_DATA:
+		return "NO_DATA";
+	case DMUB_NOTIFICATION_AUX_REPLY:
+		return "AUX_REPLY";
+	case DMUB_NOTIFICATION_HPD:
+		return "HPD";
+	case DMUB_NOTIFICATION_HPD_IRQ:
+		return "HPD_IRQ";
+	case DMUB_NOTIFICATION_SET_CONFIG_REPLY:
+		return "SET_CONFIG_REPLY";
+	case DMUB_NOTIFICATION_DPIA_NOTIFICATION:
+		return "DPIA_NOTIFICATION";
+	case DMUB_NOTIFICATION_HPD_SENSE_NOTIFY:
+		return "HPD_SENSE_NOTIFY";
+	case DMUB_NOTIFICATION_FUSED_IO:
+		return "FUSED_IO";
+	default:
+		return "<unknown>";
+	}
+}
+
 #define DMUB_TRACE_MAX_READ 64
 /**
  * dm_dmub_outbox1_low_irq() - Handles Outbox interrupt
@@ -748,99 +989,87 @@ static void dm_handle_hpd_work(struct work_struct *work)
  */
 static void dm_dmub_outbox1_low_irq(void *interrupt_params)
 {
-	struct dmub_notification notify;
+	struct dmub_notification notify = {0};
 	struct common_irq_params *irq_params = interrupt_params;
 	struct amdgpu_device *adev = irq_params->adev;
 	struct amdgpu_display_manager *dm = &adev->dm;
 	struct dmcub_trace_buf_entry entry = { 0 };
-	uint32_t count = 0;
+	u32 count = 0;
 	struct dmub_hpd_work *dmub_hpd_wrk;
-	struct dc_link *plink = NULL;
+
+	do {
+		if (dc_dmub_srv_get_dmub_outbox0_msg(dm->dc, &entry)) {
+			trace_amdgpu_dmub_trace_high_irq(entry.trace_code, entry.tick_count,
+							entry.param0, entry.param1);
+
+			drm_dbg_driver(adev_to_drm(adev), "trace_code:%u, tick_count:%u, param0:%u, param1:%u\n",
+				 entry.trace_code, entry.tick_count, entry.param0, entry.param1);
+		} else
+			break;
+
+		count++;
+
+	} while (count <= DMUB_TRACE_MAX_READ);
+
+	if (count > DMUB_TRACE_MAX_READ)
+		drm_dbg_driver(adev_to_drm(adev), "Warning : count > DMUB_TRACE_MAX_READ");
 
 	if (dc_enable_dmub_notifications(adev->dm.dc) &&
 		irq_params->irq_src == DC_IRQ_SOURCE_DMCUB_OUTBOX) {
 
 		do {
 			dc_stat_get_dmub_notification(adev->dm.dc, &notify);
-			if (notify.type > ARRAY_SIZE(dm->dmub_thread_offload)) {
-				DRM_ERROR("DM: notify type %d invalid!", notify.type);
+			if (notify.type >= ARRAY_SIZE(dm->dmub_thread_offload)) {
+				drm_err(adev_to_drm(adev), "DM: notify type %d invalid!", notify.type);
 				continue;
 			}
 			if (!dm->dmub_callback[notify.type]) {
-				DRM_DEBUG_DRIVER("DMUB notification skipped, no handler: type=%d\n", notify.type);
+				drm_warn(adev_to_drm(adev), "DMUB notification skipped due to no handler: type=%s\n",
+					dmub_notification_type_str(notify.type));
 				continue;
 			}
 			if (dm->dmub_thread_offload[notify.type] == true) {
 				dmub_hpd_wrk = kzalloc(sizeof(*dmub_hpd_wrk), GFP_ATOMIC);
 				if (!dmub_hpd_wrk) {
-					DRM_ERROR("Failed to allocate dmub_hpd_wrk");
+					drm_err(adev_to_drm(adev), "Failed to allocate dmub_hpd_wrk");
 					return;
 				}
-				dmub_hpd_wrk->dmub_notify = kzalloc(sizeof(struct dmub_notification), GFP_ATOMIC);
+				dmub_hpd_wrk->dmub_notify = kmemdup(&notify, sizeof(struct dmub_notification),
+								    GFP_ATOMIC);
 				if (!dmub_hpd_wrk->dmub_notify) {
 					kfree(dmub_hpd_wrk);
-					DRM_ERROR("Failed to allocate dmub_hpd_wrk->dmub_notify");
+					drm_err(adev_to_drm(adev), "Failed to allocate dmub_hpd_wrk->dmub_notify");
 					return;
 				}
 				INIT_WORK(&dmub_hpd_wrk->handle_hpd_work, dm_handle_hpd_work);
-				if (dmub_hpd_wrk->dmub_notify)
-					memcpy(dmub_hpd_wrk->dmub_notify, &notify, sizeof(struct dmub_notification));
 				dmub_hpd_wrk->adev = adev;
-				if (notify.type == DMUB_NOTIFICATION_HPD) {
-					plink = adev->dm.dc->links[notify.link_index];
-					if (plink) {
-						plink->hpd_status =
-							notify.hpd_status ==
-							DP_HPD_PLUG ? true : false;
-					}
-				}
 				queue_work(adev->dm.delayed_hpd_wq, &dmub_hpd_wrk->handle_hpd_work);
 			} else {
 				dm->dmub_callback[notify.type](adev, &notify);
 			}
 		} while (notify.pending_notification);
 	}
-
-
-	do {
-		if (dc_dmub_srv_get_dmub_outbox0_msg(dm->dc, &entry)) {
-			trace_amdgpu_dmub_trace_high_irq(entry.trace_code, entry.tick_count,
-							entry.param0, entry.param1);
-
-			DRM_DEBUG_DRIVER("trace_code:%u, tick_count:%u, param0:%u, param1:%u\n",
-				 entry.trace_code, entry.tick_count, entry.param0, entry.param1);
-		} else
-			break;
-
-		count++;
-
-	} while (count <= DMUB_TRACE_MAX_READ);
-
-	if (count > DMUB_TRACE_MAX_READ)
-		DRM_DEBUG_DRIVER("Warning : count > DMUB_TRACE_MAX_READ");
 }
-#endif /* CONFIG_DRM_AMD_DC_DCN */
 
-static int dm_set_clockgating_state(void *handle,
+static int dm_set_clockgating_state(struct amdgpu_ip_block *ip_block,
 		  enum amd_clockgating_state state)
 {
 	return 0;
 }
 
-static int dm_set_powergating_state(void *handle,
+static int dm_set_powergating_state(struct amdgpu_ip_block *ip_block,
 		  enum amd_powergating_state state)
 {
 	return 0;
 }
 
 /* Prototypes of private functions */
-static int dm_early_init(void* handle);
+static int dm_early_init(struct amdgpu_ip_block *ip_block);
 
 /* Allocate memory for FBC compressed data  */
 static void amdgpu_dm_fbc_init(struct drm_connector *connector)
 {
-	struct drm_device *dev = connector->dev;
-	struct amdgpu_device *adev = drm_to_adev(dev);
+	struct amdgpu_device *adev = drm_to_adev(connector->dev);
 	struct dm_compressor_info *compressor = &adev->dm.compressor;
 	struct amdgpu_dm_connector *aconn = to_amdgpu_dm_connector(connector);
 	struct drm_display_mode *mode;
@@ -857,8 +1086,8 @@ static void amdgpu_dm_fbc_init(struct drm_connector *connector)
 
 
 	list_for_each_entry(mode, &connector->modes, head) {
-		if (max_size < mode->htotal * mode->vtotal)
-			max_size = mode->htotal * mode->vtotal;
+		if (max_size < (unsigned long) mode->htotal * mode->vtotal)
+			max_size = (unsigned long) mode->htotal * mode->vtotal;
 	}
 
 	if (max_size) {
@@ -867,10 +1096,10 @@ static void amdgpu_dm_fbc_init(struct drm_connector *connector)
 			    &compressor->gpu_addr, &compressor->cpu_addr);
 
 		if (r)
-			DRM_ERROR("DM: Failed to initialize FBC\n");
+			drm_err(adev_to_drm(adev), "DM: Failed to initialize FBC\n");
 		else {
 			adev->dm.dc->ctx->fbc_gpu_addr = compressor->gpu_addr;
-			DRM_INFO("DM: FBC alloc %lu\n", max_size*4);
+			drm_info(adev_to_drm(adev), "DM: FBC alloc %lu\n", max_size*4);
 		}
 
 	}
@@ -894,13 +1123,19 @@ static int amdgpu_dm_audio_component_get_eld(struct device *kdev, int port,
 
 	drm_connector_list_iter_begin(dev, &conn_iter);
 	drm_for_each_connector_iter(connector, &conn_iter) {
+
+		if (connector->connector_type == DRM_MODE_CONNECTOR_WRITEBACK)
+			continue;
+
 		aconnector = to_amdgpu_dm_connector(connector);
 		if (aconnector->audio_inst != port)
 			continue;
 
 		*enabled = true;
+		mutex_lock(&connector->eld_mutex);
 		ret = drm_eld_size(connector->eld);
 		memcpy(buf, connector->eld, min(max_bytes, ret));
+		mutex_unlock(&connector->eld_mutex);
 
 		break;
 	}
@@ -934,8 +1169,7 @@ static int amdgpu_dm_audio_component_bind(struct device *kdev,
 static void amdgpu_dm_audio_component_unbind(struct device *kdev,
 					  struct device *hda_kdev, void *data)
 {
-	struct drm_device *dev = dev_get_drvdata(kdev);
-	struct amdgpu_device *adev = drm_to_adev(dev);
+	struct amdgpu_device *adev = drm_to_adev(dev_get_drvdata(kdev));
 	struct drm_audio_component *acomp = data;
 
 	acomp->ops = NULL;
@@ -1018,43 +1252,47 @@ static int dm_dmub_hw_init(struct amdgpu_device *adev)
 	const struct firmware *dmub_fw = adev->dm.dmub_fw;
 	struct dmcu *dmcu = adev->dm.dc->res_pool->dmcu;
 	struct abm *abm = adev->dm.dc->res_pool->abm;
+	struct dc_context *ctx = adev->dm.dc->ctx;
 	struct dmub_srv_hw_params hw_params;
 	enum dmub_status status;
 	const unsigned char *fw_inst_const, *fw_bss_data;
-	uint32_t i, fw_inst_const_size, fw_bss_data_size;
+	u32 i, fw_inst_const_size, fw_bss_data_size;
 	bool has_hw_support;
-	struct dc *dc = adev->dm.dc;
 
 	if (!dmub_srv)
 		/* DMUB isn't supported on the ASIC. */
 		return 0;
 
 	if (!fb_info) {
-		DRM_ERROR("No framebuffer info for DMUB service.\n");
+		drm_err(adev_to_drm(adev), "No framebuffer info for DMUB service.\n");
 		return -EINVAL;
 	}
 
 	if (!dmub_fw) {
 		/* Firmware required for DMUB support. */
-		DRM_ERROR("No firmware provided for DMUB.\n");
+		drm_err(adev_to_drm(adev), "No firmware provided for DMUB.\n");
 		return -EINVAL;
 	}
 
+	/* initialize register offsets for ASICs with runtime initialization available */
+	if (dmub_srv->hw_funcs.init_reg_offsets)
+		dmub_srv->hw_funcs.init_reg_offsets(dmub_srv, ctx);
+
 	status = dmub_srv_has_hw_support(dmub_srv, &has_hw_support);
 	if (status != DMUB_STATUS_OK) {
-		DRM_ERROR("Error checking HW support for DMUB: %d\n", status);
+		drm_err(adev_to_drm(adev), "Error checking HW support for DMUB: %d\n", status);
 		return -EINVAL;
 	}
 
 	if (!has_hw_support) {
-		DRM_INFO("DMUB unsupported on ASIC\n");
+		drm_info(adev_to_drm(adev), "DMUB unsupported on ASIC\n");
 		return 0;
 	}
 
 	/* Reset DMCUB if it was previously running - before we overwrite its memory. */
 	status = dmub_srv_hw_reset(dmub_srv);
 	if (status != DMUB_STATUS_OK)
-		DRM_WARN("Error resetting DMUB HW: %d\n", status);
+		drm_warn(adev_to_drm(adev), "Error resetting DMUB HW: %d\n", status);
 
 	hdr = (const struct dmcub_firmware_header_v1_0 *)dmub_fw->data;
 
@@ -1100,10 +1338,13 @@ static int dm_dmub_hw_init(struct amdgpu_device *adev)
 	memset(fb_info->fb[DMUB_WINDOW_6_FW_STATE].cpu_addr, 0,
 	       fb_info->fb[DMUB_WINDOW_6_FW_STATE].size);
 
+	memset(fb_info->fb[DMUB_WINDOW_SHARED_STATE].cpu_addr, 0,
+	       fb_info->fb[DMUB_WINDOW_SHARED_STATE].size);
+
 	/* Initialize hardware. */
 	memset(&hw_params, 0, sizeof(hw_params));
 	hw_params.fb_base = adev->gmc.fb_start;
-	hw_params.fb_offset = adev->gmc.aper_base;
+	hw_params.fb_offset = adev->vm_manager.vram_base_offset;
 
 	/* backdoor load firmware and trigger dmub running */
 	if (adev->firmware.load_type != AMDGPU_FW_LOAD_PSP)
@@ -1115,14 +1356,26 @@ static int dm_dmub_hw_init(struct amdgpu_device *adev)
 	for (i = 0; i < fb_info->num_fb; ++i)
 		hw_params.fb[i] = &fb_info->fb[i];
 
-	switch (adev->asic_type) {
-	case CHIP_YELLOW_CARP:
-		if (dc->ctx->asic_id.hw_internal_rev != YELLOW_CARP_A0) {
-			hw_params.dpia_supported = true;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-			hw_params.disable_dpia = dc->debug.dpia_debug.bits.disable_dpia;
-#endif
-		}
+	switch (amdgpu_ip_version(adev, DCE_HWIP, 0)) {
+	case IP_VERSION(3, 1, 3):
+	case IP_VERSION(3, 1, 4):
+	case IP_VERSION(3, 5, 0):
+	case IP_VERSION(3, 5, 1):
+	case IP_VERSION(3, 6, 0):
+	case IP_VERSION(4, 0, 1):
+		hw_params.dpia_supported = true;
+		hw_params.disable_dpia = adev->dm.dc->debug.dpia_debug.bits.disable_dpia;
+		break;
+	default:
+		break;
+	}
+
+	switch (amdgpu_ip_version(adev, DCE_HWIP, 0)) {
+	case IP_VERSION(3, 5, 0):
+	case IP_VERSION(3, 5, 1):
+	case IP_VERSION(3, 6, 0):
+		hw_params.ips_sequential_ono = adev->external_rev_id > 0x10;
+		hw_params.lower_hbr3_phy_ssc = true;
 		break;
 	default:
 		break;
@@ -1130,14 +1383,14 @@ static int dm_dmub_hw_init(struct amdgpu_device *adev)
 
 	status = dmub_srv_hw_init(dmub_srv, &hw_params);
 	if (status != DMUB_STATUS_OK) {
-		DRM_ERROR("Error initializing DMUB HW: %d\n", status);
+		drm_err(adev_to_drm(adev), "Error initializing DMUB HW: %d\n", status);
 		return -EINVAL;
 	}
 
 	/* Wait for firmware load to finish. */
 	status = dmub_srv_wait_for_auto_load(dmub_srv, 100000);
 	if (status != DMUB_STATUS_OK)
-		DRM_WARN("Wait for DMUB auto-load failed: %d\n", status);
+		drm_warn(adev_to_drm(adev), "Wait for DMUB auto-load failed: %d\n", status);
 
 	/* Init DMCU and ABM if available. */
 	if (dmcu && abm) {
@@ -1148,111 +1401,160 @@ static int dm_dmub_hw_init(struct amdgpu_device *adev)
 	if (!adev->dm.dc->ctx->dmub_srv)
 		adev->dm.dc->ctx->dmub_srv = dc_dmub_srv_create(adev->dm.dc, dmub_srv);
 	if (!adev->dm.dc->ctx->dmub_srv) {
-		DRM_ERROR("Couldn't allocate DC DMUB server!\n");
+		drm_err(adev_to_drm(adev), "Couldn't allocate DC DMUB server!\n");
 		return -ENOMEM;
 	}
 
-	DRM_INFO("DMUB hardware initialized: version=0x%08X\n",
+	drm_info(adev_to_drm(adev), "DMUB hardware initialized: version=0x%08X\n",
 		 adev->dm.dmcub_fw_version);
 
+	/* Keeping sanity checks off if
+	 * DCN31 >= 4.0.59.0
+	 * DCN314 >= 8.0.16.0
+	 * Otherwise, turn on sanity checks
+	 */
+	switch (amdgpu_ip_version(adev, DCE_HWIP, 0)) {
+	case IP_VERSION(3, 1, 2):
+	case IP_VERSION(3, 1, 3):
+		if (adev->dm.dmcub_fw_version &&
+			adev->dm.dmcub_fw_version >= DMUB_FW_VERSION(4, 0, 0) &&
+			adev->dm.dmcub_fw_version < DMUB_FW_VERSION(4, 0, 59))
+				adev->dm.dc->debug.sanity_checks = true;
+		break;
+	case IP_VERSION(3, 1, 4):
+		if (adev->dm.dmcub_fw_version &&
+			adev->dm.dmcub_fw_version >= DMUB_FW_VERSION(4, 0, 0) &&
+			adev->dm.dmcub_fw_version < DMUB_FW_VERSION(8, 0, 16))
+				adev->dm.dc->debug.sanity_checks = true;
+		break;
+	default:
+		break;
+	}
+
 	return 0;
 }
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
+static void dm_dmub_hw_resume(struct amdgpu_device *adev)
+{
+	struct dmub_srv *dmub_srv = adev->dm.dmub_srv;
+	enum dmub_status status;
+	bool init;
+	int r;
+
+	if (!dmub_srv) {
+		/* DMUB isn't supported on the ASIC. */
+		return;
+	}
+
+	status = dmub_srv_is_hw_init(dmub_srv, &init);
+	if (status != DMUB_STATUS_OK)
+		drm_warn(adev_to_drm(adev), "DMUB hardware init check failed: %d\n", status);
+
+	if (status == DMUB_STATUS_OK && init) {
+		/* Wait for firmware load to finish. */
+		status = dmub_srv_wait_for_auto_load(dmub_srv, 100000);
+		if (status != DMUB_STATUS_OK)
+			drm_warn(adev_to_drm(adev), "Wait for DMUB auto-load failed: %d\n", status);
+	} else {
+		/* Perform the full hardware initialization. */
+		r = dm_dmub_hw_init(adev);
+		if (r)
+			drm_err(adev_to_drm(adev), "DMUB interface failed to initialize: status=%d\n", r);
+	}
+}
+
 static void mmhub_read_system_context(struct amdgpu_device *adev, struct dc_phy_addr_space_config *pa_config)
 {
-	uint64_t pt_base;
-	uint32_t logical_addr_low;
-	uint32_t logical_addr_high;
-	uint32_t agp_base, agp_bot, agp_top;
+	u64 pt_base;
+	u32 logical_addr_low;
+	u32 logical_addr_high;
+	u32 agp_base, agp_bot, agp_top;
 	PHYSICAL_ADDRESS_LOC page_table_start, page_table_end, page_table_base;
 
 	memset(pa_config, 0, sizeof(*pa_config));
 
-	logical_addr_low  = min(adev->gmc.fb_start, adev->gmc.agp_start) >> 18;
-	pt_base = amdgpu_gmc_pd_addr(adev->gart.bo);
-
-	if (adev->apu_flags & AMD_APU_IS_RAVEN2)
-		/*
-		 * Raven2 has a HW issue that it is unable to use the vram which
-		 * is out of MC_VM_SYSTEM_APERTURE_HIGH_ADDR. So here is the
-		 * workaround that increase system aperture high address (add 1)
-		 * to get rid of the VM fault and hardware hang.
-		 */
-		logical_addr_high = max((adev->gmc.fb_end >> 18) + 0x1, adev->gmc.agp_end >> 18);
-	else
-		logical_addr_high = max(adev->gmc.fb_end, adev->gmc.agp_end) >> 18;
-
 	agp_base = 0;
 	agp_bot = adev->gmc.agp_start >> 24;
 	agp_top = adev->gmc.agp_end >> 24;
 
+	/* AGP aperture is disabled */
+	if (agp_bot > agp_top) {
+		logical_addr_low = adev->gmc.fb_start >> 18;
+		if (adev->apu_flags & (AMD_APU_IS_RAVEN2 |
+				       AMD_APU_IS_RENOIR |
+				       AMD_APU_IS_GREEN_SARDINE))
+			/*
+			 * Raven2 has a HW issue that it is unable to use the vram which
+			 * is out of MC_VM_SYSTEM_APERTURE_HIGH_ADDR. So here is the
+			 * workaround that increase system aperture high address (add 1)
+			 * to get rid of the VM fault and hardware hang.
+			 */
+			logical_addr_high = (adev->gmc.fb_end >> 18) + 0x1;
+		else
+			logical_addr_high = adev->gmc.fb_end >> 18;
+	} else {
+		logical_addr_low = min(adev->gmc.fb_start, adev->gmc.agp_start) >> 18;
+		if (adev->apu_flags & (AMD_APU_IS_RAVEN2 |
+				       AMD_APU_IS_RENOIR |
+				       AMD_APU_IS_GREEN_SARDINE))
+			/*
+			 * Raven2 has a HW issue that it is unable to use the vram which
+			 * is out of MC_VM_SYSTEM_APERTURE_HIGH_ADDR. So here is the
+			 * workaround that increase system aperture high address (add 1)
+			 * to get rid of the VM fault and hardware hang.
+			 */
+			logical_addr_high = max((adev->gmc.fb_end >> 18) + 0x1, adev->gmc.agp_end >> 18);
+		else
+			logical_addr_high = max(adev->gmc.fb_end, adev->gmc.agp_end) >> 18;
+	}
+
+	pt_base = amdgpu_gmc_pd_addr(adev->gart.bo);
 
-	page_table_start.high_part = (u32)(adev->gmc.gart_start >> 44) & 0xF;
-	page_table_start.low_part = (u32)(adev->gmc.gart_start >> 12);
-	page_table_end.high_part = (u32)(adev->gmc.gart_end >> 44) & 0xF;
-	page_table_end.low_part = (u32)(adev->gmc.gart_end >> 12);
-	page_table_base.high_part = upper_32_bits(pt_base) & 0xF;
+	page_table_start.high_part = upper_32_bits(adev->gmc.gart_start >>
+						   AMDGPU_GPU_PAGE_SHIFT);
+	page_table_start.low_part = lower_32_bits(adev->gmc.gart_start >>
+						  AMDGPU_GPU_PAGE_SHIFT);
+	page_table_end.high_part = upper_32_bits(adev->gmc.gart_end >>
+						 AMDGPU_GPU_PAGE_SHIFT);
+	page_table_end.low_part = lower_32_bits(adev->gmc.gart_end >>
+						AMDGPU_GPU_PAGE_SHIFT);
+	page_table_base.high_part = upper_32_bits(pt_base);
 	page_table_base.low_part = lower_32_bits(pt_base);
 
 	pa_config->system_aperture.start_addr = (uint64_t)logical_addr_low << 18;
 	pa_config->system_aperture.end_addr = (uint64_t)logical_addr_high << 18;
 
-	pa_config->system_aperture.agp_base = (uint64_t)agp_base << 24 ;
+	pa_config->system_aperture.agp_base = (uint64_t)agp_base << 24;
 	pa_config->system_aperture.agp_bot = (uint64_t)agp_bot << 24;
 	pa_config->system_aperture.agp_top = (uint64_t)agp_top << 24;
 
 	pa_config->system_aperture.fb_base = adev->gmc.fb_start;
-	pa_config->system_aperture.fb_offset = adev->gmc.aper_base;
+	pa_config->system_aperture.fb_offset = adev->vm_manager.vram_base_offset;
 	pa_config->system_aperture.fb_top = adev->gmc.fb_end;
 
 	pa_config->gart_config.page_table_start_addr = page_table_start.quad_part << 12;
 	pa_config->gart_config.page_table_end_addr = page_table_end.quad_part << 12;
 	pa_config->gart_config.page_table_base_addr = page_table_base.quad_part;
 
-	pa_config->is_hvm_enabled = 0;
+	pa_config->is_hvm_enabled = adev->mode_info.gpu_vm_support;
 
 }
-#endif
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-static void vblank_control_worker(struct work_struct *work)
-{
-	struct vblank_control_work *vblank_work =
-		container_of(work, struct vblank_control_work, work);
-	struct amdgpu_display_manager *dm = vblank_work->dm;
-
-	mutex_lock(&dm->dc_lock);
-
-	if (vblank_work->enable)
-		dm->active_vblank_irq_count++;
-	else if(dm->active_vblank_irq_count)
-		dm->active_vblank_irq_count--;
 
-	dc_allow_idle_optimizations(dm->dc, dm->active_vblank_irq_count == 0);
-
-	DRM_DEBUG_KMS("Allow idle optimizations (MALL): %d\n", dm->active_vblank_irq_count == 0);
-
-	/* Control PSR based on vblank requirements from OS */
-	if (vblank_work->stream && vblank_work->stream->link) {
-		if (vblank_work->enable) {
-			if (vblank_work->stream->link->psr_settings.psr_allow_active)
-				amdgpu_dm_psr_disable(vblank_work->stream);
-		} else if (vblank_work->stream->link->psr_settings.psr_feature_enabled &&
-			   !vblank_work->stream->link->psr_settings.psr_allow_active &&
-			   vblank_work->acrtc->dm_irq_params.allow_psr_entry) {
-			amdgpu_dm_psr_enable(vblank_work->stream);
-		}
-	}
-
-	mutex_unlock(&dm->dc_lock);
+static void force_connector_state(
+	struct amdgpu_dm_connector *aconnector,
+	enum drm_connector_force force_state)
+{
+	struct drm_connector *connector = &aconnector->base;
 
-	dc_stream_release(vblank_work->stream);
+	mutex_lock(&connector->dev->mode_config.mutex);
+	aconnector->base.force = force_state;
+	mutex_unlock(&connector->dev->mode_config.mutex);
 
-	kfree(vblank_work);
+	mutex_lock(&aconnector->hpd_lock);
+	drm_kms_helper_connector_hotplug_event(connector);
+	mutex_unlock(&aconnector->hpd_lock);
 }
 
-#endif
-
 static void dm_handle_hpd_rx_offload_work(struct work_struct *work)
 {
 	struct hpd_rx_irq_offload_work *offload_work;
@@ -1261,21 +1563,24 @@ static void dm_handle_hpd_rx_offload_work(struct work_struct *work)
 	struct amdgpu_device *adev;
 	enum dc_connection_type new_connection_type = dc_connection_none;
 	unsigned long flags;
+	union test_response test_response;
+
+	memset(&test_response, 0, sizeof(test_response));
 
 	offload_work = container_of(work, struct hpd_rx_irq_offload_work, work);
 	aconnector = offload_work->offload_wq->aconnector;
+	adev = offload_work->adev;
 
 	if (!aconnector) {
-		DRM_ERROR("Can't retrieve aconnector in hpd_rx_irq_offload_work");
+		drm_err(adev_to_drm(adev), "Can't retrieve aconnector in hpd_rx_irq_offload_work");
 		goto skip;
 	}
 
-	adev = drm_to_adev(aconnector->base.dev);
 	dc_link = aconnector->dc_link;
 
 	mutex_lock(&aconnector->hpd_lock);
-	if (!dc_link_detect_sink(dc_link, &new_connection_type))
-		DRM_ERROR("KMS: Failed to detect connector\n");
+	if (!dc_link_detect_connection_type(dc_link, &new_connection_type))
+		drm_err(adev_to_drm(adev), "KMS: Failed to detect connector\n");
 	mutex_unlock(&aconnector->hpd_lock);
 
 	if (new_connection_type == dc_connection_none)
@@ -1284,16 +1589,58 @@ static void dm_handle_hpd_rx_offload_work(struct work_struct *work)
 	if (amdgpu_in_reset(adev))
 		goto skip;
 
+	if (offload_work->data.bytes.device_service_irq.bits.UP_REQ_MSG_RDY ||
+		offload_work->data.bytes.device_service_irq.bits.DOWN_REP_MSG_RDY) {
+		dm_handle_mst_sideband_msg_ready_event(&aconnector->mst_mgr, DOWN_OR_UP_MSG_RDY_EVENT);
+		spin_lock_irqsave(&offload_work->offload_wq->offload_lock, flags);
+		offload_work->offload_wq->is_handling_mst_msg_rdy_event = false;
+		spin_unlock_irqrestore(&offload_work->offload_wq->offload_lock, flags);
+		goto skip;
+	}
+
 	mutex_lock(&adev->dm.dc_lock);
-	if (offload_work->data.bytes.device_service_irq.bits.AUTOMATED_TEST)
+	if (offload_work->data.bytes.device_service_irq.bits.AUTOMATED_TEST) {
 		dc_link_dp_handle_automated_test(dc_link);
-	else if ((dc_link->connector_signal != SIGNAL_TYPE_EDP) &&
-			hpd_rx_irq_check_link_loss_status(dc_link, &offload_work->data) &&
+
+		if (aconnector->timing_changed) {
+			/* force connector disconnect and reconnect */
+			force_connector_state(aconnector, DRM_FORCE_OFF);
+			msleep(100);
+			force_connector_state(aconnector, DRM_FORCE_UNSPECIFIED);
+		}
+
+		test_response.bits.ACK = 1;
+
+		core_link_write_dpcd(
+		dc_link,
+		DP_TEST_RESPONSE,
+		&test_response.raw,
+		sizeof(test_response));
+	} else if ((dc_link->connector_signal != SIGNAL_TYPE_EDP) &&
+			dc_link_check_link_loss_status(dc_link, &offload_work->data) &&
 			dc_link_dp_allow_hpd_rx_irq(dc_link)) {
-		dc_link_dp_handle_link_loss(dc_link);
+		/* offload_work->data is from handle_hpd_rx_irq->
+		 * schedule_hpd_rx_offload_work.this is defer handle
+		 * for hpd short pulse. upon here, link status may be
+		 * changed, need get latest link status from dpcd
+		 * registers. if link status is good, skip run link
+		 * training again.
+		 */
+		union hpd_irq_data irq_data;
+
+		memset(&irq_data, 0, sizeof(irq_data));
+
+		/* before dc_link_dp_handle_link_loss, allow new link lost handle
+		 * request be added to work queue if link lost at end of dc_link_
+		 * dp_handle_link_loss
+		 */
 		spin_lock_irqsave(&offload_work->offload_wq->offload_lock, flags);
 		offload_work->offload_wq->is_handling_link_loss = false;
 		spin_unlock_irqrestore(&offload_work->offload_wq->offload_lock, flags);
+
+		if ((dc_link_dp_read_hpd_rx_irq_data(dc_link, &irq_data) == DC_OK) &&
+			dc_link_check_link_loss_status(dc_link, &irq_data))
+			dc_link_dp_handle_link_loss(dc_link);
 	}
 	mutex_unlock(&adev->dm.dc_lock);
 
@@ -1302,8 +1649,9 @@ skip:
 
 }
 
-static struct hpd_rx_irq_offload_work_queue *hpd_rx_irq_create_workqueue(struct dc *dc)
+static struct hpd_rx_irq_offload_work_queue *hpd_rx_irq_create_workqueue(struct amdgpu_device *adev)
 {
+	struct dc *dc = adev->dm.dc;
 	int max_caps = dc->caps.max_links;
 	int i = 0;
 	struct hpd_rx_irq_offload_work_queue *hpd_rx_offload_wq = NULL;
@@ -1319,14 +1667,22 @@ static struct hpd_rx_irq_offload_work_queue *hpd_rx_irq_create_workqueue(struct
 				    create_singlethread_workqueue("amdgpu_dm_hpd_rx_offload_wq");
 
 		if (hpd_rx_offload_wq[i].wq == NULL) {
-			DRM_ERROR("create amdgpu_dm_hpd_rx_offload_wq fail!");
-			return NULL;
+			drm_err(adev_to_drm(adev), "create amdgpu_dm_hpd_rx_offload_wq fail!");
+			goto out_err;
 		}
 
 		spin_lock_init(&hpd_rx_offload_wq[i].offload_lock);
 	}
 
 	return hpd_rx_offload_wq;
+
+out_err:
+	for (i = 0; i < max_caps; i++) {
+		if (hpd_rx_offload_wq[i].wq)
+			destroy_workqueue(hpd_rx_offload_wq[i].wq);
+	}
+	kfree(hpd_rx_offload_wq);
+	return NULL;
 }
 
 struct amdgpu_stutter_quirk {
@@ -1360,12 +1716,169 @@ static bool dm_should_disable_stutter(struct pci_dev *pdev)
 	return false;
 }
 
+
+void*
+dm_allocate_gpu_mem(
+		struct amdgpu_device *adev,
+		enum dc_gpu_mem_alloc_type type,
+		size_t size,
+		long long *addr)
+{
+	struct dal_allocation *da;
+	u32 domain = (type == DC_MEM_ALLOC_TYPE_GART) ?
+		AMDGPU_GEM_DOMAIN_GTT : AMDGPU_GEM_DOMAIN_VRAM;
+	int ret;
+
+	da = kzalloc(sizeof(struct dal_allocation), GFP_KERNEL);
+	if (!da)
+		return NULL;
+
+	ret = amdgpu_bo_create_kernel(adev, size, PAGE_SIZE,
+				      domain, &da->bo,
+				      &da->gpu_addr, &da->cpu_ptr);
+
+	*addr = da->gpu_addr;
+
+	if (ret) {
+		kfree(da);
+		return NULL;
+	}
+
+	/* add da to list in dm */
+	list_add(&da->list, &adev->dm.da_list);
+
+	return da->cpu_ptr;
+}
+
+void
+dm_free_gpu_mem(
+		struct amdgpu_device *adev,
+		enum dc_gpu_mem_alloc_type type,
+		void *pvMem)
+{
+	struct dal_allocation *da;
+
+	/* walk the da list in DM */
+	list_for_each_entry(da, &adev->dm.da_list, list) {
+		if (pvMem == da->cpu_ptr) {
+			amdgpu_bo_free_kernel(&da->bo, &da->gpu_addr, &da->cpu_ptr);
+			list_del(&da->list);
+			kfree(da);
+			break;
+		}
+	}
+
+}
+
+static enum dmub_status
+dm_dmub_send_vbios_gpint_command(struct amdgpu_device *adev,
+				 enum dmub_gpint_command command_code,
+				 uint16_t param,
+				 uint32_t timeout_us)
+{
+	union dmub_gpint_data_register reg, test;
+	uint32_t i;
+
+	/* Assume that VBIOS DMUB is ready to take commands */
+
+	reg.bits.status = 1;
+	reg.bits.command_code = command_code;
+	reg.bits.param = param;
+
+	cgs_write_register(adev->dm.cgs_device, 0x34c0 + 0x01f8, reg.all);
+
+	for (i = 0; i < timeout_us; ++i) {
+		udelay(1);
+
+		/* Check if our GPINT got acked */
+		reg.bits.status = 0;
+		test = (union dmub_gpint_data_register)
+			cgs_read_register(adev->dm.cgs_device, 0x34c0 + 0x01f8);
+
+		if (test.all == reg.all)
+			return DMUB_STATUS_OK;
+	}
+
+	return DMUB_STATUS_TIMEOUT;
+}
+
+static void *dm_dmub_get_vbios_bounding_box(struct amdgpu_device *adev)
+{
+	void *bb;
+	long long addr;
+	unsigned int bb_size;
+	int i = 0;
+	uint16_t chunk;
+	enum dmub_gpint_command send_addrs[] = {
+		DMUB_GPINT__SET_BB_ADDR_WORD0,
+		DMUB_GPINT__SET_BB_ADDR_WORD1,
+		DMUB_GPINT__SET_BB_ADDR_WORD2,
+		DMUB_GPINT__SET_BB_ADDR_WORD3,
+	};
+	enum dmub_status ret;
+
+	switch (amdgpu_ip_version(adev, DCE_HWIP, 0)) {
+	case IP_VERSION(4, 0, 1):
+		bb_size = sizeof(struct dml2_soc_bb);
+		break;
+	default:
+		return NULL;
+	}
+
+	bb =  dm_allocate_gpu_mem(adev,
+				  DC_MEM_ALLOC_TYPE_GART,
+				  bb_size,
+				  &addr);
+	if (!bb)
+		return NULL;
+
+	for (i = 0; i < 4; i++) {
+		/* Extract 16-bit chunk */
+		chunk = ((uint64_t) addr >> (i * 16)) & 0xFFFF;
+		/* Send the chunk */
+		ret = dm_dmub_send_vbios_gpint_command(adev, send_addrs[i], chunk, 30000);
+		if (ret != DMUB_STATUS_OK)
+			goto free_bb;
+	}
+
+	/* Now ask DMUB to copy the bb */
+	ret = dm_dmub_send_vbios_gpint_command(adev, DMUB_GPINT__BB_COPY, 1, 200000);
+	if (ret != DMUB_STATUS_OK)
+		goto free_bb;
+
+	return bb;
+
+free_bb:
+	dm_free_gpu_mem(adev, DC_MEM_ALLOC_TYPE_GART, (void *) bb);
+	return NULL;
+
+}
+
+static enum dmub_ips_disable_type dm_get_default_ips_mode(
+	struct amdgpu_device *adev)
+{
+	enum dmub_ips_disable_type ret = DMUB_IPS_ENABLE;
+
+	switch (amdgpu_ip_version(adev, DCE_HWIP, 0)) {
+	case IP_VERSION(3, 5, 0):
+	case IP_VERSION(3, 6, 0):
+	case IP_VERSION(3, 5, 1):
+		ret =  DMUB_IPS_RCG_IN_ACTIVE_IPS2_IN_OFF;
+		break;
+	default:
+		/* ASICs older than DCN35 do not have IPSs */
+		if (amdgpu_ip_version(adev, DCE_HWIP, 0) < IP_VERSION(3, 5, 0))
+			ret = DMUB_IPS_DISABLE_ALL;
+		break;
+	}
+
+	return ret;
+}
+
 static int amdgpu_dm_init(struct amdgpu_device *adev)
 {
 	struct dc_init_data init_data;
-#ifdef CONFIG_DRM_AMD_DC_HDCP
 	struct dc_callback_init init_params;
-#endif
 	int r;
 
 	adev->dm.ddev = adev_to_drm(adev);
@@ -1373,18 +1886,14 @@ static int amdgpu_dm_init(struct amdgpu_device *adev)
 
 	/* Zero all the fields */
 	memset(&init_data, 0, sizeof(init_data));
-#ifdef CONFIG_DRM_AMD_DC_HDCP
 	memset(&init_params, 0, sizeof(init_params));
-#endif
 
+	mutex_init(&adev->dm.dpia_aux_lock);
 	mutex_init(&adev->dm.dc_lock);
 	mutex_init(&adev->dm.audio_lock);
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	spin_lock_init(&adev->dm.vblank_lock);
-#endif
 
-	if(amdgpu_dm_irq_init(adev)) {
-		DRM_ERROR("amdgpu: failed to initialize DM IRQ support.\n");
+	if (amdgpu_dm_irq_init(adev)) {
+		drm_err(adev_to_drm(adev), "failed to initialize DM IRQ support.\n");
 		goto error;
 	}
 
@@ -1401,52 +1910,52 @@ static int amdgpu_dm_init(struct amdgpu_device *adev)
 
 	init_data.driver = adev;
 
-	adev->dm.cgs_device = amdgpu_cgs_create_device(adev);
-
-	if (!adev->dm.cgs_device) {
-		DRM_ERROR("amdgpu: failed to create cgs device.\n");
-		goto error;
-	}
-
+	/* cgs_device was created in dm_sw_init() */
 	init_data.cgs_device = adev->dm.cgs_device;
 
 	init_data.dce_environment = DCE_ENV_PRODUCTION_DRV;
 
-	switch (adev->asic_type) {
-	case CHIP_CARRIZO:
-	case CHIP_STONEY:
-		init_data.flags.gpu_vm_support = true;
-		break;
-	default:
-		switch (adev->ip_versions[DCE_HWIP][0]) {
-		case IP_VERSION(2, 1, 0):
-			init_data.flags.gpu_vm_support = true;
-			switch (adev->dm.dmcub_fw_version) {
-			case 0: /* development */
-			case 0x1: /* linux-firmware.git hash 6d9f399 */
-			case 0x01000000: /* linux-firmware.git hash 9a0b0f4 */
-				init_data.flags.disable_dmcu = false;
-				break;
-			default:
-				init_data.flags.disable_dmcu = true;
-			}
-			break;
-		case IP_VERSION(1, 0, 0):
-		case IP_VERSION(1, 0, 1):
-		case IP_VERSION(3, 0, 1):
-		case IP_VERSION(3, 1, 2):
-		case IP_VERSION(3, 1, 3):
-			init_data.flags.gpu_vm_support = true;
-			break;
-		case IP_VERSION(2, 0, 3):
-			init_data.flags.disable_dmcu = true;
+	switch (amdgpu_ip_version(adev, DCE_HWIP, 0)) {
+	case IP_VERSION(2, 1, 0):
+		switch (adev->dm.dmcub_fw_version) {
+		case 0: /* development */
+		case 0x1: /* linux-firmware.git hash 6d9f399 */
+		case 0x01000000: /* linux-firmware.git hash 9a0b0f4 */
+			init_data.flags.disable_dmcu = false;
 			break;
 		default:
-			break;
+			init_data.flags.disable_dmcu = true;
 		}
 		break;
+	case IP_VERSION(2, 0, 3):
+		init_data.flags.disable_dmcu = true;
+		break;
+	default:
+		break;
 	}
 
+	/* APU support S/G display by default except:
+	 * ASICs before Carrizo,
+	 * RAVEN1 (Users reported stability issue)
+	 */
+
+	if (adev->asic_type < CHIP_CARRIZO) {
+		init_data.flags.gpu_vm_support = false;
+	} else if (adev->asic_type == CHIP_RAVEN) {
+		if (adev->apu_flags & AMD_APU_IS_RAVEN)
+			init_data.flags.gpu_vm_support = false;
+		else
+			init_data.flags.gpu_vm_support = (amdgpu_sg_display != 0);
+	} else {
+		if (amdgpu_ip_version(adev, DCE_HWIP, 0) == IP_VERSION(2, 0, 3))
+			init_data.flags.gpu_vm_support = (amdgpu_sg_display == 1);
+		else
+			init_data.flags.gpu_vm_support =
+				(amdgpu_sg_display != 0) && (adev->flags & AMD_IS_APU);
+	}
+
+	adev->mode_info.gpu_vm_support = init_data.flags.gpu_vm_support;
+
 	if (amdgpu_dc_feature_mask & DC_FBC_MASK)
 		init_data.flags.fbc_support = true;
 
@@ -1459,16 +1968,63 @@ static int amdgpu_dm_init(struct amdgpu_device *adev)
 	if (amdgpu_dc_feature_mask & DC_EDP_NO_POWER_SEQUENCING)
 		init_data.flags.edp_no_power_sequencing = true;
 
-	init_data.flags.power_down_display_on_boot = true;
+	if (amdgpu_dc_feature_mask & DC_DISABLE_LTTPR_DP1_4A)
+		init_data.flags.allow_lttpr_non_transparent_mode.bits.DP1_4A = true;
+	if (amdgpu_dc_feature_mask & DC_DISABLE_LTTPR_DP2_0)
+		init_data.flags.allow_lttpr_non_transparent_mode.bits.DP2_0 = true;
+
+	init_data.flags.seamless_boot_edp_requested = false;
+
+	if (amdgpu_device_seamless_boot_supported(adev)) {
+		init_data.flags.seamless_boot_edp_requested = true;
+		init_data.flags.allow_seamless_boot_optimization = true;
+		drm_dbg(adev->dm.ddev, "Seamless boot requested\n");
+	}
+
+	init_data.flags.enable_mipi_converter_optimization = true;
+
+	init_data.dcn_reg_offsets = adev->reg_offset[DCE_HWIP][0];
+	init_data.nbio_reg_offsets = adev->reg_offset[NBIO_HWIP][0];
+	init_data.clk_reg_offsets = adev->reg_offset[CLK_HWIP][0];
+
+	if (amdgpu_dc_debug_mask & DC_DISABLE_IPS)
+		init_data.flags.disable_ips = DMUB_IPS_DISABLE_ALL;
+	else if (amdgpu_dc_debug_mask & DC_DISABLE_IPS_DYNAMIC)
+		init_data.flags.disable_ips = DMUB_IPS_DISABLE_DYNAMIC;
+	else if (amdgpu_dc_debug_mask & DC_DISABLE_IPS2_DYNAMIC)
+		init_data.flags.disable_ips = DMUB_IPS_RCG_IN_ACTIVE_IPS2_IN_OFF;
+	else if (amdgpu_dc_debug_mask & DC_FORCE_IPS_ENABLE)
+		init_data.flags.disable_ips = DMUB_IPS_ENABLE;
+	else
+		init_data.flags.disable_ips = dm_get_default_ips_mode(adev);
+
+	init_data.flags.disable_ips_in_vpb = 0;
+
+	/* DCN35 and above supports dynamic DTBCLK switch */
+	if (amdgpu_ip_version(adev, DCE_HWIP, 0) >= IP_VERSION(3, 5, 0))
+		init_data.flags.allow_0_dtb_clk = true;
+
+	/* Enable DWB for tested platforms only */
+	if (amdgpu_ip_version(adev, DCE_HWIP, 0) >= IP_VERSION(3, 0, 0))
+		init_data.num_virtual_links = 1;
+
+	retrieve_dmi_info(&adev->dm);
+	if (adev->dm.edp0_on_dp1_quirk)
+		init_data.flags.support_edp0_on_dp1 = true;
+
+	if (adev->dm.bb_from_dmub)
+		init_data.bb_from_dmub = adev->dm.bb_from_dmub;
+	else
+		init_data.bb_from_dmub = NULL;
 
-	INIT_LIST_HEAD(&adev->dm.da_list);
 	/* Display Core create. */
 	adev->dm.dc = dc_create(&init_data);
 
 	if (adev->dm.dc) {
-		DRM_INFO("Display Core initialized with v%s!\n", DC_VER);
+		drm_info(adev_to_drm(adev), "Display Core v%s initialized on %s\n", DC_VER,
+			 dce_version_to_string(adev->dm.dc->ctx->dce_version));
 	} else {
-		DRM_INFO("Display Core failed to initialize with v%s!\n", DC_VER);
+		drm_info(adev_to_drm(adev), "Display Core failed to initialize with v%s!\n", DC_VER);
 		goto error;
 	}
 
@@ -1491,21 +2047,50 @@ static int amdgpu_dm_init(struct amdgpu_device *adev)
 	if (amdgpu_dc_debug_mask & DC_DISABLE_CLOCK_GATING)
 		adev->dm.dc->debug.disable_clock_gate = true;
 
+	if (amdgpu_dc_debug_mask & DC_FORCE_SUBVP_MCLK_SWITCH)
+		adev->dm.dc->debug.force_subvp_mclk_switch = true;
+
+	if (amdgpu_dc_debug_mask & DC_DISABLE_SUBVP_FAMS) {
+		adev->dm.dc->debug.force_disable_subvp = true;
+		adev->dm.dc->debug.fams2_config.bits.enable = false;
+	}
+
+	if (amdgpu_dc_debug_mask & DC_ENABLE_DML2) {
+		adev->dm.dc->debug.using_dml2 = true;
+		adev->dm.dc->debug.using_dml21 = true;
+	}
+
+	if (amdgpu_dc_debug_mask & DC_HDCP_LC_FORCE_FW_ENABLE)
+		adev->dm.dc->debug.hdcp_lc_force_fw_enable = true;
+
+	if (amdgpu_dc_debug_mask & DC_HDCP_LC_ENABLE_SW_FALLBACK)
+		adev->dm.dc->debug.hdcp_lc_enable_sw_fallback = true;
+
+	if (amdgpu_dc_debug_mask & DC_SKIP_DETECTION_LT)
+		adev->dm.dc->debug.skip_detection_link_training = true;
+
+	adev->dm.dc->debug.visual_confirm = amdgpu_dc_visual_confirm;
+
+	/* TODO: Remove after DP2 receiver gets proper support of Cable ID feature */
+	adev->dm.dc->debug.ignore_cable_id = true;
+
+	if (adev->dm.dc->caps.dp_hdmi21_pcon_support)
+		drm_info(adev_to_drm(adev), "DP-HDMI FRL PCON supported\n");
+
 	r = dm_dmub_hw_init(adev);
 	if (r) {
-		DRM_ERROR("DMUB interface failed to initialize: status=%d\n", r);
+		drm_err(adev_to_drm(adev), "DMUB interface failed to initialize: status=%d\n", r);
 		goto error;
 	}
 
 	dc_hardware_init(adev->dm.dc);
 
-	adev->dm.hpd_rx_offload_wq = hpd_rx_irq_create_workqueue(adev->dm.dc);
+	adev->dm.hpd_rx_offload_wq = hpd_rx_irq_create_workqueue(adev);
 	if (!adev->dm.hpd_rx_offload_wq) {
-		DRM_ERROR("amdgpu: failed to create hpd rx offload workqueue.\n");
+		drm_err(adev_to_drm(adev), "failed to create hpd rx offload workqueue.\n");
 		goto error;
 	}
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	if ((adev->flags & AMD_IS_APU) && (adev->asic_type >= CHIP_CARRIZO)) {
 		struct dc_phy_addr_space_config pa_config;
 
@@ -1514,77 +2099,83 @@ static int amdgpu_dm_init(struct amdgpu_device *adev)
 		// Call the DC init_memory func
 		dc_setup_system_context(adev->dm.dc, &pa_config);
 	}
-#endif
 
 	adev->dm.freesync_module = mod_freesync_create(adev->dm.dc);
 	if (!adev->dm.freesync_module) {
-		DRM_ERROR(
-		"amdgpu: failed to initialize freesync_module.\n");
+		drm_err(adev_to_drm(adev),
+		"failed to initialize freesync_module.\n");
 	} else
-		DRM_DEBUG_DRIVER("amdgpu: freesync_module init done %p.\n",
+		drm_dbg_driver(adev_to_drm(adev), "amdgpu: freesync_module init done %p.\n",
 				adev->dm.freesync_module);
 
 	amdgpu_dm_init_color_mod();
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	if (adev->dm.dc->caps.max_links > 0) {
 		adev->dm.vblank_control_workqueue =
 			create_singlethread_workqueue("dm_vblank_control_workqueue");
 		if (!adev->dm.vblank_control_workqueue)
-			DRM_ERROR("amdgpu: failed to initialize vblank_workqueue.\n");
+			drm_err(adev_to_drm(adev), "failed to initialize vblank_workqueue.\n");
 	}
-#endif
 
-#ifdef CONFIG_DRM_AMD_DC_HDCP
+	if (adev->dm.dc->caps.ips_support &&
+	    adev->dm.dc->config.disable_ips != DMUB_IPS_DISABLE_ALL)
+		adev->dm.idle_workqueue = idle_create_workqueue(adev);
+
 	if (adev->dm.dc->caps.max_links > 0 && adev->family >= AMDGPU_FAMILY_RV) {
 		adev->dm.hdcp_workqueue = hdcp_create_workqueue(adev, &init_params.cp_psp, adev->dm.dc);
 
 		if (!adev->dm.hdcp_workqueue)
-			DRM_ERROR("amdgpu: failed to initialize hdcp_workqueue.\n");
+			drm_err(adev_to_drm(adev), "failed to initialize hdcp_workqueue.\n");
 		else
-			DRM_DEBUG_DRIVER("amdgpu: hdcp_workqueue init done %p.\n", adev->dm.hdcp_workqueue);
+			drm_dbg_driver(adev_to_drm(adev), "amdgpu: hdcp_workqueue init done %p.\n", adev->dm.hdcp_workqueue);
 
 		dc_init_callbacks(adev->dm.dc, &init_params);
 	}
-#endif
-#if defined(CONFIG_DRM_AMD_SECURE_DISPLAY)
-	adev->dm.crc_rd_wrk = amdgpu_dm_crtc_secure_display_create_work();
-#endif
-	if (dc_enable_dmub_notifications(adev->dm.dc)) {
+	if (dc_is_dmub_outbox_supported(adev->dm.dc)) {
 		init_completion(&adev->dm.dmub_aux_transfer_done);
 		adev->dm.dmub_notify = kzalloc(sizeof(struct dmub_notification), GFP_KERNEL);
 		if (!adev->dm.dmub_notify) {
-			DRM_INFO("amdgpu: fail to allocate adev->dm.dmub_notify");
+			drm_info(adev_to_drm(adev), "fail to allocate adev->dm.dmub_notify");
 			goto error;
 		}
 
 		adev->dm.delayed_hpd_wq = create_singlethread_workqueue("amdgpu_dm_hpd_wq");
 		if (!adev->dm.delayed_hpd_wq) {
-			DRM_ERROR("amdgpu: failed to create hpd offload workqueue.\n");
+			drm_err(adev_to_drm(adev), "failed to create hpd offload workqueue.\n");
 			goto error;
 		}
 
 		amdgpu_dm_outbox_init(adev);
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 		if (!register_dmub_notify_callback(adev, DMUB_NOTIFICATION_AUX_REPLY,
 			dmub_aux_setconfig_callback, false)) {
-			DRM_ERROR("amdgpu: fail to register dmub aux callback");
-			goto error;
-		}
-		if (!register_dmub_notify_callback(adev, DMUB_NOTIFICATION_HPD, dmub_hpd_callback, true)) {
-			DRM_ERROR("amdgpu: fail to register dmub hpd callback");
+			drm_err(adev_to_drm(adev), "fail to register dmub aux callback");
 			goto error;
 		}
-		if (!register_dmub_notify_callback(adev, DMUB_NOTIFICATION_HPD_IRQ, dmub_hpd_callback, true)) {
-			DRM_ERROR("amdgpu: fail to register dmub hpd callback");
+
+		for (size_t i = 0; i < ARRAY_SIZE(adev->dm.fused_io); i++)
+			init_completion(&adev->dm.fused_io[i].replied);
+
+		if (!register_dmub_notify_callback(adev, DMUB_NOTIFICATION_FUSED_IO,
+			dmub_aux_fused_io_callback, false)) {
+			drm_err(adev_to_drm(adev), "fail to register dmub fused io callback");
 			goto error;
 		}
-#endif /* CONFIG_DRM_AMD_DC_DCN */
+		/* Enable outbox notification only after IRQ handlers are registered and DMUB is alive.
+		 * It is expected that DMUB will resend any pending notifications at this point. Note
+		 * that hpd and hpd_irq handler registration are deferred to register_hpd_handlers() to
+		 * align legacy interface initialization sequence. Connection status will be proactivly
+		 * detected once in the amdgpu_dm_initialize_drm_device.
+		 */
+		dc_enable_dmub_outbox(adev->dm.dc);
+
+		/* DPIA trace goes to dmesg logs only if outbox is enabled */
+		if (amdgpu_dc_debug_mask & DC_ENABLE_DPIA_TRACE)
+			dc_dmub_srv_enable_dpia_trace(adev->dm.dc);
 	}
 
 	if (amdgpu_dm_initialize_drm_device(adev)) {
-		DRM_ERROR(
-		"amdgpu: failed to initialize sw for display support.\n");
+		drm_err(adev_to_drm(adev),
+		"failed to initialize sw for display support.\n");
 		goto error;
 	}
 
@@ -1598,13 +2189,22 @@ static int amdgpu_dm_init(struct amdgpu_device *adev)
 	adev_to_drm(adev)->mode_config.cursor_height = adev->dm.dc->caps.max_cursor_size;
 
 	if (drm_vblank_init(adev_to_drm(adev), adev->dm.display_indexes_num)) {
-		DRM_ERROR(
-		"amdgpu: failed to initialize sw for display support.\n");
+		drm_err(adev_to_drm(adev),
+		"failed to initialize vblank for display support.\n");
 		goto error;
 	}
 
+#if defined(CONFIG_DRM_AMD_SECURE_DISPLAY)
+	amdgpu_dm_crtc_secure_display_create_contexts(adev);
+	if (!adev->dm.secure_display_ctx.crtc_ctx)
+		drm_err(adev_to_drm(adev), "failed to initialize secure display contexts.\n");
+
+	if (amdgpu_ip_version(adev, DCE_HWIP, 0) >= IP_VERSION(4, 0, 1))
+		adev->dm.secure_display_ctx.support_mul_roi = true;
+
+#endif
 
-	DRM_DEBUG_DRIVER("KMS initialized.\n");
+	drm_dbg_driver(adev_to_drm(adev), "KMS initialized.\n");
 
 	return 0;
 error:
@@ -1613,9 +2213,9 @@ error:
 	return -EINVAL;
 }
 
-static int amdgpu_dm_early_fini(void *handle)
+static int amdgpu_dm_early_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	amdgpu_dm_audio_fini(adev);
 
@@ -1626,43 +2226,49 @@ static void amdgpu_dm_fini(struct amdgpu_device *adev)
 {
 	int i;
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	if (adev->dm.vblank_control_workqueue) {
 		destroy_workqueue(adev->dm.vblank_control_workqueue);
 		adev->dm.vblank_control_workqueue = NULL;
 	}
-#endif
 
-	for (i = 0; i < adev->dm.display_indexes_num; i++) {
-		drm_encoder_cleanup(&adev->dm.mst_encoders[i].base);
+	if (adev->dm.idle_workqueue) {
+		if (adev->dm.idle_workqueue->running) {
+			adev->dm.idle_workqueue->enable = false;
+			flush_work(&adev->dm.idle_workqueue->work);
+		}
+
+		kfree(adev->dm.idle_workqueue);
+		adev->dm.idle_workqueue = NULL;
 	}
 
 	amdgpu_dm_destroy_drm_device(&adev->dm);
 
 #if defined(CONFIG_DRM_AMD_SECURE_DISPLAY)
-	if (adev->dm.crc_rd_wrk) {
-		flush_work(&adev->dm.crc_rd_wrk->notify_ta_work);
-		kfree(adev->dm.crc_rd_wrk);
-		adev->dm.crc_rd_wrk = NULL;
+	if (adev->dm.secure_display_ctx.crtc_ctx) {
+		for (i = 0; i < adev->mode_info.num_crtc; i++) {
+			if (adev->dm.secure_display_ctx.crtc_ctx[i].crtc) {
+				flush_work(&adev->dm.secure_display_ctx.crtc_ctx[i].notify_ta_work);
+				flush_work(&adev->dm.secure_display_ctx.crtc_ctx[i].forward_roi_work);
+			}
+		}
+		kfree(adev->dm.secure_display_ctx.crtc_ctx);
+		adev->dm.secure_display_ctx.crtc_ctx = NULL;
 	}
 #endif
-#ifdef CONFIG_DRM_AMD_DC_HDCP
 	if (adev->dm.hdcp_workqueue) {
 		hdcp_destroy(&adev->dev->kobj, adev->dm.hdcp_workqueue);
 		adev->dm.hdcp_workqueue = NULL;
 	}
 
-	if (adev->dm.dc)
+	if (adev->dm.dc) {
 		dc_deinit_callbacks(adev->dm.dc);
-#endif
-
-	dc_dmub_srv_destroy(&adev->dm.dc->ctx->dmub_srv);
-
-	if (dc_enable_dmub_notifications(adev->dm.dc)) {
-		kfree(adev->dm.dmub_notify);
-		adev->dm.dmub_notify = NULL;
-		destroy_workqueue(adev->dm.delayed_hpd_wq);
-		adev->dm.delayed_hpd_wq = NULL;
+		dc_dmub_srv_destroy(&adev->dm.dc->ctx->dmub_srv);
+		if (dc_enable_dmub_notifications(adev->dm.dc)) {
+			kfree(adev->dm.dmub_notify);
+			adev->dm.dmub_notify = NULL;
+			destroy_workqueue(adev->dm.delayed_hpd_wq);
+			adev->dm.delayed_hpd_wq = NULL;
+		}
 	}
 
 	if (adev->dm.dmub_bo)
@@ -1670,7 +2276,7 @@ static void amdgpu_dm_fini(struct amdgpu_device *adev)
 				      &adev->dm.dmub_bo_gpu_addr,
 				      &adev->dm.dmub_bo_cpu_addr);
 
-	if (adev->dm.hpd_rx_offload_wq) {
+	if (adev->dm.hpd_rx_offload_wq && adev->dm.dc) {
 		for (i = 0; i < adev->dm.dc->caps.max_links; i++) {
 			if (adev->dm.hpd_rx_offload_wq[i].wq) {
 				destroy_workqueue(adev->dm.hpd_rx_offload_wq[i].wq);
@@ -1702,8 +2308,7 @@ static void amdgpu_dm_fini(struct amdgpu_device *adev)
 
 	mutex_destroy(&adev->dm.audio_lock);
 	mutex_destroy(&adev->dm.dc_lock);
-
-	return;
+	mutex_destroy(&adev->dm.dpia_aux_lock);
 }
 
 static int load_dmcu_fw(struct amdgpu_device *adev)
@@ -1712,7 +2317,7 @@ static int load_dmcu_fw(struct amdgpu_device *adev)
 	int r;
 	const struct dmcu_firmware_header_v1_0 *hdr;
 
-	switch(adev->asic_type) {
+	switch (adev->asic_type) {
 #if defined(CONFIG_DRM_AMD_DC_SI)
 	case CHIP_TAHITI:
 	case CHIP_PITCAIRN:
@@ -1748,7 +2353,7 @@ static int load_dmcu_fw(struct amdgpu_device *adev)
 			return 0;
 		break;
 	default:
-		switch (adev->ip_versions[DCE_HWIP][0]) {
+		switch (amdgpu_ip_version(adev, DCE_HWIP, 0)) {
 		case IP_VERSION(2, 0, 2):
 		case IP_VERSION(2, 0, 3):
 		case IP_VERSION(2, 0, 0):
@@ -1759,11 +2364,20 @@ static int load_dmcu_fw(struct amdgpu_device *adev)
 		case IP_VERSION(3, 0, 1):
 		case IP_VERSION(3, 1, 2):
 		case IP_VERSION(3, 1, 3):
+		case IP_VERSION(3, 1, 4):
+		case IP_VERSION(3, 1, 5):
+		case IP_VERSION(3, 1, 6):
+		case IP_VERSION(3, 2, 0):
+		case IP_VERSION(3, 2, 1):
+		case IP_VERSION(3, 5, 0):
+		case IP_VERSION(3, 5, 1):
+		case IP_VERSION(3, 6, 0):
+		case IP_VERSION(4, 0, 1):
 			return 0;
 		default:
 			break;
 		}
-		DRM_ERROR("Unsupported ASIC type: 0x%X\n", adev->asic_type);
+		drm_err(adev_to_drm(adev), "Unsupported ASIC type: 0x%X\n", adev->asic_type);
 		return -EINVAL;
 	}
 
@@ -1772,25 +2386,18 @@ static int load_dmcu_fw(struct amdgpu_device *adev)
 		return 0;
 	}
 
-	r = request_firmware_direct(&adev->dm.fw_dmcu, fw_name_dmcu, adev->dev);
-	if (r == -ENOENT) {
+	r = amdgpu_ucode_request(adev, &adev->dm.fw_dmcu, AMDGPU_UCODE_REQUIRED,
+				 "%s", fw_name_dmcu);
+	if (r == -ENODEV) {
 		/* DMCU firmware is not necessary, so don't raise a fuss if it's missing */
 		DRM_DEBUG_KMS("dm: DMCU firmware not found\n");
 		adev->dm.fw_dmcu = NULL;
 		return 0;
 	}
 	if (r) {
-		dev_err(adev->dev, "amdgpu_dm: Can't load firmware \"%s\"\n",
-			fw_name_dmcu);
-		return r;
-	}
-
-	r = amdgpu_ucode_validate(adev->dm.fw_dmcu);
-	if (r) {
-		dev_err(adev->dev, "amdgpu_dm: Can't validate firmware \"%s\"\n",
+		drm_err(adev_to_drm(adev), "amdgpu_dm: Can't validate firmware \"%s\"\n",
 			fw_name_dmcu);
-		release_firmware(adev->dm.fw_dmcu);
-		adev->dm.fw_dmcu = NULL;
+		amdgpu_ucode_release(&adev->dm.fw_dmcu);
 		return r;
 	}
 
@@ -1832,47 +2439,70 @@ static int dm_dmub_sw_init(struct amdgpu_device *adev)
 	struct dmub_srv_create_params create_params;
 	struct dmub_srv_region_params region_params;
 	struct dmub_srv_region_info region_info;
-	struct dmub_srv_fb_params fb_params;
+	struct dmub_srv_memory_params memory_params;
 	struct dmub_srv_fb_info *fb_info;
 	struct dmub_srv *dmub_srv;
 	const struct dmcub_firmware_header_v1_0 *hdr;
-	const char *fw_name_dmub;
 	enum dmub_asic dmub_asic;
 	enum dmub_status status;
+	static enum dmub_window_memory_type window_memory_type[DMUB_WINDOW_TOTAL] = {
+		DMUB_WINDOW_MEMORY_TYPE_FB,		//DMUB_WINDOW_0_INST_CONST
+		DMUB_WINDOW_MEMORY_TYPE_FB,		//DMUB_WINDOW_1_STACK
+		DMUB_WINDOW_MEMORY_TYPE_FB,		//DMUB_WINDOW_2_BSS_DATA
+		DMUB_WINDOW_MEMORY_TYPE_FB,		//DMUB_WINDOW_3_VBIOS
+		DMUB_WINDOW_MEMORY_TYPE_FB,		//DMUB_WINDOW_4_MAILBOX
+		DMUB_WINDOW_MEMORY_TYPE_FB,		//DMUB_WINDOW_5_TRACEBUFF
+		DMUB_WINDOW_MEMORY_TYPE_FB,		//DMUB_WINDOW_6_FW_STATE
+		DMUB_WINDOW_MEMORY_TYPE_FB,		//DMUB_WINDOW_7_SCRATCH_MEM
+		DMUB_WINDOW_MEMORY_TYPE_FB,		//DMUB_WINDOW_IB_MEM
+		DMUB_WINDOW_MEMORY_TYPE_FB,		//DMUB_WINDOW_SHARED_STATE
+	};
 	int r;
 
-	switch (adev->ip_versions[DCE_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, DCE_HWIP, 0)) {
 	case IP_VERSION(2, 1, 0):
 		dmub_asic = DMUB_ASIC_DCN21;
-		fw_name_dmub = FIRMWARE_RENOIR_DMUB;
-		if (ASICREV_IS_GREEN_SARDINE(adev->external_rev_id))
-			fw_name_dmub = FIRMWARE_GREEN_SARDINE_DMUB;
 		break;
 	case IP_VERSION(3, 0, 0):
-		if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 3, 0)) {
-			dmub_asic = DMUB_ASIC_DCN30;
-			fw_name_dmub = FIRMWARE_SIENNA_CICHLID_DMUB;
-		} else {
-			dmub_asic = DMUB_ASIC_DCN30;
-			fw_name_dmub = FIRMWARE_NAVY_FLOUNDER_DMUB;
-		}
+		dmub_asic = DMUB_ASIC_DCN30;
 		break;
 	case IP_VERSION(3, 0, 1):
 		dmub_asic = DMUB_ASIC_DCN301;
-		fw_name_dmub = FIRMWARE_VANGOGH_DMUB;
 		break;
 	case IP_VERSION(3, 0, 2):
 		dmub_asic = DMUB_ASIC_DCN302;
-		fw_name_dmub = FIRMWARE_DIMGREY_CAVEFISH_DMUB;
 		break;
 	case IP_VERSION(3, 0, 3):
 		dmub_asic = DMUB_ASIC_DCN303;
-		fw_name_dmub = FIRMWARE_BEIGE_GOBY_DMUB;
 		break;
 	case IP_VERSION(3, 1, 2):
 	case IP_VERSION(3, 1, 3):
 		dmub_asic = (adev->external_rev_id == YELLOW_CARP_B0) ? DMUB_ASIC_DCN31B : DMUB_ASIC_DCN31;
-		fw_name_dmub = FIRMWARE_YELLOW_CARP_DMUB;
+		break;
+	case IP_VERSION(3, 1, 4):
+		dmub_asic = DMUB_ASIC_DCN314;
+		break;
+	case IP_VERSION(3, 1, 5):
+		dmub_asic = DMUB_ASIC_DCN315;
+		break;
+	case IP_VERSION(3, 1, 6):
+		dmub_asic = DMUB_ASIC_DCN316;
+		break;
+	case IP_VERSION(3, 2, 0):
+		dmub_asic = DMUB_ASIC_DCN32;
+		break;
+	case IP_VERSION(3, 2, 1):
+		dmub_asic = DMUB_ASIC_DCN321;
+		break;
+	case IP_VERSION(3, 5, 0):
+	case IP_VERSION(3, 5, 1):
+		dmub_asic = DMUB_ASIC_DCN35;
+		break;
+	case IP_VERSION(3, 6, 0):
+		dmub_asic = DMUB_ASIC_DCN36;
+		break;
+	case IP_VERSION(4, 0, 1):
+		dmub_asic = DMUB_ASIC_DCN401;
 		break;
 
 	default:
@@ -1880,18 +2510,6 @@ static int dm_dmub_sw_init(struct amdgpu_device *adev)
 		return 0;
 	}
 
-	r = request_firmware_direct(&adev->dm.dmub_fw, fw_name_dmub, adev->dev);
-	if (r) {
-		DRM_ERROR("DMUB firmware loading failed: %d\n", r);
-		return 0;
-	}
-
-	r = amdgpu_ucode_validate(adev->dm.dmub_fw);
-	if (r) {
-		DRM_ERROR("Couldn't validate DMUB firmware: %d\n", r);
-		return 0;
-	}
-
 	hdr = (const struct dmcub_firmware_header_v1_0 *)adev->dm.dmub_fw->data;
 	adev->dm.dmcub_fw_version = le32_to_cpu(hdr->header.ucode_version);
 
@@ -1903,7 +2521,7 @@ static int dm_dmub_sw_init(struct amdgpu_device *adev)
 		adev->firmware.fw_size +=
 			ALIGN(le32_to_cpu(hdr->inst_const_bytes), PAGE_SIZE);
 
-		DRM_INFO("Loading DMUB firmware via PSP: version=0x%08X\n",
+		drm_info(adev_to_drm(adev), "Loading DMUB firmware via PSP: version=0x%08X\n",
 			 adev->dm.dmcub_fw_version);
 	}
 
@@ -1912,7 +2530,7 @@ static int dm_dmub_sw_init(struct amdgpu_device *adev)
 	dmub_srv = adev->dm.dmub_srv;
 
 	if (!dmub_srv) {
-		DRM_ERROR("Failed to allocate DMUB service!\n");
+		drm_err(adev_to_drm(adev), "Failed to allocate DMUB service!\n");
 		return -ENOMEM;
 	}
 
@@ -1925,7 +2543,7 @@ static int dm_dmub_sw_init(struct amdgpu_device *adev)
 	/* Create the DMUB service. */
 	status = dmub_srv_create(dmub_srv, &create_params);
 	if (status != DMUB_STATUS_OK) {
-		DRM_ERROR("Error creating DMUB service: %d\n", status);
+		drm_err(adev_to_drm(adev), "Error creating DMUB service: %d\n", status);
 		return -EINVAL;
 	}
 
@@ -1944,12 +2562,13 @@ static int dm_dmub_sw_init(struct amdgpu_device *adev)
 		adev->dm.dmub_fw->data +
 		le32_to_cpu(hdr->header.ucode_array_offset_bytes) +
 		PSP_HEADER_BYTES;
+	region_params.window_memory_type = window_memory_type;
 
 	status = dmub_srv_calc_region_info(dmub_srv, &region_params,
 					   &region_info);
 
 	if (status != DMUB_STATUS_OK) {
-		DRM_ERROR("Error calculating DMUB region info: %d\n", status);
+		drm_err(adev_to_drm(adev), "Error calculating DMUB region info: %d\n", status);
 		return -EINVAL;
 	}
 
@@ -1958,42 +2577,57 @@ static int dm_dmub_sw_init(struct amdgpu_device *adev)
 	 * TODO: Move this into GART.
 	 */
 	r = amdgpu_bo_create_kernel(adev, region_info.fb_size, PAGE_SIZE,
-				    AMDGPU_GEM_DOMAIN_VRAM, &adev->dm.dmub_bo,
+				    AMDGPU_GEM_DOMAIN_VRAM |
+				    AMDGPU_GEM_DOMAIN_GTT,
+				    &adev->dm.dmub_bo,
 				    &adev->dm.dmub_bo_gpu_addr,
 				    &adev->dm.dmub_bo_cpu_addr);
 	if (r)
 		return r;
 
 	/* Rebase the regions on the framebuffer address. */
-	memset(&fb_params, 0, sizeof(fb_params));
-	fb_params.cpu_addr = adev->dm.dmub_bo_cpu_addr;
-	fb_params.gpu_addr = adev->dm.dmub_bo_gpu_addr;
-	fb_params.region_info = &region_info;
+	memset(&memory_params, 0, sizeof(memory_params));
+	memory_params.cpu_fb_addr = adev->dm.dmub_bo_cpu_addr;
+	memory_params.gpu_fb_addr = adev->dm.dmub_bo_gpu_addr;
+	memory_params.region_info = &region_info;
+	memory_params.window_memory_type = window_memory_type;
 
 	adev->dm.dmub_fb_info =
 		kzalloc(sizeof(*adev->dm.dmub_fb_info), GFP_KERNEL);
 	fb_info = adev->dm.dmub_fb_info;
 
 	if (!fb_info) {
-		DRM_ERROR(
+		drm_err(adev_to_drm(adev),
 			"Failed to allocate framebuffer info for DMUB service!\n");
 		return -ENOMEM;
 	}
 
-	status = dmub_srv_calc_fb_info(dmub_srv, &fb_params, fb_info);
+	status = dmub_srv_calc_mem_info(dmub_srv, &memory_params, fb_info);
 	if (status != DMUB_STATUS_OK) {
-		DRM_ERROR("Error calculating DMUB FB info: %d\n", status);
+		drm_err(adev_to_drm(adev), "Error calculating DMUB FB info: %d\n", status);
 		return -EINVAL;
 	}
 
+	adev->dm.bb_from_dmub = dm_dmub_get_vbios_bounding_box(adev);
+
 	return 0;
 }
 
-static int dm_sw_init(void *handle)
+static int dm_sw_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	int r;
 
+	adev->dm.cgs_device = amdgpu_cgs_create_device(adev);
+
+	if (!adev->dm.cgs_device) {
+		drm_err(adev_to_drm(adev), "failed to create cgs device.\n");
+		return -EINVAL;
+	}
+
+	/* Moved from dm init since we need to use allocations for storing bounding box data */
+	INIT_LIST_HEAD(&adev->dm.da_list);
+
 	r = dm_dmub_sw_init(adev);
 	if (r)
 		return r;
@@ -2001,23 +2635,33 @@ static int dm_sw_init(void *handle)
 	return load_dmcu_fw(adev);
 }
 
-static int dm_sw_fini(void *handle)
+static int dm_sw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
+	struct dal_allocation *da;
+
+	list_for_each_entry(da, &adev->dm.da_list, list) {
+		if (adev->dm.bb_from_dmub == (void *) da->cpu_ptr) {
+			amdgpu_bo_free_kernel(&da->bo, &da->gpu_addr, &da->cpu_ptr);
+			list_del(&da->list);
+			kfree(da);
+			adev->dm.bb_from_dmub = NULL;
+			break;
+		}
+	}
+
 
 	kfree(adev->dm.dmub_fb_info);
 	adev->dm.dmub_fb_info = NULL;
 
 	if (adev->dm.dmub_srv) {
 		dmub_srv_destroy(adev->dm.dmub_srv);
+		kfree(adev->dm.dmub_srv);
 		adev->dm.dmub_srv = NULL;
 	}
 
-	release_firmware(adev->dm.dmub_fw);
-	adev->dm.dmub_fw = NULL;
-
-	release_firmware(adev->dm.fw_dmcu);
-	adev->dm.fw_dmcu = NULL;
+	amdgpu_ucode_release(&adev->dm.dmub_fw);
+	amdgpu_ucode_release(&adev->dm.fw_dmcu);
 
 	return 0;
 }
@@ -2031,18 +2675,24 @@ static int detect_mst_link_for_all_connectors(struct drm_device *dev)
 
 	drm_connector_list_iter_begin(dev, &iter);
 	drm_for_each_connector_iter(connector, &iter) {
+
+		if (connector->connector_type == DRM_MODE_CONNECTOR_WRITEBACK)
+			continue;
+
 		aconnector = to_amdgpu_dm_connector(connector);
 		if (aconnector->dc_link->type == dc_connection_mst_branch &&
 		    aconnector->mst_mgr.aux) {
-			DRM_DEBUG_DRIVER("DM_MST: starting TM on aconnector: %p [id: %d]\n",
+			drm_dbg_kms(dev, "DM_MST: starting TM on aconnector: %p [id: %d]\n",
 					 aconnector,
 					 aconnector->base.base.id);
 
 			ret = drm_dp_mst_topology_mgr_set_mst(&aconnector->mst_mgr, true);
 			if (ret < 0) {
-				DRM_ERROR("DM_MST: Failed to start MST\n");
+				drm_err(dev, "DM_MST: Failed to start MST\n");
 				aconnector->dc_link->type =
 					dc_connection_single;
+				ret = dm_helpers_dp_mst_stop_top_mgr(aconnector->dc_link->ctx,
+								     aconnector->dc_link);
 				break;
 			}
 		}
@@ -2052,9 +2702,9 @@ static int detect_mst_link_for_all_connectors(struct drm_device *dev)
 	return ret;
 }
 
-static int dm_late_init(void *handle)
+static int dm_late_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	struct dmcu_iram_parameters params;
 	unsigned int linear_lut[16];
@@ -2078,9 +2728,9 @@ static int dm_late_init(void *handle)
 	 */
 	params.min_abm_backlight = 0x28F;
 	/* In the case where abm is implemented on dmcub,
-	* dmcu object will be null.
-	* ABM 2.4 and up are implemented on dmcub.
-	*/
+	 * dmcu object will be null.
+	 * ABM 2.4 and up are implemented on dmcub.
+	 */
 	if (dmcu) {
 		if (!dmcu_load_iram(dmcu, params))
 			return -EINVAL;
@@ -2088,7 +2738,7 @@ static int dm_late_init(void *handle)
 		struct dc_link *edp_links[MAX_NUM_EDP];
 		int edp_num;
 
-		get_edp_links(adev->dm.dc, edp_links, &edp_num);
+		dc_get_edp_links(adev->dm.dc, edp_links, &edp_num);
 		for (i = 0; i < edp_num; i++) {
 			if (!dmub_init_abm_config(adev->dm.dc->res_pool, params, i))
 				return -EINVAL;
@@ -2098,20 +2748,115 @@ static int dm_late_init(void *handle)
 	return detect_mst_link_for_all_connectors(adev_to_drm(adev));
 }
 
+static void resume_mst_branch_status(struct drm_dp_mst_topology_mgr *mgr)
+{
+	u8 buf[UUID_SIZE];
+	guid_t guid;
+	int ret;
+
+	mutex_lock(&mgr->lock);
+	if (!mgr->mst_primary)
+		goto out_fail;
+
+	if (drm_dp_read_dpcd_caps(mgr->aux, mgr->dpcd) < 0) {
+		drm_dbg_kms(mgr->dev, "dpcd read failed - undocked during suspend?\n");
+		goto out_fail;
+	}
+
+	ret = drm_dp_dpcd_writeb(mgr->aux, DP_MSTM_CTRL,
+				 DP_MST_EN |
+				 DP_UP_REQ_EN |
+				 DP_UPSTREAM_IS_SRC);
+	if (ret < 0) {
+		drm_dbg_kms(mgr->dev, "mst write failed - undocked during suspend?\n");
+		goto out_fail;
+	}
+
+	/* Some hubs forget their guids after they resume */
+	ret = drm_dp_dpcd_read(mgr->aux, DP_GUID, buf, sizeof(buf));
+	if (ret != sizeof(buf)) {
+		drm_dbg_kms(mgr->dev, "dpcd read failed - undocked during suspend?\n");
+		goto out_fail;
+	}
+
+	import_guid(&guid, buf);
+
+	if (guid_is_null(&guid)) {
+		guid_gen(&guid);
+		export_guid(buf, &guid);
+
+		ret = drm_dp_dpcd_write(mgr->aux, DP_GUID, buf, sizeof(buf));
+
+		if (ret != sizeof(buf)) {
+			drm_dbg_kms(mgr->dev, "check mstb guid failed - undocked during suspend?\n");
+			goto out_fail;
+		}
+	}
+
+	guid_copy(&mgr->mst_primary->guid, &guid);
+
+out_fail:
+	mutex_unlock(&mgr->lock);
+}
+
+void hdmi_cec_unset_edid(struct amdgpu_dm_connector *aconnector)
+{
+	struct cec_notifier *n = aconnector->notifier;
+
+	if (!n)
+		return;
+
+	cec_notifier_phys_addr_invalidate(n);
+}
+
+void hdmi_cec_set_edid(struct amdgpu_dm_connector *aconnector)
+{
+	struct drm_connector *connector = &aconnector->base;
+	struct cec_notifier *n = aconnector->notifier;
+
+	if (!n)
+		return;
+
+	cec_notifier_set_phys_addr(n,
+				   connector->display_info.source_physical_address);
+}
+
+static void s3_handle_hdmi_cec(struct drm_device *ddev, bool suspend)
+{
+	struct amdgpu_dm_connector *aconnector;
+	struct drm_connector *connector;
+	struct drm_connector_list_iter conn_iter;
+
+	drm_connector_list_iter_begin(ddev, &conn_iter);
+	drm_for_each_connector_iter(connector, &conn_iter) {
+		if (connector->connector_type == DRM_MODE_CONNECTOR_WRITEBACK)
+			continue;
+
+		aconnector = to_amdgpu_dm_connector(connector);
+		if (suspend)
+			hdmi_cec_unset_edid(aconnector);
+		else
+			hdmi_cec_set_edid(aconnector);
+	}
+	drm_connector_list_iter_end(&conn_iter);
+}
+
 static void s3_handle_mst(struct drm_device *dev, bool suspend)
 {
 	struct amdgpu_dm_connector *aconnector;
 	struct drm_connector *connector;
 	struct drm_connector_list_iter iter;
 	struct drm_dp_mst_topology_mgr *mgr;
-	int ret;
-	bool need_hotplug = false;
 
 	drm_connector_list_iter_begin(dev, &iter);
 	drm_for_each_connector_iter(connector, &iter) {
+
+		if (connector->connector_type == DRM_MODE_CONNECTOR_WRITEBACK)
+			continue;
+
 		aconnector = to_amdgpu_dm_connector(connector);
 		if (aconnector->dc_link->type != dc_connection_mst_branch ||
-		    aconnector->mst_port)
+		    aconnector->mst_root)
 			continue;
 
 		mgr = &aconnector->mst_mgr;
@@ -2119,27 +2864,29 @@ static void s3_handle_mst(struct drm_device *dev, bool suspend)
 		if (suspend) {
 			drm_dp_mst_topology_mgr_suspend(mgr);
 		} else {
-			ret = drm_dp_mst_topology_mgr_resume(mgr, true);
-			if (ret < 0) {
-				drm_dp_mst_topology_mgr_set_mst(mgr, false);
-				need_hotplug = true;
-			}
+			/* if extended timeout is supported in hardware,
+			 * default to LTTPR timeout (3.2ms) first as a W/A for DP link layer
+			 * CTS 4.2.1.1 regression introduced by CTS specs requirement update.
+			 */
+			try_to_configure_aux_timeout(aconnector->dc_link->ddc, LINK_AUX_DEFAULT_LTTPR_TIMEOUT_PERIOD);
+			if (!dp_is_lttpr_present(aconnector->dc_link))
+				try_to_configure_aux_timeout(aconnector->dc_link->ddc, LINK_AUX_DEFAULT_TIMEOUT_PERIOD);
+
+			/* TODO: move resume_mst_branch_status() into drm mst resume again
+			 * once topology probing work is pulled out from mst resume into mst
+			 * resume 2nd step. mst resume 2nd step should be called after old
+			 * state getting restored (i.e. drm_atomic_helper_resume()).
+			 */
+			resume_mst_branch_status(mgr);
 		}
 	}
 	drm_connector_list_iter_end(&iter);
-
-	if (need_hotplug)
-		drm_kms_helper_hotplug_event(dev);
 }
 
 static int amdgpu_dm_smu_write_watermarks_table(struct amdgpu_device *adev)
 {
-	struct smu_context *smu = &adev->smu;
 	int ret = 0;
 
-	if (!is_support_sw_smu(adev))
-		return 0;
-
 	/* This interface is for dGPU Navi1x.Linux dc-pplib interface depends
 	 * on window driver dc implementation.
 	 * For Navi1x, clock settings of dcn watermarks are fixed. the settings
@@ -2170,7 +2917,7 @@ static int amdgpu_dm_smu_write_watermarks_table(struct amdgpu_device *adev)
 	 * therefore, this function apply to navi10/12/14 but not Renoir
 	 * *
 	 */
-	switch (adev->ip_versions[DCE_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, DCE_HWIP, 0)) {
 	case IP_VERSION(2, 0, 2):
 	case IP_VERSION(2, 0, 0):
 		break;
@@ -2178,18 +2925,45 @@ static int amdgpu_dm_smu_write_watermarks_table(struct amdgpu_device *adev)
 		return 0;
 	}
 
-	ret = smu_write_watermarks_table(smu);
+	ret = amdgpu_dpm_write_watermarks_table(adev);
 	if (ret) {
-		DRM_ERROR("Failed to update WMTABLE!\n");
+		drm_err(adev_to_drm(adev), "Failed to update WMTABLE!\n");
 		return ret;
 	}
 
 	return 0;
 }
 
+static int dm_oem_i2c_hw_init(struct amdgpu_device *adev)
+{
+	struct amdgpu_display_manager *dm = &adev->dm;
+	struct amdgpu_i2c_adapter *oem_i2c;
+	struct ddc_service *oem_ddc_service;
+	int r;
+
+	oem_ddc_service = dc_get_oem_i2c_device(adev->dm.dc);
+	if (oem_ddc_service) {
+		oem_i2c = create_i2c(oem_ddc_service, true);
+		if (!oem_i2c) {
+			drm_info(adev_to_drm(adev), "Failed to create oem i2c adapter data\n");
+			return -ENOMEM;
+		}
+
+		r = devm_i2c_add_adapter(adev->dev, &oem_i2c->base);
+		if (r) {
+			drm_info(adev_to_drm(adev), "Failed to register oem i2c\n");
+			kfree(oem_i2c);
+			return r;
+		}
+		dm->oem_i2c = oem_i2c;
+	}
+
+	return 0;
+}
+
 /**
  * dm_hw_init() - Initialize DC device
- * @handle: The base driver device containing the amdgpu_dm device.
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Initialize the &struct amdgpu_display_manager device. This involves calling
  * the initializers of each DM component, then populating the struct with them.
@@ -2207,27 +2981,35 @@ static int amdgpu_dm_smu_write_watermarks_table(struct amdgpu_device *adev)
  * - Vblank support
  * - Debug FS entries, if enabled
  */
-static int dm_hw_init(void *handle)
+static int dm_hw_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
+	int r;
+
 	/* Create DAL display manager */
-	amdgpu_dm_init(adev);
+	r = amdgpu_dm_init(adev);
+	if (r)
+		return r;
 	amdgpu_dm_hpd_init(adev);
 
+	r = dm_oem_i2c_hw_init(adev);
+	if (r)
+		drm_info(adev_to_drm(adev), "Failed to add OEM i2c bus\n");
+
 	return 0;
 }
 
 /**
  * dm_hw_fini() - Teardown DC device
- * @handle: The base driver device containing the amdgpu_dm device.
+ * @ip_block: Pointer to the amdgpu_ip_block for this hw instance.
  *
  * Teardown components within &struct amdgpu_display_manager that require
  * cleanup. This involves cleaning up the DRM device, DC, and any modules that
  * were loaded. Also flush IRQ workqueues and disable them.
  */
-static int dm_hw_fini(void *handle)
+static int dm_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 
 	amdgpu_dm_hpd_fini(adev);
 
@@ -2237,9 +3019,6 @@ static int dm_hw_fini(void *handle)
 }
 
 
-static int dm_enable_vblank(struct drm_crtc *crtc);
-static void dm_disable_vblank(struct drm_crtc *crtc);
-
 static void dm_gpureset_toggle_interrupts(struct amdgpu_device *adev,
 				 struct dc_state *state, bool enable)
 {
@@ -2255,40 +3034,51 @@ static void dm_gpureset_toggle_interrupts(struct amdgpu_device *adev,
 		if (acrtc && state->stream_status[i].plane_count != 0) {
 			irq_source = IRQ_TYPE_PFLIP + acrtc->otg_inst;
 			rc = dc_interrupt_set(adev->dm.dc, irq_source, enable) ? 0 : -EBUSY;
-			DRM_DEBUG_VBL("crtc %d - vupdate irq %sabling: r=%d\n",
-				      acrtc->crtc_id, enable ? "en" : "dis", rc);
 			if (rc)
-				DRM_WARN("Failed to %s pflip interrupts\n",
+				drm_warn(adev_to_drm(adev), "Failed to %s pflip interrupts\n",
 					 enable ? "enable" : "disable");
 
-			if (enable) {
-				rc = dm_enable_vblank(&acrtc->base);
+			if (dc_supports_vrr(adev->dm.dc->ctx->dce_version)) {
+				if (enable) {
+					if (amdgpu_dm_crtc_vrr_active(
+							to_dm_crtc_state(acrtc->base.state)))
+						rc = amdgpu_dm_crtc_set_vupdate_irq(
+							&acrtc->base, true);
+				} else
+					rc = amdgpu_dm_crtc_set_vupdate_irq(
+							&acrtc->base, false);
+
 				if (rc)
-					DRM_WARN("Failed to enable vblank interrupts\n");
-			} else {
-				dm_disable_vblank(&acrtc->base);
+					drm_warn(adev_to_drm(adev), "Failed to %sable vupdate interrupt\n",
+						enable ? "en" : "dis");
 			}
 
+			irq_source = IRQ_TYPE_VBLANK + acrtc->otg_inst;
+			/* During gpu-reset we disable and then enable vblank irq, so
+			 * don't use amdgpu_irq_get/put() to avoid refcount change.
+			 */
+			if (!dc_interrupt_set(adev->dm.dc, irq_source, enable))
+				drm_warn(adev_to_drm(adev), "Failed to %sable vblank interrupt\n", enable ? "en" : "dis");
 		}
 	}
 
 }
 
+DEFINE_FREE(state_release, struct dc_state *, if (_T) dc_state_release(_T))
+
 static enum dc_status amdgpu_dm_commit_zero_streams(struct dc *dc)
 {
-	struct dc_state *context = NULL;
-	enum dc_status res = DC_ERROR_UNEXPECTED;
+	struct dc_state *context __free(state_release) = NULL;
 	int i;
 	struct dc_stream_state *del_streams[MAX_PIPES];
 	int del_streams_count = 0;
+	struct dc_commit_streams_params params = {};
 
 	memset(del_streams, 0, sizeof(del_streams));
 
-	context = dc_create_state(dc);
+	context = dc_state_create_current_copy(dc);
 	if (context == NULL)
-		goto context_alloc_fail;
-
-	dc_resource_state_copy_construct_current(dc, context);
+		return DC_ERROR_UNEXPECTED;
 
 	/* First remove from context all streams */
 	for (i = 0; i < context->stream_count; i++) {
@@ -2299,71 +3089,134 @@ static enum dc_status amdgpu_dm_commit_zero_streams(struct dc *dc)
 
 	/* Remove all planes for removed streams and then remove the streams */
 	for (i = 0; i < del_streams_count; i++) {
-		if (!dc_rem_all_planes_for_stream(dc, del_streams[i], context)) {
-			res = DC_FAIL_DETACH_SURFACES;
-			goto fail;
-		}
+		enum dc_status res;
+
+		if (!dc_state_rem_all_planes_for_stream(dc, del_streams[i], context))
+			return DC_FAIL_DETACH_SURFACES;
 
-		res = dc_remove_stream_from_ctx(dc, context, del_streams[i]);
+		res = dc_state_remove_stream(dc, context, del_streams[i]);
 		if (res != DC_OK)
-			goto fail;
+			return res;
 	}
 
+	params.streams = context->streams;
+	params.stream_count = context->stream_count;
 
-	res = dc_validate_global_state(dc, context, false);
+	return dc_commit_streams(dc, &params);
+}
 
-	if (res != DC_OK) {
-		DRM_ERROR("%s:resource validation failed, dc_status:%d\n", __func__, res);
-		goto fail;
+static void hpd_rx_irq_work_suspend(struct amdgpu_display_manager *dm)
+{
+	int i;
+
+	if (dm->hpd_rx_offload_wq) {
+		for (i = 0; i < dm->dc->caps.max_links; i++)
+			flush_workqueue(dm->hpd_rx_offload_wq[i].wq);
 	}
+}
 
-	res = dc_commit_state(dc, context);
+static int dm_cache_state(struct amdgpu_device *adev)
+{
+	int r;
 
-fail:
-	dc_release_state(context);
+	adev->dm.cached_state = drm_atomic_helper_suspend(adev_to_drm(adev));
+	if (IS_ERR(adev->dm.cached_state)) {
+		r = PTR_ERR(adev->dm.cached_state);
+		adev->dm.cached_state = NULL;
+	}
 
-context_alloc_fail:
-	return res;
+	return adev->dm.cached_state ? 0 : r;
 }
 
-static void hpd_rx_irq_work_suspend(struct amdgpu_display_manager *dm)
+static void dm_destroy_cached_state(struct amdgpu_device *adev)
 {
+	struct amdgpu_display_manager *dm = &adev->dm;
+	struct drm_device *ddev = adev_to_drm(adev);
+	struct dm_plane_state *dm_new_plane_state;
+	struct drm_plane_state *new_plane_state;
+	struct dm_crtc_state *dm_new_crtc_state;
+	struct drm_crtc_state *new_crtc_state;
+	struct drm_plane *plane;
+	struct drm_crtc *crtc;
 	int i;
 
-	if (dm->hpd_rx_offload_wq) {
-		for (i = 0; i < dm->dc->caps.max_links; i++)
-			flush_workqueue(dm->hpd_rx_offload_wq[i].wq);
+	if (!dm->cached_state)
+		return;
+
+	/* Force mode set in atomic commit */
+	for_each_new_crtc_in_state(dm->cached_state, crtc, new_crtc_state, i) {
+		new_crtc_state->active_changed = true;
+		dm_new_crtc_state = to_dm_crtc_state(new_crtc_state);
+		reset_freesync_config_for_crtc(dm_new_crtc_state);
+	}
+
+	/*
+	 * atomic_check is expected to create the dc states. We need to release
+	 * them here, since they were duplicated as part of the suspend
+	 * procedure.
+	 */
+	for_each_new_crtc_in_state(dm->cached_state, crtc, new_crtc_state, i) {
+		dm_new_crtc_state = to_dm_crtc_state(new_crtc_state);
+		if (dm_new_crtc_state->stream) {
+			WARN_ON(kref_read(&dm_new_crtc_state->stream->refcount) > 1);
+			dc_stream_release(dm_new_crtc_state->stream);
+			dm_new_crtc_state->stream = NULL;
+		}
+		dm_new_crtc_state->base.color_mgmt_changed = true;
 	}
+
+	for_each_new_plane_in_state(dm->cached_state, plane, new_plane_state, i) {
+		dm_new_plane_state = to_dm_plane_state(new_plane_state);
+		if (dm_new_plane_state->dc_state) {
+			WARN_ON(kref_read(&dm_new_plane_state->dc_state->refcount) > 1);
+			dc_plane_state_release(dm_new_plane_state->dc_state);
+			dm_new_plane_state->dc_state = NULL;
+		}
+	}
+
+	drm_atomic_helper_resume(ddev, dm->cached_state);
+
+	dm->cached_state = NULL;
 }
 
-static int dm_suspend(void *handle)
+static int dm_suspend(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	struct amdgpu_display_manager *dm = &adev->dm;
-	int ret = 0;
 
 	if (amdgpu_in_reset(adev)) {
+		enum dc_status res;
+
 		mutex_lock(&dm->dc_lock);
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 		dc_allow_idle_optimizations(adev->dm.dc, false);
-#endif
 
-		dm->cached_dc_state = dc_copy_state(dm->dc->current_state);
+		dm->cached_dc_state = dc_state_create_copy(dm->dc->current_state);
 
-		dm_gpureset_toggle_interrupts(adev, dm->cached_dc_state, false);
+		if (dm->cached_dc_state)
+			dm_gpureset_toggle_interrupts(adev, dm->cached_dc_state, false);
 
-		amdgpu_dm_commit_zero_streams(dm->dc);
+		res = amdgpu_dm_commit_zero_streams(dm->dc);
+		if (res != DC_OK) {
+			drm_err(adev_to_drm(adev), "Failed to commit zero streams: %d\n", res);
+			return -EINVAL;
+		}
 
 		amdgpu_dm_irq_suspend(adev);
 
 		hpd_rx_irq_work_suspend(dm);
 
-		return ret;
+		return 0;
 	}
 
-	WARN_ON(adev->dm.cached_state);
-	adev->dm.cached_state = drm_atomic_helper_suspend(adev_to_drm(adev));
+	if (!adev->dm.cached_state) {
+		int r = dm_cache_state(adev);
+
+		if (r)
+			return r;
+	}
+
+	s3_handle_hdmi_cec(adev_to_drm(adev), true);
 
 	s3_handle_mst(adev_to_drm(adev), true);
 
@@ -2373,14 +3226,19 @@ static int dm_suspend(void *handle)
 
 	dc_set_power_state(dm->dc, DC_ACPI_CM_POWER_STATE_D3);
 
+	if (dm->dc->caps.ips_support && adev->in_s0ix)
+		dc_allow_idle_optimizations(dm->dc, true);
+
+	dc_dmub_srv_set_power_state(dm->dc->ctx->dmub_srv, DC_ACPI_CM_POWER_STATE_D3);
+
 	return 0;
 }
 
-static struct amdgpu_dm_connector *
+struct drm_connector *
 amdgpu_dm_find_first_crtc_matching_connector(struct drm_atomic_state *state,
 					     struct drm_crtc *crtc)
 {
-	uint32_t i;
+	u32 i;
 	struct drm_connector_state *new_con_state;
 	struct drm_connector *connector;
 	struct drm_crtc *crtc_from_state;
@@ -2389,7 +3247,7 @@ amdgpu_dm_find_first_crtc_matching_connector(struct drm_atomic_state *state,
 		crtc_from_state = new_con_state->crtc;
 
 		if (crtc_from_state == crtc)
-			return to_amdgpu_dm_connector(connector);
+			return connector;
 	}
 
 	return NULL;
@@ -2401,6 +3259,7 @@ static void emulated_link_detect(struct dc_link *link)
 	struct display_sink_capability sink_caps = { 0 };
 	enum dc_edid_status edid_status;
 	struct dc_context *dc_ctx = link->ctx;
+	struct drm_device *dev = adev_to_drm(dc_ctx->driver_context);
 	struct dc_sink *sink = NULL;
 	struct dc_sink *prev_sink = NULL;
 
@@ -2450,7 +3309,7 @@ static void emulated_link_detect(struct dc_link *link)
 	}
 
 	default:
-		DC_ERROR("Invalid connector type! signal:%d\n",
+		drm_err(dev, "Invalid connector type! signal:%d\n",
 			link->connector_signal);
 		return;
 	}
@@ -2460,7 +3319,7 @@ static void emulated_link_detect(struct dc_link *link)
 
 	sink = dc_sink_create(&sink_init_data);
 	if (!sink) {
-		DC_ERROR("Failed to create sink!\n");
+		drm_err(dev, "Failed to create sink!\n");
 		return;
 	}
 
@@ -2473,7 +3332,7 @@ static void emulated_link_detect(struct dc_link *link)
 			sink);
 
 	if (edid_status != EDID_OK)
-		DC_ERROR("Failed to read EDID");
+		drm_err(dev, "Failed to read EDID\n");
 
 }
 
@@ -2486,83 +3345,131 @@ static void dm_gpureset_commit_state(struct dc_state *dc_state,
 		struct dc_scaling_info scaling_infos[MAX_SURFACES];
 		struct dc_flip_addrs flip_addrs[MAX_SURFACES];
 		struct dc_stream_update stream_update;
-	} * bundle;
+	} *bundle __free(kfree);
 	int k, m;
 
 	bundle = kzalloc(sizeof(*bundle), GFP_KERNEL);
 
 	if (!bundle) {
-		dm_error("Failed to allocate update bundle\n");
-		goto cleanup;
+		drm_err(dm->ddev, "Failed to allocate update bundle\n");
+		return;
 	}
 
 	for (k = 0; k < dc_state->stream_count; k++) {
 		bundle->stream_update.stream = dc_state->streams[k];
 
-		for (m = 0; m < dc_state->stream_status->plane_count; m++) {
+		for (m = 0; m < dc_state->stream_status[k].plane_count; m++) {
 			bundle->surface_updates[m].surface =
-				dc_state->stream_status->plane_states[m];
+				dc_state->stream_status[k].plane_states[m];
 			bundle->surface_updates[m].surface->force_full_update =
 				true;
 		}
-		dc_commit_updates_for_stream(
-			dm->dc, bundle->surface_updates,
-			dc_state->stream_status->plane_count,
-			dc_state->streams[k], &bundle->stream_update, dc_state);
+
+		update_planes_and_stream_adapter(dm->dc,
+					 UPDATE_TYPE_FULL,
+					 dc_state->stream_status[k].plane_count,
+					 dc_state->streams[k],
+					 &bundle->stream_update,
+					 bundle->surface_updates);
 	}
+}
 
-cleanup:
-	kfree(bundle);
+static void apply_delay_after_dpcd_poweroff(struct amdgpu_device *adev,
+					    struct dc_sink *sink)
+{
+	struct dc_panel_patch *ppatch = NULL;
 
-	return;
+	if (!sink)
+		return;
+
+	ppatch = &sink->edid_caps.panel_patch;
+	if (ppatch->wait_after_dpcd_poweroff_ms) {
+		msleep(ppatch->wait_after_dpcd_poweroff_ms);
+		drm_dbg_driver(adev_to_drm(adev),
+			       "%s: adding a %ds delay as w/a for panel\n",
+			       __func__,
+			       ppatch->wait_after_dpcd_poweroff_ms / 1000);
+	}
 }
 
-static void dm_set_dpms_off(struct dc_link *link, struct dm_crtc_state *acrtc_state)
+/**
+ * amdgpu_dm_dump_links_and_sinks - Debug dump of all DC links and their sinks
+ * @adev: amdgpu device pointer
+ *
+ * Iterates through all DC links and dumps information about local and remote
+ * (MST) sinks. Should be called after connector detection is complete to see
+ * the final state of all links.
+ */
+static void amdgpu_dm_dump_links_and_sinks(struct amdgpu_device *adev)
 {
-	struct dc_stream_state *stream_state;
-	struct amdgpu_dm_connector *aconnector = link->priv;
-	struct amdgpu_device *adev = drm_to_adev(aconnector->base.dev);
-	struct dc_stream_update stream_update;
-	bool dpms_off = true;
+	struct dc *dc = adev->dm.dc;
+	struct drm_device *dev = adev_to_drm(adev);
+	int li;
 
-	memset(&stream_update, 0, sizeof(stream_update));
-	stream_update.dpms_off = &dpms_off;
+	if (!dc)
+		return;
 
-	mutex_lock(&adev->dm.dc_lock);
-	stream_state = dc_stream_find_from_link(link);
+	for (li = 0; li < dc->link_count; li++) {
+		struct dc_link *l = dc->links[li];
+		const char *name = NULL;
+		int rs;
 
-	if (stream_state == NULL) {
-		DRM_DEBUG_DRIVER("Error finding stream state associated with link!\n");
-		mutex_unlock(&adev->dm.dc_lock);
-		return;
+		if (!l)
+			continue;
+		if (l->local_sink && l->local_sink->edid_caps.display_name[0])
+			name = l->local_sink->edid_caps.display_name;
+		else
+			name = "n/a";
+
+		drm_dbg_kms(dev,
+			"LINK_DUMP[%d]: local_sink=%p type=%d sink_signal=%d sink_count=%u edid_name=%s mst_capable=%d mst_alloc_streams=%d\n",
+			li,
+			l->local_sink,
+			l->type,
+			l->local_sink ? l->local_sink->sink_signal : SIGNAL_TYPE_NONE,
+			l->sink_count,
+			name,
+			l->dpcd_caps.is_mst_capable,
+			l->mst_stream_alloc_table.stream_count);
+
+		/* Dump remote (MST) sinks if any */
+		for (rs = 0; rs < l->sink_count; rs++) {
+			struct dc_sink *rsink = l->remote_sinks[rs];
+			const char *rname = NULL;
+
+			if (!rsink)
+				continue;
+			if (rsink->edid_caps.display_name[0])
+				rname = rsink->edid_caps.display_name;
+			else
+				rname = "n/a";
+			drm_dbg_kms(dev,
+				"  REMOTE_SINK[%d:%d]: sink=%p signal=%d edid_name=%s\n",
+				li, rs,
+				rsink,
+				rsink->sink_signal,
+				rname);
+		}
 	}
-
-	stream_update.stream = stream_state;
-	acrtc_state->force_dpms_off = true;
-	dc_commit_updates_for_stream(stream_state->ctx->dc, NULL, 0,
-				     stream_state, &stream_update,
-				     stream_state->ctx->dc->current_state);
-	mutex_unlock(&adev->dm.dc_lock);
 }
 
-static int dm_resume(void *handle)
+static int dm_resume(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	struct drm_device *ddev = adev_to_drm(adev);
 	struct amdgpu_display_manager *dm = &adev->dm;
 	struct amdgpu_dm_connector *aconnector;
 	struct drm_connector *connector;
 	struct drm_connector_list_iter iter;
-	struct drm_crtc *crtc;
-	struct drm_crtc_state *new_crtc_state;
-	struct dm_crtc_state *dm_new_crtc_state;
-	struct drm_plane *plane;
-	struct drm_plane_state *new_plane_state;
-	struct dm_plane_state *dm_new_plane_state;
 	struct dm_atomic_state *dm_state = to_dm_atomic_state(dm->atomic_obj.state);
 	enum dc_connection_type new_connection_type = dc_connection_none;
 	struct dc_state *dc_state;
 	int i, r, j;
+	struct dc_commit_streams_params commit_params = {};
+
+	if (dm->dc->caps.ips_support) {
+		dc_dmub_srv_apply_idle_power_optimizations(dm->dc, false);
+	}
 
 	if (amdgpu_in_reset(adev)) {
 		dc_state = dm->cached_dc_state;
@@ -2576,19 +3483,23 @@ static int dm_resume(void *handle)
 		 * before the 0 streams commit.
 		 *
 		 * DC expects that link encoder assignments are *not* valid
-		 * when committing a state, so as a workaround it needs to be
-		 * cleared here.
+		 * when committing a state, so as a workaround we can copy
+		 * off of the current state.
+		 *
+		 * We lose the previous assignments, but we had already
+		 * commit 0 streams anyway.
 		 */
-		link_enc_cfg_init(dm->dc, dc_state);
-
-		if (dc_enable_dmub_notifications(adev->dm.dc))
-			amdgpu_dm_outbox_init(adev);
+		link_enc_cfg_copy(adev->dm.dc->current_state, dc_state);
 
 		r = dm_dmub_hw_init(adev);
-		if (r)
-			DRM_ERROR("DMUB interface failed to initialize: status=%d\n", r);
+		if (r) {
+			drm_err(adev_to_drm(adev), "DMUB interface failed to initialize: status=%d\n", r);
+			return r;
+		}
 
+		dc_dmub_srv_set_power_state(dm->dc->ctx->dmub_srv, DC_ACPI_CM_POWER_STATE_D0);
 		dc_set_power_state(dm->dc, DC_ACPI_CM_POWER_STATE_D0);
+
 		dc_resume(dm->dc);
 
 		amdgpu_dm_irq_resume_early(adev);
@@ -2600,47 +3511,52 @@ static int dm_resume(void *handle)
 					= 0xffffffff;
 			}
 		}
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-		/*
-		 * Resource allocation happens for link encoders for newer ASIC in
-		 * dc_validate_global_state, so we need to revalidate it.
-		 *
-		 * This shouldn't fail (it passed once before), so warn if it does.
-		 */
-		WARN_ON(dc_validate_global_state(dm->dc, dc_state, false) != DC_OK);
-#endif
 
-		WARN_ON(!dc_commit_state(dm->dc, dc_state));
+		if (dc_is_dmub_outbox_supported(adev->dm.dc)) {
+			amdgpu_dm_outbox_init(adev);
+			dc_enable_dmub_outbox(adev->dm.dc);
+		}
+
+		commit_params.streams = dc_state->streams;
+		commit_params.stream_count = dc_state->stream_count;
+		dc_exit_ips_for_hw_access(dm->dc);
+		WARN_ON(!dc_commit_streams(dm->dc, &commit_params));
 
 		dm_gpureset_commit_state(dm->cached_dc_state, dm);
 
 		dm_gpureset_toggle_interrupts(adev, dm->cached_dc_state, true);
 
-		dc_release_state(dm->cached_dc_state);
+		dc_state_release(dm->cached_dc_state);
 		dm->cached_dc_state = NULL;
 
 		amdgpu_dm_irq_resume_late(adev);
 
 		mutex_unlock(&dm->dc_lock);
 
+		/* set the backlight after a reset */
+		for (i = 0; i < dm->num_of_edps; i++) {
+			if (dm->backlight_dev[i])
+				amdgpu_dm_backlight_set_level(dm, i, dm->brightness[i]);
+		}
+
 		return 0;
 	}
 	/* Recreate dc_state - DC invalidates it when setting power state to S3. */
-	dc_release_state(dm_state->context);
-	dm_state->context = dc_create_state(dm->dc);
+	dc_state_release(dm_state->context);
+	dm_state->context = dc_state_create(dm->dc, NULL);
 	/* TODO: Remove dc_state->dccg, use dc->dccg directly. */
-	dc_resource_state_construct(dm->dc, dm_state->context);
+
+	/* Before powering on DC we need to re-initialize DMUB. */
+	dm_dmub_hw_resume(adev);
 
 	/* Re-enable outbox interrupts for DPIA. */
-	if (dc_enable_dmub_notifications(adev->dm.dc))
+	if (dc_is_dmub_outbox_supported(adev->dm.dc)) {
 		amdgpu_dm_outbox_init(adev);
-
-	/* Before powering on DC we need to re-initialize DMUB. */
-	r = dm_dmub_hw_init(adev);
-	if (r)
-		DRM_ERROR("DMUB interface failed to initialize: status=%d\n", r);
+		dc_enable_dmub_outbox(adev->dm.dc);
+	}
 
 	/* power on hardware */
+	dc_dmub_srv_set_power_state(dm->dc->ctx->dmub_srv, DC_ACPI_CM_POWER_STATE_D0);
 	dc_set_power_state(dm->dc, DC_ACPI_CM_POWER_STATE_D0);
 
 	/* program HPD filter */
@@ -2652,29 +3568,46 @@ static int dm_resume(void *handle)
 	 */
 	amdgpu_dm_irq_resume_early(adev);
 
+	s3_handle_hdmi_cec(ddev, false);
+
 	/* On resume we need to rewrite the MSTM control bits to enable MST*/
 	s3_handle_mst(ddev, false);
 
 	/* Do detection*/
 	drm_connector_list_iter_begin(ddev, &iter);
 	drm_for_each_connector_iter(connector, &iter) {
+		bool ret;
+
+		if (connector->connector_type == DRM_MODE_CONNECTOR_WRITEBACK)
+			continue;
+
 		aconnector = to_amdgpu_dm_connector(connector);
 
+		if (!aconnector->dc_link)
+			continue;
+
 		/*
-		 * this is the case when traversing through already created
+		 * this is the case when traversing through already created end sink
 		 * MST connectors, should be skipped
 		 */
-		if (aconnector->mst_port)
+		if (aconnector->mst_root)
 			continue;
 
-		mutex_lock(&aconnector->hpd_lock);
-		if (!dc_link_detect_sink(aconnector->dc_link, &new_connection_type))
-			DRM_ERROR("KMS: Failed to detect connector\n");
+		guard(mutex)(&aconnector->hpd_lock);
+		if (!dc_link_detect_connection_type(aconnector->dc_link, &new_connection_type))
+			drm_err(adev_to_drm(adev), "KMS: Failed to detect connector\n");
 
-		if (aconnector->base.force && new_connection_type == dc_connection_none)
+		if (aconnector->base.force && new_connection_type == dc_connection_none) {
 			emulated_link_detect(aconnector->dc_link);
-		else
-			dc_link_detect(aconnector->dc_link, DETECT_REASON_HPD);
+		} else {
+			guard(mutex)(&dm->dc_lock);
+			dc_exit_ips_for_hw_access(dm->dc);
+			ret = dc_link_detect(aconnector->dc_link, DETECT_REASON_RESUMEFROMS3S4);
+			if (ret) {
+				/* w/a delay for certain panels */
+				apply_delay_after_dpcd_poweroff(adev, aconnector->dc_sink);
+			}
+		}
 
 		if (aconnector->fake_enable && aconnector->dc_link->local_sink)
 			aconnector->fake_enable = false;
@@ -2683,45 +3616,47 @@ static int dm_resume(void *handle)
 			dc_sink_release(aconnector->dc_sink);
 		aconnector->dc_sink = NULL;
 		amdgpu_dm_update_connector_after_detect(aconnector);
-		mutex_unlock(&aconnector->hpd_lock);
 	}
 	drm_connector_list_iter_end(&iter);
 
-	/* Force mode set in atomic commit */
-	for_each_new_crtc_in_state(dm->cached_state, crtc, new_crtc_state, i)
-		new_crtc_state->active_changed = true;
+	dm_destroy_cached_state(adev);
 
-	/*
-	 * atomic_check is expected to create the dc states. We need to release
-	 * them here, since they were duplicated as part of the suspend
-	 * procedure.
-	 */
-	for_each_new_crtc_in_state(dm->cached_state, crtc, new_crtc_state, i) {
-		dm_new_crtc_state = to_dm_crtc_state(new_crtc_state);
-		if (dm_new_crtc_state->stream) {
-			WARN_ON(kref_read(&dm_new_crtc_state->stream->refcount) > 1);
-			dc_stream_release(dm_new_crtc_state->stream);
-			dm_new_crtc_state->stream = NULL;
-		}
-	}
+	/* Do mst topology probing after resuming cached state*/
+	drm_connector_list_iter_begin(ddev, &iter);
+	drm_for_each_connector_iter(connector, &iter) {
+		bool init = false;
 
-	for_each_new_plane_in_state(dm->cached_state, plane, new_plane_state, i) {
-		dm_new_plane_state = to_dm_plane_state(new_plane_state);
-		if (dm_new_plane_state->dc_state) {
-			WARN_ON(kref_read(&dm_new_plane_state->dc_state->refcount) > 1);
-			dc_plane_state_release(dm_new_plane_state->dc_state);
-			dm_new_plane_state->dc_state = NULL;
+		if (connector->connector_type == DRM_MODE_CONNECTOR_WRITEBACK)
+			continue;
+
+		aconnector = to_amdgpu_dm_connector(connector);
+		if (aconnector->dc_link->type != dc_connection_mst_branch ||
+		    aconnector->mst_root)
+			continue;
+
+		scoped_guard(mutex, &aconnector->mst_mgr.lock) {
+			init = !aconnector->mst_mgr.mst_primary;
 		}
+		if (init)
+			dm_helpers_dp_mst_start_top_mgr(aconnector->dc_link->ctx,
+				aconnector->dc_link, false);
+		else
+			drm_dp_mst_topology_queue_probe(&aconnector->mst_mgr);
 	}
+	drm_connector_list_iter_end(&iter);
 
-	drm_atomic_helper_resume(ddev, dm->cached_state);
-
-	dm->cached_state = NULL;
+	/* Debug dump: list all DC links and their associated sinks after detection
+	 * is complete for all connectors. This provides a comprehensive view of the
+	 * final state without repeating the dump for each connector.
+	 */
+	amdgpu_dm_dump_links_and_sinks(adev);
 
 	amdgpu_dm_irq_resume_late(adev);
 
 	amdgpu_dm_smu_write_watermarks_table(adev);
 
+	drm_kms_helper_hotplug_event(ddev);
+
 	return 0;
 }
 
@@ -2754,8 +3689,7 @@ static const struct amd_ip_funcs amdgpu_dm_funcs = {
 	.set_powergating_state = dm_set_powergating_state,
 };
 
-const struct amdgpu_ip_block_version dm_ip_block =
-{
+const struct amdgpu_ip_block_version dm_ip_block = {
 	.type = AMD_IP_BLOCK_TYPE_DCE,
 	.major = 1,
 	.minor = 0,
@@ -2772,94 +3706,92 @@ const struct amdgpu_ip_block_version dm_ip_block =
 
 static const struct drm_mode_config_funcs amdgpu_dm_mode_funcs = {
 	.fb_create = amdgpu_display_user_framebuffer_create,
-	.get_format_info = amd_get_format_info,
-	.output_poll_changed = drm_fb_helper_output_poll_changed,
+	.get_format_info = amdgpu_dm_plane_get_format_info,
 	.atomic_check = amdgpu_dm_atomic_check,
 	.atomic_commit = drm_atomic_helper_commit,
 };
 
 static struct drm_mode_config_helper_funcs amdgpu_dm_mode_config_helperfuncs = {
-	.atomic_commit_tail = amdgpu_dm_atomic_commit_tail
+	.atomic_commit_tail = amdgpu_dm_atomic_commit_tail,
+	.atomic_commit_setup = amdgpu_dm_atomic_setup_commit,
 };
 
 static void update_connector_ext_caps(struct amdgpu_dm_connector *aconnector)
 {
-	u32 max_cll, min_cll, max, min, q, r;
+	const struct drm_panel_backlight_quirk *panel_backlight_quirk;
 	struct amdgpu_dm_backlight_caps *caps;
-	struct amdgpu_display_manager *dm;
 	struct drm_connector *conn_base;
 	struct amdgpu_device *adev;
-	struct dc_link *link = NULL;
-	static const u8 pre_computed_values[] = {
-		50, 51, 52, 53, 55, 56, 57, 58, 59, 61, 62, 63, 65, 66, 68, 69,
-		71, 72, 74, 75, 77, 79, 81, 82, 84, 86, 88, 90, 92, 94, 96, 98};
-	int i;
-
-	if (!aconnector || !aconnector->dc_link)
-		return;
+	struct drm_luminance_range_info *luminance_range;
+	struct drm_device *drm;
 
-	link = aconnector->dc_link;
-	if (link->connector_signal != SIGNAL_TYPE_EDP)
+	if (aconnector->bl_idx == -1 ||
+	    aconnector->dc_link->connector_signal != SIGNAL_TYPE_EDP)
 		return;
 
 	conn_base = &aconnector->base;
-	adev = drm_to_adev(conn_base->dev);
-	dm = &adev->dm;
-	for (i = 0; i < dm->num_of_edps; i++) {
-		if (link == dm->backlight_link[i])
-			break;
-	}
-	if (i >= dm->num_of_edps)
-		return;
-	caps = &dm->backlight_caps[i];
+	drm = conn_base->dev;
+	adev = drm_to_adev(drm);
+
+	caps = &adev->dm.backlight_caps[aconnector->bl_idx];
 	caps->ext_caps = &aconnector->dc_link->dpcd_sink_ext_caps;
 	caps->aux_support = false;
-	max_cll = conn_base->hdr_sink_metadata.hdmi_type1.max_cll;
-	min_cll = conn_base->hdr_sink_metadata.hdmi_type1.min_cll;
 
-	if (caps->ext_caps->bits.oled == 1 /*||
-	    caps->ext_caps->bits.sdr_aux_backlight_control == 1 ||
-	    caps->ext_caps->bits.hdr_aux_backlight_control == 1*/)
+	if (caps->ext_caps->bits.oled == 1
+	    /*
+	     * ||
+	     * caps->ext_caps->bits.sdr_aux_backlight_control == 1 ||
+	     * caps->ext_caps->bits.hdr_aux_backlight_control == 1
+	     */)
 		caps->aux_support = true;
 
 	if (amdgpu_backlight == 0)
 		caps->aux_support = false;
 	else if (amdgpu_backlight == 1)
 		caps->aux_support = true;
+	if (caps->aux_support)
+		aconnector->dc_link->backlight_control_type = BACKLIGHT_CONTROL_AMD_AUX;
 
-	/* From the specification (CTA-861-G), for calculating the maximum
-	 * luminance we need to use:
-	 *	Luminance = 50*2**(CV/32)
-	 * Where CV is a one-byte value.
-	 * For calculating this expression we may need float point precision;
-	 * to avoid this complexity level, we take advantage that CV is divided
-	 * by a constant. From the Euclids division algorithm, we know that CV
-	 * can be written as: CV = 32*q + r. Next, we replace CV in the
-	 * Luminance expression and get 50*(2**q)*(2**(r/32)), hence we just
-	 * need to pre-compute the value of r/32. For pre-computing the values
-	 * We just used the following Ruby line:
-	 *	(0...32).each {|cv| puts (50*2**(cv/32.0)).round}
-	 * The results of the above expressions can be verified at
-	 * pre_computed_values.
-	 */
-	q = max_cll >> 5;
-	r = max_cll % 32;
-	max = (1 << q) * pre_computed_values[r];
+	luminance_range = &conn_base->display_info.luminance_range;
 
-	// min luminance: maxLum * (CV/255)^2 / 100
-	q = DIV_ROUND_CLOSEST(min_cll, 255);
-	min = max * DIV_ROUND_CLOSEST((q * q), 100);
+	if (luminance_range->max_luminance)
+		caps->aux_max_input_signal = luminance_range->max_luminance;
+	else
+		caps->aux_max_input_signal = 512;
+
+	if (luminance_range->min_luminance)
+		caps->aux_min_input_signal = luminance_range->min_luminance;
+	else
+		caps->aux_min_input_signal = 1;
 
-	caps->aux_max_input_signal = max;
-	caps->aux_min_input_signal = min;
+	panel_backlight_quirk =
+		drm_get_panel_backlight_quirk(aconnector->drm_edid);
+	if (!IS_ERR_OR_NULL(panel_backlight_quirk)) {
+		if (panel_backlight_quirk->min_brightness) {
+			caps->min_input_signal =
+				panel_backlight_quirk->min_brightness - 1;
+			drm_info(drm,
+				 "Applying panel backlight quirk, min_brightness: %d\n",
+				 caps->min_input_signal);
+		}
+		if (panel_backlight_quirk->brightness_mask) {
+			drm_info(drm,
+				 "Applying panel backlight quirk, brightness_mask: 0x%X\n",
+				 panel_backlight_quirk->brightness_mask);
+			caps->brightness_mask =
+				panel_backlight_quirk->brightness_mask;
+		}
+	}
 }
 
+DEFINE_FREE(sink_release, struct dc_sink *, if (_T) dc_sink_release(_T))
+
 void amdgpu_dm_update_connector_after_detect(
 		struct amdgpu_dm_connector *aconnector)
 {
 	struct drm_connector *connector = &aconnector->base;
+	struct dc_sink *sink __free(sink_release) = NULL;
 	struct drm_device *dev = connector->dev;
-	struct dc_sink *sink;
 
 	/* MST handled by drm_mst framework */
 	if (aconnector->mst_mgr.mst_state == true)
@@ -2881,7 +3813,7 @@ void amdgpu_dm_update_connector_after_detect(
 		 * For S3 resume with headless use eml_sink to fake stream
 		 * because on resume connector->sink is set to NULL
 		 */
-		mutex_lock(&dev->mode_config.mutex);
+		guard(mutex)(&dev->mode_config.mutex);
 
 		if (sink) {
 			if (aconnector->dc_sink) {
@@ -2897,7 +3829,7 @@ void amdgpu_dm_update_connector_after_detect(
 			aconnector->dc_sink = sink;
 			dc_sink_retain(aconnector->dc_sink);
 			amdgpu_dm_update_freesync_caps(connector,
-					aconnector->edid);
+					aconnector->drm_edid);
 		} else {
 			amdgpu_dm_update_freesync_caps(connector, NULL);
 			if (!aconnector->dc_sink) {
@@ -2906,10 +3838,6 @@ void amdgpu_dm_update_connector_after_detect(
 			}
 		}
 
-		mutex_unlock(&dev->mode_config.mutex);
-
-		if (sink)
-			dc_sink_release(sink);
 		return;
 	}
 
@@ -2917,27 +3845,25 @@ void amdgpu_dm_update_connector_after_detect(
 	 * TODO: temporary guard to look for proper fix
 	 * if this sink is MST sink, we should not do anything
 	 */
-	if (sink && sink->sink_signal == SIGNAL_TYPE_DISPLAY_PORT_MST) {
-		dc_sink_release(sink);
+	if (sink && sink->sink_signal == SIGNAL_TYPE_DISPLAY_PORT_MST)
 		return;
-	}
 
 	if (aconnector->dc_sink == sink) {
 		/*
 		 * We got a DP short pulse (Link Loss, DP CTS, etc...).
 		 * Do nothing!!
 		 */
-		DRM_DEBUG_DRIVER("DCHPD: connector_id=%d: dc_sink didn't change.\n",
-				aconnector->connector_id);
-		if (sink)
-			dc_sink_release(sink);
+		drm_dbg_kms(dev, "DCHPD: connector_id=%d: dc_sink didn't change.\n",
+				 aconnector->connector_id);
 		return;
 	}
 
-	DRM_DEBUG_DRIVER("DCHPD: connector_id=%d: Old sink=%p New sink=%p\n",
-		aconnector->connector_id, aconnector->dc_sink, sink);
+	drm_dbg_kms(dev, "DCHPD: connector_id=%d: Old sink=%p New sink=%p\n",
+		    aconnector->connector_id, aconnector->dc_sink, sink);
 
-	mutex_lock(&dev->mode_config.mutex);
+	/* When polling, DRM has already locked the mutex for us. */
+	if (!drm_kms_helper_is_poll_worker())
+		mutex_lock(&dev->mode_config.mutex);
 
 	/*
 	 * 1. Update status of the drm connector
@@ -2956,45 +3882,145 @@ void amdgpu_dm_update_connector_after_detect(
 		aconnector->dc_sink = sink;
 		dc_sink_retain(aconnector->dc_sink);
 		if (sink->dc_edid.length == 0) {
-			aconnector->edid = NULL;
+			aconnector->drm_edid = NULL;
+			hdmi_cec_unset_edid(aconnector);
 			if (aconnector->dc_link->aux_mode) {
-				drm_dp_cec_unset_edid(
-					&aconnector->dm_dp_aux.aux);
+				drm_dp_cec_unset_edid(&aconnector->dm_dp_aux.aux);
 			}
 		} else {
-			aconnector->edid =
-				(struct edid *)sink->dc_edid.raw_edid;
+			const struct edid *edid = (const struct edid *)sink->dc_edid.raw_edid;
+
+			aconnector->drm_edid = drm_edid_alloc(edid, sink->dc_edid.length);
+			drm_edid_connector_update(connector, aconnector->drm_edid);
 
-			drm_connector_update_edid_property(connector,
-							   aconnector->edid);
+			hdmi_cec_set_edid(aconnector);
 			if (aconnector->dc_link->aux_mode)
-				drm_dp_cec_set_edid(&aconnector->dm_dp_aux.aux,
-						    aconnector->edid);
+				drm_dp_cec_attach(&aconnector->dm_dp_aux.aux,
+						  connector->display_info.source_physical_address);
+		}
+
+		if (!aconnector->timing_requested) {
+			aconnector->timing_requested =
+				kzalloc(sizeof(struct dc_crtc_timing), GFP_KERNEL);
+			if (!aconnector->timing_requested)
+				drm_err(dev,
+					"failed to create aconnector->requested_timing\n");
 		}
 
-		amdgpu_dm_update_freesync_caps(connector, aconnector->edid);
+		amdgpu_dm_update_freesync_caps(connector, aconnector->drm_edid);
 		update_connector_ext_caps(aconnector);
 	} else {
+		hdmi_cec_unset_edid(aconnector);
 		drm_dp_cec_unset_edid(&aconnector->dm_dp_aux.aux);
 		amdgpu_dm_update_freesync_caps(connector, NULL);
-		drm_connector_update_edid_property(connector, NULL);
 		aconnector->num_modes = 0;
 		dc_sink_release(aconnector->dc_sink);
 		aconnector->dc_sink = NULL;
-		aconnector->edid = NULL;
-#ifdef CONFIG_DRM_AMD_DC_HDCP
+		drm_edid_free(aconnector->drm_edid);
+		aconnector->drm_edid = NULL;
+		kfree(aconnector->timing_requested);
+		aconnector->timing_requested = NULL;
 		/* Set CP to DESIRED if it was ENABLED, so we can re-enable it again on hotplug */
 		if (connector->state->content_protection == DRM_MODE_CONTENT_PROTECTION_ENABLED)
 			connector->state->content_protection = DRM_MODE_CONTENT_PROTECTION_DESIRED;
-#endif
 	}
 
-	mutex_unlock(&dev->mode_config.mutex);
-
 	update_subconnector_property(aconnector);
 
-	if (sink)
-		dc_sink_release(sink);
+	/* When polling, the mutex will be unlocked for us by DRM. */
+	if (!drm_kms_helper_is_poll_worker())
+		mutex_unlock(&dev->mode_config.mutex);
+}
+
+static bool are_sinks_equal(const struct dc_sink *sink1, const struct dc_sink *sink2)
+{
+	if (!sink1 || !sink2)
+		return false;
+	if (sink1->sink_signal != sink2->sink_signal)
+		return false;
+
+	if (sink1->dc_edid.length != sink2->dc_edid.length)
+		return false;
+
+	if (memcmp(sink1->dc_edid.raw_edid, sink2->dc_edid.raw_edid,
+		   sink1->dc_edid.length) != 0)
+		return false;
+	return true;
+}
+
+
+/**
+ * DOC: hdmi_hpd_debounce_work
+ *
+ * HDMI HPD debounce delay in milliseconds. When an HDMI display toggles HPD
+ * (such as during power save transitions), this delay determines how long to
+ * wait before processing the HPD event. This allows distinguishing between a
+ * physical unplug (>hdmi_hpd_debounce_delay)
+ * and a spontaneous RX HPD toggle (<hdmi_hpd_debounce_delay).
+ *
+ * If the toggle is less than this delay, the driver compares sink capabilities
+ * and permits a hotplug event if they changed.
+ *
+ * The default value of 1500ms was chosen based on experimental testing with
+ * various monitors that exhibit spontaneous HPD toggling behavior.
+ */
+static void hdmi_hpd_debounce_work(struct work_struct *work)
+{
+	struct amdgpu_dm_connector *aconnector =
+		container_of(to_delayed_work(work), struct amdgpu_dm_connector,
+			     hdmi_hpd_debounce_work);
+	struct drm_connector *connector = &aconnector->base;
+	struct drm_device *dev = connector->dev;
+	struct amdgpu_device *adev = drm_to_adev(dev);
+	struct dc *dc = aconnector->dc_link->ctx->dc;
+	bool fake_reconnect = false;
+	bool reallow_idle = false;
+	bool ret = false;
+	guard(mutex)(&aconnector->hpd_lock);
+
+	/* Re-detect the display */
+	scoped_guard(mutex, &adev->dm.dc_lock) {
+		if (dc->caps.ips_support && dc->ctx->dmub_srv->idle_allowed) {
+			dc_allow_idle_optimizations(dc, false);
+			reallow_idle = true;
+		}
+		ret = dc_link_detect(aconnector->dc_link, DETECT_REASON_HPD);
+	}
+
+	if (ret) {
+		/* Apply workaround delay for certain panels */
+		apply_delay_after_dpcd_poweroff(adev, aconnector->dc_sink);
+		/* Compare sinks to determine if this was a spontaneous HPD toggle */
+		if (are_sinks_equal(aconnector->dc_link->local_sink, aconnector->hdmi_prev_sink)) {
+			/*
+			* Sinks match - this was a spontaneous HDMI HPD toggle.
+			*/
+			drm_dbg_kms(dev, "HDMI HPD: Sink unchanged after debounce, internal re-enable\n");
+			fake_reconnect = true;
+		}
+
+		/* Update connector state */
+		amdgpu_dm_update_connector_after_detect(aconnector);
+
+		drm_modeset_lock_all(dev);
+		dm_restore_drm_connector_state(dev, connector);
+		drm_modeset_unlock_all(dev);
+
+		/* Only notify OS if sink actually changed */
+		if (!fake_reconnect && aconnector->base.force == DRM_FORCE_UNSPECIFIED)
+			drm_kms_helper_hotplug_event(dev);
+	}
+
+	/* Release the cached sink reference */
+	if (aconnector->hdmi_prev_sink) {
+		dc_sink_release(aconnector->hdmi_prev_sink);
+		aconnector->hdmi_prev_sink = NULL;
+	}
+
+	scoped_guard(mutex, &adev->dm.dc_lock) {
+		if (reallow_idle && dc->caps.ips_support)
+			dc_allow_idle_optimizations(dc, true);
+	}
 }
 
 static void handle_hpd_irq_helper(struct amdgpu_dm_connector *aconnector)
@@ -3004,32 +4030,38 @@ static void handle_hpd_irq_helper(struct amdgpu_dm_connector *aconnector)
 	enum dc_connection_type new_connection_type = dc_connection_none;
 	struct amdgpu_device *adev = drm_to_adev(dev);
 	struct dm_connector_state *dm_con_state = to_dm_connector_state(connector->state);
-	struct dm_crtc_state *dm_crtc_state = NULL;
+	struct dc *dc = aconnector->dc_link->ctx->dc;
+	bool ret = false;
+	bool debounce_required = false;
 
 	if (adev->dm.disable_hpd_irq)
 		return;
 
-	if (dm_con_state->base.state && dm_con_state->base.crtc)
-		dm_crtc_state = to_dm_crtc_state(drm_atomic_get_crtc_state(
-					dm_con_state->base.state,
-					dm_con_state->base.crtc));
 	/*
 	 * In case of failure or MST no need to update connector status or notify the OS
 	 * since (for MST case) MST does this in its own context.
 	 */
-	mutex_lock(&aconnector->hpd_lock);
+	guard(mutex)(&aconnector->hpd_lock);
 
-#ifdef CONFIG_DRM_AMD_DC_HDCP
 	if (adev->dm.hdcp_workqueue) {
 		hdcp_reset_display(adev->dm.hdcp_workqueue, aconnector->dc_link->link_index);
 		dm_con_state->update_hdcp = true;
 	}
-#endif
 	if (aconnector->fake_enable)
 		aconnector->fake_enable = false;
 
-	if (!dc_link_detect_sink(aconnector->dc_link, &new_connection_type))
-		DRM_ERROR("KMS: Failed to detect connector\n");
+	aconnector->timing_changed = false;
+
+	if (!dc_link_detect_connection_type(aconnector->dc_link, &new_connection_type))
+		drm_err(adev_to_drm(adev), "KMS: Failed to detect connector\n");
+
+	/*
+	 * Check for HDMI disconnect with debounce enabled.
+	 */
+	debounce_required = (aconnector->hdmi_hpd_debounce_delay_ms > 0 &&
+			      dc_is_hdmi_signal(aconnector->dc_link->connector_signal) &&
+			      new_connection_type == dc_connection_none &&
+			      aconnector->dc_link->local_sink != NULL);
 
 	if (aconnector->base.force && new_connection_type == dc_connection_none) {
 		emulated_link_detect(aconnector->dc_link);
@@ -3039,25 +4071,52 @@ static void handle_hpd_irq_helper(struct amdgpu_dm_connector *aconnector)
 		drm_modeset_unlock_all(dev);
 
 		if (aconnector->base.force == DRM_FORCE_UNSPECIFIED)
-			drm_kms_helper_hotplug_event(dev);
+			drm_kms_helper_connector_hotplug_event(connector);
+	} else if (debounce_required) {
+		/*
+		 * HDMI disconnect detected - schedule delayed work instead of
+		 * processing immediately. This allows us to coalesce spurious
+		 * HDMI signals from physical unplugs.
+		 */
+		drm_dbg_kms(dev, "HDMI HPD: Disconnect detected, scheduling debounce work (%u ms)\n",
+			    aconnector->hdmi_hpd_debounce_delay_ms);
+
+		/* Cache the current sink for later comparison */
+		if (aconnector->hdmi_prev_sink)
+			dc_sink_release(aconnector->hdmi_prev_sink);
+		aconnector->hdmi_prev_sink = aconnector->dc_link->local_sink;
+		if (aconnector->hdmi_prev_sink)
+			dc_sink_retain(aconnector->hdmi_prev_sink);
+
+		/* Schedule delayed detection. */
+		if (mod_delayed_work(system_wq,
+				 &aconnector->hdmi_hpd_debounce_work,
+				 msecs_to_jiffies(aconnector->hdmi_hpd_debounce_delay_ms)))
+			drm_dbg_kms(dev, "HDMI HPD: Re-scheduled debounce work\n");
 
-	} else if (dc_link_detect(aconnector->dc_link, DETECT_REASON_HPD)) {
-		if (new_connection_type == dc_connection_none &&
-		    aconnector->dc_link->type == dc_connection_none &&
-		    dm_crtc_state)
-			dm_set_dpms_off(aconnector->dc_link, dm_crtc_state);
+	} else {
 
-		amdgpu_dm_update_connector_after_detect(aconnector);
+		/* If the aconnector->hdmi_hpd_debounce_work is scheduled, exit early */
+		if (delayed_work_pending(&aconnector->hdmi_hpd_debounce_work))
+			return;
 
-		drm_modeset_lock_all(dev);
-		dm_restore_drm_connector_state(dev, connector);
-		drm_modeset_unlock_all(dev);
+		scoped_guard(mutex, &adev->dm.dc_lock) {
+			dc_exit_ips_for_hw_access(dc);
+			ret = dc_link_detect(aconnector->dc_link, DETECT_REASON_HPD);
+		}
+		if (ret) {
+			/* w/a delay for certain panels */
+			apply_delay_after_dpcd_poweroff(adev, aconnector->dc_sink);
+			amdgpu_dm_update_connector_after_detect(aconnector);
 
-		if (aconnector->base.force == DRM_FORCE_UNSPECIFIED)
-			drm_kms_helper_hotplug_event(dev);
-	}
-	mutex_unlock(&aconnector->hpd_lock);
+			drm_modeset_lock_all(dev);
+			dm_restore_drm_connector_state(dev, connector);
+			drm_modeset_unlock_all(dev);
 
+			if (aconnector->base.force == DRM_FORCE_UNSPECIFIED)
+				drm_kms_helper_connector_hotplug_event(connector);
+		}
+	}
 }
 
 static void handle_hpd_irq(void *param)
@@ -3068,98 +4127,21 @@ static void handle_hpd_irq(void *param)
 
 }
 
-static void dm_handle_mst_sideband_msg(struct amdgpu_dm_connector *aconnector)
-{
-	uint8_t esi[DP_PSR_ERROR_STATUS - DP_SINK_COUNT_ESI] = { 0 };
-	uint8_t dret;
-	bool new_irq_handled = false;
-	int dpcd_addr;
-	int dpcd_bytes_to_read;
-
-	const int max_process_count = 30;
-	int process_count = 0;
-
-	const struct dc_link_status *link_status = dc_link_get_status(aconnector->dc_link);
-
-	if (link_status->dpcd_caps->dpcd_rev.raw < 0x12) {
-		dpcd_bytes_to_read = DP_LANE0_1_STATUS - DP_SINK_COUNT;
-		/* DPCD 0x200 - 0x201 for downstream IRQ */
-		dpcd_addr = DP_SINK_COUNT;
-	} else {
-		dpcd_bytes_to_read = DP_PSR_ERROR_STATUS - DP_SINK_COUNT_ESI;
-		/* DPCD 0x2002 - 0x2005 for downstream IRQ */
-		dpcd_addr = DP_SINK_COUNT_ESI;
-	}
-
-	dret = drm_dp_dpcd_read(
-		&aconnector->dm_dp_aux.aux,
-		dpcd_addr,
-		esi,
-		dpcd_bytes_to_read);
-
-	while (dret == dpcd_bytes_to_read &&
-		process_count < max_process_count) {
-		uint8_t retry;
-		dret = 0;
-
-		process_count++;
-
-		DRM_DEBUG_DRIVER("ESI %02x %02x %02x\n", esi[0], esi[1], esi[2]);
-		/* handle HPD short pulse irq */
-		if (aconnector->mst_mgr.mst_state)
-			drm_dp_mst_hpd_irq(
-				&aconnector->mst_mgr,
-				esi,
-				&new_irq_handled);
-
-		if (new_irq_handled) {
-			/* ACK at DPCD to notify down stream */
-			const int ack_dpcd_bytes_to_write =
-				dpcd_bytes_to_read - 1;
-
-			for (retry = 0; retry < 3; retry++) {
-				uint8_t wret;
-
-				wret = drm_dp_dpcd_write(
-					&aconnector->dm_dp_aux.aux,
-					dpcd_addr + 1,
-					&esi[1],
-					ack_dpcd_bytes_to_write);
-				if (wret == ack_dpcd_bytes_to_write)
-					break;
-			}
-
-			/* check if there is new irq to be handled */
-			dret = drm_dp_dpcd_read(
-				&aconnector->dm_dp_aux.aux,
-				dpcd_addr,
-				esi,
-				dpcd_bytes_to_read);
-
-			new_irq_handled = false;
-		} else {
-			break;
-		}
-	}
-
-	if (process_count == max_process_count)
-		DRM_DEBUG_DRIVER("Loop exceeded max iterations\n");
-}
-
-static void schedule_hpd_rx_offload_work(struct hpd_rx_irq_offload_work_queue *offload_wq,
+static void schedule_hpd_rx_offload_work(struct amdgpu_device *adev, struct hpd_rx_irq_offload_work_queue *offload_wq,
 							union hpd_irq_data hpd_irq_data)
 {
 	struct hpd_rx_irq_offload_work *offload_work =
 				kzalloc(sizeof(*offload_work), GFP_KERNEL);
 
 	if (!offload_work) {
-		DRM_ERROR("Failed to allocate hpd_rx_irq_offload_work.\n");
+		drm_err(adev_to_drm(adev), "Failed to allocate hpd_rx_irq_offload_work.\n");
 		return;
 	}
 
 	INIT_WORK(&offload_work->work, dm_handle_hpd_rx_offload_work);
 	offload_work->data = hpd_irq_data;
 	offload_work->offload_wq = offload_wq;
+	offload_work->adev = adev;
 
 	queue_work(offload_wq->wq, &offload_work->work);
 	DRM_DEBUG_KMS("queue work to handle hpd_rx offload work");
@@ -3178,8 +4160,9 @@ static void handle_hpd_rx_irq(void *param)
 	union hpd_irq_data hpd_irq_data;
 	bool link_loss = false;
 	bool has_left_work = false;
-	int idx = aconnector->base.index;
+	int idx = dc_link->link_index;
 	struct hpd_rx_irq_offload_work_queue *offload_wq = &adev->dm.hpd_rx_offload_wq[idx];
+	struct dc *dc = aconnector->dc_link->ctx->dc;
 
 	memset(&hpd_irq_data, 0, sizeof(hpd_irq_data));
 
@@ -3200,14 +4183,30 @@ static void handle_hpd_rx_irq(void *param)
 		goto out;
 
 	if (hpd_irq_data.bytes.device_service_irq.bits.AUTOMATED_TEST) {
-		schedule_hpd_rx_offload_work(offload_wq, hpd_irq_data);
+		schedule_hpd_rx_offload_work(adev, offload_wq, hpd_irq_data);
 		goto out;
 	}
 
 	if (dc_link_dp_allow_hpd_rx_irq(dc_link)) {
 		if (hpd_irq_data.bytes.device_service_irq.bits.UP_REQ_MSG_RDY ||
 			hpd_irq_data.bytes.device_service_irq.bits.DOWN_REP_MSG_RDY) {
-			dm_handle_mst_sideband_msg(aconnector);
+			bool skip = false;
+
+			/*
+			 * DOWN_REP_MSG_RDY is also handled by polling method
+			 * mgr->cbs->poll_hpd_irq()
+			 */
+			spin_lock(&offload_wq->offload_lock);
+			skip = offload_wq->is_handling_mst_msg_rdy_event;
+
+			if (!skip)
+				offload_wq->is_handling_mst_msg_rdy_event = true;
+
+			spin_unlock(&offload_wq->offload_lock);
+
+			if (!skip)
+				schedule_hpd_rx_offload_work(adev, offload_wq, hpd_irq_data);
+
 			goto out;
 		}
 
@@ -3223,7 +4222,7 @@ static void handle_hpd_rx_irq(void *param)
 			spin_unlock(&offload_wq->offload_lock);
 
 			if (!skip)
-				schedule_hpd_rx_offload_work(offload_wq, hpd_irq_data);
+				schedule_hpd_rx_offload_work(adev, offload_wq, hpd_irq_data);
 
 			goto out;
 		}
@@ -3232,8 +4231,8 @@ static void handle_hpd_rx_irq(void *param)
 out:
 	if (result && !is_mst_root_connector) {
 		/* Downstream Port status changed. */
-		if (!dc_link_detect_sink(dc_link, &new_connection_type))
-			DRM_ERROR("KMS: Failed to detect connector\n");
+		if (!dc_link_detect_connection_type(dc_link, &new_connection_type))
+			drm_err(adev_to_drm(adev), "KMS: Failed to detect connector\n");
 
 		if (aconnector->base.force && new_connection_type == dc_connection_none) {
 			emulated_link_detect(dc_link);
@@ -3248,28 +4247,33 @@ out:
 			dm_restore_drm_connector_state(dev, connector);
 			drm_modeset_unlock_all(dev);
 
-			drm_kms_helper_hotplug_event(dev);
-		} else if (dc_link_detect(dc_link, DETECT_REASON_HPDRX)) {
+			drm_kms_helper_connector_hotplug_event(connector);
+		} else {
+			bool ret = false;
 
-			if (aconnector->fake_enable)
-				aconnector->fake_enable = false;
+			mutex_lock(&adev->dm.dc_lock);
+			dc_exit_ips_for_hw_access(dc);
+			ret = dc_link_detect(dc_link, DETECT_REASON_HPDRX);
+			mutex_unlock(&adev->dm.dc_lock);
 
-			amdgpu_dm_update_connector_after_detect(aconnector);
+			if (ret) {
+				if (aconnector->fake_enable)
+					aconnector->fake_enable = false;
 
+				amdgpu_dm_update_connector_after_detect(aconnector);
 
-			drm_modeset_lock_all(dev);
-			dm_restore_drm_connector_state(dev, connector);
-			drm_modeset_unlock_all(dev);
+				drm_modeset_lock_all(dev);
+				dm_restore_drm_connector_state(dev, connector);
+				drm_modeset_unlock_all(dev);
 
-			drm_kms_helper_hotplug_event(dev);
+				drm_kms_helper_connector_hotplug_event(connector);
+			}
 		}
 	}
-#ifdef CONFIG_DRM_AMD_DC_HDCP
 	if (hpd_irq_data.bytes.device_service_irq.bits.CP_IRQ) {
 		if (adev->dm.hdcp_workqueue)
 			hdcp_handle_cpirq(adev->dm.hdcp_workqueue,  aconnector->base.index);
 	}
-#endif
 
 	if (dc_link->type != dc_connection_mst_branch)
 		drm_dp_cec_irq(&aconnector->dm_dp_aux.aux);
@@ -3277,7 +4281,7 @@ out:
 	mutex_unlock(&aconnector->hpd_lock);
 }
 
-static void register_hpd_handlers(struct amdgpu_device *adev)
+static int register_hpd_handlers(struct amdgpu_device *adev)
 {
 	struct drm_device *dev = adev_to_drm(adev);
 	struct drm_connector *connector;
@@ -3288,36 +4292,70 @@ static void register_hpd_handlers(struct amdgpu_device *adev)
 	int_params.requested_polarity = INTERRUPT_POLARITY_DEFAULT;
 	int_params.current_polarity = INTERRUPT_POLARITY_DEFAULT;
 
+	if (dc_is_dmub_outbox_supported(adev->dm.dc)) {
+		if (!register_dmub_notify_callback(adev, DMUB_NOTIFICATION_HPD,
+			dmub_hpd_callback, true)) {
+			drm_err(adev_to_drm(adev), "fail to register dmub hpd callback");
+			return -EINVAL;
+		}
+
+		if (!register_dmub_notify_callback(adev, DMUB_NOTIFICATION_HPD_IRQ,
+			dmub_hpd_callback, true)) {
+			drm_err(adev_to_drm(adev), "fail to register dmub hpd callback");
+			return -EINVAL;
+		}
+
+		if (!register_dmub_notify_callback(adev, DMUB_NOTIFICATION_HPD_SENSE_NOTIFY,
+			dmub_hpd_sense_callback, true)) {
+			drm_err(adev_to_drm(adev), "fail to register dmub hpd sense callback");
+			return -EINVAL;
+		}
+	}
+
 	list_for_each_entry(connector,
 			&dev->mode_config.connector_list, head)	{
 
+		if (connector->connector_type == DRM_MODE_CONNECTOR_WRITEBACK)
+			continue;
+
 		aconnector = to_amdgpu_dm_connector(connector);
 		dc_link = aconnector->dc_link;
 
-		if (DC_IRQ_SOURCE_INVALID != dc_link->irq_source_hpd) {
+		if (dc_link->irq_source_hpd != DC_IRQ_SOURCE_INVALID) {
 			int_params.int_context = INTERRUPT_LOW_IRQ_CONTEXT;
 			int_params.irq_source = dc_link->irq_source_hpd;
 
-			amdgpu_dm_irq_register_interrupt(adev, &int_params,
-					handle_hpd_irq,
-					(void *) aconnector);
+			if (int_params.irq_source == DC_IRQ_SOURCE_INVALID ||
+				int_params.irq_source  < DC_IRQ_SOURCE_HPD1 ||
+				int_params.irq_source  > DC_IRQ_SOURCE_HPD6) {
+				drm_err(adev_to_drm(adev), "Failed to register hpd irq!\n");
+				return -EINVAL;
+			}
+
+			if (!amdgpu_dm_irq_register_interrupt(adev, &int_params,
+				handle_hpd_irq, (void *) aconnector))
+				return -ENOMEM;
 		}
 
-		if (DC_IRQ_SOURCE_INVALID != dc_link->irq_source_hpd_rx) {
+		if (dc_link->irq_source_hpd_rx != DC_IRQ_SOURCE_INVALID) {
 
 			/* Also register for DP short pulse (hpd_rx). */
 			int_params.int_context = INTERRUPT_LOW_IRQ_CONTEXT;
 			int_params.irq_source =	dc_link->irq_source_hpd_rx;
 
-			amdgpu_dm_irq_register_interrupt(adev, &int_params,
-					handle_hpd_rx_irq,
-					(void *) aconnector);
+			if (int_params.irq_source == DC_IRQ_SOURCE_INVALID ||
+				int_params.irq_source  < DC_IRQ_SOURCE_HPD1RX ||
+				int_params.irq_source  > DC_IRQ_SOURCE_HPD6RX) {
+				drm_err(adev_to_drm(adev), "Failed to register hpd rx irq!\n");
+				return -EINVAL;
+			}
 
-			if (adev->dm.hpd_rx_offload_wq)
-				adev->dm.hpd_rx_offload_wq[connector->index].aconnector =
-					aconnector;
+			if (!amdgpu_dm_irq_register_interrupt(adev, &int_params,
+				handle_hpd_rx_irq, (void *) aconnector))
+				return -ENOMEM;
 		}
 	}
+	return 0;
 }
 
 #if defined(CONFIG_DRM_AMD_DC_SI)
@@ -3329,7 +4367,7 @@ static int dce60_register_irq_handlers(struct amdgpu_device *adev)
 	struct dc_interrupt_params int_params = {0};
 	int r;
 	int i;
-	unsigned client_id = AMDGPU_IRQ_CLIENTID_LEGACY;
+	unsigned int client_id = AMDGPU_IRQ_CLIENTID_LEGACY;
 
 	int_params.requested_polarity = INTERRUPT_POLARITY_DEFAULT;
 	int_params.current_polarity = INTERRUPT_POLARITY_DEFAULT;
@@ -3343,27 +4381,36 @@ static int dce60_register_irq_handlers(struct amdgpu_device *adev)
 	 *    Base driver will call amdgpu_dm_irq_handler() for ALL interrupts
 	 *    coming from DC hardware.
 	 *    amdgpu_dm_irq_handler() will re-direct the interrupt to DC
-	 *    for acknowledging and handling. */
+	 *    for acknowledging and handling.
+	 */
 
 	/* Use VBLANK interrupt */
 	for (i = 0; i < adev->mode_info.num_crtc; i++) {
-		r = amdgpu_irq_add_id(adev, client_id, i+1 , &adev->crtc_irq);
+		r = amdgpu_irq_add_id(adev, client_id, i + 1, &adev->crtc_irq);
 		if (r) {
-			DRM_ERROR("Failed to add crtc irq id!\n");
+			drm_err(adev_to_drm(adev), "Failed to add crtc irq id!\n");
 			return r;
 		}
 
 		int_params.int_context = INTERRUPT_HIGH_IRQ_CONTEXT;
 		int_params.irq_source =
-			dc_interrupt_to_irq_source(dc, i+1 , 0);
+			dc_interrupt_to_irq_source(dc, i + 1, 0);
+
+		if (int_params.irq_source == DC_IRQ_SOURCE_INVALID ||
+			int_params.irq_source  < DC_IRQ_SOURCE_VBLANK1 ||
+			int_params.irq_source  > DC_IRQ_SOURCE_VBLANK6) {
+			drm_err(adev_to_drm(adev), "Failed to register vblank irq!\n");
+			return -EINVAL;
+		}
 
 		c_irq_params = &adev->dm.vblank_params[int_params.irq_source - DC_IRQ_SOURCE_VBLANK1];
 
 		c_irq_params->adev = adev;
 		c_irq_params->irq_src = int_params.irq_source;
 
-		amdgpu_dm_irq_register_interrupt(adev, &int_params,
-				dm_crtc_high_irq, c_irq_params);
+		if (!amdgpu_dm_irq_register_interrupt(adev, &int_params,
+			dm_crtc_high_irq, c_irq_params))
+			return -ENOMEM;
 	}
 
 	/* Use GRPH_PFLIP interrupt */
@@ -3371,7 +4418,7 @@ static int dce60_register_irq_handlers(struct amdgpu_device *adev)
 			i <= VISLANDS30_IV_SRCID_D6_GRPH_PFLIP; i += 2) {
 		r = amdgpu_irq_add_id(adev, client_id, i, &adev->pageflip_irq);
 		if (r) {
-			DRM_ERROR("Failed to add page flip irq id!\n");
+			drm_err(adev_to_drm(adev), "Failed to add page flip irq id!\n");
 			return r;
 		}
 
@@ -3379,27 +4426,34 @@ static int dce60_register_irq_handlers(struct amdgpu_device *adev)
 		int_params.irq_source =
 			dc_interrupt_to_irq_source(dc, i, 0);
 
+		if (int_params.irq_source == DC_IRQ_SOURCE_INVALID ||
+			int_params.irq_source  < DC_IRQ_SOURCE_PFLIP_FIRST ||
+			int_params.irq_source  > DC_IRQ_SOURCE_PFLIP_LAST) {
+			drm_err(adev_to_drm(adev), "Failed to register pflip irq!\n");
+			return -EINVAL;
+		}
+
 		c_irq_params = &adev->dm.pflip_params[int_params.irq_source - DC_IRQ_SOURCE_PFLIP_FIRST];
 
 		c_irq_params->adev = adev;
 		c_irq_params->irq_src = int_params.irq_source;
 
-		amdgpu_dm_irq_register_interrupt(adev, &int_params,
-				dm_pflip_high_irq, c_irq_params);
-
+		if (!amdgpu_dm_irq_register_interrupt(adev, &int_params,
+			dm_pflip_high_irq, c_irq_params))
+			return -ENOMEM;
 	}
 
 	/* HPD */
 	r = amdgpu_irq_add_id(adev, client_id,
 			VISLANDS30_IV_SRCID_HOTPLUG_DETECT_A, &adev->hpd_irq);
 	if (r) {
-		DRM_ERROR("Failed to add hpd irq id!\n");
+		drm_err(adev_to_drm(adev), "Failed to add hpd irq id!\n");
 		return r;
 	}
 
-	register_hpd_handlers(adev);
+	r = register_hpd_handlers(adev);
 
-	return 0;
+	return r;
 }
 #endif
 
@@ -3411,7 +4465,7 @@ static int dce110_register_irq_handlers(struct amdgpu_device *adev)
 	struct dc_interrupt_params int_params = {0};
 	int r;
 	int i;
-	unsigned client_id = AMDGPU_IRQ_CLIENTID_LEGACY;
+	unsigned int client_id = AMDGPU_IRQ_CLIENTID_LEGACY;
 
 	if (adev->family >= AMDGPU_FAMILY_AI)
 		client_id = SOC15_IH_CLIENTID_DCE;
@@ -3428,13 +4482,14 @@ static int dce110_register_irq_handlers(struct amdgpu_device *adev)
 	 *    Base driver will call amdgpu_dm_irq_handler() for ALL interrupts
 	 *    coming from DC hardware.
 	 *    amdgpu_dm_irq_handler() will re-direct the interrupt to DC
-	 *    for acknowledging and handling. */
+	 *    for acknowledging and handling.
+	 */
 
 	/* Use VBLANK interrupt */
 	for (i = VISLANDS30_IV_SRCID_D1_VERTICAL_INTERRUPT0; i <= VISLANDS30_IV_SRCID_D6_VERTICAL_INTERRUPT0; i++) {
 		r = amdgpu_irq_add_id(adev, client_id, i, &adev->crtc_irq);
 		if (r) {
-			DRM_ERROR("Failed to add crtc irq id!\n");
+			drm_err(adev_to_drm(adev), "Failed to add crtc irq id!\n");
 			return r;
 		}
 
@@ -3442,20 +4497,28 @@ static int dce110_register_irq_handlers(struct amdgpu_device *adev)
 		int_params.irq_source =
 			dc_interrupt_to_irq_source(dc, i, 0);
 
+		if (int_params.irq_source == DC_IRQ_SOURCE_INVALID ||
+			int_params.irq_source  < DC_IRQ_SOURCE_VBLANK1 ||
+			int_params.irq_source  > DC_IRQ_SOURCE_VBLANK6) {
+			drm_err(adev_to_drm(adev), "Failed to register vblank irq!\n");
+			return -EINVAL;
+		}
+
 		c_irq_params = &adev->dm.vblank_params[int_params.irq_source - DC_IRQ_SOURCE_VBLANK1];
 
 		c_irq_params->adev = adev;
 		c_irq_params->irq_src = int_params.irq_source;
 
-		amdgpu_dm_irq_register_interrupt(adev, &int_params,
-				dm_crtc_high_irq, c_irq_params);
+		if (!amdgpu_dm_irq_register_interrupt(adev, &int_params,
+			dm_crtc_high_irq, c_irq_params))
+			return -ENOMEM;
 	}
 
 	/* Use VUPDATE interrupt */
 	for (i = VISLANDS30_IV_SRCID_D1_V_UPDATE_INT; i <= VISLANDS30_IV_SRCID_D6_V_UPDATE_INT; i += 2) {
 		r = amdgpu_irq_add_id(adev, client_id, i, &adev->vupdate_irq);
 		if (r) {
-			DRM_ERROR("Failed to add vupdate irq id!\n");
+			drm_err(adev_to_drm(adev), "Failed to add vupdate irq id!\n");
 			return r;
 		}
 
@@ -3463,13 +4526,21 @@ static int dce110_register_irq_handlers(struct amdgpu_device *adev)
 		int_params.irq_source =
 			dc_interrupt_to_irq_source(dc, i, 0);
 
+		if (int_params.irq_source == DC_IRQ_SOURCE_INVALID ||
+			int_params.irq_source  < DC_IRQ_SOURCE_VUPDATE1 ||
+			int_params.irq_source  > DC_IRQ_SOURCE_VUPDATE6) {
+			drm_err(adev_to_drm(adev), "Failed to register vupdate irq!\n");
+			return -EINVAL;
+		}
+
 		c_irq_params = &adev->dm.vupdate_params[int_params.irq_source - DC_IRQ_SOURCE_VUPDATE1];
 
 		c_irq_params->adev = adev;
 		c_irq_params->irq_src = int_params.irq_source;
 
-		amdgpu_dm_irq_register_interrupt(adev, &int_params,
-				dm_vupdate_high_irq, c_irq_params);
+		if (!amdgpu_dm_irq_register_interrupt(adev, &int_params,
+			dm_vupdate_high_irq, c_irq_params))
+			return -ENOMEM;
 	}
 
 	/* Use GRPH_PFLIP interrupt */
@@ -3477,7 +4548,7 @@ static int dce110_register_irq_handlers(struct amdgpu_device *adev)
 			i <= VISLANDS30_IV_SRCID_D6_GRPH_PFLIP; i += 2) {
 		r = amdgpu_irq_add_id(adev, client_id, i, &adev->pageflip_irq);
 		if (r) {
-			DRM_ERROR("Failed to add page flip irq id!\n");
+			drm_err(adev_to_drm(adev), "Failed to add page flip irq id!\n");
 			return r;
 		}
 
@@ -3485,30 +4556,36 @@ static int dce110_register_irq_handlers(struct amdgpu_device *adev)
 		int_params.irq_source =
 			dc_interrupt_to_irq_source(dc, i, 0);
 
+		if (int_params.irq_source == DC_IRQ_SOURCE_INVALID ||
+			int_params.irq_source  < DC_IRQ_SOURCE_PFLIP_FIRST ||
+			int_params.irq_source  > DC_IRQ_SOURCE_PFLIP_LAST) {
+			drm_err(adev_to_drm(adev), "Failed to register pflip irq!\n");
+			return -EINVAL;
+		}
+
 		c_irq_params = &adev->dm.pflip_params[int_params.irq_source - DC_IRQ_SOURCE_PFLIP_FIRST];
 
 		c_irq_params->adev = adev;
 		c_irq_params->irq_src = int_params.irq_source;
 
-		amdgpu_dm_irq_register_interrupt(adev, &int_params,
-				dm_pflip_high_irq, c_irq_params);
-
+		if (!amdgpu_dm_irq_register_interrupt(adev, &int_params,
+			dm_pflip_high_irq, c_irq_params))
+			return -ENOMEM;
 	}
 
 	/* HPD */
 	r = amdgpu_irq_add_id(adev, client_id,
 			VISLANDS30_IV_SRCID_HOTPLUG_DETECT_A, &adev->hpd_irq);
 	if (r) {
-		DRM_ERROR("Failed to add hpd irq id!\n");
+		drm_err(adev_to_drm(adev), "Failed to add hpd irq id!\n");
 		return r;
 	}
 
-	register_hpd_handlers(adev);
+	r = register_hpd_handlers(adev);
 
-	return 0;
+	return r;
 }
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 /* Register IRQ sources and initialize IRQ callbacks */
 static int dcn10_register_irq_handlers(struct amdgpu_device *adev)
 {
@@ -3550,7 +4627,7 @@ static int dcn10_register_irq_handlers(struct amdgpu_device *adev)
 		r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_DCE, i, &adev->crtc_irq);
 
 		if (r) {
-			DRM_ERROR("Failed to add crtc irq id!\n");
+			drm_err(adev_to_drm(adev), "Failed to add crtc irq id!\n");
 			return r;
 		}
 
@@ -3558,13 +4635,21 @@ static int dcn10_register_irq_handlers(struct amdgpu_device *adev)
 		int_params.irq_source =
 			dc_interrupt_to_irq_source(dc, i, 0);
 
+		if (int_params.irq_source == DC_IRQ_SOURCE_INVALID ||
+			int_params.irq_source  < DC_IRQ_SOURCE_VBLANK1 ||
+			int_params.irq_source  > DC_IRQ_SOURCE_VBLANK6) {
+			drm_err(adev_to_drm(adev), "Failed to register vblank irq!\n");
+			return -EINVAL;
+		}
+
 		c_irq_params = &adev->dm.vblank_params[int_params.irq_source - DC_IRQ_SOURCE_VBLANK1];
 
 		c_irq_params->adev = adev;
 		c_irq_params->irq_src = int_params.irq_source;
 
-		amdgpu_dm_irq_register_interrupt(
-			adev, &int_params, dm_crtc_high_irq, c_irq_params);
+		if (!amdgpu_dm_irq_register_interrupt(adev, &int_params,
+			dm_crtc_high_irq, c_irq_params))
+			return -ENOMEM;
 	}
 
 	/* Use otg vertical line interrupt */
@@ -3574,7 +4659,7 @@ static int dcn10_register_irq_handlers(struct amdgpu_device *adev)
 				vrtl_int_srcid[i], &adev->vline0_irq);
 
 		if (r) {
-			DRM_ERROR("Failed to add vline0 irq id!\n");
+			drm_err(adev_to_drm(adev), "Failed to add vline0 irq id!\n");
 			return r;
 		}
 
@@ -3582,9 +4667,11 @@ static int dcn10_register_irq_handlers(struct amdgpu_device *adev)
 		int_params.irq_source =
 			dc_interrupt_to_irq_source(dc, vrtl_int_srcid[i], 0);
 
-		if (int_params.irq_source == DC_IRQ_SOURCE_INVALID) {
-			DRM_ERROR("Failed to register vline0 irq %d!\n", vrtl_int_srcid[i]);
-			break;
+		if (int_params.irq_source == DC_IRQ_SOURCE_INVALID ||
+			int_params.irq_source < DC_IRQ_SOURCE_DC1_VLINE0 ||
+			int_params.irq_source > DC_IRQ_SOURCE_DC6_VLINE0) {
+			drm_err(adev_to_drm(adev), "Failed to register vline0 irq!\n");
+			return -EINVAL;
 		}
 
 		c_irq_params = &adev->dm.vline0_params[int_params.irq_source
@@ -3593,8 +4680,10 @@ static int dcn10_register_irq_handlers(struct amdgpu_device *adev)
 		c_irq_params->adev = adev;
 		c_irq_params->irq_src = int_params.irq_source;
 
-		amdgpu_dm_irq_register_interrupt(adev, &int_params,
-				dm_dcn_vertical_interrupt0_high_irq, c_irq_params);
+		if (!amdgpu_dm_irq_register_interrupt(adev, &int_params,
+			dm_dcn_vertical_interrupt0_high_irq,
+			c_irq_params))
+			return -ENOMEM;
 	}
 #endif
 
@@ -3609,7 +4698,7 @@ static int dcn10_register_irq_handlers(struct amdgpu_device *adev)
 		r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_DCE, i, &adev->vupdate_irq);
 
 		if (r) {
-			DRM_ERROR("Failed to add vupdate irq id!\n");
+			drm_err(adev_to_drm(adev), "Failed to add vupdate irq id!\n");
 			return r;
 		}
 
@@ -3617,22 +4706,30 @@ static int dcn10_register_irq_handlers(struct amdgpu_device *adev)
 		int_params.irq_source =
 			dc_interrupt_to_irq_source(dc, i, 0);
 
+		if (int_params.irq_source == DC_IRQ_SOURCE_INVALID ||
+			int_params.irq_source  < DC_IRQ_SOURCE_VUPDATE1 ||
+			int_params.irq_source  > DC_IRQ_SOURCE_VUPDATE6) {
+			drm_err(adev_to_drm(adev), "Failed to register vupdate irq!\n");
+			return -EINVAL;
+		}
+
 		c_irq_params = &adev->dm.vupdate_params[int_params.irq_source - DC_IRQ_SOURCE_VUPDATE1];
 
 		c_irq_params->adev = adev;
 		c_irq_params->irq_src = int_params.irq_source;
 
-		amdgpu_dm_irq_register_interrupt(adev, &int_params,
-				dm_vupdate_high_irq, c_irq_params);
+		if (!amdgpu_dm_irq_register_interrupt(adev, &int_params,
+			dm_vupdate_high_irq, c_irq_params))
+			return -ENOMEM;
 	}
 
 	/* Use GRPH_PFLIP interrupt */
 	for (i = DCN_1_0__SRCID__HUBP0_FLIP_INTERRUPT;
-			i <= DCN_1_0__SRCID__HUBP0_FLIP_INTERRUPT + adev->mode_info.num_crtc - 1;
+			i <= DCN_1_0__SRCID__HUBP0_FLIP_INTERRUPT + dc->caps.max_otg_num - 1;
 			i++) {
 		r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_DCE, i, &adev->pageflip_irq);
 		if (r) {
-			DRM_ERROR("Failed to add page flip irq id!\n");
+			drm_err(adev_to_drm(adev), "Failed to add page flip irq id!\n");
 			return r;
 		}
 
@@ -3640,27 +4737,34 @@ static int dcn10_register_irq_handlers(struct amdgpu_device *adev)
 		int_params.irq_source =
 			dc_interrupt_to_irq_source(dc, i, 0);
 
+		if (int_params.irq_source == DC_IRQ_SOURCE_INVALID ||
+			int_params.irq_source  < DC_IRQ_SOURCE_PFLIP_FIRST ||
+			int_params.irq_source  > DC_IRQ_SOURCE_PFLIP_LAST) {
+			drm_err(adev_to_drm(adev), "Failed to register pflip irq!\n");
+			return -EINVAL;
+		}
+
 		c_irq_params = &adev->dm.pflip_params[int_params.irq_source - DC_IRQ_SOURCE_PFLIP_FIRST];
 
 		c_irq_params->adev = adev;
 		c_irq_params->irq_src = int_params.irq_source;
 
-		amdgpu_dm_irq_register_interrupt(adev, &int_params,
-				dm_pflip_high_irq, c_irq_params);
-
+		if (!amdgpu_dm_irq_register_interrupt(adev, &int_params,
+			dm_pflip_high_irq, c_irq_params))
+			return -ENOMEM;
 	}
 
 	/* HPD */
 	r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_DCE, DCN_1_0__SRCID__DC_HPD1_INT,
 			&adev->hpd_irq);
 	if (r) {
-		DRM_ERROR("Failed to add hpd irq id!\n");
+		drm_err(adev_to_drm(adev), "Failed to add hpd irq id!\n");
 		return r;
 	}
 
-	register_hpd_handlers(adev);
+	r = register_hpd_handlers(adev);
 
-	return 0;
+	return r;
 }
 /* Register Outbox IRQ sources and initialize IRQ callbacks */
 static int register_outbox_irq_handlers(struct amdgpu_device *adev)
@@ -3676,7 +4780,7 @@ static int register_outbox_irq_handlers(struct amdgpu_device *adev)
 	r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_DCE, DCN_1_0__SRCID__DMCUB_OUTBOX_LOW_PRIORITY_READY_INT,
 			&adev->dmub_outbox_irq);
 	if (r) {
-		DRM_ERROR("Failed to add outbox irq id!\n");
+		drm_err(adev_to_drm(adev), "Failed to add outbox irq id!\n");
 		return r;
 	}
 
@@ -3691,13 +4795,13 @@ static int register_outbox_irq_handlers(struct amdgpu_device *adev)
 		c_irq_params->adev = adev;
 		c_irq_params->irq_src = int_params.irq_source;
 
-		amdgpu_dm_irq_register_interrupt(adev, &int_params,
-				dm_dmub_outbox1_low_irq, c_irq_params);
+		if (!amdgpu_dm_irq_register_interrupt(adev, &int_params,
+			dm_dmub_outbox1_low_irq, c_irq_params))
+			return -ENOMEM;
 	}
 
 	return 0;
 }
-#endif
 
 /*
  * Acquires the lock for the atomic state object and returns
@@ -3705,8 +4809,8 @@ static int register_outbox_irq_handlers(struct amdgpu_device *adev)
  *
  * This should only be called during atomic check.
  */
-static int dm_atomic_get_state(struct drm_atomic_state *state,
-			       struct dm_atomic_state **dm_state)
+int dm_atomic_get_state(struct drm_atomic_state *state,
+			struct dm_atomic_state **dm_state)
 {
 	struct drm_device *dev = state->dev;
 	struct amdgpu_device *adev = drm_to_adev(dev);
@@ -3757,7 +4861,7 @@ dm_atomic_duplicate_state(struct drm_private_obj *obj)
 	old_state = to_dm_atomic_state(obj->state);
 
 	if (old_state && old_state->context)
-		new_state->context = dc_copy_state(old_state->context);
+		new_state->context = dc_state_create_copy(old_state->context);
 
 	if (!new_state->context) {
 		kfree(new_state);
@@ -3773,7 +4877,7 @@ static void dm_atomic_destroy_state(struct drm_private_obj *obj,
 	struct dm_atomic_state *dm_state = to_dm_atomic_state(state);
 
 	if (dm_state && dm_state->context)
-		dc_release_state(dm_state->context);
+		dc_state_release(dm_state->context);
 
 	kfree(dm_state);
 }
@@ -3797,24 +4901,24 @@ static int amdgpu_dm_mode_config_init(struct amdgpu_device *adev)
 	adev_to_drm(adev)->mode_config.max_height = 16384;
 
 	adev_to_drm(adev)->mode_config.preferred_depth = 24;
-	adev_to_drm(adev)->mode_config.prefer_shadow = 1;
+	if (adev->asic_type == CHIP_HAWAII)
+		/* disable prefer shadow for now due to hibernation issues */
+		adev_to_drm(adev)->mode_config.prefer_shadow = 0;
+	else
+		adev_to_drm(adev)->mode_config.prefer_shadow = 1;
 	/* indicates support for immediate flip */
 	adev_to_drm(adev)->mode_config.async_page_flip = true;
 
-	adev_to_drm(adev)->mode_config.fb_base = adev->gmc.aper_base;
-
 	state = kzalloc(sizeof(*state), GFP_KERNEL);
 	if (!state)
 		return -ENOMEM;
 
-	state->context = dc_create_state(adev->dm.dc);
+	state->context = dc_state_create_current_copy(adev->dm.dc);
 	if (!state->context) {
 		kfree(state);
 		return -ENOMEM;
 	}
 
-	dc_resource_state_copy_construct_current(adev->dm.dc, state->context);
-
 	drm_atomic_private_obj_init(adev_to_drm(adev),
 				    &adev->dm.atomic_obj,
 				    &state->base,
@@ -3822,14 +4926,22 @@ static int amdgpu_dm_mode_config_init(struct amdgpu_device *adev)
 
 	r = amdgpu_display_modeset_create_props(adev);
 	if (r) {
-		dc_release_state(state->context);
+		dc_state_release(state->context);
 		kfree(state);
 		return r;
 	}
 
+#ifdef AMD_PRIVATE_COLOR
+	if (amdgpu_dm_create_color_properties(adev)) {
+		dc_state_release(state->context);
+		kfree(state);
+		return -ENOMEM;
+	}
+#endif
+
 	r = amdgpu_dm_audio_init(adev);
 	if (r) {
-		dc_release_state(state->context);
+		dc_state_release(state->context);
 		kfree(state);
 		return r;
 	}
@@ -3839,46 +4951,51 @@ static int amdgpu_dm_mode_config_init(struct amdgpu_device *adev)
 
 #define AMDGPU_DM_DEFAULT_MIN_BACKLIGHT 12
 #define AMDGPU_DM_DEFAULT_MAX_BACKLIGHT 255
+#define AMDGPU_DM_MIN_SPREAD ((AMDGPU_DM_DEFAULT_MAX_BACKLIGHT - AMDGPU_DM_DEFAULT_MIN_BACKLIGHT) / 2)
 #define AUX_BL_DEFAULT_TRANSITION_TIME_MS 50
 
-#if defined(CONFIG_BACKLIGHT_CLASS_DEVICE) ||\
-	defined(CONFIG_BACKLIGHT_CLASS_DEVICE_MODULE)
-
 static void amdgpu_dm_update_backlight_caps(struct amdgpu_display_manager *dm,
 					    int bl_idx)
 {
+	struct amdgpu_dm_backlight_caps *caps = &dm->backlight_caps[bl_idx];
+
+	if (caps->caps_valid)
+		return;
+
 #if defined(CONFIG_ACPI)
-	struct amdgpu_dm_backlight_caps caps;
+	amdgpu_acpi_get_backlight_caps(caps);
 
-	memset(&caps, 0, sizeof(caps));
+	/* validate the firmware value is sane */
+	if (caps->caps_valid) {
+		int spread = caps->max_input_signal - caps->min_input_signal;
 
-	if (dm->backlight_caps[bl_idx].caps_valid)
-		return;
+		if (caps->max_input_signal > AMDGPU_DM_DEFAULT_MAX_BACKLIGHT ||
+		    caps->min_input_signal < 0 ||
+		    spread > AMDGPU_DM_DEFAULT_MAX_BACKLIGHT ||
+		    spread < AMDGPU_DM_MIN_SPREAD) {
+			DRM_DEBUG_KMS("DM: Invalid backlight caps: min=%d, max=%d\n",
+				      caps->min_input_signal, caps->max_input_signal);
+			caps->caps_valid = false;
+		}
+	}
 
-	amdgpu_acpi_get_backlight_caps(&caps);
-	if (caps.caps_valid) {
-		dm->backlight_caps[bl_idx].caps_valid = true;
-		if (caps.aux_support)
-			return;
-		dm->backlight_caps[bl_idx].min_input_signal = caps.min_input_signal;
-		dm->backlight_caps[bl_idx].max_input_signal = caps.max_input_signal;
-	} else {
-		dm->backlight_caps[bl_idx].min_input_signal =
-				AMDGPU_DM_DEFAULT_MIN_BACKLIGHT;
-		dm->backlight_caps[bl_idx].max_input_signal =
-				AMDGPU_DM_DEFAULT_MAX_BACKLIGHT;
+	if (!caps->caps_valid) {
+		caps->min_input_signal = AMDGPU_DM_DEFAULT_MIN_BACKLIGHT;
+		caps->max_input_signal = AMDGPU_DM_DEFAULT_MAX_BACKLIGHT;
+		caps->caps_valid = true;
 	}
 #else
-	if (dm->backlight_caps[bl_idx].aux_support)
+	if (caps->aux_support)
 		return;
 
-	dm->backlight_caps[bl_idx].min_input_signal = AMDGPU_DM_DEFAULT_MIN_BACKLIGHT;
-	dm->backlight_caps[bl_idx].max_input_signal = AMDGPU_DM_DEFAULT_MAX_BACKLIGHT;
+	caps->min_input_signal = AMDGPU_DM_DEFAULT_MIN_BACKLIGHT;
+	caps->max_input_signal = AMDGPU_DM_DEFAULT_MAX_BACKLIGHT;
+	caps->caps_valid = true;
 #endif
 }
 
 static int get_brightness_range(const struct amdgpu_dm_backlight_caps *caps,
-				unsigned *min, unsigned *max)
+				unsigned int *min, unsigned int *max)
 {
 	if (!caps)
 		return 0;
@@ -3895,66 +5012,187 @@ static int get_brightness_range(const struct amdgpu_dm_backlight_caps *caps,
 	return 1;
 }
 
+/* Rescale from [min..max] to [0..AMDGPU_MAX_BL_LEVEL] */
+static inline u32 scale_input_to_fw(int min, int max, u64 input)
+{
+	return DIV_ROUND_CLOSEST_ULL(input * AMDGPU_MAX_BL_LEVEL, max - min);
+}
+
+/* Rescale from [0..AMDGPU_MAX_BL_LEVEL] to [min..max] */
+static inline u32 scale_fw_to_input(int min, int max, u64 input)
+{
+	return min + DIV_ROUND_CLOSEST_ULL(input * (max - min), AMDGPU_MAX_BL_LEVEL);
+}
+
+static void convert_custom_brightness(const struct amdgpu_dm_backlight_caps *caps,
+				      unsigned int min, unsigned int max,
+				      uint32_t *user_brightness)
+{
+	u32 brightness = scale_input_to_fw(min, max, *user_brightness);
+	u8 lower_signal, upper_signal, upper_lum, lower_lum, lum;
+	int left, right;
+
+	if (amdgpu_dc_debug_mask & DC_DISABLE_CUSTOM_BRIGHTNESS_CURVE)
+		return;
+
+	if (!caps->data_points)
+		return;
+
+	/*
+	 * Handle the case where brightness is below the first data point
+	 * Interpolate between (0,0) and (first_signal, first_lum)
+	 */
+	if (brightness < caps->luminance_data[0].input_signal) {
+		lum = DIV_ROUND_CLOSEST(caps->luminance_data[0].luminance * brightness,
+					caps->luminance_data[0].input_signal);
+		goto scale;
+	}
+
+	left = 0;
+	right = caps->data_points - 1;
+	while (left <= right) {
+		int mid = left + (right - left) / 2;
+		u8 signal = caps->luminance_data[mid].input_signal;
+
+		/* Exact match found */
+		if (signal == brightness) {
+			lum = caps->luminance_data[mid].luminance;
+			goto scale;
+		}
+
+		if (signal < brightness)
+			left = mid + 1;
+		else
+			right = mid - 1;
+	}
+
+	/* verify bound */
+	if (left >= caps->data_points)
+		left = caps->data_points - 1;
+
+	/* At this point, left > right */
+	lower_signal = caps->luminance_data[right].input_signal;
+	upper_signal = caps->luminance_data[left].input_signal;
+	lower_lum = caps->luminance_data[right].luminance;
+	upper_lum = caps->luminance_data[left].luminance;
+
+	/* interpolate */
+	if (right == left || !lower_lum)
+		lum = upper_lum;
+	else
+		lum = lower_lum + DIV_ROUND_CLOSEST((upper_lum - lower_lum) *
+						    (brightness - lower_signal),
+						    upper_signal - lower_signal);
+scale:
+	*user_brightness = scale_fw_to_input(min, max,
+					     DIV_ROUND_CLOSEST(lum * brightness, 101));
+}
+
 static u32 convert_brightness_from_user(const struct amdgpu_dm_backlight_caps *caps,
 					uint32_t brightness)
 {
-	unsigned min, max;
+	unsigned int min, max;
 
 	if (!get_brightness_range(caps, &min, &max))
 		return brightness;
 
-	// Rescale 0..255 to min..max
-	return min + DIV_ROUND_CLOSEST((max - min) * brightness,
-				       AMDGPU_MAX_BL_LEVEL);
+	convert_custom_brightness(caps, min, max, &brightness);
+
+	// Rescale 0..max to min..max
+	return min + DIV_ROUND_CLOSEST_ULL((u64)(max - min) * brightness, max);
 }
 
 static u32 convert_brightness_to_user(const struct amdgpu_dm_backlight_caps *caps,
 				      uint32_t brightness)
 {
-	unsigned min, max;
+	unsigned int min, max;
 
 	if (!get_brightness_range(caps, &min, &max))
 		return brightness;
 
 	if (brightness < min)
 		return 0;
-	// Rescale min..max to 0..255
-	return DIV_ROUND_CLOSEST(AMDGPU_MAX_BL_LEVEL * (brightness - min),
+	// Rescale min..max to 0..max
+	return DIV_ROUND_CLOSEST_ULL((u64)max * (brightness - min),
 				 max - min);
 }
 
-static int amdgpu_dm_backlight_set_level(struct amdgpu_display_manager *dm,
+static void amdgpu_dm_backlight_set_level(struct amdgpu_display_manager *dm,
 					 int bl_idx,
 					 u32 user_brightness)
 {
-	struct amdgpu_dm_backlight_caps caps;
+	struct amdgpu_dm_backlight_caps *caps;
 	struct dc_link *link;
 	u32 brightness;
-	bool rc;
+	bool rc, reallow_idle = false;
+	struct drm_connector *connector;
+
+	list_for_each_entry(connector, &dm->ddev->mode_config.connector_list, head) {
+		struct amdgpu_dm_connector *aconnector = to_amdgpu_dm_connector(connector);
+
+		if (aconnector->bl_idx != bl_idx)
+			continue;
+
+		/* if connector is off, save the brightness for next time it's on */
+		if (!aconnector->base.encoder) {
+			dm->brightness[bl_idx] = user_brightness;
+			dm->actual_brightness[bl_idx] = 0;
+			return;
+		}
+	}
 
 	amdgpu_dm_update_backlight_caps(dm, bl_idx);
-	caps = dm->backlight_caps[bl_idx];
+	caps = &dm->backlight_caps[bl_idx];
 
 	dm->brightness[bl_idx] = user_brightness;
 	/* update scratch register */
 	if (bl_idx == 0)
 		amdgpu_atombios_scratch_regs_set_backlight_level(dm->adev, dm->brightness[bl_idx]);
-	brightness = convert_brightness_from_user(&caps, dm->brightness[bl_idx]);
+	brightness = convert_brightness_from_user(caps, dm->brightness[bl_idx]);
 	link = (struct dc_link *)dm->backlight_link[bl_idx];
 
+	/* Apply brightness quirk */
+	if (caps->brightness_mask)
+		brightness |= caps->brightness_mask;
+
 	/* Change brightness based on AUX property */
-	if (caps.aux_support) {
+	mutex_lock(&dm->dc_lock);
+	if (dm->dc->caps.ips_support && dm->dc->ctx->dmub_srv->idle_allowed) {
+		dc_allow_idle_optimizations(dm->dc, false);
+		reallow_idle = true;
+	}
+
+	if (trace_amdgpu_dm_brightness_enabled()) {
+		trace_amdgpu_dm_brightness(__builtin_return_address(0),
+					   user_brightness,
+					   brightness,
+					   caps->aux_support,
+					   power_supply_is_system_supplied() > 0);
+	}
+
+	if (caps->aux_support) {
 		rc = dc_link_set_backlight_level_nits(link, true, brightness,
 						      AUX_BL_DEFAULT_TRANSITION_TIME_MS);
 		if (!rc)
 			DRM_DEBUG("DM: Failed to update backlight via AUX on eDP[%d]\n", bl_idx);
 	} else {
-		rc = dc_link_set_backlight_level(link, brightness, 0);
+		struct set_backlight_level_params backlight_level_params = { 0 };
+
+		backlight_level_params.backlight_pwm_u16_16 = brightness;
+		backlight_level_params.transition_time_in_ms = 0;
+
+		rc = dc_link_set_backlight_level(link, &backlight_level_params);
 		if (!rc)
 			DRM_DEBUG("DM: Failed to update backlight on eDP[%d]\n", bl_idx);
 	}
 
-	return rc ? 0 : 1;
+	if (dm->dc->caps.ips_support && reallow_idle)
+		dc_allow_idle_optimizations(dm->dc, true);
+
+	mutex_unlock(&dm->dc_lock);
+
+	if (rc)
+		dm->actual_brightness[bl_idx] = user_brightness;
 }
 
 static int amdgpu_dm_backlight_update_status(struct backlight_device *bd)
@@ -3976,6 +5214,7 @@ static int amdgpu_dm_backlight_update_status(struct backlight_device *bd)
 static u32 amdgpu_dm_backlight_get_level(struct amdgpu_display_manager *dm,
 					 int bl_idx)
 {
+	int ret;
 	struct amdgpu_dm_backlight_caps caps;
 	struct dc_link *link = (struct dc_link *)dm->backlight_link[bl_idx];
 
@@ -3984,19 +5223,18 @@ static u32 amdgpu_dm_backlight_get_level(struct amdgpu_display_manager *dm,
 
 	if (caps.aux_support) {
 		u32 avg, peak;
-		bool rc;
 
-		rc = dc_link_get_backlight_level_nits(link, &avg, &peak);
-		if (!rc)
+		if (!dc_link_get_backlight_level_nits(link, &avg, &peak))
 			return dm->brightness[bl_idx];
 		return convert_brightness_to_user(&caps, avg);
-	} else {
-		int ret = dc_link_get_backlight_level(link);
-
-		if (ret == DC_ERROR_UNEXPECTED)
-			return dm->brightness[bl_idx];
-		return convert_brightness_to_user(&caps, ret);
 	}
+
+	ret = dc_link_get_backlight_level(link);
+
+	if (ret == DC_ERROR_UNEXPECTED)
+		return dm->brightness[bl_idx];
+
+	return convert_brightness_to_user(&caps, ret);
 }
 
 static int amdgpu_dm_backlight_get_brightness(struct backlight_device *bd)
@@ -4020,33 +5258,59 @@ static const struct backlight_ops amdgpu_dm_backlight_ops = {
 };
 
 static void
-amdgpu_dm_register_backlight_device(struct amdgpu_display_manager *dm)
+amdgpu_dm_register_backlight_device(struct amdgpu_dm_connector *aconnector)
 {
-	char bl_name[16];
+	struct drm_device *drm = aconnector->base.dev;
+	struct amdgpu_display_manager *dm = &drm_to_adev(drm)->dm;
 	struct backlight_properties props = { 0 };
+	struct amdgpu_dm_backlight_caps *caps;
+	char bl_name[16];
+	int min, max;
 
-	amdgpu_dm_update_backlight_caps(dm, dm->num_of_edps);
-	dm->brightness[dm->num_of_edps] = AMDGPU_MAX_BL_LEVEL;
+	if (aconnector->bl_idx == -1)
+		return;
+
+	if (!acpi_video_backlight_use_native()) {
+		drm_info(drm, "Skipping amdgpu DM backlight registration\n");
+		/* Try registering an ACPI video backlight device instead. */
+		acpi_video_register_backlight();
+		return;
+	}
 
-	props.max_brightness = AMDGPU_MAX_BL_LEVEL;
-	props.brightness = AMDGPU_MAX_BL_LEVEL;
+	caps = &dm->backlight_caps[aconnector->bl_idx];
+	if (get_brightness_range(caps, &min, &max)) {
+		if (power_supply_is_system_supplied() > 0)
+			props.brightness = DIV_ROUND_CLOSEST((max - min) * caps->ac_level, 100);
+		else
+			props.brightness = DIV_ROUND_CLOSEST((max - min) * caps->dc_level, 100);
+		/* min is zero, so max needs to be adjusted */
+		props.max_brightness = max - min;
+		drm_dbg(drm, "Backlight caps: min: %d, max: %d, ac %d, dc %d\n", min, max,
+			caps->ac_level, caps->dc_level);
+	} else
+		props.brightness = props.max_brightness = MAX_BACKLIGHT_LEVEL;
+
+	if (caps->data_points && !(amdgpu_dc_debug_mask & DC_DISABLE_CUSTOM_BRIGHTNESS_CURVE)) {
+		drm_info(drm, "Using custom brightness curve\n");
+		props.scale = BACKLIGHT_SCALE_NON_LINEAR;
+	} else
+		props.scale = BACKLIGHT_SCALE_LINEAR;
 	props.type = BACKLIGHT_RAW;
 
 	snprintf(bl_name, sizeof(bl_name), "amdgpu_bl%d",
-		 adev_to_drm(dm->adev)->primary->index + dm->num_of_edps);
+		 drm->primary->index + aconnector->bl_idx);
 
-	dm->backlight_dev[dm->num_of_edps] = backlight_device_register(bl_name,
-								       adev_to_drm(dm->adev)->dev,
-								       dm,
-								       &amdgpu_dm_backlight_ops,
-								       &props);
+	dm->backlight_dev[aconnector->bl_idx] =
+		backlight_device_register(bl_name, aconnector->base.kdev, dm,
+					  &amdgpu_dm_backlight_ops, &props);
+	dm->brightness[aconnector->bl_idx] = props.brightness;
 
-	if (IS_ERR(dm->backlight_dev[dm->num_of_edps]))
-		DRM_ERROR("DM: Backlight registration failed!\n");
-	else
-		DRM_DEBUG_DRIVER("DM: Registered Backlight device: %s\n", bl_name);
+	if (IS_ERR(dm->backlight_dev[aconnector->bl_idx])) {
+		drm_err(drm, "DM: Backlight registration failed!\n");
+		dm->backlight_dev[aconnector->bl_idx] = NULL;
+	} else
+		drm_dbg_driver(drm, "DM: Registered Backlight device: %s\n", bl_name);
 }
-#endif
 
 static int initialize_plane(struct amdgpu_display_manager *dm,
 			    struct amdgpu_mode_info *mode_info, int plane_id,
@@ -4059,7 +5323,7 @@ static int initialize_plane(struct amdgpu_display_manager *dm,
 
 	plane = kzalloc(sizeof(struct drm_plane), GFP_KERNEL);
 	if (!plane) {
-		DRM_ERROR("KMS: Failed to allocate plane\n");
+		drm_err(adev_to_drm(dm->adev), "KMS: Failed to allocate plane\n");
 		return -ENOMEM;
 	}
 	plane->type = plane_type;
@@ -4077,7 +5341,7 @@ static int initialize_plane(struct amdgpu_display_manager *dm,
 	ret = amdgpu_dm_plane_init(dm, plane, possible_crtcs, plane_cap);
 
 	if (ret) {
-		DRM_ERROR("KMS: Failed to initialize plane\n");
+		drm_err(adev_to_drm(dm->adev), "KMS: Failed to initialize plane\n");
 		kfree(plane);
 		return ret;
 	}
@@ -4089,30 +5353,39 @@ static int initialize_plane(struct amdgpu_display_manager *dm,
 }
 
 
-static void register_backlight_device(struct amdgpu_display_manager *dm,
-				      struct dc_link *link)
+static void setup_backlight_device(struct amdgpu_display_manager *dm,
+				   struct amdgpu_dm_connector *aconnector)
 {
-#if defined(CONFIG_BACKLIGHT_CLASS_DEVICE) ||\
-	defined(CONFIG_BACKLIGHT_CLASS_DEVICE_MODULE)
+	struct amdgpu_dm_backlight_caps *caps;
+	struct dc_link *link = aconnector->dc_link;
+	int bl_idx = dm->num_of_edps;
 
-	if ((link->connector_signal & (SIGNAL_TYPE_EDP | SIGNAL_TYPE_LVDS)) &&
-	    link->type != dc_connection_none) {
-		/*
-		 * Event if registration failed, we should continue with
-		 * DM initialization because not having a backlight control
-		 * is better then a black screen.
-		 */
-		if (!dm->backlight_dev[dm->num_of_edps])
-			amdgpu_dm_register_backlight_device(dm);
+	if (!(link->connector_signal & (SIGNAL_TYPE_EDP | SIGNAL_TYPE_LVDS)) ||
+	    link->type == dc_connection_none)
+		return;
 
-		if (dm->backlight_dev[dm->num_of_edps]) {
-			dm->backlight_link[dm->num_of_edps] = link;
-			dm->num_of_edps++;
-		}
+	if (dm->num_of_edps >= AMDGPU_DM_MAX_NUM_EDP) {
+		drm_warn(adev_to_drm(dm->adev), "Too much eDP connections, skipping backlight setup for additional eDPs\n");
+		return;
 	}
-#endif
+
+	aconnector->bl_idx = bl_idx;
+
+	amdgpu_dm_update_backlight_caps(dm, bl_idx);
+	dm->backlight_link[bl_idx] = link;
+	dm->num_of_edps++;
+
+	update_connector_ext_caps(aconnector);
+	caps = &dm->backlight_caps[aconnector->bl_idx];
+
+	/* Only offer ABM property when non-OLED and user didn't turn off by module parameter */
+	if (!caps->ext_caps->bits.oled && amdgpu_dm_abm_level < 0)
+		drm_object_attach_property(&aconnector->base.base,
+					   dm->adev->mode_info.abm_level_property,
+					   ABM_SYSFS_CONTROL);
 }
 
+static void amdgpu_set_panel_orientation(struct drm_connector *connector);
 
 /*
  * In this architecture, the association
@@ -4125,29 +5398,36 @@ static void register_backlight_device(struct amdgpu_display_manager *dm,
 static int amdgpu_dm_initialize_drm_device(struct amdgpu_device *adev)
 {
 	struct amdgpu_display_manager *dm = &adev->dm;
-	int32_t i;
+	s32 i;
 	struct amdgpu_dm_connector *aconnector = NULL;
 	struct amdgpu_encoder *aencoder = NULL;
 	struct amdgpu_mode_info *mode_info = &adev->mode_info;
-	uint32_t link_cnt;
-	int32_t primary_planes;
+	u32 link_cnt;
+	s32 primary_planes;
 	enum dc_connection_type new_connection_type = dc_connection_none;
 	const struct dc_plane_cap *plane;
 	bool psr_feature_enabled = false;
+	bool replay_feature_enabled = false;
+	int max_overlay = dm->dc->caps.max_slave_planes;
 
 	dm->display_indexes_num = dm->dc->caps.max_streams;
 	/* Update the actual used number of crtc */
 	adev->mode_info.num_crtc = adev->dm.display_indexes_num;
 
+	amdgpu_dm_set_irq_funcs(adev);
+
 	link_cnt = dm->dc->caps.max_links;
 	if (amdgpu_dm_mode_config_init(dm->adev)) {
-		DRM_ERROR("DM: Failed to initialize mode config\n");
+		drm_err(adev_to_drm(adev), "DM: Failed to initialize mode config\n");
 		return -EINVAL;
 	}
 
 	/* There is one primary plane per CRTC */
 	primary_planes = dm->dc->caps.max_streams;
-	ASSERT(primary_planes <= AMDGPU_MAX_PLANES);
+	if (primary_planes > AMDGPU_MAX_PLANES) {
+		drm_err(adev_to_drm(adev), "DM: Plane nums out of 6 planes\n");
+		return -EINVAL;
+	}
 
 	/*
 	 * Initialize primary planes, implicit planes for legacy IOCTLS.
@@ -4158,7 +5438,7 @@ static int amdgpu_dm_initialize_drm_device(struct amdgpu_device *adev)
 
 		if (initialize_plane(dm, mode_info, i,
 				     DRM_PLANE_TYPE_PRIMARY, plane)) {
-			DRM_ERROR("KMS: Failed to initialize primary plane\n");
+			drm_err(adev_to_drm(adev), "KMS: Failed to initialize primary plane\n");
 			goto fail;
 		}
 	}
@@ -4175,53 +5455,71 @@ static int amdgpu_dm_initialize_drm_device(struct amdgpu_device *adev)
 	for (i = 0; i < dm->dc->caps.max_planes; ++i) {
 		struct dc_plane_cap *plane = &dm->dc->caps.planes[i];
 
-		if (plane->type != DC_PLANE_TYPE_DCN_UNIVERSAL)
-			continue;
+		/* Do not create overlay if MPO disabled */
+		if (amdgpu_dc_debug_mask & DC_DISABLE_MPO)
+			break;
 
-		if (!plane->blends_with_above || !plane->blends_with_below)
+		if (plane->type != DC_PLANE_TYPE_DCN_UNIVERSAL)
 			continue;
 
 		if (!plane->pixel_format_support.argb8888)
 			continue;
 
+		if (max_overlay-- == 0)
+			break;
+
 		if (initialize_plane(dm, NULL, primary_planes + i,
 				     DRM_PLANE_TYPE_OVERLAY, plane)) {
-			DRM_ERROR("KMS: Failed to initialize overlay plane\n");
+			drm_err(adev_to_drm(adev), "KMS: Failed to initialize overlay plane\n");
 			goto fail;
 		}
-
-		/* Only create one overlay plane. */
-		break;
 	}
 
 	for (i = 0; i < dm->dc->caps.max_streams; i++)
 		if (amdgpu_dm_crtc_init(dm, mode_info->planes[i], i)) {
-			DRM_ERROR("KMS: Failed to initialize crtc\n");
+			drm_err(adev_to_drm(adev), "KMS: Failed to initialize crtc\n");
 			goto fail;
 		}
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	/* Use Outbox interrupt */
-	switch (adev->ip_versions[DCE_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, DCE_HWIP, 0)) {
 	case IP_VERSION(3, 0, 0):
 	case IP_VERSION(3, 1, 2):
 	case IP_VERSION(3, 1, 3):
+	case IP_VERSION(3, 1, 4):
+	case IP_VERSION(3, 1, 5):
+	case IP_VERSION(3, 1, 6):
+	case IP_VERSION(3, 2, 0):
+	case IP_VERSION(3, 2, 1):
 	case IP_VERSION(2, 1, 0):
+	case IP_VERSION(3, 5, 0):
+	case IP_VERSION(3, 5, 1):
+	case IP_VERSION(3, 6, 0):
+	case IP_VERSION(4, 0, 1):
 		if (register_outbox_irq_handlers(dm->adev)) {
-			DRM_ERROR("DM: Failed to initialize IRQ\n");
+			drm_err(adev_to_drm(adev), "DM: Failed to initialize IRQ\n");
 			goto fail;
 		}
 		break;
 	default:
 		DRM_DEBUG_KMS("Unsupported DCN IP version for outbox: 0x%X\n",
-			      adev->ip_versions[DCE_HWIP][0]);
+			      amdgpu_ip_version(adev, DCE_HWIP, 0));
 	}
 
 	/* Determine whether to enable PSR support by default. */
 	if (!(amdgpu_dc_debug_mask & DC_DISABLE_PSR)) {
-		switch (adev->ip_versions[DCE_HWIP][0]) {
+		switch (amdgpu_ip_version(adev, DCE_HWIP, 0)) {
 		case IP_VERSION(3, 1, 2):
 		case IP_VERSION(3, 1, 3):
+		case IP_VERSION(3, 1, 4):
+		case IP_VERSION(3, 1, 5):
+		case IP_VERSION(3, 1, 6):
+		case IP_VERSION(3, 2, 0):
+		case IP_VERSION(3, 2, 1):
+		case IP_VERSION(3, 5, 0):
+		case IP_VERSION(3, 5, 1):
+		case IP_VERSION(3, 6, 0):
+		case IP_VERSION(4, 0, 1):
 			psr_feature_enabled = true;
 			break;
 		default:
@@ -4229,16 +5527,55 @@ static int amdgpu_dm_initialize_drm_device(struct amdgpu_device *adev)
 			break;
 		}
 	}
-#endif
+
+	/* Determine whether to enable Replay support by default. */
+	if (!(amdgpu_dc_debug_mask & DC_DISABLE_REPLAY)) {
+		switch (amdgpu_ip_version(adev, DCE_HWIP, 0)) {
+		case IP_VERSION(3, 1, 4):
+		case IP_VERSION(3, 2, 0):
+		case IP_VERSION(3, 2, 1):
+		case IP_VERSION(3, 5, 0):
+		case IP_VERSION(3, 5, 1):
+		case IP_VERSION(3, 6, 0):
+			replay_feature_enabled = true;
+			break;
+
+		default:
+			replay_feature_enabled = amdgpu_dc_feature_mask & DC_REPLAY_MASK;
+			break;
+		}
+	}
+
+	if (link_cnt > MAX_LINKS) {
+		drm_err(adev_to_drm(adev),
+			"KMS: Cannot support more than %d display indexes\n",
+				MAX_LINKS);
+		goto fail;
+	}
 
 	/* loops over all connectors on the board */
 	for (i = 0; i < link_cnt; i++) {
 		struct dc_link *link = NULL;
 
-		if (i > AMDGPU_DM_MAX_DISPLAY_INDEX) {
-			DRM_ERROR(
-				"KMS: Cannot support more than %d display indexes\n",
-					AMDGPU_DM_MAX_DISPLAY_INDEX);
+		link = dc_get_link_at_index(dm->dc, i);
+
+		if (link->connector_signal == SIGNAL_TYPE_VIRTUAL) {
+			struct amdgpu_dm_wb_connector *wbcon = kzalloc(sizeof(*wbcon), GFP_KERNEL);
+
+			if (!wbcon) {
+				drm_err(adev_to_drm(adev), "KMS: Failed to allocate writeback connector\n");
+				continue;
+			}
+
+			if (amdgpu_dm_wb_connector_init(dm, wbcon, i)) {
+				drm_err(adev_to_drm(adev), "KMS: Failed to initialize writeback connector\n");
+				kfree(wbcon);
+				continue;
+			}
+
+			link->psr_settings.psr_feature_enabled = false;
+			link->psr_settings.psr_version = DC_PSR_VERSION_UNSUPPORTED;
+
 			continue;
 		}
 
@@ -4251,36 +5588,62 @@ static int amdgpu_dm_initialize_drm_device(struct amdgpu_device *adev)
 			goto fail;
 
 		if (amdgpu_dm_encoder_init(dm->ddev, aencoder, i)) {
-			DRM_ERROR("KMS: Failed to initialize encoder\n");
+			drm_err(adev_to_drm(adev), "KMS: Failed to initialize encoder\n");
 			goto fail;
 		}
 
 		if (amdgpu_dm_connector_init(dm, aconnector, i, aencoder)) {
-			DRM_ERROR("KMS: Failed to initialize connector\n");
+			drm_err(adev_to_drm(adev), "KMS: Failed to initialize connector\n");
 			goto fail;
 		}
 
-		link = dc_get_link_at_index(dm->dc, i);
+		if (dm->hpd_rx_offload_wq)
+			dm->hpd_rx_offload_wq[aconnector->base.index].aconnector =
+				aconnector;
 
-		if (!dc_link_detect_sink(link, &new_connection_type))
-			DRM_ERROR("KMS: Failed to detect connector\n");
+		if (!dc_link_detect_connection_type(link, &new_connection_type))
+			drm_err(adev_to_drm(adev), "KMS: Failed to detect connector\n");
 
 		if (aconnector->base.force && new_connection_type == dc_connection_none) {
 			emulated_link_detect(link);
 			amdgpu_dm_update_connector_after_detect(aconnector);
+		} else {
+			bool ret = false;
 
-		} else if (dc_link_detect(link, DETECT_REASON_BOOT)) {
-			amdgpu_dm_update_connector_after_detect(aconnector);
-			register_backlight_device(dm, link);
-			if (dm->num_of_edps)
-				update_connector_ext_caps(aconnector);
-			if (psr_feature_enabled)
-				amdgpu_dm_set_psr_caps(link);
-		}
-
+			mutex_lock(&dm->dc_lock);
+			dc_exit_ips_for_hw_access(dm->dc);
+			ret = dc_link_detect(link, DETECT_REASON_BOOT);
+			mutex_unlock(&dm->dc_lock);
 
+			if (ret) {
+				amdgpu_dm_update_connector_after_detect(aconnector);
+				setup_backlight_device(dm, aconnector);
+
+				/* Disable PSR if Replay can be enabled */
+				if (replay_feature_enabled)
+					if (amdgpu_dm_set_replay_caps(link, aconnector))
+						psr_feature_enabled = false;
+
+				if (psr_feature_enabled) {
+					amdgpu_dm_set_psr_caps(link);
+					drm_info(adev_to_drm(adev), "PSR support %d, DC PSR ver %d, sink PSR ver %d DPCD caps 0x%x su_y_granularity %d\n",
+						 link->psr_settings.psr_feature_enabled,
+						 link->psr_settings.psr_version,
+						 link->dpcd_caps.psr_info.psr_version,
+						 link->dpcd_caps.psr_info.psr_dpcd_caps.raw,
+						 link->dpcd_caps.psr_info.psr2_su_y_granularity_cap);
+				}
+			}
+		}
+		amdgpu_set_panel_orientation(&aconnector->base);
 	}
 
+	/* Debug dump: list all DC links and their associated sinks after detection
+	 * is complete for all connectors. This provides a comprehensive view of the
+	 * final state without repeating the dump for each connector.
+	 */
+	amdgpu_dm_dump_links_and_sinks(adev);
+
 	/* Software is initialized. Now we can register interrupt handlers. */
 	switch (adev->asic_type) {
 #if defined(CONFIG_DRM_AMD_DC_SI)
@@ -4289,7 +5652,7 @@ static int amdgpu_dm_initialize_drm_device(struct amdgpu_device *adev)
 	case CHIP_VERDE:
 	case CHIP_OLAND:
 		if (dce60_register_irq_handlers(dm->adev)) {
-			DRM_ERROR("DM: Failed to initialize IRQ\n");
+			drm_err(adev_to_drm(adev), "DM: Failed to initialize IRQ\n");
 			goto fail;
 		}
 		break;
@@ -4311,13 +5674,12 @@ static int amdgpu_dm_initialize_drm_device(struct amdgpu_device *adev)
 	case CHIP_VEGA12:
 	case CHIP_VEGA20:
 		if (dce110_register_irq_handlers(dm->adev)) {
-			DRM_ERROR("DM: Failed to initialize IRQ\n");
+			drm_err(adev_to_drm(adev), "DM: Failed to initialize IRQ\n");
 			goto fail;
 		}
 		break;
 	default:
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-		switch (adev->ip_versions[DCE_HWIP][0]) {
+		switch (amdgpu_ip_version(adev, DCE_HWIP, 0)) {
 		case IP_VERSION(1, 0, 0):
 		case IP_VERSION(1, 0, 1):
 		case IP_VERSION(2, 0, 2):
@@ -4330,17 +5692,25 @@ static int amdgpu_dm_initialize_drm_device(struct amdgpu_device *adev)
 		case IP_VERSION(3, 0, 1):
 		case IP_VERSION(3, 1, 2):
 		case IP_VERSION(3, 1, 3):
+		case IP_VERSION(3, 1, 4):
+		case IP_VERSION(3, 1, 5):
+		case IP_VERSION(3, 1, 6):
+		case IP_VERSION(3, 2, 0):
+		case IP_VERSION(3, 2, 1):
+		case IP_VERSION(3, 5, 0):
+		case IP_VERSION(3, 5, 1):
+		case IP_VERSION(3, 6, 0):
+		case IP_VERSION(4, 0, 1):
 			if (dcn10_register_irq_handlers(dm->adev)) {
-				DRM_ERROR("DM: Failed to initialize IRQ\n");
+				drm_err(adev_to_drm(adev), "DM: Failed to initialize IRQ\n");
 				goto fail;
 			}
 			break;
 		default:
-			DRM_ERROR("Unsupported DCE IP versions: 0x%X\n",
-					adev->ip_versions[DCE_HWIP][0]);
+			drm_err(adev_to_drm(adev), "Unsupported DCE IP versions: 0x%X\n",
+					amdgpu_ip_version(adev, DCE_HWIP, 0));
 			goto fail;
 		}
-#endif
 		break;
 	}
 
@@ -4354,8 +5724,8 @@ fail:
 
 static void amdgpu_dm_destroy_drm_device(struct amdgpu_display_manager *dm)
 {
-	drm_atomic_private_obj_fini(&dm->atomic_obj);
-	return;
+	if (dm->atomic_obj.state)
+		drm_atomic_private_obj_fini(&dm->atomic_obj);
 }
 
 /******************************************************************************
@@ -4399,15 +5769,20 @@ static ssize_t s3_debug_store(struct device *device,
 	int s3_state;
 	struct drm_device *drm_dev = dev_get_drvdata(device);
 	struct amdgpu_device *adev = drm_to_adev(drm_dev);
+	struct amdgpu_ip_block *ip_block;
+
+	ip_block = amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_DCE);
+	if (!ip_block)
+		return -EINVAL;
 
 	ret = kstrtoint(buf, 0, &s3_state);
 
 	if (ret == 0) {
 		if (s3_state) {
-			dm_resume(adev);
+			dm_resume(ip_block);
 			drm_kms_helper_hotplug_event(adev_to_drm(adev));
 		} else
-			dm_suspend(adev);
+			dm_suspend(ip_block);
 	}
 
 	return ret == 0 ? count : 0;
@@ -4417,9 +5792,86 @@ DEVICE_ATTR_WO(s3_debug);
 
 #endif
 
-static int dm_early_init(void *handle)
+static int dm_init_microcode(struct amdgpu_device *adev)
+{
+	char *fw_name_dmub;
+	int r;
+
+	switch (amdgpu_ip_version(adev, DCE_HWIP, 0)) {
+	case IP_VERSION(2, 1, 0):
+		fw_name_dmub = FIRMWARE_RENOIR_DMUB;
+		if (ASICREV_IS_GREEN_SARDINE(adev->external_rev_id))
+			fw_name_dmub = FIRMWARE_GREEN_SARDINE_DMUB;
+		break;
+	case IP_VERSION(3, 0, 0):
+		if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(10, 3, 0))
+			fw_name_dmub = FIRMWARE_SIENNA_CICHLID_DMUB;
+		else
+			fw_name_dmub = FIRMWARE_NAVY_FLOUNDER_DMUB;
+		break;
+	case IP_VERSION(3, 0, 1):
+		fw_name_dmub = FIRMWARE_VANGOGH_DMUB;
+		break;
+	case IP_VERSION(3, 0, 2):
+		fw_name_dmub = FIRMWARE_DIMGREY_CAVEFISH_DMUB;
+		break;
+	case IP_VERSION(3, 0, 3):
+		fw_name_dmub = FIRMWARE_BEIGE_GOBY_DMUB;
+		break;
+	case IP_VERSION(3, 1, 2):
+	case IP_VERSION(3, 1, 3):
+		fw_name_dmub = FIRMWARE_YELLOW_CARP_DMUB;
+		break;
+	case IP_VERSION(3, 1, 4):
+		fw_name_dmub = FIRMWARE_DCN_314_DMUB;
+		break;
+	case IP_VERSION(3, 1, 5):
+		fw_name_dmub = FIRMWARE_DCN_315_DMUB;
+		break;
+	case IP_VERSION(3, 1, 6):
+		fw_name_dmub = FIRMWARE_DCN316_DMUB;
+		break;
+	case IP_VERSION(3, 2, 0):
+		fw_name_dmub = FIRMWARE_DCN_V3_2_0_DMCUB;
+		break;
+	case IP_VERSION(3, 2, 1):
+		fw_name_dmub = FIRMWARE_DCN_V3_2_1_DMCUB;
+		break;
+	case IP_VERSION(3, 5, 0):
+		fw_name_dmub = FIRMWARE_DCN_35_DMUB;
+		break;
+	case IP_VERSION(3, 5, 1):
+		fw_name_dmub = FIRMWARE_DCN_351_DMUB;
+		break;
+	case IP_VERSION(3, 6, 0):
+		fw_name_dmub = FIRMWARE_DCN_36_DMUB;
+		break;
+	case IP_VERSION(4, 0, 1):
+		fw_name_dmub = FIRMWARE_DCN_401_DMUB;
+		break;
+	default:
+		/* ASIC doesn't support DMUB. */
+		return 0;
+	}
+	r = amdgpu_ucode_request(adev, &adev->dm.dmub_fw, AMDGPU_UCODE_REQUIRED,
+				 "%s", fw_name_dmub);
+	return r;
+}
+
+static int dm_early_init(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
+	struct amdgpu_mode_info *mode_info = &adev->mode_info;
+	struct atom_context *ctx = mode_info->atom_context;
+	int index = GetIndexIntoMasterTable(DATA, Object_Header);
+	u16 data_offset;
+
+	/* if there is no object header, skip DM */
+	if (!amdgpu_atom_parse_data_header(ctx, index, NULL, NULL, NULL, &data_offset)) {
+		adev->harvest_ip_mask |= AMD_HARVEST_IP_DMU_MASK;
+		drm_info(adev_to_drm(adev), "No object header, skipping DM\n");
+		return -ENOENT;
+	}
 
 	switch (adev->asic_type) {
 #if defined(CONFIG_DRM_AMD_DC_SI)
@@ -4489,8 +5941,8 @@ static int dm_early_init(void *handle)
 		adev->mode_info.num_dig = 6;
 		break;
 	default:
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-		switch (adev->ip_versions[DCE_HWIP][0]) {
+
+		switch (amdgpu_ip_version(adev, DCE_HWIP, 0)) {
 		case IP_VERSION(2, 0, 2):
 		case IP_VERSION(3, 0, 0):
 			adev->mode_info.num_crtc = 6;
@@ -4515,21 +5967,27 @@ static int dm_early_init(void *handle)
 		case IP_VERSION(2, 1, 0):
 		case IP_VERSION(3, 1, 2):
 		case IP_VERSION(3, 1, 3):
+		case IP_VERSION(3, 1, 4):
+		case IP_VERSION(3, 1, 5):
+		case IP_VERSION(3, 1, 6):
+		case IP_VERSION(3, 2, 0):
+		case IP_VERSION(3, 2, 1):
+		case IP_VERSION(3, 5, 0):
+		case IP_VERSION(3, 5, 1):
+		case IP_VERSION(3, 6, 0):
+		case IP_VERSION(4, 0, 1):
 			adev->mode_info.num_crtc = 4;
 			adev->mode_info.num_hpd = 4;
 			adev->mode_info.num_dig = 4;
 			break;
 		default:
-			DRM_ERROR("Unsupported DCE IP versions: 0x%x\n",
-					adev->ip_versions[DCE_HWIP][0]);
+			drm_err(adev_to_drm(adev), "Unsupported DCE IP versions: 0x%x\n",
+					amdgpu_ip_version(adev, DCE_HWIP, 0));
 			return -EINVAL;
 		}
-#endif
 		break;
 	}
 
-	amdgpu_dm_set_irq_funcs(adev);
-
 	if (adev->mode_info.funcs == NULL)
 		adev->mode_info.funcs = &dm_display_funcs;
 
@@ -4543,15 +6001,9 @@ static int dm_early_init(void *handle)
 		adev_to_drm(adev)->dev,
 		&dev_attr_s3_debug);
 #endif
+	adev->dc_enabled = true;
 
-	return 0;
-}
-
-static bool modeset_required(struct drm_crtc_state *crtc_state,
-			     struct dc_stream_state *new_stream,
-			     struct dc_stream_state *old_stream)
-{
-	return crtc_state->active && drm_atomic_crtc_needs_modeset(crtc_state);
+	return dm_init_microcode(adev);
 }
 
 static bool modereset_required(struct drm_crtc_state *crtc_state)
@@ -4569,811 +6021,6 @@ static const struct drm_encoder_funcs amdgpu_dm_encoder_funcs = {
 	.destroy = amdgpu_dm_encoder_destroy,
 };
 
-
-static void get_min_max_dc_plane_scaling(struct drm_device *dev,
-					 struct drm_framebuffer *fb,
-					 int *min_downscale, int *max_upscale)
-{
-	struct amdgpu_device *adev = drm_to_adev(dev);
-	struct dc *dc = adev->dm.dc;
-	/* Caps for all supported planes are the same on DCE and DCN 1 - 3 */
-	struct dc_plane_cap *plane_cap = &dc->caps.planes[0];
-
-	switch (fb->format->format) {
-	case DRM_FORMAT_P010:
-	case DRM_FORMAT_NV12:
-	case DRM_FORMAT_NV21:
-		*max_upscale = plane_cap->max_upscale_factor.nv12;
-		*min_downscale = plane_cap->max_downscale_factor.nv12;
-		break;
-
-	case DRM_FORMAT_XRGB16161616F:
-	case DRM_FORMAT_ARGB16161616F:
-	case DRM_FORMAT_XBGR16161616F:
-	case DRM_FORMAT_ABGR16161616F:
-		*max_upscale = plane_cap->max_upscale_factor.fp16;
-		*min_downscale = plane_cap->max_downscale_factor.fp16;
-		break;
-
-	default:
-		*max_upscale = plane_cap->max_upscale_factor.argb8888;
-		*min_downscale = plane_cap->max_downscale_factor.argb8888;
-		break;
-	}
-
-	/*
-	 * A factor of 1 in the plane_cap means to not allow scaling, ie. use a
-	 * scaling factor of 1.0 == 1000 units.
-	 */
-	if (*max_upscale == 1)
-		*max_upscale = 1000;
-
-	if (*min_downscale == 1)
-		*min_downscale = 1000;
-}
-
-
-static int fill_dc_scaling_info(struct amdgpu_device *adev,
-				const struct drm_plane_state *state,
-				struct dc_scaling_info *scaling_info)
-{
-	int scale_w, scale_h, min_downscale, max_upscale;
-
-	memset(scaling_info, 0, sizeof(*scaling_info));
-
-	/* Source is fixed 16.16 but we ignore mantissa for now... */
-	scaling_info->src_rect.x = state->src_x >> 16;
-	scaling_info->src_rect.y = state->src_y >> 16;
-
-	/*
-	 * For reasons we don't (yet) fully understand a non-zero
-	 * src_y coordinate into an NV12 buffer can cause a
-	 * system hang on DCN1x.
-	 * To avoid hangs (and maybe be overly cautious)
-	 * let's reject both non-zero src_x and src_y.
-	 *
-	 * We currently know of only one use-case to reproduce a
-	 * scenario with non-zero src_x and src_y for NV12, which
-	 * is to gesture the YouTube Android app into full screen
-	 * on ChromeOS.
-	 */
-	if (((adev->ip_versions[DCE_HWIP][0] == IP_VERSION(1, 0, 0)) ||
-	    (adev->ip_versions[DCE_HWIP][0] == IP_VERSION(1, 0, 1))) &&
-	    (state->fb && state->fb->format->format == DRM_FORMAT_NV12 &&
-	    (scaling_info->src_rect.x != 0 || scaling_info->src_rect.y != 0)))
-		return -EINVAL;
-
-	scaling_info->src_rect.width = state->src_w >> 16;
-	if (scaling_info->src_rect.width == 0)
-		return -EINVAL;
-
-	scaling_info->src_rect.height = state->src_h >> 16;
-	if (scaling_info->src_rect.height == 0)
-		return -EINVAL;
-
-	scaling_info->dst_rect.x = state->crtc_x;
-	scaling_info->dst_rect.y = state->crtc_y;
-
-	if (state->crtc_w == 0)
-		return -EINVAL;
-
-	scaling_info->dst_rect.width = state->crtc_w;
-
-	if (state->crtc_h == 0)
-		return -EINVAL;
-
-	scaling_info->dst_rect.height = state->crtc_h;
-
-	/* DRM doesn't specify clipping on destination output. */
-	scaling_info->clip_rect = scaling_info->dst_rect;
-
-	/* Validate scaling per-format with DC plane caps */
-	if (state->plane && state->plane->dev && state->fb) {
-		get_min_max_dc_plane_scaling(state->plane->dev, state->fb,
-					     &min_downscale, &max_upscale);
-	} else {
-		min_downscale = 250;
-		max_upscale = 16000;
-	}
-
-	scale_w = scaling_info->dst_rect.width * 1000 /
-		  scaling_info->src_rect.width;
-
-	if (scale_w < min_downscale || scale_w > max_upscale)
-		return -EINVAL;
-
-	scale_h = scaling_info->dst_rect.height * 1000 /
-		  scaling_info->src_rect.height;
-
-	if (scale_h < min_downscale || scale_h > max_upscale)
-		return -EINVAL;
-
-	/*
-	 * The "scaling_quality" can be ignored for now, quality = 0 has DC
-	 * assume reasonable defaults based on the format.
-	 */
-
-	return 0;
-}
-
-static void
-fill_gfx8_tiling_info_from_flags(union dc_tiling_info *tiling_info,
-				 uint64_t tiling_flags)
-{
-	/* Fill GFX8 params */
-	if (AMDGPU_TILING_GET(tiling_flags, ARRAY_MODE) == DC_ARRAY_2D_TILED_THIN1) {
-		unsigned int bankw, bankh, mtaspect, tile_split, num_banks;
-
-		bankw = AMDGPU_TILING_GET(tiling_flags, BANK_WIDTH);
-		bankh = AMDGPU_TILING_GET(tiling_flags, BANK_HEIGHT);
-		mtaspect = AMDGPU_TILING_GET(tiling_flags, MACRO_TILE_ASPECT);
-		tile_split = AMDGPU_TILING_GET(tiling_flags, TILE_SPLIT);
-		num_banks = AMDGPU_TILING_GET(tiling_flags, NUM_BANKS);
-
-		/* XXX fix me for VI */
-		tiling_info->gfx8.num_banks = num_banks;
-		tiling_info->gfx8.array_mode =
-				DC_ARRAY_2D_TILED_THIN1;
-		tiling_info->gfx8.tile_split = tile_split;
-		tiling_info->gfx8.bank_width = bankw;
-		tiling_info->gfx8.bank_height = bankh;
-		tiling_info->gfx8.tile_aspect = mtaspect;
-		tiling_info->gfx8.tile_mode =
-				DC_ADDR_SURF_MICRO_TILING_DISPLAY;
-	} else if (AMDGPU_TILING_GET(tiling_flags, ARRAY_MODE)
-			== DC_ARRAY_1D_TILED_THIN1) {
-		tiling_info->gfx8.array_mode = DC_ARRAY_1D_TILED_THIN1;
-	}
-
-	tiling_info->gfx8.pipe_config =
-			AMDGPU_TILING_GET(tiling_flags, PIPE_CONFIG);
-}
-
-static void
-fill_gfx9_tiling_info_from_device(const struct amdgpu_device *adev,
-				  union dc_tiling_info *tiling_info)
-{
-	tiling_info->gfx9.num_pipes =
-		adev->gfx.config.gb_addr_config_fields.num_pipes;
-	tiling_info->gfx9.num_banks =
-		adev->gfx.config.gb_addr_config_fields.num_banks;
-	tiling_info->gfx9.pipe_interleave =
-		adev->gfx.config.gb_addr_config_fields.pipe_interleave_size;
-	tiling_info->gfx9.num_shader_engines =
-		adev->gfx.config.gb_addr_config_fields.num_se;
-	tiling_info->gfx9.max_compressed_frags =
-		adev->gfx.config.gb_addr_config_fields.max_compress_frags;
-	tiling_info->gfx9.num_rb_per_se =
-		adev->gfx.config.gb_addr_config_fields.num_rb_per_se;
-	tiling_info->gfx9.shaderEnable = 1;
-	if (adev->ip_versions[GC_HWIP][0] >= IP_VERSION(10, 3, 0))
-		tiling_info->gfx9.num_pkrs = adev->gfx.config.gb_addr_config_fields.num_pkrs;
-}
-
-static int
-validate_dcc(struct amdgpu_device *adev,
-	     const enum surface_pixel_format format,
-	     const enum dc_rotation_angle rotation,
-	     const union dc_tiling_info *tiling_info,
-	     const struct dc_plane_dcc_param *dcc,
-	     const struct dc_plane_address *address,
-	     const struct plane_size *plane_size)
-{
-	struct dc *dc = adev->dm.dc;
-	struct dc_dcc_surface_param input;
-	struct dc_surface_dcc_cap output;
-
-	memset(&input, 0, sizeof(input));
-	memset(&output, 0, sizeof(output));
-
-	if (!dcc->enable)
-		return 0;
-
-	if (format >= SURFACE_PIXEL_FORMAT_VIDEO_BEGIN ||
-	    !dc->cap_funcs.get_dcc_compression_cap)
-		return -EINVAL;
-
-	input.format = format;
-	input.surface_size.width = plane_size->surface_size.width;
-	input.surface_size.height = plane_size->surface_size.height;
-	input.swizzle_mode = tiling_info->gfx9.swizzle;
-
-	if (rotation == ROTATION_ANGLE_0 || rotation == ROTATION_ANGLE_180)
-		input.scan = SCAN_DIRECTION_HORIZONTAL;
-	else if (rotation == ROTATION_ANGLE_90 || rotation == ROTATION_ANGLE_270)
-		input.scan = SCAN_DIRECTION_VERTICAL;
-
-	if (!dc->cap_funcs.get_dcc_compression_cap(dc, &input, &output))
-		return -EINVAL;
-
-	if (!output.capable)
-		return -EINVAL;
-
-	if (dcc->independent_64b_blks == 0 &&
-	    output.grph.rgb.independent_64b_blks != 0)
-		return -EINVAL;
-
-	return 0;
-}
-
-static bool
-modifier_has_dcc(uint64_t modifier)
-{
-	return IS_AMD_FMT_MOD(modifier) && AMD_FMT_MOD_GET(DCC, modifier);
-}
-
-static unsigned
-modifier_gfx9_swizzle_mode(uint64_t modifier)
-{
-	if (modifier == DRM_FORMAT_MOD_LINEAR)
-		return 0;
-
-	return AMD_FMT_MOD_GET(TILE, modifier);
-}
-
-static const struct drm_format_info *
-amd_get_format_info(const struct drm_mode_fb_cmd2 *cmd)
-{
-	return amdgpu_lookup_format_info(cmd->pixel_format, cmd->modifier[0]);
-}
-
-static void
-fill_gfx9_tiling_info_from_modifier(const struct amdgpu_device *adev,
-				    union dc_tiling_info *tiling_info,
-				    uint64_t modifier)
-{
-	unsigned int mod_bank_xor_bits = AMD_FMT_MOD_GET(BANK_XOR_BITS, modifier);
-	unsigned int mod_pipe_xor_bits = AMD_FMT_MOD_GET(PIPE_XOR_BITS, modifier);
-	unsigned int pkrs_log2 = AMD_FMT_MOD_GET(PACKERS, modifier);
-	unsigned int pipes_log2 = min(4u, mod_pipe_xor_bits);
-
-	fill_gfx9_tiling_info_from_device(adev, tiling_info);
-
-	if (!IS_AMD_FMT_MOD(modifier))
-		return;
-
-	tiling_info->gfx9.num_pipes = 1u << pipes_log2;
-	tiling_info->gfx9.num_shader_engines = 1u << (mod_pipe_xor_bits - pipes_log2);
-
-	if (adev->family >= AMDGPU_FAMILY_NV) {
-		tiling_info->gfx9.num_pkrs = 1u << pkrs_log2;
-	} else {
-		tiling_info->gfx9.num_banks = 1u << mod_bank_xor_bits;
-
-		/* for DCC we know it isn't rb aligned, so rb_per_se doesn't matter. */
-	}
-}
-
-enum dm_micro_swizzle {
-	MICRO_SWIZZLE_Z = 0,
-	MICRO_SWIZZLE_S = 1,
-	MICRO_SWIZZLE_D = 2,
-	MICRO_SWIZZLE_R = 3
-};
-
-static bool dm_plane_format_mod_supported(struct drm_plane *plane,
-					  uint32_t format,
-					  uint64_t modifier)
-{
-	struct amdgpu_device *adev = drm_to_adev(plane->dev);
-	const struct drm_format_info *info = drm_format_info(format);
-	int i;
-
-	enum dm_micro_swizzle microtile = modifier_gfx9_swizzle_mode(modifier) & 3;
-
-	if (!info)
-		return false;
-
-	/*
-	 * We always have to allow these modifiers:
-	 * 1. Core DRM checks for LINEAR support if userspace does not provide modifiers.
-	 * 2. Not passing any modifiers is the same as explicitly passing INVALID.
-	 */
-	if (modifier == DRM_FORMAT_MOD_LINEAR ||
-	    modifier == DRM_FORMAT_MOD_INVALID) {
-		return true;
-	}
-
-	/* Check that the modifier is on the list of the plane's supported modifiers. */
-	for (i = 0; i < plane->modifier_count; i++) {
-		if (modifier == plane->modifiers[i])
-			break;
-	}
-	if (i == plane->modifier_count)
-		return false;
-
-	/*
-	 * For D swizzle the canonical modifier depends on the bpp, so check
-	 * it here.
-	 */
-	if (AMD_FMT_MOD_GET(TILE_VERSION, modifier) == AMD_FMT_MOD_TILE_VER_GFX9 &&
-	    adev->family >= AMDGPU_FAMILY_NV) {
-		if (microtile == MICRO_SWIZZLE_D && info->cpp[0] == 4)
-			return false;
-	}
-
-	if (adev->family >= AMDGPU_FAMILY_RV && microtile == MICRO_SWIZZLE_D &&
-	    info->cpp[0] < 8)
-		return false;
-
-	if (modifier_has_dcc(modifier)) {
-		/* Per radeonsi comments 16/64 bpp are more complicated. */
-		if (info->cpp[0] != 4)
-			return false;
-		/* We support multi-planar formats, but not when combined with
-		 * additional DCC metadata planes. */
-		if (info->num_planes > 1)
-			return false;
-	}
-
-	return true;
-}
-
-static void
-add_modifier(uint64_t **mods, uint64_t *size, uint64_t *cap, uint64_t mod)
-{
-	if (!*mods)
-		return;
-
-	if (*cap - *size < 1) {
-		uint64_t new_cap = *cap * 2;
-		uint64_t *new_mods = kmalloc(new_cap * sizeof(uint64_t), GFP_KERNEL);
-
-		if (!new_mods) {
-			kfree(*mods);
-			*mods = NULL;
-			return;
-		}
-
-		memcpy(new_mods, *mods, sizeof(uint64_t) * *size);
-		kfree(*mods);
-		*mods = new_mods;
-		*cap = new_cap;
-	}
-
-	(*mods)[*size] = mod;
-	*size += 1;
-}
-
-static void
-add_gfx9_modifiers(const struct amdgpu_device *adev,
-		   uint64_t **mods, uint64_t *size, uint64_t *capacity)
-{
-	int pipes = ilog2(adev->gfx.config.gb_addr_config_fields.num_pipes);
-	int pipe_xor_bits = min(8, pipes +
-				ilog2(adev->gfx.config.gb_addr_config_fields.num_se));
-	int bank_xor_bits = min(8 - pipe_xor_bits,
-				ilog2(adev->gfx.config.gb_addr_config_fields.num_banks));
-	int rb = ilog2(adev->gfx.config.gb_addr_config_fields.num_se) +
-		 ilog2(adev->gfx.config.gb_addr_config_fields.num_rb_per_se);
-
-
-	if (adev->family == AMDGPU_FAMILY_RV) {
-		/* Raven2 and later */
-		bool has_constant_encode = adev->asic_type > CHIP_RAVEN || adev->external_rev_id >= 0x81;
-
-		/*
-		 * No _D DCC swizzles yet because we only allow 32bpp, which
-		 * doesn't support _D on DCN
-		 */
-
-		if (has_constant_encode) {
-			add_modifier(mods, size, capacity, AMD_FMT_MOD |
-				    AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_S_X) |
-				    AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX9) |
-				    AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
-				    AMD_FMT_MOD_SET(BANK_XOR_BITS, bank_xor_bits) |
-				    AMD_FMT_MOD_SET(DCC, 1) |
-				    AMD_FMT_MOD_SET(DCC_INDEPENDENT_64B, 1) |
-				    AMD_FMT_MOD_SET(DCC_MAX_COMPRESSED_BLOCK, AMD_FMT_MOD_DCC_BLOCK_64B) |
-				    AMD_FMT_MOD_SET(DCC_CONSTANT_ENCODE, 1));
-		}
-
-		add_modifier(mods, size, capacity, AMD_FMT_MOD |
-			    AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_S_X) |
-			    AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX9) |
-			    AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
-			    AMD_FMT_MOD_SET(BANK_XOR_BITS, bank_xor_bits) |
-			    AMD_FMT_MOD_SET(DCC, 1) |
-			    AMD_FMT_MOD_SET(DCC_INDEPENDENT_64B, 1) |
-			    AMD_FMT_MOD_SET(DCC_MAX_COMPRESSED_BLOCK, AMD_FMT_MOD_DCC_BLOCK_64B) |
-			    AMD_FMT_MOD_SET(DCC_CONSTANT_ENCODE, 0));
-
-		if (has_constant_encode) {
-			add_modifier(mods, size, capacity, AMD_FMT_MOD |
-				    AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_S_X) |
-				    AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX9) |
-				    AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
-				    AMD_FMT_MOD_SET(BANK_XOR_BITS, bank_xor_bits) |
-				    AMD_FMT_MOD_SET(DCC, 1) |
-				    AMD_FMT_MOD_SET(DCC_RETILE, 1) |
-				    AMD_FMT_MOD_SET(DCC_INDEPENDENT_64B, 1) |
-				    AMD_FMT_MOD_SET(DCC_MAX_COMPRESSED_BLOCK, AMD_FMT_MOD_DCC_BLOCK_64B) |
-
-				    AMD_FMT_MOD_SET(DCC_CONSTANT_ENCODE, 1) |
-				    AMD_FMT_MOD_SET(RB, rb) |
-				    AMD_FMT_MOD_SET(PIPE, pipes));
-		}
-
-		add_modifier(mods, size, capacity, AMD_FMT_MOD |
-			    AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_S_X) |
-			    AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX9) |
-			    AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
-			    AMD_FMT_MOD_SET(BANK_XOR_BITS, bank_xor_bits) |
-			    AMD_FMT_MOD_SET(DCC, 1) |
-			    AMD_FMT_MOD_SET(DCC_RETILE, 1) |
-			    AMD_FMT_MOD_SET(DCC_INDEPENDENT_64B, 1) |
-			    AMD_FMT_MOD_SET(DCC_MAX_COMPRESSED_BLOCK, AMD_FMT_MOD_DCC_BLOCK_64B) |
-			    AMD_FMT_MOD_SET(DCC_CONSTANT_ENCODE, 0) |
-			    AMD_FMT_MOD_SET(RB, rb) |
-			    AMD_FMT_MOD_SET(PIPE, pipes));
-	}
-
-	/*
-	 * Only supported for 64bpp on Raven, will be filtered on format in
-	 * dm_plane_format_mod_supported.
-	 */
-	add_modifier(mods, size, capacity, AMD_FMT_MOD |
-		    AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_D_X) |
-		    AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX9) |
-		    AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
-		    AMD_FMT_MOD_SET(BANK_XOR_BITS, bank_xor_bits));
-
-	if (adev->family == AMDGPU_FAMILY_RV) {
-		add_modifier(mods, size, capacity, AMD_FMT_MOD |
-			    AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_S_X) |
-			    AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX9) |
-			    AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
-			    AMD_FMT_MOD_SET(BANK_XOR_BITS, bank_xor_bits));
-	}
-
-	/*
-	 * Only supported for 64bpp on Raven, will be filtered on format in
-	 * dm_plane_format_mod_supported.
-	 */
-	add_modifier(mods, size, capacity, AMD_FMT_MOD |
-		    AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_D) |
-		    AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX9));
-
-	if (adev->family == AMDGPU_FAMILY_RV) {
-		add_modifier(mods, size, capacity, AMD_FMT_MOD |
-			    AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_S) |
-			    AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX9));
-	}
-}
-
-static void
-add_gfx10_1_modifiers(const struct amdgpu_device *adev,
-		      uint64_t **mods, uint64_t *size, uint64_t *capacity)
-{
-	int pipe_xor_bits = ilog2(adev->gfx.config.gb_addr_config_fields.num_pipes);
-
-	add_modifier(mods, size, capacity, AMD_FMT_MOD |
-		    AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_R_X) |
-		    AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX10) |
-		    AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
-		    AMD_FMT_MOD_SET(DCC, 1) |
-		    AMD_FMT_MOD_SET(DCC_CONSTANT_ENCODE, 1) |
-		    AMD_FMT_MOD_SET(DCC_INDEPENDENT_64B, 1) |
-		    AMD_FMT_MOD_SET(DCC_MAX_COMPRESSED_BLOCK, AMD_FMT_MOD_DCC_BLOCK_64B));
-
-	add_modifier(mods, size, capacity, AMD_FMT_MOD |
-		    AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_R_X) |
-		    AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX10) |
-		    AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
-		    AMD_FMT_MOD_SET(DCC, 1) |
-		    AMD_FMT_MOD_SET(DCC_RETILE, 1) |
-		    AMD_FMT_MOD_SET(DCC_CONSTANT_ENCODE, 1) |
-		    AMD_FMT_MOD_SET(DCC_INDEPENDENT_64B, 1) |
-		    AMD_FMT_MOD_SET(DCC_MAX_COMPRESSED_BLOCK, AMD_FMT_MOD_DCC_BLOCK_64B));
-
-	add_modifier(mods, size, capacity, AMD_FMT_MOD |
-		    AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_R_X) |
-		    AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX10) |
-		    AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits));
-
-	add_modifier(mods, size, capacity, AMD_FMT_MOD |
-		    AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_S_X) |
-		    AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX10) |
-		    AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits));
-
-
-	/* Only supported for 64bpp, will be filtered in dm_plane_format_mod_supported */
-	add_modifier(mods, size, capacity, AMD_FMT_MOD |
-		    AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_D) |
-		    AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX9));
-
-	add_modifier(mods, size, capacity, AMD_FMT_MOD |
-		    AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_S) |
-		    AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX9));
-}
-
-static void
-add_gfx10_3_modifiers(const struct amdgpu_device *adev,
-		      uint64_t **mods, uint64_t *size, uint64_t *capacity)
-{
-	int pipe_xor_bits = ilog2(adev->gfx.config.gb_addr_config_fields.num_pipes);
-	int pkrs = ilog2(adev->gfx.config.gb_addr_config_fields.num_pkrs);
-
-	add_modifier(mods, size, capacity, AMD_FMT_MOD |
-		    AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_R_X) |
-		    AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX10_RBPLUS) |
-		    AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
-		    AMD_FMT_MOD_SET(PACKERS, pkrs) |
-		    AMD_FMT_MOD_SET(DCC, 1) |
-		    AMD_FMT_MOD_SET(DCC_CONSTANT_ENCODE, 1) |
-		    AMD_FMT_MOD_SET(DCC_INDEPENDENT_64B, 1) |
-		    AMD_FMT_MOD_SET(DCC_INDEPENDENT_128B, 1) |
-		    AMD_FMT_MOD_SET(DCC_MAX_COMPRESSED_BLOCK, AMD_FMT_MOD_DCC_BLOCK_64B));
-
-	add_modifier(mods, size, capacity, AMD_FMT_MOD |
-		    AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_R_X) |
-		    AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX10_RBPLUS) |
-		    AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
-		    AMD_FMT_MOD_SET(PACKERS, pkrs) |
-		    AMD_FMT_MOD_SET(DCC, 1) |
-		    AMD_FMT_MOD_SET(DCC_CONSTANT_ENCODE, 1) |
-		    AMD_FMT_MOD_SET(DCC_INDEPENDENT_128B, 1) |
-		    AMD_FMT_MOD_SET(DCC_MAX_COMPRESSED_BLOCK, AMD_FMT_MOD_DCC_BLOCK_128B));
-
-	add_modifier(mods, size, capacity, AMD_FMT_MOD |
-		    AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_R_X) |
-		    AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX10_RBPLUS) |
-		    AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
-		    AMD_FMT_MOD_SET(PACKERS, pkrs) |
-		    AMD_FMT_MOD_SET(DCC, 1) |
-		    AMD_FMT_MOD_SET(DCC_RETILE, 1) |
-		    AMD_FMT_MOD_SET(DCC_CONSTANT_ENCODE, 1) |
-		    AMD_FMT_MOD_SET(DCC_INDEPENDENT_64B, 1) |
-		    AMD_FMT_MOD_SET(DCC_INDEPENDENT_128B, 1) |
-		    AMD_FMT_MOD_SET(DCC_MAX_COMPRESSED_BLOCK, AMD_FMT_MOD_DCC_BLOCK_64B));
-
-	add_modifier(mods, size, capacity, AMD_FMT_MOD |
-		    AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_R_X) |
-		    AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX10_RBPLUS) |
-		    AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
-		    AMD_FMT_MOD_SET(PACKERS, pkrs) |
-		    AMD_FMT_MOD_SET(DCC, 1) |
-		    AMD_FMT_MOD_SET(DCC_RETILE, 1) |
-		    AMD_FMT_MOD_SET(DCC_CONSTANT_ENCODE, 1) |
-		    AMD_FMT_MOD_SET(DCC_INDEPENDENT_128B, 1) |
-		    AMD_FMT_MOD_SET(DCC_MAX_COMPRESSED_BLOCK, AMD_FMT_MOD_DCC_BLOCK_128B));
-
-	add_modifier(mods, size, capacity, AMD_FMT_MOD |
-		    AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_R_X) |
-		    AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX10_RBPLUS) |
-		    AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
-		    AMD_FMT_MOD_SET(PACKERS, pkrs));
-
-	add_modifier(mods, size, capacity, AMD_FMT_MOD |
-		    AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_S_X) |
-		    AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX10_RBPLUS) |
-		    AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
-		    AMD_FMT_MOD_SET(PACKERS, pkrs));
-
-	/* Only supported for 64bpp, will be filtered in dm_plane_format_mod_supported */
-	add_modifier(mods, size, capacity, AMD_FMT_MOD |
-		    AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_D) |
-		    AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX9));
-
-	add_modifier(mods, size, capacity, AMD_FMT_MOD |
-		    AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_S) |
-		    AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX9));
-}
-
-static int
-get_plane_modifiers(const struct amdgpu_device *adev, unsigned int plane_type, uint64_t **mods)
-{
-	uint64_t size = 0, capacity = 128;
-	*mods = NULL;
-
-	/* We have not hooked up any pre-GFX9 modifiers. */
-	if (adev->family < AMDGPU_FAMILY_AI)
-		return 0;
-
-	*mods = kmalloc(capacity * sizeof(uint64_t), GFP_KERNEL);
-
-	if (plane_type == DRM_PLANE_TYPE_CURSOR) {
-		add_modifier(mods, &size, &capacity, DRM_FORMAT_MOD_LINEAR);
-		add_modifier(mods, &size, &capacity, DRM_FORMAT_MOD_INVALID);
-		return *mods ? 0 : -ENOMEM;
-	}
-
-	switch (adev->family) {
-	case AMDGPU_FAMILY_AI:
-	case AMDGPU_FAMILY_RV:
-		add_gfx9_modifiers(adev, mods, &size, &capacity);
-		break;
-	case AMDGPU_FAMILY_NV:
-	case AMDGPU_FAMILY_VGH:
-	case AMDGPU_FAMILY_YC:
-		if (adev->ip_versions[GC_HWIP][0] >= IP_VERSION(10, 3, 0))
-			add_gfx10_3_modifiers(adev, mods, &size, &capacity);
-		else
-			add_gfx10_1_modifiers(adev, mods, &size, &capacity);
-		break;
-	}
-
-	add_modifier(mods, &size, &capacity, DRM_FORMAT_MOD_LINEAR);
-
-	/* INVALID marks the end of the list. */
-	add_modifier(mods, &size, &capacity, DRM_FORMAT_MOD_INVALID);
-
-	if (!*mods)
-		return -ENOMEM;
-
-	return 0;
-}
-
-static int
-fill_gfx9_plane_attributes_from_modifiers(struct amdgpu_device *adev,
-					  const struct amdgpu_framebuffer *afb,
-					  const enum surface_pixel_format format,
-					  const enum dc_rotation_angle rotation,
-					  const struct plane_size *plane_size,
-					  union dc_tiling_info *tiling_info,
-					  struct dc_plane_dcc_param *dcc,
-					  struct dc_plane_address *address,
-					  const bool force_disable_dcc)
-{
-	const uint64_t modifier = afb->base.modifier;
-	int ret = 0;
-
-	fill_gfx9_tiling_info_from_modifier(adev, tiling_info, modifier);
-	tiling_info->gfx9.swizzle = modifier_gfx9_swizzle_mode(modifier);
-
-	if (modifier_has_dcc(modifier) && !force_disable_dcc) {
-		uint64_t dcc_address = afb->address + afb->base.offsets[1];
-		bool independent_64b_blks = AMD_FMT_MOD_GET(DCC_INDEPENDENT_64B, modifier);
-		bool independent_128b_blks = AMD_FMT_MOD_GET(DCC_INDEPENDENT_128B, modifier);
-
-		dcc->enable = 1;
-		dcc->meta_pitch = afb->base.pitches[1];
-		dcc->independent_64b_blks = independent_64b_blks;
-		if (AMD_FMT_MOD_GET(TILE_VERSION, modifier) == AMD_FMT_MOD_TILE_VER_GFX10_RBPLUS) {
-			if (independent_64b_blks && independent_128b_blks)
-				dcc->dcc_ind_blk = hubp_ind_block_64b_no_128bcl;
-			else if (independent_128b_blks)
-				dcc->dcc_ind_blk = hubp_ind_block_128b;
-			else if (independent_64b_blks && !independent_128b_blks)
-				dcc->dcc_ind_blk = hubp_ind_block_64b;
-			else
-				dcc->dcc_ind_blk = hubp_ind_block_unconstrained;
-		} else {
-			if (independent_64b_blks)
-				dcc->dcc_ind_blk = hubp_ind_block_64b;
-			else
-				dcc->dcc_ind_blk = hubp_ind_block_unconstrained;
-		}
-
-		address->grph.meta_addr.low_part = lower_32_bits(dcc_address);
-		address->grph.meta_addr.high_part = upper_32_bits(dcc_address);
-	}
-
-	ret = validate_dcc(adev, format, rotation, tiling_info, dcc, address, plane_size);
-	if (ret)
-		drm_dbg_kms(adev_to_drm(adev), "validate_dcc: returned error: %d\n", ret);
-
-	return ret;
-}
-
-static int
-fill_plane_buffer_attributes(struct amdgpu_device *adev,
-			     const struct amdgpu_framebuffer *afb,
-			     const enum surface_pixel_format format,
-			     const enum dc_rotation_angle rotation,
-			     const uint64_t tiling_flags,
-			     union dc_tiling_info *tiling_info,
-			     struct plane_size *plane_size,
-			     struct dc_plane_dcc_param *dcc,
-			     struct dc_plane_address *address,
-			     bool tmz_surface,
-			     bool force_disable_dcc)
-{
-	const struct drm_framebuffer *fb = &afb->base;
-	int ret;
-
-	memset(tiling_info, 0, sizeof(*tiling_info));
-	memset(plane_size, 0, sizeof(*plane_size));
-	memset(dcc, 0, sizeof(*dcc));
-	memset(address, 0, sizeof(*address));
-
-	address->tmz_surface = tmz_surface;
-
-	if (format < SURFACE_PIXEL_FORMAT_VIDEO_BEGIN) {
-		uint64_t addr = afb->address + fb->offsets[0];
-
-		plane_size->surface_size.x = 0;
-		plane_size->surface_size.y = 0;
-		plane_size->surface_size.width = fb->width;
-		plane_size->surface_size.height = fb->height;
-		plane_size->surface_pitch =
-			fb->pitches[0] / fb->format->cpp[0];
-
-		address->type = PLN_ADDR_TYPE_GRAPHICS;
-		address->grph.addr.low_part = lower_32_bits(addr);
-		address->grph.addr.high_part = upper_32_bits(addr);
-	} else if (format < SURFACE_PIXEL_FORMAT_INVALID) {
-		uint64_t luma_addr = afb->address + fb->offsets[0];
-		uint64_t chroma_addr = afb->address + fb->offsets[1];
-
-		plane_size->surface_size.x = 0;
-		plane_size->surface_size.y = 0;
-		plane_size->surface_size.width = fb->width;
-		plane_size->surface_size.height = fb->height;
-		plane_size->surface_pitch =
-			fb->pitches[0] / fb->format->cpp[0];
-
-		plane_size->chroma_size.x = 0;
-		plane_size->chroma_size.y = 0;
-		/* TODO: set these based on surface format */
-		plane_size->chroma_size.width = fb->width / 2;
-		plane_size->chroma_size.height = fb->height / 2;
-
-		plane_size->chroma_pitch =
-			fb->pitches[1] / fb->format->cpp[1];
-
-		address->type = PLN_ADDR_TYPE_VIDEO_PROGRESSIVE;
-		address->video_progressive.luma_addr.low_part =
-			lower_32_bits(luma_addr);
-		address->video_progressive.luma_addr.high_part =
-			upper_32_bits(luma_addr);
-		address->video_progressive.chroma_addr.low_part =
-			lower_32_bits(chroma_addr);
-		address->video_progressive.chroma_addr.high_part =
-			upper_32_bits(chroma_addr);
-	}
-
-	if (adev->family >= AMDGPU_FAMILY_AI) {
-		ret = fill_gfx9_plane_attributes_from_modifiers(adev, afb, format,
-								rotation, plane_size,
-								tiling_info, dcc,
-								address,
-								force_disable_dcc);
-		if (ret)
-			return ret;
-	} else {
-		fill_gfx8_tiling_info_from_flags(tiling_info, tiling_flags);
-	}
-
-	return 0;
-}
-
-static void
-fill_blending_from_plane_state(const struct drm_plane_state *plane_state,
-			       bool *per_pixel_alpha, bool *global_alpha,
-			       int *global_alpha_value)
-{
-	*per_pixel_alpha = false;
-	*global_alpha = false;
-	*global_alpha_value = 0xff;
-
-	if (plane_state->plane->type != DRM_PLANE_TYPE_OVERLAY)
-		return;
-
-	if (plane_state->pixel_blend_mode == DRM_MODE_BLEND_PREMULTI) {
-		static const uint32_t alpha_formats[] = {
-			DRM_FORMAT_ARGB8888,
-			DRM_FORMAT_RGBA8888,
-			DRM_FORMAT_ABGR8888,
-		};
-		uint32_t format = plane_state->fb->format->format;
-		unsigned int i;
-
-		for (i = 0; i < ARRAY_SIZE(alpha_formats); ++i) {
-			if (format == alpha_formats[i]) {
-				*per_pixel_alpha = true;
-				break;
-			}
-		}
-	}
-
-	if (plane_state->alpha < 0xffff) {
-		*global_alpha = true;
-		*global_alpha_value = plane_state->alpha >> 8;
-	}
-}
-
 static int
 fill_plane_color_attributes(const struct drm_plane_state *plane_state,
 			    const enum surface_pixel_format format,
@@ -5383,6 +6030,10 @@ fill_plane_color_attributes(const struct drm_plane_state *plane_state,
 
 	*color_space = COLOR_SPACE_SRGB;
 
+	/* Ignore properties when DRM_CLIENT_CAP_PLANE_COLOR_PIPELINE is set */
+	if (plane_state->state && plane_state->state->plane_color_pipeline)
+		return 0;
+
 	/* DRM color properties only affect non-RGB formats. */
 	if (format < SURFACE_PIXEL_FORMAT_VIDEO_BEGIN)
 		return 0;
@@ -5406,9 +6057,9 @@ fill_plane_color_attributes(const struct drm_plane_state *plane_state,
 
 	case DRM_COLOR_YCBCR_BT2020:
 		if (full_range)
-			*color_space = COLOR_SPACE_2020_YCBCR;
+			*color_space = COLOR_SPACE_2020_YCBCR_FULL;
 		else
-			return -EINVAL;
+			*color_space = COLOR_SPACE_2020_YCBCR_LIMITED;
 		break;
 
 	default:
@@ -5421,11 +6072,10 @@ fill_plane_color_attributes(const struct drm_plane_state *plane_state,
 static int
 fill_dc_plane_info_and_addr(struct amdgpu_device *adev,
 			    const struct drm_plane_state *plane_state,
-			    const uint64_t tiling_flags,
+			    const u64 tiling_flags,
 			    struct dc_plane_info *plane_info,
 			    struct dc_plane_address *address,
-			    bool tmz_surface,
-			    bool force_disable_dcc)
+			    bool tmz_surface)
 {
 	const struct drm_framebuffer *fb = plane_state->fb;
 	const struct amdgpu_framebuffer *afb =
@@ -5484,7 +6134,7 @@ fill_dc_plane_info_and_addr(struct amdgpu_device *adev,
 		plane_info->format = SURFACE_PIXEL_FORMAT_GRPH_ABGR16161616;
 		break;
 	default:
-		DRM_ERROR(
+		drm_err(adev_to_drm(adev),
 			"Unsupported screen format %p4cc\n",
 			&fb->format->format);
 		return -EINVAL;
@@ -5508,27 +6158,28 @@ fill_dc_plane_info_and_addr(struct amdgpu_device *adev,
 		break;
 	}
 
+
 	plane_info->visible = true;
 	plane_info->stereo_format = PLANE_STEREO_FORMAT_NONE;
 
-	plane_info->layer_index = 0;
+	plane_info->layer_index = plane_state->normalized_zpos;
 
 	ret = fill_plane_color_attributes(plane_state, plane_info->format,
 					  &plane_info->color_space);
 	if (ret)
 		return ret;
 
-	ret = fill_plane_buffer_attributes(adev, afb, plane_info->format,
+	ret = amdgpu_dm_plane_fill_plane_buffer_attributes(adev, afb, plane_info->format,
 					   plane_info->rotation, tiling_flags,
 					   &plane_info->tiling_info,
 					   &plane_info->plane_size,
-					   &plane_info->dcc, address, tmz_surface,
-					   force_disable_dcc);
+					   &plane_info->dcc, address,
+					   tmz_surface);
 	if (ret)
 		return ret;
 
-	fill_blending_from_plane_state(
-		plane_state, &plane_info->per_pixel_alpha,
+	amdgpu_dm_plane_fill_blending_from_plane_state(
+		plane_state, &plane_info->per_pixel_alpha, &plane_info->pre_multiplied_alpha,
 		&plane_info->global_alpha, &plane_info->global_alpha_value);
 
 	return 0;
@@ -5544,9 +6195,8 @@ static int fill_dc_plane_attributes(struct amdgpu_device *adev,
 	struct dc_scaling_info scaling_info;
 	struct dc_plane_info plane_info;
 	int ret;
-	bool force_disable_dcc = false;
 
-	ret = fill_dc_scaling_info(adev, plane_state, &scaling_info);
+	ret = amdgpu_dm_plane_fill_dc_scaling_info(adev, plane_state, &scaling_info);
 	if (ret)
 		return ret;
 
@@ -5555,13 +6205,11 @@ static int fill_dc_plane_attributes(struct amdgpu_device *adev,
 	dc_plane_state->clip_rect = scaling_info.clip_rect;
 	dc_plane_state->scaling_quality = scaling_info.scaling_quality;
 
-	force_disable_dcc = adev->asic_type == CHIP_RAVEN && adev->in_suspend;
 	ret = fill_dc_plane_info_and_addr(adev, plane_state,
 					  afb->tiling_flags,
 					  &plane_info,
 					  &dc_plane_state->address,
-					  afb->tmz_surface,
-					  force_disable_dcc);
+					  afb->tmz_surface);
 	if (ret)
 		return ret;
 
@@ -5575,23 +6223,187 @@ static int fill_dc_plane_attributes(struct amdgpu_device *adev,
 	dc_plane_state->tiling_info = plane_info.tiling_info;
 	dc_plane_state->visible = plane_info.visible;
 	dc_plane_state->per_pixel_alpha = plane_info.per_pixel_alpha;
+	dc_plane_state->pre_multiplied_alpha = plane_info.pre_multiplied_alpha;
 	dc_plane_state->global_alpha = plane_info.global_alpha;
 	dc_plane_state->global_alpha_value = plane_info.global_alpha_value;
 	dc_plane_state->dcc = plane_info.dcc;
-	dc_plane_state->layer_index = plane_info.layer_index; // Always returns 0
+	dc_plane_state->layer_index = plane_info.layer_index;
 	dc_plane_state->flip_int_enabled = true;
 
 	/*
 	 * Always set input transfer function, since plane state is refreshed
 	 * every time.
 	 */
-	ret = amdgpu_dm_update_plane_color_mgmt(dm_crtc_state, dc_plane_state);
+	ret = amdgpu_dm_update_plane_color_mgmt(dm_crtc_state,
+						plane_state,
+						dc_plane_state);
 	if (ret)
 		return ret;
 
 	return 0;
 }
 
+static inline void fill_dc_dirty_rect(struct drm_plane *plane,
+				      struct rect *dirty_rect, int32_t x,
+				      s32 y, s32 width, s32 height,
+				      int *i, bool ffu)
+{
+	WARN_ON(*i >= DC_MAX_DIRTY_RECTS);
+
+	dirty_rect->x = x;
+	dirty_rect->y = y;
+	dirty_rect->width = width;
+	dirty_rect->height = height;
+
+	if (ffu)
+		drm_dbg(plane->dev,
+			"[PLANE:%d] PSR FFU dirty rect size (%d, %d)\n",
+			plane->base.id, width, height);
+	else
+		drm_dbg(plane->dev,
+			"[PLANE:%d] PSR SU dirty rect at (%d, %d) size (%d, %d)",
+			plane->base.id, x, y, width, height);
+
+	(*i)++;
+}
+
+/**
+ * fill_dc_dirty_rects() - Fill DC dirty regions for PSR selective updates
+ *
+ * @plane: DRM plane containing dirty regions that need to be flushed to the eDP
+ *         remote fb
+ * @old_plane_state: Old state of @plane
+ * @new_plane_state: New state of @plane
+ * @crtc_state: New state of CRTC connected to the @plane
+ * @flip_addrs: DC flip tracking struct, which also tracts dirty rects
+ * @is_psr_su: Flag indicating whether Panel Self Refresh Selective Update (PSR SU) is enabled.
+ *             If PSR SU is enabled and damage clips are available, only the regions of the screen
+ *             that have changed will be updated. If PSR SU is not enabled,
+ *             or if damage clips are not available, the entire screen will be updated.
+ * @dirty_regions_changed: dirty regions changed
+ *
+ * For PSR SU, DC informs the DMUB uController of dirty rectangle regions
+ * (referred to as "damage clips" in DRM nomenclature) that require updating on
+ * the eDP remote buffer. The responsibility of specifying the dirty regions is
+ * amdgpu_dm's.
+ *
+ * A damage-aware DRM client should fill the FB_DAMAGE_CLIPS property on the
+ * plane with regions that require flushing to the eDP remote buffer. In
+ * addition, certain use cases - such as cursor and multi-plane overlay (MPO) -
+ * implicitly provide damage clips without any client support via the plane
+ * bounds.
+ */
+static void fill_dc_dirty_rects(struct drm_plane *plane,
+				struct drm_plane_state *old_plane_state,
+				struct drm_plane_state *new_plane_state,
+				struct drm_crtc_state *crtc_state,
+				struct dc_flip_addrs *flip_addrs,
+				bool is_psr_su,
+				bool *dirty_regions_changed)
+{
+	struct dm_crtc_state *dm_crtc_state = to_dm_crtc_state(crtc_state);
+	struct rect *dirty_rects = flip_addrs->dirty_rects;
+	u32 num_clips;
+	struct drm_mode_rect *clips;
+	bool bb_changed;
+	bool fb_changed;
+	u32 i = 0;
+	*dirty_regions_changed = false;
+
+	/*
+	 * Cursor plane has it's own dirty rect update interface. See
+	 * dcn10_dmub_update_cursor_data and dmub_cmd_update_cursor_info_data
+	 */
+	if (plane->type == DRM_PLANE_TYPE_CURSOR)
+		return;
+
+	if (new_plane_state->rotation != DRM_MODE_ROTATE_0)
+		goto ffu;
+
+	num_clips = drm_plane_get_damage_clips_count(new_plane_state);
+	clips = drm_plane_get_damage_clips(new_plane_state);
+
+	if (num_clips && (!amdgpu_damage_clips || (amdgpu_damage_clips < 0 &&
+						   is_psr_su)))
+		goto ffu;
+
+	if (!dm_crtc_state->mpo_requested) {
+		if (!num_clips || num_clips > DC_MAX_DIRTY_RECTS)
+			goto ffu;
+
+		for (; flip_addrs->dirty_rect_count < num_clips; clips++)
+			fill_dc_dirty_rect(new_plane_state->plane,
+					   &dirty_rects[flip_addrs->dirty_rect_count],
+					   clips->x1, clips->y1,
+					   clips->x2 - clips->x1, clips->y2 - clips->y1,
+					   &flip_addrs->dirty_rect_count,
+					   false);
+		return;
+	}
+
+	/*
+	 * MPO is requested. Add entire plane bounding box to dirty rects if
+	 * flipped to or damaged.
+	 *
+	 * If plane is moved or resized, also add old bounding box to dirty
+	 * rects.
+	 */
+	fb_changed = old_plane_state->fb->base.id !=
+		     new_plane_state->fb->base.id;
+	bb_changed = (old_plane_state->crtc_x != new_plane_state->crtc_x ||
+		      old_plane_state->crtc_y != new_plane_state->crtc_y ||
+		      old_plane_state->crtc_w != new_plane_state->crtc_w ||
+		      old_plane_state->crtc_h != new_plane_state->crtc_h);
+
+	drm_dbg(plane->dev,
+		"[PLANE:%d] PSR bb_changed:%d fb_changed:%d num_clips:%d\n",
+		new_plane_state->plane->base.id,
+		bb_changed, fb_changed, num_clips);
+
+	*dirty_regions_changed = bb_changed;
+
+	if ((num_clips + (bb_changed ? 2 : 0)) > DC_MAX_DIRTY_RECTS)
+		goto ffu;
+
+	if (bb_changed) {
+		fill_dc_dirty_rect(new_plane_state->plane, &dirty_rects[i],
+				   new_plane_state->crtc_x,
+				   new_plane_state->crtc_y,
+				   new_plane_state->crtc_w,
+				   new_plane_state->crtc_h, &i, false);
+
+		/* Add old plane bounding-box if plane is moved or resized */
+		fill_dc_dirty_rect(new_plane_state->plane, &dirty_rects[i],
+				   old_plane_state->crtc_x,
+				   old_plane_state->crtc_y,
+				   old_plane_state->crtc_w,
+				   old_plane_state->crtc_h, &i, false);
+	}
+
+	if (num_clips) {
+		for (; i < num_clips; clips++)
+			fill_dc_dirty_rect(new_plane_state->plane,
+					   &dirty_rects[i], clips->x1,
+					   clips->y1, clips->x2 - clips->x1,
+					   clips->y2 - clips->y1, &i, false);
+	} else if (fb_changed && !bb_changed) {
+		fill_dc_dirty_rect(new_plane_state->plane, &dirty_rects[i],
+				   new_plane_state->crtc_x,
+				   new_plane_state->crtc_y,
+				   new_plane_state->crtc_w,
+				   new_plane_state->crtc_h, &i, false);
+	}
+
+	flip_addrs->dirty_rect_count = i;
+	return;
+
+ffu:
+	fill_dc_dirty_rect(new_plane_state->plane, &dirty_rects[0], 0, 0,
+			   dm_crtc_state->base.mode.crtc_hdisplay,
+			   dm_crtc_state->base.mode.crtc_vdisplay,
+			   &flip_addrs->dirty_rect_count, true);
+}
+
 static void update_stream_scaling_settings(const struct drm_display_mode *mode,
 					   const struct dm_connector_state *dm_state,
 					   struct dc_stream_state *stream)
@@ -5651,7 +6463,7 @@ static enum dc_color_depth
 convert_color_depth_from_display_info(const struct drm_connector *connector,
 				      bool is_y420, int requested_bpc)
 {
-	uint8_t bpc;
+	u8 bpc;
 
 	if (is_y420) {
 		bpc = 8;
@@ -5717,21 +6529,46 @@ get_aspect_ratio(const struct drm_display_mode *mode_in)
 }
 
 static enum dc_color_space
-get_output_color_space(const struct dc_crtc_timing *dc_crtc_timing)
+get_output_color_space(const struct dc_crtc_timing *dc_crtc_timing,
+		       const struct drm_connector_state *connector_state)
 {
 	enum dc_color_space color_space = COLOR_SPACE_SRGB;
 
-	switch (dc_crtc_timing->pixel_encoding)	{
-	case PIXEL_ENCODING_YCBCR422:
-	case PIXEL_ENCODING_YCBCR444:
-	case PIXEL_ENCODING_YCBCR420:
-	{
+	switch (connector_state->colorspace) {
+	case DRM_MODE_COLORIMETRY_BT601_YCC:
+		if (dc_crtc_timing->flags.Y_ONLY)
+			color_space = COLOR_SPACE_YCBCR601_LIMITED;
+		else
+			color_space = COLOR_SPACE_YCBCR601;
+		break;
+	case DRM_MODE_COLORIMETRY_BT709_YCC:
+		if (dc_crtc_timing->flags.Y_ONLY)
+			color_space = COLOR_SPACE_YCBCR709_LIMITED;
+		else
+			color_space = COLOR_SPACE_YCBCR709;
+		break;
+	case DRM_MODE_COLORIMETRY_OPRGB:
+		color_space = COLOR_SPACE_ADOBERGB;
+		break;
+	case DRM_MODE_COLORIMETRY_BT2020_RGB:
+	case DRM_MODE_COLORIMETRY_BT2020_YCC:
+		if (dc_crtc_timing->pixel_encoding == PIXEL_ENCODING_RGB)
+			color_space = COLOR_SPACE_2020_RGB_FULLRANGE;
+		else
+			color_space = COLOR_SPACE_2020_YCBCR_LIMITED;
+		break;
+	case DRM_MODE_COLORIMETRY_DEFAULT: // ITU601
+	default:
+		if (dc_crtc_timing->pixel_encoding == PIXEL_ENCODING_RGB) {
+			color_space = COLOR_SPACE_SRGB;
+			if (connector_state->hdmi.broadcast_rgb == DRM_HDMI_BROADCAST_RGB_LIMITED)
+				color_space = COLOR_SPACE_SRGB_LIMITED;
 		/*
 		 * 27030khz is the separation point between HDTV and SDTV
 		 * according to HDMI spec, we use YCbCr709 and YCbCr601
 		 * respectively
 		 */
-		if (dc_crtc_timing->pix_clk_100hz > 270300) {
+		} else if (dc_crtc_timing->pix_clk_100hz > 270300) {
 			if (dc_crtc_timing->flags.Y_ONLY)
 				color_space =
 					COLOR_SPACE_YCBCR709_LIMITED;
@@ -5744,27 +6581,37 @@ get_output_color_space(const struct dc_crtc_timing *dc_crtc_timing)
 			else
 				color_space = COLOR_SPACE_YCBCR601;
 		}
-
-	}
-	break;
-	case PIXEL_ENCODING_RGB:
-		color_space = COLOR_SPACE_SRGB;
-		break;
-
-	default:
-		WARN_ON(1);
 		break;
 	}
 
 	return color_space;
 }
 
+static enum display_content_type
+get_output_content_type(const struct drm_connector_state *connector_state)
+{
+	switch (connector_state->content_type) {
+	default:
+	case DRM_MODE_CONTENT_TYPE_NO_DATA:
+		return DISPLAY_CONTENT_TYPE_NO_DATA;
+	case DRM_MODE_CONTENT_TYPE_GRAPHICS:
+		return DISPLAY_CONTENT_TYPE_GRAPHICS;
+	case DRM_MODE_CONTENT_TYPE_PHOTO:
+		return DISPLAY_CONTENT_TYPE_PHOTO;
+	case DRM_MODE_CONTENT_TYPE_CINEMA:
+		return DISPLAY_CONTENT_TYPE_CINEMA;
+	case DRM_MODE_CONTENT_TYPE_GAME:
+		return DISPLAY_CONTENT_TYPE_GAME;
+	}
+}
+
 static bool adjust_colour_depth_from_display_info(
 	struct dc_crtc_timing *timing_out,
 	const struct drm_display_info *info)
 {
 	enum dc_color_depth depth = timing_out->display_color_depth;
 	int normalized_clk;
+
 	do {
 		normalized_clk = timing_out->pix_clk_100hz / 10;
 		/* YCbCr 4:2:0 requires additional adjustment of 1/2 */
@@ -5805,9 +6652,13 @@ static void fill_stream_properties_from_drm_display_mode(
 {
 	struct dc_crtc_timing *timing_out = &stream->timing;
 	const struct drm_display_info *info = &connector->display_info;
-	struct amdgpu_dm_connector *aconnector = to_amdgpu_dm_connector(connector);
+	struct amdgpu_dm_connector *aconnector = NULL;
 	struct hdmi_vendor_infoframe hv_frame;
 	struct hdmi_avi_infoframe avi_frame;
+	ssize_t err;
+
+	if (connector->connector_type != DRM_MODE_CONNECTOR_WRITEBACK)
+		aconnector = to_amdgpu_dm_connector(connector);
 
 	memset(&hv_frame, 0, sizeof(hv_frame));
 	memset(&avi_frame, 0, sizeof(avi_frame));
@@ -5821,9 +6672,14 @@ static void fill_stream_properties_from_drm_display_mode(
 			&& stream->signal == SIGNAL_TYPE_HDMI_TYPE_A)
 		timing_out->pixel_encoding = PIXEL_ENCODING_YCBCR420;
 	else if (drm_mode_is_420_also(info, mode_in)
+			&& aconnector
 			&& aconnector->force_yuv420_output)
 		timing_out->pixel_encoding = PIXEL_ENCODING_YCBCR420;
-	else if ((connector->display_info.color_formats & DRM_COLOR_FORMAT_YCRCB444)
+	else if ((connector->display_info.color_formats & DRM_COLOR_FORMAT_YCBCR422)
+			&& aconnector
+			&& aconnector->force_yuv422_output)
+		timing_out->pixel_encoding = PIXEL_ENCODING_YCBCR422;
+	else if ((connector->display_info.color_formats & DRM_COLOR_FORMAT_YCBCR444)
 			&& stream->signal == SIGNAL_TYPE_HDMI_TYPE_A)
 		timing_out->pixel_encoding = PIXEL_ENCODING_YCBCR444;
 	else
@@ -5837,7 +6693,7 @@ static void fill_stream_properties_from_drm_display_mode(
 	timing_out->scan_type = SCANNING_TYPE_NODATA;
 	timing_out->hdmi_vic = 0;
 
-	if(old_stream) {
+	if (old_stream) {
 		timing_out->vic = old_stream->timing.vic;
 		timing_out->flags.HSYNC_POSITIVE_POLARITY = old_stream->timing.flags.HSYNC_POSITIVE_POLARITY;
 		timing_out->flags.VSYNC_POSITIVE_POLARITY = old_stream->timing.flags.VSYNC_POSITIVE_POLARITY;
@@ -5850,13 +6706,23 @@ static void fill_stream_properties_from_drm_display_mode(
 	}
 
 	if (stream->signal == SIGNAL_TYPE_HDMI_TYPE_A) {
-		drm_hdmi_avi_infoframe_from_display_mode(&avi_frame, (struct drm_connector *)connector, mode_in);
+		err = drm_hdmi_avi_infoframe_from_display_mode(&avi_frame,
+							       (struct drm_connector *)connector,
+							       mode_in);
+		if (err < 0)
+			drm_warn_once(connector->dev, "Failed to setup avi infoframe on connector %s: %zd\n",
+				      connector->name, err);
 		timing_out->vic = avi_frame.video_code;
-		drm_hdmi_vendor_infoframe_from_display_mode(&hv_frame, (struct drm_connector *)connector, mode_in);
+		err = drm_hdmi_vendor_infoframe_from_display_mode(&hv_frame,
+								  (struct drm_connector *)connector,
+								  mode_in);
+		if (err < 0)
+			drm_warn_once(connector->dev, "Failed to setup vendor infoframe on connector %s: %zd\n",
+				      connector->name, err);
 		timing_out->hdmi_vic = hv_frame.vic;
 	}
 
-	if (is_freesync_video_mode(mode_in, aconnector)) {
+	if (aconnector && is_freesync_video_mode(mode_in, aconnector)) {
 		timing_out->h_addressable = mode_in->hdisplay;
 		timing_out->h_total = mode_in->htotal;
 		timing_out->h_sync_width = mode_in->hsync_end - mode_in->hsync_start;
@@ -5880,10 +6746,8 @@ static void fill_stream_properties_from_drm_display_mode(
 
 	timing_out->aspect_ratio = get_aspect_ratio(mode_in);
 
-	stream->output_color_space = get_output_color_space(timing_out);
-
-	stream->out_transfer_func->type = TF_TYPE_PREDEFINED;
-	stream->out_transfer_func->tf = TRANSFER_FUNCTION_SRGB;
+	stream->out_transfer_func.type = TF_TYPE_PREDEFINED;
+	stream->out_transfer_func.tf = TRANSFER_FUNCTION_SRGB;
 	if (stream->signal == SIGNAL_TYPE_HDMI_TYPE_A) {
 		if (!adjust_colour_depth_from_display_info(timing_out, info) &&
 		    drm_mode_is_420_also(info, mode_in) &&
@@ -5892,6 +6756,9 @@ static void fill_stream_properties_from_drm_display_mode(
 			adjust_colour_depth_from_display_info(timing_out, info);
 		}
 	}
+
+	stream->output_color_space = get_output_color_space(timing_out, connector_state);
+	stream->content_type = get_output_content_type(connector_state);
 }
 
 static void fill_audio_info(struct audio_info *audio_info,
@@ -5964,28 +6831,29 @@ decide_crtc_timing_for_drm_display_mode(struct drm_display_mode *drm_mode,
 					const struct drm_display_mode *native_mode,
 					bool scale_enabled)
 {
-	if (scale_enabled) {
-		copy_crtc_timing_for_drm_display_mode(native_mode, drm_mode);
-	} else if (native_mode->clock == drm_mode->clock &&
-			native_mode->htotal == drm_mode->htotal &&
-			native_mode->vtotal == drm_mode->vtotal) {
-		copy_crtc_timing_for_drm_display_mode(native_mode, drm_mode);
+	if (scale_enabled || (
+	    native_mode->clock == drm_mode->clock &&
+	    native_mode->htotal == drm_mode->htotal &&
+	    native_mode->vtotal == drm_mode->vtotal)) {
+		if (native_mode->crtc_clock)
+			copy_crtc_timing_for_drm_display_mode(native_mode, drm_mode);
 	} else {
 		/* no scaling nor amdgpu inserted, no need to patch */
 	}
 }
 
 static struct dc_sink *
-create_fake_sink(struct amdgpu_dm_connector *aconnector)
+create_fake_sink(struct drm_device *dev, struct dc_link *link)
 {
 	struct dc_sink_init_data sink_init_data = { 0 };
 	struct dc_sink *sink = NULL;
-	sink_init_data.link = aconnector->dc_link;
-	sink_init_data.sink_signal = aconnector->dc_link->connector_signal;
+
+	sink_init_data.link = link;
+	sink_init_data.sink_signal = link->connector_signal;
 
 	sink = dc_sink_create(&sink_init_data);
 	if (!sink) {
-		DRM_ERROR("Failed to create sink!\n");
+		drm_err(dev, "Failed to create sink!\n");
 		return NULL;
 	}
 	sink->sink_signal = SIGNAL_TYPE_VIRTUAL;
@@ -6059,69 +6927,6 @@ static void dm_enable_per_frame_crtc_master_sync(struct dc_state *context)
 	}
 }
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-static void update_dsc_caps(struct amdgpu_dm_connector *aconnector,
-							struct dc_sink *sink, struct dc_stream_state *stream,
-							struct dsc_dec_dpcd_caps *dsc_caps)
-{
-	stream->timing.flags.DSC = 0;
-
-	if (aconnector->dc_link && sink->sink_signal == SIGNAL_TYPE_DISPLAY_PORT) {
-		dc_dsc_parse_dsc_dpcd(aconnector->dc_link->ctx->dc,
-				      aconnector->dc_link->dpcd_caps.dsc_caps.dsc_basic_caps.raw,
-				      aconnector->dc_link->dpcd_caps.dsc_caps.dsc_branch_decoder_caps.raw,
-				      dsc_caps);
-	}
-}
-
-static void apply_dsc_policy_for_stream(struct amdgpu_dm_connector *aconnector,
-										struct dc_sink *sink, struct dc_stream_state *stream,
-										struct dsc_dec_dpcd_caps *dsc_caps)
-{
-	struct drm_connector *drm_connector = &aconnector->base;
-	uint32_t link_bandwidth_kbps;
-	uint32_t max_dsc_target_bpp_limit_override = 0;
-
-	link_bandwidth_kbps = dc_link_bandwidth_kbps(aconnector->dc_link,
-							dc_link_get_link_cap(aconnector->dc_link));
-
-	if (stream->link && stream->link->local_sink)
-		max_dsc_target_bpp_limit_override =
-			stream->link->local_sink->edid_caps.panel_patch.max_dsc_target_bpp_limit;
-	
-	/* Set DSC policy according to dsc_clock_en */
-	dc_dsc_policy_set_enable_dsc_when_not_needed(
-		aconnector->dsc_settings.dsc_force_enable == DSC_CLK_FORCE_ENABLE);
-
-	if (aconnector->dc_link && sink->sink_signal == SIGNAL_TYPE_DISPLAY_PORT) {
-
-		if (dc_dsc_compute_config(aconnector->dc_link->ctx->dc->res_pool->dscs[0],
-						dsc_caps,
-						aconnector->dc_link->ctx->dc->debug.dsc_min_slice_height_override,
-						max_dsc_target_bpp_limit_override,
-						link_bandwidth_kbps,
-						&stream->timing,
-						&stream->timing.dsc_cfg)) {
-			stream->timing.flags.DSC = 1;
-			DRM_DEBUG_DRIVER("%s: [%s] DSC is selected from SST RX\n", __func__, drm_connector->name);
-		}
-	}
-
-	/* Overwrite the stream flag if DSC is enabled through debugfs */
-	if (aconnector->dsc_settings.dsc_force_enable == DSC_CLK_FORCE_ENABLE)
-		stream->timing.flags.DSC = 1;
-
-	if (stream->timing.flags.DSC && aconnector->dsc_settings.dsc_num_slices_h)
-		stream->timing.dsc_cfg.num_slices_h = aconnector->dsc_settings.dsc_num_slices_h;
-
-	if (stream->timing.flags.DSC && aconnector->dsc_settings.dsc_num_slices_v)
-		stream->timing.dsc_cfg.num_slices_v = aconnector->dsc_settings.dsc_num_slices_v;
-
-	if (stream->timing.flags.DSC && aconnector->dsc_settings.dsc_bits_per_pixel)
-		stream->timing.dsc_cfg.bits_per_pixel = aconnector->dsc_settings.dsc_bits_per_pixel;
-}
-#endif /* CONFIG_DRM_AMD_DC_DCN */
-
 /**
  * DOC: FreeSync Video
  *
@@ -6138,7 +6943,7 @@ static void apply_dsc_policy_for_stream(struct amdgpu_dm_connector *aconnector,
  * - Cinema HFR (48 FPS)
  * - TV/PAL (50 FPS)
  * - Commonly used (60 FPS)
- * - Multiples of 24 (48,72,96,120 FPS)
+ * - Multiples of 24 (48,72,96 FPS)
  *
  * The list of standards video format is not huge and can be added to the
  * connector modeset list beforehand. With that, userspace can leverage
@@ -6154,19 +6959,22 @@ static void apply_dsc_policy_for_stream(struct amdgpu_dm_connector *aconnector,
  */
 static struct drm_display_mode *
 get_highest_refresh_rate_mode(struct amdgpu_dm_connector *aconnector,
-			  bool use_probed_modes)
+		bool use_probed_modes)
 {
 	struct drm_display_mode *m, *m_pref = NULL;
 	u16 current_refresh, highest_refresh;
 	struct list_head *list_head = use_probed_modes ?
-						    &aconnector->base.probed_modes :
-						    &aconnector->base.modes;
+		&aconnector->base.probed_modes :
+		&aconnector->base.modes;
+
+	if (aconnector->base.connector_type == DRM_MODE_CONNECTOR_WRITEBACK)
+		return NULL;
 
 	if (aconnector->freesync_vid_base.clock != 0)
 		return &aconnector->freesync_vid_base;
 
 	/* Find the preferred mode */
-	list_for_each_entry (m, list_head, head) {
+	list_for_each_entry(m, list_head, head) {
 		if (m->type & DRM_MODE_TYPE_PREFERRED) {
 			m_pref = m;
 			break;
@@ -6176,9 +6984,9 @@ get_highest_refresh_rate_mode(struct amdgpu_dm_connector *aconnector,
 	if (!m_pref) {
 		/* Probably an EDID with no preferred mode. Fallback to first entry */
 		m_pref = list_first_entry_or_null(
-			&aconnector->base.modes, struct drm_display_mode, head);
+				&aconnector->base.modes, struct drm_display_mode, head);
 		if (!m_pref) {
-			DRM_DEBUG_DRIVER("No preferred mode found in EDID\n");
+			drm_dbg_driver(aconnector->base.dev, "No preferred mode found in EDID\n");
 			return NULL;
 		}
 	}
@@ -6190,7 +6998,7 @@ get_highest_refresh_rate_mode(struct amdgpu_dm_connector *aconnector,
 	 * For some monitors, preferred mode is not the mode with highest
 	 * supported refresh rate.
 	 */
-	list_for_each_entry (m, list_head, head) {
+	list_for_each_entry(m, list_head, head) {
 		current_refresh  = drm_mode_vrefresh(m);
 
 		if (m->hdisplay == m_pref->hdisplay &&
@@ -6201,12 +7009,12 @@ get_highest_refresh_rate_mode(struct amdgpu_dm_connector *aconnector,
 		}
 	}
 
-	aconnector->freesync_vid_base = *m_pref;
+	drm_mode_copy(&aconnector->freesync_vid_base, m_pref);
 	return m_pref;
 }
 
 static bool is_freesync_video_mode(const struct drm_display_mode *mode,
-				   struct amdgpu_dm_connector *aconnector)
+		struct amdgpu_dm_connector *aconnector)
 {
 	struct drm_display_mode *high_mode;
 	int timing_diff;
@@ -6232,44 +7040,216 @@ static bool is_freesync_video_mode(const struct drm_display_mode *mode,
 		return true;
 }
 
+#if defined(CONFIG_DRM_AMD_DC_FP)
+static void update_dsc_caps(struct amdgpu_dm_connector *aconnector,
+			    struct dc_sink *sink, struct dc_stream_state *stream,
+			    struct dsc_dec_dpcd_caps *dsc_caps)
+{
+	stream->timing.flags.DSC = 0;
+	dsc_caps->is_dsc_supported = false;
+
+	if (aconnector->dc_link && (sink->sink_signal == SIGNAL_TYPE_DISPLAY_PORT ||
+	    sink->sink_signal == SIGNAL_TYPE_EDP)) {
+		if (sink->link->dpcd_caps.dongle_type == DISPLAY_DONGLE_NONE ||
+			sink->link->dpcd_caps.dongle_type == DISPLAY_DONGLE_DP_HDMI_CONVERTER)
+			dc_dsc_parse_dsc_dpcd(aconnector->dc_link->ctx->dc,
+				aconnector->dc_link->dpcd_caps.dsc_caps.dsc_basic_caps.raw,
+				aconnector->dc_link->dpcd_caps.dsc_caps.dsc_branch_decoder_caps.raw,
+				dsc_caps);
+	}
+}
+
+static void apply_dsc_policy_for_edp(struct amdgpu_dm_connector *aconnector,
+				    struct dc_sink *sink, struct dc_stream_state *stream,
+				    struct dsc_dec_dpcd_caps *dsc_caps,
+				    uint32_t max_dsc_target_bpp_limit_override)
+{
+	const struct dc_link_settings *verified_link_cap = NULL;
+	u32 link_bw_in_kbps;
+	u32 edp_min_bpp_x16, edp_max_bpp_x16;
+	struct dc *dc = sink->ctx->dc;
+	struct dc_dsc_bw_range bw_range = {0};
+	struct dc_dsc_config dsc_cfg = {0};
+	struct dc_dsc_config_options dsc_options = {0};
+
+	dc_dsc_get_default_config_option(dc, &dsc_options);
+	dsc_options.max_target_bpp_limit_override_x16 = max_dsc_target_bpp_limit_override * 16;
+
+	verified_link_cap = dc_link_get_link_cap(stream->link);
+	link_bw_in_kbps = dc_link_bandwidth_kbps(stream->link, verified_link_cap);
+	edp_min_bpp_x16 = 8 * 16;
+	edp_max_bpp_x16 = 8 * 16;
+
+	if (edp_max_bpp_x16 > dsc_caps->edp_max_bits_per_pixel)
+		edp_max_bpp_x16 = dsc_caps->edp_max_bits_per_pixel;
+
+	if (edp_max_bpp_x16 < edp_min_bpp_x16)
+		edp_min_bpp_x16 = edp_max_bpp_x16;
+
+	if (dc_dsc_compute_bandwidth_range(dc->res_pool->dscs[0],
+				dc->debug.dsc_min_slice_height_override,
+				edp_min_bpp_x16, edp_max_bpp_x16,
+				dsc_caps,
+				&stream->timing,
+				dc_link_get_highest_encoding_format(aconnector->dc_link),
+				&bw_range)) {
+
+		if (bw_range.max_kbps < link_bw_in_kbps) {
+			if (dc_dsc_compute_config(dc->res_pool->dscs[0],
+					dsc_caps,
+					&dsc_options,
+					0,
+					&stream->timing,
+					dc_link_get_highest_encoding_format(aconnector->dc_link),
+					&dsc_cfg)) {
+				stream->timing.dsc_cfg = dsc_cfg;
+				stream->timing.flags.DSC = 1;
+				stream->timing.dsc_cfg.bits_per_pixel = edp_max_bpp_x16;
+			}
+			return;
+		}
+	}
+
+	if (dc_dsc_compute_config(dc->res_pool->dscs[0],
+				dsc_caps,
+				&dsc_options,
+				link_bw_in_kbps,
+				&stream->timing,
+				dc_link_get_highest_encoding_format(aconnector->dc_link),
+				&dsc_cfg)) {
+		stream->timing.dsc_cfg = dsc_cfg;
+		stream->timing.flags.DSC = 1;
+	}
+}
+
+static void apply_dsc_policy_for_stream(struct amdgpu_dm_connector *aconnector,
+					struct dc_sink *sink, struct dc_stream_state *stream,
+					struct dsc_dec_dpcd_caps *dsc_caps)
+{
+	struct drm_connector *drm_connector = &aconnector->base;
+	u32 link_bandwidth_kbps;
+	struct dc *dc = sink->ctx->dc;
+	u32 max_supported_bw_in_kbps, timing_bw_in_kbps;
+	u32 dsc_max_supported_bw_in_kbps;
+	u32 max_dsc_target_bpp_limit_override =
+		drm_connector->display_info.max_dsc_bpp;
+	struct dc_dsc_config_options dsc_options = {0};
+
+	dc_dsc_get_default_config_option(dc, &dsc_options);
+	dsc_options.max_target_bpp_limit_override_x16 = max_dsc_target_bpp_limit_override * 16;
+
+	link_bandwidth_kbps = dc_link_bandwidth_kbps(aconnector->dc_link,
+							dc_link_get_link_cap(aconnector->dc_link));
+
+	/* Set DSC policy according to dsc_clock_en */
+	dc_dsc_policy_set_enable_dsc_when_not_needed(
+		aconnector->dsc_settings.dsc_force_enable == DSC_CLK_FORCE_ENABLE);
+
+	if (sink->sink_signal == SIGNAL_TYPE_EDP &&
+	    !aconnector->dc_link->panel_config.dsc.disable_dsc_edp &&
+	    dc->caps.edp_dsc_support && aconnector->dsc_settings.dsc_force_enable != DSC_CLK_FORCE_DISABLE) {
+
+		apply_dsc_policy_for_edp(aconnector, sink, stream, dsc_caps, max_dsc_target_bpp_limit_override);
+
+	} else if (sink->sink_signal == SIGNAL_TYPE_DISPLAY_PORT) {
+		if (sink->link->dpcd_caps.dongle_type == DISPLAY_DONGLE_NONE) {
+			if (dc_dsc_compute_config(aconnector->dc_link->ctx->dc->res_pool->dscs[0],
+						dsc_caps,
+						&dsc_options,
+						link_bandwidth_kbps,
+						&stream->timing,
+						dc_link_get_highest_encoding_format(aconnector->dc_link),
+						&stream->timing.dsc_cfg)) {
+				stream->timing.flags.DSC = 1;
+				drm_dbg_driver(drm_connector->dev, "%s: SST_DSC [%s] DSC is selected from SST RX\n",
+							__func__, drm_connector->name);
+			}
+		} else if (sink->link->dpcd_caps.dongle_type == DISPLAY_DONGLE_DP_HDMI_CONVERTER) {
+			timing_bw_in_kbps = dc_bandwidth_in_kbps_from_timing(&stream->timing,
+					dc_link_get_highest_encoding_format(aconnector->dc_link));
+			max_supported_bw_in_kbps = link_bandwidth_kbps;
+			dsc_max_supported_bw_in_kbps = link_bandwidth_kbps;
+
+			if (timing_bw_in_kbps > max_supported_bw_in_kbps &&
+					max_supported_bw_in_kbps > 0 &&
+					dsc_max_supported_bw_in_kbps > 0)
+				if (dc_dsc_compute_config(aconnector->dc_link->ctx->dc->res_pool->dscs[0],
+						dsc_caps,
+						&dsc_options,
+						dsc_max_supported_bw_in_kbps,
+						&stream->timing,
+						dc_link_get_highest_encoding_format(aconnector->dc_link),
+						&stream->timing.dsc_cfg)) {
+					stream->timing.flags.DSC = 1;
+					drm_dbg_driver(drm_connector->dev, "%s: SST_DSC [%s] DSC is selected from DP-HDMI PCON\n",
+									 __func__, drm_connector->name);
+				}
+		}
+	}
+
+	/* Overwrite the stream flag if DSC is enabled through debugfs */
+	if (aconnector->dsc_settings.dsc_force_enable == DSC_CLK_FORCE_ENABLE)
+		stream->timing.flags.DSC = 1;
+
+	if (stream->timing.flags.DSC && aconnector->dsc_settings.dsc_num_slices_h)
+		stream->timing.dsc_cfg.num_slices_h = aconnector->dsc_settings.dsc_num_slices_h;
+
+	if (stream->timing.flags.DSC && aconnector->dsc_settings.dsc_num_slices_v)
+		stream->timing.dsc_cfg.num_slices_v = aconnector->dsc_settings.dsc_num_slices_v;
+
+	if (stream->timing.flags.DSC && aconnector->dsc_settings.dsc_bits_per_pixel)
+		stream->timing.dsc_cfg.bits_per_pixel = aconnector->dsc_settings.dsc_bits_per_pixel;
+}
+#endif
+
 static struct dc_stream_state *
-create_stream_for_sink(struct amdgpu_dm_connector *aconnector,
+create_stream_for_sink(struct drm_connector *connector,
 		       const struct drm_display_mode *drm_mode,
 		       const struct dm_connector_state *dm_state,
 		       const struct dc_stream_state *old_stream,
 		       int requested_bpc)
 {
+	struct drm_device *dev = connector->dev;
+	struct amdgpu_dm_connector *aconnector = NULL;
 	struct drm_display_mode *preferred_mode = NULL;
-	struct drm_connector *drm_connector;
-	const struct drm_connector_state *con_state =
-		dm_state ? &dm_state->base : NULL;
+	const struct drm_connector_state *con_state = &dm_state->base;
 	struct dc_stream_state *stream = NULL;
-	struct drm_display_mode mode = *drm_mode;
+	struct drm_display_mode mode;
 	struct drm_display_mode saved_mode;
 	struct drm_display_mode *freesync_mode = NULL;
 	bool native_mode_found = false;
 	bool recalculate_timing = false;
-	bool scale = dm_state ? (dm_state->scaling != RMX_OFF) : false;
+	bool scale = dm_state->scaling != RMX_OFF;
 	int mode_refresh;
 	int preferred_refresh = 0;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
+	enum color_transfer_func tf = TRANSFER_FUNC_UNKNOWN;
+#if defined(CONFIG_DRM_AMD_DC_FP)
 	struct dsc_dec_dpcd_caps dsc_caps;
 #endif
+	struct dc_link *link = NULL;
 	struct dc_sink *sink = NULL;
 
+	drm_mode_init(&mode, drm_mode);
 	memset(&saved_mode, 0, sizeof(saved_mode));
 
-	if (aconnector == NULL) {
-		DRM_ERROR("aconnector is NULL!\n");
-		return stream;
-	}
+	if (connector->connector_type != DRM_MODE_CONNECTOR_WRITEBACK) {
+		aconnector = NULL;
+		aconnector = to_amdgpu_dm_connector(connector);
+		link = aconnector->dc_link;
+	} else {
+		struct drm_writeback_connector *wbcon = NULL;
+		struct amdgpu_dm_wb_connector *dm_wbcon = NULL;
 
-	drm_connector = &aconnector->base;
+		wbcon = drm_connector_to_writeback(connector);
+		dm_wbcon = to_amdgpu_dm_wb_connector(wbcon);
+		link = dm_wbcon->link;
+	}
 
-	if (!aconnector->dc_sink) {
-		sink = create_fake_sink(aconnector);
+	if (!aconnector || !aconnector->dc_sink) {
+		sink = create_fake_sink(dev, link);
 		if (!sink)
 			return stream;
+
 	} else {
 		sink = aconnector->dc_sink;
 		dc_sink_retain(sink);
@@ -6278,16 +7258,17 @@ create_stream_for_sink(struct amdgpu_dm_connector *aconnector,
 	stream = dc_create_stream_for_sink(sink);
 
 	if (stream == NULL) {
-		DRM_ERROR("Failed to create stream for sink!\n");
+		drm_err(dev, "Failed to create stream for sink!\n");
 		goto finish;
 	}
 
+	/* We leave this NULL for writeback connectors */
 	stream->dm_stream_context = aconnector;
 
 	stream->timing.flags.LTE_340MCSC_SCRAMBLE =
-		drm_connector->display_info.hdmi.scdc.scrambling.low_rates;
+		connector->display_info.hdmi.scdc.scrambling.low_rates;
 
-	list_for_each_entry(preferred_mode, &aconnector->base.modes, head) {
+	list_for_each_entry(preferred_mode, &connector->modes, head) {
 		/* Search for preferred mode */
 		if (preferred_mode->type & DRM_MODE_TYPE_PREFERRED) {
 			native_mode_found = true;
@@ -6296,7 +7277,7 @@ create_stream_for_sink(struct amdgpu_dm_connector *aconnector,
 	}
 	if (!native_mode_found)
 		preferred_mode = list_first_entry_or_null(
-				&aconnector->base.modes,
+				&connector->modes,
 				struct drm_display_mode,
 				head);
 
@@ -6307,19 +7288,21 @@ create_stream_for_sink(struct amdgpu_dm_connector *aconnector,
 		 * This may not be an error, the use case is when we have no
 		 * usermode calls to reset and set mode upon hotplug. In this
 		 * case, we call set mode ourselves to restore the previous mode
-		 * and the modelist may not be filled in in time.
+		 * and the modelist may not be filled in time.
 		 */
-		DRM_DEBUG_DRIVER("No preferred mode found\n");
-	} else {
+		drm_dbg_driver(dev, "No preferred mode found\n");
+	} else if (aconnector) {
 		recalculate_timing = amdgpu_freesync_vid_mode &&
 				 is_freesync_video_mode(&mode, aconnector);
 		if (recalculate_timing) {
 			freesync_mode = get_highest_refresh_rate_mode(aconnector, false);
-			saved_mode = mode;
-			mode = *freesync_mode;
+			drm_mode_copy(&saved_mode, &mode);
+			saved_mode.picture_aspect_ratio = mode.picture_aspect_ratio;
+			drm_mode_copy(&mode, freesync_mode);
+			mode.picture_aspect_ratio = saved_mode.picture_aspect_ratio;
 		} else {
 			decide_crtc_timing_for_drm_display_mode(
-				&mode, preferred_mode, scale);
+					&mode, preferred_mode, scale);
 
 			preferred_refresh = drm_mode_vrefresh(preferred_mode);
 		}
@@ -6327,23 +7310,33 @@ create_stream_for_sink(struct amdgpu_dm_connector *aconnector,
 
 	if (recalculate_timing)
 		drm_mode_set_crtcinfo(&saved_mode, 0);
-	else if (!dm_state)
-		drm_mode_set_crtcinfo(&mode, 0);
 
-       /*
-	* If scaling is enabled and refresh rate didn't change
-	* we copy the vic and polarities of the old timings
-	*/
+	/*
+	 * If scaling is enabled and refresh rate didn't change
+	 * we copy the vic and polarities of the old timings
+	 */
 	if (!scale || mode_refresh != preferred_refresh)
 		fill_stream_properties_from_drm_display_mode(
-			stream, &mode, &aconnector->base, con_state, NULL,
+			stream, &mode, connector, con_state, NULL,
 			requested_bpc);
 	else
 		fill_stream_properties_from_drm_display_mode(
-			stream, &mode, &aconnector->base, con_state, old_stream,
+			stream, &mode, connector, con_state, old_stream,
 			requested_bpc);
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
+	/* The rest isn't needed for writeback connectors */
+	if (!aconnector)
+		goto finish;
+
+	if (aconnector->timing_changed) {
+		drm_dbg(aconnector->base.dev,
+			"overriding timing for automated test, bpc %d, changing to %d\n",
+			stream->timing.display_color_depth,
+			aconnector->timing_requested->display_color_depth);
+		stream->timing = *aconnector->timing_requested;
+	}
+
+#if defined(CONFIG_DRM_AMD_DC_FP)
 	/* SST DSC determination policy */
 	update_dsc_caps(aconnector, sink, stream, &dsc_caps);
 	if (aconnector->dsc_settings.dsc_force_enable != DSC_CLK_FORCE_DISABLE && dsc_caps.is_dsc_supported)
@@ -6354,7 +7347,7 @@ create_stream_for_sink(struct amdgpu_dm_connector *aconnector,
 
 	fill_audio_info(
 		&stream->audio_info,
-		drm_connector,
+		connector,
 		sink);
 
 	update_stream_signal(stream, sink);
@@ -6362,21 +7355,29 @@ create_stream_for_sink(struct amdgpu_dm_connector *aconnector,
 	if (stream->signal == SIGNAL_TYPE_HDMI_TYPE_A)
 		mod_build_hf_vsif_infopacket(stream, &stream->vsp_infopacket);
 
-	if (stream->link->psr_settings.psr_feature_enabled) {
+	if (stream->signal == SIGNAL_TYPE_DISPLAY_PORT ||
+	    stream->signal == SIGNAL_TYPE_DISPLAY_PORT_MST ||
+	    stream->signal == SIGNAL_TYPE_EDP) {
+		const struct dc_edid_caps *edid_caps;
+		unsigned int disable_colorimetry = 0;
+
+		if (aconnector->dc_sink) {
+			edid_caps = &aconnector->dc_sink->edid_caps;
+			disable_colorimetry = edid_caps->panel_patch.disable_colorimetry;
+		}
+
 		//
 		// should decide stream support vsc sdp colorimetry capability
 		// before building vsc info packet
 		//
-		stream->use_vsc_sdp_for_colorimetry = false;
-		if (aconnector->dc_sink->sink_signal == SIGNAL_TYPE_DISPLAY_PORT_MST) {
-			stream->use_vsc_sdp_for_colorimetry =
-				aconnector->dc_sink->is_vsc_sdp_colorimetry_supported;
-		} else {
-			if (stream->link->dpcd_caps.dprx_feature.bits.VSC_SDP_COLORIMETRY_SUPPORTED)
-				stream->use_vsc_sdp_for_colorimetry = true;
-		}
-		mod_build_vsc_infopacket(stream, &stream->vsc_infopacket);
-		aconnector->psr_skip_count = AMDGPU_DM_PSR_ENTRY_DELAY;
+		stream->use_vsc_sdp_for_colorimetry = stream->link->dpcd_caps.dpcd_rev.raw >= 0x14 &&
+						      stream->link->dpcd_caps.dprx_feature.bits.VSC_SDP_COLORIMETRY_SUPPORTED &&
+						      !disable_colorimetry;
+
+		if (stream->out_transfer_func.tf == TRANSFER_FUNCTION_GAMMA22)
+			tf = TRANSFER_FUNC_GAMMA_22;
+		mod_build_vsc_infopacket(stream, &stream->vsc_infopacket, stream->output_color_space, tf);
+		aconnector->sr_skip_count = AMDGPU_DM_PSR_ENTRY_DELAY;
 
 	}
 finish:
@@ -6385,208 +7386,117 @@ finish:
 	return stream;
 }
 
-static void amdgpu_dm_crtc_destroy(struct drm_crtc *crtc)
-{
-	drm_crtc_cleanup(crtc);
-	kfree(crtc);
-}
-
-static void dm_crtc_destroy_state(struct drm_crtc *crtc,
-				  struct drm_crtc_state *state)
-{
-	struct dm_crtc_state *cur = to_dm_crtc_state(state);
-
-	/* TODO Destroy dc_stream objects are stream object is flattened */
-	if (cur->stream)
-		dc_stream_release(cur->stream);
-
-
-	__drm_atomic_helper_crtc_destroy_state(state);
-
-
-	kfree(state);
-}
-
-static void dm_crtc_reset_state(struct drm_crtc *crtc)
-{
-	struct dm_crtc_state *state;
-
-	if (crtc->state)
-		dm_crtc_destroy_state(crtc, crtc->state);
-
-	state = kzalloc(sizeof(*state), GFP_KERNEL);
-	if (WARN_ON(!state))
-		return;
-
-	__drm_atomic_helper_crtc_reset(crtc, &state->base);
-}
-
-static struct drm_crtc_state *
-dm_crtc_duplicate_state(struct drm_crtc *crtc)
-{
-	struct dm_crtc_state *state, *cur;
-
-	cur = to_dm_crtc_state(crtc->state);
-
-	if (WARN_ON(!crtc->state))
-		return NULL;
-
-	state = kzalloc(sizeof(*state), GFP_KERNEL);
-	if (!state)
-		return NULL;
-
-	__drm_atomic_helper_crtc_duplicate_state(crtc, &state->base);
-
-	if (cur->stream) {
-		state->stream = cur->stream;
-		dc_stream_retain(state->stream);
-	}
-
-	state->active_planes = cur->active_planes;
-	state->vrr_infopacket = cur->vrr_infopacket;
-	state->abm_level = cur->abm_level;
-	state->vrr_supported = cur->vrr_supported;
-	state->freesync_config = cur->freesync_config;
-	state->cm_has_degamma = cur->cm_has_degamma;
-	state->cm_is_degamma_srgb = cur->cm_is_degamma_srgb;
-	state->force_dpms_off = cur->force_dpms_off;
-	/* TODO Duplicate dc_stream after objects are stream object is flattened */
-
-	return &state->base;
-}
-
-#ifdef CONFIG_DRM_AMD_SECURE_DISPLAY
-static int amdgpu_dm_crtc_late_register(struct drm_crtc *crtc)
-{
-	crtc_debugfs_init(crtc);
-
-	return 0;
-}
-#endif
-
-static inline int dm_set_vupdate_irq(struct drm_crtc *crtc, bool enable)
+/**
+ * amdgpu_dm_connector_poll - Poll a connector to see if it's connected to a display
+ * @aconnector: DM connector to poll (owns @base drm_connector and @dc_link)
+ * @force: if true, force polling even when DAC load detection was used
+ *
+ * Used for connectors that don't support HPD (hotplug detection) to
+ * periodically check whether the connector is connected to a display.
+ *
+ * When connection was determined via DAC load detection, we avoid
+ * re-running it on normal polls to prevent visible glitches, unless
+ * @force is set.
+ *
+ * Return: The probed connector status (connected/disconnected/unknown).
+ */
+static enum drm_connector_status
+amdgpu_dm_connector_poll(struct amdgpu_dm_connector *aconnector, bool force)
 {
-	enum dc_irq_source irq_source;
-	struct amdgpu_crtc *acrtc = to_amdgpu_crtc(crtc);
-	struct amdgpu_device *adev = drm_to_adev(crtc->dev);
-	int rc;
+	struct drm_connector *connector = &aconnector->base;
+	struct drm_device *dev = connector->dev;
+	struct amdgpu_device *adev = drm_to_adev(dev);
+	struct dc_link *link = aconnector->dc_link;
+	enum dc_connection_type conn_type = dc_connection_none;
+	enum drm_connector_status status = connector_status_disconnected;
 
-	irq_source = IRQ_TYPE_VUPDATE + acrtc->otg_inst;
+	/* When we determined the connection using DAC load detection,
+	 * do NOT poll the connector do detect disconnect because
+	 * that would run DAC load detection again which can cause
+	 * visible visual glitches.
+	 *
+	 * Only allow to poll such a connector again when forcing.
+	 */
+	if (!force && link->local_sink && link->type == dc_connection_dac_load)
+		return connector->status;
 
-	rc = dc_interrupt_set(adev->dm.dc, irq_source, enable) ? 0 : -EBUSY;
+	mutex_lock(&aconnector->hpd_lock);
 
-	DRM_DEBUG_VBL("crtc %d - vupdate irq %sabling: r=%d\n",
-		      acrtc->crtc_id, enable ? "en" : "dis", rc);
-	return rc;
-}
+	if (dc_link_detect_connection_type(aconnector->dc_link, &conn_type) &&
+	    conn_type != dc_connection_none) {
+		mutex_lock(&adev->dm.dc_lock);
 
-static inline int dm_set_vblank(struct drm_crtc *crtc, bool enable)
-{
-	enum dc_irq_source irq_source;
-	struct amdgpu_crtc *acrtc = to_amdgpu_crtc(crtc);
-	struct amdgpu_device *adev = drm_to_adev(crtc->dev);
-	struct dm_crtc_state *acrtc_state = to_dm_crtc_state(crtc->state);
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	struct amdgpu_display_manager *dm = &adev->dm;
-	struct vblank_control_work *work;
-#endif
-	int rc = 0;
+		/* Only call full link detection when a sink isn't created yet,
+		 * ie. just when the display is plugged in, otherwise we risk flickering.
+		 */
+		if (link->local_sink ||
+			dc_link_detect(link, DETECT_REASON_HPD))
+			status = connector_status_connected;
 
-	if (enable) {
-		/* vblank irq on -> Only need vupdate irq in vrr mode */
-		if (amdgpu_dm_vrr_active(acrtc_state))
-			rc = dm_set_vupdate_irq(crtc, true);
-	} else {
-		/* vblank irq off -> vupdate irq off */
-		rc = dm_set_vupdate_irq(crtc, false);
+		mutex_unlock(&adev->dm.dc_lock);
 	}
 
-	if (rc)
-		return rc;
-
-	irq_source = IRQ_TYPE_VBLANK + acrtc->otg_inst;
-
-	if (!dc_interrupt_set(adev->dm.dc, irq_source, enable))
-		return -EBUSY;
-
-	if (amdgpu_in_reset(adev))
-		return 0;
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	if (dm->vblank_control_workqueue) {
-		work = kzalloc(sizeof(*work), GFP_ATOMIC);
-		if (!work)
-			return -ENOMEM;
-
-		INIT_WORK(&work->work, vblank_control_worker);
-		work->dm = dm;
-		work->acrtc = acrtc;
-		work->enable = enable;
+	if (connector->status != status) {
+		if (status == connector_status_disconnected) {
+			if (link->local_sink)
+				dc_sink_release(link->local_sink);
 
-		if (acrtc_state->stream) {
-			dc_stream_retain(acrtc_state->stream);
-			work->stream = acrtc_state->stream;
+			link->local_sink = NULL;
+			link->dpcd_sink_count = 0;
+			link->type = dc_connection_none;
 		}
 
-		queue_work(dm->vblank_control_workqueue, &work->work);
+		amdgpu_dm_update_connector_after_detect(aconnector);
 	}
-#endif
-
-	return 0;
-}
 
-static int dm_enable_vblank(struct drm_crtc *crtc)
-{
-	return dm_set_vblank(crtc, true);
-}
-
-static void dm_disable_vblank(struct drm_crtc *crtc)
-{
-	dm_set_vblank(crtc, false);
+	mutex_unlock(&aconnector->hpd_lock);
+	return status;
 }
 
-/* Implemented only the options currently availible for the driver */
-static const struct drm_crtc_funcs amdgpu_dm_crtc_funcs = {
-	.reset = dm_crtc_reset_state,
-	.destroy = amdgpu_dm_crtc_destroy,
-	.set_config = drm_atomic_helper_set_config,
-	.page_flip = drm_atomic_helper_page_flip,
-	.atomic_duplicate_state = dm_crtc_duplicate_state,
-	.atomic_destroy_state = dm_crtc_destroy_state,
-	.set_crc_source = amdgpu_dm_crtc_set_crc_source,
-	.verify_crc_source = amdgpu_dm_crtc_verify_crc_source,
-	.get_crc_sources = amdgpu_dm_crtc_get_crc_sources,
-	.get_vblank_counter = amdgpu_get_vblank_counter_kms,
-	.enable_vblank = dm_enable_vblank,
-	.disable_vblank = dm_disable_vblank,
-	.get_vblank_timestamp = drm_crtc_vblank_helper_get_vblank_timestamp,
-#if defined(CONFIG_DRM_AMD_SECURE_DISPLAY)
-	.late_register = amdgpu_dm_crtc_late_register,
-#endif
-};
-
+/**
+ * amdgpu_dm_connector_detect() - Detect whether a DRM connector is connected to a display
+ *
+ * A connector is considered connected when it has a sink that is not NULL.
+ * For connectors that support HPD (hotplug detection), the connection is
+ * handled in the HPD interrupt.
+ * For connectors that may not support HPD, such as analog connectors,
+ * DRM will call this function repeatedly to poll them.
+ *
+ * Notes:
+ * 1. This interface is NOT called in context of HPD irq.
+ * 2. This interface *is called* in context of user-mode ioctl. Which
+ *    makes it a bad place for *any* MST-related activity.
+ *
+ * @connector: The DRM connector we are checking. We convert it to
+ *             amdgpu_dm_connector so we can read the DC link and state.
+ * @force:     If true, do a full detect again. This is used even when
+ *             a lighter check would normally be used to avoid flicker.
+ *
+ * Return: The connector status (connected, disconnected, or unknown).
+ *
+ */
 static enum drm_connector_status
 amdgpu_dm_connector_detect(struct drm_connector *connector, bool force)
 {
-	bool connected;
 	struct amdgpu_dm_connector *aconnector = to_amdgpu_dm_connector(connector);
 
-	/*
-	 * Notes:
-	 * 1. This interface is NOT called in context of HPD irq.
-	 * 2. This interface *is called* in context of user-mode ioctl. Which
-	 * makes it a bad place for *any* MST-related activity.
-	 */
+	update_subconnector_property(aconnector);
 
-	if (aconnector->base.force == DRM_FORCE_UNSPECIFIED &&
-	    !aconnector->fake_enable)
-		connected = (aconnector->dc_sink != NULL);
-	else
-		connected = (aconnector->base.force == DRM_FORCE_ON);
+	if (aconnector->base.force == DRM_FORCE_ON ||
+		aconnector->base.force == DRM_FORCE_ON_DIGITAL)
+		return connector_status_connected;
+	else if (aconnector->base.force == DRM_FORCE_OFF)
+		return connector_status_disconnected;
 
-	update_subconnector_property(aconnector);
+	/* Poll analog connectors and only when either
+	 * disconnected or connected to an analog display.
+	 */
+	if (drm_kms_helper_is_poll_worker() &&
+		dc_connector_supports_analog(aconnector->dc_link->link_id.id) &&
+		(!aconnector->dc_sink || aconnector->dc_sink->edid_caps.analog))
+		return amdgpu_dm_connector_poll(aconnector, force);
 
-	return (connected ? connector_status_connected :
+	return (aconnector->dc_sink ? connector_status_connected :
 			connector_status_disconnected);
 }
 
@@ -6638,7 +7548,18 @@ int amdgpu_dm_connector_atomic_set_property(struct drm_connector *connector,
 		dm_new_state->underscan_enable = val;
 		ret = 0;
 	} else if (property == adev->mode_info.abm_level_property) {
-		dm_new_state->abm_level = val;
+		switch (val) {
+		case ABM_SYSFS_CONTROL:
+			dm_new_state->abm_sysfs_forbidden = false;
+			break;
+		case ABM_LEVEL_OFF:
+			dm_new_state->abm_sysfs_forbidden = true;
+			dm_new_state->abm_level = ABM_LEVEL_IMMEDIATE_DISABLE;
+			break;
+		default:
+			dm_new_state->abm_sysfs_forbidden = true;
+			dm_new_state->abm_level = val;
+		}
 		ret = 0;
 	}
 
@@ -6683,44 +7604,149 @@ int amdgpu_dm_connector_atomic_get_property(struct drm_connector *connector,
 		*val = dm_state->underscan_enable;
 		ret = 0;
 	} else if (property == adev->mode_info.abm_level_property) {
-		*val = dm_state->abm_level;
+		if (!dm_state->abm_sysfs_forbidden)
+			*val = ABM_SYSFS_CONTROL;
+		else
+			*val = (dm_state->abm_level != ABM_LEVEL_IMMEDIATE_DISABLE) ?
+				dm_state->abm_level : 0;
 		ret = 0;
 	}
 
 	return ret;
 }
 
+/**
+ * DOC: panel power savings
+ *
+ * The display manager allows you to set your desired **panel power savings**
+ * level (between 0-4, with 0 representing off), e.g. using the following::
+ *
+ *   # echo 3 > /sys/class/drm/card0-eDP-1/amdgpu/panel_power_savings
+ *
+ * Modifying this value can have implications on color accuracy, so tread
+ * carefully.
+ */
+
+static ssize_t panel_power_savings_show(struct device *device,
+					struct device_attribute *attr,
+					char *buf)
+{
+	struct drm_connector *connector = dev_get_drvdata(device);
+	struct drm_device *dev = connector->dev;
+	u8 val;
+
+	drm_modeset_lock(&dev->mode_config.connection_mutex, NULL);
+	val = to_dm_connector_state(connector->state)->abm_level ==
+		ABM_LEVEL_IMMEDIATE_DISABLE ? 0 :
+		to_dm_connector_state(connector->state)->abm_level;
+	drm_modeset_unlock(&dev->mode_config.connection_mutex);
+
+	return sysfs_emit(buf, "%u\n", val);
+}
+
+static ssize_t panel_power_savings_store(struct device *device,
+					 struct device_attribute *attr,
+					 const char *buf, size_t count)
+{
+	struct drm_connector *connector = dev_get_drvdata(device);
+	struct drm_device *dev = connector->dev;
+	long val;
+	int ret;
+
+	ret = kstrtol(buf, 0, &val);
+
+	if (ret)
+		return ret;
+
+	if (val < 0 || val > 4)
+		return -EINVAL;
+
+	drm_modeset_lock(&dev->mode_config.connection_mutex, NULL);
+	if (to_dm_connector_state(connector->state)->abm_sysfs_forbidden)
+		ret = -EBUSY;
+	else
+		to_dm_connector_state(connector->state)->abm_level = val ?:
+			ABM_LEVEL_IMMEDIATE_DISABLE;
+	drm_modeset_unlock(&dev->mode_config.connection_mutex);
+
+	if (ret)
+		return ret;
+
+	drm_kms_helper_hotplug_event(dev);
+
+	return count;
+}
+
+static DEVICE_ATTR_RW(panel_power_savings);
+
+static struct attribute *amdgpu_attrs[] = {
+	&dev_attr_panel_power_savings.attr,
+	NULL
+};
+
+static const struct attribute_group amdgpu_group = {
+	.name = "amdgpu",
+	.attrs = amdgpu_attrs
+};
+
+static bool
+amdgpu_dm_should_create_sysfs(struct amdgpu_dm_connector *amdgpu_dm_connector)
+{
+	if (amdgpu_dm_abm_level >= 0)
+		return false;
+
+	if (amdgpu_dm_connector->base.connector_type != DRM_MODE_CONNECTOR_eDP)
+		return false;
+
+	/* check for OLED panels */
+	if (amdgpu_dm_connector->bl_idx >= 0) {
+		struct drm_device *drm = amdgpu_dm_connector->base.dev;
+		struct amdgpu_display_manager *dm = &drm_to_adev(drm)->dm;
+		struct amdgpu_dm_backlight_caps *caps;
+
+		caps = &dm->backlight_caps[amdgpu_dm_connector->bl_idx];
+		if (caps->aux_support)
+			return false;
+	}
+
+	return true;
+}
+
 static void amdgpu_dm_connector_unregister(struct drm_connector *connector)
 {
 	struct amdgpu_dm_connector *amdgpu_dm_connector = to_amdgpu_dm_connector(connector);
 
+	if (amdgpu_dm_should_create_sysfs(amdgpu_dm_connector))
+		sysfs_remove_group(&connector->kdev->kobj, &amdgpu_group);
+
+	cec_notifier_conn_unregister(amdgpu_dm_connector->notifier);
 	drm_dp_aux_unregister(&amdgpu_dm_connector->dm_dp_aux.aux);
 }
 
 static void amdgpu_dm_connector_destroy(struct drm_connector *connector)
 {
 	struct amdgpu_dm_connector *aconnector = to_amdgpu_dm_connector(connector);
-	const struct dc_link *link = aconnector->dc_link;
 	struct amdgpu_device *adev = drm_to_adev(connector->dev);
 	struct amdgpu_display_manager *dm = &adev->dm;
-	int i;
 
 	/*
-	 * Call only if mst_mgr was iniitalized before since it's not done
+	 * Call only if mst_mgr was initialized before since it's not done
 	 * for all connector types.
 	 */
 	if (aconnector->mst_mgr.dev)
 		drm_dp_mst_topology_mgr_destroy(&aconnector->mst_mgr);
 
-#if defined(CONFIG_BACKLIGHT_CLASS_DEVICE) ||\
-	defined(CONFIG_BACKLIGHT_CLASS_DEVICE_MODULE)
-	for (i = 0; i < dm->num_of_edps; i++) {
-		if ((link == dm->backlight_link[i]) && dm->backlight_dev[i]) {
-			backlight_device_unregister(dm->backlight_dev[i]);
-			dm->backlight_dev[i] = NULL;
-		}
+	/* Cancel and flush any pending HDMI HPD debounce work */
+	cancel_delayed_work_sync(&aconnector->hdmi_hpd_debounce_work);
+	if (aconnector->hdmi_prev_sink) {
+		dc_sink_release(aconnector->hdmi_prev_sink);
+		aconnector->hdmi_prev_sink = NULL;
+	}
+
+	if (aconnector->bl_idx != -1) {
+		backlight_device_unregister(dm->backlight_dev[aconnector->bl_idx]);
+		dm->backlight_dev[aconnector->bl_idx] = NULL;
 	}
-#endif
 
 	if (aconnector->dc_em_sink)
 		dc_sink_release(aconnector->dc_em_sink);
@@ -6732,10 +7758,6 @@ static void amdgpu_dm_connector_destroy(struct drm_connector *connector)
 	drm_dp_cec_unregister_connector(&aconnector->dm_dp_aux.aux);
 	drm_connector_unregister(connector);
 	drm_connector_cleanup(connector);
-	if (aconnector->i2c) {
-		i2c_del_adapter(&aconnector->i2c->base);
-		kfree(aconnector->i2c);
-	}
 	kfree(aconnector->dm_dp_aux.aux.name);
 
 	kfree(connector);
@@ -6761,8 +7783,13 @@ void amdgpu_dm_connector_funcs_reset(struct drm_connector *connector)
 		state->base.max_requested_bpc = 8;
 		state->vcpi_slots = 0;
 		state->pbn = 0;
-		if (connector->connector_type == DRM_MODE_CONNECTOR_eDP)
-			state->abm_level = amdgpu_dm_abm_level;
+
+		if (connector->connector_type == DRM_MODE_CONNECTOR_eDP) {
+			if (amdgpu_dm_abm_level <= 0)
+				state->abm_level = ABM_LEVEL_IMMEDIATE_DISABLE;
+			else
+				state->abm_level = amdgpu_dm_abm_level;
+		}
 
 		__drm_atomic_helper_connector_reset(connector, &state->base);
 	}
@@ -6800,6 +7827,15 @@ amdgpu_dm_connector_late_register(struct drm_connector *connector)
 		to_amdgpu_dm_connector(connector);
 	int r;
 
+	if (amdgpu_dm_should_create_sysfs(amdgpu_dm_connector)) {
+		r = sysfs_create_group(&connector->kdev->kobj,
+				       &amdgpu_group);
+		if (r)
+			return r;
+	}
+
+	amdgpu_dm_register_backlight_device(amdgpu_dm_connector);
+
 	if ((connector->connector_type == DRM_MODE_CONNECTOR_DisplayPort) ||
 	    (connector->connector_type == DRM_MODE_CONNECTOR_eDP)) {
 		amdgpu_dm_connector->dm_dp_aux.aux.dev = connector->kdev;
@@ -6815,6 +7851,43 @@ amdgpu_dm_connector_late_register(struct drm_connector *connector)
 	return 0;
 }
 
+static void amdgpu_dm_connector_funcs_force(struct drm_connector *connector)
+{
+	struct amdgpu_dm_connector *aconnector = to_amdgpu_dm_connector(connector);
+	struct dc_link *dc_link = aconnector->dc_link;
+	struct dc_sink *dc_em_sink = aconnector->dc_em_sink;
+	const struct drm_edid *drm_edid;
+	struct i2c_adapter *ddc;
+	struct drm_device *dev = connector->dev;
+
+	if (dc_link && dc_link->aux_mode)
+		ddc = &aconnector->dm_dp_aux.aux.ddc;
+	else
+		ddc = &aconnector->i2c->base;
+
+	drm_edid = drm_edid_read_ddc(connector, ddc);
+	drm_edid_connector_update(connector, drm_edid);
+	if (!drm_edid) {
+		drm_err(dev, "No EDID found on connector: %s.\n", connector->name);
+		return;
+	}
+
+	aconnector->drm_edid = drm_edid;
+	/* Update emulated (virtual) sink's EDID */
+	if (dc_em_sink && dc_link) {
+		// FIXME: Get rid of drm_edid_raw()
+		const struct edid *edid = drm_edid_raw(drm_edid);
+
+		memset(&dc_em_sink->edid_caps, 0, sizeof(struct dc_edid_caps));
+		memmove(dc_em_sink->dc_edid.raw_edid, edid,
+			(edid->extensions + 1) * EDID_LENGTH);
+		dm_helpers_parse_edid_caps(
+			dc_link,
+			&dc_em_sink->dc_edid,
+			&dc_em_sink->edid_caps);
+	}
+}
+
 static const struct drm_connector_funcs amdgpu_dm_connector_funcs = {
 	.reset = amdgpu_dm_connector_funcs_reset,
 	.detect = amdgpu_dm_connector_detect,
@@ -6825,7 +7898,8 @@ static const struct drm_connector_funcs amdgpu_dm_connector_funcs = {
 	.atomic_set_property = amdgpu_dm_connector_atomic_set_property,
 	.atomic_get_property = amdgpu_dm_connector_atomic_get_property,
 	.late_register = amdgpu_dm_connector_late_register,
-	.early_unregister = amdgpu_dm_connector_unregister
+	.early_unregister = amdgpu_dm_connector_unregister,
+	.force = amdgpu_dm_connector_funcs_force
 };
 
 static int get_modes(struct drm_connector *connector)
@@ -6835,25 +7909,34 @@ static int get_modes(struct drm_connector *connector)
 
 static void create_eml_sink(struct amdgpu_dm_connector *aconnector)
 {
+	struct drm_connector *connector = &aconnector->base;
+	struct dc_link *dc_link = aconnector->dc_link;
 	struct dc_sink_init_data init_params = {
 			.link = aconnector->dc_link,
 			.sink_signal = SIGNAL_TYPE_VIRTUAL
 	};
-	struct edid *edid;
+	const struct drm_edid *drm_edid;
+	const struct edid *edid;
+	struct i2c_adapter *ddc;
 
-	if (!aconnector->base.edid_blob_ptr) {
-		DRM_ERROR("No EDID firmware found on connector: %s ,forcing to OFF!\n",
-				aconnector->base.name);
+	if (dc_link && dc_link->aux_mode)
+		ddc = &aconnector->dm_dp_aux.aux.ddc;
+	else
+		ddc = &aconnector->i2c->base;
 
-		aconnector->base.force = DRM_FORCE_OFF;
-		aconnector->base.override_edid = false;
+	drm_edid = drm_edid_read_ddc(connector, ddc);
+	drm_edid_connector_update(connector, drm_edid);
+	if (!drm_edid) {
+		drm_err(connector->dev, "No EDID found on connector: %s.\n", connector->name);
 		return;
 	}
 
-	edid = (struct edid *) aconnector->base.edid_blob_ptr->data;
+	if (connector->display_info.is_hdmi)
+		init_params.sink_signal = SIGNAL_TYPE_HDMI_TYPE_A;
 
-	aconnector->edid = edid;
+	aconnector->drm_edid = drm_edid;
 
+	edid = drm_edid_raw(drm_edid); // FIXME: Get rid of drm_edid_raw()
 	aconnector->dc_em_sink = dc_link_add_remote_sink(
 		aconnector->dc_link,
 		(uint8_t *)edid,
@@ -6864,7 +7947,8 @@ static void create_eml_sink(struct amdgpu_dm_connector *aconnector)
 		aconnector->dc_sink = aconnector->dc_link->local_sink ?
 		aconnector->dc_link->local_sink :
 		aconnector->dc_em_sink;
-		dc_sink_retain(aconnector->dc_sink);
+		if (aconnector->dc_sink)
+			dc_sink_retain(aconnector->dc_sink);
 	}
 }
 
@@ -6881,41 +7965,125 @@ static void handle_edid_mgmt(struct amdgpu_dm_connector *aconnector)
 		link->verified_link_cap.link_rate = LINK_RATE_HIGH2;
 	}
 
-
-	aconnector->base.override_edid = true;
 	create_eml_sink(aconnector);
 }
 
-static struct dc_stream_state *
-create_validate_stream_for_sink(struct amdgpu_dm_connector *aconnector,
+static enum dc_status dm_validate_stream_and_context(struct dc *dc,
+						struct dc_stream_state *stream)
+{
+	enum dc_status dc_result = DC_ERROR_UNEXPECTED;
+	struct dc_plane_state *dc_plane_state = NULL;
+	struct dc_state *dc_state = NULL;
+
+	if (!stream)
+		goto cleanup;
+
+	dc_plane_state = dc_create_plane_state(dc);
+	if (!dc_plane_state)
+		goto cleanup;
+
+	dc_state = dc_state_create(dc, NULL);
+	if (!dc_state)
+		goto cleanup;
+
+	/* populate stream to plane */
+	dc_plane_state->src_rect.height  = stream->src.height;
+	dc_plane_state->src_rect.width   = stream->src.width;
+	dc_plane_state->dst_rect.height  = stream->src.height;
+	dc_plane_state->dst_rect.width   = stream->src.width;
+	dc_plane_state->clip_rect.height = stream->src.height;
+	dc_plane_state->clip_rect.width  = stream->src.width;
+	dc_plane_state->plane_size.surface_pitch = ((stream->src.width + 255) / 256) * 256;
+	dc_plane_state->plane_size.surface_size.height = stream->src.height;
+	dc_plane_state->plane_size.surface_size.width  = stream->src.width;
+	dc_plane_state->plane_size.chroma_size.height  = stream->src.height;
+	dc_plane_state->plane_size.chroma_size.width   = stream->src.width;
+	dc_plane_state->format = SURFACE_PIXEL_FORMAT_GRPH_ARGB8888;
+	dc_plane_state->tiling_info.gfx9.swizzle = DC_SW_UNKNOWN;
+	dc_plane_state->rotation = ROTATION_ANGLE_0;
+	dc_plane_state->is_tiling_rotated = false;
+	dc_plane_state->tiling_info.gfx8.array_mode = DC_ARRAY_LINEAR_GENERAL;
+
+	dc_result = dc_validate_stream(dc, stream);
+	if (dc_result == DC_OK)
+		dc_result = dc_validate_plane(dc, dc_plane_state);
+
+	if (dc_result == DC_OK)
+		dc_result = dc_state_add_stream(dc, dc_state, stream);
+
+	if (dc_result == DC_OK && !dc_state_add_plane(
+						dc,
+						stream,
+						dc_plane_state,
+						dc_state))
+		dc_result = DC_FAIL_ATTACH_SURFACES;
+
+	if (dc_result == DC_OK)
+		dc_result = dc_validate_global_state(dc, dc_state, DC_VALIDATE_MODE_ONLY);
+
+cleanup:
+	if (dc_state)
+		dc_state_release(dc_state);
+
+	if (dc_plane_state)
+		dc_plane_state_release(dc_plane_state);
+
+	return dc_result;
+}
+
+struct dc_stream_state *
+create_validate_stream_for_sink(struct drm_connector *connector,
 				const struct drm_display_mode *drm_mode,
 				const struct dm_connector_state *dm_state,
 				const struct dc_stream_state *old_stream)
 {
-	struct drm_connector *connector = &aconnector->base;
+	struct amdgpu_dm_connector *aconnector = NULL;
 	struct amdgpu_device *adev = drm_to_adev(connector->dev);
 	struct dc_stream_state *stream;
 	const struct drm_connector_state *drm_state = dm_state ? &dm_state->base : NULL;
 	int requested_bpc = drm_state ? drm_state->max_requested_bpc : 8;
 	enum dc_status dc_result = DC_OK;
+	uint8_t bpc_limit = 6;
+
+	if (!dm_state)
+		return NULL;
+
+	if (connector->connector_type != DRM_MODE_CONNECTOR_WRITEBACK)
+		aconnector = to_amdgpu_dm_connector(connector);
+
+	if (aconnector &&
+	    (aconnector->dc_link->connector_signal == SIGNAL_TYPE_HDMI_TYPE_A ||
+	     aconnector->dc_link->dpcd_caps.dongle_type == DISPLAY_DONGLE_DP_HDMI_CONVERTER))
+		bpc_limit = 8;
 
 	do {
-		stream = create_stream_for_sink(aconnector, drm_mode,
+		drm_dbg_kms(connector->dev, "Trying with %d bpc\n", requested_bpc);
+		stream = create_stream_for_sink(connector, drm_mode,
 						dm_state, old_stream,
 						requested_bpc);
 		if (stream == NULL) {
-			DRM_ERROR("Failed to create stream for sink!\n");
+			drm_err(adev_to_drm(adev), "Failed to create stream for sink!\n");
 			break;
 		}
 
 		dc_result = dc_validate_stream(adev->dm.dc, stream);
 
+		if (!aconnector) /* writeback connector */
+			return stream;
+
+		if (dc_result == DC_OK && stream->signal == SIGNAL_TYPE_DISPLAY_PORT_MST)
+			dc_result = dm_dp_mst_is_port_support_mode(aconnector, stream);
+
+		if (dc_result == DC_OK)
+			dc_result = dm_validate_stream_and_context(adev->dm.dc, stream);
+
 		if (dc_result != DC_OK) {
-			DRM_DEBUG_KMS("Mode %dx%d (clk %d) failed DC validation with error %d (%s)\n",
+			DRM_DEBUG_KMS("Pruned mode %d x %d (clk %d) %s %s -- %s\n",
 				      drm_mode->hdisplay,
 				      drm_mode->vdisplay,
 				      drm_mode->clock,
-				      dc_result,
+				      dc_pixel_encoding_to_str(stream->timing.pixel_encoding),
+				      dc_color_depth_to_str(stream->timing.display_color_depth),
 				      dc_status_to_str(dc_result));
 
 			dc_stream_release(stream);
@@ -6923,27 +8091,59 @@ create_validate_stream_for_sink(struct amdgpu_dm_connector *aconnector,
 			requested_bpc -= 2; /* lower bpc to retry validation */
 		}
 
-	} while (stream == NULL && requested_bpc >= 6);
-
-	if (dc_result == DC_FAIL_ENC_VALIDATE && !aconnector->force_yuv420_output) {
-		DRM_DEBUG_KMS("Retry forcing YCbCr420 encoding\n");
+	} while (stream == NULL && requested_bpc >= bpc_limit);
 
-		aconnector->force_yuv420_output = true;
-		stream = create_validate_stream_for_sink(aconnector, drm_mode,
-						dm_state, old_stream);
+	switch (dc_result) {
+	/*
+	 * If we failed to validate DP bandwidth stream with the requested RGB color depth,
+	 * we try to fallback and configure in order:
+	 * YUV422 (8bpc, 6bpc)
+	 * YUV420 (8bpc, 6bpc)
+	 */
+	case DC_FAIL_ENC_VALIDATE:
+	case DC_EXCEED_DONGLE_CAP:
+	case DC_NO_DP_LINK_BANDWIDTH:
+		/* recursively entered twice and already tried both YUV422 and YUV420 */
+		if (aconnector->force_yuv422_output && aconnector->force_yuv420_output)
+			break;
+		/* first failure; try YUV422 */
+		if (!aconnector->force_yuv422_output) {
+			drm_dbg_kms(connector->dev, "%s:%d Validation failed with %d, retrying w/ YUV422\n",
+				    __func__, __LINE__, dc_result);
+			aconnector->force_yuv422_output = true;
+		/* recursively entered and YUV422 failed, try YUV420 */
+		} else if (!aconnector->force_yuv420_output) {
+			drm_dbg_kms(connector->dev, "%s:%d Validation failed with %d, retrying w/ YUV420\n",
+				    __func__, __LINE__, dc_result);
+			aconnector->force_yuv420_output = true;
+		}
+		stream = create_validate_stream_for_sink(connector, drm_mode,
+							 dm_state, old_stream);
+		aconnector->force_yuv422_output = false;
 		aconnector->force_yuv420_output = false;
+		break;
+	case DC_OK:
+		break;
+	default:
+		drm_dbg_kms(connector->dev, "%s:%d Unhandled validation failure %d\n",
+			    __func__, __LINE__, dc_result);
+		break;
 	}
 
 	return stream;
 }
 
 enum drm_mode_status amdgpu_dm_connector_mode_valid(struct drm_connector *connector,
-				   struct drm_display_mode *mode)
+				   const struct drm_display_mode *mode)
 {
 	int result = MODE_ERROR;
 	struct dc_sink *dc_sink;
+	struct drm_display_mode *test_mode;
 	/* TODO: Unhardcode stream count */
 	struct dc_stream_state *stream;
+	/* we always have an amdgpu_dm_connector here since we got
+	 * here via the amdgpu_dm_connector_helper_funcs
+	 */
 	struct amdgpu_dm_connector *aconnector = to_amdgpu_dm_connector(connector);
 
 	if ((mode->flags & DRM_MODE_FLAG_INTERLACE) ||
@@ -6962,11 +8162,20 @@ enum drm_mode_status amdgpu_dm_connector_mode_valid(struct drm_connector *connec
 
 	if (dc_sink == NULL && aconnector->base.force != DRM_FORCE_ON_DIGITAL &&
 				aconnector->base.force != DRM_FORCE_ON) {
-		DRM_ERROR("dc_sink is NULL!\n");
+		drm_err(connector->dev, "dc_sink is NULL!\n");
 		goto fail;
 	}
 
-	stream = create_validate_stream_for_sink(aconnector, mode, NULL, NULL);
+	test_mode = drm_mode_duplicate(connector->dev, mode);
+	if (!test_mode)
+		goto fail;
+
+	drm_mode_set_crtcinfo(test_mode, 0);
+
+	stream = create_validate_stream_for_sink(connector, test_mode,
+						 to_dm_connector_state(connector->state),
+						 NULL);
+	drm_mode_destroy(connector->dev, test_mode);
 	if (stream) {
 		dc_stream_release(stream);
 		result = MODE_OK;
@@ -7046,13 +8255,47 @@ amdgpu_dm_connector_atomic_check(struct drm_connector *conn,
 		drm_atomic_get_old_connector_state(state, conn);
 	struct drm_crtc *crtc = new_con_state->crtc;
 	struct drm_crtc_state *new_crtc_state;
+	struct amdgpu_dm_connector *aconn = to_amdgpu_dm_connector(conn);
 	int ret;
 
+	if (WARN_ON(unlikely(!old_con_state || !new_con_state)))
+		return -EINVAL;
+
 	trace_amdgpu_dm_connector_atomic_check(new_con_state);
 
+	if (conn->connector_type == DRM_MODE_CONNECTOR_DisplayPort) {
+		ret = drm_dp_mst_root_conn_atomic_check(new_con_state, &aconn->mst_mgr);
+		if (ret < 0)
+			return ret;
+	}
+
 	if (!crtc)
 		return 0;
 
+	if (new_con_state->privacy_screen_sw_state != old_con_state->privacy_screen_sw_state) {
+		new_crtc_state = drm_atomic_get_crtc_state(state, crtc);
+		if (IS_ERR(new_crtc_state))
+			return PTR_ERR(new_crtc_state);
+
+		new_crtc_state->mode_changed = true;
+	}
+
+	if (new_con_state->colorspace != old_con_state->colorspace) {
+		new_crtc_state = drm_atomic_get_crtc_state(state, crtc);
+		if (IS_ERR(new_crtc_state))
+			return PTR_ERR(new_crtc_state);
+
+		new_crtc_state->mode_changed = true;
+	}
+
+	if (new_con_state->content_type != old_con_state->content_type) {
+		new_crtc_state = drm_atomic_get_crtc_state(state, crtc);
+		if (IS_ERR(new_crtc_state))
+			return PTR_ERR(new_crtc_state);
+
+		new_crtc_state->mode_changed = true;
+	}
+
 	if (!drm_connector_atomic_hdr_metadata_equal(old_con_state, new_con_state)) {
 		struct dc_info_packet hdr_infopacket;
 
@@ -7075,7 +8318,7 @@ amdgpu_dm_connector_atomic_check(struct drm_connector *conn,
 		 * set is permissible, however. So only force a
 		 * modeset if we're entering or exiting HDR.
 		 */
-		new_crtc_state->mode_changed =
+		new_crtc_state->mode_changed = new_crtc_state->mode_changed ||
 			!old_con_state->hdr_output_metadata ||
 			!new_con_state->hdr_output_metadata;
 	}
@@ -7096,136 +8339,29 @@ amdgpu_dm_connector_helper_funcs = {
 	.atomic_check = amdgpu_dm_connector_atomic_check,
 };
 
-static void dm_crtc_helper_disable(struct drm_crtc *crtc)
-{
-}
-
-static int count_crtc_active_planes(struct drm_crtc_state *new_crtc_state)
-{
-	struct drm_atomic_state *state = new_crtc_state->state;
-	struct drm_plane *plane;
-	int num_active = 0;
-
-	drm_for_each_plane_mask(plane, state->dev, new_crtc_state->plane_mask) {
-		struct drm_plane_state *new_plane_state;
-
-		/* Cursor planes are "fake". */
-		if (plane->type == DRM_PLANE_TYPE_CURSOR)
-			continue;
-
-		new_plane_state = drm_atomic_get_new_plane_state(state, plane);
-
-		if (!new_plane_state) {
-			/*
-			 * The plane is enable on the CRTC and hasn't changed
-			 * state. This means that it previously passed
-			 * validation and is therefore enabled.
-			 */
-			num_active += 1;
-			continue;
-		}
-
-		/* We need a framebuffer to be considered enabled. */
-		num_active += (new_plane_state->fb != NULL);
-	}
-
-	return num_active;
-}
-
-static void dm_update_crtc_active_planes(struct drm_crtc *crtc,
-					 struct drm_crtc_state *new_crtc_state)
-{
-	struct dm_crtc_state *dm_new_crtc_state =
-		to_dm_crtc_state(new_crtc_state);
-
-	dm_new_crtc_state->active_planes = 0;
-
-	if (!dm_new_crtc_state->stream)
-		return;
-
-	dm_new_crtc_state->active_planes =
-		count_crtc_active_planes(new_crtc_state);
-}
-
-static int dm_crtc_helper_atomic_check(struct drm_crtc *crtc,
-				       struct drm_atomic_state *state)
-{
-	struct drm_crtc_state *crtc_state = drm_atomic_get_new_crtc_state(state,
-									  crtc);
-	struct amdgpu_device *adev = drm_to_adev(crtc->dev);
-	struct dc *dc = adev->dm.dc;
-	struct dm_crtc_state *dm_crtc_state = to_dm_crtc_state(crtc_state);
-	int ret = -EINVAL;
-
-	trace_amdgpu_dm_crtc_atomic_check(crtc_state);
-
-	dm_update_crtc_active_planes(crtc, crtc_state);
-
-	if (WARN_ON(unlikely(!dm_crtc_state->stream &&
-		     modeset_required(crtc_state, NULL, dm_crtc_state->stream)))) {
-		return ret;
-	}
-
-	/*
-	 * We require the primary plane to be enabled whenever the CRTC is, otherwise
-	 * drm_mode_cursor_universal may end up trying to enable the cursor plane while all other
-	 * planes are disabled, which is not supported by the hardware. And there is legacy
-	 * userspace which stops using the HW cursor altogether in response to the resulting EINVAL.
-	 */
-	if (crtc_state->enable &&
-	    !(crtc_state->plane_mask & drm_plane_mask(crtc->primary))) {
-		DRM_DEBUG_ATOMIC("Can't enable a CRTC without enabling the primary plane\n");
-		return -EINVAL;
-	}
-
-	/* In some use cases, like reset, no stream is attached */
-	if (!dm_crtc_state->stream)
-		return 0;
-
-	if (dc_validate_stream(dc, dm_crtc_state->stream) == DC_OK)
-		return 0;
-
-	DRM_DEBUG_ATOMIC("Failed DC stream validation\n");
-	return ret;
-}
-
-static bool dm_crtc_helper_mode_fixup(struct drm_crtc *crtc,
-				      const struct drm_display_mode *mode,
-				      struct drm_display_mode *adjusted_mode)
-{
-	return true;
-}
-
-static const struct drm_crtc_helper_funcs amdgpu_dm_crtc_helper_funcs = {
-	.disable = dm_crtc_helper_disable,
-	.atomic_check = dm_crtc_helper_atomic_check,
-	.mode_fixup = dm_crtc_helper_mode_fixup,
-	.get_scanout_position = amdgpu_crtc_get_scanout_position,
-};
-
 static void dm_encoder_helper_disable(struct drm_encoder *encoder)
 {
 
 }
 
-static int convert_dc_color_depth_into_bpc (enum dc_color_depth display_color_depth)
+int convert_dc_color_depth_into_bpc(enum dc_color_depth display_color_depth)
 {
 	switch (display_color_depth) {
-		case COLOR_DEPTH_666:
-			return 6;
-		case COLOR_DEPTH_888:
-			return 8;
-		case COLOR_DEPTH_101010:
-			return 10;
-		case COLOR_DEPTH_121212:
-			return 12;
-		case COLOR_DEPTH_141414:
-			return 14;
-		case COLOR_DEPTH_161616:
-			return 16;
-		default:
-			break;
-		}
+	case COLOR_DEPTH_666:
+		return 6;
+	case COLOR_DEPTH_888:
+		return 8;
+	case COLOR_DEPTH_101010:
+		return 10;
+	case COLOR_DEPTH_121212:
+		return 12;
+	case COLOR_DEPTH_141414:
+		return 14;
+	case COLOR_DEPTH_161616:
+		return 16;
+	default:
+		break;
+	}
 	return 0;
 }
 
@@ -7240,35 +8376,59 @@ static int dm_encoder_helper_atomic_check(struct drm_encoder *encoder,
 	const struct drm_display_mode *adjusted_mode = &crtc_state->adjusted_mode;
 	struct drm_dp_mst_topology_mgr *mst_mgr;
 	struct drm_dp_mst_port *mst_port;
+	struct drm_dp_mst_topology_state *mst_state;
 	enum dc_color_depth color_depth;
 	int clock, bpp = 0;
 	bool is_y420 = false;
 
-	if (!aconnector->port || !aconnector->dc_sink)
+	if ((connector->connector_type == DRM_MODE_CONNECTOR_eDP) ||
+	    (connector->connector_type == DRM_MODE_CONNECTOR_LVDS)) {
+		struct amdgpu_encoder *amdgpu_encoder = to_amdgpu_encoder(encoder);
+		struct drm_display_mode *native_mode = &amdgpu_encoder->native_mode;
+		enum drm_mode_status result;
+
+		result = drm_crtc_helper_mode_valid_fixed(encoder->crtc, adjusted_mode, native_mode);
+		if (result != MODE_OK && dm_new_connector_state->scaling == RMX_OFF) {
+			drm_dbg_driver(encoder->dev,
+				       "mode %dx%d@%dHz is not native, enabling scaling\n",
+				       adjusted_mode->hdisplay, adjusted_mode->vdisplay,
+				       drm_mode_vrefresh(adjusted_mode));
+			dm_new_connector_state->scaling = RMX_ASPECT;
+		}
 		return 0;
+	}
 
-	mst_port = aconnector->port;
-	mst_mgr = &aconnector->mst_port->mst_mgr;
+	if (!aconnector->mst_output_port)
+		return 0;
+
+	mst_port = aconnector->mst_output_port;
+	mst_mgr = &aconnector->mst_root->mst_mgr;
 
 	if (!crtc_state->connectors_changed && !crtc_state->mode_changed)
 		return 0;
 
+	mst_state = drm_atomic_get_mst_topology_state(state, mst_mgr);
+	if (IS_ERR(mst_state))
+		return PTR_ERR(mst_state);
+
+	mst_state->pbn_div.full = dm_mst_get_pbn_divider(aconnector->mst_root->dc_link);
+
 	if (!state->duplicated) {
 		int max_bpc = conn_state->max_requested_bpc;
+
 		is_y420 = drm_mode_is_420_also(&connector->display_info, adjusted_mode) &&
-				aconnector->force_yuv420_output;
+			  aconnector->force_yuv420_output;
 		color_depth = convert_color_depth_from_display_info(connector,
 								    is_y420,
 								    max_bpc);
 		bpp = convert_dc_color_depth_into_bpc(color_depth) * 3;
 		clock = adjusted_mode->clock;
-		dm_new_connector_state->pbn = drm_dp_calc_pbn_mode(clock, bpp, false);
+		dm_new_connector_state->pbn = drm_dp_calc_pbn_mode(clock, bpp << 4);
 	}
-	dm_new_connector_state->vcpi_slots = drm_dp_atomic_find_vcpi_slots(state,
-									   mst_mgr,
-									   mst_port,
-									   dm_new_connector_state->pbn,
-									   dm_mst_get_pbn_divider(aconnector->dc_link));
+
+	dm_new_connector_state->vcpi_slots =
+		drm_dp_atomic_find_time_slots(state, mst_mgr, mst_port,
+					      dm_new_connector_state->pbn);
 	if (dm_new_connector_state->vcpi_slots < 0) {
 		DRM_DEBUG_ATOMIC("failed finding vcpi slots: %d\n", (int)dm_new_connector_state->vcpi_slots);
 		return dm_new_connector_state->vcpi_slots;
@@ -7281,7 +8441,6 @@ const struct drm_encoder_helper_funcs amdgpu_dm_encoder_helper_funcs = {
 	.atomic_check = dm_encoder_helper_atomic_check
 };
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 static int dm_update_mst_vcpi_slots_for_dsc(struct drm_atomic_state *state,
 					    struct dc_state *dc_state,
 					    struct dsc_mst_fairness_vars *vars)
@@ -7291,14 +8450,17 @@ static int dm_update_mst_vcpi_slots_for_dsc(struct drm_atomic_state *state,
 	struct drm_connector_state *new_con_state;
 	struct amdgpu_dm_connector *aconnector;
 	struct dm_connector_state *dm_conn_state;
-	int i, j;
-	int vcpi, pbn_div, pbn, slot_num = 0;
+	int i, j, ret;
+	int vcpi, pbn_div, pbn = 0, slot_num = 0;
 
 	for_each_new_connector_in_state(state, connector, new_con_state, i) {
 
+		if (connector->connector_type == DRM_MODE_CONNECTOR_WRITEBACK)
+			continue;
+
 		aconnector = to_amdgpu_dm_connector(connector);
 
-		if (!aconnector->port)
+		if (!aconnector->mst_output_port)
 			continue;
 
 		if (!new_con_state || !new_con_state->crtc)
@@ -7311,7 +8473,7 @@ static int dm_update_mst_vcpi_slots_for_dsc(struct drm_atomic_state *state,
 			if (!stream)
 				continue;
 
-			if ((struct amdgpu_dm_connector*)stream->dm_stream_context == aconnector)
+			if ((struct amdgpu_dm_connector *)stream->dm_stream_context == aconnector)
 				break;
 
 			stream = NULL;
@@ -7329,7 +8491,7 @@ static int dm_update_mst_vcpi_slots_for_dsc(struct drm_atomic_state *state,
 			}
 		}
 
-		if (j == dc_state->stream_count)
+		if (j == dc_state->stream_count || pbn_div == 0)
 			continue;
 
 		slot_num = DIV_ROUND_UP(pbn, pbn_div);
@@ -7338,18 +8500,15 @@ static int dm_update_mst_vcpi_slots_for_dsc(struct drm_atomic_state *state,
 			dm_conn_state->pbn = pbn;
 			dm_conn_state->vcpi_slots = slot_num;
 
-			drm_dp_mst_atomic_enable_dsc(state,
-						     aconnector->port,
-						     dm_conn_state->pbn,
-						     0,
-						     false);
+			ret = drm_dp_mst_atomic_enable_dsc(state, aconnector->mst_output_port,
+							   dm_conn_state->pbn, false);
+			if (ret < 0)
+				return ret;
+
 			continue;
 		}
 
-		vcpi = drm_dp_mst_atomic_enable_dsc(state,
-						    aconnector->port,
-						    pbn, pbn_div,
-						    true);
+		vcpi = drm_dp_mst_atomic_enable_dsc(state, aconnector->mst_output_port, pbn, true);
 		if (vcpi < 0)
 			return vcpi;
 
@@ -7358,525 +8517,8 @@ static int dm_update_mst_vcpi_slots_for_dsc(struct drm_atomic_state *state,
 	}
 	return 0;
 }
-#endif
-
-static void dm_drm_plane_reset(struct drm_plane *plane)
-{
-	struct dm_plane_state *amdgpu_state = NULL;
-
-	if (plane->state)
-		plane->funcs->atomic_destroy_state(plane, plane->state);
-
-	amdgpu_state = kzalloc(sizeof(*amdgpu_state), GFP_KERNEL);
-	WARN_ON(amdgpu_state == NULL);
-
-	if (amdgpu_state)
-		__drm_atomic_helper_plane_reset(plane, &amdgpu_state->base);
-}
-
-static struct drm_plane_state *
-dm_drm_plane_duplicate_state(struct drm_plane *plane)
-{
-	struct dm_plane_state *dm_plane_state, *old_dm_plane_state;
-
-	old_dm_plane_state = to_dm_plane_state(plane->state);
-	dm_plane_state = kzalloc(sizeof(*dm_plane_state), GFP_KERNEL);
-	if (!dm_plane_state)
-		return NULL;
-
-	__drm_atomic_helper_plane_duplicate_state(plane, &dm_plane_state->base);
-
-	if (old_dm_plane_state->dc_state) {
-		dm_plane_state->dc_state = old_dm_plane_state->dc_state;
-		dc_plane_state_retain(dm_plane_state->dc_state);
-	}
-
-	return &dm_plane_state->base;
-}
-
-static void dm_drm_plane_destroy_state(struct drm_plane *plane,
-				struct drm_plane_state *state)
-{
-	struct dm_plane_state *dm_plane_state = to_dm_plane_state(state);
-
-	if (dm_plane_state->dc_state)
-		dc_plane_state_release(dm_plane_state->dc_state);
-
-	drm_atomic_helper_plane_destroy_state(plane, state);
-}
-
-static const struct drm_plane_funcs dm_plane_funcs = {
-	.update_plane	= drm_atomic_helper_update_plane,
-	.disable_plane	= drm_atomic_helper_disable_plane,
-	.destroy	= drm_primary_helper_destroy,
-	.reset = dm_drm_plane_reset,
-	.atomic_duplicate_state = dm_drm_plane_duplicate_state,
-	.atomic_destroy_state = dm_drm_plane_destroy_state,
-	.format_mod_supported = dm_plane_format_mod_supported,
-};
-
-static int dm_plane_helper_prepare_fb(struct drm_plane *plane,
-				      struct drm_plane_state *new_state)
-{
-	struct amdgpu_framebuffer *afb;
-	struct drm_gem_object *obj;
-	struct amdgpu_device *adev;
-	struct amdgpu_bo *rbo;
-	struct dm_plane_state *dm_plane_state_new, *dm_plane_state_old;
-	struct list_head list;
-	struct ttm_validate_buffer tv;
-	struct ww_acquire_ctx ticket;
-	uint32_t domain;
-	int r;
-
-	if (!new_state->fb) {
-		DRM_DEBUG_KMS("No FB bound\n");
-		return 0;
-	}
-
-	afb = to_amdgpu_framebuffer(new_state->fb);
-	obj = new_state->fb->obj[0];
-	rbo = gem_to_amdgpu_bo(obj);
-	adev = amdgpu_ttm_adev(rbo->tbo.bdev);
-	INIT_LIST_HEAD(&list);
-
-	tv.bo = &rbo->tbo;
-	tv.num_shared = 1;
-	list_add(&tv.head, &list);
-
-	r = ttm_eu_reserve_buffers(&ticket, &list, false, NULL);
-	if (r) {
-		dev_err(adev->dev, "fail to reserve bo (%d)\n", r);
-		return r;
-	}
-
-	if (plane->type != DRM_PLANE_TYPE_CURSOR)
-		domain = amdgpu_display_supported_domains(adev, rbo->flags);
-	else
-		domain = AMDGPU_GEM_DOMAIN_VRAM;
-
-	r = amdgpu_bo_pin(rbo, domain);
-	if (unlikely(r != 0)) {
-		if (r != -ERESTARTSYS)
-			DRM_ERROR("Failed to pin framebuffer with error %d\n", r);
-		ttm_eu_backoff_reservation(&ticket, &list);
-		return r;
-	}
-
-	r = amdgpu_ttm_alloc_gart(&rbo->tbo);
-	if (unlikely(r != 0)) {
-		amdgpu_bo_unpin(rbo);
-		ttm_eu_backoff_reservation(&ticket, &list);
-		DRM_ERROR("%p bind failed\n", rbo);
-		return r;
-	}
-
-	ttm_eu_backoff_reservation(&ticket, &list);
-
-	afb->address = amdgpu_bo_gpu_offset(rbo);
-
-	amdgpu_bo_ref(rbo);
-
-	/**
-	 * We don't do surface updates on planes that have been newly created,
-	 * but we also don't have the afb->address during atomic check.
-	 *
-	 * Fill in buffer attributes depending on the address here, but only on
-	 * newly created planes since they're not being used by DC yet and this
-	 * won't modify global state.
-	 */
-	dm_plane_state_old = to_dm_plane_state(plane->state);
-	dm_plane_state_new = to_dm_plane_state(new_state);
-
-	if (dm_plane_state_new->dc_state &&
-	    dm_plane_state_old->dc_state != dm_plane_state_new->dc_state) {
-		struct dc_plane_state *plane_state =
-			dm_plane_state_new->dc_state;
-		bool force_disable_dcc = !plane_state->dcc.enable;
-
-		fill_plane_buffer_attributes(
-			adev, afb, plane_state->format, plane_state->rotation,
-			afb->tiling_flags,
-			&plane_state->tiling_info, &plane_state->plane_size,
-			&plane_state->dcc, &plane_state->address,
-			afb->tmz_surface, force_disable_dcc);
-	}
-
-	return 0;
-}
 
-static void dm_plane_helper_cleanup_fb(struct drm_plane *plane,
-				       struct drm_plane_state *old_state)
-{
-	struct amdgpu_bo *rbo;
-	int r;
-
-	if (!old_state->fb)
-		return;
-
-	rbo = gem_to_amdgpu_bo(old_state->fb->obj[0]);
-	r = amdgpu_bo_reserve(rbo, false);
-	if (unlikely(r)) {
-		DRM_ERROR("failed to reserve rbo before unpin\n");
-		return;
-	}
-
-	amdgpu_bo_unpin(rbo);
-	amdgpu_bo_unreserve(rbo);
-	amdgpu_bo_unref(&rbo);
-}
-
-static int dm_plane_helper_check_state(struct drm_plane_state *state,
-				       struct drm_crtc_state *new_crtc_state)
-{
-	struct drm_framebuffer *fb = state->fb;
-	int min_downscale, max_upscale;
-	int min_scale = 0;
-	int max_scale = INT_MAX;
-
-	/* Plane enabled? Validate viewport and get scaling factors from plane caps. */
-	if (fb && state->crtc) {
-		/* Validate viewport to cover the case when only the position changes */
-		if (state->plane->type != DRM_PLANE_TYPE_CURSOR) {
-			int viewport_width = state->crtc_w;
-			int viewport_height = state->crtc_h;
-
-			if (state->crtc_x < 0)
-				viewport_width += state->crtc_x;
-			else if (state->crtc_x + state->crtc_w > new_crtc_state->mode.crtc_hdisplay)
-				viewport_width = new_crtc_state->mode.crtc_hdisplay - state->crtc_x;
-
-			if (state->crtc_y < 0)
-				viewport_height += state->crtc_y;
-			else if (state->crtc_y + state->crtc_h > new_crtc_state->mode.crtc_vdisplay)
-				viewport_height = new_crtc_state->mode.crtc_vdisplay - state->crtc_y;
-
-			if (viewport_width < 0 || viewport_height < 0) {
-				DRM_DEBUG_ATOMIC("Plane completely outside of screen\n");
-				return -EINVAL;
-			} else if (viewport_width < MIN_VIEWPORT_SIZE*2) { /* x2 for width is because of pipe-split. */
-				DRM_DEBUG_ATOMIC("Viewport width %d smaller than %d\n", viewport_width, MIN_VIEWPORT_SIZE*2);
-				return -EINVAL;
-			} else if (viewport_height < MIN_VIEWPORT_SIZE) {
-				DRM_DEBUG_ATOMIC("Viewport height %d smaller than %d\n", viewport_height, MIN_VIEWPORT_SIZE);
-				return -EINVAL;
-			}
-
-		}
-
-		/* Get min/max allowed scaling factors from plane caps. */
-		get_min_max_dc_plane_scaling(state->crtc->dev, fb,
-					     &min_downscale, &max_upscale);
-		/*
-		 * Convert to drm convention: 16.16 fixed point, instead of dc's
-		 * 1.0 == 1000. Also drm scaling is src/dst instead of dc's
-		 * dst/src, so min_scale = 1.0 / max_upscale, etc.
-		 */
-		min_scale = (1000 << 16) / max_upscale;
-		max_scale = (1000 << 16) / min_downscale;
-	}
-
-	return drm_atomic_helper_check_plane_state(
-		state, new_crtc_state, min_scale, max_scale, true, true);
-}
-
-static int dm_plane_atomic_check(struct drm_plane *plane,
-				 struct drm_atomic_state *state)
-{
-	struct drm_plane_state *new_plane_state = drm_atomic_get_new_plane_state(state,
-										 plane);
-	struct amdgpu_device *adev = drm_to_adev(plane->dev);
-	struct dc *dc = adev->dm.dc;
-	struct dm_plane_state *dm_plane_state;
-	struct dc_scaling_info scaling_info;
-	struct drm_crtc_state *new_crtc_state;
-	int ret;
-
-	trace_amdgpu_dm_plane_atomic_check(new_plane_state);
-
-	dm_plane_state = to_dm_plane_state(new_plane_state);
-
-	if (!dm_plane_state->dc_state)
-		return 0;
-
-	new_crtc_state =
-		drm_atomic_get_new_crtc_state(state,
-					      new_plane_state->crtc);
-	if (!new_crtc_state)
-		return -EINVAL;
-
-	ret = dm_plane_helper_check_state(new_plane_state, new_crtc_state);
-	if (ret)
-		return ret;
-
-	ret = fill_dc_scaling_info(adev, new_plane_state, &scaling_info);
-	if (ret)
-		return ret;
-
-	if (dc_validate_plane(dc, dm_plane_state->dc_state) == DC_OK)
-		return 0;
-
-	return -EINVAL;
-}
-
-static int dm_plane_atomic_async_check(struct drm_plane *plane,
-				       struct drm_atomic_state *state)
-{
-	/* Only support async updates on cursor planes. */
-	if (plane->type != DRM_PLANE_TYPE_CURSOR)
-		return -EINVAL;
-
-	return 0;
-}
-
-static void dm_plane_atomic_async_update(struct drm_plane *plane,
-					 struct drm_atomic_state *state)
-{
-	struct drm_plane_state *new_state = drm_atomic_get_new_plane_state(state,
-									   plane);
-	struct drm_plane_state *old_state =
-		drm_atomic_get_old_plane_state(state, plane);
-
-	trace_amdgpu_dm_atomic_update_cursor(new_state);
-
-	swap(plane->state->fb, new_state->fb);
-
-	plane->state->src_x = new_state->src_x;
-	plane->state->src_y = new_state->src_y;
-	plane->state->src_w = new_state->src_w;
-	plane->state->src_h = new_state->src_h;
-	plane->state->crtc_x = new_state->crtc_x;
-	plane->state->crtc_y = new_state->crtc_y;
-	plane->state->crtc_w = new_state->crtc_w;
-	plane->state->crtc_h = new_state->crtc_h;
-
-	handle_cursor_update(plane, old_state);
-}
-
-static const struct drm_plane_helper_funcs dm_plane_helper_funcs = {
-	.prepare_fb = dm_plane_helper_prepare_fb,
-	.cleanup_fb = dm_plane_helper_cleanup_fb,
-	.atomic_check = dm_plane_atomic_check,
-	.atomic_async_check = dm_plane_atomic_async_check,
-	.atomic_async_update = dm_plane_atomic_async_update
-};
-
-/*
- * TODO: these are currently initialized to rgb formats only.
- * For future use cases we should either initialize them dynamically based on
- * plane capabilities, or initialize this array to all formats, so internal drm
- * check will succeed, and let DC implement proper check
- */
-static const uint32_t rgb_formats[] = {
-	DRM_FORMAT_XRGB8888,
-	DRM_FORMAT_ARGB8888,
-	DRM_FORMAT_RGBA8888,
-	DRM_FORMAT_XRGB2101010,
-	DRM_FORMAT_XBGR2101010,
-	DRM_FORMAT_ARGB2101010,
-	DRM_FORMAT_ABGR2101010,
-	DRM_FORMAT_XRGB16161616,
-	DRM_FORMAT_XBGR16161616,
-	DRM_FORMAT_ARGB16161616,
-	DRM_FORMAT_ABGR16161616,
-	DRM_FORMAT_XBGR8888,
-	DRM_FORMAT_ABGR8888,
-	DRM_FORMAT_RGB565,
-};
-
-static const uint32_t overlay_formats[] = {
-	DRM_FORMAT_XRGB8888,
-	DRM_FORMAT_ARGB8888,
-	DRM_FORMAT_RGBA8888,
-	DRM_FORMAT_XBGR8888,
-	DRM_FORMAT_ABGR8888,
-	DRM_FORMAT_RGB565
-};
-
-static const u32 cursor_formats[] = {
-	DRM_FORMAT_ARGB8888
-};
-
-static int get_plane_formats(const struct drm_plane *plane,
-			     const struct dc_plane_cap *plane_cap,
-			     uint32_t *formats, int max_formats)
-{
-	int i, num_formats = 0;
-
-	/*
-	 * TODO: Query support for each group of formats directly from
-	 * DC plane caps. This will require adding more formats to the
-	 * caps list.
-	 */
-
-	switch (plane->type) {
-	case DRM_PLANE_TYPE_PRIMARY:
-		for (i = 0; i < ARRAY_SIZE(rgb_formats); ++i) {
-			if (num_formats >= max_formats)
-				break;
-
-			formats[num_formats++] = rgb_formats[i];
-		}
-
-		if (plane_cap && plane_cap->pixel_format_support.nv12)
-			formats[num_formats++] = DRM_FORMAT_NV12;
-		if (plane_cap && plane_cap->pixel_format_support.p010)
-			formats[num_formats++] = DRM_FORMAT_P010;
-		if (plane_cap && plane_cap->pixel_format_support.fp16) {
-			formats[num_formats++] = DRM_FORMAT_XRGB16161616F;
-			formats[num_formats++] = DRM_FORMAT_ARGB16161616F;
-			formats[num_formats++] = DRM_FORMAT_XBGR16161616F;
-			formats[num_formats++] = DRM_FORMAT_ABGR16161616F;
-		}
-		break;
-
-	case DRM_PLANE_TYPE_OVERLAY:
-		for (i = 0; i < ARRAY_SIZE(overlay_formats); ++i) {
-			if (num_formats >= max_formats)
-				break;
-
-			formats[num_formats++] = overlay_formats[i];
-		}
-		break;
-
-	case DRM_PLANE_TYPE_CURSOR:
-		for (i = 0; i < ARRAY_SIZE(cursor_formats); ++i) {
-			if (num_formats >= max_formats)
-				break;
-
-			formats[num_formats++] = cursor_formats[i];
-		}
-		break;
-	}
-
-	return num_formats;
-}
-
-static int amdgpu_dm_plane_init(struct amdgpu_display_manager *dm,
-				struct drm_plane *plane,
-				unsigned long possible_crtcs,
-				const struct dc_plane_cap *plane_cap)
-{
-	uint32_t formats[32];
-	int num_formats;
-	int res = -EPERM;
-	unsigned int supported_rotations;
-	uint64_t *modifiers = NULL;
-
-	num_formats = get_plane_formats(plane, plane_cap, formats,
-					ARRAY_SIZE(formats));
-
-	res = get_plane_modifiers(dm->adev, plane->type, &modifiers);
-	if (res)
-		return res;
-
-	res = drm_universal_plane_init(adev_to_drm(dm->adev), plane, possible_crtcs,
-				       &dm_plane_funcs, formats, num_formats,
-				       modifiers, plane->type, NULL);
-	kfree(modifiers);
-	if (res)
-		return res;
-
-	if (plane->type == DRM_PLANE_TYPE_OVERLAY &&
-	    plane_cap && plane_cap->per_pixel_alpha) {
-		unsigned int blend_caps = BIT(DRM_MODE_BLEND_PIXEL_NONE) |
-					  BIT(DRM_MODE_BLEND_PREMULTI);
-
-		drm_plane_create_alpha_property(plane);
-		drm_plane_create_blend_mode_property(plane, blend_caps);
-	}
-
-	if (plane->type == DRM_PLANE_TYPE_PRIMARY &&
-	    plane_cap &&
-	    (plane_cap->pixel_format_support.nv12 ||
-	     plane_cap->pixel_format_support.p010)) {
-		/* This only affects YUV formats. */
-		drm_plane_create_color_properties(
-			plane,
-			BIT(DRM_COLOR_YCBCR_BT601) |
-			BIT(DRM_COLOR_YCBCR_BT709) |
-			BIT(DRM_COLOR_YCBCR_BT2020),
-			BIT(DRM_COLOR_YCBCR_LIMITED_RANGE) |
-			BIT(DRM_COLOR_YCBCR_FULL_RANGE),
-			DRM_COLOR_YCBCR_BT709, DRM_COLOR_YCBCR_LIMITED_RANGE);
-	}
-
-	supported_rotations =
-		DRM_MODE_ROTATE_0 | DRM_MODE_ROTATE_90 |
-		DRM_MODE_ROTATE_180 | DRM_MODE_ROTATE_270;
-
-	if (dm->adev->asic_type >= CHIP_BONAIRE &&
-	    plane->type != DRM_PLANE_TYPE_CURSOR)
-		drm_plane_create_rotation_property(plane, DRM_MODE_ROTATE_0,
-						   supported_rotations);
-
-	drm_plane_helper_add(plane, &dm_plane_helper_funcs);
-
-	/* Create (reset) the plane state */
-	if (plane->funcs->reset)
-		plane->funcs->reset(plane);
-
-	return 0;
-}
-
-static int amdgpu_dm_crtc_init(struct amdgpu_display_manager *dm,
-			       struct drm_plane *plane,
-			       uint32_t crtc_index)
-{
-	struct amdgpu_crtc *acrtc = NULL;
-	struct drm_plane *cursor_plane;
-
-	int res = -ENOMEM;
-
-	cursor_plane = kzalloc(sizeof(*cursor_plane), GFP_KERNEL);
-	if (!cursor_plane)
-		goto fail;
-
-	cursor_plane->type = DRM_PLANE_TYPE_CURSOR;
-	res = amdgpu_dm_plane_init(dm, cursor_plane, 0, NULL);
-
-	acrtc = kzalloc(sizeof(struct amdgpu_crtc), GFP_KERNEL);
-	if (!acrtc)
-		goto fail;
-
-	res = drm_crtc_init_with_planes(
-			dm->ddev,
-			&acrtc->base,
-			plane,
-			cursor_plane,
-			&amdgpu_dm_crtc_funcs, NULL);
-
-	if (res)
-		goto fail;
-
-	drm_crtc_helper_add(&acrtc->base, &amdgpu_dm_crtc_helper_funcs);
-
-	/* Create (reset) the plane state */
-	if (acrtc->base.funcs->reset)
-		acrtc->base.funcs->reset(&acrtc->base);
-
-	acrtc->max_cursor_width = dm->adev->dm.dc->caps.max_cursor_size;
-	acrtc->max_cursor_height = dm->adev->dm.dc->caps.max_cursor_size;
-
-	acrtc->crtc_id = crtc_index;
-	acrtc->base.enabled = false;
-	acrtc->otg_inst = -1;
-
-	dm->adev->mode_info.crtcs[crtc_index] = acrtc;
-	drm_crtc_enable_color_mgmt(&acrtc->base, MAX_COLOR_LUT_ENTRIES,
-				   true, MAX_COLOR_LUT_ENTRIES);
-	drm_mode_crtc_set_gamma_size(&acrtc->base, MAX_COLOR_LEGACY_LUT_ENTRIES);
-
-	return 0;
-
-fail:
-	kfree(acrtc);
-	kfree(cursor_plane);
-	return res;
-}
-
-
-static int to_drm_connector_type(enum signal_type st)
+static int to_drm_connector_type(enum signal_type st, uint32_t connector_id)
 {
 	switch (st) {
 	case SIGNAL_TYPE_HDMI_TYPE_A:
@@ -7892,6 +8534,10 @@ static int to_drm_connector_type(enum signal_type st)
 		return DRM_MODE_CONNECTOR_DisplayPort;
 	case SIGNAL_TYPE_DVI_DUAL_LINK:
 	case SIGNAL_TYPE_DVI_SINGLE_LINK:
+		if (connector_id == CONNECTOR_ID_SINGLE_LINK_DVII ||
+			connector_id == CONNECTOR_ID_DUAL_LINK_DVII)
+			return DRM_MODE_CONNECTOR_DVII;
+
 		return DRM_MODE_CONNECTOR_DVID;
 	case SIGNAL_TYPE_VIRTUAL:
 		return DRM_MODE_CONNECTOR_VIRTUAL;
@@ -7943,7 +8589,7 @@ static void amdgpu_dm_get_native_mode(struct drm_connector *connector)
 
 static struct drm_display_mode *
 amdgpu_dm_create_common_mode(struct drm_encoder *encoder,
-			     char *name,
+			     const char *name,
 			     int hdisplay, int vdisplay)
 {
 	struct drm_device *dev = encoder->dev;
@@ -7965,6 +8611,24 @@ amdgpu_dm_create_common_mode(struct drm_encoder *encoder,
 
 }
 
+static const struct amdgpu_dm_mode_size {
+	char name[DRM_DISPLAY_MODE_LEN];
+	int w;
+	int h;
+} common_modes[] = {
+	{  "640x480",  640,  480},
+	{  "800x600",  800,  600},
+	{ "1024x768", 1024,  768},
+	{ "1280x720", 1280,  720},
+	{ "1280x800", 1280,  800},
+	{"1280x1024", 1280, 1024},
+	{ "1440x900", 1440,  900},
+	{"1680x1050", 1680, 1050},
+	{"1600x1200", 1600, 1200},
+	{"1920x1080", 1920, 1080},
+	{"1920x1200", 1920, 1200}
+};
+
 static void amdgpu_dm_connector_add_common_modes(struct drm_encoder *encoder,
 						 struct drm_connector *connector)
 {
@@ -7975,23 +8639,10 @@ static void amdgpu_dm_connector_add_common_modes(struct drm_encoder *encoder,
 				to_amdgpu_dm_connector(connector);
 	int i;
 	int n;
-	struct mode_size {
-		char name[DRM_DISPLAY_MODE_LEN];
-		int w;
-		int h;
-	} common_modes[] = {
-		{  "640x480",  640,  480},
-		{  "800x600",  800,  600},
-		{ "1024x768", 1024,  768},
-		{ "1280x720", 1280,  720},
-		{ "1280x800", 1280,  800},
-		{"1280x1024", 1280, 1024},
-		{ "1440x900", 1440,  900},
-		{"1680x1050", 1680, 1050},
-		{"1600x1200", 1600, 1200},
-		{"1920x1080", 1920, 1080},
-		{"1920x1200", 1920, 1200}
-	};
+
+	if ((connector->connector_type != DRM_MODE_CONNECTOR_eDP) &&
+	    (connector->connector_type != DRM_MODE_CONNECTOR_LVDS))
+		return;
 
 	n = ARRAY_SIZE(common_modes);
 
@@ -8019,6 +8670,9 @@ static void amdgpu_dm_connector_add_common_modes(struct drm_encoder *encoder,
 		mode = amdgpu_dm_create_common_mode(encoder,
 				common_modes[i].name, common_modes[i].w,
 				common_modes[i].h);
+		if (!mode)
+			continue;
+
 		drm_mode_probed_add(connector, mode);
 		amdgpu_dm_connector->num_modes++;
 	}
@@ -8034,6 +8688,10 @@ static void amdgpu_set_panel_orientation(struct drm_connector *connector)
 	    connector->connector_type != DRM_MODE_CONNECTOR_LVDS)
 		return;
 
+	mutex_lock(&connector->dev->mode_config.mutex);
+	amdgpu_dm_connector_get_modes(connector);
+	mutex_unlock(&connector->dev->mode_config.mutex);
+
 	encoder = amdgpu_dm_connector_to_encoder(connector);
 	if (!encoder)
 		return;
@@ -8051,16 +8709,16 @@ static void amdgpu_set_panel_orientation(struct drm_connector *connector)
 }
 
 static void amdgpu_dm_connector_ddc_get_modes(struct drm_connector *connector,
-					      struct edid *edid)
+					      const struct drm_edid *drm_edid)
 {
 	struct amdgpu_dm_connector *amdgpu_dm_connector =
 			to_amdgpu_dm_connector(connector);
 
-	if (edid) {
+	if (drm_edid) {
 		/* empty probed_modes */
 		INIT_LIST_HEAD(&connector->probed_modes);
 		amdgpu_dm_connector->num_modes =
-				drm_add_edid_modes(connector, edid);
+				drm_edid_connector_add_modes(connector);
 
 		/* sorting the probed modes before calling function
 		 * amdgpu_dm_get_native_mode() since EDID can have
@@ -8074,12 +8732,10 @@ static void amdgpu_dm_connector_ddc_get_modes(struct drm_connector *connector,
 		amdgpu_dm_get_native_mode(connector);
 
 		/* Freesync capabilities are reset by calling
-		 * drm_add_edid_modes() and need to be
+		 * drm_edid_connector_add_modes() and need to be
 		 * restored here.
 		 */
-		amdgpu_dm_update_freesync_caps(connector, edid);
-
-		amdgpu_set_panel_orientation(connector);
+		amdgpu_dm_update_freesync_caps(connector, drm_edid);
 	} else {
 		amdgpu_dm_connector->num_modes = 0;
 	}
@@ -8090,7 +8746,7 @@ static bool is_duplicate_mode(struct amdgpu_dm_connector *aconnector,
 {
 	struct drm_display_mode *m;
 
-	list_for_each_entry (m, &aconnector->base.probed_modes, head) {
+	list_for_each_entry(m, &aconnector->base.probed_modes, head) {
 		if (drm_mode_equal(m, mode))
 			return true;
 	}
@@ -8103,21 +8759,21 @@ static uint add_fs_modes(struct amdgpu_dm_connector *aconnector)
 	const struct drm_display_mode *m;
 	struct drm_display_mode *new_mode;
 	uint i;
-	uint32_t new_modes_count = 0;
+	u32 new_modes_count = 0;
 
 	/* Standard FPS values
 	 *
 	 * 23.976       - TV/NTSC
-	 * 24 	        - Cinema
-	 * 25 	        - TV/PAL
+	 * 24           - Cinema
+	 * 25           - TV/PAL
 	 * 29.97        - TV/NTSC
-	 * 30 	        - TV/NTSC
-	 * 48 	        - Cinema HFR
-	 * 50 	        - TV/PAL
-	 * 60 	        - Commonly used
+	 * 30           - TV/NTSC
+	 * 48           - Cinema HFR
+	 * 50           - TV/PAL
+	 * 60           - Commonly used
 	 * 48,72,96,120 - Multiples of 24
 	 */
-	static const uint32_t common_rates[] = {
+	static const u32 common_rates[] = {
 		23976, 24000, 25000, 29970, 30000,
 		48000, 50000, 60000, 72000, 96000, 120000
 	};
@@ -8133,8 +8789,8 @@ static uint add_fs_modes(struct amdgpu_dm_connector *aconnector)
 		return 0;
 
 	for (i = 0; i < ARRAY_SIZE(common_rates); i++) {
-		uint64_t target_vtotal, target_vtotal_diff;
-		uint64_t num, den;
+		u64 target_vtotal, target_vtotal_diff;
+		u64 num, den;
 
 		if (drm_mode_vrefresh(m) * 1000 < common_rates[i])
 			continue;
@@ -8175,12 +8831,22 @@ static uint add_fs_modes(struct amdgpu_dm_connector *aconnector)
 }
 
 static void amdgpu_dm_connector_add_freesync_modes(struct drm_connector *connector,
-						   struct edid *edid)
+						   const struct drm_edid *drm_edid)
 {
 	struct amdgpu_dm_connector *amdgpu_dm_connector =
 		to_amdgpu_dm_connector(connector);
 
-	if (!(amdgpu_freesync_vid_mode && edid))
+	if (!(amdgpu_freesync_vid_mode && drm_edid))
+		return;
+
+	if (!amdgpu_dm_connector->dc_sink || !amdgpu_dm_connector->dc_link)
+		return;
+
+	if (!dc_supports_vrr(amdgpu_dm_connector->dc_sink->ctx->dce_version))
+		return;
+
+	if (dc_connector_supports_analog(amdgpu_dm_connector->dc_link->link_id.id) &&
+	    amdgpu_dm_connector->dc_sink->edid_caps.analog)
 		return;
 
 	if (amdgpu_dm_connector->max_vfreq - amdgpu_dm_connector->min_vfreq > 10)
@@ -8192,24 +8858,48 @@ static int amdgpu_dm_connector_get_modes(struct drm_connector *connector)
 {
 	struct amdgpu_dm_connector *amdgpu_dm_connector =
 			to_amdgpu_dm_connector(connector);
+	struct dc_link *dc_link = amdgpu_dm_connector->dc_link;
 	struct drm_encoder *encoder;
-	struct edid *edid = amdgpu_dm_connector->edid;
+	const struct drm_edid *drm_edid = amdgpu_dm_connector->drm_edid;
+	struct dc_link_settings *verified_link_cap = &dc_link->verified_link_cap;
+	const struct dc *dc = dc_link->dc;
 
 	encoder = amdgpu_dm_connector_to_encoder(connector);
 
-	if (!drm_edid_is_valid(edid)) {
+	if (!drm_edid) {
 		amdgpu_dm_connector->num_modes =
 				drm_add_modes_noedid(connector, 640, 480);
+		if (dc->link_srv->dp_get_encoding_format(verified_link_cap) == DP_128b_132b_ENCODING)
+			amdgpu_dm_connector->num_modes +=
+				drm_add_modes_noedid(connector, 1920, 1080);
+
+		if (amdgpu_dm_connector->dc_sink &&
+		    amdgpu_dm_connector->dc_sink->edid_caps.analog &&
+		    dc_connector_supports_analog(dc_link->link_id.id)) {
+			/* Analog monitor connected by DAC load detection.
+			 * Add common modes. It will be up to the user to select one that works.
+			 */
+			for (int i = 0; i < ARRAY_SIZE(common_modes); i++)
+				amdgpu_dm_connector->num_modes += drm_add_modes_noedid(
+					connector, common_modes[i].w, common_modes[i].h);
+		}
 	} else {
-		amdgpu_dm_connector_ddc_get_modes(connector, edid);
-		amdgpu_dm_connector_add_common_modes(encoder, connector);
-		amdgpu_dm_connector_add_freesync_modes(connector, edid);
+		amdgpu_dm_connector_ddc_get_modes(connector, drm_edid);
+		if (encoder)
+			amdgpu_dm_connector_add_common_modes(encoder, connector);
+		amdgpu_dm_connector_add_freesync_modes(connector, drm_edid);
 	}
 	amdgpu_dm_fbc_init(connector);
 
 	return amdgpu_dm_connector->num_modes;
 }
 
+static const u32 supported_colorspaces =
+	BIT(DRM_MODE_COLORIMETRY_BT709_YCC) |
+	BIT(DRM_MODE_COLORIMETRY_OPRGB) |
+	BIT(DRM_MODE_COLORIMETRY_BT2020_RGB) |
+	BIT(DRM_MODE_COLORIMETRY_BT2020_YCC);
+
 void amdgpu_dm_connector_init_helper(struct amdgpu_display_manager *dm,
 				     struct amdgpu_dm_connector *aconnector,
 				     int connector_type,
@@ -8226,6 +8916,7 @@ void amdgpu_dm_connector_init_helper(struct amdgpu_display_manager *dm,
 		aconnector->base.funcs->reset(&aconnector->base);
 
 	aconnector->connector_id = link_index;
+	aconnector->bl_idx = -1;
 	aconnector->dc_link = link;
 	aconnector->base.interlace_allowed = false;
 	aconnector->base.doublescan_allowed = false;
@@ -8233,7 +8924,15 @@ void amdgpu_dm_connector_init_helper(struct amdgpu_display_manager *dm,
 	aconnector->base.dpms = DRM_MODE_DPMS_OFF;
 	aconnector->hpd.hpd = AMDGPU_HPD_NONE; /* not used */
 	aconnector->audio_inst = -1;
+	aconnector->pack_sdp_v1_3 = false;
+	aconnector->as_type = ADAPTIVE_SYNC_TYPE_NONE;
+	memset(&aconnector->vsdb_info, 0, sizeof(aconnector->vsdb_info));
 	mutex_init(&aconnector->hpd_lock);
+	mutex_init(&aconnector->handle_mst_msg_ready);
+
+	aconnector->hdmi_hpd_debounce_delay_ms = AMDGPU_DM_HDMI_HPD_DEBOUNCE_MS;
+	INIT_DELAYED_WORK(&aconnector->hdmi_hpd_debounce_work, hdmi_hpd_debounce_work);
+	aconnector->hdmi_prev_sink = NULL;
 
 	/*
 	 * configure support HPD hot plug connector_>polled default value is 0
@@ -8247,15 +8946,8 @@ void amdgpu_dm_connector_init_helper(struct amdgpu_display_manager *dm,
 		break;
 	case DRM_MODE_CONNECTOR_DisplayPort:
 		aconnector->base.polled = DRM_CONNECTOR_POLL_HPD;
-		if (link->is_dig_mapping_flexible &&
-		    link->dc->res_pool->funcs->link_encs_assign) {
-			link->link_enc =
-				link_enc_cfg_get_link_enc_used_by_link(link->ctx->dc, link);
-			if (!link->link_enc)
-				link->link_enc =
-					link_enc_cfg_get_next_avail_link_enc(link->ctx->dc);
-		}
-
+		link->link_enc = link_enc_cfg_get_link_enc(link);
+		ASSERT(link->link_enc);
 		if (link->link_enc)
 			aconnector->base.ycbcr_420_allowed =
 			link->link_enc->features.dp_ycbcr420_supported ? true : false;
@@ -8263,6 +8955,11 @@ void amdgpu_dm_connector_init_helper(struct amdgpu_display_manager *dm,
 	case DRM_MODE_CONNECTOR_DVID:
 		aconnector->base.polled = DRM_CONNECTOR_POLL_HPD;
 		break;
+	case DRM_MODE_CONNECTOR_DVII:
+	case DRM_MODE_CONNECTOR_VGA:
+		aconnector->base.polled =
+			DRM_CONNECTOR_POLL_CONNECT | DRM_CONNECTOR_POLL_DISCONNECT;
+		break;
 	default:
 		break;
 	}
@@ -8271,6 +8968,10 @@ void amdgpu_dm_connector_init_helper(struct amdgpu_display_manager *dm,
 				dm->ddev->mode_config.scaling_mode_property,
 				DRM_MODE_SCALE_NONE);
 
+	if (connector_type == DRM_MODE_CONNECTOR_HDMIA
+		|| (connector_type == DRM_MODE_CONNECTOR_DisplayPort && !aconnector->mst_root))
+		drm_connector_attach_broadcast_rgb_property(&aconnector->base);
+
 	drm_object_attach_property(&aconnector->base.base,
 				adev->mode_info.underscan_property,
 				UNDERSCAN_OFF);
@@ -8281,17 +8982,24 @@ void amdgpu_dm_connector_init_helper(struct amdgpu_display_manager *dm,
 				adev->mode_info.underscan_vborder_property,
 				0);
 
-	if (!aconnector->mst_port)
+	if (!aconnector->mst_root)
 		drm_connector_attach_max_bpc_property(&aconnector->base, 8, 16);
 
-	/* This defaults to the max in the range, but we want 8bpc for non-edp. */
-	aconnector->base.state->max_bpc = (connector_type == DRM_MODE_CONNECTOR_eDP) ? 16 : 8;
+	aconnector->base.state->max_bpc = 16;
 	aconnector->base.state->max_requested_bpc = aconnector->base.state->max_bpc;
 
-	if (connector_type == DRM_MODE_CONNECTOR_eDP &&
-	    (dc_is_dmcu_initialized(adev->dm.dc) || adev->dm.dc->ctx->dmub_srv)) {
-		drm_object_attach_property(&aconnector->base.base,
-				adev->mode_info.abm_level_property, 0);
+	if (connector_type == DRM_MODE_CONNECTOR_HDMIA) {
+		/* Content Type is currently only implemented for HDMI. */
+		drm_connector_attach_content_type_property(&aconnector->base);
+	}
+
+	if (connector_type == DRM_MODE_CONNECTOR_HDMIA) {
+		if (!drm_mode_create_hdmi_colorspace_property(&aconnector->base, supported_colorspaces))
+			drm_connector_attach_colorspace_property(&aconnector->base);
+	} else if ((connector_type == DRM_MODE_CONNECTOR_DisplayPort && !aconnector->mst_root) ||
+		   connector_type == DRM_MODE_CONNECTOR_eDP) {
+		if (!drm_mode_create_dp_colorspace_property(&aconnector->base, supported_colorspaces))
+			drm_connector_attach_colorspace_property(&aconnector->base);
 	}
 
 	if (connector_type == DRM_MODE_CONNECTOR_HDMIA ||
@@ -8299,13 +9007,23 @@ void amdgpu_dm_connector_init_helper(struct amdgpu_display_manager *dm,
 	    connector_type == DRM_MODE_CONNECTOR_eDP) {
 		drm_connector_attach_hdr_output_metadata_property(&aconnector->base);
 
-		if (!aconnector->mst_port)
+		if (!aconnector->mst_root)
 			drm_connector_attach_vrr_capable_property(&aconnector->base);
 
-#ifdef CONFIG_DRM_AMD_DC_HDCP
 		if (adev->dm.hdcp_workqueue)
 			drm_connector_attach_content_protection_property(&aconnector->base, true);
-#endif
+	}
+
+	if (connector_type == DRM_MODE_CONNECTOR_eDP) {
+		struct drm_privacy_screen *privacy_screen;
+
+		privacy_screen = drm_privacy_screen_get(adev_to_drm(adev)->dev, NULL);
+		if (!IS_ERR(privacy_screen)) {
+			drm_connector_attach_privacy_screen_provider(&aconnector->base,
+								     privacy_screen);
+		} else if (PTR_ERR(privacy_screen) != -ENODEV) {
+			drm_warn(adev_to_drm(adev), "Error getting privacy-screen\n");
+		}
 	}
 }
 
@@ -8318,6 +9036,9 @@ static int amdgpu_dm_i2c_xfer(struct i2c_adapter *i2c_adap,
 	int i;
 	int result = -EIO;
 
+	if (!ddc_service->ddc_pin)
+		return result;
+
 	cmd.payloads = kcalloc(num, sizeof(struct i2c_payload), GFP_KERNEL);
 
 	if (!cmd.payloads)
@@ -8334,11 +9055,18 @@ static int amdgpu_dm_i2c_xfer(struct i2c_adapter *i2c_adap,
 		cmd.payloads[i].data = msgs[i].buf;
 	}
 
-	if (dc_submit_i2c(
-			ddc_service->ctx->dc,
-			ddc_service->ddc_pin->hw_info.ddc_channel,
-			&cmd))
-		result = num;
+	if (i2c->oem) {
+		if (dc_submit_i2c_oem(
+			    ddc_service->ctx->dc,
+			    &cmd))
+			result = num;
+	} else {
+		if (dc_submit_i2c(
+			    ddc_service->ctx->dc,
+			    ddc_service->link->link_index,
+			    &cmd))
+			result = num;
+	}
 
 	kfree(cmd.payloads);
 	return result;
@@ -8355,9 +9083,7 @@ static const struct i2c_algorithm amdgpu_dm_i2c_algo = {
 };
 
 static struct amdgpu_i2c_adapter *
-create_i2c(struct ddc_service *ddc_service,
-	   int link_index,
-	   int *res)
+create_i2c(struct ddc_service *ddc_service, bool oem)
 {
 	struct amdgpu_device *adev = ddc_service->ctx->driver_context;
 	struct amdgpu_i2c_adapter *i2c;
@@ -8366,18 +9092,41 @@ create_i2c(struct ddc_service *ddc_service,
 	if (!i2c)
 		return NULL;
 	i2c->base.owner = THIS_MODULE;
-	i2c->base.class = I2C_CLASS_DDC;
 	i2c->base.dev.parent = &adev->pdev->dev;
 	i2c->base.algo = &amdgpu_dm_i2c_algo;
-	snprintf(i2c->base.name, sizeof(i2c->base.name), "AMDGPU DM i2c hw bus %d", link_index);
+	if (oem)
+		snprintf(i2c->base.name, sizeof(i2c->base.name), "AMDGPU DM i2c OEM bus");
+	else
+		snprintf(i2c->base.name, sizeof(i2c->base.name), "AMDGPU DM i2c hw bus %d",
+			 ddc_service->link->link_index);
 	i2c_set_adapdata(&i2c->base, i2c);
 	i2c->ddc_service = ddc_service;
-	if (i2c->ddc_service->ddc_pin)
-		i2c->ddc_service->ddc_pin->hw_info.ddc_channel = link_index;
+	i2c->oem = oem;
 
 	return i2c;
 }
 
+int amdgpu_dm_initialize_hdmi_connector(struct amdgpu_dm_connector *aconnector)
+{
+	struct cec_connector_info conn_info;
+	struct drm_device *ddev = aconnector->base.dev;
+	struct device *hdmi_dev = ddev->dev;
+
+	if (amdgpu_dc_debug_mask & DC_DISABLE_HDMI_CEC) {
+		drm_info(ddev, "HDMI-CEC feature masked\n");
+		return -EINVAL;
+	}
+
+	cec_fill_conn_info_from_drm(&conn_info, &aconnector->base);
+	aconnector->notifier =
+		cec_notifier_conn_register(hdmi_dev, NULL, &conn_info);
+	if (!aconnector->notifier) {
+		drm_err(ddev, "Failed to create cec notifier\n");
+		return -ENOMEM;
+	}
+
+	return 0;
+}
 
 /*
  * Note: this function assumes that dc_link_detect() was called for the
@@ -8385,7 +9134,7 @@ create_i2c(struct ddc_service *ddc_service,
  */
 static int amdgpu_dm_connector_init(struct amdgpu_display_manager *dm,
 				    struct amdgpu_dm_connector *aconnector,
-				    uint32_t link_index,
+				    u32 link_index,
 				    struct amdgpu_encoder *aencoder)
 {
 	int res = 0;
@@ -8394,25 +9143,25 @@ static int amdgpu_dm_connector_init(struct amdgpu_display_manager *dm,
 	struct dc_link *link = dc_get_link_at_index(dc, link_index);
 	struct amdgpu_i2c_adapter *i2c;
 
+	/* Not needed for writeback connector */
 	link->priv = aconnector;
 
-	DRM_DEBUG_DRIVER("%s()\n", __func__);
 
-	i2c = create_i2c(link->ddc, link->link_index, &res);
+	i2c = create_i2c(link->ddc, false);
 	if (!i2c) {
-		DRM_ERROR("Failed to create i2c adapter data\n");
+		drm_err(adev_to_drm(dm->adev), "Failed to create i2c adapter data\n");
 		return -ENOMEM;
 	}
 
 	aconnector->i2c = i2c;
-	res = i2c_add_adapter(&i2c->base);
+	res = devm_i2c_add_adapter(dm->adev->dev, &i2c->base);
 
 	if (res) {
-		DRM_ERROR("Failed to register hw i2c %d\n", link->link_index);
+		drm_err(adev_to_drm(dm->adev), "Failed to register hw i2c %d\n", link->link_index);
 		goto out_free;
 	}
 
-	connector_type = to_drm_connector_type(link->connector_signal);
+	connector_type = to_drm_connector_type(link->connector_signal, link->link_id.id);
 
 	res = drm_connector_init_with_ddc(
 			dm->ddev,
@@ -8422,7 +9171,7 @@ static int amdgpu_dm_connector_init(struct amdgpu_display_manager *dm,
 			&i2c->base);
 
 	if (res) {
-		DRM_ERROR("connector_init failed\n");
+		drm_err(adev_to_drm(dm->adev), "connector_init failed\n");
 		aconnector->connector_id = -1;
 		goto out_free;
 	}
@@ -8441,6 +9190,10 @@ static int amdgpu_dm_connector_init(struct amdgpu_display_manager *dm,
 	drm_connector_attach_encoder(
 		&aconnector->base, &aencoder->base);
 
+	if (connector_type == DRM_MODE_CONNECTOR_HDMIA ||
+	    connector_type == DRM_MODE_CONNECTOR_HDMIB)
+		amdgpu_dm_initialize_hdmi_connector(aconnector);
+
 	if (connector_type == DRM_MODE_CONNECTOR_DisplayPort
 		|| connector_type == DRM_MODE_CONNECTOR_eDP)
 		amdgpu_dm_initialize_dp_connector(dm, aconnector, link->link_index);
@@ -8498,42 +9251,98 @@ static int amdgpu_dm_encoder_init(struct drm_device *dev,
 
 static void manage_dm_interrupts(struct amdgpu_device *adev,
 				 struct amdgpu_crtc *acrtc,
-				 bool enable)
-{
-	/*
-	 * We have no guarantee that the frontend index maps to the same
+				 struct dm_crtc_state *acrtc_state)
+{	/*
+	 * We cannot be sure that the frontend index maps to the same
 	 * backend index - some even map to more than one.
-	 *
-	 * TODO: Use a different interrupt or check DC itself for the mapping.
+	 * So we have to go through the CRTC to find the right IRQ.
 	 */
-	int irq_type =
-		amdgpu_display_crtc_idx_to_irq_type(
+	int irq_type = amdgpu_display_crtc_idx_to_irq_type(
 			adev,
 			acrtc->crtc_id);
+	struct drm_device *dev = adev_to_drm(adev);
 
-	if (enable) {
-		drm_crtc_vblank_on(&acrtc->base);
-		amdgpu_irq_get(
-			adev,
-			&adev->pageflip_irq,
-			irq_type);
+	struct drm_vblank_crtc_config config = {0};
+	struct dc_crtc_timing *timing;
+	int offdelay;
+
+	if (acrtc_state) {
+		timing = &acrtc_state->stream->timing;
+
+		/*
+		 * Depending on when the HW latching event of double-buffered
+		 * registers happen relative to the PSR SDP deadline, and how
+		 * bad the Panel clock has drifted since the last ALPM off
+		 * event, there can be up to 3 frames of delay between sending
+		 * the PSR exit cmd to DMUB fw, and when the panel starts
+		 * displaying live frames.
+		 *
+		 * We can set:
+		 *
+		 * 20/100 * offdelay_ms = 3_frames_ms
+		 * => offdelay_ms = 5 * 3_frames_ms
+		 *
+		 * This ensures that `3_frames_ms` will only be experienced as a
+		 * 20% delay on top how long the display has been static, and
+		 * thus make the delay less perceivable.
+		 */
+		if (acrtc_state->stream->link->psr_settings.psr_version <
+		    DC_PSR_VERSION_UNSUPPORTED) {
+			offdelay = DIV64_U64_ROUND_UP((u64)5 * 3 * 10 *
+						      timing->v_total *
+						      timing->h_total,
+						      timing->pix_clk_100hz);
+			config.offdelay_ms = offdelay ?: 30;
+		} else if (amdgpu_ip_version(adev, DCE_HWIP, 0) <
+			   IP_VERSION(3, 5, 0) ||
+			   !(adev->flags & AMD_IS_APU)) {
+			/*
+			 * Older HW and DGPU have issues with instant off;
+			 * use a 2 frame offdelay.
+			 */
+			offdelay = DIV64_U64_ROUND_UP((u64)20 *
+						      timing->v_total *
+						      timing->h_total,
+						      timing->pix_clk_100hz);
+
+			config.offdelay_ms = offdelay ?: 30;
+		} else {
+			/* offdelay_ms = 0 will never disable vblank */
+			config.offdelay_ms = 1;
+			config.disable_immediate = true;
+		}
+
+		drm_crtc_vblank_on_config(&acrtc->base,
+					  &config);
+		/* Allow RX6xxx, RX7700, RX7800 GPUs to call amdgpu_irq_get.*/
+		switch (amdgpu_ip_version(adev, DCE_HWIP, 0)) {
+		case IP_VERSION(3, 0, 0):
+		case IP_VERSION(3, 0, 2):
+		case IP_VERSION(3, 0, 3):
+		case IP_VERSION(3, 2, 0):
+			if (amdgpu_irq_get(adev, &adev->pageflip_irq, irq_type))
+				drm_err(dev, "DM_IRQ: Cannot get pageflip irq!\n");
 #if defined(CONFIG_DRM_AMD_SECURE_DISPLAY)
-		amdgpu_irq_get(
-			adev,
-			&adev->vline0_irq,
-			irq_type);
+			if (amdgpu_irq_get(adev, &adev->vline0_irq, irq_type))
+				drm_err(dev, "DM_IRQ: Cannot get vline0 irq!\n");
 #endif
+		}
+
 	} else {
+		/* Allow RX6xxx, RX7700, RX7800 GPUs to call amdgpu_irq_put.*/
+		switch (amdgpu_ip_version(adev, DCE_HWIP, 0)) {
+		case IP_VERSION(3, 0, 0):
+		case IP_VERSION(3, 0, 2):
+		case IP_VERSION(3, 0, 3):
+		case IP_VERSION(3, 2, 0):
 #if defined(CONFIG_DRM_AMD_SECURE_DISPLAY)
-		amdgpu_irq_put(
-			adev,
-			&adev->vline0_irq,
-			irq_type);
+			if (amdgpu_irq_put(adev, &adev->vline0_irq, irq_type))
+				drm_err(dev, "DM_IRQ: Cannot put vline0 irq!\n");
 #endif
-		amdgpu_irq_put(
-			adev,
-			&adev->pageflip_irq,
-			irq_type);
+			if (amdgpu_irq_put(adev, &adev->pageflip_irq, irq_type))
+				drm_err(dev, "DM_IRQ: Cannot put pageflip irq!\n");
+		}
+
 		drm_crtc_vblank_off(&acrtc->base);
 	}
 }
@@ -8569,28 +9378,55 @@ is_scaling_state_different(const struct dm_connector_state *dm_state,
 	return false;
 }
 
-#ifdef CONFIG_DRM_AMD_DC_HDCP
-static bool is_content_protection_different(struct drm_connector_state *state,
-					    const struct drm_connector_state *old_state,
-					    const struct drm_connector *connector, struct hdcp_workqueue *hdcp_w)
+static bool is_content_protection_different(struct drm_crtc_state *new_crtc_state,
+					    struct drm_crtc_state *old_crtc_state,
+					    struct drm_connector_state *new_conn_state,
+					    struct drm_connector_state *old_conn_state,
+					    const struct drm_connector *connector,
+					    struct hdcp_workqueue *hdcp_w)
 {
 	struct amdgpu_dm_connector *aconnector = to_amdgpu_dm_connector(connector);
 	struct dm_connector_state *dm_con_state = to_dm_connector_state(connector->state);
 
-	/* Handle: Type0/1 change */
-	if (old_state->hdcp_content_type != state->hdcp_content_type &&
-	    state->content_protection != DRM_MODE_CONTENT_PROTECTION_UNDESIRED) {
-		state->content_protection = DRM_MODE_CONTENT_PROTECTION_DESIRED;
+	pr_debug("[HDCP_DM] connector->index: %x connect_status: %x dpms: %x\n",
+		connector->index, connector->status, connector->dpms);
+	pr_debug("[HDCP_DM] state protection old: %x new: %x\n",
+		old_conn_state->content_protection, new_conn_state->content_protection);
+
+	if (old_crtc_state)
+		pr_debug("[HDCP_DM] old crtc en: %x a: %x m: %x a-chg: %x c-chg: %x\n",
+		old_crtc_state->enable,
+		old_crtc_state->active,
+		old_crtc_state->mode_changed,
+		old_crtc_state->active_changed,
+		old_crtc_state->connectors_changed);
+
+	if (new_crtc_state)
+		pr_debug("[HDCP_DM] NEW crtc en: %x a: %x m: %x a-chg: %x c-chg: %x\n",
+		new_crtc_state->enable,
+		new_crtc_state->active,
+		new_crtc_state->mode_changed,
+		new_crtc_state->active_changed,
+		new_crtc_state->connectors_changed);
+
+	/* hdcp content type change */
+	if (old_conn_state->hdcp_content_type != new_conn_state->hdcp_content_type &&
+	    new_conn_state->content_protection != DRM_MODE_CONTENT_PROTECTION_UNDESIRED) {
+		new_conn_state->content_protection = DRM_MODE_CONTENT_PROTECTION_DESIRED;
+		pr_debug("[HDCP_DM] Type0/1 change %s :true\n", __func__);
 		return true;
 	}
 
-	/* CP is being re enabled, ignore this
-	 *
-	 * Handles:	ENABLED -> DESIRED
-	 */
-	if (old_state->content_protection == DRM_MODE_CONTENT_PROTECTION_ENABLED &&
-	    state->content_protection == DRM_MODE_CONTENT_PROTECTION_DESIRED) {
-		state->content_protection = DRM_MODE_CONTENT_PROTECTION_ENABLED;
+	/* CP is being re enabled, ignore this */
+	if (old_conn_state->content_protection == DRM_MODE_CONTENT_PROTECTION_ENABLED &&
+	    new_conn_state->content_protection == DRM_MODE_CONTENT_PROTECTION_DESIRED) {
+		if (new_crtc_state && new_crtc_state->mode_changed) {
+			new_conn_state->content_protection = DRM_MODE_CONTENT_PROTECTION_DESIRED;
+			pr_debug("[HDCP_DM] ENABLED->DESIRED & mode_changed %s :true\n", __func__);
+			return true;
+		}
+		new_conn_state->content_protection = DRM_MODE_CONTENT_PROTECTION_ENABLED;
+		pr_debug("[HDCP_DM] ENABLED -> DESIRED %s :false\n", __func__);
 		return false;
 	}
 
@@ -8598,9 +9434,9 @@ static bool is_content_protection_different(struct drm_connector_state *state,
 	 *
 	 * Handles:	UNDESIRED -> ENABLED
 	 */
-	if (old_state->content_protection == DRM_MODE_CONTENT_PROTECTION_UNDESIRED &&
-	    state->content_protection == DRM_MODE_CONTENT_PROTECTION_ENABLED)
-		state->content_protection = DRM_MODE_CONTENT_PROTECTION_DESIRED;
+	if (old_conn_state->content_protection == DRM_MODE_CONTENT_PROTECTION_UNDESIRED &&
+	    new_conn_state->content_protection == DRM_MODE_CONTENT_PROTECTION_ENABLED)
+		new_conn_state->content_protection = DRM_MODE_CONTENT_PROTECTION_DESIRED;
 
 	/* Stream removed and re-enabled
 	 *
@@ -8610,10 +9446,12 @@ static bool is_content_protection_different(struct drm_connector_state *state,
 	 *
 	 * Handles:	DESIRED -> DESIRED (Special case)
 	 */
-	if (!(old_state->crtc && old_state->crtc->enabled) &&
-		state->crtc && state->crtc->enabled &&
+	if (!(old_conn_state->crtc && old_conn_state->crtc->enabled) &&
+		new_conn_state->crtc && new_conn_state->crtc->enabled &&
 		connector->state->content_protection == DRM_MODE_CONTENT_PROTECTION_DESIRED) {
 		dm_con_state->update_hdcp = false;
+		pr_debug("[HDCP_DM] DESIRED->DESIRED (Stream removed and re-enabled) %s :true\n",
+			__func__);
 		return true;
 	}
 
@@ -8625,35 +9463,41 @@ static bool is_content_protection_different(struct drm_connector_state *state,
 	 *
 	 * Handles:	DESIRED -> DESIRED (Special case)
 	 */
-	if (dm_con_state->update_hdcp && state->content_protection == DRM_MODE_CONTENT_PROTECTION_DESIRED &&
-	    connector->dpms == DRM_MODE_DPMS_ON && aconnector->dc_sink != NULL) {
+	if (dm_con_state->update_hdcp &&
+	new_conn_state->content_protection == DRM_MODE_CONTENT_PROTECTION_DESIRED &&
+	connector->dpms == DRM_MODE_DPMS_ON && aconnector->dc_sink != NULL) {
 		dm_con_state->update_hdcp = false;
+		pr_debug("[HDCP_DM] DESIRED->DESIRED (Hot-plug, headless s3, dpms) %s :true\n",
+			__func__);
 		return true;
 	}
 
-	/*
-	 * Handles:	UNDESIRED -> UNDESIRED
-	 *		DESIRED -> DESIRED
-	 *		ENABLED -> ENABLED
-	 */
-	if (old_state->content_protection == state->content_protection)
+	if (old_conn_state->content_protection == new_conn_state->content_protection) {
+		if (new_conn_state->content_protection >= DRM_MODE_CONTENT_PROTECTION_DESIRED) {
+			if (new_crtc_state && new_crtc_state->mode_changed) {
+				pr_debug("[HDCP_DM] DESIRED->DESIRED or ENABLE->ENABLE mode_change %s :true\n",
+					__func__);
+				return true;
+			}
+			pr_debug("[HDCP_DM] DESIRED->DESIRED & ENABLE->ENABLE %s :false\n",
+				__func__);
+			return false;
+		}
+
+		pr_debug("[HDCP_DM] UNDESIRED->UNDESIRED %s :false\n", __func__);
 		return false;
+	}
 
-	/*
-	 * Handles:	UNDESIRED -> DESIRED
-	 *		DESIRED -> UNDESIRED
-	 *		ENABLED -> UNDESIRED
-	 */
-	if (state->content_protection != DRM_MODE_CONTENT_PROTECTION_ENABLED)
+	if (new_conn_state->content_protection != DRM_MODE_CONTENT_PROTECTION_ENABLED) {
+		pr_debug("[HDCP_DM] UNDESIRED->DESIRED or DESIRED->UNDESIRED or ENABLED->UNDESIRED %s :true\n",
+			__func__);
 		return true;
+	}
 
-	/*
-	 * Handles:	DESIRED -> ENABLED
-	 */
+	pr_debug("[HDCP_DM] DESIRED->ENABLED %s :false\n", __func__);
 	return false;
 }
 
-#endif
 static void remove_stream(struct amdgpu_device *adev,
 			  struct amdgpu_crtc *acrtc,
 			  struct dc_stream_state *stream)
@@ -8664,114 +9508,6 @@ static void remove_stream(struct amdgpu_device *adev,
 	acrtc->enabled = false;
 }
 
-static int get_cursor_position(struct drm_plane *plane, struct drm_crtc *crtc,
-			       struct dc_cursor_position *position)
-{
-	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
-	int x, y;
-	int xorigin = 0, yorigin = 0;
-
-	if (!crtc || !plane->state->fb)
-		return 0;
-
-	if ((plane->state->crtc_w > amdgpu_crtc->max_cursor_width) ||
-	    (plane->state->crtc_h > amdgpu_crtc->max_cursor_height)) {
-		DRM_ERROR("%s: bad cursor width or height %d x %d\n",
-			  __func__,
-			  plane->state->crtc_w,
-			  plane->state->crtc_h);
-		return -EINVAL;
-	}
-
-	x = plane->state->crtc_x;
-	y = plane->state->crtc_y;
-
-	if (x <= -amdgpu_crtc->max_cursor_width ||
-	    y <= -amdgpu_crtc->max_cursor_height)
-		return 0;
-
-	if (x < 0) {
-		xorigin = min(-x, amdgpu_crtc->max_cursor_width - 1);
-		x = 0;
-	}
-	if (y < 0) {
-		yorigin = min(-y, amdgpu_crtc->max_cursor_height - 1);
-		y = 0;
-	}
-	position->enable = true;
-	position->translate_by_source = true;
-	position->x = x;
-	position->y = y;
-	position->x_hotspot = xorigin;
-	position->y_hotspot = yorigin;
-
-	return 0;
-}
-
-static void handle_cursor_update(struct drm_plane *plane,
-				 struct drm_plane_state *old_plane_state)
-{
-	struct amdgpu_device *adev = drm_to_adev(plane->dev);
-	struct amdgpu_framebuffer *afb = to_amdgpu_framebuffer(plane->state->fb);
-	struct drm_crtc *crtc = afb ? plane->state->crtc : old_plane_state->crtc;
-	struct dm_crtc_state *crtc_state = crtc ? to_dm_crtc_state(crtc->state) : NULL;
-	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
-	uint64_t address = afb ? afb->address : 0;
-	struct dc_cursor_position position = {0};
-	struct dc_cursor_attributes attributes;
-	int ret;
-
-	if (!plane->state->fb && !old_plane_state->fb)
-		return;
-
-	DC_LOG_CURSOR("%s: crtc_id=%d with size %d to %d\n",
-		      __func__,
-		      amdgpu_crtc->crtc_id,
-		      plane->state->crtc_w,
-		      plane->state->crtc_h);
-
-	ret = get_cursor_position(plane, crtc, &position);
-	if (ret)
-		return;
-
-	if (!position.enable) {
-		/* turn off cursor */
-		if (crtc_state && crtc_state->stream) {
-			mutex_lock(&adev->dm.dc_lock);
-			dc_stream_set_cursor_position(crtc_state->stream,
-						      &position);
-			mutex_unlock(&adev->dm.dc_lock);
-		}
-		return;
-	}
-
-	amdgpu_crtc->cursor_width = plane->state->crtc_w;
-	amdgpu_crtc->cursor_height = plane->state->crtc_h;
-
-	memset(&attributes, 0, sizeof(attributes));
-	attributes.address.high_part = upper_32_bits(address);
-	attributes.address.low_part  = lower_32_bits(address);
-	attributes.width             = plane->state->crtc_w;
-	attributes.height            = plane->state->crtc_h;
-	attributes.color_format      = CURSOR_MODE_COLOR_PRE_MULTIPLIED_ALPHA;
-	attributes.rotation_angle    = 0;
-	attributes.attribute_flags.value = 0;
-
-	attributes.pitch = afb->base.pitches[0] / afb->base.format->cpp[0];
-
-	if (crtc_state->stream) {
-		mutex_lock(&adev->dm.dc_lock);
-		if (!dc_stream_set_cursor_attributes(crtc_state->stream,
-							 &attributes))
-			DRM_ERROR("DC failed to set cursor attributes\n");
-
-		if (!dc_stream_set_cursor_position(crtc_state->stream,
-						   &position))
-			DRM_ERROR("DC failed to set cursor position\n");
-		mutex_unlock(&adev->dm.dc_lock);
-	}
-}
-
 static void prepare_flip_isr(struct amdgpu_crtc *acrtc)
 {
 
@@ -8786,8 +9522,9 @@ static void prepare_flip_isr(struct amdgpu_crtc *acrtc)
 	/* Mark this event as consumed */
 	acrtc->base.state->event = NULL;
 
-	DC_LOG_PFLIP("crtc:%d, pflip_stat:AMDGPU_FLIP_SUBMITTED\n",
-		     acrtc->crtc_id);
+	drm_dbg_state(acrtc->base.dev,
+		      "crtc:%d, pflip_stat:AMDGPU_FLIP_SUBMITTED\n",
+		      acrtc->crtc_id);
 }
 
 static void update_freesync_state_on_stream(
@@ -8803,6 +9540,8 @@ static void update_freesync_state_on_stream(
 	struct amdgpu_crtc *acrtc = to_amdgpu_crtc(new_crtc_state->base.crtc);
 	unsigned long flags;
 	bool pack_sdp_v1_3 = false;
+	struct amdgpu_dm_connector *aconn;
+	enum vrr_packet_type packet_type = PACKET_TYPE_VRR;
 
 	if (!new_stream)
 		return;
@@ -8816,7 +9555,7 @@ static void update_freesync_state_on_stream(
 		return;
 
 	spin_lock_irqsave(&adev_to_drm(adev)->event_lock, flags);
-        vrr_params = acrtc->dm_irq_params.vrr_params;
+	vrr_params = acrtc->dm_irq_params.vrr_params;
 
 	if (surface) {
 		mod_freesync_handle_preflip(
@@ -8827,7 +9566,7 @@ static void update_freesync_state_on_stream(
 			&vrr_params);
 
 		if (adev->family < AMDGPU_FAMILY_AI &&
-		    amdgpu_dm_vrr_active(new_crtc_state)) {
+		    amdgpu_dm_crtc_vrr_active(new_crtc_state)) {
 			mod_freesync_handle_v_update(dm->freesync_module,
 						     new_stream, &vrr_params);
 
@@ -8838,20 +9577,31 @@ static void update_freesync_state_on_stream(
 		}
 	}
 
+	aconn = (struct amdgpu_dm_connector *)new_stream->dm_stream_context;
+
+	if (aconn && (aconn->as_type == FREESYNC_TYPE_PCON_IN_WHITELIST || aconn->vsdb_info.replay_mode)) {
+		pack_sdp_v1_3 = aconn->pack_sdp_v1_3;
+
+		if (aconn->vsdb_info.amd_vsdb_version == 1)
+			packet_type = PACKET_TYPE_FS_V1;
+		else if (aconn->vsdb_info.amd_vsdb_version == 2)
+			packet_type = PACKET_TYPE_FS_V2;
+		else if (aconn->vsdb_info.amd_vsdb_version == 3)
+			packet_type = PACKET_TYPE_FS_V3;
+
+		mod_build_adaptive_sync_infopacket(new_stream, aconn->as_type, NULL,
+					&new_stream->adaptive_sync_infopacket);
+	}
+
 	mod_freesync_build_vrr_infopacket(
 		dm->freesync_module,
 		new_stream,
 		&vrr_params,
-		PACKET_TYPE_VRR,
+		packet_type,
 		TRANSFER_FUNC_UNKNOWN,
 		&vrr_infopacket,
 		pack_sdp_v1_3);
 
-	new_crtc_state->freesync_timing_changed |=
-		(memcmp(&acrtc->dm_irq_params.vrr_params.adjust,
-			&vrr_params.adjust,
-			sizeof(vrr_params.adjust)) != 0);
-
 	new_crtc_state->freesync_vrr_info_changed |=
 		(memcmp(&new_crtc_state->vrr_infopacket,
 			&vrr_infopacket,
@@ -8860,8 +9610,8 @@ static void update_freesync_state_on_stream(
 	acrtc->dm_irq_params.vrr_params = vrr_params;
 	new_crtc_state->vrr_infopacket = vrr_infopacket;
 
-	new_stream->adjust = acrtc->dm_irq_params.vrr_params.adjust;
 	new_stream->vrr_infopacket = vrr_infopacket;
+	new_stream->allow_freesync = mod_freesync_get_freesync_enabled(&vrr_params);
 
 	if (new_crtc_state->freesync_vrr_info_changed)
 		DRM_DEBUG_KMS("VRR packet update: crtc=%u enabled=%d state=%d",
@@ -8923,10 +9673,6 @@ static void update_stream_irq_parameters(
 				      new_stream,
 				      &config, &vrr_params);
 
-	new_crtc_state->freesync_timing_changed |=
-		(memcmp(&acrtc->dm_irq_params.vrr_params.adjust,
-			&vrr_params.adjust, sizeof(vrr_params.adjust)) != 0);
-
 	new_crtc_state->freesync_config = config;
 	/* Copy state for access from DM IRQ handler */
 	acrtc->dm_irq_params.freesync_config = config;
@@ -8938,8 +9684,8 @@ static void update_stream_irq_parameters(
 static void amdgpu_dm_handle_vrr_transition(struct dm_crtc_state *old_state,
 					    struct dm_crtc_state *new_state)
 {
-	bool old_vrr_active = amdgpu_dm_vrr_active(old_state);
-	bool new_vrr_active = amdgpu_dm_vrr_active(new_state);
+	bool old_vrr_active = amdgpu_dm_crtc_vrr_active(old_state);
+	bool new_vrr_active = amdgpu_dm_crtc_vrr_active(new_state);
 
 	if (!old_vrr_active && new_vrr_active) {
 		/* Transition VRR inactive -> active:
@@ -8950,17 +9696,17 @@ static void amdgpu_dm_handle_vrr_transition(struct dm_crtc_state *old_state,
 		 * We also need vupdate irq for the actual core vblank handling
 		 * at end of vblank.
 		 */
-		dm_set_vupdate_irq(new_state->base.crtc, true);
-		drm_crtc_vblank_get(new_state->base.crtc);
-		DRM_DEBUG_DRIVER("%s: crtc=%u VRR off->on: Get vblank ref\n",
+		WARN_ON(amdgpu_dm_crtc_set_vupdate_irq(new_state->base.crtc, true) != 0);
+		WARN_ON(drm_crtc_vblank_get(new_state->base.crtc) != 0);
+		drm_dbg_driver(new_state->base.crtc->dev, "%s: crtc=%u VRR off->on: Get vblank ref\n",
 				 __func__, new_state->base.crtc->base.id);
 	} else if (old_vrr_active && !new_vrr_active) {
 		/* Transition VRR active -> inactive:
 		 * Allow vblank irq disable again for fixed refresh rate.
 		 */
-		dm_set_vupdate_irq(new_state->base.crtc, false);
+		WARN_ON(amdgpu_dm_crtc_set_vupdate_irq(new_state->base.crtc, false) != 0);
 		drm_crtc_vblank_put(new_state->base.crtc);
-		DRM_DEBUG_DRIVER("%s: crtc=%u VRR on->off: Drop vblank ref\n",
+		drm_dbg_driver(new_state->base.crtc->dev, "%s: crtc=%u VRR on->off: Drop vblank ref\n",
 				 __func__, new_state->base.crtc->base.id);
 	}
 }
@@ -8977,18 +9723,148 @@ static void amdgpu_dm_commit_cursors(struct drm_atomic_state *state)
 	 */
 	for_each_old_plane_in_state(state, plane, old_plane_state, i)
 		if (plane->type == DRM_PLANE_TYPE_CURSOR)
-			handle_cursor_update(plane, old_plane_state);
+			amdgpu_dm_plane_handle_cursor_update(plane, old_plane_state);
+}
+
+static inline uint32_t get_mem_type(struct drm_framebuffer *fb)
+{
+	struct amdgpu_bo *abo = gem_to_amdgpu_bo(fb->obj[0]);
+
+	return abo->tbo.resource ? abo->tbo.resource->mem_type : 0;
+}
+
+static void amdgpu_dm_update_cursor(struct drm_plane *plane,
+				    struct drm_plane_state *old_plane_state,
+				    struct dc_stream_update *update)
+{
+	struct amdgpu_device *adev = drm_to_adev(plane->dev);
+	struct amdgpu_framebuffer *afb = to_amdgpu_framebuffer(plane->state->fb);
+	struct drm_crtc *crtc = afb ? plane->state->crtc : old_plane_state->crtc;
+	struct dm_crtc_state *crtc_state = crtc ? to_dm_crtc_state(crtc->state) : NULL;
+	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
+	uint64_t address = afb ? afb->address : 0;
+	struct dc_cursor_position position = {0};
+	struct dc_cursor_attributes attributes;
+	int ret;
+
+	if (!plane->state->fb && !old_plane_state->fb)
+		return;
+
+	drm_dbg_atomic(plane->dev, "crtc_id=%d with size %d to %d\n",
+		       amdgpu_crtc->crtc_id, plane->state->crtc_w,
+		       plane->state->crtc_h);
+
+	ret = amdgpu_dm_plane_get_cursor_position(plane, crtc, &position);
+	if (ret)
+		return;
+
+	if (!position.enable) {
+		/* turn off cursor */
+		if (crtc_state && crtc_state->stream) {
+			dc_stream_set_cursor_position(crtc_state->stream,
+						      &position);
+			update->cursor_position = &crtc_state->stream->cursor_position;
+		}
+		return;
+	}
+
+	amdgpu_crtc->cursor_width = plane->state->crtc_w;
+	amdgpu_crtc->cursor_height = plane->state->crtc_h;
+
+	memset(&attributes, 0, sizeof(attributes));
+	attributes.address.high_part = upper_32_bits(address);
+	attributes.address.low_part  = lower_32_bits(address);
+	attributes.width             = plane->state->crtc_w;
+	attributes.height            = plane->state->crtc_h;
+	attributes.color_format      = CURSOR_MODE_COLOR_PRE_MULTIPLIED_ALPHA;
+	attributes.rotation_angle    = 0;
+	attributes.attribute_flags.value = 0;
+
+	/* Enable cursor degamma ROM on DCN3+ for implicit sRGB degamma in DRM
+	 * legacy gamma setup.
+	 */
+	if (crtc_state->cm_is_degamma_srgb &&
+	    adev->dm.dc->caps.color.dpp.gamma_corr)
+		attributes.attribute_flags.bits.ENABLE_CURSOR_DEGAMMA = 1;
+
+	if (afb)
+		attributes.pitch = afb->base.pitches[0] / afb->base.format->cpp[0];
+
+	if (crtc_state->stream) {
+		if (!dc_stream_set_cursor_attributes(crtc_state->stream,
+						     &attributes))
+			drm_err(adev_to_drm(adev), "DC failed to set cursor attributes\n");
+
+		update->cursor_attributes = &crtc_state->stream->cursor_attributes;
+
+		if (!dc_stream_set_cursor_position(crtc_state->stream,
+						   &position))
+			drm_err(adev_to_drm(adev), "DC failed to set cursor position\n");
+
+		update->cursor_position = &crtc_state->stream->cursor_position;
+	}
+}
+
+static void amdgpu_dm_enable_self_refresh(struct amdgpu_crtc *acrtc_attach,
+					  const struct dm_crtc_state *acrtc_state,
+					  const u64 current_ts)
+{
+	struct psr_settings *psr = &acrtc_state->stream->link->psr_settings;
+	struct replay_settings *pr = &acrtc_state->stream->link->replay_settings;
+	struct amdgpu_dm_connector *aconn =
+		(struct amdgpu_dm_connector *)acrtc_state->stream->dm_stream_context;
+	bool vrr_active = amdgpu_dm_crtc_vrr_active(acrtc_state);
+
+	if (acrtc_state->update_type > UPDATE_TYPE_FAST) {
+		if (pr->config.replay_supported && !pr->replay_feature_enabled)
+			amdgpu_dm_link_setup_replay(acrtc_state->stream->link, aconn);
+		else if (psr->psr_version != DC_PSR_VERSION_UNSUPPORTED &&
+			     !psr->psr_feature_enabled)
+			if (!aconn->disallow_edp_enter_psr)
+				amdgpu_dm_link_setup_psr(acrtc_state->stream);
+	}
+
+	/* Decrement skip count when SR is enabled and we're doing fast updates. */
+	if (acrtc_state->update_type == UPDATE_TYPE_FAST &&
+	    (psr->psr_feature_enabled || pr->config.replay_supported)) {
+		if (aconn->sr_skip_count > 0)
+			aconn->sr_skip_count--;
+
+		/* Allow SR when skip count is 0. */
+		acrtc_attach->dm_irq_params.allow_sr_entry = !aconn->sr_skip_count;
+
+		/*
+		 * If sink supports PSR SU/Panel Replay, there is no need to rely on
+		 * a vblank event disable request to enable PSR/RP. PSR SU/RP
+		 * can be enabled immediately once OS demonstrates an
+		 * adequate number of fast atomic commits to notify KMD
+		 * of update events. See `vblank_control_worker()`.
+		 */
+		if (!vrr_active &&
+		    acrtc_attach->dm_irq_params.allow_sr_entry &&
+#ifdef CONFIG_DRM_AMD_SECURE_DISPLAY
+		    !amdgpu_dm_crc_window_is_activated(acrtc_state->base.crtc) &&
+#endif
+		    (current_ts - psr->psr_dirty_rects_change_timestamp_ns) > 500000000) {
+			if (pr->replay_feature_enabled && !pr->replay_allow_active)
+				amdgpu_dm_replay_enable(acrtc_state->stream, true);
+			if (psr->psr_version == DC_PSR_VERSION_SU_1 &&
+			    !psr->psr_allow_active && !aconn->disallow_edp_enter_psr)
+				amdgpu_dm_psr_enable(acrtc_state->stream);
+		}
+	} else {
+		acrtc_attach->dm_irq_params.allow_sr_entry = false;
+	}
 }
 
 static void amdgpu_dm_commit_planes(struct drm_atomic_state *state,
-				    struct dc_state *dc_state,
 				    struct drm_device *dev,
 				    struct amdgpu_display_manager *dm,
 				    struct drm_crtc *pcrtc,
 				    bool wait_for_vblank)
 {
-	uint32_t i;
-	uint64_t timestamp_ns;
+	u32 i;
+	u64 timestamp_ns = ktime_get_ns();
 	struct drm_plane *plane;
 	struct drm_plane_state *old_plane_state, *new_plane_state;
 	struct amdgpu_crtc *acrtc_attach = to_amdgpu_crtc(pcrtc);
@@ -8998,12 +9874,13 @@ static void amdgpu_dm_commit_planes(struct drm_atomic_state *state,
 	struct dm_crtc_state *dm_old_crtc_state =
 			to_dm_crtc_state(drm_atomic_get_old_crtc_state(state, pcrtc));
 	int planes_count = 0, vpos, hpos;
-	long r;
 	unsigned long flags;
-	struct amdgpu_bo *abo;
-	uint32_t target_vblank, last_flip_vblank;
-	bool vrr_active = amdgpu_dm_vrr_active(acrtc_state);
+	u32 target_vblank, last_flip_vblank;
+	bool vrr_active = amdgpu_dm_crtc_vrr_active(acrtc_state);
+	bool cursor_update = false;
 	bool pflip_present = false;
+	bool dirty_rects_changed = false;
+	bool updated_planes_and_streams = false;
 	struct {
 		struct dc_surface_update surface_updates[MAX_SURFACES];
 		struct dc_plane_info plane_infos[MAX_SURFACES];
@@ -9015,7 +9892,7 @@ static void amdgpu_dm_commit_planes(struct drm_atomic_state *state,
 	bundle = kzalloc(sizeof(*bundle), GFP_KERNEL);
 
 	if (!bundle) {
-		dm_error("Failed to allocate update bundle\n");
+		drm_err(dev, "Failed to allocate update bundle\n");
 		goto cleanup;
 	}
 
@@ -9023,8 +9900,24 @@ static void amdgpu_dm_commit_planes(struct drm_atomic_state *state,
 	 * Disable the cursor first if we're disabling all the planes.
 	 * It'll remain on the screen after the planes are re-enabled
 	 * if we don't.
+	 *
+	 * If the cursor is transitioning from native to overlay mode, the
+	 * native cursor needs to be disabled first.
 	 */
-	if (acrtc_state->active_planes == 0)
+	if (acrtc_state->cursor_mode == DM_CURSOR_OVERLAY_MODE &&
+	    dm_old_crtc_state->cursor_mode == DM_CURSOR_NATIVE_MODE) {
+		struct dc_cursor_position cursor_position = {0};
+
+		if (!dc_stream_set_cursor_position(acrtc_state->stream,
+						   &cursor_position))
+			drm_err(dev, "DC failed to disable native cursor\n");
+
+		bundle->stream_update.cursor_position =
+				&acrtc_state->stream->cursor_position;
+	}
+
+	if (acrtc_state->active_planes == 0 &&
+	    dm_old_crtc_state->cursor_mode == DM_CURSOR_NATIVE_MODE)
 		amdgpu_dm_commit_cursors(state);
 
 	/* update planes when needed */
@@ -9038,8 +9931,17 @@ static void amdgpu_dm_commit_planes(struct drm_atomic_state *state,
 		struct dm_plane_state *dm_new_plane_state = to_dm_plane_state(new_plane_state);
 
 		/* Cursor plane is handled after stream updates */
-		if (plane->type == DRM_PLANE_TYPE_CURSOR)
+		if (plane->type == DRM_PLANE_TYPE_CURSOR &&
+		    acrtc_state->cursor_mode == DM_CURSOR_NATIVE_MODE) {
+			if ((fb && crtc == pcrtc) ||
+			    (old_plane_state->fb && old_plane_state->crtc == pcrtc)) {
+				cursor_update = true;
+				if (amdgpu_ip_version(dm->adev, DCE_HWIP, 0) != 0)
+					amdgpu_dm_update_cursor(plane, old_plane_state, &bundle->stream_update);
+			}
+
 			continue;
+		}
 
 		if (!fb || !crtc || pcrtc != crtc)
 			continue;
@@ -9049,15 +9951,21 @@ static void amdgpu_dm_commit_planes(struct drm_atomic_state *state,
 			continue;
 
 		dc_plane = dm_new_plane_state->dc_state;
+		if (!dc_plane)
+			continue;
 
 		bundle->surface_updates[planes_count].surface = dc_plane;
 		if (new_pcrtc_state->color_mgmt_changed) {
-			bundle->surface_updates[planes_count].gamma = dc_plane->gamma_correction;
-			bundle->surface_updates[planes_count].in_transfer_func = dc_plane->in_transfer_func;
+			bundle->surface_updates[planes_count].gamma = &dc_plane->gamma_correction;
+			bundle->surface_updates[planes_count].in_transfer_func = &dc_plane->in_transfer_func;
 			bundle->surface_updates[planes_count].gamut_remap_matrix = &dc_plane->gamut_remap_matrix;
+			bundle->surface_updates[planes_count].hdr_mult = dc_plane->hdr_mult;
+			bundle->surface_updates[planes_count].func_shaper = &dc_plane->in_shaper_func;
+			bundle->surface_updates[planes_count].lut3d_func = &dc_plane->lut3d_func;
+			bundle->surface_updates[planes_count].blend_tf = &dc_plane->blend_tf;
 		}
 
-		fill_dc_scaling_info(dm->adev, new_plane_state,
+		amdgpu_dm_plane_fill_dc_scaling_info(dm->adev, new_plane_state,
 				     &bundle->scaling_infos[planes_count]);
 
 		bundle->surface_updates[planes_count].scaling_info =
@@ -9072,39 +9980,69 @@ static void amdgpu_dm_commit_planes(struct drm_atomic_state *state,
 			continue;
 		}
 
-		abo = gem_to_amdgpu_bo(fb->obj[0]);
-
-		/*
-		 * Wait for all fences on this FB. Do limited wait to avoid
-		 * deadlock during GPU reset when this fence will not signal
-		 * but we hold reservation lock for the BO.
-		 */
-		r = dma_resv_wait_timeout(abo->tbo.base.resv, true, false,
-					  msecs_to_jiffies(5000));
-		if (unlikely(r <= 0))
-			DRM_ERROR("Waiting for fences timed out!");
-
 		fill_dc_plane_info_and_addr(
 			dm->adev, new_plane_state,
 			afb->tiling_flags,
 			&bundle->plane_infos[planes_count],
 			&bundle->flip_addrs[planes_count].address,
-			afb->tmz_surface, false);
+			afb->tmz_surface);
 
-		DRM_DEBUG_ATOMIC("plane: id=%d dcc_en=%d\n",
+		drm_dbg_state(state->dev, "plane: id=%d dcc_en=%d\n",
 				 new_plane_state->plane->index,
 				 bundle->plane_infos[planes_count].dcc.enable);
 
 		bundle->surface_updates[planes_count].plane_info =
 			&bundle->plane_infos[planes_count];
 
+		if (acrtc_state->stream->link->psr_settings.psr_feature_enabled ||
+		    acrtc_state->stream->link->replay_settings.replay_feature_enabled) {
+			fill_dc_dirty_rects(plane, old_plane_state,
+					    new_plane_state, new_crtc_state,
+					    &bundle->flip_addrs[planes_count],
+					    acrtc_state->stream->link->psr_settings.psr_version ==
+					    DC_PSR_VERSION_SU_1,
+					    &dirty_rects_changed);
+
+			/*
+			 * If the dirty regions changed, PSR-SU need to be disabled temporarily
+			 * and enabled it again after dirty regions are stable to avoid video glitch.
+			 * PSR-SU will be enabled in vblank_control_worker() if user pause the video
+			 * during the PSR-SU was disabled.
+			 */
+			if (acrtc_state->stream->link->psr_settings.psr_version >= DC_PSR_VERSION_SU_1 &&
+			    acrtc_attach->dm_irq_params.allow_sr_entry &&
+#ifdef CONFIG_DRM_AMD_SECURE_DISPLAY
+			    !amdgpu_dm_crc_window_is_activated(acrtc_state->base.crtc) &&
+#endif
+			    dirty_rects_changed) {
+				mutex_lock(&dm->dc_lock);
+				acrtc_state->stream->link->psr_settings.psr_dirty_rects_change_timestamp_ns =
+				timestamp_ns;
+				if (acrtc_state->stream->link->psr_settings.psr_allow_active)
+					amdgpu_dm_psr_disable(acrtc_state->stream, true);
+				mutex_unlock(&dm->dc_lock);
+			}
+		}
+
 		/*
 		 * Only allow immediate flips for fast updates that don't
-		 * change FB pitch, DCC state, rotation or mirroing.
+		 * change memory domain, FB pitch, DCC state, rotation or
+		 * mirroring.
+		 *
+		 * dm_crtc_helper_atomic_check() only accepts async flips with
+		 * fast updates.
 		 */
+		if (crtc->state->async_flip &&
+		    (acrtc_state->update_type != UPDATE_TYPE_FAST ||
+		     get_mem_type(old_plane_state->fb) != get_mem_type(fb)))
+			drm_warn_once(state->dev,
+				      "[PLANE:%d:%s] async flip with non-fast update\n",
+				      plane->base.id, plane->name);
+
 		bundle->flip_addrs[planes_count].flip_immediate =
 			crtc->state->async_flip &&
-			acrtc_state->update_type == UPDATE_TYPE_FAST;
+			acrtc_state->update_type == UPDATE_TYPE_FAST &&
+			get_mem_type(old_plane_state->fb) == get_mem_type(fb);
 
 		timestamp_ns = ktime_get_ns();
 		bundle->flip_addrs[planes_count].flip_timestamp_in_us = div_u64(timestamp_ns, 1000);
@@ -9112,7 +10050,7 @@ static void amdgpu_dm_commit_planes(struct drm_atomic_state *state,
 		bundle->surface_updates[planes_count].surface = dc_plane;
 
 		if (!bundle->surface_updates[planes_count].surface) {
-			DRM_ERROR("No surface for CRTC: id=%d\n",
+			drm_err(dev, "No surface for CRTC: id=%d\n",
 					acrtc_attach->crtc_id);
 			continue;
 		}
@@ -9125,7 +10063,7 @@ static void amdgpu_dm_commit_planes(struct drm_atomic_state *state,
 				dc_plane,
 				bundle->flip_addrs[planes_count].flip_timestamp_in_us);
 
-		DRM_DEBUG_ATOMIC("%s Flipping to hi: 0x%x, low: 0x%x\n",
+		drm_dbg_state(state->dev, "%s Flipping to hi: 0x%x, low: 0x%x\n",
 				 __func__,
 				 bundle->flip_addrs[planes_count].address.grph.addr.high_part,
 				 bundle->flip_addrs[planes_count].address.grph.addr.low_part);
@@ -9143,8 +10081,7 @@ static void amdgpu_dm_commit_planes(struct drm_atomic_state *state,
 			 * DRI3/Present extension with defined target_msc.
 			 */
 			last_flip_vblank = amdgpu_get_vblank_counter_kms(pcrtc);
-		}
-		else {
+		} else {
 			/* For variable refresh rate mode only:
 			 * Get vblank of last completed flip to avoid > 1 vrr
 			 * flips per video frame by use of throttling, but allow
@@ -9184,8 +10121,7 @@ static void amdgpu_dm_commit_planes(struct drm_atomic_state *state,
 		 * and rely on sending it from software.
 		 */
 		if (acrtc_attach->base.state->event &&
-		    acrtc_state->active_planes > 0 &&
-		    !acrtc_state->force_dpms_off) {
+		    acrtc_state->active_planes > 0) {
 			drm_crtc_vblank_get(pcrtc);
 
 			spin_lock_irqsave(&pcrtc->dev->event_lock, flags);
@@ -9201,19 +10137,25 @@ static void amdgpu_dm_commit_planes(struct drm_atomic_state *state,
 				bundle->stream_update.vrr_infopacket =
 					&acrtc_state->stream->vrr_infopacket;
 		}
+	} else if (cursor_update && acrtc_state->active_planes > 0) {
+		spin_lock_irqsave(&pcrtc->dev->event_lock, flags);
+		if (acrtc_attach->base.state->event) {
+			drm_crtc_vblank_get(pcrtc);
+			acrtc_attach->event = acrtc_attach->base.state->event;
+			acrtc_attach->base.state->event = NULL;
+		}
+		spin_unlock_irqrestore(&pcrtc->dev->event_lock, flags);
 	}
 
 	/* Update the planes if changed or disable if we don't have any. */
 	if ((planes_count || acrtc_state->active_planes == 0) &&
 		acrtc_state->stream) {
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 		/*
 		 * If PSR or idle optimizations are enabled then flush out
 		 * any pending work before hardware programming.
 		 */
 		if (dm->vblank_control_workqueue)
 			flush_workqueue(dm->vblank_control_workqueue);
-#endif
 
 		bundle->stream_update.stream = acrtc_state->stream;
 		if (new_pcrtc_state->mode_changed) {
@@ -9231,13 +10173,26 @@ static void amdgpu_dm_commit_planes(struct drm_atomic_state *state,
 			bundle->stream_update.output_csc_transform =
 				&acrtc_state->stream->csc_color_matrix;
 			bundle->stream_update.out_transfer_func =
-				acrtc_state->stream->out_transfer_func;
+				&acrtc_state->stream->out_transfer_func;
+			bundle->stream_update.lut3d_func =
+				(struct dc_3dlut *) acrtc_state->stream->lut3d_func;
+			bundle->stream_update.func_shaper =
+				(struct dc_transfer_func *) acrtc_state->stream->func_shaper;
 		}
 
 		acrtc_state->stream->abm_level = acrtc_state->abm_level;
 		if (acrtc_state->abm_level != dm_old_crtc_state->abm_level)
 			bundle->stream_update.abm_level = &acrtc_state->abm_level;
 
+		mutex_lock(&dm->dc_lock);
+		if ((acrtc_state->update_type > UPDATE_TYPE_FAST) || vrr_active) {
+			if (acrtc_state->stream->link->replay_settings.replay_allow_active)
+				amdgpu_dm_replay_disable(acrtc_state->stream);
+			if (acrtc_state->stream->link->psr_settings.psr_allow_active)
+				amdgpu_dm_psr_disable(acrtc_state->stream, true);
+		}
+		mutex_unlock(&dm->dc_lock);
+
 		/*
 		 * If FreeSync state on the stream has changed then we need to
 		 * re-adjust the min/max bounds now that DC doesn't handle this
@@ -9251,16 +10206,13 @@ static void amdgpu_dm_commit_planes(struct drm_atomic_state *state,
 			spin_unlock_irqrestore(&pcrtc->dev->event_lock, flags);
 		}
 		mutex_lock(&dm->dc_lock);
-		if ((acrtc_state->update_type > UPDATE_TYPE_FAST) &&
-				acrtc_state->stream->link->psr_settings.psr_allow_active)
-			amdgpu_dm_psr_disable(acrtc_state->stream);
-
-		dc_commit_updates_for_stream(dm->dc,
-						     bundle->surface_updates,
-						     planes_count,
-						     acrtc_state->stream,
-						     &bundle->stream_update,
-						     dc_state);
+		update_planes_and_stream_adapter(dm->dc,
+					 acrtc_state->update_type,
+					 planes_count,
+					 acrtc_state->stream,
+					 &bundle->stream_update,
+					 bundle->surface_updates);
+		updated_planes_and_streams = true;
 
 		/**
 		 * Enable or disable the interrupts on the backend.
@@ -9279,26 +10231,7 @@ static void amdgpu_dm_commit_planes(struct drm_atomic_state *state,
 			dm_update_pflip_irq_state(drm_to_adev(dev),
 						  acrtc_attach);
 
-		if ((acrtc_state->update_type > UPDATE_TYPE_FAST) &&
-				acrtc_state->stream->link->psr_settings.psr_version != DC_PSR_VERSION_UNSUPPORTED &&
-				!acrtc_state->stream->link->psr_settings.psr_feature_enabled)
-			amdgpu_dm_link_setup_psr(acrtc_state->stream);
-
-		/* Decrement skip count when PSR is enabled and we're doing fast updates. */
-		if (acrtc_state->update_type == UPDATE_TYPE_FAST &&
-		    acrtc_state->stream->link->psr_settings.psr_feature_enabled) {
-			struct amdgpu_dm_connector *aconn =
-				(struct amdgpu_dm_connector *)acrtc_state->stream->dm_stream_context;
-
-			if (aconn->psr_skip_count > 0)
-				aconn->psr_skip_count--;
-
-			/* Allow PSR when skip count is 0. */
-			acrtc_attach->dm_irq_params.allow_psr_entry = !aconn->psr_skip_count;
-		} else {
-			acrtc_attach->dm_irq_params.allow_psr_entry = false;
-		}
-
+		amdgpu_dm_enable_self_refresh(acrtc_attach, acrtc_state, timestamp_ns);
 		mutex_unlock(&dm->dc_lock);
 	}
 
@@ -9307,7 +10240,9 @@ static void amdgpu_dm_commit_planes(struct drm_atomic_state *state,
 	 * This avoids redundant programming in the case where we're going
 	 * to be disabling a single plane - those pipes are being disabled.
 	 */
-	if (acrtc_state->active_planes)
+	if (acrtc_state->active_planes &&
+	    (!updated_planes_and_streams || amdgpu_ip_version(dm->adev, DCE_HWIP, 0) == 0) &&
+	    acrtc_state->cursor_mode == DM_CURSOR_NATIVE_MODE)
 		amdgpu_dm_commit_cursors(state);
 
 cleanup:
@@ -9345,7 +10280,10 @@ static void amdgpu_dm_commit_audio(struct drm_device *dev,
 		if (!drm_atomic_crtc_needs_modeset(new_crtc_state))
 			continue;
 
-	notify:
+notify:
+		if (connector->connector_type == DRM_MODE_CONNECTOR_WRITEBACK)
+			continue;
+
 		aconnector = to_amdgpu_dm_connector(connector);
 
 		mutex_lock(&adev->dm.audio_lock);
@@ -9378,6 +10316,9 @@ static void amdgpu_dm_commit_audio(struct drm_device *dev,
 		if (!status)
 			continue;
 
+		if (connector->connector_type == DRM_MODE_CONNECTOR_WRITEBACK)
+			continue;
+
 		aconnector = to_amdgpu_dm_connector(connector);
 
 		mutex_lock(&adev->dm.audio_lock);
@@ -9403,49 +10344,57 @@ static void amdgpu_dm_crtc_copy_transient_flags(struct drm_crtc_state *crtc_stat
 	stream_state->mode_changed = drm_atomic_crtc_needs_modeset(crtc_state);
 }
 
-/**
- * amdgpu_dm_atomic_commit_tail() - AMDgpu DM's commit tail implementation.
- * @state: The atomic state to commit
- *
- * This will tell DC to commit the constructed DC state from atomic_check,
- * programming the hardware. Any failures here implies a hardware failure, since
- * atomic check should have filtered anything non-kosher.
- */
-static void amdgpu_dm_atomic_commit_tail(struct drm_atomic_state *state)
+static void dm_clear_writeback(struct amdgpu_display_manager *dm,
+			      struct dm_crtc_state *crtc_state)
+{
+	dc_stream_remove_writeback(dm->dc, crtc_state->stream, 0);
+}
+
+static void amdgpu_dm_commit_streams(struct drm_atomic_state *state,
+					struct dc_state *dc_state)
 {
 	struct drm_device *dev = state->dev;
 	struct amdgpu_device *adev = drm_to_adev(dev);
 	struct amdgpu_display_manager *dm = &adev->dm;
-	struct dm_atomic_state *dm_state;
-	struct dc_state *dc_state = NULL, *dc_state_temp = NULL;
-	uint32_t i, j;
 	struct drm_crtc *crtc;
 	struct drm_crtc_state *old_crtc_state, *new_crtc_state;
-	unsigned long flags;
-	bool wait_for_vblank = true;
-	struct drm_connector *connector;
-	struct drm_connector_state *old_con_state, *new_con_state;
 	struct dm_crtc_state *dm_old_crtc_state, *dm_new_crtc_state;
-	int crtc_disable_count = 0;
+	struct drm_connector_state *old_con_state;
+	struct drm_connector *connector;
 	bool mode_set_reset_required = false;
+	u32 i;
+	struct dc_commit_streams_params params = {dc_state->streams, dc_state->stream_count};
+	bool set_backlight_level = false;
 
-	trace_amdgpu_dm_atomic_commit_tail_begin(state);
+	/* Disable writeback */
+	for_each_old_connector_in_state(state, connector, old_con_state, i) {
+		struct dm_connector_state *dm_old_con_state;
+		struct amdgpu_crtc *acrtc;
 
-	drm_atomic_helper_update_legacy_modeset_state(dev, state);
+		if (connector->connector_type != DRM_MODE_CONNECTOR_WRITEBACK)
+			continue;
 
-	dm_state = dm_atomic_get_new_state(state);
-	if (dm_state && dm_state->context) {
-		dc_state = dm_state->context;
-	} else {
-		/* No state changes, retain current state. */
-		dc_state_temp = dc_create_state(dm->dc);
-		ASSERT(dc_state_temp);
-		dc_state = dc_state_temp;
-		dc_resource_state_copy_construct_current(dm->dc, dc_state);
+		old_crtc_state = NULL;
+
+		dm_old_con_state = to_dm_connector_state(old_con_state);
+		if (!dm_old_con_state->base.crtc)
+			continue;
+
+		acrtc = to_amdgpu_crtc(dm_old_con_state->base.crtc);
+		if (acrtc)
+			old_crtc_state = drm_atomic_get_old_crtc_state(state, &acrtc->base);
+
+		if (!acrtc || !acrtc->wb_enabled)
+			continue;
+
+		dm_old_crtc_state = to_dm_crtc_state(old_crtc_state);
+
+		dm_clear_writeback(dm, dm_old_crtc_state);
+		acrtc->wb_enabled = false;
 	}
 
-	for_each_oldnew_crtc_in_state (state, crtc, old_crtc_state,
-				       new_crtc_state, i) {
+	for_each_oldnew_crtc_in_state(state, crtc, old_crtc_state,
+				      new_crtc_state, i) {
 		struct amdgpu_crtc *acrtc = to_amdgpu_crtc(crtc);
 
 		dm_old_crtc_state = to_dm_crtc_state(old_crtc_state);
@@ -9453,7 +10402,7 @@ static void amdgpu_dm_atomic_commit_tail(struct drm_atomic_state *state)
 		if (old_crtc_state->active &&
 		    (!new_crtc_state->active ||
 		     drm_atomic_crtc_needs_modeset(new_crtc_state))) {
-			manage_dm_interrupts(adev, acrtc, false);
+			manage_dm_interrupts(adev, acrtc, NULL);
 			dc_stream_release(dm_old_crtc_state->stream);
 		}
 	}
@@ -9467,10 +10416,8 @@ static void amdgpu_dm_atomic_commit_tail(struct drm_atomic_state *state)
 		dm_new_crtc_state = to_dm_crtc_state(new_crtc_state);
 		dm_old_crtc_state = to_dm_crtc_state(old_crtc_state);
 
-		DRM_DEBUG_ATOMIC(
-			"amdgpu_crtc id:%d crtc_state_flags: enable:%d, active:%d, "
-			"planes_changed:%d, mode_changed:%d,active_changed:%d,"
-			"connectors_changed:%d\n",
+		drm_dbg_state(state->dev,
+			"amdgpu_crtc id:%d crtc_state_flags: enable:%d, active:%d, planes_changed:%d, mode_changed:%d,active_changed:%d,connectors_changed:%d\n",
 			acrtc->crtc_id,
 			new_crtc_state->enable,
 			new_crtc_state->active,
@@ -9485,7 +10432,8 @@ static void amdgpu_dm_atomic_commit_tail(struct drm_atomic_state *state)
 
 			memset(&position, 0, sizeof(position));
 			mutex_lock(&dm->dc_lock);
-			dc_stream_set_cursor_position(dm_old_crtc_state->stream, &position);
+			dc_exit_ips_for_hw_access(dm->dc);
+			dc_stream_program_cursor_position(dm_old_crtc_state->stream, &position);
 			mutex_unlock(&dm->dc_lock);
 		}
 
@@ -9499,9 +10447,11 @@ static void amdgpu_dm_atomic_commit_tail(struct drm_atomic_state *state)
 		 * aconnector as needed
 		 */
 
-		if (modeset_required(new_crtc_state, dm_new_crtc_state->stream, dm_old_crtc_state->stream)) {
+		if (amdgpu_dm_crtc_modeset_required(new_crtc_state, dm_new_crtc_state->stream, dm_old_crtc_state->stream)) {
 
-			DRM_DEBUG_ATOMIC("Atomic commit: SET crtc id %d: [%p]\n", acrtc->crtc_id, acrtc);
+			drm_dbg_atomic(dev,
+				       "Atomic commit: SET crtc id %d: [%p]\n",
+				       acrtc->crtc_id, acrtc);
 
 			if (!dm_new_crtc_state->stream) {
 				/*
@@ -9519,8 +10469,9 @@ static void amdgpu_dm_atomic_commit_tail(struct drm_atomic_state *state)
 				 * have a sink to keep the pipe running so that
 				 * hw state is consistent with the sw state
 				 */
-				DRM_DEBUG_DRIVER("%s: Failed to create new stream for crtc %d\n",
-						__func__, acrtc->base.base.id);
+				drm_dbg_atomic(dev,
+					       "Failed to create new stream for crtc %d\n",
+						acrtc->base.base.id);
 				continue;
 			}
 
@@ -9533,8 +10484,11 @@ static void amdgpu_dm_atomic_commit_tail(struct drm_atomic_state *state)
 			acrtc->hw_mode = new_crtc_state->mode;
 			crtc->hwmode = new_crtc_state->mode;
 			mode_set_reset_required = true;
+			set_backlight_level = true;
 		} else if (modereset_required(new_crtc_state)) {
-			DRM_DEBUG_ATOMIC("Atomic commit: RESET. crtc id %d:[%p]\n", acrtc->crtc_id, acrtc);
+			drm_dbg_atomic(dev,
+				       "Atomic commit: RESET. crtc id %d:[%p]\n",
+				       acrtc->crtc_id, acrtc);
 			/* i.e. reset mode */
 			if (dm_old_crtc_state->stream)
 				remove_stream(adev, acrtc, dm_old_crtc_state->stream);
@@ -9543,27 +10497,25 @@ static void amdgpu_dm_atomic_commit_tail(struct drm_atomic_state *state)
 		}
 	} /* for_each_crtc_in_state() */
 
-	if (dc_state) {
-		/* if there mode set or reset, disable eDP PSR */
-		if (mode_set_reset_required) {
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-			if (dm->vblank_control_workqueue)
-				flush_workqueue(dm->vblank_control_workqueue);
-#endif
-			amdgpu_dm_psr_disable_all(dm);
-		}
+	/* if there mode set or reset, disable eDP PSR, Replay */
+	if (mode_set_reset_required) {
+		if (dm->vblank_control_workqueue)
+			flush_workqueue(dm->vblank_control_workqueue);
 
-		dm_enable_per_frame_crtc_master_sync(dc_state);
-		mutex_lock(&dm->dc_lock);
-		WARN_ON(!dc_commit_state(dm->dc, dc_state));
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-               /* Allow idle optimization when vblank count is 0 for display off */
-               if (dm->active_vblank_irq_count == 0)
-                   dc_allow_idle_optimizations(dm->dc,true);
-#endif
-		mutex_unlock(&dm->dc_lock);
+		amdgpu_dm_replay_disable_all(dm);
+		amdgpu_dm_psr_disable_all(dm);
 	}
 
+	dm_enable_per_frame_crtc_master_sync(dc_state);
+	mutex_lock(&dm->dc_lock);
+	dc_exit_ips_for_hw_access(dm->dc);
+	WARN_ON(!dc_commit_streams(dm->dc, &params));
+
+	/* Allow idle optimization when vblank count is 0 for display off */
+	if ((dm->active_vblank_irq_count == 0) && amdgpu_dm_is_headless(dm->adev))
+		dc_allow_idle_optimizations(dm->dc, true);
+	mutex_unlock(&dm->dc_lock);
+
 	for_each_new_crtc_in_state(state, crtc, new_crtc_state, i) {
 		struct amdgpu_crtc *acrtc = to_amdgpu_crtc(crtc);
 
@@ -9574,24 +10526,192 @@ static void amdgpu_dm_atomic_commit_tail(struct drm_atomic_state *state)
 					dc_stream_get_status(dm_new_crtc_state->stream);
 
 			if (!status)
-				status = dc_stream_get_status_from_state(dc_state,
+				status = dc_state_get_stream_status(dc_state,
 									 dm_new_crtc_state->stream);
 			if (!status)
-				DC_ERR("got no status for stream %p on acrtc%p\n", dm_new_crtc_state->stream, acrtc);
+				drm_err(dev,
+					"got no status for stream %p on acrtc%p\n",
+					dm_new_crtc_state->stream, acrtc);
 			else
 				acrtc->otg_inst = status->primary_otg_inst;
 		}
 	}
-#ifdef CONFIG_DRM_AMD_DC_HDCP
+
+	/* During boot up and resume the DC layer will reset the panel brightness
+	 * to fix a flicker issue.
+	 * It will cause the dm->actual_brightness is not the current panel brightness
+	 * level. (the dm->brightness is the correct panel level)
+	 * So we set the backlight level with dm->brightness value after set mode
+	 */
+	if (set_backlight_level) {
+		for (i = 0; i < dm->num_of_edps; i++) {
+			if (dm->backlight_dev[i])
+				amdgpu_dm_backlight_set_level(dm, i, dm->brightness[i]);
+		}
+	}
+}
+
+static void dm_set_writeback(struct amdgpu_display_manager *dm,
+			      struct dm_crtc_state *crtc_state,
+			      struct drm_connector *connector,
+			      struct drm_connector_state *new_con_state)
+{
+	struct drm_writeback_connector *wb_conn = drm_connector_to_writeback(connector);
+	struct amdgpu_device *adev = dm->adev;
+	struct amdgpu_crtc *acrtc;
+	struct dc_writeback_info *wb_info;
+	struct pipe_ctx *pipe = NULL;
+	struct amdgpu_framebuffer *afb;
+	int i = 0;
+
+	wb_info = kzalloc(sizeof(*wb_info), GFP_KERNEL);
+	if (!wb_info) {
+		drm_err(adev_to_drm(adev), "Failed to allocate wb_info\n");
+		return;
+	}
+
+	acrtc = to_amdgpu_crtc(wb_conn->encoder.crtc);
+	if (!acrtc) {
+		drm_err(adev_to_drm(adev), "no amdgpu_crtc found\n");
+		kfree(wb_info);
+		return;
+	}
+
+	afb = to_amdgpu_framebuffer(new_con_state->writeback_job->fb);
+	if (!afb) {
+		drm_err(adev_to_drm(adev), "No amdgpu_framebuffer found\n");
+		kfree(wb_info);
+		return;
+	}
+
+	for (i = 0; i < MAX_PIPES; i++) {
+		if (dm->dc->current_state->res_ctx.pipe_ctx[i].stream == crtc_state->stream) {
+			pipe = &dm->dc->current_state->res_ctx.pipe_ctx[i];
+			break;
+		}
+	}
+
+	/* fill in wb_info */
+	wb_info->wb_enabled = true;
+
+	wb_info->dwb_pipe_inst = 0;
+	wb_info->dwb_params.dwbscl_black_color = 0;
+	wb_info->dwb_params.hdr_mult = 0x1F000;
+	wb_info->dwb_params.csc_params.gamut_adjust_type = CM_GAMUT_ADJUST_TYPE_BYPASS;
+	wb_info->dwb_params.csc_params.gamut_coef_format = CM_GAMUT_REMAP_COEF_FORMAT_S2_13;
+	wb_info->dwb_params.output_depth = DWB_OUTPUT_PIXEL_DEPTH_10BPC;
+	wb_info->dwb_params.cnv_params.cnv_out_bpc = DWB_CNV_OUT_BPC_10BPC;
+
+	/* width & height from crtc */
+	wb_info->dwb_params.cnv_params.src_width = acrtc->base.mode.crtc_hdisplay;
+	wb_info->dwb_params.cnv_params.src_height = acrtc->base.mode.crtc_vdisplay;
+	wb_info->dwb_params.dest_width = acrtc->base.mode.crtc_hdisplay;
+	wb_info->dwb_params.dest_height = acrtc->base.mode.crtc_vdisplay;
+
+	wb_info->dwb_params.cnv_params.crop_en = false;
+	wb_info->dwb_params.stereo_params.stereo_enabled = false;
+
+	wb_info->dwb_params.cnv_params.out_max_pix_val = 0x3ff;	// 10 bits
+	wb_info->dwb_params.cnv_params.out_min_pix_val = 0;
+	wb_info->dwb_params.cnv_params.fc_out_format = DWB_OUT_FORMAT_32BPP_ARGB;
+	wb_info->dwb_params.cnv_params.out_denorm_mode = DWB_OUT_DENORM_BYPASS;
+
+	wb_info->dwb_params.out_format = dwb_scaler_mode_bypass444;
+
+	wb_info->dwb_params.capture_rate = dwb_capture_rate_0;
+
+	wb_info->dwb_params.scaler_taps.h_taps = 4;
+	wb_info->dwb_params.scaler_taps.v_taps = 4;
+	wb_info->dwb_params.scaler_taps.h_taps_c = 2;
+	wb_info->dwb_params.scaler_taps.v_taps_c = 2;
+	wb_info->dwb_params.subsample_position = DWB_INTERSTITIAL_SUBSAMPLING;
+
+	wb_info->mcif_buf_params.luma_pitch = afb->base.pitches[0];
+	wb_info->mcif_buf_params.chroma_pitch = afb->base.pitches[1];
+
+	for (i = 0; i < DWB_MCIF_BUF_COUNT; i++) {
+		wb_info->mcif_buf_params.luma_address[i] = afb->address;
+		wb_info->mcif_buf_params.chroma_address[i] = 0;
+	}
+
+	wb_info->mcif_buf_params.p_vmid = 1;
+	if (amdgpu_ip_version(adev, DCE_HWIP, 0) >= IP_VERSION(3, 0, 0)) {
+		wb_info->mcif_warmup_params.start_address.quad_part = afb->address;
+		wb_info->mcif_warmup_params.region_size =
+			wb_info->mcif_buf_params.luma_pitch * wb_info->dwb_params.dest_height;
+	}
+	wb_info->mcif_warmup_params.p_vmid = 1;
+	wb_info->writeback_source_plane = pipe->plane_state;
+
+	dc_stream_add_writeback(dm->dc, crtc_state->stream, wb_info);
+
+	acrtc->wb_pending = true;
+	acrtc->wb_conn = wb_conn;
+	drm_writeback_queue_job(wb_conn, new_con_state);
+}
+
+static void amdgpu_dm_update_hdcp(struct drm_atomic_state *state)
+{
+	struct drm_connector_state *old_con_state, *new_con_state;
+	struct drm_device *dev = state->dev;
+	struct drm_connector *connector;
+	struct amdgpu_device *adev = drm_to_adev(dev);
+	int i;
+
+	if (!adev->dm.hdcp_workqueue)
+		return;
+
 	for_each_oldnew_connector_in_state(state, connector, old_con_state, new_con_state, i) {
 		struct dm_connector_state *dm_new_con_state = to_dm_connector_state(new_con_state);
 		struct amdgpu_crtc *acrtc = to_amdgpu_crtc(dm_new_con_state->base.crtc);
-		struct amdgpu_dm_connector *aconnector = to_amdgpu_dm_connector(connector);
+		struct drm_crtc_state *old_crtc_state, *new_crtc_state;
+		struct dm_crtc_state *dm_new_crtc_state;
+		struct amdgpu_dm_connector *aconnector;
+
+		if (!connector || connector->connector_type == DRM_MODE_CONNECTOR_WRITEBACK)
+			continue;
+
+		aconnector = to_amdgpu_dm_connector(connector);
+
+		drm_dbg(dev, "[HDCP_DM] -------------- i : %x ----------\n", i);
+
+		drm_dbg(dev, "[HDCP_DM] connector->index: %x connect_status: %x dpms: %x\n",
+			connector->index, connector->status, connector->dpms);
+		drm_dbg(dev, "[HDCP_DM] state protection old: %x new: %x\n",
+			old_con_state->content_protection, new_con_state->content_protection);
+
+		if (aconnector->dc_sink) {
+			if (aconnector->dc_sink->sink_signal != SIGNAL_TYPE_VIRTUAL &&
+				aconnector->dc_sink->sink_signal != SIGNAL_TYPE_NONE) {
+				drm_dbg(dev, "[HDCP_DM] pipe_ctx dispname=%s\n",
+				aconnector->dc_sink->edid_caps.display_name);
+			}
+		}
 
 		new_crtc_state = NULL;
+		old_crtc_state = NULL;
 
-		if (acrtc)
+		if (acrtc) {
 			new_crtc_state = drm_atomic_get_new_crtc_state(state, &acrtc->base);
+			old_crtc_state = drm_atomic_get_old_crtc_state(state, &acrtc->base);
+		}
+
+		if (old_crtc_state)
+			drm_dbg(dev, "old crtc en: %x a: %x m: %x a-chg: %x c-chg: %x\n",
+			old_crtc_state->enable,
+			old_crtc_state->active,
+			old_crtc_state->mode_changed,
+			old_crtc_state->active_changed,
+			old_crtc_state->connectors_changed);
+
+		if (new_crtc_state)
+			drm_dbg(dev, "NEW crtc en: %x a: %x m: %x a-chg: %x c-chg: %x\n",
+			new_crtc_state->enable,
+			new_crtc_state->active,
+			new_crtc_state->mode_changed,
+			new_crtc_state->active_changed,
+			new_crtc_state->connectors_changed);
+
 
 		dm_new_crtc_state = to_dm_crtc_state(new_crtc_state);
 
@@ -9603,26 +10723,130 @@ static void amdgpu_dm_atomic_commit_tail(struct drm_atomic_state *state)
 			continue;
 		}
 
-		if (is_content_protection_different(new_con_state, old_con_state, connector, adev->dm.hdcp_workqueue))
-			hdcp_update_display(
-				adev->dm.hdcp_workqueue, aconnector->dc_link->link_index, aconnector,
-				new_con_state->hdcp_content_type,
-				new_con_state->content_protection == DRM_MODE_CONTENT_PROTECTION_DESIRED);
+		if (is_content_protection_different(new_crtc_state, old_crtc_state, new_con_state,
+											old_con_state, connector, adev->dm.hdcp_workqueue)) {
+			/* when display is unplugged from mst hub, connctor will
+			 * be destroyed within dm_dp_mst_connector_destroy. connector
+			 * hdcp perperties, like type, undesired, desired, enabled,
+			 * will be lost. So, save hdcp properties into hdcp_work within
+			 * amdgpu_dm_atomic_commit_tail. if the same display is
+			 * plugged back with same display index, its hdcp properties
+			 * will be retrieved from hdcp_work within dm_dp_mst_get_modes
+			 */
+
+			bool enable_encryption = false;
+
+			if (new_con_state->content_protection == DRM_MODE_CONTENT_PROTECTION_DESIRED)
+				enable_encryption = true;
+
+			if (aconnector->dc_link && aconnector->dc_sink &&
+				aconnector->dc_link->type == dc_connection_mst_branch) {
+				struct hdcp_workqueue *hdcp_work = adev->dm.hdcp_workqueue;
+				struct hdcp_workqueue *hdcp_w =
+					&hdcp_work[aconnector->dc_link->link_index];
+
+				hdcp_w->hdcp_content_type[connector->index] =
+					new_con_state->hdcp_content_type;
+				hdcp_w->content_protection[connector->index] =
+					new_con_state->content_protection;
+			}
+
+			if (new_crtc_state && new_crtc_state->mode_changed &&
+				new_con_state->content_protection >= DRM_MODE_CONTENT_PROTECTION_DESIRED)
+				enable_encryption = true;
+
+			drm_info(dev, "[HDCP_DM] hdcp_update_display enable_encryption = %x\n", enable_encryption);
+
+			if (aconnector->dc_link)
+				hdcp_update_display(
+					adev->dm.hdcp_workqueue, aconnector->dc_link->link_index, aconnector,
+					new_con_state->hdcp_content_type, enable_encryption);
+		}
 	}
-#endif
+}
+
+static int amdgpu_dm_atomic_setup_commit(struct drm_atomic_state *state)
+{
+	struct drm_crtc *crtc;
+	struct drm_crtc_state *old_crtc_state, *new_crtc_state;
+	struct dm_crtc_state *dm_old_crtc_state, *dm_new_crtc_state;
+	int i, ret;
+
+	ret = drm_dp_mst_atomic_setup_commit(state);
+	if (ret)
+		return ret;
+
+	for_each_oldnew_crtc_in_state(state, crtc, old_crtc_state, new_crtc_state, i) {
+		dm_old_crtc_state = to_dm_crtc_state(old_crtc_state);
+		dm_new_crtc_state = to_dm_crtc_state(new_crtc_state);
+		/*
+		 * Color management settings. We also update color properties
+		 * when a modeset is needed, to ensure it gets reprogrammed.
+		 */
+		if (dm_new_crtc_state->base.active && dm_new_crtc_state->stream &&
+		    (dm_new_crtc_state->base.color_mgmt_changed ||
+		     dm_old_crtc_state->regamma_tf != dm_new_crtc_state->regamma_tf ||
+		     drm_atomic_crtc_needs_modeset(new_crtc_state))) {
+			ret = amdgpu_dm_update_crtc_color_mgmt(dm_new_crtc_state);
+			if (ret) {
+				drm_dbg_atomic(state->dev, "Failed to update color state\n");
+				return ret;
+			}
+		}
+	}
+
+	return 0;
+}
+
+/**
+ * amdgpu_dm_atomic_commit_tail() - AMDgpu DM's commit tail implementation.
+ * @state: The atomic state to commit
+ *
+ * This will tell DC to commit the constructed DC state from atomic_check,
+ * programming the hardware. Any failures here implies a hardware failure, since
+ * atomic check should have filtered anything non-kosher.
+ */
+static void amdgpu_dm_atomic_commit_tail(struct drm_atomic_state *state)
+{
+	struct drm_device *dev = state->dev;
+	struct amdgpu_device *adev = drm_to_adev(dev);
+	struct amdgpu_display_manager *dm = &adev->dm;
+	struct dm_atomic_state *dm_state;
+	struct dc_state *dc_state = NULL;
+	u32 i, j;
+	struct drm_crtc *crtc;
+	struct drm_crtc_state *old_crtc_state, *new_crtc_state;
+	unsigned long flags;
+	bool wait_for_vblank = true;
+	struct drm_connector *connector;
+	struct drm_connector_state *old_con_state = NULL, *new_con_state = NULL;
+	struct dm_crtc_state *dm_old_crtc_state, *dm_new_crtc_state;
+	int crtc_disable_count = 0;
+
+	trace_amdgpu_dm_atomic_commit_tail_begin(state);
+
+	drm_atomic_helper_update_legacy_modeset_state(dev, state);
+	drm_dp_mst_atomic_wait_for_dependencies(state);
+
+	dm_state = dm_atomic_get_new_state(state);
+	if (dm_state && dm_state->context) {
+		dc_state = dm_state->context;
+		amdgpu_dm_commit_streams(state, dc_state);
+	}
+
+	amdgpu_dm_update_hdcp(state);
 
 	/* Handle connector state changes */
 	for_each_oldnew_connector_in_state(state, connector, old_con_state, new_con_state, i) {
 		struct dm_connector_state *dm_new_con_state = to_dm_connector_state(new_con_state);
 		struct dm_connector_state *dm_old_con_state = to_dm_connector_state(old_con_state);
 		struct amdgpu_crtc *acrtc = to_amdgpu_crtc(dm_new_con_state->base.crtc);
-		struct dc_surface_update dummy_updates[MAX_SURFACES];
+		struct dc_surface_update *dummy_updates;
 		struct dc_stream_update stream_update;
 		struct dc_info_packet hdr_packet;
 		struct dc_stream_status *status = NULL;
-		bool abm_changed, hdr_changed, scaling_changed;
+		bool abm_changed, hdr_changed, scaling_changed, output_color_space_changed = false;
 
-		memset(&dummy_updates, 0, sizeof(dummy_updates));
 		memset(&stream_update, 0, sizeof(stream_update));
 
 		if (acrtc) {
@@ -9640,13 +10864,18 @@ static void amdgpu_dm_atomic_commit_tail(struct drm_atomic_state *state)
 		scaling_changed = is_scaling_state_different(dm_new_con_state,
 							     dm_old_con_state);
 
+		if ((new_con_state->hdmi.broadcast_rgb != old_con_state->hdmi.broadcast_rgb) &&
+			(dm_old_crtc_state->stream->output_color_space !=
+				get_output_color_space(&dm_new_crtc_state->stream->timing, new_con_state)))
+			output_color_space_changed = true;
+
 		abm_changed = dm_new_crtc_state->abm_level !=
 			      dm_old_crtc_state->abm_level;
 
 		hdr_changed =
 			!drm_connector_atomic_hdr_metadata_equal(old_con_state, new_con_state);
 
-		if (!scaling_changed && !abm_changed && !hdr_changed)
+		if (!scaling_changed && !abm_changed && !hdr_changed && !output_color_space_changed)
 			continue;
 
 		stream_update.stream = dm_new_crtc_state->stream;
@@ -9658,6 +10887,13 @@ static void amdgpu_dm_atomic_commit_tail(struct drm_atomic_state *state)
 			stream_update.dst = dm_new_crtc_state->stream->dst;
 		}
 
+		if (output_color_space_changed) {
+			dm_new_crtc_state->stream->output_color_space
+				= get_output_color_space(&dm_new_crtc_state->stream->timing, new_con_state);
+
+			stream_update.output_color_space = &dm_new_crtc_state->stream->output_color_space;
+		}
+
 		if (abm_changed) {
 			dm_new_crtc_state->stream->abm_level = dm_new_crtc_state->abm_level;
 
@@ -9681,35 +10917,28 @@ static void amdgpu_dm_atomic_commit_tail(struct drm_atomic_state *state)
 		 * Here we create an empty update on each plane.
 		 * To fix this, DC should permit updating only stream properties.
 		 */
+		dummy_updates = kzalloc(sizeof(struct dc_surface_update) * MAX_SURFACES, GFP_KERNEL);
+		if (!dummy_updates) {
+			drm_err(adev_to_drm(adev), "Failed to allocate memory for dummy_updates.\n");
+			continue;
+		}
 		for (j = 0; j < status->plane_count; j++)
 			dummy_updates[j].surface = status->plane_states[0];
 
+		sort(dummy_updates, status->plane_count,
+		     sizeof(*dummy_updates), dm_plane_layer_index_cmp, NULL);
 
 		mutex_lock(&dm->dc_lock);
-		dc_commit_updates_for_stream(dm->dc,
-						     dummy_updates,
-						     status->plane_count,
-						     dm_new_crtc_state->stream,
-						     &stream_update,
-						     dc_state);
+		dc_exit_ips_for_hw_access(dm->dc);
+		dc_update_planes_and_stream(dm->dc,
+					    dummy_updates,
+					    status->plane_count,
+					    dm_new_crtc_state->stream,
+					    &stream_update);
 		mutex_unlock(&dm->dc_lock);
-	}
+		kfree(dummy_updates);
 
-	/* Count number of newly disabled CRTCs for dropping PM refs later. */
-	for_each_oldnew_crtc_in_state(state, crtc, old_crtc_state,
-				      new_crtc_state, i) {
-		if (old_crtc_state->active && !new_crtc_state->active)
-			crtc_disable_count++;
-
-		dm_new_crtc_state = to_dm_crtc_state(new_crtc_state);
-		dm_old_crtc_state = to_dm_crtc_state(old_crtc_state);
-
-		/* For freesync config update on crtc state and params for irq */
-		update_stream_irq_parameters(dm, dm_new_crtc_state);
-
-		/* Handle vrr on->off / off->on transitions */
-		amdgpu_dm_handle_vrr_transition(dm_old_crtc_state,
-						dm_new_crtc_state);
+		drm_connector_update_privacy_screen(new_con_state);
 	}
 
 	/**
@@ -9721,52 +10950,68 @@ static void amdgpu_dm_atomic_commit_tail(struct drm_atomic_state *state)
 	for_each_oldnew_crtc_in_state(state, crtc, old_crtc_state, new_crtc_state, i) {
 		struct amdgpu_crtc *acrtc = to_amdgpu_crtc(crtc);
 #ifdef CONFIG_DEBUG_FS
-		bool configure_crc = false;
 		enum amdgpu_dm_pipe_crc_source cur_crc_src;
-#if defined(CONFIG_DRM_AMD_SECURE_DISPLAY)
-		struct crc_rd_work *crc_rd_wrk = dm->crc_rd_wrk;
 #endif
+		/* Count number of newly disabled CRTCs for dropping PM refs later. */
+		if (old_crtc_state->active && !new_crtc_state->active)
+			crtc_disable_count++;
+
+		dm_new_crtc_state = to_dm_crtc_state(new_crtc_state);
+		dm_old_crtc_state = to_dm_crtc_state(old_crtc_state);
+
+		/* For freesync config update on crtc state and params for irq */
+		update_stream_irq_parameters(dm, dm_new_crtc_state);
+
+#ifdef CONFIG_DEBUG_FS
 		spin_lock_irqsave(&adev_to_drm(adev)->event_lock, flags);
 		cur_crc_src = acrtc->dm_irq_params.crc_src;
 		spin_unlock_irqrestore(&adev_to_drm(adev)->event_lock, flags);
 #endif
-		dm_new_crtc_state = to_dm_crtc_state(new_crtc_state);
 
 		if (new_crtc_state->active &&
 		    (!old_crtc_state->active ||
 		     drm_atomic_crtc_needs_modeset(new_crtc_state))) {
 			dc_stream_retain(dm_new_crtc_state->stream);
 			acrtc->dm_irq_params.stream = dm_new_crtc_state->stream;
-			manage_dm_interrupts(adev, acrtc, true);
+			manage_dm_interrupts(adev, acrtc, dm_new_crtc_state);
+		}
+		/* Handle vrr on->off / off->on transitions */
+		amdgpu_dm_handle_vrr_transition(dm_old_crtc_state, dm_new_crtc_state);
 
 #ifdef CONFIG_DEBUG_FS
+		if (new_crtc_state->active &&
+		    (!old_crtc_state->active ||
+		     drm_atomic_crtc_needs_modeset(new_crtc_state))) {
 			/**
 			 * Frontend may have changed so reapply the CRC capture
 			 * settings for the stream.
 			 */
-			dm_new_crtc_state = to_dm_crtc_state(new_crtc_state);
-
 			if (amdgpu_dm_is_valid_crc_source(cur_crc_src)) {
-				configure_crc = true;
 #if defined(CONFIG_DRM_AMD_SECURE_DISPLAY)
 				if (amdgpu_dm_crc_window_is_activated(crtc)) {
+					uint8_t cnt;
+
 					spin_lock_irqsave(&adev_to_drm(adev)->event_lock, flags);
-					acrtc->dm_irq_params.crc_window.update_win = true;
-					acrtc->dm_irq_params.crc_window.skip_frame_cnt = 2;
-					spin_lock_irq(&crc_rd_wrk->crc_rd_work_lock);
-					crc_rd_wrk->crtc = crtc;
-					spin_unlock_irq(&crc_rd_wrk->crc_rd_work_lock);
+					for (cnt = 0; cnt < MAX_CRC_WINDOW_NUM; cnt++) {
+						if (acrtc->dm_irq_params.window_param[cnt].enable) {
+							acrtc->dm_irq_params.window_param[cnt].update_win = true;
+
+							/**
+							 * It takes 2 frames for HW to stably generate CRC when
+							 * resuming from suspend, so we set skip_frame_cnt 2.
+							 */
+							acrtc->dm_irq_params.window_param[cnt].skip_frame_cnt = 2;
+						}
+					}
 					spin_unlock_irqrestore(&adev_to_drm(adev)->event_lock, flags);
 				}
 #endif
-			}
-
-			if (configure_crc)
 				if (amdgpu_dm_crtc_configure_crc_source(
 					crtc, dm_new_crtc_state, cur_crc_src))
-					DRM_DEBUG_DRIVER("Failed to configure crc source");
-#endif
+					drm_dbg_atomic(dev, "Failed to configure crc source");
+			}
 		}
+#endif
 	}
 
 	for_each_new_crtc_in_state(state, crtc, new_crtc_state, j)
@@ -9778,22 +11023,44 @@ static void amdgpu_dm_atomic_commit_tail(struct drm_atomic_state *state)
 		dm_new_crtc_state = to_dm_crtc_state(new_crtc_state);
 
 		if (dm_new_crtc_state->stream)
-			amdgpu_dm_commit_planes(state, dc_state, dev,
-						dm, crtc, wait_for_vblank);
+			amdgpu_dm_commit_planes(state, dev, dm, crtc, wait_for_vblank);
+	}
+
+	/* Enable writeback */
+	for_each_new_connector_in_state(state, connector, new_con_state, i) {
+		struct dm_connector_state *dm_new_con_state = to_dm_connector_state(new_con_state);
+		struct amdgpu_crtc *acrtc = to_amdgpu_crtc(dm_new_con_state->base.crtc);
+
+		if (connector->connector_type != DRM_MODE_CONNECTOR_WRITEBACK)
+			continue;
+
+		if (!new_con_state->writeback_job)
+			continue;
+
+		new_crtc_state = drm_atomic_get_new_crtc_state(state, &acrtc->base);
+
+		if (!new_crtc_state)
+			continue;
+
+		if (acrtc->wb_enabled)
+			continue;
+
+		dm_new_crtc_state = to_dm_crtc_state(new_crtc_state);
+
+		dm_set_writeback(dm, dm_new_crtc_state, connector, new_con_state);
+		acrtc->wb_enabled = true;
 	}
 
 	/* Update audio instances for each connector. */
 	amdgpu_dm_commit_audio(dev, state);
 
-#if defined(CONFIG_BACKLIGHT_CLASS_DEVICE) ||		\
-	defined(CONFIG_BACKLIGHT_CLASS_DEVICE_MODULE)
 	/* restore the backlight level */
 	for (i = 0; i < dm->num_of_edps; i++) {
 		if (dm->backlight_dev[i] &&
-		    (amdgpu_dm_backlight_get_level(dm, i) != dm->brightness[i]))
+		    (dm->actual_brightness[i] != dm->brightness[i]))
 			amdgpu_dm_backlight_set_level(dm, i, dm->brightness[i]);
 	}
-#endif
+
 	/*
 	 * send vblank event on all events not handled in flip and
 	 * mark consumed event for drm_atomic_helper_commit_hw_done
@@ -9816,10 +11083,17 @@ static void amdgpu_dm_atomic_commit_tail(struct drm_atomic_state *state)
 
 	drm_atomic_helper_cleanup_planes(dev, state);
 
-	/* return the stolen vga memory back to VRAM */
-	if (!adev->mman.keep_stolen_vga_memory)
-		amdgpu_bo_free_kernel(&adev->mman.stolen_vga_memory, NULL, NULL);
-	amdgpu_bo_free_kernel(&adev->mman.stolen_extended_memory, NULL, NULL);
+	/* Don't free the memory if we are hitting this as part of suspend.
+	 * This way we don't free any memory during suspend; see
+	 * amdgpu_bo_free_kernel().  The memory will be freed in the first
+	 * non-suspend modeset or when the driver is torn down.
+	 */
+	if (!adev->in_suspend) {
+		/* return the stolen vga memory back to VRAM */
+		if (!adev->mman.keep_stolen_vga_memory)
+			amdgpu_bo_free_kernel(&adev->mman.stolen_vga_memory, NULL, NULL);
+		amdgpu_bo_free_kernel(&adev->mman.stolen_extended_memory, NULL, NULL);
+	}
 
 	/*
 	 * Finally, drop a runtime PM reference for each newly disabled CRTC,
@@ -9830,11 +11104,9 @@ static void amdgpu_dm_atomic_commit_tail(struct drm_atomic_state *state)
 		pm_runtime_put_autosuspend(dev->dev);
 	pm_runtime_mark_last_busy(dev->dev);
 
-	if (dc_state_temp)
-		dc_release_state(dc_state_temp);
+	trace_amdgpu_dm_atomic_commit_tail_finish(state);
 }
 
-
 static int dm_force_atomic_commit(struct drm_connector *connector)
 {
 	int ret = 0;
@@ -9858,16 +11130,20 @@ static int dm_force_atomic_commit(struct drm_connector *connector)
 	 */
 	conn_state = drm_atomic_get_connector_state(state, connector);
 
-	ret = PTR_ERR_OR_ZERO(conn_state);
-	if (ret)
+	/* Check for error in getting connector state */
+	if (IS_ERR(conn_state)) {
+		ret = PTR_ERR(conn_state);
 		goto out;
+	}
 
 	/* Attach crtc to drm_atomic_state*/
 	crtc_state = drm_atomic_get_crtc_state(state, &disconnected_acrtc->base);
 
-	ret = PTR_ERR_OR_ZERO(crtc_state);
-	if (ret)
+	/* Check for error in getting crtc state */
+	if (IS_ERR(crtc_state)) {
+		ret = PTR_ERR(crtc_state);
 		goto out;
+	}
 
 	/* force a restore */
 	crtc_state->mode_changed = true;
@@ -9875,9 +11151,11 @@ static int dm_force_atomic_commit(struct drm_connector *connector)
 	/* Attach plane to drm_atomic_state */
 	plane_state = drm_atomic_get_plane_state(state, plane);
 
-	ret = PTR_ERR_OR_ZERO(plane_state);
-	if (ret)
+	/* Check for error in getting plane state */
+	if (IS_ERR(plane_state)) {
+		ret = PTR_ERR(plane_state);
 		goto out;
+	}
 
 	/* Call commit internally with the state we just constructed */
 	ret = drm_atomic_commit(state);
@@ -9885,7 +11163,7 @@ static int dm_force_atomic_commit(struct drm_connector *connector)
 out:
 	drm_atomic_state_put(state);
 	if (ret)
-		DRM_ERROR("Restoring old state failed with %i\n", ret);
+		drm_err(ddev, "Restoring old state failed with %i\n", ret);
 
 	return ret;
 }
@@ -9898,10 +11176,15 @@ out:
 void dm_restore_drm_connector_state(struct drm_device *dev,
 				    struct drm_connector *connector)
 {
-	struct amdgpu_dm_connector *aconnector = to_amdgpu_dm_connector(connector);
+	struct amdgpu_dm_connector *aconnector;
 	struct amdgpu_crtc *disconnected_acrtc;
 	struct dm_crtc_state *acrtc_state;
 
+	if (connector->connector_type == DRM_MODE_CONNECTOR_WRITEBACK)
+		return;
+
+	aconnector = to_amdgpu_dm_connector(connector);
+
 	if (!aconnector->dc_sink || !connector->state || !connector->encoder)
 		return;
 
@@ -9964,8 +11247,8 @@ static int do_aquire_global_lock(struct drm_device *dev,
 					&commit->flip_done, 10*HZ);
 
 		if (ret == 0)
-			DRM_ERROR("[CRTC:%d:%s] hw_done or flip_done "
-				  "timed out\n", crtc->base.id, crtc->name);
+			drm_err(dev, "[CRTC:%d:%s] hw_done or flip_done timed out\n",
+				  crtc->base.id, crtc->name);
 
 		drm_crtc_commit_put(commit);
 	}
@@ -9978,12 +11261,16 @@ static void get_freesync_config_for_crtc(
 	struct dm_connector_state *new_con_state)
 {
 	struct mod_freesync_config config = {0};
-	struct amdgpu_dm_connector *aconnector =
-			to_amdgpu_dm_connector(new_con_state->base.connector);
+	struct amdgpu_dm_connector *aconnector;
 	struct drm_display_mode *mode = &new_crtc_state->base.mode;
 	int vrefresh = drm_mode_vrefresh(mode);
 	bool fs_vid_mode = false;
 
+	if (new_con_state->base.connector->connector_type == DRM_MODE_CONNECTOR_WRITEBACK)
+		return;
+
+	aconnector = to_amdgpu_dm_connector(new_con_state->base.connector);
+
 	new_crtc_state->vrr_supported = new_con_state->freesync_capable &&
 					vrefresh >= aconnector->min_vfreq &&
 					vrefresh <= aconnector->max_vfreq;
@@ -10006,6 +11293,8 @@ static void get_freesync_config_for_crtc(
 		} else {
 			config.state = VRR_STATE_INACTIVE;
 		}
+	} else {
+		config.state = VRR_STATE_UNSUPPORTED;
 	}
 out:
 	new_crtc_state->freesync_config = config;
@@ -10024,34 +11313,35 @@ static bool
 is_timing_unchanged_for_freesync(struct drm_crtc_state *old_crtc_state,
 				 struct drm_crtc_state *new_crtc_state)
 {
-	struct drm_display_mode old_mode, new_mode;
+	const struct drm_display_mode *old_mode, *new_mode;
 
 	if (!old_crtc_state || !new_crtc_state)
 		return false;
 
-	old_mode = old_crtc_state->mode;
-	new_mode = new_crtc_state->mode;
-
-	if (old_mode.clock       == new_mode.clock &&
-	    old_mode.hdisplay    == new_mode.hdisplay &&
-	    old_mode.vdisplay    == new_mode.vdisplay &&
-	    old_mode.htotal      == new_mode.htotal &&
-	    old_mode.vtotal      != new_mode.vtotal &&
-	    old_mode.hsync_start == new_mode.hsync_start &&
-	    old_mode.vsync_start != new_mode.vsync_start &&
-	    old_mode.hsync_end   == new_mode.hsync_end &&
-	    old_mode.vsync_end   != new_mode.vsync_end &&
-	    old_mode.hskew       == new_mode.hskew &&
-	    old_mode.vscan       == new_mode.vscan &&
-	    (old_mode.vsync_end - old_mode.vsync_start) ==
-	    (new_mode.vsync_end - new_mode.vsync_start))
+	old_mode = &old_crtc_state->mode;
+	new_mode = &new_crtc_state->mode;
+
+	if (old_mode->clock       == new_mode->clock &&
+	    old_mode->hdisplay    == new_mode->hdisplay &&
+	    old_mode->vdisplay    == new_mode->vdisplay &&
+	    old_mode->htotal      == new_mode->htotal &&
+	    old_mode->vtotal      != new_mode->vtotal &&
+	    old_mode->hsync_start == new_mode->hsync_start &&
+	    old_mode->vsync_start != new_mode->vsync_start &&
+	    old_mode->hsync_end   == new_mode->hsync_end &&
+	    old_mode->vsync_end   != new_mode->vsync_end &&
+	    old_mode->hskew       == new_mode->hskew &&
+	    old_mode->vscan       == new_mode->vscan &&
+	    (old_mode->vsync_end - old_mode->vsync_start) ==
+	    (new_mode->vsync_end - new_mode->vsync_start))
 		return true;
 
 	return false;
 }
 
-static void set_freesync_fixed_config(struct dm_crtc_state *dm_new_crtc_state) {
-	uint64_t num, den, res;
+static void set_freesync_fixed_config(struct dm_crtc_state *dm_new_crtc_state)
+{
+	u64 num, den, res;
 	struct drm_crtc_state *new_crtc_state = &dm_new_crtc_state->base;
 
 	dm_new_crtc_state->freesync_config.state = VRR_STATE_ACTIVE_FIXED;
@@ -10065,16 +11355,17 @@ static void set_freesync_fixed_config(struct dm_crtc_state *dm_new_crtc_state) {
 }
 
 static int dm_update_crtc_state(struct amdgpu_display_manager *dm,
-				struct drm_atomic_state *state,
-				struct drm_crtc *crtc,
-				struct drm_crtc_state *old_crtc_state,
-				struct drm_crtc_state *new_crtc_state,
-				bool enable,
-				bool *lock_and_validation_needed)
+			 struct drm_atomic_state *state,
+			 struct drm_crtc *crtc,
+			 struct drm_crtc_state *old_crtc_state,
+			 struct drm_crtc_state *new_crtc_state,
+			 bool enable,
+			 bool *lock_and_validation_needed)
 {
 	struct dm_atomic_state *dm_state = NULL;
 	struct dm_crtc_state *dm_old_crtc_state, *dm_new_crtc_state;
 	struct dc_stream_state *new_stream;
+	struct amdgpu_device *adev = dm->adev;
 	int ret = 0;
 
 	/*
@@ -10082,6 +11373,7 @@ static int dm_update_crtc_state(struct amdgpu_display_manager *dm,
 	 * update changed items
 	 */
 	struct amdgpu_crtc *acrtc = NULL;
+	struct drm_connector *connector = NULL;
 	struct amdgpu_dm_connector *aconnector = NULL;
 	struct drm_connector_state *drm_new_conn_state = NULL, *drm_old_conn_state = NULL;
 	struct dm_connector_state *dm_new_conn_state = NULL, *dm_old_conn_state = NULL;
@@ -10091,18 +11383,20 @@ static int dm_update_crtc_state(struct amdgpu_display_manager *dm,
 	dm_old_crtc_state = to_dm_crtc_state(old_crtc_state);
 	dm_new_crtc_state = to_dm_crtc_state(new_crtc_state);
 	acrtc = to_amdgpu_crtc(crtc);
-	aconnector = amdgpu_dm_find_first_crtc_matching_connector(state, crtc);
+	connector = amdgpu_dm_find_first_crtc_matching_connector(state, crtc);
+	if (connector)
+		aconnector = to_amdgpu_dm_connector(connector);
 
 	/* TODO This hack should go away */
-	if (aconnector && enable) {
+	if (connector && enable) {
 		/* Make sure fake sink is created in plug-in scenario */
 		drm_new_conn_state = drm_atomic_get_new_connector_state(state,
-							    &aconnector->base);
+									connector);
 		drm_old_conn_state = drm_atomic_get_old_connector_state(state,
-							    &aconnector->base);
+									connector);
 
-		if (IS_ERR(drm_new_conn_state)) {
-			ret = PTR_ERR_OR_ZERO(drm_new_conn_state);
+		if (WARN_ON(!drm_new_conn_state)) {
+			ret = -EINVAL;
 			goto fail;
 		}
 
@@ -10112,7 +11406,7 @@ static int dm_update_crtc_state(struct amdgpu_display_manager *dm,
 		if (!drm_atomic_crtc_needs_modeset(new_crtc_state))
 			goto skip_modeset;
 
-		new_stream = create_validate_stream_for_sink(aconnector,
+		new_stream = create_validate_stream_for_sink(connector,
 							     &new_crtc_state->mode,
 							     dm_new_conn_state,
 							     dm_old_crtc_state->stream);
@@ -10125,7 +11419,7 @@ static int dm_update_crtc_state(struct amdgpu_display_manager *dm,
 		 */
 
 		if (!new_stream) {
-			DRM_DEBUG_DRIVER("%s: Failed to create new stream for crtc %d\n",
+			drm_dbg_driver(adev_to_drm(adev), "%s: Failed to create new stream for crtc %d\n",
 					__func__, acrtc->base.base.id);
 			ret = -ENOMEM;
 			goto fail;
@@ -10163,7 +11457,7 @@ static int dm_update_crtc_state(struct amdgpu_display_manager *dm,
 		    dc_is_stream_unchanged(new_stream, dm_old_crtc_state->stream) &&
 		    dc_is_stream_scaling_unchanged(new_stream, dm_old_crtc_state->stream)) {
 			new_crtc_state->mode_changed = false;
-			DRM_DEBUG_DRIVER("Mode change not required, setting mode_changed to %d",
+			drm_dbg_driver(adev_to_drm(adev), "Mode change not required, setting mode_changed to %d",
 					 new_crtc_state->mode_changed);
 		}
 	}
@@ -10172,10 +11466,8 @@ static int dm_update_crtc_state(struct amdgpu_display_manager *dm,
 	if (!drm_atomic_crtc_needs_modeset(new_crtc_state))
 		goto skip_modeset;
 
-	DRM_DEBUG_ATOMIC(
-		"amdgpu_crtc id:%d crtc_state_flags: enable:%d, active:%d, "
-		"planes_changed:%d, mode_changed:%d,active_changed:%d,"
-		"connectors_changed:%d\n",
+	drm_dbg_state(state->dev,
+		"amdgpu_crtc id:%d crtc_state_flags: enable:%d, active:%d, planes_changed:%d, mode_changed:%d,active_changed:%d,connectors_changed:%d\n",
 		acrtc->crtc_id,
 		new_crtc_state->enable,
 		new_crtc_state->active,
@@ -10190,13 +11482,21 @@ static int dm_update_crtc_state(struct amdgpu_display_manager *dm,
 		if (!dm_old_crtc_state->stream)
 			goto skip_modeset;
 
+		/* Unset freesync video if it was active before */
+		if (dm_old_crtc_state->freesync_config.state == VRR_STATE_ACTIVE_FIXED) {
+			dm_new_crtc_state->freesync_config.state = VRR_STATE_INACTIVE;
+			dm_new_crtc_state->freesync_config.fixed_refresh_in_uhz = 0;
+		}
+
+		/* Now check if we should set freesync video mode */
 		if (amdgpu_freesync_vid_mode && dm_new_crtc_state->stream &&
+		    dc_is_stream_unchanged(new_stream, dm_old_crtc_state->stream) &&
+		    dc_is_stream_scaling_unchanged(new_stream, dm_old_crtc_state->stream) &&
 		    is_timing_unchanged_for_freesync(new_crtc_state,
 						     old_crtc_state)) {
 			new_crtc_state->mode_changed = false;
-			DRM_DEBUG_DRIVER(
-				"Mode change not required for front porch change, "
-				"setting mode_changed to %d",
+			drm_dbg_driver(adev_to_drm(adev),
+				"Mode change not required for front porch change, setting mode_changed to %d",
 				new_crtc_state->mode_changed);
 
 			set_freesync_fixed_config(dm_new_crtc_state);
@@ -10208,20 +11508,19 @@ static int dm_update_crtc_state(struct amdgpu_display_manager *dm,
 			struct drm_display_mode *high_mode;
 
 			high_mode = get_highest_refresh_rate_mode(aconnector, false);
-			if (!drm_mode_equal(&new_crtc_state->mode, high_mode)) {
+			if (!drm_mode_equal(&new_crtc_state->mode, high_mode))
 				set_freesync_fixed_config(dm_new_crtc_state);
-			}
 		}
 
 		ret = dm_atomic_get_state(state, &dm_state);
 		if (ret)
 			goto fail;
 
-		DRM_DEBUG_DRIVER("Disabling DRM crtc: %d\n",
+		drm_dbg_driver(adev_to_drm(adev), "Disabling DRM crtc: %d\n",
 				crtc->base.id);
 
 		/* i.e. reset mode */
-		if (dc_remove_stream_from_ctx(
+		if (dc_state_remove_stream(
 				dm->dc,
 				dm_state->context,
 				dm_old_crtc_state->stream) != DC_OK) {
@@ -10242,13 +11541,13 @@ static int dm_update_crtc_state(struct amdgpu_display_manager *dm,
 		 * added MST connectors not found in existing crtc_state in the chained mode
 		 * TODO: need to dig out the root cause of that
 		 */
-		if (!aconnector || (!aconnector->dc_sink && aconnector->mst_port))
+		if (!connector)
 			goto skip_modeset;
 
 		if (modereset_required(new_crtc_state))
 			goto skip_modeset;
 
-		if (modeset_required(new_crtc_state, new_stream,
+		if (amdgpu_dm_crtc_modeset_required(new_crtc_state, new_stream,
 				     dm_old_crtc_state->stream)) {
 
 			WARN_ON(dm_new_crtc_state->stream);
@@ -10264,7 +11563,7 @@ static int dm_update_crtc_state(struct amdgpu_display_manager *dm,
 			DRM_DEBUG_ATOMIC("Enabling DRM crtc: %d\n",
 					 crtc->base.id);
 
-			if (dc_add_stream_to_ctx(
+			if (dc_state_add_stream(
 					dm->dc,
 					dm_state->context,
 					dm_new_crtc_state->stream) != DC_OK) {
@@ -10279,13 +11578,13 @@ static int dm_update_crtc_state(struct amdgpu_display_manager *dm,
 skip_modeset:
 	/* Release extra reference */
 	if (new_stream)
-		 dc_stream_release(new_stream);
+		dc_stream_release(new_stream);
 
 	/*
 	 * We want to do dc stream updates that do not require a
 	 * full modeset below.
 	 */
-	if (!(enable && aconnector && new_crtc_state->active))
+	if (!(enable && connector && new_crtc_state->active))
 		return 0;
 	/*
 	 * Given above conditions, the dc state cannot be NULL because:
@@ -10311,8 +11610,9 @@ skip_modeset:
 	 * when a modeset is needed, to ensure it gets reprogrammed.
 	 */
 	if (dm_new_crtc_state->base.color_mgmt_changed ||
+	    dm_old_crtc_state->regamma_tf != dm_new_crtc_state->regamma_tf ||
 	    drm_atomic_crtc_needs_modeset(new_crtc_state)) {
-		ret = amdgpu_dm_update_crtc_color_mgmt(dm_new_crtc_state);
+		ret = amdgpu_dm_check_crtc_color_mgmt(dm_new_crtc_state, true);
 		if (ret)
 			goto fail;
 	}
@@ -10336,15 +11636,20 @@ static bool should_reset_plane(struct drm_atomic_state *state,
 {
 	struct drm_plane *other;
 	struct drm_plane_state *old_other_state, *new_other_state;
-	struct drm_crtc_state *new_crtc_state;
+	struct drm_crtc_state *old_crtc_state, *new_crtc_state;
+	struct dm_crtc_state *old_dm_crtc_state, *new_dm_crtc_state;
+	struct amdgpu_device *adev = drm_to_adev(plane->dev);
 	int i;
 
 	/*
-	 * TODO: Remove this hack once the checks below are sufficient
-	 * enough to determine when we need to reset all the planes on
-	 * the stream.
+	 * TODO: Remove this hack for all asics once it proves that the
+	 * fast updates works fine on DCN3.2+.
 	 */
-	if (state->allow_modeset)
+	if (amdgpu_ip_version(adev, DCE_HWIP, 0) < IP_VERSION(3, 2, 0) &&
+	    state->allow_modeset)
+		return true;
+
+	if (amdgpu_in_reset(adev) && state->allow_modeset)
 		return true;
 
 	/* Exit early if we know that we're adding or removing the plane. */
@@ -10357,14 +11662,38 @@ static bool should_reset_plane(struct drm_atomic_state *state,
 
 	new_crtc_state =
 		drm_atomic_get_new_crtc_state(state, new_plane_state->crtc);
+	old_crtc_state =
+		drm_atomic_get_old_crtc_state(state, old_plane_state->crtc);
 
 	if (!new_crtc_state)
 		return true;
 
+	/*
+	 * A change in cursor mode means a new dc pipe needs to be acquired or
+	 * released from the state
+	 */
+	old_dm_crtc_state = to_dm_crtc_state(old_crtc_state);
+	new_dm_crtc_state = to_dm_crtc_state(new_crtc_state);
+	if (plane->type == DRM_PLANE_TYPE_CURSOR &&
+	    old_dm_crtc_state != NULL &&
+	    old_dm_crtc_state->cursor_mode != new_dm_crtc_state->cursor_mode) {
+		return true;
+	}
+
 	/* CRTC Degamma changes currently require us to recreate planes. */
 	if (new_crtc_state->color_mgmt_changed)
 		return true;
 
+	/*
+	 * On zpos change, planes need to be reordered by removing and re-adding
+	 * them one by one to the dc state, in order of descending zpos.
+	 *
+	 * TODO: We can likely skip bandwidth validation if the only thing that
+	 * changed about the plane was it'z z-ordering.
+	 */
+	if (old_plane_state->normalized_zpos != new_plane_state->normalized_zpos)
+		return true;
+
 	if (drm_atomic_crtc_needs_modeset(new_crtc_state))
 		return true;
 
@@ -10378,6 +11707,11 @@ static bool should_reset_plane(struct drm_atomic_state *state,
 	 */
 	for_each_oldnew_plane_in_state(state, other, old_other_state, new_other_state, i) {
 		struct amdgpu_framebuffer *old_afb, *new_afb;
+		struct dm_plane_state *dm_new_other_state, *dm_old_other_state;
+
+		dm_new_other_state = to_dm_plane_state(new_other_state);
+		dm_old_other_state = to_dm_plane_state(old_other_state);
+
 		if (other->type == DRM_PLANE_TYPE_CURSOR)
 			continue;
 
@@ -10413,6 +11747,18 @@ static bool should_reset_plane(struct drm_atomic_state *state,
 		    old_other_state->color_encoding != new_other_state->color_encoding)
 			return true;
 
+		/* HDR/Transfer Function changes. */
+		if (dm_old_other_state->degamma_tf != dm_new_other_state->degamma_tf ||
+		    dm_old_other_state->degamma_lut != dm_new_other_state->degamma_lut ||
+		    dm_old_other_state->hdr_mult != dm_new_other_state->hdr_mult ||
+		    dm_old_other_state->ctm != dm_new_other_state->ctm ||
+		    dm_old_other_state->shaper_lut != dm_new_other_state->shaper_lut ||
+		    dm_old_other_state->shaper_tf != dm_new_other_state->shaper_tf ||
+		    dm_old_other_state->lut3d != dm_new_other_state->lut3d ||
+		    dm_old_other_state->blend_lut != dm_new_other_state->blend_lut ||
+		    dm_old_other_state->blend_tf != dm_new_other_state->blend_tf)
+			return true;
+
 		/* Framebuffer checks fall at the end. */
 		if (!old_other_state->fb || !new_other_state->fb)
 			continue;
@@ -10476,14 +11822,17 @@ static int dm_check_cursor_fb(struct amdgpu_crtc *new_acrtc,
 	}
 
 	/* Core DRM takes care of checking FB modifiers, so we only need to
-	 * check tiling flags when the FB doesn't have a modifier. */
+	 * check tiling flags when the FB doesn't have a modifier.
+	 */
 	if (!(fb->flags & DRM_MODE_FB_MODIFIERS)) {
-		if (adev->family < AMDGPU_FAMILY_AI) {
+		if (adev->family >= AMDGPU_FAMILY_GC_12_0_0) {
+			linear = AMDGPU_TILING_GET(afb->tiling_flags, GFX12_SWIZZLE_MODE) == 0;
+		} else if (adev->family >= AMDGPU_FAMILY_AI) {
+			linear = AMDGPU_TILING_GET(afb->tiling_flags, SWIZZLE_MODE) == 0;
+		} else {
 			linear = AMDGPU_TILING_GET(afb->tiling_flags, ARRAY_MODE) != DC_ARRAY_2D_TILED_THIN1 &&
-			         AMDGPU_TILING_GET(afb->tiling_flags, ARRAY_MODE) != DC_ARRAY_1D_TILED_THIN1 &&
+				 AMDGPU_TILING_GET(afb->tiling_flags, ARRAY_MODE) != DC_ARRAY_1D_TILED_THIN1 &&
 				 AMDGPU_TILING_GET(afb->tiling_flags, MICRO_TILE_MODE) == 0;
-		} else {
-			linear = AMDGPU_TILING_GET(afb->tiling_flags, SWIZZLE_MODE) == 0;
 		}
 		if (!linear) {
 			DRM_DEBUG_ATOMIC("Cursor FB not linear");
@@ -10494,13 +11843,76 @@ static int dm_check_cursor_fb(struct amdgpu_crtc *new_acrtc,
 	return 0;
 }
 
+/*
+ * Helper function for checking the cursor in native mode
+ */
+static int dm_check_native_cursor_state(struct drm_crtc *new_plane_crtc,
+					struct drm_plane *plane,
+					struct drm_plane_state *new_plane_state,
+					bool enable)
+{
+
+	struct amdgpu_crtc *new_acrtc;
+	int ret;
+
+	if (!enable || !new_plane_crtc ||
+	    drm_atomic_plane_disabling(plane->state, new_plane_state))
+		return 0;
+
+	new_acrtc = to_amdgpu_crtc(new_plane_crtc);
+
+	if (new_plane_state->src_x != 0 || new_plane_state->src_y != 0) {
+		DRM_DEBUG_ATOMIC("Cropping not supported for cursor plane\n");
+		return -EINVAL;
+	}
+
+	if (new_plane_state->fb) {
+		ret = dm_check_cursor_fb(new_acrtc, new_plane_state,
+						new_plane_state->fb);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+static bool dm_should_update_native_cursor(struct drm_atomic_state *state,
+					   struct drm_crtc *old_plane_crtc,
+					   struct drm_crtc *new_plane_crtc,
+					   bool enable)
+{
+	struct drm_crtc_state *old_crtc_state, *new_crtc_state;
+	struct dm_crtc_state *dm_old_crtc_state, *dm_new_crtc_state;
+
+	if (!enable) {
+		if (old_plane_crtc == NULL)
+			return true;
+
+		old_crtc_state = drm_atomic_get_old_crtc_state(
+			state, old_plane_crtc);
+		dm_old_crtc_state = to_dm_crtc_state(old_crtc_state);
+
+		return dm_old_crtc_state->cursor_mode == DM_CURSOR_NATIVE_MODE;
+	} else {
+		if (new_plane_crtc == NULL)
+			return true;
+
+		new_crtc_state = drm_atomic_get_new_crtc_state(
+			state, new_plane_crtc);
+		dm_new_crtc_state = to_dm_crtc_state(new_crtc_state);
+
+		return dm_new_crtc_state->cursor_mode == DM_CURSOR_NATIVE_MODE;
+	}
+}
+
 static int dm_update_plane_state(struct dc *dc,
 				 struct drm_atomic_state *state,
 				 struct drm_plane *plane,
 				 struct drm_plane_state *old_plane_state,
 				 struct drm_plane_state *new_plane_state,
 				 bool enable,
-				 bool *lock_and_validation_needed)
+				 bool *lock_and_validation_needed,
+				 bool *is_top_most_overlay)
 {
 
 	struct dm_atomic_state *dm_state = NULL;
@@ -10508,8 +11920,7 @@ static int dm_update_plane_state(struct dc *dc,
 	struct drm_crtc_state *old_crtc_state, *new_crtc_state;
 	struct dm_crtc_state *dm_new_crtc_state, *dm_old_crtc_state;
 	struct dm_plane_state *dm_new_plane_state, *dm_old_plane_state;
-	struct amdgpu_crtc *new_acrtc;
-	bool needs_reset;
+	bool needs_reset, update_native_cursor;
 	int ret = 0;
 
 
@@ -10518,24 +11929,16 @@ static int dm_update_plane_state(struct dc *dc,
 	dm_new_plane_state = to_dm_plane_state(new_plane_state);
 	dm_old_plane_state = to_dm_plane_state(old_plane_state);
 
-	if (plane->type == DRM_PLANE_TYPE_CURSOR) {
-		if (!enable || !new_plane_crtc ||
-			drm_atomic_plane_disabling(plane->state, new_plane_state))
-			return 0;
-
-		new_acrtc = to_amdgpu_crtc(new_plane_crtc);
+	update_native_cursor = dm_should_update_native_cursor(state,
+							      old_plane_crtc,
+							      new_plane_crtc,
+							      enable);
 
-		if (new_plane_state->src_x != 0 || new_plane_state->src_y != 0) {
-			DRM_DEBUG_ATOMIC("Cropping not supported for cursor plane\n");
-			return -EINVAL;
-		}
-
-		if (new_plane_state->fb) {
-			ret = dm_check_cursor_fb(new_acrtc, new_plane_state,
-						 new_plane_state->fb);
-			if (ret)
-				return ret;
-		}
+	if (plane->type == DRM_PLANE_TYPE_CURSOR && update_native_cursor) {
+		ret = dm_check_native_cursor_state(new_plane_crtc, plane,
+						    new_plane_state, enable);
+		if (ret)
+			return ret;
 
 		return 0;
 	}
@@ -10565,7 +11968,7 @@ static int dm_update_plane_state(struct dc *dc,
 		if (ret)
 			return ret;
 
-		if (!dc_remove_plane_from_context(
+		if (!dc_state_remove_plane(
 				dc,
 				dm_old_crtc_state->stream,
 				dm_old_plane_state->dc_state,
@@ -10574,8 +11977,9 @@ static int dm_update_plane_state(struct dc *dc,
 			return -EINVAL;
 		}
 
+		if (dm_old_plane_state->dc_state)
+			dc_plane_state_release(dm_old_plane_state->dc_state);
 
-		dc_plane_state_release(dm_old_plane_state->dc_state);
 		dm_new_plane_state->dc_state = NULL;
 
 		*lock_and_validation_needed = true;
@@ -10598,15 +12002,17 @@ static int dm_update_plane_state(struct dc *dc,
 		if (!needs_reset)
 			return 0;
 
-		ret = dm_plane_helper_check_state(new_plane_state, new_crtc_state);
+		ret = amdgpu_dm_plane_helper_check_state(new_plane_state, new_crtc_state);
 		if (ret)
-			return ret;
+			goto out;
 
 		WARN_ON(dm_new_plane_state->dc_state);
 
 		dc_new_plane_state = dc_create_plane_state(dc);
-		if (!dc_new_plane_state)
-			return -ENOMEM;
+		if (!dc_new_plane_state) {
+			ret = -ENOMEM;
+			goto out;
+		}
 
 		DRM_DEBUG_ATOMIC("Enabling DRM plane: %d on DRM crtc %d\n",
 				 plane->base.id, new_plane_crtc->base.id);
@@ -10618,13 +12024,13 @@ static int dm_update_plane_state(struct dc *dc,
 			new_crtc_state);
 		if (ret) {
 			dc_plane_state_release(dc_new_plane_state);
-			return ret;
+			goto out;
 		}
 
 		ret = dm_atomic_get_state(state, &dm_state);
 		if (ret) {
 			dc_plane_state_release(dc_new_plane_state);
-			return ret;
+			goto out;
 		}
 
 		/*
@@ -10634,18 +12040,21 @@ static int dm_update_plane_state(struct dc *dc,
 		 * state. It'll be released when the atomic state is
 		 * cleaned.
 		 */
-		if (!dc_add_plane_to_context(
+		if (!dc_state_add_plane(
 				dc,
 				dm_new_crtc_state->stream,
 				dc_new_plane_state,
 				dm_state->context)) {
 
 			dc_plane_state_release(dc_new_plane_state);
-			return -EINVAL;
+			ret = -EINVAL;
+			goto out;
 		}
 
 		dm_new_plane_state->dc_state = dc_new_plane_state;
 
+		dm_new_crtc_state->mpo_requested |= (plane->type == DRM_PLANE_TYPE_OVERLAY);
+
 		/* Tell DC to do a full surface update every time there
 		 * is a plane change. Inefficient, but works for now.
 		 */
@@ -10654,80 +12063,126 @@ static int dm_update_plane_state(struct dc *dc,
 		*lock_and_validation_needed = true;
 	}
 
+out:
+	/* If enabling cursor overlay failed, attempt fallback to native mode */
+	if (enable && ret == -EINVAL && plane->type == DRM_PLANE_TYPE_CURSOR) {
+		ret = dm_check_native_cursor_state(new_plane_crtc, plane,
+						    new_plane_state, enable);
+		if (ret)
+			return ret;
+
+		dm_new_crtc_state->cursor_mode = DM_CURSOR_NATIVE_MODE;
+	}
 
 	return ret;
 }
 
-static int dm_check_crtc_cursor(struct drm_atomic_state *state,
-				struct drm_crtc *crtc,
-				struct drm_crtc_state *new_crtc_state)
+static void dm_get_oriented_plane_size(struct drm_plane_state *plane_state,
+				       int *src_w, int *src_h)
 {
-	struct drm_plane *cursor = crtc->cursor, *underlying;
-	struct drm_plane_state *new_cursor_state, *new_underlying_state;
-	int i;
-	int cursor_scale_w, cursor_scale_h, underlying_scale_w, underlying_scale_h;
+	switch (plane_state->rotation & DRM_MODE_ROTATE_MASK) {
+	case DRM_MODE_ROTATE_90:
+	case DRM_MODE_ROTATE_270:
+		*src_w = plane_state->src_h >> 16;
+		*src_h = plane_state->src_w >> 16;
+		break;
+	case DRM_MODE_ROTATE_0:
+	case DRM_MODE_ROTATE_180:
+	default:
+		*src_w = plane_state->src_w >> 16;
+		*src_h = plane_state->src_h >> 16;
+		break;
+	}
+}
 
-	/* On DCE and DCN there is no dedicated hardware cursor plane. We get a
-	 * cursor per pipe but it's going to inherit the scaling and
-	 * positioning from the underlying pipe. Check the cursor plane's
-	 * blending properties match the underlying planes'. */
+static void
+dm_get_plane_scale(struct drm_plane_state *plane_state,
+		   int *out_plane_scale_w, int *out_plane_scale_h)
+{
+	int plane_src_w, plane_src_h;
 
-	new_cursor_state = drm_atomic_get_new_plane_state(state, cursor);
-	if (!new_cursor_state || !new_cursor_state->fb) {
-		return 0;
-	}
+	dm_get_oriented_plane_size(plane_state, &plane_src_w, &plane_src_h);
+	*out_plane_scale_w = plane_src_w ? plane_state->crtc_w * 1000 / plane_src_w : 0;
+	*out_plane_scale_h = plane_src_h ? plane_state->crtc_h * 1000 / plane_src_h : 0;
+}
 
-	cursor_scale_w = new_cursor_state->crtc_w * 1000 /
-			 (new_cursor_state->src_w >> 16);
-	cursor_scale_h = new_cursor_state->crtc_h * 1000 /
-			 (new_cursor_state->src_h >> 16);
+/*
+ * The normalized_zpos value cannot be used by this iterator directly. It's only
+ * calculated for enabled planes, potentially causing normalized_zpos collisions
+ * between enabled/disabled planes in the atomic state. We need a unique value
+ * so that the iterator will not generate the same object twice, or loop
+ * indefinitely.
+ */
+static inline struct __drm_planes_state *__get_next_zpos(
+	struct drm_atomic_state *state,
+	struct __drm_planes_state *prev)
+{
+	unsigned int highest_zpos = 0, prev_zpos = 256;
+	uint32_t highest_id = 0, prev_id = UINT_MAX;
+	struct drm_plane_state *new_plane_state;
+	struct drm_plane *plane;
+	int i, highest_i = -1;
 
-	for_each_new_plane_in_state_reverse(state, underlying, new_underlying_state, i) {
-		/* Narrow down to non-cursor planes on the same CRTC as the cursor */
-		if (new_underlying_state->crtc != crtc || underlying == crtc->cursor)
-			continue;
+	if (prev != NULL) {
+		prev_zpos = prev->new_state->zpos;
+		prev_id = prev->ptr->base.id;
+	}
 
-		/* Ignore disabled planes */
-		if (!new_underlying_state->fb)
+	for_each_new_plane_in_state(state, plane, new_plane_state, i) {
+		/* Skip planes with higher zpos than the previously returned */
+		if (new_plane_state->zpos > prev_zpos ||
+		    (new_plane_state->zpos == prev_zpos &&
+		     plane->base.id >= prev_id))
 			continue;
 
-		underlying_scale_w = new_underlying_state->crtc_w * 1000 /
-				     (new_underlying_state->src_w >> 16);
-		underlying_scale_h = new_underlying_state->crtc_h * 1000 /
-				     (new_underlying_state->src_h >> 16);
-
-		if (cursor_scale_w != underlying_scale_w ||
-		    cursor_scale_h != underlying_scale_h) {
-			drm_dbg_atomic(crtc->dev,
-				       "Cursor [PLANE:%d:%s] scaling doesn't match underlying [PLANE:%d:%s]\n",
-				       cursor->base.id, cursor->name, underlying->base.id, underlying->name);
-			return -EINVAL;
+		/* Save the index of the plane with highest zpos */
+		if (new_plane_state->zpos > highest_zpos ||
+		    (new_plane_state->zpos == highest_zpos &&
+		     plane->base.id > highest_id)) {
+			highest_zpos = new_plane_state->zpos;
+			highest_id = plane->base.id;
+			highest_i = i;
 		}
-
-		/* If this plane covers the whole CRTC, no need to check planes underneath */
-		if (new_underlying_state->crtc_x <= 0 &&
-		    new_underlying_state->crtc_y <= 0 &&
-		    new_underlying_state->crtc_x + new_underlying_state->crtc_w >= new_crtc_state->mode.hdisplay &&
-		    new_underlying_state->crtc_y + new_underlying_state->crtc_h >= new_crtc_state->mode.vdisplay)
-			break;
 	}
 
-	return 0;
+	if (highest_i < 0)
+		return NULL;
+
+	return &state->planes[highest_i];
 }
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
+/*
+ * Use the uniqueness of the plane's (zpos, drm obj ID) combination to iterate
+ * by descending zpos, as read from the new plane state. This is the same
+ * ordering as defined by drm_atomic_normalize_zpos().
+ */
+#define for_each_oldnew_plane_in_descending_zpos(__state, plane, old_plane_state, new_plane_state) \
+	for (struct __drm_planes_state *__i = __get_next_zpos((__state), NULL); \
+	     __i != NULL; __i = __get_next_zpos((__state), __i))		\
+		for_each_if(((plane) = __i->ptr,				\
+			     (void)(plane) /* Only to avoid unused-but-set-variable warning */, \
+			     (old_plane_state) = __i->old_state,		\
+			     (new_plane_state) = __i->new_state, 1))
+
 static int add_affected_mst_dsc_crtcs(struct drm_atomic_state *state, struct drm_crtc *crtc)
 {
 	struct drm_connector *connector;
-	struct drm_connector_state *conn_state;
+	struct drm_connector_state *conn_state, *old_conn_state;
 	struct amdgpu_dm_connector *aconnector = NULL;
 	int i;
-	for_each_new_connector_in_state(state, connector, conn_state, i) {
+
+	for_each_oldnew_connector_in_state(state, connector, old_conn_state, conn_state, i) {
+		if (!conn_state->crtc)
+			conn_state = old_conn_state;
+
 		if (conn_state->crtc != crtc)
 			continue;
 
+		if (connector->connector_type == DRM_MODE_CONNECTOR_WRITEBACK)
+			continue;
+
 		aconnector = to_amdgpu_dm_connector(connector);
-		if (!aconnector->port || !aconnector->mst_port)
+		if (!aconnector->mst_output_port || !aconnector->mst_root)
 			aconnector = NULL;
 		else
 			break;
@@ -10736,12 +12191,199 @@ static int add_affected_mst_dsc_crtcs(struct drm_atomic_state *state, struct drm
 	if (!aconnector)
 		return 0;
 
-	return drm_dp_mst_add_affected_dsc_crtcs(state, &aconnector->mst_port->mst_mgr);
+	return drm_dp_mst_add_affected_dsc_crtcs(state, &aconnector->mst_root->mst_mgr);
+}
+
+/**
+ * DOC: Cursor Modes - Native vs Overlay
+ *
+ * In native mode, the cursor uses a integrated cursor pipe within each DCN hw
+ * plane. It does not require a dedicated hw plane to enable, but it is
+ * subjected to the same z-order and scaling as the hw plane. It also has format
+ * restrictions, a RGB cursor in native mode cannot be enabled within a non-RGB
+ * hw plane.
+ *
+ * In overlay mode, the cursor uses a separate DCN hw plane, and thus has its
+ * own scaling and z-pos. It also has no blending restrictions. It lends to a
+ * cursor behavior more akin to a DRM client's expectations. However, it does
+ * occupy an extra DCN plane, and therefore will only be used if a DCN plane is
+ * available.
+ */
+
+/**
+ * dm_crtc_get_cursor_mode() - Determine the required cursor mode on crtc
+ * @adev: amdgpu device
+ * @state: DRM atomic state
+ * @dm_crtc_state: amdgpu state for the CRTC containing the cursor
+ * @cursor_mode: Returns the required cursor mode on dm_crtc_state
+ *
+ * Get whether the cursor should be enabled in native mode, or overlay mode, on
+ * the dm_crtc_state.
+ *
+ * The cursor should be enabled in overlay mode if there exists an underlying
+ * plane - on which the cursor may be blended - that is either YUV formatted, or
+ * scaled differently from the cursor.
+ *
+ * Since zpos info is required, drm_atomic_normalize_zpos must be called before
+ * calling this function.
+ *
+ * Return: 0 on success, or an error code if getting the cursor plane state
+ * failed.
+ */
+static int dm_crtc_get_cursor_mode(struct amdgpu_device *adev,
+				   struct drm_atomic_state *state,
+				   struct dm_crtc_state *dm_crtc_state,
+				   enum amdgpu_dm_cursor_mode *cursor_mode)
+{
+	struct drm_plane_state *old_plane_state, *plane_state, *cursor_state;
+	struct drm_crtc_state *crtc_state = &dm_crtc_state->base;
+	struct drm_plane *plane;
+	bool consider_mode_change = false;
+	bool entire_crtc_covered = false;
+	bool cursor_changed = false;
+	int underlying_scale_w, underlying_scale_h;
+	int cursor_scale_w, cursor_scale_h;
+	int i;
+
+	/* Overlay cursor not supported on HW before DCN
+	 * DCN401 does not have the cursor-on-scaled-plane or cursor-on-yuv-plane restrictions
+	 * as previous DCN generations, so enable native mode on DCN401 in addition to DCE
+	 */
+	if (amdgpu_ip_version(adev, DCE_HWIP, 0) == 0 ||
+	    amdgpu_ip_version(adev, DCE_HWIP, 0) == IP_VERSION(4, 0, 1)) {
+		*cursor_mode = DM_CURSOR_NATIVE_MODE;
+		return 0;
+	}
+
+	/* Init cursor_mode to be the same as current */
+	*cursor_mode = dm_crtc_state->cursor_mode;
+
+	/*
+	 * Cursor mode can change if a plane's format changes, scale changes, is
+	 * enabled/disabled, or z-order changes.
+	 */
+	for_each_oldnew_plane_in_state(state, plane, old_plane_state, plane_state, i) {
+		int new_scale_w, new_scale_h, old_scale_w, old_scale_h;
+
+		/* Only care about planes on this CRTC */
+		if ((drm_plane_mask(plane) & crtc_state->plane_mask) == 0)
+			continue;
+
+		if (plane->type == DRM_PLANE_TYPE_CURSOR)
+			cursor_changed = true;
+
+		if (drm_atomic_plane_enabling(old_plane_state, plane_state) ||
+		    drm_atomic_plane_disabling(old_plane_state, plane_state) ||
+		    old_plane_state->fb->format != plane_state->fb->format) {
+			consider_mode_change = true;
+			break;
+		}
+
+		dm_get_plane_scale(plane_state, &new_scale_w, &new_scale_h);
+		dm_get_plane_scale(old_plane_state, &old_scale_w, &old_scale_h);
+		if (new_scale_w != old_scale_w || new_scale_h != old_scale_h) {
+			consider_mode_change = true;
+			break;
+		}
+	}
+
+	if (!consider_mode_change && !crtc_state->zpos_changed)
+		return 0;
+
+	/*
+	 * If no cursor change on this CRTC, and not enabled on this CRTC, then
+	 * no need to set cursor mode. This avoids needlessly locking the cursor
+	 * state.
+	 */
+	if (!cursor_changed &&
+	    !(drm_plane_mask(crtc_state->crtc->cursor) & crtc_state->plane_mask)) {
+		return 0;
+	}
+
+	cursor_state = drm_atomic_get_plane_state(state,
+						  crtc_state->crtc->cursor);
+	if (IS_ERR(cursor_state))
+		return PTR_ERR(cursor_state);
+
+	/* Cursor is disabled */
+	if (!cursor_state->fb)
+		return 0;
+
+	/* For all planes in descending z-order (all of which are below cursor
+	 * as per zpos definitions), check their scaling and format
+	 */
+	for_each_oldnew_plane_in_descending_zpos(state, plane, old_plane_state, plane_state) {
+
+		/* Only care about non-cursor planes on this CRTC */
+		if ((drm_plane_mask(plane) & crtc_state->plane_mask) == 0 ||
+		    plane->type == DRM_PLANE_TYPE_CURSOR)
+			continue;
+
+		/* Underlying plane is YUV format - use overlay cursor */
+		if (amdgpu_dm_plane_is_video_format(plane_state->fb->format->format)) {
+			*cursor_mode = DM_CURSOR_OVERLAY_MODE;
+			return 0;
+		}
+
+		dm_get_plane_scale(plane_state,
+				   &underlying_scale_w, &underlying_scale_h);
+		dm_get_plane_scale(cursor_state,
+				   &cursor_scale_w, &cursor_scale_h);
+
+		/* Underlying plane has different scale - use overlay cursor */
+		if (cursor_scale_w != underlying_scale_w &&
+		    cursor_scale_h != underlying_scale_h) {
+			*cursor_mode = DM_CURSOR_OVERLAY_MODE;
+			return 0;
+		}
+
+		/* If this plane covers the whole CRTC, no need to check planes underneath */
+		if (plane_state->crtc_x <= 0 && plane_state->crtc_y <= 0 &&
+		    plane_state->crtc_x + plane_state->crtc_w >= crtc_state->mode.hdisplay &&
+		    plane_state->crtc_y + plane_state->crtc_h >= crtc_state->mode.vdisplay) {
+			entire_crtc_covered = true;
+			break;
+		}
+	}
+
+	/* If planes do not cover the entire CRTC, use overlay mode to enable
+	 * cursor over holes
+	 */
+	if (entire_crtc_covered)
+		*cursor_mode = DM_CURSOR_NATIVE_MODE;
+	else
+		*cursor_mode = DM_CURSOR_OVERLAY_MODE;
+
+	return 0;
+}
+
+static bool amdgpu_dm_crtc_mem_type_changed(struct drm_device *dev,
+					    struct drm_atomic_state *state,
+					    struct drm_crtc_state *crtc_state)
+{
+	struct drm_plane *plane;
+	struct drm_plane_state *new_plane_state, *old_plane_state;
+
+	drm_for_each_plane_mask(plane, dev, crtc_state->plane_mask) {
+		new_plane_state = drm_atomic_get_plane_state(state, plane);
+		old_plane_state = drm_atomic_get_plane_state(state, plane);
+
+		if (IS_ERR(new_plane_state) || IS_ERR(old_plane_state)) {
+			drm_err(dev, "Failed to get plane state for plane %s\n", plane->name);
+			return false;
+		}
+
+		if (old_plane_state->fb && new_plane_state->fb &&
+		    get_mem_type(old_plane_state->fb) != get_mem_type(new_plane_state->fb))
+			return true;
+	}
+
+	return false;
 }
-#endif
 
 /**
  * amdgpu_dm_atomic_check() - Atomic check implementation for AMDgpu DM.
+ *
  * @dev: The DRM device
  * @state: The atomic state to commit
  *
@@ -10775,22 +12417,23 @@ static int amdgpu_dm_atomic_check(struct drm_device *dev,
 	struct drm_crtc *crtc;
 	struct drm_crtc_state *old_crtc_state, *new_crtc_state;
 	struct drm_plane *plane;
-	struct drm_plane_state *old_plane_state, *new_plane_state;
+	struct drm_plane_state *old_plane_state, *new_plane_state, *new_cursor_state;
 	enum dc_status status;
 	int ret, i;
 	bool lock_and_validation_needed = false;
-	struct dm_crtc_state *dm_old_crtc_state;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	struct dsc_mst_fairness_vars vars[MAX_PIPES];
-	struct drm_dp_mst_topology_state *mst_state;
+	bool is_top_most_overlay = true;
+	struct dm_crtc_state *dm_old_crtc_state, *dm_new_crtc_state;
 	struct drm_dp_mst_topology_mgr *mgr;
-#endif
+	struct drm_dp_mst_topology_state *mst_state;
+	struct dsc_mst_fairness_vars vars[MAX_PIPES] = {0};
 
 	trace_amdgpu_dm_atomic_check_begin(state);
 
 	ret = drm_atomic_helper_check_modeset(dev, state);
-	if (ret)
+	if (ret) {
+		drm_dbg_atomic(dev, "drm_atomic_helper_check_modeset() failed\n");
 		goto fail;
+	}
 
 	/* Check connector changes */
 	for_each_oldnew_connector_in_state(state, connector, old_con_state, new_con_state, i) {
@@ -10798,34 +12441,32 @@ static int amdgpu_dm_atomic_check(struct drm_device *dev,
 		struct dm_connector_state *dm_new_con_state = to_dm_connector_state(new_con_state);
 
 		/* Skip connectors that are disabled or part of modeset already. */
-		if (!old_con_state->crtc && !new_con_state->crtc)
-			continue;
-
 		if (!new_con_state->crtc)
 			continue;
 
 		new_crtc_state = drm_atomic_get_crtc_state(state, new_con_state->crtc);
 		if (IS_ERR(new_crtc_state)) {
+			drm_dbg_atomic(dev, "drm_atomic_get_crtc_state() failed\n");
 			ret = PTR_ERR(new_crtc_state);
 			goto fail;
 		}
 
-		if (dm_old_con_state->abm_level !=
-		    dm_new_con_state->abm_level)
+		if (dm_old_con_state->abm_level != dm_new_con_state->abm_level ||
+		    dm_old_con_state->scaling != dm_new_con_state->scaling)
 			new_crtc_state->connectors_changed = true;
 	}
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	if (dc_resource_is_dsc_encoding_supported(dc)) {
 		for_each_oldnew_crtc_in_state(state, crtc, old_crtc_state, new_crtc_state, i) {
 			if (drm_atomic_crtc_needs_modeset(new_crtc_state)) {
 				ret = add_affected_mst_dsc_crtcs(state, crtc);
-				if (ret)
+				if (ret) {
+					drm_dbg_atomic(dev, "add_affected_mst_dsc_crtcs() failed\n");
 					goto fail;
+				}
 			}
 		}
 	}
-#endif
 	for_each_oldnew_crtc_in_state(state, crtc, old_crtc_state, new_crtc_state, i) {
 		dm_old_crtc_state = to_dm_crtc_state(old_crtc_state);
 
@@ -10836,19 +12477,25 @@ static int amdgpu_dm_atomic_check(struct drm_device *dev,
 			continue;
 
 		ret = amdgpu_dm_verify_lut_sizes(new_crtc_state);
-		if (ret)
+		if (ret) {
+			drm_dbg_atomic(dev, "amdgpu_dm_verify_lut_sizes() failed\n");
 			goto fail;
+		}
 
 		if (!new_crtc_state->enable)
 			continue;
 
 		ret = drm_atomic_add_affected_connectors(state, crtc);
-		if (ret)
+		if (ret) {
+			drm_dbg_atomic(dev, "drm_atomic_add_affected_connectors() failed\n");
 			goto fail;
+		}
 
 		ret = drm_atomic_add_affected_planes(state, crtc);
-		if (ret)
+		if (ret) {
+			drm_dbg_atomic(dev, "drm_atomic_add_affected_planes() failed\n");
 			goto fail;
+		}
 
 		if (dm_old_crtc_state->dsc_force_changed)
 			new_crtc_state->mode_changed = true;
@@ -10885,20 +12532,63 @@ static int amdgpu_dm_atomic_check(struct drm_device *dev,
 
 			if (IS_ERR(new_plane_state)) {
 				ret = PTR_ERR(new_plane_state);
+				drm_dbg_atomic(dev, "new_plane_state is BAD\n");
 				goto fail;
 			}
 		}
 	}
 
+	/*
+	 * DC consults the zpos (layer_index in DC terminology) to determine the
+	 * hw plane on which to enable the hw cursor (see
+	 * `dcn10_can_pipe_disable_cursor`). By now, all modified planes are in
+	 * atomic state, so call drm helper to normalize zpos.
+	 */
+	ret = drm_atomic_normalize_zpos(dev, state);
+	if (ret) {
+		drm_dbg(dev, "drm_atomic_normalize_zpos() failed\n");
+		goto fail;
+	}
+
+	/*
+	 * Determine whether cursors on each CRTC should be enabled in native or
+	 * overlay mode.
+	 */
+	for_each_new_crtc_in_state(state, crtc, new_crtc_state, i) {
+		dm_new_crtc_state = to_dm_crtc_state(new_crtc_state);
+
+		ret = dm_crtc_get_cursor_mode(adev, state, dm_new_crtc_state,
+					      &dm_new_crtc_state->cursor_mode);
+		if (ret) {
+			drm_dbg(dev, "Failed to determine cursor mode\n");
+			goto fail;
+		}
+
+		/*
+		 * If overlay cursor is needed, DC cannot go through the
+		 * native cursor update path. All enabled planes on the CRTC
+		 * need to be added for DC to not disable a plane by mistake
+		 */
+		if (dm_new_crtc_state->cursor_mode == DM_CURSOR_OVERLAY_MODE) {
+			ret = drm_atomic_add_affected_planes(state, crtc);
+			if (ret)
+				goto fail;
+		}
+	}
+
 	/* Remove exiting planes if they are modified */
-	for_each_oldnew_plane_in_state_reverse(state, plane, old_plane_state, new_plane_state, i) {
+	for_each_oldnew_plane_in_descending_zpos(state, plane, old_plane_state, new_plane_state) {
+
 		ret = dm_update_plane_state(dc, state, plane,
 					    old_plane_state,
 					    new_plane_state,
 					    false,
-					    &lock_and_validation_needed);
-		if (ret)
+					    &lock_and_validation_needed,
+					    &is_top_most_overlay);
+		if (ret) {
+			drm_dbg_atomic(dev, "dm_update_plane_state() failed\n");
 			goto fail;
+		}
 	}
 
 	/* Disable all crtcs which require disable */
@@ -10908,8 +12598,10 @@ static int amdgpu_dm_atomic_check(struct drm_device *dev,
 					   new_crtc_state,
 					   false,
 					   &lock_and_validation_needed);
-		if (ret)
+		if (ret) {
+			drm_dbg_atomic(dev, "DISABLE: dm_update_crtc_state() failed\n");
 			goto fail;
+		}
 	}
 
 	/* Enable all crtcs which require enable */
@@ -10919,31 +12611,92 @@ static int amdgpu_dm_atomic_check(struct drm_device *dev,
 					   new_crtc_state,
 					   true,
 					   &lock_and_validation_needed);
-		if (ret)
+		if (ret) {
+			drm_dbg_atomic(dev, "ENABLE: dm_update_crtc_state() failed\n");
 			goto fail;
+		}
 	}
 
 	/* Add new/modified planes */
-	for_each_oldnew_plane_in_state_reverse(state, plane, old_plane_state, new_plane_state, i) {
+	for_each_oldnew_plane_in_descending_zpos(state, plane, old_plane_state, new_plane_state) {
 		ret = dm_update_plane_state(dc, state, plane,
 					    old_plane_state,
 					    new_plane_state,
 					    true,
-					    &lock_and_validation_needed);
-		if (ret)
+					    &lock_and_validation_needed,
+					    &is_top_most_overlay);
+		if (ret) {
+			drm_dbg_atomic(dev, "dm_update_plane_state() failed\n");
 			goto fail;
+		}
 	}
 
+#if defined(CONFIG_DRM_AMD_DC_FP)
+	if (dc_resource_is_dsc_encoding_supported(dc)) {
+		ret = pre_validate_dsc(state, &dm_state, vars);
+		if (ret != 0)
+			goto fail;
+	}
+#endif
+
 	/* Run this here since we want to validate the streams we created */
 	ret = drm_atomic_helper_check_planes(dev, state);
-	if (ret)
+	if (ret) {
+		drm_dbg_atomic(dev, "drm_atomic_helper_check_planes() failed\n");
 		goto fail;
+	}
 
-	/* Check cursor planes scaling */
 	for_each_new_crtc_in_state(state, crtc, new_crtc_state, i) {
-		ret = dm_check_crtc_cursor(state, crtc, new_crtc_state);
-		if (ret)
+		dm_new_crtc_state = to_dm_crtc_state(new_crtc_state);
+		if (dm_new_crtc_state->mpo_requested)
+			drm_dbg_atomic(dev, "MPO enablement requested on crtc:[%p]\n", crtc);
+	}
+
+	/* Check cursor restrictions */
+	for_each_new_crtc_in_state(state, crtc, new_crtc_state, i) {
+		enum amdgpu_dm_cursor_mode required_cursor_mode;
+		int is_rotated, is_scaled;
+
+		/* Overlay cusor not subject to native cursor restrictions */
+		dm_new_crtc_state = to_dm_crtc_state(new_crtc_state);
+		if (dm_new_crtc_state->cursor_mode == DM_CURSOR_OVERLAY_MODE)
+			continue;
+
+		/* Check if rotation or scaling is enabled on DCN401 */
+		if ((drm_plane_mask(crtc->cursor) & new_crtc_state->plane_mask) &&
+		    amdgpu_ip_version(adev, DCE_HWIP, 0) == IP_VERSION(4, 0, 1)) {
+			new_cursor_state = drm_atomic_get_new_plane_state(state, crtc->cursor);
+
+			is_rotated = new_cursor_state &&
+				((new_cursor_state->rotation & DRM_MODE_ROTATE_MASK) != DRM_MODE_ROTATE_0);
+			is_scaled = new_cursor_state && ((new_cursor_state->src_w >> 16 != new_cursor_state->crtc_w) ||
+				(new_cursor_state->src_h >> 16 != new_cursor_state->crtc_h));
+
+			if (is_rotated || is_scaled) {
+				drm_dbg_driver(
+					crtc->dev,
+					"[CRTC:%d:%s] cannot enable hardware cursor due to rotation/scaling\n",
+					crtc->base.id, crtc->name);
+				ret = -EINVAL;
+				goto fail;
+			}
+		}
+
+		/* If HW can only do native cursor, check restrictions again */
+		ret = dm_crtc_get_cursor_mode(adev, state, dm_new_crtc_state,
+					      &required_cursor_mode);
+		if (ret) {
+			drm_dbg_driver(crtc->dev,
+				       "[CRTC:%d:%s] Checking cursor mode failed\n",
+				       crtc->base.id, crtc->name);
+			goto fail;
+		} else if (required_cursor_mode == DM_CURSOR_OVERLAY_MODE) {
+			drm_dbg_driver(crtc->dev,
+				       "[CRTC:%d:%s] Cannot enable native cursor due to scaling or YUV restrictions\n",
+				       crtc->base.id, crtc->name);
+			ret = -EINVAL;
 			goto fail;
+		}
 	}
 
 	if (state->legacy_cursor_update) {
@@ -10988,7 +12741,6 @@ static int amdgpu_dm_atomic_check(struct drm_device *dev,
 		lock_and_validation_needed = true;
 	}
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	/* set the slot info for each mst_state based on the link encoding format */
 	for_each_new_mst_mgr_in_state(state, mgr, mst_state, i) {
 		struct amdgpu_dm_connector *aconnector;
@@ -10996,14 +12748,9 @@ static int amdgpu_dm_atomic_check(struct drm_device *dev,
 		struct drm_connector_list_iter iter;
 		u8 link_coding_cap;
 
-		if (!mgr->mst_state )
-			continue;
-
 		drm_connector_list_iter_begin(dev, &iter);
 		drm_for_each_connector_iter(connector, &iter) {
-			int id = connector->index;
-
-			if (id == mst_state->mgr->conn_base_id) {
+			if (connector->index == mst_state->mgr->conn_base_id) {
 				aconnector = to_amdgpu_dm_connector(connector);
 				link_coding_cap = dc_link_dp_mst_decide_link_encoding_format(aconnector->dc_link);
 				drm_dp_mst_update_slots(mst_state, link_coding_cap);
@@ -11012,9 +12759,8 @@ static int amdgpu_dm_atomic_check(struct drm_device *dev,
 			}
 		}
 		drm_connector_list_iter_end(&iter);
-
 	}
-#endif
+
 	/**
 	 * Streams and planes are reset when there are changes that affect
 	 * bandwidth. Anything that affects bandwidth needs to go through
@@ -11030,21 +12776,33 @@ static int amdgpu_dm_atomic_check(struct drm_device *dev,
 	 */
 	if (lock_and_validation_needed) {
 		ret = dm_atomic_get_state(state, &dm_state);
-		if (ret)
+		if (ret) {
+			drm_dbg_atomic(dev, "dm_atomic_get_state() failed\n");
 			goto fail;
+		}
 
 		ret = do_aquire_global_lock(dev, state);
-		if (ret)
+		if (ret) {
+			drm_dbg_atomic(dev, "do_aquire_global_lock() failed\n");
 			goto fail;
+		}
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-		if (!compute_mst_dsc_configs_for_state(state, dm_state->context, vars))
-			goto fail;
+#if defined(CONFIG_DRM_AMD_DC_FP)
+		if (dc_resource_is_dsc_encoding_supported(dc)) {
+			ret = compute_mst_dsc_configs_for_state(state, dm_state->context, vars);
+			if (ret) {
+				drm_dbg_atomic(dev, "MST_DSC compute_mst_dsc_configs_for_state() failed\n");
+				ret = -EINVAL;
+				goto fail;
+			}
+		}
+#endif
 
 		ret = dm_update_mst_vcpi_slots_for_dsc(state, dm_state->context, vars);
-		if (ret)
+		if (ret) {
+			drm_dbg_atomic(dev, "dm_update_mst_vcpi_slots_for_dsc() failed\n");
 			goto fail;
-#endif
+		}
 
 		/*
 		 * Perform validation of MST topology in the state:
@@ -11053,12 +12811,13 @@ static int amdgpu_dm_atomic_check(struct drm_device *dev,
 		 * to get stuck in an infinite loop and hang eventually.
 		 */
 		ret = drm_dp_mst_atomic_check(state);
-		if (ret)
+		if (ret) {
+			drm_dbg_atomic(dev, "MST drm_dp_mst_atomic_check() failed\n");
 			goto fail;
-		status = dc_validate_global_state(dc, dm_state->context, false);
+		}
+		status = dc_validate_global_state(dc, dm_state->context, DC_VALIDATE_MODE_ONLY);
 		if (status != DC_OK) {
-			drm_dbg_atomic(dev,
-				       "DC global validation failure: %s (%d)",
+			drm_dbg_atomic(dev, "DC global validation failure: %s (%d)",
 				       dc_status_to_str(status), status);
 			ret = -EINVAL;
 			goto fail;
@@ -11085,7 +12844,7 @@ static int amdgpu_dm_atomic_check(struct drm_device *dev,
 				int j = state->num_private_objs-1;
 
 				dm_atomic_destroy_state(obj,
-						state->private_objs[i].state);
+						state->private_objs[i].state_to_destroy);
 
 				/* If i is not at the end of the array then the
 				 * last element needs to be moved to where i was
@@ -11096,7 +12855,7 @@ static int amdgpu_dm_atomic_check(struct drm_device *dev,
 						state->private_objs[j];
 
 				state->private_objs[j].ptr = NULL;
-				state->private_objs[j].state = NULL;
+				state->private_objs[j].state_to_destroy = NULL;
 				state->private_objs[j].old_state = NULL;
 				state->private_objs[j].new_state = NULL;
 
@@ -11107,13 +12866,26 @@ static int amdgpu_dm_atomic_check(struct drm_device *dev,
 	}
 
 	/* Store the overall update type for use later in atomic check. */
-	for_each_new_crtc_in_state (state, crtc, new_crtc_state, i) {
+	for_each_new_crtc_in_state(state, crtc, new_crtc_state, i) {
 		struct dm_crtc_state *dm_new_crtc_state =
 			to_dm_crtc_state(new_crtc_state);
 
+		/*
+		 * Only allow async flips for fast updates that don't change
+		 * the FB pitch, the DCC state, rotation, mem_type, etc.
+		 */
+		if (new_crtc_state->async_flip &&
+		    (lock_and_validation_needed ||
+		     amdgpu_dm_crtc_mem_type_changed(dev, state, new_crtc_state))) {
+			drm_dbg_atomic(crtc->dev,
+				       "[CRTC:%d:%s] async flips are only supported for fast updates\n",
+				       crtc->base.id, crtc->name);
+			ret = -EINVAL;
+			goto fail;
+		}
+
 		dm_new_crtc_state->update_type = lock_and_validation_needed ?
-							 UPDATE_TYPE_FULL :
-							 UPDATE_TYPE_FAST;
+			UPDATE_TYPE_FULL : UPDATE_TYPE_FAST;
 	}
 
 	/* Must be success */
@@ -11125,40 +12897,21 @@ static int amdgpu_dm_atomic_check(struct drm_device *dev,
 
 fail:
 	if (ret == -EDEADLK)
-		DRM_DEBUG_DRIVER("Atomic check stopped to avoid deadlock.\n");
+		drm_dbg_atomic(dev, "Atomic check stopped to avoid deadlock.\n");
 	else if (ret == -EINTR || ret == -EAGAIN || ret == -ERESTARTSYS)
-		DRM_DEBUG_DRIVER("Atomic check stopped due to signal.\n");
+		drm_dbg_atomic(dev, "Atomic check stopped due to signal.\n");
 	else
-		DRM_DEBUG_DRIVER("Atomic check failed with err: %d \n", ret);
+		drm_dbg_atomic(dev, "Atomic check failed with err: %d\n", ret);
 
 	trace_amdgpu_dm_atomic_check_finish(state, ret);
 
 	return ret;
 }
 
-static bool is_dp_capable_without_timing_msa(struct dc *dc,
-					     struct amdgpu_dm_connector *amdgpu_dm_connector)
-{
-	uint8_t dpcd_data;
-	bool capable = false;
-
-	if (amdgpu_dm_connector->dc_link &&
-		dm_helpers_dp_read_dpcd(
-				NULL,
-				amdgpu_dm_connector->dc_link,
-				DP_DOWN_STREAM_PORT_COUNT,
-				&dpcd_data,
-				sizeof(dpcd_data))) {
-		capable = (dpcd_data & DP_MSA_TIMING_PAR_IGNORED) ? true:false;
-	}
-
-	return capable;
-}
-
 static bool dm_edid_parser_send_cea(struct amdgpu_display_manager *dm,
 		unsigned int offset,
 		unsigned int total_length,
-		uint8_t *data,
+		u8 *data,
 		unsigned int length,
 		struct amdgpu_hdmi_vsdb_info *vsdb)
 {
@@ -11180,12 +12933,12 @@ static bool dm_edid_parser_send_cea(struct amdgpu_display_manager *dm,
 		sizeof(cmd.edid_cea) - sizeof(cmd.edid_cea.header);
 	input->offset = offset;
 	input->length = length;
-	input->total_length = total_length;
+	input->cea_total_length = total_length;
 	memcpy(input->payload, data, length);
 
-	res = dc_dmub_srv_cmd_with_reply_data(dm->dc->ctx->dmub_srv, &cmd);
+	res = dc_wake_and_execute_dmub_cmd(dm->dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY);
 	if (!res) {
-		DRM_ERROR("EDID CEA parser failed\n");
+		drm_err(adev_to_drm(dm->adev), "EDID CEA parser failed\n");
 		return false;
 	}
 
@@ -11193,7 +12946,7 @@ static bool dm_edid_parser_send_cea(struct amdgpu_display_manager *dm,
 
 	if (output->type == DMUB_CMD__EDID_CEA_ACK) {
 		if (!output->ack.success) {
-			DRM_ERROR("EDID CEA ack failed at offset %d\n",
+			drm_err(adev_to_drm(dm->adev), "EDID CEA ack failed at offset %d\n",
 					output->ack.offset);
 		}
 	} else if (output->type == DMUB_CMD__EDID_CEA_AMD_VSDB) {
@@ -11205,7 +12958,7 @@ static bool dm_edid_parser_send_cea(struct amdgpu_display_manager *dm,
 		vsdb->min_refresh_rate_hz = output->amd_vsdb.min_frame_rate;
 		vsdb->max_refresh_rate_hz = output->amd_vsdb.max_frame_rate;
 	} else {
-		DRM_WARN("Unknown EDID CEA parser results\n");
+		drm_warn(adev_to_drm(dm->adev), "Unknown EDID CEA parser results\n");
 		return false;
 	}
 
@@ -11213,7 +12966,7 @@ static bool dm_edid_parser_send_cea(struct amdgpu_display_manager *dm,
 }
 
 static bool parse_edid_cea_dmcu(struct amdgpu_display_manager *dm,
-		uint8_t *edid_ext, int len,
+		u8 *edid_ext, int len,
 		struct amdgpu_hdmi_vsdb_info *vsdb_info)
 {
 	int i;
@@ -11254,7 +13007,7 @@ static bool parse_edid_cea_dmcu(struct amdgpu_display_manager *dm,
 }
 
 static bool parse_edid_cea_dmub(struct amdgpu_display_manager *dm,
-		uint8_t *edid_ext, int len,
+		u8 *edid_ext, int len,
 		struct amdgpu_hdmi_vsdb_info *vsdb_info)
 {
 	int i;
@@ -11270,21 +13023,104 @@ static bool parse_edid_cea_dmub(struct amdgpu_display_manager *dm,
 }
 
 static bool parse_edid_cea(struct amdgpu_dm_connector *aconnector,
-		uint8_t *edid_ext, int len,
+		u8 *edid_ext, int len,
 		struct amdgpu_hdmi_vsdb_info *vsdb_info)
 {
 	struct amdgpu_device *adev = drm_to_adev(aconnector->base.dev);
+	bool ret;
 
+	mutex_lock(&adev->dm.dc_lock);
 	if (adev->dm.dmub_srv)
-		return parse_edid_cea_dmub(&adev->dm, edid_ext, len, vsdb_info);
+		ret = parse_edid_cea_dmub(&adev->dm, edid_ext, len, vsdb_info);
 	else
-		return parse_edid_cea_dmcu(&adev->dm, edid_ext, len, vsdb_info);
+		ret = parse_edid_cea_dmcu(&adev->dm, edid_ext, len, vsdb_info);
+	mutex_unlock(&adev->dm.dc_lock);
+	return ret;
+}
+
+static void parse_edid_displayid_vrr(struct drm_connector *connector,
+				     const struct edid *edid)
+{
+	u8 *edid_ext = NULL;
+	int i;
+	int j = 0;
+	u16 min_vfreq;
+	u16 max_vfreq;
+
+	if (edid == NULL || edid->extensions == 0)
+		return;
+
+	/* Find DisplayID extension */
+	for (i = 0; i < edid->extensions; i++) {
+		edid_ext = (void *)(edid + (i + 1));
+		if (edid_ext[0] == DISPLAYID_EXT)
+			break;
+	}
+
+	if (edid_ext == NULL)
+		return;
+
+	while (j < EDID_LENGTH) {
+		/* Get dynamic video timing range from DisplayID if available */
+		if (EDID_LENGTH - j > 13 && edid_ext[j] == 0x25	&&
+		    (edid_ext[j+1] & 0xFE) == 0 && (edid_ext[j+2] == 9)) {
+			min_vfreq = edid_ext[j+9];
+			if (edid_ext[j+1] & 7)
+				max_vfreq = edid_ext[j+10] + ((edid_ext[j+11] & 3) << 8);
+			else
+				max_vfreq = edid_ext[j+10];
+
+			if (max_vfreq && min_vfreq) {
+				connector->display_info.monitor_range.max_vfreq = max_vfreq;
+				connector->display_info.monitor_range.min_vfreq = min_vfreq;
+
+				return;
+			}
+		}
+		j++;
+	}
+}
+
+static int parse_amd_vsdb(struct amdgpu_dm_connector *aconnector,
+			  const struct edid *edid, struct amdgpu_hdmi_vsdb_info *vsdb_info)
+{
+	u8 *edid_ext = NULL;
+	int i;
+	int j = 0;
+
+	if (edid == NULL || edid->extensions == 0)
+		return -ENODEV;
+
+	/* Find DisplayID extension */
+	for (i = 0; i < edid->extensions; i++) {
+		edid_ext = (void *)(edid + (i + 1));
+		if (edid_ext[0] == DISPLAYID_EXT)
+			break;
+	}
+
+	while (j < EDID_LENGTH - sizeof(struct amd_vsdb_block)) {
+		struct amd_vsdb_block *amd_vsdb = (struct amd_vsdb_block *)&edid_ext[j];
+		unsigned int ieeeId = (amd_vsdb->ieee_id[2] << 16) | (amd_vsdb->ieee_id[1] << 8) | (amd_vsdb->ieee_id[0]);
+
+		if (ieeeId == HDMI_AMD_VENDOR_SPECIFIC_DATA_BLOCK_IEEE_REGISTRATION_ID &&
+				amd_vsdb->version == HDMI_AMD_VENDOR_SPECIFIC_DATA_BLOCK_VERSION_3) {
+			vsdb_info->replay_mode = (amd_vsdb->feature_caps & AMD_VSDB_VERSION_3_FEATURECAP_REPLAYMODE) ? true : false;
+			vsdb_info->amd_vsdb_version = HDMI_AMD_VENDOR_SPECIFIC_DATA_BLOCK_VERSION_3;
+			DRM_DEBUG_KMS("Panel supports Replay Mode: %d\n", vsdb_info->replay_mode);
+
+			return true;
+		}
+		j++;
+	}
+
+	return false;
 }
 
 static int parse_hdmi_amd_vsdb(struct amdgpu_dm_connector *aconnector,
-		struct edid *edid, struct amdgpu_hdmi_vsdb_info *vsdb_info)
+			       const struct edid *edid,
+			       struct amdgpu_hdmi_vsdb_info *vsdb_info)
 {
-	uint8_t *edid_ext = NULL;
+	u8 *edid_ext = NULL;
 	int i;
 	bool valid_vsdb_found = false;
 
@@ -11312,25 +13148,33 @@ static int parse_hdmi_amd_vsdb(struct amdgpu_dm_connector *aconnector,
 	return valid_vsdb_found ? i : -ENODEV;
 }
 
+/**
+ * amdgpu_dm_update_freesync_caps - Update Freesync capabilities
+ *
+ * @connector: Connector to query.
+ * @drm_edid: DRM EDID from monitor
+ *
+ * Amdgpu supports Freesync in DP and HDMI displays, and it is required to keep
+ * track of some of the display information in the internal data struct used by
+ * amdgpu_dm. This function checks which type of connector we need to set the
+ * FreeSync parameters.
+ */
 void amdgpu_dm_update_freesync_caps(struct drm_connector *connector,
-					struct edid *edid)
+				    const struct drm_edid *drm_edid)
 {
 	int i = 0;
-	struct detailed_timing *timing;
-	struct detailed_non_pixel *data;
-	struct detailed_data_monitor_range *range;
 	struct amdgpu_dm_connector *amdgpu_dm_connector =
 			to_amdgpu_dm_connector(connector);
 	struct dm_connector_state *dm_con_state = NULL;
 	struct dc_sink *sink;
-
-	struct drm_device *dev = connector->dev;
-	struct amdgpu_device *adev = drm_to_adev(dev);
-	bool freesync_capable = false;
+	struct amdgpu_device *adev = drm_to_adev(connector->dev);
 	struct amdgpu_hdmi_vsdb_info vsdb_info = {0};
+	const struct edid *edid;
+	bool freesync_capable = false;
+	enum adaptive_sync_type as_type = ADAPTIVE_SYNC_TYPE_NONE;
 
 	if (!connector->state) {
-		DRM_ERROR("%s - Connector has no state", __func__);
+		drm_err(adev_to_drm(adev), "%s - Connector has no state", __func__);
 		goto update;
 	}
 
@@ -11338,14 +13182,13 @@ void amdgpu_dm_update_freesync_caps(struct drm_connector *connector,
 		amdgpu_dm_connector->dc_sink :
 		amdgpu_dm_connector->dc_em_sink;
 
-	if (!edid || !sink) {
+	drm_edid_connector_update(connector, drm_edid);
+
+	if (!drm_edid || !sink) {
 		dm_con_state = to_dm_connector_state(connector->state);
 
 		amdgpu_dm_connector->min_vfreq = 0;
 		amdgpu_dm_connector->max_vfreq = 0;
-		amdgpu_dm_connector->pixel_clock_mhz = 0;
-		connector->display_info.monitor_range.min_vfreq = 0;
-		connector->display_info.monitor_range.max_vfreq = 0;
 		freesync_capable = false;
 
 		goto update;
@@ -11353,63 +13196,57 @@ void amdgpu_dm_update_freesync_caps(struct drm_connector *connector,
 
 	dm_con_state = to_dm_connector_state(connector->state);
 
-	if (!adev->dm.freesync_module)
+	if (!adev->dm.freesync_module || !dc_supports_vrr(sink->ctx->dce_version))
 		goto update;
 
+	edid = drm_edid_raw(drm_edid); // FIXME: Get rid of drm_edid_raw()
 
-	if (sink->sink_signal == SIGNAL_TYPE_DISPLAY_PORT
-		|| sink->sink_signal == SIGNAL_TYPE_EDP) {
-		bool edid_check_required = false;
+	/* Some eDP panels only have the refresh rate range info in DisplayID */
+	if ((connector->display_info.monitor_range.min_vfreq == 0 ||
+	     connector->display_info.monitor_range.max_vfreq == 0))
+		parse_edid_displayid_vrr(connector, edid);
 
-		if (edid) {
-			edid_check_required = is_dp_capable_without_timing_msa(
-						adev->dm.dc,
-						amdgpu_dm_connector);
+	if (edid && (sink->sink_signal == SIGNAL_TYPE_DISPLAY_PORT ||
+		     sink->sink_signal == SIGNAL_TYPE_EDP)) {
+		if (amdgpu_dm_connector->dc_link &&
+		    amdgpu_dm_connector->dc_link->dpcd_caps.allow_invalid_MSA_timing_param) {
+			amdgpu_dm_connector->min_vfreq = connector->display_info.monitor_range.min_vfreq;
+			amdgpu_dm_connector->max_vfreq = connector->display_info.monitor_range.max_vfreq;
+			if (amdgpu_dm_connector->max_vfreq - amdgpu_dm_connector->min_vfreq > 10)
+				freesync_capable = true;
 		}
 
-		if (edid_check_required == true && (edid->version > 1 ||
-		   (edid->version == 1 && edid->revision > 1))) {
-			for (i = 0; i < 4; i++) {
+		parse_amd_vsdb(amdgpu_dm_connector, edid, &vsdb_info);
 
-				timing	= &edid->detailed_timings[i];
-				data	= &timing->data.other_data;
-				range	= &data->data.range;
-				/*
-				 * Check if monitor has continuous frequency mode
-				 */
-				if (data->type != EDID_DETAIL_MONITOR_RANGE)
-					continue;
-				/*
-				 * Check for flag range limits only. If flag == 1 then
-				 * no additional timing information provided.
-				 * Default GTF, GTF Secondary curve and CVT are not
-				 * supported
-				 */
-				if (range->flags != 1)
-					continue;
-
-				amdgpu_dm_connector->min_vfreq = range->min_vfreq;
-				amdgpu_dm_connector->max_vfreq = range->max_vfreq;
-				amdgpu_dm_connector->pixel_clock_mhz =
-					range->pixel_clock_mhz * 10;
+		if (vsdb_info.replay_mode) {
+			amdgpu_dm_connector->vsdb_info.replay_mode = vsdb_info.replay_mode;
+			amdgpu_dm_connector->vsdb_info.amd_vsdb_version = vsdb_info.amd_vsdb_version;
+			amdgpu_dm_connector->as_type = ADAPTIVE_SYNC_TYPE_EDP;
+		}
 
-				connector->display_info.monitor_range.min_vfreq = range->min_vfreq;
-				connector->display_info.monitor_range.max_vfreq = range->max_vfreq;
+	} else if (drm_edid && sink->sink_signal == SIGNAL_TYPE_HDMI_TYPE_A) {
+		i = parse_hdmi_amd_vsdb(amdgpu_dm_connector, edid, &vsdb_info);
+		if (i >= 0 && vsdb_info.freesync_supported) {
+			amdgpu_dm_connector->min_vfreq = vsdb_info.min_refresh_rate_hz;
+			amdgpu_dm_connector->max_vfreq = vsdb_info.max_refresh_rate_hz;
+			if (amdgpu_dm_connector->max_vfreq - amdgpu_dm_connector->min_vfreq > 10)
+				freesync_capable = true;
 
-				break;
-			}
+			connector->display_info.monitor_range.min_vfreq = vsdb_info.min_refresh_rate_hz;
+			connector->display_info.monitor_range.max_vfreq = vsdb_info.max_refresh_rate_hz;
+		}
+	}
 
-			if (amdgpu_dm_connector->max_vfreq -
-			    amdgpu_dm_connector->min_vfreq > 10) {
+	if (amdgpu_dm_connector->dc_link)
+		as_type = dm_get_adaptive_sync_support_type(amdgpu_dm_connector->dc_link);
 
-				freesync_capable = true;
-			}
-		}
-	} else if (edid && sink->sink_signal == SIGNAL_TYPE_HDMI_TYPE_A) {
+	if (as_type == FREESYNC_TYPE_PCON_IN_WHITELIST) {
 		i = parse_hdmi_amd_vsdb(amdgpu_dm_connector, edid, &vsdb_info);
-		if (i >= 0 && vsdb_info.freesync_supported) {
-			timing  = &edid->detailed_timings[i];
-			data    = &timing->data.other_data;
+		if (i >= 0 && vsdb_info.freesync_supported && vsdb_info.amd_vsdb_version > 0) {
+
+			amdgpu_dm_connector->pack_sdp_v1_3 = true;
+			amdgpu_dm_connector->as_type = as_type;
+			amdgpu_dm_connector->vsdb_info = vsdb_info;
 
 			amdgpu_dm_connector->min_vfreq = vsdb_info.min_refresh_rate_hz;
 			amdgpu_dm_connector->max_vfreq = vsdb_info.max_refresh_rate_hz;
@@ -11425,6 +13262,12 @@ update:
 	if (dm_con_state)
 		dm_con_state->freesync_capable = freesync_capable;
 
+	if (connector->state && amdgpu_dm_connector->dc_link && !freesync_capable &&
+	    amdgpu_dm_connector->dc_link->replay_settings.config.replay_supported) {
+		amdgpu_dm_connector->dc_link->replay_settings.config.replay_supported = false;
+		amdgpu_dm_connector->dc_link->replay_settings.replay_feature_enabled = false;
+	}
+
 	if (connector->vrr_capable_property)
 		drm_connector_set_vrr_capable_property(connector,
 						       freesync_capable);
@@ -11449,15 +13292,24 @@ void amdgpu_dm_trigger_timing_sync(struct drm_device *dev)
 	mutex_unlock(&adev->dm.dc_lock);
 }
 
+static inline void amdgpu_dm_exit_ips_for_hw_access(struct dc *dc)
+{
+	if (dc->ctx->dmub_srv && !dc->ctx->dmub_srv->idle_exit_counter)
+		dc_exit_ips_for_hw_access(dc);
+}
+
 void dm_write_reg_func(const struct dc_context *ctx, uint32_t address,
-		       uint32_t value, const char *func_name)
+		       u32 value, const char *func_name)
 {
 #ifdef DM_CHECK_ADDR_0
 	if (address == 0) {
-		DC_ERR("invalid register write. address = 0");
+		drm_err(adev_to_drm(ctx->driver_context),
+			"invalid register write. address = 0");
 		return;
 	}
 #endif
+
+	amdgpu_dm_exit_ips_for_hw_access(ctx->dc);
 	cgs_write_register(ctx->cgs_device, address, value);
 	trace_amdgpu_dc_wreg(&ctx->perf_trace->write_count, address, value);
 }
@@ -11465,10 +13317,11 @@ void dm_write_reg_func(const struct dc_context *ctx, uint32_t address,
 uint32_t dm_read_reg_func(const struct dc_context *ctx, uint32_t address,
 			  const char *func_name)
 {
-	uint32_t value;
+	u32 value;
 #ifdef DM_CHECK_ADDR_0
 	if (address == 0) {
-		DC_ERR("invalid register read; address = 0\n");
+		drm_err(adev_to_drm(ctx->driver_context),
+			"invalid register read; address = 0\n");
 		return 0;
 	}
 #endif
@@ -11480,6 +13333,8 @@ uint32_t dm_read_reg_func(const struct dc_context *ctx, uint32_t address,
 		return 0;
 	}
 
+	amdgpu_dm_exit_ips_for_hw_access(ctx->dc);
+
 	value = cgs_read_register(ctx->cgs_device, address);
 
 	trace_amdgpu_dc_rreg(&ctx->perf_trace->read_count, address, value);
@@ -11487,75 +13342,176 @@ uint32_t dm_read_reg_func(const struct dc_context *ctx, uint32_t address,
 	return value;
 }
 
-int amdgpu_dm_set_dmub_async_sync_status(bool is_cmd_aux, struct dc_context *ctx,
-	uint8_t status_type, uint32_t *operation_result)
+int amdgpu_dm_process_dmub_aux_transfer_sync(
+		struct dc_context *ctx,
+		unsigned int link_index,
+		struct aux_payload *payload,
+		enum aux_return_code_type *operation_result)
 {
 	struct amdgpu_device *adev = ctx->driver_context;
-	int return_status = -1;
 	struct dmub_notification *p_notify = adev->dm.dmub_notify;
+	int ret = -1;
 
-	if (is_cmd_aux) {
-		if (status_type == DMUB_ASYNC_TO_SYNC_ACCESS_SUCCESS) {
-			return_status = p_notify->aux_reply.length;
-			*operation_result = p_notify->result;
-		} else if (status_type == DMUB_ASYNC_TO_SYNC_ACCESS_TIMEOUT) {
-			*operation_result = AUX_RET_ERROR_TIMEOUT;
-		} else if (status_type == DMUB_ASYNC_TO_SYNC_ACCESS_FAIL) {
-			*operation_result = AUX_RET_ERROR_ENGINE_ACQUIRE;
-		} else {
-			*operation_result = AUX_RET_ERROR_UNKNOWN;
+	mutex_lock(&adev->dm.dpia_aux_lock);
+	if (!dc_process_dmub_aux_transfer_async(ctx->dc, link_index, payload)) {
+		*operation_result = AUX_RET_ERROR_ENGINE_ACQUIRE;
+		goto out;
+	}
+
+	if (!wait_for_completion_timeout(&adev->dm.dmub_aux_transfer_done, 10 * HZ)) {
+		drm_err(adev_to_drm(adev), "wait_for_completion_timeout timeout!");
+		*operation_result = AUX_RET_ERROR_TIMEOUT;
+		goto out;
+	}
+
+	if (p_notify->result != AUX_RET_SUCCESS) {
+		/*
+		 * Transient states before tunneling is enabled could
+		 * lead to this error. We can ignore this for now.
+		 */
+		if (p_notify->result == AUX_RET_ERROR_PROTOCOL_ERROR) {
+			drm_warn(adev_to_drm(adev), "DPIA AUX failed on 0x%x(%d), error %d\n",
+					payload->address, payload->length,
+					p_notify->result);
 		}
-	} else {
-		if (status_type == DMUB_ASYNC_TO_SYNC_ACCESS_SUCCESS) {
-			return_status = 0;
-			*operation_result = p_notify->sc_status;
-		} else {
-			*operation_result = SET_CONFIG_UNKNOWN_ERROR;
+		*operation_result = p_notify->result;
+		goto out;
+	}
+
+	payload->reply[0] = adev->dm.dmub_notify->aux_reply.command & 0xF;
+	if (adev->dm.dmub_notify->aux_reply.command & 0xF0)
+		/* The reply is stored in the top nibble of the command. */
+		payload->reply[0] = (adev->dm.dmub_notify->aux_reply.command >> 4) & 0xF;
+
+	/*write req may receive a byte indicating partially written number as well*/
+	if (p_notify->aux_reply.length)
+		memcpy(payload->data, p_notify->aux_reply.data,
+				p_notify->aux_reply.length);
+
+	/* success */
+	ret = p_notify->aux_reply.length;
+	*operation_result = p_notify->result;
+out:
+	reinit_completion(&adev->dm.dmub_aux_transfer_done);
+	mutex_unlock(&adev->dm.dpia_aux_lock);
+	return ret;
+}
+
+static void abort_fused_io(
+		struct dc_context *ctx,
+		const struct dmub_cmd_fused_request *request
+)
+{
+	union dmub_rb_cmd command = { 0 };
+	struct dmub_rb_cmd_fused_io *io = &command.fused_io;
+
+	io->header.type = DMUB_CMD__FUSED_IO;
+	io->header.sub_type = DMUB_CMD__FUSED_IO_ABORT;
+	io->header.payload_bytes = sizeof(*io) - sizeof(io->header);
+	io->request = *request;
+	dm_execute_dmub_cmd(ctx, &command, DM_DMUB_WAIT_TYPE_NO_WAIT);
+}
+
+static bool execute_fused_io(
+		struct amdgpu_device *dev,
+		struct dc_context *ctx,
+		union dmub_rb_cmd *commands,
+		uint8_t count,
+		uint32_t timeout_us
+)
+{
+	const uint8_t ddc_line = commands[0].fused_io.request.u.aux.ddc_line;
+
+	if (ddc_line >= ARRAY_SIZE(dev->dm.fused_io))
+		return false;
+
+	struct fused_io_sync *sync = &dev->dm.fused_io[ddc_line];
+	struct dmub_rb_cmd_fused_io *first = &commands[0].fused_io;
+	const bool result = dm_execute_dmub_cmd_list(ctx, count, commands, DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY)
+			&& first->header.ret_status
+			&& first->request.status == FUSED_REQUEST_STATUS_SUCCESS;
+
+	if (!result)
+		return false;
+
+	while (wait_for_completion_timeout(&sync->replied, usecs_to_jiffies(timeout_us))) {
+		reinit_completion(&sync->replied);
+
+		struct dmub_cmd_fused_request *reply = (struct dmub_cmd_fused_request *) sync->reply_data;
+
+		static_assert(sizeof(*reply) <= sizeof(sync->reply_data), "Size mismatch");
+
+		if (reply->identifier == first->request.identifier) {
+			first->request = *reply;
+			return true;
 		}
 	}
 
-	return return_status;
+	reinit_completion(&sync->replied);
+	first->request.status = FUSED_REQUEST_STATUS_TIMEOUT;
+	abort_fused_io(ctx, &first->request);
+	return false;
 }
 
-int amdgpu_dm_process_dmub_aux_transfer_sync(bool is_cmd_aux, struct dc_context *ctx,
-	unsigned int link_index, void *cmd_payload, void *operation_result)
+bool amdgpu_dm_execute_fused_io(
+		struct amdgpu_device *dev,
+		struct dc_link *link,
+		union dmub_rb_cmd *commands,
+		uint8_t count,
+		uint32_t timeout_us)
+{
+	struct amdgpu_display_manager *dm = &dev->dm;
+
+	mutex_lock(&dm->dpia_aux_lock);
+
+	const bool result = execute_fused_io(dev, link->ctx, commands, count, timeout_us);
+
+	mutex_unlock(&dm->dpia_aux_lock);
+	return result;
+}
+
+int amdgpu_dm_process_dmub_set_config_sync(
+		struct dc_context *ctx,
+		unsigned int link_index,
+		struct set_config_cmd_payload *payload,
+		enum set_config_status *operation_result)
 {
 	struct amdgpu_device *adev = ctx->driver_context;
-	int ret = 0;
+	bool is_cmd_complete;
+	int ret;
 
-	if (is_cmd_aux) {
-		dc_process_dmub_aux_transfer_async(ctx->dc,
-			link_index, (struct aux_payload *)cmd_payload);
-	} else if (dc_process_dmub_set_config_async(ctx->dc, link_index,
-					(struct set_config_cmd_payload *)cmd_payload,
-					adev->dm.dmub_notify)) {
-		return amdgpu_dm_set_dmub_async_sync_status(is_cmd_aux,
-					ctx, DMUB_ASYNC_TO_SYNC_ACCESS_SUCCESS,
-					(uint32_t *)operation_result);
-	}
+	mutex_lock(&adev->dm.dpia_aux_lock);
+	is_cmd_complete = dc_process_dmub_set_config_async(ctx->dc,
+			link_index, payload, adev->dm.dmub_notify);
 
-	ret = wait_for_completion_timeout(&adev->dm.dmub_aux_transfer_done, 10 * HZ);
-	if (ret == 0) {
-		DRM_ERROR("wait_for_completion_timeout timeout!");
-		return amdgpu_dm_set_dmub_async_sync_status(is_cmd_aux,
-				ctx, DMUB_ASYNC_TO_SYNC_ACCESS_TIMEOUT,
-				(uint32_t *)operation_result);
+	if (is_cmd_complete || wait_for_completion_timeout(&adev->dm.dmub_aux_transfer_done, 10 * HZ)) {
+		ret = 0;
+		*operation_result = adev->dm.dmub_notify->sc_status;
+	} else {
+		drm_err(adev_to_drm(adev), "wait_for_completion_timeout timeout!");
+		ret = -1;
+		*operation_result = SET_CONFIG_UNKNOWN_ERROR;
 	}
 
-	if (is_cmd_aux) {
-		if (adev->dm.dmub_notify->result == AUX_RET_SUCCESS) {
-			struct aux_payload *payload = (struct aux_payload *)cmd_payload;
+	if (!is_cmd_complete)
+		reinit_completion(&adev->dm.dmub_aux_transfer_done);
+	mutex_unlock(&adev->dm.dpia_aux_lock);
+	return ret;
+}
 
-			payload->reply[0] = adev->dm.dmub_notify->aux_reply.command;
-			if (!payload->write && adev->dm.dmub_notify->aux_reply.length &&
-			    payload->reply[0] == AUX_TRANSACTION_REPLY_AUX_ACK) {
-				memcpy(payload->data, adev->dm.dmub_notify->aux_reply.data,
-				       adev->dm.dmub_notify->aux_reply.length);
-			}
-		}
-	}
+bool dm_execute_dmub_cmd(const struct dc_context *ctx, union dmub_rb_cmd *cmd, enum dm_dmub_wait_type wait_type)
+{
+	return dc_dmub_srv_cmd_run(ctx->dmub_srv, cmd, wait_type);
+}
+
+bool dm_execute_dmub_cmd_list(const struct dc_context *ctx, unsigned int count, union dmub_rb_cmd *cmd, enum dm_dmub_wait_type wait_type)
+{
+	return dc_dmub_srv_cmd_run_list(ctx->dmub_srv, count, cmd, wait_type);
+}
 
-	return amdgpu_dm_set_dmub_async_sync_status(is_cmd_aux,
-			ctx, DMUB_ASYNC_TO_SYNC_ACCESS_SUCCESS,
-			(uint32_t *)operation_result);
+void dm_acpi_process_phy_transition_interlock(
+	const struct dc_context *ctx,
+	struct dm_process_phy_transition_init_params process_phy_transition_init_params)
+{
+	// Not yet implemented
 }
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h
index 37e61a88d49e..ef97cede9926 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h
@@ -1,3 +1,4 @@
+/* SPDX-License-Identifier: MIT */
 /*
  * Copyright (C) 2015-2020 Advanced Micro Devices, Inc. All rights reserved.
  *
@@ -26,11 +27,13 @@
 #ifndef __AMDGPU_DM_H__
 #define __AMDGPU_DM_H__
 
+#include <drm/display/drm_dp_mst_helper.h>
 #include <drm/drm_atomic.h>
 #include <drm/drm_connector.h>
 #include <drm/drm_crtc.h>
-#include <drm/drm_dp_mst_helper.h>
 #include <drm/drm_plane.h>
+#include "link_service_types.h"
+#include <drm/drm_writeback.h>
 
 /*
  * This file contains the definition for amdgpu_display_manager
@@ -48,14 +51,15 @@
 
 #define AMDGPU_DM_MAX_NUM_EDP 2
 
-#define AMDGPU_DMUB_NOTIFICATION_MAX 5
+#define AMDGPU_DMUB_NOTIFICATION_MAX 8
 
-/**
- * DMUB Async to Sync Mechanism Status
- **/
-#define DMUB_ASYNC_TO_SYNC_ACCESS_FAIL 1
-#define DMUB_ASYNC_TO_SYNC_ACCESS_TIMEOUT 2
-#define DMUB_ASYNC_TO_SYNC_ACCESS_SUCCESS 3
+#define HDMI_AMD_VENDOR_SPECIFIC_DATA_BLOCK_IEEE_REGISTRATION_ID 0x00001A
+#define AMD_VSDB_VERSION_3_FEATURECAP_REPLAYMODE 0x40
+#define HDMI_AMD_VENDOR_SPECIFIC_DATA_BLOCK_VERSION_3 0x3
+
+#define AMDGPU_HDR_MULT_DEFAULT (0x100000000LL)
+
+#define AMDGPU_DM_HDMI_HPD_DEBOUNCE_MS 1500
 /*
 #include "include/amdgpu_dal_power_if.h"
 #include "amdgpu_dm_irq.h"
@@ -64,8 +68,11 @@
 #include "irq_types.h"
 #include "signal_types.h"
 #include "amdgpu_dm_crc.h"
+#include "mod_info_packet.h"
 struct aux_payload;
+struct set_config_cmd_payload;
 enum aux_return_code_type;
+enum set_config_status;
 
 /* Forward declarations */
 struct amdgpu_device;
@@ -76,6 +83,13 @@ struct amdgpu_bo;
 struct dmub_srv;
 struct dc_plane_state;
 struct dmub_notification;
+struct dmub_cmd_fused_request;
+
+struct amd_vsdb_block {
+	unsigned char ieee_id[3];
+	unsigned char version;
+	unsigned char feature_caps;
+};
 
 struct common_irq_params {
 	struct amdgpu_device *adev;
@@ -127,6 +141,46 @@ struct vblank_control_work {
 };
 
 /**
+ * struct idle_workqueue - Work data for periodic action in idle
+ * @work: Kernel work data for the work event
+ * @dm: amdgpu display manager device
+ * @enable: true if idle worker is enabled
+ * @running: true if idle worker is running
+ */
+struct idle_workqueue {
+	struct work_struct work;
+	struct amdgpu_display_manager *dm;
+	bool enable;
+	bool running;
+};
+
+/**
+ * struct vupdate_offload_work - Work data for offloading task from vupdate handler
+ * @work: Kernel work data for the work event
+ * @adev: amdgpu_device back pointer
+ * @stream: DC stream associated with the crtc
+ * @adjust: DC CRTC timing adjust to be applied to the crtc
+ */
+struct vupdate_offload_work {
+	struct work_struct work;
+	struct amdgpu_device *adev;
+	struct dc_stream_state *stream;
+	struct dc_crtc_timing_adjust *adjust;
+};
+
+#define MAX_LUMINANCE_DATA_POINTS 99
+
+/**
+ * struct amdgpu_dm_luminance_data - Custom luminance data
+ * @luminance: Luminance in percent
+ * @input_signal: Input signal in range 0-255
+ */
+struct amdgpu_dm_luminance_data {
+	u8 luminance;
+	u8 input_signal;
+} __packed;
+
+/**
  * struct amdgpu_dm_backlight_caps - Information about backlight
  *
  * Describe the backlight support for ACPI or eDP AUX.
@@ -162,6 +216,27 @@ struct amdgpu_dm_backlight_caps {
 	 * @aux_support: Describes if the display supports AUX backlight.
 	 */
 	bool aux_support;
+	/**
+	 * @brightness_mask: After deriving brightness, OR it with this mask.
+	 * Workaround for panels with issues with certain brightness values.
+	 */
+	u32 brightness_mask;
+	/**
+	 * @ac_level: the default brightness if booted on AC
+	 */
+	u8 ac_level;
+	/**
+	 * @dc_level: the default brightness if booted on DC
+	 */
+	u8 dc_level;
+	/**
+	 * @data_points: the number of custom luminance data points
+	 */
+	u8 data_points;
+	/**
+	 * @luminance_data: custom luminance data
+	 */
+	struct amdgpu_dm_luminance_data luminance_data[MAX_LUMINANCE_DATA_POINTS];
 };
 
 /**
@@ -197,6 +272,11 @@ struct hpd_rx_irq_offload_work_queue {
 	 */
 	bool is_handling_link_loss;
 	/**
+	 * @is_handling_mst_msg_rdy_event: Used to prevent inserting mst message
+	 * ready event when we're already handling mst message ready event
+	 */
+	bool is_handling_mst_msg_rdy_event;
+	/**
 	 * @aconnector: The aconnector that this work queue is attached to
 	 */
 	struct amdgpu_dm_connector *aconnector;
@@ -218,6 +298,10 @@ struct hpd_rx_irq_offload_work {
 	 * @offload_wq: offload work queue that this work is queued to
 	 */
 	struct hpd_rx_irq_offload_work_queue *offload_wq;
+	/**
+	 * @adev: amdgpu_device pointer
+	 */
+	struct amdgpu_device *adev;
 };
 
 /**
@@ -242,6 +326,13 @@ struct hpd_rx_irq_offload_work {
  * @force_timing_sync: set via debugfs. When set, indicates that all connected
  *		       displays will be forced to synchronize.
  * @dmcub_trace_event_en: enable dmcub trace events
+ * @dmub_outbox_params: DMUB Outbox parameters
+ * @num_of_edps: number of backlight eDPs
+ * @disable_hpd_irq: disables all HPD and HPD RX interrupt handling in the
+ *		     driver when true
+ * @dmub_aux_transfer_done: struct completion used to indicate when DMUB
+ * 			    transfers are done
+ * @delayed_hpd_wq: work queue used to delay DMUB HPD work
  */
 struct amdgpu_display_manager {
 
@@ -358,15 +449,6 @@ struct amdgpu_display_manager {
 	 */
 	struct mutex audio_lock;
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	/**
-	 * @vblank_lock:
-	 *
-	 * Guards access to deferred vblank work state.
-	 */
-	spinlock_t vblank_lock;
-#endif
-
 	/**
 	 * @audio_component:
 	 *
@@ -465,18 +547,21 @@ struct amdgpu_display_manager {
 	struct amdgpu_dm_backlight_caps backlight_caps[AMDGPU_DM_MAX_NUM_EDP];
 
 	struct mod_freesync *freesync_module;
-#ifdef CONFIG_DRM_AMD_DC_HDCP
 	struct hdcp_workqueue *hdcp_workqueue;
-#endif
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	/**
 	 * @vblank_control_workqueue:
 	 *
 	 * Deferred work for vblank control events.
 	 */
 	struct workqueue_struct *vblank_control_workqueue;
-#endif
+
+	/**
+	 * @idle_workqueue:
+	 *
+	 * Periodic work for idle events.
+	 */
+	struct idle_workqueue *idle_workqueue;
 
 	struct drm_atomic_state *cached_state;
 	struct dc_state *cached_dc_state;
@@ -493,22 +578,21 @@ struct amdgpu_display_manager {
 	 */
 	const struct gpu_info_soc_bounding_box_v1_0 *soc_bounding_box;
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	/**
 	 * @active_vblank_irq_count:
 	 *
 	 * number of currently active vblank irqs
 	 */
 	uint32_t active_vblank_irq_count;
-#endif
 
 #if defined(CONFIG_DRM_AMD_SECURE_DISPLAY)
 	/**
-	 * @crc_rd_wrk:
+	 * @secure_display_ctx:
 	 *
-	 * Work to be executed in a separate thread to communicate with PSP.
+	 * Store secure display relevant info. e.g. the ROI information
+	 * , the work_struct to command dmub, etc.
 	 */
-	struct crc_rd_work *crc_rd_wrk;
+	struct secure_display_context secure_display_ctx;
 #endif
 	/**
 	 * @hpd_rx_offload_wq:
@@ -540,6 +624,59 @@ struct amdgpu_display_manager {
 	 * cached backlight values.
 	 */
 	u32 brightness[AMDGPU_DM_MAX_NUM_EDP];
+	/**
+	 * @actual_brightness:
+	 *
+	 * last successfully applied backlight values.
+	 */
+	u32 actual_brightness[AMDGPU_DM_MAX_NUM_EDP];
+
+	/**
+	 * @aux_hpd_discon_quirk:
+	 *
+	 * quirk for hpd discon while aux is on-going.
+	 * occurred on certain intel platform
+	 */
+	bool aux_hpd_discon_quirk;
+
+	/**
+	 * @edp0_on_dp1_quirk:
+	 *
+	 * quirk for platforms that put edp0 on DP1.
+	 */
+	bool edp0_on_dp1_quirk;
+
+	/**
+	 * @dpia_aux_lock:
+	 *
+	 * Guards access to DPIA AUX
+	 */
+	struct mutex dpia_aux_lock;
+
+	/**
+	 * @bb_from_dmub:
+	 *
+	 * Bounding box data read from dmub during early initialization for DCN4+
+	 * Data is stored as a byte array that should be casted to the appropriate bb struct
+	 */
+	void *bb_from_dmub;
+
+	/**
+	 * @oem_i2c:
+	 *
+	 * OEM i2c bus
+	 */
+	struct amdgpu_i2c_adapter *oem_i2c;
+
+	/**
+	 * @fused_io:
+	 *
+	 * dmub fused io interface
+	 */
+	struct fused_io_sync {
+		struct completion replied;
+		char reply_data[0x40];  // Cannot include dmub_cmd here
+	} fused_io[8];
 };
 
 enum dsc_clock_force_state {
@@ -556,14 +693,60 @@ struct dsc_preferred_settings {
 	bool dsc_force_disable_passthrough;
 };
 
+enum mst_progress_status {
+	MST_STATUS_DEFAULT = 0,
+	MST_PROBE = BIT(0),
+	MST_REMOTE_EDID = BIT(1),
+	MST_ALLOCATE_NEW_PAYLOAD = BIT(2),
+	MST_CLEAR_ALLOCATED_PAYLOAD = BIT(3),
+};
+
+/**
+ * struct amdgpu_hdmi_vsdb_info - Keep track of the VSDB info
+ *
+ * AMDGPU supports FreeSync over HDMI by using the VSDB section, and this
+ * struct is useful to keep track of the display-specific information about
+ * FreeSync.
+ */
+struct amdgpu_hdmi_vsdb_info {
+	/**
+	 * @amd_vsdb_version: Vendor Specific Data Block Version, should be
+	 * used to determine which Vendor Specific InfoFrame (VSIF) to send.
+	 */
+	unsigned int amd_vsdb_version;
+
+	/**
+	 * @freesync_supported: FreeSync Supported.
+	 */
+	bool freesync_supported;
+
+	/**
+	 * @min_refresh_rate_hz: FreeSync Minimum Refresh Rate in Hz.
+	 */
+	unsigned int min_refresh_rate_hz;
+
+	/**
+	 * @max_refresh_rate_hz: FreeSync Maximum Refresh Rate in Hz
+	 */
+	unsigned int max_refresh_rate_hz;
+
+	/**
+	 * @replay_mode: Replay supported
+	 */
+	bool replay_mode;
+};
+
 struct amdgpu_dm_connector {
 
 	struct drm_connector base;
 	uint32_t connector_id;
+	int bl_idx;
+
+	struct cec_notifier *notifier;
 
 	/* we need to mind the EDID between detect
 	   and get modes due to analog/digital/tvencoder */
-	struct edid *edid;
+	const struct drm_edid *drm_edid;
 
 	/* shared with amdgpu */
 	struct amdgpu_hpd hpd;
@@ -575,22 +758,40 @@ struct amdgpu_dm_connector {
 	 * The 'current' sink is in dc_link->sink. */
 	struct dc_sink *dc_sink;
 	struct dc_link *dc_link;
+
+	/**
+	 * @dc_em_sink: Reference to the emulated (virtual) sink.
+	 */
 	struct dc_sink *dc_em_sink;
 
 	/* DM only */
 	struct drm_dp_mst_topology_mgr mst_mgr;
 	struct amdgpu_dm_dp_aux dm_dp_aux;
-	struct drm_dp_mst_port *port;
-	struct amdgpu_dm_connector *mst_port;
+	struct drm_dp_mst_port *mst_output_port;
+	struct amdgpu_dm_connector *mst_root;
 	struct drm_dp_aux *dsc_aux;
+	uint32_t mst_local_bw;
+	uint16_t vc_full_pbn;
+	struct mutex handle_mst_msg_ready;
+
+	/* branch device specific data */
+	uint32_t branch_ieee_oui;
 
 	/* TODO see if we can merge with ddc_bus or make a dm_connector */
 	struct amdgpu_i2c_adapter *i2c;
 
 	/* Monitor range limits */
-	int min_vfreq ;
+	/**
+	 * @min_vfreq: Minimal frequency supported by the display in Hz. This
+	 * value is set to zero when there is no FreeSync support.
+	 */
+	int min_vfreq;
+
+	/**
+	 * @max_vfreq: Maximum frequency supported by the display in Hz. This
+	 * value is set to zero when there is no FreeSync support.
+	 */
 	int max_vfreq ;
-	int pixel_clock_mhz;
 
 	/* Audio instance - protected by audio_lock. */
 	int audio_inst;
@@ -598,25 +799,151 @@ struct amdgpu_dm_connector {
 	struct mutex hpd_lock;
 
 	bool fake_enable;
-#ifdef CONFIG_DEBUG_FS
-	uint32_t debugfs_dpcd_address;
-	uint32_t debugfs_dpcd_size;
-#endif
 	bool force_yuv420_output;
+	bool force_yuv422_output;
 	struct dsc_preferred_settings dsc_settings;
+	union dp_downstream_port_present mst_downstream_port_present;
 	/* Cached display modes */
 	struct drm_display_mode freesync_vid_base;
 
-	int psr_skip_count;
+	int sr_skip_count;
+	bool disallow_edp_enter_psr;
+
+	/* Record progress status of mst*/
+	uint8_t mst_status;
+
+	/* Automated testing */
+	bool timing_changed;
+	struct dc_crtc_timing *timing_requested;
+
+	/* Adaptive Sync */
+	bool pack_sdp_v1_3;
+	enum adaptive_sync_type as_type;
+	struct amdgpu_hdmi_vsdb_info vsdb_info;
+
+	/* HDMI HPD debounce support */
+	unsigned int hdmi_hpd_debounce_delay_ms;
+	struct delayed_work hdmi_hpd_debounce_work;
+	struct dc_sink *hdmi_prev_sink;
 };
 
+static inline void amdgpu_dm_set_mst_status(uint8_t *status,
+		uint8_t flags, bool set)
+{
+	if (set)
+		*status |= flags;
+	else
+		*status &= ~flags;
+}
+
 #define to_amdgpu_dm_connector(x) container_of(x, struct amdgpu_dm_connector, base)
 
+struct amdgpu_dm_wb_connector {
+	struct drm_writeback_connector base;
+	struct dc_link *link;
+};
+
+#define to_amdgpu_dm_wb_connector(x) container_of(x, struct amdgpu_dm_wb_connector, base)
+
 extern const struct amdgpu_ip_block_version dm_ip_block;
 
+/* enum amdgpu_transfer_function: pre-defined transfer function supported by AMD.
+ *
+ * It includes standardized transfer functions and pure power functions. The
+ * transfer function coefficients are available at modules/color/color_gamma.c
+ */
+enum amdgpu_transfer_function {
+	AMDGPU_TRANSFER_FUNCTION_DEFAULT,
+	AMDGPU_TRANSFER_FUNCTION_SRGB_EOTF,
+	AMDGPU_TRANSFER_FUNCTION_BT709_INV_OETF,
+	AMDGPU_TRANSFER_FUNCTION_PQ_EOTF,
+	AMDGPU_TRANSFER_FUNCTION_IDENTITY,
+	AMDGPU_TRANSFER_FUNCTION_GAMMA22_EOTF,
+	AMDGPU_TRANSFER_FUNCTION_GAMMA24_EOTF,
+	AMDGPU_TRANSFER_FUNCTION_GAMMA26_EOTF,
+	AMDGPU_TRANSFER_FUNCTION_SRGB_INV_EOTF,
+	AMDGPU_TRANSFER_FUNCTION_BT709_OETF,
+	AMDGPU_TRANSFER_FUNCTION_PQ_INV_EOTF,
+	AMDGPU_TRANSFER_FUNCTION_GAMMA22_INV_EOTF,
+	AMDGPU_TRANSFER_FUNCTION_GAMMA24_INV_EOTF,
+	AMDGPU_TRANSFER_FUNCTION_GAMMA26_INV_EOTF,
+	AMDGPU_TRANSFER_FUNCTION_COUNT
+};
+
 struct dm_plane_state {
 	struct drm_plane_state base;
 	struct dc_plane_state *dc_state;
+
+	/* Plane color mgmt */
+	/**
+	 * @degamma_lut:
+	 *
+	 * 1D LUT for mapping framebuffer/plane pixel data before sampling or
+	 * blending operations. It's usually applied to linearize input space.
+	 * The blob (if not NULL) is an array of &struct drm_color_lut.
+	 */
+	struct drm_property_blob *degamma_lut;
+	/**
+	 * @degamma_tf:
+	 *
+	 * Predefined transfer function to tell DC driver the input space to
+	 * linearize.
+	 */
+	enum amdgpu_transfer_function degamma_tf;
+	/**
+	 * @hdr_mult:
+	 *
+	 * Multiplier to 'gain' the plane.  When PQ is decoded using the fixed
+	 * func transfer function to the internal FP16 fb, 1.0 -> 80 nits (on
+	 * AMD at least). When sRGB is decoded, 1.0 -> 1.0, obviously.
+	 * Therefore, 1.0 multiplier = 80 nits for SDR content.  So if you
+	 * want, 203 nits for SDR content, pass in (203.0 / 80.0).  Format is
+	 * S31.32 sign-magnitude.
+	 *
+	 * HDR multiplier can wide range beyond [0.0, 1.0]. This means that PQ
+	 * TF is needed for any subsequent linear-to-non-linear transforms.
+	 */
+	__u64 hdr_mult;
+	/**
+	 * @ctm:
+	 *
+	 * Color transformation matrix. The blob (if not NULL) is a &struct
+	 * drm_color_ctm_3x4.
+	 */
+	struct drm_property_blob *ctm;
+	/**
+	 * @shaper_lut: shaper lookup table blob. The blob (if not NULL) is an
+	 * array of &struct drm_color_lut.
+	 */
+	struct drm_property_blob *shaper_lut;
+	/**
+	 * @shaper_tf:
+	 *
+	 * Predefined transfer function to delinearize color space.
+	 */
+	enum amdgpu_transfer_function shaper_tf;
+	/**
+	 * @lut3d: 3D lookup table blob. The blob (if not NULL) is an array of
+	 * &struct drm_color_lut.
+	 */
+	struct drm_property_blob *lut3d;
+	/**
+	 * @blend_lut: blend lut lookup table blob. The blob (if not NULL) is an
+	 * array of &struct drm_color_lut.
+	 */
+	struct drm_property_blob *blend_lut;
+	/**
+	 * @blend_tf:
+	 *
+	 * Pre-defined transfer function for converting plane pixel data before
+	 * applying blend LUT.
+	 */
+	enum amdgpu_transfer_function blend_tf;
+};
+
+enum amdgpu_dm_cursor_mode {
+	DM_CURSOR_NATIVE_MODE = 0,
+	DM_CURSOR_OVERLAY_MODE,
 };
 
 struct dm_crtc_state {
@@ -626,22 +953,31 @@ struct dm_crtc_state {
 	bool cm_has_degamma;
 	bool cm_is_degamma_srgb;
 
+	bool mpo_requested;
+
 	int update_type;
 	int active_planes;
 
 	int crc_skip_count;
 
-	bool freesync_timing_changed;
 	bool freesync_vrr_info_changed;
 
 	bool dsc_force_changed;
 	bool vrr_supported;
-
-	bool force_dpms_off;
 	struct mod_freesync_config freesync_config;
 	struct dc_info_packet vrr_infopacket;
 
 	int abm_level;
+
+	/**
+	 * @regamma_tf:
+	 *
+	 * Pre-defined transfer function for converting internal FB -> wire
+	 * encoding.
+	 */
+	enum amdgpu_transfer_function regamma_tf;
+
+	enum amdgpu_dm_cursor_mode cursor_mode;
 };
 
 #define to_dm_crtc_state(x) container_of(x, struct dm_crtc_state, base)
@@ -662,22 +998,13 @@ struct dm_connector_state {
 	uint8_t underscan_hborder;
 	bool underscan_enable;
 	bool freesync_capable;
-#ifdef CONFIG_DRM_AMD_DC_HDCP
 	bool update_hdcp;
-#endif
+	bool abm_sysfs_forbidden;
 	uint8_t abm_level;
 	int vcpi_slots;
 	uint64_t pbn;
 };
 
-struct amdgpu_hdmi_vsdb_info {
-	unsigned int amd_vsdb_version;		/* VSDB version, should be used to determine which VSIF to send */
-	bool freesync_supported;		/* FreeSync Supported */
-	unsigned int min_refresh_rate_hz;	/* FreeSync Minimum Refresh Rate in Hz */
-	unsigned int max_refresh_rate_hz;	/* FreeSync Maximum Refresh Rate in Hz */
-};
-
-
 #define to_dm_connector_state(x)\
 	container_of((x), struct dm_connector_state, base)
 
@@ -703,24 +1030,34 @@ void amdgpu_dm_connector_init_helper(struct amdgpu_display_manager *dm,
 				     int link_index);
 
 enum drm_mode_status amdgpu_dm_connector_mode_valid(struct drm_connector *connector,
-				   struct drm_display_mode *mode);
+				   const struct drm_display_mode *mode);
 
 void dm_restore_drm_connector_state(struct drm_device *dev,
 				    struct drm_connector *connector);
 
 void amdgpu_dm_update_freesync_caps(struct drm_connector *connector,
-					struct edid *edid);
+				    const struct drm_edid *drm_edid);
 
 void amdgpu_dm_trigger_timing_sync(struct drm_device *dev);
 
+/* 3D LUT max size is 17x17x17 (4913 entries) */
+#define MAX_COLOR_3DLUT_SIZE 17
+#define MAX_COLOR_3DLUT_BITDEPTH 12
+int amdgpu_dm_verify_lut3d_size(struct amdgpu_device *adev,
+				struct drm_plane_state *plane_state);
+/* 1D LUT size */
 #define MAX_COLOR_LUT_ENTRIES 4096
 /* Legacy gamm LUT users such as X doesn't like large LUT sizes */
 #define MAX_COLOR_LEGACY_LUT_ENTRIES 256
 
 void amdgpu_dm_init_color_mod(void);
+int amdgpu_dm_create_color_properties(struct amdgpu_device *adev);
 int amdgpu_dm_verify_lut_sizes(const struct drm_crtc_state *crtc_state);
 int amdgpu_dm_update_crtc_color_mgmt(struct dm_crtc_state *crtc);
+int amdgpu_dm_check_crtc_color_mgmt(struct dm_crtc_state *crtc,
+				    bool check_only);
 int amdgpu_dm_update_plane_color_mgmt(struct dm_crtc_state *crtc,
+				      struct drm_plane_state *plane_state,
 				      struct dc_plane_state *dc_plane_state);
 
 void amdgpu_dm_update_connector_after_detect(
@@ -728,7 +1065,50 @@ void amdgpu_dm_update_connector_after_detect(
 
 extern const struct drm_encoder_helper_funcs amdgpu_dm_encoder_helper_funcs;
 
-int amdgpu_dm_process_dmub_aux_transfer_sync(bool is_cmd_aux,
-					struct dc_context *ctx, unsigned int link_index,
-					void *payload, void *operation_result);
+int amdgpu_dm_process_dmub_aux_transfer_sync(struct dc_context *ctx, unsigned int link_index,
+					struct aux_payload *payload, enum aux_return_code_type *operation_result);
+
+bool amdgpu_dm_execute_fused_io(
+		struct amdgpu_device *dev,
+		struct dc_link *link,
+		union dmub_rb_cmd *commands,
+		uint8_t count,
+		uint32_t timeout_us
+);
+
+int amdgpu_dm_process_dmub_set_config_sync(struct dc_context *ctx, unsigned int link_index,
+					struct set_config_cmd_payload *payload, enum set_config_status *operation_result);
+
+struct dc_stream_state *
+	create_validate_stream_for_sink(struct drm_connector *connector,
+					const struct drm_display_mode *drm_mode,
+					const struct dm_connector_state *dm_state,
+					const struct dc_stream_state *old_stream);
+
+int dm_atomic_get_state(struct drm_atomic_state *state,
+			struct dm_atomic_state **dm_state);
+
+struct drm_connector *
+amdgpu_dm_find_first_crtc_matching_connector(struct drm_atomic_state *state,
+					     struct drm_crtc *crtc);
+
+int convert_dc_color_depth_into_bpc(enum dc_color_depth display_color_depth);
+struct idle_workqueue *idle_create_workqueue(struct amdgpu_device *adev);
+
+void *dm_allocate_gpu_mem(struct amdgpu_device *adev,
+						  enum dc_gpu_mem_alloc_type type,
+						  size_t size,
+						  long long *addr);
+void dm_free_gpu_mem(struct amdgpu_device *adev,
+						  enum dc_gpu_mem_alloc_type type,
+						  void *addr);
+
+bool amdgpu_dm_is_headless(struct amdgpu_device *adev);
+
+void hdmi_cec_set_edid(struct amdgpu_dm_connector *aconnector);
+void hdmi_cec_unset_edid(struct amdgpu_dm_connector *aconnector);
+int amdgpu_dm_initialize_hdmi_connector(struct amdgpu_dm_connector *aconnector);
+
+void retrieve_dmi_info(struct amdgpu_display_manager *dm);
+
 #endif /* __AMDGPU_DM_H__ */
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_color.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_color.c
index a022e5bb30a5..1dcc79b35225 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_color.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_color.c
@@ -1,3 +1,4 @@
+// SPDX-License-Identifier: MIT
 /*
  * Copyright 2018 Advanced Micro Devices, Inc.
  *
@@ -25,11 +26,39 @@
 #include "amdgpu.h"
 #include "amdgpu_mode.h"
 #include "amdgpu_dm.h"
+#include "amdgpu_dm_colorop.h"
 #include "dc.h"
 #include "modules/color/color_gamma.h"
-#include "basics/conversion.h"
 
-/*
+/**
+ * DOC: overview
+ *
+ * We have three types of color management in the AMD display driver.
+ * 1. the legacy &drm_crtc DEGAMMA, CTM, and GAMMA properties
+ * 2. AMD driver private color management on &drm_plane and &drm_crtc
+ * 3. AMD plane color pipeline
+ *
+ * The CRTC properties are the original color management. When they were
+ * implemented per-plane color management was not a thing yet. Because
+ * of that we could get away with plumbing the DEGAMMA and CTM
+ * properties to pre-blending HW functions. This is incompatible with
+ * per-plane color management, such as via the AMD private properties or
+ * the new drm_plane color pipeline. The only compatible CRTC property
+ * with per-plane color management is the GAMMA property as it is
+ * applied post-blending.
+ *
+ * The AMD driver private color management properties are only exposed
+ * when the kernel is built explicitly with -DAMD_PRIVATE_COLOR. They
+ * are temporary building blocks on the path to full-fledged &drm_plane
+ * and &drm_crtc color pipelines and lay the driver's groundwork for the
+ * color pipelines.
+ *
+ * The AMD plane color pipeline describes AMD's &drm_colorops via the
+ * &drm_plane's COLOR_PIPELINE property.
+ *
+ * drm_crtc Properties
+ * -------------------
+ *
  * The DC interface to HW gives us the following color management blocks
  * per pipe (surface):
  *
@@ -40,39 +69,97 @@
  * - Surface regamma LUT (normalized)
  * - Output CSC (normalized)
  *
- * But these aren't a direct mapping to DRM color properties. The current DRM
- * interface exposes CRTC degamma, CRTC CTM and CRTC regamma while our hardware
- * is essentially giving:
+ * But these aren't a direct mapping to DRM color properties. The
+ * current DRM interface exposes CRTC degamma, CRTC CTM and CRTC regamma
+ * while our hardware is essentially giving:
  *
  * Plane CTM -> Plane degamma -> Plane CTM -> Plane regamma -> Plane CTM
  *
- * The input gamma LUT block isn't really applicable here since it operates
- * on the actual input data itself rather than the HW fp representation. The
- * input and output CSC blocks are technically available to use as part of
- * the DC interface but are typically used internally by DC for conversions
- * between color spaces. These could be blended together with user
- * adjustments in the future but for now these should remain untouched.
+ * The input gamma LUT block isn't really applicable here since it
+ * operates on the actual input data itself rather than the HW fp
+ * representation. The input and output CSC blocks are technically
+ * available to use as part of the DC interface but are typically used
+ * internally by DC for conversions between color spaces. These could be
+ * blended together with user adjustments in the future but for now
+ * these should remain untouched.
+ *
+ * The pipe blending also happens after these blocks so we don't
+ * actually support any CRTC props with correct blending with multiple
+ * planes - but we can still support CRTC color management properties in
+ * DM in most single plane cases correctly with clever management of the
+ * DC interface in DM.
+ *
+ * As per DRM documentation, blocks should be in hardware bypass when
+ * their respective property is set to NULL. A linear DGM/RGM LUT should
+ * also considered as putting the respective block into bypass mode.
+ *
+ * This means that the following configuration is assumed to be the
+ * default:
+ *
+ * Plane DGM Bypass -> Plane CTM Bypass -> Plane RGM Bypass -> ... CRTC
+ * DGM Bypass -> CRTC CTM Bypass -> CRTC RGM Bypass
+ *
+ * AMD Private Color Management on drm_plane
+ * -----------------------------------------
+ *
+ * The AMD private color management properties on a &drm_plane are:
+ *
+ * - AMD_PLANE_DEGAMMA_LUT
+ * - AMD_PLANE_DEGAMMA_LUT_SIZE
+ * - AMD_PLANE_DEGAMMA_TF
+ * - AMD_PLANE_HDR_MULT
+ * - AMD_PLANE_CTM
+ * - AMD_PLANE_SHAPER_LUT
+ * - AMD_PLANE_SHAPER_LUT_SIZE
+ * - AMD_PLANE_SHAPER_TF
+ * - AMD_PLANE_LUT3D
+ * - AMD_PLANE_LUT3D_SIZE
+ * - AMD_PLANE_BLEND_LUT
+ * - AMD_PLANE_BLEND_LUT_SIZE
+ * - AMD_PLANE_BLEND_TF
+ *
+ * The AMD private color management property on a &drm_crtc is:
+ *
+ * - AMD_CRTC_REGAMMA_TF
+ *
+ * Use of these properties is discouraged.
  *
- * The pipe blending also happens after these blocks so we don't actually
- * support any CRTC props with correct blending with multiple planes - but we
- * can still support CRTC color management properties in DM in most single
- * plane cases correctly with clever management of the DC interface in DM.
+ * AMD plane color pipeline
+ * ------------------------
  *
- * As per DRM documentation, blocks should be in hardware bypass when their
- * respective property is set to NULL. A linear DGM/RGM LUT should also
- * considered as putting the respective block into bypass mode.
+ * The AMD &drm_plane color pipeline is advertised for DCN generations
+ * 3.0 and newer. It exposes these elements in this order:
  *
- * This means that the following
- * configuration is assumed to be the default:
+ * 1. 1D curve colorop
+ * 2. Multiplier
+ * 3. 3x4 CTM
+ * 4. 1D curve colorop
+ * 5. 1D LUT
+ * 6. 3D LUT
+ * 7. 1D curve colorop
+ * 8. 1D LUT
+ *
+ * The multiplier (#2) is a simple multiplier that is applied to all
+ * channels.
+ *
+ * The 3x4 CTM (#3) is a simple 3x4 matrix.
+ *
+ * #1, and #7 are non-linear to linear curves. #4 is a linear to
+ * non-linear curve. They support sRGB, PQ, and BT.709/BT.2020 EOTFs or
+ * their inverse.
+ *
+ * The 1D LUTs (#5 and #8) are plain 4096 entry LUTs.
+ *
+ * The 3DLUT (#6) is a tetrahedrally interpolated 17 cube LUT.
  *
- * Plane DGM Bypass -> Plane CTM Bypass -> Plane RGM Bypass -> ...
- * CRTC DGM Bypass -> CRTC CTM Bypass -> CRTC RGM Bypass
  */
 
 #define MAX_DRM_LUT_VALUE 0xFFFF
+#define MAX_DRM_LUT32_VALUE 0xFFFFFFFF
+#define SDR_WHITE_LEVEL_INIT_VALUE 80
 
-/*
- * Initialize the color module.
+/**
+ * amdgpu_dm_init_color_mod - Initialize the color module.
  *
  * We're not using the full color module, only certain components.
  * Only call setup functions for components that we need.
@@ -82,7 +169,255 @@ void amdgpu_dm_init_color_mod(void)
 	setup_x_points_distribution();
 }
 
-/* Extracts the DRM lut and lut size from a blob. */
+static inline struct fixed31_32 amdgpu_dm_fixpt_from_s3132(__u64 x)
+{
+	struct fixed31_32 val;
+
+	/* If negative, convert to 2's complement. */
+	if (x & (1ULL << 63))
+		x = -(x & ~(1ULL << 63));
+
+	val.value = x;
+	return val;
+}
+
+#ifdef AMD_PRIVATE_COLOR
+/* Pre-defined Transfer Functions (TF)
+ *
+ * AMD driver supports pre-defined mathematical functions for transferring
+ * between encoded values and optical/linear space. Depending on HW color caps,
+ * ROMs and curves built by the AMD color module support these transforms.
+ *
+ * The driver-specific color implementation exposes properties for pre-blending
+ * degamma TF, shaper TF (before 3D LUT), and blend(dpp.ogam) TF and
+ * post-blending regamma (mpc.ogam) TF. However, only pre-blending degamma
+ * supports ROM curves. AMD color module uses pre-defined coefficients to build
+ * curves for the other blocks. What can be done by each color block is
+ * described by struct dpp_color_capsand struct mpc_color_caps.
+ *
+ * AMD driver-specific color API exposes the following pre-defined transfer
+ * functions:
+ *
+ * - Identity: linear/identity relationship between pixel value and
+ *   luminance value;
+ * - Gamma 2.2, Gamma 2.4, Gamma 2.6: pure power functions;
+ * - sRGB: 2.4: The piece-wise transfer function from IEC 61966-2-1:1999;
+ * - BT.709: has a linear segment in the bottom part and then a power function
+ *   with a 0.45 (~1/2.22) gamma for the rest of the range; standardized by
+ *   ITU-R BT.709-6;
+ * - PQ (Perceptual Quantizer): used for HDR display, allows luminance range
+ *   capability of 0 to 10,000 nits; standardized by SMPTE ST 2084.
+ *
+ * The AMD color model is designed with an assumption that SDR (sRGB, BT.709,
+ * Gamma 2.2, etc.) peak white maps (normalized to 1.0 FP) to 80 nits in the PQ
+ * system. This has the implication that PQ EOTF (non-linear to linear) maps to
+ * [0.0..125.0] where 125.0 = 10,000 nits / 80 nits.
+ *
+ * Non-linear and linear forms are described in the table below:
+ *
+ * ┌───────────┬─────────────────────┬──────────────────────┐
+ * │           │     Non-linear      │   Linear             │
+ * ├───────────┼─────────────────────┼──────────────────────┤
+ * │      sRGB │ UNORM or [0.0, 1.0] │ [0.0, 1.0]           │
+ * ├───────────┼─────────────────────┼──────────────────────┤
+ * │     BT709 │ UNORM or [0.0, 1.0] │ [0.0, 1.0]           │
+ * ├───────────┼─────────────────────┼──────────────────────┤
+ * │ Gamma 2.x │ UNORM or [0.0, 1.0] │ [0.0, 1.0]           │
+ * ├───────────┼─────────────────────┼──────────────────────┤
+ * │        PQ │ UNORM or FP16 CCCS* │ [0.0, 125.0]         │
+ * ├───────────┼─────────────────────┼──────────────────────┤
+ * │  Identity │ UNORM or FP16 CCCS* │ [0.0, 1.0] or CCCS** │
+ * └───────────┴─────────────────────┴──────────────────────┘
+ * * CCCS: Windows canonical composition color space
+ * ** Respectively
+ *
+ * In the driver-specific API, color block names attached to TF properties
+ * suggest the intention regarding non-linear encoding pixel's luminance
+ * values. As some newer encodings don't use gamma curve, we make encoding and
+ * decoding explicit by defining an enum list of transfer functions supported
+ * in terms of EOTF and inverse EOTF, where:
+ *
+ * - EOTF (electro-optical transfer function): is the transfer function to go
+ *   from the encoded value to an optical (linear) value. De-gamma functions
+ *   traditionally do this.
+ * - Inverse EOTF (simply the inverse of the EOTF): is usually intended to go
+ *   from an optical/linear space (which might have been used for blending)
+ *   back to the encoded values. Gamma functions traditionally do this.
+ */
+static const char * const
+amdgpu_transfer_function_names[] = {
+	[AMDGPU_TRANSFER_FUNCTION_DEFAULT]		= "Default",
+	[AMDGPU_TRANSFER_FUNCTION_IDENTITY]		= "Identity",
+	[AMDGPU_TRANSFER_FUNCTION_SRGB_EOTF]		= "sRGB EOTF",
+	[AMDGPU_TRANSFER_FUNCTION_BT709_INV_OETF]	= "BT.709 inv_OETF",
+	[AMDGPU_TRANSFER_FUNCTION_PQ_EOTF]		= "PQ EOTF",
+	[AMDGPU_TRANSFER_FUNCTION_GAMMA22_EOTF]		= "Gamma 2.2 EOTF",
+	[AMDGPU_TRANSFER_FUNCTION_GAMMA24_EOTF]		= "Gamma 2.4 EOTF",
+	[AMDGPU_TRANSFER_FUNCTION_GAMMA26_EOTF]		= "Gamma 2.6 EOTF",
+	[AMDGPU_TRANSFER_FUNCTION_SRGB_INV_EOTF]	= "sRGB inv_EOTF",
+	[AMDGPU_TRANSFER_FUNCTION_BT709_OETF]		= "BT.709 OETF",
+	[AMDGPU_TRANSFER_FUNCTION_PQ_INV_EOTF]		= "PQ inv_EOTF",
+	[AMDGPU_TRANSFER_FUNCTION_GAMMA22_INV_EOTF]	= "Gamma 2.2 inv_EOTF",
+	[AMDGPU_TRANSFER_FUNCTION_GAMMA24_INV_EOTF]	= "Gamma 2.4 inv_EOTF",
+	[AMDGPU_TRANSFER_FUNCTION_GAMMA26_INV_EOTF]	= "Gamma 2.6 inv_EOTF",
+};
+
+static const u32 amdgpu_eotf =
+	BIT(AMDGPU_TRANSFER_FUNCTION_SRGB_EOTF) |
+	BIT(AMDGPU_TRANSFER_FUNCTION_BT709_INV_OETF) |
+	BIT(AMDGPU_TRANSFER_FUNCTION_PQ_EOTF) |
+	BIT(AMDGPU_TRANSFER_FUNCTION_GAMMA22_EOTF) |
+	BIT(AMDGPU_TRANSFER_FUNCTION_GAMMA24_EOTF) |
+	BIT(AMDGPU_TRANSFER_FUNCTION_GAMMA26_EOTF);
+
+static const u32 amdgpu_inv_eotf =
+	BIT(AMDGPU_TRANSFER_FUNCTION_SRGB_INV_EOTF) |
+	BIT(AMDGPU_TRANSFER_FUNCTION_BT709_OETF) |
+	BIT(AMDGPU_TRANSFER_FUNCTION_PQ_INV_EOTF) |
+	BIT(AMDGPU_TRANSFER_FUNCTION_GAMMA22_INV_EOTF) |
+	BIT(AMDGPU_TRANSFER_FUNCTION_GAMMA24_INV_EOTF) |
+	BIT(AMDGPU_TRANSFER_FUNCTION_GAMMA26_INV_EOTF);
+
+static struct drm_property *
+amdgpu_create_tf_property(struct drm_device *dev,
+			  const char *name,
+			  u32 supported_tf)
+{
+	u32 transfer_functions = supported_tf |
+				 BIT(AMDGPU_TRANSFER_FUNCTION_DEFAULT) |
+				 BIT(AMDGPU_TRANSFER_FUNCTION_IDENTITY);
+	struct drm_prop_enum_list enum_list[AMDGPU_TRANSFER_FUNCTION_COUNT];
+	int i, len;
+
+	len = 0;
+	for (i = 0; i < AMDGPU_TRANSFER_FUNCTION_COUNT; i++) {
+		if ((transfer_functions & BIT(i)) == 0)
+			continue;
+
+		enum_list[len].type = i;
+		enum_list[len].name = amdgpu_transfer_function_names[i];
+		len++;
+	}
+
+	return drm_property_create_enum(dev, DRM_MODE_PROP_ENUM,
+					name, enum_list, len);
+}
+
+int
+amdgpu_dm_create_color_properties(struct amdgpu_device *adev)
+{
+	struct drm_property *prop;
+
+	prop = drm_property_create(adev_to_drm(adev),
+				   DRM_MODE_PROP_BLOB,
+				   "AMD_PLANE_DEGAMMA_LUT", 0);
+	if (!prop)
+		return -ENOMEM;
+	adev->mode_info.plane_degamma_lut_property = prop;
+
+	prop = drm_property_create_range(adev_to_drm(adev),
+					 DRM_MODE_PROP_IMMUTABLE,
+					 "AMD_PLANE_DEGAMMA_LUT_SIZE",
+					 0, UINT_MAX);
+	if (!prop)
+		return -ENOMEM;
+	adev->mode_info.plane_degamma_lut_size_property = prop;
+
+	prop = amdgpu_create_tf_property(adev_to_drm(adev),
+					 "AMD_PLANE_DEGAMMA_TF",
+					 amdgpu_eotf);
+	if (!prop)
+		return -ENOMEM;
+	adev->mode_info.plane_degamma_tf_property = prop;
+
+	prop = drm_property_create_range(adev_to_drm(adev),
+					 0, "AMD_PLANE_HDR_MULT", 0, U64_MAX);
+	if (!prop)
+		return -ENOMEM;
+	adev->mode_info.plane_hdr_mult_property = prop;
+
+	prop = drm_property_create(adev_to_drm(adev),
+				   DRM_MODE_PROP_BLOB,
+				   "AMD_PLANE_CTM", 0);
+	if (!prop)
+		return -ENOMEM;
+	adev->mode_info.plane_ctm_property = prop;
+
+	prop = drm_property_create(adev_to_drm(adev),
+				   DRM_MODE_PROP_BLOB,
+				   "AMD_PLANE_SHAPER_LUT", 0);
+	if (!prop)
+		return -ENOMEM;
+	adev->mode_info.plane_shaper_lut_property = prop;
+
+	prop = drm_property_create_range(adev_to_drm(adev),
+					 DRM_MODE_PROP_IMMUTABLE,
+					 "AMD_PLANE_SHAPER_LUT_SIZE", 0, UINT_MAX);
+	if (!prop)
+		return -ENOMEM;
+	adev->mode_info.plane_shaper_lut_size_property = prop;
+
+	prop = amdgpu_create_tf_property(adev_to_drm(adev),
+					 "AMD_PLANE_SHAPER_TF",
+					 amdgpu_inv_eotf);
+	if (!prop)
+		return -ENOMEM;
+	adev->mode_info.plane_shaper_tf_property = prop;
+
+	prop = drm_property_create(adev_to_drm(adev),
+				   DRM_MODE_PROP_BLOB,
+				   "AMD_PLANE_LUT3D", 0);
+	if (!prop)
+		return -ENOMEM;
+	adev->mode_info.plane_lut3d_property = prop;
+
+	prop = drm_property_create_range(adev_to_drm(adev),
+					 DRM_MODE_PROP_IMMUTABLE,
+					 "AMD_PLANE_LUT3D_SIZE", 0, UINT_MAX);
+	if (!prop)
+		return -ENOMEM;
+	adev->mode_info.plane_lut3d_size_property = prop;
+
+	prop = drm_property_create(adev_to_drm(adev),
+				   DRM_MODE_PROP_BLOB,
+				   "AMD_PLANE_BLEND_LUT", 0);
+	if (!prop)
+		return -ENOMEM;
+	adev->mode_info.plane_blend_lut_property = prop;
+
+	prop = drm_property_create_range(adev_to_drm(adev),
+					 DRM_MODE_PROP_IMMUTABLE,
+					 "AMD_PLANE_BLEND_LUT_SIZE", 0, UINT_MAX);
+	if (!prop)
+		return -ENOMEM;
+	adev->mode_info.plane_blend_lut_size_property = prop;
+
+	prop = amdgpu_create_tf_property(adev_to_drm(adev),
+					 "AMD_PLANE_BLEND_TF",
+					 amdgpu_eotf);
+	if (!prop)
+		return -ENOMEM;
+	adev->mode_info.plane_blend_tf_property = prop;
+
+	prop = amdgpu_create_tf_property(adev_to_drm(adev),
+					 "AMD_CRTC_REGAMMA_TF",
+					 amdgpu_inv_eotf);
+	if (!prop)
+		return -ENOMEM;
+	adev->mode_info.regamma_tf_property = prop;
+
+	return 0;
+}
+#endif
+
+/**
+ * __extract_blob_lut - Extracts the DRM lut and lut size from a blob.
+ * @blob: DRM color mgmt property blob
+ * @size: lut size
+ *
+ * Returns:
+ * DRM LUT or NULL
+ */
 static const struct drm_color_lut *
 __extract_blob_lut(const struct drm_property_blob *blob, uint32_t *size)
 {
@@ -90,13 +425,33 @@ __extract_blob_lut(const struct drm_property_blob *blob, uint32_t *size)
 	return blob ? (struct drm_color_lut *)blob->data : NULL;
 }
 
-/*
- * Return true if the given lut is a linear mapping of values, i.e. it acts
- * like a bypass LUT.
+/**
+ * __extract_blob_lut32 - Extracts the DRM lut and lut size from a blob.
+ * @blob: DRM color mgmt property blob
+ * @size: lut size
+ *
+ * Returns:
+ * DRM LUT or NULL
+ */
+static const struct drm_color_lut32 *
+__extract_blob_lut32(const struct drm_property_blob *blob, uint32_t *size)
+{
+	*size = blob ? drm_color_lut32_size(blob) : 0;
+	return blob ? (struct drm_color_lut32 *)blob->data : NULL;
+}
+
+/**
+ * __is_lut_linear - check if the given lut is a linear mapping of values
+ * @lut: given lut to check values
+ * @size: lut size
  *
  * It is considered linear if the lut represents:
- * f(a) = (0xFF00/MAX_COLOR_LUT_ENTRIES-1)a; for integer a in
- *                                           [0, MAX_COLOR_LUT_ENTRIES)
+ * f(a) = (0xFF00/MAX_COLOR_LUT_ENTRIES-1)a; for integer a in [0,
+ * MAX_COLOR_LUT_ENTRIES)
+ *
+ * Returns:
+ * True if the given lut is a linear mapping of values, i.e. it acts like a
+ * bypass LUT. Otherwise, false.
  */
 static bool __is_lut_linear(const struct drm_color_lut *lut, uint32_t size)
 {
@@ -119,9 +474,13 @@ static bool __is_lut_linear(const struct drm_color_lut *lut, uint32_t size)
 	return true;
 }
 
-/*
- * Convert the drm_color_lut to dc_gamma. The conversion depends on the size
- * of the lut - whether or not it's legacy.
+/**
+ * __drm_lut_to_dc_gamma - convert the drm_color_lut to dc_gamma.
+ * @lut: DRM lookup table for color conversion
+ * @gamma: DC gamma to set entries
+ * @is_legacy: legacy or atomic gamma
+ *
+ * The conversion depends on the size of the lut - whether or not it's legacy.
  */
 static void __drm_lut_to_dc_gamma(const struct drm_color_lut *lut,
 				  struct dc_gamma *gamma, bool is_legacy)
@@ -154,14 +513,34 @@ static void __drm_lut_to_dc_gamma(const struct drm_color_lut *lut,
 	}
 }
 
-/*
- * Converts a DRM CTM to a DC CSC float matrix.
+/**
+ * __drm_lut32_to_dc_gamma - convert the drm_color_lut to dc_gamma.
+ * @lut: DRM lookup table for color conversion
+ * @gamma: DC gamma to set entries
+ *
+ * The conversion depends on the size of the lut - whether or not it's legacy.
+ */
+static void __drm_lut32_to_dc_gamma(const struct drm_color_lut32 *lut, struct dc_gamma *gamma)
+{
+	int i;
+
+	for (i = 0; i < MAX_COLOR_LUT_ENTRIES; i++) {
+		gamma->entries.red[i] = dc_fixpt_from_fraction(lut[i].red, MAX_DRM_LUT32_VALUE);
+		gamma->entries.green[i] = dc_fixpt_from_fraction(lut[i].green, MAX_DRM_LUT32_VALUE);
+		gamma->entries.blue[i] = dc_fixpt_from_fraction(lut[i].blue, MAX_DRM_LUT32_VALUE);
+	}
+}
+
+/**
+ * __drm_ctm_to_dc_matrix - converts a DRM CTM to a DC CSC float matrix
+ * @ctm: DRM color transformation matrix
+ * @matrix: DC CSC float matrix
+ *
  * The matrix needs to be a 3x4 (12 entry) matrix.
  */
 static void __drm_ctm_to_dc_matrix(const struct drm_color_ctm *ctm,
 				   struct fixed31_32 *matrix)
 {
-	int64_t val;
 	int i;
 
 	/*
@@ -180,16 +559,44 @@ static void __drm_ctm_to_dc_matrix(const struct drm_color_ctm *ctm,
 		}
 
 		/* gamut_remap_matrix[i] = ctm[i - floor(i/4)] */
-		val = ctm->matrix[i - (i / 4)];
-		/* If negative, convert to 2's complement. */
-		if (val & (1ULL << 63))
-			val = -(val & ~(1ULL << 63));
+		matrix[i] = amdgpu_dm_fixpt_from_s3132(ctm->matrix[i - (i / 4)]);
+	}
+}
+
+/**
+ * __drm_ctm_3x4_to_dc_matrix - converts a DRM CTM 3x4 to a DC CSC float matrix
+ * @ctm: DRM color transformation matrix with 3x4 dimensions
+ * @matrix: DC CSC float matrix
+ *
+ * The matrix needs to be a 3x4 (12 entry) matrix.
+ */
+static void __drm_ctm_3x4_to_dc_matrix(const struct drm_color_ctm_3x4 *ctm,
+				       struct fixed31_32 *matrix)
+{
+	int i;
 
-		matrix[i].value = val;
+	/* The format provided is S31.32, using signed-magnitude representation.
+	 * Our fixed31_32 is also S31.32, but is using 2's complement. We have
+	 * to convert from signed-magnitude to 2's complement.
+	 */
+	for (i = 0; i < 12; i++) {
+		/* gamut_remap_matrix[i] = ctm[i - floor(i/4)] */
+		matrix[i] = amdgpu_dm_fixpt_from_s3132(ctm->matrix[i]);
 	}
 }
 
-/* Calculates the legacy transfer function - only for sRGB input space. */
+/**
+ * __set_legacy_tf - Calculates the legacy transfer function
+ * @func: transfer function
+ * @lut: lookup table that defines the color space
+ * @lut_size: size of respective lut
+ * @has_rom: if ROM can be used for hardcoded curve
+ *
+ * Only for sRGB input space
+ *
+ * Returns:
+ * 0 in case of success, -ENOMEM if fails
+ */
 static int __set_legacy_tf(struct dc_transfer_func *func,
 			   const struct drm_color_lut *lut, uint32_t lut_size,
 			   bool has_rom)
@@ -218,7 +625,16 @@ static int __set_legacy_tf(struct dc_transfer_func *func,
 	return res ? 0 : -ENOMEM;
 }
 
-/* Calculates the output transfer function based on expected input space. */
+/**
+ * __set_output_tf - calculates the output transfer function based on expected input space.
+ * @func: transfer function
+ * @lut: lookup table that defines the color space
+ * @lut_size: size of respective lut
+ * @has_rom: if ROM can be used for hardcoded curve
+ *
+ * Returns:
+ * 0 in case of success. -ENOMEM if fails.
+ */
 static int __set_output_tf(struct dc_transfer_func *func,
 			   const struct drm_color_lut *lut, uint32_t lut_size,
 			   bool has_rom)
@@ -227,16 +643,74 @@ static int __set_output_tf(struct dc_transfer_func *func,
 	struct calculate_buffer cal_buffer = {0};
 	bool res;
 
-	ASSERT(lut && lut_size == MAX_COLOR_LUT_ENTRIES);
+	cal_buffer.buffer_index = -1;
+
+	if (lut_size) {
+		ASSERT(lut && lut_size == MAX_COLOR_LUT_ENTRIES);
+
+		gamma = dc_create_gamma();
+		if (!gamma)
+			return -ENOMEM;
+
+		gamma->num_entries = lut_size;
+		__drm_lut_to_dc_gamma(lut, gamma, false);
+	}
+
+	if (func->tf == TRANSFER_FUNCTION_LINEAR) {
+		/*
+		 * Color module doesn't like calculating regamma params
+		 * on top of a linear input. But degamma params can be used
+		 * instead to simulate this.
+		 */
+		if (gamma)
+			gamma->type = GAMMA_CUSTOM;
+		res = mod_color_calculate_degamma_params(NULL, func,
+							 gamma, gamma != NULL);
+	} else {
+		/*
+		 * Assume sRGB. The actual mapping will depend on whether the
+		 * input was legacy or not.
+		 */
+		if (gamma)
+			gamma->type = GAMMA_CS_TFM_1D;
+		res = mod_color_calculate_regamma_params(func, gamma, gamma != NULL,
+							 has_rom, NULL, &cal_buffer);
+	}
+
+	if (gamma)
+		dc_gamma_release(&gamma);
+
+	return res ? 0 : -ENOMEM;
+}
+
+/**
+ * __set_output_tf_32 - calculates the output transfer function based on expected input space.
+ * @func: transfer function
+ * @lut: lookup table that defines the color space
+ * @lut_size: size of respective lut
+ * @has_rom: if ROM can be used for hardcoded curve
+ *
+ * Returns:
+ * 0 in case of success. -ENOMEM if fails.
+ */
+static int __set_output_tf_32(struct dc_transfer_func *func,
+			      const struct drm_color_lut32 *lut, uint32_t lut_size,
+			      bool has_rom)
+{
+	struct dc_gamma *gamma = NULL;
+	struct calculate_buffer cal_buffer = {0};
+	bool res;
 
 	cal_buffer.buffer_index = -1;
 
-	gamma = dc_create_gamma();
-	if (!gamma)
-		return -ENOMEM;
+	if (lut_size) {
+		gamma = dc_create_gamma();
+		if (!gamma)
+			return -ENOMEM;
 
-	gamma->num_entries = lut_size;
-	__drm_lut_to_dc_gamma(lut, gamma, false);
+		gamma->num_entries = lut_size;
+		__drm_lut32_to_dc_gamma(lut, gamma);
+	}
 
 	if (func->tf == TRANSFER_FUNCTION_LINEAR) {
 		/*
@@ -244,50 +718,428 @@ static int __set_output_tf(struct dc_transfer_func *func,
 		 * on top of a linear input. But degamma params can be used
 		 * instead to simulate this.
 		 */
-		gamma->type = GAMMA_CUSTOM;
+		if (gamma)
+			gamma->type = GAMMA_CUSTOM;
 		res = mod_color_calculate_degamma_params(NULL, func,
-							gamma, true);
+							 gamma, gamma != NULL);
 	} else {
 		/*
 		 * Assume sRGB. The actual mapping will depend on whether the
 		 * input was legacy or not.
 		 */
-		gamma->type = GAMMA_CS_TFM_1D;
-		res = mod_color_calculate_regamma_params(func, gamma, false,
+		if (gamma)
+			gamma->type = GAMMA_CS_TFM_1D;
+		res = mod_color_calculate_regamma_params(func, gamma, gamma != NULL,
 							 has_rom, NULL, &cal_buffer);
 	}
 
-	dc_gamma_release(&gamma);
+	if (gamma)
+		dc_gamma_release(&gamma);
 
 	return res ? 0 : -ENOMEM;
 }
 
-/* Caculates the input transfer function based on expected input space. */
-static int __set_input_tf(struct dc_transfer_func *func,
+
+static int amdgpu_dm_set_atomic_regamma(struct dc_transfer_func *out_tf,
+					const struct drm_color_lut *regamma_lut,
+					uint32_t regamma_size, bool has_rom,
+					enum dc_transfer_func_predefined tf)
+{
+	int ret = 0;
+
+	if (regamma_size || tf != TRANSFER_FUNCTION_LINEAR) {
+		/*
+		 * CRTC RGM goes into RGM LUT.
+		 *
+		 * Note: there is no implicit sRGB regamma here. We are using
+		 * degamma calculation from color module to calculate the curve
+		 * from a linear base if gamma TF is not set. However, if gamma
+		 * TF (!= Linear) and LUT are set at the same time, we will use
+		 * regamma calculation, and the color module will combine the
+		 * pre-defined TF and the custom LUT values into the LUT that's
+		 * actually programmed.
+		 */
+		out_tf->type = TF_TYPE_DISTRIBUTED_POINTS;
+		out_tf->tf = tf;
+		out_tf->sdr_ref_white_level = SDR_WHITE_LEVEL_INIT_VALUE;
+
+		ret = __set_output_tf(out_tf, regamma_lut, regamma_size, has_rom);
+	} else {
+		/*
+		 * No CRTC RGM means we can just put the block into bypass
+		 * since we don't have any plane level adjustments using it.
+		 */
+		out_tf->type = TF_TYPE_BYPASS;
+		out_tf->tf = TRANSFER_FUNCTION_LINEAR;
+	}
+
+	return ret;
+}
+
+/**
+ * __set_input_tf - calculates the input transfer function based on expected
+ * input space.
+ * @caps: dc color capabilities
+ * @func: transfer function
+ * @lut: lookup table that defines the color space
+ * @lut_size: size of respective lut.
+ *
+ * Returns:
+ * 0 in case of success. -ENOMEM if fails.
+ */
+static int __set_input_tf(struct dc_color_caps *caps, struct dc_transfer_func *func,
 			  const struct drm_color_lut *lut, uint32_t lut_size)
 {
 	struct dc_gamma *gamma = NULL;
 	bool res;
 
-	gamma = dc_create_gamma();
-	if (!gamma)
-		return -ENOMEM;
+	if (lut_size) {
+		gamma = dc_create_gamma();
+		if (!gamma)
+			return -ENOMEM;
 
-	gamma->type = GAMMA_CUSTOM;
-	gamma->num_entries = lut_size;
+		gamma->type = GAMMA_CUSTOM;
+		gamma->num_entries = lut_size;
 
-	__drm_lut_to_dc_gamma(lut, gamma, false);
+		__drm_lut_to_dc_gamma(lut, gamma, false);
+	}
 
-	res = mod_color_calculate_degamma_params(NULL, func, gamma, true);
-	dc_gamma_release(&gamma);
+	res = mod_color_calculate_degamma_params(caps, func, gamma, gamma != NULL);
+
+	if (gamma)
+		dc_gamma_release(&gamma);
+
+	return res ? 0 : -ENOMEM;
+}
+
+/**
+ * __set_input_tf_32 - calculates the input transfer function based on expected
+ * input space.
+ * @caps: dc color capabilities
+ * @func: transfer function
+ * @lut: lookup table that defines the color space
+ * @lut_size: size of respective lut.
+ *
+ * Returns:
+ * 0 in case of success. -ENOMEM if fails.
+ */
+static int __set_input_tf_32(struct dc_color_caps *caps, struct dc_transfer_func *func,
+			     const struct drm_color_lut32 *lut, uint32_t lut_size)
+{
+	struct dc_gamma *gamma = NULL;
+	bool res;
+
+	if (lut_size) {
+		gamma = dc_create_gamma();
+		if (!gamma)
+			return -ENOMEM;
+
+		gamma->type = GAMMA_CUSTOM;
+		gamma->num_entries = lut_size;
+
+		__drm_lut32_to_dc_gamma(lut, gamma);
+	}
+
+	res = mod_color_calculate_degamma_params(caps, func, gamma, gamma != NULL);
+
+	if (gamma)
+		dc_gamma_release(&gamma);
 
 	return res ? 0 : -ENOMEM;
 }
 
+static enum dc_transfer_func_predefined
+amdgpu_tf_to_dc_tf(enum amdgpu_transfer_function tf)
+{
+	switch (tf) {
+	default:
+	case AMDGPU_TRANSFER_FUNCTION_DEFAULT:
+	case AMDGPU_TRANSFER_FUNCTION_IDENTITY:
+		return TRANSFER_FUNCTION_LINEAR;
+	case AMDGPU_TRANSFER_FUNCTION_SRGB_EOTF:
+	case AMDGPU_TRANSFER_FUNCTION_SRGB_INV_EOTF:
+		return TRANSFER_FUNCTION_SRGB;
+	case AMDGPU_TRANSFER_FUNCTION_BT709_OETF:
+	case AMDGPU_TRANSFER_FUNCTION_BT709_INV_OETF:
+		return TRANSFER_FUNCTION_BT709;
+	case AMDGPU_TRANSFER_FUNCTION_PQ_EOTF:
+	case AMDGPU_TRANSFER_FUNCTION_PQ_INV_EOTF:
+		return TRANSFER_FUNCTION_PQ;
+	case AMDGPU_TRANSFER_FUNCTION_GAMMA22_EOTF:
+	case AMDGPU_TRANSFER_FUNCTION_GAMMA22_INV_EOTF:
+		return TRANSFER_FUNCTION_GAMMA22;
+	case AMDGPU_TRANSFER_FUNCTION_GAMMA24_EOTF:
+	case AMDGPU_TRANSFER_FUNCTION_GAMMA24_INV_EOTF:
+		return TRANSFER_FUNCTION_GAMMA24;
+	case AMDGPU_TRANSFER_FUNCTION_GAMMA26_EOTF:
+	case AMDGPU_TRANSFER_FUNCTION_GAMMA26_INV_EOTF:
+		return TRANSFER_FUNCTION_GAMMA26;
+	}
+}
+
+static enum dc_transfer_func_predefined
+amdgpu_colorop_tf_to_dc_tf(enum drm_colorop_curve_1d_type tf)
+{
+	switch (tf) {
+	case DRM_COLOROP_1D_CURVE_SRGB_EOTF:
+	case DRM_COLOROP_1D_CURVE_SRGB_INV_EOTF:
+		return TRANSFER_FUNCTION_SRGB;
+	case DRM_COLOROP_1D_CURVE_PQ_125_EOTF:
+	case DRM_COLOROP_1D_CURVE_PQ_125_INV_EOTF:
+		return TRANSFER_FUNCTION_PQ;
+	case DRM_COLOROP_1D_CURVE_BT2020_INV_OETF:
+	case DRM_COLOROP_1D_CURVE_BT2020_OETF:
+		return TRANSFER_FUNCTION_BT709;
+	case DRM_COLOROP_1D_CURVE_GAMMA22:
+	case DRM_COLOROP_1D_CURVE_GAMMA22_INV:
+		return TRANSFER_FUNCTION_GAMMA22;
+	default:
+		return TRANSFER_FUNCTION_LINEAR;
+	}
+}
+
+static void __to_dc_lut3d_color(struct dc_rgb *rgb,
+				const struct drm_color_lut lut,
+				int bit_precision)
+{
+	rgb->red = drm_color_lut_extract(lut.red, bit_precision);
+	rgb->green = drm_color_lut_extract(lut.green, bit_precision);
+	rgb->blue  = drm_color_lut_extract(lut.blue, bit_precision);
+}
+
+static void __drm_3dlut_to_dc_3dlut(const struct drm_color_lut *lut,
+				    uint32_t lut3d_size,
+				    struct tetrahedral_params *params,
+				    bool use_tetrahedral_9,
+				    int bit_depth)
+{
+	struct dc_rgb *lut0;
+	struct dc_rgb *lut1;
+	struct dc_rgb *lut2;
+	struct dc_rgb *lut3;
+	int lut_i, i;
+
+
+	if (use_tetrahedral_9) {
+		lut0 = params->tetrahedral_9.lut0;
+		lut1 = params->tetrahedral_9.lut1;
+		lut2 = params->tetrahedral_9.lut2;
+		lut3 = params->tetrahedral_9.lut3;
+	} else {
+		lut0 = params->tetrahedral_17.lut0;
+		lut1 = params->tetrahedral_17.lut1;
+		lut2 = params->tetrahedral_17.lut2;
+		lut3 = params->tetrahedral_17.lut3;
+	}
+
+	for (lut_i = 0, i = 0; i < lut3d_size - 4; lut_i++, i += 4) {
+		/*
+		 * We should consider the 3D LUT RGB values are distributed
+		 * along four arrays lut0-3 where the first sizes 1229 and the
+		 * other 1228. The bit depth supported for 3dlut channel is
+		 * 12-bit, but DC also supports 10-bit.
+		 *
+		 * TODO: improve color pipeline API to enable the userspace set
+		 * bit depth and 3D LUT size/stride, as specified by VA-API.
+		 */
+		__to_dc_lut3d_color(&lut0[lut_i], lut[i], bit_depth);
+		__to_dc_lut3d_color(&lut1[lut_i], lut[i + 1], bit_depth);
+		__to_dc_lut3d_color(&lut2[lut_i], lut[i + 2], bit_depth);
+		__to_dc_lut3d_color(&lut3[lut_i], lut[i + 3], bit_depth);
+	}
+	/* lut0 has 1229 points (lut_size/4 + 1) */
+	__to_dc_lut3d_color(&lut0[lut_i], lut[i], bit_depth);
+}
+
+static void __to_dc_lut3d_32_color(struct dc_rgb *rgb,
+				   const struct drm_color_lut32 lut,
+				   int bit_precision)
+{
+	rgb->red = drm_color_lut32_extract(lut.red, bit_precision);
+	rgb->green = drm_color_lut32_extract(lut.green, bit_precision);
+	rgb->blue  = drm_color_lut32_extract(lut.blue, bit_precision);
+}
+
+static void __drm_3dlut32_to_dc_3dlut(const struct drm_color_lut32 *lut,
+				       uint32_t lut3d_size,
+				       struct tetrahedral_params *params,
+				       bool use_tetrahedral_9,
+				       int bit_depth)
+{
+	struct dc_rgb *lut0;
+	struct dc_rgb *lut1;
+	struct dc_rgb *lut2;
+	struct dc_rgb *lut3;
+	int lut_i, i;
+
+
+	if (use_tetrahedral_9) {
+		lut0 = params->tetrahedral_9.lut0;
+		lut1 = params->tetrahedral_9.lut1;
+		lut2 = params->tetrahedral_9.lut2;
+		lut3 = params->tetrahedral_9.lut3;
+	} else {
+		lut0 = params->tetrahedral_17.lut0;
+		lut1 = params->tetrahedral_17.lut1;
+		lut2 = params->tetrahedral_17.lut2;
+		lut3 = params->tetrahedral_17.lut3;
+	}
+
+	for (lut_i = 0, i = 0; i < lut3d_size - 4; lut_i++, i += 4) {
+		/*
+		 * We should consider the 3D LUT RGB values are distributed
+		 * along four arrays lut0-3 where the first sizes 1229 and the
+		 * other 1228. The bit depth supported for 3dlut channel is
+		 * 12-bit, but DC also supports 10-bit.
+		 *
+		 * TODO: improve color pipeline API to enable the userspace set
+		 * bit depth and 3D LUT size/stride, as specified by VA-API.
+		 */
+		__to_dc_lut3d_32_color(&lut0[lut_i], lut[i], bit_depth);
+		__to_dc_lut3d_32_color(&lut1[lut_i], lut[i + 1], bit_depth);
+		__to_dc_lut3d_32_color(&lut2[lut_i], lut[i + 2], bit_depth);
+		__to_dc_lut3d_32_color(&lut3[lut_i], lut[i + 3], bit_depth);
+	}
+	/* lut0 has 1229 points (lut_size/4 + 1) */
+	__to_dc_lut3d_32_color(&lut0[lut_i], lut[i], bit_depth);
+}
+
+/* amdgpu_dm_atomic_lut3d - set DRM 3D LUT to DC stream
+ * @drm_lut3d: user 3D LUT
+ * @drm_lut3d_size: size of 3D LUT
+ * @lut3d: DC 3D LUT
+ *
+ * Map user 3D LUT data to DC 3D LUT and all necessary bits to program it
+ * on DCN accordingly.
+ */
+static void amdgpu_dm_atomic_lut3d(const struct drm_color_lut *drm_lut3d,
+				   uint32_t drm_lut3d_size,
+				   struct dc_3dlut *lut)
+{
+	if (!drm_lut3d_size) {
+		lut->state.bits.initialized = 0;
+	} else {
+		/* Stride and bit depth are not programmable by API yet.
+		 * Therefore, only supports 17x17x17 3D LUT (12-bit).
+		 */
+		lut->lut_3d.use_tetrahedral_9 = false;
+		lut->lut_3d.use_12bits = true;
+		lut->state.bits.initialized = 1;
+		__drm_3dlut_to_dc_3dlut(drm_lut3d, drm_lut3d_size, &lut->lut_3d,
+					lut->lut_3d.use_tetrahedral_9,
+					MAX_COLOR_3DLUT_BITDEPTH);
+	}
+}
+
+static int amdgpu_dm_atomic_shaper_lut(const struct drm_color_lut *shaper_lut,
+				       bool has_rom,
+				       enum dc_transfer_func_predefined tf,
+				       uint32_t shaper_size,
+				       struct dc_transfer_func *func_shaper)
+{
+	int ret = 0;
+
+	if (shaper_size || tf != TRANSFER_FUNCTION_LINEAR) {
+		/*
+		 * If user shaper LUT is set, we assume a linear color space
+		 * (linearized by degamma 1D LUT or not).
+		 */
+		func_shaper->type = TF_TYPE_DISTRIBUTED_POINTS;
+		func_shaper->tf = tf;
+		func_shaper->sdr_ref_white_level = SDR_WHITE_LEVEL_INIT_VALUE;
+
+		ret = __set_output_tf(func_shaper, shaper_lut, shaper_size, has_rom);
+	} else {
+		func_shaper->type = TF_TYPE_BYPASS;
+		func_shaper->tf = TRANSFER_FUNCTION_LINEAR;
+	}
+
+	return ret;
+}
+
+static int amdgpu_dm_atomic_blend_lut(const struct drm_color_lut *blend_lut,
+				       bool has_rom,
+				       enum dc_transfer_func_predefined tf,
+				       uint32_t blend_size,
+				       struct dc_transfer_func *func_blend)
+{
+	int ret = 0;
+
+	if (blend_size || tf != TRANSFER_FUNCTION_LINEAR) {
+		/*
+		 * DRM plane gamma LUT or TF means we are linearizing color
+		 * space before blending (similar to degamma programming). As
+		 * we don't have hardcoded curve support, or we use AMD color
+		 * module to fill the parameters that will be translated to HW
+		 * points.
+		 */
+		func_blend->type = TF_TYPE_DISTRIBUTED_POINTS;
+		func_blend->tf = tf;
+		func_blend->sdr_ref_white_level = SDR_WHITE_LEVEL_INIT_VALUE;
+
+		ret = __set_input_tf(NULL, func_blend, blend_lut, blend_size);
+	} else {
+		func_blend->type = TF_TYPE_BYPASS;
+		func_blend->tf = TRANSFER_FUNCTION_LINEAR;
+	}
+
+	return ret;
+}
+
 /**
- * Verifies that the Degamma and Gamma LUTs attached to the |crtc_state| are of
- * the expected size.
- * Returns 0 on success.
+ * amdgpu_dm_verify_lut3d_size - verifies if 3D LUT is supported and if user
+ * shaper and 3D LUTs match the hw supported size
+ * @adev: amdgpu device
+ * @plane_state: the DRM plane state
+ *
+ * Verifies if pre-blending (DPP) 3D LUT is supported by the HW (DCN 2.0 or
+ * newer) and if the user shaper and 3D LUTs match the supported size.
+ *
+ * Returns:
+ * 0 on success. -EINVAL if lut size are invalid.
+ */
+int amdgpu_dm_verify_lut3d_size(struct amdgpu_device *adev,
+				struct drm_plane_state *plane_state)
+{
+	struct dm_plane_state *dm_plane_state = to_dm_plane_state(plane_state);
+	const struct drm_color_lut *shaper = NULL, *lut3d = NULL;
+	uint32_t exp_size, size, dim_size = MAX_COLOR_3DLUT_SIZE;
+	bool has_3dlut = adev->dm.dc->caps.color.dpp.hw_3d_lut || adev->dm.dc->caps.color.mpc.preblend;
+
+	/* shaper LUT is only available if 3D LUT color caps */
+	exp_size = has_3dlut ? MAX_COLOR_LUT_ENTRIES : 0;
+	shaper = __extract_blob_lut(dm_plane_state->shaper_lut, &size);
+
+	if (shaper && size != exp_size) {
+		drm_dbg(&adev->ddev,
+			"Invalid Shaper LUT size. Should be %u but got %u.\n",
+			exp_size, size);
+		return -EINVAL;
+	}
+
+	/* The number of 3D LUT entries is the dimension size cubed */
+	exp_size = has_3dlut ? dim_size * dim_size * dim_size : 0;
+	lut3d = __extract_blob_lut(dm_plane_state->lut3d, &size);
+
+	if (lut3d && size != exp_size) {
+		drm_dbg(&adev->ddev,
+			"Invalid 3D LUT size. Should be %u but got %u.\n",
+			exp_size, size);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+/**
+ * amdgpu_dm_verify_lut_sizes - verifies if DRM luts match the hw supported sizes
+ * @crtc_state: the DRM CRTC state
+ *
+ * Verifies that the Degamma and Gamma LUTs attached to the &crtc_state
+ * are of the expected size.
+ *
+ * Returns:
+ * 0 on success. -EINVAL if any lut sizes are invalid.
  */
 int amdgpu_dm_verify_lut_sizes(const struct drm_crtc_state *crtc_state)
 {
@@ -316,38 +1168,42 @@ int amdgpu_dm_verify_lut_sizes(const struct drm_crtc_state *crtc_state)
 }
 
 /**
- * amdgpu_dm_update_crtc_color_mgmt: Maps DRM color management to DC stream.
+ * amdgpu_dm_check_crtc_color_mgmt: Check if DRM color props are programmable by DC.
  * @crtc: amdgpu_dm crtc state
+ * @check_only: only check color state without update dc stream
  *
- * With no plane level color management properties we're free to use any
- * of the HW blocks as long as the CRTC CTM always comes before the
- * CRTC RGM and after the CRTC DGM.
- *
- * The CRTC RGM block will be placed in the RGM LUT block if it is non-linear.
- * The CRTC DGM block will be placed in the DGM LUT block if it is non-linear.
- * The CRTC CTM will be placed in the gamut remap block if it is non-linear.
+ * This function just verifies CRTC LUT sizes, if there is enough space for
+ * output transfer function and if its parameters can be calculated by AMD
+ * color module. It also adjusts some settings for programming CRTC degamma at
+ * plane stage, using plane DGM block.
  *
  * The RGM block is typically more fully featured and accurate across
  * all ASICs - DCE can't support a custom non-linear CRTC DGM.
  *
  * For supporting both plane level color management and CRTC level color
- * management at once we have to either restrict the usage of CRTC properties
- * or blend adjustments together.
+ * management at once we have to either restrict the usage of some CRTC
+ * properties or blend adjustments together.
  *
- * Returns 0 on success.
+ * Returns:
+ * 0 on success. Error code if validation fails.
  */
-int amdgpu_dm_update_crtc_color_mgmt(struct dm_crtc_state *crtc)
+
+int amdgpu_dm_check_crtc_color_mgmt(struct dm_crtc_state *crtc,
+				    bool check_only)
 {
 	struct dc_stream_state *stream = crtc->stream;
 	struct amdgpu_device *adev = drm_to_adev(crtc->base.state->dev);
 	bool has_rom = adev->asic_type <= CHIP_RAVEN;
-	struct drm_color_ctm *ctm = NULL;
+	struct dc_transfer_func *out_tf;
 	const struct drm_color_lut *degamma_lut, *regamma_lut;
 	uint32_t degamma_size, regamma_size;
 	bool has_regamma, has_degamma;
+	enum dc_transfer_func_predefined tf = TRANSFER_FUNCTION_LINEAR;
 	bool is_legacy;
 	int r;
 
+	tf = amdgpu_tf_to_dc_tf(crtc->regamma_tf);
+
 	r = amdgpu_dm_verify_lut_sizes(&crtc->base);
 	if (r)
 		return r;
@@ -367,6 +1223,14 @@ int amdgpu_dm_update_crtc_color_mgmt(struct dm_crtc_state *crtc)
 	crtc->cm_has_degamma = false;
 	crtc->cm_is_degamma_srgb = false;
 
+	if (check_only) {
+		out_tf = kvzalloc(sizeof(*out_tf), GFP_KERNEL);
+		if (!out_tf)
+			return -ENOMEM;
+	} else {
+		out_tf = &stream->out_transfer_func;
+	}
+
 	/* Setup regamma and degamma. */
 	if (is_legacy) {
 		/*
@@ -381,29 +1245,21 @@ int amdgpu_dm_update_crtc_color_mgmt(struct dm_crtc_state *crtc)
 		 * inverse color ramp in legacy userspace.
 		 */
 		crtc->cm_is_degamma_srgb = true;
-		stream->out_transfer_func->type = TF_TYPE_DISTRIBUTED_POINTS;
-		stream->out_transfer_func->tf = TRANSFER_FUNCTION_SRGB;
-
-		r = __set_legacy_tf(stream->out_transfer_func, regamma_lut,
-				    regamma_size, has_rom);
-		if (r)
-			return r;
-	} else if (has_regamma) {
-		/* CRTC RGM goes into RGM LUT. */
-		stream->out_transfer_func->type = TF_TYPE_DISTRIBUTED_POINTS;
-		stream->out_transfer_func->tf = TRANSFER_FUNCTION_LINEAR;
-
-		r = __set_output_tf(stream->out_transfer_func, regamma_lut,
-				    regamma_size, has_rom);
-		if (r)
-			return r;
-	} else {
+		out_tf->type = TF_TYPE_DISTRIBUTED_POINTS;
+		out_tf->tf = TRANSFER_FUNCTION_SRGB;
 		/*
-		 * No CRTC RGM means we can just put the block into bypass
-		 * since we don't have any plane level adjustments using it.
+		 * Note: although we pass has_rom as parameter here, we never
+		 * actually use ROM because the color module only takes the ROM
+		 * path if transfer_func->type == PREDEFINED.
+		 *
+		 * See more in mod_color_calculate_regamma_params()
 		 */
-		stream->out_transfer_func->type = TF_TYPE_BYPASS;
-		stream->out_transfer_func->tf = TRANSFER_FUNCTION_LINEAR;
+		r = __set_legacy_tf(out_tf, regamma_lut,
+				    regamma_size, has_rom);
+	} else {
+		regamma_size = has_regamma ? regamma_size : 0;
+		r = amdgpu_dm_set_atomic_regamma(out_tf, regamma_lut,
+						 regamma_size, has_rom, tf);
 	}
 
 	/*
@@ -412,6 +1268,43 @@ int amdgpu_dm_update_crtc_color_mgmt(struct dm_crtc_state *crtc)
 	 * have to place the CTM in the OCSC in that case.
 	 */
 	crtc->cm_has_degamma = has_degamma;
+	if (check_only)
+		kvfree(out_tf);
+
+	return r;
+}
+
+/**
+ * amdgpu_dm_update_crtc_color_mgmt: Maps DRM color management to DC stream.
+ * @crtc: amdgpu_dm crtc state
+ *
+ * With no plane level color management properties we're free to use any
+ * of the HW blocks as long as the CRTC CTM always comes before the
+ * CRTC RGM and after the CRTC DGM.
+ *
+ * - The CRTC RGM block will be placed in the RGM LUT block if it is non-linear.
+ * - The CRTC DGM block will be placed in the DGM LUT block if it is non-linear.
+ * - The CRTC CTM will be placed in the gamut remap block if it is non-linear.
+ *
+ * The RGM block is typically more fully featured and accurate across
+ * all ASICs - DCE can't support a custom non-linear CRTC DGM.
+ *
+ * For supporting both plane level color management and CRTC level color
+ * management at once we have to either restrict the usage of CRTC properties
+ * or blend adjustments together.
+ *
+ * Returns:
+ * 0 on success. Error code if setup fails.
+ */
+int amdgpu_dm_update_crtc_color_mgmt(struct dm_crtc_state *crtc)
+{
+	struct dc_stream_state *stream = crtc->stream;
+	struct drm_color_ctm *ctm = NULL;
+	int ret;
+
+	ret = amdgpu_dm_check_crtc_color_mgmt(crtc, false);
+	if (ret)
+		return ret;
 
 	/* Setup CRTC CTM. */
 	if (crtc->base.ctm) {
@@ -439,19 +1332,10 @@ int amdgpu_dm_update_crtc_color_mgmt(struct dm_crtc_state *crtc)
 	return 0;
 }
 
-/**
- * amdgpu_dm_update_plane_color_mgmt: Maps DRM color management to DC plane.
- * @crtc: amdgpu_dm crtc state
- * @dc_plane_state: target DC surface
- *
- * Update the underlying dc_stream_state's input transfer function (ITF) in
- * preparation for hardware commit. The transfer function used depends on
- * the prepartion done on the stream for color management.
- *
- * Returns 0 on success.
- */
-int amdgpu_dm_update_plane_color_mgmt(struct dm_crtc_state *crtc,
-				      struct dc_plane_state *dc_plane_state)
+static int
+map_crtc_degamma_to_dc_plane(struct dm_crtc_state *crtc,
+			     struct dc_plane_state *dc_plane_state,
+			     struct dc_color_caps *caps)
 {
 	const struct drm_color_lut *degamma_lut;
 	enum dc_transfer_func_predefined tf = TRANSFER_FUNCTION_SRGB;
@@ -474,8 +1358,7 @@ int amdgpu_dm_update_plane_color_mgmt(struct dm_crtc_state *crtc,
 						 &degamma_size);
 		ASSERT(degamma_size == MAX_COLOR_LUT_ENTRIES);
 
-		dc_plane_state->in_transfer_func->type =
-			TF_TYPE_DISTRIBUTED_POINTS;
+		dc_plane_state->in_transfer_func.type = TF_TYPE_DISTRIBUTED_POINTS;
 
 		/*
 		 * This case isn't fully correct, but also fairly
@@ -502,32 +1385,671 @@ int amdgpu_dm_update_plane_color_mgmt(struct dm_crtc_state *crtc,
 		 * map these to the atomic one instead.
 		 */
 		if (crtc->cm_is_degamma_srgb)
-			dc_plane_state->in_transfer_func->tf = tf;
+			dc_plane_state->in_transfer_func.tf = tf;
 		else
-			dc_plane_state->in_transfer_func->tf =
+			dc_plane_state->in_transfer_func.tf =
 				TRANSFER_FUNCTION_LINEAR;
 
-		r = __set_input_tf(dc_plane_state->in_transfer_func,
+		r = __set_input_tf(caps, &dc_plane_state->in_transfer_func,
 				   degamma_lut, degamma_size);
 		if (r)
 			return r;
-	} else if (crtc->cm_is_degamma_srgb) {
+	} else {
 		/*
 		 * For legacy gamma support we need the regamma input
 		 * in linear space. Assume that the input is sRGB.
 		 */
-		dc_plane_state->in_transfer_func->type = TF_TYPE_PREDEFINED;
-		dc_plane_state->in_transfer_func->tf = tf;
+		dc_plane_state->in_transfer_func.type = TF_TYPE_PREDEFINED;
+		dc_plane_state->in_transfer_func.tf = tf;
 
 		if (tf != TRANSFER_FUNCTION_SRGB &&
-		    !mod_color_calculate_degamma_params(NULL,
-			    dc_plane_state->in_transfer_func, NULL, false))
+		    !mod_color_calculate_degamma_params(caps,
+							&dc_plane_state->in_transfer_func,
+							NULL, false))
 			return -ENOMEM;
-	} else {
-		/* ...Otherwise we can just bypass the DGM block. */
-		dc_plane_state->in_transfer_func->type = TF_TYPE_BYPASS;
-		dc_plane_state->in_transfer_func->tf = TRANSFER_FUNCTION_LINEAR;
 	}
 
 	return 0;
 }
+
+static int
+__set_dm_plane_degamma(struct drm_plane_state *plane_state,
+		       struct dc_plane_state *dc_plane_state,
+		       struct dc_color_caps *color_caps)
+{
+	struct dm_plane_state *dm_plane_state = to_dm_plane_state(plane_state);
+	const struct drm_color_lut *degamma_lut;
+	enum amdgpu_transfer_function tf = AMDGPU_TRANSFER_FUNCTION_DEFAULT;
+	uint32_t degamma_size;
+	bool has_degamma_lut;
+	int ret;
+
+	degamma_lut = __extract_blob_lut(dm_plane_state->degamma_lut,
+					 &degamma_size);
+
+	has_degamma_lut = degamma_lut &&
+			  !__is_lut_linear(degamma_lut, degamma_size);
+
+	tf = dm_plane_state->degamma_tf;
+
+	/* If we don't have plane degamma LUT nor TF to set on DC, we have
+	 * nothing to do here, return.
+	 */
+	if (!has_degamma_lut && tf == AMDGPU_TRANSFER_FUNCTION_DEFAULT)
+		return -EINVAL;
+
+	dc_plane_state->in_transfer_func.tf = amdgpu_tf_to_dc_tf(tf);
+
+	if (has_degamma_lut) {
+		ASSERT(degamma_size == MAX_COLOR_LUT_ENTRIES);
+
+		dc_plane_state->in_transfer_func.type =
+			TF_TYPE_DISTRIBUTED_POINTS;
+
+		ret = __set_input_tf(color_caps, &dc_plane_state->in_transfer_func,
+				     degamma_lut, degamma_size);
+		if (ret)
+			return ret;
+       } else {
+		dc_plane_state->in_transfer_func.type =
+			TF_TYPE_PREDEFINED;
+
+		if (!mod_color_calculate_degamma_params(color_caps,
+		    &dc_plane_state->in_transfer_func, NULL, false))
+			return -ENOMEM;
+	}
+	return 0;
+}
+
+static int
+__set_colorop_in_tf_1d_curve(struct dc_plane_state *dc_plane_state,
+			     struct drm_colorop_state *colorop_state)
+{
+	struct dc_transfer_func *tf = &dc_plane_state->in_transfer_func;
+	struct drm_colorop *colorop = colorop_state->colorop;
+	struct drm_device *drm = colorop->dev;
+
+	if (colorop->type != DRM_COLOROP_1D_CURVE)
+		return -EINVAL;
+
+	if (!(BIT(colorop_state->curve_1d_type) & amdgpu_dm_supported_degam_tfs))
+		return -EINVAL;
+
+	if (colorop_state->bypass) {
+		tf->type = TF_TYPE_BYPASS;
+		tf->tf = TRANSFER_FUNCTION_LINEAR;
+		return 0;
+	}
+
+	drm_dbg(drm, "Degamma colorop with ID: %d\n", colorop->base.id);
+
+	tf->type = TF_TYPE_PREDEFINED;
+	tf->tf = amdgpu_colorop_tf_to_dc_tf(colorop_state->curve_1d_type);
+
+	return 0;
+}
+
+static int
+__set_dm_plane_colorop_degamma(struct drm_plane_state *plane_state,
+			       struct dc_plane_state *dc_plane_state,
+			       struct drm_colorop *colorop)
+{
+	struct drm_colorop *old_colorop;
+	struct drm_colorop_state *colorop_state = NULL, *new_colorop_state;
+	struct drm_atomic_state *state = plane_state->state;
+	int i = 0;
+
+	old_colorop = colorop;
+
+	/* 1st op: 1d curve - degamma */
+	for_each_new_colorop_in_state(state, colorop, new_colorop_state, i) {
+		if (new_colorop_state->colorop == old_colorop &&
+		    (BIT(new_colorop_state->curve_1d_type) & amdgpu_dm_supported_degam_tfs)) {
+			colorop_state = new_colorop_state;
+			break;
+		}
+	}
+
+	if (!colorop_state)
+		return -EINVAL;
+
+	return __set_colorop_in_tf_1d_curve(dc_plane_state, colorop_state);
+}
+
+static int
+__set_dm_plane_colorop_3x4_matrix(struct drm_plane_state *plane_state,
+				  struct dc_plane_state *dc_plane_state,
+				  struct drm_colorop *colorop)
+{
+	struct drm_colorop *old_colorop;
+	struct drm_colorop_state *colorop_state = NULL, *new_colorop_state;
+	struct drm_atomic_state *state = plane_state->state;
+	const struct drm_device *dev = colorop->dev;
+	const struct drm_property_blob *blob;
+	struct drm_color_ctm_3x4 *ctm = NULL;
+	int i = 0;
+
+	/* 3x4 matrix */
+	old_colorop = colorop;
+	for_each_new_colorop_in_state(state, colorop, new_colorop_state, i) {
+		if (new_colorop_state->colorop == old_colorop &&
+		    new_colorop_state->colorop->type == DRM_COLOROP_CTM_3X4) {
+			colorop_state = new_colorop_state;
+			break;
+		}
+	}
+
+	if (colorop_state && !colorop_state->bypass && colorop->type == DRM_COLOROP_CTM_3X4) {
+		drm_dbg(dev, "3x4 matrix colorop with ID: %d\n", colorop->base.id);
+		blob = colorop_state->data;
+		if (blob->length == sizeof(struct drm_color_ctm_3x4)) {
+			ctm = (struct drm_color_ctm_3x4 *) blob->data;
+			__drm_ctm_3x4_to_dc_matrix(ctm, dc_plane_state->gamut_remap_matrix.matrix);
+			dc_plane_state->gamut_remap_matrix.enable_remap = true;
+			dc_plane_state->input_csc_color_matrix.enable_adjustment = false;
+		} else {
+			drm_warn(dev, "blob->length (%zu) isn't equal to drm_color_ctm_3x4 (%zu)\n",
+				 blob->length, sizeof(struct drm_color_ctm_3x4));
+			return -EINVAL;
+		}
+	}
+
+	return 0;
+}
+
+static int
+__set_dm_plane_colorop_multiplier(struct drm_plane_state *plane_state,
+				  struct dc_plane_state *dc_plane_state,
+				  struct drm_colorop *colorop)
+{
+	struct drm_colorop *old_colorop;
+	struct drm_colorop_state *colorop_state = NULL, *new_colorop_state;
+	struct drm_atomic_state *state = plane_state->state;
+	const struct drm_device *dev = colorop->dev;
+	int i = 0;
+
+	/* Multiplier */
+	old_colorop = colorop;
+	for_each_new_colorop_in_state(state, colorop, new_colorop_state, i) {
+		if (new_colorop_state->colorop == old_colorop &&
+		    new_colorop_state->colorop->type == DRM_COLOROP_MULTIPLIER) {
+			colorop_state = new_colorop_state;
+			break;
+		}
+	}
+
+	if (colorop_state && !colorop_state->bypass && colorop->type == DRM_COLOROP_MULTIPLIER) {
+		drm_dbg(dev, "Multiplier colorop with ID: %d\n", colorop->base.id);
+		dc_plane_state->hdr_mult = amdgpu_dm_fixpt_from_s3132(colorop_state->multiplier);
+	}
+
+	return 0;
+}
+
+static int
+__set_dm_plane_colorop_shaper(struct drm_plane_state *plane_state,
+			      struct dc_plane_state *dc_plane_state,
+			      struct drm_colorop *colorop)
+{
+	struct drm_colorop *old_colorop;
+	struct drm_colorop_state *colorop_state = NULL, *new_colorop_state;
+	struct drm_atomic_state *state = plane_state->state;
+	enum dc_transfer_func_predefined default_tf = TRANSFER_FUNCTION_LINEAR;
+	struct dc_transfer_func *tf = &dc_plane_state->in_shaper_func;
+	const struct drm_color_lut32 *shaper_lut;
+	struct drm_device *dev = colorop->dev;
+	bool enabled = false;
+	u32 shaper_size;
+	int i = 0, ret = 0;
+
+	/* 1D Curve - SHAPER TF */
+	old_colorop = colorop;
+	for_each_new_colorop_in_state(state, colorop, new_colorop_state, i) {
+		if (new_colorop_state->colorop == old_colorop &&
+		    (BIT(new_colorop_state->curve_1d_type) & amdgpu_dm_supported_shaper_tfs)) {
+			colorop_state = new_colorop_state;
+			break;
+		}
+	}
+
+	if (colorop_state && !colorop_state->bypass && colorop->type == DRM_COLOROP_1D_CURVE) {
+		drm_dbg(dev, "Shaper TF colorop with ID: %d\n", colorop->base.id);
+		tf->type = TF_TYPE_DISTRIBUTED_POINTS;
+		tf->tf = default_tf = amdgpu_colorop_tf_to_dc_tf(colorop_state->curve_1d_type);
+		tf->sdr_ref_white_level = SDR_WHITE_LEVEL_INIT_VALUE;
+		ret = __set_output_tf(tf, 0, 0, false);
+		if (ret)
+			return ret;
+		enabled = true;
+	}
+
+	/* 1D LUT - SHAPER LUT */
+	colorop = old_colorop->next;
+	if (!colorop) {
+		drm_dbg(dev, "no Shaper LUT colorop found\n");
+		return -EINVAL;
+	}
+
+	old_colorop = colorop;
+	for_each_new_colorop_in_state(state, colorop, new_colorop_state, i) {
+		if (new_colorop_state->colorop == old_colorop &&
+		    new_colorop_state->colorop->type == DRM_COLOROP_1D_LUT) {
+			colorop_state = new_colorop_state;
+			break;
+		}
+	}
+
+	if (colorop_state && !colorop_state->bypass && colorop->type == DRM_COLOROP_1D_LUT) {
+		drm_dbg(dev, "Shaper LUT colorop with ID: %d\n", colorop->base.id);
+		tf->type = TF_TYPE_DISTRIBUTED_POINTS;
+		tf->tf = default_tf;
+		tf->sdr_ref_white_level = SDR_WHITE_LEVEL_INIT_VALUE;
+		shaper_lut = __extract_blob_lut32(colorop_state->data, &shaper_size);
+		shaper_size = shaper_lut != NULL ? shaper_size : 0;
+
+		/* Custom LUT size must be the same as supported size */
+		if (shaper_size == colorop->size) {
+			ret = __set_output_tf_32(tf, shaper_lut, shaper_size, false);
+			if (ret)
+				return ret;
+			enabled = true;
+		}
+	}
+
+	if (!enabled)
+		tf->type = TF_TYPE_BYPASS;
+
+	return 0;
+}
+
+/* __set_colorop_3dlut - set DRM 3D LUT to DC stream
+ * @drm_lut3d: user 3D LUT
+ * @drm_lut3d_size: size of 3D LUT
+ * @lut3d: DC 3D LUT
+ *
+ * Map user 3D LUT data to DC 3D LUT and all necessary bits to program it
+ * on DCN accordingly.
+ *
+ * Returns:
+ * 0 on success. -EINVAL if drm_lut3d_size is zero.
+ */
+static int __set_colorop_3dlut(const struct drm_color_lut32 *drm_lut3d,
+				uint32_t drm_lut3d_size,
+				struct dc_3dlut *lut)
+{
+	if (!drm_lut3d_size) {
+		lut->state.bits.initialized = 0;
+		return -EINVAL;
+	}
+
+	/* Only supports 17x17x17 3D LUT (12-bit) now */
+	lut->lut_3d.use_12bits = true;
+	lut->lut_3d.use_tetrahedral_9 = false;
+
+	lut->state.bits.initialized = 1;
+	__drm_3dlut32_to_dc_3dlut(drm_lut3d, drm_lut3d_size, &lut->lut_3d,
+				   lut->lut_3d.use_tetrahedral_9, 12);
+
+	return 0;
+}
+
+static int
+__set_dm_plane_colorop_3dlut(struct drm_plane_state *plane_state,
+			     struct dc_plane_state *dc_plane_state,
+			     struct drm_colorop *colorop)
+{
+	struct drm_colorop *old_colorop;
+	struct drm_colorop_state *colorop_state = NULL, *new_colorop_state;
+	struct dc_transfer_func *tf = &dc_plane_state->in_shaper_func;
+	struct drm_atomic_state *state = plane_state->state;
+	const struct amdgpu_device *adev = drm_to_adev(colorop->dev);
+	const struct drm_device *dev = colorop->dev;
+	const struct drm_color_lut32 *lut3d;
+	uint32_t lut3d_size;
+	int i = 0, ret = 0;
+
+	/* 3D LUT */
+	old_colorop = colorop;
+	for_each_new_colorop_in_state(state, colorop, new_colorop_state, i) {
+		if (new_colorop_state->colorop == old_colorop &&
+		    new_colorop_state->colorop->type == DRM_COLOROP_3D_LUT) {
+			colorop_state = new_colorop_state;
+			break;
+		}
+	}
+
+	if (colorop_state && !colorop_state->bypass && colorop->type == DRM_COLOROP_3D_LUT) {
+		if (!adev->dm.dc->caps.color.dpp.hw_3d_lut) {
+			drm_dbg(dev, "3D LUT is not supported by hardware\n");
+			return -EINVAL;
+		}
+
+		drm_dbg(dev, "3D LUT colorop with ID: %d\n", colorop->base.id);
+		lut3d = __extract_blob_lut32(colorop_state->data, &lut3d_size);
+		lut3d_size = lut3d != NULL ? lut3d_size : 0;
+		ret = __set_colorop_3dlut(lut3d, lut3d_size, &dc_plane_state->lut3d_func);
+		if (ret) {
+			drm_dbg(dev, "3D LUT colorop with ID: %d has LUT size = %d\n",
+				colorop->base.id, lut3d_size);
+			return ret;
+		}
+
+		/* 3D LUT requires shaper. If shaper colorop is bypassed, enable shaper curve
+		 * with TRANSFER_FUNCTION_LINEAR
+		 */
+		if (tf->type == TF_TYPE_BYPASS) {
+			tf->type = TF_TYPE_DISTRIBUTED_POINTS;
+			tf->tf = TRANSFER_FUNCTION_LINEAR;
+			tf->sdr_ref_white_level = SDR_WHITE_LEVEL_INIT_VALUE;
+			ret = __set_output_tf_32(tf, NULL, 0, false);
+		}
+	}
+
+	return ret;
+}
+
+static int
+__set_dm_plane_colorop_blend(struct drm_plane_state *plane_state,
+			     struct dc_plane_state *dc_plane_state,
+			     struct drm_colorop *colorop)
+{
+	struct drm_colorop *old_colorop;
+	struct drm_colorop_state *colorop_state = NULL, *new_colorop_state;
+	struct drm_atomic_state *state = plane_state->state;
+	enum dc_transfer_func_predefined default_tf = TRANSFER_FUNCTION_LINEAR;
+	struct dc_transfer_func *tf = &dc_plane_state->blend_tf;
+	const struct drm_color_lut32 *blend_lut = NULL;
+	struct drm_device *dev = colorop->dev;
+	uint32_t blend_size = 0;
+	int i = 0;
+
+	/* 1D Curve - BLND TF */
+	old_colorop = colorop;
+	for_each_new_colorop_in_state(state, colorop, new_colorop_state, i) {
+		if (new_colorop_state->colorop == old_colorop &&
+		    (BIT(new_colorop_state->curve_1d_type) & amdgpu_dm_supported_blnd_tfs)) {
+			colorop_state = new_colorop_state;
+			break;
+		}
+	}
+
+	if (colorop_state && !colorop_state->bypass && colorop->type == DRM_COLOROP_1D_CURVE &&
+	    (BIT(colorop_state->curve_1d_type) & amdgpu_dm_supported_blnd_tfs)) {
+		drm_dbg(dev, "Blend TF colorop with ID: %d\n", colorop->base.id);
+		tf->type = TF_TYPE_DISTRIBUTED_POINTS;
+		tf->tf = default_tf = amdgpu_colorop_tf_to_dc_tf(colorop_state->curve_1d_type);
+		tf->sdr_ref_white_level = SDR_WHITE_LEVEL_INIT_VALUE;
+		__set_input_tf_32(NULL, tf, blend_lut, blend_size);
+	}
+
+	/* 1D Curve - BLND LUT */
+	colorop = old_colorop->next;
+	if (!colorop) {
+		drm_dbg(dev, "no Blend LUT colorop found\n");
+		return -EINVAL;
+	}
+
+	old_colorop = colorop;
+	for_each_new_colorop_in_state(state, colorop, new_colorop_state, i) {
+		if (new_colorop_state->colorop == old_colorop &&
+		    new_colorop_state->colorop->type == DRM_COLOROP_1D_LUT) {
+			colorop_state = new_colorop_state;
+			break;
+		}
+	}
+
+	if (colorop_state && !colorop_state->bypass && colorop->type == DRM_COLOROP_1D_LUT &&
+	    (BIT(colorop_state->curve_1d_type) & amdgpu_dm_supported_blnd_tfs)) {
+		drm_dbg(dev, "Blend LUT colorop with ID: %d\n", colorop->base.id);
+		tf->type = TF_TYPE_DISTRIBUTED_POINTS;
+		tf->tf = default_tf;
+		tf->sdr_ref_white_level = SDR_WHITE_LEVEL_INIT_VALUE;
+		blend_lut = __extract_blob_lut32(colorop_state->data, &blend_size);
+		blend_size = blend_lut != NULL ? blend_size : 0;
+
+		/* Custom LUT size must be the same as supported size */
+		if (blend_size == colorop->size)
+			__set_input_tf_32(NULL, tf, blend_lut, blend_size);
+	}
+
+	return 0;
+}
+
+static int
+amdgpu_dm_plane_set_color_properties(struct drm_plane_state *plane_state,
+				     struct dc_plane_state *dc_plane_state)
+{
+	struct dm_plane_state *dm_plane_state = to_dm_plane_state(plane_state);
+	enum amdgpu_transfer_function shaper_tf = AMDGPU_TRANSFER_FUNCTION_DEFAULT;
+	enum amdgpu_transfer_function blend_tf = AMDGPU_TRANSFER_FUNCTION_DEFAULT;
+	const struct drm_color_lut *shaper_lut, *lut3d, *blend_lut;
+	uint32_t shaper_size, lut3d_size, blend_size;
+	int ret;
+
+	dc_plane_state->hdr_mult = amdgpu_dm_fixpt_from_s3132(dm_plane_state->hdr_mult);
+
+	shaper_lut = __extract_blob_lut(dm_plane_state->shaper_lut, &shaper_size);
+	shaper_size = shaper_lut != NULL ? shaper_size : 0;
+	shaper_tf = dm_plane_state->shaper_tf;
+	lut3d = __extract_blob_lut(dm_plane_state->lut3d, &lut3d_size);
+	lut3d_size = lut3d != NULL ? lut3d_size : 0;
+
+	amdgpu_dm_atomic_lut3d(lut3d, lut3d_size, &dc_plane_state->lut3d_func);
+	ret = amdgpu_dm_atomic_shaper_lut(shaper_lut, false,
+					  amdgpu_tf_to_dc_tf(shaper_tf),
+					  shaper_size,
+					  &dc_plane_state->in_shaper_func);
+	if (ret) {
+		drm_dbg_kms(plane_state->plane->dev,
+			    "setting plane %d shaper LUT failed.\n",
+			    plane_state->plane->index);
+
+		return ret;
+	}
+
+	blend_tf = dm_plane_state->blend_tf;
+	blend_lut = __extract_blob_lut(dm_plane_state->blend_lut, &blend_size);
+	blend_size = blend_lut != NULL ? blend_size : 0;
+
+	ret = amdgpu_dm_atomic_blend_lut(blend_lut, false,
+					 amdgpu_tf_to_dc_tf(blend_tf),
+					 blend_size, &dc_plane_state->blend_tf);
+	if (ret) {
+		drm_dbg_kms(plane_state->plane->dev,
+			    "setting plane %d gamma lut failed.\n",
+			    plane_state->plane->index);
+
+		return ret;
+	}
+
+	return 0;
+}
+
+static int
+amdgpu_dm_plane_set_colorop_properties(struct drm_plane_state *plane_state,
+				       struct dc_plane_state *dc_plane_state)
+{
+	struct drm_colorop *colorop = plane_state->color_pipeline;
+	struct drm_device *dev = plane_state->plane->dev;
+	struct amdgpu_device *adev = drm_to_adev(dev);
+	int ret;
+
+	/* 1D Curve - DEGAM TF */
+	if (!colorop)
+		return -EINVAL;
+
+	ret = __set_dm_plane_colorop_degamma(plane_state, dc_plane_state, colorop);
+	if (ret)
+		return ret;
+
+	/* Multiplier */
+	colorop = colorop->next;
+	if (!colorop) {
+		drm_dbg(dev, "no multiplier colorop found\n");
+		return -EINVAL;
+	}
+
+	ret = __set_dm_plane_colorop_multiplier(plane_state, dc_plane_state, colorop);
+	if (ret)
+		return ret;
+
+	/* 3x4 matrix */
+	colorop = colorop->next;
+	if (!colorop) {
+		drm_dbg(dev, "no 3x4 matrix colorop found\n");
+		return -EINVAL;
+	}
+
+	ret = __set_dm_plane_colorop_3x4_matrix(plane_state, dc_plane_state, colorop);
+	if (ret)
+		return ret;
+
+	if (adev->dm.dc->caps.color.dpp.hw_3d_lut) {
+		/* 1D Curve & LUT - SHAPER TF & LUT */
+		colorop = colorop->next;
+		if (!colorop) {
+			drm_dbg(dev, "no Shaper TF colorop found\n");
+			return -EINVAL;
+		}
+
+		ret = __set_dm_plane_colorop_shaper(plane_state, dc_plane_state, colorop);
+		if (ret)
+			return ret;
+
+		/* Shaper LUT colorop is already handled, just skip here */
+		colorop = colorop->next;
+		if (!colorop)
+			return -EINVAL;
+
+		/* 3D LUT */
+		colorop = colorop->next;
+		if (!colorop) {
+			drm_dbg(dev, "no 3D LUT colorop found\n");
+			return -EINVAL;
+		}
+
+		ret = __set_dm_plane_colorop_3dlut(plane_state, dc_plane_state, colorop);
+		if (ret)
+			return ret;
+	}
+
+	/* 1D Curve & LUT - BLND TF & LUT */
+	colorop = colorop->next;
+	if (!colorop) {
+		drm_dbg(dev, "no Blend TF colorop found\n");
+		return -EINVAL;
+	}
+
+	ret = __set_dm_plane_colorop_blend(plane_state, dc_plane_state, colorop);
+	if (ret)
+		return ret;
+
+	/* BLND LUT colorop is already handled, just skip here */
+	colorop = colorop->next;
+	if (!colorop)
+		return -EINVAL;
+
+	return 0;
+}
+
+/**
+ * amdgpu_dm_update_plane_color_mgmt: Maps DRM color management to DC plane.
+ * @crtc: amdgpu_dm crtc state
+ * @plane_state: DRM plane state
+ * @dc_plane_state: target DC surface
+ *
+ * Update the underlying dc_stream_state's input transfer function (ITF) in
+ * preparation for hardware commit. The transfer function used depends on
+ * the preparation done on the stream for color management.
+ *
+ * Returns:
+ * 0 on success. -ENOMEM if mem allocation fails.
+ */
+int amdgpu_dm_update_plane_color_mgmt(struct dm_crtc_state *crtc,
+				      struct drm_plane_state *plane_state,
+				      struct dc_plane_state *dc_plane_state)
+{
+	struct amdgpu_device *adev = drm_to_adev(crtc->base.state->dev);
+	struct dm_plane_state *dm_plane_state = to_dm_plane_state(plane_state);
+	struct drm_color_ctm_3x4 *ctm = NULL;
+	struct dc_color_caps *color_caps = NULL;
+	bool has_crtc_cm_degamma;
+	int ret;
+
+	ret = amdgpu_dm_verify_lut3d_size(adev, plane_state);
+	if (ret) {
+		drm_dbg_driver(&adev->ddev, "amdgpu_dm_verify_lut3d_size() failed\n");
+		return ret;
+	}
+
+	if (dc_plane_state->ctx && dc_plane_state->ctx->dc)
+		color_caps = &dc_plane_state->ctx->dc->caps.color;
+
+	/* Initially, we can just bypass the DGM block. */
+	dc_plane_state->in_transfer_func.type = TF_TYPE_BYPASS;
+	dc_plane_state->in_transfer_func.tf = TRANSFER_FUNCTION_LINEAR;
+
+	/* After, we start to update values according to color props */
+	has_crtc_cm_degamma = (crtc->cm_has_degamma || crtc->cm_is_degamma_srgb);
+
+	ret = __set_dm_plane_degamma(plane_state, dc_plane_state, color_caps);
+	if (ret == -ENOMEM)
+		return ret;
+
+	/* We only have one degamma block available (pre-blending) for the
+	 * whole color correction pipeline, so that we can't actually perform
+	 * plane and CRTC degamma at the same time. Explicitly reject atomic
+	 * updates when userspace sets both plane and CRTC degamma properties.
+	 */
+	if (has_crtc_cm_degamma && ret != -EINVAL) {
+		drm_dbg_kms(crtc->base.crtc->dev,
+			    "doesn't support plane and CRTC degamma at the same time\n");
+		return -EINVAL;
+	}
+
+	/* If we are here, it means we don't have plane degamma settings, check
+	 * if we have CRTC degamma waiting for mapping to pre-blending degamma
+	 * block
+	 */
+	if (has_crtc_cm_degamma) {
+		/*
+		 * AMD HW doesn't have post-blending degamma caps. When DRM
+		 * CRTC atomic degamma is set, we maps it to DPP degamma block
+		 * (pre-blending) or, on legacy gamma, we use DPP degamma to
+		 * linearize (implicit degamma) from sRGB/BT709 according to
+		 * the input space.
+		 */
+		ret = map_crtc_degamma_to_dc_plane(crtc, dc_plane_state, color_caps);
+		if (ret)
+			return ret;
+	}
+
+	/* Setup CRTC CTM. */
+	if (dm_plane_state->ctm) {
+		ctm = (struct drm_color_ctm_3x4 *)dm_plane_state->ctm->data;
+		/*
+		 * DCN2 and older don't support both pre-blending and
+		 * post-blending gamut remap. For this HW family, if we have
+		 * the plane and CRTC CTMs simultaneously, CRTC CTM takes
+		 * priority, and we discard plane CTM, as implemented in
+		 * dcn10_program_gamut_remap(). However, DCN3+ has DPP
+		 * (pre-blending) and MPC (post-blending) `gamut remap` blocks;
+		 * therefore, we can program plane and CRTC CTMs together by
+		 * mapping CRTC CTM to MPC and keeping plane CTM setup at DPP,
+		 * as it's done by dcn30_program_gamut_remap().
+		 */
+		__drm_ctm_3x4_to_dc_matrix(ctm, dc_plane_state->gamut_remap_matrix.matrix);
+
+		dc_plane_state->gamut_remap_matrix.enable_remap = true;
+		dc_plane_state->input_csc_color_matrix.enable_adjustment = false;
+	} else {
+		/* Bypass CTM. */
+		dc_plane_state->gamut_remap_matrix.enable_remap = false;
+		dc_plane_state->input_csc_color_matrix.enable_adjustment = false;
+	}
+
+	if (!amdgpu_dm_plane_set_colorop_properties(plane_state, dc_plane_state))
+		return 0;
+
+	return amdgpu_dm_plane_set_color_properties(plane_state, dc_plane_state);
+}
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_colorop.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_colorop.c
new file mode 100644
index 000000000000..d585618b8064
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_colorop.c
@@ -0,0 +1,209 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#include <drm/drm_print.h>
+#include <drm/drm_plane.h>
+#include <drm/drm_property.h>
+#include <drm/drm_colorop.h>
+
+#include "amdgpu.h"
+#include "amdgpu_dm_colorop.h"
+#include "dc.h"
+
+const u64 amdgpu_dm_supported_degam_tfs =
+	BIT(DRM_COLOROP_1D_CURVE_SRGB_EOTF) |
+	BIT(DRM_COLOROP_1D_CURVE_PQ_125_EOTF) |
+	BIT(DRM_COLOROP_1D_CURVE_BT2020_INV_OETF) |
+	BIT(DRM_COLOROP_1D_CURVE_GAMMA22_INV);
+
+const u64 amdgpu_dm_supported_shaper_tfs =
+	BIT(DRM_COLOROP_1D_CURVE_SRGB_INV_EOTF) |
+	BIT(DRM_COLOROP_1D_CURVE_PQ_125_INV_EOTF) |
+	BIT(DRM_COLOROP_1D_CURVE_BT2020_OETF) |
+	BIT(DRM_COLOROP_1D_CURVE_GAMMA22);
+
+const u64 amdgpu_dm_supported_blnd_tfs =
+	BIT(DRM_COLOROP_1D_CURVE_SRGB_EOTF) |
+	BIT(DRM_COLOROP_1D_CURVE_PQ_125_EOTF) |
+	BIT(DRM_COLOROP_1D_CURVE_BT2020_INV_OETF) |
+	BIT(DRM_COLOROP_1D_CURVE_GAMMA22_INV);
+
+#define MAX_COLOR_PIPELINE_OPS 10
+
+#define LUT3D_SIZE		17
+
+int amdgpu_dm_initialize_default_pipeline(struct drm_plane *plane, struct drm_prop_enum_list *list)
+{
+	struct drm_colorop *ops[MAX_COLOR_PIPELINE_OPS];
+	struct drm_device *dev = plane->dev;
+	struct amdgpu_device *adev = drm_to_adev(dev);
+	int ret;
+	int i = 0;
+
+	memset(ops, 0, sizeof(ops));
+
+	/* 1D curve - DEGAM TF */
+	ops[i] = kzalloc(sizeof(*ops[0]), GFP_KERNEL);
+	if (!ops[i]) {
+		ret = -ENOMEM;
+		goto cleanup;
+	}
+
+	ret = drm_plane_colorop_curve_1d_init(dev, ops[i], plane,
+					      amdgpu_dm_supported_degam_tfs,
+					      DRM_COLOROP_FLAG_ALLOW_BYPASS);
+	if (ret)
+		goto cleanup;
+
+	list->type = ops[i]->base.id;
+	list->name = kasprintf(GFP_KERNEL, "Color Pipeline %d", ops[i]->base.id);
+
+	i++;
+
+	/* Multiplier */
+	ops[i] = kzalloc(sizeof(struct drm_colorop), GFP_KERNEL);
+	if (!ops[i]) {
+		ret = -ENOMEM;
+		goto cleanup;
+	}
+
+	ret = drm_plane_colorop_mult_init(dev, ops[i], plane, DRM_COLOROP_FLAG_ALLOW_BYPASS);
+	if (ret)
+		goto cleanup;
+
+	drm_colorop_set_next_property(ops[i-1], ops[i]);
+
+	i++;
+
+	/* 3x4 matrix */
+	ops[i] = kzalloc(sizeof(struct drm_colorop), GFP_KERNEL);
+	if (!ops[i]) {
+		ret = -ENOMEM;
+		goto cleanup;
+	}
+
+	ret = drm_plane_colorop_ctm_3x4_init(dev, ops[i], plane, DRM_COLOROP_FLAG_ALLOW_BYPASS);
+	if (ret)
+		goto cleanup;
+
+	drm_colorop_set_next_property(ops[i-1], ops[i]);
+
+	i++;
+
+	if (adev->dm.dc->caps.color.dpp.hw_3d_lut) {
+		/* 1D curve - SHAPER TF */
+		ops[i] = kzalloc(sizeof(*ops[0]), GFP_KERNEL);
+		if (!ops[i]) {
+			ret = -ENOMEM;
+			goto cleanup;
+		}
+
+		ret = drm_plane_colorop_curve_1d_init(dev, ops[i], plane,
+						amdgpu_dm_supported_shaper_tfs,
+						DRM_COLOROP_FLAG_ALLOW_BYPASS);
+		if (ret)
+			goto cleanup;
+
+		drm_colorop_set_next_property(ops[i-1], ops[i]);
+
+		i++;
+
+		/* 1D LUT - SHAPER LUT */
+		ops[i] = kzalloc(sizeof(*ops[0]), GFP_KERNEL);
+		if (!ops[i]) {
+			ret = -ENOMEM;
+			goto cleanup;
+		}
+
+		ret = drm_plane_colorop_curve_1d_lut_init(dev, ops[i], plane, MAX_COLOR_LUT_ENTRIES,
+							DRM_COLOROP_LUT1D_INTERPOLATION_LINEAR,
+							DRM_COLOROP_FLAG_ALLOW_BYPASS);
+		if (ret)
+			goto cleanup;
+
+		drm_colorop_set_next_property(ops[i-1], ops[i]);
+
+		i++;
+
+		/* 3D LUT */
+		ops[i] = kzalloc(sizeof(*ops[0]), GFP_KERNEL);
+		if (!ops[i]) {
+			ret = -ENOMEM;
+			goto cleanup;
+		}
+
+		ret = drm_plane_colorop_3dlut_init(dev, ops[i], plane, LUT3D_SIZE,
+					DRM_COLOROP_LUT3D_INTERPOLATION_TETRAHEDRAL,
+					DRM_COLOROP_FLAG_ALLOW_BYPASS);
+		if (ret)
+			goto cleanup;
+
+		drm_colorop_set_next_property(ops[i-1], ops[i]);
+
+		i++;
+	}
+
+	/* 1D curve - BLND TF */
+	ops[i] = kzalloc(sizeof(*ops[0]), GFP_KERNEL);
+	if (!ops[i]) {
+		ret = -ENOMEM;
+		goto cleanup;
+	}
+
+	ret = drm_plane_colorop_curve_1d_init(dev, ops[i], plane,
+					      amdgpu_dm_supported_blnd_tfs,
+					      DRM_COLOROP_FLAG_ALLOW_BYPASS);
+	if (ret)
+		goto cleanup;
+
+	drm_colorop_set_next_property(ops[i - 1], ops[i]);
+
+	i++;
+
+	/* 1D LUT - BLND LUT */
+	ops[i] = kzalloc(sizeof(struct drm_colorop), GFP_KERNEL);
+	if (!ops[i]) {
+		ret = -ENOMEM;
+		goto cleanup;
+	}
+
+	ret = drm_plane_colorop_curve_1d_lut_init(dev, ops[i], plane, MAX_COLOR_LUT_ENTRIES,
+						  DRM_COLOROP_LUT1D_INTERPOLATION_LINEAR,
+						  DRM_COLOROP_FLAG_ALLOW_BYPASS);
+	if (ret)
+		goto cleanup;
+
+	drm_colorop_set_next_property(ops[i-1], ops[i]);
+	return 0;
+
+cleanup:
+	if (ret == -ENOMEM)
+		drm_err(plane->dev, "KMS: Failed to allocate colorop\n");
+
+	drm_colorop_pipeline_destroy(dev);
+
+	return ret;
+}
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_colorop.h b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_colorop.h
new file mode 100644
index 000000000000..2e1617ffc8ee
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_colorop.h
@@ -0,0 +1,36 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __AMDGPU_DM_COLOROP_H__
+#define __AMDGPU_DM_COLOROP_H__
+
+extern const u64 amdgpu_dm_supported_degam_tfs;
+extern const u64 amdgpu_dm_supported_shaper_tfs;
+extern const u64 amdgpu_dm_supported_blnd_tfs;
+
+int amdgpu_dm_initialize_default_pipeline(struct drm_plane *plane, struct drm_prop_enum_list *list);
+
+#endif /* __AMDGPU_DM_COLOROP_H__*/
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.c
index 8a441a22c46e..e20aa7438066 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.c
@@ -1,3 +1,4 @@
+// SPDX-License-Identifier: MIT
 /*
  * Copyright 2015 Advanced Micro Devices, Inc.
  *
@@ -30,6 +31,7 @@
 #include "amdgpu_dm.h"
 #include "dc.h"
 #include "amdgpu_securedisplay.h"
+#include "amdgpu_dm_psr.h"
 
 static const char *const pipe_crc_sources[] = {
 	"none",
@@ -83,56 +85,383 @@ const char *const *amdgpu_dm_crtc_get_crc_sources(struct drm_crtc *crtc,
 }
 
 #ifdef CONFIG_DRM_AMD_SECURE_DISPLAY
-static void amdgpu_dm_set_crc_window_default(struct drm_crtc *crtc)
+static void update_phy_id_mapping(struct amdgpu_device *adev)
+{
+	struct drm_device *ddev = adev_to_drm(adev);
+	struct amdgpu_display_manager *dm = &adev->dm;
+	struct drm_connector *connector;
+	struct amdgpu_dm_connector *aconnector;
+	struct amdgpu_dm_connector *sort_connector[AMDGPU_DM_MAX_CRTC] = {NULL};
+	struct drm_connector_list_iter iter;
+	uint8_t idx = 0, idx_2 = 0, connector_cnt = 0;
+
+	dm->secure_display_ctx.phy_mapping_updated = false;
+
+	mutex_lock(&ddev->mode_config.mutex);
+	drm_connector_list_iter_begin(ddev, &iter);
+	drm_for_each_connector_iter(connector, &iter) {
+
+		if (connector->status != connector_status_connected)
+			continue;
+
+		if (idx >= AMDGPU_DM_MAX_CRTC) {
+			DRM_WARN("%s connected connectors exceed max crtc\n", __func__);
+			mutex_unlock(&ddev->mode_config.mutex);
+			return;
+		}
+
+		aconnector = to_amdgpu_dm_connector(connector);
+
+		sort_connector[idx] = aconnector;
+		idx++;
+		connector_cnt++;
+	}
+	drm_connector_list_iter_end(&iter);
+
+	/* sort connectors by link_enc_hw_instance first */
+	for (idx = connector_cnt; idx > 1 ; idx--) {
+		for (idx_2 = 0; idx_2 < (idx - 1); idx_2++) {
+			if (sort_connector[idx_2]->dc_link->link_enc_hw_inst >
+			    sort_connector[idx_2 + 1]->dc_link->link_enc_hw_inst)
+				swap(sort_connector[idx_2], sort_connector[idx_2 + 1]);
+		}
+	}
+
+	/*
+	 * Sort mst connectors by RAD. mst connectors with the same enc_hw_instance are already
+	 * sorted together above.
+	 */
+	for (idx = 0; idx < connector_cnt; /*Do nothing*/) {
+		if (sort_connector[idx]->mst_root) {
+			uint8_t i, j, k;
+			uint8_t mst_con_cnt = 1;
+
+			for (idx_2 = (idx + 1); idx_2 < connector_cnt; idx_2++) {
+				if (sort_connector[idx_2]->mst_root == sort_connector[idx]->mst_root)
+					mst_con_cnt++;
+				else
+					break;
+			}
+
+			for (i = mst_con_cnt; i > 1; i--) {
+				for (j = idx; j < (idx + i - 2); j++) {
+					int mstb_lct = sort_connector[j]->mst_output_port->parent->lct;
+					int next_mstb_lct = sort_connector[j + 1]->mst_output_port->parent->lct;
+					u8 *rad;
+					u8 *next_rad;
+					bool swap = false;
+
+					/* Sort by mst tree depth first. Then compare RAD if depth is the same*/
+					if (mstb_lct > next_mstb_lct) {
+						swap = true;
+					} else if (mstb_lct == next_mstb_lct) {
+						if (mstb_lct == 1) {
+							if (sort_connector[j]->mst_output_port->port_num > sort_connector[j + 1]->mst_output_port->port_num)
+								swap = true;
+						} else if (mstb_lct > 1) {
+							rad = sort_connector[j]->mst_output_port->parent->rad;
+							next_rad = sort_connector[j + 1]->mst_output_port->parent->rad;
+
+							for (k = 0; k < mstb_lct - 1; k++) {
+								int shift = (k % 2) ? 0 : 4;
+								int port_num = (rad[k / 2] >> shift) & 0xf;
+								int next_port_num = (next_rad[k / 2] >> shift) & 0xf;
+
+								if (port_num > next_port_num) {
+									swap = true;
+									break;
+								}
+							}
+						} else {
+							DRM_ERROR("MST LCT shouldn't be set as < 1");
+							mutex_unlock(&ddev->mode_config.mutex);
+							return;
+						}
+					}
+
+					if (swap)
+						swap(sort_connector[j], sort_connector[j + 1]);
+				}
+			}
+
+			idx += mst_con_cnt;
+		} else {
+			idx++;
+		}
+	}
+
+	/* Complete sorting. Assign relavant result to dm->secure_display_ctx.phy_id_mapping[]*/
+	memset(dm->secure_display_ctx.phy_id_mapping, 0, sizeof(dm->secure_display_ctx.phy_id_mapping));
+	for (idx = 0; idx < connector_cnt; idx++) {
+		aconnector = sort_connector[idx];
+
+		dm->secure_display_ctx.phy_id_mapping[idx].assigned = true;
+		dm->secure_display_ctx.phy_id_mapping[idx].is_mst = false;
+		dm->secure_display_ctx.phy_id_mapping[idx].enc_hw_inst = aconnector->dc_link->link_enc_hw_inst;
+
+		if (sort_connector[idx]->mst_root) {
+			dm->secure_display_ctx.phy_id_mapping[idx].is_mst = true;
+			dm->secure_display_ctx.phy_id_mapping[idx].lct = aconnector->mst_output_port->parent->lct;
+			dm->secure_display_ctx.phy_id_mapping[idx].port_num = aconnector->mst_output_port->port_num;
+			memcpy(dm->secure_display_ctx.phy_id_mapping[idx].rad,
+				aconnector->mst_output_port->parent->rad, sizeof(aconnector->mst_output_port->parent->rad));
+		}
+	}
+	mutex_unlock(&ddev->mode_config.mutex);
+
+	dm->secure_display_ctx.phy_id_mapping_cnt = connector_cnt;
+	dm->secure_display_ctx.phy_mapping_updated = true;
+}
+
+static bool get_phy_id(struct amdgpu_display_manager *dm,
+			struct amdgpu_dm_connector *aconnector, uint8_t *phy_id)
+{
+	int idx, idx_2;
+	bool found = false;
+
+	/*
+	 * Assume secure display start after all connectors are probed. The connection
+	 * config is static as well
+	 */
+	if (!dm->secure_display_ctx.phy_mapping_updated) {
+		DRM_WARN("%s Should update the phy id table before get it's value", __func__);
+		return false;
+	}
+
+	for (idx = 0; idx < dm->secure_display_ctx.phy_id_mapping_cnt; idx++) {
+		if (!dm->secure_display_ctx.phy_id_mapping[idx].assigned) {
+			DRM_ERROR("phy_id_mapping[%d] should be assigned", idx);
+			return false;
+		}
+
+		if (aconnector->dc_link->link_enc_hw_inst ==
+				dm->secure_display_ctx.phy_id_mapping[idx].enc_hw_inst) {
+			if (!dm->secure_display_ctx.phy_id_mapping[idx].is_mst) {
+				found = true;
+				goto out;
+			} else {
+				/* Could caused by wrongly pass mst root connector */
+				if (!aconnector->mst_output_port) {
+					DRM_ERROR("%s Check mst case but connector without a port assigned", __func__);
+					return false;
+				}
+
+				if (aconnector->mst_root &&
+					aconnector->mst_root->mst_mgr.mst_primary == NULL) {
+					DRM_WARN("%s pass in a stale mst connector", __func__);
+				}
+
+				if (aconnector->mst_output_port->parent->lct == dm->secure_display_ctx.phy_id_mapping[idx].lct &&
+					aconnector->mst_output_port->port_num == dm->secure_display_ctx.phy_id_mapping[idx].port_num) {
+					if (aconnector->mst_output_port->parent->lct == 1) {
+						found = true;
+						goto out;
+					} else if (aconnector->mst_output_port->parent->lct > 1) {
+						/* Check RAD */
+						for (idx_2 = 0; idx_2 < aconnector->mst_output_port->parent->lct - 1; idx_2++) {
+							int shift = (idx_2 % 2) ? 0 : 4;
+							int port_num = (aconnector->mst_output_port->parent->rad[idx_2 / 2] >> shift) & 0xf;
+							int port_num2 = (dm->secure_display_ctx.phy_id_mapping[idx].rad[idx_2 / 2] >> shift) & 0xf;
+
+							if (port_num != port_num2)
+								break;
+						}
+
+						if (idx_2 == aconnector->mst_output_port->parent->lct - 1) {
+							found = true;
+							goto out;
+						}
+					} else {
+						DRM_ERROR("lCT should be >= 1");
+						return false;
+					}
+				}
+			}
+		}
+	}
+
+out:
+	if (found) {
+		DRM_DEBUG_DRIVER("Associated secure display PHY ID as %d", idx);
+		*phy_id = idx;
+	} else {
+		DRM_WARN("Can't find associated phy ID");
+		return false;
+	}
+
+	return true;
+}
+
+static void amdgpu_dm_set_crc_window_default(struct drm_crtc *crtc, struct dc_stream_state *stream)
 {
 	struct drm_device *drm_dev = crtc->dev;
+	struct amdgpu_display_manager *dm = &drm_to_adev(drm_dev)->dm;
 	struct amdgpu_crtc *acrtc = to_amdgpu_crtc(crtc);
+	struct amdgpu_dm_connector *aconnector;
+	bool was_activated;
+	uint8_t phy_id;
+	unsigned long flags;
+	int i;
 
-	spin_lock_irq(&drm_dev->event_lock);
-	acrtc->dm_irq_params.crc_window.x_start = 0;
-	acrtc->dm_irq_params.crc_window.y_start = 0;
-	acrtc->dm_irq_params.crc_window.x_end = 0;
-	acrtc->dm_irq_params.crc_window.y_end = 0;
-	acrtc->dm_irq_params.crc_window.activated = false;
-	acrtc->dm_irq_params.crc_window.update_win = false;
-	acrtc->dm_irq_params.crc_window.skip_frame_cnt = 0;
-	spin_unlock_irq(&drm_dev->event_lock);
+	spin_lock_irqsave(&drm_dev->event_lock, flags);
+	was_activated = acrtc->dm_irq_params.crc_window_activated;
+	for (i = 0; i < MAX_CRC_WINDOW_NUM; i++) {
+		acrtc->dm_irq_params.window_param[i].x_start = 0;
+		acrtc->dm_irq_params.window_param[i].y_start = 0;
+		acrtc->dm_irq_params.window_param[i].x_end = 0;
+		acrtc->dm_irq_params.window_param[i].y_end = 0;
+		acrtc->dm_irq_params.window_param[i].enable = false;
+		acrtc->dm_irq_params.window_param[i].update_win = false;
+		acrtc->dm_irq_params.window_param[i].skip_frame_cnt = 0;
+	}
+	acrtc->dm_irq_params.crc_window_activated = false;
+	spin_unlock_irqrestore(&drm_dev->event_lock, flags);
+
+	/* Disable secure_display if it was enabled */
+	if (was_activated && dm->secure_display_ctx.op_mode == LEGACY_MODE) {
+		/* stop ROI update on this crtc */
+		flush_work(&dm->secure_display_ctx.crtc_ctx[crtc->index].notify_ta_work);
+		flush_work(&dm->secure_display_ctx.crtc_ctx[crtc->index].forward_roi_work);
+		aconnector = (struct amdgpu_dm_connector *)stream->dm_stream_context;
+
+		if (aconnector && get_phy_id(dm, aconnector, &phy_id)) {
+			if (dm->secure_display_ctx.support_mul_roi)
+				dc_stream_forward_multiple_crc_window(stream, NULL, phy_id, true);
+			else
+				dc_stream_forward_crc_window(stream, NULL, phy_id, true);
+		} else {
+			DRM_DEBUG_DRIVER("%s Can't find matching phy id", __func__);
+		}
+	}
 }
 
 static void amdgpu_dm_crtc_notify_ta_to_read(struct work_struct *work)
 {
-	struct crc_rd_work *crc_rd_wrk;
-	struct amdgpu_device *adev;
+	struct secure_display_crtc_context *crtc_ctx;
 	struct psp_context *psp;
-	struct securedisplay_cmd *securedisplay_cmd;
+	struct ta_securedisplay_cmd *securedisplay_cmd;
 	struct drm_crtc *crtc;
-	uint8_t phy_id;
+	struct dc_stream_state *stream;
+	struct amdgpu_dm_connector *aconnector;
+	uint8_t phy_inst;
+	struct amdgpu_display_manager *dm;
+	struct crc_data crc_cpy[MAX_CRC_WINDOW_NUM];
+	unsigned long flags;
+	uint8_t roi_idx = 0;
 	int ret;
+	int i;
 
-	crc_rd_wrk = container_of(work, struct crc_rd_work, notify_ta_work);
-	spin_lock_irq(&crc_rd_wrk->crc_rd_work_lock);
-	crtc = crc_rd_wrk->crtc;
+	crtc_ctx = container_of(work, struct secure_display_crtc_context, notify_ta_work);
+	crtc = crtc_ctx->crtc;
 
-	if (!crtc) {
-		spin_unlock_irq(&crc_rd_wrk->crc_rd_work_lock);
+	if (!crtc)
+		return;
+
+	psp = &drm_to_adev(crtc->dev)->psp;
+
+	if (!psp->securedisplay_context.context.initialized) {
+		DRM_DEBUG_DRIVER("Secure Display fails to notify PSP TA\n");
 		return;
 	}
 
-	adev = drm_to_adev(crtc->dev);
-	psp = &adev->psp;
-	phy_id = crc_rd_wrk->phy_inst;
-	spin_unlock_irq(&crc_rd_wrk->crc_rd_work_lock);
-
-	psp_prep_securedisplay_cmd_buf(psp, &securedisplay_cmd,
-						TA_SECUREDISPLAY_COMMAND__SEND_ROI_CRC);
-	securedisplay_cmd->securedisplay_in_message.send_roi_crc.phy_id =
-						phy_id;
-	ret = psp_securedisplay_invoke(psp, TA_SECUREDISPLAY_COMMAND__SEND_ROI_CRC);
+	dm = &drm_to_adev(crtc->dev)->dm;
+	stream = to_amdgpu_crtc(crtc)->dm_irq_params.stream;
+	aconnector = (struct amdgpu_dm_connector *)stream->dm_stream_context;
+	if (!aconnector)
+		return;
+
+	mutex_lock(&crtc->dev->mode_config.mutex);
+	if (!get_phy_id(dm, aconnector, &phy_inst)) {
+		DRM_WARN("%s Can't find mapping phy id!", __func__);
+		mutex_unlock(&crtc->dev->mode_config.mutex);
+		return;
+	}
+	mutex_unlock(&crtc->dev->mode_config.mutex);
+
+	spin_lock_irqsave(&crtc->dev->event_lock, flags);
+	memcpy(crc_cpy, crtc_ctx->crc_info.crc, sizeof(struct crc_data) * MAX_CRC_WINDOW_NUM);
+	spin_unlock_irqrestore(&crtc->dev->event_lock, flags);
+
+	/* need lock for multiple crtcs to use the command buffer */
+	mutex_lock(&psp->securedisplay_context.mutex);
+	/* PSP TA is expected to finish data transmission over I2C within current frame,
+	 * even there are up to 4 crtcs request to send in this frame.
+	 */
+	if (dm->secure_display_ctx.support_mul_roi) {
+		psp_prep_securedisplay_cmd_buf(psp, &securedisplay_cmd,
+							TA_SECUREDISPLAY_COMMAND__SEND_ROI_CRC_V2);
+
+		securedisplay_cmd->securedisplay_in_message.send_roi_crc_v2.phy_id = phy_inst;
+
+		for (i = 0; i < MAX_CRC_WINDOW_NUM; i++) {
+			if (crc_cpy[i].crc_ready)
+				roi_idx |= 1 << i;
+		}
+		securedisplay_cmd->securedisplay_in_message.send_roi_crc_v2.roi_idx = roi_idx;
+
+		ret = psp_securedisplay_invoke(psp, TA_SECUREDISPLAY_COMMAND__SEND_ROI_CRC_V2);
+	} else {
+		psp_prep_securedisplay_cmd_buf(psp, &securedisplay_cmd,
+							TA_SECUREDISPLAY_COMMAND__SEND_ROI_CRC);
+
+		securedisplay_cmd->securedisplay_in_message.send_roi_crc.phy_id = phy_inst;
+
+		ret = psp_securedisplay_invoke(psp, TA_SECUREDISPLAY_COMMAND__SEND_ROI_CRC);
+	}
+
 	if (!ret) {
-		if (securedisplay_cmd->status != TA_SECUREDISPLAY_STATUS__SUCCESS) {
+		if (securedisplay_cmd->status != TA_SECUREDISPLAY_STATUS__SUCCESS)
 			psp_securedisplay_parse_resp_status(psp, securedisplay_cmd->status);
-		}
 	}
+
+	mutex_unlock(&psp->securedisplay_context.mutex);
+}
+
+static void
+amdgpu_dm_forward_crc_window(struct work_struct *work)
+{
+	struct secure_display_crtc_context *crtc_ctx;
+	struct amdgpu_display_manager *dm;
+	struct drm_crtc *crtc;
+	struct dc_stream_state *stream;
+	struct amdgpu_dm_connector *aconnector;
+	struct crc_window roi_cpy[MAX_CRC_WINDOW_NUM];
+	unsigned long flags;
+	uint8_t phy_id;
+
+	crtc_ctx = container_of(work, struct secure_display_crtc_context, forward_roi_work);
+	crtc = crtc_ctx->crtc;
+
+	if (!crtc)
+		return;
+
+	dm = &drm_to_adev(crtc->dev)->dm;
+	stream = to_amdgpu_crtc(crtc)->dm_irq_params.stream;
+	aconnector = (struct amdgpu_dm_connector *)stream->dm_stream_context;
+
+	if (!aconnector)
+		return;
+
+	mutex_lock(&crtc->dev->mode_config.mutex);
+	if (!get_phy_id(dm, aconnector, &phy_id)) {
+		DRM_WARN("%s Can't find mapping phy id!", __func__);
+		mutex_unlock(&crtc->dev->mode_config.mutex);
+		return;
+	}
+	mutex_unlock(&crtc->dev->mode_config.mutex);
+
+	spin_lock_irqsave(&crtc->dev->event_lock, flags);
+	memcpy(roi_cpy, crtc_ctx->roi, sizeof(struct crc_window) * MAX_CRC_WINDOW_NUM);
+	spin_unlock_irqrestore(&crtc->dev->event_lock, flags);
+
+	mutex_lock(&dm->dc_lock);
+	if (dm->secure_display_ctx.support_mul_roi)
+		dc_stream_forward_multiple_crc_window(stream, roi_cpy,
+			phy_id, false);
+	else
+		dc_stream_forward_crc_window(stream, &roi_cpy[0].rect,
+			phy_id, false);
+	mutex_unlock(&dm->dc_lock);
 }
 
 bool amdgpu_dm_crc_window_is_activated(struct drm_crtc *crtc)
@@ -142,7 +471,7 @@ bool amdgpu_dm_crc_window_is_activated(struct drm_crtc *crtc)
 	bool ret = false;
 
 	spin_lock_irq(&drm_dev->event_lock);
-	ret = acrtc->dm_irq_params.crc_window.activated;
+	ret = acrtc->dm_irq_params.crc_window_activated;
 	spin_unlock_irq(&drm_dev->event_lock);
 
 	return ret;
@@ -180,24 +509,14 @@ int amdgpu_dm_crtc_configure_crc_source(struct drm_crtc *crtc,
 
 	mutex_lock(&adev->dm.dc_lock);
 
-	/* Enable CRTC CRC generation if necessary. */
+	/* For PSR1, check that the panel has exited PSR */
+	if (stream_state->link->psr_settings.psr_version < DC_PSR_VERSION_SU_1)
+		amdgpu_dm_psr_wait_disable(stream_state);
+
+	/* Enable or disable CRTC CRC generation */
 	if (dm_is_crc_source_crtc(source) || source == AMDGPU_DM_PIPE_CRC_SOURCE_NONE) {
-#if defined(CONFIG_DRM_AMD_SECURE_DISPLAY)
-		if (!enable) {
-			if (adev->dm.crc_rd_wrk) {
-				flush_work(&adev->dm.crc_rd_wrk->notify_ta_work);
-				spin_lock_irq(&adev->dm.crc_rd_wrk->crc_rd_work_lock);
-				if (adev->dm.crc_rd_wrk->crtc == crtc) {
-					dc_stream_stop_dmcu_crc_win_update(stream_state->ctx->dc,
-									dm_crtc_state->stream);
-					adev->dm.crc_rd_wrk->crtc = NULL;
-				}
-				spin_unlock_irq(&adev->dm.crc_rd_wrk->crc_rd_work_lock);
-			}
-		}
-#endif
 		if (!dc_stream_configure_crc(stream_state->ctx->dc,
-					     stream_state, NULL, enable, enable)) {
+					     stream_state, NULL, enable, enable, 0, true)) {
 			ret = -EINVAL;
 			goto unlock;
 		}
@@ -228,6 +547,10 @@ int amdgpu_dm_crtc_set_crc_source(struct drm_crtc *crtc, const char *src_name)
 	struct drm_crtc_commit *commit;
 	struct dm_crtc_state *crtc_state;
 	struct drm_device *drm_dev = crtc->dev;
+#if defined(CONFIG_DRM_AMD_SECURE_DISPLAY)
+	struct amdgpu_device *adev = drm_to_adev(drm_dev);
+	struct amdgpu_display_manager *dm = &adev->dm;
+#endif
 	struct amdgpu_crtc *acrtc = to_amdgpu_crtc(crtc);
 	struct drm_dp_aux *aux = NULL;
 	bool enable = false;
@@ -296,6 +619,9 @@ int amdgpu_dm_crtc_set_crc_source(struct drm_crtc *crtc, const char *src_name)
 			if (!connector->state || connector->state->crtc != crtc)
 				continue;
 
+			if (connector->connector_type == DRM_MODE_CONNECTOR_WRITEBACK)
+				continue;
+
 			aconn = to_amdgpu_dm_connector(connector);
 			break;
 		}
@@ -307,7 +633,7 @@ int amdgpu_dm_crtc_set_crc_source(struct drm_crtc *crtc, const char *src_name)
 			goto cleanup;
 		}
 
-		aux = (aconn->port) ? &aconn->port->aux : &aconn->dm_dp_aux.aux;
+		aux = (aconn->mst_output_port) ? &aconn->mst_output_port->aux : &aconn->dm_dp_aux.aux;
 
 		if (!aux) {
 			DRM_DEBUG_DRIVER("No dp aux for amd connector\n");
@@ -324,15 +650,6 @@ int amdgpu_dm_crtc_set_crc_source(struct drm_crtc *crtc, const char *src_name)
 
 	}
 
-#if defined(CONFIG_DRM_AMD_SECURE_DISPLAY)
-	amdgpu_dm_set_crc_window_default(crtc);
-#endif
-
-	if (amdgpu_dm_crtc_configure_crc_source(crtc, crtc_state, source)) {
-		ret = -EINVAL;
-		goto cleanup;
-	}
-
 	/*
 	 * Reading the CRC requires the vblank interrupt handler to be
 	 * enabled. Keep a reference until CRC capture stops.
@@ -342,7 +659,19 @@ int amdgpu_dm_crtc_set_crc_source(struct drm_crtc *crtc, const char *src_name)
 		ret = drm_crtc_vblank_get(crtc);
 		if (ret)
 			goto cleanup;
+	}
+
+#if defined(CONFIG_DRM_AMD_SECURE_DISPLAY)
+	/* Reset secure_display when we change crc source from debugfs */
+	amdgpu_dm_set_crc_window_default(crtc, crtc_state->stream);
+#endif
+
+	if (amdgpu_dm_crtc_configure_crc_source(crtc, crtc_state, source)) {
+		ret = -EINVAL;
+		goto cleanup;
+	}
 
+	if (!enabled && enable) {
 		if (dm_is_crc_source_dprx(source)) {
 			if (drm_dp_start_crc(aux, crtc)) {
 				DRM_DEBUG_DRIVER("dp start crc failed\n");
@@ -368,6 +697,13 @@ int amdgpu_dm_crtc_set_crc_source(struct drm_crtc *crtc, const char *src_name)
 	/* Reset crc_skipped on dm state */
 	crtc_state->crc_skip_count = 0;
 
+#if defined(CONFIG_DRM_AMD_SECURE_DISPLAY)
+	/* Initialize phy id mapping table for secure display*/
+	if (dm->secure_display_ctx.op_mode == LEGACY_MODE &&
+		!dm->secure_display_ctx.phy_mapping_updated)
+		update_phy_id_mapping(adev);
+#endif
+
 cleanup:
 	if (commit)
 		drm_crtc_commit_put(commit);
@@ -422,7 +758,7 @@ void amdgpu_dm_crtc_handle_crc_irq(struct drm_crtc *crtc)
 	}
 
 	if (dm_is_crc_source_crtc(cur_crc_src)) {
-		if (!dc_stream_get_crc(stream_state->ctx->dc, stream_state,
+		if (!dc_stream_get_crc(stream_state->ctx->dc, stream_state, 0,
 				       &crcs[0], &crcs[1], &crcs[2]))
 			return;
 
@@ -434,19 +770,21 @@ void amdgpu_dm_crtc_handle_crc_irq(struct drm_crtc *crtc)
 #if defined(CONFIG_DRM_AMD_SECURE_DISPLAY)
 void amdgpu_dm_crtc_handle_crc_window_irq(struct drm_crtc *crtc)
 {
-	struct dc_stream_state *stream_state;
 	struct drm_device *drm_dev = NULL;
 	enum amdgpu_dm_pipe_crc_source cur_crc_src;
 	struct amdgpu_crtc *acrtc = NULL;
 	struct amdgpu_device *adev = NULL;
-	struct crc_rd_work *crc_rd_wrk = NULL;
-	struct crc_params *crc_window = NULL, tmp_window;
-	unsigned long flags1, flags2;
-	struct crtc_position position;
-	uint32_t v_blank;
-	uint32_t v_back_porch;
-	uint32_t crc_window_latch_up_line;
-	struct dc_crtc_timing *timing_out;
+	struct secure_display_crtc_context *crtc_ctx = NULL;
+	bool reset_crc_frame_count[MAX_CRC_WINDOW_NUM] = {false};
+	uint32_t crc_r[MAX_CRC_WINDOW_NUM] = {0};
+	uint32_t crc_g[MAX_CRC_WINDOW_NUM] = {0};
+	uint32_t crc_b[MAX_CRC_WINDOW_NUM] = {0};
+	unsigned long flags1;
+	bool forward_roi_change = false;
+	bool notify_ta = false;
+	bool all_crc_ready = true;
+	struct dc_stream_state *stream_state;
+	int i;
 
 	if (crtc == NULL)
 		return;
@@ -454,97 +792,160 @@ void amdgpu_dm_crtc_handle_crc_window_irq(struct drm_crtc *crtc)
 	acrtc = to_amdgpu_crtc(crtc);
 	adev = drm_to_adev(crtc->dev);
 	drm_dev = crtc->dev;
+	stream_state = to_dm_crtc_state(crtc->state)->stream;
 
 	spin_lock_irqsave(&drm_dev->event_lock, flags1);
-	stream_state = acrtc->dm_irq_params.stream;
 	cur_crc_src = acrtc->dm_irq_params.crc_src;
-	timing_out = &stream_state->timing;
 
 	/* Early return if CRC capture is not enabled. */
-	if (!amdgpu_dm_is_valid_crc_source(cur_crc_src))
-		goto cleanup;
+	if (!amdgpu_dm_is_valid_crc_source(cur_crc_src) ||
+	    !dm_is_crc_source_crtc(cur_crc_src)) {
+		spin_unlock_irqrestore(&drm_dev->event_lock, flags1);
+		return;
+	}
 
-	if (dm_is_crc_source_crtc(cur_crc_src)) {
-		if (acrtc->dm_irq_params.crc_window.activated) {
-			if (acrtc->dm_irq_params.crc_window.update_win) {
-				if (acrtc->dm_irq_params.crc_window.skip_frame_cnt) {
-					acrtc->dm_irq_params.crc_window.skip_frame_cnt -= 1;
-					goto cleanup;
-				}
-				crc_window = &tmp_window;
-
-				tmp_window.windowa_x_start =
-							acrtc->dm_irq_params.crc_window.x_start;
-				tmp_window.windowa_y_start =
-							acrtc->dm_irq_params.crc_window.y_start;
-				tmp_window.windowa_x_end =
-							acrtc->dm_irq_params.crc_window.x_end;
-				tmp_window.windowa_y_end =
-							acrtc->dm_irq_params.crc_window.y_end;
-				tmp_window.windowb_x_start =
-							acrtc->dm_irq_params.crc_window.x_start;
-				tmp_window.windowb_y_start =
-							acrtc->dm_irq_params.crc_window.y_start;
-				tmp_window.windowb_x_end =
-							acrtc->dm_irq_params.crc_window.x_end;
-				tmp_window.windowb_y_end =
-							acrtc->dm_irq_params.crc_window.y_end;
-
-				dc_stream_forward_dmcu_crc_window(stream_state->ctx->dc,
-									stream_state, crc_window);
-
-				acrtc->dm_irq_params.crc_window.update_win = false;
-
-				dc_stream_get_crtc_position(stream_state->ctx->dc, &stream_state, 1,
-					&position.vertical_count,
-					&position.nominal_vcount);
-
-				v_blank = timing_out->v_total - timing_out->v_border_top -
-					timing_out->v_addressable - timing_out->v_border_bottom;
-
-				v_back_porch = v_blank - timing_out->v_front_porch -
-					timing_out->v_sync_width;
-
-				crc_window_latch_up_line = v_back_porch + timing_out->v_sync_width;
-
-				/* take 3 lines margin*/
-				if ((position.vertical_count + 3) >= crc_window_latch_up_line)
-					acrtc->dm_irq_params.crc_window.skip_frame_cnt = 1;
-				else
-					acrtc->dm_irq_params.crc_window.skip_frame_cnt = 0;
-			} else {
-				if (acrtc->dm_irq_params.crc_window.skip_frame_cnt == 0) {
-					if (adev->dm.crc_rd_wrk) {
-						crc_rd_wrk = adev->dm.crc_rd_wrk;
-						spin_lock_irqsave(&crc_rd_wrk->crc_rd_work_lock, flags2);
-						crc_rd_wrk->phy_inst =
-							stream_state->link->link_enc_hw_inst;
-						spin_unlock_irqrestore(&crc_rd_wrk->crc_rd_work_lock, flags2);
-						schedule_work(&crc_rd_wrk->notify_ta_work);
-					}
-				} else {
-					acrtc->dm_irq_params.crc_window.skip_frame_cnt -= 1;
-				}
-			}
+	if (!acrtc->dm_irq_params.crc_window_activated) {
+		spin_unlock_irqrestore(&drm_dev->event_lock, flags1);
+		return;
+	}
+
+	crtc_ctx = &adev->dm.secure_display_ctx.crtc_ctx[acrtc->crtc_id];
+	if (WARN_ON(crtc_ctx->crtc != crtc)) {
+		/* We have set the crtc when creating secure_display_crtc_context,
+		 * don't expect it to be changed here.
+		 */
+		crtc_ctx->crtc = crtc;
+	}
+
+	for (i = 0; i < MAX_CRC_WINDOW_NUM; i++) {
+		struct crc_params crc_window = {
+			.windowa_x_start = acrtc->dm_irq_params.window_param[i].x_start,
+			.windowa_y_start = acrtc->dm_irq_params.window_param[i].y_start,
+			.windowa_x_end = acrtc->dm_irq_params.window_param[i].x_end,
+			.windowa_y_end = acrtc->dm_irq_params.window_param[i].y_end,
+			.windowb_x_start = acrtc->dm_irq_params.window_param[i].x_start,
+			.windowb_y_start = acrtc->dm_irq_params.window_param[i].y_start,
+			.windowb_x_end = acrtc->dm_irq_params.window_param[i].x_end,
+			.windowb_y_end = acrtc->dm_irq_params.window_param[i].y_end,
+		};
+
+		crtc_ctx->roi[i].enable = acrtc->dm_irq_params.window_param[i].enable;
+
+		if (!acrtc->dm_irq_params.window_param[i].enable) {
+			crtc_ctx->crc_info.crc[i].crc_ready = false;
+			continue;
+		}
+
+		if (acrtc->dm_irq_params.window_param[i].skip_frame_cnt) {
+			acrtc->dm_irq_params.window_param[i].skip_frame_cnt -= 1;
+			crtc_ctx->crc_info.crc[i].crc_ready = false;
+			continue;
+		}
+
+		if (acrtc->dm_irq_params.window_param[i].update_win) {
+			crtc_ctx->roi[i].rect.x = crc_window.windowa_x_start;
+			crtc_ctx->roi[i].rect.y = crc_window.windowa_y_start;
+			crtc_ctx->roi[i].rect.width = crc_window.windowa_x_end -
+						crc_window.windowa_x_start;
+			crtc_ctx->roi[i].rect.height = crc_window.windowa_y_end -
+						crc_window.windowa_y_start;
+
+			if (adev->dm.secure_display_ctx.op_mode == LEGACY_MODE)
+				/* forward task to dmub to update ROI */
+				forward_roi_change = true;
+			else if (adev->dm.secure_display_ctx.op_mode == DISPLAY_CRC_MODE)
+				/* update ROI via dm*/
+				dc_stream_configure_crc(stream_state->ctx->dc, stream_state,
+					&crc_window, true, true, i, false);
+
+			reset_crc_frame_count[i] = true;
+
+			acrtc->dm_irq_params.window_param[i].update_win = false;
+
+			/* Statically skip 1 frame, because we may need to wait below things
+			 * before sending ROI to dmub:
+			 * 1. We defer the work by using system workqueue.
+			 * 2. We may need to wait for dc_lock before accessing dmub.
+			 */
+			acrtc->dm_irq_params.window_param[i].skip_frame_cnt = 1;
+			crtc_ctx->crc_info.crc[i].crc_ready = false;
+		} else {
+			if (!dc_stream_get_crc(stream_state->ctx->dc, stream_state, i,
+						&crc_r[i], &crc_g[i], &crc_b[i]))
+				DRM_ERROR("Secure Display: fail to get crc from engine %d\n", i);
+
+			if (adev->dm.secure_display_ctx.op_mode == LEGACY_MODE)
+				/* forward task to psp to read ROI/CRC and output via I2C */
+				notify_ta = true;
+			else if (adev->dm.secure_display_ctx.op_mode == DISPLAY_CRC_MODE)
+				/* Avoid ROI window get changed, keep overwriting. */
+				dc_stream_configure_crc(stream_state->ctx->dc, stream_state,
+						&crc_window, true, true, i, false);
+
+			/* crc ready for psp to read out */
+			crtc_ctx->crc_info.crc[i].crc_ready = true;
 		}
 	}
 
-cleanup:
 	spin_unlock_irqrestore(&drm_dev->event_lock, flags1);
+
+	if (forward_roi_change)
+		schedule_work(&crtc_ctx->forward_roi_work);
+
+	if (notify_ta)
+		schedule_work(&crtc_ctx->notify_ta_work);
+
+	spin_lock_irqsave(&crtc_ctx->crc_info.lock, flags1);
+	for (i = 0; i < MAX_CRC_WINDOW_NUM; i++) {
+		crtc_ctx->crc_info.crc[i].crc_R = crc_r[i];
+		crtc_ctx->crc_info.crc[i].crc_G = crc_g[i];
+		crtc_ctx->crc_info.crc[i].crc_B = crc_b[i];
+
+		if (!crtc_ctx->roi[i].enable) {
+			crtc_ctx->crc_info.crc[i].frame_count = 0;
+			continue;
+		}
+
+		if (!crtc_ctx->crc_info.crc[i].crc_ready)
+			all_crc_ready = false;
+
+		if (reset_crc_frame_count[i] || crtc_ctx->crc_info.crc[i].frame_count == UINT_MAX)
+			/* Reset the reference frame count after user update the ROI
+			 * or it reaches the maximum value.
+			 */
+			crtc_ctx->crc_info.crc[i].frame_count = 0;
+		else
+			crtc_ctx->crc_info.crc[i].frame_count += 1;
+	}
+	spin_unlock_irqrestore(&crtc_ctx->crc_info.lock, flags1);
+
+	if (all_crc_ready)
+		complete_all(&crtc_ctx->crc_info.completion);
 }
 
-struct crc_rd_work *amdgpu_dm_crtc_secure_display_create_work(void)
+void amdgpu_dm_crtc_secure_display_create_contexts(struct amdgpu_device *adev)
 {
-	struct crc_rd_work *crc_rd_wrk = NULL;
+	struct secure_display_crtc_context *crtc_ctx = NULL;
+	int i;
 
-	crc_rd_wrk = kzalloc(sizeof(*crc_rd_wrk), GFP_KERNEL);
+	crtc_ctx = kcalloc(adev->mode_info.num_crtc,
+				      sizeof(struct secure_display_crtc_context),
+				      GFP_KERNEL);
 
-	if (!crc_rd_wrk)
-		return NULL;
+	if (!crtc_ctx) {
+		adev->dm.secure_display_ctx.crtc_ctx = NULL;
+		return;
+	}
+
+	for (i = 0; i < adev->mode_info.num_crtc; i++) {
+		INIT_WORK(&crtc_ctx[i].forward_roi_work, amdgpu_dm_forward_crc_window);
+		INIT_WORK(&crtc_ctx[i].notify_ta_work, amdgpu_dm_crtc_notify_ta_to_read);
+		crtc_ctx[i].crtc = &adev->mode_info.crtcs[i]->base;
+		spin_lock_init(&crtc_ctx[i].crc_info.lock);
+	}
 
-	spin_lock_init(&crc_rd_wrk->crc_rd_work_lock);
-	INIT_WORK(&crc_rd_wrk->notify_ta_work, amdgpu_dm_crtc_notify_ta_to_read);
+	adev->dm.secure_display_ctx.crtc_ctx = crtc_ctx;
 
-	return crc_rd_wrk;
+	adev->dm.secure_display_ctx.op_mode = DISPLAY_CRC_MODE;
 }
 #endif
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.h b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.h
index f07850db60a6..95bdb8699d7f 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.h
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.h
@@ -1,3 +1,4 @@
+/* SPDX-License-Identifier: MIT */
 /*
  * Copyright 2019 Advanced Micro Devices, Inc.
  *
@@ -40,25 +41,76 @@ enum amdgpu_dm_pipe_crc_source {
 };
 
 #ifdef CONFIG_DRM_AMD_SECURE_DISPLAY
-struct crc_window_parm {
+#define MAX_CRTC 6
+
+enum secure_display_mode {
+	/* via dmub + psp */
+	LEGACY_MODE = 0,
+	/* driver directly */
+	DISPLAY_CRC_MODE,
+	SECURE_DISPLAY_MODE_MAX,
+};
+
+struct phy_id_mapping {
+	bool assigned;
+	bool is_mst;
+	uint8_t enc_hw_inst;
+	u8 lct;
+	u8 port_num;
+	u8 rad[8];
+};
+
+struct crc_data {
+	uint32_t crc_R;
+	uint32_t crc_G;
+	uint32_t crc_B;
+	uint32_t frame_count;
+	bool crc_ready;
+};
+
+struct crc_info {
+	struct crc_data crc[MAX_CRC_WINDOW_NUM];
+	struct completion completion;
+	spinlock_t lock;
+};
+
+struct crc_window_param {
 	uint16_t x_start;
 	uint16_t y_start;
 	uint16_t x_end;
 	uint16_t y_end;
-	/* CRC windwo is activated or not*/
-	bool activated;
+	/* CRC window is activated or not*/
+	bool enable;
 	/* Update crc window during vertical blank or not */
 	bool update_win;
 	/* skip reading/writing for few frames */
 	int skip_frame_cnt;
 };
 
-struct crc_rd_work {
+struct secure_display_crtc_context {
+	/* work to notify PSP TA*/
 	struct work_struct notify_ta_work;
-	/* To protect crc_rd_work carried fields*/
-	spinlock_t crc_rd_work_lock;
+
+	/* work to forward ROI to dmcu/dmub */
+	struct work_struct forward_roi_work;
+
 	struct drm_crtc *crtc;
-	uint8_t phy_inst;
+
+	/* Region of Interest (ROI) */
+	struct crc_window roi[MAX_CRC_WINDOW_NUM];
+
+	struct crc_info crc_info;
+};
+
+struct secure_display_context {
+
+	struct secure_display_crtc_context *crtc_ctx;
+    /* Whether dmub support multiple ROI setting */
+	bool support_mul_roi;
+	enum secure_display_mode op_mode;
+	bool phy_mapping_updated;
+	int phy_id_mapping_cnt;
+	struct phy_id_mapping phy_id_mapping[MAX_CRTC];
 };
 #endif
 
@@ -90,11 +142,11 @@ void amdgpu_dm_crtc_handle_crc_irq(struct drm_crtc *crtc);
 #ifdef CONFIG_DRM_AMD_SECURE_DISPLAY
 bool amdgpu_dm_crc_window_is_activated(struct drm_crtc *crtc);
 void amdgpu_dm_crtc_handle_crc_window_irq(struct drm_crtc *crtc);
-struct crc_rd_work *amdgpu_dm_crtc_secure_display_create_work(void);
+void amdgpu_dm_crtc_secure_display_create_contexts(struct amdgpu_device *adev);
 #else
 #define amdgpu_dm_crc_window_is_activated(x)
 #define amdgpu_dm_crtc_handle_crc_window_irq(x)
-#define amdgpu_dm_crtc_secure_display_create_work()
+#define amdgpu_dm_crtc_secure_display_create_contexts(x)
 #endif
 
 #endif /* AMD_DAL_DEV_AMDGPU_DM_AMDGPU_DM_CRC_H_ */
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crtc.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crtc.c
new file mode 100644
index 000000000000..697e232acebf
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crtc.c
@@ -0,0 +1,804 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+#include <drm/drm_vblank.h>
+#include <drm/drm_atomic_helper.h>
+
+#include "dc.h"
+#include "amdgpu.h"
+#include "amdgpu_dm_psr.h"
+#include "amdgpu_dm_replay.h"
+#include "amdgpu_dm_crtc.h"
+#include "amdgpu_dm_plane.h"
+#include "amdgpu_dm_trace.h"
+#include "amdgpu_dm_debugfs.h"
+
+#define HPD_DETECTION_PERIOD_uS 2000000
+#define HPD_DETECTION_TIME_uS 100000
+
+void amdgpu_dm_crtc_handle_vblank(struct amdgpu_crtc *acrtc)
+{
+	struct drm_crtc *crtc = &acrtc->base;
+	struct drm_device *dev = crtc->dev;
+	unsigned long flags;
+
+	drm_crtc_handle_vblank(crtc);
+
+	spin_lock_irqsave(&dev->event_lock, flags);
+
+	/* Send completion event for cursor-only commits */
+	if (acrtc->event && acrtc->pflip_status != AMDGPU_FLIP_SUBMITTED) {
+		drm_crtc_send_vblank_event(crtc, acrtc->event);
+		drm_crtc_vblank_put(crtc);
+		acrtc->event = NULL;
+	}
+
+	spin_unlock_irqrestore(&dev->event_lock, flags);
+}
+
+bool amdgpu_dm_crtc_modeset_required(struct drm_crtc_state *crtc_state,
+			     struct dc_stream_state *new_stream,
+			     struct dc_stream_state *old_stream)
+{
+	return crtc_state->active && drm_atomic_crtc_needs_modeset(crtc_state);
+}
+
+bool amdgpu_dm_crtc_vrr_active_irq(struct amdgpu_crtc *acrtc)
+
+{
+	return acrtc->dm_irq_params.freesync_config.state ==
+		       VRR_STATE_ACTIVE_VARIABLE ||
+	       acrtc->dm_irq_params.freesync_config.state ==
+		       VRR_STATE_ACTIVE_FIXED;
+}
+
+int amdgpu_dm_crtc_set_vupdate_irq(struct drm_crtc *crtc, bool enable)
+{
+	enum dc_irq_source irq_source;
+	struct amdgpu_crtc *acrtc = to_amdgpu_crtc(crtc);
+	struct amdgpu_device *adev = drm_to_adev(crtc->dev);
+	int rc;
+
+	if (acrtc->otg_inst == -1)
+		return 0;
+
+	irq_source = IRQ_TYPE_VUPDATE + acrtc->otg_inst;
+
+	rc = dc_interrupt_set(adev->dm.dc, irq_source, enable) ? 0 : -EBUSY;
+
+	DRM_DEBUG_VBL("crtc %d - vupdate irq %sabling: r=%d\n",
+		      acrtc->crtc_id, enable ? "en" : "dis", rc);
+	return rc;
+}
+
+bool amdgpu_dm_crtc_vrr_active(const struct dm_crtc_state *dm_state)
+{
+	return dm_state->freesync_config.state == VRR_STATE_ACTIVE_VARIABLE ||
+	       dm_state->freesync_config.state == VRR_STATE_ACTIVE_FIXED;
+}
+
+/**
+ * amdgpu_dm_crtc_set_panel_sr_feature() - Manage panel self-refresh features.
+ *
+ * @vblank_work:    is a pointer to a struct vblank_control_work object.
+ * @vblank_enabled: indicates whether the DRM vblank counter is currently
+ *                  enabled (true) or disabled (false).
+ * @allow_sr_entry: represents whether entry into the self-refresh mode is
+ *                  allowed (true) or not allowed (false).
+ *
+ * The DRM vblank counter enable/disable action is used as the trigger to enable
+ * or disable various panel self-refresh features:
+ *
+ * Panel Replay and PSR SU
+ * - Enable when:
+ *      - VRR is disabled
+ *      - vblank counter is disabled
+ *      - entry is allowed: usermode demonstrates an adequate number of fast
+ *        commits)
+ *     - CRC capture window isn't active
+ * - Keep enabled even when vblank counter gets enabled
+ *
+ * PSR1
+ * - Enable condition same as above
+ * - Disable when vblank counter is enabled
+ */
+static void amdgpu_dm_crtc_set_panel_sr_feature(
+	struct vblank_control_work *vblank_work,
+	bool vblank_enabled, bool allow_sr_entry)
+{
+	struct dc_link *link = vblank_work->stream->link;
+	bool is_sr_active = (link->replay_settings.replay_allow_active ||
+				 link->psr_settings.psr_allow_active);
+	bool is_crc_window_active = false;
+	bool vrr_active = amdgpu_dm_crtc_vrr_active_irq(vblank_work->acrtc);
+
+#ifdef CONFIG_DRM_AMD_SECURE_DISPLAY
+	is_crc_window_active =
+		amdgpu_dm_crc_window_is_activated(&vblank_work->acrtc->base);
+#endif
+
+	if (link->replay_settings.replay_feature_enabled && !vrr_active &&
+		allow_sr_entry && !is_sr_active && !is_crc_window_active) {
+		amdgpu_dm_replay_enable(vblank_work->stream, true);
+	} else if (vblank_enabled) {
+		if (link->psr_settings.psr_version < DC_PSR_VERSION_SU_1 && is_sr_active)
+			amdgpu_dm_psr_disable(vblank_work->stream, false);
+	} else if (link->psr_settings.psr_feature_enabled && !vrr_active &&
+		allow_sr_entry && !is_sr_active && !is_crc_window_active) {
+
+		struct amdgpu_dm_connector *aconn =
+			(struct amdgpu_dm_connector *) vblank_work->stream->dm_stream_context;
+
+		if (!aconn->disallow_edp_enter_psr) {
+			struct amdgpu_display_manager *dm = vblank_work->dm;
+
+			amdgpu_dm_psr_enable(vblank_work->stream);
+			if (dm->idle_workqueue &&
+			    (dm->dc->config.disable_ips == DMUB_IPS_ENABLE) &&
+			    dm->dc->idle_optimizations_allowed &&
+			    dm->idle_workqueue->enable &&
+			    !dm->idle_workqueue->running)
+				schedule_work(&dm->idle_workqueue->work);
+		}
+	}
+}
+
+bool amdgpu_dm_is_headless(struct amdgpu_device *adev)
+{
+	struct drm_connector *connector;
+	struct drm_connector_list_iter iter;
+	struct drm_device *dev;
+	bool is_headless = true;
+
+	if (adev == NULL)
+		return true;
+
+	dev = adev->dm.ddev;
+
+	drm_connector_list_iter_begin(dev, &iter);
+	drm_for_each_connector_iter(connector, &iter) {
+
+		if (connector->connector_type == DRM_MODE_CONNECTOR_WRITEBACK)
+			continue;
+
+		if (connector->status == connector_status_connected) {
+			is_headless = false;
+			break;
+		}
+	}
+	drm_connector_list_iter_end(&iter);
+	return is_headless;
+}
+
+static void amdgpu_dm_idle_worker(struct work_struct *work)
+{
+	struct idle_workqueue *idle_work;
+
+	idle_work = container_of(work, struct idle_workqueue, work);
+	idle_work->dm->idle_workqueue->running = true;
+
+	while (idle_work->enable) {
+		fsleep(HPD_DETECTION_PERIOD_uS);
+		mutex_lock(&idle_work->dm->dc_lock);
+		if (!idle_work->dm->dc->idle_optimizations_allowed) {
+			mutex_unlock(&idle_work->dm->dc_lock);
+			break;
+		}
+		dc_allow_idle_optimizations(idle_work->dm->dc, false);
+
+		mutex_unlock(&idle_work->dm->dc_lock);
+		fsleep(HPD_DETECTION_TIME_uS);
+		mutex_lock(&idle_work->dm->dc_lock);
+
+		if (!amdgpu_dm_is_headless(idle_work->dm->adev) &&
+		    !amdgpu_dm_psr_is_active_allowed(idle_work->dm)) {
+			mutex_unlock(&idle_work->dm->dc_lock);
+			break;
+		}
+
+		if (idle_work->enable) {
+			dc_post_update_surfaces_to_stream(idle_work->dm->dc);
+			dc_allow_idle_optimizations(idle_work->dm->dc, true);
+		}
+		mutex_unlock(&idle_work->dm->dc_lock);
+	}
+	idle_work->dm->idle_workqueue->running = false;
+}
+
+struct idle_workqueue *idle_create_workqueue(struct amdgpu_device *adev)
+{
+	struct idle_workqueue *idle_work;
+
+	idle_work = kzalloc(sizeof(*idle_work), GFP_KERNEL);
+	if (ZERO_OR_NULL_PTR(idle_work))
+		return NULL;
+
+	idle_work->dm = &adev->dm;
+	idle_work->enable = false;
+	idle_work->running = false;
+	INIT_WORK(&idle_work->work, amdgpu_dm_idle_worker);
+
+	return idle_work;
+}
+
+static void amdgpu_dm_crtc_vblank_control_worker(struct work_struct *work)
+{
+	struct vblank_control_work *vblank_work =
+		container_of(work, struct vblank_control_work, work);
+	struct amdgpu_display_manager *dm = vblank_work->dm;
+	struct amdgpu_device *adev = drm_to_adev(dm->ddev);
+	int r;
+
+	mutex_lock(&dm->dc_lock);
+
+	if (vblank_work->enable)
+		dm->active_vblank_irq_count++;
+	else if (dm->active_vblank_irq_count)
+		dm->active_vblank_irq_count--;
+
+	if (dm->active_vblank_irq_count > 0)
+		dc_allow_idle_optimizations(dm->dc, false);
+
+	/*
+	 * Control PSR based on vblank requirements from OS
+	 *
+	 * If panel supports PSR SU, there's no need to disable PSR when OS is
+	 * submitting fast atomic commits (we infer this by whether the OS
+	 * requests vblank events). Fast atomic commits will simply trigger a
+	 * full-frame-update (FFU); a specific case of selective-update (SU)
+	 * where the SU region is the full hactive*vactive region. See
+	 * fill_dc_dirty_rects().
+	 */
+	if (vblank_work->stream && vblank_work->stream->link && vblank_work->acrtc) {
+		amdgpu_dm_crtc_set_panel_sr_feature(
+			vblank_work, vblank_work->enable,
+			vblank_work->acrtc->dm_irq_params.allow_sr_entry);
+	}
+
+	if (dm->active_vblank_irq_count == 0) {
+		dc_post_update_surfaces_to_stream(dm->dc);
+
+		r = amdgpu_dpm_pause_power_profile(adev, true);
+		if (r)
+			dev_warn(adev->dev, "failed to set default power profile mode\n");
+
+		dc_allow_idle_optimizations(dm->dc, true);
+
+		r = amdgpu_dpm_pause_power_profile(adev, false);
+		if (r)
+			dev_warn(adev->dev, "failed to restore the power profile mode\n");
+	}
+
+	mutex_unlock(&dm->dc_lock);
+
+	dc_stream_release(vblank_work->stream);
+
+	kfree(vblank_work);
+}
+
+static inline int amdgpu_dm_crtc_set_vblank(struct drm_crtc *crtc, bool enable)
+{
+	struct amdgpu_crtc *acrtc = to_amdgpu_crtc(crtc);
+	struct amdgpu_device *adev = drm_to_adev(crtc->dev);
+	struct dm_crtc_state *acrtc_state = to_dm_crtc_state(crtc->state);
+	struct amdgpu_display_manager *dm = &adev->dm;
+	struct vblank_control_work *work;
+	int irq_type;
+	int rc = 0;
+
+	if (enable && !acrtc->base.enabled) {
+		drm_dbg_vbl(crtc->dev,
+				"Reject vblank enable on unconfigured CRTC %d (enabled=%d)\n",
+				acrtc->crtc_id, acrtc->base.enabled);
+		return -EINVAL;
+	}
+
+	irq_type = amdgpu_display_crtc_idx_to_irq_type(adev, acrtc->crtc_id);
+
+	if (enable) {
+		struct dc *dc = adev->dm.dc;
+		struct drm_vblank_crtc *vblank = drm_crtc_vblank_crtc(crtc);
+		struct psr_settings *psr = &acrtc_state->stream->link->psr_settings;
+		struct replay_settings *pr = &acrtc_state->stream->link->replay_settings;
+		bool sr_supported = (psr->psr_version != DC_PSR_VERSION_UNSUPPORTED) ||
+								pr->config.replay_supported;
+
+		/*
+		 * IPS & self-refresh feature can cause vblank counter resets between
+		 * vblank disable and enable.
+		 * It may cause system stuck due to waiting for the vblank counter.
+		 * Call this function to estimate missed vblanks by using timestamps and
+		 * update the vblank counter in DRM.
+		 */
+		if (dc->caps.ips_support &&
+			dc->config.disable_ips != DMUB_IPS_DISABLE_ALL &&
+			sr_supported && vblank->config.disable_immediate)
+			drm_crtc_vblank_restore(crtc);
+	}
+
+	if (dc_supports_vrr(dm->dc->ctx->dce_version)) {
+		if (enable) {
+			/* vblank irq on -> Only need vupdate irq in vrr mode */
+			if (amdgpu_dm_crtc_vrr_active(acrtc_state))
+				rc = amdgpu_dm_crtc_set_vupdate_irq(crtc, true);
+		} else {
+			/* vblank irq off -> vupdate irq off */
+			rc = amdgpu_dm_crtc_set_vupdate_irq(crtc, false);
+		}
+	}
+
+	if (rc)
+		return rc;
+
+	/* crtc vblank or vstartup interrupt */
+	if (enable) {
+		rc = amdgpu_irq_get(adev, &adev->crtc_irq, irq_type);
+		drm_dbg_vbl(crtc->dev, "Get crtc_irq ret=%d\n", rc);
+	} else {
+		rc = amdgpu_irq_put(adev, &adev->crtc_irq, irq_type);
+		drm_dbg_vbl(crtc->dev, "Put crtc_irq ret=%d\n", rc);
+	}
+
+	if (rc)
+		return rc;
+
+	/*
+	 * hubp surface flip interrupt
+	 *
+	 * We have no guarantee that the frontend index maps to the same
+	 * backend index - some even map to more than one.
+	 *
+	 * TODO: Use a different interrupt or check DC itself for the mapping.
+	 */
+	if (enable) {
+		rc = amdgpu_irq_get(adev, &adev->pageflip_irq, irq_type);
+		drm_dbg_vbl(crtc->dev, "Get pageflip_irq ret=%d\n", rc);
+	} else {
+		rc = amdgpu_irq_put(adev, &adev->pageflip_irq, irq_type);
+		drm_dbg_vbl(crtc->dev, "Put pageflip_irq ret=%d\n", rc);
+	}
+
+	if (rc)
+		return rc;
+
+#if defined(CONFIG_DRM_AMD_SECURE_DISPLAY)
+	/* crtc vline0 interrupt, only available on DCN+ */
+	if (amdgpu_ip_version(adev, DCE_HWIP, 0) != 0) {
+		if (enable) {
+			rc = amdgpu_irq_get(adev, &adev->vline0_irq, irq_type);
+			drm_dbg_vbl(crtc->dev, "Get vline0_irq ret=%d\n", rc);
+		} else {
+			rc = amdgpu_irq_put(adev, &adev->vline0_irq, irq_type);
+			drm_dbg_vbl(crtc->dev, "Put vline0_irq ret=%d\n", rc);
+		}
+
+		if (rc)
+			return rc;
+	}
+#endif
+
+	if (amdgpu_in_reset(adev))
+		return 0;
+
+	if (dm->vblank_control_workqueue) {
+		work = kzalloc(sizeof(*work), GFP_ATOMIC);
+		if (!work)
+			return -ENOMEM;
+
+		INIT_WORK(&work->work, amdgpu_dm_crtc_vblank_control_worker);
+		work->dm = dm;
+		work->acrtc = acrtc;
+		work->enable = enable;
+
+		if (acrtc_state->stream) {
+			dc_stream_retain(acrtc_state->stream);
+			work->stream = acrtc_state->stream;
+		}
+
+		queue_work(dm->vblank_control_workqueue, &work->work);
+	}
+
+	return 0;
+}
+
+int amdgpu_dm_crtc_enable_vblank(struct drm_crtc *crtc)
+{
+	return amdgpu_dm_crtc_set_vblank(crtc, true);
+}
+
+void amdgpu_dm_crtc_disable_vblank(struct drm_crtc *crtc)
+{
+	amdgpu_dm_crtc_set_vblank(crtc, false);
+}
+
+static void amdgpu_dm_crtc_destroy_state(struct drm_crtc *crtc,
+				  struct drm_crtc_state *state)
+{
+	struct dm_crtc_state *cur = to_dm_crtc_state(state);
+
+	/* TODO Destroy dc_stream objects are stream object is flattened */
+	if (cur->stream)
+		dc_stream_release(cur->stream);
+
+
+	__drm_atomic_helper_crtc_destroy_state(state);
+
+
+	kfree(state);
+}
+
+static struct drm_crtc_state *amdgpu_dm_crtc_duplicate_state(struct drm_crtc *crtc)
+{
+	struct dm_crtc_state *state, *cur;
+
+	cur = to_dm_crtc_state(crtc->state);
+
+	if (WARN_ON(!crtc->state))
+		return NULL;
+
+	state = kzalloc(sizeof(*state), GFP_KERNEL);
+	if (!state)
+		return NULL;
+
+	__drm_atomic_helper_crtc_duplicate_state(crtc, &state->base);
+
+	if (cur->stream) {
+		state->stream = cur->stream;
+		dc_stream_retain(state->stream);
+	}
+
+	state->active_planes = cur->active_planes;
+	state->vrr_infopacket = cur->vrr_infopacket;
+	state->abm_level = cur->abm_level;
+	state->vrr_supported = cur->vrr_supported;
+	state->freesync_config = cur->freesync_config;
+	state->cm_has_degamma = cur->cm_has_degamma;
+	state->cm_is_degamma_srgb = cur->cm_is_degamma_srgb;
+	state->regamma_tf = cur->regamma_tf;
+	state->crc_skip_count = cur->crc_skip_count;
+	state->mpo_requested = cur->mpo_requested;
+	state->cursor_mode = cur->cursor_mode;
+	/* TODO Duplicate dc_stream after objects are stream object is flattened */
+
+	return &state->base;
+}
+
+static void amdgpu_dm_crtc_destroy(struct drm_crtc *crtc)
+{
+	drm_crtc_cleanup(crtc);
+	kfree(crtc);
+}
+
+static void amdgpu_dm_crtc_reset_state(struct drm_crtc *crtc)
+{
+	struct dm_crtc_state *state;
+
+	if (crtc->state)
+		amdgpu_dm_crtc_destroy_state(crtc, crtc->state);
+
+	state = kzalloc(sizeof(*state), GFP_KERNEL);
+	if (WARN_ON(!state))
+		return;
+
+	__drm_atomic_helper_crtc_reset(crtc, &state->base);
+}
+
+#ifdef CONFIG_DEBUG_FS
+static int amdgpu_dm_crtc_late_register(struct drm_crtc *crtc)
+{
+	crtc_debugfs_init(crtc);
+
+	return 0;
+}
+#endif
+
+#ifdef AMD_PRIVATE_COLOR
+/**
+ * dm_crtc_additional_color_mgmt - enable additional color properties
+ * @crtc: DRM CRTC
+ *
+ * This function lets the driver enable post-blending CRTC regamma transfer
+ * function property in addition to DRM CRTC gamma LUT. Default value means
+ * linear transfer function, which is the default CRTC gamma LUT behaviour
+ * without this property.
+ */
+static void
+dm_crtc_additional_color_mgmt(struct drm_crtc *crtc)
+{
+	struct amdgpu_device *adev = drm_to_adev(crtc->dev);
+
+	if (adev->dm.dc->caps.color.mpc.ogam_ram)
+		drm_object_attach_property(&crtc->base,
+					   adev->mode_info.regamma_tf_property,
+					   AMDGPU_TRANSFER_FUNCTION_DEFAULT);
+}
+
+static int
+amdgpu_dm_atomic_crtc_set_property(struct drm_crtc *crtc,
+				   struct drm_crtc_state *state,
+				   struct drm_property *property,
+				   uint64_t val)
+{
+	struct amdgpu_device *adev = drm_to_adev(crtc->dev);
+	struct dm_crtc_state *acrtc_state = to_dm_crtc_state(state);
+
+	if (property == adev->mode_info.regamma_tf_property) {
+		if (acrtc_state->regamma_tf != val) {
+			acrtc_state->regamma_tf = val;
+			acrtc_state->base.color_mgmt_changed |= 1;
+		}
+	} else {
+		drm_dbg_atomic(crtc->dev,
+			       "[CRTC:%d:%s] unknown property [PROP:%d:%s]]\n",
+			       crtc->base.id, crtc->name,
+			       property->base.id, property->name);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int
+amdgpu_dm_atomic_crtc_get_property(struct drm_crtc *crtc,
+				   const struct drm_crtc_state *state,
+				   struct drm_property *property,
+				   uint64_t *val)
+{
+	struct amdgpu_device *adev = drm_to_adev(crtc->dev);
+	struct dm_crtc_state *acrtc_state = to_dm_crtc_state(state);
+
+	if (property == adev->mode_info.regamma_tf_property)
+		*val = acrtc_state->regamma_tf;
+	else
+		return -EINVAL;
+
+	return 0;
+}
+#endif
+
+/* Implemented only the options currently available for the driver */
+static const struct drm_crtc_funcs amdgpu_dm_crtc_funcs = {
+	.reset = amdgpu_dm_crtc_reset_state,
+	.destroy = amdgpu_dm_crtc_destroy,
+	.set_config = drm_atomic_helper_set_config,
+	.page_flip = drm_atomic_helper_page_flip,
+	.atomic_duplicate_state = amdgpu_dm_crtc_duplicate_state,
+	.atomic_destroy_state = amdgpu_dm_crtc_destroy_state,
+	.set_crc_source = amdgpu_dm_crtc_set_crc_source,
+	.verify_crc_source = amdgpu_dm_crtc_verify_crc_source,
+	.get_crc_sources = amdgpu_dm_crtc_get_crc_sources,
+	.get_vblank_counter = amdgpu_get_vblank_counter_kms,
+	.enable_vblank = amdgpu_dm_crtc_enable_vblank,
+	.disable_vblank = amdgpu_dm_crtc_disable_vblank,
+	.get_vblank_timestamp = drm_crtc_vblank_helper_get_vblank_timestamp,
+#if defined(CONFIG_DEBUG_FS)
+	.late_register = amdgpu_dm_crtc_late_register,
+#endif
+#ifdef AMD_PRIVATE_COLOR
+	.atomic_set_property = amdgpu_dm_atomic_crtc_set_property,
+	.atomic_get_property = amdgpu_dm_atomic_crtc_get_property,
+#endif
+};
+
+static void amdgpu_dm_crtc_helper_disable(struct drm_crtc *crtc)
+{
+}
+
+static int amdgpu_dm_crtc_count_crtc_active_planes(struct drm_crtc_state *new_crtc_state)
+{
+	struct drm_atomic_state *state = new_crtc_state->state;
+	struct drm_plane *plane;
+	int num_active = 0;
+
+	drm_for_each_plane_mask(plane, state->dev, new_crtc_state->plane_mask) {
+		struct drm_plane_state *new_plane_state;
+
+		/* Cursor planes are "fake". */
+		if (plane->type == DRM_PLANE_TYPE_CURSOR)
+			continue;
+
+		new_plane_state = drm_atomic_get_new_plane_state(state, plane);
+
+		if (!new_plane_state) {
+			/*
+			 * The plane is enable on the CRTC and hasn't changed
+			 * state. This means that it previously passed
+			 * validation and is therefore enabled.
+			 */
+			num_active += 1;
+			continue;
+		}
+
+		/* We need a framebuffer to be considered enabled. */
+		num_active += (new_plane_state->fb != NULL);
+	}
+
+	return num_active;
+}
+
+static void amdgpu_dm_crtc_update_crtc_active_planes(struct drm_crtc *crtc,
+						     struct drm_crtc_state *new_crtc_state)
+{
+	struct dm_crtc_state *dm_new_crtc_state =
+		to_dm_crtc_state(new_crtc_state);
+
+	dm_new_crtc_state->active_planes = 0;
+
+	if (!dm_new_crtc_state->stream)
+		return;
+
+	dm_new_crtc_state->active_planes =
+		amdgpu_dm_crtc_count_crtc_active_planes(new_crtc_state);
+}
+
+static bool amdgpu_dm_crtc_helper_mode_fixup(struct drm_crtc *crtc,
+				      const struct drm_display_mode *mode,
+				      struct drm_display_mode *adjusted_mode)
+{
+	return true;
+}
+
+static int amdgpu_dm_crtc_helper_atomic_check(struct drm_crtc *crtc,
+					      struct drm_atomic_state *state)
+{
+	struct drm_crtc_state *crtc_state = drm_atomic_get_new_crtc_state(state,
+										crtc);
+	struct amdgpu_device *adev = drm_to_adev(crtc->dev);
+	struct dc *dc = adev->dm.dc;
+	struct dm_crtc_state *dm_crtc_state = to_dm_crtc_state(crtc_state);
+	int ret = -EINVAL;
+
+	trace_amdgpu_dm_crtc_atomic_check(crtc_state);
+
+	amdgpu_dm_crtc_update_crtc_active_planes(crtc, crtc_state);
+
+	if (WARN_ON(unlikely(!dm_crtc_state->stream &&
+			amdgpu_dm_crtc_modeset_required(crtc_state, NULL, dm_crtc_state->stream)))) {
+		return ret;
+	}
+
+	/*
+	 * We require the primary plane to be enabled whenever the CRTC is, otherwise
+	 * drm_mode_cursor_universal may end up trying to enable the cursor plane while all other
+	 * planes are disabled, which is not supported by the hardware. And there is legacy
+	 * userspace which stops using the HW cursor altogether in response to the resulting EINVAL.
+	 */
+	if (crtc_state->enable &&
+		!(crtc_state->plane_mask & drm_plane_mask(crtc->primary))) {
+		DRM_DEBUG_ATOMIC("Can't enable a CRTC without enabling the primary plane\n");
+		return -EINVAL;
+	}
+
+	/*
+	 * Only allow async flips for fast updates that don't change the FB
+	 * pitch, the DCC state, rotation, etc.
+	 */
+	if (crtc_state->async_flip &&
+	    dm_crtc_state->update_type != UPDATE_TYPE_FAST) {
+		drm_dbg_atomic(crtc->dev,
+			       "[CRTC:%d:%s] async flips are only supported for fast updates\n",
+			       crtc->base.id, crtc->name);
+		return -EINVAL;
+	}
+
+	if (!state->legacy_cursor_update && amdgpu_dm_crtc_vrr_active(dm_crtc_state)) {
+		struct drm_plane_state *primary_state;
+
+		/* Pull in primary plane for correct VRR handling */
+		primary_state = drm_atomic_get_plane_state(state, crtc->primary);
+		if (IS_ERR(primary_state))
+			return PTR_ERR(primary_state);
+	}
+
+	/* In some use cases, like reset, no stream is attached */
+	if (!dm_crtc_state->stream)
+		return 0;
+
+	if (dc_validate_stream(dc, dm_crtc_state->stream) == DC_OK)
+		return 0;
+
+	DRM_DEBUG_ATOMIC("Failed DC stream validation\n");
+	return ret;
+}
+
+static const struct drm_crtc_helper_funcs amdgpu_dm_crtc_helper_funcs = {
+	.disable = amdgpu_dm_crtc_helper_disable,
+	.atomic_check = amdgpu_dm_crtc_helper_atomic_check,
+	.mode_fixup = amdgpu_dm_crtc_helper_mode_fixup,
+	.get_scanout_position = amdgpu_crtc_get_scanout_position,
+};
+
+int amdgpu_dm_crtc_init(struct amdgpu_display_manager *dm,
+			       struct drm_plane *plane,
+			       uint32_t crtc_index)
+{
+	struct amdgpu_crtc *acrtc = NULL;
+	struct drm_plane *cursor_plane;
+	bool has_degamma;
+	int res = -ENOMEM;
+
+	cursor_plane = kzalloc(sizeof(*cursor_plane), GFP_KERNEL);
+	if (!cursor_plane)
+		goto fail;
+
+	cursor_plane->type = DRM_PLANE_TYPE_CURSOR;
+	res = amdgpu_dm_plane_init(dm, cursor_plane, 0, NULL);
+
+	acrtc = kzalloc(sizeof(struct amdgpu_crtc), GFP_KERNEL);
+	if (!acrtc)
+		goto fail;
+
+	res = drm_crtc_init_with_planes(
+			dm->ddev,
+			&acrtc->base,
+			plane,
+			cursor_plane,
+			&amdgpu_dm_crtc_funcs, NULL);
+
+	if (res)
+		goto fail;
+
+	drm_crtc_helper_add(&acrtc->base, &amdgpu_dm_crtc_helper_funcs);
+
+	/* Create (reset) the plane state */
+	if (acrtc->base.funcs->reset)
+		acrtc->base.funcs->reset(&acrtc->base);
+
+	acrtc->max_cursor_width = dm->adev->dm.dc->caps.max_cursor_size;
+	acrtc->max_cursor_height = dm->adev->dm.dc->caps.max_cursor_size;
+
+	acrtc->crtc_id = crtc_index;
+	acrtc->base.enabled = false;
+	acrtc->otg_inst = -1;
+
+	dm->adev->mode_info.crtcs[crtc_index] = acrtc;
+
+	/* Don't enable DRM CRTC degamma property for
+	 * 1. Degamma is replaced by color pipeline.
+	 * 2. DCE since it doesn't support programmable degamma anywhere.
+	 * 3. DCN401 since pre-blending degamma LUT doesn't apply to cursor.
+	 */
+	if (plane->color_pipeline_property)
+		has_degamma = false;
+	else
+		has_degamma = dm->adev->dm.dc->caps.color.dpp.dcn_arch &&
+			      dm->adev->dm.dc->ctx->dce_version != DCN_VERSION_4_01;
+
+	drm_crtc_enable_color_mgmt(&acrtc->base, has_degamma ? MAX_COLOR_LUT_ENTRIES : 0,
+				   true, MAX_COLOR_LUT_ENTRIES);
+
+	drm_mode_crtc_set_gamma_size(&acrtc->base, MAX_COLOR_LEGACY_LUT_ENTRIES);
+
+#ifdef AMD_PRIVATE_COLOR
+	dm_crtc_additional_color_mgmt(&acrtc->base);
+#endif
+	return 0;
+
+fail:
+	kfree(acrtc);
+	kfree(cursor_plane);
+	return res;
+}
+
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crtc.h b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crtc.h
new file mode 100644
index 000000000000..c1212947a77b
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crtc.h
@@ -0,0 +1,51 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __AMDGPU_DM_CRTC_H__
+#define __AMDGPU_DM_CRTC_H__
+
+void amdgpu_dm_crtc_handle_vblank(struct amdgpu_crtc *acrtc);
+
+bool amdgpu_dm_crtc_modeset_required(struct drm_crtc_state *crtc_state,
+		      struct dc_stream_state *new_stream,
+		      struct dc_stream_state *old_stream);
+
+int amdgpu_dm_crtc_set_vupdate_irq(struct drm_crtc *crtc, bool enable);
+
+bool amdgpu_dm_crtc_vrr_active_irq(struct amdgpu_crtc *acrtc);
+
+bool amdgpu_dm_crtc_vrr_active(const struct dm_crtc_state *dm_state);
+
+int amdgpu_dm_crtc_enable_vblank(struct drm_crtc *crtc);
+
+void amdgpu_dm_crtc_disable_vblank(struct drm_crtc *crtc);
+
+int amdgpu_dm_crtc_init(struct amdgpu_display_manager *dm,
+			struct drm_plane *plane,
+			uint32_t link_index);
+
+#endif
+
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.c
index 9d43ecb1f692..a9839485f2a2 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.c
@@ -1,3 +1,4 @@
+// SPDX-License-Identifier: MIT
 /*
  * Copyright 2018 Advanced Micro Devices, Inc.
  *
@@ -23,19 +24,27 @@
  *
  */
 
+#include <linux/string_helpers.h>
 #include <linux/uaccess.h>
+#include <media/cec-notifier.h>
 
 #include "dc.h"
 #include "amdgpu.h"
 #include "amdgpu_dm.h"
 #include "amdgpu_dm_debugfs.h"
+#include "amdgpu_dm_replay.h"
 #include "dm_helpers.h"
 #include "dmub/dmub_srv.h"
 #include "resource.h"
 #include "dsc.h"
-#include "dc_link_dp.h"
 #include "link_hwss.h"
 #include "dc/dc_dmub_srv.h"
+#include "link/protocols/link_dp_capability.h"
+#include "inc/hw/dchubbub.h"
+
+#ifdef CONFIG_DRM_AMD_SECURE_DISPLAY
+#include "amdgpu_dm_psr.h"
+#endif
 
 struct dmub_debugfs_trace_header {
 	uint32_t entry_count;
@@ -49,10 +58,12 @@ struct dmub_debugfs_trace_entry {
 	uint32_t param1;
 };
 
-static inline const char *yesno(bool v)
-{
-	return v ? "yes" : "no";
-}
+static const char *const mst_progress_status[] = {
+	"probe",
+	"remote_edid",
+	"allocate_new_payload",
+	"clear_allocated_payload",
+};
 
 /* parse_write_buffer_into_params - Helper function to parse debugfs write buffer into an array
  *
@@ -227,8 +238,10 @@ static ssize_t dp_link_settings_read(struct file *f, char __user *buf,
 			break;
 
 		r = put_user(*(rd_buf + result), buf);
-		if (r)
+		if (r) {
+			kfree(rd_buf);
 			return r; /* r = -EFAULT */
+		}
 
 		buf += 1;
 		size -= 1;
@@ -245,8 +258,9 @@ static ssize_t dp_link_settings_write(struct file *f, const char __user *buf,
 {
 	struct amdgpu_dm_connector *connector = file_inode(f)->i_private;
 	struct dc_link *link = connector->dc_link;
+	struct amdgpu_device *adev = drm_to_adev(connector->base.dev);
 	struct dc *dc = (struct dc *)link->dc;
-	struct dc_link_settings prefer_link_settings;
+	struct dc_link_settings prefer_link_settings = {0};
 	char *wr_buf = NULL;
 	const uint32_t wr_buf_size = 40;
 	/* 0: lane_count; 1: link_rate */
@@ -292,9 +306,9 @@ static ssize_t dp_link_settings_write(struct file *f, const char __user *buf,
 	case LINK_RATE_RBR2:
 	case LINK_RATE_HIGH2:
 	case LINK_RATE_HIGH3:
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	case LINK_RATE_UHBR10:
-#endif
+	case LINK_RATE_UHBR13_5:
+	case LINK_RATE_UHBR20:
 		break;
 	default:
 		valid_input = false;
@@ -304,6 +318,9 @@ static ssize_t dp_link_settings_write(struct file *f, const char __user *buf,
 	if (!valid_input) {
 		kfree(wr_buf);
 		DRM_DEBUG_DRIVER("Invalid Input value No HW will be programmed\n");
+		mutex_lock(&adev->dm.dc_lock);
+		dc_link_set_preferred_training_settings(dc, NULL, NULL, link, false);
+		mutex_unlock(&adev->dm.dc_lock);
 		return size;
 	}
 
@@ -315,7 +332,156 @@ static ssize_t dp_link_settings_write(struct file *f, const char __user *buf,
 	prefer_link_settings.lane_count = param[0];
 	prefer_link_settings.link_rate = param[1];
 
+	mutex_lock(&adev->dm.dc_lock);
+	dc_link_set_preferred_training_settings(dc, &prefer_link_settings, NULL, link, false);
+	mutex_unlock(&adev->dm.dc_lock);
+
+	kfree(wr_buf);
+	return size;
+}
+
+static bool dp_mst_is_end_device(struct amdgpu_dm_connector *aconnector)
+{
+	bool is_end_device = false;
+	struct drm_dp_mst_topology_mgr *mgr = NULL;
+	struct drm_dp_mst_port *port = NULL;
+
+	if (aconnector->mst_root && aconnector->mst_root->mst_mgr.mst_state) {
+		mgr = &aconnector->mst_root->mst_mgr;
+		port = aconnector->mst_output_port;
+
+		drm_modeset_lock(&mgr->base.lock, NULL);
+		if (port->pdt == DP_PEER_DEVICE_SST_SINK ||
+			port->pdt == DP_PEER_DEVICE_DP_LEGACY_CONV)
+			is_end_device = true;
+		drm_modeset_unlock(&mgr->base.lock);
+	}
+
+	return is_end_device;
+}
+
+/* Change MST link setting
+ *
+ * valid lane count value: 1, 2, 4
+ * valid link rate value:
+ * 06h = 1.62Gbps per lane
+ * 0Ah = 2.7Gbps per lane
+ * 0Ch = 3.24Gbps per lane
+ * 14h = 5.4Gbps per lane
+ * 1Eh = 8.1Gbps per lane
+ * 3E8h = 10.0Gbps per lane
+ * 546h = 13.5Gbps per lane
+ * 7D0h = 20.0Gbps per lane
+ *
+ * debugfs is located at /sys/kernel/debug/dri/0/DP-x/mst_link_settings
+ *
+ * for example, to force to  2 lane, 10.0GHz,
+ * echo 2 0x3e8 > /sys/kernel/debug/dri/0/DP-x/mst_link_settings
+ *
+ * Valid input will trigger hotplug event to get new link setting applied
+ * Invalid input will trigger training setting reset
+ *
+ * The usage can be referred to link_settings entry
+ *
+ */
+static ssize_t dp_mst_link_setting(struct file *f, const char __user *buf,
+				 size_t size, loff_t *pos)
+{
+	struct amdgpu_dm_connector *aconnector = file_inode(f)->i_private;
+	struct dc_link *link = aconnector->dc_link;
+	struct amdgpu_device *adev = drm_to_adev(aconnector->base.dev);
+	struct dc *dc = (struct dc *)link->dc;
+	struct dc_link_settings prefer_link_settings = {0};
+	char *wr_buf = NULL;
+	const uint32_t wr_buf_size = 40;
+	/* 0: lane_count; 1: link_rate */
+	int max_param_num = 2;
+	uint8_t param_nums = 0;
+	long param[2];
+	bool valid_input = true;
+
+	if (!dp_mst_is_end_device(aconnector))
+		return -EINVAL;
+
+	if (size == 0)
+		return -EINVAL;
+
+	wr_buf = kcalloc(wr_buf_size, sizeof(char), GFP_KERNEL);
+	if (!wr_buf)
+		return -ENOSPC;
+
+	if (parse_write_buffer_into_params(wr_buf, wr_buf_size,
+					   (long *)param, buf,
+					   max_param_num,
+					   &param_nums)) {
+		kfree(wr_buf);
+		return -EINVAL;
+	}
+
+	if (param_nums <= 0) {
+		kfree(wr_buf);
+		DRM_DEBUG_DRIVER("user data not be read\n");
+		return -EINVAL;
+	}
+
+	switch (param[0]) {
+	case LANE_COUNT_ONE:
+	case LANE_COUNT_TWO:
+	case LANE_COUNT_FOUR:
+		break;
+	default:
+		valid_input = false;
+		break;
+	}
+
+	switch (param[1]) {
+	case LINK_RATE_LOW:
+	case LINK_RATE_HIGH:
+	case LINK_RATE_RBR2:
+	case LINK_RATE_HIGH2:
+	case LINK_RATE_HIGH3:
+	case LINK_RATE_UHBR10:
+	case LINK_RATE_UHBR13_5:
+	case LINK_RATE_UHBR20:
+		break;
+	default:
+		valid_input = false;
+		break;
+	}
+
+	if (!valid_input) {
+		kfree(wr_buf);
+		DRM_DEBUG_DRIVER("Invalid Input value No HW will be programmed\n");
+		mutex_lock(&adev->dm.dc_lock);
+		dc_link_set_preferred_training_settings(dc, NULL, NULL, link, false);
+		mutex_unlock(&adev->dm.dc_lock);
+		return -EINVAL;
+	}
+
+	/* save user force lane_count, link_rate to preferred settings
+	 * spread spectrum will not be changed
+	 */
+	prefer_link_settings.link_spread = link->cur_link_settings.link_spread;
+	prefer_link_settings.use_link_rate_set = false;
+	prefer_link_settings.lane_count = param[0];
+	prefer_link_settings.link_rate = param[1];
+
+	/* skip immediate retrain, and train to new link setting after hotplug event triggered */
+	mutex_lock(&adev->dm.dc_lock);
 	dc_link_set_preferred_training_settings(dc, &prefer_link_settings, NULL, link, true);
+	mutex_unlock(&adev->dm.dc_lock);
+
+	mutex_lock(&aconnector->base.dev->mode_config.mutex);
+	aconnector->base.force = DRM_FORCE_OFF;
+	mutex_unlock(&aconnector->base.dev->mode_config.mutex);
+	drm_kms_helper_hotplug_event(aconnector->base.dev);
+
+	msleep(100);
+
+	mutex_lock(&aconnector->base.dev->mode_config.mutex);
+	aconnector->base.force = DRM_FORCE_UNSPECIFIED;
+	mutex_unlock(&aconnector->base.dev->mode_config.mutex);
+	drm_kms_helper_hotplug_event(aconnector->base.dev);
 
 	kfree(wr_buf);
 	return size;
@@ -389,8 +555,10 @@ static ssize_t dp_phy_settings_read(struct file *f, char __user *buf,
 			break;
 
 		r = put_user((*(rd_buf + result)), buf);
-		if (r)
+		if (r) {
+			kfree(rd_buf);
 			return r; /* r = -EFAULT */
+		}
 
 		buf += 1;
 		size -= 1;
@@ -402,67 +570,38 @@ static ssize_t dp_phy_settings_read(struct file *f, char __user *buf,
 	return result;
 }
 
-static int dp_lttpr_status_show(struct seq_file *m, void *d)
+static int dp_lttpr_status_show(struct seq_file *m, void *unused)
 {
-	char *data;
-	struct amdgpu_dm_connector *connector = file_inode(m->file)->i_private;
-	struct dc_link *link = connector->dc_link;
-	uint32_t read_size = 1;
-	uint8_t repeater_count = 0;
+	struct drm_connector *connector = m->private;
+	struct amdgpu_dm_connector *aconnector =
+		to_amdgpu_dm_connector(connector);
+	struct dc_lttpr_caps caps = aconnector->dc_link->dpcd_caps.lttpr_caps;
 
-	data = kzalloc(read_size, GFP_KERNEL);
-	if (!data)
-		return 0;
+	if (connector->status != connector_status_connected)
+		return -ENODEV;
 
-	dm_helpers_dp_read_dpcd(link->ctx, link, 0xF0002, data, read_size);
+	seq_printf(m, "phy repeater count: %u (raw: 0x%x)\n",
+		   dp_parse_lttpr_repeater_count(caps.phy_repeater_cnt),
+		   caps.phy_repeater_cnt);
 
-	switch ((uint8_t)*data) {
-	case 0x80:
-		repeater_count = 1;
-		break;
-	case 0x40:
-		repeater_count = 2;
-		break;
-	case 0x20:
-		repeater_count = 3;
-		break;
-	case 0x10:
-		repeater_count = 4;
-		break;
-	case 0x8:
-		repeater_count = 5;
-		break;
-	case 0x4:
-		repeater_count = 6;
-		break;
-	case 0x2:
-		repeater_count = 7;
+	seq_puts(m, "phy repeater mode: ");
+
+	switch (caps.mode) {
+	case DP_PHY_REPEATER_MODE_TRANSPARENT:
+		seq_puts(m, "transparent");
 		break;
-	case 0x1:
-		repeater_count = 8;
+	case DP_PHY_REPEATER_MODE_NON_TRANSPARENT:
+		seq_puts(m, "non-transparent");
 		break;
-	case 0x0:
-		repeater_count = 0;
+	case 0x00:
+		seq_puts(m, "non lttpr");
 		break;
 	default:
-		repeater_count = (uint8_t)*data;
+		seq_printf(m, "read error (raw: 0x%x)", caps.mode);
 		break;
 	}
 
-	seq_printf(m, "phy repeater count: %d\n", repeater_count);
-
-	dm_helpers_dp_read_dpcd(link->ctx, link, 0xF0003, data, read_size);
-
-	if ((uint8_t)*data == 0x55)
-		seq_printf(m, "phy repeater mode: transparent\n");
-	else if ((uint8_t)*data == 0xAA)
-		seq_printf(m, "phy repeater mode: non-transparent\n");
-	else if ((uint8_t)*data == 0x00)
-		seq_printf(m, "phy repeater mode: non lttpr\n");
-	else
-		seq_printf(m, "phy repeater mode: read error\n");
-
-	kfree(data);
+	seq_puts(m, "\n");
 	return 0;
 }
 
@@ -476,7 +615,7 @@ static ssize_t dp_phy_settings_write(struct file *f, const char __user *buf,
 	uint32_t wr_buf_size = 40;
 	long param[3];
 	bool use_prefer_link_setting;
-	struct link_training_settings link_lane_settings;
+	struct link_training_settings link_lane_settings = {0};
 	int max_param_num = 3;
 	uint8_t param_nums = 0;
 	int r = 0;
@@ -536,11 +675,11 @@ static ssize_t dp_phy_settings_write(struct file *f, const char __user *buf,
 
 	/* apply phy settings from user */
 	for (r = 0; r < link_lane_settings.link_settings.lane_count; r++) {
-		link_lane_settings.lane_settings[r].VOLTAGE_SWING =
+		link_lane_settings.hw_lane_settings[r].VOLTAGE_SWING =
 				(enum dc_voltage_swing) (param[0]);
-		link_lane_settings.lane_settings[r].PRE_EMPHASIS =
+		link_lane_settings.hw_lane_settings[r].PRE_EMPHASIS =
 				(enum dc_pre_emphasis) (param[1]);
-		link_lane_settings.lane_settings[r].POST_CURSOR2 =
+		link_lane_settings.hw_lane_settings[r].POST_CURSOR2 =
 				(enum dc_post_cursor2) (param[2]);
 	}
 
@@ -620,6 +759,7 @@ static ssize_t dp_phy_test_pattern_debugfs_write(struct file *f, const char __us
 	int max_param_num = 11;
 	enum dp_test_pattern test_pattern = DP_TEST_PATTERN_UNSUPPORTED;
 	bool disable_hpd = false;
+	bool supports_hpd = link->irq_source_hpd != DC_IRQ_SOURCE_INVALID;
 	bool valid_test_pattern = false;
 	uint8_t param_nums = 0;
 	/* init with default 80bit custom pattern */
@@ -631,7 +771,7 @@ static ssize_t dp_phy_test_pattern_debugfs_write(struct file *f, const char __us
 			LINK_RATE_UNKNOWN, LINK_SPREAD_DISABLED};
 	struct dc_link_settings cur_link_settings = {LANE_COUNT_UNKNOWN,
 			LINK_RATE_UNKNOWN, LINK_SPREAD_DISABLED};
-	struct link_training_settings link_training_settings;
+	struct link_training_settings link_training_settings = {0};
 	int i;
 
 	if (size == 0)
@@ -711,7 +851,7 @@ static ssize_t dp_phy_test_pattern_debugfs_write(struct file *f, const char __us
 	 * because it might have been disabled after a test pattern was set.
 	 * AUX depends on HPD * sequence dependent, do not move!
 	 */
-	if (!disable_hpd)
+	if (supports_hpd && !disable_hpd)
 		dc_link_enable_hpd(link);
 
 	prefer_link_settings.lane_count = link->verified_link_cap.lane_count;
@@ -734,9 +874,9 @@ static ssize_t dp_phy_test_pattern_debugfs_write(struct file *f, const char __us
 	}
 
 	for (i = 0; i < (unsigned int)(link_training_settings.link_settings.lane_count); i++)
-		link_training_settings.lane_settings[i] = link->cur_lane_setting[i];
+		link_training_settings.hw_lane_settings[i] = link->cur_lane_setting[i];
 
-	dc_link_set_test_pattern(
+	dc_link_dp_set_test_pattern(
 		link,
 		test_pattern,
 		DP_TEST_PATTERN_COLOR_SPACE_RGB,
@@ -749,7 +889,7 @@ static ssize_t dp_phy_test_pattern_debugfs_write(struct file *f, const char __us
 	 * Need disable interrupt to avoid SW driver disable DP output. This is
 	 * done after the test pattern is set.
 	 */
-	if (valid_test_pattern && disable_hpd)
+	if (valid_test_pattern && supports_hpd && disable_hpd)
 		dc_link_disable_hpd(link);
 
 	kfree(wr_buf);
@@ -765,9 +905,10 @@ static int dmub_tracebuffer_show(struct seq_file *m, void *data)
 {
 	struct amdgpu_device *adev = m->private;
 	struct dmub_srv_fb_info *fb_info = adev->dm.dmub_fb_info;
+	struct dmub_fw_meta_info *fw_meta_info = NULL;
 	struct dmub_debugfs_trace_entry *entries;
 	uint8_t *tbuf_base;
-	uint32_t tbuf_size, max_entries, num_entries, i;
+	uint32_t tbuf_size, max_entries, num_entries, first_entry, i;
 
 	if (!fb_info)
 		return 0;
@@ -776,20 +917,42 @@ static int dmub_tracebuffer_show(struct seq_file *m, void *data)
 	if (!tbuf_base)
 		return 0;
 
-	tbuf_size = fb_info->fb[DMUB_WINDOW_5_TRACEBUFF].size;
+	if (adev->dm.dmub_srv)
+		fw_meta_info = &adev->dm.dmub_srv->meta_info;
+
+	tbuf_size = fw_meta_info ? fw_meta_info->trace_buffer_size :
+				   DMUB_TRACE_BUFFER_SIZE;
 	max_entries = (tbuf_size - sizeof(struct dmub_debugfs_trace_header)) /
 		      sizeof(struct dmub_debugfs_trace_entry);
 
 	num_entries =
 		((struct dmub_debugfs_trace_header *)tbuf_base)->entry_count;
 
+	/* DMCUB tracebuffer is a ring. If it rolled over, print a hint that
+	 * entries are being overwritten.
+	 */
+	if (num_entries > max_entries)
+		seq_printf(m, "...\n");
+
+	first_entry = num_entries % max_entries;
 	num_entries = min(num_entries, max_entries);
 
 	entries = (struct dmub_debugfs_trace_entry
 			   *)(tbuf_base +
 			      sizeof(struct dmub_debugfs_trace_header));
 
-	for (i = 0; i < num_entries; ++i) {
+	/* To print entries chronologically, start from the first entry till the
+	 * top of buffer, then from base of buffer to first entry.
+	 */
+	for (i = first_entry; i < num_entries; ++i) {
+		struct dmub_debugfs_trace_entry *entry = &entries[i];
+
+		seq_printf(m,
+			   "trace_code=%u tick_count=%u param0=%u param1=%u\n",
+			   entry->trace_code, entry->tick_count, entry->param0,
+			   entry->param1);
+	}
+	for (i = 0; i < first_entry; ++i) {
 		struct dmub_debugfs_trace_entry *entry = &entries[i];
 
 		seq_printf(m,
@@ -824,29 +987,113 @@ static int dmub_fw_state_show(struct seq_file *m, void *data)
 	return seq_write(m, state_base, state_size);
 }
 
-/*
- * Returns the current and maximum output bpc for the connector.
- * Example usage: cat /sys/kernel/debug/dri/0/DP-1/output_bpc
+/* replay_capability_show() - show eDP panel replay capability
+ *
+ * The read function: replay_capability_show
+ * Shows if sink and driver has Replay capability or not.
+ *
+ *	cat /sys/kernel/debug/dri/0/eDP-X/replay_capability
+ *
+ * Expected output:
+ * "Sink support: no\n" - if panel doesn't support Replay
+ * "Sink support: yes\n" - if panel supports Replay
+ * "Driver support: no\n" - if driver doesn't support Replay
+ * "Driver support: yes\n" - if driver supports Replay
  */
-static int output_bpc_show(struct seq_file *m, void *data)
+static int replay_capability_show(struct seq_file *m, void *data)
 {
 	struct drm_connector *connector = m->private;
-	struct drm_device *dev = connector->dev;
-	struct drm_crtc *crtc = NULL;
+	struct amdgpu_dm_connector *aconnector = to_amdgpu_dm_connector(connector);
+	struct dc_link *link = aconnector->dc_link;
+	bool sink_support_replay = false;
+	bool driver_support_replay = false;
+
+	if (!link)
+		return -ENODEV;
+
+	if (link->type == dc_connection_none)
+		return -ENODEV;
+
+	if (!(link->connector_signal & SIGNAL_TYPE_EDP))
+		return -ENODEV;
+
+	/* If Replay is already set to support, skip the checks */
+	if (link->replay_settings.config.replay_supported) {
+		sink_support_replay = true;
+		driver_support_replay = true;
+	} else if ((amdgpu_dc_debug_mask & DC_DISABLE_REPLAY)) {
+		sink_support_replay = amdgpu_dm_link_supports_replay(link, aconnector);
+	} else {
+		struct dc *dc = link->ctx->dc;
+
+		sink_support_replay = amdgpu_dm_link_supports_replay(link, aconnector);
+		if (dc->ctx->dmub_srv && dc->ctx->dmub_srv->dmub)
+			driver_support_replay =
+				(bool)dc->ctx->dmub_srv->dmub->feature_caps.replay_supported;
+	}
+
+	seq_printf(m, "Sink support: %s\n", str_yes_no(sink_support_replay));
+	seq_printf(m, "Driver support: %s\n", str_yes_no(driver_support_replay));
+	seq_printf(m, "Config support: %s\n", str_yes_no(link->replay_settings.config.replay_supported));
+
+	return 0;
+}
+
+/* psr_capability_show() - show eDP panel PSR capability
+ *
+ * The read function: sink_psr_capability_show
+ * Shows if sink has PSR capability or not.
+ * If yes - the PSR version is appended
+ *
+ *	cat /sys/kernel/debug/dri/0/eDP-X/psr_capability
+ *
+ * Expected output:
+ * "Sink support: no\n" - if panel doesn't support PSR
+ * "Sink support: yes [0x01]\n" - if panel supports PSR1
+ * "Driver support: no\n" - if driver doesn't support PSR
+ * "Driver support: yes [0x01]\n" - if driver supports PSR1
+ */
+static int psr_capability_show(struct seq_file *m, void *data)
+{
+	struct drm_connector *connector = m->private;
+	struct amdgpu_dm_connector *aconnector = to_amdgpu_dm_connector(connector);
+	struct dc_link *link = aconnector->dc_link;
+
+	if (!link)
+		return -ENODEV;
+
+	if (link->type == dc_connection_none)
+		return -ENODEV;
+
+	if (!(link->connector_signal & SIGNAL_TYPE_EDP))
+		return -ENODEV;
+
+	seq_printf(m, "Sink support: %s", str_yes_no(link->dpcd_caps.psr_info.psr_version != 0));
+	if (link->dpcd_caps.psr_info.psr_version)
+		seq_printf(m, " [0x%02x]", link->dpcd_caps.psr_info.psr_version);
+	seq_puts(m, "\n");
+
+	seq_printf(m, "Driver support: %s", str_yes_no(link->psr_settings.psr_feature_enabled));
+	if (link->psr_settings.psr_version)
+		seq_printf(m, " [0x%02x]", link->psr_settings.psr_version);
+	seq_puts(m, "\n");
+
+	return 0;
+}
+
+/*
+ * Returns the current bpc for the crtc.
+ * Example usage: cat /sys/kernel/debug/dri/0/crtc-0/amdgpu_current_bpc
+ */
+static int amdgpu_current_bpc_show(struct seq_file *m, void *data)
+{
+	struct drm_crtc *crtc = m->private;
+	struct drm_device *dev = crtc->dev;
 	struct dm_crtc_state *dm_crtc_state = NULL;
 	int res = -ENODEV;
 	unsigned int bpc;
 
 	mutex_lock(&dev->mode_config.mutex);
-	drm_modeset_lock(&dev->mode_config.connection_mutex, NULL);
-
-	if (connector->state == NULL)
-		goto unlock;
-
-	crtc = connector->state->crtc;
-	if (crtc == NULL)
-		goto unlock;
-
 	drm_modeset_lock(&crtc->mutex, NULL);
 	if (crtc->state == NULL)
 		goto unlock;
@@ -876,18 +1123,70 @@ static int output_bpc_show(struct seq_file *m, void *data)
 	}
 
 	seq_printf(m, "Current: %u\n", bpc);
-	seq_printf(m, "Maximum: %u\n", connector->display_info.bpc);
 	res = 0;
 
 unlock:
-	if (crtc)
-		drm_modeset_unlock(&crtc->mutex);
+	drm_modeset_unlock(&crtc->mutex);
+	mutex_unlock(&dev->mode_config.mutex);
 
-	drm_modeset_unlock(&dev->mode_config.connection_mutex);
+	return res;
+}
+DEFINE_SHOW_ATTRIBUTE(amdgpu_current_bpc);
+
+/*
+ * Returns the current colorspace for the crtc.
+ * Example usage: cat /sys/kernel/debug/dri/0/crtc-0/amdgpu_current_colorspace
+ */
+static int amdgpu_current_colorspace_show(struct seq_file *m, void *data)
+{
+	struct drm_crtc *crtc = m->private;
+	struct drm_device *dev = crtc->dev;
+	struct dm_crtc_state *dm_crtc_state = NULL;
+	int res = -ENODEV;
+
+	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_lock(&crtc->mutex, NULL);
+	if (crtc->state == NULL)
+		goto unlock;
+
+	dm_crtc_state = to_dm_crtc_state(crtc->state);
+	if (dm_crtc_state->stream == NULL)
+		goto unlock;
+
+	switch (dm_crtc_state->stream->output_color_space) {
+	case COLOR_SPACE_SRGB:
+		seq_puts(m, "sRGB");
+		break;
+	case COLOR_SPACE_YCBCR601:
+	case COLOR_SPACE_YCBCR601_LIMITED:
+		seq_puts(m, "BT601_YCC");
+		break;
+	case COLOR_SPACE_YCBCR709:
+	case COLOR_SPACE_YCBCR709_LIMITED:
+		seq_puts(m, "BT709_YCC");
+		break;
+	case COLOR_SPACE_ADOBERGB:
+		seq_puts(m, "opRGB");
+		break;
+	case COLOR_SPACE_2020_RGB_FULLRANGE:
+		seq_puts(m, "BT2020_RGB");
+		break;
+	case COLOR_SPACE_2020_YCBCR_LIMITED:
+		seq_puts(m, "BT2020_YCC");
+		break;
+	default:
+		goto unlock;
+	}
+	res = 0;
+
+unlock:
+	drm_modeset_unlock(&crtc->mutex);
 	mutex_unlock(&dev->mode_config.mutex);
 
 	return res;
 }
+DEFINE_SHOW_ATTRIBUTE(amdgpu_current_colorspace);
+
 
 /*
  * Example usage:
@@ -930,7 +1229,6 @@ static ssize_t dp_dsc_passthrough_set(struct file *f, const char __user *buf,
 	return 0;
 }
 
-#ifdef CONFIG_DRM_AMD_DC_HDCP
 /*
  * Returns the HDCP capability of the Display (1.4 for now).
  *
@@ -967,7 +1265,6 @@ static int hdcp_sink_capability_show(struct seq_file *m, void *data)
 
 	return 0;
 }
-#endif
 
 /*
  * Returns whether the connected display is internal and not hotpluggable.
@@ -984,6 +1281,36 @@ static int internal_display_show(struct seq_file *m, void *data)
 	return 0;
 }
 
+/*
+ * Returns the number of segments used if ODM Combine mode is enabled.
+ * Example usage: cat /sys/kernel/debug/dri/0/DP-1/odm_combine_segments
+ */
+static int odm_combine_segments_show(struct seq_file *m, void *unused)
+{
+	struct drm_connector *connector = m->private;
+	struct amdgpu_dm_connector *aconnector = to_amdgpu_dm_connector(connector);
+	struct dc_link *link = aconnector->dc_link;
+	struct pipe_ctx *pipe_ctx = NULL;
+	int i, segments = -EOPNOTSUPP;
+
+	for (i = 0; i < MAX_PIPES; i++) {
+		pipe_ctx = &link->dc->current_state->res_ctx.pipe_ctx[i];
+		if (pipe_ctx->stream &&
+		    pipe_ctx->stream->link == link)
+			break;
+	}
+
+	if (connector->status != connector_status_connected)
+		return -ENODEV;
+
+	if (pipe_ctx && pipe_ctx->stream_res.tg &&
+	    pipe_ctx->stream_res.tg->funcs->get_odm_combine_segments)
+		pipe_ctx->stream_res.tg->funcs->get_odm_combine_segments(pipe_ctx->stream_res.tg, &segments);
+
+	seq_printf(m, "%d\n", segments);
+	return 0;
+}
+
 /* function description
  *
  * generic SDP message access for testing
@@ -1002,7 +1329,7 @@ static ssize_t dp_sdp_message_debugfs_write(struct file *f, const char __user *b
 				 size_t size, loff_t *pos)
 {
 	int r;
-	uint8_t data[36];
+	uint8_t data[36] = {0};
 	struct amdgpu_dm_connector *connector = file_inode(f)->i_private;
 	struct dm_crtc_state *acrtc_state;
 	uint32_t write_size = 36;
@@ -1024,88 +1351,6 @@ static ssize_t dp_sdp_message_debugfs_write(struct file *f, const char __user *b
 	return write_size;
 }
 
-static ssize_t dp_dpcd_address_write(struct file *f, const char __user *buf,
-				 size_t size, loff_t *pos)
-{
-	int r;
-	struct amdgpu_dm_connector *connector = file_inode(f)->i_private;
-
-	if (size < sizeof(connector->debugfs_dpcd_address))
-		return -EINVAL;
-
-	r = copy_from_user(&connector->debugfs_dpcd_address,
-			buf, sizeof(connector->debugfs_dpcd_address));
-
-	return size - r;
-}
-
-static ssize_t dp_dpcd_size_write(struct file *f, const char __user *buf,
-				 size_t size, loff_t *pos)
-{
-	int r;
-	struct amdgpu_dm_connector *connector = file_inode(f)->i_private;
-
-	if (size < sizeof(connector->debugfs_dpcd_size))
-		return -EINVAL;
-
-	r = copy_from_user(&connector->debugfs_dpcd_size,
-			buf, sizeof(connector->debugfs_dpcd_size));
-
-	if (connector->debugfs_dpcd_size > 256)
-		connector->debugfs_dpcd_size = 0;
-
-	return size - r;
-}
-
-static ssize_t dp_dpcd_data_write(struct file *f, const char __user *buf,
-				 size_t size, loff_t *pos)
-{
-	int r;
-	char *data;
-	struct amdgpu_dm_connector *connector = file_inode(f)->i_private;
-	struct dc_link *link = connector->dc_link;
-	uint32_t write_size = connector->debugfs_dpcd_size;
-
-	if (!write_size || size < write_size)
-		return -EINVAL;
-
-	data = kzalloc(write_size, GFP_KERNEL);
-	if (!data)
-		return 0;
-
-	r = copy_from_user(data, buf, write_size);
-
-	dm_helpers_dp_write_dpcd(link->ctx, link,
-			connector->debugfs_dpcd_address, data, write_size - r);
-	kfree(data);
-	return write_size - r;
-}
-
-static ssize_t dp_dpcd_data_read(struct file *f, char __user *buf,
-				 size_t size, loff_t *pos)
-{
-	int r;
-	char *data;
-	struct amdgpu_dm_connector *connector = file_inode(f)->i_private;
-	struct dc_link *link = connector->dc_link;
-	uint32_t read_size = connector->debugfs_dpcd_size;
-
-	if (!read_size || size < read_size)
-		return 0;
-
-	data = kzalloc(read_size, GFP_KERNEL);
-	if (!data)
-		return 0;
-
-	dm_helpers_dp_read_dpcd(link->ctx, link,
-			connector->debugfs_dpcd_address, data, read_size);
-
-	r = copy_to_user(buf, data, read_size);
-
-	kfree(data);
-	return read_size - r;
-}
-
 /* function: Read link's DSC & FEC capabilities
  *
  *
@@ -1146,7 +1391,7 @@ static int dp_dsc_fec_support_show(struct seq_file *m, void *data)
 			break;
 		}
 		dpcd_caps = aconnector->dc_link->dpcd_caps;
-		if (aconnector->port) {
+		if (aconnector->mst_output_port) {
 			/* aconnector sets dsc_aux during get_modes call
 			 * if MST connector has it means it can either
 			 * enable DSC on the sink device or on MST branch
@@ -1165,8 +1410,8 @@ static int dp_dsc_fec_support_show(struct seq_file *m, void *data)
 	drm_modeset_drop_locks(&ctx);
 	drm_modeset_acquire_fini(&ctx);
 
-	seq_printf(m, "FEC_Sink_Support: %s\n", yesno(is_fec_supported));
-	seq_printf(m, "DSC_Sink_Support: %s\n", yesno(is_dsc_supported));
+	seq_printf(m, "FEC_Sink_Support: %s\n", str_yes_no(is_fec_supported));
+	seq_printf(m, "DSC_Sink_Support: %s\n", str_yes_no(is_dsc_supported));
 
 	return ret;
 }
@@ -1193,14 +1438,16 @@ static ssize_t trigger_hotplug(struct file *f, const char __user *buf,
 	struct drm_connector *connector = &aconnector->base;
 	struct dc_link *link = NULL;
 	struct drm_device *dev = connector->dev;
+	struct amdgpu_device *adev = drm_to_adev(dev);
 	enum dc_connection_type new_connection_type = dc_connection_none;
 	char *wr_buf = NULL;
 	uint32_t wr_buf_size = 42;
 	int max_param_num = 1;
 	long param[1] = {0};
 	uint8_t param_nums = 0;
+	bool ret = false;
 
-	if (!aconnector || !aconnector->dc_link)
+	if (!aconnector->dc_link)
 		return -EINVAL;
 
 	if (size == 0)
@@ -1221,20 +1468,32 @@ static ssize_t trigger_hotplug(struct file *f, const char __user *buf,
 		return -EINVAL;
 	}
 
+	kfree(wr_buf);
+
 	if (param_nums <= 0) {
 		DRM_DEBUG_DRIVER("user data not be read\n");
-		kfree(wr_buf);
+		return -EINVAL;
+	}
+
+	mutex_lock(&aconnector->hpd_lock);
+
+	/* Don't support for mst end device*/
+	if (aconnector->mst_root) {
+		mutex_unlock(&aconnector->hpd_lock);
 		return -EINVAL;
 	}
 
 	if (param[0] == 1) {
-		mutex_lock(&aconnector->hpd_lock);
 
-		if (!dc_link_detect_sink(aconnector->dc_link, &new_connection_type) &&
+		if (!dc_link_detect_connection_type(aconnector->dc_link, &new_connection_type) &&
 			new_connection_type != dc_connection_none)
 			goto unlock;
 
-		if (!dc_link_detect(aconnector->dc_link, DETECT_REASON_HPD))
+		mutex_lock(&adev->dm.dc_lock);
+		ret = dc_link_detect(aconnector->dc_link, DETECT_REASON_HPD);
+		mutex_unlock(&adev->dm.dc_lock);
+
+		if (!ret)
 			goto unlock;
 
 		amdgpu_dm_update_connector_after_detect(aconnector);
@@ -1243,7 +1502,7 @@ static ssize_t trigger_hotplug(struct file *f, const char __user *buf,
 		dm_restore_drm_connector_state(dev, connector);
 		drm_modeset_unlock_all(dev);
 
-		drm_kms_helper_hotplug_event(dev);
+		drm_kms_helper_connector_hotplug_event(connector);
 	} else if (param[0] == 0) {
 		if (!aconnector->dc_link)
 			goto unlock;
@@ -1261,17 +1520,20 @@ static ssize_t trigger_hotplug(struct file *f, const char __user *buf,
 
 		amdgpu_dm_update_connector_after_detect(aconnector);
 
+		/* If the aconnector is the root node in mst topology */
+		if (aconnector->mst_mgr.mst_state == true)
+			dc_link_reset_cur_dp_mst_topology(link);
+
 		drm_modeset_lock_all(dev);
 		dm_restore_drm_connector_state(dev, connector);
 		drm_modeset_unlock_all(dev);
 
-		drm_kms_helper_hotplug_event(dev);
+		drm_kms_helper_connector_hotplug_event(connector);
 	}
 
 unlock:
 	mutex_unlock(&aconnector->hpd_lock);
 
-	kfree(wr_buf);
 	return size;
 }
 
@@ -1294,48 +1556,45 @@ static ssize_t dp_dsc_clock_en_read(struct file *f, char __user *buf,
 				    size_t size, loff_t *pos)
 {
 	char *rd_buf = NULL;
-	char *rd_buf_ptr = NULL;
 	struct amdgpu_dm_connector *aconnector = file_inode(f)->i_private;
 	struct display_stream_compressor *dsc;
 	struct dcn_dsc_state dsc_state = {0};
 	const uint32_t rd_buf_size = 10;
 	struct pipe_ctx *pipe_ctx;
 	ssize_t result = 0;
-	int i, r, str_len = 30;
+	int i, r, str_len = 10;
 
 	rd_buf = kcalloc(rd_buf_size, sizeof(char), GFP_KERNEL);
 
 	if (!rd_buf)
 		return -ENOMEM;
 
-	rd_buf_ptr = rd_buf;
-
 	for (i = 0; i < MAX_PIPES; i++) {
 		pipe_ctx = &aconnector->dc_link->dc->current_state->res_ctx.pipe_ctx[i];
-			if (pipe_ctx && pipe_ctx->stream &&
-			    pipe_ctx->stream->link == aconnector->dc_link)
-				break;
+		if (pipe_ctx->stream &&
+		    pipe_ctx->stream->link == aconnector->dc_link &&
+		    pipe_ctx->stream->sink &&
+		    pipe_ctx->stream->sink == aconnector->dc_sink)
+			break;
 	}
 
-	if (!pipe_ctx)
-		return -ENXIO;
-
 	dsc = pipe_ctx->stream_res.dsc;
 	if (dsc)
 		dsc->funcs->dsc_read_state(dsc, &dsc_state);
 
-	snprintf(rd_buf_ptr, str_len,
+	snprintf(rd_buf, str_len,
 		"%d\n",
 		dsc_state.dsc_clock_en);
-	rd_buf_ptr += str_len;
 
 	while (size) {
 		if (*pos >= rd_buf_size)
 			break;
 
 		r = put_user(*(rd_buf + result), buf);
-		if (r)
+		if (r) {
+			kfree(rd_buf);
 			return r; /* r = -EFAULT */
+		}
 
 		buf += 1;
 		size -= 1;
@@ -1414,12 +1673,14 @@ static ssize_t dp_dsc_clock_en_write(struct file *f, const char __user *buf,
 
 	for (i = 0; i < MAX_PIPES; i++) {
 		pipe_ctx = &aconnector->dc_link->dc->current_state->res_ctx.pipe_ctx[i];
-			if (pipe_ctx && pipe_ctx->stream &&
-			    pipe_ctx->stream->link == aconnector->dc_link)
-				break;
+		if (pipe_ctx->stream &&
+		    pipe_ctx->stream->link == aconnector->dc_link &&
+		    pipe_ctx->stream->sink &&
+		    pipe_ctx->stream->sink == aconnector->dc_sink)
+			break;
 	}
 
-	if (!pipe_ctx || !pipe_ctx->stream)
+	if (!pipe_ctx->stream)
 		goto done;
 
 	// Get CRTC state
@@ -1481,7 +1742,6 @@ static ssize_t dp_dsc_slice_width_read(struct file *f, char __user *buf,
 				    size_t size, loff_t *pos)
 {
 	char *rd_buf = NULL;
-	char *rd_buf_ptr = NULL;
 	struct amdgpu_dm_connector *aconnector = file_inode(f)->i_private;
 	struct display_stream_compressor *dsc;
 	struct dcn_dsc_state dsc_state = {0};
@@ -1495,34 +1755,32 @@ static ssize_t dp_dsc_slice_width_read(struct file *f, char __user *buf,
 	if (!rd_buf)
 		return -ENOMEM;
 
-	rd_buf_ptr = rd_buf;
-
 	for (i = 0; i < MAX_PIPES; i++) {
 		pipe_ctx = &aconnector->dc_link->dc->current_state->res_ctx.pipe_ctx[i];
-			if (pipe_ctx && pipe_ctx->stream &&
-			    pipe_ctx->stream->link == aconnector->dc_link)
-				break;
+		if (pipe_ctx->stream &&
+		    pipe_ctx->stream->link == aconnector->dc_link &&
+		    pipe_ctx->stream->sink &&
+		    pipe_ctx->stream->sink == aconnector->dc_sink)
+			break;
 	}
 
-	if (!pipe_ctx)
-		return -ENXIO;
-
 	dsc = pipe_ctx->stream_res.dsc;
 	if (dsc)
 		dsc->funcs->dsc_read_state(dsc, &dsc_state);
 
-	snprintf(rd_buf_ptr, str_len,
+	snprintf(rd_buf, str_len,
 		"%d\n",
 		dsc_state.dsc_slice_width);
-	rd_buf_ptr += str_len;
 
 	while (size) {
 		if (*pos >= rd_buf_size)
 			break;
 
 		r = put_user(*(rd_buf + result), buf);
-		if (r)
+		if (r) {
+			kfree(rd_buf);
 			return r; /* r = -EFAULT */
+		}
 
 		buf += 1;
 		size -= 1;
@@ -1599,12 +1857,14 @@ static ssize_t dp_dsc_slice_width_write(struct file *f, const char __user *buf,
 
 	for (i = 0; i < MAX_PIPES; i++) {
 		pipe_ctx = &aconnector->dc_link->dc->current_state->res_ctx.pipe_ctx[i];
-			if (pipe_ctx && pipe_ctx->stream &&
-			    pipe_ctx->stream->link == aconnector->dc_link)
-				break;
+		if (pipe_ctx->stream &&
+		    pipe_ctx->stream->link == aconnector->dc_link &&
+		    pipe_ctx->stream->sink &&
+		    pipe_ctx->stream->sink == aconnector->dc_sink)
+			break;
 	}
 
-	if (!pipe_ctx || !pipe_ctx->stream)
+	if (!pipe_ctx->stream)
 		goto done;
 
 	// Safely get CRTC state
@@ -1666,7 +1926,6 @@ static ssize_t dp_dsc_slice_height_read(struct file *f, char __user *buf,
 				    size_t size, loff_t *pos)
 {
 	char *rd_buf = NULL;
-	char *rd_buf_ptr = NULL;
 	struct amdgpu_dm_connector *aconnector = file_inode(f)->i_private;
 	struct display_stream_compressor *dsc;
 	struct dcn_dsc_state dsc_state = {0};
@@ -1680,34 +1939,32 @@ static ssize_t dp_dsc_slice_height_read(struct file *f, char __user *buf,
 	if (!rd_buf)
 		return -ENOMEM;
 
-	rd_buf_ptr = rd_buf;
-
 	for (i = 0; i < MAX_PIPES; i++) {
 		pipe_ctx = &aconnector->dc_link->dc->current_state->res_ctx.pipe_ctx[i];
-			if (pipe_ctx && pipe_ctx->stream &&
-			    pipe_ctx->stream->link == aconnector->dc_link)
-				break;
+		if (pipe_ctx->stream &&
+		    pipe_ctx->stream->link == aconnector->dc_link &&
+		    pipe_ctx->stream->sink &&
+		    pipe_ctx->stream->sink == aconnector->dc_sink)
+			break;
 	}
 
-	if (!pipe_ctx)
-		return -ENXIO;
-
 	dsc = pipe_ctx->stream_res.dsc;
 	if (dsc)
 		dsc->funcs->dsc_read_state(dsc, &dsc_state);
 
-	snprintf(rd_buf_ptr, str_len,
+	snprintf(rd_buf, str_len,
 		"%d\n",
 		dsc_state.dsc_slice_height);
-	rd_buf_ptr += str_len;
 
 	while (size) {
 		if (*pos >= rd_buf_size)
 			break;
 
 		r = put_user(*(rd_buf + result), buf);
-		if (r)
+		if (r) {
+			kfree(rd_buf);
 			return r; /* r = -EFAULT */
+		}
 
 		buf += 1;
 		size -= 1;
@@ -1784,12 +2041,14 @@ static ssize_t dp_dsc_slice_height_write(struct file *f, const char __user *buf,
 
 	for (i = 0; i < MAX_PIPES; i++) {
 		pipe_ctx = &aconnector->dc_link->dc->current_state->res_ctx.pipe_ctx[i];
-			if (pipe_ctx && pipe_ctx->stream &&
-			    pipe_ctx->stream->link == aconnector->dc_link)
-				break;
+		if (pipe_ctx->stream &&
+		    pipe_ctx->stream->link == aconnector->dc_link &&
+		    pipe_ctx->stream->sink &&
+		    pipe_ctx->stream->sink == aconnector->dc_sink)
+			break;
 	}
 
-	if (!pipe_ctx || !pipe_ctx->stream)
+	if (!pipe_ctx->stream)
 		goto done;
 
 	// Get CRTC state
@@ -1847,7 +2106,6 @@ static ssize_t dp_dsc_bits_per_pixel_read(struct file *f, char __user *buf,
 				    size_t size, loff_t *pos)
 {
 	char *rd_buf = NULL;
-	char *rd_buf_ptr = NULL;
 	struct amdgpu_dm_connector *aconnector = file_inode(f)->i_private;
 	struct display_stream_compressor *dsc;
 	struct dcn_dsc_state dsc_state = {0};
@@ -1861,34 +2119,32 @@ static ssize_t dp_dsc_bits_per_pixel_read(struct file *f, char __user *buf,
 	if (!rd_buf)
 		return -ENOMEM;
 
-	rd_buf_ptr = rd_buf;
-
 	for (i = 0; i < MAX_PIPES; i++) {
 		pipe_ctx = &aconnector->dc_link->dc->current_state->res_ctx.pipe_ctx[i];
-			if (pipe_ctx && pipe_ctx->stream &&
-			    pipe_ctx->stream->link == aconnector->dc_link)
-				break;
+		if (pipe_ctx->stream &&
+		    pipe_ctx->stream->link == aconnector->dc_link &&
+		    pipe_ctx->stream->sink &&
+		    pipe_ctx->stream->sink == aconnector->dc_sink)
+			break;
 	}
 
-	if (!pipe_ctx)
-		return -ENXIO;
-
 	dsc = pipe_ctx->stream_res.dsc;
 	if (dsc)
 		dsc->funcs->dsc_read_state(dsc, &dsc_state);
 
-	snprintf(rd_buf_ptr, str_len,
+	snprintf(rd_buf, str_len,
 		"%d\n",
 		dsc_state.dsc_bits_per_pixel);
-	rd_buf_ptr += str_len;
 
 	while (size) {
 		if (*pos >= rd_buf_size)
 			break;
 
 		r = put_user(*(rd_buf + result), buf);
-		if (r)
+		if (r) {
+			kfree(rd_buf);
 			return r; /* r = -EFAULT */
+		}
 
 		buf += 1;
 		size -= 1;
@@ -1962,12 +2218,14 @@ static ssize_t dp_dsc_bits_per_pixel_write(struct file *f, const char __user *bu
 
 	for (i = 0; i < MAX_PIPES; i++) {
 		pipe_ctx = &aconnector->dc_link->dc->current_state->res_ctx.pipe_ctx[i];
-			if (pipe_ctx && pipe_ctx->stream &&
-			    pipe_ctx->stream->link == aconnector->dc_link)
-				break;
+		if (pipe_ctx->stream &&
+		    pipe_ctx->stream->link == aconnector->dc_link &&
+		    pipe_ctx->stream->sink &&
+		    pipe_ctx->stream->sink == aconnector->dc_sink)
+			break;
 	}
 
-	if (!pipe_ctx || !pipe_ctx->stream)
+	if (!pipe_ctx->stream)
 		goto done;
 
 	// Get CRTC state
@@ -2023,7 +2281,6 @@ static ssize_t dp_dsc_pic_width_read(struct file *f, char __user *buf,
 				    size_t size, loff_t *pos)
 {
 	char *rd_buf = NULL;
-	char *rd_buf_ptr = NULL;
 	struct amdgpu_dm_connector *aconnector = file_inode(f)->i_private;
 	struct display_stream_compressor *dsc;
 	struct dcn_dsc_state dsc_state = {0};
@@ -2037,34 +2294,32 @@ static ssize_t dp_dsc_pic_width_read(struct file *f, char __user *buf,
 	if (!rd_buf)
 		return -ENOMEM;
 
-	rd_buf_ptr = rd_buf;
-
 	for (i = 0; i < MAX_PIPES; i++) {
 		pipe_ctx = &aconnector->dc_link->dc->current_state->res_ctx.pipe_ctx[i];
-			if (pipe_ctx && pipe_ctx->stream &&
-			    pipe_ctx->stream->link == aconnector->dc_link)
-				break;
+		if (pipe_ctx->stream &&
+		    pipe_ctx->stream->link == aconnector->dc_link &&
+		    pipe_ctx->stream->sink &&
+		    pipe_ctx->stream->sink == aconnector->dc_sink)
+			break;
 	}
 
-	if (!pipe_ctx)
-		return -ENXIO;
-
 	dsc = pipe_ctx->stream_res.dsc;
 	if (dsc)
 		dsc->funcs->dsc_read_state(dsc, &dsc_state);
 
-	snprintf(rd_buf_ptr, str_len,
+	snprintf(rd_buf, str_len,
 		"%d\n",
 		dsc_state.dsc_pic_width);
-	rd_buf_ptr += str_len;
 
 	while (size) {
 		if (*pos >= rd_buf_size)
 			break;
 
 		r = put_user(*(rd_buf + result), buf);
-		if (r)
+		if (r) {
+			kfree(rd_buf);
 			return r; /* r = -EFAULT */
+		}
 
 		buf += 1;
 		size -= 1;
@@ -2080,7 +2335,6 @@ static ssize_t dp_dsc_pic_height_read(struct file *f, char __user *buf,
 				    size_t size, loff_t *pos)
 {
 	char *rd_buf = NULL;
-	char *rd_buf_ptr = NULL;
 	struct amdgpu_dm_connector *aconnector = file_inode(f)->i_private;
 	struct display_stream_compressor *dsc;
 	struct dcn_dsc_state dsc_state = {0};
@@ -2094,34 +2348,32 @@ static ssize_t dp_dsc_pic_height_read(struct file *f, char __user *buf,
 	if (!rd_buf)
 		return -ENOMEM;
 
-	rd_buf_ptr = rd_buf;
-
 	for (i = 0; i < MAX_PIPES; i++) {
 		pipe_ctx = &aconnector->dc_link->dc->current_state->res_ctx.pipe_ctx[i];
-			if (pipe_ctx && pipe_ctx->stream &&
-			    pipe_ctx->stream->link == aconnector->dc_link)
-				break;
+		if (pipe_ctx->stream &&
+		    pipe_ctx->stream->link == aconnector->dc_link &&
+		    pipe_ctx->stream->sink &&
+		    pipe_ctx->stream->sink == aconnector->dc_sink)
+			break;
 	}
 
-	if (!pipe_ctx)
-		return -ENXIO;
-
 	dsc = pipe_ctx->stream_res.dsc;
 	if (dsc)
 		dsc->funcs->dsc_read_state(dsc, &dsc_state);
 
-	snprintf(rd_buf_ptr, str_len,
+	snprintf(rd_buf, str_len,
 		"%d\n",
 		dsc_state.dsc_pic_height);
-	rd_buf_ptr += str_len;
 
 	while (size) {
 		if (*pos >= rd_buf_size)
 			break;
 
 		r = put_user(*(rd_buf + result), buf);
-		if (r)
+		if (r) {
+			kfree(rd_buf);
 			return r; /* r = -EFAULT */
+		}
 
 		buf += 1;
 		size -= 1;
@@ -2152,7 +2404,6 @@ static ssize_t dp_dsc_chunk_size_read(struct file *f, char __user *buf,
 				    size_t size, loff_t *pos)
 {
 	char *rd_buf = NULL;
-	char *rd_buf_ptr = NULL;
 	struct amdgpu_dm_connector *aconnector = file_inode(f)->i_private;
 	struct display_stream_compressor *dsc;
 	struct dcn_dsc_state dsc_state = {0};
@@ -2166,34 +2417,32 @@ static ssize_t dp_dsc_chunk_size_read(struct file *f, char __user *buf,
 	if (!rd_buf)
 		return -ENOMEM;
 
-	rd_buf_ptr = rd_buf;
-
 	for (i = 0; i < MAX_PIPES; i++) {
 		pipe_ctx = &aconnector->dc_link->dc->current_state->res_ctx.pipe_ctx[i];
-			if (pipe_ctx && pipe_ctx->stream &&
-			    pipe_ctx->stream->link == aconnector->dc_link)
-				break;
+		if (pipe_ctx->stream &&
+		    pipe_ctx->stream->link == aconnector->dc_link &&
+		    pipe_ctx->stream->sink &&
+		    pipe_ctx->stream->sink == aconnector->dc_sink)
+			break;
 	}
 
-	if (!pipe_ctx)
-		return -ENXIO;
-
 	dsc = pipe_ctx->stream_res.dsc;
 	if (dsc)
 		dsc->funcs->dsc_read_state(dsc, &dsc_state);
 
-	snprintf(rd_buf_ptr, str_len,
+	snprintf(rd_buf, str_len,
 		"%d\n",
 		dsc_state.dsc_chunk_size);
-	rd_buf_ptr += str_len;
 
 	while (size) {
 		if (*pos >= rd_buf_size)
 			break;
 
 		r = put_user(*(rd_buf + result), buf);
-		if (r)
+		if (r) {
+			kfree(rd_buf);
 			return r; /* r = -EFAULT */
+		}
 
 		buf += 1;
 		size -= 1;
@@ -2224,7 +2473,6 @@ static ssize_t dp_dsc_slice_bpg_offset_read(struct file *f, char __user *buf,
 				    size_t size, loff_t *pos)
 {
 	char *rd_buf = NULL;
-	char *rd_buf_ptr = NULL;
 	struct amdgpu_dm_connector *aconnector = file_inode(f)->i_private;
 	struct display_stream_compressor *dsc;
 	struct dcn_dsc_state dsc_state = {0};
@@ -2238,34 +2486,32 @@ static ssize_t dp_dsc_slice_bpg_offset_read(struct file *f, char __user *buf,
 	if (!rd_buf)
 		return -ENOMEM;
 
-	rd_buf_ptr = rd_buf;
-
 	for (i = 0; i < MAX_PIPES; i++) {
 		pipe_ctx = &aconnector->dc_link->dc->current_state->res_ctx.pipe_ctx[i];
-			if (pipe_ctx && pipe_ctx->stream &&
-			    pipe_ctx->stream->link == aconnector->dc_link)
-				break;
+		if (pipe_ctx->stream &&
+		    pipe_ctx->stream->link == aconnector->dc_link &&
+		    pipe_ctx->stream->sink &&
+		    pipe_ctx->stream->sink == aconnector->dc_sink)
+			break;
 	}
 
-	if (!pipe_ctx)
-		return -ENXIO;
-
 	dsc = pipe_ctx->stream_res.dsc;
 	if (dsc)
 		dsc->funcs->dsc_read_state(dsc, &dsc_state);
 
-	snprintf(rd_buf_ptr, str_len,
+	snprintf(rd_buf, str_len,
 		"%d\n",
 		dsc_state.dsc_slice_bpg_offset);
-	rd_buf_ptr += str_len;
 
 	while (size) {
 		if (*pos >= rd_buf_size)
 			break;
 
 		r = put_user(*(rd_buf + result), buf);
-		if (r)
+		if (r) {
+			kfree(rd_buf);
 			return r; /* r = -EFAULT */
+		}
 
 		buf += 1;
 		size -= 1;
@@ -2421,6 +2667,49 @@ unlock:
 }
 
 /*
+ * IPS status.  Read only.
+ *
+ * Example usage: cat /sys/kernel/debug/dri/0/amdgpu_dm_ips_status
+ */
+static int ips_status_show(struct seq_file *m, void *unused)
+{
+	struct amdgpu_device *adev = m->private;
+	struct dc *dc = adev->dm.dc;
+	struct dc_dmub_srv *dc_dmub_srv;
+
+	seq_printf(m, "IPS config: %d\n", dc->config.disable_ips);
+	seq_printf(m, "Idle optimization: %d\n", dc->idle_optimizations_allowed);
+
+	if (adev->dm.idle_workqueue) {
+		seq_printf(m, "Idle workqueue - enabled: %d\n", adev->dm.idle_workqueue->enable);
+		seq_printf(m, "Idle workqueue - running: %d\n", adev->dm.idle_workqueue->running);
+	}
+
+	dc_dmub_srv = dc->ctx->dmub_srv;
+	if (dc_dmub_srv && dc_dmub_srv->dmub) {
+		uint32_t rcg_count, ips1_count, ips2_count;
+		volatile const struct dmub_shared_state_ips_fw *ips_fw =
+			&dc_dmub_srv->dmub->shared_state[DMUB_SHARED_SHARE_FEATURE__IPS_FW].data.ips_fw;
+		rcg_count = ips_fw->rcg_entry_count;
+		ips1_count = ips_fw->ips1_entry_count;
+		ips2_count = ips_fw->ips2_entry_count;
+		seq_printf(m, "entry counts: rcg=%u ips1=%u ips2=%u\n",
+			   rcg_count,
+			   ips1_count,
+			   ips2_count);
+		rcg_count = ips_fw->rcg_exit_count;
+		ips1_count = ips_fw->ips1_exit_count;
+		ips2_count = ips_fw->ips2_exit_count;
+		seq_printf(m, "exit counts: rcg=%u ips1=%u ips2=%u",
+			   rcg_count,
+			   ips1_count,
+			   ips2_count);
+		seq_puts(m, "\n");
+	}
+	return 0;
+}
+
+/*
  * Backlight at this moment.  Read only.
  * As written to display, taking ABM and backlight lut into account.
  * Ranges from 0x0 to 0x10000 (= 100% PWM)
@@ -2458,15 +2747,185 @@ static int target_backlight_show(struct seq_file *m, void *unused)
 	return 0;
 }
 
+/*
+ * function description: Determine if the connector is mst connector
+ *
+ * This function helps to determine whether a connector is a mst connector.
+ * - "root" stands for the root connector of the topology
+ * - "branch" stands for branch device of the topology
+ * - "end" stands for leaf node connector of the topology
+ * - "no" stands for the connector is not a device of a mst topology
+ * Access it with the following command:
+ *
+ *	cat /sys/kernel/debug/dri/0/DP-X/is_mst_connector
+ *
+ */
+static int dp_is_mst_connector_show(struct seq_file *m, void *unused)
+{
+	struct drm_connector *connector = m->private;
+	struct amdgpu_dm_connector *aconnector = to_amdgpu_dm_connector(connector);
+	struct drm_dp_mst_topology_mgr *mgr = NULL;
+	struct drm_dp_mst_port *port = NULL;
+	char *role = NULL;
+
+	mutex_lock(&aconnector->hpd_lock);
+
+	if (aconnector->mst_mgr.mst_state) {
+		role = "root";
+	} else if (aconnector->mst_root &&
+		aconnector->mst_root->mst_mgr.mst_state) {
+
+		role = "end";
+
+		mgr = &aconnector->mst_root->mst_mgr;
+		port = aconnector->mst_output_port;
+
+		drm_modeset_lock(&mgr->base.lock, NULL);
+		if (port->pdt == DP_PEER_DEVICE_MST_BRANCHING &&
+			port->mcs)
+			role = "branch";
+		drm_modeset_unlock(&mgr->base.lock);
+
+	} else {
+		role = "no";
+	}
+
+	seq_printf(m, "%s\n", role);
+
+	mutex_unlock(&aconnector->hpd_lock);
+
+	return 0;
+}
+
+/*
+ * function description: Read out the mst progress status
+ *
+ * This function helps to determine the mst progress status of
+ * a mst connector.
+ *
+ * Access it with the following command:
+ *
+ *	cat /sys/kernel/debug/dri/0/DP-X/mst_progress_status
+ *
+ */
+static int dp_mst_progress_status_show(struct seq_file *m, void *unused)
+{
+	struct drm_connector *connector = m->private;
+	struct amdgpu_dm_connector *aconnector = to_amdgpu_dm_connector(connector);
+	struct amdgpu_device *adev = drm_to_adev(connector->dev);
+	int i;
+
+	mutex_lock(&aconnector->hpd_lock);
+	mutex_lock(&adev->dm.dc_lock);
+
+	if (aconnector->mst_status == MST_STATUS_DEFAULT) {
+		seq_puts(m, "disabled\n");
+	} else {
+		for (i = 0; i < sizeof(mst_progress_status)/sizeof(char *); i++)
+			seq_printf(m, "%s:%s\n",
+				mst_progress_status[i],
+				aconnector->mst_status & BIT(i) ? "done" : "not_done");
+	}
+
+	mutex_unlock(&adev->dm.dc_lock);
+	mutex_unlock(&aconnector->hpd_lock);
+
+	return 0;
+}
+
+/*
+ * Reports whether the connected display is a USB4 DPIA tunneled display
+ * Example usage: cat /sys/kernel/debug/dri/0/DP-8/is_dpia_link
+ */
+static int is_dpia_link_show(struct seq_file *m, void *data)
+{
+	struct drm_connector *connector = m->private;
+	struct amdgpu_dm_connector *aconnector = to_amdgpu_dm_connector(connector);
+	struct dc_link *link = aconnector->dc_link;
+
+	if (connector->status != connector_status_connected)
+		return -ENODEV;
+
+	seq_printf(m, "%s\n", (link->ep_type == DISPLAY_ENDPOINT_USB4_DPIA) ? "yes" :
+				(link->ep_type == DISPLAY_ENDPOINT_PHY) ? "no" : "unknown");
+
+	return 0;
+}
+
+/**
+ * hdmi_cec_state_show - Read out the HDMI-CEC feature status
+ * @m: sequence file.
+ * @data: unused.
+ *
+ * Return 0 on success
+ */
+static int hdmi_cec_state_show(struct seq_file *m, void *data)
+{
+	struct drm_connector *connector = m->private;
+	struct amdgpu_dm_connector *aconnector = to_amdgpu_dm_connector(connector);
+
+	seq_printf(m, "%s:%d\n", connector->name, connector->base.id);
+	seq_printf(m, "HDMI-CEC status: %d\n", aconnector->notifier ? 1 : 0);
+
+	return 0;
+}
+
+/**
+ * hdmi_cec_state_write - Enable/Disable HDMI-CEC feature from driver side
+ * @f: file structure.
+ * @buf: userspace buffer. set to '1' to enable; '0' to disable cec feature.
+ * @size: size of buffer from userpsace.
+ * @pos: unused.
+ *
+ * Return size on success, error code on failure
+ */
+static ssize_t hdmi_cec_state_write(struct file *f, const char __user *buf,
+				    size_t size, loff_t *pos)
+{
+	int ret;
+	bool enable;
+	struct amdgpu_dm_connector *aconnector = file_inode(f)->i_private;
+	struct drm_device *ddev = aconnector->base.dev;
+
+	if (size == 0)
+		return -EINVAL;
+
+	ret = kstrtobool_from_user(buf, size, &enable);
+	if (ret) {
+		drm_dbg_driver(ddev, "invalid user data !\n");
+		return ret;
+	}
+
+	if (enable) {
+		if (aconnector->notifier)
+			return -EINVAL;
+		ret = amdgpu_dm_initialize_hdmi_connector(aconnector);
+		if (ret)
+			return ret;
+		hdmi_cec_set_edid(aconnector);
+	} else {
+		if (!aconnector->notifier)
+			return -EINVAL;
+		cec_notifier_conn_unregister(aconnector->notifier);
+		aconnector->notifier = NULL;
+	}
+
+	return size;
+}
+
 DEFINE_SHOW_ATTRIBUTE(dp_dsc_fec_support);
 DEFINE_SHOW_ATTRIBUTE(dmub_fw_state);
 DEFINE_SHOW_ATTRIBUTE(dmub_tracebuffer);
-DEFINE_SHOW_ATTRIBUTE(output_bpc);
 DEFINE_SHOW_ATTRIBUTE(dp_lttpr_status);
-#ifdef CONFIG_DRM_AMD_DC_HDCP
 DEFINE_SHOW_ATTRIBUTE(hdcp_sink_capability);
-#endif
 DEFINE_SHOW_ATTRIBUTE(internal_display);
+DEFINE_SHOW_ATTRIBUTE(odm_combine_segments);
+DEFINE_SHOW_ATTRIBUTE(replay_capability);
+DEFINE_SHOW_ATTRIBUTE(psr_capability);
+DEFINE_SHOW_ATTRIBUTE(dp_is_mst_connector);
+DEFINE_SHOW_ATTRIBUTE(dp_mst_progress_status);
+DEFINE_SHOW_ATTRIBUTE(is_dpia_link);
+DEFINE_SHOW_STORE_ATTRIBUTE(hdmi_cec_state);
 
 static const struct file_operations dp_dsc_clock_en_debugfs_fops = {
 	.owner = THIS_MODULE,
@@ -2552,25 +3011,6 @@ static const struct file_operations sdp_message_fops = {
 	.llseek = default_llseek
 };
 
-static const struct file_operations dp_dpcd_address_debugfs_fops = {
-	.owner = THIS_MODULE,
-	.write = dp_dpcd_address_write,
-	.llseek = default_llseek
-};
-
-static const struct file_operations dp_dpcd_size_debugfs_fops = {
-	.owner = THIS_MODULE,
-	.write = dp_dpcd_size_write,
-	.llseek = default_llseek
-};
-
-static const struct file_operations dp_dpcd_data_debugfs_fops = {
-	.owner = THIS_MODULE,
-	.read = dp_dpcd_data_read,
-	.write = dp_dpcd_data_write,
-	.llseek = default_llseek
-};
-
 static const struct file_operations dp_max_bpc_debugfs_fops = {
 	.owner = THIS_MODULE,
 	.read = dp_max_bpc_read,
@@ -2584,6 +3024,12 @@ static const struct file_operations dp_dsc_disable_passthrough_debugfs_fops = {
 	.llseek = default_llseek
 };
 
+static const struct file_operations dp_mst_link_settings_debugfs_fops = {
+	.owner = THIS_MODULE,
+	.write = dp_mst_link_setting,
+	.llseek = default_llseek
+};
+
 static const struct {
 	char *name;
 	const struct file_operations *fops;
@@ -2592,13 +3038,8 @@ static const struct {
 		{"phy_settings", &dp_phy_settings_debugfs_fop},
 		{"lttpr_status", &dp_lttpr_status_fops},
 		{"test_pattern", &dp_phy_test_pattern_fops},
-#ifdef CONFIG_DRM_AMD_DC_HDCP
 		{"hdcp_sink_capability", &hdcp_sink_capability_fops},
-#endif
 		{"sdp_message", &sdp_message_fops},
-		{"aux_dpcd_address", &dp_dpcd_address_debugfs_fops},
-		{"aux_dpcd_size", &dp_dpcd_size_debugfs_fops},
-		{"aux_dpcd_data", &dp_dpcd_data_debugfs_fops},
 		{"dsc_clock_en", &dp_dsc_clock_en_debugfs_fops},
 		{"dsc_slice_width", &dp_dsc_slice_width_debugfs_fops},
 		{"dsc_slice_height", &dp_dsc_slice_height_debugfs_fops},
@@ -2610,16 +3051,20 @@ static const struct {
 		{"dp_dsc_fec_support", &dp_dsc_fec_support_fops},
 		{"max_bpc", &dp_max_bpc_debugfs_fops},
 		{"dsc_disable_passthrough", &dp_dsc_disable_passthrough_debugfs_fops},
+		{"is_mst_connector", &dp_is_mst_connector_fops},
+		{"mst_progress_status", &dp_mst_progress_status_fops},
+		{"is_dpia_link", &is_dpia_link_fops},
+		{"mst_link_settings", &dp_mst_link_settings_debugfs_fops}
 };
 
-#ifdef CONFIG_DRM_AMD_DC_HDCP
 static const struct {
 	char *name;
 	const struct file_operations *fops;
 } hdmi_debugfs_entries[] = {
-		{"hdcp_sink_capability", &hdcp_sink_capability_fops}
+		{"hdcp_sink_capability", &hdcp_sink_capability_fops},
+		{"hdmi_cec_state", &hdmi_cec_state_fops}
 };
-#endif
+
 /*
  * Force YUV420 output if available from the given mode
  */
@@ -2648,6 +3093,51 @@ DEFINE_DEBUGFS_ATTRIBUTE(force_yuv420_output_fops, force_yuv420_output_get,
 			 force_yuv420_output_set, "%llu\n");
 
 /*
+ *  Read Replay state
+ */
+static int replay_get_state(void *data, u64 *val)
+{
+	struct amdgpu_dm_connector *connector = data;
+	struct dc_link *link = connector->dc_link;
+	uint64_t state = REPLAY_STATE_INVALID;
+
+	dc_link_get_replay_state(link, &state);
+
+	*val = state;
+
+	return 0;
+}
+
+/*
+ *  Start / Stop capture Replay residency
+ */
+static int replay_set_residency(void *data, u64 val)
+{
+	struct amdgpu_dm_connector *connector = data;
+	struct dc_link *link = connector->dc_link;
+	bool is_start = (val != 0);
+	u32 residency = 0;
+
+	link->dc->link_srv->edp_replay_residency(link, &residency, is_start, PR_RESIDENCY_MODE_PHY);
+	return 0;
+}
+
+/*
+ *  Read Replay residency
+ */
+static int replay_get_residency(void *data, u64 *val)
+{
+	struct amdgpu_dm_connector *connector = data;
+	struct dc_link *link = connector->dc_link;
+	u32 residency = 0;
+
+	link->dc->link_srv->edp_replay_residency(link, &residency, false, PR_RESIDENCY_MODE_PHY);
+	*val = (u64)residency;
+
+	return 0;
+}
+
+/*
  *  Read PSR state
  */
 static int psr_get(void *data, u64 *val)
@@ -2664,6 +3154,174 @@ static int psr_get(void *data, u64 *val)
 }
 
 /*
+ *  Read PSR state residency
+ */
+static int psr_read_residency(void *data, u64 *val)
+{
+	struct amdgpu_dm_connector *connector = data;
+	struct dc_link *link = connector->dc_link;
+	u32 residency = 0;
+
+	link->dc->link_srv->edp_get_psr_residency(link, &residency, PSR_RESIDENCY_MODE_PHY);
+
+	*val = (u64)residency;
+
+	return 0;
+}
+
+/* read allow_edp_hotplug_detection */
+static int allow_edp_hotplug_detection_get(void *data, u64 *val)
+{
+	struct amdgpu_dm_connector *aconnector = data;
+	struct drm_connector *connector = &aconnector->base;
+	struct drm_device *dev = connector->dev;
+	struct amdgpu_device *adev = drm_to_adev(dev);
+
+	*val = adev->dm.dc->config.allow_edp_hotplug_detection;
+
+	return 0;
+}
+
+/* set allow_edp_hotplug_detection */
+static int allow_edp_hotplug_detection_set(void *data, u64 val)
+{
+	struct amdgpu_dm_connector *aconnector = data;
+	struct drm_connector *connector = &aconnector->base;
+	struct drm_device *dev = connector->dev;
+	struct amdgpu_device *adev = drm_to_adev(dev);
+
+	adev->dm.dc->config.allow_edp_hotplug_detection = (uint32_t) val;
+
+	return 0;
+}
+
+/* check if kernel disallow eDP enter psr state
+ * cat /sys/kernel/debug/dri/0/eDP-X/disallow_edp_enter_psr
+ * 0: allow edp enter psr; 1: disallow
+ */
+static int disallow_edp_enter_psr_get(void *data, u64 *val)
+{
+	struct amdgpu_dm_connector *aconnector = data;
+
+	*val = (u64) aconnector->disallow_edp_enter_psr;
+	return 0;
+}
+
+/* set kernel disallow eDP enter psr state
+ * echo 0x0 /sys/kernel/debug/dri/0/eDP-X/disallow_edp_enter_psr
+ * 0: allow edp enter psr; 1: disallow
+ *
+ * usage: test app read crc from PSR eDP rx.
+ *
+ * during kernel boot up, kernel write dpcd 0x170 = 5.
+ * this notify eDP rx psr enable and let rx check crc.
+ * rx fw will start checking crc for rx internal logic.
+ * crc read count within dpcd 0x246 is not updated and
+ * value is 0. when eDP tx driver wants to read rx crc
+ * from dpcd 0x246, 0x270, read count 0 lead tx driver
+ * timeout.
+ *
+ * to avoid this, we add this debugfs to let test app to disbable
+ * rx crc checking for rx internal logic. then test app can read
+ * non-zero crc read count.
+ *
+ * expected app sequence is as below:
+ * 1. disable eDP PHY and notify eDP rx with dpcd 0x600 = 2.
+ * 2. echo 0x1 /sys/kernel/debug/dri/0/eDP-X/disallow_edp_enter_psr
+ * 3. enable eDP PHY and notify eDP rx with dpcd 0x600 = 1 but
+ *    without dpcd 0x170 = 5.
+ * 4. read crc from rx dpcd 0x270, 0x246, etc.
+ * 5. echo 0x0 /sys/kernel/debug/dri/0/eDP-X/disallow_edp_enter_psr.
+ *    this will let eDP back to normal with psr setup dpcd 0x170 = 5.
+ */
+static int disallow_edp_enter_psr_set(void *data, u64 val)
+{
+	struct amdgpu_dm_connector *aconnector = data;
+
+	aconnector->disallow_edp_enter_psr = val ? true : false;
+	return 0;
+}
+
+static int dmub_trace_mask_set(void *data, u64 val)
+{
+	struct amdgpu_device *adev = data;
+	struct dmub_srv *srv = adev->dm.dc->ctx->dmub_srv->dmub;
+	enum dmub_gpint_command cmd;
+	u64 mask = 0xffff;
+	u8 shift = 0;
+	u32 res;
+	int i;
+
+	if (!srv->fw_version)
+		return -EINVAL;
+
+	for (i = 0;  i < 4; i++) {
+		res = (val & mask) >> shift;
+
+		switch (i) {
+		case 0:
+			cmd = DMUB_GPINT__SET_TRACE_BUFFER_MASK_WORD0;
+			break;
+		case 1:
+			cmd = DMUB_GPINT__SET_TRACE_BUFFER_MASK_WORD1;
+			break;
+		case 2:
+			cmd = DMUB_GPINT__SET_TRACE_BUFFER_MASK_WORD2;
+			break;
+		case 3:
+			cmd = DMUB_GPINT__SET_TRACE_BUFFER_MASK_WORD3;
+			break;
+		}
+
+		if (!dc_wake_and_execute_gpint(adev->dm.dc->ctx, cmd, res, NULL, DM_DMUB_WAIT_TYPE_WAIT))
+			return -EIO;
+
+		usleep_range(100, 1000);
+
+		mask <<= 16;
+		shift += 16;
+	}
+
+	return 0;
+}
+
+static int dmub_trace_mask_show(void *data, u64 *val)
+{
+	enum dmub_gpint_command cmd = DMUB_GPINT__GET_TRACE_BUFFER_MASK_WORD0;
+	struct amdgpu_device *adev = data;
+	struct dmub_srv *srv = adev->dm.dc->ctx->dmub_srv->dmub;
+	u8 shift = 0;
+	u64 raw = 0;
+	u64 res = 0;
+	int i = 0;
+
+	if (!srv->fw_version)
+		return -EINVAL;
+
+	while (i < 4) {
+		uint32_t response;
+
+		if (!dc_wake_and_execute_gpint(adev->dm.dc->ctx, cmd, 0, &response, DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY))
+			return -EIO;
+
+		raw = response;
+		usleep_range(100, 1000);
+
+		cmd++;
+		res |= (raw << shift);
+		shift += 16;
+		i++;
+	}
+
+	*val = res;
+
+	return 0;
+}
+
+DEFINE_DEBUGFS_ATTRIBUTE(dmub_trace_mask_fops, dmub_trace_mask_show,
+			 dmub_trace_mask_set, "0x%llx\n");
+
+/*
  * Set dmcub trace event IRQ enable or disable.
  * Usage to enable dmcub trace event IRQ: echo 1 > /sys/kernel/debug/dri/0/amdgpu_dm_dmcub_trace_event_en
  * Usage to disable dmcub trace event IRQ: echo 0 > /sys/kernel/debug/dri/0/amdgpu_dm_dmcub_trace_event_en
@@ -2697,19 +3355,167 @@ static int dmcub_trace_event_state_get(void *data, u64 *val)
 DEFINE_DEBUGFS_ATTRIBUTE(dmcub_trace_event_state_fops, dmcub_trace_event_state_get,
 			 dmcub_trace_event_state_set, "%llu\n");
 
+DEFINE_DEBUGFS_ATTRIBUTE(replay_state_fops, replay_get_state, NULL, "%llu\n");
+DEFINE_DEBUGFS_ATTRIBUTE(replay_residency_fops, replay_get_residency, replay_set_residency,
+			 "%llu\n");
 DEFINE_DEBUGFS_ATTRIBUTE(psr_fops, psr_get, NULL, "%llu\n");
+DEFINE_DEBUGFS_ATTRIBUTE(psr_residency_fops, psr_read_residency, NULL,
+			 "%llu\n");
+
+DEFINE_DEBUGFS_ATTRIBUTE(allow_edp_hotplug_detection_fops,
+			allow_edp_hotplug_detection_get,
+			allow_edp_hotplug_detection_set, "%llu\n");
+
+DEFINE_DEBUGFS_ATTRIBUTE(disallow_edp_enter_psr_fops,
+			disallow_edp_enter_psr_get,
+			disallow_edp_enter_psr_set, "%llu\n");
 
 DEFINE_SHOW_ATTRIBUTE(current_backlight);
 DEFINE_SHOW_ATTRIBUTE(target_backlight);
+DEFINE_SHOW_ATTRIBUTE(ips_status);
 
 static const struct {
 	char *name;
 	const struct file_operations *fops;
 } connector_debugfs_entries[] = {
 		{"force_yuv420_output", &force_yuv420_output_fops},
-		{"output_bpc", &output_bpc_fops},
 		{"trigger_hotplug", &trigger_hotplug_debugfs_fops},
-		{"internal_display", &internal_display_fops}
+		{"internal_display", &internal_display_fops},
+		{"odm_combine_segments", &odm_combine_segments_fops}
+};
+
+/*
+ * Returns supported customized link rates by this eDP panel.
+ * Example usage: cat /sys/kernel/debug/dri/0/eDP-x/ilr_setting
+ */
+static int edp_ilr_show(struct seq_file *m, void *unused)
+{
+	struct amdgpu_dm_connector *aconnector = to_amdgpu_dm_connector(m->private);
+	struct dc_link *link = aconnector->dc_link;
+	uint8_t supported_link_rates[16];
+	uint32_t link_rate_in_khz;
+	uint32_t entry = 0;
+	uint8_t dpcd_rev;
+
+	memset(supported_link_rates, 0, sizeof(supported_link_rates));
+	dm_helpers_dp_read_dpcd(link->ctx, link, DP_SUPPORTED_LINK_RATES,
+		supported_link_rates, sizeof(supported_link_rates));
+
+	dpcd_rev = link->dpcd_caps.dpcd_rev.raw;
+
+	if (dpcd_rev >= DP_DPCD_REV_13 &&
+		(supported_link_rates[entry+1] != 0 || supported_link_rates[entry] != 0)) {
+
+		for (entry = 0; entry < 16; entry += 2) {
+			link_rate_in_khz = (supported_link_rates[entry+1] * 0x100 +
+										supported_link_rates[entry]) * 200;
+			seq_printf(m, "[%d] %d kHz\n", entry/2, link_rate_in_khz);
+		}
+	} else {
+		seq_puts(m, "ILR is not supported by this eDP panel.\n");
+	}
+
+	return 0;
+}
+
+/*
+ * Set supported customized link rate to eDP panel.
+ *
+ * echo <lane_count>  <link_rate option> > ilr_setting
+ *
+ * for example, supported ILR : [0] 1620000 kHz [1] 2160000 kHz [2] 2430000 kHz ...
+ * echo 4 1 > /sys/kernel/debug/dri/0/eDP-x/ilr_setting
+ * to set 4 lanes and 2.16 GHz
+ */
+static ssize_t edp_ilr_write(struct file *f, const char __user *buf,
+				 size_t size, loff_t *pos)
+{
+	struct amdgpu_dm_connector *connector = file_inode(f)->i_private;
+	struct dc_link *link = connector->dc_link;
+	struct amdgpu_device *adev = drm_to_adev(connector->base.dev);
+	struct dc *dc = (struct dc *)link->dc;
+	struct dc_link_settings prefer_link_settings;
+	char *wr_buf = NULL;
+	const uint32_t wr_buf_size = 40;
+	/* 0: lane_count; 1: link_rate */
+	int max_param_num = 2;
+	uint8_t param_nums = 0;
+	long param[2];
+	bool valid_input = true;
+
+	if (size == 0)
+		return -EINVAL;
+
+	wr_buf = kcalloc(wr_buf_size, sizeof(char), GFP_KERNEL);
+	if (!wr_buf)
+		return -ENOMEM;
+
+	if (parse_write_buffer_into_params(wr_buf, wr_buf_size,
+					   (long *)param, buf,
+					   max_param_num,
+					   &param_nums)) {
+		kfree(wr_buf);
+		return -EINVAL;
+	}
+
+	if (param_nums <= 0) {
+		kfree(wr_buf);
+		return -EINVAL;
+	}
+
+	switch (param[0]) {
+	case LANE_COUNT_ONE:
+	case LANE_COUNT_TWO:
+	case LANE_COUNT_FOUR:
+		break;
+	default:
+		valid_input = false;
+		break;
+	}
+
+	if (param[1] >= link->dpcd_caps.edp_supported_link_rates_count)
+		valid_input = false;
+
+	if (!valid_input) {
+		kfree(wr_buf);
+		DRM_DEBUG_DRIVER("Invalid Input value. No HW will be programmed\n");
+		prefer_link_settings.use_link_rate_set = false;
+		mutex_lock(&adev->dm.dc_lock);
+		dc_link_set_preferred_training_settings(dc, NULL, NULL, link, false);
+		mutex_unlock(&adev->dm.dc_lock);
+		return size;
+	}
+
+	/* save user force lane_count, link_rate to preferred settings
+	 * spread spectrum will not be changed
+	 */
+	prefer_link_settings.link_spread = link->cur_link_settings.link_spread;
+	prefer_link_settings.lane_count = param[0];
+	prefer_link_settings.use_link_rate_set = true;
+	prefer_link_settings.link_rate_set = param[1];
+	prefer_link_settings.link_rate = link->dpcd_caps.edp_supported_link_rates[param[1]];
+
+	mutex_lock(&adev->dm.dc_lock);
+	dc_link_set_preferred_training_settings(dc, &prefer_link_settings,
+						NULL, link, false);
+	mutex_unlock(&adev->dm.dc_lock);
+
+	kfree(wr_buf);
+	return size;
+}
+
+static int edp_ilr_open(struct inode *inode, struct file *file)
+{
+	return single_open(file, edp_ilr_show, inode->i_private);
+}
+
+static const struct file_operations edp_ilr_debugfs_fops = {
+	.owner = THIS_MODULE,
+	.open = edp_ilr_open,
+	.read = seq_read,
+	.llseek = seq_lseek,
+	.release = single_release,
+	.write = edp_ilr_write
 };
 
 void connector_debugfs_init(struct amdgpu_dm_connector *connector)
@@ -2726,11 +3532,25 @@ void connector_debugfs_init(struct amdgpu_dm_connector *connector)
 		}
 	}
 	if (connector->base.connector_type == DRM_MODE_CONNECTOR_eDP) {
+		debugfs_create_file("replay_capability", 0444, dir, connector,
+					&replay_capability_fops);
+		debugfs_create_file("replay_state", 0444, dir, connector, &replay_state_fops);
+		debugfs_create_file_unsafe("replay_residency", 0444, dir,
+					   connector, &replay_residency_fops);
+		debugfs_create_file_unsafe("psr_capability", 0444, dir, connector, &psr_capability_fops);
 		debugfs_create_file_unsafe("psr_state", 0444, dir, connector, &psr_fops);
+		debugfs_create_file_unsafe("psr_residency", 0444, dir,
+					   connector, &psr_residency_fops);
 		debugfs_create_file("amdgpu_current_backlight_pwm", 0444, dir, connector,
 				    &current_backlight_fops);
 		debugfs_create_file("amdgpu_target_backlight_pwm", 0444, dir, connector,
 				    &target_backlight_fops);
+		debugfs_create_file("ilr_setting", 0644, dir, connector,
+					&edp_ilr_debugfs_fops);
+		debugfs_create_file("allow_edp_hotplug_detection", 0644, dir, connector,
+					&allow_edp_hotplug_detection_fops);
+		debugfs_create_file("disallow_edp_enter_psr", 0644, dir, connector,
+					&disallow_edp_enter_psr_fops);
 	}
 
 	for (i = 0; i < ARRAY_SIZE(connector_debugfs_entries); i++) {
@@ -2739,10 +3559,6 @@ void connector_debugfs_init(struct amdgpu_dm_connector *connector)
 				    connector_debugfs_entries[i].fops);
 	}
 
-	connector->debugfs_dpcd_address = 0;
-	connector->debugfs_dpcd_size = 0;
-
-#ifdef CONFIG_DRM_AMD_DC_HDCP
 	if (connector->base.connector_type == DRM_MODE_CONNECTOR_HDMIA) {
 		for (i = 0; i < ARRAY_SIZE(hdmi_debugfs_entries); i++) {
 			debugfs_create_file(hdmi_debugfs_entries[i].name,
@@ -2750,7 +3566,6 @@ void connector_debugfs_init(struct amdgpu_dm_connector *connector)
 					    hdmi_debugfs_entries[i].fops);
 		}
 	}
-#endif
 }
 
 #ifdef CONFIG_DRM_AMD_SECURE_DISPLAY
@@ -2764,8 +3579,8 @@ static int crc_win_x_start_set(void *data, u64 val)
 	struct amdgpu_crtc *acrtc = to_amdgpu_crtc(crtc);
 
 	spin_lock_irq(&drm_dev->event_lock);
-	acrtc->dm_irq_params.crc_window.x_start = (uint16_t) val;
-	acrtc->dm_irq_params.crc_window.update_win = false;
+	acrtc->dm_irq_params.window_param[0].x_start = (uint16_t) val;
+	acrtc->dm_irq_params.window_param[0].update_win = false;
 	spin_unlock_irq(&drm_dev->event_lock);
 
 	return 0;
@@ -2781,7 +3596,7 @@ static int crc_win_x_start_get(void *data, u64 *val)
 	struct amdgpu_crtc *acrtc = to_amdgpu_crtc(crtc);
 
 	spin_lock_irq(&drm_dev->event_lock);
-	*val = acrtc->dm_irq_params.crc_window.x_start;
+	*val = acrtc->dm_irq_params.window_param[0].x_start;
 	spin_unlock_irq(&drm_dev->event_lock);
 
 	return 0;
@@ -2801,8 +3616,8 @@ static int crc_win_y_start_set(void *data, u64 val)
 	struct amdgpu_crtc *acrtc = to_amdgpu_crtc(crtc);
 
 	spin_lock_irq(&drm_dev->event_lock);
-	acrtc->dm_irq_params.crc_window.y_start = (uint16_t) val;
-	acrtc->dm_irq_params.crc_window.update_win = false;
+	acrtc->dm_irq_params.window_param[0].y_start = (uint16_t) val;
+	acrtc->dm_irq_params.window_param[0].update_win = false;
 	spin_unlock_irq(&drm_dev->event_lock);
 
 	return 0;
@@ -2818,7 +3633,7 @@ static int crc_win_y_start_get(void *data, u64 *val)
 	struct amdgpu_crtc *acrtc = to_amdgpu_crtc(crtc);
 
 	spin_lock_irq(&drm_dev->event_lock);
-	*val = acrtc->dm_irq_params.crc_window.y_start;
+	*val = acrtc->dm_irq_params.window_param[0].y_start;
 	spin_unlock_irq(&drm_dev->event_lock);
 
 	return 0;
@@ -2837,8 +3652,8 @@ static int crc_win_x_end_set(void *data, u64 val)
 	struct amdgpu_crtc *acrtc = to_amdgpu_crtc(crtc);
 
 	spin_lock_irq(&drm_dev->event_lock);
-	acrtc->dm_irq_params.crc_window.x_end = (uint16_t) val;
-	acrtc->dm_irq_params.crc_window.update_win = false;
+	acrtc->dm_irq_params.window_param[0].x_end = (uint16_t) val;
+	acrtc->dm_irq_params.window_param[0].update_win = false;
 	spin_unlock_irq(&drm_dev->event_lock);
 
 	return 0;
@@ -2854,7 +3669,7 @@ static int crc_win_x_end_get(void *data, u64 *val)
 	struct amdgpu_crtc *acrtc = to_amdgpu_crtc(crtc);
 
 	spin_lock_irq(&drm_dev->event_lock);
-	*val = acrtc->dm_irq_params.crc_window.x_end;
+	*val = acrtc->dm_irq_params.window_param[0].x_end;
 	spin_unlock_irq(&drm_dev->event_lock);
 
 	return 0;
@@ -2873,8 +3688,8 @@ static int crc_win_y_end_set(void *data, u64 val)
 	struct amdgpu_crtc *acrtc = to_amdgpu_crtc(crtc);
 
 	spin_lock_irq(&drm_dev->event_lock);
-	acrtc->dm_irq_params.crc_window.y_end = (uint16_t) val;
-	acrtc->dm_irq_params.crc_window.update_win = false;
+	acrtc->dm_irq_params.window_param[0].y_end = (uint16_t) val;
+	acrtc->dm_irq_params.window_param[0].update_win = false;
 	spin_unlock_irq(&drm_dev->event_lock);
 
 	return 0;
@@ -2890,7 +3705,7 @@ static int crc_win_y_end_get(void *data, u64 *val)
 	struct amdgpu_crtc *acrtc = to_amdgpu_crtc(crtc);
 
 	spin_lock_irq(&drm_dev->event_lock);
-	*val = acrtc->dm_irq_params.crc_window.y_end;
+	*val = acrtc->dm_irq_params.window_param[0].y_end;
 	spin_unlock_irq(&drm_dev->event_lock);
 
 	return 0;
@@ -2903,38 +3718,27 @@ DEFINE_DEBUGFS_ATTRIBUTE(crc_win_y_end_fops, crc_win_y_end_get,
  */
 static int crc_win_update_set(void *data, u64 val)
 {
-	struct drm_crtc *new_crtc = data;
-	struct drm_crtc *old_crtc = NULL;
-	struct amdgpu_crtc *new_acrtc, *old_acrtc;
-	struct amdgpu_device *adev = drm_to_adev(new_crtc->dev);
-	struct crc_rd_work *crc_rd_wrk = adev->dm.crc_rd_wrk;
+	struct drm_crtc *crtc = data;
+	struct amdgpu_crtc *acrtc;
+	struct amdgpu_device *adev = drm_to_adev(crtc->dev);
 
 	if (val) {
+		acrtc = to_amdgpu_crtc(crtc);
+		mutex_lock(&adev->dm.dc_lock);
+		/* PSR may write to OTG CRC window control register,
+		 * so close it before starting secure_display.
+		 */
+		amdgpu_dm_psr_disable(acrtc->dm_irq_params.stream, true);
+
 		spin_lock_irq(&adev_to_drm(adev)->event_lock);
-		spin_lock_irq(&crc_rd_wrk->crc_rd_work_lock);
-		if (crc_rd_wrk && crc_rd_wrk->crtc) {
-			old_crtc = crc_rd_wrk->crtc;
-			old_acrtc = to_amdgpu_crtc(old_crtc);
-		}
-		new_acrtc = to_amdgpu_crtc(new_crtc);
 
-		if (old_crtc && old_crtc != new_crtc) {
-			old_acrtc->dm_irq_params.crc_window.activated = false;
-			old_acrtc->dm_irq_params.crc_window.update_win = false;
-			old_acrtc->dm_irq_params.crc_window.skip_frame_cnt = 0;
+		acrtc->dm_irq_params.window_param[0].enable = true;
+		acrtc->dm_irq_params.window_param[0].update_win = true;
+		acrtc->dm_irq_params.window_param[0].skip_frame_cnt = 0;
+		acrtc->dm_irq_params.crc_window_activated = true;
 
-			new_acrtc->dm_irq_params.crc_window.activated = true;
-			new_acrtc->dm_irq_params.crc_window.update_win = true;
-			new_acrtc->dm_irq_params.crc_window.skip_frame_cnt = 0;
-			crc_rd_wrk->crtc = new_crtc;
-		} else {
-			new_acrtc->dm_irq_params.crc_window.activated = true;
-			new_acrtc->dm_irq_params.crc_window.update_win = true;
-			new_acrtc->dm_irq_params.crc_window.skip_frame_cnt = 0;
-			crc_rd_wrk->crtc = new_crtc;
-		}
-		spin_unlock_irq(&crc_rd_wrk->crc_rd_work_lock);
 		spin_unlock_irq(&adev_to_drm(adev)->event_lock);
+		mutex_unlock(&adev->dm.dc_lock);
 	}
 
 	return 0;
@@ -2951,9 +3755,10 @@ static int crc_win_update_get(void *data, u64 *val)
 
 DEFINE_DEBUGFS_ATTRIBUTE(crc_win_update_fops, crc_win_update_get,
 			 crc_win_update_set, "%llu\n");
-
+#endif
 void crtc_debugfs_init(struct drm_crtc *crtc)
 {
+#ifdef CONFIG_DRM_AMD_SECURE_DISPLAY
 	struct dentry *dir = debugfs_lookup("crc", crtc->debugfs_entry);
 
 	if (!dir)
@@ -2969,9 +3774,14 @@ void crtc_debugfs_init(struct drm_crtc *crtc)
 				   &crc_win_y_end_fops);
 	debugfs_create_file_unsafe("crc_win_update", 0644, dir, crtc,
 				   &crc_win_update_fops);
-
-}
+	dput(dir);
 #endif
+	debugfs_create_file("amdgpu_current_bpc", 0644, crtc->debugfs_entry,
+			    crtc, &amdgpu_current_bpc_fops);
+	debugfs_create_file("amdgpu_current_colorspace", 0644, crtc->debugfs_entry,
+			    crtc, &amdgpu_current_colorspace_fops);
+}
+
 /*
  * Writes DTN log state to the user supplied buffer.
  * Example usage: cat /sys/kernel/debug/dri/0/amdgpu_dm_dtn_log
@@ -3076,6 +3886,7 @@ static int trigger_hpd_mst_set(void *data, u64 val)
 	struct amdgpu_dm_connector *aconnector;
 	struct drm_connector *connector;
 	struct dc_link *link = NULL;
+	int ret;
 
 	if (val == 1) {
 		drm_connector_list_iter_begin(dev, &iter);
@@ -3083,8 +3894,16 @@ static int trigger_hpd_mst_set(void *data, u64 val)
 			aconnector = to_amdgpu_dm_connector(connector);
 			if (aconnector->dc_link->type == dc_connection_mst_branch &&
 			    aconnector->mst_mgr.aux) {
-				dc_link_detect(aconnector->dc_link, DETECT_REASON_HPD);
-				drm_dp_mst_topology_mgr_set_mst(&aconnector->mst_mgr, true);
+				mutex_lock(&adev->dm.dc_lock);
+				ret = dc_link_detect(aconnector->dc_link, DETECT_REASON_HPD);
+				mutex_unlock(&adev->dm.dc_lock);
+
+				if (!ret)
+					DRM_ERROR("DM_MST: Failed to detect dc link!");
+
+				ret = drm_dp_mst_topology_mgr_set_mst(&aconnector->mst_mgr, true);
+				if (ret < 0)
+					DRM_ERROR("DM_MST: Failed to set the device into MST mode!");
 			}
 		}
 	} else if (val == 0) {
@@ -3094,12 +3913,12 @@ static int trigger_hpd_mst_set(void *data, u64 val)
 			if (!aconnector->dc_link)
 				continue;
 
-			if (!aconnector->mst_port)
+			if (!aconnector->mst_root)
 				continue;
 
 			link = aconnector->dc_link;
-			dp_receiver_power_ctrl(link, false);
-			drm_dp_mst_topology_mgr_set_mst(&aconnector->mst_port->mst_mgr, false);
+			dc_link_dp_receiver_power_ctrl(link, false);
+			drm_dp_mst_topology_mgr_set_mst(&aconnector->mst_root->mst_mgr, false);
 			link->mst_stream_alloc_table.stream_count = 0;
 			memset(link->mst_stream_alloc_table.stream_allocations, 0,
 					sizeof(link->mst_stream_alloc_table.stream_allocations));
@@ -3191,6 +4010,84 @@ DEFINE_DEBUGFS_ATTRIBUTE(disable_hpd_ops, disable_hpd_get,
 			 disable_hpd_set, "%llu\n");
 
 /*
+ * Prints hardware capabilities. These are used for IGT testing.
+ */
+static int capabilities_show(struct seq_file *m, void *unused)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)m->private;
+	struct dc *dc = adev->dm.dc;
+	bool mall_supported = dc->caps.mall_size_total;
+	bool subvp_supported = dc->caps.subvp_fw_processing_delay_us;
+	unsigned int mall_in_use = false;
+	unsigned int subvp_in_use = false;
+
+	struct hubbub *hubbub = dc->res_pool->hubbub;
+
+	if (hubbub && hubbub->funcs->get_mall_en)
+		hubbub->funcs->get_mall_en(hubbub, &mall_in_use);
+
+	if (dc->cap_funcs.get_subvp_en)
+		subvp_in_use = dc->cap_funcs.get_subvp_en(dc, dc->current_state);
+
+	seq_printf(m, "mall supported: %s, enabled: %s\n",
+			   mall_supported ? "yes" : "no", mall_in_use ? "yes" : "no");
+	seq_printf(m, "sub-viewport supported: %s, enabled: %s\n",
+			   subvp_supported ? "yes" : "no", subvp_in_use ? "yes" : "no");
+
+	return 0;
+}
+
+DEFINE_SHOW_ATTRIBUTE(capabilities);
+
+/*
+ * Temporary w/a to force sst sequence in M42D DP2 mst receiver
+ * Example usage: echo 1 > /sys/kernel/debug/dri/0/amdgpu_dm_dp_set_mst_en_for_sst
+ */
+static int dp_force_sst_set(void *data, u64 val)
+{
+	struct amdgpu_device *adev = data;
+
+	adev->dm.dc->debug.set_mst_en_for_sst = val;
+
+	return 0;
+}
+
+static int dp_force_sst_get(void *data, u64 *val)
+{
+	struct amdgpu_device *adev = data;
+
+	*val = adev->dm.dc->debug.set_mst_en_for_sst;
+
+	return 0;
+}
+DEFINE_DEBUGFS_ATTRIBUTE(dp_set_mst_en_for_sst_ops, dp_force_sst_get,
+			 dp_force_sst_set, "%llu\n");
+
+/*
+ * Force DP2 sequence without VESA certified cable.
+ * Example usage: echo 1 > /sys/kernel/debug/dri/0/amdgpu_dm_dp_ignore_cable_id
+ */
+static int dp_ignore_cable_id_set(void *data, u64 val)
+{
+	struct amdgpu_device *adev = data;
+
+	adev->dm.dc->debug.ignore_cable_id = val;
+
+	return 0;
+}
+
+static int dp_ignore_cable_id_get(void *data, u64 *val)
+{
+	struct amdgpu_device *adev = data;
+
+	*val = adev->dm.dc->debug.ignore_cable_id;
+
+	return 0;
+}
+DEFINE_DEBUGFS_ATTRIBUTE(dp_ignore_cable_id_ops, dp_ignore_cable_id_get,
+			 dp_ignore_cable_id_set, "%llu\n");
+
+/*
  * Sets the DC visual confirm debug option from the given string.
  * Example usage: echo 1 > /sys/kernel/debug/dri/0/amdgpu_visual_confirm
  */
@@ -3220,6 +4117,40 @@ DEFINE_SHOW_ATTRIBUTE(mst_topo);
 DEFINE_DEBUGFS_ATTRIBUTE(visual_confirm_fops, visual_confirm_get,
 			 visual_confirm_set, "%llu\n");
 
+
+/*
+ * Sets the DC skip_detection_link_training debug option from the given string.
+ * Example usage: echo 1 > /sys/kernel/debug/dri/0/amdgpu_skip_detection_link_training
+ */
+static int skip_detection_link_training_set(void *data, u64 val)
+{
+	struct amdgpu_device *adev = data;
+
+	if (val == 0)
+		adev->dm.dc->debug.skip_detection_link_training = false;
+	else
+		adev->dm.dc->debug.skip_detection_link_training = true;
+
+	return 0;
+}
+
+/*
+ * Reads the DC skip_detection_link_training debug option value into the given buffer.
+ * Example usage: cat /sys/kernel/debug/dri/0/amdgpu_dm_skip_detection_link_training
+ */
+static int skip_detection_link_training_get(void *data, u64 *val)
+{
+	struct amdgpu_device *adev = data;
+
+	*val = adev->dm.dc->debug.skip_detection_link_training;
+
+	return 0;
+}
+
+DEFINE_DEBUGFS_ATTRIBUTE(skip_detection_link_training_fops,
+			 skip_detection_link_training_get,
+			 skip_detection_link_training_set, "%llu\n");
+
 /*
  * Dumps the DCC_EN bit for each pipe.
  * Example usage: cat /sys/kernel/debug/dri/0/amdgpu_dm_dcc_en
@@ -3252,8 +4183,10 @@ static ssize_t dcc_en_bits_read(
 	dc->hwss.get_dcc_en_bits(dc, dcc_en_bits);
 
 	rd_buf = kcalloc(rd_buf_size, sizeof(char), GFP_KERNEL);
-	if (!rd_buf)
+	if (!rd_buf) {
+		kfree(dcc_en_bits);
 		return -ENOMEM;
+	}
 
 	for (i = 0; i < num_pipes; i++)
 		offset += snprintf(rd_buf + offset, rd_buf_size - offset,
@@ -3266,8 +4199,10 @@ static ssize_t dcc_en_bits_read(
 		if (*pos >= rd_buf_size)
 			break;
 		r = put_user(*(rd_buf + result), buf);
-		if (r)
+		if (r) {
+			kfree(rd_buf);
 			return r; /* r = -EFAULT */
+		}
 		buf += 1;
 		size -= 1;
 		*pos += 1;
@@ -3297,12 +4232,21 @@ void dtn_debugfs_init(struct amdgpu_device *adev)
 
 	debugfs_create_file("amdgpu_mst_topology", 0444, root,
 			    adev, &mst_topo_fops);
+	debugfs_create_file("amdgpu_dm_capabilities", 0444, root,
+			    adev, &capabilities_fops);
 	debugfs_create_file("amdgpu_dm_dtn_log", 0644, root, adev,
 			    &dtn_log_fops);
+	debugfs_create_file("amdgpu_dm_dp_set_mst_en_for_sst", 0644, root, adev,
+				&dp_set_mst_en_for_sst_ops);
+	debugfs_create_file("amdgpu_dm_dp_ignore_cable_id", 0644, root, adev,
+				&dp_ignore_cable_id_ops);
 
 	debugfs_create_file_unsafe("amdgpu_dm_visual_confirm", 0644, root, adev,
 				   &visual_confirm_fops);
 
+	debugfs_create_file_unsafe("amdgpu_dm_skip_detection_link_training", 0644, root, adev,
+				   &skip_detection_link_training_fops);
+
 	debugfs_create_file_unsafe("amdgpu_dm_dmub_tracebuffer", 0644, root,
 				   adev, &dmub_tracebuffer_fops);
 
@@ -3312,6 +4256,9 @@ void dtn_debugfs_init(struct amdgpu_device *adev)
 	debugfs_create_file_unsafe("amdgpu_dm_force_timing_sync", 0644, root,
 				   adev, &force_timing_sync_ops);
 
+	debugfs_create_file_unsafe("amdgpu_dm_dmub_trace_mask", 0644, root,
+				   adev, &dmub_trace_mask_fops);
+
 	debugfs_create_file_unsafe("amdgpu_dm_dmcub_trace_event_en", 0644, root,
 				   adev, &dmcub_trace_event_state_fops);
 
@@ -3324,4 +4271,7 @@ void dtn_debugfs_init(struct amdgpu_device *adev)
 	debugfs_create_file_unsafe("amdgpu_dm_disable_hpd", 0644, root, adev,
 				   &disable_hpd_ops);
 
+	if (adev->dm.dc->caps.ips_support)
+		debugfs_create_file_unsafe("amdgpu_dm_ips_status", 0644, root, adev,
+					   &ips_status_fops);
 }
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.h b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.h
index 3366cb644053..122cdc124b3b 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.h
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.h
@@ -1,3 +1,4 @@
+/* SPDX-License-Identifier: MIT */
 /*
  * Copyright 2018 Advanced Micro Devices, Inc.
  *
@@ -31,8 +32,6 @@
 
 void connector_debugfs_init(struct amdgpu_dm_connector *connector);
 void dtn_debugfs_init(struct amdgpu_device *adev);
-#if defined(CONFIG_DRM_AMD_SECURE_DISPLAY)
 void crtc_debugfs_init(struct drm_crtc *crtc);
-#endif
 
 #endif
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_hdcp.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_hdcp.c
index 5bfdc66b5867..85ce558cefc5 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_hdcp.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_hdcp.c
@@ -1,3 +1,4 @@
+// SPDX-License-Identifier: MIT
 /*
  * Copyright 2019 Advanced Micro Devices, Inc.
  *
@@ -26,8 +27,9 @@
 #include "amdgpu_dm_hdcp.h"
 #include "amdgpu.h"
 #include "amdgpu_dm.h"
+#include "dc_fused_io.h"
 #include "dm_helpers.h"
-#include <drm/drm_hdcp.h>
+#include <drm/display/drm_hdcp_helper.h>
 #include "hdcp_psp.h"
 
 /*
@@ -39,10 +41,10 @@
 static bool
 lp_write_i2c(void *handle, uint32_t address, const uint8_t *data, uint32_t size)
 {
-
 	struct dc_link *link = handle;
 	struct i2c_payload i2c_payloads[] = {{true, address, size, (void *)data} };
-	struct i2c_command cmd = {i2c_payloads, 1, I2C_COMMAND_ENGINE_HW, link->dc->caps.i2c_speed_in_khz};
+	struct i2c_command cmd = {i2c_payloads, 1, I2C_COMMAND_ENGINE_HW,
+				  link->dc->caps.i2c_speed_in_khz};
 
 	return dm_helpers_submit_i2c(link->ctx, link, &cmd);
 }
@@ -52,8 +54,10 @@ lp_read_i2c(void *handle, uint32_t address, uint8_t offset, uint8_t *data, uint3
 {
 	struct dc_link *link = handle;
 
-	struct i2c_payload i2c_payloads[] = {{true, address, 1, &offset}, {false, address, size, data} };
-	struct i2c_command cmd = {i2c_payloads, 2, I2C_COMMAND_ENGINE_HW, link->dc->caps.i2c_speed_in_khz};
+	struct i2c_payload i2c_payloads[] = {{true, address, 1, &offset},
+					     {false, address, size, data} };
+	struct i2c_command cmd = {i2c_payloads, 2, I2C_COMMAND_ENGINE_HW,
+				  link->dc->caps.i2c_speed_in_khz};
 
 	return dm_helpers_submit_i2c(link->ctx, link, &cmd);
 }
@@ -74,9 +78,36 @@ lp_read_dpcd(void *handle, uint32_t address, uint8_t *data, uint32_t size)
 	return dm_helpers_dp_read_dpcd(link->ctx, link, address, data, size);
 }
 
-static uint8_t *psp_get_srm(struct psp_context *psp, uint32_t *srm_version, uint32_t *srm_size)
+static bool lp_atomic_write_poll_read_i2c(
+		void *handle,
+		const struct mod_hdcp_atomic_op_i2c *write,
+		const struct mod_hdcp_atomic_op_i2c *poll,
+		struct mod_hdcp_atomic_op_i2c *read,
+		uint32_t poll_timeout_us,
+		uint8_t poll_mask_msb
+)
+{
+	struct dc_link *link = handle;
+
+	return dm_atomic_write_poll_read_i2c(link, write, poll, read, poll_timeout_us, poll_mask_msb);
+}
+
+static bool lp_atomic_write_poll_read_aux(
+		void *handle,
+		const struct mod_hdcp_atomic_op_aux *write,
+		const struct mod_hdcp_atomic_op_aux *poll,
+		struct mod_hdcp_atomic_op_aux *read,
+		uint32_t poll_timeout_us,
+		uint8_t poll_mask_msb
+)
 {
+	struct dc_link *link = handle;
+
+	return dm_atomic_write_poll_read_aux(link, write, poll, read, poll_timeout_us, poll_mask_msb);
+}
 
+static uint8_t *psp_get_srm(struct psp_context *psp, uint32_t *srm_version, uint32_t *srm_size)
+{
 	struct ta_hdcp_shared_memory *hdcp_cmd;
 
 	if (!psp->hdcp_context.context.initialized) {
@@ -96,13 +127,12 @@ static uint8_t *psp_get_srm(struct psp_context *psp, uint32_t *srm_version, uint
 	*srm_version = hdcp_cmd->out_msg.hdcp_get_srm.srm_version;
 	*srm_size = hdcp_cmd->out_msg.hdcp_get_srm.srm_buf_size;
 
-
 	return hdcp_cmd->out_msg.hdcp_get_srm.srm_buf;
 }
 
-static int psp_set_srm(struct psp_context *psp, uint8_t *srm, uint32_t srm_size, uint32_t *srm_version)
+static int psp_set_srm(struct psp_context *psp,
+		       u8 *srm, uint32_t srm_size, uint32_t *srm_version)
 {
-
 	struct ta_hdcp_shared_memory *hdcp_cmd;
 
 	if (!psp->hdcp_context.context.initialized) {
@@ -119,7 +149,8 @@ static int psp_set_srm(struct psp_context *psp, uint8_t *srm, uint32_t srm_size,
 
 	psp_hdcp_invoke(psp, hdcp_cmd->cmd_id);
 
-	if (hdcp_cmd->hdcp_status != TA_HDCP_STATUS__SUCCESS || hdcp_cmd->out_msg.hdcp_set_srm.valid_signature != 1 ||
+	if (hdcp_cmd->hdcp_status != TA_HDCP_STATUS__SUCCESS ||
+	    hdcp_cmd->out_msg.hdcp_set_srm.valid_signature != 1 ||
 	    hdcp_cmd->out_msg.hdcp_set_srm.srm_version == PSP_SRM_VERSION_MAX)
 		return -EINVAL;
 
@@ -150,7 +181,6 @@ static void process_output(struct hdcp_workqueue *hdcp_work)
 
 static void link_lock(struct hdcp_workqueue *work, bool lock)
 {
-
 	int i = 0;
 
 	for (i = 0; i < work->max_link; i++) {
@@ -160,72 +190,74 @@ static void link_lock(struct hdcp_workqueue *work, bool lock)
 			mutex_unlock(&work[i].mutex);
 	}
 }
+
 void hdcp_update_display(struct hdcp_workqueue *hdcp_work,
 			 unsigned int link_index,
 			 struct amdgpu_dm_connector *aconnector,
-			 uint8_t content_type,
+			 u8 content_type,
 			 bool enable_encryption)
 {
 	struct hdcp_workqueue *hdcp_w = &hdcp_work[link_index];
-	struct mod_hdcp_display *display = &hdcp_work[link_index].display;
-	struct mod_hdcp_link *link = &hdcp_work[link_index].link;
-	struct mod_hdcp_display_query query;
-
-	mutex_lock(&hdcp_w->mutex);
-	hdcp_w->aconnector = aconnector;
-
-	query.display = NULL;
-	mod_hdcp_query_display(&hdcp_w->hdcp, aconnector->base.index, &query);
-
-	if (query.display != NULL) {
-		memcpy(display, query.display, sizeof(struct mod_hdcp_display));
-		mod_hdcp_remove_display(&hdcp_w->hdcp, aconnector->base.index, &hdcp_w->output);
-
-		hdcp_w->link.adjust.hdcp2.force_type = MOD_HDCP_FORCE_TYPE_0;
-
-		if (enable_encryption) {
-			/* Explicitly set the saved SRM as sysfs call will be after we already enabled hdcp
-			 * (s3 resume case)
-			 */
-			if (hdcp_work->srm_size > 0)
-				psp_set_srm(hdcp_work->hdcp.config.psp.handle, hdcp_work->srm, hdcp_work->srm_size,
-					    &hdcp_work->srm_version);
-
-			display->adjust.disable = MOD_HDCP_DISPLAY_NOT_DISABLE;
-			if (content_type == DRM_MODE_HDCP_CONTENT_TYPE0) {
-				hdcp_w->link.adjust.hdcp1.disable = 0;
-				hdcp_w->link.adjust.hdcp2.force_type = MOD_HDCP_FORCE_TYPE_0;
-			} else if (content_type == DRM_MODE_HDCP_CONTENT_TYPE1) {
-				hdcp_w->link.adjust.hdcp1.disable = 1;
-				hdcp_w->link.adjust.hdcp2.force_type = MOD_HDCP_FORCE_TYPE_1;
-			}
-
-			schedule_delayed_work(&hdcp_w->property_validate_dwork,
-					      msecs_to_jiffies(DRM_HDCP_CHECK_PERIOD_MS));
-		} else {
-			display->adjust.disable = MOD_HDCP_DISPLAY_DISABLE_AUTHENTICATION;
-			hdcp_w->encryption_status = MOD_HDCP_ENCRYPTION_STATUS_HDCP_OFF;
-			cancel_delayed_work(&hdcp_w->property_validate_dwork);
+	struct mod_hdcp_link_adjustment link_adjust;
+	struct mod_hdcp_display_adjustment display_adjust;
+	unsigned int conn_index = aconnector->base.index;
+	const struct dc *dc = aconnector->dc_link->dc;
+
+	guard(mutex)(&hdcp_w->mutex);
+	drm_connector_get(&aconnector->base);
+	if (hdcp_w->aconnector[conn_index])
+		drm_connector_put(&hdcp_w->aconnector[conn_index]->base);
+	hdcp_w->aconnector[conn_index] = aconnector;
+
+	memset(&link_adjust, 0, sizeof(link_adjust));
+	memset(&display_adjust, 0, sizeof(display_adjust));
+
+	if (enable_encryption) {
+		/* Explicitly set the saved SRM as sysfs call will be after we already enabled hdcp
+		 * (s3 resume case)
+		 */
+		if (hdcp_work->srm_size > 0)
+			psp_set_srm(hdcp_work->hdcp.config.psp.handle, hdcp_work->srm,
+				    hdcp_work->srm_size,
+				    &hdcp_work->srm_version);
+
+		display_adjust.disable = MOD_HDCP_DISPLAY_NOT_DISABLE;
+
+		link_adjust.auth_delay = 2;
+		link_adjust.retry_limit = MAX_NUM_OF_ATTEMPTS;
+
+		if (content_type == DRM_MODE_HDCP_CONTENT_TYPE0) {
+			link_adjust.hdcp2.force_type = MOD_HDCP_FORCE_TYPE_0;
+		} else if (content_type == DRM_MODE_HDCP_CONTENT_TYPE1) {
+			link_adjust.hdcp1.disable = 1;
+			link_adjust.hdcp2.force_type = MOD_HDCP_FORCE_TYPE_1;
 		}
+		link_adjust.hdcp2.use_fw_locality_check =
+				(dc->caps.fused_io_supported || dc->debug.hdcp_lc_force_fw_enable);
+		link_adjust.hdcp2.use_sw_locality_fallback = dc->debug.hdcp_lc_enable_sw_fallback;
 
-		display->state = MOD_HDCP_DISPLAY_ACTIVE;
+		schedule_delayed_work(&hdcp_w->property_validate_dwork,
+				      msecs_to_jiffies(DRM_HDCP_CHECK_PERIOD_MS));
+	} else {
+		display_adjust.disable = MOD_HDCP_DISPLAY_DISABLE_AUTHENTICATION;
+		hdcp_w->encryption_status[conn_index] = MOD_HDCP_ENCRYPTION_STATUS_HDCP_OFF;
+		cancel_delayed_work(&hdcp_w->property_validate_dwork);
 	}
 
-	mod_hdcp_add_display(&hdcp_w->hdcp, link, display, &hdcp_w->output);
+	mod_hdcp_update_display(&hdcp_w->hdcp, conn_index, &link_adjust, &display_adjust, &hdcp_w->output);
 
 	process_output(hdcp_w);
-	mutex_unlock(&hdcp_w->mutex);
 }
 
 static void hdcp_remove_display(struct hdcp_workqueue *hdcp_work,
-			 unsigned int link_index,
+				unsigned int link_index,
 			 struct amdgpu_dm_connector *aconnector)
 {
 	struct hdcp_workqueue *hdcp_w = &hdcp_work[link_index];
 	struct drm_connector_state *conn_state = aconnector->base.state;
+	unsigned int conn_index = aconnector->base.index;
 
-	mutex_lock(&hdcp_w->mutex);
-	hdcp_w->aconnector = aconnector;
+	guard(mutex)(&hdcp_w->mutex);
 
 	/* the removal of display will invoke auth reset -> hdcp destroy and
 	 * we'd expect the Content Protection (CP) property changed back to
@@ -236,28 +268,39 @@ static void hdcp_remove_display(struct hdcp_workqueue *hdcp_work,
 		conn_state->content_protection = DRM_MODE_CONTENT_PROTECTION_DESIRED;
 
 		DRM_DEBUG_DRIVER("[HDCP_DM] display %d, CP 2 -> 1, type %u, DPMS %u\n",
-			 aconnector->base.index, conn_state->hdcp_content_type, aconnector->base.dpms);
+				 aconnector->base.index, conn_state->hdcp_content_type,
+				 aconnector->base.dpms);
 	}
 
 	mod_hdcp_remove_display(&hdcp_w->hdcp, aconnector->base.index, &hdcp_w->output);
-
+	if (hdcp_w->aconnector[conn_index]) {
+		drm_connector_put(&hdcp_w->aconnector[conn_index]->base);
+		hdcp_w->aconnector[conn_index] = NULL;
+	}
 	process_output(hdcp_w);
-	mutex_unlock(&hdcp_w->mutex);
 }
+
 void hdcp_reset_display(struct hdcp_workqueue *hdcp_work, unsigned int link_index)
 {
 	struct hdcp_workqueue *hdcp_w = &hdcp_work[link_index];
+	unsigned int conn_index;
 
-	mutex_lock(&hdcp_w->mutex);
+	guard(mutex)(&hdcp_w->mutex);
 
 	mod_hdcp_reset_connection(&hdcp_w->hdcp,  &hdcp_w->output);
 
 	cancel_delayed_work(&hdcp_w->property_validate_dwork);
-	hdcp_w->encryption_status = MOD_HDCP_ENCRYPTION_STATUS_HDCP_OFF;
 
-	process_output(hdcp_w);
+	for (conn_index = 0; conn_index < AMDGPU_DM_MAX_DISPLAY_INDEX; conn_index++) {
+		hdcp_w->encryption_status[conn_index] =
+			MOD_HDCP_ENCRYPTION_STATUS_HDCP_OFF;
+		if (hdcp_w->aconnector[conn_index]) {
+			drm_connector_put(&hdcp_w->aconnector[conn_index]->base);
+			hdcp_w->aconnector[conn_index] = NULL;
+		}
+	}
 
-	mutex_unlock(&hdcp_w->mutex);
+	process_output(hdcp_w);
 }
 
 void hdcp_handle_cpirq(struct hdcp_workqueue *hdcp_work, unsigned int link_index)
@@ -267,17 +310,14 @@ void hdcp_handle_cpirq(struct hdcp_workqueue *hdcp_work, unsigned int link_index
 	schedule_work(&hdcp_w->cpirq_work);
 }
 
-
-
-
 static void event_callback(struct work_struct *work)
 {
 	struct hdcp_workqueue *hdcp_work;
 
 	hdcp_work = container_of(to_delayed_work(work), struct hdcp_workqueue,
-				      callback_dwork);
+				 callback_dwork);
 
-	mutex_lock(&hdcp_work->mutex);
+	guard(mutex)(&hdcp_work->mutex);
 
 	cancel_delayed_work(&hdcp_work->callback_dwork);
 
@@ -285,46 +325,76 @@ static void event_callback(struct work_struct *work)
 			       &hdcp_work->output);
 
 	process_output(hdcp_work);
-
-	mutex_unlock(&hdcp_work->mutex);
-
-
 }
+
 static void event_property_update(struct work_struct *work)
 {
-
-	struct hdcp_workqueue *hdcp_work = container_of(work, struct hdcp_workqueue, property_update_work);
-	struct amdgpu_dm_connector *aconnector = hdcp_work->aconnector;
-	struct drm_device *dev = hdcp_work->aconnector->base.dev;
+	struct hdcp_workqueue *hdcp_work = container_of(work, struct hdcp_workqueue,
+							property_update_work);
+	struct amdgpu_dm_connector *aconnector = NULL;
+	struct drm_device *dev;
 	long ret;
+	unsigned int conn_index;
+	struct drm_connector *connector;
+	struct drm_connector_state *conn_state;
 
-	drm_modeset_lock(&dev->mode_config.connection_mutex, NULL);
-	mutex_lock(&hdcp_work->mutex);
+	for (conn_index = 0; conn_index < AMDGPU_DM_MAX_DISPLAY_INDEX; conn_index++) {
+		aconnector = hdcp_work->aconnector[conn_index];
 
+		if (!aconnector)
+			continue;
 
-	if (aconnector->base.state->commit) {
-		ret = wait_for_completion_interruptible_timeout(&aconnector->base.state->commit->hw_done, 10 * HZ);
+		connector = &aconnector->base;
 
-		if (ret == 0) {
-			DRM_ERROR("HDCP state unknown! Setting it to DESIRED");
-			hdcp_work->encryption_status = MOD_HDCP_ENCRYPTION_STATUS_HDCP_OFF;
-		}
-	}
+		/* check if display connected */
+		if (connector->status != connector_status_connected)
+			continue;
 
-	if (hdcp_work->encryption_status != MOD_HDCP_ENCRYPTION_STATUS_HDCP_OFF) {
-		if (aconnector->base.state->hdcp_content_type == DRM_MODE_HDCP_CONTENT_TYPE0 &&
-		    hdcp_work->encryption_status <= MOD_HDCP_ENCRYPTION_STATUS_HDCP2_TYPE0_ON)
-			drm_hdcp_update_content_protection(&aconnector->base, DRM_MODE_CONTENT_PROTECTION_ENABLED);
-		else if (aconnector->base.state->hdcp_content_type == DRM_MODE_HDCP_CONTENT_TYPE1 &&
-			 hdcp_work->encryption_status == MOD_HDCP_ENCRYPTION_STATUS_HDCP2_TYPE1_ON)
-			drm_hdcp_update_content_protection(&aconnector->base, DRM_MODE_CONTENT_PROTECTION_ENABLED);
-	} else {
-		drm_hdcp_update_content_protection(&aconnector->base, DRM_MODE_CONTENT_PROTECTION_DESIRED);
-	}
+		conn_state = aconnector->base.state;
 
+		if (!conn_state)
+			continue;
 
-	mutex_unlock(&hdcp_work->mutex);
-	drm_modeset_unlock(&dev->mode_config.connection_mutex);
+		dev = connector->dev;
+
+		if (!dev)
+			continue;
+
+		drm_modeset_lock(&dev->mode_config.connection_mutex, NULL);
+		guard(mutex)(&hdcp_work->mutex);
+
+		if (conn_state->commit) {
+			ret = wait_for_completion_interruptible_timeout(&conn_state->commit->hw_done,
+									10 * HZ);
+			if (ret == 0) {
+				DRM_ERROR("HDCP state unknown! Setting it to DESIRED\n");
+				hdcp_work->encryption_status[conn_index] =
+					MOD_HDCP_ENCRYPTION_STATUS_HDCP_OFF;
+			}
+		}
+		if (hdcp_work->encryption_status[conn_index] !=
+			MOD_HDCP_ENCRYPTION_STATUS_HDCP_OFF) {
+			if (conn_state->hdcp_content_type ==
+				DRM_MODE_HDCP_CONTENT_TYPE0 &&
+				hdcp_work->encryption_status[conn_index] <=
+				MOD_HDCP_ENCRYPTION_STATUS_HDCP2_TYPE0_ON) {
+				DRM_DEBUG_DRIVER("[HDCP_DM] DRM_MODE_CONTENT_PROTECTION_ENABLED\n");
+				drm_hdcp_update_content_protection(connector,
+								   DRM_MODE_CONTENT_PROTECTION_ENABLED);
+			} else if (conn_state->hdcp_content_type ==
+					DRM_MODE_HDCP_CONTENT_TYPE1 &&
+					hdcp_work->encryption_status[conn_index] ==
+					MOD_HDCP_ENCRYPTION_STATUS_HDCP2_TYPE1_ON) {
+				drm_hdcp_update_content_protection(connector,
+								   DRM_MODE_CONTENT_PROTECTION_ENABLED);
+			}
+		} else {
+			DRM_DEBUG_DRIVER("[HDCP_DM] DRM_MODE_CONTENT_PROTECTION_DESIRED\n");
+			drm_hdcp_update_content_protection(connector,
+							   DRM_MODE_CONTENT_PROTECTION_DESIRED);
+		}
+		drm_modeset_unlock(&dev->mode_config.connection_mutex);
+	}
 }
 
 static void event_property_validate(struct work_struct *work)
@@ -332,22 +402,49 @@ static void event_property_validate(struct work_struct *work)
 	struct hdcp_workqueue *hdcp_work =
 		container_of(to_delayed_work(work), struct hdcp_workqueue, property_validate_dwork);
 	struct mod_hdcp_display_query query;
-	struct amdgpu_dm_connector *aconnector = hdcp_work->aconnector;
+	struct amdgpu_dm_connector *aconnector;
+	unsigned int conn_index;
 
-	if (!aconnector)
-		return;
+	guard(mutex)(&hdcp_work->mutex);
 
-	mutex_lock(&hdcp_work->mutex);
+	for (conn_index = 0; conn_index < AMDGPU_DM_MAX_DISPLAY_INDEX;
+	     conn_index++) {
+		aconnector = hdcp_work->aconnector[conn_index];
 
-	query.encryption_status = MOD_HDCP_ENCRYPTION_STATUS_HDCP_OFF;
-	mod_hdcp_query_display(&hdcp_work->hdcp, aconnector->base.index, &query);
+		if (!aconnector)
+			continue;
 
-	if (query.encryption_status != hdcp_work->encryption_status) {
-		hdcp_work->encryption_status = query.encryption_status;
-		schedule_work(&hdcp_work->property_update_work);
-	}
+		/* check if display connected */
+		if (aconnector->base.status != connector_status_connected)
+			continue;
+
+		if (!aconnector->base.state)
+			continue;
 
-	mutex_unlock(&hdcp_work->mutex);
+		query.encryption_status = MOD_HDCP_ENCRYPTION_STATUS_HDCP_OFF;
+		mod_hdcp_query_display(&hdcp_work->hdcp, aconnector->base.index,
+				       &query);
+
+		DRM_DEBUG_DRIVER("[HDCP_DM] disp %d, connector->CP %u, (query, work): (%d, %d)\n",
+				 aconnector->base.index,
+			aconnector->base.state->content_protection,
+			query.encryption_status,
+			hdcp_work->encryption_status[conn_index]);
+
+		if (query.encryption_status !=
+		    hdcp_work->encryption_status[conn_index]) {
+			DRM_DEBUG_DRIVER("[HDCP_DM] encryption_status change from %x to %x\n",
+					 hdcp_work->encryption_status[conn_index],
+					 query.encryption_status);
+
+			hdcp_work->encryption_status[conn_index] =
+				query.encryption_status;
+
+			DRM_DEBUG_DRIVER("[HDCP_DM] trigger property_update_work\n");
+
+			schedule_work(&hdcp_work->property_update_work);
+		}
+	}
 }
 
 static void event_watchdog_timer(struct work_struct *work)
@@ -355,10 +452,10 @@ static void event_watchdog_timer(struct work_struct *work)
 	struct hdcp_workqueue *hdcp_work;
 
 	hdcp_work = container_of(to_delayed_work(work),
-				      struct hdcp_workqueue,
+				 struct hdcp_workqueue,
 				      watchdog_timer_dwork);
 
-	mutex_lock(&hdcp_work->mutex);
+	guard(mutex)(&hdcp_work->mutex);
 
 	cancel_delayed_work(&hdcp_work->watchdog_timer_dwork);
 
@@ -367,9 +464,6 @@ static void event_watchdog_timer(struct work_struct *work)
 			       &hdcp_work->output);
 
 	process_output(hdcp_work);
-
-	mutex_unlock(&hdcp_work->mutex);
-
 }
 
 static void event_cpirq(struct work_struct *work)
@@ -378,17 +472,13 @@ static void event_cpirq(struct work_struct *work)
 
 	hdcp_work = container_of(work, struct hdcp_workqueue, cpirq_work);
 
-	mutex_lock(&hdcp_work->mutex);
+	guard(mutex)(&hdcp_work->mutex);
 
 	mod_hdcp_process_event(&hdcp_work->hdcp, MOD_HDCP_EVENT_CPIRQ, &hdcp_work->output);
 
 	process_output(hdcp_work);
-
-	mutex_unlock(&hdcp_work->mutex);
-
 }
 
-
 void hdcp_destroy(struct kobject *kobj, struct hdcp_workqueue *hdcp_work)
 {
 	int i = 0;
@@ -396,6 +486,7 @@ void hdcp_destroy(struct kobject *kobj, struct hdcp_workqueue *hdcp_work)
 	for (i = 0; i < hdcp_work->max_link; i++) {
 		cancel_delayed_work_sync(&hdcp_work[i].callback_dwork);
 		cancel_delayed_work_sync(&hdcp_work[i].watchdog_timer_dwork);
+		cancel_delayed_work_sync(&hdcp_work[i].property_validate_dwork);
 	}
 
 	sysfs_remove_bin_file(kobj, &hdcp_work[0].attr);
@@ -404,15 +495,12 @@ void hdcp_destroy(struct kobject *kobj, struct hdcp_workqueue *hdcp_work)
 	kfree(hdcp_work);
 }
 
-
 static bool enable_assr(void *handle, struct dc_link *link)
 {
-
 	struct hdcp_workqueue *hdcp_work = handle;
 	struct mod_hdcp hdcp = hdcp_work->hdcp;
 	struct psp_context *psp = hdcp.config.psp.handle;
 	struct ta_dtm_shared_memory *dtm_cmd;
-	bool res = true;
 
 	if (!psp->dtm_context.context.initialized) {
 		DRM_INFO("Failed to enable ASSR, DTM TA is not initialized.");
@@ -421,23 +509,22 @@ static bool enable_assr(void *handle, struct dc_link *link)
 
 	dtm_cmd = (struct ta_dtm_shared_memory *)psp->dtm_context.context.mem_context.shared_buf;
 
-	mutex_lock(&psp->dtm_context.mutex);
+	guard(mutex)(&psp->dtm_context.mutex);
 	memset(dtm_cmd, 0, sizeof(struct ta_dtm_shared_memory));
 
 	dtm_cmd->cmd_id = TA_DTM_COMMAND__TOPOLOGY_ASSR_ENABLE;
-	dtm_cmd->dtm_in_message.topology_assr_enable.display_topology_dig_be_index = link->link_enc_hw_inst;
+	dtm_cmd->dtm_in_message.topology_assr_enable.display_topology_dig_be_index =
+		link->link_enc_hw_inst;
 	dtm_cmd->dtm_status = TA_DTM_STATUS__GENERIC_FAILURE;
 
 	psp_dtm_invoke(psp, dtm_cmd->cmd_id);
 
 	if (dtm_cmd->dtm_status != TA_DTM_STATUS__SUCCESS) {
 		DRM_INFO("Failed to enable ASSR");
-		res = false;
+		return false;
 	}
 
-	mutex_unlock(&psp->dtm_context.mutex);
-
-	return res;
+	return true;
 }
 
 static void update_config(void *handle, struct cp_psp_stream_config *config)
@@ -445,11 +532,13 @@ static void update_config(void *handle, struct cp_psp_stream_config *config)
 	struct hdcp_workqueue *hdcp_work = handle;
 	struct amdgpu_dm_connector *aconnector = config->dm_stream_ctx;
 	int link_index = aconnector->dc_link->link_index;
+	unsigned int conn_index = aconnector->base.index;
 	struct mod_hdcp_display *display = &hdcp_work[link_index].display;
 	struct mod_hdcp_link *link = &hdcp_work[link_index].link;
-	struct drm_connector_state *conn_state;
+	struct hdcp_workqueue *hdcp_w = &hdcp_work[link_index];
 	struct dc_sink *sink = NULL;
 	bool link_is_hdcp14 = false;
+	const struct dc *dc = aconnector->dc_link->dc;
 
 	if (config->dpms_off) {
 		hdcp_remove_display(hdcp_work, link_index, aconnector);
@@ -467,7 +556,7 @@ static void update_config(void *handle, struct cp_psp_stream_config *config)
 	else if (aconnector->dc_em_sink)
 		sink = aconnector->dc_em_sink;
 
-	if (sink != NULL)
+	if (sink)
 		link->mode = mod_hdcp_signal_type_to_operation_mode(sink->sink_signal);
 
 	display->controller = CONTROLLER_ID_D0 + config->otg_inst;
@@ -476,27 +565,45 @@ static void update_config(void *handle, struct cp_psp_stream_config *config)
 	link->ddc_line = aconnector->dc_link->ddc_hw_inst + 1;
 	display->stream_enc_idx = config->stream_enc_idx;
 	link->link_enc_idx = config->link_enc_idx;
+	link->dio_output_id = config->dio_output_idx;
 	link->phy_idx = config->phy_idx;
+
 	if (sink)
 		link_is_hdcp14 = dc_link_is_hdcp14(aconnector->dc_link, sink->sink_signal);
 	link->hdcp_supported_informational = link_is_hdcp14;
 	link->dp.rev = aconnector->dc_link->dpcd_caps.dpcd_rev.raw;
 	link->dp.assr_enabled = config->assr_enabled;
 	link->dp.mst_enabled = config->mst_enabled;
+	link->dp.dp2_enabled = config->dp2_enabled;
+	link->dp.usb4_enabled = config->usb4_enabled;
 	display->adjust.disable = MOD_HDCP_DISPLAY_DISABLE_AUTHENTICATION;
-	link->adjust.auth_delay = 3;
+	link->adjust.auth_delay = 2;
+	link->adjust.retry_limit = MAX_NUM_OF_ATTEMPTS;
 	link->adjust.hdcp1.disable = 0;
-	conn_state = aconnector->base.state;
+	link->adjust.hdcp2.use_fw_locality_check = (dc->caps.fused_io_supported || dc->debug.hdcp_lc_force_fw_enable);
+	link->adjust.hdcp2.use_sw_locality_fallback = dc->debug.hdcp_lc_enable_sw_fallback;
+	hdcp_w->encryption_status[display->index] = MOD_HDCP_ENCRYPTION_STATUS_HDCP_OFF;
 
 	DRM_DEBUG_DRIVER("[HDCP_DM] display %d, CP %d, type %d\n", aconnector->base.index,
-			(!!aconnector->base.state) ? aconnector->base.state->content_protection : -1,
-			(!!aconnector->base.state) ? aconnector->base.state->hdcp_content_type : -1);
+			 (!!aconnector->base.state) ?
+			 aconnector->base.state->content_protection : -1,
+			 (!!aconnector->base.state) ?
+			 aconnector->base.state->hdcp_content_type : -1);
 
-	hdcp_update_display(hdcp_work, link_index, aconnector, conn_state->hdcp_content_type, false);
-}
+	guard(mutex)(&hdcp_w->mutex);
 
+	mod_hdcp_add_display(&hdcp_w->hdcp, link, display, &hdcp_w->output);
+	drm_connector_get(&aconnector->base);
+	if (hdcp_w->aconnector[conn_index])
+		drm_connector_put(&hdcp_w->aconnector[conn_index]->base);
+	hdcp_w->aconnector[conn_index] = aconnector;
+	process_output(hdcp_w);
+}
 
-/* NOTE: From the usermodes prospective you only need to call write *ONCE*, the kernel
+/**
+ * DOC: Add sysfs interface for set/get srm
+ *
+ * NOTE: From the usermodes prospective you only need to call write *ONCE*, the kernel
  *      will automatically call once or twice depending on the size
  *
  * call: "cat file > /sys/class/drm/card0/device/hdcp_srm" from usermode no matter what the size is
@@ -507,23 +614,23 @@ static void update_config(void *handle, struct cp_psp_stream_config *config)
  * sysfs interface doesn't tell us the size we will get so we are sending partial SRMs to psp and on
  * the last call we will send the full SRM. PSP will fail on every call before the last.
  *
- * This means we don't know if the SRM is good until the last call. And because of this limitation we
- * cannot throw errors early as it will stop the kernel from writing to sysfs
+ * This means we don't know if the SRM is good until the last call. And because of this
+ * limitation we cannot throw errors early as it will stop the kernel from writing to sysfs
  *
  * Example 1:
- * 	Good SRM size = 5096
- * 	first call to write 4096 -> PSP fails
- * 	Second call to write 1000 -> PSP Pass -> SRM is set
+ *	Good SRM size = 5096
+ *	first call to write 4096 -> PSP fails
+ *	Second call to write 1000 -> PSP Pass -> SRM is set
  *
  * Example 2:
- * 	Bad SRM size = 4096
- * 	first call to write 4096 -> PSP fails (This is the same as above, but we don't know if this
- * 	is the last call)
+ *	Bad SRM size = 4096
+ *	first call to write 4096 -> PSP fails (This is the same as above, but we don't know if this
+ *	is the last call)
  *
  * Solution?:
- * 	1: Parse the SRM? -> It is signed so we don't know the EOF
- * 	2: We can have another sysfs that passes the size before calling set. -> simpler solution
- * 	below
+ *	1: Parse the SRM? -> It is signed so we don't know the EOF
+ *	2: We can have another sysfs that passes the size before calling set. -> simpler solution
+ *	below
  *
  * Easy Solution:
  * Always call get after Set to verify if set was successful.
@@ -532,20 +639,21 @@ static void update_config(void *handle, struct cp_psp_stream_config *config)
  * +----------------------+
  * PSP will only update its srm if its older than the one we are trying to load.
  * Always do set first than get.
- * 	-if we try to "1. SET" a older version PSP will reject it and we can "2. GET" the newer
- * 	version and save it
+ *	-if we try to "1. SET" a older version PSP will reject it and we can "2. GET" the newer
+ *	version and save it
  *
- * 	-if we try to "1. SET" a newer version PSP will accept it and we can "2. GET" the
- * 	same(newer) version back and save it
+ *	-if we try to "1. SET" a newer version PSP will accept it and we can "2. GET" the
+ *	same(newer) version back and save it
  *
- * 	-if we try to "1. SET" a newer version and PSP rejects it. That means the format is
- * 	incorrect/corrupted and we should correct our SRM by getting it from PSP
+ *	-if we try to "1. SET" a newer version and PSP rejects it. That means the format is
+ *	incorrect/corrupted and we should correct our SRM by getting it from PSP
  */
-static ssize_t srm_data_write(struct file *filp, struct kobject *kobj, struct bin_attribute *bin_attr, char *buffer,
+static ssize_t srm_data_write(struct file *filp, struct kobject *kobj,
+			      const struct bin_attribute *bin_attr, char *buffer,
 			      loff_t pos, size_t count)
 {
 	struct hdcp_workqueue *work;
-	uint32_t srm_version = 0;
+	u32 srm_version = 0;
 
 	work = container_of(bin_attr, struct hdcp_workqueue, attr);
 	link_lock(work, true);
@@ -559,19 +667,19 @@ static ssize_t srm_data_write(struct file *filp, struct kobject *kobj, struct bi
 		work->srm_version = srm_version;
 	}
 
-
 	link_lock(work, false);
 
 	return count;
 }
 
-static ssize_t srm_data_read(struct file *filp, struct kobject *kobj, struct bin_attribute *bin_attr, char *buffer,
+static ssize_t srm_data_read(struct file *filp, struct kobject *kobj,
+			     const struct bin_attribute *bin_attr, char *buffer,
 			     loff_t pos, size_t count)
 {
 	struct hdcp_workqueue *work;
-	uint8_t *srm = NULL;
-	uint32_t srm_version;
-	uint32_t srm_size;
+	u8 *srm = NULL;
+	u32 srm_version;
+	u32 srm_size;
 	size_t ret = count;
 
 	work = container_of(bin_attr, struct hdcp_workqueue, attr);
@@ -604,12 +712,12 @@ ret:
 /* From the hdcp spec (5.Renewability) SRM needs to be stored in a non-volatile memory.
  *
  * For example,
- * 	if Application "A" sets the SRM (ver 2) and we reboot/suspend and later when Application "B"
- * 	needs to use HDCP, the version in PSP should be SRM(ver 2). So SRM should be persistent
- * 	across boot/reboots/suspend/resume/shutdown
+ *	if Application "A" sets the SRM (ver 2) and we reboot/suspend and later when Application "B"
+ *	needs to use HDCP, the version in PSP should be SRM(ver 2). So SRM should be persistent
+ *	across boot/reboots/suspend/resume/shutdown
  *
- * Currently when the system goes down (suspend/shutdown) the SRM is cleared from PSP. For HDCP we need
- * to make the SRM persistent.
+ * Currently when the system goes down (suspend/shutdown) the SRM is cleared from PSP. For HDCP
+ * we need to make the SRM persistent.
  *
  * -PSP owns the checking of SRM but doesn't have the ability to store it in a non-volatile memory.
  * -The kernel cannot write to the file systems.
@@ -619,8 +727,8 @@ ret:
  *
  * Usermode can read/write to/from PSP using the sysfs interface
  * For example:
- * 	to save SRM from PSP to storage : cat /sys/class/drm/card0/device/hdcp_srm > srmfile
- * 	to load from storage to PSP: cat srmfile > /sys/class/drm/card0/device/hdcp_srm
+ *	to save SRM from PSP to storage : cat /sys/class/drm/card0/device/hdcp_srm > srmfile
+ *	to load from storage to PSP: cat srmfile > /sys/class/drm/card0/device/hdcp_srm
  */
 static const struct bin_attribute data_attr = {
 	.attr = {.name = "hdcp_srm", .mode = 0664},
@@ -629,10 +737,9 @@ static const struct bin_attribute data_attr = {
 	.read = srm_data_read,
 };
 
-
-struct hdcp_workqueue *hdcp_create_workqueue(struct amdgpu_device *adev, struct cp_psp *cp_psp, struct dc *dc)
+struct hdcp_workqueue *hdcp_create_workqueue(struct amdgpu_device *adev,
+					     struct cp_psp *cp_psp, struct dc *dc)
 {
-
 	int max_caps = dc->caps.max_links;
 	struct hdcp_workqueue *hdcp_work;
 	int i = 0;
@@ -641,14 +748,16 @@ struct hdcp_workqueue *hdcp_create_workqueue(struct amdgpu_device *adev, struct
 	if (ZERO_OR_NULL_PTR(hdcp_work))
 		return NULL;
 
-	hdcp_work->srm = kcalloc(PSP_HDCP_SRM_FIRST_GEN_MAX_SIZE, sizeof(*hdcp_work->srm), GFP_KERNEL);
+	hdcp_work->srm = kcalloc(PSP_HDCP_SRM_FIRST_GEN_MAX_SIZE,
+				 sizeof(*hdcp_work->srm), GFP_KERNEL);
 
-	if (hdcp_work->srm == NULL)
+	if (!hdcp_work->srm)
 		goto fail_alloc_context;
 
-	hdcp_work->srm_temp = kcalloc(PSP_HDCP_SRM_FIRST_GEN_MAX_SIZE, sizeof(*hdcp_work->srm_temp), GFP_KERNEL);
+	hdcp_work->srm_temp = kcalloc(PSP_HDCP_SRM_FIRST_GEN_MAX_SIZE,
+				      sizeof(*hdcp_work->srm_temp), GFP_KERNEL);
 
-	if (hdcp_work->srm_temp == NULL)
+	if (!hdcp_work->srm_temp)
 		goto fail_alloc_context;
 
 	hdcp_work->max_link = max_caps;
@@ -662,14 +771,37 @@ struct hdcp_workqueue *hdcp_create_workqueue(struct amdgpu_device *adev, struct
 		INIT_DELAYED_WORK(&hdcp_work[i].watchdog_timer_dwork, event_watchdog_timer);
 		INIT_DELAYED_WORK(&hdcp_work[i].property_validate_dwork, event_property_validate);
 
-		hdcp_work[i].hdcp.config.psp.handle = &adev->psp;
-		if (dc->ctx->dce_version == DCN_VERSION_3_1)
-			hdcp_work[i].hdcp.config.psp.caps.dtm_v3_supported = 1;
-		hdcp_work[i].hdcp.config.ddc.handle = dc_get_link_at_index(dc, i);
-		hdcp_work[i].hdcp.config.ddc.funcs.write_i2c = lp_write_i2c;
-		hdcp_work[i].hdcp.config.ddc.funcs.read_i2c = lp_read_i2c;
-		hdcp_work[i].hdcp.config.ddc.funcs.write_dpcd = lp_write_dpcd;
-		hdcp_work[i].hdcp.config.ddc.funcs.read_dpcd = lp_read_dpcd;
+		struct mod_hdcp_config *config = &hdcp_work[i].hdcp.config;
+		struct mod_hdcp_ddc_funcs *ddc_funcs = &config->ddc.funcs;
+
+		config->psp.handle = &adev->psp;
+		if (dc->ctx->dce_version == DCN_VERSION_3_1  ||
+		    dc->ctx->dce_version == DCN_VERSION_3_14 ||
+		    dc->ctx->dce_version == DCN_VERSION_3_15 ||
+		    dc->ctx->dce_version == DCN_VERSION_3_16 ||
+		    dc->ctx->dce_version == DCN_VERSION_3_2  ||
+		    dc->ctx->dce_version == DCN_VERSION_3_21 ||
+		    dc->ctx->dce_version == DCN_VERSION_3_5  ||
+		    dc->ctx->dce_version == DCN_VERSION_3_51 ||
+		    dc->ctx->dce_version == DCN_VERSION_3_6  ||
+		    dc->ctx->dce_version == DCN_VERSION_4_01)
+			config->psp.caps.dtm_v3_supported = 1;
+
+		config->ddc.handle = dc_get_link_at_index(dc, i);
+
+		ddc_funcs->write_i2c = lp_write_i2c;
+		ddc_funcs->read_i2c = lp_read_i2c;
+		ddc_funcs->write_dpcd = lp_write_dpcd;
+		ddc_funcs->read_dpcd = lp_read_dpcd;
+		ddc_funcs->atomic_write_poll_read_i2c = lp_atomic_write_poll_read_i2c;
+		ddc_funcs->atomic_write_poll_read_aux = lp_atomic_write_poll_read_aux;
+
+		memset(hdcp_work[i].aconnector, 0,
+		       sizeof(struct amdgpu_dm_connector *) *
+			       AMDGPU_DM_MAX_DISPLAY_INDEX);
+		memset(hdcp_work[i].encryption_status, 0,
+		       sizeof(enum mod_hdcp_encryption_status) *
+			       AMDGPU_DM_MAX_DISPLAY_INDEX);
 	}
 
 	cp_psp->funcs.update_stream_config = update_config;
@@ -691,10 +823,5 @@ fail_alloc_context:
 	kfree(hdcp_work);
 
 	return NULL;
-
-
-
 }
 
-
-
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_hdcp.h b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_hdcp.h
index 09294ff122fe..4faa344f196e 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_hdcp.h
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_hdcp.h
@@ -1,3 +1,4 @@
+/* SPDX-License-Identifier: MIT */
 /*
  * Copyright 2019 Advanced Micro Devices, Inc.
  *
@@ -43,7 +44,7 @@ struct hdcp_workqueue {
 	struct delayed_work callback_dwork;
 	struct delayed_work watchdog_timer_dwork;
 	struct delayed_work property_validate_dwork;
-	struct amdgpu_dm_connector *aconnector;
+	struct amdgpu_dm_connector *aconnector[AMDGPU_DM_MAX_DISPLAY_INDEX];
 	struct mutex mutex;
 
 	struct mod_hdcp hdcp;
@@ -51,7 +52,20 @@ struct hdcp_workqueue {
 	struct mod_hdcp_display display;
 	struct mod_hdcp_link link;
 
-	enum mod_hdcp_encryption_status encryption_status;
+	enum mod_hdcp_encryption_status encryption_status[AMDGPU_DM_MAX_DISPLAY_INDEX];
+	/* when display is unplugged from mst hub, connctor will be
+	 * destroyed within dm_dp_mst_connector_destroy. connector
+	 * hdcp perperties, like type, undesired, desired, enabled,
+	 * will be lost. So, save hdcp properties into hdcp_work within
+	 * amdgpu_dm_atomic_commit_tail. if the same display is
+	 * plugged back with same display index, its hdcp properties
+	 * will be retrieved from hdcp_work within dm_dp_mst_get_modes
+	 */
+	/* un-desired, desired, enabled */
+	unsigned int content_protection[AMDGPU_DM_MAX_DISPLAY_INDEX];
+	/* hdcp1.x, hdcp2.x */
+	unsigned int hdcp_content_type[AMDGPU_DM_MAX_DISPLAY_INDEX];
+
 	uint8_t max_link;
 
 	uint8_t *srm;
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c
index 8cbeeb7c986d..ac98c746c3de 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c
@@ -1,3 +1,4 @@
+// SPDX-License-Identifier: MIT
 /*
  * Copyright 2015 Advanced Micro Devices, Inc.
  *
@@ -23,13 +24,17 @@
  *
  */
 
+#include <acpi/video.h>
+
 #include <linux/string.h>
 #include <linux/acpi.h>
 #include <linux/i2c.h>
 
+#include <drm/drm_atomic.h>
 #include <drm/drm_probe_helper.h>
 #include <drm/amdgpu_drm.h>
 #include <drm/drm_edid.h>
+#include <drm/drm_fixed.h>
 
 #include "dm_services.h"
 #include "amdgpu.h"
@@ -37,62 +42,78 @@
 #include "amdgpu_dm.h"
 #include "amdgpu_dm_irq.h"
 #include "amdgpu_dm_mst_types.h"
+#include "dpcd_defs.h"
+#include "dc/inc/core_types.h"
 
 #include "dm_helpers.h"
+#include "ddc_service_types.h"
+#include "clk_mgr.h"
 
-struct monitor_patch_info {
-	unsigned int manufacturer_id;
-	unsigned int product_id;
-	void (*patch_func)(struct dc_edid_caps *edid_caps, unsigned int param);
-	unsigned int patch_param;
-};
-static void set_max_dsc_bpp_limit(struct dc_edid_caps *edid_caps, unsigned int param);
-
-static const struct monitor_patch_info monitor_patch_table[] = {
-{0x6D1E, 0x5BBF, set_max_dsc_bpp_limit, 15},
-{0x6D1E, 0x5B9A, set_max_dsc_bpp_limit, 15},
-};
-
-static void set_max_dsc_bpp_limit(struct dc_edid_caps *edid_caps, unsigned int param)
+static u32 edid_extract_panel_id(struct edid *edid)
 {
-	if (edid_caps)
-		edid_caps->panel_patch.max_dsc_target_bpp_limit = param;
+	return (u32)edid->mfg_id[0] << 24   |
+	       (u32)edid->mfg_id[1] << 16   |
+	       (u32)EDID_PRODUCT_ID(edid);
 }
 
-static int amdgpu_dm_patch_edid_caps(struct dc_edid_caps *edid_caps)
+static void apply_edid_quirks(struct drm_device *dev, struct edid *edid, struct dc_edid_caps *edid_caps)
 {
-	int i, ret = 0;
-
-	for (i = 0; i < ARRAY_SIZE(monitor_patch_table); i++)
-		if ((edid_caps->manufacturer_id == monitor_patch_table[i].manufacturer_id)
-			&&  (edid_caps->product_id == monitor_patch_table[i].product_id)) {
-			monitor_patch_table[i].patch_func(edid_caps, monitor_patch_table[i].patch_param);
-			ret++;
-		}
-
-	return ret;
+	uint32_t panel_id = edid_extract_panel_id(edid);
+
+	switch (panel_id) {
+	/* Workaround for monitors that need a delay after detecting the link */
+	case drm_edid_encode_panel_id('G', 'B', 'T', 0x3215):
+		drm_dbg_driver(dev, "Add 10s delay for link detection for panel id %X\n", panel_id);
+		edid_caps->panel_patch.wait_after_dpcd_poweroff_ms = 10000;
+		break;
+	/* Workaround for some monitors which does not work well with FAMS */
+	case drm_edid_encode_panel_id('S', 'A', 'M', 0x0E5E):
+	case drm_edid_encode_panel_id('S', 'A', 'M', 0x7053):
+	case drm_edid_encode_panel_id('S', 'A', 'M', 0x71AC):
+		drm_dbg_driver(dev, "Disabling FAMS on monitor with panel id %X\n", panel_id);
+		edid_caps->panel_patch.disable_fams = true;
+		break;
+	/* Workaround for some monitors that do not clear DPCD 0x317 if FreeSync is unsupported */
+	case drm_edid_encode_panel_id('A', 'U', 'O', 0xA7AB):
+	case drm_edid_encode_panel_id('A', 'U', 'O', 0xE69B):
+	case drm_edid_encode_panel_id('B', 'O', 'E', 0x092A):
+	case drm_edid_encode_panel_id('L', 'G', 'D', 0x06D1):
+	case drm_edid_encode_panel_id('M', 'S', 'F', 0x1003):
+		drm_dbg_driver(dev, "Clearing DPCD 0x317 on monitor with panel id %X\n", panel_id);
+		edid_caps->panel_patch.remove_sink_ext_caps = true;
+		break;
+	case drm_edid_encode_panel_id('S', 'D', 'C', 0x4154):
+	case drm_edid_encode_panel_id('S', 'D', 'C', 0x4171):
+		drm_dbg_driver(dev, "Disabling VSC on monitor with panel id %X\n", panel_id);
+		edid_caps->panel_patch.disable_colorimetry = true;
+		break;
+	default:
+		return;
+	}
 }
 
-/* dm_helpers_parse_edid_caps
- *
- * Parse edid caps
+/**
+ * dm_helpers_parse_edid_caps() - Parse edid caps
  *
+ * @link: current detected link
  * @edid:	[in] pointer to edid
- *  edid_caps:	[in] pointer to edid caps
- * @return
- *	void
- * */
+ * @edid_caps:	[in] pointer to edid caps
+ *
+ * Return: void
+ */
 enum dc_edid_status dm_helpers_parse_edid_caps(
-		struct dc_context *ctx,
+		struct dc_link *link,
 		const struct dc_edid *edid,
 		struct dc_edid_caps *edid_caps)
 {
-	struct edid *edid_buf = (struct edid *) edid->raw_edid;
+	struct amdgpu_dm_connector *aconnector = link->priv;
+	struct drm_connector *connector = &aconnector->base;
+	struct drm_device *dev = connector->dev;
+	struct edid *edid_buf = edid ? (struct edid *) edid->raw_edid : NULL;
 	struct cea_sad *sads;
 	int sad_count = -1;
 	int sadb_count = -1;
 	int i = 0;
-	int j = 0;
 	uint8_t *sadb = NULL;
 
 	enum dc_edid_status result = EDID_OK;
@@ -110,30 +131,21 @@ enum dc_edid_status dm_helpers_parse_edid_caps(
 	edid_caps->serial_number = edid_buf->serial;
 	edid_caps->manufacture_week = edid_buf->mfg_week;
 	edid_caps->manufacture_year = edid_buf->mfg_year;
+	edid_caps->analog = !(edid_buf->input & DRM_EDID_INPUT_DIGITAL);
 
-	/* One of the four detailed_timings stores the monitor name. It's
-	 * stored in an array of length 13. */
-	for (i = 0; i < 4; i++) {
-		if (edid_buf->detailed_timings[i].data.other_data.type == 0xfc) {
-			while (j < 13 && edid_buf->detailed_timings[i].data.other_data.data.str.str[j]) {
-				if (edid_buf->detailed_timings[i].data.other_data.data.str.str[j] == '\n')
-					break;
-
-				edid_caps->display_name[j] =
-					edid_buf->detailed_timings[i].data.other_data.data.str.str[j];
-				j++;
-			}
-		}
-	}
+	drm_edid_get_monitor_name(edid_buf,
+				  edid_caps->display_name,
+				  AUDIO_INFO_DISPLAY_NAME_SIZE_IN_CHARS);
 
-	edid_caps->edid_hdmi = drm_detect_hdmi_monitor(
-			(struct edid *) edid->raw_edid);
+	edid_caps->edid_hdmi = connector->display_info.is_hdmi;
+
+	apply_edid_quirks(dev, edid_buf, edid_caps);
 
 	sad_count = drm_edid_to_sad((struct edid *) edid->raw_edid, &sads);
 	if (sad_count <= 0)
 		return result;
 
-	edid_caps->audio_mode_count = sad_count < DC_MAX_AUDIO_DESC_COUNT ? sad_count : DC_MAX_AUDIO_DESC_COUNT;
+	edid_caps->audio_mode_count = min(sad_count, DC_MAX_AUDIO_DESC_COUNT);
 	for (i = 0; i < edid_caps->audio_mode_count; ++i) {
 		struct cea_sad *sad = &sads[i];
 
@@ -158,46 +170,58 @@ enum dc_edid_status dm_helpers_parse_edid_caps(
 	kfree(sads);
 	kfree(sadb);
 
-	amdgpu_dm_patch_edid_caps(edid_caps);
-
 	return result;
 }
 
-static void get_payload_table(
-		struct amdgpu_dm_connector *aconnector,
-		struct dp_mst_stream_allocation_table *proposed_table)
+static void
+fill_dc_mst_payload_table_from_drm(struct dc_link *link,
+				   bool enable,
+				   struct drm_dp_mst_atomic_payload *target_payload,
+				   struct dc_dp_mst_stream_allocation_table *table)
 {
-	int i;
-	struct drm_dp_mst_topology_mgr *mst_mgr =
-			&aconnector->mst_port->mst_mgr;
-
-	mutex_lock(&mst_mgr->payload_lock);
-
-	proposed_table->stream_count = 0;
+	struct dc_dp_mst_stream_allocation_table new_table = { 0 };
+	struct dc_dp_mst_stream_allocation *sa;
+	struct link_mst_stream_allocation_table copy_of_link_table =
+										link->mst_stream_alloc_table;
 
-	/* number of active streams */
-	for (i = 0; i < mst_mgr->max_payloads; i++) {
-		if (mst_mgr->payloads[i].num_slots == 0)
-			break; /* end of vcp_id table */
-
-		ASSERT(mst_mgr->payloads[i].payload_state !=
-				DP_PAYLOAD_DELETE_LOCAL);
-
-		if (mst_mgr->payloads[i].payload_state == DP_PAYLOAD_LOCAL ||
-			mst_mgr->payloads[i].payload_state ==
-					DP_PAYLOAD_REMOTE) {
+	int i;
+	int current_hw_table_stream_cnt = copy_of_link_table.stream_count;
+	struct link_mst_stream_allocation *dc_alloc;
 
-			struct dp_mst_stream_allocation *sa =
-					&proposed_table->stream_allocations[
-						proposed_table->stream_count];
+	/* TODO: refactor to set link->mst_stream_alloc_table directly if possible.*/
+	if (enable) {
+		dc_alloc =
+		&copy_of_link_table.stream_allocations[current_hw_table_stream_cnt];
+		dc_alloc->vcp_id = target_payload->vcpi;
+		dc_alloc->slot_count = target_payload->time_slots;
+	} else {
+		for (i = 0; i < copy_of_link_table.stream_count; i++) {
+			dc_alloc =
+			&copy_of_link_table.stream_allocations[i];
+
+			if (dc_alloc->vcp_id == target_payload->vcpi) {
+				dc_alloc->vcp_id = 0;
+				dc_alloc->slot_count = 0;
+				break;
+			}
+		}
+		ASSERT(i != copy_of_link_table.stream_count);
+	}
 
-			sa->slot_count = mst_mgr->payloads[i].num_slots;
-			sa->vcp_id = mst_mgr->proposed_vcpis[i]->vcpi;
-			proposed_table->stream_count++;
+	/* Fill payload info*/
+	for (i = 0; i < MAX_CONTROLLER_NUM; i++) {
+		dc_alloc =
+			&copy_of_link_table.stream_allocations[i];
+		if (dc_alloc->vcp_id > 0 && dc_alloc->slot_count > 0) {
+			sa = &new_table.stream_allocations[new_table.stream_count];
+			sa->slot_count = dc_alloc->slot_count;
+			sa->vcp_id = dc_alloc->vcp_id;
+			new_table.stream_count++;
 		}
 	}
 
-	mutex_unlock(&mst_mgr->payload_lock);
+	/* Overwrite the old table */
+	*table = new_table;
 }
 
 void dm_helpers_dp_update_branch_info(
@@ -205,65 +229,86 @@ void dm_helpers_dp_update_branch_info(
 	const struct dc_link *link)
 {}
 
+static void dm_helpers_construct_old_payload(
+			struct drm_dp_mst_topology_mgr *mgr,
+			struct drm_dp_mst_topology_state *mst_state,
+			struct drm_dp_mst_atomic_payload *new_payload,
+			struct drm_dp_mst_atomic_payload *old_payload)
+{
+	struct drm_dp_mst_atomic_payload *pos;
+	int pbn_per_slot = dfixed_trunc(mst_state->pbn_div);
+	u8 next_payload_vc_start = mgr->next_start_slot;
+	u8 payload_vc_start = new_payload->vc_start_slot;
+	u8 allocated_time_slots;
+
+	*old_payload = *new_payload;
+
+	/* Set correct time_slots/PBN of old payload.
+	 * other fields (delete & dsc_enabled) in
+	 * struct drm_dp_mst_atomic_payload are don't care fields
+	 * while calling drm_dp_remove_payload_part2()
+	 */
+	list_for_each_entry(pos, &mst_state->payloads, next) {
+		if (pos != new_payload &&
+		    pos->vc_start_slot > payload_vc_start &&
+		    pos->vc_start_slot < next_payload_vc_start)
+			next_payload_vc_start = pos->vc_start_slot;
+	}
+
+	allocated_time_slots = next_payload_vc_start - payload_vc_start;
+
+	old_payload->time_slots = allocated_time_slots;
+	old_payload->pbn = allocated_time_slots * pbn_per_slot;
+}
+
 /*
  * Writes payload allocation table in immediate downstream device.
  */
 bool dm_helpers_dp_mst_write_payload_allocation_table(
 		struct dc_context *ctx,
 		const struct dc_stream_state *stream,
-		struct dp_mst_stream_allocation_table *proposed_table,
+		struct dc_dp_mst_stream_allocation_table *proposed_table,
 		bool enable)
 {
 	struct amdgpu_dm_connector *aconnector;
-	struct dm_connector_state *dm_conn_state;
+	struct drm_dp_mst_topology_state *mst_state;
+	struct drm_dp_mst_atomic_payload *target_payload, *new_payload, old_payload;
 	struct drm_dp_mst_topology_mgr *mst_mgr;
-	struct drm_dp_mst_port *mst_port;
-	bool ret;
-	u8 link_coding_cap = DP_8b_10b_ENCODING;
 
 	aconnector = (struct amdgpu_dm_connector *)stream->dm_stream_context;
 	/* Accessing the connector state is required for vcpi_slots allocation
 	 * and directly relies on behaviour in commit check
 	 * that blocks before commit guaranteeing that the state
-	 * is not gonna be swapped while still in use in commit tail */
-
-	if (!aconnector || !aconnector->mst_port)
-		return false;
-
-	dm_conn_state = to_dm_connector_state(aconnector->base.state);
-
-	mst_mgr = &aconnector->mst_port->mst_mgr;
+	 * is not gonna be swapped while still in use in commit tail
+	 */
 
-	if (!mst_mgr->mst_state)
+	if (!aconnector || !aconnector->mst_root)
 		return false;
 
-	mst_port = aconnector->port;
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	link_coding_cap = dc_link_dp_mst_decide_link_encoding_format(aconnector->dc_link);
-#endif
+	mst_mgr = &aconnector->mst_root->mst_mgr;
+	mst_state = to_drm_dp_mst_topology_state(mst_mgr->base.state);
+	new_payload = drm_atomic_get_mst_payload_state(mst_state, aconnector->mst_output_port);
 
 	if (enable) {
+		target_payload = new_payload;
 
-		ret = drm_dp_mst_allocate_vcpi(mst_mgr, mst_port,
-					       dm_conn_state->pbn,
-					       dm_conn_state->vcpi_slots);
-		if (!ret)
-			return false;
-
+		/* It's OK for this to fail */
+		drm_dp_add_payload_part1(mst_mgr, mst_state, new_payload);
 	} else {
-		drm_dp_mst_reset_vcpi_slots(mst_mgr, mst_port);
-	}
+		/* construct old payload by VCPI*/
+		dm_helpers_construct_old_payload(mst_mgr, mst_state,
+						 new_payload, &old_payload);
+		target_payload = &old_payload;
 
-	/* It's OK for this to fail */
-	drm_dp_update_payload_part1(mst_mgr, (link_coding_cap == DP_CAP_ANSI_128B132B) ? 0:1);
+		drm_dp_remove_payload_part1(mst_mgr, mst_state, new_payload);
+	}
 
 	/* mst_mgr->->payloads are VC payload notify MST branch using DPCD or
 	 * AUX message. The sequence is slot 1-63 allocated sequence for each
 	 * stream. AMD ASIC stream slot allocation should follow the same
-	 * sequence. copy DRM MST allocation to dc */
-
-	get_payload_table(aconnector, proposed_table);
+	 * sequence. copy DRM MST allocation to dc
+	 */
+	fill_dc_mst_payload_table_from_drm(stream->link, enable, target_payload, proposed_table);
 
 	return true;
 }
@@ -298,10 +343,10 @@ enum act_return_status dm_helpers_dp_mst_poll_for_allocation_change_trigger(
 
 	aconnector = (struct amdgpu_dm_connector *)stream->dm_stream_context;
 
-	if (!aconnector || !aconnector->mst_port)
+	if (!aconnector || !aconnector->mst_root)
 		return ACT_FAILED;
 
-	mst_mgr = &aconnector->mst_port->mst_mgr;
+	mst_mgr = &aconnector->mst_root->mst_mgr;
 
 	if (!mst_mgr->mst_state)
 		return ACT_FAILED;
@@ -314,35 +359,67 @@ enum act_return_status dm_helpers_dp_mst_poll_for_allocation_change_trigger(
 	return ACT_SUCCESS;
 }
 
-bool dm_helpers_dp_mst_send_payload_allocation(
+void dm_helpers_dp_mst_send_payload_allocation(
 		struct dc_context *ctx,
-		const struct dc_stream_state *stream,
-		bool enable)
+		const struct dc_stream_state *stream)
 {
 	struct amdgpu_dm_connector *aconnector;
+	struct drm_dp_mst_topology_state *mst_state;
 	struct drm_dp_mst_topology_mgr *mst_mgr;
-	struct drm_dp_mst_port *mst_port;
+	struct drm_dp_mst_atomic_payload *new_payload;
+	enum mst_progress_status set_flag = MST_ALLOCATE_NEW_PAYLOAD;
+	enum mst_progress_status clr_flag = MST_CLEAR_ALLOCATED_PAYLOAD;
+	int ret = 0;
 
 	aconnector = (struct amdgpu_dm_connector *)stream->dm_stream_context;
 
-	if (!aconnector || !aconnector->mst_port)
-		return false;
+	if (!aconnector || !aconnector->mst_root)
+		return;
 
-	mst_port = aconnector->port;
+	mst_mgr = &aconnector->mst_root->mst_mgr;
+	mst_state = to_drm_dp_mst_topology_state(mst_mgr->base.state);
+	new_payload = drm_atomic_get_mst_payload_state(mst_state, aconnector->mst_output_port);
 
-	mst_mgr = &aconnector->mst_port->mst_mgr;
+	ret = drm_dp_add_payload_part2(mst_mgr, new_payload);
 
-	if (!mst_mgr->mst_state)
-		return false;
+	if (ret) {
+		amdgpu_dm_set_mst_status(&aconnector->mst_status,
+			set_flag, false);
+	} else {
+		amdgpu_dm_set_mst_status(&aconnector->mst_status,
+			set_flag, true);
+		amdgpu_dm_set_mst_status(&aconnector->mst_status,
+			clr_flag, false);
+	}
+}
 
-	/* It's OK for this to fail */
-	drm_dp_update_payload_part2(mst_mgr);
+void dm_helpers_dp_mst_update_mst_mgr_for_deallocation(
+		struct dc_context *ctx,
+		const struct dc_stream_state *stream)
+{
+	struct amdgpu_dm_connector *aconnector;
+	struct drm_dp_mst_topology_state *mst_state;
+	struct drm_dp_mst_topology_mgr *mst_mgr;
+	struct drm_dp_mst_atomic_payload *new_payload, old_payload;
+	enum mst_progress_status set_flag = MST_CLEAR_ALLOCATED_PAYLOAD;
+	enum mst_progress_status clr_flag = MST_ALLOCATE_NEW_PAYLOAD;
 
-	if (!enable)
-		drm_dp_mst_deallocate_vcpi(mst_mgr, mst_port);
+	aconnector = (struct amdgpu_dm_connector *)stream->dm_stream_context;
 
-	return true;
-}
+	if (!aconnector || !aconnector->mst_root)
+		return;
+
+	mst_mgr = &aconnector->mst_root->mst_mgr;
+	mst_state = to_drm_dp_mst_topology_state(mst_mgr->base.state);
+	new_payload = drm_atomic_get_mst_payload_state(mst_state, aconnector->mst_output_port);
+	dm_helpers_construct_old_payload(mst_mgr, mst_state,
+					 new_payload, &old_payload);
+
+	drm_dp_remove_payload_part2(mst_mgr, mst_state, &old_payload, new_payload);
+
+	amdgpu_dm_set_mst_status(&aconnector->mst_status, set_flag, true);
+	amdgpu_dm_set_mst_status(&aconnector->mst_status, clr_flag, false);
+ }
 
 void dm_dtn_log_begin(struct dc_context *ctx,
 	struct dc_log_buffer_ctx *log_ctx)
@@ -392,7 +469,7 @@ void dm_dtn_log_append_v(struct dc_context *ctx,
 	total = log_ctx->pos + n + 1;
 
 	if (total > log_ctx->size) {
-		char *buf = (char *)kvcalloc(total, sizeof(char), GFP_KERNEL);
+		char *buf = kvcalloc(total, sizeof(char), GFP_KERNEL);
 
 		if (buf) {
 			memcpy(buf, log_ctx->buf, log_ctx->pos);
@@ -438,6 +515,7 @@ bool dm_helpers_dp_mst_start_top_mgr(
 		bool boot)
 {
 	struct amdgpu_dm_connector *aconnector = link->priv;
+	int ret;
 
 	if (!aconnector) {
 		DRM_ERROR("Failed to find connector for link!");
@@ -453,19 +531,27 @@ bool dm_helpers_dp_mst_start_top_mgr(
 	DRM_INFO("DM_MST: starting TM on aconnector: %p [id: %d]\n",
 			aconnector, aconnector->base.base.id);
 
-	return (drm_dp_mst_topology_mgr_set_mst(&aconnector->mst_mgr, true) == 0);
+	ret = drm_dp_mst_topology_mgr_set_mst(&aconnector->mst_mgr, true);
+	if (ret < 0) {
+		DRM_ERROR("DM_MST: Failed to set the device into MST mode!");
+		return false;
+	}
+
+	DRM_INFO("DM_MST: DP%x, %d-lane link detected\n", aconnector->mst_mgr.dpcd[0],
+		aconnector->mst_mgr.dpcd[2] & DP_MAX_LANE_COUNT_MASK);
+
+	return true;
 }
 
-void dm_helpers_dp_mst_stop_top_mgr(
+bool dm_helpers_dp_mst_stop_top_mgr(
 		struct dc_context *ctx,
 		struct dc_link *link)
 {
 	struct amdgpu_dm_connector *aconnector = link->priv;
-	uint8_t i;
 
 	if (!aconnector) {
 		DRM_ERROR("Failed to find connector for link!");
-		return;
+		return false;
 	}
 
 	DRM_INFO("DM_MST: stopping TM on aconnector: %p [id: %d]\n",
@@ -473,23 +559,10 @@ void dm_helpers_dp_mst_stop_top_mgr(
 
 	if (aconnector->mst_mgr.mst_state == true) {
 		drm_dp_mst_topology_mgr_set_mst(&aconnector->mst_mgr, false);
-
-		for (i = 0; i < MAX_SINKS_PER_LINK; i++) {
-			if (link->remote_sinks[i] == NULL)
-				continue;
-
-			if (link->remote_sinks[i]->sink_signal ==
-			    SIGNAL_TYPE_DISPLAY_PORT_MST) {
-				dc_link_remove_remote_sink(link, link->remote_sinks[i]);
-
-				if (aconnector->dc_sink) {
-					dc_sink_release(aconnector->dc_sink);
-					aconnector->dc_sink = NULL;
-					aconnector->dc_link->cur_link_settings.lane_count = 0;
-				}
-			}
-		}
+		link->cur_link_settings.lane_count = 0;
 	}
+
+	return false;
 }
 
 bool dm_helpers_dp_read_dpcd(
@@ -502,13 +575,11 @@ bool dm_helpers_dp_read_dpcd(
 
 	struct amdgpu_dm_connector *aconnector = link->priv;
 
-	if (!aconnector) {
-		DC_LOG_DC("Failed to find connector for link!\n");
+	if (!aconnector)
 		return false;
-	}
 
-	return drm_dp_dpcd_read(&aconnector->dm_dp_aux.aux, address,
-			data, size) > 0;
+	return drm_dp_dpcd_read(&aconnector->dm_dp_aux.aux, address, data,
+				size) == size;
 }
 
 bool dm_helpers_dp_write_dpcd(
@@ -520,10 +591,8 @@ bool dm_helpers_dp_write_dpcd(
 {
 	struct amdgpu_dm_connector *aconnector = link->priv;
 
-	if (!aconnector) {
-		DRM_ERROR("Failed to find connector for link!");
+	if (!aconnector)
 		return false;
-	}
 
 	return drm_dp_dpcd_write(&aconnector->dm_dp_aux.aux,
 			address, (uint8_t *)data, size) > 0;
@@ -563,33 +632,285 @@ bool dm_helpers_submit_i2c(
 
 	return result;
 }
-bool dm_helpers_dp_write_dsc_enable(
+
+bool dm_helpers_execute_fused_io(
 		struct dc_context *ctx,
+		struct dc_link *link,
+		union dmub_rb_cmd *commands,
+		uint8_t count,
+		uint32_t timeout_us
+)
+{
+	struct amdgpu_device *dev = ctx->driver_context;
+
+	return amdgpu_dm_execute_fused_io(dev, link, commands, count, timeout_us);
+}
+
+static bool execute_synaptics_rc_command(struct drm_dp_aux *aux,
+		bool is_write_cmd,
+		unsigned char cmd,
+		unsigned int length,
+		unsigned int offset,
+		unsigned char *data)
+{
+	bool success = false;
+	unsigned char rc_data[16] = {0};
+	unsigned char rc_offset[4] = {0};
+	unsigned char rc_length[2] = {0};
+	unsigned char rc_cmd = 0;
+	unsigned char rc_result = 0xFF;
+	unsigned char i = 0;
+	int ret;
+
+	if (is_write_cmd) {
+		// write rc data
+		memmove(rc_data, data, length);
+		ret = drm_dp_dpcd_write(aux, SYNAPTICS_RC_DATA, rc_data, sizeof(rc_data));
+		if (ret < 0)
+			goto err;
+	}
+
+	// write rc offset
+	rc_offset[0] = (unsigned char) offset & 0xFF;
+	rc_offset[1] = (unsigned char) (offset >> 8) & 0xFF;
+	rc_offset[2] = (unsigned char) (offset >> 16) & 0xFF;
+	rc_offset[3] = (unsigned char) (offset >> 24) & 0xFF;
+	ret = drm_dp_dpcd_write(aux, SYNAPTICS_RC_OFFSET, rc_offset, sizeof(rc_offset));
+	if (ret < 0)
+		goto err;
+
+	// write rc length
+	rc_length[0] = (unsigned char) length & 0xFF;
+	rc_length[1] = (unsigned char) (length >> 8) & 0xFF;
+	ret = drm_dp_dpcd_write(aux, SYNAPTICS_RC_LENGTH, rc_length, sizeof(rc_length));
+	if (ret < 0)
+		goto err;
+
+	// write rc cmd
+	rc_cmd = cmd | 0x80;
+	ret = drm_dp_dpcd_write(aux, SYNAPTICS_RC_COMMAND, &rc_cmd, sizeof(rc_cmd));
+	if (ret < 0)
+		goto err;
+
+	// poll until active is 0
+	for (i = 0; i < 10; i++) {
+		drm_dp_dpcd_read(aux, SYNAPTICS_RC_COMMAND, &rc_cmd, sizeof(rc_cmd));
+		if (rc_cmd == cmd)
+			// active is 0
+			break;
+		msleep(10);
+	}
+
+	// read rc result
+	drm_dp_dpcd_read(aux, SYNAPTICS_RC_RESULT, &rc_result, sizeof(rc_result));
+	success = (rc_result == 0);
+
+	if (success && !is_write_cmd) {
+		// read rc data
+		drm_dp_dpcd_read(aux, SYNAPTICS_RC_DATA, data, length);
+	}
+
+	drm_dbg_dp(aux->drm_dev, "success = %d\n", success);
+
+	return success;
+
+err:
+	DRM_ERROR("%s: write cmd ..., err = %d\n",  __func__, ret);
+	return false;
+}
+
+static void apply_synaptics_fifo_reset_wa(struct drm_dp_aux *aux)
+{
+	unsigned char data[16] = {0};
+
+	drm_dbg_dp(aux->drm_dev, "Start\n");
+
+	// Step 2
+	data[0] = 'P';
+	data[1] = 'R';
+	data[2] = 'I';
+	data[3] = 'U';
+	data[4] = 'S';
+
+	if (!execute_synaptics_rc_command(aux, true, 0x01, 5, 0, data))
+		return;
+
+	// Step 3 and 4
+	if (!execute_synaptics_rc_command(aux, false, 0x31, 4, 0x220998, data))
+		return;
+
+	data[0] &= (~(1 << 1)); // set bit 1 to 0
+	if (!execute_synaptics_rc_command(aux, true, 0x21, 4, 0x220998, data))
+		return;
+
+	if (!execute_synaptics_rc_command(aux, false, 0x31, 4, 0x220D98, data))
+		return;
+
+	data[0] &= (~(1 << 1)); // set bit 1 to 0
+	if (!execute_synaptics_rc_command(aux, true, 0x21, 4, 0x220D98, data))
+		return;
+
+	if (!execute_synaptics_rc_command(aux, false, 0x31, 4, 0x221198, data))
+		return;
+
+	data[0] &= (~(1 << 1)); // set bit 1 to 0
+	if (!execute_synaptics_rc_command(aux, true, 0x21, 4, 0x221198, data))
+		return;
+
+	// Step 3 and 5
+	if (!execute_synaptics_rc_command(aux, false, 0x31, 4, 0x220998, data))
+		return;
+
+	data[0] |= (1 << 1); // set bit 1 to 1
+	if (!execute_synaptics_rc_command(aux, true, 0x21, 4, 0x220998, data))
+		return;
+
+	if (!execute_synaptics_rc_command(aux, false, 0x31, 4, 0x220D98, data))
+		return;
+
+	data[0] |= (1 << 1); // set bit 1 to 1
+
+	if (!execute_synaptics_rc_command(aux, false, 0x31, 4, 0x221198, data))
+		return;
+
+	data[0] |= (1 << 1); // set bit 1 to 1
+	if (!execute_synaptics_rc_command(aux, true, 0x21, 4, 0x221198, data))
+		return;
+
+	// Step 6
+	if (!execute_synaptics_rc_command(aux, true, 0x02, 0, 0, NULL))
+		return;
+
+	drm_dbg_dp(aux->drm_dev, "Done\n");
+}
+
+/* MST Dock */
+static const uint8_t SYNAPTICS_DEVICE_ID[] = "SYNA";
+
+static uint8_t write_dsc_enable_synaptics_non_virtual_dpcd_mst(
+		struct drm_dp_aux *aux,
 		const struct dc_stream_state *stream,
 		bool enable)
 {
-	uint8_t enable_dsc = enable ? 1 : 0;
-	struct amdgpu_dm_connector *aconnector;
 	uint8_t ret = 0;
 
-	if (!stream)
-		return false;
+	drm_dbg_dp(aux->drm_dev,
+		   "MST_DSC Configure DSC to non-virtual dpcd synaptics\n");
 
-	if (stream->signal == SIGNAL_TYPE_DISPLAY_PORT_MST) {
-		aconnector = (struct amdgpu_dm_connector *)stream->dm_stream_context;
+	if (enable) {
+		/* When DSC is enabled on previous boot and reboot with the hub,
+		 * there is a chance that Synaptics hub gets stuck during reboot sequence.
+		 * Applying a workaround to reset Synaptics SDP fifo before enabling the first stream
+		 */
+		if (!stream->link->link_status.link_active &&
+			memcmp(stream->link->dpcd_caps.branch_dev_name,
+				(int8_t *)SYNAPTICS_DEVICE_ID, 4) == 0)
+			apply_synaptics_fifo_reset_wa(aux);
+
+		ret = drm_dp_dpcd_write(aux, DP_DSC_ENABLE, &enable, 1);
+		DRM_INFO("MST_DSC Send DSC enable to synaptics\n");
+
+	} else {
+		/* Synaptics hub not support virtual dpcd,
+		 * external monitor occur garbage while disable DSC,
+		 * Disable DSC only when entire link status turn to false,
+		 */
+		if (!stream->link->link_status.link_active) {
+			ret = drm_dp_dpcd_write(aux, DP_DSC_ENABLE, &enable, 1);
+			DRM_INFO("MST_DSC Send DSC disable to synaptics\n");
+		}
+	}
+
+	return ret;
+}
 
+bool dm_helpers_dp_write_dsc_enable(
+		struct dc_context *ctx,
+		const struct dc_stream_state *stream,
+		bool enable)
+{
+	static const uint8_t DSC_DISABLE;
+	static const uint8_t DSC_DECODING = 0x01;
+	static const uint8_t DSC_PASSTHROUGH = 0x02;
+
+	struct amdgpu_dm_connector *aconnector =
+		(struct amdgpu_dm_connector *)stream->dm_stream_context;
+	struct drm_device *dev = aconnector->base.dev;
+	struct drm_dp_mst_port *port;
+	uint8_t enable_dsc = enable ? DSC_DECODING : DSC_DISABLE;
+	uint8_t enable_passthrough = enable ? DSC_PASSTHROUGH : DSC_DISABLE;
+	uint8_t ret = 0;
+
+	if (stream->signal == SIGNAL_TYPE_DISPLAY_PORT_MST) {
 		if (!aconnector->dsc_aux)
 			return false;
 
-		ret = drm_dp_dpcd_write(aconnector->dsc_aux, DP_DSC_ENABLE, &enable_dsc, 1);
+		// apply w/a to synaptics
+		if (needs_dsc_aux_workaround(aconnector->dc_link) &&
+		    (aconnector->mst_downstream_port_present.byte & 0x7) != 0x3)
+			return write_dsc_enable_synaptics_non_virtual_dpcd_mst(
+				aconnector->dsc_aux, stream, enable_dsc);
+
+		port = aconnector->mst_output_port;
+
+		if (enable) {
+			if (port->passthrough_aux) {
+				ret = drm_dp_dpcd_write(port->passthrough_aux,
+							DP_DSC_ENABLE,
+							&enable_passthrough, 1);
+				drm_dbg_dp(dev,
+					   "MST_DSC Sent DSC pass-through enable to virtual dpcd port, ret = %u\n",
+					   ret);
+			}
+
+			ret = drm_dp_dpcd_write(aconnector->dsc_aux,
+						DP_DSC_ENABLE, &enable_dsc, 1);
+			drm_dbg_dp(dev,
+				   "MST_DSC Sent DSC decoding enable to %s port, ret = %u\n",
+				   (port->passthrough_aux) ? "remote RX" :
+				   "virtual dpcd",
+				   ret);
+		} else {
+			ret = drm_dp_dpcd_write(aconnector->dsc_aux,
+						DP_DSC_ENABLE, &enable_dsc, 1);
+			drm_dbg_dp(dev,
+				   "MST_DSC Sent DSC decoding disable to %s port, ret = %u\n",
+				   (port->passthrough_aux) ? "remote RX" :
+				   "virtual dpcd",
+				   ret);
+
+			if (port->passthrough_aux) {
+				ret = drm_dp_dpcd_write(port->passthrough_aux,
+							DP_DSC_ENABLE,
+							&enable_passthrough, 1);
+				drm_dbg_dp(dev,
+					   "MST_DSC Sent DSC pass-through disable to virtual dpcd port, ret = %u\n",
+					   ret);
+			}
+		}
 	}
 
-	if (stream->signal == SIGNAL_TYPE_DISPLAY_PORT) {
-		ret = dm_helpers_dp_write_dpcd(ctx, stream->link, DP_DSC_ENABLE, &enable_dsc, 1);
-		DC_LOG_DC("Send DSC %s to sst display\n", enable_dsc ? "enable" : "disable");
+	if (stream->signal == SIGNAL_TYPE_DISPLAY_PORT || stream->signal == SIGNAL_TYPE_EDP) {
+		if (stream->sink->link->dpcd_caps.dongle_type == DISPLAY_DONGLE_NONE) {
+			ret = dm_helpers_dp_write_dpcd(ctx, stream->link, DP_DSC_ENABLE, &enable_dsc, 1);
+			drm_dbg_dp(dev,
+				   "SST_DSC Send DSC %s to SST RX\n",
+				   enable_dsc ? "enable" : "disable");
+		} else if (stream->sink->link->dpcd_caps.dongle_type == DISPLAY_DONGLE_DP_HDMI_CONVERTER) {
+			ret = dm_helpers_dp_write_dpcd(ctx, stream->link, DP_DSC_ENABLE, &enable_dsc, 1);
+			drm_dbg_dp(dev,
+				   "SST_DSC Send DSC %s to DP-HDMI PCON\n",
+				   enable_dsc ? "enable" : "disable");
+		}
 	}
 
-	return (ret > 0);
+	return ret;
+}
+
+bool dm_helpers_dp_write_hblank_reduction(struct dc_context *ctx, const struct dc_stream_state *stream)
+{
+	// TODO
+	return false;
 }
 
 bool dm_helpers_is_dp_sink_present(struct dc_link *link)
@@ -608,6 +929,67 @@ bool dm_helpers_is_dp_sink_present(struct dc_link *link)
 	return dp_sink_present;
 }
 
+static int
+dm_helpers_probe_acpi_edid(void *data, u8 *buf, unsigned int block, size_t len)
+{
+	struct drm_connector *connector = data;
+	struct acpi_device *acpidev = ACPI_COMPANION(connector->dev->dev);
+	unsigned short start = block * EDID_LENGTH;
+	struct edid *edid;
+	int r;
+
+	if (!acpidev)
+		return -ENODEV;
+
+	/* fetch the entire edid from BIOS */
+	r = acpi_video_get_edid(acpidev, ACPI_VIDEO_DISPLAY_LCD, -1, (void *)&edid);
+	if (r < 0) {
+		drm_dbg(connector->dev, "Failed to get EDID from ACPI: %d\n", r);
+		return r;
+	}
+	if (len > r || start > r || start + len > r) {
+		r = -EINVAL;
+		goto cleanup;
+	}
+
+	/* sanity check */
+	if (edid->revision < 4 || !(edid->input & DRM_EDID_INPUT_DIGITAL) ||
+	    (edid->input & DRM_EDID_DIGITAL_TYPE_MASK) == DRM_EDID_DIGITAL_TYPE_UNDEF) {
+		r = -EINVAL;
+		goto cleanup;
+	}
+
+	memcpy(buf, (void *)edid + start, len);
+	r = 0;
+
+cleanup:
+	kfree(edid);
+
+	return r;
+}
+
+static const struct drm_edid *
+dm_helpers_read_acpi_edid(struct amdgpu_dm_connector *aconnector)
+{
+	struct drm_connector *connector = &aconnector->base;
+
+	if (amdgpu_dc_debug_mask & DC_DISABLE_ACPI_EDID)
+		return NULL;
+
+	switch (connector->connector_type) {
+	case DRM_MODE_CONNECTOR_LVDS:
+	case DRM_MODE_CONNECTOR_eDP:
+		break;
+	default:
+		return NULL;
+	}
+
+	if (connector->force == DRM_FORCE_OFF)
+		return NULL;
+
+	return drm_edid_read_custom(connector, dm_helpers_probe_acpi_edid, connector);
+}
+
 enum dc_edid_status dm_helpers_read_local_edid(
 		struct dc_context *ctx,
 		struct dc_link *link,
@@ -616,9 +998,10 @@ enum dc_edid_status dm_helpers_read_local_edid(
 	struct amdgpu_dm_connector *aconnector = link->priv;
 	struct drm_connector *connector = &aconnector->base;
 	struct i2c_adapter *ddc;
-	int retry = 3;
-	enum dc_edid_status edid_status;
-	struct edid *edid;
+	int retry = 25;
+	enum dc_edid_status edid_status = EDID_NO_RESPONSE;
+	const struct drm_edid *drm_edid;
+	const struct edid *edid;
 
 	if (link->aux_mode)
 		ddc = &aconnector->dm_dp_aux.aux.ddc;
@@ -629,56 +1012,75 @@ enum dc_edid_status dm_helpers_read_local_edid(
 	 * do check sum and retry to make sure read correct edid.
 	 */
 	do {
-
-		edid = drm_get_edid(&aconnector->base, ddc);
+		drm_edid = dm_helpers_read_acpi_edid(aconnector);
+		if (drm_edid)
+			drm_info(connector->dev, "Using ACPI provided EDID for %s\n", connector->name);
+		else
+			drm_edid = drm_edid_read_ddc(connector, ddc);
+		drm_edid_connector_update(connector, drm_edid);
 
 		/* DP Compliance Test 4.2.2.6 */
 		if (link->aux_mode && connector->edid_corrupt)
 			drm_dp_send_real_edid_checksum(&aconnector->dm_dp_aux.aux, connector->real_edid_checksum);
 
-		if (!edid && connector->edid_corrupt) {
+		if (!drm_edid && connector->edid_corrupt) {
 			connector->edid_corrupt = false;
 			return EDID_BAD_CHECKSUM;
 		}
 
-		if (!edid)
-			return EDID_NO_RESPONSE;
+		if (!drm_edid)
+			continue;
+
+		edid = drm_edid_raw(drm_edid); // FIXME: Get rid of drm_edid_raw()
+		if (!edid ||
+		    edid->extensions >= sizeof(sink->dc_edid.raw_edid) / EDID_LENGTH)
+			return EDID_BAD_INPUT;
 
 		sink->dc_edid.length = EDID_LENGTH * (edid->extensions + 1);
 		memmove(sink->dc_edid.raw_edid, (uint8_t *)edid, sink->dc_edid.length);
 
 		/* We don't need the original edid anymore */
-		kfree(edid);
-
-		/* connector->display_info will be parsed from EDID and saved
-		 * into drm_connector->display_info from edid by call stack
-		 * below:
-		 * drm_parse_ycbcr420_deep_color_info
-		 * drm_parse_hdmi_forum_vsdb
-		 * drm_parse_cea_ext
-		 * drm_add_display_info
-		 * drm_connector_update_edid_property
-		 *
-		 * drm_connector->display_info will be used by amdgpu_dm funcs,
-		 * like fill_stream_properties_from_drm_display_mode
-		 */
-		amdgpu_dm_update_connector_after_detect(aconnector);
+		drm_edid_free(drm_edid);
 
 		edid_status = dm_helpers_parse_edid_caps(
-						ctx,
+						link,
 						&sink->dc_edid,
 						&sink->edid_caps);
 
-	} while (edid_status == EDID_BAD_CHECKSUM && --retry > 0);
+	} while ((edid_status == EDID_BAD_CHECKSUM || edid_status == EDID_NO_RESPONSE) && --retry > 0);
 
 	if (edid_status != EDID_OK)
 		DRM_ERROR("EDID err: %d, on connector: %s",
 				edid_status,
 				aconnector->base.name);
+	if (link->aux_mode) {
+		union test_request test_request = {0};
+		union test_response test_response = {0};
 
-	/* DP Compliance Test 4.2.2.3 */
-	if (link->aux_mode)
-		drm_dp_send_real_edid_checksum(&aconnector->dm_dp_aux.aux, sink->dc_edid.raw_edid[sink->dc_edid.length-1]);
+		dm_helpers_dp_read_dpcd(ctx,
+					link,
+					DP_TEST_REQUEST,
+					&test_request.raw,
+					sizeof(union test_request));
+
+		if (!test_request.bits.EDID_READ)
+			return edid_status;
+
+		test_response.bits.EDID_CHECKSUM_WRITE = 1;
+
+		dm_helpers_dp_write_dpcd(ctx,
+					link,
+					DP_TEST_EDID_CHECKSUM,
+					&sink->dc_edid.raw_edid[sink->dc_edid.length-1],
+					1);
+
+		dm_helpers_dp_write_dpcd(ctx,
+					link,
+					DP_TEST_RESPONSE,
+					&test_response.raw,
+					sizeof(test_response));
+
+	}
 
 	return edid_status;
 }
@@ -688,9 +1090,13 @@ int dm_helper_dmub_aux_transfer_sync(
 		struct aux_payload *payload,
 		enum aux_return_code_type *operation_result)
 {
-	return amdgpu_dm_process_dmub_aux_transfer_sync(true, ctx,
-			link->link_index, (void *)payload,
-			(void *)operation_result);
+	if (!link->hpd_status) {
+		*operation_result = AUX_RET_ERROR_HPD_DISCON;
+		return -1;
+	}
+
+	return amdgpu_dm_process_dmub_aux_transfer_sync(ctx, link->link_index, payload,
+			operation_result);
 }
 
 int dm_helpers_dmub_set_config_sync(struct dc_context *ctx,
@@ -698,9 +1104,8 @@ int dm_helpers_dmub_set_config_sync(struct dc_context *ctx,
 		struct set_config_cmd_payload *payload,
 		enum set_config_status *operation_result)
 {
-	return amdgpu_dm_process_dmub_aux_transfer_sync(false, ctx,
-			link->link_index, (void *)payload,
-			(void *)operation_result);
+	return amdgpu_dm_process_dmub_set_config_sync(ctx, link->link_index, payload,
+			operation_result);
 }
 
 void dm_set_dcn_clocks(struct dc_context *ctx, struct dc_clocks *clks)
@@ -714,6 +1119,33 @@ void dm_helpers_smu_timeout(struct dc_context *ctx, unsigned int msg_id, unsigne
 	//amdgpu_device_gpu_recover(dc_context->driver-context, NULL);
 }
 
+void dm_helpers_init_panel_settings(
+	struct dc_context *ctx,
+	struct dc_panel_config *panel_config,
+	struct dc_sink *sink)
+{
+	// Extra Panel Power Sequence
+	panel_config->pps.extra_t3_ms = sink->edid_caps.panel_patch.extra_t3_ms;
+	panel_config->pps.extra_t7_ms = sink->edid_caps.panel_patch.extra_t7_ms;
+	panel_config->pps.extra_delay_backlight_off = sink->edid_caps.panel_patch.extra_delay_backlight_off;
+	panel_config->pps.extra_post_t7_ms = 0;
+	panel_config->pps.extra_pre_t11_ms = 0;
+	panel_config->pps.extra_t12_ms = sink->edid_caps.panel_patch.extra_t12_ms;
+	panel_config->pps.extra_post_OUI_ms = 0;
+	// Feature DSC
+	panel_config->dsc.disable_dsc_edp = false;
+	panel_config->dsc.force_dsc_edp_policy = 0;
+}
+
+void dm_helpers_override_panel_settings(
+	struct dc_context *ctx,
+	struct dc_panel_config *panel_config)
+{
+	// Feature DSC
+	if (amdgpu_dc_debug_mask & DC_DISABLE_DSC)
+		panel_config->dsc.disable_dsc_edp = true;
+}
+
 void *dm_helpers_allocate_gpu_mem(
 		struct dc_context *ctx,
 		enum dc_gpu_mem_alloc_type type,
@@ -721,30 +1153,8 @@ void *dm_helpers_allocate_gpu_mem(
 		long long *addr)
 {
 	struct amdgpu_device *adev = ctx->driver_context;
-	struct dal_allocation *da;
-	u32 domain = (type == DC_MEM_ALLOC_TYPE_GART) ?
-		AMDGPU_GEM_DOMAIN_GTT : AMDGPU_GEM_DOMAIN_VRAM;
-	int ret;
-
-	da = kzalloc(sizeof(struct dal_allocation), GFP_KERNEL);
-	if (!da)
-		return NULL;
-
-	ret = amdgpu_bo_create_kernel(adev, size, PAGE_SIZE,
-				      domain, &da->bo,
-				      &da->gpu_addr, &da->cpu_ptr);
-
-	*addr = da->gpu_addr;
-
-	if (ret) {
-		kfree(da);
-		return NULL;
-	}
-
-	/* add da to list in dm */
-	list_add(&da->list, &adev->dm.da_list);
 
-	return da->cpu_ptr;
+	return dm_allocate_gpu_mem(adev, type, size, addr);
 }
 
 void dm_helpers_free_gpu_mem(
@@ -753,17 +1163,8 @@ void dm_helpers_free_gpu_mem(
 		void *pvMem)
 {
 	struct amdgpu_device *adev = ctx->driver_context;
-	struct dal_allocation *da;
-
-	/* walk the da list in DM */
-	list_for_each_entry(da, &adev->dm.da_list, list) {
-		if (pvMem == da->cpu_ptr) {
-			amdgpu_bo_free_kernel(&da->bo, &da->gpu_addr, &da->cpu_ptr);
-			list_del(&da->list);
-			kfree(da);
-			break;
-		}
-	}
+
+	dm_free_gpu_mem(adev, type, pvMem);
 }
 
 bool dm_helpers_dmub_outbox_interrupt_control(struct dc_context *ctx, bool enable)
@@ -805,16 +1206,211 @@ void dm_helpers_mst_enable_stream_features(const struct dc_stream_state *stream)
 					 sizeof(new_downspread));
 }
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-void dm_set_phyd32clk(struct dc_context *ctx, int freq_khz)
+bool dm_helpers_dp_handle_test_pattern_request(
+		struct dc_context *ctx,
+		const struct dc_link *link,
+		union link_test_pattern dpcd_test_pattern,
+		union test_misc dpcd_test_params)
 {
-       // FPGA programming for this clock in diags framework that
-       // needs to go through dm layer, therefore leave dummy interace here
+	enum dp_test_pattern test_pattern;
+	enum dp_test_pattern_color_space test_pattern_color_space =
+			DP_TEST_PATTERN_COLOR_SPACE_UNDEFINED;
+	enum dc_color_depth requestColorDepth = COLOR_DEPTH_UNDEFINED;
+	enum dc_pixel_encoding requestPixelEncoding = PIXEL_ENCODING_UNDEFINED;
+	struct pipe_ctx *pipes = link->dc->current_state->res_ctx.pipe_ctx;
+	struct pipe_ctx *pipe_ctx = NULL;
+	struct amdgpu_dm_connector *aconnector = link->priv;
+	struct drm_device *dev = aconnector->base.dev;
+	struct dc_state *dc_state = ctx->dc->current_state;
+	struct clk_mgr *clk_mgr = ctx->dc->clk_mgr;
+	int i;
+
+	for (i = 0; i < MAX_PIPES; i++) {
+		if (pipes[i].stream == NULL)
+			continue;
+
+		if (pipes[i].stream->link == link && !pipes[i].top_pipe &&
+			!pipes[i].prev_odm_pipe) {
+			pipe_ctx = &pipes[i];
+			break;
+		}
+	}
+
+	if (pipe_ctx == NULL)
+		return false;
+
+	switch (dpcd_test_pattern.bits.PATTERN) {
+	case LINK_TEST_PATTERN_COLOR_RAMP:
+		test_pattern = DP_TEST_PATTERN_COLOR_RAMP;
+	break;
+	case LINK_TEST_PATTERN_VERTICAL_BARS:
+		test_pattern = DP_TEST_PATTERN_VERTICAL_BARS;
+	break; /* black and white */
+	case LINK_TEST_PATTERN_COLOR_SQUARES:
+		test_pattern = (dpcd_test_params.bits.DYN_RANGE ==
+				TEST_DYN_RANGE_VESA ?
+				DP_TEST_PATTERN_COLOR_SQUARES :
+				DP_TEST_PATTERN_COLOR_SQUARES_CEA);
+	break;
+	default:
+		test_pattern = DP_TEST_PATTERN_VIDEO_MODE;
+	break;
+	}
+
+	if (dpcd_test_params.bits.CLR_FORMAT == 0)
+		test_pattern_color_space = DP_TEST_PATTERN_COLOR_SPACE_RGB;
+	else
+		test_pattern_color_space = dpcd_test_params.bits.YCBCR_COEFS ?
+				DP_TEST_PATTERN_COLOR_SPACE_YCBCR709 :
+				DP_TEST_PATTERN_COLOR_SPACE_YCBCR601;
+
+	switch (dpcd_test_params.bits.BPC) {
+	case 0: // 6 bits
+		requestColorDepth = COLOR_DEPTH_666;
+		break;
+	case 1: // 8 bits
+		requestColorDepth = COLOR_DEPTH_888;
+		break;
+	case 2: // 10 bits
+		requestColorDepth = COLOR_DEPTH_101010;
+		break;
+	case 3: // 12 bits
+		requestColorDepth = COLOR_DEPTH_121212;
+		break;
+	default:
+		break;
+	}
+
+	switch (dpcd_test_params.bits.CLR_FORMAT) {
+	case 0:
+		requestPixelEncoding = PIXEL_ENCODING_RGB;
+		break;
+	case 1:
+		requestPixelEncoding = PIXEL_ENCODING_YCBCR422;
+		break;
+	case 2:
+		requestPixelEncoding = PIXEL_ENCODING_YCBCR444;
+		break;
+	default:
+		requestPixelEncoding = PIXEL_ENCODING_RGB;
+		break;
+	}
+
+	if ((requestColorDepth != COLOR_DEPTH_UNDEFINED
+		&& pipe_ctx->stream->timing.display_color_depth != requestColorDepth)
+		|| (requestPixelEncoding != PIXEL_ENCODING_UNDEFINED
+		&& pipe_ctx->stream->timing.pixel_encoding != requestPixelEncoding)) {
+		drm_dbg(dev,
+			"original bpc %d pix encoding %d, changing to %d  %d\n",
+			pipe_ctx->stream->timing.display_color_depth,
+			pipe_ctx->stream->timing.pixel_encoding,
+			requestColorDepth,
+			requestPixelEncoding);
+		pipe_ctx->stream->timing.display_color_depth = requestColorDepth;
+		pipe_ctx->stream->timing.pixel_encoding = requestPixelEncoding;
+
+		dc_link_update_dsc_config(pipe_ctx);
+
+		aconnector->timing_changed = true;
+		/* store current timing */
+		if (aconnector->timing_requested)
+			*aconnector->timing_requested = pipe_ctx->stream->timing;
+		else
+			drm_err(dev, "timing storage failed\n");
+
+	}
+
+	pipe_ctx->stream->test_pattern.type = test_pattern;
+	pipe_ctx->stream->test_pattern.color_space = test_pattern_color_space;
+
+	/* Temp W/A for compliance test failure */
+	dc_state->bw_ctx.bw.dcn.clk.p_state_change_support = false;
+	dc_state->bw_ctx.bw.dcn.clk.dramclk_khz = clk_mgr->dc_mode_softmax_enabled ?
+		clk_mgr->bw_params->dc_mode_softmax_memclk : clk_mgr->bw_params->max_memclk_mhz;
+	dc_state->bw_ctx.bw.dcn.clk.idle_dramclk_khz = dc_state->bw_ctx.bw.dcn.clk.dramclk_khz;
+	ctx->dc->clk_mgr->funcs->update_clocks(
+			ctx->dc->clk_mgr,
+			dc_state,
+			false);
+
+	dc_link_dp_set_test_pattern(
+		(struct dc_link *) link,
+		test_pattern,
+		test_pattern_color_space,
+		NULL,
+		NULL,
+		0);
+
+	return false;
 }
 
+void dm_set_phyd32clk(struct dc_context *ctx, int freq_khz)
+{
+       // TODO
+}
 
 void dm_helpers_enable_periodic_detection(struct dc_context *ctx, bool enable)
 {
-	/* TODO: add peridic detection implementation */
+	struct amdgpu_device *adev = ctx->driver_context;
+
+	if (adev->dm.idle_workqueue) {
+		adev->dm.idle_workqueue->enable = enable;
+		if (enable && !adev->dm.idle_workqueue->running && amdgpu_dm_is_headless(adev))
+			schedule_work(&adev->dm.idle_workqueue->work);
+	}
+}
+
+void dm_helpers_dp_mst_update_branch_bandwidth(
+		struct dc_context *ctx,
+		struct dc_link *link)
+{
+	// TODO
+}
+
+static bool dm_is_freesync_pcon_whitelist(const uint32_t branch_dev_id)
+{
+	bool ret_val = false;
+
+	switch (branch_dev_id) {
+	case DP_BRANCH_DEVICE_ID_0060AD:
+	case DP_BRANCH_DEVICE_ID_00E04C:
+	case DP_BRANCH_DEVICE_ID_90CC24:
+		ret_val = true;
+		break;
+	default:
+		break;
+	}
+
+	return ret_val;
+}
+
+enum adaptive_sync_type dm_get_adaptive_sync_support_type(struct dc_link *link)
+{
+	struct dpcd_caps *dpcd_caps = &link->dpcd_caps;
+	enum adaptive_sync_type as_type = ADAPTIVE_SYNC_TYPE_NONE;
+
+	switch (dpcd_caps->dongle_type) {
+	case DISPLAY_DONGLE_DP_HDMI_CONVERTER:
+		if (dpcd_caps->adaptive_sync_caps.dp_adap_sync_caps.bits.ADAPTIVE_SYNC_SDP_SUPPORT == true &&
+			dpcd_caps->allow_invalid_MSA_timing_param == true &&
+			dm_is_freesync_pcon_whitelist(dpcd_caps->branch_dev_id))
+			as_type = FREESYNC_TYPE_PCON_IN_WHITELIST;
+		break;
+	default:
+		break;
+	}
+
+	return as_type;
+}
+
+bool dm_helpers_is_fullscreen(struct dc_context *ctx, struct dc_stream_state *stream)
+{
+	// TODO
+	return false;
+}
+
+bool dm_helpers_is_hdr_on(struct dc_context *ctx, struct dc_stream_state *stream)
+{
+	// TODO
+	return false;
 }
-#endif
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_irq.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_irq.c
index 4aba0e8c84f8..0a2a3f233a0e 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_irq.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_irq.c
@@ -1,3 +1,4 @@
+// SPDX-License-Identifier: MIT
 /*
  * Copyright 2015 Advanced Micro Devices, Inc.
  *
@@ -120,7 +121,8 @@ static void dm_irq_work_func(struct work_struct *work)
 
 	/* Call a DAL subcomponent which registered for interrupt notification
 	 * at INTERRUPT_LOW_IRQ_CONTEXT.
-	 * (The most common use is HPD interrupt) */
+	 * (The most common use is HPD interrupt)
+	 */
 }
 
 /*
@@ -172,7 +174,8 @@ static struct list_head *remove_irq_handler(struct amdgpu_device *adev,
 
 	if (handler_removed == false) {
 		/* Not necessarily an error - caller may not
-		 * know the context. */
+		 * know the context.
+		 */
 		return NULL;
 	}
 
@@ -261,7 +264,7 @@ validate_irq_registration_params(struct dc_interrupt_params *int_params,
 static bool validate_irq_unregistration_params(enum dc_irq_source irq_source,
 					       irq_handler_idx handler_idx)
 {
-	if (DAL_INVALID_IRQ_HANDLER_IDX == handler_idx) {
+	if (handler_idx == DAL_INVALID_IRQ_HANDLER_IDX) {
 		DRM_ERROR("DM_IRQ: invalid handler_idx==NULL!\n");
 		return false;
 	}
@@ -343,7 +346,8 @@ void *amdgpu_dm_irq_register_interrupt(struct amdgpu_device *adev,
 	/* This pointer will be stored by code which requested interrupt
 	 * registration.
 	 * The same pointer will be needed in order to unregister the
-	 * interrupt. */
+	 * interrupt.
+	 */
 
 	DRM_DEBUG_KMS(
 		"DM_IRQ: added irq handler: %p for: dal_src=%d, irq context=%d\n",
@@ -390,7 +394,8 @@ void amdgpu_dm_irq_unregister_interrupt(struct amdgpu_device *adev,
 
 	if (handler_list == NULL) {
 		/* If we got here, it means we searched all irq contexts
-		 * for this irq source, but the handler was not found. */
+		 * for this irq source, but the handler was not found.
+		 */
 		DRM_ERROR(
 		"DM_IRQ: failed to find irq handler:%p for irq_source:%d!\n",
 			ih, irq_source);
@@ -450,7 +455,8 @@ void amdgpu_dm_irq_fini(struct amdgpu_device *adev)
 		DM_IRQ_TABLE_LOCK(adev, irq_table_flags);
 		/* The handler was removed from the table,
 		 * it means it is safe to flush all the 'work'
-		 * (because no code can schedule a new one). */
+		 * (because no code can schedule a new one).
+		 */
 		lh = &adev->dm.irq_handler_list_low_tab[src];
 		DM_IRQ_TABLE_UNLOCK(adev, irq_table_flags);
 
@@ -468,8 +474,9 @@ void amdgpu_dm_irq_fini(struct amdgpu_device *adev)
 	unregister_all_irq_handlers(adev);
 }
 
-int amdgpu_dm_irq_suspend(struct amdgpu_device *adev)
+void amdgpu_dm_irq_suspend(struct amdgpu_device *adev)
 {
+	struct drm_device *dev = adev_to_drm(adev);
 	int src;
 	struct list_head *hnd_list_h;
 	struct list_head *hnd_list_l;
@@ -494,7 +501,7 @@ int amdgpu_dm_irq_suspend(struct amdgpu_device *adev)
 		DM_IRQ_TABLE_UNLOCK(adev, irq_table_flags);
 
 		if (!list_empty(hnd_list_l)) {
-			list_for_each_safe (entry, tmp, hnd_list_l) {
+			list_for_each_safe(entry, tmp, hnd_list_l) {
 				handler = list_entry(
 					entry,
 					struct amdgpu_dm_irq_handler_data,
@@ -506,10 +513,12 @@ int amdgpu_dm_irq_suspend(struct amdgpu_device *adev)
 	}
 
 	DM_IRQ_TABLE_UNLOCK(adev, irq_table_flags);
-	return 0;
+
+	if (dev->mode_config.poll_enabled)
+		drm_kms_helper_poll_disable(dev);
 }
 
-int amdgpu_dm_irq_resume_early(struct amdgpu_device *adev)
+void amdgpu_dm_irq_resume_early(struct amdgpu_device *adev)
 {
 	int src;
 	struct list_head *hnd_list_h, *hnd_list_l;
@@ -517,7 +526,7 @@ int amdgpu_dm_irq_resume_early(struct amdgpu_device *adev)
 
 	DM_IRQ_TABLE_LOCK(adev, irq_table_flags);
 
-	DRM_DEBUG_KMS("DM_IRQ: early resume\n");
+	drm_dbg(adev_to_drm(adev), "DM_IRQ: early resume\n");
 
 	/* re-enable short pulse interrupts HW interrupt */
 	for (src = DC_IRQ_SOURCE_HPD1RX; src <= DC_IRQ_SOURCE_HPD6RX; src++) {
@@ -528,19 +537,18 @@ int amdgpu_dm_irq_resume_early(struct amdgpu_device *adev)
 	}
 
 	DM_IRQ_TABLE_UNLOCK(adev, irq_table_flags);
-
-	return 0;
 }
 
-int amdgpu_dm_irq_resume_late(struct amdgpu_device *adev)
+void amdgpu_dm_irq_resume_late(struct amdgpu_device *adev)
 {
+	struct drm_device *dev = adev_to_drm(adev);
 	int src;
 	struct list_head *hnd_list_h, *hnd_list_l;
 	unsigned long irq_table_flags;
 
 	DM_IRQ_TABLE_LOCK(adev, irq_table_flags);
 
-	DRM_DEBUG_KMS("DM_IRQ: resume\n");
+	drm_dbg(adev_to_drm(adev), "DM_IRQ: resume\n");
 
 	/**
 	 * Renable HW interrupt  for HPD and only since FLIP and VBLANK
@@ -554,7 +562,9 @@ int amdgpu_dm_irq_resume_late(struct amdgpu_device *adev)
 	}
 
 	DM_IRQ_TABLE_UNLOCK(adev, irq_table_flags);
-	return 0;
+
+	if (dev->mode_config.poll_enabled)
+		drm_kms_helper_poll_enable(dev);
 }
 
 /*
@@ -571,7 +581,7 @@ static void amdgpu_dm_irq_schedule_work(struct amdgpu_device *adev,
 	if (list_empty(handler_list))
 		return;
 
-	list_for_each_entry (handler_data, handler_list, list) {
+	list_for_each_entry(handler_data, handler_list, list) {
 		if (queue_work(system_highpri_wq, &handler_data->work)) {
 			work_queued = true;
 			break;
@@ -627,7 +637,8 @@ static void amdgpu_dm_irq_immediate_work(struct amdgpu_device *adev,
 			    &adev->dm.irq_handler_list_high_tab[irq_source],
 			    list) {
 		/* Call a subcomponent which registered for immediate
-		 * interrupt notification */
+		 * interrupt notification
+		 */
 		handler_data->handler(handler_data->handler_arg);
 	}
 
@@ -664,7 +675,7 @@ static int amdgpu_dm_irq_handler(struct amdgpu_device *adev,
 	return 0;
 }
 
-static enum dc_irq_source amdgpu_dm_hpd_to_dal_irq_source(unsigned type)
+static enum dc_irq_source amdgpu_dm_hpd_to_dal_irq_source(unsigned int type)
 {
 	switch (type) {
 	case AMDGPU_HPD_1:
@@ -686,7 +697,7 @@ static enum dc_irq_source amdgpu_dm_hpd_to_dal_irq_source(unsigned type)
 
 static int amdgpu_dm_set_hpd_irq_state(struct amdgpu_device *adev,
 				       struct amdgpu_irq_src *source,
-				       unsigned type,
+				       unsigned int type,
 				       enum amdgpu_interrupt_state state)
 {
 	enum dc_irq_source src = amdgpu_dm_hpd_to_dal_irq_source(type);
@@ -698,14 +709,14 @@ static int amdgpu_dm_set_hpd_irq_state(struct amdgpu_device *adev,
 
 static inline int dm_irq_state(struct amdgpu_device *adev,
 			       struct amdgpu_irq_src *source,
-			       unsigned crtc_id,
+			       unsigned int crtc_id,
 			       enum amdgpu_interrupt_state state,
 			       const enum irq_type dal_irq_type,
 			       const char *func)
 {
 	bool st;
 	enum dc_irq_source irq_source;
-
+	struct dc *dc = adev->dm.dc;
 	struct amdgpu_crtc *acrtc = adev->mode_info.crtcs[crtc_id];
 
 	if (!acrtc) {
@@ -723,13 +734,16 @@ static inline int dm_irq_state(struct amdgpu_device *adev,
 
 	st = (state == AMDGPU_IRQ_STATE_ENABLE);
 
+	if (dc && dc->caps.ips_support && dc->idle_optimizations_allowed)
+		dc_allow_idle_optimizations(dc, false);
+
 	dc_interrupt_set(adev->dm.dc, irq_source, st);
 	return 0;
 }
 
 static int amdgpu_dm_set_pflip_irq_state(struct amdgpu_device *adev,
 					 struct amdgpu_irq_src *source,
-					 unsigned crtc_id,
+					 unsigned int crtc_id,
 					 enum amdgpu_interrupt_state state)
 {
 	return dm_irq_state(
@@ -743,7 +757,7 @@ static int amdgpu_dm_set_pflip_irq_state(struct amdgpu_device *adev,
 
 static int amdgpu_dm_set_crtc_irq_state(struct amdgpu_device *adev,
 					struct amdgpu_irq_src *source,
-					unsigned crtc_id,
+					unsigned int crtc_id,
 					enum amdgpu_interrupt_state state)
 {
 	return dm_irq_state(
@@ -885,27 +899,68 @@ void amdgpu_dm_hpd_init(struct amdgpu_device *adev)
 	struct drm_device *dev = adev_to_drm(adev);
 	struct drm_connector *connector;
 	struct drm_connector_list_iter iter;
+	int irq_type;
+	int i;
+	bool use_polling = false;
+
+	/* First, clear all hpd and hpdrx interrupts */
+	for (i = DC_IRQ_SOURCE_HPD1; i <= DC_IRQ_SOURCE_HPD6RX; i++) {
+		if (!dc_interrupt_set(adev->dm.dc, i, false))
+			drm_err(dev, "Failed to clear hpd(rx) source=%d on init\n",
+				i);
+	}
 
 	drm_connector_list_iter_begin(dev, &iter);
 	drm_for_each_connector_iter(connector, &iter) {
-		struct amdgpu_dm_connector *amdgpu_dm_connector =
-				to_amdgpu_dm_connector(connector);
+		struct amdgpu_dm_connector *amdgpu_dm_connector;
+		const struct dc_link *dc_link;
 
-		const struct dc_link *dc_link = amdgpu_dm_connector->dc_link;
+		use_polling |= connector->polled != DRM_CONNECTOR_POLL_HPD;
 
-		if (DC_IRQ_SOURCE_INVALID != dc_link->irq_source_hpd) {
-			dc_interrupt_set(adev->dm.dc,
-					dc_link->irq_source_hpd,
-					true);
+		if (connector->connector_type == DRM_MODE_CONNECTOR_WRITEBACK)
+			continue;
+
+		amdgpu_dm_connector = to_amdgpu_dm_connector(connector);
+
+		dc_link = amdgpu_dm_connector->dc_link;
+
+		/*
+		 * Get a base driver irq reference for hpd ints for the lifetime
+		 * of dm. Note that only hpd interrupt types are registered with
+		 * base driver; hpd_rx types aren't. IOW, amdgpu_irq_get/put on
+		 * hpd_rx isn't available. DM currently controls hpd_rx
+		 * explicitly with dc_interrupt_set()
+		 */
+		if (dc_link->irq_source_hpd != DC_IRQ_SOURCE_INVALID) {
+			irq_type = dc_link->irq_source_hpd - DC_IRQ_SOURCE_HPD1;
+			/*
+			 * TODO: There's a mismatch between mode_info.num_hpd
+			 * and what bios reports as the # of connectors with hpd
+			 * sources. Since the # of hpd source types registered
+			 * with base driver == mode_info.num_hpd, we have to
+			 * fallback to dc_interrupt_set for the remaining types.
+			 */
+			if (irq_type < adev->mode_info.num_hpd) {
+				if (amdgpu_irq_get(adev, &adev->hpd_irq, irq_type))
+					drm_err(dev, "DM_IRQ: Failed get HPD for source=%d)!\n",
+						dc_link->irq_source_hpd);
+			} else {
+				dc_interrupt_set(adev->dm.dc,
+						 dc_link->irq_source_hpd,
+						 true);
+			}
 		}
 
-		if (DC_IRQ_SOURCE_INVALID != dc_link->irq_source_hpd_rx) {
+		if (dc_link->irq_source_hpd_rx != DC_IRQ_SOURCE_INVALID) {
 			dc_interrupt_set(adev->dm.dc,
 					dc_link->irq_source_hpd_rx,
 					true);
 		}
 	}
 	drm_connector_list_iter_end(&iter);
+
+	if (use_polling)
+		drm_kms_helper_poll_init(dev);
 }
 
 /**
@@ -921,20 +976,42 @@ void amdgpu_dm_hpd_fini(struct amdgpu_device *adev)
 	struct drm_device *dev = adev_to_drm(adev);
 	struct drm_connector *connector;
 	struct drm_connector_list_iter iter;
+	int irq_type;
 
 	drm_connector_list_iter_begin(dev, &iter);
 	drm_for_each_connector_iter(connector, &iter) {
-		struct amdgpu_dm_connector *amdgpu_dm_connector =
-				to_amdgpu_dm_connector(connector);
-		const struct dc_link *dc_link = amdgpu_dm_connector->dc_link;
+		struct amdgpu_dm_connector *amdgpu_dm_connector;
+		const struct dc_link *dc_link;
+
+		if (connector->connector_type == DRM_MODE_CONNECTOR_WRITEBACK)
+			continue;
 
-		dc_interrupt_set(adev->dm.dc, dc_link->irq_source_hpd, false);
+		amdgpu_dm_connector = to_amdgpu_dm_connector(connector);
+		dc_link = amdgpu_dm_connector->dc_link;
+
+		if (dc_link->irq_source_hpd != DC_IRQ_SOURCE_INVALID) {
+			irq_type = dc_link->irq_source_hpd - DC_IRQ_SOURCE_HPD1;
+
+			/* TODO: See same TODO in amdgpu_dm_hpd_init() */
+			if (irq_type < adev->mode_info.num_hpd) {
+				if (amdgpu_irq_put(adev, &adev->hpd_irq, irq_type))
+					drm_err(dev, "DM_IRQ: Failed put HPD for source=%d!\n",
+						dc_link->irq_source_hpd);
+			} else {
+				dc_interrupt_set(adev->dm.dc,
+						 dc_link->irq_source_hpd,
+						 false);
+			}
+		}
 
-		if (DC_IRQ_SOURCE_INVALID != dc_link->irq_source_hpd_rx) {
+		if (dc_link->irq_source_hpd_rx != DC_IRQ_SOURCE_INVALID) {
 			dc_interrupt_set(adev->dm.dc,
 					dc_link->irq_source_hpd_rx,
 					false);
 		}
 	}
 	drm_connector_list_iter_end(&iter);
+
+	if (dev->mode_config.poll_enabled)
+		drm_kms_helper_poll_fini(dev);
 }
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_irq.h b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_irq.h
index 2349238a626b..4f6b58f4f90d 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_irq.h
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_irq.h
@@ -1,3 +1,4 @@
+/* SPDX-License-Identifier: MIT */
 /*
  * Copyright 2015 Advanced Micro Devices, Inc.
  *
@@ -90,14 +91,14 @@ void amdgpu_dm_hpd_fini(struct amdgpu_device *adev);
  * amdgpu_dm_irq_suspend - disable ASIC interrupt during suspend.
  *
  */
-int amdgpu_dm_irq_suspend(struct amdgpu_device *adev);
+void amdgpu_dm_irq_suspend(struct amdgpu_device *adev);
 
 /**
  * amdgpu_dm_irq_resume_early - enable HPDRX ASIC interrupts during resume.
  * amdgpu_dm_irq_resume - enable ASIC interrupt during resume.
  *
  */
-int amdgpu_dm_irq_resume_early(struct amdgpu_device *adev);
-int amdgpu_dm_irq_resume_late(struct amdgpu_device *adev);
+void amdgpu_dm_irq_resume_early(struct amdgpu_device *adev);
+void amdgpu_dm_irq_resume_late(struct amdgpu_device *adev);
 
 #endif /* __AMDGPU_DM_IRQ_H__ */
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_irq_params.h b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_irq_params.h
index 79b5f9999fec..3c9995275cbd 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_irq_params.h
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_irq_params.h
@@ -1,3 +1,4 @@
+/* SPDX-License-Identifier: MIT */
 /*
  * Copyright 2020 Advanced Micro Devices, Inc.
  *
@@ -33,13 +34,15 @@ struct dm_irq_params {
 	struct mod_vrr_params vrr_params;
 	struct dc_stream_state *stream;
 	int active_planes;
-	bool allow_psr_entry;
+	bool allow_sr_entry;
 	struct mod_freesync_config freesync_config;
 
 #ifdef CONFIG_DEBUG_FS
 	enum amdgpu_dm_pipe_crc_source crc_src;
 #ifdef CONFIG_DRM_AMD_SECURE_DISPLAY
-	struct crc_window_parm crc_window;
+	struct crc_window_param window_param[MAX_CRC_WINDOW_NUM];
+	/* At least one CRC window is activated or not*/
+	bool crc_window_activated;
 #endif
 #endif
 };
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
index cc34a35d0bcb..dbd1da4d85d3 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
@@ -1,3 +1,4 @@
+// SPDX-License-Identifier: MIT
 /*
  * Copyright 2012-15 Advanced Micro Devices, Inc.
  *
@@ -23,38 +24,46 @@
  *
  */
 
+#include <linux/vmalloc.h>
+#include <drm/display/drm_dp_helper.h>
+#include <drm/display/drm_dp_mst_helper.h>
 #include <drm/drm_atomic.h>
 #include <drm/drm_atomic_helper.h>
-#include <drm/drm_dp_mst_helper.h>
-#include <drm/drm_dp_helper.h>
+#include <drm/drm_fixed.h>
+#include <drm/drm_edid.h>
 #include "dm_services.h"
 #include "amdgpu.h"
 #include "amdgpu_dm.h"
 #include "amdgpu_dm_mst_types.h"
+#include "amdgpu_dm_hdcp.h"
 
 #include "dc.h"
 #include "dm_helpers.h"
 
-#include "dc_link_ddc.h"
 #include "ddc_service_types.h"
 #include "dpcd_defs.h"
 
-#include "i2caux_interface.h"
 #include "dmub_cmd.h"
 #if defined(CONFIG_DEBUG_FS)
 #include "amdgpu_dm_debugfs.h"
 #endif
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-#include "dc/dcn20/dcn20_resource.h"
-#endif
+#include "dc/resource/dcn20/dcn20_resource.h"
+
+#define PEAK_FACTOR_X1000 1006
 
+/*
+ * This function handles both native AUX and I2C-Over-AUX transactions.
+ */
 static ssize_t dm_dp_aux_transfer(struct drm_dp_aux *aux,
 				  struct drm_dp_aux_msg *msg)
 {
 	ssize_t result = 0;
 	struct aux_payload payload;
 	enum aux_return_code_type operation_result;
+	struct amdgpu_device *adev;
+	struct ddc_service *ddc;
+	uint8_t copy[16];
 
 	if (WARN_ON(msg->size > 16))
 		return -E2BIG;
@@ -70,13 +79,43 @@ static ssize_t dm_dp_aux_transfer(struct drm_dp_aux *aux,
 			(msg->request & DP_AUX_I2C_WRITE_STATUS_UPDATE) != 0;
 	payload.defer_delay = 0;
 
+	if (payload.write) {
+		memcpy(copy, msg->buffer, msg->size);
+		payload.data = copy;
+	}
+
 	result = dc_link_aux_transfer_raw(TO_DM_AUX(aux)->ddc_service, &payload,
 				      &operation_result);
 
-	if (payload.write && result >= 0)
-		result = msg->size;
+	/*
+	 * w/a on certain intel platform where hpd is unexpected to pull low during
+	 * 1st sideband message transaction by return AUX_RET_ERROR_HPD_DISCON
+	 * aux transaction is succuess in such case, therefore bypass the error
+	 */
+	ddc = TO_DM_AUX(aux)->ddc_service;
+	adev = ddc->ctx->driver_context;
+	if (adev->dm.aux_hpd_discon_quirk) {
+		if (msg->address == DP_SIDEBAND_MSG_DOWN_REQ_BASE &&
+			operation_result == AUX_RET_ERROR_HPD_DISCON) {
+			result = msg->size;
+			operation_result = AUX_RET_SUCCESS;
+		}
+	}
+
+	/*
+	 * result equals to 0 includes the cases of AUX_DEFER/I2C_DEFER
+	 */
+	if (payload.write && result >= 0) {
+		if (result) {
+			/*one byte indicating partially written bytes*/
+			drm_dbg_dp(adev_to_drm(adev), "AUX partially written\n");
+			result = payload.data[0];
+		} else if (!payload.reply[0])
+			/*I2C_ACK|AUX_ACK*/
+			result = msg->size;
+	}
 
-	if (result < 0)
+	if (result < 0) {
 		switch (operation_result) {
 		case AUX_RET_SUCCESS:
 			break;
@@ -95,6 +134,13 @@ static ssize_t dm_dp_aux_transfer(struct drm_dp_aux *aux,
 			break;
 		}
 
+		drm_dbg_dp(adev_to_drm(adev), "DP AUX transfer fail:%d\n", operation_result);
+	}
+
+	if (payload.reply[0])
+		drm_dbg_dp(adev_to_drm(adev), "AUX reply command not ACK: 0x%02x.",
+			payload.reply[0]);
+
 	return result;
 }
 
@@ -110,10 +156,10 @@ dm_dp_mst_connector_destroy(struct drm_connector *connector)
 		dc_sink_release(aconnector->dc_sink);
 	}
 
-	kfree(aconnector->edid);
+	drm_edid_free(aconnector->drm_edid);
 
 	drm_connector_cleanup(connector);
-	drm_dp_mst_put_port_malloc(aconnector->port);
+	drm_dp_mst_put_port_malloc(aconnector->mst_output_port);
 	kfree(aconnector);
 }
 
@@ -125,7 +171,7 @@ amdgpu_dm_mst_connector_late_register(struct drm_connector *connector)
 	int r;
 
 	r = drm_dp_mst_connector_late_register(connector,
-					       amdgpu_dm_connector->port);
+					       amdgpu_dm_connector->mst_output_port);
 	if (r < 0)
 		return r;
 
@@ -136,14 +182,49 @@ amdgpu_dm_mst_connector_late_register(struct drm_connector *connector)
 	return 0;
 }
 
+
+static inline void
+amdgpu_dm_mst_reset_mst_connector_setting(struct amdgpu_dm_connector *aconnector)
+{
+	aconnector->drm_edid = NULL;
+	aconnector->dsc_aux = NULL;
+	aconnector->mst_output_port->passthrough_aux = NULL;
+	aconnector->mst_local_bw = 0;
+	aconnector->vc_full_pbn = 0;
+}
+
 static void
 amdgpu_dm_mst_connector_early_unregister(struct drm_connector *connector)
 {
-	struct amdgpu_dm_connector *amdgpu_dm_connector =
+	struct amdgpu_dm_connector *aconnector =
 		to_amdgpu_dm_connector(connector);
-	struct drm_dp_mst_port *port = amdgpu_dm_connector->port;
+	struct drm_dp_mst_port *port = aconnector->mst_output_port;
+	struct amdgpu_dm_connector *root = aconnector->mst_root;
+	struct dc_link *dc_link = aconnector->dc_link;
+	struct dc_sink *dc_sink = aconnector->dc_sink;
 
 	drm_dp_mst_connector_early_unregister(connector, port);
+
+	/*
+	 * Release dc_sink for connector which its attached port is
+	 * no longer in the mst topology
+	 */
+	drm_modeset_lock(&root->mst_mgr.base.lock, NULL);
+	if (dc_sink) {
+		if (dc_link->sink_count)
+			dc_link_remove_remote_sink(dc_link, dc_sink);
+
+		drm_dbg_dp(connector->dev,
+			   "DM_MST: remove remote sink 0x%p, %d remaining\n",
+			   dc_sink, dc_link->sink_count);
+
+		dc_sink_release(dc_sink);
+		aconnector->dc_sink = NULL;
+		amdgpu_dm_mst_reset_mst_connector_setting(aconnector);
+	}
+
+	aconnector->mst_status = MST_STATUS_DEFAULT;
+	drm_modeset_unlock(&root->mst_mgr.base.lock);
 }
 
 static const struct drm_connector_funcs dm_dp_mst_connector_funcs = {
@@ -158,8 +239,7 @@ static const struct drm_connector_funcs dm_dp_mst_connector_funcs = {
 	.early_unregister = amdgpu_dm_mst_connector_early_unregister,
 };
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-static bool needs_dsc_aux_workaround(struct dc_link *link)
+bool needs_dsc_aux_workaround(struct dc_link *link)
 {
 	if (link->dpcd_caps.branch_dev_id == DP_BRANCH_DEVICE_ID_90CC24 &&
 	    (link->dpcd_caps.dpcd_rev.raw == DPCD_REV_14 || link->dpcd_caps.dpcd_rev.raw == DPCD_REV_12) &&
@@ -169,10 +249,26 @@ static bool needs_dsc_aux_workaround(struct dc_link *link)
 	return false;
 }
 
+#if defined(CONFIG_DRM_AMD_DC_FP)
+static bool is_synaptics_cascaded_panamera(struct dc_link *link, struct drm_dp_mst_port *port)
+{
+	u8 branch_vendor_data[4] = { 0 }; // Vendor data 0x50C ~ 0x50F
+
+	if (drm_dp_dpcd_read(port->mgr->aux, DP_BRANCH_VENDOR_SPECIFIC_START, &branch_vendor_data, 4) == 4) {
+		if (link->dpcd_caps.branch_dev_id == DP_BRANCH_DEVICE_ID_90CC24 &&
+				IS_SYNAPTICS_CASCADED_PANAMERA(link->dpcd_caps.branch_dev_name, branch_vendor_data)) {
+			DRM_INFO("Synaptics Cascaded MST hub\n");
+			return true;
+		}
+	}
+
+	return false;
+}
+
 static bool validate_dsc_caps_on_connector(struct amdgpu_dm_connector *aconnector)
 {
 	struct dc_sink *dc_sink = aconnector->dc_sink;
-	struct drm_dp_mst_port *port = aconnector->port;
+	struct drm_dp_mst_port *port = aconnector->mst_output_port;
 	u8 dsc_caps[16] = { 0 };
 	u8 dsc_branch_dec_caps_raw[3] = { 0 };	// DSC branch decoder caps 0xA0 ~ 0xA2
 	u8 *dsc_branch_dec_caps = NULL;
@@ -190,7 +286,11 @@ static bool validate_dsc_caps_on_connector(struct amdgpu_dm_connector *aconnecto
 	 */
 	if (!aconnector->dsc_aux && !port->parent->port_parent &&
 	    needs_dsc_aux_workaround(aconnector->dc_link))
-		aconnector->dsc_aux = &aconnector->mst_port->dm_dp_aux.aux;
+		aconnector->dsc_aux = &aconnector->mst_root->dm_dp_aux.aux;
+
+	/* synaptics cascaded MST hub case */
+	if (is_synaptics_cascaded_panamera(aconnector->dc_link, port))
+		aconnector->dsc_aux = port->mgr->aux;
 
 	if (!aconnector->dsc_aux)
 		return false;
@@ -211,6 +311,53 @@ static bool validate_dsc_caps_on_connector(struct amdgpu_dm_connector *aconnecto
 }
 #endif
 
+static bool retrieve_downstream_port_device(struct amdgpu_dm_connector *aconnector)
+{
+	union dp_downstream_port_present ds_port_present;
+
+	if (!aconnector->dsc_aux)
+		return false;
+
+	if (drm_dp_dpcd_read(aconnector->dsc_aux, DP_DOWNSTREAMPORT_PRESENT, &ds_port_present, 1) < 0) {
+		DRM_INFO("Failed to read downstream_port_present 0x05 from DFP of branch device\n");
+		return false;
+	}
+
+	aconnector->mst_downstream_port_present = ds_port_present;
+	DRM_INFO("Downstream port present %d, type %d\n",
+			ds_port_present.fields.PORT_PRESENT, ds_port_present.fields.PORT_TYPE);
+
+	return true;
+}
+
+static bool retrieve_branch_specific_data(struct amdgpu_dm_connector *aconnector)
+{
+	struct drm_connector *connector = &aconnector->base;
+	struct drm_dp_mst_port *port = aconnector->mst_output_port;
+	struct drm_dp_mst_port *port_parent;
+	struct drm_dp_aux *immediate_upstream_aux;
+	struct drm_dp_desc branch_desc;
+
+	if (!port->parent)
+		return false;
+
+	port_parent = port->parent->port_parent;
+
+	immediate_upstream_aux = port_parent ? &port_parent->aux : port->mgr->aux;
+
+	if (drm_dp_read_desc(immediate_upstream_aux, &branch_desc, true))
+		return false;
+
+	aconnector->branch_ieee_oui = (branch_desc.ident.oui[0] << 16) +
+				      (branch_desc.ident.oui[1] << 8) +
+				      (branch_desc.ident.oui[2]);
+
+	drm_dbg_dp(port->aux.drm_dev, "MST branch oui 0x%x detected at %s\n",
+		   aconnector->branch_ieee_oui, connector->name);
+
+	return true;
+}
+
 static int dm_dp_mst_get_modes(struct drm_connector *connector)
 {
 	struct amdgpu_dm_connector *aconnector = to_amdgpu_dm_connector(connector);
@@ -219,12 +366,18 @@ static int dm_dp_mst_get_modes(struct drm_connector *connector)
 	if (!aconnector)
 		return drm_add_edid_modes(connector, NULL);
 
-	if (!aconnector->edid) {
-		struct edid *edid;
-		edid = drm_dp_mst_get_edid(connector, &aconnector->mst_port->mst_mgr, aconnector->port);
+	if (!aconnector->drm_edid) {
+		const struct drm_edid *drm_edid;
+
+		drm_edid = drm_dp_mst_edid_read(connector,
+						&aconnector->mst_root->mst_mgr,
+						aconnector->mst_output_port);
 
-		if (!edid) {
-			drm_connector_update_edid_property(
+		if (!drm_edid) {
+			amdgpu_dm_set_mst_status(&aconnector->mst_status,
+			MST_REMOTE_EDID, false);
+
+			drm_edid_connector_update(
 				&aconnector->base,
 				NULL);
 
@@ -246,6 +399,11 @@ static int dm_dp_mst_get_modes(struct drm_connector *connector)
 					return 0;
 				}
 
+				drm_dbg_dp(connector->dev,
+					   "DM_MST: add remote sink 0x%p, %d remaining\n",
+					   dc_sink,
+					   aconnector->dc_link->sink_count);
+
 				dc_sink->priv = aconnector;
 				aconnector->dc_sink = dc_sink;
 			}
@@ -253,7 +411,9 @@ static int dm_dp_mst_get_modes(struct drm_connector *connector)
 			return ret;
 		}
 
-		aconnector->edid = edid;
+		aconnector->drm_edid = drm_edid;
+		amdgpu_dm_set_mst_status(&aconnector->mst_status,
+			MST_REMOTE_EDID, true);
 	}
 
 	if (aconnector->dc_sink && aconnector->dc_sink->sink_signal == SIGNAL_TYPE_VIRTUAL) {
@@ -266,10 +426,13 @@ static int dm_dp_mst_get_modes(struct drm_connector *connector)
 		struct dc_sink_init_data init_params = {
 				.link = aconnector->dc_link,
 				.sink_signal = SIGNAL_TYPE_DISPLAY_PORT_MST };
+		const struct edid *edid;
+
+		edid = drm_edid_raw(aconnector->drm_edid); // FIXME: Get rid of drm_edid_raw()
 		dc_sink = dc_link_add_remote_sink(
 			aconnector->dc_link,
-			(uint8_t *)aconnector->edid,
-			(aconnector->edid->extensions + 1) * EDID_LENGTH,
+			(uint8_t *)edid,
+			(edid->extensions + 1) * EDID_LENGTH,
 			&init_params);
 
 		if (!dc_sink) {
@@ -277,26 +440,57 @@ static int dm_dp_mst_get_modes(struct drm_connector *connector)
 			return 0;
 		}
 
+		drm_dbg_dp(connector->dev,
+			   "DM_MST: add remote sink 0x%p, %d remaining\n",
+			   dc_sink, aconnector->dc_link->sink_count);
+
 		dc_sink->priv = aconnector;
 		/* dc_link_add_remote_sink returns a new reference */
 		aconnector->dc_sink = dc_sink;
 
+		/* when display is unplugged from mst hub, connctor will be
+		 * destroyed within dm_dp_mst_connector_destroy. connector
+		 * hdcp perperties, like type, undesired, desired, enabled,
+		 * will be lost. So, save hdcp properties into hdcp_work within
+		 * amdgpu_dm_atomic_commit_tail. if the same display is
+		 * plugged back with same display index, its hdcp properties
+		 * will be retrieved from hdcp_work within dm_dp_mst_get_modes
+		 */
+		if (aconnector->dc_sink && connector->state) {
+			struct drm_device *dev = connector->dev;
+			struct amdgpu_device *adev = drm_to_adev(dev);
+
+			if (adev->dm.hdcp_workqueue) {
+				struct hdcp_workqueue *hdcp_work = adev->dm.hdcp_workqueue;
+				struct hdcp_workqueue *hdcp_w =
+					&hdcp_work[aconnector->dc_link->link_index];
+
+				connector->state->hdcp_content_type =
+				hdcp_w->hdcp_content_type[connector->index];
+				connector->state->content_protection =
+				hdcp_w->content_protection[connector->index];
+			}
+		}
+
 		if (aconnector->dc_sink) {
 			amdgpu_dm_update_freesync_caps(
-					connector, aconnector->edid);
+					connector, aconnector->drm_edid);
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
+#if defined(CONFIG_DRM_AMD_DC_FP)
 			if (!validate_dsc_caps_on_connector(aconnector))
 				memset(&aconnector->dc_sink->dsc_caps,
 				       0, sizeof(aconnector->dc_sink->dsc_caps));
 #endif
+
+			if (!retrieve_downstream_port_device(aconnector))
+				memset(&aconnector->mst_downstream_port_present,
+					0, sizeof(aconnector->mst_downstream_port_present));
 		}
 	}
 
-	drm_connector_update_edid_property(
-					&aconnector->base, aconnector->edid);
+	drm_edid_connector_update(&aconnector->base, aconnector->drm_edid);
 
-	ret = drm_add_edid_modes(connector, aconnector->edid);
+	ret = drm_edid_connector_add_modes(connector);
 
 	return ret;
 }
@@ -307,8 +501,7 @@ dm_mst_atomic_best_encoder(struct drm_connector *connector,
 {
 	struct drm_connector_state *connector_state = drm_atomic_get_new_connector_state(state,
 											 connector);
-	struct drm_device *dev = connector->dev;
-	struct amdgpu_device *adev = drm_to_adev(dev);
+	struct amdgpu_device *adev = drm_to_adev(connector->dev);
 	struct amdgpu_crtc *acrtc = to_amdgpu_crtc(connector_state->crtc);
 
 	return &adev->dm.mst_encoders[acrtc->crtc_id].base;
@@ -319,44 +512,80 @@ dm_dp_mst_detect(struct drm_connector *connector,
 		 struct drm_modeset_acquire_ctx *ctx, bool force)
 {
 	struct amdgpu_dm_connector *aconnector = to_amdgpu_dm_connector(connector);
-	struct amdgpu_dm_connector *master = aconnector->mst_port;
+	struct amdgpu_dm_connector *master = aconnector->mst_root;
+	struct drm_dp_mst_port *port = aconnector->mst_output_port;
+	int connection_status;
 
 	if (drm_connector_is_unregistered(connector))
 		return connector_status_disconnected;
 
-	return drm_dp_mst_detect_port(connector, ctx, &master->mst_mgr,
-				      aconnector->port);
-}
+	connection_status = drm_dp_mst_detect_port(connector, ctx, &master->mst_mgr,
+							aconnector->mst_output_port);
 
-static int dm_dp_mst_atomic_check(struct drm_connector *connector,
-				struct drm_atomic_state *state)
-{
-	struct drm_connector_state *new_conn_state =
-			drm_atomic_get_new_connector_state(state, connector);
-	struct drm_connector_state *old_conn_state =
-			drm_atomic_get_old_connector_state(state, connector);
-	struct amdgpu_dm_connector *aconnector = to_amdgpu_dm_connector(connector);
-	struct drm_crtc_state *new_crtc_state;
-	struct drm_dp_mst_topology_mgr *mst_mgr;
-	struct drm_dp_mst_port *mst_port;
+	if (port->pdt != DP_PEER_DEVICE_NONE && !port->dpcd_rev) {
+		uint8_t dpcd_rev;
+		int ret;
 
-	mst_port = aconnector->port;
-	mst_mgr = &aconnector->mst_port->mst_mgr;
+		ret = drm_dp_dpcd_readb(&port->aux, DP_DP13_DPCD_REV, &dpcd_rev);
 
-	if (!old_conn_state->crtc)
-		return 0;
+		if (ret == 1) {
+			port->dpcd_rev = dpcd_rev;
 
-	if (new_conn_state->crtc) {
-		new_crtc_state = drm_atomic_get_new_crtc_state(state, new_conn_state->crtc);
-		if (!new_crtc_state ||
-		    !drm_atomic_crtc_needs_modeset(new_crtc_state) ||
-		    new_crtc_state->enable)
-			return 0;
+			/* Could be DP1.2 DP Rx case*/
+			if (!dpcd_rev) {
+				ret = drm_dp_dpcd_readb(&port->aux, DP_DPCD_REV, &dpcd_rev);
+
+				if (ret == 1)
+					port->dpcd_rev = dpcd_rev;
+			}
+
+			if (!dpcd_rev)
+				DRM_DEBUG_KMS("Can't decide DPCD revision number!");
 		}
 
-	return drm_dp_atomic_release_vcpi_slots(state,
-						mst_mgr,
-						mst_port);
+		/*
+		 * Could be legacy sink, logical port etc on DP1.2.
+		 * Will get Nack under these cases when issue remote
+		 * DPCD read.
+		 */
+		if (ret != 1)
+			DRM_DEBUG_KMS("Can't access DPCD");
+	} else if (port->pdt == DP_PEER_DEVICE_NONE) {
+		port->dpcd_rev = 0;
+	}
+
+	/*
+	 * Release dc_sink for connector which unplug event is notified by CSN msg
+	 */
+	if (connection_status == connector_status_disconnected && aconnector->dc_sink) {
+		if (aconnector->dc_link->sink_count)
+			dc_link_remove_remote_sink(aconnector->dc_link, aconnector->dc_sink);
+
+		drm_dbg_dp(connector->dev,
+			   "DM_MST: remove remote sink 0x%p, %d remaining\n",
+			   aconnector->dc_link,
+			   aconnector->dc_link->sink_count);
+
+		dc_sink_release(aconnector->dc_sink);
+		aconnector->dc_sink = NULL;
+		amdgpu_dm_mst_reset_mst_connector_setting(aconnector);
+
+		amdgpu_dm_set_mst_status(&aconnector->mst_status,
+			MST_REMOTE_EDID | MST_ALLOCATE_NEW_PAYLOAD | MST_CLEAR_ALLOCATED_PAYLOAD,
+			false);
+	}
+
+	return connection_status;
+}
+
+static int dm_dp_mst_atomic_check(struct drm_connector *connector,
+				  struct drm_atomic_state *state)
+{
+	struct amdgpu_dm_connector *aconnector = to_amdgpu_dm_connector(connector);
+	struct drm_dp_mst_topology_mgr *mst_mgr = &aconnector->mst_root->mst_mgr;
+	struct drm_dp_mst_port *mst_port = aconnector->mst_output_port;
+
+	return drm_dp_atomic_release_time_slots(state, mst_mgr, mst_port);
 }
 
 static const struct drm_connector_helper_funcs dm_dp_mst_connector_helper_funcs = {
@@ -370,7 +599,6 @@ static const struct drm_connector_helper_funcs dm_dp_mst_connector_helper_funcs
 static void amdgpu_dm_encoder_destroy(struct drm_encoder *encoder)
 {
 	drm_encoder_cleanup(encoder);
-	kfree(encoder);
 }
 
 static const struct drm_encoder_funcs amdgpu_dm_encoder_funcs = {
@@ -416,15 +644,20 @@ dm_dp_add_mst_connector(struct drm_dp_mst_topology_mgr *mgr,
 	if (!aconnector)
 		return NULL;
 
+	DRM_DEBUG_DRIVER("%s: Create aconnector 0x%p for port 0x%p\n", __func__, aconnector, port);
+
 	connector = &aconnector->base;
-	aconnector->port = port;
-	aconnector->mst_port = master;
+	aconnector->mst_output_port = port;
+	aconnector->mst_root = master;
+	amdgpu_dm_set_mst_status(&aconnector->mst_status,
+			MST_PROBE, true);
 
-	if (drm_connector_init(
+	if (drm_connector_dynamic_init(
 		dev,
 		connector,
 		&dm_dp_mst_connector_funcs,
-		DRM_MODE_CONNECTOR_DisplayPort)) {
+		DRM_MODE_CONNECTOR_DisplayPort,
+		NULL)) {
 		kfree(aconnector);
 		return NULL;
 	}
@@ -458,9 +691,15 @@ dm_dp_add_mst_connector(struct drm_dp_mst_topology_mgr *mgr,
 		&connector->base,
 		dev->mode_config.tile_property,
 		0);
+	connector->colorspace_property = master->base.colorspace_property;
+	if (connector->colorspace_property)
+		drm_connector_attach_colorspace_property(connector);
 
 	drm_connector_set_path_property(connector, pathprop);
 
+	if (!retrieve_branch_specific_data(aconnector))
+		aconnector->branch_ieee_oui = 0;
+
 	/*
 	 * Initialize connector state before adding the connectror to drm and
 	 * framebuffer lists
@@ -472,8 +711,118 @@ dm_dp_add_mst_connector(struct drm_dp_mst_topology_mgr *mgr,
 	return connector;
 }
 
+void dm_handle_mst_sideband_msg_ready_event(
+	struct drm_dp_mst_topology_mgr *mgr,
+	enum mst_msg_ready_type msg_rdy_type)
+{
+	uint8_t esi[DP_PSR_ERROR_STATUS - DP_SINK_COUNT_ESI] = { 0 };
+	uint8_t dret;
+	bool new_irq_handled = false;
+	int dpcd_addr;
+	uint8_t dpcd_bytes_to_read;
+	const uint8_t max_process_count = 30;
+	uint8_t process_count = 0;
+	u8 retry;
+	struct amdgpu_dm_connector *aconnector =
+			container_of(mgr, struct amdgpu_dm_connector, mst_mgr);
+
+
+	const struct dc_link_status *link_status = dc_link_get_status(aconnector->dc_link);
+
+	if (link_status->dpcd_caps->dpcd_rev.raw < 0x12) {
+		dpcd_bytes_to_read = DP_LANE0_1_STATUS - DP_SINK_COUNT;
+		/* DPCD 0x200 - 0x201 for downstream IRQ */
+		dpcd_addr = DP_SINK_COUNT;
+	} else {
+		dpcd_bytes_to_read = DP_PSR_ERROR_STATUS - DP_SINK_COUNT_ESI;
+		/* DPCD 0x2002 - 0x2005 for downstream IRQ */
+		dpcd_addr = DP_SINK_COUNT_ESI;
+	}
+
+	mutex_lock(&aconnector->handle_mst_msg_ready);
+
+	while (process_count < max_process_count) {
+		u8 ack[DP_PSR_ERROR_STATUS - DP_SINK_COUNT_ESI] = {};
+
+		process_count++;
+
+		dret = drm_dp_dpcd_read(
+			&aconnector->dm_dp_aux.aux,
+			dpcd_addr,
+			esi,
+			dpcd_bytes_to_read);
+
+		if (dret != dpcd_bytes_to_read) {
+			DRM_DEBUG_KMS("DPCD read and acked number is not as expected!");
+			break;
+		}
+
+		DRM_DEBUG_DRIVER("ESI %02x %02x %02x\n", esi[0], esi[1], esi[2]);
+
+		switch (msg_rdy_type) {
+		case DOWN_REP_MSG_RDY_EVENT:
+			/* Only handle DOWN_REP_MSG_RDY case*/
+			esi[1] &= DP_DOWN_REP_MSG_RDY;
+			break;
+		case UP_REQ_MSG_RDY_EVENT:
+			/* Only handle UP_REQ_MSG_RDY case*/
+			esi[1] &= DP_UP_REQ_MSG_RDY;
+			break;
+		default:
+			/* Handle both cases*/
+			esi[1] &= (DP_DOWN_REP_MSG_RDY | DP_UP_REQ_MSG_RDY);
+			break;
+		}
+
+		if (!esi[1])
+			break;
+
+		/* handle MST irq */
+		if (aconnector->mst_mgr.mst_state)
+			drm_dp_mst_hpd_irq_handle_event(&aconnector->mst_mgr,
+						 esi,
+						 ack,
+						 &new_irq_handled);
+
+		if (new_irq_handled) {
+			/* ACK at DPCD to notify down stream */
+			for (retry = 0; retry < 3; retry++) {
+				ssize_t wret;
+
+				wret = drm_dp_dpcd_writeb(&aconnector->dm_dp_aux.aux,
+							  dpcd_addr + 1,
+							  ack[1]);
+				if (wret == 1)
+					break;
+			}
+
+			if (retry == 3) {
+				DRM_ERROR("Failed to ack MST event.\n");
+				break;
+			}
+
+			drm_dp_mst_hpd_irq_send_new_request(&aconnector->mst_mgr);
+
+			new_irq_handled = false;
+		} else {
+			break;
+		}
+	}
+
+	mutex_unlock(&aconnector->handle_mst_msg_ready);
+
+	if (process_count == max_process_count)
+		DRM_DEBUG_DRIVER("Loop exceeded max iterations\n");
+}
+
+static void dm_handle_mst_down_rep_msg_ready(struct drm_dp_mst_topology_mgr *mgr)
+{
+	dm_handle_mst_sideband_msg_ready_event(mgr, DOWN_REP_MSG_RDY_EVENT);
+}
+
 static const struct drm_dp_mst_topology_cbs dm_mst_cbs = {
 	.add_connector = dm_dp_add_mst_connector,
+	.poll_hpd_irq = dm_handle_mst_down_rep_msg_ready,
 };
 
 void amdgpu_dm_initialize_dp_connector(struct amdgpu_display_manager *dm,
@@ -492,35 +841,34 @@ void amdgpu_dm_initialize_dp_connector(struct amdgpu_display_manager *dm,
 	drm_dp_aux_init(&aconnector->dm_dp_aux.aux);
 	drm_dp_cec_register_connector(&aconnector->dm_dp_aux.aux,
 				      &aconnector->base);
+	drm_dp_dpcd_set_probe(&aconnector->dm_dp_aux.aux, false);
 
 	if (aconnector->base.connector_type == DRM_MODE_CONNECTOR_eDP)
 		return;
 
 	dc_link_dp_get_max_link_enc_cap(aconnector->dc_link, &max_link_enc_cap);
 	aconnector->mst_mgr.cbs = &dm_mst_cbs;
-	drm_dp_mst_topology_mgr_init(
-		&aconnector->mst_mgr,
-		adev_to_drm(dm->adev),
-		&aconnector->dm_dp_aux.aux,
-		16,
-		4,
-		max_link_enc_cap.lane_count,
-		drm_dp_bw_code_to_link_rate(max_link_enc_cap.link_rate),
-		aconnector->connector_id);
+	drm_dp_mst_topology_mgr_init(&aconnector->mst_mgr, adev_to_drm(dm->adev),
+				     &aconnector->dm_dp_aux.aux, 16, 4, aconnector->connector_id);
 
 	drm_connector_attach_dp_subconnector_property(&aconnector->base);
 }
 
-int dm_mst_get_pbn_divider(struct dc_link *link)
+uint32_t dm_mst_get_pbn_divider(struct dc_link *link)
 {
+	uint32_t pbn_div_x100;
+	uint64_t dividend, divisor;
+
 	if (!link)
 		return 0;
 
-	return dc_link_bandwidth_kbps(link,
-			dc_link_get_link_cap(link)) / (8 * 1000 * 54);
-}
+	dividend = (uint64_t)dc_link_bandwidth_kbps(link, dc_link_get_link_cap(link)) * 100;
+	divisor = 8 * 1000 * 54;
+
+	pbn_div_x100 = div64_u64(dividend, divisor);
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
+	return dfixed_const(pbn_div_x100) / 100;
+}
 
 struct dsc_mst_fairness_params {
 	struct dc_crtc_timing *timing;
@@ -535,13 +883,29 @@ struct dsc_mst_fairness_params {
 	struct amdgpu_dm_connector *aconnector;
 };
 
-static int kbps_to_peak_pbn(int kbps)
+#if defined(CONFIG_DRM_AMD_DC_FP)
+static uint64_t kbps_to_pbn(int kbps, bool is_peak_pbn)
+{
+	uint64_t effective_kbps = (uint64_t)kbps;
+
+	if (is_peak_pbn) {	// add 0.6% (1006/1000) overhead into effective kbps
+		effective_kbps *= 1006;
+		effective_kbps = div_u64(effective_kbps, 1000);
+	}
+
+	return (uint64_t) DIV64_U64_ROUND_UP(effective_kbps * 64, (54 * 8 * 1000));
+}
+
+static uint32_t pbn_to_kbps(unsigned int pbn, bool with_margin)
 {
-	u64 peak_kbps = kbps;
+	uint64_t pbn_effective = (uint64_t)pbn;
+
+	if (with_margin)	// deduct 0.6% (994/1000) overhead from effective pbn
+		pbn_effective *= (1000000 / PEAK_FACTOR_X1000);
+	else
+		pbn_effective *= 1000;
 
-	peak_kbps *= 1006;
-	peak_kbps = div_u64(peak_kbps, 1000);
-	return (int) DIV64_U64_ROUND_UP(peak_kbps * 64, (54 * 8 * 1000));
+	return DIV_U64_ROUND_UP(pbn_effective * 8 * 54, 64);
 }
 
 static void set_dsc_configs_from_fairness_vars(struct dsc_mst_fairness_params *params,
@@ -549,17 +913,24 @@ static void set_dsc_configs_from_fairness_vars(struct dsc_mst_fairness_params *p
 		int count,
 		int k)
 {
+	struct drm_connector *drm_connector;
 	int i;
+	struct dc_dsc_config_options dsc_options = {0};
 
 	for (i = 0; i < count; i++) {
+		drm_connector = &params[i].aconnector->base;
+
+		dc_dsc_get_default_config_option(params[i].sink->ctx->dc, &dsc_options);
+		dsc_options.max_target_bpp_limit_override_x16 = drm_connector->display_info.max_dsc_bpp * 16;
+
 		memset(&params[i].timing->dsc_cfg, 0, sizeof(params[i].timing->dsc_cfg));
 		if (vars[i + k].dsc_enabled && dc_dsc_compute_config(
 					params[i].sink->ctx->dc->res_pool->dscs[0],
 					&params[i].sink->dsc_caps.dsc_dec_caps,
-					params[i].sink->ctx->dc->debug.dsc_min_slice_height_override,
-					params[i].sink->edid_caps.panel_patch.max_dsc_target_bpp_limit,
+					&dsc_options,
 					0,
 					params[i].timing,
+					dc_link_get_highest_encoding_format(params[i].aconnector->dc_link),
 					&params[i].timing->dsc_cfg)) {
 			params[i].timing->flags.DSC = 1;
 
@@ -576,6 +947,21 @@ static void set_dsc_configs_from_fairness_vars(struct dsc_mst_fairness_params *p
 		} else {
 			params[i].timing->flags.DSC = 0;
 		}
+		params[i].timing->dsc_cfg.mst_pbn = vars[i + k].pbn;
+	}
+
+	for (i = 0; i < count; i++) {
+		if (params[i].sink) {
+			if (params[i].sink->sink_signal != SIGNAL_TYPE_VIRTUAL &&
+				params[i].sink->sink_signal != SIGNAL_TYPE_NONE)
+				DRM_DEBUG_DRIVER("MST_DSC %s i=%d dispname=%s\n", __func__, i,
+					params[i].sink->edid_caps.display_name);
+		}
+
+		DRM_DEBUG_DRIVER("MST_DSC dsc=%d bits_per_pixel=%d pbn=%d\n",
+			params[i].timing->flags.DSC,
+			params[i].timing->dsc_cfg.bits_per_pixel,
+			vars[i + k].pbn);
 	}
 }
 
@@ -584,23 +970,31 @@ static int bpp_x16_from_pbn(struct dsc_mst_fairness_params param, int pbn)
 	struct dc_dsc_config dsc_config;
 	u64 kbps;
 
-	kbps = div_u64((u64)pbn * 994 * 8 * 54, 64);
+	struct drm_connector *drm_connector = &param.aconnector->base;
+	struct dc_dsc_config_options dsc_options = {0};
+
+	dc_dsc_get_default_config_option(param.sink->ctx->dc, &dsc_options);
+	dsc_options.max_target_bpp_limit_override_x16 = drm_connector->display_info.max_dsc_bpp * 16;
+
+	kbps = pbn_to_kbps(pbn, false);
 	dc_dsc_compute_config(
 			param.sink->ctx->dc->res_pool->dscs[0],
 			&param.sink->dsc_caps.dsc_dec_caps,
-			param.sink->ctx->dc->debug.dsc_min_slice_height_override,
-			param.sink->edid_caps.panel_patch.max_dsc_target_bpp_limit,
-			(int) kbps, param.timing, &dsc_config);
+			&dsc_options,
+			(int) kbps, param.timing,
+			dc_link_get_highest_encoding_format(param.aconnector->dc_link),
+			&dsc_config);
 
 	return dsc_config.bits_per_pixel;
 }
 
-static void increase_dsc_bpp(struct drm_atomic_state *state,
-			     struct dc_link *dc_link,
-			     struct dsc_mst_fairness_params *params,
-			     struct dsc_mst_fairness_vars *vars,
-			     int count,
-			     int k)
+static int increase_dsc_bpp(struct drm_atomic_state *state,
+			    struct drm_dp_mst_topology_state *mst_state,
+			    struct dc_link *dc_link,
+			    struct dsc_mst_fairness_params *params,
+			    struct dsc_mst_fairness_vars *vars,
+			    int count,
+			    int k)
 {
 	int i;
 	bool bpp_increased[MAX_PIPES];
@@ -608,16 +1002,14 @@ static void increase_dsc_bpp(struct drm_atomic_state *state,
 	int min_initial_slack;
 	int next_index;
 	int remaining_to_increase = 0;
-	int pbn_per_timeslot;
 	int link_timeslots_used;
 	int fair_pbn_alloc;
-
-	pbn_per_timeslot = dm_mst_get_pbn_divider(dc_link);
+	int ret = 0;
 
 	for (i = 0; i < count; i++) {
 		if (vars[i + k].dsc_enabled) {
 			initial_slack[i] =
-			kbps_to_peak_pbn(params[i].bw_range.max_kbps) - vars[i + k].pbn;
+			kbps_to_pbn(params[i].bw_range.max_kbps, false) - vars[i + k].pbn;
 			bpp_increased[i] = false;
 			remaining_to_increase += 1;
 		} else {
@@ -644,61 +1036,67 @@ static void increase_dsc_bpp(struct drm_atomic_state *state,
 		link_timeslots_used = 0;
 
 		for (i = 0; i < count; i++)
-			link_timeslots_used += DIV_ROUND_UP(vars[i + k].pbn, pbn_per_timeslot);
+			link_timeslots_used += DIV_ROUND_UP(vars[i + k].pbn, dfixed_trunc(mst_state->pbn_div));
 
-		fair_pbn_alloc = (63 - link_timeslots_used) / remaining_to_increase * pbn_per_timeslot;
+		fair_pbn_alloc =
+			(63 - link_timeslots_used) / remaining_to_increase * dfixed_trunc(mst_state->pbn_div);
 
 		if (initial_slack[next_index] > fair_pbn_alloc) {
 			vars[next_index].pbn += fair_pbn_alloc;
-			if (drm_dp_atomic_find_vcpi_slots(state,
-							  params[next_index].port->mgr,
-							  params[next_index].port,
-							  vars[next_index].pbn,
-							  pbn_per_timeslot) < 0)
-				return;
-			if (!drm_dp_mst_atomic_check(state)) {
+			ret = drm_dp_atomic_find_time_slots(state,
+							    params[next_index].port->mgr,
+							    params[next_index].port,
+							    vars[next_index].pbn);
+			if (ret < 0)
+				return ret;
+
+			ret = drm_dp_mst_atomic_check(state);
+			if (ret == 0) {
 				vars[next_index].bpp_x16 = bpp_x16_from_pbn(params[next_index], vars[next_index].pbn);
 			} else {
 				vars[next_index].pbn -= fair_pbn_alloc;
-				if (drm_dp_atomic_find_vcpi_slots(state,
-								  params[next_index].port->mgr,
-								  params[next_index].port,
-								  vars[next_index].pbn,
-								  pbn_per_timeslot) < 0)
-					return;
+				ret = drm_dp_atomic_find_time_slots(state,
+								    params[next_index].port->mgr,
+								    params[next_index].port,
+								    vars[next_index].pbn);
+				if (ret < 0)
+					return ret;
 			}
 		} else {
 			vars[next_index].pbn += initial_slack[next_index];
-			if (drm_dp_atomic_find_vcpi_slots(state,
-							  params[next_index].port->mgr,
-							  params[next_index].port,
-							  vars[next_index].pbn,
-							  pbn_per_timeslot) < 0)
-				return;
-			if (!drm_dp_mst_atomic_check(state)) {
+			ret = drm_dp_atomic_find_time_slots(state,
+							    params[next_index].port->mgr,
+							    params[next_index].port,
+							    vars[next_index].pbn);
+			if (ret < 0)
+				return ret;
+
+			ret = drm_dp_mst_atomic_check(state);
+			if (ret == 0) {
 				vars[next_index].bpp_x16 = params[next_index].bw_range.max_target_bpp_x16;
 			} else {
 				vars[next_index].pbn -= initial_slack[next_index];
-				if (drm_dp_atomic_find_vcpi_slots(state,
-								  params[next_index].port->mgr,
-								  params[next_index].port,
-								  vars[next_index].pbn,
-								  pbn_per_timeslot) < 0)
-					return;
+				ret = drm_dp_atomic_find_time_slots(state,
+								    params[next_index].port->mgr,
+								    params[next_index].port,
+								    vars[next_index].pbn);
+				if (ret < 0)
+					return ret;
 			}
 		}
 
 		bpp_increased[next_index] = true;
 		remaining_to_increase--;
 	}
+	return 0;
 }
 
-static void try_disable_dsc(struct drm_atomic_state *state,
-			    struct dc_link *dc_link,
-			    struct dsc_mst_fairness_params *params,
-			    struct dsc_mst_fairness_vars *vars,
-			    int count,
-			    int k)
+static int try_disable_dsc(struct drm_atomic_state *state,
+			   struct dc_link *dc_link,
+			   struct dsc_mst_fairness_params *params,
+			   struct dsc_mst_fairness_vars *vars,
+			   int count,
+			   int k)
 {
 	int i;
 	bool tried[MAX_PIPES];
@@ -706,6 +1104,8 @@ static void try_disable_dsc(struct drm_atomic_state *state,
 	int max_kbps_increase;
 	int next_index;
 	int remaining_to_try = 0;
+	int ret;
+	int var_pbn;
 
 	for (i = 0; i < count; i++) {
 		if (vars[i + k].dsc_enabled
@@ -735,48 +1135,77 @@ static void try_disable_dsc(struct drm_atomic_state *state,
 		if (next_index == -1)
 			break;
 
-		vars[next_index].pbn = kbps_to_peak_pbn(params[next_index].bw_range.stream_kbps);
-		if (drm_dp_atomic_find_vcpi_slots(state,
-						  params[next_index].port->mgr,
-						  params[next_index].port,
-						  vars[next_index].pbn,
-						  dm_mst_get_pbn_divider(dc_link)) < 0)
-			return;
+		DRM_DEBUG_DRIVER("MST_DSC index #%d, try no compression\n", next_index);
+		var_pbn = vars[next_index].pbn;
+		vars[next_index].pbn = kbps_to_pbn(params[next_index].bw_range.stream_kbps, true);
+		ret = drm_dp_atomic_find_time_slots(state,
+						    params[next_index].port->mgr,
+						    params[next_index].port,
+						    vars[next_index].pbn);
+		if (ret < 0) {
+			DRM_DEBUG_DRIVER("%s:%d MST_DSC index #%d, failed to set pbn to the state, %d\n",
+						__func__, __LINE__, next_index, ret);
+			vars[next_index].pbn = var_pbn;
+			return ret;
+		}
 
-		if (!drm_dp_mst_atomic_check(state)) {
+		ret = drm_dp_mst_atomic_check(state);
+		if (ret == 0) {
+			DRM_DEBUG_DRIVER("MST_DSC index #%d, greedily disable dsc\n", next_index);
 			vars[next_index].dsc_enabled = false;
 			vars[next_index].bpp_x16 = 0;
 		} else {
-			vars[next_index].pbn = kbps_to_peak_pbn(params[next_index].bw_range.max_kbps);
-			if (drm_dp_atomic_find_vcpi_slots(state,
-							  params[next_index].port->mgr,
-							  params[next_index].port,
-							  vars[next_index].pbn,
-							  dm_mst_get_pbn_divider(dc_link)) < 0)
-				return;
+			DRM_DEBUG_DRIVER("MST_DSC index #%d, restore optimized pbn value\n", next_index);
+			vars[next_index].pbn = var_pbn;
+			ret = drm_dp_atomic_find_time_slots(state,
+							    params[next_index].port->mgr,
+							    params[next_index].port,
+							    vars[next_index].pbn);
+			if (ret < 0) {
+				DRM_DEBUG_DRIVER("%s:%d MST_DSC index #%d, failed to set pbn to the state, %d\n",
+							__func__, __LINE__, next_index, ret);
+				return ret;
+			}
 		}
 
 		tried[next_index] = true;
 		remaining_to_try--;
 	}
+	return 0;
+}
+
+static void log_dsc_params(int count, struct dsc_mst_fairness_vars *vars, int k)
+{
+	int i;
+
+	for (i = 0; i < count; i++)
+		DRM_DEBUG_DRIVER("MST_DSC DSC params: stream #%d --- dsc_enabled = %d, bpp_x16 = %d, pbn = %d\n",
+				 i, vars[i + k].dsc_enabled, vars[i + k].bpp_x16, vars[i + k].pbn);
 }
 
-static bool compute_mst_dsc_configs_for_link(struct drm_atomic_state *state,
-					     struct dc_state *dc_state,
-					     struct dc_link *dc_link,
-					     struct dsc_mst_fairness_vars *vars,
-					     int *link_vars_start_index)
+static int compute_mst_dsc_configs_for_link(struct drm_atomic_state *state,
+					    struct dc_state *dc_state,
+					    struct dc_link *dc_link,
+					    struct dsc_mst_fairness_vars *vars,
+					    struct drm_dp_mst_topology_mgr *mgr,
+					    int *link_vars_start_index)
 {
-	int i, k;
 	struct dc_stream_state *stream;
 	struct dsc_mst_fairness_params params[MAX_PIPES];
 	struct amdgpu_dm_connector *aconnector;
+	struct drm_dp_mst_topology_state *mst_state = drm_atomic_get_mst_topology_state(state, mgr);
 	int count = 0;
+	int i, k, ret;
 	bool debugfs_overwrite = false;
+	struct drm_connector_state *new_conn_state;
 
 	memset(params, 0, sizeof(params));
 
+	if (IS_ERR(mst_state))
+		return PTR_ERR(mst_state);
+
 	/* Set up params */
+	DRM_DEBUG_DRIVER("%s: MST_DSC Try to set up params from %d streams\n", __func__, dc_state->stream_count);
 	for (i = 0; i < dc_state->stream_count; i++) {
 		struct dc_dsc_policy dsc_policy = {0};
 
@@ -789,15 +1218,23 @@ static bool compute_mst_dsc_configs_for_link(struct drm_atomic_state *state,
 		if (!aconnector)
 			continue;
 
-		if (!aconnector->port)
+		if (!aconnector->mst_output_port)
+			continue;
+
+		new_conn_state = drm_atomic_get_new_connector_state(state, &aconnector->base);
+
+		if (!new_conn_state) {
+			DRM_DEBUG_DRIVER("%s:%d MST_DSC Skip the stream 0x%p with invalid new_conn_state\n",
+					__func__, __LINE__, stream);
 			continue;
+		}
 
 		stream->timing.flags.DSC = 0;
 
 		params[count].timing = &stream->timing;
 		params[count].sink = stream->sink;
 		params[count].aconnector = aconnector;
-		params[count].port = aconnector->port;
+		params[count].port = aconnector->mst_output_port;
 		params[count].clock_force_enable = aconnector->dsc_settings.dsc_force_enable;
 		if (params[count].clock_force_enable == DSC_CLK_FORCE_ENABLE)
 			debugfs_overwrite = true;
@@ -805,22 +1242,30 @@ static bool compute_mst_dsc_configs_for_link(struct drm_atomic_state *state,
 		params[count].num_slices_v = aconnector->dsc_settings.dsc_num_slices_v;
 		params[count].bpp_overwrite = aconnector->dsc_settings.dsc_bits_per_pixel;
 		params[count].compression_possible = stream->sink->dsc_caps.dsc_dec_caps.is_dsc_supported;
-		dc_dsc_get_policy_for_timing(params[count].timing, 0, &dsc_policy);
+		dc_dsc_get_policy_for_timing(params[count].timing, 0, &dsc_policy, dc_link_get_highest_encoding_format(stream->link));
 		if (!dc_dsc_compute_bandwidth_range(
 				stream->sink->ctx->dc->res_pool->dscs[0],
 				stream->sink->ctx->dc->debug.dsc_min_slice_height_override,
 				dsc_policy.min_target_bpp * 16,
 				dsc_policy.max_target_bpp * 16,
 				&stream->sink->dsc_caps.dsc_dec_caps,
-				&stream->timing, &params[count].bw_range))
-			params[count].bw_range.stream_kbps = dc_bandwidth_in_kbps_from_timing(&stream->timing);
-
+				&stream->timing,
+				dc_link_get_highest_encoding_format(dc_link),
+				&params[count].bw_range))
+			params[count].bw_range.stream_kbps = dc_bandwidth_in_kbps_from_timing(&stream->timing,
+					dc_link_get_highest_encoding_format(dc_link));
+
+		DRM_DEBUG_DRIVER("MST_DSC #%d stream 0x%p - max_kbps = %u, min_kbps = %u, uncompressed_kbps = %u\n",
+			count, stream, params[count].bw_range.max_kbps, params[count].bw_range.min_kbps,
+			params[count].bw_range.stream_kbps);
 		count++;
 	}
 
+	DRM_DEBUG_DRIVER("%s: MST_DSC Params set up for %d streams\n", __func__, count);
+
 	if (count == 0) {
 		ASSERT(0);
-		return true;
+		return 0;
 	}
 
 	/* k is start index of vars for current phy link used by mst hub */
@@ -829,58 +1274,76 @@ static bool compute_mst_dsc_configs_for_link(struct drm_atomic_state *state,
 	*link_vars_start_index += count;
 
 	/* Try no compression */
+	DRM_DEBUG_DRIVER("MST_DSC Try no compression\n");
 	for (i = 0; i < count; i++) {
 		vars[i + k].aconnector = params[i].aconnector;
-		vars[i + k].pbn = kbps_to_peak_pbn(params[i].bw_range.stream_kbps);
+		vars[i + k].pbn = kbps_to_pbn(params[i].bw_range.stream_kbps, false);
 		vars[i + k].dsc_enabled = false;
 		vars[i + k].bpp_x16 = 0;
-		if (drm_dp_atomic_find_vcpi_slots(state,
-						 params[i].port->mgr,
-						 params[i].port,
-						 vars[i + k].pbn,
-						 dm_mst_get_pbn_divider(dc_link)) < 0)
-			return false;
+		ret = drm_dp_atomic_find_time_slots(state, params[i].port->mgr, params[i].port,
+						    vars[i + k].pbn);
+		if (ret < 0)
+			return ret;
 	}
-	if (!drm_dp_mst_atomic_check(state) && !debugfs_overwrite) {
+	ret = drm_dp_mst_atomic_check(state);
+	if (ret == 0 && !debugfs_overwrite) {
 		set_dsc_configs_from_fairness_vars(params, vars, count, k);
-		return true;
+		return 0;
+	} else if (ret != -ENOSPC) {
+		return ret;
 	}
 
+	log_dsc_params(count, vars, k);
+
 	/* Try max compression */
+	DRM_DEBUG_DRIVER("MST_DSC Try max compression\n");
 	for (i = 0; i < count; i++) {
 		if (params[i].compression_possible && params[i].clock_force_enable != DSC_CLK_FORCE_DISABLE) {
-			vars[i + k].pbn = kbps_to_peak_pbn(params[i].bw_range.min_kbps);
+			vars[i + k].pbn = kbps_to_pbn(params[i].bw_range.min_kbps, false);
 			vars[i + k].dsc_enabled = true;
 			vars[i + k].bpp_x16 = params[i].bw_range.min_target_bpp_x16;
-			if (drm_dp_atomic_find_vcpi_slots(state,
-							  params[i].port->mgr,
-							  params[i].port,
-							  vars[i + k].pbn,
-							  dm_mst_get_pbn_divider(dc_link)) < 0)
-				return false;
+			ret = drm_dp_atomic_find_time_slots(state, params[i].port->mgr,
+							    params[i].port, vars[i + k].pbn);
+			if (ret < 0)
+				return ret;
 		} else {
-			vars[i + k].pbn = kbps_to_peak_pbn(params[i].bw_range.stream_kbps);
+			vars[i + k].pbn = kbps_to_pbn(params[i].bw_range.stream_kbps, false);
 			vars[i + k].dsc_enabled = false;
 			vars[i + k].bpp_x16 = 0;
-			if (drm_dp_atomic_find_vcpi_slots(state,
-							  params[i].port->mgr,
-							  params[i].port,
-							  vars[i + k].pbn,
-							  dm_mst_get_pbn_divider(dc_link)) < 0)
-				return false;
+			ret = drm_dp_atomic_find_time_slots(state, params[i].port->mgr,
+							    params[i].port, vars[i + k].pbn);
+			if (ret < 0)
+				return ret;
 		}
 	}
-	if (drm_dp_mst_atomic_check(state))
-		return false;
+	ret = drm_dp_mst_atomic_check(state);
+	if (ret != 0)
+		return ret;
+
+	log_dsc_params(count, vars, k);
 
 	/* Optimize degree of compression */
-	increase_dsc_bpp(state, dc_link, params, vars, count, k);
+	DRM_DEBUG_DRIVER("MST_DSC Try optimize compression\n");
+	ret = increase_dsc_bpp(state, mst_state, dc_link, params, vars, count, k);
+	if (ret < 0) {
+		DRM_DEBUG_DRIVER("MST_DSC Failed to optimize compression\n");
+		return ret;
+	}
 
-	try_disable_dsc(state, dc_link, params, vars, count, k);
+	log_dsc_params(count, vars, k);
+
+	DRM_DEBUG_DRIVER("MST_DSC Try disable compression\n");
+	ret = try_disable_dsc(state, dc_link, params, vars, count, k);
+	if (ret < 0) {
+		DRM_DEBUG_DRIVER("MST_DSC Failed to disable compression\n");
+		return ret;
+	}
+
+	log_dsc_params(count, vars, k);
 
 	set_dsc_configs_from_fairness_vars(params, vars, count, k);
 
-	return true;
+	return 0;
 }
 
 static bool is_dsc_need_re_compute(
@@ -888,25 +1351,40 @@ static bool is_dsc_need_re_compute(
 	struct dc_state *dc_state,
 	struct dc_link *dc_link)
 {
-	int i;
+	int i, j;
 	bool is_dsc_need_re_compute = false;
+	struct amdgpu_dm_connector *stream_on_link[MAX_PIPES];
+	int new_stream_on_link_num = 0;
+	struct amdgpu_dm_connector *aconnector;
+	struct dc_stream_state *stream;
+	const struct dc *dc = dc_link->dc;
 
-	/* only check phy used by mst branch */
+	/* only check phy used by dsc mst branch */
 	if (dc_link->type != dc_connection_mst_branch)
-		return false;
+		goto out;
+
+	/* add a check for older MST DSC with no virtual DPCDs */
+	if (needs_dsc_aux_workaround(dc_link)  &&
+		(!(dc_link->dpcd_caps.dsc_caps.dsc_basic_caps.fields.dsc_support.DSC_SUPPORT ||
+		dc_link->dpcd_caps.dsc_caps.dsc_basic_caps.fields.dsc_support.DSC_PASSTHROUGH_SUPPORT)))
+		goto out;
+
+	for (i = 0; i < MAX_PIPES; i++)
+		stream_on_link[i] = NULL;
+
+	DRM_DEBUG_DRIVER("%s: MST_DSC check on %d streams in new dc_state\n", __func__, dc_state->stream_count);
 
 	/* check if there is mode change in new request */
 	for (i = 0; i < dc_state->stream_count; i++) {
-		struct amdgpu_dm_connector *aconnector;
-		struct dc_stream_state *stream;
 		struct drm_crtc_state *new_crtc_state;
 		struct drm_connector_state *new_conn_state;
 
 		stream = dc_state->streams[i];
-
 		if (!stream)
 			continue;
 
+		DRM_DEBUG_DRIVER("%s:%d MST_DSC checking #%d stream 0x%p\n", __func__, __LINE__, i, stream);
+
 		/* check if stream using the same link for mst */
 		if (stream->link != dc_link)
 			continue;
@@ -915,10 +1393,15 @@ static bool is_dsc_need_re_compute(
 		if (!aconnector)
 			continue;
 
-		new_conn_state = drm_atomic_get_new_connector_state(state, &aconnector->base);
+		stream_on_link[new_stream_on_link_num] = aconnector;
+		new_stream_on_link_num++;
 
-		if (!new_conn_state)
+		new_conn_state = drm_atomic_get_new_connector_state(state, &aconnector->base);
+		if (!new_conn_state) {
+			DRM_DEBUG_DRIVER("%s:%d MST_DSC no new_conn_state for stream 0x%p, aconnector 0x%p\n",
+					 __func__, __LINE__, stream, aconnector);
 			continue;
+		}
 
 		if (IS_ERR(new_conn_state))
 			continue;
@@ -927,45 +1410,103 @@ static bool is_dsc_need_re_compute(
 			continue;
 
 		new_crtc_state = drm_atomic_get_new_crtc_state(state, new_conn_state->crtc);
-
-		if (!new_crtc_state)
+		if (!new_crtc_state) {
+			DRM_DEBUG_DRIVER("%s:%d MST_DSC no new_crtc_state for crtc of stream 0x%p, aconnector 0x%p\n",
+						__func__, __LINE__, stream, aconnector);
 			continue;
+		}
 
 		if (IS_ERR(new_crtc_state))
 			continue;
 
 		if (new_crtc_state->enable && new_crtc_state->active) {
 			if (new_crtc_state->mode_changed || new_crtc_state->active_changed ||
-				new_crtc_state->connectors_changed)
+					new_crtc_state->connectors_changed) {
+				DRM_DEBUG_DRIVER("%s:%d MST_DSC dsc recompute required."
+						 "stream 0x%p in new dc_state\n",
+						 __func__, __LINE__, stream);
 				is_dsc_need_re_compute = true;
+				goto out;
+			}
+		}
+	}
+
+	if (new_stream_on_link_num == 0) {
+		DRM_DEBUG_DRIVER("%s:%d MST_DSC no mode change request for streams in new dc_state\n",
+				 __func__, __LINE__);
+		is_dsc_need_re_compute = false;
+		goto out;
+	}
+
+	DRM_DEBUG_DRIVER("%s: MST_DSC check on %d streams in current dc_state\n",
+			 __func__, dc->current_state->stream_count);
+
+	/* check current_state if there stream on link but it is not in
+	 * new request state
+	 */
+	for (i = 0; i < dc->current_state->stream_count; i++) {
+		stream = dc->current_state->streams[i];
+		/* only check stream on the mst hub */
+		if (stream->link != dc_link)
+			continue;
+
+		aconnector = (struct amdgpu_dm_connector *)stream->dm_stream_context;
+		if (!aconnector)
+			continue;
+
+		for (j = 0; j < new_stream_on_link_num; j++) {
+			if (stream_on_link[j]) {
+				if (aconnector == stream_on_link[j])
+					break;
+			}
+		}
+
+		if (j == new_stream_on_link_num) {
+			/* not in new state */
+			DRM_DEBUG_DRIVER("%s:%d MST_DSC dsc recompute required."
+					 "stream 0x%p in current dc_state but not in new dc_state\n",
+						__func__, __LINE__, stream);
+			is_dsc_need_re_compute = true;
+			break;
 		}
 	}
 
+out:
+	DRM_DEBUG_DRIVER("%s: MST_DSC dsc recompute %s\n",
+			 __func__, is_dsc_need_re_compute ? "required" : "not required");
+
 	return is_dsc_need_re_compute;
 }
 
-bool compute_mst_dsc_configs_for_state(struct drm_atomic_state *state,
-				       struct dc_state *dc_state,
-				       struct dsc_mst_fairness_vars *vars)
+int compute_mst_dsc_configs_for_state(struct drm_atomic_state *state,
+				      struct dc_state *dc_state,
+				      struct dsc_mst_fairness_vars *vars)
 {
 	int i, j;
 	struct dc_stream_state *stream;
 	bool computed_streams[MAX_PIPES];
 	struct amdgpu_dm_connector *aconnector;
+	struct drm_dp_mst_topology_mgr *mst_mgr;
+	struct resource_pool *res_pool;
 	int link_vars_start_index = 0;
+	int ret = 0;
 
 	for (i = 0; i < dc_state->stream_count; i++)
 		computed_streams[i] = false;
 
 	for (i = 0; i < dc_state->stream_count; i++) {
 		stream = dc_state->streams[i];
+		res_pool = stream->ctx->dc->res_pool;
 
 		if (stream->signal != SIGNAL_TYPE_DISPLAY_PORT_MST)
 			continue;
 
 		aconnector = (struct amdgpu_dm_connector *)stream->dm_stream_context;
 
-		if (!aconnector || !aconnector->dc_sink)
+		DRM_DEBUG_DRIVER("%s: MST_DSC compute mst dsc configs for stream 0x%p, aconnector 0x%p\n",
+				__func__, stream, aconnector);
+
+		if (!aconnector || !aconnector->dc_sink || !aconnector->mst_output_port)
 			continue;
 
 		if (!aconnector->dc_sink->dsc_caps.dsc_dec_caps.is_dsc_supported)
@@ -974,19 +1515,18 @@ bool compute_mst_dsc_configs_for_state(struct drm_atomic_state *state,
 		if (computed_streams[i])
 			continue;
 
-		if (dcn20_remove_stream_from_ctx(stream->ctx->dc, dc_state, stream) != DC_OK)
-			return false;
+		if (res_pool->funcs->remove_stream_from_ctx &&
+		    res_pool->funcs->remove_stream_from_ctx(stream->ctx->dc, dc_state, stream) != DC_OK)
+			return -EINVAL;
 
 		if (!is_dsc_need_re_compute(state, dc_state, stream->link))
 			continue;
 
-		mutex_lock(&aconnector->mst_mgr.lock);
-		if (!compute_mst_dsc_configs_for_link(state, dc_state, stream->link,
-			vars, &link_vars_start_index)) {
-			mutex_unlock(&aconnector->mst_mgr.lock);
-			return false;
-		}
-		mutex_unlock(&aconnector->mst_mgr.lock);
+		mst_mgr = aconnector->mst_output_port->mgr;
+		ret = compute_mst_dsc_configs_for_link(state, dc_state, stream->link, vars, mst_mgr,
+						       &link_vars_start_index);
+		if (ret != 0)
+			return ret;
 
 		for (j = 0; j < dc_state->stream_count; j++) {
 			if (dc_state->streams[j]->link == stream->link)
@@ -998,11 +1538,437 @@ bool compute_mst_dsc_configs_for_state(struct drm_atomic_state *state,
 		stream = dc_state->streams[i];
 
 		if (stream->timing.flags.DSC == 1)
-			if (dc_stream_add_dsc_to_resource(stream->ctx->dc, dc_state, stream) != DC_OK)
-				return false;
+			if (dc_stream_add_dsc_to_resource(stream->ctx->dc, dc_state, stream) != DC_OK) {
+				DRM_DEBUG_DRIVER("%s:%d MST_DSC Failed to request dsc hw resource for stream 0x%p\n",
+							__func__, __LINE__, stream);
+				return -EINVAL;
+			}
 	}
 
+	return ret;
+}
+
+static int pre_compute_mst_dsc_configs_for_state(struct drm_atomic_state *state,
+						 struct dc_state *dc_state,
+						 struct dsc_mst_fairness_vars *vars)
+{
+	int i, j;
+	struct dc_stream_state *stream;
+	bool computed_streams[MAX_PIPES];
+	struct amdgpu_dm_connector *aconnector;
+	struct drm_dp_mst_topology_mgr *mst_mgr;
+	int link_vars_start_index = 0;
+	int ret = 0;
+
+	for (i = 0; i < dc_state->stream_count; i++)
+		computed_streams[i] = false;
+
+	for (i = 0; i < dc_state->stream_count; i++) {
+		stream = dc_state->streams[i];
+
+		if (stream->signal != SIGNAL_TYPE_DISPLAY_PORT_MST)
+			continue;
+
+		aconnector = (struct amdgpu_dm_connector *)stream->dm_stream_context;
+
+		DRM_DEBUG_DRIVER("MST_DSC pre compute mst dsc configs for #%d stream 0x%p, aconnector 0x%p\n",
+					i, stream, aconnector);
+
+		if (!aconnector || !aconnector->dc_sink || !aconnector->mst_output_port)
+			continue;
+
+		if (!aconnector->dc_sink->dsc_caps.dsc_dec_caps.is_dsc_supported)
+			continue;
+
+		if (computed_streams[i])
+			continue;
+
+		if (!is_dsc_need_re_compute(state, dc_state, stream->link))
+			continue;
+
+		mst_mgr = aconnector->mst_output_port->mgr;
+		ret = compute_mst_dsc_configs_for_link(state, dc_state, stream->link, vars, mst_mgr,
+						       &link_vars_start_index);
+		if (ret != 0)
+			return ret;
+
+		for (j = 0; j < dc_state->stream_count; j++) {
+			if (dc_state->streams[j]->link == stream->link)
+				computed_streams[j] = true;
+		}
+	}
+
+	return ret;
+}
+
+static int find_crtc_index_in_state_by_stream(struct drm_atomic_state *state,
+					      struct dc_stream_state *stream)
+{
+	int i;
+	struct drm_crtc *crtc;
+	struct drm_crtc_state *new_state, *old_state;
+
+	for_each_oldnew_crtc_in_state(state, crtc, old_state, new_state, i) {
+		struct dm_crtc_state *dm_state = to_dm_crtc_state(new_state);
+
+		if (dm_state->stream == stream)
+			return i;
+	}
+	return -1;
+}
+
+static bool is_link_to_dschub(struct dc_link *dc_link)
+{
+	union dpcd_dsc_basic_capabilities *dsc_caps =
+			&dc_link->dpcd_caps.dsc_caps.dsc_basic_caps;
+
+	/* only check phy used by dsc mst branch */
+	if (dc_link->type != dc_connection_mst_branch)
+		return false;
+
+	if (!(dsc_caps->fields.dsc_support.DSC_SUPPORT ||
+	      dsc_caps->fields.dsc_support.DSC_PASSTHROUGH_SUPPORT))
+		return false;
+	return true;
+}
+
+static bool is_dsc_precompute_needed(struct drm_atomic_state *state)
+{
+	int i;
+	struct drm_crtc *crtc;
+	struct drm_crtc_state *old_crtc_state, *new_crtc_state;
+	bool ret = false;
+
+	for_each_oldnew_crtc_in_state(state, crtc, old_crtc_state, new_crtc_state, i) {
+		struct dm_crtc_state *dm_crtc_state = to_dm_crtc_state(new_crtc_state);
+
+		if (!amdgpu_dm_find_first_crtc_matching_connector(state, crtc)) {
+			ret =  false;
+			break;
+		}
+		if (dm_crtc_state->stream && dm_crtc_state->stream->link)
+			if (is_link_to_dschub(dm_crtc_state->stream->link))
+				ret = true;
+	}
+	return ret;
+}
+
+int pre_validate_dsc(struct drm_atomic_state *state,
+		     struct dm_atomic_state **dm_state_ptr,
+		     struct dsc_mst_fairness_vars *vars)
+{
+	int i;
+	struct dm_atomic_state *dm_state;
+	struct dc_state *local_dc_state = NULL;
+	int ret = 0;
+
+	if (!is_dsc_precompute_needed(state)) {
+		DRM_INFO_ONCE("%s:%d MST_DSC dsc precompute is not needed\n", __func__, __LINE__);
+		return 0;
+	}
+	ret = dm_atomic_get_state(state, dm_state_ptr);
+	if (ret != 0) {
+		DRM_INFO_ONCE("%s:%d MST_DSC dm_atomic_get_state() failed\n", __func__, __LINE__);
+		return ret;
+	}
+	dm_state = *dm_state_ptr;
+
+	/*
+	 * create local vailable for dc_state. copy content of streams of dm_state->context
+	 * to local variable. make sure stream pointer of local variable not the same as stream
+	 * from dm_state->context.
+	 */
+
+	local_dc_state = vmalloc(sizeof(struct dc_state));
+	if (!local_dc_state)
+		return -ENOMEM;
+	memcpy(local_dc_state, dm_state->context, sizeof(struct dc_state));
+
+	for (i = 0; i < local_dc_state->stream_count; i++) {
+		struct dc_stream_state *stream = dm_state->context->streams[i];
+		int ind = find_crtc_index_in_state_by_stream(state, stream);
+
+		if (ind >= 0) {
+			struct drm_connector *connector;
+			struct drm_connector_state *drm_new_conn_state;
+			struct dm_connector_state *dm_new_conn_state;
+			struct dm_crtc_state *dm_old_crtc_state;
+
+			connector =
+				amdgpu_dm_find_first_crtc_matching_connector(state,
+									     state->crtcs[ind].ptr);
+			if (!connector)
+				continue;
+
+			drm_new_conn_state =
+				drm_atomic_get_new_connector_state(state,
+								   connector);
+			dm_new_conn_state = to_dm_connector_state(drm_new_conn_state);
+			dm_old_crtc_state = to_dm_crtc_state(state->crtcs[ind].old_state);
+
+			local_dc_state->streams[i] =
+				create_validate_stream_for_sink(connector,
+								&state->crtcs[ind].new_state->mode,
+								dm_new_conn_state,
+								dm_old_crtc_state->stream);
+			if (local_dc_state->streams[i] == NULL) {
+				ret = -EINVAL;
+				break;
+			}
+		}
+	}
+
+	if (ret != 0)
+		goto clean_exit;
+
+	ret = pre_compute_mst_dsc_configs_for_state(state, local_dc_state, vars);
+	if (ret != 0) {
+		DRM_INFO_ONCE("%s:%d MST_DSC dsc pre_compute_mst_dsc_configs_for_state() failed\n",
+				__func__, __LINE__);
+		ret = -EINVAL;
+		goto clean_exit;
+	}
+
+	/*
+	 * compare local_streams -> timing  with dm_state->context,
+	 * if the same set crtc_state->mode-change = 0;
+	 */
+	for (i = 0; i < local_dc_state->stream_count; i++) {
+		struct dc_stream_state *stream = dm_state->context->streams[i];
+
+		if (local_dc_state->streams[i] &&
+		    dc_is_timing_changed(stream, local_dc_state->streams[i])) {
+			DRM_INFO_ONCE("%s:%d MST_DSC crtc[%d] needs mode_change\n", __func__, __LINE__, i);
+		} else {
+			int ind = find_crtc_index_in_state_by_stream(state, stream);
+
+			if (ind >= 0) {
+				DRM_INFO_ONCE("%s:%d MST_DSC no mode changed for stream 0x%p\n",
+						__func__, __LINE__, stream);
+				state->crtcs[ind].new_state->mode_changed = 0;
+			}
+		}
+	}
+clean_exit:
+	for (i = 0; i < local_dc_state->stream_count; i++) {
+		struct dc_stream_state *stream = dm_state->context->streams[i];
+
+		if (local_dc_state->streams[i] != stream)
+			dc_stream_release(local_dc_state->streams[i]);
+	}
+
+	vfree(local_dc_state);
+
+	return ret;
+}
+
+static bool is_dsc_common_config_possible(struct dc_stream_state *stream,
+					  struct dc_dsc_bw_range *bw_range)
+{
+	struct dc_dsc_policy dsc_policy = {0};
+	bool is_dsc_possible;
+
+	dc_dsc_get_policy_for_timing(&stream->timing, 0, &dsc_policy, dc_link_get_highest_encoding_format(stream->link));
+	is_dsc_possible = dc_dsc_compute_bandwidth_range(stream->sink->ctx->dc->res_pool->dscs[0],
+							 stream->sink->ctx->dc->debug.dsc_min_slice_height_override,
+							 dsc_policy.min_target_bpp * 16,
+							 dsc_policy.max_target_bpp * 16,
+							 &stream->sink->dsc_caps.dsc_dec_caps,
+							 &stream->timing, dc_link_get_highest_encoding_format(stream->link), bw_range);
+
+	return is_dsc_possible;
+}
+#endif
+
+#if defined(CONFIG_DRM_AMD_DC_FP)
+static bool dp_get_link_current_set_bw(struct drm_dp_aux *aux, uint32_t *cur_link_bw)
+{
+	uint32_t total_data_bw_efficiency_x10000 = 0;
+	uint32_t link_rate_per_lane_kbps = 0;
+	enum dc_link_rate link_rate;
+	union lane_count_set lane_count;
+	u8 dp_link_encoding;
+	u8 link_bw_set = 0;
+	u8 data[16] = {0};
+
+	*cur_link_bw = 0;
+
+	if (drm_dp_dpcd_read(aux, DP_LINK_BW_SET, data, 16) != 16)
+		return false;
+
+	dp_link_encoding = data[DP_MAIN_LINK_CHANNEL_CODING_SET - DP_LINK_BW_SET];
+	link_bw_set = data[DP_LINK_BW_SET - DP_LINK_BW_SET];
+	lane_count.raw = data[DP_LANE_COUNT_SET - DP_LINK_BW_SET];
+
+	drm_dbg_dp(aux->drm_dev, "MST_DSC downlink setting: %d, 0x%x x %d\n",
+		   dp_link_encoding, link_bw_set, lane_count.bits.LANE_COUNT_SET);
+
+	switch (dp_link_encoding) {
+	case DP_8b_10b_ENCODING:
+		link_rate = link_bw_set;
+		link_rate_per_lane_kbps = link_rate * LINK_RATE_REF_FREQ_IN_KHZ * BITS_PER_DP_BYTE;
+		total_data_bw_efficiency_x10000 = DATA_EFFICIENCY_8b_10b_x10000;
+		total_data_bw_efficiency_x10000 /= 100;
+		total_data_bw_efficiency_x10000 *= DATA_EFFICIENCY_8b_10b_FEC_EFFICIENCY_x100;
+		break;
+	case DP_128b_132b_ENCODING:
+		switch (link_bw_set) {
+		case DP_LINK_BW_10:
+			link_rate = LINK_RATE_UHBR10;
+			break;
+		case DP_LINK_BW_13_5:
+			link_rate = LINK_RATE_UHBR13_5;
+			break;
+		case DP_LINK_BW_20:
+			link_rate = LINK_RATE_UHBR20;
+			break;
+		default:
+			return false;
+		}
+
+		link_rate_per_lane_kbps = link_rate * 10000;
+		total_data_bw_efficiency_x10000 = DATA_EFFICIENCY_128b_132b_x10000;
+		break;
+	default:
+		return false;
+	}
+
+	*cur_link_bw = link_rate_per_lane_kbps * lane_count.bits.LANE_COUNT_SET / 10000 * total_data_bw_efficiency_x10000;
 	return true;
 }
+#endif
 
+enum dc_status dm_dp_mst_is_port_support_mode(
+	struct amdgpu_dm_connector *aconnector,
+	struct dc_stream_state *stream)
+{
+#if defined(CONFIG_DRM_AMD_DC_FP)
+	int branch_max_throughput_mps = 0;
+	struct dc_link_settings cur_link_settings;
+	uint32_t end_to_end_bw_in_kbps = 0;
+	uint32_t root_link_bw_in_kbps = 0;
+	uint32_t virtual_channel_bw_in_kbps = 0;
+	struct dc_dsc_bw_range bw_range = {0};
+	struct dc_dsc_config_options dsc_options = {0};
+	uint32_t stream_kbps;
+
+	/* DSC unnecessary case
+	 * Check if timing could be supported within end-to-end BW
+	 */
+	stream_kbps =
+		dc_bandwidth_in_kbps_from_timing(&stream->timing,
+			dc_link_get_highest_encoding_format(stream->link));
+	cur_link_settings = stream->link->verified_link_cap;
+	root_link_bw_in_kbps = dc_link_bandwidth_kbps(aconnector->dc_link, &cur_link_settings);
+	virtual_channel_bw_in_kbps = pbn_to_kbps(aconnector->mst_output_port->full_pbn, true);
+
+	/* pick the end to end bw bottleneck */
+	end_to_end_bw_in_kbps = min(root_link_bw_in_kbps, virtual_channel_bw_in_kbps);
+
+	if (stream_kbps <= end_to_end_bw_in_kbps) {
+		DRM_DEBUG_DRIVER("MST_DSC no dsc required. End-to-end bw sufficient\n");
+		return DC_OK;
+	}
+
+	/*DSC necessary case*/
+	if (!aconnector->dsc_aux)
+		return DC_FAIL_BANDWIDTH_VALIDATE;
+
+	if (is_dsc_common_config_possible(stream, &bw_range)) {
+
+		/*capable of dsc passthough. dsc bitstream along the entire path*/
+		if (aconnector->mst_output_port->passthrough_aux) {
+			if (bw_range.min_kbps > end_to_end_bw_in_kbps) {
+				DRM_DEBUG_DRIVER("MST_DSC dsc passthrough and decode at endpoint"
+						 "Max dsc compression bw can't fit into end-to-end bw\n");
+				return DC_FAIL_BANDWIDTH_VALIDATE;
+			}
+		} else {
+			/*dsc bitstream decoded at the dp last link*/
+			struct drm_dp_mst_port *immediate_upstream_port = NULL;
+			uint32_t end_link_bw = 0;
+
+			/*Get last DP link BW capability. Mode shall be supported by Legacy peer*/
+			if (aconnector->mst_output_port->pdt != DP_PEER_DEVICE_DP_LEGACY_CONV &&
+				aconnector->mst_output_port->pdt != DP_PEER_DEVICE_NONE) {
+				if (aconnector->vc_full_pbn != aconnector->mst_output_port->full_pbn) {
+					dp_get_link_current_set_bw(&aconnector->mst_output_port->aux, &end_link_bw);
+					aconnector->vc_full_pbn = aconnector->mst_output_port->full_pbn;
+					aconnector->mst_local_bw = end_link_bw;
+				} else {
+					end_link_bw = aconnector->mst_local_bw;
+				}
+
+				if (end_link_bw > 0 &&
+				    stream_kbps > end_link_bw &&
+				    aconnector->branch_ieee_oui != DP_BRANCH_DEVICE_ID_90CC24) {
+					DRM_DEBUG_DRIVER("MST_DSC dsc decode at last link. "
+							 "Mode required bw can't fit into last link\n");
+					return DC_FAIL_BANDWIDTH_VALIDATE;
+				}
+			}
+
+			/*Get virtual channel bandwidth between source and the link before the last link*/
+			if (aconnector->mst_output_port->parent->port_parent)
+				immediate_upstream_port = aconnector->mst_output_port->parent->port_parent;
+
+			if (immediate_upstream_port) {
+				virtual_channel_bw_in_kbps = pbn_to_kbps(immediate_upstream_port->full_pbn, true);
+				virtual_channel_bw_in_kbps = min(root_link_bw_in_kbps, virtual_channel_bw_in_kbps);
+			} else {
+				/* For topology LCT 1 case - only one mstb*/
+				virtual_channel_bw_in_kbps = root_link_bw_in_kbps;
+			}
+
+			if (bw_range.min_kbps > virtual_channel_bw_in_kbps) {
+				DRM_DEBUG_DRIVER("MST_DSC dsc decode at last link."
+						 "Max dsc compression can't fit into MST available bw\n");
+				return DC_FAIL_BANDWIDTH_VALIDATE;
+			}
+		}
+
+		/*Confirm if we can obtain dsc config*/
+		dc_dsc_get_default_config_option(stream->link->dc, &dsc_options);
+		dsc_options.max_target_bpp_limit_override_x16 = aconnector->base.display_info.max_dsc_bpp * 16;
+		if (dc_dsc_compute_config(stream->sink->ctx->dc->res_pool->dscs[0],
+				&stream->sink->dsc_caps.dsc_dec_caps,
+				&dsc_options,
+				end_to_end_bw_in_kbps,
+				&stream->timing,
+				dc_link_get_highest_encoding_format(stream->link),
+				&stream->timing.dsc_cfg)) {
+			stream->timing.flags.DSC = 1;
+			DRM_DEBUG_DRIVER("MST_DSC require dsc and dsc config found\n");
+		} else {
+			DRM_DEBUG_DRIVER("MST_DSC require dsc but can't find appropriate dsc config\n");
+			return DC_FAIL_BANDWIDTH_VALIDATE;
+		}
+
+		/* check is mst dsc output bandwidth branch_overall_throughput_0_mps */
+		switch (stream->timing.pixel_encoding) {
+		case PIXEL_ENCODING_RGB:
+		case PIXEL_ENCODING_YCBCR444:
+			branch_max_throughput_mps =
+				aconnector->dc_sink->dsc_caps.dsc_dec_caps.branch_overall_throughput_0_mps;
+			break;
+		case PIXEL_ENCODING_YCBCR422:
+		case PIXEL_ENCODING_YCBCR420:
+			branch_max_throughput_mps =
+				aconnector->dc_sink->dsc_caps.dsc_dec_caps.branch_overall_throughput_1_mps;
+			break;
+		default:
+			break;
+		}
+
+		if (branch_max_throughput_mps != 0 &&
+			((stream->timing.pix_clk_100hz / 10) >  branch_max_throughput_mps * 1000)) {
+			DRM_DEBUG_DRIVER("MST_DSC require dsc but max throughput mps fails\n");
+			return DC_FAIL_BANDWIDTH_VALIDATE;
+		}
+	} else {
+		DRM_DEBUG_DRIVER("MST_DSC require dsc but can't find common dsc config\n");
+		return DC_FAIL_BANDWIDTH_VALIDATE;
+	}
 #endif
+	return DC_OK;
+}
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.h b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.h
index 900d3f7a8498..6f7ea684b555 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.h
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.h
@@ -1,3 +1,4 @@
+/* SPDX-License-Identifier: MIT */
 /*
  * Copyright 2012-15 Advanced Micro Devices, Inc.
  *
@@ -26,10 +27,40 @@
 #ifndef __DAL_AMDGPU_DM_MST_TYPES_H__
 #define __DAL_AMDGPU_DM_MST_TYPES_H__
 
+#define DP_BRANCH_DEVICE_ID_90CC24 0x90CC24
+
+#define SYNAPTICS_RC_COMMAND       0x4B2
+#define SYNAPTICS_RC_RESULT        0x4B3
+#define SYNAPTICS_RC_LENGTH        0x4B8
+#define SYNAPTICS_RC_OFFSET        0x4BC
+#define SYNAPTICS_RC_DATA          0x4C0
+
+#define DP_BRANCH_VENDOR_SPECIFIC_START 0x50C
+
+/**
+ * Panamera MST Hub detection
+ * Offset DPCD 050Eh == 0x5A indicates cascaded MST hub case
+ * Check from beginning of branch device vendor specific field (050Ch)
+ */
+#define IS_SYNAPTICS_PANAMERA(branchDevName) (((int)branchDevName[4] & 0xF0) == 0x50 ? 1 : 0)
+#define BRANCH_HW_REVISION_PANAMERA_A2 0x10
+#define SYNAPTICS_CASCADED_HUB_ID  0x5A
+#define IS_SYNAPTICS_CASCADED_PANAMERA(devName, data) ((IS_SYNAPTICS_PANAMERA(devName) && ((int)data[2] == SYNAPTICS_CASCADED_HUB_ID)) ? 1 : 0)
+
+#define PBN_FEC_OVERHEAD_MULTIPLIER_8B_10B     1031
+#define PBN_FEC_OVERHEAD_MULTIPLIER_128B_132B  1000
+
+enum mst_msg_ready_type {
+	NONE_MSG_RDY_EVENT = 0,
+	DOWN_REP_MSG_RDY_EVENT = 1,
+	UP_REQ_MSG_RDY_EVENT = 2,
+	DOWN_OR_UP_MSG_RDY_EVENT = 3
+};
+
 struct amdgpu_display_manager;
 struct amdgpu_dm_connector;
 
-int dm_mst_get_pbn_divider(struct dc_link *link);
+uint32_t dm_mst_get_pbn_divider(struct dc_link *link);
 
 void amdgpu_dm_initialize_dp_connector(struct amdgpu_display_manager *dm,
 				       struct amdgpu_dm_connector *aconnector,
@@ -38,7 +69,9 @@ void amdgpu_dm_initialize_dp_connector(struct amdgpu_display_manager *dm,
 void
 dm_dp_create_fake_mst_encoders(struct amdgpu_device *adev);
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
+void dm_handle_mst_sideband_msg_ready_event(
+	struct drm_dp_mst_topology_mgr *mgr,
+	enum mst_msg_ready_type msg_rdy_type);
 
 struct dsc_mst_fairness_vars {
 	int pbn;
@@ -47,9 +80,18 @@ struct dsc_mst_fairness_vars {
 	struct amdgpu_dm_connector *aconnector;
 };
 
-bool compute_mst_dsc_configs_for_state(struct drm_atomic_state *state,
-				       struct dc_state *dc_state,
-				       struct dsc_mst_fairness_vars *vars);
-#endif
+int compute_mst_dsc_configs_for_state(struct drm_atomic_state *state,
+				      struct dc_state *dc_state,
+				      struct dsc_mst_fairness_vars *vars);
+
+bool needs_dsc_aux_workaround(struct dc_link *link);
+
+int pre_validate_dsc(struct drm_atomic_state *state,
+		     struct dm_atomic_state **dm_state_ptr,
+		     struct dsc_mst_fairness_vars *vars);
+
+enum dc_status dm_dp_mst_is_port_support_mode(
+	struct amdgpu_dm_connector *aconnector,
+	struct dc_stream_state *stream);
 
 #endif
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_plane.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_plane.c
new file mode 100644
index 000000000000..2e3ee78999d9
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_plane.c
@@ -0,0 +1,1950 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_blend.h>
+#include "drm/drm_framebuffer.h"
+#include <drm/drm_gem_atomic_helper.h>
+#include <drm/drm_plane_helper.h>
+#include <drm/drm_gem_framebuffer_helper.h>
+#include <drm/drm_fourcc.h>
+
+#include "amdgpu.h"
+#include "dal_asic_id.h"
+#include "amdgpu_display.h"
+#include "amdgpu_dm_trace.h"
+#include "amdgpu_dm_plane.h"
+#include "amdgpu_dm_colorop.h"
+#include "gc/gc_11_0_0_offset.h"
+#include "gc/gc_11_0_0_sh_mask.h"
+
+/*
+ * TODO: these are currently initialized to rgb formats only.
+ * For future use cases we should either initialize them dynamically based on
+ * plane capabilities, or initialize this array to all formats, so internal drm
+ * check will succeed, and let DC implement proper check
+ */
+static const uint32_t rgb_formats[] = {
+	DRM_FORMAT_XRGB8888,
+	DRM_FORMAT_ARGB8888,
+	DRM_FORMAT_RGBA8888,
+	DRM_FORMAT_XRGB2101010,
+	DRM_FORMAT_XBGR2101010,
+	DRM_FORMAT_ARGB2101010,
+	DRM_FORMAT_ABGR2101010,
+	DRM_FORMAT_XRGB16161616,
+	DRM_FORMAT_XBGR16161616,
+	DRM_FORMAT_ARGB16161616,
+	DRM_FORMAT_ABGR16161616,
+	DRM_FORMAT_XBGR8888,
+	DRM_FORMAT_ABGR8888,
+	DRM_FORMAT_RGB565,
+};
+
+static const uint32_t overlay_formats[] = {
+	DRM_FORMAT_XRGB8888,
+	DRM_FORMAT_ARGB8888,
+	DRM_FORMAT_RGBA8888,
+	DRM_FORMAT_XBGR8888,
+	DRM_FORMAT_ABGR8888,
+	DRM_FORMAT_RGB565,
+	DRM_FORMAT_NV21,
+	DRM_FORMAT_NV12,
+	DRM_FORMAT_P010
+};
+
+static const uint32_t video_formats[] = {
+	DRM_FORMAT_NV21,
+	DRM_FORMAT_NV12,
+	DRM_FORMAT_P010
+};
+
+static const u32 cursor_formats[] = {
+	DRM_FORMAT_ARGB8888
+};
+
+enum dm_micro_swizzle {
+	MICRO_SWIZZLE_Z = 0,
+	MICRO_SWIZZLE_S = 1,
+	MICRO_SWIZZLE_D = 2,
+	MICRO_SWIZZLE_R = 3
+};
+
+const struct drm_format_info *amdgpu_dm_plane_get_format_info(u32 pixel_format, u64 modifier)
+{
+	return amdgpu_lookup_format_info(pixel_format, modifier);
+}
+
+void amdgpu_dm_plane_fill_blending_from_plane_state(const struct drm_plane_state *plane_state,
+			       bool *per_pixel_alpha, bool *pre_multiplied_alpha,
+			       bool *global_alpha, int *global_alpha_value)
+{
+	*per_pixel_alpha = false;
+	*pre_multiplied_alpha = true;
+	*global_alpha = false;
+	*global_alpha_value = 0xff;
+
+
+	if (plane_state->pixel_blend_mode == DRM_MODE_BLEND_PREMULTI ||
+		plane_state->pixel_blend_mode == DRM_MODE_BLEND_COVERAGE) {
+		static const uint32_t alpha_formats[] = {
+			DRM_FORMAT_ARGB8888,
+			DRM_FORMAT_RGBA8888,
+			DRM_FORMAT_ABGR8888,
+			DRM_FORMAT_ARGB2101010,
+			DRM_FORMAT_ABGR2101010,
+			DRM_FORMAT_ARGB16161616,
+			DRM_FORMAT_ABGR16161616,
+			DRM_FORMAT_ARGB16161616F,
+		};
+		uint32_t format = plane_state->fb->format->format;
+		unsigned int i;
+
+		for (i = 0; i < ARRAY_SIZE(alpha_formats); ++i) {
+			if (format == alpha_formats[i]) {
+				*per_pixel_alpha = true;
+				break;
+			}
+		}
+
+		if (*per_pixel_alpha && plane_state->pixel_blend_mode == DRM_MODE_BLEND_COVERAGE)
+			*pre_multiplied_alpha = false;
+	}
+
+	if (plane_state->alpha < 0xffff) {
+		*global_alpha = true;
+		*global_alpha_value = plane_state->alpha >> 8;
+	}
+}
+
+static void amdgpu_dm_plane_add_modifier(uint64_t **mods, uint64_t *size, uint64_t *cap, uint64_t mod)
+{
+	if (!*mods)
+		return;
+
+	if (*cap - *size < 1) {
+		uint64_t new_cap = *cap * 2;
+		uint64_t *new_mods = kmalloc_array(new_cap, sizeof(uint64_t), GFP_KERNEL);
+
+		if (!new_mods) {
+			kfree(*mods);
+			*mods = NULL;
+			return;
+		}
+
+		memcpy(new_mods, *mods, sizeof(uint64_t) * *size);
+		kfree(*mods);
+		*mods = new_mods;
+		*cap = new_cap;
+	}
+
+	(*mods)[*size] = mod;
+	*size += 1;
+}
+
+static bool amdgpu_dm_plane_modifier_has_dcc(uint64_t modifier)
+{
+	return IS_AMD_FMT_MOD(modifier) && AMD_FMT_MOD_GET(DCC, modifier);
+}
+
+static unsigned int amdgpu_dm_plane_modifier_gfx9_swizzle_mode(uint64_t modifier)
+{
+	if (modifier == DRM_FORMAT_MOD_LINEAR)
+		return 0;
+
+	return AMD_FMT_MOD_GET(TILE, modifier);
+}
+
+static void amdgpu_dm_plane_fill_gfx8_tiling_info_from_flags(struct dc_tiling_info *tiling_info,
+							     uint64_t tiling_flags)
+{
+	/* Fill GFX8 params */
+	if (AMDGPU_TILING_GET(tiling_flags, ARRAY_MODE) == DC_ARRAY_2D_TILED_THIN1) {
+		unsigned int bankw, bankh, mtaspect, tile_split, num_banks;
+
+		bankw = AMDGPU_TILING_GET(tiling_flags, BANK_WIDTH);
+		bankh = AMDGPU_TILING_GET(tiling_flags, BANK_HEIGHT);
+		mtaspect = AMDGPU_TILING_GET(tiling_flags, MACRO_TILE_ASPECT);
+		tile_split = AMDGPU_TILING_GET(tiling_flags, TILE_SPLIT);
+		num_banks = AMDGPU_TILING_GET(tiling_flags, NUM_BANKS);
+
+		tiling_info->gfxversion = DcGfxVersion8;
+		/* XXX fix me for VI */
+		tiling_info->gfx8.num_banks = num_banks;
+		tiling_info->gfx8.array_mode =
+				DC_ARRAY_2D_TILED_THIN1;
+		tiling_info->gfx8.tile_split = tile_split;
+		tiling_info->gfx8.bank_width = bankw;
+		tiling_info->gfx8.bank_height = bankh;
+		tiling_info->gfx8.tile_aspect = mtaspect;
+		tiling_info->gfx8.tile_mode =
+				DC_ADDR_SURF_MICRO_TILING_DISPLAY;
+	} else if (AMDGPU_TILING_GET(tiling_flags, ARRAY_MODE)
+			== DC_ARRAY_1D_TILED_THIN1) {
+		tiling_info->gfx8.array_mode = DC_ARRAY_1D_TILED_THIN1;
+	}
+
+	tiling_info->gfx8.pipe_config =
+			AMDGPU_TILING_GET(tiling_flags, PIPE_CONFIG);
+}
+
+static void amdgpu_dm_plane_fill_gfx9_tiling_info_from_device(const struct amdgpu_device *adev,
+							      struct dc_tiling_info *tiling_info)
+{
+	/* Fill GFX9 params */
+	tiling_info->gfx9.num_pipes =
+		adev->gfx.config.gb_addr_config_fields.num_pipes;
+	tiling_info->gfx9.num_banks =
+		adev->gfx.config.gb_addr_config_fields.num_banks;
+	tiling_info->gfx9.pipe_interleave =
+		adev->gfx.config.gb_addr_config_fields.pipe_interleave_size;
+	tiling_info->gfx9.num_shader_engines =
+		adev->gfx.config.gb_addr_config_fields.num_se;
+	tiling_info->gfx9.max_compressed_frags =
+		adev->gfx.config.gb_addr_config_fields.max_compress_frags;
+	tiling_info->gfx9.num_rb_per_se =
+		adev->gfx.config.gb_addr_config_fields.num_rb_per_se;
+	tiling_info->gfx9.shaderEnable = 1;
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) >= IP_VERSION(10, 3, 0))
+		tiling_info->gfx9.num_pkrs = adev->gfx.config.gb_addr_config_fields.num_pkrs;
+}
+
+static void amdgpu_dm_plane_fill_gfx9_tiling_info_from_modifier(const struct amdgpu_device *adev,
+								struct dc_tiling_info *tiling_info,
+								uint64_t modifier)
+{
+	unsigned int mod_bank_xor_bits = AMD_FMT_MOD_GET(BANK_XOR_BITS, modifier);
+	unsigned int mod_pipe_xor_bits = AMD_FMT_MOD_GET(PIPE_XOR_BITS, modifier);
+	unsigned int pkrs_log2 = AMD_FMT_MOD_GET(PACKERS, modifier);
+	unsigned int pipes_log2;
+
+	pipes_log2 = min(5u, mod_pipe_xor_bits);
+
+	amdgpu_dm_plane_fill_gfx9_tiling_info_from_device(adev, tiling_info);
+
+	if (!IS_AMD_FMT_MOD(modifier))
+		return;
+
+	tiling_info->gfx9.num_pipes = 1u << pipes_log2;
+	tiling_info->gfx9.num_shader_engines = 1u << (mod_pipe_xor_bits - pipes_log2);
+
+	if (adev->family >= AMDGPU_FAMILY_NV) {
+		tiling_info->gfx9.num_pkrs = 1u << pkrs_log2;
+	} else {
+		tiling_info->gfx9.num_banks = 1u << mod_bank_xor_bits;
+
+		/* for DCC we know it isn't rb aligned, so rb_per_se doesn't matter. */
+	}
+}
+
+static int amdgpu_dm_plane_validate_dcc(struct amdgpu_device *adev,
+					const enum surface_pixel_format format,
+					const enum dc_rotation_angle rotation,
+					const struct dc_tiling_info *tiling_info,
+					const struct dc_plane_dcc_param *dcc,
+					const struct dc_plane_address *address,
+					const struct plane_size *plane_size)
+{
+	struct dc *dc = adev->dm.dc;
+	struct dc_dcc_surface_param input;
+	struct dc_surface_dcc_cap output;
+
+	memset(&input, 0, sizeof(input));
+	memset(&output, 0, sizeof(output));
+
+	if (!dcc->enable)
+		return 0;
+
+	if (adev->family < AMDGPU_FAMILY_GC_12_0_0 &&
+	    format >= SURFACE_PIXEL_FORMAT_VIDEO_BEGIN)
+		return -EINVAL;
+
+	if (!dc->cap_funcs.get_dcc_compression_cap)
+		return -EINVAL;
+
+	input.format = format;
+	input.surface_size.width = plane_size->surface_size.width;
+	input.surface_size.height = plane_size->surface_size.height;
+	input.swizzle_mode = tiling_info->gfx9.swizzle;
+
+	if (rotation == ROTATION_ANGLE_0 || rotation == ROTATION_ANGLE_180)
+		input.scan = SCAN_DIRECTION_HORIZONTAL;
+	else if (rotation == ROTATION_ANGLE_90 || rotation == ROTATION_ANGLE_270)
+		input.scan = SCAN_DIRECTION_VERTICAL;
+
+	if (!dc->cap_funcs.get_dcc_compression_cap(dc, &input, &output))
+		return -EINVAL;
+
+	if (!output.capable)
+		return -EINVAL;
+
+	if (dcc->independent_64b_blks == 0 &&
+	    output.grph.rgb.independent_64b_blks != 0)
+		return -EINVAL;
+
+	return 0;
+}
+
+static int amdgpu_dm_plane_fill_gfx9_plane_attributes_from_modifiers(struct amdgpu_device *adev,
+								     const struct amdgpu_framebuffer *afb,
+								     const enum surface_pixel_format format,
+								     const enum dc_rotation_angle rotation,
+								     const struct plane_size *plane_size,
+								     struct dc_tiling_info *tiling_info,
+								     struct dc_plane_dcc_param *dcc,
+								     struct dc_plane_address *address)
+{
+	const uint64_t modifier = afb->base.modifier;
+	int ret = 0;
+
+	amdgpu_dm_plane_fill_gfx9_tiling_info_from_modifier(adev, tiling_info, modifier);
+	tiling_info->gfx9.swizzle = amdgpu_dm_plane_modifier_gfx9_swizzle_mode(modifier);
+	tiling_info->gfxversion = DcGfxVersion9;
+
+	if (amdgpu_dm_plane_modifier_has_dcc(modifier)) {
+		uint64_t dcc_address = afb->address + afb->base.offsets[1];
+		bool independent_64b_blks = AMD_FMT_MOD_GET(DCC_INDEPENDENT_64B, modifier);
+		bool independent_128b_blks = AMD_FMT_MOD_GET(DCC_INDEPENDENT_128B, modifier);
+
+		dcc->enable = 1;
+		dcc->meta_pitch = afb->base.pitches[1];
+		dcc->independent_64b_blks = independent_64b_blks;
+		if (AMD_FMT_MOD_GET(TILE_VERSION, modifier) >= AMD_FMT_MOD_TILE_VER_GFX10_RBPLUS) {
+			if (independent_64b_blks && independent_128b_blks)
+				dcc->dcc_ind_blk = hubp_ind_block_64b_no_128bcl;
+			else if (independent_128b_blks)
+				dcc->dcc_ind_blk = hubp_ind_block_128b;
+			else if (independent_64b_blks && !independent_128b_blks)
+				dcc->dcc_ind_blk = hubp_ind_block_64b;
+			else
+				dcc->dcc_ind_blk = hubp_ind_block_unconstrained;
+		} else {
+			if (independent_64b_blks)
+				dcc->dcc_ind_blk = hubp_ind_block_64b;
+			else
+				dcc->dcc_ind_blk = hubp_ind_block_unconstrained;
+		}
+
+		address->grph.meta_addr.low_part = lower_32_bits(dcc_address);
+		address->grph.meta_addr.high_part = upper_32_bits(dcc_address);
+	}
+
+	ret = amdgpu_dm_plane_validate_dcc(adev, format, rotation, tiling_info, dcc, address, plane_size);
+	if (ret)
+		drm_dbg_kms(adev_to_drm(adev), "amdgpu_dm_plane_validate_dcc: returned error: %d\n", ret);
+
+	return ret;
+}
+
+static int amdgpu_dm_plane_fill_gfx12_plane_attributes_from_modifiers(struct amdgpu_device *adev,
+								      const struct amdgpu_framebuffer *afb,
+								      const enum surface_pixel_format format,
+								      const enum dc_rotation_angle rotation,
+								      const struct plane_size *plane_size,
+								      struct dc_tiling_info *tiling_info,
+								      struct dc_plane_dcc_param *dcc,
+								      struct dc_plane_address *address)
+{
+	const uint64_t modifier = afb->base.modifier;
+	int ret = 0;
+
+	/* TODO: Most of this function shouldn't be needed on GFX12. */
+	amdgpu_dm_plane_fill_gfx9_tiling_info_from_device(adev, tiling_info);
+
+	tiling_info->gfx9.swizzle = amdgpu_dm_plane_modifier_gfx9_swizzle_mode(modifier);
+	tiling_info->gfxversion = DcGfxAddr3;
+
+	if (amdgpu_dm_plane_modifier_has_dcc(modifier)) {
+		int max_compressed_block = AMD_FMT_MOD_GET(DCC_MAX_COMPRESSED_BLOCK, modifier);
+
+		dcc->enable = 1;
+		dcc->independent_64b_blks = max_compressed_block == 0;
+
+		if (max_compressed_block == 0)
+			dcc->dcc_ind_blk = hubp_ind_block_64b;
+		else if (max_compressed_block == 1)
+			dcc->dcc_ind_blk = hubp_ind_block_128b;
+		else
+			dcc->dcc_ind_blk = hubp_ind_block_unconstrained;
+	}
+
+	/* TODO: This seems wrong because there is no DCC plane on GFX12. */
+	ret = amdgpu_dm_plane_validate_dcc(adev, format, rotation, tiling_info, dcc, address, plane_size);
+	if (ret)
+		drm_dbg_kms(adev_to_drm(adev), "amdgpu_dm_plane_validate_dcc: returned error: %d\n", ret);
+
+	return ret;
+}
+
+static void amdgpu_dm_plane_add_gfx10_1_modifiers(const struct amdgpu_device *adev,
+						  uint64_t **mods,
+						  uint64_t *size,
+						  uint64_t *capacity)
+{
+	int pipe_xor_bits = ilog2(adev->gfx.config.gb_addr_config_fields.num_pipes);
+
+	amdgpu_dm_plane_add_modifier(mods, size, capacity, AMD_FMT_MOD |
+				     AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_R_X) |
+				     AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX10) |
+				     AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
+				     AMD_FMT_MOD_SET(DCC, 1) |
+				     AMD_FMT_MOD_SET(DCC_CONSTANT_ENCODE, 1) |
+				     AMD_FMT_MOD_SET(DCC_INDEPENDENT_64B, 1) |
+				     AMD_FMT_MOD_SET(DCC_MAX_COMPRESSED_BLOCK, AMD_FMT_MOD_DCC_BLOCK_64B));
+
+	amdgpu_dm_plane_add_modifier(mods, size, capacity, AMD_FMT_MOD |
+				     AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_R_X) |
+				     AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX10) |
+				     AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
+				     AMD_FMT_MOD_SET(DCC, 1) |
+				     AMD_FMT_MOD_SET(DCC_RETILE, 1) |
+				     AMD_FMT_MOD_SET(DCC_CONSTANT_ENCODE, 1) |
+				     AMD_FMT_MOD_SET(DCC_INDEPENDENT_64B, 1) |
+				     AMD_FMT_MOD_SET(DCC_MAX_COMPRESSED_BLOCK, AMD_FMT_MOD_DCC_BLOCK_64B));
+
+	amdgpu_dm_plane_add_modifier(mods, size, capacity, AMD_FMT_MOD |
+				     AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_R_X) |
+				     AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX10) |
+				     AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits));
+
+	amdgpu_dm_plane_add_modifier(mods, size, capacity, AMD_FMT_MOD |
+				     AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_S_X) |
+				     AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX10) |
+				     AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits));
+
+
+	/* Only supported for 64bpp, will be filtered in amdgpu_dm_plane_format_mod_supported */
+	amdgpu_dm_plane_add_modifier(mods, size, capacity, AMD_FMT_MOD |
+				     AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_D) |
+				     AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX9));
+
+	amdgpu_dm_plane_add_modifier(mods, size, capacity, AMD_FMT_MOD |
+				     AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_S) |
+				     AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX9));
+}
+
+static void amdgpu_dm_plane_add_gfx9_modifiers(const struct amdgpu_device *adev,
+					       uint64_t **mods,
+					       uint64_t *size,
+					       uint64_t *capacity)
+{
+	int pipes = ilog2(adev->gfx.config.gb_addr_config_fields.num_pipes);
+	int pipe_xor_bits = min(8, pipes +
+				ilog2(adev->gfx.config.gb_addr_config_fields.num_se));
+	int bank_xor_bits = min(8 - pipe_xor_bits,
+				ilog2(adev->gfx.config.gb_addr_config_fields.num_banks));
+	int rb = ilog2(adev->gfx.config.gb_addr_config_fields.num_se) +
+		 ilog2(adev->gfx.config.gb_addr_config_fields.num_rb_per_se);
+
+
+	if (adev->family == AMDGPU_FAMILY_RV) {
+		/* Raven2 and later */
+		bool has_constant_encode = adev->asic_type > CHIP_RAVEN || adev->external_rev_id >= 0x81;
+
+		/*
+		 * No _D DCC swizzles yet because we only allow 32bpp, which
+		 * doesn't support _D on DCN
+		 */
+
+		if (has_constant_encode) {
+			amdgpu_dm_plane_add_modifier(mods, size, capacity, AMD_FMT_MOD |
+						     AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_S_X) |
+						     AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX9) |
+						     AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
+						     AMD_FMT_MOD_SET(BANK_XOR_BITS, bank_xor_bits) |
+						     AMD_FMT_MOD_SET(DCC, 1) |
+						     AMD_FMT_MOD_SET(DCC_INDEPENDENT_64B, 1) |
+						     AMD_FMT_MOD_SET(DCC_MAX_COMPRESSED_BLOCK, AMD_FMT_MOD_DCC_BLOCK_64B) |
+						     AMD_FMT_MOD_SET(DCC_CONSTANT_ENCODE, 1));
+		}
+
+		amdgpu_dm_plane_add_modifier(mods, size, capacity, AMD_FMT_MOD |
+					     AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_S_X) |
+					     AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX9) |
+					     AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
+					     AMD_FMT_MOD_SET(BANK_XOR_BITS, bank_xor_bits) |
+					     AMD_FMT_MOD_SET(DCC, 1) |
+					     AMD_FMT_MOD_SET(DCC_INDEPENDENT_64B, 1) |
+					     AMD_FMT_MOD_SET(DCC_MAX_COMPRESSED_BLOCK, AMD_FMT_MOD_DCC_BLOCK_64B) |
+					     AMD_FMT_MOD_SET(DCC_CONSTANT_ENCODE, 0));
+
+		if (has_constant_encode) {
+			amdgpu_dm_plane_add_modifier(mods, size, capacity, AMD_FMT_MOD |
+						     AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_S_X) |
+						     AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX9) |
+						     AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
+						     AMD_FMT_MOD_SET(BANK_XOR_BITS, bank_xor_bits) |
+						     AMD_FMT_MOD_SET(DCC, 1) |
+						     AMD_FMT_MOD_SET(DCC_RETILE, 1) |
+						     AMD_FMT_MOD_SET(DCC_INDEPENDENT_64B, 1) |
+						     AMD_FMT_MOD_SET(DCC_MAX_COMPRESSED_BLOCK, AMD_FMT_MOD_DCC_BLOCK_64B) |
+						     AMD_FMT_MOD_SET(DCC_CONSTANT_ENCODE, 1) |
+						     AMD_FMT_MOD_SET(RB, rb) |
+						     AMD_FMT_MOD_SET(PIPE, pipes));
+		}
+
+		amdgpu_dm_plane_add_modifier(mods, size, capacity, AMD_FMT_MOD |
+					     AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_S_X) |
+					     AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX9) |
+					     AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
+					     AMD_FMT_MOD_SET(BANK_XOR_BITS, bank_xor_bits) |
+					     AMD_FMT_MOD_SET(DCC, 1) |
+					     AMD_FMT_MOD_SET(DCC_RETILE, 1) |
+					     AMD_FMT_MOD_SET(DCC_INDEPENDENT_64B, 1) |
+					     AMD_FMT_MOD_SET(DCC_MAX_COMPRESSED_BLOCK, AMD_FMT_MOD_DCC_BLOCK_64B) |
+					     AMD_FMT_MOD_SET(DCC_CONSTANT_ENCODE, 0) |
+					     AMD_FMT_MOD_SET(RB, rb) |
+					     AMD_FMT_MOD_SET(PIPE, pipes));
+	}
+
+	/*
+	 * Only supported for 64bpp on Raven, will be filtered on format in
+	 * amdgpu_dm_plane_format_mod_supported.
+	 */
+	amdgpu_dm_plane_add_modifier(mods, size, capacity, AMD_FMT_MOD |
+				     AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_D_X) |
+				     AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX9) |
+				     AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
+				     AMD_FMT_MOD_SET(BANK_XOR_BITS, bank_xor_bits));
+
+	if (adev->family == AMDGPU_FAMILY_RV) {
+		amdgpu_dm_plane_add_modifier(mods, size, capacity, AMD_FMT_MOD |
+					     AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_S_X) |
+					     AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX9) |
+					     AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
+					     AMD_FMT_MOD_SET(BANK_XOR_BITS, bank_xor_bits));
+	}
+
+	/*
+	 * Only supported for 64bpp on Raven, will be filtered on format in
+	 * amdgpu_dm_plane_format_mod_supported.
+	 */
+	amdgpu_dm_plane_add_modifier(mods, size, capacity, AMD_FMT_MOD |
+				     AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_D) |
+				     AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX9));
+
+	if (adev->family == AMDGPU_FAMILY_RV) {
+		amdgpu_dm_plane_add_modifier(mods, size, capacity, AMD_FMT_MOD |
+					     AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_S) |
+					     AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX9));
+	}
+}
+
+static void amdgpu_dm_plane_add_gfx10_3_modifiers(const struct amdgpu_device *adev,
+						  uint64_t **mods,
+						  uint64_t *size,
+						  uint64_t *capacity)
+{
+	int pipe_xor_bits = ilog2(adev->gfx.config.gb_addr_config_fields.num_pipes);
+	int pkrs = ilog2(adev->gfx.config.gb_addr_config_fields.num_pkrs);
+
+	amdgpu_dm_plane_add_modifier(mods, size, capacity, AMD_FMT_MOD |
+				     AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_R_X) |
+				     AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX10_RBPLUS) |
+				     AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
+				     AMD_FMT_MOD_SET(PACKERS, pkrs) |
+				     AMD_FMT_MOD_SET(DCC, 1) |
+				     AMD_FMT_MOD_SET(DCC_CONSTANT_ENCODE, 1) |
+				     AMD_FMT_MOD_SET(DCC_INDEPENDENT_64B, 1) |
+				     AMD_FMT_MOD_SET(DCC_INDEPENDENT_128B, 1) |
+				     AMD_FMT_MOD_SET(DCC_MAX_COMPRESSED_BLOCK, AMD_FMT_MOD_DCC_BLOCK_64B));
+
+	amdgpu_dm_plane_add_modifier(mods, size, capacity, AMD_FMT_MOD |
+				     AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_R_X) |
+				     AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX10_RBPLUS) |
+				     AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
+				     AMD_FMT_MOD_SET(PACKERS, pkrs) |
+				     AMD_FMT_MOD_SET(DCC, 1) |
+				     AMD_FMT_MOD_SET(DCC_CONSTANT_ENCODE, 1) |
+				     AMD_FMT_MOD_SET(DCC_INDEPENDENT_128B, 1) |
+				     AMD_FMT_MOD_SET(DCC_MAX_COMPRESSED_BLOCK, AMD_FMT_MOD_DCC_BLOCK_128B));
+
+	amdgpu_dm_plane_add_modifier(mods, size, capacity, AMD_FMT_MOD |
+				     AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_R_X) |
+				     AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX10_RBPLUS) |
+				     AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
+				     AMD_FMT_MOD_SET(PACKERS, pkrs) |
+				     AMD_FMT_MOD_SET(DCC, 1) |
+				     AMD_FMT_MOD_SET(DCC_RETILE, 1) |
+				     AMD_FMT_MOD_SET(DCC_CONSTANT_ENCODE, 1) |
+				     AMD_FMT_MOD_SET(DCC_INDEPENDENT_64B, 1) |
+				     AMD_FMT_MOD_SET(DCC_INDEPENDENT_128B, 1) |
+				     AMD_FMT_MOD_SET(DCC_MAX_COMPRESSED_BLOCK, AMD_FMT_MOD_DCC_BLOCK_64B));
+
+	amdgpu_dm_plane_add_modifier(mods, size, capacity, AMD_FMT_MOD |
+				     AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_R_X) |
+				     AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX10_RBPLUS) |
+				     AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
+				     AMD_FMT_MOD_SET(PACKERS, pkrs) |
+				     AMD_FMT_MOD_SET(DCC, 1) |
+				     AMD_FMT_MOD_SET(DCC_RETILE, 1) |
+				     AMD_FMT_MOD_SET(DCC_CONSTANT_ENCODE, 1) |
+				     AMD_FMT_MOD_SET(DCC_INDEPENDENT_128B, 1) |
+				     AMD_FMT_MOD_SET(DCC_MAX_COMPRESSED_BLOCK, AMD_FMT_MOD_DCC_BLOCK_128B));
+
+	amdgpu_dm_plane_add_modifier(mods, size, capacity, AMD_FMT_MOD |
+				     AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_R_X) |
+				     AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX10_RBPLUS) |
+				     AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
+				     AMD_FMT_MOD_SET(PACKERS, pkrs));
+
+	amdgpu_dm_plane_add_modifier(mods, size, capacity, AMD_FMT_MOD |
+				     AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_S_X) |
+				     AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX10_RBPLUS) |
+				     AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
+				     AMD_FMT_MOD_SET(PACKERS, pkrs));
+
+	/* Only supported for 64bpp, will be filtered in amdgpu_dm_plane_format_mod_supported */
+	amdgpu_dm_plane_add_modifier(mods, size, capacity, AMD_FMT_MOD |
+				     AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_D) |
+				     AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX9));
+
+	amdgpu_dm_plane_add_modifier(mods, size, capacity, AMD_FMT_MOD |
+				     AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_S) |
+				     AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX9));
+}
+
+static void amdgpu_dm_plane_add_gfx11_modifiers(struct amdgpu_device *adev,
+		      uint64_t **mods, uint64_t *size, uint64_t *capacity)
+{
+	int num_pipes = 0;
+	int pipe_xor_bits = 0;
+	int num_pkrs = 0;
+	int pkrs = 0;
+	u32 gb_addr_config;
+	u8 i = 0;
+	unsigned int swizzle_r_x;
+	uint64_t modifier_r_x;
+	uint64_t modifier_dcc_best;
+	uint64_t modifier_dcc_4k;
+
+	/* TODO: GFX11 IP HW init hasnt finish and we get zero if we read from
+	 * adev->gfx.config.gb_addr_config_fields.num_{pkrs,pipes}
+	 */
+	gb_addr_config = RREG32_SOC15(GC, 0, regGB_ADDR_CONFIG);
+	ASSERT(gb_addr_config != 0);
+
+	num_pkrs = 1 << REG_GET_FIELD(gb_addr_config, GB_ADDR_CONFIG, NUM_PKRS);
+	pkrs = ilog2(num_pkrs);
+	num_pipes = 1 << REG_GET_FIELD(gb_addr_config, GB_ADDR_CONFIG, NUM_PIPES);
+	pipe_xor_bits = ilog2(num_pipes);
+
+	for (i = 0; i < 2; i++) {
+		/* Insert the best one first. */
+		/* R_X swizzle modes are the best for rendering and DCC requires them. */
+		if (num_pipes > 16)
+			swizzle_r_x = !i ? AMD_FMT_MOD_TILE_GFX11_256K_R_X : AMD_FMT_MOD_TILE_GFX9_64K_R_X;
+		else
+			swizzle_r_x = !i ? AMD_FMT_MOD_TILE_GFX9_64K_R_X : AMD_FMT_MOD_TILE_GFX11_256K_R_X;
+
+		modifier_r_x = AMD_FMT_MOD |
+			       AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX11) |
+			       AMD_FMT_MOD_SET(PIPE_XOR_BITS, pipe_xor_bits) |
+			       AMD_FMT_MOD_SET(TILE, swizzle_r_x) |
+			       AMD_FMT_MOD_SET(PACKERS, pkrs);
+
+		/* DCC_CONSTANT_ENCODE is not set because it can't vary with gfx11 (it's implied to be 1). */
+		modifier_dcc_best = modifier_r_x | AMD_FMT_MOD_SET(DCC, 1) |
+				    AMD_FMT_MOD_SET(DCC_INDEPENDENT_64B, 0) |
+				    AMD_FMT_MOD_SET(DCC_INDEPENDENT_128B, 1) |
+				    AMD_FMT_MOD_SET(DCC_MAX_COMPRESSED_BLOCK, AMD_FMT_MOD_DCC_BLOCK_128B);
+
+		/* DCC settings for 4K and greater resolutions. (required by display hw) */
+		modifier_dcc_4k = modifier_r_x | AMD_FMT_MOD_SET(DCC, 1) |
+				  AMD_FMT_MOD_SET(DCC_INDEPENDENT_64B, 1) |
+				  AMD_FMT_MOD_SET(DCC_INDEPENDENT_128B, 1) |
+				  AMD_FMT_MOD_SET(DCC_MAX_COMPRESSED_BLOCK, AMD_FMT_MOD_DCC_BLOCK_64B);
+
+		amdgpu_dm_plane_add_modifier(mods, size, capacity, modifier_dcc_best);
+		amdgpu_dm_plane_add_modifier(mods, size, capacity, modifier_dcc_4k);
+
+		amdgpu_dm_plane_add_modifier(mods, size, capacity, modifier_dcc_best | AMD_FMT_MOD_SET(DCC_RETILE, 1));
+		amdgpu_dm_plane_add_modifier(mods, size, capacity, modifier_dcc_4k | AMD_FMT_MOD_SET(DCC_RETILE, 1));
+
+		amdgpu_dm_plane_add_modifier(mods, size, capacity, modifier_r_x);
+	}
+
+	amdgpu_dm_plane_add_modifier(mods, size, capacity, AMD_FMT_MOD |
+				     AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX11) |
+				     AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX9_64K_D));
+}
+
+static void amdgpu_dm_plane_add_gfx12_modifiers(struct amdgpu_device *adev,
+		      uint64_t **mods, uint64_t *size, uint64_t *capacity)
+{
+	uint64_t ver = AMD_FMT_MOD | AMD_FMT_MOD_SET(TILE_VERSION, AMD_FMT_MOD_TILE_VER_GFX12);
+	uint64_t mod_256k = ver | AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX12_256K_2D);
+	uint64_t mod_64k = ver | AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX12_64K_2D);
+	uint64_t mod_4k = ver | AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX12_4K_2D);
+	uint64_t mod_256b = ver | AMD_FMT_MOD_SET(TILE, AMD_FMT_MOD_TILE_GFX12_256B_2D);
+	uint64_t dcc = ver | AMD_FMT_MOD_SET(DCC, 1);
+	uint8_t max_comp_block[] = {2, 1, 0};
+	uint64_t max_comp_block_mod[ARRAY_SIZE(max_comp_block)] = {0};
+	uint8_t i = 0, j = 0;
+	uint64_t gfx12_modifiers[] = {mod_256k, mod_64k, mod_4k, mod_256b, DRM_FORMAT_MOD_LINEAR};
+
+	for (i = 0; i < ARRAY_SIZE(max_comp_block); i++)
+		max_comp_block_mod[i] = AMD_FMT_MOD_SET(DCC_MAX_COMPRESSED_BLOCK, max_comp_block[i]);
+
+	/* With DCC: Best choice should be kept first. Hence, add all 256k modifiers of different
+	 * max compressed blocks first and then move on to the next smaller sized layouts.
+	 * Do not add the linear modifier here, and hence the condition of size-1 for the loop
+	 */
+	for (j = 0; j < ARRAY_SIZE(gfx12_modifiers) - 1; j++)
+		for (i = 0; i < ARRAY_SIZE(max_comp_block); i++)
+			amdgpu_dm_plane_add_modifier(mods, size, capacity,
+						     ver | dcc | max_comp_block_mod[i] | gfx12_modifiers[j]);
+
+	/* Without DCC. Add all modifiers including linear at the end */
+	for (i = 0; i < ARRAY_SIZE(gfx12_modifiers); i++)
+		amdgpu_dm_plane_add_modifier(mods, size, capacity, gfx12_modifiers[i]);
+
+}
+
+static int amdgpu_dm_plane_get_plane_modifiers(struct amdgpu_device *adev, unsigned int plane_type, uint64_t **mods)
+{
+	uint64_t size = 0, capacity = 128;
+	*mods = NULL;
+
+	/* We have not hooked up any pre-GFX9 modifiers. */
+	if (adev->family < AMDGPU_FAMILY_AI)
+		return 0;
+
+	*mods = kmalloc_array(capacity, sizeof(uint64_t), GFP_KERNEL);
+
+	if (plane_type == DRM_PLANE_TYPE_CURSOR) {
+		amdgpu_dm_plane_add_modifier(mods, &size, &capacity, DRM_FORMAT_MOD_LINEAR);
+		amdgpu_dm_plane_add_modifier(mods, &size, &capacity, DRM_FORMAT_MOD_INVALID);
+		return *mods ? 0 : -ENOMEM;
+	}
+
+	switch (adev->family) {
+	case AMDGPU_FAMILY_AI:
+	case AMDGPU_FAMILY_RV:
+		amdgpu_dm_plane_add_gfx9_modifiers(adev, mods, &size, &capacity);
+		break;
+	case AMDGPU_FAMILY_NV:
+	case AMDGPU_FAMILY_VGH:
+	case AMDGPU_FAMILY_YC:
+	case AMDGPU_FAMILY_GC_10_3_6:
+	case AMDGPU_FAMILY_GC_10_3_7:
+		if (amdgpu_ip_version(adev, GC_HWIP, 0) >= IP_VERSION(10, 3, 0))
+			amdgpu_dm_plane_add_gfx10_3_modifiers(adev, mods, &size, &capacity);
+		else
+			amdgpu_dm_plane_add_gfx10_1_modifiers(adev, mods, &size, &capacity);
+		break;
+	case AMDGPU_FAMILY_GC_11_0_0:
+	case AMDGPU_FAMILY_GC_11_0_1:
+	case AMDGPU_FAMILY_GC_11_5_0:
+		amdgpu_dm_plane_add_gfx11_modifiers(adev, mods, &size, &capacity);
+		break;
+	case AMDGPU_FAMILY_GC_12_0_0:
+		amdgpu_dm_plane_add_gfx12_modifiers(adev, mods, &size, &capacity);
+		break;
+	}
+
+	amdgpu_dm_plane_add_modifier(mods, &size, &capacity, DRM_FORMAT_MOD_LINEAR);
+
+	/* INVALID marks the end of the list. */
+	amdgpu_dm_plane_add_modifier(mods, &size, &capacity, DRM_FORMAT_MOD_INVALID);
+
+	if (!*mods)
+		return -ENOMEM;
+
+	return 0;
+}
+
+static int amdgpu_dm_plane_get_plane_formats(const struct drm_plane *plane,
+					     const struct dc_plane_cap *plane_cap,
+					     uint32_t *formats, int max_formats)
+{
+	int i, num_formats = 0;
+
+	/*
+	 * TODO: Query support for each group of formats directly from
+	 * DC plane caps. This will require adding more formats to the
+	 * caps list.
+	 */
+
+	if (plane->type == DRM_PLANE_TYPE_PRIMARY ||
+		(plane_cap && plane_cap->type == DC_PLANE_TYPE_DCN_UNIVERSAL && plane->type != DRM_PLANE_TYPE_CURSOR)) {
+		for (i = 0; i < ARRAY_SIZE(rgb_formats); ++i) {
+			if (num_formats >= max_formats)
+				break;
+
+			formats[num_formats++] = rgb_formats[i];
+		}
+
+		if (plane_cap && plane_cap->pixel_format_support.nv12)
+			formats[num_formats++] = DRM_FORMAT_NV12;
+		if (plane_cap && plane_cap->pixel_format_support.p010)
+			formats[num_formats++] = DRM_FORMAT_P010;
+		if (plane_cap && plane_cap->pixel_format_support.fp16) {
+			formats[num_formats++] = DRM_FORMAT_XRGB16161616F;
+			formats[num_formats++] = DRM_FORMAT_ARGB16161616F;
+			formats[num_formats++] = DRM_FORMAT_XBGR16161616F;
+			formats[num_formats++] = DRM_FORMAT_ABGR16161616F;
+		}
+	} else {
+		switch (plane->type) {
+		case DRM_PLANE_TYPE_OVERLAY:
+			for (i = 0; i < ARRAY_SIZE(overlay_formats); ++i) {
+				if (num_formats >= max_formats)
+					break;
+
+				formats[num_formats++] = overlay_formats[i];
+			}
+			break;
+
+		case DRM_PLANE_TYPE_CURSOR:
+			for (i = 0; i < ARRAY_SIZE(cursor_formats); ++i) {
+				if (num_formats >= max_formats)
+					break;
+
+				formats[num_formats++] = cursor_formats[i];
+			}
+			break;
+
+		default:
+			break;
+		}
+	}
+
+	return num_formats;
+}
+
+int amdgpu_dm_plane_fill_plane_buffer_attributes(struct amdgpu_device *adev,
+			     const struct amdgpu_framebuffer *afb,
+			     const enum surface_pixel_format format,
+			     const enum dc_rotation_angle rotation,
+			     const uint64_t tiling_flags,
+			     struct dc_tiling_info *tiling_info,
+			     struct plane_size *plane_size,
+			     struct dc_plane_dcc_param *dcc,
+			     struct dc_plane_address *address,
+			     bool tmz_surface)
+{
+	const struct drm_framebuffer *fb = &afb->base;
+	int ret;
+
+	memset(tiling_info, 0, sizeof(*tiling_info));
+	memset(plane_size, 0, sizeof(*plane_size));
+	memset(dcc, 0, sizeof(*dcc));
+	memset(address, 0, sizeof(*address));
+
+	address->tmz_surface = tmz_surface;
+
+	if (format < SURFACE_PIXEL_FORMAT_VIDEO_BEGIN) {
+		uint64_t addr = afb->address + fb->offsets[0];
+
+		plane_size->surface_size.x = 0;
+		plane_size->surface_size.y = 0;
+		plane_size->surface_size.width = fb->width;
+		plane_size->surface_size.height = fb->height;
+		plane_size->surface_pitch =
+			fb->pitches[0] / fb->format->cpp[0];
+
+		address->type = PLN_ADDR_TYPE_GRAPHICS;
+		address->grph.addr.low_part = lower_32_bits(addr);
+		address->grph.addr.high_part = upper_32_bits(addr);
+	} else if (format < SURFACE_PIXEL_FORMAT_INVALID) {
+		uint64_t luma_addr = afb->address + fb->offsets[0];
+		uint64_t chroma_addr = afb->address + fb->offsets[1];
+
+		plane_size->surface_size.x = 0;
+		plane_size->surface_size.y = 0;
+		plane_size->surface_size.width = fb->width;
+		plane_size->surface_size.height = fb->height;
+		plane_size->surface_pitch =
+			fb->pitches[0] / fb->format->cpp[0];
+
+		plane_size->chroma_size.x = 0;
+		plane_size->chroma_size.y = 0;
+		/* TODO: set these based on surface format */
+		plane_size->chroma_size.width = fb->width / 2;
+		plane_size->chroma_size.height = fb->height / 2;
+
+		plane_size->chroma_pitch =
+			fb->pitches[1] / fb->format->cpp[1];
+
+		address->type = PLN_ADDR_TYPE_VIDEO_PROGRESSIVE;
+		address->video_progressive.luma_addr.low_part =
+			lower_32_bits(luma_addr);
+		address->video_progressive.luma_addr.high_part =
+			upper_32_bits(luma_addr);
+		address->video_progressive.chroma_addr.low_part =
+			lower_32_bits(chroma_addr);
+		address->video_progressive.chroma_addr.high_part =
+			upper_32_bits(chroma_addr);
+	}
+
+	if (adev->family >= AMDGPU_FAMILY_GC_12_0_0) {
+		ret = amdgpu_dm_plane_fill_gfx12_plane_attributes_from_modifiers(adev, afb, format,
+										 rotation, plane_size,
+										 tiling_info, dcc,
+										 address);
+		if (ret)
+			return ret;
+	} else if (adev->family >= AMDGPU_FAMILY_AI) {
+		ret = amdgpu_dm_plane_fill_gfx9_plane_attributes_from_modifiers(adev, afb, format,
+										rotation, plane_size,
+										tiling_info, dcc,
+										address);
+		if (ret)
+			return ret;
+	} else {
+		amdgpu_dm_plane_fill_gfx8_tiling_info_from_flags(tiling_info, tiling_flags);
+	}
+
+	return 0;
+}
+
+static int amdgpu_dm_plane_helper_prepare_fb(struct drm_plane *plane,
+					     struct drm_plane_state *new_state)
+{
+	struct amdgpu_framebuffer *afb;
+	struct drm_gem_object *obj;
+	struct amdgpu_device *adev;
+	struct amdgpu_bo *rbo;
+	struct dm_plane_state *dm_plane_state_new, *dm_plane_state_old;
+	uint32_t domain;
+	int r;
+
+	if (!new_state->fb) {
+		DRM_DEBUG_KMS("No FB bound\n");
+		return 0;
+	}
+
+	afb = to_amdgpu_framebuffer(new_state->fb);
+	obj = drm_gem_fb_get_obj(new_state->fb, 0);
+	if (!obj) {
+		DRM_ERROR("Failed to get obj from framebuffer\n");
+		return -EINVAL;
+	}
+
+	rbo = gem_to_amdgpu_bo(obj);
+	adev = amdgpu_ttm_adev(rbo->tbo.bdev);
+	r = amdgpu_bo_reserve(rbo, true);
+	if (r) {
+		drm_err(adev_to_drm(adev), "fail to reserve bo (%d)\n", r);
+		return r;
+	}
+
+	r = dma_resv_reserve_fences(rbo->tbo.base.resv, 1);
+	if (r) {
+		drm_err(adev_to_drm(adev), "reserving fence slot failed (%d)\n", r);
+		goto error_unlock;
+	}
+
+	if (plane->type != DRM_PLANE_TYPE_CURSOR)
+		domain = amdgpu_display_supported_domains(adev, rbo->flags);
+	else
+		domain = AMDGPU_GEM_DOMAIN_VRAM;
+
+	rbo->flags |= AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS;
+	r = amdgpu_bo_pin(rbo, domain);
+	if (unlikely(r != 0)) {
+		if (r != -ERESTARTSYS)
+			DRM_ERROR("Failed to pin framebuffer with error %d\n", r);
+		goto error_unlock;
+	}
+
+	r = amdgpu_ttm_alloc_gart(&rbo->tbo);
+	if (unlikely(r != 0)) {
+		DRM_ERROR("%p bind failed\n", rbo);
+		goto error_unpin;
+	}
+
+	r = drm_gem_plane_helper_prepare_fb(plane, new_state);
+	if (unlikely(r != 0))
+		goto error_unpin;
+
+	amdgpu_bo_unreserve(rbo);
+
+	afb->address = amdgpu_bo_gpu_offset(rbo);
+
+	amdgpu_bo_ref(rbo);
+
+	/**
+	 * We don't do surface updates on planes that have been newly created,
+	 * but we also don't have the afb->address during atomic check.
+	 *
+	 * Fill in buffer attributes depending on the address here, but only on
+	 * newly created planes since they're not being used by DC yet and this
+	 * won't modify global state.
+	 */
+	dm_plane_state_old = to_dm_plane_state(plane->state);
+	dm_plane_state_new = to_dm_plane_state(new_state);
+
+	if (dm_plane_state_new->dc_state &&
+	    dm_plane_state_old->dc_state != dm_plane_state_new->dc_state) {
+		struct dc_plane_state *plane_state =
+			dm_plane_state_new->dc_state;
+
+		amdgpu_dm_plane_fill_plane_buffer_attributes(
+			adev, afb, plane_state->format, plane_state->rotation,
+			afb->tiling_flags,
+			&plane_state->tiling_info, &plane_state->plane_size,
+			&plane_state->dcc, &plane_state->address,
+			afb->tmz_surface);
+	}
+
+	return 0;
+
+error_unpin:
+	amdgpu_bo_unpin(rbo);
+
+error_unlock:
+	amdgpu_bo_unreserve(rbo);
+	return r;
+}
+
+static void amdgpu_dm_plane_helper_cleanup_fb(struct drm_plane *plane,
+					      struct drm_plane_state *old_state)
+{
+	struct amdgpu_bo *rbo;
+	int r;
+
+	if (!old_state->fb)
+		return;
+
+	rbo = gem_to_amdgpu_bo(old_state->fb->obj[0]);
+	r = amdgpu_bo_reserve(rbo, false);
+	if (unlikely(r)) {
+		DRM_ERROR("failed to reserve rbo before unpin\n");
+		return;
+	}
+
+	amdgpu_bo_unpin(rbo);
+	amdgpu_bo_unreserve(rbo);
+	amdgpu_bo_unref(&rbo);
+}
+
+static void amdgpu_dm_plane_get_min_max_dc_plane_scaling(struct drm_device *dev,
+					 struct drm_framebuffer *fb,
+					 int *min_downscale, int *max_upscale)
+{
+	struct amdgpu_device *adev = drm_to_adev(dev);
+	struct dc *dc = adev->dm.dc;
+	/* Caps for all supported planes are the same on DCE and DCN 1 - 3 */
+	struct dc_plane_cap *plane_cap = &dc->caps.planes[0];
+
+	switch (fb->format->format) {
+	case DRM_FORMAT_P010:
+	case DRM_FORMAT_NV12:
+	case DRM_FORMAT_NV21:
+		*max_upscale = plane_cap->max_upscale_factor.nv12;
+		*min_downscale = plane_cap->max_downscale_factor.nv12;
+		break;
+
+	case DRM_FORMAT_XRGB16161616F:
+	case DRM_FORMAT_ARGB16161616F:
+	case DRM_FORMAT_XBGR16161616F:
+	case DRM_FORMAT_ABGR16161616F:
+		*max_upscale = plane_cap->max_upscale_factor.fp16;
+		*min_downscale = plane_cap->max_downscale_factor.fp16;
+		break;
+
+	default:
+		*max_upscale = plane_cap->max_upscale_factor.argb8888;
+		*min_downscale = plane_cap->max_downscale_factor.argb8888;
+		break;
+	}
+
+	/*
+	 * A factor of 1 in the plane_cap means to not allow scaling, ie. use a
+	 * scaling factor of 1.0 == 1000 units.
+	 */
+	if (*max_upscale == 1)
+		*max_upscale = 1000;
+
+	if (*min_downscale == 1)
+		*min_downscale = 1000;
+}
+
+int amdgpu_dm_plane_helper_check_state(struct drm_plane_state *state,
+				       struct drm_crtc_state *new_crtc_state)
+{
+	struct drm_framebuffer *fb = state->fb;
+	int min_downscale, max_upscale;
+	int min_scale = 0;
+	int max_scale = INT_MAX;
+
+	/* Plane enabled? Validate viewport and get scaling factors from plane caps. */
+	if (fb && state->crtc) {
+		/* Validate viewport to cover the case when only the position changes */
+		if (state->plane->type != DRM_PLANE_TYPE_CURSOR) {
+			int viewport_width = state->crtc_w;
+			int viewport_height = state->crtc_h;
+
+			if (state->crtc_x < 0)
+				viewport_width += state->crtc_x;
+			else if (state->crtc_x + state->crtc_w > new_crtc_state->mode.crtc_hdisplay)
+				viewport_width = new_crtc_state->mode.crtc_hdisplay - state->crtc_x;
+
+			if (state->crtc_y < 0)
+				viewport_height += state->crtc_y;
+			else if (state->crtc_y + state->crtc_h > new_crtc_state->mode.crtc_vdisplay)
+				viewport_height = new_crtc_state->mode.crtc_vdisplay - state->crtc_y;
+
+			if (viewport_width < 0 || viewport_height < 0) {
+				DRM_DEBUG_ATOMIC("Plane completely outside of screen\n");
+				return -EINVAL;
+			} else if (viewport_width < MIN_VIEWPORT_SIZE*2) { /* x2 for width is because of pipe-split. */
+				DRM_DEBUG_ATOMIC("Viewport width %d smaller than %d\n", viewport_width, MIN_VIEWPORT_SIZE*2);
+				return -EINVAL;
+			} else if (viewport_height < MIN_VIEWPORT_SIZE) {
+				DRM_DEBUG_ATOMIC("Viewport height %d smaller than %d\n", viewport_height, MIN_VIEWPORT_SIZE);
+				return -EINVAL;
+			}
+
+		}
+
+		/* Get min/max allowed scaling factors from plane caps. */
+		amdgpu_dm_plane_get_min_max_dc_plane_scaling(state->crtc->dev, fb,
+							     &min_downscale, &max_upscale);
+		/*
+		 * Convert to drm convention: 16.16 fixed point, instead of dc's
+		 * 1.0 == 1000. Also drm scaling is src/dst instead of dc's
+		 * dst/src, so min_scale = 1.0 / max_upscale, etc.
+		 */
+		min_scale = (1000 << 16) / max_upscale;
+		max_scale = (1000 << 16) / min_downscale;
+	}
+
+	return drm_atomic_helper_check_plane_state(
+		state, new_crtc_state, min_scale, max_scale, true, true);
+}
+
+int amdgpu_dm_plane_fill_dc_scaling_info(struct amdgpu_device *adev,
+				const struct drm_plane_state *state,
+				struct dc_scaling_info *scaling_info)
+{
+	int scale_w, scale_h, min_downscale, max_upscale;
+
+	memset(scaling_info, 0, sizeof(*scaling_info));
+
+	/* Source is fixed 16.16 but we ignore mantissa for now... */
+	scaling_info->src_rect.x = state->src_x >> 16;
+	scaling_info->src_rect.y = state->src_y >> 16;
+
+	/*
+	 * For reasons we don't (yet) fully understand a non-zero
+	 * src_y coordinate into an NV12 buffer can cause a
+	 * system hang on DCN1x.
+	 * To avoid hangs (and maybe be overly cautious)
+	 * let's reject both non-zero src_x and src_y.
+	 *
+	 * We currently know of only one use-case to reproduce a
+	 * scenario with non-zero src_x and src_y for NV12, which
+	 * is to gesture the YouTube Android app into full screen
+	 * on ChromeOS.
+	 */
+	if (((amdgpu_ip_version(adev, DCE_HWIP, 0) == IP_VERSION(1, 0, 0)) ||
+	    (amdgpu_ip_version(adev, DCE_HWIP, 0) == IP_VERSION(1, 0, 1))) &&
+	    (state->fb && state->fb->format->format == DRM_FORMAT_NV12 &&
+	    (scaling_info->src_rect.x != 0 || scaling_info->src_rect.y != 0)))
+		return -EINVAL;
+
+	scaling_info->src_rect.width = state->src_w >> 16;
+	if (scaling_info->src_rect.width == 0)
+		return -EINVAL;
+
+	scaling_info->src_rect.height = state->src_h >> 16;
+	if (scaling_info->src_rect.height == 0)
+		return -EINVAL;
+
+	scaling_info->dst_rect.x = state->crtc_x;
+	scaling_info->dst_rect.y = state->crtc_y;
+
+	if (state->crtc_w == 0)
+		return -EINVAL;
+
+	scaling_info->dst_rect.width = state->crtc_w;
+
+	if (state->crtc_h == 0)
+		return -EINVAL;
+
+	scaling_info->dst_rect.height = state->crtc_h;
+
+	/* DRM doesn't specify clipping on destination output. */
+	scaling_info->clip_rect = scaling_info->dst_rect;
+
+	/* Validate scaling per-format with DC plane caps */
+	if (state->plane && state->plane->dev && state->fb) {
+		amdgpu_dm_plane_get_min_max_dc_plane_scaling(state->plane->dev, state->fb,
+							     &min_downscale, &max_upscale);
+	} else {
+		min_downscale = 250;
+		max_upscale = 16000;
+	}
+
+	scale_w = scaling_info->dst_rect.width * 1000 /
+		  scaling_info->src_rect.width;
+
+	if (scale_w < min_downscale || scale_w > max_upscale)
+		return -EINVAL;
+
+	scale_h = scaling_info->dst_rect.height * 1000 /
+		  scaling_info->src_rect.height;
+
+	if (scale_h < min_downscale || scale_h > max_upscale)
+		return -EINVAL;
+
+	/*
+	 * The "scaling_quality" can be ignored for now, quality = 0 has DC
+	 * assume reasonable defaults based on the format.
+	 */
+
+	return 0;
+}
+
+static int amdgpu_dm_plane_atomic_check(struct drm_plane *plane,
+					struct drm_atomic_state *state)
+{
+	struct drm_plane_state *new_plane_state = drm_atomic_get_new_plane_state(state,
+										 plane);
+	struct amdgpu_device *adev = drm_to_adev(plane->dev);
+	struct dc *dc = adev->dm.dc;
+	struct dm_plane_state *dm_plane_state;
+	struct dc_scaling_info scaling_info;
+	struct drm_crtc_state *new_crtc_state;
+	int ret;
+
+	trace_amdgpu_dm_plane_atomic_check(new_plane_state);
+
+	dm_plane_state = to_dm_plane_state(new_plane_state);
+
+	if (!dm_plane_state->dc_state)
+		return 0;
+
+	new_crtc_state =
+		drm_atomic_get_new_crtc_state(state,
+					      new_plane_state->crtc);
+	if (!new_crtc_state)
+		return -EINVAL;
+
+	ret = amdgpu_dm_plane_helper_check_state(new_plane_state, new_crtc_state);
+	if (ret)
+		return ret;
+
+	ret = amdgpu_dm_plane_fill_dc_scaling_info(adev, new_plane_state, &scaling_info);
+	if (ret)
+		return ret;
+
+	if (dc_validate_plane(dc, dm_plane_state->dc_state) == DC_OK)
+		return 0;
+
+	return -EINVAL;
+}
+
+static int amdgpu_dm_plane_atomic_async_check(struct drm_plane *plane,
+					      struct drm_atomic_state *state, bool flip)
+{
+	struct drm_crtc_state *new_crtc_state;
+	struct drm_plane_state *new_plane_state;
+	struct dm_crtc_state *dm_new_crtc_state;
+
+	if (flip) {
+		if (plane->type != DRM_PLANE_TYPE_OVERLAY)
+			return -EINVAL;
+	} else if (plane->type != DRM_PLANE_TYPE_CURSOR) {
+		return -EINVAL;
+	}
+
+	new_plane_state = drm_atomic_get_new_plane_state(state, plane);
+	new_crtc_state = drm_atomic_get_new_crtc_state(state, new_plane_state->crtc);
+	dm_new_crtc_state = to_dm_crtc_state(new_crtc_state);
+	/* Reject overlay cursors for now*/
+	if (!flip && dm_new_crtc_state->cursor_mode == DM_CURSOR_OVERLAY_MODE)
+		return -EINVAL;
+
+	return 0;
+}
+
+int amdgpu_dm_plane_get_cursor_position(struct drm_plane *plane, struct drm_crtc *crtc,
+					struct dc_cursor_position *position)
+{
+	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
+	struct amdgpu_device *adev = drm_to_adev(plane->dev);
+	int x, y;
+	int xorigin = 0, yorigin = 0;
+
+	if (!crtc || !plane->state->fb)
+		return 0;
+
+	if ((plane->state->crtc_w > amdgpu_crtc->max_cursor_width) ||
+	    (plane->state->crtc_h > amdgpu_crtc->max_cursor_height)) {
+		DRM_ERROR("%s: bad cursor width or height %d x %d\n",
+			  __func__,
+			  plane->state->crtc_w,
+			  plane->state->crtc_h);
+		return -EINVAL;
+	}
+
+	x = plane->state->crtc_x;
+	y = plane->state->crtc_y;
+
+	if (x <= -amdgpu_crtc->max_cursor_width ||
+	    y <= -amdgpu_crtc->max_cursor_height)
+		return 0;
+
+	if (x < 0) {
+		xorigin = min(-x, amdgpu_crtc->max_cursor_width - 1);
+		x = 0;
+	}
+	if (y < 0) {
+		yorigin = min(-y, amdgpu_crtc->max_cursor_height - 1);
+		y = 0;
+	}
+	position->enable = true;
+	position->x = x;
+	position->y = y;
+	position->x_hotspot = xorigin;
+	position->y_hotspot = yorigin;
+
+	if (amdgpu_ip_version(adev, DCE_HWIP, 0) < IP_VERSION(4, 0, 1))
+		position->translate_by_source = true;
+
+	return 0;
+}
+
+void amdgpu_dm_plane_handle_cursor_update(struct drm_plane *plane,
+				 struct drm_plane_state *old_plane_state)
+{
+	struct amdgpu_device *adev = drm_to_adev(plane->dev);
+	struct amdgpu_framebuffer *afb = to_amdgpu_framebuffer(plane->state->fb);
+	struct drm_crtc *crtc = afb ? plane->state->crtc : old_plane_state->crtc;
+	struct dm_crtc_state *crtc_state = crtc ? to_dm_crtc_state(crtc->state) : NULL;
+	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
+	uint64_t address = afb ? afb->address : 0;
+	struct dc_cursor_position position = {0};
+	struct dc_cursor_attributes attributes;
+	int ret;
+
+	if (!plane->state->fb && !old_plane_state->fb)
+		return;
+
+	drm_dbg_atomic(plane->dev, "crtc_id=%d with size %d to %d\n",
+		       amdgpu_crtc->crtc_id, plane->state->crtc_w,
+		       plane->state->crtc_h);
+
+	ret = amdgpu_dm_plane_get_cursor_position(plane, crtc, &position);
+	if (ret)
+		return;
+
+	if (!position.enable) {
+		/* turn off cursor */
+		if (crtc_state && crtc_state->stream) {
+			mutex_lock(&adev->dm.dc_lock);
+			dc_stream_program_cursor_position(crtc_state->stream,
+						      &position);
+			mutex_unlock(&adev->dm.dc_lock);
+		}
+		return;
+	}
+
+	amdgpu_crtc->cursor_width = plane->state->crtc_w;
+	amdgpu_crtc->cursor_height = plane->state->crtc_h;
+
+	memset(&attributes, 0, sizeof(attributes));
+	attributes.address.high_part = upper_32_bits(address);
+	attributes.address.low_part  = lower_32_bits(address);
+	attributes.width             = plane->state->crtc_w;
+	attributes.height            = plane->state->crtc_h;
+	attributes.color_format      = CURSOR_MODE_COLOR_PRE_MULTIPLIED_ALPHA;
+	attributes.rotation_angle    = 0;
+	attributes.attribute_flags.value = 0;
+
+	/* Enable cursor degamma ROM on DCN3+ for implicit sRGB degamma in DRM
+	 * legacy gamma setup.
+	 */
+	if (crtc_state->cm_is_degamma_srgb &&
+	    adev->dm.dc->caps.color.dpp.gamma_corr)
+		attributes.attribute_flags.bits.ENABLE_CURSOR_DEGAMMA = 1;
+
+	if (afb)
+		attributes.pitch = afb->base.pitches[0] / afb->base.format->cpp[0];
+
+	if (crtc_state->stream) {
+		mutex_lock(&adev->dm.dc_lock);
+		if (!dc_stream_program_cursor_attributes(crtc_state->stream,
+							 &attributes))
+			DRM_ERROR("DC failed to set cursor attributes\n");
+
+		if (!dc_stream_program_cursor_position(crtc_state->stream,
+						   &position))
+			DRM_ERROR("DC failed to set cursor position\n");
+		mutex_unlock(&adev->dm.dc_lock);
+	}
+}
+
+static void amdgpu_dm_plane_atomic_async_update(struct drm_plane *plane,
+						struct drm_atomic_state *state)
+{
+	struct drm_plane_state *new_state = drm_atomic_get_new_plane_state(state,
+									   plane);
+	struct drm_plane_state *old_state =
+		drm_atomic_get_old_plane_state(state, plane);
+
+	trace_amdgpu_dm_atomic_update_cursor(new_state);
+
+	swap(plane->state->fb, new_state->fb);
+
+	plane->state->src_x = new_state->src_x;
+	plane->state->src_y = new_state->src_y;
+	plane->state->src_w = new_state->src_w;
+	plane->state->src_h = new_state->src_h;
+	plane->state->crtc_x = new_state->crtc_x;
+	plane->state->crtc_y = new_state->crtc_y;
+	plane->state->crtc_w = new_state->crtc_w;
+	plane->state->crtc_h = new_state->crtc_h;
+
+	amdgpu_dm_plane_handle_cursor_update(plane, old_state);
+}
+
+static void amdgpu_dm_plane_panic_flush(struct drm_plane *plane)
+{
+	struct dm_plane_state *dm_plane_state = to_dm_plane_state(plane->state);
+	struct drm_framebuffer *fb = plane->state->fb;
+	struct dc_plane_state *dc_plane_state;
+
+	if (!dm_plane_state || !dm_plane_state->dc_state)
+		return;
+
+	dc_plane_state = dm_plane_state->dc_state;
+
+	dc_plane_force_dcc_and_tiling_disable(dc_plane_state, fb->modifier ? true : false);
+}
+
+static const struct drm_plane_helper_funcs dm_plane_helper_funcs = {
+	.prepare_fb = amdgpu_dm_plane_helper_prepare_fb,
+	.cleanup_fb = amdgpu_dm_plane_helper_cleanup_fb,
+	.atomic_check = amdgpu_dm_plane_atomic_check,
+	.atomic_async_check = amdgpu_dm_plane_atomic_async_check,
+	.atomic_async_update = amdgpu_dm_plane_atomic_async_update
+};
+
+static const struct drm_plane_helper_funcs dm_primary_plane_helper_funcs = {
+	.prepare_fb = amdgpu_dm_plane_helper_prepare_fb,
+	.cleanup_fb = amdgpu_dm_plane_helper_cleanup_fb,
+	.atomic_check = amdgpu_dm_plane_atomic_check,
+	.atomic_async_check = amdgpu_dm_plane_atomic_async_check,
+	.atomic_async_update = amdgpu_dm_plane_atomic_async_update,
+	.get_scanout_buffer = amdgpu_display_get_scanout_buffer,
+	.panic_flush = amdgpu_dm_plane_panic_flush,
+};
+
+static void amdgpu_dm_plane_drm_plane_reset(struct drm_plane *plane)
+{
+	struct dm_plane_state *amdgpu_state = NULL;
+
+	if (plane->state)
+		plane->funcs->atomic_destroy_state(plane, plane->state);
+
+	amdgpu_state = kzalloc(sizeof(*amdgpu_state), GFP_KERNEL);
+	WARN_ON(amdgpu_state == NULL);
+
+	if (!amdgpu_state)
+		return;
+
+	__drm_atomic_helper_plane_reset(plane, &amdgpu_state->base);
+	amdgpu_state->degamma_tf = AMDGPU_TRANSFER_FUNCTION_DEFAULT;
+	amdgpu_state->hdr_mult = AMDGPU_HDR_MULT_DEFAULT;
+	amdgpu_state->shaper_tf = AMDGPU_TRANSFER_FUNCTION_DEFAULT;
+	amdgpu_state->blend_tf = AMDGPU_TRANSFER_FUNCTION_DEFAULT;
+}
+
+static struct drm_plane_state *amdgpu_dm_plane_drm_plane_duplicate_state(struct drm_plane *plane)
+{
+	struct dm_plane_state *dm_plane_state, *old_dm_plane_state;
+
+	old_dm_plane_state = to_dm_plane_state(plane->state);
+	dm_plane_state = kzalloc(sizeof(*dm_plane_state), GFP_KERNEL);
+	if (!dm_plane_state)
+		return NULL;
+
+	__drm_atomic_helper_plane_duplicate_state(plane, &dm_plane_state->base);
+
+	if (old_dm_plane_state->dc_state) {
+		dm_plane_state->dc_state = old_dm_plane_state->dc_state;
+		dc_plane_state_retain(dm_plane_state->dc_state);
+	}
+
+	if (old_dm_plane_state->degamma_lut)
+		dm_plane_state->degamma_lut =
+			drm_property_blob_get(old_dm_plane_state->degamma_lut);
+	if (old_dm_plane_state->ctm)
+		dm_plane_state->ctm =
+			drm_property_blob_get(old_dm_plane_state->ctm);
+	if (old_dm_plane_state->shaper_lut)
+		dm_plane_state->shaper_lut =
+			drm_property_blob_get(old_dm_plane_state->shaper_lut);
+	if (old_dm_plane_state->lut3d)
+		dm_plane_state->lut3d =
+			drm_property_blob_get(old_dm_plane_state->lut3d);
+	if (old_dm_plane_state->blend_lut)
+		dm_plane_state->blend_lut =
+			drm_property_blob_get(old_dm_plane_state->blend_lut);
+
+	dm_plane_state->degamma_tf = old_dm_plane_state->degamma_tf;
+	dm_plane_state->hdr_mult = old_dm_plane_state->hdr_mult;
+	dm_plane_state->shaper_tf = old_dm_plane_state->shaper_tf;
+	dm_plane_state->blend_tf = old_dm_plane_state->blend_tf;
+
+	return &dm_plane_state->base;
+}
+
+static bool amdgpu_dm_plane_format_mod_supported(struct drm_plane *plane,
+						 uint32_t format,
+						 uint64_t modifier)
+{
+	struct amdgpu_device *adev = drm_to_adev(plane->dev);
+	const struct drm_format_info *info = drm_format_info(format);
+	int i;
+
+	if (!info)
+		return false;
+
+	/*
+	 * We always have to allow these modifiers:
+	 * 1. Core DRM checks for LINEAR support if userspace does not provide modifiers.
+	 * 2. Not passing any modifiers is the same as explicitly passing INVALID.
+	 */
+	if (modifier == DRM_FORMAT_MOD_LINEAR ||
+	    modifier == DRM_FORMAT_MOD_INVALID) {
+		return true;
+	}
+
+	/* Check that the modifier is on the list of the plane's supported modifiers. */
+	for (i = 0; i < plane->modifier_count; i++) {
+		if (modifier == plane->modifiers[i])
+			break;
+	}
+	if (i == plane->modifier_count)
+		return false;
+
+	/* GFX12 doesn't have these limitations. */
+	if (AMD_FMT_MOD_GET(TILE_VERSION, modifier) <= AMD_FMT_MOD_TILE_VER_GFX11) {
+		enum dm_micro_swizzle microtile = amdgpu_dm_plane_modifier_gfx9_swizzle_mode(modifier) & 3;
+
+		/*
+		 * For D swizzle the canonical modifier depends on the bpp, so check
+		 * it here.
+		 */
+		if (AMD_FMT_MOD_GET(TILE_VERSION, modifier) == AMD_FMT_MOD_TILE_VER_GFX9 &&
+		    adev->family >= AMDGPU_FAMILY_NV) {
+			if (microtile == MICRO_SWIZZLE_D && info->cpp[0] == 4)
+				return false;
+		}
+
+		if (adev->family >= AMDGPU_FAMILY_RV && microtile == MICRO_SWIZZLE_D &&
+		    info->cpp[0] < 8)
+			return false;
+
+		if (amdgpu_dm_plane_modifier_has_dcc(modifier)) {
+			/* Per radeonsi comments 16/64 bpp are more complicated. */
+			if (info->cpp[0] != 4)
+				return false;
+			/* We support multi-planar formats, but not when combined with
+			 * additional DCC metadata planes.
+			 */
+			if (info->num_planes > 1)
+				return false;
+		}
+	}
+
+	return true;
+}
+
+static void amdgpu_dm_plane_drm_plane_destroy_state(struct drm_plane *plane,
+						    struct drm_plane_state *state)
+{
+	struct dm_plane_state *dm_plane_state = to_dm_plane_state(state);
+
+	if (dm_plane_state->degamma_lut)
+		drm_property_blob_put(dm_plane_state->degamma_lut);
+	if (dm_plane_state->ctm)
+		drm_property_blob_put(dm_plane_state->ctm);
+	if (dm_plane_state->lut3d)
+		drm_property_blob_put(dm_plane_state->lut3d);
+	if (dm_plane_state->shaper_lut)
+		drm_property_blob_put(dm_plane_state->shaper_lut);
+	if (dm_plane_state->blend_lut)
+		drm_property_blob_put(dm_plane_state->blend_lut);
+
+	if (dm_plane_state->dc_state)
+		dc_plane_state_release(dm_plane_state->dc_state);
+
+	drm_atomic_helper_plane_destroy_state(plane, state);
+}
+
+#ifdef AMD_PRIVATE_COLOR
+static void
+dm_atomic_plane_attach_color_mgmt_properties(struct amdgpu_display_manager *dm,
+					     struct drm_plane *plane)
+{
+	struct amdgpu_mode_info mode_info = dm->adev->mode_info;
+	struct dpp_color_caps dpp_color_caps = dm->dc->caps.color.dpp;
+
+	/* Check HW color pipeline capabilities on DPP block (pre-blending)
+	 * before exposing related properties.
+	 */
+	if (dpp_color_caps.dgam_ram || dpp_color_caps.gamma_corr) {
+		drm_object_attach_property(&plane->base,
+					   mode_info.plane_degamma_lut_property,
+					   0);
+		drm_object_attach_property(&plane->base,
+					   mode_info.plane_degamma_lut_size_property,
+					   MAX_COLOR_LUT_ENTRIES);
+		drm_object_attach_property(&plane->base,
+					   dm->adev->mode_info.plane_degamma_tf_property,
+					   AMDGPU_TRANSFER_FUNCTION_DEFAULT);
+	}
+	/* HDR MULT is always available */
+	drm_object_attach_property(&plane->base,
+				   dm->adev->mode_info.plane_hdr_mult_property,
+				   AMDGPU_HDR_MULT_DEFAULT);
+
+	/* Only enable plane CTM if both DPP and MPC gamut remap is available. */
+	if (dm->dc->caps.color.mpc.gamut_remap)
+		drm_object_attach_property(&plane->base,
+					   dm->adev->mode_info.plane_ctm_property, 0);
+
+	if (dpp_color_caps.hw_3d_lut || dm->dc->caps.color.mpc.preblend) {
+		drm_object_attach_property(&plane->base,
+					   mode_info.plane_shaper_lut_property, 0);
+		drm_object_attach_property(&plane->base,
+					   mode_info.plane_shaper_lut_size_property,
+					   MAX_COLOR_LUT_ENTRIES);
+		drm_object_attach_property(&plane->base,
+					   mode_info.plane_shaper_tf_property,
+					   AMDGPU_TRANSFER_FUNCTION_DEFAULT);
+		drm_object_attach_property(&plane->base,
+					   mode_info.plane_lut3d_property, 0);
+		drm_object_attach_property(&plane->base,
+					   mode_info.plane_lut3d_size_property,
+					   MAX_COLOR_3DLUT_SIZE);
+	}
+
+	if (dpp_color_caps.ogam_ram) {
+		drm_object_attach_property(&plane->base,
+					   mode_info.plane_blend_lut_property, 0);
+		drm_object_attach_property(&plane->base,
+					   mode_info.plane_blend_lut_size_property,
+					   MAX_COLOR_LUT_ENTRIES);
+		drm_object_attach_property(&plane->base,
+					   mode_info.plane_blend_tf_property,
+					   AMDGPU_TRANSFER_FUNCTION_DEFAULT);
+	}
+}
+
+static int
+dm_atomic_plane_set_property(struct drm_plane *plane,
+			     struct drm_plane_state *state,
+			     struct drm_property *property,
+			     uint64_t val)
+{
+	struct dm_plane_state *dm_plane_state = to_dm_plane_state(state);
+	struct amdgpu_device *adev = drm_to_adev(plane->dev);
+	bool replaced = false;
+	int ret;
+
+	if (property == adev->mode_info.plane_degamma_lut_property) {
+		ret = drm_property_replace_blob_from_id(plane->dev,
+							&dm_plane_state->degamma_lut,
+							val, -1,
+							sizeof(struct drm_color_lut),
+							&replaced);
+		dm_plane_state->base.color_mgmt_changed |= replaced;
+		return ret;
+	} else if (property == adev->mode_info.plane_degamma_tf_property) {
+		if (dm_plane_state->degamma_tf != val) {
+			dm_plane_state->degamma_tf = val;
+			dm_plane_state->base.color_mgmt_changed = 1;
+		}
+	} else if (property == adev->mode_info.plane_hdr_mult_property) {
+		if (dm_plane_state->hdr_mult != val) {
+			dm_plane_state->hdr_mult = val;
+			dm_plane_state->base.color_mgmt_changed = 1;
+		}
+	} else if (property == adev->mode_info.plane_ctm_property) {
+		ret = drm_property_replace_blob_from_id(plane->dev,
+							&dm_plane_state->ctm,
+							val,
+							sizeof(struct drm_color_ctm_3x4), -1,
+							&replaced);
+		dm_plane_state->base.color_mgmt_changed |= replaced;
+		return ret;
+	} else if (property == adev->mode_info.plane_shaper_lut_property) {
+		ret = drm_property_replace_blob_from_id(plane->dev,
+							&dm_plane_state->shaper_lut,
+							val, -1,
+							sizeof(struct drm_color_lut),
+							&replaced);
+		dm_plane_state->base.color_mgmt_changed |= replaced;
+		return ret;
+	} else if (property == adev->mode_info.plane_shaper_tf_property) {
+		if (dm_plane_state->shaper_tf != val) {
+			dm_plane_state->shaper_tf = val;
+			dm_plane_state->base.color_mgmt_changed = 1;
+		}
+	} else if (property == adev->mode_info.plane_lut3d_property) {
+		ret = drm_property_replace_blob_from_id(plane->dev,
+							&dm_plane_state->lut3d,
+							val, -1,
+							sizeof(struct drm_color_lut),
+							&replaced);
+		dm_plane_state->base.color_mgmt_changed |= replaced;
+		return ret;
+	} else if (property == adev->mode_info.plane_blend_lut_property) {
+		ret = drm_property_replace_blob_from_id(plane->dev,
+							&dm_plane_state->blend_lut,
+							val, -1,
+							sizeof(struct drm_color_lut),
+							&replaced);
+		dm_plane_state->base.color_mgmt_changed |= replaced;
+		return ret;
+	} else if (property == adev->mode_info.plane_blend_tf_property) {
+		if (dm_plane_state->blend_tf != val) {
+			dm_plane_state->blend_tf = val;
+			dm_plane_state->base.color_mgmt_changed = 1;
+		}
+	} else {
+		drm_dbg_atomic(plane->dev,
+			       "[PLANE:%d:%s] unknown property [PROP:%d:%s]]\n",
+			       plane->base.id, plane->name,
+			       property->base.id, property->name);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int
+dm_atomic_plane_get_property(struct drm_plane *plane,
+			     const struct drm_plane_state *state,
+			     struct drm_property *property,
+			     uint64_t *val)
+{
+	struct dm_plane_state *dm_plane_state = to_dm_plane_state(state);
+	struct amdgpu_device *adev = drm_to_adev(plane->dev);
+
+	if (property == adev->mode_info.plane_degamma_lut_property) {
+		*val = (dm_plane_state->degamma_lut) ?
+			dm_plane_state->degamma_lut->base.id : 0;
+	} else if (property == adev->mode_info.plane_degamma_tf_property) {
+		*val = dm_plane_state->degamma_tf;
+	} else if (property == adev->mode_info.plane_hdr_mult_property) {
+		*val = dm_plane_state->hdr_mult;
+	} else if (property == adev->mode_info.plane_ctm_property) {
+		*val = (dm_plane_state->ctm) ?
+			dm_plane_state->ctm->base.id : 0;
+	} else 	if (property == adev->mode_info.plane_shaper_lut_property) {
+		*val = (dm_plane_state->shaper_lut) ?
+			dm_plane_state->shaper_lut->base.id : 0;
+	} else if (property == adev->mode_info.plane_shaper_tf_property) {
+		*val = dm_plane_state->shaper_tf;
+	} else 	if (property == adev->mode_info.plane_lut3d_property) {
+		*val = (dm_plane_state->lut3d) ?
+			dm_plane_state->lut3d->base.id : 0;
+	} else 	if (property == adev->mode_info.plane_blend_lut_property) {
+		*val = (dm_plane_state->blend_lut) ?
+			dm_plane_state->blend_lut->base.id : 0;
+	} else if (property == adev->mode_info.plane_blend_tf_property) {
+		*val = dm_plane_state->blend_tf;
+
+	} else {
+		return -EINVAL;
+	}
+
+	return 0;
+}
+#else
+
+#define MAX_COLOR_PIPELINES 5
+
+static int
+dm_plane_init_colorops(struct drm_plane *plane)
+{
+	struct drm_prop_enum_list pipelines[MAX_COLOR_PIPELINES];
+	struct drm_device *dev = plane->dev;
+	struct amdgpu_device *adev = drm_to_adev(dev);
+	struct dc *dc = adev->dm.dc;
+	int len = 0;
+	int ret;
+
+	if (plane->type == DRM_PLANE_TYPE_CURSOR)
+		return 0;
+
+	/* initialize pipeline */
+	if (dc->ctx->dce_version >= DCN_VERSION_3_0) {
+		ret = amdgpu_dm_initialize_default_pipeline(plane, &pipelines[len]);
+		if (ret) {
+			drm_err(plane->dev, "Failed to create color pipeline for plane %d: %d\n",
+				plane->base.id, ret);
+			return ret;
+		}
+		len++;
+
+		/* Create COLOR_PIPELINE property and attach */
+		drm_plane_create_color_pipeline_property(plane, pipelines, len);
+	}
+
+	return 0;
+}
+#endif
+
+static const struct drm_plane_funcs dm_plane_funcs = {
+	.update_plane	= drm_atomic_helper_update_plane,
+	.disable_plane	= drm_atomic_helper_disable_plane,
+	.destroy	= drm_plane_helper_destroy,
+	.reset = amdgpu_dm_plane_drm_plane_reset,
+	.atomic_duplicate_state = amdgpu_dm_plane_drm_plane_duplicate_state,
+	.atomic_destroy_state = amdgpu_dm_plane_drm_plane_destroy_state,
+	.format_mod_supported = amdgpu_dm_plane_format_mod_supported,
+#ifdef AMD_PRIVATE_COLOR
+	.atomic_set_property = dm_atomic_plane_set_property,
+	.atomic_get_property = dm_atomic_plane_get_property,
+#endif
+};
+
+int amdgpu_dm_plane_init(struct amdgpu_display_manager *dm,
+				struct drm_plane *plane,
+				unsigned long possible_crtcs,
+				const struct dc_plane_cap *plane_cap)
+{
+	uint32_t formats[32];
+	int num_formats;
+	int res = -EPERM;
+	unsigned int supported_rotations;
+	uint64_t *modifiers = NULL;
+	unsigned int primary_zpos = dm->dc->caps.max_slave_planes;
+
+	num_formats = amdgpu_dm_plane_get_plane_formats(plane, plane_cap, formats,
+							ARRAY_SIZE(formats));
+
+	res = amdgpu_dm_plane_get_plane_modifiers(dm->adev, plane->type, &modifiers);
+	if (res)
+		return res;
+
+	if (modifiers == NULL)
+		adev_to_drm(dm->adev)->mode_config.fb_modifiers_not_supported = true;
+
+	res = drm_universal_plane_init(adev_to_drm(dm->adev), plane, possible_crtcs,
+				       &dm_plane_funcs, formats, num_formats,
+				       modifiers, plane->type, NULL);
+	kfree(modifiers);
+	if (res)
+		return res;
+
+	if (plane->type == DRM_PLANE_TYPE_OVERLAY &&
+	    plane_cap && plane_cap->per_pixel_alpha) {
+		unsigned int blend_caps = BIT(DRM_MODE_BLEND_PIXEL_NONE) |
+					  BIT(DRM_MODE_BLEND_PREMULTI) |
+					  BIT(DRM_MODE_BLEND_COVERAGE);
+
+		drm_plane_create_alpha_property(plane);
+		drm_plane_create_blend_mode_property(plane, blend_caps);
+	}
+
+	if (plane->type == DRM_PLANE_TYPE_PRIMARY) {
+		/*
+		 * Allow OVERLAY planes to be used as underlays by assigning an
+		 * immutable zpos = # of OVERLAY planes to the PRIMARY plane.
+		 */
+		drm_plane_create_zpos_immutable_property(plane, primary_zpos);
+	} else if (plane->type == DRM_PLANE_TYPE_OVERLAY) {
+		/*
+		 * OVERLAY planes can be below or above the PRIMARY, but cannot
+		 * be above the CURSOR plane.
+		 */
+		unsigned int zpos = primary_zpos + 1 + drm_plane_index(plane);
+
+		drm_plane_create_zpos_property(plane, zpos, 0, 254);
+	} else if (plane->type == DRM_PLANE_TYPE_CURSOR) {
+		drm_plane_create_zpos_immutable_property(plane, 255);
+	}
+
+	if (plane->type == DRM_PLANE_TYPE_PRIMARY &&
+	    plane_cap &&
+	    (plane_cap->pixel_format_support.nv12 ||
+	     plane_cap->pixel_format_support.p010)) {
+		/* This only affects YUV formats. */
+		drm_plane_create_color_properties(
+			plane,
+			BIT(DRM_COLOR_YCBCR_BT601) |
+			BIT(DRM_COLOR_YCBCR_BT709) |
+			BIT(DRM_COLOR_YCBCR_BT2020),
+			BIT(DRM_COLOR_YCBCR_LIMITED_RANGE) |
+			BIT(DRM_COLOR_YCBCR_FULL_RANGE),
+			DRM_COLOR_YCBCR_BT709, DRM_COLOR_YCBCR_LIMITED_RANGE);
+	}
+
+	supported_rotations =
+		DRM_MODE_ROTATE_0 | DRM_MODE_ROTATE_90 |
+		DRM_MODE_ROTATE_180 | DRM_MODE_ROTATE_270;
+
+	if (dm->adev->asic_type >= CHIP_BONAIRE &&
+	    plane->type != DRM_PLANE_TYPE_CURSOR)
+		drm_plane_create_rotation_property(plane, DRM_MODE_ROTATE_0,
+						   supported_rotations);
+
+	if (amdgpu_ip_version(dm->adev, DCE_HWIP, 0) > IP_VERSION(3, 0, 1) &&
+	    plane->type != DRM_PLANE_TYPE_CURSOR)
+		drm_plane_enable_fb_damage_clips(plane);
+
+	if (plane->type == DRM_PLANE_TYPE_PRIMARY)
+		drm_plane_helper_add(plane, &dm_primary_plane_helper_funcs);
+	else
+		drm_plane_helper_add(plane, &dm_plane_helper_funcs);
+
+#ifdef AMD_PRIVATE_COLOR
+	dm_atomic_plane_attach_color_mgmt_properties(dm, plane);
+#else
+	res = dm_plane_init_colorops(plane);
+	if (res)
+		return res;
+#endif
+
+	/* Create (reset) the plane state */
+	if (plane->funcs->reset)
+		plane->funcs->reset(plane);
+
+	return 0;
+}
+
+bool amdgpu_dm_plane_is_video_format(uint32_t format)
+{
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(video_formats); i++)
+		if (format == video_formats[i])
+			return true;
+
+	return false;
+}
+
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_plane.h b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_plane.h
new file mode 100644
index 000000000000..ea2619b507db
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_plane.h
@@ -0,0 +1,68 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __AMDGPU_DM_PLANE_H__
+#define __AMDGPU_DM_PLANE_H__
+
+#include "dc.h"
+
+int amdgpu_dm_plane_get_cursor_position(struct drm_plane *plane, struct drm_crtc *crtc,
+					struct dc_cursor_position *position);
+
+void amdgpu_dm_plane_handle_cursor_update(struct drm_plane *plane,
+			  struct drm_plane_state *old_plane_state);
+
+int amdgpu_dm_plane_fill_dc_scaling_info(struct amdgpu_device *adev,
+			 const struct drm_plane_state *state,
+			 struct dc_scaling_info *scaling_info);
+
+int amdgpu_dm_plane_helper_check_state(struct drm_plane_state *state,
+				struct drm_crtc_state *new_crtc_state);
+
+int amdgpu_dm_plane_fill_plane_buffer_attributes(struct amdgpu_device *adev,
+				 const struct amdgpu_framebuffer *afb,
+				 const enum surface_pixel_format format,
+				 const enum dc_rotation_angle rotation,
+				 const uint64_t tiling_flags,
+				 struct dc_tiling_info *tiling_info,
+				 struct plane_size *plane_size,
+				 struct dc_plane_dcc_param *dcc,
+				 struct dc_plane_address *address,
+				 bool tmz_surface);
+
+int amdgpu_dm_plane_init(struct amdgpu_display_manager *dm,
+			 struct drm_plane *plane,
+			 unsigned long possible_crtcs,
+			 const struct dc_plane_cap *plane_cap);
+
+const struct drm_format_info *amdgpu_dm_plane_get_format_info(u32 pixel_format, u64 modifier);
+
+void amdgpu_dm_plane_fill_blending_from_plane_state(const struct drm_plane_state *plane_state,
+				    bool *per_pixel_alpha, bool *pre_multiplied_alpha,
+				    bool *global_alpha, int *global_alpha_value);
+
+bool amdgpu_dm_plane_is_video_format(uint32_t format);
+#endif
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_pp_smu.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_pp_smu.c
index eba270121698..11b2ea6edf95 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_pp_smu.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_pp_smu.c
@@ -1,3 +1,4 @@
+// SPDX-License-Identifier: MIT
 /*
  * Copyright 2018 Advanced Micro Devices, Inc.
  *
@@ -97,14 +98,12 @@ bool dm_pp_apply_display_requirements(
 			const struct dm_pp_single_disp_config *dc_cfg =
 						&pp_display_cfg->disp_configs[i];
 			adev->pm.pm_display_cfg.displays[i].controller_id = dc_cfg->pipe_idx + 1;
+			adev->pm.pm_display_cfg.displays[i].pixel_clock = dc_cfg->pixel_clock;
 		}
 
-		if (adev->powerplay.pp_funcs && adev->powerplay.pp_funcs->display_configuration_change)
-			adev->powerplay.pp_funcs->display_configuration_change(
-				adev->powerplay.pp_handle,
-				&adev->pm.pm_display_cfg);
+		amdgpu_dpm_display_configuration_change(adev, &adev->pm.pm_display_cfg);
 
-		amdgpu_pm_compute_clocks(adev);
+		amdgpu_dpm_compute_clocks(adev);
 	}
 
 	return true;
@@ -298,31 +297,25 @@ bool dm_pp_get_clock_levels_by_type(
 		struct dm_pp_clock_levels *dc_clks)
 {
 	struct amdgpu_device *adev = ctx->driver_context;
-	void *pp_handle = adev->powerplay.pp_handle;
 	struct amd_pp_clocks pp_clks = { 0 };
 	struct amd_pp_simple_clock_info validation_clks = { 0 };
 	uint32_t i;
 
-	if (adev->powerplay.pp_funcs && adev->powerplay.pp_funcs->get_clock_by_type) {
-		if (adev->powerplay.pp_funcs->get_clock_by_type(pp_handle,
-			dc_to_pp_clock_type(clk_type), &pp_clks)) {
-			/* Error in pplib. Provide default values. */
-			get_default_clock_levels(clk_type, dc_clks);
-			return true;
-		}
+	if (amdgpu_dpm_get_clock_by_type(adev,
+		dc_to_pp_clock_type(clk_type), &pp_clks)) {
+		/* Error in pplib. Provide default values. */
+		get_default_clock_levels(clk_type, dc_clks);
+		return true;
 	}
 
 	pp_to_dc_clock_levels(&pp_clks, dc_clks, clk_type);
 
-	if (adev->powerplay.pp_funcs && adev->powerplay.pp_funcs->get_display_mode_validation_clocks) {
-		if (adev->powerplay.pp_funcs->get_display_mode_validation_clocks(
-						pp_handle, &validation_clks)) {
-			/* Error in pplib. Provide default values. */
-			DRM_INFO("DM_PPLIB: Warning: using default validation clocks!\n");
-			validation_clks.engine_max_clock = 72000;
-			validation_clks.memory_max_clock = 80000;
-			validation_clks.level = 0;
-		}
+	if (amdgpu_dpm_get_display_mode_validation_clks(adev, &validation_clks)) {
+		/* Error in pplib. Provide default values. */
+		DRM_INFO("DM_PPLIB: Warning: using default validation clocks!\n");
+		validation_clks.engine_max_clock = 72000;
+		validation_clks.memory_max_clock = 80000;
+		validation_clks.level = 0;
 	}
 
 	DRM_INFO("DM_PPLIB: Validation clocks:\n");
@@ -343,7 +336,8 @@ bool dm_pp_get_clock_levels_by_type(
 			if (dc_clks->clocks_in_khz[i] > validation_clks.engine_max_clock) {
 				/* This clock is higher the validation clock.
 				 * Than means the previous one is the highest
-				 * non-boosted one. */
+				 * non-boosted one.
+				 */
 				DRM_INFO("DM_PPLIB: reducing engine clock level from %d to %d\n",
 						dc_clks->num_levels, i);
 				dc_clks->num_levels = i > 0 ? i : 1;
@@ -370,18 +364,14 @@ bool dm_pp_get_clock_levels_by_type_with_latency(
 	struct dm_pp_clock_levels_with_latency *clk_level_info)
 {
 	struct amdgpu_device *adev = ctx->driver_context;
-	void *pp_handle = adev->powerplay.pp_handle;
 	struct pp_clock_levels_with_latency pp_clks = { 0 };
-	const struct amd_pm_funcs *pp_funcs = adev->powerplay.pp_funcs;
 	int ret;
 
-	if (pp_funcs && pp_funcs->get_clock_by_type_with_latency) {
-		ret = pp_funcs->get_clock_by_type_with_latency(pp_handle,
-						dc_to_pp_clock_type(clk_type),
-						&pp_clks);
-		if (ret)
-			return false;
-	}
+	ret = amdgpu_dpm_get_clock_by_type_with_latency(adev,
+					dc_to_pp_clock_type(clk_type),
+					&pp_clks);
+	if (ret)
+		return false;
 
 	pp_to_dc_clock_levels_with_latency(&pp_clks, clk_level_info, clk_type);
 
@@ -394,18 +384,14 @@ bool dm_pp_get_clock_levels_by_type_with_voltage(
 	struct dm_pp_clock_levels_with_voltage *clk_level_info)
 {
 	struct amdgpu_device *adev = ctx->driver_context;
-	void *pp_handle = adev->powerplay.pp_handle;
 	struct pp_clock_levels_with_voltage pp_clk_info = {0};
-	const struct amd_pm_funcs *pp_funcs = adev->powerplay.pp_funcs;
 	int ret;
 
-	if (pp_funcs && pp_funcs->get_clock_by_type_with_voltage) {
-		ret = pp_funcs->get_clock_by_type_with_voltage(pp_handle,
-						dc_to_pp_clock_type(clk_type),
-						&pp_clk_info);
-		if (ret)
-			return false;
-	}
+	ret = amdgpu_dpm_get_clock_by_type_with_voltage(adev,
+					dc_to_pp_clock_type(clk_type),
+					&pp_clk_info);
+	if (ret)
+		return false;
 
 	pp_to_dc_clock_levels_with_voltage(&pp_clk_info, clk_level_info, clk_type);
 
@@ -417,19 +403,16 @@ bool dm_pp_notify_wm_clock_changes(
 	struct dm_pp_wm_sets_with_clock_ranges *wm_with_clock_ranges)
 {
 	struct amdgpu_device *adev = ctx->driver_context;
-	void *pp_handle = adev->powerplay.pp_handle;
-	const struct amd_pm_funcs *pp_funcs = adev->powerplay.pp_funcs;
 
 	/*
 	 * Limit this watermark setting for Polaris for now
 	 * TODO: expand this to other ASICs
 	 */
-	if ((adev->asic_type >= CHIP_POLARIS10) && (adev->asic_type <= CHIP_VEGAM)
-	     && pp_funcs && pp_funcs->set_watermarks_for_clocks_ranges) {
-		if (!pp_funcs->set_watermarks_for_clocks_ranges(pp_handle,
-						(void *)wm_with_clock_ranges))
-			return true;
-	}
+	if ((adev->asic_type >= CHIP_POLARIS10) &&
+	    (adev->asic_type <= CHIP_VEGAM) &&
+	    !amdgpu_dpm_set_watermarks_for_clocks_ranges(adev,
+							 (void *)wm_with_clock_ranges))
+		return true;
 
 	return false;
 }
@@ -456,12 +439,10 @@ bool dm_pp_apply_clock_for_voltage_request(
 	if (!pp_clock_request.clock_type)
 		return false;
 
-	if (adev->powerplay.pp_funcs && adev->powerplay.pp_funcs->display_clock_voltage_request)
-		ret = adev->powerplay.pp_funcs->display_clock_voltage_request(
-			adev->powerplay.pp_handle,
-			&pp_clock_request);
-	if (ret)
+	ret = amdgpu_dpm_display_clock_voltage_request(adev, &pp_clock_request);
+	if (ret && (ret != -EOPNOTSUPP))
 		return false;
+
 	return true;
 }
 
@@ -471,15 +452,8 @@ bool dm_pp_get_static_clocks(
 {
 	struct amdgpu_device *adev = ctx->driver_context;
 	struct amd_pp_clock_info pp_clk_info = {0};
-	int ret = 0;
 
-	if (adev->powerplay.pp_funcs && adev->powerplay.pp_funcs->get_current_clocks)
-		ret = adev->powerplay.pp_funcs->get_current_clocks(
-			adev->powerplay.pp_handle,
-			&pp_clk_info);
-	else
-		return false;
-	if (ret)
+	if (amdgpu_dpm_get_current_clocks(adev, &pp_clk_info))
 		return false;
 
 	static_clk_info->max_clocks_state = pp_to_dc_powerlevel_state(pp_clk_info.max_clocks_state);
@@ -494,8 +468,6 @@ static void pp_rv_set_wm_ranges(struct pp_smu *pp,
 {
 	const struct dc_context *ctx = pp->dm;
 	struct amdgpu_device *adev = ctx->driver_context;
-	void *pp_handle = adev->powerplay.pp_handle;
-	const struct amd_pm_funcs *pp_funcs = adev->powerplay.pp_funcs;
 	struct dm_pp_wm_sets_with_clock_ranges_soc15 wm_with_clock_ranges;
 	struct dm_pp_clock_range_for_dmif_wm_set_soc15 *wm_dce_clocks = wm_with_clock_ranges.wm_dmif_clocks_ranges;
 	struct dm_pp_clock_range_for_mcif_wm_set_soc15 *wm_soc_clocks = wm_with_clock_ranges.wm_mcif_clocks_ranges;
@@ -536,72 +508,48 @@ static void pp_rv_set_wm_ranges(struct pp_smu *pp,
 				ranges->writer_wm_sets[i].min_drain_clk_mhz * 1000;
 	}
 
-	if (pp_funcs && pp_funcs->set_watermarks_for_clocks_ranges)
-		pp_funcs->set_watermarks_for_clocks_ranges(pp_handle,
-							   &wm_with_clock_ranges);
+	amdgpu_dpm_set_watermarks_for_clocks_ranges(adev,
+						    &wm_with_clock_ranges);
 }
 
 static void pp_rv_set_pme_wa_enable(struct pp_smu *pp)
 {
 	const struct dc_context *ctx = pp->dm;
 	struct amdgpu_device *adev = ctx->driver_context;
-	void *pp_handle = adev->powerplay.pp_handle;
-	const struct amd_pm_funcs *pp_funcs = adev->powerplay.pp_funcs;
 
-	if (pp_funcs && pp_funcs->notify_smu_enable_pwe)
-		pp_funcs->notify_smu_enable_pwe(pp_handle);
+	amdgpu_dpm_notify_smu_enable_pwe(adev);
 }
 
 static void pp_rv_set_active_display_count(struct pp_smu *pp, int count)
 {
 	const struct dc_context *ctx = pp->dm;
 	struct amdgpu_device *adev = ctx->driver_context;
-	void *pp_handle = adev->powerplay.pp_handle;
-	const struct amd_pm_funcs *pp_funcs = adev->powerplay.pp_funcs;
-
-	if (!pp_funcs || !pp_funcs->set_active_display_count)
-		return;
 
-	pp_funcs->set_active_display_count(pp_handle, count);
+	amdgpu_dpm_set_active_display_count(adev, count);
 }
 
 static void pp_rv_set_min_deep_sleep_dcfclk(struct pp_smu *pp, int clock)
 {
 	const struct dc_context *ctx = pp->dm;
 	struct amdgpu_device *adev = ctx->driver_context;
-	void *pp_handle = adev->powerplay.pp_handle;
-	const struct amd_pm_funcs *pp_funcs = adev->powerplay.pp_funcs;
-
-	if (!pp_funcs || !pp_funcs->set_min_deep_sleep_dcefclk)
-		return;
 
-	pp_funcs->set_min_deep_sleep_dcefclk(pp_handle, clock);
+	amdgpu_dpm_set_min_deep_sleep_dcefclk(adev, clock);
 }
 
 static void pp_rv_set_hard_min_dcefclk_by_freq(struct pp_smu *pp, int clock)
 {
 	const struct dc_context *ctx = pp->dm;
 	struct amdgpu_device *adev = ctx->driver_context;
-	void *pp_handle = adev->powerplay.pp_handle;
-	const struct amd_pm_funcs *pp_funcs = adev->powerplay.pp_funcs;
 
-	if (!pp_funcs || !pp_funcs->set_hard_min_dcefclk_by_freq)
-		return;
-
-	pp_funcs->set_hard_min_dcefclk_by_freq(pp_handle, clock);
+	amdgpu_dpm_set_hard_min_dcefclk_by_freq(adev, clock);
 }
 
 static void pp_rv_set_hard_min_fclk_by_freq(struct pp_smu *pp, int mhz)
 {
 	const struct dc_context *ctx = pp->dm;
 	struct amdgpu_device *adev = ctx->driver_context;
-	void *pp_handle = adev->powerplay.pp_handle;
-	const struct amd_pm_funcs *pp_funcs = adev->powerplay.pp_funcs;
-
-	if (!pp_funcs || !pp_funcs->set_hard_min_fclk_by_freq)
-		return;
 
-	pp_funcs->set_hard_min_fclk_by_freq(pp_handle, mhz);
+	amdgpu_dpm_set_hard_min_fclk_by_freq(adev, mhz);
 }
 
 static enum pp_smu_status pp_nv_set_wm_ranges(struct pp_smu *pp,
@@ -609,11 +557,8 @@ static enum pp_smu_status pp_nv_set_wm_ranges(struct pp_smu *pp,
 {
 	const struct dc_context *ctx = pp->dm;
 	struct amdgpu_device *adev = ctx->driver_context;
-	void *pp_handle = adev->powerplay.pp_handle;
-	const struct amd_pm_funcs *pp_funcs = adev->powerplay.pp_funcs;
 
-	if (pp_funcs && pp_funcs->set_watermarks_for_clocks_ranges)
-		pp_funcs->set_watermarks_for_clocks_ranges(pp_handle, ranges);
+	amdgpu_dpm_set_watermarks_for_clocks_ranges(adev, ranges);
 
 	return PP_SMU_RESULT_OK;
 }
@@ -622,14 +567,13 @@ static enum pp_smu_status pp_nv_set_display_count(struct pp_smu *pp, int count)
 {
 	const struct dc_context *ctx = pp->dm;
 	struct amdgpu_device *adev = ctx->driver_context;
-	void *pp_handle = adev->powerplay.pp_handle;
-	const struct amd_pm_funcs *pp_funcs = adev->powerplay.pp_funcs;
+	int ret = 0;
 
-	if (!pp_funcs || !pp_funcs->set_active_display_count)
+	ret = amdgpu_dpm_set_active_display_count(adev, count);
+	if (ret == -EOPNOTSUPP)
 		return PP_SMU_RESULT_UNSUPPORTED;
-
-	/* 0: successful or smu.ppt_funcs->set_display_count = NULL;  1: fail */
-	if (pp_funcs->set_active_display_count(pp_handle, count))
+	else if (ret)
+		/* 0: successful or smu.ppt_funcs->set_display_count = NULL;  1: fail */
 		return PP_SMU_RESULT_FAIL;
 
 	return PP_SMU_RESULT_OK;
@@ -640,14 +584,13 @@ pp_nv_set_min_deep_sleep_dcfclk(struct pp_smu *pp, int mhz)
 {
 	const struct dc_context *ctx = pp->dm;
 	struct amdgpu_device *adev = ctx->driver_context;
-	void *pp_handle = adev->powerplay.pp_handle;
-	const struct amd_pm_funcs *pp_funcs = adev->powerplay.pp_funcs;
-
-	if (!pp_funcs || !pp_funcs->set_min_deep_sleep_dcefclk)
-		return PP_SMU_RESULT_UNSUPPORTED;
+	int ret = 0;
 
 	/* 0: successful or smu.ppt_funcs->set_deep_sleep_dcefclk = NULL;1: fail */
-	if (pp_funcs->set_min_deep_sleep_dcefclk(pp_handle, mhz))
+	ret = amdgpu_dpm_set_min_deep_sleep_dcefclk(adev, mhz);
+	if (ret == -EOPNOTSUPP)
+		return PP_SMU_RESULT_UNSUPPORTED;
+	else if (ret)
 		return PP_SMU_RESULT_FAIL;
 
 	return PP_SMU_RESULT_OK;
@@ -658,12 +601,8 @@ static enum pp_smu_status pp_nv_set_hard_min_dcefclk_by_freq(
 {
 	const struct dc_context *ctx = pp->dm;
 	struct amdgpu_device *adev = ctx->driver_context;
-	void *pp_handle = adev->powerplay.pp_handle;
-	const struct amd_pm_funcs *pp_funcs = adev->powerplay.pp_funcs;
 	struct pp_display_clock_request clock_req;
-
-	if (!pp_funcs || !pp_funcs->display_clock_voltage_request)
-		return PP_SMU_RESULT_UNSUPPORTED;
+	int ret = 0;
 
 	clock_req.clock_type = amd_pp_dcef_clock;
 	clock_req.clock_freq_in_khz = mhz * 1000;
@@ -671,7 +610,10 @@ static enum pp_smu_status pp_nv_set_hard_min_dcefclk_by_freq(
 	/* 0: successful or smu.ppt_funcs->display_clock_voltage_request = NULL
 	 * 1: fail
 	 */
-	if (pp_funcs->display_clock_voltage_request(pp_handle, &clock_req))
+	ret = amdgpu_dpm_display_clock_voltage_request(adev, &clock_req);
+	if (ret == -EOPNOTSUPP)
+		return PP_SMU_RESULT_UNSUPPORTED;
+	else if (ret)
 		return PP_SMU_RESULT_FAIL;
 
 	return PP_SMU_RESULT_OK;
@@ -682,12 +624,8 @@ pp_nv_set_hard_min_uclk_by_freq(struct pp_smu *pp, int mhz)
 {
 	const struct dc_context *ctx = pp->dm;
 	struct amdgpu_device *adev = ctx->driver_context;
-	void *pp_handle = adev->powerplay.pp_handle;
-	const struct amd_pm_funcs *pp_funcs = adev->powerplay.pp_funcs;
 	struct pp_display_clock_request clock_req;
-
-	if (!pp_funcs || !pp_funcs->display_clock_voltage_request)
-		return PP_SMU_RESULT_UNSUPPORTED;
+	int ret = 0;
 
 	clock_req.clock_type = amd_pp_mem_clock;
 	clock_req.clock_freq_in_khz = mhz * 1000;
@@ -695,7 +633,10 @@ pp_nv_set_hard_min_uclk_by_freq(struct pp_smu *pp, int mhz)
 	/* 0: successful or smu.ppt_funcs->display_clock_voltage_request = NULL
 	 * 1: fail
 	 */
-	if (pp_funcs->display_clock_voltage_request(pp_handle, &clock_req))
+	ret = amdgpu_dpm_display_clock_voltage_request(adev, &clock_req);
+	if (ret == -EOPNOTSUPP)
+		return PP_SMU_RESULT_UNSUPPORTED;
+	else if (ret)
 		return PP_SMU_RESULT_FAIL;
 
 	return PP_SMU_RESULT_OK;
@@ -706,14 +647,10 @@ static enum pp_smu_status pp_nv_set_pstate_handshake_support(
 {
 	const struct dc_context *ctx = pp->dm;
 	struct amdgpu_device *adev = ctx->driver_context;
-	void *pp_handle = adev->powerplay.pp_handle;
-	const struct amd_pm_funcs *pp_funcs = adev->powerplay.pp_funcs;
 
-	if (pp_funcs && pp_funcs->display_disable_memory_clock_switch) {
-		if (pp_funcs->display_disable_memory_clock_switch(pp_handle,
-								  !pstate_handshake_supported))
-			return PP_SMU_RESULT_FAIL;
-	}
+	if (amdgpu_dpm_display_disable_memory_clock_switch(adev,
+							  !pstate_handshake_supported))
+		return PP_SMU_RESULT_FAIL;
 
 	return PP_SMU_RESULT_OK;
 }
@@ -723,12 +660,8 @@ static enum pp_smu_status pp_nv_set_voltage_by_freq(struct pp_smu *pp,
 {
 	const struct dc_context *ctx = pp->dm;
 	struct amdgpu_device *adev = ctx->driver_context;
-	void *pp_handle = adev->powerplay.pp_handle;
-	const struct amd_pm_funcs *pp_funcs = adev->powerplay.pp_funcs;
 	struct pp_display_clock_request clock_req;
-
-	if (!pp_funcs || !pp_funcs->display_clock_voltage_request)
-		return PP_SMU_RESULT_UNSUPPORTED;
+	int ret = 0;
 
 	switch (clock_id) {
 	case PP_SMU_NV_DISPCLK:
@@ -748,7 +681,10 @@ static enum pp_smu_status pp_nv_set_voltage_by_freq(struct pp_smu *pp,
 	/* 0: successful or smu.ppt_funcs->display_clock_voltage_request = NULL
 	 * 1: fail
 	 */
-	if (pp_funcs->display_clock_voltage_request(pp_handle, &clock_req))
+	ret = amdgpu_dpm_display_clock_voltage_request(adev, &clock_req);
+	if (ret == -EOPNOTSUPP)
+		return PP_SMU_RESULT_UNSUPPORTED;
+	else if (ret)
 		return PP_SMU_RESULT_FAIL;
 
 	return PP_SMU_RESULT_OK;
@@ -759,16 +695,16 @@ static enum pp_smu_status pp_nv_get_maximum_sustainable_clocks(
 {
 	const struct dc_context *ctx = pp->dm;
 	struct amdgpu_device *adev = ctx->driver_context;
-	void *pp_handle = adev->powerplay.pp_handle;
-	const struct amd_pm_funcs *pp_funcs = adev->powerplay.pp_funcs;
+	int ret = 0;
 
-	if (!pp_funcs || !pp_funcs->get_max_sustainable_clocks_by_dc)
+	ret = amdgpu_dpm_get_max_sustainable_clocks_by_dc(adev,
+							  max_clocks);
+	if (ret == -EOPNOTSUPP)
 		return PP_SMU_RESULT_UNSUPPORTED;
+	else if (ret)
+		return PP_SMU_RESULT_FAIL;
 
-	if (!pp_funcs->get_max_sustainable_clocks_by_dc(pp_handle, max_clocks))
-		return PP_SMU_RESULT_OK;
-
-	return PP_SMU_RESULT_FAIL;
+	return PP_SMU_RESULT_OK;
 }
 
 static enum pp_smu_status pp_nv_get_uclk_dpm_states(struct pp_smu *pp,
@@ -776,18 +712,17 @@ static enum pp_smu_status pp_nv_get_uclk_dpm_states(struct pp_smu *pp,
 {
 	const struct dc_context *ctx = pp->dm;
 	struct amdgpu_device *adev = ctx->driver_context;
-	void *pp_handle = adev->powerplay.pp_handle;
-	const struct amd_pm_funcs *pp_funcs = adev->powerplay.pp_funcs;
+	int ret = 0;
 
-	if (!pp_funcs || !pp_funcs->get_uclk_dpm_states)
+	ret = amdgpu_dpm_get_uclk_dpm_states(adev,
+					     clock_values_in_khz,
+					    num_states);
+	if (ret == -EOPNOTSUPP)
 		return PP_SMU_RESULT_UNSUPPORTED;
+	else if (ret)
+		return PP_SMU_RESULT_FAIL;
 
-	if (!pp_funcs->get_uclk_dpm_states(pp_handle,
-					   clock_values_in_khz,
-					   num_states))
-		return PP_SMU_RESULT_OK;
-
-	return PP_SMU_RESULT_FAIL;
+	return PP_SMU_RESULT_OK;
 }
 
 static enum pp_smu_status pp_rn_get_dpm_clock_table(
@@ -795,16 +730,15 @@ static enum pp_smu_status pp_rn_get_dpm_clock_table(
 {
 	const struct dc_context *ctx = pp->dm;
 	struct amdgpu_device *adev = ctx->driver_context;
-	void *pp_handle = adev->powerplay.pp_handle;
-	const struct amd_pm_funcs *pp_funcs = adev->powerplay.pp_funcs;
+	int ret = 0;
 
-	if (!pp_funcs || !pp_funcs->get_dpm_clock_table)
+	ret = amdgpu_dpm_get_dpm_clock_table(adev, clock_table);
+	if (ret == -EOPNOTSUPP)
 		return PP_SMU_RESULT_UNSUPPORTED;
+	else if (ret)
+		return PP_SMU_RESULT_FAIL;
 
-	if (!pp_funcs->get_dpm_clock_table(pp_handle, clock_table))
-		return PP_SMU_RESULT_OK;
-
-	return PP_SMU_RESULT_FAIL;
+	return PP_SMU_RESULT_OK;
 }
 
 static enum pp_smu_status pp_rn_set_wm_ranges(struct pp_smu *pp,
@@ -812,11 +746,8 @@ static enum pp_smu_status pp_rn_set_wm_ranges(struct pp_smu *pp,
 {
 	const struct dc_context *ctx = pp->dm;
 	struct amdgpu_device *adev = ctx->driver_context;
-	void *pp_handle = adev->powerplay.pp_handle;
-	const struct amd_pm_funcs *pp_funcs = adev->powerplay.pp_funcs;
 
-	if (pp_funcs && pp_funcs->set_watermarks_for_clocks_ranges)
-		pp_funcs->set_watermarks_for_clocks_ranges(pp_handle, ranges);
+	amdgpu_dpm_set_watermarks_for_clocks_ranges(adev, ranges);
 
 	return PP_SMU_RESULT_OK;
 }
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_psr.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_psr.c
index c022e56f9459..fd491b7a3cd7 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_psr.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_psr.c
@@ -1,3 +1,4 @@
+// SPDX-License-Identifier: MIT
 /*
  * Copyright 2021 Advanced Micro Devices, Inc.
  *
@@ -24,8 +25,38 @@
  */
 
 #include "amdgpu_dm_psr.h"
+#include "dc_dmub_srv.h"
 #include "dc.h"
-#include "dm_helpers.h"
+#include "amdgpu_dm.h"
+#include "modules/power/power_helpers.h"
+
+static bool link_supports_psrsu(struct dc_link *link)
+{
+	struct dc *dc = link->ctx->dc;
+
+	if (!dc->caps.dmcub_support)
+		return false;
+
+	if (dc->ctx->dce_version < DCN_VERSION_3_1)
+		return false;
+
+	if (!is_psr_su_specific_panel(link))
+		return false;
+
+	if (!link->dpcd_caps.alpm_caps.bits.AUX_WAKE_ALPM_CAP ||
+	    !link->dpcd_caps.psr_info.psr_dpcd_caps.bits.Y_COORDINATE_REQUIRED)
+		return false;
+
+	if (link->dpcd_caps.psr_info.psr_dpcd_caps.bits.SU_GRANULARITY_REQUIRED &&
+	    !link->dpcd_caps.psr_info.psr2_su_y_granularity_cap)
+		return false;
+
+	if (amdgpu_dc_debug_mask & DC_DISABLE_PSR_SU)
+		return false;
+
+	/* Temporarily disable PSR-SU to avoid glitches */
+	return false;
+}
 
 /*
  * amdgpu_dm_set_psr_caps() - set link psr capabilities
@@ -34,25 +65,27 @@
  */
 void amdgpu_dm_set_psr_caps(struct dc_link *link)
 {
-	uint8_t dpcd_data[EDP_PSR_RECEIVER_CAP_SIZE];
-
-	if (!(link->connector_signal & SIGNAL_TYPE_EDP))
+	if (!(link->connector_signal & SIGNAL_TYPE_EDP)) {
+		link->psr_settings.psr_feature_enabled = false;
 		return;
-	if (link->type == dc_connection_none)
+	}
+
+	if (link->type == dc_connection_none) {
+		link->psr_settings.psr_feature_enabled = false;
 		return;
-	if (dm_helpers_dp_read_dpcd(NULL, link, DP_PSR_SUPPORT,
-					dpcd_data, sizeof(dpcd_data))) {
-		link->dpcd_caps.psr_caps.psr_version = dpcd_data[0];
-
-		if (dpcd_data[0] == 0) {
-			link->psr_settings.psr_version = DC_PSR_VERSION_UNSUPPORTED;
-			link->psr_settings.psr_feature_enabled = false;
-		} else {
+	}
+
+	if (link->dpcd_caps.psr_info.psr_version == 0) {
+		link->psr_settings.psr_version = DC_PSR_VERSION_UNSUPPORTED;
+		link->psr_settings.psr_feature_enabled = false;
+
+	} else {
+		if (link_supports_psrsu(link))
+			link->psr_settings.psr_version = DC_PSR_VERSION_SU_1;
+		else
 			link->psr_settings.psr_version = DC_PSR_VERSION_1;
-			link->psr_settings.psr_feature_enabled = true;
-		}
 
-		DRM_INFO("PSR support:%d\n", link->psr_settings.psr_feature_enabled);
+		link->psr_settings.psr_feature_enabled = true;
 	}
 }
 
@@ -67,21 +100,29 @@ bool amdgpu_dm_link_setup_psr(struct dc_stream_state *stream)
 	struct dc_link *link = NULL;
 	struct psr_config psr_config = {0};
 	struct psr_context psr_context = {0};
+	struct dc *dc = NULL;
 	bool ret = false;
 
 	if (stream == NULL)
 		return false;
 
 	link = stream->link;
+	dc = link->ctx->dc;
 
-	psr_config.psr_version = link->dpcd_caps.psr_caps.psr_version;
+	if (link->psr_settings.psr_version != DC_PSR_VERSION_UNSUPPORTED) {
+		mod_power_calc_psr_configs(&psr_config, link, stream);
 
-	if (psr_config.psr_version > 0) {
-		psr_config.psr_exit_link_training_required = 0x1;
-		psr_config.psr_frame_capture_indication_req = 0;
-		psr_config.psr_rfb_setup_time = 0x37;
-		psr_config.psr_sdp_transmit_line_num_deadline = 0x20;
-		psr_config.allow_smu_optimizations = 0x0;
+		/* linux DM specific updating for psr config fields */
+		psr_config.allow_smu_optimizations =
+			(amdgpu_dc_feature_mask & DC_PSR_ALLOW_SMU_OPT) &&
+			mod_power_only_edp(dc->current_state, stream);
+		psr_config.allow_multi_disp_optimizations =
+			(amdgpu_dc_feature_mask & DC_PSR_ALLOW_MULTI_DISP_OPT);
+
+		if (link->psr_settings.psr_version == DC_PSR_VERSION_SU_1) {
+			if (!psr_su_set_dsc_slice_height(dc, link, stream, &psr_config))
+				return false;
+		}
 
 		ret = dc_link_setup_psr(link, stream, &psr_config, &psr_context);
 
@@ -95,9 +136,8 @@ bool amdgpu_dm_link_setup_psr(struct dc_stream_state *stream)
  * amdgpu_dm_psr_enable() - enable psr f/w
  * @stream: stream state
  *
- * Return: true if success
  */
-bool amdgpu_dm_psr_enable(struct dc_stream_state *stream)
+void amdgpu_dm_psr_enable(struct dc_stream_state *stream)
 {
 	struct dc_link *link = stream->link;
 	unsigned int vsync_rate_hz = 0;
@@ -113,7 +153,7 @@ bool amdgpu_dm_psr_enable(struct dc_stream_state *stream)
 	DRM_DEBUG_DRIVER("Enabling psr...\n");
 
 	vsync_rate_hz = div64_u64(div64_u64((
-			stream->timing.pix_clk_100hz * 100),
+			stream->timing.pix_clk_100hz * (uint64_t)100),
 			stream->timing.v_total),
 			stream->timing.h_total);
 
@@ -123,6 +163,7 @@ bool amdgpu_dm_psr_enable(struct dc_stream_state *stream)
 	 */
 	if (vsync_rate_hz != 0) {
 		unsigned int frame_time_microsec = 1000000 / vsync_rate_hz;
+
 		num_frames_static = (30000 / frame_time_microsec) + 1;
 	}
 
@@ -135,9 +176,18 @@ bool amdgpu_dm_psr_enable(struct dc_stream_state *stream)
 					   &stream, 1,
 					   &params);
 
-	power_opt |= psr_power_opt_z10_static_screen;
+	/*
+	 * Only enable static-screen optimizations for PSR1. For PSR SU, this
+	 * causes vstartup interrupt issues, used by amdgpu_dm to send vblank
+	 * events.
+	 */
+	if (link->psr_settings.psr_version < DC_PSR_VERSION_SU_1)
+		power_opt |= psr_power_opt_z10_static_screen;
+
+	dc_link_set_psr_allow_active(link, &psr_enable, false, false, &power_opt);
 
-	return dc_link_set_psr_allow_active(link, &psr_enable, false, false, &power_opt);
+	if (link->ctx->dc->caps.ips_support)
+		dc_allow_idle_optimizations(link->ctx->dc, true);
 }
 
 /*
@@ -146,18 +196,17 @@ bool amdgpu_dm_psr_enable(struct dc_stream_state *stream)
  *
  * Return: true if success
  */
-bool amdgpu_dm_psr_disable(struct dc_stream_state *stream)
+bool amdgpu_dm_psr_disable(struct dc_stream_state *stream, bool wait)
 {
-	unsigned int power_opt = 0;
 	bool psr_enable = false;
 
 	DRM_DEBUG_DRIVER("Disabling psr...\n");
 
-	return dc_link_set_psr_allow_active(stream->link, &psr_enable, true, false, &power_opt);
+	return dc_link_set_psr_allow_active(stream->link, &psr_enable, wait, false, NULL);
 }
 
 /*
- * amdgpu_dm_psr_disable() - disable psr f/w
+ * amdgpu_dm_psr_disable_all() - disable psr f/w for all streams
  * if psr is enabled on any stream
  *
  * Return: true if success
@@ -168,3 +217,61 @@ bool amdgpu_dm_psr_disable_all(struct amdgpu_display_manager *dm)
 	return dc_set_psr_allow_active(dm->dc, false);
 }
 
+/*
+ * amdgpu_dm_psr_is_active_allowed() - check if psr is allowed on any stream
+ * @dm:  pointer to amdgpu_display_manager
+ *
+ * Return: true if allowed
+ */
+
+bool amdgpu_dm_psr_is_active_allowed(struct amdgpu_display_manager *dm)
+{
+	unsigned int i;
+	bool allow_active = false;
+
+	for (i = 0; i < dm->dc->current_state->stream_count ; i++) {
+		struct dc_link *link;
+		struct dc_stream_state *stream = dm->dc->current_state->streams[i];
+
+		link = stream->link;
+		if (!link)
+			continue;
+		if (link->psr_settings.psr_feature_enabled &&
+		    link->psr_settings.psr_allow_active) {
+			allow_active = true;
+			break;
+		}
+	}
+
+	return allow_active;
+}
+
+/**
+ * amdgpu_dm_psr_wait_disable() - Wait for eDP panel to exit PSR
+ * @stream: stream state attached to the eDP link
+ *
+ * Waits for a max of 500ms for the eDP panel to exit PSR.
+ *
+ * Return: true if panel exited PSR, false otherwise.
+ */
+bool amdgpu_dm_psr_wait_disable(struct dc_stream_state *stream)
+{
+	enum dc_psr_state psr_state = PSR_STATE0;
+	struct dc_link *link = stream->link;
+	int retry_count;
+
+	if (link == NULL)
+		return false;
+
+	for (retry_count = 0; retry_count <= 1000; retry_count++) {
+		dc_link_get_psr_state(link, &psr_state);
+		if (psr_state == PSR_STATE0)
+			break;
+		udelay(500);
+	}
+
+	if (retry_count == 1000)
+		return false;
+
+	return true;
+}
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_psr.h b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_psr.h
index 6806b3c9c84b..4fb8626913cf 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_psr.h
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_psr.h
@@ -1,3 +1,4 @@
+/* SPDX-License-Identifier: MIT */
 /*
  * Copyright 2021 Advanced Micro Devices, Inc.
  *
@@ -32,9 +33,11 @@
 #define AMDGPU_DM_PSR_ENTRY_DELAY 5
 
 void amdgpu_dm_set_psr_caps(struct dc_link *link);
-bool amdgpu_dm_psr_enable(struct dc_stream_state *stream);
+void amdgpu_dm_psr_enable(struct dc_stream_state *stream);
 bool amdgpu_dm_link_setup_psr(struct dc_stream_state *stream);
-bool amdgpu_dm_psr_disable(struct dc_stream_state *stream);
+bool amdgpu_dm_psr_disable(struct dc_stream_state *stream, bool wait);
 bool amdgpu_dm_psr_disable_all(struct amdgpu_display_manager *dm);
+bool amdgpu_dm_psr_is_active_allowed(struct amdgpu_display_manager *dm);
+bool amdgpu_dm_psr_wait_disable(struct dc_stream_state *stream);
 
 #endif /* AMDGPU_DM_AMDGPU_DM_PSR_H_ */
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_quirks.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_quirks.c
new file mode 100644
index 000000000000..1da07ebf9217
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_quirks.c
@@ -0,0 +1,178 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright 2025 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#include <linux/dmi.h>
+
+#include "amdgpu.h"
+#include "amdgpu_dm.h"
+
+struct amdgpu_dm_quirks {
+	bool aux_hpd_discon;
+	bool support_edp0_on_dp1;
+};
+
+static struct amdgpu_dm_quirks quirk_entries = {
+	.aux_hpd_discon = false,
+	.support_edp0_on_dp1 = false
+};
+
+static int edp0_on_dp1_callback(const struct dmi_system_id *id)
+{
+	quirk_entries.support_edp0_on_dp1 = true;
+	return 0;
+}
+
+static int aux_hpd_discon_callback(const struct dmi_system_id *id)
+{
+	quirk_entries.aux_hpd_discon = true;
+	return 0;
+}
+
+static const struct dmi_system_id dmi_quirk_table[] = {
+	{
+		.callback = aux_hpd_discon_callback,
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."),
+			DMI_MATCH(DMI_PRODUCT_NAME, "Precision 3660"),
+		},
+	},
+	{
+		.callback = aux_hpd_discon_callback,
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."),
+			DMI_MATCH(DMI_PRODUCT_NAME, "Precision 3260"),
+		},
+	},
+	{
+		.callback = aux_hpd_discon_callback,
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."),
+			DMI_MATCH(DMI_PRODUCT_NAME, "Precision 3460"),
+		},
+	},
+	{
+		.callback = aux_hpd_discon_callback,
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."),
+			DMI_MATCH(DMI_PRODUCT_NAME, "OptiPlex Tower Plus 7010"),
+		},
+	},
+	{
+		.callback = aux_hpd_discon_callback,
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."),
+			DMI_MATCH(DMI_PRODUCT_NAME, "OptiPlex Tower 7010"),
+		},
+	},
+	{
+		.callback = aux_hpd_discon_callback,
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."),
+			DMI_MATCH(DMI_PRODUCT_NAME, "OptiPlex SFF Plus 7010"),
+		},
+	},
+	{
+		.callback = aux_hpd_discon_callback,
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."),
+			DMI_MATCH(DMI_PRODUCT_NAME, "OptiPlex SFF 7010"),
+		},
+	},
+	{
+		.callback = aux_hpd_discon_callback,
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."),
+			DMI_MATCH(DMI_PRODUCT_NAME, "OptiPlex Micro Plus 7010"),
+		},
+	},
+	{
+		.callback = aux_hpd_discon_callback,
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."),
+			DMI_MATCH(DMI_PRODUCT_NAME, "OptiPlex Micro 7010"),
+		},
+	},
+	{
+		.callback = edp0_on_dp1_callback,
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "HP"),
+			DMI_MATCH(DMI_PRODUCT_NAME, "HP Elite mt645 G8 Mobile Thin Client"),
+		},
+	},
+	{
+		.callback = edp0_on_dp1_callback,
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "HP"),
+			DMI_MATCH(DMI_PRODUCT_NAME, "HP EliteBook 645 14 inch G11 Notebook PC"),
+		},
+	},
+	{
+		.callback = edp0_on_dp1_callback,
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "HP"),
+			DMI_MATCH(DMI_PRODUCT_NAME, "HP EliteBook 665 16 inch G11 Notebook PC"),
+		},
+	},
+	{
+		.callback = edp0_on_dp1_callback,
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "HP"),
+			DMI_MATCH(DMI_PRODUCT_NAME, "HP ProBook 445 14 inch G11 Notebook PC"),
+		},
+	},
+	{
+		.callback = edp0_on_dp1_callback,
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "HP"),
+			DMI_MATCH(DMI_PRODUCT_NAME, "HP ProBook 465 16 inch G11 Notebook PC"),
+		},
+	},
+	{}
+	/* TODO: refactor this from a fixed table to a dynamic option */
+};
+
+void retrieve_dmi_info(struct amdgpu_display_manager *dm)
+{
+	struct drm_device *dev = dm->ddev;
+	int dmi_id;
+
+	dm->aux_hpd_discon_quirk = false;
+	dm->edp0_on_dp1_quirk = false;
+
+	dmi_id = dmi_check_system(dmi_quirk_table);
+
+	if (!dmi_id)
+		return;
+
+	if (quirk_entries.aux_hpd_discon) {
+		dm->aux_hpd_discon_quirk = true;
+		drm_info(dev, "aux_hpd_discon_quirk attached\n");
+	}
+	if (quirk_entries.support_edp0_on_dp1) {
+		dm->edp0_on_dp1_quirk = true;
+		drm_info(dev, "support_edp0_on_dp1 attached\n");
+	}
+}
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_replay.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_replay.c
new file mode 100644
index 000000000000..da94e3544b65
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_replay.c
@@ -0,0 +1,209 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#include "amdgpu_dm_replay.h"
+#include "dc_dmub_srv.h"
+#include "dc.h"
+#include "dm_helpers.h"
+#include "amdgpu_dm.h"
+#include "modules/power/power_helpers.h"
+#include "dmub/inc/dmub_cmd.h"
+#include "dc/inc/link_service.h"
+
+/*
+ * amdgpu_dm_link_supports_replay() - check if the link supports replay
+ * @link: link
+ * @aconnector: aconnector
+ *
+ */
+bool amdgpu_dm_link_supports_replay(struct dc_link *link, struct amdgpu_dm_connector *aconnector)
+{
+	struct dm_connector_state *state = to_dm_connector_state(aconnector->base.state);
+	struct dpcd_caps *dpcd_caps = &link->dpcd_caps;
+	struct adaptive_sync_caps *as_caps = &link->dpcd_caps.adaptive_sync_caps;
+
+	if (!state->freesync_capable)
+		return false;
+
+	if (!aconnector->vsdb_info.replay_mode)
+		return false;
+
+	// Check the eDP version
+	if (dpcd_caps->edp_rev < EDP_REVISION_13)
+		return false;
+
+	if (!dpcd_caps->alpm_caps.bits.AUX_WAKE_ALPM_CAP)
+		return false;
+
+	// Check adaptive sync support cap
+	if (!as_caps->dp_adap_sync_caps.bits.ADAPTIVE_SYNC_SDP_SUPPORT)
+		return false;
+
+	// Sink shall populate line deviation information
+	if (dpcd_caps->pr_info.pixel_deviation_per_line == 0 ||
+		dpcd_caps->pr_info.max_deviation_line == 0)
+		return false;
+
+	return true;
+}
+
+/*
+ * amdgpu_dm_set_replay_caps() - setup Replay capabilities
+ * @link: link
+ * @aconnector: aconnector
+ *
+ */
+bool amdgpu_dm_set_replay_caps(struct dc_link *link, struct amdgpu_dm_connector *aconnector)
+{
+	struct replay_config pr_config = { 0 };
+	union replay_debug_flags *debug_flags = NULL;
+	struct dc *dc = link->ctx->dc;
+
+	// If Replay is already set to support, return true to skip checks
+	if (link->replay_settings.config.replay_supported)
+		return true;
+
+	if (!dc_is_embedded_signal(link->connector_signal))
+		return false;
+
+	if (link->panel_config.psr.disallow_replay)
+		return false;
+
+	if (!amdgpu_dm_link_supports_replay(link, aconnector))
+		return false;
+
+	if (!dc->ctx->dmub_srv || !dc->ctx->dmub_srv->dmub ||
+		!dc->ctx->dmub_srv->dmub->feature_caps.replay_supported)
+		return false;
+
+	// Mark Replay is supported in pr_config
+	pr_config.replay_supported = true;
+
+	debug_flags = (union replay_debug_flags *)&pr_config.debug_flags;
+	debug_flags->u32All = 0;
+	debug_flags->bitfields.visual_confirm =
+		link->ctx->dc->debug.visual_confirm == VISUAL_CONFIRM_REPLAY;
+
+	init_replay_config(link, &pr_config);
+
+	return true;
+}
+
+/*
+ * amdgpu_dm_link_setup_replay() - configure replay link
+ * @link: link
+ * @aconnector: aconnector
+ *
+ */
+bool amdgpu_dm_link_setup_replay(struct dc_link *link, struct amdgpu_dm_connector *aconnector)
+{
+	struct replay_config *pr_config;
+
+	if (link == NULL || aconnector == NULL)
+		return false;
+
+	pr_config = &link->replay_settings.config;
+
+	if (!pr_config->replay_supported)
+		return false;
+
+	pr_config->replay_power_opt_supported = 0x11;
+	pr_config->replay_smu_opt_supported = false;
+	pr_config->replay_enable_option |= pr_enable_option_static_screen;
+	pr_config->replay_support_fast_resync_in_ultra_sleep_mode = aconnector->max_vfreq >= 2 * aconnector->min_vfreq;
+	pr_config->replay_timing_sync_supported = false;
+
+	if (!pr_config->replay_timing_sync_supported)
+		pr_config->replay_enable_option &= ~pr_enable_option_general_ui;
+
+	link->replay_settings.replay_feature_enabled = true;
+
+	return true;
+}
+
+/*
+ * amdgpu_dm_replay_enable() - enable replay f/w
+ * @stream: stream state
+ *
+ * Return: true if success
+ */
+bool amdgpu_dm_replay_enable(struct dc_stream_state *stream, bool wait)
+{
+	bool replay_active = true;
+	struct dc_link *link = NULL;
+
+	if (stream == NULL)
+		return false;
+
+	link = stream->link;
+
+	if (link) {
+		link->dc->link_srv->edp_setup_replay(link, stream);
+		link->dc->link_srv->edp_set_coasting_vtotal(link, stream->timing.v_total, 0);
+		DRM_DEBUG_DRIVER("Enabling replay...\n");
+		link->dc->link_srv->edp_set_replay_allow_active(link, &replay_active, wait, false, NULL);
+		return true;
+	}
+
+	return false;
+}
+
+/*
+ * amdgpu_dm_replay_disable() - disable replay f/w
+ * @stream:  stream state
+ *
+ * Return: true if success
+ */
+bool amdgpu_dm_replay_disable(struct dc_stream_state *stream)
+{
+	bool replay_active = false;
+	struct dc_link *link = NULL;
+
+	if (stream == NULL)
+		return false;
+
+	link = stream->link;
+
+	if (link) {
+		DRM_DEBUG_DRIVER("Disabling replay...\n");
+		link->dc->link_srv->edp_set_replay_allow_active(stream->link, &replay_active, true, false, NULL);
+		return true;
+	}
+
+	return false;
+}
+
+/*
+ * amdgpu_dm_replay_disable_all() - disable replay f/w
+ * if replay is enabled on any stream
+ *
+ * Return: true if success
+ */
+bool amdgpu_dm_replay_disable_all(struct amdgpu_display_manager *dm)
+{
+	DRM_DEBUG_DRIVER("Disabling replay if replay is enabled on any stream\n");
+	return dc_set_replay_allow_active(dm->dc, false);
+}
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_replay.h b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_replay.h
new file mode 100644
index 000000000000..73b6c67ae5e7
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_replay.h
@@ -0,0 +1,49 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef AMDGPU_DM_AMDGPU_DM_REPLAY_H_
+#define AMDGPU_DM_AMDGPU_DM_REPLAY_H_
+
+#include "amdgpu.h"
+
+enum replay_enable_option {
+	pr_enable_option_static_screen = 0x1,
+	pr_enable_option_mpo_video = 0x2,
+	pr_enable_option_full_screen_video = 0x4,
+	pr_enable_option_general_ui = 0x8,
+	pr_enable_option_static_screen_coasting = 0x10000,
+	pr_enable_option_mpo_video_coasting = 0x20000,
+	pr_enable_option_full_screen_video_coasting = 0x40000,
+};
+
+bool amdgpu_dm_link_supports_replay(struct dc_link *link, struct amdgpu_dm_connector *aconnector);
+bool amdgpu_dm_replay_enable(struct dc_stream_state *stream, bool enable);
+bool amdgpu_dm_set_replay_caps(struct dc_link *link, struct amdgpu_dm_connector *aconnector);
+bool amdgpu_dm_link_setup_replay(struct dc_link *link, struct amdgpu_dm_connector *aconnector);
+bool amdgpu_dm_replay_disable(struct dc_stream_state *stream);
+bool amdgpu_dm_replay_disable_all(struct amdgpu_display_manager *dm);
+
+#endif /* AMDGPU_DM_AMDGPU_DM_REPLAY_H_ */
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_services.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_services.c
index d9e33c6bccd9..8550d5e8b753 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_services.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_services.c
@@ -1,3 +1,4 @@
+//SPDX-License-Identifier: MIT
 /*
  * Copyright 2015 Advanced Micro Devices, Inc.
  *
@@ -52,4 +53,12 @@ void dm_perf_trace_timestamp(const char *func_name, unsigned int line, struct dc
 				    func_name, line);
 }
 
+void dm_trace_smu_enter(uint32_t msg_id, uint32_t param_in, unsigned int delay, struct dc_context *ctx)
+{
+}
+
+void dm_trace_smu_exit(bool success, uint32_t response, struct dc_context *ctx)
+{
+}
+
 /**** power component interfaces ****/
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_trace.h b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_trace.h
index fdcaea22b456..aa56fd6d56c3 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_trace.h
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_trace.h
@@ -1,3 +1,4 @@
+//SPDX-License-Identifier: MIT
 /*
  * Copyright 2018 Advanced Micro Devices, Inc.
  *
@@ -34,8 +35,10 @@
 #include <drm/drm_crtc.h>
 #include <drm/drm_plane.h>
 #include <drm/drm_fourcc.h>
+#include <drm/drm_framebuffer.h>
 #include <drm/drm_encoder.h>
 #include <drm/drm_atomic.h>
+#include "dc/inc/hw/optc.h"
 
 #include "dc/inc/core_types.h"
 
@@ -85,7 +88,7 @@ TRACE_EVENT(amdgpu_dc_performance,
 			__entry->writes = write_count;
 			__entry->read_delta = read_count - *last_read;
 			__entry->write_delta = write_count - *last_write;
-			__assign_str(func, func);
+			__assign_str(func);
 			__entry->line = line;
 			*last_read = read_count;
 			*last_write = write_count;
@@ -661,6 +664,95 @@ TRACE_EVENT(dcn_fpu,
 	    )
 );
 
+TRACE_EVENT(dcn_optc_lock_unlock_state,
+	    TP_PROTO(const struct optc *optc_state, int instance, bool lock, const char *function, const int line),
+	    TP_ARGS(optc_state, instance, lock, function, line),
+
+	    TP_STRUCT__entry(
+			     __field(const char *, function)
+			     __field(int, instance)
+			     __field(bool, lock)
+			     __field(int, line)
+			     __field(int, opp_count)
+			     __field(int, max_h_total)
+			     __field(int, max_v_total)
+			     __field(int, min_h_blank)
+			     __field(int, min_h_sync_width)
+			     __field(int, min_v_sync_width)
+			     __field(int, min_v_blank)
+			     __field(int, min_v_blank_interlace)
+			     __field(int, vstartup_start)
+			     __field(int, vupdate_offset)
+			     __field(int, vupdate_width)
+			     __field(int, vready_offset)
+	    ),
+	    TP_fast_assign(
+			   __entry->function = function;
+			   __entry->instance = instance;
+			   __entry->lock = lock;
+			   __entry->line = line;
+			   __entry->opp_count = optc_state->opp_count;
+			   __entry->max_h_total = optc_state->max_h_total;
+			   __entry->max_v_total = optc_state->max_v_total;
+			   __entry->min_h_blank = optc_state->min_h_blank;
+			   __entry->min_h_sync_width = optc_state->min_h_sync_width;
+			   __entry->min_v_sync_width = optc_state->min_v_sync_width;
+			   __entry->min_v_blank = optc_state->min_v_blank;
+			   __entry->min_v_blank_interlace = optc_state->min_v_blank_interlace;
+			   __entry->vstartup_start = optc_state->vstartup_start;
+			   __entry->vupdate_offset = optc_state->vupdate_offset;
+			   __entry->vupdate_width = optc_state->vupdate_width;
+			   __entry->vready_offset = optc_state->vupdate_offset;
+	    ),
+	    TP_printk("%s: %s()+%d: optc_instance=%d opp_count=%d max_h_total=%d max_v_total=%d "
+		      "min_h_blank=%d min_h_sync_width=%d min_v_sync_width=%d min_v_blank=%d "
+		      "min_v_blank_interlace=%d vstartup_start=%d vupdate_offset=%d vupdate_width=%d "
+		      "vready_offset=%d",
+		      __entry->lock ? "Lock" : "Unlock",
+		      __entry->function,
+		      __entry->line,
+		      __entry->instance,
+		      __entry->opp_count,
+		      __entry->max_h_total,
+		      __entry->max_v_total,
+		      __entry->min_h_blank,
+		      __entry->min_h_sync_width,
+		      __entry->min_v_sync_width,
+		      __entry->min_v_blank,
+		      __entry->min_v_blank_interlace,
+		      __entry->vstartup_start,
+		      __entry->vupdate_offset,
+		      __entry->vupdate_width,
+		      __entry->vready_offset
+	    )
+);
+
+TRACE_EVENT(amdgpu_dm_brightness,
+	TP_PROTO(void *function, u32 user_brightness, u32 converted_brightness, bool aux, bool ac),
+	TP_ARGS(function, user_brightness, converted_brightness, aux, ac),
+	TP_STRUCT__entry(
+		__field(void *, function)
+		__field(u32, user_brightness)
+		__field(u32, converted_brightness)
+		__field(bool, aux)
+		__field(bool, ac)
+	),
+	TP_fast_assign(
+		__entry->function = function;
+		__entry->user_brightness = user_brightness;
+		__entry->converted_brightness = converted_brightness;
+		__entry->aux = aux;
+		__entry->ac = ac;
+	),
+	TP_printk("%ps: brightness requested=%u converted=%u aux=%s power=%s",
+		  (void *)__entry->function,
+		  (u32)__entry->user_brightness,
+		  (u32)__entry->converted_brightness,
+		  (__entry->aux) ? "true" : "false",
+		  (__entry->ac) ? "AC" : "DC"
+	)
+);
+
 #endif /* _AMDGPU_DM_TRACE_H_ */
 
 #undef TRACE_INCLUDE_PATH
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_wb.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_wb.c
new file mode 100644
index 000000000000..d9527c05fc87
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_wb.c
@@ -0,0 +1,215 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#include "dm_services_types.h"
+
+#include "amdgpu.h"
+#include "amdgpu_dm.h"
+#include "amdgpu_dm_wb.h"
+#include "amdgpu_display.h"
+#include "dc.h"
+
+#include <drm/drm_edid.h>
+#include <drm/drm_atomic_state_helper.h>
+#include <drm/drm_modeset_helper_vtables.h>
+
+static const u32 amdgpu_dm_wb_formats[] = {
+	DRM_FORMAT_XRGB2101010,
+};
+
+static int amdgpu_dm_wb_encoder_atomic_check(struct drm_encoder *encoder,
+					struct drm_crtc_state *crtc_state,
+					struct drm_connector_state *conn_state)
+{
+	struct drm_framebuffer *fb;
+	const struct drm_display_mode *mode = &crtc_state->mode;
+	bool found = false;
+	uint8_t i;
+
+	if (!conn_state->writeback_job || !conn_state->writeback_job->fb)
+		return 0;
+
+	fb = conn_state->writeback_job->fb;
+	if (fb->width != mode->hdisplay || fb->height != mode->vdisplay) {
+		DRM_DEBUG_KMS("Invalid framebuffer size %ux%u\n",
+			      fb->width, fb->height);
+		return -EINVAL;
+	}
+
+	for (i = 0; i < sizeof(amdgpu_dm_wb_formats) / sizeof(u32); i++) {
+		if (fb->format->format == amdgpu_dm_wb_formats[i])
+			found = true;
+	}
+
+	if (!found) {
+		DRM_DEBUG_KMS("Invalid pixel format %p4cc\n",
+			      &fb->format->format);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+
+static int amdgpu_dm_wb_connector_get_modes(struct drm_connector *connector)
+{
+	/* Maximum resolution supported by DWB */
+	return drm_add_modes_noedid(connector, 3840, 2160);
+}
+
+static int amdgpu_dm_wb_prepare_job(struct drm_writeback_connector *wb_connector,
+			       struct drm_writeback_job *job)
+{
+	struct amdgpu_framebuffer *afb;
+	struct drm_gem_object *obj;
+	struct amdgpu_device *adev;
+	struct amdgpu_bo *rbo;
+	uint32_t domain;
+	int r;
+
+	if (!job->fb) {
+		DRM_DEBUG_KMS("No FB bound\n");
+		return 0;
+	}
+
+	afb = to_amdgpu_framebuffer(job->fb);
+	obj = job->fb->obj[0];
+	rbo = gem_to_amdgpu_bo(obj);
+	adev = amdgpu_ttm_adev(rbo->tbo.bdev);
+
+	r = amdgpu_bo_reserve(rbo, true);
+	if (r) {
+		drm_err(adev_to_drm(adev), "fail to reserve bo (%d)\n", r);
+		return r;
+	}
+
+	r = dma_resv_reserve_fences(rbo->tbo.base.resv, 1);
+	if (r) {
+		drm_err(adev_to_drm(adev), "reserving fence slot failed (%d)\n", r);
+		goto error_unlock;
+	}
+
+	domain = amdgpu_display_supported_domains(adev, rbo->flags);
+
+	rbo->flags |= AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS;
+	r = amdgpu_bo_pin(rbo, domain);
+	if (unlikely(r != 0)) {
+		if (r != -ERESTARTSYS)
+			DRM_ERROR("Failed to pin framebuffer with error %d\n", r);
+		goto error_unlock;
+	}
+
+	r = amdgpu_ttm_alloc_gart(&rbo->tbo);
+	if (unlikely(r != 0)) {
+		DRM_ERROR("%p bind failed\n", rbo);
+		goto error_unpin;
+	}
+
+	amdgpu_bo_unreserve(rbo);
+
+	afb->address = amdgpu_bo_gpu_offset(rbo);
+
+	amdgpu_bo_ref(rbo);
+
+	return 0;
+
+error_unpin:
+	amdgpu_bo_unpin(rbo);
+
+error_unlock:
+	amdgpu_bo_unreserve(rbo);
+	return r;
+}
+
+static void amdgpu_dm_wb_cleanup_job(struct drm_writeback_connector *connector,
+				struct drm_writeback_job *job)
+{
+	struct amdgpu_bo *rbo;
+	int r;
+
+	if (!job->fb)
+		return;
+
+	rbo = gem_to_amdgpu_bo(job->fb->obj[0]);
+	r = amdgpu_bo_reserve(rbo, false);
+	if (unlikely(r)) {
+		DRM_ERROR("failed to reserve rbo before unpin\n");
+		return;
+	}
+
+	amdgpu_bo_unpin(rbo);
+	amdgpu_bo_unreserve(rbo);
+	amdgpu_bo_unref(&rbo);
+}
+
+static const struct drm_encoder_helper_funcs amdgpu_dm_wb_encoder_helper_funcs = {
+	.atomic_check = amdgpu_dm_wb_encoder_atomic_check,
+};
+
+static const struct drm_connector_funcs amdgpu_dm_wb_connector_funcs = {
+	.fill_modes = drm_helper_probe_single_connector_modes,
+	.destroy = drm_connector_cleanup,
+	.reset = amdgpu_dm_connector_funcs_reset,
+	.atomic_duplicate_state = amdgpu_dm_connector_atomic_duplicate_state,
+	.atomic_destroy_state = drm_atomic_helper_connector_destroy_state,
+};
+
+static const struct drm_connector_helper_funcs amdgpu_dm_wb_conn_helper_funcs = {
+	.get_modes = amdgpu_dm_wb_connector_get_modes,
+	.prepare_writeback_job = amdgpu_dm_wb_prepare_job,
+	.cleanup_writeback_job = amdgpu_dm_wb_cleanup_job,
+};
+
+int amdgpu_dm_wb_connector_init(struct amdgpu_display_manager *dm,
+				struct amdgpu_dm_wb_connector *wbcon,
+				uint32_t link_index)
+{
+	struct dc *dc = dm->dc;
+	struct dc_link *link = dc_get_link_at_index(dc, link_index);
+	int res = 0;
+
+	wbcon->link = link;
+
+	drm_connector_helper_add(&wbcon->base.base, &amdgpu_dm_wb_conn_helper_funcs);
+
+	res = drm_writeback_connector_init(&dm->adev->ddev, &wbcon->base,
+					    &amdgpu_dm_wb_connector_funcs,
+					    &amdgpu_dm_wb_encoder_helper_funcs,
+					    amdgpu_dm_wb_formats,
+					    ARRAY_SIZE(amdgpu_dm_wb_formats),
+					    amdgpu_dm_get_encoder_crtc_mask(dm->adev));
+
+	if (res)
+		return res;
+	/*
+	 * Some of the properties below require access to state, like bpc.
+	 * Allocate some default initial connector state with our reset helper.
+	 */
+	if (wbcon->base.base.funcs->reset)
+		wbcon->base.base.funcs->reset(&wbcon->base.base);
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_wb.h b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_wb.h
new file mode 100644
index 000000000000..13d31c857dee
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_wb.h
@@ -0,0 +1,36 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __AMDGPU_DM_WB_H__
+#define __AMDGPU_DM_WB_H__
+
+#include <drm/drm_writeback.h>
+
+int amdgpu_dm_wb_connector_init(struct amdgpu_display_manager *dm,
+				struct amdgpu_dm_wb_connector *dm_wbcon,
+				uint32_t link_index);
+
+#endif
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/dc_fpu.c b/drivers/gpu/drm/amd/display/amdgpu_dm/dc_fpu.c
index ab0c6d191038..e46f8ce41d87 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/dc_fpu.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/dc_fpu.c
@@ -26,12 +26,7 @@
 
 #include "dc_trace.h"
 
-#if defined(CONFIG_X86)
-#include <asm/fpu/api.h>
-#elif defined(CONFIG_PPC64)
-#include <asm/switch_to.h>
-#include <asm/cputable.h>
-#endif
+#include <linux/fpu.h>
 
 /**
  * DOC: DC FPU manipulation overview
@@ -56,11 +51,9 @@ static DEFINE_PER_CPU(int, fpu_recursion_depth);
  */
 inline void dc_assert_fp_enabled(void)
 {
-	int *pcpu, depth = 0;
+	int depth;
 
-	pcpu = get_cpu_ptr(&fpu_recursion_depth);
-	depth = *pcpu;
-	put_cpu_ptr(&fpu_recursion_depth);
+	depth = __this_cpu_read(fpu_recursion_depth);
 
 	ASSERT(depth >= 1);
 }
@@ -80,30 +73,17 @@ inline void dc_assert_fp_enabled(void)
  */
 void dc_fpu_begin(const char *function_name, const int line)
 {
-	int *pcpu;
+	int depth;
 
-	pcpu = get_cpu_ptr(&fpu_recursion_depth);
-	*pcpu += 1;
-
-	if (*pcpu == 1) {
-#if defined(CONFIG_X86)
+	WARN_ON_ONCE(!in_task());
+	preempt_disable();
+	depth = __this_cpu_inc_return(fpu_recursion_depth);
+	if (depth == 1) {
+		BUG_ON(!kernel_fpu_available());
 		kernel_fpu_begin();
-#elif defined(CONFIG_PPC64)
-		if (cpu_has_feature(CPU_FTR_VSX_COMP)) {
-			preempt_disable();
-			enable_kernel_vsx();
-		} else if (cpu_has_feature(CPU_FTR_ALTIVEC_COMP)) {
-			preempt_disable();
-			enable_kernel_altivec();
-		} else if (!cpu_has_feature(CPU_FTR_FPU_UNAVAILABLE)) {
-			preempt_disable();
-			enable_kernel_fp();
-		}
-#endif
 	}
 
-	TRACE_DCN_FPU(true, function_name, line, *pcpu);
-	put_cpu_ptr(&fpu_recursion_depth);
+	TRACE_DCN_FPU(true, function_name, line, depth);
 }
 
 /**
@@ -118,27 +98,15 @@ void dc_fpu_begin(const char *function_name, const int line)
  */
 void dc_fpu_end(const char *function_name, const int line)
 {
-	int *pcpu;
+	int depth;
 
-	pcpu = get_cpu_ptr(&fpu_recursion_depth);
-	*pcpu -= 1;
-	if (*pcpu <= 0) {
-#if defined(CONFIG_X86)
+	depth = __this_cpu_dec_return(fpu_recursion_depth);
+	if (depth == 0) {
 		kernel_fpu_end();
-#elif defined(CONFIG_PPC64)
-		if (cpu_has_feature(CPU_FTR_VSX_COMP)) {
-			disable_kernel_vsx();
-			preempt_enable();
-		} else if (cpu_has_feature(CPU_FTR_ALTIVEC_COMP)) {
-			disable_kernel_altivec();
-			preempt_enable();
-		} else if (!cpu_has_feature(CPU_FTR_FPU_UNAVAILABLE)) {
-			disable_kernel_fp();
-			preempt_enable();
-		}
-#endif
+	} else {
+		WARN_ON_ONCE(depth < 0);
 	}
 
-	TRACE_DCN_FPU(false, function_name, line, *pcpu);
-	put_cpu_ptr(&fpu_recursion_depth);
+	TRACE_DCN_FPU(false, function_name, line, depth);
+	preempt_enable();
 }
diff --git a/drivers/gpu/drm/amd/display/dc/Makefile b/drivers/gpu/drm/amd/display/dc/Makefile
index b1f0d6260226..7277ed21552f 100644
--- a/drivers/gpu/drm/amd/display/dc/Makefile
+++ b/drivers/gpu/drm/amd/display/dc/Makefile
@@ -21,59 +21,63 @@
 #
 #
 # Makefile for Display Core (dc) component.
-#
 
-DC_LIBS = basics bios calcs clk_mgr dce gpio irq virtual
+DC_LIBS = basics bios dml clk_mgr dce gpio hwss irq link virtual dsc resource optc dpp hubbub dccg hubp dio dwb hpo mmhubbub mpc opp pg
+
+ifdef CONFIG_DRM_AMD_DC_FP
 
-ifdef CONFIG_DRM_AMD_DC_DCN
+KCOV_INSTRUMENT := n
+
+DC_LIBS += dcn10
 DC_LIBS += dcn20
-DC_LIBS += dsc
-DC_LIBS += dcn10 dml
 DC_LIBS += dcn21
 DC_LIBS += dcn201
 DC_LIBS += dcn30
 DC_LIBS += dcn301
-DC_LIBS += dcn302
-DC_LIBS += dcn303
 DC_LIBS += dcn31
+DC_LIBS += dml
+DC_LIBS += dml2_0
+DC_LIBS += soc_and_ip_translator
 endif
 
 DC_LIBS += dce120
 
 DC_LIBS += dce112
 DC_LIBS += dce110
-DC_LIBS += dce100
 DC_LIBS += dce80
 
 ifdef CONFIG_DRM_AMD_DC_SI
 DC_LIBS += dce60
 endif
 
-ifdef CONFIG_DRM_AMD_DC_HDCP
 DC_LIBS += hdcp
+
+ifdef CONFIG_DRM_AMD_DC_FP
+DC_LIBS += sspl
+AMD_DISPLAY_FILES += $(addprefix $(AMDDALPATH)/dc/, dc_spl_translate.o)
 endif
 
 AMD_DC = $(addsuffix /Makefile, $(addprefix $(FULL_AMD_DISPLAY_PATH)/dc/,$(DC_LIBS)))
 
 include $(AMD_DC)
 
-DISPLAY_CORE = dc.o  dc_stat.o dc_link.o dc_resource.o dc_hw_sequencer.o dc_sink.o \
-dc_surface.o dc_link_hwss.o dc_link_dp.o dc_link_ddc.o dc_debug.o dc_stream.o \
-dc_link_enc_cfg.o dc_link_dpia.o dc_link_dpcd.o
-
-ifdef CONFIG_DRM_AMD_DC_DCN
-DISPLAY_CORE += dc_vm_helper.o
-endif
-
-AMD_DISPLAY_CORE = $(addprefix $(AMDDALPATH)/dc/core/,$(DISPLAY_CORE))
-
-AMD_DM_REG_UPDATE = $(addprefix $(AMDDALPATH)/dc/,dc_helper.o)
+FILES =
+FILES += dc_dmub_srv.o
+FILES += dc_edid_parser.o
+FILES += dc_fused_io.o
+FILES += dc_helper.o
+FILES += core/dc.o
+FILES += core/dc_debug.o
+FILES += core/dc_hw_sequencer.o
+FILES += core/dc_link_enc_cfg.o
+FILES += core/dc_link_exports.o
+FILES += core/dc_resource.o
+FILES += core/dc_sink.o
+FILES += core/dc_stat.o
+FILES += core/dc_state.o
+FILES += core/dc_stream.o
+FILES += core/dc_surface.o
+FILES += core/dc_vm_helper.o
 
-AMD_DISPLAY_FILES += $(AMD_DISPLAY_CORE)
-AMD_DISPLAY_FILES += $(AMD_DM_REG_UPDATE)
+AMD_DISPLAY_FILES += $(addprefix $(AMDDALPATH)/dc/, $(FILES))
 
-DC_DMUB += dc_dmub_srv.o
-DC_EDID += dc_edid_parser.o
-AMD_DISPLAY_DMUB = $(addprefix $(AMDDALPATH)/dc/,$(DC_DMUB))
-AMD_DISPLAY_EDID = $(addprefix $(AMDDALPATH)/dc/,$(DC_EDID))
-AMD_DISPLAY_FILES += $(AMD_DISPLAY_DMUB) $(AMD_DISPLAY_EDID)
diff --git a/drivers/gpu/drm/amd/display/dc/basics/Makefile b/drivers/gpu/drm/amd/display/dc/basics/Makefile
index 01b99e0d788e..aabcebf69049 100644
--- a/drivers/gpu/drm/amd/display/dc/basics/Makefile
+++ b/drivers/gpu/drm/amd/display/dc/basics/Makefile
@@ -24,7 +24,14 @@
 # It provides the general basic services required by other DAL
 # subcomponents.
 
-BASICS = conversion.o fixpt31_32.o vector.o dc_common.o
+BASICS := \
+	conversion.o \
+	fixpt31_32.o \
+	vector.o \
+	dc_common.o \
+	dce_calcs.o \
+	custom_float.o \
+	bw_fixed.o
 
 AMD_DAL_BASICS = $(addprefix $(AMDDALPATH)/dc/basics/,$(BASICS))
 
diff --git a/drivers/gpu/drm/amd/display/dc/calcs/bw_fixed.c b/drivers/gpu/drm/amd/display/dc/basics/bw_fixed.c
index 6ca288fb5fb9..c8cb89e0d4d0 100644
--- a/drivers/gpu/drm/amd/display/dc/calcs/bw_fixed.c
+++ b/drivers/gpu/drm/amd/display/dc/basics/bw_fixed.c
@@ -1,5 +1,6 @@
+// SPDX-License-Identifier: MIT
 /*
- * Copyright 2015 Advanced Micro Devices, Inc.
+ * Copyright 2023 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -25,12 +26,11 @@
 #include "dm_services.h"
 #include "bw_fixed.h"
 
+#define MAX_I64 \
+	((int64_t)((1ULL << 63) - 1))
 
 #define MIN_I64 \
-	(int64_t)(-(1LL << 63))
-
-#define MAX_I64 \
-	(int64_t)((1ULL << 63) - 1)
+	(-MAX_I64 - 1)
 
 #define FRACTIONAL_PART_MASK \
 	((1ULL << BW_FIXED_BITS_PER_FRACTIONAL_PART) - 1)
@@ -49,6 +49,7 @@ static uint64_t abs_i64(int64_t arg)
 struct bw_fixed bw_int_to_fixed_nonconst(int64_t value)
 {
 	struct bw_fixed res;
+
 	ASSERT(value < BW_FIXED_MAX_I32 && value > BW_FIXED_MIN_I32);
 	res.value = value << BW_FIXED_BITS_PER_FRACTIONAL_PART;
 	return res;
@@ -78,14 +79,12 @@ struct bw_fixed bw_frc_to_fixed(int64_t numerator, int64_t denominator)
 	{
 		uint32_t i = BW_FIXED_BITS_PER_FRACTIONAL_PART;
 
-		do
-		{
+		do {
 			remainder <<= 1;
 
 			res_value <<= 1;
 
-			if (remainder >= arg2_value)
-			{
+			if (remainder >= arg2_value) {
 				res_value |= 1;
 				remainder -= arg2_value;
 			}
@@ -108,9 +107,8 @@ struct bw_fixed bw_frc_to_fixed(int64_t numerator, int64_t denominator)
 	return res;
 }
 
-struct bw_fixed bw_floor2(
-	const struct bw_fixed arg,
-	const struct bw_fixed significance)
+struct bw_fixed bw_floor2(const struct bw_fixed arg,
+			  const struct bw_fixed significance)
 {
 	struct bw_fixed result;
 	int64_t multiplicand;
@@ -121,9 +119,8 @@ struct bw_fixed bw_floor2(
 	return result;
 }
 
-struct bw_fixed bw_ceil2(
-	const struct bw_fixed arg,
-	const struct bw_fixed significance)
+struct bw_fixed bw_ceil2(const struct bw_fixed arg,
+			 const struct bw_fixed significance)
 {
 	struct bw_fixed result;
 	int64_t multiplicand;
diff --git a/drivers/gpu/drm/amd/display/dc/calcs/calcs_logger.h b/drivers/gpu/drm/amd/display/dc/basics/calcs_logger.h
index 62435bfc274d..62435bfc274d 100644
--- a/drivers/gpu/drm/amd/display/dc/calcs/calcs_logger.h
+++ b/drivers/gpu/drm/amd/display/dc/basics/calcs_logger.h
diff --git a/drivers/gpu/drm/amd/display/dc/basics/conversion.c b/drivers/gpu/drm/amd/display/dc/basics/conversion.c
index 6767fab55c26..bd1f60ecaba4 100644
--- a/drivers/gpu/drm/amd/display/dc/basics/conversion.c
+++ b/drivers/gpu/drm/amd/display/dc/basics/conversion.c
@@ -24,7 +24,7 @@
  */
 
 #include "dm_services.h"
-#include "conversion.h"
+#include "basics/conversion.h"
 
 #define DIVIDER 10000
 
@@ -100,3 +100,59 @@ void convert_float_matrix(
 		matrix[i] = (uint16_t)reg_value;
 	}
 }
+
+static struct fixed31_32 int_frac_to_fixed_point(uint16_t arg,
+						 uint8_t integer_bits,
+						 uint8_t fractional_bits)
+{
+	struct fixed31_32 result;
+	uint16_t sign_mask = 1 << (fractional_bits + integer_bits);
+	uint16_t value_mask = sign_mask - 1;
+
+	result.value = (long long)(arg & value_mask) <<
+		       (FIXED31_32_BITS_PER_FRACTIONAL_PART - fractional_bits);
+
+	if (arg & sign_mask)
+		result = dc_fixpt_neg(result);
+
+	return result;
+}
+
+/**
+ * convert_hw_matrix - converts HW values into fixed31_32 matrix.
+ * @matrix: fixed point 31.32 matrix
+ * @reg: array of register values
+ * @buffer_size: size of the array of register values
+ *
+ * Converts HW register spec defined format S2D13 into a fixed-point 31.32
+ * matrix.
+ */
+void convert_hw_matrix(struct fixed31_32 *matrix,
+		       uint16_t *reg,
+		       uint32_t buffer_size)
+{
+	for (int i = 0; i < buffer_size; ++i)
+		matrix[i] = int_frac_to_fixed_point(reg[i], 2, 13);
+}
+
+static uint32_t find_gcd(uint32_t a, uint32_t b)
+{
+	uint32_t remainder;
+
+	while (b != 0) {
+		remainder = a % b;
+		a = b;
+		b = remainder;
+	}
+	return a;
+}
+
+void reduce_fraction(uint32_t num, uint32_t den,
+		uint32_t *out_num, uint32_t *out_den)
+{
+	uint32_t gcd = 0;
+
+	gcd = find_gcd(num, den);
+	*out_num = num / gcd;
+	*out_den = den / gcd;
+}
diff --git a/drivers/gpu/drm/amd/display/dc/basics/conversion.h b/drivers/gpu/drm/amd/display/dc/basics/conversion.h
index ade785c4fdc7..a433cef78496 100644
--- a/drivers/gpu/drm/amd/display/dc/basics/conversion.h
+++ b/drivers/gpu/drm/amd/display/dc/basics/conversion.h
@@ -38,6 +38,13 @@ void convert_float_matrix(
 	struct fixed31_32 *flt,
 	uint32_t buffer_size);
 
+void reduce_fraction(uint32_t num, uint32_t den,
+		uint32_t *out_num, uint32_t *out_den);
+
+void convert_hw_matrix(struct fixed31_32 *matrix,
+		       uint16_t *reg,
+		       uint32_t buffer_size);
+
 static inline unsigned int log_2(unsigned int num)
 {
 	return ilog2(num);
diff --git a/drivers/gpu/drm/amd/display/dc/basics/custom_float.c b/drivers/gpu/drm/amd/display/dc/basics/custom_float.c
new file mode 100644
index 000000000000..ae05ded9a7f3
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/basics/custom_float.c
@@ -0,0 +1,173 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+#include "dm_services.h"
+#include "custom_float.h"
+
+static bool build_custom_float(struct fixed31_32 value,
+			       const struct custom_float_format *format,
+			       bool *negative,
+			       uint32_t *mantissa,
+			       uint32_t *exponenta)
+{
+	uint32_t exp_offset = (1 << (format->exponenta_bits - 1)) - 1;
+
+	const struct fixed31_32 mantissa_constant_plus_max_fraction =
+		dc_fixpt_from_fraction((1LL << (format->mantissa_bits + 1)) - 1,
+				       1LL << format->mantissa_bits);
+
+	struct fixed31_32 mantiss;
+
+	if (dc_fixpt_eq(value, dc_fixpt_zero)) {
+		*negative = false;
+		*mantissa = 0;
+		*exponenta = 0;
+		return true;
+	}
+
+	if (dc_fixpt_lt(value, dc_fixpt_zero)) {
+		*negative = format->sign;
+		value = dc_fixpt_neg(value);
+	} else {
+		*negative = false;
+	}
+
+	if (dc_fixpt_lt(value, dc_fixpt_one)) {
+		uint32_t i = 1;
+
+		do {
+			value = dc_fixpt_shl(value, 1);
+			++i;
+		} while (dc_fixpt_lt(value, dc_fixpt_one));
+
+		--i;
+
+		if (exp_offset <= i) {
+			*mantissa = 0;
+			*exponenta = 0;
+			return true;
+		}
+
+		*exponenta = exp_offset - i;
+	} else if (dc_fixpt_le(mantissa_constant_plus_max_fraction, value)) {
+		uint32_t i = 1;
+
+		do {
+			value = dc_fixpt_shr(value, 1);
+			++i;
+		} while (dc_fixpt_lt(mantissa_constant_plus_max_fraction, value));
+
+		*exponenta = exp_offset + i - 1;
+	} else {
+		*exponenta = exp_offset;
+	}
+
+	mantiss = dc_fixpt_sub(value, dc_fixpt_one);
+
+	if (dc_fixpt_lt(mantiss, dc_fixpt_zero) ||
+	    dc_fixpt_lt(dc_fixpt_one, mantiss))
+		mantiss = dc_fixpt_zero;
+	else
+		mantiss = dc_fixpt_shl(mantiss, format->mantissa_bits);
+
+	*mantissa = dc_fixpt_floor(mantiss);
+
+	return true;
+}
+
+static bool setup_custom_float(const struct custom_float_format *format,
+			       bool negative,
+			       uint32_t mantissa,
+			       uint32_t exponenta,
+			       uint32_t *result)
+{
+	uint32_t i = 0;
+	uint32_t j = 0;
+	uint32_t value = 0;
+
+	/* verification code:
+	 * once calculation is ok we can remove it
+	 */
+
+	const uint32_t mantissa_mask =
+		(1 << (format->mantissa_bits + 1)) - 1;
+
+	const uint32_t exponenta_mask =
+		(1 << (format->exponenta_bits + 1)) - 1;
+
+	if (mantissa & ~mantissa_mask) {
+		BREAK_TO_DEBUGGER();
+		mantissa = mantissa_mask;
+	}
+
+	if (exponenta & ~exponenta_mask) {
+		BREAK_TO_DEBUGGER();
+		exponenta = exponenta_mask;
+	}
+
+	/* end of verification code */
+
+	while (i < format->mantissa_bits) {
+		uint32_t mask = 1 << i;
+
+		if (mantissa & mask)
+			value |= mask;
+
+		++i;
+	}
+
+	while (j < format->exponenta_bits) {
+		uint32_t mask = 1 << j;
+
+		if (exponenta & mask)
+			value |= mask << i;
+
+		++j;
+	}
+
+	if (negative && format->sign)
+		value |= 1 << (i + j);
+
+	*result = value;
+
+	return true;
+}
+
+bool convert_to_custom_float_format(struct fixed31_32 value,
+				    const struct custom_float_format *format,
+				    uint32_t *result)
+{
+	uint32_t mantissa;
+	uint32_t exponenta;
+	bool negative;
+
+	return build_custom_float(value, format, &negative, &mantissa, &exponenta) &&
+				  setup_custom_float(format,
+						     negative,
+						     mantissa,
+						     exponenta,
+						     result);
+}
+
diff --git a/drivers/gpu/drm/amd/display/dc/basics/dc_common.c b/drivers/gpu/drm/amd/display/dc/basics/dc_common.c
index b2fc4f8e6482..a51c2701da24 100644
--- a/drivers/gpu/drm/amd/display/dc/basics/dc_common.c
+++ b/drivers/gpu/drm/amd/display/dc/basics/dc_common.c
@@ -40,7 +40,8 @@ bool is_rgb_cspace(enum dc_color_space output_color_space)
 	case COLOR_SPACE_YCBCR709:
 	case COLOR_SPACE_YCBCR601_LIMITED:
 	case COLOR_SPACE_YCBCR709_LIMITED:
-	case COLOR_SPACE_2020_YCBCR:
+	case COLOR_SPACE_2020_YCBCR_LIMITED:
+	case COLOR_SPACE_2020_YCBCR_FULL:
 		return false;
 	default:
 		/* Add a case to switch */
diff --git a/drivers/gpu/drm/amd/display/dc/calcs/dce_calcs.c b/drivers/gpu/drm/amd/display/dc/basics/dce_calcs.c
index ff5bb152ef49..4da5adab799c 100644
--- a/drivers/gpu/drm/amd/display/dc/calcs/dce_calcs.c
+++ b/drivers/gpu/drm/amd/display/dc/basics/dce_calcs.c
@@ -23,8 +23,6 @@
  *
  */
 
-#include <linux/slab.h>
-
 #include "resource.h"
 #include "dm_services.h"
 #include "dce_calcs.h"
@@ -94,7 +92,7 @@ static void calculate_bandwidth(
 	const uint32_t s_high = 7;
 	const uint32_t dmif_chunk_buff_margin = 1;
 
-	uint32_t max_chunks_fbc_mode;
+	uint32_t max_chunks_fbc_mode = 0;
 	int32_t num_cursor_lines;
 
 	int32_t i, j, k;
@@ -571,7 +569,7 @@ static void calculate_bandwidth(
 				break;
 			}
 			data->lb_partitions[i] = bw_floor2(bw_div(data->lb_size_per_component[i], data->lb_line_pitch), bw_int_to_fixed(1));
-			/*clamp the partitions to the maxium number supported by the lb*/
+			/* clamp the partitions to the maximum number supported by the lb */
 			if ((surface_type[i] != bw_def_graphics || dceip->graphics_lb_nodownscaling_multi_line_prefetching == 1)) {
 				data->lb_partitions_max[i] = bw_int_to_fixed(10);
 			}
@@ -1138,7 +1136,7 @@ static void calculate_bandwidth(
 			}
 		}
 	}
-	data->total_dmifmc_urgent_trips = bw_ceil2(bw_div(data->total_requests_for_adjusted_dmif_size, (bw_add(dceip->dmif_request_buffer_size, bw_int_to_fixed(vbios->number_of_request_slots_gmc_reserves_for_dmif_per_channel * data->number_of_dram_channels)))), bw_int_to_fixed(1));
+	data->total_dmifmc_urgent_trips = bw_ceil2(bw_div(data->total_requests_for_adjusted_dmif_size, (bw_add(dceip->dmif_request_buffer_size, bw_int_to_fixed((uint64_t)vbios->number_of_request_slots_gmc_reserves_for_dmif_per_channel * data->number_of_dram_channels)))), bw_int_to_fixed(1));
 	data->total_dmifmc_urgent_latency = bw_mul(vbios->dmifmc_urgent_latency, data->total_dmifmc_urgent_trips);
 	data->total_display_reads_required_data = bw_int_to_fixed(0);
 	data->total_display_reads_required_dram_access_data = bw_int_to_fixed(0);
@@ -1395,7 +1393,7 @@ static void calculate_bandwidth(
 						if ((bw_mtn(data->dram_speed_change_margin, bw_int_to_fixed(0)) && bw_ltn(data->dram_speed_change_margin, bw_int_to_fixed(9999)))) {
 							/*determine the minimum dram clock change margin for each set of clock frequencies*/
 							data->min_dram_speed_change_margin[i][j] = bw_min2(data->min_dram_speed_change_margin[i][j], data->dram_speed_change_margin);
-							/*compute the maximum clock frequuency required for the dram clock change at each set of clock frequencies*/
+							/*compute the maximum clock frequency required for the dram clock change at each set of clock frequencies*/
 							data->dispclk_required_for_dram_speed_change_pipe[i][j] = bw_max2(bw_div(bw_div(bw_mul(data->src_pixels_for_first_output_pixel[k], dceip->display_pipe_throughput_factor), dceip->lb_write_pixels_per_dispclk), (bw_sub(bw_sub(bw_sub(data->maximum_latency_hiding_with_cursor[k], vbios->nbp_state_change_latency), data->dmif_burst_time[i][j]), data->dram_speed_change_line_source_transfer_time[k][i][j]))), bw_div(bw_div(bw_mul(data->src_pixels_for_last_output_pixel[k], dceip->display_pipe_throughput_factor), dceip->lb_write_pixels_per_dispclk), (bw_add(bw_sub(bw_sub(bw_sub(data->maximum_latency_hiding_with_cursor[k], vbios->nbp_state_change_latency), data->dmif_burst_time[i][j]), data->dram_speed_change_line_source_transfer_time[k][i][j]), data->active_time[k]))));
 							if ((bw_ltn(data->dispclk_required_for_dram_speed_change_pipe[i][j], vbios->high_voltage_max_dispclk))) {
 								data->display_pstate_change_enable[k] = 1;
@@ -1409,7 +1407,7 @@ static void calculate_bandwidth(
 						if ((bw_mtn(data->dram_speed_change_margin, bw_int_to_fixed(0)) && bw_ltn(data->dram_speed_change_margin, bw_int_to_fixed(9999)))) {
 							/*determine the minimum dram clock change margin for each display pipe*/
 							data->min_dram_speed_change_margin[i][j] = bw_min2(data->min_dram_speed_change_margin[i][j], data->dram_speed_change_margin);
-							/*compute the maximum clock frequuency required for the dram clock change at each set of clock frequencies*/
+							/*compute the maximum clock frequency required for the dram clock change at each set of clock frequencies*/
 							data->dispclk_required_for_dram_speed_change_pipe[i][j] = bw_max2(bw_div(bw_div(bw_mul(data->src_pixels_for_first_output_pixel[k], dceip->display_pipe_throughput_factor), dceip->lb_write_pixels_per_dispclk), (bw_sub(bw_sub(bw_sub(bw_sub(data->maximum_latency_hiding_with_cursor[k], vbios->nbp_state_change_latency), data->dmif_burst_time[i][j]), data->dram_speed_change_line_source_transfer_time[k][i][j]), data->mcifwr_burst_time[i][j]))), bw_div(bw_div(bw_mul(data->src_pixels_for_last_output_pixel[k], dceip->display_pipe_throughput_factor), dceip->lb_write_pixels_per_dispclk), (bw_add(bw_sub(bw_sub(bw_sub(bw_sub(data->maximum_latency_hiding_with_cursor[k], vbios->nbp_state_change_latency), data->dmif_burst_time[i][j]), data->dram_speed_change_line_source_transfer_time[k][i][j]), data->mcifwr_burst_time[i][j]), data->active_time[k]))));
 							if ((bw_ltn(data->dispclk_required_for_dram_speed_change_pipe[i][j], vbios->high_voltage_max_dispclk))) {
 								data->display_pstate_change_enable[k] = 1;
@@ -1855,7 +1853,7 @@ static void calculate_bandwidth(
 	/*compute total time to request one chunk from each active display pipe*/
 	for (i = 0; i <= maximum_number_of_surfaces - 1; i++) {
 		if (data->enable[i]) {
-			data->chunk_request_time = bw_add(data->chunk_request_time, (bw_div((bw_div(bw_int_to_fixed(pixels_per_chunk * data->bytes_per_pixel[i]), data->useful_bytes_per_request[i])), bw_min2(sclk[data->sclk_level], bw_div(data->dispclk, bw_int_to_fixed(2))))));
+			data->chunk_request_time = bw_add(data->chunk_request_time, (bw_div((bw_div(bw_int_to_fixed(pixels_per_chunk * (int64_t)data->bytes_per_pixel[i]), data->useful_bytes_per_request[i])), bw_min2(sclk[data->sclk_level], bw_div(data->dispclk, bw_int_to_fixed(2))))));
 		}
 	}
 	/*compute total time to request cursor data*/
@@ -2033,10 +2031,10 @@ static void calculate_bandwidth(
 	kfree(surface_type);
 free_tiling_mode:
 	kfree(tiling_mode);
-free_yclk:
-	kfree(yclk);
 free_sclk:
 	kfree(sclk);
+free_yclk:
+	kfree(yclk);
 }
 
 /*******************************************************************************
@@ -3015,7 +3013,7 @@ static bool all_displays_in_sync(const struct pipe_ctx pipe[],
 	int i, num_active_pipes = 0;
 
 	for (i = 0; i < pipe_count; i++) {
-		if (!pipe[i].stream || pipe[i].top_pipe)
+		if (!resource_is_pipe_type(&pipe[i], OPP_HEAD))
 			continue;
 
 		active_pipes[num_active_pipes++] = &pipe[i];
@@ -3411,35 +3409,33 @@ bool bw_calcs(struct dc_context *ctx,
 			calcs_output->stutter_exit_wm_ns[5].c_mark =
 				bw_fixed_to_int(bw_mul(data->
 					stutter_exit_watermark[9], bw_int_to_fixed(1000)));
-
-		calcs_output->stutter_entry_wm_ns[0].c_mark =
-			bw_fixed_to_int(bw_mul(data->
-				stutter_entry_watermark[4], bw_int_to_fixed(1000)));
-		calcs_output->stutter_entry_wm_ns[1].c_mark =
-			bw_fixed_to_int(bw_mul(data->
-				stutter_entry_watermark[5], bw_int_to_fixed(1000)));
-		calcs_output->stutter_entry_wm_ns[2].c_mark =
-			bw_fixed_to_int(bw_mul(data->
-				stutter_entry_watermark[6], bw_int_to_fixed(1000)));
-		if (ctx->dc->caps.max_slave_planes) {
-			calcs_output->stutter_entry_wm_ns[3].c_mark =
+			calcs_output->stutter_entry_wm_ns[0].c_mark =
 				bw_fixed_to_int(bw_mul(data->
-					stutter_entry_watermark[0], bw_int_to_fixed(1000)));
-			calcs_output->stutter_entry_wm_ns[4].c_mark =
+					stutter_entry_watermark[4], bw_int_to_fixed(1000)));
+			calcs_output->stutter_entry_wm_ns[1].c_mark =
 				bw_fixed_to_int(bw_mul(data->
-					stutter_entry_watermark[1], bw_int_to_fixed(1000)));
-		} else {
-			calcs_output->stutter_entry_wm_ns[3].c_mark =
+					stutter_entry_watermark[5], bw_int_to_fixed(1000)));
+			calcs_output->stutter_entry_wm_ns[2].c_mark =
 				bw_fixed_to_int(bw_mul(data->
-					stutter_entry_watermark[7], bw_int_to_fixed(1000)));
-			calcs_output->stutter_entry_wm_ns[4].c_mark =
+					stutter_entry_watermark[6], bw_int_to_fixed(1000)));
+			if (ctx->dc->caps.max_slave_planes) {
+				calcs_output->stutter_entry_wm_ns[3].c_mark =
+					bw_fixed_to_int(bw_mul(data->stutter_entry_watermark[0],
+						bw_int_to_fixed(1000)));
+				calcs_output->stutter_entry_wm_ns[4].c_mark =
+					bw_fixed_to_int(bw_mul(data->stutter_entry_watermark[1],
+						bw_int_to_fixed(1000)));
+			} else {
+				calcs_output->stutter_entry_wm_ns[3].c_mark =
+					bw_fixed_to_int(bw_mul(data->stutter_entry_watermark[7],
+						bw_int_to_fixed(1000)));
+				calcs_output->stutter_entry_wm_ns[4].c_mark =
+					bw_fixed_to_int(bw_mul(data->stutter_entry_watermark[8],
+						bw_int_to_fixed(1000)));
+			}
+			calcs_output->stutter_entry_wm_ns[5].c_mark =
 				bw_fixed_to_int(bw_mul(data->
-					stutter_entry_watermark[8], bw_int_to_fixed(1000)));
-		}
-		calcs_output->stutter_entry_wm_ns[5].c_mark =
-			bw_fixed_to_int(bw_mul(data->
-				stutter_entry_watermark[9], bw_int_to_fixed(1000)));
-
+					stutter_entry_watermark[9], bw_int_to_fixed(1000)));
 			calcs_output->urgent_wm_ns[0].c_mark =
 				bw_fixed_to_int(bw_mul(data->
 					urgent_watermark[4], bw_int_to_fixed(1000)));
diff --git a/drivers/gpu/drm/amd/display/dc/basics/fixpt31_32.c b/drivers/gpu/drm/amd/display/dc/basics/fixpt31_32.c
index 1726bdf89bae..6073cadde76c 100644
--- a/drivers/gpu/drm/amd/display/dc/basics/fixpt31_32.c
+++ b/drivers/gpu/drm/amd/display/dc/basics/fixpt31_32.c
@@ -51,8 +51,6 @@ static inline unsigned long long complete_integer_division_u64(
 {
 	unsigned long long result;
 
-	ASSERT(divisor);
-
 	result = div64_u64_rem(dividend, divisor, remainder);
 
 	return result;
@@ -140,8 +138,6 @@ struct fixed31_32 dc_fixpt_mul(struct fixed31_32 arg1, struct fixed31_32 arg2)
 
 	res.value = arg1_int * arg2_int;
 
-	ASSERT(res.value <= LONG_MAX);
-
 	res.value <<= FIXED31_32_BITS_PER_FRACTIONAL_PART;
 
 	tmp = arg1_int * arg2_fra;
@@ -185,8 +181,6 @@ struct fixed31_32 dc_fixpt_sqr(struct fixed31_32 arg)
 
 	res.value = arg_int * arg_int;
 
-	ASSERT(res.value <= LONG_MAX);
-
 	res.value <<= FIXED31_32_BITS_PER_FRACTIONAL_PART;
 
 	tmp = arg_int * arg_fra;
@@ -217,9 +211,6 @@ struct fixed31_32 dc_fixpt_recip(struct fixed31_32 arg)
 	 * @note
 	 * Good idea to use Newton's method
 	 */
-
-	ASSERT(arg.value);
-
 	return dc_fixpt_from_fraction(
 		dc_fixpt_one.value,
 		arg.value);
@@ -293,7 +284,7 @@ struct fixed31_32 dc_fixpt_cos(struct fixed31_32 arg)
 				dc_fixpt_mul(
 					square,
 					res),
-				n * (n - 1)));
+				(long long)n * (n - 1)));
 
 		n -= 2;
 	} while (n != 0);
@@ -490,3 +481,30 @@ int dc_fixpt_s4d19(struct fixed31_32 arg)
 	else
 		return ux_dy(arg.value, 4, 19);
 }
+
+struct fixed31_32 dc_fixpt_from_ux_dy(unsigned int value,
+	unsigned int integer_bits,
+	unsigned int fractional_bits)
+{
+	struct fixed31_32 fixpt_value = dc_fixpt_zero;
+	struct fixed31_32 fixpt_int_value = dc_fixpt_zero;
+	long long frac_mask = ((long long)1 << (long long)integer_bits) - 1;
+
+	fixpt_value.value = (long long)value << (FIXED31_32_BITS_PER_FRACTIONAL_PART - fractional_bits);
+	frac_mask = frac_mask << fractional_bits;
+	fixpt_int_value.value = value & frac_mask;
+	fixpt_int_value.value <<= (FIXED31_32_BITS_PER_FRACTIONAL_PART - fractional_bits);
+	fixpt_value.value |= fixpt_int_value.value;
+	return fixpt_value;
+}
+
+struct fixed31_32 dc_fixpt_from_int_dy(unsigned int int_value,
+	unsigned int frac_value,
+	unsigned int integer_bits,
+	unsigned int fractional_bits)
+{
+	struct fixed31_32 fixpt_value = dc_fixpt_from_int(int_value);
+
+	fixpt_value.value |= (long long)frac_value << (FIXED31_32_BITS_PER_FRACTIONAL_PART - fractional_bits);
+	return fixpt_value;
+}
diff --git a/drivers/gpu/drm/amd/display/dc/basics/vector.c b/drivers/gpu/drm/amd/display/dc/basics/vector.c
index 706c803c4d3b..b413a672c2c0 100644
--- a/drivers/gpu/drm/amd/display/dc/basics/vector.c
+++ b/drivers/gpu/drm/amd/display/dc/basics/vector.c
@@ -23,8 +23,6 @@
  *
  */
 
-#include <linux/slab.h>
-
 #include "dm_services.h"
 #include "include/vector.h"
 
@@ -52,12 +50,11 @@ bool dal_vector_construct(
 	return true;
 }
 
-static bool dal_vector_presized_costruct(
-	struct vector *vector,
-	struct dc_context *ctx,
-	uint32_t count,
-	void *initial_value,
-	uint32_t struct_size)
+static bool dal_vector_presized_costruct(struct vector *vector,
+					 struct dc_context *ctx,
+					 uint32_t count,
+					 void *initial_value,
+					 uint32_t struct_size)
 {
 	uint32_t i;
 
@@ -173,7 +170,7 @@ bool dal_vector_remove_at_index(
 		memmove(
 			vector->container + (index * vector->struct_size),
 			vector->container + ((index + 1) * vector->struct_size),
-			(vector->count - index - 1) * vector->struct_size);
+			(size_t)(vector->count - index - 1) * vector->struct_size);
 	vector->count -= 1;
 
 	return true;
@@ -222,7 +219,7 @@ bool dal_vector_insert_at(
 		memmove(
 			insert_address + vector->struct_size,
 			insert_address,
-			vector->struct_size * (vector->count - position));
+			(size_t)vector->struct_size * (vector->count - position));
 
 	memmove(
 		insert_address,
@@ -274,7 +271,7 @@ struct vector *dal_vector_clone(
 
 	/* copy vector's data */
 	memmove(vec_cloned->container, vector->container,
-			vec_cloned->struct_size * vec_cloned->capacity);
+			(size_t)vec_cloned->struct_size * vec_cloned->capacity);
 
 	return vec_cloned;
 }
diff --git a/drivers/gpu/drm/amd/display/dc/bios/bios_parser.c b/drivers/gpu/drm/amd/display/dc/bios/bios_parser.c
index 9b8ea6e9a2b9..d1471f34e419 100644
--- a/drivers/gpu/drm/amd/display/dc/bios/bios_parser.c
+++ b/drivers/gpu/drm/amd/display/dc/bios/bios_parser.c
@@ -33,7 +33,6 @@
 #include "include/gpio_service_interface.h"
 #include "include/grph_object_ctrl_defs.h"
 #include "include/bios_parser_interface.h"
-#include "include/i2caux_interface.h"
 #include "include/logger_interface.h"
 
 #include "command_table.h"
@@ -45,8 +44,6 @@
 
 #include "bios_parser_common.h"
 
-#include "dc.h"
-
 #define THREE_PERCENT_OF_10000 300
 
 #define LAST_RECORD_TYPE 0xff
@@ -70,7 +67,9 @@ static ATOM_HPD_INT_RECORD *get_hpd_record(struct bios_parser *bp,
 	ATOM_OBJECT *object);
 static struct device_id device_type_from_device_id(uint16_t device_id);
 static uint32_t signal_to_ss_id(enum as_signal_type signal);
-static uint32_t get_support_mask_for_device_id(struct device_id device_id);
+static uint32_t get_support_mask_for_device_id(
+	enum dal_device_type device_type,
+	uint32_t enum_id);
 static ATOM_ENCODER_CAP_RECORD_V2 *get_encoder_cap_record(
 	struct bios_parser *bp,
 	ATOM_OBJECT *object);
@@ -97,7 +96,7 @@ struct dc_bios *bios_parser_create(
 	struct bp_init_data *init,
 	enum dce_version dce_version)
 {
-	struct bios_parser *bp = NULL;
+	struct bios_parser *bp;
 
 	bp = kzalloc(sizeof(struct bios_parser), GFP_KERNEL);
 	if (!bp)
@@ -138,7 +137,9 @@ static uint8_t get_number_of_objects(struct bios_parser *bp, uint32_t offset)
 
 	uint32_t object_table_offset = bp->object_info_tbl_offset + offset;
 
-	table = GET_IMAGE(ATOM_OBJECT_TABLE, object_table_offset);
+	table = ((ATOM_OBJECT_TABLE *) bios_get_image(&bp->base,
+				object_table_offset,
+				struct_size(table, asObjects, 1)));
 
 	if (!table)
 		return 0;
@@ -166,19 +167,17 @@ static struct graphics_object_id bios_parser_get_connector_id(
 	uint32_t connector_table_offset = bp->object_info_tbl_offset
 		+ le16_to_cpu(bp->object_info_tbl.v1_1->usConnectorObjectTableOffset);
 
-	ATOM_OBJECT_TABLE *tbl =
-		GET_IMAGE(ATOM_OBJECT_TABLE, connector_table_offset);
+	ATOM_OBJECT_TABLE *tbl = ((ATOM_OBJECT_TABLE *) bios_get_image(&bp->base,
+				connector_table_offset,
+				struct_size(tbl, asObjects, 1)));
 
 	if (!tbl) {
 		dm_error("Can't get connector table from atom bios.\n");
 		return object_id;
 	}
 
-	if (tbl->ucNumberOfObjects <= i) {
-		dm_error("Can't find connector id %d in connector table of size %d.\n",
-			 i, tbl->ucNumberOfObjects);
+	if (tbl->ucNumberOfObjects <= i)
 		return object_id;
-	}
 
 	id = le16_to_cpu(tbl->asObjects[i].usObjectID);
 	object_id = object_id_from_bios_object_id(id);
@@ -444,6 +443,7 @@ static enum bp_result get_firmware_info_v1_4(
 		le32_to_cpu(firmware_info->ulMinPixelClockPLL_Output) * 10;
 	info->pll_info.max_output_pxl_clk_pll_frequency =
 		le32_to_cpu(firmware_info->ulMaxPixelClockPLL_Output) * 10;
+	info->max_pixel_clock = le16_to_cpu(firmware_info->usMaxPixelClock) * 10;
 
 	if (firmware_info->usFirmwareCapability.sbfAccess.MemoryClockSS_Support)
 		/* Since there is no information on the SS, report conservative
@@ -500,6 +500,7 @@ static enum bp_result get_firmware_info_v2_1(
 	info->external_clock_source_frequency_for_dp =
 		le16_to_cpu(firmwareInfo->usUniphyDPModeExtClkFreq) * 10;
 	info->min_allowed_bl_level = firmwareInfo->ucMinAllowedBL_Level;
+	info->max_pixel_clock = le16_to_cpu(firmwareInfo->usMaxPixelClock) * 10;
 
 	/* There should be only one entry in the SS info table for Memory Clock
 	 */
@@ -662,8 +663,12 @@ static enum bp_result get_ss_info_v3_1(
 	if (!DATA_TABLES(ASIC_InternalSS_Info))
 		return BP_RESULT_UNSUPPORTED;
 
-	ss_table_header_include = GET_IMAGE(ATOM_ASIC_INTERNAL_SS_INFO_V3,
-		DATA_TABLES(ASIC_InternalSS_Info));
+	ss_table_header_include = ((ATOM_ASIC_INTERNAL_SS_INFO_V3 *) bios_get_image(&bp->base,
+				DATA_TABLES(ASIC_InternalSS_Info),
+				struct_size(ss_table_header_include, asSpreadSpectrum, 1)));
+	if (!ss_table_header_include)
+		return BP_RESULT_UNSUPPORTED;
+
 	table_size =
 		(le16_to_cpu(ss_table_header_include->sHeader.usStructureSize)
 				- sizeof(ATOM_COMMON_TABLE_HEADER))
@@ -735,18 +740,94 @@ static enum bp_result bios_parser_transmitter_control(
 	return bp->cmd_tbl.transmitter_control(bp, cntl);
 }
 
+static enum bp_result bios_parser_select_crtc_source(
+	struct dc_bios *dcb,
+	struct bp_crtc_source_select *bp_params)
+{
+	struct bios_parser *bp = BP_FROM_DCB(dcb);
+
+	if (!bp->cmd_tbl.select_crtc_source)
+		return BP_RESULT_FAILURE;
+
+	return bp->cmd_tbl.select_crtc_source(bp, bp_params);
+}
+
 static enum bp_result bios_parser_encoder_control(
 	struct dc_bios *dcb,
 	struct bp_encoder_control *cntl)
 {
 	struct bios_parser *bp = BP_FROM_DCB(dcb);
 
+	if (cntl->engine_id == ENGINE_ID_DACA) {
+		if (!bp->cmd_tbl.dac1_encoder_control)
+			return BP_RESULT_FAILURE;
+
+		return bp->cmd_tbl.dac1_encoder_control(
+			bp, cntl->action == ENCODER_CONTROL_ENABLE,
+			cntl->pixel_clock, ATOM_DAC1_PS2);
+	} else if (cntl->engine_id == ENGINE_ID_DACB) {
+		if (!bp->cmd_tbl.dac2_encoder_control)
+			return BP_RESULT_FAILURE;
+
+		return bp->cmd_tbl.dac2_encoder_control(
+			bp, cntl->action == ENCODER_CONTROL_ENABLE,
+			cntl->pixel_clock, ATOM_DAC1_PS2);
+	}
+
 	if (!bp->cmd_tbl.dig_encoder_control)
 		return BP_RESULT_FAILURE;
 
 	return bp->cmd_tbl.dig_encoder_control(bp, cntl);
 }
 
+static enum bp_result bios_parser_dac_load_detection(
+	struct dc_bios *dcb,
+	enum engine_id engine_id,
+	enum dal_device_type device_type,
+	uint32_t enum_id)
+{
+	struct bios_parser *bp = BP_FROM_DCB(dcb);
+	struct dc_context *ctx = dcb->ctx;
+	struct bp_load_detection_parameters bp_params = {0};
+	enum bp_result bp_result;
+	uint32_t bios_0_scratch;
+	uint32_t device_id_mask = 0;
+
+	bp_params.engine_id = engine_id;
+	bp_params.device_id = get_support_mask_for_device_id(device_type, enum_id);
+
+	if (engine_id != ENGINE_ID_DACA &&
+	    engine_id != ENGINE_ID_DACB)
+		return BP_RESULT_UNSUPPORTED;
+
+	if (!bp->cmd_tbl.dac_load_detection)
+		return BP_RESULT_UNSUPPORTED;
+
+	if (bp_params.device_id == ATOM_DEVICE_CRT1_SUPPORT)
+		device_id_mask = ATOM_S0_CRT1_MASK;
+	else if (bp_params.device_id == ATOM_DEVICE_CRT2_SUPPORT)
+		device_id_mask = ATOM_S0_CRT2_MASK;
+	else
+		return BP_RESULT_UNSUPPORTED;
+
+	/* BIOS will write the detected devices to BIOS_SCRATCH_0, clear corresponding bit */
+	bios_0_scratch = dm_read_reg(ctx, bp->base.regs->BIOS_SCRATCH_0);
+	bios_0_scratch &= ~device_id_mask;
+	dm_write_reg(ctx, bp->base.regs->BIOS_SCRATCH_0, bios_0_scratch);
+
+	bp_result = bp->cmd_tbl.dac_load_detection(bp, &bp_params);
+
+	if (bp_result != BP_RESULT_OK)
+		return bp_result;
+
+	bios_0_scratch = dm_read_reg(ctx, bp->base.regs->BIOS_SCRATCH_0);
+
+	if (bios_0_scratch & device_id_mask)
+		return BP_RESULT_OK;
+
+	return BP_RESULT_FAILURE;
+}
+
 static enum bp_result bios_parser_adjust_pixel_clock(
 	struct dc_bios *dcb,
 	struct bp_adjust_pixel_clock_parameters *bp_params)
@@ -857,7 +938,7 @@ static bool bios_parser_is_device_id_supported(
 {
 	struct bios_parser *bp = BP_FROM_DCB(dcb);
 
-	uint32_t mask = get_support_mask_for_device_id(id);
+	uint32_t mask = get_support_mask_for_device_id(id.device_type, id.enum_id);
 
 	return (le16_to_cpu(bp->object_info_tbl.v1_1->usDeviceSupport) & mask) != 0;
 }
@@ -1029,8 +1110,12 @@ static enum bp_result get_ss_info_from_internal_ss_info_tbl_V2_1(
 	if (!DATA_TABLES(ASIC_InternalSS_Info))
 		return result;
 
-	header = GET_IMAGE(ATOM_ASIC_INTERNAL_SS_INFO_V2,
-		DATA_TABLES(ASIC_InternalSS_Info));
+	header = ((ATOM_ASIC_INTERNAL_SS_INFO_V2 *) bios_get_image(
+				&bp->base,
+				DATA_TABLES(ASIC_InternalSS_Info),
+				struct_size(header, asSpreadSpectrum, 1)));
+	if (!header)
+		return result;
 
 	memset(info, 0, sizeof(struct spread_spectrum_info));
 
@@ -1104,6 +1189,8 @@ static enum bp_result get_ss_info_from_ss_info_table(
 	get_atom_data_table_revision(header, &revision);
 
 	tbl = GET_IMAGE(ATOM_SPREAD_SPECTRUM_INFO, DATA_TABLES(SS_Info));
+	if (!tbl)
+		return result;
 
 	if (1 != revision.major || 2 > revision.minor)
 		return result;
@@ -1631,6 +1718,8 @@ static uint32_t get_ss_entry_number_from_ss_info_tbl(
 
 	tbl = GET_IMAGE(ATOM_SPREAD_SPECTRUM_INFO,
 			DATA_TABLES(SS_Info));
+	if (!tbl)
+		return number;
 
 	if (1 != revision.major || 2 > revision.minor)
 		return number;
@@ -1709,8 +1798,12 @@ static uint32_t get_ss_entry_number_from_internal_ss_info_tbl_v2_1(
 	if (!DATA_TABLES(ASIC_InternalSS_Info))
 		return 0;
 
-	header_include = GET_IMAGE(ATOM_ASIC_INTERNAL_SS_INFO_V2,
-			DATA_TABLES(ASIC_InternalSS_Info));
+	header_include = ((ATOM_ASIC_INTERNAL_SS_INFO_V2 *) bios_get_image(
+				&bp->base,
+				DATA_TABLES(ASIC_InternalSS_Info),
+				struct_size(header_include, asSpreadSpectrum, 1)));
+	if (!header_include)
+		return 0;
 
 	size = (le16_to_cpu(header_include->sHeader.usStructureSize)
 			- sizeof(ATOM_COMMON_TABLE_HEADER))
@@ -1724,6 +1817,7 @@ static uint32_t get_ss_entry_number_from_internal_ss_info_tbl_v2_1(
 
 	return 0;
 }
+
 /**
  * get_ss_entry_number_from_internal_ss_info_tbl_V3_1
  * Get Number of SpreadSpectrum Entry from the ASIC_InternalSS_Info table of
@@ -1746,8 +1840,12 @@ static uint32_t get_ss_entry_number_from_internal_ss_info_tbl_V3_1(
 	if (!DATA_TABLES(ASIC_InternalSS_Info))
 		return number;
 
-	header_include = GET_IMAGE(ATOM_ASIC_INTERNAL_SS_INFO_V3,
-			DATA_TABLES(ASIC_InternalSS_Info));
+	header_include = ((ATOM_ASIC_INTERNAL_SS_INFO_V3 *) bios_get_image(&bp->base,
+				DATA_TABLES(ASIC_InternalSS_Info),
+				struct_size(header_include, asSpreadSpectrum, 1)));
+	if (!header_include)
+		return number;
+
 	size = (le16_to_cpu(header_include->sHeader.usStructureSize) -
 			sizeof(ATOM_COMMON_TABLE_HEADER)) /
 					sizeof(ATOM_ASIC_SS_ASSIGNMENT_V3);
@@ -1789,11 +1887,13 @@ static enum bp_result bios_parser_get_gpio_pin_info(
 	if (!DATA_TABLES(GPIO_Pin_LUT))
 		return BP_RESULT_BADBIOSTABLE;
 
-	header = GET_IMAGE(ATOM_GPIO_PIN_LUT, DATA_TABLES(GPIO_Pin_LUT));
+	header = ((ATOM_GPIO_PIN_LUT *) bios_get_image(&bp->base,
+				DATA_TABLES(GPIO_Pin_LUT),
+				struct_size(header, asGPIO_Pin, 1)));
 	if (!header)
 		return BP_RESULT_BADBIOSTABLE;
 
-	if (sizeof(ATOM_COMMON_TABLE_HEADER) + sizeof(ATOM_GPIO_PIN_LUT)
+	if (sizeof(ATOM_COMMON_TABLE_HEADER) + struct_size(header, asGPIO_Pin, 1)
 			> le16_to_cpu(header->sHeader.usStructureSize))
 		return BP_RESULT_BADBIOSTABLE;
 
@@ -1978,7 +2078,8 @@ static ATOM_OBJECT *get_bios_object(struct bios_parser *bp,
 
 	offset += bp->object_info_tbl_offset;
 
-	tbl = GET_IMAGE(ATOM_OBJECT_TABLE, offset);
+	tbl = ((ATOM_OBJECT_TABLE *) bios_get_image(&bp->base, offset,
+				struct_size(tbl, asObjects, 1)));
 	if (!tbl)
 		return NULL;
 
@@ -2128,11 +2229,10 @@ static uint32_t signal_to_ss_id(enum as_signal_type signal)
 	return clk_id_ss;
 }
 
-static uint32_t get_support_mask_for_device_id(struct device_id device_id)
+static uint32_t get_support_mask_for_device_id(
+	enum dal_device_type device_type,
+	uint32_t enum_id)
 {
-	enum dal_device_type device_type = device_id.device_type;
-	uint32_t enum_id = device_id.enum_id;
-
 	switch (device_type) {
 	case DEVICE_TYPE_LCD:
 		switch (enum_id) {
@@ -2360,10 +2460,10 @@ static enum bp_result get_integrated_info_v8(
 }
 
 /*
- * get_integrated_info_v8
+ * get_integrated_info_v9
  *
  * @brief
- * Get V8 integrated BIOS information
+ * Get V9 integrated BIOS information
  *
  * @param
  * bios_parser *bp - [in]BIOS parser handler to get master data table
@@ -2541,8 +2641,8 @@ static enum bp_result construct_integrated_info(
 
 	/* Sort voltage table from low to high*/
 	if (result == BP_RESULT_OK) {
-		uint32_t i;
-		uint32_t j;
+		int32_t i;
+		int32_t j;
 
 		for (i = 1; i < NUMBER_OF_DISP_CLK_VOLTAGE; ++i) {
 			for (j = i; j > 0; --j) {
@@ -2565,7 +2665,7 @@ static struct integrated_info *bios_parser_create_integrated_info(
 	struct dc_bios *dcb)
 {
 	struct bios_parser *bp = BP_FROM_DCB(dcb);
-	struct integrated_info *info = NULL;
+	struct integrated_info *info;
 
 	info = kzalloc(sizeof(struct integrated_info), GFP_KERNEL);
 
@@ -2582,11 +2682,10 @@ static struct integrated_info *bios_parser_create_integrated_info(
 	return NULL;
 }
 
-static enum bp_result update_slot_layout_info(
-	struct dc_bios *dcb,
-	unsigned int i,
-	struct slot_layout_info *slot_layout_info,
-	unsigned int record_offset)
+static enum bp_result update_slot_layout_info(struct dc_bios *dcb,
+					      unsigned int i,
+					      struct slot_layout_info *slot_layout_info,
+					      unsigned int record_offset)
 {
 	unsigned int j;
 	struct bios_parser *bp;
@@ -2600,8 +2699,7 @@ static enum bp_result update_slot_layout_info(
 
 	for (;;) {
 
-		record_header = (ATOM_COMMON_RECORD_HEADER *)
-			GET_IMAGE(ATOM_COMMON_RECORD_HEADER, record_offset);
+		record_header = GET_IMAGE(ATOM_COMMON_RECORD_HEADER, record_offset);
 		if (record_header == NULL) {
 			result = BP_RESULT_BADBIOSTABLE;
 			break;
@@ -2615,7 +2713,7 @@ static enum bp_result update_slot_layout_info(
 
 		if (record_header->ucRecordType ==
 			ATOM_BRACKET_LAYOUT_RECORD_TYPE &&
-			sizeof(ATOM_BRACKET_LAYOUT_RECORD)
+			struct_size(record, asConnInfo, 1)
 			<= record_header->ucRecordSize) {
 			record = (ATOM_BRACKET_LAYOUT_RECORD *)
 				(record_header);
@@ -2686,10 +2784,9 @@ static enum bp_result update_slot_layout_info(
 }
 
 
-static enum bp_result get_bracket_layout_record(
-	struct dc_bios *dcb,
-	unsigned int bracket_layout_id,
-	struct slot_layout_info *slot_layout_info)
+static enum bp_result get_bracket_layout_record(struct dc_bios *dcb,
+						unsigned int bracket_layout_id,
+						struct slot_layout_info *slot_layout_info)
 {
 	unsigned int i;
 	unsigned int record_offset;
@@ -2709,8 +2806,9 @@ static enum bp_result get_bracket_layout_record(
 
 	genericTableOffset = bp->object_info_tbl_offset +
 		bp->object_info_tbl.v1_3->usMiscObjectTableOffset;
-	object_table = (ATOM_OBJECT_TABLE *)
-		GET_IMAGE(ATOM_OBJECT_TABLE, genericTableOffset);
+	object_table = ((ATOM_OBJECT_TABLE *) bios_get_image(&bp->base,
+				genericTableOffset,
+				struct_size(object_table, asObjects, 1)));
 	if (!object_table)
 		return BP_RESULT_FAILURE;
 
@@ -2737,6 +2835,7 @@ static enum bp_result bios_get_board_layout_info(
 	struct board_layout_info *board_layout_info)
 {
 	unsigned int i;
+	struct bios_parser *bp;
 	enum bp_result record_result;
 
 	const unsigned int slot_index_to_vbios_id[MAX_BOARD_SLOTS] = {
@@ -2745,6 +2844,8 @@ static enum bp_result bios_get_board_layout_info(
 		0, 0
 	};
 
+	bp = BP_FROM_DCB(dcb);
+
 	if (board_layout_info == NULL) {
 		DC_LOG_DETECTION_EDID_PARSER("Invalid board_layout_info\n");
 		return BP_RESULT_BADINPUT;
@@ -2807,8 +2908,12 @@ static const struct dc_vbios_funcs vbios_funcs = {
 	.is_device_id_supported = bios_parser_is_device_id_supported,
 
 	/* COMMANDS */
+	.select_crtc_source = bios_parser_select_crtc_source,
+
 	.encoder_control = bios_parser_encoder_control,
 
+	.dac_load_detection = bios_parser_dac_load_detection,
+
 	.transmitter_control = bios_parser_transmitter_control,
 
 	.enable_crtc = bios_parser_enable_crtc,
diff --git a/drivers/gpu/drm/amd/display/dc/bios/bios_parser2.c b/drivers/gpu/drm/amd/display/dc/bios/bios_parser2.c
index a4bef4364afd..550a9f1d03f8 100644
--- a/drivers/gpu/drm/amd/display/dc/bios/bios_parser2.c
+++ b/drivers/gpu/drm/amd/display/dc/bios/bios_parser2.c
@@ -23,9 +23,8 @@
  *
  */
 
-#include <linux/slab.h>
-
 #include "dm_services.h"
+#include "core_types.h"
 
 #include "ObjectID.h"
 #include "atomfirmware.h"
@@ -33,7 +32,6 @@
 #include "dc_bios_types.h"
 #include "include/grph_object_ctrl_defs.h"
 #include "include/bios_parser_interface.h"
-#include "include/i2caux_interface.h"
 #include "include/logger_interface.h"
 
 #include "command_table2.h"
@@ -46,38 +44,12 @@
 
 #include "bios_parser_common.h"
 
-/* Temporarily add in defines until ObjectID.h patch is updated in a few days */
-#ifndef GENERIC_OBJECT_ID_BRACKET_LAYOUT
-#define GENERIC_OBJECT_ID_BRACKET_LAYOUT          0x05
-#endif /* GENERIC_OBJECT_ID_BRACKET_LAYOUT */
-
-#ifndef GENERICOBJECT_BRACKET_LAYOUT_ENUM_ID1
-#define GENERICOBJECT_BRACKET_LAYOUT_ENUM_ID1	\
-	(GRAPH_OBJECT_TYPE_GENERIC << OBJECT_TYPE_SHIFT |\
-	GRAPH_OBJECT_ENUM_ID1 << ENUM_ID_SHIFT |\
-	GENERIC_OBJECT_ID_BRACKET_LAYOUT << OBJECT_ID_SHIFT)
-#endif /* GENERICOBJECT_BRACKET_LAYOUT_ENUM_ID1 */
-
-#ifndef GENERICOBJECT_BRACKET_LAYOUT_ENUM_ID2
-#define GENERICOBJECT_BRACKET_LAYOUT_ENUM_ID2	\
-	(GRAPH_OBJECT_TYPE_GENERIC << OBJECT_TYPE_SHIFT |\
-	GRAPH_OBJECT_ENUM_ID2 << ENUM_ID_SHIFT |\
-	GENERIC_OBJECT_ID_BRACKET_LAYOUT << OBJECT_ID_SHIFT)
-#endif /* GENERICOBJECT_BRACKET_LAYOUT_ENUM_ID2 */
-
 #define DC_LOGGER \
 	bp->base.ctx->logger
 
 #define LAST_RECORD_TYPE 0xff
 #define SMU9_SYSPLL0_ID  0
 
-struct i2c_id_config_access {
-	uint8_t bfI2C_LineMux:4;
-	uint8_t bfHW_EngineID:3;
-	uint8_t bfHW_Capable:1;
-	uint8_t ucAccess;
-};
-
 static enum bp_result get_gpio_i2c_info(struct bios_parser *bp,
 	struct atom_i2c_record *record,
 	struct graphics_object_i2c_info *info);
@@ -103,6 +75,10 @@ static enum bp_result get_firmware_info_v3_4(
 	struct bios_parser *bp,
 	struct dc_firmware_info *info);
 
+static enum bp_result get_firmware_info_v3_5(
+	struct bios_parser *bp,
+	struct dc_firmware_info *info);
+
 static struct atom_hpd_int_record *get_hpd_record(struct bios_parser *bp,
 		struct atom_display_object_path_v2 *object);
 
@@ -165,9 +141,21 @@ static uint8_t bios_parser_get_connectors_number(struct dc_bios *dcb)
 	unsigned int count = 0;
 	unsigned int i;
 
-	for (i = 0; i < bp->object_info_tbl.v1_4->number_of_path; i++) {
-		if (bp->object_info_tbl.v1_4->display_path[i].encoderobjid != 0)
-			count++;
+	switch (bp->object_info_tbl.revision.minor) {
+	default:
+	case 4:
+		for (i = 0; i < bp->object_info_tbl.v1_4->number_of_path; i++)
+			if (bp->object_info_tbl.v1_4->display_path[i].encoderobjid != 0)
+				count++;
+
+		break;
+
+	case 5:
+		for (i = 0; i < bp->object_info_tbl.v1_5->number_of_path; i++)
+			if (bp->object_info_tbl.v1_5->display_path[i].encoderobjid != 0)
+				count++;
+
+		break;
 	}
 	return count;
 }
@@ -182,16 +170,34 @@ static struct graphics_object_id bios_parser_get_connector_id(
 	struct object_info_table *tbl = &bp->object_info_tbl;
 	struct display_object_info_table_v1_4 *v1_4 = tbl->v1_4;
 
-	if (v1_4->number_of_path > i) {
-		/* If display_objid is generic object id,  the encoderObj
-		 * /extencoderobjId should be 0
-		 */
-		if (v1_4->display_path[i].encoderobjid != 0 &&
-				v1_4->display_path[i].display_objid != 0)
-			object_id = object_id_from_bios_object_id(
+	struct display_object_info_table_v1_5 *v1_5 = tbl->v1_5;
+
+	switch (bp->object_info_tbl.revision.minor) {
+	default:
+	case 4:
+		if (v1_4->number_of_path > i) {
+			/* If display_objid is generic object id,  the encoderObj
+			 * /extencoderobjId should be 0
+			 */
+			if (v1_4->display_path[i].encoderobjid != 0 &&
+			    v1_4->display_path[i].display_objid != 0)
+				object_id = object_id_from_bios_object_id(
 					v1_4->display_path[i].display_objid);
-	}
+		}
+		break;
 
+	case 5:
+		if (v1_5->number_of_path > i) {
+			/* If display_objid is generic object id,  the encoderObjId
+		 * should be 0
+		 */
+			if (v1_5->display_path[i].encoderobjid != 0 &&
+			    v1_5->display_path[i].display_objid != 0)
+				object_id = object_id_from_bios_object_id(
+					v1_5->display_path[i].display_objid);
+		}
+		break;
+	}
 	return object_id;
 }
 
@@ -201,8 +207,8 @@ static enum bp_result bios_parser_get_src_obj(struct dc_bios *dcb,
 {
 	struct bios_parser *bp = BP_FROM_DCB(dcb);
 	unsigned int i;
-	enum bp_result  bp_result = BP_RESULT_BADINPUT;
-	struct graphics_object_id obj_id = {0};
+	enum bp_result bp_result = BP_RESULT_BADINPUT;
+	struct graphics_object_id obj_id = { 0 };
 	struct object_info_table *tbl = &bp->object_info_tbl;
 
 	if (!src_object_id)
@@ -217,37 +223,84 @@ static enum bp_result bios_parser_get_src_obj(struct dc_bios *dcb,
 		 * If found in for loop, should break.
 		 * DAL2 implementation may be changed too
 		 */
-		for (i = 0; i < tbl->v1_4->number_of_path; i++) {
-			obj_id = object_id_from_bios_object_id(
-			tbl->v1_4->display_path[i].encoderobjid);
-			if (object_id.type == obj_id.type &&
-					object_id.id == obj_id.id &&
-						object_id.enum_id ==
-							obj_id.enum_id) {
-				*src_object_id =
-				object_id_from_bios_object_id(0x1100);
-				/* break; */
+		switch (bp->object_info_tbl.revision.minor) {
+		default:
+		case 4:
+			for (i = 0; i < tbl->v1_4->number_of_path; i++) {
+				obj_id = object_id_from_bios_object_id(
+					tbl->v1_4->display_path[i].encoderobjid);
+				if (object_id.type == obj_id.type &&
+				    object_id.id == obj_id.id &&
+				    object_id.enum_id == obj_id.enum_id) {
+					*src_object_id =
+						object_id_from_bios_object_id(
+							0x1100);
+					/* break; */
+				}
 			}
+			bp_result = BP_RESULT_OK;
+			break;
+
+		case 5:
+			for (i = 0; i < tbl->v1_5->number_of_path; i++) {
+				obj_id = object_id_from_bios_object_id(
+					tbl->v1_5->display_path[i].encoderobjid);
+				if (object_id.type == obj_id.type &&
+				    object_id.id == obj_id.id &&
+				    object_id.enum_id == obj_id.enum_id) {
+					*src_object_id =
+						object_id_from_bios_object_id(
+							0x1100);
+					/* break; */
+				}
+			}
+			bp_result = BP_RESULT_OK;
+			break;
 		}
-		bp_result = BP_RESULT_OK;
 		break;
 	case OBJECT_TYPE_CONNECTOR:
-		for (i = 0; i < tbl->v1_4->number_of_path; i++) {
-			obj_id = object_id_from_bios_object_id(
-				tbl->v1_4->display_path[i].display_objid);
-
-			if (object_id.type == obj_id.type &&
-				object_id.id == obj_id.id &&
-					object_id.enum_id == obj_id.enum_id) {
-				*src_object_id =
-				object_id_from_bios_object_id(
-				tbl->v1_4->display_path[i].encoderobjid);
-				/* break; */
+		switch (bp->object_info_tbl.revision.minor) {
+		default:
+		case 4:
+			for (i = 0; i < tbl->v1_4->number_of_path; i++) {
+				obj_id = object_id_from_bios_object_id(
+					tbl->v1_4->display_path[i]
+						.display_objid);
+
+				if (object_id.type == obj_id.type &&
+				    object_id.id == obj_id.id &&
+				    object_id.enum_id == obj_id.enum_id) {
+					*src_object_id =
+						object_id_from_bios_object_id(
+							tbl->v1_4
+								->display_path[i]
+								.encoderobjid);
+					/* break; */
+				}
 			}
+			bp_result = BP_RESULT_OK;
+			break;
 		}
 		bp_result = BP_RESULT_OK;
 		break;
+		case 5:
+			for (i = 0; i < tbl->v1_5->number_of_path; i++) {
+				obj_id = object_id_from_bios_object_id(
+								       tbl->v1_5->display_path[i].display_objid);
+
+				if (object_id.type == obj_id.type &&
+				    object_id.id == obj_id.id &&
+				    object_id.enum_id == obj_id.enum_id) {
+					*src_object_id = object_id_from_bios_object_id(
+										       tbl->v1_5->display_path[i].encoderobjid);
+					/* break; */
+				}
+			}
+		bp_result = BP_RESULT_OK;
+		break;
+
 	default:
+		bp_result = BP_RESULT_OK;
 		break;
 	}
 
@@ -290,12 +343,54 @@ static struct atom_display_object_path_v2 *get_bios_object(
 	}
 }
 
+/* from graphics_object_id, find display path which includes the object_id */
+static struct atom_display_object_path_v3 *get_bios_object_from_path_v3(struct bios_parser *bp,
+									struct graphics_object_id id)
+{
+	unsigned int i;
+	struct graphics_object_id obj_id = {0};
+
+	switch (id.type) {
+	case OBJECT_TYPE_ENCODER:
+		for (i = 0; i < bp->object_info_tbl.v1_5->number_of_path; i++) {
+			obj_id = object_id_from_bios_object_id(
+					bp->object_info_tbl.v1_5->display_path[i].encoderobjid);
+			if (id.type == obj_id.type && id.id == obj_id.id
+					&& id.enum_id == obj_id.enum_id)
+				return &bp->object_info_tbl.v1_5->display_path[i];
+		}
+	break;
+
+	case OBJECT_TYPE_CONNECTOR:
+	case OBJECT_TYPE_GENERIC:
+		/* Both Generic and Connector Object ID
+		 * will be stored on display_objid
+		 */
+		for (i = 0; i < bp->object_info_tbl.v1_5->number_of_path; i++) {
+			obj_id = object_id_from_bios_object_id(
+					bp->object_info_tbl.v1_5->display_path[i].display_objid);
+			if (id.type == obj_id.type && id.id == obj_id.id
+					&& id.enum_id == obj_id.enum_id)
+				return &bp->object_info_tbl.v1_5->display_path[i];
+		}
+	break;
+
+	default:
+		return NULL;
+	}
+
+	return NULL;
+}
+
 static enum bp_result bios_parser_get_i2c_info(struct dc_bios *dcb,
 	struct graphics_object_id id,
 	struct graphics_object_i2c_info *info)
 {
 	uint32_t offset;
 	struct atom_display_object_path_v2 *object;
+
+	struct atom_display_object_path_v3 *object_path_v3;
+
 	struct atom_common_record_header *header;
 	struct atom_i2c_record *record;
 	struct atom_i2c_record dummy_record = {0};
@@ -313,12 +408,25 @@ static enum bp_result bios_parser_get_i2c_info(struct dc_bios *dcb,
 			return BP_RESULT_NORECORD;
 	}
 
-	object = get_bios_object(bp, id);
+	switch (bp->object_info_tbl.revision.minor) {
+	case 4:
+	default:
+		object = get_bios_object(bp, id);
 
-	if (!object)
-		return BP_RESULT_BADINPUT;
+		if (!object)
+			return BP_RESULT_BADINPUT;
 
-	offset = object->disp_recordoffset + bp->object_info_tbl_offset;
+		offset = object->disp_recordoffset + bp->object_info_tbl_offset;
+		break;
+	case 5:
+		object_path_v3 = get_bios_object_from_path_v3(bp, id);
+
+		if (!object_path_v3)
+			return BP_RESULT_BADINPUT;
+
+		offset = object_path_v3->disp_recordoffset + bp->object_info_tbl_offset;
+		break;
+	}
 
 	for (;;) {
 		header = GET_IMAGE(struct atom_common_record_header, offset);
@@ -356,6 +464,7 @@ static enum bp_result get_gpio_i2c_info(
 	uint32_t count = 0;
 	unsigned int table_index = 0;
 	bool find_valid = false;
+	struct atom_gpio_pin_assignment *pin;
 
 	if (!info)
 		return BP_RESULT_BADINPUT;
@@ -383,20 +492,17 @@ static enum bp_result get_gpio_i2c_info(
 			- sizeof(struct atom_common_table_header))
 				/ sizeof(struct atom_gpio_pin_assignment);
 
+	pin = (struct atom_gpio_pin_assignment *) header->gpio_pin;
+
 	for (table_index = 0; table_index < count; table_index++) {
-		if (((record->i2c_id & I2C_HW_CAP) == (
-		header->gpio_pin[table_index].gpio_id &
-						I2C_HW_CAP)) &&
-		((record->i2c_id & I2C_HW_ENGINE_ID_MASK)  ==
-		(header->gpio_pin[table_index].gpio_id &
-					I2C_HW_ENGINE_ID_MASK)) &&
-		((record->i2c_id & I2C_HW_LANE_MUX) ==
-		(header->gpio_pin[table_index].gpio_id &
-						I2C_HW_LANE_MUX))) {
+		if (((record->i2c_id & I2C_HW_CAP) 				== (pin->gpio_id & I2C_HW_CAP)) &&
+		    ((record->i2c_id & I2C_HW_ENGINE_ID_MASK)	== (pin->gpio_id & I2C_HW_ENGINE_ID_MASK)) &&
+		    ((record->i2c_id & I2C_HW_LANE_MUX) 		== (pin->gpio_id & I2C_HW_LANE_MUX))) {
 			/* still valid */
 			find_valid = true;
 			break;
 		}
+		pin = (struct atom_gpio_pin_assignment *)((uint8_t *)pin + sizeof(struct atom_gpio_pin_assignment));
 	}
 
 	/* If we don't find the entry that we are looking for then
@@ -412,15 +518,46 @@ static enum bp_result get_gpio_i2c_info(
 	info->i2c_slave_address = record->i2c_slave_addr;
 
 	/* TODO: check how to get register offset for en, Y, etc. */
-	info->gpio_info.clk_a_register_index =
-			le16_to_cpu(
-			header->gpio_pin[table_index].data_a_reg_index);
-	info->gpio_info.clk_a_shift =
-			header->gpio_pin[table_index].gpio_bitshift;
+	info->gpio_info.clk_a_register_index = le16_to_cpu(pin->data_a_reg_index);
+	info->gpio_info.clk_a_shift = pin->gpio_bitshift;
 
 	return BP_RESULT_OK;
 }
 
+static struct atom_hpd_int_record *get_hpd_record_for_path_v3(struct bios_parser *bp,
+							      struct atom_display_object_path_v3 *object)
+{
+	struct atom_common_record_header *header;
+	uint32_t offset;
+
+	if (!object) {
+		BREAK_TO_DEBUGGER(); /* Invalid object */
+		return NULL;
+	}
+
+	offset = object->disp_recordoffset + bp->object_info_tbl_offset;
+
+	for (;;) {
+		header = GET_IMAGE(struct atom_common_record_header, offset);
+
+		if (!header)
+			return NULL;
+
+		if (header->record_type == ATOM_RECORD_END_TYPE ||
+			!header->record_size)
+			break;
+
+		if (header->record_type == ATOM_HPD_INT_RECORD_TYPE
+			&& sizeof(struct atom_hpd_int_record) <=
+							header->record_size)
+			return (struct atom_hpd_int_record *) header;
+
+		offset += header->record_size;
+	}
+
+	return NULL;
+}
+
 static enum bp_result bios_parser_get_hpd_info(
 	struct dc_bios *dcb,
 	struct graphics_object_id id,
@@ -428,17 +565,31 @@ static enum bp_result bios_parser_get_hpd_info(
 {
 	struct bios_parser *bp = BP_FROM_DCB(dcb);
 	struct atom_display_object_path_v2 *object;
+	struct atom_display_object_path_v3 *object_path_v3;
 	struct atom_hpd_int_record *record = NULL;
 
 	if (!info)
 		return BP_RESULT_BADINPUT;
 
-	object = get_bios_object(bp, id);
+	switch (bp->object_info_tbl.revision.minor) {
+	case 4:
+	default:
+		object = get_bios_object(bp, id);
+
+		if (!object)
+			return BP_RESULT_BADINPUT;
 
-	if (!object)
-		return BP_RESULT_BADINPUT;
+		record = get_hpd_record(bp, object);
+		break;
+	case 5:
+		object_path_v3 = get_bios_object_from_path_v3(bp, id);
 
-	record = get_hpd_record(bp, object);
+		if (!object_path_v3)
+			return BP_RESULT_BADINPUT;
+
+		record = get_hpd_record_for_path_v3(bp, object_path_v3);
+		break;
+	}
 
 	if (record != NULL) {
 		info->hpd_int_gpio_uid = record->pin_id;
@@ -526,25 +677,9 @@ static enum bp_result bios_parser_get_gpio_pin_info(
 		return BP_RESULT_UNSUPPORTED;
 
 	/* Temporary hard code gpio pin info */
-#if defined(FOR_SIMNOW_BOOT)
-	{
-		struct  atom_gpio_pin_assignment  gpio_pin[8] = {
-				{0x5db5, 0, 0, 1, 0},
-				{0x5db5, 8, 8, 2, 0},
-				{0x5db5, 0x10, 0x10, 3, 0},
-				{0x5db5, 0x18, 0x14, 4, 0},
-				{0x5db5, 0x1A, 0x18, 5, 0},
-				{0x5db5, 0x1C, 0x1C, 6, 0},
-		};
-
-		count = 6;
-		memmove(header->gpio_pin, gpio_pin, sizeof(gpio_pin));
-	}
-#else
 	count = (le16_to_cpu(header->table_header.structuresize)
 			- sizeof(struct atom_common_table_header))
 				/ sizeof(struct atom_gpio_pin_assignment);
-#endif
 	for (i = 0; i < count; ++i) {
 		if (header->gpio_pin[i].gpio_id != gpio_id)
 			continue;
@@ -633,19 +768,37 @@ static enum bp_result bios_parser_get_device_tag(
 	struct bios_parser *bp = BP_FROM_DCB(dcb);
 	struct atom_display_object_path_v2 *object;
 
+	struct atom_display_object_path_v3 *object_path_v3;
+
+
 	if (!info)
 		return BP_RESULT_BADINPUT;
 
-	/* getBiosObject will return MXM object */
-	object = get_bios_object(bp, connector_object_id);
+	switch (bp->object_info_tbl.revision.minor) {
+	case 4:
+	default:
+	        /* getBiosObject will return MXM object */
+		object = get_bios_object(bp, connector_object_id);
 
-	if (!object) {
-		BREAK_TO_DEBUGGER(); /* Invalid object id */
-		return BP_RESULT_BADINPUT;
-	}
+		if (!object) {
+			BREAK_TO_DEBUGGER(); /* Invalid object id */
+			return BP_RESULT_BADINPUT;
+		}
 
-	info->acpi_device = 0; /* BIOS no longer provides this */
-	info->dev_id = device_type_from_device_id(object->device_tag);
+		info->acpi_device = 0; /* BIOS no longer provides this */
+		info->dev_id = device_type_from_device_id(object->device_tag);
+		break;
+	case 5:
+		object_path_v3 = get_bios_object_from_path_v3(bp, connector_object_id);
+
+		if (!object_path_v3) {
+			BREAK_TO_DEBUGGER(); /* Invalid object id */
+			return BP_RESULT_BADINPUT;
+		}
+		info->acpi_device = 0; /* BIOS no longer provides this */
+		info->dev_id = device_type_from_device_id(object_path_v3->device_tag);
+		break;
+	}
 
 	return BP_RESULT_OK;
 }
@@ -685,6 +838,8 @@ static enum bp_result get_ss_info_v4_1(
 				disp_cntl_tbl->dvi_ss_rate_10hz * 10;
 		if (disp_cntl_tbl->dvi_ss_mode & ATOM_SS_CENTRE_SPREAD_MODE)
 			ss_info->type.CENTER_MODE = true;
+
+		DC_LOG_BIOS("AS_SIGNAL_TYPE_DVI ss_percentage: %d\n", ss_info->spread_spectrum_percentage);
 		break;
 	case AS_SIGNAL_TYPE_HDMI:
 		ss_info->spread_spectrum_percentage =
@@ -693,6 +848,8 @@ static enum bp_result get_ss_info_v4_1(
 				disp_cntl_tbl->hdmi_ss_rate_10hz * 10;
 		if (disp_cntl_tbl->hdmi_ss_mode & ATOM_SS_CENTRE_SPREAD_MODE)
 			ss_info->type.CENTER_MODE = true;
+
+		DC_LOG_BIOS("AS_SIGNAL_TYPE_HDMI ss_percentage: %d\n", ss_info->spread_spectrum_percentage);
 		break;
 	/* TODO LVDS not support anymore? */
 	case AS_SIGNAL_TYPE_DISPLAY_PORT:
@@ -702,6 +859,8 @@ static enum bp_result get_ss_info_v4_1(
 				disp_cntl_tbl->dp_ss_rate_10hz * 10;
 		if (disp_cntl_tbl->dp_ss_mode & ATOM_SS_CENTRE_SPREAD_MODE)
 			ss_info->type.CENTER_MODE = true;
+
+		DC_LOG_BIOS("AS_SIGNAL_TYPE_DISPLAY_PORT ss_percentage: %d\n", ss_info->spread_spectrum_percentage);
 		break;
 	case AS_SIGNAL_TYPE_GPU_PLL:
 		/* atom_firmware: DAL only get data from dce_info table.
@@ -715,13 +874,15 @@ static enum bp_result get_ss_info_v4_1(
 				      DATA_TABLES(smu_info));
 		if (!smu_info)
 			return BP_RESULT_BADBIOSTABLE;
-
+		DC_LOG_BIOS("gpuclk_ss_percentage (unit of 0.001 percent): %d\n", smu_info->gpuclk_ss_percentage);
 		ss_info->spread_spectrum_percentage =
 				smu_info->waflclk_ss_percentage;
 		ss_info->spread_spectrum_range =
 				smu_info->gpuclk_ss_rate_10hz * 10;
 		if (smu_info->waflclk_ss_mode & ATOM_SS_CENTRE_SPREAD_MODE)
 			ss_info->type.CENTER_MODE = true;
+
+		DC_LOG_BIOS("AS_SIGNAL_TYPE_XGMI ss_percentage: %d\n", ss_info->spread_spectrum_percentage);
 		break;
 	default:
 		result = BP_RESULT_UNSUPPORTED;
@@ -758,6 +919,7 @@ static enum bp_result get_ss_info_v4_2(
 	if (!smu_info)
 		return BP_RESULT_BADBIOSTABLE;
 
+	DC_LOG_BIOS("gpuclk_ss_percentage (unit of 0.001 percent): %d\n", smu_info->gpuclk_ss_percentage);
 	ss_info->type.STEP_AND_DELAY_INFO = false;
 	ss_info->spread_percentage_divider = 1000;
 	/* BIOS no longer uses target clock.  Always enable for now */
@@ -771,6 +933,8 @@ static enum bp_result get_ss_info_v4_2(
 				disp_cntl_tbl->dvi_ss_rate_10hz * 10;
 		if (disp_cntl_tbl->dvi_ss_mode & ATOM_SS_CENTRE_SPREAD_MODE)
 			ss_info->type.CENTER_MODE = true;
+
+		DC_LOG_BIOS("AS_SIGNAL_TYPE_DVI ss_percentage: %d\n", ss_info->spread_spectrum_percentage);
 		break;
 	case AS_SIGNAL_TYPE_HDMI:
 		ss_info->spread_spectrum_percentage =
@@ -779,6 +943,8 @@ static enum bp_result get_ss_info_v4_2(
 				disp_cntl_tbl->hdmi_ss_rate_10hz * 10;
 		if (disp_cntl_tbl->hdmi_ss_mode & ATOM_SS_CENTRE_SPREAD_MODE)
 			ss_info->type.CENTER_MODE = true;
+
+		DC_LOG_BIOS("AS_SIGNAL_TYPE_HDMI ss_percentage: %d\n", ss_info->spread_spectrum_percentage);
 		break;
 	/* TODO LVDS not support anymore? */
 	case AS_SIGNAL_TYPE_DISPLAY_PORT:
@@ -788,6 +954,8 @@ static enum bp_result get_ss_info_v4_2(
 				smu_info->gpuclk_ss_rate_10hz * 10;
 		if (smu_info->gpuclk_ss_mode & ATOM_SS_CENTRE_SPREAD_MODE)
 			ss_info->type.CENTER_MODE = true;
+
+		DC_LOG_BIOS("AS_SIGNAL_TYPE_DISPLAY_PORT ss_percentage: %d\n", ss_info->spread_spectrum_percentage);
 		break;
 	case AS_SIGNAL_TYPE_GPU_PLL:
 		/* atom_firmware: DAL only get data from dce_info table.
@@ -803,6 +971,84 @@ static enum bp_result get_ss_info_v4_2(
 	return result;
 }
 
+static enum bp_result get_ss_info_v4_5(
+	struct bios_parser *bp,
+	uint32_t id,
+	uint32_t index,
+	struct spread_spectrum_info *ss_info)
+{
+	enum bp_result result = BP_RESULT_OK;
+	struct atom_display_controller_info_v4_5 *disp_cntl_tbl = NULL;
+
+	if (!ss_info)
+		return BP_RESULT_BADINPUT;
+
+	if (!DATA_TABLES(dce_info))
+		return BP_RESULT_BADBIOSTABLE;
+
+	disp_cntl_tbl =  GET_IMAGE(struct atom_display_controller_info_v4_5,
+							DATA_TABLES(dce_info));
+	if (!disp_cntl_tbl)
+		return BP_RESULT_BADBIOSTABLE;
+
+	ss_info->type.STEP_AND_DELAY_INFO = false;
+	ss_info->spread_percentage_divider = 1000;
+	/* BIOS no longer uses target clock.  Always enable for now */
+	ss_info->target_clock_range = 0xffffffff;
+
+	switch (id) {
+	case AS_SIGNAL_TYPE_DVI:
+		ss_info->spread_spectrum_percentage =
+				disp_cntl_tbl->dvi_ss_percentage;
+		ss_info->spread_spectrum_range =
+				disp_cntl_tbl->dvi_ss_rate_10hz * 10;
+		if (disp_cntl_tbl->dvi_ss_mode & ATOM_SS_CENTRE_SPREAD_MODE)
+			ss_info->type.CENTER_MODE = true;
+
+		DC_LOG_BIOS("AS_SIGNAL_TYPE_DVI ss_percentage: %d\n", ss_info->spread_spectrum_percentage);
+		break;
+	case AS_SIGNAL_TYPE_HDMI:
+		ss_info->spread_spectrum_percentage =
+				disp_cntl_tbl->hdmi_ss_percentage;
+		ss_info->spread_spectrum_range =
+				disp_cntl_tbl->hdmi_ss_rate_10hz * 10;
+		if (disp_cntl_tbl->hdmi_ss_mode & ATOM_SS_CENTRE_SPREAD_MODE)
+			ss_info->type.CENTER_MODE = true;
+
+		DC_LOG_BIOS("AS_SIGNAL_TYPE_HDMI ss_percentage: %d\n", ss_info->spread_spectrum_percentage);
+		break;
+	case AS_SIGNAL_TYPE_DISPLAY_PORT:
+		if (bp->base.integrated_info) {
+			DC_LOG_BIOS("gpuclk_ss_percentage (unit of 0.001 percent): %d\n", bp->base.integrated_info->gpuclk_ss_percentage);
+			ss_info->spread_spectrum_percentage =
+					bp->base.integrated_info->gpuclk_ss_percentage;
+			ss_info->type.CENTER_MODE =
+					bp->base.integrated_info->gpuclk_ss_type;
+		} else {
+			ss_info->spread_spectrum_percentage =
+				disp_cntl_tbl->dp_ss_percentage;
+			ss_info->spread_spectrum_range =
+				disp_cntl_tbl->dp_ss_rate_10hz * 10;
+			if (disp_cntl_tbl->dp_ss_mode & ATOM_SS_CENTRE_SPREAD_MODE)
+				ss_info->type.CENTER_MODE = true;
+		}
+		DC_LOG_BIOS("AS_SIGNAL_TYPE_DISPLAY_PORT ss_percentage: %d\n", ss_info->spread_spectrum_percentage);
+		break;
+	case AS_SIGNAL_TYPE_GPU_PLL:
+		/* atom_smu_info_v4_0 does not have fields for SS for SMU Display PLL anymore.
+		 * SMU Display PLL supposed to be without spread.
+		 * Better place for it would be in atom_display_controller_info_v4_5 table.
+		 */
+		result = BP_RESULT_UNSUPPORTED;
+		break;
+	default:
+		result = BP_RESULT_UNSUPPORTED;
+		break;
+	}
+
+	return result;
+}
+
 /**
  * bios_parser_get_spread_spectrum_info
  * Get spread spectrum information from the ASIC_InternalSS_Info(ver 2.1 or
@@ -847,6 +1093,9 @@ static enum bp_result bios_parser_get_spread_spectrum_info(
 		case 3:
 		case 4:
 			return get_ss_info_v4_2(bp, signal, index, ss_info);
+		case 5:
+			return get_ss_info_v4_5(bp, signal, index, ss_info);
+
 		default:
 			ASSERT(0);
 			break;
@@ -887,6 +1136,31 @@ static enum bp_result get_soc_bb_info_v4_4(
 	return result;
 }
 
+static enum bp_result get_soc_bb_info_v4_5(
+	struct bios_parser *bp,
+	struct bp_soc_bb_info *soc_bb_info)
+{
+	enum bp_result result = BP_RESULT_OK;
+	struct atom_display_controller_info_v4_5 *disp_cntl_tbl = NULL;
+
+	if (!soc_bb_info)
+		return BP_RESULT_BADINPUT;
+
+	if (!DATA_TABLES(dce_info))
+		return BP_RESULT_BADBIOSTABLE;
+
+	disp_cntl_tbl =  GET_IMAGE(struct atom_display_controller_info_v4_5,
+							DATA_TABLES(dce_info));
+	if (!disp_cntl_tbl)
+		return BP_RESULT_BADBIOSTABLE;
+
+	soc_bb_info->dram_clock_change_latency_100ns = disp_cntl_tbl->max_mclk_chg_lat;
+	soc_bb_info->dram_sr_enter_exit_latency_100ns = disp_cntl_tbl->max_sr_enter_exit_lat;
+	soc_bb_info->dram_sr_exit_latency_100ns = disp_cntl_tbl->max_sr_exit_lat;
+
+	return result;
+}
+
 static enum bp_result bios_parser_get_soc_bb_info(
 	struct dc_bios *dcb,
 	struct bp_soc_bb_info *soc_bb_info)
@@ -916,6 +1190,9 @@ static enum bp_result bios_parser_get_soc_bb_info(
 		case 4:
 			result = get_soc_bb_info_v4_4(bp, soc_bb_info);
 			break;
+		case 5:
+			result = get_soc_bb_info_v4_5(bp, soc_bb_info);
+			break;
 		default:
 			break;
 		}
@@ -1023,6 +1300,30 @@ static enum bp_result get_disp_caps_v4_4(
 	return result;
 }
 
+static enum bp_result get_disp_caps_v4_5(
+	struct bios_parser *bp,
+	uint8_t *dce_caps)
+{
+	enum bp_result result = BP_RESULT_OK;
+	struct atom_display_controller_info_v4_5 *disp_cntl_tbl = NULL;
+
+	if (!dce_caps)
+		return BP_RESULT_BADINPUT;
+
+	if (!DATA_TABLES(dce_info))
+		return BP_RESULT_BADBIOSTABLE;
+
+	disp_cntl_tbl = GET_IMAGE(struct atom_display_controller_info_v4_5,
+							DATA_TABLES(dce_info));
+
+	if (!disp_cntl_tbl)
+		return BP_RESULT_BADBIOSTABLE;
+
+	*dce_caps = disp_cntl_tbl->display_caps;
+
+	return result;
+}
+
 static enum bp_result bios_parser_get_lttpr_interop(
 	struct dc_bios *dcb,
 	uint8_t *dce_caps)
@@ -1057,6 +1358,11 @@ static enum bp_result bios_parser_get_lttpr_interop(
 			result = get_disp_caps_v4_4(bp, dce_caps);
 			*dce_caps = !!(*dce_caps & DCE_INFO_CAPS_VBIOS_LTTPR_TRANSPARENT_ENABLE);
 			break;
+		case 5:
+			result = get_disp_caps_v4_5(bp, dce_caps);
+			*dce_caps = !!(*dce_caps & DCE_INFO_CAPS_VBIOS_LTTPR_TRANSPARENT_ENABLE);
+			break;
+
 		default:
 			break;
 		}
@@ -1064,7 +1370,7 @@ static enum bp_result bios_parser_get_lttpr_interop(
 	default:
 		break;
 	}
-
+	DC_LOG_BIOS("DCE_INFO_CAPS_VBIOS_LTTPR_TRANSPARENT_ENABLE: %d tbl_revision.major = %d tbl_revision.minor = %d\n", *dce_caps, tbl_revision.major, tbl_revision.minor);
 	return result;
 }
 
@@ -1080,6 +1386,7 @@ static enum bp_result bios_parser_get_lttpr_caps(
 	if (!DATA_TABLES(dce_info))
 		return BP_RESULT_UNSUPPORTED;
 
+	*dce_caps  = 0;
 	header = GET_IMAGE(struct atom_common_table_header,
 						DATA_TABLES(dce_info));
 	get_atom_data_table_revision(header, &tbl_revision);
@@ -1102,6 +1409,10 @@ static enum bp_result bios_parser_get_lttpr_caps(
 			result = get_disp_caps_v4_4(bp, dce_caps);
 			*dce_caps = !!(*dce_caps & DCE_INFO_CAPS_LTTPR_SUPPORT_ENABLE);
 			break;
+		case 5:
+			result = get_disp_caps_v4_5(bp, dce_caps);
+			*dce_caps = !!(*dce_caps & DCE_INFO_CAPS_LTTPR_SUPPORT_ENABLE);
+			break;
 		default:
 			break;
 		}
@@ -1109,7 +1420,11 @@ static enum bp_result bios_parser_get_lttpr_caps(
 	default:
 		break;
 	}
-
+	DC_LOG_BIOS("DCE_INFO_CAPS_LTTPR_SUPPORT_ENABLE: %d tbl_revision.major = %d tbl_revision.minor = %d\n", *dce_caps, tbl_revision.major, tbl_revision.minor);
+	if (dcb->ctx->dc->config.force_bios_enable_lttpr && *dce_caps == 0) {
+		*dce_caps = 1;
+		DC_LOG_BIOS("DCE_INFO_CAPS_VBIOS_LTTPR_TRANSPARENT_ENABLE: forced enabled");
+	}
 	return result;
 }
 
@@ -1165,10 +1480,10 @@ static enum bp_result get_embedded_panel_info_v2_1(
 	/* not provided by VBIOS */
 	info->lcd_timing.misc_info.HORIZONTAL_CUT_OFF = 0;
 
-	info->lcd_timing.misc_info.H_SYNC_POLARITY = ~(uint32_t) (lvds->lcd_timing.miscinfo
-			& ATOM_HSYNC_POLARITY);
-	info->lcd_timing.misc_info.V_SYNC_POLARITY = ~(uint32_t) (lvds->lcd_timing.miscinfo
-			& ATOM_VSYNC_POLARITY);
+	info->lcd_timing.misc_info.H_SYNC_POLARITY = !(lvds->lcd_timing.miscinfo &
+						       ATOM_HSYNC_POLARITY);
+	info->lcd_timing.misc_info.V_SYNC_POLARITY = !(lvds->lcd_timing.miscinfo &
+						       ATOM_VSYNC_POLARITY);
 
 	/* not provided by VBIOS */
 	info->lcd_timing.misc_info.VERTICAL_CUT_OFF = 0;
@@ -1274,8 +1589,15 @@ static bool bios_parser_is_device_id_supported(
 
 	uint32_t mask = get_support_mask_for_device_id(id);
 
-	return (le16_to_cpu(bp->object_info_tbl.v1_4->supporteddevices) &
-								mask) != 0;
+	switch (bp->object_info_tbl.revision.minor) {
+	case 4:
+	default:
+		return (le16_to_cpu(bp->object_info_tbl.v1_4->supporteddevices) & mask) != 0;
+		break;
+	case 5:
+		return (le16_to_cpu(bp->object_info_tbl.v1_5->supporteddevices) & mask) != 0;
+		break;
+	}
 }
 
 static uint32_t bios_parser_get_ss_entry_number(
@@ -1378,14 +1700,15 @@ static enum bp_result bios_parser_enable_disp_power_gating(
 static enum bp_result bios_parser_enable_lvtma_control(
 	struct dc_bios *dcb,
 	uint8_t uc_pwr_on,
-	uint8_t panel_instance)
+	uint8_t pwrseq_instance,
+	uint8_t bypass_panel_control_wait)
 {
 	struct bios_parser *bp = BP_FROM_DCB(dcb);
 
 	if (!bp->cmd_tbl.enable_lvtma_control)
 		return BP_RESULT_FAILURE;
 
-	return bp->cmd_tbl.enable_lvtma_control(bp, uc_pwr_on, panel_instance);
+	return bp->cmd_tbl.enable_lvtma_control(bp, uc_pwr_on, pwrseq_instance, bypass_panel_control_wait);
 }
 
 static bool bios_parser_is_accelerated_mode(
@@ -1413,7 +1736,7 @@ static enum bp_result bios_parser_get_firmware_info(
 	struct dc_firmware_info *info)
 {
 	struct bios_parser *bp = BP_FROM_DCB(dcb);
-	enum bp_result result = BP_RESULT_BADBIOSTABLE;
+	static enum bp_result result = BP_RESULT_BADBIOSTABLE;
 	struct atom_common_table_header *header;
 
 	struct atom_data_revision revision;
@@ -1431,10 +1754,13 @@ static enum bp_result bios_parser_get_firmware_info(
 			case 2:
 			case 3:
 				result = get_firmware_info_v3_2(bp, info);
-                                break;
+				break;
 			case 4:
 				result = get_firmware_info_v3_4(bp, info);
 				break;
+			case 5:
+				result = get_firmware_info_v3_5(bp, info);
+				break;
 			default:
 				break;
 			}
@@ -1452,6 +1778,7 @@ static enum bp_result get_firmware_info_v3_1(
 	struct dc_firmware_info *info)
 {
 	struct atom_firmware_info_v3_1 *firmware_info;
+	struct atom_firmware_info_v3_2 *firmware_info32;
 	struct atom_display_controller_info_v4_1 *dce_info = NULL;
 
 	if (!info)
@@ -1459,11 +1786,13 @@ static enum bp_result get_firmware_info_v3_1(
 
 	firmware_info = GET_IMAGE(struct atom_firmware_info_v3_1,
 			DATA_TABLES(firmwareinfo));
+	firmware_info32 = GET_IMAGE(struct atom_firmware_info_v3_2,
+			DATA_TABLES(firmwareinfo));
 
 	dce_info = GET_IMAGE(struct atom_display_controller_info_v4_1,
 			DATA_TABLES(dce_info));
 
-	if (!firmware_info || !dce_info)
+	if (!firmware_info || !firmware_info32 || !dce_info)
 		return BP_RESULT_BADBIOSTABLE;
 
 	memset(info, 0, sizeof(*info));
@@ -1491,7 +1820,15 @@ static enum bp_result get_firmware_info_v3_1(
 				bp->cmd_tbl.get_smu_clock_info(bp, SMU9_SYSPLL0_ID) * 10;
 	}
 
-	info->oem_i2c_present = false;
+	/* These fields are marked as reserved in v3_1, but they appear to be populated
+	 * properly.
+	 */
+	if (firmware_info32 && firmware_info32->board_i2c_feature_id == 0x2) {
+		info->oem_i2c_present = true;
+		info->oem_i2c_obj_id = firmware_info32->board_i2c_feature_gpio_id;
+	} else {
+		info->oem_i2c_present = false;
+	}
 
 	return BP_RESULT_OK;
 }
@@ -1529,19 +1866,21 @@ static enum bp_result get_firmware_info_v3_2(
 		/* Vega12 */
 		smu_info_v3_2 = GET_IMAGE(struct atom_smu_info_v3_2,
 							DATA_TABLES(smu_info));
-
 		if (!smu_info_v3_2)
 			return BP_RESULT_BADBIOSTABLE;
 
+		DC_LOG_BIOS("gpuclk_ss_percentage (unit of 0.001 percent): %d\n", smu_info_v3_2->gpuclk_ss_percentage);
+
 		info->default_engine_clk = smu_info_v3_2->bootup_dcefclk_10khz * 10;
 	} else if (revision.minor == 3) {
 		/* Vega20 */
 		smu_info_v3_3 = GET_IMAGE(struct atom_smu_info_v3_3,
 							DATA_TABLES(smu_info));
-
 		if (!smu_info_v3_3)
 			return BP_RESULT_BADBIOSTABLE;
 
+		DC_LOG_BIOS("gpuclk_ss_percentage (unit of 0.001 percent): %d\n", smu_info_v3_3->gpuclk_ss_percentage);
+
 		info->default_engine_clk = smu_info_v3_3->bootup_dcefclk_10khz * 10;
 	}
 
@@ -1590,6 +1929,11 @@ static enum bp_result get_firmware_info_v3_4(
 	struct atom_data_revision revision;
 	struct atom_display_controller_info_v4_1 *dce_info_v4_1 = NULL;
 	struct atom_display_controller_info_v4_4 *dce_info_v4_4 = NULL;
+
+	struct atom_smu_info_v3_5 *smu_info_v3_5 = NULL;
+	struct atom_display_controller_info_v4_5 *dce_info_v4_5 = NULL;
+	struct atom_smu_info_v4_0 *smu_info_v4_0 = NULL;
+
 	if (!info)
 		return BP_RESULT_BADINPUT;
 
@@ -1609,6 +1953,22 @@ static enum bp_result get_firmware_info_v3_4(
 	switch (revision.major) {
 	case 4:
 		switch (revision.minor) {
+		case 5:
+			dce_info_v4_5 = GET_IMAGE(struct atom_display_controller_info_v4_5,
+							DATA_TABLES(dce_info));
+
+			if (!dce_info_v4_5)
+				return BP_RESULT_BADBIOSTABLE;
+
+			 /* 100MHz expected */
+			info->pll_info.crystal_frequency = dce_info_v4_5->dce_refclk_10khz * 10;
+			info->dp_phy_ref_clk             = dce_info_v4_5->dpphy_refclk_10khz * 10;
+			 /* 50MHz expected */
+			info->i2c_engine_ref_clk         = dce_info_v4_5->i2c_engine_refclk_10khz * 10;
+
+			/* For DCN32/321 Display PLL VCO Frequency from dce_info_v4_5 may not be reliable */
+			break;
+
 		case 4:
 			dce_info_v4_4 = GET_IMAGE(struct atom_display_controller_info_v4_4,
 							DATA_TABLES(dce_info));
@@ -1650,6 +2010,45 @@ static enum bp_result get_firmware_info_v3_4(
 					DATA_TABLES(smu_info));
 	get_atom_data_table_revision(header, &revision);
 
+	switch (revision.major) {
+	case 3:
+		switch (revision.minor) {
+		case 5:
+			smu_info_v3_5 = GET_IMAGE(struct atom_smu_info_v3_5,
+							DATA_TABLES(smu_info));
+
+			if (!smu_info_v3_5)
+				return BP_RESULT_BADBIOSTABLE;
+			DC_LOG_BIOS("gpuclk_ss_percentage (unit of 0.001 percent): %d\n", smu_info_v3_5->gpuclk_ss_percentage);
+			info->default_engine_clk = smu_info_v3_5->bootup_dcefclk_10khz * 10;
+			break;
+
+		default:
+			break;
+		}
+		break;
+
+	case 4:
+		switch (revision.minor) {
+		case 0:
+			smu_info_v4_0 = GET_IMAGE(struct atom_smu_info_v4_0,
+							DATA_TABLES(smu_info));
+
+			if (!smu_info_v4_0)
+				return BP_RESULT_BADBIOSTABLE;
+
+			/* For DCN32/321 bootup DCFCLK from smu_info_v4_0 may not be reliable */
+			break;
+
+		default:
+			break;
+		}
+		break;
+
+	default:
+		break;
+	}
+
 	 // We need to convert from 10KHz units into KHz units.
 	info->default_memory_clk = firmware_info->bootup_mclk_in10khz * 10;
 
@@ -1663,6 +2062,63 @@ static enum bp_result get_firmware_info_v3_4(
 	return BP_RESULT_OK;
 }
 
+static enum bp_result get_firmware_info_v3_5(
+	struct bios_parser *bp,
+	struct dc_firmware_info *info)
+{
+	struct atom_firmware_info_v3_5 *firmware_info;
+	struct atom_common_table_header *header;
+	struct atom_data_revision revision;
+	struct atom_display_controller_info_v4_5 *dce_info_v4_5 = NULL;
+
+	if (!info)
+		return BP_RESULT_BADINPUT;
+
+	firmware_info = GET_IMAGE(struct atom_firmware_info_v3_5,
+			DATA_TABLES(firmwareinfo));
+
+	if (!firmware_info)
+		return BP_RESULT_BADBIOSTABLE;
+
+	memset(info, 0, sizeof(*info));
+
+	if (firmware_info->board_i2c_feature_id == 0x2) {
+		info->oem_i2c_present = true;
+		info->oem_i2c_obj_id = firmware_info->board_i2c_feature_gpio_id;
+	} else {
+		info->oem_i2c_present = false;
+	}
+
+	header = GET_IMAGE(struct atom_common_table_header,
+					DATA_TABLES(dce_info));
+
+	get_atom_data_table_revision(header, &revision);
+
+	switch (revision.major) {
+	case 4:
+		switch (revision.minor) {
+		case 5:
+			dce_info_v4_5 = GET_IMAGE(struct atom_display_controller_info_v4_5,
+							DATA_TABLES(dce_info));
+
+			if (!dce_info_v4_5)
+				return BP_RESULT_BADBIOSTABLE;
+
+			 /* 100MHz expected */
+			info->pll_info.crystal_frequency = dce_info_v4_5->dce_refclk_10khz * 10;
+			break;
+		default:
+			break;
+		}
+		break;
+	default:
+		break;
+	}
+
+
+	return BP_RESULT_OK;
+}
+
 static enum bp_result bios_parser_get_encoder_cap_info(
 	struct dc_bios *dcb,
 	struct graphics_object_id object_id,
@@ -1675,6 +2131,12 @@ static enum bp_result bios_parser_get_encoder_cap_info(
 	if (!info)
 		return BP_RESULT_BADINPUT;
 
+#if defined(CONFIG_DRM_AMD_DC_FP)
+	/* encoder cap record not available in v1_5 */
+	if (bp->object_info_tbl.revision.minor == 5)
+		return BP_RESULT_NORECORD;
+#endif
+
 	object = get_bios_object(bp, object_id);
 
 	if (!object)
@@ -1683,6 +2145,7 @@ static enum bp_result bios_parser_get_encoder_cap_info(
 	record = get_encoder_cap_record(bp, object);
 	if (!record)
 		return BP_RESULT_NORECORD;
+	DC_LOG_BIOS("record->encodercaps 0x%x for object_id 0x%x", record->encodercaps, object_id.id);
 
 	info->DP_HBR2_CAP = (record->encodercaps &
 			ATOM_ENCODER_CAP_RECORD_HBR2) ? 1 : 0;
@@ -1692,7 +2155,6 @@ static enum bp_result bios_parser_get_encoder_cap_info(
 			ATOM_ENCODER_CAP_RECORD_HBR3_EN) ? 1 : 0;
 	info->HDMI_6GB_EN = (record->encodercaps &
 			ATOM_ENCODER_CAP_RECORD_HDMI6Gbps_EN) ? 1 : 0;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	info->IS_DP2_CAPABLE = (record->encodercaps &
 			ATOM_ENCODER_CAP_RECORD_DP2) ? 1 : 0;
 	info->DP_UHBR10_EN = (record->encodercaps &
@@ -1701,9 +2163,9 @@ static enum bp_result bios_parser_get_encoder_cap_info(
 			ATOM_ENCODER_CAP_RECORD_UHBR13_5_EN) ? 1 : 0;
 	info->DP_UHBR20_EN = (record->encodercaps &
 			ATOM_ENCODER_CAP_RECORD_UHBR20_EN) ? 1 : 0;
-#endif
 	info->DP_IS_USB_C = (record->encodercaps &
 			ATOM_ENCODER_CAP_RECORD_USB_C_TYPE) ? 1 : 0;
+	DC_LOG_BIOS("\t info->DP_IS_USB_C %d", info->DP_IS_USB_C);
 
 	return BP_RESULT_OK;
 }
@@ -1783,6 +2245,41 @@ static struct atom_disp_connector_caps_record *get_disp_connector_caps_record(
 	return NULL;
 }
 
+static struct atom_connector_caps_record *get_connector_caps_record(struct bios_parser *bp,
+								    struct atom_display_object_path_v3 *object)
+{
+	struct atom_common_record_header *header;
+	uint32_t offset;
+
+	if (!object) {
+		BREAK_TO_DEBUGGER(); /* Invalid object */
+		return NULL;
+	}
+
+	offset = object->disp_recordoffset + bp->object_info_tbl_offset;
+
+	for (;;) {
+		header = GET_IMAGE(struct atom_common_record_header, offset);
+
+		if (!header)
+			return NULL;
+
+		offset += header->record_size;
+
+		if (header->record_type == ATOM_RECORD_END_TYPE ||
+				!header->record_size)
+			break;
+
+		if (header->record_type != ATOM_CONNECTOR_CAP_RECORD_TYPE)
+			continue;
+
+		if (sizeof(struct atom_connector_caps_record) <= header->record_size)
+			return (struct atom_connector_caps_record *)header;
+	}
+
+	return NULL;
+}
+
 static enum bp_result bios_parser_get_disp_connector_caps_info(
 	struct dc_bios *dcb,
 	struct graphics_object_id object_id,
@@ -1790,25 +2287,113 @@ static enum bp_result bios_parser_get_disp_connector_caps_info(
 {
 	struct bios_parser *bp = BP_FROM_DCB(dcb);
 	struct atom_display_object_path_v2 *object;
+	struct atom_display_object_path_v3 *object_path_v3;
+	struct atom_connector_caps_record *record_path_v3;
 	struct atom_disp_connector_caps_record *record = NULL;
 
 	if (!info)
 		return BP_RESULT_BADINPUT;
 
-	object = get_bios_object(bp, object_id);
+	switch (bp->object_info_tbl.revision.minor) {
+	case 4:
+		default:
+			object = get_bios_object(bp, object_id);
 
-	if (!object)
+			if (!object)
+				return BP_RESULT_BADINPUT;
+
+			record = get_disp_connector_caps_record(bp, object);
+			if (!record)
+				return BP_RESULT_NORECORD;
+
+			info->INTERNAL_DISPLAY =
+				(record->connectcaps & ATOM_CONNECTOR_CAP_INTERNAL_DISPLAY) ? 1 : 0;
+			info->INTERNAL_DISPLAY_BL =
+				(record->connectcaps & ATOM_CONNECTOR_CAP_INTERNAL_DISPLAY_BL) ? 1 : 0;
+			break;
+	case 5:
+		object_path_v3 = get_bios_object_from_path_v3(bp, object_id);
+
+		if (!object_path_v3)
+			return BP_RESULT_BADINPUT;
+
+		record_path_v3 = get_connector_caps_record(bp, object_path_v3);
+		if (!record_path_v3)
+			return BP_RESULT_NORECORD;
+
+		info->INTERNAL_DISPLAY = (record_path_v3->connector_caps & ATOM_CONNECTOR_CAP_INTERNAL_DISPLAY)
+									? 1 : 0;
+		info->INTERNAL_DISPLAY_BL = (record_path_v3->connector_caps & ATOM_CONNECTOR_CAP_INTERNAL_DISPLAY_BL)
+										? 1 : 0;
+		break;
+	}
+
+	return BP_RESULT_OK;
+}
+
+static struct atom_connector_speed_record *get_connector_speed_cap_record(struct bios_parser *bp,
+									  struct atom_display_object_path_v3 *object)
+{
+	struct atom_common_record_header *header;
+	uint32_t offset;
+
+	if (!object) {
+		BREAK_TO_DEBUGGER(); /* Invalid object */
+		return NULL;
+	}
+
+	offset = object->disp_recordoffset + bp->object_info_tbl_offset;
+
+	for (;;) {
+		header = GET_IMAGE(struct atom_common_record_header, offset);
+
+		if (!header)
+			return NULL;
+
+		offset += header->record_size;
+
+		if (header->record_type == ATOM_RECORD_END_TYPE ||
+				!header->record_size)
+			break;
+
+		if (header->record_type != ATOM_CONNECTOR_SPEED_UPTO)
+			continue;
+
+		if (sizeof(struct atom_connector_speed_record) <= header->record_size)
+			return (struct atom_connector_speed_record *)header;
+	}
+
+	return NULL;
+}
+
+static enum bp_result bios_parser_get_connector_speed_cap_info(
+	struct dc_bios *dcb,
+	struct graphics_object_id object_id,
+	struct bp_connector_speed_cap_info *info)
+{
+	struct bios_parser *bp = BP_FROM_DCB(dcb);
+	struct atom_display_object_path_v3 *object_path_v3;
+	//struct atom_connector_speed_record *record = NULL;
+	struct atom_connector_speed_record *record;
+
+	if (!info)
+		return BP_RESULT_BADINPUT;
+
+	object_path_v3 = get_bios_object_from_path_v3(bp, object_id);
+
+	if (!object_path_v3)
 		return BP_RESULT_BADINPUT;
 
-	record = get_disp_connector_caps_record(bp, object);
+	record = get_connector_speed_cap_record(bp, object_path_v3);
 	if (!record)
 		return BP_RESULT_NORECORD;
 
-	info->INTERNAL_DISPLAY = (record->connectcaps & ATOM_CONNECTOR_CAP_INTERNAL_DISPLAY)
-									? 1 : 0;
-	info->INTERNAL_DISPLAY_BL = (record->connectcaps & ATOM_CONNECTOR_CAP_INTERNAL_DISPLAY_BL)
-											? 1 : 0;
-
+	info->DP_HBR2_EN = (record->connector_max_speed >= 5400) ? 1 : 0;
+	info->DP_HBR3_EN = (record->connector_max_speed >= 8100) ? 1 : 0;
+	info->HDMI_6GB_EN = (record->connector_max_speed >= 5940) ? 1 : 0;
+	info->DP_UHBR10_EN = (record->connector_max_speed >= 10000) ? 1 : 0;
+	info->DP_UHBR13_5_EN = (record->connector_max_speed >= 13500) ? 1 : 0;
+	info->DP_UHBR20_EN = (record->connector_max_speed >= 20000) ? 1 : 0;
 	return BP_RESULT_OK;
 }
 
@@ -1817,7 +2402,7 @@ static enum bp_result get_vram_info_v23(
 	struct dc_vram_info *info)
 {
 	struct atom_vram_info_header_v2_3 *info_v23;
-	enum bp_result result = BP_RESULT_OK;
+	static enum bp_result result = BP_RESULT_OK;
 
 	info_v23 = GET_IMAGE(struct atom_vram_info_header_v2_3,
 						DATA_TABLES(vram_info));
@@ -1836,7 +2421,7 @@ static enum bp_result get_vram_info_v24(
 	struct dc_vram_info *info)
 {
 	struct atom_vram_info_header_v2_4 *info_v24;
-	enum bp_result result = BP_RESULT_OK;
+	static enum bp_result result = BP_RESULT_OK;
 
 	info_v24 = GET_IMAGE(struct atom_vram_info_header_v2_4,
 						DATA_TABLES(vram_info));
@@ -1855,7 +2440,7 @@ static enum bp_result get_vram_info_v25(
 	struct dc_vram_info *info)
 {
 	struct atom_vram_info_header_v2_5 *info_v25;
-	enum bp_result result = BP_RESULT_OK;
+	static enum bp_result result = BP_RESULT_OK;
 
 	info_v25 = GET_IMAGE(struct atom_vram_info_header_v2_5,
 						DATA_TABLES(vram_info));
@@ -1869,6 +2454,44 @@ static enum bp_result get_vram_info_v25(
 	return result;
 }
 
+static enum bp_result get_vram_info_v30(
+	struct bios_parser *bp,
+	struct dc_vram_info *info)
+{
+	struct atom_vram_info_header_v3_0 *info_v30;
+	enum bp_result result = BP_RESULT_OK;
+
+	info_v30 = GET_IMAGE(struct atom_vram_info_header_v3_0,
+						DATA_TABLES(vram_info));
+
+	if (info_v30 == NULL)
+		return BP_RESULT_BADBIOSTABLE;
+
+	info->num_chans = info_v30->channel_num;
+	info->dram_channel_width_bytes = (1 << info_v30->channel_width) / 8;
+
+	return result;
+}
+
+static enum bp_result get_vram_info_from_umc_info_v40(
+		struct bios_parser *bp,
+		struct dc_vram_info *info)
+{
+	struct atom_umc_info_v4_0 *info_v40;
+	enum bp_result result = BP_RESULT_OK;
+
+	info_v40 = GET_IMAGE(struct atom_umc_info_v4_0,
+						DATA_TABLES(umc_info));
+
+	if (info_v40 == NULL)
+		return BP_RESULT_BADBIOSTABLE;
+
+	info->num_chans = info_v40->channel_num;
+	info->dram_channel_width_bytes = (1 << info_v40->channel_width) / 8;
+
+	return result;
+}
+
 /*
  * get_integrated_info_v11
  *
@@ -1880,7 +2503,7 @@ static enum bp_result get_vram_info_v25(
  * integrated_info *info - [out] store and output integrated info
  *
  * @return
- * enum bp_result - BP_RESULT_OK if information is available,
+ * static enum bp_result - BP_RESULT_OK if information is available,
  *                  BP_RESULT_BADBIOSTABLE otherwise.
  */
 static enum bp_result get_integrated_info_v11(
@@ -1896,6 +2519,8 @@ static enum bp_result get_integrated_info_v11(
 	if (info_v11 == NULL)
 		return BP_RESULT_BADBIOSTABLE;
 
+	DC_LOG_BIOS("gpuclk_ss_percentage (unit of 0.001 percent): %d\n", info_v11->gpuclk_ss_percentage);
+
 	info->gpu_cap_info =
 	le32_to_cpu(info_v11->gpucapinfo);
 	/*
@@ -2111,6 +2736,8 @@ static enum bp_result get_integrated_info_v2_1(
 	if (info_v2_1 == NULL)
 		return BP_RESULT_BADBIOSTABLE;
 
+	DC_LOG_BIOS("gpuclk_ss_percentage (unit of 0.001 percent): %d\n", info_v2_1->gpuclk_ss_percentage);
+
 	info->gpu_cap_info =
 	le32_to_cpu(info_v2_1->gpucapinfo);
 	/*
@@ -2271,6 +2898,8 @@ static enum bp_result get_integrated_info_v2_2(
 	if (info_v2_2 == NULL)
 		return BP_RESULT_BADBIOSTABLE;
 
+	DC_LOG_BIOS("gpuclk_ss_percentage (unit of 0.001 percent): %d\n", info_v2_2->gpuclk_ss_percentage);
+
 	info->gpu_cap_info =
 	le32_to_cpu(info_v2_2->gpucapinfo);
 	/*
@@ -2287,6 +2916,8 @@ static enum bp_result get_integrated_info_v2_2(
 	info->ma_channel_number = info_v2_2->umachannelnumber;
 	info->dp_ss_control =
 		le16_to_cpu(info_v2_2->reserved1);
+	info->gpuclk_ss_percentage = info_v2_2->gpuclk_ss_percentage;
+	info->gpuclk_ss_type = info_v2_2->gpuclk_ss_type;
 
 	for (i = 0; i < NUMBER_OF_UCHAR_FOR_GUID; ++i) {
 		info->ext_disp_conn_info.gu_id[i] =
@@ -2371,19 +3002,23 @@ static enum bp_result get_integrated_info_v2_2(
  * integrated_info *info - [out] store and output integrated info
  *
  * @return
- * enum bp_result - BP_RESULT_OK if information is available,
+ * static enum bp_result - BP_RESULT_OK if information is available,
  *                  BP_RESULT_BADBIOSTABLE otherwise.
  */
 static enum bp_result construct_integrated_info(
 	struct bios_parser *bp,
 	struct integrated_info *info)
 {
-	enum bp_result result = BP_RESULT_BADBIOSTABLE;
+	static enum bp_result result = BP_RESULT_BADBIOSTABLE;
 
 	struct atom_common_table_header *header;
 	struct atom_data_revision revision;
-	uint32_t i;
-	uint32_t j;
+
+	int32_t i;
+	int32_t j;
+
+	if (!info)
+		return result;
 
 	if (info && DATA_TABLES(integratedsysteminfo)) {
 		header = GET_IMAGE(struct atom_common_table_header,
@@ -2408,6 +3043,7 @@ static enum bp_result construct_integrated_info(
 				result = get_integrated_info_v2_1(bp, info);
 				break;
 			case 2:
+			case 3:
 				result = get_integrated_info_v2_2(bp, info);
 				break;
 			default:
@@ -2417,21 +3053,76 @@ static enum bp_result construct_integrated_info(
 		default:
 			return result;
 		}
+		if (result == BP_RESULT_OK) {
+
+			DC_LOG_BIOS("edp1:\n"
+						"\tedp_pwr_on_off_delay = %d\n"
+						"\tedp_pwr_on_vary_bl_to_blon = %d\n"
+						"\tedp_pwr_down_bloff_to_vary_bloff = %d\n"
+						"\tedp_bootup_bl_level = %d\n",
+						info->edp1_info.edp_pwr_on_off_delay,
+						info->edp1_info.edp_pwr_on_vary_bl_to_blon,
+						info->edp1_info.edp_pwr_down_bloff_to_vary_bloff,
+						info->edp1_info.edp_bootup_bl_level);
+			DC_LOG_BIOS("edp2:\n"
+						"\tedp_pwr_on_off_delayv = %d\n"
+						"\tedp_pwr_on_vary_bl_to_blon = %d\n"
+						"\tedp_pwr_down_bloff_to_vary_bloff = %d\n"
+						"\tedp_bootup_bl_level = %d\n",
+						info->edp2_info.edp_pwr_on_off_delay,
+						info->edp2_info.edp_pwr_on_vary_bl_to_blon,
+						info->edp2_info.edp_pwr_down_bloff_to_vary_bloff,
+						info->edp2_info.edp_bootup_bl_level);
+		}
 	}
 
 	if (result != BP_RESULT_OK)
 		return result;
-
+	else {
+		// Log each external path
+		for (i = 0; i < MAX_NUMBER_OF_EXT_DISPLAY_PATH; i++) {
+			if (info->ext_disp_conn_info.path[i].device_tag != 0)
+				DC_LOG_BIOS("integrated_info:For EXTERNAL DISPLAY PATH %d --------------\n"
+						"DEVICE_TAG: 0x%x\n"
+						"DEVICE_ACPI_ENUM: 0x%x\n"
+						"DEVICE_CONNECTOR_ID: 0x%x\n"
+						"EXT_AUX_DDC_LUT_INDEX: %d\n"
+						"EXT_HPD_PIN_LUT_INDEX: %d\n"
+						"EXT_ENCODER_OBJ_ID: 0x%x\n"
+						"Encoder CAPS: 0x%x\n",
+						i,
+						info->ext_disp_conn_info.path[i].device_tag,
+						info->ext_disp_conn_info.path[i].device_acpi_enum,
+						info->ext_disp_conn_info.path[i].device_connector_id.id,
+						info->ext_disp_conn_info.path[i].ext_aux_ddc_lut_index,
+						info->ext_disp_conn_info.path[i].ext_hpd_pin_lut_index,
+						info->ext_disp_conn_info.path[i].ext_encoder_obj_id.id,
+						info->ext_disp_conn_info.path[i].caps
+						);
+			if ((info->ext_disp_conn_info.path[i].caps & AMD_EXT_DISPLAY_PATH_CAPS__EXT_CHIP_MASK) == AMD_EXT_DISPLAY_PATH_CAPS__DP_FIXED_VS_EN)
+				DC_LOG_BIOS("BIOS AMD_EXT_DISPLAY_PATH_CAPS__DP_FIXED_VS_EN on path %d\n", i);
+			else if (bp->base.ctx->dc->config.force_bios_fixed_vs) {
+				info->ext_disp_conn_info.path[i].caps &= ~AMD_EXT_DISPLAY_PATH_CAPS__EXT_CHIP_MASK;
+				info->ext_disp_conn_info.path[i].caps |= AMD_EXT_DISPLAY_PATH_CAPS__DP_FIXED_VS_EN;
+				DC_LOG_BIOS("driver forced AMD_EXT_DISPLAY_PATH_CAPS__DP_FIXED_VS_EN on path %d\n", i);
+			}
+		}
+		// Log the Checksum and Voltage Swing
+		DC_LOG_BIOS("Integrated info table CHECKSUM: %d\n"
+					"Integrated info table FIX_DP_VOLTAGE_SWING: %d\n",
+					info->ext_disp_conn_info.checksum,
+					info->ext_disp_conn_info.fixdpvoltageswing);
+		if (bp->base.ctx->dc->config.force_bios_fixed_vs && info->ext_disp_conn_info.fixdpvoltageswing == 0) {
+			info->ext_disp_conn_info.fixdpvoltageswing = bp->base.ctx->dc->config.force_bios_fixed_vs & 0xF;
+			DC_LOG_BIOS("driver forced fixdpvoltageswing = %d\n", info->ext_disp_conn_info.fixdpvoltageswing);
+		}
+	}
 	/* Sort voltage table from low to high*/
 	for (i = 1; i < NUMBER_OF_DISP_CLK_VOLTAGE; ++i) {
 		for (j = i; j > 0; --j) {
 			if (info->disp_clk_voltage[j].max_supported_clk <
-				info->disp_clk_voltage[j-1].max_supported_clk
-				) {
-				/* swap j and j - 1*/
-				swap(info->disp_clk_voltage[j - 1],
-				     info->disp_clk_voltage[j]);
-			}
+			    info->disp_clk_voltage[j-1].max_supported_clk)
+				swap(info->disp_clk_voltage[j-1], info->disp_clk_voltage[j]);
 		}
 	}
 
@@ -2447,7 +3138,29 @@ static enum bp_result bios_parser_get_vram_info(
 	struct atom_common_table_header *header;
 	struct atom_data_revision revision;
 
-	if (info && DATA_TABLES(vram_info)) {
+	// vram info moved to umc_info for DCN4x
+	if (info && DATA_TABLES(umc_info)) {
+		header = GET_IMAGE(struct atom_common_table_header,
+					DATA_TABLES(umc_info));
+
+		get_atom_data_table_revision(header, &revision);
+
+		switch (revision.major) {
+		case 4:
+			switch (revision.minor) {
+			case 0:
+				result = get_vram_info_from_umc_info_v40(bp, info);
+				break;
+			default:
+				break;
+			}
+			break;
+		default:
+			break;
+		}
+	}
+
+	if (result != BP_RESULT_OK && info && DATA_TABLES(vram_info)) {
 		header = GET_IMAGE(struct atom_common_table_header,
 					DATA_TABLES(vram_info));
 
@@ -2470,6 +3183,16 @@ static enum bp_result bios_parser_get_vram_info(
 			}
 			break;
 
+		case 3:
+			switch (revision.minor) {
+			case 0:
+				result = get_vram_info_v30(bp, info);
+				break;
+			default:
+				break;
+			}
+			break;
+
 		default:
 			return result;
 		}
@@ -2482,7 +3205,7 @@ static struct integrated_info *bios_parser_create_integrated_info(
 	struct dc_bios *dcb)
 {
 	struct bios_parser *bp = BP_FROM_DCB(dcb);
-	struct integrated_info *info = NULL;
+	struct integrated_info *info;
 
 	info = kzalloc(sizeof(struct integrated_info), GFP_KERNEL);
 
@@ -2509,7 +3232,7 @@ static enum bp_result update_slot_layout_info(
 	struct atom_display_object_path_v2 *object;
 	struct atom_bracket_layout_record *record;
 	struct atom_common_record_header *record_header;
-	enum bp_result result;
+	static enum bp_result result;
 	struct bios_parser *bp;
 	struct object_info_table *tbl;
 	struct display_object_info_table_v1_4 *v1_4;
@@ -2615,6 +3338,105 @@ static enum bp_result update_slot_layout_info(
 	return result;
 }
 
+static enum bp_result update_slot_layout_info_v2(
+	struct dc_bios *dcb,
+	unsigned int i,
+	struct slot_layout_info *slot_layout_info)
+{
+	unsigned int record_offset;
+	struct atom_display_object_path_v3 *object;
+	struct atom_bracket_layout_record_v2 *record;
+	struct atom_common_record_header *record_header;
+	static enum bp_result result;
+	struct bios_parser *bp;
+	struct object_info_table *tbl;
+	struct display_object_info_table_v1_5 *v1_5;
+	struct graphics_object_id connector_id;
+
+	record = NULL;
+	record_header = NULL;
+	result = BP_RESULT_NORECORD;
+
+	bp = BP_FROM_DCB(dcb);
+	tbl = &bp->object_info_tbl;
+	v1_5 = tbl->v1_5;
+
+	object = &v1_5->display_path[i];
+	record_offset = (unsigned int)
+		(object->disp_recordoffset) +
+		(unsigned int)(bp->object_info_tbl_offset);
+
+	for (;;) {
+
+		record_header = (struct atom_common_record_header *)
+			GET_IMAGE(struct atom_common_record_header,
+			record_offset);
+		if (record_header == NULL) {
+			result = BP_RESULT_BADBIOSTABLE;
+			break;
+		}
+
+		/* the end of the list */
+		if (record_header->record_type == ATOM_RECORD_END_TYPE ||
+			record_header->record_size == 0)	{
+			break;
+		}
+
+		if (record_header->record_type ==
+			ATOM_BRACKET_LAYOUT_V2_RECORD_TYPE &&
+			sizeof(struct atom_bracket_layout_record_v2)
+			<= record_header->record_size) {
+			record = (struct atom_bracket_layout_record_v2 *)
+				(record_header);
+			result = BP_RESULT_OK;
+			break;
+		}
+
+		record_offset += record_header->record_size;
+	}
+
+	/* return if the record not found */
+	if (result != BP_RESULT_OK)
+		return result;
+
+	/* get slot sizes */
+	connector_id = object_id_from_bios_object_id(object->display_objid);
+
+	slot_layout_info->length = record->bracketlen;
+	slot_layout_info->width = record->bracketwidth;
+	slot_layout_info->num_of_connectors = v1_5->number_of_path;
+	slot_layout_info->connectors[i].position = record->conn_num;
+	slot_layout_info->connectors[i].connector_id = connector_id;
+
+	switch (connector_id.id) {
+	case CONNECTOR_ID_SINGLE_LINK_DVID:
+	case CONNECTOR_ID_DUAL_LINK_DVID:
+		slot_layout_info->connectors[i].connector_type = CONNECTOR_LAYOUT_TYPE_DVI_D;
+		slot_layout_info->connectors[i].length = CONNECTOR_SIZE_DVI;
+		break;
+
+	case CONNECTOR_ID_HDMI_TYPE_A:
+		slot_layout_info->connectors[i].connector_type = CONNECTOR_LAYOUT_TYPE_HDMI;
+		slot_layout_info->connectors[i].length = CONNECTOR_SIZE_HDMI;
+		break;
+
+	case CONNECTOR_ID_DISPLAY_PORT:
+	case CONNECTOR_ID_USBC:
+		if (record->mini_type == MINI_TYPE_NORMAL) {
+			slot_layout_info->connectors[i].connector_type = CONNECTOR_LAYOUT_TYPE_DP;
+			slot_layout_info->connectors[i].length = CONNECTOR_SIZE_DP;
+		} else {
+			slot_layout_info->connectors[i].connector_type = CONNECTOR_LAYOUT_TYPE_MINI_DP;
+			slot_layout_info->connectors[i].length = CONNECTOR_SIZE_MINI_DP;
+		}
+		break;
+
+	default:
+		slot_layout_info->connectors[i].connector_type = CONNECTOR_LAYOUT_TYPE_UNKNOWN;
+		slot_layout_info->connectors[i].length = CONNECTOR_SIZE_UNKNOWN;
+	}
+	return result;
+}
 
 static enum bp_result get_bracket_layout_record(
 	struct dc_bios *dcb,
@@ -2623,27 +3445,37 @@ static enum bp_result get_bracket_layout_record(
 {
 	unsigned int i;
 	struct bios_parser *bp = BP_FROM_DCB(dcb);
-	enum bp_result result;
+	static enum bp_result result;
 	struct object_info_table *tbl;
 	struct display_object_info_table_v1_4 *v1_4;
+	struct display_object_info_table_v1_5 *v1_5;
 
 	if (slot_layout_info == NULL) {
 		DC_LOG_DETECTION_EDID_PARSER("Invalid slot_layout_info\n");
 		return BP_RESULT_BADINPUT;
 	}
+
 	tbl = &bp->object_info_tbl;
 	v1_4 = tbl->v1_4;
+	v1_5 = tbl->v1_5;
 
 	result = BP_RESULT_NORECORD;
-	for (i = 0; i < v1_4->number_of_path; ++i)	{
-
-		if (bracket_layout_id ==
-			v1_4->display_path[i].display_objid) {
-			result = update_slot_layout_info(dcb, i,
-				slot_layout_info);
-			break;
+	switch (bp->object_info_tbl.revision.minor) {
+	case 4:
+	default:
+		for (i = 0; i < v1_4->number_of_path; ++i) {
+			if (bracket_layout_id == v1_4->display_path[i].display_objid) {
+				result = update_slot_layout_info(dcb, i, slot_layout_info);
+				break;
+			}
 		}
+		break;
+	case 5:
+		for (i = 0; i < v1_5->number_of_path; ++i)
+			result = update_slot_layout_info_v2(dcb, i, slot_layout_info);
+		break;
 	}
+
 	return result;
 }
 
@@ -2652,7 +3484,9 @@ static enum bp_result bios_get_board_layout_info(
 	struct board_layout_info *board_layout_info)
 {
 	unsigned int i;
-	enum bp_result record_result;
+	struct bios_parser *bp;
+	static enum bp_result record_result;
+	unsigned int max_slots;
 
 	const unsigned int slot_index_to_vbios_id[MAX_BOARD_SLOTS] = {
 		GENERICOBJECT_BRACKET_LAYOUT_ENUM_ID1,
@@ -2660,14 +3494,22 @@ static enum bp_result bios_get_board_layout_info(
 		0, 0
 	};
 
+	bp = BP_FROM_DCB(dcb);
+
 	if (board_layout_info == NULL) {
 		DC_LOG_DETECTION_EDID_PARSER("Invalid board_layout_info\n");
 		return BP_RESULT_BADINPUT;
 	}
 
 	board_layout_info->num_of_slots = 0;
+	max_slots = MAX_BOARD_SLOTS;
+
+	// Assume single slot on v1_5
+	if (bp->object_info_tbl.revision.minor == 5) {
+		max_slots = 1;
+	}
 
-	for (i = 0; i < MAX_BOARD_SLOTS; ++i) {
+	for (i = 0; i < max_slots; ++i) {
 		record_result = get_bracket_layout_record(dcb,
 			slot_index_to_vbios_id[i],
 			&board_layout_info->slots[i]);
@@ -2694,99 +3536,6 @@ static uint16_t bios_parser_pack_data_tables(
 	struct dc_bios *dcb,
 	void *dst)
 {
-#ifdef PACK_BIOS_DATA
-	struct bios_parser *bp = BP_FROM_DCB(dcb);
-	struct atom_rom_header_v2_2 *rom_header = NULL;
-	struct atom_rom_header_v2_2 *packed_rom_header = NULL;
-	struct atom_common_table_header *data_tbl_header = NULL;
-	struct atom_master_list_of_data_tables_v2_1 *data_tbl_list = NULL;
-	struct atom_master_data_table_v2_1 *packed_master_data_tbl = NULL;
-	struct atom_data_revision tbl_rev = {0};
-	uint16_t *rom_header_offset = NULL;
-	const uint8_t *bios = bp->base.bios;
-	uint8_t *bios_dst = (uint8_t *)dst;
-	uint16_t packed_rom_header_offset;
-	uint16_t packed_masterdatatable_offset;
-	uint16_t packed_data_tbl_offset;
-	uint16_t data_tbl_offset;
-	unsigned int i;
-
-	rom_header_offset =
-		GET_IMAGE(uint16_t, OFFSET_TO_ATOM_ROM_HEADER_POINTER);
-
-	if (!rom_header_offset)
-		return 0;
-
-	rom_header = GET_IMAGE(struct atom_rom_header_v2_2, *rom_header_offset);
-
-	if (!rom_header)
-		return 0;
-
-	get_atom_data_table_revision(&rom_header->table_header, &tbl_rev);
-	if (!(tbl_rev.major >= 2 && tbl_rev.minor >= 2))
-		return 0;
-
-	get_atom_data_table_revision(&bp->master_data_tbl->table_header, &tbl_rev);
-	if (!(tbl_rev.major >= 2 && tbl_rev.minor >= 1))
-		return 0;
-
-	packed_rom_header_offset =
-		OFFSET_TO_ATOM_ROM_HEADER_POINTER + sizeof(*rom_header_offset);
-
-	packed_masterdatatable_offset =
-		packed_rom_header_offset + rom_header->table_header.structuresize;
-
-	packed_data_tbl_offset =
-		packed_masterdatatable_offset +
-		bp->master_data_tbl->table_header.structuresize;
-
-	packed_rom_header =
-		(struct atom_rom_header_v2_2 *)(bios_dst + packed_rom_header_offset);
-
-	packed_master_data_tbl =
-		(struct atom_master_data_table_v2_1 *)(bios_dst +
-		packed_masterdatatable_offset);
-
-	memcpy(bios_dst, bios, OFFSET_TO_ATOM_ROM_HEADER_POINTER);
-
-	*((uint16_t *)(bios_dst + OFFSET_TO_ATOM_ROM_HEADER_POINTER)) =
-		packed_rom_header_offset;
-
-	memcpy(bios_dst + packed_rom_header_offset, rom_header,
-		rom_header->table_header.structuresize);
-
-	packed_rom_header->masterdatatable_offset = packed_masterdatatable_offset;
-
-	memcpy(&packed_master_data_tbl->table_header,
-		&bp->master_data_tbl->table_header,
-		sizeof(bp->master_data_tbl->table_header));
-
-	data_tbl_list = &bp->master_data_tbl->listOfdatatables;
-
-	/* Each data table offset in data table list is 2 bytes,
-	 * we can use that to iterate through listOfdatatables
-	 * without knowing the name of each member.
-	 */
-	for (i = 0; i < sizeof(*data_tbl_list)/sizeof(uint16_t); i++) {
-		data_tbl_offset = *((uint16_t *)data_tbl_list + i);
-
-		if (data_tbl_offset) {
-			data_tbl_header =
-				(struct atom_common_table_header *)(bios + data_tbl_offset);
-
-			memcpy(bios_dst + packed_data_tbl_offset, data_tbl_header,
-				data_tbl_header->structuresize);
-
-			*((uint16_t *)&packed_master_data_tbl->listOfdatatables + i) =
-				packed_data_tbl_offset;
-
-			packed_data_tbl_offset += data_tbl_header->structuresize;
-		} else {
-			*((uint16_t *)&packed_master_data_tbl->listOfdatatables + i) = 0;
-		}
-	}
-	return packed_data_tbl_offset;
-#endif
 	// TODO: There is data bytes alignment issue, disable it for now.
 	return 0;
 }
@@ -2816,6 +3565,13 @@ static struct atom_dc_golden_table_v1 *bios_get_golden_table(
 			dc_golden_offset = DATA_TABLES(dce_info) + disp_cntl_tbl_4_4->dc_golden_table_offset;
 			*dc_golden_table_ver = disp_cntl_tbl_4_4->dc_golden_table_ver;
 			break;
+		case 5:
+		default:
+			/* For atom_display_controller_info_v4_5 there is no need to get golden table from
+			 * dc_golden_table_offset as all these fields previously in golden table used for AUX
+			 * pre-charge settings are now available directly in atom_display_controller_info_v4_5.
+			 */
+			break;
 		}
 		break;
 	}
@@ -2931,6 +3687,8 @@ static const struct dc_vbios_funcs vbios_funcs = {
 	.get_lttpr_caps = bios_parser_get_lttpr_caps,
 
 	.get_lttpr_interop = bios_parser_get_lttpr_interop,
+
+	.get_connector_speed_cap_info = bios_parser_get_connector_speed_cap_info,
 };
 
 static bool bios_parser2_construct(
@@ -2995,7 +3753,7 @@ static bool bios_parser2_construct(
 		&bp->object_info_tbl.revision);
 
 	if (bp->object_info_tbl.revision.major == 1
-		&& bp->object_info_tbl.revision.minor >= 4) {
+		&& bp->object_info_tbl.revision.minor == 4) {
 		struct display_object_info_table_v1_4 *tbl_v1_4;
 
 		tbl_v1_4 = GET_IMAGE(struct display_object_info_table_v1_4,
@@ -3004,8 +3762,20 @@ static bool bios_parser2_construct(
 			return false;
 
 		bp->object_info_tbl.v1_4 = tbl_v1_4;
-	} else
+	} else if (bp->object_info_tbl.revision.major == 1
+		&& bp->object_info_tbl.revision.minor == 5) {
+		struct display_object_info_table_v1_5 *tbl_v1_5;
+
+		tbl_v1_5 = GET_IMAGE(struct display_object_info_table_v1_5,
+			bp->object_info_tbl_offset);
+		if (!tbl_v1_5)
+			return false;
+
+		bp->object_info_tbl.v1_5 = tbl_v1_5;
+	} else {
+		ASSERT(0);
 		return false;
+	}
 
 	dal_firmware_parser_init_cmd_tbl(bp);
 	dal_bios_parser_init_cmd_tbl_helper2(&bp->cmd_helper, dce_version);
@@ -3013,7 +3783,7 @@ static bool bios_parser2_construct(
 	bp->base.integrated_info = bios_parser_create_integrated_info(&bp->base);
 	bp->base.fw_info_valid = bios_parser_get_firmware_info(&bp->base, &bp->base.fw_info) == BP_RESULT_OK;
 	bios_parser_get_vram_info(&bp->base, &bp->base.vram_info);
-
+	bios_parser_get_soc_bb_info(&bp->base, &bp->base.bb_info);
 	return true;
 }
 
@@ -3021,7 +3791,7 @@ struct dc_bios *firmware_parser_create(
 	struct bp_init_data *init,
 	enum dce_version dce_version)
 {
-	struct bios_parser *bp = NULL;
+	struct bios_parser *bp;
 
 	bp = kzalloc(sizeof(struct bios_parser), GFP_KERNEL);
 	if (!bp)
diff --git a/drivers/gpu/drm/amd/display/dc/bios/bios_parser_common.c b/drivers/gpu/drm/amd/display/dc/bios/bios_parser_common.c
index a8cb039d2572..34e3a64f556e 100644
--- a/drivers/gpu/drm/amd/display/dc/bios/bios_parser_common.c
+++ b/drivers/gpu/drm/amd/display/dc/bios/bios_parser_common.c
@@ -213,6 +213,9 @@ static enum connector_id connector_id_from_bios_object_id(
 	case CONNECTOR_OBJECT_ID_MXM:
 		id = CONNECTOR_ID_MXM;
 		break;
+	case CONNECTOR_OBJECT_ID_USBC:
+		id = CONNECTOR_ID_USBC;
+		break;
 	default:
 		id = CONNECTOR_ID_UNKNOWN;
 		break;
diff --git a/drivers/gpu/drm/amd/display/dc/bios/bios_parser_helper.c b/drivers/gpu/drm/amd/display/dc/bios/bios_parser_helper.c
index adc710fe4a45..8d2cf95ae739 100644
--- a/drivers/gpu/drm/amd/display/dc/bios/bios_parser_helper.c
+++ b/drivers/gpu/drm/amd/display/dc/bios/bios_parser_helper.c
@@ -78,10 +78,3 @@ void bios_set_scratch_critical_state(
 	uint32_t critial_state = state ? 1 : 0;
 	REG_UPDATE(BIOS_SCRATCH_6, S6_CRITICAL_STATE, critial_state);
 }
-
-uint32_t bios_get_vga_enabled_displays(
-	struct dc_bios *bios)
-{
-	return REG_READ(BIOS_SCRATCH_3) & 0XFFFF;
-}
-
diff --git a/drivers/gpu/drm/amd/display/dc/bios/bios_parser_helper.h b/drivers/gpu/drm/amd/display/dc/bios/bios_parser_helper.h
index e1b4a40a353d..ab162f2fe577 100644
--- a/drivers/gpu/drm/amd/display/dc/bios/bios_parser_helper.h
+++ b/drivers/gpu/drm/amd/display/dc/bios/bios_parser_helper.h
@@ -34,7 +34,6 @@ uint8_t *bios_get_image(struct dc_bios *bp, uint32_t offset,
 bool bios_is_accelerated_mode(struct dc_bios *bios);
 void bios_set_scratch_acc_mode_change(struct dc_bios *bios, uint32_t state);
 void bios_set_scratch_critical_state(struct dc_bios *bios, bool state);
-uint32_t bios_get_vga_enabled_displays(struct dc_bios *bios);
 
 #define GET_IMAGE(type, offset) ((type *) bios_get_image(&bp->base, offset, sizeof(type)))
 
diff --git a/drivers/gpu/drm/amd/display/dc/bios/bios_parser_types_internal2.h b/drivers/gpu/drm/amd/display/dc/bios/bios_parser_types_internal2.h
index bf1f5c86e65c..41d02d473082 100644
--- a/drivers/gpu/drm/amd/display/dc/bios/bios_parser_types_internal2.h
+++ b/drivers/gpu/drm/amd/display/dc/bios/bios_parser_types_internal2.h
@@ -40,6 +40,7 @@ struct object_info_table {
 	struct atom_data_revision revision;
 	union {
 		struct display_object_info_table_v1_4 *v1_4;
+		struct display_object_info_table_v1_5 *v1_5;
 	};
 };
 
diff --git a/drivers/gpu/drm/amd/display/dc/bios/command_table.c b/drivers/gpu/drm/amd/display/dc/bios/command_table.c
index ad13e4e36d77..22457f417e65 100644
--- a/drivers/gpu/drm/amd/display/dc/bios/command_table.c
+++ b/drivers/gpu/drm/amd/display/dc/bios/command_table.c
@@ -37,7 +37,7 @@
 #define EXEC_BIOS_CMD_TABLE(command, params)\
 	(amdgpu_atom_execute_table(((struct amdgpu_device *)bp->base.ctx->driver_context)->mode_info.atom_context, \
 		GetIndexIntoMasterTable(COMMAND, command), \
-		(uint32_t *)&params) == 0)
+		(uint32_t *)&params, sizeof(params)) == 0)
 
 #define BIOS_CMD_TABLE_REVISION(command, frev, crev)\
 	amdgpu_atom_parse_cmd_header(((struct amdgpu_device *)bp->base.ctx->driver_context)->mode_info.atom_context, \
@@ -52,7 +52,9 @@ static void init_transmitter_control(struct bios_parser *bp);
 static void init_set_pixel_clock(struct bios_parser *bp);
 static void init_enable_spread_spectrum_on_ppll(struct bios_parser *bp);
 static void init_adjust_display_pll(struct bios_parser *bp);
+static void init_select_crtc_source(struct bios_parser *bp);
 static void init_dac_encoder_control(struct bios_parser *bp);
+static void init_dac_load_detection(struct bios_parser *bp);
 static void init_dac_output_control(struct bios_parser *bp);
 static void init_set_crtc_timing(struct bios_parser *bp);
 static void init_enable_crtc(struct bios_parser *bp);
@@ -69,7 +71,9 @@ void dal_bios_parser_init_cmd_tbl(struct bios_parser *bp)
 	init_set_pixel_clock(bp);
 	init_enable_spread_spectrum_on_ppll(bp);
 	init_adjust_display_pll(bp);
+	init_select_crtc_source(bp);
 	init_dac_encoder_control(bp);
+	init_dac_load_detection(bp);
 	init_dac_output_control(bp);
 	init_set_crtc_timing(bp);
 	init_enable_crtc(bp);
@@ -399,7 +403,7 @@ static enum bp_result transmitter_control_v1_6(
 static void init_transmitter_control(struct bios_parser *bp)
 {
 	uint8_t frev;
-	uint8_t crev;
+	uint8_t crev = 0;
 
 	if (BIOS_CMD_TABLE_REVISION(UNIPHYTransmitterControl,
 			frev, crev) == false)
@@ -456,7 +460,7 @@ static enum bp_result transmitter_control_v2(
 		if ((CONNECTOR_ID_DUAL_LINK_DVII == connector_id) ||
 				(CONNECTOR_ID_DUAL_LINK_DVID == connector_id))
 			/* on INIT this bit should be set according to the
-			 * phisycal connector
+			 * physical connector
 			 * Bit0: dual link connector flag
 			 * =0 connector is single link connector
 			 * =1 connector is dual link connector
@@ -468,7 +472,7 @@ static enum bp_result transmitter_control_v2(
 				cpu_to_le16((uint8_t)cntl->connector_obj_id.id);
 		break;
 	case TRANSMITTER_CONTROL_SET_VOLTAGE_AND_PREEMPASIS:
-		/* votage swing and pre-emphsis */
+		/* voltage swing and pre-emphsis */
 		params.asMode.ucLaneSel = (uint8_t)cntl->lane_select;
 		params.asMode.ucLaneSet = (uint8_t)cntl->lane_settings;
 		break;
@@ -522,7 +526,8 @@ static enum bp_result transmitter_control_v2(
 		 */
 		params.acConfig.ucEncoderSel = 1;
 
-	if (CONNECTOR_ID_DISPLAY_PORT == connector_id)
+	if (CONNECTOR_ID_DISPLAY_PORT == connector_id ||
+	    CONNECTOR_ID_USBC == connector_id)
 		/* Bit4: DP connector flag
 		 * =0 connector is none-DP connector
 		 * =1 connector is DP connector
@@ -992,7 +997,7 @@ static enum bp_result set_pixel_clock_v3(
 	allocation.sPCLKInput.usFbDiv =
 			cpu_to_le16((uint16_t)bp_params->feedback_divider);
 	allocation.sPCLKInput.ucFracFbDiv =
-			(uint8_t)bp_params->fractional_feedback_divider;
+			(uint8_t)(bp_params->fractional_feedback_divider / 100000);
 	allocation.sPCLKInput.ucPostDiv =
 			(uint8_t)bp_params->pixel_clock_post_divider;
 
@@ -1611,6 +1616,198 @@ static enum bp_result adjust_display_pll_v3(
 /*******************************************************************************
  ********************************************************************************
  **
+ **                  SELECT CRTC SOURCE
+ **
+ ********************************************************************************
+ *******************************************************************************/
+
+static enum bp_result select_crtc_source_v1(
+	struct bios_parser *bp,
+	struct bp_crtc_source_select *bp_params);
+static enum bp_result select_crtc_source_v2(
+	struct bios_parser *bp,
+	struct bp_crtc_source_select *bp_params);
+static enum bp_result select_crtc_source_v3(
+	struct bios_parser *bp,
+	struct bp_crtc_source_select *bp_params);
+
+static void init_select_crtc_source(struct bios_parser *bp)
+{
+	switch (BIOS_CMD_TABLE_PARA_REVISION(SelectCRTC_Source)) {
+	case 1:
+		bp->cmd_tbl.select_crtc_source = select_crtc_source_v1;
+		break;
+	case 2:
+		bp->cmd_tbl.select_crtc_source = select_crtc_source_v2;
+		break;
+	case 3:
+		bp->cmd_tbl.select_crtc_source = select_crtc_source_v3;
+		break;
+	default:
+		bp->cmd_tbl.select_crtc_source = NULL;
+		break;
+	}
+}
+
+static enum bp_result select_crtc_source_v1(
+	struct bios_parser *bp,
+	struct bp_crtc_source_select *bp_params)
+{
+	enum bp_result result = BP_RESULT_FAILURE;
+	SELECT_CRTC_SOURCE_PS_ALLOCATION params;
+
+	if (!bp->cmd_helper->controller_id_to_atom(bp_params->controller_id, &params.ucCRTC))
+		return BP_RESULT_BADINPUT;
+
+	switch (bp_params->engine_id) {
+	case ENGINE_ID_DACA:
+		params.ucDevice = ATOM_DEVICE_CRT1_INDEX;
+		break;
+	case ENGINE_ID_DACB:
+		params.ucDevice = ATOM_DEVICE_CRT2_INDEX;
+		break;
+	default:
+		return BP_RESULT_BADINPUT;
+	}
+
+	if (EXEC_BIOS_CMD_TABLE(SelectCRTC_Source, params))
+		result = BP_RESULT_OK;
+
+	return result;
+}
+
+static bool select_crtc_source_v2_encoder_id(
+	enum engine_id engine_id, uint8_t *out_encoder_id)
+{
+	uint8_t encoder_id = 0;
+
+	switch (engine_id) {
+	case ENGINE_ID_DIGA:
+		encoder_id = ASIC_INT_DIG1_ENCODER_ID;
+		break;
+	case ENGINE_ID_DIGB:
+		encoder_id = ASIC_INT_DIG2_ENCODER_ID;
+		break;
+	case ENGINE_ID_DIGC:
+		encoder_id = ASIC_INT_DIG3_ENCODER_ID;
+		break;
+	case ENGINE_ID_DIGD:
+		encoder_id = ASIC_INT_DIG4_ENCODER_ID;
+		break;
+	case ENGINE_ID_DIGE:
+		encoder_id = ASIC_INT_DIG5_ENCODER_ID;
+		break;
+	case ENGINE_ID_DIGF:
+		encoder_id = ASIC_INT_DIG6_ENCODER_ID;
+		break;
+	case ENGINE_ID_DIGG:
+		encoder_id = ASIC_INT_DIG7_ENCODER_ID;
+		break;
+	case ENGINE_ID_DACA:
+		encoder_id = ASIC_INT_DAC1_ENCODER_ID;
+		break;
+	case ENGINE_ID_DACB:
+		encoder_id = ASIC_INT_DAC2_ENCODER_ID;
+		break;
+	default:
+		return false;
+	}
+
+	*out_encoder_id = encoder_id;
+	return true;
+}
+
+static bool select_crtc_source_v2_encoder_mode(
+	enum signal_type signal_type, uint8_t *out_encoder_mode)
+{
+	uint8_t encoder_mode = 0;
+
+	switch (signal_type) {
+	case SIGNAL_TYPE_DVI_SINGLE_LINK:
+	case SIGNAL_TYPE_DVI_DUAL_LINK:
+		encoder_mode = ATOM_ENCODER_MODE_DVI;
+		break;
+	case SIGNAL_TYPE_HDMI_TYPE_A:
+		encoder_mode = ATOM_ENCODER_MODE_HDMI;
+		break;
+	case SIGNAL_TYPE_LVDS:
+		encoder_mode = ATOM_ENCODER_MODE_LVDS;
+		break;
+	case SIGNAL_TYPE_RGB:
+		encoder_mode = ATOM_ENCODER_MODE_CRT;
+		break;
+	case SIGNAL_TYPE_DISPLAY_PORT:
+		encoder_mode = ATOM_ENCODER_MODE_DP;
+		break;
+	case SIGNAL_TYPE_DISPLAY_PORT_MST:
+		encoder_mode = ATOM_ENCODER_MODE_DP_MST;
+		break;
+	case SIGNAL_TYPE_EDP:
+		encoder_mode = ATOM_ENCODER_MODE_DP;
+		break;
+	default:
+		return false;
+	}
+
+	*out_encoder_mode = encoder_mode;
+	return true;
+}
+
+static enum bp_result select_crtc_source_v2(
+	struct bios_parser *bp,
+	struct bp_crtc_source_select *bp_params)
+{
+	enum bp_result result = BP_RESULT_FAILURE;
+	SELECT_CRTC_SOURCE_PARAMETERS_V3 params;
+
+	if (!bp->cmd_helper->controller_id_to_atom(bp_params->controller_id, &params.ucCRTC))
+		return BP_RESULT_BADINPUT;
+
+	if (!select_crtc_source_v2_encoder_id(
+		bp_params->engine_id,
+		&params.ucEncoderID))
+		return BP_RESULT_BADINPUT;
+	if (!select_crtc_source_v2_encoder_mode(
+		bp_params->sink_signal,
+		&params.ucEncodeMode))
+		return BP_RESULT_BADINPUT;
+
+	if (EXEC_BIOS_CMD_TABLE(SelectCRTC_Source, params))
+		result = BP_RESULT_OK;
+
+	return result;
+}
+
+static enum bp_result select_crtc_source_v3(
+	struct bios_parser *bp,
+	struct bp_crtc_source_select *bp_params)
+{
+	enum bp_result result = BP_RESULT_FAILURE;
+	SELECT_CRTC_SOURCE_PARAMETERS_V3 params;
+
+	if (!bp->cmd_helper->controller_id_to_atom(bp_params->controller_id, &params.ucCRTC))
+		return BP_RESULT_BADINPUT;
+
+	if (!select_crtc_source_v2_encoder_id(
+		bp_params->engine_id,
+		&params.ucEncoderID))
+		return BP_RESULT_BADINPUT;
+	if (!select_crtc_source_v2_encoder_mode(
+		bp_params->sink_signal,
+		&params.ucEncodeMode))
+		return BP_RESULT_BADINPUT;
+
+	params.ucDstBpc = bp_params->bit_depth;
+
+	if (EXEC_BIOS_CMD_TABLE(SelectCRTC_Source, params))
+		result = BP_RESULT_OK;
+
+	return result;
+}
+
+/*******************************************************************************
+ ********************************************************************************
+ **
  **                  DAC ENCODER CONTROL
  **
  ********************************************************************************
@@ -1710,6 +1907,96 @@ static enum bp_result dac2_encoder_control_v1(
 /*******************************************************************************
  ********************************************************************************
  **
+ **                  DAC LOAD DETECTION
+ **
+ ********************************************************************************
+ *******************************************************************************/
+
+static enum bp_result dac_load_detection_v1(
+	struct bios_parser *bp,
+	struct bp_load_detection_parameters *bp_params);
+
+static enum bp_result dac_load_detection_v3(
+	struct bios_parser *bp,
+	struct bp_load_detection_parameters *bp_params);
+
+static void init_dac_load_detection(struct bios_parser *bp)
+{
+	switch (BIOS_CMD_TABLE_PARA_REVISION(DAC_LoadDetection)) {
+	case 1:
+	case 2:
+		bp->cmd_tbl.dac_load_detection = dac_load_detection_v1;
+		break;
+	case 3:
+	default:
+		bp->cmd_tbl.dac_load_detection = dac_load_detection_v3;
+		break;
+	}
+}
+
+static void dac_load_detect_prepare_params(
+	struct _DAC_LOAD_DETECTION_PS_ALLOCATION *params,
+	enum engine_id engine_id,
+	uint16_t device_id,
+	uint8_t misc)
+{
+	uint8_t dac_type = ENGINE_ID_DACA;
+
+	if (engine_id == ENGINE_ID_DACB)
+		dac_type = ATOM_DAC_B;
+
+	params->sDacload.usDeviceID = cpu_to_le16(device_id);
+	params->sDacload.ucDacType = dac_type;
+	params->sDacload.ucMisc = misc;
+}
+
+static enum bp_result dac_load_detection_v1(
+	struct bios_parser *bp,
+	struct bp_load_detection_parameters *bp_params)
+{
+	enum bp_result result = BP_RESULT_FAILURE;
+	DAC_LOAD_DETECTION_PS_ALLOCATION params;
+
+	dac_load_detect_prepare_params(
+		&params,
+		bp_params->engine_id,
+		bp_params->device_id,
+		0);
+
+	if (EXEC_BIOS_CMD_TABLE(DAC_LoadDetection, params))
+		result = BP_RESULT_OK;
+
+	return result;
+}
+
+static enum bp_result dac_load_detection_v3(
+	struct bios_parser *bp,
+	struct bp_load_detection_parameters *bp_params)
+{
+	enum bp_result result = BP_RESULT_FAILURE;
+	DAC_LOAD_DETECTION_PS_ALLOCATION params;
+
+	uint8_t misc = 0;
+
+	if (bp_params->device_id == ATOM_DEVICE_CV_SUPPORT ||
+	    bp_params->device_id == ATOM_DEVICE_TV1_SUPPORT)
+		misc = DAC_LOAD_MISC_YPrPb;
+
+	dac_load_detect_prepare_params(
+		&params,
+		bp_params->engine_id,
+		bp_params->device_id,
+		misc);
+
+	if (EXEC_BIOS_CMD_TABLE(DAC_LoadDetection, params))
+		result = BP_RESULT_OK;
+
+	return result;
+}
+
+/*******************************************************************************
+ ********************************************************************************
+ **
  **                  DAC OUTPUT CONTROL
  **
  ********************************************************************************
@@ -2120,7 +2407,7 @@ static enum bp_result program_clock_v5(
 	memset(&params, 0, sizeof(params));
 	if (!bp->cmd_helper->clock_source_id_to_atom(
 			bp_params->pll_id, &atom_pll_id)) {
-		BREAK_TO_DEBUGGER(); /* Invalid Inpute!! */
+		BREAK_TO_DEBUGGER(); /* Invalid Input!! */
 		return BP_RESULT_BADINPUT;
 	}
 
diff --git a/drivers/gpu/drm/amd/display/dc/bios/command_table.h b/drivers/gpu/drm/amd/display/dc/bios/command_table.h
index ad533775e724..e89b1ba0048b 100644
--- a/drivers/gpu/drm/amd/display/dc/bios/command_table.h
+++ b/drivers/gpu/drm/amd/display/dc/bios/command_table.h
@@ -52,6 +52,9 @@ struct cmd_tbl {
 	enum bp_result (*adjust_display_pll)(
 		struct bios_parser *bp,
 		struct bp_adjust_pixel_clock_parameters *bp_params);
+	enum bp_result (*select_crtc_source)(
+		struct bios_parser *bp,
+		struct bp_crtc_source_select *bp_params);
 	enum bp_result (*dac1_encoder_control)(
 		struct bios_parser *bp,
 		bool enable,
@@ -68,6 +71,9 @@ struct cmd_tbl {
 	enum bp_result (*dac2_output_control)(
 		struct bios_parser *bp,
 		bool enable);
+	enum bp_result (*dac_load_detection)(
+		struct bios_parser *bp,
+		struct bp_load_detection_parameters *bp_params);
 	enum bp_result (*set_crtc_timing)(
 		struct bios_parser *bp,
 		struct bp_hw_crtc_timing_parameters *bp_params);
diff --git a/drivers/gpu/drm/amd/display/dc/bios/command_table2.c b/drivers/gpu/drm/amd/display/dc/bios/command_table2.c
index 9afa5eb2e6d3..f2b1720a6a66 100644
--- a/drivers/gpu/drm/amd/display/dc/bios/command_table2.c
+++ b/drivers/gpu/drm/amd/display/dc/bios/command_table2.c
@@ -49,7 +49,7 @@
 #define EXEC_BIOS_CMD_TABLE(fname, params)\
 	(amdgpu_atom_execute_table(((struct amdgpu_device *)bp->base.ctx->driver_context)->mode_info.atom_context, \
 		GET_INDEX_INTO_MASTER_TABLE(command, fname), \
-		(uint32_t *)&params) == 0)
+		(uint32_t *)&params, sizeof(params)) == 0)
 
 #define BIOS_CMD_TABLE_REVISION(fname, frev, crev)\
 	amdgpu_atom_parse_cmd_header(((struct amdgpu_device *)bp->base.ctx->driver_context)->mode_info.atom_context, \
@@ -101,7 +101,6 @@ static void init_dig_encoder_control(struct bios_parser *bp)
 		bp->cmd_tbl.dig_encoder_control = encoder_control_digx_v1_5;
 		break;
 	default:
-		dm_output_to_console("Don't have dig_encoder_control for v%d\n", version);
 		bp->cmd_tbl.dig_encoder_control = encoder_control_fallback;
 		break;
 	}
@@ -123,9 +122,7 @@ static void encoder_control_dmcub(
 		sizeof(cmd.digx_encoder_control.header);
 	cmd.digx_encoder_control.encoder_control.dig.stream_param = *dig;
 
-	dc_dmub_srv_cmd_queue(dmcub, &cmd);
-	dc_dmub_srv_cmd_execute(dmcub);
-	dc_dmub_srv_wait_idle(dmcub);
+	dc_wake_and_execute_dmub_cmd(dmcub->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
 }
 
 static enum bp_result encoder_control_digx_v1_5(
@@ -212,6 +209,7 @@ static enum bp_result encoder_control_fallback(
  ******************************************************************************
  *****************************************************************************/
 
+
 static enum bp_result transmitter_control_v1_6(
 	struct bios_parser *bp,
 	struct bp_transmitter_control *cntl);
@@ -227,9 +225,10 @@ static enum bp_result transmitter_control_fallback(
 static void init_transmitter_control(struct bios_parser *bp)
 {
 	uint8_t frev;
-	uint8_t crev;
+	uint8_t crev = 0;
 
-	BIOS_CMD_TABLE_REVISION(dig1transmittercontrol, frev, crev);
+	if (!BIOS_CMD_TABLE_REVISION(dig1transmittercontrol, frev, crev) && (bp->base.ctx->dc->ctx->dce_version <= DCN_VERSION_2_0))
+		BREAK_TO_DEBUGGER();
 
 	switch (crev) {
 	case 6:
@@ -239,7 +238,6 @@ static void init_transmitter_control(struct bios_parser *bp)
 		bp->cmd_tbl.transmitter_control = transmitter_control_v1_7;
 		break;
 	default:
-		dm_output_to_console("Don't have transmitter_control for v%d\n", crev);
 		bp->cmd_tbl.transmitter_control = transmitter_control_fallback;
 		break;
 	}
@@ -261,9 +259,7 @@ static void transmitter_control_dmcub(
 		sizeof(cmd.dig1_transmitter_control.header);
 	cmd.dig1_transmitter_control.transmitter_control.dig = *dig;
 
-	dc_dmub_srv_cmd_queue(dmcub, &cmd);
-	dc_dmub_srv_cmd_execute(dmcub);
-	dc_dmub_srv_wait_idle(dmcub);
+	dc_wake_and_execute_dmub_cmd(dmcub->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
 }
 
 static enum bp_result transmitter_control_v1_6(
@@ -325,9 +321,22 @@ static void transmitter_control_dmcub_v1_7(
 		sizeof(cmd.dig1_transmitter_control.header);
 	cmd.dig1_transmitter_control.transmitter_control.dig_v1_7 = *dig;
 
-	dc_dmub_srv_cmd_queue(dmcub, &cmd);
-	dc_dmub_srv_cmd_execute(dmcub);
-	dc_dmub_srv_wait_idle(dmcub);
+	dc_wake_and_execute_dmub_cmd(dmcub->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
+}
+
+static struct dc_link *get_link_by_phy_id(struct dc *p_dc, uint32_t phy_id)
+{
+	struct dc_link *link = NULL;
+
+	// Get Transition Bitmask from dc_link structure associated with PHY
+	for (uint8_t link_id = 0; link_id < MAX_LINKS; link_id++) {
+		if (phy_id == p_dc->links[link_id]->link_enc->transmitter) {
+			link = p_dc->links[link_id];
+			break;
+		}
+	}
+
+	return link;
 }
 
 static enum bp_result transmitter_control_v1_7(
@@ -338,12 +347,10 @@ static enum bp_result transmitter_control_v1_7(
 	const struct command_table_helper *cmd = bp->cmd_helper;
 	struct dmub_dig_transmitter_control_data_v1_7 dig_v1_7 = {0};
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	uint8_t hpo_instance = (uint8_t)cntl->hpo_engine_id - ENGINE_ID_HPO_0;
 
 	if (dc_is_dp_signal(cntl->signal))
 		hpo_instance = (uint8_t)cntl->hpo_engine_id - ENGINE_ID_HPO_DP_0;
-#endif
 
 	dig_v1_7.phyid = cmd->phy_id_to_atom(cntl->transmitter);
 	dig_v1_7.action = (uint8_t)cntl->action;
@@ -358,9 +365,7 @@ static enum bp_result transmitter_control_v1_7(
 	dig_v1_7.hpdsel = cmd->hpd_sel_to_atom(cntl->hpd_sel);
 	dig_v1_7.digfe_sel = cmd->dig_encoder_sel_to_atom(cntl->engine_id);
 	dig_v1_7.connobj_id = (uint8_t)cntl->connector_obj_id.id;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	dig_v1_7.HPO_instance = hpo_instance;
-#endif
 	dig_v1_7.symclk_units.symclk_10khz = cntl->pixel_clock/10;
 
 	if (cntl->action == TRANSMITTER_CONTROL_ENABLE ||
@@ -372,7 +377,38 @@ static enum bp_result transmitter_control_v1_7(
 
 	if (bp->base.ctx->dc->ctx->dmub_srv &&
 		bp->base.ctx->dc->debug.dmub_command_table) {
+		struct dm_process_phy_transition_init_params process_phy_transition_init_params = {0};
+		struct dc_link *link = get_link_by_phy_id(bp->base.ctx->dc, dig_v1_7.phyid);
+		bool is_phy_transition_interlock_allowed = false;
+		uint8_t action = dig_v1_7.action;
+
+		if (link) {
+			if (link->phy_transition_bitmask &&
+				(action == TRANSMITTER_CONTROL_ENABLE || action == TRANSMITTER_CONTROL_DISABLE)) {
+				is_phy_transition_interlock_allowed = true;
+
+				// Prepare input parameters for processing ACPI retimers
+				process_phy_transition_init_params.action                   = action;
+				process_phy_transition_init_params.display_port_lanes_count = cntl->lanes_number;
+				process_phy_transition_init_params.phy_id                   = dig_v1_7.phyid;
+				process_phy_transition_init_params.signal                   = cntl->signal;
+				process_phy_transition_init_params.sym_clock_10khz          = dig_v1_7.symclk_units.symclk_10khz;
+				process_phy_transition_init_params.display_port_link_rate   = link->cur_link_settings.link_rate;
+				process_phy_transition_init_params.transition_bitmask       = link->phy_transition_bitmask;
+			}
+			dig_v1_7.skip_phy_ssc_reduction = link->wa_flags.skip_phy_ssc_reduction;
+		}
+
+		// Handle PRE_OFF_TO_ON: Process ACPI PHY Transition Interlock
+		if (is_phy_transition_interlock_allowed && action == TRANSMITTER_CONTROL_ENABLE)
+			dm_acpi_process_phy_transition_interlock(bp->base.ctx, process_phy_transition_init_params);
+
 		transmitter_control_dmcub_v1_7(bp->base.ctx->dmub_srv, &dig_v1_7);
+
+		// Handle POST_ON_TO_OFF: Process ACPI PHY Transition Interlock
+		if (is_phy_transition_interlock_allowed && action == TRANSMITTER_CONTROL_DISABLE)
+			dm_acpi_process_phy_transition_interlock(bp->base.ctx, process_phy_transition_init_params);
+
 		return BP_RESULT_OK;
 	}
 
@@ -417,8 +453,6 @@ static void init_set_pixel_clock(struct bios_parser *bp)
 		bp->cmd_tbl.set_pixel_clock = set_pixel_clock_v7;
 		break;
 	default:
-		dm_output_to_console("Don't have set_pixel_clock for v%d\n",
-			 BIOS_CMD_TABLE_PARA_REVISION(setpixelclock));
 		bp->cmd_tbl.set_pixel_clock = set_pixel_clock_fallback;
 		break;
 	}
@@ -439,9 +473,7 @@ static void set_pixel_clock_dmcub(
 		sizeof(cmd.set_pixel_clock.header);
 	cmd.set_pixel_clock.pixel_clock.clk = *clk;
 
-	dc_dmub_srv_cmd_queue(dmcub, &cmd);
-	dc_dmub_srv_cmd_execute(dmcub);
-	dc_dmub_srv_wait_idle(dmcub);
+	dc_wake_and_execute_dmub_cmd(dmcub->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
 }
 
 static enum bp_result set_pixel_clock_v7(
@@ -565,7 +597,6 @@ static void init_set_crtc_timing(struct bios_parser *bp)
 			set_crtc_using_dtd_timing_v3;
 		break;
 	default:
-		dm_output_to_console("Don't have set_crtc_timing for v%d\n", dtd_version);
 		bp->cmd_tbl.set_crtc_timing = NULL;
 		break;
 	}
@@ -682,8 +713,6 @@ static void init_enable_crtc(struct bios_parser *bp)
 		bp->cmd_tbl.enable_crtc = enable_crtc_v1;
 		break;
 	default:
-		dm_output_to_console("Don't have enable_crtc for v%d\n",
-			 BIOS_CMD_TABLE_PARA_REVISION(enablecrtc));
 		bp->cmd_tbl.enable_crtc = NULL;
 		break;
 	}
@@ -808,9 +837,7 @@ static void enable_disp_power_gating_dmcub(
 		sizeof(cmd.enable_disp_power_gating.header);
 	cmd.enable_disp_power_gating.power_gating.pwr = *pwr;
 
-	dc_dmub_srv_cmd_queue(dmcub, &cmd);
-	dc_dmub_srv_cmd_execute(dmcub);
-	dc_dmub_srv_wait_idle(dmcub);
+	dc_wake_and_execute_dmub_cmd(dmcub->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
 }
 
 static enum bp_result enable_disp_power_gating_v2_1(
@@ -877,8 +904,6 @@ static void init_set_dce_clock(struct bios_parser *bp)
 		bp->cmd_tbl.set_dce_clock = set_dce_clock_v2_1;
 		break;
 	default:
-		dm_output_to_console("Don't have set_dce_clock for v%d\n",
-			 BIOS_CMD_TABLE_PARA_REVISION(setdceclock));
 		bp->cmd_tbl.set_dce_clock = NULL;
 		break;
 	}
@@ -990,7 +1015,8 @@ static unsigned int get_smu_clock_info_v3_1(struct bios_parser *bp, uint8_t id)
 static enum bp_result enable_lvtma_control(
 	struct bios_parser *bp,
 	uint8_t uc_pwr_on,
-	uint8_t panel_instance);
+	uint8_t pwrseq_instance,
+	uint8_t bypass_panel_control_wait);
 
 static void init_enable_lvtma_control(struct bios_parser *bp)
 {
@@ -1002,7 +1028,8 @@ static void init_enable_lvtma_control(struct bios_parser *bp)
 static void enable_lvtma_control_dmcub(
 	struct dc_dmub_srv *dmcub,
 	uint8_t uc_pwr_on,
-	uint8_t panel_instance)
+	uint8_t pwrseq_instance,
+	uint8_t bypass_panel_control_wait)
 {
 
 	union dmub_rb_cmd cmd;
@@ -1014,18 +1041,18 @@ static void enable_lvtma_control_dmcub(
 			DMUB_CMD__VBIOS_LVTMA_CONTROL;
 	cmd.lvtma_control.data.uc_pwr_action =
 			uc_pwr_on;
-	cmd.lvtma_control.data.panel_inst =
-			panel_instance;
-	dc_dmub_srv_cmd_queue(dmcub, &cmd);
-	dc_dmub_srv_cmd_execute(dmcub);
-	dc_dmub_srv_wait_idle(dmcub);
-
+	cmd.lvtma_control.data.pwrseq_inst =
+			pwrseq_instance;
+	cmd.lvtma_control.data.bypass_panel_control_wait =
+			bypass_panel_control_wait;
+	dc_wake_and_execute_dmub_cmd(dmcub->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
 }
 
 static enum bp_result enable_lvtma_control(
 	struct bios_parser *bp,
 	uint8_t uc_pwr_on,
-	uint8_t panel_instance)
+	uint8_t pwrseq_instance,
+	uint8_t bypass_panel_control_wait)
 {
 	enum bp_result result = BP_RESULT_FAILURE;
 
@@ -1033,7 +1060,8 @@ static enum bp_result enable_lvtma_control(
 	    bp->base.ctx->dc->debug.dmub_command_table) {
 		enable_lvtma_control_dmcub(bp->base.ctx->dmub_srv,
 				uc_pwr_on,
-				panel_instance);
+				pwrseq_instance,
+				bypass_panel_control_wait);
 		return BP_RESULT_OK;
 	}
 	return result;
@@ -1056,3 +1084,4 @@ void dal_firmware_parser_init_cmd_tbl(struct bios_parser *bp)
 
 	init_enable_lvtma_control(bp);
 }
+
diff --git a/drivers/gpu/drm/amd/display/dc/bios/command_table2.h b/drivers/gpu/drm/amd/display/dc/bios/command_table2.h
index be060b4b87db..41c8c014397f 100644
--- a/drivers/gpu/drm/amd/display/dc/bios/command_table2.h
+++ b/drivers/gpu/drm/amd/display/dc/bios/command_table2.h
@@ -96,7 +96,8 @@ struct cmd_tbl {
 			struct bios_parser *bp, uint8_t id);
 	enum bp_result (*enable_lvtma_control)(struct bios_parser *bp,
 			uint8_t uc_pwr_on,
-			uint8_t panel_instance);
+			uint8_t pwrseq_instance,
+			uint8_t bypass_panel_control_wait);
 };
 
 void dal_firmware_parser_init_cmd_tbl(struct bios_parser *bp);
diff --git a/drivers/gpu/drm/amd/display/dc/bios/command_table_helper.c b/drivers/gpu/drm/amd/display/dc/bios/command_table_helper.c
index e317a3615147..91bc8a06e2cf 100644
--- a/drivers/gpu/drm/amd/display/dc/bios/command_table_helper.c
+++ b/drivers/gpu/drm/amd/display/dc/bios/command_table_helper.c
@@ -293,3 +293,107 @@ uint8_t dal_cmd_table_helper_encoder_id_to_atom(
 		return ENCODER_OBJECT_ID_NONE;
 	}
 }
+
+uint8_t phy_id_to_atom(enum transmitter t)
+{
+	uint8_t atom_phy_id;
+
+	switch (t) {
+	case TRANSMITTER_UNIPHY_A:
+		atom_phy_id = ATOM_PHY_ID_UNIPHYA;
+		break;
+	case TRANSMITTER_UNIPHY_B:
+		atom_phy_id = ATOM_PHY_ID_UNIPHYB;
+		break;
+	case TRANSMITTER_UNIPHY_C:
+		atom_phy_id = ATOM_PHY_ID_UNIPHYC;
+		break;
+	case TRANSMITTER_UNIPHY_D:
+		atom_phy_id = ATOM_PHY_ID_UNIPHYD;
+		break;
+	case TRANSMITTER_UNIPHY_E:
+		atom_phy_id = ATOM_PHY_ID_UNIPHYE;
+		break;
+	case TRANSMITTER_UNIPHY_F:
+		atom_phy_id = ATOM_PHY_ID_UNIPHYF;
+		break;
+	case TRANSMITTER_UNIPHY_G:
+		atom_phy_id = ATOM_PHY_ID_UNIPHYG;
+		break;
+	default:
+		atom_phy_id = ATOM_PHY_ID_UNIPHYA;
+		break;
+	}
+	return atom_phy_id;
+}
+
+uint8_t clock_source_id_to_atom_phy_clk_src_id(
+		enum clock_source_id id)
+{
+	uint8_t atom_phy_clk_src_id = 0;
+
+	switch (id) {
+	case CLOCK_SOURCE_ID_PLL0:
+		atom_phy_clk_src_id = ATOM_TRANSMITTER_CONFIG_V5_P0PLL;
+		break;
+	case CLOCK_SOURCE_ID_PLL1:
+		atom_phy_clk_src_id = ATOM_TRANSMITTER_CONFIG_V5_P1PLL;
+		break;
+	case CLOCK_SOURCE_ID_PLL2:
+		atom_phy_clk_src_id = ATOM_TRANSMITTER_CONFIG_V5_P2PLL;
+		break;
+	case CLOCK_SOURCE_ID_EXTERNAL:
+		atom_phy_clk_src_id = ATOM_TRANSMITTER_CONFIG_V5_REFCLK_SRC_EXT;
+		break;
+	default:
+		atom_phy_clk_src_id = ATOM_TRANSMITTER_CONFIG_V5_P1PLL;
+		break;
+	}
+
+	return atom_phy_clk_src_id >> 2;
+}
+
+bool engine_bp_to_atom(enum engine_id id, uint32_t *atom_engine_id)
+{
+	bool result = false;
+
+	if (atom_engine_id != NULL)
+		switch (id) {
+		case ENGINE_ID_DIGA:
+			*atom_engine_id = ASIC_INT_DIG1_ENCODER_ID;
+			result = true;
+			break;
+		case ENGINE_ID_DIGB:
+			*atom_engine_id = ASIC_INT_DIG2_ENCODER_ID;
+			result = true;
+			break;
+		case ENGINE_ID_DIGC:
+			*atom_engine_id = ASIC_INT_DIG3_ENCODER_ID;
+			result = true;
+			break;
+		case ENGINE_ID_DIGD:
+			*atom_engine_id = ASIC_INT_DIG4_ENCODER_ID;
+			result = true;
+			break;
+		case ENGINE_ID_DIGE:
+			*atom_engine_id = ASIC_INT_DIG5_ENCODER_ID;
+			result = true;
+			break;
+		case ENGINE_ID_DIGF:
+			*atom_engine_id = ASIC_INT_DIG6_ENCODER_ID;
+			result = true;
+			break;
+		case ENGINE_ID_DIGG:
+			*atom_engine_id = ASIC_INT_DIG7_ENCODER_ID;
+			result = true;
+			break;
+		case ENGINE_ID_DACA:
+			*atom_engine_id = ASIC_INT_DAC1_ENCODER_ID;
+			result = true;
+			break;
+		default:
+			break;
+		}
+
+	return result;
+}
diff --git a/drivers/gpu/drm/amd/display/dc/bios/command_table_helper.h b/drivers/gpu/drm/amd/display/dc/bios/command_table_helper.h
index dfd30aaf4032..547700e119a6 100644
--- a/drivers/gpu/drm/amd/display/dc/bios/command_table_helper.h
+++ b/drivers/gpu/drm/amd/display/dc/bios/command_table_helper.h
@@ -59,4 +59,12 @@ uint8_t dal_cmd_table_helper_transmitter_bp_to_atom(
 
 uint8_t dal_cmd_table_helper_encoder_id_to_atom(
 	enum encoder_id id);
+
+uint8_t phy_id_to_atom(enum transmitter t);
+
+uint8_t clock_source_id_to_atom_phy_clk_src_id(
+		enum clock_source_id id);
+
+bool engine_bp_to_atom(enum engine_id id, uint32_t *atom_engine_id);
+
 #endif
diff --git a/drivers/gpu/drm/amd/display/dc/bios/command_table_helper2.c b/drivers/gpu/drm/amd/display/dc/bios/command_table_helper2.c
index eedc553f340e..268e2414b34f 100644
--- a/drivers/gpu/drm/amd/display/dc/bios/command_table_helper2.c
+++ b/drivers/gpu/drm/amd/display/dc/bios/command_table_helper2.c
@@ -65,7 +65,6 @@ bool dal_bios_parser_init_cmd_tbl_helper2(
 	case DCE_VERSION_12_1:
 		*h = dal_cmd_tbl_helper_dce112_get_table2();
 		return true;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	case DCN_VERSION_1_0:
 	case DCN_VERSION_1_01:
 	case DCN_VERSION_2_0:
@@ -76,12 +75,20 @@ bool dal_bios_parser_init_cmd_tbl_helper2(
 	case DCN_VERSION_3_02:
 	case DCN_VERSION_3_03:
 	case DCN_VERSION_3_1:
+	case DCN_VERSION_3_14:
+	case DCN_VERSION_3_15:
+	case DCN_VERSION_3_16:
+	case DCN_VERSION_3_2:
+	case DCN_VERSION_3_21:
+	case DCN_VERSION_3_5:
+	case DCN_VERSION_3_51:
+	case DCN_VERSION_3_6:
+	case DCN_VERSION_4_01:
 		*h = dal_cmd_tbl_helper_dce112_get_table2();
 		return true;
-#endif
+
 	default:
-		/* Unsupported DCE */
-		BREAK_TO_DEBUGGER();
+		*h = dal_cmd_tbl_helper_dce112_get_table2();
 		return false;
 	}
 }
diff --git a/drivers/gpu/drm/amd/display/dc/bios/dce110/command_table_helper_dce110.c b/drivers/gpu/drm/amd/display/dc/bios/dce110/command_table_helper_dce110.c
index 11bf247bb180..3099128223df 100644
--- a/drivers/gpu/drm/amd/display/dc/bios/dce110/command_table_helper_dce110.c
+++ b/drivers/gpu/drm/amd/display/dc/bios/dce110/command_table_helper_dce110.c
@@ -31,39 +31,6 @@
 
 #include "../command_table_helper.h"
 
-static uint8_t phy_id_to_atom(enum transmitter t)
-{
-	uint8_t atom_phy_id;
-
-	switch (t) {
-	case TRANSMITTER_UNIPHY_A:
-		atom_phy_id = ATOM_PHY_ID_UNIPHYA;
-		break;
-	case TRANSMITTER_UNIPHY_B:
-		atom_phy_id = ATOM_PHY_ID_UNIPHYB;
-		break;
-	case TRANSMITTER_UNIPHY_C:
-		atom_phy_id = ATOM_PHY_ID_UNIPHYC;
-		break;
-	case TRANSMITTER_UNIPHY_D:
-		atom_phy_id = ATOM_PHY_ID_UNIPHYD;
-		break;
-	case TRANSMITTER_UNIPHY_E:
-		atom_phy_id = ATOM_PHY_ID_UNIPHYE;
-		break;
-	case TRANSMITTER_UNIPHY_F:
-		atom_phy_id = ATOM_PHY_ID_UNIPHYF;
-		break;
-	case TRANSMITTER_UNIPHY_G:
-		atom_phy_id = ATOM_PHY_ID_UNIPHYG;
-		break;
-	default:
-		atom_phy_id = ATOM_PHY_ID_UNIPHYA;
-		break;
-	}
-	return atom_phy_id;
-}
-
 static uint8_t signal_type_to_atom_dig_mode(enum signal_type s)
 {
 	uint8_t atom_dig_mode = ATOM_TRANSMITTER_DIGMODE_V5_DP;
@@ -94,32 +61,6 @@ static uint8_t signal_type_to_atom_dig_mode(enum signal_type s)
 	return atom_dig_mode;
 }
 
-static uint8_t clock_source_id_to_atom_phy_clk_src_id(
-		enum clock_source_id id)
-{
-	uint8_t atom_phy_clk_src_id = 0;
-
-	switch (id) {
-	case CLOCK_SOURCE_ID_PLL0:
-		atom_phy_clk_src_id = ATOM_TRANSMITTER_CONFIG_V5_P0PLL;
-		break;
-	case CLOCK_SOURCE_ID_PLL1:
-		atom_phy_clk_src_id = ATOM_TRANSMITTER_CONFIG_V5_P1PLL;
-		break;
-	case CLOCK_SOURCE_ID_PLL2:
-		atom_phy_clk_src_id = ATOM_TRANSMITTER_CONFIG_V5_P2PLL;
-		break;
-	case CLOCK_SOURCE_ID_EXTERNAL:
-		atom_phy_clk_src_id = ATOM_TRANSMITTER_CONFIG_V5_REFCLK_SRC_EXT;
-		break;
-	default:
-		atom_phy_clk_src_id = ATOM_TRANSMITTER_CONFIG_V5_P1PLL;
-		break;
-	}
-
-	return atom_phy_clk_src_id >> 2;
-}
-
 static uint8_t hpd_sel_to_atom(enum hpd_source_id id)
 {
 	uint8_t atom_hpd_sel = 0;
@@ -207,51 +148,6 @@ static bool clock_source_id_to_atom(
 	return result;
 }
 
-static bool engine_bp_to_atom(enum engine_id id, uint32_t *atom_engine_id)
-{
-	bool result = false;
-
-	if (atom_engine_id != NULL)
-		switch (id) {
-		case ENGINE_ID_DIGA:
-			*atom_engine_id = ASIC_INT_DIG1_ENCODER_ID;
-			result = true;
-			break;
-		case ENGINE_ID_DIGB:
-			*atom_engine_id = ASIC_INT_DIG2_ENCODER_ID;
-			result = true;
-			break;
-		case ENGINE_ID_DIGC:
-			*atom_engine_id = ASIC_INT_DIG3_ENCODER_ID;
-			result = true;
-			break;
-		case ENGINE_ID_DIGD:
-			*atom_engine_id = ASIC_INT_DIG4_ENCODER_ID;
-			result = true;
-			break;
-		case ENGINE_ID_DIGE:
-			*atom_engine_id = ASIC_INT_DIG5_ENCODER_ID;
-			result = true;
-			break;
-		case ENGINE_ID_DIGF:
-			*atom_engine_id = ASIC_INT_DIG6_ENCODER_ID;
-			result = true;
-			break;
-		case ENGINE_ID_DIGG:
-			*atom_engine_id = ASIC_INT_DIG7_ENCODER_ID;
-			result = true;
-			break;
-		case ENGINE_ID_DACA:
-			*atom_engine_id = ASIC_INT_DAC1_ENCODER_ID;
-			result = true;
-			break;
-		default:
-			break;
-		}
-
-	return result;
-}
-
 static uint8_t encoder_action_to_atom(enum bp_encoder_control_action action)
 {
 	uint8_t atom_action = 0;
diff --git a/drivers/gpu/drm/amd/display/dc/bios/dce112/command_table_helper2_dce112.c b/drivers/gpu/drm/amd/display/dc/bios/dce112/command_table_helper2_dce112.c
index 755b6e33140a..349f0e5d5856 100644
--- a/drivers/gpu/drm/amd/display/dc/bios/dce112/command_table_helper2_dce112.c
+++ b/drivers/gpu/drm/amd/display/dc/bios/dce112/command_table_helper2_dce112.c
@@ -29,40 +29,9 @@
 
 #include "include/bios_parser_types.h"
 
-#include "../command_table_helper2.h"
-
-static uint8_t phy_id_to_atom(enum transmitter t)
-{
-	uint8_t atom_phy_id;
+#include "../command_table_helper.h"
 
-	switch (t) {
-	case TRANSMITTER_UNIPHY_A:
-		atom_phy_id = ATOM_PHY_ID_UNIPHYA;
-		break;
-	case TRANSMITTER_UNIPHY_B:
-		atom_phy_id = ATOM_PHY_ID_UNIPHYB;
-		break;
-	case TRANSMITTER_UNIPHY_C:
-		atom_phy_id = ATOM_PHY_ID_UNIPHYC;
-		break;
-	case TRANSMITTER_UNIPHY_D:
-		atom_phy_id = ATOM_PHY_ID_UNIPHYD;
-		break;
-	case TRANSMITTER_UNIPHY_E:
-		atom_phy_id = ATOM_PHY_ID_UNIPHYE;
-		break;
-	case TRANSMITTER_UNIPHY_F:
-		atom_phy_id = ATOM_PHY_ID_UNIPHYF;
-		break;
-	case TRANSMITTER_UNIPHY_G:
-		atom_phy_id = ATOM_PHY_ID_UNIPHYG;
-		break;
-	default:
-		atom_phy_id = ATOM_PHY_ID_UNIPHYA;
-		break;
-	}
-	return atom_phy_id;
-}
+#include "../command_table_helper2.h"
 
 static uint8_t signal_type_to_atom_dig_mode(enum signal_type s)
 {
@@ -91,32 +60,6 @@ static uint8_t signal_type_to_atom_dig_mode(enum signal_type s)
 	return atom_dig_mode;
 }
 
-static uint8_t clock_source_id_to_atom_phy_clk_src_id(
-		enum clock_source_id id)
-{
-	uint8_t atom_phy_clk_src_id = 0;
-
-	switch (id) {
-	case CLOCK_SOURCE_ID_PLL0:
-		atom_phy_clk_src_id = ATOM_TRANSMITTER_CONFIG_V5_P0PLL;
-		break;
-	case CLOCK_SOURCE_ID_PLL1:
-		atom_phy_clk_src_id = ATOM_TRANSMITTER_CONFIG_V5_P1PLL;
-		break;
-	case CLOCK_SOURCE_ID_PLL2:
-		atom_phy_clk_src_id = ATOM_TRANSMITTER_CONFIG_V5_P2PLL;
-		break;
-	case CLOCK_SOURCE_ID_EXTERNAL:
-		atom_phy_clk_src_id = ATOM_TRANSMITTER_CONFIG_V5_REFCLK_SRC_EXT;
-		break;
-	default:
-		atom_phy_clk_src_id = ATOM_TRANSMITTER_CONFIG_V5_P1PLL;
-		break;
-	}
-
-	return atom_phy_clk_src_id >> 2;
-}
-
 static uint8_t hpd_sel_to_atom(enum hpd_source_id id)
 {
 	uint8_t atom_hpd_sel = 0;
@@ -209,51 +152,6 @@ static bool clock_source_id_to_atom(
 	return result;
 }
 
-static bool engine_bp_to_atom(enum engine_id id, uint32_t *atom_engine_id)
-{
-	bool result = false;
-
-	if (atom_engine_id != NULL)
-		switch (id) {
-		case ENGINE_ID_DIGA:
-			*atom_engine_id = ASIC_INT_DIG1_ENCODER_ID;
-			result = true;
-			break;
-		case ENGINE_ID_DIGB:
-			*atom_engine_id = ASIC_INT_DIG2_ENCODER_ID;
-			result = true;
-			break;
-		case ENGINE_ID_DIGC:
-			*atom_engine_id = ASIC_INT_DIG3_ENCODER_ID;
-			result = true;
-			break;
-		case ENGINE_ID_DIGD:
-			*atom_engine_id = ASIC_INT_DIG4_ENCODER_ID;
-			result = true;
-			break;
-		case ENGINE_ID_DIGE:
-			*atom_engine_id = ASIC_INT_DIG5_ENCODER_ID;
-			result = true;
-			break;
-		case ENGINE_ID_DIGF:
-			*atom_engine_id = ASIC_INT_DIG6_ENCODER_ID;
-			result = true;
-			break;
-		case ENGINE_ID_DIGG:
-			*atom_engine_id = ASIC_INT_DIG7_ENCODER_ID;
-			result = true;
-			break;
-		case ENGINE_ID_DACA:
-			*atom_engine_id = ASIC_INT_DAC1_ENCODER_ID;
-			result = true;
-			break;
-		default:
-			break;
-		}
-
-	return result;
-}
-
 static uint8_t encoder_action_to_atom(enum bp_encoder_control_action action)
 {
 	uint8_t atom_action = 0;
diff --git a/drivers/gpu/drm/amd/display/dc/bios/dce112/command_table_helper_dce112.c b/drivers/gpu/drm/amd/display/dc/bios/dce112/command_table_helper_dce112.c
index 06b4f7fa4a50..1a5fefcde8af 100644
--- a/drivers/gpu/drm/amd/display/dc/bios/dce112/command_table_helper_dce112.c
+++ b/drivers/gpu/drm/amd/display/dc/bios/dce112/command_table_helper_dce112.c
@@ -31,39 +31,6 @@
 
 #include "../command_table_helper.h"
 
-static uint8_t phy_id_to_atom(enum transmitter t)
-{
-	uint8_t atom_phy_id;
-
-	switch (t) {
-	case TRANSMITTER_UNIPHY_A:
-		atom_phy_id = ATOM_PHY_ID_UNIPHYA;
-		break;
-	case TRANSMITTER_UNIPHY_B:
-		atom_phy_id = ATOM_PHY_ID_UNIPHYB;
-		break;
-	case TRANSMITTER_UNIPHY_C:
-		atom_phy_id = ATOM_PHY_ID_UNIPHYC;
-		break;
-	case TRANSMITTER_UNIPHY_D:
-		atom_phy_id = ATOM_PHY_ID_UNIPHYD;
-		break;
-	case TRANSMITTER_UNIPHY_E:
-		atom_phy_id = ATOM_PHY_ID_UNIPHYE;
-		break;
-	case TRANSMITTER_UNIPHY_F:
-		atom_phy_id = ATOM_PHY_ID_UNIPHYF;
-		break;
-	case TRANSMITTER_UNIPHY_G:
-		atom_phy_id = ATOM_PHY_ID_UNIPHYG;
-		break;
-	default:
-		atom_phy_id = ATOM_PHY_ID_UNIPHYA;
-		break;
-	}
-	return atom_phy_id;
-}
-
 static uint8_t signal_type_to_atom_dig_mode(enum signal_type s)
 {
 	uint8_t atom_dig_mode = ATOM_TRANSMITTER_DIGMODE_V6_DP;
@@ -91,32 +58,6 @@ static uint8_t signal_type_to_atom_dig_mode(enum signal_type s)
 	return atom_dig_mode;
 }
 
-static uint8_t clock_source_id_to_atom_phy_clk_src_id(
-		enum clock_source_id id)
-{
-	uint8_t atom_phy_clk_src_id = 0;
-
-	switch (id) {
-	case CLOCK_SOURCE_ID_PLL0:
-		atom_phy_clk_src_id = ATOM_TRANSMITTER_CONFIG_V5_P0PLL;
-		break;
-	case CLOCK_SOURCE_ID_PLL1:
-		atom_phy_clk_src_id = ATOM_TRANSMITTER_CONFIG_V5_P1PLL;
-		break;
-	case CLOCK_SOURCE_ID_PLL2:
-		atom_phy_clk_src_id = ATOM_TRANSMITTER_CONFIG_V5_P2PLL;
-		break;
-	case CLOCK_SOURCE_ID_EXTERNAL:
-		atom_phy_clk_src_id = ATOM_TRANSMITTER_CONFIG_V5_REFCLK_SRC_EXT;
-		break;
-	default:
-		atom_phy_clk_src_id = ATOM_TRANSMITTER_CONFIG_V5_P1PLL;
-		break;
-	}
-
-	return atom_phy_clk_src_id >> 2;
-}
-
 static uint8_t hpd_sel_to_atom(enum hpd_source_id id)
 {
 	uint8_t atom_hpd_sel = 0;
@@ -209,51 +150,6 @@ static bool clock_source_id_to_atom(
 	return result;
 }
 
-static bool engine_bp_to_atom(enum engine_id id, uint32_t *atom_engine_id)
-{
-	bool result = false;
-
-	if (atom_engine_id != NULL)
-		switch (id) {
-		case ENGINE_ID_DIGA:
-			*atom_engine_id = ASIC_INT_DIG1_ENCODER_ID;
-			result = true;
-			break;
-		case ENGINE_ID_DIGB:
-			*atom_engine_id = ASIC_INT_DIG2_ENCODER_ID;
-			result = true;
-			break;
-		case ENGINE_ID_DIGC:
-			*atom_engine_id = ASIC_INT_DIG3_ENCODER_ID;
-			result = true;
-			break;
-		case ENGINE_ID_DIGD:
-			*atom_engine_id = ASIC_INT_DIG4_ENCODER_ID;
-			result = true;
-			break;
-		case ENGINE_ID_DIGE:
-			*atom_engine_id = ASIC_INT_DIG5_ENCODER_ID;
-			result = true;
-			break;
-		case ENGINE_ID_DIGF:
-			*atom_engine_id = ASIC_INT_DIG6_ENCODER_ID;
-			result = true;
-			break;
-		case ENGINE_ID_DIGG:
-			*atom_engine_id = ASIC_INT_DIG7_ENCODER_ID;
-			result = true;
-			break;
-		case ENGINE_ID_DACA:
-			*atom_engine_id = ASIC_INT_DAC1_ENCODER_ID;
-			result = true;
-			break;
-		default:
-			break;
-		}
-
-	return result;
-}
-
 static uint8_t encoder_action_to_atom(enum bp_encoder_control_action action)
 {
 	uint8_t atom_action = 0;
diff --git a/drivers/gpu/drm/amd/display/dc/bios/dce60/command_table_helper_dce60.c b/drivers/gpu/drm/amd/display/dc/bios/dce60/command_table_helper_dce60.c
index 710221b4f5c5..01ccc803040c 100644
--- a/drivers/gpu/drm/amd/display/dc/bios/dce60/command_table_helper_dce60.c
+++ b/drivers/gpu/drm/amd/display/dc/bios/dce60/command_table_helper_dce60.c
@@ -58,51 +58,6 @@ static uint8_t encoder_action_to_atom(enum bp_encoder_control_action action)
 	return atom_action;
 }
 
-static bool engine_bp_to_atom(enum engine_id id, uint32_t *atom_engine_id)
-{
-	bool result = false;
-
-	if (atom_engine_id != NULL)
-		switch (id) {
-		case ENGINE_ID_DIGA:
-			*atom_engine_id = ASIC_INT_DIG1_ENCODER_ID;
-			result = true;
-			break;
-		case ENGINE_ID_DIGB:
-			*atom_engine_id = ASIC_INT_DIG2_ENCODER_ID;
-			result = true;
-			break;
-		case ENGINE_ID_DIGC:
-			*atom_engine_id = ASIC_INT_DIG3_ENCODER_ID;
-			result = true;
-			break;
-		case ENGINE_ID_DIGD:
-			*atom_engine_id = ASIC_INT_DIG4_ENCODER_ID;
-			result = true;
-			break;
-		case ENGINE_ID_DIGE:
-			*atom_engine_id = ASIC_INT_DIG5_ENCODER_ID;
-			result = true;
-			break;
-		case ENGINE_ID_DIGF:
-			*atom_engine_id = ASIC_INT_DIG6_ENCODER_ID;
-			result = true;
-			break;
-		case ENGINE_ID_DIGG:
-			*atom_engine_id = ASIC_INT_DIG7_ENCODER_ID;
-			result = true;
-			break;
-		case ENGINE_ID_DACA:
-			*atom_engine_id = ASIC_INT_DAC1_ENCODER_ID;
-			result = true;
-			break;
-		default:
-			break;
-		}
-
-	return result;
-}
-
 static bool clock_source_id_to_atom(
 	enum clock_source_id id,
 	uint32_t *atom_pll_id)
@@ -149,32 +104,6 @@ static bool clock_source_id_to_atom(
 	return result;
 }
 
-static uint8_t clock_source_id_to_atom_phy_clk_src_id(
-		enum clock_source_id id)
-{
-	uint8_t atom_phy_clk_src_id = 0;
-
-	switch (id) {
-	case CLOCK_SOURCE_ID_PLL0:
-		atom_phy_clk_src_id = ATOM_TRANSMITTER_CONFIG_V5_P0PLL;
-		break;
-	case CLOCK_SOURCE_ID_PLL1:
-		atom_phy_clk_src_id = ATOM_TRANSMITTER_CONFIG_V5_P1PLL;
-		break;
-	case CLOCK_SOURCE_ID_PLL2:
-		atom_phy_clk_src_id = ATOM_TRANSMITTER_CONFIG_V5_P2PLL;
-		break;
-	case CLOCK_SOURCE_ID_EXTERNAL:
-		atom_phy_clk_src_id = ATOM_TRANSMITTER_CONFIG_V5_REFCLK_SRC_EXT;
-		break;
-	default:
-		atom_phy_clk_src_id = ATOM_TRANSMITTER_CONFIG_V5_P1PLL;
-		break;
-	}
-
-	return atom_phy_clk_src_id >> 2;
-}
-
 static uint8_t signal_type_to_atom_dig_mode(enum signal_type s)
 {
 	uint8_t atom_dig_mode = ATOM_TRANSMITTER_DIGMODE_V5_DP;
@@ -270,39 +199,6 @@ static uint8_t dig_encoder_sel_to_atom(enum engine_id id)
 	return atom_dig_encoder_sel;
 }
 
-static uint8_t phy_id_to_atom(enum transmitter t)
-{
-	uint8_t atom_phy_id;
-
-	switch (t) {
-	case TRANSMITTER_UNIPHY_A:
-		atom_phy_id = ATOM_PHY_ID_UNIPHYA;
-		break;
-	case TRANSMITTER_UNIPHY_B:
-		atom_phy_id = ATOM_PHY_ID_UNIPHYB;
-		break;
-	case TRANSMITTER_UNIPHY_C:
-		atom_phy_id = ATOM_PHY_ID_UNIPHYC;
-		break;
-	case TRANSMITTER_UNIPHY_D:
-		atom_phy_id = ATOM_PHY_ID_UNIPHYD;
-		break;
-	case TRANSMITTER_UNIPHY_E:
-		atom_phy_id = ATOM_PHY_ID_UNIPHYE;
-		break;
-	case TRANSMITTER_UNIPHY_F:
-		atom_phy_id = ATOM_PHY_ID_UNIPHYF;
-		break;
-	case TRANSMITTER_UNIPHY_G:
-		atom_phy_id = ATOM_PHY_ID_UNIPHYG;
-		break;
-	default:
-		atom_phy_id = ATOM_PHY_ID_UNIPHYA;
-		break;
-	}
-	return atom_phy_id;
-}
-
 static uint8_t disp_power_gating_action_to_atom(
 	enum bp_pipe_control_action action)
 {
diff --git a/drivers/gpu/drm/amd/display/dc/bios/dce80/command_table_helper_dce80.c b/drivers/gpu/drm/amd/display/dc/bios/dce80/command_table_helper_dce80.c
index 8b30b558cf1f..2ec5264536c7 100644
--- a/drivers/gpu/drm/amd/display/dc/bios/dce80/command_table_helper_dce80.c
+++ b/drivers/gpu/drm/amd/display/dc/bios/dce80/command_table_helper_dce80.c
@@ -58,51 +58,6 @@ static uint8_t encoder_action_to_atom(enum bp_encoder_control_action action)
 	return atom_action;
 }
 
-static bool engine_bp_to_atom(enum engine_id id, uint32_t *atom_engine_id)
-{
-	bool result = false;
-
-	if (atom_engine_id != NULL)
-		switch (id) {
-		case ENGINE_ID_DIGA:
-			*atom_engine_id = ASIC_INT_DIG1_ENCODER_ID;
-			result = true;
-			break;
-		case ENGINE_ID_DIGB:
-			*atom_engine_id = ASIC_INT_DIG2_ENCODER_ID;
-			result = true;
-			break;
-		case ENGINE_ID_DIGC:
-			*atom_engine_id = ASIC_INT_DIG3_ENCODER_ID;
-			result = true;
-			break;
-		case ENGINE_ID_DIGD:
-			*atom_engine_id = ASIC_INT_DIG4_ENCODER_ID;
-			result = true;
-			break;
-		case ENGINE_ID_DIGE:
-			*atom_engine_id = ASIC_INT_DIG5_ENCODER_ID;
-			result = true;
-			break;
-		case ENGINE_ID_DIGF:
-			*atom_engine_id = ASIC_INT_DIG6_ENCODER_ID;
-			result = true;
-			break;
-		case ENGINE_ID_DIGG:
-			*atom_engine_id = ASIC_INT_DIG7_ENCODER_ID;
-			result = true;
-			break;
-		case ENGINE_ID_DACA:
-			*atom_engine_id = ASIC_INT_DAC1_ENCODER_ID;
-			result = true;
-			break;
-		default:
-			break;
-		}
-
-	return result;
-}
-
 static bool clock_source_id_to_atom(
 	enum clock_source_id id,
 	uint32_t *atom_pll_id)
@@ -149,32 +104,6 @@ static bool clock_source_id_to_atom(
 	return result;
 }
 
-static uint8_t clock_source_id_to_atom_phy_clk_src_id(
-		enum clock_source_id id)
-{
-	uint8_t atom_phy_clk_src_id = 0;
-
-	switch (id) {
-	case CLOCK_SOURCE_ID_PLL0:
-		atom_phy_clk_src_id = ATOM_TRANSMITTER_CONFIG_V5_P0PLL;
-		break;
-	case CLOCK_SOURCE_ID_PLL1:
-		atom_phy_clk_src_id = ATOM_TRANSMITTER_CONFIG_V5_P1PLL;
-		break;
-	case CLOCK_SOURCE_ID_PLL2:
-		atom_phy_clk_src_id = ATOM_TRANSMITTER_CONFIG_V5_P2PLL;
-		break;
-	case CLOCK_SOURCE_ID_EXTERNAL:
-		atom_phy_clk_src_id = ATOM_TRANSMITTER_CONFIG_V5_REFCLK_SRC_EXT;
-		break;
-	default:
-		atom_phy_clk_src_id = ATOM_TRANSMITTER_CONFIG_V5_P1PLL;
-		break;
-	}
-
-	return atom_phy_clk_src_id >> 2;
-}
-
 static uint8_t signal_type_to_atom_dig_mode(enum signal_type s)
 {
 	uint8_t atom_dig_mode = ATOM_TRANSMITTER_DIGMODE_V5_DP;
@@ -270,39 +199,6 @@ static uint8_t dig_encoder_sel_to_atom(enum engine_id id)
 	return atom_dig_encoder_sel;
 }
 
-static uint8_t phy_id_to_atom(enum transmitter t)
-{
-	uint8_t atom_phy_id;
-
-	switch (t) {
-	case TRANSMITTER_UNIPHY_A:
-		atom_phy_id = ATOM_PHY_ID_UNIPHYA;
-		break;
-	case TRANSMITTER_UNIPHY_B:
-		atom_phy_id = ATOM_PHY_ID_UNIPHYB;
-		break;
-	case TRANSMITTER_UNIPHY_C:
-		atom_phy_id = ATOM_PHY_ID_UNIPHYC;
-		break;
-	case TRANSMITTER_UNIPHY_D:
-		atom_phy_id = ATOM_PHY_ID_UNIPHYD;
-		break;
-	case TRANSMITTER_UNIPHY_E:
-		atom_phy_id = ATOM_PHY_ID_UNIPHYE;
-		break;
-	case TRANSMITTER_UNIPHY_F:
-		atom_phy_id = ATOM_PHY_ID_UNIPHYF;
-		break;
-	case TRANSMITTER_UNIPHY_G:
-		atom_phy_id = ATOM_PHY_ID_UNIPHYG;
-		break;
-	default:
-		atom_phy_id = ATOM_PHY_ID_UNIPHYA;
-		break;
-	}
-	return atom_phy_id;
-}
-
 static uint8_t disp_power_gating_action_to_atom(
 	enum bp_pipe_control_action action)
 {
diff --git a/drivers/gpu/drm/amd/display/dc/calcs/Makefile b/drivers/gpu/drm/amd/display/dc/calcs/Makefile
deleted file mode 100644
index f3c00f479e1c..000000000000
--- a/drivers/gpu/drm/amd/display/dc/calcs/Makefile
+++ /dev/null
@@ -1,68 +0,0 @@
-#
-# Copyright 2017 Advanced Micro Devices, Inc.
-# Copyright 2019 Raptor Engineering, LLC
-#
-# Permission is hereby granted, free of charge, to any person obtaining a
-# copy of this software and associated documentation files (the "Software"),
-# to deal in the Software without restriction, including without limitation
-# the rights to use, copy, modify, merge, publish, distribute, sublicense,
-# and/or sell copies of the Software, and to permit persons to whom the
-# Software is furnished to do so, subject to the following conditions:
-#
-# The above copyright notice and this permission notice shall be included in
-# all copies or substantial portions of the Software.
-#
-# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
-# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
-# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
-# THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
-# OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
-# ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
-# OTHER DEALINGS IN THE SOFTWARE.
-#
-#
-# Makefile for the 'calcs' sub-component of DAL.
-# It calculates Bandwidth and Watermarks values for HW programming
-#
-
-ifdef CONFIG_X86
-calcs_ccflags := -mhard-float -msse
-endif
-
-ifdef CONFIG_PPC64
-calcs_ccflags := -mhard-float -maltivec
-endif
-
-ifdef CONFIG_CC_IS_GCC
-ifeq ($(call cc-ifversion, -lt, 0701, y), y)
-IS_OLD_GCC = 1
-endif
-endif
-
-ifdef CONFIG_X86
-ifdef IS_OLD_GCC
-# Stack alignment mismatch, proceed with caution.
-# GCC < 7.1 cannot compile code using `double` and -mpreferred-stack-boundary=3
-# (8B stack alignment).
-calcs_ccflags += -mpreferred-stack-boundary=4
-else
-calcs_ccflags += -msse2
-endif
-endif
-
-CFLAGS_$(AMDDALPATH)/dc/calcs/dcn_calcs.o := $(calcs_ccflags)
-CFLAGS_$(AMDDALPATH)/dc/calcs/dcn_calc_auto.o := $(calcs_ccflags)
-CFLAGS_$(AMDDALPATH)/dc/calcs/dcn_calc_math.o := $(calcs_ccflags) -Wno-tautological-compare
-CFLAGS_REMOVE_$(AMDDALPATH)/dc/calcs/dcn_calcs.o := $(calcs_rcflags)
-CFLAGS_REMOVE_$(AMDDALPATH)/dc/calcs/dcn_calc_auto.o := $(calcs_rcflags)
-CFLAGS_REMOVE_$(AMDDALPATH)/dc/calcs/dcn_calc_math.o := $(calcs_rcflags)
-
-BW_CALCS = dce_calcs.o bw_fixed.o custom_float.o
-
-ifdef CONFIG_DRM_AMD_DC_DCN
-BW_CALCS += dcn_calcs.o dcn_calc_math.o dcn_calc_auto.o
-endif
-
-AMD_DAL_BW_CALCS = $(addprefix $(AMDDALPATH)/dc/calcs/,$(BW_CALCS))
-
-AMD_DISPLAY_FILES += $(AMD_DAL_BW_CALCS)
diff --git a/drivers/gpu/drm/amd/display/dc/calcs/custom_float.c b/drivers/gpu/drm/amd/display/dc/calcs/custom_float.c
deleted file mode 100644
index 31d167bc548f..000000000000
--- a/drivers/gpu/drm/amd/display/dc/calcs/custom_float.c
+++ /dev/null
@@ -1,197 +0,0 @@
-/*
- * Copyright 2017 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: AMD
- *
- */
-#include "dm_services.h"
-#include "custom_float.h"
-
-
-static bool build_custom_float(
-	struct fixed31_32 value,
-	const struct custom_float_format *format,
-	bool *negative,
-	uint32_t *mantissa,
-	uint32_t *exponenta)
-{
-	uint32_t exp_offset = (1 << (format->exponenta_bits - 1)) - 1;
-
-	const struct fixed31_32 mantissa_constant_plus_max_fraction =
-		dc_fixpt_from_fraction(
-			(1LL << (format->mantissa_bits + 1)) - 1,
-			1LL << format->mantissa_bits);
-
-	struct fixed31_32 mantiss;
-
-	if (dc_fixpt_eq(
-		value,
-		dc_fixpt_zero)) {
-		*negative = false;
-		*mantissa = 0;
-		*exponenta = 0;
-		return true;
-	}
-
-	if (dc_fixpt_lt(
-		value,
-		dc_fixpt_zero)) {
-		*negative = format->sign;
-		value = dc_fixpt_neg(value);
-	} else {
-		*negative = false;
-	}
-
-	if (dc_fixpt_lt(
-		value,
-		dc_fixpt_one)) {
-		uint32_t i = 1;
-
-		do {
-			value = dc_fixpt_shl(value, 1);
-			++i;
-		} while (dc_fixpt_lt(
-			value,
-			dc_fixpt_one));
-
-		--i;
-
-		if (exp_offset <= i) {
-			*mantissa = 0;
-			*exponenta = 0;
-			return true;
-		}
-
-		*exponenta = exp_offset - i;
-	} else if (dc_fixpt_le(
-		mantissa_constant_plus_max_fraction,
-		value)) {
-		uint32_t i = 1;
-
-		do {
-			value = dc_fixpt_shr(value, 1);
-			++i;
-		} while (dc_fixpt_lt(
-			mantissa_constant_plus_max_fraction,
-			value));
-
-		*exponenta = exp_offset + i - 1;
-	} else {
-		*exponenta = exp_offset;
-	}
-
-	mantiss = dc_fixpt_sub(
-		value,
-		dc_fixpt_one);
-
-	if (dc_fixpt_lt(
-			mantiss,
-			dc_fixpt_zero) ||
-		dc_fixpt_lt(
-			dc_fixpt_one,
-			mantiss))
-		mantiss = dc_fixpt_zero;
-	else
-		mantiss = dc_fixpt_shl(
-			mantiss,
-			format->mantissa_bits);
-
-	*mantissa = dc_fixpt_floor(mantiss);
-
-	return true;
-}
-
-static bool setup_custom_float(
-	const struct custom_float_format *format,
-	bool negative,
-	uint32_t mantissa,
-	uint32_t exponenta,
-	uint32_t *result)
-{
-	uint32_t i = 0;
-	uint32_t j = 0;
-
-	uint32_t value = 0;
-
-	/* verification code:
-	 * once calculation is ok we can remove it
-	 */
-
-	const uint32_t mantissa_mask =
-		(1 << (format->mantissa_bits + 1)) - 1;
-
-	const uint32_t exponenta_mask =
-		(1 << (format->exponenta_bits + 1)) - 1;
-
-	if (mantissa & ~mantissa_mask) {
-		BREAK_TO_DEBUGGER();
-		mantissa = mantissa_mask;
-	}
-
-	if (exponenta & ~exponenta_mask) {
-		BREAK_TO_DEBUGGER();
-		exponenta = exponenta_mask;
-	}
-
-	/* end of verification code */
-
-	while (i < format->mantissa_bits) {
-		uint32_t mask = 1 << i;
-
-		if (mantissa & mask)
-			value |= mask;
-
-		++i;
-	}
-
-	while (j < format->exponenta_bits) {
-		uint32_t mask = 1 << j;
-
-		if (exponenta & mask)
-			value |= mask << i;
-
-		++j;
-	}
-
-	if (negative && format->sign)
-		value |= 1 << (i + j);
-
-	*result = value;
-
-	return true;
-}
-
-bool convert_to_custom_float_format(
-	struct fixed31_32 value,
-	const struct custom_float_format *format,
-	uint32_t *result)
-{
-	uint32_t mantissa;
-	uint32_t exponenta;
-	bool negative;
-
-	return build_custom_float(
-		value, format, &negative, &mantissa, &exponenta) &&
-	setup_custom_float(
-		format, negative, mantissa, exponenta, result);
-}
-
-
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/Makefile b/drivers/gpu/drm/amd/display/dc/clk_mgr/Makefile
index 6bd73e49a6d2..60021671b386 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/Makefile
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/Makefile
@@ -74,7 +74,7 @@ CLK_MGR_DCE120 = dce120_clk_mgr.o
 AMD_DAL_CLK_MGR_DCE120 = $(addprefix $(AMDDALPATH)/dc/clk_mgr/dce120/,$(CLK_MGR_DCE120))
 
 AMD_DISPLAY_FILES += $(AMD_DAL_CLK_MGR_DCE120)
-ifdef CONFIG_DRM_AMD_DC_DCN
+ifdef CONFIG_DRM_AMD_DC_FP
 ###############################################################################
 # DCN10
 ###############################################################################
@@ -83,7 +83,6 @@ CLK_MGR_DCN10 = rv1_clk_mgr.o rv1_clk_mgr_vbios_smu.o rv2_clk_mgr.o
 AMD_DAL_CLK_MGR_DCN10 = $(addprefix $(AMDDALPATH)/dc/clk_mgr/dcn10/,$(CLK_MGR_DCN10))
 
 AMD_DISPLAY_FILES += $(AMD_DAL_CLK_MGR_DCN10)
-
 ###############################################################################
 # DCN20
 ###############################################################################
@@ -107,25 +106,13 @@ AMD_DISPLAY_FILES += $(AMD_DAL_CLK_MGR_DCN201)
 ###############################################################################
 CLK_MGR_DCN21 = rn_clk_mgr.o rn_clk_mgr_vbios_smu.o
 
-# prevent build errors regarding soft-float vs hard-float FP ABI tags
-# this code is currently unused on ppc64, as it applies to Renoir APUs only
-ifdef CONFIG_PPC64
-CFLAGS_$(AMDDALPATH)/dc/clk_mgr/dcn21/rn_clk_mgr.o := $(call cc-option,-mno-gnu-attribute)
-endif
-
 AMD_DAL_CLK_MGR_DCN21 = $(addprefix $(AMDDALPATH)/dc/clk_mgr/dcn21/,$(CLK_MGR_DCN21))
 
 AMD_DISPLAY_FILES += $(AMD_DAL_CLK_MGR_DCN21)
 ###############################################################################
 # DCN30
 ###############################################################################
-CLK_MGR_DCN30 = dcn30_clk_mgr.o dcn30_clk_mgr_smu_msg.o
-
-# prevent build errors regarding soft-float vs hard-float FP ABI tags
-# this code is currently unused on ppc64, as it applies to VanGogh APUs only
-ifdef CONFIG_PPC64
-CFLAGS_$(AMDDALPATH)/dc/clk_mgr/dcn30/dcn30_clk_mgr.o := $(call cc-option,-mno-gnu-attribute)
-endif
+CLK_MGR_DCN30 = dcn30_clk_mgr.o dcn30_clk_mgr_smu_msg.o dcn30m_clk_mgr.o dcn30m_clk_mgr_smu_msg.o
 
 AMD_DAL_CLK_MGR_DCN30 = $(addprefix $(AMDDALPATH)/dc/clk_mgr/dcn30/,$(CLK_MGR_DCN30))
 
@@ -135,12 +122,6 @@ AMD_DISPLAY_FILES += $(AMD_DAL_CLK_MGR_DCN30)
 ###############################################################################
 CLK_MGR_DCN301 = vg_clk_mgr.o dcn301_smu.o
 
-# prevent build errors regarding soft-float vs hard-float FP ABI tags
-# this code is currently unused on ppc64, as it applies to VanGogh APUs only
-ifdef CONFIG_PPC64
-CFLAGS_$(AMDDALPATH)/dc/clk_mgr/dcn301/vg_clk_mgr.o := $(call cc-option,-mno-gnu-attribute)
-endif
-
 AMD_DAL_CLK_MGR_DCN301 = $(addprefix $(AMDDALPATH)/dc/clk_mgr/dcn301/,$(CLK_MGR_DCN301))
 
 AMD_DISPLAY_FILES += $(AMD_DAL_CLK_MGR_DCN301)
@@ -153,4 +134,58 @@ CLK_MGR_DCN31 = dcn31_smu.o dcn31_clk_mgr.o
 AMD_DAL_CLK_MGR_DCN31 = $(addprefix $(AMDDALPATH)/dc/clk_mgr/dcn31/,$(CLK_MGR_DCN31))
 
 AMD_DISPLAY_FILES += $(AMD_DAL_CLK_MGR_DCN31)
+
+###############################################################################
+# DCN314
+###############################################################################
+CLK_MGR_DCN314 = dcn314_smu.o dcn314_clk_mgr.o
+
+AMD_DAL_CLK_MGR_DCN314 = $(addprefix $(AMDDALPATH)/dc/clk_mgr/dcn314/,$(CLK_MGR_DCN314))
+
+AMD_DISPLAY_FILES += $(AMD_DAL_CLK_MGR_DCN314)
+
+###############################################################################
+# DCN315
+###############################################################################
+CLK_MGR_DCN315 = dcn315_smu.o dcn315_clk_mgr.o
+
+AMD_DAL_CLK_MGR_DCN315 = $(addprefix $(AMDDALPATH)/dc/clk_mgr/dcn315/,$(CLK_MGR_DCN315))
+
+AMD_DISPLAY_FILES += $(AMD_DAL_CLK_MGR_DCN315)
+
+###############################################################################
+# DCN316
+###############################################################################
+CLK_MGR_DCN316 = dcn316_smu.o dcn316_clk_mgr.o
+
+AMD_DAL_CLK_MGR_DCN316 = $(addprefix $(AMDDALPATH)/dc/clk_mgr/dcn316/,$(CLK_MGR_DCN316))
+
+AMD_DISPLAY_FILES += $(AMD_DAL_CLK_MGR_DCN316)
+
+###############################################################################
+# DCN32
+###############################################################################
+CLK_MGR_DCN32 = dcn32_clk_mgr.o dcn32_clk_mgr_smu_msg.o
+
+AMD_DAL_CLK_MGR_DCN32 = $(addprefix $(AMDDALPATH)/dc/clk_mgr/dcn32/,$(CLK_MGR_DCN32))
+
+AMD_DISPLAY_FILES += $(AMD_DAL_CLK_MGR_DCN32)
+
+###############################################################################
+# DCN35
+###############################################################################
+CLK_MGR_DCN35 = dcn35_smu.o dcn351_clk_mgr.o dcn35_clk_mgr.o
+
+AMD_DAL_CLK_MGR_DCN35 = $(addprefix $(AMDDALPATH)/dc/clk_mgr/dcn35/,$(CLK_MGR_DCN35))
+
+AMD_DISPLAY_FILES += $(AMD_DAL_CLK_MGR_DCN35)
+
+###############################################################################
+# DCN401
+###############################################################################
+CLK_MGR_DCN401 = dcn401_clk_mgr.o dcn401_clk_mgr_smu_msg.o
+
+AMD_DAL_CLK_MGR_DCN401 = $(addprefix $(AMDDALPATH)/dc/clk_mgr/dcn401/,$(CLK_MGR_DCN401))
+
+AMD_DISPLAY_FILES += $(AMD_DAL_CLK_MGR_DCN401)
 endif
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/clk_mgr.c
index 26f96ee32472..15cf13ec5302 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/clk_mgr.c
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/clk_mgr.c
@@ -23,12 +23,12 @@
  *
  */
 
-#include <linux/slab.h>
-
 #include "dal_asic_id.h"
 #include "dc_types.h"
 #include "dccg.h"
 #include "clk_mgr_internal.h"
+#include "dc_state_priv.h"
+#include "link_service.h"
 
 #include "dce100/dce_clk_mgr.h"
 #include "dce110/dce110_clk_mgr.h"
@@ -43,7 +43,12 @@
 #include "dcn30/dcn30_clk_mgr.h"
 #include "dcn301/vg_clk_mgr.h"
 #include "dcn31/dcn31_clk_mgr.h"
-
+#include "dcn314/dcn314_clk_mgr.h"
+#include "dcn315/dcn315_clk_mgr.h"
+#include "dcn316/dcn316_clk_mgr.h"
+#include "dcn32/dcn32_clk_mgr.h"
+#include "dcn35/dcn35_clk_mgr.h"
+#include "dcn401/dcn401_clk_mgr.h"
 
 int clk_mgr_helper_get_active_display_cnt(
 		struct dc *dc,
@@ -54,14 +59,15 @@ int clk_mgr_helper_get_active_display_cnt(
 	display_count = 0;
 	for (i = 0; i < context->stream_count; i++) {
 		const struct dc_stream_state *stream = context->streams[i];
+		const struct dc_stream_status *stream_status = &context->stream_status[i];
 
-		/*
-		 * Only notify active stream or virtual stream.
-		 * Need to notify virtual stream to work around
-		 * headless case. HPD does not fire when system is in
-		 * S0i2.
+		/* Don't count SubVP phantom pipes as part of active
+		 * display count
 		 */
-		if (!stream->dpms_off || stream->signal == SIGNAL_TYPE_VIRTUAL)
+		if (dc_state_get_stream_subvp_type(context, stream) == SUBVP_PHANTOM)
+			continue;
+
+		if (!stream->dpms_off || dc->is_switch_in_progress_dest || (stream_status && stream_status->plane_count))
 			display_count++;
 	}
 
@@ -94,7 +100,7 @@ void clk_mgr_exit_optimized_pwr_state(const struct dc *dc, struct clk_mgr *clk_m
 	int edp_num;
 	unsigned int panel_inst;
 
-	get_edp_links(dc, edp_links, &edp_num);
+	dc_get_edp_links(dc, edp_links, &edp_num);
 	if (dc->hwss.exit_optimized_pwr_state)
 		dc->hwss.exit_optimized_pwr_state(dc, dc->current_state);
 
@@ -106,7 +112,8 @@ void clk_mgr_exit_optimized_pwr_state(const struct dc *dc, struct clk_mgr *clk_m
 			if (!edp_link->psr_settings.psr_feature_enabled)
 				continue;
 			clk_mgr->psr_allow_active_cache = edp_link->psr_settings.psr_allow_active;
-			dc_link_set_psr_allow_active(edp_link, &allow_active, false, false, NULL);
+			dc->link_srv->edp_set_psr_allow_active(edp_link, &allow_active, false, false, NULL);
+			dc->link_srv->edp_set_replay_allow_active(edp_link, &allow_active, false, false, NULL);
 		}
 	}
 
@@ -119,13 +126,15 @@ void clk_mgr_optimize_pwr_state(const struct dc *dc, struct clk_mgr *clk_mgr)
 	int edp_num;
 	unsigned int panel_inst;
 
-	get_edp_links(dc, edp_links, &edp_num);
+	dc_get_edp_links(dc, edp_links, &edp_num);
 	if (edp_num) {
 		for (panel_inst = 0; panel_inst < edp_num; panel_inst++) {
 			edp_link = edp_links[panel_inst];
 			if (!edp_link->psr_settings.psr_feature_enabled)
 				continue;
-			dc_link_set_psr_allow_active(edp_link,
+			dc->link_srv->edp_set_psr_allow_active(edp_link,
+					&clk_mgr->psr_allow_active_cache, false, false, NULL);
+			dc->link_srv->edp_set_replay_allow_active(edp_link,
 					&clk_mgr->psr_allow_active_cache, false, false, NULL);
 		}
 	}
@@ -149,7 +158,6 @@ struct clk_mgr *dc_clk_mgr_create(struct dc_context *ctx, struct pp_smu_funcs *p
 			return NULL;
 		}
 		dce60_clk_mgr_construct(ctx, clk_mgr);
-		dce_clk_mgr_construct(ctx, clk_mgr);
 		return &clk_mgr->base;
 	}
 #endif
@@ -211,7 +219,7 @@ struct clk_mgr *dc_clk_mgr_create(struct dc_context *ctx, struct pp_smu_funcs *p
 			dce120_clk_mgr_construct(ctx, clk_mgr);
 		return &clk_mgr->base;
 	}
-#if defined(CONFIG_DRM_AMD_DC_DCN)
+#if defined(CONFIG_DRM_AMD_DC_FP)
 	case FAMILY_RV: {
 		struct clk_mgr_internal *clk_mgr = kzalloc(sizeof(*clk_mgr), GFP_KERNEL);
 
@@ -259,7 +267,7 @@ struct clk_mgr *dc_clk_mgr_create(struct dc_context *ctx, struct pp_smu_funcs *p
 			dcn3_clk_mgr_construct(ctx, clk_mgr, pp_smu, dccg);
 			return &clk_mgr->base;
 		}
-		if (asic_id.chip_id == DEVICE_ID_NV_13FE) {
+		if (ctx->dce_version == DCN_VERSION_2_01) {
 			dcn201_clk_mgr_construct(ctx, clk_mgr, pp_smu, dccg);
 			return &clk_mgr->base;
 		}
@@ -278,6 +286,7 @@ struct clk_mgr *dc_clk_mgr_create(struct dc_context *ctx, struct pp_smu_funcs *p
 			return &clk_mgr->base.base;
 		}
 		break;
+
 	case FAMILY_YELLOW_CARP: {
 		struct clk_mgr_dcn31 *clk_mgr = kzalloc(sizeof(*clk_mgr), GFP_KERNEL);
 
@@ -289,8 +298,83 @@ struct clk_mgr *dc_clk_mgr_create(struct dc_context *ctx, struct pp_smu_funcs *p
 		dcn31_clk_mgr_construct(ctx, clk_mgr, pp_smu, dccg);
 		return &clk_mgr->base.base;
 	}
-#endif
+		break;
+	case AMDGPU_FAMILY_GC_10_3_6: {
+		struct clk_mgr_dcn315 *clk_mgr = kzalloc(sizeof(*clk_mgr), GFP_KERNEL);
+
+		if (clk_mgr == NULL) {
+			BREAK_TO_DEBUGGER();
+			return NULL;
+		}
+
+		dcn315_clk_mgr_construct(ctx, clk_mgr, pp_smu, dccg);
+		return &clk_mgr->base.base;
+	}
+		break;
+	case AMDGPU_FAMILY_GC_10_3_7: {
+		struct clk_mgr_dcn316 *clk_mgr = kzalloc(sizeof(*clk_mgr), GFP_KERNEL);
+
+		if (clk_mgr == NULL) {
+			BREAK_TO_DEBUGGER();
+			return NULL;
+		}
+
+		dcn316_clk_mgr_construct(ctx, clk_mgr, pp_smu, dccg);
+		return &clk_mgr->base.base;
+	}
+		break;
+	case AMDGPU_FAMILY_GC_11_0_0: {
+		struct clk_mgr_internal *clk_mgr = kzalloc(sizeof(*clk_mgr), GFP_KERNEL);
 
+		if (clk_mgr == NULL) {
+			BREAK_TO_DEBUGGER();
+			return NULL;
+		}
+		dcn32_clk_mgr_construct(ctx, clk_mgr, pp_smu, dccg);
+		return &clk_mgr->base;
+	}
+
+	case AMDGPU_FAMILY_GC_11_0_1: {
+		struct clk_mgr_dcn314 *clk_mgr = kzalloc(sizeof(*clk_mgr), GFP_KERNEL);
+
+		if (clk_mgr == NULL) {
+			BREAK_TO_DEBUGGER();
+			return NULL;
+		}
+
+		dcn314_clk_mgr_construct(ctx, clk_mgr, pp_smu, dccg);
+		return &clk_mgr->base.base;
+	}
+	break;
+
+	case AMDGPU_FAMILY_GC_11_5_0: {
+		struct clk_mgr_dcn35 *clk_mgr = kzalloc(sizeof(*clk_mgr), GFP_KERNEL);
+
+		if (clk_mgr == NULL) {
+			BREAK_TO_DEBUGGER();
+			return NULL;
+		}
+		if (ctx->dce_version == DCN_VERSION_3_51)
+			dcn351_clk_mgr_construct(ctx, clk_mgr, pp_smu, dccg);
+		else
+			dcn35_clk_mgr_construct(ctx, clk_mgr, pp_smu, dccg);
+
+		return &clk_mgr->base.base;
+	}
+	break;
+
+	case AMDGPU_FAMILY_GC_12_0_0: {
+		struct clk_mgr_internal *clk_mgr = dcn401_clk_mgr_construct(ctx, dccg);
+
+		if (clk_mgr == NULL) {
+			BREAK_TO_DEBUGGER();
+			return NULL;
+		}
+
+		return &clk_mgr->base;
+	}
+	break;
+#endif	/* CONFIG_DRM_AMD_DC_FP */
 	default:
 		ASSERT(0); /* Unknown Asic */
 		break;
@@ -303,13 +387,12 @@ void dc_destroy_clk_mgr(struct clk_mgr *clk_mgr_base)
 {
 	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
 
-#ifdef CONFIG_DRM_AMD_DC_DCN
+#ifdef CONFIG_DRM_AMD_DC_FP
 	switch (clk_mgr_base->ctx->asic_id.chip_family) {
 	case FAMILY_NV:
 		if (ASICREV_IS_SIENNA_CICHLID_P(clk_mgr_base->ctx->asic_id.hw_internal_rev)) {
 			dcn3_clk_mgr_destroy(clk_mgr);
-		}
-		if (ASICREV_IS_DIMGREY_CAVEFISH_P(clk_mgr_base->ctx->asic_id.hw_internal_rev)) {
+		} else if (ASICREV_IS_DIMGREY_CAVEFISH_P(clk_mgr_base->ctx->asic_id.hw_internal_rev)) {
 			dcn3_clk_mgr_destroy(clk_mgr);
 		}
 		if (ASICREV_IS_BEIGE_GOBY_P(clk_mgr_base->ctx->asic_id.hw_internal_rev)) {
@@ -323,13 +406,36 @@ void dc_destroy_clk_mgr(struct clk_mgr *clk_mgr_base)
 		break;
 
 	case FAMILY_YELLOW_CARP:
-			dcn31_clk_mgr_destroy(clk_mgr);
+		dcn31_clk_mgr_destroy(clk_mgr);
+		break;
+
+	case AMDGPU_FAMILY_GC_10_3_6:
+		dcn315_clk_mgr_destroy(clk_mgr);
+		break;
+
+	case AMDGPU_FAMILY_GC_10_3_7:
+		dcn316_clk_mgr_destroy(clk_mgr);
+		break;
+
+	case AMDGPU_FAMILY_GC_11_0_0:
+		dcn32_clk_mgr_destroy(clk_mgr);
+		break;
+
+	case AMDGPU_FAMILY_GC_11_0_1:
+		dcn314_clk_mgr_destroy(clk_mgr);
+		break;
+
+	case AMDGPU_FAMILY_GC_11_5_0:
+		dcn35_clk_mgr_destroy(clk_mgr);
+		break;
+	case AMDGPU_FAMILY_GC_12_0_0:
+		dcn401_clk_mgr_destroy(clk_mgr);
 		break;
 
 	default:
 		break;
 	}
-#endif
+#endif /* CONFIG_DRM_AMD_DC_FP */
 
 	kfree(clk_mgr);
 }
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dce100/dce_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dce100/dce_clk_mgr.c
index b210f8e9d592..6131ede2db7a 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dce100/dce_clk_mgr.c
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dce100/dce_clk_mgr.c
@@ -72,9 +72,9 @@ static const struct state_dependent_clocks dce80_max_clks_by_state[] = {
 /* ClocksStateLow */
 { .display_clk_khz = 352000, .pixel_clk_khz = 330000},
 /* ClocksStateNominal */
-{ .display_clk_khz = 600000, .pixel_clk_khz = 400000 },
+{ .display_clk_khz = 625000, .pixel_clk_khz = 400000 },
 /* ClocksStatePerformance */
-{ .display_clk_khz = 600000, .pixel_clk_khz = 400000 } };
+{ .display_clk_khz = 625000, .pixel_clk_khz = 400000 } };
 
 int dentist_get_divider_from_did(int did)
 {
@@ -245,6 +245,11 @@ int dce_set_clock(
 	pxl_clk_params.target_pixel_clock_100hz = requested_clk_khz * 10;
 	pxl_clk_params.pll_id = CLOCK_SOURCE_ID_DFS;
 
+	/* DCE 6.0, DCE 6.4: engine clock is the same as PLL0 */
+	if (clk_mgr_base->ctx->dce_version == DCE_VERSION_6_0 ||
+	    clk_mgr_base->ctx->dce_version == DCE_VERSION_6_4)
+		pxl_clk_params.pll_id = CLOCK_SOURCE_ID_PLL0;
+
 	if (clk_mgr_dce->dfs_bypass_active)
 		pxl_clk_params.flags.SET_DISPCLK_DFS_BYPASS = true;
 
@@ -374,6 +379,8 @@ void dce_clock_read_ss_info(struct clk_mgr_internal *clk_mgr_dce)
 				clk_mgr_dce->dprefclk_ss_percentage =
 						info.spread_spectrum_percentage;
 			}
+			if (clk_mgr_dce->base.ctx->dc->config.ignore_dpref_ss)
+				clk_mgr_dce->dprefclk_ss_percentage = 0;
 		}
 	}
 }
@@ -384,8 +391,6 @@ static void dce_pplib_apply_display_requirements(
 {
 	struct dm_pp_display_configuration *pp_display_cfg = &context->pp_display_cfg;
 
-	pp_display_cfg->avail_mclk_switch_time_us = dce110_get_min_vblank_time_us(context);
-
 	dce110_fill_display_configs(context, pp_display_cfg);
 
 	if (memcmp(&dc->current_state->pp_display_cfg, pp_display_cfg, sizeof(*pp_display_cfg)) !=  0)
@@ -398,11 +403,9 @@ static void dce_update_clocks(struct clk_mgr *clk_mgr_base,
 {
 	struct clk_mgr_internal *clk_mgr_dce = TO_CLK_MGR_INTERNAL(clk_mgr_base);
 	struct dm_pp_power_level_change_request level_change_req;
-	int patched_disp_clk = context->bw_ctx.bw.dce.dispclk_khz;
-
-	/*TODO: W/A for dal3 linux, investigate why this works */
-	if (!clk_mgr_dce->dfs_bypass_active)
-		patched_disp_clk = patched_disp_clk * 115 / 100;
+	const int max_disp_clk =
+		clk_mgr_dce->max_clks_by_state[DM_PP_CLOCKS_STATE_PERFORMANCE].display_clk_khz;
+	int patched_disp_clk = MIN(max_disp_clk, context->bw_ctx.bw.dce.dispclk_khz);
 
 	level_change_req.power_level = dce_get_required_clocks_state(clk_mgr_base, context);
 	/* get max clock state from PPLIB */
@@ -460,6 +463,9 @@ void dce_clk_mgr_construct(
 		clk_mgr->max_clks_state = DM_PP_CLOCKS_STATE_NOMINAL;
 	clk_mgr->cur_min_clks_state = DM_PP_CLOCKS_STATE_INVALID;
 
+	base->clks.max_supported_dispclk_khz =
+		clk_mgr->max_clks_by_state[DM_PP_CLOCKS_STATE_PERFORMANCE].display_clk_khz;
+
 	dce_clock_read_integrated_info(clk_mgr);
 	dce_clock_read_ss_info(clk_mgr);
 }
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dce110/dce110_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dce110/dce110_clk_mgr.c
index 78df96882d6e..d50b9440210e 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dce110/dce110_clk_mgr.c
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dce110/dce110_clk_mgr.c
@@ -120,9 +120,15 @@ void dce110_fill_display_configs(
 	const struct dc_state *context,
 	struct dm_pp_display_configuration *pp_display_cfg)
 {
+	struct dc *dc = context->clk_mgr->ctx->dc;
 	int j;
 	int num_cfgs = 0;
 
+	pp_display_cfg->avail_mclk_switch_time_us = dce110_get_min_vblank_time_us(context);
+	pp_display_cfg->disp_clk_khz = dc->clk_mgr->clks.dispclk_khz;
+	pp_display_cfg->avail_mclk_switch_time_in_disp_active_us = 0;
+	pp_display_cfg->crtc_index = dc->res_pool->res_cap->num_timing_generator;
+
 	for (j = 0; j < context->stream_count; j++) {
 		int k;
 
@@ -158,12 +164,29 @@ void dce110_fill_display_configs(
 			stream->link->cur_link_settings.link_rate;
 		cfg->link_settings.link_spread =
 			stream->link->cur_link_settings.link_spread;
-		cfg->sym_clock = stream->phy_pix_clk;
+		cfg->pixel_clock = stream->phy_pix_clk;
 		/* Round v_refresh*/
 		cfg->v_refresh = stream->timing.pix_clk_100hz * 100;
 		cfg->v_refresh /= stream->timing.h_total;
 		cfg->v_refresh = (cfg->v_refresh + stream->timing.v_total / 2)
 							/ stream->timing.v_total;
+
+		/* Find first CRTC index and calculate its line time.
+		 * This is necessary for DPM on SI GPUs.
+		 */
+		if (cfg->pipe_idx < pp_display_cfg->crtc_index) {
+			const struct dc_crtc_timing *timing =
+				&context->streams[0]->timing;
+
+			pp_display_cfg->crtc_index = cfg->pipe_idx;
+			pp_display_cfg->line_time_in_us =
+				timing->h_total * 10000 / timing->pix_clk_100hz;
+		}
+	}
+
+	if (!num_cfgs) {
+		pp_display_cfg->crtc_index = 0;
+		pp_display_cfg->line_time_in_us = 0;
 	}
 
 	pp_display_cfg->display_count = num_cfgs;
@@ -195,7 +218,7 @@ void dce11_pplib_apply_display_requirements(
 	 * , then change minimum memory clock based on real-time bandwidth
 	 * limitation.
 	 */
-	if ((dc->ctx->asic_id.chip_family == FAMILY_AI) &&
+	if (dc->bw_vbios && (dc->ctx->asic_id.chip_family == FAMILY_AI) &&
 	     ASICREV_IS_VEGA20_P(dc->ctx->asic_id.hw_internal_rev) && (context->stream_count >= 2)) {
 		pp_display_cfg->min_memory_clock_khz = max(pp_display_cfg->min_memory_clock_khz,
 							   (uint32_t) div64_s64(
@@ -223,25 +246,8 @@ void dce11_pplib_apply_display_requirements(
 	pp_display_cfg->min_engine_clock_deep_sleep_khz
 			= context->bw_ctx.bw.dce.sclk_deep_sleep_khz;
 
-	pp_display_cfg->avail_mclk_switch_time_us =
-						dce110_get_min_vblank_time_us(context);
-	/* TODO: dce11.2*/
-	pp_display_cfg->avail_mclk_switch_time_in_disp_active_us = 0;
-
-	pp_display_cfg->disp_clk_khz = dc->clk_mgr->clks.dispclk_khz;
-
 	dce110_fill_display_configs(context, pp_display_cfg);
 
-	/* TODO: is this still applicable?*/
-	if (pp_display_cfg->display_count == 1) {
-		const struct dc_crtc_timing *timing =
-			&context->streams[0]->timing;
-
-		pp_display_cfg->crtc_index =
-			pp_display_cfg->disp_configs[0].pipe_idx;
-		pp_display_cfg->line_time_in_us = timing->h_total * 10000 / timing->pix_clk_100hz;
-	}
-
 	if (memcmp(&dc->current_state->pp_display_cfg, pp_display_cfg, sizeof(*pp_display_cfg)) !=  0)
 		dm_pp_apply_display_requirements(dc->ctx, pp_display_cfg);
 }
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dce112/dce112_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dce112/dce112_clk_mgr.c
index 934e6423dc1a..1f36ad8a7de4 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dce112/dce112_clk_mgr.c
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dce112/dce112_clk_mgr.c
@@ -111,12 +111,10 @@ int dce112_set_clock(struct clk_mgr *clk_mgr_base, int requested_clk_khz)
 
 	bp->funcs->set_dce_clock(bp, &dce_clk_params);
 
-	if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) {
-		if (dmcu && dmcu->funcs->is_dmcu_initialized(dmcu)) {
-			if (clk_mgr_dce->dfs_bypass_disp_clk != actual_clock)
-				dmcu->funcs->set_psr_wait_loop(dmcu,
-						actual_clock / 1000 / 7);
-		}
+	if (dmcu && dmcu->funcs->is_dmcu_initialized(dmcu)) {
+		if (clk_mgr_dce->dfs_bypass_disp_clk != actual_clock)
+			dmcu->funcs->set_psr_wait_loop(dmcu,
+					actual_clock / 1000 / 7);
 	}
 
 	clk_mgr_dce->dfs_bypass_disp_clk = actual_clock;
@@ -153,12 +151,10 @@ int dce112_set_dispclk(struct clk_mgr_internal *clk_mgr, int requested_clk_khz)
 		clk_mgr->cur_min_clks_state = DM_PP_CLOCKS_STATE_NOMINAL;
 
 
-	if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) {
-		if (dmcu && dmcu->funcs->is_dmcu_initialized(dmcu)) {
-			if (clk_mgr->dfs_bypass_disp_clk != actual_clock)
-				dmcu->funcs->set_psr_wait_loop(dmcu,
-						actual_clock / 1000 / 7);
-		}
+	if (dmcu && dmcu->funcs->is_dmcu_initialized(dmcu)) {
+		if (clk_mgr->dfs_bypass_disp_clk != actual_clock)
+			dmcu->funcs->set_psr_wait_loop(dmcu,
+					actual_clock / 1000 / 7);
 	}
 
 	clk_mgr->dfs_bypass_disp_clk = actual_clock;
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dce120/dce120_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dce120/dce120_clk_mgr.c
index 5399b8cf6b75..c9ba7b3fd2c3 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dce120/dce120_clk_mgr.c
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dce120/dce120_clk_mgr.c
@@ -30,7 +30,7 @@
 #include "dce110/dce110_clk_mgr.h"
 #include "dce120_clk_mgr.h"
 #include "dce100/dce_clk_mgr.h"
-#include "dce120/dce120_hw_sequencer.h"
+#include "dce120/dce120_hwseq.h"
 
 static const struct state_dependent_clocks dce120_max_clks_by_state[] = {
 /*ClocksStateInvalid - should not be used*/
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dce60/dce60_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dce60/dce60_clk_mgr.c
index 0267644717b2..69dd80d9f738 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dce60/dce60_clk_mgr.c
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dce60/dce60_clk_mgr.c
@@ -83,22 +83,13 @@ static const struct state_dependent_clocks dce60_max_clks_by_state[] = {
 static int dce60_get_dp_ref_freq_khz(struct clk_mgr *clk_mgr_base)
 {
 	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
-	int dprefclk_wdivider;
-	int dp_ref_clk_khz;
-	int target_div;
+	struct dc_context *ctx = clk_mgr_base->ctx;
+	int dp_ref_clk_khz = 0;
 
-	/* DCE6 has no DPREFCLK_CNTL to read DP Reference Clock source */
-
-	/* Read the mmDENTIST_DISPCLK_CNTL to get the currently
-	 * programmed DID DENTIST_DPREFCLK_WDIVIDER*/
-	REG_GET(DENTIST_DISPCLK_CNTL, DENTIST_DPREFCLK_WDIVIDER, &dprefclk_wdivider);
-
-	/* Convert DENTIST_DPREFCLK_WDIVIDERto actual divider*/
-	target_div = dentist_get_divider_from_did(dprefclk_wdivider);
-
-	/* Calculate the current DFS clock, in kHz.*/
-	dp_ref_clk_khz = (DENTIST_DIVIDER_RANGE_SCALE_FACTOR
-		* clk_mgr->base.dentist_vco_freq_khz) / target_div;
+	if (ASIC_REV_IS_TAHITI_P(ctx->asic_id.hw_internal_rev))
+		dp_ref_clk_khz = ctx->dc_bios->fw_info.default_display_engine_pll_frequency;
+	else
+		dp_ref_clk_khz = clk_mgr_base->clks.dispclk_khz;
 
 	return dce_adjust_dp_ref_freq_for_ss(clk_mgr, dp_ref_clk_khz);
 }
@@ -109,8 +100,6 @@ static void dce60_pplib_apply_display_requirements(
 {
 	struct dm_pp_display_configuration *pp_display_cfg = &context->pp_display_cfg;
 
-	pp_display_cfg->avail_mclk_switch_time_us = dce110_get_min_vblank_time_us(context);
-
 	dce110_fill_display_configs(context, pp_display_cfg);
 
 	if (memcmp(&dc->current_state->pp_display_cfg, pp_display_cfg, sizeof(*pp_display_cfg)) !=  0)
@@ -123,11 +112,9 @@ static void dce60_update_clocks(struct clk_mgr *clk_mgr_base,
 {
 	struct clk_mgr_internal *clk_mgr_dce = TO_CLK_MGR_INTERNAL(clk_mgr_base);
 	struct dm_pp_power_level_change_request level_change_req;
-	int patched_disp_clk = context->bw_ctx.bw.dce.dispclk_khz;
-
-	/*TODO: W/A for dal3 linux, investigate why this works */
-	if (!clk_mgr_dce->dfs_bypass_active)
-		patched_disp_clk = patched_disp_clk * 115 / 100;
+	const int max_disp_clk =
+		clk_mgr_dce->max_clks_by_state[DM_PP_CLOCKS_STATE_PERFORMANCE].display_clk_khz;
+	int patched_disp_clk = MIN(max_disp_clk, context->bw_ctx.bw.dce.dispclk_khz);
 
 	level_change_req.power_level = dce_get_required_clocks_state(clk_mgr_base, context);
 	/* get max clock state from PPLIB */
@@ -160,6 +147,8 @@ void dce60_clk_mgr_construct(
 		struct dc_context *ctx,
 		struct clk_mgr_internal *clk_mgr)
 {
+	struct clk_mgr *base = &clk_mgr->base;
+
 	dce_clk_mgr_construct(ctx, clk_mgr);
 
 	memcpy(clk_mgr->max_clks_by_state,
@@ -170,5 +159,8 @@ void dce60_clk_mgr_construct(
 	clk_mgr->clk_mgr_shift = &disp_clk_shift;
 	clk_mgr->clk_mgr_mask = &disp_clk_mask;
 	clk_mgr->base.funcs = &dce60_funcs;
+
+	base->clks.max_supported_dispclk_khz =
+		clk_mgr->max_clks_by_state[DM_PP_CLOCKS_STATE_PERFORMANCE].display_clk_khz;
 }
 
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn10/rv1_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn10/rv1_clk_mgr.c
index 76ec8ec92efd..369421e46c52 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn10/rv1_clk_mgr.c
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn10/rv1_clk_mgr.c
@@ -23,9 +23,6 @@
  *
  */
 
-#include <linux/slab.h>
-
-#include "reg_helper.h"
 #include "core_types.h"
 #include "clk_mgr_internal.h"
 #include "rv1_clk_mgr.h"
@@ -34,7 +31,7 @@
 #include "rv1_clk_mgr_vbios_smu.h"
 #include "rv1_clk_mgr_clk.h"
 
-void rv1_init_clocks(struct clk_mgr *clk_mgr)
+static void rv1_init_clocks(struct clk_mgr *clk_mgr)
 {
 	memset(&(clk_mgr->clks), 0, sizeof(struct dc_clocks));
 }
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn10/rv1_clk_mgr_clk.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn10/rv1_clk_mgr_clk.c
deleted file mode 100644
index 61dd12198a3c..000000000000
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn10/rv1_clk_mgr_clk.c
+++ /dev/null
@@ -1,79 +0,0 @@
-/*
- * Copyright 2012-16 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: AMD
- *
- */
-
-#include "reg_helper.h"
-#include "clk_mgr_internal.h"
-#include "rv1_clk_mgr_clk.h"
-
-#include "ip/Discovery/hwid.h"
-#include "ip/Discovery/v1/ip_offset_1.h"
-#include "ip/CLK/clk_10_0_default.h"
-#include "ip/CLK/clk_10_0_offset.h"
-#include "ip/CLK/clk_10_0_reg.h"
-#include "ip/CLK/clk_10_0_sh_mask.h"
-
-#include "dce100/dce_clk_mgr.h"
-
-#define CLK_BASE_INNER(inst) \
-	CLK_BASE__INST ## inst ## _SEG0
-
-
-#define CLK_REG(reg_name, block, inst)\
-	CLK_BASE(mm ## block ## _ ## inst ## _ ## reg_name ## _BASE_IDX) + \
-					mm ## block ## _ ## inst ## _ ## reg_name
-
-#define REG(reg_name) \
-	CLK_REG(reg_name, CLK0, 0)
-
-
-/* Only used by testing framework*/
-void rv1_dump_clk_registers(struct clk_state_registers *regs, struct clk_bypass *bypass, struct clk_mgr *clk_mgr_base)
-{
-	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
-
-		regs->CLK0_CLK8_CURRENT_CNT = REG_READ(CLK0_CLK8_CURRENT_CNT) / 10; //dcf clk
-
-		bypass->dcfclk_bypass = REG_READ(CLK0_CLK8_BYPASS_CNTL) & 0x0007;
-		if (bypass->dcfclk_bypass < 0 || bypass->dcfclk_bypass > 4)
-			bypass->dcfclk_bypass = 0;
-
-
-		regs->CLK0_CLK8_DS_CNTL = REG_READ(CLK0_CLK8_DS_CNTL) / 10;	//dcf deep sleep divider
-
-		regs->CLK0_CLK8_ALLOW_DS = REG_READ(CLK0_CLK8_ALLOW_DS); //dcf deep sleep allow
-
-		regs->CLK0_CLK10_CURRENT_CNT = REG_READ(CLK0_CLK10_CURRENT_CNT) / 10; //dpref clk
-
-		bypass->dispclk_pypass = REG_READ(CLK0_CLK10_BYPASS_CNTL) & 0x0007;
-		if (bypass->dispclk_pypass < 0 || bypass->dispclk_pypass > 4)
-			bypass->dispclk_pypass = 0;
-
-		regs->CLK0_CLK11_CURRENT_CNT = REG_READ(CLK0_CLK11_CURRENT_CNT) / 10; //disp clk
-
-		bypass->dprefclk_bypass = REG_READ(CLK0_CLK11_BYPASS_CNTL) & 0x0007;
-		if (bypass->dprefclk_bypass < 0 || bypass->dprefclk_bypass > 4)
-			bypass->dprefclk_bypass = 0;
-
-}
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn10/rv1_clk_mgr_vbios_smu.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn10/rv1_clk_mgr_vbios_smu.c
index fe18bb9e19aa..d82a52319088 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn10/rv1_clk_mgr_vbios_smu.c
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn10/rv1_clk_mgr_vbios_smu.c
@@ -26,7 +26,8 @@
 #include "core_types.h"
 #include "clk_mgr_internal.h"
 #include "reg_helper.h"
-#include <linux/delay.h>
+
+#include "rv1_clk_mgr_vbios_smu.h"
 
 #define MAX_INSTANCE	5
 #define MAX_SEGMENT		5
@@ -99,7 +100,8 @@ static uint32_t rv1_smu_wait_for_response(struct clk_mgr_internal *clk_mgr, unsi
 	return res_val;
 }
 
-int rv1_vbios_smu_send_msg_with_param(struct clk_mgr_internal *clk_mgr, unsigned int msg_id, unsigned int param)
+static int rv1_vbios_smu_send_msg_with_param(struct clk_mgr_internal *clk_mgr,
+		unsigned int msg_id, unsigned int param)
 {
 	uint32_t result;
 
@@ -132,27 +134,11 @@ int rv1_vbios_smu_set_dispclk(struct clk_mgr_internal *clk_mgr, int requested_di
 			VBIOSSMC_MSG_SetDispclkFreq,
 			khz_to_mhz_ceil(requested_dispclk_khz));
 
-	if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) {
-		if (dmcu && dmcu->funcs->is_dmcu_initialized(dmcu)) {
-			if (clk_mgr->dfs_bypass_disp_clk != actual_dispclk_set_mhz)
-				dmcu->funcs->set_psr_wait_loop(dmcu,
-						actual_dispclk_set_mhz / 7);
-		}
+	if (dmcu && dmcu->funcs->is_dmcu_initialized(dmcu)) {
+		if (clk_mgr->dfs_bypass_disp_clk != actual_dispclk_set_mhz)
+			dmcu->funcs->set_psr_wait_loop(dmcu,
+					actual_dispclk_set_mhz / 7);
 	}
 
 	return actual_dispclk_set_mhz * 1000;
 }
-
-int rv1_vbios_smu_set_dprefclk(struct clk_mgr_internal *clk_mgr)
-{
-	int actual_dprefclk_set_mhz = -1;
-
-	actual_dprefclk_set_mhz = rv1_vbios_smu_send_msg_with_param(
-			clk_mgr,
-			VBIOSSMC_MSG_SetDprefclkFreq,
-			khz_to_mhz_ceil(clk_mgr->base.dprefclk_khz));
-
-	/* TODO: add code for programing DP DTO, currently this is down by command table */
-
-	return actual_dprefclk_set_mhz * 1000;
-}
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn10/rv1_clk_mgr_vbios_smu.h b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn10/rv1_clk_mgr_vbios_smu.h
index 083cb3158859..81d7c912549c 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn10/rv1_clk_mgr_vbios_smu.h
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn10/rv1_clk_mgr_vbios_smu.h
@@ -27,6 +27,5 @@
 #define DAL_DC_DCN10_RV1_CLK_MGR_VBIOS_SMU_H_
 
 int rv1_vbios_smu_set_dispclk(struct clk_mgr_internal *clk_mgr, int requested_dispclk_khz);
-int rv1_vbios_smu_set_dprefclk(struct clk_mgr_internal *clk_mgr);
 
 #endif /* DAL_DC_DCN10_RV1_CLK_MGR_VBIOS_SMU_H_ */
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn20/dcn20_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn20/dcn20_clk_mgr.c
index 2108bff49d4e..bb4f3bd7532e 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn20/dcn20_clk_mgr.c
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn20/dcn20_clk_mgr.c
@@ -38,7 +38,6 @@
 #include "clk/clk_11_0_0_offset.h"
 #include "clk/clk_11_0_0_sh_mask.h"
 
-#include "irq/dcn20/irq_service_dcn20.h"
 
 #undef FN
 #define FN(reg_name, field_name) \
@@ -127,16 +126,24 @@ void dcn20_update_clocks_update_dpp_dto(struct clk_mgr_internal *clk_mgr,
 
 void dcn20_update_clocks_update_dentist(struct clk_mgr_internal *clk_mgr, struct dc_state *context)
 {
-	int dpp_divider = DENTIST_DIVIDER_RANGE_SCALE_FACTOR
-			* clk_mgr->base.dentist_vco_freq_khz / clk_mgr->base.clks.dppclk_khz;
-	int disp_divider = DENTIST_DIVIDER_RANGE_SCALE_FACTOR
-			* clk_mgr->base.dentist_vco_freq_khz / clk_mgr->base.clks.dispclk_khz;
-
-	uint32_t dppclk_wdivider = dentist_get_did_from_divider(dpp_divider);
-	uint32_t dispclk_wdivider = dentist_get_did_from_divider(disp_divider);
+	int dpp_divider = 0;
+	int disp_divider = 0;
+	uint32_t dppclk_wdivider = 0;
+	uint32_t dispclk_wdivider = 0;
 	uint32_t current_dispclk_wdivider;
 	uint32_t i;
 
+	if (clk_mgr->base.clks.dppclk_khz == 0 || clk_mgr->base.clks.dispclk_khz == 0)
+		return;
+
+	dpp_divider = DENTIST_DIVIDER_RANGE_SCALE_FACTOR
+		* clk_mgr->base.dentist_vco_freq_khz / clk_mgr->base.clks.dppclk_khz;
+	disp_divider = DENTIST_DIVIDER_RANGE_SCALE_FACTOR
+		* clk_mgr->base.dentist_vco_freq_khz / clk_mgr->base.clks.dispclk_khz;
+
+	dppclk_wdivider = dentist_get_did_from_divider(dpp_divider);
+	dispclk_wdivider = dentist_get_did_from_divider(disp_divider);
+
 	REG_GET(DENTIST_DISPCLK_CNTL,
 			DENTIST_DISPCLK_WDIVIDER, &current_dispclk_wdivider);
 
@@ -150,7 +157,7 @@ void dcn20_update_clocks_update_dentist(struct clk_mgr_internal *clk_mgr, struct
 			int32_t N;
 			int32_t j;
 
-			if (!pipe_ctx->stream)
+			if (!resource_is_pipe_type(pipe_ctx, OTG_MASTER))
 				continue;
 			/* Virtual encoders don't have this function */
 			if (!stream_enc->funcs->get_fifo_cal_average_level)
@@ -172,7 +179,7 @@ void dcn20_update_clocks_update_dentist(struct clk_mgr_internal *clk_mgr, struct
 	} else if (dispclk_wdivider == 127 && current_dispclk_wdivider != 127) {
 		REG_UPDATE(DENTIST_DISPCLK_CNTL,
 				DENTIST_DISPCLK_WDIVIDER, 126);
-		REG_WAIT(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_CHG_DONE, 1, 50, 100);
+		REG_WAIT(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_CHG_DONE, 1, 50, 2000);
 		for (i = 0; i < clk_mgr->base.ctx->dc->res_pool->pipe_count; i++) {
 			struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i];
 			struct dccg *dccg = clk_mgr->base.ctx->dc->res_pool->dccg;
@@ -181,7 +188,7 @@ void dcn20_update_clocks_update_dentist(struct clk_mgr_internal *clk_mgr, struct
 			int32_t N;
 			int32_t j;
 
-			if (!pipe_ctx->stream)
+			if (!resource_is_pipe_type(pipe_ctx, OTG_MASTER))
 				continue;
 			/* Virtual encoders don't have this function */
 			if (!stream_enc->funcs->get_fifo_cal_average_level)
@@ -199,7 +206,7 @@ void dcn20_update_clocks_update_dentist(struct clk_mgr_internal *clk_mgr, struct
 
 	REG_UPDATE(DENTIST_DISPCLK_CNTL,
 			DENTIST_DISPCLK_WDIVIDER, dispclk_wdivider);
-	REG_WAIT(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_CHG_DONE, 1, 50, 1000);
+	REG_WAIT(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_CHG_DONE, 1, 50, 2000);
 	REG_UPDATE(DENTIST_DISPCLK_CNTL,
 			DENTIST_DPPCLK_WDIVIDER, dppclk_wdivider);
 	REG_WAIT(DENTIST_DISPCLK_CNTL, DENTIST_DPPCLK_CHG_DONE, 1, 5, 100);
@@ -223,8 +230,6 @@ void dcn2_update_clocks(struct clk_mgr *clk_mgr_base,
 	bool force_reset = false;
 	bool p_state_change_support;
 	int total_plane_count;
-	int irq_src;
-	uint32_t hpd_state;
 
 	if (dc->work_arounds.skip_clock_update)
 		return;
@@ -242,13 +247,7 @@ void dcn2_update_clocks(struct clk_mgr *clk_mgr_base,
 	if (dc->res_pool->pp_smu)
 		pp_smu = &dc->res_pool->pp_smu->nv_funcs;
 
-	for (irq_src = DC_IRQ_SOURCE_HPD1; irq_src <= DC_IRQ_SOURCE_HPD6; irq_src++) {
-		hpd_state = dc_get_hpd_state_dcn20(dc->res_pool->irqs, irq_src);
-		if (hpd_state)
-			break;
-	}
-
-	if (display_count == 0 && !hpd_state)
+	if (display_count == 0)
 		enter_display_off = true;
 
 	if (enter_display_off == safe_to_lower) {
@@ -409,7 +408,7 @@ void dcn2_init_clocks(struct clk_mgr *clk_mgr)
 	clk_mgr->clks.prev_p_state_change_support = true;
 }
 
-void dcn2_enable_pme_wa(struct clk_mgr *clk_mgr_base)
+static void dcn2_enable_pme_wa(struct clk_mgr *clk_mgr_base)
 {
 	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
 	struct pp_smu_funcs_nv *pp_smu = NULL;
@@ -445,7 +444,6 @@ void dcn2_read_clocks_from_hw_dentist(struct clk_mgr *clk_mgr_base)
 		clk_mgr_base->clks.dppclk_khz = (DENTIST_DIVIDER_RANGE_SCALE_FACTOR
 				* clk_mgr->base.dentist_vco_freq_khz) / dpp_divider;
 	}
-
 }
 
 void dcn2_get_clock(struct clk_mgr *clk_mgr,
@@ -505,7 +503,7 @@ static void dcn2_notify_link_rate_change(struct clk_mgr *clk_mgr_base, struct dc
 
 	clk_mgr->cur_phyclk_req_table[link->link_index] = link->cur_link_settings.link_rate * LINK_RATE_REF_FREQ_IN_KHZ;
 
-	for (i = 0; i < MAX_PIPES * 2; i++) {
+	for (i = 0; i < MAX_LINKS; i++) {
 		if (clk_mgr->cur_phyclk_req_table[i] > max_phyclk_req)
 			max_phyclk_req = clk_mgr->cur_phyclk_req_table[i];
 	}
@@ -533,6 +531,11 @@ void dcn20_clk_mgr_construct(
 		struct pp_smu_funcs *pp_smu,
 		struct dccg *dccg)
 {
+	int dprefclk_did;
+	int target_div;
+	uint32_t pll_req_reg;
+	struct fixed31_32 pll_req;
+
 	clk_mgr->base.ctx = ctx;
 	clk_mgr->pp_smu = pp_smu;
 	clk_mgr->base.funcs = &dcn2_funcs;
@@ -549,42 +552,34 @@ void dcn20_clk_mgr_construct(
 
 	clk_mgr->base.dprefclk_khz = 700000; // 700 MHz planned if VCO is 3.85 GHz, will be retrieved
 
-	if (IS_FPGA_MAXIMUS_DC(ctx->dce_environment)) {
-		dcn2_funcs.update_clocks = dcn2_update_clocks_fpga;
-		clk_mgr->base.dentist_vco_freq_khz = 3850000;
+	/* DFS Slice 2 should be used for DPREFCLK */
+	dprefclk_did = REG_READ(CLK3_CLK2_DFS_CNTL);
+	/* Convert DPREFCLK DFS Slice DID to actual divider */
+	target_div = dentist_get_divider_from_did(dprefclk_did);
+	/* get FbMult value */
+	pll_req_reg = REG_READ(CLK3_CLK_PLL_REQ);
 
-	} else {
-		/* DFS Slice 2 should be used for DPREFCLK */
-		int dprefclk_did = REG_READ(CLK3_CLK2_DFS_CNTL);
-		/* Convert DPREFCLK DFS Slice DID to actual divider*/
-		int target_div = dentist_get_divider_from_did(dprefclk_did);
-
-		/* get FbMult value */
-		uint32_t pll_req_reg = REG_READ(CLK3_CLK_PLL_REQ);
-		struct fixed31_32 pll_req;
-
-		/* set up a fixed-point number
-		 * this works because the int part is on the right edge of the register
-		 * and the frac part is on the left edge
-		 */
+	/* set up a fixed-point number
+	 * this works because the int part is on the right edge of the register
+	 * and the frac part is on the left edge
+	 */
 
-		pll_req = dc_fixpt_from_int(pll_req_reg & clk_mgr->clk_mgr_mask->FbMult_int);
-		pll_req.value |= pll_req_reg & clk_mgr->clk_mgr_mask->FbMult_frac;
+	pll_req = dc_fixpt_from_int(pll_req_reg & clk_mgr->clk_mgr_mask->FbMult_int);
+	pll_req.value |= pll_req_reg & clk_mgr->clk_mgr_mask->FbMult_frac;
 
-		/* multiply by REFCLK period */
-		pll_req = dc_fixpt_mul_int(pll_req, 100000);
+	/* multiply by REFCLK period */
+	pll_req = dc_fixpt_mul_int(pll_req, 100000);
 
-		/* integer part is now VCO frequency in kHz */
-		clk_mgr->base.dentist_vco_freq_khz = dc_fixpt_floor(pll_req);
+	/* integer part is now VCO frequency in kHz */
+	clk_mgr->base.dentist_vco_freq_khz = dc_fixpt_floor(pll_req);
 
-		/* in case we don't get a value from the register, use default */
-		if (clk_mgr->base.dentist_vco_freq_khz == 0)
-			clk_mgr->base.dentist_vco_freq_khz = 3850000;
+	/* in case we don't get a value from the register, use default */
+	if (clk_mgr->base.dentist_vco_freq_khz == 0)
+		clk_mgr->base.dentist_vco_freq_khz = 3850000;
 
-		/* Calculate the DPREFCLK in kHz.*/
-		clk_mgr->base.dprefclk_khz = (DENTIST_DIVIDER_RANGE_SCALE_FACTOR
-			* clk_mgr->base.dentist_vco_freq_khz) / target_div;
-	}
+	/* Calculate the DPREFCLK in kHz.*/
+	clk_mgr->base.dprefclk_khz = (DENTIST_DIVIDER_RANGE_SCALE_FACTOR
+		* clk_mgr->base.dentist_vco_freq_khz) / target_div;
 	//Integrated_info table does not exist on dGPU projects so should not be referenced
 	//anywhere in code for dGPUs.
 	//Also there is no plan for now that DFS BYPASS will be used on NV10/12/14.
@@ -592,4 +587,3 @@ void dcn20_clk_mgr_construct(
 
 	dce_clock_read_ss_info(clk_mgr);
 }
-
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn201/dcn201_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn201/dcn201_clk_mgr.c
index db9950244c7b..76c612ecfe3c 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn201/dcn201_clk_mgr.c
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn201/dcn201_clk_mgr.c
@@ -34,8 +34,8 @@
 #include "dm_services.h"
 
 #include "cyan_skillfish_ip_offset.h"
-#include "dcn/dcn_2_0_3_offset.h"
-#include "dcn/dcn_2_0_3_sh_mask.h"
+#include "dcn/dcn_2_0_1_offset.h"
+#include "dcn/dcn_2_0_1_sh_mask.h"
 #include "clk/clk_11_0_1_offset.h"
 #include "clk/clk_11_0_1_sh_mask.h"
 
@@ -59,8 +59,6 @@
 
 #define CTX \
 	clk_mgr->base.ctx
-#define DC_LOGGER \
-	clk_mgr->base.ctx->logger
 
 static const struct clk_mgr_registers clk_mgr_regs = {
 		CLK_COMMON_REG_LIST_DCN_201()
@@ -74,42 +72,6 @@ static const struct clk_mgr_mask clk_mgr_mask = {
 	CLK_COMMON_MASK_SH_LIST_DCN201_BASE(_MASK)
 };
 
-void dcn201_update_clocks_vbios(struct clk_mgr *clk_mgr,
-			struct dc_state *context,
-			bool safe_to_lower)
-{
-	struct dc_clocks *new_clocks = &context->bw_ctx.bw.dcn.clk;
-
-	bool update_dppclk = false;
-	bool update_dispclk = false;
-
-	if (should_set_clock(safe_to_lower, new_clocks->dppclk_khz, clk_mgr->clks.dppclk_khz)) {
-		clk_mgr->clks.dppclk_khz = new_clocks->dppclk_khz;
-		update_dppclk = true;
-	}
-
-	if (should_set_clock(safe_to_lower, new_clocks->dispclk_khz, clk_mgr->clks.dispclk_khz)) {
-		clk_mgr->clks.dispclk_khz = new_clocks->dispclk_khz;
-		update_dispclk = true;
-	}
-
-	if (update_dppclk || update_dispclk) {
-		struct bp_set_dce_clock_parameters dce_clk_params;
-		struct dc_bios *bp = clk_mgr->ctx->dc_bios;
-
-		if (update_dispclk) {
-			memset(&dce_clk_params, 0, sizeof(dce_clk_params));
-			dce_clk_params.target_clock_frequency = new_clocks->dispclk_khz;
-			dce_clk_params.pll_id = CLOCK_SOURCE_ID_DFS;
-			dce_clk_params.clock_type = DCECLOCK_TYPE_DISPLAY_CLOCK;
-			bp->funcs->set_dce_clock(bp, &dce_clk_params);
-		}
-		/* currently there is no DCECLOCK_TYPE_DPPCLK type defined in VBIOS interface.
-		 * vbios program DPPCLK to the same DispCLK limitation
-		 */
-	}
-}
-
 static void dcn201_init_clocks(struct clk_mgr *clk_mgr)
 {
 	memset(&(clk_mgr->clks), 0, sizeof(struct dc_clocks));
@@ -126,10 +88,8 @@ static void dcn201_update_clocks(struct clk_mgr *clk_mgr_base,
 	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
 	struct dc_clocks *new_clocks = &context->bw_ctx.bw.dcn.clk;
 	struct dc *dc = clk_mgr_base->ctx->dc;
-	int display_count;
 	bool update_dppclk = false;
 	bool update_dispclk = false;
-	bool enter_display_off = false;
 	bool dpp_clock_lowered = false;
 	bool force_reset = false;
 	bool p_state_change_support;
@@ -139,17 +99,20 @@ static void dcn201_update_clocks(struct clk_mgr *clk_mgr_base,
 		return;
 
 	if (clk_mgr_base->clks.dispclk_khz == 0 ||
-		dc->debug.force_clock_mode & 0x1) {
-			force_reset = true;
+	    dc->debug.force_clock_mode & 0x1) {
+		/* this is from resume or boot up, if forced_clock cfg option
+		 * used, we bypass program dispclk and DPPCLK, but need set them
+		 * for S3.
+		 */
+
+		force_reset = true;
+		/* force_clock_mode 0x1:  force reset the clock even it is the
+		 * same clock as long as it is in Passive level.
+		 */
 
 		dcn2_read_clocks_from_hw_dentist(clk_mgr_base);
 	}
 
-	display_count = clk_mgr_helper_get_active_display_cnt(dc, context);
-
-	if (display_count == 0)
-		enter_display_off = true;
-
 	if (should_set_clock(safe_to_lower, new_clocks->phyclk_khz, clk_mgr_base->clks.phyclk_khz))
 		clk_mgr_base->clks.phyclk_khz = new_clocks->phyclk_khz;
 
@@ -193,18 +156,20 @@ static void dcn201_update_clocks(struct clk_mgr *clk_mgr_base,
 
 	if (dc->config.forced_clocks == false || (force_reset && safe_to_lower)) {
 		if (dpp_clock_lowered) {
+			// if clock is being lowered, increase DTO before lowering refclk
 			dcn20_update_clocks_update_dpp_dto(clk_mgr, context, safe_to_lower);
 			dcn20_update_clocks_update_dentist(clk_mgr, context);
 		} else {
+			// if clock is being raised, increase refclk before lowering DTO
 			if (update_dppclk || update_dispclk)
 				dcn20_update_clocks_update_dentist(clk_mgr, context);
-			if (new_clocks->dppclk_khz >= dc->current_state->bw_ctx.bw.dcn.clk.dppclk_khz)
-				dcn20_update_clocks_update_dpp_dto(clk_mgr, context, safe_to_lower);
+			// always update dtos unless clock is lowered and not safe to lower
+			dcn20_update_clocks_update_dpp_dto(clk_mgr, context, safe_to_lower);
 		}
 	}
 }
 
-struct clk_mgr_funcs dcn201_funcs = {
+static struct clk_mgr_funcs dcn201_funcs = {
 	.get_dp_ref_clk_frequency = dce12_get_dp_ref_freq_khz,
 	.update_clocks = dcn201_update_clocks,
 	.init_clocks = dcn201_init_clocks,
@@ -232,23 +197,17 @@ void dcn201_clk_mgr_construct(struct dc_context *ctx,
 	clk_mgr->dprefclk_ss_divider = 1000;
 	clk_mgr->ss_on_dprefclk = false;
 
-	if (IS_FPGA_MAXIMUS_DC(ctx->dce_environment)) {
-		dcn201_funcs.update_clocks = dcn2_update_clocks_fpga;
-		clk_mgr->base.dprefclk_khz = 600000;
-		clk_mgr->base.dentist_vco_freq_khz = 3000000;
-	} else {
-		clk_mgr->base.dprefclk_khz = REG_READ(CLK4_CLK2_CURRENT_CNT);
-		clk_mgr->base.dprefclk_khz *= 100;
+	clk_mgr->base.dprefclk_khz = REG_READ(CLK4_CLK2_CURRENT_CNT);
+	clk_mgr->base.dprefclk_khz *= 100;
 
-		if (clk_mgr->base.dprefclk_khz == 0)
-			clk_mgr->base.dprefclk_khz = 600000;
+	if (clk_mgr->base.dprefclk_khz == 0)
+		clk_mgr->base.dprefclk_khz = 600000;
 
-		REG_GET(CLK4_CLK_PLL_REQ, FbMult_int, &clk_mgr->base.dentist_vco_freq_khz);
-		clk_mgr->base.dentist_vco_freq_khz *= 100000;
+	REG_GET(CLK4_CLK_PLL_REQ, FbMult_int, &clk_mgr->base.dentist_vco_freq_khz);
+	clk_mgr->base.dentist_vco_freq_khz *= 100000;
 
-		if (clk_mgr->base.dentist_vco_freq_khz == 0)
-			clk_mgr->base.dentist_vco_freq_khz = 3000000;
-	}
+	if (clk_mgr->base.dentist_vco_freq_khz == 0)
+		clk_mgr->base.dentist_vco_freq_khz = 3000000;
 
 	if (!debug->disable_dfs_bypass && bp->integrated_info)
 		if (bp->integrated_info->gpu_cap_info & DFS_BYPASS_ENABLE)
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn21/rn_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn21/rn_clk_mgr.c
index ac2d4c4f04e4..e18097f82091 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn21/rn_clk_mgr.c
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn21/rn_clk_mgr.c
@@ -24,12 +24,10 @@
  */
 
 #include "dccg.h"
-#include "clk_mgr_internal.h"
-
-
-#include "dcn20/dcn20_clk_mgr.h"
 #include "rn_clk_mgr.h"
 
+#include "dcn20/dcn20_clk_mgr.h"
+#include "dml/dcn20/dcn20_fpu.h"
 
 #include "dce100/dce_clk_mgr.h"
 #include "rn_clk_mgr_vbios_smu.h"
@@ -42,11 +40,9 @@
 #include "clk/clk_10_0_2_sh_mask.h"
 #include "renoir_ip_offset.h"
 
-#include "irq/dcn21/irq_service_dcn21.h"
 
 /* Constants */
 
-#define LPDDR_MEM_RETRAIN_LATENCY 4.977 /* Number obtained from LPDDR4 Training Counter Requirement doc */
 #define SMU_VER_55_51_0 0x373300 /* SMU Version that is able to set DISPCLK below 100MHz */
 
 /* Macros */
@@ -56,9 +52,7 @@
 
 
 /* TODO: evaluate how to lower or disable all dcn clocks in screen off case */
-int rn_get_active_display_cnt_wa(
-		struct dc *dc,
-		struct dc_state *context)
+static int rn_get_active_display_cnt_wa(struct dc *dc, struct dc_state *context)
 {
 	int i, display_count;
 	bool tmds_present = false;
@@ -77,7 +71,8 @@ int rn_get_active_display_cnt_wa(
 		const struct dc_link *link = dc->links[i];
 
 		/* abusing the fact that the dig and phy are coupled to see if the phy is enabled */
-		if (link->link_enc->funcs->is_dig_enabled(link->link_enc))
+		if (link->link_enc->funcs->is_dig_enabled &&
+		    link->link_enc->funcs->is_dig_enabled(link->link_enc))
 			display_count++;
 	}
 
@@ -88,13 +83,24 @@ int rn_get_active_display_cnt_wa(
 	return display_count;
 }
 
-void rn_set_low_power_state(struct clk_mgr *clk_mgr_base)
+static void rn_set_low_power_state(struct clk_mgr *clk_mgr_base)
 {
+	int display_count;
 	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+	struct dc *dc = clk_mgr_base->ctx->dc;
+	struct dc_state *context = dc->current_state;
+
+	if (clk_mgr_base->clks.pwr_state != DCN_PWR_STATE_LOW_POWER) {
 
-	rn_vbios_smu_set_dcn_low_power_state(clk_mgr, DCN_PWR_STATE_LOW_POWER);
-	/* update power state */
-	clk_mgr_base->clks.pwr_state = DCN_PWR_STATE_LOW_POWER;
+		display_count = rn_get_active_display_cnt_wa(dc, context);
+
+		/* if we can go lower, go lower */
+		if (display_count == 0) {
+			rn_vbios_smu_set_dcn_low_power_state(clk_mgr, DCN_PWR_STATE_LOW_POWER);
+			/* update power state */
+			clk_mgr_base->clks.pwr_state = DCN_PWR_STATE_LOW_POWER;
+		}
+	}
 }
 
 static void rn_update_clocks_update_dpp_dto(struct clk_mgr_internal *clk_mgr,
@@ -113,7 +119,7 @@ static void rn_update_clocks_update_dpp_dto(struct clk_mgr_internal *clk_mgr,
 		dpp_inst = clk_mgr->base.ctx->dc->res_pool->dpps[i]->inst;
 		dppclk_khz = context->res_ctx.pipe_ctx[i].plane_res.bw.dppclk_khz;
 
-		prev_dppclk_khz = clk_mgr->dccg->pipe_dppclk_khz[i];
+		prev_dppclk_khz = clk_mgr->dccg->pipe_dppclk_khz[dpp_inst];
 
 		if (safe_to_lower || prev_dppclk_khz < dppclk_khz)
 			clk_mgr->dccg->funcs->update_dpp_dto(
@@ -122,7 +128,7 @@ static void rn_update_clocks_update_dpp_dto(struct clk_mgr_internal *clk_mgr,
 }
 
 
-void rn_update_clocks(struct clk_mgr *clk_mgr_base,
+static void rn_update_clocks(struct clk_mgr *clk_mgr_base,
 			struct dc_state *context,
 			bool safe_to_lower)
 {
@@ -130,11 +136,9 @@ void rn_update_clocks(struct clk_mgr *clk_mgr_base,
 	struct dc_clocks *new_clocks = &context->bw_ctx.bw.dcn.clk;
 	struct dc *dc = clk_mgr_base->ctx->dc;
 	int display_count;
-	int irq_src;
 	bool update_dppclk = false;
 	bool update_dispclk = false;
 	bool dpp_clock_lowered = false;
-	uint32_t hpd_state;
 
 	struct dmcu *dmcu = clk_mgr_base->ctx->dc->res_pool->dmcu;
 
@@ -151,14 +155,8 @@ void rn_update_clocks(struct clk_mgr *clk_mgr_base,
 
 			display_count = rn_get_active_display_cnt_wa(dc, context);
 
-			for (irq_src = DC_IRQ_SOURCE_HPD1; irq_src <= DC_IRQ_SOURCE_HPD5; irq_src++) {
-				hpd_state = dc_get_hpd_state_dcn21(dc->res_pool->irqs, irq_src);
-				if (hpd_state)
-					break;
-			}
-
 			/* if we can go lower, go lower */
-			if (display_count == 0 && !hpd_state) {
+			if (display_count == 0) {
 				rn_vbios_smu_set_dcn_low_power_state(clk_mgr, DCN_PWR_STATE_LOW_POWER);
 				/* update power state */
 				clk_mgr_base->clks.pwr_state = DCN_PWR_STATE_LOW_POWER;
@@ -322,16 +320,16 @@ static void rn_dump_clk_registers(struct clk_state_registers_and_bypass *regs_an
 	regs_and_bypass->dppclk = internal.CLK1_CLK1_CURRENT_CNT / 10;
 
 	regs_and_bypass->dppclk_bypass = internal.CLK1_CLK1_BYPASS_CNTL & 0x0007;
-	if (regs_and_bypass->dppclk_bypass < 0 || regs_and_bypass->dppclk_bypass > 4)
+	if (regs_and_bypass->dppclk_bypass > 4)
 		regs_and_bypass->dppclk_bypass = 0;
 	regs_and_bypass->dcfclk_bypass = internal.CLK1_CLK3_BYPASS_CNTL & 0x0007;
-	if (regs_and_bypass->dcfclk_bypass < 0 || regs_and_bypass->dcfclk_bypass > 4)
+	if (regs_and_bypass->dcfclk_bypass > 4)
 		regs_and_bypass->dcfclk_bypass = 0;
 	regs_and_bypass->dispclk_bypass = internal.CLK1_CLK0_BYPASS_CNTL & 0x0007;
-	if (regs_and_bypass->dispclk_bypass < 0 || regs_and_bypass->dispclk_bypass > 4)
+	if (regs_and_bypass->dispclk_bypass > 4)
 		regs_and_bypass->dispclk_bypass = 0;
 	regs_and_bypass->dprefclk_bypass = internal.CLK1_CLK2_BYPASS_CNTL & 0x0007;
-	if (regs_and_bypass->dprefclk_bypass < 0 || regs_and_bypass->dprefclk_bypass > 4)
+	if (regs_and_bypass->dprefclk_bypass > 4)
 		regs_and_bypass->dprefclk_bypass = 0;
 
 	if (log_info->enabled) {
@@ -437,25 +435,14 @@ static void rn_dump_clk_registers(struct clk_state_registers_and_bypass *regs_an
 	}
 }
 
-/* This function produce translated logical clk state values*/
-void rn_get_clk_states(struct clk_mgr *clk_mgr_base, struct clk_states *s)
-{
-	struct clk_state_registers_and_bypass sb = { 0 };
-	struct clk_log_info log_info = { 0 };
-
-	rn_dump_clk_registers(&sb, clk_mgr_base, &log_info);
-
-	s->dprefclk_khz = sb.dprefclk * 1000;
-}
-
-void rn_enable_pme_wa(struct clk_mgr *clk_mgr_base)
+static void rn_enable_pme_wa(struct clk_mgr *clk_mgr_base)
 {
 	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
 
 	rn_vbios_smu_enable_pme_wa(clk_mgr);
 }
 
-void rn_init_clocks(struct clk_mgr *clk_mgr)
+static void rn_init_clocks(struct clk_mgr *clk_mgr)
 {
 	memset(&(clk_mgr->clks), 0, sizeof(struct dc_clocks));
 	// Assumption is that boot state always supports pstate
@@ -497,7 +484,8 @@ static void build_watermark_ranges(struct clk_bw_params *bw_params, struct pp_sm
 			ranges->reader_wm_sets[num_valid_sets].max_fill_clk_mhz = PP_SMU_WM_SET_RANGE_CLK_UNCONSTRAINED_MAX;
 
 			/* Modify previous watermark range to cover up to max */
-			ranges->reader_wm_sets[num_valid_sets - 1].max_fill_clk_mhz = PP_SMU_WM_SET_RANGE_CLK_UNCONSTRAINED_MAX;
+			if (num_valid_sets > 0)
+				ranges->reader_wm_sets[num_valid_sets - 1].max_fill_clk_mhz = PP_SMU_WM_SET_RANGE_CLK_UNCONSTRAINED_MAX;
 		}
 		num_valid_sets++;
 	}
@@ -561,7 +549,7 @@ static void rn_notify_link_rate_change(struct clk_mgr *clk_mgr_base, struct dc_l
 
 	clk_mgr->cur_phyclk_req_table[link->link_index] = link->cur_link_settings.link_rate * LINK_RATE_REF_FREQ_IN_KHZ;
 
-	for (i = 0; i < MAX_PIPES * 2; i++) {
+	for (i = 0; i < MAX_LINKS; i++) {
 		if (clk_mgr->cur_phyclk_req_table[i] > max_phyclk_req)
 			max_phyclk_req = clk_mgr->cur_phyclk_req_table[i];
 	}
@@ -623,228 +611,6 @@ static struct clk_bw_params rn_bw_params = {
 
 };
 
-static struct wm_table ddr4_wm_table_gs = {
-	.entries = {
-		{
-			.wm_inst = WM_A,
-			.wm_type = WM_TYPE_PSTATE_CHG,
-			.pstate_latency_us = 11.72,
-			.sr_exit_time_us = 7.09,
-			.sr_enter_plus_exit_time_us = 8.14,
-			.valid = true,
-		},
-		{
-			.wm_inst = WM_B,
-			.wm_type = WM_TYPE_PSTATE_CHG,
-			.pstate_latency_us = 11.72,
-			.sr_exit_time_us = 10.12,
-			.sr_enter_plus_exit_time_us = 11.48,
-			.valid = true,
-		},
-		{
-			.wm_inst = WM_C,
-			.wm_type = WM_TYPE_PSTATE_CHG,
-			.pstate_latency_us = 11.72,
-			.sr_exit_time_us = 10.12,
-			.sr_enter_plus_exit_time_us = 11.48,
-			.valid = true,
-		},
-		{
-			.wm_inst = WM_D,
-			.wm_type = WM_TYPE_PSTATE_CHG,
-			.pstate_latency_us = 11.72,
-			.sr_exit_time_us = 10.12,
-			.sr_enter_plus_exit_time_us = 11.48,
-			.valid = true,
-		},
-	}
-};
-
-static struct wm_table lpddr4_wm_table_gs = {
-	.entries = {
-		{
-			.wm_inst = WM_A,
-			.wm_type = WM_TYPE_PSTATE_CHG,
-			.pstate_latency_us = 11.65333,
-			.sr_exit_time_us = 5.32,
-			.sr_enter_plus_exit_time_us = 6.38,
-			.valid = true,
-		},
-		{
-			.wm_inst = WM_B,
-			.wm_type = WM_TYPE_PSTATE_CHG,
-			.pstate_latency_us = 11.65333,
-			.sr_exit_time_us = 9.82,
-			.sr_enter_plus_exit_time_us = 11.196,
-			.valid = true,
-		},
-		{
-			.wm_inst = WM_C,
-			.wm_type = WM_TYPE_PSTATE_CHG,
-			.pstate_latency_us = 11.65333,
-			.sr_exit_time_us = 9.89,
-			.sr_enter_plus_exit_time_us = 11.24,
-			.valid = true,
-		},
-		{
-			.wm_inst = WM_D,
-			.wm_type = WM_TYPE_PSTATE_CHG,
-			.pstate_latency_us = 11.65333,
-			.sr_exit_time_us = 9.748,
-			.sr_enter_plus_exit_time_us = 11.102,
-			.valid = true,
-		},
-	}
-};
-
-static struct wm_table lpddr4_wm_table_with_disabled_ppt = {
-	.entries = {
-		{
-			.wm_inst = WM_A,
-			.wm_type = WM_TYPE_PSTATE_CHG,
-			.pstate_latency_us = 11.65333,
-			.sr_exit_time_us = 8.32,
-			.sr_enter_plus_exit_time_us = 9.38,
-			.valid = true,
-		},
-		{
-			.wm_inst = WM_B,
-			.wm_type = WM_TYPE_PSTATE_CHG,
-			.pstate_latency_us = 11.65333,
-			.sr_exit_time_us = 9.82,
-			.sr_enter_plus_exit_time_us = 11.196,
-			.valid = true,
-		},
-		{
-			.wm_inst = WM_C,
-			.wm_type = WM_TYPE_PSTATE_CHG,
-			.pstate_latency_us = 11.65333,
-			.sr_exit_time_us = 9.89,
-			.sr_enter_plus_exit_time_us = 11.24,
-			.valid = true,
-		},
-		{
-			.wm_inst = WM_D,
-			.wm_type = WM_TYPE_PSTATE_CHG,
-			.pstate_latency_us = 11.65333,
-			.sr_exit_time_us = 9.748,
-			.sr_enter_plus_exit_time_us = 11.102,
-			.valid = true,
-		},
-	}
-};
-
-static struct wm_table ddr4_wm_table_rn = {
-	.entries = {
-		{
-			.wm_inst = WM_A,
-			.wm_type = WM_TYPE_PSTATE_CHG,
-			.pstate_latency_us = 11.72,
-			.sr_exit_time_us = 11.90,
-			.sr_enter_plus_exit_time_us = 12.80,
-			.valid = true,
-		},
-		{
-			.wm_inst = WM_B,
-			.wm_type = WM_TYPE_PSTATE_CHG,
-			.pstate_latency_us = 11.72,
-			.sr_exit_time_us = 13.18,
-			.sr_enter_plus_exit_time_us = 14.30,
-			.valid = true,
-		},
-		{
-			.wm_inst = WM_C,
-			.wm_type = WM_TYPE_PSTATE_CHG,
-			.pstate_latency_us = 11.72,
-			.sr_exit_time_us = 13.18,
-			.sr_enter_plus_exit_time_us = 14.30,
-			.valid = true,
-		},
-		{
-			.wm_inst = WM_D,
-			.wm_type = WM_TYPE_PSTATE_CHG,
-			.pstate_latency_us = 11.72,
-			.sr_exit_time_us = 13.18,
-			.sr_enter_plus_exit_time_us = 14.30,
-			.valid = true,
-		},
-	}
-};
-
-static struct wm_table ddr4_1R_wm_table_rn = {
-	.entries = {
-		{
-			.wm_inst = WM_A,
-			.wm_type = WM_TYPE_PSTATE_CHG,
-			.pstate_latency_us = 11.72,
-			.sr_exit_time_us = 13.90,
-			.sr_enter_plus_exit_time_us = 14.80,
-			.valid = true,
-		},
-		{
-			.wm_inst = WM_B,
-			.wm_type = WM_TYPE_PSTATE_CHG,
-			.pstate_latency_us = 11.72,
-			.sr_exit_time_us = 13.90,
-			.sr_enter_plus_exit_time_us = 14.80,
-			.valid = true,
-		},
-		{
-			.wm_inst = WM_C,
-			.wm_type = WM_TYPE_PSTATE_CHG,
-			.pstate_latency_us = 11.72,
-			.sr_exit_time_us = 13.90,
-			.sr_enter_plus_exit_time_us = 14.80,
-			.valid = true,
-		},
-		{
-			.wm_inst = WM_D,
-			.wm_type = WM_TYPE_PSTATE_CHG,
-			.pstate_latency_us = 11.72,
-			.sr_exit_time_us = 13.90,
-			.sr_enter_plus_exit_time_us = 14.80,
-			.valid = true,
-		},
-	}
-};
-
-static struct wm_table lpddr4_wm_table_rn = {
-	.entries = {
-		{
-			.wm_inst = WM_A,
-			.wm_type = WM_TYPE_PSTATE_CHG,
-			.pstate_latency_us = 11.65333,
-			.sr_exit_time_us = 7.32,
-			.sr_enter_plus_exit_time_us = 8.38,
-			.valid = true,
-		},
-		{
-			.wm_inst = WM_B,
-			.wm_type = WM_TYPE_PSTATE_CHG,
-			.pstate_latency_us = 11.65333,
-			.sr_exit_time_us = 9.82,
-			.sr_enter_plus_exit_time_us = 11.196,
-			.valid = true,
-		},
-		{
-			.wm_inst = WM_C,
-			.wm_type = WM_TYPE_PSTATE_CHG,
-			.pstate_latency_us = 11.65333,
-			.sr_exit_time_us = 9.89,
-			.sr_enter_plus_exit_time_us = 11.24,
-			.valid = true,
-		},
-		{
-			.wm_inst = WM_D,
-			.wm_type = WM_TYPE_PSTATE_CHG,
-			.pstate_latency_us = 11.65333,
-			.sr_exit_time_us = 9.748,
-			.sr_enter_plus_exit_time_us = 11.102,
-			.valid = true,
-		},
-	}
-};
-
 static unsigned int find_socclk_for_voltage(struct dpm_clocks *clock_table, unsigned int voltage)
 {
 	int i;
@@ -877,7 +643,8 @@ static void rn_clk_mgr_helper_populate_bw_params(struct clk_bw_params *bw_params
 
 	j = -1;
 
-	ASSERT(PP_SMU_NUM_FCLK_DPM_LEVELS <= MAX_NUM_DPM_LVL);
+	static_assert(PP_SMU_NUM_FCLK_DPM_LEVELS <= MAX_NUM_DPM_LVL,
+		"number of reported FCLK DPM levels exceed maximum");
 
 	/* Find lowest DPM, FCLK is filled in reverse order*/
 
@@ -924,12 +691,10 @@ static void rn_clk_mgr_helper_populate_bw_params(struct clk_bw_params *bw_params
 		/*
 		 * WM set D will be re-purposed for memory retraining
 		 */
-		bw_params->wm_table.entries[WM_D].pstate_latency_us = LPDDR_MEM_RETRAIN_LATENCY;
-		bw_params->wm_table.entries[WM_D].wm_inst = WM_D;
-		bw_params->wm_table.entries[WM_D].wm_type = WM_TYPE_RETRAINING;
-		bw_params->wm_table.entries[WM_D].valid = true;
+		DC_FP_START();
+		dcn21_clk_mgr_set_bw_params_wm_table(bw_params);
+		DC_FP_END();
 	}
-
 }
 
 void rn_clk_mgr_construct(
@@ -942,10 +707,9 @@ void rn_clk_mgr_construct(
 	struct dpm_clocks clock_table = { 0 };
 	enum pp_smu_status status = 0;
 	int is_green_sardine = 0;
+	struct clk_log_info log_info = {0};
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	is_green_sardine = ASICREV_IS_GREEN_SARDINE(ctx->asic_id.hw_internal_rev);
-#endif
 
 	clk_mgr->base.ctx = ctx;
 	clk_mgr->base.funcs = &dcn21_funcs;
@@ -962,48 +726,41 @@ void rn_clk_mgr_construct(
 
 	clk_mgr->smu_ver = rn_vbios_smu_get_smu_version(clk_mgr);
 
-	if (IS_FPGA_MAXIMUS_DC(ctx->dce_environment)) {
-		dcn21_funcs.update_clocks = dcn2_update_clocks_fpga;
+	clk_mgr->periodic_retraining_disabled = rn_vbios_smu_is_periodic_retraining_disabled(clk_mgr);
+
+	/* SMU Version 55.51.0 and up no longer have an issue
+	 * that needs to limit minimum dispclk */
+	if (clk_mgr->smu_ver >= SMU_VER_55_51_0)
+		debug->min_disp_clk_khz = 0;
+
+	/* TODO: Check we get what we expect during bringup */
+	clk_mgr->base.dentist_vco_freq_khz = get_vco_frequency_from_reg(clk_mgr);
+
+	/* in case we don't get a value from the register, use default */
+	if (clk_mgr->base.dentist_vco_freq_khz == 0)
 		clk_mgr->base.dentist_vco_freq_khz = 3600000;
-	} else {
-		struct clk_log_info log_info = {0};
-
-		clk_mgr->periodic_retraining_disabled = rn_vbios_smu_is_periodic_retraining_disabled(clk_mgr);
-
-		/* SMU Version 55.51.0 and up no longer have an issue
-		 * that needs to limit minimum dispclk */
-		if (clk_mgr->smu_ver >= SMU_VER_55_51_0)
-			debug->min_disp_clk_khz = 0;
-
-		/* TODO: Check we get what we expect during bringup */
-		clk_mgr->base.dentist_vco_freq_khz = get_vco_frequency_from_reg(clk_mgr);
-
-		/* in case we don't get a value from the register, use default */
-		if (clk_mgr->base.dentist_vco_freq_khz == 0)
-			clk_mgr->base.dentist_vco_freq_khz = 3600000;
-
-		if (ctx->dc_bios->integrated_info->memory_type == LpDdr4MemType) {
-			if (clk_mgr->periodic_retraining_disabled) {
-				rn_bw_params.wm_table = lpddr4_wm_table_with_disabled_ppt;
-			} else {
-				if (is_green_sardine)
-					rn_bw_params.wm_table = lpddr4_wm_table_gs;
-				else
-					rn_bw_params.wm_table = lpddr4_wm_table_rn;
-			}
+
+	if (ctx->dc_bios->integrated_info->memory_type == LpDdr4MemType) {
+		if (clk_mgr->periodic_retraining_disabled) {
+			rn_bw_params.wm_table = lpddr4_wm_table_with_disabled_ppt;
 		} else {
 			if (is_green_sardine)
-				rn_bw_params.wm_table = ddr4_wm_table_gs;
-			else {
-				if (ctx->dc->config.is_single_rank_dimm)
-					rn_bw_params.wm_table = ddr4_1R_wm_table_rn;
-				else
-					rn_bw_params.wm_table = ddr4_wm_table_rn;
-			}
+				rn_bw_params.wm_table = lpddr4_wm_table_gs;
+			else
+				rn_bw_params.wm_table = lpddr4_wm_table_rn;
+		}
+	} else {
+		if (is_green_sardine)
+			rn_bw_params.wm_table = ddr4_wm_table_gs;
+		else {
+			if (ctx->dc->config.is_single_rank_dimm)
+				rn_bw_params.wm_table = ddr4_1R_wm_table_rn;
+			else
+				rn_bw_params.wm_table = ddr4_wm_table_rn;
 		}
-		/* Saved clocks configured at boot for debug purposes */
-		rn_dump_clk_registers(&clk_mgr->base.boot_snapshot, &clk_mgr->base, &log_info);
 	}
+	/* Saved clocks configured at boot for debug purposes */
+	rn_dump_clk_registers(&clk_mgr->base.boot_snapshot, &clk_mgr->base, &log_info);
 
 	clk_mgr->base.dprefclk_khz = 600000;
 	dce_clock_read_ss_info(clk_mgr);
@@ -1015,7 +772,7 @@ void rn_clk_mgr_construct(
 		status = pp_smu->rn_funcs.get_dpm_clock_table(&pp_smu->rn_funcs.pp_smu, &clock_table);
 
 		if (status == PP_SMU_RESULT_OK &&
-		    ctx->dc_bios && ctx->dc_bios->integrated_info) {
+		    ctx->dc_bios->integrated_info) {
 			rn_clk_mgr_helper_populate_bw_params (clk_mgr->base.bw_params, &clock_table, ctx->dc_bios->integrated_info);
 			/* treat memory config as single channel if memory is asymmetrics. */
 			if (ctx->dc->config.is_asymmetric_memory)
@@ -1023,9 +780,8 @@ void rn_clk_mgr_construct(
 		}
 	}
 
-	if (!IS_FPGA_MAXIMUS_DC(ctx->dce_environment) && clk_mgr->smu_ver >= 0x00371500) {
-		/* enable powerfeatures when displaycount goes to 0 */
+	/* enable powerfeatures when displaycount goes to 0 */
+	if (clk_mgr->smu_ver >= 0x00371500)
 		rn_vbios_smu_enable_48mhz_tmdp_refclk_pwrdwn(clk_mgr, !debug->disable_48mhz_pwrdwn);
-	}
 }
 
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn21/rn_clk_mgr.h b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn21/rn_clk_mgr.h
index e4322fa5475b..f1319957e400 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn21/rn_clk_mgr.h
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn21/rn_clk_mgr.h
@@ -28,6 +28,14 @@
 
 #include "clk_mgr.h"
 #include "dm_pp_smu.h"
+#include "clk_mgr_internal.h"
+
+extern struct wm_table ddr4_wm_table_gs;
+extern struct wm_table lpddr4_wm_table_gs;
+extern struct wm_table lpddr4_wm_table_with_disabled_ppt;
+extern struct wm_table ddr4_wm_table_rn;
+extern struct wm_table ddr4_1R_wm_table_rn;
+extern struct wm_table lpddr4_wm_table_rn;
 
 struct rn_clk_registers {
 	uint32_t CLK1_CLK0_CURRENT_CNT; /* DPREFCLK */
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn21/rn_clk_mgr_vbios_smu.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn21/rn_clk_mgr_vbios_smu.c
index 9f7eed6688c4..5a633333dbb5 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn21/rn_clk_mgr_vbios_smu.c
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn21/rn_clk_mgr_vbios_smu.c
@@ -26,6 +26,10 @@
 #include "core_types.h"
 #include "clk_mgr_internal.h"
 #include "reg_helper.h"
+#include "dm_helpers.h"
+
+#include "rn_clk_mgr_vbios_smu.h"
+
 #include <linux/delay.h>
 
 #include "renoir_ip_offset.h"
@@ -39,6 +43,12 @@
 #define FN(reg_name, field) \
 	FD(reg_name##__##field)
 
+#include "logger_types.h"
+#undef DC_LOGGER
+#define DC_LOGGER \
+	CTX->logger
+#define smu_print(str, ...) {DC_LOG_SMU(str, ##__VA_ARGS__); }
+
 #define VBIOSSMC_MSG_TestMessage                  0x1
 #define VBIOSSMC_MSG_GetSmuVersion                0x2
 #define VBIOSSMC_MSG_PowerUpGfx                   0x3
@@ -86,10 +96,21 @@ static uint32_t rn_smu_wait_for_response(struct clk_mgr_internal *clk_mgr, unsig
 }
 
 
-int rn_vbios_smu_send_msg_with_param(struct clk_mgr_internal *clk_mgr, unsigned int msg_id, unsigned int param)
+static int rn_vbios_smu_send_msg_with_param(struct clk_mgr_internal *clk_mgr,
+					    unsigned int msg_id,
+					    unsigned int param)
 {
 	uint32_t result;
 
+	result = rn_smu_wait_for_response(clk_mgr, 10, 200000);
+
+	if (result != VBIOSSMC_Result_OK)
+		smu_print("SMU Response was not OK. SMU response after wait received is: %d\n", result);
+
+	if (result == VBIOSSMC_Status_BUSY) {
+		return -1;
+	}
+
 	/* First clear response register */
 	REG_WRITE(MP1_SMN_C2PMSG_91, VBIOSSMC_Status_BUSY);
 
@@ -101,7 +122,10 @@ int rn_vbios_smu_send_msg_with_param(struct clk_mgr_internal *clk_mgr, unsigned
 
 	result = rn_smu_wait_for_response(clk_mgr, 10, 200000);
 
-	ASSERT(result == VBIOSSMC_Result_OK || result == VBIOSSMC_Result_UnknownCmd);
+	if (IS_SMU_TIMEOUT(result)) {
+		ASSERT(0);
+		dm_helpers_smu_timeout(CTX, msg_id, param, 10 * 200000);
+	}
 
 	/* Actual dispclk set is returned in the parameter register */
 	return REG_READ(MP1_SMN_C2PMSG_83);
@@ -128,35 +152,18 @@ int rn_vbios_smu_set_dispclk(struct clk_mgr_internal *clk_mgr, int requested_dis
 			VBIOSSMC_MSG_SetDispclkFreq,
 			khz_to_mhz_ceil(requested_dispclk_khz));
 
-	if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) {
-		if (dmcu && dmcu->funcs->is_dmcu_initialized(dmcu)) {
-			if (clk_mgr->dfs_bypass_disp_clk != actual_dispclk_set_mhz)
-				dmcu->funcs->set_psr_wait_loop(dmcu,
-						actual_dispclk_set_mhz / 7);
-		}
+	if (dmcu && dmcu->funcs->is_dmcu_initialized(dmcu)) {
+		if (clk_mgr->dfs_bypass_disp_clk != actual_dispclk_set_mhz)
+			dmcu->funcs->set_psr_wait_loop(dmcu,
+					actual_dispclk_set_mhz / 7);
 	}
 
 	// pmfw always set clock more than or equal requested clock
-	if (!IS_DIAG_DC(dc->ctx->dce_environment))
-		ASSERT(actual_dispclk_set_mhz >= khz_to_mhz_ceil(requested_dispclk_khz));
+	ASSERT(actual_dispclk_set_mhz >= khz_to_mhz_ceil(requested_dispclk_khz));
 
 	return actual_dispclk_set_mhz * 1000;
 }
 
-int rn_vbios_smu_set_dprefclk(struct clk_mgr_internal *clk_mgr)
-{
-	int actual_dprefclk_set_mhz = -1;
-
-	actual_dprefclk_set_mhz = rn_vbios_smu_send_msg_with_param(
-			clk_mgr,
-			VBIOSSMC_MSG_SetDprefclkFreq,
-			khz_to_mhz_ceil(clk_mgr->base.dprefclk_khz));
-
-	/* TODO: add code for programing DP DTO, currently this is down by command table */
-
-	return actual_dprefclk_set_mhz * 1000;
-}
-
 int rn_vbios_smu_set_hard_min_dcfclk(struct clk_mgr_internal *clk_mgr, int requested_dcfclk_khz)
 {
 	int actual_dcfclk_set_mhz = -1;
@@ -198,15 +205,13 @@ void rn_vbios_smu_set_phyclk(struct clk_mgr_internal *clk_mgr, int requested_phy
 int rn_vbios_smu_set_dppclk(struct clk_mgr_internal *clk_mgr, int requested_dpp_khz)
 {
 	int actual_dppclk_set_mhz = -1;
-	struct dc *dc = clk_mgr->base.ctx->dc;
 
 	actual_dppclk_set_mhz = rn_vbios_smu_send_msg_with_param(
 			clk_mgr,
 			VBIOSSMC_MSG_SetDppclkFreq,
 			khz_to_mhz_ceil(requested_dpp_khz));
 
-	if (!IS_DIAG_DC(dc->ctx->dce_environment))
-		ASSERT(actual_dppclk_set_mhz >= khz_to_mhz_ceil(requested_dpp_khz));
+	ASSERT(actual_dppclk_set_mhz >= khz_to_mhz_ceil(requested_dpp_khz));
 
 	return actual_dppclk_set_mhz * 1000;
 }
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn21/rn_clk_mgr_vbios_smu.h b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn21/rn_clk_mgr_vbios_smu.h
index 3e5df27aa96f..f76fad87f0e1 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn21/rn_clk_mgr_vbios_smu.h
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn21/rn_clk_mgr_vbios_smu.h
@@ -26,14 +26,15 @@
 #ifndef DAL_DC_RN_CLK_MGR_VBIOS_SMU_H_
 #define DAL_DC_RN_CLK_MGR_VBIOS_SMU_H_
 
+enum dcn_pwr_state;
+
 int rn_vbios_smu_get_smu_version(struct clk_mgr_internal *clk_mgr);
 int rn_vbios_smu_set_dispclk(struct clk_mgr_internal *clk_mgr, int requested_dispclk_khz);
-int rn_vbios_smu_set_dprefclk(struct clk_mgr_internal *clk_mgr);
 int rn_vbios_smu_set_hard_min_dcfclk(struct clk_mgr_internal *clk_mgr, int requested_dcfclk_khz);
 int rn_vbios_smu_set_min_deep_sleep_dcfclk(struct clk_mgr_internal *clk_mgr, int requested_min_ds_dcfclk_khz);
 void rn_vbios_smu_set_phyclk(struct clk_mgr_internal *clk_mgr, int requested_phyclk_khz);
 int rn_vbios_smu_set_dppclk(struct clk_mgr_internal *clk_mgr, int requested_dpp_khz);
-void rn_vbios_smu_set_dcn_low_power_state(struct clk_mgr_internal *clk_mgr, int display_count);
+void rn_vbios_smu_set_dcn_low_power_state(struct clk_mgr_internal *clk_mgr, enum dcn_pwr_state);
 void rn_vbios_smu_enable_48mhz_tmdp_refclk_pwrdwn(struct clk_mgr_internal *clk_mgr, bool enable);
 void rn_vbios_smu_enable_pme_wa(struct clk_mgr_internal *clk_mgr);
 int rn_vbios_smu_is_periodic_retraining_disabled(struct clk_mgr_internal *clk_mgr);
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dalsmc.h b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dalsmc.h
index fa09c594fd36..06da34676965 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dalsmc.h
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dalsmc.h
@@ -56,6 +56,7 @@
 #define DALSMC_MSG_SetDisplayRefreshFromMall      0xF
 #define DALSMC_MSG_SetExternalClientDfCstateAllow 0x10
 #define DALSMC_MSG_BacoAudioD3PME                 0x11
-#define DALSMC_Message_Count                      0x12
+#define DALSMC_MSG_SmartAccess                    0x12
+#define DALSMC_Message_Count                      0x13
 
 #endif
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dcn30_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dcn30_clk_mgr.c
index 1861a147a7fa..ef77fcd164ed 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dcn30_clk_mgr.c
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dcn30_clk_mgr.c
@@ -25,30 +25,25 @@
 
 #include "dccg.h"
 #include "clk_mgr_internal.h"
-
 #include "dcn30_clk_mgr_smu_msg.h"
 #include "dcn20/dcn20_clk_mgr.h"
 #include "dce100/dce_clk_mgr.h"
+#include "dcn30/dcn30_clk_mgr.h"
+#include "dml/dcn30/dcn30_fpu.h"
+#include "dcn30/dcn30m_clk_mgr.h"
 #include "reg_helper.h"
 #include "core_types.h"
 #include "dm_helpers.h"
-
 #include "atomfirmware.h"
-
-
 #include "sienna_cichlid_ip_offset.h"
 #include "dcn/dcn_3_0_0_offset.h"
 #include "dcn/dcn_3_0_0_sh_mask.h"
-
 #include "nbio/nbio_7_4_offset.h"
-
-#include "dcn/dpcs_3_0_0_offset.h"
-#include "dcn/dpcs_3_0_0_sh_mask.h"
-
+#include "dpcs/dpcs_3_0_0_offset.h"
+#include "dpcs/dpcs_3_0_0_sh_mask.h"
 #include "mmhub/mmhub_2_0_0_offset.h"
 #include "mmhub/mmhub_2_0_0_sh_mask.h"
-/*we don't have clk folder yet*/
-#include "dcn30/dcn30_clk_mgr.h"
+#include "dcn30_smu11_driver_if.h"
 
 #undef FN
 #define FN(reg_name, field_name) \
@@ -83,7 +78,7 @@ static const struct clk_mgr_mask clk_mgr_mask = {
 
 
 /* Query SMU for all clock states for a particular clock */
-static void dcn3_init_single_clock(struct clk_mgr_internal *clk_mgr, PPCLK_e clk, unsigned int *entry_0, unsigned int *num_levels)
+static void dcn3_init_single_clock(struct clk_mgr_internal *clk_mgr, uint32_t clk, unsigned int *entry_0, unsigned int *num_levels)
 {
 	unsigned int i;
 	char *entry_i = (char *)entry_0;
@@ -104,57 +99,11 @@ static void dcn3_init_single_clock(struct clk_mgr_internal *clk_mgr, PPCLK_e clk
 	}
 }
 
-static noinline void dcn3_build_wm_range_table(struct clk_mgr_internal *clk_mgr)
+static void dcn3_build_wm_range_table(struct clk_mgr_internal *clk_mgr)
 {
-	/* defaults */
-	double pstate_latency_us = clk_mgr->base.ctx->dc->dml.soc.dram_clock_change_latency_us;
-	double sr_exit_time_us = clk_mgr->base.ctx->dc->dml.soc.sr_exit_time_us;
-	double sr_enter_plus_exit_time_us = clk_mgr->base.ctx->dc->dml.soc.sr_enter_plus_exit_time_us;
-	uint16_t min_uclk_mhz = clk_mgr->base.bw_params->clk_table.entries[0].memclk_mhz;
-
-	/* Set A - Normal - default values*/
-	clk_mgr->base.bw_params->wm_table.nv_entries[WM_A].valid = true;
-	clk_mgr->base.bw_params->wm_table.nv_entries[WM_A].dml_input.pstate_latency_us = pstate_latency_us;
-	clk_mgr->base.bw_params->wm_table.nv_entries[WM_A].dml_input.sr_exit_time_us = sr_exit_time_us;
-	clk_mgr->base.bw_params->wm_table.nv_entries[WM_A].dml_input.sr_enter_plus_exit_time_us = sr_enter_plus_exit_time_us;
-	clk_mgr->base.bw_params->wm_table.nv_entries[WM_A].pmfw_breakdown.wm_type = WATERMARKS_CLOCK_RANGE;
-	clk_mgr->base.bw_params->wm_table.nv_entries[WM_A].pmfw_breakdown.min_dcfclk = 0;
-	clk_mgr->base.bw_params->wm_table.nv_entries[WM_A].pmfw_breakdown.max_dcfclk = 0xFFFF;
-	clk_mgr->base.bw_params->wm_table.nv_entries[WM_A].pmfw_breakdown.min_uclk = min_uclk_mhz;
-	clk_mgr->base.bw_params->wm_table.nv_entries[WM_A].pmfw_breakdown.max_uclk = 0xFFFF;
-
-	/* Set B - Performance - higher minimum clocks */
-//	clk_mgr->base.bw_params->wm_table.nv_entries[WM_B].valid = true;
-//	clk_mgr->base.bw_params->wm_table.nv_entries[WM_B].dml_input.pstate_latency_us = pstate_latency_us;
-//	clk_mgr->base.bw_params->wm_table.nv_entries[WM_B].dml_input.sr_exit_time_us = sr_exit_time_us;
-//	clk_mgr->base.bw_params->wm_table.nv_entries[WM_B].dml_input.sr_enter_plus_exit_time_us = sr_enter_plus_exit_time_us;
-//	clk_mgr->base.bw_params->wm_table.nv_entries[WM_B].pmfw_breakdown.wm_type = WATERMARKS_CLOCK_RANGE;
-//	clk_mgr->base.bw_params->wm_table.nv_entries[WM_B].pmfw_breakdown.min_dcfclk = TUNED VALUE;
-//	clk_mgr->base.bw_params->wm_table.nv_entries[WM_B].pmfw_breakdown.max_dcfclk = 0xFFFF;
-//	clk_mgr->base.bw_params->wm_table.nv_entries[WM_B].pmfw_breakdown.min_uclk = TUNED VALUE;
-//	clk_mgr->base.bw_params->wm_table.nv_entries[WM_B].pmfw_breakdown.max_uclk = 0xFFFF;
-
-	/* Set C - Dummy P-State - P-State latency set to "dummy p-state" value */
-	clk_mgr->base.bw_params->wm_table.nv_entries[WM_C].valid = true;
-	clk_mgr->base.bw_params->wm_table.nv_entries[WM_C].dml_input.pstate_latency_us = clk_mgr->base.ctx->dc->dml.soc.dummy_pstate_latency_us;
-	clk_mgr->base.bw_params->wm_table.nv_entries[WM_C].dml_input.sr_exit_time_us = sr_exit_time_us;
-	clk_mgr->base.bw_params->wm_table.nv_entries[WM_C].dml_input.sr_enter_plus_exit_time_us = sr_enter_plus_exit_time_us;
-	clk_mgr->base.bw_params->wm_table.nv_entries[WM_C].pmfw_breakdown.wm_type = WATERMARKS_DUMMY_PSTATE;
-	clk_mgr->base.bw_params->wm_table.nv_entries[WM_C].pmfw_breakdown.min_dcfclk = 0;
-	clk_mgr->base.bw_params->wm_table.nv_entries[WM_C].pmfw_breakdown.max_dcfclk = 0xFFFF;
-	clk_mgr->base.bw_params->wm_table.nv_entries[WM_C].pmfw_breakdown.min_uclk = min_uclk_mhz;
-	clk_mgr->base.bw_params->wm_table.nv_entries[WM_C].pmfw_breakdown.max_uclk = 0xFFFF;
-
-	/* Set D - MALL - SR enter and exit times adjusted for MALL */
-	clk_mgr->base.bw_params->wm_table.nv_entries[WM_D].valid = true;
-	clk_mgr->base.bw_params->wm_table.nv_entries[WM_D].dml_input.pstate_latency_us = pstate_latency_us;
-	clk_mgr->base.bw_params->wm_table.nv_entries[WM_D].dml_input.sr_exit_time_us = 2;
-	clk_mgr->base.bw_params->wm_table.nv_entries[WM_D].dml_input.sr_enter_plus_exit_time_us = 4;
-	clk_mgr->base.bw_params->wm_table.nv_entries[WM_D].pmfw_breakdown.wm_type = WATERMARKS_MALL;
-	clk_mgr->base.bw_params->wm_table.nv_entries[WM_D].pmfw_breakdown.min_dcfclk = 0;
-	clk_mgr->base.bw_params->wm_table.nv_entries[WM_D].pmfw_breakdown.max_dcfclk = 0xFFFF;
-	clk_mgr->base.bw_params->wm_table.nv_entries[WM_D].pmfw_breakdown.min_uclk = min_uclk_mhz;
-	clk_mgr->base.bw_params->wm_table.nv_entries[WM_D].pmfw_breakdown.max_uclk = 0xFFFF;
+	DC_FP_START();
+	dcn3_fpu_build_wm_range_table(&clk_mgr->base);
+	DC_FP_END();
 }
 
 void dcn3_init_clocks(struct clk_mgr *clk_mgr_base)
@@ -184,6 +133,7 @@ void dcn3_init_clocks(struct clk_mgr *clk_mgr_base)
 	dcn3_init_single_clock(clk_mgr, PPCLK_DCEFCLK,
 			&clk_mgr_base->bw_params->clk_table.entries[0].dcfclk_mhz,
 			&num_levels);
+	dcn30_smu_set_min_deep_sleep_dcef_clk(clk_mgr, 0);
 
 	/* DTBCLK */
 	dcn3_init_single_clock(clk_mgr, PPCLK_DTBCLK,
@@ -252,11 +202,11 @@ static void dcn3_update_clocks(struct clk_mgr *clk_mgr_base,
 	bool update_dispclk = false;
 	bool enter_display_off = false;
 	bool dpp_clock_lowered = false;
+	bool update_pstate_unsupported_clk = false;
 	struct dmcu *dmcu = clk_mgr_base->ctx->dc->res_pool->dmcu;
 	bool force_reset = false;
 	bool update_uclk = false;
 	bool p_state_change_support;
-	int total_plane_count;
 
 	if (dc->work_arounds.skip_clock_update || !clk_mgr->smu_present)
 		return;
@@ -297,15 +247,29 @@ static void dcn3_update_clocks(struct clk_mgr *clk_mgr_base,
 		clk_mgr_base->clks.socclk_khz = new_clocks->socclk_khz;
 
 	clk_mgr_base->clks.prev_p_state_change_support = clk_mgr_base->clks.p_state_change_support;
-	total_plane_count = clk_mgr_helper_get_active_plane_cnt(dc, context);
-	p_state_change_support = new_clocks->p_state_change_support || (total_plane_count == 0);
-	if (should_update_pstate_support(safe_to_lower, p_state_change_support, clk_mgr_base->clks.p_state_change_support)) {
+	p_state_change_support = new_clocks->p_state_change_support;
+
+	// invalidate the current P-State forced min in certain dc_mode_softmax situations
+	if (dc->clk_mgr->dc_mode_softmax_enabled && safe_to_lower && !p_state_change_support) {
+		if ((new_clocks->dramclk_khz <= dc->clk_mgr->bw_params->dc_mode_softmax_memclk * 1000) !=
+				(clk_mgr_base->clks.dramclk_khz <= dc->clk_mgr->bw_params->dc_mode_softmax_memclk * 1000))
+			update_pstate_unsupported_clk = true;
+	}
+
+	if (should_update_pstate_support(safe_to_lower, p_state_change_support, clk_mgr_base->clks.p_state_change_support) ||
+			update_pstate_unsupported_clk) {
 		clk_mgr_base->clks.p_state_change_support = p_state_change_support;
 
 		/* to disable P-State switching, set UCLK min = max */
-		if (!clk_mgr_base->clks.p_state_change_support)
-			dcn30_smu_set_hard_min_by_freq(clk_mgr, PPCLK_UCLK,
+		if (!clk_mgr_base->clks.p_state_change_support) {
+			if (dc->clk_mgr->dc_mode_softmax_enabled &&
+				new_clocks->dramclk_khz <= dc->clk_mgr->bw_params->dc_mode_softmax_memclk * 1000)
+				dcn30_smu_set_hard_min_by_freq(clk_mgr, PPCLK_UCLK,
+					dc->clk_mgr->bw_params->dc_mode_softmax_memclk);
+			else
+				dcn30_smu_set_hard_min_by_freq(clk_mgr, PPCLK_UCLK,
 					clk_mgr_base->bw_params->clk_table.entries[clk_mgr_base->bw_params->clk_table.num_entries - 1].memclk_mhz);
+		}
 	}
 
 	/* Always update saved value, even if new value not set due to P-State switching unsupported */
@@ -421,6 +385,24 @@ static void dcn3_set_hard_max_memclk(struct clk_mgr *clk_mgr_base)
 			clk_mgr_base->bw_params->clk_table.entries[clk_mgr_base->bw_params->clk_table.num_entries - 1].memclk_mhz);
 }
 
+static void dcn3_set_max_memclk(struct clk_mgr *clk_mgr_base, unsigned int memclk_mhz)
+{
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+
+	if (!clk_mgr->smu_present)
+		return;
+
+	dcn30_smu_set_hard_max_by_freq(clk_mgr, PPCLK_UCLK, memclk_mhz);
+}
+static void dcn3_set_min_memclk(struct clk_mgr *clk_mgr_base, unsigned int memclk_mhz)
+{
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+
+	if (!clk_mgr->smu_present)
+		return;
+	dcn30_smu_set_hard_min_by_freq(clk_mgr, PPCLK_UCLK, memclk_mhz);
+}
+
 /* Get current memclk states, update bounding box */
 static void dcn3_get_memclk_states_from_smu(struct clk_mgr *clk_mgr_base)
 {
@@ -436,9 +418,13 @@ static void dcn3_get_memclk_states_from_smu(struct clk_mgr *clk_mgr_base)
 			&num_levels);
 	clk_mgr_base->bw_params->clk_table.num_entries = num_levels ? num_levels : 1;
 
+	clk_mgr_base->bw_params->dc_mode_softmax_memclk = dcn30_smu_get_dc_mode_max_dpm_freq(clk_mgr, PPCLK_UCLK);
+
 	/* Refresh bounding box */
+	DC_FP_START();
 	clk_mgr_base->ctx->dc->res_pool->funcs->update_bw_bounding_box(
 			clk_mgr->base.ctx->dc, clk_mgr_base->bw_params);
+	DC_FP_END();
 }
 
 static bool dcn3_is_smu_present(struct clk_mgr *clk_mgr_base)
@@ -485,9 +471,11 @@ static void dcn30_notify_link_rate_change(struct clk_mgr *clk_mgr_base, struct d
 	if (!clk_mgr->smu_present)
 		return;
 
+	/* TODO - DP2.0 HW: calculate link 128b/132 link rate in clock manager with new formula */
+
 	clk_mgr->cur_phyclk_req_table[link->link_index] = link->cur_link_settings.link_rate * LINK_RATE_REF_FREQ_IN_KHZ;
 
-	for (i = 0; i < MAX_PIPES * 2; i++) {
+	for (i = 0; i < MAX_LINKS; i++) {
 		if (clk_mgr->cur_phyclk_req_table[i] > max_phyclk_req)
 			max_phyclk_req = clk_mgr->cur_phyclk_req_table[i];
 	}
@@ -505,11 +493,14 @@ static struct clk_mgr_funcs dcn3_funcs = {
 		.notify_wm_ranges = dcn3_notify_wm_ranges,
 		.set_hard_min_memclk = dcn3_set_hard_min_memclk,
 		.set_hard_max_memclk = dcn3_set_hard_max_memclk,
+		.set_max_memclk = dcn3_set_max_memclk,
+		.set_min_memclk = dcn3_set_min_memclk,
 		.get_memclk_states_from_smu = dcn3_get_memclk_states_from_smu,
 		.are_clock_states_equal = dcn3_are_clock_states_equal,
 		.enable_pme_wa = dcn3_enable_pme_wa,
 		.notify_link_rate_change = dcn30_notify_link_rate_change,
-		.is_smu_present = dcn3_is_smu_present
+		.is_smu_present = dcn3_is_smu_present,
+		.set_smartmux_switch = dcn30m_set_smartmux_switch
 };
 
 static void dcn3_init_clocks_fpga(struct clk_mgr *clk_mgr)
@@ -532,6 +523,8 @@ void dcn3_clk_mgr_construct(
 		struct pp_smu_funcs *pp_smu,
 		struct dccg *dccg)
 {
+	struct clk_state_registers_and_bypass s = { 0 };
+
 	clk_mgr->base.ctx = ctx;
 	clk_mgr->base.funcs = &dcn3_funcs;
 	clk_mgr->regs = &clk_mgr_regs;
@@ -548,27 +541,19 @@ void dcn3_clk_mgr_construct(
 
 	clk_mgr->base.dprefclk_khz = 730000; // 700 MHz planned if VCO is 3.85 GHz, will be retrieved
 
-	if (IS_FPGA_MAXIMUS_DC(ctx->dce_environment)) {
-		clk_mgr->base.funcs  = &dcn3_fpga_funcs;
-		clk_mgr->base.dentist_vco_freq_khz = 3650000;
-
-	} else {
-		struct clk_state_registers_and_bypass s = { 0 };
+	/* integer part is now VCO frequency in kHz */
+	clk_mgr->base.dentist_vco_freq_khz = dcn30_get_vco_frequency_from_reg(clk_mgr);
 
-		/* integer part is now VCO frequency in kHz */
-		clk_mgr->base.dentist_vco_freq_khz = dcn30_get_vco_frequency_from_reg(clk_mgr);
+	/* in case we don't get a value from the register, use default */
+	if (clk_mgr->base.dentist_vco_freq_khz == 0)
+		clk_mgr->base.dentist_vco_freq_khz = 3650000;
+	/* Convert dprefclk units from MHz to KHz */
+	/* Value already divided by 10, some resolution lost */
 
-		/* in case we don't get a value from the register, use default */
-		if (clk_mgr->base.dentist_vco_freq_khz == 0)
-			clk_mgr->base.dentist_vco_freq_khz = 3650000;
-		/* Convert dprefclk units from MHz to KHz */
-		/* Value already divided by 10, some resolution lost */
-
-		/*TODO: uncomment assert once dcn3_dump_clk_registers is implemented */
-		//ASSERT(s.dprefclk != 0);
-		if (s.dprefclk != 0)
-			clk_mgr->base.dprefclk_khz = s.dprefclk * 1000;
-	}
+	/*TODO: uncomment assert once dcn3_dump_clk_registers is implemented */
+	//ASSERT(s.dprefclk != 0);
+	if (s.dprefclk != 0)
+		clk_mgr->base.dprefclk_khz = s.dprefclk * 1000;
 
 	clk_mgr->dfs_bypass_enabled = false;
 
@@ -577,11 +562,19 @@ void dcn3_clk_mgr_construct(
 	dce_clock_read_ss_info(clk_mgr);
 
 	clk_mgr->base.bw_params = kzalloc(sizeof(*clk_mgr->base.bw_params), GFP_KERNEL);
+	if (!clk_mgr->base.bw_params) {
+		BREAK_TO_DEBUGGER();
+		return;
+	}
 
 	/* need physical address of table to give to PMFW */
 	clk_mgr->wm_range_table = dm_helpers_allocate_gpu_mem(clk_mgr->base.ctx,
 			DC_MEM_ALLOC_TYPE_GART, sizeof(WatermarksExternal_t),
 			&clk_mgr->wm_range_table_addr);
+	if (!clk_mgr->wm_range_table) {
+		BREAK_TO_DEBUGGER();
+		return;
+	}
 }
 
 void dcn3_clk_mgr_destroy(struct clk_mgr_internal *clk_mgr)
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dcn30_clk_mgr.h b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dcn30_clk_mgr.h
index dd4a0bd72458..2cd95ec38266 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dcn30_clk_mgr.h
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dcn30_clk_mgr.h
@@ -26,6 +26,66 @@
 #ifndef __DCN30_CLK_MGR_H__
 #define __DCN30_CLK_MGR_H__
 
+//CLK1_CLK_PLL_REQ
+#ifndef CLK11_CLK1_CLK_PLL_REQ__FbMult_int__SHIFT
+#define CLK11_CLK1_CLK_PLL_REQ__FbMult_int__SHIFT                                                                   0x0
+#define CLK11_CLK1_CLK_PLL_REQ__PllSpineDiv__SHIFT                                                                  0xc
+#define CLK11_CLK1_CLK_PLL_REQ__FbMult_frac__SHIFT                                                                  0x10
+#define CLK11_CLK1_CLK_PLL_REQ__FbMult_int_MASK                                                                     0x000001FFL
+#define CLK11_CLK1_CLK_PLL_REQ__PllSpineDiv_MASK                                                                    0x0000F000L
+#define CLK11_CLK1_CLK_PLL_REQ__FbMult_frac_MASK                                                                    0xFFFF0000L
+//CLK1_CLK0_DFS_CNTL
+#define CLK11_CLK1_CLK0_DFS_CNTL__CLK0_DIVIDER__SHIFT                                                               0x0
+#define CLK11_CLK1_CLK0_DFS_CNTL__CLK0_DIVIDER_MASK                                                                 0x0000007FL
+/*DPREF clock related*/
+#define CLK0_CLK3_DFS_CNTL__CLK3_DIVIDER__SHIFT                                                               0x0
+#define CLK0_CLK3_DFS_CNTL__CLK3_DIVIDER_MASK                                                                 0x0000007FL
+#define CLK1_CLK3_DFS_CNTL__CLK3_DIVIDER__SHIFT                                                               0x0
+#define CLK1_CLK3_DFS_CNTL__CLK3_DIVIDER_MASK                                                                 0x0000007FL
+#define CLK2_CLK3_DFS_CNTL__CLK3_DIVIDER__SHIFT                                                               0x0
+#define CLK2_CLK3_DFS_CNTL__CLK3_DIVIDER_MASK                                                                 0x0000007FL
+#define CLK3_CLK3_DFS_CNTL__CLK3_DIVIDER__SHIFT                                                               0x0
+#define CLK3_CLK3_DFS_CNTL__CLK3_DIVIDER_MASK                                                                 0x0000007FL
+
+//CLK3_0_CLK3_CLK_PLL_REQ
+#define CLK3_0_CLK3_CLK_PLL_REQ__FbMult_int__SHIFT                                                            0x0
+#define CLK3_0_CLK3_CLK_PLL_REQ__PllSpineDiv__SHIFT                                                           0xc
+#define CLK3_0_CLK3_CLK_PLL_REQ__FbMult_frac__SHIFT                                                           0x10
+#define CLK3_0_CLK3_CLK_PLL_REQ__FbMult_int_MASK                                                              0x000001FFL
+#define CLK3_0_CLK3_CLK_PLL_REQ__PllSpineDiv_MASK                                                             0x0000F000L
+#define CLK3_0_CLK3_CLK_PLL_REQ__FbMult_frac_MASK                                                             0xFFFF0000L
+
+#define mmCLK0_CLK2_DFS_CNTL                            0x16C55
+#define mmCLK00_CLK0_CLK2_DFS_CNTL                      0x16C55
+#define mmCLK01_CLK0_CLK2_DFS_CNTL                      0x16E55
+#define mmCLK02_CLK0_CLK2_DFS_CNTL                      0x17055
+
+#define mmCLK0_CLK3_DFS_CNTL                            0x16C60
+#define mmCLK00_CLK0_CLK3_DFS_CNTL                      0x16C60
+#define mmCLK01_CLK0_CLK3_DFS_CNTL                      0x16E60
+#define mmCLK02_CLK0_CLK3_DFS_CNTL                      0x17060
+#define mmCLK03_CLK0_CLK3_DFS_CNTL                      0x17260
+
+#define mmCLK0_CLK_PLL_REQ                              0x16C10
+#define mmCLK00_CLK0_CLK_PLL_REQ                        0x16C10
+#define mmCLK01_CLK0_CLK_PLL_REQ                        0x16E10
+#define mmCLK02_CLK0_CLK_PLL_REQ                        0x17010
+#define mmCLK03_CLK0_CLK_PLL_REQ                        0x17210
+
+#define mmCLK1_CLK_PLL_REQ                              0x1B00D
+#define mmCLK10_CLK1_CLK_PLL_REQ                        0x1B00D
+#define mmCLK11_CLK1_CLK_PLL_REQ                        0x1B20D
+#define mmCLK12_CLK1_CLK_PLL_REQ                        0x1B40D
+#define mmCLK13_CLK1_CLK_PLL_REQ                        0x1B60D
+
+#define mmCLK2_CLK_PLL_REQ                              0x17E0D
+
+/*AMCLK*/
+
+#define mmCLK11_CLK1_CLK0_DFS_CNTL                      0x1B23F
+#define mmCLK11_CLK1_CLK_PLL_REQ                        0x1B20D
+
+#endif
 void dcn3_init_clocks(struct clk_mgr *clk_mgr_base);
 
 void dcn3_clk_mgr_construct(struct dc_context *ctx,
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dcn30_clk_mgr_smu_msg.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dcn30_clk_mgr_smu_msg.c
index 8ecc708bcd9e..827bc2431d5d 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dcn30_clk_mgr_smu_msg.c
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dcn30_clk_mgr_smu_msg.c
@@ -23,12 +23,14 @@
  *
  */
 
-#include <linux/delay.h>
 #include "dcn30_clk_mgr_smu_msg.h"
 
 #include "clk_mgr_internal.h"
 #include "reg_helper.h"
+#include "dm_helpers.h"
+
 #include "dalsmc.h"
+#include "dcn30_smu11_driver_if.h"
 
 #define mmDAL_MSG_REG  0x1628A
 #define mmDAL_ARG_REG  0x16273
@@ -51,6 +53,7 @@
  */
 static uint32_t dcn30_smu_wait_for_response(struct clk_mgr_internal *clk_mgr, unsigned int delay_us, unsigned int max_retries)
 {
+	const uint32_t initial_max_retries = max_retries;
 	uint32_t reg = 0;
 
 	do {
@@ -66,13 +69,14 @@ static uint32_t dcn30_smu_wait_for_response(struct clk_mgr_internal *clk_mgr, un
 
 	/* handle DALSMC_Result_CmdRejectedBusy? */
 
-	/* Log? */
+	TRACE_SMU_MSG_DELAY(0, 0, delay_us * (initial_max_retries - max_retries), clk_mgr->base.ctx);
 
 	return reg;
 }
 
 static bool dcn30_smu_send_msg_with_param(struct clk_mgr_internal *clk_mgr, uint32_t msg_id, uint32_t param_in, uint32_t *param_out)
 {
+	uint32_t result;
 	/* Wait for response register to be ready */
 	dcn30_smu_wait_for_response(clk_mgr, 10, 200000);
 
@@ -85,8 +89,16 @@ static bool dcn30_smu_send_msg_with_param(struct clk_mgr_internal *clk_mgr, uint
 	/* Trigger the message transaction by writing the message ID */
 	REG_WRITE(DAL_MSG_REG, msg_id);
 
+	TRACE_SMU_MSG(msg_id, param_in, clk_mgr->base.ctx);
+
+	result = dcn30_smu_wait_for_response(clk_mgr, 10, 200000);
+
+	if (IS_SMU_TIMEOUT(result)) {
+		dm_helpers_smu_timeout(CTX, msg_id, param_in, 10 * 200000);
+	}
+
 	/* Wait for response */
-	if (dcn30_smu_wait_for_response(clk_mgr, 10, 200000) == DALSMC_Result_OK) {
+	if (result == DALSMC_Result_OK) {
 		if (param_out)
 			*param_out = REG_READ(DAL_ARG_REG);
 
@@ -197,7 +209,7 @@ void dcn30_smu_transfer_wm_table_dram_2_smu(struct clk_mgr_internal *clk_mgr)
 }
 
 /* Returns the actual frequency that was set in MHz, 0 on failure */
-unsigned int dcn30_smu_set_hard_min_by_freq(struct clk_mgr_internal *clk_mgr, PPCLK_e clk, uint16_t freq_mhz)
+unsigned int dcn30_smu_set_hard_min_by_freq(struct clk_mgr_internal *clk_mgr, uint32_t clk, uint16_t freq_mhz)
 {
 	uint32_t response = 0;
 
@@ -215,7 +227,7 @@ unsigned int dcn30_smu_set_hard_min_by_freq(struct clk_mgr_internal *clk_mgr, PP
 }
 
 /* Returns the actual frequency that was set in MHz, 0 on failure */
-unsigned int dcn30_smu_set_hard_max_by_freq(struct clk_mgr_internal *clk_mgr, PPCLK_e clk, uint16_t freq_mhz)
+unsigned int dcn30_smu_set_hard_max_by_freq(struct clk_mgr_internal *clk_mgr, uint32_t clk, uint16_t freq_mhz)
 {
 	uint32_t response = 0;
 
@@ -246,7 +258,7 @@ unsigned int dcn30_smu_set_hard_max_by_freq(struct clk_mgr_internal *clk_mgr, PP
  *
  * Returns 0 on failure
  */
-unsigned int dcn30_smu_get_dpm_freq_by_index(struct clk_mgr_internal *clk_mgr, PPCLK_e clk, uint8_t dpm_level)
+unsigned int dcn30_smu_get_dpm_freq_by_index(struct clk_mgr_internal *clk_mgr, uint32_t clk, uint8_t dpm_level)
 {
 	uint32_t response = 0;
 
@@ -264,7 +276,7 @@ unsigned int dcn30_smu_get_dpm_freq_by_index(struct clk_mgr_internal *clk_mgr, P
 }
 
 /* Returns the max DPM frequency in DC mode in MHz, 0 on failure */
-unsigned int dcn30_smu_get_dc_mode_max_dpm_freq(struct clk_mgr_internal *clk_mgr, PPCLK_e clk)
+unsigned int dcn30_smu_get_dc_mode_max_dpm_freq(struct clk_mgr_internal *clk_mgr, uint32_t clk)
 {
 	uint32_t response = 0;
 
@@ -302,6 +314,9 @@ void dcn30_smu_set_display_refresh_from_mall(struct clk_mgr_internal *clk_mgr, b
 	/* bits 8:7 for cache timer scale, bits 6:1 for cache timer delay, bit 0 = 1 for enable, = 0 for disable */
 	uint32_t param = (cache_timer_scale << 7) | (cache_timer_delay << 1) | (enable ? 1 : 0);
 
+	smu_print("SMU Set display refresh from mall: enable = %d, cache_timer_delay = %d, cache_timer_scale = %d\n",
+		enable, cache_timer_delay, cache_timer_scale);
+
 	dcn30_smu_send_msg_with_param(clk_mgr,
 			DALSMC_MSG_SetDisplayRefreshFromMall, param, NULL);
 }
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dcn30_clk_mgr_smu_msg.h b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dcn30_clk_mgr_smu_msg.h
index dd2640a3ce5d..ca9f5296be94 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dcn30_clk_mgr_smu_msg.h
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dcn30_clk_mgr_smu_msg.h
@@ -28,65 +28,6 @@
 
 #include "core_types.h"
 
-#define SMU11_DRIVER_IF_VERSION 0x1F
-
-typedef enum {
-	PPCLK_GFXCLK = 0,
-	PPCLK_SOCCLK,
-	PPCLK_UCLK,
-	PPCLK_FCLK,
-	PPCLK_DCLK_0,
-	PPCLK_VCLK_0,
-	PPCLK_DCLK_1,
-	PPCLK_VCLK_1,
-	PPCLK_DCEFCLK,
-	PPCLK_DISPCLK,
-	PPCLK_PIXCLK,
-	PPCLK_PHYCLK,
-	PPCLK_DTBCLK,
-	PPCLK_COUNT,
-} PPCLK_e;
-
-typedef struct {
-	uint16_t MinClock; // This is either DCEFCLK or SOCCLK (in MHz)
-	uint16_t MaxClock; // This is either DCEFCLK or SOCCLK (in MHz)
-	uint16_t MinUclk;
-	uint16_t MaxUclk;
-
-	uint8_t  WmSetting;
-	uint8_t  Flags;
-	uint8_t  Padding[2];
-
-} WatermarkRowGeneric_t;
-
-#define NUM_WM_RANGES 4
-
-typedef enum {
-	WM_SOCCLK = 0,
-	WM_DCEFCLK,
-	WM_COUNT,
-} WM_CLOCK_e;
-
-typedef enum {
-	WATERMARKS_CLOCK_RANGE = 0,
-	WATERMARKS_DUMMY_PSTATE,
-	WATERMARKS_MALL,
-	WATERMARKS_COUNT,
-} WATERMARKS_FLAGS_e;
-
-typedef struct {
-	// Watermarks
-	WatermarkRowGeneric_t WatermarkRow[WM_COUNT][NUM_WM_RANGES];
-} Watermarks_t;
-
-typedef struct {
-	Watermarks_t Watermarks;
-
-	uint32_t     MmHubPadding[8]; // SMU internal use
-} WatermarksExternal_t;
-
-#define TABLE_WATERMARKS 1
-
 struct clk_mgr_internal;
 
 bool         dcn30_smu_test_message(struct clk_mgr_internal *clk_mgr, uint32_t input);
@@ -97,10 +38,10 @@ void         dcn30_smu_set_dram_addr_high(struct clk_mgr_internal *clk_mgr, uint
 void         dcn30_smu_set_dram_addr_low(struct clk_mgr_internal *clk_mgr, uint32_t addr_low);
 void         dcn30_smu_transfer_wm_table_smu_2_dram(struct clk_mgr_internal *clk_mgr);
 void         dcn30_smu_transfer_wm_table_dram_2_smu(struct clk_mgr_internal *clk_mgr);
-unsigned int dcn30_smu_set_hard_min_by_freq(struct clk_mgr_internal *clk_mgr, PPCLK_e clk, uint16_t freq_mhz);
-unsigned int dcn30_smu_set_hard_max_by_freq(struct clk_mgr_internal *clk_mgr, PPCLK_e clk, uint16_t freq_mhz);
-unsigned int dcn30_smu_get_dpm_freq_by_index(struct clk_mgr_internal *clk_mgr, PPCLK_e clk, uint8_t dpm_level);
-unsigned int dcn30_smu_get_dc_mode_max_dpm_freq(struct clk_mgr_internal *clk_mgr, PPCLK_e clk);
+unsigned int dcn30_smu_set_hard_min_by_freq(struct clk_mgr_internal *clk_mgr, uint32_t clk, uint16_t freq_mhz);
+unsigned int dcn30_smu_set_hard_max_by_freq(struct clk_mgr_internal *clk_mgr, uint32_t clk, uint16_t freq_mhz);
+unsigned int dcn30_smu_get_dpm_freq_by_index(struct clk_mgr_internal *clk_mgr, uint32_t clk, uint8_t dpm_level);
+unsigned int dcn30_smu_get_dc_mode_max_dpm_freq(struct clk_mgr_internal *clk_mgr, uint32_t clk);
 void         dcn30_smu_set_min_deep_sleep_dcef_clk(struct clk_mgr_internal *clk_mgr, uint32_t freq_mhz);
 void         dcn30_smu_set_num_of_displays(struct clk_mgr_internal *clk_mgr, uint32_t num_displays);
 void         dcn30_smu_set_display_refresh_from_mall(struct clk_mgr_internal *clk_mgr, bool enable, uint8_t cache_timer_delay, uint8_t cache_timer_scale);
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dcn30_smu11_driver_if.h b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dcn30_smu11_driver_if.h
new file mode 100644
index 000000000000..1bfd6f66f035
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dcn30_smu11_driver_if.h
@@ -0,0 +1,75 @@
+// SPDX-License-Identifier: MIT
+/* Copyright © 2022-2024 Advanced Micro Devices, Inc. All rights reserved. */
+
+#define SMU11_DRIVER_IF_VERSION 0x40
+
+//Only Clks that have DPM descriptors are listed here
+typedef enum {
+	PPCLK_GFXCLK = 0,
+	PPCLK_SOCCLK,
+	PPCLK_UCLK,
+	PPCLK_FCLK,
+	PPCLK_DCLK_0,
+	PPCLK_VCLK_0,
+	PPCLK_DCLK_1,
+	PPCLK_VCLK_1,
+	PPCLK_DCEFCLK,
+	PPCLK_DISPCLK,
+	PPCLK_PIXCLK,
+	PPCLK_PHYCLK,
+	PPCLK_DTBCLK,
+	PPCLK_COUNT,
+} PPCLK_e;
+
+typedef struct {
+	uint16_t MinClock; // This is either DCEFCLK or SOCCLK (in MHz)
+	uint16_t MaxClock; // This is either DCEFCLK or SOCCLK (in MHz)
+	uint16_t MinUclk;
+	uint16_t MaxUclk;
+
+	uint8_t  WmSetting;
+	uint8_t  Flags;
+	uint8_t  Padding[2];
+
+} WatermarkRowGeneric_t;
+
+#define NUM_WM_RANGES 4
+
+typedef enum {
+	WM_SOCCLK = 0,
+	WM_DCEFCLK,
+	WM_COUNT,
+} WM_CLOCK_e;
+
+typedef enum {
+	WATERMARKS_CLOCK_RANGE = 0,
+	WATERMARKS_DUMMY_PSTATE,
+	WATERMARKS_MALL,
+	WATERMARKS_COUNT,
+} WATERMARKS_FLAGS_e;
+
+typedef struct {
+	// Watermarks
+	WatermarkRowGeneric_t WatermarkRow[WM_COUNT][NUM_WM_RANGES];
+} Watermarks_t;
+
+typedef struct {
+	Watermarks_t Watermarks;
+
+	uint32_t     MmHubPadding[8]; // SMU internal use
+} WatermarksExternal_t;
+
+// Table types
+#define TABLE_PPTABLE                 0
+#define TABLE_WATERMARKS              1
+#define TABLE_AVFS_PSM_DEBUG          2
+#define TABLE_AVFS_FUSE_OVERRIDE      3
+#define TABLE_PMSTATUSLOG             4
+#define TABLE_SMU_METRICS             5
+#define TABLE_DRIVER_SMU_CONFIG       6
+#define TABLE_ACTIVITY_MONITOR_COEFF  7
+#define TABLE_OVERDRIVE               8
+#define TABLE_I2C_COMMANDS            9
+#define TABLE_PACE                   10
+#define TABLE_ECCINFO                11
+#define TABLE_COUNT                  12
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dcn30m_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dcn30m_clk_mgr.c
new file mode 100644
index 000000000000..8e8a11c7437e
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dcn30m_clk_mgr.c
@@ -0,0 +1,36 @@
+/*
+ * Copyright 2019 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#include "clk_mgr_internal.h"
+#include "dcn30/dcn30m_clk_mgr.h"
+#include "dcn30m_clk_mgr_smu_msg.h"
+
+
+uint32_t dcn30m_set_smartmux_switch(struct clk_mgr *clk_mgr_base, uint32_t pins_to_set)
+{
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+
+	return dcn30m_smu_set_smart_mux_switch(clk_mgr, pins_to_set);
+}
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dcn30m_clk_mgr.h b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dcn30m_clk_mgr.h
new file mode 100644
index 000000000000..757985b2eadc
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dcn30m_clk_mgr.h
@@ -0,0 +1,31 @@
+/*
+ * Copyright 2019 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __DCN30M_CLK_MGR_H__
+#define __DCN30M_CLK_MGR_H__
+
+uint32_t dcn30m_set_smartmux_switch(struct clk_mgr *clk_mgr_base, uint32_t pins_to_set);
+
+#endif //__DCN30M_CLK_MGR_H__
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dcn30m_clk_mgr_smu_msg.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dcn30m_clk_mgr_smu_msg.c
new file mode 100644
index 000000000000..0dd0583ff21e
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dcn30m_clk_mgr_smu_msg.c
@@ -0,0 +1,118 @@
+/*
+ * Copyright 2020 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#include "dcn30m_clk_mgr_smu_msg.h"
+
+#include "clk_mgr_internal.h"
+#include "reg_helper.h"
+#include "dm_helpers.h"
+
+#include "dalsmc.h"
+
+#define mmDAL_MSG_REG  0x1628A
+#define mmDAL_ARG_REG  0x16273
+#define mmDAL_RESP_REG 0x16274
+
+#define REG(reg_name) \
+	mm ## reg_name
+
+#include "logger_types.h"
+#undef DC_LOGGER
+#define DC_LOGGER \
+	CTX->logger
+#define smu_print(str, ...) {DC_LOG_SMU(str, ##__VA_ARGS__); }
+
+
+/*
+ * Function to be used instead of REG_WAIT macro because the wait ends when
+ * the register is NOT EQUAL to zero, and because the translation in msg_if.h
+ * won't work with REG_WAIT.
+ */
+static uint32_t dcn30m_smu_wait_for_response(struct clk_mgr_internal *clk_mgr,
+	unsigned int delay_us, unsigned int max_retries)
+{
+	uint32_t reg = 0;
+
+	do {
+		reg = REG_READ(DAL_RESP_REG);
+		if (reg)
+			break;
+
+		if (delay_us >= 1000)
+			msleep(delay_us/1000);
+		else if (delay_us > 0)
+			udelay(delay_us);
+	} while (max_retries--);
+
+	/* handle DALSMC_Result_CmdRejectedBusy? */
+
+	/* Log? */
+
+	return reg;
+}
+
+static bool dcn30m_smu_send_msg_with_param(struct clk_mgr_internal *clk_mgr,
+	uint32_t msg_id, uint32_t param_in, uint32_t *param_out)
+{
+	uint32_t result;
+	/* Wait for response register to be ready */
+	dcn30m_smu_wait_for_response(clk_mgr, 10, 200000);
+
+	/* Clear response register */
+	REG_WRITE(DAL_RESP_REG, 0);
+
+	/* Set the parameter register for the SMU message */
+	REG_WRITE(DAL_ARG_REG, param_in);
+
+	/* Trigger the message transaction by writing the message ID */
+	REG_WRITE(DAL_MSG_REG, msg_id);
+
+	result = dcn30m_smu_wait_for_response(clk_mgr, 10, 200000);
+
+	if (IS_SMU_TIMEOUT(result))
+		dm_helpers_smu_timeout(CTX, msg_id, param_in, 10 * 200000);
+
+	/* Wait for response */
+	if (result == DALSMC_Result_OK) {
+		if (param_out)
+			*param_out = REG_READ(DAL_ARG_REG);
+
+		return true;
+	}
+
+	return false;
+}
+
+uint32_t dcn30m_smu_set_smart_mux_switch(struct clk_mgr_internal *clk_mgr, uint32_t pins_to_set)
+{
+	uint32_t response = 0;
+
+	smu_print("SMU Set SmartMux Switch: switch_dgpu = %d\n", pins_to_set);
+
+	dcn30m_smu_send_msg_with_param(clk_mgr,
+			DALSMC_MSG_SmartAccess, pins_to_set, &response);
+
+	return response;
+}
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dcn30m_clk_mgr_smu_msg.h b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dcn30m_clk_mgr_smu_msg.h
new file mode 100644
index 000000000000..8a59a473fc5e
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn30/dcn30m_clk_mgr_smu_msg.h
@@ -0,0 +1,34 @@
+/*
+ * Copyright 2020 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef DAL_DC_DCN30M_CLK_MGR_SMU_MSG_H_
+#define DAL_DC_DCN30M_CLK_MGR_SMU_MSG_H_
+
+#include "core_types.h"
+
+struct clk_mgr_internal;
+
+uint32_t     dcn30m_smu_set_smart_mux_switch(struct clk_mgr_internal *clk_mgr, uint32_t pins_to_set);
+#endif /* DAL_DC_DCN30M_CLK_MGR_SMU_MSG_H_ */
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn301/dcn301_smu.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn301/dcn301_smu.c
index 6ea642615854..b4fb17b7a096 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn301/dcn301_smu.c
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn301/dcn301_smu.c
@@ -29,6 +29,7 @@
 #include <linux/delay.h>
 
 #include "dcn301_smu.h"
+#include "dm_helpers.h"
 
 #include "vangogh_ip_offset.h"
 
@@ -41,6 +42,12 @@
 #define FN(reg_name, field) \
 	FD(reg_name##__##field)
 
+#include "logger_types.h"
+#undef DC_LOGGER
+#define DC_LOGGER \
+	CTX->logger
+#define smu_print(str, ...) {DC_LOG_SMU(str, ##__VA_ARGS__); }
+
 #define VBIOSSMC_MSG_GetSmuVersion                0x2
 #define VBIOSSMC_MSG_SetDispclkFreq               0x4
 #define VBIOSSMC_MSG_SetDprefclkFreq              0x5
@@ -88,12 +95,21 @@ static uint32_t dcn301_smu_wait_for_response(struct clk_mgr_internal *clk_mgr, u
 	return res_val;
 }
 
-int dcn301_smu_send_msg_with_param(
-		struct clk_mgr_internal *clk_mgr,
-		unsigned int msg_id, unsigned int param)
+static int dcn301_smu_send_msg_with_param(struct clk_mgr_internal *clk_mgr,
+					  unsigned int msg_id,
+					  unsigned int param)
 {
 	uint32_t result;
 
+	result = dcn301_smu_wait_for_response(clk_mgr, 10, 200000);
+
+	if (result != VBIOSSMC_Result_OK)
+		smu_print("SMU Response was not OK. SMU response after wait received is: %d\n", result);
+
+	if (result == VBIOSSMC_Status_BUSY) {
+		return -1;
+	}
+
 	/* First clear response register */
 	REG_WRITE(MP1_SMN_C2PMSG_91, VBIOSSMC_Status_BUSY);
 
@@ -105,7 +121,10 @@ int dcn301_smu_send_msg_with_param(
 
 	result = dcn301_smu_wait_for_response(clk_mgr, 10, 200000);
 
-	ASSERT(result == VBIOSSMC_Result_OK);
+	if (IS_SMU_TIMEOUT(result)) {
+		ASSERT(0);
+		dm_helpers_smu_timeout(CTX, msg_id, param, 10 * 200000);
+	}
 
 	/* Actual dispclk set is returned in the parameter register */
 	return REG_READ(MP1_SMN_C2PMSG_83);
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn301/vg_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn301/vg_clk_mgr.c
index 3eee32faa208..7aee02d56292 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn301/vg_clk_mgr.c
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn301/vg_clk_mgr.c
@@ -32,6 +32,9 @@
 // For dcn20_update_clocks_update_dpp_dto
 #include "dcn20/dcn20_clk_mgr.h"
 
+// For DML FPU code
+#include "dml/dcn20/dcn20_fpu.h"
+
 #include "vg_clk_mgr.h"
 #include "dcn301_smu.h"
 #include "reg_helper.h"
@@ -89,9 +92,9 @@ static int vg_get_active_display_cnt_wa(
 	return display_count;
 }
 
-void vg_update_clocks(struct clk_mgr *clk_mgr_base,
-			struct dc_state *context,
-			bool safe_to_lower)
+static void vg_update_clocks(struct clk_mgr *clk_mgr_base,
+			     struct dc_state *context,
+			     bool safe_to_lower)
 {
 	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
 	struct dc_clocks *new_clocks = &context->bw_ctx.bw.dcn.clk;
@@ -114,7 +117,7 @@ void vg_update_clocks(struct clk_mgr *clk_mgr_base,
 
 			display_count = vg_get_active_display_cnt_wa(dc, context);
 			/* if we can go lower, go lower */
-			if (display_count == 0 && !IS_DIAG_DC(dc->ctx->dce_environment)) {
+			if (display_count == 0) {
 				union display_idle_optimization_u idle_info = { 0 };
 
 				idle_info.idle_info.df_request_disabled = 1;
@@ -148,10 +151,8 @@ void vg_update_clocks(struct clk_mgr *clk_mgr_base,
 	}
 
 	// workaround: Limit dppclk to 100Mhz to avoid lower eDP panel switch to plus 4K monitor underflow.
-	if (!IS_DIAG_DC(dc->ctx->dce_environment)) {
-		if (new_clocks->dppclk_khz < 100000)
-			new_clocks->dppclk_khz = 100000;
-	}
+	if (new_clocks->dppclk_khz < 100000)
+		new_clocks->dppclk_khz = 100000;
 
 	if (should_set_clock(safe_to_lower, new_clocks->dppclk_khz, clk_mgr->base.clks.dppclk_khz)) {
 		if (clk_mgr->base.clks.dppclk_khz > new_clocks->dppclk_khz)
@@ -176,8 +177,7 @@ void vg_update_clocks(struct clk_mgr *clk_mgr_base,
 		if (update_dppclk || update_dispclk)
 			dcn301_smu_set_dppclk(clk_mgr, clk_mgr_base->clks.dppclk_khz);
 		// always update dtos unless clock is lowered and not safe to lower
-		if (new_clocks->dppclk_khz >= dc->current_state->bw_ctx.bw.dcn.clk.dppclk_khz)
-			dcn20_update_clocks_update_dpp_dto(clk_mgr, context, safe_to_lower);
+		dcn20_update_clocks_update_dpp_dto(clk_mgr, context, safe_to_lower);
 	}
 }
 
@@ -252,16 +252,16 @@ static void vg_dump_clk_registers(struct clk_state_registers_and_bypass *regs_an
 	regs_and_bypass->dppclk = internal.CLK1_CLK1_CURRENT_CNT / 10;
 
 	regs_and_bypass->dppclk_bypass = internal.CLK1_CLK1_BYPASS_CNTL & 0x0007;
-	if (regs_and_bypass->dppclk_bypass < 0 || regs_and_bypass->dppclk_bypass > 4)
+	if (regs_and_bypass->dppclk_bypass > 4)
 		regs_and_bypass->dppclk_bypass = 0;
 	regs_and_bypass->dcfclk_bypass = internal.CLK1_CLK3_BYPASS_CNTL & 0x0007;
-	if (regs_and_bypass->dcfclk_bypass < 0 || regs_and_bypass->dcfclk_bypass > 4)
+	if (regs_and_bypass->dcfclk_bypass > 4)
 		regs_and_bypass->dcfclk_bypass = 0;
 	regs_and_bypass->dispclk_bypass = internal.CLK1_CLK0_BYPASS_CNTL & 0x0007;
-	if (regs_and_bypass->dispclk_bypass < 0 || regs_and_bypass->dispclk_bypass > 4)
+	if (regs_and_bypass->dispclk_bypass > 4)
 		regs_and_bypass->dispclk_bypass = 0;
 	regs_and_bypass->dprefclk_bypass = internal.CLK1_CLK2_BYPASS_CNTL & 0x0007;
-	if (regs_and_bypass->dprefclk_bypass < 0 || regs_and_bypass->dprefclk_bypass > 4)
+	if (regs_and_bypass->dprefclk_bypass > 4)
 		regs_and_bypass->dprefclk_bypass = 0;
 
 	if (log_info->enabled) {
@@ -367,18 +367,6 @@ static void vg_dump_clk_registers(struct clk_state_registers_and_bypass *regs_an
 	}
 }
 
-/* This function produce translated logical clk state values*/
-void vg_get_clk_states(struct clk_mgr *clk_mgr_base, struct clk_states *s)
-{
-
-	struct clk_state_registers_and_bypass sb = { 0 };
-	struct clk_log_info log_info = { 0 };
-
-	vg_dump_clk_registers(&sb, clk_mgr_base, &log_info);
-
-	s->dprefclk_khz = sb.dprefclk * 1000;
-}
-
 static void vg_enable_pme_wa(struct clk_mgr *clk_mgr_base)
 {
 	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
@@ -386,7 +374,7 @@ static void vg_enable_pme_wa(struct clk_mgr *clk_mgr_base)
 	dcn301_smu_enable_pme_wa(clk_mgr);
 }
 
-void vg_init_clocks(struct clk_mgr *clk_mgr)
+static void vg_init_clocks(struct clk_mgr *clk_mgr)
 {
 	memset(&(clk_mgr->clks), 0, sizeof(struct dc_clocks));
 	// Assumption is that boot state always supports pstate
@@ -539,80 +527,18 @@ static struct clk_bw_params vg_bw_params = {
 
 };
 
-static struct wm_table ddr4_wm_table = {
-	.entries = {
-		{
-			.wm_inst = WM_A,
-			.wm_type = WM_TYPE_PSTATE_CHG,
-			.pstate_latency_us = 11.72,
-			.sr_exit_time_us = 6.09,
-			.sr_enter_plus_exit_time_us = 7.14,
-			.valid = true,
-		},
-		{
-			.wm_inst = WM_B,
-			.wm_type = WM_TYPE_PSTATE_CHG,
-			.pstate_latency_us = 11.72,
-			.sr_exit_time_us = 10.12,
-			.sr_enter_plus_exit_time_us = 11.48,
-			.valid = true,
-		},
-		{
-			.wm_inst = WM_C,
-			.wm_type = WM_TYPE_PSTATE_CHG,
-			.pstate_latency_us = 11.72,
-			.sr_exit_time_us = 10.12,
-			.sr_enter_plus_exit_time_us = 11.48,
-			.valid = true,
-		},
-		{
-			.wm_inst = WM_D,
-			.wm_type = WM_TYPE_PSTATE_CHG,
-			.pstate_latency_us = 11.72,
-			.sr_exit_time_us = 10.12,
-			.sr_enter_plus_exit_time_us = 11.48,
-			.valid = true,
-		},
-	}
-};
+static uint32_t find_max_clk_value(const uint32_t clocks[], uint32_t num_clocks)
+{
+	uint32_t max = 0;
+	int i;
 
-static struct wm_table lpddr5_wm_table = {
-	.entries = {
-		{
-			.wm_inst = WM_A,
-			.wm_type = WM_TYPE_PSTATE_CHG,
-			.pstate_latency_us = 11.65333,
-			.sr_exit_time_us = 7.95,
-			.sr_enter_plus_exit_time_us = 9,
-			.valid = true,
-		},
-		{
-			.wm_inst = WM_B,
-			.wm_type = WM_TYPE_PSTATE_CHG,
-			.pstate_latency_us = 11.65333,
-			.sr_exit_time_us = 9.82,
-			.sr_enter_plus_exit_time_us = 11.196,
-			.valid = true,
-		},
-		{
-			.wm_inst = WM_C,
-			.wm_type = WM_TYPE_PSTATE_CHG,
-			.pstate_latency_us = 11.65333,
-			.sr_exit_time_us = 9.89,
-			.sr_enter_plus_exit_time_us = 11.24,
-			.valid = true,
-		},
-		{
-			.wm_inst = WM_D,
-			.wm_type = WM_TYPE_PSTATE_CHG,
-			.pstate_latency_us = 11.65333,
-			.sr_exit_time_us = 9.748,
-			.sr_enter_plus_exit_time_us = 11.102,
-			.valid = true,
-		},
+	for (i = 0; i < num_clocks; ++i) {
+		if (clocks[i] > max)
+			max = clocks[i];
 	}
-};
 
+	return max;
+}
 
 static unsigned int find_dcfclk_for_voltage(const struct vg_dpm_clocks *clock_table,
 		unsigned int voltage)
@@ -620,6 +546,8 @@ static unsigned int find_dcfclk_for_voltage(const struct vg_dpm_clocks *clock_ta
 	int i;
 
 	for (i = 0; i < VG_NUM_SOC_VOLTAGE_LEVELS; i++) {
+		if (i >= VG_NUM_DCFCLK_DPM_LEVELS)
+			break;
 		if (clock_table->SocVoltage[i] == voltage)
 			return clock_table->DcfClocks[i];
 	}
@@ -635,10 +563,12 @@ static void vg_clk_mgr_helper_populate_bw_params(
 {
 	int i, j;
 	struct clk_bw_params *bw_params = clk_mgr->base.bw_params;
+	uint32_t max_dispclk = 0, max_dppclk = 0;
 
 	j = -1;
 
-	ASSERT(VG_NUM_FCLK_DPM_LEVELS <= MAX_NUM_DPM_LVL);
+	static_assert(VG_NUM_FCLK_DPM_LEVELS <= MAX_NUM_DPM_LVL,
+		"number of reported FCLK DPM levels exceeds maximum");
 
 	/* Find lowest DPM, FCLK is filled in reverse order*/
 
@@ -655,14 +585,33 @@ static void vg_clk_mgr_helper_populate_bw_params(
 		return;
 	}
 
+	/* dispclk and dppclk can be max at any voltage, same number of levels for both */
+	if (clock_table->NumDispClkLevelsEnabled <= VG_NUM_DISPCLK_DPM_LEVELS &&
+	    clock_table->NumDispClkLevelsEnabled <= VG_NUM_DPPCLK_DPM_LEVELS) {
+		max_dispclk = find_max_clk_value(clock_table->DispClocks, clock_table->NumDispClkLevelsEnabled);
+		max_dppclk = find_max_clk_value(clock_table->DppClocks, clock_table->NumDispClkLevelsEnabled);
+	} else {
+		ASSERT(0);
+	}
+
 	bw_params->clk_table.num_entries = j + 1;
 
-	for (i = 0; i < bw_params->clk_table.num_entries; i++, j--) {
+	for (i = 0; i < bw_params->clk_table.num_entries - 1; i++, j--) {
 		bw_params->clk_table.entries[i].fclk_mhz = clock_table->DfPstateTable[j].fclk;
 		bw_params->clk_table.entries[i].memclk_mhz = clock_table->DfPstateTable[j].memclk;
 		bw_params->clk_table.entries[i].voltage = clock_table->DfPstateTable[j].voltage;
 		bw_params->clk_table.entries[i].dcfclk_mhz = find_dcfclk_for_voltage(clock_table, clock_table->DfPstateTable[j].voltage);
+
+		/* Now update clocks we do read */
+		bw_params->clk_table.entries[i].dispclk_mhz = max_dispclk;
+		bw_params->clk_table.entries[i].dppclk_mhz = max_dppclk;
 	}
+	bw_params->clk_table.entries[i].fclk_mhz = clock_table->DfPstateTable[j].fclk;
+	bw_params->clk_table.entries[i].memclk_mhz = clock_table->DfPstateTable[j].memclk;
+	bw_params->clk_table.entries[i].voltage = clock_table->DfPstateTable[j].voltage;
+	bw_params->clk_table.entries[i].dcfclk_mhz = find_max_clk_value(clock_table->DcfClocks, VG_NUM_DCFCLK_DPM_LEVELS);
+	bw_params->clk_table.entries[i].dispclk_mhz = find_max_clk_value(clock_table->DispClocks, VG_NUM_DISPCLK_DPM_LEVELS);
+	bw_params->clk_table.entries[i].dppclk_mhz = find_max_clk_value(clock_table->DppClocks, VG_NUM_DPPCLK_DPM_LEVELS);
 
 	bw_params->vram_type = bios_info->memory_type;
 	bw_params->num_channels = bios_info->ma_channel_number;
@@ -683,10 +632,9 @@ static void vg_clk_mgr_helper_populate_bw_params(
 		/*
 		 * WM set D will be re-purposed for memory retraining
 		 */
-		bw_params->wm_table.entries[WM_D].pstate_latency_us = LPDDR_MEM_RETRAIN_LATENCY;
-		bw_params->wm_table.entries[WM_D].wm_inst = WM_D;
-		bw_params->wm_table.entries[WM_D].wm_type = WM_TYPE_RETRAINING;
-		bw_params->wm_table.entries[WM_D].valid = true;
+		DC_FP_START();
+		dcn21_clk_mgr_set_bw_params_wm_table(bw_params);
+		DC_FP_END();
 	}
 
 }
@@ -733,6 +681,7 @@ void vg_clk_mgr_construct(
 		struct dccg *dccg)
 {
 	struct smu_dpm_clks smu_dpm_clks = { 0 };
+	struct clk_log_info log_info = {0};
 
 	clk_mgr->base.base.ctx = ctx;
 	clk_mgr->base.base.funcs = &vg_funcs;
@@ -753,7 +702,7 @@ void vg_clk_mgr_construct(
 				sizeof(struct watermarks),
 				&clk_mgr->smu_wm_set.mc_address.quad_part);
 
-	if (clk_mgr->smu_wm_set.wm_set == 0) {
+	if (!clk_mgr->smu_wm_set.wm_set) {
 		clk_mgr->smu_wm_set.wm_set = &dummy_wms;
 		clk_mgr->smu_wm_set.mc_address.quad_part = 0;
 	}
@@ -772,32 +721,25 @@ void vg_clk_mgr_construct(
 
 	ASSERT(smu_dpm_clks.dpm_clks);
 
-	if (IS_FPGA_MAXIMUS_DC(ctx->dce_environment)) {
-		vg_funcs.update_clocks = dcn2_update_clocks_fpga;
-		clk_mgr->base.base.dentist_vco_freq_khz = 3600000;
-	} else {
-		struct clk_log_info log_info = {0};
-
-		clk_mgr->base.smu_ver = dcn301_smu_get_smu_version(&clk_mgr->base);
+	clk_mgr->base.smu_ver = dcn301_smu_get_smu_version(&clk_mgr->base);
 
-		if (clk_mgr->base.smu_ver)
-			clk_mgr->base.smu_present = true;
+	if (clk_mgr->base.smu_ver)
+		clk_mgr->base.smu_present = true;
 
-		/* TODO: Check we get what we expect during bringup */
-		clk_mgr->base.base.dentist_vco_freq_khz = get_vco_frequency_from_reg(&clk_mgr->base);
+	/* TODO: Check we get what we expect during bringup */
+	clk_mgr->base.base.dentist_vco_freq_khz = get_vco_frequency_from_reg(&clk_mgr->base);
 
-		/* in case we don't get a value from the register, use default */
-		if (clk_mgr->base.base.dentist_vco_freq_khz == 0)
-			clk_mgr->base.base.dentist_vco_freq_khz = 3600000;
+	/* in case we don't get a value from the register, use default */
+	if (clk_mgr->base.base.dentist_vco_freq_khz == 0)
+		clk_mgr->base.base.dentist_vco_freq_khz = 3600000;
 
-		if (ctx->dc_bios->integrated_info->memory_type == LpDdr5MemType) {
-			vg_bw_params.wm_table = lpddr5_wm_table;
-		} else {
-			vg_bw_params.wm_table = ddr4_wm_table;
-		}
-		/* Saved clocks configured at boot for debug purposes */
-		vg_dump_clk_registers(&clk_mgr->base.base.boot_snapshot, &clk_mgr->base.base, &log_info);
+	if (ctx->dc_bios->integrated_info->memory_type == LpDdr5MemType) {
+		vg_bw_params.wm_table = lpddr5_wm_table;
+	} else {
+		vg_bw_params.wm_table = ddr4_wm_table;
 	}
+	/* Saved clocks configured at boot for debug purposes */
+	vg_dump_clk_registers(&clk_mgr->base.base.boot_snapshot, &clk_mgr->base.base, &log_info);
 
 	clk_mgr->base.base.dprefclk_khz = 600000;
 	dce_clock_read_ss_info(&clk_mgr->base);
@@ -805,7 +747,7 @@ void vg_clk_mgr_construct(
 	clk_mgr->base.base.bw_params = &vg_bw_params;
 
 	vg_get_dpm_table_from_smu(&clk_mgr->base, &smu_dpm_clks);
-	if (ctx->dc_bios && ctx->dc_bios->integrated_info) {
+	if (ctx->dc_bios->integrated_info) {
 		vg_clk_mgr_helper_populate_bw_params(
 				&clk_mgr->base,
 				ctx->dc_bios->integrated_info,
@@ -815,12 +757,6 @@ void vg_clk_mgr_construct(
 	if (smu_dpm_clks.dpm_clks && smu_dpm_clks.mc_address.quad_part != 0)
 		dm_helpers_free_gpu_mem(clk_mgr->base.base.ctx, DC_MEM_ALLOC_TYPE_FRAME_BUFFER,
 				smu_dpm_clks.dpm_clks);
-/*
-	if (!IS_FPGA_MAXIMUS_DC(ctx->dce_environment) && clk_mgr->base.smu_ver) {
-		 enable powerfeatures when displaycount goes to 0
-		dcn301_smu_enable_phy_refclk_pwrdwn(clk_mgr, !debug->disable_48mhz_pwrdwn);
-	}
-*/
 }
 
 void vg_clk_mgr_destroy(struct clk_mgr_internal *clk_mgr_int)
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn301/vg_clk_mgr.h b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn301/vg_clk_mgr.h
index 7255477307f1..75884f572989 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn301/vg_clk_mgr.h
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn301/vg_clk_mgr.h
@@ -29,6 +29,9 @@
 
 struct watermarks;
 
+extern struct wm_table ddr4_wm_table;
+extern struct wm_table lpddr5_wm_table;
+
 struct smu_watermark_set {
 	struct watermarks *wm_set;
 	union large_integer mc_address;
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn31/dcn31_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn31/dcn31_clk_mgr.c
index f4c9a458ace8..051052bd10c9 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn31/dcn31_clk_mgr.c
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn31/dcn31_clk_mgr.c
@@ -47,8 +47,15 @@
 #include "dcn30/dcn30_clk_mgr.h"
 
 #include "dc_dmub_srv.h"
+#include "link_service.h"
+
+#include "logger_types.h"
+
 
 #include "yellow_carp_offset.h"
+#undef DC_LOGGER
+#define DC_LOGGER \
+	clk_mgr->base.base.ctx->logger
 
 #define regCLK1_CLK_PLL_REQ			0x0237
 #define regCLK1_CLK_PLL_REQ_BASE_IDX		0
@@ -66,7 +73,7 @@
 #define TO_CLK_MGR_DCN31(clk_mgr)\
 	container_of(clk_mgr, struct clk_mgr_dcn31, base)
 
-int dcn31_get_active_display_cnt_wa(
+static int dcn31_get_active_display_cnt_wa(
 		struct dc *dc,
 		struct dc_state *context)
 {
@@ -81,6 +88,11 @@ int dcn31_get_active_display_cnt_wa(
 				stream->signal == SIGNAL_TYPE_DVI_SINGLE_LINK ||
 				stream->signal == SIGNAL_TYPE_DVI_DUAL_LINK)
 			tmds_present = true;
+
+		/* Checking stream / link detection ensuring that PHY is active*/
+		if (dc_is_dp_signal(stream->signal) && !stream->dpms_off)
+			display_count++;
+
 	}
 
 	for (i = 0; i < dc->link_count; i++) {
@@ -99,7 +111,7 @@ int dcn31_get_active_display_cnt_wa(
 	return display_count;
 }
 
-static void dcn31_disable_otg_wa(struct clk_mgr *clk_mgr_base, bool disable)
+static void dcn31_disable_otg_wa(struct clk_mgr *clk_mgr_base, struct dc_state *context, bool disable)
 {
 	struct dc *dc = clk_mgr_base->ctx->dc;
 	int i;
@@ -110,15 +122,16 @@ static void dcn31_disable_otg_wa(struct clk_mgr *clk_mgr_base, bool disable)
 		if (pipe->top_pipe || pipe->prev_odm_pipe)
 			continue;
 		if (pipe->stream && (pipe->stream->dpms_off || dc_is_virtual_signal(pipe->stream->signal))) {
-			if (disable)
+			if (disable) {
 				pipe->stream_res.tg->funcs->immediate_disable_crtc(pipe->stream_res.tg);
-			else
+				reset_sync_context_for_pipe(dc, context, i);
+			} else
 				pipe->stream_res.tg->funcs->enable_crtc(pipe->stream_res.tg);
 		}
 	}
 }
 
-static void dcn31_update_clocks(struct clk_mgr *clk_mgr_base,
+void dcn31_update_clocks(struct clk_mgr *clk_mgr_base,
 			struct dc_state *context,
 			bool safe_to_lower)
 {
@@ -139,9 +152,9 @@ static void dcn31_update_clocks(struct clk_mgr *clk_mgr_base,
 	 * also if safe to lower is false, we just go in the higher state
 	 */
 	if (safe_to_lower) {
-		if (new_clocks->zstate_support == DCN_ZSTATE_SUPPORT_ALLOW &&
+		if (new_clocks->zstate_support != DCN_ZSTATE_SUPPORT_DISALLOW &&
 				new_clocks->zstate_support != clk_mgr_base->clks.zstate_support) {
-			dcn31_smu_set_Z9_support(clk_mgr, true);
+			dcn31_smu_set_zstate_support(clk_mgr, new_clocks->zstate_support);
 			dm_helpers_enable_periodic_detection(clk_mgr_base->ctx, true);
 			clk_mgr_base->clks.zstate_support = new_clocks->zstate_support;
 		}
@@ -158,6 +171,7 @@ static void dcn31_update_clocks(struct clk_mgr *clk_mgr_base,
 				union display_idle_optimization_u idle_info = { 0 };
 				idle_info.idle_info.df_request_disabled = 1;
 				idle_info.idle_info.phy_ref_clk_off = 1;
+				idle_info.idle_info.s0i2_rdy = 1;
 				dcn31_smu_set_display_idle_optimization(clk_mgr, idle_info.data);
 				/* update power state */
 				clk_mgr_base->clks.pwr_state = DCN_PWR_STATE_LOW_POWER;
@@ -166,7 +180,7 @@ static void dcn31_update_clocks(struct clk_mgr *clk_mgr_base,
 	} else {
 		if (new_clocks->zstate_support == DCN_ZSTATE_SUPPORT_DISALLOW &&
 				new_clocks->zstate_support != clk_mgr_base->clks.zstate_support) {
-			dcn31_smu_set_Z9_support(clk_mgr, false);
+			dcn31_smu_set_zstate_support(clk_mgr, DCN_ZSTATE_SUPPORT_DISALLOW);
 			dm_helpers_enable_periodic_detection(clk_mgr_base->ctx, false);
 			clk_mgr_base->clks.zstate_support = new_clocks->zstate_support;
 		}
@@ -197,10 +211,8 @@ static void dcn31_update_clocks(struct clk_mgr *clk_mgr_base,
 	}
 
 	// workaround: Limit dppclk to 100Mhz to avoid lower eDP panel switch to plus 4K monitor underflow.
-	if (!IS_DIAG_DC(dc->ctx->dce_environment)) {
-		if (new_clocks->dppclk_khz < 100000)
-			new_clocks->dppclk_khz = 100000;
-	}
+	if (new_clocks->dppclk_khz < 100000)
+		new_clocks->dppclk_khz = 100000;
 
 	if (should_set_clock(safe_to_lower, new_clocks->dppclk_khz, clk_mgr->base.clks.dppclk_khz)) {
 		if (clk_mgr->base.clks.dppclk_khz > new_clocks->dppclk_khz)
@@ -210,11 +222,11 @@ static void dcn31_update_clocks(struct clk_mgr *clk_mgr_base,
 	}
 
 	if (should_set_clock(safe_to_lower, new_clocks->dispclk_khz, clk_mgr_base->clks.dispclk_khz)) {
-		dcn31_disable_otg_wa(clk_mgr_base, true);
+		dcn31_disable_otg_wa(clk_mgr_base, context, true);
 
 		clk_mgr_base->clks.dispclk_khz = new_clocks->dispclk_khz;
 		dcn31_smu_set_dispclk(clk_mgr, clk_mgr_base->clks.dispclk_khz);
-		dcn31_disable_otg_wa(clk_mgr_base, false);
+		dcn31_disable_otg_wa(clk_mgr_base, context, false);
 
 		update_dispclk = true;
 	}
@@ -242,9 +254,7 @@ static void dcn31_update_clocks(struct clk_mgr *clk_mgr_base,
 	cmd.notify_clocks.clocks.dispclk_khz = clk_mgr_base->clks.dispclk_khz;
 	cmd.notify_clocks.clocks.dppclk_khz = clk_mgr_base->clks.dppclk_khz;
 
-	dc_dmub_srv_cmd_queue(dc->ctx->dmub_srv, &cmd);
-	dc_dmub_srv_cmd_execute(dc->ctx->dmub_srv);
-	dc_dmub_srv_wait_idle(dc->ctx->dmub_srv);
+	dc_wake_and_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
 }
 
 static int get_vco_frequency_from_reg(struct clk_mgr_internal *clk_mgr)
@@ -284,17 +294,20 @@ static void dcn31_enable_pme_wa(struct clk_mgr *clk_mgr_base)
 	dcn31_smu_enable_pme_wa(clk_mgr);
 }
 
-static void dcn31_init_clocks(struct clk_mgr *clk_mgr)
+void dcn31_init_clocks(struct clk_mgr *clk_mgr)
 {
+	uint32_t ref_dtbclk = clk_mgr->clks.ref_dtbclk_khz;
+
 	memset(&(clk_mgr->clks), 0, sizeof(struct dc_clocks));
 	// Assumption is that boot state always supports pstate
+	clk_mgr->clks.ref_dtbclk_khz = ref_dtbclk;	// restore ref_dtbclk
 	clk_mgr->clks.p_state_change_support = true;
 	clk_mgr->clks.prev_p_state_change_support = true;
 	clk_mgr->clks.pwr_state = DCN_PWR_STATE_UNKNOWN;
 	clk_mgr->clks.zstate_support = DCN_ZSTATE_SUPPORT_UNKNOWN;
 }
 
-static bool dcn31_are_clock_states_equal(struct dc_clocks *a,
+bool dcn31_are_clock_states_equal(struct dc_clocks *a,
 		struct dc_clocks *b)
 {
 	if (a->dispclk_khz != b->dispclk_khz)
@@ -328,38 +341,38 @@ static struct clk_bw_params dcn31_bw_params = {
 
 };
 
-static struct wm_table ddr4_wm_table = {
+static struct wm_table ddr5_wm_table = {
 	.entries = {
 		{
 			.wm_inst = WM_A,
 			.wm_type = WM_TYPE_PSTATE_CHG,
 			.pstate_latency_us = 11.72,
-			.sr_exit_time_us = 6.09,
-			.sr_enter_plus_exit_time_us = 7.14,
+			.sr_exit_time_us = 9,
+			.sr_enter_plus_exit_time_us = 11,
 			.valid = true,
 		},
 		{
 			.wm_inst = WM_B,
 			.wm_type = WM_TYPE_PSTATE_CHG,
 			.pstate_latency_us = 11.72,
-			.sr_exit_time_us = 10.12,
-			.sr_enter_plus_exit_time_us = 11.48,
+			.sr_exit_time_us = 9,
+			.sr_enter_plus_exit_time_us = 11,
 			.valid = true,
 		},
 		{
 			.wm_inst = WM_C,
 			.wm_type = WM_TYPE_PSTATE_CHG,
 			.pstate_latency_us = 11.72,
-			.sr_exit_time_us = 10.12,
-			.sr_enter_plus_exit_time_us = 11.48,
+			.sr_exit_time_us = 9,
+			.sr_enter_plus_exit_time_us = 11,
 			.valid = true,
 		},
 		{
 			.wm_inst = WM_D,
 			.wm_type = WM_TYPE_PSTATE_CHG,
 			.pstate_latency_us = 11.72,
-			.sr_exit_time_us = 10.12,
-			.sr_enter_plus_exit_time_us = 11.48,
+			.sr_exit_time_us = 9,
+			.sr_enter_plus_exit_time_us = 11,
 			.valid = true,
 		},
 	}
@@ -540,10 +553,9 @@ static unsigned int find_clk_for_voltage(
 	return clock;
 }
 
-void dcn31_clk_mgr_helper_populate_bw_params(
-		struct clk_mgr_internal *clk_mgr,
-		struct integrated_info *bios_info,
-		const DpmClocks_t *clock_table)
+static void dcn31_clk_mgr_helper_populate_bw_params(struct clk_mgr_internal *clk_mgr,
+						    struct integrated_info *bios_info,
+						    const DpmClocks_t *clock_table)
 {
 	int i, j;
 	struct clk_bw_params *bw_params = clk_mgr->base.bw_params;
@@ -551,7 +563,8 @@ void dcn31_clk_mgr_helper_populate_bw_params(
 
 	j = -1;
 
-	ASSERT(NUM_DF_PSTATE_LEVELS <= MAX_NUM_DPM_LVL);
+	static_assert(NUM_DF_PSTATE_LEVELS <= MAX_NUM_DPM_LVL,
+		"number of reported pstate levels exceeds maximum");
 
 	/* Find lowest DPM, FCLK is filled in reverse order*/
 
@@ -600,8 +613,10 @@ void dcn31_clk_mgr_helper_populate_bw_params(
 	}
 
 	bw_params->vram_type = bios_info->memory_type;
-	bw_params->num_channels = bios_info->ma_channel_number;
 
+	bw_params->dram_channel_width_bytes = bios_info->memory_type == 0x22 ? 8 : 4;
+	//bw_params->dram_channel_width_bytes = dc->ctx->asic_id.vram_width;
+	bw_params->num_channels = bios_info->ma_channel_number ? bios_info->ma_channel_number : 4;
 	for (i = 0; i < WM_SET_COUNT; i++) {
 		bw_params->wm_table.entries[i].wm_inst = i;
 
@@ -615,13 +630,43 @@ void dcn31_clk_mgr_helper_populate_bw_params(
 	}
 }
 
+static void dcn31_set_low_power_state(struct clk_mgr *clk_mgr_base)
+{
+	int display_count;
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+	struct dc *dc = clk_mgr_base->ctx->dc;
+	struct dc_state *context = dc->current_state;
+
+	if (clk_mgr_base->clks.pwr_state != DCN_PWR_STATE_LOW_POWER) {
+		display_count = dcn31_get_active_display_cnt_wa(dc, context);
+		/* if we can go lower, go lower */
+		if (display_count == 0) {
+			union display_idle_optimization_u idle_info = { 0 };
+
+			idle_info.idle_info.df_request_disabled = 1;
+			idle_info.idle_info.phy_ref_clk_off = 1;
+			idle_info.idle_info.s0i2_rdy = 1;
+			dcn31_smu_set_display_idle_optimization(clk_mgr, idle_info.data);
+			/* update power state */
+			clk_mgr_base->clks.pwr_state = DCN_PWR_STATE_LOW_POWER;
+		}
+	}
+}
+
+int dcn31_get_dtb_ref_freq_khz(struct clk_mgr *clk_mgr_base)
+{
+	return clk_mgr_base->clks.ref_dtbclk_khz;
+}
+
 static struct clk_mgr_funcs dcn31_funcs = {
 	.get_dp_ref_clk_frequency = dce12_get_dp_ref_freq_khz,
+	.get_dtb_ref_clk_frequency = dcn31_get_dtb_ref_freq_khz,
 	.update_clocks = dcn31_update_clocks,
 	.init_clocks = dcn31_init_clocks,
 	.enable_pme_wa = dcn31_enable_pme_wa,
 	.are_clock_states_equal = dcn31_are_clock_states_equal,
-	.notify_wm_ranges = dcn31_notify_wm_ranges
+	.notify_wm_ranges = dcn31_notify_wm_ranges,
+	.set_low_power_state = dcn31_set_low_power_state
 };
 extern struct clk_mgr_funcs dcn3_fpga_funcs;
 
@@ -632,6 +677,7 @@ void dcn31_clk_mgr_construct(
 		struct dccg *dccg)
 {
 	struct dcn31_smu_dpm_clks smu_dpm_clks = { 0 };
+	struct clk_log_info log_info = {0};
 
 	clk_mgr->base.base.ctx = ctx;
 	clk_mgr->base.base.funcs = &dcn31_funcs;
@@ -671,39 +717,76 @@ void dcn31_clk_mgr_construct(
 
 	ASSERT(smu_dpm_clks.dpm_clks);
 
-	if (IS_FPGA_MAXIMUS_DC(ctx->dce_environment)) {
-		clk_mgr->base.base.funcs = &dcn3_fpga_funcs;
-	} else {
-		struct clk_log_info log_info = {0};
-
-		clk_mgr->base.smu_ver = dcn31_smu_get_smu_version(&clk_mgr->base);
+	clk_mgr->base.smu_ver = dcn31_smu_get_smu_version(&clk_mgr->base);
 
-		if (clk_mgr->base.smu_ver)
-			clk_mgr->base.smu_present = true;
+	if (clk_mgr->base.smu_ver)
+		clk_mgr->base.smu_present = true;
 
-		/* TODO: Check we get what we expect during bringup */
-		clk_mgr->base.base.dentist_vco_freq_khz = get_vco_frequency_from_reg(&clk_mgr->base);
-
-		if (ctx->dc_bios->integrated_info->memory_type == LpDdr5MemType) {
-			dcn31_bw_params.wm_table = lpddr5_wm_table;
-		} else {
-			dcn31_bw_params.wm_table = ddr4_wm_table;
-		}
-		/* Saved clocks configured at boot for debug purposes */
-		 dcn31_dump_clk_registers(&clk_mgr->base.base.boot_snapshot, &clk_mgr->base.base, &log_info);
+	/* TODO: Check we get what we expect during bringup */
+	clk_mgr->base.base.dentist_vco_freq_khz = get_vco_frequency_from_reg(&clk_mgr->base);
 
+	if (ctx->dc_bios->integrated_info->memory_type == LpDdr5MemType) {
+		dcn31_bw_params.wm_table = lpddr5_wm_table;
+	} else {
+		dcn31_bw_params.wm_table = ddr5_wm_table;
 	}
+	/* Saved clocks configured at boot for debug purposes */
+	dcn31_dump_clk_registers(&clk_mgr->base.base.boot_snapshot,
+				 &clk_mgr->base.base, &log_info);
 
 	clk_mgr->base.base.dprefclk_khz = 600000;
-	clk_mgr->base.dccg->ref_dtbclk_khz = 600000;
+	clk_mgr->base.base.clks.ref_dtbclk_khz = 600000;
 	dce_clock_read_ss_info(&clk_mgr->base);
+	/*if bios enabled SS, driver needs to adjust dtb clock, only enable with correct bios*/
+	//clk_mgr->base.dccg->ref_dtbclk_khz = dce_adjust_dp_ref_freq_for_ss(clk_mgr_internal, clk_mgr->base.base.dprefclk_khz);
 
 	clk_mgr->base.base.bw_params = &dcn31_bw_params;
 
 	if (clk_mgr->base.base.ctx->dc->debug.pstate_enabled) {
+		int i;
+
 		dcn31_get_dpm_table_from_smu(&clk_mgr->base, &smu_dpm_clks);
 
-		if (ctx->dc_bios && ctx->dc_bios->integrated_info) {
+		DC_LOG_SMU("NumDcfClkLevelsEnabled: %d\n"
+				   "NumDispClkLevelsEnabled: %d\n"
+				   "NumSocClkLevelsEnabled: %d\n"
+				   "VcnClkLevelsEnabled: %d\n"
+				   "NumDfPst atesEnabled: %d\n"
+				   "MinGfxClk: %d\n"
+				   "MaxGfxClk: %d\n",
+				   smu_dpm_clks.dpm_clks->NumDcfClkLevelsEnabled,
+				   smu_dpm_clks.dpm_clks->NumDispClkLevelsEnabled,
+				   smu_dpm_clks.dpm_clks->NumSocClkLevelsEnabled,
+				   smu_dpm_clks.dpm_clks->VcnClkLevelsEnabled,
+				   smu_dpm_clks.dpm_clks->NumDfPstatesEnabled,
+				   smu_dpm_clks.dpm_clks->MinGfxClk,
+				   smu_dpm_clks.dpm_clks->MaxGfxClk);
+		for (i = 0; i < smu_dpm_clks.dpm_clks->NumDcfClkLevelsEnabled; i++) {
+			DC_LOG_SMU("smu_dpm_clks.dpm_clks->DcfClocks[%d] = %d\n",
+					   i,
+					   smu_dpm_clks.dpm_clks->DcfClocks[i]);
+		}
+		for (i = 0; i < smu_dpm_clks.dpm_clks->NumDispClkLevelsEnabled; i++) {
+			DC_LOG_SMU("smu_dpm_clks.dpm_clks->DispClocks[%d] = %d\n",
+					   i, smu_dpm_clks.dpm_clks->DispClocks[i]);
+		}
+		for (i = 0; i < smu_dpm_clks.dpm_clks->NumSocClkLevelsEnabled; i++) {
+			DC_LOG_SMU("smu_dpm_clks.dpm_clks->SocClocks[%d] = %d\n",
+					   i, smu_dpm_clks.dpm_clks->SocClocks[i]);
+		}
+		for (i = 0; i < NUM_SOC_VOLTAGE_LEVELS; i++)
+			DC_LOG_SMU("smu_dpm_clks.dpm_clks->SocVoltage[%d] = %d\n",
+					   i, smu_dpm_clks.dpm_clks->SocVoltage[i]);
+
+		for (i = 0; i < NUM_DF_PSTATE_LEVELS; i++) {
+			DC_LOG_SMU("smu_dpm_clks.dpm_clks.DfPstateTable[%d].FClk = %d\n"
+					   "smu_dpm_clks.dpm_clks->DfPstateTable[%d].MemClk= %d\n"
+					   "smu_dpm_clks.dpm_clks->DfPstateTable[%d].Voltage = %d\n",
+					   i, smu_dpm_clks.dpm_clks->DfPstateTable[i].FClk,
+					   i, smu_dpm_clks.dpm_clks->DfPstateTable[i].MemClk,
+					   i, smu_dpm_clks.dpm_clks->DfPstateTable[i].Voltage);
+		}
+		if (ctx->dc_bios->integrated_info) {
 			dcn31_clk_mgr_helper_populate_bw_params(
 					&clk_mgr->base,
 					ctx->dc_bios->integrated_info,
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn31/dcn31_clk_mgr.h b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn31/dcn31_clk_mgr.h
index f8f100535526..be06fdbd0c22 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn31/dcn31_clk_mgr.h
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn31/dcn31_clk_mgr.h
@@ -39,11 +39,20 @@ struct clk_mgr_dcn31 {
 	struct dcn31_smu_watermark_set smu_wm_set;
 };
 
+bool dcn31_are_clock_states_equal(struct dc_clocks *a,
+		struct dc_clocks *b);
+void dcn31_init_clocks(struct clk_mgr *clk_mgr);
+void dcn31_update_clocks(struct clk_mgr *clk_mgr_base,
+			struct dc_state *context,
+			bool safe_to_lower);
+
 void dcn31_clk_mgr_construct(struct dc_context *ctx,
 		struct clk_mgr_dcn31 *clk_mgr,
 		struct pp_smu_funcs *pp_smu,
 		struct dccg *dccg);
 
+int dcn31_get_dtb_ref_freq_khz(struct clk_mgr *clk_mgr_base);
+
 void dcn31_clk_mgr_destroy(struct clk_mgr_internal *clk_mgr_int);
 
 #endif //__DCN31_CLK_MGR_H__
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn31/dcn31_smu.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn31/dcn31_smu.c
index 8c2b77eb9459..f201628e4e98 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn31/dcn31_smu.c
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn31/dcn31_smu.c
@@ -23,7 +23,6 @@
  *
  */
 
-#include <linux/delay.h>
 #include "core_types.h"
 #include "clk_mgr_internal.h"
 #include "reg_helper.h"
@@ -40,6 +39,12 @@
 #define FN(reg_name, field) \
 	FD(reg_name##__##field)
 
+#include "logger_types.h"
+#undef DC_LOGGER
+#define DC_LOGGER \
+	CTX->logger
+#define smu_print(str, ...) {DC_LOG_SMU(str, ##__VA_ARGS__); }
+
 #define VBIOSSMC_MSG_TestMessage                  0x1
 #define VBIOSSMC_MSG_GetSmuVersion                0x2
 #define VBIOSSMC_MSG_PowerUpGfx                   0x3
@@ -95,14 +100,16 @@ static uint32_t dcn31_smu_wait_for_response(struct clk_mgr_internal *clk_mgr, un
 	return res_val;
 }
 
-int dcn31_smu_send_msg_with_param(
-		struct clk_mgr_internal *clk_mgr,
-		unsigned int msg_id, unsigned int param)
+static int dcn31_smu_send_msg_with_param(struct clk_mgr_internal *clk_mgr,
+					 unsigned int msg_id,
+					 unsigned int param)
 {
 	uint32_t result;
 
 	result = dcn31_smu_wait_for_response(clk_mgr, 10, 200000);
-	ASSERT(result == VBIOSSMC_Result_OK);
+
+	if (result != VBIOSSMC_Result_OK)
+		smu_print("SMU Response was not OK. SMU response after wait received is: %d\n", result);
 
 	if (result == VBIOSSMC_Status_BUSY) {
 		return -1;
@@ -119,6 +126,16 @@ int dcn31_smu_send_msg_with_param(
 
 	result = dcn31_smu_wait_for_response(clk_mgr, 10, 200000);
 
+	if (result == VBIOSSMC_Result_Failed) {
+		if (msg_id == VBIOSSMC_MSG_TransferTableDram2Smu &&
+		    param == TABLE_WATERMARKS)
+			DC_LOG_DEBUG("Watermarks table not configured properly by SMU");
+		else
+			ASSERT(0);
+		REG_WRITE(MP1_SMN_C2PMSG_91, VBIOSSMC_Result_OK);
+		return -1;
+	}
+
 	if (IS_SMU_TIMEOUT(result)) {
 		ASSERT(0);
 		dm_helpers_smu_timeout(CTX, msg_id, param, 10 * 200000);
@@ -300,23 +317,32 @@ void dcn31_smu_transfer_wm_table_dram_2_smu(struct clk_mgr_internal *clk_mgr)
 			VBIOSSMC_MSG_TransferTableDram2Smu, TABLE_WATERMARKS);
 }
 
-void dcn31_smu_set_Z9_support(struct clk_mgr_internal *clk_mgr, bool support)
+void dcn31_smu_set_zstate_support(struct clk_mgr_internal *clk_mgr, enum dcn_zstate_support_state support)
 {
-	//TODO: Work with smu team to define optimization options.
-	unsigned int msg_id;
+	unsigned int msg_id, param;
 
 	if (!clk_mgr->smu_present)
 		return;
 
-	if (support)
-		msg_id = VBIOSSMC_MSG_AllowZstatesEntry;
+	if (!clk_mgr->base.ctx->dc->debug.enable_z9_disable_interface &&
+			(support == DCN_ZSTATE_SUPPORT_ALLOW_Z10_ONLY))
+		support = DCN_ZSTATE_SUPPORT_DISALLOW;
+
+	if (support == DCN_ZSTATE_SUPPORT_ALLOW_Z10_ONLY ||
+	    support == DCN_ZSTATE_SUPPORT_ALLOW_Z8_Z10_ONLY)
+		param = 1;
 	else
+		param = 0;
+
+	if (support == DCN_ZSTATE_SUPPORT_DISALLOW)
 		msg_id = VBIOSSMC_MSG_DisallowZstatesEntry;
+	else
+		msg_id = VBIOSSMC_MSG_AllowZstatesEntry;
 
 	dcn31_smu_send_msg_with_param(
 		clk_mgr,
 		msg_id,
-		0);
+		param);
 
 }
 
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn31/dcn31_smu.h b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn31/dcn31_smu.h
index cd0b7e1e685f..dfa25a76a6d1 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn31/dcn31_smu.h
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn31/dcn31_smu.h
@@ -265,7 +265,7 @@ void dcn31_smu_set_dram_addr_low(struct clk_mgr_internal *clk_mgr, uint32_t addr
 void dcn31_smu_transfer_dpm_table_smu_2_dram(struct clk_mgr_internal *clk_mgr);
 void dcn31_smu_transfer_wm_table_dram_2_smu(struct clk_mgr_internal *clk_mgr);
 
-void dcn31_smu_set_Z9_support(struct clk_mgr_internal *clk_mgr, bool support);
+void dcn31_smu_set_zstate_support(struct clk_mgr_internal *clk_mgr, enum dcn_zstate_support_state support);
 void dcn31_smu_set_dtbclk(struct clk_mgr_internal *clk_mgr, bool enable);
 
 #endif /* DAL_DC_31_SMU_H_ */
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn314/dcn314_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn314/dcn314_clk_mgr.c
new file mode 100644
index 000000000000..db687a13174d
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn314/dcn314_clk_mgr.c
@@ -0,0 +1,1051 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+
+
+#include "dcn314_clk_mgr.h"
+
+#include "dccg.h"
+#include "clk_mgr_internal.h"
+
+// For dce12_get_dp_ref_freq_khz
+#include "dce100/dce_clk_mgr.h"
+
+// For dcn20_update_clocks_update_dpp_dto
+#include "dcn20/dcn20_clk_mgr.h"
+
+
+
+#include "reg_helper.h"
+#include "core_types.h"
+#include "dm_helpers.h"
+
+/* TODO: remove this include once we ported over remaining clk mgr functions*/
+#include "dcn30/dcn30_clk_mgr.h"
+#include "dcn31/dcn31_clk_mgr.h"
+
+#include "dc_dmub_srv.h"
+#include "link_service.h"
+#include "dcn314_smu.h"
+
+
+#include "logger_types.h"
+
+
+#define MAX_INSTANCE                                        7
+#define MAX_SEGMENT                                         8
+
+struct IP_BASE_INSTANCE {
+	unsigned int segment[MAX_SEGMENT];
+};
+
+struct IP_BASE {
+	struct IP_BASE_INSTANCE instance[MAX_INSTANCE];
+};
+
+static const struct IP_BASE CLK_BASE = { { { { 0x00016C00, 0x02401800, 0, 0, 0, 0, 0, 0 } },
+					{ { 0x00016E00, 0x02401C00, 0, 0, 0, 0, 0, 0 } },
+					{ { 0x00017000, 0x02402000, 0, 0, 0, 0, 0, 0 } },
+					{ { 0x00017200, 0x02402400, 0, 0, 0, 0, 0, 0 } },
+					{ { 0x0001B000, 0x0242D800, 0, 0, 0, 0, 0, 0 } },
+					{ { 0x0001B200, 0x0242DC00, 0, 0, 0, 0, 0, 0 } },
+					{ { 0x0001B400, 0x0242E000, 0, 0, 0, 0, 0, 0 } } } };
+
+#undef DC_LOGGER
+#define DC_LOGGER \
+	clk_mgr->base.base.ctx->logger
+
+#define regCLK1_CLK_PLL_REQ			0x0237
+#define regCLK1_CLK_PLL_REQ_BASE_IDX		0
+
+#define CLK1_CLK_PLL_REQ__FbMult_int__SHIFT	0x0
+#define CLK1_CLK_PLL_REQ__PllSpineDiv__SHIFT	0xc
+#define CLK1_CLK_PLL_REQ__FbMult_frac__SHIFT	0x10
+#define CLK1_CLK_PLL_REQ__FbMult_int_MASK	0x000001FFL
+#define CLK1_CLK_PLL_REQ__PllSpineDiv_MASK	0x0000F000L
+#define CLK1_CLK_PLL_REQ__FbMult_frac_MASK	0xFFFF0000L
+
+#define regCLK1_CLK0_DFS_CNTL				0x0269
+#define regCLK1_CLK0_DFS_CNTL_BASE_IDX		0
+#define regCLK1_CLK1_DFS_CNTL				0x026c
+#define regCLK1_CLK1_DFS_CNTL_BASE_IDX		0
+#define regCLK1_CLK2_DFS_CNTL				0x026f
+#define regCLK1_CLK2_DFS_CNTL_BASE_IDX		0
+#define regCLK1_CLK3_DFS_CNTL				0x0272
+#define regCLK1_CLK3_DFS_CNTL_BASE_IDX		0
+#define regCLK1_CLK4_DFS_CNTL				0x0275
+#define regCLK1_CLK4_DFS_CNTL_BASE_IDX		0
+#define regCLK1_CLK5_DFS_CNTL				0x0278
+#define regCLK1_CLK5_DFS_CNTL_BASE_IDX		0
+
+#define regCLK1_CLK0_CURRENT_CNT			0x02fb
+#define regCLK1_CLK0_CURRENT_CNT_BASE_IDX	0
+#define regCLK1_CLK1_CURRENT_CNT			0x02fc
+#define regCLK1_CLK1_CURRENT_CNT_BASE_IDX	0
+#define regCLK1_CLK2_CURRENT_CNT			0x02fd
+#define regCLK1_CLK2_CURRENT_CNT_BASE_IDX	0
+#define regCLK1_CLK3_CURRENT_CNT			0x02fe
+#define regCLK1_CLK3_CURRENT_CNT_BASE_IDX	0
+#define regCLK1_CLK4_CURRENT_CNT			0x02ff
+#define regCLK1_CLK4_CURRENT_CNT_BASE_IDX	0
+#define regCLK1_CLK5_CURRENT_CNT			0x0300
+#define regCLK1_CLK5_CURRENT_CNT_BASE_IDX	0
+
+#define regCLK1_CLK0_BYPASS_CNTL			0x028a
+#define regCLK1_CLK0_BYPASS_CNTL_BASE_IDX	0
+#define regCLK1_CLK1_BYPASS_CNTL			0x0293
+#define regCLK1_CLK1_BYPASS_CNTL_BASE_IDX	0
+#define regCLK1_CLK2_BYPASS_CNTL			0x029c
+#define regCLK1_CLK2_BYPASS_CNTL_BASE_IDX	0
+#define regCLK1_CLK3_BYPASS_CNTL			0x02a5
+#define regCLK1_CLK3_BYPASS_CNTL_BASE_IDX	0
+#define regCLK1_CLK4_BYPASS_CNTL			0x02ae
+#define regCLK1_CLK4_BYPASS_CNTL_BASE_IDX	0
+#define regCLK1_CLK5_BYPASS_CNTL			0x02b7
+#define regCLK1_CLK5_BYPASS_CNTL_BASE_IDX	0
+
+#define regCLK1_CLK0_DS_CNTL				0x0283
+#define regCLK1_CLK0_DS_CNTL_BASE_IDX		0
+#define regCLK1_CLK1_DS_CNTL				0x028c
+#define regCLK1_CLK1_DS_CNTL_BASE_IDX		0
+#define regCLK1_CLK2_DS_CNTL				0x0295
+#define regCLK1_CLK2_DS_CNTL_BASE_IDX		0
+#define regCLK1_CLK3_DS_CNTL				0x029e
+#define regCLK1_CLK3_DS_CNTL_BASE_IDX		0
+#define regCLK1_CLK4_DS_CNTL				0x02a7
+#define regCLK1_CLK4_DS_CNTL_BASE_IDX		0
+#define regCLK1_CLK5_DS_CNTL				0x02b0
+#define regCLK1_CLK5_DS_CNTL_BASE_IDX		0
+
+#define regCLK1_CLK0_ALLOW_DS				0x0284
+#define regCLK1_CLK0_ALLOW_DS_BASE_IDX		0
+#define regCLK1_CLK1_ALLOW_DS				0x028d
+#define regCLK1_CLK1_ALLOW_DS_BASE_IDX		0
+#define regCLK1_CLK2_ALLOW_DS				0x0296
+#define regCLK1_CLK2_ALLOW_DS_BASE_IDX		0
+#define regCLK1_CLK3_ALLOW_DS				0x029f
+#define regCLK1_CLK3_ALLOW_DS_BASE_IDX		0
+#define regCLK1_CLK4_ALLOW_DS				0x02a8
+#define regCLK1_CLK4_ALLOW_DS_BASE_IDX		0
+#define regCLK1_CLK5_ALLOW_DS				0x02b1
+#define regCLK1_CLK5_ALLOW_DS_BASE_IDX		0
+
+#define CLK1_CLK2_BYPASS_CNTL__CLK2_BYPASS_SEL__SHIFT	0x0
+#define CLK1_CLK2_BYPASS_CNTL__CLK2_BYPASS_DIV__SHIFT	0x10
+#define CLK1_CLK2_BYPASS_CNTL__CLK2_BYPASS_SEL_MASK		0x00000007L
+#define CLK1_CLK2_BYPASS_CNTL__CLK2_BYPASS_DIV_MASK		0x000F0000L
+
+#define regCLK6_0_CLK6_spll_field_8				0x464b
+#define regCLK6_0_CLK6_spll_field_8_BASE_IDX	0
+
+#define CLK6_0_CLK6_spll_field_8__spll_ssc_en__SHIFT	0xd
+#define CLK6_0_CLK6_spll_field_8__spll_ssc_en_MASK		0x00002000L
+
+#define REG(reg_name) \
+	(CLK_BASE.instance[0].segment[reg ## reg_name ## _BASE_IDX] + reg ## reg_name)
+
+#define TO_CLK_MGR_DCN314(clk_mgr)\
+	container_of(clk_mgr, struct clk_mgr_dcn314, base)
+
+static int dcn314_get_active_display_cnt_wa(
+		struct dc *dc,
+		struct dc_state *context)
+{
+	int i, display_count;
+	bool tmds_present = false;
+
+	display_count = 0;
+	for (i = 0; i < context->stream_count; i++) {
+		const struct dc_stream_state *stream = context->streams[i];
+
+		if (stream->signal == SIGNAL_TYPE_HDMI_TYPE_A ||
+				stream->signal == SIGNAL_TYPE_DVI_SINGLE_LINK ||
+				stream->signal == SIGNAL_TYPE_DVI_DUAL_LINK)
+			tmds_present = true;
+
+		/* Checking stream / link detection ensuring that PHY is active*/
+		if (dc_is_dp_signal(stream->signal) && !stream->dpms_off)
+			display_count++;
+
+	}
+
+	for (i = 0; i < dc->link_count; i++) {
+		const struct dc_link *link = dc->links[i];
+
+		/* abusing the fact that the dig and phy are coupled to see if the phy is enabled */
+		if (link->link_enc && link->link_enc->funcs->is_dig_enabled &&
+				link->link_enc->funcs->is_dig_enabled(link->link_enc))
+			display_count++;
+	}
+
+	/* WA for hang on HDMI after display off back on*/
+	if (display_count == 0 && tmds_present)
+		display_count = 1;
+
+	return display_count;
+}
+
+static void dcn314_disable_otg_wa(struct clk_mgr *clk_mgr_base, struct dc_state *context,
+				  bool safe_to_lower, bool disable)
+{
+	struct dc *dc = clk_mgr_base->ctx->dc;
+	int i;
+
+	for (i = 0; i < dc->res_pool->pipe_count; ++i) {
+		struct pipe_ctx *pipe = safe_to_lower
+			? &context->res_ctx.pipe_ctx[i]
+			: &dc->current_state->res_ctx.pipe_ctx[i];
+
+		if (pipe->top_pipe || pipe->prev_odm_pipe)
+			continue;
+		if (pipe->stream && (pipe->stream->dpms_off || dc_is_virtual_signal(pipe->stream->signal))) {
+			if (disable) {
+				if (pipe->stream_res.tg && pipe->stream_res.tg->funcs->immediate_disable_crtc)
+					pipe->stream_res.tg->funcs->immediate_disable_crtc(pipe->stream_res.tg);
+
+				reset_sync_context_for_pipe(dc, context, i);
+			} else {
+				pipe->stream_res.tg->funcs->enable_crtc(pipe->stream_res.tg);
+			}
+		}
+	}
+}
+
+bool dcn314_is_spll_ssc_enabled(struct clk_mgr *clk_mgr_base)
+{
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+	uint32_t ssc_enable;
+
+	REG_GET(CLK6_0_CLK6_spll_field_8, spll_ssc_en, &ssc_enable);
+
+	return ssc_enable == 1;
+}
+
+void dcn314_init_clocks(struct clk_mgr *clk_mgr)
+{
+	struct clk_mgr_internal *clk_mgr_int = TO_CLK_MGR_INTERNAL(clk_mgr);
+	uint32_t ref_dtbclk = clk_mgr->clks.ref_dtbclk_khz;
+	struct clk_mgr_dcn314 *clk_mgr_dcn314 = TO_CLK_MGR_DCN314(clk_mgr_int);
+	struct clk_log_info log_info = {0};
+
+	memset(&(clk_mgr->clks), 0, sizeof(struct dc_clocks));
+	// Assumption is that boot state always supports pstate
+	clk_mgr->clks.ref_dtbclk_khz = ref_dtbclk;	// restore ref_dtbclk
+	clk_mgr->clks.p_state_change_support = true;
+	clk_mgr->clks.prev_p_state_change_support = true;
+	clk_mgr->clks.pwr_state = DCN_PWR_STATE_UNKNOWN;
+	clk_mgr->clks.zstate_support = DCN_ZSTATE_SUPPORT_UNKNOWN;
+
+	// to adjust dp_dto reference clock if ssc is enable otherwise to apply dprefclk
+	if (dcn314_is_spll_ssc_enabled(clk_mgr))
+		clk_mgr->dp_dto_source_clock_in_khz =
+			dce_adjust_dp_ref_freq_for_ss(clk_mgr_int, clk_mgr->dprefclk_khz);
+	else
+		clk_mgr->dp_dto_source_clock_in_khz = clk_mgr->dprefclk_khz;
+
+	dcn314_dump_clk_registers(&clk_mgr->boot_snapshot, &clk_mgr_dcn314->base.base, &log_info);
+	clk_mgr->clks.dispclk_khz =  clk_mgr->boot_snapshot.dispclk * 1000;
+}
+
+void dcn314_update_clocks(struct clk_mgr *clk_mgr_base,
+			struct dc_state *context,
+			bool safe_to_lower)
+{
+	union dmub_rb_cmd cmd;
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+	struct dc_clocks *new_clocks = &context->bw_ctx.bw.dcn.clk;
+	struct dc *dc = clk_mgr_base->ctx->dc;
+	int display_count;
+	bool update_dppclk = false;
+	bool update_dispclk = false;
+	bool dpp_clock_lowered = false;
+
+	if (dc->work_arounds.skip_clock_update)
+		return;
+
+	display_count = dcn314_get_active_display_cnt_wa(dc, context);
+
+	/*
+	 * if it is safe to lower, but we are already in the lower state, we don't have to do anything
+	 * also if safe to lower is false, we just go in the higher state
+	 */
+	if (safe_to_lower) {
+		if (new_clocks->zstate_support != DCN_ZSTATE_SUPPORT_DISALLOW &&
+				new_clocks->zstate_support != clk_mgr_base->clks.zstate_support) {
+			dcn314_smu_set_zstate_support(clk_mgr, new_clocks->zstate_support);
+			dm_helpers_enable_periodic_detection(clk_mgr_base->ctx, true);
+			clk_mgr_base->clks.zstate_support = new_clocks->zstate_support;
+		}
+
+		if (clk_mgr_base->clks.dtbclk_en && !new_clocks->dtbclk_en) {
+			dcn314_smu_set_dtbclk(clk_mgr, false);
+			clk_mgr_base->clks.dtbclk_en = new_clocks->dtbclk_en;
+		}
+		/* check that we're not already in lower */
+		if (clk_mgr_base->clks.pwr_state != DCN_PWR_STATE_LOW_POWER) {
+			/* if we can go lower, go lower */
+			if (display_count == 0) {
+				union display_idle_optimization_u idle_info = { 0 };
+				idle_info.idle_info.df_request_disabled = 1;
+				idle_info.idle_info.phy_ref_clk_off = 1;
+				idle_info.idle_info.s0i2_rdy = 1;
+				dcn314_smu_set_display_idle_optimization(clk_mgr, idle_info.data);
+				/* update power state */
+				clk_mgr_base->clks.pwr_state = DCN_PWR_STATE_LOW_POWER;
+			}
+		}
+	} else {
+		if (new_clocks->zstate_support == DCN_ZSTATE_SUPPORT_DISALLOW &&
+				new_clocks->zstate_support != clk_mgr_base->clks.zstate_support) {
+			dcn314_smu_set_zstate_support(clk_mgr, DCN_ZSTATE_SUPPORT_DISALLOW);
+			dm_helpers_enable_periodic_detection(clk_mgr_base->ctx, false);
+			clk_mgr_base->clks.zstate_support = new_clocks->zstate_support;
+		}
+
+		if (!clk_mgr_base->clks.dtbclk_en && new_clocks->dtbclk_en) {
+			dcn314_smu_set_dtbclk(clk_mgr, true);
+			clk_mgr_base->clks.dtbclk_en = new_clocks->dtbclk_en;
+		}
+
+		/* check that we're not already in D0 */
+		if (clk_mgr_base->clks.pwr_state != DCN_PWR_STATE_MISSION_MODE) {
+			union display_idle_optimization_u idle_info = { 0 };
+
+			dcn314_smu_set_display_idle_optimization(clk_mgr, idle_info.data);
+			/* update power state */
+			clk_mgr_base->clks.pwr_state = DCN_PWR_STATE_MISSION_MODE;
+		}
+	}
+
+	if (should_set_clock(safe_to_lower, new_clocks->dcfclk_khz, clk_mgr_base->clks.dcfclk_khz)) {
+		clk_mgr_base->clks.dcfclk_khz = new_clocks->dcfclk_khz;
+		dcn314_smu_set_hard_min_dcfclk(clk_mgr, clk_mgr_base->clks.dcfclk_khz);
+	}
+
+	if (should_set_clock(safe_to_lower,
+			new_clocks->dcfclk_deep_sleep_khz, clk_mgr_base->clks.dcfclk_deep_sleep_khz)) {
+		clk_mgr_base->clks.dcfclk_deep_sleep_khz = new_clocks->dcfclk_deep_sleep_khz;
+		dcn314_smu_set_min_deep_sleep_dcfclk(clk_mgr, clk_mgr_base->clks.dcfclk_deep_sleep_khz);
+	}
+
+	// workaround: Limit dppclk to 100Mhz to avoid lower eDP panel switch to plus 4K monitor underflow.
+	if (new_clocks->dppclk_khz < 100000)
+		new_clocks->dppclk_khz = 100000;
+
+	if (should_set_clock(safe_to_lower, new_clocks->dppclk_khz, clk_mgr->base.clks.dppclk_khz)) {
+		if (clk_mgr->base.clks.dppclk_khz > new_clocks->dppclk_khz)
+			dpp_clock_lowered = true;
+		clk_mgr_base->clks.dppclk_khz = new_clocks->dppclk_khz;
+		update_dppclk = true;
+	}
+
+	if (should_set_clock(safe_to_lower, new_clocks->dispclk_khz, clk_mgr_base->clks.dispclk_khz) &&
+	    (new_clocks->dispclk_khz > 0 || (safe_to_lower && display_count == 0))) {
+		int requested_dispclk_khz = new_clocks->dispclk_khz;
+
+		dcn314_disable_otg_wa(clk_mgr_base, context, safe_to_lower, true);
+
+		/* Clamp the requested clock to PMFW based on their limit. */
+		if (dc->debug.min_disp_clk_khz > 0 && requested_dispclk_khz < dc->debug.min_disp_clk_khz)
+			requested_dispclk_khz = dc->debug.min_disp_clk_khz;
+
+		dcn314_smu_set_dispclk(clk_mgr, requested_dispclk_khz);
+		clk_mgr_base->clks.dispclk_khz = new_clocks->dispclk_khz;
+
+		dcn314_disable_otg_wa(clk_mgr_base, context, safe_to_lower, false);
+
+		update_dispclk = true;
+	}
+
+	if (dpp_clock_lowered) {
+		// increase per DPP DTO before lowering global dppclk
+		dcn20_update_clocks_update_dpp_dto(clk_mgr, context, safe_to_lower);
+		dcn314_smu_set_dppclk(clk_mgr, clk_mgr_base->clks.dppclk_khz);
+	} else {
+		// increase global DPPCLK before lowering per DPP DTO
+		if (update_dppclk || update_dispclk)
+			dcn314_smu_set_dppclk(clk_mgr, clk_mgr_base->clks.dppclk_khz);
+		// always update dtos unless clock is lowered and not safe to lower
+		if (new_clocks->dppclk_khz >= dc->current_state->bw_ctx.bw.dcn.clk.dppclk_khz)
+			dcn20_update_clocks_update_dpp_dto(clk_mgr, context, safe_to_lower);
+	}
+
+	// notify DMCUB of latest clocks
+	memset(&cmd, 0, sizeof(cmd));
+	cmd.notify_clocks.header.type = DMUB_CMD__CLK_MGR;
+	cmd.notify_clocks.header.sub_type = DMUB_CMD__CLK_MGR_NOTIFY_CLOCKS;
+	cmd.notify_clocks.clocks.dcfclk_khz = clk_mgr_base->clks.dcfclk_khz;
+	cmd.notify_clocks.clocks.dcfclk_deep_sleep_khz =
+		clk_mgr_base->clks.dcfclk_deep_sleep_khz;
+	cmd.notify_clocks.clocks.dispclk_khz = clk_mgr_base->clks.dispclk_khz;
+	cmd.notify_clocks.clocks.dppclk_khz = clk_mgr_base->clks.dppclk_khz;
+
+	dc_wake_and_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
+}
+
+static int get_vco_frequency_from_reg(struct clk_mgr_internal *clk_mgr)
+{
+	/* get FbMult value */
+	struct fixed31_32 pll_req;
+	unsigned int fbmult_frac_val = 0;
+	unsigned int fbmult_int_val = 0;
+
+	/*
+	 * Register value of fbmult is in 8.16 format, we are converting to 314.32
+	 * to leverage the fix point operations available in driver
+	 */
+
+	REG_GET(CLK1_CLK_PLL_REQ, FbMult_frac, &fbmult_frac_val); /* 16 bit fractional part*/
+	REG_GET(CLK1_CLK_PLL_REQ, FbMult_int, &fbmult_int_val); /* 8 bit integer part */
+
+	pll_req = dc_fixpt_from_int(fbmult_int_val);
+
+	/*
+	 * since fractional part is only 16 bit in register definition but is 32 bit
+	 * in our fix point definiton, need to shift left by 16 to obtain correct value
+	 */
+	pll_req.value |= fbmult_frac_val << 16;
+
+	/* multiply by REFCLK period */
+	pll_req = dc_fixpt_mul_int(pll_req, clk_mgr->dfs_ref_freq_khz);
+
+	/* integer part is now VCO frequency in kHz */
+	return dc_fixpt_floor(pll_req);
+}
+
+static void dcn314_enable_pme_wa(struct clk_mgr *clk_mgr_base)
+{
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+
+	dcn314_smu_enable_pme_wa(clk_mgr);
+}
+
+bool dcn314_are_clock_states_equal(struct dc_clocks *a,
+		struct dc_clocks *b)
+{
+	if (a->dispclk_khz != b->dispclk_khz)
+		return false;
+	else if (a->dppclk_khz != b->dppclk_khz)
+		return false;
+	else if (a->dcfclk_khz != b->dcfclk_khz)
+		return false;
+	else if (a->dcfclk_deep_sleep_khz != b->dcfclk_deep_sleep_khz)
+		return false;
+	else if (a->zstate_support != b->zstate_support)
+		return false;
+	else if (a->dtbclk_en != b->dtbclk_en)
+		return false;
+
+	return true;
+}
+
+
+static void dcn314_dump_clk_registers_internal(struct dcn35_clk_internal *internal, struct clk_mgr *clk_mgr_base)
+{
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+
+	// read dtbclk
+	internal->CLK1_CLK4_CURRENT_CNT = REG_READ(CLK1_CLK4_CURRENT_CNT);
+	internal->CLK1_CLK4_BYPASS_CNTL = REG_READ(CLK1_CLK4_BYPASS_CNTL);
+
+	// read dcfclk
+	internal->CLK1_CLK3_CURRENT_CNT = REG_READ(CLK1_CLK3_CURRENT_CNT);
+	internal->CLK1_CLK3_BYPASS_CNTL = REG_READ(CLK1_CLK3_BYPASS_CNTL);
+
+	// read dcf deep sleep divider
+	internal->CLK1_CLK3_DS_CNTL = REG_READ(CLK1_CLK3_DS_CNTL);
+	internal->CLK1_CLK3_ALLOW_DS = REG_READ(CLK1_CLK3_ALLOW_DS);
+
+	// read dppclk
+	internal->CLK1_CLK1_CURRENT_CNT = REG_READ(CLK1_CLK1_CURRENT_CNT);
+	internal->CLK1_CLK1_BYPASS_CNTL = REG_READ(CLK1_CLK1_BYPASS_CNTL);
+
+	// read dprefclk
+	internal->CLK1_CLK2_CURRENT_CNT = REG_READ(CLK1_CLK2_CURRENT_CNT);
+	internal->CLK1_CLK2_BYPASS_CNTL = REG_READ(CLK1_CLK2_BYPASS_CNTL);
+
+	// read dispclk
+	internal->CLK1_CLK0_CURRENT_CNT = REG_READ(CLK1_CLK0_CURRENT_CNT);
+	internal->CLK1_CLK0_BYPASS_CNTL = REG_READ(CLK1_CLK0_BYPASS_CNTL);
+}
+
+void dcn314_dump_clk_registers(struct clk_state_registers_and_bypass *regs_and_bypass,
+		struct clk_mgr *clk_mgr_base, struct clk_log_info *log_info)
+{
+
+	struct dcn35_clk_internal internal = {0};
+
+	dcn314_dump_clk_registers_internal(&internal, clk_mgr_base);
+
+	regs_and_bypass->dcfclk = internal.CLK1_CLK3_CURRENT_CNT / 10;
+	regs_and_bypass->dcf_deep_sleep_divider = internal.CLK1_CLK3_DS_CNTL / 10;
+	regs_and_bypass->dcf_deep_sleep_allow = internal.CLK1_CLK3_ALLOW_DS;
+	regs_and_bypass->dprefclk = internal.CLK1_CLK2_CURRENT_CNT / 10;
+	regs_and_bypass->dispclk = internal.CLK1_CLK0_CURRENT_CNT / 10;
+	regs_and_bypass->dppclk = internal.CLK1_CLK1_CURRENT_CNT / 10;
+	regs_and_bypass->dtbclk = internal.CLK1_CLK4_CURRENT_CNT / 10;
+
+	regs_and_bypass->dppclk_bypass = internal.CLK1_CLK1_BYPASS_CNTL & 0x0007;
+	if (regs_and_bypass->dppclk_bypass > 4)
+		regs_and_bypass->dppclk_bypass = 0;
+	regs_and_bypass->dcfclk_bypass = internal.CLK1_CLK3_BYPASS_CNTL & 0x0007;
+	if (regs_and_bypass->dcfclk_bypass > 4)
+		regs_and_bypass->dcfclk_bypass = 0;
+	regs_and_bypass->dispclk_bypass = internal.CLK1_CLK0_BYPASS_CNTL & 0x0007;
+	if (regs_and_bypass->dispclk_bypass > 4)
+		regs_and_bypass->dispclk_bypass = 0;
+	regs_and_bypass->dprefclk_bypass = internal.CLK1_CLK2_BYPASS_CNTL & 0x0007;
+	if (regs_and_bypass->dprefclk_bypass > 4)
+		regs_and_bypass->dprefclk_bypass = 0;
+
+}
+
+static struct clk_bw_params dcn314_bw_params = {
+	.vram_type = Ddr4MemType,
+	.num_channels = 1,
+	.clk_table = {
+		.num_entries = 4,
+	},
+
+};
+
+static struct wm_table ddr5_wm_table = {
+	.entries = {
+		{
+			.wm_inst = WM_A,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.72,
+			.sr_exit_time_us = 12.5,
+			.sr_enter_plus_exit_time_us = 14.5,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_B,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.72,
+			.sr_exit_time_us = 12.5,
+			.sr_enter_plus_exit_time_us = 14.5,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_C,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.72,
+			.sr_exit_time_us = 12.5,
+			.sr_enter_plus_exit_time_us = 14.5,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_D,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.72,
+			.sr_exit_time_us = 12.5,
+			.sr_enter_plus_exit_time_us = 14.5,
+			.valid = true,
+		},
+	}
+};
+
+static struct wm_table lpddr5_wm_table = {
+	.entries = {
+		{
+			.wm_inst = WM_A,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.65333,
+			.sr_exit_time_us = 30.0,
+			.sr_enter_plus_exit_time_us = 32.0,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_B,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.65333,
+			.sr_exit_time_us = 30.0,
+			.sr_enter_plus_exit_time_us = 32.0,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_C,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.65333,
+			.sr_exit_time_us = 30.0,
+			.sr_enter_plus_exit_time_us = 32.0,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_D,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.65333,
+			.sr_exit_time_us = 30.0,
+			.sr_enter_plus_exit_time_us = 32.0,
+			.valid = true,
+		},
+	}
+};
+
+static DpmClocks314_t dummy_clocks;
+
+static struct dcn314_watermarks dummy_wms = { 0 };
+
+static struct dcn314_ss_info_table ss_info_table = {
+	.ss_divider = 1000,
+	.ss_percentage = {0, 0, 375, 375, 375}
+};
+
+static void dcn314_build_watermark_ranges(struct clk_bw_params *bw_params, struct dcn314_watermarks *table)
+{
+	int i, num_valid_sets;
+
+	num_valid_sets = 0;
+
+	for (i = 0; i < WM_SET_COUNT; i++) {
+		/* skip empty entries, the smu array has no holes*/
+		if (!bw_params->wm_table.entries[i].valid)
+			continue;
+
+		table->WatermarkRow[WM_DCFCLK][num_valid_sets].WmSetting = bw_params->wm_table.entries[i].wm_inst;
+		table->WatermarkRow[WM_DCFCLK][num_valid_sets].WmType = bw_params->wm_table.entries[i].wm_type;
+		/* We will not select WM based on fclk, so leave it as unconstrained */
+		table->WatermarkRow[WM_DCFCLK][num_valid_sets].MinClock = 0;
+		table->WatermarkRow[WM_DCFCLK][num_valid_sets].MaxClock = 0xFFFF;
+
+		if (table->WatermarkRow[WM_DCFCLK][num_valid_sets].WmType == WM_TYPE_PSTATE_CHG) {
+			if (i == 0)
+				table->WatermarkRow[WM_DCFCLK][num_valid_sets].MinMclk = 0;
+			else {
+				/* add 1 to make it non-overlapping with next lvl */
+				table->WatermarkRow[WM_DCFCLK][num_valid_sets].MinMclk =
+						bw_params->clk_table.entries[i - 1].dcfclk_mhz + 1;
+			}
+			table->WatermarkRow[WM_DCFCLK][num_valid_sets].MaxMclk =
+					bw_params->clk_table.entries[i].dcfclk_mhz;
+
+		} else {
+			/* unconstrained for memory retraining */
+			table->WatermarkRow[WM_DCFCLK][num_valid_sets].MinClock = 0;
+			table->WatermarkRow[WM_DCFCLK][num_valid_sets].MaxClock = 0xFFFF;
+
+			/* Modify previous watermark range to cover up to max */
+			table->WatermarkRow[WM_DCFCLK][num_valid_sets - 1].MaxClock = 0xFFFF;
+		}
+		num_valid_sets++;
+	}
+
+	ASSERT(num_valid_sets != 0); /* Must have at least one set of valid watermarks */
+
+	/* modify the min and max to make sure we cover the whole range*/
+	table->WatermarkRow[WM_DCFCLK][0].MinMclk = 0;
+	table->WatermarkRow[WM_DCFCLK][0].MinClock = 0;
+	table->WatermarkRow[WM_DCFCLK][num_valid_sets - 1].MaxMclk = 0xFFFF;
+	table->WatermarkRow[WM_DCFCLK][num_valid_sets - 1].MaxClock = 0xFFFF;
+
+	/* This is for writeback only, does not matter currently as no writeback support*/
+	table->WatermarkRow[WM_SOCCLK][0].WmSetting = WM_A;
+	table->WatermarkRow[WM_SOCCLK][0].MinClock = 0;
+	table->WatermarkRow[WM_SOCCLK][0].MaxClock = 0xFFFF;
+	table->WatermarkRow[WM_SOCCLK][0].MinMclk = 0;
+	table->WatermarkRow[WM_SOCCLK][0].MaxMclk = 0xFFFF;
+}
+
+static void dcn314_notify_wm_ranges(struct clk_mgr *clk_mgr_base)
+{
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+	struct clk_mgr_dcn314 *clk_mgr_dcn314 = TO_CLK_MGR_DCN314(clk_mgr);
+	struct dcn314_watermarks *table = clk_mgr_dcn314->smu_wm_set.wm_set;
+
+	if (!clk_mgr->smu_ver)
+		return;
+
+	if (!table || clk_mgr_dcn314->smu_wm_set.mc_address.quad_part == 0)
+		return;
+
+	memset(table, 0, sizeof(*table));
+
+	dcn314_build_watermark_ranges(clk_mgr_base->bw_params, table);
+
+	dcn314_smu_set_dram_addr_high(clk_mgr,
+			clk_mgr_dcn314->smu_wm_set.mc_address.high_part);
+	dcn314_smu_set_dram_addr_low(clk_mgr,
+			clk_mgr_dcn314->smu_wm_set.mc_address.low_part);
+	dcn314_smu_transfer_wm_table_dram_2_smu(clk_mgr);
+}
+
+static void dcn314_get_dpm_table_from_smu(struct clk_mgr_internal *clk_mgr,
+		struct dcn314_smu_dpm_clks *smu_dpm_clks)
+{
+	DpmClocks314_t *table = smu_dpm_clks->dpm_clks;
+
+	if (!clk_mgr->smu_ver)
+		return;
+
+	if (!table || smu_dpm_clks->mc_address.quad_part == 0)
+		return;
+
+	memset(table, 0, sizeof(*table));
+
+	dcn314_smu_set_dram_addr_high(clk_mgr,
+			smu_dpm_clks->mc_address.high_part);
+	dcn314_smu_set_dram_addr_low(clk_mgr,
+			smu_dpm_clks->mc_address.low_part);
+	dcn314_smu_transfer_dpm_table_smu_2_dram(clk_mgr);
+}
+
+static inline bool is_valid_clock_value(uint32_t clock_value)
+{
+	return clock_value > 1 && clock_value < 100000;
+}
+
+static unsigned int convert_wck_ratio(uint8_t wck_ratio)
+{
+	switch (wck_ratio) {
+	case WCK_RATIO_1_2:
+		return 2;
+
+	case WCK_RATIO_1_4:
+		return 4;
+
+	default:
+		break;
+	}
+	return 1;
+}
+
+static uint32_t find_max_clk_value(const uint32_t clocks[], uint32_t num_clocks)
+{
+	uint32_t max = 0;
+	int i;
+
+	for (i = 0; i < num_clocks; ++i) {
+		if (clocks[i] > max)
+			max = clocks[i];
+	}
+
+	return max;
+}
+
+static void dcn314_clk_mgr_helper_populate_bw_params(struct clk_mgr_internal *clk_mgr,
+						    struct integrated_info *bios_info,
+						    const DpmClocks314_t *clock_table)
+{
+	struct clk_bw_params *bw_params = clk_mgr->base.bw_params;
+	struct clk_limit_table_entry def_max = bw_params->clk_table.entries[bw_params->clk_table.num_entries - 1];
+	uint32_t max_pstate = 0,  max_fclk = 0,  min_pstate = 0, max_dispclk = 0, max_dppclk = 0;
+	int i;
+
+	/* Find highest valid fclk pstate */
+	for (i = 0; i < clock_table->NumDfPstatesEnabled; i++) {
+		if (is_valid_clock_value(clock_table->DfPstateTable[i].FClk) &&
+		    clock_table->DfPstateTable[i].FClk > max_fclk) {
+			max_fclk = clock_table->DfPstateTable[i].FClk;
+			max_pstate = i;
+		}
+	}
+
+	/* We expect the table to contain at least one valid fclk entry. */
+	ASSERT(is_valid_clock_value(max_fclk));
+
+	/* Dispclk and dppclk can be max at any voltage, same number of levels for both */
+	if (clock_table->NumDispClkLevelsEnabled <= NUM_DISPCLK_DPM_LEVELS &&
+	    clock_table->NumDispClkLevelsEnabled <= NUM_DPPCLK_DPM_LEVELS) {
+		max_dispclk = find_max_clk_value(clock_table->DispClocks, clock_table->NumDispClkLevelsEnabled);
+		max_dppclk = find_max_clk_value(clock_table->DppClocks, clock_table->NumDispClkLevelsEnabled);
+	} else {
+		/* Invalid number of entries in the table from PMFW. */
+		ASSERT(0);
+	}
+
+	/* Base the clock table on dcfclk, need at least one entry regardless of pmfw table */
+	for (i = 0; i < clock_table->NumDcfClkLevelsEnabled; i++) {
+		uint32_t min_fclk = clock_table->DfPstateTable[0].FClk;
+		int j;
+
+		for (j = 1; j < clock_table->NumDfPstatesEnabled; j++) {
+			if (is_valid_clock_value(clock_table->DfPstateTable[j].FClk) &&
+			    clock_table->DfPstateTable[j].FClk < min_fclk &&
+			    clock_table->DfPstateTable[j].Voltage <= clock_table->SocVoltage[i]) {
+				min_fclk = clock_table->DfPstateTable[j].FClk;
+				min_pstate = j;
+			}
+		}
+
+		/* First search defaults for the clocks we don't read using closest lower or equal default dcfclk */
+		for (j = bw_params->clk_table.num_entries - 1; j > 0; j--)
+			if (bw_params->clk_table.entries[j].dcfclk_mhz <= clock_table->DcfClocks[i])
+				break;
+
+		bw_params->clk_table.entries[i].phyclk_mhz = bw_params->clk_table.entries[j].phyclk_mhz;
+		bw_params->clk_table.entries[i].phyclk_d18_mhz = bw_params->clk_table.entries[j].phyclk_d18_mhz;
+		bw_params->clk_table.entries[i].dtbclk_mhz = bw_params->clk_table.entries[j].dtbclk_mhz;
+
+		/* Now update clocks we do read */
+		bw_params->clk_table.entries[i].fclk_mhz = min_fclk;
+		bw_params->clk_table.entries[i].memclk_mhz = clock_table->DfPstateTable[min_pstate].MemClk;
+		bw_params->clk_table.entries[i].voltage = clock_table->DfPstateTable[min_pstate].Voltage;
+		bw_params->clk_table.entries[i].dcfclk_mhz = clock_table->DcfClocks[i];
+		bw_params->clk_table.entries[i].socclk_mhz = clock_table->SocClocks[i];
+		bw_params->clk_table.entries[i].dispclk_mhz = max_dispclk;
+		bw_params->clk_table.entries[i].dppclk_mhz = max_dppclk;
+		bw_params->clk_table.entries[i].wck_ratio = convert_wck_ratio(
+			clock_table->DfPstateTable[min_pstate].WckRatio);
+	}
+
+	/* Make sure to include at least one entry at highest pstate */
+	if (max_pstate != min_pstate || i == 0) {
+		if (i > MAX_NUM_DPM_LVL - 1)
+			i = MAX_NUM_DPM_LVL - 1;
+
+		bw_params->clk_table.entries[i].fclk_mhz = max_fclk;
+		bw_params->clk_table.entries[i].memclk_mhz = clock_table->DfPstateTable[max_pstate].MemClk;
+		bw_params->clk_table.entries[i].voltage = clock_table->DfPstateTable[max_pstate].Voltage;
+		bw_params->clk_table.entries[i].dcfclk_mhz = find_max_clk_value(clock_table->DcfClocks, NUM_DCFCLK_DPM_LEVELS);
+		bw_params->clk_table.entries[i].socclk_mhz = find_max_clk_value(clock_table->SocClocks, NUM_SOCCLK_DPM_LEVELS);
+		bw_params->clk_table.entries[i].dispclk_mhz = max_dispclk;
+		bw_params->clk_table.entries[i].dppclk_mhz = max_dppclk;
+		bw_params->clk_table.entries[i].wck_ratio = convert_wck_ratio(
+			clock_table->DfPstateTable[max_pstate].WckRatio);
+		i++;
+	}
+	bw_params->clk_table.num_entries = i--;
+
+	/* Make sure all highest clocks are included*/
+	bw_params->clk_table.entries[i].socclk_mhz = find_max_clk_value(clock_table->SocClocks, NUM_SOCCLK_DPM_LEVELS);
+	bw_params->clk_table.entries[i].dispclk_mhz = find_max_clk_value(clock_table->DispClocks, NUM_DISPCLK_DPM_LEVELS);
+	bw_params->clk_table.entries[i].dppclk_mhz = find_max_clk_value(clock_table->DppClocks, NUM_DPPCLK_DPM_LEVELS);
+	ASSERT(clock_table->DcfClocks[i] == find_max_clk_value(clock_table->DcfClocks, NUM_DCFCLK_DPM_LEVELS));
+	bw_params->clk_table.entries[i].phyclk_mhz = def_max.phyclk_mhz;
+	bw_params->clk_table.entries[i].phyclk_d18_mhz = def_max.phyclk_d18_mhz;
+	bw_params->clk_table.entries[i].dtbclk_mhz = def_max.dtbclk_mhz;
+
+	/*
+	 * Set any 0 clocks to max default setting. Not an issue for
+	 * power since we aren't doing switching in such case anyway
+	 */
+	for (i = 0; i < bw_params->clk_table.num_entries; i++) {
+		if (!bw_params->clk_table.entries[i].fclk_mhz) {
+			bw_params->clk_table.entries[i].fclk_mhz = def_max.fclk_mhz;
+			bw_params->clk_table.entries[i].memclk_mhz = def_max.memclk_mhz;
+			bw_params->clk_table.entries[i].voltage = def_max.voltage;
+		}
+		if (!bw_params->clk_table.entries[i].dcfclk_mhz)
+			bw_params->clk_table.entries[i].dcfclk_mhz = def_max.dcfclk_mhz;
+		if (!bw_params->clk_table.entries[i].socclk_mhz)
+			bw_params->clk_table.entries[i].socclk_mhz = def_max.socclk_mhz;
+		if (!bw_params->clk_table.entries[i].dispclk_mhz)
+			bw_params->clk_table.entries[i].dispclk_mhz = def_max.dispclk_mhz;
+		if (!bw_params->clk_table.entries[i].dppclk_mhz)
+			bw_params->clk_table.entries[i].dppclk_mhz = def_max.dppclk_mhz;
+		if (!bw_params->clk_table.entries[i].phyclk_mhz)
+			bw_params->clk_table.entries[i].phyclk_mhz = def_max.phyclk_mhz;
+		if (!bw_params->clk_table.entries[i].phyclk_d18_mhz)
+			bw_params->clk_table.entries[i].phyclk_d18_mhz = def_max.phyclk_d18_mhz;
+		if (!bw_params->clk_table.entries[i].dtbclk_mhz)
+			bw_params->clk_table.entries[i].dtbclk_mhz = def_max.dtbclk_mhz;
+	}
+	ASSERT(bw_params->clk_table.entries[i-1].dcfclk_mhz);
+	bw_params->vram_type = bios_info->memory_type;
+
+	bw_params->dram_channel_width_bytes = bios_info->memory_type == 0x22 ? 8 : 4;
+	bw_params->num_channels = bios_info->ma_channel_number ? bios_info->ma_channel_number : 4;
+
+	for (i = 0; i < WM_SET_COUNT; i++) {
+		bw_params->wm_table.entries[i].wm_inst = i;
+
+		if (i >= bw_params->clk_table.num_entries) {
+			bw_params->wm_table.entries[i].valid = false;
+			continue;
+		}
+
+		bw_params->wm_table.entries[i].wm_type = WM_TYPE_PSTATE_CHG;
+		bw_params->wm_table.entries[i].valid = true;
+	}
+}
+
+static struct clk_mgr_funcs dcn314_funcs = {
+	.get_dp_ref_clk_frequency = dce12_get_dp_ref_freq_khz,
+	.get_dtb_ref_clk_frequency = dcn31_get_dtb_ref_freq_khz,
+	.update_clocks = dcn314_update_clocks,
+	.init_clocks = dcn314_init_clocks,
+	.enable_pme_wa = dcn314_enable_pme_wa,
+	.are_clock_states_equal = dcn314_are_clock_states_equal,
+	.notify_wm_ranges = dcn314_notify_wm_ranges
+};
+extern struct clk_mgr_funcs dcn3_fpga_funcs;
+
+static void dcn314_read_ss_info_from_lut(struct clk_mgr_internal *clk_mgr)
+{
+	uint32_t clock_source;
+	//uint32_t ssc_enable;
+
+	REG_GET(CLK1_CLK2_BYPASS_CNTL, CLK2_BYPASS_SEL, &clock_source);
+	//REG_GET(CLK6_0_CLK6_spll_field_8, spll_ssc_en, &ssc_enable);
+
+	if (dcn314_is_spll_ssc_enabled(&clk_mgr->base) && (clock_source < ARRAY_SIZE(ss_info_table.ss_percentage))) {
+		clk_mgr->dprefclk_ss_percentage = ss_info_table.ss_percentage[clock_source];
+
+		if (clk_mgr->dprefclk_ss_percentage != 0) {
+			clk_mgr->ss_on_dprefclk = true;
+			clk_mgr->dprefclk_ss_divider = ss_info_table.ss_divider;
+		}
+	}
+}
+
+void dcn314_clk_mgr_construct(
+		struct dc_context *ctx,
+		struct clk_mgr_dcn314 *clk_mgr,
+		struct pp_smu_funcs *pp_smu,
+		struct dccg *dccg)
+{
+	struct dcn314_smu_dpm_clks smu_dpm_clks = { 0 };
+	struct clk_log_info log_info = {0};
+
+	clk_mgr->base.base.ctx = ctx;
+	clk_mgr->base.base.funcs = &dcn314_funcs;
+
+	clk_mgr->base.pp_smu = pp_smu;
+
+	clk_mgr->base.dccg = dccg;
+	clk_mgr->base.dfs_bypass_disp_clk = 0;
+
+	clk_mgr->base.dprefclk_ss_percentage = 0;
+	clk_mgr->base.dprefclk_ss_divider = 1000;
+	clk_mgr->base.ss_on_dprefclk = false;
+	clk_mgr->base.dfs_ref_freq_khz = 48000;
+
+	clk_mgr->smu_wm_set.wm_set = (struct dcn314_watermarks *)dm_helpers_allocate_gpu_mem(
+				clk_mgr->base.base.ctx,
+				DC_MEM_ALLOC_TYPE_FRAME_BUFFER,
+				sizeof(struct dcn314_watermarks),
+				&clk_mgr->smu_wm_set.mc_address.quad_part);
+
+	if (!clk_mgr->smu_wm_set.wm_set) {
+		clk_mgr->smu_wm_set.wm_set = &dummy_wms;
+		clk_mgr->smu_wm_set.mc_address.quad_part = 0;
+	}
+	ASSERT(clk_mgr->smu_wm_set.wm_set);
+
+	smu_dpm_clks.dpm_clks = (DpmClocks314_t *)dm_helpers_allocate_gpu_mem(
+				clk_mgr->base.base.ctx,
+				DC_MEM_ALLOC_TYPE_FRAME_BUFFER,
+				sizeof(DpmClocks314_t),
+				&smu_dpm_clks.mc_address.quad_part);
+
+	if (smu_dpm_clks.dpm_clks == NULL) {
+		smu_dpm_clks.dpm_clks = &dummy_clocks;
+		smu_dpm_clks.mc_address.quad_part = 0;
+	}
+
+	ASSERT(smu_dpm_clks.dpm_clks);
+
+	clk_mgr->base.smu_ver = dcn314_smu_get_smu_version(&clk_mgr->base);
+
+	if (clk_mgr->base.smu_ver)
+		clk_mgr->base.smu_present = true;
+
+	/* TODO: Check we get what we expect during bringup */
+	clk_mgr->base.base.dentist_vco_freq_khz = get_vco_frequency_from_reg(&clk_mgr->base);
+
+	if (ctx->dc_bios->integrated_info->memory_type == LpDdr5MemType)
+		dcn314_bw_params.wm_table = lpddr5_wm_table;
+	else
+		dcn314_bw_params.wm_table = ddr5_wm_table;
+
+	/* Saved clocks configured at boot for debug purposes */
+	dcn314_dump_clk_registers(&clk_mgr->base.base.boot_snapshot,
+				  &clk_mgr->base.base, &log_info);
+
+	clk_mgr->base.base.dprefclk_khz = 600000;
+	clk_mgr->base.base.clks.ref_dtbclk_khz = 600000;
+	dce_clock_read_ss_info(&clk_mgr->base);
+	dcn314_read_ss_info_from_lut(&clk_mgr->base);
+	/*if bios enabled SS, driver needs to adjust dtb clock, only enable with correct bios*/
+
+	clk_mgr->base.base.bw_params = &dcn314_bw_params;
+
+	if (clk_mgr->base.base.ctx->dc->debug.pstate_enabled) {
+		int i;
+
+		dcn314_get_dpm_table_from_smu(&clk_mgr->base, &smu_dpm_clks);
+		DC_LOG_SMU("NumDcfClkLevelsEnabled: %d\n"
+				   "NumDispClkLevelsEnabled: %d\n"
+				   "NumSocClkLevelsEnabled: %d\n"
+				   "VcnClkLevelsEnabled: %d\n"
+				   "NumDfPst atesEnabled: %d\n"
+				   "MinGfxClk: %d\n"
+				   "MaxGfxClk: %d\n",
+				   smu_dpm_clks.dpm_clks->NumDcfClkLevelsEnabled,
+				   smu_dpm_clks.dpm_clks->NumDispClkLevelsEnabled,
+				   smu_dpm_clks.dpm_clks->NumSocClkLevelsEnabled,
+				   smu_dpm_clks.dpm_clks->VcnClkLevelsEnabled,
+				   smu_dpm_clks.dpm_clks->NumDfPstatesEnabled,
+				   smu_dpm_clks.dpm_clks->MinGfxClk,
+				   smu_dpm_clks.dpm_clks->MaxGfxClk);
+		for (i = 0; i < smu_dpm_clks.dpm_clks->NumDcfClkLevelsEnabled; i++) {
+			DC_LOG_SMU("smu_dpm_clks.dpm_clks->DcfClocks[%d] = %d\n",
+					   i,
+					   smu_dpm_clks.dpm_clks->DcfClocks[i]);
+		}
+		for (i = 0; i < smu_dpm_clks.dpm_clks->NumDispClkLevelsEnabled; i++) {
+			DC_LOG_SMU("smu_dpm_clks.dpm_clks->DispClocks[%d] = %d\n",
+					   i, smu_dpm_clks.dpm_clks->DispClocks[i]);
+		}
+		for (i = 0; i < smu_dpm_clks.dpm_clks->NumSocClkLevelsEnabled; i++) {
+			DC_LOG_SMU("smu_dpm_clks.dpm_clks->SocClocks[%d] = %d\n",
+					   i, smu_dpm_clks.dpm_clks->SocClocks[i]);
+		}
+		for (i = 0; i < NUM_SOC_VOLTAGE_LEVELS; i++)
+			DC_LOG_SMU("smu_dpm_clks.dpm_clks->SocVoltage[%d] = %d\n",
+					   i, smu_dpm_clks.dpm_clks->SocVoltage[i]);
+
+		for (i = 0; i < NUM_DF_PSTATE_LEVELS; i++) {
+			DC_LOG_SMU("smu_dpm_clks.dpm_clks.DfPstateTable[%d].FClk = %d\n"
+					   "smu_dpm_clks.dpm_clks->DfPstateTable[%d].MemClk= %d\n"
+					   "smu_dpm_clks.dpm_clks->DfPstateTable[%d].Voltage = %d\n",
+					   i, smu_dpm_clks.dpm_clks->DfPstateTable[i].FClk,
+					   i, smu_dpm_clks.dpm_clks->DfPstateTable[i].MemClk,
+					   i, smu_dpm_clks.dpm_clks->DfPstateTable[i].Voltage);
+		}
+
+		if (ctx->dc_bios->integrated_info && ctx->dc->config.use_default_clock_table == false) {
+			dcn314_clk_mgr_helper_populate_bw_params(
+					&clk_mgr->base,
+					ctx->dc_bios->integrated_info,
+					smu_dpm_clks.dpm_clks);
+		}
+	}
+
+	if (smu_dpm_clks.dpm_clks && smu_dpm_clks.mc_address.quad_part != 0)
+		dm_helpers_free_gpu_mem(clk_mgr->base.base.ctx, DC_MEM_ALLOC_TYPE_FRAME_BUFFER,
+				smu_dpm_clks.dpm_clks);
+}
+
+void dcn314_clk_mgr_destroy(struct clk_mgr_internal *clk_mgr_int)
+{
+	struct clk_mgr_dcn314 *clk_mgr = TO_CLK_MGR_DCN314(clk_mgr_int);
+
+	if (clk_mgr->smu_wm_set.wm_set && clk_mgr->smu_wm_set.mc_address.quad_part != 0)
+		dm_helpers_free_gpu_mem(clk_mgr_int->base.ctx, DC_MEM_ALLOC_TYPE_FRAME_BUFFER,
+				clk_mgr->smu_wm_set.wm_set);
+}
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn314/dcn314_clk_mgr.h b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn314/dcn314_clk_mgr.h
new file mode 100644
index 000000000000..0577eb527bc3
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn314/dcn314_clk_mgr.h
@@ -0,0 +1,73 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __DCN314_CLK_MGR_H__
+#define __DCN314_CLK_MGR_H__
+#include "clk_mgr_internal.h"
+
+#define DCN314_NUM_CLOCK_SOURCES   5
+
+struct dcn314_watermarks;
+
+struct dcn314_smu_watermark_set {
+	struct dcn314_watermarks *wm_set;
+	union large_integer mc_address;
+};
+
+struct clk_mgr_dcn314 {
+	struct clk_mgr_internal base;
+	struct dcn314_smu_watermark_set smu_wm_set;
+};
+
+struct dcn314_ss_info_table {
+	uint32_t ss_divider;
+	uint32_t ss_percentage[DCN314_NUM_CLOCK_SOURCES];
+};
+
+bool dcn314_are_clock_states_equal(struct dc_clocks *a,
+		struct dc_clocks *b);
+
+bool dcn314_is_spll_ssc_enabled(struct clk_mgr *clk_mgr_base);
+
+void dcn314_init_clocks(struct clk_mgr *clk_mgr);
+
+void dcn314_update_clocks(struct clk_mgr *clk_mgr_base,
+			struct dc_state *context,
+			bool safe_to_lower);
+
+void dcn314_clk_mgr_construct(struct dc_context *ctx,
+		struct clk_mgr_dcn314 *clk_mgr,
+		struct pp_smu_funcs *pp_smu,
+		struct dccg *dccg);
+
+void dcn314_clk_mgr_destroy(struct clk_mgr_internal *clk_mgr_int);
+
+
+void dcn314_dump_clk_registers(struct clk_state_registers_and_bypass *regs_and_bypass,
+		struct clk_mgr *clk_mgr_base, struct clk_log_info *log_info);
+
+
+#endif //__DCN314_CLK_MGR_H__
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn314/dcn314_smu.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn314/dcn314_smu.c
new file mode 100644
index 000000000000..c4af406146b7
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn314/dcn314_smu.c
@@ -0,0 +1,398 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+
+
+#include "core_types.h"
+#include "clk_mgr_internal.h"
+#include "reg_helper.h"
+#include "dm_helpers.h"
+#include "dcn314_smu.h"
+
+#include "mp/mp_13_0_5_offset.h"
+
+/* TODO: Use the real headers when they're correct */
+#define MP1_BASE__INST0_SEG0                       0x00016000
+#define MP1_BASE__INST0_SEG1                       0x0243FC00
+#define MP1_BASE__INST0_SEG2                       0x00DC0000
+#define MP1_BASE__INST0_SEG3                       0x00E00000
+#define MP1_BASE__INST0_SEG4                       0x00E40000
+#define MP1_BASE__INST0_SEG5                       0
+
+#ifdef BASE_INNER
+#undef BASE_INNER
+#endif
+
+#define BASE_INNER(seg) MP1_BASE__INST0_SEG ## seg
+
+#define BASE(seg) BASE_INNER(seg)
+
+#define REG(reg_name) (BASE(reg##reg_name##_BASE_IDX) + reg##reg_name)
+
+#define FN(reg_name, field) \
+	FD(reg_name##__##field)
+
+#include "logger_types.h"
+#undef DC_LOGGER
+#define DC_LOGGER \
+	CTX->logger
+#define smu_print(str, ...) {DC_LOG_SMU(str, ##__VA_ARGS__); }
+
+#define VBIOSSMC_MSG_TestMessage                  0x1
+#define VBIOSSMC_MSG_GetSmuVersion                0x2
+#define VBIOSSMC_MSG_PowerUpGfx                   0x3
+#define VBIOSSMC_MSG_SetDispclkFreq               0x4
+#define VBIOSSMC_MSG_SetDprefclkFreq              0x5   //Not used. DPRef is constant
+#define VBIOSSMC_MSG_SetDppclkFreq                0x6
+#define VBIOSSMC_MSG_SetHardMinDcfclkByFreq       0x7
+#define VBIOSSMC_MSG_SetMinDeepSleepDcfclk        0x8
+#define VBIOSSMC_MSG_SetPhyclkVoltageByFreq       0x9	//Keep it in case VMIN dees not support phy clk
+#define VBIOSSMC_MSG_GetFclkFrequency             0xA
+#define VBIOSSMC_MSG_SetDisplayCount              0xB   //Not used anymore
+#define VBIOSSMC_MSG_EnableTmdp48MHzRefclkPwrDown 0xC   //Not used anymore
+#define VBIOSSMC_MSG_UpdatePmeRestore             0xD
+#define VBIOSSMC_MSG_SetVbiosDramAddrHigh         0xE   //Used for WM table txfr
+#define VBIOSSMC_MSG_SetVbiosDramAddrLow          0xF
+#define VBIOSSMC_MSG_TransferTableSmu2Dram        0x10
+#define VBIOSSMC_MSG_TransferTableDram2Smu        0x11
+#define VBIOSSMC_MSG_SetDisplayIdleOptimizations  0x12
+#define VBIOSSMC_MSG_GetDprefclkFreq              0x13
+#define VBIOSSMC_MSG_GetDtbclkFreq                0x14
+#define VBIOSSMC_MSG_AllowZstatesEntry            0x15
+#define VBIOSSMC_MSG_DisallowZstatesEntry	  0x16
+#define VBIOSSMC_MSG_SetDtbClk                    0x17
+#define VBIOSSMC_Message_Count                    0x18
+
+#define VBIOSSMC_Status_BUSY                      0x0
+#define VBIOSSMC_Result_OK                        0x1
+#define VBIOSSMC_Result_Failed                    0xFF
+#define VBIOSSMC_Result_UnknownCmd                0xFE
+#define VBIOSSMC_Result_CmdRejectedPrereq         0xFD
+#define VBIOSSMC_Result_CmdRejectedBusy           0xFC
+
+/*
+ * Function to be used instead of REG_WAIT macro because the wait ends when
+ * the register is NOT EQUAL to zero, and because the translation in msg_if.h
+ * won't work with REG_WAIT.
+ */
+static uint32_t dcn314_smu_wait_for_response(struct clk_mgr_internal *clk_mgr, unsigned int delay_us, unsigned int max_retries)
+{
+	uint32_t res_val = VBIOSSMC_Status_BUSY;
+
+	do {
+		res_val = REG_READ(MP1_SMN_C2PMSG_91);
+		if (res_val != VBIOSSMC_Status_BUSY)
+			break;
+
+		if (delay_us >= 1000)
+			msleep(delay_us/1000);
+		else if (delay_us > 0)
+			udelay(delay_us);
+	} while (max_retries--);
+
+	return res_val;
+}
+
+static int dcn314_smu_send_msg_with_param(struct clk_mgr_internal *clk_mgr,
+					 unsigned int msg_id,
+					 unsigned int param)
+{
+	uint32_t result;
+
+	result = dcn314_smu_wait_for_response(clk_mgr, 10, 200000);
+
+	if (result != VBIOSSMC_Result_OK)
+		smu_print("SMU Response was not OK. SMU response after wait received is: %d\n",
+				result);
+
+	if (result == VBIOSSMC_Status_BUSY)
+		return -1;
+
+	/* First clear response register */
+	REG_WRITE(MP1_SMN_C2PMSG_91, VBIOSSMC_Status_BUSY);
+
+	/* Set the parameter register for the SMU message, unit is Mhz */
+	REG_WRITE(MP1_SMN_C2PMSG_83, param);
+
+	/* Trigger the message transaction by writing the message ID */
+	REG_WRITE(MP1_SMN_C2PMSG_67, msg_id);
+
+	result = dcn314_smu_wait_for_response(clk_mgr, 10, 200000);
+
+	if (result == VBIOSSMC_Result_Failed) {
+		if (msg_id == VBIOSSMC_MSG_TransferTableDram2Smu &&
+		    param == TABLE_WATERMARKS)
+			DC_LOG_DEBUG("Watermarks table not configured properly by SMU");
+		else if (msg_id == VBIOSSMC_MSG_SetHardMinDcfclkByFreq ||
+			 msg_id == VBIOSSMC_MSG_SetMinDeepSleepDcfclk)
+			DC_LOG_WARNING("DCFCLK_DPM is not enabled by BIOS");
+		else
+			ASSERT(0);
+		REG_WRITE(MP1_SMN_C2PMSG_91, VBIOSSMC_Result_OK);
+		return -1;
+	}
+
+	if (IS_SMU_TIMEOUT(result)) {
+		ASSERT(0);
+		dm_helpers_smu_timeout(CTX, msg_id, param, 10 * 200000);
+	}
+
+	return REG_READ(MP1_SMN_C2PMSG_83);
+}
+
+int dcn314_smu_get_smu_version(struct clk_mgr_internal *clk_mgr)
+{
+	return dcn314_smu_send_msg_with_param(
+			clk_mgr,
+			VBIOSSMC_MSG_GetSmuVersion,
+			0);
+}
+
+
+int dcn314_smu_set_dispclk(struct clk_mgr_internal *clk_mgr, int requested_dispclk_khz)
+{
+	int actual_dispclk_set_mhz = -1;
+
+	if (!clk_mgr->smu_present)
+		return requested_dispclk_khz;
+
+	/*  Unit of SMU msg parameter is Mhz */
+	actual_dispclk_set_mhz = dcn314_smu_send_msg_with_param(
+			clk_mgr,
+			VBIOSSMC_MSG_SetDispclkFreq,
+			khz_to_mhz_ceil(requested_dispclk_khz));
+
+	return actual_dispclk_set_mhz * 1000;
+}
+
+int dcn314_smu_set_dprefclk(struct clk_mgr_internal *clk_mgr)
+{
+	int actual_dprefclk_set_mhz = -1;
+
+	if (!clk_mgr->smu_present)
+		return clk_mgr->base.dprefclk_khz;
+
+	actual_dprefclk_set_mhz = dcn314_smu_send_msg_with_param(
+			clk_mgr,
+			VBIOSSMC_MSG_SetDprefclkFreq,
+			khz_to_mhz_ceil(clk_mgr->base.dprefclk_khz));
+
+	/* TODO: add code for programing DP DTO, currently this is down by command table */
+
+	return actual_dprefclk_set_mhz * 1000;
+}
+
+int dcn314_smu_set_hard_min_dcfclk(struct clk_mgr_internal *clk_mgr, int requested_dcfclk_khz)
+{
+	int actual_dcfclk_set_mhz = -1;
+
+	if (!clk_mgr->base.ctx->dc->debug.pstate_enabled)
+		return -1;
+
+	if (!clk_mgr->smu_present)
+		return requested_dcfclk_khz;
+
+	actual_dcfclk_set_mhz = dcn314_smu_send_msg_with_param(
+			clk_mgr,
+			VBIOSSMC_MSG_SetHardMinDcfclkByFreq,
+			khz_to_mhz_ceil(requested_dcfclk_khz));
+
+	return actual_dcfclk_set_mhz * 1000;
+}
+
+int dcn314_smu_set_min_deep_sleep_dcfclk(struct clk_mgr_internal *clk_mgr, int requested_min_ds_dcfclk_khz)
+{
+	int actual_min_ds_dcfclk_mhz = -1;
+
+	if (!clk_mgr->base.ctx->dc->debug.pstate_enabled)
+		return -1;
+
+	if (!clk_mgr->smu_present)
+		return requested_min_ds_dcfclk_khz;
+
+	actual_min_ds_dcfclk_mhz = dcn314_smu_send_msg_with_param(
+			clk_mgr,
+			VBIOSSMC_MSG_SetMinDeepSleepDcfclk,
+			khz_to_mhz_ceil(requested_min_ds_dcfclk_khz));
+
+	return actual_min_ds_dcfclk_mhz * 1000;
+}
+
+int dcn314_smu_set_dppclk(struct clk_mgr_internal *clk_mgr, int requested_dpp_khz)
+{
+	int actual_dppclk_set_mhz = -1;
+
+	if (!clk_mgr->smu_present)
+		return requested_dpp_khz;
+
+	actual_dppclk_set_mhz = dcn314_smu_send_msg_with_param(
+			clk_mgr,
+			VBIOSSMC_MSG_SetDppclkFreq,
+			khz_to_mhz_ceil(requested_dpp_khz));
+
+	return actual_dppclk_set_mhz * 1000;
+}
+
+void dcn314_smu_set_display_idle_optimization(struct clk_mgr_internal *clk_mgr, uint32_t idle_info)
+{
+	if (!clk_mgr->base.ctx->dc->debug.pstate_enabled)
+		return;
+
+	if (!clk_mgr->smu_present)
+		return;
+
+	//TODO: Work with smu team to define optimization options.
+	dcn314_smu_send_msg_with_param(
+		clk_mgr,
+		VBIOSSMC_MSG_SetDisplayIdleOptimizations,
+		idle_info);
+}
+
+void dcn314_smu_enable_phy_refclk_pwrdwn(struct clk_mgr_internal *clk_mgr, bool enable)
+{
+	union display_idle_optimization_u idle_info = { 0 };
+
+	if (!clk_mgr->smu_present)
+		return;
+
+	if (enable) {
+		idle_info.idle_info.df_request_disabled = 1;
+		idle_info.idle_info.phy_ref_clk_off = 1;
+	}
+
+	dcn314_smu_send_msg_with_param(
+			clk_mgr,
+			VBIOSSMC_MSG_SetDisplayIdleOptimizations,
+			idle_info.data);
+}
+
+void dcn314_smu_enable_pme_wa(struct clk_mgr_internal *clk_mgr)
+{
+	if (!clk_mgr->smu_present)
+		return;
+
+	dcn314_smu_send_msg_with_param(
+			clk_mgr,
+			VBIOSSMC_MSG_UpdatePmeRestore,
+			0);
+}
+
+void dcn314_smu_set_dram_addr_high(struct clk_mgr_internal *clk_mgr, uint32_t addr_high)
+{
+	if (!clk_mgr->smu_present)
+		return;
+
+	dcn314_smu_send_msg_with_param(clk_mgr,
+			VBIOSSMC_MSG_SetVbiosDramAddrHigh, addr_high);
+}
+
+void dcn314_smu_set_dram_addr_low(struct clk_mgr_internal *clk_mgr, uint32_t addr_low)
+{
+	if (!clk_mgr->smu_present)
+		return;
+
+	dcn314_smu_send_msg_with_param(clk_mgr,
+			VBIOSSMC_MSG_SetVbiosDramAddrLow, addr_low);
+}
+
+void dcn314_smu_transfer_dpm_table_smu_2_dram(struct clk_mgr_internal *clk_mgr)
+{
+	if (!clk_mgr->smu_present)
+		return;
+
+	dcn314_smu_send_msg_with_param(clk_mgr,
+			VBIOSSMC_MSG_TransferTableSmu2Dram, TABLE_DPMCLOCKS);
+}
+
+void dcn314_smu_transfer_wm_table_dram_2_smu(struct clk_mgr_internal *clk_mgr)
+{
+	if (!clk_mgr->smu_present)
+		return;
+
+	dcn314_smu_send_msg_with_param(clk_mgr,
+			VBIOSSMC_MSG_TransferTableDram2Smu, TABLE_WATERMARKS);
+}
+
+void dcn314_smu_set_zstate_support(struct clk_mgr_internal *clk_mgr, enum dcn_zstate_support_state support)
+{
+	unsigned int msg_id, param;
+
+	if (!clk_mgr->smu_present)
+		return;
+
+	switch (support) {
+
+	case DCN_ZSTATE_SUPPORT_ALLOW:
+		msg_id = VBIOSSMC_MSG_AllowZstatesEntry;
+		param = (1 << 10) | (1 << 9) | (1 << 8);
+		break;
+
+	case DCN_ZSTATE_SUPPORT_DISALLOW:
+		msg_id = VBIOSSMC_MSG_AllowZstatesEntry;
+		param = 0;
+		break;
+
+
+	case DCN_ZSTATE_SUPPORT_ALLOW_Z10_ONLY:
+		msg_id = VBIOSSMC_MSG_AllowZstatesEntry;
+		param = (1 << 10);
+		break;
+
+	case DCN_ZSTATE_SUPPORT_ALLOW_Z8_Z10_ONLY:
+		msg_id = VBIOSSMC_MSG_AllowZstatesEntry;
+		param = (1 << 10) | (1 << 8);
+		break;
+
+	case DCN_ZSTATE_SUPPORT_ALLOW_Z8_ONLY:
+		msg_id = VBIOSSMC_MSG_AllowZstatesEntry;
+		param = (1 << 8);
+		break;
+
+	default: //DCN_ZSTATE_SUPPORT_UNKNOWN
+		msg_id = VBIOSSMC_MSG_AllowZstatesEntry;
+		param = 0;
+		break;
+	}
+
+
+	dcn314_smu_send_msg_with_param(
+		clk_mgr,
+		msg_id,
+		param);
+
+}
+
+/* Arg = 1: Turn DTB on; 0: Turn DTB CLK OFF. when it is on, it is 600MHZ */
+void dcn314_smu_set_dtbclk(struct clk_mgr_internal *clk_mgr, bool enable)
+{
+	if (!clk_mgr->smu_present)
+		return;
+
+	dcn314_smu_send_msg_with_param(
+			clk_mgr,
+			VBIOSSMC_MSG_SetDtbClk,
+			enable);
+}
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn314/dcn314_smu.h b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn314/dcn314_smu.h
new file mode 100644
index 000000000000..78ca1e5c5e9e
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn314/dcn314_smu.h
@@ -0,0 +1,110 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef DAL_DC_314_SMU_H_
+#define DAL_DC_314_SMU_H_
+
+#include "smu13_driver_if_v13_0_4.h"
+
+typedef enum {
+	WCK_RATIO_1_1 = 0,  // DDR5, Wck:ck is always 1:1;
+	WCK_RATIO_1_2,
+	WCK_RATIO_1_4,
+	WCK_RATIO_MAX
+} WCK_RATIO_e;
+
+typedef struct {
+	uint32_t FClk;
+	uint32_t MemClk;
+	uint32_t Voltage;
+	uint8_t  WckRatio;
+	uint8_t  Spare[3];
+} DfPstateTable314_t;
+
+//Freq in MHz
+//Voltage in milli volts with 2 fractional bits
+typedef struct {
+	uint32_t DcfClocks[NUM_DCFCLK_DPM_LEVELS];
+	uint32_t DispClocks[NUM_DISPCLK_DPM_LEVELS];
+	uint32_t DppClocks[NUM_DPPCLK_DPM_LEVELS];
+	uint32_t SocClocks[NUM_SOCCLK_DPM_LEVELS];
+	uint32_t VClocks[NUM_VCN_DPM_LEVELS];
+	uint32_t DClocks[NUM_VCN_DPM_LEVELS];
+	uint32_t SocVoltage[NUM_SOC_VOLTAGE_LEVELS];
+	DfPstateTable314_t DfPstateTable[NUM_DF_PSTATE_LEVELS];
+
+	uint8_t  NumDcfClkLevelsEnabled;
+	uint8_t  NumDispClkLevelsEnabled; //Applies to both Dispclk and Dppclk
+	uint8_t  NumSocClkLevelsEnabled;
+	uint8_t  VcnClkLevelsEnabled;     //Applies to both Vclk and Dclk
+	uint8_t  NumDfPstatesEnabled;
+	uint8_t  spare[3];
+
+	uint32_t MinGfxClk;
+	uint32_t MaxGfxClk;
+} DpmClocks314_t;
+
+struct dcn314_watermarks {
+	// Watermarks
+	WatermarkRowGeneric_t WatermarkRow[WM_COUNT][NUM_WM_RANGES];
+	uint32_t MmHubPadding[7]; // SMU internal use
+};
+
+struct dcn314_smu_dpm_clks {
+	DpmClocks314_t *dpm_clks;
+	union large_integer mc_address;
+};
+
+struct display_idle_optimization {
+	unsigned int df_request_disabled : 1;
+	unsigned int phy_ref_clk_off     : 1;
+	unsigned int s0i2_rdy            : 1;
+	unsigned int reserved            : 29;
+};
+
+union display_idle_optimization_u {
+	struct display_idle_optimization idle_info;
+	uint32_t data;
+};
+
+int dcn314_smu_get_smu_version(struct clk_mgr_internal *clk_mgr);
+int dcn314_smu_set_dispclk(struct clk_mgr_internal *clk_mgr, int requested_dispclk_khz);
+int dcn314_smu_set_dprefclk(struct clk_mgr_internal *clk_mgr);
+int dcn314_smu_set_hard_min_dcfclk(struct clk_mgr_internal *clk_mgr, int requested_dcfclk_khz);
+int dcn314_smu_set_min_deep_sleep_dcfclk(struct clk_mgr_internal *clk_mgr, int requested_min_ds_dcfclk_khz);
+int dcn314_smu_set_dppclk(struct clk_mgr_internal *clk_mgr, int requested_dpp_khz);
+void dcn314_smu_set_display_idle_optimization(struct clk_mgr_internal *clk_mgr, uint32_t idle_info);
+void dcn314_smu_enable_phy_refclk_pwrdwn(struct clk_mgr_internal *clk_mgr, bool enable);
+void dcn314_smu_enable_pme_wa(struct clk_mgr_internal *clk_mgr);
+void dcn314_smu_set_dram_addr_high(struct clk_mgr_internal *clk_mgr, uint32_t addr_high);
+void dcn314_smu_set_dram_addr_low(struct clk_mgr_internal *clk_mgr, uint32_t addr_low);
+void dcn314_smu_transfer_dpm_table_smu_2_dram(struct clk_mgr_internal *clk_mgr);
+void dcn314_smu_transfer_wm_table_dram_2_smu(struct clk_mgr_internal *clk_mgr);
+
+void dcn314_smu_set_zstate_support(struct clk_mgr_internal *clk_mgr, enum dcn_zstate_support_state support);
+void dcn314_smu_set_dtbclk(struct clk_mgr_internal *clk_mgr, bool enable);
+
+#endif /* DAL_DC_314_SMU_H_ */
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn315/dcn315_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn315/dcn315_clk_mgr.c
new file mode 100644
index 000000000000..3a881451e9da
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn315/dcn315_clk_mgr.c
@@ -0,0 +1,820 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+
+
+#include "dccg.h"
+#include "clk_mgr_internal.h"
+
+// For dce12_get_dp_ref_freq_khz
+#include "dce100/dce_clk_mgr.h"
+// For dcn20_update_clocks_update_dpp_dto
+#include "dcn20/dcn20_clk_mgr.h"
+#include "dcn31/dcn31_clk_mgr.h"
+#include "dcn315_clk_mgr.h"
+
+#include "core_types.h"
+#include "dcn315_smu.h"
+#include "dm_helpers.h"
+
+#include "dc_dmub_srv.h"
+#include "reg_helper.h"
+#include "logger_types.h"
+#undef DC_LOGGER
+#define DC_LOGGER \
+	clk_mgr->base.base.ctx->logger
+
+#include "link_service.h"
+
+#define MAX_INSTANCE                                        7
+#define MAX_SEGMENT                                         8
+
+struct IP_BASE_INSTANCE {
+	unsigned int segment[MAX_SEGMENT];
+};
+
+struct IP_BASE {
+	struct IP_BASE_INSTANCE instance[MAX_INSTANCE];
+};
+
+static const struct IP_BASE CLK_BASE = { { { { 0x00016C00, 0x02401800, 0, 0, 0, 0, 0, 0 } },
+					{ { 0x00016E00, 0x02401C00, 0, 0, 0, 0, 0, 0 } },
+					{ { 0x00017000, 0x02402000, 0, 0, 0, 0, 0, 0 } },
+					{ { 0x00017200, 0x02402400, 0, 0, 0, 0, 0, 0 } },
+					{ { 0x0001B000, 0x0242D800, 0, 0, 0, 0, 0, 0 } },
+					{ { 0x0001B200, 0x0242DC00, 0, 0, 0, 0, 0, 0 } } } };
+
+#define regCLK1_CLK0_CURRENT_CNT			0x0314
+#define regCLK1_CLK0_CURRENT_CNT_BASE_IDX	0
+#define regCLK1_CLK1_CURRENT_CNT			0x0315
+#define regCLK1_CLK1_CURRENT_CNT_BASE_IDX	0
+#define regCLK1_CLK2_CURRENT_CNT			0x0316
+#define regCLK1_CLK2_CURRENT_CNT_BASE_IDX	0
+#define regCLK1_CLK3_CURRENT_CNT			0x0317
+#define regCLK1_CLK3_CURRENT_CNT_BASE_IDX	0
+#define regCLK1_CLK4_CURRENT_CNT			0x0318
+#define regCLK1_CLK4_CURRENT_CNT_BASE_IDX	0
+#define regCLK1_CLK5_CURRENT_CNT			0x0319
+#define regCLK1_CLK5_CURRENT_CNT_BASE_IDX	0
+
+#define TO_CLK_MGR_DCN315(clk_mgr)\
+	container_of(clk_mgr, struct clk_mgr_dcn315, base)
+
+#define REG(reg_name) \
+	(CLK_BASE.instance[0].segment[reg ## reg_name ## _BASE_IDX] + reg ## reg_name)
+
+#define UNSUPPORTED_DCFCLK 10000000
+#define MIN_DPP_DISP_CLK     100000
+
+static int dcn315_get_active_display_cnt_wa(
+		struct dc *dc,
+		struct dc_state *context)
+{
+	int i, display_count;
+	bool tmds_present = false;
+
+	display_count = 0;
+	for (i = 0; i < context->stream_count; i++) {
+		const struct dc_stream_state *stream = context->streams[i];
+
+		if (stream->signal == SIGNAL_TYPE_HDMI_TYPE_A ||
+				stream->signal == SIGNAL_TYPE_DVI_SINGLE_LINK ||
+				stream->signal == SIGNAL_TYPE_DVI_DUAL_LINK)
+			tmds_present = true;
+	}
+
+	for (i = 0; i < dc->link_count; i++) {
+		const struct dc_link *link = dc->links[i];
+
+		/* abusing the fact that the dig and phy are coupled to see if the phy is enabled */
+		if (link->link_enc && link->link_enc->funcs->is_dig_enabled &&
+				link->link_enc->funcs->is_dig_enabled(link->link_enc))
+			display_count++;
+	}
+
+	/* WA for hang on HDMI after display off back back on*/
+	if (display_count == 0 && tmds_present)
+		display_count = 1;
+
+	return display_count;
+}
+
+static bool should_disable_otg(struct pipe_ctx *pipe)
+{
+	bool ret = true;
+
+	if (pipe->stream->link->link_enc && pipe->stream->link->link_enc->funcs->is_dig_enabled &&
+			pipe->stream->link->link_enc->funcs->is_dig_enabled(pipe->stream->link->link_enc))
+		ret = false;
+	return ret;
+}
+
+static void dcn315_disable_otg_wa(struct clk_mgr *clk_mgr_base, struct dc_state *context, bool disable)
+{
+	struct dc *dc = clk_mgr_base->ctx->dc;
+	int i;
+
+	for (i = 0; i < dc->res_pool->pipe_count; ++i) {
+		struct pipe_ctx *pipe = &dc->current_state->res_ctx.pipe_ctx[i];
+
+		if (pipe->top_pipe || pipe->prev_odm_pipe)
+			continue;
+		if (pipe->stream && (pipe->stream->dpms_off || pipe->plane_state == NULL ||
+					dc_is_virtual_signal(pipe->stream->signal))) {
+
+			/* This w/a should not trigger when we have a dig active */
+			if (should_disable_otg(pipe)) {
+				if (disable) {
+					pipe->stream_res.tg->funcs->immediate_disable_crtc(pipe->stream_res.tg);
+					reset_sync_context_for_pipe(dc, context, i);
+				} else
+					pipe->stream_res.tg->funcs->enable_crtc(pipe->stream_res.tg);
+			}
+		}
+	}
+}
+
+static void dcn315_update_clocks(struct clk_mgr *clk_mgr_base,
+			struct dc_state *context,
+			bool safe_to_lower)
+{
+	union dmub_rb_cmd cmd;
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+	struct dc_clocks *new_clocks = &context->bw_ctx.bw.dcn.clk;
+	struct dc *dc = clk_mgr_base->ctx->dc;
+	int display_count = 0;
+	bool update_dppclk = false;
+	bool update_dispclk = false;
+	bool dpp_clock_lowered = false;
+
+	if (dc->work_arounds.skip_clock_update)
+		return;
+
+	clk_mgr_base->clks.zstate_support = new_clocks->zstate_support;
+	/*
+	 * if it is safe to lower, but we are already in the lower state, we don't have to do anything
+	 * also if safe to lower is false, we just go in the higher state
+	 */
+	clk_mgr_base->clks.zstate_support = new_clocks->zstate_support;
+	if (safe_to_lower) {
+		if (clk_mgr_base->clks.dtbclk_en && !new_clocks->dtbclk_en) {
+			dcn315_smu_set_dtbclk(clk_mgr, false);
+			clk_mgr_base->clks.dtbclk_en = new_clocks->dtbclk_en;
+		}
+		/* check that we're not already in lower */
+		if (clk_mgr_base->clks.pwr_state != DCN_PWR_STATE_LOW_POWER) {
+			display_count = dcn315_get_active_display_cnt_wa(dc, context);
+			/* if we can go lower, go lower */
+			if (display_count == 0) {
+				union display_idle_optimization_u idle_info = { 0 };
+				idle_info.idle_info.df_request_disabled = 1;
+				idle_info.idle_info.phy_ref_clk_off = 1;
+				idle_info.idle_info.s0i2_rdy = 1;
+				dcn315_smu_set_display_idle_optimization(clk_mgr, idle_info.data);
+				/* update power state */
+				clk_mgr_base->clks.pwr_state = DCN_PWR_STATE_LOW_POWER;
+			}
+		}
+	} else {
+		if (!clk_mgr_base->clks.dtbclk_en && new_clocks->dtbclk_en) {
+			dcn315_smu_set_dtbclk(clk_mgr, true);
+			clk_mgr_base->clks.dtbclk_en = new_clocks->dtbclk_en;
+		}
+		/* check that we're not already in D0 */
+		if (clk_mgr_base->clks.pwr_state != DCN_PWR_STATE_MISSION_MODE) {
+			union display_idle_optimization_u idle_info = { 0 };
+			dcn315_smu_set_display_idle_optimization(clk_mgr, idle_info.data);
+			/* update power state */
+			clk_mgr_base->clks.pwr_state = DCN_PWR_STATE_MISSION_MODE;
+		}
+	}
+
+	/* Lock pstate by requesting unsupported dcfclk if change is unsupported */
+	if (!new_clocks->p_state_change_support)
+		new_clocks->dcfclk_khz = UNSUPPORTED_DCFCLK;
+	if (should_set_clock(safe_to_lower, new_clocks->dcfclk_khz, clk_mgr_base->clks.dcfclk_khz)) {
+		clk_mgr_base->clks.dcfclk_khz = new_clocks->dcfclk_khz;
+		dcn315_smu_set_hard_min_dcfclk(clk_mgr, clk_mgr_base->clks.dcfclk_khz);
+	}
+
+	if (should_set_clock(safe_to_lower,
+			new_clocks->dcfclk_deep_sleep_khz, clk_mgr_base->clks.dcfclk_deep_sleep_khz)) {
+		clk_mgr_base->clks.dcfclk_deep_sleep_khz = new_clocks->dcfclk_deep_sleep_khz;
+		dcn315_smu_set_min_deep_sleep_dcfclk(clk_mgr, clk_mgr_base->clks.dcfclk_deep_sleep_khz);
+	}
+
+	// workaround: Limit dppclk to 100Mhz to avoid lower eDP panel switch to plus 4K monitor underflow.
+	if (new_clocks->dppclk_khz < MIN_DPP_DISP_CLK)
+		new_clocks->dppclk_khz = MIN_DPP_DISP_CLK;
+
+	if (should_set_clock(safe_to_lower, new_clocks->dppclk_khz, clk_mgr->base.clks.dppclk_khz)) {
+		if (clk_mgr->base.clks.dppclk_khz > new_clocks->dppclk_khz)
+			dpp_clock_lowered = true;
+		clk_mgr_base->clks.dppclk_khz = new_clocks->dppclk_khz;
+		update_dppclk = true;
+	}
+
+	if (should_set_clock(safe_to_lower, new_clocks->dispclk_khz, clk_mgr_base->clks.dispclk_khz) &&
+	    (new_clocks->dispclk_khz > 0 || (safe_to_lower && display_count == 0))) {
+		int requested_dispclk_khz = new_clocks->dispclk_khz;
+
+		dcn315_disable_otg_wa(clk_mgr_base, context, true);
+
+		/* Clamp the requested clock to PMFW based on their limit. */
+		if (dc->debug.min_disp_clk_khz > 0 && requested_dispclk_khz < dc->debug.min_disp_clk_khz)
+			requested_dispclk_khz = dc->debug.min_disp_clk_khz;
+
+		dcn315_smu_set_dispclk(clk_mgr, requested_dispclk_khz);
+		clk_mgr_base->clks.dispclk_khz = new_clocks->dispclk_khz;
+		dcn315_disable_otg_wa(clk_mgr_base, context, false);
+
+		update_dispclk = true;
+	}
+
+	if (dpp_clock_lowered) {
+		// increase per DPP DTO before lowering global dppclk
+		dcn20_update_clocks_update_dpp_dto(clk_mgr, context, safe_to_lower);
+		dcn315_smu_set_dppclk(clk_mgr, clk_mgr_base->clks.dppclk_khz);
+	} else {
+		// increase global DPPCLK before lowering per DPP DTO
+		if (update_dppclk || update_dispclk)
+			dcn315_smu_set_dppclk(clk_mgr, clk_mgr_base->clks.dppclk_khz);
+		// always update dtos unless clock is lowered and not safe to lower
+		if (new_clocks->dppclk_khz >= dc->current_state->bw_ctx.bw.dcn.clk.dppclk_khz)
+			dcn20_update_clocks_update_dpp_dto(clk_mgr, context, safe_to_lower);
+	}
+
+	// notify DMCUB of latest clocks
+	memset(&cmd, 0, sizeof(cmd));
+	cmd.notify_clocks.header.type = DMUB_CMD__CLK_MGR;
+	cmd.notify_clocks.header.sub_type = DMUB_CMD__CLK_MGR_NOTIFY_CLOCKS;
+	cmd.notify_clocks.clocks.dcfclk_khz = clk_mgr_base->clks.dcfclk_khz;
+	cmd.notify_clocks.clocks.dcfclk_deep_sleep_khz =
+		clk_mgr_base->clks.dcfclk_deep_sleep_khz;
+	cmd.notify_clocks.clocks.dispclk_khz = clk_mgr_base->clks.dispclk_khz;
+	cmd.notify_clocks.clocks.dppclk_khz = clk_mgr_base->clks.dppclk_khz;
+
+	dc_wake_and_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
+}
+
+static void dcn315_dump_clk_registers_internal(struct dcn35_clk_internal *internal, struct clk_mgr *clk_mgr_base)
+{
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+
+	// read dtbclk
+	internal->CLK1_CLK4_CURRENT_CNT = REG_READ(CLK1_CLK4_CURRENT_CNT);
+
+	// read dcfclk
+	internal->CLK1_CLK3_CURRENT_CNT = REG_READ(CLK1_CLK3_CURRENT_CNT);
+
+	// read dppclk
+	internal->CLK1_CLK1_CURRENT_CNT = REG_READ(CLK1_CLK1_CURRENT_CNT);
+
+	// read dprefclk
+	internal->CLK1_CLK2_CURRENT_CNT = REG_READ(CLK1_CLK2_CURRENT_CNT);
+
+	// read dispclk
+	internal->CLK1_CLK0_CURRENT_CNT = REG_READ(CLK1_CLK0_CURRENT_CNT);
+}
+
+static void dcn315_dump_clk_registers(struct clk_state_registers_and_bypass *regs_and_bypass,
+		struct clk_mgr *clk_mgr_base, struct clk_log_info *log_info)
+{
+	struct dcn35_clk_internal internal = {0};
+
+	dcn315_dump_clk_registers_internal(&internal, clk_mgr_base);
+
+	regs_and_bypass->dcfclk = internal.CLK1_CLK3_CURRENT_CNT / 10;
+	regs_and_bypass->dprefclk = internal.CLK1_CLK2_CURRENT_CNT / 10;
+	regs_and_bypass->dispclk = internal.CLK1_CLK0_CURRENT_CNT / 10;
+	regs_and_bypass->dppclk = internal.CLK1_CLK1_CURRENT_CNT / 10;
+	regs_and_bypass->dtbclk = internal.CLK1_CLK4_CURRENT_CNT / 10;
+	return;
+}
+
+static struct clk_bw_params dcn315_bw_params = {
+	.vram_type = Ddr4MemType,
+	.num_channels = 2,
+	.clk_table = {
+		.entries = {
+			{
+				.voltage = 0,
+				.dispclk_mhz = 640,
+				.dppclk_mhz = 640,
+				.phyclk_mhz = 810,
+				.phyclk_d18_mhz = 667,
+				.dtbclk_mhz = 600,
+			},
+			{
+				.voltage = 1,
+				.dispclk_mhz = 739,
+				.dppclk_mhz = 739,
+				.phyclk_mhz = 810,
+				.phyclk_d18_mhz = 667,
+				.dtbclk_mhz = 600,
+			},
+			{
+				.voltage = 2,
+				.dispclk_mhz = 960,
+				.dppclk_mhz = 960,
+				.phyclk_mhz = 810,
+				.phyclk_d18_mhz = 667,
+				.dtbclk_mhz = 600,
+			},
+			{
+				.voltage = 3,
+				.dispclk_mhz = 1200,
+				.dppclk_mhz = 1200,
+				.phyclk_mhz = 810,
+				.phyclk_d18_mhz = 667,
+				.dtbclk_mhz = 600,
+			},
+			{
+				.voltage = 4,
+				.dispclk_mhz = 1372,
+				.dppclk_mhz = 1372,
+				.phyclk_mhz = 810,
+				.phyclk_d18_mhz = 667,
+				.dtbclk_mhz = 600,
+			},
+		},
+		.num_entries = 5,
+	},
+
+};
+
+static struct wm_table ddr5_wm_table = {
+	.entries = {
+		{
+			.wm_inst = WM_A,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 129.0,
+			.sr_exit_time_us = 11.5,
+			.sr_enter_plus_exit_time_us = 14.5,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_B,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 129.0,
+			.sr_exit_time_us = 11.5,
+			.sr_enter_plus_exit_time_us = 14.5,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_C,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 129.0,
+			.sr_exit_time_us = 11.5,
+			.sr_enter_plus_exit_time_us = 14.5,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_D,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 129.0,
+			.sr_exit_time_us = 11.5,
+			.sr_enter_plus_exit_time_us = 14.5,
+			.valid = true,
+		},
+	}
+};
+
+static struct wm_table lpddr5_wm_table = {
+	.entries = {
+		{
+			.wm_inst = WM_A,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 129.0,
+			.sr_exit_time_us = 11.5,
+			.sr_enter_plus_exit_time_us = 14.5,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_B,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 129.0,
+			.sr_exit_time_us = 11.5,
+			.sr_enter_plus_exit_time_us = 14.5,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_C,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 129.0,
+			.sr_exit_time_us = 11.5,
+			.sr_enter_plus_exit_time_us = 14.5,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_D,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 129.0,
+			.sr_exit_time_us = 11.5,
+			.sr_enter_plus_exit_time_us = 14.5,
+			.valid = true,
+		},
+	}
+};
+
+/* Temporary Place holder until we can get them from fuse */
+static DpmClocks_315_t dummy_clocks = { 0 };
+static struct dcn315_watermarks dummy_wms = { 0 };
+
+static void dcn315_build_watermark_ranges(struct clk_bw_params *bw_params, struct dcn315_watermarks *table)
+{
+	int i, num_valid_sets;
+
+	num_valid_sets = 0;
+
+	for (i = 0; i < WM_SET_COUNT; i++) {
+		/* skip empty entries, the smu array has no holes*/
+		if (!bw_params->wm_table.entries[i].valid)
+			continue;
+
+		table->WatermarkRow[WM_DCFCLK][num_valid_sets].WmSetting = bw_params->wm_table.entries[i].wm_inst;
+		table->WatermarkRow[WM_DCFCLK][num_valid_sets].WmType = bw_params->wm_table.entries[i].wm_type;
+		/* We will not select WM based on fclk, so leave it as unconstrained */
+		table->WatermarkRow[WM_DCFCLK][num_valid_sets].MinClock = 0;
+		table->WatermarkRow[WM_DCFCLK][num_valid_sets].MaxClock = 0xFFFF;
+
+		if (table->WatermarkRow[WM_DCFCLK][num_valid_sets].WmType == WM_TYPE_PSTATE_CHG) {
+			if (i == 0)
+				table->WatermarkRow[WM_DCFCLK][num_valid_sets].MinMclk = 0;
+			else {
+				/* add 1 to make it non-overlapping with next lvl */
+				table->WatermarkRow[WM_DCFCLK][num_valid_sets].MinMclk =
+						bw_params->clk_table.entries[i - 1].dcfclk_mhz + 1;
+			}
+			table->WatermarkRow[WM_DCFCLK][num_valid_sets].MaxMclk =
+					bw_params->clk_table.entries[i].dcfclk_mhz;
+
+		} else {
+			/* unconstrained for memory retraining */
+			table->WatermarkRow[WM_DCFCLK][num_valid_sets].MinClock = 0;
+			table->WatermarkRow[WM_DCFCLK][num_valid_sets].MaxClock = 0xFFFF;
+
+			/* Modify previous watermark range to cover up to max */
+			table->WatermarkRow[WM_DCFCLK][num_valid_sets - 1].MaxClock = 0xFFFF;
+		}
+		num_valid_sets++;
+	}
+
+	ASSERT(num_valid_sets != 0); /* Must have at least one set of valid watermarks */
+
+	/* modify the min and max to make sure we cover the whole range*/
+	table->WatermarkRow[WM_DCFCLK][0].MinMclk = 0;
+	table->WatermarkRow[WM_DCFCLK][0].MinClock = 0;
+	table->WatermarkRow[WM_DCFCLK][num_valid_sets - 1].MaxMclk = 0xFFFF;
+	table->WatermarkRow[WM_DCFCLK][num_valid_sets - 1].MaxClock = 0xFFFF;
+
+	/* This is for writeback only, does not matter currently as no writeback support*/
+	table->WatermarkRow[WM_SOCCLK][0].WmSetting = WM_A;
+	table->WatermarkRow[WM_SOCCLK][0].MinClock = 0;
+	table->WatermarkRow[WM_SOCCLK][0].MaxClock = 0xFFFF;
+	table->WatermarkRow[WM_SOCCLK][0].MinMclk = 0;
+	table->WatermarkRow[WM_SOCCLK][0].MaxMclk = 0xFFFF;
+}
+
+static void dcn315_notify_wm_ranges(struct clk_mgr *clk_mgr_base)
+{
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+	struct clk_mgr_dcn315 *clk_mgr_dcn315 = TO_CLK_MGR_DCN315(clk_mgr);
+	struct dcn315_watermarks *table = clk_mgr_dcn315->smu_wm_set.wm_set;
+
+	if (!clk_mgr->smu_ver)
+		return;
+
+	if (!table || clk_mgr_dcn315->smu_wm_set.mc_address.quad_part == 0)
+		return;
+
+	memset(table, 0, sizeof(*table));
+
+	dcn315_build_watermark_ranges(clk_mgr_base->bw_params, table);
+
+	dcn315_smu_set_dram_addr_high(clk_mgr,
+			clk_mgr_dcn315->smu_wm_set.mc_address.high_part);
+	dcn315_smu_set_dram_addr_low(clk_mgr,
+			clk_mgr_dcn315->smu_wm_set.mc_address.low_part);
+	dcn315_smu_transfer_wm_table_dram_2_smu(clk_mgr);
+}
+
+static void dcn315_get_dpm_table_from_smu(struct clk_mgr_internal *clk_mgr,
+		struct dcn315_smu_dpm_clks *smu_dpm_clks)
+{
+	DpmClocks_315_t *table = smu_dpm_clks->dpm_clks;
+
+	if (!clk_mgr->smu_ver)
+		return;
+
+	if (!table || smu_dpm_clks->mc_address.quad_part == 0)
+		return;
+
+	memset(table, 0, sizeof(*table));
+
+	dcn315_smu_set_dram_addr_high(clk_mgr,
+			smu_dpm_clks->mc_address.high_part);
+	dcn315_smu_set_dram_addr_low(clk_mgr,
+			smu_dpm_clks->mc_address.low_part);
+	dcn315_smu_transfer_dpm_table_smu_2_dram(clk_mgr);
+}
+
+static void dcn315_clk_mgr_helper_populate_bw_params(
+		struct clk_mgr_internal *clk_mgr,
+		struct integrated_info *bios_info,
+		const DpmClocks_315_t *clock_table)
+{
+	int i;
+	struct clk_bw_params *bw_params = clk_mgr->base.bw_params;
+	uint32_t max_pstate = clock_table->NumDfPstatesEnabled - 1;
+	struct clk_limit_table_entry def_max = bw_params->clk_table.entries[bw_params->clk_table.num_entries - 1];
+
+	/* For 315 we want to base clock table on dcfclk, need at least one entry regardless of pmfw table */
+	for (i = 0; i < clock_table->NumDcfClkLevelsEnabled; i++) {
+		int j;
+
+		/* DF table is sorted with clocks decreasing */
+		for (j = clock_table->NumDfPstatesEnabled - 2; j >= 0; j--) {
+			if (clock_table->DfPstateTable[j].Voltage <= clock_table->SocVoltage[i])
+				max_pstate = j;
+		}
+		/* Max DCFCLK should match up with max pstate */
+		if (i == clock_table->NumDcfClkLevelsEnabled - 1)
+			max_pstate = 0;
+
+		/* First search defaults for the clocks we don't read using closest lower or equal default dcfclk */
+		for (j = bw_params->clk_table.num_entries - 1; j > 0; j--)
+			if (bw_params->clk_table.entries[j].dcfclk_mhz <= clock_table->DcfClocks[i])
+				break;
+		bw_params->clk_table.entries[i].phyclk_mhz = bw_params->clk_table.entries[j].phyclk_mhz;
+		bw_params->clk_table.entries[i].phyclk_d18_mhz = bw_params->clk_table.entries[j].phyclk_d18_mhz;
+		bw_params->clk_table.entries[i].dtbclk_mhz = bw_params->clk_table.entries[j].dtbclk_mhz;
+
+		/* Now update clocks we do read */
+		bw_params->clk_table.entries[i].fclk_mhz = clock_table->DfPstateTable[max_pstate].FClk;
+		bw_params->clk_table.entries[i].memclk_mhz = clock_table->DfPstateTable[max_pstate].MemClk;
+		bw_params->clk_table.entries[i].voltage = clock_table->SocVoltage[i];
+		bw_params->clk_table.entries[i].dcfclk_mhz = clock_table->DcfClocks[i];
+		bw_params->clk_table.entries[i].socclk_mhz = clock_table->SocClocks[i];
+		bw_params->clk_table.entries[i].dispclk_mhz = clock_table->DispClocks[i];
+		bw_params->clk_table.entries[i].dppclk_mhz = clock_table->DppClocks[i];
+		bw_params->clk_table.entries[i].wck_ratio = 1;
+	}
+
+	/* Make sure to include at least one entry */
+	if (i == 0) {
+		bw_params->clk_table.entries[i].fclk_mhz = clock_table->DfPstateTable[0].FClk;
+		bw_params->clk_table.entries[i].memclk_mhz = clock_table->DfPstateTable[0].MemClk;
+		bw_params->clk_table.entries[i].voltage = clock_table->DfPstateTable[0].Voltage;
+		bw_params->clk_table.entries[i].dcfclk_mhz = clock_table->DcfClocks[0];
+		bw_params->clk_table.entries[i].wck_ratio = 1;
+		i++;
+	} else if (clock_table->NumDcfClkLevelsEnabled != clock_table->NumSocClkLevelsEnabled) {
+		bw_params->clk_table.entries[i-1].voltage = clock_table->SocVoltage[clock_table->NumSocClkLevelsEnabled - 1];
+		bw_params->clk_table.entries[i-1].socclk_mhz = clock_table->SocClocks[clock_table->NumSocClkLevelsEnabled - 1];
+		bw_params->clk_table.entries[i-1].dispclk_mhz = clock_table->DispClocks[clock_table->NumDispClkLevelsEnabled - 1];
+		bw_params->clk_table.entries[i-1].dppclk_mhz = clock_table->DppClocks[clock_table->NumDispClkLevelsEnabled - 1];
+	}
+	bw_params->clk_table.num_entries = i;
+
+	/* Set any 0 clocks to max default setting. Not an issue for
+	 * power since we aren't doing switching in such case anyway
+	 */
+	for (i = 0; i < bw_params->clk_table.num_entries; i++) {
+		if (!bw_params->clk_table.entries[i].fclk_mhz) {
+			bw_params->clk_table.entries[i].fclk_mhz = def_max.fclk_mhz;
+			bw_params->clk_table.entries[i].memclk_mhz = def_max.memclk_mhz;
+			bw_params->clk_table.entries[i].voltage = def_max.voltage;
+		}
+		if (!bw_params->clk_table.entries[i].dcfclk_mhz)
+			bw_params->clk_table.entries[i].dcfclk_mhz = def_max.dcfclk_mhz;
+		if (!bw_params->clk_table.entries[i].socclk_mhz)
+			bw_params->clk_table.entries[i].socclk_mhz = def_max.socclk_mhz;
+		if (!bw_params->clk_table.entries[i].dispclk_mhz)
+			bw_params->clk_table.entries[i].dispclk_mhz = def_max.dispclk_mhz;
+		if (!bw_params->clk_table.entries[i].dppclk_mhz)
+			bw_params->clk_table.entries[i].dppclk_mhz = def_max.dppclk_mhz;
+		if (!bw_params->clk_table.entries[i].phyclk_mhz)
+			bw_params->clk_table.entries[i].phyclk_mhz = def_max.phyclk_mhz;
+		if (!bw_params->clk_table.entries[i].phyclk_d18_mhz)
+			bw_params->clk_table.entries[i].phyclk_d18_mhz = def_max.phyclk_d18_mhz;
+		if (!bw_params->clk_table.entries[i].dtbclk_mhz)
+			bw_params->clk_table.entries[i].dtbclk_mhz = def_max.dtbclk_mhz;
+	}
+
+	/* Make sure all highest default clocks are included*/
+	ASSERT(bw_params->clk_table.entries[i-1].phyclk_mhz == def_max.phyclk_mhz);
+	ASSERT(bw_params->clk_table.entries[i-1].phyclk_d18_mhz == def_max.phyclk_d18_mhz);
+	ASSERT(bw_params->clk_table.entries[i-1].dtbclk_mhz == def_max.dtbclk_mhz);
+	ASSERT(bw_params->clk_table.entries[i-1].dcfclk_mhz);
+	bw_params->vram_type = bios_info->memory_type;
+	bw_params->num_channels = bios_info->ma_channel_number;
+	bw_params->dram_channel_width_bytes = bios_info->memory_type == 0x22 ? 8 : 4;
+
+	for (i = 0; i < WM_SET_COUNT; i++) {
+		bw_params->wm_table.entries[i].wm_inst = i;
+
+		if (i >= bw_params->clk_table.num_entries) {
+			bw_params->wm_table.entries[i].valid = false;
+			continue;
+		}
+
+		bw_params->wm_table.entries[i].wm_type = WM_TYPE_PSTATE_CHG;
+		bw_params->wm_table.entries[i].valid = true;
+	}
+}
+
+static void dcn315_enable_pme_wa(struct clk_mgr *clk_mgr_base)
+{
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+
+	dcn315_smu_enable_pme_wa(clk_mgr);
+}
+
+static struct clk_mgr_funcs dcn315_funcs = {
+	.get_dp_ref_clk_frequency = dce12_get_dp_ref_freq_khz,
+	.get_dtb_ref_clk_frequency = dcn31_get_dtb_ref_freq_khz,
+	.update_clocks = dcn315_update_clocks,
+	.init_clocks = dcn315_init_clocks,
+	.enable_pme_wa = dcn315_enable_pme_wa,
+	.are_clock_states_equal = dcn31_are_clock_states_equal,
+	.notify_wm_ranges = dcn315_notify_wm_ranges
+};
+extern struct clk_mgr_funcs dcn3_fpga_funcs;
+
+void dcn315_init_clocks(struct clk_mgr *clk_mgr)
+{
+	struct clk_mgr_internal *clk_mgr_int = TO_CLK_MGR_INTERNAL(clk_mgr);
+	uint32_t ref_dtbclk = clk_mgr->clks.ref_dtbclk_khz;
+	struct clk_mgr_dcn315 *clk_mgr_dcn315 = TO_CLK_MGR_DCN315(clk_mgr_int);
+	struct clk_log_info log_info = {0};
+
+	memset(&(clk_mgr->clks), 0, sizeof(struct dc_clocks));
+	// Assumption is that boot state always supports pstate
+	clk_mgr->clks.ref_dtbclk_khz = ref_dtbclk;	// restore ref_dtbclk
+	clk_mgr->clks.p_state_change_support = true;
+	clk_mgr->clks.prev_p_state_change_support = true;
+	clk_mgr->clks.pwr_state = DCN_PWR_STATE_UNKNOWN;
+	clk_mgr->clks.zstate_support = DCN_ZSTATE_SUPPORT_UNKNOWN;
+
+	dcn315_dump_clk_registers(&clk_mgr->boot_snapshot, &clk_mgr_dcn315->base.base, &log_info);
+	clk_mgr->clks.dispclk_khz =  clk_mgr->boot_snapshot.dispclk * 1000;
+}
+
+void dcn315_clk_mgr_construct(
+		struct dc_context *ctx,
+		struct clk_mgr_dcn315 *clk_mgr,
+		struct pp_smu_funcs *pp_smu,
+		struct dccg *dccg)
+{
+	struct dcn315_smu_dpm_clks smu_dpm_clks = { 0 };
+	struct clk_log_info log_info = {0};
+
+	clk_mgr->base.base.ctx = ctx;
+	clk_mgr->base.base.funcs = &dcn315_funcs;
+
+	clk_mgr->base.pp_smu = pp_smu;
+
+	clk_mgr->base.dccg = dccg;
+	clk_mgr->base.dfs_bypass_disp_clk = 0;
+
+	clk_mgr->base.dprefclk_ss_percentage = 0;
+	clk_mgr->base.dprefclk_ss_divider = 1000;
+	clk_mgr->base.ss_on_dprefclk = false;
+	clk_mgr->base.dfs_ref_freq_khz = 48000;
+
+	clk_mgr->smu_wm_set.wm_set = (struct dcn315_watermarks *)dm_helpers_allocate_gpu_mem(
+				clk_mgr->base.base.ctx,
+				DC_MEM_ALLOC_TYPE_FRAME_BUFFER,
+				sizeof(struct dcn315_watermarks),
+				&clk_mgr->smu_wm_set.mc_address.quad_part);
+
+	if (!clk_mgr->smu_wm_set.wm_set) {
+		clk_mgr->smu_wm_set.wm_set = &dummy_wms;
+		clk_mgr->smu_wm_set.mc_address.quad_part = 0;
+	}
+	ASSERT(clk_mgr->smu_wm_set.wm_set);
+
+	smu_dpm_clks.dpm_clks = (DpmClocks_315_t *)dm_helpers_allocate_gpu_mem(
+				clk_mgr->base.base.ctx,
+				DC_MEM_ALLOC_TYPE_FRAME_BUFFER,
+				sizeof(DpmClocks_315_t),
+				&smu_dpm_clks.mc_address.quad_part);
+
+	if (smu_dpm_clks.dpm_clks == NULL) {
+		smu_dpm_clks.dpm_clks = &dummy_clocks;
+		smu_dpm_clks.mc_address.quad_part = 0;
+	}
+
+	ASSERT(smu_dpm_clks.dpm_clks);
+
+	clk_mgr->base.smu_ver = dcn315_smu_get_smu_version(&clk_mgr->base);
+
+	if (clk_mgr->base.smu_ver > 0)
+		clk_mgr->base.smu_present = true;
+
+	if (ctx->dc_bios->integrated_info->memory_type == LpDdr5MemType) {
+		dcn315_bw_params.wm_table = lpddr5_wm_table;
+	} else {
+		dcn315_bw_params.wm_table = ddr5_wm_table;
+	}
+	/* Saved clocks configured at boot for debug purposes */
+	dcn315_dump_clk_registers(&clk_mgr->base.base.boot_snapshot,
+				  &clk_mgr->base.base, &log_info);
+	clk_mgr->base.base.clks.dispclk_khz =  clk_mgr->base.base.boot_snapshot.dispclk * 1000;
+
+	clk_mgr->base.base.dprefclk_khz = 600000;
+	clk_mgr->base.base.dprefclk_khz = dcn315_smu_get_dpref_clk(&clk_mgr->base);
+	clk_mgr->base.base.clks.ref_dtbclk_khz = clk_mgr->base.base.dprefclk_khz;
+	dce_clock_read_ss_info(&clk_mgr->base);
+	clk_mgr->base.base.clks.ref_dtbclk_khz = dce_adjust_dp_ref_freq_for_ss(&clk_mgr->base, clk_mgr->base.base.dprefclk_khz);
+
+	clk_mgr->base.base.bw_params = &dcn315_bw_params;
+
+	if (clk_mgr->base.base.ctx->dc->debug.pstate_enabled) {
+		int i;
+
+		dcn315_get_dpm_table_from_smu(&clk_mgr->base, &smu_dpm_clks);
+		DC_LOG_SMU("NumDcfClkLevelsEnabled: %d\n"
+				   "NumDispClkLevelsEnabled: %d\n"
+				   "NumSocClkLevelsEnabled: %d\n"
+				   "VcnClkLevelsEnabled: %d\n"
+				   "NumDfPst atesEnabled: %d\n"
+				   "MinGfxClk: %d\n"
+				   "MaxGfxClk: %d\n",
+				   smu_dpm_clks.dpm_clks->NumDcfClkLevelsEnabled,
+				   smu_dpm_clks.dpm_clks->NumDispClkLevelsEnabled,
+				   smu_dpm_clks.dpm_clks->NumSocClkLevelsEnabled,
+				   smu_dpm_clks.dpm_clks->VcnClkLevelsEnabled,
+				   smu_dpm_clks.dpm_clks->NumDfPstatesEnabled,
+				   smu_dpm_clks.dpm_clks->MinGfxClk,
+				   smu_dpm_clks.dpm_clks->MaxGfxClk);
+		for (i = 0; i < smu_dpm_clks.dpm_clks->NumDcfClkLevelsEnabled; i++) {
+			DC_LOG_SMU("smu_dpm_clks.dpm_clks->DcfClocks[%d] = %d\n",
+					   i,
+					   smu_dpm_clks.dpm_clks->DcfClocks[i]);
+		}
+		for (i = 0; i < smu_dpm_clks.dpm_clks->NumDispClkLevelsEnabled; i++) {
+			DC_LOG_SMU("smu_dpm_clks.dpm_clks->DispClocks[%d] = %d\n",
+					   i, smu_dpm_clks.dpm_clks->DispClocks[i]);
+		}
+		for (i = 0; i < smu_dpm_clks.dpm_clks->NumSocClkLevelsEnabled; i++) {
+			DC_LOG_SMU("smu_dpm_clks.dpm_clks->SocClocks[%d] = %d\n",
+					   i, smu_dpm_clks.dpm_clks->SocClocks[i]);
+		}
+		for (i = 0; i < NUM_SOC_VOLTAGE_LEVELS; i++)
+			DC_LOG_SMU("smu_dpm_clks.dpm_clks->SocVoltage[%d] = %d\n",
+					   i, smu_dpm_clks.dpm_clks->SocVoltage[i]);
+
+		for (i = 0; i < NUM_DF_PSTATE_LEVELS; i++) {
+			DC_LOG_SMU("smu_dpm_clks.dpm_clks.DfPstateTable[%d].FClk = %d\n"
+					   "smu_dpm_clks.dpm_clks->DfPstateTable[%d].MemClk= %d\n"
+					   "smu_dpm_clks.dpm_clks->DfPstateTable[%d].Voltage = %d\n",
+					   i, smu_dpm_clks.dpm_clks->DfPstateTable[i].FClk,
+					   i, smu_dpm_clks.dpm_clks->DfPstateTable[i].MemClk,
+					   i, smu_dpm_clks.dpm_clks->DfPstateTable[i].Voltage);
+		}
+
+		if (ctx->dc_bios->integrated_info) {
+			dcn315_clk_mgr_helper_populate_bw_params(
+					&clk_mgr->base,
+					ctx->dc_bios->integrated_info,
+					smu_dpm_clks.dpm_clks);
+		}
+	}
+
+	if (smu_dpm_clks.dpm_clks && smu_dpm_clks.mc_address.quad_part != 0)
+		dm_helpers_free_gpu_mem(clk_mgr->base.base.ctx, DC_MEM_ALLOC_TYPE_FRAME_BUFFER,
+				smu_dpm_clks.dpm_clks);
+}
+
+void dcn315_clk_mgr_destroy(struct clk_mgr_internal *clk_mgr_int)
+{
+	struct clk_mgr_dcn315 *clk_mgr = TO_CLK_MGR_DCN315(clk_mgr_int);
+
+	if (clk_mgr->smu_wm_set.wm_set && clk_mgr->smu_wm_set.mc_address.quad_part != 0)
+		dm_helpers_free_gpu_mem(clk_mgr_int->base.ctx, DC_MEM_ALLOC_TYPE_FRAME_BUFFER,
+				clk_mgr->smu_wm_set.wm_set);
+}
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn315/dcn315_clk_mgr.h b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn315/dcn315_clk_mgr.h
new file mode 100644
index 000000000000..642ae3d4a790
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn315/dcn315_clk_mgr.h
@@ -0,0 +1,50 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __DCN315_CLK_MGR_H__
+#define __DCN315_CLK_MGR_H__
+#include "clk_mgr_internal.h"
+
+struct dcn315_watermarks;
+
+struct dcn315_smu_watermark_set {
+	struct dcn315_watermarks *wm_set;
+	union large_integer mc_address;
+};
+
+struct clk_mgr_dcn315 {
+	struct clk_mgr_internal base;
+	struct dcn315_smu_watermark_set smu_wm_set;
+};
+
+void dcn315_clk_mgr_construct(struct dc_context *ctx,
+		struct clk_mgr_dcn315 *clk_mgr,
+		struct pp_smu_funcs *pp_smu,
+		struct dccg *dccg);
+
+void dcn315_init_clocks(struct clk_mgr *clk_mgr);
+void dcn315_clk_mgr_destroy(struct clk_mgr_internal *clk_mgr_int);
+
+#endif //__DCN315_CLK_MGR_H__
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn315/dcn315_smu.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn315/dcn315_smu.c
new file mode 100644
index 000000000000..478b4d6a3544
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn315/dcn315_smu.c
@@ -0,0 +1,366 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#include "core_types.h"
+#include "clk_mgr_internal.h"
+#include "reg_helper.h"
+#include "dm_helpers.h"
+#include "dcn315_smu.h"
+#include "mp/mp_13_0_5_offset.h"
+#include "logger_types.h"
+
+#define MAX_INSTANCE                                        6
+#define MAX_SEGMENT                                         6
+#define SMU_REGISTER_WRITE_RETRY_COUNT                      5
+
+struct IP_BASE_INSTANCE {
+    unsigned int segment[MAX_SEGMENT];
+};
+
+struct IP_BASE {
+    struct IP_BASE_INSTANCE instance[MAX_INSTANCE];
+};
+
+static const struct IP_BASE MP0_BASE = { { { { 0x00016000, 0x00DC0000, 0x00E00000, 0x00E40000, 0x0243FC00, 0 } },
+					{ { 0, 0, 0, 0, 0, 0 } },
+					{ { 0, 0, 0, 0, 0, 0 } },
+					{ { 0, 0, 0, 0, 0, 0 } },
+					{ { 0, 0, 0, 0, 0, 0 } },
+					{ { 0, 0, 0, 0, 0, 0 } } } };
+
+#define CTX clk_mgr->base.ctx
+#define IND_REG(offset)	offset
+
+#define regBIF_BX_PF2_RSMU_INDEX                                                                        0x0000
+#define regBIF_BX_PF2_RSMU_INDEX_BASE_IDX                                                               1
+#define regBIF_BX_PF2_RSMU_DATA                                                                         0x0001
+#define regBIF_BX_PF2_RSMU_DATA_BASE_IDX                                                                1
+
+#define REG(reg_name) \
+	(MP0_BASE.instance[0].segment[reg ## reg_name ## _BASE_IDX] + reg ## reg_name)
+
+#define FN(reg_name, field) \
+	FD(reg_name##__##field)
+
+#undef DC_LOGGER
+#define DC_LOGGER \
+	CTX->logger
+#define smu_print(str, ...) {DC_LOG_SMU(str, ##__VA_ARGS__); }
+
+#define mmMP1_C2PMSG_3                            0x3B1050C
+
+#define reg__MP1_C2PMSG_3_MASK					(0xFFFFFFFF)
+#define reg__MP1_C2PMSG_3__SHIFT					(0)
+
+
+#define data_reg_name__MP1_C2PMSG_3_MASK		(0xFFFFFFFF)
+#define data_reg_name__MP1_C2PMSG_3__SHIFT		(0)
+
+#define VBIOSSMC_MSG_TestMessage                  0x01 ///< To check if PMFW is alive and responding. Requirement specified by PMFW team
+#define VBIOSSMC_MSG_GetPmfwVersion               0x02 ///< Get PMFW version
+#define VBIOSSMC_MSG_Spare0                       0x03 ///< Spare0
+#define VBIOSSMC_MSG_SetDispclkFreq               0x04 ///< Set display clock frequency in MHZ
+#define VBIOSSMC_MSG_Spare1                       0x05 ///< Spare1
+#define VBIOSSMC_MSG_SetDppclkFreq                0x06 ///< Set DPP clock frequency in MHZ
+#define VBIOSSMC_MSG_SetHardMinDcfclkByFreq       0x07 ///< Set DCF clock frequency hard min in MHZ
+#define VBIOSSMC_MSG_SetMinDeepSleepDcfclk        0x08 ///< Set DCF clock minimum frequency in deep sleep in MHZ
+#define VBIOSSMC_MSG_GetDtbclkFreq                0x09 ///< Get display dtb clock frequency in MHZ in case VMIN does not support phy frequency
+#define VBIOSSMC_MSG_SetDtbClk                    0x0A ///< Set dtb clock frequency, return frequemcy in MHZ
+#define VBIOSSMC_MSG_SetDisplayCount              0x0B ///< Inform PMFW of number of display connected
+#define VBIOSSMC_MSG_EnableTmdp48MHzRefclkPwrDown 0x0C ///< To ask PMFW turn off TMDP 48MHz refclk during display off to save power
+#define VBIOSSMC_MSG_UpdatePmeRestore             0x0D ///< To ask PMFW to write into Azalia for PME wake up event
+#define VBIOSSMC_MSG_SetVbiosDramAddrHigh         0x0E ///< Set DRAM address high 32 bits for WM table transfer
+#define VBIOSSMC_MSG_SetVbiosDramAddrLow          0x0F ///< Set DRAM address low 32 bits for WM table transfer
+#define VBIOSSMC_MSG_TransferTableSmu2Dram        0x10 ///< Transfer table from PMFW SRAM to system DRAM
+#define VBIOSSMC_MSG_TransferTableDram2Smu        0x11 ///< Transfer table from system DRAM to PMFW
+#define VBIOSSMC_MSG_SetDisplayIdleOptimizations  0x12 ///< Set Idle state optimization for display off
+#define VBIOSSMC_MSG_GetDprefclkFreq              0x13 ///< Get DPREF clock frequency. Return in MHZ
+#define VBIOSSMC_Message_Count                    0x14 ///< Total number of VBIS and DAL messages
+
+#define VBIOSSMC_Status_BUSY                      0x0
+#define VBIOSSMC_Result_OK                        0x01 ///< Message Response OK
+#define VBIOSSMC_Result_Failed                    0xFF ///< Message Response Failed
+#define VBIOSSMC_Result_UnknownCmd                0xFE ///< Message Response Unknown Command
+#define VBIOSSMC_Result_CmdRejectedPrereq         0xFD ///< Message Response Command Failed Prerequisite
+#define VBIOSSMC_Result_CmdRejectedBusy           0xFC ///< Message Response Command Rejected due to PMFW is busy. Sender should retry sending this message
+
+/*
+ * Function to be used instead of REG_WAIT macro because the wait ends when
+ * the register is NOT EQUAL to zero, and because the translation in msg_if.h
+ * won't work with REG_WAIT.
+ */
+static uint32_t dcn315_smu_wait_for_response(struct clk_mgr_internal *clk_mgr, unsigned int delay_us, unsigned int max_retries)
+{
+	uint32_t res_val = VBIOSSMC_Status_BUSY;
+
+	do {
+		res_val = REG_READ(MP1_SMN_C2PMSG_38);
+		if (res_val != VBIOSSMC_Status_BUSY)
+			break;
+
+		if (delay_us >= 1000)
+			msleep(delay_us/1000);
+		else if (delay_us > 0)
+			udelay(delay_us);
+	} while (max_retries--);
+
+	return res_val;
+}
+
+static int dcn315_smu_send_msg_with_param(
+		struct clk_mgr_internal *clk_mgr,
+		unsigned int msg_id, unsigned int param)
+{
+	uint32_t result;
+	uint32_t i = 0;
+	uint32_t read_back_data;
+
+	result = dcn315_smu_wait_for_response(clk_mgr, 10, 200000);
+
+	if (result != VBIOSSMC_Result_OK)
+		smu_print("SMU Response was not OK. SMU response after wait received is: %d\n", result);
+
+	if (result == VBIOSSMC_Status_BUSY) {
+		return -1;
+	}
+
+	/* First clear response register */
+	REG_WRITE(MP1_SMN_C2PMSG_38, VBIOSSMC_Status_BUSY);
+
+	/* Set the parameter register for the SMU message, unit is Mhz */
+	REG_WRITE(MP1_SMN_C2PMSG_37, param);
+
+	for (i = 0; i < SMU_REGISTER_WRITE_RETRY_COUNT; i++) {
+		/* Trigger the message transaction by writing the message ID */
+		IX_REG_SET_SYNC(mmMP1_C2PMSG_3, 0,
+			MP1_C2PMSG_3, msg_id);
+		IX_REG_GET_SYNC(mmMP1_C2PMSG_3,
+			MP1_C2PMSG_3, &read_back_data);
+		if (read_back_data == msg_id)
+			break;
+		udelay(2);
+		smu_print("SMU msg id write fail %x times. \n", i + 1);
+	}
+
+	result = dcn315_smu_wait_for_response(clk_mgr, 10, 200000);
+
+	if (result == VBIOSSMC_Status_BUSY) {
+		ASSERT(0);
+		dm_helpers_smu_timeout(CTX, msg_id, param, 10 * 200000);
+	}
+
+	return REG_READ(MP1_SMN_C2PMSG_37);
+}
+
+int dcn315_smu_get_smu_version(struct clk_mgr_internal *clk_mgr)
+{
+	return dcn315_smu_send_msg_with_param(
+			clk_mgr,
+			VBIOSSMC_MSG_GetPmfwVersion,
+			0);
+}
+
+
+int dcn315_smu_set_dispclk(struct clk_mgr_internal *clk_mgr, int requested_dispclk_khz)
+{
+	int actual_dispclk_set_mhz = -1;
+
+	if (!clk_mgr->smu_present)
+		return requested_dispclk_khz;
+
+	/*  Unit of SMU msg parameter is Mhz */
+	actual_dispclk_set_mhz = dcn315_smu_send_msg_with_param(
+			clk_mgr,
+			VBIOSSMC_MSG_SetDispclkFreq,
+			khz_to_mhz_ceil(requested_dispclk_khz));
+
+	return actual_dispclk_set_mhz * 1000;
+}
+
+int dcn315_smu_set_hard_min_dcfclk(struct clk_mgr_internal *clk_mgr, int requested_dcfclk_khz)
+{
+	int actual_dcfclk_set_mhz = -1;
+
+	if (!clk_mgr->base.ctx->dc->debug.pstate_enabled)
+		return -1;
+
+	if (!clk_mgr->smu_present)
+		return requested_dcfclk_khz;
+
+	actual_dcfclk_set_mhz = dcn315_smu_send_msg_with_param(
+			clk_mgr,
+			VBIOSSMC_MSG_SetHardMinDcfclkByFreq,
+			khz_to_mhz_ceil(requested_dcfclk_khz));
+
+	return actual_dcfclk_set_mhz * 1000;
+}
+
+int dcn315_smu_set_min_deep_sleep_dcfclk(struct clk_mgr_internal *clk_mgr, int requested_min_ds_dcfclk_khz)
+{
+	int actual_min_ds_dcfclk_mhz = -1;
+
+	if (!clk_mgr->base.ctx->dc->debug.pstate_enabled)
+		return -1;
+
+	if (!clk_mgr->smu_present)
+		return requested_min_ds_dcfclk_khz;
+
+	actual_min_ds_dcfclk_mhz = dcn315_smu_send_msg_with_param(
+			clk_mgr,
+			VBIOSSMC_MSG_SetMinDeepSleepDcfclk,
+			khz_to_mhz_ceil(requested_min_ds_dcfclk_khz));
+
+	return actual_min_ds_dcfclk_mhz * 1000;
+}
+
+int dcn315_smu_set_dppclk(struct clk_mgr_internal *clk_mgr, int requested_dpp_khz)
+{
+	int actual_dppclk_set_mhz = -1;
+
+	if (!clk_mgr->smu_present)
+		return requested_dpp_khz;
+
+	actual_dppclk_set_mhz = dcn315_smu_send_msg_with_param(
+			clk_mgr,
+			VBIOSSMC_MSG_SetDppclkFreq,
+			khz_to_mhz_ceil(requested_dpp_khz));
+
+	return actual_dppclk_set_mhz * 1000;
+}
+
+void dcn315_smu_set_display_idle_optimization(struct clk_mgr_internal *clk_mgr, uint32_t idle_info)
+{
+	if (!clk_mgr->base.ctx->dc->debug.pstate_enabled)
+		return;
+
+	if (!clk_mgr->smu_present)
+		return;
+
+	//TODO: Work with smu team to define optimization options.
+	dcn315_smu_send_msg_with_param(
+		clk_mgr,
+		VBIOSSMC_MSG_SetDisplayIdleOptimizations,
+		idle_info);
+}
+
+void dcn315_smu_enable_phy_refclk_pwrdwn(struct clk_mgr_internal *clk_mgr, bool enable)
+{
+	union display_idle_optimization_u idle_info = { 0 };
+
+	if (!clk_mgr->smu_present)
+		return;
+
+	if (enable) {
+		idle_info.idle_info.df_request_disabled = 1;
+		idle_info.idle_info.phy_ref_clk_off = 1;
+	}
+
+	dcn315_smu_send_msg_with_param(
+			clk_mgr,
+			VBIOSSMC_MSG_SetDisplayIdleOptimizations,
+			idle_info.data);
+}
+
+void dcn315_smu_enable_pme_wa(struct clk_mgr_internal *clk_mgr)
+{
+	if (!clk_mgr->smu_present)
+		return;
+
+	dcn315_smu_send_msg_with_param(
+			clk_mgr,
+			VBIOSSMC_MSG_UpdatePmeRestore,
+			0);
+}
+void dcn315_smu_set_dram_addr_high(struct clk_mgr_internal *clk_mgr, uint32_t addr_high)
+{
+	if (!clk_mgr->smu_present)
+		return;
+
+	dcn315_smu_send_msg_with_param(clk_mgr,
+			VBIOSSMC_MSG_SetVbiosDramAddrHigh, addr_high);
+}
+
+void dcn315_smu_set_dram_addr_low(struct clk_mgr_internal *clk_mgr, uint32_t addr_low)
+{
+	if (!clk_mgr->smu_present)
+		return;
+
+	dcn315_smu_send_msg_with_param(clk_mgr,
+			VBIOSSMC_MSG_SetVbiosDramAddrLow, addr_low);
+}
+
+void dcn315_smu_transfer_dpm_table_smu_2_dram(struct clk_mgr_internal *clk_mgr)
+{
+	if (!clk_mgr->smu_present)
+		return;
+
+	dcn315_smu_send_msg_with_param(clk_mgr,
+			VBIOSSMC_MSG_TransferTableSmu2Dram, TABLE_DPMCLOCKS);
+}
+
+void dcn315_smu_transfer_wm_table_dram_2_smu(struct clk_mgr_internal *clk_mgr)
+{
+	if (!clk_mgr->smu_present)
+		return;
+
+	dcn315_smu_send_msg_with_param(clk_mgr,
+			VBIOSSMC_MSG_TransferTableDram2Smu, TABLE_WATERMARKS);
+}
+
+int dcn315_smu_get_dpref_clk(struct clk_mgr_internal *clk_mgr)
+{
+	int dprefclk_get_mhz = -1;
+	if (clk_mgr->smu_present) {
+		dprefclk_get_mhz = dcn315_smu_send_msg_with_param(
+			clk_mgr,
+			VBIOSSMC_MSG_GetDprefclkFreq,
+			0);
+	}
+	return (dprefclk_get_mhz * 1000);
+}
+
+int dcn315_smu_get_dtbclk(struct clk_mgr_internal *clk_mgr)
+{
+	int fclk_get_mhz = -1;
+
+	if (clk_mgr->smu_present) {
+		fclk_get_mhz = dcn315_smu_send_msg_with_param(
+			clk_mgr,
+			VBIOSSMC_MSG_GetDtbclkFreq,
+			0);
+	}
+	return (fclk_get_mhz * 1000);
+}
+
+void dcn315_smu_set_dtbclk(struct clk_mgr_internal *clk_mgr, bool enable)
+{
+	if (!clk_mgr->smu_present)
+		return;
+
+	dcn315_smu_send_msg_with_param(
+			clk_mgr,
+			VBIOSSMC_MSG_SetDtbClk,
+			enable);
+}
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn315/dcn315_smu.h b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn315/dcn315_smu.h
new file mode 100644
index 000000000000..5aa3275ac7d8
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn315/dcn315_smu.h
@@ -0,0 +1,130 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef DAL_DC_315_SMU_H_
+#define DAL_DC_315_SMU_H_
+#include "os_types.h"
+
+#define PMFW_DRIVER_IF_VERSION 4
+
+#define NUM_DCFCLK_DPM_LEVELS   4
+#define NUM_DISPCLK_DPM_LEVELS  4
+#define NUM_DPPCLK_DPM_LEVELS   4
+#define NUM_SOCCLK_DPM_LEVELS   4
+#define NUM_VCN_DPM_LEVELS      4
+#define NUM_SOC_VOLTAGE_LEVELS  4
+#define NUM_DF_PSTATE_LEVELS    4
+
+
+typedef struct {
+  uint16_t MinClock; // This is either DCFCLK or SOCCLK (in MHz)
+  uint16_t MaxClock; // This is either DCFCLK or SOCCLK (in MHz)
+  uint16_t MinMclk;
+  uint16_t MaxMclk;
+  uint8_t  WmSetting;
+  uint8_t  WmType;  // Used for normal pstate change or memory retraining
+  uint8_t  Padding[2];
+} WatermarkRowGeneric_t;
+
+#define NUM_WM_RANGES 4
+#define WM_PSTATE_CHG 0
+#define WM_RETRAINING 1
+
+typedef enum {
+  WM_SOCCLK = 0,
+  WM_DCFCLK,
+  WM_COUNT,
+} WM_CLOCK_e;
+
+typedef struct {
+  uint32_t FClk;
+  uint32_t MemClk;
+  uint32_t Voltage;
+} DfPstateTable_t;
+
+//Freq in MHz
+//Voltage in milli volts with 2 fractional bits
+typedef struct {
+  uint32_t DcfClocks[NUM_DCFCLK_DPM_LEVELS];
+  uint32_t DispClocks[NUM_DISPCLK_DPM_LEVELS];
+  uint32_t DppClocks[NUM_DPPCLK_DPM_LEVELS];
+  uint32_t SocClocks[NUM_SOCCLK_DPM_LEVELS];
+  uint32_t VClocks[NUM_VCN_DPM_LEVELS];
+  uint32_t DClocks[NUM_VCN_DPM_LEVELS];
+  uint32_t SocVoltage[NUM_SOC_VOLTAGE_LEVELS];
+  DfPstateTable_t DfPstateTable[NUM_DF_PSTATE_LEVELS];
+  uint8_t  NumDcfClkLevelsEnabled;
+  uint8_t  NumDispClkLevelsEnabled; //Applies to both Dispclk and Dppclk
+  uint8_t  NumSocClkLevelsEnabled;
+  uint8_t  VcnClkLevelsEnabled;     //Applies to both Vclk and Dclk
+  uint8_t  NumDfPstatesEnabled;
+  uint8_t  spare[3];
+  uint32_t MinGfxClk;
+  uint32_t MaxGfxClk;
+} DpmClocks_315_t;
+
+struct dcn315_watermarks {
+  // Watermarks
+  WatermarkRowGeneric_t WatermarkRow[WM_COUNT][NUM_WM_RANGES];
+  uint32_t MmHubPadding[7]; // SMU internal use
+};
+
+struct dcn315_smu_dpm_clks {
+	DpmClocks_315_t *dpm_clks;
+	union large_integer mc_address;
+};
+
+#define TABLE_WATERMARKS         1 // Called by DAL through VBIOS
+#define TABLE_DPMCLOCKS          4 // Called by Driver and VBIOS
+
+struct display_idle_optimization {
+	unsigned int df_request_disabled : 1;
+	unsigned int phy_ref_clk_off     : 1;
+	unsigned int s0i2_rdy            : 1;
+	unsigned int reserved            : 29;
+};
+
+union display_idle_optimization_u {
+	struct display_idle_optimization idle_info;
+	uint32_t data;
+};
+
+int dcn315_smu_get_smu_version(struct clk_mgr_internal *clk_mgr);
+int dcn315_smu_set_dispclk(struct clk_mgr_internal *clk_mgr, int requested_dispclk_khz);
+int dcn315_smu_set_hard_min_dcfclk(struct clk_mgr_internal *clk_mgr, int requested_dcfclk_khz);
+int dcn315_smu_set_min_deep_sleep_dcfclk(struct clk_mgr_internal *clk_mgr, int requested_min_ds_dcfclk_khz);
+int dcn315_smu_set_dppclk(struct clk_mgr_internal *clk_mgr, int requested_dpp_khz);
+void dcn315_smu_set_display_idle_optimization(struct clk_mgr_internal *clk_mgr, uint32_t idle_info);
+void dcn315_smu_enable_phy_refclk_pwrdwn(struct clk_mgr_internal *clk_mgr, bool enable);
+void dcn315_smu_set_dram_addr_high(struct clk_mgr_internal *clk_mgr, uint32_t addr_high);
+void dcn315_smu_set_dram_addr_low(struct clk_mgr_internal *clk_mgr, uint32_t addr_low);
+void dcn315_smu_transfer_dpm_table_smu_2_dram(struct clk_mgr_internal *clk_mgr);
+void dcn315_smu_transfer_wm_table_dram_2_smu(struct clk_mgr_internal *clk_mgr);
+void dcn315_smu_request_voltage_via_phyclk(struct clk_mgr_internal *clk_mgr, int requested_phyclk_khz);
+void dcn315_smu_enable_pme_wa(struct clk_mgr_internal *clk_mgr);
+int dcn315_smu_get_dpref_clk(struct clk_mgr_internal *clk_mgr);
+int dcn315_smu_get_dtbclk(struct clk_mgr_internal *clk_mgr);
+void dcn315_smu_set_dtbclk(struct clk_mgr_internal *clk_mgr, bool enable);
+#endif /* DAL_DC_315_SMU_H_ */
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn316/dcn316_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn316/dcn316_clk_mgr.c
new file mode 100644
index 000000000000..1769b1f26e75
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn316/dcn316_clk_mgr.c
@@ -0,0 +1,680 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+
+
+#include "dccg.h"
+#include "clk_mgr_internal.h"
+
+// For dce12_get_dp_ref_freq_khz
+#include "dce100/dce_clk_mgr.h"
+// For dcn20_update_clocks_update_dpp_dto
+#include "dcn20/dcn20_clk_mgr.h"
+#include "dcn31/dcn31_clk_mgr.h"
+#include "dcn316_clk_mgr.h"
+#include "reg_helper.h"
+#include "core_types.h"
+#include "dcn316_smu.h"
+#include "dm_helpers.h"
+#include "dc_dmub_srv.h"
+#include "link_service.h"
+
+// DCN316 this is CLK1 instance
+#define MAX_INSTANCE                                        7
+#define MAX_SEGMENT                                         6
+
+struct IP_BASE_INSTANCE {
+    unsigned int segment[MAX_SEGMENT];
+};
+
+struct IP_BASE {
+    struct IP_BASE_INSTANCE instance[MAX_INSTANCE];
+};
+
+#define regCLK1_CLK_PLL_REQ						0x0237
+#define regCLK1_CLK_PLL_REQ_BASE_IDX			0
+
+#define CLK1_CLK_PLL_REQ__FbMult_int__SHIFT		0x0
+#define CLK1_CLK_PLL_REQ__PllSpineDiv__SHIFT	0xc
+#define CLK1_CLK_PLL_REQ__FbMult_frac__SHIFT	0x10
+#define CLK1_CLK_PLL_REQ__FbMult_int_MASK		0x000001FFL
+#define CLK1_CLK_PLL_REQ__PllSpineDiv_MASK		0x0000F000L
+#define CLK1_CLK_PLL_REQ__FbMult_frac_MASK		0xFFFF0000L
+
+#define TO_CLK_MGR_DCN316(clk_mgr)\
+	container_of(clk_mgr, struct clk_mgr_dcn316, base)
+
+static int dcn316_get_active_display_cnt_wa(
+		struct dc *dc,
+		struct dc_state *context)
+{
+	int i, display_count;
+	bool tmds_present = false;
+
+	display_count = 0;
+	for (i = 0; i < context->stream_count; i++) {
+		const struct dc_stream_state *stream = context->streams[i];
+
+		if (stream->signal == SIGNAL_TYPE_HDMI_TYPE_A ||
+				stream->signal == SIGNAL_TYPE_DVI_SINGLE_LINK ||
+				stream->signal == SIGNAL_TYPE_DVI_DUAL_LINK)
+			tmds_present = true;
+	}
+
+	for (i = 0; i < dc->link_count; i++) {
+		const struct dc_link *link = dc->links[i];
+
+		/* abusing the fact that the dig and phy are coupled to see if the phy is enabled */
+		if (link->link_enc && link->link_enc->funcs->is_dig_enabled &&
+				link->link_enc->funcs->is_dig_enabled(link->link_enc))
+			display_count++;
+	}
+
+	/* WA for hang on HDMI after display off back back on*/
+	if (display_count == 0 && tmds_present)
+		display_count = 1;
+
+	return display_count;
+}
+
+static void dcn316_disable_otg_wa(struct clk_mgr *clk_mgr_base, struct dc_state *context,
+		bool safe_to_lower, bool disable)
+{
+	struct dc *dc = clk_mgr_base->ctx->dc;
+	int i;
+
+	for (i = 0; i < dc->res_pool->pipe_count; ++i) {
+		struct pipe_ctx *pipe = safe_to_lower
+			? &context->res_ctx.pipe_ctx[i]
+			: &dc->current_state->res_ctx.pipe_ctx[i];
+
+		if (pipe->top_pipe || pipe->prev_odm_pipe)
+			continue;
+		if (pipe->stream && (pipe->stream->dpms_off || dc_is_virtual_signal(pipe->stream->signal) ||
+				     !pipe->stream->link_enc)) {
+			if (disable) {
+				if (pipe->stream_res.tg && pipe->stream_res.tg->funcs->immediate_disable_crtc)
+					pipe->stream_res.tg->funcs->immediate_disable_crtc(pipe->stream_res.tg);
+
+				reset_sync_context_for_pipe(dc, context, i);
+			} else
+				pipe->stream_res.tg->funcs->enable_crtc(pipe->stream_res.tg);
+		}
+	}
+}
+
+static void dcn316_enable_pme_wa(struct clk_mgr *clk_mgr_base)
+{
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+
+	dcn316_smu_enable_pme_wa(clk_mgr);
+}
+
+static void dcn316_update_clocks(struct clk_mgr *clk_mgr_base,
+			struct dc_state *context,
+			bool safe_to_lower)
+{
+	union dmub_rb_cmd cmd;
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+	struct dc_clocks *new_clocks = &context->bw_ctx.bw.dcn.clk;
+	struct dc *dc = clk_mgr_base->ctx->dc;
+	int display_count = 0;
+	bool update_dppclk = false;
+	bool update_dispclk = false;
+	bool dpp_clock_lowered = false;
+
+	if (dc->work_arounds.skip_clock_update)
+		return;
+
+	/*
+	 * if it is safe to lower, but we are already in the lower state, we don't have to do anything
+	 * also if safe to lower is false, we just go in the higher state
+	 */
+	clk_mgr_base->clks.zstate_support = new_clocks->zstate_support;
+	if (safe_to_lower) {
+		if (clk_mgr_base->clks.dtbclk_en && !new_clocks->dtbclk_en) {
+			dcn316_smu_set_dtbclk(clk_mgr, false);
+			clk_mgr_base->clks.dtbclk_en = new_clocks->dtbclk_en;
+		}
+		/* check that we're not already in lower */
+		if (clk_mgr_base->clks.pwr_state != DCN_PWR_STATE_LOW_POWER) {
+			display_count = dcn316_get_active_display_cnt_wa(dc, context);
+			/* if we can go lower, go lower */
+			if (display_count == 0) {
+				union display_idle_optimization_u idle_info = { 0 };
+				idle_info.idle_info.df_request_disabled = 1;
+				idle_info.idle_info.phy_ref_clk_off = 1;
+				idle_info.idle_info.s0i2_rdy = 1;
+				dcn316_smu_set_display_idle_optimization(clk_mgr, idle_info.data);
+				/* update power state */
+				clk_mgr_base->clks.pwr_state = DCN_PWR_STATE_LOW_POWER;
+			}
+		}
+	} else {
+		if (!clk_mgr_base->clks.dtbclk_en && new_clocks->dtbclk_en) {
+			dcn316_smu_set_dtbclk(clk_mgr, true);
+			clk_mgr_base->clks.dtbclk_en = new_clocks->dtbclk_en;
+		}
+
+		/* check that we're not already in D0 */
+		if (clk_mgr_base->clks.pwr_state != DCN_PWR_STATE_MISSION_MODE) {
+			union display_idle_optimization_u idle_info = { 0 };
+			dcn316_smu_set_display_idle_optimization(clk_mgr, idle_info.data);
+			/* update power state */
+			clk_mgr_base->clks.pwr_state = DCN_PWR_STATE_MISSION_MODE;
+		}
+	}
+
+	if (should_set_clock(safe_to_lower, new_clocks->dcfclk_khz, clk_mgr_base->clks.dcfclk_khz)) {
+		clk_mgr_base->clks.dcfclk_khz = new_clocks->dcfclk_khz;
+		dcn316_smu_set_hard_min_dcfclk(clk_mgr, clk_mgr_base->clks.dcfclk_khz);
+	}
+
+	if (should_set_clock(safe_to_lower,
+			new_clocks->dcfclk_deep_sleep_khz, clk_mgr_base->clks.dcfclk_deep_sleep_khz)) {
+		clk_mgr_base->clks.dcfclk_deep_sleep_khz = new_clocks->dcfclk_deep_sleep_khz;
+		dcn316_smu_set_min_deep_sleep_dcfclk(clk_mgr, clk_mgr_base->clks.dcfclk_deep_sleep_khz);
+	}
+
+	// workaround: Limit dppclk to 100Mhz to avoid lower eDP panel switch to plus 4K monitor underflow.
+	if (new_clocks->dppclk_khz < 100000)
+		new_clocks->dppclk_khz = 100000;
+
+	if (should_set_clock(safe_to_lower, new_clocks->dppclk_khz, clk_mgr->base.clks.dppclk_khz)) {
+		if (clk_mgr->base.clks.dppclk_khz > new_clocks->dppclk_khz)
+			dpp_clock_lowered = true;
+		clk_mgr_base->clks.dppclk_khz = new_clocks->dppclk_khz;
+		update_dppclk = true;
+	}
+
+	if (should_set_clock(safe_to_lower, new_clocks->dispclk_khz, clk_mgr_base->clks.dispclk_khz) &&
+	    (new_clocks->dispclk_khz > 0 || (safe_to_lower && display_count == 0))) {
+		int requested_dispclk_khz = new_clocks->dispclk_khz;
+
+		dcn316_disable_otg_wa(clk_mgr_base, context, safe_to_lower, true);
+
+		/* Clamp the requested clock to PMFW based on their limit. */
+		if (dc->debug.min_disp_clk_khz > 0 && requested_dispclk_khz < dc->debug.min_disp_clk_khz)
+			requested_dispclk_khz = dc->debug.min_disp_clk_khz;
+
+		dcn316_smu_set_dispclk(clk_mgr, requested_dispclk_khz);
+		clk_mgr_base->clks.dispclk_khz = new_clocks->dispclk_khz;
+		dcn316_disable_otg_wa(clk_mgr_base, context, safe_to_lower, false);
+
+		update_dispclk = true;
+	}
+
+	if (dpp_clock_lowered) {
+		// increase per DPP DTO before lowering global dppclk
+		dcn20_update_clocks_update_dpp_dto(clk_mgr, context, safe_to_lower);
+		dcn316_smu_set_dppclk(clk_mgr, clk_mgr_base->clks.dppclk_khz);
+	} else {
+		// increase global DPPCLK before lowering per DPP DTO
+		if (update_dppclk || update_dispclk)
+			dcn316_smu_set_dppclk(clk_mgr, clk_mgr_base->clks.dppclk_khz);
+		// always update dtos unless clock is lowered and not safe to lower
+		if (new_clocks->dppclk_khz >= dc->current_state->bw_ctx.bw.dcn.clk.dppclk_khz)
+			dcn20_update_clocks_update_dpp_dto(clk_mgr, context, safe_to_lower);
+	}
+
+	// notify DMCUB of latest clocks
+	memset(&cmd, 0, sizeof(cmd));
+	cmd.notify_clocks.header.type = DMUB_CMD__CLK_MGR;
+	cmd.notify_clocks.header.sub_type = DMUB_CMD__CLK_MGR_NOTIFY_CLOCKS;
+	cmd.notify_clocks.clocks.dcfclk_khz = clk_mgr_base->clks.dcfclk_khz;
+	cmd.notify_clocks.clocks.dcfclk_deep_sleep_khz =
+		clk_mgr_base->clks.dcfclk_deep_sleep_khz;
+	cmd.notify_clocks.clocks.dispclk_khz = clk_mgr_base->clks.dispclk_khz;
+	cmd.notify_clocks.clocks.dppclk_khz = clk_mgr_base->clks.dppclk_khz;
+
+	dc_wake_and_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
+}
+
+static void dcn316_dump_clk_registers(struct clk_state_registers_and_bypass *regs_and_bypass,
+		struct clk_mgr *clk_mgr_base, struct clk_log_info *log_info)
+{
+	return;
+}
+
+static struct clk_bw_params dcn316_bw_params = {
+	.vram_type = Ddr4MemType,
+	.num_channels = 1,
+	.clk_table = {
+		.num_entries = 5,
+	},
+
+};
+
+static struct wm_table ddr4_wm_table = {
+	.entries = {
+		{
+			.wm_inst = WM_A,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.72,
+			.sr_exit_time_us = 6.09,
+			.sr_enter_plus_exit_time_us = 7.14,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_B,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.72,
+			.sr_exit_time_us = 10.12,
+			.sr_enter_plus_exit_time_us = 11.48,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_C,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.72,
+			.sr_exit_time_us = 10.12,
+			.sr_enter_plus_exit_time_us = 11.48,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_D,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.72,
+			.sr_exit_time_us = 10.12,
+			.sr_enter_plus_exit_time_us = 11.48,
+			.valid = true,
+		},
+	}
+};
+
+static struct wm_table lpddr5_wm_table = {
+	.entries = {
+		{
+			.wm_inst = WM_A,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.65333,
+			.sr_exit_time_us = 11.5,
+			.sr_enter_plus_exit_time_us = 14.5,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_B,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.65333,
+			.sr_exit_time_us = 11.5,
+			.sr_enter_plus_exit_time_us = 14.5,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_C,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.65333,
+			.sr_exit_time_us = 11.5,
+			.sr_enter_plus_exit_time_us = 14.5,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_D,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.65333,
+			.sr_exit_time_us = 11.5,
+			.sr_enter_plus_exit_time_us = 14.5,
+			.valid = true,
+		},
+	}
+};
+
+static DpmClocks_316_t dummy_clocks;
+
+static struct dcn316_watermarks dummy_wms = { 0 };
+
+static void dcn316_build_watermark_ranges(struct clk_bw_params *bw_params, struct dcn316_watermarks *table)
+{
+	int i, num_valid_sets;
+
+	num_valid_sets = 0;
+
+	for (i = 0; i < WM_SET_COUNT; i++) {
+		/* skip empty entries, the smu array has no holes*/
+		if (!bw_params->wm_table.entries[i].valid)
+			continue;
+
+		table->WatermarkRow[WM_DCFCLK][num_valid_sets].WmSetting = bw_params->wm_table.entries[i].wm_inst;
+		table->WatermarkRow[WM_DCFCLK][num_valid_sets].WmType = bw_params->wm_table.entries[i].wm_type;
+		/* We will not select WM based on fclk, so leave it as unconstrained */
+		table->WatermarkRow[WM_DCFCLK][num_valid_sets].MinClock = 0;
+		table->WatermarkRow[WM_DCFCLK][num_valid_sets].MaxClock = 0xFFFF;
+
+		if (table->WatermarkRow[WM_DCFCLK][num_valid_sets].WmType == WM_TYPE_PSTATE_CHG) {
+			if (i == 0)
+				table->WatermarkRow[WM_DCFCLK][num_valid_sets].MinMclk = 0;
+			else {
+				/* add 1 to make it non-overlapping with next lvl */
+				table->WatermarkRow[WM_DCFCLK][num_valid_sets].MinMclk =
+						bw_params->clk_table.entries[i - 1].dcfclk_mhz + 1;
+			}
+			table->WatermarkRow[WM_DCFCLK][num_valid_sets].MaxMclk =
+					bw_params->clk_table.entries[i].dcfclk_mhz;
+
+		} else {
+			/* unconstrained for memory retraining */
+			table->WatermarkRow[WM_DCFCLK][num_valid_sets].MinClock = 0;
+			table->WatermarkRow[WM_DCFCLK][num_valid_sets].MaxClock = 0xFFFF;
+
+			/* Modify previous watermark range to cover up to max */
+			table->WatermarkRow[WM_DCFCLK][num_valid_sets - 1].MaxClock = 0xFFFF;
+		}
+		num_valid_sets++;
+	}
+
+	ASSERT(num_valid_sets != 0); /* Must have at least one set of valid watermarks */
+
+	/* modify the min and max to make sure we cover the whole range*/
+	table->WatermarkRow[WM_DCFCLK][0].MinMclk = 0;
+	table->WatermarkRow[WM_DCFCLK][0].MinClock = 0;
+	table->WatermarkRow[WM_DCFCLK][num_valid_sets - 1].MaxMclk = 0xFFFF;
+	table->WatermarkRow[WM_DCFCLK][num_valid_sets - 1].MaxClock = 0xFFFF;
+
+	/* This is for writeback only, does not matter currently as no writeback support*/
+	table->WatermarkRow[WM_SOCCLK][0].WmSetting = WM_A;
+	table->WatermarkRow[WM_SOCCLK][0].MinClock = 0;
+	table->WatermarkRow[WM_SOCCLK][0].MaxClock = 0xFFFF;
+	table->WatermarkRow[WM_SOCCLK][0].MinMclk = 0;
+	table->WatermarkRow[WM_SOCCLK][0].MaxMclk = 0xFFFF;
+}
+
+static void dcn316_notify_wm_ranges(struct clk_mgr *clk_mgr_base)
+{
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+	struct clk_mgr_dcn316 *clk_mgr_dcn316 = TO_CLK_MGR_DCN316(clk_mgr);
+	struct dcn316_watermarks *table = clk_mgr_dcn316->smu_wm_set.wm_set;
+
+	if (!clk_mgr->smu_ver)
+		return;
+
+	if (!table || clk_mgr_dcn316->smu_wm_set.mc_address.quad_part == 0)
+		return;
+
+	memset(table, 0, sizeof(*table));
+
+	dcn316_build_watermark_ranges(clk_mgr_base->bw_params, table);
+
+	dcn316_smu_set_dram_addr_high(clk_mgr,
+			clk_mgr_dcn316->smu_wm_set.mc_address.high_part);
+	dcn316_smu_set_dram_addr_low(clk_mgr,
+			clk_mgr_dcn316->smu_wm_set.mc_address.low_part);
+	dcn316_smu_transfer_wm_table_dram_2_smu(clk_mgr);
+}
+
+static void dcn316_get_dpm_table_from_smu(struct clk_mgr_internal *clk_mgr,
+		struct dcn316_smu_dpm_clks *smu_dpm_clks)
+{
+	DpmClocks_316_t *table = smu_dpm_clks->dpm_clks;
+
+	if (!clk_mgr->smu_ver)
+		return;
+
+	if (!table || smu_dpm_clks->mc_address.quad_part == 0)
+		return;
+
+	memset(table, 0, sizeof(*table));
+
+	dcn316_smu_set_dram_addr_high(clk_mgr,
+			smu_dpm_clks->mc_address.high_part);
+	dcn316_smu_set_dram_addr_low(clk_mgr,
+			smu_dpm_clks->mc_address.low_part);
+	dcn316_smu_transfer_dpm_table_smu_2_dram(clk_mgr);
+}
+
+static uint32_t find_max_clk_value(const uint32_t clocks[], uint32_t num_clocks)
+{
+	uint32_t max = 0;
+	int i;
+
+	for (i = 0; i < num_clocks; ++i) {
+		if (clocks[i] > max)
+			max = clocks[i];
+	}
+
+	return max;
+}
+
+static unsigned int find_clk_for_voltage(
+		const DpmClocks_316_t *clock_table,
+		const uint32_t clocks[],
+		unsigned int voltage)
+{
+	int i;
+	int max_voltage = 0;
+	int clock = 0;
+
+	for (i = 0; i < NUM_SOC_VOLTAGE_LEVELS; i++) {
+		if (clock_table->SocVoltage[i] == voltage) {
+			return clocks[i];
+		} else if (clock_table->SocVoltage[i] >= max_voltage &&
+				clock_table->SocVoltage[i] < voltage) {
+			max_voltage = clock_table->SocVoltage[i];
+			clock = clocks[i];
+		}
+	}
+
+	ASSERT(clock);
+	return clock;
+}
+
+static void dcn316_clk_mgr_helper_populate_bw_params(
+		struct clk_mgr_internal *clk_mgr,
+		struct integrated_info *bios_info,
+		const DpmClocks_316_t *clock_table)
+{
+	int i, j;
+	struct clk_bw_params *bw_params = clk_mgr->base.bw_params;
+	uint32_t max_dispclk = 0, max_dppclk = 0;
+
+	j = -1;
+
+	static_assert(NUM_DF_PSTATE_LEVELS <= MAX_NUM_DPM_LVL,
+		"number of reported pstate levels exceeds maximum");
+
+	/* Find lowest DPM, FCLK is filled in reverse order*/
+
+	for (i = NUM_DF_PSTATE_LEVELS - 1; i >= 0; i--) {
+		if (clock_table->DfPstateTable[i].FClk != 0) {
+			j = i;
+			break;
+		}
+	}
+
+	if (j == -1) {
+		/* clock table is all 0s, just use our own hardcode */
+		ASSERT(0);
+		return;
+	}
+
+	bw_params->clk_table.num_entries = j + 1;
+
+	/* dispclk and dppclk can be max at any voltage, same number of levels for both */
+	if (clock_table->NumDispClkLevelsEnabled <= NUM_DISPCLK_DPM_LEVELS &&
+	    clock_table->NumDispClkLevelsEnabled <= NUM_DPPCLK_DPM_LEVELS) {
+		max_dispclk = find_max_clk_value(clock_table->DispClocks, clock_table->NumDispClkLevelsEnabled);
+		max_dppclk = find_max_clk_value(clock_table->DppClocks, clock_table->NumDispClkLevelsEnabled);
+	} else {
+		ASSERT(0);
+	}
+
+	for (i = 0; i < bw_params->clk_table.num_entries; i++, j--) {
+		int temp;
+
+		bw_params->clk_table.entries[i].fclk_mhz = clock_table->DfPstateTable[j].FClk;
+		bw_params->clk_table.entries[i].memclk_mhz = clock_table->DfPstateTable[j].MemClk;
+		bw_params->clk_table.entries[i].voltage = clock_table->DfPstateTable[j].Voltage;
+		switch (clock_table->DfPstateTable[j].WckRatio) {
+		case WCK_RATIO_1_2:
+			bw_params->clk_table.entries[i].wck_ratio = 2;
+			break;
+		case WCK_RATIO_1_4:
+			bw_params->clk_table.entries[i].wck_ratio = 4;
+			break;
+		default:
+			bw_params->clk_table.entries[i].wck_ratio = 1;
+		}
+		temp = find_clk_for_voltage(clock_table, clock_table->DcfClocks, clock_table->DfPstateTable[j].Voltage);
+		if (temp)
+			bw_params->clk_table.entries[i].dcfclk_mhz = temp;
+		temp = find_clk_for_voltage(clock_table, clock_table->SocClocks, clock_table->DfPstateTable[j].Voltage);
+		if (temp)
+			bw_params->clk_table.entries[i].socclk_mhz = temp;
+		bw_params->clk_table.entries[i].dispclk_mhz = max_dispclk;
+		bw_params->clk_table.entries[i].dppclk_mhz = max_dppclk;
+	}
+
+	bw_params->vram_type = bios_info->memory_type;
+	bw_params->num_channels = bios_info->ma_channel_number;
+	bw_params->dram_channel_width_bytes = bios_info->memory_type == 0x22 ? 8 : 4;
+
+	for (i = 0; i < WM_SET_COUNT; i++) {
+		bw_params->wm_table.entries[i].wm_inst = i;
+
+		if (i >= bw_params->clk_table.num_entries) {
+			bw_params->wm_table.entries[i].valid = false;
+			continue;
+		}
+
+		bw_params->wm_table.entries[i].wm_type = WM_TYPE_PSTATE_CHG;
+		bw_params->wm_table.entries[i].valid = true;
+	}
+}
+
+
+
+static struct clk_mgr_funcs dcn316_funcs = {
+	.enable_pme_wa = dcn316_enable_pme_wa,
+	.get_dp_ref_clk_frequency = dce12_get_dp_ref_freq_khz,
+	.get_dtb_ref_clk_frequency = dcn31_get_dtb_ref_freq_khz,
+	.update_clocks = dcn316_update_clocks,
+	.init_clocks = dcn31_init_clocks,
+	.are_clock_states_equal = dcn31_are_clock_states_equal,
+	.notify_wm_ranges = dcn316_notify_wm_ranges
+};
+extern struct clk_mgr_funcs dcn3_fpga_funcs;
+
+void dcn316_clk_mgr_construct(
+		struct dc_context *ctx,
+		struct clk_mgr_dcn316 *clk_mgr,
+		struct pp_smu_funcs *pp_smu,
+		struct dccg *dccg)
+{
+	struct dcn316_smu_dpm_clks smu_dpm_clks = { 0 };
+	struct clk_log_info log_info = {0};
+
+	clk_mgr->base.base.ctx = ctx;
+	clk_mgr->base.base.funcs = &dcn316_funcs;
+
+	clk_mgr->base.pp_smu = pp_smu;
+
+	clk_mgr->base.dccg = dccg;
+	clk_mgr->base.dfs_bypass_disp_clk = 0;
+
+	clk_mgr->base.dprefclk_ss_percentage = 0;
+	clk_mgr->base.dprefclk_ss_divider = 1000;
+	clk_mgr->base.ss_on_dprefclk = false;
+	clk_mgr->base.dfs_ref_freq_khz = 48000;
+
+	clk_mgr->smu_wm_set.wm_set = (struct dcn316_watermarks *)dm_helpers_allocate_gpu_mem(
+				clk_mgr->base.base.ctx,
+				DC_MEM_ALLOC_TYPE_FRAME_BUFFER,
+				sizeof(struct dcn316_watermarks),
+				&clk_mgr->smu_wm_set.mc_address.quad_part);
+
+	if (!clk_mgr->smu_wm_set.wm_set) {
+		clk_mgr->smu_wm_set.wm_set = &dummy_wms;
+		clk_mgr->smu_wm_set.mc_address.quad_part = 0;
+	}
+	ASSERT(clk_mgr->smu_wm_set.wm_set);
+
+	smu_dpm_clks.dpm_clks = (DpmClocks_316_t *)dm_helpers_allocate_gpu_mem(
+				clk_mgr->base.base.ctx,
+				DC_MEM_ALLOC_TYPE_FRAME_BUFFER,
+				sizeof(DpmClocks_316_t),
+				&smu_dpm_clks.mc_address.quad_part);
+
+	if (smu_dpm_clks.dpm_clks == NULL) {
+		smu_dpm_clks.dpm_clks = &dummy_clocks;
+		smu_dpm_clks.mc_address.quad_part = 0;
+	}
+
+	ASSERT(smu_dpm_clks.dpm_clks);
+
+	clk_mgr->base.smu_ver = dcn316_smu_get_smu_version(&clk_mgr->base);
+
+	if (clk_mgr->base.smu_ver > 0)
+		clk_mgr->base.smu_present = true;
+
+	// Skip this for now as it did not work on DCN315, renable during bring up
+	//clk_mgr->base.base.dentist_vco_freq_khz = get_vco_frequency_from_reg(&clk_mgr->base);
+	clk_mgr->base.base.dentist_vco_freq_khz = 2500000;
+
+	/* in case we don't get a value from the register, use default */
+	if (clk_mgr->base.base.dentist_vco_freq_khz == 0)
+		clk_mgr->base.base.dentist_vco_freq_khz = 2500000; /* 2400MHz */
+
+
+	if (ctx->dc_bios->integrated_info->memory_type == LpDdr5MemType) {
+		dcn316_bw_params.wm_table = lpddr5_wm_table;
+	} else {
+		dcn316_bw_params.wm_table = ddr4_wm_table;
+	}
+	/* Saved clocks configured at boot for debug purposes */
+	dcn316_dump_clk_registers(&clk_mgr->base.base.boot_snapshot,
+				  &clk_mgr->base.base, &log_info);
+
+	clk_mgr->base.base.dprefclk_khz = 600000;
+	clk_mgr->base.base.dprefclk_khz = dcn316_smu_get_dpref_clk(&clk_mgr->base);
+	clk_mgr->base.base.clks.ref_dtbclk_khz = clk_mgr->base.base.dprefclk_khz;
+	dce_clock_read_ss_info(&clk_mgr->base);
+	/*clk_mgr->base.dccg->ref_dtbclk_khz =
+	dce_adjust_dp_ref_freq_for_ss(&clk_mgr->base, clk_mgr->base.base.dprefclk_khz);*/
+
+	clk_mgr->base.base.bw_params = &dcn316_bw_params;
+
+	if (clk_mgr->base.base.ctx->dc->debug.pstate_enabled) {
+		dcn316_get_dpm_table_from_smu(&clk_mgr->base, &smu_dpm_clks);
+
+		if (ctx->dc_bios->integrated_info) {
+			dcn316_clk_mgr_helper_populate_bw_params(
+					&clk_mgr->base,
+					ctx->dc_bios->integrated_info,
+					smu_dpm_clks.dpm_clks);
+		}
+	}
+
+	if (smu_dpm_clks.dpm_clks && smu_dpm_clks.mc_address.quad_part != 0)
+		dm_helpers_free_gpu_mem(clk_mgr->base.base.ctx, DC_MEM_ALLOC_TYPE_FRAME_BUFFER,
+				smu_dpm_clks.dpm_clks);
+}
+
+void dcn316_clk_mgr_destroy(struct clk_mgr_internal *clk_mgr_int)
+{
+	struct clk_mgr_dcn316 *clk_mgr = TO_CLK_MGR_DCN316(clk_mgr_int);
+
+	if (clk_mgr->smu_wm_set.wm_set && clk_mgr->smu_wm_set.mc_address.quad_part != 0)
+		dm_helpers_free_gpu_mem(clk_mgr_int->base.ctx, DC_MEM_ALLOC_TYPE_FRAME_BUFFER,
+				clk_mgr->smu_wm_set.wm_set);
+}
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn316/dcn316_clk_mgr.h b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn316/dcn316_clk_mgr.h
new file mode 100644
index 000000000000..864d1f6cef26
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn316/dcn316_clk_mgr.h
@@ -0,0 +1,49 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __DCN316_CLK_MGR_H__
+#define __DCN316_CLK_MGR_H__
+#include "clk_mgr_internal.h"
+
+struct dcn316_watermarks;
+
+struct dcn316_smu_watermark_set {
+	struct dcn316_watermarks *wm_set;
+	union large_integer mc_address;
+};
+
+struct clk_mgr_dcn316 {
+	struct clk_mgr_internal base;
+	struct dcn316_smu_watermark_set smu_wm_set;
+};
+
+void dcn316_clk_mgr_construct(struct dc_context *ctx,
+		struct clk_mgr_dcn316 *clk_mgr,
+		struct pp_smu_funcs *pp_smu,
+		struct dccg *dccg);
+
+void dcn316_clk_mgr_destroy(struct clk_mgr_internal *clk_mgr_int);
+
+#endif //__DCN316_CLK_MGR_H__
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn316/dcn316_smu.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn316/dcn316_smu.c
new file mode 100644
index 000000000000..8b82092b91cd
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn316/dcn316_smu.c
@@ -0,0 +1,344 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#include "core_types.h"
+#include "clk_mgr_internal.h"
+#include "reg_helper.h"
+#include "dm_helpers.h"
+#include "dcn316_smu.h"
+#include "mp/mp_13_0_8_offset.h"
+#include "mp/mp_13_0_8_sh_mask.h"
+
+#define MAX_INSTANCE                                        7
+#define MAX_SEGMENT                                         6
+
+struct IP_BASE_INSTANCE {
+    unsigned int segment[MAX_SEGMENT];
+};
+
+struct IP_BASE {
+    struct IP_BASE_INSTANCE instance[MAX_INSTANCE];
+};
+
+static const struct IP_BASE MP0_BASE = { { { { 0x00016000, 0x00DC0000, 0x00E00000, 0x00E40000, 0x0243FC00, 0 } },
+					{ { 0, 0, 0, 0, 0, 0 } },
+					{ { 0, 0, 0, 0, 0, 0 } },
+					{ { 0, 0, 0, 0, 0, 0 } },
+					{ { 0, 0, 0, 0, 0, 0 } },
+					{ { 0, 0, 0, 0, 0, 0 } },
+					{ { 0, 0, 0, 0, 0, 0 } } } };
+
+#define REG(reg_name) \
+	(MP0_BASE.instance[0].segment[reg ## reg_name ## _BASE_IDX] + reg ## reg_name)
+
+#define FN(reg_name, field) \
+	FD(reg_name##__##field)
+
+#include "logger_types.h"
+#undef DC_LOGGER
+#define DC_LOGGER \
+	CTX->logger
+#define smu_print(str, ...) {DC_LOG_SMU(str, ##__VA_ARGS__); }
+
+#define VBIOSSMC_MSG_TestMessage                  0x01 ///< To check if PMFW is alive and responding. Requirement specified by PMFW team
+#define VBIOSSMC_MSG_GetPmfwVersion               0x02 ///< Get PMFW version
+#define VBIOSSMC_MSG_Spare0                       0x03 ///< Spare0
+#define VBIOSSMC_MSG_SetDispclkFreq               0x04 ///< Set display clock frequency in MHZ
+#define VBIOSSMC_MSG_Spare1                       0x05 ///< Spare1
+#define VBIOSSMC_MSG_SetDppclkFreq                0x06 ///< Set DPP clock frequency in MHZ
+#define VBIOSSMC_MSG_SetHardMinDcfclkByFreq       0x07 ///< Set DCF clock frequency hard min in MHZ
+#define VBIOSSMC_MSG_SetMinDeepSleepDcfclk        0x08 ///< Set DCF clock minimum frequency in deep sleep in MHZ
+#define VBIOSSMC_MSG_SetPhyclkVoltageByFreq       0x09 ///< Set display phy clock frequency in MHZ in case VMIN does not support phy frequency
+#define VBIOSSMC_MSG_GetFclkFrequency             0x0A ///< Get FCLK frequency, return frequemcy in MHZ
+#define VBIOSSMC_MSG_SetDisplayCount              0x0B ///< Inform PMFW of number of display connected
+#define VBIOSSMC_MSG_SPARE						  0x0C ///< SPARE
+#define VBIOSSMC_MSG_UpdatePmeRestore             0x0D ///< To ask PMFW to write into Azalia for PME wake up event
+#define VBIOSSMC_MSG_SetVbiosDramAddrHigh         0x0E ///< Set DRAM address high 32 bits for WM table transfer
+#define VBIOSSMC_MSG_SetVbiosDramAddrLow          0x0F ///< Set DRAM address low 32 bits for WM table transfer
+#define VBIOSSMC_MSG_TransferTableSmu2Dram        0x10 ///< Transfer table from PMFW SRAM to system DRAM
+#define VBIOSSMC_MSG_TransferTableDram2Smu        0x11 ///< Transfer table from system DRAM to PMFW
+#define VBIOSSMC_MSG_SetDisplayIdleOptimizations  0x12 ///< Set Idle state optimization for display off
+#define VBIOSSMC_MSG_GetDprefclkFreq              0x13 ///< Get DPREF clock frequency. Return in MHZ
+#define VBIOSSMC_MSG_GetDtbclkFreq                0x14 ///< Get DPREF clock frequency. Return in MHZ
+#define VBIOSSMC_MSG_SetDtbclkFreq                0x15 ///< Inform PMFW to turn on/off DTB clock arg = 1, turn DTB clock on 600MHz/ arg = 0 turn DTB clock off
+#define VBIOSSMC_Message_Count                    0x16 ///< Total number of VBIS and DAL messages
+
+#define VBIOSSMC_Status_BUSY                      0x0
+#define VBIOSSMC_Result_OK                        0x01 ///< Message Response OK
+#define VBIOSSMC_Result_Failed                    0xFF ///< Message Response Failed
+#define VBIOSSMC_Result_UnknownCmd                0xFE ///< Message Response Unknown Command
+#define VBIOSSMC_Result_CmdRejectedPrereq         0xFD ///< Message Response Command Failed Prerequisite
+#define VBIOSSMC_Result_CmdRejectedBusy           0xFC ///< Message Response Command Rejected due to PMFW is busy. Sender should retry sending this message
+
+/*
+ * Function to be used instead of REG_WAIT macro because the wait ends when
+ * the register is NOT EQUAL to zero, and because the translation in msg_if.h
+ * won't work with REG_WAIT.
+ */
+static uint32_t dcn316_smu_wait_for_response(struct clk_mgr_internal *clk_mgr, unsigned int delay_us, unsigned int max_retries)
+{
+	uint32_t res_val = VBIOSSMC_Status_BUSY;
+
+	do {
+		res_val = REG_READ(MP1_SMN_C2PMSG_91);
+		if (res_val != VBIOSSMC_Status_BUSY)
+			break;
+
+		if (delay_us >= 1000)
+			msleep(delay_us/1000);
+		else if (delay_us > 0)
+			udelay(delay_us);
+	} while (max_retries--);
+
+	return res_val;
+}
+
+static int dcn316_smu_send_msg_with_param(
+		struct clk_mgr_internal *clk_mgr,
+		unsigned int msg_id, unsigned int param)
+{
+	uint32_t result;
+
+	result = dcn316_smu_wait_for_response(clk_mgr, 10, 200000);
+
+	if (result != VBIOSSMC_Result_OK)
+		smu_print("SMU Response was not OK. SMU response after wait received is: %d\n", result);
+
+	if (result == VBIOSSMC_Status_BUSY) {
+		return -1;
+	}
+
+	/* First clear response register */
+	REG_WRITE(MP1_SMN_C2PMSG_91, VBIOSSMC_Status_BUSY);
+
+	/* Set the parameter register for the SMU message, unit is Mhz */
+	REG_WRITE(MP1_SMN_C2PMSG_83, param);
+
+	/* Trigger the message transaction by writing the message ID */
+	REG_WRITE(MP1_SMN_C2PMSG_67, msg_id);
+
+	result = dcn316_smu_wait_for_response(clk_mgr, 10, 200000);
+
+	if (result == VBIOSSMC_Status_BUSY) {
+		ASSERT(0);
+		dm_helpers_smu_timeout(CTX, msg_id, param, 10 * 200000);
+	}
+
+	return REG_READ(MP1_SMN_C2PMSG_83);
+}
+
+int dcn316_smu_get_smu_version(struct clk_mgr_internal *clk_mgr)
+{
+	return dcn316_smu_send_msg_with_param(
+			clk_mgr,
+			VBIOSSMC_MSG_GetPmfwVersion,
+			0);
+}
+
+
+int dcn316_smu_set_dispclk(struct clk_mgr_internal *clk_mgr, int requested_dispclk_khz)
+{
+	int actual_dispclk_set_mhz = -1;
+
+	if (!clk_mgr->smu_present)
+		return requested_dispclk_khz;
+
+	/*  Unit of SMU msg parameter is Mhz */
+	actual_dispclk_set_mhz = dcn316_smu_send_msg_with_param(
+			clk_mgr,
+			VBIOSSMC_MSG_SetDispclkFreq,
+			khz_to_mhz_ceil(requested_dispclk_khz));
+
+	return actual_dispclk_set_mhz * 1000;
+}
+
+int dcn316_smu_set_hard_min_dcfclk(struct clk_mgr_internal *clk_mgr, int requested_dcfclk_khz)
+{
+	int actual_dcfclk_set_mhz = -1;
+
+	if (!clk_mgr->base.ctx->dc->debug.pstate_enabled)
+		return -1;
+
+	if (!clk_mgr->smu_present)
+		return requested_dcfclk_khz;
+
+	actual_dcfclk_set_mhz = dcn316_smu_send_msg_with_param(
+			clk_mgr,
+			VBIOSSMC_MSG_SetHardMinDcfclkByFreq,
+			khz_to_mhz_ceil(requested_dcfclk_khz));
+
+	return actual_dcfclk_set_mhz * 1000;
+}
+
+int dcn316_smu_set_min_deep_sleep_dcfclk(struct clk_mgr_internal *clk_mgr, int requested_min_ds_dcfclk_khz)
+{
+	int actual_min_ds_dcfclk_mhz = -1;
+
+	if (!clk_mgr->base.ctx->dc->debug.pstate_enabled)
+		return -1;
+
+	if (!clk_mgr->smu_present)
+		return requested_min_ds_dcfclk_khz;
+
+	actual_min_ds_dcfclk_mhz = dcn316_smu_send_msg_with_param(
+			clk_mgr,
+			VBIOSSMC_MSG_SetMinDeepSleepDcfclk,
+			khz_to_mhz_ceil(requested_min_ds_dcfclk_khz));
+
+	return actual_min_ds_dcfclk_mhz * 1000;
+}
+
+int dcn316_smu_set_dppclk(struct clk_mgr_internal *clk_mgr, int requested_dpp_khz)
+{
+	int actual_dppclk_set_mhz = -1;
+
+	if (!clk_mgr->smu_present)
+		return requested_dpp_khz;
+
+	actual_dppclk_set_mhz = dcn316_smu_send_msg_with_param(
+			clk_mgr,
+			VBIOSSMC_MSG_SetDppclkFreq,
+			khz_to_mhz_ceil(requested_dpp_khz));
+
+	return actual_dppclk_set_mhz * 1000;
+}
+
+void dcn316_smu_set_display_idle_optimization(struct clk_mgr_internal *clk_mgr, uint32_t idle_info)
+{
+	if (!clk_mgr->base.ctx->dc->debug.pstate_enabled)
+		return;
+
+	if (!clk_mgr->smu_present)
+		return;
+
+	//TODO: Work with smu team to define optimization options.
+	dcn316_smu_send_msg_with_param(
+		clk_mgr,
+		VBIOSSMC_MSG_SetDisplayIdleOptimizations,
+		idle_info);
+}
+
+void dcn316_smu_enable_phy_refclk_pwrdwn(struct clk_mgr_internal *clk_mgr, bool enable)
+{
+	union display_idle_optimization_u idle_info = { 0 };
+
+	if (!clk_mgr->smu_present)
+		return;
+
+	if (enable) {
+		idle_info.idle_info.df_request_disabled = 1;
+		idle_info.idle_info.phy_ref_clk_off = 1;
+	}
+
+	dcn316_smu_send_msg_with_param(
+			clk_mgr,
+			VBIOSSMC_MSG_SetDisplayIdleOptimizations,
+			idle_info.data);
+}
+
+void dcn316_smu_set_dram_addr_high(struct clk_mgr_internal *clk_mgr, uint32_t addr_high)
+{
+	if (!clk_mgr->smu_present)
+		return;
+
+	dcn316_smu_send_msg_with_param(clk_mgr,
+			VBIOSSMC_MSG_SetVbiosDramAddrHigh, addr_high);
+}
+
+void dcn316_smu_set_dram_addr_low(struct clk_mgr_internal *clk_mgr, uint32_t addr_low)
+{
+	if (!clk_mgr->smu_present)
+		return;
+
+	dcn316_smu_send_msg_with_param(clk_mgr,
+			VBIOSSMC_MSG_SetVbiosDramAddrLow, addr_low);
+}
+
+void dcn316_smu_transfer_dpm_table_smu_2_dram(struct clk_mgr_internal *clk_mgr)
+{
+	if (!clk_mgr->smu_present)
+		return;
+
+	dcn316_smu_send_msg_with_param(clk_mgr,
+			VBIOSSMC_MSG_TransferTableSmu2Dram, TABLE_DPMCLOCKS);
+}
+
+void dcn316_smu_transfer_wm_table_dram_2_smu(struct clk_mgr_internal *clk_mgr)
+{
+	if (!clk_mgr->smu_present)
+		return;
+
+	dcn316_smu_send_msg_with_param(clk_mgr,
+			VBIOSSMC_MSG_TransferTableDram2Smu, TABLE_WATERMARKS);
+}
+
+void dcn316_smu_enable_pme_wa(struct clk_mgr_internal *clk_mgr)
+{
+	if (!clk_mgr->smu_present)
+		return;
+
+	dcn316_smu_send_msg_with_param(
+			clk_mgr,
+			VBIOSSMC_MSG_UpdatePmeRestore,
+			0);
+}
+
+/* Arg = 1: Turn DTB on; 0: Turn DTB CLK OFF. when it is on, it is 600MHZ */
+void dcn316_smu_set_dtbclk(struct clk_mgr_internal *clk_mgr, bool enable)
+{
+	if (!clk_mgr->smu_present)
+		return;
+
+	dcn316_smu_send_msg_with_param(
+			clk_mgr,
+			VBIOSSMC_MSG_SetDtbclkFreq,
+			enable);
+}
+
+int dcn316_smu_get_dpref_clk(struct clk_mgr_internal *clk_mgr)
+{
+	int dprefclk_get_mhz = -1;
+
+	if (clk_mgr->smu_present) {
+		dprefclk_get_mhz = dcn316_smu_send_msg_with_param(
+			clk_mgr,
+			VBIOSSMC_MSG_GetDprefclkFreq,
+			0);
+	}
+	return (dprefclk_get_mhz * 1000);
+}
+
+int dcn316_smu_get_smu_fclk(struct clk_mgr_internal *clk_mgr)
+{
+	int fclk_get_mhz = -1;
+
+	if (clk_mgr->smu_present) {
+		fclk_get_mhz = dcn316_smu_send_msg_with_param(
+			clk_mgr,
+			VBIOSSMC_MSG_GetFclkFrequency,
+			0);
+	}
+	return (fclk_get_mhz * 1000);
+}
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn316/dcn316_smu.h b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn316/dcn316_smu.h
new file mode 100644
index 000000000000..2a7293f66515
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn316/dcn316_smu.h
@@ -0,0 +1,139 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef DAL_DC_316_SMU_H_
+#define DAL_DC_316_SMU_H_
+#include "os_types.h"
+
+#define PMFW_DRIVER_IF_VERSION 4
+
+#define NUM_DCFCLK_DPM_LEVELS   8
+#define NUM_DISPCLK_DPM_LEVELS  8
+#define NUM_DPPCLK_DPM_LEVELS   8
+#define NUM_SOCCLK_DPM_LEVELS   8
+#define NUM_VCN_DPM_LEVELS      8
+#define NUM_SOC_VOLTAGE_LEVELS  8
+#define NUM_DF_PSTATE_LEVELS    4
+
+typedef struct {
+  uint16_t MinClock; // This is either DCFCLK or SOCCLK (in MHz)
+  uint16_t MaxClock; // This is either DCFCLK or SOCCLK (in MHz)
+  uint16_t MinMclk;
+  uint16_t MaxMclk;
+  uint8_t  WmSetting;
+  uint8_t  WmType;  // Used for normal pstate change or memory retraining
+  uint8_t  Padding[2];
+} WatermarkRowGeneric_t;
+
+#define NUM_WM_RANGES 4
+#define WM_PSTATE_CHG 0
+#define WM_RETRAINING 1
+
+typedef enum {
+  WM_SOCCLK = 0,
+  WM_DCFCLK,
+  WM_COUNT,
+} WM_CLOCK_e;
+
+typedef enum{
+  WCK_RATIO_1_1 = 0,  // DDR5, Wck:ck is always 1:1;
+  WCK_RATIO_1_2,
+  WCK_RATIO_1_4,
+  WCK_RATIO_MAX
+} WCK_RATIO_e;
+
+typedef struct {
+  uint32_t FClk;
+  uint32_t MemClk;
+  uint32_t Voltage;
+  uint8_t  WckRatio;
+  uint8_t  Spare[3];
+} DfPstateTable_t;
+
+//Freq in MHz
+//Voltage in milli volts with 2 fractional bits
+typedef struct {
+  uint32_t DcfClocks[NUM_DCFCLK_DPM_LEVELS];
+  uint32_t DispClocks[NUM_DISPCLK_DPM_LEVELS];
+  uint32_t DppClocks[NUM_DPPCLK_DPM_LEVELS];
+  uint32_t SocClocks[NUM_SOCCLK_DPM_LEVELS];
+  uint32_t VClocks[NUM_VCN_DPM_LEVELS];
+  uint32_t DClocks[NUM_VCN_DPM_LEVELS];
+  uint32_t SocVoltage[NUM_SOC_VOLTAGE_LEVELS];
+  DfPstateTable_t DfPstateTable[NUM_DF_PSTATE_LEVELS];
+  uint8_t  NumDcfClkLevelsEnabled;
+  uint8_t  NumDispClkLevelsEnabled; //Applies to both Dispclk and Dppclk
+  uint8_t  NumSocClkLevelsEnabled;
+  uint8_t  VcnClkLevelsEnabled;     //Applies to both Vclk and Dclk
+  uint8_t  NumDfPstatesEnabled;
+  uint8_t  spare[3];
+  uint32_t MinGfxClk;
+  uint32_t MaxGfxClk;
+} DpmClocks_316_t;
+
+struct dcn316_watermarks {
+  // Watermarks
+  WatermarkRowGeneric_t WatermarkRow[WM_COUNT][NUM_WM_RANGES];
+  uint32_t MmHubPadding[7]; // SMU internal use
+};
+
+struct dcn316_smu_dpm_clks {
+	DpmClocks_316_t *dpm_clks;
+	union large_integer mc_address;
+};
+
+#define TABLE_WATERMARKS         1 // Called by DAL through VBIOS
+#define TABLE_DPMCLOCKS          4 // Called by Driver and VBIOS
+
+struct display_idle_optimization {
+	unsigned int df_request_disabled : 1;
+	unsigned int phy_ref_clk_off     : 1;
+	unsigned int s0i2_rdy            : 1;
+	unsigned int reserved            : 29;
+};
+
+union display_idle_optimization_u {
+	struct display_idle_optimization idle_info;
+	uint32_t data;
+};
+
+int dcn316_smu_get_smu_version(struct clk_mgr_internal *clk_mgr);
+int dcn316_smu_set_dispclk(struct clk_mgr_internal *clk_mgr, int requested_dispclk_khz);
+int dcn316_smu_set_hard_min_dcfclk(struct clk_mgr_internal *clk_mgr, int requested_dcfclk_khz);
+int dcn316_smu_set_min_deep_sleep_dcfclk(struct clk_mgr_internal *clk_mgr, int requested_min_ds_dcfclk_khz);
+int dcn316_smu_set_dppclk(struct clk_mgr_internal *clk_mgr, int requested_dpp_khz);
+void dcn316_smu_set_display_idle_optimization(struct clk_mgr_internal *clk_mgr, uint32_t idle_info);
+void dcn316_smu_enable_phy_refclk_pwrdwn(struct clk_mgr_internal *clk_mgr, bool enable);
+void dcn316_smu_set_dram_addr_high(struct clk_mgr_internal *clk_mgr, uint32_t addr_high);
+void dcn316_smu_set_dram_addr_low(struct clk_mgr_internal *clk_mgr, uint32_t addr_low);
+void dcn316_smu_transfer_dpm_table_smu_2_dram(struct clk_mgr_internal *clk_mgr);
+void dcn316_smu_transfer_wm_table_dram_2_smu(struct clk_mgr_internal *clk_mgr);
+void dcn316_smu_request_voltage_via_phyclk(struct clk_mgr_internal *clk_mgr, int requested_phyclk_khz);
+void dcn316_smu_enable_pme_wa(struct clk_mgr_internal *clk_mgr);
+void dcn316_smu_set_dtbclk(struct clk_mgr_internal *clk_mgr, bool enable);
+int dcn316_smu_get_dpref_clk(struct clk_mgr_internal *clk_mgr);
+int dcn316_smu_get_smu_fclk(struct clk_mgr_internal *clk_mgr);
+
+#endif /* DAL_DC_316_SMU_H_ */
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn32/dalsmc.h b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn32/dalsmc.h
new file mode 100644
index 000000000000..724a508b0adb
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn32/dalsmc.h
@@ -0,0 +1,74 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+#ifndef DALSMC_H
+#define DALSMC_H
+
+#define DALSMC_VERSION 0x1
+
+// SMU Response Codes:
+#define DALSMC_Result_OK                   0x1
+#define DALSMC_Result_Failed               0xFF
+#define DALSMC_Result_UnknownCmd           0xFE
+#define DALSMC_Result_CmdRejectedPrereq    0xFD
+#define DALSMC_Result_CmdRejectedBusy      0xFC
+
+// Message Definitions:
+#define DALSMC_MSG_TestMessage                    0x1
+#define DALSMC_MSG_GetSmuVersion                  0x2
+#define DALSMC_MSG_GetDriverIfVersion             0x3
+#define DALSMC_MSG_GetMsgHeaderVersion            0x4
+#define DALSMC_MSG_SetDalDramAddrHigh             0x5
+#define DALSMC_MSG_SetDalDramAddrLow              0x6
+#define DALSMC_MSG_TransferTableSmu2Dram          0x7
+#define DALSMC_MSG_TransferTableDram2Smu          0x8
+#define DALSMC_MSG_SetHardMinByFreq               0x9
+#define DALSMC_MSG_SetHardMaxByFreq               0xA
+#define DALSMC_MSG_GetDpmFreqByIndex              0xB
+#define DALSMC_MSG_GetDcModeMaxDpmFreq            0xC
+#define DALSMC_MSG_SetMinDeepSleepDcfclk          0xD
+#define DALSMC_MSG_NumOfDisplays                  0xE
+#define DALSMC_MSG_SetExternalClientDfCstateAllow 0xF
+#define DALSMC_MSG_BacoAudioD3PME                 0x10
+#define DALSMC_MSG_SetFclkSwitchAllow             0x11
+#define DALSMC_MSG_SetCabForUclkPstate            0x12
+#define DALSMC_MSG_SetWorstCaseUclkLatency        0x13
+#define DALSMC_MSG_SetAlwaysWaitDmcubResp         0x14
+#define DALSMC_MSG_ReturnHardMinStatus            0x15
+#define DALSMC_Message_Count                      0x16
+
+#define CHECK_HARD_MIN_CLK_DISPCLK                0x1
+#define CHECK_HARD_MIN_CLK_DPPCLK                 0x2
+#define CHECK_HARD_MIN_CLK_DPREFCLK               0x4
+#define CHECK_HARD_MIN_CLK_DCFCLK                 0x8
+#define CHECK_HARD_MIN_CLK_DTBCLK                 0x10
+#define CHECK_HARD_MIN_CLK_UCLK                   0x20
+
+typedef enum {
+	FCLK_SWITCH_DISALLOW,
+	FCLK_SWITCH_ALLOW,
+} FclkSwitchAllow_e;
+
+#endif
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn32/dcn32_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn32/dcn32_clk_mgr.c
new file mode 100644
index 000000000000..7da7b41bd092
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn32/dcn32_clk_mgr.c
@@ -0,0 +1,1233 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#include "dccg.h"
+#include "clk_mgr_internal.h"
+#include "dcn32/dcn32_clk_mgr_smu_msg.h"
+#include "dcn20/dcn20_clk_mgr.h"
+#include "dce100/dce_clk_mgr.h"
+#include "dcn31/dcn31_clk_mgr.h"
+#include "dcn32/dcn32_clk_mgr.h"
+#include "reg_helper.h"
+#include "core_types.h"
+#include "dm_helpers.h"
+#include "link_service.h"
+#include "dc_state_priv.h"
+#include "atomfirmware.h"
+#include "dcn32_smu13_driver_if.h"
+
+#include "dcn/dcn_3_2_0_offset.h"
+#include "dcn/dcn_3_2_0_sh_mask.h"
+
+#include "dml/dcn32/dcn32_fpu.h"
+
+#define DCN_BASE__INST0_SEG1                       0x000000C0
+
+#define mmCLK1_CLK_PLL_REQ                              0x16E37
+#define mmCLK1_CLK0_DFS_CNTL                            0x16E69
+#define mmCLK1_CLK1_DFS_CNTL                            0x16E6C
+#define mmCLK1_CLK2_DFS_CNTL                            0x16E6F
+#define mmCLK1_CLK3_DFS_CNTL                            0x16E72
+#define mmCLK1_CLK4_DFS_CNTL                            0x16E75
+
+#define mmCLK1_CLK0_CURRENT_CNT                         0x16EE7
+#define mmCLK1_CLK1_CURRENT_CNT                         0x16EE8
+#define mmCLK1_CLK2_CURRENT_CNT                         0x16EE9
+#define mmCLK1_CLK3_CURRENT_CNT                         0x16EEA
+#define mmCLK1_CLK4_CURRENT_CNT                         0x16EEB
+
+#define mmCLK4_CLK0_CURRENT_CNT                         0x1B0C9
+
+#define CLK1_CLK_PLL_REQ__FbMult_int_MASK               0x000001ffUL
+#define CLK1_CLK_PLL_REQ__PllSpineDiv_MASK              0x0000f000UL
+#define CLK1_CLK_PLL_REQ__FbMult_frac_MASK              0xffff0000UL
+#define CLK1_CLK_PLL_REQ__FbMult_int__SHIFT             0x00000000
+#define CLK1_CLK_PLL_REQ__PllSpineDiv__SHIFT            0x0000000c
+#define CLK1_CLK_PLL_REQ__FbMult_frac__SHIFT            0x00000010
+
+#define mmCLK01_CLK0_CLK_PLL_REQ                        0x16E37
+#define mmCLK01_CLK0_CLK0_DFS_CNTL                      0x16E64
+#define mmCLK01_CLK0_CLK1_DFS_CNTL                      0x16E67
+#define mmCLK01_CLK0_CLK2_DFS_CNTL                      0x16E6A
+#define mmCLK01_CLK0_CLK3_DFS_CNTL                      0x16E6D
+#define mmCLK01_CLK0_CLK4_DFS_CNTL                      0x16E70
+
+#define CLK0_CLK_PLL_REQ__FbMult_int_MASK               0x000001ffL
+#define CLK0_CLK_PLL_REQ__PllSpineDiv_MASK              0x0000f000L
+#define CLK0_CLK_PLL_REQ__FbMult_frac_MASK              0xffff0000L
+#define CLK0_CLK_PLL_REQ__FbMult_int__SHIFT             0x00000000
+#define CLK0_CLK_PLL_REQ__PllSpineDiv__SHIFT            0x0000000c
+#define CLK0_CLK_PLL_REQ__FbMult_frac__SHIFT            0x00000010
+
+#undef FN
+#define FN(reg_name, field_name) \
+	clk_mgr->clk_mgr_shift->field_name, clk_mgr->clk_mgr_mask->field_name
+
+#define REG(reg) \
+	(clk_mgr->regs->reg)
+
+#define BASE_INNER(seg) DCN_BASE__INST0_SEG ## seg
+
+#define BASE(seg) BASE_INNER(seg)
+
+#define SR(reg_name)\
+		.reg_name = BASE(reg ## reg_name ## _BASE_IDX) +  \
+					reg ## reg_name
+
+#define CLK_SR_DCN32(reg_name)\
+	.reg_name = mm ## reg_name
+
+static const struct clk_mgr_registers clk_mgr_regs_dcn32 = {
+	CLK_REG_LIST_DCN32()
+};
+
+static const struct clk_mgr_shift clk_mgr_shift_dcn32 = {
+	CLK_COMMON_MASK_SH_LIST_DCN32(__SHIFT)
+};
+
+static const struct clk_mgr_mask clk_mgr_mask_dcn32 = {
+	CLK_COMMON_MASK_SH_LIST_DCN32(_MASK)
+};
+
+
+#define CLK_SR_DCN321(reg_name, block, inst)\
+	.reg_name = mm ## block ## _ ## reg_name
+
+static const struct clk_mgr_registers clk_mgr_regs_dcn321 = {
+	CLK_REG_LIST_DCN321()
+};
+
+static const struct clk_mgr_shift clk_mgr_shift_dcn321 = {
+	CLK_COMMON_MASK_SH_LIST_DCN321(__SHIFT)
+};
+
+static const struct clk_mgr_mask clk_mgr_mask_dcn321 = {
+	CLK_COMMON_MASK_SH_LIST_DCN321(_MASK)
+};
+
+
+/* Query SMU for all clock states for a particular clock */
+static void dcn32_init_single_clock(struct clk_mgr_internal *clk_mgr, PPCLK_e clk, unsigned int *entry_0,
+		unsigned int *num_levels)
+{
+	unsigned int i;
+	char *entry_i = (char *)entry_0;
+
+	uint32_t ret = dcn30_smu_get_dpm_freq_by_index(clk_mgr, clk, 0xFF);
+
+	if (ret & (1 << 31))
+		/* fine-grained, only min and max */
+		*num_levels = 2;
+	else
+		/* discrete, a number of fixed states */
+		/* will set num_levels to 0 on failure */
+		*num_levels = ret & 0xFF;
+
+	/* if the initial message failed, num_levels will be 0 */
+	for (i = 0; i < *num_levels; i++) {
+		*((unsigned int *)entry_i) = (dcn30_smu_get_dpm_freq_by_index(clk_mgr, clk, i) & 0xFFFF);
+		entry_i += sizeof(clk_mgr->base.bw_params->clk_table.entries[0]);
+	}
+}
+
+static void dcn32_build_wm_range_table(struct clk_mgr_internal *clk_mgr)
+{
+	DC_FP_START();
+	dcn32_build_wm_range_table_fpu(clk_mgr);
+	DC_FP_END();
+}
+
+void dcn32_init_clocks(struct clk_mgr *clk_mgr_base)
+{
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+	unsigned int num_levels;
+	struct clk_limit_num_entries *num_entries_per_clk;
+	unsigned int i;
+
+	if (!clk_mgr_base->bw_params)
+		return;
+
+	num_entries_per_clk = &clk_mgr_base->bw_params->clk_table.num_entries_per_clk;
+
+	memset(&(clk_mgr_base->clks), 0, sizeof(struct dc_clocks));
+	clk_mgr_base->clks.p_state_change_support = true;
+	clk_mgr_base->clks.prev_p_state_change_support = true;
+	clk_mgr_base->clks.fclk_prev_p_state_change_support = true;
+	clk_mgr->smu_present = false;
+	clk_mgr->dpm_present = false;
+
+	if (!clk_mgr_base->force_smu_not_present && dcn30_smu_get_smu_version(clk_mgr, &clk_mgr->smu_ver))
+		clk_mgr->smu_present = true;
+
+	if (!clk_mgr->smu_present)
+		return;
+
+	dcn30_smu_check_driver_if_version(clk_mgr);
+	dcn30_smu_check_msg_header_version(clk_mgr);
+
+	/* DCFCLK */
+	dcn32_init_single_clock(clk_mgr, PPCLK_DCFCLK,
+			&clk_mgr_base->bw_params->clk_table.entries[0].dcfclk_mhz,
+			&num_entries_per_clk->num_dcfclk_levels);
+	clk_mgr_base->bw_params->dc_mode_limit.dcfclk_mhz = dcn30_smu_get_dc_mode_max_dpm_freq(clk_mgr, PPCLK_DCFCLK);
+
+	/* SOCCLK */
+	dcn32_init_single_clock(clk_mgr, PPCLK_SOCCLK,
+					&clk_mgr_base->bw_params->clk_table.entries[0].socclk_mhz,
+					&num_entries_per_clk->num_socclk_levels);
+	clk_mgr_base->bw_params->dc_mode_limit.socclk_mhz = dcn30_smu_get_dc_mode_max_dpm_freq(clk_mgr, PPCLK_SOCCLK);
+
+	/* DTBCLK */
+	if (!clk_mgr->base.ctx->dc->debug.disable_dtb_ref_clk_switch) {
+		dcn32_init_single_clock(clk_mgr, PPCLK_DTBCLK,
+				&clk_mgr_base->bw_params->clk_table.entries[0].dtbclk_mhz,
+				&num_entries_per_clk->num_dtbclk_levels);
+		clk_mgr_base->bw_params->dc_mode_limit.dtbclk_mhz =
+				dcn30_smu_get_dc_mode_max_dpm_freq(clk_mgr, PPCLK_DTBCLK);
+	}
+
+	/* DISPCLK */
+	dcn32_init_single_clock(clk_mgr, PPCLK_DISPCLK,
+			&clk_mgr_base->bw_params->clk_table.entries[0].dispclk_mhz,
+			&num_entries_per_clk->num_dispclk_levels);
+	num_levels = num_entries_per_clk->num_dispclk_levels;
+	clk_mgr_base->bw_params->dc_mode_limit.dispclk_mhz = dcn30_smu_get_dc_mode_max_dpm_freq(clk_mgr, PPCLK_DISPCLK);
+	//HW recommends limit of 1950 MHz in display clock for all DCN3.2.x
+	if (clk_mgr_base->bw_params->dc_mode_limit.dispclk_mhz > 1950)
+		clk_mgr_base->bw_params->dc_mode_limit.dispclk_mhz = 1950;
+
+	/* DPPCLK */
+	dcn32_init_single_clock(clk_mgr, PPCLK_DPPCLK,
+			&clk_mgr_base->bw_params->clk_table.entries[0].dppclk_mhz,
+			&num_entries_per_clk->num_dppclk_levels);
+	num_levels = num_entries_per_clk->num_dppclk_levels;
+	clk_mgr_base->bw_params->dc_mode_limit.dppclk_mhz = dcn30_smu_get_dc_mode_max_dpm_freq(clk_mgr, PPCLK_DPPCLK);
+	//HW recommends limit of 1950 MHz in display clock for all DCN3.2.x
+	if (clk_mgr_base->bw_params->dc_mode_limit.dppclk_mhz > 1950)
+		clk_mgr_base->bw_params->dc_mode_limit.dppclk_mhz = 1950;
+
+	if (num_entries_per_clk->num_dcfclk_levels &&
+			num_entries_per_clk->num_dtbclk_levels &&
+			num_entries_per_clk->num_dispclk_levels)
+		clk_mgr->dpm_present = true;
+
+	if (clk_mgr_base->ctx->dc->debug.min_disp_clk_khz) {
+		for (i = 0; i < num_levels; i++)
+			if (clk_mgr_base->bw_params->clk_table.entries[i].dispclk_mhz
+					< khz_to_mhz_ceil(clk_mgr_base->ctx->dc->debug.min_disp_clk_khz))
+				clk_mgr_base->bw_params->clk_table.entries[i].dispclk_mhz
+					= khz_to_mhz_ceil(clk_mgr_base->ctx->dc->debug.min_disp_clk_khz);
+	}
+	for (i = 0; i < num_levels; i++)
+		if (clk_mgr_base->bw_params->clk_table.entries[i].dispclk_mhz > 1950)
+			clk_mgr_base->bw_params->clk_table.entries[i].dispclk_mhz = 1950;
+
+	if (clk_mgr_base->ctx->dc->debug.min_dpp_clk_khz) {
+		for (i = 0; i < num_levels; i++)
+			if (clk_mgr_base->bw_params->clk_table.entries[i].dppclk_mhz
+					< khz_to_mhz_ceil(clk_mgr_base->ctx->dc->debug.min_dpp_clk_khz))
+				clk_mgr_base->bw_params->clk_table.entries[i].dppclk_mhz
+					= khz_to_mhz_ceil(clk_mgr_base->ctx->dc->debug.min_dpp_clk_khz);
+	}
+
+	for (i = 0; i < num_levels; i++)
+		if (clk_mgr_base->bw_params->clk_table.entries[i].dppclk_mhz > 1950)
+			clk_mgr_base->bw_params->clk_table.entries[i].dppclk_mhz = 1950;
+
+	/* Get UCLK, update bounding box */
+	clk_mgr_base->funcs->get_memclk_states_from_smu(clk_mgr_base);
+
+	/* WM range table */
+	dcn32_build_wm_range_table(clk_mgr);
+}
+
+static void dcn32_update_clocks_update_dtb_dto(struct clk_mgr_internal *clk_mgr,
+			struct dc_state *context,
+			int ref_dtbclk_khz)
+{
+	struct dccg *dccg = clk_mgr->dccg;
+	uint32_t tg_mask = 0;
+	int i;
+
+	for (i = 0; i < clk_mgr->base.ctx->dc->res_pool->pipe_count; i++) {
+		struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i];
+		struct dtbclk_dto_params dto_params = {0};
+
+		/* use mask to program DTO once per tg */
+		if (pipe_ctx->stream_res.tg &&
+				!(tg_mask & (1 << pipe_ctx->stream_res.tg->inst))) {
+			tg_mask |= (1 << pipe_ctx->stream_res.tg->inst);
+
+			dto_params.otg_inst = pipe_ctx->stream_res.tg->inst;
+			dto_params.ref_dtbclk_khz = ref_dtbclk_khz;
+
+			dccg->funcs->set_dtbclk_dto(clk_mgr->dccg, &dto_params);
+			//dccg->funcs->set_audio_dtbclk_dto(clk_mgr->dccg, &dto_params);
+		}
+	}
+}
+
+/* Since DPPCLK request to PMFW needs to be exact (due to DPP DTO programming),
+ * update DPPCLK to be the exact frequency that will be set after the DPPCLK
+ * divider is updated. This will prevent rounding issues that could cause DPP
+ * refclk and DPP DTO to not match up.
+ */
+static void dcn32_update_dppclk_dispclk_freq(struct clk_mgr_internal *clk_mgr, struct dc_clocks *new_clocks)
+{
+	int dpp_divider = 0;
+	int disp_divider = 0;
+
+	if (new_clocks->dppclk_khz) {
+		dpp_divider = DENTIST_DIVIDER_RANGE_SCALE_FACTOR
+				* clk_mgr->base.dentist_vco_freq_khz / new_clocks->dppclk_khz;
+		new_clocks->dppclk_khz = (DENTIST_DIVIDER_RANGE_SCALE_FACTOR * clk_mgr->base.dentist_vco_freq_khz) / dpp_divider;
+	}
+	if (new_clocks->dispclk_khz > 0) {
+		disp_divider = DENTIST_DIVIDER_RANGE_SCALE_FACTOR
+				* clk_mgr->base.dentist_vco_freq_khz / new_clocks->dispclk_khz;
+		new_clocks->dispclk_khz = (DENTIST_DIVIDER_RANGE_SCALE_FACTOR * clk_mgr->base.dentist_vco_freq_khz) / disp_divider;
+	}
+}
+
+void dcn32_update_clocks_update_dpp_dto(struct clk_mgr_internal *clk_mgr,
+		struct dc_state *context, bool safe_to_lower)
+{
+	int i;
+
+	clk_mgr->dccg->ref_dppclk = clk_mgr->base.clks.dppclk_khz;
+	for (i = 0; i < clk_mgr->base.ctx->dc->res_pool->pipe_count; i++) {
+		int dpp_inst = 0, dppclk_khz, prev_dppclk_khz;
+
+		dppclk_khz = context->res_ctx.pipe_ctx[i].plane_res.bw.dppclk_khz;
+
+		if (context->res_ctx.pipe_ctx[i].plane_res.dpp)
+			dpp_inst = context->res_ctx.pipe_ctx[i].plane_res.dpp->inst;
+		else if (!context->res_ctx.pipe_ctx[i].plane_res.dpp && dppclk_khz == 0) {
+			/* dpp == NULL && dppclk_khz == 0 is valid because of pipe harvesting.
+			 * In this case just continue in loop
+			 */
+			continue;
+		} else if (!context->res_ctx.pipe_ctx[i].plane_res.dpp && dppclk_khz > 0) {
+			/* The software state is not valid if dpp resource is NULL and
+			 * dppclk_khz > 0.
+			 */
+			ASSERT(false);
+			continue;
+		}
+
+		prev_dppclk_khz = clk_mgr->dccg->pipe_dppclk_khz[i];
+
+		if (safe_to_lower || prev_dppclk_khz < dppclk_khz)
+			clk_mgr->dccg->funcs->update_dpp_dto(
+							clk_mgr->dccg, dpp_inst, dppclk_khz);
+	}
+}
+
+static void dcn32_update_clocks_update_dentist(
+		struct clk_mgr_internal *clk_mgr,
+		struct dc_state *context)
+{
+	uint32_t new_disp_divider = 0;
+	uint32_t new_dispclk_wdivider = 0;
+	uint32_t old_dispclk_wdivider = 0;
+	uint32_t i;
+	uint32_t dentist_dispclk_wdivider_readback = 0;
+	struct dc *dc = clk_mgr->base.ctx->dc;
+
+	if (clk_mgr->base.clks.dispclk_khz == 0)
+		return;
+
+	new_disp_divider = DENTIST_DIVIDER_RANGE_SCALE_FACTOR
+			* clk_mgr->base.dentist_vco_freq_khz / clk_mgr->base.clks.dispclk_khz;
+
+	new_dispclk_wdivider = dentist_get_did_from_divider(new_disp_divider);
+	REG_GET(DENTIST_DISPCLK_CNTL,
+			DENTIST_DISPCLK_WDIVIDER, &old_dispclk_wdivider);
+
+	/* When changing divider to or from 127, some extra programming is required to prevent corruption */
+	if (old_dispclk_wdivider == 127 && new_dispclk_wdivider != 127) {
+		for (i = 0; i < clk_mgr->base.ctx->dc->res_pool->pipe_count; i++) {
+			struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i];
+			uint32_t fifo_level;
+			struct dccg *dccg = clk_mgr->base.ctx->dc->res_pool->dccg;
+			struct stream_encoder *stream_enc = pipe_ctx->stream_res.stream_enc;
+			int32_t N;
+			int32_t j;
+
+			if (!resource_is_pipe_type(pipe_ctx, OTG_MASTER))
+				continue;
+			/* Virtual encoders don't have this function */
+			if (!stream_enc->funcs->get_fifo_cal_average_level)
+				continue;
+			fifo_level = stream_enc->funcs->get_fifo_cal_average_level(
+					stream_enc);
+			N = fifo_level / 4;
+			dccg->funcs->set_fifo_errdet_ovr_en(
+					dccg,
+					true);
+			for (j = 0; j < N - 4; j++)
+				dccg->funcs->otg_drop_pixel(
+						dccg,
+						pipe_ctx->stream_res.tg->inst);
+			dccg->funcs->set_fifo_errdet_ovr_en(
+					dccg,
+					false);
+		}
+	} else if (new_dispclk_wdivider == 127 && old_dispclk_wdivider != 127) {
+		/* request clock with 126 divider first */
+		uint32_t temp_disp_divider = dentist_get_divider_from_did(126);
+		uint32_t temp_dispclk_khz = (DENTIST_DIVIDER_RANGE_SCALE_FACTOR * clk_mgr->base.dentist_vco_freq_khz) / temp_disp_divider;
+
+		if (clk_mgr->smu_present)
+			/*
+			 * SMU uses discrete dispclk presets. We applied
+			 * the same formula to increase our dppclk_khz
+			 * to the next matching discrete value. By
+			 * contract, we should use the preset dispclk
+			 * floored in Mhz to describe the intended clock.
+			 */
+			dcn32_smu_set_hard_min_by_freq(clk_mgr, PPCLK_DISPCLK,
+					khz_to_mhz_floor(temp_dispclk_khz));
+
+		if (dc->debug.override_dispclk_programming) {
+			REG_GET(DENTIST_DISPCLK_CNTL,
+					DENTIST_DISPCLK_WDIVIDER, &dentist_dispclk_wdivider_readback);
+
+			if (dentist_dispclk_wdivider_readback != 126) {
+				REG_UPDATE(DENTIST_DISPCLK_CNTL,
+						DENTIST_DISPCLK_WDIVIDER, 126);
+				REG_WAIT(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_CHG_DONE, 1, 50, 2000);
+			}
+		}
+
+		for (i = 0; i < clk_mgr->base.ctx->dc->res_pool->pipe_count; i++) {
+			struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i];
+			struct dccg *dccg = clk_mgr->base.ctx->dc->res_pool->dccg;
+			struct stream_encoder *stream_enc = pipe_ctx->stream_res.stream_enc;
+			uint32_t fifo_level;
+			int32_t N;
+			int32_t j;
+
+			if (!resource_is_pipe_type(pipe_ctx, OTG_MASTER))
+				continue;
+			/* Virtual encoders don't have this function */
+			if (!stream_enc->funcs->get_fifo_cal_average_level)
+				continue;
+			fifo_level = stream_enc->funcs->get_fifo_cal_average_level(
+					stream_enc);
+			N = fifo_level / 4;
+			dccg->funcs->set_fifo_errdet_ovr_en(dccg, true);
+			for (j = 0; j < 12 - N; j++)
+				dccg->funcs->otg_add_pixel(dccg,
+						pipe_ctx->stream_res.tg->inst);
+			dccg->funcs->set_fifo_errdet_ovr_en(dccg, false);
+		}
+	}
+
+	/* do requested DISPCLK updates*/
+	if (clk_mgr->smu_present)
+		/*
+		 * SMU uses discrete dispclk presets. We applied
+		 * the same formula to increase our dppclk_khz
+		 * to the next matching discrete value. By
+		 * contract, we should use the preset dispclk
+		 * floored in Mhz to describe the intended clock.
+		 */
+		dcn32_smu_set_hard_min_by_freq(clk_mgr, PPCLK_DISPCLK,
+				khz_to_mhz_floor(clk_mgr->base.clks.dispclk_khz));
+
+	if (dc->debug.override_dispclk_programming) {
+		REG_GET(DENTIST_DISPCLK_CNTL,
+				DENTIST_DISPCLK_WDIVIDER, &dentist_dispclk_wdivider_readback);
+
+		if (dentist_dispclk_wdivider_readback > new_dispclk_wdivider) {
+			REG_UPDATE(DENTIST_DISPCLK_CNTL,
+					DENTIST_DISPCLK_WDIVIDER, new_dispclk_wdivider);
+			REG_WAIT(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_CHG_DONE, 1, 50, 2000);
+		}
+	}
+
+}
+
+static int dcn32_get_dispclk_from_dentist(struct clk_mgr *clk_mgr_base)
+{
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+	uint32_t dispclk_wdivider;
+	int disp_divider;
+
+	REG_GET(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_WDIVIDER, &dispclk_wdivider);
+	disp_divider = dentist_get_divider_from_did(dispclk_wdivider);
+
+	/* Return DISPCLK freq in Khz */
+	if (disp_divider)
+		return (DENTIST_DIVIDER_RANGE_SCALE_FACTOR * clk_mgr->base.dentist_vco_freq_khz) / disp_divider;
+
+	return 0;
+}
+
+static bool dcn32_check_native_scaling(struct pipe_ctx *pipe)
+{
+	bool is_native_scaling = false;
+	int width = pipe->plane_state->src_rect.width;
+	int height = pipe->plane_state->src_rect.height;
+
+	if (pipe->stream->timing.h_addressable == width &&
+			pipe->stream->timing.v_addressable == height &&
+			pipe->plane_state->dst_rect.width == width &&
+			pipe->plane_state->dst_rect.height == height)
+		is_native_scaling = true;
+
+	return is_native_scaling;
+}
+
+static void dcn32_auto_dpm_test_log(
+		struct dc_clocks *new_clocks,
+		struct clk_mgr_internal *clk_mgr,
+		struct dc_state *context)
+{
+	unsigned int dispclk_khz_reg, dppclk_khz_reg, dprefclk_khz_reg, dcfclk_khz_reg, dtbclk_khz_reg,
+				 fclk_khz_reg, mall_ss_size_bytes;
+	int dramclk_khz_override, fclk_khz_override, num_fclk_levels;
+
+	struct pipe_ctx *pipe_ctx_list[MAX_PIPES];
+	int active_pipe_count = 0;
+
+	for (int i = 0; i < MAX_PIPES; i++) {
+		struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i];
+
+		if (pipe_ctx->stream && dc_state_get_pipe_subvp_type(context, pipe_ctx) != SUBVP_PHANTOM) {
+			pipe_ctx_list[active_pipe_count] = pipe_ctx;
+			active_pipe_count++;
+		}
+	}
+
+	msleep(5);
+
+	mall_ss_size_bytes = context->bw_ctx.bw.dcn.mall_ss_size_bytes;
+
+    dispclk_khz_reg    = REG_READ(CLK1_CLK0_CURRENT_CNT); // DISPCLK
+    dppclk_khz_reg     = REG_READ(CLK1_CLK1_CURRENT_CNT); // DPPCLK
+    dprefclk_khz_reg   = REG_READ(CLK1_CLK2_CURRENT_CNT); // DPREFCLK
+    dcfclk_khz_reg     = REG_READ(CLK1_CLK3_CURRENT_CNT); // DCFCLK
+    dtbclk_khz_reg     = REG_READ(CLK1_CLK4_CURRENT_CNT); // DTBCLK
+    fclk_khz_reg       = REG_READ(CLK4_CLK0_CURRENT_CNT); // FCLK
+
+    // Overrides for these clocks in case there is no p_state change support
+    dramclk_khz_override = new_clocks->dramclk_khz;
+    fclk_khz_override = new_clocks->fclk_khz;
+
+    num_fclk_levels = clk_mgr->base.bw_params->clk_table.num_entries_per_clk.num_fclk_levels - 1;
+
+    if (!new_clocks->p_state_change_support) {
+	    dramclk_khz_override = clk_mgr->base.bw_params->max_memclk_mhz * 1000;
+    }
+    if (!new_clocks->fclk_p_state_change_support) {
+	    fclk_khz_override = clk_mgr->base.bw_params->clk_table.entries[num_fclk_levels].fclk_mhz * 1000;
+    }
+
+	////////////////////////////////////////////////////////////////////////////
+	//	IMPORTANT: 	When adding more clocks to these logs, do NOT put a newline
+	//	 			anywhere other than at the very end of the string.
+	//
+	//	Formatting example (make sure to have " - " between each entry):
+	//
+	//				AutoDPMTest: clk1:%d - clk2:%d - clk3:%d - clk4:%d\n"
+	////////////////////////////////////////////////////////////////////////////
+	if (active_pipe_count > 0 &&
+		new_clocks->dramclk_khz > 0 &&
+		new_clocks->fclk_khz > 0 &&
+		new_clocks->dcfclk_khz > 0 &&
+		new_clocks->dppclk_khz > 0) {
+
+		uint32_t pix_clk_list[MAX_PIPES] = {0};
+		int p_state_list[MAX_PIPES] = {0};
+		int disp_src_width_list[MAX_PIPES] = {0};
+		int disp_src_height_list[MAX_PIPES] = {0};
+		uint64_t disp_src_refresh_list[MAX_PIPES] = {0};
+		bool is_scaled_list[MAX_PIPES] = {0};
+
+		for (int i = 0; i < active_pipe_count; i++) {
+			struct pipe_ctx *curr_pipe_ctx = pipe_ctx_list[i];
+			uint64_t refresh_rate;
+
+			pix_clk_list[i] = curr_pipe_ctx->stream->timing.pix_clk_100hz;
+			p_state_list[i] = curr_pipe_ctx->p_state_type;
+
+			refresh_rate = (curr_pipe_ctx->stream->timing.pix_clk_100hz * (uint64_t)100 +
+				curr_pipe_ctx->stream->timing.v_total * (uint64_t)curr_pipe_ctx->stream->timing.h_total - (uint64_t)1);
+			refresh_rate = div_u64(refresh_rate, curr_pipe_ctx->stream->timing.v_total);
+			refresh_rate = div_u64(refresh_rate, curr_pipe_ctx->stream->timing.h_total);
+			disp_src_refresh_list[i] = refresh_rate;
+
+			if (curr_pipe_ctx->plane_state) {
+				is_scaled_list[i] = !(dcn32_check_native_scaling(curr_pipe_ctx));
+				disp_src_width_list[i] = curr_pipe_ctx->plane_state->src_rect.width;
+				disp_src_height_list[i] = curr_pipe_ctx->plane_state->src_rect.height;
+			}
+		}
+
+		DC_LOG_AUTO_DPM_TEST("AutoDPMTest: dramclk:%d - fclk:%d - "
+			"dcfclk:%d - dppclk:%d - dispclk_hw:%d - "
+			"dppclk_hw:%d - dprefclk_hw:%d - dcfclk_hw:%d - "
+			"dtbclk_hw:%d - fclk_hw:%d - pix_clk_0:%d - pix_clk_1:%d - "
+			"pix_clk_2:%d - pix_clk_3:%d - mall_ss_size:%d - p_state_type_0:%d - "
+			"p_state_type_1:%d - p_state_type_2:%d - p_state_type_3:%d - "
+			"pix_width_0:%d - pix_height_0:%d - refresh_rate_0:%lld - is_scaled_0:%d - "
+			"pix_width_1:%d - pix_height_1:%d - refresh_rate_1:%lld - is_scaled_1:%d - "
+			"pix_width_2:%d - pix_height_2:%d - refresh_rate_2:%lld - is_scaled_2:%d - "
+			"pix_width_3:%d - pix_height_3:%d - refresh_rate_3:%lld - is_scaled_3:%d - LOG_END\n",
+			dramclk_khz_override,
+			fclk_khz_override,
+			new_clocks->dcfclk_khz,
+			new_clocks->dppclk_khz,
+			dispclk_khz_reg,
+			dppclk_khz_reg,
+			dprefclk_khz_reg,
+			dcfclk_khz_reg,
+			dtbclk_khz_reg,
+			fclk_khz_reg,
+			pix_clk_list[0], pix_clk_list[1], pix_clk_list[3], pix_clk_list[2],
+			mall_ss_size_bytes,
+			p_state_list[0], p_state_list[1], p_state_list[2], p_state_list[3],
+			disp_src_width_list[0], disp_src_height_list[0], disp_src_refresh_list[0], is_scaled_list[0],
+			disp_src_width_list[1], disp_src_height_list[1], disp_src_refresh_list[1], is_scaled_list[1],
+			disp_src_width_list[2], disp_src_height_list[2], disp_src_refresh_list[2], is_scaled_list[2],
+			disp_src_width_list[3], disp_src_height_list[3], disp_src_refresh_list[3], is_scaled_list[3]);
+	}
+}
+
+static void dcn32_update_clocks(struct clk_mgr *clk_mgr_base,
+			struct dc_state *context,
+			bool safe_to_lower)
+{
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+	struct dc_clocks *new_clocks = &context->bw_ctx.bw.dcn.clk;
+	struct dc *dc = clk_mgr_base->ctx->dc;
+	int display_count;
+	bool update_dppclk = false;
+	bool update_dispclk = false;
+	bool enter_display_off = false;
+	bool dpp_clock_lowered = false;
+	struct dmcu *dmcu = clk_mgr_base->ctx->dc->res_pool->dmcu;
+	bool force_reset = false;
+	bool update_uclk = false, update_fclk = false;
+	bool p_state_change_support;
+	bool fclk_p_state_change_support;
+
+	if (clk_mgr_base->clks.dispclk_khz == 0 ||
+			(dc->debug.force_clock_mode & 0x1)) {
+		/* This is from resume or boot up, if forced_clock cfg option used,
+		 * we bypass program dispclk and DPPCLK, but need set them for S3.
+		 */
+		force_reset = true;
+
+		dcn2_read_clocks_from_hw_dentist(clk_mgr_base);
+
+		/* Force_clock_mode 0x1:  force reset the clock even it is the same clock
+		 * as long as it is in Passive level.
+		 */
+	}
+	display_count = clk_mgr_helper_get_active_display_cnt(dc, context);
+
+	if (display_count == 0)
+		enter_display_off = true;
+
+	if (clk_mgr->smu_present) {
+		if (enter_display_off == safe_to_lower)
+			dcn30_smu_set_num_of_displays(clk_mgr, display_count);
+
+		clk_mgr_base->clks.fclk_prev_p_state_change_support = clk_mgr_base->clks.fclk_p_state_change_support;
+
+		fclk_p_state_change_support = new_clocks->fclk_p_state_change_support;
+
+		if (should_update_pstate_support(safe_to_lower, fclk_p_state_change_support, clk_mgr_base->clks.fclk_p_state_change_support) &&
+				!dc->work_arounds.clock_update_disable_mask.fclk) {
+			clk_mgr_base->clks.fclk_p_state_change_support = fclk_p_state_change_support;
+
+			/* To enable FCLK P-state switching, send FCLK_PSTATE_SUPPORTED message to PMFW */
+			if (clk_mgr_base->ctx->dce_version != DCN_VERSION_3_21 && clk_mgr_base->clks.fclk_p_state_change_support) {
+				/* Handle the code for sending a message to PMFW that FCLK P-state change is supported */
+				dcn32_smu_send_fclk_pstate_message(clk_mgr, FCLK_PSTATE_SUPPORTED);
+			}
+		}
+
+		if (dc->debug.force_min_dcfclk_mhz > 0)
+			new_clocks->dcfclk_khz = (new_clocks->dcfclk_khz > (dc->debug.force_min_dcfclk_mhz * 1000)) ?
+					new_clocks->dcfclk_khz : (dc->debug.force_min_dcfclk_mhz * 1000);
+
+		if (should_set_clock(safe_to_lower, new_clocks->dcfclk_khz, clk_mgr_base->clks.dcfclk_khz) &&
+				!dc->work_arounds.clock_update_disable_mask.dcfclk) {
+			clk_mgr_base->clks.dcfclk_khz = new_clocks->dcfclk_khz;
+			dcn32_smu_set_hard_min_by_freq(clk_mgr, PPCLK_DCFCLK, khz_to_mhz_ceil(clk_mgr_base->clks.dcfclk_khz));
+		}
+
+		if (should_set_clock(safe_to_lower, new_clocks->dcfclk_deep_sleep_khz, clk_mgr_base->clks.dcfclk_deep_sleep_khz) &&
+				!dc->work_arounds.clock_update_disable_mask.dcfclk_ds) {
+			clk_mgr_base->clks.dcfclk_deep_sleep_khz = new_clocks->dcfclk_deep_sleep_khz;
+			dcn30_smu_set_min_deep_sleep_dcef_clk(clk_mgr, khz_to_mhz_ceil(clk_mgr_base->clks.dcfclk_deep_sleep_khz));
+		}
+
+		if (should_set_clock(safe_to_lower, new_clocks->socclk_khz, clk_mgr_base->clks.socclk_khz))
+			/* We don't actually care about socclk, don't notify SMU of hard min */
+			clk_mgr_base->clks.socclk_khz = new_clocks->socclk_khz;
+
+		clk_mgr_base->clks.prev_p_state_change_support = clk_mgr_base->clks.p_state_change_support;
+		clk_mgr_base->clks.prev_num_ways = clk_mgr_base->clks.num_ways;
+
+		if (clk_mgr_base->clks.num_ways != new_clocks->num_ways &&
+				clk_mgr_base->clks.num_ways < new_clocks->num_ways) {
+			clk_mgr_base->clks.num_ways = new_clocks->num_ways;
+			dcn32_smu_send_cab_for_uclk_message(clk_mgr, clk_mgr_base->clks.num_ways);
+		}
+
+		p_state_change_support = new_clocks->p_state_change_support;
+		if (should_update_pstate_support(safe_to_lower, p_state_change_support, clk_mgr_base->clks.p_state_change_support) &&
+				!dc->work_arounds.clock_update_disable_mask.uclk) {
+			clk_mgr_base->clks.p_state_change_support = p_state_change_support;
+
+			/* to disable P-State switching, set UCLK min = max */
+			if (!clk_mgr_base->clks.p_state_change_support) {
+				if (dc->clk_mgr->dc_mode_softmax_enabled) {
+					/* On DCN32x we will never have the functional UCLK min above the softmax
+					 * since we calculate mode support based on softmax being the max UCLK
+					 * frequency.
+					 */
+					if (dc->debug.disable_dc_mode_overwrite) {
+						dcn30_smu_set_hard_max_by_freq(clk_mgr, PPCLK_UCLK, dc->clk_mgr->bw_params->max_memclk_mhz);
+						dcn32_smu_set_hard_min_by_freq(clk_mgr, PPCLK_UCLK, dc->clk_mgr->bw_params->max_memclk_mhz);
+					} else
+						dcn32_smu_set_hard_min_by_freq(clk_mgr, PPCLK_UCLK,
+								dc->clk_mgr->bw_params->dc_mode_softmax_memclk);
+				} else {
+					dcn32_smu_set_hard_min_by_freq(clk_mgr, PPCLK_UCLK, dc->clk_mgr->bw_params->max_memclk_mhz);
+				}
+			}
+		}
+
+		if (context->bw_ctx.bw.dcn.clk.fw_based_mclk_switching)
+			dcn32_smu_wait_for_dmub_ack_mclk(clk_mgr, true);
+		else
+			dcn32_smu_wait_for_dmub_ack_mclk(clk_mgr, false);
+
+		/* Always update saved value, even if new value not set due to P-State switching unsupported. Also check safe_to_lower for FCLK */
+		if (safe_to_lower && (clk_mgr_base->clks.fclk_p_state_change_support != clk_mgr_base->clks.fclk_prev_p_state_change_support)) {
+			update_fclk = true;
+		}
+
+		if (clk_mgr_base->ctx->dce_version != DCN_VERSION_3_21 && !clk_mgr_base->clks.fclk_p_state_change_support && update_fclk &&
+				!dc->work_arounds.clock_update_disable_mask.fclk) {
+			/* Handle code for sending a message to PMFW that FCLK P-state change is not supported */
+			dcn32_smu_send_fclk_pstate_message(clk_mgr, FCLK_PSTATE_NOTSUPPORTED);
+		}
+
+		/* Always update saved value, even if new value not set due to P-State switching unsupported */
+		if (should_set_clock(safe_to_lower, new_clocks->dramclk_khz, clk_mgr_base->clks.dramclk_khz) &&
+				!dc->work_arounds.clock_update_disable_mask.uclk) {
+			clk_mgr_base->clks.dramclk_khz = new_clocks->dramclk_khz;
+			update_uclk = true;
+		}
+
+		/* set UCLK to requested value if P-State switching is supported, or to re-enable P-State switching */
+		if (clk_mgr_base->clks.p_state_change_support &&
+				(update_uclk || !clk_mgr_base->clks.prev_p_state_change_support) &&
+				!dc->work_arounds.clock_update_disable_mask.uclk) {
+			if (dc->clk_mgr->dc_mode_softmax_enabled && dc->debug.disable_dc_mode_overwrite)
+				dcn30_smu_set_hard_max_by_freq(clk_mgr, PPCLK_UCLK,
+						max((int)dc->clk_mgr->bw_params->dc_mode_softmax_memclk, khz_to_mhz_ceil(clk_mgr_base->clks.dramclk_khz)));
+
+			dcn32_smu_set_hard_min_by_freq(clk_mgr, PPCLK_UCLK, khz_to_mhz_ceil(clk_mgr_base->clks.dramclk_khz));
+		}
+
+		if (clk_mgr_base->clks.num_ways != new_clocks->num_ways &&
+				clk_mgr_base->clks.num_ways > new_clocks->num_ways) {
+			clk_mgr_base->clks.num_ways = new_clocks->num_ways;
+			dcn32_smu_send_cab_for_uclk_message(clk_mgr, clk_mgr_base->clks.num_ways);
+		}
+	}
+
+	dcn32_update_dppclk_dispclk_freq(clk_mgr, new_clocks);
+	if (should_set_clock(safe_to_lower, new_clocks->dppclk_khz, clk_mgr_base->clks.dppclk_khz)) {
+		if (clk_mgr_base->clks.dppclk_khz > new_clocks->dppclk_khz)
+			dpp_clock_lowered = true;
+
+		clk_mgr_base->clks.dppclk_khz = new_clocks->dppclk_khz;
+
+		if (clk_mgr->smu_present && !dpp_clock_lowered)
+			/*
+			 * SMU uses discrete dppclk presets. We applied
+			 * the same formula to increase our dppclk_khz
+			 * to the next matching discrete value. By
+			 * contract, we should use the preset dppclk
+			 * floored in Mhz to describe the intended clock.
+			 */
+			dcn32_smu_set_hard_min_by_freq(clk_mgr, PPCLK_DPPCLK,
+					khz_to_mhz_floor(clk_mgr_base->clks.dppclk_khz));
+
+		update_dppclk = true;
+	}
+
+	if (should_set_clock(safe_to_lower, new_clocks->dispclk_khz, clk_mgr_base->clks.dispclk_khz)) {
+		clk_mgr_base->clks.dispclk_khz = new_clocks->dispclk_khz;
+
+		update_dispclk = true;
+	}
+
+	if (!new_clocks->dtbclk_en) {
+		new_clocks->ref_dtbclk_khz = clk_mgr_base->bw_params->clk_table.entries[0].dtbclk_mhz * 1000;
+	}
+
+	/* clock limits are received with MHz precision, divide by 1000 to prevent setting clocks at every call */
+	if (!dc->debug.disable_dtb_ref_clk_switch &&
+			should_set_clock(safe_to_lower, new_clocks->ref_dtbclk_khz / 1000, clk_mgr_base->clks.ref_dtbclk_khz / 1000)) {
+		/* DCCG requires KHz precision for DTBCLK */
+		clk_mgr_base->clks.ref_dtbclk_khz =
+				dcn32_smu_set_hard_min_by_freq(clk_mgr, PPCLK_DTBCLK, khz_to_mhz_ceil(new_clocks->ref_dtbclk_khz));
+
+		dcn32_update_clocks_update_dtb_dto(clk_mgr, context, clk_mgr_base->clks.ref_dtbclk_khz);
+	}
+
+	if (dc->config.forced_clocks == false || (force_reset && safe_to_lower)) {
+		if (dpp_clock_lowered) {
+			/* if clock is being lowered, increase DTO before lowering refclk */
+			dcn32_update_clocks_update_dpp_dto(clk_mgr, context, safe_to_lower);
+			dcn32_update_clocks_update_dentist(clk_mgr, context);
+			if (clk_mgr->smu_present)
+				/*
+				 * SMU uses discrete dppclk presets. We applied
+				 * the same formula to increase our dppclk_khz
+				 * to the next matching discrete value. By
+				 * contract, we should use the preset dppclk
+				 * floored in Mhz to describe the intended clock.
+				 */
+				dcn32_smu_set_hard_min_by_freq(clk_mgr, PPCLK_DPPCLK,
+						khz_to_mhz_floor(clk_mgr_base->clks.dppclk_khz));
+		} else {
+			/* if clock is being raised, increase refclk before lowering DTO */
+			if (update_dppclk || update_dispclk)
+				dcn32_update_clocks_update_dentist(clk_mgr, context);
+			/* There is a check inside dcn20_update_clocks_update_dpp_dto which ensures
+			 * that we do not lower dto when it is not safe to lower. We do not need to
+			 * compare the current and new dppclk before calling this function.
+			 */
+			dcn32_update_clocks_update_dpp_dto(clk_mgr, context, safe_to_lower);
+		}
+	}
+
+	if (update_dispclk && dmcu && dmcu->funcs->is_dmcu_initialized(dmcu))
+		/*update dmcu for wait_loop count*/
+		dmcu->funcs->set_psr_wait_loop(dmcu,
+				clk_mgr_base->clks.dispclk_khz / 1000 / 7);
+
+	if (dc->config.enable_auto_dpm_test_logs) {
+	    dcn32_auto_dpm_test_log(new_clocks, clk_mgr, context);
+	}
+}
+
+static uint32_t dcn32_get_vco_frequency_from_reg(struct clk_mgr_internal *clk_mgr)
+{
+		struct fixed31_32 pll_req;
+		uint32_t pll_req_reg = 0;
+
+		/* get FbMult value */
+		if (ASICREV_IS_GC_11_0_2(clk_mgr->base.ctx->asic_id.hw_internal_rev))
+			pll_req_reg = REG_READ(CLK0_CLK_PLL_REQ);
+		else
+			pll_req_reg = REG_READ(CLK1_CLK_PLL_REQ);
+
+		/* set up a fixed-point number
+		 * this works because the int part is on the right edge of the register
+		 * and the frac part is on the left edge
+		 */
+		pll_req = dc_fixpt_from_int(pll_req_reg & clk_mgr->clk_mgr_mask->FbMult_int);
+		pll_req.value |= pll_req_reg & clk_mgr->clk_mgr_mask->FbMult_frac;
+
+		/* multiply by REFCLK period */
+		pll_req = dc_fixpt_mul_int(pll_req, clk_mgr->dfs_ref_freq_khz);
+
+		return dc_fixpt_floor(pll_req);
+}
+
+static void dcn32_dump_clk_registers(struct clk_state_registers_and_bypass *regs_and_bypass,
+		struct clk_mgr *clk_mgr_base, struct clk_log_info *log_info)
+{
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+	uint32_t dprefclk_did = 0;
+	uint32_t dcfclk_did = 0;
+	uint32_t dtbclk_did = 0;
+	uint32_t dispclk_did = 0;
+	uint32_t dppclk_did = 0;
+	uint32_t target_div = 0;
+
+	if (ASICREV_IS_GC_11_0_2(clk_mgr->base.ctx->asic_id.hw_internal_rev)) {
+		/* DFS Slice 0 is used for DISPCLK */
+		dispclk_did = REG_READ(CLK0_CLK0_DFS_CNTL);
+		/* DFS Slice 1 is used for DPPCLK */
+		dppclk_did = REG_READ(CLK0_CLK1_DFS_CNTL);
+		/* DFS Slice 2 is used for DPREFCLK */
+		dprefclk_did = REG_READ(CLK0_CLK2_DFS_CNTL);
+		/* DFS Slice 3 is used for DCFCLK */
+		dcfclk_did = REG_READ(CLK0_CLK3_DFS_CNTL);
+		/* DFS Slice 4 is used for DTBCLK */
+		dtbclk_did = REG_READ(CLK0_CLK4_DFS_CNTL);
+	} else {
+		/* DFS Slice 0 is used for DISPCLK */
+		dispclk_did = REG_READ(CLK1_CLK0_DFS_CNTL);
+		/* DFS Slice 1 is used for DPPCLK */
+		dppclk_did = REG_READ(CLK1_CLK1_DFS_CNTL);
+		/* DFS Slice 2 is used for DPREFCLK */
+		dprefclk_did = REG_READ(CLK1_CLK2_DFS_CNTL);
+		/* DFS Slice 3 is used for DCFCLK */
+		dcfclk_did = REG_READ(CLK1_CLK3_DFS_CNTL);
+		/* DFS Slice 4 is used for DTBCLK */
+		dtbclk_did = REG_READ(CLK1_CLK4_DFS_CNTL);
+	}
+
+	/* Convert DISPCLK DFS Slice DID to divider*/
+	target_div = dentist_get_divider_from_did(dispclk_did);
+	//Get dispclk in khz
+	regs_and_bypass->dispclk = (DENTIST_DIVIDER_RANGE_SCALE_FACTOR
+			* clk_mgr->base.dentist_vco_freq_khz) / target_div;
+
+	/* Convert DISPCLK DFS Slice DID to divider*/
+	target_div = dentist_get_divider_from_did(dppclk_did);
+	//Get dppclk in khz
+	regs_and_bypass->dppclk = (DENTIST_DIVIDER_RANGE_SCALE_FACTOR
+			* clk_mgr->base.dentist_vco_freq_khz) / target_div;
+
+	/* Convert DPREFCLK DFS Slice DID to divider*/
+	target_div = dentist_get_divider_from_did(dprefclk_did);
+	//Get dprefclk in khz
+	regs_and_bypass->dprefclk = (DENTIST_DIVIDER_RANGE_SCALE_FACTOR
+			* clk_mgr->base.dentist_vco_freq_khz) / target_div;
+
+	/* Convert DCFCLK DFS Slice DID to divider*/
+	target_div = dentist_get_divider_from_did(dcfclk_did);
+	//Get dcfclk in khz
+	regs_and_bypass->dcfclk = (DENTIST_DIVIDER_RANGE_SCALE_FACTOR
+			* clk_mgr->base.dentist_vco_freq_khz) / target_div;
+
+	/* Convert DTBCLK DFS Slice DID to divider*/
+	target_div = dentist_get_divider_from_did(dtbclk_did);
+	//Get dtbclk in khz
+	regs_and_bypass->dtbclk = (DENTIST_DIVIDER_RANGE_SCALE_FACTOR
+			* clk_mgr->base.dentist_vco_freq_khz) / target_div;
+}
+
+static void dcn32_clock_read_ss_info(struct clk_mgr_internal *clk_mgr)
+{
+	struct dc_bios *bp = clk_mgr->base.ctx->dc_bios;
+	int ss_info_num = bp->funcs->get_ss_entry_number(
+			bp, AS_SIGNAL_TYPE_GPU_PLL);
+
+	if (ss_info_num) {
+		struct spread_spectrum_info info = { { 0 } };
+		enum bp_result result = bp->funcs->get_spread_spectrum_info(
+				bp, AS_SIGNAL_TYPE_GPU_PLL, 0, &info);
+
+		/* SSInfo.spreadSpectrumPercentage !=0 would be sign
+		 * that SS is enabled
+		 */
+		if (result == BP_RESULT_OK &&
+				info.spread_spectrum_percentage != 0) {
+			clk_mgr->ss_on_dprefclk = true;
+			clk_mgr->dprefclk_ss_divider = info.spread_percentage_divider;
+
+			if (info.type.CENTER_MODE == 0) {
+				/* Currently for DP Reference clock we
+				 * need only SS percentage for
+				 * downspread
+				 */
+				clk_mgr->dprefclk_ss_percentage =
+						info.spread_spectrum_percentage;
+			}
+		}
+	}
+}
+static void dcn32_notify_wm_ranges(struct clk_mgr *clk_mgr_base)
+{
+	unsigned int i;
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+	WatermarksExternal_t *table = (WatermarksExternal_t *) clk_mgr->wm_range_table;
+
+	if (!clk_mgr->smu_present)
+		return;
+
+	if (!table)
+		return;
+
+	memset(table, 0, sizeof(*table));
+
+	/* collect valid ranges, place in pmfw table */
+	for (i = 0; i < WM_SET_COUNT; i++)
+		if (clk_mgr->base.bw_params->wm_table.nv_entries[i].valid) {
+			table->Watermarks.WatermarkRow[i].WmSetting = i;
+			table->Watermarks.WatermarkRow[i].Flags = clk_mgr->base.bw_params->wm_table.nv_entries[i].pmfw_breakdown.wm_type;
+		}
+	dcn30_smu_set_dram_addr_high(clk_mgr, clk_mgr->wm_range_table_addr >> 32);
+	dcn30_smu_set_dram_addr_low(clk_mgr, clk_mgr->wm_range_table_addr & 0xFFFFFFFF);
+	dcn32_smu_transfer_wm_table_dram_2_smu(clk_mgr);
+}
+
+/* Set min memclk to minimum, either constrained by the current mode or DPM0 */
+static void dcn32_set_hard_min_memclk(struct clk_mgr *clk_mgr_base, bool current_mode)
+{
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+
+	if (!clk_mgr->smu_present)
+		return;
+
+	if (current_mode) {
+		if (clk_mgr_base->clks.p_state_change_support)
+			dcn32_smu_set_hard_min_by_freq(clk_mgr, PPCLK_UCLK,
+					khz_to_mhz_ceil(clk_mgr_base->clks.dramclk_khz));
+		else
+			dcn32_smu_set_hard_min_by_freq(clk_mgr, PPCLK_UCLK,
+					clk_mgr_base->bw_params->max_memclk_mhz);
+	} else {
+		dcn32_smu_set_hard_min_by_freq(clk_mgr, PPCLK_UCLK,
+				clk_mgr_base->bw_params->clk_table.entries[0].memclk_mhz);
+	}
+}
+
+/* Set max memclk to highest DPM value */
+static void dcn32_set_hard_max_memclk(struct clk_mgr *clk_mgr_base)
+{
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+
+	if (!clk_mgr->smu_present)
+		return;
+
+	dcn30_smu_set_hard_max_by_freq(clk_mgr, PPCLK_UCLK, clk_mgr_base->bw_params->max_memclk_mhz);
+}
+
+/* Get current memclk states, update bounding box */
+static void dcn32_get_memclk_states_from_smu(struct clk_mgr *clk_mgr_base)
+{
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+	struct clk_limit_num_entries *num_entries_per_clk = &clk_mgr_base->bw_params->clk_table.num_entries_per_clk;
+	unsigned int num_levels;
+
+	if (!clk_mgr->smu_present)
+		return;
+
+	/* Refresh memclk and fclk states */
+	dcn32_init_single_clock(clk_mgr, PPCLK_UCLK,
+			&clk_mgr_base->bw_params->clk_table.entries[0].memclk_mhz,
+			&num_entries_per_clk->num_memclk_levels);
+	clk_mgr_base->bw_params->dc_mode_limit.memclk_mhz = dcn30_smu_get_dc_mode_max_dpm_freq(clk_mgr, PPCLK_UCLK);
+	clk_mgr_base->bw_params->dc_mode_softmax_memclk = clk_mgr_base->bw_params->dc_mode_limit.memclk_mhz;
+
+	/* memclk must have at least one level */
+	num_entries_per_clk->num_memclk_levels = num_entries_per_clk->num_memclk_levels ? num_entries_per_clk->num_memclk_levels : 1;
+
+	dcn32_init_single_clock(clk_mgr, PPCLK_FCLK,
+			&clk_mgr_base->bw_params->clk_table.entries[0].fclk_mhz,
+			&num_entries_per_clk->num_fclk_levels);
+	clk_mgr_base->bw_params->dc_mode_limit.fclk_mhz = dcn30_smu_get_dc_mode_max_dpm_freq(clk_mgr, PPCLK_FCLK);
+
+	num_levels = max(num_entries_per_clk->num_memclk_levels, num_entries_per_clk->num_fclk_levels);
+
+	clk_mgr_base->bw_params->max_memclk_mhz =
+			clk_mgr_base->bw_params->clk_table.entries[num_entries_per_clk->num_memclk_levels - 1].memclk_mhz;
+	clk_mgr_base->bw_params->clk_table.num_entries = num_levels ? num_levels : 1;
+
+	if (clk_mgr->dpm_present && !num_levels)
+		clk_mgr->dpm_present = false;
+
+	if (!clk_mgr->dpm_present)
+		dcn32_patch_dpm_table(clk_mgr_base->bw_params);
+
+	DC_FP_START();
+	/* Refresh bounding box */
+	clk_mgr_base->ctx->dc->res_pool->funcs->update_bw_bounding_box(
+			clk_mgr->base.ctx->dc, clk_mgr_base->bw_params);
+	DC_FP_END();
+}
+
+static bool dcn32_are_clock_states_equal(struct dc_clocks *a,
+					struct dc_clocks *b)
+{
+	if (a->dispclk_khz != b->dispclk_khz)
+		return false;
+	else if (a->dppclk_khz != b->dppclk_khz)
+		return false;
+	else if (a->dcfclk_khz != b->dcfclk_khz)
+		return false;
+	else if (a->dcfclk_deep_sleep_khz != b->dcfclk_deep_sleep_khz)
+		return false;
+	else if (a->dramclk_khz != b->dramclk_khz)
+		return false;
+	else if (a->p_state_change_support != b->p_state_change_support)
+		return false;
+	else if (a->fclk_p_state_change_support != b->fclk_p_state_change_support)
+		return false;
+
+	return true;
+}
+
+static void dcn32_enable_pme_wa(struct clk_mgr *clk_mgr_base)
+{
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+
+	if (!clk_mgr->smu_present)
+		return;
+
+	dcn32_smu_set_pme_workaround(clk_mgr);
+}
+
+static bool dcn32_is_smu_present(struct clk_mgr *clk_mgr_base)
+{
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+	return clk_mgr->smu_present;
+}
+
+static void dcn32_set_max_memclk(struct clk_mgr *clk_mgr_base, unsigned int memclk_mhz)
+{
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+
+	if (!clk_mgr->smu_present)
+		return;
+
+	dcn30_smu_set_hard_max_by_freq(clk_mgr, PPCLK_UCLK, memclk_mhz);
+}
+
+static void dcn32_set_min_memclk(struct clk_mgr *clk_mgr_base, unsigned int memclk_mhz)
+{
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+
+	if (!clk_mgr->smu_present)
+		return;
+
+	dcn32_smu_set_hard_min_by_freq(clk_mgr, PPCLK_UCLK, memclk_mhz);
+}
+
+static struct clk_mgr_funcs dcn32_funcs = {
+		.get_dp_ref_clk_frequency = dce12_get_dp_ref_freq_khz,
+		.get_dtb_ref_clk_frequency = dcn31_get_dtb_ref_freq_khz,
+		.update_clocks = dcn32_update_clocks,
+		.dump_clk_registers = dcn32_dump_clk_registers,
+		.init_clocks = dcn32_init_clocks,
+		.notify_wm_ranges = dcn32_notify_wm_ranges,
+		.set_hard_min_memclk = dcn32_set_hard_min_memclk,
+		.set_hard_max_memclk = dcn32_set_hard_max_memclk,
+		.set_max_memclk = dcn32_set_max_memclk,
+		.set_min_memclk = dcn32_set_min_memclk,
+		.get_memclk_states_from_smu = dcn32_get_memclk_states_from_smu,
+		.are_clock_states_equal = dcn32_are_clock_states_equal,
+		.enable_pme_wa = dcn32_enable_pme_wa,
+		.is_smu_present = dcn32_is_smu_present,
+		.get_dispclk_from_dentist = dcn32_get_dispclk_from_dentist,
+};
+
+void dcn32_clk_mgr_construct(
+		struct dc_context *ctx,
+		struct clk_mgr_internal *clk_mgr,
+		struct pp_smu_funcs *pp_smu,
+		struct dccg *dccg)
+{
+	struct clk_log_info log_info = {0};
+
+	clk_mgr->base.ctx = ctx;
+	clk_mgr->base.funcs = &dcn32_funcs;
+	if (ASICREV_IS_GC_11_0_2(clk_mgr->base.ctx->asic_id.hw_internal_rev)) {
+		clk_mgr->regs = &clk_mgr_regs_dcn321;
+		clk_mgr->clk_mgr_shift = &clk_mgr_shift_dcn321;
+		clk_mgr->clk_mgr_mask = &clk_mgr_mask_dcn321;
+	} else {
+		clk_mgr->regs = &clk_mgr_regs_dcn32;
+		clk_mgr->clk_mgr_shift = &clk_mgr_shift_dcn32;
+		clk_mgr->clk_mgr_mask = &clk_mgr_mask_dcn32;
+	}
+
+	clk_mgr->dccg = dccg;
+	clk_mgr->dfs_bypass_disp_clk = 0;
+
+	clk_mgr->dprefclk_ss_percentage = 0;
+	clk_mgr->dprefclk_ss_divider = 1000;
+	clk_mgr->ss_on_dprefclk = false;
+	clk_mgr->dfs_ref_freq_khz = 100000;
+
+	/* Changed from DCN3.2_clock_frequency doc to match
+	 * dcn32_dump_clk_registers from 4 * dentist_vco_freq_khz /
+	 * dprefclk DID divider
+	 */
+	clk_mgr->base.dprefclk_khz = 716666;
+	if (ctx->dc->debug.disable_dtb_ref_clk_switch) {
+		//initialize DTB ref clock value if DPM disabled
+		if (ctx->dce_version == DCN_VERSION_3_21)
+			clk_mgr->base.clks.ref_dtbclk_khz = 477800;
+		else
+			clk_mgr->base.clks.ref_dtbclk_khz = 268750;
+	}
+
+
+	/* integer part is now VCO frequency in kHz */
+	clk_mgr->base.dentist_vco_freq_khz = dcn32_get_vco_frequency_from_reg(clk_mgr);
+
+	/* in case we don't get a value from the register, use default */
+	if (clk_mgr->base.dentist_vco_freq_khz == 0)
+		clk_mgr->base.dentist_vco_freq_khz = 4300000; /* Updated as per HW docs */
+
+	dcn32_dump_clk_registers(&clk_mgr->base.boot_snapshot, &clk_mgr->base, &log_info);
+
+	if (ctx->dc->debug.disable_dtb_ref_clk_switch &&
+			clk_mgr->base.clks.ref_dtbclk_khz != clk_mgr->base.boot_snapshot.dtbclk) {
+		clk_mgr->base.clks.ref_dtbclk_khz = clk_mgr->base.boot_snapshot.dtbclk;
+	}
+
+	if (clk_mgr->base.boot_snapshot.dprefclk != 0) {
+		clk_mgr->base.dprefclk_khz = clk_mgr->base.boot_snapshot.dprefclk;
+	}
+	dcn32_clock_read_ss_info(clk_mgr);
+
+	clk_mgr->dfs_bypass_enabled = false;
+
+	clk_mgr->smu_present = false;
+
+	clk_mgr->base.bw_params = kzalloc(sizeof(*clk_mgr->base.bw_params), GFP_KERNEL);
+	if (!clk_mgr->base.bw_params) {
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+
+	/* need physical address of table to give to PMFW */
+	clk_mgr->wm_range_table = dm_helpers_allocate_gpu_mem(clk_mgr->base.ctx,
+			DC_MEM_ALLOC_TYPE_GART, sizeof(WatermarksExternal_t),
+			&clk_mgr->wm_range_table_addr);
+	if (!clk_mgr->wm_range_table) {
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+}
+
+void dcn32_clk_mgr_destroy(struct clk_mgr_internal *clk_mgr)
+{
+	kfree(clk_mgr->base.bw_params);
+
+	if (clk_mgr->wm_range_table)
+		dm_helpers_free_gpu_mem(clk_mgr->base.ctx, DC_MEM_ALLOC_TYPE_GART,
+				clk_mgr->wm_range_table);
+}
+
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn32/dcn32_clk_mgr.h b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn32/dcn32_clk_mgr.h
new file mode 100644
index 000000000000..186daada7b03
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn32/dcn32_clk_mgr.h
@@ -0,0 +1,42 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+#ifndef __DCN32_CLK_MGR_H_
+#define __DCN32_CLK_MGR_H_
+
+void dcn32_init_clocks(struct clk_mgr *clk_mgr_base);
+
+void dcn32_clk_mgr_construct(struct dc_context *ctx,
+		struct clk_mgr_internal *clk_mgr,
+		struct pp_smu_funcs *pp_smu,
+		struct dccg *dccg);
+
+void dcn32_update_clocks_update_dpp_dto(struct clk_mgr_internal *clk_mgr,
+		struct dc_state *context, bool safe_to_lower);
+
+void dcn32_clk_mgr_destroy(struct clk_mgr_internal *clk_mgr);
+
+
+
+#endif /* __DCN32_CLK_MGR_H_ */
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn32/dcn32_clk_mgr_smu_msg.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn32/dcn32_clk_mgr_smu_msg.c
new file mode 100644
index 000000000000..5d80fdf63ffc
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn32/dcn32_clk_mgr_smu_msg.c
@@ -0,0 +1,308 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#include "dcn32_clk_mgr_smu_msg.h"
+
+#include "clk_mgr_internal.h"
+#include "reg_helper.h"
+#include "dalsmc.h"
+#include "dcn32_smu13_driver_if.h"
+
+#define mmDAL_MSG_REG  0x1628A
+#define mmDAL_ARG_REG  0x16273
+#define mmDAL_RESP_REG 0x16274
+
+#define REG(reg_name) \
+	mm ## reg_name
+
+#include "logger_types.h"
+
+#define smu_print(str, ...) {DC_LOG_SMU(str, ##__VA_ARGS__); }
+
+
+/*
+ * Function to be used instead of REG_WAIT macro because the wait ends when
+ * the register is NOT EQUAL to zero, and because the translation in msg_if.h
+ * won't work with REG_WAIT.
+ */
+static uint32_t dcn32_smu_wait_for_response(struct clk_mgr_internal *clk_mgr, unsigned int delay_us, unsigned int max_retries)
+{
+	const uint32_t initial_max_retries = max_retries;
+	uint32_t reg = 0;
+
+	do {
+		reg = REG_READ(DAL_RESP_REG);
+		if (reg)
+			break;
+
+		if (delay_us >= 1000)
+			msleep(delay_us/1000);
+		else if (delay_us > 0)
+			udelay(delay_us);
+	} while (max_retries--);
+
+	TRACE_SMU_MSG_DELAY(0, 0, delay_us * (initial_max_retries - max_retries), clk_mgr->base.ctx);
+
+
+	return reg;
+}
+
+static bool dcn32_smu_send_msg_with_param(struct clk_mgr_internal *clk_mgr, uint32_t msg_id, uint32_t param_in, uint32_t *param_out)
+{
+	/* Wait for response register to be ready */
+	dcn32_smu_wait_for_response(clk_mgr, 10, 200000);
+
+	/* Clear response register */
+	REG_WRITE(DAL_RESP_REG, 0);
+
+	/* Set the parameter register for the SMU message */
+	REG_WRITE(DAL_ARG_REG, param_in);
+
+	/* Trigger the message transaction by writing the message ID */
+	REG_WRITE(DAL_MSG_REG, msg_id);
+
+	TRACE_SMU_MSG(msg_id, param_in, clk_mgr->base.ctx);
+
+	/* Wait for response */
+	if (dcn32_smu_wait_for_response(clk_mgr, 10, 200000) == DALSMC_Result_OK) {
+		if (param_out)
+			*param_out = REG_READ(DAL_ARG_REG);
+
+		return true;
+	}
+
+	return false;
+}
+
+/*
+ * Use these functions to return back delay information so we can aggregate the total
+ *  delay when requesting hardmin clk
+ *
+ * dcn32_smu_wait_for_response_delay
+ * dcn32_smu_send_msg_with_param_delay
+ *
+ */
+static uint32_t dcn32_smu_wait_for_response_delay(struct clk_mgr_internal *clk_mgr, unsigned int delay_us, unsigned int max_retries, unsigned int *total_delay_us)
+{
+	uint32_t reg = 0;
+	*total_delay_us = 0;
+
+	do {
+		reg = REG_READ(DAL_RESP_REG);
+		if (reg)
+			break;
+
+		if (delay_us >= 1000)
+			msleep(delay_us/1000);
+		else if (delay_us > 0)
+			udelay(delay_us);
+		*total_delay_us += delay_us;
+	} while (max_retries--);
+
+	TRACE_SMU_MSG_DELAY(0, 0, *total_delay_us, clk_mgr->base.ctx);
+
+	return reg;
+}
+
+static bool dcn32_smu_send_msg_with_param_delay(struct clk_mgr_internal *clk_mgr, uint32_t msg_id, uint32_t param_in, uint32_t *param_out, unsigned int *total_delay_us)
+{
+	unsigned int delay1_us, delay2_us;
+	*total_delay_us = 0;
+
+	/* Wait for response register to be ready */
+	dcn32_smu_wait_for_response_delay(clk_mgr, 10, 200000, &delay1_us);
+
+	/* Clear response register */
+	REG_WRITE(DAL_RESP_REG, 0);
+
+	/* Set the parameter register for the SMU message */
+	REG_WRITE(DAL_ARG_REG, param_in);
+
+	/* Trigger the message transaction by writing the message ID */
+	REG_WRITE(DAL_MSG_REG, msg_id);
+
+	TRACE_SMU_MSG(msg_id, param_in, clk_mgr->base.ctx);
+
+	/* Wait for response */
+	if (dcn32_smu_wait_for_response_delay(clk_mgr, 10, 200000, &delay2_us) == DALSMC_Result_OK) {
+		if (param_out)
+			*param_out = REG_READ(DAL_ARG_REG);
+
+		*total_delay_us = delay1_us + delay2_us;
+		return true;
+	}
+
+	*total_delay_us = delay1_us + 2000000;
+	return false;
+}
+
+void dcn32_smu_send_fclk_pstate_message(struct clk_mgr_internal *clk_mgr, bool enable)
+{
+	smu_print("FCLK P-state support value is : %d\n", enable);
+
+	dcn32_smu_send_msg_with_param(clk_mgr,
+			DALSMC_MSG_SetFclkSwitchAllow, enable ? FCLK_PSTATE_SUPPORTED : FCLK_PSTATE_NOTSUPPORTED, NULL);
+}
+
+void dcn32_smu_send_cab_for_uclk_message(struct clk_mgr_internal *clk_mgr, unsigned int num_ways)
+{
+	uint32_t param = (num_ways << 1) | (num_ways > 0);
+
+	dcn32_smu_send_msg_with_param(clk_mgr, DALSMC_MSG_SetCabForUclkPstate, param, NULL);
+	smu_print("Numways for SubVP : %d\n", num_ways);
+}
+
+void dcn32_smu_transfer_wm_table_dram_2_smu(struct clk_mgr_internal *clk_mgr)
+{
+	smu_print("SMU Transfer WM table DRAM 2 SMU\n");
+
+	dcn32_smu_send_msg_with_param(clk_mgr,
+			DALSMC_MSG_TransferTableDram2Smu, TABLE_WATERMARKS, NULL);
+}
+
+void dcn32_smu_set_pme_workaround(struct clk_mgr_internal *clk_mgr)
+{
+	smu_print("SMU Set PME workaround\n");
+
+	dcn32_smu_send_msg_with_param(clk_mgr,
+		DALSMC_MSG_BacoAudioD3PME, 0, NULL);
+}
+
+/* Check PMFW version if it supports ReturnHardMinStatus message */
+static bool dcn32_get_hard_min_status_supported(struct clk_mgr_internal *clk_mgr)
+{
+	if (ASICREV_IS_GC_11_0_0(clk_mgr->base.ctx->asic_id.hw_internal_rev)) {
+		if (clk_mgr->smu_ver >= 0x4e6a00)
+			return true;
+	} else if (ASICREV_IS_GC_11_0_2(clk_mgr->base.ctx->asic_id.hw_internal_rev)) {
+		if (clk_mgr->smu_ver >= 0x524e00)
+			return true;
+	} else { /* ASICREV_IS_GC_11_0_3 */
+		if (clk_mgr->smu_ver >= 0x503900)
+			return true;
+	}
+	return false;
+}
+
+/* Returns the clocks which were fulfilled by the DAL hard min arbiter in PMFW */
+static unsigned int dcn32_smu_get_hard_min_status(struct clk_mgr_internal *clk_mgr, bool *no_timeout, unsigned int *total_delay_us)
+{
+	uint32_t response = 0;
+
+	/* bits 23:16 for clock type, lower 16 bits for frequency in MHz */
+	uint32_t param = 0;
+
+	*no_timeout = dcn32_smu_send_msg_with_param_delay(clk_mgr,
+			DALSMC_MSG_ReturnHardMinStatus, param, &response, total_delay_us);
+
+	smu_print("SMU Get hard min status: no_timeout %d delay %d us clk bits %x\n",
+		*no_timeout, *total_delay_us, response);
+
+	return response;
+}
+
+static bool dcn32_smu_wait_get_hard_min_status(struct clk_mgr_internal *clk_mgr,
+	uint32_t clk)
+{
+	int readDalHardMinClkBits, checkDalHardMinClkBits;
+	unsigned int total_delay_us, read_total_delay_us;
+	bool no_timeout, hard_min_done;
+
+	static unsigned int cur_wait_get_hard_min_max_us;
+	static unsigned int cur_wait_get_hard_min_max_timeouts;
+
+	checkDalHardMinClkBits = CHECK_HARD_MIN_CLK_DPREFCLK;
+	if (clk == PPCLK_DISPCLK)
+		checkDalHardMinClkBits |= CHECK_HARD_MIN_CLK_DISPCLK;
+	if (clk == PPCLK_DPPCLK)
+		checkDalHardMinClkBits |= CHECK_HARD_MIN_CLK_DPPCLK;
+	if (clk == PPCLK_DCFCLK)
+		checkDalHardMinClkBits |= CHECK_HARD_MIN_CLK_DCFCLK;
+	if (clk == PPCLK_DTBCLK)
+		checkDalHardMinClkBits |= CHECK_HARD_MIN_CLK_DTBCLK;
+	if (clk == PPCLK_UCLK)
+		checkDalHardMinClkBits |= CHECK_HARD_MIN_CLK_UCLK;
+
+	if (checkDalHardMinClkBits == CHECK_HARD_MIN_CLK_DPREFCLK)
+		return 0;
+
+	total_delay_us = 0;
+	hard_min_done = false;
+	while (1) {
+		readDalHardMinClkBits = dcn32_smu_get_hard_min_status(clk_mgr, &no_timeout, &read_total_delay_us);
+		total_delay_us += read_total_delay_us;
+		if (checkDalHardMinClkBits == (readDalHardMinClkBits & checkDalHardMinClkBits)) {
+			hard_min_done = true;
+			break;
+		}
+
+
+		if (total_delay_us >= 2000000) {
+			cur_wait_get_hard_min_max_timeouts++;
+			smu_print("SMU Wait get hard min status: %d timeouts\n", cur_wait_get_hard_min_max_timeouts);
+			break;
+		}
+		msleep(1);
+		total_delay_us += 1000;
+	}
+
+	if (total_delay_us > cur_wait_get_hard_min_max_us)
+		cur_wait_get_hard_min_max_us = total_delay_us;
+
+	smu_print("SMU Wait get hard min status: no_timeout %d, delay %d us, max %d us, read %x, check %x\n",
+		no_timeout, total_delay_us, cur_wait_get_hard_min_max_us, readDalHardMinClkBits, checkDalHardMinClkBits);
+
+	return hard_min_done;
+}
+
+/* Returns the actual frequency that was set in MHz, 0 on failure */
+unsigned int dcn32_smu_set_hard_min_by_freq(struct clk_mgr_internal *clk_mgr, uint32_t clk, uint16_t freq_mhz)
+{
+	uint32_t response = 0;
+	bool hard_min_done = false;
+
+	/* bits 23:16 for clock type, lower 16 bits for frequency in MHz */
+	uint32_t param = (clk << 16) | freq_mhz;
+
+	smu_print("SMU Set hard min by freq: clk = %d, freq_mhz = %d MHz\n", clk, freq_mhz);
+
+	dcn32_smu_send_msg_with_param(clk_mgr,
+		DALSMC_MSG_SetHardMinByFreq, param, &response);
+
+	if (dcn32_get_hard_min_status_supported(clk_mgr)) {
+		hard_min_done = dcn32_smu_wait_get_hard_min_status(clk_mgr, clk);
+		smu_print("SMU Frequency set = %d KHz hard_min_done %d\n", response, hard_min_done);
+	} else
+		smu_print("SMU Frequency set = %d KHz\n", response);
+
+	return response;
+}
+
+void dcn32_smu_wait_for_dmub_ack_mclk(struct clk_mgr_internal *clk_mgr, bool enable)
+{
+	smu_print("PMFW to wait for DMCUB ack for MCLK : %d\n", enable);
+
+	dcn32_smu_send_msg_with_param(clk_mgr, 0x14, enable ? 1 : 0, NULL);
+}
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn32/dcn32_clk_mgr_smu_msg.h b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn32/dcn32_clk_mgr_smu_msg.h
new file mode 100644
index 000000000000..5c44ab0e8667
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn32/dcn32_clk_mgr_smu_msg.h
@@ -0,0 +1,46 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __DCN32_CLK_MGR_SMU_MSG_H_
+#define __DCN32_CLK_MGR_SMU_MSG_H_
+
+#include "core_types.h"
+#include "dcn30/dcn30_clk_mgr_smu_msg.h"
+
+#define FCLK_PSTATE_NOTSUPPORTED       0x00
+#define FCLK_PSTATE_SUPPORTED          0x01
+
+/* TODO Remove this MSG ID define after it becomes available in dalsmc */
+#define DALSMC_MSG_SetCabForUclkPstate	0x12
+#define DALSMC_Result_OK				0x1
+
+void dcn32_smu_send_fclk_pstate_message(struct clk_mgr_internal *clk_mgr, bool enable);
+void dcn32_smu_send_cab_for_uclk_message(struct clk_mgr_internal *clk_mgr, unsigned int num_ways);
+void dcn32_smu_transfer_wm_table_dram_2_smu(struct clk_mgr_internal *clk_mgr);
+void dcn32_smu_set_pme_workaround(struct clk_mgr_internal *clk_mgr);
+unsigned int dcn32_smu_set_hard_min_by_freq(struct clk_mgr_internal *clk_mgr, uint32_t clk, uint16_t freq_mhz);
+void dcn32_smu_wait_for_dmub_ack_mclk(struct clk_mgr_internal *clk_mgr, bool enable);
+
+#endif /* __DCN32_CLK_MGR_SMU_MSG_H_ */
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn32/dcn32_smu13_driver_if.h b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn32/dcn32_smu13_driver_if.h
new file mode 100644
index 000000000000..8d54865bbd5d
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn32/dcn32_smu13_driver_if.h
@@ -0,0 +1,64 @@
+// SPDX-License-Identifier: MIT
+/* Copyright © 2022-2024 Advanced Micro Devices, Inc. All rights reserved. */
+
+#define SMU13_DRIVER_IF_VERSION  0x18
+
+//Only Clks that have DPM descriptors are listed here
+typedef enum {
+	PPCLK_GFXCLK = 0,
+	PPCLK_SOCCLK,
+	PPCLK_UCLK,
+	PPCLK_FCLK,
+	PPCLK_DCLK_0,
+	PPCLK_VCLK_0,
+	PPCLK_DCLK_1,
+	PPCLK_VCLK_1,
+	PPCLK_DISPCLK,
+	PPCLK_DPPCLK,
+	PPCLK_DPREFCLK,
+	PPCLK_DCFCLK,
+	PPCLK_DTBCLK,
+	PPCLK_COUNT,
+} PPCLK_e;
+
+typedef struct {
+	uint8_t  WmSetting;
+	uint8_t  Flags;
+	uint8_t  Padding[2];
+
+} WatermarkRowGeneric_t;
+
+#define NUM_WM_RANGES 4
+
+typedef enum {
+	WATERMARKS_CLOCK_RANGE = 0,
+	WATERMARKS_DUMMY_PSTATE,
+	WATERMARKS_MALL,
+	WATERMARKS_COUNT,
+} WATERMARKS_FLAGS_e;
+
+typedef struct {
+	// Watermarks
+	WatermarkRowGeneric_t WatermarkRow[NUM_WM_RANGES];
+} Watermarks_t;
+
+typedef struct {
+	Watermarks_t Watermarks;
+	uint32_t  Spare[16];
+
+	uint32_t     MmHubPadding[8]; // SMU internal use
+} WatermarksExternal_t;
+
+// Table types
+#define TABLE_PMFW_PPTABLE            0
+#define TABLE_COMBO_PPTABLE           1
+#define TABLE_WATERMARKS              2
+#define TABLE_AVFS_PSM_DEBUG          3
+#define TABLE_PMSTATUSLOG             4
+#define TABLE_SMU_METRICS             5
+#define TABLE_DRIVER_SMU_CONFIG       6
+#define TABLE_ACTIVITY_MONITOR_COEFF  7
+#define TABLE_OVERDRIVE               8
+#define TABLE_I2C_COMMANDS            9
+#define TABLE_DRIVER_INFO             10
+#define TABLE_COUNT                   11
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn351_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn351_clk_mgr.c
new file mode 100644
index 000000000000..4607eff07253
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn351_clk_mgr.c
@@ -0,0 +1,141 @@
+/*
+ * Copyright 2024 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#include "core_types.h"
+#include "dcn35_clk_mgr.h"
+
+#define DCN_BASE__INST0_SEG1 0x000000C0
+#define mmCLK1_CLK_PLL_REQ 0x16E37
+
+#define mmCLK1_CLK0_DFS_CNTL 0x16E69
+#define mmCLK1_CLK1_DFS_CNTL 0x16E6C
+#define mmCLK1_CLK2_DFS_CNTL 0x16E6F
+#define mmCLK1_CLK3_DFS_CNTL 0x16E72
+#define mmCLK1_CLK4_DFS_CNTL 0x16E75
+#define mmCLK1_CLK5_DFS_CNTL 0x16E78
+
+#define mmCLK1_CLK0_CURRENT_CNT 0x16EFC
+#define mmCLK1_CLK1_CURRENT_CNT 0x16EFD
+#define mmCLK1_CLK2_CURRENT_CNT 0x16EFE
+#define mmCLK1_CLK3_CURRENT_CNT 0x16EFF
+#define mmCLK1_CLK4_CURRENT_CNT 0x16F00
+#define mmCLK1_CLK5_CURRENT_CNT 0x16F01
+
+#define mmCLK1_CLK0_BYPASS_CNTL 0x16E8A
+#define mmCLK1_CLK1_BYPASS_CNTL 0x16E93
+#define mmCLK1_CLK2_BYPASS_CNTL 0x16E9C
+#define mmCLK1_CLK3_BYPASS_CNTL 0x16EA5
+#define mmCLK1_CLK4_BYPASS_CNTL 0x16EAE
+#define mmCLK1_CLK5_BYPASS_CNTL 0x16EB7
+
+#define mmCLK1_CLK0_DS_CNTL 0x16E83
+#define mmCLK1_CLK1_DS_CNTL 0x16E8C
+#define mmCLK1_CLK2_DS_CNTL 0x16E95
+#define mmCLK1_CLK3_DS_CNTL 0x16E9E
+#define mmCLK1_CLK4_DS_CNTL 0x16EA7
+#define mmCLK1_CLK5_DS_CNTL 0x16EB0
+
+#define mmCLK1_CLK0_ALLOW_DS 0x16E84
+#define mmCLK1_CLK1_ALLOW_DS 0x16E8D
+#define mmCLK1_CLK2_ALLOW_DS 0x16E96
+#define mmCLK1_CLK3_ALLOW_DS 0x16E9F
+#define mmCLK1_CLK4_ALLOW_DS 0x16EA8
+#define mmCLK1_CLK5_ALLOW_DS 0x16EB1
+
+#define mmCLK5_spll_field_8 0x1B04B
+#define mmCLK6_spll_field_8 0x1B24B
+#define mmDENTIST_DISPCLK_CNTL 0x0124
+#define regDENTIST_DISPCLK_CNTL 0x0064
+#define regDENTIST_DISPCLK_CNTL_BASE_IDX 1
+
+#define CLK1_CLK_PLL_REQ__FbMult_int__SHIFT 0x0
+#define CLK1_CLK_PLL_REQ__PllSpineDiv__SHIFT 0xc
+#define CLK1_CLK_PLL_REQ__FbMult_frac__SHIFT 0x10
+#define CLK1_CLK_PLL_REQ__FbMult_int_MASK 0x000001FFL
+#define CLK1_CLK_PLL_REQ__PllSpineDiv_MASK 0x0000F000L
+#define CLK1_CLK_PLL_REQ__FbMult_frac_MASK 0xFFFF0000L
+
+#define CLK1_CLK2_BYPASS_CNTL__CLK2_BYPASS_SEL_MASK 0x00000007L
+
+// DENTIST_DISPCLK_CNTL
+#define DENTIST_DISPCLK_CNTL__DENTIST_DISPCLK_WDIVIDER__SHIFT 0x0
+#define DENTIST_DISPCLK_CNTL__DENTIST_DISPCLK_RDIVIDER__SHIFT 0x8
+#define DENTIST_DISPCLK_CNTL__DENTIST_DISPCLK_CHG_DONE__SHIFT 0x13
+#define DENTIST_DISPCLK_CNTL__DENTIST_DPPCLK_CHG_DONE__SHIFT 0x14
+#define DENTIST_DISPCLK_CNTL__DENTIST_DPPCLK_WDIVIDER__SHIFT 0x18
+#define DENTIST_DISPCLK_CNTL__DENTIST_DISPCLK_WDIVIDER_MASK 0x0000007FL
+#define DENTIST_DISPCLK_CNTL__DENTIST_DISPCLK_RDIVIDER_MASK 0x00007F00L
+#define DENTIST_DISPCLK_CNTL__DENTIST_DISPCLK_CHG_DONE_MASK 0x00080000L
+#define DENTIST_DISPCLK_CNTL__DENTIST_DPPCLK_CHG_DONE_MASK 0x00100000L
+#define DENTIST_DISPCLK_CNTL__DENTIST_DPPCLK_WDIVIDER_MASK 0x7F000000L
+
+#define CLK5_spll_field_8__spll_ssc_en_MASK 0x00002000L
+
+#define REG(reg) \
+	(clk_mgr->regs->reg)
+
+#define BASE_INNER(seg) DCN_BASE__INST0_SEG ## seg
+
+#define BASE(seg) BASE_INNER(seg)
+
+#define SR(reg_name)\
+		.reg_name = BASE(reg ## reg_name ## _BASE_IDX) +  \
+					reg ## reg_name
+
+#define CLK_SR_DCN35(reg_name)\
+	.reg_name = mm ## reg_name
+
+static const struct clk_mgr_registers clk_mgr_regs_dcn351 = {
+	CLK_REG_LIST_DCN35()
+};
+
+static const struct clk_mgr_shift clk_mgr_shift_dcn351 = {
+	CLK_COMMON_MASK_SH_LIST_DCN32(__SHIFT)
+};
+
+static const struct clk_mgr_mask clk_mgr_mask_dcn351 = {
+	CLK_COMMON_MASK_SH_LIST_DCN32(_MASK)
+};
+
+#define TO_CLK_MGR_DCN35(clk_mgr)\
+	container_of(clk_mgr, struct clk_mgr_dcn35, base)
+
+
+void dcn351_clk_mgr_construct(
+		struct dc_context *ctx,
+		struct clk_mgr_dcn35 *clk_mgr,
+		struct pp_smu_funcs *pp_smu,
+		struct dccg *dccg)
+{
+	/*register offset changed*/
+	clk_mgr->base.regs = &clk_mgr_regs_dcn351;
+	clk_mgr->base.clk_mgr_shift = &clk_mgr_shift_dcn351;
+	clk_mgr->base.clk_mgr_mask = &clk_mgr_mask_dcn351;
+
+	dcn35_clk_mgr_construct(ctx,  clk_mgr, pp_smu, dccg);
+
+}
+
+
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c
new file mode 100644
index 000000000000..dfd0c9505af0
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c
@@ -0,0 +1,1588 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+
+#include "dcn35_clk_mgr.h"
+
+#include "dccg.h"
+#include "clk_mgr_internal.h"
+
+// For dce12_get_dp_ref_freq_khz
+#include "dce100/dce_clk_mgr.h"
+
+// For dcn20_update_clocks_update_dpp_dto
+#include "dcn20/dcn20_clk_mgr.h"
+
+
+#include "reg_helper.h"
+#include "core_types.h"
+#include "dcn35_smu.h"
+#include "dm_helpers.h"
+
+#include "dcn31/dcn31_clk_mgr.h"
+
+#include "dc_dmub_srv.h"
+#include "link_service.h"
+#include "logger_types.h"
+
+#undef DC_LOGGER
+#define DC_LOGGER \
+	clk_mgr->base.base.ctx->logger
+
+#define DCN_BASE__INST0_SEG1 0x000000C0
+#define mmCLK1_CLK_PLL_REQ 0x16E37
+
+#define mmCLK1_CLK0_DFS_CNTL 0x16E69
+#define mmCLK1_CLK1_DFS_CNTL 0x16E6C
+#define mmCLK1_CLK2_DFS_CNTL 0x16E6F
+#define mmCLK1_CLK3_DFS_CNTL 0x16E72
+#define mmCLK1_CLK4_DFS_CNTL 0x16E75
+#define mmCLK1_CLK5_DFS_CNTL 0x16E78
+
+#define mmCLK1_CLK0_CURRENT_CNT 0x16EFB
+#define mmCLK1_CLK1_CURRENT_CNT 0x16EFC
+#define mmCLK1_CLK2_CURRENT_CNT 0x16EFD
+#define mmCLK1_CLK3_CURRENT_CNT 0x16EFE
+#define mmCLK1_CLK4_CURRENT_CNT 0x16EFF
+#define mmCLK1_CLK5_CURRENT_CNT 0x16F00
+
+#define mmCLK1_CLK0_BYPASS_CNTL 0x16E8A
+#define mmCLK1_CLK1_BYPASS_CNTL 0x16E93
+#define mmCLK1_CLK2_BYPASS_CNTL 0x16E9C
+#define mmCLK1_CLK3_BYPASS_CNTL 0x16EA5
+#define mmCLK1_CLK4_BYPASS_CNTL 0x16EAE
+#define mmCLK1_CLK5_BYPASS_CNTL 0x16EB7
+
+#define mmCLK1_CLK0_DS_CNTL 0x16E83
+#define mmCLK1_CLK1_DS_CNTL 0x16E8C
+#define mmCLK1_CLK2_DS_CNTL 0x16E95
+#define mmCLK1_CLK3_DS_CNTL 0x16E9E
+#define mmCLK1_CLK4_DS_CNTL 0x16EA7
+#define mmCLK1_CLK5_DS_CNTL 0x16EB0
+
+#define mmCLK1_CLK0_ALLOW_DS 0x16E84
+#define mmCLK1_CLK1_ALLOW_DS 0x16E8D
+#define mmCLK1_CLK2_ALLOW_DS 0x16E96
+#define mmCLK1_CLK3_ALLOW_DS 0x16E9F
+#define mmCLK1_CLK4_ALLOW_DS 0x16EA8
+#define mmCLK1_CLK5_ALLOW_DS 0x16EB1
+
+#define mmCLK5_spll_field_8 0x1B24B
+#define mmCLK6_spll_field_8 0x1B24B
+#define mmDENTIST_DISPCLK_CNTL 0x0124
+#define regDENTIST_DISPCLK_CNTL 0x0064
+#define regDENTIST_DISPCLK_CNTL_BASE_IDX 1
+
+#define CLK1_CLK_PLL_REQ__FbMult_int__SHIFT 0x0
+#define CLK1_CLK_PLL_REQ__PllSpineDiv__SHIFT 0xc
+#define CLK1_CLK_PLL_REQ__FbMult_frac__SHIFT 0x10
+#define CLK1_CLK_PLL_REQ__FbMult_int_MASK 0x000001FFL
+#define CLK1_CLK_PLL_REQ__PllSpineDiv_MASK 0x0000F000L
+#define CLK1_CLK_PLL_REQ__FbMult_frac_MASK 0xFFFF0000L
+
+#define CLK1_CLK2_BYPASS_CNTL__CLK2_BYPASS_SEL_MASK 0x00000007L
+#define CLK1_CLK2_BYPASS_CNTL__CLK2_BYPASS_DIV_MASK 0x000F0000L
+// DENTIST_DISPCLK_CNTL
+#define DENTIST_DISPCLK_CNTL__DENTIST_DISPCLK_WDIVIDER__SHIFT 0x0
+#define DENTIST_DISPCLK_CNTL__DENTIST_DISPCLK_RDIVIDER__SHIFT 0x8
+#define DENTIST_DISPCLK_CNTL__DENTIST_DISPCLK_CHG_DONE__SHIFT 0x13
+#define DENTIST_DISPCLK_CNTL__DENTIST_DPPCLK_CHG_DONE__SHIFT 0x14
+#define DENTIST_DISPCLK_CNTL__DENTIST_DPPCLK_WDIVIDER__SHIFT 0x18
+#define DENTIST_DISPCLK_CNTL__DENTIST_DISPCLK_WDIVIDER_MASK 0x0000007FL
+#define DENTIST_DISPCLK_CNTL__DENTIST_DISPCLK_RDIVIDER_MASK 0x00007F00L
+#define DENTIST_DISPCLK_CNTL__DENTIST_DISPCLK_CHG_DONE_MASK 0x00080000L
+#define DENTIST_DISPCLK_CNTL__DENTIST_DPPCLK_CHG_DONE_MASK 0x00100000L
+#define DENTIST_DISPCLK_CNTL__DENTIST_DPPCLK_WDIVIDER_MASK 0x7F000000L
+
+#define CLK5_spll_field_8__spll_ssc_en_MASK 0x00002000L
+#define CLK6_spll_field_8__spll_ssc_en_MASK 0x00002000L
+
+#define SMU_VER_THRESHOLD 0x5D4A00 //93.74.0
+#undef FN
+#define FN(reg_name, field_name) \
+	clk_mgr->clk_mgr_shift->field_name, clk_mgr->clk_mgr_mask->field_name
+
+#define REG(reg) \
+	(clk_mgr->regs->reg)
+
+#define BASE_INNER(seg) DCN_BASE__INST0_SEG ## seg
+
+#define BASE(seg) BASE_INNER(seg)
+
+#define SR(reg_name)\
+		.reg_name = BASE(reg ## reg_name ## _BASE_IDX) +  \
+					reg ## reg_name
+
+#define CLK_SR_DCN35(reg_name)\
+	.reg_name = mm ## reg_name
+
+static const struct clk_mgr_registers clk_mgr_regs_dcn35 = {
+	CLK_REG_LIST_DCN35()
+};
+
+static const struct clk_mgr_shift clk_mgr_shift_dcn35 = {
+	CLK_COMMON_MASK_SH_LIST_DCN32(__SHIFT)
+};
+
+static const struct clk_mgr_mask clk_mgr_mask_dcn35 = {
+	CLK_COMMON_MASK_SH_LIST_DCN32(_MASK)
+};
+
+#define TO_CLK_MGR_DCN35(clk_mgr)\
+	container_of(clk_mgr, struct clk_mgr_dcn35, base)
+
+static int dcn35_get_active_display_cnt_wa(
+		struct dc *dc,
+		struct dc_state *context,
+		int *all_active_disps)
+{
+	int i, display_count = 0;
+	bool tmds_present = false;
+
+	for (i = 0; i < context->stream_count; i++) {
+		const struct dc_stream_state *stream = context->streams[i];
+
+		if (stream->signal == SIGNAL_TYPE_HDMI_TYPE_A ||
+				stream->signal == SIGNAL_TYPE_DVI_SINGLE_LINK ||
+				stream->signal == SIGNAL_TYPE_DVI_DUAL_LINK)
+			tmds_present = true;
+	}
+
+	for (i = 0; i < dc->link_count; i++) {
+		const struct dc_link *link = dc->links[i];
+
+		/* abusing the fact that the dig and phy are coupled to see if the phy is enabled */
+		if (link->link_enc && link->link_enc->funcs->is_dig_enabled &&
+				link->link_enc->funcs->is_dig_enabled(link->link_enc))
+			display_count++;
+	}
+	if (all_active_disps != NULL)
+		*all_active_disps = display_count;
+	/* WA for hang on HDMI after display off back on*/
+	if (display_count == 0 && tmds_present)
+		display_count = 1;
+
+	return display_count;
+}
+static void dcn35_disable_otg_wa(struct clk_mgr *clk_mgr_base, struct dc_state *context,
+		bool safe_to_lower, bool disable)
+{
+	struct dc *dc = clk_mgr_base->ctx->dc;
+	int i;
+
+	if (dc->ctx->dce_environment == DCE_ENV_DIAG)
+		return;
+
+	for (i = 0; i < dc->res_pool->pipe_count; ++i) {
+		struct pipe_ctx *old_pipe = &dc->current_state->res_ctx.pipe_ctx[i];
+		struct pipe_ctx *new_pipe = &context->res_ctx.pipe_ctx[i];
+		struct clk_mgr_internal *clk_mgr_internal = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+		struct dccg *dccg = clk_mgr_internal->dccg;
+		struct pipe_ctx *pipe = safe_to_lower
+			? &context->res_ctx.pipe_ctx[i]
+			: &dc->current_state->res_ctx.pipe_ctx[i];
+		struct link_encoder *new_pipe_link_enc = new_pipe->link_res.dio_link_enc;
+		struct link_encoder *pipe_link_enc = pipe->link_res.dio_link_enc;
+		bool stream_changed_otg_dig_on = false;
+		bool has_active_hpo = false;
+
+		if (pipe->top_pipe || pipe->prev_odm_pipe)
+			continue;
+
+		if (!dc->config.unify_link_enc_assignment) {
+			if (new_pipe->stream)
+				new_pipe_link_enc = new_pipe->stream->link_enc;
+			if (pipe->stream)
+				pipe_link_enc = pipe->stream->link_enc;
+		}
+
+		stream_changed_otg_dig_on = old_pipe->stream && new_pipe->stream &&
+		old_pipe->stream != new_pipe->stream &&
+		old_pipe->stream_res.tg == new_pipe->stream_res.tg &&
+		new_pipe_link_enc && !new_pipe->stream->dpms_off &&
+		new_pipe_link_enc->funcs->is_dig_enabled &&
+		new_pipe_link_enc->funcs->is_dig_enabled(
+		new_pipe_link_enc) &&
+		new_pipe->stream_res.stream_enc &&
+		new_pipe->stream_res.stream_enc->funcs->is_fifo_enabled &&
+		new_pipe->stream_res.stream_enc->funcs->is_fifo_enabled(new_pipe->stream_res.stream_enc);
+
+		if (old_pipe->stream && new_pipe->stream && old_pipe->stream == new_pipe->stream) {
+			has_active_hpo =  dccg->ctx->dc->link_srv->dp_is_128b_132b_signal(old_pipe) &&
+			dccg->ctx->dc->link_srv->dp_is_128b_132b_signal(new_pipe);
+
+		}
+
+		if (!has_active_hpo && !stream_changed_otg_dig_on && pipe->stream &&
+		    (pipe->stream->dpms_off || dc_is_virtual_signal(pipe->stream->signal) || !pipe_link_enc) &&
+		    !dccg->ctx->dc->link_srv->dp_is_128b_132b_signal(pipe)) {
+			/* This w/a should not trigger when we have a dig active */
+			if (disable) {
+				if (pipe->stream_res.tg && pipe->stream_res.tg->funcs->immediate_disable_crtc)
+					pipe->stream_res.tg->funcs->immediate_disable_crtc(pipe->stream_res.tg);
+
+				reset_sync_context_for_pipe(dc, context, i);
+			} else {
+				pipe->stream_res.tg->funcs->enable_crtc(pipe->stream_res.tg);
+			}
+		}
+	}
+}
+
+static void dcn35_update_clocks_update_dtb_dto(struct clk_mgr_internal *clk_mgr,
+			struct dc_state *context,
+			int ref_dtbclk_khz)
+{
+	struct dccg *dccg = clk_mgr->dccg;
+	uint32_t tg_mask = 0;
+	int i;
+
+	for (i = 0; i < clk_mgr->base.ctx->dc->res_pool->pipe_count; i++) {
+		struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i];
+		struct dtbclk_dto_params dto_params = {0};
+
+		/* use mask to program DTO once per tg */
+		if (pipe_ctx->stream_res.tg &&
+				!(tg_mask & (1 << pipe_ctx->stream_res.tg->inst))) {
+			tg_mask |= (1 << pipe_ctx->stream_res.tg->inst);
+
+			dto_params.otg_inst = pipe_ctx->stream_res.tg->inst;
+			dto_params.ref_dtbclk_khz = ref_dtbclk_khz;
+
+			dccg->funcs->set_dtbclk_dto(clk_mgr->dccg, &dto_params);
+			//dccg->funcs->set_audio_dtbclk_dto(clk_mgr->dccg, &dto_params);
+		}
+	}
+}
+
+static void dcn35_update_clocks_update_dpp_dto(struct clk_mgr_internal *clk_mgr,
+		struct dc_state *context, bool safe_to_lower)
+{
+	int i;
+	bool dppclk_active[MAX_PIPES] = {0};
+
+
+	clk_mgr->dccg->ref_dppclk = clk_mgr->base.clks.dppclk_khz;
+	for (i = 0; i < clk_mgr->base.ctx->dc->res_pool->pipe_count; i++) {
+		int dpp_inst = 0, dppclk_khz, prev_dppclk_khz;
+
+		dppclk_khz = context->res_ctx.pipe_ctx[i].plane_res.bw.dppclk_khz;
+
+		if (context->res_ctx.pipe_ctx[i].plane_res.dpp)
+			dpp_inst = context->res_ctx.pipe_ctx[i].plane_res.dpp->inst;
+		else if (!context->res_ctx.pipe_ctx[i].plane_res.dpp && dppclk_khz == 0) {
+			/* dpp == NULL && dppclk_khz == 0 is valid because of pipe harvesting.
+			 * In this case just continue in loop
+			 */
+			continue;
+		} else if (!context->res_ctx.pipe_ctx[i].plane_res.dpp && dppclk_khz > 0) {
+			/* The software state is not valid if dpp resource is NULL and
+			 * dppclk_khz > 0.
+			 */
+			ASSERT(false);
+			continue;
+		}
+
+		prev_dppclk_khz = clk_mgr->dccg->pipe_dppclk_khz[i];
+
+		if (safe_to_lower || prev_dppclk_khz < dppclk_khz)
+			clk_mgr->dccg->funcs->update_dpp_dto(
+							clk_mgr->dccg, dpp_inst, dppclk_khz);
+		dppclk_active[dpp_inst] = true;
+	}
+	if (safe_to_lower)
+		for (i = 0; i < clk_mgr->base.ctx->dc->res_pool->pipe_count; i++) {
+			struct dpp *old_dpp = clk_mgr->base.ctx->dc->current_state->res_ctx.pipe_ctx[i].plane_res.dpp;
+
+			if (old_dpp && !dppclk_active[old_dpp->inst])
+				clk_mgr->dccg->funcs->update_dpp_dto(clk_mgr->dccg, old_dpp->inst, 0);
+		}
+}
+
+static uint8_t get_lowest_dpia_index(const struct dc_link *link)
+{
+	const struct dc *dc_struct = link->dc;
+	uint8_t idx = 0xFF;
+	int i;
+
+	for (i = 0; i < MAX_PIPES * 2; ++i) {
+		if (!dc_struct->links[i] || dc_struct->links[i]->ep_type != DISPLAY_ENDPOINT_USB4_DPIA)
+			continue;
+
+		if (idx > dc_struct->links[i]->link_index)
+			idx = dc_struct->links[i]->link_index;
+	}
+
+	return idx;
+}
+
+static void dcn35_notify_host_router_bw(struct clk_mgr *clk_mgr_base, struct dc_state *context,
+					bool safe_to_lower)
+{
+	struct dc_clocks *new_clocks = &context->bw_ctx.bw.dcn.clk;
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+	uint32_t host_router_bw_kbps[MAX_HOST_ROUTERS_NUM] = { 0 };
+	int i;
+	for (i = 0; i < context->stream_count; ++i) {
+		const struct dc_stream_state *stream = context->streams[i];
+		const struct dc_link *link = stream->link;
+		uint8_t lowest_dpia_index = 0;
+		unsigned int hr_index = 0;
+
+		if (!link)
+			continue;
+
+		lowest_dpia_index = get_lowest_dpia_index(link);
+		if (link->link_index < lowest_dpia_index)
+			continue;
+
+		hr_index = (link->link_index - lowest_dpia_index) / 2;
+		if (hr_index >= MAX_HOST_ROUTERS_NUM)
+			continue;
+		host_router_bw_kbps[hr_index] += dc_bandwidth_in_kbps_from_timing(
+			&stream->timing, dc_link_get_highest_encoding_format(link));
+	}
+
+	for (i = 0; i < MAX_HOST_ROUTERS_NUM; ++i) {
+		new_clocks->host_router_bw_kbps[i] = host_router_bw_kbps[i];
+		if (should_set_clock(safe_to_lower, new_clocks->host_router_bw_kbps[i], clk_mgr_base->clks.host_router_bw_kbps[i])) {
+			clk_mgr_base->clks.host_router_bw_kbps[i] = new_clocks->host_router_bw_kbps[i];
+			dcn35_smu_notify_host_router_bw(clk_mgr, i, new_clocks->host_router_bw_kbps[i]);
+		}
+	}
+}
+
+void dcn35_update_clocks(struct clk_mgr *clk_mgr_base,
+			struct dc_state *context,
+			bool safe_to_lower)
+{
+	union dmub_rb_cmd cmd;
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+	struct dc_clocks *new_clocks = &context->bw_ctx.bw.dcn.clk;
+	struct dc *dc = clk_mgr_base->ctx->dc;
+	int display_count = 0;
+	bool update_dppclk = false;
+	bool update_dispclk = false;
+	bool dpp_clock_lowered = false;
+	int all_active_disps = 0;
+
+	if (dc->work_arounds.skip_clock_update)
+		return;
+
+	display_count = dcn35_get_active_display_cnt_wa(dc, context, &all_active_disps);
+	if (new_clocks->dtbclk_en && !new_clocks->ref_dtbclk_khz)
+		new_clocks->ref_dtbclk_khz = 600000;
+	else if (!new_clocks->dtbclk_en && new_clocks->ref_dtbclk_khz > 590000)
+		new_clocks->ref_dtbclk_khz = 0;
+
+	/*
+	 * if it is safe to lower, but we are already in the lower state, we don't have to do anything
+	 * also if safe to lower is false, we just go in the higher state
+	 */
+	if (safe_to_lower) {
+		if (new_clocks->zstate_support != DCN_ZSTATE_SUPPORT_DISALLOW &&
+				new_clocks->zstate_support != clk_mgr_base->clks.zstate_support) {
+			dcn35_smu_set_zstate_support(clk_mgr, new_clocks->zstate_support);
+			dm_helpers_enable_periodic_detection(clk_mgr_base->ctx, true);
+			clk_mgr_base->clks.zstate_support = new_clocks->zstate_support;
+		}
+
+		if (clk_mgr_base->clks.dtbclk_en && !new_clocks->dtbclk_en) {
+			if (clk_mgr->base.ctx->dc->config.allow_0_dtb_clk)
+				dcn35_smu_set_dtbclk(clk_mgr, false);
+
+			clk_mgr_base->clks.dtbclk_en = new_clocks->dtbclk_en;
+		}
+		/* check that we're not already in lower */
+		if (clk_mgr_base->clks.pwr_state != DCN_PWR_STATE_LOW_POWER) {
+			/* if we can go lower, go lower */
+			if (display_count == 0)
+				clk_mgr_base->clks.pwr_state = DCN_PWR_STATE_LOW_POWER;
+		}
+	} else {
+		if (new_clocks->zstate_support == DCN_ZSTATE_SUPPORT_DISALLOW &&
+				new_clocks->zstate_support != clk_mgr_base->clks.zstate_support) {
+			dcn35_smu_set_zstate_support(clk_mgr, DCN_ZSTATE_SUPPORT_DISALLOW);
+			dm_helpers_enable_periodic_detection(clk_mgr_base->ctx, false);
+			clk_mgr_base->clks.zstate_support = new_clocks->zstate_support;
+		}
+
+		if (!clk_mgr_base->clks.dtbclk_en && new_clocks->dtbclk_en) {
+			int actual_dtbclk = 0;
+
+			dcn35_update_clocks_update_dtb_dto(clk_mgr, context, new_clocks->ref_dtbclk_khz);
+			dcn35_smu_set_dtbclk(clk_mgr, true);
+
+			actual_dtbclk = REG_READ(CLK1_CLK4_CURRENT_CNT);
+
+			if (actual_dtbclk > 590000) {
+				clk_mgr_base->clks.ref_dtbclk_khz = new_clocks->ref_dtbclk_khz;
+				clk_mgr_base->clks.dtbclk_en = new_clocks->dtbclk_en;
+			}
+		}
+
+		/* check that we're not already in D0 */
+		if (clk_mgr_base->clks.pwr_state != DCN_PWR_STATE_MISSION_MODE) {
+			union display_idle_optimization_u idle_info = { 0 };
+
+			dcn35_smu_set_display_idle_optimization(clk_mgr, idle_info.data);
+			/* update power state */
+			clk_mgr_base->clks.pwr_state = DCN_PWR_STATE_MISSION_MODE;
+		}
+	}
+	if (dc->debug.force_min_dcfclk_mhz > 0)
+		new_clocks->dcfclk_khz = (new_clocks->dcfclk_khz > (dc->debug.force_min_dcfclk_mhz * 1000)) ?
+				new_clocks->dcfclk_khz : (dc->debug.force_min_dcfclk_mhz * 1000);
+
+	if (should_set_clock(safe_to_lower, new_clocks->dcfclk_khz, clk_mgr_base->clks.dcfclk_khz)) {
+		clk_mgr_base->clks.dcfclk_khz = new_clocks->dcfclk_khz;
+		dcn35_smu_set_hard_min_dcfclk(clk_mgr, clk_mgr_base->clks.dcfclk_khz);
+	}
+
+	if (should_set_clock(safe_to_lower,
+			new_clocks->dcfclk_deep_sleep_khz, clk_mgr_base->clks.dcfclk_deep_sleep_khz)) {
+		clk_mgr_base->clks.dcfclk_deep_sleep_khz = new_clocks->dcfclk_deep_sleep_khz;
+		dcn35_smu_set_min_deep_sleep_dcfclk(clk_mgr, clk_mgr_base->clks.dcfclk_deep_sleep_khz);
+	}
+
+	// workaround: Limit dppclk to 100Mhz to avoid lower eDP panel switch to plus 4K monitor underflow.
+	if (new_clocks->dppclk_khz < 100000)
+		new_clocks->dppclk_khz = 100000;
+
+	if (should_set_clock(safe_to_lower, new_clocks->dppclk_khz, clk_mgr->base.clks.dppclk_khz)) {
+		if (clk_mgr->base.clks.dppclk_khz > new_clocks->dppclk_khz)
+			dpp_clock_lowered = true;
+		clk_mgr_base->clks.dppclk_khz = new_clocks->dppclk_khz;
+		update_dppclk = true;
+	}
+
+	if (should_set_clock(safe_to_lower, new_clocks->dispclk_khz, clk_mgr_base->clks.dispclk_khz) &&
+	    (new_clocks->dispclk_khz > 0 || (safe_to_lower && display_count == 0))) {
+		int requested_dispclk_khz = new_clocks->dispclk_khz;
+
+		dcn35_disable_otg_wa(clk_mgr_base, context, safe_to_lower, true);
+
+		/* Clamp the requested clock to PMFW based on their limit. */
+		if (dc->debug.min_disp_clk_khz > 0 && requested_dispclk_khz < dc->debug.min_disp_clk_khz)
+			requested_dispclk_khz = dc->debug.min_disp_clk_khz;
+
+		dcn35_smu_set_dispclk(clk_mgr, requested_dispclk_khz);
+		clk_mgr_base->clks.dispclk_khz = new_clocks->dispclk_khz;
+
+		dcn35_disable_otg_wa(clk_mgr_base, context, safe_to_lower, false);
+
+		update_dispclk = true;
+	}
+
+	/* clock limits are received with MHz precision, divide by 1000 to prevent setting clocks at every call */
+	if (!dc->debug.disable_dtb_ref_clk_switch &&
+	    should_set_clock(safe_to_lower, new_clocks->ref_dtbclk_khz / 1000,
+			     clk_mgr_base->clks.ref_dtbclk_khz / 1000)) {
+		dcn35_update_clocks_update_dtb_dto(clk_mgr, context, new_clocks->ref_dtbclk_khz);
+		clk_mgr_base->clks.ref_dtbclk_khz = new_clocks->ref_dtbclk_khz;
+	}
+
+	if (dpp_clock_lowered) {
+		// increase per DPP DTO before lowering global dppclk
+		dcn35_update_clocks_update_dpp_dto(clk_mgr, context, safe_to_lower);
+		dcn35_smu_set_dppclk(clk_mgr, clk_mgr_base->clks.dppclk_khz);
+	} else {
+		// increase global DPPCLK before lowering per DPP DTO
+		if (update_dppclk || update_dispclk)
+			dcn35_smu_set_dppclk(clk_mgr, clk_mgr_base->clks.dppclk_khz);
+		dcn35_update_clocks_update_dpp_dto(clk_mgr, context, safe_to_lower);
+	}
+
+	// notify PMFW of bandwidth per DPIA tunnel
+	if (dc->debug.notify_dpia_hr_bw)
+		dcn35_notify_host_router_bw(clk_mgr_base, context, safe_to_lower);
+
+	// notify DMCUB of latest clocks
+	memset(&cmd, 0, sizeof(cmd));
+	cmd.notify_clocks.header.type = DMUB_CMD__CLK_MGR;
+	cmd.notify_clocks.header.sub_type = DMUB_CMD__CLK_MGR_NOTIFY_CLOCKS;
+	cmd.notify_clocks.clocks.dcfclk_khz = clk_mgr_base->clks.dcfclk_khz;
+	cmd.notify_clocks.clocks.dcfclk_deep_sleep_khz =
+		clk_mgr_base->clks.dcfclk_deep_sleep_khz;
+	cmd.notify_clocks.clocks.dispclk_khz = clk_mgr_base->clks.dispclk_khz;
+	cmd.notify_clocks.clocks.dppclk_khz = clk_mgr_base->clks.dppclk_khz;
+
+	dc_wake_and_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
+}
+
+static int get_vco_frequency_from_reg(struct clk_mgr_internal *clk_mgr)
+{
+	/* get FbMult value */
+	struct fixed31_32 pll_req;
+	unsigned int fbmult_frac_val = 0;
+	unsigned int fbmult_int_val = 0;
+
+	/*
+	 * Register value of fbmult is in 8.16 format, we are converting to 314.32
+	 * to leverage the fix point operations available in driver
+	 */
+
+	REG_GET(CLK1_CLK_PLL_REQ, FbMult_frac, &fbmult_frac_val); /* 16 bit fractional part*/
+	REG_GET(CLK1_CLK_PLL_REQ, FbMult_int, &fbmult_int_val); /* 8 bit integer part */
+
+	pll_req = dc_fixpt_from_int(fbmult_int_val);
+
+	/*
+	 * since fractional part is only 16 bit in register definition but is 32 bit
+	 * in our fix point definiton, need to shift left by 16 to obtain correct value
+	 */
+	pll_req.value |= fbmult_frac_val << 16;
+
+	/* multiply by REFCLK period */
+	pll_req = dc_fixpt_mul_int(pll_req, clk_mgr->dfs_ref_freq_khz);
+
+	/* integer part is now VCO frequency in kHz */
+	return dc_fixpt_floor(pll_req);
+}
+
+static void dcn35_enable_pme_wa(struct clk_mgr *clk_mgr_base)
+{
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+
+	dcn35_smu_enable_pme_wa(clk_mgr);
+}
+
+
+bool dcn35_are_clock_states_equal(struct dc_clocks *a,
+		struct dc_clocks *b)
+{
+	if (a->dispclk_khz != b->dispclk_khz)
+		return false;
+	else if (a->dppclk_khz != b->dppclk_khz)
+		return false;
+	else if (a->dcfclk_khz != b->dcfclk_khz)
+		return false;
+	else if (a->dcfclk_deep_sleep_khz != b->dcfclk_deep_sleep_khz)
+		return false;
+	else if (a->zstate_support != b->zstate_support)
+		return false;
+	else if (a->dtbclk_en != b->dtbclk_en)
+		return false;
+
+	return true;
+}
+
+static void dcn35_save_clk_registers_internal(struct dcn35_clk_internal *internal, struct clk_mgr *clk_mgr_base)
+{
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+
+	// read dtbclk
+	internal->CLK1_CLK4_CURRENT_CNT = REG_READ(CLK1_CLK4_CURRENT_CNT);
+	internal->CLK1_CLK4_BYPASS_CNTL = REG_READ(CLK1_CLK4_BYPASS_CNTL);
+
+	// read dcfclk
+	internal->CLK1_CLK3_CURRENT_CNT = REG_READ(CLK1_CLK3_CURRENT_CNT);
+	internal->CLK1_CLK3_BYPASS_CNTL = REG_READ(CLK1_CLK3_BYPASS_CNTL);
+
+	// read dcf deep sleep divider
+	internal->CLK1_CLK3_DS_CNTL = REG_READ(CLK1_CLK3_DS_CNTL);
+	internal->CLK1_CLK3_ALLOW_DS = REG_READ(CLK1_CLK3_ALLOW_DS);
+
+	// read dppclk
+	internal->CLK1_CLK1_CURRENT_CNT = REG_READ(CLK1_CLK1_CURRENT_CNT);
+	internal->CLK1_CLK1_BYPASS_CNTL = REG_READ(CLK1_CLK1_BYPASS_CNTL);
+
+	// read dprefclk
+	internal->CLK1_CLK2_CURRENT_CNT = REG_READ(CLK1_CLK2_CURRENT_CNT);
+	internal->CLK1_CLK2_BYPASS_CNTL = REG_READ(CLK1_CLK2_BYPASS_CNTL);
+
+	// read dispclk
+	internal->CLK1_CLK0_CURRENT_CNT = REG_READ(CLK1_CLK0_CURRENT_CNT);
+	internal->CLK1_CLK0_BYPASS_CNTL = REG_READ(CLK1_CLK0_BYPASS_CNTL);
+}
+
+static void dcn35_save_clk_registers(struct clk_state_registers_and_bypass *regs_and_bypass,
+		struct clk_mgr_dcn35 *clk_mgr)
+{
+	struct dcn35_clk_internal internal = {0};
+	char *bypass_clks[5] = {"0x0 DFS", "0x1 REFCLK", "0x2 ERROR", "0x3 400 FCH", "0x4 600 FCH"};
+
+	dcn35_save_clk_registers_internal(&internal, &clk_mgr->base.base);
+
+	regs_and_bypass->dcfclk = internal.CLK1_CLK3_CURRENT_CNT / 10;
+	regs_and_bypass->dcf_deep_sleep_divider = internal.CLK1_CLK3_DS_CNTL / 10;
+	regs_and_bypass->dcf_deep_sleep_allow = internal.CLK1_CLK3_ALLOW_DS;
+	regs_and_bypass->dprefclk = internal.CLK1_CLK2_CURRENT_CNT / 10;
+	regs_and_bypass->dispclk = internal.CLK1_CLK0_CURRENT_CNT / 10;
+	regs_and_bypass->dppclk = internal.CLK1_CLK1_CURRENT_CNT / 10;
+	regs_and_bypass->dtbclk = internal.CLK1_CLK4_CURRENT_CNT / 10;
+
+	regs_and_bypass->dppclk_bypass = internal.CLK1_CLK1_BYPASS_CNTL & 0x0007;
+	if (regs_and_bypass->dppclk_bypass > 4)
+		regs_and_bypass->dppclk_bypass = 0;
+	regs_and_bypass->dcfclk_bypass = internal.CLK1_CLK3_BYPASS_CNTL & 0x0007;
+	if (regs_and_bypass->dcfclk_bypass > 4)
+		regs_and_bypass->dcfclk_bypass = 0;
+	regs_and_bypass->dispclk_bypass = internal.CLK1_CLK0_BYPASS_CNTL & 0x0007;
+	if (regs_and_bypass->dispclk_bypass > 4)
+		regs_and_bypass->dispclk_bypass = 0;
+	regs_and_bypass->dprefclk_bypass = internal.CLK1_CLK2_BYPASS_CNTL & 0x0007;
+	if (regs_and_bypass->dprefclk_bypass > 4)
+		regs_and_bypass->dprefclk_bypass = 0;
+
+	if (clk_mgr->base.base.ctx->dc->debug.pstate_enabled) {
+		DC_LOG_SMU("clk_type,clk_value,deepsleep_cntl,deepsleep_allow,bypass\n");
+
+		DC_LOG_SMU("dcfclk,%d,%d,%d,%s\n",
+				   regs_and_bypass->dcfclk,
+				   regs_and_bypass->dcf_deep_sleep_divider,
+				   regs_and_bypass->dcf_deep_sleep_allow,
+				   bypass_clks[(int) regs_and_bypass->dcfclk_bypass]);
+
+		DC_LOG_SMU("dprefclk,%d,N/A,N/A,%s\n",
+			regs_and_bypass->dprefclk,
+			bypass_clks[(int) regs_and_bypass->dprefclk_bypass]);
+
+		DC_LOG_SMU("dispclk,%d,N/A,N/A,%s\n",
+			regs_and_bypass->dispclk,
+			bypass_clks[(int) regs_and_bypass->dispclk_bypass]);
+
+		// REGISTER VALUES
+		DC_LOG_SMU("reg_name,value,clk_type");
+
+		DC_LOG_SMU("CLK1_CLK3_CURRENT_CNT,%d,dcfclk",
+				internal.CLK1_CLK3_CURRENT_CNT);
+
+		DC_LOG_SMU("CLK1_CLK4_CURRENT_CNT,%d,dtbclk",
+					internal.CLK1_CLK4_CURRENT_CNT);
+
+		DC_LOG_SMU("CLK1_CLK3_DS_CNTL,%d,dcf_deep_sleep_divider",
+					internal.CLK1_CLK3_DS_CNTL);
+
+		DC_LOG_SMU("CLK1_CLK3_ALLOW_DS,%d,dcf_deep_sleep_allow",
+					internal.CLK1_CLK3_ALLOW_DS);
+
+		DC_LOG_SMU("CLK1_CLK2_CURRENT_CNT,%d,dprefclk",
+					internal.CLK1_CLK2_CURRENT_CNT);
+
+		DC_LOG_SMU("CLK1_CLK0_CURRENT_CNT,%d,dispclk",
+					internal.CLK1_CLK0_CURRENT_CNT);
+
+		DC_LOG_SMU("CLK1_CLK1_CURRENT_CNT,%d,dppclk",
+					internal.CLK1_CLK1_CURRENT_CNT);
+
+		DC_LOG_SMU("CLK1_CLK3_BYPASS_CNTL,%d,dcfclk_bypass",
+					internal.CLK1_CLK3_BYPASS_CNTL);
+
+		DC_LOG_SMU("CLK1_CLK2_BYPASS_CNTL,%d,dprefclk_bypass",
+					internal.CLK1_CLK2_BYPASS_CNTL);
+
+		DC_LOG_SMU("CLK1_CLK0_BYPASS_CNTL,%d,dispclk_bypass",
+					internal.CLK1_CLK0_BYPASS_CNTL);
+
+		DC_LOG_SMU("CLK1_CLK1_BYPASS_CNTL,%d,dppclk_bypass",
+					internal.CLK1_CLK1_BYPASS_CNTL);
+
+	}
+}
+
+static bool dcn35_is_spll_ssc_enabled(struct clk_mgr *clk_mgr_base)
+{
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+
+	uint32_t ssc_enable;
+
+	if (clk_mgr_base->ctx->dce_version == DCN_VERSION_3_51) {
+		ssc_enable = REG_READ(CLK6_spll_field_8) & CLK6_spll_field_8__spll_ssc_en_MASK;
+	} else {
+		ssc_enable = REG_READ(CLK5_spll_field_8) & CLK5_spll_field_8__spll_ssc_en_MASK;
+	}
+
+	return ssc_enable != 0;
+}
+
+static void init_clk_states(struct clk_mgr *clk_mgr)
+{
+	uint32_t ref_dtbclk = clk_mgr->clks.ref_dtbclk_khz;
+
+	memset(&(clk_mgr->clks), 0, sizeof(struct dc_clocks));
+
+	clk_mgr->clks.ref_dtbclk_khz = ref_dtbclk;	// restore ref_dtbclk
+	clk_mgr->clks.p_state_change_support = true;
+	clk_mgr->clks.prev_p_state_change_support = true;
+	clk_mgr->clks.pwr_state = DCN_PWR_STATE_UNKNOWN;
+	clk_mgr->clks.zstate_support = DCN_ZSTATE_SUPPORT_UNKNOWN;
+}
+
+void dcn35_init_clocks(struct clk_mgr *clk_mgr)
+{
+	struct clk_mgr_internal *clk_mgr_int = TO_CLK_MGR_INTERNAL(clk_mgr);
+	struct clk_mgr_dcn35 *clk_mgr_dcn35 = TO_CLK_MGR_DCN35(clk_mgr_int);
+
+	init_clk_states(clk_mgr);
+
+	// to adjust dp_dto reference clock if ssc is enable otherwise to apply dprefclk
+	if (dcn35_is_spll_ssc_enabled(clk_mgr))
+		clk_mgr->dp_dto_source_clock_in_khz =
+			dce_adjust_dp_ref_freq_for_ss(clk_mgr_int, clk_mgr->dprefclk_khz);
+	else
+		clk_mgr->dp_dto_source_clock_in_khz = clk_mgr->dprefclk_khz;
+
+	dcn35_save_clk_registers(&clk_mgr->boot_snapshot, clk_mgr_dcn35);
+
+	clk_mgr->clks.ref_dtbclk_khz =  clk_mgr->boot_snapshot.dtbclk * 10;
+	if (clk_mgr->boot_snapshot.dtbclk > 59000) {
+		/*dtbclk enabled based on */
+		clk_mgr->clks.dtbclk_en = true;
+	}
+}
+static struct clk_bw_params dcn35_bw_params = {
+	.vram_type = Ddr4MemType,
+	.num_channels = 1,
+	.clk_table = {
+		.num_entries = 4,
+	},
+
+};
+
+static struct wm_table ddr5_wm_table = {
+	.entries = {
+		{
+			.wm_inst = WM_A,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.72,
+			.sr_exit_time_us = 28.0,
+			.sr_enter_plus_exit_time_us = 30.0,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_B,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.72,
+			.sr_exit_time_us = 28.0,
+			.sr_enter_plus_exit_time_us = 30.0,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_C,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.72,
+			.sr_exit_time_us = 28.0,
+			.sr_enter_plus_exit_time_us = 30.0,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_D,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.72,
+			.sr_exit_time_us = 28.0,
+			.sr_enter_plus_exit_time_us = 30.0,
+			.valid = true,
+		},
+	}
+};
+
+static struct wm_table lpddr5_wm_table = {
+	.entries = {
+		{
+			.wm_inst = WM_A,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.65333,
+			.sr_exit_time_us = 28.0,
+			.sr_enter_plus_exit_time_us = 30.0,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_B,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.65333,
+			.sr_exit_time_us = 28.0,
+			.sr_enter_plus_exit_time_us = 30.0,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_C,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.65333,
+			.sr_exit_time_us = 28.0,
+			.sr_enter_plus_exit_time_us = 30.0,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_D,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.65333,
+			.sr_exit_time_us = 28.0,
+			.sr_enter_plus_exit_time_us = 30.0,
+			.valid = true,
+		},
+	}
+};
+
+static DpmClocks_t_dcn35 dummy_clocks;
+static DpmClocks_t_dcn351 dummy_clocks_dcn351;
+
+static struct dcn35_watermarks dummy_wms = { 0 };
+
+static struct dcn35_ss_info_table ss_info_table = {
+	.ss_divider = 1000,
+	.ss_percentage = {0, 0, 375, 375, 375}
+};
+
+static void dcn35_read_ss_info_from_lut(struct clk_mgr_internal *clk_mgr)
+{
+	uint32_t clock_source = 0;
+
+	clock_source = REG_READ(CLK1_CLK2_BYPASS_CNTL) & CLK1_CLK2_BYPASS_CNTL__CLK2_BYPASS_SEL_MASK;
+
+	// If it's DFS mode, clock_source is 0.
+	if (dcn35_is_spll_ssc_enabled(&clk_mgr->base) && (clock_source < ARRAY_SIZE(ss_info_table.ss_percentage))) {
+		clk_mgr->dprefclk_ss_percentage = ss_info_table.ss_percentage[clock_source];
+
+		if (clk_mgr->dprefclk_ss_percentage != 0) {
+			clk_mgr->ss_on_dprefclk = true;
+			clk_mgr->dprefclk_ss_divider = ss_info_table.ss_divider;
+		}
+	}
+}
+
+static void dcn35_build_watermark_ranges(struct clk_bw_params *bw_params, struct dcn35_watermarks *table)
+{
+	int i, num_valid_sets;
+
+	num_valid_sets = 0;
+
+	for (i = 0; i < WM_SET_COUNT; i++) {
+		/* skip empty entries, the smu array has no holes*/
+		if (!bw_params->wm_table.entries[i].valid)
+			continue;
+
+		table->WatermarkRow[WM_DCFCLK][num_valid_sets].WmSetting = bw_params->wm_table.entries[i].wm_inst;
+		table->WatermarkRow[WM_DCFCLK][num_valid_sets].WmType = bw_params->wm_table.entries[i].wm_type;
+		/* We will not select WM based on fclk, so leave it as unconstrained */
+		table->WatermarkRow[WM_DCFCLK][num_valid_sets].MinClock = 0;
+		table->WatermarkRow[WM_DCFCLK][num_valid_sets].MaxClock = 0xFFFF;
+
+		if (table->WatermarkRow[WM_DCFCLK][num_valid_sets].WmType == WM_TYPE_PSTATE_CHG) {
+			if (i == 0)
+				table->WatermarkRow[WM_DCFCLK][num_valid_sets].MinMclk = 0;
+			else {
+				/* add 1 to make it non-overlapping with next lvl */
+				table->WatermarkRow[WM_DCFCLK][num_valid_sets].MinMclk =
+						bw_params->clk_table.entries[i - 1].dcfclk_mhz + 1;
+			}
+			table->WatermarkRow[WM_DCFCLK][num_valid_sets].MaxMclk =
+					bw_params->clk_table.entries[i].dcfclk_mhz;
+
+		} else {
+			/* unconstrained for memory retraining */
+			table->WatermarkRow[WM_DCFCLK][num_valid_sets].MinClock = 0;
+			table->WatermarkRow[WM_DCFCLK][num_valid_sets].MaxClock = 0xFFFF;
+
+			/* Modify previous watermark range to cover up to max */
+			table->WatermarkRow[WM_DCFCLK][num_valid_sets - 1].MaxClock = 0xFFFF;
+		}
+		num_valid_sets++;
+	}
+
+	ASSERT(num_valid_sets != 0); /* Must have at least one set of valid watermarks */
+
+	/* modify the min and max to make sure we cover the whole range*/
+	table->WatermarkRow[WM_DCFCLK][0].MinMclk = 0;
+	table->WatermarkRow[WM_DCFCLK][0].MinClock = 0;
+	table->WatermarkRow[WM_DCFCLK][num_valid_sets - 1].MaxMclk = 0xFFFF;
+	table->WatermarkRow[WM_DCFCLK][num_valid_sets - 1].MaxClock = 0xFFFF;
+
+	/* This is for writeback only, does not matter currently as no writeback support*/
+	table->WatermarkRow[WM_SOCCLK][0].WmSetting = WM_A;
+	table->WatermarkRow[WM_SOCCLK][0].MinClock = 0;
+	table->WatermarkRow[WM_SOCCLK][0].MaxClock = 0xFFFF;
+	table->WatermarkRow[WM_SOCCLK][0].MinMclk = 0;
+	table->WatermarkRow[WM_SOCCLK][0].MaxMclk = 0xFFFF;
+}
+
+static void dcn35_notify_wm_ranges(struct clk_mgr *clk_mgr_base)
+{
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+	struct clk_mgr_dcn35 *clk_mgr_dcn35 = TO_CLK_MGR_DCN35(clk_mgr);
+	struct dcn35_watermarks *table = clk_mgr_dcn35->smu_wm_set.wm_set;
+
+	if (!clk_mgr->smu_ver)
+		return;
+
+	if (!table || clk_mgr_dcn35->smu_wm_set.mc_address.quad_part == 0)
+		return;
+
+	memset(table, 0, sizeof(*table));
+
+	dcn35_build_watermark_ranges(clk_mgr_base->bw_params, table);
+
+	dcn35_smu_set_dram_addr_high(clk_mgr,
+			clk_mgr_dcn35->smu_wm_set.mc_address.high_part);
+	dcn35_smu_set_dram_addr_low(clk_mgr,
+			clk_mgr_dcn35->smu_wm_set.mc_address.low_part);
+	dcn35_smu_transfer_wm_table_dram_2_smu(clk_mgr);
+}
+
+static void dcn35_get_dpm_table_from_smu(struct clk_mgr_internal *clk_mgr,
+		struct dcn35_smu_dpm_clks *smu_dpm_clks)
+{
+	DpmClocks_t_dcn35 *table = smu_dpm_clks->dpm_clks;
+
+	if (!clk_mgr->smu_ver)
+		return;
+
+	if (!table || smu_dpm_clks->mc_address.quad_part == 0)
+		return;
+
+	memset(table, 0, sizeof(*table));
+
+	dcn35_smu_set_dram_addr_high(clk_mgr,
+			smu_dpm_clks->mc_address.high_part);
+	dcn35_smu_set_dram_addr_low(clk_mgr,
+			smu_dpm_clks->mc_address.low_part);
+	dcn35_smu_transfer_dpm_table_smu_2_dram(clk_mgr);
+}
+
+static void dcn351_get_dpm_table_from_smu(struct clk_mgr_internal *clk_mgr,
+		struct dcn351_smu_dpm_clks *smu_dpm_clks)
+{
+	DpmClocks_t_dcn351 *table = smu_dpm_clks->dpm_clks;
+
+	if (!clk_mgr->smu_ver)
+		return;
+	if (!table || smu_dpm_clks->mc_address.quad_part == 0)
+		return;
+	memset(table, 0, sizeof(*table));
+	dcn35_smu_set_dram_addr_high(clk_mgr,
+			smu_dpm_clks->mc_address.high_part);
+	dcn35_smu_set_dram_addr_low(clk_mgr,
+			smu_dpm_clks->mc_address.low_part);
+	dcn35_smu_transfer_dpm_table_smu_2_dram(clk_mgr);
+}
+static uint32_t find_max_clk_value(const uint32_t clocks[], uint32_t num_clocks)
+{
+	uint32_t max = 0;
+	int i;
+
+	for (i = 0; i < num_clocks; ++i) {
+		if (clocks[i] > max)
+			max = clocks[i];
+	}
+
+	return max;
+}
+
+static inline bool is_valid_clock_value(uint32_t clock_value)
+{
+	return clock_value > 1 && clock_value < 100000;
+}
+
+static unsigned int convert_wck_ratio(uint8_t wck_ratio)
+{
+	switch (wck_ratio) {
+	case WCK_RATIO_1_2:
+		return 2;
+
+	case WCK_RATIO_1_4:
+		return 4;
+	/* Find lowest DPM, FCLK is filled in reverse order*/
+
+	default:
+			break;
+	}
+
+	return 1;
+}
+
+static inline uint32_t calc_dram_speed_mts(const MemPstateTable_t *entry)
+{
+	return entry->UClk * convert_wck_ratio(entry->WckRatio) * 2;
+}
+
+static void dcn35_clk_mgr_helper_populate_bw_params(struct clk_mgr_internal *clk_mgr,
+						    struct integrated_info *bios_info,
+						    DpmClocks_t_dcn35 *clock_table)
+{
+	struct clk_bw_params *bw_params = clk_mgr->base.bw_params;
+	struct clk_limit_table_entry def_max = bw_params->clk_table.entries[bw_params->clk_table.num_entries - 1];
+	uint32_t max_fclk = 0, min_pstate = 0, max_dispclk = 0, max_dppclk = 0;
+	uint32_t max_pstate = 0, max_dram_speed_mts = 0, min_dram_speed_mts = 0;
+	uint32_t num_memps, num_fclk, num_dcfclk;
+	int i;
+
+	/* Determine min/max p-state values. */
+	num_memps = (clock_table->NumMemPstatesEnabled > NUM_MEM_PSTATE_LEVELS) ? NUM_MEM_PSTATE_LEVELS :
+		clock_table->NumMemPstatesEnabled;
+	for (i = 0; i < num_memps; i++) {
+		uint32_t dram_speed_mts = calc_dram_speed_mts(&clock_table->MemPstateTable[i]);
+
+		if (is_valid_clock_value(dram_speed_mts) && dram_speed_mts > max_dram_speed_mts) {
+			max_dram_speed_mts = dram_speed_mts;
+			max_pstate = i;
+		}
+	}
+
+	min_dram_speed_mts = max_dram_speed_mts;
+	min_pstate = max_pstate;
+
+	for (i = 0; i < num_memps; i++) {
+		uint32_t dram_speed_mts = calc_dram_speed_mts(&clock_table->MemPstateTable[i]);
+
+		if (is_valid_clock_value(dram_speed_mts) && dram_speed_mts < min_dram_speed_mts) {
+			min_dram_speed_mts = dram_speed_mts;
+			min_pstate = i;
+		}
+	}
+
+	/* We expect the table to contain at least one valid P-state entry. */
+	ASSERT(clock_table->NumMemPstatesEnabled &&
+	       is_valid_clock_value(max_dram_speed_mts) &&
+	       is_valid_clock_value(min_dram_speed_mts));
+
+	/* dispclk and dppclk can be max at any voltage, same number of levels for both */
+	if (clock_table->NumDispClkLevelsEnabled <= NUM_DISPCLK_DPM_LEVELS &&
+	    clock_table->NumDispClkLevelsEnabled <= NUM_DPPCLK_DPM_LEVELS) {
+		max_dispclk = find_max_clk_value(clock_table->DispClocks,
+			clock_table->NumDispClkLevelsEnabled);
+		max_dppclk = find_max_clk_value(clock_table->DppClocks,
+			clock_table->NumDispClkLevelsEnabled);
+	} else {
+		/* Invalid number of entries in the table from PMFW. */
+		ASSERT(0);
+	}
+
+	/* Base the clock table on dcfclk, need at least one entry regardless of pmfw table */
+	ASSERT(clock_table->NumDcfClkLevelsEnabled > 0);
+
+	num_fclk = (clock_table->NumFclkLevelsEnabled > NUM_FCLK_DPM_LEVELS) ? NUM_FCLK_DPM_LEVELS :
+		clock_table->NumFclkLevelsEnabled;
+	max_fclk = find_max_clk_value(clock_table->FclkClocks_Freq, num_fclk);
+
+	num_dcfclk = (clock_table->NumDcfClkLevelsEnabled > NUM_DCFCLK_DPM_LEVELS) ? NUM_DCFCLK_DPM_LEVELS :
+		clock_table->NumDcfClkLevelsEnabled;
+	for (i = 0; i < num_dcfclk; i++) {
+		int j;
+
+		/* First search defaults for the clocks we don't read using closest lower or equal default dcfclk */
+		for (j = bw_params->clk_table.num_entries - 1; j > 0; j--)
+			if (bw_params->clk_table.entries[j].dcfclk_mhz <= clock_table->DcfClocks[i])
+				break;
+
+		bw_params->clk_table.entries[i].phyclk_mhz = bw_params->clk_table.entries[j].phyclk_mhz;
+		bw_params->clk_table.entries[i].phyclk_d18_mhz = bw_params->clk_table.entries[j].phyclk_d18_mhz;
+		bw_params->clk_table.entries[i].dtbclk_mhz = bw_params->clk_table.entries[j].dtbclk_mhz;
+
+		/* Now update clocks we do read */
+		bw_params->clk_table.entries[i].memclk_mhz = clock_table->MemPstateTable[min_pstate].MemClk;
+		bw_params->clk_table.entries[i].voltage = clock_table->MemPstateTable[min_pstate].Voltage;
+		bw_params->clk_table.entries[i].dcfclk_mhz = clock_table->DcfClocks[i];
+		bw_params->clk_table.entries[i].socclk_mhz = clock_table->SocClocks[i];
+		bw_params->clk_table.entries[i].dispclk_mhz = max_dispclk;
+		bw_params->clk_table.entries[i].dppclk_mhz = max_dppclk;
+		bw_params->clk_table.entries[i].wck_ratio =
+			convert_wck_ratio(clock_table->MemPstateTable[min_pstate].WckRatio);
+
+		/* Dcfclk and Fclk are tied, but at a different ratio */
+		bw_params->clk_table.entries[i].fclk_mhz = min(max_fclk, 2 * clock_table->DcfClocks[i]);
+	}
+
+	/* Make sure to include at least one entry at highest pstate */
+	if (max_pstate != min_pstate || i == 0) {
+		if (i > MAX_NUM_DPM_LVL - 1)
+			i = MAX_NUM_DPM_LVL - 1;
+
+		bw_params->clk_table.entries[i].fclk_mhz = max_fclk;
+		bw_params->clk_table.entries[i].memclk_mhz = clock_table->MemPstateTable[max_pstate].MemClk;
+		bw_params->clk_table.entries[i].voltage = clock_table->MemPstateTable[max_pstate].Voltage;
+		bw_params->clk_table.entries[i].dcfclk_mhz =
+			find_max_clk_value(clock_table->DcfClocks, NUM_DCFCLK_DPM_LEVELS);
+		bw_params->clk_table.entries[i].socclk_mhz =
+			find_max_clk_value(clock_table->SocClocks, NUM_SOCCLK_DPM_LEVELS);
+		bw_params->clk_table.entries[i].dispclk_mhz = max_dispclk;
+		bw_params->clk_table.entries[i].dppclk_mhz = max_dppclk;
+		bw_params->clk_table.entries[i].wck_ratio = convert_wck_ratio(
+			clock_table->MemPstateTable[max_pstate].WckRatio);
+		i++;
+	}
+	bw_params->clk_table.num_entries = i--;
+
+	/* Make sure all highest clocks are included*/
+	bw_params->clk_table.entries[i].socclk_mhz =
+		find_max_clk_value(clock_table->SocClocks, NUM_SOCCLK_DPM_LEVELS);
+	bw_params->clk_table.entries[i].dispclk_mhz =
+		find_max_clk_value(clock_table->DispClocks, NUM_DISPCLK_DPM_LEVELS);
+	bw_params->clk_table.entries[i].dppclk_mhz =
+		find_max_clk_value(clock_table->DppClocks, NUM_DPPCLK_DPM_LEVELS);
+	bw_params->clk_table.entries[i].fclk_mhz =
+		find_max_clk_value(clock_table->FclkClocks_Freq, NUM_FCLK_DPM_LEVELS);
+	ASSERT(clock_table->DcfClocks[i] == find_max_clk_value(clock_table->DcfClocks, NUM_DCFCLK_DPM_LEVELS));
+	bw_params->clk_table.entries[i].phyclk_mhz = def_max.phyclk_mhz;
+	bw_params->clk_table.entries[i].phyclk_d18_mhz = def_max.phyclk_d18_mhz;
+	bw_params->clk_table.entries[i].dtbclk_mhz = def_max.dtbclk_mhz;
+	bw_params->clk_table.num_entries_per_clk.num_dcfclk_levels = clock_table->NumDcfClkLevelsEnabled;
+	bw_params->clk_table.num_entries_per_clk.num_dispclk_levels = clock_table->NumDispClkLevelsEnabled;
+	bw_params->clk_table.num_entries_per_clk.num_dppclk_levels = clock_table->NumDispClkLevelsEnabled;
+	bw_params->clk_table.num_entries_per_clk.num_fclk_levels = clock_table->NumFclkLevelsEnabled;
+	bw_params->clk_table.num_entries_per_clk.num_memclk_levels = clock_table->NumMemPstatesEnabled;
+	bw_params->clk_table.num_entries_per_clk.num_socclk_levels = clock_table->NumSocClkLevelsEnabled;
+
+	/*
+	 * Set any 0 clocks to max default setting. Not an issue for
+	 * power since we aren't doing switching in such case anyway
+	 */
+	for (i = 0; i < bw_params->clk_table.num_entries; i++) {
+		if (!bw_params->clk_table.entries[i].fclk_mhz) {
+			bw_params->clk_table.entries[i].fclk_mhz = def_max.fclk_mhz;
+			bw_params->clk_table.entries[i].memclk_mhz = def_max.memclk_mhz;
+			bw_params->clk_table.entries[i].voltage = def_max.voltage;
+		}
+		if (!bw_params->clk_table.entries[i].dcfclk_mhz)
+			bw_params->clk_table.entries[i].dcfclk_mhz = def_max.dcfclk_mhz;
+		if (!bw_params->clk_table.entries[i].socclk_mhz)
+			bw_params->clk_table.entries[i].socclk_mhz = def_max.socclk_mhz;
+		if (!bw_params->clk_table.entries[i].dispclk_mhz)
+			bw_params->clk_table.entries[i].dispclk_mhz = def_max.dispclk_mhz;
+		if (!bw_params->clk_table.entries[i].dppclk_mhz)
+			bw_params->clk_table.entries[i].dppclk_mhz = def_max.dppclk_mhz;
+		if (!bw_params->clk_table.entries[i].fclk_mhz)
+			bw_params->clk_table.entries[i].fclk_mhz = def_max.fclk_mhz;
+		if (!bw_params->clk_table.entries[i].phyclk_mhz)
+			bw_params->clk_table.entries[i].phyclk_mhz = def_max.phyclk_mhz;
+		if (!bw_params->clk_table.entries[i].phyclk_d18_mhz)
+			bw_params->clk_table.entries[i].phyclk_d18_mhz = def_max.phyclk_d18_mhz;
+		if (!bw_params->clk_table.entries[i].dtbclk_mhz)
+			bw_params->clk_table.entries[i].dtbclk_mhz = def_max.dtbclk_mhz;
+	}
+	ASSERT(bw_params->clk_table.entries[i-1].dcfclk_mhz);
+	bw_params->vram_type = bios_info->memory_type;
+	bw_params->dram_channel_width_bytes = bios_info->memory_type == 0x22 ? 8 : 4;
+	bw_params->num_channels = bios_info->ma_channel_number ? bios_info->ma_channel_number : 4;
+
+	for (i = 0; i < WM_SET_COUNT; i++) {
+		bw_params->wm_table.entries[i].wm_inst = i;
+
+		if (i >= bw_params->clk_table.num_entries) {
+			bw_params->wm_table.entries[i].valid = false;
+			continue;
+		}
+
+		bw_params->wm_table.entries[i].wm_type = WM_TYPE_PSTATE_CHG;
+		bw_params->wm_table.entries[i].valid = true;
+	}
+}
+
+static void dcn35_set_low_power_state(struct clk_mgr *clk_mgr_base)
+{
+	int display_count;
+	struct dc *dc = clk_mgr_base->ctx->dc;
+	struct dc_state *context = dc->current_state;
+
+	if (clk_mgr_base->clks.pwr_state != DCN_PWR_STATE_LOW_POWER) {
+		display_count = dcn35_get_active_display_cnt_wa(dc, context, NULL);
+		/* if we can go lower, go lower */
+		if (display_count == 0)
+			clk_mgr_base->clks.pwr_state = DCN_PWR_STATE_LOW_POWER;
+	}
+}
+
+static void dcn35_exit_low_power_state(struct clk_mgr *clk_mgr_base)
+{
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+
+	//SMU optimization is performed part of low power state exit.
+	dcn35_smu_exit_low_power_state(clk_mgr);
+
+}
+
+static bool dcn35_is_ips_supported(struct clk_mgr *clk_mgr_base)
+{
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+
+	return dcn35_smu_get_ips_supported(clk_mgr) ? true : false;
+}
+
+static void dcn35_init_clocks_fpga(struct clk_mgr *clk_mgr)
+{
+	init_clk_states(clk_mgr);
+
+/* TODO: Implement the functions and remove the ifndef guard */
+}
+
+static void dcn35_update_clocks_fpga(struct clk_mgr *clk_mgr,
+		struct dc_state *context,
+		bool safe_to_lower)
+{
+	struct clk_mgr_internal *clk_mgr_int = TO_CLK_MGR_INTERNAL(clk_mgr);
+	struct dc_clocks *new_clocks = &context->bw_ctx.bw.dcn.clk;
+	int fclk_adj = new_clocks->fclk_khz;
+
+	/* TODO: remove this after correctly set by DML */
+	new_clocks->dcfclk_khz = 400000;
+	new_clocks->socclk_khz = 400000;
+
+	/* Min fclk = 1.2GHz since all the extra scemi logic seems to run off of it */
+	//int fclk_adj = new_clocks->fclk_khz > 1200000 ? new_clocks->fclk_khz : 1200000;
+	new_clocks->fclk_khz = 4320000;
+
+	if (should_set_clock(safe_to_lower, new_clocks->phyclk_khz, clk_mgr->clks.phyclk_khz)) {
+		clk_mgr->clks.phyclk_khz = new_clocks->phyclk_khz;
+	}
+
+	if (should_set_clock(safe_to_lower, new_clocks->dcfclk_khz, clk_mgr->clks.dcfclk_khz)) {
+		clk_mgr->clks.dcfclk_khz = new_clocks->dcfclk_khz;
+	}
+
+	if (should_set_clock(safe_to_lower,
+			new_clocks->dcfclk_deep_sleep_khz, clk_mgr->clks.dcfclk_deep_sleep_khz)) {
+		clk_mgr->clks.dcfclk_deep_sleep_khz = new_clocks->dcfclk_deep_sleep_khz;
+	}
+
+	if (should_set_clock(safe_to_lower, new_clocks->socclk_khz, clk_mgr->clks.socclk_khz)) {
+		clk_mgr->clks.socclk_khz = new_clocks->socclk_khz;
+	}
+
+	if (should_set_clock(safe_to_lower, new_clocks->dramclk_khz, clk_mgr->clks.dramclk_khz)) {
+		clk_mgr->clks.dramclk_khz = new_clocks->dramclk_khz;
+	}
+
+	if (should_set_clock(safe_to_lower, new_clocks->dppclk_khz, clk_mgr->clks.dppclk_khz)) {
+		clk_mgr->clks.dppclk_khz = new_clocks->dppclk_khz;
+	}
+
+	if (should_set_clock(safe_to_lower, fclk_adj, clk_mgr->clks.fclk_khz)) {
+		clk_mgr->clks.fclk_khz = fclk_adj;
+	}
+
+	if (should_set_clock(safe_to_lower, new_clocks->dispclk_khz, clk_mgr->clks.dispclk_khz)) {
+		clk_mgr->clks.dispclk_khz = new_clocks->dispclk_khz;
+	}
+
+	/* Both fclk and ref_dppclk run on the same scemi clock.
+	 * So take the higher value since the DPP DTO is typically programmed
+	 * such that max dppclk is 1:1 with ref_dppclk.
+	 */
+	if (clk_mgr->clks.fclk_khz > clk_mgr->clks.dppclk_khz)
+		clk_mgr->clks.dppclk_khz = clk_mgr->clks.fclk_khz;
+	if (clk_mgr->clks.dppclk_khz > clk_mgr->clks.fclk_khz)
+		clk_mgr->clks.fclk_khz = clk_mgr->clks.dppclk_khz;
+
+	// Both fclk and ref_dppclk run on the same scemi clock.
+	clk_mgr_int->dccg->ref_dppclk = clk_mgr->clks.fclk_khz;
+
+	/* TODO: set dtbclk in correct place */
+	clk_mgr->clks.dtbclk_en = true;
+	dm_set_dcn_clocks(clk_mgr->ctx, &clk_mgr->clks);
+	dcn35_update_clocks_update_dpp_dto(clk_mgr_int, context, safe_to_lower);
+
+	dcn35_update_clocks_update_dtb_dto(clk_mgr_int, context, clk_mgr->clks.ref_dtbclk_khz);
+}
+
+static unsigned int dcn35_get_max_clock_khz(struct clk_mgr *clk_mgr_base, enum clk_type clk_type)
+{
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+
+	unsigned int num_clk_levels;
+
+	switch (clk_type) {
+	case CLK_TYPE_DISPCLK:
+		num_clk_levels = clk_mgr->base.bw_params->clk_table.num_entries_per_clk.num_dispclk_levels;
+		return num_clk_levels ?
+				clk_mgr->base.bw_params->clk_table.entries[num_clk_levels - 1].dispclk_mhz * 1000 :
+				clk_mgr->base.boot_snapshot.dispclk;
+	case CLK_TYPE_DPPCLK:
+		num_clk_levels = clk_mgr->base.bw_params->clk_table.num_entries_per_clk.num_dppclk_levels;
+		return num_clk_levels ?
+				clk_mgr->base.bw_params->clk_table.entries[num_clk_levels - 1].dppclk_mhz * 1000 :
+				clk_mgr->base.boot_snapshot.dppclk;
+	case CLK_TYPE_DSCCLK:
+		num_clk_levels = clk_mgr->base.bw_params->clk_table.num_entries_per_clk.num_dispclk_levels;
+		return num_clk_levels ?
+				clk_mgr->base.bw_params->clk_table.entries[num_clk_levels - 1].dispclk_mhz * 1000 / 3 :
+				clk_mgr->base.boot_snapshot.dispclk / 3;
+	default:
+		break;
+	}
+
+	return 0;
+}
+
+static struct clk_mgr_funcs dcn35_funcs = {
+	.get_dp_ref_clk_frequency = dce12_get_dp_ref_freq_khz,
+	.get_dtb_ref_clk_frequency = dcn31_get_dtb_ref_freq_khz,
+	.update_clocks = dcn35_update_clocks,
+	.init_clocks = dcn35_init_clocks,
+	.enable_pme_wa = dcn35_enable_pme_wa,
+	.are_clock_states_equal = dcn35_are_clock_states_equal,
+	.notify_wm_ranges = dcn35_notify_wm_ranges,
+	.set_low_power_state = dcn35_set_low_power_state,
+	.exit_low_power_state = dcn35_exit_low_power_state,
+	.is_ips_supported = dcn35_is_ips_supported,
+	.get_max_clock_khz = dcn35_get_max_clock_khz,
+};
+
+struct clk_mgr_funcs dcn35_fpga_funcs = {
+	.get_dp_ref_clk_frequency = dce12_get_dp_ref_freq_khz,
+	.update_clocks = dcn35_update_clocks_fpga,
+	.init_clocks = dcn35_init_clocks_fpga,
+	.get_dtb_ref_clk_frequency = dcn31_get_dtb_ref_freq_khz,
+};
+
+static void translate_to_DpmClocks_t_dcn35(struct dcn351_smu_dpm_clks *smu_dpm_clks_a,
+		struct dcn35_smu_dpm_clks *smu_dpm_clks_b)
+{
+	/*translate two structures and only take need clock tables*/
+	uint8_t i;
+
+	if (smu_dpm_clks_a == NULL || smu_dpm_clks_b == NULL ||
+		smu_dpm_clks_a->dpm_clks == NULL || smu_dpm_clks_b->dpm_clks == NULL)
+		return;
+
+	for (i = 0; i < NUM_DCFCLK_DPM_LEVELS; i++)
+		smu_dpm_clks_b->dpm_clks->DcfClocks[i] = smu_dpm_clks_a->dpm_clks->DcfClocks[i];
+
+	for (i = 0; i < NUM_DISPCLK_DPM_LEVELS; i++)
+		smu_dpm_clks_b->dpm_clks->DispClocks[i] = smu_dpm_clks_a->dpm_clks->DispClocks[i];
+
+	for (i = 0; i < NUM_DPPCLK_DPM_LEVELS; i++)
+		smu_dpm_clks_b->dpm_clks->DppClocks[i] = smu_dpm_clks_a->dpm_clks->DppClocks[i];
+
+	for (i = 0; i < NUM_FCLK_DPM_LEVELS; i++) {
+		smu_dpm_clks_b->dpm_clks->FclkClocks_Freq[i] = smu_dpm_clks_a->dpm_clks->FclkClocks_Freq[i];
+		smu_dpm_clks_b->dpm_clks->FclkClocks_Voltage[i] = smu_dpm_clks_a->dpm_clks->FclkClocks_Voltage[i];
+	}
+	for (i = 0; i < NUM_MEM_PSTATE_LEVELS; i++) {
+		smu_dpm_clks_b->dpm_clks->MemPstateTable[i].MemClk =
+			smu_dpm_clks_a->dpm_clks->MemPstateTable[i].MemClk;
+		smu_dpm_clks_b->dpm_clks->MemPstateTable[i].UClk =
+			smu_dpm_clks_a->dpm_clks->MemPstateTable[i].UClk;
+		smu_dpm_clks_b->dpm_clks->MemPstateTable[i].Voltage =
+			smu_dpm_clks_a->dpm_clks->MemPstateTable[i].Voltage;
+		smu_dpm_clks_b->dpm_clks->MemPstateTable[i].WckRatio =
+			smu_dpm_clks_a->dpm_clks->MemPstateTable[i].WckRatio;
+	}
+	smu_dpm_clks_b->dpm_clks->MaxGfxClk = smu_dpm_clks_a->dpm_clks->MaxGfxClk;
+	smu_dpm_clks_b->dpm_clks->MinGfxClk = smu_dpm_clks_a->dpm_clks->MinGfxClk;
+	smu_dpm_clks_b->dpm_clks->NumDcfClkLevelsEnabled =
+		smu_dpm_clks_a->dpm_clks->NumDcfClkLevelsEnabled;
+	smu_dpm_clks_b->dpm_clks->NumDispClkLevelsEnabled =
+		smu_dpm_clks_a->dpm_clks->NumDispClkLevelsEnabled;
+	smu_dpm_clks_b->dpm_clks->NumFclkLevelsEnabled =
+		smu_dpm_clks_a->dpm_clks->NumFclkLevelsEnabled;
+	smu_dpm_clks_b->dpm_clks->NumMemPstatesEnabled =
+		smu_dpm_clks_a->dpm_clks->NumMemPstatesEnabled;
+	smu_dpm_clks_b->dpm_clks->NumSocClkLevelsEnabled =
+		smu_dpm_clks_a->dpm_clks->NumSocClkLevelsEnabled;
+
+	for (i = 0; i < NUM_SOC_VOLTAGE_LEVELS; i++) {
+		smu_dpm_clks_b->dpm_clks->SocClocks[i] = smu_dpm_clks_a->dpm_clks->SocClocks[i];
+		smu_dpm_clks_b->dpm_clks->SocVoltage[i] = smu_dpm_clks_a->dpm_clks->SocVoltage[i];
+	}
+}
+void dcn35_clk_mgr_construct(
+		struct dc_context *ctx,
+		struct clk_mgr_dcn35 *clk_mgr,
+		struct pp_smu_funcs *pp_smu,
+		struct dccg *dccg)
+{
+	struct dcn35_smu_dpm_clks smu_dpm_clks = { 0 };
+	struct dcn351_smu_dpm_clks smu_dpm_clks_dcn351 = { 0 };
+	clk_mgr->base.base.ctx = ctx;
+	clk_mgr->base.base.funcs = &dcn35_funcs;
+
+	clk_mgr->base.pp_smu = pp_smu;
+
+	clk_mgr->base.dccg = dccg;
+	clk_mgr->base.dfs_bypass_disp_clk = 0;
+
+	clk_mgr->base.dprefclk_ss_percentage = 0;
+	clk_mgr->base.dprefclk_ss_divider = 1000;
+	clk_mgr->base.ss_on_dprefclk = false;
+	clk_mgr->base.dfs_ref_freq_khz = 48000;
+	if (ctx->dce_version != DCN_VERSION_3_51) {
+		clk_mgr->base.regs = &clk_mgr_regs_dcn35;
+		clk_mgr->base.clk_mgr_shift = &clk_mgr_shift_dcn35;
+		clk_mgr->base.clk_mgr_mask = &clk_mgr_mask_dcn35;
+	}
+
+
+	clk_mgr->smu_wm_set.wm_set = (struct dcn35_watermarks *)dm_helpers_allocate_gpu_mem(
+				clk_mgr->base.base.ctx,
+				DC_MEM_ALLOC_TYPE_GART,
+				sizeof(struct dcn35_watermarks),
+				&clk_mgr->smu_wm_set.mc_address.quad_part);
+
+	if (!clk_mgr->smu_wm_set.wm_set) {
+		clk_mgr->smu_wm_set.wm_set = &dummy_wms;
+		clk_mgr->smu_wm_set.mc_address.quad_part = 0;
+	}
+	ASSERT(clk_mgr->smu_wm_set.wm_set);
+
+	smu_dpm_clks.dpm_clks = (DpmClocks_t_dcn35 *)dm_helpers_allocate_gpu_mem(
+				clk_mgr->base.base.ctx,
+				DC_MEM_ALLOC_TYPE_GART,
+				sizeof(DpmClocks_t_dcn35),
+				&smu_dpm_clks.mc_address.quad_part);
+	if (smu_dpm_clks.dpm_clks == NULL) {
+		smu_dpm_clks.dpm_clks = &dummy_clocks;
+		smu_dpm_clks.mc_address.quad_part = 0;
+	}
+	ASSERT(smu_dpm_clks.dpm_clks);
+
+	if (ctx->dce_version == DCN_VERSION_3_51) {
+		smu_dpm_clks_dcn351.dpm_clks = (DpmClocks_t_dcn351 *)dm_helpers_allocate_gpu_mem(
+				clk_mgr->base.base.ctx,
+				DC_MEM_ALLOC_TYPE_GART,
+				sizeof(DpmClocks_t_dcn351),
+				&smu_dpm_clks_dcn351.mc_address.quad_part);
+		if (smu_dpm_clks_dcn351.dpm_clks == NULL) {
+			smu_dpm_clks_dcn351.dpm_clks = &dummy_clocks_dcn351;
+			smu_dpm_clks_dcn351.mc_address.quad_part = 0;
+		}
+	}
+
+	clk_mgr->base.smu_ver = dcn35_smu_get_smu_version(&clk_mgr->base);
+
+	if (clk_mgr->base.smu_ver)
+		clk_mgr->base.smu_present = true;
+
+	/* TODO: Check we get what we expect during bringup */
+	clk_mgr->base.base.dentist_vco_freq_khz = get_vco_frequency_from_reg(&clk_mgr->base);
+
+	if (ctx->dc_bios->integrated_info->memory_type == LpDdr5MemType) {
+		dcn35_bw_params.wm_table = lpddr5_wm_table;
+	} else {
+		dcn35_bw_params.wm_table = ddr5_wm_table;
+	}
+	/* Saved clocks configured at boot for debug purposes */
+	dcn35_save_clk_registers(&clk_mgr->base.base.boot_snapshot, clk_mgr);
+
+	clk_mgr->base.base.dprefclk_khz = dcn35_smu_get_dprefclk(&clk_mgr->base);
+	clk_mgr->base.base.clks.ref_dtbclk_khz = 600000;
+
+	dce_clock_read_ss_info(&clk_mgr->base);
+	/*when clk src is from FCH, it could have ss, same clock src as DPREF clk*/
+
+	dcn35_read_ss_info_from_lut(&clk_mgr->base);
+
+	clk_mgr->base.base.bw_params = &dcn35_bw_params;
+
+	if (clk_mgr->base.base.ctx->dc->debug.pstate_enabled) {
+		int i;
+		if (ctx->dce_version == DCN_VERSION_3_51) {
+			dcn351_get_dpm_table_from_smu(&clk_mgr->base, &smu_dpm_clks_dcn351);
+			translate_to_DpmClocks_t_dcn35(&smu_dpm_clks_dcn351, &smu_dpm_clks);
+		} else
+			dcn35_get_dpm_table_from_smu(&clk_mgr->base, &smu_dpm_clks);
+		DC_LOG_SMU("NumDcfClkLevelsEnabled: %d\n"
+				   "NumDispClkLevelsEnabled: %d\n"
+				   "NumSocClkLevelsEnabled: %d\n"
+				   "VcnClkLevelsEnabled: %d\n"
+				   "FClkLevelsEnabled: %d\n"
+				   "NumMemPstatesEnabled: %d\n"
+				   "MinGfxClk: %d\n"
+				   "MaxGfxClk: %d\n",
+				   smu_dpm_clks.dpm_clks->NumDcfClkLevelsEnabled,
+				   smu_dpm_clks.dpm_clks->NumDispClkLevelsEnabled,
+				   smu_dpm_clks.dpm_clks->NumSocClkLevelsEnabled,
+				   smu_dpm_clks.dpm_clks->VcnClkLevelsEnabled,
+				   smu_dpm_clks.dpm_clks->NumFclkLevelsEnabled,
+				   smu_dpm_clks.dpm_clks->NumMemPstatesEnabled,
+				   smu_dpm_clks.dpm_clks->MinGfxClk,
+				   smu_dpm_clks.dpm_clks->MaxGfxClk);
+		for (i = 0; i < smu_dpm_clks.dpm_clks->NumDcfClkLevelsEnabled; i++) {
+			DC_LOG_SMU("smu_dpm_clks.dpm_clks->DcfClocks[%d] = %d\n",
+					   i,
+					   smu_dpm_clks.dpm_clks->DcfClocks[i]);
+		}
+		for (i = 0; i < smu_dpm_clks.dpm_clks->NumDispClkLevelsEnabled; i++) {
+			DC_LOG_SMU("smu_dpm_clks.dpm_clks->DispClocks[%d] = %d\n",
+					   i, smu_dpm_clks.dpm_clks->DispClocks[i]);
+		}
+		for (i = 0; i < smu_dpm_clks.dpm_clks->NumSocClkLevelsEnabled; i++) {
+			DC_LOG_SMU("smu_dpm_clks.dpm_clks->SocClocks[%d] = %d\n",
+					   i, smu_dpm_clks.dpm_clks->SocClocks[i]);
+		}
+		for (i = 0; i < smu_dpm_clks.dpm_clks->NumFclkLevelsEnabled; i++) {
+			DC_LOG_SMU("smu_dpm_clks.dpm_clks->FclkClocks_Freq[%d] = %d\n",
+					   i, smu_dpm_clks.dpm_clks->FclkClocks_Freq[i]);
+			DC_LOG_SMU("smu_dpm_clks.dpm_clks->FclkClocks_Voltage[%d] = %d\n",
+					   i, smu_dpm_clks.dpm_clks->FclkClocks_Voltage[i]);
+		}
+		for (i = 0; i < smu_dpm_clks.dpm_clks->NumSocClkLevelsEnabled; i++)
+			DC_LOG_SMU("smu_dpm_clks.dpm_clks->SocVoltage[%d] = %d\n",
+					   i, smu_dpm_clks.dpm_clks->SocVoltage[i]);
+
+		for (i = 0; i < smu_dpm_clks.dpm_clks->NumMemPstatesEnabled; i++) {
+			DC_LOG_SMU("smu_dpm_clks.dpm_clks.MemPstateTable[%d].UClk = %d\n"
+					   "smu_dpm_clks.dpm_clks->MemPstateTable[%d].MemClk= %d\n"
+					   "smu_dpm_clks.dpm_clks->MemPstateTable[%d].Voltage = %d\n",
+					   i, smu_dpm_clks.dpm_clks->MemPstateTable[i].UClk,
+					   i, smu_dpm_clks.dpm_clks->MemPstateTable[i].MemClk,
+					   i, smu_dpm_clks.dpm_clks->MemPstateTable[i].Voltage);
+		}
+
+		if (ctx->dc_bios->integrated_info && ctx->dc->config.use_default_clock_table == false) {
+			dcn35_clk_mgr_helper_populate_bw_params(
+					&clk_mgr->base,
+					ctx->dc_bios->integrated_info,
+					smu_dpm_clks.dpm_clks);
+		}
+	}
+
+	if (smu_dpm_clks.dpm_clks && smu_dpm_clks.mc_address.quad_part != 0)
+		dm_helpers_free_gpu_mem(clk_mgr->base.base.ctx, DC_MEM_ALLOC_TYPE_GART,
+				smu_dpm_clks.dpm_clks);
+
+	if (smu_dpm_clks_dcn351.dpm_clks && smu_dpm_clks_dcn351.mc_address.quad_part != 0)
+		dm_helpers_free_gpu_mem(clk_mgr->base.base.ctx, DC_MEM_ALLOC_TYPE_GART,
+				smu_dpm_clks_dcn351.dpm_clks);
+
+	if (ctx->dc->config.disable_ips != DMUB_IPS_DISABLE_ALL) {
+		bool ips_support = false;
+
+		/*avoid call pmfw at init*/
+		ips_support = dcn35_smu_get_ips_supported(&clk_mgr->base);
+		if (ips_support) {
+			ctx->dc->debug.ignore_pg = false;
+			ctx->dc->debug.disable_dpp_power_gate = false;
+			ctx->dc->debug.disable_hubp_power_gate = false;
+			ctx->dc->debug.disable_dsc_power_gate = false;
+
+			/* Disable dynamic IPS2 in older PMFW (93.12) for Z8 interop. */
+			if (ctx->dc->config.disable_ips == DMUB_IPS_ENABLE &&
+			    ctx->dce_version != DCN_VERSION_3_51 &&
+			    ((clk_mgr->base.smu_ver & 0x00FFFFFF) <= 0x005d0c00))
+				ctx->dc->config.disable_ips = DMUB_IPS_RCG_IN_ACTIVE_IPS2_IN_OFF;
+		} else {
+			/*let's reset the config control flag*/
+			ctx->dc->config.disable_ips = DMUB_IPS_DISABLE_ALL; /*pmfw not support it, disable it all*/
+		}
+	}
+}
+
+void dcn35_clk_mgr_destroy(struct clk_mgr_internal *clk_mgr_int)
+{
+	struct clk_mgr_dcn35 *clk_mgr = TO_CLK_MGR_DCN35(clk_mgr_int);
+
+	if (clk_mgr->smu_wm_set.wm_set && clk_mgr->smu_wm_set.mc_address.quad_part != 0)
+		dm_helpers_free_gpu_mem(clk_mgr_int->base.ctx, DC_MEM_ALLOC_TYPE_FRAME_BUFFER,
+				clk_mgr->smu_wm_set.wm_set);
+}
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.h b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.h
new file mode 100644
index 000000000000..a12a9bf90806
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.h
@@ -0,0 +1,67 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __DCN35_CLK_MGR_H__
+#define __DCN35_CLK_MGR_H__
+#include "clk_mgr_internal.h"
+
+#define NUM_CLOCK_SOURCES   5
+
+struct dcn35_watermarks;
+
+struct dcn35_smu_watermark_set {
+	struct dcn35_watermarks *wm_set;
+	union large_integer mc_address;
+};
+
+struct dcn35_ss_info_table {
+	uint32_t ss_divider;
+	uint32_t ss_percentage[NUM_CLOCK_SOURCES];
+};
+
+struct clk_mgr_dcn35 {
+	struct clk_mgr_internal base;
+	struct dcn35_smu_watermark_set smu_wm_set;
+};
+
+bool dcn35_are_clock_states_equal(struct dc_clocks *a,
+		struct dc_clocks *b);
+void dcn35_init_clocks(struct clk_mgr *clk_mgr);
+void dcn35_update_clocks(struct clk_mgr *clk_mgr_base,
+			struct dc_state *context,
+			bool safe_to_lower);
+
+void dcn35_clk_mgr_construct(struct dc_context *ctx,
+		struct clk_mgr_dcn35 *clk_mgr,
+		struct pp_smu_funcs *pp_smu,
+		struct dccg *dccg);
+
+void dcn35_clk_mgr_destroy(struct clk_mgr_internal *clk_mgr_int);
+
+void dcn351_clk_mgr_construct(struct dc_context *ctx,
+		struct clk_mgr_dcn35 *clk_mgr,
+		struct pp_smu_funcs *pp_smu,
+		struct dccg *dccg);
+#endif //__DCN35_CLK_MGR_H__
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_smu.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_smu.c
new file mode 100644
index 000000000000..604d256cb47a
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_smu.c
@@ -0,0 +1,508 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+
+
+#include "core_types.h"
+#include "clk_mgr_internal.h"
+#include "reg_helper.h"
+#include "dm_helpers.h"
+#include "dcn35_smu.h"
+
+#include "mp/mp_14_0_0_offset.h"
+#include "mp/mp_14_0_0_sh_mask.h"
+
+/* TODO: Use the real headers when they're correct */
+#define MP1_BASE__INST0_SEG0                       0x00016000
+#define MP1_BASE__INST0_SEG1                       0x0243FC00
+#define MP1_BASE__INST0_SEG2                       0x00DC0000
+#define MP1_BASE__INST0_SEG3                       0x00E00000
+#define MP1_BASE__INST0_SEG4                       0x00E40000
+#define MP1_BASE__INST0_SEG5                       0
+
+#ifdef BASE_INNER
+#undef BASE_INNER
+#endif
+
+#define BASE_INNER(seg) MP1_BASE__INST0_SEG ## seg
+
+#define BASE(seg) BASE_INNER(seg)
+
+#define REG(reg_name) (BASE(reg##reg_name##_BASE_IDX) + reg##reg_name)
+
+#define FN(reg_name, field) \
+	FD(reg_name##__##field)
+
+#include "logger_types.h"
+#undef DC_LOGGER
+#define DC_LOGGER \
+	CTX->logger
+#define smu_print(str, ...) {DC_LOG_SMU(str, ##__VA_ARGS__); }
+
+#define VBIOSSMC_MSG_TestMessage                  0x1
+#define VBIOSSMC_MSG_GetSmuVersion                0x2
+#define VBIOSSMC_MSG_PowerUpGfx                   0x3
+#define VBIOSSMC_MSG_SetDispclkFreq               0x4
+#define VBIOSSMC_MSG_SetDprefclkFreq              0x5   //Not used. DPRef is constant
+#define VBIOSSMC_MSG_SetDppclkFreq                0x6
+#define VBIOSSMC_MSG_SetHardMinDcfclkByFreq       0x7
+#define VBIOSSMC_MSG_SetMinDeepSleepDcfclk        0x8
+#define VBIOSSMC_MSG_SetPhyclkVoltageByFreq       0x9	//Keep it in case VMIN dees not support phy clk
+#define VBIOSSMC_MSG_GetFclkFrequency             0xA
+#define VBIOSSMC_MSG_SetDisplayCount              0xB   //Not used anymore
+#define VBIOSSMC_MSG_EnableTmdp48MHzRefclkPwrDown 0xC   //To ask PMFW turn off TMDP 48MHz refclk during display off to save power
+#define VBIOSSMC_MSG_UpdatePmeRestore             0xD
+#define VBIOSSMC_MSG_SetVbiosDramAddrHigh         0xE   //Used for WM table txfr
+#define VBIOSSMC_MSG_SetVbiosDramAddrLow          0xF
+#define VBIOSSMC_MSG_TransferTableSmu2Dram        0x10
+#define VBIOSSMC_MSG_TransferTableDram2Smu        0x11
+#define VBIOSSMC_MSG_SetDisplayIdleOptimizations  0x12
+#define VBIOSSMC_MSG_GetDprefclkFreq              0x13
+#define VBIOSSMC_MSG_GetDtbclkFreq                0x14
+#define VBIOSSMC_MSG_AllowZstatesEntry            0x15
+#define VBIOSSMC_MSG_DisallowZstatesEntry     	  0x16
+#define VBIOSSMC_MSG_SetDtbClk                    0x17
+#define VBIOSSMC_MSG_DispIPS2Entry                0x18 ///< Display IPS2 entry, DMU
+#define VBIOSSMC_MSG_DispIPS2Exit                 0x19 ///< Display IPS2 exit, DMU
+#define VBIOSSMC_MSG_DisableLSdma                 0x1A ///< Disable LSDMA; only sent by VBIOS
+#define VBIOSSMC_MSG_DpControllerPhyStatus        0x1B ///< Inform PMFW about the pre conditions for turning SLDO2 on/off . bit[0]==1 precondition is met, bit[1-2] are for DPPHY number
+#define VBIOSSMC_MSG_QueryIPS2Support             0x1C ///< Return 1: support; else not supported
+#define VBIOSSMC_MSG_NotifyHostRouterBW           0x1D
+#define VBIOSSMC_Message_Count                    0x1E
+
+#define VBIOSSMC_Status_BUSY                      0x0
+#define VBIOSSMC_Result_OK                        0x1
+#define VBIOSSMC_Result_Failed                    0xFF
+#define VBIOSSMC_Result_UnknownCmd                0xFE
+#define VBIOSSMC_Result_CmdRejectedPrereq         0xFD
+#define VBIOSSMC_Result_CmdRejectedBusy           0xFC
+
+union dcn35_dpia_host_router_bw {
+	struct {
+		uint32_t hr_id : 16;
+		uint32_t bw_mbps : 16;
+	} bits;
+	uint32_t all;
+};
+
+/*
+ * Function to be used instead of REG_WAIT macro because the wait ends when
+ * the register is NOT EQUAL to zero, and because `the translation in msg_if.h
+ * won't work with REG_WAIT.
+ */
+static uint32_t dcn35_smu_wait_for_response(struct clk_mgr_internal *clk_mgr, unsigned int delay_us, unsigned int max_retries)
+{
+	uint32_t res_val = VBIOSSMC_Status_BUSY;
+
+	do {
+		res_val = REG_READ(MP1_SMN_C2PMSG_91);
+		if (res_val != VBIOSSMC_Status_BUSY)
+			break;
+
+		if (delay_us >= 1000)
+			msleep(delay_us/1000);
+		else if (delay_us > 0)
+			udelay(delay_us);
+
+		if (clk_mgr->base.ctx->dc->debug.disable_timeout)
+			max_retries++;
+	} while (max_retries--);
+
+	return res_val;
+}
+
+static int dcn35_smu_send_msg_with_param(struct clk_mgr_internal *clk_mgr,
+					 unsigned int msg_id,
+					 unsigned int param)
+{
+	uint32_t result;
+
+	result = dcn35_smu_wait_for_response(clk_mgr, 10, 2000000);
+	ASSERT(result == VBIOSSMC_Result_OK);
+
+	if (result != VBIOSSMC_Result_OK) {
+		DC_LOG_WARNING("SMU response after wait: %d, msg id = %d\n", result, msg_id);
+
+		if (result == VBIOSSMC_Status_BUSY)
+			return -1;
+	}
+
+	/* First clear response register */
+	REG_WRITE(MP1_SMN_C2PMSG_91, VBIOSSMC_Status_BUSY);
+
+	/* Set the parameter register for the SMU message, unit is Mhz */
+	REG_WRITE(MP1_SMN_C2PMSG_83, param);
+
+	/* Trigger the message transaction by writing the message ID */
+	REG_WRITE(MP1_SMN_C2PMSG_67, msg_id);
+
+	result = dcn35_smu_wait_for_response(clk_mgr, 10, 2000000);
+
+	if (result == VBIOSSMC_Result_Failed) {
+		if (msg_id == VBIOSSMC_MSG_TransferTableDram2Smu &&
+		    param == TABLE_WATERMARKS)
+			DC_LOG_WARNING("Watermarks table not configured properly by SMU");
+		else
+			ASSERT(0);
+		REG_WRITE(MP1_SMN_C2PMSG_91, VBIOSSMC_Result_OK);
+		DC_LOG_WARNING("SMU response after wait: %d, msg id = %d\n", result, msg_id);
+		return -1;
+	}
+
+	if (IS_SMU_TIMEOUT(result)) {
+		ASSERT(0);
+		result = dcn35_smu_wait_for_response(clk_mgr, 10, 2000000);
+		//dm_helpers_smu_timeout(CTX, msg_id, param, 10 * 200000);
+		DC_LOG_WARNING("SMU response after wait: %d, msg id = %d\n", result, msg_id);
+	}
+
+	return REG_READ(MP1_SMN_C2PMSG_83);
+}
+
+int dcn35_smu_get_smu_version(struct clk_mgr_internal *clk_mgr)
+{
+	return dcn35_smu_send_msg_with_param(
+			clk_mgr,
+			VBIOSSMC_MSG_GetSmuVersion,
+			0);
+}
+
+
+int dcn35_smu_set_dispclk(struct clk_mgr_internal *clk_mgr, int requested_dispclk_khz)
+{
+	int actual_dispclk_set_mhz = -1;
+
+	if (!clk_mgr->smu_present)
+		return requested_dispclk_khz;
+
+	/*  Unit of SMU msg parameter is Mhz */
+	actual_dispclk_set_mhz = dcn35_smu_send_msg_with_param(
+			clk_mgr,
+			VBIOSSMC_MSG_SetDispclkFreq,
+			khz_to_mhz_ceil(requested_dispclk_khz));
+
+	smu_print("requested_dispclk_khz = %d, actual_dispclk_set_mhz: %d\n", requested_dispclk_khz, actual_dispclk_set_mhz);
+	return actual_dispclk_set_mhz * 1000;
+}
+
+int dcn35_smu_set_dprefclk(struct clk_mgr_internal *clk_mgr)
+{
+	int actual_dprefclk_set_mhz = -1;
+
+	if (!clk_mgr->smu_present)
+		return clk_mgr->base.dprefclk_khz;
+
+	actual_dprefclk_set_mhz = dcn35_smu_send_msg_with_param(
+			clk_mgr,
+			VBIOSSMC_MSG_SetDprefclkFreq,
+			khz_to_mhz_ceil(clk_mgr->base.dprefclk_khz));
+
+	/* TODO: add code for programing DP DTO, currently this is down by command table */
+
+	return actual_dprefclk_set_mhz * 1000;
+}
+
+int dcn35_smu_set_hard_min_dcfclk(struct clk_mgr_internal *clk_mgr, int requested_dcfclk_khz)
+{
+	int actual_dcfclk_set_mhz = -1;
+
+	if (!clk_mgr->smu_present)
+		return requested_dcfclk_khz;
+
+	actual_dcfclk_set_mhz = dcn35_smu_send_msg_with_param(
+			clk_mgr,
+			VBIOSSMC_MSG_SetHardMinDcfclkByFreq,
+			khz_to_mhz_ceil(requested_dcfclk_khz));
+
+	smu_print("requested_dcfclk_khz = %d, actual_dcfclk_set_mhz: %d\n", requested_dcfclk_khz, actual_dcfclk_set_mhz);
+
+	return actual_dcfclk_set_mhz * 1000;
+}
+
+int dcn35_smu_set_min_deep_sleep_dcfclk(struct clk_mgr_internal *clk_mgr, int requested_min_ds_dcfclk_khz)
+{
+	int actual_min_ds_dcfclk_mhz = -1;
+
+	if (!clk_mgr->smu_present)
+		return requested_min_ds_dcfclk_khz;
+
+	actual_min_ds_dcfclk_mhz = dcn35_smu_send_msg_with_param(
+			clk_mgr,
+			VBIOSSMC_MSG_SetMinDeepSleepDcfclk,
+			khz_to_mhz_ceil(requested_min_ds_dcfclk_khz));
+
+	smu_print("requested_min_ds_dcfclk_khz = %d, actual_min_ds_dcfclk_mhz: %d\n", requested_min_ds_dcfclk_khz, actual_min_ds_dcfclk_mhz);
+
+	return actual_min_ds_dcfclk_mhz * 1000;
+}
+
+int dcn35_smu_set_dppclk(struct clk_mgr_internal *clk_mgr, int requested_dpp_khz)
+{
+	int actual_dppclk_set_mhz = -1;
+
+	if (!clk_mgr->smu_present)
+		return requested_dpp_khz;
+
+	actual_dppclk_set_mhz = dcn35_smu_send_msg_with_param(
+			clk_mgr,
+			VBIOSSMC_MSG_SetDppclkFreq,
+			khz_to_mhz_ceil(requested_dpp_khz));
+
+	smu_print("requested_dpp_khz = %d, actual_dppclk_set_mhz: %d\n", requested_dpp_khz, actual_dppclk_set_mhz);
+
+	return actual_dppclk_set_mhz * 1000;
+}
+
+void dcn35_smu_set_display_idle_optimization(struct clk_mgr_internal *clk_mgr, uint32_t idle_info)
+{
+	if (!clk_mgr->base.ctx->dc->debug.pstate_enabled)
+		return;
+
+	if (!clk_mgr->smu_present)
+		return;
+
+	//TODO: Work with smu team to define optimization options.
+	dcn35_smu_send_msg_with_param(
+		clk_mgr,
+		VBIOSSMC_MSG_SetDisplayIdleOptimizations,
+		idle_info);
+	smu_print("%s: VBIOSSMC_MSG_SetDisplayIdleOptimizations idle_info  = %x\n", __func__, idle_info);
+}
+
+void dcn35_smu_enable_phy_refclk_pwrdwn(struct clk_mgr_internal *clk_mgr, bool enable)
+{
+	union display_idle_optimization_u idle_info = { 0 };
+
+	if (!clk_mgr->smu_present)
+		return;
+
+	if (enable) {
+		idle_info.idle_info.df_request_disabled = 1;
+		idle_info.idle_info.phy_ref_clk_off = 1;
+	}
+
+	dcn35_smu_send_msg_with_param(
+			clk_mgr,
+			VBIOSSMC_MSG_SetDisplayIdleOptimizations,
+			idle_info.data);
+	smu_print("%s smu_enable_phy_refclk_pwrdwn  = %d\n", __func__, enable ? 1 : 0);
+}
+
+void dcn35_smu_enable_pme_wa(struct clk_mgr_internal *clk_mgr)
+{
+	if (!clk_mgr->smu_present)
+		return;
+
+	dcn35_smu_send_msg_with_param(
+			clk_mgr,
+			VBIOSSMC_MSG_UpdatePmeRestore,
+			0);
+	smu_print("%s: SMC_MSG_UpdatePmeRestore\n", __func__);
+}
+
+void dcn35_smu_set_dram_addr_high(struct clk_mgr_internal *clk_mgr, uint32_t addr_high)
+{
+	if (!clk_mgr->smu_present)
+		return;
+
+	dcn35_smu_send_msg_with_param(clk_mgr,
+			VBIOSSMC_MSG_SetVbiosDramAddrHigh, addr_high);
+}
+
+void dcn35_smu_set_dram_addr_low(struct clk_mgr_internal *clk_mgr, uint32_t addr_low)
+{
+	if (!clk_mgr->smu_present)
+		return;
+
+	dcn35_smu_send_msg_with_param(clk_mgr,
+			VBIOSSMC_MSG_SetVbiosDramAddrLow, addr_low);
+}
+
+void dcn35_smu_transfer_dpm_table_smu_2_dram(struct clk_mgr_internal *clk_mgr)
+{
+	if (!clk_mgr->smu_present)
+		return;
+
+	dcn35_smu_send_msg_with_param(clk_mgr,
+			VBIOSSMC_MSG_TransferTableSmu2Dram, TABLE_DPMCLOCKS);
+}
+
+void dcn35_smu_transfer_wm_table_dram_2_smu(struct clk_mgr_internal *clk_mgr)
+{
+	if (!clk_mgr->smu_present)
+		return;
+
+	dcn35_smu_send_msg_with_param(clk_mgr,
+			VBIOSSMC_MSG_TransferTableDram2Smu, TABLE_WATERMARKS);
+}
+
+void dcn35_smu_set_zstate_support(struct clk_mgr_internal *clk_mgr, enum dcn_zstate_support_state support)
+{
+	unsigned int msg_id, param, retv;
+
+	if (!clk_mgr->smu_present)
+		return;
+
+	switch (support) {
+
+	case DCN_ZSTATE_SUPPORT_ALLOW:
+		msg_id = VBIOSSMC_MSG_AllowZstatesEntry;
+		param = (1 << 10) | (1 << 9) | (1 << 8);
+		smu_print("%s: SMC_MSG_AllowZstatesEntry msg = ALLOW, param = 0x%x\n", __func__, param);
+		break;
+
+	case DCN_ZSTATE_SUPPORT_DISALLOW:
+		msg_id = VBIOSSMC_MSG_AllowZstatesEntry;
+		param = 0;
+		smu_print("%s: SMC_MSG_AllowZstatesEntry msg_id = DISALLOW, param = 0x%x\n",  __func__, param);
+		break;
+
+
+	case DCN_ZSTATE_SUPPORT_ALLOW_Z10_ONLY:
+		msg_id = VBIOSSMC_MSG_AllowZstatesEntry;
+		param = (1 << 10);
+		smu_print("%s: SMC_MSG_AllowZstatesEntry msg = ALLOW_Z10_ONLY, param = 0x%x\n", __func__, param);
+		break;
+
+	case DCN_ZSTATE_SUPPORT_ALLOW_Z8_Z10_ONLY:
+		msg_id = VBIOSSMC_MSG_AllowZstatesEntry;
+		param = (1 << 10) | (1 << 8);
+		smu_print("%s: SMC_MSG_AllowZstatesEntry msg = ALLOW_Z8_Z10_ONLY, param = 0x%x\n", __func__, param);
+		break;
+
+	case DCN_ZSTATE_SUPPORT_ALLOW_Z8_ONLY:
+		msg_id = VBIOSSMC_MSG_AllowZstatesEntry;
+		param = (1 << 8);
+		smu_print("%s: SMC_MSG_AllowZstatesEntry msg = ALLOW_Z8_ONLY, param = 0x%x\n", __func__, param);
+		break;
+
+	default: //DCN_ZSTATE_SUPPORT_UNKNOWN
+		msg_id = VBIOSSMC_MSG_AllowZstatesEntry;
+		param = 0;
+		break;
+	}
+
+
+	retv = dcn35_smu_send_msg_with_param(
+		clk_mgr,
+		msg_id,
+		param);
+	smu_print("%s:  msg_id = %d, param = 0x%x, return = 0x%x\n", __func__, msg_id, param, retv);
+}
+
+int dcn35_smu_get_dprefclk(struct clk_mgr_internal *clk_mgr)
+{
+	int dprefclk;
+
+	if (!clk_mgr->smu_present)
+		return 0;
+
+	dprefclk = dcn35_smu_send_msg_with_param(clk_mgr,
+						 VBIOSSMC_MSG_GetDprefclkFreq,
+						 0);
+
+	smu_print("%s:  SMU DPREF clk  = %d mhz\n",  __func__, dprefclk);
+	return dprefclk * 1000;
+}
+
+int dcn35_smu_get_dtbclk(struct clk_mgr_internal *clk_mgr)
+{
+	int dtbclk;
+
+	if (!clk_mgr->smu_present)
+		return 0;
+
+	dtbclk = dcn35_smu_send_msg_with_param(clk_mgr,
+					       VBIOSSMC_MSG_GetDtbclkFreq,
+					       0);
+
+	smu_print("%s: get_dtbclk  = %dmhz\n", __func__, dtbclk);
+	return dtbclk * 1000;
+}
+/* Arg = 1: Turn DTB on; 0: Turn DTB CLK OFF. when it is on, it is 600MHZ */
+void dcn35_smu_set_dtbclk(struct clk_mgr_internal *clk_mgr, bool enable)
+{
+	if (!clk_mgr->smu_present)
+		return;
+
+	dcn35_smu_send_msg_with_param(
+			clk_mgr,
+			VBIOSSMC_MSG_SetDtbClk,
+			enable);
+	smu_print("%s: smu_set_dtbclk = %d\n", __func__, enable ? 1 : 0);
+}
+
+void dcn35_vbios_smu_enable_48mhz_tmdp_refclk_pwrdwn(struct clk_mgr_internal *clk_mgr, bool enable)
+{
+	if (!clk_mgr->smu_present)
+		return;
+
+	dcn35_smu_send_msg_with_param(
+			clk_mgr,
+			VBIOSSMC_MSG_EnableTmdp48MHzRefclkPwrDown,
+			enable);
+	smu_print("%s: smu_enable_48mhz_tmdp_refclk_pwrdwn = %d\n", __func__, enable ? 1 : 0);
+}
+
+int dcn35_smu_exit_low_power_state(struct clk_mgr_internal *clk_mgr)
+{
+	int retv;
+
+	if (!clk_mgr->smu_present)
+		return 0;
+
+	retv = dcn35_smu_send_msg_with_param(
+		clk_mgr,
+		VBIOSSMC_MSG_DispIPS2Exit,
+		0);
+	smu_print("%s: smu_exit_low_power_state return = %d\n", __func__, retv);
+	return retv;
+}
+
+int dcn35_smu_get_ips_supported(struct clk_mgr_internal *clk_mgr)
+{
+	int retv;
+
+	if (!clk_mgr->smu_present)
+		return 0;
+
+	retv = dcn35_smu_send_msg_with_param(
+			clk_mgr,
+			VBIOSSMC_MSG_QueryIPS2Support,
+			0);
+
+	//smu_print("%s: VBIOSSMC_MSG_QueryIPS2Support return = %x\n", __func__, retv);
+	return retv;
+}
+
+void dcn35_smu_notify_host_router_bw(struct clk_mgr_internal *clk_mgr, uint32_t hr_id, uint32_t bw_kbps)
+{
+	union dcn35_dpia_host_router_bw msg_data = { 0 };
+
+	msg_data.bits.hr_id = hr_id;
+	msg_data.bits.bw_mbps = bw_kbps / 1000;
+
+	dcn35_smu_send_msg_with_param(clk_mgr, VBIOSSMC_MSG_NotifyHostRouterBW, msg_data.all);
+}
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_smu.h b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_smu.h
new file mode 100644
index 000000000000..ab9d21ba0c43
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_smu.h
@@ -0,0 +1,220 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef DAL_DC_35_SMU_H_
+#define DAL_DC_35_SMU_H_
+
+#include "os_types.h"
+
+#ifndef PMFW_DRIVER_IF_H
+#define PMFW_DRIVER_IF_H
+#define PMFW_DRIVER_IF_VERSION 4
+
+typedef enum {
+  DSPCLK_DCFCLK = 0,
+  DSPCLK_DISPCLK,
+  DSPCLK_PIXCLK,
+  DSPCLK_PHYCLK,
+  DSPCLK_COUNT,
+} DSPCLK_e;
+
+typedef struct {
+  uint16_t Freq; // in MHz
+  uint16_t Vid;  // min voltage in SVI3 VID
+} DisplayClockTable_t;
+
+typedef struct {
+  uint16_t MinClock; // This is either DCFCLK or SOCCLK (in MHz)
+  uint16_t MaxClock; // This is either DCFCLK or SOCCLK (in MHz)
+  uint16_t MinMclk;
+  uint16_t MaxMclk;
+
+  uint8_t  WmSetting;
+  uint8_t  WmType;  // Used for normal pstate change or memory retraining
+  uint8_t  Padding[2];
+} WatermarkRowGeneric_t;
+
+#define NUM_WM_RANGES 4
+#define WM_PSTATE_CHG 0
+#define WM_RETRAINING 1
+
+typedef enum {
+  WM_SOCCLK = 0,
+  WM_DCFCLK,
+  WM_COUNT,
+} WM_CLOCK_e;
+
+typedef struct {
+  // Watermarks
+  WatermarkRowGeneric_t WatermarkRow[WM_COUNT][NUM_WM_RANGES];
+
+  uint32_t MmHubPadding[7]; // SMU internal use
+} Watermarks_t;
+
+#define NUM_DCFCLK_DPM_LEVELS   8
+#define NUM_DISPCLK_DPM_LEVELS  8
+#define NUM_DPPCLK_DPM_LEVELS   8
+#define NUM_SOCCLK_DPM_LEVELS   8
+#define NUM_VCN_DPM_LEVELS      8
+#define NUM_SOC_VOLTAGE_LEVELS  8
+#define NUM_VPE_DPM_LEVELS        8
+#define NUM_FCLK_DPM_LEVELS       8
+#define NUM_MEM_PSTATE_LEVELS     4
+
+typedef enum{
+  WCK_RATIO_1_1 = 0,  // DDR5, Wck:ck is always 1:1;
+  WCK_RATIO_1_2,
+  WCK_RATIO_1_4,
+  WCK_RATIO_MAX
+} WCK_RATIO_e;
+
+typedef struct {
+  uint32_t UClk;
+  uint32_t MemClk;
+  uint32_t Voltage;
+  uint8_t  WckRatio;
+  uint8_t  Spare[3];
+} MemPstateTable_t;
+
+//Freq in MHz
+//Voltage in milli volts with 2 fractional bits
+typedef struct {
+  uint32_t DcfClocks[NUM_DCFCLK_DPM_LEVELS];
+  uint32_t DispClocks[NUM_DISPCLK_DPM_LEVELS];
+  uint32_t DppClocks[NUM_DPPCLK_DPM_LEVELS];
+  uint32_t SocClocks[NUM_SOCCLK_DPM_LEVELS];
+  uint32_t VClocks[NUM_VCN_DPM_LEVELS];
+  uint32_t DClocks[NUM_VCN_DPM_LEVELS];
+  uint32_t VPEClocks[NUM_VPE_DPM_LEVELS];
+  uint32_t FclkClocks_Freq[NUM_FCLK_DPM_LEVELS];
+  uint32_t FclkClocks_Voltage[NUM_FCLK_DPM_LEVELS];
+  uint32_t SocVoltage[NUM_SOC_VOLTAGE_LEVELS];
+  MemPstateTable_t MemPstateTable[NUM_MEM_PSTATE_LEVELS];
+
+  uint8_t  NumDcfClkLevelsEnabled;
+  uint8_t  NumDispClkLevelsEnabled; //Applies to both Dispclk and Dppclk
+  uint8_t  NumSocClkLevelsEnabled;
+  uint8_t  VcnClkLevelsEnabled;     //Applies to both Vclk and Dclk
+  uint8_t  VpeClkLevelsEnabled;
+  uint8_t  NumMemPstatesEnabled;
+  uint8_t  NumFclkLevelsEnabled;
+  uint8_t  spare[2];
+
+  uint32_t MinGfxClk;
+  uint32_t MaxGfxClk;
+} DpmClocks_t_dcn35;
+
+typedef struct {
+	uint32_t DcfClocks[NUM_DCFCLK_DPM_LEVELS];
+	uint32_t DispClocks[NUM_DISPCLK_DPM_LEVELS];
+	uint32_t DppClocks[NUM_DPPCLK_DPM_LEVELS];
+	uint32_t SocClocks[NUM_SOCCLK_DPM_LEVELS];
+	uint32_t VClocks0[NUM_VCN_DPM_LEVELS];
+	uint32_t VClocks1[NUM_VCN_DPM_LEVELS];
+	uint32_t DClocks0[NUM_VCN_DPM_LEVELS];
+	uint32_t DClocks1[NUM_VCN_DPM_LEVELS];
+	uint32_t VPEClocks[NUM_VPE_DPM_LEVELS];
+	uint32_t FclkClocks_Freq[NUM_FCLK_DPM_LEVELS];
+	uint32_t FclkClocks_Voltage[NUM_FCLK_DPM_LEVELS];
+	uint32_t SocVoltage[NUM_SOC_VOLTAGE_LEVELS];
+	MemPstateTable_t MemPstateTable[NUM_MEM_PSTATE_LEVELS];
+	uint8_t NumDcfClkLevelsEnabled;
+	uint8_t NumDispClkLevelsEnabled; // Applies to both Dispclk and Dppclk
+	uint8_t NumSocClkLevelsEnabled;
+	uint8_t Vcn0ClkLevelsEnabled; // Applies to both Vclk0 and Dclk0
+	uint8_t Vcn1ClkLevelsEnabled; // Applies to both Vclk1 and Dclk1
+	uint8_t VpeClkLevelsEnabled;
+	uint8_t NumMemPstatesEnabled;
+	uint8_t NumFclkLevelsEnabled;
+	uint32_t MinGfxClk;
+	uint32_t MaxGfxClk;
+} DpmClocks_t_dcn351;
+
+#define TABLE_BIOS_IF            0 // Called by BIOS
+#define TABLE_WATERMARKS         1 // Called by DAL through VBIOS
+#define TABLE_CUSTOM_DPM         2 // Called by Driver
+#define TABLE_SPARE1             3
+#define TABLE_DPMCLOCKS          4 // Called by Driver
+#define TABLE_MOMENTARY_PM       5 // Called by Tools
+#define TABLE_MODERN_STDBY       6 // Called by Tools for Modern Standby Log
+#define TABLE_SMU_METRICS        7 // Called by Driver
+#define TABLE_COUNT              8
+
+#endif
+
+struct dcn35_watermarks {
+  // Watermarks
+  WatermarkRowGeneric_t WatermarkRow[WM_COUNT][NUM_WM_RANGES];
+
+  uint32_t MmHubPadding[7]; // SMU internal use
+};
+
+struct dcn35_smu_dpm_clks {
+	DpmClocks_t_dcn35 *dpm_clks;
+	union large_integer mc_address;
+};
+
+struct dcn351_smu_dpm_clks {
+	DpmClocks_t_dcn351 *dpm_clks;
+	union large_integer mc_address;
+};
+/* TODO: taken from vgh, may not be correct */
+struct display_idle_optimization {
+	unsigned int df_request_disabled : 1;
+	unsigned int phy_ref_clk_off     : 1;
+	unsigned int s0i2_rdy            : 1;
+	unsigned int reserved            : 29;
+};
+
+union display_idle_optimization_u {
+	struct display_idle_optimization idle_info;
+	uint32_t data;
+};
+
+int dcn35_smu_get_smu_version(struct clk_mgr_internal *clk_mgr);
+int dcn35_smu_set_dispclk(struct clk_mgr_internal *clk_mgr, int requested_dispclk_khz);
+int dcn35_smu_set_dprefclk(struct clk_mgr_internal *clk_mgr);
+int dcn35_smu_set_hard_min_dcfclk(struct clk_mgr_internal *clk_mgr, int requested_dcfclk_khz);
+int dcn35_smu_set_min_deep_sleep_dcfclk(struct clk_mgr_internal *clk_mgr, int requested_min_ds_dcfclk_khz);
+int dcn35_smu_set_dppclk(struct clk_mgr_internal *clk_mgr, int requested_dpp_khz);
+void dcn35_smu_set_display_idle_optimization(struct clk_mgr_internal *clk_mgr, uint32_t idle_info);
+void dcn35_smu_enable_phy_refclk_pwrdwn(struct clk_mgr_internal *clk_mgr, bool enable);
+void dcn35_smu_enable_pme_wa(struct clk_mgr_internal *clk_mgr);
+void dcn35_smu_set_dram_addr_high(struct clk_mgr_internal *clk_mgr, uint32_t addr_high);
+void dcn35_smu_set_dram_addr_low(struct clk_mgr_internal *clk_mgr, uint32_t addr_low);
+void dcn35_smu_transfer_dpm_table_smu_2_dram(struct clk_mgr_internal *clk_mgr);
+void dcn35_smu_transfer_wm_table_dram_2_smu(struct clk_mgr_internal *clk_mgr);
+
+void dcn35_smu_set_zstate_support(struct clk_mgr_internal *clk_mgr, enum dcn_zstate_support_state support);
+void dcn35_smu_set_dtbclk(struct clk_mgr_internal *clk_mgr, bool enable);
+void dcn35_vbios_smu_enable_48mhz_tmdp_refclk_pwrdwn(struct clk_mgr_internal *clk_mgr, bool enable);
+
+int dcn35_smu_exit_low_power_state(struct clk_mgr_internal *clk_mgr);
+int dcn35_smu_get_ips_supported(struct clk_mgr_internal *clk_mgr);
+int dcn35_smu_get_dtbclk(struct clk_mgr_internal *clk_mgr);
+int dcn35_smu_get_dprefclk(struct clk_mgr_internal *clk_mgr);
+void dcn35_smu_notify_host_router_bw(struct clk_mgr_internal *clk_mgr, uint32_t hr_id, uint32_t bw_kbps);
+
+#endif /* DAL_DC_35_SMU_H_ */
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dalsmc.h b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dalsmc.h
new file mode 100644
index 000000000000..2e0d34fd7512
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dalsmc.h
@@ -0,0 +1,55 @@
+// SPDX-License-Identifier: MIT
+//
+// Copyright 2024 Advanced Micro Devices, Inc.
+
+#ifndef DALSMC_H
+#define DALSMC_H
+
+#define DALSMC_VERSION 0x1
+
+// SMU Response Codes:
+#define DALSMC_Result_OK                   0x1
+#define DALSMC_Result_Failed               0xFF
+#define DALSMC_Result_UnknownCmd           0xFE
+#define DALSMC_Result_CmdRejectedPrereq    0xFD
+#define DALSMC_Result_CmdRejectedBusy      0xFC
+
+
+
+// Message Definitions:
+#define DALSMC_MSG_TestMessage                    0x1
+#define DALSMC_MSG_GetSmuVersion                  0x2
+#define DALSMC_MSG_GetDriverIfVersion             0x3
+#define DALSMC_MSG_GetMsgHeaderVersion            0x4
+#define DALSMC_MSG_SetDalDramAddrHigh             0x5
+#define DALSMC_MSG_SetDalDramAddrLow              0x6
+#define DALSMC_MSG_TransferTableSmu2Dram          0x7
+#define DALSMC_MSG_TransferTableDram2Smu          0x8
+#define DALSMC_MSG_SetHardMinByFreq               0x9
+#define DALSMC_MSG_SetHardMaxByFreq               0xA
+#define DALSMC_MSG_GetDpmFreqByIndex              0xB
+#define DALSMC_MSG_GetDcModeMaxDpmFreq            0xC
+#define DALSMC_MSG_SetMinDeepSleepDcfclk          0xD
+#define DALSMC_MSG_NumOfDisplays                  0xE
+#define DALSMC_MSG_SetExternalClientDfCstateAllow 0xF
+#define DALSMC_MSG_BacoAudioD3PME                 0x10
+#define DALSMC_MSG_SetFclkSwitchAllow             0x11
+#define DALSMC_MSG_SetCabForUclkPstate            0x12
+#define DALSMC_MSG_SetWorstCaseUclkLatency        0x13
+#define DALSMC_MSG_DcnExitReset                   0x14
+#define DALSMC_MSG_ReturnHardMinStatus            0x15
+#define DALSMC_MSG_SetAlwaysWaitDmcubResp         0x16
+#define DALSMC_MSG_IndicateDrrStatus              0x17  // PMFW 15811
+#define DALSMC_MSG_ActiveUclkFclk                 0x18
+#define DALSMC_MSG_IdleUclkFclk                   0x19
+#define DALSMC_MSG_SetUclkPstateAllow             0x1A
+#define DALSMC_MSG_SubvpUclkFclk                  0x1B
+#define DALSMC_MSG_GetNumUmcChannels              0x1C
+#define DALSMC_Message_Count                      0x1D
+
+typedef enum {
+  FCLK_SWITCH_DISALLOW,
+  FCLK_SWITCH_ALLOW,
+} FclkSwitchAllow_e;
+
+#endif
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr.c
new file mode 100644
index 000000000000..306016c1f109
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr.c
@@ -0,0 +1,1631 @@
+// SPDX-License-Identifier: MIT
+//
+// Copyright 2024 Advanced Micro Devices, Inc.
+
+#include "dccg.h"
+#include "clk_mgr_internal.h"
+#include "dcn401/dcn401_clk_mgr_smu_msg.h"
+#include "dcn20/dcn20_clk_mgr.h"
+#include "dce100/dce_clk_mgr.h"
+#include "dcn31/dcn31_clk_mgr.h"
+#include "dcn32/dcn32_clk_mgr.h"
+#include "dcn401/dcn401_clk_mgr.h"
+#include "reg_helper.h"
+#include "core_types.h"
+#include "dm_helpers.h"
+#include "link_service.h"
+#include "dc_state_priv.h"
+#include "atomfirmware.h"
+
+#include "dcn401_smu14_driver_if.h"
+
+#include "dcn/dcn_4_1_0_offset.h"
+#include "dcn/dcn_4_1_0_sh_mask.h"
+
+#define DCN_BASE__INST0_SEG1                       0x000000C0
+
+#define mmCLK01_CLK0_CLK_PLL_REQ                        0x16E37
+#define mmCLK01_CLK0_CLK0_DFS_CNTL                      0x16E69
+#define mmCLK01_CLK0_CLK1_DFS_CNTL                      0x16E6C
+#define mmCLK01_CLK0_CLK2_DFS_CNTL                      0x16E6F
+#define mmCLK01_CLK0_CLK3_DFS_CNTL                      0x16E72
+#define mmCLK01_CLK0_CLK4_DFS_CNTL                      0x16E75
+#define mmCLK20_CLK2_CLK2_DFS_CNTL                      0x1B051
+
+#define CLK0_CLK_PLL_REQ__FbMult_int_MASK                  0x000001ffUL
+#define CLK0_CLK_PLL_REQ__PllSpineDiv_MASK                 0x0000f000UL
+#define CLK0_CLK_PLL_REQ__FbMult_frac_MASK                 0xffff0000UL
+#define CLK0_CLK_PLL_REQ__FbMult_int__SHIFT                0x00000000
+#define CLK0_CLK_PLL_REQ__PllSpineDiv__SHIFT               0x0000000c
+#define CLK0_CLK_PLL_REQ__FbMult_frac__SHIFT               0x00000010
+
+#undef FN
+#define FN(reg_name, field_name) \
+	clk_mgr->clk_mgr_shift->field_name, clk_mgr->clk_mgr_mask->field_name
+
+#define REG(reg) \
+	(clk_mgr->regs->reg)
+
+#define BASE_INNER(seg) DCN_BASE__INST0_SEG ## seg
+
+#define BASE(seg) BASE_INNER(seg)
+
+#define SR(reg_name)\
+		.reg_name = BASE(reg ## reg_name ## _BASE_IDX) +  \
+					reg ## reg_name
+
+#define CLK_SR_DCN401(reg_name, block, inst)\
+	.reg_name = mm ## block ## _ ## reg_name
+
+static const struct clk_mgr_registers clk_mgr_regs_dcn401 = {
+	CLK_REG_LIST_DCN401()
+};
+
+static const struct clk_mgr_shift clk_mgr_shift_dcn401 = {
+	CLK_COMMON_MASK_SH_LIST_DCN401(__SHIFT)
+};
+
+static const struct clk_mgr_mask clk_mgr_mask_dcn401 = {
+	CLK_COMMON_MASK_SH_LIST_DCN401(_MASK)
+};
+
+#define TO_DCN401_CLK_MGR(clk_mgr)\
+	container_of(clk_mgr, struct dcn401_clk_mgr, base)
+
+static bool dcn401_is_ppclk_dpm_enabled(struct clk_mgr_internal *clk_mgr, PPCLK_e clk)
+{
+	bool ppclk_dpm_enabled = false;
+
+	switch (clk) {
+	case PPCLK_SOCCLK:
+		ppclk_dpm_enabled =
+				clk_mgr->base.bw_params->clk_table.num_entries_per_clk.num_socclk_levels > 1;
+		break;
+	case PPCLK_UCLK:
+		ppclk_dpm_enabled =
+				clk_mgr->base.bw_params->clk_table.num_entries_per_clk.num_memclk_levels > 1;
+		break;
+	case PPCLK_FCLK:
+		ppclk_dpm_enabled =
+				clk_mgr->base.bw_params->clk_table.num_entries_per_clk.num_fclk_levels > 1;
+		break;
+	case PPCLK_DISPCLK:
+		ppclk_dpm_enabled =
+				clk_mgr->base.bw_params->clk_table.num_entries_per_clk.num_dispclk_levels > 1;
+		break;
+	case PPCLK_DPPCLK:
+		ppclk_dpm_enabled =
+				clk_mgr->base.bw_params->clk_table.num_entries_per_clk.num_dppclk_levels > 1;
+		break;
+	case PPCLK_DPREFCLK:
+		ppclk_dpm_enabled = false;
+		break;
+	case PPCLK_DCFCLK:
+		ppclk_dpm_enabled =
+				clk_mgr->base.bw_params->clk_table.num_entries_per_clk.num_dcfclk_levels > 1;
+		break;
+	case PPCLK_DTBCLK:
+		ppclk_dpm_enabled =
+				clk_mgr->base.bw_params->clk_table.num_entries_per_clk.num_dtbclk_levels > 1;
+		break;
+	default:
+		ppclk_dpm_enabled = false;
+	}
+
+	ppclk_dpm_enabled &= clk_mgr->smu_present;
+
+	return ppclk_dpm_enabled;
+}
+
+static bool dcn401_is_ppclk_idle_dpm_enabled(struct clk_mgr_internal *clk_mgr, PPCLK_e clk)
+{
+	bool ppclk_idle_dpm_enabled = false;
+
+	switch (clk) {
+	case PPCLK_UCLK:
+	case PPCLK_FCLK:
+		if (ASICREV_IS_GC_12_0_0_A0(clk_mgr->base.ctx->asic_id.hw_internal_rev) &&
+				clk_mgr->smu_ver >= 0x681800) {
+			ppclk_idle_dpm_enabled = true;
+		} else if (ASICREV_IS_GC_12_0_1_A0(clk_mgr->base.ctx->asic_id.hw_internal_rev) &&
+				clk_mgr->smu_ver >= 0x661300) {
+			ppclk_idle_dpm_enabled = true;
+		}
+		break;
+	default:
+		ppclk_idle_dpm_enabled = false;
+	}
+
+	ppclk_idle_dpm_enabled &= clk_mgr->smu_present;
+
+	return ppclk_idle_dpm_enabled;
+}
+
+static bool dcn401_is_df_throttle_opt_enabled(struct clk_mgr_internal *clk_mgr)
+{
+	bool is_df_throttle_opt_enabled = false;
+
+	if (ASICREV_IS_GC_12_0_1_A0(clk_mgr->base.ctx->asic_id.hw_internal_rev) &&
+			clk_mgr->smu_ver >= 0x663500) {
+		is_df_throttle_opt_enabled = !clk_mgr->base.ctx->dc->debug.force_subvp_df_throttle;
+	}
+
+	is_df_throttle_opt_enabled &= clk_mgr->smu_present;
+
+	return is_df_throttle_opt_enabled;
+}
+
+/* Query SMU for all clock states for a particular clock */
+static void dcn401_init_single_clock(struct clk_mgr_internal *clk_mgr, PPCLK_e clk, unsigned int *entry_0,
+		unsigned int *num_levels)
+{
+	unsigned int i;
+	char *entry_i = (char *)entry_0;
+
+	uint32_t ret = dcn401_smu_get_dpm_freq_by_index(clk_mgr, clk, 0xFF);
+
+	if (ret & (1 << 31))
+		/* fine-grained, only min and max */
+		*num_levels = 2;
+	else
+		/* discrete, a number of fixed states */
+		/* will set num_levels to 0 on failure */
+		*num_levels = ret & 0xFF;
+
+	/* if the initial message failed, num_levels will be 0 */
+	for (i = 0; i < *num_levels && i < ARRAY_SIZE(clk_mgr->base.bw_params->clk_table.entries); i++) {
+		*((unsigned int *)entry_i) = (dcn401_smu_get_dpm_freq_by_index(clk_mgr, clk, i) & 0xFFFF);
+		entry_i += sizeof(clk_mgr->base.bw_params->clk_table.entries[0]);
+	}
+}
+
+static void dcn401_build_wm_range_table(struct clk_mgr *clk_mgr)
+{
+	/* For min clocks use as reported by PM FW and report those as min */
+	uint16_t min_uclk_mhz = clk_mgr->bw_params->clk_table.entries[0].memclk_mhz;
+	uint16_t min_dcfclk_mhz	= clk_mgr->bw_params->clk_table.entries[0].dcfclk_mhz;
+
+	/* Set A - Normal - default values */
+	clk_mgr->bw_params->wm_table.nv_entries[WM_A].valid = true;
+	clk_mgr->bw_params->wm_table.nv_entries[WM_A].pmfw_breakdown.wm_type = WATERMARKS_CLOCK_RANGE;
+	clk_mgr->bw_params->wm_table.nv_entries[WM_A].pmfw_breakdown.min_dcfclk = min_dcfclk_mhz;
+	clk_mgr->bw_params->wm_table.nv_entries[WM_A].pmfw_breakdown.max_dcfclk = 0xFFFF;
+	clk_mgr->bw_params->wm_table.nv_entries[WM_A].pmfw_breakdown.min_uclk = min_uclk_mhz;
+	clk_mgr->bw_params->wm_table.nv_entries[WM_A].pmfw_breakdown.max_uclk = 0xFFFF;
+
+	/* Set B - Unused on dcn4 */
+	clk_mgr->bw_params->wm_table.nv_entries[WM_B].valid = false;
+
+	/* Set 1A - Dummy P-State - P-State latency set to "dummy p-state" value */
+	/* 'DalDummyClockChangeLatencyNs' registry key option set to 0x7FFFFFFF can be used to disable Set C for dummy p-state */
+	if (clk_mgr->ctx->dc->bb_overrides.dummy_clock_change_latency_ns != 0x7FFFFFFF) {
+		clk_mgr->bw_params->wm_table.nv_entries[WM_1A].valid = true;
+		clk_mgr->bw_params->wm_table.nv_entries[WM_1A].pmfw_breakdown.wm_type = WATERMARKS_DUMMY_PSTATE;
+		clk_mgr->bw_params->wm_table.nv_entries[WM_1A].pmfw_breakdown.min_dcfclk = min_dcfclk_mhz;
+		clk_mgr->bw_params->wm_table.nv_entries[WM_1A].pmfw_breakdown.max_dcfclk = 0xFFFF;
+		clk_mgr->bw_params->wm_table.nv_entries[WM_1A].pmfw_breakdown.min_uclk = min_uclk_mhz;
+		clk_mgr->bw_params->wm_table.nv_entries[WM_1A].pmfw_breakdown.max_uclk = 0xFFFF;
+	} else {
+		clk_mgr->bw_params->wm_table.nv_entries[WM_1A].valid = false;
+	}
+
+	/* Set 1B - Unused on dcn4 */
+	clk_mgr->bw_params->wm_table.nv_entries[WM_1B].valid = false;
+}
+
+void dcn401_init_clocks(struct clk_mgr *clk_mgr_base)
+{
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+	struct clk_limit_num_entries *num_entries_per_clk;
+	unsigned int i;
+
+	if (!clk_mgr_base->bw_params)
+		return;
+
+	num_entries_per_clk = &clk_mgr_base->bw_params->clk_table.num_entries_per_clk;
+
+	memset(&(clk_mgr_base->clks), 0, sizeof(struct dc_clocks));
+	clk_mgr_base->clks.p_state_change_support = true;
+	clk_mgr_base->clks.prev_p_state_change_support = true;
+	clk_mgr_base->clks.fclk_prev_p_state_change_support = true;
+	clk_mgr->smu_present = false;
+	clk_mgr->dpm_present = false;
+
+	if (!clk_mgr_base->force_smu_not_present && dcn401_smu_get_smu_version(clk_mgr, &clk_mgr->smu_ver))
+		clk_mgr->smu_present = true;
+
+	if (!clk_mgr->smu_present)
+		return;
+
+	dcn401_smu_check_driver_if_version(clk_mgr);
+	dcn401_smu_check_msg_header_version(clk_mgr);
+
+	/* DCFCLK */
+	dcn401_init_single_clock(clk_mgr, PPCLK_DCFCLK,
+			&clk_mgr_base->bw_params->clk_table.entries[0].dcfclk_mhz,
+			&num_entries_per_clk->num_dcfclk_levels);
+	clk_mgr_base->bw_params->dc_mode_limit.dcfclk_mhz = dcn401_smu_get_dc_mode_max_dpm_freq(clk_mgr, PPCLK_DCFCLK);
+	if (num_entries_per_clk->num_dcfclk_levels && clk_mgr_base->bw_params->dc_mode_limit.dcfclk_mhz ==
+			clk_mgr_base->bw_params->clk_table.entries[num_entries_per_clk->num_dcfclk_levels - 1].dcfclk_mhz)
+		clk_mgr_base->bw_params->dc_mode_limit.dcfclk_mhz = 0;
+
+	/* SOCCLK */
+	dcn401_init_single_clock(clk_mgr, PPCLK_SOCCLK,
+					&clk_mgr_base->bw_params->clk_table.entries[0].socclk_mhz,
+					&num_entries_per_clk->num_socclk_levels);
+	clk_mgr_base->bw_params->dc_mode_limit.socclk_mhz = dcn401_smu_get_dc_mode_max_dpm_freq(clk_mgr, PPCLK_SOCCLK);
+	if (num_entries_per_clk->num_socclk_levels && clk_mgr_base->bw_params->dc_mode_limit.socclk_mhz ==
+			clk_mgr_base->bw_params->clk_table.entries[num_entries_per_clk->num_socclk_levels - 1].socclk_mhz)
+		clk_mgr_base->bw_params->dc_mode_limit.socclk_mhz = 0;
+
+	/* DTBCLK */
+	if (!clk_mgr->base.ctx->dc->debug.disable_dtb_ref_clk_switch) {
+		dcn401_init_single_clock(clk_mgr, PPCLK_DTBCLK,
+				&clk_mgr_base->bw_params->clk_table.entries[0].dtbclk_mhz,
+				&num_entries_per_clk->num_dtbclk_levels);
+		clk_mgr_base->bw_params->dc_mode_limit.dtbclk_mhz = dcn401_smu_get_dc_mode_max_dpm_freq(clk_mgr, PPCLK_DTBCLK);
+		if (num_entries_per_clk->num_dtbclk_levels && clk_mgr_base->bw_params->dc_mode_limit.dtbclk_mhz ==
+				clk_mgr_base->bw_params->clk_table.entries[num_entries_per_clk->num_dtbclk_levels - 1].dtbclk_mhz)
+			clk_mgr_base->bw_params->dc_mode_limit.dtbclk_mhz = 0;
+	}
+
+	/* DISPCLK */
+	dcn401_init_single_clock(clk_mgr, PPCLK_DISPCLK,
+			&clk_mgr_base->bw_params->clk_table.entries[0].dispclk_mhz,
+			&num_entries_per_clk->num_dispclk_levels);
+	clk_mgr_base->bw_params->dc_mode_limit.dispclk_mhz = dcn401_smu_get_dc_mode_max_dpm_freq(clk_mgr, PPCLK_DISPCLK);
+	if (num_entries_per_clk->num_dispclk_levels && clk_mgr_base->bw_params->dc_mode_limit.dispclk_mhz ==
+			clk_mgr_base->bw_params->clk_table.entries[num_entries_per_clk->num_dispclk_levels - 1].dispclk_mhz)
+		clk_mgr_base->bw_params->dc_mode_limit.dispclk_mhz = 0;
+
+	/* DPPCLK */
+	dcn401_init_single_clock(clk_mgr, PPCLK_DPPCLK,
+			&clk_mgr_base->bw_params->clk_table.entries[0].dppclk_mhz,
+			&num_entries_per_clk->num_dppclk_levels);
+
+	if (num_entries_per_clk->num_dcfclk_levels &&
+			num_entries_per_clk->num_dtbclk_levels &&
+			num_entries_per_clk->num_dispclk_levels)
+		clk_mgr->dpm_present = true;
+
+	if (clk_mgr_base->ctx->dc->debug.min_disp_clk_khz) {
+		for (i = 0; i < num_entries_per_clk->num_dispclk_levels; i++)
+			if (clk_mgr_base->bw_params->clk_table.entries[i].dispclk_mhz
+					< khz_to_mhz_ceil(clk_mgr_base->ctx->dc->debug.min_disp_clk_khz))
+				clk_mgr_base->bw_params->clk_table.entries[i].dispclk_mhz
+					= khz_to_mhz_ceil(clk_mgr_base->ctx->dc->debug.min_disp_clk_khz);
+	}
+
+	if (clk_mgr_base->ctx->dc->debug.min_dpp_clk_khz) {
+		for (i = 0; i < num_entries_per_clk->num_dppclk_levels; i++)
+			if (clk_mgr_base->bw_params->clk_table.entries[i].dppclk_mhz
+					< khz_to_mhz_ceil(clk_mgr_base->ctx->dc->debug.min_dpp_clk_khz))
+				clk_mgr_base->bw_params->clk_table.entries[i].dppclk_mhz
+					= khz_to_mhz_ceil(clk_mgr_base->ctx->dc->debug.min_dpp_clk_khz);
+	}
+
+	/* Get UCLK, update bounding box */
+	clk_mgr_base->funcs->get_memclk_states_from_smu(clk_mgr_base);
+
+	/* WM range table */
+	dcn401_build_wm_range_table(clk_mgr_base);
+}
+
+bool dcn401_is_dc_mode_present(struct clk_mgr *clk_mgr_base)
+{
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+
+	return clk_mgr->smu_present && clk_mgr->dpm_present &&
+			((clk_mgr_base->bw_params->clk_table.num_entries_per_clk.num_dcfclk_levels &&
+			clk_mgr_base->bw_params->dc_mode_limit.dcfclk_mhz) ||
+			(clk_mgr_base->bw_params->clk_table.num_entries_per_clk.num_dispclk_levels &&
+			clk_mgr_base->bw_params->dc_mode_limit.dispclk_mhz) ||
+			(clk_mgr_base->bw_params->clk_table.num_entries_per_clk.num_dtbclk_levels &&
+			clk_mgr_base->bw_params->dc_mode_limit.dtbclk_mhz) ||
+			(clk_mgr_base->bw_params->clk_table.num_entries_per_clk.num_fclk_levels &&
+			clk_mgr_base->bw_params->dc_mode_limit.fclk_mhz) ||
+			(clk_mgr_base->bw_params->clk_table.num_entries_per_clk.num_memclk_levels &&
+			clk_mgr_base->bw_params->dc_mode_limit.memclk_mhz) ||
+			(clk_mgr_base->bw_params->clk_table.num_entries_per_clk.num_socclk_levels &&
+			clk_mgr_base->bw_params->dc_mode_limit.socclk_mhz));
+}
+
+static void dcn401_dump_clk_registers(struct clk_state_registers_and_bypass *regs_and_bypass,
+		struct clk_mgr *clk_mgr_base, struct clk_log_info *log_info)
+{
+		struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+		uint32_t dprefclk_did = 0;
+		uint32_t dcfclk_did = 0;
+		uint32_t dtbclk_did = 0;
+		uint32_t dispclk_did = 0;
+		uint32_t dppclk_did = 0;
+		uint32_t fclk_did = 0;
+		uint32_t target_div = 0;
+
+		/* DFS Slice 0 is used for DISPCLK */
+		dispclk_did = REG_READ(CLK0_CLK0_DFS_CNTL);
+		/* DFS Slice 1 is used for DPPCLK */
+		dppclk_did = REG_READ(CLK0_CLK1_DFS_CNTL);
+		/* DFS Slice 2 is used for DPREFCLK */
+		dprefclk_did = REG_READ(CLK0_CLK2_DFS_CNTL);
+		/* DFS Slice 3 is used for DCFCLK */
+		dcfclk_did = REG_READ(CLK0_CLK3_DFS_CNTL);
+		/* DFS Slice 4 is used for DTBCLK */
+		dtbclk_did = REG_READ(CLK0_CLK4_DFS_CNTL);
+		/* DFS Slice _ is used for FCLK */
+		fclk_did = REG_READ(CLK2_CLK2_DFS_CNTL);
+
+		/* Convert DISPCLK DFS Slice DID to divider*/
+		target_div = dentist_get_divider_from_did(dispclk_did);
+		//Get dispclk in khz
+		regs_and_bypass->dispclk = (DENTIST_DIVIDER_RANGE_SCALE_FACTOR
+				* clk_mgr->base.dentist_vco_freq_khz) / target_div;
+
+		/* Convert DISPCLK DFS Slice DID to divider*/
+		target_div = dentist_get_divider_from_did(dppclk_did);
+		//Get dppclk in khz
+		regs_and_bypass->dppclk = (DENTIST_DIVIDER_RANGE_SCALE_FACTOR
+				* clk_mgr->base.dentist_vco_freq_khz) / target_div;
+
+		/* Convert DPREFCLK DFS Slice DID to divider*/
+		target_div = dentist_get_divider_from_did(dprefclk_did);
+		//Get dprefclk in khz
+		regs_and_bypass->dprefclk = (DENTIST_DIVIDER_RANGE_SCALE_FACTOR
+				* clk_mgr->base.dentist_vco_freq_khz) / target_div;
+
+		/* Convert DCFCLK DFS Slice DID to divider*/
+		target_div = dentist_get_divider_from_did(dcfclk_did);
+		//Get dcfclk in khz
+		regs_and_bypass->dcfclk = (DENTIST_DIVIDER_RANGE_SCALE_FACTOR
+				* clk_mgr->base.dentist_vco_freq_khz) / target_div;
+
+		/* Convert DTBCLK DFS Slice DID to divider*/
+		target_div = dentist_get_divider_from_did(dtbclk_did);
+		//Get dtbclk in khz
+		regs_and_bypass->dtbclk = (DENTIST_DIVIDER_RANGE_SCALE_FACTOR
+				* clk_mgr->base.dentist_vco_freq_khz) / target_div;
+
+		/* Convert DTBCLK DFS Slice DID to divider*/
+		target_div = dentist_get_divider_from_did(fclk_did);
+		//Get fclk in khz
+		regs_and_bypass->fclk = (DENTIST_DIVIDER_RANGE_SCALE_FACTOR
+				* clk_mgr->base.dentist_vco_freq_khz) / target_div;
+}
+
+static bool dcn401_check_native_scaling(struct pipe_ctx *pipe)
+{
+	bool is_native_scaling = false;
+	int width = pipe->plane_state->src_rect.width;
+	int height = pipe->plane_state->src_rect.height;
+
+	if (pipe->stream->timing.h_addressable == width &&
+			pipe->stream->timing.v_addressable == height &&
+			pipe->plane_state->dst_rect.width == width &&
+			pipe->plane_state->dst_rect.height == height)
+		is_native_scaling = true;
+
+	return is_native_scaling;
+}
+
+static void dcn401_auto_dpm_test_log(
+		struct dc_clocks *new_clocks,
+		struct clk_mgr_internal *clk_mgr,
+		struct dc_state *context)
+{
+	unsigned int mall_ss_size_bytes;
+	int dramclk_khz_override, fclk_khz_override, num_fclk_levels;
+
+	struct pipe_ctx *pipe_ctx_list[MAX_PIPES];
+	int active_pipe_count = 0;
+
+	for (int i = 0; i < MAX_PIPES; i++) {
+		struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i];
+
+		if (pipe_ctx->stream && dc_state_get_pipe_subvp_type(context, pipe_ctx) != SUBVP_PHANTOM) {
+			pipe_ctx_list[active_pipe_count] = pipe_ctx;
+			active_pipe_count++;
+		}
+	}
+
+	msleep(5);
+
+	mall_ss_size_bytes = context->bw_ctx.bw.dcn.mall_ss_size_bytes;
+
+	struct clk_log_info log_info = {0};
+	struct clk_state_registers_and_bypass clk_register_dump;
+
+	dcn401_dump_clk_registers(&clk_register_dump, &clk_mgr->base, &log_info);
+
+	// Overrides for these clocks in case there is no p_state change support
+	dramclk_khz_override = new_clocks->dramclk_khz;
+	fclk_khz_override = new_clocks->fclk_khz;
+
+	num_fclk_levels = clk_mgr->base.bw_params->clk_table.num_entries_per_clk.num_fclk_levels - 1;
+
+	if (!new_clocks->p_state_change_support)
+		dramclk_khz_override = clk_mgr->base.bw_params->max_memclk_mhz * 1000;
+
+	if (!new_clocks->fclk_p_state_change_support)
+		fclk_khz_override = clk_mgr->base.bw_params->clk_table.entries[num_fclk_levels].fclk_mhz * 1000;
+
+
+	////////////////////////////////////////////////////////////////////////////
+	//	IMPORTANT: 	When adding more clocks to these logs, do NOT put a newline
+	//	 			anywhere other than at the very end of the string.
+	//
+	//	Formatting example (make sure to have " - " between each entry):
+	//
+	//				AutoDPMTest: clk1:%d - clk2:%d - clk3:%d - clk4:%d\n"
+	////////////////////////////////////////////////////////////////////////////
+	if (active_pipe_count > 0 &&
+		new_clocks->dramclk_khz > 0 &&
+		new_clocks->fclk_khz > 0 &&
+		new_clocks->dcfclk_khz > 0 &&
+		new_clocks->dppclk_khz > 0) {
+
+		uint32_t pix_clk_list[MAX_PIPES] = {0};
+		int p_state_list[MAX_PIPES] = {0};
+		int disp_src_width_list[MAX_PIPES] = {0};
+		int disp_src_height_list[MAX_PIPES] = {0};
+		uint64_t disp_src_refresh_list[MAX_PIPES] = {0};
+		bool is_scaled_list[MAX_PIPES] = {0};
+
+		for (int i = 0; i < active_pipe_count; i++) {
+			struct pipe_ctx *curr_pipe_ctx = pipe_ctx_list[i];
+			uint64_t refresh_rate;
+
+			pix_clk_list[i] = curr_pipe_ctx->stream->timing.pix_clk_100hz;
+			p_state_list[i] = curr_pipe_ctx->p_state_type;
+
+			refresh_rate = (curr_pipe_ctx->stream->timing.pix_clk_100hz * (uint64_t)100 +
+				curr_pipe_ctx->stream->timing.v_total
+				* (uint64_t) curr_pipe_ctx->stream->timing.h_total - (uint64_t)1);
+			refresh_rate = div_u64(refresh_rate, curr_pipe_ctx->stream->timing.v_total);
+			refresh_rate = div_u64(refresh_rate, curr_pipe_ctx->stream->timing.h_total);
+			disp_src_refresh_list[i] = refresh_rate;
+
+			if (curr_pipe_ctx->plane_state) {
+				is_scaled_list[i] = !(dcn401_check_native_scaling(curr_pipe_ctx));
+				disp_src_width_list[i] = curr_pipe_ctx->plane_state->src_rect.width;
+				disp_src_height_list[i] = curr_pipe_ctx->plane_state->src_rect.height;
+			}
+		}
+
+		DC_LOG_AUTO_DPM_TEST("AutoDPMTest: dramclk:%d - fclk:%d - "
+			"dcfclk:%d - dppclk:%d - dispclk_hw:%d - "
+			"dppclk_hw:%d - dprefclk_hw:%d - dcfclk_hw:%d - "
+			"dtbclk_hw:%d - fclk_hw:%d - pix_clk_0:%d - pix_clk_1:%d - "
+			"pix_clk_2:%d - pix_clk_3:%d - mall_ss_size:%d - p_state_type_0:%d - "
+			"p_state_type_1:%d - p_state_type_2:%d - p_state_type_3:%d - "
+			"pix_width_0:%d - pix_height_0:%d - refresh_rate_0:%lld - is_scaled_0:%d - "
+			"pix_width_1:%d - pix_height_1:%d - refresh_rate_1:%lld - is_scaled_1:%d - "
+			"pix_width_2:%d - pix_height_2:%d - refresh_rate_2:%lld - is_scaled_2:%d - "
+			"pix_width_3:%d - pix_height_3:%d - refresh_rate_3:%lld - is_scaled_3:%d - LOG_END\n",
+			dramclk_khz_override,
+			fclk_khz_override,
+			new_clocks->dcfclk_khz,
+			new_clocks->dppclk_khz,
+			clk_register_dump.dispclk,
+			clk_register_dump.dppclk,
+			clk_register_dump.dprefclk,
+			clk_register_dump.dcfclk,
+			clk_register_dump.dtbclk,
+			clk_register_dump.fclk,
+			pix_clk_list[0], pix_clk_list[1], pix_clk_list[3], pix_clk_list[2],
+			mall_ss_size_bytes,
+			p_state_list[0], p_state_list[1], p_state_list[2], p_state_list[3],
+			disp_src_width_list[0], disp_src_height_list[0], disp_src_refresh_list[0], is_scaled_list[0],
+			disp_src_width_list[1], disp_src_height_list[1], disp_src_refresh_list[1], is_scaled_list[1],
+			disp_src_width_list[2], disp_src_height_list[2], disp_src_refresh_list[2], is_scaled_list[2],
+			disp_src_width_list[3], disp_src_height_list[3], disp_src_refresh_list[3], is_scaled_list[3]);
+	}
+}
+
+static void dcn401_update_clocks_update_dtb_dto(struct clk_mgr_internal *clk_mgr,
+			struct dc_state *context,
+			int ref_dtbclk_khz)
+{
+	int i;
+	struct dccg *dccg = clk_mgr->dccg;
+	struct pipe_ctx *otg_master;
+	bool use_hpo_encoder;
+
+
+	for (i = 0; i < context->stream_count; i++) {
+		otg_master = resource_get_otg_master_for_stream(
+				&context->res_ctx, context->streams[i]);
+		ASSERT(otg_master);
+		ASSERT(otg_master->clock_source);
+		ASSERT(otg_master->clock_source->funcs->program_pix_clk);
+		ASSERT(otg_master->stream_res.pix_clk_params.controller_id >= CONTROLLER_ID_D0);
+
+		use_hpo_encoder = dccg->ctx->dc->link_srv->dp_is_128b_132b_signal(otg_master);
+		if (!use_hpo_encoder)
+			continue;
+
+		if (otg_master->stream_res.pix_clk_params.controller_id > CONTROLLER_ID_UNDEFINED)
+			otg_master->clock_source->funcs->program_pix_clk(
+				otg_master->clock_source,
+				&otg_master->stream_res.pix_clk_params,
+				dccg->ctx->dc->link_srv->dp_get_encoding_format(
+					&otg_master->link_config.dp_link_settings),
+				&otg_master->pll_settings);
+	}
+}
+
+static void dcn401_update_clocks_update_dpp_dto(struct clk_mgr_internal *clk_mgr,
+		struct dc_state *context, bool safe_to_lower, int ref_dppclk_khz)
+{
+	int i;
+
+	clk_mgr->dccg->ref_dppclk = ref_dppclk_khz;
+	for (i = 0; i < clk_mgr->base.ctx->dc->res_pool->pipe_count; i++) {
+		int dpp_inst = 0, dppclk_khz, prev_dppclk_khz;
+
+		dppclk_khz = context->res_ctx.pipe_ctx[i].plane_res.bw.dppclk_khz;
+
+		if (context->res_ctx.pipe_ctx[i].plane_res.dpp)
+			dpp_inst = context->res_ctx.pipe_ctx[i].plane_res.dpp->inst;
+		else if (!context->res_ctx.pipe_ctx[i].plane_res.dpp && dppclk_khz == 0) {
+			/* dpp == NULL && dppclk_khz == 0 is valid because of pipe harvesting.
+			 * In this case just continue in loop
+			 */
+			continue;
+		} else if (!context->res_ctx.pipe_ctx[i].plane_res.dpp && dppclk_khz > 0) {
+			/* The software state is not valid if dpp resource is NULL and
+			 * dppclk_khz > 0.
+			 */
+			ASSERT(false);
+			continue;
+		}
+
+		prev_dppclk_khz = clk_mgr->dccg->pipe_dppclk_khz[i];
+
+		if (safe_to_lower || prev_dppclk_khz < dppclk_khz)
+			clk_mgr->dccg->funcs->update_dpp_dto(
+							clk_mgr->dccg, dpp_inst, dppclk_khz);
+	}
+}
+
+static int dcn401_set_hard_min_by_freq_optimized(struct clk_mgr_internal *clk_mgr, PPCLK_e clk, int requested_clk_khz)
+{
+	if (!clk_mgr->smu_present || !dcn401_is_ppclk_dpm_enabled(clk_mgr, clk))
+		return 0;
+
+	/*
+	 * SMU set hard min interface takes requested clock in mhz and return
+	 * actual clock configured in khz. If we floor requested clk to mhz,
+	 * there is a chance that the actual clock configured in khz is less
+	 * than requested. If we ceil it to mhz, there is a chance that it
+	 * unnecessarily dumps up to a higher dpm level, which burns more power.
+	 * The solution is to set by flooring it to mhz first. If the actual
+	 * clock returned is less than requested, then we will ceil the
+	 * requested value to mhz and call it again.
+	 */
+	int actual_clk_khz = dcn401_smu_set_hard_min_by_freq(clk_mgr, clk, khz_to_mhz_floor(requested_clk_khz));
+
+	if (actual_clk_khz < requested_clk_khz)
+		actual_clk_khz = dcn401_smu_set_hard_min_by_freq(clk_mgr, clk, khz_to_mhz_ceil(requested_clk_khz));
+
+	return actual_clk_khz;
+}
+
+static void dcn401_update_clocks_update_dentist(
+		struct clk_mgr_internal *clk_mgr,
+		struct dc_state *context)
+{
+	uint32_t new_disp_divider = 0;
+	uint32_t new_dispclk_wdivider = 0;
+	uint32_t dentist_dispclk_wdivider_readback = 0;
+	struct dc *dc = clk_mgr->base.ctx->dc;
+
+	if (clk_mgr->base.clks.dispclk_khz == 0)
+		return;
+
+	new_disp_divider = DENTIST_DIVIDER_RANGE_SCALE_FACTOR
+			* clk_mgr->base.dentist_vco_freq_khz / clk_mgr->base.clks.dispclk_khz;
+
+	new_dispclk_wdivider = dentist_get_did_from_divider(new_disp_divider);
+
+	if (dc->debug.override_dispclk_programming) {
+		REG_GET(DENTIST_DISPCLK_CNTL,
+				DENTIST_DISPCLK_WDIVIDER, &dentist_dispclk_wdivider_readback);
+
+		if (dentist_dispclk_wdivider_readback > new_dispclk_wdivider) {
+			REG_UPDATE(DENTIST_DISPCLK_CNTL,
+					DENTIST_DISPCLK_WDIVIDER, new_dispclk_wdivider);
+			REG_WAIT(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_CHG_DONE, 1, 50, 2000);
+		}
+	}
+
+}
+
+static void dcn401_execute_block_sequence(struct clk_mgr *clk_mgr_base, unsigned int num_steps)
+{
+	struct clk_mgr_internal *clk_mgr_internal = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+	struct dcn401_clk_mgr *clk_mgr401 = TO_DCN401_CLK_MGR(clk_mgr_internal);
+
+	unsigned int i;
+	union dcn401_clk_mgr_block_sequence_params *params;
+
+	/* execute sequence */
+	for (i = 0; i < num_steps; i++) {
+		params = &clk_mgr401->block_sequence[i].params;
+
+		switch (clk_mgr401->block_sequence[i].func) {
+		case CLK_MGR401_READ_CLOCKS_FROM_DENTIST:
+			dcn2_read_clocks_from_hw_dentist(clk_mgr_base);
+			break;
+		case CLK_MGR401_UPDATE_NUM_DISPLAYS:
+			dcn401_smu_set_num_of_displays(clk_mgr_internal,
+					params->update_num_displays_params.num_displays);
+			break;
+		case CLK_MGR401_UPDATE_HARDMIN_PPCLK:
+			if (params->update_hardmin_params.response)
+				*params->update_hardmin_params.response = dcn401_smu_set_hard_min_by_freq(
+						clk_mgr_internal,
+						params->update_hardmin_params.ppclk,
+						params->update_hardmin_params.freq_mhz);
+			else
+				dcn401_smu_set_hard_min_by_freq(clk_mgr_internal,
+						params->update_hardmin_params.ppclk,
+						params->update_hardmin_params.freq_mhz);
+			break;
+		case CLK_MGR401_UPDATE_HARDMIN_PPCLK_OPTIMIZED:
+			if (params->update_hardmin_optimized_params.response)
+				*params->update_hardmin_optimized_params.response = dcn401_set_hard_min_by_freq_optimized(
+						clk_mgr_internal,
+						params->update_hardmin_optimized_params.ppclk,
+						params->update_hardmin_optimized_params.freq_khz);
+			else
+				dcn401_set_hard_min_by_freq_optimized(clk_mgr_internal,
+						params->update_hardmin_optimized_params.ppclk,
+						params->update_hardmin_optimized_params.freq_khz);
+			break;
+		case CLK_MGR401_UPDATE_ACTIVE_HARDMINS:
+			dcn401_smu_set_active_uclk_fclk_hardmin(
+					clk_mgr_internal,
+					params->update_idle_hardmin_params.uclk_mhz,
+					params->update_idle_hardmin_params.fclk_mhz);
+			break;
+		case CLK_MGR401_UPDATE_IDLE_HARDMINS:
+			dcn401_smu_set_idle_uclk_fclk_hardmin(
+					clk_mgr_internal,
+					params->update_idle_hardmin_params.uclk_mhz,
+					params->update_idle_hardmin_params.fclk_mhz);
+			break;
+		case CLK_MGR401_UPDATE_SUBVP_HARDMINS:
+			dcn401_smu_set_subvp_uclk_fclk_hardmin(
+					clk_mgr_internal,
+					params->update_idle_hardmin_params.uclk_mhz,
+					params->update_idle_hardmin_params.fclk_mhz);
+			break;
+		case CLK_MGR401_UPDATE_DEEP_SLEEP_DCFCLK:
+			dcn401_smu_set_min_deep_sleep_dcef_clk(
+					clk_mgr_internal,
+					params->update_deep_sleep_dcfclk_params.freq_mhz);
+			break;
+		case CLK_MGR401_UPDATE_FCLK_PSTATE_SUPPORT:
+			dcn401_smu_send_fclk_pstate_message(
+					clk_mgr_internal,
+					params->update_pstate_support_params.support);
+			break;
+		case CLK_MGR401_UPDATE_UCLK_PSTATE_SUPPORT:
+			dcn401_smu_send_uclk_pstate_message(
+					clk_mgr_internal,
+					params->update_pstate_support_params.support);
+			break;
+		case CLK_MGR401_UPDATE_CAB_FOR_UCLK:
+			dcn401_smu_send_cab_for_uclk_message(
+				clk_mgr_internal,
+				params->update_cab_for_uclk_params.num_ways);
+			break;
+		case CLK_MGR401_UPDATE_WAIT_FOR_DMUB_ACK:
+			dcn401_smu_wait_for_dmub_ack_mclk(
+					clk_mgr_internal,
+					params->update_wait_for_dmub_ack_params.enable);
+			break;
+		case CLK_MGR401_INDICATE_DRR_STATUS:
+			dcn401_smu_indicate_drr_status(
+					clk_mgr_internal,
+					params->indicate_drr_status_params.mod_drr_for_pstate);
+			break;
+		case CLK_MGR401_UPDATE_DPPCLK_DTO:
+			dcn401_update_clocks_update_dpp_dto(
+					clk_mgr_internal,
+					params->update_dppclk_dto_params.context,
+					params->update_dppclk_dto_params.safe_to_lower,
+					*params->update_dppclk_dto_params.ref_dppclk_khz);
+			break;
+		case CLK_MGR401_UPDATE_DTBCLK_DTO:
+			dcn401_update_clocks_update_dtb_dto(
+					clk_mgr_internal,
+					params->update_dtbclk_dto_params.context,
+					*params->update_dtbclk_dto_params.ref_dtbclk_khz);
+			break;
+		case CLK_MGR401_UPDATE_DENTIST:
+			dcn401_update_clocks_update_dentist(
+					clk_mgr_internal,
+					params->update_dentist_params.context);
+			break;
+		case CLK_MGR401_UPDATE_PSR_WAIT_LOOP:
+			params->update_psr_wait_loop_params.dmcu->funcs->set_psr_wait_loop(
+					params->update_psr_wait_loop_params.dmcu,
+					params->update_psr_wait_loop_params.wait);
+			break;
+		default:
+			/* this should never happen */
+			BREAK_TO_DEBUGGER();
+			break;
+		}
+	}
+}
+
+static unsigned int dcn401_build_update_bandwidth_clocks_sequence(
+		struct clk_mgr *clk_mgr_base,
+		struct dc_state *context,
+		struct dc_clocks *new_clocks,
+		bool safe_to_lower)
+{
+	struct clk_mgr_internal *clk_mgr_internal = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+	struct dcn401_clk_mgr *clk_mgr401 = TO_DCN401_CLK_MGR(clk_mgr_internal);
+	struct dc *dc = clk_mgr_base->ctx->dc;
+	struct dcn401_clk_mgr_block_sequence *block_sequence = clk_mgr401->block_sequence;
+	bool enter_display_off = false;
+	bool update_active_fclk = false;
+	bool update_active_uclk = false;
+	bool update_idle_fclk = false;
+	bool update_idle_uclk = false;
+	bool update_subvp_prefetch_dramclk = false;
+	bool update_subvp_prefetch_fclk = false;
+	bool is_idle_dpm_enabled = dcn401_is_ppclk_dpm_enabled(clk_mgr_internal, PPCLK_UCLK) &&
+			dcn401_is_ppclk_dpm_enabled(clk_mgr_internal, PPCLK_FCLK) &&
+			dcn401_is_ppclk_idle_dpm_enabled(clk_mgr_internal, PPCLK_UCLK) &&
+			dcn401_is_ppclk_idle_dpm_enabled(clk_mgr_internal, PPCLK_FCLK);
+	bool is_df_throttle_opt_enabled = is_idle_dpm_enabled &&
+		dcn401_is_df_throttle_opt_enabled(clk_mgr_internal);
+	int total_plane_count = clk_mgr_helper_get_active_plane_cnt(dc, context);
+	int active_uclk_mhz = khz_to_mhz_ceil(clk_mgr_base->clks.dramclk_khz);
+	int active_fclk_mhz = khz_to_mhz_ceil(clk_mgr_base->clks.fclk_khz);
+	int idle_uclk_mhz = khz_to_mhz_ceil(clk_mgr_base->clks.idle_dramclk_khz);
+	int idle_fclk_mhz = khz_to_mhz_ceil(clk_mgr_base->clks.idle_fclk_khz);
+	int subvp_prefetch_dramclk_mhz = khz_to_mhz_ceil(clk_mgr_base->clks.subvp_prefetch_dramclk_khz);
+	int subvp_prefetch_fclk_mhz = khz_to_mhz_ceil(clk_mgr_base->clks.subvp_prefetch_fclk_khz);
+
+	unsigned int num_steps = 0;
+
+	int display_count;
+	bool fclk_p_state_change_support, uclk_p_state_change_support;
+
+	/* CLK_MGR401_UPDATE_NUM_DISPLAYS */
+	if (clk_mgr_internal->smu_present) {
+		display_count = clk_mgr_helper_get_active_display_cnt(dc, context);
+
+		if (display_count == 0)
+			enter_display_off = true;
+
+		if (enter_display_off == safe_to_lower) {
+			block_sequence[num_steps].params.update_num_displays_params.num_displays = display_count;
+			block_sequence[num_steps].func = CLK_MGR401_UPDATE_NUM_DISPLAYS;
+			num_steps++;
+		}
+	}
+
+	/* CLK_MGR401_UPDATE_FCLK_PSTATE_SUPPORT */
+	clk_mgr_base->clks.fclk_prev_p_state_change_support = clk_mgr_base->clks.fclk_p_state_change_support;
+	fclk_p_state_change_support = new_clocks->fclk_p_state_change_support || (total_plane_count == 0);
+	if (should_update_pstate_support(safe_to_lower, fclk_p_state_change_support, clk_mgr_base->clks.fclk_prev_p_state_change_support)) {
+		clk_mgr_base->clks.fclk_p_state_change_support = fclk_p_state_change_support;
+		update_active_fclk = true;
+		update_idle_fclk = true;
+
+		/* To enable FCLK P-state switching, send PSTATE_SUPPORTED message to PMFW (message not supported on DCN401)*/
+		// if (clk_mgr_base->clks.fclk_p_state_change_support) {
+		// 	/* Handle the code for sending a message to PMFW that FCLK P-state change is supported */
+		// 	if (dcn401_is_ppclk_dpm_enabled(clk_mgr_internal, PPCLK_FCLK)) {
+		// 		block_sequence[num_steps].params.update_pstate_support_params.support = true;
+		// 		block_sequence[num_steps].func = CLK_MGR401_UPDATE_FCLK_PSTATE_SUPPORT;
+		// 		num_steps++;
+		// 	}
+		// }
+	}
+
+	if (!clk_mgr_base->clks.fclk_p_state_change_support && dcn401_is_ppclk_dpm_enabled(clk_mgr_internal, PPCLK_FCLK)) {
+		/* when P-State switching disabled, set UCLK min = max */
+		idle_fclk_mhz =
+				clk_mgr_base->bw_params->clk_table.entries[clk_mgr_base->bw_params->clk_table.num_entries_per_clk.num_fclk_levels - 1].fclk_mhz;
+		active_fclk_mhz = idle_fclk_mhz;
+	}
+
+	/* UPDATE DCFCLK */
+	if (dc->debug.force_min_dcfclk_mhz > 0)
+		new_clocks->dcfclk_khz = (new_clocks->dcfclk_khz > (dc->debug.force_min_dcfclk_mhz * 1000)) ?
+				new_clocks->dcfclk_khz : (dc->debug.force_min_dcfclk_mhz * 1000);
+
+	if (should_set_clock(safe_to_lower, new_clocks->dcfclk_khz, clk_mgr_base->clks.dcfclk_khz)) {
+		clk_mgr_base->clks.dcfclk_khz = new_clocks->dcfclk_khz;
+		if (dcn401_is_ppclk_dpm_enabled(clk_mgr_internal, PPCLK_DCFCLK)) {
+			block_sequence[num_steps].params.update_hardmin_params.ppclk = PPCLK_DCFCLK;
+			block_sequence[num_steps].params.update_hardmin_params.freq_mhz = khz_to_mhz_ceil(clk_mgr_base->clks.dcfclk_khz);
+			block_sequence[num_steps].params.update_hardmin_params.response = NULL;
+			block_sequence[num_steps].func = CLK_MGR401_UPDATE_HARDMIN_PPCLK;
+			num_steps++;
+		}
+	}
+
+	/* CLK_MGR401_UPDATE_DEEP_SLEEP_DCFCLK */
+	if (should_set_clock(safe_to_lower, new_clocks->dcfclk_deep_sleep_khz, clk_mgr_base->clks.dcfclk_deep_sleep_khz)) {
+		clk_mgr_base->clks.dcfclk_deep_sleep_khz = new_clocks->dcfclk_deep_sleep_khz;
+		if (dcn401_is_ppclk_dpm_enabled(clk_mgr_internal, PPCLK_DCFCLK)) {
+			block_sequence[num_steps].params.update_deep_sleep_dcfclk_params.freq_mhz = khz_to_mhz_ceil(clk_mgr_base->clks.dcfclk_deep_sleep_khz);
+			block_sequence[num_steps].func = CLK_MGR401_UPDATE_DEEP_SLEEP_DCFCLK;
+			num_steps++;
+		}
+	}
+
+	/* SOCCLK */
+	if (should_set_clock(safe_to_lower, new_clocks->socclk_khz, clk_mgr_base->clks.socclk_khz))
+		/* We don't actually care about socclk, don't notify SMU of hard min */
+		clk_mgr_base->clks.socclk_khz = new_clocks->socclk_khz;
+
+	/* UCLK */
+	if (new_clocks->fw_based_mclk_switching != clk_mgr_base->clks.fw_based_mclk_switching &&
+			new_clocks->fw_based_mclk_switching) {
+		/* enable FAMS features */
+		clk_mgr_base->clks.fw_based_mclk_switching = new_clocks->fw_based_mclk_switching;
+
+		block_sequence[num_steps].params.update_wait_for_dmub_ack_params.enable = clk_mgr_base->clks.fw_based_mclk_switching;
+		block_sequence[num_steps].func = CLK_MGR401_UPDATE_WAIT_FOR_DMUB_ACK;
+		num_steps++;
+
+		block_sequence[num_steps].params.indicate_drr_status_params.mod_drr_for_pstate = clk_mgr_base->clks.fw_based_mclk_switching;
+		block_sequence[num_steps].func = CLK_MGR401_INDICATE_DRR_STATUS;
+		num_steps++;
+	}
+
+	/* CLK_MGR401_UPDATE_CAB_FOR_UCLK */
+	clk_mgr_base->clks.prev_num_ways = clk_mgr_base->clks.num_ways;
+	if (clk_mgr_base->clks.num_ways != new_clocks->num_ways &&
+			clk_mgr_base->clks.num_ways < new_clocks->num_ways) {
+		/* increase num ways for subvp */
+		clk_mgr_base->clks.num_ways = new_clocks->num_ways;
+		if (dcn401_is_ppclk_dpm_enabled(clk_mgr_internal, PPCLK_UCLK)) {
+			block_sequence[num_steps].params.update_cab_for_uclk_params.num_ways = clk_mgr_base->clks.num_ways;
+			block_sequence[num_steps].func = CLK_MGR401_UPDATE_CAB_FOR_UCLK;
+			num_steps++;
+		}
+	}
+
+	clk_mgr_base->clks.prev_p_state_change_support = clk_mgr_base->clks.p_state_change_support;
+	uclk_p_state_change_support = new_clocks->p_state_change_support || (total_plane_count == 0);
+	if (should_update_pstate_support(safe_to_lower, uclk_p_state_change_support, clk_mgr_base->clks.prev_p_state_change_support)) {
+		clk_mgr_base->clks.p_state_change_support = uclk_p_state_change_support;
+		update_active_uclk = true;
+		update_idle_uclk = true;
+
+		if (clk_mgr_base->clks.p_state_change_support) {
+			/* enable UCLK switching  */
+			if (dcn401_is_ppclk_dpm_enabled(clk_mgr_internal, PPCLK_UCLK)) {
+				block_sequence[num_steps].params.update_pstate_support_params.support = true;
+				block_sequence[num_steps].func = CLK_MGR401_UPDATE_UCLK_PSTATE_SUPPORT;
+				num_steps++;
+			}
+		}
+	}
+
+	if (!clk_mgr_base->clks.p_state_change_support && dcn401_is_ppclk_dpm_enabled(clk_mgr_internal, PPCLK_UCLK)) {
+		/* when P-State switching disabled, set UCLK min = max */
+		if (dc->clk_mgr->dc_mode_softmax_enabled) {
+			/* will never have the functional UCLK min above the softmax
+			* since we calculate mode support based on softmax being the max UCLK
+			* frequency.
+			*/
+			active_uclk_mhz = clk_mgr_base->bw_params->dc_mode_softmax_memclk;
+		} else {
+			active_uclk_mhz = clk_mgr_base->bw_params->max_memclk_mhz;
+		}
+		idle_uclk_mhz = active_uclk_mhz;
+	}
+
+	/* Always update saved value, even if new value not set due to P-State switching unsupported */
+	if (should_set_clock(safe_to_lower, new_clocks->dramclk_khz, clk_mgr_base->clks.dramclk_khz)) {
+		clk_mgr_base->clks.dramclk_khz = new_clocks->dramclk_khz;
+
+		if (clk_mgr_base->clks.p_state_change_support) {
+			update_active_uclk = true;
+			active_uclk_mhz = khz_to_mhz_ceil(clk_mgr_base->clks.dramclk_khz);
+		}
+	}
+
+	if (should_set_clock(safe_to_lower, new_clocks->idle_dramclk_khz, clk_mgr_base->clks.idle_dramclk_khz)) {
+		clk_mgr_base->clks.idle_dramclk_khz = new_clocks->idle_dramclk_khz;
+
+		if (clk_mgr_base->clks.p_state_change_support) {
+			update_idle_uclk = true;
+			idle_uclk_mhz = khz_to_mhz_ceil(clk_mgr_base->clks.idle_dramclk_khz);
+		}
+	}
+
+	if (should_set_clock(safe_to_lower, new_clocks->subvp_prefetch_dramclk_khz, clk_mgr_base->clks.subvp_prefetch_dramclk_khz)) {
+		clk_mgr_base->clks.subvp_prefetch_dramclk_khz = new_clocks->subvp_prefetch_dramclk_khz;
+		update_subvp_prefetch_dramclk = true;
+		subvp_prefetch_dramclk_mhz = khz_to_mhz_ceil(clk_mgr_base->clks.subvp_prefetch_dramclk_khz);
+	}
+
+	/* FCLK */
+	/* Always update saved value, even if new value not set due to P-State switching unsupported */
+	if (should_set_clock(safe_to_lower, new_clocks->fclk_khz, clk_mgr_base->clks.fclk_khz)) {
+		clk_mgr_base->clks.fclk_khz = new_clocks->fclk_khz;
+
+		if (clk_mgr_base->clks.fclk_p_state_change_support) {
+			update_active_fclk = true;
+			active_fclk_mhz = khz_to_mhz_ceil(clk_mgr_base->clks.fclk_khz);
+		}
+	}
+
+	if (should_set_clock(safe_to_lower, new_clocks->idle_fclk_khz, clk_mgr_base->clks.idle_fclk_khz)) {
+		clk_mgr_base->clks.idle_fclk_khz = new_clocks->idle_fclk_khz;
+
+		if (clk_mgr_base->clks.fclk_p_state_change_support) {
+			update_idle_fclk = true;
+			idle_fclk_mhz = khz_to_mhz_ceil(clk_mgr_base->clks.idle_fclk_khz);
+		}
+	}
+
+	if (should_set_clock(safe_to_lower, new_clocks->subvp_prefetch_fclk_khz, clk_mgr_base->clks.subvp_prefetch_fclk_khz)) {
+		clk_mgr_base->clks.subvp_prefetch_fclk_khz = new_clocks->subvp_prefetch_fclk_khz;
+		update_subvp_prefetch_fclk = true;
+		subvp_prefetch_fclk_mhz = khz_to_mhz_ceil(clk_mgr_base->clks.subvp_prefetch_fclk_khz);
+	}
+
+	/* When idle DPM is enabled, need to send active and idle hardmins separately */
+	/* CLK_MGR401_UPDATE_ACTIVE_HARDMINS */
+	if ((update_active_uclk || update_active_fclk) && is_idle_dpm_enabled) {
+		block_sequence[num_steps].params.update_idle_hardmin_params.uclk_mhz = active_uclk_mhz;
+		block_sequence[num_steps].params.update_idle_hardmin_params.fclk_mhz = active_fclk_mhz;
+		block_sequence[num_steps].func = CLK_MGR401_UPDATE_ACTIVE_HARDMINS;
+		num_steps++;
+	}
+
+	/* CLK_MGR401_UPDATE_IDLE_HARDMINS */
+	if ((update_idle_uclk || update_idle_fclk) && is_idle_dpm_enabled) {
+		block_sequence[num_steps].params.update_idle_hardmin_params.uclk_mhz = idle_uclk_mhz;
+		block_sequence[num_steps].params.update_idle_hardmin_params.fclk_mhz = idle_fclk_mhz;
+		block_sequence[num_steps].func = CLK_MGR401_UPDATE_IDLE_HARDMINS;
+		num_steps++;
+	}
+
+	/* CLK_MGR401_UPDATE_SUBVP_HARDMINS */
+	if ((update_subvp_prefetch_dramclk || update_subvp_prefetch_fclk) && is_df_throttle_opt_enabled) {
+		block_sequence[num_steps].params.update_idle_hardmin_params.uclk_mhz = subvp_prefetch_dramclk_mhz;
+		block_sequence[num_steps].params.update_idle_hardmin_params.fclk_mhz = subvp_prefetch_fclk_mhz;
+		block_sequence[num_steps].func = CLK_MGR401_UPDATE_SUBVP_HARDMINS;
+		num_steps++;
+	}
+
+	/* set UCLK to requested value if P-State switching is supported, or to re-enable P-State switching */
+	if (update_active_uclk || update_idle_uclk) {
+		if (!is_idle_dpm_enabled) {
+			block_sequence[num_steps].params.update_hardmin_params.ppclk = PPCLK_UCLK;
+			block_sequence[num_steps].params.update_hardmin_params.freq_mhz = active_uclk_mhz;
+			block_sequence[num_steps].params.update_hardmin_params.response = NULL;
+			block_sequence[num_steps].func = CLK_MGR401_UPDATE_HARDMIN_PPCLK;
+			num_steps++;
+		}
+
+		/* disable UCLK P-State support if needed */
+		if (!uclk_p_state_change_support &&
+				should_update_pstate_support(safe_to_lower, uclk_p_state_change_support, clk_mgr_base->clks.prev_p_state_change_support) &&
+				dcn401_is_ppclk_dpm_enabled(clk_mgr_internal, PPCLK_UCLK)) {
+			block_sequence[num_steps].params.update_pstate_support_params.support = false;
+			block_sequence[num_steps].func = CLK_MGR401_UPDATE_UCLK_PSTATE_SUPPORT;
+			num_steps++;
+		}
+	}
+
+	/* set FCLK to requested value if P-State switching is supported, or to re-enable P-State switching */
+	if (update_active_fclk || update_idle_fclk) {
+		/* No need to send active FCLK hardmin, automatically set based on DCFCLK */
+		// if (!is_idle_dpm_enabled) {
+		// 	block_sequence[*num_steps].update_hardmin_params.clk_mgr = clk_mgr;
+		// 	block_sequence[*num_steps].update_hardmin_params.ppclk = PPCLK_FCLK;
+		// 	block_sequence[*num_steps].update_hardmin_params.freq_mhz = active_fclk_mhz;
+		// 	block_sequence[*num_steps].update_hardmin_params.response = NULL;
+		// 	block_sequence[*num_steps].func = CLK_MGR401_UPDATE_HARDMIN_PPCLK;
+		// 	(*num_steps)++;
+		// }
+
+		/* disable FCLK P-State support if needed (message not supported on DCN401)*/
+		// if (!fclk_p_state_change_support &&
+		// 		should_update_pstate_support(safe_to_lower, fclk_p_state_change_support, clk_mgr_base->clks.fclk_prev_p_state_change_support) &&
+		// 		dcn401_is_ppclk_dpm_enabled(clk_mgr_internal, PPCLK_FCLK)) {
+		// 	block_sequence[num_steps].params.update_pstate_support_params.support = false;
+		// 	block_sequence[num_steps].func = CLK_MGR401_UPDATE_FCLK_PSTATE_SUPPORT;
+		// 	num_steps++;
+		// }
+	}
+
+	if (new_clocks->fw_based_mclk_switching != clk_mgr_base->clks.fw_based_mclk_switching &&
+			safe_to_lower && !new_clocks->fw_based_mclk_switching) {
+		/* disable FAMS features */
+		clk_mgr_base->clks.fw_based_mclk_switching = new_clocks->fw_based_mclk_switching;
+
+		block_sequence[num_steps].params.update_wait_for_dmub_ack_params.enable = clk_mgr_base->clks.fw_based_mclk_switching;
+		block_sequence[num_steps].func = CLK_MGR401_UPDATE_WAIT_FOR_DMUB_ACK;
+		num_steps++;
+
+		block_sequence[num_steps].params.indicate_drr_status_params.mod_drr_for_pstate = clk_mgr_base->clks.fw_based_mclk_switching;
+		block_sequence[num_steps].func = CLK_MGR401_INDICATE_DRR_STATUS;
+		num_steps++;
+	}
+
+	/* CLK_MGR401_UPDATE_CAB_FOR_UCLK */
+	if (clk_mgr_base->clks.num_ways != new_clocks->num_ways &&
+			safe_to_lower && clk_mgr_base->clks.num_ways > new_clocks->num_ways) {
+		/* decrease num ways for subvp */
+		clk_mgr_base->clks.num_ways = new_clocks->num_ways;
+		if (dcn401_is_ppclk_dpm_enabled(clk_mgr_internal, PPCLK_UCLK)) {
+			block_sequence[num_steps].params.update_cab_for_uclk_params.num_ways = clk_mgr_base->clks.num_ways;
+			block_sequence[num_steps].func = CLK_MGR401_UPDATE_CAB_FOR_UCLK;
+			num_steps++;
+		}
+	}
+
+	return num_steps;
+}
+
+static unsigned int dcn401_build_update_display_clocks_sequence(
+		struct clk_mgr *clk_mgr_base,
+		struct dc_state *context,
+		struct dc_clocks *new_clocks,
+		bool safe_to_lower)
+{
+	struct clk_mgr_internal *clk_mgr_internal = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+	struct dcn401_clk_mgr *clk_mgr401 = TO_DCN401_CLK_MGR(clk_mgr_internal);
+	struct dc *dc = clk_mgr_base->ctx->dc;
+	struct dmcu *dmcu = clk_mgr_base->ctx->dc->res_pool->dmcu;
+	struct dcn401_clk_mgr_block_sequence *block_sequence = clk_mgr401->block_sequence;
+	bool force_reset = false;
+	bool update_dispclk = false;
+	bool update_dppclk = false;
+	bool dppclk_lowered = false;
+
+	unsigned int num_steps = 0;
+
+	/* CLK_MGR401_READ_CLOCKS_FROM_DENTIST */
+	if (clk_mgr_base->clks.dispclk_khz == 0 ||
+			(dc->debug.force_clock_mode & 0x1)) {
+		/* This is from resume or boot up, if forced_clock cfg option used,
+		 * we bypass program dispclk and DPPCLK, but need set them for S3.
+		 * Force_clock_mode 0x1:  force reset the clock even it is the same clock
+		 * as long as it is in Passive level.
+		 */
+		force_reset = true;
+
+		clk_mgr_base->clks.dispclk_khz = clk_mgr_base->boot_snapshot.dispclk;
+		clk_mgr_base->clks.actual_dispclk_khz = clk_mgr_base->clks.dispclk_khz;
+
+		clk_mgr_base->clks.dppclk_khz = clk_mgr_base->boot_snapshot.dppclk;
+		clk_mgr_base->clks.actual_dppclk_khz = clk_mgr_base->clks.dppclk_khz;
+	}
+
+	/* DTBCLK */
+	if (!new_clocks->dtbclk_en && dcn401_is_ppclk_dpm_enabled(clk_mgr_internal, PPCLK_DTBCLK)) {
+		new_clocks->ref_dtbclk_khz = clk_mgr_base->bw_params->clk_table.entries[0].dtbclk_mhz * 1000;
+	}
+
+	/* clock limits are received with MHz precision, divide by 1000 to prevent setting clocks at every call */
+	if (!dc->debug.disable_dtb_ref_clk_switch &&
+			should_set_clock(safe_to_lower, new_clocks->ref_dtbclk_khz / 1000, clk_mgr_base->clks.ref_dtbclk_khz / 1000) && //TODO these should be ceiled
+			dcn401_is_ppclk_dpm_enabled(clk_mgr_internal, PPCLK_DTBCLK)) {
+		/* DCCG requires KHz precision for DTBCLK */
+		block_sequence[num_steps].params.update_hardmin_params.ppclk = PPCLK_DTBCLK;
+		block_sequence[num_steps].params.update_hardmin_params.freq_mhz = khz_to_mhz_ceil(new_clocks->ref_dtbclk_khz);
+		block_sequence[num_steps].params.update_hardmin_params.response = &clk_mgr_base->clks.ref_dtbclk_khz;
+		block_sequence[num_steps].func = CLK_MGR401_UPDATE_HARDMIN_PPCLK;
+		num_steps++;
+
+		/* Update DTO in DCCG */
+		block_sequence[num_steps].params.update_dtbclk_dto_params.context = context;
+		block_sequence[num_steps].params.update_dtbclk_dto_params.ref_dtbclk_khz = &clk_mgr_base->clks.ref_dtbclk_khz;
+		block_sequence[num_steps].func = CLK_MGR401_UPDATE_DTBCLK_DTO;
+		num_steps++;
+	}
+
+	if (should_set_clock(safe_to_lower, new_clocks->dppclk_khz, clk_mgr_base->clks.dppclk_khz)) {
+		if (clk_mgr_base->clks.dppclk_khz > new_clocks->dppclk_khz)
+			dppclk_lowered = true;
+
+		clk_mgr_base->clks.dppclk_khz = new_clocks->dppclk_khz;
+		clk_mgr_base->clks.actual_dppclk_khz = new_clocks->dppclk_khz;
+
+		update_dppclk = true;
+	}
+
+	if (should_set_clock(safe_to_lower, new_clocks->dispclk_khz, clk_mgr_base->clks.dispclk_khz)) {
+		clk_mgr_base->clks.dispclk_khz = new_clocks->dispclk_khz;
+
+		block_sequence[num_steps].params.update_hardmin_optimized_params.ppclk = PPCLK_DISPCLK;
+		block_sequence[num_steps].params.update_hardmin_optimized_params.freq_khz = clk_mgr_base->clks.dispclk_khz;
+		block_sequence[num_steps].params.update_hardmin_optimized_params.response = &clk_mgr_base->clks.actual_dispclk_khz;
+		block_sequence[num_steps].func = CLK_MGR401_UPDATE_HARDMIN_PPCLK_OPTIMIZED;
+		num_steps++;
+
+		update_dispclk = true;
+	}
+
+	if (dc->config.forced_clocks == false || (force_reset && safe_to_lower)) {
+		if (dppclk_lowered) {
+			/* if clock is being lowered, increase DTO before lowering refclk */
+			block_sequence[num_steps].params.update_dppclk_dto_params.context = context;
+			block_sequence[num_steps].params.update_dppclk_dto_params.ref_dppclk_khz = &clk_mgr_base->clks.dppclk_khz;
+			block_sequence[num_steps].params.update_dppclk_dto_params.safe_to_lower = safe_to_lower;
+			block_sequence[num_steps].func = CLK_MGR401_UPDATE_DPPCLK_DTO;
+			num_steps++;
+
+			block_sequence[num_steps].params.update_dentist_params.context = context;
+			block_sequence[num_steps].func = CLK_MGR401_UPDATE_DENTIST;
+			num_steps++;
+
+			if (dcn401_is_ppclk_dpm_enabled(clk_mgr_internal, PPCLK_DPPCLK)) {
+				block_sequence[num_steps].params.update_hardmin_optimized_params.ppclk = PPCLK_DPPCLK;
+				block_sequence[num_steps].params.update_hardmin_optimized_params.freq_khz = clk_mgr_base->clks.dppclk_khz;
+				block_sequence[num_steps].params.update_hardmin_optimized_params.response = &clk_mgr_base->clks.actual_dppclk_khz;
+				block_sequence[num_steps].func = CLK_MGR401_UPDATE_HARDMIN_PPCLK_OPTIMIZED;
+				num_steps++;
+
+				block_sequence[num_steps].params.update_dppclk_dto_params.context = context;
+				block_sequence[num_steps].params.update_dppclk_dto_params.ref_dppclk_khz = &clk_mgr_base->clks.actual_dppclk_khz;
+				block_sequence[num_steps].params.update_dppclk_dto_params.safe_to_lower = safe_to_lower;
+				block_sequence[num_steps].func = CLK_MGR401_UPDATE_DPPCLK_DTO;
+				num_steps++;
+			}
+		} else {
+			/* if clock is being raised, increase refclk before lowering DTO */
+			if (update_dppclk && dcn401_is_ppclk_dpm_enabled(clk_mgr_internal, PPCLK_DPPCLK)) {
+				block_sequence[num_steps].params.update_hardmin_optimized_params.ppclk = PPCLK_DPPCLK;
+				block_sequence[num_steps].params.update_hardmin_optimized_params.freq_khz = clk_mgr_base->clks.dppclk_khz;
+				block_sequence[num_steps].params.update_hardmin_optimized_params.response = &clk_mgr_base->clks.actual_dppclk_khz;
+				block_sequence[num_steps].func = CLK_MGR401_UPDATE_HARDMIN_PPCLK_OPTIMIZED;
+				num_steps++;
+			}
+
+			if (update_dppclk || update_dispclk) {
+				block_sequence[num_steps].params.update_dentist_params.context = context;
+				block_sequence[num_steps].func = CLK_MGR401_UPDATE_DENTIST;
+				num_steps++;
+			}
+
+			block_sequence[num_steps].params.update_dppclk_dto_params.context = context;
+			block_sequence[num_steps].params.update_dppclk_dto_params.ref_dppclk_khz = &clk_mgr_base->clks.actual_dppclk_khz;
+			block_sequence[num_steps].params.update_dppclk_dto_params.safe_to_lower = safe_to_lower;
+			block_sequence[num_steps].func = CLK_MGR401_UPDATE_DPPCLK_DTO;
+			num_steps++;
+		}
+	}
+
+	if (update_dispclk && dmcu && dmcu->funcs->is_dmcu_initialized(dmcu)) {
+		/*update dmcu for wait_loop count*/
+		block_sequence[num_steps].params.update_psr_wait_loop_params.dmcu = dmcu;
+		block_sequence[num_steps].params.update_psr_wait_loop_params.wait = clk_mgr_base->clks.dispclk_khz / 1000 / 7;
+		block_sequence[num_steps].func = CLK_MGR401_UPDATE_PSR_WAIT_LOOP;
+		num_steps++;
+	}
+
+	return num_steps;
+}
+
+static void dcn401_update_clocks(struct clk_mgr *clk_mgr_base,
+		struct dc_state *context,
+		bool safe_to_lower)
+{
+	struct dc *dc = clk_mgr_base->ctx->dc;
+
+	unsigned int num_steps = 0;
+
+	/* build bandwidth related clocks update sequence */
+	num_steps = dcn401_build_update_bandwidth_clocks_sequence(clk_mgr_base,
+			context,
+			&context->bw_ctx.bw.dcn.clk,
+			safe_to_lower);
+
+	/* execute sequence */
+	dcn401_execute_block_sequence(clk_mgr_base,	num_steps);
+
+	/* build display related clocks update sequence */
+	num_steps = dcn401_build_update_display_clocks_sequence(clk_mgr_base,
+			context,
+			&context->bw_ctx.bw.dcn.clk,
+			safe_to_lower);
+
+	/* execute sequence */
+	dcn401_execute_block_sequence(clk_mgr_base,	num_steps);
+
+	if (dc->config.enable_auto_dpm_test_logs)
+		dcn401_auto_dpm_test_log(&context->bw_ctx.bw.dcn.clk, TO_CLK_MGR_INTERNAL(clk_mgr_base), context);
+
+}
+
+
+static uint32_t dcn401_get_vco_frequency_from_reg(struct clk_mgr_internal *clk_mgr)
+{
+		struct fixed31_32 pll_req;
+		uint32_t pll_req_reg = 0;
+
+		/* get FbMult value */
+		pll_req_reg = REG_READ(CLK0_CLK_PLL_REQ);
+
+		/* set up a fixed-point number
+		 * this works because the int part is on the right edge of the register
+		 * and the frac part is on the left edge
+		 */
+		pll_req = dc_fixpt_from_int(pll_req_reg & clk_mgr->clk_mgr_mask->FbMult_int);
+		pll_req.value |= pll_req_reg & clk_mgr->clk_mgr_mask->FbMult_frac;
+
+		/* multiply by REFCLK period */
+		pll_req = dc_fixpt_mul_int(pll_req, clk_mgr->dfs_ref_freq_khz);
+
+		return dc_fixpt_floor(pll_req);
+}
+
+static void dcn401_clock_read_ss_info(struct clk_mgr_internal *clk_mgr)
+{
+	struct dc_bios *bp = clk_mgr->base.ctx->dc_bios;
+	int ss_info_num = bp->funcs->get_ss_entry_number(
+			bp, AS_SIGNAL_TYPE_GPU_PLL);
+
+	if (ss_info_num) {
+		struct spread_spectrum_info info = { { 0 } };
+		enum bp_result result = bp->funcs->get_spread_spectrum_info(
+				bp, AS_SIGNAL_TYPE_GPU_PLL, 0, &info);
+
+		/* SSInfo.spreadSpectrumPercentage !=0 would be sign
+		 * that SS is enabled
+		 */
+		if (result == BP_RESULT_OK &&
+				info.spread_spectrum_percentage != 0) {
+			clk_mgr->ss_on_dprefclk = true;
+			clk_mgr->dprefclk_ss_divider = info.spread_percentage_divider;
+
+			if (info.type.CENTER_MODE == 0) {
+				/* Currently for DP Reference clock we
+				 * need only SS percentage for
+				 * downspread
+				 */
+				clk_mgr->dprefclk_ss_percentage =
+						info.spread_spectrum_percentage;
+			}
+		}
+	}
+}
+static void dcn401_notify_wm_ranges(struct clk_mgr *clk_mgr_base)
+{
+	unsigned int i;
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+	WatermarksExternal_t *table = (WatermarksExternal_t *) clk_mgr->wm_range_table;
+
+	if (!clk_mgr->smu_present)
+		return;
+
+	if (!table)
+		return;
+
+	memset(table, 0, sizeof(*table));
+
+	/* collect valid ranges, place in pmfw table */
+	for (i = 0; i < WM_SET_COUNT; i++)
+		if (clk_mgr->base.bw_params->wm_table.nv_entries[i].valid) {
+			table->Watermarks.WatermarkRow[i].WmSetting = i;
+			table->Watermarks.WatermarkRow[i].Flags = clk_mgr->base.bw_params->wm_table.nv_entries[i].pmfw_breakdown.wm_type;
+		}
+	dcn401_smu_set_dram_addr_high(clk_mgr, clk_mgr->wm_range_table_addr >> 32);
+	dcn401_smu_set_dram_addr_low(clk_mgr, clk_mgr->wm_range_table_addr & 0xFFFFFFFF);
+	dcn401_smu_transfer_wm_table_dram_2_smu(clk_mgr);
+}
+
+/* Set min memclk to minimum, either constrained by the current mode or DPM0 */
+static void dcn401_set_hard_min_memclk(struct clk_mgr *clk_mgr_base, bool current_mode)
+{
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+	const struct dc *dc = clk_mgr->base.ctx->dc;
+	struct dc_state *context = dc->current_state;
+	struct dc_clocks new_clocks;
+	int num_steps;
+
+	if (!clk_mgr->smu_present || !dcn401_is_ppclk_dpm_enabled(clk_mgr, PPCLK_UCLK))
+		return;
+
+	/* build clock update */
+	memcpy(&new_clocks, &clk_mgr_base->clks, sizeof(struct dc_clocks));
+
+	if (current_mode) {
+		new_clocks.dramclk_khz = context->bw_ctx.bw.dcn.clk.dramclk_khz;
+		new_clocks.idle_dramclk_khz = context->bw_ctx.bw.dcn.clk.idle_dramclk_khz;
+		new_clocks.p_state_change_support = context->bw_ctx.bw.dcn.clk.p_state_change_support;
+	} else {
+		new_clocks.dramclk_khz = clk_mgr_base->bw_params->clk_table.entries[0].memclk_mhz * 1000;
+		new_clocks.idle_dramclk_khz = new_clocks.dramclk_khz;
+		new_clocks.p_state_change_support = true;
+	}
+
+	num_steps = dcn401_build_update_bandwidth_clocks_sequence(clk_mgr_base,
+			context,
+			&new_clocks,
+			true);
+
+	/* execute sequence */
+	dcn401_execute_block_sequence(clk_mgr_base,	num_steps);
+}
+
+static int dcn401_get_hard_min_memclk(struct clk_mgr *clk_mgr_base)
+{
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+
+	return clk_mgr->base.ctx->dc->current_state->bw_ctx.bw.dcn.clk.dramclk_khz;
+}
+
+static int dcn401_get_hard_min_fclk(struct clk_mgr *clk_mgr_base)
+{
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+
+	return clk_mgr->base.ctx->dc->current_state->bw_ctx.bw.dcn.clk.fclk_khz;
+}
+
+/* Get current memclk states, update bounding box */
+static void dcn401_get_memclk_states_from_smu(struct clk_mgr *clk_mgr_base)
+{
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+	struct clk_limit_num_entries *num_entries_per_clk = &clk_mgr_base->bw_params->clk_table.num_entries_per_clk;
+	unsigned int num_levels;
+
+	if (!clk_mgr->smu_present)
+		return;
+
+	/* Refresh memclk and fclk states */
+	dcn401_init_single_clock(clk_mgr, PPCLK_UCLK,
+			&clk_mgr_base->bw_params->clk_table.entries[0].memclk_mhz,
+			&num_entries_per_clk->num_memclk_levels);
+	if (num_entries_per_clk->num_memclk_levels) {
+		clk_mgr_base->bw_params->max_memclk_mhz =
+				clk_mgr_base->bw_params->clk_table.entries[num_entries_per_clk->num_memclk_levels - 1].memclk_mhz;
+	}
+
+	clk_mgr_base->bw_params->dc_mode_limit.memclk_mhz = dcn401_smu_get_dc_mode_max_dpm_freq(clk_mgr, PPCLK_UCLK);
+	if (num_entries_per_clk->num_memclk_levels && clk_mgr_base->bw_params->dc_mode_limit.memclk_mhz ==
+			clk_mgr_base->bw_params->clk_table.entries[num_entries_per_clk->num_memclk_levels - 1].memclk_mhz)
+		clk_mgr_base->bw_params->dc_mode_limit.memclk_mhz = 0;
+	clk_mgr_base->bw_params->dc_mode_softmax_memclk = clk_mgr_base->bw_params->dc_mode_limit.memclk_mhz;
+
+	dcn401_init_single_clock(clk_mgr, PPCLK_FCLK,
+			&clk_mgr_base->bw_params->clk_table.entries[0].fclk_mhz,
+			&num_entries_per_clk->num_fclk_levels);
+	clk_mgr_base->bw_params->dc_mode_limit.fclk_mhz = dcn401_smu_get_dc_mode_max_dpm_freq(clk_mgr, PPCLK_FCLK);
+	if (num_entries_per_clk->num_fclk_levels && clk_mgr_base->bw_params->dc_mode_limit.fclk_mhz ==
+			clk_mgr_base->bw_params->clk_table.entries[num_entries_per_clk->num_fclk_levels - 1].fclk_mhz)
+		clk_mgr_base->bw_params->dc_mode_limit.fclk_mhz = 0;
+
+	num_levels = max(num_entries_per_clk->num_memclk_levels, num_entries_per_clk->num_fclk_levels);
+
+	clk_mgr_base->bw_params->clk_table.num_entries = num_levels ? num_levels : 1;
+
+	if (clk_mgr->dpm_present && !num_levels)
+		clk_mgr->dpm_present = false;
+
+	clk_mgr_base->bw_params->num_channels = dcn401_smu_get_num_of_umc_channels(clk_mgr);
+	if (clk_mgr_base->ctx->dc_bios) {
+		/* use BIOS values if none provided by PMFW */
+		if (clk_mgr_base->bw_params->num_channels == 0) {
+			clk_mgr_base->bw_params->num_channels = clk_mgr_base->ctx->dc_bios->vram_info.num_chans;
+		}
+		clk_mgr_base->bw_params->dram_channel_width_bytes = clk_mgr_base->ctx->dc_bios->vram_info.dram_channel_width_bytes;
+	}
+
+	/* Refresh bounding box */
+	clk_mgr_base->ctx->dc->res_pool->funcs->update_bw_bounding_box(
+			clk_mgr->base.ctx->dc, clk_mgr_base->bw_params);
+}
+
+static bool dcn401_are_clock_states_equal(struct dc_clocks *a,
+					struct dc_clocks *b)
+{
+	if (a->dispclk_khz != b->dispclk_khz)
+		return false;
+	else if (a->dppclk_khz != b->dppclk_khz)
+		return false;
+	else if (a->dcfclk_khz != b->dcfclk_khz)
+		return false;
+	else if (a->dcfclk_deep_sleep_khz != b->dcfclk_deep_sleep_khz)
+		return false;
+	else if (a->dramclk_khz != b->dramclk_khz)
+		return false;
+	else if (a->p_state_change_support != b->p_state_change_support)
+		return false;
+	else if (a->fclk_p_state_change_support != b->fclk_p_state_change_support)
+		return false;
+
+	return true;
+}
+
+static void dcn401_enable_pme_wa(struct clk_mgr *clk_mgr_base)
+{
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+
+	if (!clk_mgr->smu_present)
+		return;
+
+	dcn401_smu_set_pme_workaround(clk_mgr);
+}
+
+static bool dcn401_is_smu_present(struct clk_mgr *clk_mgr_base)
+{
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+	return clk_mgr->smu_present;
+}
+
+
+static int dcn401_get_dtb_ref_freq_khz(struct clk_mgr *clk_mgr_base)
+{
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+
+	int dtb_ref_clk_khz = 0;
+
+	if (clk_mgr->smu_present && dcn401_is_ppclk_dpm_enabled(clk_mgr, PPCLK_DTBCLK)) {
+		/* DPM enabled, use currently set value */
+		dtb_ref_clk_khz = clk_mgr_base->clks.ref_dtbclk_khz;
+	} else {
+		/* DPM disabled, so use boot snapshot */
+		dtb_ref_clk_khz = clk_mgr_base->boot_snapshot.dtbclk;
+	}
+
+	return dtb_ref_clk_khz;
+}
+
+static int dcn401_get_dispclk_from_dentist(struct clk_mgr *clk_mgr_base)
+{
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+	uint32_t dispclk_wdivider;
+	int disp_divider;
+
+	REG_GET(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_WDIVIDER, &dispclk_wdivider);
+	disp_divider = dentist_get_divider_from_did(dispclk_wdivider);
+
+	/* Return DISPCLK freq in Khz */
+	if (disp_divider)
+		return (DENTIST_DIVIDER_RANGE_SCALE_FACTOR * clk_mgr->base.dentist_vco_freq_khz) / disp_divider;
+
+	return 0;
+}
+
+unsigned int dcn401_get_max_clock_khz(struct clk_mgr *clk_mgr_base, enum clk_type clk_type)
+{
+	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+
+	unsigned int num_clk_levels;
+
+	switch (clk_type) {
+	case CLK_TYPE_DISPCLK:
+		num_clk_levels = clk_mgr->base.bw_params->clk_table.num_entries_per_clk.num_dispclk_levels;
+		return dcn401_is_ppclk_dpm_enabled(clk_mgr, PPCLK_DISPCLK) ?
+				clk_mgr->base.bw_params->clk_table.entries[num_clk_levels - 1].dispclk_mhz * 1000 :
+				clk_mgr->base.boot_snapshot.dispclk;
+	case CLK_TYPE_DPPCLK:
+		num_clk_levels = clk_mgr->base.bw_params->clk_table.num_entries_per_clk.num_dppclk_levels;
+		return dcn401_is_ppclk_dpm_enabled(clk_mgr, PPCLK_DPPCLK) ?
+				clk_mgr->base.bw_params->clk_table.entries[num_clk_levels - 1].dppclk_mhz * 1000 :
+				clk_mgr->base.boot_snapshot.dppclk;
+	case CLK_TYPE_DSCCLK:
+		num_clk_levels = clk_mgr->base.bw_params->clk_table.num_entries_per_clk.num_dispclk_levels;
+		return dcn401_is_ppclk_dpm_enabled(clk_mgr, PPCLK_DISPCLK) ?
+				clk_mgr->base.bw_params->clk_table.entries[num_clk_levels - 1].dispclk_mhz * 1000 / 3 :
+				clk_mgr->base.boot_snapshot.dispclk / 3;
+	default:
+		break;
+	}
+
+	return 0;
+}
+
+static struct clk_mgr_funcs dcn401_funcs = {
+		.get_dp_ref_clk_frequency = dce12_get_dp_ref_freq_khz,
+		.get_dtb_ref_clk_frequency = dcn401_get_dtb_ref_freq_khz,
+		.update_clocks = dcn401_update_clocks,
+		.dump_clk_registers = dcn401_dump_clk_registers,
+		.init_clocks = dcn401_init_clocks,
+		.notify_wm_ranges = dcn401_notify_wm_ranges,
+		.set_hard_min_memclk = dcn401_set_hard_min_memclk,
+		.get_memclk_states_from_smu = dcn401_get_memclk_states_from_smu,
+		.are_clock_states_equal = dcn401_are_clock_states_equal,
+		.enable_pme_wa = dcn401_enable_pme_wa,
+		.is_smu_present = dcn401_is_smu_present,
+		.get_dispclk_from_dentist = dcn401_get_dispclk_from_dentist,
+		.get_hard_min_memclk = dcn401_get_hard_min_memclk,
+		.get_hard_min_fclk = dcn401_get_hard_min_fclk,
+		.is_dc_mode_present = dcn401_is_dc_mode_present,
+		.get_max_clock_khz = dcn401_get_max_clock_khz,
+};
+
+struct clk_mgr_internal *dcn401_clk_mgr_construct(
+		struct dc_context *ctx,
+		struct dccg *dccg)
+{
+	struct clk_log_info log_info = {0};
+	struct dcn401_clk_mgr *clk_mgr401 = kzalloc(sizeof(struct dcn401_clk_mgr), GFP_KERNEL);
+	struct clk_mgr_internal *clk_mgr;
+
+	if (!clk_mgr401)
+		return NULL;
+
+	clk_mgr = &clk_mgr401->base;
+	clk_mgr->base.ctx = ctx;
+	clk_mgr->base.funcs = &dcn401_funcs;
+	clk_mgr->regs = &clk_mgr_regs_dcn401;
+	clk_mgr->clk_mgr_shift = &clk_mgr_shift_dcn401;
+	clk_mgr->clk_mgr_mask = &clk_mgr_mask_dcn401;
+
+	clk_mgr->dccg = dccg;
+	clk_mgr->dfs_bypass_disp_clk = 0;
+
+	clk_mgr->dprefclk_ss_percentage = 0;
+	clk_mgr->dprefclk_ss_divider = 1000;
+	clk_mgr->ss_on_dprefclk = false;
+	clk_mgr->dfs_ref_freq_khz = 100000;
+
+	/* Changed from DCN3.2_clock_frequency doc to match
+	 * dcn401_dump_clk_registers from 4 * dentist_vco_freq_khz /
+	 * dprefclk DID divider
+	 */
+	clk_mgr->base.dprefclk_khz = 720000; //TODO update from VBIOS
+
+	/* integer part is now VCO frequency in kHz */
+	clk_mgr->base.dentist_vco_freq_khz = dcn401_get_vco_frequency_from_reg(clk_mgr);
+
+	/* in case we don't get a value from the register, use default */
+	if (clk_mgr->base.dentist_vco_freq_khz == 0)
+		clk_mgr->base.dentist_vco_freq_khz = 4500000; //TODO Update from VBIOS
+
+	dcn401_dump_clk_registers(&clk_mgr->base.boot_snapshot, &clk_mgr->base, &log_info);
+
+	if (ctx->dc->debug.disable_dtb_ref_clk_switch &&
+			clk_mgr->base.clks.ref_dtbclk_khz != clk_mgr->base.boot_snapshot.dtbclk) {
+		clk_mgr->base.clks.ref_dtbclk_khz = clk_mgr->base.boot_snapshot.dtbclk;
+	}
+
+	if (clk_mgr->base.boot_snapshot.dprefclk != 0) {
+		clk_mgr->base.dprefclk_khz = clk_mgr->base.boot_snapshot.dprefclk;
+	}
+	dcn401_clock_read_ss_info(clk_mgr);
+
+	clk_mgr->dfs_bypass_enabled = false;
+
+	clk_mgr->smu_present = false;
+
+	clk_mgr->base.bw_params = kzalloc(sizeof(*clk_mgr->base.bw_params), GFP_KERNEL);
+	if (!clk_mgr->base.bw_params) {
+		BREAK_TO_DEBUGGER();
+		kfree(clk_mgr401);
+		return NULL;
+	}
+
+	/* need physical address of table to give to PMFW */
+	clk_mgr->wm_range_table = dm_helpers_allocate_gpu_mem(clk_mgr->base.ctx,
+			DC_MEM_ALLOC_TYPE_GART, sizeof(WatermarksExternal_t),
+			&clk_mgr->wm_range_table_addr);
+	if (!clk_mgr->wm_range_table) {
+		BREAK_TO_DEBUGGER();
+		kfree(clk_mgr->base.bw_params);
+		kfree(clk_mgr401);
+		return NULL;
+	}
+
+	return &clk_mgr401->base;
+}
+
+void dcn401_clk_mgr_destroy(struct clk_mgr_internal *clk_mgr)
+{
+	kfree(clk_mgr->base.bw_params);
+
+	if (clk_mgr->wm_range_table)
+		dm_helpers_free_gpu_mem(clk_mgr->base.ctx, DC_MEM_ALLOC_TYPE_GART,
+				clk_mgr->wm_range_table);
+}
+
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr.h b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr.h
new file mode 100644
index 000000000000..97a1ce1e8a9e
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr.h
@@ -0,0 +1,117 @@
+// SPDX-License-Identifier: MIT
+//
+// Copyright 2024 Advanced Micro Devices, Inc.
+
+#ifndef __DCN401_CLK_MGR_H_
+#define __DCN401_CLK_MGR_H_
+
+#define DCN401_CLK_MGR_MAX_SEQUENCE_SIZE 30
+
+union dcn401_clk_mgr_block_sequence_params {
+	struct {
+		/* inputs */
+		uint32_t num_displays;
+	} update_num_displays_params;
+	struct {
+		/* inputs */
+		uint32_t ppclk;
+		uint16_t freq_mhz;
+		/* outputs */
+		uint32_t *response;
+	} update_hardmin_params;
+	struct {
+		/* inputs */
+		uint32_t ppclk;
+		int freq_khz;
+		/* outputs */
+		uint32_t *response;
+	} update_hardmin_optimized_params;
+	struct {
+		/* inputs */
+		uint16_t uclk_mhz;
+		uint16_t fclk_mhz;
+	} update_idle_hardmin_params;
+	struct {
+		/* inputs */
+		uint16_t freq_mhz;
+	} update_deep_sleep_dcfclk_params;
+	struct {
+		/* inputs */
+		bool support;
+	} update_pstate_support_params;
+	struct {
+		/* inputs */
+		unsigned int num_ways;
+	} update_cab_for_uclk_params;
+	struct {
+		/* inputs */
+		bool enable;
+	} update_wait_for_dmub_ack_params;
+	struct {
+		/* inputs */
+		bool mod_drr_for_pstate;
+	} indicate_drr_status_params;
+	struct {
+		/* inputs */
+		struct dc_state *context;
+		int *ref_dppclk_khz;
+		bool safe_to_lower;
+	} update_dppclk_dto_params;
+	struct {
+		/* inputs */
+		struct dc_state *context;
+		int *ref_dtbclk_khz;
+	} update_dtbclk_dto_params;
+	struct {
+		/* inputs */
+		struct dc_state *context;
+	} update_dentist_params;
+	struct {
+		/* inputs */
+		struct dmcu *dmcu;
+		unsigned int wait;
+	} update_psr_wait_loop_params;
+};
+
+enum dcn401_clk_mgr_block_sequence_func {
+	CLK_MGR401_READ_CLOCKS_FROM_DENTIST,
+	CLK_MGR401_UPDATE_NUM_DISPLAYS,
+	CLK_MGR401_UPDATE_HARDMIN_PPCLK,
+	CLK_MGR401_UPDATE_HARDMIN_PPCLK_OPTIMIZED,
+	CLK_MGR401_UPDATE_ACTIVE_HARDMINS,
+	CLK_MGR401_UPDATE_IDLE_HARDMINS,
+	CLK_MGR401_UPDATE_DEEP_SLEEP_DCFCLK,
+	CLK_MGR401_UPDATE_FCLK_PSTATE_SUPPORT,
+	CLK_MGR401_UPDATE_UCLK_PSTATE_SUPPORT,
+	CLK_MGR401_UPDATE_CAB_FOR_UCLK,
+	CLK_MGR401_UPDATE_WAIT_FOR_DMUB_ACK,
+	CLK_MGR401_INDICATE_DRR_STATUS,
+	CLK_MGR401_UPDATE_DPPCLK_DTO,
+	CLK_MGR401_UPDATE_DTBCLK_DTO,
+	CLK_MGR401_UPDATE_DENTIST,
+	CLK_MGR401_UPDATE_PSR_WAIT_LOOP,
+	CLK_MGR401_UPDATE_SUBVP_HARDMINS,
+};
+
+struct dcn401_clk_mgr_block_sequence {
+	union dcn401_clk_mgr_block_sequence_params params;
+	enum dcn401_clk_mgr_block_sequence_func func;
+};
+
+struct dcn401_clk_mgr {
+	struct clk_mgr_internal base;
+
+	struct dcn401_clk_mgr_block_sequence block_sequence[DCN401_CLK_MGR_MAX_SEQUENCE_SIZE];
+};
+
+void dcn401_init_clocks(struct clk_mgr *clk_mgr_base);
+bool dcn401_is_dc_mode_present(struct clk_mgr *clk_mgr_base);
+
+struct clk_mgr_internal *dcn401_clk_mgr_construct(struct dc_context *ctx,
+		struct dccg *dccg);
+
+void dcn401_clk_mgr_destroy(struct clk_mgr_internal *clk_mgr);
+
+unsigned int dcn401_get_max_clock_khz(struct clk_mgr *clk_mgr_base, enum clk_type clk_type);
+
+#endif /* __DCN401_CLK_MGR_H_ */
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr_smu_msg.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr_smu_msg.c
new file mode 100644
index 000000000000..3a263840893e
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr_smu_msg.c
@@ -0,0 +1,472 @@
+// SPDX-License-Identifier: MIT
+//
+// Copyright 2024 Advanced Micro Devices, Inc.
+
+#include "dcn401_clk_mgr_smu_msg.h"
+
+#include "clk_mgr_internal.h"
+#include "reg_helper.h"
+
+#include "dalsmc.h"
+#include "dcn401_smu14_driver_if.h"
+
+#define mmDAL_MSG_REG  0x1628A
+#define mmDAL_ARG_REG  0x16273
+#define mmDAL_RESP_REG 0x16274
+
+#define REG(reg_name) \
+	mm ## reg_name
+
+#include "logger_types.h"
+
+#define smu_print(str, ...) {DC_LOG_SMU(str, ##__VA_ARGS__); }
+
+/* temporary define */
+#ifndef DALSMC_MSG_SubvpUclkFclk
+#define DALSMC_MSG_SubvpUclkFclk 0x1B
+#endif
+#ifndef DALSMC_MSG_GetNumUmcChannels
+#define DALSMC_MSG_GetNumUmcChannels 0x1C
+#endif
+
+/*
+ * Function to be used instead of REG_WAIT macro because the wait ends when
+ * the register is NOT EQUAL to zero, and because the translation in msg_if.h
+ * won't work with REG_WAIT.
+ */
+static uint32_t dcn401_smu_wait_for_response(struct clk_mgr_internal *clk_mgr, unsigned int delay_us, unsigned int max_retries)
+{
+	uint32_t reg = 0;
+
+	do {
+		reg = REG_READ(DAL_RESP_REG);
+		if (reg)
+			break;
+
+		if (delay_us >= 1000)
+			msleep(delay_us/1000);
+		else if (delay_us > 0)
+			udelay(delay_us);
+	} while (max_retries--);
+
+	return reg;
+}
+
+static bool dcn401_smu_send_msg_with_param(struct clk_mgr_internal *clk_mgr, uint32_t msg_id, uint32_t param_in, uint32_t *param_out)
+{
+	/* Wait for response register to be ready */
+	dcn401_smu_wait_for_response(clk_mgr, 10, 200000);
+
+	TRACE_SMU_MSG_ENTER(msg_id, param_in, clk_mgr->base.ctx);
+
+	/* Clear response register */
+	REG_WRITE(DAL_RESP_REG, 0);
+
+	/* Set the parameter register for the SMU message */
+	REG_WRITE(DAL_ARG_REG, param_in);
+
+	/* Trigger the message transaction by writing the message ID */
+	REG_WRITE(DAL_MSG_REG, msg_id);
+
+	/* Wait for response */
+	if (dcn401_smu_wait_for_response(clk_mgr, 10, 200000) == DALSMC_Result_OK) {
+		if (param_out)
+			*param_out = REG_READ(DAL_ARG_REG);
+
+		TRACE_SMU_MSG_EXIT(true, param_out ? *param_out : 0, clk_mgr->base.ctx);
+		return true;
+	}
+
+	TRACE_SMU_MSG_EXIT(false, 0, clk_mgr->base.ctx);
+	return false;
+}
+
+/*
+ * Use these functions to return back delay information so we can aggregate the total
+ *  delay when requesting hardmin clk
+ *
+ * dcn401_smu_wait_for_response_delay
+ * dcn401_smu_send_msg_with_param_delay
+ *
+ */
+static uint32_t dcn401_smu_wait_for_response_delay(struct clk_mgr_internal *clk_mgr, unsigned int delay_us, unsigned int max_retries, unsigned int *total_delay_us)
+{
+	uint32_t reg = 0;
+	*total_delay_us = 0;
+
+	do {
+		reg = REG_READ(DAL_RESP_REG);
+		if (reg)
+			break;
+
+		if (delay_us >= 1000)
+			msleep(delay_us/1000);
+		else if (delay_us > 0)
+			udelay(delay_us);
+		*total_delay_us += delay_us;
+	} while (max_retries--);
+
+	return reg;
+}
+
+static bool dcn401_smu_send_msg_with_param_delay(struct clk_mgr_internal *clk_mgr, uint32_t msg_id, uint32_t param_in, uint32_t *param_out, unsigned int *total_delay_us)
+{
+	unsigned int delay1_us, delay2_us;
+	*total_delay_us = 0;
+
+	/* Wait for response register to be ready */
+	dcn401_smu_wait_for_response_delay(clk_mgr, 10, 200000, &delay1_us);
+
+	TRACE_SMU_MSG_ENTER(msg_id, param_in, clk_mgr->base.ctx);
+
+	/* Clear response register */
+	REG_WRITE(DAL_RESP_REG, 0);
+
+	/* Set the parameter register for the SMU message */
+	REG_WRITE(DAL_ARG_REG, param_in);
+
+	/* Trigger the message transaction by writing the message ID */
+	REG_WRITE(DAL_MSG_REG, msg_id);
+
+	/* Wait for response */
+	if (dcn401_smu_wait_for_response_delay(clk_mgr, 10, 200000, &delay2_us) == DALSMC_Result_OK) {
+		if (param_out)
+			*param_out = REG_READ(DAL_ARG_REG);
+
+		*total_delay_us = delay1_us + delay2_us;
+		TRACE_SMU_MSG_EXIT(true, param_out ? *param_out : 0, clk_mgr->base.ctx);
+		return true;
+	}
+
+	*total_delay_us = delay1_us + 2000000;
+	TRACE_SMU_MSG_EXIT(false, 0, clk_mgr->base.ctx);
+	return false;
+}
+
+bool dcn401_smu_get_smu_version(struct clk_mgr_internal *clk_mgr, unsigned int *version)
+{
+	smu_print("SMU Get SMU version\n");
+
+	if (dcn401_smu_send_msg_with_param(clk_mgr,
+			DALSMC_MSG_GetSmuVersion, 0, version)) {
+
+		smu_print("SMU version: %d\n", *version);
+
+		return true;
+	}
+
+	return false;
+}
+
+/* Message output should match SMU11_DRIVER_IF_VERSION in smu11_driver_if.h */
+bool dcn401_smu_check_driver_if_version(struct clk_mgr_internal *clk_mgr)
+{
+	uint32_t response = 0;
+
+	smu_print("SMU Check driver if version\n");
+
+	if (dcn401_smu_send_msg_with_param(clk_mgr,
+			DALSMC_MSG_GetDriverIfVersion, 0, &response)) {
+
+		smu_print("SMU driver if version: %d\n", response);
+
+		if (response == SMU14_DRIVER_IF_VERSION)
+			return true;
+	}
+
+	return false;
+}
+
+/* Message output should match DALSMC_VERSION in dalsmc.h */
+bool dcn401_smu_check_msg_header_version(struct clk_mgr_internal *clk_mgr)
+{
+	uint32_t response = 0;
+
+	smu_print("SMU Check msg header version\n");
+
+	if (dcn401_smu_send_msg_with_param(clk_mgr,
+			DALSMC_MSG_GetMsgHeaderVersion, 0, &response)) {
+
+		smu_print("SMU msg header version: %d\n", response);
+
+		if (response == DALSMC_VERSION)
+			return true;
+	}
+
+	return false;
+}
+
+void dcn401_smu_send_fclk_pstate_message(struct clk_mgr_internal *clk_mgr, bool support)
+{
+	smu_print("FCLK P-state support value is : %d\n", support);
+
+	dcn401_smu_send_msg_with_param(clk_mgr,
+			DALSMC_MSG_SetFclkSwitchAllow, support, NULL);
+}
+
+void dcn401_smu_send_uclk_pstate_message(struct clk_mgr_internal *clk_mgr, bool support)
+{
+	smu_print("UCLK P-state support value is : %d\n", support);
+
+	dcn401_smu_send_msg_with_param(clk_mgr,
+			DALSMC_MSG_SetUclkPstateAllow, support, NULL);
+}
+
+void dcn401_smu_send_cab_for_uclk_message(struct clk_mgr_internal *clk_mgr, unsigned int num_ways)
+{
+	uint32_t param = (num_ways << 1) | (num_ways > 0);
+
+	dcn401_smu_send_msg_with_param(clk_mgr, DALSMC_MSG_SetCabForUclkPstate, param, NULL);
+	smu_print("Numways for SubVP : %d\n", num_ways);
+}
+
+void dcn401_smu_set_dram_addr_high(struct clk_mgr_internal *clk_mgr, uint32_t addr_high)
+{
+	smu_print("SMU Set DRAM addr high: %d\n", addr_high);
+
+	dcn401_smu_send_msg_with_param(clk_mgr,
+			DALSMC_MSG_SetDalDramAddrHigh, addr_high, NULL);
+}
+
+void dcn401_smu_set_dram_addr_low(struct clk_mgr_internal *clk_mgr, uint32_t addr_low)
+{
+	smu_print("SMU Set DRAM addr low: %d\n", addr_low);
+
+	dcn401_smu_send_msg_with_param(clk_mgr,
+			DALSMC_MSG_SetDalDramAddrLow, addr_low, NULL);
+}
+
+void dcn401_smu_transfer_wm_table_dram_2_smu(struct clk_mgr_internal *clk_mgr)
+{
+	smu_print("SMU Transfer WM table DRAM 2 SMU\n");
+
+	dcn401_smu_send_msg_with_param(clk_mgr,
+			DALSMC_MSG_TransferTableDram2Smu, TABLE_WATERMARKS, NULL);
+}
+
+void dcn401_smu_set_pme_workaround(struct clk_mgr_internal *clk_mgr)
+{
+	smu_print("SMU Set PME workaround\n");
+
+	dcn401_smu_send_msg_with_param(clk_mgr,
+		DALSMC_MSG_BacoAudioD3PME, 0, NULL);
+}
+
+static unsigned int dcn401_smu_get_hard_min_status(struct clk_mgr_internal *clk_mgr, bool *no_timeout, unsigned int *total_delay_us)
+{
+	uint32_t response = 0;
+
+	/* bits 23:16 for clock type, lower 16 bits for frequency in MHz */
+	uint32_t param = 0;
+
+	*no_timeout = dcn401_smu_send_msg_with_param_delay(clk_mgr,
+			DALSMC_MSG_ReturnHardMinStatus, param, &response, total_delay_us);
+
+	smu_print("SMU Get hard min status: no_timeout %d delay %d us clk bits %x\n",
+		*no_timeout, *total_delay_us, response);
+
+	return response;
+}
+
+static bool dcn401_smu_wait_hard_min_status(struct clk_mgr_internal *clk_mgr, uint32_t ppclk)
+{
+	const unsigned int max_delay_us = 1000000;
+
+	unsigned int hardmin_status_mask = (1 << ppclk);
+	unsigned int total_delay_us = 0;
+	bool hardmin_done = false;
+
+	while (!hardmin_done && total_delay_us < max_delay_us) {
+		unsigned int hardmin_status;
+		unsigned int read_total_delay_us;
+		bool no_timeout;
+
+		if (!hardmin_done && total_delay_us > 0) {
+			/* hardmin not yet fulfilled, wait 500us and retry*/
+			udelay(500);
+			total_delay_us += 500;
+
+			smu_print("SMU Wait hard min status for %d us\n", total_delay_us);
+		}
+
+		hardmin_status = dcn401_smu_get_hard_min_status(clk_mgr, &no_timeout, &read_total_delay_us);
+		total_delay_us += read_total_delay_us;
+		hardmin_done = hardmin_status & hardmin_status_mask;
+	}
+
+	return hardmin_done;
+}
+
+/* Returns the actual frequency that was set in MHz, 0 on failure */
+unsigned int dcn401_smu_set_hard_min_by_freq(struct clk_mgr_internal *clk_mgr, uint32_t clk, uint16_t freq_mhz)
+{
+	uint32_t response = 0;
+	bool hard_min_done = false;
+
+	/* bits 23:16 for clock type, lower 16 bits for frequency in MHz */
+	uint32_t param = (clk << 16) | freq_mhz;
+
+	smu_print("SMU Set hard min by freq: clk = %d, freq_mhz = %d MHz\n", clk, freq_mhz);
+
+	dcn401_smu_send_msg_with_param(clk_mgr,
+			DALSMC_MSG_SetHardMinByFreq, param, &response);
+
+	/* wait until hardmin acknowledged */
+	hard_min_done = dcn401_smu_wait_hard_min_status(clk_mgr, clk);
+	smu_print("SMU Frequency set = %d KHz hard_min_done %d\n", response, hard_min_done);
+
+	return response;
+}
+
+void dcn401_smu_wait_for_dmub_ack_mclk(struct clk_mgr_internal *clk_mgr, bool enable)
+{
+	smu_print("SMU to wait for DMCUB ack for MCLK : %d\n", enable);
+
+	dcn401_smu_send_msg_with_param(clk_mgr, DALSMC_MSG_SetAlwaysWaitDmcubResp, enable ? 1 : 0, NULL);
+}
+
+void dcn401_smu_indicate_drr_status(struct clk_mgr_internal *clk_mgr, bool mod_drr_for_pstate)
+{
+	smu_print("SMU Set indicate drr status = %d\n", mod_drr_for_pstate);
+
+	dcn401_smu_send_msg_with_param(clk_mgr,
+			DALSMC_MSG_IndicateDrrStatus, mod_drr_for_pstate ? 1 : 0, NULL);
+}
+
+bool dcn401_smu_set_idle_uclk_fclk_hardmin(struct clk_mgr_internal *clk_mgr,
+		uint16_t uclk_freq_mhz,
+		uint16_t fclk_freq_mhz)
+{
+	uint32_t response = 0;
+	bool success;
+
+	/* 15:0 for uclk, 32:16 for fclk */
+	uint32_t param = (fclk_freq_mhz << 16) | uclk_freq_mhz;
+
+	smu_print("SMU Set idle hardmin by freq: uclk_freq_mhz = %d MHz, fclk_freq_mhz = %d MHz\n", uclk_freq_mhz, fclk_freq_mhz);
+
+	success = dcn401_smu_send_msg_with_param(clk_mgr,
+			DALSMC_MSG_IdleUclkFclk, param, &response);
+
+	/* wait until hardmin acknowledged */
+	success &= dcn401_smu_wait_hard_min_status(clk_mgr, PPCLK_UCLK);
+	smu_print("SMU hard_min_done %d\n", success);
+
+	return success;
+}
+
+bool dcn401_smu_set_active_uclk_fclk_hardmin(struct clk_mgr_internal *clk_mgr,
+		uint16_t uclk_freq_mhz,
+		uint16_t fclk_freq_mhz)
+{
+	uint32_t response = 0;
+	bool success;
+
+	/* 15:0 for uclk, 32:16 for fclk */
+	uint32_t param = (fclk_freq_mhz << 16) | uclk_freq_mhz;
+
+	smu_print("SMU Set active hardmin by freq: uclk_freq_mhz = %d MHz, fclk_freq_mhz = %d MHz\n", uclk_freq_mhz, fclk_freq_mhz);
+
+	success = dcn401_smu_send_msg_with_param(clk_mgr,
+			DALSMC_MSG_ActiveUclkFclk, param, &response);
+
+	/* wait until hardmin acknowledged */
+	success &= dcn401_smu_wait_hard_min_status(clk_mgr, PPCLK_UCLK);
+	smu_print("SMU hard_min_done %d\n", success);
+
+	return success;
+}
+
+bool dcn401_smu_set_subvp_uclk_fclk_hardmin(struct clk_mgr_internal *clk_mgr,
+		uint16_t uclk_freq_mhz,
+		uint16_t fclk_freq_mhz)
+{
+	uint32_t response = 0;
+	bool success;
+
+	/* 15:0 for uclk, 32:16 for fclk */
+	uint32_t param = (fclk_freq_mhz << 16) | uclk_freq_mhz;
+
+	smu_print("SMU Set active hardmin by freq: uclk_freq_mhz = %d MHz, fclk_freq_mhz = %d MHz\n", uclk_freq_mhz, fclk_freq_mhz);
+
+	success = dcn401_smu_send_msg_with_param(clk_mgr,
+			DALSMC_MSG_SubvpUclkFclk, param, &response);
+
+	return success;
+}
+
+void dcn401_smu_set_min_deep_sleep_dcef_clk(struct clk_mgr_internal *clk_mgr, uint32_t freq_mhz)
+{
+	smu_print("SMU Set min deep sleep dcef clk: freq_mhz = %d MHz\n", freq_mhz);
+
+	dcn401_smu_send_msg_with_param(clk_mgr,
+			DALSMC_MSG_SetMinDeepSleepDcfclk, freq_mhz, NULL);
+}
+
+void dcn401_smu_set_num_of_displays(struct clk_mgr_internal *clk_mgr, uint32_t num_displays)
+{
+	smu_print("SMU Set num of displays: num_displays = %d\n", num_displays);
+
+	dcn401_smu_send_msg_with_param(clk_mgr,
+			DALSMC_MSG_NumOfDisplays, num_displays, NULL);
+}
+
+unsigned int dcn401_smu_get_num_of_umc_channels(struct clk_mgr_internal *clk_mgr)
+{
+	unsigned int response = 0;
+
+	dcn401_smu_send_msg_with_param(clk_mgr, DALSMC_MSG_GetNumUmcChannels, 0, &response);
+
+	smu_print("SMU Get Num UMC Channels: num_umc_channels = %d\n", response);
+
+	return response;
+}
+
+/*
+ * Frequency in MHz returned in lower 16 bits for valid DPM level
+ *
+ * Call with dpm_level = 0xFF to query features, return value will be:
+ *     Bits 7:0 - number of DPM levels
+ *     Bit   28 - 1 = auto DPM on
+ *     Bit   29 - 1 = sweep DPM on
+ *     Bit   30 - 1 = forced DPM on
+ *     Bit   31 - 0 = discrete, 1 = fine-grained
+ *
+ * With fine-grained DPM, only min and max frequencies will be reported
+ *
+ * Returns 0 on failure
+ */
+unsigned int dcn401_smu_get_dpm_freq_by_index(struct clk_mgr_internal *clk_mgr, uint32_t clk, uint8_t dpm_level)
+{
+	uint32_t response = 0;
+
+	/* bits 23:16 for clock type, lower 8 bits for DPM level */
+	uint32_t param = (clk << 16) | dpm_level;
+
+	smu_print("SMU Get dpm freq by index: clk = %d, dpm_level = %d\n", clk, dpm_level);
+
+	dcn401_smu_send_msg_with_param(clk_mgr,
+			DALSMC_MSG_GetDpmFreqByIndex, param, &response);
+
+	smu_print("SMU dpm freq: %d MHz\n", response);
+
+	return response;
+}
+
+/* Returns the max DPM frequency in DC mode in MHz, 0 on failure */
+unsigned int dcn401_smu_get_dc_mode_max_dpm_freq(struct clk_mgr_internal *clk_mgr, uint32_t clk)
+{
+	uint32_t response = 0;
+
+	/* bits 23:16 for clock type */
+	uint32_t param = clk << 16;
+
+	smu_print("SMU Get DC mode max DPM freq: clk = %d\n", clk);
+
+	dcn401_smu_send_msg_with_param(clk_mgr,
+			DALSMC_MSG_GetDcModeMaxDpmFreq, param, &response);
+
+	smu_print("SMU DC mode max DMP freq: %d MHz\n", response);
+
+	return response;
+}
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr_smu_msg.h b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr_smu_msg.h
new file mode 100644
index 000000000000..4f5ac603e822
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr_smu_msg.h
@@ -0,0 +1,41 @@
+// SPDX-License-Identifier: MIT
+//
+// Copyright 2024 Advanced Micro Devices, Inc.
+
+#ifndef __DCN401_CLK_MGR_SMU_MSG_H_
+#define __DCN401_CLK_MGR_SMU_MSG_H_
+
+#include "os_types.h"
+#include "core_types.h"
+
+struct clk_mgr_internal;
+
+bool dcn401_smu_get_smu_version(struct clk_mgr_internal *clk_mgr, unsigned int *version);
+bool dcn401_smu_check_driver_if_version(struct clk_mgr_internal *clk_mgr);
+bool dcn401_smu_check_msg_header_version(struct clk_mgr_internal *clk_mgr);
+void dcn401_smu_send_fclk_pstate_message(struct clk_mgr_internal *clk_mgr, bool support);
+void dcn401_smu_send_uclk_pstate_message(struct clk_mgr_internal *clk_mgr, bool support);
+void dcn401_smu_send_cab_for_uclk_message(struct clk_mgr_internal *clk_mgr, unsigned int num_ways);
+void dcn401_smu_set_dram_addr_high(struct clk_mgr_internal *clk_mgr, uint32_t addr_high);
+void dcn401_smu_set_dram_addr_low(struct clk_mgr_internal *clk_mgr, uint32_t addr_low);
+void dcn401_smu_transfer_wm_table_dram_2_smu(struct clk_mgr_internal *clk_mgr);
+void dcn401_smu_set_pme_workaround(struct clk_mgr_internal *clk_mgr);
+unsigned int dcn401_smu_set_hard_min_by_freq(struct clk_mgr_internal *clk_mgr, uint32_t clk, uint16_t freq_mhz);
+void dcn401_smu_wait_for_dmub_ack_mclk(struct clk_mgr_internal *clk_mgr, bool enable);
+void dcn401_smu_indicate_drr_status(struct clk_mgr_internal *clk_mgr, bool mod_drr_for_pstate);
+bool dcn401_smu_set_idle_uclk_fclk_hardmin(struct clk_mgr_internal *clk_mgr,
+		uint16_t uclk_freq_mhz,
+		uint16_t fclk_freq_mhz);
+bool dcn401_smu_set_active_uclk_fclk_hardmin(struct clk_mgr_internal *clk_mgr,
+		uint16_t uclk_freq_mhz,
+		uint16_t fclk_freq_mhz);
+bool dcn401_smu_set_subvp_uclk_fclk_hardmin(struct clk_mgr_internal *clk_mgr,
+		uint16_t uclk_freq_mhz,
+		uint16_t fclk_freq_mhz);
+void dcn401_smu_set_min_deep_sleep_dcef_clk(struct clk_mgr_internal *clk_mgr, uint32_t freq_mhz);
+void dcn401_smu_set_num_of_displays(struct clk_mgr_internal *clk_mgr, uint32_t num_displays);
+unsigned int dcn401_smu_get_num_of_umc_channels(struct clk_mgr_internal *clk_mgr);
+unsigned int dcn401_smu_get_dc_mode_max_dpm_freq(struct clk_mgr_internal *clk_mgr, uint32_t clk);
+unsigned int dcn401_smu_get_dpm_freq_by_index(struct clk_mgr_internal *clk_mgr, uint32_t clk, uint8_t dpm_level);
+
+#endif /* __DCN401_CLK_MGR_SMU_MSG_H_ */
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_smu14_driver_if.h b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_smu14_driver_if.h
new file mode 100644
index 000000000000..36034b32870c
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_smu14_driver_if.h
@@ -0,0 +1,66 @@
+// SPDX-License-Identifier: MIT
+//
+// Copyright 2024 Advanced Micro Devices, Inc.
+//
+// This is a stripped-down version of the smu13_driver_if.h file for the relevant DAL interfaces.
+
+#define SMU14_DRIVER_IF_VERSION  0x1
+
+//Only Clks that have DPM descriptors are listed here
+typedef enum {
+  PPCLK_GFXCLK = 0,
+  PPCLK_SOCCLK,
+  PPCLK_UCLK,
+  PPCLK_FCLK,
+  PPCLK_DCLK_0,
+  PPCLK_VCLK_0,
+  PPCLK_DISPCLK,
+  PPCLK_DPPCLK,
+  PPCLK_DPREFCLK,
+  PPCLK_DCFCLK,
+  PPCLK_DTBCLK,
+  PPCLK_COUNT,
+} PPCLK_e;
+
+typedef struct {
+  uint8_t  WmSetting;
+  uint8_t  Flags;
+  uint8_t  Padding[2];
+
+} WatermarkRowGeneric_t;
+
+#define NUM_WM_RANGES 4
+
+typedef enum {
+  WATERMARKS_CLOCK_RANGE = 0,
+  WATERMARKS_DUMMY_PSTATE,
+  WATERMARKS_MALL,
+  WATERMARKS_COUNT,
+} WATERMARKS_FLAGS_e;
+
+typedef struct {
+  // Watermarks
+  WatermarkRowGeneric_t WatermarkRow[NUM_WM_RANGES];
+} Watermarks_t;
+
+typedef struct {
+  Watermarks_t Watermarks;
+  uint32_t  Spare[16];
+
+  uint32_t     MmHubPadding[8]; // SMU internal use
+} WatermarksExternal_t;
+
+// Table types
+#define TABLE_PMFW_PPTABLE            0
+#define TABLE_COMBO_PPTABLE           1
+#define TABLE_WATERMARKS              2
+#define TABLE_AVFS_PSM_DEBUG          3
+#define TABLE_PMSTATUSLOG             4
+#define TABLE_SMU_METRICS             5
+#define TABLE_DRIVER_SMU_CONFIG       6
+#define TABLE_ACTIVITY_MONITOR_COEFF  7
+#define TABLE_OVERDRIVE               8
+#define TABLE_I2C_COMMANDS            9
+#define TABLE_DRIVER_INFO             10
+#define TABLE_ECCINFO                 11
+#define TABLE_COUNT                   12
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc.c b/drivers/gpu/drm/amd/display/dc/core/dc.c
index 0ded4decee05..8be9cbd43e18 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc.c
@@ -22,11 +22,10 @@
  * Authors: AMD
  */
 
-#include <linux/slab.h>
-#include <linux/mm.h>
-
 #include "dm_services.h"
 
+#include "amdgpu.h"
+
 #include "dc.h"
 
 #include "core_status.h"
@@ -35,7 +34,13 @@
 #include "dce/dce_hwseq.h"
 
 #include "resource.h"
+#include "dc_state.h"
+#include "dc_state_priv.h"
+#include "dc_plane.h"
+#include "dc_plane_priv.h"
+#include "dc_stream_priv.h"
 
+#include "gpio_service_interface.h"
 #include "clk_mgr.h"
 #include "clock_source.h"
 #include "dc_bios_types.h"
@@ -55,12 +60,10 @@
 #include "link_encoder.h"
 #include "link_enc_cfg.h"
 
-#include "dc_link.h"
-#include "dc_link_ddc.h"
+#include "link_service.h"
 #include "dm_helpers.h"
 #include "mem_input.h"
 
-#include "dc_link_dp.h"
 #include "dc_dmub_srv.h"
 
 #include "dsc.h"
@@ -71,11 +74,21 @@
 
 #include "dmub/dmub_srv.h"
 
-#include "i2caux_interface.h"
+#include "dce/dmub_psr.h"
+
 #include "dce/dmub_hw_lock_mgr.h"
 
 #include "dc_trace.h"
 
+#include "hw_sequencer_private.h"
+
+#if defined(CONFIG_DRM_AMD_DC_FP)
+#include "dml2_0/dml2_internal_types.h"
+#include "soc_and_ip_translator.h"
+#endif
+
+#include "dce/dmub_outbox.h"
+
 #define CTX \
 	dc->ctx
 
@@ -133,14 +146,18 @@ static const char DC_BUILD_ID[] = "production-build";
  * one or two (in the pipe-split case).
  */
 
-/*******************************************************************************
- * Private functions
- ******************************************************************************/
+/* Private functions */
 
-static inline void elevate_update_type(enum surface_update_type *original, enum surface_update_type new)
+static inline void elevate_update_type(
+		struct surface_update_descriptor *descriptor,
+		enum surface_update_type new_type,
+		enum dc_lock_descriptor new_locks
+)
 {
-	if (new > *original)
-		*original = new;
+	if (new_type > descriptor->update_type)
+		descriptor->update_type = new_type;
+
+	descriptor->lock_descriptor |= new_locks;
 }
 
 static void destroy_links(struct dc *dc)
@@ -149,7 +166,7 @@ static void destroy_links(struct dc *dc)
 
 	for (i = 0; i < dc->link_count; i++) {
 		if (NULL != dc->links[i])
-			link_destroy(&dc->links[i]);
+			dc->link_srv->destroy_link(&dc->links[i]);
 	}
 }
 
@@ -207,10 +224,24 @@ static bool create_links(
 		connectors_num,
 		num_virtual_links);
 
-	for (i = 0; i < connectors_num; i++) {
+	/* When getting the number of connectors, the VBIOS reports the number of valid indices,
+	 * but it doesn't say which indices are valid, and not every index has an actual connector.
+	 * So, if we don't find a connector on an index, that is not an error.
+	 *
+	 * - There is no guarantee that the first N indices will be valid
+	 * - VBIOS may report a higher amount of valid indices than there are actual connectors
+	 * - Some VBIOS have valid configurations for more connectors than there actually are
+	 *   on the card. This may be because the manufacturer used the same VBIOS for different
+	 *   variants of the same card.
+	 */
+	for (i = 0; dc->link_count < connectors_num && i < MAX_LINKS; i++) {
+		struct graphics_object_id connector_id = bios->funcs->get_connector_id(bios, i);
 		struct link_init_data link_init_params = {0};
 		struct dc_link *link;
 
+		if (connector_id.id == CONNECTOR_ID_UNKNOWN)
+			continue;
+
 		DC_LOG_DC("BIOS object table - printing link object info for connector number: %d, link_index: %d", i, dc->link_count);
 
 		link_init_params.ctx = dc->ctx;
@@ -218,18 +249,19 @@ static bool create_links(
 		link_init_params.connector_index = i;
 		link_init_params.link_index = dc->link_count;
 		link_init_params.dc = dc;
-		link = link_create(&link_init_params);
+		link = dc->link_srv->create_link(&link_init_params);
 
 		if (link) {
-				dc->links[dc->link_count] = link;
-				link->dc = dc;
-				++dc->link_count;
+			dc->links[dc->link_count] = link;
+			link->dc = dc;
+			++dc->link_count;
 		}
 	}
 
 	DC_LOG_DC("BIOS object table - end");
 
 	/* Create a link for each usb4 dpia port */
+	dc->lowest_dpia_link_index = MAX_LINKS;
 	for (i = 0; i < dc->res_pool->usb4_dpia_count; i++) {
 		struct link_init_data link_init_params = {0};
 		struct dc_link *link;
@@ -240,8 +272,11 @@ static bool create_links(
 		link_init_params.dc = dc;
 		link_init_params.is_dpia_link = true;
 
-		link = link_create(&link_init_params);
+		link = dc->link_srv->create_link(&link_init_params);
 		if (link) {
+			if (dc->lowest_dpia_link_index > dc->link_count)
+				dc->lowest_dpia_link_index = dc->link_count;
+
 			dc->links[dc->link_count] = link;
 			link->dc = dc;
 			++dc->link_count;
@@ -267,6 +302,8 @@ static bool create_links(
 		link->link_id.type = OBJECT_TYPE_CONNECTOR;
 		link->link_id.id = CONNECTOR_ID_VIRTUAL;
 		link->link_id.enum_id = ENUM_ID_1;
+		link->psr_settings.psr_version = DC_PSR_VERSION_UNSUPPORTED;
+		link->replay_settings.config.replay_version = DC_REPLAY_VERSION_UNSUPPORTED;
 		link->link_enc = kzalloc(sizeof(*link->link_enc), GFP_KERNEL);
 
 		if (!link->link_enc) {
@@ -274,24 +311,6 @@ static bool create_links(
 			goto failed_alloc;
 		}
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-		if (IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment) &&
-				dc->caps.dp_hpo &&
-				link->dc->res_pool->res_cap->num_hpo_dp_link_encoder > 0) {
-			/* FPGA case - Allocate HPO DP link encoder */
-			if (i < link->dc->res_pool->res_cap->num_hpo_dp_link_encoder) {
-				link->hpo_dp_link_enc = link->dc->res_pool->hpo_dp_link_enc[i];
-
-				if (link->hpo_dp_link_enc == NULL) {
-					BREAK_TO_DEBUGGER();
-					goto failed_alloc;
-				}
-				link->hpo_dp_link_enc->hpd_source = link->link_enc->hpd_source;
-				link->hpo_dp_link_enc->transmitter = link->link_enc->transmitter;
-			}
-		}
-#endif
-
 		link->link_status.dpcd_caps = &link->dpcd_caps;
 
 		enc_init.ctx = dc->ctx;
@@ -360,10 +379,16 @@ static bool create_link_encoders(struct dc *dc)
  */
 static void destroy_link_encoders(struct dc *dc)
 {
-	unsigned int num_usb4_dpia = dc->res_pool->res_cap->num_usb4_dpia;
-	unsigned int num_dig_link_enc = dc->res_pool->res_cap->num_dig_link_enc;
+	unsigned int num_usb4_dpia;
+	unsigned int num_dig_link_enc;
 	int i;
 
+	if (!dc->res_pool)
+		return;
+
+	num_usb4_dpia = dc->res_pool->res_cap->num_usb4_dpia;
+	num_dig_link_enc = dc->res_pool->res_cap->num_dig_link_enc;
+
 	/* A platform without USB4 DPIA endpoints has a fixed mapping between DIG
 	 * link encoders and physical display endpoints and does not require
 	 * additional link encoder objects.
@@ -393,29 +418,79 @@ static void dc_perf_trace_destroy(struct dc_perf_trace **perf_trace)
 	*perf_trace = NULL;
 }
 
+static bool set_long_vtotal(struct dc *dc, struct dc_stream_state *stream, struct dc_crtc_timing_adjust *adjust)
+{
+	if (!dc || !stream || !adjust)
+		return false;
+
+	if (!dc->current_state)
+		return false;
+
+	int i;
+
+	for (i = 0; i < MAX_PIPES; i++) {
+		struct pipe_ctx *pipe = &dc->current_state->res_ctx.pipe_ctx[i];
+
+		if (pipe->stream == stream && pipe->stream_res.tg) {
+			if (dc->hwss.set_long_vtotal)
+				dc->hwss.set_long_vtotal(&pipe, 1, adjust->v_total_min, adjust->v_total_max);
+
+			return true;
+		}
+	}
+
+	return false;
+}
+
 /**
- *  dc_stream_adjust_vmin_vmax:
+ *  dc_stream_adjust_vmin_vmax - look up pipe context & update parts of DRR
+ *  @dc:     dc reference
+ *  @stream: Initial dc stream state
+ *  @adjust: Updated parameters for vertical_total_min and vertical_total_max
  *
  *  Looks up the pipe context of dc_stream_state and updates the
  *  vertical_total_min and vertical_total_max of the DRR, Dynamic Refresh
  *  Rate, which is a power-saving feature that targets reducing panel
  *  refresh rate while the screen is static
  *
- *  @dc:     dc reference
- *  @stream: Initial dc stream state
- *  @adjust: Updated parameters for vertical_total_min and vertical_total_max
+ *  Return: %true if the pipe context is found and adjusted;
+ *          %false if the pipe context is not found.
  */
 bool dc_stream_adjust_vmin_vmax(struct dc *dc,
 		struct dc_stream_state *stream,
 		struct dc_crtc_timing_adjust *adjust)
 {
 	int i;
-	bool ret = false;
+
+	/*
+	 * Don't adjust DRR while there's bandwidth optimizations pending to
+	 * avoid conflicting with firmware updates.
+	 */
+	if (dc->ctx->dce_version > DCE_VERSION_MAX) {
+		if (dc->optimized_required &&
+			(stream->adjust.v_total_max != adjust->v_total_max ||
+			stream->adjust.v_total_min != adjust->v_total_min)) {
+			stream->adjust.timing_adjust_pending = true;
+			return false;
+		}
+	}
+
+	dc_exit_ips_for_hw_access(dc);
 
 	stream->adjust.v_total_max = adjust->v_total_max;
 	stream->adjust.v_total_mid = adjust->v_total_mid;
 	stream->adjust.v_total_mid_frame_num = adjust->v_total_mid_frame_num;
 	stream->adjust.v_total_min = adjust->v_total_min;
+	stream->adjust.allow_otg_v_count_halt = adjust->allow_otg_v_count_halt;
+
+	if (dc->caps.max_v_total != 0 &&
+		(adjust->v_total_max > dc->caps.max_v_total || adjust->v_total_min > dc->caps.max_v_total)) {
+		stream->adjust.timing_adjust_pending = false;
+		if (adjust->allow_otg_v_count_halt)
+			return set_long_vtotal(dc, stream, adjust);
+		else
+			return false;
+	}
 
 	for (i = 0; i < MAX_PIPES; i++) {
 		struct pipe_ctx *pipe = &dc->current_state->res_ctx.pipe_ctx[i];
@@ -424,26 +499,30 @@ bool dc_stream_adjust_vmin_vmax(struct dc *dc,
 			dc->hwss.set_drr(&pipe,
 					1,
 					*adjust);
+			stream->adjust.timing_adjust_pending = false;
 
-			ret = true;
+			if (dc->hwss.notify_cursor_offload_drr_update)
+				dc->hwss.notify_cursor_offload_drr_update(dc, dc->current_state, stream);
+
+			return true;
 		}
 	}
-	return ret;
+
+	return false;
 }
 
 /**
- *****************************************************************************
- *  Function: dc_stream_get_last_vrr_vtotal
- *
- *  @brief
- *     Looks up the pipe context of dc_stream_state and gets the
- *     last VTOTAL used by DRR (Dynamic Refresh Rate)
- *
- *  @param [in] dc: dc reference
- *  @param [in] stream: Initial dc stream state
- *  @param [in] adjust: Updated parameters for vertical_total_min and
- *  vertical_total_max
- *****************************************************************************
+ * dc_stream_get_last_used_drr_vtotal - Looks up the pipe context of
+ * dc_stream_state and gets the last VTOTAL used by DRR (Dynamic Refresh Rate)
+ *
+ * @dc: [in] dc reference
+ * @stream: [in] Initial dc stream state
+ * @refresh_rate: [in] new refresh_rate
+ *
+ * Return: %true if the pipe context is found and there is an associated
+ *         timing_generator for the DC;
+ *         %false if the pipe context is not found or there is no
+ *         timing_generator for the DC.
  */
 bool dc_stream_get_last_used_drr_vtotal(struct dc *dc,
 		struct dc_stream_state *stream,
@@ -453,6 +532,8 @@ bool dc_stream_get_last_used_drr_vtotal(struct dc *dc,
 
 	int i = 0;
 
+	dc_exit_ips_for_hw_access(dc);
+
 	for (i = 0; i < MAX_PIPES; i++) {
 		struct pipe_ctx *pipe = &dc->current_state->res_ctx.pipe_ctx[i];
 
@@ -473,142 +554,176 @@ bool dc_stream_get_last_used_drr_vtotal(struct dc *dc,
 	return status;
 }
 
-bool dc_stream_get_crtc_position(struct dc *dc,
-		struct dc_stream_state **streams, int num_streams,
-		unsigned int *v_pos, unsigned int *nom_v_pos)
+#if defined(CONFIG_DRM_AMD_SECURE_DISPLAY)
+static inline void
+dc_stream_forward_dmub_crc_window(struct dc_dmub_srv *dmub_srv,
+		struct rect *rect, struct otg_phy_mux *mux_mapping, bool is_stop)
 {
-	/* TODO: Support multiple streams */
-	const struct dc_stream_state *stream = streams[0];
-	int i;
-	bool ret = false;
-	struct crtc_position position;
-
-	for (i = 0; i < MAX_PIPES; i++) {
-		struct pipe_ctx *pipe =
-				&dc->current_state->res_ctx.pipe_ctx[i];
+	union dmub_rb_cmd cmd = {0};
 
-		if (pipe->stream == stream && pipe->stream_res.stream_enc) {
-			dc->hwss.get_position(&pipe, 1, &position);
+	cmd.secure_display.roi_info.phy_id = mux_mapping->phy_output_num;
+	cmd.secure_display.roi_info.otg_id = mux_mapping->otg_output_num;
 
-			*v_pos = position.vertical_count;
-			*nom_v_pos = position.nominal_vcount;
-			ret = true;
-		}
+	if (is_stop) {
+		cmd.secure_display.header.type = DMUB_CMD__SECURE_DISPLAY;
+		cmd.secure_display.header.sub_type = DMUB_CMD__SECURE_DISPLAY_CRC_STOP_UPDATE;
+	} else {
+		cmd.secure_display.header.type = DMUB_CMD__SECURE_DISPLAY;
+		cmd.secure_display.header.sub_type = DMUB_CMD__SECURE_DISPLAY_CRC_WIN_NOTIFY;
+		cmd.secure_display.roi_info.x_start = rect->x;
+		cmd.secure_display.roi_info.y_start = rect->y;
+		cmd.secure_display.roi_info.x_end = rect->x + rect->width;
+		cmd.secure_display.roi_info.y_end = rect->y + rect->height;
 	}
-	return ret;
+
+	dc_wake_and_execute_dmub_cmd(dmub_srv->ctx, &cmd, DM_DMUB_WAIT_TYPE_NO_WAIT);
 }
 
-#if defined(CONFIG_DRM_AMD_SECURE_DISPLAY)
-bool dc_stream_forward_dmcu_crc_window(struct dc *dc, struct dc_stream_state *stream,
-			     struct crc_params *crc_window)
+static inline void
+dc_stream_forward_dmcu_crc_window(struct dmcu *dmcu,
+		struct rect *rect, struct otg_phy_mux *mux_mapping, bool is_stop)
 {
-	int i;
-	struct dmcu *dmcu = dc->res_pool->dmcu;
+	if (is_stop)
+		dmcu->funcs->stop_crc_win_update(dmcu, mux_mapping);
+	else
+		dmcu->funcs->forward_crc_window(dmcu, rect, mux_mapping);
+}
+
+bool
+dc_stream_forward_crc_window(struct dc_stream_state *stream,
+		struct rect *rect, uint8_t phy_id, bool is_stop)
+{
+	struct dmcu *dmcu;
+	struct dc_dmub_srv *dmub_srv;
+	struct otg_phy_mux mux_mapping;
 	struct pipe_ctx *pipe;
-	struct crc_region tmp_win, *crc_win;
-	struct otg_phy_mux mapping_tmp, *mux_mapping;
+	int i;
+	struct dc *dc = stream->ctx->dc;
+
+	for (i = 0; i < MAX_PIPES; i++) {
+		pipe = &dc->current_state->res_ctx.pipe_ctx[i];
+		if (pipe->stream == stream && !pipe->top_pipe && !pipe->prev_odm_pipe)
+			break;
+	}
 
-	/*crc window can't be null*/
-	if (!crc_window)
+	/* Stream not found */
+	if (i == MAX_PIPES)
 		return false;
 
-	if ((dmcu != NULL && dmcu->funcs->is_dmcu_initialized(dmcu))) {
-		crc_win = &tmp_win;
-		mux_mapping = &mapping_tmp;
-		/*set crc window*/
-		tmp_win.x_start = crc_window->windowa_x_start;
-		tmp_win.y_start = crc_window->windowa_y_start;
-		tmp_win.x_end = crc_window->windowa_x_end;
-		tmp_win.y_end = crc_window->windowa_y_end;
+	mux_mapping.phy_output_num = phy_id;
+	mux_mapping.otg_output_num = pipe->stream_res.tg->inst;
 
-		for (i = 0; i < MAX_PIPES; i++) {
-			pipe = &dc->current_state->res_ctx.pipe_ctx[i];
-			if (pipe->stream == stream && !pipe->top_pipe && !pipe->prev_odm_pipe)
-				break;
-		}
+	dmcu = dc->res_pool->dmcu;
+	dmub_srv = dc->ctx->dmub_srv;
 
-		/* Stream not found */
-		if (i == MAX_PIPES)
-			return false;
+	/* forward to dmub */
+	if (dmub_srv)
+		dc_stream_forward_dmub_crc_window(dmub_srv, rect, &mux_mapping, is_stop);
+	/* forward to dmcu */
+	else if (dmcu && dmcu->funcs->is_dmcu_initialized(dmcu))
+		dc_stream_forward_dmcu_crc_window(dmcu, rect, &mux_mapping, is_stop);
+	else
+		return false;
+
+	return true;
+}
 
+static void
+dc_stream_forward_dmub_multiple_crc_window(struct dc_dmub_srv *dmub_srv,
+		struct crc_window *window, struct otg_phy_mux *mux_mapping, bool stop)
+{
+	int i;
+	union dmub_rb_cmd cmd = {0};
 
-		/*set mux routing info*/
-		mapping_tmp.phy_output_num = stream->link->link_enc_hw_inst;
-		mapping_tmp.otg_output_num = pipe->stream_res.tg->inst;
+	cmd.secure_display.mul_roi_ctl.phy_id = mux_mapping->phy_output_num;
+	cmd.secure_display.mul_roi_ctl.otg_id = mux_mapping->otg_output_num;
 
-		dmcu->funcs->forward_crc_window(dmcu, crc_win, mux_mapping);
+	cmd.secure_display.header.type = DMUB_CMD__SECURE_DISPLAY;
+
+	if (stop) {
+		cmd.secure_display.header.sub_type = DMUB_CMD__SECURE_DISPLAY_MULTIPLE_CRC_STOP_UPDATE;
 	} else {
-		DC_LOG_DC("dmcu is not initialized");
-		return false;
+		cmd.secure_display.header.sub_type = DMUB_CMD__SECURE_DISPLAY_MULTIPLE_CRC_WIN_NOTIFY;
+		for (i = 0; i < MAX_CRC_WINDOW_NUM; i++) {
+			cmd.secure_display.mul_roi_ctl.roi_ctl[i].x_start = window[i].rect.x;
+			cmd.secure_display.mul_roi_ctl.roi_ctl[i].y_start = window[i].rect.y;
+			cmd.secure_display.mul_roi_ctl.roi_ctl[i].x_end = window[i].rect.x + window[i].rect.width;
+			cmd.secure_display.mul_roi_ctl.roi_ctl[i].y_end = window[i].rect.y + window[i].rect.height;
+			cmd.secure_display.mul_roi_ctl.roi_ctl[i].enable = window[i].enable;
+		}
 	}
 
-	return true;
+	dc_wake_and_execute_dmub_cmd(dmub_srv->ctx, &cmd, DM_DMUB_WAIT_TYPE_NO_WAIT);
 }
 
-bool dc_stream_stop_dmcu_crc_win_update(struct dc *dc, struct dc_stream_state *stream)
+bool
+dc_stream_forward_multiple_crc_window(struct dc_stream_state *stream,
+		struct crc_window *window, uint8_t phy_id, bool stop)
 {
-	int i;
-	struct dmcu *dmcu = dc->res_pool->dmcu;
+	struct dc_dmub_srv *dmub_srv;
+	struct otg_phy_mux mux_mapping;
 	struct pipe_ctx *pipe;
-	struct otg_phy_mux mapping_tmp, *mux_mapping;
-
-	if ((dmcu != NULL && dmcu->funcs->is_dmcu_initialized(dmcu))) {
-		mux_mapping = &mapping_tmp;
+	int i;
+	struct dc *dc = stream->ctx->dc;
 
-		for (i = 0; i < MAX_PIPES; i++) {
-			pipe = &dc->current_state->res_ctx.pipe_ctx[i];
-			if (pipe->stream == stream && !pipe->top_pipe && !pipe->prev_odm_pipe)
-				break;
-		}
+	for (i = 0; i < MAX_PIPES; i++) {
+		pipe = &dc->current_state->res_ctx.pipe_ctx[i];
+		if (pipe->stream == stream && !pipe->top_pipe && !pipe->prev_odm_pipe)
+			break;
+	}
 
-		/* Stream not found */
-		if (i == MAX_PIPES)
-			return false;
+	/* Stream not found */
+	if (i == MAX_PIPES)
+		return false;
 
+	mux_mapping.phy_output_num = phy_id;
+	mux_mapping.otg_output_num = pipe->stream_res.tg->inst;
 
-		/*set mux routing info*/
-		mapping_tmp.phy_output_num = stream->link->link_enc_hw_inst;
-		mapping_tmp.otg_output_num = pipe->stream_res.tg->inst;
+	dmub_srv = dc->ctx->dmub_srv;
 
-		dmcu->funcs->stop_crc_win_update(dmcu, mux_mapping);
-	} else {
-		DC_LOG_DC("dmcu is not initialized");
+	/* forward to dmub only. no dmcu support*/
+	if (dmub_srv)
+		dc_stream_forward_dmub_multiple_crc_window(dmub_srv, window, &mux_mapping, stop);
+	else
 		return false;
-	}
 
 	return true;
 }
-#endif
+#endif /* CONFIG_DRM_AMD_SECURE_DISPLAY */
 
 /**
  * dc_stream_configure_crc() - Configure CRC capture for the given stream.
  * @dc: DC Object
  * @stream: The stream to configure CRC on.
- * @enable: Enable CRC if true, disable otherwise.
  * @crc_window: CRC window (x/y start/end) information
+ * @enable: Enable CRC if true, disable otherwise.
  * @continuous: Capture CRC on every frame if true. Otherwise, only capture
  *              once.
+ * @idx: Capture CRC on which CRC engine instance
+ * @reset: Reset CRC engine before the configuration
+ *
+ * By default, the entire frame is used to calculate the CRC.
  *
- * By default, only CRC0 is configured, and the entire frame is used to
- * calculate the crc.
+ * Return: %false if the stream is not found or CRC capture is not supported;
+ *         %true if the stream has been configured.
  */
 bool dc_stream_configure_crc(struct dc *dc, struct dc_stream_state *stream,
-			     struct crc_params *crc_window, bool enable, bool continuous)
+			     struct crc_params *crc_window, bool enable, bool continuous,
+			     uint8_t idx, bool reset)
 {
-	int i;
 	struct pipe_ctx *pipe;
 	struct crc_params param;
 	struct timing_generator *tg;
 
-	for (i = 0; i < MAX_PIPES; i++) {
-		pipe = &dc->current_state->res_ctx.pipe_ctx[i];
-		if (pipe->stream == stream && !pipe->top_pipe && !pipe->prev_odm_pipe)
-			break;
-	}
+	pipe = resource_get_otg_master_for_stream(
+			&dc->current_state->res_ctx, stream);
+
 	/* Stream not found */
-	if (i == MAX_PIPES)
+	if (pipe == NULL)
 		return false;
 
+	dc_exit_ips_for_hw_access(dc);
+
 	/* By default, capture the full frame */
 	param.windowa_x_start = 0;
 	param.windowa_y_start = 0;
@@ -638,6 +753,9 @@ bool dc_stream_configure_crc(struct dc *dc, struct dc_stream_state *stream,
 	param.continuous_mode = continuous;
 	param.enable = enable;
 
+	param.crc_eng_inst = idx;
+	param.reset = reset;
+
 	tg = pipe->stream_res.tg;
 
 	/* Only call if supported */
@@ -649,22 +767,28 @@ bool dc_stream_configure_crc(struct dc *dc, struct dc_stream_state *stream,
 
 /**
  * dc_stream_get_crc() - Get CRC values for the given stream.
- * @dc: DC object
+ *
+ * @dc: DC object.
  * @stream: The DC stream state of the stream to get CRCs from.
- * @r_cr: CRC value for the first of the 3 channels stored here.
- * @g_y:  CRC value for the second of the 3 channels stored here.
- * @b_cb: CRC value for the third of the 3 channels stored here.
+ * @idx: index of crc engine to get CRC from
+ * @r_cr: CRC value for the red component.
+ * @g_y:  CRC value for the green component.
+ * @b_cb: CRC value for the blue component.
  *
  * dc_stream_configure_crc needs to be called beforehand to enable CRCs.
- * Return false if stream is not found, or if CRCs are not enabled.
+ *
+ * Return:
+ * %false if stream is not found, or if CRCs are not enabled.
  */
-bool dc_stream_get_crc(struct dc *dc, struct dc_stream_state *stream,
+bool dc_stream_get_crc(struct dc *dc, struct dc_stream_state *stream, uint8_t idx,
 		       uint32_t *r_cr, uint32_t *g_y, uint32_t *b_cb)
 {
 	int i;
 	struct pipe_ctx *pipe;
 	struct timing_generator *tg;
 
+	dc_exit_ips_for_hw_access(dc);
+
 	for (i = 0; i < MAX_PIPES; i++) {
 		pipe = &dc->current_state->res_ctx.pipe_ctx[i];
 		if (pipe->stream == stream)
@@ -677,7 +801,7 @@ bool dc_stream_get_crc(struct dc *dc, struct dc_stream_state *stream,
 	tg = pipe->stream_res.tg;
 
 	if (tg->funcs->get_crc)
-		return tg->funcs->get_crc(tg, r_cr, g_y, b_cb);
+		return tg->funcs->get_crc(tg, idx, r_cr, g_y, b_cb);
 	DC_LOG_WARNING("CRC capture not supported.");
 	return false;
 }
@@ -689,6 +813,8 @@ void dc_stream_set_dyn_expansion(struct dc *dc, struct dc_stream_state *stream,
 	int i;
 	struct pipe_ctx *pipe_ctx;
 
+	dc_exit_ips_for_hw_access(dc);
+
 	for (i = 0; i < MAX_PIPES; i++) {
 		if (dc->current_state->res_ctx.pipe_ctx[i].stream
 				== stream) {
@@ -724,6 +850,8 @@ void dc_stream_set_dither_option(struct dc_stream_state *stream,
 	if (option > DITHER_OPTION_MAX)
 		return;
 
+	dc_exit_ips_for_hw_access(stream->ctx->dc);
+
 	stream->dither_option = option;
 
 	memset(&params, 0, sizeof(params));
@@ -748,6 +876,8 @@ bool dc_stream_set_gamut_remap(struct dc *dc, const struct dc_stream_state *stre
 	bool ret = false;
 	struct pipe_ctx *pipes;
 
+	dc_exit_ips_for_hw_access(dc);
+
 	for (i = 0; i < MAX_PIPES; i++) {
 		if (dc->current_state->res_ctx.pipe_ctx[i].stream == stream) {
 			pipes = &dc->current_state->res_ctx.pipe_ctx[i];
@@ -765,6 +895,8 @@ bool dc_stream_program_csc_matrix(struct dc *dc, struct dc_stream_state *stream)
 	bool ret = false;
 	struct pipe_ctx *pipes;
 
+	dc_exit_ips_for_hw_access(dc);
+
 	for (i = 0; i < MAX_PIPES; i++) {
 		if (dc->current_state->res_ctx.pipe_ctx[i].stream
 				== stream) {
@@ -791,6 +923,8 @@ void dc_stream_set_static_screen_params(struct dc *dc,
 	struct pipe_ctx *pipes_affected[MAX_PIPES];
 	int num_pipes_affected = 0;
 
+	dc_exit_ips_for_hw_access(dc);
+
 	for (i = 0; i < num_streams; i++) {
 		struct dc_stream_state *stream = streams[i];
 
@@ -808,8 +942,13 @@ void dc_stream_set_static_screen_params(struct dc *dc,
 
 static void dc_destruct(struct dc *dc)
 {
+	// reset link encoder assignment table on destruct
+	if (dc->res_pool && dc->res_pool->funcs->link_encs_assign &&
+			!dc->config.unify_link_enc_assignment)
+		link_enc_cfg_init(dc, dc->current_state);
+
 	if (dc->current_state) {
-		dc_release_state(dc->current_state);
+		dc_state_release(dc->current_state);
 		dc->current_state = NULL;
 	}
 
@@ -823,17 +962,24 @@ static void dc_destruct(struct dc *dc)
 	}
 
 	dc_destroy_resource_pool(dc);
+#ifdef CONFIG_DRM_AMD_DC_FP
+	dc_destroy_soc_and_ip_translator(&dc->soc_and_ip_translator);
+#endif
+	if (dc->link_srv)
+		link_destroy_link_service(&dc->link_srv);
 
-	if (dc->ctx->gpio_service)
-		dal_gpio_service_destroy(&dc->ctx->gpio_service);
-
-	if (dc->ctx->created_bios)
-		dal_bios_parser_destroy(&dc->ctx->dc_bios);
+	if (dc->ctx) {
+		if (dc->ctx->gpio_service)
+			dal_gpio_service_destroy(&dc->ctx->gpio_service);
 
-	dc_perf_trace_destroy(&dc->ctx->perf_trace);
+		if (dc->ctx->created_bios)
+			dal_bios_parser_destroy(&dc->ctx->dc_bios);
+		kfree(dc->ctx->logger);
+		dc_perf_trace_destroy(&dc->ctx->perf_trace);
 
-	kfree(dc->ctx);
-	dc->ctx = NULL;
+		kfree(dc->ctx);
+		dc->ctx = NULL;
+	}
 
 	kfree(dc->bw_vbios);
 	dc->bw_vbios = NULL;
@@ -841,14 +987,12 @@ static void dc_destruct(struct dc *dc)
 	kfree(dc->bw_dceip);
 	dc->bw_dceip = NULL;
 
-#ifdef CONFIG_DRM_AMD_DC_DCN
 	kfree(dc->dcn_soc);
 	dc->dcn_soc = NULL;
 
 	kfree(dc->dcn_ip);
 	dc->dcn_ip = NULL;
 
-#endif
 	kfree(dc->vm_helper);
 	dc->vm_helper = NULL;
 
@@ -858,12 +1002,13 @@ static bool dc_construct_ctx(struct dc *dc,
 		const struct dc_init_data *init_params)
 {
 	struct dc_context *dc_ctx;
-	enum dce_version dc_version = DCE_VERSION_UNKNOWN;
 
 	dc_ctx = kzalloc(sizeof(*dc_ctx), GFP_KERNEL);
 	if (!dc_ctx)
 		return false;
 
+	dc_stream_init_rmcm_3dlut(dc);
+
 	dc_ctx->cgs_device = init_params->cgs_device;
 	dc_ctx->driver_context = init_params->driver;
 	dc_ctx->dc = dc;
@@ -871,20 +1016,36 @@ static bool dc_construct_ctx(struct dc *dc,
 	dc_ctx->dc_sink_id_count = 0;
 	dc_ctx->dc_stream_id_count = 0;
 	dc_ctx->dce_environment = init_params->dce_environment;
+	dc_ctx->dcn_reg_offsets = init_params->dcn_reg_offsets;
+	dc_ctx->nbio_reg_offsets = init_params->nbio_reg_offsets;
+	dc_ctx->clk_reg_offsets = init_params->clk_reg_offsets;
 
 	/* Create logger */
+	dc_ctx->logger = kmalloc(sizeof(*dc_ctx->logger), GFP_KERNEL);
 
-	dc_version = resource_parse_asic_id(init_params->asic_id);
-	dc_ctx->dce_version = dc_version;
+	if (!dc_ctx->logger) {
+		kfree(dc_ctx);
+		return false;
+	}
+
+	dc_ctx->logger->dev = adev_to_drm(init_params->driver);
+	dc->dml.logger = dc_ctx->logger;
+
+	dc_ctx->dce_version = resource_parse_asic_id(init_params->asic_id);
 
 	dc_ctx->perf_trace = dc_perf_trace_create();
 	if (!dc_ctx->perf_trace) {
+		kfree(dc_ctx);
 		ASSERT_CRITICAL(false);
 		return false;
 	}
 
 	dc->ctx = dc_ctx;
 
+	dc->link_srv = link_create_link_service();
+	if (!dc->link_srv)
+		return false;
+
 	return true;
 }
 
@@ -894,10 +1055,8 @@ static bool dc_construct(struct dc *dc,
 	struct dc_context *dc_ctx;
 	struct bw_calcs_dceip *dc_dceip;
 	struct bw_calcs_vbios *dc_vbios;
-#ifdef CONFIG_DRM_AMD_DC_DCN
 	struct dcn_soc_bounding_box *dcn_soc;
 	struct dcn_ip_params *dcn_ip;
-#endif
 
 	dc->config = init_params->flags;
 
@@ -925,7 +1084,6 @@ static bool dc_construct(struct dc *dc,
 	}
 
 	dc->bw_vbios = dc_vbios;
-#ifdef CONFIG_DRM_AMD_DC_DCN
 	dcn_soc = kzalloc(sizeof(*dcn_soc), GFP_KERNEL);
 	if (!dcn_soc) {
 		dm_error("%s: failed to create dcn_soc\n", __func__);
@@ -941,14 +1099,18 @@ static bool dc_construct(struct dc *dc,
 	}
 
 	dc->dcn_ip = dcn_ip;
-#endif
+
+	if (init_params->bb_from_dmub)
+		dc->dml2_options.bb_from_dmub = init_params->bb_from_dmub;
+	else
+		dc->dml2_options.bb_from_dmub = NULL;
 
 	if (!dc_construct_ctx(dc, init_params)) {
 		dm_error("%s: failed to create ctx\n", __func__);
 		goto fail;
 	}
 
-        dc_ctx = dc->ctx;
+	dc_ctx = dc->ctx;
 
 	/* Resource should construct all asic specific resources.
 	 * This should be the only place where we need to parse the asic id
@@ -993,30 +1155,23 @@ static bool dc_construct(struct dc *dc,
 	/* set i2c speed if not done by the respective dcnxxx__resource.c */
 	if (dc->caps.i2c_speed_in_khz_hdcp == 0)
 		dc->caps.i2c_speed_in_khz_hdcp = dc->caps.i2c_speed_in_khz;
-
+	if (dc->check_config.max_optimizable_video_width == 0)
+		dc->check_config.max_optimizable_video_width = 5120;
 	dc->clk_mgr = dc_clk_mgr_create(dc->ctx, dc->res_pool->pp_smu, dc->res_pool->dccg);
 	if (!dc->clk_mgr)
 		goto fail;
-#ifdef CONFIG_DRM_AMD_DC_DCN
+#ifdef CONFIG_DRM_AMD_DC_FP
 	dc->clk_mgr->force_smu_not_present = init_params->force_smu_not_present;
-#endif
 
-	if (dc->res_pool->funcs->update_bw_bounding_box)
+	if (dc->res_pool->funcs->update_bw_bounding_box) {
+		DC_FP_START();
 		dc->res_pool->funcs->update_bw_bounding_box(dc, dc->clk_mgr->bw_params);
-
-	/* Creation of current_state must occur after dc->dml
-	 * is initialized in dc_create_resource_pool because
-	 * on creation it copies the contents of dc->dml
-	 */
-
-	dc->current_state = dc_create_state(dc);
-
-	if (!dc->current_state) {
-		dm_error("%s: failed to create validate ctx\n", __func__);
-		goto fail;
+		DC_FP_END();
 	}
-
-	dc_resource_state_construct(dc, dc->current_state);
+	dc->soc_and_ip_translator = dc_create_soc_and_ip_translator(dc_ctx->dce_version);
+	if (!dc->soc_and_ip_translator)
+		goto fail;
+#endif
 
 	if (!create_links(dc, init_params->num_virtual_links))
 		goto fail;
@@ -1027,8 +1182,16 @@ static bool dc_construct(struct dc *dc,
 	if (!create_link_encoders(dc))
 		goto fail;
 
-	/* Initialise DIG link encoder resource tracking variables. */
-	link_enc_cfg_init(dc, dc->current_state);
+	/* Creation of current_state must occur after dc->dml
+	 * is initialized in dc_create_resource_pool because
+	 * on creation it copies the contents of dc->dml
+	 */
+	dc->current_state = dc_state_create(dc, NULL);
+
+	if (!dc->current_state) {
+		dm_error("%s: failed to create validate ctx\n", __func__);
+		goto fail;
+	}
 
 	return true;
 
@@ -1047,8 +1210,10 @@ static void disable_all_writeback_pipes_for_stream(
 		stream->writeback_info[i].wb_enabled = false;
 }
 
-static void apply_ctx_interdependent_lock(struct dc *dc, struct dc_state *context,
-					  struct dc_stream_state *stream, bool lock)
+static void apply_ctx_interdependent_lock(struct dc *dc,
+					  struct dc_state *context,
+					  struct dc_stream_state *stream,
+					  bool lock)
 {
 	int i;
 
@@ -1062,7 +1227,7 @@ static void apply_ctx_interdependent_lock(struct dc *dc, struct dc_state *contex
 
 			// Copied conditions that were previously in dce110_apply_ctx_for_surface
 			if (stream == pipe_ctx->stream) {
-				if (!pipe_ctx->top_pipe &&
+				if (resource_is_pipe_type(pipe_ctx, OPP_HEAD) &&
 					(pipe_ctx->plane_state || old_pipe_ctx->plane_state))
 					dc->hwss.pipe_control_lock(dc, pipe_ctx, lock);
 			}
@@ -1070,23 +1235,116 @@ static void apply_ctx_interdependent_lock(struct dc *dc, struct dc_state *contex
 	}
 }
 
+static void dc_update_visual_confirm_color(struct dc *dc, struct dc_state *context, struct pipe_ctx *pipe_ctx)
+{
+	if (dc->debug.visual_confirm & VISUAL_CONFIRM_EXPLICIT) {
+		memcpy(&pipe_ctx->visual_confirm_color, &pipe_ctx->plane_state->visual_confirm_color,
+		sizeof(pipe_ctx->visual_confirm_color));
+		return;
+	}
+
+	if (dc->ctx->dce_version >= DCN_VERSION_1_0) {
+		memset(&pipe_ctx->visual_confirm_color, 0, sizeof(struct tg_color));
+
+		if (dc->debug.visual_confirm == VISUAL_CONFIRM_HDR)
+			get_hdr_visual_confirm_color(pipe_ctx, &(pipe_ctx->visual_confirm_color));
+		else if (dc->debug.visual_confirm == VISUAL_CONFIRM_SURFACE)
+			get_surface_visual_confirm_color(pipe_ctx, &(pipe_ctx->visual_confirm_color));
+		else if (dc->debug.visual_confirm == VISUAL_CONFIRM_SWIZZLE)
+			get_surface_tile_visual_confirm_color(pipe_ctx, &(pipe_ctx->visual_confirm_color));
+		else if (dc->debug.visual_confirm == VISUAL_CONFIRM_HW_CURSOR)
+			get_cursor_visual_confirm_color(pipe_ctx, &(pipe_ctx->visual_confirm_color));
+		else if (dc->debug.visual_confirm == VISUAL_CONFIRM_DCC)
+			get_dcc_visual_confirm_color(dc, pipe_ctx, &(pipe_ctx->visual_confirm_color));
+		else {
+			if (dc->ctx->dce_version < DCN_VERSION_2_0)
+				color_space_to_black_color(
+					dc, pipe_ctx->stream->output_color_space, &(pipe_ctx->visual_confirm_color));
+		}
+		if (dc->ctx->dce_version >= DCN_VERSION_2_0) {
+			if (dc->debug.visual_confirm == VISUAL_CONFIRM_MPCTREE)
+				get_mpctree_visual_confirm_color(pipe_ctx, &(pipe_ctx->visual_confirm_color));
+			else if (dc->debug.visual_confirm == VISUAL_CONFIRM_SUBVP)
+				get_subvp_visual_confirm_color(pipe_ctx, &(pipe_ctx->visual_confirm_color));
+			else if (dc->debug.visual_confirm == VISUAL_CONFIRM_MCLK_SWITCH)
+				get_mclk_switch_visual_confirm_color(pipe_ctx, &(pipe_ctx->visual_confirm_color));
+			else if (dc->debug.visual_confirm == VISUAL_CONFIRM_FAMS2)
+				get_fams2_visual_confirm_color(dc, context, pipe_ctx, &(pipe_ctx->visual_confirm_color));
+			else if (dc->debug.visual_confirm == VISUAL_CONFIRM_VABC)
+				get_vabc_visual_confirm_color(pipe_ctx, &(pipe_ctx->visual_confirm_color));
+		}
+	}
+}
+
+void dc_get_visual_confirm_for_stream(
+	struct dc *dc,
+	struct dc_stream_state *stream_state,
+	struct tg_color *color)
+{
+	struct dc_stream_status *stream_status = dc_stream_get_status(stream_state);
+	struct pipe_ctx *pipe_ctx;
+	int i;
+	struct dc_plane_state *plane_state = NULL;
+
+	if (!stream_status)
+		return;
+
+	switch (dc->debug.visual_confirm) {
+	case VISUAL_CONFIRM_DISABLE:
+		return;
+	case VISUAL_CONFIRM_PSR:
+	case VISUAL_CONFIRM_FAMS:
+		pipe_ctx = dc_stream_get_pipe_ctx(stream_state);
+		if (!pipe_ctx)
+			return;
+		dc_dmub_srv_get_visual_confirm_color_cmd(dc, pipe_ctx);
+		memcpy(color, &dc->ctx->dmub_srv->dmub->visual_confirm_color, sizeof(struct tg_color));
+		return;
+
+	default:
+		/* find plane with highest layer_index */
+		for (i = 0; i < stream_status->plane_count; i++) {
+			if (stream_status->plane_states[i]->visible)
+				plane_state = stream_status->plane_states[i];
+		}
+		if (!plane_state)
+			return;
+		/* find pipe that contains plane with highest layer index */
+		for (i = 0; i < MAX_PIPES; i++) {
+			struct pipe_ctx *pipe = &dc->current_state->res_ctx.pipe_ctx[i];
+
+			if (pipe->plane_state == plane_state) {
+				memcpy(color, &pipe->visual_confirm_color, sizeof(struct tg_color));
+				return;
+			}
+		}
+	}
+}
+
 static void disable_dangling_plane(struct dc *dc, struct dc_state *context)
 {
 	int i, j;
-	struct dc_state *dangling_context = dc_create_state(dc);
+	struct dc_state *dangling_context = dc_state_create_current_copy(dc);
 	struct dc_state *current_ctx;
+	struct pipe_ctx *pipe;
+	struct timing_generator *tg;
 
 	if (dangling_context == NULL)
 		return;
 
-	dc_resource_state_copy_construct(dc->current_state, dangling_context);
-
 	for (i = 0; i < dc->res_pool->pipe_count; i++) {
 		struct dc_stream_state *old_stream =
 				dc->current_state->res_ctx.pipe_ctx[i].stream;
 		bool should_disable = true;
-		bool pipe_split_change =
-			context->res_ctx.pipe_ctx[i].top_pipe != dc->current_state->res_ctx.pipe_ctx[i].top_pipe;
+		bool pipe_split_change = false;
+
+		if ((context->res_ctx.pipe_ctx[i].top_pipe) &&
+			(dc->current_state->res_ctx.pipe_ctx[i].top_pipe))
+			pipe_split_change = context->res_ctx.pipe_ctx[i].top_pipe->pipe_idx !=
+				dc->current_state->res_ctx.pipe_ctx[i].top_pipe->pipe_idx;
+		else
+			pipe_split_change = context->res_ctx.pipe_ctx[i].top_pipe !=
+				dc->current_state->res_ctx.pipe_ctx[i].top_pipe;
 
 		for (j = 0; j < context->stream_count; j++) {
 			if (old_stream == context->streams[j]) {
@@ -1094,31 +1352,81 @@ static void disable_dangling_plane(struct dc *dc, struct dc_state *context)
 				break;
 			}
 		}
-		if (!should_disable && pipe_split_change)
+		if (!should_disable && pipe_split_change &&
+				dc->current_state->stream_count != context->stream_count)
 			should_disable = true;
 
+		if (old_stream && !dc->current_state->res_ctx.pipe_ctx[i].top_pipe &&
+				!dc->current_state->res_ctx.pipe_ctx[i].prev_odm_pipe) {
+			struct pipe_ctx *old_pipe, *new_pipe;
+
+			old_pipe = &dc->current_state->res_ctx.pipe_ctx[i];
+			new_pipe = &context->res_ctx.pipe_ctx[i];
+
+			if (old_pipe->plane_state && !new_pipe->plane_state)
+				should_disable = true;
+		}
+
 		if (should_disable && old_stream) {
-			dc_rem_all_planes_for_stream(dc, old_stream, dangling_context);
+			bool is_phantom = dc_state_get_stream_subvp_type(dc->current_state, old_stream) == SUBVP_PHANTOM;
+			pipe = &dc->current_state->res_ctx.pipe_ctx[i];
+			tg = pipe->stream_res.tg;
+			/* When disabling plane for a phantom pipe, we must turn on the
+			 * phantom OTG so the disable programming gets the double buffer
+			 * update. Otherwise the pipe will be left in a partially disabled
+			 * state that can result in underflow or hang when enabling it
+			 * again for different use.
+			 */
+			if (is_phantom) {
+				if (tg->funcs->enable_crtc) {
+					if (dc->hwseq->funcs.blank_pixel_data)
+						dc->hwseq->funcs.blank_pixel_data(dc, pipe, true);
+					tg->funcs->enable_crtc(tg);
+				}
+			}
+
+			if (is_phantom)
+				dc_state_rem_all_phantom_planes_for_stream(dc, old_stream, dangling_context, true);
+			else
+				dc_state_rem_all_planes_for_stream(dc, old_stream, dangling_context);
 			disable_all_writeback_pipes_for_stream(dc, old_stream, dangling_context);
 
+			if (pipe->stream && pipe->plane_state) {
+				if (!dc->debug.using_dml2)
+					set_p_state_switch_method(dc, context, pipe);
+				dc_update_visual_confirm_color(dc, context, pipe);
+			}
+
 			if (dc->hwss.apply_ctx_for_surface) {
 				apply_ctx_interdependent_lock(dc, dc->current_state, old_stream, true);
 				dc->hwss.apply_ctx_for_surface(dc, old_stream, 0, dangling_context);
 				apply_ctx_interdependent_lock(dc, dc->current_state, old_stream, false);
 				dc->hwss.post_unlock_program_front_end(dc, dangling_context);
 			}
+
+			if (dc->res_pool->funcs->prepare_mcache_programming)
+				dc->res_pool->funcs->prepare_mcache_programming(dc, dangling_context);
 			if (dc->hwss.program_front_end_for_ctx) {
 				dc->hwss.interdependent_update_lock(dc, dc->current_state, true);
 				dc->hwss.program_front_end_for_ctx(dc, dangling_context);
 				dc->hwss.interdependent_update_lock(dc, dc->current_state, false);
 				dc->hwss.post_unlock_program_front_end(dc, dangling_context);
 			}
+			/* We need to put the phantom OTG back into it's default (disabled) state or we
+			 * can get corruption when transition from one SubVP config to a different one.
+			 * The OTG is set to disable on falling edge of VUPDATE so the plane disable
+			 * will still get it's double buffer update.
+			 */
+			if (is_phantom) {
+				if (tg->funcs->disable_phantom_crtc)
+					tg->funcs->disable_phantom_crtc(tg);
+			}
 		}
 	}
 
 	current_ctx = dc->current_state;
 	dc->current_state = dangling_context;
-	dc_release_state(current_ctx);
+	dc_state_release(current_ctx);
 }
 
 static void disable_vbios_mode_if_required(
@@ -1138,6 +1446,9 @@ static void disable_vbios_mode_if_required(
 		if (stream == NULL)
 			continue;
 
+		if (stream->apply_seamless_boot_optimization)
+			continue;
+
 		// only looking for first odm pipe
 		if (pipe->prev_odm_pipe)
 			continue;
@@ -1149,7 +1460,7 @@ static void disable_vbios_mode_if_required(
 
 		if (link != NULL && link->link_enc->funcs->is_dig_enabled(link->link_enc)) {
 			unsigned int enc_inst, tg_inst = 0;
-			unsigned int pix_clk_100hz;
+			unsigned int pix_clk_100hz = 0;
 
 			enc_inst = link->link_enc->funcs->get_dig_frontend(link->link_enc);
 			if (enc_inst != ENGINE_ID_UNKNOWN) {
@@ -1170,7 +1481,7 @@ static void disable_vbios_mode_if_required(
 						pipe->stream_res.pix_clk_params.requested_pix_clk_100hz;
 
 					if (pix_clk_100hz != requested_pix_clk_100hz) {
-						core_link_disable_stream(pipe);
+						dc->link_srv->set_dpms_off(pipe);
 						pipe->stream->dpms_off = false;
 					}
 				}
@@ -1179,35 +1490,7 @@ static void disable_vbios_mode_if_required(
 	}
 }
 
-static void wait_for_no_pipes_pending(struct dc *dc, struct dc_state *context)
-{
-	int i;
-	PERF_TRACE();
-	for (i = 0; i < MAX_PIPES; i++) {
-		int count = 0;
-		struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
-
-		if (!pipe->plane_state)
-			continue;
-
-		/* Timeout 100 ms */
-		while (count < 100000) {
-			/* Must set to false to start with, due to OR in update function */
-			pipe->plane_state->status.is_flip_pending = false;
-			dc->hwss.update_pending_status(pipe);
-			if (!pipe->plane_state->status.is_flip_pending)
-				break;
-			udelay(1);
-			count++;
-		}
-		ASSERT(!pipe->plane_state->status.is_flip_pending);
-	}
-	PERF_TRACE();
-}
-
-/*******************************************************************************
- * Public functions
- ******************************************************************************/
+/* Public functions */
 
 struct dc *dc_create(const struct dc_init_data *init_params)
 {
@@ -1218,6 +1501,7 @@ struct dc *dc_create(const struct dc_init_data *init_params)
 		return NULL;
 
 	if (init_params->dce_environment == DCE_ENV_VIRTUAL_HW) {
+		dc->caps.linear_pitch_alignment = 64;
 		if (!dc_construct_ctx(dc, init_params))
 			goto destruct_dc;
 	} else {
@@ -1237,10 +1521,16 @@ struct dc *dc_create(const struct dc_init_data *init_params)
 
 		dc->caps.max_dp_protocol_version = DP_VERSION_1_4;
 
+		dc->caps.max_otg_num = dc->res_pool->res_cap->num_timing_generator;
+
 		if (dc->res_pool->dmcu != NULL)
 			dc->versions.dmcu_version = dc->res_pool->dmcu->dmcu_version;
 	}
 
+	dc->dcn_reg_offsets = init_params->dcn_reg_offsets;
+	dc->nbio_reg_offsets = init_params->nbio_reg_offsets;
+	dc->clk_reg_offsets = init_params->clk_reg_offsets;
+
 	/* Populate versioning information */
 	dc->versions.dc_ver = DC_VER;
 
@@ -1248,8 +1538,6 @@ struct dc *dc_create(const struct dc_init_data *init_params)
 
 	DC_LOG_DC("Display Core initialized\n");
 
-
-
 	return dc;
 
 destruct_dc:
@@ -1266,7 +1554,7 @@ static void detect_edp_presence(struct dc *dc)
 	int i;
 	int edp_num;
 
-	get_edp_links(dc, edp_links, &edp_num);
+	dc_get_edp_links(dc, edp_links, &edp_num);
 	if (!edp_num)
 		return;
 
@@ -1275,7 +1563,7 @@ static void detect_edp_presence(struct dc *dc)
 		if (dc->config.edp_not_connected) {
 			edp_link->edp_sink_present = false;
 		} else {
-			dc_link_detect_sink(edp_link, &type);
+			dc_link_detect_connection_type(edp_link, &type);
 			edp_link->edp_sink_present = (type != dc_connection_none);
 		}
 	}
@@ -1287,21 +1575,18 @@ void dc_hardware_init(struct dc *dc)
 	detect_edp_presence(dc);
 	if (dc->ctx->dce_environment != DCE_ENV_VIRTUAL_HW)
 		dc->hwss.init_hw(dc);
+	dc_dmub_srv_notify_fw_dc_power_state(dc->ctx->dmub_srv, DC_ACPI_CM_POWER_STATE_D0);
 }
 
 void dc_init_callbacks(struct dc *dc,
 		const struct dc_callback_init *init_params)
 {
-#ifdef CONFIG_DRM_AMD_DC_HDCP
 	dc->ctx->cp_psp = init_params->cp_psp;
-#endif
 }
 
 void dc_deinit_callbacks(struct dc *dc)
 {
-#ifdef CONFIG_DRM_AMD_DC_HDCP
 	memset(&dc->ctx->cp_psp, 0, sizeof(dc->ctx->cp_psp));
-#endif
 }
 
 void dc_destroy(struct dc **dc)
@@ -1346,7 +1631,9 @@ static void program_timing_sync(
 	struct pipe_ctx *unsynced_pipes[MAX_PIPES] = { NULL };
 
 	for (i = 0; i < pipe_count; i++) {
-		if (!ctx->res_ctx.pipe_ctx[i].stream || ctx->res_ctx.pipe_ctx[i].top_pipe)
+		if (!ctx->res_ctx.pipe_ctx[i].stream
+				|| ctx->res_ctx.pipe_ctx[i].top_pipe
+				|| ctx->res_ctx.pipe_ctx[i].prev_odm_pipe)
 			continue;
 
 		unsynced_pipes[i] = &ctx->res_ctx.pipe_ctx[i];
@@ -1411,7 +1698,10 @@ static void program_timing_sync(
 		}
 
 		for (k = 0; k < group_size; k++) {
-			struct dc_stream_status *status = dc_stream_get_status_from_state(ctx, pipe_set[k]->stream);
+			struct dc_stream_status *status = dc_state_get_stream_status(ctx, pipe_set[k]->stream);
+
+			if (!status)
+				continue;
 
 			status->timing_sync_info.group_id = num_group;
 			status->timing_sync_info.group_size = group_size;
@@ -1421,27 +1711,42 @@ static void program_timing_sync(
 				status->timing_sync_info.master = false;
 
 		}
+
 		/* remove any other unblanked pipes as they have already been synced */
-		for (j = j + 1; j < group_size; j++) {
-			bool is_blanked;
+		if (dc->config.use_pipe_ctx_sync_logic) {
+			/* check pipe's syncd to decide which pipe to be removed */
+			for (j = 1; j < group_size; j++) {
+				if (pipe_set[j]->pipe_idx_syncd == pipe_set[0]->pipe_idx_syncd) {
+					group_size--;
+					pipe_set[j] = pipe_set[group_size];
+					j--;
+				} else
+					/* link slave pipe's syncd with master pipe */
+					pipe_set[j]->pipe_idx_syncd = pipe_set[0]->pipe_idx_syncd;
+			}
+		} else {
+			/* remove any other pipes by checking valid plane */
+			for (j = j + 1; j < group_size; j++) {
+				bool is_blanked;
 
-			if (pipe_set[j]->stream_res.opp->funcs->dpg_is_blanked)
-				is_blanked =
-					pipe_set[j]->stream_res.opp->funcs->dpg_is_blanked(pipe_set[j]->stream_res.opp);
-			else
-				is_blanked =
-					pipe_set[j]->stream_res.tg->funcs->is_blanked(pipe_set[j]->stream_res.tg);
-			if (!is_blanked) {
-				group_size--;
-				pipe_set[j] = pipe_set[group_size];
-				j--;
+				if (pipe_set[j]->stream_res.opp->funcs->dpg_is_blanked)
+					is_blanked =
+						pipe_set[j]->stream_res.opp->funcs->dpg_is_blanked(pipe_set[j]->stream_res.opp);
+				else
+					is_blanked =
+						pipe_set[j]->stream_res.tg->funcs->is_blanked(pipe_set[j]->stream_res.tg);
+				if (!is_blanked) {
+					group_size--;
+					pipe_set[j] = pipe_set[group_size];
+					j--;
+				}
 			}
 		}
 
 		if (group_size > 1) {
 			if (sync_type == TIMING_SYNCHRONIZABLE) {
 				dc->hwss.enable_timing_synchronization(
-					dc, group_index, group_size, pipe_set);
+					dc, ctx, group_index, group_size, pipe_set);
 			} else
 				if (sync_type == VBLANK_SYNCHRONIZABLE) {
 				dc->hwss.enable_vblanks_synchronization(
@@ -1453,24 +1758,26 @@ static void program_timing_sync(
 	}
 }
 
-static bool context_changed(
-		struct dc *dc,
-		struct dc_state *context)
+static bool streams_changed(struct dc *dc,
+			    struct dc_stream_state *streams[],
+			    uint8_t stream_count)
 {
 	uint8_t i;
 
-	if (context->stream_count != dc->current_state->stream_count)
+	if (stream_count != dc->current_state->stream_count)
 		return true;
 
 	for (i = 0; i < dc->current_state->stream_count; i++) {
-		if (dc->current_state->streams[i] != context->streams[i])
+		if (dc->current_state->streams[i] != streams[i])
+			return true;
+		if (!streams[i]->link->link_state_valid)
 			return true;
 	}
 
 	return false;
 }
 
-bool dc_validate_seamless_boot_timing(const struct dc *dc,
+bool dc_validate_boot_timing(const struct dc *dc,
 				const struct dc_sink *sink,
 				struct dc_crtc_timing *crtc_timing)
 {
@@ -1487,14 +1794,23 @@ bool dc_validate_seamless_boot_timing(const struct dc *dc,
 		return false;
 	}
 
+	if (dc->debug.force_odm_combine) {
+		DC_LOG_DEBUG("boot timing validation failed due to force_odm_combine\n");
+		return false;
+	}
+
 	/* Check for enabled DIG to identify enabled display */
-	if (!link->link_enc->funcs->is_dig_enabled(link->link_enc))
+	if (!link->link_enc->funcs->is_dig_enabled(link->link_enc)) {
+		DC_LOG_DEBUG("boot timing validation failed due to disabled DIG\n");
 		return false;
+	}
 
 	enc_inst = link->link_enc->funcs->get_dig_frontend(link->link_enc);
 
-	if (enc_inst == ENGINE_ID_UNKNOWN)
+	if (enc_inst == ENGINE_ID_UNKNOWN) {
+		DC_LOG_DEBUG("boot timing validation failed due to unknown DIG engine ID\n");
 		return false;
+	}
 
 	for (i = 0; i < dc->res_pool->stream_enc_count; i++) {
 		if (dc->res_pool->stream_enc[i]->id == enc_inst) {
@@ -1508,91 +1824,170 @@ bool dc_validate_seamless_boot_timing(const struct dc *dc,
 	}
 
 	// tg_inst not found
-	if (i == dc->res_pool->stream_enc_count)
+	if (i == dc->res_pool->stream_enc_count) {
+		DC_LOG_DEBUG("boot timing validation failed due to timing generator instance not found\n");
 		return false;
+	}
 
-	if (tg_inst >= dc->res_pool->timing_generator_count)
+	if (tg_inst >= dc->res_pool->timing_generator_count) {
+		DC_LOG_DEBUG("boot timing validation failed due to invalid timing generator count\n");
 		return false;
+	}
+
+	if (tg_inst != link->link_enc->preferred_engine) {
+		DC_LOG_DEBUG("boot timing validation failed due to non-preferred timing generator\n");
+		return false;
+	}
 
 	tg = dc->res_pool->timing_generators[tg_inst];
 
-	if (!tg->funcs->get_hw_timing)
+	if (!tg->funcs->get_hw_timing) {
+		DC_LOG_DEBUG("boot timing validation failed due to missing get_hw_timing callback\n");
 		return false;
+	}
 
-	if (!tg->funcs->get_hw_timing(tg, &hw_crtc_timing))
+	if (!tg->funcs->get_hw_timing(tg, &hw_crtc_timing)) {
+		DC_LOG_DEBUG("boot timing validation failed due to failed get_hw_timing return\n");
 		return false;
+	}
 
-	if (crtc_timing->h_total != hw_crtc_timing.h_total)
+	if (crtc_timing->h_total != hw_crtc_timing.h_total) {
+		DC_LOG_DEBUG("boot timing validation failed due to h_total mismatch\n");
 		return false;
+	}
 
-	if (crtc_timing->h_border_left != hw_crtc_timing.h_border_left)
+	if (crtc_timing->h_border_left != hw_crtc_timing.h_border_left) {
+		DC_LOG_DEBUG("boot timing validation failed due to h_border_left mismatch\n");
 		return false;
+	}
 
-	if (crtc_timing->h_addressable != hw_crtc_timing.h_addressable)
+	if (crtc_timing->h_addressable != hw_crtc_timing.h_addressable) {
+		DC_LOG_DEBUG("boot timing validation failed due to h_addressable mismatch\n");
 		return false;
+	}
 
-	if (crtc_timing->h_border_right != hw_crtc_timing.h_border_right)
+	if (crtc_timing->h_border_right != hw_crtc_timing.h_border_right) {
+		DC_LOG_DEBUG("boot timing validation failed due to h_border_right mismatch\n");
 		return false;
+	}
 
-	if (crtc_timing->h_front_porch != hw_crtc_timing.h_front_porch)
+	if (crtc_timing->h_front_porch != hw_crtc_timing.h_front_porch) {
+		DC_LOG_DEBUG("boot timing validation failed due to h_front_porch mismatch\n");
 		return false;
+	}
 
-	if (crtc_timing->h_sync_width != hw_crtc_timing.h_sync_width)
+	if (crtc_timing->h_sync_width != hw_crtc_timing.h_sync_width) {
+		DC_LOG_DEBUG("boot timing validation failed due to h_sync_width mismatch\n");
 		return false;
+	}
 
-	if (crtc_timing->v_total != hw_crtc_timing.v_total)
+	if (crtc_timing->v_total != hw_crtc_timing.v_total) {
+		DC_LOG_DEBUG("boot timing validation failed due to v_total mismatch\n");
 		return false;
+	}
 
-	if (crtc_timing->v_border_top != hw_crtc_timing.v_border_top)
+	if (crtc_timing->v_border_top != hw_crtc_timing.v_border_top) {
+		DC_LOG_DEBUG("boot timing validation failed due to v_border_top mismatch\n");
 		return false;
+	}
 
-	if (crtc_timing->v_addressable != hw_crtc_timing.v_addressable)
+	if (crtc_timing->v_addressable != hw_crtc_timing.v_addressable) {
+		DC_LOG_DEBUG("boot timing validation failed due to v_addressable mismatch\n");
 		return false;
+	}
 
-	if (crtc_timing->v_border_bottom != hw_crtc_timing.v_border_bottom)
+	if (crtc_timing->v_border_bottom != hw_crtc_timing.v_border_bottom) {
+		DC_LOG_DEBUG("boot timing validation failed due to v_border_bottom mismatch\n");
 		return false;
+	}
 
-	if (crtc_timing->v_front_porch != hw_crtc_timing.v_front_porch)
+	if (crtc_timing->v_front_porch != hw_crtc_timing.v_front_porch) {
+		DC_LOG_DEBUG("boot timing validation failed due to v_front_porch mismatch\n");
 		return false;
+	}
 
-	if (crtc_timing->v_sync_width != hw_crtc_timing.v_sync_width)
+	if (crtc_timing->v_sync_width != hw_crtc_timing.v_sync_width) {
+		DC_LOG_DEBUG("boot timing validation failed due to v_sync_width mismatch\n");
 		return false;
+	}
 
 	/* block DSC for now, as VBIOS does not currently support DSC timings */
-	if (crtc_timing->flags.DSC)
+	if (crtc_timing->flags.DSC) {
+		DC_LOG_DEBUG("boot timing validation failed due to DSC\n");
 		return false;
+	}
 
 	if (dc_is_dp_signal(link->connector_signal)) {
-		unsigned int pix_clk_100hz;
+		unsigned int pix_clk_100hz = 0;
+		uint32_t numOdmPipes = 1;
+		uint32_t id_src[4] = {0};
 
 		dc->res_pool->dp_clock_source->funcs->get_pixel_clk_frequency_100hz(
 			dc->res_pool->dp_clock_source,
 			tg_inst, &pix_clk_100hz);
 
-		if (crtc_timing->pix_clk_100hz != pix_clk_100hz)
+		if (tg->funcs->get_optc_source)
+			tg->funcs->get_optc_source(tg,
+						&numOdmPipes, &id_src[0], &id_src[1]);
+
+		if (numOdmPipes == 2) {
+			pix_clk_100hz *= 2;
+		} else if (numOdmPipes == 4) {
+			pix_clk_100hz *= 4;
+		} else if (se && se->funcs->get_pixels_per_cycle) {
+			uint32_t pixels_per_cycle = se->funcs->get_pixels_per_cycle(se);
+
+			if (pixels_per_cycle != 1 && !dc->debug.enable_dp_dig_pixel_rate_div_policy) {
+				DC_LOG_DEBUG("boot timing validation failed due to pixels_per_cycle\n");
+				return false;
+			}
+
+			pix_clk_100hz *= pixels_per_cycle;
+		}
+
+		// Note: In rare cases, HW pixclk may differ from crtc's pixclk
+		// slightly due to rounding issues in 10 kHz units.
+		if (crtc_timing->pix_clk_100hz != pix_clk_100hz) {
+			DC_LOG_DEBUG("boot timing validation failed due to pix_clk_100hz mismatch\n");
 			return false;
+		}
 
-		if (!se->funcs->dp_get_pixel_format)
+		if (!se || !se->funcs->dp_get_pixel_format) {
+			DC_LOG_DEBUG("boot timing validation failed due to missing dp_get_pixel_format\n");
 			return false;
+		}
 
 		if (!se->funcs->dp_get_pixel_format(
 			se,
 			&hw_crtc_timing.pixel_encoding,
-			&hw_crtc_timing.display_color_depth))
+			&hw_crtc_timing.display_color_depth)) {
+			DC_LOG_DEBUG("boot timing validation failed due to dp_get_pixel_format failure\n");
 			return false;
+		}
 
-		if (hw_crtc_timing.display_color_depth != crtc_timing->display_color_depth)
+		if (hw_crtc_timing.display_color_depth != crtc_timing->display_color_depth) {
+			DC_LOG_DEBUG("boot timing validation failed due to display_color_depth mismatch\n");
 			return false;
+		}
 
-		if (hw_crtc_timing.pixel_encoding != crtc_timing->pixel_encoding)
+		if (hw_crtc_timing.pixel_encoding != crtc_timing->pixel_encoding) {
+			DC_LOG_DEBUG("boot timing validation failed due to pixel_encoding mismatch\n");
 			return false;
+		}
 	}
 
+
 	if (link->dpcd_caps.dprx_feature.bits.VSC_SDP_COLORIMETRY_SUPPORTED) {
+		DC_LOG_DEBUG("boot timing validation failed due to VSC SDP colorimetry\n");
+		return false;
+	}
+
+	if (link->dpcd_caps.channel_coding_cap.bits.DP_128b_132b_SUPPORTED) {
+		DC_LOG_DEBUG("boot timing validation failed due to DP 128b/132b\n");
 		return false;
 	}
 
-	if (is_edp_ilr_optimization_required(link, crtc_timing)) {
+	if (dc->link_srv->edp_is_ilr_optimization_required(link, crtc_timing)) {
 		DC_LOG_EVENT_LINK_TRAINING("Seamless boot disabled to optimize eDP link rate\n");
 		return false;
 	}
@@ -1625,6 +2020,8 @@ void dc_enable_stereo(
 	int i, j;
 	struct pipe_ctx *pipe;
 
+	dc_exit_ips_for_hw_access(dc);
+
 	for (i = 0; i < MAX_PIPES; i++) {
 		if (context != NULL) {
 			pipe = &context->res_ctx.pipe_ctx[i];
@@ -1644,6 +2041,8 @@ void dc_enable_stereo(
 void dc_trigger_sync(struct dc *dc, struct dc_state *context)
 {
 	if (context->stream_count > 1 && !dc->debug.disable_timing_sync) {
+		dc_exit_ips_for_hw_access(dc);
+
 		enable_timing_multisync(dc, context);
 		program_timing_sync(dc, context);
 	}
@@ -1662,7 +2061,6 @@ static uint8_t get_stream_mask(struct dc *dc, struct dc_state *context)
 	return stream_mask;
 }
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 void dc_z10_restore(const struct dc *dc)
 {
 	if (dc->hwss.z10_restore)
@@ -1674,10 +2072,52 @@ void dc_z10_save_init(struct dc *dc)
 	if (dc->hwss.z10_save_init)
 		dc->hwss.z10_save_init(dc);
 }
-#endif
-/*
- * Applies given context to HW and copy it into current context.
+
+/* Set a pipe unlock order based on the change in DET allocation and stores it in dc scratch memory
+ * Prevents over allocation of DET during unlock process
+ * e.g. 2 pipe config with different streams with a max of 20 DET segments
+ *	Before:								After:
+ *		- Pipe0: 10 DET segments			- Pipe0: 12 DET segments
+ *		- Pipe1: 10 DET segments			- Pipe1: 8 DET segments
+ * If Pipe0 gets updated first, 22 DET segments will be allocated
+ */
+static void determine_pipe_unlock_order(struct dc *dc, struct dc_state *context)
+{
+	unsigned int i = 0;
+	struct pipe_ctx *pipe = NULL;
+	struct timing_generator *tg = NULL;
+
+	if (!dc->config.set_pipe_unlock_order)
+		return;
+
+	memset(dc->scratch.pipes_to_unlock_first, 0, sizeof(dc->scratch.pipes_to_unlock_first));
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		pipe = &context->res_ctx.pipe_ctx[i];
+		tg = pipe->stream_res.tg;
+
+		if (!resource_is_pipe_type(pipe, OTG_MASTER) ||
+				!tg->funcs->is_tg_enabled(tg) ||
+				dc_state_get_pipe_subvp_type(context, pipe) == SUBVP_PHANTOM) {
+			continue;
+		}
+
+		if (resource_calculate_det_for_stream(context, pipe) <
+				resource_calculate_det_for_stream(dc->current_state, &dc->current_state->res_ctx.pipe_ctx[i])) {
+			dc->scratch.pipes_to_unlock_first[i] = true;
+		}
+	}
+}
+
+/**
+ * dc_commit_state_no_check - Apply context to the hardware
+ *
+ * @dc: DC object with the current status to be updated
+ * @context: New state that will become the current status at the end of this function
+ *
+ * Applies given context to the hardware and copy it into current context.
  * It's up to the user to release the src context afterwards.
+ *
+ * Return: an enum dc_status result code for the operation
  */
 static enum dc_status dc_commit_state_no_check(struct dc *dc, struct dc_state *context)
 {
@@ -1686,11 +2126,20 @@ static enum dc_status dc_commit_state_no_check(struct dc *dc, struct dc_state *c
 	struct pipe_ctx *pipe;
 	int i, k, l;
 	struct dc_stream_state *dc_streams[MAX_STREAMS] = {0};
+	struct dc_state *old_state;
+	bool subvp_prev_use = false;
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	dc_z10_restore(dc);
 	dc_allow_idle_optimizations(dc, false);
-#endif
+
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		struct pipe_ctx *old_pipe = &dc->current_state->res_ctx.pipe_ctx[i];
+
+		/* Check old context for SubVP */
+		subvp_prev_use |= (dc_state_get_pipe_subvp_type(dc->current_state, old_pipe) == SUBVP_PHANTOM);
+		if (subvp_prev_use)
+			break;
+	}
 
 	for (i = 0; i < context->stream_count; i++)
 		dc_streams[i] =  context->streams[i];
@@ -1698,12 +2147,43 @@ static enum dc_status dc_commit_state_no_check(struct dc *dc, struct dc_state *c
 	if (!dcb->funcs->is_accelerated_mode(dcb)) {
 		disable_vbios_mode_if_required(dc, context);
 		dc->hwss.enable_accelerated_mode(dc, context);
+	} else if (get_seamless_boot_stream_count(dc->current_state) > 0) {
+		/* If the previous Stream still retains the apply seamless boot flag,
+		 * it means the OS has not actually performed a flip yet.
+		 * At this point, if we receive dc_commit_streams again, we should
+		 * once more check whether the actual HW timing matches what the OS
+		 * has provided
+		 */
+		disable_vbios_mode_if_required(dc, context);
+	}
+
+	if (dc->hwseq->funcs.wait_for_pipe_update_if_needed) {
+		for (i = 0; i < dc->res_pool->pipe_count; i++) {
+			pipe = &context->res_ctx.pipe_ctx[i];
+			//Only delay otg master for a given config
+			if (resource_is_pipe_type(pipe, OTG_MASTER)) {
+				//dc_commit_state_no_check is always a full update
+				dc->hwseq->funcs.wait_for_pipe_update_if_needed(dc, pipe, false);
+				break;
+			}
+		}
 	}
 
 	if (context->stream_count > get_seamless_boot_stream_count(context) ||
 		context->stream_count == 0)
 		dc->hwss.prepare_bandwidth(dc, context);
 
+	/* When SubVP is active, all HW programming must be done while
+	 * SubVP lock is acquired
+	 */
+	if (dc->hwss.subvp_pipe_control_lock)
+		dc->hwss.subvp_pipe_control_lock(dc, context, true, true, NULL, subvp_prev_use);
+	if (dc->hwss.dmub_hw_control_lock)
+		dc->hwss.dmub_hw_control_lock(dc, context, true);
+
+	if (dc->hwss.update_dsc_pg)
+		dc->hwss.update_dsc_pg(dc, context, false);
+
 	disable_dangling_plane(dc, context);
 	/* re-program planes for existing stream, in case we need to
 	 * free up plane resource for later use
@@ -1728,20 +2208,56 @@ static enum dc_status dc_commit_state_no_check(struct dc *dc, struct dc_state *c
 		dc->hwss.wait_for_mpcc_disconnect(dc, dc->res_pool, pipe);
 	}
 
+	for (i = 0; i < dc->current_state->stream_count; i++)
+		dc_dmub_srv_control_cursor_offload(dc, dc->current_state, dc->current_state->streams[i], false);
+
 	result = dc->hwss.apply_ctx_to_hw(dc, context);
 
-	if (result != DC_OK)
+	for (i = 0; i < context->stream_count; i++)
+		dc_dmub_srv_control_cursor_offload(dc, context, context->streams[i], true);
+
+	if (result != DC_OK) {
+		/* Application of dc_state to hardware stopped. */
+		dc->current_state->res_ctx.link_enc_cfg_ctx.mode = LINK_ENC_CFG_STEADY;
 		return result;
+	}
 
 	dc_trigger_sync(dc, context);
 
+	/* Full update should unconditionally be triggered when dc_commit_state_no_check is called */
+	for (i = 0; i < context->stream_count; i++) {
+		uint32_t prev_dsc_changed = context->streams[i]->update_flags.bits.dsc_changed;
+
+		context->streams[i]->update_flags.raw = 0xFFFFFFFF;
+		context->streams[i]->update_flags.bits.dsc_changed = prev_dsc_changed;
+	}
+
+	determine_pipe_unlock_order(dc, context);
 	/* Program all planes within new context*/
+	if (dc->res_pool->funcs->prepare_mcache_programming)
+		dc->res_pool->funcs->prepare_mcache_programming(dc, context);
 	if (dc->hwss.program_front_end_for_ctx) {
 		dc->hwss.interdependent_update_lock(dc, context, true);
 		dc->hwss.program_front_end_for_ctx(dc, context);
+
+		if (dc->hwseq->funcs.set_wait_for_update_needed_for_pipe) {
+			for (i = 0; i < dc->res_pool->pipe_count; i++) {
+				pipe = &context->res_ctx.pipe_ctx[i];
+				dc->hwseq->funcs.set_wait_for_update_needed_for_pipe(dc, pipe);
+			}
+		}
+
 		dc->hwss.interdependent_update_lock(dc, context, false);
 		dc->hwss.post_unlock_program_front_end(dc, context);
 	}
+
+	if (dc->hwss.commit_subvp_config)
+		dc->hwss.commit_subvp_config(dc, context);
+	if (dc->hwss.subvp_pipe_control_lock)
+		dc->hwss.subvp_pipe_control_lock(dc, context, false, true, NULL, subvp_prev_use);
+	if (dc->hwss.dmub_hw_control_lock)
+		dc->hwss.dmub_hw_control_lock(dc, context, false);
+
 	for (i = 0; i < context->stream_count; i++) {
 		const struct dc_link *link = context->streams[i]->link;
 
@@ -1783,14 +2299,26 @@ static enum dc_status dc_commit_state_no_check(struct dc *dc, struct dc_state *c
 
 	dc_enable_stereo(dc, context, dc_streams, context->stream_count);
 
-	if (context->stream_count > get_seamless_boot_stream_count(context) ||
+	if (get_seamless_boot_stream_count(context) == 0 ||
 		context->stream_count == 0) {
 		/* Must wait for no flips to be pending before doing optimize bw */
-		wait_for_no_pipes_pending(dc, context);
+		hwss_wait_for_no_pipes_pending(dc, context);
+		/*
+		 * optimized dispclk depends on ODM setup. Need to wait for ODM
+		 * update pending complete before optimizing bandwidth.
+		 */
+		hwss_wait_for_odm_update_pending_complete(dc, context);
 		/* pplib is notified if disp_num changed */
 		dc->hwss.optimize_bandwidth(dc, context);
+		/* Need to do otg sync again as otg could be out of sync due to otg
+		 * workaround applied during clock update
+		 */
+		dc_trigger_sync(dc, context);
 	}
 
+	if (dc->hwss.update_dsc_pg)
+		dc->hwss.update_dsc_pg(dc, context, true);
+
 	if (dc->ctx->dce_version >= DCE_VERSION_MAX)
 		TRACE_DCN_CLOCK_STATE(&context->bw_ctx.bw.dcn.clk);
 	else
@@ -1804,38 +2332,162 @@ static enum dc_status dc_commit_state_no_check(struct dc *dc, struct dc_state *c
 	for (i = 0; i < context->stream_count; i++)
 		context->streams[i]->mode_changed = false;
 
-	dc_release_state(dc->current_state);
+	/* Clear update flags that were set earlier to avoid redundant programming */
+	for (i = 0; i < context->stream_count; i++) {
+		context->streams[i]->update_flags.raw = 0x0;
+	}
 
+	old_state = dc->current_state;
 	dc->current_state = context;
 
-	dc_retain_state(dc->current_state);
+	dc_state_release(old_state);
+
+	dc_state_retain(dc->current_state);
 
 	return result;
 }
 
-bool dc_commit_state(struct dc *dc, struct dc_state *context)
+static bool commit_minimal_transition_state(struct dc *dc,
+		struct dc_state *transition_base_context);
+
+/**
+ * dc_commit_streams - Commit current stream state
+ *
+ * @dc: DC object with the commit state to be configured in the hardware
+ * @params: Parameters for the commit, including the streams to be committed
+ *
+ * Function responsible for commit streams change to the hardware.
+ *
+ * Return:
+ * Return DC_OK if everything work as expected, otherwise, return a dc_status
+ * code.
+ */
+enum dc_status dc_commit_streams(struct dc *dc, struct dc_commit_streams_params *params)
 {
-	enum dc_status result = DC_ERROR_UNEXPECTED;
-	int i;
+	int i, j;
+	struct dc_state *context;
+	enum dc_status res = DC_OK;
+	struct dc_validation_set set[MAX_STREAMS] = {0};
+	struct pipe_ctx *pipe;
+	bool handle_exit_odm2to1 = false;
+
+	if (!params)
+		return DC_ERROR_UNEXPECTED;
 
-	if (!context_changed(dc, context))
-		return DC_OK;
+	if (dc->ctx->dce_environment == DCE_ENV_VIRTUAL_HW)
+		return res;
+
+	if (!streams_changed(dc, params->streams, params->stream_count) &&
+			dc->current_state->power_source == params->power_source)
+		return res;
 
-	DC_LOG_DC("%s: %d streams\n",
-				__func__, context->stream_count);
+	dc_exit_ips_for_hw_access(dc);
+
+	DC_LOG_DC("%s: %d streams\n", __func__, params->stream_count);
+
+	for (i = 0; i < params->stream_count; i++) {
+		struct dc_stream_state *stream = params->streams[i];
+		struct dc_stream_status *status = dc_stream_get_status(stream);
+		struct dc_sink *sink = stream->sink;
+
+		/* revalidate streams */
+		if (!dc_is_virtual_signal(sink->sink_signal)) {
+			res = dc_validate_stream(dc, stream);
+			if (res != DC_OK)
+				return res;
+		}
 
-	for (i = 0; i < context->stream_count; i++) {
-		struct dc_stream_state *stream = context->streams[i];
 
 		dc_stream_log(dc, stream);
+
+		set[i].stream = stream;
+
+		if (status) {
+			set[i].plane_count = status->plane_count;
+			for (j = 0; j < status->plane_count; j++)
+				set[i].plane_states[j] = status->plane_states[j];
+		}
+	}
+
+	/* ODM Combine 2:1 power optimization is only applied for single stream
+	 * scenario, it uses extra pipes than needed to reduce power consumption
+	 * We need to switch off this feature to make room for new streams.
+	 */
+	if (params->stream_count > dc->current_state->stream_count &&
+			dc->current_state->stream_count == 1) {
+		for (i = 0; i < dc->res_pool->pipe_count; i++) {
+			pipe = &dc->current_state->res_ctx.pipe_ctx[i];
+			if (pipe->next_odm_pipe)
+				handle_exit_odm2to1 = true;
+		}
+	}
+
+	if (handle_exit_odm2to1)
+		res = commit_minimal_transition_state(dc, dc->current_state);
+
+	context = dc_state_create_current_copy(dc);
+	if (!context)
+		goto context_alloc_fail;
+
+	context->power_source = params->power_source;
+
+	res = dc_validate_with_context(dc, set, params->stream_count, context, DC_VALIDATE_MODE_AND_PROGRAMMING);
+
+	/*
+	 * Only update link encoder to stream assignment after bandwidth validation passed.
+	 */
+	if (res == DC_OK && dc->res_pool->funcs->link_encs_assign && !dc->config.unify_link_enc_assignment)
+		dc->res_pool->funcs->link_encs_assign(
+			dc, context, context->streams, context->stream_count);
+
+	if (res != DC_OK) {
+		BREAK_TO_DEBUGGER();
+		goto fail;
 	}
 
-	result = dc_commit_state_no_check(dc, context);
+	/*
+	 * If not already seamless, make transition seamless by inserting intermediate minimal transition
+	 */
+	if (dc->hwss.is_pipe_topology_transition_seamless &&
+			!dc->hwss.is_pipe_topology_transition_seamless(dc, dc->current_state, context)) {
+		res = commit_minimal_transition_state(dc, context);
+		if (res != DC_OK) {
+			BREAK_TO_DEBUGGER();
+			goto fail;
+		}
+	}
+
+	res = dc_commit_state_no_check(dc, context);
+
+	for (i = 0; i < params->stream_count; i++) {
+		for (j = 0; j < context->stream_count; j++) {
+			if (params->streams[i]->stream_id == context->streams[j]->stream_id)
+				params->streams[i]->out.otg_offset = context->stream_status[j].primary_otg_inst;
+
+			if (dc_is_embedded_signal(params->streams[i]->signal)) {
+				struct dc_stream_status *status = dc_state_get_stream_status(context, params->streams[i]);
+
+				if (!status)
+					continue;
+
+				if (dc->hwss.is_abm_supported)
+					status->is_abm_supported = dc->hwss.is_abm_supported(dc, context, params->streams[i]);
+				else
+					status->is_abm_supported = true;
+			}
+		}
+	}
 
-	return (result == DC_OK);
+fail:
+	dc_state_release(context);
+
+context_alloc_fail:
+
+	DC_LOG_DC("%s Finished.\n", __func__);
+
+	return res;
 }
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 bool dc_acquire_release_mpc_3dlut(
 		struct dc *dc, bool acquire,
 		struct dc_stream_state *stream,
@@ -1871,7 +2523,7 @@ bool dc_acquire_release_mpc_3dlut(
 	}
 	return ret;
 }
-#endif
+
 static bool is_flip_pending_in_pipes(struct dc *dc, struct dc_state *context)
 {
 	int i;
@@ -1880,7 +2532,8 @@ static bool is_flip_pending_in_pipes(struct dc *dc, struct dc_state *context)
 	for (i = 0; i < MAX_PIPES; i++) {
 		pipe = &context->res_ctx.pipe_ctx[i];
 
-		if (!pipe->plane_state)
+		// Don't check flip pending on phantom pipes
+		if (!pipe->plane_state || (dc_state_get_pipe_subvp_type(context, pipe) == SUBVP_PHANTOM))
 			continue;
 
 		/* Must set to false to start with, due to OR in update function */
@@ -1892,7 +2545,6 @@ static bool is_flip_pending_in_pipes(struct dc *dc, struct dc_state *context)
 	return false;
 }
 
-#ifdef CONFIG_DRM_AMD_DC_DCN
 /* Perform updates here which need to be deferred until next vupdate
  *
  * i.e. blnd lut, 3dlut, and shaper lut bypass regs are double buffered
@@ -1911,7 +2563,6 @@ static void process_deferred_updates(struct dc *dc)
 				dc->res_pool->dpps[i]->funcs->dpp_deferred_update(dc->res_pool->dpps[i]);
 	}
 }
-#endif /* CONFIG_DRM_AMD_DC_DCN */
 
 void dc_post_update_surfaces_to_stream(struct dc *dc)
 {
@@ -1923,110 +2574,35 @@ void dc_post_update_surfaces_to_stream(struct dc *dc)
 
 	post_surface_trace(dc);
 
-	if (dc->ctx->dce_version >= DCE_VERSION_MAX)
-		TRACE_DCN_CLOCK_STATE(&context->bw_ctx.bw.dcn.clk);
-	else
-		TRACE_DCE_CLOCK_STATE(&context->bw_ctx.bw.dce);
-
-	if (is_flip_pending_in_pipes(dc, context))
-		return;
-
-	for (i = 0; i < dc->res_pool->pipe_count; i++)
-		if (context->res_ctx.pipe_ctx[i].stream == NULL ||
-		    context->res_ctx.pipe_ctx[i].plane_state == NULL) {
-			context->res_ctx.pipe_ctx[i].pipe_idx = i;
-			dc->hwss.disable_plane(dc, &context->res_ctx.pipe_ctx[i]);
-		}
-
-#ifdef CONFIG_DRM_AMD_DC_DCN
-	process_deferred_updates(dc);
-#endif
-
-	dc->hwss.optimize_bandwidth(dc, context);
-
-	dc->optimized_required = false;
-	dc->wm_optimized_required = false;
-}
-
-static void init_state(struct dc *dc, struct dc_state *context)
-{
-	/* Each context must have their own instance of VBA and in order to
-	 * initialize and obtain IP and SOC the base DML instance from DC is
-	 * initially copied into every context
+	/*
+	 * Only relevant for DCN behavior where we can guarantee the optimization
+	 * is safe to apply - retain the legacy behavior for DCE.
 	 */
-#ifdef CONFIG_DRM_AMD_DC_DCN
-	memcpy(&context->bw_ctx.dml, &dc->dml, sizeof(struct display_mode_lib));
-#endif
-}
 
-struct dc_state *dc_create_state(struct dc *dc)
-{
-	struct dc_state *context = kvzalloc(sizeof(struct dc_state),
-					    GFP_KERNEL);
-
-	if (!context)
-		return NULL;
-
-	init_state(dc, context);
-
-	kref_init(&context->refcount);
-
-	return context;
-}
-
-struct dc_state *dc_copy_state(struct dc_state *src_ctx)
-{
-	int i, j;
-	struct dc_state *new_ctx = kvmalloc(sizeof(struct dc_state), GFP_KERNEL);
-
-	if (!new_ctx)
-		return NULL;
-	memcpy(new_ctx, src_ctx, sizeof(struct dc_state));
-
-	for (i = 0; i < MAX_PIPES; i++) {
-			struct pipe_ctx *cur_pipe = &new_ctx->res_ctx.pipe_ctx[i];
-
-			if (cur_pipe->top_pipe)
-				cur_pipe->top_pipe =  &new_ctx->res_ctx.pipe_ctx[cur_pipe->top_pipe->pipe_idx];
+	if (dc->ctx->dce_version < DCE_VERSION_MAX)
+		TRACE_DCE_CLOCK_STATE(&context->bw_ctx.bw.dce);
+	else {
+		TRACE_DCN_CLOCK_STATE(&context->bw_ctx.bw.dcn.clk);
 
-			if (cur_pipe->bottom_pipe)
-				cur_pipe->bottom_pipe = &new_ctx->res_ctx.pipe_ctx[cur_pipe->bottom_pipe->pipe_idx];
+		if (is_flip_pending_in_pipes(dc, context))
+			return;
 
-			if (cur_pipe->prev_odm_pipe)
-				cur_pipe->prev_odm_pipe =  &new_ctx->res_ctx.pipe_ctx[cur_pipe->prev_odm_pipe->pipe_idx];
+		for (i = 0; i < dc->res_pool->pipe_count; i++)
+			if (context->res_ctx.pipe_ctx[i].stream == NULL ||
+					context->res_ctx.pipe_ctx[i].plane_state == NULL) {
+				context->res_ctx.pipe_ctx[i].pipe_idx = i;
+				dc->hwss.disable_plane(dc, context, &context->res_ctx.pipe_ctx[i]);
+			}
 
-			if (cur_pipe->next_odm_pipe)
-				cur_pipe->next_odm_pipe = &new_ctx->res_ctx.pipe_ctx[cur_pipe->next_odm_pipe->pipe_idx];
+		process_deferred_updates(dc);
 
-	}
+		dc->hwss.optimize_bandwidth(dc, context);
 
-	for (i = 0; i < new_ctx->stream_count; i++) {
-			dc_stream_retain(new_ctx->streams[i]);
-			for (j = 0; j < new_ctx->stream_status[i].plane_count; j++)
-				dc_plane_state_retain(
-					new_ctx->stream_status[i].plane_states[j]);
+		if (dc->hwss.update_dsc_pg)
+			dc->hwss.update_dsc_pg(dc, context, true);
 	}
 
-	kref_init(&new_ctx->refcount);
-
-	return new_ctx;
-}
-
-void dc_retain_state(struct dc_state *context)
-{
-	kref_get(&context->refcount);
-}
-
-static void dc_state_free(struct kref *kref)
-{
-	struct dc_state *context = container_of(kref, struct dc_state, refcount);
-	dc_resource_state_destruct(context);
-	kvfree(context);
-}
-
-void dc_release_state(struct dc_state *context)
-{
-	kref_put(&context->refcount, dc_state_free);
+	dc->optimized_required = false;
 }
 
 bool dc_set_generic_gpio_for_stereo(bool enable,
@@ -2095,47 +2671,50 @@ static bool is_surface_in_context(
 	return false;
 }
 
-static enum surface_update_type get_plane_info_update_type(const struct dc_surface_update *u)
+static struct surface_update_descriptor get_plane_info_update_type(const struct dc_surface_update *u)
 {
 	union surface_update_flags *update_flags = &u->surface->update_flags;
-	enum surface_update_type update_type = UPDATE_TYPE_FAST;
+	struct surface_update_descriptor update_type = { UPDATE_TYPE_FAST, LOCK_DESCRIPTOR_NONE };
 
 	if (!u->plane_info)
-		return UPDATE_TYPE_FAST;
+		return update_type;
+
+	// `plane_info` present means at least `STREAM` lock is required
+	elevate_update_type(&update_type, UPDATE_TYPE_FAST, LOCK_DESCRIPTOR_STREAM);
 
 	if (u->plane_info->color_space != u->surface->color_space) {
 		update_flags->bits.color_space_change = 1;
-		elevate_update_type(&update_type, UPDATE_TYPE_MED);
+		elevate_update_type(&update_type, UPDATE_TYPE_MED, LOCK_DESCRIPTOR_STREAM);
 	}
 
 	if (u->plane_info->horizontal_mirror != u->surface->horizontal_mirror) {
 		update_flags->bits.horizontal_mirror_change = 1;
-		elevate_update_type(&update_type, UPDATE_TYPE_MED);
+		elevate_update_type(&update_type, UPDATE_TYPE_MED, LOCK_DESCRIPTOR_STREAM);
 	}
 
 	if (u->plane_info->rotation != u->surface->rotation) {
 		update_flags->bits.rotation_change = 1;
-		elevate_update_type(&update_type, UPDATE_TYPE_FULL);
+		elevate_update_type(&update_type, UPDATE_TYPE_FULL, LOCK_DESCRIPTOR_GLOBAL);
 	}
 
 	if (u->plane_info->format != u->surface->format) {
 		update_flags->bits.pixel_format_change = 1;
-		elevate_update_type(&update_type, UPDATE_TYPE_FULL);
+		elevate_update_type(&update_type, UPDATE_TYPE_FULL, LOCK_DESCRIPTOR_GLOBAL);
 	}
 
 	if (u->plane_info->stereo_format != u->surface->stereo_format) {
 		update_flags->bits.stereo_format_change = 1;
-		elevate_update_type(&update_type, UPDATE_TYPE_FULL);
+		elevate_update_type(&update_type, UPDATE_TYPE_FULL, LOCK_DESCRIPTOR_GLOBAL);
 	}
 
 	if (u->plane_info->per_pixel_alpha != u->surface->per_pixel_alpha) {
 		update_flags->bits.per_pixel_alpha_change = 1;
-		elevate_update_type(&update_type, UPDATE_TYPE_MED);
+		elevate_update_type(&update_type, UPDATE_TYPE_MED, LOCK_DESCRIPTOR_STREAM);
 	}
 
 	if (u->plane_info->global_alpha_value != u->surface->global_alpha_value) {
 		update_flags->bits.global_alpha_change = 1;
-		elevate_update_type(&update_type, UPDATE_TYPE_MED);
+		elevate_update_type(&update_type, UPDATE_TYPE_MED, LOCK_DESCRIPTOR_STREAM);
 	}
 
 	if (u->plane_info->dcc.enable != u->surface->dcc.enable
@@ -2147,7 +2726,7 @@ static enum surface_update_type get_plane_info_update_type(const struct dc_surfa
 		 * recalculate stutter period.
 		 */
 		update_flags->bits.dcc_change = 1;
-		elevate_update_type(&update_type, UPDATE_TYPE_FULL);
+		elevate_update_type(&update_type, UPDATE_TYPE_FULL, LOCK_DESCRIPTOR_GLOBAL);
 	}
 
 	if (resource_pixel_format_to_bpp(u->plane_info->format) !=
@@ -2156,30 +2735,41 @@ static enum surface_update_type get_plane_info_update_type(const struct dc_surfa
 		 * and DML calculation
 		 */
 		update_flags->bits.bpp_change = 1;
-		elevate_update_type(&update_type, UPDATE_TYPE_FULL);
+		elevate_update_type(&update_type, UPDATE_TYPE_FULL, LOCK_DESCRIPTOR_GLOBAL);
 	}
 
 	if (u->plane_info->plane_size.surface_pitch != u->surface->plane_size.surface_pitch
 			|| u->plane_info->plane_size.chroma_pitch != u->surface->plane_size.chroma_pitch) {
 		update_flags->bits.plane_size_change = 1;
-		elevate_update_type(&update_type, UPDATE_TYPE_MED);
+		elevate_update_type(&update_type, UPDATE_TYPE_MED, LOCK_DESCRIPTOR_STREAM);
 	}
 
+	const struct dc_tiling_info *tiling = &u->plane_info->tiling_info;
 
-	if (memcmp(&u->plane_info->tiling_info, &u->surface->tiling_info,
-			sizeof(union dc_tiling_info)) != 0) {
+	if (memcmp(tiling, &u->surface->tiling_info, sizeof(*tiling)) != 0) {
 		update_flags->bits.swizzle_change = 1;
-		elevate_update_type(&update_type, UPDATE_TYPE_MED);
-
-		/* todo: below are HW dependent, we should add a hook to
-		 * DCE/N resource and validated there.
-		 */
-		if (u->plane_info->tiling_info.gfx9.swizzle != DC_SW_LINEAR) {
-			/* swizzled mode requires RQ to be setup properly,
-			 * thus need to run DML to calculate RQ settings
-			 */
-			update_flags->bits.bandwidth_change = 1;
-			elevate_update_type(&update_type, UPDATE_TYPE_FULL);
+		elevate_update_type(&update_type, UPDATE_TYPE_MED, LOCK_DESCRIPTOR_STREAM);
+
+		switch (tiling->gfxversion) {
+		case DcGfxVersion9:
+		case DcGfxVersion10:
+		case DcGfxVersion11:
+			if (tiling->gfx9.swizzle != DC_SW_LINEAR) {
+				update_flags->bits.bandwidth_change = 1;
+				elevate_update_type(&update_type, UPDATE_TYPE_FULL, LOCK_DESCRIPTOR_GLOBAL);
+			}
+			break;
+		case DcGfxAddr3:
+			if (tiling->gfx_addr3.swizzle != DC_ADDR3_SW_LINEAR) {
+				update_flags->bits.bandwidth_change = 1;
+				elevate_update_type(&update_type, UPDATE_TYPE_FULL, LOCK_DESCRIPTOR_GLOBAL);
+			}
+			break;
+		case DcGfxVersion7:
+		case DcGfxVersion8:
+		case DcGfxVersionUnknown:
+		default:
+			break;
 		}
 	}
 
@@ -2187,22 +2777,34 @@ static enum surface_update_type get_plane_info_update_type(const struct dc_surfa
 	return update_type;
 }
 
-static enum surface_update_type get_scaling_info_update_type(
+static struct surface_update_descriptor get_scaling_info_update_type(
+		const struct dc_check_config *check_config,
 		const struct dc_surface_update *u)
 {
 	union surface_update_flags *update_flags = &u->surface->update_flags;
+	struct surface_update_descriptor update_type = { UPDATE_TYPE_FAST, LOCK_DESCRIPTOR_NONE };
 
 	if (!u->scaling_info)
-		return UPDATE_TYPE_FAST;
+		return update_type;
 
-	if (u->scaling_info->clip_rect.width != u->surface->clip_rect.width
-			|| u->scaling_info->clip_rect.height != u->surface->clip_rect.height
+	// `scaling_info` present means at least `STREAM` lock is required
+	elevate_update_type(&update_type, UPDATE_TYPE_FAST, LOCK_DESCRIPTOR_STREAM);
+
+	if (u->scaling_info->src_rect.width != u->surface->src_rect.width
+			|| u->scaling_info->src_rect.height != u->surface->src_rect.height
 			|| u->scaling_info->dst_rect.width != u->surface->dst_rect.width
 			|| u->scaling_info->dst_rect.height != u->surface->dst_rect.height
+			|| u->scaling_info->clip_rect.width != u->surface->clip_rect.width
+			|| u->scaling_info->clip_rect.height != u->surface->clip_rect.height
 			|| u->scaling_info->scaling_quality.integer_scaling !=
-				u->surface->scaling_quality.integer_scaling
-			) {
+					u->surface->scaling_quality.integer_scaling) {
 		update_flags->bits.scaling_change = 1;
+		elevate_update_type(&update_type, UPDATE_TYPE_FULL, LOCK_DESCRIPTOR_GLOBAL);
+
+		if (u->scaling_info->src_rect.width > u->surface->src_rect.width
+				|| u->scaling_info->src_rect.height > u->surface->src_rect.height)
+			/* Making src rect bigger requires a bandwidth change */
+			update_flags->bits.clock_change = 1;
 
 		if ((u->scaling_info->dst_rect.width < u->surface->dst_rect.width
 			|| u->scaling_info->dst_rect.height < u->surface->dst_rect.height)
@@ -2210,16 +2812,12 @@ static enum surface_update_type get_scaling_info_update_type(
 					|| u->scaling_info->dst_rect.height < u->surface->src_rect.height))
 			/* Making dst rect smaller requires a bandwidth change */
 			update_flags->bits.bandwidth_change = 1;
-	}
 
-	if (u->scaling_info->src_rect.width != u->surface->src_rect.width
-		|| u->scaling_info->src_rect.height != u->surface->src_rect.height) {
-
-		update_flags->bits.scaling_change = 1;
-		if (u->scaling_info->src_rect.width > u->surface->src_rect.width
-				|| u->scaling_info->src_rect.height > u->surface->src_rect.height)
-			/* Making src rect bigger requires a bandwidth change */
-			update_flags->bits.clock_change = 1;
+		if (u->scaling_info->src_rect.width > check_config->max_optimizable_video_width &&
+			(u->scaling_info->clip_rect.width > u->surface->clip_rect.width ||
+			 u->scaling_info->clip_rect.height > u->surface->clip_rect.height))
+			 /* Changing clip size of a large surface may result in MPC slice count change */
+			update_flags->bits.bandwidth_change = 1;
 	}
 
 	if (u->scaling_info->src_rect.x != u->surface->src_rect.x
@@ -2227,121 +2825,156 @@ static enum surface_update_type get_scaling_info_update_type(
 			|| u->scaling_info->clip_rect.x != u->surface->clip_rect.x
 			|| u->scaling_info->clip_rect.y != u->surface->clip_rect.y
 			|| u->scaling_info->dst_rect.x != u->surface->dst_rect.x
-			|| u->scaling_info->dst_rect.y != u->surface->dst_rect.y)
+			|| u->scaling_info->dst_rect.y != u->surface->dst_rect.y) {
+		elevate_update_type(&update_type, UPDATE_TYPE_MED, LOCK_DESCRIPTOR_STREAM);
 		update_flags->bits.position_change = 1;
+	}
 
-	if (update_flags->bits.clock_change
-			|| update_flags->bits.bandwidth_change
-			|| update_flags->bits.scaling_change)
-		return UPDATE_TYPE_FULL;
-
-	if (update_flags->bits.position_change)
-		return UPDATE_TYPE_MED;
-
-	return UPDATE_TYPE_FAST;
+	return update_type;
 }
 
-static enum surface_update_type det_surface_update(const struct dc *dc,
-		const struct dc_surface_update *u)
+static struct surface_update_descriptor det_surface_update(
+		const struct dc_check_config *check_config,
+		struct dc_surface_update *u)
 {
-	const struct dc_state *context = dc->current_state;
-	enum surface_update_type type;
-	enum surface_update_type overall_type = UPDATE_TYPE_FAST;
+	struct surface_update_descriptor overall_type = { UPDATE_TYPE_FAST, LOCK_DESCRIPTOR_NONE };
 	union surface_update_flags *update_flags = &u->surface->update_flags;
 
-	if (u->flip_addr)
-		update_flags->bits.addr_update = 1;
-
-	if (!is_surface_in_context(context, u->surface) || u->surface->force_full_update) {
+	if (u->surface->force_full_update) {
 		update_flags->raw = 0xFFFFFFFF;
-		return UPDATE_TYPE_FULL;
+		elevate_update_type(&overall_type, UPDATE_TYPE_FULL, LOCK_DESCRIPTOR_GLOBAL);
+		return overall_type;
 	}
 
 	update_flags->raw = 0; // Reset all flags
 
-	type = get_plane_info_update_type(u);
-	elevate_update_type(&overall_type, type);
+	struct surface_update_descriptor inner_type = get_plane_info_update_type(u);
+
+	elevate_update_type(&overall_type, inner_type.update_type, inner_type.lock_descriptor);
 
-	type = get_scaling_info_update_type(u);
-	elevate_update_type(&overall_type, type);
+	inner_type = get_scaling_info_update_type(check_config, u);
+	elevate_update_type(&overall_type, inner_type.update_type, inner_type.lock_descriptor);
 
-	if (u->flip_addr)
+	if (u->flip_addr) {
 		update_flags->bits.addr_update = 1;
+		elevate_update_type(&overall_type, UPDATE_TYPE_FAST, LOCK_DESCRIPTOR_STREAM);
 
-	if (u->in_transfer_func)
+		if (u->flip_addr->address.tmz_surface != u->surface->address.tmz_surface) {
+			update_flags->bits.tmz_changed = 1;
+			elevate_update_type(&overall_type, UPDATE_TYPE_FULL, LOCK_DESCRIPTOR_GLOBAL);
+		}
+	}
+	if (u->in_transfer_func) {
 		update_flags->bits.in_transfer_func_change = 1;
+		elevate_update_type(&overall_type, UPDATE_TYPE_MED, LOCK_DESCRIPTOR_STREAM);
+	}
 
-	if (u->input_csc_color_matrix)
+	if (u->input_csc_color_matrix) {
 		update_flags->bits.input_csc_change = 1;
+		elevate_update_type(&overall_type, UPDATE_TYPE_FAST, LOCK_DESCRIPTOR_STREAM);
+	}
 
-	if (u->coeff_reduction_factor)
+	if (u->coeff_reduction_factor) {
 		update_flags->bits.coeff_reduction_change = 1;
+		elevate_update_type(&overall_type, UPDATE_TYPE_FAST, LOCK_DESCRIPTOR_STREAM);
+	}
 
-	if (u->gamut_remap_matrix)
+	if (u->gamut_remap_matrix) {
 		update_flags->bits.gamut_remap_change = 1;
+		elevate_update_type(&overall_type, UPDATE_TYPE_FAST, LOCK_DESCRIPTOR_STREAM);
+	}
 
-	if (u->gamma) {
-		enum surface_pixel_format format = SURFACE_PIXEL_FORMAT_GRPH_BEGIN;
-
-		if (u->plane_info)
-			format = u->plane_info->format;
-		else if (u->surface)
-			format = u->surface->format;
-
-		if (dce_use_lut(format))
-			update_flags->bits.gamma_change = 1;
+	if (u->blend_tf || (u->gamma && dce_use_lut(u->plane_info ? u->plane_info->format : u->surface->format))) {
+		update_flags->bits.gamma_change = 1;
+		elevate_update_type(&overall_type, UPDATE_TYPE_FAST, LOCK_DESCRIPTOR_STREAM);
 	}
 
-	if (u->lut3d_func || u->func_shaper)
+	if (u->lut3d_func || u->func_shaper) {
 		update_flags->bits.lut_3d = 1;
+		elevate_update_type(&overall_type, UPDATE_TYPE_FAST, LOCK_DESCRIPTOR_STREAM);
+	}
 
 	if (u->hdr_mult.value)
 		if (u->hdr_mult.value != u->surface->hdr_mult.value) {
+			// TODO: Should be fast?
 			update_flags->bits.hdr_mult = 1;
-			elevate_update_type(&overall_type, UPDATE_TYPE_MED);
+			elevate_update_type(&overall_type, UPDATE_TYPE_MED, LOCK_DESCRIPTOR_STREAM);
 		}
 
-	if (update_flags->bits.in_transfer_func_change) {
-		type = UPDATE_TYPE_MED;
-		elevate_update_type(&overall_type, type);
+	if (u->sdr_white_level_nits)
+		if (u->sdr_white_level_nits != u->surface->sdr_white_level_nits) {
+			// TODO: Should be fast?
+			update_flags->bits.sdr_white_level_nits = 1;
+			elevate_update_type(&overall_type, UPDATE_TYPE_FULL, LOCK_DESCRIPTOR_GLOBAL);
+		}
+
+	if (u->cm2_params) {
+		if (u->cm2_params->component_settings.shaper_3dlut_setting != u->surface->mcm_shaper_3dlut_setting
+				|| u->cm2_params->component_settings.lut1d_enable != u->surface->mcm_lut1d_enable
+				|| u->cm2_params->cm2_luts.lut3d_data.lut3d_src != u->surface->mcm_luts.lut3d_data.lut3d_src) {
+			update_flags->bits.mcm_transfer_function_enable_change = 1;
+			elevate_update_type(&overall_type, UPDATE_TYPE_FULL, LOCK_DESCRIPTOR_GLOBAL);
+		}
 	}
 
-	if (update_flags->bits.input_csc_change
-			|| update_flags->bits.coeff_reduction_change
-			|| update_flags->bits.lut_3d
-			|| update_flags->bits.gamma_change
-			|| update_flags->bits.gamut_remap_change) {
-		type = UPDATE_TYPE_FULL;
-		elevate_update_type(&overall_type, type);
+	if (update_flags->bits.lut_3d &&
+			u->surface->mcm_luts.lut3d_data.lut3d_src != DC_CM2_TRANSFER_FUNC_SOURCE_VIDMEM) {
+		elevate_update_type(&overall_type, UPDATE_TYPE_FULL, LOCK_DESCRIPTOR_GLOBAL);
 	}
 
+	if (check_config->enable_legacy_fast_update &&
+			(update_flags->bits.gamma_change ||
+			update_flags->bits.gamut_remap_change ||
+			update_flags->bits.input_csc_change ||
+			update_flags->bits.coeff_reduction_change)) {
+		elevate_update_type(&overall_type, UPDATE_TYPE_FULL, LOCK_DESCRIPTOR_GLOBAL);
+	}
 	return overall_type;
 }
 
-static enum surface_update_type check_update_surfaces_for_stream(
-		struct dc *dc,
-		struct dc_surface_update *updates,
-		int surface_count,
-		struct dc_stream_update *stream_update,
-		const struct dc_stream_status *stream_status)
+/* May need to flip the desktop plane in cases where MPO plane receives a flip but desktop plane doesn't
+ * while both planes are flip_immediate
+ */
+static void force_immediate_gsl_plane_flip(struct dc *dc, struct dc_surface_update *updates, int surface_count)
 {
+	bool has_flip_immediate_plane = false;
 	int i;
-	enum surface_update_type overall_type = UPDATE_TYPE_FAST;
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	if (dc->idle_optimizations_allowed)
-		overall_type = UPDATE_TYPE_FULL;
+	for (i = 0; i < surface_count; i++) {
+		if (updates[i].surface->flip_immediate) {
+			has_flip_immediate_plane = true;
+			break;
+		}
+	}
 
-#endif
-	if (stream_status == NULL || stream_status->plane_count != surface_count)
-		overall_type = UPDATE_TYPE_FULL;
+	if (has_flip_immediate_plane && surface_count > 1) {
+		for (i = 0; i < surface_count; i++) {
+			if (updates[i].surface->flip_immediate)
+				updates[i].surface->update_flags.bits.addr_update = 1;
+		}
+	}
+}
+
+static struct surface_update_descriptor check_update_surfaces_for_stream(
+		const struct dc_check_config *check_config,
+		struct dc_surface_update *updates,
+		int surface_count,
+		struct dc_stream_update *stream_update)
+{
+	struct surface_update_descriptor overall_type = { UPDATE_TYPE_FAST, LOCK_DESCRIPTOR_NONE };
 
 	if (stream_update && stream_update->pending_test_pattern) {
-		overall_type = UPDATE_TYPE_FULL;
+		elevate_update_type(&overall_type, UPDATE_TYPE_FULL, LOCK_DESCRIPTOR_GLOBAL);
+	}
+
+	if (stream_update && stream_update->hw_cursor_req) {
+		elevate_update_type(&overall_type, UPDATE_TYPE_FULL, LOCK_DESCRIPTOR_GLOBAL);
 	}
 
 	/* some stream updates require passive update */
 	if (stream_update) {
+		elevate_update_type(&overall_type, UPDATE_TYPE_FAST, LOCK_DESCRIPTOR_STREAM);
+
 		union stream_update_flags *su_flags = &stream_update->stream->update_flags;
 
 		if ((stream_update->src.height != 0 && stream_update->src.width != 0) ||
@@ -2349,14 +2982,16 @@ static enum surface_update_type check_update_surfaces_for_stream(
 			stream_update->integer_scaling_update)
 			su_flags->bits.scaling = 1;
 
-		if (stream_update->out_transfer_func)
+		if (check_config->enable_legacy_fast_update && stream_update->out_transfer_func)
 			su_flags->bits.out_tf = 1;
 
 		if (stream_update->abm_level)
 			su_flags->bits.abm_level = 1;
 
-		if (stream_update->dpms_off)
+		if (stream_update->dpms_off) {
 			su_flags->bits.dpms_off = 1;
+			elevate_update_type(&overall_type, UPDATE_TYPE_FULL, LOCK_DESCRIPTOR_GLOBAL | LOCK_DESCRIPTOR_LINK);
+		}
 
 		if (stream_update->gamut_remap)
 			su_flags->bits.gamut_remap = 1;
@@ -2367,23 +3002,44 @@ static enum surface_update_type check_update_surfaces_for_stream(
 		if (stream_update->dsc_config)
 			su_flags->bits.dsc_changed = 1;
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 		if (stream_update->mst_bw_update)
 			su_flags->bits.mst_bw = 1;
-#endif
 
-		if (su_flags->raw != 0)
-			overall_type = UPDATE_TYPE_FULL;
+		if (stream_update->stream->freesync_on_desktop &&
+			(stream_update->vrr_infopacket || stream_update->allow_freesync ||
+				stream_update->vrr_active_variable || stream_update->vrr_active_fixed))
+			su_flags->bits.fams_changed = 1;
+
+		if (stream_update->scaler_sharpener_update)
+			su_flags->bits.scaler_sharpener = 1;
+
+		if (stream_update->sharpening_required)
+			su_flags->bits.sharpening_required = 1;
+
+		if (stream_update->output_color_space)
+			su_flags->bits.out_csc = 1;
+
+		// TODO: Make each elevation explicit, as to not override fast stream in crct_timing_adjust
+		if (su_flags->raw)
+			elevate_update_type(&overall_type, UPDATE_TYPE_FULL, LOCK_DESCRIPTOR_GLOBAL);
 
-		if (stream_update->output_csc_transform || stream_update->output_color_space)
+		// Non-global cases
+		if (stream_update->output_csc_transform) {
 			su_flags->bits.out_csc = 1;
+			elevate_update_type(&overall_type, UPDATE_TYPE_FAST, LOCK_DESCRIPTOR_STREAM);
+		}
+
+		if (!check_config->enable_legacy_fast_update && stream_update->out_transfer_func) {
+			su_flags->bits.out_tf = 1;
+			elevate_update_type(&overall_type, UPDATE_TYPE_FAST, LOCK_DESCRIPTOR_STREAM);
+		}
 	}
 
-	for (i = 0 ; i < surface_count; i++) {
-		enum surface_update_type type =
-				det_surface_update(dc, &updates[i]);
+	for (int i = 0 ; i < surface_count; i++) {
+		struct surface_update_descriptor inner_type =
+				det_surface_update(check_config, &updates[i]);
 
-		elevate_update_type(&overall_type, type);
+		elevate_update_type(&overall_type, inner_type.update_type, inner_type.lock_descriptor);
 	}
 
 	return overall_type;
@@ -2394,46 +3050,18 @@ static enum surface_update_type check_update_surfaces_for_stream(
  *
  * See :c:type:`enum surface_update_type <surface_update_type>` for explanation of update types
  */
-enum surface_update_type dc_check_update_surfaces_for_stream(
-		struct dc *dc,
+struct surface_update_descriptor dc_check_update_surfaces_for_stream(
+		const struct dc_check_config *check_config,
 		struct dc_surface_update *updates,
 		int surface_count,
-		struct dc_stream_update *stream_update,
-		const struct dc_stream_status *stream_status)
+		struct dc_stream_update *stream_update)
 {
-	int i;
-	enum surface_update_type type;
-
 	if (stream_update)
 		stream_update->stream->update_flags.raw = 0;
-	for (i = 0; i < surface_count; i++)
+	for (size_t i = 0; i < surface_count; i++)
 		updates[i].surface->update_flags.raw = 0;
 
-	type = check_update_surfaces_for_stream(dc, updates, surface_count, stream_update, stream_status);
-	if (type == UPDATE_TYPE_FULL) {
-		if (stream_update) {
-			uint32_t dsc_changed = stream_update->stream->update_flags.bits.dsc_changed;
-			stream_update->stream->update_flags.raw = 0xFFFFFFFF;
-			stream_update->stream->update_flags.bits.dsc_changed = dsc_changed;
-		}
-		for (i = 0; i < surface_count; i++)
-			updates[i].surface->update_flags.raw = 0xFFFFFFFF;
-	}
-
-	if (type == UPDATE_TYPE_FAST) {
-		// If there's an available clock comparator, we use that.
-		if (dc->clk_mgr->funcs->are_clock_states_equal) {
-			if (!dc->clk_mgr->funcs->are_clock_states_equal(&dc->clk_mgr->clks, &dc->current_state->bw_ctx.bw.dcn.clk))
-				dc->optimized_required = true;
-		// Else we fallback to mem compare.
-		} else if (memcmp(&dc->current_state->bw_ctx.bw.dcn.clk, &dc->clk_mgr->clks, offsetof(struct dc_clocks, prev_p_state_change_support)) != 0) {
-			dc->optimized_required = true;
-		}
-
-		dc->optimized_required |= dc->wm_optimized_required;
-	}
-
-	return type;
+	return check_update_surfaces_for_stream(check_config, updates, surface_count, stream_update);
 }
 
 static struct dc_stream_status *stream_get_status(
@@ -2513,55 +3141,66 @@ static void copy_surface_update_to_plane(
 				srf_update->plane_info->layer_index;
 	}
 
-	if (srf_update->gamma &&
-			(surface->gamma_correction !=
-					srf_update->gamma)) {
-		memcpy(&surface->gamma_correction->entries,
+	if (srf_update->gamma) {
+		memcpy(&surface->gamma_correction.entries,
 			&srf_update->gamma->entries,
 			sizeof(struct dc_gamma_entries));
-		surface->gamma_correction->is_identity =
+		surface->gamma_correction.is_identity =
 			srf_update->gamma->is_identity;
-		surface->gamma_correction->num_entries =
+		surface->gamma_correction.num_entries =
 			srf_update->gamma->num_entries;
-		surface->gamma_correction->type =
+		surface->gamma_correction.type =
 			srf_update->gamma->type;
 	}
 
-	if (srf_update->in_transfer_func &&
-			(surface->in_transfer_func !=
-				srf_update->in_transfer_func)) {
-		surface->in_transfer_func->sdr_ref_white_level =
+	if (srf_update->in_transfer_func) {
+		surface->in_transfer_func.sdr_ref_white_level =
 			srf_update->in_transfer_func->sdr_ref_white_level;
-		surface->in_transfer_func->tf =
+		surface->in_transfer_func.tf =
 			srf_update->in_transfer_func->tf;
-		surface->in_transfer_func->type =
+		surface->in_transfer_func.type =
 			srf_update->in_transfer_func->type;
-		memcpy(&surface->in_transfer_func->tf_pts,
+		memcpy(&surface->in_transfer_func.tf_pts,
 			&srf_update->in_transfer_func->tf_pts,
 			sizeof(struct dc_transfer_func_distributed_points));
 	}
 
-	if (srf_update->func_shaper &&
-			(surface->in_shaper_func !=
-			srf_update->func_shaper))
-		memcpy(surface->in_shaper_func, srf_update->func_shaper,
-		sizeof(*surface->in_shaper_func));
+	if (srf_update->cm2_params) {
+		surface->mcm_shaper_3dlut_setting = srf_update->cm2_params->component_settings.shaper_3dlut_setting;
+		surface->mcm_lut1d_enable = srf_update->cm2_params->component_settings.lut1d_enable;
+		surface->mcm_luts = srf_update->cm2_params->cm2_luts;
+	}
+
+	if (srf_update->func_shaper) {
+		memcpy(&surface->in_shaper_func, srf_update->func_shaper,
+		sizeof(surface->in_shaper_func));
+
+		if (surface->mcm_shaper_3dlut_setting >= DC_CM2_SHAPER_3DLUT_SETTING_ENABLE_SHAPER)
+			surface->mcm_luts.shaper = &surface->in_shaper_func;
+	}
 
-	if (srf_update->lut3d_func &&
-			(surface->lut3d_func !=
-			srf_update->lut3d_func))
-		memcpy(surface->lut3d_func, srf_update->lut3d_func,
-		sizeof(*surface->lut3d_func));
+	if (srf_update->lut3d_func)
+		memcpy(&surface->lut3d_func, srf_update->lut3d_func,
+		sizeof(surface->lut3d_func));
 
 	if (srf_update->hdr_mult.value)
 		surface->hdr_mult =
 				srf_update->hdr_mult;
 
-	if (srf_update->blend_tf &&
-			(surface->blend_tf !=
-			srf_update->blend_tf))
-		memcpy(surface->blend_tf, srf_update->blend_tf,
-		sizeof(*surface->blend_tf));
+	if (srf_update->sdr_white_level_nits)
+		surface->sdr_white_level_nits =
+				srf_update->sdr_white_level_nits;
+
+	if (srf_update->blend_tf) {
+		memcpy(&surface->blend_tf, srf_update->blend_tf,
+		sizeof(surface->blend_tf));
+
+		if (surface->mcm_lut1d_enable)
+			surface->mcm_luts.lut1d_func = &surface->blend_tf;
+	}
+
+	if (srf_update->cm2_params || srf_update->blend_tf)
+		surface->lut_bank_a = !surface->lut_bank_a;
 
 	if (srf_update->input_csc_color_matrix)
 		surface->input_csc_color_matrix =
@@ -2574,6 +3213,14 @@ static void copy_surface_update_to_plane(
 	if (srf_update->gamut_remap_matrix)
 		surface->gamut_remap_matrix =
 			*srf_update->gamut_remap_matrix;
+
+	if (srf_update->cursor_csc_color_matrix)
+		surface->cursor_csc_color_matrix =
+			*srf_update->cursor_csc_color_matrix;
+
+	if (srf_update->bias_and_scale.bias_and_scale_valid)
+			surface->bias_and_scale =
+					srf_update->bias_and_scale;
 }
 
 static void copy_stream_update_to_stream(struct dc *dc,
@@ -2592,14 +3239,13 @@ static void copy_stream_update_to_stream(struct dc *dc,
 	if (update->dst.height && update->dst.width)
 		stream->dst = update->dst;
 
-	if (update->out_transfer_func &&
-	    stream->out_transfer_func != update->out_transfer_func) {
-		stream->out_transfer_func->sdr_ref_white_level =
+	if (update->out_transfer_func) {
+		stream->out_transfer_func.sdr_ref_white_level =
 			update->out_transfer_func->sdr_ref_white_level;
-		stream->out_transfer_func->tf = update->out_transfer_func->tf;
-		stream->out_transfer_func->type =
+		stream->out_transfer_func.tf = update->out_transfer_func->tf;
+		stream->out_transfer_func.type =
 			update->out_transfer_func->type;
-		memcpy(&stream->out_transfer_func->tf_pts,
+		memcpy(&stream->out_transfer_func.tf_pts,
 		       &update->out_transfer_func->tf_pts,
 		       sizeof(struct dc_transfer_func_distributed_points));
 	}
@@ -2610,11 +3256,8 @@ static void copy_stream_update_to_stream(struct dc *dc,
 	if (update->abm_level)
 		stream->abm_level = *update->abm_level;
 
-	if (update->periodic_interrupt0)
-		stream->periodic_interrupt0 = *update->periodic_interrupt0;
-
-	if (update->periodic_interrupt1)
-		stream->periodic_interrupt1 = *update->periodic_interrupt1;
+	if (update->periodic_interrupt)
+		stream->periodic_interrupt = *update->periodic_interrupt;
 
 	if (update->gamut_remap)
 		stream->gamut_remap_matrix = *update->gamut_remap;
@@ -2632,15 +3275,48 @@ static void copy_stream_update_to_stream(struct dc *dc,
 	if (update->vrr_infopacket)
 		stream->vrr_infopacket = *update->vrr_infopacket;
 
+	if (update->hw_cursor_req)
+		stream->hw_cursor_req = *update->hw_cursor_req;
+
+	if (update->allow_freesync)
+		stream->allow_freesync = *update->allow_freesync;
+
+	if (update->vrr_active_variable)
+		stream->vrr_active_variable = *update->vrr_active_variable;
+
+	if (update->vrr_active_fixed)
+		stream->vrr_active_fixed = *update->vrr_active_fixed;
+
+	if (update->crtc_timing_adjust) {
+		if (stream->adjust.v_total_min != update->crtc_timing_adjust->v_total_min ||
+			stream->adjust.v_total_max != update->crtc_timing_adjust->v_total_max ||
+			stream->adjust.timing_adjust_pending)
+			update->crtc_timing_adjust->timing_adjust_pending = true;
+		stream->adjust = *update->crtc_timing_adjust;
+		update->crtc_timing_adjust->timing_adjust_pending = false;
+	}
+
 	if (update->dpms_off)
 		stream->dpms_off = *update->dpms_off;
 
+	if (update->hfvsif_infopacket)
+		stream->hfvsif_infopacket = *update->hfvsif_infopacket;
+
+	if (update->vtem_infopacket)
+		stream->vtem_infopacket = *update->vtem_infopacket;
+
 	if (update->vsc_infopacket)
 		stream->vsc_infopacket = *update->vsc_infopacket;
 
 	if (update->vsp_infopacket)
 		stream->vsp_infopacket = *update->vsp_infopacket;
 
+	if (update->adaptive_sync_infopacket)
+		stream->adaptive_sync_infopacket = *update->adaptive_sync_infopacket;
+
+	if (update->avi_infopacket)
+		stream->avi_infopacket = *update->avi_infopacket;
+
 	if (update->dither_option)
 		stream->dither_option = *update->dither_option;
 
@@ -2663,25 +3339,277 @@ static void copy_stream_update_to_stream(struct dc *dc,
 				       update->dsc_config->num_slices_v != 0);
 
 		/* Use temporarry context for validating new DSC config */
-		struct dc_state *dsc_validate_context = dc_create_state(dc);
+		struct dc_state *dsc_validate_context = dc_state_create_copy(dc->current_state);
 
 		if (dsc_validate_context) {
-			dc_resource_state_copy_construct(dc->current_state, dsc_validate_context);
-
 			stream->timing.dsc_cfg = *update->dsc_config;
 			stream->timing.flags.DSC = enable_dsc;
-			if (!dc->res_pool->funcs->validate_bandwidth(dc, dsc_validate_context, true)) {
+			if (dc->res_pool->funcs->validate_bandwidth(dc, dsc_validate_context,
+				DC_VALIDATE_MODE_ONLY) != DC_OK) {
 				stream->timing.dsc_cfg = old_dsc_cfg;
 				stream->timing.flags.DSC = old_dsc_enabled;
 				update->dsc_config = NULL;
 			}
 
-			dc_release_state(dsc_validate_context);
+			dc_state_release(dsc_validate_context);
 		} else {
 			DC_ERROR("Failed to allocate new validate context for DSC change\n");
 			update->dsc_config = NULL;
 		}
 	}
+	if (update->scaler_sharpener_update)
+		stream->scaler_sharpener_update = *update->scaler_sharpener_update;
+	if (update->sharpening_required)
+		stream->sharpening_required = *update->sharpening_required;
+}
+
+static void backup_planes_and_stream_state(
+		struct dc_scratch_space *scratch,
+		struct dc_stream_state *stream)
+{
+	int i;
+	struct dc_stream_status *status = dc_stream_get_status(stream);
+
+	if (!status)
+		return;
+
+	for (i = 0; i < status->plane_count; i++) {
+		dc_plane_copy_config(&scratch->plane_states[i], status->plane_states[i]);
+	}
+	scratch->stream_state = *stream;
+}
+
+static void restore_planes_and_stream_state(
+		struct dc_scratch_space *scratch,
+		struct dc_stream_state *stream)
+{
+	int i;
+	struct dc_stream_status *status = dc_stream_get_status(stream);
+
+	if (!status)
+		return;
+
+	for (i = 0; i < status->plane_count; i++) {
+		dc_plane_copy_config(status->plane_states[i], &scratch->plane_states[i]);
+	}
+
+	// refcount is persistent
+	struct kref temp_refcount = stream->refcount;
+	*stream = scratch->stream_state;
+	stream->refcount = temp_refcount;
+}
+
+/**
+ * update_seamless_boot_flags() - Helper function for updating seamless boot flags
+ *
+ * @dc: Current DC state
+ * @context: New DC state to be programmed
+ * @surface_count: Number of surfaces that have an updated
+ * @stream: Corresponding stream to be updated in the current flip
+ *
+ * Updating seamless boot flags do not need to be part of the commit sequence. This
+ * helper function will update the seamless boot flags on each flip (if required)
+ * outside of the HW commit sequence (fast or slow).
+ *
+ * Return: void
+ */
+static void update_seamless_boot_flags(struct dc *dc,
+		struct dc_state *context,
+		int surface_count,
+		struct dc_stream_state *stream)
+{
+	if (get_seamless_boot_stream_count(context) > 0 && (surface_count > 0 || stream->dpms_off)) {
+		/* Optimize seamless boot flag keeps clocks and watermarks high until
+		 * first flip. After first flip, optimization is required to lower
+		 * bandwidth. Important to note that it is expected UEFI will
+		 * only light up a single display on POST, therefore we only expect
+		 * one stream with seamless boot flag set.
+		 */
+		if (stream->apply_seamless_boot_optimization) {
+			stream->apply_seamless_boot_optimization = false;
+
+			if (get_seamless_boot_stream_count(context) == 0)
+				dc->optimized_required = true;
+		}
+	}
+}
+
+static bool full_update_required_weak(
+		const struct dc *dc,
+		const struct dc_surface_update *srf_updates,
+		int surface_count,
+		const struct dc_stream_update *stream_update,
+		const struct dc_stream_state *stream);
+
+/**
+ * update_planes_and_stream_state() - The function takes planes and stream
+ * updates as inputs and determines the appropriate update type. If update type
+ * is FULL, the function allocates a new context, populates and validates it.
+ * Otherwise, it updates current dc context. The function will return both
+ * new_context and new_update_type back to the caller. The function also backs
+ * up both current and new contexts into corresponding dc state scratch memory.
+ * TODO: The function does too many things, and even conditionally allocates dc
+ * context memory implicitly. We should consider to break it down.
+ *
+ * @dc: Current DC state
+ * @srf_updates: an array of surface updates
+ * @surface_count: surface update count
+ * @stream: Corresponding stream to be updated
+ * @stream_update: stream update
+ * @new_update_type: [out] determined update type by the function
+ * @new_context: [out] new context allocated and validated if update type is
+ * FULL, reference to current context if update type is less than FULL.
+ *
+ * Return: true if a valid update is populated into new_context, false
+ * otherwise.
+ */
+static bool update_planes_and_stream_state(struct dc *dc,
+		struct dc_surface_update *srf_updates, int surface_count,
+		struct dc_stream_state *stream,
+		struct dc_stream_update *stream_update,
+		enum surface_update_type *new_update_type,
+		struct dc_state **new_context)
+{
+	struct dc_state *context;
+	int i, j;
+	enum surface_update_type update_type;
+	const struct dc_stream_status *stream_status;
+	struct dc_context *dc_ctx = dc->ctx;
+
+	stream_status = dc_stream_get_status(stream);
+
+	if (!stream_status) {
+		if (surface_count) /* Only an error condition if surf_count non-zero*/
+			ASSERT(false);
+
+		return false; /* Cannot commit surface to stream that is not committed */
+	}
+
+	context = dc->current_state;
+	update_type = dc_check_update_surfaces_for_stream(
+			&dc->check_config, srf_updates, surface_count, stream_update).update_type;
+	if (full_update_required_weak(dc, srf_updates, surface_count, stream_update, stream))
+		update_type = UPDATE_TYPE_FULL;
+
+	/* It is possible to receive a flip for one plane while there are multiple flip_immediate planes in the same stream.
+	 * E.g. Desktop and MPO plane are flip_immediate but only the MPO plane received a flip
+	 * Force the other flip_immediate planes to flip so GSL doesn't wait for a flip that won't come.
+	 */
+	force_immediate_gsl_plane_flip(dc, srf_updates, surface_count);
+	if (update_type == UPDATE_TYPE_FULL)
+		backup_planes_and_stream_state(&dc->scratch.current_state, stream);
+
+	/* update current stream with the new updates */
+	copy_stream_update_to_stream(dc, context, stream, stream_update);
+
+	/* do not perform surface update if surface has invalid dimensions
+	 * (all zero) and no scaling_info is provided
+	 */
+	if (surface_count > 0) {
+		for (i = 0; i < surface_count; i++) {
+			if ((srf_updates[i].surface->src_rect.width == 0 ||
+				 srf_updates[i].surface->src_rect.height == 0 ||
+				 srf_updates[i].surface->dst_rect.width == 0 ||
+				 srf_updates[i].surface->dst_rect.height == 0) &&
+				(!srf_updates[i].scaling_info ||
+				  srf_updates[i].scaling_info->src_rect.width == 0 ||
+				  srf_updates[i].scaling_info->src_rect.height == 0 ||
+				  srf_updates[i].scaling_info->dst_rect.width == 0 ||
+				  srf_updates[i].scaling_info->dst_rect.height == 0)) {
+				DC_ERROR("Invalid src/dst rects in surface update!\n");
+				return false;
+			}
+		}
+	}
+
+	if (update_type == UPDATE_TYPE_FULL) {
+		if (stream_update) {
+			uint32_t dsc_changed = stream_update->stream->update_flags.bits.dsc_changed;
+			stream_update->stream->update_flags.raw = 0xFFFFFFFF;
+			stream_update->stream->update_flags.bits.dsc_changed = dsc_changed;
+		}
+		for (i = 0; i < surface_count; i++)
+			srf_updates[i].surface->update_flags.raw = 0xFFFFFFFF;
+	}
+
+	if (update_type >= update_surface_trace_level)
+		update_surface_trace(dc, srf_updates, surface_count);
+
+	for (i = 0; i < surface_count; i++)
+		copy_surface_update_to_plane(srf_updates[i].surface, &srf_updates[i]);
+
+	if (update_type >= UPDATE_TYPE_FULL) {
+		struct dc_plane_state *new_planes[MAX_SURFACES] = {0};
+
+		for (i = 0; i < surface_count; i++)
+			new_planes[i] = srf_updates[i].surface;
+
+		/* initialize scratch memory for building context */
+		context = dc_state_create_copy(dc->current_state);
+		if (context == NULL) {
+			DC_ERROR("Failed to allocate new validate context!\n");
+			return false;
+		}
+
+		/* For each full update, remove all existing phantom pipes first.
+		 * Ensures that we have enough pipes for newly added MPO planes
+		 */
+		dc_state_remove_phantom_streams_and_planes(dc, context);
+		dc_state_release_phantom_streams_and_planes(dc, context);
+
+		/*remove old surfaces from context */
+		if (!dc_state_rem_all_planes_for_stream(dc, stream, context)) {
+
+			BREAK_TO_DEBUGGER();
+			goto fail;
+		}
+
+		/* add surface to context */
+		if (!dc_state_add_all_planes_for_stream(dc, stream, new_planes, surface_count, context)) {
+
+			BREAK_TO_DEBUGGER();
+			goto fail;
+		}
+	}
+
+	/* save update parameters into surface */
+	for (i = 0; i < surface_count; i++) {
+		struct dc_plane_state *surface = srf_updates[i].surface;
+
+		if (update_type != UPDATE_TYPE_MED)
+			continue;
+		if (surface->update_flags.bits.position_change) {
+			for (j = 0; j < dc->res_pool->pipe_count; j++) {
+				struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[j];
+
+				if (pipe_ctx->plane_state != surface)
+					continue;
+
+				resource_build_scaling_params(pipe_ctx);
+			}
+		}
+	}
+
+	if (update_type == UPDATE_TYPE_FULL) {
+		if (dc->res_pool->funcs->validate_bandwidth(dc, context, DC_VALIDATE_MODE_AND_PROGRAMMING) != DC_OK) {
+			BREAK_TO_DEBUGGER();
+			goto fail;
+		}
+	}
+	update_seamless_boot_flags(dc, context, surface_count, stream);
+
+	*new_context = context;
+	*new_update_type = update_type;
+	if (update_type == UPDATE_TYPE_FULL)
+		backup_planes_and_stream_state(&dc->scratch.new_state, stream);
+
+	return true;
+
+fail:
+	dc_state_release(context);
+
+	return false;
+
 }
 
 static void commit_planes_do_stream_update(struct dc *dc,
@@ -2696,22 +3624,26 @@ static void commit_planes_do_stream_update(struct dc *dc,
 	for (j = 0; j < dc->res_pool->pipe_count; j++) {
 		struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[j];
 
-		if (!pipe_ctx->top_pipe &&  !pipe_ctx->prev_odm_pipe && pipe_ctx->stream == stream) {
-
-			if (stream_update->periodic_interrupt0 &&
-					dc->hwss.setup_periodic_interrupt)
-				dc->hwss.setup_periodic_interrupt(dc, pipe_ctx, VLINE0);
+		if (resource_is_pipe_type(pipe_ctx, OTG_MASTER) && pipe_ctx->stream == stream) {
 
-			if (stream_update->periodic_interrupt1 &&
-					dc->hwss.setup_periodic_interrupt)
-				dc->hwss.setup_periodic_interrupt(dc, pipe_ctx, VLINE1);
+			if (stream_update->periodic_interrupt && dc->hwss.setup_periodic_interrupt)
+				dc->hwss.setup_periodic_interrupt(dc, pipe_ctx);
 
 			if ((stream_update->hdr_static_metadata && !stream->use_dynamic_meta) ||
 					stream_update->vrr_infopacket ||
 					stream_update->vsc_infopacket ||
-					stream_update->vsp_infopacket) {
+					stream_update->vsp_infopacket ||
+					stream_update->hfvsif_infopacket ||
+					stream_update->adaptive_sync_infopacket ||
+					stream_update->vtem_infopacket ||
+					stream_update->avi_infopacket) {
 				resource_build_info_frame(pipe_ctx);
 				dc->hwss.update_info_frame(pipe_ctx);
+
+				if (dc_is_dp_signal(pipe_ctx->stream->signal))
+					dc->link_srv->dp_trace_source_sequence(
+							pipe_ctx->stream->link,
+							DPCD_SOURCE_SEQ_AFTER_UPDATE_INFO_FRAME);
 			}
 
 			if (stream_update->hdr_static_metadata &&
@@ -2741,35 +3673,54 @@ static void commit_planes_do_stream_update(struct dc *dc,
 				}
 			}
 
+			if (stream_update->cursor_attributes)
+				program_cursor_attributes(dc, stream);
+
+			if (stream_update->cursor_position)
+				program_cursor_position(dc, stream);
 
 			/* Full fe update*/
 			if (update_type == UPDATE_TYPE_FAST)
 				continue;
 
 			if (stream_update->dsc_config)
-				dp_update_dsc_config(pipe_ctx);
+				dc->link_srv->update_dsc_config(pipe_ctx);
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 			if (stream_update->mst_bw_update) {
 				if (stream_update->mst_bw_update->is_increase)
-					dc_link_increase_mst_payload(pipe_ctx, stream_update->mst_bw_update->mst_stream_bw);
+					dc->link_srv->increase_mst_payload(pipe_ctx,
+							stream_update->mst_bw_update->mst_stream_bw);
 				else
-					dc_link_reduce_mst_payload(pipe_ctx, stream_update->mst_bw_update->mst_stream_bw);
+					dc->link_srv->reduce_mst_payload(pipe_ctx,
+							stream_update->mst_bw_update->mst_stream_bw);
 			}
-#endif
 
 			if (stream_update->pending_test_pattern) {
-				dc_link_dp_set_test_pattern(stream->link,
+				/*
+				 * test pattern params depends on ODM topology
+				 * changes that we could be applying to front
+				 * end. Since at the current stage front end
+				 * changes are not yet applied. We can only
+				 * apply test pattern in hw based on current
+				 * state and populate the final test pattern
+				 * params in new state. If current and new test
+				 * pattern params are different as result of
+				 * different ODM topology being used, it will be
+				 * detected and handle during front end
+				 * programming update.
+				 */
+				dc->link_srv->dp_set_test_pattern(stream->link,
 					stream->test_pattern.type,
 					stream->test_pattern.color_space,
 					stream->test_pattern.p_link_settings,
 					stream->test_pattern.p_custom_pattern,
 					stream->test_pattern.cust_pattern_size);
+				resource_build_test_pattern_params(&context->res_ctx, pipe_ctx);
 			}
 
 			if (stream_update->dpms_off) {
 				if (*stream_update->dpms_off) {
-					core_link_disable_stream(pipe_ctx);
+					dc->link_srv->set_dpms_off(pipe_ctx);
 					/* for dpms, keep acquired resources*/
 					if (pipe_ctx->stream_res.audio && !dc->debug.az_endpoint_mute_only)
 						pipe_ctx->stream_res.audio->funcs->az_disable(pipe_ctx->stream_res.audio);
@@ -2779,9 +3730,15 @@ static void commit_planes_do_stream_update(struct dc *dc,
 				} else {
 					if (get_seamless_boot_stream_count(context) == 0)
 						dc->hwss.prepare_bandwidth(dc, dc->current_state);
-
-					core_link_enable_stream(dc->current_state, pipe_ctx);
+					dc->link_srv->set_dpms_on(dc->current_state, pipe_ctx);
 				}
+			} else if (pipe_ctx->stream->link->wa_flags.blank_stream_on_ocs_change && stream_update->output_color_space
+					&& !stream->dpms_off && dc_is_dp_signal(pipe_ctx->stream->signal)) {
+				/*
+				 * Workaround for firmware issue in some receivers where they don't pick up
+				 * correct output color space unless DP link is disabled/re-enabled
+				 */
+				dc->link_srv->set_dpms_on(dc->current_state, pipe_ctx);
 			}
 
 			if (stream_update->abm_level && pipe_ctx->stream_res.abm) {
@@ -2805,6 +3762,329 @@ static void commit_planes_do_stream_update(struct dc *dc,
 	}
 }
 
+static bool dc_dmub_should_send_dirty_rect_cmd(struct dc *dc, struct dc_stream_state *stream)
+{
+	if ((stream->link->psr_settings.psr_version == DC_PSR_VERSION_SU_1
+			|| stream->link->psr_settings.psr_version == DC_PSR_VERSION_1)
+			&& stream->ctx->dce_version >= DCN_VERSION_3_1)
+		return true;
+
+	if (stream->link->replay_settings.config.replay_supported)
+		return true;
+
+	if (stream->ctx->dce_version >= DCN_VERSION_3_5 && stream->abm_level)
+		return true;
+
+	return false;
+}
+
+void dc_dmub_update_dirty_rect(struct dc *dc,
+			       int surface_count,
+			       struct dc_stream_state *stream,
+			       struct dc_surface_update *srf_updates,
+			       struct dc_state *context)
+{
+	union dmub_rb_cmd cmd;
+	struct dmub_cmd_update_dirty_rect_data *update_dirty_rect;
+	unsigned int i, j;
+	unsigned int panel_inst = 0;
+
+	if (!dc_dmub_should_send_dirty_rect_cmd(dc, stream))
+		return;
+
+	if (!dc_get_edp_link_panel_inst(dc, stream->link, &panel_inst))
+		return;
+
+	memset(&cmd, 0x0, sizeof(cmd));
+	cmd.update_dirty_rect.header.type = DMUB_CMD__UPDATE_DIRTY_RECT;
+	cmd.update_dirty_rect.header.sub_type = 0;
+	cmd.update_dirty_rect.header.payload_bytes =
+		sizeof(cmd.update_dirty_rect) -
+		sizeof(cmd.update_dirty_rect.header);
+	update_dirty_rect = &cmd.update_dirty_rect.update_dirty_rect_data;
+	for (i = 0; i < surface_count; i++) {
+		struct dc_plane_state *plane_state = srf_updates[i].surface;
+		const struct dc_flip_addrs *flip_addr = srf_updates[i].flip_addr;
+
+		if (!srf_updates[i].surface || !flip_addr)
+			continue;
+		/* Do not send in immediate flip mode */
+		if (srf_updates[i].surface->flip_immediate)
+			continue;
+
+		update_dirty_rect->cmd_version = DMUB_CMD_PSR_CONTROL_VERSION_1;
+		update_dirty_rect->dirty_rect_count = flip_addr->dirty_rect_count;
+		memcpy(update_dirty_rect->src_dirty_rects, flip_addr->dirty_rects,
+				sizeof(flip_addr->dirty_rects));
+		for (j = 0; j < dc->res_pool->pipe_count; j++) {
+			struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[j];
+
+			if (pipe_ctx->stream != stream)
+				continue;
+			if (pipe_ctx->plane_state != plane_state)
+				continue;
+
+			update_dirty_rect->panel_inst = panel_inst;
+			update_dirty_rect->pipe_idx = j;
+			dc_wake_and_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_NO_WAIT);
+		}
+	}
+}
+
+static void build_dmub_update_dirty_rect(
+		struct dc *dc,
+		int surface_count,
+		struct dc_stream_state *stream,
+		struct dc_surface_update *srf_updates,
+		struct dc_state *context,
+		struct dc_dmub_cmd dc_dmub_cmd[],
+		unsigned int *dmub_cmd_count)
+{
+	union dmub_rb_cmd cmd;
+	struct dmub_cmd_update_dirty_rect_data *update_dirty_rect;
+	unsigned int i, j;
+	unsigned int panel_inst = 0;
+
+	if (!dc_dmub_should_send_dirty_rect_cmd(dc, stream))
+		return;
+
+	if (!dc_get_edp_link_panel_inst(dc, stream->link, &panel_inst))
+		return;
+
+	memset(&cmd, 0x0, sizeof(cmd));
+	cmd.update_dirty_rect.header.type = DMUB_CMD__UPDATE_DIRTY_RECT;
+	cmd.update_dirty_rect.header.sub_type = 0;
+	cmd.update_dirty_rect.header.payload_bytes =
+		sizeof(cmd.update_dirty_rect) -
+		sizeof(cmd.update_dirty_rect.header);
+	update_dirty_rect = &cmd.update_dirty_rect.update_dirty_rect_data;
+	for (i = 0; i < surface_count; i++) {
+		struct dc_plane_state *plane_state = srf_updates[i].surface;
+		const struct dc_flip_addrs *flip_addr = srf_updates[i].flip_addr;
+
+		if (!srf_updates[i].surface || !flip_addr)
+			continue;
+		/* Do not send in immediate flip mode */
+		if (srf_updates[i].surface->flip_immediate)
+			continue;
+		update_dirty_rect->cmd_version = DMUB_CMD_PSR_CONTROL_VERSION_1;
+		update_dirty_rect->dirty_rect_count = flip_addr->dirty_rect_count;
+		memcpy(update_dirty_rect->src_dirty_rects, flip_addr->dirty_rects,
+				sizeof(flip_addr->dirty_rects));
+		for (j = 0; j < dc->res_pool->pipe_count; j++) {
+			struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[j];
+
+			if (pipe_ctx->stream != stream)
+				continue;
+			if (pipe_ctx->plane_state != plane_state)
+				continue;
+			update_dirty_rect->panel_inst = panel_inst;
+			update_dirty_rect->pipe_idx = j;
+			dc_dmub_cmd[*dmub_cmd_count].dmub_cmd = cmd;
+			dc_dmub_cmd[*dmub_cmd_count].wait_type = DM_DMUB_WAIT_TYPE_NO_WAIT;
+			(*dmub_cmd_count)++;
+		}
+	}
+}
+
+static bool check_address_only_update(union surface_update_flags update_flags)
+{
+	union surface_update_flags addr_only_update_flags;
+	addr_only_update_flags.raw = 0;
+	addr_only_update_flags.bits.addr_update = 1;
+
+	return update_flags.bits.addr_update &&
+			!(update_flags.raw & ~addr_only_update_flags.raw);
+}
+
+/**
+ * build_dmub_cmd_list() - Build an array of DMCUB commands to be sent to DMCUB
+ *
+ * @dc: Current DC state
+ * @srf_updates: Array of surface updates
+ * @surface_count: Number of surfaces that have an updated
+ * @stream: Corresponding stream to be updated in the current flip
+ * @context: New DC state to be programmed
+ *
+ * @dc_dmub_cmd: Array of DMCUB commands to be sent to DMCUB
+ * @dmub_cmd_count: Count indicating the number of DMCUB commands in dc_dmub_cmd array
+ *
+ * This function builds an array of DMCUB commands to be sent to DMCUB. This function is required
+ * to build an array of commands and have them sent while the OTG lock is acquired.
+ *
+ * Return: void
+ */
+static void build_dmub_cmd_list(struct dc *dc,
+		struct dc_surface_update *srf_updates,
+		int surface_count,
+		struct dc_stream_state *stream,
+		struct dc_state *context,
+		struct dc_dmub_cmd dc_dmub_cmd[],
+		unsigned int *dmub_cmd_count)
+{
+	// Initialize cmd count to 0
+	*dmub_cmd_count = 0;
+	build_dmub_update_dirty_rect(dc, surface_count, stream, srf_updates, context, dc_dmub_cmd, dmub_cmd_count);
+}
+
+static void commit_plane_for_stream_offload_fams2_flip(struct dc *dc,
+		struct dc_surface_update *srf_updates,
+		int surface_count,
+		struct dc_stream_state *stream,
+		struct dc_state *context)
+{
+	int i, j;
+
+	/* update dirty rect for PSR */
+	dc_dmub_update_dirty_rect(dc, surface_count, stream,
+			srf_updates, context);
+
+	/* Perform requested Updates */
+	for (i = 0; i < surface_count; i++) {
+		struct dc_plane_state *plane_state = srf_updates[i].surface;
+
+		for (j = 0; j < dc->res_pool->pipe_count; j++) {
+			struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[j];
+
+			if (!should_update_pipe_for_stream(context, pipe_ctx, stream))
+				continue;
+
+			if (!should_update_pipe_for_plane(context, pipe_ctx, plane_state))
+				continue;
+
+			/* update pipe context for plane */
+			if (pipe_ctx->plane_state->update_flags.bits.addr_update)
+				dc->hwss.update_plane_addr(dc, pipe_ctx);
+		}
+	}
+
+	/* Send commands to DMCUB */
+	dc_dmub_srv_fams2_passthrough_flip(dc,
+				context,
+				stream,
+				srf_updates,
+				surface_count);
+}
+
+static void commit_planes_for_stream_fast(struct dc *dc,
+		struct dc_surface_update *srf_updates,
+		int surface_count,
+		struct dc_stream_state *stream,
+		struct dc_stream_update *stream_update,
+		enum surface_update_type update_type,
+		struct dc_state *context)
+{
+	int i, j;
+	struct pipe_ctx *top_pipe_to_program = NULL;
+	struct dc_stream_status *stream_status = NULL;
+	bool should_offload_fams2_flip = false;
+	bool should_lock_all_pipes = (update_type != UPDATE_TYPE_FAST);
+
+	if (should_lock_all_pipes)
+		determine_pipe_unlock_order(dc, context);
+
+	if (dc->debug.fams2_config.bits.enable &&
+			dc->debug.fams2_config.bits.enable_offload_flip &&
+			dc_state_is_fams2_in_use(dc, context)) {
+		/* if not offloading to HWFQ, offload to FAMS2 if needed */
+		should_offload_fams2_flip = true;
+		for (i = 0; i < surface_count; i++) {
+			if (srf_updates[i].surface &&
+					srf_updates[i].surface->update_flags.raw &&
+					!check_address_only_update(srf_updates[i].surface->update_flags)) {
+				/* more than address update, need to acquire FAMS2 lock */
+				should_offload_fams2_flip = false;
+				break;
+			}
+		}
+		if (stream_update) {
+			/* more than address update, need to acquire FAMS2 lock */
+			should_offload_fams2_flip = false;
+		}
+	}
+
+	dc_exit_ips_for_hw_access(dc);
+
+	dc_z10_restore(dc);
+
+	top_pipe_to_program = resource_get_otg_master_for_stream(
+			&context->res_ctx,
+			stream);
+
+	if (!top_pipe_to_program)
+		return;
+
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
+
+		if (pipe->stream && pipe->plane_state) {
+			if (!dc->debug.using_dml2)
+				set_p_state_switch_method(dc, context, pipe);
+
+			if (dc->debug.visual_confirm)
+				dc_update_visual_confirm_color(dc, context, pipe);
+		}
+	}
+
+	for (i = 0; i < surface_count; i++) {
+		struct dc_plane_state *plane_state = srf_updates[i].surface;
+		/*set logical flag for lock/unlock use*/
+		for (j = 0; j < dc->res_pool->pipe_count; j++) {
+			struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[j];
+
+			if (!pipe_ctx->plane_state)
+				continue;
+			if (!should_update_pipe_for_plane(context, pipe_ctx, plane_state))
+				continue;
+
+			pipe_ctx->plane_state->triplebuffer_flips = false;
+			if (update_type == UPDATE_TYPE_FAST &&
+					dc->hwss.program_triplebuffer != NULL &&
+					!pipe_ctx->plane_state->flip_immediate && dc->debug.enable_tri_buf) {
+				/*triple buffer for VUpdate only*/
+				pipe_ctx->plane_state->triplebuffer_flips = true;
+			}
+		}
+	}
+
+	stream_status = dc_state_get_stream_status(context, stream);
+
+	if (should_offload_fams2_flip) {
+		commit_plane_for_stream_offload_fams2_flip(dc,
+				srf_updates,
+				surface_count,
+				stream,
+				context);
+	} else if (stream_status) {
+		build_dmub_cmd_list(dc,
+				srf_updates,
+				surface_count,
+				stream,
+				context,
+				context->dc_dmub_cmd,
+				&(context->dmub_cmd_count));
+		hwss_build_fast_sequence(dc,
+				context->dc_dmub_cmd,
+				context->dmub_cmd_count,
+				context->block_sequence,
+				&(context->block_sequence_steps),
+				top_pipe_to_program,
+				stream_status,
+				context);
+		hwss_execute_sequence(dc,
+				context->block_sequence,
+				context->block_sequence_steps);
+	}
+
+	/* Clear update flags so next flip doesn't have redundant programming
+	 * (if there's no stream update, the update flags are not cleared).
+	 * Surface updates are cleared unconditionally at the beginning of each flip,
+	 * so no need to clear here.
+	 */
+	if (top_pipe_to_program->stream)
+		top_pipe_to_program->stream->update_flags.raw = 0;
+}
+
 static void commit_planes_for_stream(struct dc *dc,
 		struct dc_surface_update *srf_updates,
 		int surface_count,
@@ -2816,49 +4096,75 @@ static void commit_planes_for_stream(struct dc *dc,
 	int i, j;
 	struct pipe_ctx *top_pipe_to_program = NULL;
 	bool should_lock_all_pipes = (update_type != UPDATE_TYPE_FAST);
+	bool subvp_prev_use = false;
+	bool subvp_curr_use = false;
+	uint8_t current_stream_mask = 0;
+
+	if (should_lock_all_pipes)
+		determine_pipe_unlock_order(dc, context);
+	// Once we apply the new subvp context to hardware it won't be in the
+	// dc->current_state anymore, so we have to cache it before we apply
+	// the new SubVP context
+	subvp_prev_use = false;
+	dc_exit_ips_for_hw_access(dc);
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	dc_z10_restore(dc);
-#endif
+	if (update_type == UPDATE_TYPE_FULL && dc->optimized_required)
+		hwss_process_outstanding_hw_updates(dc, dc->current_state);
 
-	if (get_seamless_boot_stream_count(context) > 0 && surface_count > 0) {
-		/* Optimize seamless boot flag keeps clocks and watermarks high until
-		 * first flip. After first flip, optimization is required to lower
-		 * bandwidth. Important to note that it is expected UEFI will
-		 * only light up a single display on POST, therefore we only expect
-		 * one stream with seamless boot flag set.
-		 */
-		if (stream->apply_seamless_boot_optimization) {
-			stream->apply_seamless_boot_optimization = false;
+	if (update_type != UPDATE_TYPE_FAST && dc->res_pool->funcs->prepare_mcache_programming)
+		dc->res_pool->funcs->prepare_mcache_programming(dc, context);
 
-			if (get_seamless_boot_stream_count(context) == 0)
-				dc->optimized_required = true;
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
+
+		if (pipe->stream && pipe->plane_state) {
+			if (!dc->debug.using_dml2)
+				set_p_state_switch_method(dc, context, pipe);
+
+			if (dc->debug.visual_confirm)
+				dc_update_visual_confirm_color(dc, context, pipe);
 		}
 	}
 
 	if (update_type == UPDATE_TYPE_FULL) {
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 		dc_allow_idle_optimizations(dc, false);
 
-#endif
 		if (get_seamless_boot_stream_count(context) == 0)
 			dc->hwss.prepare_bandwidth(dc, context);
 
+		if (dc->hwss.update_dsc_pg)
+			dc->hwss.update_dsc_pg(dc, context, false);
+
 		context_clock_trace(dc, context);
 	}
 
-	for (j = 0; j < dc->res_pool->pipe_count; j++) {
-		struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[j];
+	if (update_type == UPDATE_TYPE_FULL)
+		hwss_wait_for_outstanding_hw_updates(dc, dc->current_state);
 
-		if (!pipe_ctx->top_pipe &&
-			!pipe_ctx->prev_odm_pipe &&
-			pipe_ctx->stream &&
-			pipe_ctx->stream == stream) {
-			top_pipe_to_program = pipe_ctx;
+	top_pipe_to_program = resource_get_otg_master_for_stream(
+				&context->res_ctx,
+				stream);
+	ASSERT(top_pipe_to_program != NULL);
+
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		struct pipe_ctx *old_pipe = &dc->current_state->res_ctx.pipe_ctx[i];
+
+		// Check old context for SubVP
+		subvp_prev_use |= (dc_state_get_pipe_subvp_type(dc->current_state, old_pipe) == SUBVP_PHANTOM);
+		if (subvp_prev_use)
+			break;
+	}
+
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
+
+		if (dc_state_get_pipe_subvp_type(context, pipe) == SUBVP_PHANTOM) {
+			subvp_curr_use = true;
+			break;
 		}
 	}
 
-#ifdef CONFIG_DRM_AMD_DC_DCN
 	if (stream->test_pattern.type != DP_TEST_PATTERN_VIDEO_MODE) {
 		struct pipe_ctx *mpcc_pipe;
 		struct pipe_ctx *odm_pipe;
@@ -2867,11 +4173,11 @@ static void commit_planes_for_stream(struct dc *dc,
 			for (odm_pipe = mpcc_pipe; odm_pipe; odm_pipe = odm_pipe->next_odm_pipe)
 				odm_pipe->ttu_regs.min_ttu_vblank = MAX_TTU;
 	}
-#endif
 
 	if ((update_type != UPDATE_TYPE_FAST) && stream->update_flags.bits.dsc_changed)
-		if (top_pipe_to_program->stream_res.tg->funcs->lock_doublebuffer_enable) {
-			if (should_use_dmub_lock(stream->link)) {
+		if (top_pipe_to_program &&
+			top_pipe_to_program->stream_res.tg->funcs->lock_doublebuffer_enable) {
+			if (should_use_dmub_inbox1_lock(dc, stream->link)) {
 				union dmub_hw_lock_flags hw_locks = { 0 };
 				struct dmub_hw_lock_inst_flags inst_flags = { 0 };
 
@@ -2887,14 +4193,36 @@ static void commit_planes_for_stream(struct dc *dc,
 						top_pipe_to_program->stream_res.tg);
 		}
 
-	if (should_lock_all_pipes && dc->hwss.interdependent_update_lock)
+	if (dc->hwss.wait_for_dcc_meta_propagation) {
+		dc->hwss.wait_for_dcc_meta_propagation(dc, top_pipe_to_program);
+	}
+
+	if (dc->hwseq->funcs.wait_for_pipe_update_if_needed)
+		dc->hwseq->funcs.wait_for_pipe_update_if_needed(dc, top_pipe_to_program, update_type < UPDATE_TYPE_FULL);
+
+	if (should_lock_all_pipes && dc->hwss.interdependent_update_lock) {
+		if (dc->hwss.subvp_pipe_control_lock)
+			dc->hwss.subvp_pipe_control_lock(dc, context, true, should_lock_all_pipes, NULL, subvp_prev_use);
+
+		if (dc->hwss.dmub_hw_control_lock)
+			dc->hwss.dmub_hw_control_lock(dc, context, true);
+
 		dc->hwss.interdependent_update_lock(dc, context, true);
-	else
+	} else {
+		if (dc->hwss.subvp_pipe_control_lock)
+			dc->hwss.subvp_pipe_control_lock(dc, context, true, should_lock_all_pipes, top_pipe_to_program, subvp_prev_use);
+
+		if (dc->hwss.dmub_hw_control_lock)
+			dc->hwss.dmub_hw_control_lock(dc, context, true);
+
 		/* Lock the top pipe while updating plane addrs, since freesync requires
 		 *  plane addr update event triggers to be synchronized.
 		 *  top_pipe_to_program is expected to never be NULL
 		 */
 		dc->hwss.pipe_control_lock(dc, top_pipe_to_program, true);
+	}
+
+	dc_dmub_update_dirty_rect(dc, surface_count, stream, srf_updates, context);
 
 	// Stream updates
 	if (stream_update)
@@ -2910,37 +4238,69 @@ static void commit_planes_for_stream(struct dc *dc,
 		if (dc->hwss.program_front_end_for_ctx)
 			dc->hwss.program_front_end_for_ctx(dc, context);
 
-		if (should_lock_all_pipes && dc->hwss.interdependent_update_lock)
+		if (should_lock_all_pipes && dc->hwss.interdependent_update_lock) {
 			dc->hwss.interdependent_update_lock(dc, context, false);
-		else
+		} else {
 			dc->hwss.pipe_control_lock(dc, top_pipe_to_program, false);
+		}
 		dc->hwss.post_unlock_program_front_end(dc, context);
+
+		if (update_type != UPDATE_TYPE_FAST)
+			if (dc->hwss.commit_subvp_config)
+				dc->hwss.commit_subvp_config(dc, context);
+
+		/* Since phantom pipe programming is moved to post_unlock_program_front_end,
+		 * move the SubVP lock to after the phantom pipes have been setup
+		 */
+		if (dc->hwss.subvp_pipe_control_lock)
+			dc->hwss.subvp_pipe_control_lock(dc, context, false, should_lock_all_pipes,
+							 NULL, subvp_prev_use);
+
+		if (dc->hwss.dmub_hw_control_lock)
+			dc->hwss.dmub_hw_control_lock(dc, context, false);
 		return;
 	}
 
-	if (!IS_DIAG_DC(dc->ctx->dce_environment)) {
-		for (i = 0; i < surface_count; i++) {
-			struct dc_plane_state *plane_state = srf_updates[i].surface;
-			/*set logical flag for lock/unlock use*/
-			for (j = 0; j < dc->res_pool->pipe_count; j++) {
-				struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[j];
-				if (!pipe_ctx->plane_state)
-					continue;
-				if (should_update_pipe_for_plane(context, pipe_ctx, plane_state))
-					continue;
-				pipe_ctx->plane_state->triplebuffer_flips = false;
-				if (update_type == UPDATE_TYPE_FAST &&
+	if (update_type != UPDATE_TYPE_FAST) {
+		for (j = 0; j < dc->res_pool->pipe_count; j++) {
+			struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[j];
+
+			if ((dc->debug.visual_confirm == VISUAL_CONFIRM_SUBVP ||
+				dc->debug.visual_confirm == VISUAL_CONFIRM_MCLK_SWITCH) &&
+				pipe_ctx->stream && pipe_ctx->plane_state) {
+				/* Only update visual confirm for SUBVP and Mclk switching here.
+				 * The bar appears on all pipes, so we need to update the bar on all displays,
+				 * so the information doesn't get stale.
+				 */
+				dc->hwss.update_visual_confirm_color(dc, pipe_ctx,
+						pipe_ctx->plane_res.hubp->inst);
+			}
+		}
+	}
+
+	for (i = 0; i < surface_count; i++) {
+		struct dc_plane_state *plane_state = srf_updates[i].surface;
+
+		/*set logical flag for lock/unlock use*/
+		for (j = 0; j < dc->res_pool->pipe_count; j++) {
+			struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[j];
+			if (!pipe_ctx->plane_state)
+				continue;
+			if (!should_update_pipe_for_plane(context, pipe_ctx, plane_state))
+				continue;
+			pipe_ctx->plane_state->triplebuffer_flips = false;
+			if (update_type == UPDATE_TYPE_FAST &&
 					dc->hwss.program_triplebuffer != NULL &&
 					!pipe_ctx->plane_state->flip_immediate && dc->debug.enable_tri_buf) {
-						/*triple buffer for VUpdate  only*/
-						pipe_ctx->plane_state->triplebuffer_flips = true;
-				}
-			}
-			if (update_type == UPDATE_TYPE_FULL) {
-				/* force vsync flip when reconfiguring pipes to prevent underflow */
-				plane_state->flip_immediate = false;
+				/*triple buffer for VUpdate only*/
+				pipe_ctx->plane_state->triplebuffer_flips = true;
 			}
 		}
+		if (update_type == UPDATE_TYPE_FULL) {
+			/* force vsync flip when reconfiguring pipes to prevent underflow */
+			plane_state->flip_immediate = false;
+			plane_state->triplebuffer_flips = false;
+		}
 	}
 
 	// Update Type FULL, Surface updates
@@ -2959,38 +4319,58 @@ static void commit_planes_for_stream(struct dc *dc,
 			if (update_type == UPDATE_TYPE_FAST)
 				continue;
 
-			ASSERT(!pipe_ctx->plane_state->triplebuffer_flips);
-
-			if (dc->hwss.program_triplebuffer != NULL && dc->debug.enable_tri_buf) {
-				/*turn off triple buffer for full update*/
-				dc->hwss.program_triplebuffer(
-					dc, pipe_ctx, pipe_ctx->plane_state->triplebuffer_flips);
-			}
 			stream_status =
 				stream_get_status(context, pipe_ctx->stream);
 
-			if (dc->hwss.apply_ctx_for_surface)
+			if (dc->hwss.apply_ctx_for_surface && stream_status)
 				dc->hwss.apply_ctx_for_surface(
 					dc, pipe_ctx->stream, stream_status->plane_count, context);
 		}
 	}
+
+	for (j = 0; j < dc->res_pool->pipe_count; j++) {
+		struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[j];
+
+		if (!pipe_ctx->plane_state)
+			continue;
+
+		/* Full fe update*/
+		if (update_type == UPDATE_TYPE_FAST)
+			continue;
+
+		ASSERT(!pipe_ctx->plane_state->triplebuffer_flips);
+		if (dc->hwss.program_triplebuffer != NULL && dc->debug.enable_tri_buf) {
+			/*turn off triple buffer for full update*/
+			dc->hwss.program_triplebuffer(
+				dc, pipe_ctx, pipe_ctx->plane_state->triplebuffer_flips);
+		}
+	}
+
 	if (dc->hwss.program_front_end_for_ctx && update_type != UPDATE_TYPE_FAST) {
 		dc->hwss.program_front_end_for_ctx(dc, context);
-#ifdef CONFIG_DRM_AMD_DC_DCN
+
+		//Pipe busy until some frame and line #
+		if (dc->hwseq->funcs.set_wait_for_update_needed_for_pipe && update_type == UPDATE_TYPE_FULL) {
+			for (j = 0; j < dc->res_pool->pipe_count; j++) {
+				struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[j];
+
+				dc->hwseq->funcs.set_wait_for_update_needed_for_pipe(dc, pipe_ctx);
+			}
+		}
+
 		if (dc->debug.validate_dml_output) {
 			for (i = 0; i < dc->res_pool->pipe_count; i++) {
-				struct pipe_ctx cur_pipe = context->res_ctx.pipe_ctx[i];
-				if (cur_pipe.stream == NULL)
+				struct pipe_ctx *cur_pipe = &context->res_ctx.pipe_ctx[i];
+				if (cur_pipe->stream == NULL)
 					continue;
 
-				cur_pipe.plane_res.hubp->funcs->validate_dml_output(
-						cur_pipe.plane_res.hubp, dc->ctx,
+				cur_pipe->plane_res.hubp->funcs->validate_dml_output(
+						cur_pipe->plane_res.hubp, dc->ctx,
 						&context->res_ctx.pipe_ctx[i].rq_regs,
 						&context->res_ctx.pipe_ctx[i].dlg_regs,
 						&context->res_ctx.pipe_ctx[i].ttu_regs);
 			}
 		}
-#endif
 	}
 
 	// Update Type FAST, Surface updates
@@ -3027,9 +4407,17 @@ static void commit_planes_for_stream(struct dc *dc,
 				if (!should_update_pipe_for_plane(context, pipe_ctx, plane_state))
 					continue;
 
+				if (srf_updates[i].cm2_params &&
+						srf_updates[i].cm2_params->cm2_luts.lut3d_data.lut3d_src ==
+								DC_CM2_TRANSFER_FUNC_SOURCE_VIDMEM &&
+						srf_updates[i].cm2_params->component_settings.shaper_3dlut_setting ==
+								DC_CM2_SHAPER_3DLUT_SETTING_ENABLE_SHAPER_3DLUT &&
+						dc->hwss.trigger_3dlut_dma_load)
+					dc->hwss.trigger_3dlut_dma_load(dc, pipe_ctx);
+
 				/*program triple buffer after lock based on flip type*/
 				if (dc->hwss.program_triplebuffer != NULL && dc->debug.enable_tri_buf) {
-					/*only enable triplebuffer for  fast_update*/
+					/*only enable triplebuffer for fast_update*/
 					dc->hwss.program_triplebuffer(
 						dc, pipe_ctx, pipe_ctx->plane_state->triplebuffer_flips);
 				}
@@ -3037,27 +4425,28 @@ static void commit_planes_for_stream(struct dc *dc,
 					dc->hwss.update_plane_addr(dc, pipe_ctx);
 			}
 		}
-
 	}
 
-	if (should_lock_all_pipes && dc->hwss.interdependent_update_lock)
+	if (should_lock_all_pipes && dc->hwss.interdependent_update_lock) {
 		dc->hwss.interdependent_update_lock(dc, context, false);
-	else
+	} else {
 		dc->hwss.pipe_control_lock(dc, top_pipe_to_program, false);
+	}
 
 	if ((update_type != UPDATE_TYPE_FAST) && stream->update_flags.bits.dsc_changed)
-		if (top_pipe_to_program->stream_res.tg->funcs->lock_doublebuffer_enable) {
+		if (top_pipe_to_program &&
+		    top_pipe_to_program->stream_res.tg->funcs->lock_doublebuffer_enable) {
 			top_pipe_to_program->stream_res.tg->funcs->wait_for_state(
-					top_pipe_to_program->stream_res.tg,
-					CRTC_STATE_VACTIVE);
+				top_pipe_to_program->stream_res.tg,
+				CRTC_STATE_VACTIVE);
 			top_pipe_to_program->stream_res.tg->funcs->wait_for_state(
-					top_pipe_to_program->stream_res.tg,
-					CRTC_STATE_VBLANK);
+				top_pipe_to_program->stream_res.tg,
+				CRTC_STATE_VBLANK);
 			top_pipe_to_program->stream_res.tg->funcs->wait_for_state(
-					top_pipe_to_program->stream_res.tg,
-					CRTC_STATE_VACTIVE);
+				top_pipe_to_program->stream_res.tg,
+				CRTC_STATE_VACTIVE);
 
-			if (stream && should_use_dmub_lock(stream->link)) {
+			if (should_use_dmub_inbox1_lock(dc, stream->link)) {
 				union dmub_hw_lock_flags hw_locks = { 0 };
 				struct dmub_hw_lock_inst_flags inst_flags = { 0 };
 
@@ -3073,9 +4462,47 @@ static void commit_planes_for_stream(struct dc *dc,
 					top_pipe_to_program->stream_res.tg);
 		}
 
+	if (subvp_curr_use) {
+		/* If enabling subvp or transitioning from subvp->subvp, enable the
+		 * phantom streams before we program front end for the phantom pipes.
+		 */
+		if (update_type != UPDATE_TYPE_FAST) {
+			if (dc->hwss.enable_phantom_streams)
+				dc->hwss.enable_phantom_streams(dc, context);
+		}
+	}
+
 	if (update_type != UPDATE_TYPE_FAST)
 		dc->hwss.post_unlock_program_front_end(dc, context);
 
+	if (subvp_prev_use && !subvp_curr_use) {
+		/* If disabling subvp, disable phantom streams after front end
+		 * programming has completed (we turn on phantom OTG in order
+		 * to complete the plane disable for phantom pipes).
+		 */
+
+		if (dc->hwss.disable_phantom_streams)
+			dc->hwss.disable_phantom_streams(dc, context);
+	}
+
+	if (update_type != UPDATE_TYPE_FAST)
+		if (dc->hwss.commit_subvp_config)
+			dc->hwss.commit_subvp_config(dc, context);
+	/* Since phantom pipe programming is moved to post_unlock_program_front_end,
+	 * move the SubVP lock to after the phantom pipes have been setup
+	 */
+	if (should_lock_all_pipes && dc->hwss.interdependent_update_lock) {
+		if (dc->hwss.subvp_pipe_control_lock)
+			dc->hwss.subvp_pipe_control_lock(dc, context, false, should_lock_all_pipes, NULL, subvp_prev_use);
+		if (dc->hwss.dmub_hw_control_lock)
+			dc->hwss.dmub_hw_control_lock(dc, context, false);
+	} else {
+		if (dc->hwss.subvp_pipe_control_lock)
+			dc->hwss.subvp_pipe_control_lock(dc, context, false, should_lock_all_pipes, top_pipe_to_program, subvp_prev_use);
+		if (dc->hwss.dmub_hw_control_lock)
+			dc->hwss.dmub_hw_control_lock(dc, context, false);
+	}
+
 	// Fire manual trigger only when bottom plane is flipped
 	for (j = 0; j < dc->res_pool->pipe_count; j++) {
 		struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[j];
@@ -3089,94 +4516,732 @@ static void commit_planes_for_stream(struct dc *dc,
 				pipe_ctx->plane_state->skip_manual_trigger)
 			continue;
 
+		if (dc->hwss.program_cursor_offload_now)
+			dc->hwss.program_cursor_offload_now(dc, pipe_ctx);
 		if (pipe_ctx->stream_res.tg->funcs->program_manual_trigger)
 			pipe_ctx->stream_res.tg->funcs->program_manual_trigger(pipe_ctx->stream_res.tg);
 	}
+
+	current_stream_mask = get_stream_mask(dc, context);
+	if (current_stream_mask != context->stream_mask) {
+		context->stream_mask = current_stream_mask;
+		dc_dmub_srv_notify_stream_mask(dc->ctx->dmub_srv, current_stream_mask);
+	}
 }
 
-void dc_commit_updates_for_stream(struct dc *dc,
+/**
+ * could_mpcc_tree_change_for_active_pipes - Check if an OPP associated with MPCC might change
+ *
+ * @dc: Used to get the current state status
+ * @stream: Target stream, which we want to remove the attached planes
+ * @srf_updates: Array of surface updates
+ * @surface_count: Number of surface update
+ * @is_plane_addition: [in] Fill out with true if it is a plane addition case
+ *
+ * DCN32x and newer support a feature named Dynamic ODM which can conflict with
+ * the MPO if used simultaneously in some specific configurations (e.g.,
+ * 4k@144). This function checks if the incoming context requires applying a
+ * transition state with unnecessary pipe splitting and ODM disabled to
+ * circumvent our hardware limitations to prevent this edge case. If the OPP
+ * associated with an MPCC might change due to plane additions, this function
+ * returns true.
+ *
+ * Return:
+ * Return true if OPP and MPCC might change, otherwise, return false.
+ */
+static bool could_mpcc_tree_change_for_active_pipes(struct dc *dc,
+		struct dc_stream_state *stream,
 		struct dc_surface_update *srf_updates,
 		int surface_count,
-		struct dc_stream_state *stream,
-		struct dc_stream_update *stream_update,
-		struct dc_state *state)
+		bool *is_plane_addition)
 {
-	const struct dc_stream_status *stream_status;
-	enum surface_update_type update_type;
-	struct dc_state *context;
-	struct dc_context *dc_ctx = dc->ctx;
-	int i, j;
 
-	stream_status = dc_stream_get_status(stream);
-	context = dc->current_state;
+	struct dc_stream_status *cur_stream_status = stream_get_status(dc->current_state, stream);
+	bool force_minimal_pipe_splitting = false;
+	bool subvp_active = false;
+	uint32_t i;
 
-	update_type = dc_check_update_surfaces_for_stream(
-				dc, srf_updates, surface_count, stream_update, stream_status);
+	*is_plane_addition = false;
+
+	if (cur_stream_status &&
+			dc->current_state->stream_count > 0 &&
+			dc->debug.pipe_split_policy != MPC_SPLIT_AVOID) {
+		/* determine if minimal transition is required due to MPC*/
+		if (surface_count > 0) {
+			if (cur_stream_status->plane_count > surface_count) {
+				force_minimal_pipe_splitting = true;
+			} else if (cur_stream_status->plane_count < surface_count) {
+				force_minimal_pipe_splitting = true;
+				*is_plane_addition = true;
+			}
+		}
+	}
 
-	if (update_type >= update_surface_trace_level)
-		update_surface_trace(dc, srf_updates, surface_count);
+	if (cur_stream_status &&
+			dc->current_state->stream_count == 1 &&
+			dc->debug.enable_single_display_2to1_odm_policy) {
+		/* determine if minimal transition is required due to dynamic ODM*/
+		if (surface_count > 0) {
+			if (cur_stream_status->plane_count > 2 && cur_stream_status->plane_count > surface_count) {
+				force_minimal_pipe_splitting = true;
+			} else if (surface_count > 2 && cur_stream_status->plane_count < surface_count) {
+				force_minimal_pipe_splitting = true;
+				*is_plane_addition = true;
+			}
+		}
+	}
 
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		struct pipe_ctx *pipe = &dc->current_state->res_ctx.pipe_ctx[i];
 
-	if (update_type >= UPDATE_TYPE_FULL) {
+		if (dc_state_get_pipe_subvp_type(dc->current_state, pipe) != SUBVP_NONE) {
+			subvp_active = true;
+			break;
+		}
+	}
 
-		/* initialize scratch memory for building context */
-		context = dc_create_state(dc);
-		if (context == NULL) {
-			DC_ERROR("Failed to allocate new validate context!\n");
-			return;
+	/* For SubVP when adding or removing planes we need to add a minimal transition
+	 * (even when disabling all planes). Whenever disabling a phantom pipe, we
+	 * must use the minimal transition path to disable the pipe correctly.
+	 *
+	 * We want to use the minimal transition whenever subvp is active, not only if
+	 * a plane is being added / removed from a subvp stream (MPO plane can be added
+	 * to a DRR pipe of SubVP + DRR config, in which case we still want to run through
+	 * a min transition to disable subvp.
+	 */
+	if (cur_stream_status && subvp_active) {
+		/* determine if minimal transition is required due to SubVP*/
+		if (cur_stream_status->plane_count > surface_count) {
+			force_minimal_pipe_splitting = true;
+		} else if (cur_stream_status->plane_count < surface_count) {
+			force_minimal_pipe_splitting = true;
+			*is_plane_addition = true;
 		}
+	}
 
-		dc_resource_state_copy_construct(state, context);
+	return force_minimal_pipe_splitting;
+}
 
-		for (i = 0; i < dc->res_pool->pipe_count; i++) {
-			struct pipe_ctx *new_pipe = &context->res_ctx.pipe_ctx[i];
-			struct pipe_ctx *old_pipe = &dc->current_state->res_ctx.pipe_ctx[i];
+struct pipe_split_policy_backup {
+	bool dynamic_odm_policy;
+	bool subvp_policy;
+	enum pipe_split_policy mpc_policy;
+	char force_odm[MAX_PIPES];
+};
 
-			if (new_pipe->plane_state && new_pipe->plane_state != old_pipe->plane_state)
-				new_pipe->plane_state->force_full_update = true;
-		}
-	} else if (update_type == UPDATE_TYPE_FAST && dc_ctx->dce_version >= DCE_VERSION_MAX) {
+static void backup_and_set_minimal_pipe_split_policy(struct dc *dc,
+		struct dc_state *context,
+		struct pipe_split_policy_backup *policy)
+{
+	int i;
+
+	if (!dc->config.is_vmin_only_asic) {
+		policy->mpc_policy = dc->debug.pipe_split_policy;
+		dc->debug.pipe_split_policy = MPC_SPLIT_AVOID;
+	}
+	policy->dynamic_odm_policy = dc->debug.enable_single_display_2to1_odm_policy;
+	dc->debug.enable_single_display_2to1_odm_policy = false;
+	policy->subvp_policy = dc->debug.force_disable_subvp;
+	dc->debug.force_disable_subvp = true;
+	for (i = 0; i < context->stream_count; i++) {
+		policy->force_odm[i] = context->streams[i]->debug.force_odm_combine_segments;
+		if (context->streams[i]->debug.allow_transition_for_forced_odm)
+			context->streams[i]->debug.force_odm_combine_segments = 0;
+	}
+}
+
+static void restore_minimal_pipe_split_policy(struct dc *dc,
+		struct dc_state *context,
+		struct pipe_split_policy_backup *policy)
+{
+	uint8_t i;
+
+	if (!dc->config.is_vmin_only_asic)
+		dc->debug.pipe_split_policy = policy->mpc_policy;
+	dc->debug.enable_single_display_2to1_odm_policy =
+			policy->dynamic_odm_policy;
+	dc->debug.force_disable_subvp = policy->subvp_policy;
+	for (i = 0; i < context->stream_count; i++)
+		context->streams[i]->debug.force_odm_combine_segments = policy->force_odm[i];
+}
+
+static void release_minimal_transition_state(struct dc *dc,
+		struct dc_state *minimal_transition_context,
+		struct dc_state *base_context,
+		struct pipe_split_policy_backup *policy)
+{
+	restore_minimal_pipe_split_policy(dc, base_context, policy);
+	dc_state_release(minimal_transition_context);
+}
+
+static void force_vsync_flip_in_minimal_transition_context(struct dc_state *context)
+{
+	uint8_t i;
+	int j;
+	struct dc_stream_status *stream_status;
+
+	for (i = 0; i < context->stream_count; i++) {
+		stream_status = &context->stream_status[i];
+
+		for (j = 0; j < stream_status->plane_count; j++)
+			stream_status->plane_states[j]->flip_immediate = false;
+	}
+}
+
+static struct dc_state *create_minimal_transition_state(struct dc *dc,
+		struct dc_state *base_context, struct pipe_split_policy_backup *policy)
+{
+	struct dc_state *minimal_transition_context = NULL;
+
+	minimal_transition_context = dc_state_create_copy(base_context);
+	if (!minimal_transition_context)
+		return NULL;
+
+	backup_and_set_minimal_pipe_split_policy(dc, base_context, policy);
+	/* commit minimal state */
+	if (dc->res_pool->funcs->validate_bandwidth(dc, minimal_transition_context,
+		DC_VALIDATE_MODE_AND_PROGRAMMING) == DC_OK) {
+		/* prevent underflow and corruption when reconfiguring pipes */
+		force_vsync_flip_in_minimal_transition_context(minimal_transition_context);
+	} else {
 		/*
-		 * Previous frame finished and HW is ready for optimization.
-		 *
-		 * Only relevant for DCN behavior where we can guarantee the optimization
-		 * is safe to apply - retain the legacy behavior for DCE.
+		 * This should never happen, minimal transition state should
+		 * always be validated first before adding pipe split features.
 		 */
-		dc_post_update_surfaces_to_stream(dc);
+		release_minimal_transition_state(dc, minimal_transition_context, base_context, policy);
+		BREAK_TO_DEBUGGER();
+		minimal_transition_context = NULL;
 	}
+	return minimal_transition_context;
+}
 
+static bool is_pipe_topology_transition_seamless_with_intermediate_step(
+		struct dc *dc,
+		struct dc_state *initial_state,
+		struct dc_state *intermediate_state,
+		struct dc_state *final_state)
+{
+	return dc->hwss.is_pipe_topology_transition_seamless(dc, initial_state,
+			intermediate_state) &&
+			dc->hwss.is_pipe_topology_transition_seamless(dc,
+					intermediate_state, final_state);
+}
 
-	for (i = 0; i < surface_count; i++) {
-		struct dc_plane_state *surface = srf_updates[i].surface;
+static void swap_and_release_current_context(struct dc *dc,
+		struct dc_state *new_context, struct dc_stream_state *stream)
+{
 
-		copy_surface_update_to_plane(surface, &srf_updates[i]);
+	int i;
+	struct dc_state *old = dc->current_state;
+	struct pipe_ctx *pipe_ctx;
 
-		if (update_type >= UPDATE_TYPE_MED) {
-			for (j = 0; j < dc->res_pool->pipe_count; j++) {
-				struct pipe_ctx *pipe_ctx =
-					&context->res_ctx.pipe_ctx[j];
+	/* Since memory free requires elevated IRQ, an interrupt
+	 * request is generated by mem free. If this happens
+	 * between freeing and reassigning the context, our vsync
+	 * interrupt will call into dc and cause a memory
+	 * corruption. Hence, we first reassign the context,
+	 * then free the old context.
+	 */
+	dc->current_state = new_context;
+	dc_state_release(old);
 
-				if (pipe_ctx->plane_state != surface)
-					continue;
+	// clear any forced full updates
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		pipe_ctx = &new_context->res_ctx.pipe_ctx[i];
 
-				resource_build_scaling_params(pipe_ctx);
-			}
+		if (pipe_ctx->plane_state && pipe_ctx->stream == stream)
+			pipe_ctx->plane_state->force_full_update = false;
+	}
+}
+
+static int initialize_empty_surface_updates(
+		struct dc_stream_state *stream,
+		struct dc_surface_update *srf_updates)
+{
+	struct dc_stream_status *status = dc_stream_get_status(stream);
+	int i;
+
+	if (!status)
+		return 0;
+
+	for (i = 0; i < status->plane_count; i++)
+		srf_updates[i].surface = status->plane_states[i];
+
+	return status->plane_count;
+}
+
+static bool commit_minimal_transition_based_on_new_context(struct dc *dc,
+		struct dc_state *new_context,
+		struct dc_stream_state *stream,
+		struct dc_surface_update *srf_updates,
+		int surface_count)
+{
+	bool success = false;
+	struct pipe_split_policy_backup policy;
+	struct dc_state *intermediate_context =
+			create_minimal_transition_state(dc, new_context,
+					&policy);
+
+	if (intermediate_context) {
+		if (is_pipe_topology_transition_seamless_with_intermediate_step(
+				dc,
+				dc->current_state,
+				intermediate_context,
+				new_context)) {
+			DC_LOG_DC("commit minimal transition state: base = new state\n");
+			commit_planes_for_stream(dc, srf_updates,
+					surface_count, stream, NULL,
+					UPDATE_TYPE_FULL, intermediate_context);
+			swap_and_release_current_context(
+					dc, intermediate_context, stream);
+			dc_state_retain(dc->current_state);
+			success = true;
 		}
+		release_minimal_transition_state(
+				dc, intermediate_context, new_context, &policy);
 	}
+	return success;
+}
 
-	copy_stream_update_to_stream(dc, context, stream, stream_update);
+static bool commit_minimal_transition_based_on_current_context(struct dc *dc,
+		struct dc_state *new_context, struct dc_stream_state *stream)
+{
+	bool success = false;
+	struct pipe_split_policy_backup policy;
+	struct dc_state *intermediate_context;
+	struct dc_state *old_current_state = dc->current_state;
+	struct dc_surface_update srf_updates[MAX_SURFACES] = {0};
+	int surface_count;
 
-	if (update_type >= UPDATE_TYPE_FULL) {
-		if (!dc->res_pool->funcs->validate_bandwidth(dc, context, false)) {
-			DC_ERROR("Mode validation failed for stream update!\n");
-			dc_release_state(context);
-			return;
+	/*
+	 * Both current and new contexts share the same stream and plane state
+	 * pointers. When new context is validated, stream and planes get
+	 * populated with new updates such as new plane addresses. This makes
+	 * the current context no longer valid because stream and planes are
+	 * modified from the original. We backup current stream and plane states
+	 * into scratch space whenever we are populating new context. So we can
+	 * restore the original values back by calling the restore function now.
+	 * This restores back the original stream and plane states associated
+	 * with the current state.
+	 */
+	restore_planes_and_stream_state(&dc->scratch.current_state, stream);
+	dc_state_retain(old_current_state);
+	intermediate_context = create_minimal_transition_state(dc,
+			old_current_state, &policy);
+
+	if (intermediate_context) {
+		if (is_pipe_topology_transition_seamless_with_intermediate_step(
+				dc,
+				dc->current_state,
+				intermediate_context,
+				new_context)) {
+			DC_LOG_DC("commit minimal transition state: base = current state\n");
+			surface_count = initialize_empty_surface_updates(
+					stream, srf_updates);
+			commit_planes_for_stream(dc, srf_updates,
+					surface_count, stream, NULL,
+					UPDATE_TYPE_FULL, intermediate_context);
+			swap_and_release_current_context(
+					dc, intermediate_context, stream);
+			dc_state_retain(dc->current_state);
+			success = true;
 		}
+		release_minimal_transition_state(dc, intermediate_context,
+				old_current_state, &policy);
 	}
+	dc_state_release(old_current_state);
+	/*
+	 * Restore stream and plane states back to the values associated with
+	 * new context.
+	 */
+	restore_planes_and_stream_state(&dc->scratch.new_state, stream);
+	return success;
+}
 
-	TRACE_DC_PIPE_STATE(pipe_ctx, i, MAX_PIPES);
+/**
+ * commit_minimal_transition_state_in_dc_update - Commit a minimal state based
+ * on current or new context
+ *
+ * @dc: DC structure, used to get the current state
+ * @new_context: New context
+ * @stream: Stream getting the update for the flip
+ * @srf_updates: Surface updates
+ * @surface_count: Number of surfaces
+ *
+ * The function takes in current state and new state and determine a minimal
+ * transition state as the intermediate step which could make the transition
+ * between current and new states seamless. If found, it will commit the minimal
+ * transition state and update current state to this minimal transition state
+ * and return true, if not, it will return false.
+ *
+ * Return:
+ * Return True if the minimal transition succeeded, false otherwise
+ */
+static bool commit_minimal_transition_state_in_dc_update(struct dc *dc,
+		struct dc_state *new_context,
+		struct dc_stream_state *stream,
+		struct dc_surface_update *srf_updates,
+		int surface_count)
+{
+	bool success = commit_minimal_transition_based_on_new_context(
+				dc, new_context, stream, srf_updates,
+				surface_count);
+	if (!success)
+		success = commit_minimal_transition_based_on_current_context(dc,
+				new_context, stream);
+	if (!success)
+		DC_LOG_ERROR("Fail to commit a seamless minimal transition state between current and new states.\nThis pipe topology update is non-seamless!\n");
+	return success;
+}
 
-	commit_planes_for_stream(
+/**
+ * commit_minimal_transition_state - Create a transition pipe split state
+ *
+ * @dc: Used to get the current state status
+ * @transition_base_context: New transition state
+ *
+ * In some specific configurations, such as pipe split on multi-display with
+ * MPO and/or Dynamic ODM, removing a plane may cause unsupported pipe
+ * programming when moving to new planes. To mitigate those types of problems,
+ * this function adds a transition state that minimizes pipe usage before
+ * programming the new configuration. When adding a new plane, the current
+ * state requires the least pipes, so it is applied without splitting. When
+ * removing a plane, the new state requires the least pipes, so it is applied
+ * without splitting.
+ *
+ * Return:
+ * Return false if something is wrong in the transition state.
+ */
+static bool commit_minimal_transition_state(struct dc *dc,
+		struct dc_state *transition_base_context)
+{
+	struct dc_state *transition_context;
+	struct pipe_split_policy_backup policy;
+	enum dc_status ret = DC_ERROR_UNEXPECTED;
+	unsigned int i, j;
+	unsigned int pipe_in_use = 0;
+	bool subvp_in_use = false;
+	bool odm_in_use = false;
+
+	/* check current pipes in use*/
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		struct pipe_ctx *pipe = &transition_base_context->res_ctx.pipe_ctx[i];
+
+		if (pipe->plane_state)
+			pipe_in_use++;
+	}
+
+	/* If SubVP is enabled and we are adding or removing planes from any main subvp
+	 * pipe, we must use the minimal transition.
+	 */
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		struct pipe_ctx *pipe = &dc->current_state->res_ctx.pipe_ctx[i];
+
+		if (pipe->stream && dc_state_get_pipe_subvp_type(dc->current_state, pipe) == SUBVP_PHANTOM) {
+			subvp_in_use = true;
+			break;
+		}
+	}
+
+	/* If ODM is enabled and we are adding or removing planes from any ODM
+	 * pipe, we must use the minimal transition.
+	 */
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		struct pipe_ctx *pipe = &transition_base_context->res_ctx.pipe_ctx[i];
+
+		if (resource_is_pipe_type(pipe, OTG_MASTER)) {
+			odm_in_use = resource_get_odm_slice_count(pipe) > 1;
+			break;
+		}
+	}
+
+	/* When the OS add a new surface if we have been used all of pipes with odm combine
+	 * and mpc split feature, it need use commit_minimal_transition_state to transition safely.
+	 * After OS exit MPO, it will back to use odm and mpc split with all of pipes, we need
+	 * call it again. Otherwise return true to skip.
+	 *
+	 * Reduce the scenarios to use dc_commit_state_no_check in the stage of flip. Especially
+	 * enter/exit MPO when DCN still have enough resources.
+	 */
+	if (pipe_in_use != dc->res_pool->pipe_count && !subvp_in_use && !odm_in_use)
+		return true;
+
+	DC_LOG_DC("%s base = %s state, reason = %s\n", __func__,
+			dc->current_state == transition_base_context ? "current" : "new",
+			subvp_in_use ? "Subvp In Use" :
+			odm_in_use ? "ODM in Use" :
+			dc->debug.pipe_split_policy != MPC_SPLIT_AVOID ? "MPC in Use" :
+			"Unknown");
+
+	dc_state_retain(transition_base_context);
+	transition_context = create_minimal_transition_state(dc,
+			transition_base_context, &policy);
+	if (transition_context) {
+		ret = dc_commit_state_no_check(dc, transition_context);
+		release_minimal_transition_state(dc, transition_context, transition_base_context, &policy);
+	}
+	dc_state_release(transition_base_context);
+
+	if (ret != DC_OK) {
+		/* this should never happen */
+		BREAK_TO_DEBUGGER();
+		return false;
+	}
+
+	/* force full surface update */
+	for (i = 0; i < dc->current_state->stream_count; i++) {
+		for (j = 0; j < dc->current_state->stream_status[i].plane_count; j++) {
+			dc->current_state->stream_status[i].plane_states[j]->update_flags.raw = 0xFFFFFFFF;
+		}
+	}
+
+	return true;
+}
+
+void populate_fast_updates(struct dc_fast_update *fast_update,
+		struct dc_surface_update *srf_updates,
+		int surface_count,
+		struct dc_stream_update *stream_update)
+{
+	int i = 0;
+
+	if (stream_update) {
+		fast_update[0].out_transfer_func = stream_update->out_transfer_func;
+		fast_update[0].output_csc_transform = stream_update->output_csc_transform;
+	} else {
+		fast_update[0].out_transfer_func = NULL;
+		fast_update[0].output_csc_transform = NULL;
+	}
+
+	for (i = 0; i < surface_count; i++) {
+		fast_update[i].flip_addr = srf_updates[i].flip_addr;
+		fast_update[i].gamma = srf_updates[i].gamma;
+		fast_update[i].gamut_remap_matrix = srf_updates[i].gamut_remap_matrix;
+		fast_update[i].input_csc_color_matrix = srf_updates[i].input_csc_color_matrix;
+		fast_update[i].coeff_reduction_factor = srf_updates[i].coeff_reduction_factor;
+		fast_update[i].cursor_csc_color_matrix = srf_updates[i].cursor_csc_color_matrix;
+	}
+}
+
+static bool fast_updates_exist(const struct dc_fast_update *fast_update, int surface_count)
+{
+	int i;
+
+	if (fast_update[0].out_transfer_func ||
+		fast_update[0].output_csc_transform)
+		return true;
+
+	for (i = 0; i < surface_count; i++) {
+		if (fast_update[i].flip_addr ||
+				fast_update[i].gamma ||
+				fast_update[i].gamut_remap_matrix ||
+				fast_update[i].input_csc_color_matrix ||
+				fast_update[i].cursor_csc_color_matrix ||
+				fast_update[i].coeff_reduction_factor)
+			return true;
+	}
+
+	return false;
+}
+
+bool fast_nonaddr_updates_exist(struct dc_fast_update *fast_update, int surface_count)
+{
+	int i;
+
+	if (fast_update[0].out_transfer_func ||
+		fast_update[0].output_csc_transform)
+		return true;
+
+	for (i = 0; i < surface_count; i++) {
+		if (fast_update[i].input_csc_color_matrix ||
+				fast_update[i].gamma ||
+				fast_update[i].gamut_remap_matrix ||
+				fast_update[i].coeff_reduction_factor ||
+				fast_update[i].cursor_csc_color_matrix)
+			return true;
+	}
+
+	return false;
+}
+
+static bool full_update_required_weak(
+		const struct dc *dc,
+		const struct dc_surface_update *srf_updates,
+		int surface_count,
+		const struct dc_stream_update *stream_update,
+		const struct dc_stream_state *stream)
+{
+	const struct dc_state *context = dc->current_state;
+	if (srf_updates)
+		for (int i = 0; i < surface_count; i++)
+			if (!is_surface_in_context(context, srf_updates[i].surface))
+				return true;
+
+	if (stream) {
+		const struct dc_stream_status *stream_status = dc_stream_get_status_const(stream);
+		if (stream_status == NULL || stream_status->plane_count != surface_count)
+			return true;
+	}
+	if (dc->idle_optimizations_allowed)
+		return true;
+
+	if (dc_can_clear_cursor_limit(dc))
+		return true;
+
+	return false;
+}
+
+static bool full_update_required(
+		const struct dc *dc,
+		const struct dc_surface_update *srf_updates,
+		int surface_count,
+		const struct dc_stream_update *stream_update,
+		const struct dc_stream_state *stream)
+{
+	if (full_update_required_weak(dc, srf_updates, surface_count, stream_update, stream))
+		return true;
+
+	for (int i = 0; i < surface_count; i++) {
+		if (srf_updates &&
+				(srf_updates[i].plane_info ||
+				srf_updates[i].scaling_info ||
+				(srf_updates[i].hdr_mult.value &&
+				srf_updates[i].hdr_mult.value != srf_updates->surface->hdr_mult.value) ||
+				(srf_updates[i].sdr_white_level_nits &&
+				srf_updates[i].sdr_white_level_nits != srf_updates->surface->sdr_white_level_nits) ||
+				srf_updates[i].in_transfer_func ||
+				srf_updates[i].func_shaper ||
+				srf_updates[i].lut3d_func ||
+				srf_updates[i].surface->force_full_update ||
+				(srf_updates[i].flip_addr &&
+				srf_updates[i].flip_addr->address.tmz_surface != srf_updates[i].surface->address.tmz_surface) ||
+				(srf_updates[i].cm2_params &&
+				 (srf_updates[i].cm2_params->component_settings.shaper_3dlut_setting != srf_updates[i].surface->mcm_shaper_3dlut_setting ||
+				  srf_updates[i].cm2_params->component_settings.lut1d_enable != srf_updates[i].surface->mcm_lut1d_enable))))
+			return true;
+	}
+
+	if (stream_update &&
+			(((stream_update->src.height != 0 && stream_update->src.width != 0) ||
+			(stream_update->dst.height != 0 && stream_update->dst.width != 0) ||
+			stream_update->integer_scaling_update) ||
+			stream_update->hdr_static_metadata ||
+			stream_update->abm_level ||
+			stream_update->periodic_interrupt ||
+			stream_update->vrr_infopacket ||
+			stream_update->vsc_infopacket ||
+			stream_update->vsp_infopacket ||
+			stream_update->hfvsif_infopacket ||
+			stream_update->vtem_infopacket ||
+			stream_update->adaptive_sync_infopacket ||
+			stream_update->avi_infopacket ||
+			stream_update->dpms_off ||
+			stream_update->allow_freesync ||
+			stream_update->vrr_active_variable ||
+			stream_update->vrr_active_fixed ||
+			stream_update->gamut_remap ||
+			stream_update->output_color_space ||
+			stream_update->dither_option ||
+			stream_update->wb_update ||
+			stream_update->dsc_config ||
+			stream_update->mst_bw_update ||
+			stream_update->func_shaper ||
+			stream_update->lut3d_func ||
+			stream_update->pending_test_pattern ||
+			stream_update->crtc_timing_adjust ||
+			stream_update->scaler_sharpener_update ||
+			stream_update->hw_cursor_req))
+		return true;
+
+	return false;
+}
+
+static bool fast_update_only(
+		const struct dc *dc,
+		const struct dc_fast_update *fast_update,
+		const struct dc_surface_update *srf_updates,
+		int surface_count,
+		const struct dc_stream_update *stream_update,
+		const struct dc_stream_state *stream)
+{
+	return fast_updates_exist(fast_update, surface_count)
+			&& !full_update_required(dc, srf_updates, surface_count, stream_update, stream);
+}
+
+static bool update_planes_and_stream_v2(struct dc *dc,
+		struct dc_surface_update *srf_updates, int surface_count,
+		struct dc_stream_state *stream,
+		struct dc_stream_update *stream_update)
+{
+	struct dc_state *context;
+	enum surface_update_type update_type;
+	struct dc_fast_update fast_update[MAX_SURFACES] = {0};
+
+	/* In cases where MPO and split or ODM are used transitions can
+	 * cause underflow. Apply stream configuration with minimal pipe
+	 * split first to avoid unsupported transitions for active pipes.
+	 */
+	bool force_minimal_pipe_splitting = 0;
+	bool is_plane_addition = 0;
+	bool is_fast_update_only;
+
+	populate_fast_updates(fast_update, srf_updates, surface_count, stream_update);
+	is_fast_update_only = fast_update_only(dc, fast_update, srf_updates,
+			surface_count, stream_update, stream);
+	force_minimal_pipe_splitting = could_mpcc_tree_change_for_active_pipes(
+			dc,
+			stream,
+			srf_updates,
+			surface_count,
+			&is_plane_addition);
+
+	/* on plane addition, minimal state is the current one */
+	if (force_minimal_pipe_splitting && is_plane_addition &&
+		!commit_minimal_transition_state(dc, dc->current_state))
+		return false;
+
+	if (!update_planes_and_stream_state(
+			dc,
+			srf_updates,
+			surface_count,
+			stream,
+			stream_update,
+			&update_type,
+			&context))
+		return false;
+
+	/* on plane removal, minimal state is the new one */
+	if (force_minimal_pipe_splitting && !is_plane_addition) {
+		if (!commit_minimal_transition_state(dc, context)) {
+			dc_state_release(context);
+			return false;
+		}
+		update_type = UPDATE_TYPE_FULL;
+	}
+
+	if (dc->hwss.is_pipe_topology_transition_seamless &&
+			!dc->hwss.is_pipe_topology_transition_seamless(
+					dc, dc->current_state, context))
+		commit_minimal_transition_state_in_dc_update(dc, context, stream,
+				srf_updates, surface_count);
+
+	if (is_fast_update_only && !dc->check_config.enable_legacy_fast_update) {
+		commit_planes_for_stream_fast(dc,
+				srf_updates,
+				surface_count,
+				stream,
+				stream_update,
+				update_type,
+				context);
+	} else {
+		if (!stream_update &&
+				dc->hwss.is_pipe_topology_transition_seamless &&
+				!dc->hwss.is_pipe_topology_transition_seamless(
+						dc, dc->current_state, context)) {
+			DC_LOG_ERROR("performing non-seamless pipe topology transition with surface only update!\n");
+			BREAK_TO_DEBUGGER();
+		}
+		commit_planes_for_stream(
 				dc,
 				srf_updates,
 				surface_count,
@@ -3184,30 +5249,190 @@ void dc_commit_updates_for_stream(struct dc *dc,
 				stream_update,
 				update_type,
 				context);
-	/*update current_State*/
-	if (dc->current_state != context) {
+	}
+	if (dc->current_state != context)
+		swap_and_release_current_context(dc, context, stream);
+	return true;
+}
+
+static void commit_planes_and_stream_update_on_current_context(struct dc *dc,
+		struct dc_surface_update *srf_updates, int surface_count,
+		struct dc_stream_state *stream,
+		struct dc_stream_update *stream_update,
+		enum surface_update_type update_type)
+{
+	struct dc_fast_update fast_update[MAX_SURFACES] = {0};
+
+	ASSERT(update_type < UPDATE_TYPE_FULL);
+	populate_fast_updates(fast_update, srf_updates, surface_count,
+			stream_update);
+	if (fast_update_only(dc, fast_update, srf_updates, surface_count,
+			stream_update, stream) &&
+			!dc->check_config.enable_legacy_fast_update)
+		commit_planes_for_stream_fast(dc,
+				srf_updates,
+				surface_count,
+				stream,
+				stream_update,
+				update_type,
+				dc->current_state);
+	else
+		commit_planes_for_stream(
+				dc,
+				srf_updates,
+				surface_count,
+				stream,
+				stream_update,
+				update_type,
+				dc->current_state);
+}
+
+static void commit_planes_and_stream_update_with_new_context(struct dc *dc,
+		struct dc_surface_update *srf_updates, int surface_count,
+		struct dc_stream_state *stream,
+		struct dc_stream_update *stream_update,
+		enum surface_update_type update_type,
+		struct dc_state *new_context)
+{
+	ASSERT(update_type >= UPDATE_TYPE_FULL);
+	if (!dc->hwss.is_pipe_topology_transition_seamless(dc,
+			dc->current_state, new_context))
+		/*
+		 * It is required by the feature design that all pipe topologies
+		 * using extra free pipes for power saving purposes such as
+		 * dynamic ODM or SubVp shall only be enabled when it can be
+		 * transitioned seamlessly to AND from its minimal transition
+		 * state. A minimal transition state is defined as the same dc
+		 * state but with all power saving features disabled. So it uses
+		 * the minimum pipe topology. When we can't seamlessly
+		 * transition from state A to state B, we will insert the
+		 * minimal transition state A' or B' in between so seamless
+		 * transition between A and B can be made possible.
+		 */
+		commit_minimal_transition_state_in_dc_update(dc, new_context,
+				stream, srf_updates, surface_count);
 
-		struct dc_state *old = dc->current_state;
+	commit_planes_for_stream(
+			dc,
+			srf_updates,
+			surface_count,
+			stream,
+			stream_update,
+			update_type,
+			new_context);
+}
 
-		dc->current_state = context;
-		dc_release_state(old);
+static bool update_planes_and_stream_v3(struct dc *dc,
+		struct dc_surface_update *srf_updates, int surface_count,
+		struct dc_stream_state *stream,
+		struct dc_stream_update *stream_update)
+{
+	struct dc_state *new_context;
+	enum surface_update_type update_type;
 
-		for (i = 0; i < dc->res_pool->pipe_count; i++) {
-			struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i];
+	/*
+	 * When this function returns true and new_context is not equal to
+	 * current state, the function allocates and validates a new dc state
+	 * and assigns it to new_context. The function expects that the caller
+	 * is responsible to free this memory when new_context is no longer
+	 * used. We swap current with new context and free current instead. So
+	 * new_context's memory will live until the next full update after it is
+	 * replaced by a newer context. Refer to the use of
+	 * swap_and_free_current_context below.
+	 */
+	if (!update_planes_and_stream_state(dc, srf_updates, surface_count,
+				stream, stream_update, &update_type,
+				&new_context))
+		return false;
 
-			if (pipe_ctx->plane_state && pipe_ctx->stream == stream)
-				pipe_ctx->plane_state->force_full_update = false;
-		}
+	if (new_context == dc->current_state) {
+		commit_planes_and_stream_update_on_current_context(dc,
+				srf_updates, surface_count, stream,
+				stream_update, update_type);
+	} else {
+		commit_planes_and_stream_update_with_new_context(dc,
+				srf_updates, surface_count, stream,
+				stream_update, update_type, new_context);
+		swap_and_release_current_context(dc, new_context, stream);
 	}
 
-	/* Legacy optimization path for DCE. */
-	if (update_type >= UPDATE_TYPE_FULL && dc_ctx->dce_version < DCE_VERSION_MAX) {
-		dc_post_update_surfaces_to_stream(dc);
-		TRACE_DCE_CLOCK_STATE(&context->bw_ctx.bw.dce);
-	}
+	return true;
+}
 
-	return;
+static void clear_update_flags(struct dc_surface_update *srf_updates,
+	int surface_count, struct dc_stream_state *stream)
+{
+	int i;
 
+	if (stream)
+		stream->update_flags.raw = 0;
+
+	for (i = 0; i < surface_count; i++)
+		if (srf_updates[i].surface)
+			srf_updates[i].surface->update_flags.raw = 0;
+}
+
+bool dc_update_planes_and_stream(struct dc *dc,
+		struct dc_surface_update *srf_updates, int surface_count,
+		struct dc_stream_state *stream,
+		struct dc_stream_update *stream_update)
+{
+	bool ret = false;
+
+	dc_exit_ips_for_hw_access(dc);
+	/*
+	 * update planes and stream version 3 separates FULL and FAST updates
+	 * to their own sequences. It aims to clean up frequent checks for
+	 * update type resulting unnecessary branching in logic flow. It also
+	 * adds a new commit minimal transition sequence, which detects the need
+	 * for minimal transition based on the actual comparison of current and
+	 * new states instead of "predicting" it based on per feature software
+	 * policy.i.e could_mpcc_tree_change_for_active_pipes. The new commit
+	 * minimal transition sequence is made universal to any power saving
+	 * features that would use extra free pipes such as Dynamic ODM/MPC
+	 * Combine, MPO or SubVp. Therefore there is no longer a need to
+	 * specially handle compatibility problems with transitions among those
+	 * features as they are now transparent to the new sequence.
+	 */
+	if (dc->ctx->dce_version >= DCN_VERSION_4_01 || dc->ctx->dce_version == DCN_VERSION_3_2 ||
+			dc->ctx->dce_version == DCN_VERSION_3_21)
+		ret = update_planes_and_stream_v3(dc, srf_updates,
+				surface_count, stream, stream_update);
+	else
+		ret = update_planes_and_stream_v2(dc, srf_updates,
+			surface_count, stream, stream_update);
+	if (ret && (dc->ctx->dce_version >= DCN_VERSION_3_2 ||
+		dc->ctx->dce_version == DCN_VERSION_3_01))
+		clear_update_flags(srf_updates, surface_count, stream);
+
+	return ret;
+}
+
+void dc_commit_updates_for_stream(struct dc *dc,
+		struct dc_surface_update *srf_updates,
+		int surface_count,
+		struct dc_stream_state *stream,
+		struct dc_stream_update *stream_update,
+		struct dc_state *state)
+{
+	bool ret = false;
+
+	dc_exit_ips_for_hw_access(dc);
+	/* TODO: Since change commit sequence can have a huge impact,
+	 * we decided to only enable it for DCN3x. However, as soon as
+	 * we get more confident about this change we'll need to enable
+	 * the new sequence for all ASICs.
+	 */
+	if (dc->ctx->dce_version >= DCN_VERSION_4_01) {
+		ret = update_planes_and_stream_v3(dc, srf_updates, surface_count,
+				stream, stream_update);
+	} else {
+		ret = update_planes_and_stream_v2(dc, srf_updates, surface_count,
+				stream, stream_update);
+	}
+
+	if (ret && dc->ctx->dce_version >= DCN_VERSION_3_2)
+		clear_update_flags(srf_updates, surface_count, stream);
 }
 
 uint8_t dc_get_current_stream_count(struct dc *dc)
@@ -3222,19 +5447,6 @@ struct dc_stream_state *dc_get_stream_at_index(struct dc *dc, uint8_t i)
 	return NULL;
 }
 
-struct dc_stream_state *dc_stream_find_from_link(const struct dc_link *link)
-{
-	uint8_t i;
-	struct dc_context *ctx = link->ctx;
-
-	for (i = 0; i < ctx->dc->current_state->stream_count; i++) {
-		if (ctx->dc->current_state->streams[i]->link == link)
-			return ctx->dc->current_state->streams[i];
-	}
-
-	return NULL;
-}
-
 enum dc_irq_source dc_interrupt_to_irq_source(
 		struct dc *dc,
 		uint32_t src_id,
@@ -3263,29 +5475,27 @@ void dc_interrupt_ack(struct dc *dc, enum dc_irq_source src)
 void dc_power_down_on_boot(struct dc *dc)
 {
 	if (dc->ctx->dce_environment != DCE_ENV_VIRTUAL_HW &&
-			dc->hwss.power_down_on_boot)
+	    dc->hwss.power_down_on_boot) {
+		if (dc->caps.ips_support)
+			dc_exit_ips_for_hw_access(dc);
 		dc->hwss.power_down_on_boot(dc);
+	}
 }
 
-void dc_set_power_state(
-	struct dc *dc,
-	enum dc_acpi_cm_power_state power_state)
+void dc_set_power_state(struct dc *dc, enum dc_acpi_cm_power_state power_state)
 {
-	struct kref refcount;
-	struct display_mode_lib *dml;
-
 	if (!dc->current_state)
 		return;
 
 	switch (power_state) {
 	case DC_ACPI_CM_POWER_STATE_D0:
-		dc_resource_state_construct(dc, dc->current_state);
+		dc_state_construct(dc, dc->current_state);
+
+		dc_exit_ips_for_hw_access(dc);
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 		dc_z10_restore(dc);
-#endif
-		if (dc->ctx->dmub_srv)
-			dc_dmub_srv_wait_phy_init(dc->ctx->dmub_srv);
+
+		dc_dmub_srv_notify_fw_dc_power_state(dc->ctx->dmub_srv, power_state);
 
 		dc->hwss.init_hw(dc);
 
@@ -3293,34 +5503,21 @@ void dc_set_power_state(
 			dc->vm_pa_config.valid) {
 			dc->hwss.init_sys_ctx(dc->hwseq, dc, &dc->vm_pa_config);
 		}
+		break;
+	case DC_ACPI_CM_POWER_STATE_D3:
+		if (dc->caps.ips_support)
+			dc_dmub_srv_notify_fw_dc_power_state(dc->ctx->dmub_srv, DC_ACPI_CM_POWER_STATE_D3);
 
+		if (dc->caps.ips_v2_support) {
+			if (dc->clk_mgr->funcs->set_low_power_state)
+				dc->clk_mgr->funcs->set_low_power_state(dc->clk_mgr);
+		}
 		break;
 	default:
 		ASSERT(dc->current_state->stream_count == 0);
-		/* Zero out the current context so that on resume we start with
-		 * clean state, and dc hw programming optimizations will not
-		 * cause any trouble.
-		 */
-		dml = kzalloc(sizeof(struct display_mode_lib),
-				GFP_KERNEL);
-
-		ASSERT(dml);
-		if (!dml)
-			return;
+		dc_dmub_srv_notify_fw_dc_power_state(dc->ctx->dmub_srv, power_state);
 
-		/* Preserve refcount */
-		refcount = dc->current_state->refcount;
-		/* Preserve display mode lib */
-		memcpy(dml, &dc->current_state->bw_ctx.dml, sizeof(struct display_mode_lib));
-
-		dc_resource_state_destruct(dc->current_state);
-		memset(dc->current_state, 0,
-				sizeof(*dc->current_state));
-
-		dc->current_state->refcount = refcount;
-		dc->current_state->bw_ctx.dml = *dml;
-
-		kfree(dml);
+		dc_state_destruct(dc->current_state);
 
 		break;
 	}
@@ -3331,7 +5528,7 @@ void dc_resume(struct dc *dc)
 	uint32_t i;
 
 	for (i = 0; i < dc->link_count; i++)
-		core_link_resume(dc->links[i]);
+		dc->link_srv->resume(dc->links[i]);
 }
 
 bool dc_is_dmcu_initialized(struct dc *dc)
@@ -3343,153 +5540,6 @@ bool dc_is_dmcu_initialized(struct dc *dc)
 	return false;
 }
 
-bool dc_submit_i2c(
-		struct dc *dc,
-		uint32_t link_index,
-		struct i2c_command *cmd)
-{
-
-	struct dc_link *link = dc->links[link_index];
-	struct ddc_service *ddc = link->ddc;
-	return dce_i2c_submit_command(
-		dc->res_pool,
-		ddc->ddc_pin,
-		cmd);
-}
-
-bool dc_submit_i2c_oem(
-		struct dc *dc,
-		struct i2c_command *cmd)
-{
-	struct ddc_service *ddc = dc->res_pool->oem_device;
-	return dce_i2c_submit_command(
-		dc->res_pool,
-		ddc->ddc_pin,
-		cmd);
-}
-
-static bool link_add_remote_sink_helper(struct dc_link *dc_link, struct dc_sink *sink)
-{
-	if (dc_link->sink_count >= MAX_SINKS_PER_LINK) {
-		BREAK_TO_DEBUGGER();
-		return false;
-	}
-
-	dc_sink_retain(sink);
-
-	dc_link->remote_sinks[dc_link->sink_count] = sink;
-	dc_link->sink_count++;
-
-	return true;
-}
-
-/*
- * dc_link_add_remote_sink() - Create a sink and attach it to an existing link
- *
- * EDID length is in bytes
- */
-struct dc_sink *dc_link_add_remote_sink(
-		struct dc_link *link,
-		const uint8_t *edid,
-		int len,
-		struct dc_sink_init_data *init_data)
-{
-	struct dc_sink *dc_sink;
-	enum dc_edid_status edid_status;
-
-	if (len > DC_MAX_EDID_BUFFER_SIZE) {
-		dm_error("Max EDID buffer size breached!\n");
-		return NULL;
-	}
-
-	if (!init_data) {
-		BREAK_TO_DEBUGGER();
-		return NULL;
-	}
-
-	if (!init_data->link) {
-		BREAK_TO_DEBUGGER();
-		return NULL;
-	}
-
-	dc_sink = dc_sink_create(init_data);
-
-	if (!dc_sink)
-		return NULL;
-
-	memmove(dc_sink->dc_edid.raw_edid, edid, len);
-	dc_sink->dc_edid.length = len;
-
-	if (!link_add_remote_sink_helper(
-			link,
-			dc_sink))
-		goto fail_add_sink;
-
-	edid_status = dm_helpers_parse_edid_caps(
-			link->ctx,
-			&dc_sink->dc_edid,
-			&dc_sink->edid_caps);
-
-	/*
-	 * Treat device as no EDID device if EDID
-	 * parsing fails
-	 */
-	if (edid_status != EDID_OK) {
-		dc_sink->dc_edid.length = 0;
-		dm_error("Bad EDID, status%d!\n", edid_status);
-	}
-
-	return dc_sink;
-
-fail_add_sink:
-	dc_sink_release(dc_sink);
-	return NULL;
-}
-
-/*
- * dc_link_remove_remote_sink() - Remove a remote sink from a dc_link
- *
- * Note that this just removes the struct dc_sink - it doesn't
- * program hardware or alter other members of dc_link
- */
-void dc_link_remove_remote_sink(struct dc_link *link, struct dc_sink *sink)
-{
-	int i;
-
-	if (!link->sink_count) {
-		BREAK_TO_DEBUGGER();
-		return;
-	}
-
-	for (i = 0; i < link->sink_count; i++) {
-		if (link->remote_sinks[i] == sink) {
-			dc_sink_release(sink);
-			link->remote_sinks[i] = NULL;
-
-			/* shrink array to remove empty place */
-			while (i < link->sink_count - 1) {
-				link->remote_sinks[i] = link->remote_sinks[i+1];
-				i++;
-			}
-			link->remote_sinks[i] = NULL;
-			link->sink_count--;
-			return;
-		}
-	}
-}
-
-void get_clock_requirements_for_state(struct dc_state *state, struct AsicStateEx *info)
-{
-	info->displayClock				= (unsigned int)state->bw_ctx.bw.dcn.clk.dispclk_khz;
-	info->engineClock				= (unsigned int)state->bw_ctx.bw.dcn.clk.dcfclk_khz;
-	info->memoryClock				= (unsigned int)state->bw_ctx.bw.dcn.clk.dramclk_khz;
-	info->maxSupportedDppClock		= (unsigned int)state->bw_ctx.bw.dcn.clk.max_supported_dppclk_khz;
-	info->dppClock					= (unsigned int)state->bw_ctx.bw.dcn.clk.dppclk_khz;
-	info->socClock					= (unsigned int)state->bw_ctx.bw.dcn.clk.socclk_khz;
-	info->dcfClockDeepSleep			= (unsigned int)state->bw_ctx.bw.dcn.clk.dcfclk_deep_sleep_khz;
-	info->fClock					= (unsigned int)state->bw_ctx.bw.dcn.clk.fclk_khz;
-	info->phyClock					= (unsigned int)state->bw_ctx.bw.dcn.clk.phyclk_khz;
-}
 enum dc_status dc_set_clock(struct dc *dc, enum dc_clock_type clock_type, uint32_t clk_khz, uint32_t stepping)
 {
 	if (dc->hwss.set_clock)
@@ -3532,11 +5582,65 @@ bool dc_set_psr_allow_active(struct dc *dc, bool enable)
 	return true;
 }
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
+/* enable/disable eDP Replay without specify stream for eDP */
+bool dc_set_replay_allow_active(struct dc *dc, bool active)
+{
+	int i;
+	bool allow_active;
 
-void dc_allow_idle_optimizations(struct dc *dc, bool allow)
+	for (i = 0; i < dc->current_state->stream_count; i++) {
+		struct dc_link *link;
+		struct dc_stream_state *stream = dc->current_state->streams[i];
+
+		link = stream->link;
+		if (!link)
+			continue;
+
+		if (link->replay_settings.replay_feature_enabled) {
+			if (active && !link->replay_settings.replay_allow_active) {
+				allow_active = true;
+				if (!dc_link_set_replay_allow_active(link, &allow_active,
+					false, false, NULL))
+					return false;
+			} else if (!active && link->replay_settings.replay_allow_active) {
+				allow_active = false;
+				if (!dc_link_set_replay_allow_active(link, &allow_active,
+					true, false, NULL))
+					return false;
+			}
+		}
+	}
+
+	return true;
+}
+
+/* set IPS disable state */
+bool dc_set_ips_disable(struct dc *dc, unsigned int disable_ips)
 {
-	if (dc->debug.disable_idle_power_optimizations)
+	dc_exit_ips_for_hw_access(dc);
+
+	dc->config.disable_ips = disable_ips;
+
+	return true;
+}
+
+void dc_allow_idle_optimizations_internal(struct dc *dc, bool allow, char const *caller_name)
+{
+	int idle_fclk_khz = 0, idle_dramclk_khz = 0, i = 0;
+	enum mall_stream_type subvp_pipe_type[MAX_PIPES] = {0};
+	struct pipe_ctx *pipe = NULL;
+	struct dc_state *context = dc->current_state;
+
+	if (dc->debug.disable_idle_power_optimizations) {
+		DC_LOG_DEBUG("%s: disabled\n", __func__);
+		return;
+	}
+
+	if (allow != dc->idle_optimizations_allowed)
+		DC_LOG_IPS("%s: allow_idle old=%d new=%d (caller=%s)\n", __func__,
+			   dc->idle_optimizations_allowed, allow, caller_name);
+
+	if (dc->caps.ips_support && (dc->config.disable_ips == DMUB_IPS_DISABLE_ALL))
 		return;
 
 	if (dc->clk_mgr != NULL && dc->clk_mgr->funcs->is_smu_present)
@@ -3546,47 +5650,177 @@ void dc_allow_idle_optimizations(struct dc *dc, bool allow)
 	if (allow == dc->idle_optimizations_allowed)
 		return;
 
-	if (dc->hwss.apply_idle_power_optimizations && dc->hwss.apply_idle_power_optimizations(dc, allow))
+	if (dc->hwss.apply_idle_power_optimizations && dc->clk_mgr != NULL &&
+	    dc->hwss.apply_idle_power_optimizations(dc, allow)) {
 		dc->idle_optimizations_allowed = allow;
+		DC_LOG_DEBUG("%s: %s\n", __func__, allow ? "enabled" : "disabled");
+	}
+
+	// log idle clocks and sub vp pipe types at idle optimization time
+	if (dc->clk_mgr != NULL && dc->clk_mgr->funcs->get_hard_min_fclk)
+		idle_fclk_khz = dc->clk_mgr->funcs->get_hard_min_fclk(dc->clk_mgr);
+
+	if (dc->clk_mgr != NULL && dc->clk_mgr->funcs->get_hard_min_memclk)
+		idle_dramclk_khz = dc->clk_mgr->funcs->get_hard_min_memclk(dc->clk_mgr);
+
+	if (dc->res_pool && context) {
+		for (i = 0; i < dc->res_pool->pipe_count; i++) {
+			pipe = &context->res_ctx.pipe_ctx[i];
+			subvp_pipe_type[i] = dc_state_get_pipe_subvp_type(context, pipe);
+		}
+	}
+	if (!dc->caps.is_apu)
+		DC_LOG_DC("%s: allow_idle=%d\n HardMinUClk_Khz=%d HardMinDramclk_Khz=%d\n Pipe_0=%d Pipe_1=%d Pipe_2=%d Pipe_3=%d Pipe_4=%d Pipe_5=%d (caller=%s)\n",
+			__func__, allow, idle_fclk_khz, idle_dramclk_khz, subvp_pipe_type[0], subvp_pipe_type[1], subvp_pipe_type[2],
+			subvp_pipe_type[3], subvp_pipe_type[4], subvp_pipe_type[5], caller_name);
+
 }
 
-/*
- * blank all streams, and set min and max memory clock to
- * lowest and highest DPM level, respectively
- */
-void dc_unlock_memory_clock_frequency(struct dc *dc)
+void dc_exit_ips_for_hw_access_internal(struct dc *dc, const char *caller_name)
 {
-	unsigned int i;
+	if (dc->caps.ips_support)
+		dc_allow_idle_optimizations_internal(dc, false, caller_name);
+}
+
+bool dc_dmub_is_ips_idle_state(struct dc *dc)
+{
+	if (dc->debug.disable_idle_power_optimizations)
+		return false;
+
+	if (!dc->caps.ips_support || (dc->config.disable_ips == DMUB_IPS_DISABLE_ALL))
+		return false;
 
-	for (i = 0; i < MAX_PIPES; i++)
-		if (dc->current_state->res_ctx.pipe_ctx[i].plane_state)
-			core_link_disable_stream(&dc->current_state->res_ctx.pipe_ctx[i]);
+	if (!dc->ctx->dmub_srv)
+		return false;
 
-	dc->clk_mgr->funcs->set_hard_min_memclk(dc->clk_mgr, false);
-	dc->clk_mgr->funcs->set_hard_max_memclk(dc->clk_mgr);
+	return dc->ctx->dmub_srv->idle_allowed;
 }
 
-/*
- * set min memory clock to the min required for current mode,
- * max to maxDPM, and unblank streams
- */
+/* set min and max memory clock to lowest and highest DPM level, respectively */
+void dc_unlock_memory_clock_frequency(struct dc *dc)
+{
+	if (dc->clk_mgr->funcs->set_hard_min_memclk)
+		dc->clk_mgr->funcs->set_hard_min_memclk(dc->clk_mgr, false);
+
+	if (dc->clk_mgr->funcs->set_hard_max_memclk)
+		dc->clk_mgr->funcs->set_hard_max_memclk(dc->clk_mgr);
+}
+
+/* set min memory clock to the min required for current mode, max to maxDPM */
 void dc_lock_memory_clock_frequency(struct dc *dc)
 {
-	unsigned int i;
+	if (dc->clk_mgr->funcs->get_memclk_states_from_smu)
+		dc->clk_mgr->funcs->get_memclk_states_from_smu(dc->clk_mgr);
 
-	dc->clk_mgr->funcs->get_memclk_states_from_smu(dc->clk_mgr);
-	dc->clk_mgr->funcs->set_hard_min_memclk(dc->clk_mgr, true);
-	dc->clk_mgr->funcs->set_hard_max_memclk(dc->clk_mgr);
+	if (dc->clk_mgr->funcs->set_hard_min_memclk)
+		dc->clk_mgr->funcs->set_hard_min_memclk(dc->clk_mgr, true);
 
-	for (i = 0; i < MAX_PIPES; i++)
-		if (dc->current_state->res_ctx.pipe_ctx[i].plane_state)
-			core_link_enable_stream(dc->current_state, &dc->current_state->res_ctx.pipe_ctx[i]);
+	if (dc->clk_mgr->funcs->set_hard_max_memclk)
+		dc->clk_mgr->funcs->set_hard_max_memclk(dc->clk_mgr);
 }
 
-bool dc_is_plane_eligible_for_idle_optimizations(struct dc *dc, struct dc_plane_state *plane,
+static void blank_and_force_memclk(struct dc *dc, bool apply, unsigned int memclk_mhz)
+{
+	struct dc_state *context = dc->current_state;
+	struct hubp *hubp;
+	struct pipe_ctx *pipe;
+	int i;
+
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		pipe = &context->res_ctx.pipe_ctx[i];
+
+		if (pipe->stream != NULL) {
+			dc->hwss.disable_pixel_data(dc, pipe, true);
+
+			// wait for double buffer
+			pipe->stream_res.tg->funcs->wait_for_state(pipe->stream_res.tg, CRTC_STATE_VACTIVE);
+			pipe->stream_res.tg->funcs->wait_for_state(pipe->stream_res.tg, CRTC_STATE_VBLANK);
+			pipe->stream_res.tg->funcs->wait_for_state(pipe->stream_res.tg, CRTC_STATE_VACTIVE);
+
+			hubp = pipe->plane_res.hubp;
+			hubp->funcs->set_blank_regs(hubp, true);
+		}
+	}
+	if (dc->clk_mgr->funcs->set_max_memclk)
+		dc->clk_mgr->funcs->set_max_memclk(dc->clk_mgr, memclk_mhz);
+	if (dc->clk_mgr->funcs->set_min_memclk)
+		dc->clk_mgr->funcs->set_min_memclk(dc->clk_mgr, memclk_mhz);
+
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		pipe = &context->res_ctx.pipe_ctx[i];
+
+		if (pipe->stream != NULL) {
+			dc->hwss.disable_pixel_data(dc, pipe, false);
+
+			hubp = pipe->plane_res.hubp;
+			hubp->funcs->set_blank_regs(hubp, false);
+		}
+	}
+}
+
+
+/**
+ * dc_enable_dcmode_clk_limit() - lower clocks in dc (battery) mode
+ * @dc: pointer to dc of the dm calling this
+ * @enable: True = transition to DC mode, false = transition back to AC mode
+ *
+ * Some SoCs define additional clock limits when in DC mode, DM should
+ * invoke this function when the platform undergoes a power source transition
+ * so DC can apply/unapply the limit. This interface may be disruptive to
+ * the onscreen content.
+ *
+ * Context: Triggered by OS through DM interface, or manually by escape calls.
+ * Need to hold a dclock when doing so.
+ *
+ * Return: none (void function)
+ *
+ */
+void dc_enable_dcmode_clk_limit(struct dc *dc, bool enable)
+{
+	unsigned int softMax = 0, maxDPM = 0, funcMin = 0, i;
+	bool p_state_change_support;
+
+	if (!dc->config.dc_mode_clk_limit_support)
+		return;
+
+	softMax = dc->clk_mgr->bw_params->dc_mode_softmax_memclk;
+	for (i = 0; i < dc->clk_mgr->bw_params->clk_table.num_entries; i++) {
+		if (dc->clk_mgr->bw_params->clk_table.entries[i].memclk_mhz > maxDPM)
+			maxDPM = dc->clk_mgr->bw_params->clk_table.entries[i].memclk_mhz;
+	}
+	funcMin = (dc->clk_mgr->clks.dramclk_khz + 999) / 1000;
+	p_state_change_support = dc->clk_mgr->clks.p_state_change_support;
+
+	if (enable && !dc->clk_mgr->dc_mode_softmax_enabled) {
+		if (p_state_change_support) {
+			if (funcMin <= softMax && dc->clk_mgr->funcs->set_max_memclk)
+				dc->clk_mgr->funcs->set_max_memclk(dc->clk_mgr, softMax);
+			// else: No-Op
+		} else {
+			if (funcMin <= softMax)
+				blank_and_force_memclk(dc, true, softMax);
+			// else: No-Op
+		}
+	} else if (!enable && dc->clk_mgr->dc_mode_softmax_enabled) {
+		if (p_state_change_support) {
+			if (funcMin <= softMax && dc->clk_mgr->funcs->set_max_memclk)
+				dc->clk_mgr->funcs->set_max_memclk(dc->clk_mgr, maxDPM);
+			// else: No-Op
+		} else {
+			if (funcMin <= softMax)
+				blank_and_force_memclk(dc, true, maxDPM);
+			// else: No-Op
+		}
+	}
+	dc->clk_mgr->dc_mode_softmax_enabled = enable;
+}
+bool dc_is_plane_eligible_for_idle_optimizations(struct dc *dc,
+		unsigned int pitch,
+		unsigned int height,
+		enum surface_pixel_format format,
 		struct dc_cursor_attributes *cursor_attr)
 {
-	if (dc->hwss.does_plane_fit_in_mall && dc->hwss.does_plane_fit_in_mall(dc, plane, cursor_attr))
+	if (dc->hwss.does_plane_fit_in_mall && dc->hwss.does_plane_fit_in_mall(dc, pitch, height, format, cursor_attr))
 		return true;
 	return false;
 }
@@ -3594,28 +5828,89 @@ bool dc_is_plane_eligible_for_idle_optimizations(struct dc *dc, struct dc_plane_
 /* cleanup on driver unload */
 void dc_hardware_release(struct dc *dc)
 {
+	dc_mclk_switch_using_fw_based_vblank_stretch_shut_down(dc);
+
 	if (dc->hwss.hardware_release)
 		dc->hwss.hardware_release(dc);
 }
-#endif
+
+void dc_mclk_switch_using_fw_based_vblank_stretch_shut_down(struct dc *dc)
+{
+	if (dc->current_state)
+		dc->current_state->bw_ctx.bw.dcn.clk.fw_based_mclk_switching_shut_down = true;
+}
 
 /**
- * dc_enable_dmub_notifications - Returns whether dmub notification can be enabled
- * @dc: dc structure
+ * dc_is_dmub_outbox_supported - Check if DMUB firmware support outbox notification
+ *
+ * @dc: [in] dc structure
  *
- * Returns: True to enable dmub notifications, False otherwise
+ * Checks whether DMUB FW supports outbox notifications, if supported DM
+ * should register outbox interrupt prior to actually enabling interrupts
+ * via dc_enable_dmub_outbox
+ *
+ * Return:
+ * True if DMUB FW supports outbox notifications, False otherwise
  */
-bool dc_enable_dmub_notifications(struct dc *dc)
+bool dc_is_dmub_outbox_supported(struct dc *dc)
 {
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	/* YELLOW_CARP B0 USB4 DPIA needs dmub notifications for interrupts */
-	if (dc->ctx->asic_id.chip_family == FAMILY_YELLOW_CARP &&
-	    dc->ctx->asic_id.hw_internal_rev == YELLOW_CARP_B0 &&
-	    !dc->debug.dpia_debug.bits.disable_dpia)
-		return true;
-#endif
+	if (!dc->caps.dmcub_support)
+		return false;
+
+	switch (dc->ctx->asic_id.chip_family) {
+
+	case FAMILY_YELLOW_CARP:
+		/* DCN31 B0 USB4 DPIA needs dmub notifications for interrupts */
+		if (dc->ctx->asic_id.hw_internal_rev == YELLOW_CARP_B0 &&
+		    !dc->debug.dpia_debug.bits.disable_dpia)
+			return true;
+	break;
+
+	case AMDGPU_FAMILY_GC_11_0_1:
+	case AMDGPU_FAMILY_GC_11_5_0:
+		if (!dc->debug.dpia_debug.bits.disable_dpia)
+			return true;
+	break;
+
+	default:
+		break;
+	}
+
 	/* dmub aux needs dmub notifications to be enabled */
 	return dc->debug.enable_dmub_aux_for_legacy_ddc;
+
+}
+
+/**
+ * dc_enable_dmub_notifications - Check if dmub fw supports outbox
+ *
+ * @dc: [in] dc structure
+ *
+ * Calls dc_is_dmub_outbox_supported to check if dmub fw supports outbox
+ * notifications. All DMs shall switch to dc_is_dmub_outbox_supported.  This
+ * API shall be removed after switching.
+ *
+ * Return:
+ * True if DMUB FW supports outbox notifications, False otherwise
+ */
+bool dc_enable_dmub_notifications(struct dc *dc)
+{
+	return dc_is_dmub_outbox_supported(dc);
+}
+
+/**
+ * dc_enable_dmub_outbox - Enables DMUB unsolicited notification
+ *
+ * @dc: [in] dc structure
+ *
+ * Enables DMUB unsolicited notifications to x86 via outbox.
+ */
+void dc_enable_dmub_outbox(struct dc *dc)
+{
+	struct dc_context *dc_ctx = dc->ctx;
+
+	dmub_enable_outbox_notification(dc_ctx->dmub_srv);
+	DC_LOG_DC("%s: dmub outbox notifications enabled\n", __func__);
 }
 
 /**
@@ -3633,7 +5928,6 @@ bool dc_process_dmub_aux_transfer_async(struct dc *dc,
 {
 	uint8_t action;
 	union dmub_rb_cmd cmd = {0};
-	struct dc_dmub_srv *dmub_srv = dc->ctx->dmub_srv;
 
 	ASSERT(payload->length <= 16);
 
@@ -3681,13 +5975,106 @@ bool dc_process_dmub_aux_transfer_async(struct dc *dc,
 			);
 	}
 
-	dc_dmub_srv_cmd_queue(dmub_srv, &cmd);
-	dc_dmub_srv_cmd_execute(dmub_srv);
-	dc_dmub_srv_wait_idle(dmub_srv);
+	dc_wake_and_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
 
 	return true;
 }
 
+bool dc_smart_power_oled_enable(const struct dc_link *link, bool enable, uint16_t peak_nits,
+			uint8_t debug_control, uint16_t fixed_CLL, uint32_t triggerline)
+{
+	bool status = false;
+	struct dc *dc = link->ctx->dc;
+	union dmub_rb_cmd cmd;
+	uint8_t otg_inst = 0;
+	unsigned int panel_inst = 0;
+	struct pipe_ctx *pipe_ctx = NULL;
+	struct resource_context *res_ctx = &link->ctx->dc->current_state->res_ctx;
+	int i = 0;
+
+	// get panel_inst
+	if (!dc_get_edp_link_panel_inst(dc, link, &panel_inst))
+		return status;
+
+	// get otg_inst
+	for (i = 0; i < MAX_PIPES; i++) {
+		if (res_ctx &&
+			res_ctx->pipe_ctx[i].stream &&
+			res_ctx->pipe_ctx[i].stream->link &&
+			res_ctx->pipe_ctx[i].stream->link == link &&
+			res_ctx->pipe_ctx[i].stream->link->connector_signal == SIGNAL_TYPE_EDP) {
+			pipe_ctx = &res_ctx->pipe_ctx[i];
+			//TODO: refactor for multi edp support
+			break;
+		}
+	}
+
+	if (pipe_ctx)
+		otg_inst = pipe_ctx->stream_res.tg->inst;
+
+	// before enable smart power OLED, we need to call set pipe for DMUB to set ABM config
+	if (enable) {
+		if (dc->hwss.set_pipe && pipe_ctx)
+			dc->hwss.set_pipe(pipe_ctx);
+	}
+
+	// fill in cmd
+	memset(&cmd, 0, sizeof(cmd));
+
+	cmd.smart_power_oled_enable.header.type = DMUB_CMD__SMART_POWER_OLED;
+	cmd.smart_power_oled_enable.header.sub_type = DMUB_CMD__SMART_POWER_OLED_ENABLE;
+	cmd.smart_power_oled_enable.header.payload_bytes =
+		sizeof(struct dmub_rb_cmd_smart_power_oled_enable_data) - sizeof(struct dmub_cmd_header);
+	cmd.smart_power_oled_enable.header.ret_status = 1;
+	cmd.smart_power_oled_enable.data.enable = enable;
+	cmd.smart_power_oled_enable.data.panel_inst = panel_inst;
+	cmd.smart_power_oled_enable.data.peak_nits = peak_nits;
+	cmd.smart_power_oled_enable.data.otg_inst = otg_inst;
+	cmd.smart_power_oled_enable.data.digfe_inst = link->link_enc->preferred_engine;
+	cmd.smart_power_oled_enable.data.digbe_inst = link->link_enc->transmitter;
+
+	cmd.smart_power_oled_enable.data.debugcontrol = debug_control;
+	cmd.smart_power_oled_enable.data.triggerline = triggerline;
+	cmd.smart_power_oled_enable.data.fixed_max_cll = fixed_CLL;
+
+	// send cmd
+	status = dc_wake_and_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
+
+	return status;
+}
+
+bool dc_smart_power_oled_get_max_cll(const struct dc_link *link, unsigned int *pCurrent_MaxCLL)
+{
+	struct dc *dc = link->ctx->dc;
+	union dmub_rb_cmd cmd;
+	bool status = false;
+	unsigned int panel_inst = 0;
+
+	// get panel_inst
+	if (!dc_get_edp_link_panel_inst(dc, link, &panel_inst))
+		return status;
+
+	// fill in cmd
+	memset(&cmd, 0, sizeof(cmd));
+
+	cmd.smart_power_oled_getmaxcll.header.type = DMUB_CMD__SMART_POWER_OLED;
+	cmd.smart_power_oled_getmaxcll.header.sub_type = DMUB_CMD__SMART_POWER_OLED_GETMAXCLL;
+	cmd.smart_power_oled_getmaxcll.header.payload_bytes = sizeof(cmd.smart_power_oled_getmaxcll.data);
+	cmd.smart_power_oled_getmaxcll.header.ret_status = 1;
+
+	cmd.smart_power_oled_getmaxcll.data.input.panel_inst = panel_inst;
+
+	// send cmd and wait for reply
+	status = dc_wake_and_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY);
+
+	if (status)
+		*pCurrent_MaxCLL = cmd.smart_power_oled_getmaxcll.data.output.current_max_cll;
+	else
+		*pCurrent_MaxCLL = 0;
+
+	return status;
+}
+
 uint8_t get_link_index_from_dpia_port_index(const struct dc *dc,
 					    uint8_t dpia_port_index)
 {
@@ -3709,21 +6096,17 @@ uint8_t get_link_index_from_dpia_port_index(const struct dc *dc,
 }
 
 /**
- *****************************************************************************
- *  Function: dc_process_dmub_set_config_async
+ * dc_process_dmub_set_config_async - Submits set_config command
  *
- *  @brief
- *		Submits set_config command to dmub via inbox message
+ * @dc: [in] dc structure
+ * @link_index: [in] link_index: link index
+ * @payload: [in] aux payload
+ * @notify: [out] set_config immediate reply
  *
- *  @param
- *		[in] dc: dc structure
- *		[in] link_index: link index
- *		[in] payload: aux payload
- *		[out] notify: set_config immediate reply
+ * Submits set_config command to dmub via inbox message.
  *
- *	@return
- *		True if successful, False if failure
- *****************************************************************************
+ * Return:
+ * True if successful, False if failure
  */
 bool dc_process_dmub_set_config_async(struct dc *dc,
 				uint32_t link_index,
@@ -3731,7 +6114,6 @@ bool dc_process_dmub_set_config_async(struct dc *dc,
 				struct dmub_notification *notify)
 {
 	union dmub_rb_cmd cmd = {0};
-	struct dc_dmub_srv *dmub_srv = dc->ctx->dmub_srv;
 	bool is_cmd_complete = true;
 
 	/* prepare SET_CONFIG command */
@@ -3742,7 +6124,7 @@ bool dc_process_dmub_set_config_async(struct dc *dc,
 	cmd.set_config_access.set_config_control.cmd_pkt.msg_type = payload->msg_type;
 	cmd.set_config_access.set_config_control.cmd_pkt.msg_data = payload->msg_data;
 
-	if (!dc_dmub_srv_cmd_with_reply_data(dmub_srv, &cmd)) {
+	if (!dc_wake_and_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY)) {
 		/* command is not processed by dmub */
 		notify->sc_status = SET_CONFIG_UNKNOWN_ERROR;
 		return is_cmd_complete;
@@ -3759,21 +6141,17 @@ bool dc_process_dmub_set_config_async(struct dc *dc,
 }
 
 /**
- *****************************************************************************
- *  Function: dc_process_dmub_set_mst_slots
+ * dc_process_dmub_set_mst_slots - Submits MST solt allocation
  *
- *  @brief
- *		Submits mst slot allocation command to dmub via inbox message
+ * @dc: [in] dc structure
+ * @link_index: [in] link index
+ * @mst_alloc_slots: [in] mst slots to be allotted
+ * @mst_slots_in_use: [out] mst slots in use returned in failure case
  *
- *  @param
- *		[in] dc: dc structure
- *		[in] link_index: link index
- *		[in] mst_alloc_slots: mst slots to be allotted
- *		[out] mst_slots_in_use: mst slots in use returned in failure case
+ * Submits mst slot allocation command to dmub via inbox message
  *
- *	@return
- *		DC_OK if successful, DC_ERROR if failure
- *****************************************************************************
+ * Return:
+ * DC_OK if successful, DC_ERROR if failure
  */
 enum dc_status dc_process_dmub_set_mst_slots(const struct dc *dc,
 				uint32_t link_index,
@@ -3781,7 +6159,6 @@ enum dc_status dc_process_dmub_set_mst_slots(const struct dc *dc,
 				uint8_t *mst_slots_in_use)
 {
 	union dmub_rb_cmd cmd = {0};
-	struct dc_dmub_srv *dmub_srv = dc->ctx->dmub_srv;
 
 	/* prepare MST_ALLOC_SLOTS command */
 	cmd.set_mst_alloc_slots.header.type = DMUB_CMD__DPIA;
@@ -3790,7 +6167,7 @@ enum dc_status dc_process_dmub_set_mst_slots(const struct dc *dc,
 	cmd.set_mst_alloc_slots.mst_slots_control.instance = dc->links[link_index]->ddc_hw_inst;
 	cmd.set_mst_alloc_slots.mst_slots_control.mst_alloc_slots = mst_alloc_slots;
 
-	if (!dc_dmub_srv_cmd_with_reply_data(dmub_srv, &cmd))
+	if (!dc_wake_and_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY))
 		/* command is not processed by dmub */
 		return DC_ERROR_UNEXPECTED;
 
@@ -3813,6 +6190,60 @@ enum dc_status dc_process_dmub_set_mst_slots(const struct dc *dc,
 }
 
 /**
+ * dc_process_dmub_dpia_set_tps_notification - Submits tps notification
+ *
+ * @dc: [in] dc structure
+ * @link_index: [in] link index
+ * @tps: [in] request tps
+ *
+ * Submits set_tps_notification command to dmub via inbox message
+ */
+void dc_process_dmub_dpia_set_tps_notification(const struct dc *dc, uint32_t link_index, uint8_t tps)
+{
+	union dmub_rb_cmd cmd = {0};
+
+	cmd.set_tps_notification.header.type = DMUB_CMD__DPIA;
+	cmd.set_tps_notification.header.sub_type = DMUB_CMD__DPIA_SET_TPS_NOTIFICATION;
+	cmd.set_tps_notification.tps_notification.instance = dc->links[link_index]->ddc_hw_inst;
+	cmd.set_tps_notification.tps_notification.tps = tps;
+
+	dc_wake_and_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
+}
+
+/**
+ * dc_process_dmub_dpia_hpd_int_enable - Submits DPIA DPD interruption
+ *
+ * @dc: [in] dc structure
+ * @hpd_int_enable: [in] 1 for hpd int enable, 0 to disable
+ *
+ * Submits dpia hpd int enable command to dmub via inbox message
+ */
+void dc_process_dmub_dpia_hpd_int_enable(const struct dc *dc,
+				uint32_t hpd_int_enable)
+{
+	union dmub_rb_cmd cmd = {0};
+
+	cmd.dpia_hpd_int_enable.header.type = DMUB_CMD__DPIA_HPD_INT_ENABLE;
+	cmd.dpia_hpd_int_enable.enable = hpd_int_enable;
+
+	dc_wake_and_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
+
+	DC_LOG_DEBUG("%s: hpd_int_enable(%d)\n", __func__, hpd_int_enable);
+}
+
+/**
+ * dc_print_dmub_diagnostic_data - Print DMUB diagnostic data for debugging
+ *
+ * @dc: [in] dc structure
+ *
+ *
+ */
+void dc_print_dmub_diagnostic_data(const struct dc *dc)
+{
+	dc_dmub_srv_log_diagnostic_data(dc->ctx->dmub_srv);
+}
+
+/**
  * dc_disable_accelerated_mode - disable accelerated mode
  * @dc: dc structure
  */
@@ -3823,16 +6254,13 @@ void dc_disable_accelerated_mode(struct dc *dc)
 
 
 /**
- *****************************************************************************
- *  dc_notify_vsync_int_state() - notifies vsync enable/disable state
+ *  dc_notify_vsync_int_state - notifies vsync enable/disable state
  *  @dc: dc structure
- *	@stream: stream where vsync int state changed
- *	@enable: whether vsync is enabled or disabled
- *
- *  Called when vsync is enabled/disabled
- *	Will notify DMUB to start/stop ABM interrupts after steady state is reached
+ *  @stream: stream where vsync int state changed
+ *  @enable: whether vsync is enabled or disabled
  *
- *****************************************************************************
+ *  Called when vsync is enabled/disabled Will notify DMUB to start/stop ABM
+ *  interrupts after steady state is reached.
  */
 void dc_notify_vsync_int_state(struct dc *dc, struct dc_stream_state *stream, bool enable)
 {
@@ -3846,6 +6274,9 @@ void dc_notify_vsync_int_state(struct dc *dc, struct dc_stream_state *stream, bo
 	if (link->psr_settings.psr_feature_enabled)
 		return;
 
+	if (link->replay_settings.replay_feature_enabled)
+		return;
+
 	/*find primary pipe associated with stream*/
 	for (i = 0; i < MAX_PIPES; i++) {
 		pipe = &dc->current_state->res_ctx.pipe_ctx[i];
@@ -3859,7 +6290,7 @@ void dc_notify_vsync_int_state(struct dc *dc, struct dc_stream_state *stream, bo
 		return;
 	}
 
-	get_edp_links(dc, edp_links, &edp_num);
+	dc_get_edp_links(dc, edp_links, &edp_num);
 
 	/* Determine panel inst */
 	for (i = 0; i < edp_num; i++) {
@@ -3874,3 +6305,789 @@ void dc_notify_vsync_int_state(struct dc *dc, struct dc_stream_state *stream, bo
 	if (pipe->stream_res.abm && pipe->stream_res.abm->funcs->set_abm_pause)
 		pipe->stream_res.abm->funcs->set_abm_pause(pipe->stream_res.abm, !enable, i, pipe->stream_res.tg->inst);
 }
+
+/*****************************************************************************
+ *  dc_abm_save_restore() - Interface to DC for save+pause and restore+un-pause
+ *                          ABM
+ *  @dc: dc structure
+ *	@stream: stream where vsync int state changed
+ *  @pData: abm hw states
+ *
+ ****************************************************************************/
+bool dc_abm_save_restore(
+		struct dc *dc,
+		struct dc_stream_state *stream,
+		struct abm_save_restore *pData)
+{
+	int i;
+	int edp_num;
+	struct pipe_ctx *pipe = NULL;
+	struct dc_link *link = stream->sink->link;
+	struct dc_link *edp_links[MAX_NUM_EDP];
+
+	if (link->replay_settings.replay_feature_enabled)
+		return false;
+
+	/*find primary pipe associated with stream*/
+	for (i = 0; i < MAX_PIPES; i++) {
+		pipe = &dc->current_state->res_ctx.pipe_ctx[i];
+
+		if (pipe->stream == stream && pipe->stream_res.tg)
+			break;
+	}
+
+	if (i == MAX_PIPES) {
+		ASSERT(0);
+		return false;
+	}
+
+	dc_get_edp_links(dc, edp_links, &edp_num);
+
+	/* Determine panel inst */
+	for (i = 0; i < edp_num; i++)
+		if (edp_links[i] == link)
+			break;
+
+	if (i == edp_num)
+		return false;
+
+	if (pipe->stream_res.abm &&
+		pipe->stream_res.abm->funcs->save_restore)
+		return pipe->stream_res.abm->funcs->save_restore(
+				pipe->stream_res.abm,
+				i,
+				pData);
+	return false;
+}
+
+void dc_query_current_properties(struct dc *dc, struct dc_current_properties *properties)
+{
+	unsigned int i;
+	unsigned int max_cursor_size = dc->caps.max_cursor_size;
+	unsigned int stream_cursor_size;
+
+	if (dc->debug.allow_sw_cursor_fallback && dc->res_pool->funcs->get_max_hw_cursor_size) {
+		for (i = 0; i < dc->current_state->stream_count; i++) {
+			stream_cursor_size = dc->res_pool->funcs->get_max_hw_cursor_size(dc,
+					dc->current_state,
+					dc->current_state->streams[i]);
+
+			if (stream_cursor_size < max_cursor_size) {
+				max_cursor_size = stream_cursor_size;
+			}
+		}
+	}
+
+	properties->cursor_size_limit = max_cursor_size;
+}
+
+/**
+ * dc_set_edp_power() - DM controls eDP power to be ON/OFF
+ *
+ * Called when DM wants to power on/off eDP.
+ *     Only work on links with flag skip_implict_edp_power_control is set.
+ *
+ * @dc: Current DC state
+ * @edp_link: a link with eDP connector signal type
+ * @powerOn: power on/off eDP
+ *
+ * Return: void
+ */
+void dc_set_edp_power(const struct dc *dc, struct dc_link *edp_link,
+				 bool powerOn)
+{
+	if (edp_link->connector_signal != SIGNAL_TYPE_EDP)
+		return;
+
+	if (edp_link->skip_implict_edp_power_control == false)
+		return;
+
+	edp_link->dc->link_srv->edp_set_panel_power(edp_link, powerOn);
+}
+
+/**
+ * dc_get_power_profile_for_dc_state() - extracts power profile from dc state
+ *
+ * Called when DM wants to make power policy decisions based on dc_state
+ *
+ * @context: Pointer to the dc_state from which the power profile is extracted.
+ *
+ * Return: The power profile structure containing the power level information.
+ */
+struct dc_power_profile dc_get_power_profile_for_dc_state(const struct dc_state *context)
+{
+	struct dc_power_profile profile = { 0 };
+
+	profile.power_level = !context->bw_ctx.bw.dcn.clk.p_state_change_support;
+	if (!context->clk_mgr || !context->clk_mgr->ctx || !context->clk_mgr->ctx->dc)
+		return profile;
+	struct dc *dc = context->clk_mgr->ctx->dc;
+
+	if (dc->res_pool->funcs->get_power_profile)
+		profile.power_level = dc->res_pool->funcs->get_power_profile(context);
+	return profile;
+}
+
+/**
+ * dc_get_det_buffer_size_from_state() - extracts detile buffer size from dc state
+ *
+ * This function is called to log the detile buffer size from the dc_state.
+ *
+ * @context: a pointer to the dc_state from which the detile buffer size is extracted.
+ *
+ * Return: the size of the detile buffer, or 0 if not available.
+ */
+unsigned int dc_get_det_buffer_size_from_state(const struct dc_state *context)
+{
+	struct dc *dc = context->clk_mgr->ctx->dc;
+
+	if (dc->res_pool->funcs->get_det_buffer_size)
+		return dc->res_pool->funcs->get_det_buffer_size(context);
+	else
+		return 0;
+}
+
+/**
+ * dc_get_host_router_index: Get index of host router from a dpia link
+ *
+ * This function return a host router index of the target link. If the target link is dpia link.
+ *
+ * @link: Pointer to the target link (input)
+ * @host_router_index: Pointer to store the host router index of the target link (output).
+ *
+ * Return: true if the host router index is found and valid.
+ *
+ */
+bool dc_get_host_router_index(const struct dc_link *link, unsigned int *host_router_index)
+{
+	struct dc *dc;
+
+	if (!link || !host_router_index || link->ep_type != DISPLAY_ENDPOINT_USB4_DPIA)
+		return false;
+
+	dc = link->ctx->dc;
+
+	if (link->link_index < dc->lowest_dpia_link_index)
+		return false;
+
+	*host_router_index = (link->link_index - dc->lowest_dpia_link_index) / dc->caps.num_of_dpias_per_host_router;
+	if (*host_router_index < dc->caps.num_of_host_routers)
+		return true;
+	else
+		return false;
+}
+
+bool dc_is_cursor_limit_pending(struct dc *dc)
+{
+	uint32_t i;
+
+	for (i = 0; i < dc->current_state->stream_count; i++) {
+		if (dc_stream_is_cursor_limit_pending(dc, dc->current_state->streams[i]))
+			return true;
+	}
+
+	return false;
+}
+
+bool dc_can_clear_cursor_limit(const struct dc *dc)
+{
+	uint32_t i;
+
+	for (i = 0; i < dc->current_state->stream_count; i++) {
+		if (dc_state_can_clear_stream_cursor_subvp_limit(dc->current_state->streams[i], dc->current_state))
+			return true;
+	}
+
+	return false;
+}
+
+void dc_get_underflow_debug_data_for_otg(struct dc *dc, int primary_otg_inst,
+				struct dc_underflow_debug_data *out_data)
+{
+	struct timing_generator *tg = NULL;
+
+	for (int i = 0; i < MAX_PIPES; i++) {
+		if (dc->res_pool->timing_generators[i] &&
+			dc->res_pool->timing_generators[i]->inst == primary_otg_inst) {
+				tg = dc->res_pool->timing_generators[i];
+				break;
+		}
+	}
+
+	dc_exit_ips_for_hw_access(dc);
+	if (dc->hwss.get_underflow_debug_data)
+		dc->hwss.get_underflow_debug_data(dc, tg, out_data);
+}
+
+void dc_get_power_feature_status(struct dc *dc, int primary_otg_inst,
+				struct power_features *out_data)
+{
+	out_data->uclk_p_state = dc->current_state->clk_mgr->clks.p_state_change_support;
+	out_data->fams = dc->current_state->bw_ctx.bw.dcn.clk.fw_based_mclk_switching;
+}
+
+bool dc_capture_register_software_state(struct dc *dc, struct dc_register_software_state *state)
+{
+	struct dc_state *context;
+	struct resource_context *res_ctx;
+	int i;
+
+	if (!dc || !dc->current_state || !state) {
+		if (state)
+			state->state_valid = false;
+		return false;
+	}
+
+	/* Initialize the state structure */
+	memset(state, 0, sizeof(struct dc_register_software_state));
+
+	context = dc->current_state;
+	res_ctx = &context->res_ctx;
+
+	/* Count active pipes and streams */
+	state->active_pipe_count = 0;
+	state->active_stream_count = context->stream_count;
+
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		if (res_ctx->pipe_ctx[i].stream)
+			state->active_pipe_count++;
+	}
+
+	/* Capture HUBP programming state for each pipe */
+	for (i = 0; i < MAX_PIPES && i < dc->res_pool->pipe_count; i++) {
+		struct pipe_ctx *pipe_ctx = &res_ctx->pipe_ctx[i];
+
+		state->hubp[i].valid_stream = false;
+		if (!pipe_ctx->stream)
+			continue;
+
+		state->hubp[i].valid_stream = true;
+
+		/* HUBP register programming variables */
+		if (pipe_ctx->stream_res.tg)
+			state->hubp[i].vtg_sel = pipe_ctx->stream_res.tg->inst;
+
+		state->hubp[i].hubp_clock_enable = (pipe_ctx->plane_res.hubp != NULL) ? 1 : 0;
+
+		state->hubp[i].valid_plane_state = false;
+		if (pipe_ctx->plane_state) {
+			state->hubp[i].valid_plane_state = true;
+			state->hubp[i].surface_pixel_format = pipe_ctx->plane_state->format;
+			state->hubp[i].rotation_angle = pipe_ctx->plane_state->rotation;
+			state->hubp[i].h_mirror_en = pipe_ctx->plane_state->horizontal_mirror ? 1 : 0;
+
+			/* Surface size */
+			if (pipe_ctx->plane_state->plane_size.surface_size.width > 0) {
+				state->hubp[i].surface_size_width = pipe_ctx->plane_state->plane_size.surface_size.width;
+				state->hubp[i].surface_size_height = pipe_ctx->plane_state->plane_size.surface_size.height;
+			}
+
+			/* Viewport dimensions from scaler data */
+			if (pipe_ctx->plane_state->src_rect.width > 0) {
+				state->hubp[i].pri_viewport_width = pipe_ctx->plane_state->src_rect.width;
+				state->hubp[i].pri_viewport_height = pipe_ctx->plane_state->src_rect.height;
+				state->hubp[i].pri_viewport_x_start = pipe_ctx->plane_state->src_rect.x;
+				state->hubp[i].pri_viewport_y_start = pipe_ctx->plane_state->src_rect.y;
+			}
+
+			/* DCC settings */
+			state->hubp[i].surface_dcc_en = (pipe_ctx->plane_state->dcc.enable) ? 1 : 0;
+			state->hubp[i].surface_dcc_ind_64b_blk = pipe_ctx->plane_state->dcc.independent_64b_blks;
+			state->hubp[i].surface_dcc_ind_128b_blk = pipe_ctx->plane_state->dcc.dcc_ind_blk;
+
+			/* Surface pitch */
+			state->hubp[i].surface_pitch = pipe_ctx->plane_state->plane_size.surface_pitch;
+			state->hubp[i].meta_pitch = pipe_ctx->plane_state->dcc.meta_pitch;
+			state->hubp[i].chroma_pitch = pipe_ctx->plane_state->plane_size.chroma_pitch;
+			state->hubp[i].meta_pitch_c = pipe_ctx->plane_state->dcc.meta_pitch_c;
+
+			/* Surface addresses - primary */
+			state->hubp[i].primary_surface_address_low = pipe_ctx->plane_state->address.grph.addr.low_part;
+			state->hubp[i].primary_surface_address_high = pipe_ctx->plane_state->address.grph.addr.high_part;
+			state->hubp[i].primary_meta_surface_address_low = pipe_ctx->plane_state->address.grph.meta_addr.low_part;
+			state->hubp[i].primary_meta_surface_address_high = pipe_ctx->plane_state->address.grph.meta_addr.high_part;
+
+			/* TMZ settings */
+			state->hubp[i].primary_surface_tmz = pipe_ctx->plane_state->address.tmz_surface;
+			state->hubp[i].primary_meta_surface_tmz = pipe_ctx->plane_state->address.tmz_surface;
+
+			/* Tiling configuration */
+			state->hubp[i].min_dc_gfx_version9 = false;
+			if (pipe_ctx->plane_state->tiling_info.gfxversion >= DcGfxVersion9) {
+				state->hubp[i].min_dc_gfx_version9 = true;
+				state->hubp[i].sw_mode = pipe_ctx->plane_state->tiling_info.gfx9.swizzle;
+				state->hubp[i].num_pipes = pipe_ctx->plane_state->tiling_info.gfx9.num_pipes;
+				state->hubp[i].num_banks = pipe_ctx->plane_state->tiling_info.gfx9.num_banks;
+				state->hubp[i].pipe_interleave = pipe_ctx->plane_state->tiling_info.gfx9.pipe_interleave;
+				state->hubp[i].num_shader_engines = pipe_ctx->plane_state->tiling_info.gfx9.num_shader_engines;
+				state->hubp[i].num_rb_per_se = pipe_ctx->plane_state->tiling_info.gfx9.num_rb_per_se;
+				state->hubp[i].num_pkrs = pipe_ctx->plane_state->tiling_info.gfx9.num_pkrs;
+			}
+		}
+
+		/* DML Request Size Configuration */
+		if (pipe_ctx->rq_regs.rq_regs_l.chunk_size > 0) {
+			state->hubp[i].rq_chunk_size = pipe_ctx->rq_regs.rq_regs_l.chunk_size;
+			state->hubp[i].rq_min_chunk_size = pipe_ctx->rq_regs.rq_regs_l.min_chunk_size;
+			state->hubp[i].rq_meta_chunk_size = pipe_ctx->rq_regs.rq_regs_l.meta_chunk_size;
+			state->hubp[i].rq_min_meta_chunk_size = pipe_ctx->rq_regs.rq_regs_l.min_meta_chunk_size;
+			state->hubp[i].rq_dpte_group_size = pipe_ctx->rq_regs.rq_regs_l.dpte_group_size;
+			state->hubp[i].rq_mpte_group_size = pipe_ctx->rq_regs.rq_regs_l.mpte_group_size;
+			state->hubp[i].rq_swath_height_l = pipe_ctx->rq_regs.rq_regs_l.swath_height;
+			state->hubp[i].rq_pte_row_height_l = pipe_ctx->rq_regs.rq_regs_l.pte_row_height_linear;
+		}
+
+		/* Chroma request size configuration */
+		if (pipe_ctx->rq_regs.rq_regs_c.chunk_size > 0) {
+			state->hubp[i].rq_chunk_size_c = pipe_ctx->rq_regs.rq_regs_c.chunk_size;
+			state->hubp[i].rq_min_chunk_size_c = pipe_ctx->rq_regs.rq_regs_c.min_chunk_size;
+			state->hubp[i].rq_meta_chunk_size_c = pipe_ctx->rq_regs.rq_regs_c.meta_chunk_size;
+			state->hubp[i].rq_min_meta_chunk_size_c = pipe_ctx->rq_regs.rq_regs_c.min_meta_chunk_size;
+			state->hubp[i].rq_dpte_group_size_c = pipe_ctx->rq_regs.rq_regs_c.dpte_group_size;
+			state->hubp[i].rq_mpte_group_size_c = pipe_ctx->rq_regs.rq_regs_c.mpte_group_size;
+			state->hubp[i].rq_swath_height_c = pipe_ctx->rq_regs.rq_regs_c.swath_height;
+			state->hubp[i].rq_pte_row_height_c = pipe_ctx->rq_regs.rq_regs_c.pte_row_height_linear;
+		}
+
+		/* DML expansion modes */
+		state->hubp[i].drq_expansion_mode = pipe_ctx->rq_regs.drq_expansion_mode;
+		state->hubp[i].prq_expansion_mode = pipe_ctx->rq_regs.prq_expansion_mode;
+		state->hubp[i].mrq_expansion_mode = pipe_ctx->rq_regs.mrq_expansion_mode;
+		state->hubp[i].crq_expansion_mode = pipe_ctx->rq_regs.crq_expansion_mode;
+
+		/* DML DLG parameters - nominal */
+		state->hubp[i].dst_y_per_vm_vblank = pipe_ctx->dlg_regs.dst_y_per_vm_vblank;
+		state->hubp[i].dst_y_per_row_vblank = pipe_ctx->dlg_regs.dst_y_per_row_vblank;
+		state->hubp[i].dst_y_per_vm_flip = pipe_ctx->dlg_regs.dst_y_per_vm_flip;
+		state->hubp[i].dst_y_per_row_flip = pipe_ctx->dlg_regs.dst_y_per_row_flip;
+
+		/* DML prefetch settings */
+		state->hubp[i].dst_y_prefetch = pipe_ctx->dlg_regs.dst_y_prefetch;
+		state->hubp[i].vratio_prefetch = pipe_ctx->dlg_regs.vratio_prefetch;
+		state->hubp[i].vratio_prefetch_c = pipe_ctx->dlg_regs.vratio_prefetch_c;
+
+		/* TTU parameters */
+		state->hubp[i].qos_level_low_wm = pipe_ctx->ttu_regs.qos_level_low_wm;
+		state->hubp[i].qos_level_high_wm = pipe_ctx->ttu_regs.qos_level_high_wm;
+		state->hubp[i].qos_level_flip = pipe_ctx->ttu_regs.qos_level_flip;
+		state->hubp[i].min_ttu_vblank = pipe_ctx->ttu_regs.min_ttu_vblank;
+	}
+
+	/* Capture HUBBUB programming state */
+	if (dc->res_pool->hubbub) {
+		/* Individual DET buffer sizes - software state variables that program DET registers */
+		for (i = 0; i < 4 && i < dc->res_pool->pipe_count; i++) {
+			uint32_t det_size = res_ctx->pipe_ctx[i].det_buffer_size_kb;
+			switch (i) {
+			case 0:
+				state->hubbub.det0_size = det_size;
+				break;
+			case 1:
+				state->hubbub.det1_size = det_size;
+				break;
+			case 2:
+				state->hubbub.det2_size = det_size;
+				break;
+			case 3:
+				state->hubbub.det3_size = det_size;
+				break;
+			}
+		}
+
+		/* Compression buffer configuration - software state that programs COMPBUF_SIZE register */
+		// TODO: Handle logic for legacy DCN pre-DCN401
+		state->hubbub.compbuf_size = context->bw_ctx.bw.dcn.arb_regs.compbuf_size;
+	}
+
+	/* Capture DPP programming state for each pipe */
+	for (i = 0; i < MAX_PIPES && i < dc->res_pool->pipe_count; i++) {
+		struct pipe_ctx *pipe_ctx = &res_ctx->pipe_ctx[i];
+
+		if (!pipe_ctx->stream)
+			continue;
+
+		state->dpp[i].dpp_clock_enable = (pipe_ctx->plane_res.dpp != NULL) ? 1 : 0;
+
+		if (pipe_ctx->plane_state && pipe_ctx->plane_res.scl_data.recout.width > 0) {
+			/* Access dscl_prog_data directly - this contains the actual software state used for register programming */
+			struct dscl_prog_data *dscl_data = &pipe_ctx->plane_res.scl_data.dscl_prog_data;
+
+			/* Recout (Rectangle of Interest) configuration - software state that programs RECOUT registers */
+			state->dpp[i].recout_start_x = dscl_data->recout.x;
+			state->dpp[i].recout_start_y = dscl_data->recout.y;
+			state->dpp[i].recout_width = dscl_data->recout.width;
+			state->dpp[i].recout_height = dscl_data->recout.height;
+
+			/* MPC (Multiple Pipe/Plane Combiner) size - software state that programs MPC_SIZE registers */
+			state->dpp[i].mpc_width = dscl_data->mpc_size.width;
+			state->dpp[i].mpc_height = dscl_data->mpc_size.height;
+
+			/* DSCL mode - software state that programs SCL_MODE registers */
+			state->dpp[i].dscl_mode = dscl_data->dscl_mode;
+
+			/* Scaler ratios - software state that programs scale ratio registers (use actual programmed ratios) */
+			state->dpp[i].horz_ratio_int = dscl_data->ratios.h_scale_ratio >> 19; // Extract integer part from programmed ratio
+			state->dpp[i].vert_ratio_int = dscl_data->ratios.v_scale_ratio >> 19; // Extract integer part from programmed ratio
+
+			/* Basic scaler taps - software state that programs tap control registers (use actual programmed taps) */
+			state->dpp[i].h_taps = dscl_data->taps.h_taps + 1; // dscl_prog_data.taps stores (taps - 1), so add 1 back
+			state->dpp[i].v_taps = dscl_data->taps.v_taps + 1; // dscl_prog_data.taps stores (taps - 1), so add 1 back
+		}
+	}
+
+	/* Capture essential clock state for underflow analysis */
+	if (dc->clk_mgr && dc->clk_mgr->clks.dispclk_khz > 0) {
+		/* Core display clocks affecting bandwidth and timing */
+		state->dccg.dispclk_khz = dc->clk_mgr->clks.dispclk_khz;
+
+		/* Per-pipe clock configuration - only capture what's essential */
+		for (i = 0; i < MAX_PIPES && i < dc->res_pool->pipe_count; i++) {
+			struct pipe_ctx *pipe_ctx = &res_ctx->pipe_ctx[i];
+			if (pipe_ctx->stream) {
+				/* Essential clocks that directly affect underflow risk */
+				state->dccg.dppclk_khz[i] = dc->clk_mgr->clks.dppclk_khz;
+				state->dccg.pixclk_khz[i] = pipe_ctx->stream->timing.pix_clk_100hz / 10;
+				state->dccg.dppclk_enable[i] = 1;
+
+				/* DP stream clock only for DP signals */
+				if (pipe_ctx->stream->signal == SIGNAL_TYPE_DISPLAY_PORT ||
+						pipe_ctx->stream->signal == SIGNAL_TYPE_DISPLAY_PORT_MST) {
+					state->dccg.dpstreamclk_enable[i] = 1;
+				} else {
+					state->dccg.dpstreamclk_enable[i] = 0;
+				}
+			} else {
+				/* Inactive pipe - no clocks */
+				state->dccg.dppclk_khz[i] = 0;
+				state->dccg.pixclk_khz[i] = 0;
+				state->dccg.dppclk_enable[i] = 0;
+				if (i < 4) {
+					state->dccg.dpstreamclk_enable[i] = 0;
+				}
+			}
+		}
+
+		/* DSC clock state - only when actually using DSC */
+		for (i = 0; i < MAX_PIPES; i++) {
+			struct pipe_ctx *pipe_ctx = (i < dc->res_pool->pipe_count) ? &res_ctx->pipe_ctx[i] : NULL;
+			if (pipe_ctx && pipe_ctx->stream && pipe_ctx->stream->timing.dsc_cfg.num_slices_h > 0) {
+				state->dccg.dscclk_khz[i] = 400000; /* Typical DSC clock frequency */
+			} else {
+				state->dccg.dscclk_khz[i] = 0;
+			}
+		}
+
+		/* SYMCLK32 LE Control - only the essential HPO state for underflow analysis */
+		for (i = 0; i < 2; i++) {
+			state->dccg.symclk32_le_enable[i] = 0; /* Default: disabled */
+		}
+
+	}
+
+	/* Capture essential DSC configuration for underflow analysis */
+	for (i = 0; i < MAX_PIPES && i < dc->res_pool->pipe_count; i++) {
+		struct pipe_ctx *pipe_ctx = &res_ctx->pipe_ctx[i];
+
+		if (pipe_ctx->stream && pipe_ctx->stream->timing.dsc_cfg.num_slices_h > 0) {
+			/* DSC is enabled - capture essential configuration */
+			state->dsc[i].dsc_clock_enable = 1;
+
+			/* DSC configuration affecting bandwidth and timing */
+			struct dc_dsc_config *dsc_cfg = &pipe_ctx->stream->timing.dsc_cfg;
+			state->dsc[i].dsc_num_slices_h = dsc_cfg->num_slices_h;
+			state->dsc[i].dsc_num_slices_v = dsc_cfg->num_slices_v;
+			state->dsc[i].dsc_bits_per_pixel = dsc_cfg->bits_per_pixel;
+
+			/* OPP pipe source for DSC forwarding */
+			if (pipe_ctx->stream_res.opp) {
+				state->dsc[i].dscrm_dsc_forward_enable = 1;
+				state->dsc[i].dscrm_dsc_opp_pipe_source = pipe_ctx->stream_res.opp->inst;
+			} else {
+				state->dsc[i].dscrm_dsc_forward_enable = 0;
+				state->dsc[i].dscrm_dsc_opp_pipe_source = 0;
+			}
+		} else {
+			/* DSC not enabled - clear all fields */
+			memset(&state->dsc[i], 0, sizeof(state->dsc[i]));
+		}
+	}
+
+	/* Capture MPC programming state - comprehensive register field coverage */
+	for (i = 0; i < MAX_PIPES && i < dc->res_pool->pipe_count; i++) {
+		struct pipe_ctx *pipe_ctx = &res_ctx->pipe_ctx[i];
+
+		if (pipe_ctx->plane_state && pipe_ctx->stream) {
+			struct dc_plane_state *plane_state = pipe_ctx->plane_state;
+
+			/* MPCC blending tree and mode control - capture actual blend configuration */
+			state->mpc.mpcc_mode[i] = (plane_state->blend_tf.type != TF_TYPE_BYPASS) ? 1 : 0;
+			state->mpc.mpcc_alpha_blend_mode[i] = plane_state->per_pixel_alpha ? 1 : 0;
+			state->mpc.mpcc_alpha_multiplied_mode[i] = plane_state->pre_multiplied_alpha ? 1 : 0;
+			state->mpc.mpcc_blnd_active_overlap_only[i] = 0; /* Default - no overlap restriction */
+			state->mpc.mpcc_global_alpha[i] = plane_state->global_alpha_value;
+			state->mpc.mpcc_global_gain[i] = plane_state->global_alpha ? 255 : 0;
+			state->mpc.mpcc_bg_bpc[i] = 8; /* Standard 8-bit background */
+			state->mpc.mpcc_bot_gain_mode[i] = 0; /* Standard gain mode */
+
+			/* MPCC blending tree connections - capture tree topology */
+			if (pipe_ctx->bottom_pipe) {
+				state->mpc.mpcc_bot_sel[i] = pipe_ctx->bottom_pipe->pipe_idx;
+			} else {
+				state->mpc.mpcc_bot_sel[i] = 0xF; /* No bottom connection */
+			}
+			state->mpc.mpcc_top_sel[i] = pipe_ctx->pipe_idx; /* This pipe's DPP ID */
+
+			/* MPCC output gamma control - capture gamma programming */
+			if (plane_state->gamma_correction.type != GAMMA_CS_TFM_1D && plane_state->gamma_correction.num_entries > 0) {
+				state->mpc.mpcc_ogam_mode[i] = 1; /* Gamma enabled */
+				state->mpc.mpcc_ogam_select[i] = 0; /* Bank A selection */
+				state->mpc.mpcc_ogam_pwl_disable[i] = 0; /* PWL enabled */
+			} else {
+				state->mpc.mpcc_ogam_mode[i] = 0; /* Bypass mode */
+				state->mpc.mpcc_ogam_select[i] = 0;
+				state->mpc.mpcc_ogam_pwl_disable[i] = 1; /* PWL disabled */
+			}
+
+			/* MPCC pipe assignment and operational status */
+			if (pipe_ctx->stream_res.opp) {
+				state->mpc.mpcc_opp_id[i] = pipe_ctx->stream_res.opp->inst;
+			} else {
+				state->mpc.mpcc_opp_id[i] = 0xF; /* No OPP assignment */
+			}
+
+			/* MPCC status indicators - active pipe state */
+			state->mpc.mpcc_idle[i] = 0; /* Active pipe - not idle */
+			state->mpc.mpcc_busy[i] = 1; /* Active pipe - busy processing */
+
+		} else {
+			/* Pipe not active - set disabled/idle state for all fields */
+			state->mpc.mpcc_mode[i] = 0;
+			state->mpc.mpcc_alpha_blend_mode[i] = 0;
+			state->mpc.mpcc_alpha_multiplied_mode[i] = 0;
+			state->mpc.mpcc_blnd_active_overlap_only[i] = 0;
+			state->mpc.mpcc_global_alpha[i] = 0;
+			state->mpc.mpcc_global_gain[i] = 0;
+			state->mpc.mpcc_bg_bpc[i] = 0;
+			state->mpc.mpcc_bot_gain_mode[i] = 0;
+			state->mpc.mpcc_bot_sel[i] = 0xF; /* No bottom connection */
+			state->mpc.mpcc_top_sel[i] = 0xF; /* No top connection */
+			state->mpc.mpcc_ogam_mode[i] = 0; /* Bypass */
+			state->mpc.mpcc_ogam_select[i] = 0;
+			state->mpc.mpcc_ogam_pwl_disable[i] = 1; /* PWL disabled */
+			state->mpc.mpcc_opp_id[i] = 0xF; /* No OPP assignment */
+			state->mpc.mpcc_idle[i] = 1; /* Idle */
+			state->mpc.mpcc_busy[i] = 0; /* Not busy */
+		}
+	}
+
+	/* Capture OPP programming state for each pipe - comprehensive register field coverage */
+	for (i = 0; i < MAX_PIPES && i < dc->res_pool->pipe_count; i++) {
+		struct pipe_ctx *pipe_ctx = &res_ctx->pipe_ctx[i];
+
+		if (!pipe_ctx->stream)
+			continue;
+
+		if (pipe_ctx->stream_res.opp) {
+			struct dc_crtc_timing *timing = &pipe_ctx->stream->timing;
+
+			/* OPP Pipe Control */
+			state->opp[i].opp_pipe_clock_enable = 1; /* Active pipe has clock enabled */
+
+			/* Display Pattern Generator (DPG) Control - 19 fields */
+			if (pipe_ctx->stream->test_pattern.type != DP_TEST_PATTERN_VIDEO_MODE) {
+				state->opp[i].dpg_enable = 1;
+			} else {
+				/* Video mode - DPG disabled */
+				state->opp[i].dpg_enable = 0;
+			}
+
+			/* Format Control (FMT) - 18 fields */
+			state->opp[i].fmt_pixel_encoding = timing->pixel_encoding;
+
+			/* Chroma subsampling mode based on pixel encoding */
+			if (timing->pixel_encoding == PIXEL_ENCODING_YCBCR420) {
+				state->opp[i].fmt_subsampling_mode = 1; /* 4:2:0 subsampling */
+			} else if (timing->pixel_encoding == PIXEL_ENCODING_YCBCR422) {
+				state->opp[i].fmt_subsampling_mode = 2; /* 4:2:2 subsampling */
+			} else {
+				state->opp[i].fmt_subsampling_mode = 0; /* No subsampling (4:4:4) */
+			}
+
+			state->opp[i].fmt_cbcr_bit_reduction_bypass = (timing->pixel_encoding == PIXEL_ENCODING_RGB) ? 1 : 0;
+			state->opp[i].fmt_stereosync_override = (timing->timing_3d_format != TIMING_3D_FORMAT_NONE) ? 1 : 0;
+
+			/* Dithering control based on bit depth */
+			if (timing->display_color_depth < COLOR_DEPTH_121212) {
+				state->opp[i].fmt_spatial_dither_frame_counter_max = 15; /* Typical frame counter max */
+				state->opp[i].fmt_spatial_dither_frame_counter_bit_swap = 0; /* No bit swapping */
+				state->opp[i].fmt_spatial_dither_enable = 1;
+				state->opp[i].fmt_spatial_dither_mode = 0; /* Spatial dithering mode */
+				state->opp[i].fmt_spatial_dither_depth = timing->display_color_depth;
+				state->opp[i].fmt_temporal_dither_enable = 0; /* Spatial dithering preferred */
+			} else {
+				state->opp[i].fmt_spatial_dither_frame_counter_max = 0;
+				state->opp[i].fmt_spatial_dither_frame_counter_bit_swap = 0;
+				state->opp[i].fmt_spatial_dither_enable = 0;
+				state->opp[i].fmt_spatial_dither_mode = 0;
+				state->opp[i].fmt_spatial_dither_depth = 0;
+				state->opp[i].fmt_temporal_dither_enable = 0;
+			}
+
+			/* Truncation control for bit depth reduction */
+			if (timing->display_color_depth < COLOR_DEPTH_121212) {
+				state->opp[i].fmt_truncate_enable = 1;
+				state->opp[i].fmt_truncate_depth = timing->display_color_depth;
+				state->opp[i].fmt_truncate_mode = 0; /* Round mode */
+			} else {
+				state->opp[i].fmt_truncate_enable = 0;
+				state->opp[i].fmt_truncate_depth = 0;
+				state->opp[i].fmt_truncate_mode = 0;
+			}
+
+			/* Data clamping control */
+			state->opp[i].fmt_clamp_data_enable = 1; /* Clamping typically enabled */
+			state->opp[i].fmt_clamp_color_format = timing->pixel_encoding;
+
+			/* Dynamic expansion for limited range content */
+			if (timing->pixel_encoding != PIXEL_ENCODING_RGB) {
+				state->opp[i].fmt_dynamic_exp_enable = 1; /* YCbCr typically needs expansion */
+				state->opp[i].fmt_dynamic_exp_mode = 0; /* Standard expansion */
+			} else {
+				state->opp[i].fmt_dynamic_exp_enable = 0; /* RGB typically full range */
+				state->opp[i].fmt_dynamic_exp_mode = 0;
+			}
+
+			/* Legacy field for compatibility */
+			state->opp[i].fmt_bit_depth_control = timing->display_color_depth;
+
+			/* Output Buffer (OPPBUF) Control - 6 fields */
+			state->opp[i].oppbuf_active_width = timing->h_addressable;
+			state->opp[i].oppbuf_pixel_repetition = 0; /* No pixel repetition by default */
+
+			/* Multi-Stream Output (MSO) / ODM segmentation */
+			if (pipe_ctx->next_odm_pipe) {
+				state->opp[i].oppbuf_display_segmentation = 1; /* Segmented display */
+				state->opp[i].oppbuf_overlap_pixel_num = 0; /* ODM overlap pixels */
+			} else {
+				state->opp[i].oppbuf_display_segmentation = 0; /* Single segment */
+				state->opp[i].oppbuf_overlap_pixel_num = 0;
+			}
+
+			/* 3D/Stereo control */
+			if (timing->timing_3d_format != TIMING_3D_FORMAT_NONE) {
+				state->opp[i].oppbuf_3d_vact_space1_size = 30; /* Typical stereo blanking */
+				state->opp[i].oppbuf_3d_vact_space2_size = 30;
+			} else {
+				state->opp[i].oppbuf_3d_vact_space1_size = 0;
+				state->opp[i].oppbuf_3d_vact_space2_size = 0;
+			}
+
+			/* DSC Forward Config - 3 fields */
+			if (timing->dsc_cfg.num_slices_h > 0) {
+				state->opp[i].dscrm_dsc_forward_enable = 1;
+				state->opp[i].dscrm_dsc_opp_pipe_source = pipe_ctx->stream_res.opp->inst;
+				state->opp[i].dscrm_dsc_forward_enable_status = 1; /* Status follows enable */
+			} else {
+				state->opp[i].dscrm_dsc_forward_enable = 0;
+				state->opp[i].dscrm_dsc_opp_pipe_source = 0;
+				state->opp[i].dscrm_dsc_forward_enable_status = 0;
+			}
+		} else {
+			/* No OPP resource - set all fields to disabled state */
+			memset(&state->opp[i], 0, sizeof(state->opp[i]));
+		}
+	}
+
+	/* Capture OPTC programming state for each pipe - comprehensive register field coverage */
+	for (i = 0; i < MAX_PIPES && i < dc->res_pool->pipe_count; i++) {
+		struct pipe_ctx *pipe_ctx = &res_ctx->pipe_ctx[i];
+
+		if (!pipe_ctx->stream)
+			continue;
+
+		if (pipe_ctx->stream_res.tg) {
+			struct dc_crtc_timing *timing = &pipe_ctx->stream->timing;
+
+			state->optc[i].otg_master_inst = pipe_ctx->stream_res.tg->inst;
+
+			/* OTG_CONTROL register - 5 fields */
+			state->optc[i].otg_master_enable = 1; /* Active stream */
+			state->optc[i].otg_disable_point_cntl = 0; /* Normal operation */
+			state->optc[i].otg_start_point_cntl = 0; /* Normal start */
+			state->optc[i].otg_field_number_cntl = (timing->flags.INTERLACE) ? 1 : 0;
+			state->optc[i].otg_out_mux = 0; /* Direct output */
+
+			/* OTG Horizontal Timing - 7 fields */
+			state->optc[i].otg_h_total = timing->h_total;
+			state->optc[i].otg_h_blank_start = timing->h_addressable;
+			state->optc[i].otg_h_blank_end = timing->h_total - timing->h_front_porch;
+			state->optc[i].otg_h_sync_start = timing->h_addressable + timing->h_front_porch;
+			state->optc[i].otg_h_sync_end = timing->h_addressable + timing->h_front_porch + timing->h_sync_width;
+			state->optc[i].otg_h_sync_polarity = timing->flags.HSYNC_POSITIVE_POLARITY ? 0 : 1;
+			state->optc[i].otg_h_timing_div_mode = (pipe_ctx->next_odm_pipe) ? 1 : 0; /* ODM divide mode */
+
+			/* OTG Vertical Timing - 7 fields */
+			state->optc[i].otg_v_total = timing->v_total;
+			state->optc[i].otg_v_blank_start = timing->v_addressable;
+			state->optc[i].otg_v_blank_end = timing->v_total - timing->v_front_porch;
+			state->optc[i].otg_v_sync_start = timing->v_addressable + timing->v_front_porch;
+			state->optc[i].otg_v_sync_end = timing->v_addressable + timing->v_front_porch + timing->v_sync_width;
+			state->optc[i].otg_v_sync_polarity = timing->flags.VSYNC_POSITIVE_POLARITY ? 0 : 1;
+			state->optc[i].otg_v_sync_mode = 0; /* Normal sync mode */
+
+			/* Initialize remaining core fields with appropriate defaults */
+			// TODO: Update logic for accurate vtotal min/max
+			state->optc[i].otg_v_total_max = timing->v_total + 100; /* Typical DRR range */
+			state->optc[i].otg_v_total_min = timing->v_total - 50;
+			state->optc[i].otg_v_total_mid = timing->v_total;
+
+			/* ODM configuration */
+			// TODO: Update logic to have complete ODM mappings (e.g. 3:1 and 4:1) stored in single pipe
+			if (pipe_ctx->next_odm_pipe) {
+				state->optc[i].optc_seg0_src_sel = pipe_ctx->stream_res.opp ? pipe_ctx->stream_res.opp->inst : 0;
+				state->optc[i].optc_seg1_src_sel = pipe_ctx->next_odm_pipe->stream_res.opp ? pipe_ctx->next_odm_pipe->stream_res.opp->inst : 0;
+				state->optc[i].optc_num_of_input_segment = 1; /* 2 segments - 1 */
+			} else {
+				state->optc[i].optc_seg0_src_sel = pipe_ctx->stream_res.opp ? pipe_ctx->stream_res.opp->inst : 0;
+				state->optc[i].optc_seg1_src_sel = 0;
+				state->optc[i].optc_num_of_input_segment = 0; /* Single segment */
+			}
+
+			/* DSC configuration */
+			if (timing->dsc_cfg.num_slices_h > 0) {
+				state->optc[i].optc_dsc_mode = 1; /* DSC enabled */
+				state->optc[i].optc_dsc_bytes_per_pixel = timing->dsc_cfg.bits_per_pixel / 16; /* Convert to bytes */
+				state->optc[i].optc_dsc_slice_width = timing->h_addressable / timing->dsc_cfg.num_slices_h;
+			} else {
+				state->optc[i].optc_dsc_mode = 0;
+				state->optc[i].optc_dsc_bytes_per_pixel = 0;
+				state->optc[i].optc_dsc_slice_width = 0;
+			}
+
+			/* Essential control fields */
+			state->optc[i].otg_stereo_enable = (timing->timing_3d_format != TIMING_3D_FORMAT_NONE) ? 1 : 0;
+			state->optc[i].otg_interlace_enable = timing->flags.INTERLACE ? 1 : 0;
+			state->optc[i].otg_clock_enable = 1; /* OTG clock enabled */
+			state->optc[i].vtg0_enable = 1; /* VTG enabled for timing generation */
+
+			/* Initialize other key fields to defaults */
+			state->optc[i].optc_input_pix_clk_en = 1;
+			state->optc[i].optc_segment_width = (pipe_ctx->next_odm_pipe) ? (timing->h_addressable / 2) : timing->h_addressable;
+			state->optc[i].otg_vready_offset = 1;
+			state->optc[i].otg_vstartup_start = timing->v_addressable + 10;
+			state->optc[i].otg_vupdate_offset = 0;
+			state->optc[i].otg_vupdate_width = 5;
+		} else {
+			/* No timing generator resource - initialize all fields to 0 */
+			memset(&state->optc[i], 0, sizeof(state->optc[i]));
+		}
+	}
+
+	state->state_valid = true;
+	return true;
+}
+
+void dc_log_preos_dmcub_info(const struct dc *dc)
+{
+	dc_dmub_srv_log_preos_dmcub_info(dc->ctx->dmub_srv);
+}
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_debug.c b/drivers/gpu/drm/amd/display/dc/core/dc_debug.c
index 21be2a684393..bbce751b485f 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_debug.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_debug.c
@@ -36,6 +36,8 @@
 
 #include "resource.h"
 
+#define DC_LOGGER \
+	dc->ctx->logger
 #define DC_LOGGER_INIT(logger)
 
 
@@ -44,136 +46,11 @@
 			DC_LOG_IF_TRACE(__VA_ARGS__); \
 } while (0)
 
-#define TIMING_TRACE(...) do {\
-	if (dc->debug.timing_trace) \
-		DC_LOG_SYNC(__VA_ARGS__); \
-} while (0)
-
 #define CLOCK_TRACE(...) do {\
 	if (dc->debug.clock_trace) \
 		DC_LOG_BANDWIDTH_CALCS(__VA_ARGS__); \
 } while (0)
 
-void pre_surface_trace(
-		struct dc *dc,
-		const struct dc_plane_state *const *plane_states,
-		int surface_count)
-{
-	int i;
-	DC_LOGGER_INIT(dc->ctx->logger);
-
-	for (i = 0; i < surface_count; i++) {
-		const struct dc_plane_state *plane_state = plane_states[i];
-
-		SURFACE_TRACE("Planes %d:\n", i);
-
-		SURFACE_TRACE(
-				"plane_state->visible = %d;\n"
-				"plane_state->flip_immediate = %d;\n"
-				"plane_state->address.type = %d;\n"
-				"plane_state->address.grph.addr.quad_part = 0x%llX;\n"
-				"plane_state->address.grph.meta_addr.quad_part = 0x%llX;\n"
-				"plane_state->scaling_quality.h_taps = %d;\n"
-				"plane_state->scaling_quality.v_taps = %d;\n"
-				"plane_state->scaling_quality.h_taps_c = %d;\n"
-				"plane_state->scaling_quality.v_taps_c = %d;\n",
-				plane_state->visible,
-				plane_state->flip_immediate,
-				plane_state->address.type,
-				plane_state->address.grph.addr.quad_part,
-				plane_state->address.grph.meta_addr.quad_part,
-				plane_state->scaling_quality.h_taps,
-				plane_state->scaling_quality.v_taps,
-				plane_state->scaling_quality.h_taps_c,
-				plane_state->scaling_quality.v_taps_c);
-
-		SURFACE_TRACE(
-				"plane_state->src_rect.x = %d;\n"
-				"plane_state->src_rect.y = %d;\n"
-				"plane_state->src_rect.width = %d;\n"
-				"plane_state->src_rect.height = %d;\n"
-				"plane_state->dst_rect.x = %d;\n"
-				"plane_state->dst_rect.y = %d;\n"
-				"plane_state->dst_rect.width = %d;\n"
-				"plane_state->dst_rect.height = %d;\n"
-				"plane_state->clip_rect.x = %d;\n"
-				"plane_state->clip_rect.y = %d;\n"
-				"plane_state->clip_rect.width = %d;\n"
-				"plane_state->clip_rect.height = %d;\n",
-				plane_state->src_rect.x,
-				plane_state->src_rect.y,
-				plane_state->src_rect.width,
-				plane_state->src_rect.height,
-				plane_state->dst_rect.x,
-				plane_state->dst_rect.y,
-				plane_state->dst_rect.width,
-				plane_state->dst_rect.height,
-				plane_state->clip_rect.x,
-				plane_state->clip_rect.y,
-				plane_state->clip_rect.width,
-				plane_state->clip_rect.height);
-
-		SURFACE_TRACE(
-				"plane_state->plane_size.surface_size.x = %d;\n"
-				"plane_state->plane_size.surface_size.y = %d;\n"
-				"plane_state->plane_size.surface_size.width = %d;\n"
-				"plane_state->plane_size.surface_size.height = %d;\n"
-				"plane_state->plane_size.surface_pitch = %d;\n",
-				plane_state->plane_size.surface_size.x,
-				plane_state->plane_size.surface_size.y,
-				plane_state->plane_size.surface_size.width,
-				plane_state->plane_size.surface_size.height,
-				plane_state->plane_size.surface_pitch);
-
-
-		SURFACE_TRACE(
-				"plane_state->tiling_info.gfx8.num_banks = %d;\n"
-				"plane_state->tiling_info.gfx8.bank_width = %d;\n"
-				"plane_state->tiling_info.gfx8.bank_width_c = %d;\n"
-				"plane_state->tiling_info.gfx8.bank_height = %d;\n"
-				"plane_state->tiling_info.gfx8.bank_height_c = %d;\n"
-				"plane_state->tiling_info.gfx8.tile_aspect = %d;\n"
-				"plane_state->tiling_info.gfx8.tile_aspect_c = %d;\n"
-				"plane_state->tiling_info.gfx8.tile_split = %d;\n"
-				"plane_state->tiling_info.gfx8.tile_split_c = %d;\n"
-				"plane_state->tiling_info.gfx8.tile_mode = %d;\n"
-				"plane_state->tiling_info.gfx8.tile_mode_c = %d;\n",
-				plane_state->tiling_info.gfx8.num_banks,
-				plane_state->tiling_info.gfx8.bank_width,
-				plane_state->tiling_info.gfx8.bank_width_c,
-				plane_state->tiling_info.gfx8.bank_height,
-				plane_state->tiling_info.gfx8.bank_height_c,
-				plane_state->tiling_info.gfx8.tile_aspect,
-				plane_state->tiling_info.gfx8.tile_aspect_c,
-				plane_state->tiling_info.gfx8.tile_split,
-				plane_state->tiling_info.gfx8.tile_split_c,
-				plane_state->tiling_info.gfx8.tile_mode,
-				plane_state->tiling_info.gfx8.tile_mode_c);
-
-		SURFACE_TRACE(
-				"plane_state->tiling_info.gfx8.pipe_config = %d;\n"
-				"plane_state->tiling_info.gfx8.array_mode = %d;\n"
-				"plane_state->color_space = %d;\n"
-				"plane_state->dcc.enable = %d;\n"
-				"plane_state->format = %d;\n"
-				"plane_state->rotation = %d;\n"
-				"plane_state->stereo_format = %d;\n",
-				plane_state->tiling_info.gfx8.pipe_config,
-				plane_state->tiling_info.gfx8.array_mode,
-				plane_state->color_space,
-				plane_state->dcc.enable,
-				plane_state->format,
-				plane_state->rotation,
-				plane_state->stereo_format);
-
-		SURFACE_TRACE("plane_state->tiling_info.gfx9.swizzle = %d;\n",
-				plane_state->tiling_info.gfx9.swizzle);
-
-		SURFACE_TRACE("\n");
-	}
-	SURFACE_TRACE("\n");
-}
-
 void update_surface_trace(
 		struct dc *dc,
 		const struct dc_surface_update *updates,
@@ -304,48 +181,10 @@ void post_surface_trace(struct dc *dc)
 
 }
 
-void context_timing_trace(
-		struct dc *dc,
-		struct resource_context *res_ctx)
-{
-	int i;
-	int h_pos[MAX_PIPES] = {0}, v_pos[MAX_PIPES] = {0};
-	struct crtc_position position;
-	unsigned int underlay_idx = dc->res_pool->underlay_pipe_index;
-	DC_LOGGER_INIT(dc->ctx->logger);
-
-
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		struct pipe_ctx *pipe_ctx = &res_ctx->pipe_ctx[i];
-		/* get_position() returns CRTC vertical/horizontal counter
-		 * hence not applicable for underlay pipe
-		 */
-		if (pipe_ctx->stream == NULL || pipe_ctx->pipe_idx == underlay_idx)
-			continue;
-
-		pipe_ctx->stream_res.tg->funcs->get_position(pipe_ctx->stream_res.tg, &position);
-		h_pos[i] = position.horizontal_count;
-		v_pos[i] = position.vertical_count;
-	}
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		struct pipe_ctx *pipe_ctx = &res_ctx->pipe_ctx[i];
-
-		if (pipe_ctx->stream == NULL || pipe_ctx->pipe_idx == underlay_idx)
-			continue;
-
-		TIMING_TRACE("OTG_%d   H_tot:%d  V_tot:%d   H_pos:%d  V_pos:%d\n",
-				pipe_ctx->stream_res.tg->inst,
-				pipe_ctx->stream->timing.h_total,
-				pipe_ctx->stream->timing.v_total,
-				h_pos[i], v_pos[i]);
-	}
-}
-
 void context_clock_trace(
 		struct dc *dc,
 		struct dc_state *context)
 {
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	DC_LOGGER_INIT(dc->ctx->logger);
 	CLOCK_TRACE("Current: dispclk_khz:%d  max_dppclk_khz:%d  dcfclk_khz:%d\n"
 			"dcfclk_deep_sleep_khz:%d  fclk_khz:%d  socclk_khz:%d\n",
@@ -363,7 +202,6 @@ void context_clock_trace(
 			context->bw_ctx.bw.dcn.clk.dcfclk_deep_sleep_khz,
 			context->bw_ctx.bw.dcn.clk.fclk_khz,
 			context->bw_ctx.bw.dcn.clk.socclk_khz);
-#endif
 }
 
 /**
@@ -422,9 +260,59 @@ char *dc_status_to_str(enum dc_status status)
 		return "The operation is not supported.";
 	case DC_UNSUPPORTED_VALUE:
 		return "The value specified is not supported.";
+	case DC_NO_LINK_ENC_RESOURCE:
+		return "No link encoder resource";
+	case DC_FAIL_DP_PAYLOAD_ALLOCATION:
+		return "Fail dp payload allocation";
+	case DC_FAIL_DP_LINK_BANDWIDTH:
+		return "Insufficient DP link bandwidth";
+	case DC_FAIL_HW_CURSOR_SUPPORT:
+		return "HW Cursor not supported";
+	case DC_FAIL_DP_TUNNEL_BW_VALIDATE:
+		return "Fail DP Tunnel BW validation";
 	case DC_ERROR_UNEXPECTED:
 		return "Unexpected error";
 	}
 
 	return "Unexpected status error";
 }
+
+char *dc_pixel_encoding_to_str(enum dc_pixel_encoding pixel_encoding)
+{
+	switch (pixel_encoding) {
+	case PIXEL_ENCODING_RGB:
+		return "RGB";
+	case PIXEL_ENCODING_YCBCR422:
+		return "YUV422";
+	case PIXEL_ENCODING_YCBCR444:
+		return "YUV444";
+	case PIXEL_ENCODING_YCBCR420:
+		return "YUV420";
+	default:
+		return "Unknown";
+	}
+}
+
+char *dc_color_depth_to_str(enum dc_color_depth color_depth)
+{
+	switch (color_depth) {
+	case COLOR_DEPTH_666:
+		return "6-bpc";
+	case COLOR_DEPTH_888:
+		return "8-bpc";
+	case COLOR_DEPTH_101010:
+		return "10-bpc";
+	case COLOR_DEPTH_121212:
+		return "12-bpc";
+	case COLOR_DEPTH_141414:
+		return "14-bpc";
+	case COLOR_DEPTH_161616:
+		return "16-bpc";
+	case COLOR_DEPTH_999:
+		return "9-bpc";
+	case COLOR_DEPTH_111111:
+		return "11-bpc";
+	default:
+		return "Unknown";
+	}
+}
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/core/dc_hw_sequencer.c
index 9039fb134db5..e2763b60482a 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_hw_sequencer.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_hw_sequencer.c
@@ -23,14 +23,25 @@
  *
  */
 
-#include <linux/delay.h>
-
 #include "dm_services.h"
 #include "core_types.h"
 #include "timing_generator.h"
 #include "hw_sequencer.h"
+#include "hw_sequencer_private.h"
+#include "basics/dc_common.h"
+#include "resource.h"
+#include "dc_dmub_srv.h"
+#include "dc_state_priv.h"
+#include "opp.h"
+#include "dsc.h"
+#include "dchubbub.h"
+#include "dccg.h"
+#include "abm.h"
+#include "dcn10/dcn10_hubbub.h"
+#include "dce/dmub_hw_lock_mgr.h"
 
 #define NUM_ELEMENTS(a) (sizeof(a) / sizeof((a)[0]))
+#define MAX_NUM_MCACHE 8
 
 /* used as index in array of black_color_format */
 enum black_color_format {
@@ -75,28 +86,38 @@ struct out_csc_color_matrix_type {
 
 static const struct out_csc_color_matrix_type output_csc_matrix[] = {
 	{ COLOR_SPACE_RGB_TYPE,
-		{ 0x2000, 0, 0, 0, 0, 0x2000, 0, 0, 0, 0, 0x2000, 0} },
+		{ 0x2000, 0,      0,      0,
+		  0,      0x2000, 0,      0,
+		  0,      0,      0x2000, 0} },
 	{ COLOR_SPACE_RGB_LIMITED_TYPE,
-		{ 0x1B67, 0, 0, 0x201, 0, 0x1B67, 0, 0x201, 0, 0, 0x1B67, 0x201} },
+		{ 0x1B67, 0,      0,      0x201,
+		  0,      0x1B67, 0,      0x201,
+		  0,      0,      0x1B67, 0x201} },
 	{ COLOR_SPACE_YCBCR601_TYPE,
-		{ 0xE04, 0xF444, 0xFDB9, 0x1004, 0x831, 0x1016, 0x320, 0x201, 0xFB45,
-				0xF6B7, 0xE04, 0x1004} },
+		{ 0xE04,  0xF444, 0xFDB9, 0x1004,
+		  0x831,  0x1016, 0x320,  0x201,
+		  0xFB45, 0xF6B7, 0xE04,  0x1004} },
 	{ COLOR_SPACE_YCBCR709_TYPE,
-		{ 0xE04, 0xF345, 0xFEB7, 0x1004, 0x5D3, 0x1399, 0x1FA,
-				0x201, 0xFCCA, 0xF533, 0xE04, 0x1004} },
+		{ 0xE04,  0xF345, 0xFEB7, 0x1004,
+		  0x5D3,  0x1399, 0x1FA,  0x201,
+		  0xFCCA, 0xF533, 0xE04,  0x1004} },
 	/* TODO: correct values below */
 	{ COLOR_SPACE_YCBCR601_LIMITED_TYPE,
-		{ 0xE00, 0xF447, 0xFDB9, 0x1000, 0x991,
-				0x12C9, 0x3A6, 0x200, 0xFB47, 0xF6B9, 0xE00, 0x1000} },
+		{ 0xE00,  0xF447, 0xFDB9, 0x1000,
+		  0x991,  0x12C9, 0x3A6,  0x200,
+		  0xFB47, 0xF6B9, 0xE00,  0x1000} },
 	{ COLOR_SPACE_YCBCR709_LIMITED_TYPE,
-		{ 0xE00, 0xF349, 0xFEB7, 0x1000, 0x6CE, 0x16E3,
-				0x24F, 0x200, 0xFCCB, 0xF535, 0xE00, 0x1000} },
+		{ 0xE00, 0xF349, 0xFEB7, 0x1000,
+		  0x6CE, 0x16E3, 0x24F,  0x200,
+		  0xFCCB, 0xF535, 0xE00, 0x1000} },
 	{ COLOR_SPACE_YCBCR2020_TYPE,
-		{ 0x1000, 0xF149, 0xFEB7, 0x0000, 0x0868, 0x15B2,
-				0x01E6, 0x0000, 0xFB88, 0xF478, 0x1000, 0x0000} },
+		{ 0x1000, 0xF149, 0xFEB7, 0x1004,
+		  0x0868, 0x15B2, 0x01E6, 0x201,
+		  0xFB88, 0xF478, 0x1000, 0x1004} },
 	{ COLOR_SPACE_YCBCR709_BLACK_TYPE,
-		{ 0x0000, 0x0000, 0x0000, 0x1000, 0x0000, 0x0000,
-				0x0000, 0x0200, 0x0000, 0x0000, 0x0000, 0x1000} },
+		{ 0x0000, 0x0000, 0x0000, 0x1000,
+		  0x0000, 0x0000, 0x0000, 0x0200,
+		  0x0000, 0x0000, 0x0000, 0x1000} },
 };
 
 static bool is_rgb_type(
@@ -163,7 +184,7 @@ static bool is_ycbcr2020_type(
 {
 	bool ret = false;
 
-	if (color_space == COLOR_SPACE_2020_YCBCR)
+	if (color_space == COLOR_SPACE_2020_YCBCR_LIMITED || color_space == COLOR_SPACE_2020_YCBCR_FULL)
 		ret = true;
 	return ret;
 }
@@ -177,6 +198,7 @@ static bool is_ycbcr709_limited_type(
 		ret = true;
 	return ret;
 }
+
 static enum dc_color_space_type get_color_space_type(enum dc_color_space color_space)
 {
 	enum dc_color_space_type type = COLOR_SPACE_RGB_TYPE;
@@ -233,7 +255,8 @@ void color_space_to_black_color(
 	case COLOR_SPACE_YCBCR709_BLACK:
 	case COLOR_SPACE_YCBCR601_LIMITED:
 	case COLOR_SPACE_YCBCR709_LIMITED:
-	case COLOR_SPACE_2020_YCBCR:
+	case COLOR_SPACE_2020_YCBCR_LIMITED:
+	case COLOR_SPACE_2020_YCBCR_FULL:
 		*black_color = black_color_format[BLACK_COLOR_FORMAT_YUV_CV];
 		break;
 
@@ -242,7 +265,7 @@ void color_space_to_black_color(
 			black_color_format[BLACK_COLOR_FORMAT_RGB_LIMITED];
 		break;
 
-	/**
+	/*
 	 * Remove default and add case for all color space
 	 * so when we forget to add new color space
 	 * compiler will give a warning
@@ -298,11 +321,11 @@ void get_mpctree_visual_confirm_color(
 {
 	const struct tg_color pipe_colors[6] = {
 			{MAX_TG_COLOR_VALUE, 0, 0}, /* red */
-			{MAX_TG_COLOR_VALUE, MAX_TG_COLOR_VALUE / 4, 0}, /* orange */
 			{MAX_TG_COLOR_VALUE, MAX_TG_COLOR_VALUE, 0}, /* yellow */
 			{0, MAX_TG_COLOR_VALUE, 0}, /* green */
+			{0, MAX_TG_COLOR_VALUE, MAX_TG_COLOR_VALUE}, /* cyan */
 			{0, 0, MAX_TG_COLOR_VALUE}, /* blue */
-			{MAX_TG_COLOR_VALUE / 2, 0, MAX_TG_COLOR_VALUE / 2}, /* purple */
+			{MAX_TG_COLOR_VALUE, 0, MAX_TG_COLOR_VALUE}, /* magenta */
 	};
 
 	struct pipe_ctx *top_pipe = pipe_ctx;
@@ -368,6 +391,7 @@ void get_hdr_visual_confirm_color(
 		struct tg_color *color)
 {
 	uint32_t color_value = MAX_TG_COLOR_VALUE;
+	bool is_sdr = false;
 
 	/* Determine the overscan color based on the top-most (desktop) plane's context */
 	struct pipe_ctx *top_pipe_ctx  = pipe_ctx;
@@ -377,33 +401,1657 @@ void get_hdr_visual_confirm_color(
 
 	switch (top_pipe_ctx->plane_res.scl_data.format) {
 	case PIXEL_FORMAT_ARGB2101010:
-		if (top_pipe_ctx->stream->out_transfer_func->tf == TRANSFER_FUNCTION_PQ) {
+		if (top_pipe_ctx->stream->out_transfer_func.tf == TRANSFER_FUNCTION_PQ) {
 			/* HDR10, ARGB2101010 - set border color to red */
 			color->color_r_cr = color_value;
-		} else if (top_pipe_ctx->stream->out_transfer_func->tf == TRANSFER_FUNCTION_GAMMA22) {
+		} else if (top_pipe_ctx->stream->out_transfer_func.tf == TRANSFER_FUNCTION_GAMMA22) {
 			/* FreeSync 2 ARGB2101010 - set border color to pink */
 			color->color_r_cr = color_value;
 			color->color_b_cb = color_value;
-		}
+		} else
+			is_sdr = true;
 		break;
 	case PIXEL_FORMAT_FP16:
-		if (top_pipe_ctx->stream->out_transfer_func->tf == TRANSFER_FUNCTION_PQ) {
+		if (top_pipe_ctx->stream->out_transfer_func.tf == TRANSFER_FUNCTION_PQ) {
 			/* HDR10, FP16 - set border color to blue */
 			color->color_b_cb = color_value;
-		} else if (top_pipe_ctx->stream->out_transfer_func->tf == TRANSFER_FUNCTION_GAMMA22) {
+		} else if (top_pipe_ctx->stream->out_transfer_func.tf == TRANSFER_FUNCTION_GAMMA22) {
 			/* FreeSync 2 HDR - set border color to green */
 			color->color_g_y = color_value;
-		}
+		} else
+			is_sdr = true;
 		break;
 	default:
+		is_sdr = true;
+		break;
+	}
+
+	if (is_sdr) {
 		/* SDR - set border color to Gray */
 		color->color_r_cr = color_value/2;
 		color->color_b_cb = color_value/2;
 		color->color_g_y = color_value/2;
-		break;
 	}
 }
 
+/* Visual Confirm color definition for Smart Mux */
+void get_smartmux_visual_confirm_color(
+	struct dc *dc,
+	struct tg_color *color)
+{
+	uint32_t color_value = MAX_TG_COLOR_VALUE;
+
+	const struct tg_color sm_ver_colors[5] = {
+			{0, 0, 0},					/* SMUX_MUXCONTROL_UNSUPPORTED - Black */
+			{0, MAX_TG_COLOR_VALUE, 0},			/* SMUX_MUXCONTROL_v10 - Green */
+			{0, MAX_TG_COLOR_VALUE, MAX_TG_COLOR_VALUE},	/* SMUX_MUXCONTROL_v15 - Cyan */
+			{MAX_TG_COLOR_VALUE, MAX_TG_COLOR_VALUE, 0}, 	/* SMUX_MUXCONTROL_MDM - Yellow */
+			{MAX_TG_COLOR_VALUE, 0, MAX_TG_COLOR_VALUE}, 	/* SMUX_MUXCONTROL_vUNKNOWN - Magenta*/
+	};
+
+	if (dc->caps.is_apu) {
+		/* APU driving the eDP */
+		*color = sm_ver_colors[dc->config.smart_mux_version];
+	} else {
+		/* dGPU driving the eDP - red */
+		color->color_r_cr = color_value;
+		color->color_g_y = 0;
+		color->color_b_cb = 0;
+	}
+}
+
+/* Visual Confirm color definition for VABC */
+void get_vabc_visual_confirm_color(
+	struct pipe_ctx *pipe_ctx,
+	struct tg_color *color)
+{
+	uint32_t color_value = MAX_TG_COLOR_VALUE;
+	struct dc_link *edp_link = NULL;
+
+	if (pipe_ctx && pipe_ctx->stream && pipe_ctx->stream->link) {
+		if (pipe_ctx->stream->link->connector_signal == SIGNAL_TYPE_EDP)
+			edp_link = pipe_ctx->stream->link;
+	}
+
+	if (edp_link) {
+		switch (edp_link->backlight_control_type) {
+		case BACKLIGHT_CONTROL_PWM:
+			color->color_r_cr = color_value;
+			color->color_g_y = 0;
+			color->color_b_cb = 0;
+			break;
+		case BACKLIGHT_CONTROL_AMD_AUX:
+			color->color_r_cr = 0;
+			color->color_g_y = color_value;
+			color->color_b_cb = 0;
+			break;
+		case BACKLIGHT_CONTROL_VESA_AUX:
+			color->color_r_cr = 0;
+			color->color_g_y = 0;
+			color->color_b_cb = color_value;
+			break;
+		}
+	} else {
+		color->color_r_cr = 0;
+		color->color_g_y = 0;
+		color->color_b_cb = 0;
+	}
+}
+
+void get_subvp_visual_confirm_color(
+		struct pipe_ctx *pipe_ctx,
+		struct tg_color *color)
+{
+	uint32_t color_value = MAX_TG_COLOR_VALUE;
+	if (pipe_ctx) {
+		switch (pipe_ctx->p_state_type) {
+		case P_STATE_SUB_VP:
+			color->color_r_cr = color_value;
+			color->color_g_y  = 0;
+			color->color_b_cb = 0;
+			break;
+		case P_STATE_DRR_SUB_VP:
+			color->color_r_cr = 0;
+			color->color_g_y  = color_value;
+			color->color_b_cb = 0;
+			break;
+		case P_STATE_V_BLANK_SUB_VP:
+			color->color_r_cr = 0;
+			color->color_g_y  = 0;
+			color->color_b_cb = color_value;
+			break;
+		default:
+			break;
+		}
+	}
+}
+
+void get_mclk_switch_visual_confirm_color(
+		struct pipe_ctx *pipe_ctx,
+		struct tg_color *color)
+{
+	uint32_t color_value = MAX_TG_COLOR_VALUE;
+
+	if (pipe_ctx) {
+		switch (pipe_ctx->p_state_type) {
+		case P_STATE_V_BLANK:
+			color->color_r_cr = color_value;
+			color->color_g_y = color_value;
+			color->color_b_cb = 0;
+			break;
+		case P_STATE_FPO:
+			color->color_r_cr = 0;
+			color->color_g_y  = color_value;
+			color->color_b_cb = color_value;
+			break;
+		case P_STATE_V_ACTIVE:
+			color->color_r_cr = color_value;
+			color->color_g_y  = 0;
+			color->color_b_cb = color_value;
+			break;
+		case P_STATE_SUB_VP:
+			color->color_r_cr = color_value;
+			color->color_g_y  = 0;
+			color->color_b_cb = 0;
+			break;
+		case P_STATE_DRR_SUB_VP:
+			color->color_r_cr = 0;
+			color->color_g_y  = color_value;
+			color->color_b_cb = 0;
+			break;
+		case P_STATE_V_BLANK_SUB_VP:
+			color->color_r_cr = 0;
+			color->color_g_y  = 0;
+			color->color_b_cb = color_value;
+			break;
+		default:
+			break;
+		}
+	}
+}
+
+void get_cursor_visual_confirm_color(
+		struct pipe_ctx *pipe_ctx,
+		struct tg_color *color)
+{
+	uint32_t color_value = MAX_TG_COLOR_VALUE;
+
+	if (pipe_ctx->stream && pipe_ctx->stream->cursor_position.enable) {
+		color->color_r_cr = color_value;
+		color->color_g_y = 0;
+		color->color_b_cb = 0;
+	} else {
+		color->color_r_cr = 0;
+		color->color_g_y = 0;
+		color->color_b_cb = color_value;
+	}
+}
+
+void get_dcc_visual_confirm_color(
+	struct dc *dc,
+	struct pipe_ctx *pipe_ctx,
+	struct tg_color *color)
+{
+	const uint32_t MCACHE_ID_UNASSIGNED = 0xF;
+
+	if (!pipe_ctx->plane_state->dcc.enable) {
+		color->color_r_cr = 0; /* black - DCC disabled */
+		color->color_g_y = 0;
+		color->color_b_cb = 0;
+		return;
+	}
+
+	if (dc->ctx->dce_version < DCN_VERSION_4_01) {
+		color->color_r_cr = MAX_TG_COLOR_VALUE; /* red - DCC enabled */
+		color->color_g_y = 0;
+		color->color_b_cb = 0;
+		return;
+	}
+
+	uint32_t first_id = pipe_ctx->mcache_regs.main.p0.mcache_id_first;
+	uint32_t second_id = pipe_ctx->mcache_regs.main.p0.mcache_id_second;
+
+	if (first_id != MCACHE_ID_UNASSIGNED && second_id != MCACHE_ID_UNASSIGNED && first_id != second_id) {
+		color->color_r_cr = MAX_TG_COLOR_VALUE/2; /* grey - 2 mcache */
+		color->color_g_y = MAX_TG_COLOR_VALUE/2;
+		color->color_b_cb = MAX_TG_COLOR_VALUE/2;
+	}
+
+	else if (first_id != MCACHE_ID_UNASSIGNED || second_id != MCACHE_ID_UNASSIGNED) {
+		const struct tg_color id_colors[MAX_NUM_MCACHE] = {
+		{0, MAX_TG_COLOR_VALUE, 0}, /* green */
+		{0, 0, MAX_TG_COLOR_VALUE}, /* blue */
+		{MAX_TG_COLOR_VALUE, MAX_TG_COLOR_VALUE, 0}, /* yellow */
+		{MAX_TG_COLOR_VALUE, 0, MAX_TG_COLOR_VALUE}, /* magenta */
+		{0, MAX_TG_COLOR_VALUE, MAX_TG_COLOR_VALUE}, /* cyan */
+		{MAX_TG_COLOR_VALUE, MAX_TG_COLOR_VALUE, MAX_TG_COLOR_VALUE}, /* white */
+		{MAX_TG_COLOR_VALUE/2, 0, 0}, /* dark red */
+		{0, MAX_TG_COLOR_VALUE/2, 0}, /* dark green */
+		};
+
+		uint32_t assigned_id = (first_id != MCACHE_ID_UNASSIGNED) ? first_id : second_id;
+		*color = id_colors[assigned_id];
+	}
+}
+
+void set_p_state_switch_method(
+		struct dc *dc,
+		struct dc_state *context,
+		struct pipe_ctx *pipe_ctx)
+{
+	struct vba_vars_st *vba = &context->bw_ctx.dml.vba;
+	bool enable_subvp;
+
+	if (!dc->ctx || !dc->ctx->dmub_srv || !pipe_ctx || !vba)
+		return;
+
+	pipe_ctx->p_state_type = P_STATE_UNKNOWN;
+	if (vba->DRAMClockChangeSupport[vba->VoltageLevel][vba->maxMpcComb] !=
+			dm_dram_clock_change_unsupported) {
+		/* MCLK switching is supported */
+		if (!pipe_ctx->has_vactive_margin) {
+			/* In Vblank - yellow */
+			pipe_ctx->p_state_type = P_STATE_V_BLANK;
+
+			if (context->bw_ctx.bw.dcn.clk.fw_based_mclk_switching) {
+				/* FPO + Vblank - cyan */
+				pipe_ctx->p_state_type = P_STATE_FPO;
+			}
+		} else {
+			/* In Vactive - pink */
+			pipe_ctx->p_state_type = P_STATE_V_ACTIVE;
+		}
+
+		/* SubVP */
+		enable_subvp = false;
+
+		for (int i = 0; i < dc->res_pool->pipe_count; i++) {
+			struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
+
+			if (pipe->stream && dc_state_get_paired_subvp_stream(context, pipe->stream) &&
+					dc_state_get_pipe_subvp_type(context, pipe) == SUBVP_MAIN) {
+				/* SubVP enable - red */
+				pipe_ctx->p_state_type = P_STATE_SUB_VP;
+				enable_subvp = true;
+
+				if (pipe_ctx->stream == pipe->stream)
+					return;
+				break;
+			}
+		}
+
+		if (enable_subvp && dc_state_get_pipe_subvp_type(context, pipe_ctx) == SUBVP_NONE) {
+			if (pipe_ctx->stream->allow_freesync == 1) {
+				/* SubVP enable and DRR on - green */
+				pipe_ctx->p_state_type = P_STATE_DRR_SUB_VP;
+			} else {
+				/* SubVP enable and No DRR - blue */
+				pipe_ctx->p_state_type = P_STATE_V_BLANK_SUB_VP;
+			}
+		}
+	}
+}
+
+void set_drr_and_clear_adjust_pending(
+		struct pipe_ctx *pipe_ctx,
+		struct dc_stream_state *stream,
+		struct drr_params *params)
+{
+	/* params can be null.*/
+	if (pipe_ctx && pipe_ctx->stream_res.tg &&
+			pipe_ctx->stream_res.tg->funcs->set_drr)
+		pipe_ctx->stream_res.tg->funcs->set_drr(
+				pipe_ctx->stream_res.tg, params);
+
+	if (stream)
+		stream->adjust.timing_adjust_pending = false;
+}
+
+void get_fams2_visual_confirm_color(
+		struct dc *dc,
+		struct dc_state *context,
+		struct pipe_ctx *pipe_ctx,
+		struct tg_color *color)
+{
+	uint32_t color_value = MAX_TG_COLOR_VALUE;
+
+	if (!dc->ctx || !dc->ctx->dmub_srv || !pipe_ctx || !context || !dc->debug.fams2_config.bits.enable)
+		return;
+
+	/* driver only handles visual confirm when FAMS2 is disabled */
+	if (!dc_state_is_fams2_in_use(dc, context)) {
+		/* when FAMS2 is disabled, all pipes are grey */
+		color->color_g_y = color_value / 2;
+		color->color_b_cb = color_value / 2;
+		color->color_r_cr = color_value / 2;
+	}
+}
+
+void hwss_build_fast_sequence(struct dc *dc,
+		struct dc_dmub_cmd *dc_dmub_cmd,
+		unsigned int dmub_cmd_count,
+		struct block_sequence block_sequence[MAX_HWSS_BLOCK_SEQUENCE_SIZE],
+		unsigned int *num_steps,
+		struct pipe_ctx *pipe_ctx,
+		struct dc_stream_status *stream_status,
+		struct dc_state *context)
+{
+	struct dc_plane_state *plane = pipe_ctx->plane_state;
+	struct dc_stream_state *stream = pipe_ctx->stream;
+	struct dce_hwseq *hws = dc->hwseq;
+	struct pipe_ctx *current_pipe = NULL;
+	struct pipe_ctx *current_mpc_pipe = NULL;
+	unsigned int i = 0;
+
+	*num_steps = 0; // Initialize to 0
+
+	if (!plane || !stream)
+		return;
+
+	if (dc->hwss.wait_for_dcc_meta_propagation) {
+		block_sequence[*num_steps].params.wait_for_dcc_meta_propagation_params.dc = dc;
+		block_sequence[*num_steps].params.wait_for_dcc_meta_propagation_params.top_pipe_to_program = pipe_ctx;
+		block_sequence[*num_steps].func = HUBP_WAIT_FOR_DCC_META_PROP;
+		(*num_steps)++;
+	}
+	if (dc->hwss.subvp_pipe_control_lock_fast) {
+		block_sequence[*num_steps].params.subvp_pipe_control_lock_fast_params.dc = dc;
+		block_sequence[*num_steps].params.subvp_pipe_control_lock_fast_params.lock = true;
+		block_sequence[*num_steps].params.subvp_pipe_control_lock_fast_params.subvp_immediate_flip =
+				plane->flip_immediate && stream_status->mall_stream_config.type == SUBVP_MAIN;
+		block_sequence[*num_steps].func = DMUB_SUBVP_PIPE_CONTROL_LOCK_FAST;
+		(*num_steps)++;
+	}
+	if (dc->hwss.dmub_hw_control_lock_fast) {
+		block_sequence[*num_steps].params.dmub_hw_control_lock_fast_params.dc = dc;
+		block_sequence[*num_steps].params.dmub_hw_control_lock_fast_params.lock = true;
+		block_sequence[*num_steps].params.dmub_hw_control_lock_fast_params.is_required =
+			dc_state_is_fams2_in_use(dc, context) ||
+			dmub_hw_lock_mgr_does_link_require_lock(dc, stream->link);
+		block_sequence[*num_steps].func = DMUB_HW_CONTROL_LOCK_FAST;
+		(*num_steps)++;
+	}
+	if (dc->hwss.pipe_control_lock) {
+		block_sequence[*num_steps].params.pipe_control_lock_params.dc = dc;
+		block_sequence[*num_steps].params.pipe_control_lock_params.lock = true;
+		block_sequence[*num_steps].params.pipe_control_lock_params.pipe_ctx = pipe_ctx;
+		block_sequence[*num_steps].func = OPTC_PIPE_CONTROL_LOCK;
+		(*num_steps)++;
+	}
+
+	for (i = 0; i < dmub_cmd_count; i++) {
+		block_sequence[*num_steps].params.send_dmcub_cmd_params.ctx = dc->ctx;
+		block_sequence[*num_steps].params.send_dmcub_cmd_params.cmd = &(dc_dmub_cmd[i].dmub_cmd);
+		block_sequence[*num_steps].params.send_dmcub_cmd_params.wait_type = dc_dmub_cmd[i].wait_type;
+		block_sequence[*num_steps].func = DMUB_SEND_DMCUB_CMD;
+		(*num_steps)++;
+	}
+
+	current_pipe = pipe_ctx;
+	while (current_pipe) {
+		current_mpc_pipe = current_pipe;
+		while (current_mpc_pipe) {
+			if (current_mpc_pipe->plane_state) {
+				if (dc->hwss.set_flip_control_gsl && current_mpc_pipe->plane_state->update_flags.raw) {
+					block_sequence[*num_steps].params.set_flip_control_gsl_params.hubp = current_mpc_pipe->plane_res.hubp;
+					block_sequence[*num_steps].params.set_flip_control_gsl_params.flip_immediate = current_mpc_pipe->plane_state->flip_immediate;
+					block_sequence[*num_steps].func = HUBP_SET_FLIP_CONTROL_GSL;
+					(*num_steps)++;
+				}
+				if (dc->hwss.program_triplebuffer && dc->debug.enable_tri_buf && current_mpc_pipe->plane_state->update_flags.raw) {
+					block_sequence[*num_steps].params.program_triplebuffer_params.dc = dc;
+					block_sequence[*num_steps].params.program_triplebuffer_params.pipe_ctx = current_mpc_pipe;
+					block_sequence[*num_steps].params.program_triplebuffer_params.enableTripleBuffer = current_mpc_pipe->plane_state->triplebuffer_flips;
+					block_sequence[*num_steps].func = HUBP_PROGRAM_TRIPLEBUFFER;
+					(*num_steps)++;
+				}
+				if (dc->hwss.update_plane_addr && current_mpc_pipe->plane_state->update_flags.bits.addr_update) {
+					if (resource_is_pipe_type(current_mpc_pipe, OTG_MASTER) &&
+							stream_status->mall_stream_config.type == SUBVP_MAIN) {
+						block_sequence[*num_steps].params.subvp_save_surf_addr.dc_dmub_srv = dc->ctx->dmub_srv;
+						block_sequence[*num_steps].params.subvp_save_surf_addr.addr = &current_mpc_pipe->plane_state->address;
+						block_sequence[*num_steps].params.subvp_save_surf_addr.subvp_index = current_mpc_pipe->subvp_index;
+						block_sequence[*num_steps].func = DMUB_SUBVP_SAVE_SURF_ADDR;
+						(*num_steps)++;
+					}
+
+					block_sequence[*num_steps].params.update_plane_addr_params.dc = dc;
+					block_sequence[*num_steps].params.update_plane_addr_params.pipe_ctx = current_mpc_pipe;
+					block_sequence[*num_steps].func = HUBP_UPDATE_PLANE_ADDR;
+					(*num_steps)++;
+				}
+
+				if (hws->funcs.set_input_transfer_func && current_mpc_pipe->plane_state->update_flags.bits.gamma_change) {
+					block_sequence[*num_steps].params.set_input_transfer_func_params.dc = dc;
+					block_sequence[*num_steps].params.set_input_transfer_func_params.pipe_ctx = current_mpc_pipe;
+					block_sequence[*num_steps].params.set_input_transfer_func_params.plane_state = current_mpc_pipe->plane_state;
+					block_sequence[*num_steps].func = DPP_SET_INPUT_TRANSFER_FUNC;
+					(*num_steps)++;
+				}
+
+				if (dc->hwss.program_gamut_remap && current_mpc_pipe->plane_state->update_flags.bits.gamut_remap_change) {
+					block_sequence[*num_steps].params.program_gamut_remap_params.pipe_ctx = current_mpc_pipe;
+					block_sequence[*num_steps].func = DPP_PROGRAM_GAMUT_REMAP;
+					(*num_steps)++;
+				}
+				if (current_mpc_pipe->plane_state->update_flags.bits.input_csc_change) {
+					block_sequence[*num_steps].params.setup_dpp_params.pipe_ctx = current_mpc_pipe;
+					block_sequence[*num_steps].func = DPP_SETUP_DPP;
+					(*num_steps)++;
+				}
+				if (current_mpc_pipe->plane_state->update_flags.bits.coeff_reduction_change) {
+					block_sequence[*num_steps].params.program_bias_and_scale_params.pipe_ctx = current_mpc_pipe;
+					block_sequence[*num_steps].func = DPP_PROGRAM_BIAS_AND_SCALE;
+					(*num_steps)++;
+				}
+			}
+			if (hws->funcs.set_output_transfer_func && current_mpc_pipe->stream->update_flags.bits.out_tf) {
+				block_sequence[*num_steps].params.set_output_transfer_func_params.dc = dc;
+				block_sequence[*num_steps].params.set_output_transfer_func_params.pipe_ctx = current_mpc_pipe;
+				block_sequence[*num_steps].params.set_output_transfer_func_params.stream = current_mpc_pipe->stream;
+				block_sequence[*num_steps].func = DPP_SET_OUTPUT_TRANSFER_FUNC;
+				(*num_steps)++;
+			}
+			if (dc->debug.visual_confirm != VISUAL_CONFIRM_DISABLE &&
+				dc->hwss.update_visual_confirm_color) {
+				block_sequence[*num_steps].params.update_visual_confirm_params.dc = dc;
+				block_sequence[*num_steps].params.update_visual_confirm_params.pipe_ctx = current_mpc_pipe;
+				block_sequence[*num_steps].params.update_visual_confirm_params.mpcc_id = current_mpc_pipe->plane_res.hubp->inst;
+				block_sequence[*num_steps].func = MPC_UPDATE_VISUAL_CONFIRM;
+				(*num_steps)++;
+			}
+			if (current_mpc_pipe->stream->update_flags.bits.out_csc) {
+				block_sequence[*num_steps].params.power_on_mpc_mem_pwr_params.mpc = dc->res_pool->mpc;
+				block_sequence[*num_steps].params.power_on_mpc_mem_pwr_params.mpcc_id = current_mpc_pipe->plane_res.hubp->inst;
+				block_sequence[*num_steps].params.power_on_mpc_mem_pwr_params.power_on = true;
+				block_sequence[*num_steps].func = MPC_POWER_ON_MPC_MEM_PWR;
+				(*num_steps)++;
+
+				if (current_mpc_pipe->stream->csc_color_matrix.enable_adjustment == true) {
+					block_sequence[*num_steps].params.set_output_csc_params.mpc = dc->res_pool->mpc;
+					block_sequence[*num_steps].params.set_output_csc_params.opp_id = current_mpc_pipe->stream_res.opp->inst;
+					block_sequence[*num_steps].params.set_output_csc_params.regval = current_mpc_pipe->stream->csc_color_matrix.matrix;
+					block_sequence[*num_steps].params.set_output_csc_params.ocsc_mode = MPC_OUTPUT_CSC_COEF_A;
+					block_sequence[*num_steps].func = MPC_SET_OUTPUT_CSC;
+					(*num_steps)++;
+				} else {
+					block_sequence[*num_steps].params.set_ocsc_default_params.mpc = dc->res_pool->mpc;
+					block_sequence[*num_steps].params.set_ocsc_default_params.opp_id = current_mpc_pipe->stream_res.opp->inst;
+					block_sequence[*num_steps].params.set_ocsc_default_params.color_space = current_mpc_pipe->stream->output_color_space;
+					block_sequence[*num_steps].params.set_ocsc_default_params.ocsc_mode = MPC_OUTPUT_CSC_COEF_A;
+					block_sequence[*num_steps].func = MPC_SET_OCSC_DEFAULT;
+					(*num_steps)++;
+				}
+			}
+			current_mpc_pipe = current_mpc_pipe->bottom_pipe;
+		}
+		current_pipe = current_pipe->next_odm_pipe;
+	}
+
+	if (dc->hwss.pipe_control_lock) {
+		block_sequence[*num_steps].params.pipe_control_lock_params.dc = dc;
+		block_sequence[*num_steps].params.pipe_control_lock_params.lock = false;
+		block_sequence[*num_steps].params.pipe_control_lock_params.pipe_ctx = pipe_ctx;
+		block_sequence[*num_steps].func = OPTC_PIPE_CONTROL_LOCK;
+		(*num_steps)++;
+	}
+	if (dc->hwss.subvp_pipe_control_lock_fast) {
+		block_sequence[*num_steps].params.subvp_pipe_control_lock_fast_params.dc = dc;
+		block_sequence[*num_steps].params.subvp_pipe_control_lock_fast_params.lock = false;
+		block_sequence[*num_steps].params.subvp_pipe_control_lock_fast_params.subvp_immediate_flip =
+				plane->flip_immediate && stream_status->mall_stream_config.type == SUBVP_MAIN;
+		block_sequence[*num_steps].func = DMUB_SUBVP_PIPE_CONTROL_LOCK_FAST;
+		(*num_steps)++;
+	}
+	if (dc->hwss.dmub_hw_control_lock_fast) {
+		block_sequence[*num_steps].params.dmub_hw_control_lock_fast_params.dc = dc;
+		block_sequence[*num_steps].params.dmub_hw_control_lock_fast_params.lock = false;
+		block_sequence[*num_steps].params.dmub_hw_control_lock_fast_params.is_required = dc_state_is_fams2_in_use(dc, context);
+		block_sequence[*num_steps].func = DMUB_HW_CONTROL_LOCK_FAST;
+		(*num_steps)++;
+	}
+
+	current_pipe = pipe_ctx;
+	while (current_pipe) {
+		current_mpc_pipe = current_pipe;
+
+		while (current_mpc_pipe) {
+			if (!current_mpc_pipe->bottom_pipe && !current_mpc_pipe->next_odm_pipe &&
+					current_mpc_pipe->stream && current_mpc_pipe->plane_state &&
+					current_mpc_pipe->plane_state->update_flags.bits.addr_update &&
+					!current_mpc_pipe->plane_state->skip_manual_trigger) {
+				if (dc->hwss.program_cursor_offload_now) {
+					block_sequence[*num_steps].params.program_cursor_update_now_params.dc = dc;
+					block_sequence[*num_steps].params.program_cursor_update_now_params.pipe_ctx = current_mpc_pipe;
+					block_sequence[*num_steps].func = PROGRAM_CURSOR_UPDATE_NOW;
+					(*num_steps)++;
+				}
+
+				block_sequence[*num_steps].params.program_manual_trigger_params.pipe_ctx = current_mpc_pipe;
+				block_sequence[*num_steps].func = OPTC_PROGRAM_MANUAL_TRIGGER;
+				(*num_steps)++;
+			}
+			current_mpc_pipe = current_mpc_pipe->bottom_pipe;
+		}
+		current_pipe = current_pipe->next_odm_pipe;
+	}
+}
+
+void hwss_execute_sequence(struct dc *dc,
+		struct block_sequence block_sequence[MAX_HWSS_BLOCK_SEQUENCE_SIZE],
+		int num_steps)
+{
+	unsigned int i;
+	union block_sequence_params *params;
+	struct dce_hwseq *hws = dc->hwseq;
+
+	for (i = 0; i < num_steps; i++) {
+		params = &(block_sequence[i].params);
+		switch (block_sequence[i].func) {
+
+		case DMUB_SUBVP_PIPE_CONTROL_LOCK_FAST:
+			dc->hwss.subvp_pipe_control_lock_fast(params);
+			break;
+		case OPTC_PIPE_CONTROL_LOCK:
+			dc->hwss.pipe_control_lock(params->pipe_control_lock_params.dc,
+					params->pipe_control_lock_params.pipe_ctx,
+					params->pipe_control_lock_params.lock);
+			break;
+		case HUBP_SET_FLIP_CONTROL_GSL:
+			params->set_flip_control_gsl_params.hubp->funcs->hubp_set_flip_control_surface_gsl(
+				params->set_flip_control_gsl_params.hubp,
+				params->set_flip_control_gsl_params.flip_immediate);
+			break;
+		case HUBP_PROGRAM_TRIPLEBUFFER:
+			dc->hwss.program_triplebuffer(params->program_triplebuffer_params.dc,
+					params->program_triplebuffer_params.pipe_ctx,
+					params->program_triplebuffer_params.enableTripleBuffer);
+			break;
+		case HUBP_UPDATE_PLANE_ADDR:
+			dc->hwss.update_plane_addr(params->update_plane_addr_params.dc,
+					params->update_plane_addr_params.pipe_ctx);
+			break;
+		case DPP_SET_INPUT_TRANSFER_FUNC:
+			hws->funcs.set_input_transfer_func(params->set_input_transfer_func_params.dc,
+					params->set_input_transfer_func_params.pipe_ctx,
+					params->set_input_transfer_func_params.plane_state);
+			break;
+		case DPP_PROGRAM_GAMUT_REMAP:
+			dc->hwss.program_gamut_remap(params->program_gamut_remap_params.pipe_ctx);
+			break;
+		case DPP_SETUP_DPP:
+			hwss_setup_dpp(params);
+			break;
+		case DPP_PROGRAM_BIAS_AND_SCALE:
+			hwss_program_bias_and_scale(params);
+			break;
+		case OPTC_PROGRAM_MANUAL_TRIGGER:
+			hwss_program_manual_trigger(params);
+			break;
+		case DPP_SET_OUTPUT_TRANSFER_FUNC:
+			hws->funcs.set_output_transfer_func(params->set_output_transfer_func_params.dc,
+					params->set_output_transfer_func_params.pipe_ctx,
+					params->set_output_transfer_func_params.stream);
+			break;
+		case MPC_UPDATE_VISUAL_CONFIRM:
+			dc->hwss.update_visual_confirm_color(params->update_visual_confirm_params.dc,
+					params->update_visual_confirm_params.pipe_ctx,
+					params->update_visual_confirm_params.mpcc_id);
+			break;
+		case MPC_POWER_ON_MPC_MEM_PWR:
+			hwss_power_on_mpc_mem_pwr(params);
+			break;
+		case MPC_SET_OUTPUT_CSC:
+			hwss_set_output_csc(params);
+			break;
+		case MPC_SET_OCSC_DEFAULT:
+			hwss_set_ocsc_default(params);
+			break;
+		case DMUB_SEND_DMCUB_CMD:
+			hwss_send_dmcub_cmd(params);
+			break;
+		case DMUB_SUBVP_SAVE_SURF_ADDR:
+			hwss_subvp_save_surf_addr(params);
+			break;
+		case HUBP_WAIT_FOR_DCC_META_PROP:
+			dc->hwss.wait_for_dcc_meta_propagation(
+					params->wait_for_dcc_meta_propagation_params.dc,
+					params->wait_for_dcc_meta_propagation_params.top_pipe_to_program);
+			break;
+		case DMUB_HW_CONTROL_LOCK_FAST:
+			dc->hwss.dmub_hw_control_lock_fast(params);
+			break;
+		case HUBP_PROGRAM_SURFACE_CONFIG:
+			hwss_program_surface_config(params);
+			break;
+		case HUBP_PROGRAM_MCACHE_ID:
+			hwss_program_mcache_id_and_split_coordinate(params);
+			break;
+		case PROGRAM_CURSOR_UPDATE_NOW:
+			dc->hwss.program_cursor_offload_now(
+				params->program_cursor_update_now_params.dc,
+				params->program_cursor_update_now_params.pipe_ctx);
+			break;
+		case HUBP_WAIT_PIPE_READ_START:
+			params->hubp_wait_pipe_read_start_params.hubp->funcs->hubp_wait_pipe_read_start(
+				params->hubp_wait_pipe_read_start_params.hubp);
+			break;
+		case HWS_APPLY_UPDATE_FLAGS_FOR_PHANTOM:
+			dc->hwss.apply_update_flags_for_phantom(params->apply_update_flags_for_phantom_params.pipe_ctx);
+			break;
+		case HWS_UPDATE_PHANTOM_VP_POSITION:
+			dc->hwss.update_phantom_vp_position(params->update_phantom_vp_position_params.dc,
+				params->update_phantom_vp_position_params.context,
+				params->update_phantom_vp_position_params.pipe_ctx);
+			break;
+		case OPTC_SET_ODM_COMBINE:
+			hwss_set_odm_combine(params);
+			break;
+		case OPTC_SET_ODM_BYPASS:
+			hwss_set_odm_bypass(params);
+			break;
+		case OPP_PIPE_CLOCK_CONTROL:
+			hwss_opp_pipe_clock_control(params);
+			break;
+		case OPP_PROGRAM_LEFT_EDGE_EXTRA_PIXEL:
+			hwss_opp_program_left_edge_extra_pixel(params);
+			break;
+		case DCCG_SET_DTO_DSCCLK:
+			hwss_dccg_set_dto_dscclk(params);
+			break;
+		case DSC_SET_CONFIG:
+			hwss_dsc_set_config(params);
+			break;
+		case DSC_ENABLE:
+			hwss_dsc_enable(params);
+			break;
+		case TG_SET_DSC_CONFIG:
+			hwss_tg_set_dsc_config(params);
+			break;
+		case DSC_DISCONNECT:
+			hwss_dsc_disconnect(params);
+			break;
+		case DSC_READ_STATE:
+			hwss_dsc_read_state(params);
+			break;
+		case DSC_CALCULATE_AND_SET_CONFIG:
+			hwss_dsc_calculate_and_set_config(params);
+			break;
+		case DSC_ENABLE_WITH_OPP:
+			hwss_dsc_enable_with_opp(params);
+			break;
+		case TG_PROGRAM_GLOBAL_SYNC:
+			hwss_tg_program_global_sync(params);
+			break;
+		case TG_WAIT_FOR_STATE:
+			hwss_tg_wait_for_state(params);
+			break;
+		case TG_SET_VTG_PARAMS:
+			hwss_tg_set_vtg_params(params);
+			break;
+		case TG_SETUP_VERTICAL_INTERRUPT2:
+			hwss_tg_setup_vertical_interrupt2(params);
+			break;
+		case DPP_SET_HDR_MULTIPLIER:
+			hwss_dpp_set_hdr_multiplier(params);
+			break;
+		case HUBP_PROGRAM_DET_SIZE:
+			hwss_program_det_size(params);
+			break;
+		case HUBP_PROGRAM_DET_SEGMENTS:
+			hwss_program_det_segments(params);
+			break;
+		case OPP_SET_DYN_EXPANSION:
+			hwss_opp_set_dyn_expansion(params);
+			break;
+		case OPP_PROGRAM_FMT:
+			hwss_opp_program_fmt(params);
+			break;
+		case OPP_PROGRAM_BIT_DEPTH_REDUCTION:
+			hwss_opp_program_bit_depth_reduction(params);
+			break;
+		case OPP_SET_DISP_PATTERN_GENERATOR:
+			hwss_opp_set_disp_pattern_generator(params);
+			break;
+		case ABM_SET_PIPE:
+			hwss_set_abm_pipe(params);
+			break;
+		case ABM_SET_LEVEL:
+			hwss_set_abm_level(params);
+			break;
+		case ABM_SET_IMMEDIATE_DISABLE:
+			hwss_set_abm_immediate_disable(params);
+			break;
+		case MPC_REMOVE_MPCC:
+			hwss_mpc_remove_mpcc(params);
+			break;
+		case OPP_SET_MPCC_DISCONNECT_PENDING:
+			hwss_opp_set_mpcc_disconnect_pending(params);
+			break;
+		case DC_SET_OPTIMIZED_REQUIRED:
+			hwss_dc_set_optimized_required(params);
+			break;
+		case HUBP_DISCONNECT:
+			hwss_hubp_disconnect(params);
+			break;
+		case HUBBUB_FORCE_PSTATE_CHANGE_CONTROL:
+			hwss_hubbub_force_pstate_change_control(params);
+			break;
+		case TG_ENABLE_CRTC:
+			hwss_tg_enable_crtc(params);
+			break;
+		case TG_SET_GSL:
+			hwss_tg_set_gsl(params);
+			break;
+		case TG_SET_GSL_SOURCE_SELECT:
+			hwss_tg_set_gsl_source_select(params);
+			break;
+		case HUBP_WAIT_FLIP_PENDING:
+			hwss_hubp_wait_flip_pending(params);
+			break;
+		case TG_WAIT_DOUBLE_BUFFER_PENDING:
+			hwss_tg_wait_double_buffer_pending(params);
+			break;
+		case UPDATE_FORCE_PSTATE:
+			hwss_update_force_pstate(params);
+			break;
+		case HUBBUB_APPLY_DEDCN21_147_WA:
+			hwss_hubbub_apply_dedcn21_147_wa(params);
+			break;
+		case HUBBUB_ALLOW_SELF_REFRESH_CONTROL:
+			hwss_hubbub_allow_self_refresh_control(params);
+			break;
+		case TG_GET_FRAME_COUNT:
+			hwss_tg_get_frame_count(params);
+			break;
+		case MPC_SET_DWB_MUX:
+			hwss_mpc_set_dwb_mux(params);
+			break;
+		case MPC_DISABLE_DWB_MUX:
+			hwss_mpc_disable_dwb_mux(params);
+			break;
+		case MCIF_WB_CONFIG_BUF:
+			hwss_mcif_wb_config_buf(params);
+			break;
+		case MCIF_WB_CONFIG_ARB:
+			hwss_mcif_wb_config_arb(params);
+			break;
+		case MCIF_WB_ENABLE:
+			hwss_mcif_wb_enable(params);
+			break;
+		case MCIF_WB_DISABLE:
+			hwss_mcif_wb_disable(params);
+			break;
+		case DWBC_ENABLE:
+			hwss_dwbc_enable(params);
+			break;
+		case DWBC_DISABLE:
+			hwss_dwbc_disable(params);
+			break;
+		case DWBC_UPDATE:
+			hwss_dwbc_update(params);
+			break;
+		case HUBP_UPDATE_MALL_SEL:
+			hwss_hubp_update_mall_sel(params);
+			break;
+		case HUBP_PREPARE_SUBVP_BUFFERING:
+			hwss_hubp_prepare_subvp_buffering(params);
+			break;
+		case HUBP_SET_BLANK_EN:
+			hwss_hubp_set_blank_en(params);
+			break;
+		case HUBP_DISABLE_CONTROL:
+			hwss_hubp_disable_control(params);
+			break;
+		case HUBBUB_SOFT_RESET:
+			hwss_hubbub_soft_reset(params);
+			break;
+		case HUBP_CLK_CNTL:
+			hwss_hubp_clk_cntl(params);
+			break;
+		case HUBP_INIT:
+			hwss_hubp_init(params);
+			break;
+		case HUBP_SET_VM_SYSTEM_APERTURE_SETTINGS:
+			hwss_hubp_set_vm_system_aperture_settings(params);
+			break;
+		case HUBP_SET_FLIP_INT:
+			hwss_hubp_set_flip_int(params);
+			break;
+		case DPP_DPPCLK_CONTROL:
+			hwss_dpp_dppclk_control(params);
+			break;
+		case DISABLE_PHANTOM_CRTC:
+			hwss_disable_phantom_crtc(params);
+			break;
+		case DSC_PG_STATUS:
+			hwss_dsc_pg_status(params);
+			break;
+		case DSC_WAIT_DISCONNECT_PENDING_CLEAR:
+			hwss_dsc_wait_disconnect_pending_clear(params);
+			break;
+		case DSC_DISABLE:
+			hwss_dsc_disable(params);
+			break;
+		case DCCG_SET_REF_DSCCLK:
+			hwss_dccg_set_ref_dscclk(params);
+			break;
+		case DPP_PG_CONTROL:
+			hwss_dpp_pg_control(params);
+			break;
+		case HUBP_PG_CONTROL:
+			hwss_hubp_pg_control(params);
+			break;
+		case HUBP_RESET:
+			hwss_hubp_reset(params);
+			break;
+		case DPP_RESET:
+			hwss_dpp_reset(params);
+			break;
+		case DPP_ROOT_CLOCK_CONTROL:
+			hwss_dpp_root_clock_control(params);
+			break;
+		case DC_IP_REQUEST_CNTL:
+			hwss_dc_ip_request_cntl(params);
+			break;
+		case DCCG_UPDATE_DPP_DTO:
+			hwss_dccg_update_dpp_dto(params);
+			break;
+		case HUBP_VTG_SEL:
+			hwss_hubp_vtg_sel(params);
+			break;
+		case HUBP_SETUP2:
+			hwss_hubp_setup2(params);
+			break;
+		case HUBP_SETUP:
+			hwss_hubp_setup(params);
+			break;
+		case HUBP_SET_UNBOUNDED_REQUESTING:
+			hwss_hubp_set_unbounded_requesting(params);
+			break;
+		case HUBP_SETUP_INTERDEPENDENT2:
+			hwss_hubp_setup_interdependent2(params);
+			break;
+		case HUBP_SETUP_INTERDEPENDENT:
+			hwss_hubp_setup_interdependent(params);
+			break;
+		case DPP_SET_CURSOR_MATRIX:
+			hwss_dpp_set_cursor_matrix(params);
+			break;
+		case MPC_UPDATE_BLENDING:
+			hwss_mpc_update_blending(params);
+			break;
+		case MPC_ASSERT_IDLE_MPCC:
+			hwss_mpc_assert_idle_mpcc(params);
+			break;
+		case MPC_INSERT_PLANE:
+			hwss_mpc_insert_plane(params);
+			break;
+		case DPP_SET_SCALER:
+			hwss_dpp_set_scaler(params);
+			break;
+		case HUBP_MEM_PROGRAM_VIEWPORT:
+			hwss_hubp_mem_program_viewport(params);
+			break;
+		case ABORT_CURSOR_OFFLOAD_UPDATE:
+			hwss_abort_cursor_offload_update(params);
+			break;
+		case SET_CURSOR_ATTRIBUTE:
+			hwss_set_cursor_attribute(params);
+			break;
+		case SET_CURSOR_POSITION:
+			hwss_set_cursor_position(params);
+			break;
+		case SET_CURSOR_SDR_WHITE_LEVEL:
+			hwss_set_cursor_sdr_white_level(params);
+			break;
+		case PROGRAM_OUTPUT_CSC:
+			hwss_program_output_csc(params);
+			break;
+		case HUBP_SET_BLANK:
+			hwss_hubp_set_blank(params);
+			break;
+		case PHANTOM_HUBP_POST_ENABLE:
+			hwss_phantom_hubp_post_enable(params);
+			break;
+		default:
+			ASSERT(false);
+			break;
+		}
+	}
+}
+
+/*
+ * Helper function to add OPTC pipe control lock to block sequence
+ */
+void hwss_add_optc_pipe_control_lock(struct block_sequence_state *seq_state,
+		struct dc *dc,
+		struct pipe_ctx *pipe_ctx,
+		bool lock)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].params.pipe_control_lock_params.dc = dc;
+		seq_state->steps[*seq_state->num_steps].params.pipe_control_lock_params.pipe_ctx = pipe_ctx;
+		seq_state->steps[*seq_state->num_steps].params.pipe_control_lock_params.lock = lock;
+		seq_state->steps[*seq_state->num_steps].func = OPTC_PIPE_CONTROL_LOCK;
+		(*seq_state->num_steps)++;
+	}
+}
+
+/*
+ * Helper function to add HUBP set flip control GSL to block sequence
+ */
+void hwss_add_hubp_set_flip_control_gsl(struct block_sequence_state *seq_state,
+		struct hubp *hubp,
+		bool flip_immediate)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].params.set_flip_control_gsl_params.hubp = hubp;
+		seq_state->steps[*seq_state->num_steps].params.set_flip_control_gsl_params.flip_immediate = flip_immediate;
+		seq_state->steps[*seq_state->num_steps].func = HUBP_SET_FLIP_CONTROL_GSL;
+		(*seq_state->num_steps)++;
+	}
+}
+
+/*
+ * Helper function to add HUBP program triplebuffer to block sequence
+ */
+void hwss_add_hubp_program_triplebuffer(struct block_sequence_state *seq_state,
+		struct dc *dc,
+		struct pipe_ctx *pipe_ctx,
+		bool enableTripleBuffer)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].params.program_triplebuffer_params.dc = dc;
+		seq_state->steps[*seq_state->num_steps].params.program_triplebuffer_params.pipe_ctx = pipe_ctx;
+		seq_state->steps[*seq_state->num_steps].params.program_triplebuffer_params.enableTripleBuffer = enableTripleBuffer;
+		seq_state->steps[*seq_state->num_steps].func = HUBP_PROGRAM_TRIPLEBUFFER;
+		(*seq_state->num_steps)++;
+	}
+}
+
+/*
+ * Helper function to add HUBP update plane address to block sequence
+ */
+void hwss_add_hubp_update_plane_addr(struct block_sequence_state *seq_state,
+		struct dc *dc,
+		struct pipe_ctx *pipe_ctx)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].params.update_plane_addr_params.dc = dc;
+		seq_state->steps[*seq_state->num_steps].params.update_plane_addr_params.pipe_ctx = pipe_ctx;
+		seq_state->steps[*seq_state->num_steps].func = HUBP_UPDATE_PLANE_ADDR;
+		(*seq_state->num_steps)++;
+	}
+}
+
+/*
+ * Helper function to add DPP set input transfer function to block sequence
+ */
+void hwss_add_dpp_set_input_transfer_func(struct block_sequence_state *seq_state,
+		struct dc *dc,
+		struct pipe_ctx *pipe_ctx,
+		struct dc_plane_state *plane_state)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].params.set_input_transfer_func_params.dc = dc;
+		seq_state->steps[*seq_state->num_steps].params.set_input_transfer_func_params.pipe_ctx = pipe_ctx;
+		seq_state->steps[*seq_state->num_steps].params.set_input_transfer_func_params.plane_state = plane_state;
+		seq_state->steps[*seq_state->num_steps].func = DPP_SET_INPUT_TRANSFER_FUNC;
+		(*seq_state->num_steps)++;
+	}
+}
+
+/*
+ * Helper function to add DPP program gamut remap to block sequence
+ */
+void hwss_add_dpp_program_gamut_remap(struct block_sequence_state *seq_state,
+		struct pipe_ctx *pipe_ctx)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].params.program_gamut_remap_params.pipe_ctx = pipe_ctx;
+		seq_state->steps[*seq_state->num_steps].func = DPP_PROGRAM_GAMUT_REMAP;
+		(*seq_state->num_steps)++;
+	}
+}
+
+/*
+ * Helper function to add DPP program bias and scale to block sequence
+ */
+void hwss_add_dpp_program_bias_and_scale(struct block_sequence_state *seq_state, struct pipe_ctx *pipe_ctx)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].params.program_bias_and_scale_params.pipe_ctx = pipe_ctx;
+		seq_state->steps[*seq_state->num_steps].func = DPP_PROGRAM_BIAS_AND_SCALE;
+		(*seq_state->num_steps)++;
+	}
+}
+
+/*
+ * Helper function to add OPTC program manual trigger to block sequence
+ */
+void hwss_add_optc_program_manual_trigger(struct block_sequence_state *seq_state,
+		struct pipe_ctx *pipe_ctx)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].params.program_manual_trigger_params.pipe_ctx = pipe_ctx;
+		seq_state->steps[*seq_state->num_steps].func = OPTC_PROGRAM_MANUAL_TRIGGER;
+		(*seq_state->num_steps)++;
+	}
+}
+
+/*
+ * Helper function to add DPP set output transfer function to block sequence
+ */
+void hwss_add_dpp_set_output_transfer_func(struct block_sequence_state *seq_state,
+		struct dc *dc,
+		struct pipe_ctx *pipe_ctx,
+		struct dc_stream_state *stream)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].params.set_output_transfer_func_params.dc = dc;
+		seq_state->steps[*seq_state->num_steps].params.set_output_transfer_func_params.pipe_ctx = pipe_ctx;
+		seq_state->steps[*seq_state->num_steps].params.set_output_transfer_func_params.stream = stream;
+		seq_state->steps[*seq_state->num_steps].func = DPP_SET_OUTPUT_TRANSFER_FUNC;
+		(*seq_state->num_steps)++;
+	}
+}
+
+/*
+ * Helper function to add MPC update visual confirm to block sequence
+ */
+void hwss_add_mpc_update_visual_confirm(struct block_sequence_state *seq_state,
+		struct dc *dc,
+		struct pipe_ctx *pipe_ctx,
+		int mpcc_id)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].params.update_visual_confirm_params.dc = dc;
+		seq_state->steps[*seq_state->num_steps].params.update_visual_confirm_params.pipe_ctx = pipe_ctx;
+		seq_state->steps[*seq_state->num_steps].params.update_visual_confirm_params.mpcc_id = mpcc_id;
+		seq_state->steps[*seq_state->num_steps].func = MPC_UPDATE_VISUAL_CONFIRM;
+		(*seq_state->num_steps)++;
+	}
+}
+
+/*
+ * Helper function to add MPC power on MPC mem PWR to block sequence
+ */
+void hwss_add_mpc_power_on_mpc_mem_pwr(struct block_sequence_state *seq_state,
+		struct mpc *mpc,
+		int mpcc_id,
+		bool power_on)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].params.power_on_mpc_mem_pwr_params.mpc = mpc;
+		seq_state->steps[*seq_state->num_steps].params.power_on_mpc_mem_pwr_params.mpcc_id = mpcc_id;
+		seq_state->steps[*seq_state->num_steps].params.power_on_mpc_mem_pwr_params.power_on = power_on;
+		seq_state->steps[*seq_state->num_steps].func = MPC_POWER_ON_MPC_MEM_PWR;
+		(*seq_state->num_steps)++;
+	}
+}
+
+/*
+ * Helper function to add MPC set output CSC to block sequence
+ */
+void hwss_add_mpc_set_output_csc(struct block_sequence_state *seq_state,
+		struct mpc *mpc,
+		int opp_id,
+		const uint16_t *regval,
+		enum mpc_output_csc_mode ocsc_mode)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].params.set_output_csc_params.mpc = mpc;
+		seq_state->steps[*seq_state->num_steps].params.set_output_csc_params.opp_id = opp_id;
+		seq_state->steps[*seq_state->num_steps].params.set_output_csc_params.regval = regval;
+		seq_state->steps[*seq_state->num_steps].params.set_output_csc_params.ocsc_mode = ocsc_mode;
+		seq_state->steps[*seq_state->num_steps].func = MPC_SET_OUTPUT_CSC;
+		(*seq_state->num_steps)++;
+	}
+}
+
+/*
+ * Helper function to add MPC set OCSC default to block sequence
+ */
+void hwss_add_mpc_set_ocsc_default(struct block_sequence_state *seq_state,
+		struct mpc *mpc,
+		int opp_id,
+		enum dc_color_space colorspace,
+		enum mpc_output_csc_mode ocsc_mode)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].params.set_ocsc_default_params.mpc = mpc;
+		seq_state->steps[*seq_state->num_steps].params.set_ocsc_default_params.opp_id = opp_id;
+		seq_state->steps[*seq_state->num_steps].params.set_ocsc_default_params.color_space = colorspace;
+		seq_state->steps[*seq_state->num_steps].params.set_ocsc_default_params.ocsc_mode = ocsc_mode;
+		seq_state->steps[*seq_state->num_steps].func = MPC_SET_OCSC_DEFAULT;
+		(*seq_state->num_steps)++;
+	}
+}
+
+/*
+ * Helper function to add DMUB send DMCUB command to block sequence
+ */
+void hwss_add_dmub_send_dmcub_cmd(struct block_sequence_state *seq_state,
+		struct dc_context *ctx,
+		union dmub_rb_cmd *cmd,
+		enum dm_dmub_wait_type wait_type)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].params.send_dmcub_cmd_params.ctx = ctx;
+		seq_state->steps[*seq_state->num_steps].params.send_dmcub_cmd_params.cmd = cmd;
+		seq_state->steps[*seq_state->num_steps].params.send_dmcub_cmd_params.wait_type = wait_type;
+		seq_state->steps[*seq_state->num_steps].func = DMUB_SEND_DMCUB_CMD;
+		(*seq_state->num_steps)++;
+	}
+}
+
+/*
+ * Helper function to add DMUB SubVP save surface address to block sequence
+ */
+void hwss_add_dmub_subvp_save_surf_addr(struct block_sequence_state *seq_state,
+		struct dc_dmub_srv *dc_dmub_srv,
+		struct dc_plane_address *addr,
+		uint8_t subvp_index)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].params.subvp_save_surf_addr.dc_dmub_srv = dc_dmub_srv;
+		seq_state->steps[*seq_state->num_steps].params.subvp_save_surf_addr.addr = addr;
+		seq_state->steps[*seq_state->num_steps].params.subvp_save_surf_addr.subvp_index = subvp_index;
+		seq_state->steps[*seq_state->num_steps].func = DMUB_SUBVP_SAVE_SURF_ADDR;
+		(*seq_state->num_steps)++;
+	}
+}
+
+/*
+ * Helper function to add HUBP wait for DCC meta propagation to block sequence
+ */
+void hwss_add_hubp_wait_for_dcc_meta_prop(struct block_sequence_state *seq_state,
+		struct dc *dc,
+		struct pipe_ctx *top_pipe_to_program)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].params.wait_for_dcc_meta_propagation_params.dc = dc;
+		seq_state->steps[*seq_state->num_steps].params.wait_for_dcc_meta_propagation_params.top_pipe_to_program = top_pipe_to_program;
+		seq_state->steps[*seq_state->num_steps].func = HUBP_WAIT_FOR_DCC_META_PROP;
+		(*seq_state->num_steps)++;
+	}
+}
+
+/*
+ * Helper function to add HUBP wait pipe read start to block sequence
+ */
+void hwss_add_hubp_wait_pipe_read_start(struct block_sequence_state *seq_state,
+		struct hubp *hubp)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].params.hubp_wait_pipe_read_start_params.hubp = hubp;
+		seq_state->steps[*seq_state->num_steps].func = HUBP_WAIT_PIPE_READ_START;
+		(*seq_state->num_steps)++;
+	}
+}
+
+/*
+ * Helper function to add HWS apply update flags for phantom to block sequence
+ */
+void hwss_add_hws_apply_update_flags_for_phantom(struct block_sequence_state *seq_state,
+		struct pipe_ctx *pipe_ctx)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].params.apply_update_flags_for_phantom_params.pipe_ctx = pipe_ctx;
+		seq_state->steps[*seq_state->num_steps].func = HWS_APPLY_UPDATE_FLAGS_FOR_PHANTOM;
+		(*seq_state->num_steps)++;
+	}
+}
+
+/*
+ * Helper function to add HWS update phantom VP position to block sequence
+ */
+void hwss_add_hws_update_phantom_vp_position(struct block_sequence_state *seq_state,
+		struct dc *dc,
+		struct dc_state *context,
+		struct pipe_ctx *pipe_ctx)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].params.update_phantom_vp_position_params.dc = dc;
+		seq_state->steps[*seq_state->num_steps].params.update_phantom_vp_position_params.context = context;
+		seq_state->steps[*seq_state->num_steps].params.update_phantom_vp_position_params.pipe_ctx = pipe_ctx;
+		seq_state->steps[*seq_state->num_steps].func = HWS_UPDATE_PHANTOM_VP_POSITION;
+		(*seq_state->num_steps)++;
+	}
+}
+
+/*
+ * Helper function to add OPTC set ODM combine to block sequence
+ */
+void hwss_add_optc_set_odm_combine(struct block_sequence_state *seq_state,
+		struct timing_generator *tg, int opp_inst[MAX_PIPES], int opp_head_count,
+		int odm_slice_width, int last_odm_slice_width)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].params.set_odm_combine_params.tg = tg;
+		memcpy(seq_state->steps[*seq_state->num_steps].params.set_odm_combine_params.opp_inst, opp_inst, sizeof(int) * MAX_PIPES);
+		seq_state->steps[*seq_state->num_steps].params.set_odm_combine_params.opp_head_count = opp_head_count;
+		seq_state->steps[*seq_state->num_steps].params.set_odm_combine_params.odm_slice_width = odm_slice_width;
+		seq_state->steps[*seq_state->num_steps].params.set_odm_combine_params.last_odm_slice_width = last_odm_slice_width;
+		seq_state->steps[*seq_state->num_steps].func = OPTC_SET_ODM_COMBINE;
+		(*seq_state->num_steps)++;
+	}
+}
+
+/*
+ * Helper function to add OPTC set ODM bypass to block sequence
+ */
+void hwss_add_optc_set_odm_bypass(struct block_sequence_state *seq_state,
+		struct timing_generator *tg, struct dc_crtc_timing *timing)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].params.set_odm_bypass_params.tg = tg;
+		seq_state->steps[*seq_state->num_steps].params.set_odm_bypass_params.timing = timing;
+		seq_state->steps[*seq_state->num_steps].func = OPTC_SET_ODM_BYPASS;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_send_dmcub_cmd(union block_sequence_params *params)
+{
+	struct dc_context *ctx = params->send_dmcub_cmd_params.ctx;
+	union dmub_rb_cmd *cmd = params->send_dmcub_cmd_params.cmd;
+	enum dm_dmub_wait_type wait_type = params->send_dmcub_cmd_params.wait_type;
+
+	dc_wake_and_execute_dmub_cmd(ctx, cmd, wait_type);
+}
+
+/*
+ * Helper function to add TG program global sync to block sequence
+ */
+void hwss_add_tg_program_global_sync(struct block_sequence_state *seq_state,
+		struct timing_generator *tg,
+		int vready_offset,
+		unsigned int vstartup_lines,
+		unsigned int vupdate_offset_pixels,
+		unsigned int vupdate_vupdate_width_pixels,
+		unsigned int pstate_keepout_start_lines)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].params.tg_program_global_sync_params.tg = tg;
+		seq_state->steps[*seq_state->num_steps].params.tg_program_global_sync_params.vready_offset = vready_offset;
+		seq_state->steps[*seq_state->num_steps].params.tg_program_global_sync_params.vstartup_lines = vstartup_lines;
+		seq_state->steps[*seq_state->num_steps].params.tg_program_global_sync_params.vupdate_offset_pixels = vupdate_offset_pixels;
+		seq_state->steps[*seq_state->num_steps].params.tg_program_global_sync_params.vupdate_vupdate_width_pixels = vupdate_vupdate_width_pixels;
+		seq_state->steps[*seq_state->num_steps].params.tg_program_global_sync_params.pstate_keepout_start_lines = pstate_keepout_start_lines;
+		seq_state->steps[*seq_state->num_steps].func = TG_PROGRAM_GLOBAL_SYNC;
+		(*seq_state->num_steps)++;
+	}
+}
+
+/*
+ * Helper function to add TG wait for state to block sequence
+ */
+void hwss_add_tg_wait_for_state(struct block_sequence_state *seq_state,
+		struct timing_generator *tg,
+		enum crtc_state state)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].params.tg_wait_for_state_params.tg = tg;
+		seq_state->steps[*seq_state->num_steps].params.tg_wait_for_state_params.state = state;
+		seq_state->steps[*seq_state->num_steps].func = TG_WAIT_FOR_STATE;
+		(*seq_state->num_steps)++;
+	}
+}
+
+/*
+ * Helper function to add TG set VTG params to block sequence
+ */
+void hwss_add_tg_set_vtg_params(struct block_sequence_state *seq_state,
+		struct timing_generator *tg,
+		struct dc_crtc_timing *dc_crtc_timing,
+		bool program_fp2)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].params.tg_set_vtg_params_params.tg = tg;
+		seq_state->steps[*seq_state->num_steps].params.tg_set_vtg_params_params.timing = dc_crtc_timing;
+		seq_state->steps[*seq_state->num_steps].params.tg_set_vtg_params_params.program_fp2 = program_fp2;
+		seq_state->steps[*seq_state->num_steps].func = TG_SET_VTG_PARAMS;
+		(*seq_state->num_steps)++;
+	}
+}
+
+/*
+ * Helper function to add TG setup vertical interrupt2 to block sequence
+ */
+void hwss_add_tg_setup_vertical_interrupt2(struct block_sequence_state *seq_state,
+		struct timing_generator *tg, int start_line)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].params.tg_setup_vertical_interrupt2_params.tg = tg;
+		seq_state->steps[*seq_state->num_steps].params.tg_setup_vertical_interrupt2_params.start_line = start_line;
+		seq_state->steps[*seq_state->num_steps].func = TG_SETUP_VERTICAL_INTERRUPT2;
+		(*seq_state->num_steps)++;
+	}
+}
+
+/*
+ * Helper function to add DPP set HDR multiplier to block sequence
+ */
+void hwss_add_dpp_set_hdr_multiplier(struct block_sequence_state *seq_state,
+		struct dpp *dpp, uint32_t hw_mult)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].params.dpp_set_hdr_multiplier_params.dpp = dpp;
+		seq_state->steps[*seq_state->num_steps].params.dpp_set_hdr_multiplier_params.hw_mult = hw_mult;
+		seq_state->steps[*seq_state->num_steps].func = DPP_SET_HDR_MULTIPLIER;
+		(*seq_state->num_steps)++;
+	}
+}
+
+/*
+ * Helper function to add HUBP program DET size to block sequence
+ */
+void hwss_add_hubp_program_det_size(struct block_sequence_state *seq_state,
+		struct hubbub *hubbub,
+		unsigned int hubp_inst,
+		unsigned int det_buffer_size_kb)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].params.program_det_size_params.hubbub = hubbub;
+		seq_state->steps[*seq_state->num_steps].params.program_det_size_params.hubp_inst = hubp_inst;
+		seq_state->steps[*seq_state->num_steps].params.program_det_size_params.det_buffer_size_kb = det_buffer_size_kb;
+		seq_state->steps[*seq_state->num_steps].func = HUBP_PROGRAM_DET_SIZE;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_hubp_program_mcache_id(struct block_sequence_state *seq_state,
+		struct hubp *hubp,
+		struct dml2_hubp_pipe_mcache_regs *mcache_regs)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].params.program_mcache_id_and_split_coordinate.hubp = hubp;
+		seq_state->steps[*seq_state->num_steps].params.program_mcache_id_and_split_coordinate.mcache_regs = mcache_regs;
+		seq_state->steps[*seq_state->num_steps].func = HUBP_PROGRAM_MCACHE_ID;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_hubbub_force_pstate_change_control(struct block_sequence_state *seq_state,
+		struct hubbub *hubbub,
+		bool enable,
+		bool wait)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].params.hubbub_force_pstate_change_control_params.hubbub = hubbub;
+		seq_state->steps[*seq_state->num_steps].params.hubbub_force_pstate_change_control_params.enable = enable;
+		seq_state->steps[*seq_state->num_steps].params.hubbub_force_pstate_change_control_params.wait = wait;
+		seq_state->steps[*seq_state->num_steps].func = HUBBUB_FORCE_PSTATE_CHANGE_CONTROL;
+		(*seq_state->num_steps)++;
+	}
+}
+
+/*
+ * Helper function to add HUBP program DET segments to block sequence
+ */
+void hwss_add_hubp_program_det_segments(struct block_sequence_state *seq_state,
+		struct hubbub *hubbub,
+		unsigned int hubp_inst,
+		unsigned int det_size)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].params.program_det_segments_params.hubbub = hubbub;
+		seq_state->steps[*seq_state->num_steps].params.program_det_segments_params.hubp_inst = hubp_inst;
+		seq_state->steps[*seq_state->num_steps].params.program_det_segments_params.det_size = det_size;
+		seq_state->steps[*seq_state->num_steps].func = HUBP_PROGRAM_DET_SEGMENTS;
+		(*seq_state->num_steps)++;
+	}
+}
+
+/*
+ * Helper function to add OPP set dynamic expansion to block sequence
+ */
+void hwss_add_opp_set_dyn_expansion(struct block_sequence_state *seq_state,
+		struct output_pixel_processor *opp,
+		enum dc_color_space color_space,
+		enum dc_color_depth color_depth,
+		enum signal_type signal)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].params.opp_set_dyn_expansion_params.opp = opp;
+		seq_state->steps[*seq_state->num_steps].params.opp_set_dyn_expansion_params.color_space = color_space;
+		seq_state->steps[*seq_state->num_steps].params.opp_set_dyn_expansion_params.color_depth = color_depth;
+		seq_state->steps[*seq_state->num_steps].params.opp_set_dyn_expansion_params.signal = signal;
+		seq_state->steps[*seq_state->num_steps].func = OPP_SET_DYN_EXPANSION;
+		(*seq_state->num_steps)++;
+	}
+}
+
+/*
+ * Helper function to add OPP program FMT to block sequence
+ */
+void hwss_add_opp_program_fmt(struct block_sequence_state *seq_state,
+		struct output_pixel_processor *opp,
+		struct bit_depth_reduction_params *fmt_bit_depth,
+		struct clamping_and_pixel_encoding_params *clamping)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].params.opp_program_fmt_params.opp = opp;
+		seq_state->steps[*seq_state->num_steps].params.opp_program_fmt_params.fmt_bit_depth = fmt_bit_depth;
+		seq_state->steps[*seq_state->num_steps].params.opp_program_fmt_params.clamping = clamping;
+		seq_state->steps[*seq_state->num_steps].func = OPP_PROGRAM_FMT;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_opp_program_left_edge_extra_pixel(struct block_sequence_state *seq_state,
+		struct output_pixel_processor *opp,
+		enum dc_pixel_encoding pixel_encoding,
+		bool is_otg_master)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = OPP_PROGRAM_LEFT_EDGE_EXTRA_PIXEL;
+		seq_state->steps[*seq_state->num_steps].params.opp_program_left_edge_extra_pixel_params.opp = opp;
+		seq_state->steps[*seq_state->num_steps].params.opp_program_left_edge_extra_pixel_params.pixel_encoding = pixel_encoding;
+		seq_state->steps[*seq_state->num_steps].params.opp_program_left_edge_extra_pixel_params.is_otg_master = is_otg_master;
+		(*seq_state->num_steps)++;
+	}
+}
+
+/*
+ * Helper function to add ABM set pipe to block sequence
+ */
+void hwss_add_abm_set_pipe(struct block_sequence_state *seq_state,
+		struct dc *dc,
+		struct pipe_ctx *pipe_ctx)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].params.set_abm_pipe_params.dc = dc;
+		seq_state->steps[*seq_state->num_steps].params.set_abm_pipe_params.pipe_ctx = pipe_ctx;
+		seq_state->steps[*seq_state->num_steps].func = ABM_SET_PIPE;
+		(*seq_state->num_steps)++;
+	}
+}
+
+/*
+ * Helper function to add ABM set level to block sequence
+ */
+void hwss_add_abm_set_level(struct block_sequence_state *seq_state,
+		struct abm *abm,
+		uint32_t abm_level)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].params.set_abm_level_params.abm = abm;
+		seq_state->steps[*seq_state->num_steps].params.set_abm_level_params.abm_level = abm_level;
+		seq_state->steps[*seq_state->num_steps].func = ABM_SET_LEVEL;
+		(*seq_state->num_steps)++;
+	}
+}
+
+/*
+ * Helper function to add TG enable CRTC to block sequence
+ */
+void hwss_add_tg_enable_crtc(struct block_sequence_state *seq_state,
+		struct timing_generator *tg)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].params.tg_enable_crtc_params.tg = tg;
+		seq_state->steps[*seq_state->num_steps].func = TG_ENABLE_CRTC;
+		(*seq_state->num_steps)++;
+	}
+}
+
+/*
+ * Helper function to add HUBP wait flip pending to block sequence
+ */
+void hwss_add_hubp_wait_flip_pending(struct block_sequence_state *seq_state,
+		struct hubp *hubp,
+		unsigned int timeout_us,
+		unsigned int polling_interval_us)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].params.hubp_wait_flip_pending_params.hubp = hubp;
+		seq_state->steps[*seq_state->num_steps].params.hubp_wait_flip_pending_params.timeout_us = timeout_us;
+		seq_state->steps[*seq_state->num_steps].params.hubp_wait_flip_pending_params.polling_interval_us = polling_interval_us;
+		seq_state->steps[*seq_state->num_steps].func = HUBP_WAIT_FLIP_PENDING;
+		(*seq_state->num_steps)++;
+	}
+}
+
+/*
+ * Helper function to add TG wait double buffer pending to block sequence
+ */
+void hwss_add_tg_wait_double_buffer_pending(struct block_sequence_state *seq_state,
+		struct timing_generator *tg,
+		unsigned int timeout_us,
+		unsigned int polling_interval_us)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].params.tg_wait_double_buffer_pending_params.tg = tg;
+		seq_state->steps[*seq_state->num_steps].params.tg_wait_double_buffer_pending_params.timeout_us = timeout_us;
+		seq_state->steps[*seq_state->num_steps].params.tg_wait_double_buffer_pending_params.polling_interval_us = polling_interval_us;
+		seq_state->steps[*seq_state->num_steps].func = TG_WAIT_DOUBLE_BUFFER_PENDING;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_program_manual_trigger(union block_sequence_params *params)
+{
+	struct pipe_ctx *pipe_ctx = params->program_manual_trigger_params.pipe_ctx;
+
+	if (pipe_ctx->stream_res.tg->funcs->program_manual_trigger)
+		pipe_ctx->stream_res.tg->funcs->program_manual_trigger(pipe_ctx->stream_res.tg);
+}
+
+void hwss_setup_dpp(union block_sequence_params *params)
+{
+	struct pipe_ctx *pipe_ctx = params->setup_dpp_params.pipe_ctx;
+	struct dpp *dpp = pipe_ctx->plane_res.dpp;
+	struct dc_plane_state *plane_state = pipe_ctx->plane_state;
+
+	if (!plane_state)
+		return;
+
+	if (dpp && dpp->funcs->dpp_setup) {
+		// program the input csc
+		dpp->funcs->dpp_setup(dpp,
+				plane_state->format,
+				EXPANSION_MODE_ZERO,
+				plane_state->input_csc_color_matrix,
+				plane_state->color_space,
+				NULL);
+	}
+}
+
+void hwss_program_bias_and_scale(union block_sequence_params *params)
+{
+	struct pipe_ctx *pipe_ctx = params->program_bias_and_scale_params.pipe_ctx;
+	struct dpp *dpp = pipe_ctx->plane_res.dpp;
+	struct dc_plane_state *plane_state = pipe_ctx->plane_state;
+	struct dc_bias_and_scale bns_params = plane_state->bias_and_scale;
+
+	//TODO :for CNVC set scale and bias registers if necessary
+	if (dpp->funcs->dpp_program_bias_and_scale)
+		dpp->funcs->dpp_program_bias_and_scale(dpp, &bns_params);
+}
+
+void hwss_power_on_mpc_mem_pwr(union block_sequence_params *params)
+{
+	struct mpc *mpc = params->power_on_mpc_mem_pwr_params.mpc;
+	int mpcc_id = params->power_on_mpc_mem_pwr_params.mpcc_id;
+	bool power_on = params->power_on_mpc_mem_pwr_params.power_on;
+
+	if (mpc->funcs->power_on_mpc_mem_pwr)
+		mpc->funcs->power_on_mpc_mem_pwr(mpc, mpcc_id, power_on);
+}
+
+void hwss_set_output_csc(union block_sequence_params *params)
+{
+	struct mpc *mpc = params->set_output_csc_params.mpc;
+	int opp_id = params->set_output_csc_params.opp_id;
+	const uint16_t *matrix = params->set_output_csc_params.regval;
+	enum mpc_output_csc_mode ocsc_mode = params->set_output_csc_params.ocsc_mode;
+
+	if (mpc->funcs->set_output_csc != NULL)
+		mpc->funcs->set_output_csc(mpc,
+				opp_id,
+				matrix,
+				ocsc_mode);
+}
+
+void hwss_set_ocsc_default(union block_sequence_params *params)
+{
+	struct mpc *mpc = params->set_ocsc_default_params.mpc;
+	int opp_id = params->set_ocsc_default_params.opp_id;
+	enum dc_color_space colorspace = params->set_ocsc_default_params.color_space;
+	enum mpc_output_csc_mode ocsc_mode = params->set_ocsc_default_params.ocsc_mode;
+
+	if (mpc->funcs->set_ocsc_default != NULL)
+		mpc->funcs->set_ocsc_default(mpc,
+				opp_id,
+				colorspace,
+				ocsc_mode);
+}
+
+void hwss_subvp_save_surf_addr(union block_sequence_params *params)
+{
+	struct dc_dmub_srv *dc_dmub_srv = params->subvp_save_surf_addr.dc_dmub_srv;
+	const struct dc_plane_address *addr = params->subvp_save_surf_addr.addr;
+	uint8_t subvp_index = params->subvp_save_surf_addr.subvp_index;
+
+	dc_dmub_srv_subvp_save_surf_addr(dc_dmub_srv, addr, subvp_index);
+}
+
+void hwss_program_surface_config(union block_sequence_params *params)
+{
+	struct hubp *hubp = params->program_surface_config_params.hubp;
+	enum surface_pixel_format format = params->program_surface_config_params.format;
+	struct dc_tiling_info *tiling_info = params->program_surface_config_params.tiling_info;
+	struct plane_size size = params->program_surface_config_params.plane_size;
+	enum dc_rotation_angle rotation = params->program_surface_config_params.rotation;
+	struct dc_plane_dcc_param *dcc = params->program_surface_config_params.dcc;
+	bool horizontal_mirror = params->program_surface_config_params.horizontal_mirror;
+	int compat_level = params->program_surface_config_params.compat_level;
+
+	hubp->funcs->hubp_program_surface_config(
+		hubp,
+		format,
+		tiling_info,
+		&size,
+		rotation,
+		dcc,
+		horizontal_mirror,
+		compat_level);
+
+	hubp->power_gated = false;
+}
+
+void hwss_program_mcache_id_and_split_coordinate(union block_sequence_params *params)
+{
+	struct hubp *hubp = params->program_mcache_id_and_split_coordinate.hubp;
+	struct dml2_hubp_pipe_mcache_regs *mcache_regs = params->program_mcache_id_and_split_coordinate.mcache_regs;
+
+	hubp->funcs->hubp_program_mcache_id_and_split_coordinate(hubp, mcache_regs);
+
+}
+
 void get_surface_tile_visual_confirm_color(
 		struct pipe_ctx *pipe_ctx,
 		struct tg_color *color)
@@ -424,3 +2072,2000 @@ void get_surface_tile_visual_confirm_color(
 		break;
 	}
 }
+
+/**
+ * hwss_wait_for_all_blank_complete - wait for all active OPPs to finish pending blank
+ * pattern updates
+ *
+ * @dc: [in] dc reference
+ * @context: [in] hardware context in use
+ */
+void hwss_wait_for_all_blank_complete(struct dc *dc,
+		struct dc_state *context)
+{
+	struct pipe_ctx *opp_head;
+	struct dce_hwseq *hws = dc->hwseq;
+	int i;
+
+	if (!hws->funcs.wait_for_blank_complete)
+		return;
+
+	for (i = 0; i < MAX_PIPES; i++) {
+		opp_head = &context->res_ctx.pipe_ctx[i];
+
+		if (!resource_is_pipe_type(opp_head, OPP_HEAD) ||
+				dc_state_get_pipe_subvp_type(context, opp_head) == SUBVP_PHANTOM)
+			continue;
+
+		hws->funcs.wait_for_blank_complete(opp_head->stream_res.opp);
+	}
+}
+
+void hwss_wait_for_odm_update_pending_complete(struct dc *dc, struct dc_state *context)
+{
+	struct pipe_ctx *otg_master;
+	struct timing_generator *tg;
+	int i;
+
+	for (i = 0; i < MAX_PIPES; i++) {
+		otg_master = &context->res_ctx.pipe_ctx[i];
+		if (!resource_is_pipe_type(otg_master, OTG_MASTER) ||
+				dc_state_get_pipe_subvp_type(context, otg_master) == SUBVP_PHANTOM)
+			continue;
+		tg = otg_master->stream_res.tg;
+		if (tg->funcs->wait_odm_doublebuffer_pending_clear)
+			tg->funcs->wait_odm_doublebuffer_pending_clear(tg);
+		if (tg->funcs->wait_otg_disable)
+			tg->funcs->wait_otg_disable(tg);
+	}
+
+	/* ODM update may require to reprogram blank pattern for each OPP */
+	hwss_wait_for_all_blank_complete(dc, context);
+}
+
+void hwss_wait_for_no_pipes_pending(struct dc *dc, struct dc_state *context)
+{
+	int i;
+
+	for (i = 0; i < MAX_PIPES; i++) {
+		int count = 0;
+		struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
+
+		if (!pipe->plane_state || dc_state_get_pipe_subvp_type(context, pipe) == SUBVP_PHANTOM)
+			continue;
+
+		/* Timeout 100 ms */
+		while (count < 100000) {
+			/* Must set to false to start with, due to OR in update function */
+			pipe->plane_state->status.is_flip_pending = false;
+			dc->hwss.update_pending_status(pipe);
+			if (!pipe->plane_state->status.is_flip_pending)
+				break;
+			udelay(1);
+			count++;
+		}
+		ASSERT(!pipe->plane_state->status.is_flip_pending);
+	}
+}
+
+void hwss_wait_for_outstanding_hw_updates(struct dc *dc, struct dc_state *dc_context)
+{
+/*
+ * This function calls HWSS to wait for any potentially double buffered
+ * operations to complete. It should be invoked as a pre-amble prior
+ * to full update programming before asserting any HW locks.
+ */
+	int pipe_idx;
+	int opp_inst;
+	int opp_count = dc->res_pool->res_cap->num_opp;
+	struct hubp *hubp;
+	int mpcc_inst;
+	const struct pipe_ctx *pipe_ctx;
+
+	for (pipe_idx = 0; pipe_idx < dc->res_pool->pipe_count; pipe_idx++) {
+		pipe_ctx = &dc_context->res_ctx.pipe_ctx[pipe_idx];
+
+		if (!pipe_ctx->stream)
+			continue;
+
+		/* For full update we must wait for all double buffer updates, not just DRR updates. This
+		 * is particularly important for minimal transitions. Only check for OTG_MASTER pipes,
+		 * as non-OTG Master pipes share the same OTG as
+		 */
+		if (resource_is_pipe_type(pipe_ctx, OTG_MASTER) && dc->hwss.wait_for_all_pending_updates) {
+			dc->hwss.wait_for_all_pending_updates(pipe_ctx);
+		}
+
+		hubp = pipe_ctx->plane_res.hubp;
+		if (!hubp)
+			continue;
+
+		mpcc_inst = hubp->inst;
+		// MPCC inst is equal to pipe index in practice
+		for (opp_inst = 0; opp_inst < opp_count; opp_inst++) {
+			if ((dc->res_pool->opps[opp_inst] != NULL) &&
+				(dc->res_pool->opps[opp_inst]->mpcc_disconnect_pending[mpcc_inst])) {
+				dc->res_pool->mpc->funcs->wait_for_idle(dc->res_pool->mpc, mpcc_inst);
+				dc->res_pool->opps[opp_inst]->mpcc_disconnect_pending[mpcc_inst] = false;
+				break;
+			}
+		}
+	}
+	hwss_wait_for_odm_update_pending_complete(dc, dc_context);
+}
+
+void hwss_process_outstanding_hw_updates(struct dc *dc, struct dc_state *dc_context)
+{
+	/* wait for outstanding updates */
+	hwss_wait_for_outstanding_hw_updates(dc, dc_context);
+
+	/* perform outstanding post update programming */
+	if (dc->hwss.program_outstanding_updates)
+		dc->hwss.program_outstanding_updates(dc, dc_context);
+}
+
+void hwss_set_odm_combine(union block_sequence_params *params)
+{
+	struct timing_generator *tg = params->set_odm_combine_params.tg;
+	int *opp_inst = params->set_odm_combine_params.opp_inst;
+	int opp_head_count = params->set_odm_combine_params.opp_head_count;
+	int odm_slice_width = params->set_odm_combine_params.odm_slice_width;
+	int last_odm_slice_width = params->set_odm_combine_params.last_odm_slice_width;
+
+	if (tg && tg->funcs->set_odm_combine)
+		tg->funcs->set_odm_combine(tg, opp_inst, opp_head_count,
+				odm_slice_width, last_odm_slice_width);
+}
+
+void hwss_set_odm_bypass(union block_sequence_params *params)
+{
+	struct timing_generator *tg = params->set_odm_bypass_params.tg;
+	const struct dc_crtc_timing *timing = params->set_odm_bypass_params.timing;
+
+	if (tg && tg->funcs->set_odm_bypass)
+		tg->funcs->set_odm_bypass(tg, timing);
+}
+
+void hwss_opp_pipe_clock_control(union block_sequence_params *params)
+{
+	struct output_pixel_processor *opp = params->opp_pipe_clock_control_params.opp;
+	bool enable = params->opp_pipe_clock_control_params.enable;
+
+	if (opp && opp->funcs->opp_pipe_clock_control)
+		opp->funcs->opp_pipe_clock_control(opp, enable);
+}
+
+void hwss_opp_program_left_edge_extra_pixel(union block_sequence_params *params)
+{
+	struct output_pixel_processor *opp = params->opp_program_left_edge_extra_pixel_params.opp;
+	enum dc_pixel_encoding pixel_encoding = params->opp_program_left_edge_extra_pixel_params.pixel_encoding;
+	bool is_otg_master = params->opp_program_left_edge_extra_pixel_params.is_otg_master;
+
+	if (opp && opp->funcs->opp_program_left_edge_extra_pixel)
+		opp->funcs->opp_program_left_edge_extra_pixel(opp, pixel_encoding, is_otg_master);
+}
+
+void hwss_dccg_set_dto_dscclk(union block_sequence_params *params)
+{
+	struct dccg *dccg = params->dccg_set_dto_dscclk_params.dccg;
+	int inst = params->dccg_set_dto_dscclk_params.inst;
+	int num_slices_h = params->dccg_set_dto_dscclk_params.num_slices_h;
+
+	if (dccg && dccg->funcs->set_dto_dscclk)
+		dccg->funcs->set_dto_dscclk(dccg, inst, num_slices_h);
+}
+
+void hwss_dsc_set_config(union block_sequence_params *params)
+{
+	struct display_stream_compressor *dsc = params->dsc_set_config_params.dsc;
+	struct dsc_config *dsc_cfg = params->dsc_set_config_params.dsc_cfg;
+	struct dsc_optc_config *dsc_optc_cfg = params->dsc_set_config_params.dsc_optc_cfg;
+
+	if (dsc && dsc->funcs->dsc_set_config)
+		dsc->funcs->dsc_set_config(dsc, dsc_cfg, dsc_optc_cfg);
+}
+
+void hwss_dsc_enable(union block_sequence_params *params)
+{
+	struct display_stream_compressor *dsc = params->dsc_enable_params.dsc;
+	int opp_inst = params->dsc_enable_params.opp_inst;
+
+	if (dsc && dsc->funcs->dsc_enable)
+		dsc->funcs->dsc_enable(dsc, opp_inst);
+}
+
+void hwss_tg_set_dsc_config(union block_sequence_params *params)
+{
+	struct timing_generator *tg = params->tg_set_dsc_config_params.tg;
+	enum optc_dsc_mode optc_dsc_mode = OPTC_DSC_DISABLED;
+	uint32_t bytes_per_pixel = 0;
+	uint32_t slice_width = 0;
+
+	if (params->tg_set_dsc_config_params.enable) {
+		struct dsc_optc_config *dsc_optc_cfg = params->tg_set_dsc_config_params.dsc_optc_cfg;
+
+		if (dsc_optc_cfg) {
+			bytes_per_pixel = dsc_optc_cfg->bytes_per_pixel;
+			slice_width = dsc_optc_cfg->slice_width;
+			optc_dsc_mode = dsc_optc_cfg->is_pixel_format_444 ?
+				OPTC_DSC_ENABLED_444 : OPTC_DSC_ENABLED_NATIVE_SUBSAMPLED;
+		}
+	}
+
+	if (tg && tg->funcs->set_dsc_config)
+		tg->funcs->set_dsc_config(tg, optc_dsc_mode, bytes_per_pixel, slice_width);
+}
+
+void hwss_dsc_disconnect(union block_sequence_params *params)
+{
+	struct display_stream_compressor *dsc = params->dsc_disconnect_params.dsc;
+
+	if (dsc && dsc->funcs->dsc_disconnect)
+		dsc->funcs->dsc_disconnect(dsc);
+}
+
+void hwss_dsc_read_state(union block_sequence_params *params)
+{
+	struct display_stream_compressor *dsc = params->dsc_read_state_params.dsc;
+	struct dcn_dsc_state *dsc_state = params->dsc_read_state_params.dsc_state;
+
+	if (dsc && dsc->funcs->dsc_read_state)
+		dsc->funcs->dsc_read_state(dsc, dsc_state);
+}
+
+void hwss_dsc_calculate_and_set_config(union block_sequence_params *params)
+{
+	struct pipe_ctx *pipe_ctx = params->dsc_calculate_and_set_config_params.pipe_ctx;
+	struct pipe_ctx *top_pipe = pipe_ctx;
+	bool enable = params->dsc_calculate_and_set_config_params.enable;
+	int opp_cnt = params->dsc_calculate_and_set_config_params.opp_cnt;
+
+	struct display_stream_compressor *dsc = pipe_ctx->stream_res.dsc;
+	struct dc_stream_state *stream = pipe_ctx->stream;
+
+	if (!dsc || !enable)
+		return;
+
+	/* Calculate DSC configuration - extracted from dcn32_update_dsc_on_stream */
+	struct dsc_config dsc_cfg;
+
+	while (top_pipe->prev_odm_pipe)
+		top_pipe = top_pipe->prev_odm_pipe;
+
+	dsc_cfg.pic_width = (stream->timing.h_addressable + top_pipe->dsc_padding_params.dsc_hactive_padding +
+			stream->timing.h_border_left + stream->timing.h_border_right) / opp_cnt;
+	dsc_cfg.pic_height = stream->timing.v_addressable + stream->timing.v_border_top + stream->timing.v_border_bottom;
+	dsc_cfg.pixel_encoding = stream->timing.pixel_encoding;
+	dsc_cfg.color_depth = stream->timing.display_color_depth;
+	dsc_cfg.is_odm = top_pipe->next_odm_pipe ? true : false;
+	dsc_cfg.dc_dsc_cfg = stream->timing.dsc_cfg;
+	dsc_cfg.dc_dsc_cfg.num_slices_h /= opp_cnt;
+	dsc_cfg.dsc_padding = top_pipe->dsc_padding_params.dsc_hactive_padding;
+
+	/* Set DSC configuration */
+	if (dsc->funcs->dsc_set_config)
+		dsc->funcs->dsc_set_config(dsc, &dsc_cfg,
+			&params->dsc_calculate_and_set_config_params.dsc_optc_cfg);
+}
+
+void hwss_dsc_enable_with_opp(union block_sequence_params *params)
+{
+	struct pipe_ctx *pipe_ctx = params->dsc_enable_with_opp_params.pipe_ctx;
+	struct display_stream_compressor *dsc = pipe_ctx->stream_res.dsc;
+
+	if (dsc && dsc->funcs->dsc_enable)
+		dsc->funcs->dsc_enable(dsc, pipe_ctx->stream_res.opp->inst);
+}
+
+void hwss_tg_program_global_sync(union block_sequence_params *params)
+{
+	struct timing_generator *tg = params->tg_program_global_sync_params.tg;
+	int vready_offset = params->tg_program_global_sync_params.vready_offset;
+	unsigned int vstartup_lines = params->tg_program_global_sync_params.vstartup_lines;
+	unsigned int vupdate_offset_pixels = params->tg_program_global_sync_params.vupdate_offset_pixels;
+	unsigned int vupdate_vupdate_width_pixels = params->tg_program_global_sync_params.vupdate_vupdate_width_pixels;
+	unsigned int pstate_keepout_start_lines = params->tg_program_global_sync_params.pstate_keepout_start_lines;
+
+	if (tg->funcs->program_global_sync) {
+		tg->funcs->program_global_sync(tg, vready_offset, vstartup_lines,
+			vupdate_offset_pixels, vupdate_vupdate_width_pixels, pstate_keepout_start_lines);
+	}
+}
+
+void hwss_tg_wait_for_state(union block_sequence_params *params)
+{
+	struct timing_generator *tg = params->tg_wait_for_state_params.tg;
+	enum crtc_state state = params->tg_wait_for_state_params.state;
+
+	if (tg->funcs->wait_for_state)
+		tg->funcs->wait_for_state(tg, state);
+}
+
+void hwss_tg_set_vtg_params(union block_sequence_params *params)
+{
+	struct timing_generator *tg = params->tg_set_vtg_params_params.tg;
+	struct dc_crtc_timing *timing = params->tg_set_vtg_params_params.timing;
+	bool program_fp2 = params->tg_set_vtg_params_params.program_fp2;
+
+	if (tg->funcs->set_vtg_params)
+		tg->funcs->set_vtg_params(tg, timing, program_fp2);
+}
+
+void hwss_tg_setup_vertical_interrupt2(union block_sequence_params *params)
+{
+	struct timing_generator *tg = params->tg_setup_vertical_interrupt2_params.tg;
+	int start_line = params->tg_setup_vertical_interrupt2_params.start_line;
+
+	if (tg->funcs->setup_vertical_interrupt2)
+		tg->funcs->setup_vertical_interrupt2(tg, start_line);
+}
+
+void hwss_dpp_set_hdr_multiplier(union block_sequence_params *params)
+{
+	struct dpp *dpp = params->dpp_set_hdr_multiplier_params.dpp;
+	uint32_t hw_mult = params->dpp_set_hdr_multiplier_params.hw_mult;
+
+	if (dpp->funcs->dpp_set_hdr_multiplier)
+		dpp->funcs->dpp_set_hdr_multiplier(dpp, hw_mult);
+}
+
+void hwss_program_det_size(union block_sequence_params *params)
+{
+	struct hubbub *hubbub = params->program_det_size_params.hubbub;
+	unsigned int hubp_inst = params->program_det_size_params.hubp_inst;
+	unsigned int det_buffer_size_kb = params->program_det_size_params.det_buffer_size_kb;
+
+	if (hubbub->funcs->program_det_size)
+		hubbub->funcs->program_det_size(hubbub, hubp_inst, det_buffer_size_kb);
+}
+
+void hwss_program_det_segments(union block_sequence_params *params)
+{
+	struct hubbub *hubbub = params->program_det_segments_params.hubbub;
+	unsigned int hubp_inst = params->program_det_segments_params.hubp_inst;
+	unsigned int det_size = params->program_det_segments_params.det_size;
+
+	if (hubbub->funcs->program_det_segments)
+		hubbub->funcs->program_det_segments(hubbub, hubp_inst, det_size);
+}
+
+void hwss_opp_set_dyn_expansion(union block_sequence_params *params)
+{
+	struct output_pixel_processor *opp = params->opp_set_dyn_expansion_params.opp;
+	enum dc_color_space color_space = params->opp_set_dyn_expansion_params.color_space;
+	enum dc_color_depth color_depth = params->opp_set_dyn_expansion_params.color_depth;
+	enum signal_type signal = params->opp_set_dyn_expansion_params.signal;
+
+	if (opp->funcs->opp_set_dyn_expansion)
+		opp->funcs->opp_set_dyn_expansion(opp, color_space, color_depth, signal);
+}
+
+void hwss_opp_program_fmt(union block_sequence_params *params)
+{
+	struct output_pixel_processor *opp = params->opp_program_fmt_params.opp;
+	struct bit_depth_reduction_params *fmt_bit_depth = params->opp_program_fmt_params.fmt_bit_depth;
+	struct clamping_and_pixel_encoding_params *clamping = params->opp_program_fmt_params.clamping;
+
+	if (opp->funcs->opp_program_fmt)
+		opp->funcs->opp_program_fmt(opp, fmt_bit_depth, clamping);
+}
+
+void hwss_opp_program_bit_depth_reduction(union block_sequence_params *params)
+{
+	struct output_pixel_processor *opp = params->opp_program_bit_depth_reduction_params.opp;
+	bool use_default_params = params->opp_program_bit_depth_reduction_params.use_default_params;
+	struct pipe_ctx *pipe_ctx = params->opp_program_bit_depth_reduction_params.pipe_ctx;
+	struct bit_depth_reduction_params bit_depth_params;
+
+	if (use_default_params)
+		memset(&bit_depth_params, 0, sizeof(bit_depth_params));
+	else
+		resource_build_bit_depth_reduction_params(pipe_ctx->stream, &bit_depth_params);
+
+	if (opp->funcs->opp_program_bit_depth_reduction)
+		opp->funcs->opp_program_bit_depth_reduction(opp, &bit_depth_params);
+}
+
+void hwss_opp_set_disp_pattern_generator(union block_sequence_params *params)
+{
+	struct output_pixel_processor *opp = params->opp_set_disp_pattern_generator_params.opp;
+	enum controller_dp_test_pattern test_pattern = params->opp_set_disp_pattern_generator_params.test_pattern;
+	enum controller_dp_color_space color_space = params->opp_set_disp_pattern_generator_params.color_space;
+	enum dc_color_depth color_depth = params->opp_set_disp_pattern_generator_params.color_depth;
+	struct tg_color *solid_color = params->opp_set_disp_pattern_generator_params.use_solid_color ?
+		&params->opp_set_disp_pattern_generator_params.solid_color : NULL;
+	int width = params->opp_set_disp_pattern_generator_params.width;
+	int height = params->opp_set_disp_pattern_generator_params.height;
+	int offset = params->opp_set_disp_pattern_generator_params.offset;
+
+	if (opp && opp->funcs->opp_set_disp_pattern_generator) {
+		opp->funcs->opp_set_disp_pattern_generator(opp, test_pattern, color_space,
+			color_depth, solid_color, width, height, offset);
+	}
+}
+
+void hwss_set_abm_pipe(union block_sequence_params *params)
+{
+	struct dc *dc = params->set_abm_pipe_params.dc;
+	struct pipe_ctx *pipe_ctx = params->set_abm_pipe_params.pipe_ctx;
+
+	dc->hwss.set_pipe(pipe_ctx);
+}
+
+void hwss_set_abm_level(union block_sequence_params *params)
+{
+	struct abm *abm = params->set_abm_level_params.abm;
+	unsigned int abm_level = params->set_abm_level_params.abm_level;
+
+	if (abm->funcs->set_abm_level)
+		abm->funcs->set_abm_level(abm, abm_level);
+}
+
+void hwss_set_abm_immediate_disable(union block_sequence_params *params)
+{
+	struct dc *dc = params->set_abm_immediate_disable_params.dc;
+	struct pipe_ctx *pipe_ctx = params->set_abm_immediate_disable_params.pipe_ctx;
+
+	if (dc && dc->hwss.set_abm_immediate_disable)
+		dc->hwss.set_abm_immediate_disable(pipe_ctx);
+}
+
+void hwss_mpc_remove_mpcc(union block_sequence_params *params)
+{
+	struct mpc *mpc = params->mpc_remove_mpcc_params.mpc;
+	struct mpc_tree *mpc_tree_params = params->mpc_remove_mpcc_params.mpc_tree_params;
+	struct mpcc *mpcc_to_remove = params->mpc_remove_mpcc_params.mpcc_to_remove;
+
+	mpc->funcs->remove_mpcc(mpc, mpc_tree_params, mpcc_to_remove);
+}
+
+void hwss_opp_set_mpcc_disconnect_pending(union block_sequence_params *params)
+{
+	struct output_pixel_processor *opp = params->opp_set_mpcc_disconnect_pending_params.opp;
+	int mpcc_inst = params->opp_set_mpcc_disconnect_pending_params.mpcc_inst;
+	bool pending = params->opp_set_mpcc_disconnect_pending_params.pending;
+
+	opp->mpcc_disconnect_pending[mpcc_inst] = pending;
+}
+
+void hwss_dc_set_optimized_required(union block_sequence_params *params)
+{
+	struct dc *dc = params->dc_set_optimized_required_params.dc;
+	bool optimized_required = params->dc_set_optimized_required_params.optimized_required;
+
+	dc->optimized_required = optimized_required;
+}
+
+void hwss_hubp_disconnect(union block_sequence_params *params)
+{
+	struct hubp *hubp = params->hubp_disconnect_params.hubp;
+
+	if (hubp->funcs->hubp_disconnect)
+		hubp->funcs->hubp_disconnect(hubp);
+}
+
+void hwss_hubbub_force_pstate_change_control(union block_sequence_params *params)
+{
+	struct hubbub *hubbub = params->hubbub_force_pstate_change_control_params.hubbub;
+	bool enable = params->hubbub_force_pstate_change_control_params.enable;
+	bool wait = params->hubbub_force_pstate_change_control_params.wait;
+
+	if (hubbub->funcs->force_pstate_change_control) {
+		hubbub->funcs->force_pstate_change_control(hubbub, enable, wait);
+		/* Add delay when enabling pstate change control */
+		if (enable)
+			udelay(500);
+	}
+}
+
+void hwss_tg_enable_crtc(union block_sequence_params *params)
+{
+	struct timing_generator *tg = params->tg_enable_crtc_params.tg;
+
+	if (tg->funcs->enable_crtc)
+		tg->funcs->enable_crtc(tg);
+}
+
+void hwss_tg_set_gsl(union block_sequence_params *params)
+{
+	struct timing_generator *tg = params->tg_set_gsl_params.tg;
+	struct gsl_params *gsl = &params->tg_set_gsl_params.gsl;
+
+	if (tg->funcs->set_gsl)
+		tg->funcs->set_gsl(tg, gsl);
+}
+
+void hwss_tg_set_gsl_source_select(union block_sequence_params *params)
+{
+	struct timing_generator *tg = params->tg_set_gsl_source_select_params.tg;
+	int group_idx = params->tg_set_gsl_source_select_params.group_idx;
+	uint32_t gsl_ready_signal = params->tg_set_gsl_source_select_params.gsl_ready_signal;
+
+	if (tg->funcs->set_gsl_source_select)
+		tg->funcs->set_gsl_source_select(tg, group_idx, gsl_ready_signal);
+}
+
+void hwss_hubp_wait_flip_pending(union block_sequence_params *params)
+{
+	struct hubp *hubp = params->hubp_wait_flip_pending_params.hubp;
+	unsigned int timeout_us = params->hubp_wait_flip_pending_params.timeout_us;
+	unsigned int polling_interval_us = params->hubp_wait_flip_pending_params.polling_interval_us;
+	int j = 0;
+
+	for (j = 0; j < timeout_us / polling_interval_us
+		&& hubp->funcs->hubp_is_flip_pending(hubp); j++)
+		udelay(polling_interval_us);
+}
+
+void hwss_tg_wait_double_buffer_pending(union block_sequence_params *params)
+{
+	struct timing_generator *tg = params->tg_wait_double_buffer_pending_params.tg;
+	unsigned int timeout_us = params->tg_wait_double_buffer_pending_params.timeout_us;
+	unsigned int polling_interval_us = params->tg_wait_double_buffer_pending_params.polling_interval_us;
+	int j = 0;
+
+	if (tg->funcs->get_optc_double_buffer_pending) {
+		for (j = 0; j < timeout_us / polling_interval_us
+			&& tg->funcs->get_optc_double_buffer_pending(tg); j++)
+			udelay(polling_interval_us);
+	}
+}
+
+void hwss_update_force_pstate(union block_sequence_params *params)
+{
+	struct dc *dc = params->update_force_pstate_params.dc;
+	struct dc_state *context = params->update_force_pstate_params.context;
+	struct dce_hwseq *hwseq = dc->hwseq;
+
+	if (hwseq->funcs.update_force_pstate)
+		hwseq->funcs.update_force_pstate(dc, context);
+}
+
+void hwss_hubbub_apply_dedcn21_147_wa(union block_sequence_params *params)
+{
+	struct hubbub *hubbub = params->hubbub_apply_dedcn21_147_wa_params.hubbub;
+
+	hubbub->funcs->apply_DEDCN21_147_wa(hubbub);
+}
+
+void hwss_hubbub_allow_self_refresh_control(union block_sequence_params *params)
+{
+	struct hubbub *hubbub = params->hubbub_allow_self_refresh_control_params.hubbub;
+	bool allow = params->hubbub_allow_self_refresh_control_params.allow;
+
+	hubbub->funcs->allow_self_refresh_control(hubbub, allow);
+
+	if (!allow && params->hubbub_allow_self_refresh_control_params.disallow_self_refresh_applied)
+		*params->hubbub_allow_self_refresh_control_params.disallow_self_refresh_applied = true;
+}
+
+void hwss_tg_get_frame_count(union block_sequence_params *params)
+{
+	struct timing_generator *tg = params->tg_get_frame_count_params.tg;
+	unsigned int *frame_count = params->tg_get_frame_count_params.frame_count;
+
+	*frame_count = tg->funcs->get_frame_count(tg);
+}
+
+void hwss_mpc_set_dwb_mux(union block_sequence_params *params)
+{
+	struct mpc *mpc = params->mpc_set_dwb_mux_params.mpc;
+	int dwb_id = params->mpc_set_dwb_mux_params.dwb_id;
+	int mpcc_id = params->mpc_set_dwb_mux_params.mpcc_id;
+
+	if (mpc->funcs->set_dwb_mux)
+		mpc->funcs->set_dwb_mux(mpc, dwb_id, mpcc_id);
+}
+
+void hwss_mpc_disable_dwb_mux(union block_sequence_params *params)
+{
+	struct mpc *mpc = params->mpc_disable_dwb_mux_params.mpc;
+	unsigned int dwb_id = params->mpc_disable_dwb_mux_params.dwb_id;
+
+	if (mpc->funcs->disable_dwb_mux)
+		mpc->funcs->disable_dwb_mux(mpc, dwb_id);
+}
+
+void hwss_mcif_wb_config_buf(union block_sequence_params *params)
+{
+	struct mcif_wb *mcif_wb = params->mcif_wb_config_buf_params.mcif_wb;
+	struct mcif_buf_params *mcif_buf_params = params->mcif_wb_config_buf_params.mcif_buf_params;
+	unsigned int dest_height = params->mcif_wb_config_buf_params.dest_height;
+
+	if (mcif_wb->funcs->config_mcif_buf)
+		mcif_wb->funcs->config_mcif_buf(mcif_wb, mcif_buf_params, dest_height);
+}
+
+void hwss_mcif_wb_config_arb(union block_sequence_params *params)
+{
+	struct mcif_wb *mcif_wb = params->mcif_wb_config_arb_params.mcif_wb;
+	struct mcif_arb_params *mcif_arb_params = params->mcif_wb_config_arb_params.mcif_arb_params;
+
+	if (mcif_wb->funcs->config_mcif_arb)
+		mcif_wb->funcs->config_mcif_arb(mcif_wb, mcif_arb_params);
+}
+
+void hwss_mcif_wb_enable(union block_sequence_params *params)
+{
+	struct mcif_wb *mcif_wb = params->mcif_wb_enable_params.mcif_wb;
+
+	if (mcif_wb->funcs->enable_mcif)
+		mcif_wb->funcs->enable_mcif(mcif_wb);
+}
+
+void hwss_mcif_wb_disable(union block_sequence_params *params)
+{
+	struct mcif_wb *mcif_wb = params->mcif_wb_disable_params.mcif_wb;
+
+	if (mcif_wb->funcs->disable_mcif)
+		mcif_wb->funcs->disable_mcif(mcif_wb);
+}
+
+void hwss_dwbc_enable(union block_sequence_params *params)
+{
+	struct dwbc *dwb = params->dwbc_enable_params.dwb;
+	struct dc_dwb_params *dwb_params = params->dwbc_enable_params.dwb_params;
+
+	if (dwb->funcs->enable)
+		dwb->funcs->enable(dwb, dwb_params);
+}
+
+void hwss_dwbc_disable(union block_sequence_params *params)
+{
+	struct dwbc *dwb = params->dwbc_disable_params.dwb;
+
+	if (dwb->funcs->disable)
+		dwb->funcs->disable(dwb);
+}
+
+void hwss_dwbc_update(union block_sequence_params *params)
+{
+	struct dwbc *dwb = params->dwbc_update_params.dwb;
+	struct dc_dwb_params *dwb_params = params->dwbc_update_params.dwb_params;
+
+	if (dwb->funcs->update)
+		dwb->funcs->update(dwb, dwb_params);
+}
+
+void hwss_hubp_update_mall_sel(union block_sequence_params *params)
+{
+	struct hubp *hubp = params->hubp_update_mall_sel_params.hubp;
+	uint32_t mall_sel = params->hubp_update_mall_sel_params.mall_sel;
+	bool cache_cursor = params->hubp_update_mall_sel_params.cache_cursor;
+
+	if (hubp && hubp->funcs->hubp_update_mall_sel)
+		hubp->funcs->hubp_update_mall_sel(hubp, mall_sel, cache_cursor);
+}
+
+void hwss_hubp_prepare_subvp_buffering(union block_sequence_params *params)
+{
+	struct hubp *hubp = params->hubp_prepare_subvp_buffering_params.hubp;
+	bool enable = params->hubp_prepare_subvp_buffering_params.enable;
+
+	if (hubp && hubp->funcs->hubp_prepare_subvp_buffering)
+		hubp->funcs->hubp_prepare_subvp_buffering(hubp, enable);
+}
+
+void hwss_hubp_set_blank_en(union block_sequence_params *params)
+{
+	struct hubp *hubp = params->hubp_set_blank_en_params.hubp;
+	bool enable = params->hubp_set_blank_en_params.enable;
+
+	if (hubp && hubp->funcs->set_hubp_blank_en)
+		hubp->funcs->set_hubp_blank_en(hubp, enable);
+}
+
+void hwss_hubp_disable_control(union block_sequence_params *params)
+{
+	struct hubp *hubp = params->hubp_disable_control_params.hubp;
+	bool disable = params->hubp_disable_control_params.disable;
+
+	if (hubp && hubp->funcs->hubp_disable_control)
+		hubp->funcs->hubp_disable_control(hubp, disable);
+}
+
+void hwss_hubbub_soft_reset(union block_sequence_params *params)
+{
+	struct hubbub *hubbub = params->hubbub_soft_reset_params.hubbub;
+	bool reset = params->hubbub_soft_reset_params.reset;
+
+	if (hubbub)
+		params->hubbub_soft_reset_params.hubbub_soft_reset(hubbub, reset);
+}
+
+void hwss_hubp_clk_cntl(union block_sequence_params *params)
+{
+	struct hubp *hubp = params->hubp_clk_cntl_params.hubp;
+	bool enable = params->hubp_clk_cntl_params.enable;
+
+	if (hubp && hubp->funcs->hubp_clk_cntl) {
+		hubp->funcs->hubp_clk_cntl(hubp, enable);
+		hubp->power_gated = !enable;
+	}
+}
+
+void hwss_hubp_init(union block_sequence_params *params)
+{
+	struct hubp *hubp = params->hubp_init_params.hubp;
+
+	if (hubp && hubp->funcs->hubp_init)
+		hubp->funcs->hubp_init(hubp);
+}
+
+void hwss_hubp_set_vm_system_aperture_settings(union block_sequence_params *params)
+{
+	struct hubp *hubp = params->hubp_set_vm_system_aperture_settings_params.hubp;
+	struct vm_system_aperture_param apt;
+
+	apt.sys_default = params->hubp_set_vm_system_aperture_settings_params.sys_default;
+	apt.sys_high = params->hubp_set_vm_system_aperture_settings_params.sys_high;
+	apt.sys_low = params->hubp_set_vm_system_aperture_settings_params.sys_low;
+
+	if (hubp && hubp->funcs->hubp_set_vm_system_aperture_settings)
+		hubp->funcs->hubp_set_vm_system_aperture_settings(hubp, &apt);
+}
+
+void hwss_hubp_set_flip_int(union block_sequence_params *params)
+{
+	struct hubp *hubp = params->hubp_set_flip_int_params.hubp;
+
+	if (hubp && hubp->funcs->hubp_set_flip_int)
+		hubp->funcs->hubp_set_flip_int(hubp);
+}
+
+void hwss_dpp_dppclk_control(union block_sequence_params *params)
+{
+	struct dpp *dpp = params->dpp_dppclk_control_params.dpp;
+	bool dppclk_div = params->dpp_dppclk_control_params.dppclk_div;
+	bool enable = params->dpp_dppclk_control_params.enable;
+
+	if (dpp && dpp->funcs->dpp_dppclk_control)
+		dpp->funcs->dpp_dppclk_control(dpp, dppclk_div, enable);
+}
+
+void hwss_disable_phantom_crtc(union block_sequence_params *params)
+{
+	struct timing_generator *tg = params->disable_phantom_crtc_params.tg;
+
+	if (tg && tg->funcs->disable_phantom_crtc)
+		tg->funcs->disable_phantom_crtc(tg);
+}
+
+void hwss_dsc_pg_status(union block_sequence_params *params)
+{
+	struct dce_hwseq *hws = params->dsc_pg_status_params.hws;
+	int dsc_inst = params->dsc_pg_status_params.dsc_inst;
+
+	if (hws && hws->funcs.dsc_pg_status)
+		params->dsc_pg_status_params.is_ungated = hws->funcs.dsc_pg_status(hws, dsc_inst);
+}
+
+void hwss_dsc_wait_disconnect_pending_clear(union block_sequence_params *params)
+{
+	struct display_stream_compressor *dsc = params->dsc_wait_disconnect_pending_clear_params.dsc;
+
+	if (!params->dsc_wait_disconnect_pending_clear_params.is_ungated)
+		return;
+	if (*params->dsc_wait_disconnect_pending_clear_params.is_ungated == false)
+		return;
+
+	if (dsc && dsc->funcs->dsc_wait_disconnect_pending_clear)
+		dsc->funcs->dsc_wait_disconnect_pending_clear(dsc);
+}
+
+void hwss_dsc_disable(union block_sequence_params *params)
+{
+	struct display_stream_compressor *dsc = params->dsc_disable_params.dsc;
+
+	if (!params->dsc_disable_params.is_ungated)
+		return;
+	if (*params->dsc_disable_params.is_ungated == false)
+		return;
+
+	if (dsc && dsc->funcs->dsc_disable)
+		dsc->funcs->dsc_disable(dsc);
+}
+
+void hwss_dccg_set_ref_dscclk(union block_sequence_params *params)
+{
+	struct dccg *dccg = params->dccg_set_ref_dscclk_params.dccg;
+	int dsc_inst = params->dccg_set_ref_dscclk_params.dsc_inst;
+
+	if (!params->dccg_set_ref_dscclk_params.is_ungated)
+		return;
+	if (*params->dccg_set_ref_dscclk_params.is_ungated == false)
+		return;
+
+	if (dccg && dccg->funcs->set_ref_dscclk)
+		dccg->funcs->set_ref_dscclk(dccg, dsc_inst);
+}
+
+void hwss_dpp_pg_control(union block_sequence_params *params)
+{
+	struct dce_hwseq *hws = params->dpp_pg_control_params.hws;
+	unsigned int dpp_inst = params->dpp_pg_control_params.dpp_inst;
+	bool power_on = params->dpp_pg_control_params.power_on;
+
+	if (hws->funcs.dpp_pg_control)
+		hws->funcs.dpp_pg_control(hws, dpp_inst, power_on);
+}
+
+void hwss_hubp_pg_control(union block_sequence_params *params)
+{
+	struct dce_hwseq *hws = params->hubp_pg_control_params.hws;
+	unsigned int hubp_inst = params->hubp_pg_control_params.hubp_inst;
+	bool power_on = params->hubp_pg_control_params.power_on;
+
+	if (hws->funcs.hubp_pg_control)
+		hws->funcs.hubp_pg_control(hws, hubp_inst, power_on);
+}
+
+void hwss_hubp_reset(union block_sequence_params *params)
+{
+	struct hubp *hubp = params->hubp_reset_params.hubp;
+
+	if (hubp && hubp->funcs->hubp_reset)
+		hubp->funcs->hubp_reset(hubp);
+}
+
+void hwss_dpp_reset(union block_sequence_params *params)
+{
+	struct dpp *dpp = params->dpp_reset_params.dpp;
+
+	if (dpp && dpp->funcs->dpp_reset)
+		dpp->funcs->dpp_reset(dpp);
+}
+
+void hwss_dpp_root_clock_control(union block_sequence_params *params)
+{
+	struct dce_hwseq *hws = params->dpp_root_clock_control_params.hws;
+	unsigned int dpp_inst = params->dpp_root_clock_control_params.dpp_inst;
+	bool clock_on = params->dpp_root_clock_control_params.clock_on;
+
+	if (hws->funcs.dpp_root_clock_control)
+		hws->funcs.dpp_root_clock_control(hws, dpp_inst, clock_on);
+}
+
+void hwss_dc_ip_request_cntl(union block_sequence_params *params)
+{
+	struct dc *dc = params->dc_ip_request_cntl_params.dc;
+	bool enable = params->dc_ip_request_cntl_params.enable;
+	struct dce_hwseq *hws = dc->hwseq;
+
+	if (hws->funcs.dc_ip_request_cntl)
+		hws->funcs.dc_ip_request_cntl(dc, enable);
+}
+
+void hwss_dccg_update_dpp_dto(union block_sequence_params *params)
+{
+	struct dccg *dccg = params->dccg_update_dpp_dto_params.dccg;
+	int dpp_inst = params->dccg_update_dpp_dto_params.dpp_inst;
+	int dppclk_khz = params->dccg_update_dpp_dto_params.dppclk_khz;
+
+	if (dccg && dccg->funcs->update_dpp_dto)
+		dccg->funcs->update_dpp_dto(dccg, dpp_inst, dppclk_khz);
+}
+
+void hwss_hubp_vtg_sel(union block_sequence_params *params)
+{
+	struct hubp *hubp = params->hubp_vtg_sel_params.hubp;
+	uint32_t otg_inst = params->hubp_vtg_sel_params.otg_inst;
+
+	if (hubp && hubp->funcs->hubp_vtg_sel)
+		hubp->funcs->hubp_vtg_sel(hubp, otg_inst);
+}
+
+void hwss_hubp_setup2(union block_sequence_params *params)
+{
+	struct hubp *hubp = params->hubp_setup2_params.hubp;
+	struct dml2_dchub_per_pipe_register_set *hubp_regs = params->hubp_setup2_params.hubp_regs;
+	union dml2_global_sync_programming *global_sync = params->hubp_setup2_params.global_sync;
+	struct dc_crtc_timing *timing = params->hubp_setup2_params.timing;
+
+	if (hubp && hubp->funcs->hubp_setup2)
+		hubp->funcs->hubp_setup2(hubp, hubp_regs, global_sync, timing);
+}
+
+void hwss_hubp_setup(union block_sequence_params *params)
+{
+	struct hubp *hubp = params->hubp_setup_params.hubp;
+	struct _vcs_dpi_display_dlg_regs_st *dlg_regs = params->hubp_setup_params.dlg_regs;
+	struct _vcs_dpi_display_ttu_regs_st *ttu_regs = params->hubp_setup_params.ttu_regs;
+	struct _vcs_dpi_display_rq_regs_st *rq_regs = params->hubp_setup_params.rq_regs;
+	struct _vcs_dpi_display_pipe_dest_params_st *pipe_dest = params->hubp_setup_params.pipe_dest;
+
+	if (hubp && hubp->funcs->hubp_setup)
+		hubp->funcs->hubp_setup(hubp, dlg_regs, ttu_regs, rq_regs, pipe_dest);
+}
+
+void hwss_hubp_set_unbounded_requesting(union block_sequence_params *params)
+{
+	struct hubp *hubp = params->hubp_set_unbounded_requesting_params.hubp;
+	bool unbounded_req = params->hubp_set_unbounded_requesting_params.unbounded_req;
+
+	if (hubp && hubp->funcs->set_unbounded_requesting)
+		hubp->funcs->set_unbounded_requesting(hubp, unbounded_req);
+}
+
+void hwss_hubp_setup_interdependent2(union block_sequence_params *params)
+{
+	struct hubp *hubp = params->hubp_setup_interdependent2_params.hubp;
+	struct dml2_dchub_per_pipe_register_set *hubp_regs = params->hubp_setup_interdependent2_params.hubp_regs;
+
+	if (hubp && hubp->funcs->hubp_setup_interdependent2)
+		hubp->funcs->hubp_setup_interdependent2(hubp, hubp_regs);
+}
+
+void hwss_hubp_setup_interdependent(union block_sequence_params *params)
+{
+	struct hubp *hubp = params->hubp_setup_interdependent_params.hubp;
+	struct _vcs_dpi_display_dlg_regs_st *dlg_regs = params->hubp_setup_interdependent_params.dlg_regs;
+	struct _vcs_dpi_display_ttu_regs_st *ttu_regs = params->hubp_setup_interdependent_params.ttu_regs;
+
+	if (hubp && hubp->funcs->hubp_setup_interdependent)
+		hubp->funcs->hubp_setup_interdependent(hubp, dlg_regs, ttu_regs);
+}
+
+void hwss_dpp_set_cursor_matrix(union block_sequence_params *params)
+{
+	struct dpp *dpp = params->dpp_set_cursor_matrix_params.dpp;
+	enum dc_color_space color_space = params->dpp_set_cursor_matrix_params.color_space;
+	struct dc_csc_transform *cursor_csc_color_matrix = params->dpp_set_cursor_matrix_params.cursor_csc_color_matrix;
+
+	if (dpp && dpp->funcs->set_cursor_matrix)
+		dpp->funcs->set_cursor_matrix(dpp, color_space, *cursor_csc_color_matrix);
+}
+
+void hwss_mpc_update_mpcc(union block_sequence_params *params)
+{
+	struct dc *dc = params->mpc_update_mpcc_params.dc;
+	struct pipe_ctx *pipe_ctx = params->mpc_update_mpcc_params.pipe_ctx;
+	struct dce_hwseq *hws = dc->hwseq;
+
+	if (hws->funcs.update_mpcc)
+		hws->funcs.update_mpcc(dc, pipe_ctx);
+}
+
+void hwss_mpc_update_blending(union block_sequence_params *params)
+{
+	struct mpc *mpc = params->mpc_update_blending_params.mpc;
+	struct mpcc_blnd_cfg *blnd_cfg = &params->mpc_update_blending_params.blnd_cfg;
+	int mpcc_id = params->mpc_update_blending_params.mpcc_id;
+
+	if (mpc && mpc->funcs->update_blending)
+		mpc->funcs->update_blending(mpc, blnd_cfg, mpcc_id);
+}
+
+void hwss_mpc_assert_idle_mpcc(union block_sequence_params *params)
+{
+	struct mpc *mpc = params->mpc_assert_idle_mpcc_params.mpc;
+	int mpcc_id = params->mpc_assert_idle_mpcc_params.mpcc_id;
+
+	if (mpc && mpc->funcs->wait_for_idle)
+		mpc->funcs->wait_for_idle(mpc, mpcc_id);
+}
+
+void hwss_mpc_insert_plane(union block_sequence_params *params)
+{
+	struct mpc *mpc = params->mpc_insert_plane_params.mpc;
+	struct mpc_tree *tree = params->mpc_insert_plane_params.mpc_tree_params;
+	struct mpcc_blnd_cfg *blnd_cfg = &params->mpc_insert_plane_params.blnd_cfg;
+	struct mpcc_sm_cfg *sm_cfg = params->mpc_insert_plane_params.sm_cfg;
+	struct mpcc *insert_above_mpcc = params->mpc_insert_plane_params.insert_above_mpcc;
+	int mpcc_id = params->mpc_insert_plane_params.mpcc_id;
+	int dpp_id = params->mpc_insert_plane_params.dpp_id;
+
+	if (mpc && mpc->funcs->insert_plane)
+		mpc->funcs->insert_plane(mpc, tree, blnd_cfg, sm_cfg, insert_above_mpcc,
+			dpp_id, mpcc_id);
+}
+
+void hwss_dpp_set_scaler(union block_sequence_params *params)
+{
+	struct dpp *dpp = params->dpp_set_scaler_params.dpp;
+	const struct scaler_data *scl_data = params->dpp_set_scaler_params.scl_data;
+
+	if (dpp && dpp->funcs->dpp_set_scaler)
+		dpp->funcs->dpp_set_scaler(dpp, scl_data);
+}
+
+void hwss_hubp_mem_program_viewport(union block_sequence_params *params)
+{
+	struct hubp *hubp = params->hubp_mem_program_viewport_params.hubp;
+	const struct rect *viewport = params->hubp_mem_program_viewport_params.viewport;
+	const struct rect *viewport_c = params->hubp_mem_program_viewport_params.viewport_c;
+
+	if (hubp && hubp->funcs->mem_program_viewport)
+		hubp->funcs->mem_program_viewport(hubp, viewport, viewport_c);
+}
+
+void hwss_abort_cursor_offload_update(union block_sequence_params *params)
+{
+	struct dc *dc = params->abort_cursor_offload_update_params.dc;
+	struct pipe_ctx *pipe_ctx = params->abort_cursor_offload_update_params.pipe_ctx;
+
+	if (dc && dc->hwss.abort_cursor_offload_update)
+		dc->hwss.abort_cursor_offload_update(dc, pipe_ctx);
+}
+
+void hwss_set_cursor_attribute(union block_sequence_params *params)
+{
+	struct dc *dc = params->set_cursor_attribute_params.dc;
+	struct pipe_ctx *pipe_ctx = params->set_cursor_attribute_params.pipe_ctx;
+
+	if (dc && dc->hwss.set_cursor_attribute)
+		dc->hwss.set_cursor_attribute(pipe_ctx);
+}
+
+void hwss_set_cursor_position(union block_sequence_params *params)
+{
+	struct dc *dc = params->set_cursor_position_params.dc;
+	struct pipe_ctx *pipe_ctx = params->set_cursor_position_params.pipe_ctx;
+
+	if (dc && dc->hwss.set_cursor_position)
+		dc->hwss.set_cursor_position(pipe_ctx);
+}
+
+void hwss_set_cursor_sdr_white_level(union block_sequence_params *params)
+{
+	struct dc *dc = params->set_cursor_sdr_white_level_params.dc;
+	struct pipe_ctx *pipe_ctx = params->set_cursor_sdr_white_level_params.pipe_ctx;
+
+	if (dc && dc->hwss.set_cursor_sdr_white_level)
+		dc->hwss.set_cursor_sdr_white_level(pipe_ctx);
+}
+
+void hwss_program_output_csc(union block_sequence_params *params)
+{
+	struct dc *dc = params->program_output_csc_params.dc;
+	struct pipe_ctx *pipe_ctx = params->program_output_csc_params.pipe_ctx;
+	enum dc_color_space colorspace = params->program_output_csc_params.colorspace;
+	uint16_t *matrix = params->program_output_csc_params.matrix;
+	int opp_id = params->program_output_csc_params.opp_id;
+
+	if (dc && dc->hwss.program_output_csc)
+		dc->hwss.program_output_csc(dc, pipe_ctx, colorspace, matrix, opp_id);
+}
+
+void hwss_hubp_set_blank(union block_sequence_params *params)
+{
+	struct hubp *hubp = params->hubp_set_blank_params.hubp;
+	bool blank = params->hubp_set_blank_params.blank;
+
+	if (hubp && hubp->funcs->set_blank)
+		hubp->funcs->set_blank(hubp, blank);
+}
+
+void hwss_phantom_hubp_post_enable(union block_sequence_params *params)
+{
+	struct hubp *hubp = params->phantom_hubp_post_enable_params.hubp;
+
+	if (hubp && hubp->funcs->phantom_hubp_post_enable)
+		hubp->funcs->phantom_hubp_post_enable(hubp);
+}
+
+void hwss_add_dccg_set_dto_dscclk(struct block_sequence_state *seq_state,
+		struct dccg *dccg, int inst, int num_slices_h)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = DCCG_SET_DTO_DSCCLK;
+		seq_state->steps[*seq_state->num_steps].params.dccg_set_dto_dscclk_params.dccg = dccg;
+		seq_state->steps[*seq_state->num_steps].params.dccg_set_dto_dscclk_params.inst = inst;
+		seq_state->steps[*seq_state->num_steps].params.dccg_set_dto_dscclk_params.num_slices_h = num_slices_h;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_dsc_calculate_and_set_config(struct block_sequence_state *seq_state,
+		struct pipe_ctx *pipe_ctx, bool enable, int opp_cnt)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = DSC_CALCULATE_AND_SET_CONFIG;
+		seq_state->steps[*seq_state->num_steps].params.dsc_calculate_and_set_config_params.pipe_ctx = pipe_ctx;
+		seq_state->steps[*seq_state->num_steps].params.dsc_calculate_and_set_config_params.enable = enable;
+		seq_state->steps[*seq_state->num_steps].params.dsc_calculate_and_set_config_params.opp_cnt = opp_cnt;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_mpc_remove_mpcc(struct block_sequence_state *seq_state,
+		struct mpc *mpc, struct mpc_tree *mpc_tree_params, struct mpcc *mpcc_to_remove)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = MPC_REMOVE_MPCC;
+		seq_state->steps[*seq_state->num_steps].params.mpc_remove_mpcc_params.mpc = mpc;
+		seq_state->steps[*seq_state->num_steps].params.mpc_remove_mpcc_params.mpc_tree_params = mpc_tree_params;
+		seq_state->steps[*seq_state->num_steps].params.mpc_remove_mpcc_params.mpcc_to_remove = mpcc_to_remove;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_opp_set_mpcc_disconnect_pending(struct block_sequence_state *seq_state,
+		struct output_pixel_processor *opp, int mpcc_inst, bool pending)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = OPP_SET_MPCC_DISCONNECT_PENDING;
+		seq_state->steps[*seq_state->num_steps].params.opp_set_mpcc_disconnect_pending_params.opp = opp;
+		seq_state->steps[*seq_state->num_steps].params.opp_set_mpcc_disconnect_pending_params.mpcc_inst = mpcc_inst;
+		seq_state->steps[*seq_state->num_steps].params.opp_set_mpcc_disconnect_pending_params.pending = pending;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_hubp_disconnect(struct block_sequence_state *seq_state,
+		struct hubp *hubp)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = HUBP_DISCONNECT;
+		seq_state->steps[*seq_state->num_steps].params.hubp_disconnect_params.hubp = hubp;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_dsc_enable_with_opp(struct block_sequence_state *seq_state,
+		struct pipe_ctx *pipe_ctx)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = DSC_ENABLE_WITH_OPP;
+		seq_state->steps[*seq_state->num_steps].params.dsc_enable_with_opp_params.pipe_ctx = pipe_ctx;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_tg_set_dsc_config(struct block_sequence_state *seq_state,
+		struct timing_generator *tg, struct dsc_optc_config *dsc_optc_cfg, bool enable)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = TG_SET_DSC_CONFIG;
+		seq_state->steps[*seq_state->num_steps].params.tg_set_dsc_config_params.tg = tg;
+		seq_state->steps[*seq_state->num_steps].params.tg_set_dsc_config_params.dsc_optc_cfg = dsc_optc_cfg;
+		seq_state->steps[*seq_state->num_steps].params.tg_set_dsc_config_params.enable = enable;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_dsc_disconnect(struct block_sequence_state *seq_state,
+		struct display_stream_compressor *dsc)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = DSC_DISCONNECT;
+		seq_state->steps[*seq_state->num_steps].params.dsc_disconnect_params.dsc = dsc;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_dc_set_optimized_required(struct block_sequence_state *seq_state,
+		struct dc *dc, bool optimized_required)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = DC_SET_OPTIMIZED_REQUIRED;
+		seq_state->steps[*seq_state->num_steps].params.dc_set_optimized_required_params.dc = dc;
+		seq_state->steps[*seq_state->num_steps].params.dc_set_optimized_required_params.optimized_required = optimized_required;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_abm_set_immediate_disable(struct block_sequence_state *seq_state,
+		struct dc *dc, struct pipe_ctx *pipe_ctx)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = ABM_SET_IMMEDIATE_DISABLE;
+		seq_state->steps[*seq_state->num_steps].params.set_abm_immediate_disable_params.dc = dc;
+		seq_state->steps[*seq_state->num_steps].params.set_abm_immediate_disable_params.pipe_ctx = pipe_ctx;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_opp_set_disp_pattern_generator(struct block_sequence_state *seq_state,
+		struct output_pixel_processor *opp,
+		enum controller_dp_test_pattern test_pattern,
+		enum controller_dp_color_space color_space,
+		enum dc_color_depth color_depth,
+		struct tg_color solid_color,
+		bool use_solid_color,
+		int width,
+		int height,
+		int offset)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = OPP_SET_DISP_PATTERN_GENERATOR;
+		seq_state->steps[*seq_state->num_steps].params.opp_set_disp_pattern_generator_params.opp = opp;
+		seq_state->steps[*seq_state->num_steps].params.opp_set_disp_pattern_generator_params.test_pattern = test_pattern;
+		seq_state->steps[*seq_state->num_steps].params.opp_set_disp_pattern_generator_params.color_space = color_space;
+		seq_state->steps[*seq_state->num_steps].params.opp_set_disp_pattern_generator_params.color_depth = color_depth;
+		seq_state->steps[*seq_state->num_steps].params.opp_set_disp_pattern_generator_params.solid_color = solid_color;
+		seq_state->steps[*seq_state->num_steps].params.opp_set_disp_pattern_generator_params.use_solid_color = use_solid_color;
+		seq_state->steps[*seq_state->num_steps].params.opp_set_disp_pattern_generator_params.width = width;
+		seq_state->steps[*seq_state->num_steps].params.opp_set_disp_pattern_generator_params.height = height;
+		seq_state->steps[*seq_state->num_steps].params.opp_set_disp_pattern_generator_params.offset = offset;
+		(*seq_state->num_steps)++;
+	}
+}
+
+/*
+ * Helper function to add MPC update blending to block sequence
+ */
+void hwss_add_mpc_update_blending(struct block_sequence_state *seq_state,
+		struct mpc *mpc,
+		struct mpcc_blnd_cfg blnd_cfg,
+		int mpcc_id)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = MPC_UPDATE_BLENDING;
+		seq_state->steps[*seq_state->num_steps].params.mpc_update_blending_params.mpc = mpc;
+		seq_state->steps[*seq_state->num_steps].params.mpc_update_blending_params.blnd_cfg = blnd_cfg;
+		seq_state->steps[*seq_state->num_steps].params.mpc_update_blending_params.mpcc_id = mpcc_id;
+		(*seq_state->num_steps)++;
+	}
+}
+
+/*
+ * Helper function to add MPC insert plane to block sequence
+ */
+void hwss_add_mpc_insert_plane(struct block_sequence_state *seq_state,
+		struct mpc *mpc,
+		struct mpc_tree *mpc_tree_params,
+		struct mpcc_blnd_cfg blnd_cfg,
+		struct mpcc_sm_cfg *sm_cfg,
+		struct mpcc *insert_above_mpcc,
+		int dpp_id,
+		int mpcc_id)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = MPC_INSERT_PLANE;
+		seq_state->steps[*seq_state->num_steps].params.mpc_insert_plane_params.mpc = mpc;
+		seq_state->steps[*seq_state->num_steps].params.mpc_insert_plane_params.mpc_tree_params = mpc_tree_params;
+		seq_state->steps[*seq_state->num_steps].params.mpc_insert_plane_params.blnd_cfg = blnd_cfg;
+		seq_state->steps[*seq_state->num_steps].params.mpc_insert_plane_params.sm_cfg = sm_cfg;
+		seq_state->steps[*seq_state->num_steps].params.mpc_insert_plane_params.insert_above_mpcc = insert_above_mpcc;
+		seq_state->steps[*seq_state->num_steps].params.mpc_insert_plane_params.dpp_id = dpp_id;
+		seq_state->steps[*seq_state->num_steps].params.mpc_insert_plane_params.mpcc_id = mpcc_id;
+		(*seq_state->num_steps)++;
+	}
+}
+
+/*
+ * Helper function to add MPC assert idle MPCC to block sequence
+ */
+void hwss_add_mpc_assert_idle_mpcc(struct block_sequence_state *seq_state,
+		struct mpc *mpc,
+		int mpcc_id)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = MPC_ASSERT_IDLE_MPCC;
+		seq_state->steps[*seq_state->num_steps].params.mpc_assert_idle_mpcc_params.mpc = mpc;
+		seq_state->steps[*seq_state->num_steps].params.mpc_assert_idle_mpcc_params.mpcc_id = mpcc_id;
+		(*seq_state->num_steps)++;
+	}
+}
+
+/*
+ * Helper function to add HUBP set blank to block sequence
+ */
+void hwss_add_hubp_set_blank(struct block_sequence_state *seq_state,
+		struct hubp *hubp,
+		bool blank)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = HUBP_SET_BLANK;
+		seq_state->steps[*seq_state->num_steps].params.hubp_set_blank_params.hubp = hubp;
+		seq_state->steps[*seq_state->num_steps].params.hubp_set_blank_params.blank = blank;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_opp_program_bit_depth_reduction(struct block_sequence_state *seq_state,
+		struct output_pixel_processor *opp,
+		bool use_default_params,
+		struct pipe_ctx *pipe_ctx)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = OPP_PROGRAM_BIT_DEPTH_REDUCTION;
+		seq_state->steps[*seq_state->num_steps].params.opp_program_bit_depth_reduction_params.opp = opp;
+		seq_state->steps[*seq_state->num_steps].params.opp_program_bit_depth_reduction_params.use_default_params = use_default_params;
+		seq_state->steps[*seq_state->num_steps].params.opp_program_bit_depth_reduction_params.pipe_ctx = pipe_ctx;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_dc_ip_request_cntl(struct block_sequence_state *seq_state,
+		struct dc *dc,
+		bool enable)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = DC_IP_REQUEST_CNTL;
+		seq_state->steps[*seq_state->num_steps].params.dc_ip_request_cntl_params.dc = dc;
+		seq_state->steps[*seq_state->num_steps].params.dc_ip_request_cntl_params.enable = enable;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_dwbc_update(struct block_sequence_state *seq_state,
+		struct dwbc *dwb,
+		struct dc_dwb_params *dwb_params)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = DWBC_UPDATE;
+		seq_state->steps[*seq_state->num_steps].params.dwbc_update_params.dwb = dwb;
+		seq_state->steps[*seq_state->num_steps].params.dwbc_update_params.dwb_params = dwb_params;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_mcif_wb_config_buf(struct block_sequence_state *seq_state,
+		struct mcif_wb *mcif_wb,
+		struct mcif_buf_params *mcif_buf_params,
+		unsigned int dest_height)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = MCIF_WB_CONFIG_BUF;
+		seq_state->steps[*seq_state->num_steps].params.mcif_wb_config_buf_params.mcif_wb = mcif_wb;
+		seq_state->steps[*seq_state->num_steps].params.mcif_wb_config_buf_params.mcif_buf_params = mcif_buf_params;
+		seq_state->steps[*seq_state->num_steps].params.mcif_wb_config_buf_params.dest_height = dest_height;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_mcif_wb_config_arb(struct block_sequence_state *seq_state,
+		struct mcif_wb *mcif_wb,
+		struct mcif_arb_params *mcif_arb_params)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = MCIF_WB_CONFIG_ARB;
+		seq_state->steps[*seq_state->num_steps].params.mcif_wb_config_arb_params.mcif_wb = mcif_wb;
+		seq_state->steps[*seq_state->num_steps].params.mcif_wb_config_arb_params.mcif_arb_params = mcif_arb_params;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_mcif_wb_enable(struct block_sequence_state *seq_state,
+		struct mcif_wb *mcif_wb)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = MCIF_WB_ENABLE;
+		seq_state->steps[*seq_state->num_steps].params.mcif_wb_enable_params.mcif_wb = mcif_wb;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_mcif_wb_disable(struct block_sequence_state *seq_state,
+		struct mcif_wb *mcif_wb)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = MCIF_WB_DISABLE;
+		seq_state->steps[*seq_state->num_steps].params.mcif_wb_disable_params.mcif_wb = mcif_wb;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_mpc_set_dwb_mux(struct block_sequence_state *seq_state,
+		struct mpc *mpc,
+		int dwb_id,
+		int mpcc_id)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = MPC_SET_DWB_MUX;
+		seq_state->steps[*seq_state->num_steps].params.mpc_set_dwb_mux_params.mpc = mpc;
+		seq_state->steps[*seq_state->num_steps].params.mpc_set_dwb_mux_params.dwb_id = dwb_id;
+		seq_state->steps[*seq_state->num_steps].params.mpc_set_dwb_mux_params.mpcc_id = mpcc_id;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_mpc_disable_dwb_mux(struct block_sequence_state *seq_state,
+		struct mpc *mpc,
+		unsigned int dwb_id)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = MPC_DISABLE_DWB_MUX;
+		seq_state->steps[*seq_state->num_steps].params.mpc_disable_dwb_mux_params.mpc = mpc;
+		seq_state->steps[*seq_state->num_steps].params.mpc_disable_dwb_mux_params.dwb_id = dwb_id;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_dwbc_enable(struct block_sequence_state *seq_state,
+		struct dwbc *dwb,
+		struct dc_dwb_params *dwb_params)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = DWBC_ENABLE;
+		seq_state->steps[*seq_state->num_steps].params.dwbc_enable_params.dwb = dwb;
+		seq_state->steps[*seq_state->num_steps].params.dwbc_enable_params.dwb_params = dwb_params;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_dwbc_disable(struct block_sequence_state *seq_state,
+		struct dwbc *dwb)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = DWBC_DISABLE;
+		seq_state->steps[*seq_state->num_steps].params.dwbc_disable_params.dwb = dwb;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_tg_set_gsl(struct block_sequence_state *seq_state,
+		struct timing_generator *tg,
+		struct gsl_params gsl)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = TG_SET_GSL;
+		seq_state->steps[*seq_state->num_steps].params.tg_set_gsl_params.tg = tg;
+		seq_state->steps[*seq_state->num_steps].params.tg_set_gsl_params.gsl = gsl;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_tg_set_gsl_source_select(struct block_sequence_state *seq_state,
+		struct timing_generator *tg,
+		int group_idx,
+		uint32_t gsl_ready_signal)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = TG_SET_GSL_SOURCE_SELECT;
+		seq_state->steps[*seq_state->num_steps].params.tg_set_gsl_source_select_params.tg = tg;
+		seq_state->steps[*seq_state->num_steps].params.tg_set_gsl_source_select_params.group_idx = group_idx;
+		seq_state->steps[*seq_state->num_steps].params.tg_set_gsl_source_select_params.gsl_ready_signal = gsl_ready_signal;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_hubp_update_mall_sel(struct block_sequence_state *seq_state,
+		struct hubp *hubp,
+		uint32_t mall_sel,
+		bool cache_cursor)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = HUBP_UPDATE_MALL_SEL;
+		seq_state->steps[*seq_state->num_steps].params.hubp_update_mall_sel_params.hubp = hubp;
+		seq_state->steps[*seq_state->num_steps].params.hubp_update_mall_sel_params.mall_sel = mall_sel;
+		seq_state->steps[*seq_state->num_steps].params.hubp_update_mall_sel_params.cache_cursor = cache_cursor;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_hubp_prepare_subvp_buffering(struct block_sequence_state *seq_state,
+		struct hubp *hubp,
+		bool enable)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = HUBP_PREPARE_SUBVP_BUFFERING;
+		seq_state->steps[*seq_state->num_steps].params.hubp_prepare_subvp_buffering_params.hubp = hubp;
+		seq_state->steps[*seq_state->num_steps].params.hubp_prepare_subvp_buffering_params.enable = enable;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_hubp_set_blank_en(struct block_sequence_state *seq_state,
+		struct hubp *hubp,
+		bool enable)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = HUBP_SET_BLANK_EN;
+		seq_state->steps[*seq_state->num_steps].params.hubp_set_blank_en_params.hubp = hubp;
+		seq_state->steps[*seq_state->num_steps].params.hubp_set_blank_en_params.enable = enable;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_hubp_disable_control(struct block_sequence_state *seq_state,
+		struct hubp *hubp,
+		bool disable)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = HUBP_DISABLE_CONTROL;
+		seq_state->steps[*seq_state->num_steps].params.hubp_disable_control_params.hubp = hubp;
+		seq_state->steps[*seq_state->num_steps].params.hubp_disable_control_params.disable = disable;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_hubbub_soft_reset(struct block_sequence_state *seq_state,
+		struct hubbub *hubbub,
+		void (*hubbub_soft_reset)(struct hubbub *hubbub, bool reset),
+		bool reset)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = HUBBUB_SOFT_RESET;
+		seq_state->steps[*seq_state->num_steps].params.hubbub_soft_reset_params.hubbub = hubbub;
+		seq_state->steps[*seq_state->num_steps].params.hubbub_soft_reset_params.hubbub_soft_reset = hubbub_soft_reset;
+		seq_state->steps[*seq_state->num_steps].params.hubbub_soft_reset_params.reset = reset;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_hubp_clk_cntl(struct block_sequence_state *seq_state,
+		struct hubp *hubp,
+		bool enable)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = HUBP_CLK_CNTL;
+		seq_state->steps[*seq_state->num_steps].params.hubp_clk_cntl_params.hubp = hubp;
+		seq_state->steps[*seq_state->num_steps].params.hubp_clk_cntl_params.enable = enable;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_dpp_dppclk_control(struct block_sequence_state *seq_state,
+		struct dpp *dpp,
+		bool dppclk_div,
+		bool enable)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = DPP_DPPCLK_CONTROL;
+		seq_state->steps[*seq_state->num_steps].params.dpp_dppclk_control_params.dpp = dpp;
+		seq_state->steps[*seq_state->num_steps].params.dpp_dppclk_control_params.dppclk_div = dppclk_div;
+		seq_state->steps[*seq_state->num_steps].params.dpp_dppclk_control_params.enable = enable;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_disable_phantom_crtc(struct block_sequence_state *seq_state,
+		struct timing_generator *tg)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = DISABLE_PHANTOM_CRTC;
+		seq_state->steps[*seq_state->num_steps].params.disable_phantom_crtc_params.tg = tg;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_dsc_pg_status(struct block_sequence_state *seq_state,
+		struct dce_hwseq *hws,
+		int dsc_inst,
+		bool is_ungated)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = DSC_PG_STATUS;
+		seq_state->steps[*seq_state->num_steps].params.dsc_pg_status_params.hws = hws;
+		seq_state->steps[*seq_state->num_steps].params.dsc_pg_status_params.dsc_inst = dsc_inst;
+		seq_state->steps[*seq_state->num_steps].params.dsc_pg_status_params.is_ungated = is_ungated;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_dsc_wait_disconnect_pending_clear(struct block_sequence_state *seq_state,
+		struct display_stream_compressor *dsc,
+		bool *is_ungated)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = DSC_WAIT_DISCONNECT_PENDING_CLEAR;
+		seq_state->steps[*seq_state->num_steps].params.dsc_wait_disconnect_pending_clear_params.dsc = dsc;
+		seq_state->steps[*seq_state->num_steps].params.dsc_wait_disconnect_pending_clear_params.is_ungated = is_ungated;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_dsc_disable(struct block_sequence_state *seq_state,
+		struct display_stream_compressor *dsc,
+		bool *is_ungated)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = DSC_DISABLE;
+		seq_state->steps[*seq_state->num_steps].params.dsc_disable_params.dsc = dsc;
+		seq_state->steps[*seq_state->num_steps].params.dsc_disable_params.is_ungated = is_ungated;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_dccg_set_ref_dscclk(struct block_sequence_state *seq_state,
+		struct dccg *dccg,
+		int dsc_inst,
+		bool *is_ungated)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = DCCG_SET_REF_DSCCLK;
+		seq_state->steps[*seq_state->num_steps].params.dccg_set_ref_dscclk_params.dccg = dccg;
+		seq_state->steps[*seq_state->num_steps].params.dccg_set_ref_dscclk_params.dsc_inst = dsc_inst;
+		seq_state->steps[*seq_state->num_steps].params.dccg_set_ref_dscclk_params.is_ungated = is_ungated;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_dpp_root_clock_control(struct block_sequence_state *seq_state,
+		struct dce_hwseq *hws,
+		unsigned int dpp_inst,
+		bool clock_on)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = DPP_ROOT_CLOCK_CONTROL;
+		seq_state->steps[*seq_state->num_steps].params.dpp_root_clock_control_params.hws = hws;
+		seq_state->steps[*seq_state->num_steps].params.dpp_root_clock_control_params.dpp_inst = dpp_inst;
+		seq_state->steps[*seq_state->num_steps].params.dpp_root_clock_control_params.clock_on = clock_on;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_dpp_pg_control(struct block_sequence_state *seq_state,
+		struct dce_hwseq *hws,
+		unsigned int dpp_inst,
+		bool power_on)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = DPP_PG_CONTROL;
+		seq_state->steps[*seq_state->num_steps].params.dpp_pg_control_params.hws = hws;
+		seq_state->steps[*seq_state->num_steps].params.dpp_pg_control_params.dpp_inst = dpp_inst;
+		seq_state->steps[*seq_state->num_steps].params.dpp_pg_control_params.power_on = power_on;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_hubp_pg_control(struct block_sequence_state *seq_state,
+		struct dce_hwseq *hws,
+		unsigned int hubp_inst,
+		bool power_on)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = HUBP_PG_CONTROL;
+		seq_state->steps[*seq_state->num_steps].params.hubp_pg_control_params.hws = hws;
+		seq_state->steps[*seq_state->num_steps].params.hubp_pg_control_params.hubp_inst = hubp_inst;
+		seq_state->steps[*seq_state->num_steps].params.hubp_pg_control_params.power_on = power_on;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_hubp_init(struct block_sequence_state *seq_state,
+		struct hubp *hubp)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = HUBP_INIT;
+		seq_state->steps[*seq_state->num_steps].params.hubp_init_params.hubp = hubp;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_hubp_reset(struct block_sequence_state *seq_state,
+		struct hubp *hubp)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = HUBP_RESET;
+		seq_state->steps[*seq_state->num_steps].params.hubp_reset_params.hubp = hubp;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_dpp_reset(struct block_sequence_state *seq_state,
+		struct dpp *dpp)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = DPP_RESET;
+		seq_state->steps[*seq_state->num_steps].params.dpp_reset_params.dpp = dpp;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_opp_pipe_clock_control(struct block_sequence_state *seq_state,
+		struct output_pixel_processor *opp,
+		bool enable)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = OPP_PIPE_CLOCK_CONTROL;
+		seq_state->steps[*seq_state->num_steps].params.opp_pipe_clock_control_params.opp = opp;
+		seq_state->steps[*seq_state->num_steps].params.opp_pipe_clock_control_params.enable = enable;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_hubp_set_vm_system_aperture_settings(struct block_sequence_state *seq_state,
+		struct hubp *hubp,
+		uint64_t sys_default,
+		uint64_t sys_low,
+		uint64_t sys_high)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = HUBP_SET_VM_SYSTEM_APERTURE_SETTINGS;
+		seq_state->steps[*seq_state->num_steps].params.hubp_set_vm_system_aperture_settings_params.hubp = hubp;
+		seq_state->steps[*seq_state->num_steps].params.hubp_set_vm_system_aperture_settings_params.sys_default.quad_part = sys_default;
+		seq_state->steps[*seq_state->num_steps].params.hubp_set_vm_system_aperture_settings_params.sys_low.quad_part = sys_low;
+		seq_state->steps[*seq_state->num_steps].params.hubp_set_vm_system_aperture_settings_params.sys_high.quad_part = sys_high;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_hubp_set_flip_int(struct block_sequence_state *seq_state,
+		struct hubp *hubp)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = HUBP_SET_FLIP_INT;
+		seq_state->steps[*seq_state->num_steps].params.hubp_set_flip_int_params.hubp = hubp;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_dccg_update_dpp_dto(struct block_sequence_state *seq_state,
+		struct dccg *dccg,
+		int dpp_inst,
+		int dppclk_khz)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = DCCG_UPDATE_DPP_DTO;
+		seq_state->steps[*seq_state->num_steps].params.dccg_update_dpp_dto_params.dccg = dccg;
+		seq_state->steps[*seq_state->num_steps].params.dccg_update_dpp_dto_params.dpp_inst = dpp_inst;
+		seq_state->steps[*seq_state->num_steps].params.dccg_update_dpp_dto_params.dppclk_khz = dppclk_khz;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_hubp_vtg_sel(struct block_sequence_state *seq_state,
+		struct hubp *hubp,
+		uint32_t otg_inst)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = HUBP_VTG_SEL;
+		seq_state->steps[*seq_state->num_steps].params.hubp_vtg_sel_params.hubp = hubp;
+		seq_state->steps[*seq_state->num_steps].params.hubp_vtg_sel_params.otg_inst = otg_inst;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_hubp_setup2(struct block_sequence_state *seq_state,
+		struct hubp *hubp,
+		struct dml2_dchub_per_pipe_register_set *hubp_regs,
+		union dml2_global_sync_programming *global_sync,
+		struct dc_crtc_timing *timing)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = HUBP_SETUP2;
+		seq_state->steps[*seq_state->num_steps].params.hubp_setup2_params.hubp = hubp;
+		seq_state->steps[*seq_state->num_steps].params.hubp_setup2_params.hubp_regs = hubp_regs;
+		seq_state->steps[*seq_state->num_steps].params.hubp_setup2_params.global_sync = global_sync;
+		seq_state->steps[*seq_state->num_steps].params.hubp_setup2_params.timing = timing;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_hubp_setup(struct block_sequence_state *seq_state,
+		struct hubp *hubp,
+		struct _vcs_dpi_display_dlg_regs_st *dlg_regs,
+		struct _vcs_dpi_display_ttu_regs_st *ttu_regs,
+		struct _vcs_dpi_display_rq_regs_st *rq_regs,
+		struct _vcs_dpi_display_pipe_dest_params_st *pipe_dest)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = HUBP_SETUP;
+		seq_state->steps[*seq_state->num_steps].params.hubp_setup_params.hubp = hubp;
+		seq_state->steps[*seq_state->num_steps].params.hubp_setup_params.dlg_regs = dlg_regs;
+		seq_state->steps[*seq_state->num_steps].params.hubp_setup_params.ttu_regs = ttu_regs;
+		seq_state->steps[*seq_state->num_steps].params.hubp_setup_params.rq_regs = rq_regs;
+		seq_state->steps[*seq_state->num_steps].params.hubp_setup_params.pipe_dest = pipe_dest;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_hubp_set_unbounded_requesting(struct block_sequence_state *seq_state,
+		struct hubp *hubp,
+		bool unbounded_req)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = HUBP_SET_UNBOUNDED_REQUESTING;
+		seq_state->steps[*seq_state->num_steps].params.hubp_set_unbounded_requesting_params.hubp = hubp;
+		seq_state->steps[*seq_state->num_steps].params.hubp_set_unbounded_requesting_params.unbounded_req = unbounded_req;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_hubp_setup_interdependent2(struct block_sequence_state *seq_state,
+		struct hubp *hubp,
+		struct dml2_dchub_per_pipe_register_set *hubp_regs)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = HUBP_SETUP_INTERDEPENDENT2;
+		seq_state->steps[*seq_state->num_steps].params.hubp_setup_interdependent2_params.hubp = hubp;
+		seq_state->steps[*seq_state->num_steps].params.hubp_setup_interdependent2_params.hubp_regs = hubp_regs;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_hubp_setup_interdependent(struct block_sequence_state *seq_state,
+		struct hubp *hubp,
+		struct _vcs_dpi_display_dlg_regs_st *dlg_regs,
+		struct _vcs_dpi_display_ttu_regs_st *ttu_regs)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = HUBP_SETUP_INTERDEPENDENT;
+		seq_state->steps[*seq_state->num_steps].params.hubp_setup_interdependent_params.hubp = hubp;
+		seq_state->steps[*seq_state->num_steps].params.hubp_setup_interdependent_params.dlg_regs = dlg_regs;
+		seq_state->steps[*seq_state->num_steps].params.hubp_setup_interdependent_params.ttu_regs = ttu_regs;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_hubp_program_surface_config(struct block_sequence_state *seq_state,
+		struct hubp *hubp,
+		enum surface_pixel_format format,
+		struct dc_tiling_info *tiling_info,
+		struct plane_size plane_size,
+		enum dc_rotation_angle rotation,
+		struct dc_plane_dcc_param *dcc,
+		bool horizontal_mirror,
+		int compat_level)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = HUBP_PROGRAM_SURFACE_CONFIG;
+		seq_state->steps[*seq_state->num_steps].params.program_surface_config_params.hubp = hubp;
+		seq_state->steps[*seq_state->num_steps].params.program_surface_config_params.format = format;
+		seq_state->steps[*seq_state->num_steps].params.program_surface_config_params.tiling_info = tiling_info;
+		seq_state->steps[*seq_state->num_steps].params.program_surface_config_params.plane_size = plane_size;
+		seq_state->steps[*seq_state->num_steps].params.program_surface_config_params.rotation = rotation;
+		seq_state->steps[*seq_state->num_steps].params.program_surface_config_params.dcc = dcc;
+		seq_state->steps[*seq_state->num_steps].params.program_surface_config_params.horizontal_mirror = horizontal_mirror;
+		seq_state->steps[*seq_state->num_steps].params.program_surface_config_params.compat_level = compat_level;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_dpp_setup_dpp(struct block_sequence_state *seq_state,
+		struct pipe_ctx *pipe_ctx)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = DPP_SETUP_DPP;
+		seq_state->steps[*seq_state->num_steps].params.setup_dpp_params.pipe_ctx = pipe_ctx;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_dpp_set_cursor_matrix(struct block_sequence_state *seq_state,
+		struct dpp *dpp,
+		enum dc_color_space color_space,
+		struct dc_csc_transform *cursor_csc_color_matrix)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = DPP_SET_CURSOR_MATRIX;
+		seq_state->steps[*seq_state->num_steps].params.dpp_set_cursor_matrix_params.dpp = dpp;
+		seq_state->steps[*seq_state->num_steps].params.dpp_set_cursor_matrix_params.color_space = color_space;
+		seq_state->steps[*seq_state->num_steps].params.dpp_set_cursor_matrix_params.cursor_csc_color_matrix = cursor_csc_color_matrix;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_dpp_set_scaler(struct block_sequence_state *seq_state,
+		struct dpp *dpp,
+		const struct scaler_data *scl_data)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = DPP_SET_SCALER;
+		seq_state->steps[*seq_state->num_steps].params.dpp_set_scaler_params.dpp = dpp;
+		seq_state->steps[*seq_state->num_steps].params.dpp_set_scaler_params.scl_data = scl_data;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_hubp_mem_program_viewport(struct block_sequence_state *seq_state,
+		struct hubp *hubp,
+		const struct rect *viewport,
+		const struct rect *viewport_c)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = HUBP_MEM_PROGRAM_VIEWPORT;
+		seq_state->steps[*seq_state->num_steps].params.hubp_mem_program_viewport_params.hubp = hubp;
+		seq_state->steps[*seq_state->num_steps].params.hubp_mem_program_viewport_params.viewport = viewport;
+		seq_state->steps[*seq_state->num_steps].params.hubp_mem_program_viewport_params.viewport_c = viewport_c;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_abort_cursor_offload_update(struct block_sequence_state *seq_state,
+		struct dc *dc,
+		struct pipe_ctx *pipe_ctx)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = ABORT_CURSOR_OFFLOAD_UPDATE;
+		seq_state->steps[*seq_state->num_steps].params.abort_cursor_offload_update_params.dc = dc;
+		seq_state->steps[*seq_state->num_steps].params.abort_cursor_offload_update_params.pipe_ctx = pipe_ctx;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_set_cursor_attribute(struct block_sequence_state *seq_state,
+		struct dc *dc,
+		struct pipe_ctx *pipe_ctx)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = SET_CURSOR_ATTRIBUTE;
+		seq_state->steps[*seq_state->num_steps].params.set_cursor_attribute_params.dc = dc;
+		seq_state->steps[*seq_state->num_steps].params.set_cursor_attribute_params.pipe_ctx = pipe_ctx;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_set_cursor_position(struct block_sequence_state *seq_state,
+		struct dc *dc,
+		struct pipe_ctx *pipe_ctx)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = SET_CURSOR_POSITION;
+		seq_state->steps[*seq_state->num_steps].params.set_cursor_position_params.dc = dc;
+		seq_state->steps[*seq_state->num_steps].params.set_cursor_position_params.pipe_ctx = pipe_ctx;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_set_cursor_sdr_white_level(struct block_sequence_state *seq_state,
+		struct dc *dc,
+		struct pipe_ctx *pipe_ctx)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = SET_CURSOR_SDR_WHITE_LEVEL;
+		seq_state->steps[*seq_state->num_steps].params.set_cursor_sdr_white_level_params.dc = dc;
+		seq_state->steps[*seq_state->num_steps].params.set_cursor_sdr_white_level_params.pipe_ctx = pipe_ctx;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_program_output_csc(struct block_sequence_state *seq_state,
+		struct dc *dc,
+		struct pipe_ctx *pipe_ctx,
+		enum dc_color_space colorspace,
+		uint16_t *matrix,
+		int opp_id)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = PROGRAM_OUTPUT_CSC;
+		seq_state->steps[*seq_state->num_steps].params.program_output_csc_params.dc = dc;
+		seq_state->steps[*seq_state->num_steps].params.program_output_csc_params.pipe_ctx = pipe_ctx;
+		seq_state->steps[*seq_state->num_steps].params.program_output_csc_params.colorspace = colorspace;
+		seq_state->steps[*seq_state->num_steps].params.program_output_csc_params.matrix = matrix;
+		seq_state->steps[*seq_state->num_steps].params.program_output_csc_params.opp_id = opp_id;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_phantom_hubp_post_enable(struct block_sequence_state *seq_state,
+		struct hubp *hubp)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = PHANTOM_HUBP_POST_ENABLE;
+		seq_state->steps[*seq_state->num_steps].params.phantom_hubp_post_enable_params.hubp = hubp;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_update_force_pstate(struct block_sequence_state *seq_state,
+		struct dc *dc,
+		struct dc_state *context)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = UPDATE_FORCE_PSTATE;
+		seq_state->steps[*seq_state->num_steps].params.update_force_pstate_params.dc = dc;
+		seq_state->steps[*seq_state->num_steps].params.update_force_pstate_params.context = context;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_hubbub_apply_dedcn21_147_wa(struct block_sequence_state *seq_state,
+		struct hubbub *hubbub)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = HUBBUB_APPLY_DEDCN21_147_WA;
+		seq_state->steps[*seq_state->num_steps].params.hubbub_apply_dedcn21_147_wa_params.hubbub = hubbub;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_hubbub_allow_self_refresh_control(struct block_sequence_state *seq_state,
+		struct hubbub *hubbub,
+		bool allow,
+		bool *disallow_self_refresh_applied)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = HUBBUB_ALLOW_SELF_REFRESH_CONTROL;
+		seq_state->steps[*seq_state->num_steps].params.hubbub_allow_self_refresh_control_params.hubbub = hubbub;
+		seq_state->steps[*seq_state->num_steps].params.hubbub_allow_self_refresh_control_params.allow = allow;
+		seq_state->steps[*seq_state->num_steps].params.hubbub_allow_self_refresh_control_params.disallow_self_refresh_applied = disallow_self_refresh_applied;
+		(*seq_state->num_steps)++;
+	}
+}
+
+void hwss_add_tg_get_frame_count(struct block_sequence_state *seq_state,
+		struct timing_generator *tg,
+		unsigned int *frame_count)
+{
+	if (*seq_state->num_steps < MAX_HWSS_BLOCK_SEQUENCE_SIZE) {
+		seq_state->steps[*seq_state->num_steps].func = TG_GET_FRAME_COUNT;
+		seq_state->steps[*seq_state->num_steps].params.tg_get_frame_count_params.tg = tg;
+		seq_state->steps[*seq_state->num_steps].params.tg_get_frame_count_params.frame_count = frame_count;
+		(*seq_state->num_steps)++;
+	}
+}
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_link.c b/drivers/gpu/drm/amd/display/dc/core/dc_link.c
deleted file mode 100644
index c8457babfdea..000000000000
--- a/drivers/gpu/drm/amd/display/dc/core/dc_link.c
+++ /dev/null
@@ -1,4830 +0,0 @@
-/*
- * Copyright 2012-15 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: AMD
- *
- */
-
-#include <linux/slab.h>
-
-#include "dm_services.h"
-#include "atomfirmware.h"
-#include "dm_helpers.h"
-#include "dc.h"
-#include "grph_object_id.h"
-#include "gpio_service_interface.h"
-#include "core_status.h"
-#include "dc_link_dp.h"
-#include "dc_link_ddc.h"
-#include "link_hwss.h"
-#include "opp.h"
-
-#include "link_encoder.h"
-#include "hw_sequencer.h"
-#include "resource.h"
-#include "abm.h"
-#include "fixed31_32.h"
-#include "dpcd_defs.h"
-#include "dmcu.h"
-#include "hw/clk_mgr.h"
-#include "dce/dmub_psr.h"
-#include "dmub/dmub_srv.h"
-#include "inc/hw/panel_cntl.h"
-#include "inc/link_enc_cfg.h"
-#include "inc/link_dpcd.h"
-
-#include "dc/dcn30/dcn30_vpg.h"
-
-#define DC_LOGGER_INIT(logger)
-
-#define LINK_INFO(...) \
-	DC_LOG_HW_HOTPLUG(  \
-		__VA_ARGS__)
-
-#define RETIMER_REDRIVER_INFO(...) \
-	DC_LOG_RETIMER_REDRIVER(  \
-		__VA_ARGS__)
-
-/*******************************************************************************
- * Private functions
- ******************************************************************************/
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-static bool add_dp_hpo_link_encoder_to_link(struct dc_link *link)
-{
-	struct hpo_dp_link_encoder *enc = resource_get_unused_hpo_dp_link_encoder(
-					link->dc->res_pool);
-
-	if (!link->hpo_dp_link_enc && enc) {
-		link->hpo_dp_link_enc = enc;
-		link->hpo_dp_link_enc->transmitter = link->link_enc->transmitter;
-		link->hpo_dp_link_enc->hpd_source = link->link_enc->hpd_source;
-	}
-
-	return (link->hpo_dp_link_enc != NULL);
-}
-
-static void remove_dp_hpo_link_encoder_from_link(struct dc_link *link)
-{
-	if (link->hpo_dp_link_enc) {
-		link->hpo_dp_link_enc->hpd_source = HPD_SOURCEID_UNKNOWN;
-		link->hpo_dp_link_enc->transmitter = TRANSMITTER_UNKNOWN;
-		link->hpo_dp_link_enc = NULL;
-	}
-}
-#endif
-
-static void dc_link_destruct(struct dc_link *link)
-{
-	int i;
-
-	if (link->hpd_gpio) {
-		dal_gpio_destroy_irq(&link->hpd_gpio);
-		link->hpd_gpio = NULL;
-	}
-
-	if (link->ddc)
-		dal_ddc_service_destroy(&link->ddc);
-
-	if (link->panel_cntl)
-		link->panel_cntl->funcs->destroy(&link->panel_cntl);
-
-	if (link->link_enc) {
-		/* Update link encoder resource tracking variables. These are used for
-		 * the dynamic assignment of link encoders to streams. Virtual links
-		 * are not assigned encoder resources on creation.
-		 */
-		if (link->link_id.id != CONNECTOR_ID_VIRTUAL) {
-			link->dc->res_pool->link_encoders[link->eng_id - ENGINE_ID_DIGA] = NULL;
-			link->dc->res_pool->dig_link_enc_count--;
-		}
-		link->link_enc->funcs->destroy(&link->link_enc);
-	}
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	if (link->hpo_dp_link_enc) {
-		remove_dp_hpo_link_encoder_from_link(link);
-	}
-#endif
-
-	if (link->local_sink)
-		dc_sink_release(link->local_sink);
-
-	for (i = 0; i < link->sink_count; ++i)
-		dc_sink_release(link->remote_sinks[i]);
-}
-
-struct gpio *get_hpd_gpio(struct dc_bios *dcb,
-			  struct graphics_object_id link_id,
-			  struct gpio_service *gpio_service)
-{
-	enum bp_result bp_result;
-	struct graphics_object_hpd_info hpd_info;
-	struct gpio_pin_info pin_info;
-
-	if (dcb->funcs->get_hpd_info(dcb, link_id, &hpd_info) != BP_RESULT_OK)
-		return NULL;
-
-	bp_result = dcb->funcs->get_gpio_pin_info(dcb,
-		hpd_info.hpd_int_gpio_uid, &pin_info);
-
-	if (bp_result != BP_RESULT_OK) {
-		ASSERT(bp_result == BP_RESULT_NORECORD);
-		return NULL;
-	}
-
-	return dal_gpio_service_create_irq(gpio_service,
-					   pin_info.offset,
-					   pin_info.mask);
-}
-
-/*
- *  Function: program_hpd_filter
- *
- *  @brief
- *     Programs HPD filter on associated HPD line
- *
- *  @param [in] delay_on_connect_in_ms: Connect filter timeout
- *  @param [in] delay_on_disconnect_in_ms: Disconnect filter timeout
- *
- *  @return
- *     true on success, false otherwise
- */
-static bool program_hpd_filter(const struct dc_link *link)
-{
-	bool result = false;
-	struct gpio *hpd;
-	int delay_on_connect_in_ms = 0;
-	int delay_on_disconnect_in_ms = 0;
-
-	if (link->is_hpd_filter_disabled)
-		return false;
-	/* Verify feature is supported */
-	switch (link->connector_signal) {
-	case SIGNAL_TYPE_DVI_SINGLE_LINK:
-	case SIGNAL_TYPE_DVI_DUAL_LINK:
-	case SIGNAL_TYPE_HDMI_TYPE_A:
-		/* Program hpd filter */
-		delay_on_connect_in_ms = 500;
-		delay_on_disconnect_in_ms = 100;
-		break;
-	case SIGNAL_TYPE_DISPLAY_PORT:
-	case SIGNAL_TYPE_DISPLAY_PORT_MST:
-		/* Program hpd filter to allow DP signal to settle */
-		/* 500:	not able to detect MST <-> SST switch as HPD is low for
-		 * only 100ms on DELL U2413
-		 * 0: some passive dongle still show aux mode instead of i2c
-		 * 20-50: not enough to hide bouncing HPD with passive dongle.
-		 * also see intermittent i2c read issues.
-		 */
-		delay_on_connect_in_ms = 80;
-		delay_on_disconnect_in_ms = 0;
-		break;
-	case SIGNAL_TYPE_LVDS:
-	case SIGNAL_TYPE_EDP:
-	default:
-		/* Don't program hpd filter */
-		return false;
-	}
-
-	/* Obtain HPD handle */
-	hpd = get_hpd_gpio(link->ctx->dc_bios, link->link_id,
-			   link->ctx->gpio_service);
-
-	if (!hpd)
-		return result;
-
-	/* Setup HPD filtering */
-	if (dal_gpio_open(hpd, GPIO_MODE_INTERRUPT) == GPIO_RESULT_OK) {
-		struct gpio_hpd_config config;
-
-		config.delay_on_connect = delay_on_connect_in_ms;
-		config.delay_on_disconnect = delay_on_disconnect_in_ms;
-
-		dal_irq_setup_hpd_filter(hpd, &config);
-
-		dal_gpio_close(hpd);
-
-		result = true;
-	} else {
-		ASSERT_CRITICAL(false);
-	}
-
-	/* Release HPD handle */
-	dal_gpio_destroy_irq(&hpd);
-
-	return result;
-}
-
-bool dc_link_wait_for_t12(struct dc_link *link)
-{
-	if (link->connector_signal == SIGNAL_TYPE_EDP && link->dc->hwss.edp_wait_for_T12) {
-		link->dc->hwss.edp_wait_for_T12(link);
-
-		return true;
-	}
-
-	return false;
-}
-
-/**
- * dc_link_detect_sink() - Determine if there is a sink connected
- *
- * @link: pointer to the dc link
- * @type: Returned connection type
- * Does not detect downstream devices, such as MST sinks
- * or display connected through active dongles
- */
-bool dc_link_detect_sink(struct dc_link *link, enum dc_connection_type *type)
-{
-	uint32_t is_hpd_high = 0;
-	struct gpio *hpd_pin;
-
-	if (link->connector_signal == SIGNAL_TYPE_LVDS) {
-		*type = dc_connection_single;
-		return true;
-	}
-
-	if (link->connector_signal == SIGNAL_TYPE_EDP) {
-		/*in case it is not on*/
-		link->dc->hwss.edp_power_control(link, true);
-		link->dc->hwss.edp_wait_for_hpd_ready(link, true);
-	}
-
-	/* Link may not have physical HPD pin. */
-	if (link->ep_type != DISPLAY_ENDPOINT_PHY) {
-		if (link->hpd_status)
-			*type = dc_connection_single;
-		else
-			*type = dc_connection_none;
-
-		return true;
-	}
-
-	/* todo: may need to lock gpio access */
-	hpd_pin = get_hpd_gpio(link->ctx->dc_bios, link->link_id,
-			       link->ctx->gpio_service);
-	if (!hpd_pin)
-		goto hpd_gpio_failure;
-
-	dal_gpio_open(hpd_pin, GPIO_MODE_INTERRUPT);
-	dal_gpio_get_value(hpd_pin, &is_hpd_high);
-	dal_gpio_close(hpd_pin);
-	dal_gpio_destroy_irq(&hpd_pin);
-
-	if (is_hpd_high) {
-		*type = dc_connection_single;
-		/* TODO: need to do the actual detection */
-	} else {
-		*type = dc_connection_none;
-	}
-
-	return true;
-
-hpd_gpio_failure:
-	return false;
-}
-
-static enum ddc_transaction_type get_ddc_transaction_type(enum signal_type sink_signal)
-{
-	enum ddc_transaction_type transaction_type = DDC_TRANSACTION_TYPE_NONE;
-
-	switch (sink_signal) {
-	case SIGNAL_TYPE_DVI_SINGLE_LINK:
-	case SIGNAL_TYPE_DVI_DUAL_LINK:
-	case SIGNAL_TYPE_HDMI_TYPE_A:
-	case SIGNAL_TYPE_LVDS:
-	case SIGNAL_TYPE_RGB:
-		transaction_type = DDC_TRANSACTION_TYPE_I2C;
-		break;
-
-	case SIGNAL_TYPE_DISPLAY_PORT:
-	case SIGNAL_TYPE_EDP:
-		transaction_type = DDC_TRANSACTION_TYPE_I2C_OVER_AUX;
-		break;
-
-	case SIGNAL_TYPE_DISPLAY_PORT_MST:
-		/* MST does not use I2COverAux, but there is the
-		 * SPECIAL use case for "immediate dwnstrm device
-		 * access" (EPR#370830).
-		 */
-		transaction_type = DDC_TRANSACTION_TYPE_I2C_OVER_AUX;
-		break;
-
-	default:
-		break;
-	}
-
-	return transaction_type;
-}
-
-static enum signal_type get_basic_signal_type(struct graphics_object_id encoder,
-					      struct graphics_object_id downstream)
-{
-	if (downstream.type == OBJECT_TYPE_CONNECTOR) {
-		switch (downstream.id) {
-		case CONNECTOR_ID_SINGLE_LINK_DVII:
-			switch (encoder.id) {
-			case ENCODER_ID_INTERNAL_DAC1:
-			case ENCODER_ID_INTERNAL_KLDSCP_DAC1:
-			case ENCODER_ID_INTERNAL_DAC2:
-			case ENCODER_ID_INTERNAL_KLDSCP_DAC2:
-				return SIGNAL_TYPE_RGB;
-			default:
-				return SIGNAL_TYPE_DVI_SINGLE_LINK;
-			}
-		break;
-		case CONNECTOR_ID_DUAL_LINK_DVII:
-		{
-			switch (encoder.id) {
-			case ENCODER_ID_INTERNAL_DAC1:
-			case ENCODER_ID_INTERNAL_KLDSCP_DAC1:
-			case ENCODER_ID_INTERNAL_DAC2:
-			case ENCODER_ID_INTERNAL_KLDSCP_DAC2:
-				return SIGNAL_TYPE_RGB;
-			default:
-				return SIGNAL_TYPE_DVI_DUAL_LINK;
-			}
-		}
-		break;
-		case CONNECTOR_ID_SINGLE_LINK_DVID:
-			return SIGNAL_TYPE_DVI_SINGLE_LINK;
-		case CONNECTOR_ID_DUAL_LINK_DVID:
-			return SIGNAL_TYPE_DVI_DUAL_LINK;
-		case CONNECTOR_ID_VGA:
-			return SIGNAL_TYPE_RGB;
-		case CONNECTOR_ID_HDMI_TYPE_A:
-			return SIGNAL_TYPE_HDMI_TYPE_A;
-		case CONNECTOR_ID_LVDS:
-			return SIGNAL_TYPE_LVDS;
-		case CONNECTOR_ID_DISPLAY_PORT:
-			return SIGNAL_TYPE_DISPLAY_PORT;
-		case CONNECTOR_ID_EDP:
-			return SIGNAL_TYPE_EDP;
-		default:
-			return SIGNAL_TYPE_NONE;
-		}
-	} else if (downstream.type == OBJECT_TYPE_ENCODER) {
-		switch (downstream.id) {
-		case ENCODER_ID_EXTERNAL_NUTMEG:
-		case ENCODER_ID_EXTERNAL_TRAVIS:
-			return SIGNAL_TYPE_DISPLAY_PORT;
-		default:
-			return SIGNAL_TYPE_NONE;
-		}
-	}
-
-	return SIGNAL_TYPE_NONE;
-}
-
-/*
- * dc_link_is_dp_sink_present() - Check if there is a native DP
- * or passive DP-HDMI dongle connected
- */
-bool dc_link_is_dp_sink_present(struct dc_link *link)
-{
-	enum gpio_result gpio_result;
-	uint32_t clock_pin = 0;
-	uint8_t retry = 0;
-	struct ddc *ddc;
-
-	enum connector_id connector_id =
-		dal_graphics_object_id_get_connector_id(link->link_id);
-
-	bool present =
-		((connector_id == CONNECTOR_ID_DISPLAY_PORT) ||
-		(connector_id == CONNECTOR_ID_EDP));
-
-	ddc = dal_ddc_service_get_ddc_pin(link->ddc);
-
-	if (!ddc) {
-		BREAK_TO_DEBUGGER();
-		return present;
-	}
-
-	/* Open GPIO and set it to I2C mode */
-	/* Note: this GpioMode_Input will be converted
-	 * to GpioConfigType_I2cAuxDualMode in GPIO component,
-	 * which indicates we need additional delay
-	 */
-
-	if (dal_ddc_open(ddc, GPIO_MODE_INPUT,
-			 GPIO_DDC_CONFIG_TYPE_MODE_I2C) != GPIO_RESULT_OK) {
-		dal_ddc_close(ddc);
-
-		return present;
-	}
-
-	/*
-	 * Read GPIO: DP sink is present if both clock and data pins are zero
-	 *
-	 * [W/A] plug-unplug DP cable, sometimes customer board has
-	 * one short pulse on clk_pin(1V, < 1ms). DP will be config to HDMI/DVI
-	 * then monitor can't br light up. Add retry 3 times
-	 * But in real passive dongle, it need additional 3ms to detect
-	 */
-	do {
-		gpio_result = dal_gpio_get_value(ddc->pin_clock, &clock_pin);
-		ASSERT(gpio_result == GPIO_RESULT_OK);
-		if (clock_pin)
-			udelay(1000);
-		else
-			break;
-	} while (retry++ < 3);
-
-	present = (gpio_result == GPIO_RESULT_OK) && !clock_pin;
-
-	dal_ddc_close(ddc);
-
-	return present;
-}
-
-/*
- * @brief
- * Detect output sink type
- */
-static enum signal_type link_detect_sink(struct dc_link *link,
-					 enum dc_detect_reason reason)
-{
-	enum signal_type result;
-	struct graphics_object_id enc_id;
-
-	if (link->is_dig_mapping_flexible)
-		enc_id = (struct graphics_object_id){.id = ENCODER_ID_UNKNOWN};
-	else
-		enc_id = link->link_enc->id;
-	result = get_basic_signal_type(enc_id, link->link_id);
-
-	/* Use basic signal type for link without physical connector. */
-	if (link->ep_type != DISPLAY_ENDPOINT_PHY)
-		return result;
-
-	/* Internal digital encoder will detect only dongles
-	 * that require digital signal
-	 */
-
-	/* Detection mechanism is different
-	 * for different native connectors.
-	 * LVDS connector supports only LVDS signal;
-	 * PCIE is a bus slot, the actual connector needs to be detected first;
-	 * eDP connector supports only eDP signal;
-	 * HDMI should check straps for audio
-	 */
-
-	/* PCIE detects the actual connector on add-on board */
-	if (link->link_id.id == CONNECTOR_ID_PCIE) {
-		/* ZAZTODO implement PCIE add-on card detection */
-	}
-
-	switch (link->link_id.id) {
-	case CONNECTOR_ID_HDMI_TYPE_A: {
-		/* check audio support:
-		 * if native HDMI is not supported, switch to DVI
-		 */
-		struct audio_support *aud_support =
-					&link->dc->res_pool->audio_support;
-
-		if (!aud_support->hdmi_audio_native)
-			if (link->link_id.id == CONNECTOR_ID_HDMI_TYPE_A)
-				result = SIGNAL_TYPE_DVI_SINGLE_LINK;
-	}
-	break;
-	case CONNECTOR_ID_DISPLAY_PORT: {
-		/* DP HPD short pulse. Passive DP dongle will not
-		 * have short pulse
-		 */
-		if (reason != DETECT_REASON_HPDRX) {
-			/* Check whether DP signal detected: if not -
-			 * we assume signal is DVI; it could be corrected
-			 * to HDMI after dongle detection
-			 */
-			if (!dm_helpers_is_dp_sink_present(link))
-				result = SIGNAL_TYPE_DVI_SINGLE_LINK;
-		}
-	}
-	break;
-	default:
-	break;
-	}
-
-	return result;
-}
-
-static enum signal_type decide_signal_from_strap_and_dongle_type(enum display_dongle_type dongle_type,
-								 struct audio_support *audio_support)
-{
-	enum signal_type signal = SIGNAL_TYPE_NONE;
-
-	switch (dongle_type) {
-	case DISPLAY_DONGLE_DP_HDMI_DONGLE:
-		if (audio_support->hdmi_audio_on_dongle)
-			signal = SIGNAL_TYPE_HDMI_TYPE_A;
-		else
-			signal = SIGNAL_TYPE_DVI_SINGLE_LINK;
-		break;
-	case DISPLAY_DONGLE_DP_DVI_DONGLE:
-		signal = SIGNAL_TYPE_DVI_SINGLE_LINK;
-		break;
-	case DISPLAY_DONGLE_DP_HDMI_MISMATCHED_DONGLE:
-		if (audio_support->hdmi_audio_native)
-			signal =  SIGNAL_TYPE_HDMI_TYPE_A;
-		else
-			signal = SIGNAL_TYPE_DVI_SINGLE_LINK;
-		break;
-	default:
-		signal = SIGNAL_TYPE_NONE;
-		break;
-	}
-
-	return signal;
-}
-
-static enum signal_type dp_passive_dongle_detection(struct ddc_service *ddc,
-						    struct display_sink_capability *sink_cap,
-						    struct audio_support *audio_support)
-{
-	dal_ddc_service_i2c_query_dp_dual_mode_adaptor(ddc, sink_cap);
-
-	return decide_signal_from_strap_and_dongle_type(sink_cap->dongle_type,
-							audio_support);
-}
-
-static void link_disconnect_sink(struct dc_link *link)
-{
-	if (link->local_sink) {
-		dc_sink_release(link->local_sink);
-		link->local_sink = NULL;
-	}
-
-	link->dpcd_sink_count = 0;
-	//link->dpcd_caps.dpcd_rev.raw = 0;
-}
-
-static void link_disconnect_remap(struct dc_sink *prev_sink, struct dc_link *link)
-{
-	dc_sink_release(link->local_sink);
-	link->local_sink = prev_sink;
-}
-
-#if defined(CONFIG_DRM_AMD_DC_HDCP)
-bool dc_link_is_hdcp14(struct dc_link *link, enum signal_type signal)
-{
-	bool ret = false;
-
-	switch (signal)	{
-	case SIGNAL_TYPE_DISPLAY_PORT:
-	case SIGNAL_TYPE_DISPLAY_PORT_MST:
-		ret = link->hdcp_caps.bcaps.bits.HDCP_CAPABLE;
-		break;
-	case SIGNAL_TYPE_DVI_SINGLE_LINK:
-	case SIGNAL_TYPE_DVI_DUAL_LINK:
-	case SIGNAL_TYPE_HDMI_TYPE_A:
-	/* HDMI doesn't tell us its HDCP(1.4) capability, so assume to always be capable,
-	 * we can poll for bksv but some displays have an issue with this. Since its so rare
-	 * for a display to not be 1.4 capable, this assumtion is ok
-	 */
-		ret = true;
-		break;
-	default:
-		break;
-	}
-	return ret;
-}
-
-bool dc_link_is_hdcp22(struct dc_link *link, enum signal_type signal)
-{
-	bool ret = false;
-
-	switch (signal)	{
-	case SIGNAL_TYPE_DISPLAY_PORT:
-	case SIGNAL_TYPE_DISPLAY_PORT_MST:
-		ret = (link->hdcp_caps.bcaps.bits.HDCP_CAPABLE &&
-				link->hdcp_caps.rx_caps.fields.byte0.hdcp_capable &&
-				(link->hdcp_caps.rx_caps.fields.version == 0x2)) ? 1 : 0;
-		break;
-	case SIGNAL_TYPE_DVI_SINGLE_LINK:
-	case SIGNAL_TYPE_DVI_DUAL_LINK:
-	case SIGNAL_TYPE_HDMI_TYPE_A:
-		ret = (link->hdcp_caps.rx_caps.fields.version == 0x4) ? 1:0;
-		break;
-	default:
-		break;
-	}
-
-	return ret;
-}
-
-static void query_hdcp_capability(enum signal_type signal, struct dc_link *link)
-{
-	struct hdcp_protection_message msg22;
-	struct hdcp_protection_message msg14;
-
-	memset(&msg22, 0, sizeof(struct hdcp_protection_message));
-	memset(&msg14, 0, sizeof(struct hdcp_protection_message));
-	memset(link->hdcp_caps.rx_caps.raw, 0,
-		sizeof(link->hdcp_caps.rx_caps.raw));
-
-	if ((link->connector_signal == SIGNAL_TYPE_DISPLAY_PORT &&
-			link->ddc->transaction_type ==
-			DDC_TRANSACTION_TYPE_I2C_OVER_AUX) ||
-			link->connector_signal == SIGNAL_TYPE_EDP) {
-		msg22.data = link->hdcp_caps.rx_caps.raw;
-		msg22.length = sizeof(link->hdcp_caps.rx_caps.raw);
-		msg22.msg_id = HDCP_MESSAGE_ID_RX_CAPS;
-	} else {
-		msg22.data = &link->hdcp_caps.rx_caps.fields.version;
-		msg22.length = sizeof(link->hdcp_caps.rx_caps.fields.version);
-		msg22.msg_id = HDCP_MESSAGE_ID_HDCP2VERSION;
-	}
-	msg22.version = HDCP_VERSION_22;
-	msg22.link = HDCP_LINK_PRIMARY;
-	msg22.max_retries = 5;
-	dc_process_hdcp_msg(signal, link, &msg22);
-
-	if (signal == SIGNAL_TYPE_DISPLAY_PORT || signal == SIGNAL_TYPE_DISPLAY_PORT_MST) {
-		msg14.data = &link->hdcp_caps.bcaps.raw;
-		msg14.length = sizeof(link->hdcp_caps.bcaps.raw);
-		msg14.msg_id = HDCP_MESSAGE_ID_READ_BCAPS;
-		msg14.version = HDCP_VERSION_14;
-		msg14.link = HDCP_LINK_PRIMARY;
-		msg14.max_retries = 5;
-
-		dc_process_hdcp_msg(signal, link, &msg14);
-	}
-
-}
-#endif
-
-static void read_current_link_settings_on_detect(struct dc_link *link)
-{
-	union lane_count_set lane_count_set = {0};
-	uint8_t link_bw_set;
-	uint8_t link_rate_set;
-	uint32_t read_dpcd_retry_cnt = 10;
-	enum dc_status status = DC_ERROR_UNEXPECTED;
-	int i;
-	union max_down_spread max_down_spread = {0};
-
-	// Read DPCD 00101h to find out the number of lanes currently set
-	for (i = 0; i < read_dpcd_retry_cnt; i++) {
-		status = core_link_read_dpcd(link,
-					     DP_LANE_COUNT_SET,
-					     &lane_count_set.raw,
-					     sizeof(lane_count_set));
-		/* First DPCD read after VDD ON can fail if the particular board
-		 * does not have HPD pin wired correctly. So if DPCD read fails,
-		 * which it should never happen, retry a few times. Target worst
-		 * case scenario of 80 ms.
-		 */
-		if (status == DC_OK) {
-			link->cur_link_settings.lane_count =
-					lane_count_set.bits.LANE_COUNT_SET;
-			break;
-		}
-
-		msleep(8);
-	}
-
-	// Read DPCD 00100h to find if standard link rates are set
-	core_link_read_dpcd(link, DP_LINK_BW_SET,
-			    &link_bw_set, sizeof(link_bw_set));
-
-	if (link_bw_set == 0) {
-		if (link->connector_signal == SIGNAL_TYPE_EDP) {
-			/* If standard link rates are not being used,
-			 * Read DPCD 00115h to find the edp link rate set used
-			 */
-			core_link_read_dpcd(link, DP_LINK_RATE_SET,
-					    &link_rate_set, sizeof(link_rate_set));
-
-			// edp_supported_link_rates_count = 0 for DP
-			if (link_rate_set < link->dpcd_caps.edp_supported_link_rates_count) {
-				link->cur_link_settings.link_rate =
-					link->dpcd_caps.edp_supported_link_rates[link_rate_set];
-				link->cur_link_settings.link_rate_set = link_rate_set;
-				link->cur_link_settings.use_link_rate_set = true;
-			}
-		} else {
-			// Link Rate not found. Seamless boot may not work.
-			ASSERT(false);
-		}
-	} else {
-		link->cur_link_settings.link_rate = link_bw_set;
-		link->cur_link_settings.use_link_rate_set = false;
-	}
-	// Read DPCD 00003h to find the max down spread.
-	core_link_read_dpcd(link, DP_MAX_DOWNSPREAD,
-			    &max_down_spread.raw, sizeof(max_down_spread));
-	link->cur_link_settings.link_spread =
-		max_down_spread.bits.MAX_DOWN_SPREAD ?
-		LINK_SPREAD_05_DOWNSPREAD_30KHZ : LINK_SPREAD_DISABLED;
-}
-
-static bool detect_dp(struct dc_link *link,
-		      struct display_sink_capability *sink_caps,
-		      enum dc_detect_reason reason)
-{
-	struct audio_support *audio_support = &link->dc->res_pool->audio_support;
-
-	sink_caps->signal = link_detect_sink(link, reason);
-	sink_caps->transaction_type =
-		get_ddc_transaction_type(sink_caps->signal);
-
-	if (sink_caps->transaction_type == DDC_TRANSACTION_TYPE_I2C_OVER_AUX) {
-		sink_caps->signal = SIGNAL_TYPE_DISPLAY_PORT;
-		if (!detect_dp_sink_caps(link))
-			return false;
-		if (is_mst_supported(link)) {
-			sink_caps->signal = SIGNAL_TYPE_DISPLAY_PORT_MST;
-			link->type = dc_connection_mst_branch;
-
-			dal_ddc_service_set_transaction_type(link->ddc,
-							     sink_caps->transaction_type);
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-			/* Apply work around for tunneled MST on certain USB4 docks. Always use DSC if dock
-			 * reports DSC support.
-			 */
-			if (link->ep_type == DISPLAY_ENDPOINT_USB4_DPIA &&
-					link->type == dc_connection_mst_branch &&
-					link->dpcd_caps.branch_dev_id == DP_BRANCH_DEVICE_ID_90CC24 &&
-					link->dpcd_caps.dsc_caps.dsc_basic_caps.fields.dsc_support.DSC_SUPPORT &&
-					!link->dc->debug.dpia_debug.bits.disable_mst_dsc_work_around)
-				link->wa_flags.dpia_mst_dsc_always_on = true;
-#endif
-
-#if defined(CONFIG_DRM_AMD_DC_HDCP)
-			/* In case of fallback to SST when topology discovery below fails
-			 * HDCP caps will be querried again later by the upper layer (caller
-			 * of this function). */
-			query_hdcp_capability(SIGNAL_TYPE_DISPLAY_PORT_MST, link);
-#endif
-		}
-
-		if (link->type != dc_connection_mst_branch &&
-				is_dp_branch_device(link))
-			/* DP SST branch */
-			link->type = dc_connection_sst_branch;
-	} else {
-		/* DP passive dongles */
-		sink_caps->signal = dp_passive_dongle_detection(link->ddc,
-								sink_caps,
-								audio_support);
-		link->dpcd_caps.dongle_type = sink_caps->dongle_type;
-		link->dpcd_caps.dpcd_rev.raw = 0;
-	}
-
-	return true;
-}
-
-static bool is_same_edid(struct dc_edid *old_edid, struct dc_edid *new_edid)
-{
-	if (old_edid->length != new_edid->length)
-		return false;
-
-	if (new_edid->length == 0)
-		return false;
-
-	return (memcmp(old_edid->raw_edid,
-		       new_edid->raw_edid, new_edid->length) == 0);
-}
-
-static bool wait_for_entering_dp_alt_mode(struct dc_link *link)
-{
-	/**
-	 * something is terribly wrong if time out is > 200ms. (5Hz)
-	 * 500 microseconds * 400 tries us 200 ms
-	 **/
-	unsigned int sleep_time_in_microseconds = 500;
-	unsigned int tries_allowed = 400;
-	bool is_in_alt_mode;
-	unsigned long long enter_timestamp;
-	unsigned long long finish_timestamp;
-	unsigned long long time_taken_in_ns;
-	int tries_taken;
-
-	DC_LOGGER_INIT(link->ctx->logger);
-
-	if (!link->link_enc->funcs->is_in_alt_mode)
-		return true;
-
-	is_in_alt_mode = link->link_enc->funcs->is_in_alt_mode(link->link_enc);
-	DC_LOG_WARNING("DP Alt mode state on HPD: %d\n", is_in_alt_mode);
-
-	if (is_in_alt_mode)
-		return true;
-
-	enter_timestamp = dm_get_timestamp(link->ctx);
-
-	for (tries_taken = 0; tries_taken < tries_allowed; tries_taken++) {
-		udelay(sleep_time_in_microseconds);
-		/* ask the link if alt mode is enabled, if so return ok */
-		if (link->link_enc->funcs->is_in_alt_mode(link->link_enc)) {
-			finish_timestamp = dm_get_timestamp(link->ctx);
-			time_taken_in_ns =
-				dm_get_elapse_time_in_ns(link->ctx,
-							 finish_timestamp,
-							 enter_timestamp);
-			DC_LOG_WARNING("Alt mode entered finished after %llu ms\n",
-				       div_u64(time_taken_in_ns, 1000000));
-			return true;
-		}
-	}
-	finish_timestamp = dm_get_timestamp(link->ctx);
-	time_taken_in_ns = dm_get_elapse_time_in_ns(link->ctx, finish_timestamp,
-						    enter_timestamp);
-	DC_LOG_WARNING("Alt mode has timed out after %llu ms\n",
-		       div_u64(time_taken_in_ns, 1000000));
-	return false;
-}
-
-/*
- * dc_link_detect() - Detect if a sink is attached to a given link
- *
- * link->local_sink is created or destroyed as needed.
- *
- * This does not create remote sinks but will trigger DM
- * to start MST detection if a branch is detected.
- */
-static bool dc_link_detect_helper(struct dc_link *link,
-				  enum dc_detect_reason reason)
-{
-	struct dc_sink_init_data sink_init_data = { 0 };
-	struct display_sink_capability sink_caps = { 0 };
-	uint32_t i;
-	bool converter_disable_audio = false;
-	struct audio_support *aud_support = &link->dc->res_pool->audio_support;
-	bool same_edid = false;
-	enum dc_edid_status edid_status;
-	struct dc_context *dc_ctx = link->ctx;
-	struct dc_sink *sink = NULL;
-	struct dc_sink *prev_sink = NULL;
-	struct dpcd_caps prev_dpcd_caps;
-	enum dc_connection_type new_connection_type = dc_connection_none;
-	enum dc_connection_type pre_connection_type = dc_connection_none;
-	bool perform_dp_seamless_boot = false;
-	const uint32_t post_oui_delay = 30; // 30ms
-
-	DC_LOGGER_INIT(link->ctx->logger);
-
-	if (dc_is_virtual_signal(link->connector_signal))
-		return false;
-
-	if (((link->connector_signal == SIGNAL_TYPE_LVDS ||
-		link->connector_signal == SIGNAL_TYPE_EDP) &&
-		(!link->dc->config.allow_edp_hotplug_detection)) &&
-		link->local_sink) {
-		// need to re-write OUI and brightness in resume case
-		if (link->connector_signal == SIGNAL_TYPE_EDP) {
-			dpcd_set_source_specific_data(link);
-			msleep(post_oui_delay);
-			dc_link_set_default_brightness_aux(link);
-			//TODO: use cached
-		}
-
-		return true;
-	}
-
-	if (!dc_link_detect_sink(link, &new_connection_type)) {
-		BREAK_TO_DEBUGGER();
-		return false;
-	}
-
-	prev_sink = link->local_sink;
-	if (prev_sink) {
-		dc_sink_retain(prev_sink);
-		memcpy(&prev_dpcd_caps, &link->dpcd_caps, sizeof(struct dpcd_caps));
-	}
-
-	link_disconnect_sink(link);
-	if (new_connection_type != dc_connection_none) {
-		pre_connection_type = link->type;
-		link->type = new_connection_type;
-		link->link_state_valid = false;
-
-		/* From Disconnected-to-Connected. */
-		switch (link->connector_signal) {
-		case SIGNAL_TYPE_HDMI_TYPE_A: {
-			sink_caps.transaction_type = DDC_TRANSACTION_TYPE_I2C;
-			if (aud_support->hdmi_audio_native)
-				sink_caps.signal = SIGNAL_TYPE_HDMI_TYPE_A;
-			else
-				sink_caps.signal = SIGNAL_TYPE_DVI_SINGLE_LINK;
-			break;
-		}
-
-		case SIGNAL_TYPE_DVI_SINGLE_LINK: {
-			sink_caps.transaction_type = DDC_TRANSACTION_TYPE_I2C;
-			sink_caps.signal = SIGNAL_TYPE_DVI_SINGLE_LINK;
-			break;
-		}
-
-		case SIGNAL_TYPE_DVI_DUAL_LINK: {
-			sink_caps.transaction_type = DDC_TRANSACTION_TYPE_I2C;
-			sink_caps.signal = SIGNAL_TYPE_DVI_DUAL_LINK;
-			break;
-		}
-
-		case SIGNAL_TYPE_LVDS: {
-			sink_caps.transaction_type = DDC_TRANSACTION_TYPE_I2C;
-			sink_caps.signal = SIGNAL_TYPE_LVDS;
-			break;
-		}
-
-		case SIGNAL_TYPE_EDP: {
-			read_current_link_settings_on_detect(link);
-
-			detect_edp_sink_caps(link);
-			read_current_link_settings_on_detect(link);
-			sink_caps.transaction_type = DDC_TRANSACTION_TYPE_I2C_OVER_AUX;
-			sink_caps.signal = SIGNAL_TYPE_EDP;
-			break;
-		}
-
-		case SIGNAL_TYPE_DISPLAY_PORT: {
-			/* wa HPD high coming too early*/
-			if (link->ep_type == DISPLAY_ENDPOINT_PHY &&
-			    link->link_enc->features.flags.bits.DP_IS_USB_C == 1) {
-				/* if alt mode times out, return false */
-				if (!wait_for_entering_dp_alt_mode(link))
-					return false;
-			}
-
-			if (!detect_dp(link, &sink_caps, reason)) {
-				if (prev_sink)
-					dc_sink_release(prev_sink);
-				return false;
-			}
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-			if (dp_get_link_encoding_format(&link->reported_link_cap) == DP_128b_132b_ENCODING)
-				add_dp_hpo_link_encoder_to_link(link);
-#endif
-
-			if (link->type == dc_connection_mst_branch) {
-				LINK_INFO("link=%d, mst branch is now Connected\n",
-					  link->link_index);
-				/* Need to setup mst link_cap struct here
-				 * otherwise dc_link_detect() will leave mst link_cap
-				 * empty which leads to allocate_mst_payload() has "0"
-				 * pbn_per_slot value leading to exception on dc_fixpt_div()
-				 */
-				dp_verify_mst_link_cap(link);
-
-				/*
-				 * This call will initiate MST topology discovery. Which
-				 * will detect MST ports and add new DRM connector DRM
-				 * framework. Then read EDID via remote i2c over aux. In
-				 * the end, will notify DRM detect result and save EDID
-				 * into DRM framework.
-				 *
-				 * .detect is called by .fill_modes.
-				 * .fill_modes is called by user mode ioctl
-				 * DRM_IOCTL_MODE_GETCONNECTOR.
-				 *
-				 * .get_modes is called by .fill_modes.
-				 *
-				 * call .get_modes, AMDGPU DM implementation will create
-				 * new dc_sink and add to dc_link. For long HPD plug
-				 * in/out, MST has its own handle.
-				 *
-				 * Therefore, just after dc_create, link->sink is not
-				 * created for MST until user mode app calls
-				 * DRM_IOCTL_MODE_GETCONNECTOR.
-				 *
-				 * Need check ->sink usages in case ->sink = NULL
-				 * TODO: s3 resume check
-				 */
-
-				dm_helpers_dp_update_branch_info(link->ctx, link);
-				if (dm_helpers_dp_mst_start_top_mgr(link->ctx,
-						link, reason == DETECT_REASON_BOOT)) {
-					if (prev_sink)
-						dc_sink_release(prev_sink);
-					return false;
-				} else {
-					link->type = dc_connection_sst_branch;
-					sink_caps.signal = SIGNAL_TYPE_DISPLAY_PORT;
-				}
-			}
-
-			/* Active SST downstream branch device unplug*/
-			if (link->type == dc_connection_sst_branch &&
-			    link->dpcd_caps.sink_count.bits.SINK_COUNT == 0) {
-				if (prev_sink)
-					/* Downstream unplug */
-					dc_sink_release(prev_sink);
-				return true;
-			}
-
-			/* disable audio for non DP to HDMI active sst converter */
-			if (link->type == dc_connection_sst_branch &&
-					is_dp_active_dongle(link) &&
-					(link->dpcd_caps.dongle_type !=
-							DISPLAY_DONGLE_DP_HDMI_CONVERTER))
-				converter_disable_audio = true;
-
-			// link switch from MST to non-MST stop topology manager
-			if (pre_connection_type == dc_connection_mst_branch &&
-					link->type != dc_connection_mst_branch)
-				dm_helpers_dp_mst_stop_top_mgr(link->ctx, link);
-
-
-			// For seamless boot, to skip verify link cap, we read UEFI settings and set them as verified.
-			if (reason == DETECT_REASON_BOOT &&
-					!dc_ctx->dc->config.power_down_display_on_boot &&
-					link->link_status.link_active)
-				perform_dp_seamless_boot = true;
-
-			if (perform_dp_seamless_boot) {
-				read_current_link_settings_on_detect(link);
-				link->verified_link_cap = link->reported_link_cap;
-			}
-
-			break;
-		}
-
-		default:
-			DC_ERROR("Invalid connector type! signal:%d\n",
-				 link->connector_signal);
-			if (prev_sink)
-				dc_sink_release(prev_sink);
-			return false;
-		} /* switch() */
-
-		if (link->dpcd_caps.sink_count.bits.SINK_COUNT)
-			link->dpcd_sink_count =
-				link->dpcd_caps.sink_count.bits.SINK_COUNT;
-		else
-			link->dpcd_sink_count = 1;
-
-		dal_ddc_service_set_transaction_type(link->ddc,
-						     sink_caps.transaction_type);
-
-		link->aux_mode =
-			dal_ddc_service_is_in_aux_transaction_mode(link->ddc);
-
-		sink_init_data.link = link;
-		sink_init_data.sink_signal = sink_caps.signal;
-
-		sink = dc_sink_create(&sink_init_data);
-		if (!sink) {
-			DC_ERROR("Failed to create sink!\n");
-			if (prev_sink)
-				dc_sink_release(prev_sink);
-			return false;
-		}
-
-		sink->link->dongle_max_pix_clk = sink_caps.max_hdmi_pixel_clock;
-		sink->converter_disable_audio = converter_disable_audio;
-
-		/* dc_sink_create returns a new reference */
-		link->local_sink = sink;
-
-		edid_status = dm_helpers_read_local_edid(link->ctx,
-							 link, sink);
-
-		switch (edid_status) {
-		case EDID_BAD_CHECKSUM:
-			DC_LOG_ERROR("EDID checksum invalid.\n");
-			break;
-		case EDID_NO_RESPONSE:
-			DC_LOG_ERROR("No EDID read.\n");
-			/*
-			 * Abort detection for non-DP connectors if we have
-			 * no EDID
-			 *
-			 * DP needs to report as connected if HDP is high
-			 * even if we have no EDID in order to go to
-			 * fail-safe mode
-			 */
-			if (dc_is_hdmi_signal(link->connector_signal) ||
-			    dc_is_dvi_signal(link->connector_signal)) {
-				if (prev_sink)
-					dc_sink_release(prev_sink);
-
-				return false;
-			}
-			break;
-		default:
-			break;
-		}
-
-		// Check if edid is the same
-		if ((prev_sink) &&
-		    (edid_status == EDID_THE_SAME || edid_status == EDID_OK))
-			same_edid = is_same_edid(&prev_sink->dc_edid,
-						 &sink->dc_edid);
-
-		if (sink->edid_caps.panel_patch.skip_scdc_overwrite)
-			link->ctx->dc->debug.hdmi20_disable = true;
-
-		if (link->connector_signal == SIGNAL_TYPE_DISPLAY_PORT &&
-		    sink_caps.transaction_type ==
-		    DDC_TRANSACTION_TYPE_I2C_OVER_AUX) {
-			/*
-			 * TODO debug why Dell 2413 doesn't like
-			 *  two link trainings
-			 */
-#if defined(CONFIG_DRM_AMD_DC_HDCP)
-			query_hdcp_capability(sink->sink_signal, link);
-#endif
-
-			// verify link cap for SST non-seamless boot
-			if (!perform_dp_seamless_boot)
-				dp_verify_link_cap_with_retries(link,
-								&link->reported_link_cap,
-								LINK_TRAINING_MAX_VERIFY_RETRY);
-		} else {
-			// If edid is the same, then discard new sink and revert back to original sink
-			if (same_edid) {
-				link_disconnect_remap(prev_sink, link);
-				sink = prev_sink;
-				prev_sink = NULL;
-			}
-#if defined(CONFIG_DRM_AMD_DC_HDCP)
-			query_hdcp_capability(sink->sink_signal, link);
-#endif
-		}
-
-		/* HDMI-DVI Dongle */
-		if (sink->sink_signal == SIGNAL_TYPE_HDMI_TYPE_A &&
-		    !sink->edid_caps.edid_hdmi)
-			sink->sink_signal = SIGNAL_TYPE_DVI_SINGLE_LINK;
-
-		/* Connectivity log: detection */
-		for (i = 0; i < sink->dc_edid.length / DC_EDID_BLOCK_SIZE; i++) {
-			CONN_DATA_DETECT(link,
-					 &sink->dc_edid.raw_edid[i * DC_EDID_BLOCK_SIZE],
-					 DC_EDID_BLOCK_SIZE,
-					 "%s: [Block %d] ", sink->edid_caps.display_name, i);
-		}
-
-		DC_LOG_DETECTION_EDID_PARSER("%s: "
-			"manufacturer_id = %X, "
-			"product_id = %X, "
-			"serial_number = %X, "
-			"manufacture_week = %d, "
-			"manufacture_year = %d, "
-			"display_name = %s, "
-			"speaker_flag = %d, "
-			"audio_mode_count = %d\n",
-			__func__,
-			sink->edid_caps.manufacturer_id,
-			sink->edid_caps.product_id,
-			sink->edid_caps.serial_number,
-			sink->edid_caps.manufacture_week,
-			sink->edid_caps.manufacture_year,
-			sink->edid_caps.display_name,
-			sink->edid_caps.speaker_flags,
-			sink->edid_caps.audio_mode_count);
-
-		for (i = 0; i < sink->edid_caps.audio_mode_count; i++) {
-			DC_LOG_DETECTION_EDID_PARSER("%s: mode number = %d, "
-				"format_code = %d, "
-				"channel_count = %d, "
-				"sample_rate = %d, "
-				"sample_size = %d\n",
-				__func__,
-				i,
-				sink->edid_caps.audio_modes[i].format_code,
-				sink->edid_caps.audio_modes[i].channel_count,
-				sink->edid_caps.audio_modes[i].sample_rate,
-				sink->edid_caps.audio_modes[i].sample_size);
-		}
-	} else {
-		/* From Connected-to-Disconnected. */
-		if (link->type == dc_connection_mst_branch) {
-			LINK_INFO("link=%d, mst branch is now Disconnected\n",
-				  link->link_index);
-
-			/* Disable work around which keeps DSC on for tunneled MST on certain USB4 docks. */
-			if (link->ep_type == DISPLAY_ENDPOINT_USB4_DPIA)
-				link->wa_flags.dpia_mst_dsc_always_on = false;
-
-			dm_helpers_dp_mst_stop_top_mgr(link->ctx, link);
-
-			link->mst_stream_alloc_table.stream_count = 0;
-			memset(link->mst_stream_alloc_table.stream_allocations,
-			       0,
-			       sizeof(link->mst_stream_alloc_table.stream_allocations));
-		}
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-		if (dp_get_link_encoding_format(&link->cur_link_settings) == DP_128b_132b_ENCODING)
-			reset_dp_hpo_stream_encoders_for_link(link);
-#endif
-
-		link->type = dc_connection_none;
-		sink_caps.signal = SIGNAL_TYPE_NONE;
-		/* When we unplug a passive DP-HDMI dongle connection, dongle_max_pix_clk
-		 *  is not cleared. If we emulate a DP signal on this connection, it thinks
-		 *  the dongle is still there and limits the number of modes we can emulate.
-		 *  Clear dongle_max_pix_clk on disconnect to fix this
-		 */
-		link->dongle_max_pix_clk = 0;
-	}
-
-	LINK_INFO("link=%d, dc_sink_in=%p is now %s prev_sink=%p edid same=%d\n",
-		  link->link_index, sink,
-		  (sink_caps.signal ==
-		   SIGNAL_TYPE_NONE ? "Disconnected" : "Connected"),
-		  prev_sink, same_edid);
-
-	if (prev_sink)
-		dc_sink_release(prev_sink);
-
-	return true;
-}
-
-bool dc_link_detect(struct dc_link *link, enum dc_detect_reason reason)
-{
-	const struct dc *dc = link->dc;
-	bool ret;
-	bool can_apply_seamless_boot = false;
-	int i;
-
-	for (i = 0; i < dc->current_state->stream_count; i++) {
-		if (dc->current_state->streams[i]->apply_seamless_boot_optimization) {
-			can_apply_seamless_boot = true;
-			break;
-		}
-	}
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	dc_z10_restore(dc);
-#endif
-
-	/* get out of low power state */
-	if (!can_apply_seamless_boot && reason != DETECT_REASON_BOOT)
-		clk_mgr_exit_optimized_pwr_state(dc, dc->clk_mgr);
-
-	ret = dc_link_detect_helper(link, reason);
-
-	/* Go back to power optimized state */
-	if (!can_apply_seamless_boot && reason != DETECT_REASON_BOOT)
-		clk_mgr_optimize_pwr_state(dc, dc->clk_mgr);
-
-	return ret;
-}
-
-bool dc_link_get_hpd_state(struct dc_link *dc_link)
-{
-	uint32_t state;
-
-	dal_gpio_lock_pin(dc_link->hpd_gpio);
-	dal_gpio_get_value(dc_link->hpd_gpio, &state);
-	dal_gpio_unlock_pin(dc_link->hpd_gpio);
-
-	return state;
-}
-
-static enum hpd_source_id get_hpd_line(struct dc_link *link)
-{
-	struct gpio *hpd;
-	enum hpd_source_id hpd_id = HPD_SOURCEID_UNKNOWN;
-
-	hpd = get_hpd_gpio(link->ctx->dc_bios, link->link_id,
-			   link->ctx->gpio_service);
-
-	if (hpd) {
-		switch (dal_irq_get_source(hpd)) {
-		case DC_IRQ_SOURCE_HPD1:
-			hpd_id = HPD_SOURCEID1;
-		break;
-		case DC_IRQ_SOURCE_HPD2:
-			hpd_id = HPD_SOURCEID2;
-		break;
-		case DC_IRQ_SOURCE_HPD3:
-			hpd_id = HPD_SOURCEID3;
-		break;
-		case DC_IRQ_SOURCE_HPD4:
-			hpd_id = HPD_SOURCEID4;
-		break;
-		case DC_IRQ_SOURCE_HPD5:
-			hpd_id = HPD_SOURCEID5;
-		break;
-		case DC_IRQ_SOURCE_HPD6:
-			hpd_id = HPD_SOURCEID6;
-		break;
-		default:
-			BREAK_TO_DEBUGGER();
-		break;
-		}
-
-		dal_gpio_destroy_irq(&hpd);
-	}
-
-	return hpd_id;
-}
-
-static enum channel_id get_ddc_line(struct dc_link *link)
-{
-	struct ddc *ddc;
-	enum channel_id channel = CHANNEL_ID_UNKNOWN;
-
-	ddc = dal_ddc_service_get_ddc_pin(link->ddc);
-
-	if (ddc) {
-		switch (dal_ddc_get_line(ddc)) {
-		case GPIO_DDC_LINE_DDC1:
-			channel = CHANNEL_ID_DDC1;
-			break;
-		case GPIO_DDC_LINE_DDC2:
-			channel = CHANNEL_ID_DDC2;
-			break;
-		case GPIO_DDC_LINE_DDC3:
-			channel = CHANNEL_ID_DDC3;
-			break;
-		case GPIO_DDC_LINE_DDC4:
-			channel = CHANNEL_ID_DDC4;
-			break;
-		case GPIO_DDC_LINE_DDC5:
-			channel = CHANNEL_ID_DDC5;
-			break;
-		case GPIO_DDC_LINE_DDC6:
-			channel = CHANNEL_ID_DDC6;
-			break;
-		case GPIO_DDC_LINE_DDC_VGA:
-			channel = CHANNEL_ID_DDC_VGA;
-			break;
-		case GPIO_DDC_LINE_I2C_PAD:
-			channel = CHANNEL_ID_I2C_PAD;
-			break;
-		default:
-			BREAK_TO_DEBUGGER();
-			break;
-		}
-	}
-
-	return channel;
-}
-
-static enum transmitter translate_encoder_to_transmitter(struct graphics_object_id encoder)
-{
-	switch (encoder.id) {
-	case ENCODER_ID_INTERNAL_UNIPHY:
-		switch (encoder.enum_id) {
-		case ENUM_ID_1:
-			return TRANSMITTER_UNIPHY_A;
-		case ENUM_ID_2:
-			return TRANSMITTER_UNIPHY_B;
-		default:
-			return TRANSMITTER_UNKNOWN;
-		}
-	break;
-	case ENCODER_ID_INTERNAL_UNIPHY1:
-		switch (encoder.enum_id) {
-		case ENUM_ID_1:
-			return TRANSMITTER_UNIPHY_C;
-		case ENUM_ID_2:
-			return TRANSMITTER_UNIPHY_D;
-		default:
-			return TRANSMITTER_UNKNOWN;
-		}
-	break;
-	case ENCODER_ID_INTERNAL_UNIPHY2:
-		switch (encoder.enum_id) {
-		case ENUM_ID_1:
-			return TRANSMITTER_UNIPHY_E;
-		case ENUM_ID_2:
-			return TRANSMITTER_UNIPHY_F;
-		default:
-			return TRANSMITTER_UNKNOWN;
-		}
-	break;
-	case ENCODER_ID_INTERNAL_UNIPHY3:
-		switch (encoder.enum_id) {
-		case ENUM_ID_1:
-			return TRANSMITTER_UNIPHY_G;
-		default:
-			return TRANSMITTER_UNKNOWN;
-		}
-	break;
-	case ENCODER_ID_EXTERNAL_NUTMEG:
-		switch (encoder.enum_id) {
-		case ENUM_ID_1:
-			return TRANSMITTER_NUTMEG_CRT;
-		default:
-			return TRANSMITTER_UNKNOWN;
-		}
-	break;
-	case ENCODER_ID_EXTERNAL_TRAVIS:
-		switch (encoder.enum_id) {
-		case ENUM_ID_1:
-			return TRANSMITTER_TRAVIS_CRT;
-		case ENUM_ID_2:
-			return TRANSMITTER_TRAVIS_LCD;
-		default:
-			return TRANSMITTER_UNKNOWN;
-		}
-	break;
-	default:
-		return TRANSMITTER_UNKNOWN;
-	}
-}
-
-static bool dc_link_construct_legacy(struct dc_link *link,
-				     const struct link_init_data *init_params)
-{
-	uint8_t i;
-	struct ddc_service_init_data ddc_service_init_data = { { 0 } };
-	struct dc_context *dc_ctx = init_params->ctx;
-	struct encoder_init_data enc_init_data = { 0 };
-	struct panel_cntl_init_data panel_cntl_init_data = { 0 };
-	struct integrated_info *info;
-	struct dc_bios *bios = init_params->dc->ctx->dc_bios;
-	const struct dc_vbios_funcs *bp_funcs = bios->funcs;
-	struct bp_disp_connector_caps_info disp_connect_caps_info = { 0 };
-
-	DC_LOGGER_INIT(dc_ctx->logger);
-
-	info = kzalloc(sizeof(*info), GFP_KERNEL);
-	if (!info)
-		goto create_fail;
-
-	link->irq_source_hpd = DC_IRQ_SOURCE_INVALID;
-	link->irq_source_hpd_rx = DC_IRQ_SOURCE_INVALID;
-
-	link->link_status.dpcd_caps = &link->dpcd_caps;
-
-	link->dc = init_params->dc;
-	link->ctx = dc_ctx;
-	link->link_index = init_params->link_index;
-
-	memset(&link->preferred_training_settings, 0,
-	       sizeof(struct dc_link_training_overrides));
-	memset(&link->preferred_link_setting, 0,
-	       sizeof(struct dc_link_settings));
-
-	link->link_id =
-		bios->funcs->get_connector_id(bios, init_params->connector_index);
-
-	link->ep_type = DISPLAY_ENDPOINT_PHY;
-
-	DC_LOG_DC("BIOS object table - link_id: %d", link->link_id.id);
-
-	if (bios->funcs->get_disp_connector_caps_info) {
-		bios->funcs->get_disp_connector_caps_info(bios, link->link_id, &disp_connect_caps_info);
-		link->is_internal_display = disp_connect_caps_info.INTERNAL_DISPLAY;
-		DC_LOG_DC("BIOS object table - is_internal_display: %d", link->is_internal_display);
-	}
-
-	if (link->link_id.type != OBJECT_TYPE_CONNECTOR) {
-		dm_output_to_console("%s: Invalid Connector ObjectID from Adapter Service for connector index:%d! type %d expected %d\n",
-				     __func__, init_params->connector_index,
-				     link->link_id.type, OBJECT_TYPE_CONNECTOR);
-		goto create_fail;
-	}
-
-	if (link->dc->res_pool->funcs->link_init)
-		link->dc->res_pool->funcs->link_init(link);
-
-	link->hpd_gpio = get_hpd_gpio(link->ctx->dc_bios, link->link_id,
-				      link->ctx->gpio_service);
-
-	if (link->hpd_gpio) {
-		dal_gpio_open(link->hpd_gpio, GPIO_MODE_INTERRUPT);
-		dal_gpio_unlock_pin(link->hpd_gpio);
-		link->irq_source_hpd = dal_irq_get_source(link->hpd_gpio);
-
-		DC_LOG_DC("BIOS object table - hpd_gpio id: %d", link->hpd_gpio->id);
-		DC_LOG_DC("BIOS object table - hpd_gpio en: %d", link->hpd_gpio->en);
-	}
-
-	switch (link->link_id.id) {
-	case CONNECTOR_ID_HDMI_TYPE_A:
-		link->connector_signal = SIGNAL_TYPE_HDMI_TYPE_A;
-
-		break;
-	case CONNECTOR_ID_SINGLE_LINK_DVID:
-	case CONNECTOR_ID_SINGLE_LINK_DVII:
-		link->connector_signal = SIGNAL_TYPE_DVI_SINGLE_LINK;
-		break;
-	case CONNECTOR_ID_DUAL_LINK_DVID:
-	case CONNECTOR_ID_DUAL_LINK_DVII:
-		link->connector_signal = SIGNAL_TYPE_DVI_DUAL_LINK;
-		break;
-	case CONNECTOR_ID_DISPLAY_PORT:
-		link->connector_signal = SIGNAL_TYPE_DISPLAY_PORT;
-
-		if (link->hpd_gpio)
-			link->irq_source_hpd_rx =
-					dal_irq_get_rx_source(link->hpd_gpio);
-
-		break;
-	case CONNECTOR_ID_EDP:
-		link->connector_signal = SIGNAL_TYPE_EDP;
-
-		if (link->hpd_gpio) {
-			if (!link->dc->config.allow_edp_hotplug_detection)
-				link->irq_source_hpd = DC_IRQ_SOURCE_INVALID;
-			link->irq_source_hpd_rx =
-					dal_irq_get_rx_source(link->hpd_gpio);
-		}
-
-		break;
-	case CONNECTOR_ID_LVDS:
-		link->connector_signal = SIGNAL_TYPE_LVDS;
-		break;
-	default:
-		DC_LOG_WARNING("Unsupported Connector type:%d!\n",
-			       link->link_id.id);
-		goto create_fail;
-	}
-
-	/* TODO: #DAL3 Implement id to str function.*/
-	LINK_INFO("Connector[%d] description:"
-		  "signal %d\n",
-		  init_params->connector_index,
-		  link->connector_signal);
-
-	ddc_service_init_data.ctx = link->ctx;
-	ddc_service_init_data.id = link->link_id;
-	ddc_service_init_data.link = link;
-	link->ddc = dal_ddc_service_create(&ddc_service_init_data);
-
-	if (!link->ddc) {
-		DC_ERROR("Failed to create ddc_service!\n");
-		goto ddc_create_fail;
-	}
-
-	if (!link->ddc->ddc_pin) {
-		DC_ERROR("Failed to get I2C info for connector!\n");
-		goto ddc_create_fail;
-	}
-
-	link->ddc_hw_inst =
-		dal_ddc_get_line(dal_ddc_service_get_ddc_pin(link->ddc));
-
-
-	if (link->dc->res_pool->funcs->panel_cntl_create &&
-		(link->link_id.id == CONNECTOR_ID_EDP ||
-			link->link_id.id == CONNECTOR_ID_LVDS)) {
-		panel_cntl_init_data.ctx = dc_ctx;
-		panel_cntl_init_data.inst =
-			panel_cntl_init_data.ctx->dc_edp_id_count;
-		link->panel_cntl =
-			link->dc->res_pool->funcs->panel_cntl_create(
-								&panel_cntl_init_data);
-		panel_cntl_init_data.ctx->dc_edp_id_count++;
-
-		if (link->panel_cntl == NULL) {
-			DC_ERROR("Failed to create link panel_cntl!\n");
-			goto panel_cntl_create_fail;
-		}
-	}
-
-	enc_init_data.ctx = dc_ctx;
-	bp_funcs->get_src_obj(dc_ctx->dc_bios, link->link_id, 0,
-			      &enc_init_data.encoder);
-	enc_init_data.connector = link->link_id;
-	enc_init_data.channel = get_ddc_line(link);
-	enc_init_data.hpd_source = get_hpd_line(link);
-
-	link->hpd_src = enc_init_data.hpd_source;
-
-	enc_init_data.transmitter =
-		translate_encoder_to_transmitter(enc_init_data.encoder);
-	link->link_enc =
-		link->dc->res_pool->funcs->link_enc_create(&enc_init_data);
-
-	if (!link->link_enc) {
-		DC_ERROR("Failed to create link encoder!\n");
-		goto link_enc_create_fail;
-	}
-
-	DC_LOG_DC("BIOS object table - DP_IS_USB_C: %d", link->link_enc->features.flags.bits.DP_IS_USB_C);
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	DC_LOG_DC("BIOS object table - IS_DP2_CAPABLE: %d", link->link_enc->features.flags.bits.IS_DP2_CAPABLE);
-#endif
-
-	/* Update link encoder tracking variables. These are used for the dynamic
-	 * assignment of link encoders to streams.
-	 */
-	link->eng_id = link->link_enc->preferred_engine;
-	link->dc->res_pool->link_encoders[link->eng_id - ENGINE_ID_DIGA] = link->link_enc;
-	link->dc->res_pool->dig_link_enc_count++;
-
-	link->link_enc_hw_inst = link->link_enc->transmitter;
-
-	for (i = 0; i < 4; i++) {
-		if (bp_funcs->get_device_tag(dc_ctx->dc_bios,
-					     link->link_id, i,
-					     &link->device_tag) != BP_RESULT_OK) {
-			DC_ERROR("Failed to find device tag!\n");
-			goto device_tag_fail;
-		}
-
-		/* Look for device tag that matches connector signal,
-		 * CRT for rgb, LCD for other supported signal tyes
-		 */
-		if (!bp_funcs->is_device_id_supported(dc_ctx->dc_bios,
-						      link->device_tag.dev_id))
-			continue;
-		if (link->device_tag.dev_id.device_type == DEVICE_TYPE_CRT &&
-		    link->connector_signal != SIGNAL_TYPE_RGB)
-			continue;
-		if (link->device_tag.dev_id.device_type == DEVICE_TYPE_LCD &&
-		    link->connector_signal == SIGNAL_TYPE_RGB)
-			continue;
-
-		DC_LOG_DC("BIOS object table - device_tag.acpi_device: %d", link->device_tag.acpi_device);
-		DC_LOG_DC("BIOS object table - device_tag.dev_id.device_type: %d", link->device_tag.dev_id.device_type);
-		DC_LOG_DC("BIOS object table - device_tag.dev_id.enum_id: %d", link->device_tag.dev_id.enum_id);
-		break;
-	}
-
-	if (bios->integrated_info)
-		memcpy(info, bios->integrated_info, sizeof(*info));
-
-	/* Look for channel mapping corresponding to connector and device tag */
-	for (i = 0; i < MAX_NUMBER_OF_EXT_DISPLAY_PATH; i++) {
-		struct external_display_path *path =
-			&info->ext_disp_conn_info.path[i];
-
-		if (path->device_connector_id.enum_id == link->link_id.enum_id &&
-		    path->device_connector_id.id == link->link_id.id &&
-		    path->device_connector_id.type == link->link_id.type) {
-			if (link->device_tag.acpi_device != 0 &&
-			    path->device_acpi_enum == link->device_tag.acpi_device) {
-				link->ddi_channel_mapping = path->channel_mapping;
-				link->chip_caps = path->caps;
-				DC_LOG_DC("BIOS object table - ddi_channel_mapping: 0x%04X", link->ddi_channel_mapping.raw);
-				DC_LOG_DC("BIOS object table - chip_caps: %d", link->chip_caps);
-			} else if (path->device_tag ==
-				   link->device_tag.dev_id.raw_device_tag) {
-				link->ddi_channel_mapping = path->channel_mapping;
-				link->chip_caps = path->caps;
-				DC_LOG_DC("BIOS object table - ddi_channel_mapping: 0x%04X", link->ddi_channel_mapping.raw);
-				DC_LOG_DC("BIOS object table - chip_caps: %d", link->chip_caps);
-			}
-
-			if (link->chip_caps & EXT_DISPLAY_PATH_CAPS__DP_FIXED_VS_EN) {
-				link->bios_forced_drive_settings.VOLTAGE_SWING =
-						(info->ext_disp_conn_info.fixdpvoltageswing & 0x3);
-				link->bios_forced_drive_settings.PRE_EMPHASIS =
-						((info->ext_disp_conn_info.fixdpvoltageswing >> 2) & 0x3);
-			}
-
-			break;
-		}
-	}
-
-	if (bios->funcs->get_atom_dc_golden_table)
-		bios->funcs->get_atom_dc_golden_table(bios);
-
-	/*
-	 * TODO check if GPIO programmed correctly
-	 *
-	 * If GPIO isn't programmed correctly HPD might not rise or drain
-	 * fast enough, leading to bounces.
-	 */
-	program_hpd_filter(link);
-
-	link->psr_settings.psr_version = DC_PSR_VERSION_UNSUPPORTED;
-
-	DC_LOG_DC("BIOS object table - %s finished successfully.\n", __func__);
-	kfree(info);
-	return true;
-device_tag_fail:
-	link->link_enc->funcs->destroy(&link->link_enc);
-link_enc_create_fail:
-	if (link->panel_cntl != NULL)
-		link->panel_cntl->funcs->destroy(&link->panel_cntl);
-panel_cntl_create_fail:
-	dal_ddc_service_destroy(&link->ddc);
-ddc_create_fail:
-create_fail:
-
-	if (link->hpd_gpio) {
-		dal_gpio_destroy_irq(&link->hpd_gpio);
-		link->hpd_gpio = NULL;
-	}
-
-	DC_LOG_DC("BIOS object table - %s failed.\n", __func__);
-	kfree(info);
-
-	return false;
-}
-
-static bool dc_link_construct_dpia(struct dc_link *link,
-				   const struct link_init_data *init_params)
-{
-	struct ddc_service_init_data ddc_service_init_data = { { 0 } };
-	struct dc_context *dc_ctx = init_params->ctx;
-
-	DC_LOGGER_INIT(dc_ctx->logger);
-
-	/* Initialized irq source for hpd and hpd rx */
-	link->irq_source_hpd = DC_IRQ_SOURCE_INVALID;
-	link->irq_source_hpd_rx = DC_IRQ_SOURCE_INVALID;
-	link->link_status.dpcd_caps = &link->dpcd_caps;
-
-	link->dc = init_params->dc;
-	link->ctx = dc_ctx;
-	link->link_index = init_params->link_index;
-
-	memset(&link->preferred_training_settings, 0,
-	       sizeof(struct dc_link_training_overrides));
-	memset(&link->preferred_link_setting, 0,
-	       sizeof(struct dc_link_settings));
-
-	/* Dummy Init for linkid */
-	link->link_id.type = OBJECT_TYPE_CONNECTOR;
-	link->link_id.id = CONNECTOR_ID_DISPLAY_PORT;
-	link->link_id.enum_id = ENUM_ID_1 + init_params->connector_index;
-	link->is_internal_display = false;
-	link->connector_signal = SIGNAL_TYPE_DISPLAY_PORT;
-	LINK_INFO("Connector[%d] description:signal %d\n",
-		  init_params->connector_index,
-		  link->connector_signal);
-
-	link->ep_type = DISPLAY_ENDPOINT_USB4_DPIA;
-	link->is_dig_mapping_flexible = true;
-
-	/* TODO: Initialize link : funcs->link_init */
-
-	ddc_service_init_data.ctx = link->ctx;
-	ddc_service_init_data.id = link->link_id;
-	ddc_service_init_data.link = link;
-	/* Set indicator for dpia link so that ddc won't be created */
-	ddc_service_init_data.is_dpia_link = true;
-
-	link->ddc = dal_ddc_service_create(&ddc_service_init_data);
-	if (!link->ddc) {
-		DC_ERROR("Failed to create ddc_service!\n");
-		goto ddc_create_fail;
-	}
-
-	/* Set dpia port index : 0 to number of dpia ports */
-	link->ddc_hw_inst = init_params->connector_index;
-
-	/* TODO: Create link encoder */
-
-	link->psr_settings.psr_version = DC_PSR_VERSION_UNSUPPORTED;
-
-	/* Some docks seem to NAK I2C writes to segment pointer with mot=0. */
-	link->wa_flags.dp_mot_reset_segment = true;
-
-	return true;
-
-ddc_create_fail:
-	return false;
-}
-
-static bool dc_link_construct(struct dc_link *link,
-			      const struct link_init_data *init_params)
-{
-	/* Handle dpia case */
-	if (init_params->is_dpia_link)
-		return dc_link_construct_dpia(link, init_params);
-	else
-		return dc_link_construct_legacy(link, init_params);
-}
-/*******************************************************************************
- * Public functions
- ******************************************************************************/
-struct dc_link *link_create(const struct link_init_data *init_params)
-{
-	struct dc_link *link =
-			kzalloc(sizeof(*link), GFP_KERNEL);
-
-	if (NULL == link)
-		goto alloc_fail;
-
-	if (false == dc_link_construct(link, init_params))
-		goto construct_fail;
-
-	/*
-	 * Must use preferred_link_setting, not reported_link_cap or verified_link_cap,
-	 * since struct preferred_link_setting won't be reset after S3.
-	 */
-	link->preferred_link_setting.dpcd_source_device_specific_field_support = true;
-
-	return link;
-
-construct_fail:
-	kfree(link);
-
-alloc_fail:
-	return NULL;
-}
-
-void link_destroy(struct dc_link **link)
-{
-	dc_link_destruct(*link);
-	kfree(*link);
-	*link = NULL;
-}
-
-static void enable_stream_features(struct pipe_ctx *pipe_ctx)
-{
-	struct dc_stream_state *stream = pipe_ctx->stream;
-
-	if (pipe_ctx->stream->signal != SIGNAL_TYPE_DISPLAY_PORT_MST) {
-		struct dc_link *link = stream->link;
-		union down_spread_ctrl old_downspread;
-		union down_spread_ctrl new_downspread;
-
-		core_link_read_dpcd(link, DP_DOWNSPREAD_CTRL,
-				&old_downspread.raw, sizeof(old_downspread));
-
-		new_downspread.raw = old_downspread.raw;
-
-		new_downspread.bits.IGNORE_MSA_TIMING_PARAM =
-				(stream->ignore_msa_timing_param) ? 1 : 0;
-
-		if (new_downspread.raw != old_downspread.raw) {
-			core_link_write_dpcd(link, DP_DOWNSPREAD_CTRL,
-				&new_downspread.raw, sizeof(new_downspread));
-		}
-
-	} else {
-		dm_helpers_mst_enable_stream_features(stream);
-	}
-}
-
-static enum dc_status enable_link_dp(struct dc_state *state,
-				     struct pipe_ctx *pipe_ctx)
-{
-	struct dc_stream_state *stream = pipe_ctx->stream;
-	enum dc_status status;
-	bool skip_video_pattern;
-	struct dc_link *link = stream->link;
-	struct dc_link_settings link_settings = {0};
-	bool fec_enable;
-	int i;
-	bool apply_seamless_boot_optimization = false;
-	uint32_t bl_oled_enable_delay = 50; // in ms
-	const uint32_t post_oui_delay = 30; // 30ms
-	/* Reduce link bandwidth between failed link training attempts. */
-	bool do_fallback = false;
-
-	// check for seamless boot
-	for (i = 0; i < state->stream_count; i++) {
-		if (state->streams[i]->apply_seamless_boot_optimization) {
-			apply_seamless_boot_optimization = true;
-			break;
-		}
-	}
-
-	/* get link settings for video mode timing */
-	decide_link_settings(stream, &link_settings);
-
-	/* Train with fallback when enabling DPIA link. Conventional links are
-	 * trained with fallback during sink detection.
-	 */
-	if (link->ep_type == DISPLAY_ENDPOINT_USB4_DPIA)
-		do_fallback = true;
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	/*
-	 * Temporary w/a to get DP2.0 link rates to work with SST.
-	 * TODO DP2.0 - Workaround: Remove w/a if and when the issue is resolved.
-	 */
-	if (dp_get_link_encoding_format(&link_settings) == DP_128b_132b_ENCODING &&
-			pipe_ctx->stream->signal == SIGNAL_TYPE_DISPLAY_PORT &&
-			link->dc->debug.set_mst_en_for_sst) {
-		dp_enable_mst_on_sink(link, true);
-	}
-#endif
-
-	if (pipe_ctx->stream->signal == SIGNAL_TYPE_EDP) {
-		/*in case it is not on*/
-		link->dc->hwss.edp_power_control(link, true);
-		link->dc->hwss.edp_wait_for_hpd_ready(link, true);
-	}
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	if (dp_get_link_encoding_format(&link_settings) == DP_128b_132b_ENCODING) {
-		/* TODO - DP2.0 HW: calculate 32 symbol clock for HPO encoder */
-	} else {
-		pipe_ctx->stream_res.pix_clk_params.requested_sym_clk =
-				link_settings.link_rate * LINK_RATE_REF_FREQ_IN_KHZ;
-		if (state->clk_mgr && !apply_seamless_boot_optimization)
-			state->clk_mgr->funcs->update_clocks(state->clk_mgr,
-					state, false);
-	}
-#else
-	pipe_ctx->stream_res.pix_clk_params.requested_sym_clk =
-		link_settings.link_rate * LINK_RATE_REF_FREQ_IN_KHZ;
-	if (state->clk_mgr && !apply_seamless_boot_optimization)
-		state->clk_mgr->funcs->update_clocks(state->clk_mgr,
-											state, false);
-#endif
-
-	// during mode switch we do DP_SET_POWER off then on, and OUI is lost
-	dpcd_set_source_specific_data(link);
-	if (link->dpcd_sink_ext_caps.raw != 0)
-		msleep(post_oui_delay);
-
-	skip_video_pattern = true;
-
-	if (link_settings.link_rate == LINK_RATE_LOW)
-		skip_video_pattern = false;
-
-	if (perform_link_training_with_retries(&link_settings,
-					       skip_video_pattern,
-					       LINK_TRAINING_ATTEMPTS,
-					       pipe_ctx,
-					       pipe_ctx->stream->signal,
-					       do_fallback)) {
-		link->cur_link_settings = link_settings;
-		status = DC_OK;
-	} else {
-		status = DC_FAIL_DP_LINK_TRAINING;
-	}
-
-	if (link->preferred_training_settings.fec_enable)
-		fec_enable = *link->preferred_training_settings.fec_enable;
-	else
-		fec_enable = true;
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	if (dp_get_link_encoding_format(&link_settings) == DP_8b_10b_ENCODING)
-		dp_set_fec_enable(link, fec_enable);
-#else
-	dp_set_fec_enable(link, fec_enable);
-#endif
-
-	// during mode set we do DP_SET_POWER off then on, aux writes are lost
-	if (link->dpcd_sink_ext_caps.bits.oled == 1 ||
-		link->dpcd_sink_ext_caps.bits.sdr_aux_backlight_control == 1 ||
-		link->dpcd_sink_ext_caps.bits.hdr_aux_backlight_control == 1) {
-		dc_link_set_default_brightness_aux(link); // TODO: use cached if known
-		if (link->dpcd_sink_ext_caps.bits.oled == 1)
-			msleep(bl_oled_enable_delay);
-		dc_link_backlight_enable_aux(link, true);
-	}
-
-	return status;
-}
-
-static enum dc_status enable_link_edp(
-		struct dc_state *state,
-		struct pipe_ctx *pipe_ctx)
-{
-	enum dc_status status;
-
-	status = enable_link_dp(state, pipe_ctx);
-
-	return status;
-}
-
-static enum dc_status enable_link_dp_mst(
-		struct dc_state *state,
-		struct pipe_ctx *pipe_ctx)
-{
-	struct dc_link *link = pipe_ctx->stream->link;
-
-	/* sink signal type after MST branch is MST. Multiple MST sinks
-	 * share one link. Link DP PHY is enable or training only once.
-	 */
-	if (link->link_status.link_active)
-		return DC_OK;
-
-	/* clear payload table */
-	dm_helpers_dp_mst_clear_payload_allocation_table(link->ctx, link);
-
-	/* to make sure the pending down rep can be processed
-	 * before enabling the link
-	 */
-	dm_helpers_dp_mst_poll_pending_down_reply(link->ctx, link);
-
-	/* set the sink to MST mode before enabling the link */
-	dp_enable_mst_on_sink(link, true);
-
-	return enable_link_dp(state, pipe_ctx);
-}
-
-static bool get_ext_hdmi_settings(struct pipe_ctx *pipe_ctx,
-		enum engine_id eng_id,
-		struct ext_hdmi_settings *settings)
-{
-	bool result = false;
-	int i = 0;
-	struct integrated_info *integrated_info =
-			pipe_ctx->stream->ctx->dc_bios->integrated_info;
-
-	if (integrated_info == NULL)
-		return false;
-
-	/*
-	 * Get retimer settings from sbios for passing SI eye test for DCE11
-	 * The setting values are varied based on board revision and port id
-	 * Therefore the setting values of each ports is passed by sbios.
-	 */
-
-	// Check if current bios contains ext Hdmi settings
-	if (integrated_info->gpu_cap_info & 0x20) {
-		switch (eng_id) {
-		case ENGINE_ID_DIGA:
-			settings->slv_addr = integrated_info->dp0_ext_hdmi_slv_addr;
-			settings->reg_num = integrated_info->dp0_ext_hdmi_6g_reg_num;
-			settings->reg_num_6g = integrated_info->dp0_ext_hdmi_6g_reg_num;
-			memmove(settings->reg_settings,
-					integrated_info->dp0_ext_hdmi_reg_settings,
-					sizeof(integrated_info->dp0_ext_hdmi_reg_settings));
-			memmove(settings->reg_settings_6g,
-					integrated_info->dp0_ext_hdmi_6g_reg_settings,
-					sizeof(integrated_info->dp0_ext_hdmi_6g_reg_settings));
-			result = true;
-			break;
-		case ENGINE_ID_DIGB:
-			settings->slv_addr = integrated_info->dp1_ext_hdmi_slv_addr;
-			settings->reg_num = integrated_info->dp1_ext_hdmi_6g_reg_num;
-			settings->reg_num_6g = integrated_info->dp1_ext_hdmi_6g_reg_num;
-			memmove(settings->reg_settings,
-					integrated_info->dp1_ext_hdmi_reg_settings,
-					sizeof(integrated_info->dp1_ext_hdmi_reg_settings));
-			memmove(settings->reg_settings_6g,
-					integrated_info->dp1_ext_hdmi_6g_reg_settings,
-					sizeof(integrated_info->dp1_ext_hdmi_6g_reg_settings));
-			result = true;
-			break;
-		case ENGINE_ID_DIGC:
-			settings->slv_addr = integrated_info->dp2_ext_hdmi_slv_addr;
-			settings->reg_num = integrated_info->dp2_ext_hdmi_6g_reg_num;
-			settings->reg_num_6g = integrated_info->dp2_ext_hdmi_6g_reg_num;
-			memmove(settings->reg_settings,
-					integrated_info->dp2_ext_hdmi_reg_settings,
-					sizeof(integrated_info->dp2_ext_hdmi_reg_settings));
-			memmove(settings->reg_settings_6g,
-					integrated_info->dp2_ext_hdmi_6g_reg_settings,
-					sizeof(integrated_info->dp2_ext_hdmi_6g_reg_settings));
-			result = true;
-			break;
-		case ENGINE_ID_DIGD:
-			settings->slv_addr = integrated_info->dp3_ext_hdmi_slv_addr;
-			settings->reg_num = integrated_info->dp3_ext_hdmi_6g_reg_num;
-			settings->reg_num_6g = integrated_info->dp3_ext_hdmi_6g_reg_num;
-			memmove(settings->reg_settings,
-					integrated_info->dp3_ext_hdmi_reg_settings,
-					sizeof(integrated_info->dp3_ext_hdmi_reg_settings));
-			memmove(settings->reg_settings_6g,
-					integrated_info->dp3_ext_hdmi_6g_reg_settings,
-					sizeof(integrated_info->dp3_ext_hdmi_6g_reg_settings));
-			result = true;
-			break;
-		default:
-			break;
-		}
-
-		if (result == true) {
-			// Validate settings from bios integrated info table
-			if (settings->slv_addr == 0)
-				return false;
-			if (settings->reg_num > 9)
-				return false;
-			if (settings->reg_num_6g > 3)
-				return false;
-
-			for (i = 0; i < settings->reg_num; i++) {
-				if (settings->reg_settings[i].i2c_reg_index > 0x20)
-					return false;
-			}
-
-			for (i = 0; i < settings->reg_num_6g; i++) {
-				if (settings->reg_settings_6g[i].i2c_reg_index > 0x20)
-					return false;
-			}
-		}
-	}
-
-	return result;
-}
-
-static bool i2c_write(struct pipe_ctx *pipe_ctx,
-		uint8_t address, uint8_t *buffer, uint32_t length)
-{
-	struct i2c_command cmd = {0};
-	struct i2c_payload payload = {0};
-
-	memset(&payload, 0, sizeof(payload));
-	memset(&cmd, 0, sizeof(cmd));
-
-	cmd.number_of_payloads = 1;
-	cmd.engine = I2C_COMMAND_ENGINE_DEFAULT;
-	cmd.speed = pipe_ctx->stream->ctx->dc->caps.i2c_speed_in_khz;
-
-	payload.address = address;
-	payload.data = buffer;
-	payload.length = length;
-	payload.write = true;
-	cmd.payloads = &payload;
-
-	if (dm_helpers_submit_i2c(pipe_ctx->stream->ctx,
-			pipe_ctx->stream->link, &cmd))
-		return true;
-
-	return false;
-}
-
-static void write_i2c_retimer_setting(
-		struct pipe_ctx *pipe_ctx,
-		bool is_vga_mode,
-		bool is_over_340mhz,
-		struct ext_hdmi_settings *settings)
-{
-	uint8_t slave_address = (settings->slv_addr >> 1);
-	uint8_t buffer[2];
-	const uint8_t apply_rx_tx_change = 0x4;
-	uint8_t offset = 0xA;
-	uint8_t value = 0;
-	int i = 0;
-	bool i2c_success = false;
-	DC_LOGGER_INIT(pipe_ctx->stream->ctx->logger);
-
-	memset(&buffer, 0, sizeof(buffer));
-
-	/* Start Ext-Hdmi programming*/
-
-	for (i = 0; i < settings->reg_num; i++) {
-		/* Apply 3G settings */
-		if (settings->reg_settings[i].i2c_reg_index <= 0x20) {
-
-			buffer[0] = settings->reg_settings[i].i2c_reg_index;
-			buffer[1] = settings->reg_settings[i].i2c_reg_val;
-			i2c_success = i2c_write(pipe_ctx, slave_address,
-						buffer, sizeof(buffer));
-			RETIMER_REDRIVER_INFO("retimer write to slave_address = 0x%x,\
-				offset = 0x%x, reg_val= 0x%x, i2c_success = %d\n",
-				slave_address, buffer[0], buffer[1], i2c_success?1:0);
-
-			if (!i2c_success)
-				goto i2c_write_fail;
-
-			/* Based on DP159 specs, APPLY_RX_TX_CHANGE bit in 0x0A
-			 * needs to be set to 1 on every 0xA-0xC write.
-			 */
-			if (settings->reg_settings[i].i2c_reg_index == 0xA ||
-				settings->reg_settings[i].i2c_reg_index == 0xB ||
-				settings->reg_settings[i].i2c_reg_index == 0xC) {
-
-				/* Query current value from offset 0xA */
-				if (settings->reg_settings[i].i2c_reg_index == 0xA)
-					value = settings->reg_settings[i].i2c_reg_val;
-				else {
-					i2c_success =
-						dal_ddc_service_query_ddc_data(
-						pipe_ctx->stream->link->ddc,
-						slave_address, &offset, 1, &value, 1);
-					if (!i2c_success)
-						goto i2c_write_fail;
-				}
-
-				buffer[0] = offset;
-				/* Set APPLY_RX_TX_CHANGE bit to 1 */
-				buffer[1] = value | apply_rx_tx_change;
-				i2c_success = i2c_write(pipe_ctx, slave_address,
-						buffer, sizeof(buffer));
-				RETIMER_REDRIVER_INFO("retimer write to slave_address = 0x%x,\
-					offset = 0x%x, reg_val = 0x%x, i2c_success = %d\n",
-					slave_address, buffer[0], buffer[1], i2c_success?1:0);
-				if (!i2c_success)
-					goto i2c_write_fail;
-			}
-		}
-	}
-
-	/* Apply 3G settings */
-	if (is_over_340mhz) {
-		for (i = 0; i < settings->reg_num_6g; i++) {
-			/* Apply 3G settings */
-			if (settings->reg_settings[i].i2c_reg_index <= 0x20) {
-
-				buffer[0] = settings->reg_settings_6g[i].i2c_reg_index;
-				buffer[1] = settings->reg_settings_6g[i].i2c_reg_val;
-				i2c_success = i2c_write(pipe_ctx, slave_address,
-							buffer, sizeof(buffer));
-				RETIMER_REDRIVER_INFO("above 340Mhz: retimer write to slave_address = 0x%x,\
-					offset = 0x%x, reg_val = 0x%x, i2c_success = %d\n",
-					slave_address, buffer[0], buffer[1], i2c_success?1:0);
-
-				if (!i2c_success)
-					goto i2c_write_fail;
-
-				/* Based on DP159 specs, APPLY_RX_TX_CHANGE bit in 0x0A
-				 * needs to be set to 1 on every 0xA-0xC write.
-				 */
-				if (settings->reg_settings_6g[i].i2c_reg_index == 0xA ||
-					settings->reg_settings_6g[i].i2c_reg_index == 0xB ||
-					settings->reg_settings_6g[i].i2c_reg_index == 0xC) {
-
-					/* Query current value from offset 0xA */
-					if (settings->reg_settings_6g[i].i2c_reg_index == 0xA)
-						value = settings->reg_settings_6g[i].i2c_reg_val;
-					else {
-						i2c_success =
-								dal_ddc_service_query_ddc_data(
-								pipe_ctx->stream->link->ddc,
-								slave_address, &offset, 1, &value, 1);
-						if (!i2c_success)
-							goto i2c_write_fail;
-					}
-
-					buffer[0] = offset;
-					/* Set APPLY_RX_TX_CHANGE bit to 1 */
-					buffer[1] = value | apply_rx_tx_change;
-					i2c_success = i2c_write(pipe_ctx, slave_address,
-							buffer, sizeof(buffer));
-					RETIMER_REDRIVER_INFO("retimer write to slave_address = 0x%x,\
-						offset = 0x%x, reg_val = 0x%x, i2c_success = %d\n",
-						slave_address, buffer[0], buffer[1], i2c_success?1:0);
-					if (!i2c_success)
-						goto i2c_write_fail;
-				}
-			}
-		}
-	}
-
-	if (is_vga_mode) {
-		/* Program additional settings if using 640x480 resolution */
-
-		/* Write offset 0xFF to 0x01 */
-		buffer[0] = 0xff;
-		buffer[1] = 0x01;
-		i2c_success = i2c_write(pipe_ctx, slave_address,
-				buffer, sizeof(buffer));
-		RETIMER_REDRIVER_INFO("retimer write to slave_address = 0x%x,\
-				offset = 0x%x, reg_val = 0x%x, i2c_success = %d\n",
-				slave_address, buffer[0], buffer[1], i2c_success?1:0);
-		if (!i2c_success)
-			goto i2c_write_fail;
-
-		/* Write offset 0x00 to 0x23 */
-		buffer[0] = 0x00;
-		buffer[1] = 0x23;
-		i2c_success = i2c_write(pipe_ctx, slave_address,
-				buffer, sizeof(buffer));
-		RETIMER_REDRIVER_INFO("retimer write to slave_address = 0x%x,\
-			offset = 0x%x, reg_val = 0x%x, i2c_success = %d\n",
-			slave_address, buffer[0], buffer[1], i2c_success?1:0);
-		if (!i2c_success)
-			goto i2c_write_fail;
-
-		/* Write offset 0xff to 0x00 */
-		buffer[0] = 0xff;
-		buffer[1] = 0x00;
-		i2c_success = i2c_write(pipe_ctx, slave_address,
-				buffer, sizeof(buffer));
-		RETIMER_REDRIVER_INFO("retimer write to slave_address = 0x%x,\
-			offset = 0x%x, reg_val = 0x%x, i2c_success = %d\n",
-			slave_address, buffer[0], buffer[1], i2c_success?1:0);
-		if (!i2c_success)
-			goto i2c_write_fail;
-
-	}
-
-	return;
-
-i2c_write_fail:
-	DC_LOG_DEBUG("Set retimer failed");
-}
-
-static void write_i2c_default_retimer_setting(
-		struct pipe_ctx *pipe_ctx,
-		bool is_vga_mode,
-		bool is_over_340mhz)
-{
-	uint8_t slave_address = (0xBA >> 1);
-	uint8_t buffer[2];
-	bool i2c_success = false;
-	DC_LOGGER_INIT(pipe_ctx->stream->ctx->logger);
-
-	memset(&buffer, 0, sizeof(buffer));
-
-	/* Program Slave Address for tuning single integrity */
-	/* Write offset 0x0A to 0x13 */
-	buffer[0] = 0x0A;
-	buffer[1] = 0x13;
-	i2c_success = i2c_write(pipe_ctx, slave_address,
-			buffer, sizeof(buffer));
-	RETIMER_REDRIVER_INFO("retimer writes default setting to slave_address = 0x%x,\
-		offset = 0x%x, reg_val = 0x%x, i2c_success = %d\n",
-		slave_address, buffer[0], buffer[1], i2c_success?1:0);
-	if (!i2c_success)
-		goto i2c_write_fail;
-
-	/* Write offset 0x0A to 0x17 */
-	buffer[0] = 0x0A;
-	buffer[1] = 0x17;
-	i2c_success = i2c_write(pipe_ctx, slave_address,
-			buffer, sizeof(buffer));
-	RETIMER_REDRIVER_INFO("retimer write to slave_addr = 0x%x,\
-		offset = 0x%x, reg_val = 0x%x, i2c_success = %d\n",
-		slave_address, buffer[0], buffer[1], i2c_success?1:0);
-	if (!i2c_success)
-		goto i2c_write_fail;
-
-	/* Write offset 0x0B to 0xDA or 0xD8 */
-	buffer[0] = 0x0B;
-	buffer[1] = is_over_340mhz ? 0xDA : 0xD8;
-	i2c_success = i2c_write(pipe_ctx, slave_address,
-			buffer, sizeof(buffer));
-	RETIMER_REDRIVER_INFO("retimer write to slave_addr = 0x%x,\
-		offset = 0x%x, reg_val = 0x%x, i2c_success = %d\n",
-		slave_address, buffer[0], buffer[1], i2c_success?1:0);
-	if (!i2c_success)
-		goto i2c_write_fail;
-
-	/* Write offset 0x0A to 0x17 */
-	buffer[0] = 0x0A;
-	buffer[1] = 0x17;
-	i2c_success = i2c_write(pipe_ctx, slave_address,
-			buffer, sizeof(buffer));
-	RETIMER_REDRIVER_INFO("retimer write to slave_addr = 0x%x,\
-		offset = 0x%x, reg_val= 0x%x, i2c_success = %d\n",
-		slave_address, buffer[0], buffer[1], i2c_success?1:0);
-	if (!i2c_success)
-		goto i2c_write_fail;
-
-	/* Write offset 0x0C to 0x1D or 0x91 */
-	buffer[0] = 0x0C;
-	buffer[1] = is_over_340mhz ? 0x1D : 0x91;
-	i2c_success = i2c_write(pipe_ctx, slave_address,
-			buffer, sizeof(buffer));
-	RETIMER_REDRIVER_INFO("retimer write to slave_addr = 0x%x,\
-		offset = 0x%x, reg_val = 0x%x, i2c_success = %d\n",
-		slave_address, buffer[0], buffer[1], i2c_success?1:0);
-	if (!i2c_success)
-		goto i2c_write_fail;
-
-	/* Write offset 0x0A to 0x17 */
-	buffer[0] = 0x0A;
-	buffer[1] = 0x17;
-	i2c_success = i2c_write(pipe_ctx, slave_address,
-			buffer, sizeof(buffer));
-	RETIMER_REDRIVER_INFO("retimer write to slave_addr = 0x%x,\
-		offset = 0x%x, reg_val = 0x%x, i2c_success = %d\n",
-		slave_address, buffer[0], buffer[1], i2c_success?1:0);
-	if (!i2c_success)
-		goto i2c_write_fail;
-
-
-	if (is_vga_mode) {
-		/* Program additional settings if using 640x480 resolution */
-
-		/* Write offset 0xFF to 0x01 */
-		buffer[0] = 0xff;
-		buffer[1] = 0x01;
-		i2c_success = i2c_write(pipe_ctx, slave_address,
-				buffer, sizeof(buffer));
-		RETIMER_REDRIVER_INFO("retimer write to slave_addr = 0x%x,\
-			offset = 0x%x, reg_val = 0x%x, i2c_success = %d\n",
-			slave_address, buffer[0], buffer[1], i2c_success?1:0);
-		if (!i2c_success)
-			goto i2c_write_fail;
-
-		/* Write offset 0x00 to 0x23 */
-		buffer[0] = 0x00;
-		buffer[1] = 0x23;
-		i2c_success = i2c_write(pipe_ctx, slave_address,
-				buffer, sizeof(buffer));
-		RETIMER_REDRIVER_INFO("retimer write to slave_addr = 0x%x,\
-			offset = 0x%x, reg_val= 0x%x, i2c_success = %d\n",
-			slave_address, buffer[0], buffer[1], i2c_success?1:0);
-		if (!i2c_success)
-			goto i2c_write_fail;
-
-		/* Write offset 0xff to 0x00 */
-		buffer[0] = 0xff;
-		buffer[1] = 0x00;
-		i2c_success = i2c_write(pipe_ctx, slave_address,
-				buffer, sizeof(buffer));
-		RETIMER_REDRIVER_INFO("retimer write default setting to slave_addr = 0x%x,\
-			offset = 0x%x, reg_val= 0x%x, i2c_success = %d end here\n",
-			slave_address, buffer[0], buffer[1], i2c_success?1:0);
-		if (!i2c_success)
-			goto i2c_write_fail;
-	}
-
-	return;
-
-i2c_write_fail:
-	DC_LOG_DEBUG("Set default retimer failed");
-}
-
-static void write_i2c_redriver_setting(
-		struct pipe_ctx *pipe_ctx,
-		bool is_over_340mhz)
-{
-	uint8_t slave_address = (0xF0 >> 1);
-	uint8_t buffer[16];
-	bool i2c_success = false;
-	DC_LOGGER_INIT(pipe_ctx->stream->ctx->logger);
-
-	memset(&buffer, 0, sizeof(buffer));
-
-	// Program Slave Address for tuning single integrity
-	buffer[3] = 0x4E;
-	buffer[4] = 0x4E;
-	buffer[5] = 0x4E;
-	buffer[6] = is_over_340mhz ? 0x4E : 0x4A;
-
-	i2c_success = i2c_write(pipe_ctx, slave_address,
-					buffer, sizeof(buffer));
-	RETIMER_REDRIVER_INFO("redriver write 0 to all 16 reg offset expect following:\n\
-		\t slave_addr = 0x%x, offset[3] = 0x%x, offset[4] = 0x%x,\
-		offset[5] = 0x%x,offset[6] is_over_340mhz = 0x%x,\
-		i2c_success = %d\n",
-		slave_address, buffer[3], buffer[4], buffer[5], buffer[6], i2c_success?1:0);
-
-	if (!i2c_success)
-		DC_LOG_DEBUG("Set redriver failed");
-}
-
-static void disable_link(struct dc_link *link, enum signal_type signal)
-{
-	/*
-	 * TODO: implement call for dp_set_hw_test_pattern
-	 * it is needed for compliance testing
-	 */
-
-	/* Here we need to specify that encoder output settings
-	 * need to be calculated as for the set mode,
-	 * it will lead to querying dynamic link capabilities
-	 * which should be done before enable output
-	 */
-
-	if (dc_is_dp_signal(signal)) {
-		/* SST DP, eDP */
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-		struct dc_link_settings link_settings = link->cur_link_settings;
-#endif
-		if (dc_is_dp_sst_signal(signal))
-			dp_disable_link_phy(link, signal);
-		else
-			dp_disable_link_phy_mst(link, signal);
-
-		if (dc_is_dp_sst_signal(signal) ||
-				link->mst_stream_alloc_table.stream_count == 0) {
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-			if (dp_get_link_encoding_format(&link_settings) == DP_8b_10b_ENCODING) {
-				dp_set_fec_enable(link, false);
-				dp_set_fec_ready(link, false);
-			}
-#else
-			dp_set_fec_enable(link, false);
-			dp_set_fec_ready(link, false);
-#endif
-		}
-	} else {
-		if (signal != SIGNAL_TYPE_VIRTUAL)
-			link->link_enc->funcs->disable_output(link->link_enc, signal);
-	}
-
-	if (signal == SIGNAL_TYPE_DISPLAY_PORT_MST) {
-		/* MST disable link only when no stream use the link */
-		if (link->mst_stream_alloc_table.stream_count <= 0)
-			link->link_status.link_active = false;
-	} else {
-		link->link_status.link_active = false;
-	}
-}
-
-static void enable_link_hdmi(struct pipe_ctx *pipe_ctx)
-{
-	struct dc_stream_state *stream = pipe_ctx->stream;
-	struct dc_link *link = stream->link;
-	enum dc_color_depth display_color_depth;
-	enum engine_id eng_id;
-	struct ext_hdmi_settings settings = {0};
-	bool is_over_340mhz = false;
-	bool is_vga_mode = (stream->timing.h_addressable == 640)
-			&& (stream->timing.v_addressable == 480);
-
-	if (stream->phy_pix_clk == 0)
-		stream->phy_pix_clk = stream->timing.pix_clk_100hz / 10;
-	if (stream->phy_pix_clk > 340000)
-		is_over_340mhz = true;
-
-	if (dc_is_hdmi_signal(pipe_ctx->stream->signal)) {
-		unsigned short masked_chip_caps = pipe_ctx->stream->link->chip_caps &
-				EXT_DISPLAY_PATH_CAPS__EXT_CHIP_MASK;
-		if (masked_chip_caps == EXT_DISPLAY_PATH_CAPS__HDMI20_TISN65DP159RSBT) {
-			/* DP159, Retimer settings */
-			eng_id = pipe_ctx->stream_res.stream_enc->id;
-
-			if (get_ext_hdmi_settings(pipe_ctx, eng_id, &settings)) {
-				write_i2c_retimer_setting(pipe_ctx,
-						is_vga_mode, is_over_340mhz, &settings);
-			} else {
-				write_i2c_default_retimer_setting(pipe_ctx,
-						is_vga_mode, is_over_340mhz);
-			}
-		} else if (masked_chip_caps == EXT_DISPLAY_PATH_CAPS__HDMI20_PI3EQX1204) {
-			/* PI3EQX1204, Redriver settings */
-			write_i2c_redriver_setting(pipe_ctx, is_over_340mhz);
-		}
-	}
-
-	if (dc_is_hdmi_signal(pipe_ctx->stream->signal))
-		dal_ddc_service_write_scdc_data(
-			stream->link->ddc,
-			stream->phy_pix_clk,
-			stream->timing.flags.LTE_340MCSC_SCRAMBLE);
-
-	memset(&stream->link->cur_link_settings, 0,
-			sizeof(struct dc_link_settings));
-
-	display_color_depth = stream->timing.display_color_depth;
-	if (stream->timing.pixel_encoding == PIXEL_ENCODING_YCBCR422)
-		display_color_depth = COLOR_DEPTH_888;
-
-	link->link_enc->funcs->enable_tmds_output(
-			link->link_enc,
-			pipe_ctx->clock_source->id,
-			display_color_depth,
-			pipe_ctx->stream->signal,
-			stream->phy_pix_clk);
-
-	if (dc_is_hdmi_signal(pipe_ctx->stream->signal))
-		dal_ddc_service_read_scdc_data(link->ddc);
-}
-
-static void enable_link_lvds(struct pipe_ctx *pipe_ctx)
-{
-	struct dc_stream_state *stream = pipe_ctx->stream;
-	struct dc_link *link = stream->link;
-
-	if (stream->phy_pix_clk == 0)
-		stream->phy_pix_clk = stream->timing.pix_clk_100hz / 10;
-
-	memset(&stream->link->cur_link_settings, 0,
-			sizeof(struct dc_link_settings));
-
-	link->link_enc->funcs->enable_lvds_output(
-			link->link_enc,
-			pipe_ctx->clock_source->id,
-			stream->phy_pix_clk);
-
-}
-
-/****************************enable_link***********************************/
-static enum dc_status enable_link(
-		struct dc_state *state,
-		struct pipe_ctx *pipe_ctx)
-{
-	enum dc_status status = DC_ERROR_UNEXPECTED;
-	struct dc_stream_state *stream = pipe_ctx->stream;
-	struct dc_link *link = stream->link;
-
-	/* There's some scenarios where driver is unloaded with display
-	 * still enabled. When driver is reloaded, it may cause a display
-	 * to not light up if there is a mismatch between old and new
-	 * link settings. Need to call disable first before enabling at
-	 * new link settings.
-	 */
-	if (link->link_status.link_active) {
-		disable_link(link, pipe_ctx->stream->signal);
-	}
-
-	switch (pipe_ctx->stream->signal) {
-	case SIGNAL_TYPE_DISPLAY_PORT:
-		status = enable_link_dp(state, pipe_ctx);
-		break;
-	case SIGNAL_TYPE_EDP:
-		status = enable_link_edp(state, pipe_ctx);
-		break;
-	case SIGNAL_TYPE_DISPLAY_PORT_MST:
-		status = enable_link_dp_mst(state, pipe_ctx);
-		msleep(200);
-		break;
-	case SIGNAL_TYPE_DVI_SINGLE_LINK:
-	case SIGNAL_TYPE_DVI_DUAL_LINK:
-	case SIGNAL_TYPE_HDMI_TYPE_A:
-		enable_link_hdmi(pipe_ctx);
-		status = DC_OK;
-		break;
-	case SIGNAL_TYPE_LVDS:
-		enable_link_lvds(pipe_ctx);
-		status = DC_OK;
-		break;
-	case SIGNAL_TYPE_VIRTUAL:
-		status = DC_OK;
-		break;
-	default:
-		break;
-	}
-
-	if (status == DC_OK)
-		pipe_ctx->stream->link->link_status.link_active = true;
-
-	return status;
-}
-
-static uint32_t get_timing_pixel_clock_100hz(const struct dc_crtc_timing *timing)
-{
-
-	uint32_t pxl_clk = timing->pix_clk_100hz;
-
-	if (timing->pixel_encoding == PIXEL_ENCODING_YCBCR420)
-		pxl_clk /= 2;
-	else if (timing->pixel_encoding == PIXEL_ENCODING_YCBCR422)
-		pxl_clk = pxl_clk * 2 / 3;
-
-	if (timing->display_color_depth == COLOR_DEPTH_101010)
-		pxl_clk = pxl_clk * 10 / 8;
-	else if (timing->display_color_depth == COLOR_DEPTH_121212)
-		pxl_clk = pxl_clk * 12 / 8;
-
-	return pxl_clk;
-}
-
-static bool dp_active_dongle_validate_timing(
-		const struct dc_crtc_timing *timing,
-		const struct dpcd_caps *dpcd_caps)
-{
-	const struct dc_dongle_caps *dongle_caps = &dpcd_caps->dongle_caps;
-
-	switch (dpcd_caps->dongle_type) {
-	case DISPLAY_DONGLE_DP_VGA_CONVERTER:
-	case DISPLAY_DONGLE_DP_DVI_CONVERTER:
-	case DISPLAY_DONGLE_DP_DVI_DONGLE:
-		if (timing->pixel_encoding == PIXEL_ENCODING_RGB)
-			return true;
-		else
-			return false;
-	default:
-		break;
-	}
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	if (dpcd_caps->dongle_type == DISPLAY_DONGLE_DP_HDMI_CONVERTER &&
-			dongle_caps->extendedCapValid == true) {
-#else
-	if (dpcd_caps->dongle_type != DISPLAY_DONGLE_DP_HDMI_CONVERTER ||
-		dongle_caps->extendedCapValid == false)
-		return true;
-#endif
-
-	/* Check Pixel Encoding */
-	switch (timing->pixel_encoding) {
-	case PIXEL_ENCODING_RGB:
-	case PIXEL_ENCODING_YCBCR444:
-		break;
-	case PIXEL_ENCODING_YCBCR422:
-		if (!dongle_caps->is_dp_hdmi_ycbcr422_pass_through)
-			return false;
-		break;
-	case PIXEL_ENCODING_YCBCR420:
-		if (!dongle_caps->is_dp_hdmi_ycbcr420_pass_through)
-			return false;
-		break;
-	default:
-		/* Invalid Pixel Encoding*/
-		return false;
-	}
-
-	switch (timing->display_color_depth) {
-	case COLOR_DEPTH_666:
-	case COLOR_DEPTH_888:
-		/*888 and 666 should always be supported*/
-		break;
-	case COLOR_DEPTH_101010:
-		if (dongle_caps->dp_hdmi_max_bpc < 10)
-			return false;
-		break;
-	case COLOR_DEPTH_121212:
-		if (dongle_caps->dp_hdmi_max_bpc < 12)
-			return false;
-		break;
-	case COLOR_DEPTH_141414:
-	case COLOR_DEPTH_161616:
-	default:
-		/* These color depths are currently not supported */
-		return false;
-	}
-
-	if (get_timing_pixel_clock_100hz(timing) > (dongle_caps->dp_hdmi_max_pixel_clk_in_khz * 10))
-		return false;
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	}
-
-	if (dpcd_caps->channel_coding_cap.bits.DP_128b_132b_SUPPORTED == 0 &&
-			dpcd_caps->dsc_caps.dsc_basic_caps.fields.dsc_support.DSC_PASSTHROUGH_SUPPORT == 0 &&
-			dongle_caps->dfp_cap_ext.supported) {
-
-		if (dongle_caps->dfp_cap_ext.max_pixel_rate_in_mps < (timing->pix_clk_100hz / 10000))
-			return false;
-
-		if (dongle_caps->dfp_cap_ext.max_video_h_active_width < timing->h_addressable)
-			return false;
-
-		if (dongle_caps->dfp_cap_ext.max_video_v_active_height < timing->v_addressable)
-			return false;
-
-		if (timing->pixel_encoding == PIXEL_ENCODING_RGB) {
-			if (!dongle_caps->dfp_cap_ext.encoding_format_caps.support_rgb)
-				return false;
-			if (timing->display_color_depth == COLOR_DEPTH_666 &&
-					!dongle_caps->dfp_cap_ext.rgb_color_depth_caps.support_6bpc)
-				return false;
-			else if (timing->display_color_depth == COLOR_DEPTH_888 &&
-					!dongle_caps->dfp_cap_ext.rgb_color_depth_caps.support_8bpc)
-				return false;
-			else if (timing->display_color_depth == COLOR_DEPTH_101010 &&
-					!dongle_caps->dfp_cap_ext.rgb_color_depth_caps.support_10bpc)
-				return false;
-			else if (timing->display_color_depth == COLOR_DEPTH_121212 &&
-					!dongle_caps->dfp_cap_ext.rgb_color_depth_caps.support_12bpc)
-				return false;
-			else if (timing->display_color_depth == COLOR_DEPTH_161616 &&
-					!dongle_caps->dfp_cap_ext.rgb_color_depth_caps.support_16bpc)
-				return false;
-		} else if (timing->pixel_encoding == PIXEL_ENCODING_YCBCR444) {
-			if (!dongle_caps->dfp_cap_ext.encoding_format_caps.support_rgb)
-				return false;
-			if (timing->display_color_depth == COLOR_DEPTH_888 &&
-					!dongle_caps->dfp_cap_ext.ycbcr444_color_depth_caps.support_8bpc)
-				return false;
-			else if (timing->display_color_depth == COLOR_DEPTH_101010 &&
-					!dongle_caps->dfp_cap_ext.ycbcr444_color_depth_caps.support_10bpc)
-				return false;
-			else if (timing->display_color_depth == COLOR_DEPTH_121212 &&
-					!dongle_caps->dfp_cap_ext.ycbcr444_color_depth_caps.support_12bpc)
-				return false;
-			else if (timing->display_color_depth == COLOR_DEPTH_161616 &&
-					!dongle_caps->dfp_cap_ext.ycbcr444_color_depth_caps.support_16bpc)
-				return false;
-		} else if (timing->pixel_encoding == PIXEL_ENCODING_YCBCR422) {
-			if (!dongle_caps->dfp_cap_ext.encoding_format_caps.support_rgb)
-				return false;
-			if (timing->display_color_depth == COLOR_DEPTH_888 &&
-					!dongle_caps->dfp_cap_ext.ycbcr422_color_depth_caps.support_8bpc)
-				return false;
-			else if (timing->display_color_depth == COLOR_DEPTH_101010 &&
-					!dongle_caps->dfp_cap_ext.ycbcr422_color_depth_caps.support_10bpc)
-				return false;
-			else if (timing->display_color_depth == COLOR_DEPTH_121212 &&
-					!dongle_caps->dfp_cap_ext.ycbcr422_color_depth_caps.support_12bpc)
-				return false;
-			else if (timing->display_color_depth == COLOR_DEPTH_161616 &&
-					!dongle_caps->dfp_cap_ext.ycbcr422_color_depth_caps.support_16bpc)
-				return false;
-		} else if (timing->pixel_encoding == PIXEL_ENCODING_YCBCR420) {
-			if (!dongle_caps->dfp_cap_ext.encoding_format_caps.support_rgb)
-				return false;
-			if (timing->display_color_depth == COLOR_DEPTH_888 &&
-					!dongle_caps->dfp_cap_ext.ycbcr420_color_depth_caps.support_8bpc)
-				return false;
-			else if (timing->display_color_depth == COLOR_DEPTH_101010 &&
-					!dongle_caps->dfp_cap_ext.ycbcr420_color_depth_caps.support_10bpc)
-				return false;
-			else if (timing->display_color_depth == COLOR_DEPTH_121212 &&
-					!dongle_caps->dfp_cap_ext.ycbcr420_color_depth_caps.support_12bpc)
-				return false;
-			else if (timing->display_color_depth == COLOR_DEPTH_161616 &&
-					!dongle_caps->dfp_cap_ext.ycbcr420_color_depth_caps.support_16bpc)
-				return false;
-		}
-	}
-#endif
-
-	return true;
-}
-
-enum dc_status dc_link_validate_mode_timing(
-		const struct dc_stream_state *stream,
-		struct dc_link *link,
-		const struct dc_crtc_timing *timing)
-{
-	uint32_t max_pix_clk = stream->link->dongle_max_pix_clk * 10;
-	struct dpcd_caps *dpcd_caps = &link->dpcd_caps;
-
-	/* A hack to avoid failing any modes for EDID override feature on
-	 * topology change such as lower quality cable for DP or different dongle
-	 */
-	if (link->remote_sinks[0] && link->remote_sinks[0]->sink_signal == SIGNAL_TYPE_VIRTUAL)
-		return DC_OK;
-
-	/* Passive Dongle */
-	if (max_pix_clk != 0 && get_timing_pixel_clock_100hz(timing) > max_pix_clk)
-		return DC_EXCEED_DONGLE_CAP;
-
-	/* Active Dongle*/
-	if (!dp_active_dongle_validate_timing(timing, dpcd_caps))
-		return DC_EXCEED_DONGLE_CAP;
-
-	switch (stream->signal) {
-	case SIGNAL_TYPE_EDP:
-	case SIGNAL_TYPE_DISPLAY_PORT:
-		if (!dp_validate_mode_timing(
-				link,
-				timing))
-			return DC_NO_DP_LINK_BANDWIDTH;
-		break;
-
-	default:
-		break;
-	}
-
-	return DC_OK;
-}
-
-static struct abm *get_abm_from_stream_res(const struct dc_link *link)
-{
-	int i;
-	struct dc *dc = NULL;
-	struct abm *abm = NULL;
-
-	if (!link || !link->ctx)
-		return NULL;
-
-	dc = link->ctx->dc;
-
-	for (i = 0; i < MAX_PIPES; i++) {
-		struct pipe_ctx pipe_ctx = dc->current_state->res_ctx.pipe_ctx[i];
-		struct dc_stream_state *stream = pipe_ctx.stream;
-
-		if (stream && stream->link == link) {
-			abm = pipe_ctx.stream_res.abm;
-			break;
-		}
-	}
-	return abm;
-}
-
-int dc_link_get_backlight_level(const struct dc_link *link)
-{
-	struct abm *abm = get_abm_from_stream_res(link);
-	struct panel_cntl *panel_cntl = link->panel_cntl;
-	struct dc  *dc = link->ctx->dc;
-	struct dmcu *dmcu = dc->res_pool->dmcu;
-	bool fw_set_brightness = true;
-
-	if (dmcu)
-		fw_set_brightness = dmcu->funcs->is_dmcu_initialized(dmcu);
-
-	if (!fw_set_brightness && panel_cntl->funcs->get_current_backlight)
-		return panel_cntl->funcs->get_current_backlight(panel_cntl);
-	else if (abm != NULL && abm->funcs->get_current_backlight != NULL)
-		return (int) abm->funcs->get_current_backlight(abm);
-	else
-		return DC_ERROR_UNEXPECTED;
-}
-
-int dc_link_get_target_backlight_pwm(const struct dc_link *link)
-{
-	struct abm *abm = get_abm_from_stream_res(link);
-
-	if (abm == NULL || abm->funcs->get_target_backlight == NULL)
-		return DC_ERROR_UNEXPECTED;
-
-	return (int) abm->funcs->get_target_backlight(abm);
-}
-
-static struct pipe_ctx *get_pipe_from_link(const struct dc_link *link)
-{
-	int i;
-	struct dc *dc = link->ctx->dc;
-	struct pipe_ctx *pipe_ctx = NULL;
-
-	for (i = 0; i < MAX_PIPES; i++) {
-		if (dc->current_state->res_ctx.pipe_ctx[i].stream) {
-			if (dc->current_state->res_ctx.pipe_ctx[i].stream->link == link) {
-				pipe_ctx = &dc->current_state->res_ctx.pipe_ctx[i];
-				break;
-			}
-		}
-	}
-
-	return pipe_ctx;
-}
-
-bool dc_link_set_backlight_level(const struct dc_link *link,
-		uint32_t backlight_pwm_u16_16,
-		uint32_t frame_ramp)
-{
-	struct dc  *dc = link->ctx->dc;
-
-	DC_LOGGER_INIT(link->ctx->logger);
-	DC_LOG_BACKLIGHT("New Backlight level: %d (0x%X)\n",
-			backlight_pwm_u16_16, backlight_pwm_u16_16);
-
-	if (dc_is_embedded_signal(link->connector_signal)) {
-		struct pipe_ctx *pipe_ctx = get_pipe_from_link(link);
-
-		if (pipe_ctx) {
-			/* Disable brightness ramping when the display is blanked
-			 * as it can hang the DMCU
-			 */
-			if (pipe_ctx->plane_state == NULL)
-				frame_ramp = 0;
-		} else {
-			return false;
-		}
-
-		dc->hwss.set_backlight_level(
-				pipe_ctx,
-				backlight_pwm_u16_16,
-				frame_ramp);
-	}
-	return true;
-}
-
-bool dc_link_set_psr_allow_active(struct dc_link *link, const bool *allow_active,
-		bool wait, bool force_static, const unsigned int *power_opts)
-{
-	struct dc  *dc = link->ctx->dc;
-	struct dmcu *dmcu = dc->res_pool->dmcu;
-	struct dmub_psr *psr = dc->res_pool->psr;
-	unsigned int panel_inst;
-
-	if (psr == NULL && force_static)
-		return false;
-
-	if (!dc_get_edp_link_panel_inst(dc, link, &panel_inst))
-		return false;
-
-	/* Set power optimization flag */
-	if (power_opts && link->psr_settings.psr_power_opt != *power_opts) {
-		link->psr_settings.psr_power_opt = *power_opts;
-
-		if (psr != NULL && link->psr_settings.psr_feature_enabled && psr->funcs->psr_set_power_opt)
-			psr->funcs->psr_set_power_opt(psr, link->psr_settings.psr_power_opt);
-	}
-
-	/* Enable or Disable PSR */
-	if (allow_active && link->psr_settings.psr_allow_active != *allow_active) {
-		link->psr_settings.psr_allow_active = *allow_active;
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-		if (!link->psr_settings.psr_allow_active)
-			dc_z10_restore(dc);
-#endif
-
-		if (psr != NULL && link->psr_settings.psr_feature_enabled) {
-			if (force_static && psr->funcs->psr_force_static)
-				psr->funcs->psr_force_static(psr, panel_inst);
-			psr->funcs->psr_enable(psr, link->psr_settings.psr_allow_active, wait, panel_inst);
-		} else if ((dmcu != NULL && dmcu->funcs->is_dmcu_initialized(dmcu)) &&
-			link->psr_settings.psr_feature_enabled)
-			dmcu->funcs->set_psr_enable(dmcu, link->psr_settings.psr_allow_active, wait);
-		else
-			return false;
-	}
-
-	return true;
-}
-
-bool dc_link_get_psr_state(const struct dc_link *link, enum dc_psr_state *state)
-{
-	struct dc  *dc = link->ctx->dc;
-	struct dmcu *dmcu = dc->res_pool->dmcu;
-	struct dmub_psr *psr = dc->res_pool->psr;
-	unsigned int panel_inst;
-
-	if (!dc_get_edp_link_panel_inst(dc, link, &panel_inst))
-		return false;
-
-	if (psr != NULL && link->psr_settings.psr_feature_enabled)
-		psr->funcs->psr_get_state(psr, state, panel_inst);
-	else if (dmcu != NULL && link->psr_settings.psr_feature_enabled)
-		dmcu->funcs->get_psr_state(dmcu, state);
-
-	return true;
-}
-
-static inline enum physical_phy_id
-transmitter_to_phy_id(enum transmitter transmitter_value)
-{
-	switch (transmitter_value) {
-	case TRANSMITTER_UNIPHY_A:
-		return PHYLD_0;
-	case TRANSMITTER_UNIPHY_B:
-		return PHYLD_1;
-	case TRANSMITTER_UNIPHY_C:
-		return PHYLD_2;
-	case TRANSMITTER_UNIPHY_D:
-		return PHYLD_3;
-	case TRANSMITTER_UNIPHY_E:
-		return PHYLD_4;
-	case TRANSMITTER_UNIPHY_F:
-		return PHYLD_5;
-	case TRANSMITTER_NUTMEG_CRT:
-		return PHYLD_6;
-	case TRANSMITTER_TRAVIS_CRT:
-		return PHYLD_7;
-	case TRANSMITTER_TRAVIS_LCD:
-		return PHYLD_8;
-	case TRANSMITTER_UNIPHY_G:
-		return PHYLD_9;
-	case TRANSMITTER_COUNT:
-		return PHYLD_COUNT;
-	case TRANSMITTER_UNKNOWN:
-		return PHYLD_UNKNOWN;
-	default:
-		WARN_ONCE(1, "Unknown transmitter value %d\n",
-			  transmitter_value);
-		return PHYLD_UNKNOWN;
-	}
-}
-
-bool dc_link_setup_psr(struct dc_link *link,
-		const struct dc_stream_state *stream, struct psr_config *psr_config,
-		struct psr_context *psr_context)
-{
-	struct dc *dc;
-	struct dmcu *dmcu;
-	struct dmub_psr *psr;
-	int i;
-	unsigned int panel_inst;
-	/* updateSinkPsrDpcdConfig*/
-	union dpcd_psr_configuration psr_configuration;
-
-	psr_context->controllerId = CONTROLLER_ID_UNDEFINED;
-
-	if (!link)
-		return false;
-
-	dc = link->ctx->dc;
-	dmcu = dc->res_pool->dmcu;
-	psr = dc->res_pool->psr;
-
-	if (!dmcu && !psr)
-		return false;
-
-	if (!dc_get_edp_link_panel_inst(dc, link, &panel_inst))
-		return false;
-
-
-	memset(&psr_configuration, 0, sizeof(psr_configuration));
-
-	psr_configuration.bits.ENABLE                    = 1;
-	psr_configuration.bits.CRC_VERIFICATION          = 1;
-	psr_configuration.bits.FRAME_CAPTURE_INDICATION  =
-			psr_config->psr_frame_capture_indication_req;
-
-	/* Check for PSR v2*/
-	if (psr_config->psr_version == 0x2) {
-		/* For PSR v2 selective update.
-		 * Indicates whether sink should start capturing
-		 * immediately following active scan line,
-		 * or starting with the 2nd active scan line.
-		 */
-		psr_configuration.bits.LINE_CAPTURE_INDICATION = 0;
-		/*For PSR v2, determines whether Sink should generate
-		 * IRQ_HPD when CRC mismatch is detected.
-		 */
-		psr_configuration.bits.IRQ_HPD_WITH_CRC_ERROR    = 1;
-	}
-
-	dm_helpers_dp_write_dpcd(
-		link->ctx,
-		link,
-		368,
-		&psr_configuration.raw,
-		sizeof(psr_configuration.raw));
-
-	psr_context->channel = link->ddc->ddc_pin->hw_info.ddc_channel;
-	psr_context->transmitterId = link->link_enc->transmitter;
-	psr_context->engineId = link->link_enc->preferred_engine;
-
-	for (i = 0; i < MAX_PIPES; i++) {
-		if (dc->current_state->res_ctx.pipe_ctx[i].stream
-				== stream) {
-			/* dmcu -1 for all controller id values,
-			 * therefore +1 here
-			 */
-			psr_context->controllerId =
-				dc->current_state->res_ctx.
-				pipe_ctx[i].stream_res.tg->inst + 1;
-			break;
-		}
-	}
-
-	/* Hardcoded for now.  Can be Pcie or Uniphy (or Unknown)*/
-	psr_context->phyType = PHY_TYPE_UNIPHY;
-	/*PhyId is associated with the transmitter id*/
-	psr_context->smuPhyId =
-		transmitter_to_phy_id(link->link_enc->transmitter);
-
-	psr_context->crtcTimingVerticalTotal = stream->timing.v_total;
-	psr_context->vsync_rate_hz = div64_u64(div64_u64((stream->
-					timing.pix_clk_100hz * 100),
-					stream->timing.v_total),
-					stream->timing.h_total);
-
-	psr_context->psrSupportedDisplayConfig = true;
-	psr_context->psrExitLinkTrainingRequired =
-		psr_config->psr_exit_link_training_required;
-	psr_context->sdpTransmitLineNumDeadline =
-		psr_config->psr_sdp_transmit_line_num_deadline;
-	psr_context->psrFrameCaptureIndicationReq =
-		psr_config->psr_frame_capture_indication_req;
-
-	psr_context->skipPsrWaitForPllLock = 0; /* only = 1 in KV */
-
-	psr_context->numberOfControllers =
-			link->dc->res_pool->timing_generator_count;
-
-	psr_context->rfb_update_auto_en = true;
-
-	/* 2 frames before enter PSR. */
-	psr_context->timehyst_frames = 2;
-	/* half a frame
-	 * (units in 100 lines, i.e. a value of 1 represents 100 lines)
-	 */
-	psr_context->hyst_lines = stream->timing.v_total / 2 / 100;
-	psr_context->aux_repeats = 10;
-
-	psr_context->psr_level.u32all = 0;
-
-	/*skip power down the single pipe since it blocks the cstate*/
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	if (link->ctx->asic_id.chip_family >= FAMILY_RV) {
-		psr_context->psr_level.bits.SKIP_CRTC_DISABLE = true;
-		if (link->ctx->asic_id.chip_family == FAMILY_YELLOW_CARP && !dc->debug.disable_z10)
-			psr_context->psr_level.bits.SKIP_CRTC_DISABLE = false;
-	}
-#else
-	if (link->ctx->asic_id.chip_family >= FAMILY_RV)
-		psr_context->psr_level.bits.SKIP_CRTC_DISABLE = true;
-#endif
-
-	/* SMU will perform additional powerdown sequence.
-	 * For unsupported ASICs, set psr_level flag to skip PSR
-	 *  static screen notification to SMU.
-	 *  (Always set for DAL2, did not check ASIC)
-	 */
-	psr_context->allow_smu_optimizations = psr_config->allow_smu_optimizations;
-	psr_context->allow_multi_disp_optimizations = psr_config->allow_multi_disp_optimizations;
-
-	/* Complete PSR entry before aborting to prevent intermittent
-	 * freezes on certain eDPs
-	 */
-	psr_context->psr_level.bits.DISABLE_PSR_ENTRY_ABORT = 1;
-
-	/* Controls additional delay after remote frame capture before
-	 * continuing power down, default = 0
-	 */
-	psr_context->frame_delay = 0;
-
-	if (psr)
-		link->psr_settings.psr_feature_enabled = psr->funcs->psr_copy_settings(psr,
-			link, psr_context, panel_inst);
-	else
-		link->psr_settings.psr_feature_enabled = dmcu->funcs->setup_psr(dmcu, link, psr_context);
-
-	/* psr_enabled == 0 indicates setup_psr did not succeed, but this
-	 * should not happen since firmware should be running at this point
-	 */
-	if (link->psr_settings.psr_feature_enabled == 0)
-		ASSERT(0);
-
-	return true;
-
-}
-
-void dc_link_get_psr_residency(const struct dc_link *link, uint32_t *residency)
-{
-	struct dc  *dc = link->ctx->dc;
-	struct dmub_psr *psr = dc->res_pool->psr;
-	unsigned int panel_inst;
-
-	if (!dc_get_edp_link_panel_inst(dc, link, &panel_inst))
-		return;
-
-	/* PSR residency measurements only supported on DMCUB */
-	if (psr != NULL && link->psr_settings.psr_feature_enabled)
-		psr->funcs->psr_get_residency(psr, residency, panel_inst);
-	else
-		*residency = 0;
-}
-
-const struct dc_link_status *dc_link_get_status(const struct dc_link *link)
-{
-	return &link->link_status;
-}
-
-void core_link_resume(struct dc_link *link)
-{
-	if (link->connector_signal != SIGNAL_TYPE_VIRTUAL)
-		program_hpd_filter(link);
-}
-
-static struct fixed31_32 get_pbn_per_slot(struct dc_stream_state *stream)
-{
-	struct fixed31_32 mbytes_per_sec;
-	uint32_t link_rate_in_mbytes_per_sec = dc_link_bandwidth_kbps(stream->link,
-			&stream->link->cur_link_settings);
-	link_rate_in_mbytes_per_sec /= 8000; /* Kbits to MBytes */
-
-	mbytes_per_sec = dc_fixpt_from_int(link_rate_in_mbytes_per_sec);
-
-	return dc_fixpt_div_int(mbytes_per_sec, 54);
-}
-
-static struct fixed31_32 get_pbn_from_bw_in_kbps(uint64_t kbps)
-{
-	struct fixed31_32 peak_kbps;
-	uint32_t numerator = 0;
-	uint32_t denominator = 1;
-
-	/*
-	 * margin 5300ppm + 300ppm ~ 0.6% as per spec, factor is 1.006
-	 * The unit of 54/64Mbytes/sec is an arbitrary unit chosen based on
-	 * common multiplier to render an integer PBN for all link rate/lane
-	 * counts combinations
-	 * calculate
-	 * peak_kbps *= (1006/1000)
-	 * peak_kbps *= (64/54)
-	 * peak_kbps *= 8    convert to bytes
-	 */
-
-	numerator = 64 * PEAK_FACTOR_X1000;
-	denominator = 54 * 8 * 1000 * 1000;
-	kbps *= numerator;
-	peak_kbps = dc_fixpt_from_fraction(kbps, denominator);
-
-	return peak_kbps;
-}
-
-static struct fixed31_32 get_pbn_from_timing(struct pipe_ctx *pipe_ctx)
-{
-	uint64_t kbps;
-
-	kbps = dc_bandwidth_in_kbps_from_timing(&pipe_ctx->stream->timing);
-	return get_pbn_from_bw_in_kbps(kbps);
-}
-
-static void update_mst_stream_alloc_table(
-	struct dc_link *link,
-	struct stream_encoder *stream_enc,
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	struct hpo_dp_stream_encoder *hpo_dp_stream_enc, // TODO: Rename stream_enc to dio_stream_enc?
-#endif
-	const struct dp_mst_stream_allocation_table *proposed_table)
-{
-	struct link_mst_stream_allocation work_table[MAX_CONTROLLER_NUM] = { 0 };
-	struct link_mst_stream_allocation *dc_alloc;
-
-	int i;
-	int j;
-
-	/* if DRM proposed_table has more than one new payload */
-	ASSERT(proposed_table->stream_count -
-			link->mst_stream_alloc_table.stream_count < 2);
-
-	/* copy proposed_table to link, add stream encoder */
-	for (i = 0; i < proposed_table->stream_count; i++) {
-
-		for (j = 0; j < link->mst_stream_alloc_table.stream_count; j++) {
-			dc_alloc =
-			&link->mst_stream_alloc_table.stream_allocations[j];
-
-			if (dc_alloc->vcp_id ==
-				proposed_table->stream_allocations[i].vcp_id) {
-
-				work_table[i] = *dc_alloc;
-				work_table[i].slot_count = proposed_table->stream_allocations[i].slot_count;
-				break; /* exit j loop */
-			}
-		}
-
-		/* new vcp_id */
-		if (j == link->mst_stream_alloc_table.stream_count) {
-			work_table[i].vcp_id =
-				proposed_table->stream_allocations[i].vcp_id;
-			work_table[i].slot_count =
-				proposed_table->stream_allocations[i].slot_count;
-			work_table[i].stream_enc = stream_enc;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-			work_table[i].hpo_dp_stream_enc = hpo_dp_stream_enc;
-#endif
-		}
-	}
-
-	/* update link->mst_stream_alloc_table with work_table */
-	link->mst_stream_alloc_table.stream_count =
-			proposed_table->stream_count;
-	for (i = 0; i < MAX_CONTROLLER_NUM; i++)
-		link->mst_stream_alloc_table.stream_allocations[i] =
-				work_table[i];
-}
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-static void dc_log_vcp_x_y(const struct dc_link *link, struct fixed31_32 avg_time_slots_per_mtp)
-{
-	const uint32_t VCP_Y_PRECISION = 1000;
-	uint64_t vcp_x, vcp_y;
-
-	// Add 0.5*(1/VCP_Y_PRECISION) to round up to decimal precision
-	avg_time_slots_per_mtp = dc_fixpt_add(
-			avg_time_slots_per_mtp, dc_fixpt_from_fraction(1, 2 * VCP_Y_PRECISION));
-
-	vcp_x = dc_fixpt_floor(avg_time_slots_per_mtp);
-	vcp_y = dc_fixpt_floor(
-			dc_fixpt_mul_int(
-				dc_fixpt_sub_int(avg_time_slots_per_mtp, dc_fixpt_floor(avg_time_slots_per_mtp)),
-				VCP_Y_PRECISION));
-
-	if (link->type == dc_connection_mst_branch)
-		DC_LOG_DP2("MST Update Payload: set_throttled_vcp_size slot X.Y for MST stream "
-				"X: %lld Y: %lld/%d", vcp_x, vcp_y, VCP_Y_PRECISION);
-	else
-		DC_LOG_DP2("SST Update Payload: set_throttled_vcp_size slot X.Y for SST stream "
-				"X: %lld Y: %lld/%d", vcp_x, vcp_y, VCP_Y_PRECISION);
-}
-
-/*
- * Payload allocation/deallocation for SST introduced in DP2.0
- */
-enum dc_status dc_link_update_sst_payload(struct pipe_ctx *pipe_ctx, bool allocate)
-{
-	struct dc_stream_state *stream = pipe_ctx->stream;
-	struct dc_link *link = stream->link;
-	struct hpo_dp_link_encoder *hpo_dp_link_encoder = link->hpo_dp_link_enc;
-	struct hpo_dp_stream_encoder *hpo_dp_stream_encoder = pipe_ctx->stream_res.hpo_dp_stream_enc;
-	struct link_mst_stream_allocation_table proposed_table = {0};
-	struct fixed31_32 avg_time_slots_per_mtp;
-	DC_LOGGER_INIT(link->ctx->logger);
-
-	/* slot X.Y for SST payload deallocate */
-	if (!allocate) {
-		avg_time_slots_per_mtp = dc_fixpt_from_int(0);
-
-		dc_log_vcp_x_y(link, avg_time_slots_per_mtp);
-
-		hpo_dp_link_encoder->funcs->set_throttled_vcp_size(
-				hpo_dp_link_encoder,
-				hpo_dp_stream_encoder->inst,
-				avg_time_slots_per_mtp);
-	}
-
-	/* calculate VC payload and update branch with new payload allocation table*/
-	if (!dpcd_write_128b_132b_sst_payload_allocation_table(
-			stream,
-			link,
-			&proposed_table,
-			allocate)) {
-		DC_LOG_ERROR("SST Update Payload: Failed to update "
-						"allocation table for "
-						"pipe idx: %d\n",
-						pipe_ctx->pipe_idx);
-	}
-
-	proposed_table.stream_allocations[0].hpo_dp_stream_enc = hpo_dp_stream_encoder;
-
-	ASSERT(proposed_table.stream_count == 1);
-
-	//TODO - DP2.0 Logging: Instead of hpo_dp_stream_enc pointer, log instance id
-	DC_LOG_DP2("SST Update Payload: hpo_dp_stream_enc: %p      "
-		"vcp_id: %d      "
-		"slot_count: %d\n",
-		(void *) proposed_table.stream_allocations[0].hpo_dp_stream_enc,
-		proposed_table.stream_allocations[0].vcp_id,
-		proposed_table.stream_allocations[0].slot_count);
-
-	/* program DP source TX for payload */
-	hpo_dp_link_encoder->funcs->update_stream_allocation_table(
-			hpo_dp_link_encoder,
-			&proposed_table);
-
-	/* poll for ACT handled */
-	if (!dpcd_poll_for_allocation_change_trigger(link)) {
-		// Failures will result in blackscreen and errors logged
-		BREAK_TO_DEBUGGER();
-	}
-
-	/* slot X.Y for SST payload allocate */
-	if (allocate) {
-		avg_time_slots_per_mtp = calculate_sst_avg_time_slots_per_mtp(stream, link);
-
-		dc_log_vcp_x_y(link, avg_time_slots_per_mtp);
-
-		hpo_dp_link_encoder->funcs->set_throttled_vcp_size(
-				hpo_dp_link_encoder,
-				hpo_dp_stream_encoder->inst,
-				avg_time_slots_per_mtp);
-	}
-
-	/* Always return DC_OK.
-	 * If part of sequence fails, log failure(s) and show blackscreen
-	 */
-	return DC_OK;
-}
-#endif
-
-/* convert link_mst_stream_alloc_table to dm dp_mst_stream_alloc_table
- * because stream_encoder is not exposed to dm
- */
-enum dc_status dc_link_allocate_mst_payload(struct pipe_ctx *pipe_ctx)
-{
-	struct dc_stream_state *stream = pipe_ctx->stream;
-	struct dc_link *link = stream->link;
-	struct link_encoder *link_encoder = NULL;
-	struct stream_encoder *stream_encoder = pipe_ctx->stream_res.stream_enc;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	struct hpo_dp_link_encoder *hpo_dp_link_encoder = link->hpo_dp_link_enc;
-	struct hpo_dp_stream_encoder *hpo_dp_stream_encoder = pipe_ctx->stream_res.hpo_dp_stream_enc;
-#endif
-	struct dp_mst_stream_allocation_table proposed_table = {0};
-	struct fixed31_32 avg_time_slots_per_mtp;
-	struct fixed31_32 pbn;
-	struct fixed31_32 pbn_per_slot;
-	int i;
-	enum act_return_status ret;
-	DC_LOGGER_INIT(link->ctx->logger);
-
-	/* Link encoder may have been dynamically assigned to non-physical display endpoint. */
-	if (link->ep_type == DISPLAY_ENDPOINT_PHY)
-		link_encoder = link->link_enc;
-	else if (link->dc->res_pool->funcs->link_encs_assign)
-		link_encoder = link_enc_cfg_get_link_enc_used_by_stream(pipe_ctx->stream->ctx->dc, stream);
-	ASSERT(link_encoder);
-
-	/* enable_link_dp_mst already check link->enabled_stream_count
-	 * and stream is in link->stream[]. This is called during set mode,
-	 * stream_enc is available.
-	 */
-
-	/* get calculate VC payload for stream: stream_alloc */
-	if (dm_helpers_dp_mst_write_payload_allocation_table(
-		stream->ctx,
-		stream,
-		&proposed_table,
-		true)) {
-		update_mst_stream_alloc_table(
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-					link,
-					pipe_ctx->stream_res.stream_enc,
-					pipe_ctx->stream_res.hpo_dp_stream_enc,
-					&proposed_table);
-#else
-					link, pipe_ctx->stream_res.stream_enc, &proposed_table);
-#endif
-	}
-	else
-		DC_LOG_WARNING("Failed to update"
-				"MST allocation table for"
-				"pipe idx:%d\n",
-				pipe_ctx->pipe_idx);
-
-	DC_LOG_MST("%s  "
-			"stream_count: %d: \n ",
-			__func__,
-			link->mst_stream_alloc_table.stream_count);
-
-	for (i = 0; i < MAX_CONTROLLER_NUM; i++) {
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-		DC_LOG_MST("stream_enc[%d]: %p      "
-		"stream[%d].hpo_dp_stream_enc: %p      "
-		"stream[%d].vcp_id: %d      "
-		"stream[%d].slot_count: %d\n",
-		i,
-		(void *) link->mst_stream_alloc_table.stream_allocations[i].stream_enc,
-		i,
-		(void *) link->mst_stream_alloc_table.stream_allocations[i].hpo_dp_stream_enc,
-		i,
-		link->mst_stream_alloc_table.stream_allocations[i].vcp_id,
-		i,
-		link->mst_stream_alloc_table.stream_allocations[i].slot_count);
-#else
-		DC_LOG_MST("stream_enc[%d]: %p      "
-		"stream[%d].vcp_id: %d      "
-		"stream[%d].slot_count: %d\n",
-		i,
-		(void *) link->mst_stream_alloc_table.stream_allocations[i].stream_enc,
-		i,
-		link->mst_stream_alloc_table.stream_allocations[i].vcp_id,
-		i,
-		link->mst_stream_alloc_table.stream_allocations[i].slot_count);
-#endif
-	}
-
-	ASSERT(proposed_table.stream_count > 0);
-
-	if (link->ep_type == DISPLAY_ENDPOINT_USB4_DPIA) {
-		static enum dc_status status;
-		uint8_t mst_alloc_slots = 0, prev_mst_slots_in_use = 0xFF;
-
-		for (i = 0; i < link->mst_stream_alloc_table.stream_count; i++)
-			mst_alloc_slots += link->mst_stream_alloc_table.stream_allocations[i].slot_count;
-
-		status = dc_process_dmub_set_mst_slots(link->dc, link->link_index,
-			mst_alloc_slots, &prev_mst_slots_in_use);
-		ASSERT(status == DC_OK);
-		DC_LOG_MST("dpia : status[%d]: alloc_slots[%d]: used_slots[%d]\n",
-				status, mst_alloc_slots, prev_mst_slots_in_use);
-	}
-
-	/* program DP source TX for payload */
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	switch (dp_get_link_encoding_format(&link->cur_link_settings)) {
-	case DP_8b_10b_ENCODING:
-		link_encoder->funcs->update_mst_stream_allocation_table(
-			link_encoder,
-			&link->mst_stream_alloc_table);
-		break;
-	case DP_128b_132b_ENCODING:
-		hpo_dp_link_encoder->funcs->update_stream_allocation_table(
-				hpo_dp_link_encoder,
-				&link->mst_stream_alloc_table);
-		break;
-	case DP_UNKNOWN_ENCODING:
-		DC_LOG_ERROR("Failure: unknown encoding format\n");
-		return DC_ERROR_UNEXPECTED;
-	}
-#else
-	link_encoder->funcs->update_mst_stream_allocation_table(
-		link_encoder,
-		&link->mst_stream_alloc_table);
-#endif
-
-	/* send down message */
-	ret = dm_helpers_dp_mst_poll_for_allocation_change_trigger(
-			stream->ctx,
-			stream);
-
-	if (ret != ACT_LINK_LOST) {
-		dm_helpers_dp_mst_send_payload_allocation(
-				stream->ctx,
-				stream,
-				true);
-	}
-
-	/* slot X.Y for only current stream */
-	pbn_per_slot = get_pbn_per_slot(stream);
-	if (pbn_per_slot.value == 0) {
-		DC_LOG_ERROR("Failure: pbn_per_slot==0 not allowed. Cannot continue, returning DC_UNSUPPORTED_VALUE.\n");
-		return DC_UNSUPPORTED_VALUE;
-	}
-	pbn = get_pbn_from_timing(pipe_ctx);
-	avg_time_slots_per_mtp = dc_fixpt_div(pbn, pbn_per_slot);
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	switch (dp_get_link_encoding_format(&link->cur_link_settings)) {
-	case DP_8b_10b_ENCODING:
-		stream_encoder->funcs->set_throttled_vcp_size(
-			stream_encoder,
-			avg_time_slots_per_mtp);
-		break;
-	case DP_128b_132b_ENCODING:
-		hpo_dp_link_encoder->funcs->set_throttled_vcp_size(
-				hpo_dp_link_encoder,
-				hpo_dp_stream_encoder->inst,
-				avg_time_slots_per_mtp);
-		break;
-	case DP_UNKNOWN_ENCODING:
-		DC_LOG_ERROR("Failure: unknown encoding format\n");
-		return DC_ERROR_UNEXPECTED;
-	}
-#else
-	stream_encoder->funcs->set_throttled_vcp_size(
-		stream_encoder,
-		avg_time_slots_per_mtp);
-#endif
-
-	return DC_OK;
-
-}
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-enum dc_status dc_link_reduce_mst_payload(struct pipe_ctx *pipe_ctx, uint32_t bw_in_kbps)
-{
-	struct dc_stream_state *stream = pipe_ctx->stream;
-	struct dc_link *link = stream->link;
-	struct fixed31_32 avg_time_slots_per_mtp;
-	struct fixed31_32 pbn;
-	struct fixed31_32 pbn_per_slot;
-	struct link_encoder *link_encoder = link->link_enc;
-	struct stream_encoder *stream_encoder = pipe_ctx->stream_res.stream_enc;
-	struct dp_mst_stream_allocation_table proposed_table = {0};
-	uint8_t i;
-	enum act_return_status ret;
-	DC_LOGGER_INIT(link->ctx->logger);
-
-	/* decrease throttled vcp size */
-	pbn_per_slot = get_pbn_per_slot(stream);
-	pbn = get_pbn_from_bw_in_kbps(bw_in_kbps);
-	avg_time_slots_per_mtp = dc_fixpt_div(pbn, pbn_per_slot);
-
-	stream_encoder->funcs->set_throttled_vcp_size(
-				stream_encoder,
-				avg_time_slots_per_mtp);
-
-	/* send ALLOCATE_PAYLOAD sideband message with updated pbn */
-	dm_helpers_dp_mst_send_payload_allocation(
-			stream->ctx,
-			stream,
-			true);
-
-	/* notify immediate branch device table update */
-	if (dm_helpers_dp_mst_write_payload_allocation_table(
-			stream->ctx,
-			stream,
-			&proposed_table,
-			true)) {
-		/* update mst stream allocation table software state */
-		update_mst_stream_alloc_table(
-				link,
-				pipe_ctx->stream_res.stream_enc,
-				pipe_ctx->stream_res.hpo_dp_stream_enc,
-				&proposed_table);
-	} else {
-		DC_LOG_WARNING("Failed to update"
-				"MST allocation table for"
-				"pipe idx:%d\n",
-				pipe_ctx->pipe_idx);
-	}
-
-	DC_LOG_MST("%s  "
-			"stream_count: %d: \n ",
-			__func__,
-			link->mst_stream_alloc_table.stream_count);
-
-	for (i = 0; i < MAX_CONTROLLER_NUM; i++) {
-		DC_LOG_MST("stream_enc[%d]: %p      "
-				"stream[%d].vcp_id: %d      "
-				"stream[%d].slot_count: %d\n",
-				i,
-				(void *) link->mst_stream_alloc_table.stream_allocations[i].stream_enc,
-				i,
-				link->mst_stream_alloc_table.stream_allocations[i].vcp_id,
-				i,
-				link->mst_stream_alloc_table.stream_allocations[i].slot_count);
-	}
-
-	ASSERT(proposed_table.stream_count > 0);
-
-	/* update mst stream allocation table hardware state */
-	link_encoder->funcs->update_mst_stream_allocation_table(
-			link_encoder,
-			&link->mst_stream_alloc_table);
-
-	/* poll for immediate branch device ACT handled */
-	ret = dm_helpers_dp_mst_poll_for_allocation_change_trigger(
-			stream->ctx,
-			stream);
-
-	return DC_OK;
-}
-
-enum dc_status dc_link_increase_mst_payload(struct pipe_ctx *pipe_ctx, uint32_t bw_in_kbps)
-{
-	struct dc_stream_state *stream = pipe_ctx->stream;
-	struct dc_link *link = stream->link;
-	struct fixed31_32 avg_time_slots_per_mtp;
-	struct fixed31_32 pbn;
-	struct fixed31_32 pbn_per_slot;
-	struct link_encoder *link_encoder = link->link_enc;
-	struct stream_encoder *stream_encoder = pipe_ctx->stream_res.stream_enc;
-	struct dp_mst_stream_allocation_table proposed_table = {0};
-	uint8_t i;
-	enum act_return_status ret;
-	DC_LOGGER_INIT(link->ctx->logger);
-
-	/* notify immediate branch device table update */
-	if (dm_helpers_dp_mst_write_payload_allocation_table(
-				stream->ctx,
-				stream,
-				&proposed_table,
-				true)) {
-		/* update mst stream allocation table software state */
-		update_mst_stream_alloc_table(
-				link,
-				pipe_ctx->stream_res.stream_enc,
-				pipe_ctx->stream_res.hpo_dp_stream_enc,
-				&proposed_table);
-	}
-
-	DC_LOG_MST("%s  "
-			"stream_count: %d: \n ",
-			__func__,
-			link->mst_stream_alloc_table.stream_count);
-
-	for (i = 0; i < MAX_CONTROLLER_NUM; i++) {
-		DC_LOG_MST("stream_enc[%d]: %p      "
-				"stream[%d].vcp_id: %d      "
-				"stream[%d].slot_count: %d\n",
-				i,
-				(void *) link->mst_stream_alloc_table.stream_allocations[i].stream_enc,
-				i,
-				link->mst_stream_alloc_table.stream_allocations[i].vcp_id,
-				i,
-				link->mst_stream_alloc_table.stream_allocations[i].slot_count);
-	}
-
-	ASSERT(proposed_table.stream_count > 0);
-
-	/* update mst stream allocation table hardware state */
-	link_encoder->funcs->update_mst_stream_allocation_table(
-			link_encoder,
-			&link->mst_stream_alloc_table);
-
-	/* poll for immediate branch device ACT handled */
-	ret = dm_helpers_dp_mst_poll_for_allocation_change_trigger(
-			stream->ctx,
-			stream);
-
-	if (ret != ACT_LINK_LOST) {
-		/* send ALLOCATE_PAYLOAD sideband message with updated pbn */
-		dm_helpers_dp_mst_send_payload_allocation(
-				stream->ctx,
-				stream,
-				true);
-	}
-
-	/* increase throttled vcp size */
-	pbn = get_pbn_from_bw_in_kbps(bw_in_kbps);
-	pbn_per_slot = get_pbn_per_slot(stream);
-	avg_time_slots_per_mtp = dc_fixpt_div(pbn, pbn_per_slot);
-
-	stream_encoder->funcs->set_throttled_vcp_size(
-				stream_encoder,
-				avg_time_slots_per_mtp);
-
-	return DC_OK;
-}
-#endif
-
-static enum dc_status deallocate_mst_payload(struct pipe_ctx *pipe_ctx)
-{
-	struct dc_stream_state *stream = pipe_ctx->stream;
-	struct dc_link *link = stream->link;
-	struct link_encoder *link_encoder = NULL;
-	struct stream_encoder *stream_encoder = pipe_ctx->stream_res.stream_enc;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	struct hpo_dp_link_encoder *hpo_dp_link_encoder = link->hpo_dp_link_enc;
-	struct hpo_dp_stream_encoder *hpo_dp_stream_encoder = pipe_ctx->stream_res.hpo_dp_stream_enc;
-#endif
-	struct dp_mst_stream_allocation_table proposed_table = {0};
-	struct fixed31_32 avg_time_slots_per_mtp = dc_fixpt_from_int(0);
-	int i;
-	bool mst_mode = (link->type == dc_connection_mst_branch);
-	DC_LOGGER_INIT(link->ctx->logger);
-
-	/* Link encoder may have been dynamically assigned to non-physical display endpoint. */
-	if (link->ep_type == DISPLAY_ENDPOINT_PHY)
-		link_encoder = link->link_enc;
-	else if (link->dc->res_pool->funcs->link_encs_assign)
-		link_encoder = link_enc_cfg_get_link_enc_used_by_stream(pipe_ctx->stream->ctx->dc, stream);
-	ASSERT(link_encoder);
-
-	/* deallocate_mst_payload is called before disable link. When mode or
-	 * disable/enable monitor, new stream is created which is not in link
-	 * stream[] yet. For this, payload is not allocated yet, so de-alloc
-	 * should not done. For new mode set, map_resources will get engine
-	 * for new stream, so stream_enc->id should be validated until here.
-	 */
-
-	/* slot X.Y */
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	switch (dp_get_link_encoding_format(&link->cur_link_settings)) {
-	case DP_8b_10b_ENCODING:
-		stream_encoder->funcs->set_throttled_vcp_size(
-			stream_encoder,
-			avg_time_slots_per_mtp);
-		break;
-	case DP_128b_132b_ENCODING:
-		hpo_dp_link_encoder->funcs->set_throttled_vcp_size(
-				hpo_dp_link_encoder,
-				hpo_dp_stream_encoder->inst,
-				avg_time_slots_per_mtp);
-		break;
-	case DP_UNKNOWN_ENCODING:
-		DC_LOG_ERROR("Failure: unknown encoding format\n");
-		return DC_ERROR_UNEXPECTED;
-	}
-#else
-	stream_encoder->funcs->set_throttled_vcp_size(
-		stream_encoder,
-		avg_time_slots_per_mtp);
-#endif
-
-	/* TODO: which component is responsible for remove payload table? */
-	if (mst_mode) {
-		if (dm_helpers_dp_mst_write_payload_allocation_table(
-				stream->ctx,
-				stream,
-				&proposed_table,
-				false)) {
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-			update_mst_stream_alloc_table(
-						link,
-						pipe_ctx->stream_res.stream_enc,
-						pipe_ctx->stream_res.hpo_dp_stream_enc,
-						&proposed_table);
-#else
-			update_mst_stream_alloc_table(
-				link, pipe_ctx->stream_res.stream_enc, &proposed_table);
-#endif
-		}
-		else {
-				DC_LOG_WARNING("Failed to update"
-						"MST allocation table for"
-						"pipe idx:%d\n",
-						pipe_ctx->pipe_idx);
-		}
-	}
-
-	DC_LOG_MST("%s"
-			"stream_count: %d: ",
-			__func__,
-			link->mst_stream_alloc_table.stream_count);
-
-	for (i = 0; i < MAX_CONTROLLER_NUM; i++) {
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-		DC_LOG_MST("stream_enc[%d]: %p      "
-		"stream[%d].hpo_dp_stream_enc: %p      "
-		"stream[%d].vcp_id: %d      "
-		"stream[%d].slot_count: %d\n",
-		i,
-		(void *) link->mst_stream_alloc_table.stream_allocations[i].stream_enc,
-		i,
-		(void *) link->mst_stream_alloc_table.stream_allocations[i].hpo_dp_stream_enc,
-		i,
-		link->mst_stream_alloc_table.stream_allocations[i].vcp_id,
-		i,
-		link->mst_stream_alloc_table.stream_allocations[i].slot_count);
-#else
-		DC_LOG_MST("stream_enc[%d]: %p      "
-		"stream[%d].vcp_id: %d      "
-		"stream[%d].slot_count: %d\n",
-		i,
-		(void *) link->mst_stream_alloc_table.stream_allocations[i].stream_enc,
-		i,
-		link->mst_stream_alloc_table.stream_allocations[i].vcp_id,
-		i,
-		link->mst_stream_alloc_table.stream_allocations[i].slot_count);
-#endif
-	}
-
-	if (link->ep_type == DISPLAY_ENDPOINT_USB4_DPIA) {
-		enum dc_status status;
-		uint8_t mst_alloc_slots = 0, prev_mst_slots_in_use = 0xFF;
-
-		for (i = 0; i < link->mst_stream_alloc_table.stream_count; i++)
-			mst_alloc_slots += link->mst_stream_alloc_table.stream_allocations[i].slot_count;
-
-		status = dc_process_dmub_set_mst_slots(link->dc, link->link_index,
-			mst_alloc_slots, &prev_mst_slots_in_use);
-		ASSERT(status != DC_NOT_SUPPORTED);
-		DC_LOG_MST("dpia : status[%d]: alloc_slots[%d]: used_slots[%d]\n",
-				status, mst_alloc_slots, prev_mst_slots_in_use);
-	}
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	switch (dp_get_link_encoding_format(&link->cur_link_settings)) {
-	case DP_8b_10b_ENCODING:
-		link_encoder->funcs->update_mst_stream_allocation_table(
-			link_encoder,
-			&link->mst_stream_alloc_table);
-		break;
-	case DP_128b_132b_ENCODING:
-		hpo_dp_link_encoder->funcs->update_stream_allocation_table(
-				hpo_dp_link_encoder,
-				&link->mst_stream_alloc_table);
-		break;
-	case DP_UNKNOWN_ENCODING:
-		DC_LOG_ERROR("Failure: unknown encoding format\n");
-		return DC_ERROR_UNEXPECTED;
-	}
-#else
-	link_encoder->funcs->update_mst_stream_allocation_table(
-		link_encoder,
-		&link->mst_stream_alloc_table);
-#endif
-
-	if (mst_mode) {
-		dm_helpers_dp_mst_poll_for_allocation_change_trigger(
-			stream->ctx,
-			stream);
-
-		dm_helpers_dp_mst_send_payload_allocation(
-			stream->ctx,
-			stream,
-			false);
-	}
-
-	return DC_OK;
-}
-
-
-#if defined(CONFIG_DRM_AMD_DC_HDCP)
-static void update_psp_stream_config(struct pipe_ctx *pipe_ctx, bool dpms_off)
-{
-	struct cp_psp *cp_psp = &pipe_ctx->stream->ctx->cp_psp;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	struct link_encoder *link_enc = NULL;
-	struct dc_state *state = pipe_ctx->stream->ctx->dc->current_state;
-	struct link_enc_assignment link_enc_assign;
-	int i;
-#endif
-
-	if (cp_psp && cp_psp->funcs.update_stream_config) {
-		struct cp_psp_stream_config config = {0};
-		enum dp_panel_mode panel_mode =
-				dp_get_panel_mode(pipe_ctx->stream->link);
-
-		config.otg_inst = (uint8_t) pipe_ctx->stream_res.tg->inst;
-		/*stream_enc_inst*/
-		config.dig_fe = (uint8_t) pipe_ctx->stream_res.stream_enc->stream_enc_inst;
-		config.dig_be = pipe_ctx->stream->link->link_enc_hw_inst;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-		config.stream_enc_idx = pipe_ctx->stream_res.stream_enc->id - ENGINE_ID_DIGA;
-		
-		if (pipe_ctx->stream->link->ep_type == DISPLAY_ENDPOINT_PHY ||
-				pipe_ctx->stream->link->ep_type == DISPLAY_ENDPOINT_USB4_DPIA) {
-			link_enc = pipe_ctx->stream->link->link_enc;
-			config.dio_output_type = pipe_ctx->stream->link->ep_type;
-			config.dio_output_idx = link_enc->transmitter - TRANSMITTER_UNIPHY_A;
-			if (pipe_ctx->stream->link->ep_type == DISPLAY_ENDPOINT_PHY)
-				link_enc = pipe_ctx->stream->link->link_enc;
-			else if (pipe_ctx->stream->link->ep_type == DISPLAY_ENDPOINT_USB4_DPIA)
-				if (pipe_ctx->stream->link->dc->res_pool->funcs->link_encs_assign) {
-					link_enc = link_enc_cfg_get_link_enc_used_by_stream(
-							pipe_ctx->stream->ctx->dc,
-							pipe_ctx->stream);
-			}
-			// Initialize PHY ID with ABCDE - 01234 mapping except when it is B0
-			config.phy_idx = link_enc->transmitter - TRANSMITTER_UNIPHY_A;
-
-			//look up the link_enc_assignment for the current pipe_ctx
-			for (i = 0; i < state->stream_count; i++) {
-				if (pipe_ctx->stream == state->streams[i]) {
-					link_enc_assign = state->res_ctx.link_enc_cfg_ctx.link_enc_assignments[i];
-				}
-			}
-			// Add flag to guard new A0 DIG mapping
-			if (pipe_ctx->stream->ctx->dc->enable_c20_dtm_b0 == true) {
-				config.dig_be = link_enc_assign.eng_id;
-				config.dio_output_type = pipe_ctx->stream->link->ep_type;
-				config.dio_output_idx = link_enc->transmitter - TRANSMITTER_UNIPHY_A;
-			} else {
-				config.dio_output_type = 0;
-				config.dio_output_idx = 0;
-			}
-
-			// Add flag to guard B0 implementation
-			if (pipe_ctx->stream->ctx->dc->enable_c20_dtm_b0 == true &&
-					link_enc->ctx->asic_id.hw_internal_rev == YELLOW_CARP_B0) {
-				if (pipe_ctx->stream->link->ep_type == DISPLAY_ENDPOINT_USB4_DPIA) {
-					link_enc = link_enc_assign.stream->link_enc;
-
-					// enum ID 1-4 maps to DPIA PHY ID 0-3
-					config.phy_idx = link_enc_assign.ep_id.link_id.enum_id - ENUM_ID_1;
-				} else {  // for non DPIA mode over B0, ABCDE maps to 01564
-
-					switch (link_enc->transmitter) {
-					case TRANSMITTER_UNIPHY_A:
-						config.phy_idx = 0;
-						break;
-					case TRANSMITTER_UNIPHY_B:
-						config.phy_idx = 1;
-						break;
-					case TRANSMITTER_UNIPHY_C:
-						config.phy_idx = 5;
-						break;
-					case TRANSMITTER_UNIPHY_D:
-						config.phy_idx = 6;
-						break;
-					case TRANSMITTER_UNIPHY_E:
-						config.phy_idx = 4;
-						break;
-					default:
-						config.phy_idx = 0;
-						break;
-					}
-
-				}
-			}
-		} else if (pipe_ctx->stream->link->dc->res_pool->funcs->link_encs_assign) {
-			link_enc = link_enc_cfg_get_link_enc_used_by_stream(
-					pipe_ctx->stream->ctx->dc,
-					pipe_ctx->stream);
-			config.phy_idx = 0; /* Clear phy_idx for non-physical display endpoints. */
-		}
-		ASSERT(link_enc);
-		if (link_enc)
-			config.link_enc_idx = link_enc->transmitter - TRANSMITTER_UNIPHY_A;
-		if (is_dp_128b_132b_signal(pipe_ctx)) {
-			config.stream_enc_idx = pipe_ctx->stream_res.hpo_dp_stream_enc->id - ENGINE_ID_HPO_DP_0;
-			config.link_enc_idx = pipe_ctx->stream->link->hpo_dp_link_enc->inst;
-			config.dp2_enabled = 1;
-		}
-#endif
-		config.dpms_off = dpms_off;
-		config.dm_stream_ctx = pipe_ctx->stream->dm_stream_context;
-		config.assr_enabled = (panel_mode == DP_PANEL_MODE_EDP);
-		config.mst_enabled = (pipe_ctx->stream->signal ==
-				SIGNAL_TYPE_DISPLAY_PORT_MST);
-		cp_psp->funcs.update_stream_config(cp_psp->handle, &config);
-	}
-}
-#endif
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-static void fpga_dp_hpo_enable_link_and_stream(struct dc_state *state, struct pipe_ctx *pipe_ctx)
-{
-	struct dc *dc = pipe_ctx->stream->ctx->dc;
-	struct dc_stream_state *stream = pipe_ctx->stream;
-	struct link_mst_stream_allocation_table proposed_table = {0};
-	struct fixed31_32 avg_time_slots_per_mtp;
-	uint8_t req_slot_count = 0;
-	uint8_t vc_id = 1; /// VC ID always 1 for SST
-
-	struct dc_link_settings link_settings = {0};
-	DC_LOGGER_INIT(pipe_ctx->stream->ctx->logger);
-
-	decide_link_settings(stream, &link_settings);
-	stream->link->cur_link_settings = link_settings;
-
-	/*  Enable clock, Configure lane count, and Enable Link Encoder*/
-	enable_dp_hpo_output(stream->link, &stream->link->cur_link_settings);
-
-#ifdef DIAGS_BUILD
-	/* Workaround for FPGA HPO capture DP link data:
-	 * HPO capture will set link to active mode
-	 * This workaround is required to get a capture from start of frame
-	 */
-	if (!dc->debug.fpga_hpo_capture_en) {
-		struct encoder_set_dp_phy_pattern_param params = {0};
-		params.dp_phy_pattern = DP_TEST_PATTERN_VIDEO_MODE;
-
-		/* Set link active */
-		stream->link->hpo_dp_link_enc->funcs->set_link_test_pattern(
-				stream->link->hpo_dp_link_enc,
-				&params);
-	}
-#endif
-
-	/* Enable DP_STREAM_ENC */
-	dc->hwss.enable_stream(pipe_ctx);
-
-	/* Set DPS PPS SDP (AKA "info frames") */
-	if (pipe_ctx->stream->timing.flags.DSC) {
-		dp_set_dsc_pps_sdp(pipe_ctx, true, true);
-	}
-
-	/* Allocate Payload */
-	if ((stream->signal == SIGNAL_TYPE_DISPLAY_PORT_MST) && (state->stream_count > 1)) {
-		// MST case
-		uint8_t i;
-
-		proposed_table.stream_count = state->stream_count;
-		for (i = 0; i < state->stream_count; i++) {
-			avg_time_slots_per_mtp = calculate_sst_avg_time_slots_per_mtp(state->streams[i], state->streams[i]->link);
-			req_slot_count = dc_fixpt_ceil(avg_time_slots_per_mtp);
-			proposed_table.stream_allocations[i].slot_count = req_slot_count;
-			proposed_table.stream_allocations[i].vcp_id = i+1;
-			/* NOTE: This makes assumption that pipe_ctx index is same as stream index */
-			proposed_table.stream_allocations[i].hpo_dp_stream_enc = state->res_ctx.pipe_ctx[i].stream_res.hpo_dp_stream_enc;
-		}
-	} else {
-		// SST case
-		avg_time_slots_per_mtp = calculate_sst_avg_time_slots_per_mtp(stream, stream->link);
-		req_slot_count = dc_fixpt_ceil(avg_time_slots_per_mtp);
-		proposed_table.stream_count = 1; /// Always 1 stream for SST
-		proposed_table.stream_allocations[0].slot_count = req_slot_count;
-		proposed_table.stream_allocations[0].vcp_id = vc_id;
-		proposed_table.stream_allocations[0].hpo_dp_stream_enc = pipe_ctx->stream_res.hpo_dp_stream_enc;
-	}
-
-	stream->link->hpo_dp_link_enc->funcs->update_stream_allocation_table(
-			stream->link->hpo_dp_link_enc,
-			&proposed_table);
-
-	stream->link->hpo_dp_link_enc->funcs->set_throttled_vcp_size(
-			stream->link->hpo_dp_link_enc,
-			pipe_ctx->stream_res.hpo_dp_stream_enc->inst,
-			avg_time_slots_per_mtp);
-
-
-
-	dc->hwss.unblank_stream(pipe_ctx, &stream->link->cur_link_settings);
-}
-#endif
-
-void core_link_enable_stream(
-		struct dc_state *state,
-		struct pipe_ctx *pipe_ctx)
-{
-	struct dc *dc = pipe_ctx->stream->ctx->dc;
-	struct dc_stream_state *stream = pipe_ctx->stream;
-	struct dc_link *link = stream->sink->link;
-	enum dc_status status;
-	struct link_encoder *link_enc;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	enum otg_out_mux_dest otg_out_dest = OUT_MUX_DIO;
-	struct vpg *vpg = pipe_ctx->stream_res.stream_enc->vpg;
-
-	if (is_dp_128b_132b_signal(pipe_ctx))
-		vpg = pipe_ctx->stream_res.hpo_dp_stream_enc->vpg;
-#endif
-	DC_LOGGER_INIT(pipe_ctx->stream->ctx->logger);
-
-	if (!IS_DIAG_DC(dc->ctx->dce_environment) &&
-			dc_is_virtual_signal(pipe_ctx->stream->signal))
-		return;
-
-	if (dc->res_pool->funcs->link_encs_assign && stream->link->ep_type != DISPLAY_ENDPOINT_PHY)
-		link_enc = link_enc_cfg_get_link_enc_used_by_stream(dc, stream);
-	else
-		link_enc = stream->link->link_enc;
-	ASSERT(link_enc);
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	if (!dc_is_virtual_signal(pipe_ctx->stream->signal)
-			&& !is_dp_128b_132b_signal(pipe_ctx)) {
-#else
-	if (!dc_is_virtual_signal(pipe_ctx->stream->signal)) {
-#endif
-		if (link_enc)
-			link_enc->funcs->setup(
-				link_enc,
-				pipe_ctx->stream->signal);
-		pipe_ctx->stream_res.stream_enc->funcs->setup_stereo_sync(
-			pipe_ctx->stream_res.stream_enc,
-			pipe_ctx->stream_res.tg->inst,
-			stream->timing.timing_3d_format != TIMING_3D_FORMAT_NONE);
-	}
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	if (is_dp_128b_132b_signal(pipe_ctx)) {
-		pipe_ctx->stream_res.hpo_dp_stream_enc->funcs->set_stream_attribute(
-				pipe_ctx->stream_res.hpo_dp_stream_enc,
-				&stream->timing,
-				stream->output_color_space,
-				stream->use_vsc_sdp_for_colorimetry,
-				stream->timing.flags.DSC,
-				false);
-		otg_out_dest = OUT_MUX_HPO_DP;
-	} else if (dc_is_dp_signal(pipe_ctx->stream->signal)) {
-		pipe_ctx->stream_res.stream_enc->funcs->dp_set_stream_attribute(
-				pipe_ctx->stream_res.stream_enc,
-				&stream->timing,
-				stream->output_color_space,
-				stream->use_vsc_sdp_for_colorimetry,
-				stream->link->dpcd_caps.dprx_feature.bits.SST_SPLIT_SDP_CAP);
-	}
-#else
-	pipe_ctx->stream_res.stream_enc->funcs->dp_set_stream_attribute(
-			pipe_ctx->stream_res.stream_enc,
-			&stream->timing,
-			stream->output_color_space,
-			stream->use_vsc_sdp_for_colorimetry,
-			stream->link->dpcd_caps.dprx_feature.bits.SST_SPLIT_SDP_CAP);
-#endif
-
-	if (dc_is_dp_signal(pipe_ctx->stream->signal))
-		dp_source_sequence_trace(link, DPCD_SOURCE_SEQ_AFTER_DP_STREAM_ATTR);
-
-	if (dc_is_hdmi_tmds_signal(pipe_ctx->stream->signal))
-		pipe_ctx->stream_res.stream_enc->funcs->hdmi_set_stream_attribute(
-			pipe_ctx->stream_res.stream_enc,
-			&stream->timing,
-			stream->phy_pix_clk,
-			pipe_ctx->stream_res.audio != NULL);
-
-	pipe_ctx->stream->link->link_state_valid = true;
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	if (pipe_ctx->stream_res.tg->funcs->set_out_mux)
-		pipe_ctx->stream_res.tg->funcs->set_out_mux(pipe_ctx->stream_res.tg, otg_out_dest);
-#endif
-
-	if (dc_is_dvi_signal(pipe_ctx->stream->signal))
-		pipe_ctx->stream_res.stream_enc->funcs->dvi_set_stream_attribute(
-			pipe_ctx->stream_res.stream_enc,
-			&stream->timing,
-			(pipe_ctx->stream->signal == SIGNAL_TYPE_DVI_DUAL_LINK) ?
-			true : false);
-
-	if (dc_is_lvds_signal(pipe_ctx->stream->signal))
-		pipe_ctx->stream_res.stream_enc->funcs->lvds_set_stream_attribute(
-			pipe_ctx->stream_res.stream_enc,
-			&stream->timing);
-
-	if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) {
-		bool apply_edp_fast_boot_optimization =
-			pipe_ctx->stream->apply_edp_fast_boot_optimization;
-
-		pipe_ctx->stream->apply_edp_fast_boot_optimization = false;
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-		// Enable VPG before building infoframe
-		if (vpg && vpg->funcs->vpg_poweron)
-			vpg->funcs->vpg_poweron(vpg);
-#endif
-
-		resource_build_info_frame(pipe_ctx);
-		dc->hwss.update_info_frame(pipe_ctx);
-
-		if (dc_is_dp_signal(pipe_ctx->stream->signal))
-			dp_source_sequence_trace(link, DPCD_SOURCE_SEQ_AFTER_UPDATE_INFO_FRAME);
-
-		/* Do not touch link on seamless boot optimization. */
-		if (pipe_ctx->stream->apply_seamless_boot_optimization) {
-			pipe_ctx->stream->dpms_off = false;
-
-			/* Still enable stream features & audio on seamless boot for DP external displays */
-			if (pipe_ctx->stream->signal == SIGNAL_TYPE_DISPLAY_PORT) {
-				enable_stream_features(pipe_ctx);
-				if (pipe_ctx->stream_res.audio != NULL) {
-					pipe_ctx->stream_res.stream_enc->funcs->dp_audio_enable(pipe_ctx->stream_res.stream_enc);
-					dc->hwss.enable_audio_stream(pipe_ctx);
-				}
-			}
-
-#if defined(CONFIG_DRM_AMD_DC_HDCP)
-			update_psp_stream_config(pipe_ctx, false);
-#endif
-			return;
-		}
-
-		/* eDP lit up by bios already, no need to enable again. */
-		if (pipe_ctx->stream->signal == SIGNAL_TYPE_EDP &&
-					apply_edp_fast_boot_optimization &&
-					!pipe_ctx->stream->timing.flags.DSC) {
-			pipe_ctx->stream->dpms_off = false;
-#if defined(CONFIG_DRM_AMD_DC_HDCP)
-			update_psp_stream_config(pipe_ctx, false);
-#endif
-			return;
-		}
-
-		if (pipe_ctx->stream->dpms_off)
-			return;
-
-		/* Have to setup DSC before DIG FE and BE are connected (which happens before the
-		 * link training). This is to make sure the bandwidth sent to DIG BE won't be
-		 * bigger than what the link and/or DIG BE can handle. VBID[6]/CompressedStream_flag
-		 * will be automatically set at a later time when the video is enabled
-		 * (DP_VID_STREAM_EN = 1).
-		 */
-		if (pipe_ctx->stream->timing.flags.DSC) {
-			if (dc_is_dp_signal(pipe_ctx->stream->signal) ||
-					dc_is_virtual_signal(pipe_ctx->stream->signal))
-				dp_set_dsc_enable(pipe_ctx, true);
-		}
-
-		status = enable_link(state, pipe_ctx);
-
-		if (status != DC_OK) {
-			DC_LOG_WARNING("enabling link %u failed: %d\n",
-			pipe_ctx->stream->link->link_index,
-			status);
-
-			/* Abort stream enable *unless* the failure was due to
-			 * DP link training - some DP monitors will recover and
-			 * show the stream anyway. But MST displays can't proceed
-			 * without link training.
-			 */
-			if (status != DC_FAIL_DP_LINK_TRAINING ||
-					pipe_ctx->stream->signal == SIGNAL_TYPE_DISPLAY_PORT_MST) {
-				if (false == stream->link->link_status.link_active)
-					disable_link(stream->link, pipe_ctx->stream->signal);
-				BREAK_TO_DEBUGGER();
-				return;
-			}
-		}
-
-		/* turn off otg test pattern if enable */
-		if (pipe_ctx->stream_res.tg->funcs->set_test_pattern)
-			pipe_ctx->stream_res.tg->funcs->set_test_pattern(pipe_ctx->stream_res.tg,
-					CONTROLLER_DP_TEST_PATTERN_VIDEOMODE,
-					COLOR_DEPTH_UNDEFINED);
-
-		/* This second call is needed to reconfigure the DIG
-		 * as a workaround for the incorrect value being applied
-		 * from transmitter control.
-		 */
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-		if (!(dc_is_virtual_signal(pipe_ctx->stream->signal) ||
-				is_dp_128b_132b_signal(pipe_ctx)))
-#else
-		if (!dc_is_virtual_signal(pipe_ctx->stream->signal))
-#endif
-			if (link_enc)
-				link_enc->funcs->setup(
-					link_enc,
-					pipe_ctx->stream->signal);
-
-		dc->hwss.enable_stream(pipe_ctx);
-
-		/* Set DPS PPS SDP (AKA "info frames") */
-		if (pipe_ctx->stream->timing.flags.DSC) {
-			if (dc_is_dp_signal(pipe_ctx->stream->signal) ||
-					dc_is_virtual_signal(pipe_ctx->stream->signal)) {
-				dp_set_dsc_on_rx(pipe_ctx, true);
-				dp_set_dsc_pps_sdp(pipe_ctx, true, true);
-			}
-		}
-
-		if (pipe_ctx->stream->signal == SIGNAL_TYPE_DISPLAY_PORT_MST)
-			dc_link_allocate_mst_payload(pipe_ctx);
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-		else if (pipe_ctx->stream->signal == SIGNAL_TYPE_DISPLAY_PORT &&
-				is_dp_128b_132b_signal(pipe_ctx))
-			dc_link_update_sst_payload(pipe_ctx, true);
-#endif
-
-		dc->hwss.unblank_stream(pipe_ctx,
-			&pipe_ctx->stream->link->cur_link_settings);
-
-		if (stream->sink_patches.delay_ignore_msa > 0)
-			msleep(stream->sink_patches.delay_ignore_msa);
-
-		if (dc_is_dp_signal(pipe_ctx->stream->signal))
-			enable_stream_features(pipe_ctx);
-#if defined(CONFIG_DRM_AMD_DC_HDCP)
-		update_psp_stream_config(pipe_ctx, false);
-#endif
-
-		dc->hwss.enable_audio_stream(pipe_ctx);
-
-	} else { // if (IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment))
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-		if (is_dp_128b_132b_signal(pipe_ctx)) {
-			fpga_dp_hpo_enable_link_and_stream(state, pipe_ctx);
-		}
-#endif
-		if (dc_is_dp_signal(pipe_ctx->stream->signal) ||
-				dc_is_virtual_signal(pipe_ctx->stream->signal))
-			dp_set_dsc_enable(pipe_ctx, true);
-
-	}
-
-	if (pipe_ctx->stream->signal == SIGNAL_TYPE_HDMI_TYPE_A) {
-		core_link_set_avmute(pipe_ctx, false);
-	}
-}
-
-void core_link_disable_stream(struct pipe_ctx *pipe_ctx)
-{
-	struct dc  *dc = pipe_ctx->stream->ctx->dc;
-	struct dc_stream_state *stream = pipe_ctx->stream;
-	struct dc_link *link = stream->sink->link;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	struct vpg *vpg = pipe_ctx->stream_res.stream_enc->vpg;
-
-	if (is_dp_128b_132b_signal(pipe_ctx))
-		vpg = pipe_ctx->stream_res.hpo_dp_stream_enc->vpg;
-#endif
-
-	if (!IS_DIAG_DC(dc->ctx->dce_environment) &&
-			dc_is_virtual_signal(pipe_ctx->stream->signal))
-		return;
-
-	if (!pipe_ctx->stream->sink->edid_caps.panel_patch.skip_avmute) {
-		if (dc_is_hdmi_signal(pipe_ctx->stream->signal))
-			core_link_set_avmute(pipe_ctx, true);
-	}
-
-	dc->hwss.disable_audio_stream(pipe_ctx);
-
-#if defined(CONFIG_DRM_AMD_DC_HDCP)
-	update_psp_stream_config(pipe_ctx, true);
-#endif
-	dc->hwss.blank_stream(pipe_ctx);
-
-	if (pipe_ctx->stream->signal == SIGNAL_TYPE_DISPLAY_PORT_MST)
-		deallocate_mst_payload(pipe_ctx);
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	else if (pipe_ctx->stream->signal == SIGNAL_TYPE_DISPLAY_PORT &&
-			is_dp_128b_132b_signal(pipe_ctx))
-		dc_link_update_sst_payload(pipe_ctx, false);
-#endif
-
-	if (dc_is_hdmi_signal(pipe_ctx->stream->signal)) {
-		struct ext_hdmi_settings settings = {0};
-		enum engine_id eng_id = pipe_ctx->stream_res.stream_enc->id;
-
-		unsigned short masked_chip_caps = link->chip_caps &
-				EXT_DISPLAY_PATH_CAPS__EXT_CHIP_MASK;
-		//Need to inform that sink is going to use legacy HDMI mode.
-		dal_ddc_service_write_scdc_data(
-			link->ddc,
-			165000,//vbios only handles 165Mhz.
-			false);
-		if (masked_chip_caps == EXT_DISPLAY_PATH_CAPS__HDMI20_TISN65DP159RSBT) {
-			/* DP159, Retimer settings */
-			if (get_ext_hdmi_settings(pipe_ctx, eng_id, &settings))
-				write_i2c_retimer_setting(pipe_ctx,
-						false, false, &settings);
-			else
-				write_i2c_default_retimer_setting(pipe_ctx,
-						false, false);
-		} else if (masked_chip_caps == EXT_DISPLAY_PATH_CAPS__HDMI20_PI3EQX1204) {
-			/* PI3EQX1204, Redriver settings */
-			write_i2c_redriver_setting(pipe_ctx, false);
-		}
-	}
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	if (pipe_ctx->stream->signal == SIGNAL_TYPE_DISPLAY_PORT &&
-			!is_dp_128b_132b_signal(pipe_ctx)) {
-
-		/* In DP1.x SST mode, our encoder will go to TPS1
-		 * when link is on but stream is off.
-		 * Disabling link before stream will avoid exposing TPS1 pattern
-		 * during the disable sequence as it will confuse some receivers
-		 * state machine.
-		 * In DP2 or MST mode, our encoder will stay video active
-		 */
-		disable_link(pipe_ctx->stream->link, pipe_ctx->stream->signal);
-		dc->hwss.disable_stream(pipe_ctx);
-	} else {
-		dc->hwss.disable_stream(pipe_ctx);
-		disable_link(pipe_ctx->stream->link, pipe_ctx->stream->signal);
-	}
-#else
-	disable_link(pipe_ctx->stream->link, pipe_ctx->stream->signal);
-
-	dc->hwss.disable_stream(pipe_ctx);
-#endif
-
-	if (pipe_ctx->stream->timing.flags.DSC) {
-		if (dc_is_dp_signal(pipe_ctx->stream->signal))
-			dp_set_dsc_enable(pipe_ctx, false);
-	}
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	if (is_dp_128b_132b_signal(pipe_ctx)) {
-		if (pipe_ctx->stream_res.tg->funcs->set_out_mux)
-			pipe_ctx->stream_res.tg->funcs->set_out_mux(pipe_ctx->stream_res.tg, OUT_MUX_DIO);
-	}
-#endif
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	if (vpg && vpg->funcs->vpg_powerdown)
-		vpg->funcs->vpg_powerdown(vpg);
-#endif
-}
-
-void core_link_set_avmute(struct pipe_ctx *pipe_ctx, bool enable)
-{
-	struct dc  *dc = pipe_ctx->stream->ctx->dc;
-
-	if (!dc_is_hdmi_signal(pipe_ctx->stream->signal))
-		return;
-
-	dc->hwss.set_avmute(pipe_ctx, enable);
-}
-
-/**
- *  dc_link_enable_hpd_filter:
- *     If enable is true, programs HPD filter on associated HPD line using
- *     delay_on_disconnect/delay_on_connect values dependent on
- *     link->connector_signal
- *
- *     If enable is false, programs HPD filter on associated HPD line with no
- *     delays on connect or disconnect
- *
- *  @link:   pointer to the dc link
- *  @enable: boolean specifying whether to enable hbd
- */
-void dc_link_enable_hpd_filter(struct dc_link *link, bool enable)
-{
-	struct gpio *hpd;
-
-	if (enable) {
-		link->is_hpd_filter_disabled = false;
-		program_hpd_filter(link);
-	} else {
-		link->is_hpd_filter_disabled = true;
-		/* Obtain HPD handle */
-		hpd = get_hpd_gpio(link->ctx->dc_bios, link->link_id, link->ctx->gpio_service);
-
-		if (!hpd)
-			return;
-
-		/* Setup HPD filtering */
-		if (dal_gpio_open(hpd, GPIO_MODE_INTERRUPT) == GPIO_RESULT_OK) {
-			struct gpio_hpd_config config;
-
-			config.delay_on_connect = 0;
-			config.delay_on_disconnect = 0;
-
-			dal_irq_setup_hpd_filter(hpd, &config);
-
-			dal_gpio_close(hpd);
-		} else {
-			ASSERT_CRITICAL(false);
-		}
-		/* Release HPD handle */
-		dal_gpio_destroy_irq(&hpd);
-	}
-}
-
-void dc_link_set_drive_settings(struct dc *dc,
-				struct link_training_settings *lt_settings,
-				const struct dc_link *link)
-{
-
-	int i;
-
-	for (i = 0; i < dc->link_count; i++) {
-		if (dc->links[i] == link)
-			break;
-	}
-
-	if (i >= dc->link_count)
-		ASSERT_CRITICAL(false);
-
-	dc_link_dp_set_drive_settings(dc->links[i], lt_settings);
-}
-
-void dc_link_set_preferred_link_settings(struct dc *dc,
-					 struct dc_link_settings *link_setting,
-					 struct dc_link *link)
-{
-	int i;
-	struct pipe_ctx *pipe;
-	struct dc_stream_state *link_stream;
-	struct dc_link_settings store_settings = *link_setting;
-
-	link->preferred_link_setting = store_settings;
-
-	/* Retrain with preferred link settings only relevant for
-	 * DP signal type
-	 * Check for non-DP signal or if passive dongle present
-	 */
-	if (!dc_is_dp_signal(link->connector_signal) ||
-		link->dongle_max_pix_clk > 0)
-		return;
-
-	for (i = 0; i < MAX_PIPES; i++) {
-		pipe = &dc->current_state->res_ctx.pipe_ctx[i];
-		if (pipe->stream && pipe->stream->link) {
-			if (pipe->stream->link == link) {
-				link_stream = pipe->stream;
-				break;
-			}
-		}
-	}
-
-	/* Stream not found */
-	if (i == MAX_PIPES)
-		return;
-
-	/* Cannot retrain link if backend is off */
-	if (link_stream->dpms_off)
-		return;
-
-	decide_link_settings(link_stream, &store_settings);
-
-	if ((store_settings.lane_count != LANE_COUNT_UNKNOWN) &&
-		(store_settings.link_rate != LINK_RATE_UNKNOWN))
-		dp_retrain_link_dp_test(link, &store_settings, false);
-}
-
-void dc_link_set_preferred_training_settings(struct dc *dc,
-						 struct dc_link_settings *link_setting,
-						 struct dc_link_training_overrides *lt_overrides,
-						 struct dc_link *link,
-						 bool skip_immediate_retrain)
-{
-	if (lt_overrides != NULL)
-		link->preferred_training_settings = *lt_overrides;
-	else
-		memset(&link->preferred_training_settings, 0, sizeof(link->preferred_training_settings));
-
-	if (link_setting != NULL) {
-		link->preferred_link_setting = *link_setting;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-		if (dp_get_link_encoding_format(link_setting) ==
-				DP_128b_132b_ENCODING && !link->hpo_dp_link_enc) {
-			if (!add_dp_hpo_link_encoder_to_link(link))
-				memset(&link->preferred_link_setting, 0, sizeof(link->preferred_link_setting));
-		}
-#endif
-	} else {
-		link->preferred_link_setting.lane_count = LANE_COUNT_UNKNOWN;
-		link->preferred_link_setting.link_rate = LINK_RATE_UNKNOWN;
-	}
-
-	/* Retrain now, or wait until next stream update to apply */
-	if (skip_immediate_retrain == false)
-		dc_link_set_preferred_link_settings(dc, &link->preferred_link_setting, link);
-}
-
-void dc_link_enable_hpd(const struct dc_link *link)
-{
-	dc_link_dp_enable_hpd(link);
-}
-
-void dc_link_disable_hpd(const struct dc_link *link)
-{
-	dc_link_dp_disable_hpd(link);
-}
-
-void dc_link_set_test_pattern(struct dc_link *link,
-			      enum dp_test_pattern test_pattern,
-			      enum dp_test_pattern_color_space test_pattern_color_space,
-			      const struct link_training_settings *p_link_settings,
-			      const unsigned char *p_custom_pattern,
-			      unsigned int cust_pattern_size)
-{
-	if (link != NULL)
-		dc_link_dp_set_test_pattern(
-			link,
-			test_pattern,
-			test_pattern_color_space,
-			p_link_settings,
-			p_custom_pattern,
-			cust_pattern_size);
-}
-
-uint32_t dc_link_bandwidth_kbps(
-	const struct dc_link *link,
-	const struct dc_link_settings *link_setting)
-{
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	uint32_t total_data_bw_efficiency_x10000 = 0;
-	uint32_t link_rate_per_lane_kbps = 0;
-
-	switch (dp_get_link_encoding_format(link_setting)) {
-	case DP_8b_10b_ENCODING:
-		/* For 8b/10b encoding:
-		 * link rate is defined in the unit of LINK_RATE_REF_FREQ_IN_KHZ per DP byte per lane.
-		 * data bandwidth efficiency is 80% with additional 3% overhead if FEC is supported.
-		 */
-		link_rate_per_lane_kbps = link_setting->link_rate * LINK_RATE_REF_FREQ_IN_KHZ * BITS_PER_DP_BYTE;
-		total_data_bw_efficiency_x10000 = DATA_EFFICIENCY_8b_10b_x10000;
-		if (dc_link_should_enable_fec(link)) {
-			total_data_bw_efficiency_x10000 /= 100;
-			total_data_bw_efficiency_x10000 *= DATA_EFFICIENCY_8b_10b_FEC_EFFICIENCY_x100;
-		}
-		break;
-	case DP_128b_132b_ENCODING:
-		/* For 128b/132b encoding:
-		 * link rate is defined in the unit of 10mbps per lane.
-		 * total data bandwidth efficiency is always 96.71%.
-		 */
-		link_rate_per_lane_kbps = link_setting->link_rate * 10000;
-		total_data_bw_efficiency_x10000 = DATA_EFFICIENCY_128b_132b_x10000;
-		break;
-	default:
-		break;
-	}
-
-	/* overall effective link bandwidth = link rate per lane * lane count * total data bandwidth efficiency */
-	return link_rate_per_lane_kbps * link_setting->lane_count / 10000 * total_data_bw_efficiency_x10000;
-#else
-	uint32_t link_bw_kbps =
-		link_setting->link_rate * LINK_RATE_REF_FREQ_IN_KHZ; /* bytes per sec */
-
-	link_bw_kbps *= 8;   /* 8 bits per byte*/
-	link_bw_kbps *= link_setting->lane_count;
-
-	if (dc_link_should_enable_fec(link)) {
-		/* Account for FEC overhead.
-		 * We have to do it based on caps,
-		 * and not based on FEC being set ready,
-		 * because FEC is set ready too late in
-		 * the process to correctly be picked up
-		 * by mode enumeration.
-		 *
-		 * There's enough zeros at the end of 'kbps'
-		 * that make the below operation 100% precise
-		 * for our purposes.
-		 * 'long long' makes it work even for HDMI 2.1
-		 * max bandwidth (and much, much bigger bandwidths
-		 * than that, actually).
-		 *
-		 * NOTE: Reducing link BW by 3% may not be precise
-		 * because it may be a stream BT that increases by 3%, and so
-		 * 1/1.03 = 0.970873 factor should have been used instead,
-		 * but the difference is minimal and is in a safe direction,
-		 * which all works well around potential ambiguity of DP 1.4a spec.
-		 */
-		long long fec_link_bw_kbps = link_bw_kbps * 970LL;
-		link_bw_kbps = (uint32_t)(div64_s64(fec_link_bw_kbps, 1000LL));
-	}
-	return link_bw_kbps;
-
-#endif
-}
-
-const struct dc_link_settings *dc_link_get_link_cap(
-		const struct dc_link *link)
-{
-	if (link->preferred_link_setting.lane_count != LANE_COUNT_UNKNOWN &&
-			link->preferred_link_setting.link_rate != LINK_RATE_UNKNOWN)
-		return &link->preferred_link_setting;
-	return &link->verified_link_cap;
-}
-
-void dc_link_overwrite_extended_receiver_cap(
-		struct dc_link *link)
-{
-	dp_overwrite_extended_receiver_cap(link);
-}
-
-bool dc_link_is_fec_supported(const struct dc_link *link)
-{
-	struct link_encoder *link_enc = NULL;
-
-	/* Links supporting dynamically assigned link encoder will be assigned next
-	 * available encoder if one not already assigned.
-	 */
-	if (link->is_dig_mapping_flexible &&
-			link->dc->res_pool->funcs->link_encs_assign) {
-		link_enc = link_enc_cfg_get_link_enc_used_by_link(link->ctx->dc, link);
-		if (link_enc == NULL)
-			link_enc = link_enc_cfg_get_next_avail_link_enc(link->ctx->dc);
-	} else
-		link_enc = link->link_enc;
-	ASSERT(link_enc);
-
-	return (dc_is_dp_signal(link->connector_signal) && link_enc &&
-			link_enc->features.fec_supported &&
-			link->dpcd_caps.fec_cap.bits.FEC_CAPABLE &&
-			!IS_FPGA_MAXIMUS_DC(link->ctx->dce_environment));
-}
-
-bool dc_link_should_enable_fec(const struct dc_link *link)
-{
-	bool is_fec_disable = false;
-	bool ret = false;
-
-	if ((link->connector_signal != SIGNAL_TYPE_DISPLAY_PORT_MST &&
-			link->local_sink &&
-			link->local_sink->edid_caps.panel_patch.disable_fec) ||
-			(link->connector_signal == SIGNAL_TYPE_EDP
-				))
-		is_fec_disable = true;
-
-	if (dc_link_is_fec_supported(link) && !link->dc->debug.disable_fec && !is_fec_disable)
-		ret = true;
-
-	return ret;
-}
-
-uint32_t dc_bandwidth_in_kbps_from_timing(
-		const struct dc_crtc_timing *timing)
-{
-	uint32_t bits_per_channel = 0;
-	uint32_t kbps;
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	if (timing->flags.DSC)
-		return dc_dsc_stream_bandwidth_in_kbps(timing,
-				timing->dsc_cfg.bits_per_pixel,
-				timing->dsc_cfg.num_slices_h,
-				timing->dsc_cfg.is_dp);
-#endif /* CONFIG_DRM_AMD_DC_DCN */
-
-	switch (timing->display_color_depth) {
-	case COLOR_DEPTH_666:
-		bits_per_channel = 6;
-		break;
-	case COLOR_DEPTH_888:
-		bits_per_channel = 8;
-		break;
-	case COLOR_DEPTH_101010:
-		bits_per_channel = 10;
-		break;
-	case COLOR_DEPTH_121212:
-		bits_per_channel = 12;
-		break;
-	case COLOR_DEPTH_141414:
-		bits_per_channel = 14;
-		break;
-	case COLOR_DEPTH_161616:
-		bits_per_channel = 16;
-		break;
-	default:
-		ASSERT(bits_per_channel != 0);
-		bits_per_channel = 8;
-		break;
-	}
-
-	kbps = timing->pix_clk_100hz / 10;
-	kbps *= bits_per_channel;
-
-	if (timing->flags.Y_ONLY != 1) {
-		/*Only YOnly make reduce bandwidth by 1/3 compares to RGB*/
-		kbps *= 3;
-		if (timing->pixel_encoding == PIXEL_ENCODING_YCBCR420)
-			kbps /= 2;
-		else if (timing->pixel_encoding == PIXEL_ENCODING_YCBCR422)
-			kbps = kbps * 2 / 3;
-	}
-
-	return kbps;
-
-}
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_link_ddc.c b/drivers/gpu/drm/amd/display/dc/core/dc_link_ddc.c
deleted file mode 100644
index 60539b1f2a80..000000000000
--- a/drivers/gpu/drm/amd/display/dc/core/dc_link_ddc.c
+++ /dev/null
@@ -1,780 +0,0 @@
-/*
- * Copyright 2012-15 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: AMD
- *
- */
-
-#include <linux/slab.h>
-
-#include "dm_services.h"
-#include "dm_helpers.h"
-#include "gpio_service_interface.h"
-#include "include/ddc_service_types.h"
-#include "include/grph_object_id.h"
-#include "include/dpcd_defs.h"
-#include "include/logger_interface.h"
-#include "include/vector.h"
-#include "core_types.h"
-#include "dc_link_ddc.h"
-#include "dce/dce_aux.h"
-#include "dmub/inc/dmub_cmd.h"
-
-#define DC_LOGGER_INIT(logger)
-
-static const uint8_t DP_VGA_DONGLE_BRANCH_DEV_NAME[] = "DpVga";
-/* DP to Dual link DVI converter */
-static const uint8_t DP_DVI_CONVERTER_ID_4[] = "m2DVIa";
-static const uint8_t DP_DVI_CONVERTER_ID_5[] = "3393N2";
-
-#define AUX_POWER_UP_WA_DELAY 500
-#define I2C_OVER_AUX_DEFER_WA_DELAY 70
-#define DPVGA_DONGLE_AUX_DEFER_WA_DELAY 40
-#define I2C_OVER_AUX_DEFER_WA_DELAY_1MS 1
-
-/* CV smart dongle slave address for retrieving supported HDTV modes*/
-#define CV_SMART_DONGLE_ADDRESS 0x20
-/* DVI-HDMI dongle slave address for retrieving dongle signature*/
-#define DVI_HDMI_DONGLE_ADDRESS 0x68
-struct dvi_hdmi_dongle_signature_data {
-	int8_t vendor[3];/* "AMD" */
-	uint8_t version[2];
-	uint8_t size;
-	int8_t id[11];/* "6140063500G"*/
-};
-/* DP-HDMI dongle slave address for retrieving dongle signature*/
-#define DP_HDMI_DONGLE_ADDRESS 0x40
-static const uint8_t dp_hdmi_dongle_signature_str[] = "DP-HDMI ADAPTOR";
-#define DP_HDMI_DONGLE_SIGNATURE_EOT 0x04
-
-struct dp_hdmi_dongle_signature_data {
-	int8_t id[15];/* "DP-HDMI ADAPTOR"*/
-	uint8_t eot;/* end of transmition '\x4' */
-};
-
-/* SCDC Address defines (HDMI 2.0)*/
-#define HDMI_SCDC_WRITE_UPDATE_0_ARRAY 3
-#define HDMI_SCDC_ADDRESS  0x54
-#define HDMI_SCDC_SINK_VERSION 0x01
-#define HDMI_SCDC_SOURCE_VERSION 0x02
-#define HDMI_SCDC_UPDATE_0 0x10
-#define HDMI_SCDC_TMDS_CONFIG 0x20
-#define HDMI_SCDC_SCRAMBLER_STATUS 0x21
-#define HDMI_SCDC_CONFIG_0 0x30
-#define HDMI_SCDC_STATUS_FLAGS 0x40
-#define HDMI_SCDC_ERR_DETECT 0x50
-#define HDMI_SCDC_TEST_CONFIG 0xC0
-
-union hdmi_scdc_update_read_data {
-	uint8_t byte[2];
-	struct {
-		uint8_t STATUS_UPDATE:1;
-		uint8_t CED_UPDATE:1;
-		uint8_t RR_TEST:1;
-		uint8_t RESERVED:5;
-		uint8_t RESERVED2:8;
-	} fields;
-};
-
-union hdmi_scdc_status_flags_data {
-	uint8_t byte[2];
-	struct {
-		uint8_t CLOCK_DETECTED:1;
-		uint8_t CH0_LOCKED:1;
-		uint8_t CH1_LOCKED:1;
-		uint8_t CH2_LOCKED:1;
-		uint8_t RESERVED:4;
-		uint8_t RESERVED2:8;
-		uint8_t RESERVED3:8;
-
-	} fields;
-};
-
-union hdmi_scdc_ced_data {
-	uint8_t byte[7];
-	struct {
-		uint8_t CH0_8LOW:8;
-		uint8_t CH0_7HIGH:7;
-		uint8_t CH0_VALID:1;
-		uint8_t CH1_8LOW:8;
-		uint8_t CH1_7HIGH:7;
-		uint8_t CH1_VALID:1;
-		uint8_t CH2_8LOW:8;
-		uint8_t CH2_7HIGH:7;
-		uint8_t CH2_VALID:1;
-		uint8_t CHECKSUM:8;
-		uint8_t RESERVED:8;
-		uint8_t RESERVED2:8;
-		uint8_t RESERVED3:8;
-		uint8_t RESERVED4:4;
-	} fields;
-};
-
-struct i2c_payloads {
-	struct vector payloads;
-};
-
-struct aux_payloads {
-	struct vector payloads;
-};
-
-static bool dal_ddc_i2c_payloads_create(
-		struct dc_context *ctx,
-		struct i2c_payloads *payloads,
-		uint32_t count)
-{
-	if (dal_vector_construct(
-		&payloads->payloads, ctx, count, sizeof(struct i2c_payload)))
-		return true;
-
-	return false;
-}
-
-static struct i2c_payload *dal_ddc_i2c_payloads_get(struct i2c_payloads *p)
-{
-	return (struct i2c_payload *)p->payloads.container;
-}
-
-static uint32_t dal_ddc_i2c_payloads_get_count(struct i2c_payloads *p)
-{
-	return p->payloads.count;
-}
-
-#define DDC_MIN(a, b) (((a) < (b)) ? (a) : (b))
-
-void dal_ddc_i2c_payloads_add(
-	struct i2c_payloads *payloads,
-	uint32_t address,
-	uint32_t len,
-	uint8_t *data,
-	bool write)
-{
-	uint32_t payload_size = EDID_SEGMENT_SIZE;
-	uint32_t pos;
-
-	for (pos = 0; pos < len; pos += payload_size) {
-		struct i2c_payload payload = {
-			.write = write,
-			.address = address,
-			.length = DDC_MIN(payload_size, len - pos),
-			.data = data + pos };
-		dal_vector_append(&payloads->payloads, &payload);
-	}
-
-}
-
-static void ddc_service_construct(
-	struct ddc_service *ddc_service,
-	struct ddc_service_init_data *init_data)
-{
-	enum connector_id connector_id =
-		dal_graphics_object_id_get_connector_id(init_data->id);
-
-	struct gpio_service *gpio_service = init_data->ctx->gpio_service;
-	struct graphics_object_i2c_info i2c_info;
-	struct gpio_ddc_hw_info hw_info;
-	struct dc_bios *dcb = init_data->ctx->dc_bios;
-
-	ddc_service->link = init_data->link;
-	ddc_service->ctx = init_data->ctx;
-
-	if (init_data->is_dpia_link ||
-	    dcb->funcs->get_i2c_info(dcb, init_data->id, &i2c_info) != BP_RESULT_OK) {
-		ddc_service->ddc_pin = NULL;
-	} else {
-		DC_LOGGER_INIT(ddc_service->ctx->logger);
-		DC_LOG_DC("BIOS object table - i2c_line: %d", i2c_info.i2c_line);
-		DC_LOG_DC("BIOS object table - i2c_engine_id: %d", i2c_info.i2c_engine_id);
-
-		hw_info.ddc_channel = i2c_info.i2c_line;
-		if (ddc_service->link != NULL)
-			hw_info.hw_supported = i2c_info.i2c_hw_assist;
-		else
-			hw_info.hw_supported = false;
-
-		ddc_service->ddc_pin = dal_gpio_create_ddc(
-			gpio_service,
-			i2c_info.gpio_info.clk_a_register_index,
-			1 << i2c_info.gpio_info.clk_a_shift,
-			&hw_info);
-	}
-
-	ddc_service->flags.EDID_QUERY_DONE_ONCE = false;
-	ddc_service->flags.FORCE_READ_REPEATED_START = false;
-	ddc_service->flags.EDID_STRESS_READ = false;
-
-	ddc_service->flags.IS_INTERNAL_DISPLAY =
-		connector_id == CONNECTOR_ID_EDP ||
-		connector_id == CONNECTOR_ID_LVDS;
-
-	ddc_service->wa.raw = 0;
-}
-
-struct ddc_service *dal_ddc_service_create(
-	struct ddc_service_init_data *init_data)
-{
-	struct ddc_service *ddc_service;
-
-	ddc_service = kzalloc(sizeof(struct ddc_service), GFP_KERNEL);
-
-	if (!ddc_service)
-		return NULL;
-
-	ddc_service_construct(ddc_service, init_data);
-	return ddc_service;
-}
-
-static void ddc_service_destruct(struct ddc_service *ddc)
-{
-	if (ddc->ddc_pin)
-		dal_gpio_destroy_ddc(&ddc->ddc_pin);
-}
-
-void dal_ddc_service_destroy(struct ddc_service **ddc)
-{
-	if (!ddc || !*ddc) {
-		BREAK_TO_DEBUGGER();
-		return;
-	}
-	ddc_service_destruct(*ddc);
-	kfree(*ddc);
-	*ddc = NULL;
-}
-
-enum ddc_service_type dal_ddc_service_get_type(struct ddc_service *ddc)
-{
-	return DDC_SERVICE_TYPE_CONNECTOR;
-}
-
-void dal_ddc_service_set_transaction_type(
-	struct ddc_service *ddc,
-	enum ddc_transaction_type type)
-{
-	ddc->transaction_type = type;
-}
-
-bool dal_ddc_service_is_in_aux_transaction_mode(struct ddc_service *ddc)
-{
-	switch (ddc->transaction_type) {
-	case DDC_TRANSACTION_TYPE_I2C_OVER_AUX:
-	case DDC_TRANSACTION_TYPE_I2C_OVER_AUX_WITH_DEFER:
-	case DDC_TRANSACTION_TYPE_I2C_OVER_AUX_RETRY_DEFER:
-		return true;
-	default:
-		break;
-	}
-	return false;
-}
-
-void ddc_service_set_dongle_type(struct ddc_service *ddc,
-		enum display_dongle_type dongle_type)
-{
-	ddc->dongle_type = dongle_type;
-}
-
-static uint32_t defer_delay_converter_wa(
-	struct ddc_service *ddc,
-	uint32_t defer_delay)
-{
-	struct dc_link *link = ddc->link;
-
-	if (link->dpcd_caps.dongle_type == DISPLAY_DONGLE_DP_VGA_CONVERTER &&
-		link->dpcd_caps.branch_dev_id == DP_BRANCH_DEVICE_ID_0080E1 &&
-		!memcmp(link->dpcd_caps.branch_dev_name,
-		    DP_VGA_DONGLE_BRANCH_DEV_NAME,
-			sizeof(link->dpcd_caps.branch_dev_name)))
-
-		return defer_delay > DPVGA_DONGLE_AUX_DEFER_WA_DELAY ?
-			defer_delay : DPVGA_DONGLE_AUX_DEFER_WA_DELAY;
-
-	if (link->dpcd_caps.branch_dev_id == DP_BRANCH_DEVICE_ID_0080E1 &&
-	    !memcmp(link->dpcd_caps.branch_dev_name,
-		    DP_DVI_CONVERTER_ID_4,
-		    sizeof(link->dpcd_caps.branch_dev_name)))
-		return defer_delay > I2C_OVER_AUX_DEFER_WA_DELAY ?
-			defer_delay : I2C_OVER_AUX_DEFER_WA_DELAY;
-	if (link->dpcd_caps.branch_dev_id == DP_BRANCH_DEVICE_ID_006037 &&
-	    !memcmp(link->dpcd_caps.branch_dev_name,
-		    DP_DVI_CONVERTER_ID_5,
-		    sizeof(link->dpcd_caps.branch_dev_name)))
-		return defer_delay > I2C_OVER_AUX_DEFER_WA_DELAY_1MS ?
-			I2C_OVER_AUX_DEFER_WA_DELAY_1MS : defer_delay;
-
-	return defer_delay;
-}
-
-#define DP_TRANSLATOR_DELAY 5
-
-uint32_t get_defer_delay(struct ddc_service *ddc)
-{
-	uint32_t defer_delay = 0;
-
-	switch (ddc->transaction_type) {
-	case DDC_TRANSACTION_TYPE_I2C_OVER_AUX:
-		if ((DISPLAY_DONGLE_DP_VGA_CONVERTER == ddc->dongle_type) ||
-			(DISPLAY_DONGLE_DP_DVI_CONVERTER == ddc->dongle_type) ||
-			(DISPLAY_DONGLE_DP_HDMI_CONVERTER ==
-				ddc->dongle_type)) {
-
-			defer_delay = DP_TRANSLATOR_DELAY;
-
-			defer_delay =
-				defer_delay_converter_wa(ddc, defer_delay);
-
-		} else /*sink has a delay different from an Active Converter*/
-			defer_delay = 0;
-		break;
-	case DDC_TRANSACTION_TYPE_I2C_OVER_AUX_WITH_DEFER:
-		defer_delay = DP_TRANSLATOR_DELAY;
-		break;
-	default:
-		break;
-	}
-	return defer_delay;
-}
-
-static bool i2c_read(
-	struct ddc_service *ddc,
-	uint32_t address,
-	uint8_t *buffer,
-	uint32_t len)
-{
-	uint8_t offs_data = 0;
-	struct i2c_payload payloads[2] = {
-		{
-		.write = true,
-		.address = address,
-		.length = 1,
-		.data = &offs_data },
-		{
-		.write = false,
-		.address = address,
-		.length = len,
-		.data = buffer } };
-
-	struct i2c_command command = {
-		.payloads = payloads,
-		.number_of_payloads = 2,
-		.engine = DDC_I2C_COMMAND_ENGINE,
-		.speed = ddc->ctx->dc->caps.i2c_speed_in_khz };
-
-	return dm_helpers_submit_i2c(
-			ddc->ctx,
-			ddc->link,
-			&command);
-}
-
-void dal_ddc_service_i2c_query_dp_dual_mode_adaptor(
-	struct ddc_service *ddc,
-	struct display_sink_capability *sink_cap)
-{
-	uint8_t i;
-	bool is_valid_hdmi_signature;
-	enum display_dongle_type *dongle = &sink_cap->dongle_type;
-	uint8_t type2_dongle_buf[DP_ADAPTOR_TYPE2_SIZE];
-	bool is_type2_dongle = false;
-	int retry_count = 2;
-	struct dp_hdmi_dongle_signature_data *dongle_signature;
-
-	/* Assume we have no valid DP passive dongle connected */
-	*dongle = DISPLAY_DONGLE_NONE;
-	sink_cap->max_hdmi_pixel_clock = DP_ADAPTOR_HDMI_SAFE_MAX_TMDS_CLK;
-
-	/* Read DP-HDMI dongle I2c (no response interpreted as DP-DVI dongle)*/
-	if (!i2c_read(
-		ddc,
-		DP_HDMI_DONGLE_ADDRESS,
-		type2_dongle_buf,
-		sizeof(type2_dongle_buf))) {
-		/* Passive HDMI dongles can sometimes fail here without retrying*/
-		while (retry_count > 0) {
-			if (i2c_read(ddc,
-				DP_HDMI_DONGLE_ADDRESS,
-				type2_dongle_buf,
-				sizeof(type2_dongle_buf)))
-				break;
-			retry_count--;
-		}
-		if (retry_count == 0) {
-			*dongle = DISPLAY_DONGLE_DP_DVI_DONGLE;
-			sink_cap->max_hdmi_pixel_clock = DP_ADAPTOR_DVI_MAX_TMDS_CLK;
-
-			CONN_DATA_DETECT(ddc->link, type2_dongle_buf, sizeof(type2_dongle_buf),
-					"DP-DVI passive dongle %dMhz: ",
-					DP_ADAPTOR_DVI_MAX_TMDS_CLK / 1000);
-			return;
-		}
-	}
-
-	/* Check if Type 2 dongle.*/
-	if (type2_dongle_buf[DP_ADAPTOR_TYPE2_REG_ID] == DP_ADAPTOR_TYPE2_ID)
-		is_type2_dongle = true;
-
-	dongle_signature =
-		(struct dp_hdmi_dongle_signature_data *)type2_dongle_buf;
-
-	is_valid_hdmi_signature = true;
-
-	/* Check EOT */
-	if (dongle_signature->eot != DP_HDMI_DONGLE_SIGNATURE_EOT) {
-		is_valid_hdmi_signature = false;
-	}
-
-	/* Check signature */
-	for (i = 0; i < sizeof(dongle_signature->id); ++i) {
-		/* If its not the right signature,
-		 * skip mismatch in subversion byte.*/
-		if (dongle_signature->id[i] !=
-			dp_hdmi_dongle_signature_str[i] && i != 3) {
-
-			if (is_type2_dongle) {
-				is_valid_hdmi_signature = false;
-				break;
-			}
-
-		}
-	}
-
-	if (is_type2_dongle) {
-		uint32_t max_tmds_clk =
-			type2_dongle_buf[DP_ADAPTOR_TYPE2_REG_MAX_TMDS_CLK];
-
-		max_tmds_clk = max_tmds_clk * 2 + max_tmds_clk / 2;
-
-		if (0 == max_tmds_clk ||
-				max_tmds_clk < DP_ADAPTOR_TYPE2_MIN_TMDS_CLK ||
-				max_tmds_clk > DP_ADAPTOR_TYPE2_MAX_TMDS_CLK) {
-			*dongle = DISPLAY_DONGLE_DP_DVI_DONGLE;
-
-			CONN_DATA_DETECT(ddc->link, type2_dongle_buf,
-					sizeof(type2_dongle_buf),
-					"DP-DVI passive dongle %dMhz: ",
-					DP_ADAPTOR_DVI_MAX_TMDS_CLK / 1000);
-		} else {
-			if (is_valid_hdmi_signature == true) {
-				*dongle = DISPLAY_DONGLE_DP_HDMI_DONGLE;
-
-				CONN_DATA_DETECT(ddc->link, type2_dongle_buf,
-						sizeof(type2_dongle_buf),
-						"Type 2 DP-HDMI passive dongle %dMhz: ",
-						max_tmds_clk);
-			} else {
-				*dongle = DISPLAY_DONGLE_DP_HDMI_MISMATCHED_DONGLE;
-
-				CONN_DATA_DETECT(ddc->link, type2_dongle_buf,
-						sizeof(type2_dongle_buf),
-						"Type 2 DP-HDMI passive dongle (no signature) %dMhz: ",
-						max_tmds_clk);
-
-			}
-
-			/* Multiply by 1000 to convert to kHz. */
-			sink_cap->max_hdmi_pixel_clock =
-				max_tmds_clk * 1000;
-		}
-
-	} else {
-		if (is_valid_hdmi_signature == true) {
-			*dongle = DISPLAY_DONGLE_DP_HDMI_DONGLE;
-
-			CONN_DATA_DETECT(ddc->link, type2_dongle_buf,
-					sizeof(type2_dongle_buf),
-					"Type 1 DP-HDMI passive dongle %dMhz: ",
-					sink_cap->max_hdmi_pixel_clock / 1000);
-		} else {
-			*dongle = DISPLAY_DONGLE_DP_HDMI_MISMATCHED_DONGLE;
-
-			CONN_DATA_DETECT(ddc->link, type2_dongle_buf,
-					sizeof(type2_dongle_buf),
-					"Type 1 DP-HDMI passive dongle (no signature) %dMhz: ",
-					sink_cap->max_hdmi_pixel_clock / 1000);
-		}
-	}
-
-	return;
-}
-
-enum {
-	DP_SINK_CAP_SIZE =
-		DP_EDP_CONFIGURATION_CAP - DP_DPCD_REV + 1
-};
-
-bool dal_ddc_service_query_ddc_data(
-	struct ddc_service *ddc,
-	uint32_t address,
-	uint8_t *write_buf,
-	uint32_t write_size,
-	uint8_t *read_buf,
-	uint32_t read_size)
-{
-	bool success = true;
-	uint32_t payload_size =
-		dal_ddc_service_is_in_aux_transaction_mode(ddc) ?
-			DEFAULT_AUX_MAX_DATA_SIZE : EDID_SEGMENT_SIZE;
-
-	uint32_t write_payloads =
-		(write_size + payload_size - 1) / payload_size;
-
-	uint32_t read_payloads =
-		(read_size + payload_size - 1) / payload_size;
-
-	uint32_t payloads_num = write_payloads + read_payloads;
-
-
-	if (write_size > EDID_SEGMENT_SIZE || read_size > EDID_SEGMENT_SIZE)
-		return false;
-
-	if (!payloads_num)
-		return false;
-
-	/*TODO: len of payload data for i2c and aux is uint8!!!!,
-	 *  but we want to read 256 over i2c!!!!*/
-	if (dal_ddc_service_is_in_aux_transaction_mode(ddc)) {
-		struct aux_payload payload;
-
-		payload.i2c_over_aux = true;
-		payload.address = address;
-		payload.reply = NULL;
-		payload.defer_delay = get_defer_delay(ddc);
-		payload.write_status_update = false;
-
-		if (write_size != 0) {
-			payload.write = true;
-			/* should not set mot (middle of transaction) to 0
-			 * if there are pending read payloads
-			 */
-			payload.mot = !(read_size == 0);
-			payload.length = write_size;
-			payload.data = write_buf;
-
-			success = dal_ddc_submit_aux_command(ddc, &payload);
-		}
-
-		if (read_size != 0 && success) {
-			payload.write = false;
-			/* should set mot (middle of transaction) to 0
-			 * since it is the last payload to send
-			 */
-			payload.mot = false;
-			payload.length = read_size;
-			payload.data = read_buf;
-
-			success = dal_ddc_submit_aux_command(ddc, &payload);
-		}
-	} else {
-		struct i2c_command command = {0};
-		struct i2c_payloads payloads;
-
-		if (!dal_ddc_i2c_payloads_create(ddc->ctx, &payloads, payloads_num))
-			return false;
-
-		command.payloads = dal_ddc_i2c_payloads_get(&payloads);
-		command.number_of_payloads = 0;
-		command.engine = DDC_I2C_COMMAND_ENGINE;
-		command.speed = ddc->ctx->dc->caps.i2c_speed_in_khz;
-
-		dal_ddc_i2c_payloads_add(
-			&payloads, address, write_size, write_buf, true);
-
-		dal_ddc_i2c_payloads_add(
-			&payloads, address, read_size, read_buf, false);
-
-		command.number_of_payloads =
-			dal_ddc_i2c_payloads_get_count(&payloads);
-
-		success = dm_helpers_submit_i2c(
-				ddc->ctx,
-				ddc->link,
-				&command);
-
-		dal_vector_destruct(&payloads.payloads);
-	}
-
-	return success;
-}
-
-bool dal_ddc_submit_aux_command(struct ddc_service *ddc,
-		struct aux_payload *payload)
-{
-	uint32_t retrieved = 0;
-	bool ret = false;
-
-	if (!ddc)
-		return false;
-
-	if (!payload)
-		return false;
-
-	do {
-		struct aux_payload current_payload;
-		bool is_end_of_payload = (retrieved + DEFAULT_AUX_MAX_DATA_SIZE) >=
-				payload->length ? true : false;
-		uint32_t payload_length = is_end_of_payload ?
-				payload->length - retrieved : DEFAULT_AUX_MAX_DATA_SIZE;
-
-		current_payload.address = payload->address;
-		current_payload.data = &payload->data[retrieved];
-		current_payload.defer_delay = payload->defer_delay;
-		current_payload.i2c_over_aux = payload->i2c_over_aux;
-		current_payload.length = payload_length;
-		/* set mot (middle of transaction) to false if it is the last payload */
-		current_payload.mot = is_end_of_payload ? payload->mot:true;
-		current_payload.write_status_update = false;
-		current_payload.reply = payload->reply;
-		current_payload.write = payload->write;
-
-		ret = dc_link_aux_transfer_with_retries(ddc, &current_payload);
-
-		retrieved += payload_length;
-	} while (retrieved < payload->length && ret == true);
-
-	return ret;
-}
-
-/* dc_link_aux_transfer_raw() - Attempt to transfer
- * the given aux payload.  This function does not perform
- * retries or handle error states.  The reply is returned
- * in the payload->reply and the result through
- * *operation_result.  Returns the number of bytes transferred,
- * or -1 on a failure.
- */
-int dc_link_aux_transfer_raw(struct ddc_service *ddc,
-		struct aux_payload *payload,
-		enum aux_return_code_type *operation_result)
-{
-	if (ddc->ctx->dc->debug.enable_dmub_aux_for_legacy_ddc ||
-	    !ddc->ddc_pin) {
-		return dce_aux_transfer_dmub_raw(ddc, payload, operation_result);
-	} else {
-		return dce_aux_transfer_raw(ddc, payload, operation_result);
-	}
-}
-
-/* dc_link_aux_transfer_with_retries() - Attempt to submit an
- * aux payload, retrying on timeouts, defers, and busy states
- * as outlined in the DP spec.  Returns true if the request
- * was successful.
- *
- * Unless you want to implement your own retry semantics, this
- * is probably the one you want.
- */
-bool dc_link_aux_transfer_with_retries(struct ddc_service *ddc,
-		struct aux_payload *payload)
-{
-	return dce_aux_transfer_with_retries(ddc, payload);
-}
-
-
-bool dc_link_aux_try_to_configure_timeout(struct ddc_service *ddc,
-		uint32_t timeout)
-{
-	bool result = false;
-	struct ddc *ddc_pin = ddc->ddc_pin;
-
-	/* Do not try to access nonexistent DDC pin. */
-	if (ddc->link->ep_type != DISPLAY_ENDPOINT_PHY)
-		return true;
-
-	if (ddc->ctx->dc->res_pool->engines[ddc_pin->pin_data->en]->funcs->configure_timeout) {
-		ddc->ctx->dc->res_pool->engines[ddc_pin->pin_data->en]->funcs->configure_timeout(ddc, timeout);
-		result = true;
-	}
-	return result;
-}
-
-/*test only function*/
-void dal_ddc_service_set_ddc_pin(
-	struct ddc_service *ddc_service,
-	struct ddc *ddc)
-{
-	ddc_service->ddc_pin = ddc;
-}
-
-struct ddc *dal_ddc_service_get_ddc_pin(struct ddc_service *ddc_service)
-{
-	return ddc_service->ddc_pin;
-}
-
-void dal_ddc_service_write_scdc_data(struct ddc_service *ddc_service,
-		uint32_t pix_clk,
-		bool lte_340_scramble)
-{
-	bool over_340_mhz = pix_clk > 340000 ? 1 : 0;
-	uint8_t slave_address = HDMI_SCDC_ADDRESS;
-	uint8_t offset = HDMI_SCDC_SINK_VERSION;
-	uint8_t sink_version = 0;
-	uint8_t write_buffer[2] = {0};
-	/*Lower than 340 Scramble bit from SCDC caps*/
-
-	if (ddc_service->link->local_sink &&
-		ddc_service->link->local_sink->edid_caps.panel_patch.skip_scdc_overwrite)
-		return;
-
-	dal_ddc_service_query_ddc_data(ddc_service, slave_address, &offset,
-			sizeof(offset), &sink_version, sizeof(sink_version));
-	if (sink_version == 1) {
-		/*Source Version = 1*/
-		write_buffer[0] = HDMI_SCDC_SOURCE_VERSION;
-		write_buffer[1] = 1;
-		dal_ddc_service_query_ddc_data(ddc_service, slave_address,
-				write_buffer, sizeof(write_buffer), NULL, 0);
-		/*Read Request from SCDC caps*/
-	}
-	write_buffer[0] = HDMI_SCDC_TMDS_CONFIG;
-
-	if (over_340_mhz) {
-		write_buffer[1] = 3;
-	} else if (lte_340_scramble) {
-		write_buffer[1] = 1;
-	} else {
-		write_buffer[1] = 0;
-	}
-	dal_ddc_service_query_ddc_data(ddc_service, slave_address, write_buffer,
-			sizeof(write_buffer), NULL, 0);
-}
-
-void dal_ddc_service_read_scdc_data(struct ddc_service *ddc_service)
-{
-	uint8_t slave_address = HDMI_SCDC_ADDRESS;
-	uint8_t offset = HDMI_SCDC_TMDS_CONFIG;
-	uint8_t tmds_config = 0;
-
-	if (ddc_service->link->local_sink &&
-		ddc_service->link->local_sink->edid_caps.panel_patch.skip_scdc_overwrite)
-		return;
-
-	dal_ddc_service_query_ddc_data(ddc_service, slave_address, &offset,
-			sizeof(offset), &tmds_config, sizeof(tmds_config));
-	if (tmds_config & 0x1) {
-		union hdmi_scdc_status_flags_data status_data = {0};
-		uint8_t scramble_status = 0;
-
-		offset = HDMI_SCDC_SCRAMBLER_STATUS;
-		dal_ddc_service_query_ddc_data(ddc_service, slave_address,
-				&offset, sizeof(offset), &scramble_status,
-				sizeof(scramble_status));
-		offset = HDMI_SCDC_STATUS_FLAGS;
-		dal_ddc_service_query_ddc_data(ddc_service, slave_address,
-				&offset, sizeof(offset), status_data.byte,
-				sizeof(status_data.byte));
-	}
-}
-
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c b/drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c
deleted file mode 100644
index 13bc69d6b679..000000000000
--- a/drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c
+++ /dev/null
@@ -1,6128 +0,0 @@
-/*
- * Copyright 2015 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: AMD
- */
-#include "dm_services.h"
-#include "dc.h"
-#include "dc_link_dp.h"
-#include "dm_helpers.h"
-#include "opp.h"
-#include "dsc.h"
-#include "resource.h"
-
-#include "inc/core_types.h"
-#include "link_hwss.h"
-#include "dc_link_ddc.h"
-#include "core_status.h"
-#include "dpcd_defs.h"
-#include "dc_dmub_srv.h"
-#include "dce/dmub_hw_lock_mgr.h"
-#include "inc/dc_link_dpia.h"
-#include "inc/link_enc_cfg.h"
-
-/*Travis*/
-static const uint8_t DP_VGA_LVDS_CONVERTER_ID_2[] = "sivarT";
-/*Nutmeg*/
-static const uint8_t DP_VGA_LVDS_CONVERTER_ID_3[] = "dnomlA";
-
-#define DC_LOGGER \
-	link->ctx->logger
-#define DC_TRACE_LEVEL_MESSAGE(...) /* do nothing */
-
-#include "link_dpcd.h"
-
-	/* maximum pre emphasis level allowed for each voltage swing level*/
-	static const enum dc_pre_emphasis
-	voltage_swing_to_pre_emphasis[] = { PRE_EMPHASIS_LEVEL3,
-					    PRE_EMPHASIS_LEVEL2,
-					    PRE_EMPHASIS_LEVEL1,
-					    PRE_EMPHASIS_DISABLED };
-
-enum {
-	POST_LT_ADJ_REQ_LIMIT = 6,
-	POST_LT_ADJ_REQ_TIMEOUT = 200
-};
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-struct dp_lt_fallback_entry {
-	enum dc_lane_count lane_count;
-	enum dc_link_rate link_rate;
-};
-
-static const struct dp_lt_fallback_entry dp_lt_fallbacks[] = {
-		/* This link training fallback array is ordered by
-		 * link bandwidth from highest to lowest.
-		 * DP specs makes it a normative policy to always
-		 * choose the next highest link bandwidth during
-		 * link training fallback.
-		 */
-		{LANE_COUNT_FOUR, LINK_RATE_UHBR20},
-		{LANE_COUNT_FOUR, LINK_RATE_UHBR13_5},
-		{LANE_COUNT_TWO, LINK_RATE_UHBR20},
-		{LANE_COUNT_FOUR, LINK_RATE_UHBR10},
-		{LANE_COUNT_TWO, LINK_RATE_UHBR13_5},
-		{LANE_COUNT_FOUR, LINK_RATE_HIGH3},
-		{LANE_COUNT_ONE, LINK_RATE_UHBR20},
-		{LANE_COUNT_TWO, LINK_RATE_UHBR10},
-		{LANE_COUNT_FOUR, LINK_RATE_HIGH2},
-		{LANE_COUNT_ONE, LINK_RATE_UHBR13_5},
-		{LANE_COUNT_TWO, LINK_RATE_HIGH3},
-		{LANE_COUNT_ONE, LINK_RATE_UHBR10},
-		{LANE_COUNT_TWO, LINK_RATE_HIGH2},
-		{LANE_COUNT_FOUR, LINK_RATE_HIGH},
-		{LANE_COUNT_ONE, LINK_RATE_HIGH3},
-		{LANE_COUNT_FOUR, LINK_RATE_LOW},
-		{LANE_COUNT_ONE, LINK_RATE_HIGH2},
-		{LANE_COUNT_TWO, LINK_RATE_HIGH},
-		{LANE_COUNT_TWO, LINK_RATE_LOW},
-		{LANE_COUNT_ONE, LINK_RATE_HIGH},
-		{LANE_COUNT_ONE, LINK_RATE_LOW},
-};
-#endif
-
-static bool decide_fallback_link_setting(
-		struct dc_link_settings initial_link_settings,
-		struct dc_link_settings *current_link_setting,
-		enum link_training_result training_result);
-static struct dc_link_settings get_common_supported_link_settings(
-		struct dc_link_settings link_setting_a,
-		struct dc_link_settings link_setting_b);
-static void maximize_lane_settings(const struct link_training_settings *lt_settings,
-		struct dc_lane_settings lane_settings[LANE_COUNT_DP_MAX]);
-static void override_lane_settings(const struct link_training_settings *lt_settings,
-		struct dc_lane_settings lane_settings[LANE_COUNT_DP_MAX]);
-
-static uint32_t get_cr_training_aux_rd_interval(struct dc_link *link,
-		const struct dc_link_settings *link_settings)
-{
-	union training_aux_rd_interval training_rd_interval;
-	uint32_t wait_in_micro_secs = 100;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	memset(&training_rd_interval, 0, sizeof(training_rd_interval));
-	if (dp_get_link_encoding_format(link_settings) == DP_8b_10b_ENCODING &&
-			link->dpcd_caps.dpcd_rev.raw >= DPCD_REV_12) {
-		core_link_read_dpcd(
-				link,
-				DP_TRAINING_AUX_RD_INTERVAL,
-				(uint8_t *)&training_rd_interval,
-				sizeof(training_rd_interval));
-		if (training_rd_interval.bits.TRAINIG_AUX_RD_INTERVAL)
-			wait_in_micro_secs = training_rd_interval.bits.TRAINIG_AUX_RD_INTERVAL * 4000;
-	}
-#else
-	core_link_read_dpcd(
-			link,
-			DP_TRAINING_AUX_RD_INTERVAL,
-			(uint8_t *)&training_rd_interval,
-			sizeof(training_rd_interval));
-	if (training_rd_interval.bits.TRAINIG_AUX_RD_INTERVAL)
-		wait_in_micro_secs = training_rd_interval.bits.TRAINIG_AUX_RD_INTERVAL * 4000;	
-#endif
-	return wait_in_micro_secs;
-}
-
-static uint32_t get_eq_training_aux_rd_interval(
-	struct dc_link *link,
-	const struct dc_link_settings *link_settings)
-{
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	union training_aux_rd_interval training_rd_interval;
-
-	memset(&training_rd_interval, 0, sizeof(training_rd_interval));
-	if (dp_get_link_encoding_format(link_settings) == DP_128b_132b_ENCODING) {
-		core_link_read_dpcd(
-				link,
-				DP_128b_132b_TRAINING_AUX_RD_INTERVAL,
-				(uint8_t *)&training_rd_interval,
-				sizeof(training_rd_interval));
-	} else if (dp_get_link_encoding_format(link_settings) == DP_8b_10b_ENCODING &&
-			link->dpcd_caps.dpcd_rev.raw >= DPCD_REV_12) {
-		core_link_read_dpcd(
-				link,
-				DP_TRAINING_AUX_RD_INTERVAL,
-				(uint8_t *)&training_rd_interval,
-				sizeof(training_rd_interval));
-	}
-
-	switch (training_rd_interval.bits.TRAINIG_AUX_RD_INTERVAL) {
-	case 0: return 400;
-	case 1: return 4000;
-	case 2: return 8000;
-	case 3: return 12000;
-	case 4: return 16000;
-	case 5: return 32000;
-	case 6: return 64000;
-	default: return 400;
-	}
-#else
-	union training_aux_rd_interval training_rd_interval;
-	uint32_t wait_in_micro_secs = 400;
-
-	memset(&training_rd_interval, 0, sizeof(training_rd_interval));
-	/* overwrite the delay if rev > 1.1*/
-	if (link->dpcd_caps.dpcd_rev.raw >= DPCD_REV_12) {
-		/* DP 1.2 or later - retrieve delay through
-		 * "DPCD_ADDR_TRAINING_AUX_RD_INTERVAL" register */
-		core_link_read_dpcd(
-			link,
-			DP_TRAINING_AUX_RD_INTERVAL,
-			(uint8_t *)&training_rd_interval,
-			sizeof(training_rd_interval));
-
-		if (training_rd_interval.bits.TRAINIG_AUX_RD_INTERVAL)
-			wait_in_micro_secs = training_rd_interval.bits.TRAINIG_AUX_RD_INTERVAL * 4000;
-	}
-
-	return wait_in_micro_secs;
-#endif
-}
-
-void dp_wait_for_training_aux_rd_interval(
-	struct dc_link *link,
-	uint32_t wait_in_micro_secs)
-{
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	if (wait_in_micro_secs > 16000)
-		msleep(wait_in_micro_secs/1000);
-	else
-		udelay(wait_in_micro_secs);
-#else
-	udelay(wait_in_micro_secs);
-#endif
-
-	DC_LOG_HW_LINK_TRAINING("%s:\n wait = %d\n",
-		__func__,
-		wait_in_micro_secs);
-}
-
-enum dpcd_training_patterns
-	dc_dp_training_pattern_to_dpcd_training_pattern(
-	struct dc_link *link,
-	enum dc_dp_training_pattern pattern)
-{
-	enum dpcd_training_patterns dpcd_tr_pattern =
-	DPCD_TRAINING_PATTERN_VIDEOIDLE;
-
-	switch (pattern) {
-	case DP_TRAINING_PATTERN_SEQUENCE_1:
-		dpcd_tr_pattern = DPCD_TRAINING_PATTERN_1;
-		break;
-	case DP_TRAINING_PATTERN_SEQUENCE_2:
-		dpcd_tr_pattern = DPCD_TRAINING_PATTERN_2;
-		break;
-	case DP_TRAINING_PATTERN_SEQUENCE_3:
-		dpcd_tr_pattern = DPCD_TRAINING_PATTERN_3;
-		break;
-	case DP_TRAINING_PATTERN_SEQUENCE_4:
-		dpcd_tr_pattern = DPCD_TRAINING_PATTERN_4;
-		break;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	case DP_128b_132b_TPS1:
-		dpcd_tr_pattern = DPCD_128b_132b_TPS1;
-		break;
-	case DP_128b_132b_TPS2:
-		dpcd_tr_pattern = DPCD_128b_132b_TPS2;
-		break;
-	case DP_128b_132b_TPS2_CDS:
-		dpcd_tr_pattern = DPCD_128b_132b_TPS2_CDS;
-		break;
-#endif
-	case DP_TRAINING_PATTERN_VIDEOIDLE:
-		dpcd_tr_pattern = DPCD_TRAINING_PATTERN_VIDEOIDLE;
-		break;
-	default:
-		ASSERT(0);
-		DC_LOG_HW_LINK_TRAINING("%s: Invalid HW Training pattern: %d\n",
-			__func__, pattern);
-		break;
-	}
-
-	return dpcd_tr_pattern;
-}
-
-static void dpcd_set_training_pattern(
-	struct dc_link *link,
-	enum dc_dp_training_pattern training_pattern)
-{
-	union dpcd_training_pattern dpcd_pattern = {0};
-
-	dpcd_pattern.v1_4.TRAINING_PATTERN_SET =
-			dc_dp_training_pattern_to_dpcd_training_pattern(
-					link, training_pattern);
-
-	core_link_write_dpcd(
-		link,
-		DP_TRAINING_PATTERN_SET,
-		&dpcd_pattern.raw,
-		1);
-
-	DC_LOG_HW_LINK_TRAINING("%s\n %x pattern = %x\n",
-		__func__,
-		DP_TRAINING_PATTERN_SET,
-		dpcd_pattern.v1_4.TRAINING_PATTERN_SET);
-}
-
-static enum dc_dp_training_pattern decide_cr_training_pattern(
-		const struct dc_link_settings *link_settings)
-{
-	switch (dp_get_link_encoding_format(link_settings)) {
-	case DP_8b_10b_ENCODING:
-	default:
-		return DP_TRAINING_PATTERN_SEQUENCE_1;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	case DP_128b_132b_ENCODING:
-		return DP_128b_132b_TPS1;
-#endif
-	}
-}
-
-static enum dc_dp_training_pattern decide_eq_training_pattern(struct dc_link *link,
-		const struct dc_link_settings *link_settings)
-{
-	struct link_encoder *link_enc;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	struct encoder_feature_support *enc_caps;
-	struct dpcd_caps *rx_caps = &link->dpcd_caps;
-	enum dc_dp_training_pattern pattern = DP_TRAINING_PATTERN_SEQUENCE_2;
-
-	/* Access link encoder capability based on whether it is statically
-	 * or dynamically assigned to a link.
-	 */
-	if (link->is_dig_mapping_flexible &&
-			link->dc->res_pool->funcs->link_encs_assign)
-		link_enc = link_enc_cfg_get_link_enc_used_by_link(link->ctx->dc, link);
-	else
-		link_enc = link->link_enc;
-	ASSERT(link_enc);
-	enc_caps = &link_enc->features;
-
-	switch (dp_get_link_encoding_format(link_settings)) {
-	case DP_8b_10b_ENCODING:
-		if (enc_caps->flags.bits.IS_TPS4_CAPABLE &&
-				rx_caps->max_down_spread.bits.TPS4_SUPPORTED)
-			pattern = DP_TRAINING_PATTERN_SEQUENCE_4;
-		else if (enc_caps->flags.bits.IS_TPS3_CAPABLE &&
-				rx_caps->max_ln_count.bits.TPS3_SUPPORTED)
-			pattern = DP_TRAINING_PATTERN_SEQUENCE_3;
-		else
-			pattern = DP_TRAINING_PATTERN_SEQUENCE_2;
-		break;
-	case DP_128b_132b_ENCODING:
-		pattern = DP_128b_132b_TPS2;
-		break;
-	default:
-		pattern = DP_TRAINING_PATTERN_SEQUENCE_2;
-		break;
-	}
-	return pattern;
-#else
-	enum dc_dp_training_pattern highest_tp = DP_TRAINING_PATTERN_SEQUENCE_2;
-	struct encoder_feature_support *features;
-	struct dpcd_caps *dpcd_caps = &link->dpcd_caps;
-
-	/* Access link encoder capability based on whether it is statically
-	 * or dynamically assigned to a link.
-	 */
-	if (link->is_dig_mapping_flexible &&
-			link->dc->res_pool->funcs->link_encs_assign)
-		link_enc = link_enc_cfg_get_link_enc_used_by_link(link->ctx->dc, link);
-	else
-		link_enc = link->link_enc;
-	ASSERT(link_enc);
-	features = &link_enc->features;
-
-	if (features->flags.bits.IS_TPS3_CAPABLE)
-		highest_tp = DP_TRAINING_PATTERN_SEQUENCE_3;
-
-	if (features->flags.bits.IS_TPS4_CAPABLE)
-		highest_tp = DP_TRAINING_PATTERN_SEQUENCE_4;
-
-	if (dpcd_caps->max_down_spread.bits.TPS4_SUPPORTED &&
-		highest_tp >= DP_TRAINING_PATTERN_SEQUENCE_4)
-		return DP_TRAINING_PATTERN_SEQUENCE_4;
-
-	if (dpcd_caps->max_ln_count.bits.TPS3_SUPPORTED &&
-		highest_tp >= DP_TRAINING_PATTERN_SEQUENCE_3)
-		return DP_TRAINING_PATTERN_SEQUENCE_3;
-
-	return DP_TRAINING_PATTERN_SEQUENCE_2;
-#endif
-}
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-static uint8_t get_dpcd_link_rate(const struct dc_link_settings *link_settings)
-{
-	uint8_t link_rate = 0;
-	enum dp_link_encoding encoding = dp_get_link_encoding_format(link_settings);
-
-	if (encoding == DP_128b_132b_ENCODING)
-		switch (link_settings->link_rate) {
-		case LINK_RATE_UHBR10:
-			link_rate = 0x1;
-			break;
-		case LINK_RATE_UHBR20:
-			link_rate = 0x2;
-			break;
-		case LINK_RATE_UHBR13_5:
-			link_rate = 0x4;
-			break;
-		default:
-			link_rate = 0;
-			break;
-		}
-	else if (encoding == DP_8b_10b_ENCODING)
-		link_rate = (uint8_t) link_settings->link_rate;
-	else
-		link_rate = 0;
-
-	return link_rate;
-}
-#endif
-
-enum dc_status dpcd_set_link_settings(
-	struct dc_link *link,
-	const struct link_training_settings *lt_settings)
-{
-	uint8_t rate;
-	enum dc_status status;
-
-	union down_spread_ctrl downspread = {0};
-	union lane_count_set lane_count_set = {0};
-
-	downspread.raw = (uint8_t)
-	(lt_settings->link_settings.link_spread);
-
-	lane_count_set.bits.LANE_COUNT_SET =
-	lt_settings->link_settings.lane_count;
-
-	lane_count_set.bits.ENHANCED_FRAMING = lt_settings->enhanced_framing;
-	lane_count_set.bits.POST_LT_ADJ_REQ_GRANTED = 0;
-
-
-	if (link->ep_type == DISPLAY_ENDPOINT_PHY &&
-			lt_settings->pattern_for_eq < DP_TRAINING_PATTERN_SEQUENCE_4) {
-		lane_count_set.bits.POST_LT_ADJ_REQ_GRANTED =
-				link->dpcd_caps.max_ln_count.bits.POST_LT_ADJ_REQ_SUPPORTED;
-	}
-
-	status = core_link_write_dpcd(link, DP_DOWNSPREAD_CTRL,
-		&downspread.raw, sizeof(downspread));
-
-	status = core_link_write_dpcd(link, DP_LANE_COUNT_SET,
-		&lane_count_set.raw, 1);
-
-	if (link->dpcd_caps.dpcd_rev.raw >= DPCD_REV_14 &&
-			lt_settings->link_settings.use_link_rate_set == true) {
-		rate = 0;
-		/* WA for some MUX chips that will power down with eDP and lose supported
-		 * link rate set for eDP 1.4. Source reads DPCD 0x010 again to ensure
-		 * MUX chip gets link rate set back before link training.
-		 */
-		if (link->connector_signal == SIGNAL_TYPE_EDP) {
-			uint8_t supported_link_rates[16];
-
-			core_link_read_dpcd(link, DP_SUPPORTED_LINK_RATES,
-					supported_link_rates, sizeof(supported_link_rates));
-		}
-		status = core_link_write_dpcd(link, DP_LINK_BW_SET, &rate, 1);
-		status = core_link_write_dpcd(link, DP_LINK_RATE_SET,
-				&lt_settings->link_settings.link_rate_set, 1);
-	} else {
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-		rate = get_dpcd_link_rate(&lt_settings->link_settings);
-#else
-		rate = (uint8_t) (lt_settings->link_settings.link_rate);
-#endif
-		status = core_link_write_dpcd(link, DP_LINK_BW_SET, &rate, 1);
-	}
-
-	if (rate) {
-		DC_LOG_HW_LINK_TRAINING("%s\n %x rate = %x\n %x lane = %x framing = %x\n %x spread = %x\n",
-			__func__,
-			DP_LINK_BW_SET,
-			lt_settings->link_settings.link_rate,
-			DP_LANE_COUNT_SET,
-			lt_settings->link_settings.lane_count,
-			lt_settings->enhanced_framing,
-			DP_DOWNSPREAD_CTRL,
-			lt_settings->link_settings.link_spread);
-	} else {
-		DC_LOG_HW_LINK_TRAINING("%s\n %x rate set = %x\n %x lane = %x framing = %x\n %x spread = %x\n",
-			__func__,
-			DP_LINK_RATE_SET,
-			lt_settings->link_settings.link_rate_set,
-			DP_LANE_COUNT_SET,
-			lt_settings->link_settings.lane_count,
-			lt_settings->enhanced_framing,
-			DP_DOWNSPREAD_CTRL,
-			lt_settings->link_settings.link_spread);
-	}
-
-	return status;
-}
-
-uint8_t dc_dp_initialize_scrambling_data_symbols(
-	struct dc_link *link,
-	enum dc_dp_training_pattern pattern)
-{
-	uint8_t disable_scrabled_data_symbols = 0;
-
-	switch (pattern) {
-	case DP_TRAINING_PATTERN_SEQUENCE_1:
-	case DP_TRAINING_PATTERN_SEQUENCE_2:
-	case DP_TRAINING_PATTERN_SEQUENCE_3:
-		disable_scrabled_data_symbols = 1;
-		break;
-	case DP_TRAINING_PATTERN_SEQUENCE_4:
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	case DP_128b_132b_TPS1:
-	case DP_128b_132b_TPS2:
-#endif
-		disable_scrabled_data_symbols = 0;
-		break;
-	default:
-		ASSERT(0);
-		DC_LOG_HW_LINK_TRAINING("%s: Invalid HW Training pattern: %d\n",
-			__func__, pattern);
-		break;
-	}
-	return disable_scrabled_data_symbols;
-}
-
-static inline bool is_repeater(struct dc_link *link, uint32_t offset)
-{
-	return (link->lttpr_mode == LTTPR_MODE_NON_TRANSPARENT) && (offset != 0);
-}
-
-static void dpcd_set_lt_pattern_and_lane_settings(
-	struct dc_link *link,
-	const struct link_training_settings *lt_settings,
-	enum dc_dp_training_pattern pattern,
-	uint32_t offset)
-{
-	uint32_t dpcd_base_lt_offset;
-
-	uint8_t dpcd_lt_buffer[5] = {0};
-	union dpcd_training_pattern dpcd_pattern = { 0 };
-	uint32_t size_in_bytes;
-	bool edp_workaround = false; /* TODO link_prop.INTERNAL */
-	dpcd_base_lt_offset = DP_TRAINING_PATTERN_SET;
-
-	if (is_repeater(link, offset))
-		dpcd_base_lt_offset = DP_TRAINING_PATTERN_SET_PHY_REPEATER1 +
-			((DP_REPEATER_CONFIGURATION_AND_STATUS_SIZE) * (offset - 1));
-
-	/*****************************************************************
-	* DpcdAddress_TrainingPatternSet
-	*****************************************************************/
-	dpcd_pattern.v1_4.TRAINING_PATTERN_SET =
-		dc_dp_training_pattern_to_dpcd_training_pattern(link, pattern);
-
-	dpcd_pattern.v1_4.SCRAMBLING_DISABLE =
-		dc_dp_initialize_scrambling_data_symbols(link, pattern);
-
-	dpcd_lt_buffer[DP_TRAINING_PATTERN_SET - DP_TRAINING_PATTERN_SET]
-		= dpcd_pattern.raw;
-
-	if (is_repeater(link, offset)) {
-		DC_LOG_HW_LINK_TRAINING("%s\n LTTPR Repeater ID: %d\n 0x%X pattern = %x\n",
-			__func__,
-			offset,
-			dpcd_base_lt_offset,
-			dpcd_pattern.v1_4.TRAINING_PATTERN_SET);
-	} else {
-		DC_LOG_HW_LINK_TRAINING("%s\n 0x%X pattern = %x\n",
-			__func__,
-			dpcd_base_lt_offset,
-			dpcd_pattern.v1_4.TRAINING_PATTERN_SET);
-	}
-
-	/* concatenate everything into one buffer*/
-	size_in_bytes = lt_settings->link_settings.lane_count *
-			sizeof(lt_settings->dpcd_lane_settings[0]);
-
-	 // 0x00103 - 0x00102
-	memmove(
-		&dpcd_lt_buffer[DP_TRAINING_LANE0_SET - DP_TRAINING_PATTERN_SET],
-		lt_settings->dpcd_lane_settings,
-		size_in_bytes);
-
-	if (is_repeater(link, offset)) {
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-		if (dp_get_link_encoding_format(&lt_settings->link_settings) ==
-				DP_128b_132b_ENCODING)
-			DC_LOG_HW_LINK_TRAINING("%s:\n LTTPR Repeater ID: %d\n"
-					" 0x%X TX_FFE_PRESET_VALUE = %x\n",
-					__func__,
-					offset,
-					dpcd_base_lt_offset,
-					lt_settings->dpcd_lane_settings[0].tx_ffe.PRESET_VALUE);
-		else if (dp_get_link_encoding_format(&lt_settings->link_settings) ==
-				DP_8b_10b_ENCODING)
-#endif
-		DC_LOG_HW_LINK_TRAINING("%s:\n LTTPR Repeater ID: %d\n"
-				" 0x%X VS set = %x PE set = %x max VS Reached = %x  max PE Reached = %x\n",
-			__func__,
-			offset,
-			dpcd_base_lt_offset,
-			lt_settings->dpcd_lane_settings[0].bits.VOLTAGE_SWING_SET,
-			lt_settings->dpcd_lane_settings[0].bits.PRE_EMPHASIS_SET,
-			lt_settings->dpcd_lane_settings[0].bits.MAX_SWING_REACHED,
-			lt_settings->dpcd_lane_settings[0].bits.MAX_PRE_EMPHASIS_REACHED);
-	} else {
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-		if (dp_get_link_encoding_format(&lt_settings->link_settings) ==
-				DP_128b_132b_ENCODING)
-			DC_LOG_HW_LINK_TRAINING("%s:\n 0x%X TX_FFE_PRESET_VALUE = %x\n",
-					__func__,
-					dpcd_base_lt_offset,
-					lt_settings->dpcd_lane_settings[0].tx_ffe.PRESET_VALUE);
-		else if (dp_get_link_encoding_format(&lt_settings->link_settings) ==
-				DP_8b_10b_ENCODING)
-#endif
-		DC_LOG_HW_LINK_TRAINING("%s:\n 0x%X VS set = %x  PE set = %x max VS Reached = %x  max PE Reached = %x\n",
-			__func__,
-			dpcd_base_lt_offset,
-			lt_settings->dpcd_lane_settings[0].bits.VOLTAGE_SWING_SET,
-			lt_settings->dpcd_lane_settings[0].bits.PRE_EMPHASIS_SET,
-			lt_settings->dpcd_lane_settings[0].bits.MAX_SWING_REACHED,
-			lt_settings->dpcd_lane_settings[0].bits.MAX_PRE_EMPHASIS_REACHED);
-	}
-	if (edp_workaround) {
-		/* for eDP write in 2 parts because the 5-byte burst is
-		* causing issues on some eDP panels (EPR#366724)
-		*/
-		core_link_write_dpcd(
-			link,
-			DP_TRAINING_PATTERN_SET,
-			&dpcd_pattern.raw,
-			sizeof(dpcd_pattern.raw));
-
-		core_link_write_dpcd(
-			link,
-			DP_TRAINING_LANE0_SET,
-			(uint8_t *)(lt_settings->dpcd_lane_settings),
-			size_in_bytes);
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	} else if (dp_get_link_encoding_format(&lt_settings->link_settings) ==
-			DP_128b_132b_ENCODING) {
-		core_link_write_dpcd(
-				link,
-				dpcd_base_lt_offset,
-				dpcd_lt_buffer,
-				sizeof(dpcd_lt_buffer));
-#endif
-		} else
-		/* write it all in (1 + number-of-lanes)-byte burst*/
-			core_link_write_dpcd(
-				link,
-				dpcd_base_lt_offset,
-				dpcd_lt_buffer,
-				size_in_bytes + sizeof(dpcd_pattern.raw));
-}
-
-bool dp_is_cr_done(enum dc_lane_count ln_count,
-	union lane_status *dpcd_lane_status)
-{
-	uint32_t lane;
-	/*LANEx_CR_DONE bits All 1's?*/
-	for (lane = 0; lane < (uint32_t)(ln_count); lane++) {
-		if (!dpcd_lane_status[lane].bits.CR_DONE_0)
-			return false;
-	}
-	return true;
-}
-
-bool dp_is_ch_eq_done(enum dc_lane_count ln_count,
-		union lane_status *dpcd_lane_status)
-{
-	bool done = true;
-	uint32_t lane;
-	for (lane = 0; lane < (uint32_t)(ln_count); lane++)
-		if (!dpcd_lane_status[lane].bits.CHANNEL_EQ_DONE_0)
-			done = false;
-	return done;
-}
-
-bool dp_is_symbol_locked(enum dc_lane_count ln_count,
-		union lane_status *dpcd_lane_status)
-{
-	bool locked = true;
-	uint32_t lane;
-	for (lane = 0; lane < (uint32_t)(ln_count); lane++)
-		if (!dpcd_lane_status[lane].bits.SYMBOL_LOCKED_0)
-			locked = false;
-	return locked;
-}
-
-bool dp_is_interlane_aligned(union lane_align_status_updated align_status)
-{
-	return align_status.bits.INTERLANE_ALIGN_DONE == 1;
-}
-
-void dp_hw_to_dpcd_lane_settings(
-		const struct link_training_settings *lt_settings,
-		const struct dc_lane_settings hw_lane_settings[LANE_COUNT_DP_MAX],
-		union dpcd_training_lane dpcd_lane_settings[LANE_COUNT_DP_MAX])
-{
-	uint8_t lane = 0;
-
-	for (lane = 0; lane < LANE_COUNT_DP_MAX; lane++) {
-		if (dp_get_link_encoding_format(&lt_settings->link_settings) ==
-				DP_8b_10b_ENCODING) {
-			dpcd_lane_settings[lane].bits.VOLTAGE_SWING_SET =
-					(uint8_t)(hw_lane_settings[lane].VOLTAGE_SWING);
-			dpcd_lane_settings[lane].bits.PRE_EMPHASIS_SET =
-					(uint8_t)(hw_lane_settings[lane].PRE_EMPHASIS);
-			dpcd_lane_settings[lane].bits.MAX_SWING_REACHED =
-					(hw_lane_settings[lane].VOLTAGE_SWING ==
-							VOLTAGE_SWING_MAX_LEVEL ? 1 : 0);
-			dpcd_lane_settings[lane].bits.MAX_PRE_EMPHASIS_REACHED =
-					(hw_lane_settings[lane].PRE_EMPHASIS ==
-							PRE_EMPHASIS_MAX_LEVEL ? 1 : 0);
-		}
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-		else if (dp_get_link_encoding_format(&lt_settings->link_settings) ==
-				DP_128b_132b_ENCODING) {
-			dpcd_lane_settings[lane].tx_ffe.PRESET_VALUE =
-					hw_lane_settings[lane].FFE_PRESET.settings.level;
-		}
-#endif
-	}
-}
-
-void dp_decide_lane_settings(
-		const struct link_training_settings *lt_settings,
-		const union lane_adjust ln_adjust[LANE_COUNT_DP_MAX],
-		struct dc_lane_settings hw_lane_settings[LANE_COUNT_DP_MAX],
-		union dpcd_training_lane dpcd_lane_settings[LANE_COUNT_DP_MAX])
-{
-	uint32_t lane;
-
-	for (lane = 0; lane < LANE_COUNT_DP_MAX; lane++) {
-		if (dp_get_link_encoding_format(&lt_settings->link_settings) ==
-				DP_8b_10b_ENCODING) {
-			hw_lane_settings[lane].VOLTAGE_SWING =
-					(enum dc_voltage_swing)(ln_adjust[lane].bits.
-							VOLTAGE_SWING_LANE);
-			hw_lane_settings[lane].PRE_EMPHASIS =
-					(enum dc_pre_emphasis)(ln_adjust[lane].bits.
-							PRE_EMPHASIS_LANE);
-		}
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-		else if (dp_get_link_encoding_format(&lt_settings->link_settings) ==
-				DP_128b_132b_ENCODING) {
-			hw_lane_settings[lane].FFE_PRESET.raw =
-					ln_adjust[lane].tx_ffe.PRESET_VALUE;
-		}
-#endif
-	}
-	dp_hw_to_dpcd_lane_settings(lt_settings, hw_lane_settings, dpcd_lane_settings);
-
-	if (lt_settings->disallow_per_lane_settings) {
-		/* we find the maximum of the requested settings across all lanes*/
-		/* and set this maximum for all lanes*/
-		maximize_lane_settings(lt_settings, hw_lane_settings);
-		override_lane_settings(lt_settings, hw_lane_settings);
-
-		if (lt_settings->always_match_dpcd_with_hw_lane_settings)
-			dp_hw_to_dpcd_lane_settings(lt_settings, hw_lane_settings, dpcd_lane_settings);
-	}
-
-}
-
-static uint8_t get_nibble_at_index(const uint8_t *buf,
-	uint32_t index)
-{
-	uint8_t nibble;
-	nibble = buf[index / 2];
-
-	if (index % 2)
-		nibble >>= 4;
-	else
-		nibble &= 0x0F;
-
-	return nibble;
-}
-
-static enum dc_pre_emphasis get_max_pre_emphasis_for_voltage_swing(
-	enum dc_voltage_swing voltage)
-{
-	enum dc_pre_emphasis pre_emphasis;
-	pre_emphasis = PRE_EMPHASIS_MAX_LEVEL;
-
-	if (voltage <= VOLTAGE_SWING_MAX_LEVEL)
-		pre_emphasis = voltage_swing_to_pre_emphasis[voltage];
-
-	return pre_emphasis;
-
-}
-
-static void maximize_lane_settings(const struct link_training_settings *lt_settings,
-		struct dc_lane_settings lane_settings[LANE_COUNT_DP_MAX])
-{
-	uint32_t lane;
-	struct dc_lane_settings max_requested;
-
-	max_requested.VOLTAGE_SWING = lane_settings[0].VOLTAGE_SWING;
-	max_requested.PRE_EMPHASIS = lane_settings[0].PRE_EMPHASIS;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	max_requested.FFE_PRESET = lane_settings[0].FFE_PRESET;
-#endif
-
-	/* Determine what the maximum of the requested settings are*/
-	for (lane = 1; lane < lt_settings->link_settings.lane_count; lane++) {
-		if (lane_settings[lane].VOLTAGE_SWING > max_requested.VOLTAGE_SWING)
-			max_requested.VOLTAGE_SWING = lane_settings[lane].VOLTAGE_SWING;
-
-		if (lane_settings[lane].PRE_EMPHASIS > max_requested.PRE_EMPHASIS)
-			max_requested.PRE_EMPHASIS = lane_settings[lane].PRE_EMPHASIS;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-		if (lane_settings[lane].FFE_PRESET.settings.level >
-				max_requested.FFE_PRESET.settings.level)
-			max_requested.FFE_PRESET.settings.level =
-					lane_settings[lane].FFE_PRESET.settings.level;
-#endif
-	}
-
-	/* make sure the requested settings are
-	 * not higher than maximum settings*/
-	if (max_requested.VOLTAGE_SWING > VOLTAGE_SWING_MAX_LEVEL)
-		max_requested.VOLTAGE_SWING = VOLTAGE_SWING_MAX_LEVEL;
-
-	if (max_requested.PRE_EMPHASIS > PRE_EMPHASIS_MAX_LEVEL)
-		max_requested.PRE_EMPHASIS = PRE_EMPHASIS_MAX_LEVEL;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	if (max_requested.FFE_PRESET.settings.level > DP_FFE_PRESET_MAX_LEVEL)
-		max_requested.FFE_PRESET.settings.level = DP_FFE_PRESET_MAX_LEVEL;
-#endif
-
-	/* make sure the pre-emphasis matches the voltage swing*/
-	if (max_requested.PRE_EMPHASIS >
-		get_max_pre_emphasis_for_voltage_swing(
-			max_requested.VOLTAGE_SWING))
-		max_requested.PRE_EMPHASIS =
-		get_max_pre_emphasis_for_voltage_swing(
-			max_requested.VOLTAGE_SWING);
-
-	for (lane = 0; lane < LANE_COUNT_DP_MAX; lane++) {
-		lane_settings[lane].VOLTAGE_SWING = max_requested.VOLTAGE_SWING;
-		lane_settings[lane].PRE_EMPHASIS = max_requested.PRE_EMPHASIS;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-		lane_settings[lane].FFE_PRESET = max_requested.FFE_PRESET;
-#endif
-	}
-}
-
-static void override_lane_settings(const struct link_training_settings *lt_settings,
-		struct dc_lane_settings lane_settings[LANE_COUNT_DP_MAX])
-{
-	uint32_t lane;
-
-	if (lt_settings->voltage_swing == NULL &&
-	    lt_settings->pre_emphasis == NULL &&
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	    lt_settings->ffe_preset == NULL &&
-#endif
-	    lt_settings->post_cursor2 == NULL)
-
-		return;
-
-	for (lane = 1; lane < LANE_COUNT_DP_MAX; lane++) {
-		if (lt_settings->voltage_swing)
-			lane_settings[lane].VOLTAGE_SWING = *lt_settings->voltage_swing;
-		if (lt_settings->pre_emphasis)
-			lane_settings[lane].PRE_EMPHASIS = *lt_settings->pre_emphasis;
-		if (lt_settings->post_cursor2)
-			lane_settings[lane].POST_CURSOR2 = *lt_settings->post_cursor2;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-		if (lt_settings->ffe_preset)
-			lane_settings[lane].FFE_PRESET = *lt_settings->ffe_preset;
-#endif
-	}
-}
-
-enum dc_status dp_get_lane_status_and_lane_adjust(
-	struct dc_link *link,
-	const struct link_training_settings *link_training_setting,
-	union lane_status ln_status[LANE_COUNT_DP_MAX],
-	union lane_align_status_updated *ln_align,
-	union lane_adjust ln_adjust[LANE_COUNT_DP_MAX],
-	uint32_t offset)
-{
-	unsigned int lane01_status_address = DP_LANE0_1_STATUS;
-	uint8_t lane_adjust_offset = 4;
-	unsigned int lane01_adjust_address;
-	uint8_t dpcd_buf[6] = {0};
-	uint32_t lane;
-	enum dc_status status;
-
-	if (is_repeater(link, offset)) {
-		lane01_status_address =
-				DP_LANE0_1_STATUS_PHY_REPEATER1 +
-				((DP_REPEATER_CONFIGURATION_AND_STATUS_SIZE) * (offset - 1));
-		lane_adjust_offset = 3;
-	}
-
-	status = core_link_read_dpcd(
-		link,
-		lane01_status_address,
-		(uint8_t *)(dpcd_buf),
-		sizeof(dpcd_buf));
-
-	for (lane = 0; lane <
-		(uint32_t)(link_training_setting->link_settings.lane_count);
-		lane++) {
-
-		ln_status[lane].raw =
-			get_nibble_at_index(&dpcd_buf[0], lane);
-		ln_adjust[lane].raw =
-			get_nibble_at_index(&dpcd_buf[lane_adjust_offset], lane);
-	}
-
-	ln_align->raw = dpcd_buf[2];
-
-	if (is_repeater(link, offset)) {
-		DC_LOG_HW_LINK_TRAINING("%s:\n LTTPR Repeater ID: %d\n"
-				" 0x%X Lane01Status = %x\n 0x%X Lane23Status = %x\n ",
-			__func__,
-			offset,
-			lane01_status_address, dpcd_buf[0],
-			lane01_status_address + 1, dpcd_buf[1]);
-	} else {
-		DC_LOG_HW_LINK_TRAINING("%s:\n 0x%X Lane01Status = %x\n 0x%X Lane23Status = %x\n ",
-			__func__,
-			lane01_status_address, dpcd_buf[0],
-			lane01_status_address + 1, dpcd_buf[1]);
-	}
-	lane01_adjust_address = DP_ADJUST_REQUEST_LANE0_1;
-
-	if (is_repeater(link, offset))
-		lane01_adjust_address = DP_ADJUST_REQUEST_LANE0_1_PHY_REPEATER1 +
-				((DP_REPEATER_CONFIGURATION_AND_STATUS_SIZE) * (offset - 1));
-
-	if (is_repeater(link, offset)) {
-		DC_LOG_HW_LINK_TRAINING("%s:\n LTTPR Repeater ID: %d\n"
-				" 0x%X Lane01AdjustRequest = %x\n 0x%X Lane23AdjustRequest = %x\n",
-					__func__,
-					offset,
-					lane01_adjust_address,
-					dpcd_buf[lane_adjust_offset],
-					lane01_adjust_address + 1,
-					dpcd_buf[lane_adjust_offset + 1]);
-	} else {
-		DC_LOG_HW_LINK_TRAINING("%s:\n 0x%X Lane01AdjustRequest = %x\n 0x%X Lane23AdjustRequest = %x\n",
-			__func__,
-			lane01_adjust_address,
-			dpcd_buf[lane_adjust_offset],
-			lane01_adjust_address + 1,
-			dpcd_buf[lane_adjust_offset + 1]);
-	}
-
-	return status;
-}
-
-enum dc_status dpcd_set_lane_settings(
-	struct dc_link *link,
-	const struct link_training_settings *link_training_setting,
-	uint32_t offset)
-{
-	unsigned int lane0_set_address;
-	enum dc_status status;
-
-	lane0_set_address = DP_TRAINING_LANE0_SET;
-
-	if (is_repeater(link, offset))
-		lane0_set_address = DP_TRAINING_LANE0_SET_PHY_REPEATER1 +
-		((DP_REPEATER_CONFIGURATION_AND_STATUS_SIZE) * (offset - 1));
-
-	status = core_link_write_dpcd(link,
-		lane0_set_address,
-		(uint8_t *)(link_training_setting->dpcd_lane_settings),
-		link_training_setting->link_settings.lane_count);
-
-	if (is_repeater(link, offset)) {
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-		if (dp_get_link_encoding_format(&link_training_setting->link_settings) ==
-				DP_128b_132b_ENCODING)
-			DC_LOG_HW_LINK_TRAINING("%s:\n LTTPR Repeater ID: %d\n"
-					" 0x%X TX_FFE_PRESET_VALUE = %x\n",
-					__func__,
-					offset,
-					lane0_set_address,
-					link_training_setting->dpcd_lane_settings[0].tx_ffe.PRESET_VALUE);
-		else if (dp_get_link_encoding_format(&link_training_setting->link_settings) ==
-				DP_8b_10b_ENCODING)
-#endif
-		DC_LOG_HW_LINK_TRAINING("%s\n LTTPR Repeater ID: %d\n"
-				" 0x%X VS set = %x  PE set = %x max VS Reached = %x  max PE Reached = %x\n",
-			__func__,
-			offset,
-			lane0_set_address,
-			link_training_setting->dpcd_lane_settings[0].bits.VOLTAGE_SWING_SET,
-			link_training_setting->dpcd_lane_settings[0].bits.PRE_EMPHASIS_SET,
-			link_training_setting->dpcd_lane_settings[0].bits.MAX_SWING_REACHED,
-			link_training_setting->dpcd_lane_settings[0].bits.MAX_PRE_EMPHASIS_REACHED);
-
-	} else {
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-		if (dp_get_link_encoding_format(&link_training_setting->link_settings) ==
-				DP_128b_132b_ENCODING)
-			DC_LOG_HW_LINK_TRAINING("%s:\n 0x%X TX_FFE_PRESET_VALUE = %x\n",
-					__func__,
-					lane0_set_address,
-					link_training_setting->dpcd_lane_settings[0].tx_ffe.PRESET_VALUE);
-		else if (dp_get_link_encoding_format(&link_training_setting->link_settings) ==
-				DP_8b_10b_ENCODING)
-#endif
-		DC_LOG_HW_LINK_TRAINING("%s\n 0x%X VS set = %x  PE set = %x max VS Reached = %x  max PE Reached = %x\n",
-			__func__,
-			lane0_set_address,
-			link_training_setting->dpcd_lane_settings[0].bits.VOLTAGE_SWING_SET,
-			link_training_setting->dpcd_lane_settings[0].bits.PRE_EMPHASIS_SET,
-			link_training_setting->dpcd_lane_settings[0].bits.MAX_SWING_REACHED,
-			link_training_setting->dpcd_lane_settings[0].bits.MAX_PRE_EMPHASIS_REACHED);
-	}
-
-	return status;
-}
-
-bool dp_is_max_vs_reached(
-	const struct link_training_settings *lt_settings)
-{
-	uint32_t lane;
-	for (lane = 0; lane <
-		(uint32_t)(lt_settings->link_settings.lane_count);
-		lane++) {
-		if (lt_settings->dpcd_lane_settings[lane].bits.VOLTAGE_SWING_SET
-			== VOLTAGE_SWING_MAX_LEVEL)
-			return true;
-	}
-	return false;
-
-}
-
-static bool perform_post_lt_adj_req_sequence(
-	struct dc_link *link,
-	struct link_training_settings *lt_settings)
-{
-	enum dc_lane_count lane_count =
-	lt_settings->link_settings.lane_count;
-
-	uint32_t adj_req_count;
-	uint32_t adj_req_timer;
-	bool req_drv_setting_changed;
-	uint32_t lane;
-
-	req_drv_setting_changed = false;
-	for (adj_req_count = 0; adj_req_count < POST_LT_ADJ_REQ_LIMIT;
-	adj_req_count++) {
-
-		req_drv_setting_changed = false;
-
-		for (adj_req_timer = 0;
-			adj_req_timer < POST_LT_ADJ_REQ_TIMEOUT;
-			adj_req_timer++) {
-
-			union lane_status dpcd_lane_status[LANE_COUNT_DP_MAX];
-			union lane_align_status_updated
-				dpcd_lane_status_updated;
-			union lane_adjust dpcd_lane_adjust[LANE_COUNT_DP_MAX] = { { {0} } };
-
-			dp_get_lane_status_and_lane_adjust(
-				link,
-				lt_settings,
-				dpcd_lane_status,
-				&dpcd_lane_status_updated,
-				dpcd_lane_adjust,
-				DPRX);
-
-			if (dpcd_lane_status_updated.bits.
-					POST_LT_ADJ_REQ_IN_PROGRESS == 0)
-				return true;
-
-			if (!dp_is_cr_done(lane_count, dpcd_lane_status))
-				return false;
-
-			if (!dp_is_ch_eq_done(lane_count, dpcd_lane_status) ||
-					!dp_is_symbol_locked(lane_count, dpcd_lane_status) ||
-					!dp_is_interlane_aligned(dpcd_lane_status_updated))
-				return false;
-
-			for (lane = 0; lane < (uint32_t)(lane_count); lane++) {
-
-				if (lt_settings->
-				dpcd_lane_settings[lane].bits.VOLTAGE_SWING_SET !=
-				dpcd_lane_adjust[lane].bits.VOLTAGE_SWING_LANE ||
-				lt_settings->dpcd_lane_settings[lane].bits.PRE_EMPHASIS_SET !=
-				dpcd_lane_adjust[lane].bits.PRE_EMPHASIS_LANE) {
-
-					req_drv_setting_changed = true;
-					break;
-				}
-			}
-
-			if (req_drv_setting_changed) {
-				dp_decide_lane_settings(lt_settings, dpcd_lane_adjust,
-						lt_settings->hw_lane_settings, lt_settings->dpcd_lane_settings);
-
-				dc_link_dp_set_drive_settings(link,
-						lt_settings);
-				break;
-			}
-
-			msleep(1);
-		}
-
-		if (!req_drv_setting_changed) {
-			DC_LOG_WARNING("%s: Post Link Training Adjust Request Timed out\n",
-				__func__);
-
-			ASSERT(0);
-			return true;
-		}
-	}
-	DC_LOG_WARNING("%s: Post Link Training Adjust Request limit reached\n",
-		__func__);
-
-	ASSERT(0);
-	return true;
-
-}
-
-/* Only used for channel equalization */
-uint32_t dp_translate_training_aux_read_interval(uint32_t dpcd_aux_read_interval)
-{
-	unsigned int aux_rd_interval_us = 400;
-
-	switch (dpcd_aux_read_interval) {
-	case 0x01:
-		aux_rd_interval_us = 4000;
-		break;
-	case 0x02:
-		aux_rd_interval_us = 8000;
-		break;
-	case 0x03:
-		aux_rd_interval_us = 12000;
-		break;
-	case 0x04:
-		aux_rd_interval_us = 16000;
-		break;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	case 0x05:
-		aux_rd_interval_us = 32000;
-		break;
-	case 0x06:
-		aux_rd_interval_us = 64000;
-		break;
-#endif
-	default:
-		break;
-	}
-
-	return aux_rd_interval_us;
-}
-
-enum link_training_result dp_get_cr_failure(enum dc_lane_count ln_count,
-					union lane_status *dpcd_lane_status)
-{
-	enum link_training_result result = LINK_TRAINING_SUCCESS;
-
-	if (ln_count >= LANE_COUNT_ONE && !dpcd_lane_status[0].bits.CR_DONE_0)
-		result = LINK_TRAINING_CR_FAIL_LANE0;
-	else if (ln_count >= LANE_COUNT_TWO && !dpcd_lane_status[1].bits.CR_DONE_0)
-		result = LINK_TRAINING_CR_FAIL_LANE1;
-	else if (ln_count >= LANE_COUNT_FOUR && !dpcd_lane_status[2].bits.CR_DONE_0)
-		result = LINK_TRAINING_CR_FAIL_LANE23;
-	else if (ln_count >= LANE_COUNT_FOUR && !dpcd_lane_status[3].bits.CR_DONE_0)
-		result = LINK_TRAINING_CR_FAIL_LANE23;
-	return result;
-}
-
-static enum link_training_result perform_channel_equalization_sequence(
-	struct dc_link *link,
-	struct link_training_settings *lt_settings,
-	uint32_t offset)
-{
-	enum dc_dp_training_pattern tr_pattern;
-	uint32_t retries_ch_eq;
-	uint32_t wait_time_microsec;
-	enum dc_lane_count lane_count = lt_settings->link_settings.lane_count;
-	union lane_align_status_updated dpcd_lane_status_updated = {0};
-	union lane_status dpcd_lane_status[LANE_COUNT_DP_MAX] = {0};
-	union lane_adjust dpcd_lane_adjust[LANE_COUNT_DP_MAX] = {0};
-
-	/* Note: also check that TPS4 is a supported feature*/
-	tr_pattern = lt_settings->pattern_for_eq;
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	if (is_repeater(link, offset) && dp_get_link_encoding_format(&lt_settings->link_settings) == DP_8b_10b_ENCODING)
-		tr_pattern = DP_TRAINING_PATTERN_SEQUENCE_4;
-#else
-	if (is_repeater(link, offset))
-		tr_pattern = DP_TRAINING_PATTERN_SEQUENCE_4;
-#endif
-
-	dp_set_hw_training_pattern(link, tr_pattern, offset);
-
-	for (retries_ch_eq = 0; retries_ch_eq <= LINK_TRAINING_MAX_RETRY_COUNT;
-		retries_ch_eq++) {
-
-		dp_set_hw_lane_settings(link, lt_settings, offset);
-
-		/* 2. update DPCD*/
-		if (!retries_ch_eq)
-			/* EPR #361076 - write as a 5-byte burst,
-			 * but only for the 1-st iteration
-			 */
-
-			dpcd_set_lt_pattern_and_lane_settings(
-				link,
-				lt_settings,
-				tr_pattern, offset);
-		else
-			dpcd_set_lane_settings(link, lt_settings, offset);
-
-		/* 3. wait for receiver to lock-on*/
-		wait_time_microsec = lt_settings->eq_pattern_time;
-
-		if (is_repeater(link, offset))
-			wait_time_microsec =
-					dp_translate_training_aux_read_interval(
-						link->dpcd_caps.lttpr_caps.aux_rd_interval[offset - 1]);
-
-		dp_wait_for_training_aux_rd_interval(
-				link,
-				wait_time_microsec);
-
-		/* 4. Read lane status and requested
-		 * drive settings as set by the sink*/
-
-		dp_get_lane_status_and_lane_adjust(
-			link,
-			lt_settings,
-			dpcd_lane_status,
-			&dpcd_lane_status_updated,
-			dpcd_lane_adjust,
-			offset);
-
-		/* 5. check CR done*/
-		if (!dp_is_cr_done(lane_count, dpcd_lane_status))
-			return LINK_TRAINING_EQ_FAIL_CR;
-
-		/* 6. check CHEQ done*/
-		if (dp_is_ch_eq_done(lane_count, dpcd_lane_status) &&
-				dp_is_symbol_locked(lane_count, dpcd_lane_status) &&
-				dp_is_interlane_aligned(dpcd_lane_status_updated))
-			return LINK_TRAINING_SUCCESS;
-
-		/* 7. update VS/PE/PC2 in lt_settings*/
-		dp_decide_lane_settings(lt_settings, dpcd_lane_adjust,
-				lt_settings->hw_lane_settings, lt_settings->dpcd_lane_settings);
-	}
-
-	return LINK_TRAINING_EQ_FAIL_EQ;
-
-}
-
-static void start_clock_recovery_pattern_early(struct dc_link *link,
-		struct link_training_settings *lt_settings,
-		uint32_t offset)
-{
-	DC_LOG_HW_LINK_TRAINING("%s\n GPU sends TPS1. Wait 400us.\n",
-			__func__);
-	dp_set_hw_training_pattern(link, lt_settings->pattern_for_cr, offset);
-	dp_set_hw_lane_settings(link, lt_settings, offset);
-	udelay(400);
-}
-
-static enum link_training_result perform_clock_recovery_sequence(
-	struct dc_link *link,
-	struct link_training_settings *lt_settings,
-	uint32_t offset)
-{
-	uint32_t retries_cr;
-	uint32_t retry_count;
-	uint32_t wait_time_microsec;
-	enum dc_lane_count lane_count = lt_settings->link_settings.lane_count;
-	union lane_status dpcd_lane_status[LANE_COUNT_DP_MAX];
-	union lane_align_status_updated dpcd_lane_status_updated;
-	union lane_adjust dpcd_lane_adjust[LANE_COUNT_DP_MAX] = { { {0} } };
-
-	retries_cr = 0;
-	retry_count = 0;
-
-	if (!link->ctx->dc->work_arounds.lt_early_cr_pattern)
-		dp_set_hw_training_pattern(link, lt_settings->pattern_for_cr, offset);
-
-	/* najeeb - The synaptics MST hub can put the LT in
-	* infinite loop by switching the VS
-	*/
-	/* between level 0 and level 1 continuously, here
-	* we try for CR lock for LinkTrainingMaxCRRetry count*/
-	while ((retries_cr < LINK_TRAINING_MAX_RETRY_COUNT) &&
-		(retry_count < LINK_TRAINING_MAX_CR_RETRY)) {
-
-		memset(&dpcd_lane_status, '\0', sizeof(dpcd_lane_status));
-		memset(&dpcd_lane_status_updated, '\0',
-		sizeof(dpcd_lane_status_updated));
-
-		/* 1. call HWSS to set lane settings*/
-		dp_set_hw_lane_settings(
-				link,
-				lt_settings,
-				offset);
-
-		/* 2. update DPCD of the receiver*/
-		if (!retry_count)
-			/* EPR #361076 - write as a 5-byte burst,
-			 * but only for the 1-st iteration.*/
-			dpcd_set_lt_pattern_and_lane_settings(
-					link,
-					lt_settings,
-					lt_settings->pattern_for_cr,
-					offset);
-		else
-			dpcd_set_lane_settings(
-					link,
-					lt_settings,
-					offset);
-
-		/* 3. wait receiver to lock-on*/
-		wait_time_microsec = lt_settings->cr_pattern_time;
-
-		if (link->lttpr_mode == LTTPR_MODE_NON_TRANSPARENT)
-			wait_time_microsec = TRAINING_AUX_RD_INTERVAL;
-
-		dp_wait_for_training_aux_rd_interval(
-				link,
-				wait_time_microsec);
-
-		/* 4. Read lane status and requested drive
-		* settings as set by the sink
-		*/
-		dp_get_lane_status_and_lane_adjust(
-				link,
-				lt_settings,
-				dpcd_lane_status,
-				&dpcd_lane_status_updated,
-				dpcd_lane_adjust,
-				offset);
-
-		/* 5. check CR done*/
-		if (dp_is_cr_done(lane_count, dpcd_lane_status))
-			return LINK_TRAINING_SUCCESS;
-
-		/* 6. max VS reached*/
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-		if ((dp_get_link_encoding_format(&lt_settings->link_settings) ==
-				DP_8b_10b_ENCODING) &&
-				dp_is_max_vs_reached(lt_settings))
-			break;
-#else
-		if (dp_is_max_vs_reached(lt_settings))
-			break;
-#endif
-
-		/* 7. same lane settings*/
-		/* Note: settings are the same for all lanes,
-		 * so comparing first lane is sufficient*/
-		if ((dp_get_link_encoding_format(&lt_settings->link_settings) == DP_8b_10b_ENCODING) &&
-				lt_settings->dpcd_lane_settings[0].bits.VOLTAGE_SWING_SET ==
-						dpcd_lane_adjust[0].bits.VOLTAGE_SWING_LANE)
-			retries_cr++;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-		else if ((dp_get_link_encoding_format(&lt_settings->link_settings) == DP_128b_132b_ENCODING) &&
-				lt_settings->dpcd_lane_settings[0].tx_ffe.PRESET_VALUE ==
-						dpcd_lane_adjust[0].tx_ffe.PRESET_VALUE)
-			retries_cr++;
-#endif
-		else
-			retries_cr = 0;
-
-		/* 8. update VS/PE/PC2 in lt_settings*/
-		dp_decide_lane_settings(lt_settings, dpcd_lane_adjust,
-				lt_settings->hw_lane_settings, lt_settings->dpcd_lane_settings);
-		retry_count++;
-	}
-
-	if (retry_count >= LINK_TRAINING_MAX_CR_RETRY) {
-		ASSERT(0);
-		DC_LOG_ERROR("%s: Link Training Error, could not get CR after %d tries. Possibly voltage swing issue",
-			__func__,
-			LINK_TRAINING_MAX_CR_RETRY);
-
-	}
-
-	return dp_get_cr_failure(lane_count, dpcd_lane_status);
-}
-
-static inline enum link_training_result dp_transition_to_video_idle(
-	struct dc_link *link,
-	struct link_training_settings *lt_settings,
-	enum link_training_result status)
-{
-	union lane_count_set lane_count_set = {0};
-
-	/* 4. mainlink output idle pattern*/
-	dp_set_hw_test_pattern(link, DP_TEST_PATTERN_VIDEO_MODE, NULL, 0);
-
-	/*
-	 * 5. post training adjust if required
-	 * If the upstream DPTX and downstream DPRX both support TPS4,
-	 * TPS4 must be used instead of POST_LT_ADJ_REQ.
-	 */
-	if (link->dpcd_caps.max_ln_count.bits.POST_LT_ADJ_REQ_SUPPORTED != 1 ||
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-			lt_settings->pattern_for_eq >= DP_TRAINING_PATTERN_SEQUENCE_4) {
-#else
-			lt_settings->pattern_for_eq == DP_TRAINING_PATTERN_SEQUENCE_4) {
-#endif
-		/* delay 5ms after Main Link output idle pattern and then check
-		 * DPCD 0202h.
-		 */
-		if (link->connector_signal != SIGNAL_TYPE_EDP && status == LINK_TRAINING_SUCCESS) {
-			msleep(5);
-			status = dp_check_link_loss_status(link, lt_settings);
-		}
-		return status;
-	}
-
-	if (status == LINK_TRAINING_SUCCESS &&
-		perform_post_lt_adj_req_sequence(link, lt_settings) == false)
-		status = LINK_TRAINING_LQA_FAIL;
-
-	lane_count_set.bits.LANE_COUNT_SET = lt_settings->link_settings.lane_count;
-	lane_count_set.bits.ENHANCED_FRAMING = lt_settings->enhanced_framing;
-	lane_count_set.bits.POST_LT_ADJ_REQ_GRANTED = 0;
-
-	core_link_write_dpcd(
-		link,
-		DP_LANE_COUNT_SET,
-		&lane_count_set.raw,
-		sizeof(lane_count_set));
-
-	return status;
-}
-
-enum link_training_result dp_check_link_loss_status(
-	struct dc_link *link,
-	const struct link_training_settings *link_training_setting)
-{
-	enum link_training_result status = LINK_TRAINING_SUCCESS;
-	union lane_status lane_status;
-	uint8_t dpcd_buf[6] = {0};
-	uint32_t lane;
-
-	core_link_read_dpcd(
-			link,
-			DP_SINK_COUNT,
-			(uint8_t *)(dpcd_buf),
-			sizeof(dpcd_buf));
-
-	/*parse lane status*/
-	for (lane = 0; lane < link->cur_link_settings.lane_count; lane++) {
-		/*
-		 * check lanes status
-		 */
-		lane_status.raw = get_nibble_at_index(&dpcd_buf[2], lane);
-
-		if (!lane_status.bits.CHANNEL_EQ_DONE_0 ||
-			!lane_status.bits.CR_DONE_0 ||
-			!lane_status.bits.SYMBOL_LOCKED_0) {
-			/* if one of the channel equalization, clock
-			 * recovery or symbol lock is dropped
-			 * consider it as (link has been
-			 * dropped) dp sink status has changed
-			 */
-			status = LINK_TRAINING_LINK_LOSS;
-			break;
-		}
-	}
-
-	return status;
-}
-
-static inline void decide_8b_10b_training_settings(
-	 struct dc_link *link,
-	const struct dc_link_settings *link_setting,
-	struct link_training_settings *lt_settings)
-{
-	memset(lt_settings, '\0', sizeof(struct link_training_settings));
-
-	/* Initialize link settings */
-	lt_settings->link_settings.use_link_rate_set = link_setting->use_link_rate_set;
-	lt_settings->link_settings.link_rate_set = link_setting->link_rate_set;
-	lt_settings->link_settings.link_rate = link_setting->link_rate;
-	lt_settings->link_settings.lane_count = link_setting->lane_count;
-	/* TODO hard coded to SS for now
-	 * lt_settings.link_settings.link_spread =
-	 * dal_display_path_is_ss_supported(
-	 * path_mode->display_path) ?
-	 * LINK_SPREAD_05_DOWNSPREAD_30KHZ :
-	 * LINK_SPREAD_DISABLED;
-	 */
-	lt_settings->link_settings.link_spread = link->dp_ss_off ?
-			LINK_SPREAD_DISABLED : LINK_SPREAD_05_DOWNSPREAD_30KHZ;
-	lt_settings->lttpr_mode = link->lttpr_mode;
-	lt_settings->cr_pattern_time = get_cr_training_aux_rd_interval(link, link_setting);
-	lt_settings->eq_pattern_time = get_eq_training_aux_rd_interval(link, link_setting);
-	lt_settings->pattern_for_cr = decide_cr_training_pattern(link_setting);
-	lt_settings->pattern_for_eq = decide_eq_training_pattern(link, link_setting);
-	lt_settings->enhanced_framing = 1;
-	lt_settings->should_set_fec_ready = true;
-	lt_settings->disallow_per_lane_settings = true;
-	lt_settings->always_match_dpcd_with_hw_lane_settings = true;
-	dp_hw_to_dpcd_lane_settings(lt_settings, lt_settings->hw_lane_settings, lt_settings->dpcd_lane_settings);
-}
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-static inline void decide_128b_132b_training_settings(struct dc_link *link,
-		const struct dc_link_settings *link_settings,
-		struct link_training_settings *lt_settings)
-{
-	memset(lt_settings, 0, sizeof(*lt_settings));
-
-	lt_settings->link_settings = *link_settings;
-	/* TODO: should decide link spread when populating link_settings */
-	lt_settings->link_settings.link_spread = link->dp_ss_off ? LINK_SPREAD_DISABLED :
-			LINK_SPREAD_05_DOWNSPREAD_30KHZ;
-
-	lt_settings->pattern_for_cr = decide_cr_training_pattern(link_settings);
-	lt_settings->pattern_for_eq = decide_eq_training_pattern(link, link_settings);
-	lt_settings->eq_pattern_time = 2500;
-	lt_settings->eq_wait_time_limit = 400000;
-	lt_settings->eq_loop_count_limit = 20;
-	lt_settings->pattern_for_cds = DP_128b_132b_TPS2_CDS;
-	lt_settings->cds_pattern_time = 2500;
-	lt_settings->cds_wait_time_limit = (dp_convert_to_count(
-			link->dpcd_caps.lttpr_caps.phy_repeater_cnt) + 1) * 20000;
-	lt_settings->lttpr_mode = dp_convert_to_count(link->dpcd_caps.lttpr_caps.phy_repeater_cnt) ?
-			LTTPR_MODE_NON_TRANSPARENT : LTTPR_MODE_TRANSPARENT;
-	lt_settings->disallow_per_lane_settings = true;
-	dp_hw_to_dpcd_lane_settings(lt_settings,
-			lt_settings->hw_lane_settings, lt_settings->dpcd_lane_settings);
-}
-#endif
-
-void dp_decide_training_settings(
-		struct dc_link *link,
-		const struct dc_link_settings *link_settings,
-		struct link_training_settings *lt_settings)
-{
-	if (dp_get_link_encoding_format(link_settings) == DP_8b_10b_ENCODING)
-		decide_8b_10b_training_settings(link, link_settings, lt_settings);
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	else if (dp_get_link_encoding_format(link_settings) == DP_128b_132b_ENCODING)
-		decide_128b_132b_training_settings(link, link_settings, lt_settings);
-#endif
-}
-
-static void override_training_settings(
-		struct dc_link *link,
-		const struct dc_link_training_overrides *overrides,
-		struct link_training_settings *lt_settings)
-{
-	uint32_t lane;
-
-	/* Override link spread */
-	if (!link->dp_ss_off && overrides->downspread != NULL)
-		lt_settings->link_settings.link_spread = *overrides->downspread ?
-				LINK_SPREAD_05_DOWNSPREAD_30KHZ
-				: LINK_SPREAD_DISABLED;
-
-	/* Override lane settings */
-	if (overrides->voltage_swing != NULL)
-		lt_settings->voltage_swing = overrides->voltage_swing;
-	if (overrides->pre_emphasis != NULL)
-		lt_settings->pre_emphasis = overrides->pre_emphasis;
-	if (overrides->post_cursor2 != NULL)
-		lt_settings->post_cursor2 = overrides->post_cursor2;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	if (overrides->ffe_preset != NULL)
-		lt_settings->ffe_preset = overrides->ffe_preset;
-#endif
-	/* Override HW lane settings with BIOS forced values if present */
-	if (link->chip_caps & EXT_DISPLAY_PATH_CAPS__DP_FIXED_VS_EN &&
-			link->lttpr_mode == LTTPR_MODE_TRANSPARENT) {
-		lt_settings->voltage_swing = &link->bios_forced_drive_settings.VOLTAGE_SWING;
-		lt_settings->pre_emphasis = &link->bios_forced_drive_settings.PRE_EMPHASIS;
-		lt_settings->always_match_dpcd_with_hw_lane_settings = false;
-	}
-	for (lane = 0; lane < LANE_COUNT_DP_MAX; lane++) {
-		lt_settings->lane_settings[lane].VOLTAGE_SWING =
-			lt_settings->voltage_swing != NULL ?
-			*lt_settings->voltage_swing :
-			VOLTAGE_SWING_LEVEL0;
-		lt_settings->lane_settings[lane].PRE_EMPHASIS =
-			lt_settings->pre_emphasis != NULL ?
-			*lt_settings->pre_emphasis
-			: PRE_EMPHASIS_DISABLED;
-		lt_settings->lane_settings[lane].POST_CURSOR2 =
-			lt_settings->post_cursor2 != NULL ?
-			*lt_settings->post_cursor2
-			: POST_CURSOR2_DISABLED;
-	}
-
-	dp_hw_to_dpcd_lane_settings(lt_settings,
-			lt_settings->hw_lane_settings, lt_settings->dpcd_lane_settings);
-
-	/* Initialize training timings */
-	if (overrides->cr_pattern_time != NULL)
-		lt_settings->cr_pattern_time = *overrides->cr_pattern_time;
-
-	if (overrides->eq_pattern_time != NULL)
-		lt_settings->eq_pattern_time = *overrides->eq_pattern_time;
-
-	if (overrides->pattern_for_cr != NULL)
-		lt_settings->pattern_for_cr = *overrides->pattern_for_cr;
-	if (overrides->pattern_for_eq != NULL)
-		lt_settings->pattern_for_eq = *overrides->pattern_for_eq;
-
-	if (overrides->enhanced_framing != NULL)
-		lt_settings->enhanced_framing = *overrides->enhanced_framing;
-
-	if (link->preferred_training_settings.fec_enable != NULL)
-		lt_settings->should_set_fec_ready = *link->preferred_training_settings.fec_enable;
-}
-
-uint8_t dp_convert_to_count(uint8_t lttpr_repeater_count)
-{
-	switch (lttpr_repeater_count) {
-	case 0x80: // 1 lttpr repeater
-		return 1;
-	case 0x40: // 2 lttpr repeaters
-		return 2;
-	case 0x20: // 3 lttpr repeaters
-		return 3;
-	case 0x10: // 4 lttpr repeaters
-		return 4;
-	case 0x08: // 5 lttpr repeaters
-		return 5;
-	case 0x04: // 6 lttpr repeaters
-		return 6;
-	case 0x02: // 7 lttpr repeaters
-		return 7;
-	case 0x01: // 8 lttpr repeaters
-		return 8;
-	default:
-		break;
-	}
-	return 0; // invalid value
-}
-
-static enum dc_status configure_lttpr_mode_transparent(struct dc_link *link)
-{
-	uint8_t repeater_mode = DP_PHY_REPEATER_MODE_TRANSPARENT;
-
-	DC_LOG_HW_LINK_TRAINING("%s\n Set LTTPR to Transparent Mode\n", __func__);
-	return core_link_write_dpcd(link,
-			DP_PHY_REPEATER_MODE,
-			(uint8_t *)&repeater_mode,
-			sizeof(repeater_mode));
-}
-
-static enum dc_status configure_lttpr_mode_non_transparent(
-		struct dc_link *link,
-		const struct link_training_settings *lt_settings)
-{
-	/* aux timeout is already set to extended */
-	/* RESET/SET lttpr mode to enable non transparent mode */
-	uint8_t repeater_cnt;
-	uint32_t aux_interval_address;
-	uint8_t repeater_id;
-	enum dc_status result = DC_ERROR_UNEXPECTED;
-	uint8_t repeater_mode = DP_PHY_REPEATER_MODE_TRANSPARENT;
-
-	enum dp_link_encoding encoding = dp_get_link_encoding_format(&lt_settings->link_settings);
-
-	if (encoding == DP_8b_10b_ENCODING) {
-		DC_LOG_HW_LINK_TRAINING("%s\n Set LTTPR to Transparent Mode\n", __func__);
-		result = core_link_write_dpcd(link,
-				DP_PHY_REPEATER_MODE,
-				(uint8_t *)&repeater_mode,
-				sizeof(repeater_mode));
-
-	}
-
-	if (result == DC_OK) {
-		link->dpcd_caps.lttpr_caps.mode = repeater_mode;
-	}
-
-	if (link->lttpr_mode == LTTPR_MODE_NON_TRANSPARENT) {
-
-		DC_LOG_HW_LINK_TRAINING("%s\n Set LTTPR to Non Transparent Mode\n", __func__);
-
-		repeater_mode = DP_PHY_REPEATER_MODE_NON_TRANSPARENT;
-		result = core_link_write_dpcd(link,
-				DP_PHY_REPEATER_MODE,
-				(uint8_t *)&repeater_mode,
-				sizeof(repeater_mode));
-
-		if (result == DC_OK) {
-			link->dpcd_caps.lttpr_caps.mode = repeater_mode;
-		}
-
-		if (encoding == DP_8b_10b_ENCODING) {
-			repeater_cnt = dp_convert_to_count(link->dpcd_caps.lttpr_caps.phy_repeater_cnt);
-
-			/* Driver does not need to train the first hop. Skip DPCD read and clear
-			 * AUX_RD_INTERVAL for DPTX-to-DPIA hop.
-			 */
-			if (link->ep_type == DISPLAY_ENDPOINT_USB4_DPIA)
-				link->dpcd_caps.lttpr_caps.aux_rd_interval[--repeater_cnt] = 0;
-
-			for (repeater_id = repeater_cnt; repeater_id > 0; repeater_id--) {
-				aux_interval_address = DP_TRAINING_AUX_RD_INTERVAL_PHY_REPEATER1 +
-							((DP_REPEATER_CONFIGURATION_AND_STATUS_SIZE) * (repeater_id - 1));
-				core_link_read_dpcd(
-					link,
-					aux_interval_address,
-					(uint8_t *)&link->dpcd_caps.lttpr_caps.aux_rd_interval[repeater_id - 1],
-					sizeof(link->dpcd_caps.lttpr_caps.aux_rd_interval[repeater_id - 1]));
-				link->dpcd_caps.lttpr_caps.aux_rd_interval[repeater_id - 1] &= 0x7F;
-			}
-		}
-	}
-
-	return result;
-}
-
-static void repeater_training_done(struct dc_link *link, uint32_t offset)
-{
-	union dpcd_training_pattern dpcd_pattern = {0};
-
-	const uint32_t dpcd_base_lt_offset =
-			DP_TRAINING_PATTERN_SET_PHY_REPEATER1 +
-				((DP_REPEATER_CONFIGURATION_AND_STATUS_SIZE) * (offset - 1));
-	/* Set training not in progress*/
-	dpcd_pattern.v1_4.TRAINING_PATTERN_SET = DPCD_TRAINING_PATTERN_VIDEOIDLE;
-
-	core_link_write_dpcd(
-		link,
-		dpcd_base_lt_offset,
-		&dpcd_pattern.raw,
-		1);
-
-	DC_LOG_HW_LINK_TRAINING("%s\n LTTPR Id: %d 0x%X pattern = %x\n",
-		__func__,
-		offset,
-		dpcd_base_lt_offset,
-		dpcd_pattern.v1_4.TRAINING_PATTERN_SET);
-}
-
-static void print_status_message(
-	struct dc_link *link,
-	const struct link_training_settings *lt_settings,
-	enum link_training_result status)
-{
-	char *link_rate = "Unknown";
-	char *lt_result = "Unknown";
-	char *lt_spread = "Disabled";
-
-	switch (lt_settings->link_settings.link_rate) {
-	case LINK_RATE_LOW:
-		link_rate = "RBR";
-		break;
-	case LINK_RATE_RATE_2:
-		link_rate = "R2";
-		break;
-	case LINK_RATE_RATE_3:
-		link_rate = "R3";
-		break;
-	case LINK_RATE_HIGH:
-		link_rate = "HBR";
-		break;
-	case LINK_RATE_RBR2:
-		link_rate = "RBR2";
-		break;
-	case LINK_RATE_RATE_6:
-		link_rate = "R6";
-		break;
-	case LINK_RATE_HIGH2:
-		link_rate = "HBR2";
-		break;
-	case LINK_RATE_HIGH3:
-		link_rate = "HBR3";
-		break;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	case LINK_RATE_UHBR10:
-		link_rate = "UHBR10";
-		break;
-	case LINK_RATE_UHBR13_5:
-		link_rate = "UHBR13.5";
-		break;
-	case LINK_RATE_UHBR20:
-		link_rate = "UHBR20";
-		break;
-#endif
-	default:
-		break;
-	}
-
-	switch (status) {
-	case LINK_TRAINING_SUCCESS:
-		lt_result = "pass";
-		break;
-	case LINK_TRAINING_CR_FAIL_LANE0:
-		lt_result = "CR failed lane0";
-		break;
-	case LINK_TRAINING_CR_FAIL_LANE1:
-		lt_result = "CR failed lane1";
-		break;
-	case LINK_TRAINING_CR_FAIL_LANE23:
-		lt_result = "CR failed lane23";
-		break;
-	case LINK_TRAINING_EQ_FAIL_CR:
-		lt_result = "CR failed in EQ";
-		break;
-	case LINK_TRAINING_EQ_FAIL_EQ:
-		lt_result = "EQ failed";
-		break;
-	case LINK_TRAINING_LQA_FAIL:
-		lt_result = "LQA failed";
-		break;
-	case LINK_TRAINING_LINK_LOSS:
-		lt_result = "Link loss";
-		break;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	case DP_128b_132b_LT_FAILED:
-		lt_result = "LT_FAILED received";
-		break;
-	case DP_128b_132b_MAX_LOOP_COUNT_REACHED:
-		lt_result = "max loop count reached";
-		break;
-	case DP_128b_132b_CHANNEL_EQ_DONE_TIMEOUT:
-		lt_result = "channel EQ timeout";
-		break;
-	case DP_128b_132b_CDS_DONE_TIMEOUT:
-		lt_result = "CDS timeout";
-		break;
-#endif
-	default:
-		break;
-	}
-
-	switch (lt_settings->link_settings.link_spread) {
-	case LINK_SPREAD_DISABLED:
-		lt_spread = "Disabled";
-		break;
-	case LINK_SPREAD_05_DOWNSPREAD_30KHZ:
-		lt_spread = "0.5% 30KHz";
-		break;
-	case LINK_SPREAD_05_DOWNSPREAD_33KHZ:
-		lt_spread = "0.5% 33KHz";
-		break;
-	default:
-		break;
-	}
-
-	/* Connectivity log: link training */
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	/* TODO - DP2.0 Log: add connectivity log for FFE PRESET */
-#endif
-	CONN_MSG_LT(link, "%sx%d %s VS=%d, PE=%d, DS=%s",
-				link_rate,
-				lt_settings->link_settings.lane_count,
-				lt_result,
-				lt_settings->lane_settings[0].VOLTAGE_SWING,
-				lt_settings->lane_settings[0].PRE_EMPHASIS,
-				lt_spread);
-}
-
-void dc_link_dp_set_drive_settings(
-	struct dc_link *link,
-	struct link_training_settings *lt_settings)
-{
-	/* program ASIC PHY settings*/
-	dp_set_hw_lane_settings(link, lt_settings, DPRX);
-
-	dp_hw_to_dpcd_lane_settings(lt_settings,
-			lt_settings->hw_lane_settings, lt_settings->dpcd_lane_settings);
-
-	/* Notify DP sink the PHY settings from source */
-	dpcd_set_lane_settings(link, lt_settings, DPRX);
-}
-
-bool dc_link_dp_perform_link_training_skip_aux(
-	struct dc_link *link,
-	const struct dc_link_settings *link_setting)
-{
-	struct link_training_settings lt_settings = {0};
-
-	dp_decide_training_settings(
-			link,
-			link_setting,
-			&lt_settings);
-	override_training_settings(
-			link,
-			&link->preferred_training_settings,
-			&lt_settings);
-
-	/* 1. Perform_clock_recovery_sequence. */
-
-	/* transmit training pattern for clock recovery */
-	dp_set_hw_training_pattern(link, lt_settings.pattern_for_cr, DPRX);
-
-	/* call HWSS to set lane settings*/
-	dp_set_hw_lane_settings(link, &lt_settings, DPRX);
-
-	/* wait receiver to lock-on*/
-	dp_wait_for_training_aux_rd_interval(link, lt_settings.cr_pattern_time);
-
-	/* 2. Perform_channel_equalization_sequence. */
-
-	/* transmit training pattern for channel equalization. */
-	dp_set_hw_training_pattern(link, lt_settings.pattern_for_eq, DPRX);
-
-	/* call HWSS to set lane settings*/
-	dp_set_hw_lane_settings(link, &lt_settings, DPRX);
-
-	/* wait receiver to lock-on. */
-	dp_wait_for_training_aux_rd_interval(link, lt_settings.eq_pattern_time);
-
-	/* 3. Perform_link_training_int. */
-
-	/* Mainlink output idle pattern. */
-	dp_set_hw_test_pattern(link, DP_TEST_PATTERN_VIDEO_MODE, NULL, 0);
-
-	print_status_message(link, &lt_settings, LINK_TRAINING_SUCCESS);
-
-	return true;
-}
-
-enum dc_status dpcd_configure_lttpr_mode(struct dc_link *link, struct link_training_settings *lt_settings)
-{
-	enum dc_status status = DC_OK;
-
-	if (lt_settings->lttpr_mode == LTTPR_MODE_TRANSPARENT)
-		status = configure_lttpr_mode_transparent(link);
-
-	else if (lt_settings->lttpr_mode == LTTPR_MODE_NON_TRANSPARENT)
-		status = configure_lttpr_mode_non_transparent(link, lt_settings);
-
-	return status;
-}
-
-static void dpcd_exit_training_mode(struct dc_link *link)
-{
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	uint8_t sink_status = 0;
-	uint8_t i;
-#endif
-
-	/* clear training pattern set */
-	dpcd_set_training_pattern(link, DP_TRAINING_PATTERN_VIDEOIDLE);
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	/* poll for intra-hop disable */
-	for (i = 0; i < 10; i++) {
-		if ((core_link_read_dpcd(link, DP_SINK_STATUS, &sink_status, 1) == DC_OK) &&
-				(sink_status & DP_INTRA_HOP_AUX_REPLY_INDICATION) == 0)
-			break;
-		udelay(1000);
-	}
-#endif
-}
-
-enum dc_status dpcd_configure_channel_coding(struct dc_link *link,
-		struct link_training_settings *lt_settings)
-{
-	enum dp_link_encoding encoding =
-			dp_get_link_encoding_format(
-					&lt_settings->link_settings);
-	enum dc_status status;
-
-	status = core_link_write_dpcd(
-			link,
-			DP_MAIN_LINK_CHANNEL_CODING_SET,
-			(uint8_t *) &encoding,
-			1);
-	DC_LOG_HW_LINK_TRAINING("%s:\n 0x%X MAIN_LINK_CHANNEL_CODING_SET = %x\n",
-					__func__,
-					DP_MAIN_LINK_CHANNEL_CODING_SET,
-					encoding);
-
-	return status;
-}
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-static void dpcd_128b_132b_get_aux_rd_interval(struct dc_link *link,
-		uint32_t *interval_in_us)
-{
-	union dp_128b_132b_training_aux_rd_interval dpcd_interval;
-	uint32_t interval_unit = 0;
-
-	dpcd_interval.raw = 0;
-	core_link_read_dpcd(link, DP_128b_132b_TRAINING_AUX_RD_INTERVAL,
-			&dpcd_interval.raw, sizeof(dpcd_interval.raw));
-	interval_unit = dpcd_interval.bits.UNIT ? 1 : 2; /* 0b = 2 ms, 1b = 1 ms */
-	/* (128b/132b_TRAINING_AUX_RD_INTERVAL value + 1) *
-	 * INTERVAL_UNIT. The maximum is 256 ms
-	 */
-	*interval_in_us = (dpcd_interval.bits.VALUE + 1) * interval_unit * 1000;
-}
-
-static enum link_training_result dp_perform_128b_132b_channel_eq_done_sequence(
-		struct dc_link *link,
-		struct link_training_settings *lt_settings)
-{
-	uint8_t loop_count;
-	uint32_t aux_rd_interval = 0;
-	uint32_t wait_time = 0;
-	union lane_align_status_updated dpcd_lane_status_updated = {0};
-	union lane_status dpcd_lane_status[LANE_COUNT_DP_MAX] = {0};
-	enum link_training_result status = LINK_TRAINING_SUCCESS;
-	union lane_adjust dpcd_lane_adjust[LANE_COUNT_DP_MAX] = {0};
-
-	/* Transmit 128b/132b_TPS1 over Main-Link */
-	dp_set_hw_training_pattern(link, lt_settings->pattern_for_cr, DPRX);
-	/* Set TRAINING_PATTERN_SET to 01h */
-	dpcd_set_training_pattern(link, lt_settings->pattern_for_cr);
-
-	/* Adjust TX_FFE_PRESET_VALUE and Transmit 128b/132b_TPS2 over Main-Link */
-	dpcd_128b_132b_get_aux_rd_interval(link, &aux_rd_interval);
-	dp_get_lane_status_and_lane_adjust(link, lt_settings, dpcd_lane_status,
-			&dpcd_lane_status_updated, dpcd_lane_adjust, DPRX);
-	dp_decide_lane_settings(lt_settings, dpcd_lane_adjust,
-			lt_settings->hw_lane_settings, lt_settings->dpcd_lane_settings);
-	dp_set_hw_lane_settings(link, lt_settings, DPRX);
-	dp_set_hw_training_pattern(link, lt_settings->pattern_for_eq, DPRX);
-
-	/* Set loop counter to start from 1 */
-	loop_count = 1;
-
-	/* Set TRAINING_PATTERN_SET to 02h and TX_FFE_PRESET_VALUE in one AUX transaction */
-	dpcd_set_lt_pattern_and_lane_settings(link, lt_settings,
-			lt_settings->pattern_for_eq, DPRX);
-
-	/* poll for channel EQ done */
-	while (status == LINK_TRAINING_SUCCESS) {
-		dp_wait_for_training_aux_rd_interval(link, aux_rd_interval);
-		wait_time += aux_rd_interval;
-		dp_get_lane_status_and_lane_adjust(link, lt_settings, dpcd_lane_status,
-				&dpcd_lane_status_updated, dpcd_lane_adjust, DPRX);
-		dp_decide_lane_settings(lt_settings, dpcd_lane_adjust,
-			lt_settings->hw_lane_settings, lt_settings->dpcd_lane_settings);
-		dpcd_128b_132b_get_aux_rd_interval(link, &aux_rd_interval);
-		if (dp_is_ch_eq_done(lt_settings->link_settings.lane_count,
-				dpcd_lane_status)) {
-			/* pass */
-			break;
-		} else if (loop_count >= lt_settings->eq_loop_count_limit) {
-			status = DP_128b_132b_MAX_LOOP_COUNT_REACHED;
-		} else if (dpcd_lane_status_updated.bits.LT_FAILED_128b_132b) {
-			status = DP_128b_132b_LT_FAILED;
-		} else {
-			dp_set_hw_lane_settings(link, lt_settings, DPRX);
-			dpcd_set_lane_settings(link, lt_settings, DPRX);
-		}
-		loop_count++;
-	}
-
-	/* poll for EQ interlane align done */
-	while (status == LINK_TRAINING_SUCCESS) {
-		if (dpcd_lane_status_updated.bits.EQ_INTERLANE_ALIGN_DONE_128b_132b) {
-			/* pass */
-			break;
-		} else if (wait_time >= lt_settings->eq_wait_time_limit) {
-			status = DP_128b_132b_CHANNEL_EQ_DONE_TIMEOUT;
-		} else if (dpcd_lane_status_updated.bits.LT_FAILED_128b_132b) {
-			status = DP_128b_132b_LT_FAILED;
-		} else {
-			dp_wait_for_training_aux_rd_interval(link,
-					lt_settings->eq_pattern_time);
-			wait_time += lt_settings->eq_pattern_time;
-			dp_get_lane_status_and_lane_adjust(link, lt_settings, dpcd_lane_status,
-					&dpcd_lane_status_updated, dpcd_lane_adjust, DPRX);
-		}
-	}
-
-	return status;
-}
-
-static enum link_training_result dp_perform_128b_132b_cds_done_sequence(
-		struct dc_link *link,
-		struct link_training_settings *lt_settings)
-{
-	/* Assumption: assume hardware has transmitted eq pattern */
-	enum link_training_result status = LINK_TRAINING_SUCCESS;
-	union lane_align_status_updated dpcd_lane_status_updated = {0};
-	union lane_status dpcd_lane_status[LANE_COUNT_DP_MAX] = {0};
-	union lane_adjust dpcd_lane_adjust[LANE_COUNT_DP_MAX] = { { {0} } };
-	uint32_t wait_time = 0;
-
-	/* initiate CDS done sequence */
-	dpcd_set_training_pattern(link, lt_settings->pattern_for_cds);
-
-	/* poll for CDS interlane align done and symbol lock */
-	while (status == LINK_TRAINING_SUCCESS) {
-		dp_wait_for_training_aux_rd_interval(link,
-				lt_settings->cds_pattern_time);
-		wait_time += lt_settings->cds_pattern_time;
-		dp_get_lane_status_and_lane_adjust(link, lt_settings, dpcd_lane_status,
-						&dpcd_lane_status_updated, dpcd_lane_adjust, DPRX);
-		if (dp_is_symbol_locked(lt_settings->link_settings.lane_count, dpcd_lane_status) &&
-				dpcd_lane_status_updated.bits.CDS_INTERLANE_ALIGN_DONE_128b_132b) {
-			/* pass */
-			break;
-		} else if (dpcd_lane_status_updated.bits.LT_FAILED_128b_132b) {
-			status = DP_128b_132b_LT_FAILED;
-		} else if (wait_time >= lt_settings->cds_wait_time_limit) {
-			status = DP_128b_132b_CDS_DONE_TIMEOUT;
-		}
-	}
-
-	return status;
-}
-#endif
-
-static enum link_training_result dp_perform_8b_10b_link_training(
-		struct dc_link *link,
-		struct link_training_settings *lt_settings)
-{
-	enum link_training_result status = LINK_TRAINING_SUCCESS;
-
-	uint8_t repeater_cnt;
-	uint8_t repeater_id;
-	uint8_t lane = 0;
-
-	if (link->ctx->dc->work_arounds.lt_early_cr_pattern)
-		start_clock_recovery_pattern_early(link, lt_settings, DPRX);
-
-	/* 1. set link rate, lane count and spread. */
-	dpcd_set_link_settings(link, lt_settings);
-
-	if (link->lttpr_mode == LTTPR_MODE_NON_TRANSPARENT) {
-
-		/* 2. perform link training (set link training done
-		 *  to false is done as well)
-		 */
-		repeater_cnt = dp_convert_to_count(link->dpcd_caps.lttpr_caps.phy_repeater_cnt);
-
-		for (repeater_id = repeater_cnt; (repeater_id > 0 && status == LINK_TRAINING_SUCCESS);
-				repeater_id--) {
-			status = perform_clock_recovery_sequence(link, lt_settings, repeater_id);
-
-			if (status != LINK_TRAINING_SUCCESS)
-				break;
-
-			status = perform_channel_equalization_sequence(link,
-					lt_settings,
-					repeater_id);
-
-			if (status != LINK_TRAINING_SUCCESS)
-				break;
-
-			repeater_training_done(link, repeater_id);
-		}
-
-		for (lane = 0; lane < (uint8_t)lt_settings->link_settings.lane_count; lane++)
-			lt_settings->dpcd_lane_settings[lane].raw = 0;
-	}
-
-	if (status == LINK_TRAINING_SUCCESS) {
-		status = perform_clock_recovery_sequence(link, lt_settings, DPRX);
-	if (status == LINK_TRAINING_SUCCESS) {
-		status = perform_channel_equalization_sequence(link,
-					lt_settings,
-					DPRX);
-		}
-	}
-
-	return status;
-}
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-static enum link_training_result dp_perform_128b_132b_link_training(
-		struct dc_link *link,
-		struct link_training_settings *lt_settings)
-{
-	enum link_training_result result = LINK_TRAINING_SUCCESS;
-
-	/* TODO - DP2.0 Link: remove legacy_dp2_lt logic */
-	if (link->dc->debug.legacy_dp2_lt) {
-		struct link_training_settings legacy_settings;
-
-		decide_8b_10b_training_settings(link,
-				&lt_settings->link_settings,
-				&legacy_settings);
-		return dp_perform_8b_10b_link_training(link, &legacy_settings);
-	}
-
-	dpcd_set_link_settings(link, lt_settings);
-
-	if (result == LINK_TRAINING_SUCCESS)
-		result = dp_perform_128b_132b_channel_eq_done_sequence(link, lt_settings);
-
-	if (result == LINK_TRAINING_SUCCESS)
-		result = dp_perform_128b_132b_cds_done_sequence(link, lt_settings);
-
-	return result;
-}
-#endif
-
-enum link_training_result dc_link_dp_perform_link_training(
-	struct dc_link *link,
-	const struct dc_link_settings *link_settings,
-	bool skip_video_pattern)
-{
-	enum link_training_result status = LINK_TRAINING_SUCCESS;
-	struct link_training_settings lt_settings = {0};
-	enum dp_link_encoding encoding =
-			dp_get_link_encoding_format(link_settings);
-
-	/* decide training settings */
-	dp_decide_training_settings(
-			link,
-			link_settings,
-			&lt_settings);
-	override_training_settings(
-			link,
-			&link->preferred_training_settings,
-			&lt_settings);
-
-	/* reset previous training states */
-	dpcd_exit_training_mode(link);
-
-	/* configure link prior to entering training mode */
-	dpcd_configure_lttpr_mode(link, &lt_settings);
-	dp_set_fec_ready(link, lt_settings.should_set_fec_ready);
-	dpcd_configure_channel_coding(link, &lt_settings);
-
-	/* enter training mode:
-	 * Per DP specs starting from here, DPTX device shall not issue
-	 * Non-LT AUX transactions inside training mode.
-	 */
-	if (encoding == DP_8b_10b_ENCODING)
-		status = dp_perform_8b_10b_link_training(link, &lt_settings);
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	else if (encoding == DP_128b_132b_ENCODING)
-		status = dp_perform_128b_132b_link_training(link, &lt_settings);
-#endif
-	else
-		ASSERT(0);
-
-	/* exit training mode and switch to video idle */
-	dpcd_exit_training_mode(link);
-	if ((status == LINK_TRAINING_SUCCESS) || !skip_video_pattern)
-		status = dp_transition_to_video_idle(link,
-				&lt_settings,
-				status);
-
-	/* dump debug data */
-	print_status_message(link, &lt_settings, status);
-	if (status != LINK_TRAINING_SUCCESS)
-		link->ctx->dc->debug_data.ltFailCount++;
-	return status;
-}
-
-bool perform_link_training_with_retries(
-	const struct dc_link_settings *link_setting,
-	bool skip_video_pattern,
-	int attempts,
-	struct pipe_ctx *pipe_ctx,
-	enum signal_type signal,
-	bool do_fallback)
-{
-	int j;
-	uint8_t delay_between_attempts = LINK_TRAINING_RETRY_DELAY;
-	struct dc_stream_state *stream = pipe_ctx->stream;
-	struct dc_link *link = stream->link;
-	enum dp_panel_mode panel_mode = dp_get_panel_mode(link);
-	struct link_encoder *link_enc;
-	enum link_training_result status = LINK_TRAINING_CR_FAIL_LANE0;
-	struct dc_link_settings current_setting = *link_setting;
-
-	/* Dynamically assigned link encoders associated with stream rather than
-	 * link.
-	 */
-	if (link->is_dig_mapping_flexible && link->dc->res_pool->funcs->link_encs_assign)
-		link_enc = link_enc_cfg_get_link_enc_used_by_stream(link->ctx->dc, pipe_ctx->stream);
-	else
-		link_enc = link->link_enc;
-
-	/* We need to do this before the link training to ensure the idle pattern in SST
-	 * mode will be sent right after the link training
-	 */
-	if (dp_get_link_encoding_format(&current_setting) == DP_8b_10b_ENCODING) {
-		link_enc->funcs->connect_dig_be_to_fe(link_enc,
-							pipe_ctx->stream_res.stream_enc->id, true);
-		dp_source_sequence_trace(link, DPCD_SOURCE_SEQ_AFTER_CONNECT_DIG_FE_BE);
-	}
-
-	for (j = 0; j < attempts; ++j) {
-
-		DC_LOG_HW_LINK_TRAINING("%s: Beginning link training attempt %u of %d\n",
-			__func__, (unsigned int)j + 1, attempts);
-
-		dp_enable_link_phy(
-			link,
-			signal,
-			pipe_ctx->clock_source->id,
-			&current_setting);
-
-		if (stream->sink_patches.dppowerup_delay > 0) {
-			int delay_dp_power_up_in_ms = stream->sink_patches.dppowerup_delay;
-
-			msleep(delay_dp_power_up_in_ms);
-		}
-
-#ifdef CONFIG_DRM_AMD_DC_HDCP
-		if (panel_mode == DP_PANEL_MODE_EDP) {
-			struct cp_psp *cp_psp = &stream->ctx->cp_psp;
-
-			if (cp_psp && cp_psp->funcs.enable_assr)
-				/* ASSR is bound to fail with unsigned PSP
-				 * verstage used during devlopment phase.
-				 * Report and continue with eDP panel mode to
-				 * perform eDP link training with right settings
-				 */
-				cp_psp->funcs.enable_assr(cp_psp->handle, link);
-		}
-#endif
-
-		dp_set_panel_mode(link, panel_mode);
-
-		if (link->aux_access_disabled) {
-			dc_link_dp_perform_link_training_skip_aux(link, &current_setting);
-			return true;
-		} else {
-			/** @todo Consolidate USB4 DP and DPx.x training. */
-			if (link->ep_type == DISPLAY_ENDPOINT_USB4_DPIA) {
-				status = dc_link_dpia_perform_link_training(link,
-									    &current_setting,
-									    skip_video_pattern);
-
-				/* Transmit idle pattern once training successful. */
-				if (status == LINK_TRAINING_SUCCESS)
-					dp_set_hw_test_pattern(link, DP_TEST_PATTERN_VIDEO_MODE,
-							       NULL, 0);
-			} else {
-				status = dc_link_dp_perform_link_training(link,
-									  &current_setting,
-									  skip_video_pattern);
-			}
-
-			if (status == LINK_TRAINING_SUCCESS)
-				return true;
-		}
-
-		/* latest link training still fail, skip delay and keep PHY on
-		 */
-		if (j == (attempts - 1) && link->ep_type == DISPLAY_ENDPOINT_PHY)
-			break;
-
-		DC_LOG_WARNING("%s: Link training attempt %u of %d failed\n",
-			__func__, (unsigned int)j + 1, attempts);
-
-		dp_disable_link_phy(link, signal);
-
-		/* Abort link training if failure due to sink being unplugged. */
-		if (status == LINK_TRAINING_ABORT) {
-			enum dc_connection_type type = dc_connection_none;
-
-			dc_link_detect_sink(link, &type);
-			if (type == dc_connection_none)
-				break;
-		} else if (do_fallback) {
-			uint32_t req_bw;
-			uint32_t link_bw;
-
-			decide_fallback_link_setting(*link_setting, &current_setting, status);
-			/* Fail link training if reduced link bandwidth no longer meets
-			 * stream requirements.
-			 */
-			req_bw = dc_bandwidth_in_kbps_from_timing(&stream->timing);
-			link_bw = dc_link_bandwidth_kbps(link, &current_setting);
-			if (req_bw > link_bw)
-				break;
-		}
-
-		msleep(delay_between_attempts);
-
-		delay_between_attempts += LINK_TRAINING_RETRY_DELAY;
-	}
-
-	return false;
-}
-
-static enum clock_source_id get_clock_source_id(struct dc_link *link)
-{
-	enum clock_source_id dp_cs_id = CLOCK_SOURCE_ID_UNDEFINED;
-	struct clock_source *dp_cs = link->dc->res_pool->dp_clock_source;
-
-	if (dp_cs != NULL) {
-		dp_cs_id = dp_cs->id;
-	} else {
-		/*
-		 * dp clock source is not initialized for some reason.
-		 * Should not happen, CLOCK_SOURCE_ID_EXTERNAL will be used
-		 */
-		ASSERT(dp_cs);
-	}
-
-	return dp_cs_id;
-}
-
-static void set_dp_mst_mode(struct dc_link *link, bool mst_enable)
-{
-	if (mst_enable == false &&
-		link->type == dc_connection_mst_branch) {
-		/* Disable MST on link. Use only local sink. */
-		dp_disable_link_phy_mst(link, link->connector_signal);
-
-		link->type = dc_connection_single;
-		link->local_sink = link->remote_sinks[0];
-		link->local_sink->sink_signal = SIGNAL_TYPE_DISPLAY_PORT;
-		dc_sink_retain(link->local_sink);
-		dm_helpers_dp_mst_stop_top_mgr(link->ctx, link);
-	} else if (mst_enable == true &&
-			link->type == dc_connection_single &&
-			link->remote_sinks[0] != NULL) {
-		/* Re-enable MST on link. */
-		dp_disable_link_phy(link, link->connector_signal);
-		dp_enable_mst_on_sink(link, true);
-
-		link->type = dc_connection_mst_branch;
-		link->local_sink->sink_signal = SIGNAL_TYPE_DISPLAY_PORT_MST;
-	}
-}
-
-bool dc_link_dp_sync_lt_begin(struct dc_link *link)
-{
-	/* Begin Sync LT. During this time,
-	 * DPCD:600h must not be powered down.
-	 */
-	link->sync_lt_in_progress = true;
-
-	/*Clear any existing preferred settings.*/
-	memset(&link->preferred_training_settings, 0,
-		sizeof(struct dc_link_training_overrides));
-	memset(&link->preferred_link_setting, 0,
-		sizeof(struct dc_link_settings));
-
-	return true;
-}
-
-enum link_training_result dc_link_dp_sync_lt_attempt(
-    struct dc_link *link,
-    struct dc_link_settings *link_settings,
-    struct dc_link_training_overrides *lt_overrides)
-{
-	struct link_training_settings lt_settings = {0};
-	enum link_training_result lt_status = LINK_TRAINING_SUCCESS;
-	enum dp_panel_mode panel_mode = DP_PANEL_MODE_DEFAULT;
-	enum clock_source_id dp_cs_id = CLOCK_SOURCE_ID_EXTERNAL;
-	bool fec_enable = false;
-
-	dp_decide_training_settings(
-			link,
-			link_settings,
-			&lt_settings);
-	override_training_settings(
-			link,
-			lt_overrides,
-			&lt_settings);
-	/* Setup MST Mode */
-	if (lt_overrides->mst_enable)
-		set_dp_mst_mode(link, *lt_overrides->mst_enable);
-
-	/* Disable link */
-	dp_disable_link_phy(link, link->connector_signal);
-
-	/* Enable link */
-	dp_cs_id = get_clock_source_id(link);
-	dp_enable_link_phy(link, link->connector_signal,
-		dp_cs_id, link_settings);
-
-	/* Set FEC enable */
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	if (dp_get_link_encoding_format(link_settings) == DP_8b_10b_ENCODING) {
-#endif
-		fec_enable = lt_overrides->fec_enable && *lt_overrides->fec_enable;
-		dp_set_fec_ready(link, fec_enable);
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	}
-#endif
-
-	if (lt_overrides->alternate_scrambler_reset) {
-		if (*lt_overrides->alternate_scrambler_reset)
-			panel_mode = DP_PANEL_MODE_EDP;
-		else
-			panel_mode = DP_PANEL_MODE_DEFAULT;
-	} else
-		panel_mode = dp_get_panel_mode(link);
-
-	dp_set_panel_mode(link, panel_mode);
-
-	/* Attempt to train with given link training settings */
-	if (link->ctx->dc->work_arounds.lt_early_cr_pattern)
-		start_clock_recovery_pattern_early(link, &lt_settings, DPRX);
-
-	/* Set link rate, lane count and spread. */
-	dpcd_set_link_settings(link, &lt_settings);
-
-	/* 2. perform link training (set link training done
-	 *  to false is done as well)
-	 */
-	lt_status = perform_clock_recovery_sequence(link, &lt_settings, DPRX);
-	if (lt_status == LINK_TRAINING_SUCCESS) {
-		lt_status = perform_channel_equalization_sequence(link,
-						&lt_settings,
-						DPRX);
-	}
-
-	/* 3. Sync LT must skip TRAINING_PATTERN_SET:0 (video pattern)*/
-	/* 4. print status message*/
-	print_status_message(link, &lt_settings, lt_status);
-
-	return lt_status;
-}
-
-bool dc_link_dp_sync_lt_end(struct dc_link *link, bool link_down)
-{
-	/* If input parameter is set, shut down phy.
-	 * Still shouldn't turn off dp_receiver (DPCD:600h)
-	 */
-	if (link_down == true) {
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-		struct dc_link_settings link_settings = link->cur_link_settings;
-#endif
-		dp_disable_link_phy(link, link->connector_signal);
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-		if (dp_get_link_encoding_format(&link_settings) == DP_8b_10b_ENCODING)
-#endif
-			dp_set_fec_ready(link, false);
-	}
-
-	link->sync_lt_in_progress = false;
-	return true;
-}
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-static enum dc_link_rate get_lttpr_max_link_rate(struct dc_link *link)
-{
-	enum dc_link_rate lttpr_max_link_rate = link->dpcd_caps.lttpr_caps.max_link_rate;
-
-	if (link->dpcd_caps.lttpr_caps.supported_128b_132b_rates.bits.UHBR20)
-		lttpr_max_link_rate = LINK_RATE_UHBR20;
-	else if (link->dpcd_caps.lttpr_caps.supported_128b_132b_rates.bits.UHBR13_5)
-		lttpr_max_link_rate = LINK_RATE_UHBR13_5;
-	else if (link->dpcd_caps.lttpr_caps.supported_128b_132b_rates.bits.UHBR10)
-		lttpr_max_link_rate = LINK_RATE_UHBR10;
-
-	return lttpr_max_link_rate;
-}
-#endif
-
-bool dc_link_dp_get_max_link_enc_cap(const struct dc_link *link, struct dc_link_settings *max_link_enc_cap)
-{
-	struct link_encoder *link_enc = NULL;
-
-	if (!max_link_enc_cap) {
-		DC_LOG_ERROR("%s: Could not return max link encoder caps", __func__);
-		return false;
-	}
-
-	/* Links supporting dynamically assigned link encoder will be assigned next
-	 * available encoder if one not already assigned.
-	 */
-	if (link->is_dig_mapping_flexible &&
-			link->dc->res_pool->funcs->link_encs_assign) {
-		link_enc = link_enc_cfg_get_link_enc_used_by_link(link->ctx->dc, link);
-		if (link_enc == NULL)
-			link_enc = link_enc_cfg_get_next_avail_link_enc(link->ctx->dc);
-	} else
-		link_enc = link->link_enc;
-	ASSERT(link_enc);
-
-	if (link_enc && link_enc->funcs->get_max_link_cap) {
-		link_enc->funcs->get_max_link_cap(link_enc, max_link_enc_cap);
-		return true;
-	}
-
-	DC_LOG_ERROR("%s: Max link encoder caps unknown", __func__);
-	max_link_enc_cap->lane_count = 1;
-	max_link_enc_cap->link_rate = 6;
-	return false;
-}
-
-static struct dc_link_settings get_max_link_cap(struct dc_link *link)
-{
-	struct dc_link_settings max_link_cap = {0};
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	enum dc_link_rate lttpr_max_link_rate;
-#endif
-	struct link_encoder *link_enc = NULL;
-
-	/* Links supporting dynamically assigned link encoder will be assigned next
-	 * available encoder if one not already assigned.
-	 */
-	if (link->is_dig_mapping_flexible &&
-			link->dc->res_pool->funcs->link_encs_assign) {
-		link_enc = link_enc_cfg_get_link_enc_used_by_link(link->ctx->dc, link);
-		if (link_enc == NULL)
-			link_enc = link_enc_cfg_get_next_avail_link_enc(link->ctx->dc);
-	} else
-		link_enc = link->link_enc;
-	ASSERT(link_enc);
-
-	/* get max link encoder capability */
-	if (link_enc)
-		link_enc->funcs->get_max_link_cap(link_enc, &max_link_cap);
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	if (max_link_cap.link_rate >= LINK_RATE_UHBR10 &&
-			!link->hpo_dp_link_enc)
-		max_link_cap.link_rate = LINK_RATE_HIGH3;
-#endif
-
-	/* Lower link settings based on sink's link cap */
-	if (link->reported_link_cap.lane_count < max_link_cap.lane_count)
-		max_link_cap.lane_count =
-				link->reported_link_cap.lane_count;
-	if (link->reported_link_cap.link_rate < max_link_cap.link_rate)
-		max_link_cap.link_rate =
-				link->reported_link_cap.link_rate;
-	if (link->reported_link_cap.link_spread <
-			max_link_cap.link_spread)
-		max_link_cap.link_spread =
-				link->reported_link_cap.link_spread;
-	/*
-	 * account for lttpr repeaters cap
-	 * notes: repeaters do not snoop in the DPRX Capabilities addresses (3.6.3).
-	 */
-	if (link->lttpr_mode == LTTPR_MODE_NON_TRANSPARENT) {
-		if (link->dpcd_caps.lttpr_caps.max_lane_count < max_link_cap.lane_count)
-			max_link_cap.lane_count = link->dpcd_caps.lttpr_caps.max_lane_count;
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-		lttpr_max_link_rate = get_lttpr_max_link_rate(link);
-
-		if (lttpr_max_link_rate < max_link_cap.link_rate)
-			max_link_cap.link_rate = lttpr_max_link_rate;
-#else
-		if (link->dpcd_caps.lttpr_caps.max_link_rate < max_link_cap.link_rate)
-			max_link_cap.link_rate = link->dpcd_caps.lttpr_caps.max_link_rate;
-#endif
-
-		DC_LOG_HW_LINK_TRAINING("%s\n Training with LTTPR,  max_lane count %d max_link rate %d \n",
-						__func__,
-						max_link_cap.lane_count,
-						max_link_cap.link_rate);
-	}
-	return max_link_cap;
-}
-
-static enum dc_status read_hpd_rx_irq_data(
-	struct dc_link *link,
-	union hpd_irq_data *irq_data)
-{
-	static enum dc_status retval;
-
-	/* The HW reads 16 bytes from 200h on HPD,
-	 * but if we get an AUX_DEFER, the HW cannot retry
-	 * and this causes the CTS tests 4.3.2.1 - 3.2.4 to
-	 * fail, so we now explicitly read 6 bytes which is
-	 * the req from the above mentioned test cases.
-	 *
-	 * For DP 1.4 we need to read those from 2002h range.
-	 */
-	if (link->dpcd_caps.dpcd_rev.raw < DPCD_REV_14)
-		retval = core_link_read_dpcd(
-			link,
-			DP_SINK_COUNT,
-			irq_data->raw,
-			sizeof(union hpd_irq_data));
-	else {
-		/* Read 14 bytes in a single read and then copy only the required fields.
-		 * This is more efficient than doing it in two separate AUX reads. */
-
-		uint8_t tmp[DP_SINK_STATUS_ESI - DP_SINK_COUNT_ESI + 1];
-
-		retval = core_link_read_dpcd(
-			link,
-			DP_SINK_COUNT_ESI,
-			tmp,
-			sizeof(tmp));
-
-		if (retval != DC_OK)
-			return retval;
-
-		irq_data->bytes.sink_cnt.raw = tmp[DP_SINK_COUNT_ESI - DP_SINK_COUNT_ESI];
-		irq_data->bytes.device_service_irq.raw = tmp[DP_DEVICE_SERVICE_IRQ_VECTOR_ESI0 - DP_SINK_COUNT_ESI];
-		irq_data->bytes.lane01_status.raw = tmp[DP_LANE0_1_STATUS_ESI - DP_SINK_COUNT_ESI];
-		irq_data->bytes.lane23_status.raw = tmp[DP_LANE2_3_STATUS_ESI - DP_SINK_COUNT_ESI];
-		irq_data->bytes.lane_status_updated.raw = tmp[DP_LANE_ALIGN_STATUS_UPDATED_ESI - DP_SINK_COUNT_ESI];
-		irq_data->bytes.sink_status.raw = tmp[DP_SINK_STATUS_ESI - DP_SINK_COUNT_ESI];
-	}
-
-	return retval;
-}
-
-bool hpd_rx_irq_check_link_loss_status(
-	struct dc_link *link,
-	union hpd_irq_data *hpd_irq_dpcd_data)
-{
-	uint8_t irq_reg_rx_power_state = 0;
-	enum dc_status dpcd_result = DC_ERROR_UNEXPECTED;
-	union lane_status lane_status;
-	uint32_t lane;
-	bool sink_status_changed;
-	bool return_code;
-
-	sink_status_changed = false;
-	return_code = false;
-
-	if (link->cur_link_settings.lane_count == 0)
-		return return_code;
-
-	/*1. Check that Link Status changed, before re-training.*/
-
-	/*parse lane status*/
-	for (lane = 0; lane < link->cur_link_settings.lane_count; lane++) {
-		/* check status of lanes 0,1
-		 * changed DpcdAddress_Lane01Status (0x202)
-		 */
-		lane_status.raw = get_nibble_at_index(
-			&hpd_irq_dpcd_data->bytes.lane01_status.raw,
-			lane);
-
-		if (!lane_status.bits.CHANNEL_EQ_DONE_0 ||
-			!lane_status.bits.CR_DONE_0 ||
-			!lane_status.bits.SYMBOL_LOCKED_0) {
-			/* if one of the channel equalization, clock
-			 * recovery or symbol lock is dropped
-			 * consider it as (link has been
-			 * dropped) dp sink status has changed
-			 */
-			sink_status_changed = true;
-			break;
-		}
-	}
-
-	/* Check interlane align.*/
-	if (sink_status_changed ||
-		!hpd_irq_dpcd_data->bytes.lane_status_updated.bits.INTERLANE_ALIGN_DONE) {
-
-		DC_LOG_HW_HPD_IRQ("%s: Link Status changed.\n", __func__);
-
-		return_code = true;
-
-		/*2. Check that we can handle interrupt: Not in FS DOS,
-		 *  Not in "Display Timeout" state, Link is trained.
-		 */
-		dpcd_result = core_link_read_dpcd(link,
-			DP_SET_POWER,
-			&irq_reg_rx_power_state,
-			sizeof(irq_reg_rx_power_state));
-
-		if (dpcd_result != DC_OK) {
-			DC_LOG_HW_HPD_IRQ("%s: DPCD read failed to obtain power state.\n",
-				__func__);
-		} else {
-			if (irq_reg_rx_power_state != DP_SET_POWER_D0)
-				return_code = false;
-		}
-	}
-
-	return return_code;
-}
-
-bool dp_verify_link_cap(
-	struct dc_link *link,
-	struct dc_link_settings *known_limit_link_setting,
-	int *fail_count)
-{
-	struct dc_link_settings max_link_cap = {0};
-	struct dc_link_settings cur_link_setting = {0};
-	struct dc_link_settings *cur = &cur_link_setting;
-	struct dc_link_settings initial_link_settings = {0};
-	bool success;
-	bool skip_link_training;
-	bool skip_video_pattern;
-	enum clock_source_id dp_cs_id = CLOCK_SOURCE_ID_EXTERNAL;
-	enum link_training_result status;
-	union hpd_irq_data irq_data;
-
-	/* link training starts with the maximum common settings
-	 * supported by both sink and ASIC.
-	 */
-	max_link_cap = get_max_link_cap(link);
-	initial_link_settings = get_common_supported_link_settings(
-			*known_limit_link_setting,
-			max_link_cap);
-
-	/* Accept reported capabilities if link supports flexible encoder mapping or encoder already in use. */
-	if (link->dc->debug.skip_detection_link_training ||
-			link->is_dig_mapping_flexible) {
-		/* TODO - should we check link encoder's max link caps here?
-		 * How do we know which link encoder to check from?
-		 */
-		link->verified_link_cap = *known_limit_link_setting;
-		return true;
-	} else if (link->link_enc && link->dc->res_pool->funcs->link_encs_assign &&
-			!link_enc_cfg_is_link_enc_avail(link->ctx->dc, link->link_enc->preferred_engine, link)) {
-		link->verified_link_cap = initial_link_settings;
-		return true;
-	}
-
-	memset(&irq_data, 0, sizeof(irq_data));
-	success = false;
-	skip_link_training = false;
-
-	/* Grant extended timeout request */
-	if ((link->lttpr_mode == LTTPR_MODE_NON_TRANSPARENT) && (link->dpcd_caps.lttpr_caps.max_ext_timeout > 0)) {
-		uint8_t grant = link->dpcd_caps.lttpr_caps.max_ext_timeout & 0x80;
-
-		core_link_write_dpcd(link, DP_PHY_REPEATER_EXTENDED_WAIT_TIMEOUT, &grant, sizeof(grant));
-	}
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	if (dp_get_link_encoding_format(&link->cur_link_settings) == DP_128b_132b_ENCODING)
-		reset_dp_hpo_stream_encoders_for_link(link);
-#endif
-	/* TODO implement override and monitor patch later */
-
-	/* try to train the link from high to low to
-	 * find the physical link capability
-	 */
-	/* disable PHY done possible by BIOS, will be done by driver itself */
-	dp_disable_link_phy(link, link->connector_signal);
-
-	dp_cs_id = get_clock_source_id(link);
-
-	cur_link_setting = initial_link_settings;
-
-	/* Temporary Renoir-specific workaround for SWDEV-215184;
-	 * PHY will sometimes be in bad state on hotplugging display from certain USB-C dongle,
-	 * so add extra cycle of enabling and disabling the PHY before first link training.
-	 */
-	if (link->link_enc && link->link_enc->features.flags.bits.DP_IS_USB_C &&
-			link->dc->debug.usbc_combo_phy_reset_wa) {
-		dp_enable_link_phy(link, link->connector_signal, dp_cs_id, cur);
-		dp_disable_link_phy(link, link->connector_signal);
-	}
-
-	do {
-		skip_video_pattern = true;
-
-		if (cur->link_rate == LINK_RATE_LOW)
-			skip_video_pattern = false;
-
-		dp_enable_link_phy(
-				link,
-				link->connector_signal,
-				dp_cs_id,
-				cur);
-
-
-		if (skip_link_training)
-			success = true;
-		else {
-			status = dc_link_dp_perform_link_training(
-							link,
-							cur,
-							skip_video_pattern);
-			if (status == LINK_TRAINING_SUCCESS)
-				success = true;
-			else
-				(*fail_count)++;
-		}
-
-		if (success) {
-			link->verified_link_cap = *cur;
-			udelay(1000);
-			if (read_hpd_rx_irq_data(link, &irq_data) == DC_OK)
-				if (hpd_rx_irq_check_link_loss_status(
-						link,
-						&irq_data))
-					(*fail_count)++;
-		}
-		/* always disable the link before trying another
-		 * setting or before returning we'll enable it later
-		 * based on the actual mode we're driving
-		 */
-		dp_disable_link_phy(link, link->connector_signal);
-	} while (!success && decide_fallback_link_setting(
-			initial_link_settings, cur, status));
-
-	/* Link Training failed for all Link Settings
-	 *  (Lane Count is still unknown)
-	 */
-	if (!success) {
-		/* If all LT fails for all settings,
-		 * set verified = failed safe (1 lane low)
-		 */
-		link->verified_link_cap.lane_count = LANE_COUNT_ONE;
-		link->verified_link_cap.link_rate = LINK_RATE_LOW;
-
-		link->verified_link_cap.link_spread =
-		LINK_SPREAD_DISABLED;
-	}
-
-
-	return success;
-}
-
-bool dp_verify_link_cap_with_retries(
-	struct dc_link *link,
-	struct dc_link_settings *known_limit_link_setting,
-	int attempts)
-{
-	int i = 0;
-	bool success = false;
-
-	for (i = 0; i < attempts; i++) {
-		int fail_count = 0;
-		enum dc_connection_type type = dc_connection_none;
-
-		memset(&link->verified_link_cap, 0,
-				sizeof(struct dc_link_settings));
-		if (!dc_link_detect_sink(link, &type) || type == dc_connection_none) {
-			link->verified_link_cap.lane_count = LANE_COUNT_ONE;
-			link->verified_link_cap.link_rate = LINK_RATE_LOW;
-			link->verified_link_cap.link_spread = LINK_SPREAD_DISABLED;
-			break;
-		} else if (dp_verify_link_cap(link,
-				known_limit_link_setting,
-				&fail_count) && fail_count == 0) {
-			success = true;
-			break;
-		}
-		msleep(10);
-	}
-	return success;
-}
-
-bool dp_verify_mst_link_cap(
-	struct dc_link *link)
-{
-	struct dc_link_settings max_link_cap = {0};
-
-	if (dp_get_link_encoding_format(&link->reported_link_cap) ==
-			DP_8b_10b_ENCODING) {
-		max_link_cap = get_max_link_cap(link);
-		link->verified_link_cap = get_common_supported_link_settings(
-				link->reported_link_cap,
-				max_link_cap);
-	}
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	else if (dp_get_link_encoding_format(&link->reported_link_cap) ==
-			DP_128b_132b_ENCODING) {
-		dp_verify_link_cap_with_retries(link,
-				&link->reported_link_cap,
-				LINK_TRAINING_MAX_VERIFY_RETRY);
-	}
-#endif
-	return true;
-}
-
-static struct dc_link_settings get_common_supported_link_settings(
-		struct dc_link_settings link_setting_a,
-		struct dc_link_settings link_setting_b)
-{
-	struct dc_link_settings link_settings = {0};
-
-	link_settings.lane_count =
-		(link_setting_a.lane_count <=
-			link_setting_b.lane_count) ?
-			link_setting_a.lane_count :
-			link_setting_b.lane_count;
-	link_settings.link_rate =
-		(link_setting_a.link_rate <=
-			link_setting_b.link_rate) ?
-			link_setting_a.link_rate :
-			link_setting_b.link_rate;
-	link_settings.link_spread = LINK_SPREAD_DISABLED;
-
-	/* in DP compliance test, DPR-120 may have
-	 * a random value in its MAX_LINK_BW dpcd field.
-	 * We map it to the maximum supported link rate that
-	 * is smaller than MAX_LINK_BW in this case.
-	 */
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	if (link_settings.link_rate > LINK_RATE_UHBR20) {
-		link_settings.link_rate = LINK_RATE_UHBR20;
-	} else if (link_settings.link_rate < LINK_RATE_UHBR20 &&
-			link_settings.link_rate > LINK_RATE_UHBR13_5) {
-		link_settings.link_rate = LINK_RATE_UHBR13_5;
-	} else if (link_settings.link_rate < LINK_RATE_UHBR10 &&
-			link_settings.link_rate > LINK_RATE_HIGH3) {
-#else
-	if (link_settings.link_rate > LINK_RATE_HIGH3) {
-#endif
-		link_settings.link_rate = LINK_RATE_HIGH3;
-	} else if (link_settings.link_rate < LINK_RATE_HIGH3
-			&& link_settings.link_rate > LINK_RATE_HIGH2) {
-		link_settings.link_rate = LINK_RATE_HIGH2;
-	} else if (link_settings.link_rate < LINK_RATE_HIGH2
-			&& link_settings.link_rate > LINK_RATE_HIGH) {
-		link_settings.link_rate = LINK_RATE_HIGH;
-	} else if (link_settings.link_rate < LINK_RATE_HIGH
-			&& link_settings.link_rate > LINK_RATE_LOW) {
-		link_settings.link_rate = LINK_RATE_LOW;
-	} else if (link_settings.link_rate < LINK_RATE_LOW) {
-		link_settings.link_rate = LINK_RATE_UNKNOWN;
-	}
-
-	return link_settings;
-}
-
-static inline bool reached_minimum_lane_count(enum dc_lane_count lane_count)
-{
-	return lane_count <= LANE_COUNT_ONE;
-}
-
-static inline bool reached_minimum_link_rate(enum dc_link_rate link_rate)
-{
-	return link_rate <= LINK_RATE_LOW;
-}
-
-static enum dc_lane_count reduce_lane_count(enum dc_lane_count lane_count)
-{
-	switch (lane_count) {
-	case LANE_COUNT_FOUR:
-		return LANE_COUNT_TWO;
-	case LANE_COUNT_TWO:
-		return LANE_COUNT_ONE;
-	case LANE_COUNT_ONE:
-		return LANE_COUNT_UNKNOWN;
-	default:
-		return LANE_COUNT_UNKNOWN;
-	}
-}
-
-static enum dc_link_rate reduce_link_rate(enum dc_link_rate link_rate)
-{
-	switch (link_rate) {
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	case LINK_RATE_UHBR20:
-		return LINK_RATE_UHBR13_5;
-	case LINK_RATE_UHBR13_5:
-		return LINK_RATE_UHBR10;
-	case LINK_RATE_UHBR10:
-		return LINK_RATE_HIGH3;
-#endif
-	case LINK_RATE_HIGH3:
-		return LINK_RATE_HIGH2;
-	case LINK_RATE_HIGH2:
-		return LINK_RATE_HIGH;
-	case LINK_RATE_HIGH:
-		return LINK_RATE_LOW;
-	case LINK_RATE_LOW:
-		return LINK_RATE_UNKNOWN;
-	default:
-		return LINK_RATE_UNKNOWN;
-	}
-}
-
-static enum dc_lane_count increase_lane_count(enum dc_lane_count lane_count)
-{
-	switch (lane_count) {
-	case LANE_COUNT_ONE:
-		return LANE_COUNT_TWO;
-	case LANE_COUNT_TWO:
-		return LANE_COUNT_FOUR;
-	default:
-		return LANE_COUNT_UNKNOWN;
-	}
-}
-
-static enum dc_link_rate increase_link_rate(enum dc_link_rate link_rate)
-{
-	switch (link_rate) {
-	case LINK_RATE_LOW:
-		return LINK_RATE_HIGH;
-	case LINK_RATE_HIGH:
-		return LINK_RATE_HIGH2;
-	case LINK_RATE_HIGH2:
-		return LINK_RATE_HIGH3;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	case LINK_RATE_HIGH3:
-		return LINK_RATE_UHBR10;
-	case LINK_RATE_UHBR10:
-		return LINK_RATE_UHBR13_5;
-	case LINK_RATE_UHBR13_5:
-		return LINK_RATE_UHBR20;
-#endif
-	default:
-		return LINK_RATE_UNKNOWN;
-	}
-}
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-static bool decide_fallback_link_setting_max_bw_policy(
-		const struct dc_link_settings *max,
-		struct dc_link_settings *cur)
-{
-	uint8_t cur_idx = 0, next_idx;
-	bool found = false;
-
-	while (cur_idx < ARRAY_SIZE(dp_lt_fallbacks))
-		/* find current index */
-		if (dp_lt_fallbacks[cur_idx].lane_count == cur->lane_count &&
-				dp_lt_fallbacks[cur_idx].link_rate == cur->link_rate)
-			break;
-		else
-			cur_idx++;
-
-	next_idx = cur_idx + 1;
-
-	while (next_idx < ARRAY_SIZE(dp_lt_fallbacks))
-		/* find next index */
-		if (dp_lt_fallbacks[next_idx].lane_count <= max->lane_count &&
-				dp_lt_fallbacks[next_idx].link_rate <= max->link_rate)
-			break;
-		else
-			next_idx++;
-
-	if (next_idx < ARRAY_SIZE(dp_lt_fallbacks)) {
-		cur->lane_count = dp_lt_fallbacks[next_idx].lane_count;
-		cur->link_rate = dp_lt_fallbacks[next_idx].link_rate;
-		found = true;
-	}
-
-	return found;
-}
-#endif
-
-/*
- * function: set link rate and lane count fallback based
- * on current link setting and last link training result
- * return value:
- *			true - link setting could be set
- *			false - has reached minimum setting
- *					and no further fallback could be done
- */
-static bool decide_fallback_link_setting(
-		struct dc_link_settings initial_link_settings,
-		struct dc_link_settings *current_link_setting,
-		enum link_training_result training_result)
-{
-	if (!current_link_setting)
-		return false;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	if (dp_get_link_encoding_format(&initial_link_settings) == DP_128b_132b_ENCODING)
-		return decide_fallback_link_setting_max_bw_policy(&initial_link_settings,
-				current_link_setting);
-#endif
-
-	switch (training_result) {
-	case LINK_TRAINING_CR_FAIL_LANE0:
-	case LINK_TRAINING_CR_FAIL_LANE1:
-	case LINK_TRAINING_CR_FAIL_LANE23:
-	case LINK_TRAINING_LQA_FAIL:
-	{
-		if (!reached_minimum_link_rate
-				(current_link_setting->link_rate)) {
-			current_link_setting->link_rate =
-				reduce_link_rate(
-					current_link_setting->link_rate);
-		} else if (!reached_minimum_lane_count
-				(current_link_setting->lane_count)) {
-			current_link_setting->link_rate =
-				initial_link_settings.link_rate;
-			if (training_result == LINK_TRAINING_CR_FAIL_LANE0)
-				return false;
-			else if (training_result == LINK_TRAINING_CR_FAIL_LANE1)
-				current_link_setting->lane_count =
-						LANE_COUNT_ONE;
-			else if (training_result ==
-					LINK_TRAINING_CR_FAIL_LANE23)
-				current_link_setting->lane_count =
-						LANE_COUNT_TWO;
-			else
-				current_link_setting->lane_count =
-					reduce_lane_count(
-					current_link_setting->lane_count);
-		} else {
-			return false;
-		}
-		break;
-	}
-	case LINK_TRAINING_EQ_FAIL_EQ:
-	{
-		if (!reached_minimum_lane_count
-				(current_link_setting->lane_count)) {
-			current_link_setting->lane_count =
-				reduce_lane_count(
-					current_link_setting->lane_count);
-		} else if (!reached_minimum_link_rate
-				(current_link_setting->link_rate)) {
-			current_link_setting->link_rate =
-				reduce_link_rate(
-					current_link_setting->link_rate);
-		} else {
-			return false;
-		}
-		break;
-	}
-	case LINK_TRAINING_EQ_FAIL_CR:
-	{
-		if (!reached_minimum_link_rate
-				(current_link_setting->link_rate)) {
-			current_link_setting->link_rate =
-				reduce_link_rate(
-					current_link_setting->link_rate);
-		} else {
-			return false;
-		}
-		break;
-	}
-	default:
-		return false;
-	}
-	return true;
-}
-
-bool dp_validate_mode_timing(
-	struct dc_link *link,
-	const struct dc_crtc_timing *timing)
-{
-	uint32_t req_bw;
-	uint32_t max_bw;
-
-	const struct dc_link_settings *link_setting;
-
-	/* According to spec, VSC SDP should be used if pixel format is YCbCr420 */
-	if (timing->pixel_encoding == PIXEL_ENCODING_YCBCR420 &&
-			!link->dpcd_caps.dprx_feature.bits.VSC_SDP_COLORIMETRY_SUPPORTED &&
-			dal_graphics_object_id_get_connector_id(link->link_id) != CONNECTOR_ID_VIRTUAL)
-		return false;
-
-	/*always DP fail safe mode*/
-	if ((timing->pix_clk_100hz / 10) == (uint32_t) 25175 &&
-		timing->h_addressable == (uint32_t) 640 &&
-		timing->v_addressable == (uint32_t) 480)
-		return true;
-
-	link_setting = dc_link_get_link_cap(link);
-
-	/* TODO: DYNAMIC_VALIDATION needs to be implemented */
-	/*if (flags.DYNAMIC_VALIDATION == 1 &&
-		link->verified_link_cap.lane_count != LANE_COUNT_UNKNOWN)
-		link_setting = &link->verified_link_cap;
-	*/
-
-	req_bw = dc_bandwidth_in_kbps_from_timing(timing);
-	max_bw = dc_link_bandwidth_kbps(link, link_setting);
-
-	if (req_bw <= max_bw) {
-		/* remember the biggest mode here, during
-		 * initial link training (to get
-		 * verified_link_cap), LS sends event about
-		 * cannot train at reported cap to upper
-		 * layer and upper layer will re-enumerate modes.
-		 * this is not necessary if the lower
-		 * verified_link_cap is enough to drive
-		 * all the modes */
-
-		/* TODO: DYNAMIC_VALIDATION needs to be implemented */
-		/* if (flags.DYNAMIC_VALIDATION == 1)
-			dpsst->max_req_bw_for_verified_linkcap = dal_max(
-				dpsst->max_req_bw_for_verified_linkcap, req_bw); */
-		return true;
-	} else
-		return false;
-}
-
-static bool decide_dp_link_settings(struct dc_link *link, struct dc_link_settings *link_setting, uint32_t req_bw)
-{
-	struct dc_link_settings initial_link_setting = {
-		LANE_COUNT_ONE, LINK_RATE_LOW, LINK_SPREAD_DISABLED, false, 0};
-	struct dc_link_settings current_link_setting =
-			initial_link_setting;
-	uint32_t link_bw;
-
-	if (req_bw > dc_link_bandwidth_kbps(link, &link->verified_link_cap))
-		return false;
-
-	/* search for the minimum link setting that:
-	 * 1. is supported according to the link training result
-	 * 2. could support the b/w requested by the timing
-	 */
-	while (current_link_setting.link_rate <=
-			link->verified_link_cap.link_rate) {
-		link_bw = dc_link_bandwidth_kbps(
-				link,
-				&current_link_setting);
-		if (req_bw <= link_bw) {
-			*link_setting = current_link_setting;
-			return true;
-		}
-
-		if (current_link_setting.lane_count <
-				link->verified_link_cap.lane_count) {
-			current_link_setting.lane_count =
-					increase_lane_count(
-							current_link_setting.lane_count);
-		} else {
-			current_link_setting.link_rate =
-					increase_link_rate(
-							current_link_setting.link_rate);
-			current_link_setting.lane_count =
-					initial_link_setting.lane_count;
-		}
-	}
-
-	return false;
-}
-
-bool decide_edp_link_settings(struct dc_link *link, struct dc_link_settings *link_setting, uint32_t req_bw)
-{
-	struct dc_link_settings initial_link_setting;
-	struct dc_link_settings current_link_setting;
-	uint32_t link_bw;
-
-	/*
-	 * edp_supported_link_rates_count is only valid for eDP v1.4 or higher.
-	 * Per VESA eDP spec, "The DPCD revision for eDP v1.4 is 13h"
-	 */
-	if (link->dpcd_caps.dpcd_rev.raw < DPCD_REV_13 ||
-			link->dpcd_caps.edp_supported_link_rates_count == 0) {
-		*link_setting = link->verified_link_cap;
-		return true;
-	}
-
-	memset(&initial_link_setting, 0, sizeof(initial_link_setting));
-	initial_link_setting.lane_count = LANE_COUNT_ONE;
-	initial_link_setting.link_rate = link->dpcd_caps.edp_supported_link_rates[0];
-	initial_link_setting.link_spread = LINK_SPREAD_DISABLED;
-	initial_link_setting.use_link_rate_set = true;
-	initial_link_setting.link_rate_set = 0;
-	current_link_setting = initial_link_setting;
-
-	/* search for the minimum link setting that:
-	 * 1. is supported according to the link training result
-	 * 2. could support the b/w requested by the timing
-	 */
-	while (current_link_setting.link_rate <=
-			link->verified_link_cap.link_rate) {
-		link_bw = dc_link_bandwidth_kbps(
-				link,
-				&current_link_setting);
-		if (req_bw <= link_bw) {
-			*link_setting = current_link_setting;
-			return true;
-		}
-
-		if (current_link_setting.lane_count <
-				link->verified_link_cap.lane_count) {
-			current_link_setting.lane_count =
-					increase_lane_count(
-							current_link_setting.lane_count);
-		} else {
-			if (current_link_setting.link_rate_set < link->dpcd_caps.edp_supported_link_rates_count) {
-				current_link_setting.link_rate_set++;
-				current_link_setting.link_rate =
-					link->dpcd_caps.edp_supported_link_rates[current_link_setting.link_rate_set];
-				current_link_setting.lane_count =
-									initial_link_setting.lane_count;
-			} else
-				break;
-		}
-	}
-	return false;
-}
-
-static bool decide_mst_link_settings(const struct dc_link *link, struct dc_link_settings *link_setting)
-{
-	*link_setting = link->verified_link_cap;
-	return true;
-}
-
-void decide_link_settings(struct dc_stream_state *stream,
-	struct dc_link_settings *link_setting)
-{
-	struct dc_link *link;
-	uint32_t req_bw;
-
-	req_bw = dc_bandwidth_in_kbps_from_timing(&stream->timing);
-
-	link = stream->link;
-
-	/* if preferred is specified through AMDDP, use it, if it's enough
-	 * to drive the mode
-	 */
-	if (link->preferred_link_setting.lane_count !=
-			LANE_COUNT_UNKNOWN &&
-			link->preferred_link_setting.link_rate !=
-					LINK_RATE_UNKNOWN) {
-		*link_setting =  link->preferred_link_setting;
-		return;
-	}
-
-	/* MST doesn't perform link training for now
-	 * TODO: add MST specific link training routine
-	 */
-	if (stream->signal == SIGNAL_TYPE_DISPLAY_PORT_MST) {
-		if (decide_mst_link_settings(link, link_setting))
-			return;
-	} else if (link->connector_signal == SIGNAL_TYPE_EDP) {
-		if (decide_edp_link_settings(link, link_setting, req_bw))
-			return;
-	} else if (decide_dp_link_settings(link, link_setting, req_bw))
-		return;
-
-	BREAK_TO_DEBUGGER();
-	ASSERT(link->verified_link_cap.lane_count != LANE_COUNT_UNKNOWN);
-
-	*link_setting = link->verified_link_cap;
-}
-
-/*************************Short Pulse IRQ***************************/
-bool dc_link_dp_allow_hpd_rx_irq(const struct dc_link *link)
-{
-	/*
-	 * Don't handle RX IRQ unless one of following is met:
-	 * 1) The link is established (cur_link_settings != unknown)
-	 * 2) We know we're dealing with a branch device, SST or MST
-	 */
-
-	if ((link->cur_link_settings.lane_count != LANE_COUNT_UNKNOWN) ||
-		is_dp_branch_device(link))
-		return true;
-
-	return false;
-}
-
-static bool handle_hpd_irq_psr_sink(struct dc_link *link)
-{
-	union dpcd_psr_configuration psr_configuration;
-
-	if (!link->psr_settings.psr_feature_enabled)
-		return false;
-
-	dm_helpers_dp_read_dpcd(
-		link->ctx,
-		link,
-		368,/*DpcdAddress_PSR_Enable_Cfg*/
-		&psr_configuration.raw,
-		sizeof(psr_configuration.raw));
-
-
-	if (psr_configuration.bits.ENABLE) {
-		unsigned char dpcdbuf[3] = {0};
-		union psr_error_status psr_error_status;
-		union psr_sink_psr_status psr_sink_psr_status;
-
-		dm_helpers_dp_read_dpcd(
-			link->ctx,
-			link,
-			0x2006, /*DpcdAddress_PSR_Error_Status*/
-			(unsigned char *) dpcdbuf,
-			sizeof(dpcdbuf));
-
-		/*DPCD 2006h   ERROR STATUS*/
-		psr_error_status.raw = dpcdbuf[0];
-		/*DPCD 2008h   SINK PANEL SELF REFRESH STATUS*/
-		psr_sink_psr_status.raw = dpcdbuf[2];
-
-		if (psr_error_status.bits.LINK_CRC_ERROR ||
-				psr_error_status.bits.RFB_STORAGE_ERROR ||
-				psr_error_status.bits.VSC_SDP_ERROR) {
-			bool allow_active;
-
-			/* Acknowledge and clear error bits */
-			dm_helpers_dp_write_dpcd(
-				link->ctx,
-				link,
-				8198,/*DpcdAddress_PSR_Error_Status*/
-				&psr_error_status.raw,
-				sizeof(psr_error_status.raw));
-
-			/* PSR error, disable and re-enable PSR */
-			allow_active = false;
-			dc_link_set_psr_allow_active(link, &allow_active, true, false, NULL);
-			allow_active = true;
-			dc_link_set_psr_allow_active(link, &allow_active, true, false, NULL);
-
-			return true;
-		} else if (psr_sink_psr_status.bits.SINK_SELF_REFRESH_STATUS ==
-				PSR_SINK_STATE_ACTIVE_DISPLAY_FROM_SINK_RFB){
-			/* No error is detect, PSR is active.
-			 * We should return with IRQ_HPD handled without
-			 * checking for loss of sync since PSR would have
-			 * powered down main link.
-			 */
-			return true;
-		}
-	}
-	return false;
-}
-
-static void dp_test_send_link_training(struct dc_link *link)
-{
-	struct dc_link_settings link_settings = {0};
-
-	core_link_read_dpcd(
-			link,
-			DP_TEST_LANE_COUNT,
-			(unsigned char *)(&link_settings.lane_count),
-			1);
-	core_link_read_dpcd(
-			link,
-			DP_TEST_LINK_RATE,
-			(unsigned char *)(&link_settings.link_rate),
-			1);
-
-	/* Set preferred link settings */
-	link->verified_link_cap.lane_count = link_settings.lane_count;
-	link->verified_link_cap.link_rate = link_settings.link_rate;
-
-	dp_retrain_link_dp_test(link, &link_settings, false);
-}
-
-/* TODO Raven hbr2 compliance eye output is unstable
- * (toggling on and off) with debugger break
- * This caueses intermittent PHY automation failure
- * Need to look into the root cause */
-static void dp_test_send_phy_test_pattern(struct dc_link *link)
-{
-	union phy_test_pattern dpcd_test_pattern;
-	union lane_adjust dpcd_lane_adjustment[2];
-	unsigned char dpcd_post_cursor_2_adjustment = 0;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	unsigned char test_pattern_buffer[
-			(DP_TEST_264BIT_CUSTOM_PATTERN_263_256 -
-			DP_TEST_264BIT_CUSTOM_PATTERN_7_0)+1] = {0};
-#else
-	unsigned char test_pattern_buffer[
-			(DP_TEST_80BIT_CUSTOM_PATTERN_79_72 -
-			DP_TEST_80BIT_CUSTOM_PATTERN_7_0)+1] = {0};
-#endif
-	unsigned int test_pattern_size = 0;
-	enum dp_test_pattern test_pattern;
-	union lane_adjust dpcd_lane_adjust;
-	unsigned int lane;
-	struct link_training_settings link_training_settings;
-
-	dpcd_test_pattern.raw = 0;
-	memset(dpcd_lane_adjustment, 0, sizeof(dpcd_lane_adjustment));
-	memset(&link_training_settings, 0, sizeof(link_training_settings));
-
-	/* get phy test pattern and pattern parameters from DP receiver */
-	core_link_read_dpcd(
-			link,
-			DP_PHY_TEST_PATTERN,
-			&dpcd_test_pattern.raw,
-			sizeof(dpcd_test_pattern));
-	core_link_read_dpcd(
-			link,
-			DP_ADJUST_REQUEST_LANE0_1,
-			&dpcd_lane_adjustment[0].raw,
-			sizeof(dpcd_lane_adjustment));
-
-	/*get post cursor 2 parameters
-	 * For DP 1.1a or eariler, this DPCD register's value is 0
-	 * For DP 1.2 or later:
-	 * Bits 1:0 = POST_CURSOR2_LANE0; Bits 3:2 = POST_CURSOR2_LANE1
-	 * Bits 5:4 = POST_CURSOR2_LANE2; Bits 7:6 = POST_CURSOR2_LANE3
-	 */
-	core_link_read_dpcd(
-			link,
-			DP_ADJUST_REQUEST_POST_CURSOR2,
-			&dpcd_post_cursor_2_adjustment,
-			sizeof(dpcd_post_cursor_2_adjustment));
-
-	/* translate request */
-	switch (dpcd_test_pattern.bits.PATTERN) {
-	case PHY_TEST_PATTERN_D10_2:
-		test_pattern = DP_TEST_PATTERN_D102;
-		break;
-	case PHY_TEST_PATTERN_SYMBOL_ERROR:
-		test_pattern = DP_TEST_PATTERN_SYMBOL_ERROR;
-		break;
-	case PHY_TEST_PATTERN_PRBS7:
-		test_pattern = DP_TEST_PATTERN_PRBS7;
-		break;
-	case PHY_TEST_PATTERN_80BIT_CUSTOM:
-		test_pattern = DP_TEST_PATTERN_80BIT_CUSTOM;
-		break;
-	case PHY_TEST_PATTERN_CP2520_1:
-		/* CP2520 pattern is unstable, temporarily use TPS4 instead */
-		test_pattern = (link->dc->caps.force_dp_tps4_for_cp2520 == 1) ?
-				DP_TEST_PATTERN_TRAINING_PATTERN4 :
-				DP_TEST_PATTERN_HBR2_COMPLIANCE_EYE;
-		break;
-	case PHY_TEST_PATTERN_CP2520_2:
-		/* CP2520 pattern is unstable, temporarily use TPS4 instead */
-		test_pattern = (link->dc->caps.force_dp_tps4_for_cp2520 == 1) ?
-				DP_TEST_PATTERN_TRAINING_PATTERN4 :
-				DP_TEST_PATTERN_HBR2_COMPLIANCE_EYE;
-		break;
-	case PHY_TEST_PATTERN_CP2520_3:
-		test_pattern = DP_TEST_PATTERN_TRAINING_PATTERN4;
-		break;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	case PHY_TEST_PATTERN_128b_132b_TPS1:
-		test_pattern = DP_TEST_PATTERN_128b_132b_TPS1;
-		break;
-	case PHY_TEST_PATTERN_128b_132b_TPS2:
-		test_pattern = DP_TEST_PATTERN_128b_132b_TPS2;
-		break;
-	case PHY_TEST_PATTERN_PRBS9:
-		test_pattern = DP_TEST_PATTERN_PRBS9;
-		break;
-	case PHY_TEST_PATTERN_PRBS11:
-		test_pattern = DP_TEST_PATTERN_PRBS11;
-		break;
-	case PHY_TEST_PATTERN_PRBS15:
-		test_pattern = DP_TEST_PATTERN_PRBS15;
-		break;
-	case PHY_TEST_PATTERN_PRBS23:
-		test_pattern = DP_TEST_PATTERN_PRBS23;
-		break;
-	case PHY_TEST_PATTERN_PRBS31:
-		test_pattern = DP_TEST_PATTERN_PRBS31;
-		break;
-	case PHY_TEST_PATTERN_264BIT_CUSTOM:
-		test_pattern = DP_TEST_PATTERN_264BIT_CUSTOM;
-		break;
-	case PHY_TEST_PATTERN_SQUARE_PULSE:
-		test_pattern = DP_TEST_PATTERN_SQUARE_PULSE;
-		break;
-#endif
-	default:
-		test_pattern = DP_TEST_PATTERN_VIDEO_MODE;
-	break;
-	}
-
-	if (test_pattern == DP_TEST_PATTERN_80BIT_CUSTOM) {
-		test_pattern_size = (DP_TEST_80BIT_CUSTOM_PATTERN_79_72 -
-				DP_TEST_80BIT_CUSTOM_PATTERN_7_0) + 1;
-		core_link_read_dpcd(
-				link,
-				DP_TEST_80BIT_CUSTOM_PATTERN_7_0,
-				test_pattern_buffer,
-				test_pattern_size);
-	}
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	if (test_pattern == DP_TEST_PATTERN_SQUARE_PULSE) {
-		test_pattern_size = 1; // Square pattern data is 1 byte (DP spec)
-		core_link_read_dpcd(
-				link,
-				DP_PHY_SQUARE_PATTERN,
-				test_pattern_buffer,
-				test_pattern_size);
-	}
-
-	if (test_pattern == DP_TEST_PATTERN_264BIT_CUSTOM) {
-		test_pattern_size = (DP_TEST_264BIT_CUSTOM_PATTERN_263_256-
-				DP_TEST_264BIT_CUSTOM_PATTERN_7_0) + 1;
-		core_link_read_dpcd(
-				link,
-				DP_TEST_264BIT_CUSTOM_PATTERN_7_0,
-				test_pattern_buffer,
-				test_pattern_size);
-	}
-#endif
-
-	/* prepare link training settings */
-	link_training_settings.link_settings = link->cur_link_settings;
-
-	for (lane = 0; lane <
-		(unsigned int)(link->cur_link_settings.lane_count);
-		lane++) {
-		dpcd_lane_adjust.raw =
-			get_nibble_at_index(&dpcd_lane_adjustment[0].raw, lane);
-		if (dp_get_link_encoding_format(&link->cur_link_settings) ==
-				DP_8b_10b_ENCODING) {
-			link_training_settings.hw_lane_settings[lane].VOLTAGE_SWING =
-				(enum dc_voltage_swing)
-				(dpcd_lane_adjust.bits.VOLTAGE_SWING_LANE);
-			link_training_settings.hw_lane_settings[lane].PRE_EMPHASIS =
-				(enum dc_pre_emphasis)
-				(dpcd_lane_adjust.bits.PRE_EMPHASIS_LANE);
-			link_training_settings.hw_lane_settings[lane].POST_CURSOR2 =
-				(enum dc_post_cursor2)
-				((dpcd_post_cursor_2_adjustment >> (lane * 2)) & 0x03);
-		}
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-		else if (dp_get_link_encoding_format(&link->cur_link_settings) ==
-				DP_128b_132b_ENCODING) {
-			link_training_settings.hw_lane_settings[lane].FFE_PRESET.raw =
-					dpcd_lane_adjust.tx_ffe.PRESET_VALUE;
-		}
-#endif
-	}
-
-	dp_hw_to_dpcd_lane_settings(&link_training_settings,
-			link_training_settings.hw_lane_settings,
-			link_training_settings.dpcd_lane_settings);
-	/*Usage: Measure DP physical lane signal
-	 * by DP SI test equipment automatically.
-	 * PHY test pattern request is generated by equipment via HPD interrupt.
-	 * HPD needs to be active all the time. HPD should be active
-	 * all the time. Do not touch it.
-	 * forward request to DS
-	 */
-	dc_link_dp_set_test_pattern(
-		link,
-		test_pattern,
-		DP_TEST_PATTERN_COLOR_SPACE_UNDEFINED,
-		&link_training_settings,
-		test_pattern_buffer,
-		test_pattern_size);
-}
-
-static void dp_test_send_link_test_pattern(struct dc_link *link)
-{
-	union link_test_pattern dpcd_test_pattern;
-	union test_misc dpcd_test_params;
-	enum dp_test_pattern test_pattern;
-	enum dp_test_pattern_color_space test_pattern_color_space =
-			DP_TEST_PATTERN_COLOR_SPACE_UNDEFINED;
-	enum dc_color_depth requestColorDepth = COLOR_DEPTH_UNDEFINED;
-	struct pipe_ctx *pipes = link->dc->current_state->res_ctx.pipe_ctx;
-	struct pipe_ctx *pipe_ctx = NULL;
-	int i;
-
-	memset(&dpcd_test_pattern, 0, sizeof(dpcd_test_pattern));
-	memset(&dpcd_test_params, 0, sizeof(dpcd_test_params));
-
-	for (i = 0; i < MAX_PIPES; i++) {
-		if (pipes[i].stream == NULL)
-			continue;
-
-		if (pipes[i].stream->link == link && !pipes[i].top_pipe && !pipes[i].prev_odm_pipe) {
-			pipe_ctx = &pipes[i];
-			break;
-		}
-	}
-
-	if (pipe_ctx == NULL)
-		return;
-
-	/* get link test pattern and pattern parameters */
-	core_link_read_dpcd(
-			link,
-			DP_TEST_PATTERN,
-			&dpcd_test_pattern.raw,
-			sizeof(dpcd_test_pattern));
-	core_link_read_dpcd(
-			link,
-			DP_TEST_MISC0,
-			&dpcd_test_params.raw,
-			sizeof(dpcd_test_params));
-
-	switch (dpcd_test_pattern.bits.PATTERN) {
-	case LINK_TEST_PATTERN_COLOR_RAMP:
-		test_pattern = DP_TEST_PATTERN_COLOR_RAMP;
-	break;
-	case LINK_TEST_PATTERN_VERTICAL_BARS:
-		test_pattern = DP_TEST_PATTERN_VERTICAL_BARS;
-	break; /* black and white */
-	case LINK_TEST_PATTERN_COLOR_SQUARES:
-		test_pattern = (dpcd_test_params.bits.DYN_RANGE ==
-				TEST_DYN_RANGE_VESA ?
-				DP_TEST_PATTERN_COLOR_SQUARES :
-				DP_TEST_PATTERN_COLOR_SQUARES_CEA);
-	break;
-	default:
-		test_pattern = DP_TEST_PATTERN_VIDEO_MODE;
-	break;
-	}
-
-	if (dpcd_test_params.bits.CLR_FORMAT == 0)
-		test_pattern_color_space = DP_TEST_PATTERN_COLOR_SPACE_RGB;
-	else
-		test_pattern_color_space = dpcd_test_params.bits.YCBCR_COEFS ?
-				DP_TEST_PATTERN_COLOR_SPACE_YCBCR709 :
-				DP_TEST_PATTERN_COLOR_SPACE_YCBCR601;
-
-	switch (dpcd_test_params.bits.BPC) {
-	case 0: // 6 bits
-		requestColorDepth = COLOR_DEPTH_666;
-		break;
-	case 1: // 8 bits
-		requestColorDepth = COLOR_DEPTH_888;
-		break;
-	case 2: // 10 bits
-		requestColorDepth = COLOR_DEPTH_101010;
-		break;
-	case 3: // 12 bits
-		requestColorDepth = COLOR_DEPTH_121212;
-		break;
-	default:
-		break;
-	}
-
-	switch (dpcd_test_params.bits.CLR_FORMAT) {
-	case 0:
-		pipe_ctx->stream->timing.pixel_encoding = PIXEL_ENCODING_RGB;
-		break;
-	case 1:
-		pipe_ctx->stream->timing.pixel_encoding = PIXEL_ENCODING_YCBCR422;
-		break;
-	case 2:
-		pipe_ctx->stream->timing.pixel_encoding = PIXEL_ENCODING_YCBCR444;
-		break;
-	default:
-		pipe_ctx->stream->timing.pixel_encoding = PIXEL_ENCODING_RGB;
-		break;
-	}
-
-
-	if (requestColorDepth != COLOR_DEPTH_UNDEFINED
-			&& pipe_ctx->stream->timing.display_color_depth != requestColorDepth) {
-		DC_LOG_DEBUG("%s: original bpc %d, changing to %d\n",
-				__func__,
-				pipe_ctx->stream->timing.display_color_depth,
-				requestColorDepth);
-		pipe_ctx->stream->timing.display_color_depth = requestColorDepth;
-	}
-
-	dp_update_dsc_config(pipe_ctx);
-
-	dc_link_dp_set_test_pattern(
-			link,
-			test_pattern,
-			test_pattern_color_space,
-			NULL,
-			NULL,
-			0);
-}
-
-static void dp_test_get_audio_test_data(struct dc_link *link, bool disable_video)
-{
-	union audio_test_mode            dpcd_test_mode = {0};
-	struct audio_test_pattern_type   dpcd_pattern_type = {0};
-	union audio_test_pattern_period  dpcd_pattern_period[AUDIO_CHANNELS_COUNT] = {0};
-	enum dp_test_pattern test_pattern = DP_TEST_PATTERN_AUDIO_OPERATOR_DEFINED;
-
-	struct pipe_ctx *pipes = link->dc->current_state->res_ctx.pipe_ctx;
-	struct pipe_ctx *pipe_ctx = &pipes[0];
-	unsigned int channel_count;
-	unsigned int channel = 0;
-	unsigned int modes = 0;
-	unsigned int sampling_rate_in_hz = 0;
-
-	// get audio test mode and test pattern parameters
-	core_link_read_dpcd(
-		link,
-		DP_TEST_AUDIO_MODE,
-		&dpcd_test_mode.raw,
-		sizeof(dpcd_test_mode));
-
-	core_link_read_dpcd(
-		link,
-		DP_TEST_AUDIO_PATTERN_TYPE,
-		&dpcd_pattern_type.value,
-		sizeof(dpcd_pattern_type));
-
-	channel_count = dpcd_test_mode.bits.channel_count + 1;
-
-	// read pattern periods for requested channels when sawTooth pattern is requested
-	if (dpcd_pattern_type.value == AUDIO_TEST_PATTERN_SAWTOOTH ||
-			dpcd_pattern_type.value == AUDIO_TEST_PATTERN_OPERATOR_DEFINED) {
-
-		test_pattern = (dpcd_pattern_type.value == AUDIO_TEST_PATTERN_SAWTOOTH) ?
-				DP_TEST_PATTERN_AUDIO_SAWTOOTH : DP_TEST_PATTERN_AUDIO_OPERATOR_DEFINED;
-		// read period for each channel
-		for (channel = 0; channel < channel_count; channel++) {
-			core_link_read_dpcd(
-							link,
-							DP_TEST_AUDIO_PERIOD_CH1 + channel,
-							&dpcd_pattern_period[channel].raw,
-							sizeof(dpcd_pattern_period[channel]));
-		}
-	}
-
-	// translate sampling rate
-	switch (dpcd_test_mode.bits.sampling_rate) {
-	case AUDIO_SAMPLING_RATE_32KHZ:
-		sampling_rate_in_hz = 32000;
-		break;
-	case AUDIO_SAMPLING_RATE_44_1KHZ:
-		sampling_rate_in_hz = 44100;
-		break;
-	case AUDIO_SAMPLING_RATE_48KHZ:
-		sampling_rate_in_hz = 48000;
-		break;
-	case AUDIO_SAMPLING_RATE_88_2KHZ:
-		sampling_rate_in_hz = 88200;
-		break;
-	case AUDIO_SAMPLING_RATE_96KHZ:
-		sampling_rate_in_hz = 96000;
-		break;
-	case AUDIO_SAMPLING_RATE_176_4KHZ:
-		sampling_rate_in_hz = 176400;
-		break;
-	case AUDIO_SAMPLING_RATE_192KHZ:
-		sampling_rate_in_hz = 192000;
-		break;
-	default:
-		sampling_rate_in_hz = 0;
-		break;
-	}
-
-	link->audio_test_data.flags.test_requested = 1;
-	link->audio_test_data.flags.disable_video = disable_video;
-	link->audio_test_data.sampling_rate = sampling_rate_in_hz;
-	link->audio_test_data.channel_count = channel_count;
-	link->audio_test_data.pattern_type = test_pattern;
-
-	if (test_pattern == DP_TEST_PATTERN_AUDIO_SAWTOOTH) {
-		for (modes = 0; modes < pipe_ctx->stream->audio_info.mode_count; modes++) {
-			link->audio_test_data.pattern_period[modes] = dpcd_pattern_period[modes].bits.pattern_period;
-		}
-	}
-}
-
-void dc_link_dp_handle_automated_test(struct dc_link *link)
-{
-	union test_request test_request;
-	union test_response test_response;
-
-	memset(&test_request, 0, sizeof(test_request));
-	memset(&test_response, 0, sizeof(test_response));
-
-	core_link_read_dpcd(
-		link,
-		DP_TEST_REQUEST,
-		&test_request.raw,
-		sizeof(union test_request));
-	if (test_request.bits.LINK_TRAINING) {
-		/* ACK first to let DP RX test box monitor LT sequence */
-		test_response.bits.ACK = 1;
-		core_link_write_dpcd(
-			link,
-			DP_TEST_RESPONSE,
-			&test_response.raw,
-			sizeof(test_response));
-		dp_test_send_link_training(link);
-		/* no acknowledge request is needed again */
-		test_response.bits.ACK = 0;
-	}
-	if (test_request.bits.LINK_TEST_PATTRN) {
-		dp_test_send_link_test_pattern(link);
-		test_response.bits.ACK = 1;
-	}
-
-	if (test_request.bits.AUDIO_TEST_PATTERN) {
-		dp_test_get_audio_test_data(link, test_request.bits.TEST_AUDIO_DISABLED_VIDEO);
-		test_response.bits.ACK = 1;
-	}
-
-	if (test_request.bits.PHY_TEST_PATTERN) {
-		dp_test_send_phy_test_pattern(link);
-		test_response.bits.ACK = 1;
-	}
-
-	/* send request acknowledgment */
-	if (test_response.bits.ACK)
-		core_link_write_dpcd(
-			link,
-			DP_TEST_RESPONSE,
-			&test_response.raw,
-			sizeof(test_response));
-}
-
-void dc_link_dp_handle_link_loss(struct dc_link *link)
-{
-	int i;
-	struct pipe_ctx *pipe_ctx;
-
-	for (i = 0; i < MAX_PIPES; i++) {
-		pipe_ctx = &link->dc->current_state->res_ctx.pipe_ctx[i];
-		if (pipe_ctx && pipe_ctx->stream && pipe_ctx->stream->link == link)
-			break;
-	}
-
-	if (pipe_ctx == NULL || pipe_ctx->stream == NULL)
-		return;
-
-	for (i = 0; i < MAX_PIPES; i++) {
-		pipe_ctx = &link->dc->current_state->res_ctx.pipe_ctx[i];
-		if (pipe_ctx && pipe_ctx->stream && !pipe_ctx->stream->dpms_off &&
-				pipe_ctx->stream->link == link && !pipe_ctx->prev_odm_pipe) {
-			core_link_disable_stream(pipe_ctx);
-		}
-	}
-
-	for (i = 0; i < MAX_PIPES; i++) {
-		pipe_ctx = &link->dc->current_state->res_ctx.pipe_ctx[i];
-		if (pipe_ctx && pipe_ctx->stream && !pipe_ctx->stream->dpms_off &&
-				pipe_ctx->stream->link == link && !pipe_ctx->prev_odm_pipe) {
-			core_link_enable_stream(link->dc->current_state, pipe_ctx);
-		}
-	}
-}
-
-bool dc_link_handle_hpd_rx_irq(struct dc_link *link, union hpd_irq_data *out_hpd_irq_dpcd_data, bool *out_link_loss,
-							bool defer_handling, bool *has_left_work)
-{
-	union hpd_irq_data hpd_irq_dpcd_data = {0};
-	union device_service_irq device_service_clear = {0};
-	enum dc_status result;
-	bool status = false;
-
-	if (out_link_loss)
-		*out_link_loss = false;
-
-	if (has_left_work)
-		*has_left_work = false;
-	/* For use cases related to down stream connection status change,
-	 * PSR and device auto test, refer to function handle_sst_hpd_irq
-	 * in DAL2.1*/
-
-	DC_LOG_HW_HPD_IRQ("%s: Got short pulse HPD on link %d\n",
-		__func__, link->link_index);
-
-
-	 /* All the "handle_hpd_irq_xxx()" methods
-		 * should be called only after
-		 * dal_dpsst_ls_read_hpd_irq_data
-		 * Order of calls is important too
-		 */
-	result = read_hpd_rx_irq_data(link, &hpd_irq_dpcd_data);
-	if (out_hpd_irq_dpcd_data)
-		*out_hpd_irq_dpcd_data = hpd_irq_dpcd_data;
-
-	if (result != DC_OK) {
-		DC_LOG_HW_HPD_IRQ("%s: DPCD read failed to obtain irq data\n",
-			__func__);
-		return false;
-	}
-
-	if (hpd_irq_dpcd_data.bytes.device_service_irq.bits.AUTOMATED_TEST) {
-		device_service_clear.bits.AUTOMATED_TEST = 1;
-		core_link_write_dpcd(
-			link,
-			DP_DEVICE_SERVICE_IRQ_VECTOR,
-			&device_service_clear.raw,
-			sizeof(device_service_clear.raw));
-		device_service_clear.raw = 0;
-		if (defer_handling && has_left_work)
-			*has_left_work = true;
-		else
-			dc_link_dp_handle_automated_test(link);
-		return false;
-	}
-
-	if (!dc_link_dp_allow_hpd_rx_irq(link)) {
-		DC_LOG_HW_HPD_IRQ("%s: skipping HPD handling on %d\n",
-			__func__, link->link_index);
-		return false;
-	}
-
-	if (handle_hpd_irq_psr_sink(link))
-		/* PSR-related error was detected and handled */
-		return true;
-
-	/* If PSR-related error handled, Main link may be off,
-	 * so do not handle as a normal sink status change interrupt.
-	 */
-
-	if (hpd_irq_dpcd_data.bytes.device_service_irq.bits.UP_REQ_MSG_RDY) {
-		if (defer_handling && has_left_work)
-			*has_left_work = true;
-		return true;
-	}
-
-	/* check if we have MST msg and return since we poll for it */
-	if (hpd_irq_dpcd_data.bytes.device_service_irq.bits.DOWN_REP_MSG_RDY) {
-		if (defer_handling && has_left_work)
-			*has_left_work = true;
-		return false;
-	}
-
-	/* For now we only handle 'Downstream port status' case.
-	 * If we got sink count changed it means
-	 * Downstream port status changed,
-	 * then DM should call DC to do the detection.
-	 * NOTE: Do not handle link loss on eDP since it is internal link*/
-	if ((link->connector_signal != SIGNAL_TYPE_EDP) &&
-		hpd_rx_irq_check_link_loss_status(
-			link,
-			&hpd_irq_dpcd_data)) {
-		/* Connectivity log: link loss */
-		CONN_DATA_LINK_LOSS(link,
-					hpd_irq_dpcd_data.raw,
-					sizeof(hpd_irq_dpcd_data),
-					"Status: ");
-
-		if (defer_handling && has_left_work)
-			*has_left_work = true;
-		else
-			dc_link_dp_handle_link_loss(link);
-
-		status = false;
-		if (out_link_loss)
-			*out_link_loss = true;
-	}
-
-	if (link->type == dc_connection_sst_branch &&
-		hpd_irq_dpcd_data.bytes.sink_cnt.bits.SINK_COUNT
-			!= link->dpcd_sink_count)
-		status = true;
-
-	/* reasons for HPD RX:
-	 * 1. Link Loss - ie Re-train the Link
-	 * 2. MST sideband message
-	 * 3. Automated Test - ie. Internal Commit
-	 * 4. CP (copy protection) - (not interesting for DM???)
-	 * 5. DRR
-	 * 6. Downstream Port status changed
-	 * -ie. Detect - this the only one
-	 * which is interesting for DM because
-	 * it must call dc_link_detect.
-	 */
-	return status;
-}
-
-/*query dpcd for version and mst cap addresses*/
-bool is_mst_supported(struct dc_link *link)
-{
-	bool mst          = false;
-	enum dc_status st = DC_OK;
-	union dpcd_rev rev;
-	union mstm_cap cap;
-
-	if (link->preferred_training_settings.mst_enable &&
-		*link->preferred_training_settings.mst_enable == false) {
-		return false;
-	}
-
-	rev.raw  = 0;
-	cap.raw  = 0;
-
-	st = core_link_read_dpcd(link, DP_DPCD_REV, &rev.raw,
-			sizeof(rev));
-
-	if (st == DC_OK && rev.raw >= DPCD_REV_12) {
-
-		st = core_link_read_dpcd(link, DP_MSTM_CAP,
-				&cap.raw, sizeof(cap));
-		if (st == DC_OK && cap.bits.MST_CAP == 1)
-			mst = true;
-	}
-	return mst;
-
-}
-
-bool is_dp_active_dongle(const struct dc_link *link)
-{
-	return (link->dpcd_caps.dongle_type >= DISPLAY_DONGLE_DP_VGA_CONVERTER) &&
-				(link->dpcd_caps.dongle_type <= DISPLAY_DONGLE_DP_HDMI_CONVERTER);
-}
-
-bool is_dp_branch_device(const struct dc_link *link)
-{
-	return link->dpcd_caps.is_branch_dev;
-}
-
-static int translate_dpcd_max_bpc(enum dpcd_downstream_port_max_bpc bpc)
-{
-	switch (bpc) {
-	case DOWN_STREAM_MAX_8BPC:
-		return 8;
-	case DOWN_STREAM_MAX_10BPC:
-		return 10;
-	case DOWN_STREAM_MAX_12BPC:
-		return 12;
-	case DOWN_STREAM_MAX_16BPC:
-		return 16;
-	default:
-		break;
-	}
-
-	return -1;
-}
-
-static void read_dp_device_vendor_id(struct dc_link *link)
-{
-	struct dp_device_vendor_id dp_id;
-
-	/* read IEEE branch device id */
-	core_link_read_dpcd(
-		link,
-		DP_BRANCH_OUI,
-		(uint8_t *)&dp_id,
-		sizeof(dp_id));
-
-	link->dpcd_caps.branch_dev_id =
-		(dp_id.ieee_oui[0] << 16) +
-		(dp_id.ieee_oui[1] << 8) +
-		dp_id.ieee_oui[2];
-
-	memmove(
-		link->dpcd_caps.branch_dev_name,
-		dp_id.ieee_device_id,
-		sizeof(dp_id.ieee_device_id));
-}
-
-
-
-static void get_active_converter_info(
-	uint8_t data, struct dc_link *link)
-{
-	union dp_downstream_port_present ds_port = { .byte = data };
-	memset(&link->dpcd_caps.dongle_caps, 0, sizeof(link->dpcd_caps.dongle_caps));
-
-	/* decode converter info*/
-	if (!ds_port.fields.PORT_PRESENT) {
-		link->dpcd_caps.dongle_type = DISPLAY_DONGLE_NONE;
-		ddc_service_set_dongle_type(link->ddc,
-				link->dpcd_caps.dongle_type);
-		link->dpcd_caps.is_branch_dev = false;
-		return;
-	}
-
-	/* DPCD 0x5 bit 0 = 1, it indicate it's branch device */
-	link->dpcd_caps.is_branch_dev = ds_port.fields.PORT_PRESENT;
-
-	switch (ds_port.fields.PORT_TYPE) {
-	case DOWNSTREAM_VGA:
-		link->dpcd_caps.dongle_type = DISPLAY_DONGLE_DP_VGA_CONVERTER;
-		break;
-	case DOWNSTREAM_DVI_HDMI_DP_PLUS_PLUS:
-		/* At this point we don't know is it DVI or HDMI or DP++,
-		 * assume DVI.*/
-		link->dpcd_caps.dongle_type = DISPLAY_DONGLE_DP_DVI_CONVERTER;
-		break;
-	default:
-		link->dpcd_caps.dongle_type = DISPLAY_DONGLE_NONE;
-		break;
-	}
-
-	if (link->dpcd_caps.dpcd_rev.raw >= DPCD_REV_11) {
-		uint8_t det_caps[16]; /* CTS 4.2.2.7 expects source to read Detailed Capabilities Info : 00080h-0008F.*/
-		union dwnstream_port_caps_byte0 *port_caps =
-			(union dwnstream_port_caps_byte0 *)det_caps;
-		if (core_link_read_dpcd(link, DP_DOWNSTREAM_PORT_0,
-				det_caps, sizeof(det_caps)) == DC_OK) {
-
-			switch (port_caps->bits.DWN_STRM_PORTX_TYPE) {
-			/*Handle DP case as DONGLE_NONE*/
-			case DOWN_STREAM_DETAILED_DP:
-				link->dpcd_caps.dongle_type = DISPLAY_DONGLE_NONE;
-				break;
-			case DOWN_STREAM_DETAILED_VGA:
-				link->dpcd_caps.dongle_type =
-					DISPLAY_DONGLE_DP_VGA_CONVERTER;
-				break;
-			case DOWN_STREAM_DETAILED_DVI:
-				link->dpcd_caps.dongle_type =
-					DISPLAY_DONGLE_DP_DVI_CONVERTER;
-				break;
-			case DOWN_STREAM_DETAILED_HDMI:
-			case DOWN_STREAM_DETAILED_DP_PLUS_PLUS:
-				/*Handle DP++ active converter case, process DP++ case as HDMI case according DP1.4 spec*/
-				link->dpcd_caps.dongle_type =
-					DISPLAY_DONGLE_DP_HDMI_CONVERTER;
-
-				link->dpcd_caps.dongle_caps.dongle_type = link->dpcd_caps.dongle_type;
-				if (ds_port.fields.DETAILED_CAPS) {
-
-					union dwnstream_port_caps_byte3_hdmi
-						hdmi_caps = {.raw = det_caps[3] };
-					union dwnstream_port_caps_byte2
-						hdmi_color_caps = {.raw = det_caps[2] };
-					link->dpcd_caps.dongle_caps.dp_hdmi_max_pixel_clk_in_khz =
-						det_caps[1] * 2500;
-
-					link->dpcd_caps.dongle_caps.is_dp_hdmi_s3d_converter =
-						hdmi_caps.bits.FRAME_SEQ_TO_FRAME_PACK;
-					/*YCBCR capability only for HDMI case*/
-					if (port_caps->bits.DWN_STRM_PORTX_TYPE
-							== DOWN_STREAM_DETAILED_HDMI) {
-						link->dpcd_caps.dongle_caps.is_dp_hdmi_ycbcr422_pass_through =
-								hdmi_caps.bits.YCrCr422_PASS_THROUGH;
-						link->dpcd_caps.dongle_caps.is_dp_hdmi_ycbcr420_pass_through =
-								hdmi_caps.bits.YCrCr420_PASS_THROUGH;
-						link->dpcd_caps.dongle_caps.is_dp_hdmi_ycbcr422_converter =
-								hdmi_caps.bits.YCrCr422_CONVERSION;
-						link->dpcd_caps.dongle_caps.is_dp_hdmi_ycbcr420_converter =
-								hdmi_caps.bits.YCrCr420_CONVERSION;
-					}
-
-					link->dpcd_caps.dongle_caps.dp_hdmi_max_bpc =
-						translate_dpcd_max_bpc(
-							hdmi_color_caps.bits.MAX_BITS_PER_COLOR_COMPONENT);
-
-					if (link->dpcd_caps.dongle_caps.dp_hdmi_max_pixel_clk_in_khz != 0)
-						link->dpcd_caps.dongle_caps.extendedCapValid = true;
-				}
-
-				break;
-			}
-		}
-	}
-
-	ddc_service_set_dongle_type(link->ddc, link->dpcd_caps.dongle_type);
-
-	{
-		struct dp_sink_hw_fw_revision dp_hw_fw_revision;
-
-		core_link_read_dpcd(
-			link,
-			DP_BRANCH_REVISION_START,
-			(uint8_t *)&dp_hw_fw_revision,
-			sizeof(dp_hw_fw_revision));
-
-		link->dpcd_caps.branch_hw_revision =
-			dp_hw_fw_revision.ieee_hw_rev;
-
-		memmove(
-			link->dpcd_caps.branch_fw_revision,
-			dp_hw_fw_revision.ieee_fw_rev,
-			sizeof(dp_hw_fw_revision.ieee_fw_rev));
-	}
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	if (link->dpcd_caps.dpcd_rev.raw >= DPCD_REV_14 &&
-			link->dpcd_caps.dongle_type != DISPLAY_DONGLE_NONE) {
-		union dp_dfp_cap_ext dfp_cap_ext;
-		memset(&dfp_cap_ext, '\0', sizeof (dfp_cap_ext));
-		core_link_read_dpcd(
-				link,
-				DP_DFP_CAPABILITY_EXTENSION_SUPPORT,
-				dfp_cap_ext.raw,
-				sizeof(dfp_cap_ext.raw));
-		link->dpcd_caps.dongle_caps.dfp_cap_ext.supported = dfp_cap_ext.fields.supported;
-		link->dpcd_caps.dongle_caps.dfp_cap_ext.max_pixel_rate_in_mps =
-				dfp_cap_ext.fields.max_pixel_rate_in_mps[0] +
-				(dfp_cap_ext.fields.max_pixel_rate_in_mps[1] << 8);
-		link->dpcd_caps.dongle_caps.dfp_cap_ext.max_video_h_active_width =
-				dfp_cap_ext.fields.max_video_h_active_width[0] +
-				(dfp_cap_ext.fields.max_video_h_active_width[1] << 8);
-		link->dpcd_caps.dongle_caps.dfp_cap_ext.max_video_v_active_height =
-				dfp_cap_ext.fields.max_video_v_active_height[0] +
-				(dfp_cap_ext.fields.max_video_v_active_height[1] << 8);
-		link->dpcd_caps.dongle_caps.dfp_cap_ext.encoding_format_caps =
-				dfp_cap_ext.fields.encoding_format_caps;
-		link->dpcd_caps.dongle_caps.dfp_cap_ext.rgb_color_depth_caps =
-				dfp_cap_ext.fields.rgb_color_depth_caps;
-		link->dpcd_caps.dongle_caps.dfp_cap_ext.ycbcr444_color_depth_caps =
-				dfp_cap_ext.fields.ycbcr444_color_depth_caps;
-		link->dpcd_caps.dongle_caps.dfp_cap_ext.ycbcr422_color_depth_caps =
-				dfp_cap_ext.fields.ycbcr422_color_depth_caps;
-		link->dpcd_caps.dongle_caps.dfp_cap_ext.ycbcr420_color_depth_caps =
-				dfp_cap_ext.fields.ycbcr420_color_depth_caps;
-		DC_LOG_DP2("DFP capability extension is read at link %d", link->link_index);
-		DC_LOG_DP2("\tdfp_cap_ext.supported = %s", link->dpcd_caps.dongle_caps.dfp_cap_ext.supported ? "true" : "false");
-		DC_LOG_DP2("\tdfp_cap_ext.max_pixel_rate_in_mps = %d", link->dpcd_caps.dongle_caps.dfp_cap_ext.max_pixel_rate_in_mps);
-		DC_LOG_DP2("\tdfp_cap_ext.max_video_h_active_width = %d", link->dpcd_caps.dongle_caps.dfp_cap_ext.max_video_h_active_width);
-		DC_LOG_DP2("\tdfp_cap_ext.max_video_v_active_height = %d", link->dpcd_caps.dongle_caps.dfp_cap_ext.max_video_v_active_height);
-	}
-#endif
-}
-
-static void dp_wa_power_up_0010FA(struct dc_link *link, uint8_t *dpcd_data,
-		int length)
-{
-	int retry = 0;
-
-	if (!link->dpcd_caps.dpcd_rev.raw) {
-		do {
-			dp_receiver_power_ctrl(link, true);
-			core_link_read_dpcd(link, DP_DPCD_REV,
-							dpcd_data, length);
-			link->dpcd_caps.dpcd_rev.raw = dpcd_data[
-				DP_DPCD_REV -
-				DP_DPCD_REV];
-		} while (retry++ < 4 && !link->dpcd_caps.dpcd_rev.raw);
-	}
-
-	if (link->dpcd_caps.dongle_type == DISPLAY_DONGLE_DP_VGA_CONVERTER) {
-		switch (link->dpcd_caps.branch_dev_id) {
-		/* 0010FA active dongles (DP-VGA, DP-DLDVI converters) power down
-		 * all internal circuits including AUX communication preventing
-		 * reading DPCD table and EDID (spec violation).
-		 * Encoder will skip DP RX power down on disable_output to
-		 * keep receiver powered all the time.*/
-		case DP_BRANCH_DEVICE_ID_0010FA:
-		case DP_BRANCH_DEVICE_ID_0080E1:
-		case DP_BRANCH_DEVICE_ID_00E04C:
-			link->wa_flags.dp_keep_receiver_powered = true;
-			break;
-
-		/* TODO: May need work around for other dongles. */
-		default:
-			link->wa_flags.dp_keep_receiver_powered = false;
-			break;
-		}
-	} else
-		link->wa_flags.dp_keep_receiver_powered = false;
-}
-
-/* Read additional sink caps defined in source specific DPCD area
- * This function currently only reads from SinkCapability address (DP_SOURCE_SINK_CAP)
- */
-static bool dpcd_read_sink_ext_caps(struct dc_link *link)
-{
-	uint8_t dpcd_data;
-
-	if (!link)
-		return false;
-
-	if (core_link_read_dpcd(link, DP_SOURCE_SINK_CAP, &dpcd_data, 1) != DC_OK)
-		return false;
-
-	link->dpcd_sink_ext_caps.raw = dpcd_data;
-	return true;
-}
-
-bool dp_retrieve_lttpr_cap(struct dc_link *link)
-{
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	uint8_t lttpr_dpcd_data[8];
-	bool allow_lttpr_non_transparent_mode = 0;
-#else
-	uint8_t lttpr_dpcd_data[6];
-#endif
-	bool vbios_lttpr_enable = link->dc->caps.vbios_lttpr_enable;
-	bool vbios_lttpr_interop = link->dc->caps.vbios_lttpr_aware;
-	enum dc_status status = DC_ERROR_UNEXPECTED;
-	bool is_lttpr_present = false;
-
-	memset(lttpr_dpcd_data, '\0', sizeof(lttpr_dpcd_data));
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	if ((link->dc->config.allow_lttpr_non_transparent_mode.bits.DP2_0 &&
-			link->dpcd_caps.channel_coding_cap.bits.DP_128b_132b_SUPPORTED)) {
-		allow_lttpr_non_transparent_mode = 1;
-	} else if (link->dc->config.allow_lttpr_non_transparent_mode.bits.DP1_4A &&
-			!link->dpcd_caps.channel_coding_cap.bits.DP_128b_132b_SUPPORTED) {
-		allow_lttpr_non_transparent_mode = 1;
-	}
-#endif
-
-	/*
-	 * Logic to determine LTTPR mode
-	 */
-	link->lttpr_mode = LTTPR_MODE_NON_LTTPR;
-	if (vbios_lttpr_enable && vbios_lttpr_interop)
-		link->lttpr_mode = LTTPR_MODE_NON_TRANSPARENT;
-	else if (!vbios_lttpr_enable && vbios_lttpr_interop) {
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-		if (allow_lttpr_non_transparent_mode)
-#else
-		if (link->dc->config.allow_lttpr_non_transparent_mode)
-#endif
-			link->lttpr_mode = LTTPR_MODE_NON_TRANSPARENT;
-		else
-			link->lttpr_mode = LTTPR_MODE_TRANSPARENT;
-	} else if (!vbios_lttpr_enable && !vbios_lttpr_interop) {
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-		if (!allow_lttpr_non_transparent_mode || !link->dc->caps.extended_aux_timeout_support)
-#else
-		if (!link->dc->config.allow_lttpr_non_transparent_mode
-			|| !link->dc->caps.extended_aux_timeout_support)
-#endif
-			link->lttpr_mode = LTTPR_MODE_NON_LTTPR;
-		else
-			link->lttpr_mode = LTTPR_MODE_NON_TRANSPARENT;
-	}
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	/* Check DP tunnel LTTPR mode debug option. */
-	if (link->ep_type == DISPLAY_ENDPOINT_USB4_DPIA &&
-	    link->dc->debug.dpia_debug.bits.force_non_lttpr)
-		link->lttpr_mode = LTTPR_MODE_NON_LTTPR;
-#endif
-
-	if (link->lttpr_mode == LTTPR_MODE_NON_TRANSPARENT || link->lttpr_mode == LTTPR_MODE_TRANSPARENT) {
-		/* By reading LTTPR capability, RX assumes that we will enable
-		 * LTTPR extended aux timeout if LTTPR is present.
-		 */
-		status = core_link_read_dpcd(
-				link,
-				DP_LT_TUNABLE_PHY_REPEATER_FIELD_DATA_STRUCTURE_REV,
-				lttpr_dpcd_data,
-				sizeof(lttpr_dpcd_data));
-		if (status != DC_OK) {
-			dm_error("%s: Read LTTPR caps data failed.\n", __func__);
-			return false;
-		}
-
-		link->dpcd_caps.lttpr_caps.revision.raw =
-				lttpr_dpcd_data[DP_LT_TUNABLE_PHY_REPEATER_FIELD_DATA_STRUCTURE_REV -
-								DP_LT_TUNABLE_PHY_REPEATER_FIELD_DATA_STRUCTURE_REV];
-
-		link->dpcd_caps.lttpr_caps.max_link_rate =
-				lttpr_dpcd_data[DP_MAX_LINK_RATE_PHY_REPEATER -
-								DP_LT_TUNABLE_PHY_REPEATER_FIELD_DATA_STRUCTURE_REV];
-
-		link->dpcd_caps.lttpr_caps.phy_repeater_cnt =
-				lttpr_dpcd_data[DP_PHY_REPEATER_CNT -
-								DP_LT_TUNABLE_PHY_REPEATER_FIELD_DATA_STRUCTURE_REV];
-
-		link->dpcd_caps.lttpr_caps.max_lane_count =
-				lttpr_dpcd_data[DP_MAX_LANE_COUNT_PHY_REPEATER -
-								DP_LT_TUNABLE_PHY_REPEATER_FIELD_DATA_STRUCTURE_REV];
-
-		link->dpcd_caps.lttpr_caps.mode =
-				lttpr_dpcd_data[DP_PHY_REPEATER_MODE -
-								DP_LT_TUNABLE_PHY_REPEATER_FIELD_DATA_STRUCTURE_REV];
-
-		link->dpcd_caps.lttpr_caps.max_ext_timeout =
-				lttpr_dpcd_data[DP_PHY_REPEATER_EXTENDED_WAIT_TIMEOUT -
-								DP_LT_TUNABLE_PHY_REPEATER_FIELD_DATA_STRUCTURE_REV];
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-		link->dpcd_caps.lttpr_caps.main_link_channel_coding.raw =
-				lttpr_dpcd_data[DP_MAIN_LINK_CHANNEL_CODING_PHY_REPEATER -
-								DP_LT_TUNABLE_PHY_REPEATER_FIELD_DATA_STRUCTURE_REV];
-
-		link->dpcd_caps.lttpr_caps.supported_128b_132b_rates.raw =
-				lttpr_dpcd_data[DP_PHY_REPEATER_128b_132b_RATES -
-								DP_LT_TUNABLE_PHY_REPEATER_FIELD_DATA_STRUCTURE_REV];
-#endif
-
-		/* Attempt to train in LTTPR transparent mode if repeater count exceeds 8. */
-		is_lttpr_present = (dp_convert_to_count(link->dpcd_caps.lttpr_caps.phy_repeater_cnt) != 0 &&
-				link->dpcd_caps.lttpr_caps.phy_repeater_cnt < 0xff &&
-				link->dpcd_caps.lttpr_caps.max_lane_count > 0 &&
-				link->dpcd_caps.lttpr_caps.max_lane_count <= 4 &&
-				link->dpcd_caps.lttpr_caps.revision.raw >= 0x14);
-		if (is_lttpr_present) {
-			CONN_DATA_DETECT(link, lttpr_dpcd_data, sizeof(lttpr_dpcd_data), "LTTPR Caps: ");
-			configure_lttpr_mode_transparent(link);
-		} else
-			link->lttpr_mode = LTTPR_MODE_NON_LTTPR;
-	}
-	return is_lttpr_present;
-}
-
-static bool retrieve_link_cap(struct dc_link *link)
-{
-	/* DP_ADAPTER_CAP - DP_DPCD_REV + 1 == 16 and also DP_DSC_BITS_PER_PIXEL_INC - DP_DSC_SUPPORT + 1 == 16,
-	 * which means size 16 will be good for both of those DPCD register block reads
-	 */
-	uint8_t dpcd_data[16];
-	/*Only need to read 1 byte starting from DP_DPRX_FEATURE_ENUMERATION_LIST.
-	 */
-	uint8_t dpcd_dprx_data = '\0';
-	uint8_t dpcd_power_state = '\0';
-
-	struct dp_device_vendor_id sink_id;
-	union down_stream_port_count down_strm_port_count;
-	union edp_configuration_cap edp_config_cap;
-	union dp_downstream_port_present ds_port = { 0 };
-	enum dc_status status = DC_ERROR_UNEXPECTED;
-	uint32_t read_dpcd_retry_cnt = 3;
-	int i;
-	struct dp_sink_hw_fw_revision dp_hw_fw_revision;
-	const uint32_t post_oui_delay = 30; // 30ms
-	bool is_lttpr_present = false;
-
-	memset(dpcd_data, '\0', sizeof(dpcd_data));
-	memset(&down_strm_port_count,
-		'\0', sizeof(union down_stream_port_count));
-	memset(&edp_config_cap, '\0',
-		sizeof(union edp_configuration_cap));
-
-	/* if extended timeout is supported in hardware,
-	 * default to LTTPR timeout (3.2ms) first as a W/A for DP link layer
-	 * CTS 4.2.1.1 regression introduced by CTS specs requirement update.
-	 */
-	dc_link_aux_try_to_configure_timeout(link->ddc,
-			LINK_AUX_DEFAULT_LTTPR_TIMEOUT_PERIOD);
-
-	is_lttpr_present = dp_retrieve_lttpr_cap(link);
-	/* Read DP tunneling information. */
-	status = dpcd_get_tunneling_device_data(link);
-
-	status = core_link_read_dpcd(link, DP_SET_POWER,
-			&dpcd_power_state, sizeof(dpcd_power_state));
-
-	/* Delay 1 ms if AUX CH is in power down state. Based on spec
-	 * section 2.3.1.2, if AUX CH may be powered down due to
-	 * write to DPCD 600h = 2. Sink AUX CH is monitoring differential
-	 * signal and may need up to 1 ms before being able to reply.
-	 */
-	if (status != DC_OK || dpcd_power_state == DP_SET_POWER_D3)
-		udelay(1000);
-
-	dpcd_set_source_specific_data(link);
-	/* Sink may need to configure internals based on vendor, so allow some
-	 * time before proceeding with possibly vendor specific transactions
-	 */
-	msleep(post_oui_delay);
-
-	for (i = 0; i < read_dpcd_retry_cnt; i++) {
-		status = core_link_read_dpcd(
-				link,
-				DP_DPCD_REV,
-				dpcd_data,
-				sizeof(dpcd_data));
-		if (status == DC_OK)
-			break;
-	}
-
-	if (status != DC_OK) {
-		dm_error("%s: Read receiver caps dpcd data failed.\n", __func__);
-		return false;
-	}
-
-	if (!is_lttpr_present)
-		dc_link_aux_try_to_configure_timeout(link->ddc, LINK_AUX_DEFAULT_TIMEOUT_PERIOD);
-
-	{
-		union training_aux_rd_interval aux_rd_interval;
-
-		aux_rd_interval.raw =
-			dpcd_data[DP_TRAINING_AUX_RD_INTERVAL];
-
-		link->dpcd_caps.ext_receiver_cap_field_present =
-				aux_rd_interval.bits.EXT_RECEIVER_CAP_FIELD_PRESENT == 1;
-
-		if (aux_rd_interval.bits.EXT_RECEIVER_CAP_FIELD_PRESENT == 1) {
-			uint8_t ext_cap_data[16];
-
-			memset(ext_cap_data, '\0', sizeof(ext_cap_data));
-			for (i = 0; i < read_dpcd_retry_cnt; i++) {
-				status = core_link_read_dpcd(
-				link,
-				DP_DP13_DPCD_REV,
-				ext_cap_data,
-				sizeof(ext_cap_data));
-				if (status == DC_OK) {
-					memcpy(dpcd_data, ext_cap_data, sizeof(dpcd_data));
-					break;
-				}
-			}
-			if (status != DC_OK)
-				dm_error("%s: Read extend caps data failed, use cap from dpcd 0.\n", __func__);
-		}
-	}
-
-	link->dpcd_caps.dpcd_rev.raw =
-			dpcd_data[DP_DPCD_REV - DP_DPCD_REV];
-
-	if (link->dpcd_caps.ext_receiver_cap_field_present) {
-		for (i = 0; i < read_dpcd_retry_cnt; i++) {
-			status = core_link_read_dpcd(
-					link,
-					DP_DPRX_FEATURE_ENUMERATION_LIST,
-					&dpcd_dprx_data,
-					sizeof(dpcd_dprx_data));
-			if (status == DC_OK)
-				break;
-		}
-
-		link->dpcd_caps.dprx_feature.raw = dpcd_dprx_data;
-
-		if (status != DC_OK)
-			dm_error("%s: Read DPRX caps data failed.\n", __func__);
-	}
-
-	else {
-		link->dpcd_caps.dprx_feature.raw = 0;
-	}
-
-
-	/* Error condition checking...
-	 * It is impossible for Sink to report Max Lane Count = 0.
-	 * It is possible for Sink to report Max Link Rate = 0, if it is
-	 * an eDP device that is reporting specialized link rates in the
-	 * SUPPORTED_LINK_RATE table.
-	 */
-	if (dpcd_data[DP_MAX_LANE_COUNT - DP_DPCD_REV] == 0)
-		return false;
-
-	ds_port.byte = dpcd_data[DP_DOWNSTREAMPORT_PRESENT -
-				 DP_DPCD_REV];
-
-	read_dp_device_vendor_id(link);
-
-	get_active_converter_info(ds_port.byte, link);
-
-	dp_wa_power_up_0010FA(link, dpcd_data, sizeof(dpcd_data));
-
-	down_strm_port_count.raw = dpcd_data[DP_DOWN_STREAM_PORT_COUNT -
-				 DP_DPCD_REV];
-
-	link->dpcd_caps.allow_invalid_MSA_timing_param =
-		down_strm_port_count.bits.IGNORE_MSA_TIMING_PARAM;
-
-	link->dpcd_caps.max_ln_count.raw = dpcd_data[
-		DP_MAX_LANE_COUNT - DP_DPCD_REV];
-
-	link->dpcd_caps.max_down_spread.raw = dpcd_data[
-		DP_MAX_DOWNSPREAD - DP_DPCD_REV];
-
-	link->reported_link_cap.lane_count =
-		link->dpcd_caps.max_ln_count.bits.MAX_LANE_COUNT;
-	link->reported_link_cap.link_rate = dpcd_data[
-		DP_MAX_LINK_RATE - DP_DPCD_REV];
-	link->reported_link_cap.link_spread =
-		link->dpcd_caps.max_down_spread.bits.MAX_DOWN_SPREAD ?
-		LINK_SPREAD_05_DOWNSPREAD_30KHZ : LINK_SPREAD_DISABLED;
-
-	edp_config_cap.raw = dpcd_data[
-		DP_EDP_CONFIGURATION_CAP - DP_DPCD_REV];
-	link->dpcd_caps.panel_mode_edp =
-		edp_config_cap.bits.ALT_SCRAMBLER_RESET;
-	link->dpcd_caps.dpcd_display_control_capable =
-		edp_config_cap.bits.DPCD_DISPLAY_CONTROL_CAPABLE;
-
-	link->test_pattern_enabled = false;
-	link->compliance_test_state.raw = 0;
-
-	/* read sink count */
-	core_link_read_dpcd(link,
-			DP_SINK_COUNT,
-			&link->dpcd_caps.sink_count.raw,
-			sizeof(link->dpcd_caps.sink_count.raw));
-
-	/* read sink ieee oui */
-	core_link_read_dpcd(link,
-			DP_SINK_OUI,
-			(uint8_t *)(&sink_id),
-			sizeof(sink_id));
-
-	link->dpcd_caps.sink_dev_id =
-			(sink_id.ieee_oui[0] << 16) +
-			(sink_id.ieee_oui[1] << 8) +
-			(sink_id.ieee_oui[2]);
-
-	memmove(
-		link->dpcd_caps.sink_dev_id_str,
-		sink_id.ieee_device_id,
-		sizeof(sink_id.ieee_device_id));
-
-	/* Quirk Apple MBP 2017 15" Retina panel: Wrong DP_MAX_LINK_RATE */
-	{
-		uint8_t str_mbp_2017[] = { 101, 68, 21, 101, 98, 97 };
-
-		if ((link->dpcd_caps.sink_dev_id == 0x0010fa) &&
-		    !memcmp(link->dpcd_caps.sink_dev_id_str, str_mbp_2017,
-			    sizeof(str_mbp_2017))) {
-			link->reported_link_cap.link_rate = 0x0c;
-		}
-	}
-
-	core_link_read_dpcd(
-		link,
-		DP_SINK_HW_REVISION_START,
-		(uint8_t *)&dp_hw_fw_revision,
-		sizeof(dp_hw_fw_revision));
-
-	link->dpcd_caps.sink_hw_revision =
-		dp_hw_fw_revision.ieee_hw_rev;
-
-	memmove(
-		link->dpcd_caps.sink_fw_revision,
-		dp_hw_fw_revision.ieee_fw_rev,
-		sizeof(dp_hw_fw_revision.ieee_fw_rev));
-
-	memset(&link->dpcd_caps.dsc_caps, '\0',
-			sizeof(link->dpcd_caps.dsc_caps));
-	memset(&link->dpcd_caps.fec_cap, '\0', sizeof(link->dpcd_caps.fec_cap));
-	/* Read DSC and FEC sink capabilities if DP revision is 1.4 and up */
-	if (link->dpcd_caps.dpcd_rev.raw >= DPCD_REV_14) {
-		status = core_link_read_dpcd(
-				link,
-				DP_FEC_CAPABILITY,
-				&link->dpcd_caps.fec_cap.raw,
-				sizeof(link->dpcd_caps.fec_cap.raw));
-		status = core_link_read_dpcd(
-				link,
-				DP_DSC_SUPPORT,
-				link->dpcd_caps.dsc_caps.dsc_basic_caps.raw,
-				sizeof(link->dpcd_caps.dsc_caps.dsc_basic_caps.raw));
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-		if (link->dpcd_caps.dongle_type != DISPLAY_DONGLE_NONE) {
-			status = core_link_read_dpcd(
-					link,
-					DP_DSC_BRANCH_OVERALL_THROUGHPUT_0,
-					link->dpcd_caps.dsc_caps.dsc_branch_decoder_caps.raw,
-					sizeof(link->dpcd_caps.dsc_caps.dsc_branch_decoder_caps.raw));
-			DC_LOG_DSC("DSC branch decoder capability is read at link %d", link->link_index);
-			DC_LOG_DSC("\tBRANCH_OVERALL_THROUGHPUT_0 = 0x%02x",
-					link->dpcd_caps.dsc_caps.dsc_branch_decoder_caps.fields.BRANCH_OVERALL_THROUGHPUT_0);
-			DC_LOG_DSC("\tBRANCH_OVERALL_THROUGHPUT_1 = 0x%02x",
-					link->dpcd_caps.dsc_caps.dsc_branch_decoder_caps.fields.BRANCH_OVERALL_THROUGHPUT_1);
-			DC_LOG_DSC("\tBRANCH_MAX_LINE_WIDTH 0x%02x",
-					link->dpcd_caps.dsc_caps.dsc_branch_decoder_caps.fields.BRANCH_MAX_LINE_WIDTH);
-		}
-#else
-		status = core_link_read_dpcd(
-				link,
-				DP_DSC_BRANCH_OVERALL_THROUGHPUT_0,
-				link->dpcd_caps.dsc_caps.dsc_branch_decoder_caps.raw,
-				sizeof(link->dpcd_caps.dsc_caps.dsc_branch_decoder_caps.raw));
-#endif
-	}
-
-	if (!dpcd_read_sink_ext_caps(link))
-		link->dpcd_sink_ext_caps.raw = 0;
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	link->dpcd_caps.channel_coding_cap.raw = dpcd_data[DP_MAIN_LINK_CHANNEL_CODING_CAP - DP_DPCD_REV];
-
-	if (link->dpcd_caps.channel_coding_cap.bits.DP_128b_132b_SUPPORTED) {
-		DC_LOG_DP2("128b/132b encoding is supported at link %d", link->link_index);
-
-		core_link_read_dpcd(link,
-				DP_128b_132b_SUPPORTED_LINK_RATES,
-				&link->dpcd_caps.dp_128b_132b_supported_link_rates.raw,
-				sizeof(link->dpcd_caps.dp_128b_132b_supported_link_rates.raw));
-		if (link->dpcd_caps.dp_128b_132b_supported_link_rates.bits.UHBR20)
-			link->reported_link_cap.link_rate = LINK_RATE_UHBR20;
-		else if (link->dpcd_caps.dp_128b_132b_supported_link_rates.bits.UHBR13_5)
-			link->reported_link_cap.link_rate = LINK_RATE_UHBR13_5;
-		else if (link->dpcd_caps.dp_128b_132b_supported_link_rates.bits.UHBR10)
-			link->reported_link_cap.link_rate = LINK_RATE_UHBR10;
-		else
-			dm_error("%s: Invalid RX 128b_132b_supported_link_rates\n", __func__);
-		DC_LOG_DP2("128b/132b supported link rates is read at link %d", link->link_index);
-		DC_LOG_DP2("\tmax 128b/132b link rate support is %d.%d GHz",
-				link->reported_link_cap.link_rate / 100,
-				link->reported_link_cap.link_rate % 100);
-
-		core_link_read_dpcd(link,
-				DP_SINK_VIDEO_FALLBACK_FORMATS,
-				&link->dpcd_caps.fallback_formats.raw,
-				sizeof(link->dpcd_caps.fallback_formats.raw));
-		DC_LOG_DP2("sink video fallback format is read at link %d", link->link_index);
-		if (link->dpcd_caps.fallback_formats.bits.dp_1920x1080_60Hz_24bpp_support)
-			DC_LOG_DP2("\t1920x1080@60Hz 24bpp fallback format supported");
-		if (link->dpcd_caps.fallback_formats.bits.dp_1280x720_60Hz_24bpp_support)
-			DC_LOG_DP2("\t1280x720@60Hz 24bpp fallback format supported");
-		if (link->dpcd_caps.fallback_formats.bits.dp_1024x768_60Hz_24bpp_support)
-			DC_LOG_DP2("\t1024x768@60Hz 24bpp fallback format supported");
-		if (link->dpcd_caps.fallback_formats.raw == 0) {
-			DC_LOG_DP2("\tno supported fallback formats, assume 1920x1080@60Hz 24bpp is supported");
-			link->dpcd_caps.fallback_formats.bits.dp_1920x1080_60Hz_24bpp_support = 1;
-		}
-
-		core_link_read_dpcd(link,
-				DP_FEC_CAPABILITY_1,
-				&link->dpcd_caps.fec_cap1.raw,
-				sizeof(link->dpcd_caps.fec_cap1.raw));
-		DC_LOG_DP2("FEC CAPABILITY 1 is read at link %d", link->link_index);
-		if (link->dpcd_caps.fec_cap1.bits.AGGREGATED_ERROR_COUNTERS_CAPABLE)
-			DC_LOG_DP2("\tFEC aggregated error counters are supported");
-	}
-#endif
-
-	/* Connectivity log: detection */
-	CONN_DATA_DETECT(link, dpcd_data, sizeof(dpcd_data), "Rx Caps: ");
-
-	return true;
-}
-
-bool dp_overwrite_extended_receiver_cap(struct dc_link *link)
-{
-	uint8_t dpcd_data[16];
-	uint32_t read_dpcd_retry_cnt = 3;
-	enum dc_status status = DC_ERROR_UNEXPECTED;
-	union dp_downstream_port_present ds_port = { 0 };
-	union down_stream_port_count down_strm_port_count;
-	union edp_configuration_cap edp_config_cap;
-
-	int i;
-
-	for (i = 0; i < read_dpcd_retry_cnt; i++) {
-		status = core_link_read_dpcd(
-				link,
-				DP_DPCD_REV,
-				dpcd_data,
-				sizeof(dpcd_data));
-		if (status == DC_OK)
-			break;
-	}
-
-	link->dpcd_caps.dpcd_rev.raw =
-		dpcd_data[DP_DPCD_REV - DP_DPCD_REV];
-
-	if (dpcd_data[DP_MAX_LANE_COUNT - DP_DPCD_REV] == 0)
-		return false;
-
-	ds_port.byte = dpcd_data[DP_DOWNSTREAMPORT_PRESENT -
-			DP_DPCD_REV];
-
-	get_active_converter_info(ds_port.byte, link);
-
-	down_strm_port_count.raw = dpcd_data[DP_DOWN_STREAM_PORT_COUNT -
-			DP_DPCD_REV];
-
-	link->dpcd_caps.allow_invalid_MSA_timing_param =
-		down_strm_port_count.bits.IGNORE_MSA_TIMING_PARAM;
-
-	link->dpcd_caps.max_ln_count.raw = dpcd_data[
-		DP_MAX_LANE_COUNT - DP_DPCD_REV];
-
-	link->dpcd_caps.max_down_spread.raw = dpcd_data[
-		DP_MAX_DOWNSPREAD - DP_DPCD_REV];
-
-	link->reported_link_cap.lane_count =
-		link->dpcd_caps.max_ln_count.bits.MAX_LANE_COUNT;
-	link->reported_link_cap.link_rate = dpcd_data[
-		DP_MAX_LINK_RATE - DP_DPCD_REV];
-	link->reported_link_cap.link_spread =
-		link->dpcd_caps.max_down_spread.bits.MAX_DOWN_SPREAD ?
-		LINK_SPREAD_05_DOWNSPREAD_30KHZ : LINK_SPREAD_DISABLED;
-
-	edp_config_cap.raw = dpcd_data[
-		DP_EDP_CONFIGURATION_CAP - DP_DPCD_REV];
-	link->dpcd_caps.panel_mode_edp =
-		edp_config_cap.bits.ALT_SCRAMBLER_RESET;
-	link->dpcd_caps.dpcd_display_control_capable =
-		edp_config_cap.bits.DPCD_DISPLAY_CONTROL_CAPABLE;
-
-	return true;
-}
-
-bool detect_dp_sink_caps(struct dc_link *link)
-{
-	return retrieve_link_cap(link);
-
-	/* dc init_hw has power encoder using default
-	 * signal for connector. For native DP, no
-	 * need to power up encoder again. If not native
-	 * DP, hw_init may need check signal or power up
-	 * encoder here.
-	 */
-	/* TODO save sink caps in link->sink */
-}
-
-static enum dc_link_rate linkRateInKHzToLinkRateMultiplier(uint32_t link_rate_in_khz)
-{
-	enum dc_link_rate link_rate;
-	// LinkRate is normally stored as a multiplier of 0.27 Gbps per lane. Do the translation.
-	switch (link_rate_in_khz) {
-	case 1620000:
-		link_rate = LINK_RATE_LOW;		// Rate_1 (RBR)		- 1.62 Gbps/Lane
-		break;
-	case 2160000:
-		link_rate = LINK_RATE_RATE_2;	// Rate_2			- 2.16 Gbps/Lane
-		break;
-	case 2430000:
-		link_rate = LINK_RATE_RATE_3;	// Rate_3			- 2.43 Gbps/Lane
-		break;
-	case 2700000:
-		link_rate = LINK_RATE_HIGH;		// Rate_4 (HBR)		- 2.70 Gbps/Lane
-		break;
-	case 3240000:
-		link_rate = LINK_RATE_RBR2;		// Rate_5 (RBR2)	- 3.24 Gbps/Lane
-		break;
-	case 4320000:
-		link_rate = LINK_RATE_RATE_6;	// Rate_6			- 4.32 Gbps/Lane
-		break;
-	case 5400000:
-		link_rate = LINK_RATE_HIGH2;	// Rate_7 (HBR2)	- 5.40 Gbps/Lane
-		break;
-	case 8100000:
-		link_rate = LINK_RATE_HIGH3;	// Rate_8 (HBR3)	- 8.10 Gbps/Lane
-		break;
-	default:
-		link_rate = LINK_RATE_UNKNOWN;
-		break;
-	}
-	return link_rate;
-}
-
-void detect_edp_sink_caps(struct dc_link *link)
-{
-	uint8_t supported_link_rates[16];
-	uint32_t entry;
-	uint32_t link_rate_in_khz;
-	enum dc_link_rate link_rate = LINK_RATE_UNKNOWN;
-	uint8_t backlight_adj_cap;
-
-	retrieve_link_cap(link);
-	link->dpcd_caps.edp_supported_link_rates_count = 0;
-	memset(supported_link_rates, 0, sizeof(supported_link_rates));
-
-	/*
-	 * edp_supported_link_rates_count is only valid for eDP v1.4 or higher.
-	 * Per VESA eDP spec, "The DPCD revision for eDP v1.4 is 13h"
-	 */
-	if (link->dpcd_caps.dpcd_rev.raw >= DPCD_REV_13 &&
-			(link->dc->debug.optimize_edp_link_rate ||
-			link->reported_link_cap.link_rate == LINK_RATE_UNKNOWN)) {
-		// Read DPCD 00010h - 0001Fh 16 bytes at one shot
-		core_link_read_dpcd(link, DP_SUPPORTED_LINK_RATES,
-							supported_link_rates, sizeof(supported_link_rates));
-
-		for (entry = 0; entry < 16; entry += 2) {
-			// DPCD register reports per-lane link rate = 16-bit link rate capability
-			// value X 200 kHz. Need multiplier to find link rate in kHz.
-			link_rate_in_khz = (supported_link_rates[entry+1] * 0x100 +
-										supported_link_rates[entry]) * 200;
-
-			if (link_rate_in_khz != 0) {
-				link_rate = linkRateInKHzToLinkRateMultiplier(link_rate_in_khz);
-				link->dpcd_caps.edp_supported_link_rates[link->dpcd_caps.edp_supported_link_rates_count] = link_rate;
-				link->dpcd_caps.edp_supported_link_rates_count++;
-
-				if (link->reported_link_cap.link_rate < link_rate)
-					link->reported_link_cap.link_rate = link_rate;
-			}
-		}
-	}
-	link->verified_link_cap = link->reported_link_cap;
-
-	core_link_read_dpcd(link, DP_EDP_BACKLIGHT_ADJUSTMENT_CAP,
-						&backlight_adj_cap, sizeof(backlight_adj_cap));
-
-	link->dpcd_caps.dynamic_backlight_capable_edp =
-				(backlight_adj_cap & DP_EDP_DYNAMIC_BACKLIGHT_CAP) ? true:false;
-
-	dc_link_set_default_brightness_aux(link);
-}
-
-void dc_link_dp_enable_hpd(const struct dc_link *link)
-{
-	struct link_encoder *encoder = link->link_enc;
-
-	if (encoder != NULL && encoder->funcs->enable_hpd != NULL)
-		encoder->funcs->enable_hpd(encoder);
-}
-
-void dc_link_dp_disable_hpd(const struct dc_link *link)
-{
-	struct link_encoder *encoder = link->link_enc;
-
-	if (encoder != NULL && encoder->funcs->enable_hpd != NULL)
-		encoder->funcs->disable_hpd(encoder);
-}
-
-static bool is_dp_phy_pattern(enum dp_test_pattern test_pattern)
-{
-	if ((DP_TEST_PATTERN_PHY_PATTERN_BEGIN <= test_pattern &&
-			test_pattern <= DP_TEST_PATTERN_PHY_PATTERN_END) ||
-			test_pattern == DP_TEST_PATTERN_VIDEO_MODE)
-		return true;
-	else
-		return false;
-}
-
-static void set_crtc_test_pattern(struct dc_link *link,
-				struct pipe_ctx *pipe_ctx,
-				enum dp_test_pattern test_pattern,
-				enum dp_test_pattern_color_space test_pattern_color_space)
-{
-	enum controller_dp_test_pattern controller_test_pattern;
-	enum dc_color_depth color_depth = pipe_ctx->
-		stream->timing.display_color_depth;
-	struct bit_depth_reduction_params params;
-	struct output_pixel_processor *opp = pipe_ctx->stream_res.opp;
-	int width = pipe_ctx->stream->timing.h_addressable +
-		pipe_ctx->stream->timing.h_border_left +
-		pipe_ctx->stream->timing.h_border_right;
-	int height = pipe_ctx->stream->timing.v_addressable +
-		pipe_ctx->stream->timing.v_border_bottom +
-		pipe_ctx->stream->timing.v_border_top;
-
-	memset(&params, 0, sizeof(params));
-
-	switch (test_pattern) {
-	case DP_TEST_PATTERN_COLOR_SQUARES:
-		controller_test_pattern =
-				CONTROLLER_DP_TEST_PATTERN_COLORSQUARES;
-	break;
-	case DP_TEST_PATTERN_COLOR_SQUARES_CEA:
-		controller_test_pattern =
-				CONTROLLER_DP_TEST_PATTERN_COLORSQUARES_CEA;
-	break;
-	case DP_TEST_PATTERN_VERTICAL_BARS:
-		controller_test_pattern =
-				CONTROLLER_DP_TEST_PATTERN_VERTICALBARS;
-	break;
-	case DP_TEST_PATTERN_HORIZONTAL_BARS:
-		controller_test_pattern =
-				CONTROLLER_DP_TEST_PATTERN_HORIZONTALBARS;
-	break;
-	case DP_TEST_PATTERN_COLOR_RAMP:
-		controller_test_pattern =
-				CONTROLLER_DP_TEST_PATTERN_COLORRAMP;
-	break;
-	default:
-		controller_test_pattern =
-				CONTROLLER_DP_TEST_PATTERN_VIDEOMODE;
-	break;
-	}
-
-	switch (test_pattern) {
-	case DP_TEST_PATTERN_COLOR_SQUARES:
-	case DP_TEST_PATTERN_COLOR_SQUARES_CEA:
-	case DP_TEST_PATTERN_VERTICAL_BARS:
-	case DP_TEST_PATTERN_HORIZONTAL_BARS:
-	case DP_TEST_PATTERN_COLOR_RAMP:
-	{
-		/* disable bit depth reduction */
-		pipe_ctx->stream->bit_depth_params = params;
-		opp->funcs->opp_program_bit_depth_reduction(opp, &params);
-		if (pipe_ctx->stream_res.tg->funcs->set_test_pattern)
-			pipe_ctx->stream_res.tg->funcs->set_test_pattern(pipe_ctx->stream_res.tg,
-				controller_test_pattern, color_depth);
-		else if (link->dc->hwss.set_disp_pattern_generator) {
-			struct pipe_ctx *odm_pipe;
-			enum controller_dp_color_space controller_color_space;
-			int opp_cnt = 1;
-			int offset = 0;
-			int dpg_width = width;
-
-			switch (test_pattern_color_space) {
-			case DP_TEST_PATTERN_COLOR_SPACE_RGB:
-				controller_color_space = CONTROLLER_DP_COLOR_SPACE_RGB;
-				break;
-			case DP_TEST_PATTERN_COLOR_SPACE_YCBCR601:
-				controller_color_space = CONTROLLER_DP_COLOR_SPACE_YCBCR601;
-				break;
-			case DP_TEST_PATTERN_COLOR_SPACE_YCBCR709:
-				controller_color_space = CONTROLLER_DP_COLOR_SPACE_YCBCR709;
-				break;
-			case DP_TEST_PATTERN_COLOR_SPACE_UNDEFINED:
-			default:
-				controller_color_space = CONTROLLER_DP_COLOR_SPACE_UDEFINED;
-				DC_LOG_ERROR("%s: Color space must be defined for test pattern", __func__);
-				ASSERT(0);
-				break;
-			}
-
-			for (odm_pipe = pipe_ctx->next_odm_pipe; odm_pipe; odm_pipe = odm_pipe->next_odm_pipe)
-				opp_cnt++;
-			dpg_width = width / opp_cnt;
-			offset = dpg_width;
-
-			link->dc->hwss.set_disp_pattern_generator(link->dc,
-					pipe_ctx,
-					controller_test_pattern,
-					controller_color_space,
-					color_depth,
-					NULL,
-					dpg_width,
-					height,
-					0);
-
-			for (odm_pipe = pipe_ctx->next_odm_pipe; odm_pipe; odm_pipe = odm_pipe->next_odm_pipe) {
-				struct output_pixel_processor *odm_opp = odm_pipe->stream_res.opp;
-
-				odm_opp->funcs->opp_program_bit_depth_reduction(odm_opp, &params);
-				link->dc->hwss.set_disp_pattern_generator(link->dc,
-						odm_pipe,
-						controller_test_pattern,
-						controller_color_space,
-						color_depth,
-						NULL,
-						dpg_width,
-						height,
-						offset);
-				offset += offset;
-			}
-		}
-	}
-	break;
-	case DP_TEST_PATTERN_VIDEO_MODE:
-	{
-		/* restore bitdepth reduction */
-		resource_build_bit_depth_reduction_params(pipe_ctx->stream, &params);
-		pipe_ctx->stream->bit_depth_params = params;
-		opp->funcs->opp_program_bit_depth_reduction(opp, &params);
-		if (pipe_ctx->stream_res.tg->funcs->set_test_pattern)
-			pipe_ctx->stream_res.tg->funcs->set_test_pattern(pipe_ctx->stream_res.tg,
-				CONTROLLER_DP_TEST_PATTERN_VIDEOMODE,
-				color_depth);
-		else if (link->dc->hwss.set_disp_pattern_generator) {
-			struct pipe_ctx *odm_pipe;
-			int opp_cnt = 1;
-			int dpg_width = width;
-
-			for (odm_pipe = pipe_ctx->next_odm_pipe; odm_pipe; odm_pipe = odm_pipe->next_odm_pipe)
-				opp_cnt++;
-
-			dpg_width = width / opp_cnt;
-			for (odm_pipe = pipe_ctx->next_odm_pipe; odm_pipe; odm_pipe = odm_pipe->next_odm_pipe) {
-				struct output_pixel_processor *odm_opp = odm_pipe->stream_res.opp;
-
-				odm_opp->funcs->opp_program_bit_depth_reduction(odm_opp, &params);
-				link->dc->hwss.set_disp_pattern_generator(link->dc,
-						odm_pipe,
-						CONTROLLER_DP_TEST_PATTERN_VIDEOMODE,
-						CONTROLLER_DP_COLOR_SPACE_UDEFINED,
-						color_depth,
-						NULL,
-						dpg_width,
-						height,
-						0);
-			}
-			link->dc->hwss.set_disp_pattern_generator(link->dc,
-					pipe_ctx,
-					CONTROLLER_DP_TEST_PATTERN_VIDEOMODE,
-					CONTROLLER_DP_COLOR_SPACE_UDEFINED,
-					color_depth,
-					NULL,
-					dpg_width,
-					height,
-					0);
-		}
-	}
-	break;
-
-	default:
-	break;
-	}
-}
-
-bool dc_link_dp_set_test_pattern(
-	struct dc_link *link,
-	enum dp_test_pattern test_pattern,
-	enum dp_test_pattern_color_space test_pattern_color_space,
-	const struct link_training_settings *p_link_settings,
-	const unsigned char *p_custom_pattern,
-	unsigned int cust_pattern_size)
-{
-	struct pipe_ctx *pipes = link->dc->current_state->res_ctx.pipe_ctx;
-	struct pipe_ctx *pipe_ctx = NULL;
-	unsigned int lane;
-	unsigned int i;
-	unsigned char link_qual_pattern[LANE_COUNT_DP_MAX] = {0};
-	union dpcd_training_pattern training_pattern;
-	enum dpcd_phy_test_patterns pattern;
-
-	memset(&training_pattern, 0, sizeof(training_pattern));
-
-	for (i = 0; i < MAX_PIPES; i++) {
-		if (pipes[i].stream == NULL)
-			continue;
-
-		if (pipes[i].stream->link == link && !pipes[i].top_pipe && !pipes[i].prev_odm_pipe) {
-			pipe_ctx = &pipes[i];
-			break;
-		}
-	}
-
-	if (pipe_ctx == NULL)
-		return false;
-
-	/* Reset CRTC Test Pattern if it is currently running and request is VideoMode */
-	if (link->test_pattern_enabled && test_pattern ==
-			DP_TEST_PATTERN_VIDEO_MODE) {
-		/* Set CRTC Test Pattern */
-		set_crtc_test_pattern(link, pipe_ctx, test_pattern, test_pattern_color_space);
-		dp_set_hw_test_pattern(link, test_pattern,
-				(uint8_t *)p_custom_pattern,
-				(uint32_t)cust_pattern_size);
-
-		/* Unblank Stream */
-		link->dc->hwss.unblank_stream(
-			pipe_ctx,
-			&link->verified_link_cap);
-		/* TODO:m_pHwss->MuteAudioEndpoint
-		 * (pPathMode->pDisplayPath, false);
-		 */
-
-		/* Reset Test Pattern state */
-		link->test_pattern_enabled = false;
-
-		return true;
-	}
-
-	/* Check for PHY Test Patterns */
-	if (is_dp_phy_pattern(test_pattern)) {
-		/* Set DPCD Lane Settings before running test pattern */
-		if (p_link_settings != NULL) {
-			dp_set_hw_lane_settings(link, p_link_settings, DPRX);
-			dpcd_set_lane_settings(link, p_link_settings, DPRX);
-		}
-
-		/* Blank stream if running test pattern */
-		if (test_pattern != DP_TEST_PATTERN_VIDEO_MODE) {
-			/*TODO:
-			 * m_pHwss->
-			 * MuteAudioEndpoint(pPathMode->pDisplayPath, true);
-			 */
-			/* Blank stream */
-			pipes->stream_res.stream_enc->funcs->dp_blank(link, pipe_ctx->stream_res.stream_enc);
-		}
-
-		dp_set_hw_test_pattern(link, test_pattern,
-				(uint8_t *)p_custom_pattern,
-				(uint32_t)cust_pattern_size);
-
-		if (test_pattern != DP_TEST_PATTERN_VIDEO_MODE) {
-			/* Set Test Pattern state */
-			link->test_pattern_enabled = true;
-			if (p_link_settings != NULL)
-				dpcd_set_link_settings(link,
-						p_link_settings);
-		}
-
-		switch (test_pattern) {
-		case DP_TEST_PATTERN_VIDEO_MODE:
-			pattern = PHY_TEST_PATTERN_NONE;
-			break;
-		case DP_TEST_PATTERN_D102:
-			pattern = PHY_TEST_PATTERN_D10_2;
-			break;
-		case DP_TEST_PATTERN_SYMBOL_ERROR:
-			pattern = PHY_TEST_PATTERN_SYMBOL_ERROR;
-			break;
-		case DP_TEST_PATTERN_PRBS7:
-			pattern = PHY_TEST_PATTERN_PRBS7;
-			break;
-		case DP_TEST_PATTERN_80BIT_CUSTOM:
-			pattern = PHY_TEST_PATTERN_80BIT_CUSTOM;
-			break;
-		case DP_TEST_PATTERN_CP2520_1:
-			pattern = PHY_TEST_PATTERN_CP2520_1;
-			break;
-		case DP_TEST_PATTERN_CP2520_2:
-			pattern = PHY_TEST_PATTERN_CP2520_2;
-			break;
-		case DP_TEST_PATTERN_CP2520_3:
-			pattern = PHY_TEST_PATTERN_CP2520_3;
-			break;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-		case DP_TEST_PATTERN_128b_132b_TPS1:
-			pattern = PHY_TEST_PATTERN_128b_132b_TPS1;
-			break;
-		case DP_TEST_PATTERN_128b_132b_TPS2:
-			pattern = PHY_TEST_PATTERN_128b_132b_TPS2;
-			break;
-		case DP_TEST_PATTERN_PRBS9:
-			pattern = PHY_TEST_PATTERN_PRBS9;
-			break;
-		case DP_TEST_PATTERN_PRBS11:
-			pattern = PHY_TEST_PATTERN_PRBS11;
-			break;
-		case DP_TEST_PATTERN_PRBS15:
-			pattern = PHY_TEST_PATTERN_PRBS15;
-			break;
-		case DP_TEST_PATTERN_PRBS23:
-			pattern = PHY_TEST_PATTERN_PRBS23;
-			break;
-		case DP_TEST_PATTERN_PRBS31:
-			pattern = PHY_TEST_PATTERN_PRBS31;
-			break;
-		case DP_TEST_PATTERN_264BIT_CUSTOM:
-			pattern = PHY_TEST_PATTERN_264BIT_CUSTOM;
-			break;
-		case DP_TEST_PATTERN_SQUARE_PULSE:
-			pattern = PHY_TEST_PATTERN_SQUARE_PULSE;
-			break;
-#endif
-		default:
-			return false;
-		}
-
-		if (test_pattern == DP_TEST_PATTERN_VIDEO_MODE
-		/*TODO:&& !pPathMode->pDisplayPath->IsTargetPoweredOn()*/)
-			return false;
-
-		if (link->dpcd_caps.dpcd_rev.raw >= DPCD_REV_12) {
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-			if (test_pattern == DP_TEST_PATTERN_SQUARE_PULSE)
-				core_link_write_dpcd(link,
-						DP_LINK_SQUARE_PATTERN,
-						p_custom_pattern,
-						1);
-
-#endif
-			/* tell receiver that we are sending qualification
-			 * pattern DP 1.2 or later - DP receiver's link quality
-			 * pattern is set using DPCD LINK_QUAL_LANEx_SET
-			 * register (0x10B~0x10E)\
-			 */
-			for (lane = 0; lane < LANE_COUNT_DP_MAX; lane++)
-				link_qual_pattern[lane] =
-						(unsigned char)(pattern);
-
-			core_link_write_dpcd(link,
-					DP_LINK_QUAL_LANE0_SET,
-					link_qual_pattern,
-					sizeof(link_qual_pattern));
-		} else if (link->dpcd_caps.dpcd_rev.raw >= DPCD_REV_10 ||
-			   link->dpcd_caps.dpcd_rev.raw == 0) {
-			/* tell receiver that we are sending qualification
-			 * pattern DP 1.1a or earlier - DP receiver's link
-			 * quality pattern is set using
-			 * DPCD TRAINING_PATTERN_SET -> LINK_QUAL_PATTERN_SET
-			 * register (0x102). We will use v_1.3 when we are
-			 * setting test pattern for DP 1.1.
-			 */
-			core_link_read_dpcd(link, DP_TRAINING_PATTERN_SET,
-					    &training_pattern.raw,
-					    sizeof(training_pattern));
-			training_pattern.v1_3.LINK_QUAL_PATTERN_SET = pattern;
-			core_link_write_dpcd(link, DP_TRAINING_PATTERN_SET,
-					     &training_pattern.raw,
-					     sizeof(training_pattern));
-		}
-	} else {
-		enum dc_color_space color_space = COLOR_SPACE_UNKNOWN;
-
-		switch (test_pattern_color_space) {
-		case DP_TEST_PATTERN_COLOR_SPACE_RGB:
-			color_space = COLOR_SPACE_SRGB;
-			if (test_pattern == DP_TEST_PATTERN_COLOR_SQUARES_CEA)
-				color_space = COLOR_SPACE_SRGB_LIMITED;
-			break;
-
-		case DP_TEST_PATTERN_COLOR_SPACE_YCBCR601:
-			color_space = COLOR_SPACE_YCBCR601;
-			if (test_pattern == DP_TEST_PATTERN_COLOR_SQUARES_CEA)
-				color_space = COLOR_SPACE_YCBCR601_LIMITED;
-			break;
-		case DP_TEST_PATTERN_COLOR_SPACE_YCBCR709:
-			color_space = COLOR_SPACE_YCBCR709;
-			if (test_pattern == DP_TEST_PATTERN_COLOR_SQUARES_CEA)
-				color_space = COLOR_SPACE_YCBCR709_LIMITED;
-			break;
-		default:
-			break;
-		}
-
-		if (pipe_ctx->stream_res.tg->funcs->lock_doublebuffer_enable) {
-			if (pipe_ctx->stream && should_use_dmub_lock(pipe_ctx->stream->link)) {
-				union dmub_hw_lock_flags hw_locks = { 0 };
-				struct dmub_hw_lock_inst_flags inst_flags = { 0 };
-
-				hw_locks.bits.lock_dig = 1;
-				inst_flags.dig_inst = pipe_ctx->stream_res.tg->inst;
-
-				dmub_hw_lock_mgr_cmd(link->ctx->dmub_srv,
-							true,
-							&hw_locks,
-							&inst_flags);
-			} else
-				pipe_ctx->stream_res.tg->funcs->lock_doublebuffer_enable(
-						pipe_ctx->stream_res.tg);
-		}
-
-		pipe_ctx->stream_res.tg->funcs->lock(pipe_ctx->stream_res.tg);
-		/* update MSA to requested color space */
-		pipe_ctx->stream_res.stream_enc->funcs->dp_set_stream_attribute(pipe_ctx->stream_res.stream_enc,
-				&pipe_ctx->stream->timing,
-				color_space,
-				pipe_ctx->stream->use_vsc_sdp_for_colorimetry,
-				link->dpcd_caps.dprx_feature.bits.SST_SPLIT_SDP_CAP);
-
-		if (pipe_ctx->stream->use_vsc_sdp_for_colorimetry) {
-			if (test_pattern == DP_TEST_PATTERN_COLOR_SQUARES_CEA)
-				pipe_ctx->stream->vsc_infopacket.sb[17] |= (1 << 7); // sb17 bit 7 Dynamic Range: 0 = VESA range, 1 = CTA range
-			else
-				pipe_ctx->stream->vsc_infopacket.sb[17] &= ~(1 << 7);
-			resource_build_info_frame(pipe_ctx);
-			link->dc->hwss.update_info_frame(pipe_ctx);
-		}
-
-		/* CRTC Patterns */
-		set_crtc_test_pattern(link, pipe_ctx, test_pattern, test_pattern_color_space);
-		pipe_ctx->stream_res.tg->funcs->unlock(pipe_ctx->stream_res.tg);
-		pipe_ctx->stream_res.tg->funcs->wait_for_state(pipe_ctx->stream_res.tg,
-				CRTC_STATE_VACTIVE);
-		pipe_ctx->stream_res.tg->funcs->wait_for_state(pipe_ctx->stream_res.tg,
-				CRTC_STATE_VBLANK);
-		pipe_ctx->stream_res.tg->funcs->wait_for_state(pipe_ctx->stream_res.tg,
-				CRTC_STATE_VACTIVE);
-
-		if (pipe_ctx->stream_res.tg->funcs->lock_doublebuffer_disable) {
-			if (pipe_ctx->stream && should_use_dmub_lock(pipe_ctx->stream->link)) {
-				union dmub_hw_lock_flags hw_locks = { 0 };
-				struct dmub_hw_lock_inst_flags inst_flags = { 0 };
-
-				hw_locks.bits.lock_dig = 1;
-				inst_flags.dig_inst = pipe_ctx->stream_res.tg->inst;
-
-				dmub_hw_lock_mgr_cmd(link->ctx->dmub_srv,
-							false,
-							&hw_locks,
-							&inst_flags);
-			} else
-				pipe_ctx->stream_res.tg->funcs->lock_doublebuffer_disable(
-						pipe_ctx->stream_res.tg);
-		}
-
-		/* Set Test Pattern state */
-		link->test_pattern_enabled = true;
-	}
-
-	return true;
-}
-
-void dp_enable_mst_on_sink(struct dc_link *link, bool enable)
-{
-	unsigned char mstmCntl;
-
-	core_link_read_dpcd(link, DP_MSTM_CTRL, &mstmCntl, 1);
-	if (enable)
-		mstmCntl |= DP_MST_EN;
-	else
-		mstmCntl &= (~DP_MST_EN);
-
-	core_link_write_dpcd(link, DP_MSTM_CTRL, &mstmCntl, 1);
-}
-
-void dp_set_panel_mode(struct dc_link *link, enum dp_panel_mode panel_mode)
-{
-	union dpcd_edp_config edp_config_set;
-	bool panel_mode_edp = false;
-
-	memset(&edp_config_set, '\0', sizeof(union dpcd_edp_config));
-
-	if (panel_mode != DP_PANEL_MODE_DEFAULT) {
-
-		switch (panel_mode) {
-		case DP_PANEL_MODE_EDP:
-		case DP_PANEL_MODE_SPECIAL:
-			panel_mode_edp = true;
-			break;
-
-		default:
-				break;
-		}
-
-		/*set edp panel mode in receiver*/
-		core_link_read_dpcd(
-			link,
-			DP_EDP_CONFIGURATION_SET,
-			&edp_config_set.raw,
-			sizeof(edp_config_set.raw));
-
-		if (edp_config_set.bits.PANEL_MODE_EDP
-			!= panel_mode_edp) {
-			enum dc_status result;
-
-			edp_config_set.bits.PANEL_MODE_EDP =
-			panel_mode_edp;
-			result = core_link_write_dpcd(
-				link,
-				DP_EDP_CONFIGURATION_SET,
-				&edp_config_set.raw,
-				sizeof(edp_config_set.raw));
-
-			ASSERT(result == DC_OK);
-		}
-	}
-	DC_LOG_DETECTION_DP_CAPS("Link: %d eDP panel mode supported: %d "
-		 "eDP panel mode enabled: %d \n",
-		 link->link_index,
-		 link->dpcd_caps.panel_mode_edp,
-		 panel_mode_edp);
-}
-
-enum dp_panel_mode dp_get_panel_mode(struct dc_link *link)
-{
-	/* We need to explicitly check that connector
-	 * is not DP. Some Travis_VGA get reported
-	 * by video bios as DP.
-	 */
-	if (link->connector_signal != SIGNAL_TYPE_DISPLAY_PORT) {
-
-		switch (link->dpcd_caps.branch_dev_id) {
-		case DP_BRANCH_DEVICE_ID_0022B9:
-			/* alternate scrambler reset is required for Travis
-			 * for the case when external chip does not
-			 * provide sink device id, alternate scrambler
-			 * scheme will  be overriden later by querying
-			 * Encoder features
-			 */
-			if (strncmp(
-				link->dpcd_caps.branch_dev_name,
-				DP_VGA_LVDS_CONVERTER_ID_2,
-				sizeof(
-				link->dpcd_caps.
-				branch_dev_name)) == 0) {
-					return DP_PANEL_MODE_SPECIAL;
-			}
-			break;
-		case DP_BRANCH_DEVICE_ID_00001A:
-			/* alternate scrambler reset is required for Travis
-			 * for the case when external chip does not provide
-			 * sink device id, alternate scrambler scheme will
-			 * be overriden later by querying Encoder feature
-			 */
-			if (strncmp(link->dpcd_caps.branch_dev_name,
-				DP_VGA_LVDS_CONVERTER_ID_3,
-				sizeof(
-				link->dpcd_caps.
-				branch_dev_name)) == 0) {
-					return DP_PANEL_MODE_SPECIAL;
-			}
-			break;
-		default:
-			break;
-		}
-	}
-
-	if (link->dpcd_caps.panel_mode_edp &&
-		(link->connector_signal == SIGNAL_TYPE_EDP ||
-		 (link->connector_signal == SIGNAL_TYPE_DISPLAY_PORT &&
-		  link->is_internal_display))) {
-		return DP_PANEL_MODE_EDP;
-	}
-
-	return DP_PANEL_MODE_DEFAULT;
-}
-
-enum dc_status dp_set_fec_ready(struct dc_link *link, bool ready)
-{
-	/* FEC has to be "set ready" before the link training.
-	 * The policy is to always train with FEC
-	 * if the sink supports it and leave it enabled on link.
-	 * If FEC is not supported, disable it.
-	 */
-	struct link_encoder *link_enc = NULL;
-	enum dc_status status = DC_OK;
-	uint8_t fec_config = 0;
-
-	/* Access link encoder based on whether it is statically
-	 * or dynamically assigned to a link.
-	 */
-	if (link->is_dig_mapping_flexible &&
-			link->dc->res_pool->funcs->link_encs_assign)
-		link_enc = link_enc_cfg_get_link_enc_used_by_link(link->ctx->dc, link);
-	else
-		link_enc = link->link_enc;
-	ASSERT(link_enc);
-
-	if (!dc_link_should_enable_fec(link))
-		return status;
-
-	if (link_enc->funcs->fec_set_ready &&
-			link->dpcd_caps.fec_cap.bits.FEC_CAPABLE) {
-		if (ready) {
-			fec_config = 1;
-			status = core_link_write_dpcd(link,
-					DP_FEC_CONFIGURATION,
-					&fec_config,
-					sizeof(fec_config));
-			if (status == DC_OK) {
-				link_enc->funcs->fec_set_ready(link_enc, true);
-				link->fec_state = dc_link_fec_ready;
-			} else {
-				link_enc->funcs->fec_set_ready(link_enc, false);
-				link->fec_state = dc_link_fec_not_ready;
-				dm_error("dpcd write failed to set fec_ready");
-			}
-		} else if (link->fec_state == dc_link_fec_ready) {
-			fec_config = 0;
-			status = core_link_write_dpcd(link,
-					DP_FEC_CONFIGURATION,
-					&fec_config,
-					sizeof(fec_config));
-			link_enc->funcs->fec_set_ready(link_enc, false);
-			link->fec_state = dc_link_fec_not_ready;
-		}
-	}
-
-	return status;
-}
-
-void dp_set_fec_enable(struct dc_link *link, bool enable)
-{
-	struct link_encoder *link_enc = NULL;
-
-	/* Access link encoder based on whether it is statically
-	 * or dynamically assigned to a link.
-	 */
-	if (link->is_dig_mapping_flexible &&
-			link->dc->res_pool->funcs->link_encs_assign)
-		link_enc = link_enc_cfg_get_link_enc_used_by_link(link->ctx->dc, link);
-	else
-		link_enc = link->link_enc;
-	ASSERT(link_enc);
-
-	if (!dc_link_should_enable_fec(link))
-		return;
-
-	if (link_enc->funcs->fec_set_enable &&
-			link->dpcd_caps.fec_cap.bits.FEC_CAPABLE) {
-		if (link->fec_state == dc_link_fec_ready && enable) {
-			/* Accord to DP spec, FEC enable sequence can first
-			 * be transmitted anytime after 1000 LL codes have
-			 * been transmitted on the link after link training
-			 * completion. Using 1 lane RBR should have the maximum
-			 * time for transmitting 1000 LL codes which is 6.173 us.
-			 * So use 7 microseconds delay instead.
-			 */
-			udelay(7);
-			link_enc->funcs->fec_set_enable(link_enc, true);
-			link->fec_state = dc_link_fec_enabled;
-		} else if (link->fec_state == dc_link_fec_enabled && !enable) {
-			link_enc->funcs->fec_set_enable(link_enc, false);
-			link->fec_state = dc_link_fec_ready;
-		}
-	}
-}
-
-void dpcd_set_source_specific_data(struct dc_link *link)
-{
-	if (!link->dc->vendor_signature.is_valid) {
-		enum dc_status __maybe_unused result_write_min_hblank = DC_NOT_SUPPORTED;
-		struct dpcd_amd_signature amd_signature = {0};
-		struct dpcd_amd_device_id amd_device_id = {0};
-
-		amd_device_id.device_id_byte1 =
-				(uint8_t)(link->ctx->asic_id.chip_id);
-		amd_device_id.device_id_byte2 =
-				(uint8_t)(link->ctx->asic_id.chip_id >> 8);
-		amd_device_id.dce_version =
-				(uint8_t)(link->ctx->dce_version);
-		amd_device_id.dal_version_byte1 = 0x0; // needed? where to get?
-		amd_device_id.dal_version_byte2 = 0x0; // needed? where to get?
-
-		core_link_read_dpcd(link, DP_SOURCE_OUI,
-				(uint8_t *)(&amd_signature),
-				sizeof(amd_signature));
-
-		if (!((amd_signature.AMD_IEEE_TxSignature_byte1 == 0x0) &&
-			(amd_signature.AMD_IEEE_TxSignature_byte2 == 0x0) &&
-			(amd_signature.AMD_IEEE_TxSignature_byte3 == 0x1A))) {
-
-			amd_signature.AMD_IEEE_TxSignature_byte1 = 0x0;
-			amd_signature.AMD_IEEE_TxSignature_byte2 = 0x0;
-			amd_signature.AMD_IEEE_TxSignature_byte3 = 0x1A;
-
-			core_link_write_dpcd(link, DP_SOURCE_OUI,
-				(uint8_t *)(&amd_signature),
-				sizeof(amd_signature));
-		}
-
-		core_link_write_dpcd(link, DP_SOURCE_OUI+0x03,
-				(uint8_t *)(&amd_device_id),
-				sizeof(amd_device_id));
-
-		if (link->ctx->dce_version >= DCN_VERSION_2_0 &&
-			link->dc->caps.min_horizontal_blanking_period != 0) {
-
-			uint8_t hblank_size = (uint8_t)link->dc->caps.min_horizontal_blanking_period;
-
-			if (link->preferred_link_setting.dpcd_source_device_specific_field_support) {
-				result_write_min_hblank = core_link_write_dpcd(link,
-					DP_SOURCE_MINIMUM_HBLANK_SUPPORTED, (uint8_t *)(&hblank_size),
-					sizeof(hblank_size));
-
-				if (result_write_min_hblank == DC_ERROR_UNEXPECTED)
-					link->preferred_link_setting.dpcd_source_device_specific_field_support = false;
-			} else {
-				DC_LOG_DC("Sink device does not support 00340h DPCD write. Skipping on purpose.\n");
-			}
-		}
-
-		DC_TRACE_LEVEL_MESSAGE(DAL_TRACE_LEVEL_INFORMATION,
-							WPP_BIT_FLAG_DC_DETECTION_DP_CAPS,
-							"result=%u link_index=%u enum dce_version=%d DPCD=0x%04X min_hblank=%u branch_dev_id=0x%x branch_dev_name='%c%c%c%c%c%c'",
-							result_write_min_hblank,
-							link->link_index,
-							link->ctx->dce_version,
-							DP_SOURCE_MINIMUM_HBLANK_SUPPORTED,
-							link->dc->caps.min_horizontal_blanking_period,
-							link->dpcd_caps.branch_dev_id,
-							link->dpcd_caps.branch_dev_name[0],
-							link->dpcd_caps.branch_dev_name[1],
-							link->dpcd_caps.branch_dev_name[2],
-							link->dpcd_caps.branch_dev_name[3],
-							link->dpcd_caps.branch_dev_name[4],
-							link->dpcd_caps.branch_dev_name[5]);
-	} else {
-		core_link_write_dpcd(link, DP_SOURCE_OUI,
-				link->dc->vendor_signature.data.raw,
-				sizeof(link->dc->vendor_signature.data.raw));
-	}
-}
-
-bool dc_link_set_backlight_level_nits(struct dc_link *link,
-		bool isHDR,
-		uint32_t backlight_millinits,
-		uint32_t transition_time_in_ms)
-{
-	struct dpcd_source_backlight_set dpcd_backlight_set;
-	uint8_t backlight_control = isHDR ? 1 : 0;
-
-	if (!link || (link->connector_signal != SIGNAL_TYPE_EDP &&
-			link->connector_signal != SIGNAL_TYPE_DISPLAY_PORT))
-		return false;
-
-	// OLEDs have no PWM, they can only use AUX
-	if (link->dpcd_sink_ext_caps.bits.oled == 1)
-		backlight_control = 1;
-
-	*(uint32_t *)&dpcd_backlight_set.backlight_level_millinits = backlight_millinits;
-	*(uint16_t *)&dpcd_backlight_set.backlight_transition_time_ms = (uint16_t)transition_time_in_ms;
-
-
-	if (core_link_write_dpcd(link, DP_SOURCE_BACKLIGHT_LEVEL,
-			(uint8_t *)(&dpcd_backlight_set),
-			sizeof(dpcd_backlight_set)) != DC_OK)
-		return false;
-
-	if (core_link_write_dpcd(link, DP_SOURCE_BACKLIGHT_CONTROL,
-			&backlight_control, 1) != DC_OK)
-		return false;
-
-	return true;
-}
-
-bool dc_link_get_backlight_level_nits(struct dc_link *link,
-		uint32_t *backlight_millinits_avg,
-		uint32_t *backlight_millinits_peak)
-{
-	union dpcd_source_backlight_get dpcd_backlight_get;
-
-	memset(&dpcd_backlight_get, 0, sizeof(union dpcd_source_backlight_get));
-
-	if (!link || (link->connector_signal != SIGNAL_TYPE_EDP &&
-			link->connector_signal != SIGNAL_TYPE_DISPLAY_PORT))
-		return false;
-
-	if (core_link_read_dpcd(link, DP_SOURCE_BACKLIGHT_CURRENT_PEAK,
-			dpcd_backlight_get.raw,
-			sizeof(union dpcd_source_backlight_get)) != DC_OK)
-		return false;
-
-	*backlight_millinits_avg =
-		dpcd_backlight_get.bytes.backlight_millinits_avg;
-	*backlight_millinits_peak =
-		dpcd_backlight_get.bytes.backlight_millinits_peak;
-
-	/* On non-supported panels dpcd_read usually succeeds with 0 returned */
-	if (*backlight_millinits_avg == 0 ||
-			*backlight_millinits_avg > *backlight_millinits_peak)
-		return false;
-
-	return true;
-}
-
-bool dc_link_backlight_enable_aux(struct dc_link *link, bool enable)
-{
-	uint8_t backlight_enable = enable ? 1 : 0;
-
-	if (!link || (link->connector_signal != SIGNAL_TYPE_EDP &&
-		link->connector_signal != SIGNAL_TYPE_DISPLAY_PORT))
-		return false;
-
-	if (core_link_write_dpcd(link, DP_SOURCE_BACKLIGHT_ENABLE,
-		&backlight_enable, 1) != DC_OK)
-		return false;
-
-	return true;
-}
-
-// we read default from 0x320 because we expect BIOS wrote it there
-// regular get_backlight_nit reads from panel set at 0x326
-bool dc_link_read_default_bl_aux(struct dc_link *link, uint32_t *backlight_millinits)
-{
-	if (!link || (link->connector_signal != SIGNAL_TYPE_EDP &&
-		link->connector_signal != SIGNAL_TYPE_DISPLAY_PORT))
-		return false;
-
-	if (core_link_read_dpcd(link, DP_SOURCE_BACKLIGHT_LEVEL,
-		(uint8_t *) backlight_millinits,
-		sizeof(uint32_t)) != DC_OK)
-		return false;
-
-	return true;
-}
-
-bool dc_link_set_default_brightness_aux(struct dc_link *link)
-{
-	uint32_t default_backlight;
-
-	if (link && link->dpcd_sink_ext_caps.bits.oled == 1) {
-		if (!dc_link_read_default_bl_aux(link, &default_backlight))
-			default_backlight = 150000;
-		// if < 5 nits or > 5000, it might be wrong readback
-		if (default_backlight < 5000 || default_backlight > 5000000)
-			default_backlight = 150000; //
-
-		return dc_link_set_backlight_level_nits(link, true,
-				default_backlight, 0);
-	}
-	return false;
-}
-
-bool is_edp_ilr_optimization_required(struct dc_link *link, struct dc_crtc_timing *crtc_timing)
-{
-	struct dc_link_settings link_setting;
-	uint8_t link_bw_set;
-	uint8_t link_rate_set;
-	uint32_t req_bw;
-	union lane_count_set lane_count_set = {0};
-
-	ASSERT(link || crtc_timing); // invalid input
-
-	if (link->dpcd_caps.edp_supported_link_rates_count == 0 ||
-			!link->dc->debug.optimize_edp_link_rate)
-		return false;
-
-
-	// Read DPCD 00100h to find if standard link rates are set
-	core_link_read_dpcd(link, DP_LINK_BW_SET,
-				&link_bw_set, sizeof(link_bw_set));
-
-	if (link_bw_set) {
-		DC_LOG_EVENT_LINK_TRAINING("eDP ILR: Optimization required, VBIOS used link_bw_set\n");
-		return true;
-	}
-
-	// Read DPCD 00115h to find the edp link rate set used
-	core_link_read_dpcd(link, DP_LINK_RATE_SET,
-			    &link_rate_set, sizeof(link_rate_set));
-
-	// Read DPCD 00101h to find out the number of lanes currently set
-	core_link_read_dpcd(link, DP_LANE_COUNT_SET,
-				&lane_count_set.raw, sizeof(lane_count_set));
-
-	req_bw = dc_bandwidth_in_kbps_from_timing(crtc_timing);
-
-	decide_edp_link_settings(link, &link_setting, req_bw);
-
-	if (link->dpcd_caps.edp_supported_link_rates[link_rate_set] != link_setting.link_rate ||
-			lane_count_set.bits.LANE_COUNT_SET != link_setting.lane_count) {
-		DC_LOG_EVENT_LINK_TRAINING("eDP ILR: Optimization required, VBIOS link_rate_set not optimal\n");
-		return true;
-	}
-
-	DC_LOG_EVENT_LINK_TRAINING("eDP ILR: No optimization required, VBIOS set optimal link_rate_set\n");
-	return false;
-}
-
-enum dp_link_encoding dp_get_link_encoding_format(const struct dc_link_settings *link_settings)
-{
-	if ((link_settings->link_rate >= LINK_RATE_LOW) &&
-			(link_settings->link_rate <= LINK_RATE_HIGH3))
-		return DP_8b_10b_ENCODING;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	else if ((link_settings->link_rate >= LINK_RATE_UHBR10) &&
-			(link_settings->link_rate <= LINK_RATE_UHBR20))
-		return DP_128b_132b_ENCODING;
-#endif
-	return DP_UNKNOWN_ENCODING;
-}
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-enum dp_link_encoding dc_link_dp_mst_decide_link_encoding_format(const struct dc_link *link)
-{
-	struct dc_link_settings link_settings = {0};
-
-	if (!dc_is_dp_signal(link->connector_signal))
-		return DP_UNKNOWN_ENCODING;
-
-	if (link->preferred_link_setting.lane_count !=
-			LANE_COUNT_UNKNOWN &&
-			link->preferred_link_setting.link_rate !=
-					LINK_RATE_UNKNOWN) {
-		link_settings = link->preferred_link_setting;
-	} else {
-		decide_mst_link_settings(link, &link_settings);
-	}
-
-	return dp_get_link_encoding_format(&link_settings);
-}
-
-// TODO - DP2.0 Link: Fix get_lane_status to handle LTTPR offset (SST and MST)
-static void get_lane_status(
-	struct dc_link *link,
-	uint32_t lane_count,
-	union lane_status *status,
-	union lane_align_status_updated *status_updated)
-{
-	unsigned int lane;
-	uint8_t dpcd_buf[3] = {0};
-
-	if (status == NULL || status_updated == NULL) {
-		return;
-	}
-
-	core_link_read_dpcd(
-			link,
-			DP_LANE0_1_STATUS,
-			dpcd_buf,
-			sizeof(dpcd_buf));
-
-	for (lane = 0; lane < lane_count; lane++) {
-		status[lane].raw = get_nibble_at_index(&dpcd_buf[0], lane);
-	}
-
-	status_updated->raw = dpcd_buf[2];
-}
-
-bool dpcd_write_128b_132b_sst_payload_allocation_table(
-		const struct dc_stream_state *stream,
-		struct dc_link *link,
-		struct link_mst_stream_allocation_table *proposed_table,
-		bool allocate)
-{
-	const uint8_t vc_id = 1; /// VC ID always 1 for SST
-	const uint8_t start_time_slot = 0; /// Always start at time slot 0 for SST
-	bool result = false;
-	uint8_t req_slot_count = 0;
-	struct fixed31_32 avg_time_slots_per_mtp = { 0 };
-	union payload_table_update_status update_status = { 0 };
-	const uint32_t max_retries = 30;
-	uint32_t retries = 0;
-
-	if (allocate)	{
-		avg_time_slots_per_mtp = calculate_sst_avg_time_slots_per_mtp(stream, link);
-		req_slot_count = dc_fixpt_ceil(avg_time_slots_per_mtp);
-	} else {
-		/// Leave req_slot_count = 0 if allocate is false.
-	}
-
-	/// Write DPCD 2C0 = 1 to start updating
-	update_status.bits.VC_PAYLOAD_TABLE_UPDATED = 1;
-	core_link_write_dpcd(
-			link,
-			DP_PAYLOAD_TABLE_UPDATE_STATUS,
-			&update_status.raw,
-			1);
-
-	/// Program the changes in DPCD 1C0 - 1C2
-	ASSERT(vc_id == 1);
-	core_link_write_dpcd(
-			link,
-			DP_PAYLOAD_ALLOCATE_SET,
-			&vc_id,
-			1);
-
-	ASSERT(start_time_slot == 0);
-	core_link_write_dpcd(
-			link,
-			DP_PAYLOAD_ALLOCATE_START_TIME_SLOT,
-			&start_time_slot,
-			1);
-
-	ASSERT(req_slot_count <= MAX_MTP_SLOT_COUNT); /// Validation should filter out modes that exceed link BW
-	core_link_write_dpcd(
-			link,
-			DP_PAYLOAD_ALLOCATE_TIME_SLOT_COUNT,
-			&req_slot_count,
-			1);
-
-	/// Poll till DPCD 2C0 read 1
-	/// Try for at least 150ms (30 retries, with 5ms delay after each attempt)
-
-	while (retries < max_retries) {
-		if (core_link_read_dpcd(
-				link,
-				DP_PAYLOAD_TABLE_UPDATE_STATUS,
-				&update_status.raw,
-				1) == DC_OK) {
-			if (update_status.bits.VC_PAYLOAD_TABLE_UPDATED == 1) {
-				DC_LOG_DP2("SST Update Payload: downstream payload table updated.");
-				result = true;
-				break;
-			}
-		} else {
-			union dpcd_rev dpcdRev;
-
-			if (core_link_read_dpcd(
-					link,
-					DP_DPCD_REV,
-					&dpcdRev.raw,
-					1) != DC_OK) {
-				DC_LOG_ERROR("SST Update Payload: Unable to read DPCD revision "
-						"of sink while polling payload table "
-						"updated status bit.");
-				break;
-			}
-		}
-		retries++;
-		udelay(5000);
-	}
-
-	if (!result && retries == max_retries) {
-		DC_LOG_ERROR("SST Update Payload: Payload table not updated after retries, "
-				"continue on. Something is wrong with the branch.");
-		// TODO - DP2.0 Payload: Read and log the payload table from downstream branch
-	}
-
-	proposed_table->stream_count = 1; /// Always 1 stream for SST
-	proposed_table->stream_allocations[0].slot_count = req_slot_count;
-	proposed_table->stream_allocations[0].vcp_id = vc_id;
-
-	return result;
-}
-
-bool dpcd_poll_for_allocation_change_trigger(struct dc_link *link)
-{
-	/*
-	 * wait for ACT handled
-	 */
-	int i;
-	const int act_retries = 30;
-	enum act_return_status result = ACT_FAILED;
-	union payload_table_update_status update_status = {0};
-	union lane_status dpcd_lane_status[LANE_COUNT_DP_MAX];
-	union lane_align_status_updated lane_status_updated;
-
-	for (i = 0; i < act_retries; i++) {
-		get_lane_status(link, link->cur_link_settings.lane_count, dpcd_lane_status, &lane_status_updated);
-
-		if (!dp_is_cr_done(link->cur_link_settings.lane_count, dpcd_lane_status) ||
-				!dp_is_ch_eq_done(link->cur_link_settings.lane_count, dpcd_lane_status) ||
-				!dp_is_symbol_locked(link->cur_link_settings.lane_count, dpcd_lane_status) ||
-				!dp_is_interlane_aligned(lane_status_updated)) {
-			DC_LOG_ERROR("SST Update Payload: Link loss occurred while "
-					"polling for ACT handled.");
-			result = ACT_LINK_LOST;
-			break;
-		}
-		core_link_read_dpcd(
-				link,
-				DP_PAYLOAD_TABLE_UPDATE_STATUS,
-				&update_status.raw,
-				1);
-
-		if (update_status.bits.ACT_HANDLED == 1) {
-			DC_LOG_DP2("SST Update Payload: ACT handled by downstream.");
-			result = ACT_SUCCESS;
-			break;
-		}
-
-		udelay(5000);
-	}
-
-	if (result == ACT_FAILED) {
-		DC_LOG_ERROR("SST Update Payload: ACT still not handled after retries, "
-				"continue on. Something is wrong with the branch.");
-	}
-
-	return (result == ACT_SUCCESS);
-}
-
-struct fixed31_32 calculate_sst_avg_time_slots_per_mtp(
-		const struct dc_stream_state *stream,
-		const struct dc_link *link)
-{
-	struct fixed31_32 link_bw_effective =
-			dc_fixpt_from_int(
-					dc_link_bandwidth_kbps(link, &link->cur_link_settings));
-	struct fixed31_32 timeslot_bw_effective =
-			dc_fixpt_div_int(link_bw_effective, MAX_MTP_SLOT_COUNT);
-	struct fixed31_32 timing_bw =
-			dc_fixpt_from_int(
-					dc_bandwidth_in_kbps_from_timing(&stream->timing));
-	struct fixed31_32 avg_time_slots_per_mtp =
-			dc_fixpt_div(timing_bw, timeslot_bw_effective);
-
-	return avg_time_slots_per_mtp;
-}
-
-bool is_dp_128b_132b_signal(struct pipe_ctx *pipe_ctx)
-{
-	return (pipe_ctx->stream_res.hpo_dp_stream_enc &&
-			pipe_ctx->stream->link->hpo_dp_link_enc &&
-			dc_is_dp_signal(pipe_ctx->stream->signal));
-}
-#endif
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_link_dpcd.c b/drivers/gpu/drm/amd/display/dc/core/dc_link_dpcd.c
deleted file mode 100644
index 7f25c11f4248..000000000000
--- a/drivers/gpu/drm/amd/display/dc/core/dc_link_dpcd.c
+++ /dev/null
@@ -1,246 +0,0 @@
-/*
- * Copyright 2021 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: AMD
- *
- */
-
-#include <inc/core_status.h>
-#include <dc_link.h>
-#include <inc/link_hwss.h>
-#include <inc/link_dpcd.h>
-#include "drm/drm_dp_helper.h"
-#include <dc_dp_types.h>
-#include "dm_helpers.h"
-
-#define END_ADDRESS(start, size) (start + size - 1)
-#define ADDRESS_RANGE_SIZE(start, end) (end - start + 1)
-struct dpcd_address_range {
-	uint32_t start;
-	uint32_t end;
-};
-
-static enum dc_status internal_link_read_dpcd(
-	struct dc_link *link,
-	uint32_t address,
-	uint8_t *data,
-	uint32_t size)
-{
-	if (!link->aux_access_disabled &&
-			!dm_helpers_dp_read_dpcd(link->ctx,
-			link, address, data, size)) {
-		return DC_ERROR_UNEXPECTED;
-	}
-
-	return DC_OK;
-}
-
-static enum dc_status internal_link_write_dpcd(
-	struct dc_link *link,
-	uint32_t address,
-	const uint8_t *data,
-	uint32_t size)
-{
-	if (!link->aux_access_disabled &&
-			!dm_helpers_dp_write_dpcd(link->ctx,
-			link, address, data, size)) {
-		return DC_ERROR_UNEXPECTED;
-	}
-
-	return DC_OK;
-}
-
-/*
- * Partition the entire DPCD address space
- * XXX: This partitioning must cover the entire DPCD address space,
- * and must contain no gaps or overlapping address ranges.
- */
-static const struct dpcd_address_range mandatory_dpcd_partitions[] = {
-	{ 0, DP_TRAINING_PATTERN_SET_PHY_REPEATER(DP_PHY_LTTPR1) - 1},
-	{ DP_TRAINING_PATTERN_SET_PHY_REPEATER(DP_PHY_LTTPR1), DP_TRAINING_PATTERN_SET_PHY_REPEATER(DP_PHY_LTTPR2) - 1 },
-	{ DP_TRAINING_PATTERN_SET_PHY_REPEATER(DP_PHY_LTTPR2), DP_TRAINING_PATTERN_SET_PHY_REPEATER(DP_PHY_LTTPR3) - 1 },
-	{ DP_TRAINING_PATTERN_SET_PHY_REPEATER(DP_PHY_LTTPR3), DP_TRAINING_PATTERN_SET_PHY_REPEATER(DP_PHY_LTTPR4) - 1 },
-	{ DP_TRAINING_PATTERN_SET_PHY_REPEATER(DP_PHY_LTTPR4), DP_TRAINING_PATTERN_SET_PHY_REPEATER(DP_PHY_LTTPR5) - 1 },
-	{ DP_TRAINING_PATTERN_SET_PHY_REPEATER(DP_PHY_LTTPR5), DP_TRAINING_PATTERN_SET_PHY_REPEATER(DP_PHY_LTTPR6) - 1 },
-	{ DP_TRAINING_PATTERN_SET_PHY_REPEATER(DP_PHY_LTTPR6), DP_TRAINING_PATTERN_SET_PHY_REPEATER(DP_PHY_LTTPR7) - 1 },
-	{ DP_TRAINING_PATTERN_SET_PHY_REPEATER(DP_PHY_LTTPR7), DP_TRAINING_PATTERN_SET_PHY_REPEATER(DP_PHY_LTTPR8) - 1 },
-	{ DP_TRAINING_PATTERN_SET_PHY_REPEATER(DP_PHY_LTTPR8), DP_FEC_STATUS_PHY_REPEATER(DP_PHY_LTTPR1) - 1 },
-	/*
-	 * The FEC registers are contiguous
-	 */
-	{ DP_FEC_STATUS_PHY_REPEATER(DP_PHY_LTTPR1), DP_FEC_STATUS_PHY_REPEATER(DP_PHY_LTTPR1) - 1 },
-	{ DP_FEC_STATUS_PHY_REPEATER(DP_PHY_LTTPR2), DP_FEC_STATUS_PHY_REPEATER(DP_PHY_LTTPR2) - 1 },
-	{ DP_FEC_STATUS_PHY_REPEATER(DP_PHY_LTTPR3), DP_FEC_STATUS_PHY_REPEATER(DP_PHY_LTTPR3) - 1 },
-	{ DP_FEC_STATUS_PHY_REPEATER(DP_PHY_LTTPR4), DP_FEC_STATUS_PHY_REPEATER(DP_PHY_LTTPR4) - 1 },
-	{ DP_FEC_STATUS_PHY_REPEATER(DP_PHY_LTTPR5), DP_FEC_STATUS_PHY_REPEATER(DP_PHY_LTTPR5) - 1 },
-	{ DP_FEC_STATUS_PHY_REPEATER(DP_PHY_LTTPR6), DP_FEC_STATUS_PHY_REPEATER(DP_PHY_LTTPR6) - 1 },
-	{ DP_FEC_STATUS_PHY_REPEATER(DP_PHY_LTTPR7), DP_FEC_STATUS_PHY_REPEATER(DP_PHY_LTTPR7) - 1 },
-	{ DP_FEC_STATUS_PHY_REPEATER(DP_PHY_LTTPR8), DP_LTTPR_MAX_ADD },
-	/* all remaining DPCD addresses */
-	{ DP_LTTPR_MAX_ADD + 1, DP_DPCD_MAX_ADD } };
-
-static inline bool do_addresses_intersect_with_range(
-		const struct dpcd_address_range *range,
-		const uint32_t start_address,
-		const uint32_t end_address)
-{
-	return start_address <= range->end && end_address >= range->start;
-}
-
-static uint32_t dpcd_get_next_partition_size(const uint32_t address, const uint32_t size)
-{
-	const uint32_t end_address = END_ADDRESS(address, size);
-	uint32_t partition_iterator = 0;
-
-	/*
-	 * find current partition
-	 * this loop spins forever if partition map above is not surjective
-	 */
-	while (!do_addresses_intersect_with_range(&mandatory_dpcd_partitions[partition_iterator],
-				address, end_address))
-		partition_iterator++;
-	if (end_address < mandatory_dpcd_partitions[partition_iterator].end)
-		return size;
-	return ADDRESS_RANGE_SIZE(address, mandatory_dpcd_partitions[partition_iterator].end);
-}
-
-/*
- * Ranges of DPCD addresses that must be read in a single transaction
- * XXX: Do not allow any two address ranges in this array to overlap
- */
-static const struct dpcd_address_range mandatory_dpcd_blocks[] = {
-	{ DP_LT_TUNABLE_PHY_REPEATER_FIELD_DATA_STRUCTURE_REV, DP_PHY_REPEATER_EXTENDED_WAIT_TIMEOUT }};
-
-/*
- * extend addresses to read all mandatory blocks together
- */
-static void dpcd_extend_address_range(
-		const uint32_t in_address,
-		uint8_t * const in_data,
-		const uint32_t in_size,
-		uint32_t *out_address,
-		uint8_t **out_data,
-		uint32_t *out_size)
-{
-	const uint32_t end_address = END_ADDRESS(in_address, in_size);
-	const struct dpcd_address_range *addr_range;
-	struct dpcd_address_range new_addr_range;
-	uint32_t i;
-
-	new_addr_range.start = in_address;
-	new_addr_range.end = end_address;
-	for (i = 0; i < ARRAY_SIZE(mandatory_dpcd_blocks); i++) {
-		addr_range = &mandatory_dpcd_blocks[i];
-		if (addr_range->start <= in_address && addr_range->end >= in_address)
-			new_addr_range.start = addr_range->start;
-
-		if (addr_range->start <= end_address && addr_range->end >= end_address)
-			new_addr_range.end = addr_range->end;
-	}
-	*out_address = in_address;
-	*out_size = in_size;
-	*out_data = in_data;
-	if (new_addr_range.start != in_address || new_addr_range.end != end_address) {
-		*out_address = new_addr_range.start;
-		*out_size = ADDRESS_RANGE_SIZE(new_addr_range.start, new_addr_range.end);
-		*out_data = kzalloc(*out_size * sizeof(**out_data), GFP_KERNEL);
-	}
-}
-
-/*
- * Reduce the AUX reply down to the values the caller requested
- */
-static void dpcd_reduce_address_range(
-		const uint32_t extended_address,
-		uint8_t * const extended_data,
-		const uint32_t extended_size,
-		const uint32_t reduced_address,
-		uint8_t * const reduced_data,
-		const uint32_t reduced_size)
-{
-	const uint32_t offset = reduced_address - extended_address;
-
-	/*
-	 * If the address is same, address was not extended.
-	 * So we do not need to free any memory.
-	 * The data is in original buffer(reduced_data).
-	 */
-	if (extended_data == reduced_data)
-		return;
-
-	memcpy(&extended_data[offset], reduced_data, reduced_size);
-	kfree(extended_data);
-}
-
-enum dc_status core_link_read_dpcd(
-	struct dc_link *link,
-	uint32_t address,
-	uint8_t *data,
-	uint32_t size)
-{
-	uint32_t extended_address;
-	uint32_t partitioned_address;
-	uint8_t *extended_data;
-	uint32_t extended_size;
-	/* size of the remaining partitioned address space */
-	uint32_t size_left_to_read;
-	enum dc_status status;
-	/* size of the next partition to be read from */
-	uint32_t partition_size;
-	uint32_t data_index = 0;
-
-	dpcd_extend_address_range(address, data, size, &extended_address, &extended_data, &extended_size);
-	partitioned_address = extended_address;
-	size_left_to_read = extended_size;
-	while (size_left_to_read) {
-		partition_size = dpcd_get_next_partition_size(partitioned_address, size_left_to_read);
-		status = internal_link_read_dpcd(link, partitioned_address, &extended_data[data_index], partition_size);
-		if (status != DC_OK)
-			break;
-		partitioned_address += partition_size;
-		data_index += partition_size;
-		size_left_to_read -= partition_size;
-	}
-	dpcd_reduce_address_range(extended_address, extended_data, extended_size, address, data, size);
-	return status;
-}
-
-enum dc_status core_link_write_dpcd(
-	struct dc_link *link,
-	uint32_t address,
-	const uint8_t *data,
-	uint32_t size)
-{
-	uint32_t partition_size;
-	uint32_t data_index = 0;
-	enum dc_status status;
-
-	while (size) {
-		partition_size = dpcd_get_next_partition_size(address, size);
-		status = internal_link_write_dpcd(link, address, &data[data_index], partition_size);
-		if (status != DC_OK)
-			break;
-		address += partition_size;
-		data_index += partition_size;
-		size -= partition_size;
-	}
-	return status;
-}
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_link_dpia.c b/drivers/gpu/drm/amd/display/dc/core/dc_link_dpia.c
deleted file mode 100644
index b1c9f77d6bf4..000000000000
--- a/drivers/gpu/drm/amd/display/dc/core/dc_link_dpia.c
+++ /dev/null
@@ -1,962 +0,0 @@
-// SPDX-License-Identifier: MIT
-/*
- * Copyright 2021 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: AMD
- *
- */
-
-#include "dc.h"
-#include "dc_link_dpia.h"
-#include "inc/core_status.h"
-#include "dc_link.h"
-#include "dc_link_dp.h"
-#include "dpcd_defs.h"
-#include "link_hwss.h"
-#include "dm_helpers.h"
-#include "dmub/inc/dmub_cmd.h"
-#include "inc/link_dpcd.h"
-
-#define DC_LOGGER \
-	link->ctx->logger
-
-enum dc_status dpcd_get_tunneling_device_data(struct dc_link *link)
-{
-	enum dc_status status = DC_OK;
-	uint8_t dpcd_dp_tun_data[3] = {0};
-	uint8_t dpcd_topology_data[DPCD_USB4_TOPOLOGY_ID_LEN] = {0};
-	uint8_t i = 0;
-
-	status = core_link_read_dpcd(link,
-			DP_TUNNELING_CAPABILITIES_SUPPORT,
-			dpcd_dp_tun_data,
-			sizeof(dpcd_dp_tun_data));
-
-	status = core_link_read_dpcd(link,
-			DP_USB4_ROUTER_TOPOLOGY_ID,
-			dpcd_topology_data,
-			sizeof(dpcd_topology_data));
-
-	link->dpcd_caps.usb4_dp_tun_info.dp_tun_cap.raw =
-			dpcd_dp_tun_data[DP_TUNNELING_CAPABILITIES_SUPPORT -
-					 DP_TUNNELING_CAPABILITIES_SUPPORT];
-	link->dpcd_caps.usb4_dp_tun_info.dpia_info.raw =
-			dpcd_dp_tun_data[DP_IN_ADAPTER_INFO - DP_TUNNELING_CAPABILITIES_SUPPORT];
-	link->dpcd_caps.usb4_dp_tun_info.usb4_driver_id =
-			dpcd_dp_tun_data[DP_USB4_DRIVER_ID - DP_TUNNELING_CAPABILITIES_SUPPORT];
-
-	for (i = 0; i < DPCD_USB4_TOPOLOGY_ID_LEN; i++)
-		link->dpcd_caps.usb4_dp_tun_info.usb4_topology_id[i] = dpcd_topology_data[i];
-
-	return status;
-}
-
-/* Configure link as prescribed in link_setting; set LTTPR mode; and
- * Initialize link training settings.
- * Abort link training if sink unplug detected.
- *
- * @param link DPIA link being trained.
- * @param[in] link_setting Lane count, link rate and downspread control.
- * @param[out] lt_settings Link settings and drive settings (voltage swing and pre-emphasis).
- */
-static enum link_training_result dpia_configure_link(struct dc_link *link,
-		const struct dc_link_settings *link_setting,
-		struct link_training_settings *lt_settings)
-{
-	enum dc_status status;
-	bool fec_enable;
-
-	DC_LOG_HW_LINK_TRAINING("%s\n DPIA(%d) configuring\n - LTTPR mode(%d)\n",
-				__func__,
-				link->link_id.enum_id - ENUM_ID_1,
-				link->lttpr_mode);
-
-	dp_decide_training_settings(link,
-		link_setting,
-		lt_settings);
-
-	status = dpcd_configure_channel_coding(link, lt_settings);
-	if (status != DC_OK && !link->hpd_status)
-		return LINK_TRAINING_ABORT;
-
-	/* Configure lttpr mode */
-	status = dpcd_configure_lttpr_mode(link, lt_settings);
-	if (status != DC_OK && !link->hpd_status)
-		return LINK_TRAINING_ABORT;
-
-	/* Set link rate, lane count and spread. */
-	status = dpcd_set_link_settings(link, lt_settings);
-	if (status != DC_OK && !link->hpd_status)
-		return LINK_TRAINING_ABORT;
-
-	if (link->preferred_training_settings.fec_enable)
-		fec_enable = *link->preferred_training_settings.fec_enable;
-	else
-		fec_enable = true;
-	status = dp_set_fec_ready(link, fec_enable);
-	if (status != DC_OK && !link->hpd_status)
-		return LINK_TRAINING_ABORT;
-
-	return LINK_TRAINING_SUCCESS;
-}
-
-static enum dc_status core_link_send_set_config(struct dc_link *link,
-	uint8_t msg_type,
-	uint8_t msg_data)
-{
-	struct set_config_cmd_payload payload;
-	enum set_config_status set_config_result = SET_CONFIG_PENDING;
-
-	/* prepare set_config payload */
-	payload.msg_type = msg_type;
-	payload.msg_data = msg_data;
-
-	if (!link->ddc->ddc_pin && !link->aux_access_disabled &&
-	    (dm_helpers_dmub_set_config_sync(link->ctx, link,
-					     &payload, &set_config_result) == -1)) {
-		return DC_ERROR_UNEXPECTED;
-	}
-
-	/* set_config should return ACK if successful */
-	return (set_config_result == SET_CONFIG_ACK_RECEIVED) ? DC_OK : DC_ERROR_UNEXPECTED;
-}
-
-/* Build SET_CONFIG message data payload for specified message type. */
-static uint8_t dpia_build_set_config_data(enum dpia_set_config_type type,
-		struct dc_link *link,
-		struct link_training_settings *lt_settings)
-{
-	union dpia_set_config_data data;
-
-	data.raw = 0;
-
-	switch (type) {
-	case DPIA_SET_CFG_SET_LINK:
-		data.set_link.mode = link->lttpr_mode == LTTPR_MODE_NON_TRANSPARENT ? 1 : 0;
-		break;
-	case DPIA_SET_CFG_SET_PHY_TEST_MODE:
-		break;
-	case DPIA_SET_CFG_SET_VSPE:
-		/* Assume all lanes have same drive settings. */
-		data.set_vspe.swing = lt_settings->lane_settings[0].VOLTAGE_SWING;
-		data.set_vspe.pre_emph = lt_settings->lane_settings[0].PRE_EMPHASIS;
-		data.set_vspe.max_swing_reached =
-			lt_settings->lane_settings[0].VOLTAGE_SWING ==
-			VOLTAGE_SWING_MAX_LEVEL ? 1 : 0;
-		data.set_vspe.max_pre_emph_reached =
-			lt_settings->lane_settings[0].PRE_EMPHASIS ==
-			PRE_EMPHASIS_MAX_LEVEL ? 1 : 0;
-		break;
-	default:
-		ASSERT(false); /* Message type not supported by helper function. */
-		break;
-	}
-
-	return data.raw;
-}
-
-/* Convert DC training pattern to DPIA training stage. */
-static enum dpia_set_config_ts convert_trng_ptn_to_trng_stg(enum dc_dp_training_pattern tps)
-{
-	enum dpia_set_config_ts ts;
-
-	switch (tps) {
-	case DP_TRAINING_PATTERN_SEQUENCE_1:
-		ts = DPIA_TS_TPS1;
-		break;
-	case DP_TRAINING_PATTERN_SEQUENCE_2:
-		ts = DPIA_TS_TPS2;
-		break;
-	case DP_TRAINING_PATTERN_SEQUENCE_3:
-		ts = DPIA_TS_TPS3;
-		break;
-	case DP_TRAINING_PATTERN_SEQUENCE_4:
-		ts = DPIA_TS_TPS4;
-		break;
-	default:
-		ts = DPIA_TS_DPRX_DONE;
-		ASSERT(false); /* TPS not supported by helper function. */
-		break;
-	}
-
-	return ts;
-}
-
-/* Write training pattern to DPCD. */
-static enum dc_status dpcd_set_lt_pattern(struct dc_link *link,
-	enum dc_dp_training_pattern pattern,
-	uint32_t hop)
-{
-	union dpcd_training_pattern dpcd_pattern = { {0} };
-	uint32_t dpcd_tps_offset = DP_TRAINING_PATTERN_SET;
-	enum dc_status status;
-
-	if (hop != DPRX)
-		dpcd_tps_offset = DP_TRAINING_PATTERN_SET_PHY_REPEATER1 +
-			((DP_REPEATER_CONFIGURATION_AND_STATUS_SIZE) * (hop - 1));
-
-	/* DpcdAddress_TrainingPatternSet */
-	dpcd_pattern.v1_4.TRAINING_PATTERN_SET =
-		dc_dp_training_pattern_to_dpcd_training_pattern(link, pattern);
-
-	dpcd_pattern.v1_4.SCRAMBLING_DISABLE =
-		dc_dp_initialize_scrambling_data_symbols(link, pattern);
-
-	if (hop != DPRX) {
-		DC_LOG_HW_LINK_TRAINING("%s\n LTTPR Repeater ID: %d\n 0x%X pattern = %x\n",
-					__func__,
-					hop,
-					dpcd_tps_offset,
-					dpcd_pattern.v1_4.TRAINING_PATTERN_SET);
-	} else {
-		DC_LOG_HW_LINK_TRAINING("%s\n 0x%X pattern = %x\n",
-					__func__,
-					dpcd_tps_offset,
-					dpcd_pattern.v1_4.TRAINING_PATTERN_SET);
-	}
-
-	status = core_link_write_dpcd(link,
-				      dpcd_tps_offset,
-				      &dpcd_pattern.raw,
-				      sizeof(dpcd_pattern.raw));
-
-	return status;
-}
-
-/* Execute clock recovery phase of link training for specified hop in display
- * path.in non-transparent mode:
- * - Driver issues both DPCD and SET_CONFIG transactions.
- * - TPS1 is transmitted for any hops downstream of DPOA.
- * - Drive (VS/PE) only transmitted for the hop immediately downstream of DPOA.
- * - CR for the first hop (DPTX-to-DPIA) is assumed to be successful.
- *
- * @param link DPIA link being trained.
- * @param lt_settings link_setting and drive settings (voltage swing and pre-emphasis).
- * @param hop The Hop in display path. DPRX = 0.
- */
-static enum link_training_result dpia_training_cr_non_transparent(struct dc_link *link,
-		struct link_training_settings *lt_settings,
-		uint32_t hop)
-{
-	enum link_training_result result = LINK_TRAINING_CR_FAIL_LANE0;
-	uint8_t repeater_cnt = 0; /* Number of hops/repeaters in display path. */
-	enum dc_status status;
-	uint32_t retries_cr = 0; /* Number of consecutive attempts with same VS or PE. */
-	uint32_t retry_count = 0;
-	/* From DP spec, CR read interval is always 100us. */
-	uint32_t wait_time_microsec = TRAINING_AUX_RD_INTERVAL;
-	enum dc_lane_count lane_count = lt_settings->link_settings.lane_count;
-	union lane_status dpcd_lane_status[LANE_COUNT_DP_MAX] = { { {0} } };
-	union lane_align_status_updated dpcd_lane_status_updated = { {0} };
-	union lane_adjust dpcd_lane_adjust[LANE_COUNT_DP_MAX] = { { {0} } };
-	uint8_t set_cfg_data;
-	enum dpia_set_config_ts ts;
-
-	repeater_cnt = dp_convert_to_count(link->dpcd_caps.lttpr_caps.phy_repeater_cnt);
-
-	/* Cap of LINK_TRAINING_MAX_CR_RETRY attempts at clock recovery.
-	 * Fix inherited from perform_clock_recovery_sequence() -
-	 * the DP equivalent of this function:
-	 * Required for Synaptics MST hub which can put the LT in
-	 * infinite loop by switching the VS between level 0 and level 1
-	 * continuously.
-	 */
-	while ((retries_cr < LINK_TRAINING_MAX_RETRY_COUNT) &&
-	       (retry_count < LINK_TRAINING_MAX_CR_RETRY)) {
-		/* DPTX-to-DPIA */
-		if (hop == repeater_cnt) {
-			/* Send SET_CONFIG(SET_LINK:LC,LR,LTTPR) to notify DPOA that
-			 * non-transparent link training has started.
-			 * This also enables the transmission of clk_sync packets.
-			 */
-			set_cfg_data = dpia_build_set_config_data(DPIA_SET_CFG_SET_LINK,
-					link,
-					lt_settings);
-			status = core_link_send_set_config(link,
-					DPIA_SET_CFG_SET_LINK,
-					set_cfg_data);
-			/* CR for this hop is considered successful as long as
-			 * SET_CONFIG message is acknowledged by DPOA.
-			 */
-			if (status == DC_OK)
-				result = LINK_TRAINING_SUCCESS;
-			else
-				result = LINK_TRAINING_ABORT;
-			break;
-		}
-
-		/* DPOA-to-x */
-		/* Instruct DPOA to transmit TPS1 then update DPCD. */
-		if (retry_count == 0) {
-			ts = convert_trng_ptn_to_trng_stg(lt_settings->pattern_for_cr);
-			status = core_link_send_set_config(link,
-					DPIA_SET_CFG_SET_TRAINING,
-					ts);
-			if (status != DC_OK) {
-				result = LINK_TRAINING_ABORT;
-				break;
-			}
-			status = dpcd_set_lt_pattern(link, lt_settings->pattern_for_cr, hop);
-			if (status != DC_OK) {
-				result = LINK_TRAINING_ABORT;
-				break;
-			}
-		}
-
-		/* Update DPOA drive settings then DPCD. DPOA does only adjusts
-		 * drive settings for hops immediately downstream.
-		 */
-		if (hop == repeater_cnt - 1) {
-			set_cfg_data = dpia_build_set_config_data(DPIA_SET_CFG_SET_VSPE,
-					link,
-					lt_settings);
-			status = core_link_send_set_config(link,
-					DPIA_SET_CFG_SET_VSPE,
-					set_cfg_data);
-			if (status != DC_OK) {
-				result = LINK_TRAINING_ABORT;
-				break;
-			}
-		}
-		status = dpcd_set_lane_settings(link, lt_settings, hop);
-		if (status != DC_OK) {
-			result = LINK_TRAINING_ABORT;
-			break;
-		}
-
-		dp_wait_for_training_aux_rd_interval(link, wait_time_microsec);
-
-		/* Read status and adjustment requests from DPCD. */
-		status = dp_get_lane_status_and_lane_adjust(
-				link,
-				lt_settings,
-				dpcd_lane_status,
-				&dpcd_lane_status_updated,
-				dpcd_lane_adjust,
-				hop);
-		if (status != DC_OK) {
-			result = LINK_TRAINING_ABORT;
-			break;
-		}
-
-		/* Check if clock recovery successful. */
-		if (dp_is_cr_done(lane_count, dpcd_lane_status)) {
-			result = LINK_TRAINING_SUCCESS;
-			break;
-		}
-
-		result = dp_get_cr_failure(lane_count, dpcd_lane_status);
-
-		if (dp_is_max_vs_reached(lt_settings))
-			break;
-
-		/* Count number of attempts with same drive settings.
-		 * Note: settings are the same for all lanes,
-		 * so comparing first lane is sufficient.
-		 */
-		if ((lt_settings->dpcd_lane_settings[0].bits.VOLTAGE_SWING_SET ==
-				dpcd_lane_adjust[0].bits.VOLTAGE_SWING_LANE)
-				&& (lt_settings->dpcd_lane_settings[0].bits.PRE_EMPHASIS_SET ==
-						dpcd_lane_adjust[0].bits.PRE_EMPHASIS_LANE))
-			retries_cr++;
-		else
-			retries_cr = 0;
-
-		/* Update VS/PE. */
-		dp_decide_lane_settings(lt_settings, dpcd_lane_adjust,
-				lt_settings->lane_settings,
-				lt_settings->dpcd_lane_settings);
-		retry_count++;
-	}
-
-	/* Abort link training if clock recovery failed due to HPD unplug. */
-	if (!link->hpd_status)
-		result = LINK_TRAINING_ABORT;
-
-	DC_LOG_HW_LINK_TRAINING("%s\n DPIA(%d) clock recovery\n"
-		" -hop(%d)\n - result(%d)\n - retries(%d)\n",
-		__func__,
-		link->link_id.enum_id - ENUM_ID_1,
-		hop,
-		result,
-		retry_count);
-
-	return result;
-}
-
-/* Execute clock recovery phase of link training in transparent LTTPR mode:
- * - Driver only issues DPCD transactions and leaves USB4 tunneling (SET_CONFIG) messages to DPIA.
- * - Driver writes TPS1 to DPCD to kick off training.
- * - Clock recovery (CR) for link is handled by DPOA, which reports result to DPIA on completion.
- * - DPIA communicates result to driver by updating CR status when driver reads DPCD.
- *
- * @param link DPIA link being trained.
- * @param lt_settings link_setting and drive settings (voltage swing and pre-emphasis).
- */
-static enum link_training_result dpia_training_cr_transparent(struct dc_link *link,
-		struct link_training_settings *lt_settings)
-{
-	enum link_training_result result = LINK_TRAINING_CR_FAIL_LANE0;
-	enum dc_status status;
-	uint32_t retries_cr = 0; /* Number of consecutive attempts with same VS or PE. */
-	uint32_t retry_count = 0;
-	uint32_t wait_time_microsec = lt_settings->cr_pattern_time;
-	enum dc_lane_count lane_count = lt_settings->link_settings.lane_count;
-	union lane_status dpcd_lane_status[LANE_COUNT_DP_MAX] = { { {0} } };
-	union lane_align_status_updated dpcd_lane_status_updated = { {0} };
-	union lane_adjust dpcd_lane_adjust[LANE_COUNT_DP_MAX] = { { {0} } };
-
-	/* Cap of LINK_TRAINING_MAX_CR_RETRY attempts at clock recovery.
-	 * Fix inherited from perform_clock_recovery_sequence() -
-	 * the DP equivalent of this function:
-	 * Required for Synaptics MST hub which can put the LT in
-	 * infinite loop by switching the VS between level 0 and level 1
-	 * continuously.
-	 */
-	while ((retries_cr < LINK_TRAINING_MAX_RETRY_COUNT) &&
-	       (retry_count < LINK_TRAINING_MAX_CR_RETRY)) {
-		/* Write TPS1 (not VS or PE) to DPCD to start CR phase.
-		 * DPIA sends SET_CONFIG(SET_LINK) to notify DPOA to
-		 * start link training.
-		 */
-		if (retry_count == 0) {
-			status = dpcd_set_lt_pattern(link, lt_settings->pattern_for_cr, DPRX);
-			if (status != DC_OK) {
-				result = LINK_TRAINING_ABORT;
-				break;
-			}
-		}
-
-		dp_wait_for_training_aux_rd_interval(link, wait_time_microsec);
-
-		/* Read status and adjustment requests from DPCD. */
-		status = dp_get_lane_status_and_lane_adjust(
-				link,
-				lt_settings,
-				dpcd_lane_status,
-				&dpcd_lane_status_updated,
-				dpcd_lane_adjust,
-				DPRX);
-		if (status != DC_OK) {
-			result = LINK_TRAINING_ABORT;
-			break;
-		}
-
-		/* Check if clock recovery successful. */
-		if (dp_is_cr_done(lane_count, dpcd_lane_status)) {
-			result = LINK_TRAINING_SUCCESS;
-			break;
-		}
-
-		result = dp_get_cr_failure(lane_count, dpcd_lane_status);
-
-		if (dp_is_max_vs_reached(lt_settings))
-			break;
-
-		/* Count number of attempts with same drive settings.
-		 * Note: settings are the same for all lanes,
-		 * so comparing first lane is sufficient.
-		 */
-		if ((lt_settings->dpcd_lane_settings[0].bits.VOLTAGE_SWING_SET ==
-				dpcd_lane_adjust[0].bits.VOLTAGE_SWING_LANE)
-				&& (lt_settings->dpcd_lane_settings[0].bits.PRE_EMPHASIS_SET ==
-						dpcd_lane_adjust[0].bits.PRE_EMPHASIS_LANE))
-			retries_cr++;
-		else
-			retries_cr = 0;
-
-		/* Update VS/PE. */
-		dp_decide_lane_settings(lt_settings, dpcd_lane_adjust,
-				lt_settings->hw_lane_settings, lt_settings->dpcd_lane_settings);
-		retry_count++;
-	}
-
-	/* Abort link training if clock recovery failed due to HPD unplug. */
-	if (!link->hpd_status)
-		result = LINK_TRAINING_ABORT;
-
-	DC_LOG_HW_LINK_TRAINING("%s\n DPIA(%d) clock recovery\n"
-		" -hop(%d)\n - result(%d)\n - retries(%d)\n",
-		__func__,
-		link->link_id.enum_id - ENUM_ID_1,
-		DPRX,
-		result,
-		retry_count);
-
-	return result;
-}
-
-/* Execute clock recovery phase of link training for specified hop in display
- * path.
- *
- * @param link DPIA link being trained.
- * @param lt_settings link_setting and drive settings (voltage swing and pre-emphasis).
- * @param hop The Hop in display path. DPRX = 0.
- */
-static enum link_training_result dpia_training_cr_phase(struct dc_link *link,
-		struct link_training_settings *lt_settings,
-		uint32_t hop)
-{
-	enum link_training_result result = LINK_TRAINING_CR_FAIL_LANE0;
-
-	if (link->lttpr_mode == LTTPR_MODE_NON_TRANSPARENT)
-		result = dpia_training_cr_non_transparent(link, lt_settings, hop);
-	else
-		result = dpia_training_cr_transparent(link, lt_settings);
-
-	return result;
-}
-
-/* Return status read interval during equalization phase. */
-static uint32_t dpia_get_eq_aux_rd_interval(const struct dc_link *link,
-		const struct link_training_settings *lt_settings,
-		uint32_t hop)
-{
-	uint32_t wait_time_microsec;
-
-	if (hop == DPRX)
-		wait_time_microsec = lt_settings->eq_pattern_time;
-	else
-		wait_time_microsec =
-				dp_translate_training_aux_read_interval(
-					link->dpcd_caps.lttpr_caps.aux_rd_interval[hop - 1]);
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	/* Check debug option for extending aux read interval. */
-	if (link->dc->debug.dpia_debug.bits.extend_aux_rd_interval)
-		wait_time_microsec = DPIA_DEBUG_EXTENDED_AUX_RD_INTERVAL_US;
-#endif
-
-	return wait_time_microsec;
-}
-
-/* Execute equalization phase of link training for specified hop in display
- * path in non-transparent mode:
- * - driver issues both DPCD and SET_CONFIG transactions.
- * - TPSx is transmitted for any hops downstream of DPOA.
- * - Drive (VS/PE) only transmitted for the hop immediately downstream of DPOA.
- * - EQ for the first hop (DPTX-to-DPIA) is assumed to be successful.
- * - DPRX EQ only reported successful when both DPRX and DPIA requirements
- * (clk sync packets sent) fulfilled.
- *
- * @param link DPIA link being trained.
- * @param lt_settings link_setting and drive settings (voltage swing and pre-emphasis).
- * @param hop The Hop in display path. DPRX = 0.
- */
-static enum link_training_result dpia_training_eq_non_transparent(struct dc_link *link,
-		struct link_training_settings *lt_settings,
-		uint32_t hop)
-{
-	enum link_training_result result = LINK_TRAINING_EQ_FAIL_EQ;
-	uint8_t repeater_cnt = 0; /* Number of hops/repeaters in display path. */
-	uint32_t retries_eq = 0;
-	enum dc_status status;
-	enum dc_dp_training_pattern tr_pattern;
-	uint32_t wait_time_microsec;
-	enum dc_lane_count lane_count = lt_settings->link_settings.lane_count;
-	union lane_align_status_updated dpcd_lane_status_updated = { {0} };
-	union lane_status dpcd_lane_status[LANE_COUNT_DP_MAX] = { { {0} } };
-	union lane_adjust dpcd_lane_adjust[LANE_COUNT_DP_MAX] = { { {0} } };
-	uint8_t set_cfg_data;
-	enum dpia_set_config_ts ts;
-
-	/* Training pattern is TPS4 for repeater;
-	 * TPS2/3/4 for DPRX depending on what it supports.
-	 */
-	if (hop == DPRX)
-		tr_pattern = lt_settings->pattern_for_eq;
-	else
-		tr_pattern = DP_TRAINING_PATTERN_SEQUENCE_4;
-
-	repeater_cnt = dp_convert_to_count(link->dpcd_caps.lttpr_caps.phy_repeater_cnt);
-
-	for (retries_eq = 0; retries_eq < LINK_TRAINING_MAX_RETRY_COUNT; retries_eq++) {
-		/* DPTX-to-DPIA equalization always successful. */
-		if (hop == repeater_cnt) {
-			result = LINK_TRAINING_SUCCESS;
-			break;
-		}
-
-		/* Instruct DPOA to transmit TPSn then update DPCD. */
-		if (retries_eq == 0) {
-			ts = convert_trng_ptn_to_trng_stg(tr_pattern);
-			status = core_link_send_set_config(link,
-					DPIA_SET_CFG_SET_TRAINING,
-					ts);
-			if (status != DC_OK) {
-				result = LINK_TRAINING_ABORT;
-				break;
-			}
-			status = dpcd_set_lt_pattern(link, tr_pattern, hop);
-			if (status != DC_OK) {
-				result = LINK_TRAINING_ABORT;
-				break;
-			}
-		}
-
-		/* Update DPOA drive settings then DPCD. DPOA only adjusts
-		 * drive settings for hop immediately downstream.
-		 */
-		if (hop == repeater_cnt - 1) {
-			set_cfg_data = dpia_build_set_config_data(DPIA_SET_CFG_SET_VSPE,
-								  link,
-								  lt_settings);
-			status = core_link_send_set_config(link,
-							   DPIA_SET_CFG_SET_VSPE,
-							   set_cfg_data);
-			if (status != DC_OK) {
-				result = LINK_TRAINING_ABORT;
-				break;
-			}
-		}
-		status = dpcd_set_lane_settings(link, lt_settings, hop);
-		if (status != DC_OK) {
-			result = LINK_TRAINING_ABORT;
-			break;
-		}
-
-		/* Extend wait time on second equalisation attempt on final hop to
-		 * ensure clock sync packets have been sent.
-		 */
-		if (hop == DPRX && retries_eq == 1)
-			wait_time_microsec = max(wait_time_microsec, (uint32_t)DPIA_CLK_SYNC_DELAY);
-		else
-			wait_time_microsec = dpia_get_eq_aux_rd_interval(link, lt_settings, hop);
-
-		dp_wait_for_training_aux_rd_interval(link, wait_time_microsec);
-
-		/* Read status and adjustment requests from DPCD. */
-		status = dp_get_lane_status_and_lane_adjust(
-				link,
-				lt_settings,
-				dpcd_lane_status,
-				&dpcd_lane_status_updated,
-				dpcd_lane_adjust,
-				hop);
-		if (status != DC_OK) {
-			result = LINK_TRAINING_ABORT;
-			break;
-		}
-
-		/* CR can still fail during EQ phase. Fail training if CR fails. */
-		if (!dp_is_cr_done(lane_count, dpcd_lane_status)) {
-			result = LINK_TRAINING_EQ_FAIL_CR;
-			break;
-		}
-
-		if (dp_is_ch_eq_done(lane_count, dpcd_lane_status) &&
-		    dp_is_symbol_locked(link->cur_link_settings.lane_count, dpcd_lane_status) &&
-		    dp_is_interlane_aligned(dpcd_lane_status_updated)) {
-			result =  LINK_TRAINING_SUCCESS;
-			break;
-		}
-
-		/* Update VS/PE. */
-		dp_decide_lane_settings(lt_settings, dpcd_lane_adjust,
-				lt_settings->hw_lane_settings, lt_settings->dpcd_lane_settings);
-	}
-
-	/* Abort link training if equalization failed due to HPD unplug. */
-	if (!link->hpd_status)
-		result = LINK_TRAINING_ABORT;
-
-	DC_LOG_HW_LINK_TRAINING("%s\n DPIA(%d) equalization\n"
-		" - hop(%d)\n - result(%d)\n - retries(%d)\n",
-		__func__,
-		link->link_id.enum_id - ENUM_ID_1,
-		hop,
-		result,
-		retries_eq);
-
-	return result;
-}
-
-/* Execute equalization phase of link training for specified hop in display
- * path in transparent LTTPR mode:
- * - driver only issues DPCD transactions leaves USB4 tunneling (SET_CONFIG) messages to DPIA.
- * - driver writes TPSx to DPCD to notify DPIA that is in equalization phase.
- * - equalization (EQ) for link is handled by DPOA, which reports result to DPIA on completion.
- * - DPIA communicates result to driver by updating EQ status when driver reads DPCD.
- *
- * @param link DPIA link being trained.
- * @param lt_settings link_setting and drive settings (voltage swing and pre-emphasis).
- * @param hop The Hop in display path. DPRX = 0.
- */
-static enum link_training_result dpia_training_eq_transparent(struct dc_link *link,
-		struct link_training_settings *lt_settings)
-{
-	enum link_training_result result = LINK_TRAINING_EQ_FAIL_EQ;
-	uint32_t retries_eq = 0;
-	enum dc_status status;
-	enum dc_dp_training_pattern tr_pattern = lt_settings->pattern_for_eq;
-	uint32_t wait_time_microsec;
-	enum dc_lane_count lane_count = lt_settings->link_settings.lane_count;
-	union lane_align_status_updated dpcd_lane_status_updated = { {0} };
-	union lane_status dpcd_lane_status[LANE_COUNT_DP_MAX] = { { {0} } };
-	union lane_adjust dpcd_lane_adjust[LANE_COUNT_DP_MAX] = { { {0} } };
-
-	wait_time_microsec = dpia_get_eq_aux_rd_interval(link, lt_settings, DPRX);
-
-	for (retries_eq = 0; retries_eq < LINK_TRAINING_MAX_RETRY_COUNT; retries_eq++) {
-		if (retries_eq == 0) {
-			status = dpcd_set_lt_pattern(link, tr_pattern, DPRX);
-			if (status != DC_OK) {
-				result = LINK_TRAINING_ABORT;
-				break;
-			}
-		}
-
-		dp_wait_for_training_aux_rd_interval(link, wait_time_microsec);
-
-		/* Read status and adjustment requests from DPCD. */
-		status = dp_get_lane_status_and_lane_adjust(
-				link,
-				lt_settings,
-				dpcd_lane_status,
-				&dpcd_lane_status_updated,
-				dpcd_lane_adjust,
-				DPRX);
-		if (status != DC_OK) {
-			result = LINK_TRAINING_ABORT;
-			break;
-		}
-
-		/* CR can still fail during EQ phase. Fail training if CR fails. */
-		if (!dp_is_cr_done(lane_count, dpcd_lane_status)) {
-			result = LINK_TRAINING_EQ_FAIL_CR;
-			break;
-		}
-
-		if (dp_is_ch_eq_done(lane_count, dpcd_lane_status) &&
-		    dp_is_symbol_locked(link->cur_link_settings.lane_count, dpcd_lane_status) &&
-		    dp_is_interlane_aligned(dpcd_lane_status_updated)) {
-			result =  LINK_TRAINING_SUCCESS;
-			break;
-		}
-
-		/* Update VS/PE. */
-		dp_decide_lane_settings(lt_settings, dpcd_lane_adjust,
-				lt_settings->hw_lane_settings, lt_settings->dpcd_lane_settings);
-	}
-
-	/* Abort link training if equalization failed due to HPD unplug. */
-	if (!link->hpd_status)
-		result = LINK_TRAINING_ABORT;
-
-	DC_LOG_HW_LINK_TRAINING("%s\n DPIA(%d) equalization\n"
-		" - hop(%d)\n - result(%d)\n - retries(%d)\n",
-		__func__,
-		link->link_id.enum_id - ENUM_ID_1,
-		DPRX,
-		result,
-		retries_eq);
-
-	return result;
-}
-
-/* Execute equalization phase of link training for specified hop in display
- * path.
- *
- * @param link DPIA link being trained.
- * @param lt_settings link_setting and drive settings (voltage swing and pre-emphasis).
- * @param hop The Hop in display path. DPRX = 0.
- */
-static enum link_training_result dpia_training_eq_phase(struct dc_link *link,
-		struct link_training_settings *lt_settings,
-		uint32_t hop)
-{
-	enum link_training_result result;
-
-	if (link->lttpr_mode == LTTPR_MODE_NON_TRANSPARENT)
-		result = dpia_training_eq_non_transparent(link, lt_settings, hop);
-	else
-		result = dpia_training_eq_transparent(link, lt_settings);
-
-	return result;
-}
-
-/* End training of specified hop in display path. */
-static enum dc_status dpcd_clear_lt_pattern(struct dc_link *link, uint32_t hop)
-{
-	union dpcd_training_pattern dpcd_pattern = { {0} };
-	uint32_t dpcd_tps_offset = DP_TRAINING_PATTERN_SET;
-	enum dc_status status;
-
-	if (hop != DPRX)
-		dpcd_tps_offset = DP_TRAINING_PATTERN_SET_PHY_REPEATER1 +
-			((DP_REPEATER_CONFIGURATION_AND_STATUS_SIZE) * (hop - 1));
-
-	status = core_link_write_dpcd(link,
-			dpcd_tps_offset,
-			&dpcd_pattern.raw,
-			sizeof(dpcd_pattern.raw));
-
-	return status;
-}
-
-/* End training of specified hop in display path.
- *
- * In transparent LTTPR mode:
- * - driver clears training pattern for the specified hop in DPCD.
- * In non-transparent LTTPR mode:
- * - in addition to clearing training pattern, driver issues USB4 tunneling
- * (SET_CONFIG) messages to notify DPOA when training is done for first hop
- * (DPTX-to-DPIA) and last hop (DPRX).
- *
- * @param link DPIA link being trained.
- * @param hop The Hop in display path. DPRX = 0.
- */
-static enum link_training_result dpia_training_end(struct dc_link *link,
-		uint32_t hop)
-{
-	enum link_training_result result = LINK_TRAINING_SUCCESS;
-	uint8_t repeater_cnt = 0; /* Number of hops/repeaters in display path. */
-	enum dc_status status;
-
-	if (link->lttpr_mode == LTTPR_MODE_NON_TRANSPARENT) {
-		repeater_cnt = dp_convert_to_count(link->dpcd_caps.lttpr_caps.phy_repeater_cnt);
-
-		if (hop == repeater_cnt) { /* DPTX-to-DPIA */
-			/* Send SET_CONFIG(SET_TRAINING:0xff) to notify DPOA that
-			 * DPTX-to-DPIA hop trained. No DPCD write needed for first hop.
-			 */
-			status = core_link_send_set_config(link,
-					DPIA_SET_CFG_SET_TRAINING,
-					DPIA_TS_UFP_DONE);
-			if (status != DC_OK)
-				result = LINK_TRAINING_ABORT;
-		} else { /* DPOA-to-x */
-			/* Write 0x0 to TRAINING_PATTERN_SET */
-			status = dpcd_clear_lt_pattern(link, hop);
-			if (status != DC_OK)
-				result = LINK_TRAINING_ABORT;
-		}
-
-		/* Notify DPOA that non-transparent link training of DPRX done. */
-		if (hop == DPRX && result != LINK_TRAINING_ABORT) {
-			status = core_link_send_set_config(link,
-					DPIA_SET_CFG_SET_TRAINING,
-					DPIA_TS_DPRX_DONE);
-			if (status != DC_OK)
-				result = LINK_TRAINING_ABORT;
-		}
-
-	} else { /* non-LTTPR or transparent LTTPR. */
-		/* Write 0x0 to TRAINING_PATTERN_SET */
-		status = dpcd_clear_lt_pattern(link, hop);
-		if (status != DC_OK)
-			result = LINK_TRAINING_ABORT;
-	}
-
-	DC_LOG_HW_LINK_TRAINING("%s\n DPIA(%d) end\n - hop(%d)\n - result(%d)\n - LTTPR mode(%d)\n",
-				__func__,
-				link->link_id.enum_id - ENUM_ID_1,
-				hop,
-				result,
-				link->lttpr_mode);
-
-	return result;
-}
-
-/* When aborting training of specified hop in display path, clean up by:
- * - Attempting to clear DPCD TRAINING_PATTERN_SET, LINK_BW_SET and LANE_COUNT_SET.
- * - Sending SET_CONFIG(SET_LINK) with lane count and link rate set to 0.
- *
- * @param link DPIA link being trained.
- * @param hop The Hop in display path. DPRX = 0.
- */
-static void dpia_training_abort(struct dc_link *link, uint32_t hop)
-{
-	uint8_t data = 0;
-	uint32_t dpcd_tps_offset = DP_TRAINING_PATTERN_SET;
-
-	DC_LOG_HW_LINK_TRAINING("%s\n DPIA(%d) aborting\n - LTTPR mode(%d)\n - HPD(%d)\n",
-				__func__,
-				link->link_id.enum_id - ENUM_ID_1,
-				link->lttpr_mode,
-				link->hpd_status);
-
-	/* Abandon clean-up if sink unplugged. */
-	if (!link->hpd_status)
-		return;
-
-	if (hop != DPRX)
-		dpcd_tps_offset = DP_TRAINING_PATTERN_SET_PHY_REPEATER1 +
-			((DP_REPEATER_CONFIGURATION_AND_STATUS_SIZE) * (hop - 1));
-
-	core_link_write_dpcd(link, dpcd_tps_offset, &data, 1);
-	core_link_write_dpcd(link, DP_LINK_BW_SET, &data, 1);
-	core_link_write_dpcd(link, DP_LANE_COUNT_SET, &data, 1);
-	core_link_send_set_config(link, DPIA_SET_CFG_SET_LINK, data);
-}
-
-enum link_training_result dc_link_dpia_perform_link_training(struct dc_link *link,
-	const struct dc_link_settings *link_setting,
-	bool skip_video_pattern)
-{
-	enum link_training_result result;
-	struct link_training_settings lt_settings;
-	uint8_t repeater_cnt = 0; /* Number of hops/repeaters in display path. */
-	int8_t repeater_id; /* Current hop. */
-
-	/* Configure link as prescribed in link_setting and set LTTPR mode. */
-	result = dpia_configure_link(link, link_setting, &lt_settings);
-	if (result != LINK_TRAINING_SUCCESS)
-		return result;
-
-	if (link->lttpr_mode == LTTPR_MODE_NON_TRANSPARENT)
-		repeater_cnt = dp_convert_to_count(link->dpcd_caps.lttpr_caps.phy_repeater_cnt);
-
-	/* Train each hop in turn starting with the one closest to DPTX.
-	 * In transparent or non-LTTPR mode, train only the final hop (DPRX).
-	 */
-	for (repeater_id = repeater_cnt; repeater_id >= 0; repeater_id--) {
-		/* Clock recovery. */
-		result = dpia_training_cr_phase(link, &lt_settings, repeater_id);
-		if (result != LINK_TRAINING_SUCCESS)
-			break;
-
-		/* Equalization. */
-		result = dpia_training_eq_phase(link, &lt_settings, repeater_id);
-		if (result != LINK_TRAINING_SUCCESS)
-			break;
-
-		/* Stop training hop. */
-		result = dpia_training_end(link, repeater_id);
-		if (result != LINK_TRAINING_SUCCESS)
-			break;
-	}
-
-	/* Double-check link status if training successful; gracefully abort
-	 * training of current hop if training failed due to message tunneling
-	 * failure; end training of hop if training ended conventionally and
-	 * falling back to lower bandwidth settings possible.
-	 */
-	if (result == LINK_TRAINING_SUCCESS) {
-		msleep(5);
-		result = dp_check_link_loss_status(link, &lt_settings);
-	} else if (result == LINK_TRAINING_ABORT) {
-		dpia_training_abort(link, repeater_id);
-	} else {
-		dpia_training_end(link, repeater_id);
-	}
-	return result;
-}
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_link_enc_cfg.c b/drivers/gpu/drm/amd/display/dc/core/dc_link_enc_cfg.c
index 25e48a8cbb78..deb23d20bca6 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_link_enc_cfg.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_link_enc_cfg.c
@@ -24,7 +24,9 @@
 
 #include "link_enc_cfg.h"
 #include "resource.h"
-#include "dc_link_dp.h"
+#include "link_service.h"
+
+#define DC_LOGGER dc->ctx->logger
 
 /* Check whether stream is supported by DIG link encoders. */
 static bool is_dig_link_enc_stream(struct dc_stream_state *stream)
@@ -42,20 +44,8 @@ static bool is_dig_link_enc_stream(struct dc_stream_state *stream)
 			 * yet match.
 			 */
 			if (link_enc && ((uint32_t)stream->link->connector_signal & link_enc->output_signals)) {
-				if (dc_is_dp_signal(stream->signal)) {
-					/* DIGs do not support DP2.0 streams with 128b/132b encoding. */
-					struct dc_link_settings link_settings = {0};
-
-					decide_link_settings(stream, &link_settings);
-					if ((link_settings.link_rate >= LINK_RATE_LOW) &&
-							link_settings.link_rate <= LINK_RATE_HIGH3) {
-						is_dig_stream = true;
-						break;
-					}
-				} else {
-					is_dig_stream = true;
-					break;
-				}
+				is_dig_stream = true;
+				break;
 			}
 		}
 	}
@@ -118,7 +108,11 @@ static void remove_link_enc_assignment(
 				 */
 				if (get_stream_using_link_enc(state, eng_id) == NULL)
 					state->res_ctx.link_enc_cfg_ctx.link_enc_avail[eng_idx] = eng_id;
+
 				stream->link_enc = NULL;
+				state->res_ctx.link_enc_cfg_ctx.link_enc_assignments[i].eng_id = ENGINE_ID_UNKNOWN;
+				state->res_ctx.link_enc_cfg_ctx.link_enc_assignments[i].stream = NULL;
+				dc_stream_release(stream);
 				break;
 			}
 		}
@@ -148,6 +142,7 @@ static void add_link_enc_assignment(
 						.ep_type = stream->link->ep_type},
 					.eng_id = eng_id,
 					.stream = stream};
+				dc_stream_retain(stream);
 				state->res_ctx.link_enc_cfg_ctx.link_enc_avail[eng_idx] = ENGINE_ID_UNKNOWN;
 				stream->link_enc = stream->ctx->dc->res_pool->link_encoders[eng_idx];
 				break;
@@ -162,11 +157,23 @@ static void add_link_enc_assignment(
 /* Return first available DIG link encoder. */
 static enum engine_id find_first_avail_link_enc(
 		const struct dc_context *ctx,
-		const struct dc_state *state)
+		const struct dc_state *state,
+		enum engine_id eng_id_requested)
 {
 	enum engine_id eng_id = ENGINE_ID_UNKNOWN;
 	int i;
 
+	if (eng_id_requested != ENGINE_ID_UNKNOWN) {
+
+		for (i = 0; i < ctx->dc->res_pool->res_cap->num_dig_link_enc; i++) {
+			eng_id = state->res_ctx.link_enc_cfg_ctx.link_enc_avail[i];
+			if (eng_id == eng_id_requested)
+				return eng_id;
+		}
+	}
+
+	eng_id = ENGINE_ID_UNKNOWN;
+
 	for (i = 0; i < ctx->dc->res_pool->res_cap->num_dig_link_enc; i++) {
 		eng_id = state->res_ctx.link_enc_cfg_ctx.link_enc_avail[i];
 		if (eng_id != ENGINE_ID_UNKNOWN)
@@ -227,8 +234,10 @@ static struct link_encoder *get_link_enc_used_by_link(
 		.link_id = link->link_id,
 		.ep_type = link->ep_type};
 
-	for (i = 0; i < state->stream_count; i++) {
+	for (i = 0; i < MAX_PIPES; i++) {
 		struct link_enc_assignment assignment = state->res_ctx.link_enc_cfg_ctx.link_enc_assignments[i];
+		if (assignment.eng_id == ENGINE_ID_UNKNOWN)
+			continue;
 
 		if (assignment.valid == true && are_ep_ids_equal(&assignment.ep_id, &ep_id))
 			link_enc = link->dc->res_pool->link_encoders[assignment.eng_id - ENGINE_ID_DIGA];
@@ -237,28 +246,18 @@ static struct link_encoder *get_link_enc_used_by_link(
 	return link_enc;
 }
 /* Clear all link encoder assignments. */
-static void clear_enc_assignments(struct dc_state *state)
+static void clear_enc_assignments(const struct dc *dc, struct dc_state *state)
 {
 	int i;
-	enum engine_id eng_id;
-	struct dc_stream_state *stream;
 
 	for (i = 0; i < MAX_PIPES; i++) {
 		state->res_ctx.link_enc_cfg_ctx.link_enc_assignments[i].valid = false;
-		eng_id = state->res_ctx.link_enc_cfg_ctx.link_enc_assignments[i].eng_id;
-		stream = state->res_ctx.link_enc_cfg_ctx.link_enc_assignments[i].stream;
-		if (eng_id != ENGINE_ID_UNKNOWN)
-			state->res_ctx.link_enc_cfg_ctx.link_enc_avail[eng_id - ENGINE_ID_DIGA] = eng_id;
-		if (stream)
-			stream->link_enc = NULL;
+		state->res_ctx.link_enc_cfg_ctx.link_enc_assignments[i].eng_id = ENGINE_ID_UNKNOWN;
+		if (state->res_ctx.link_enc_cfg_ctx.link_enc_assignments[i].stream != NULL) {
+			dc_stream_release(state->res_ctx.link_enc_cfg_ctx.link_enc_assignments[i].stream);
+			state->res_ctx.link_enc_cfg_ctx.link_enc_assignments[i].stream = NULL;
+		}
 	}
-}
-
-void link_enc_cfg_init(
-		struct dc *dc,
-		struct dc_state *state)
-{
-	int i;
 
 	for (i = 0; i < dc->res_pool->res_cap->num_dig_link_enc; i++) {
 		if (dc->res_pool->link_encoders[i])
@@ -266,30 +265,40 @@ void link_enc_cfg_init(
 		else
 			state->res_ctx.link_enc_cfg_ctx.link_enc_avail[i] = ENGINE_ID_UNKNOWN;
 	}
+}
 
-	clear_enc_assignments(state);
+void link_enc_cfg_init(
+		const struct dc *dc,
+		struct dc_state *state)
+{
+	clear_enc_assignments(dc, state);
 
 	state->res_ctx.link_enc_cfg_ctx.mode = LINK_ENC_CFG_STEADY;
 }
 
+void link_enc_cfg_copy(const struct dc_state *src_ctx, struct dc_state *dst_ctx)
+{
+	memcpy(&dst_ctx->res_ctx.link_enc_cfg_ctx,
+	       &src_ctx->res_ctx.link_enc_cfg_ctx,
+	       sizeof(dst_ctx->res_ctx.link_enc_cfg_ctx));
+}
+
 void link_enc_cfg_link_encs_assign(
 		struct dc *dc,
 		struct dc_state *state,
 		struct dc_stream_state *streams[],
 		uint8_t stream_count)
 {
-	enum engine_id eng_id = ENGINE_ID_UNKNOWN;
+	enum engine_id eng_id = ENGINE_ID_UNKNOWN, eng_id_req = ENGINE_ID_UNKNOWN;
 	int i;
 	int j;
 
 	ASSERT(state->stream_count == stream_count);
-
-	if (stream_count == 0)
-		clear_enc_assignments(state);
+	ASSERT(dc->current_state->res_ctx.link_enc_cfg_ctx.mode == LINK_ENC_CFG_STEADY);
 
 	/* Release DIG link encoder resources before running assignment algorithm. */
-	for (i = 0; i < stream_count; i++)
-		dc->res_pool->funcs->link_enc_unassign(state, streams[i]);
+	for (i = 0; i < dc->current_state->stream_count; i++)
+		dc->res_pool->funcs->link_enc_unassign(state, dc->current_state->streams[i]);
 
 	for (i = 0; i < MAX_PIPES; i++)
 		ASSERT(state->res_ctx.link_enc_cfg_ctx.link_enc_assignments[i].valid == false);
@@ -298,15 +307,17 @@ void link_enc_cfg_link_encs_assign(
 	for (i = 0; i < stream_count; i++) {
 		struct dc_stream_state *stream = streams[i];
 
+		/* skip it if the link is mappable endpoint. */
+		if (stream->link->is_dig_mapping_flexible)
+			continue;
+
 		/* Skip stream if not supported by DIG link encoder. */
 		if (!is_dig_link_enc_stream(stream))
 			continue;
 
 		/* Physical endpoints have a fixed mapping to DIG link encoders. */
-		if (!stream->link->is_dig_mapping_flexible) {
-			eng_id = stream->link->eng_id;
-			add_link_enc_assignment(state, stream, eng_id);
-		}
+		eng_id = stream->link->eng_id;
+		add_link_enc_assignment(state, stream, eng_id);
 	}
 
 	/* (b) Retain previous assignments for mappable endpoints if encoders still available. */
@@ -318,11 +329,12 @@ void link_enc_cfg_link_encs_assign(
 		for (i = 0; i < stream_count; i++) {
 			struct dc_stream_state *stream = state->streams[i];
 
-			/* Skip stream if not supported by DIG link encoder. */
-			if (!is_dig_link_enc_stream(stream))
+			/* Skip it if the link is NOT mappable endpoint. */
+			if (!stream->link->is_dig_mapping_flexible)
 				continue;
 
-			if (!stream->link->is_dig_mapping_flexible)
+			/* Skip stream if not supported by DIG link encoder. */
+			if (!is_dig_link_enc_stream(stream))
 				continue;
 
 			for (j = 0; j < prev_state->stream_count; j++) {
@@ -331,6 +343,7 @@ void link_enc_cfg_link_encs_assign(
 				if (stream == prev_stream && stream->link == prev_stream->link &&
 						prev_state->res_ctx.link_enc_cfg_ctx.link_enc_assignments[j].valid) {
 					eng_id = prev_state->res_ctx.link_enc_cfg_ctx.link_enc_assignments[j].eng_id;
+
 					if (is_avail_link_enc(state, eng_id, stream))
 						add_link_enc_assignment(state, stream, eng_id);
 				}
@@ -343,6 +356,15 @@ void link_enc_cfg_link_encs_assign(
 
 	for (i = 0; i < stream_count; i++) {
 		struct dc_stream_state *stream = streams[i];
+		struct link_encoder *link_enc = NULL;
+
+		/* Skip it if the link is NOT mappable endpoint. */
+		if (!stream->link->is_dig_mapping_flexible)
+			continue;
+
+		/* Skip if encoder assignment retained in step (b) above. */
+		if (stream->link_enc)
+			continue;
 
 		/* Skip stream if not supported by DIG link encoder. */
 		if (!is_dig_link_enc_stream(stream)) {
@@ -351,24 +373,24 @@ void link_enc_cfg_link_encs_assign(
 		}
 
 		/* Mappable endpoints have a flexible mapping to DIG link encoders. */
-		if (stream->link->is_dig_mapping_flexible) {
-			struct link_encoder *link_enc = NULL;
 
-			/* Skip if encoder assignment retained in step (b) above. */
-			if (stream->link_enc)
-				continue;
+		/* For MST, multiple streams will share the same link / display
+		 * endpoint. These streams should use the same link encoder
+		 * assigned to that endpoint.
+		 */
+		link_enc = get_link_enc_used_by_link(state, stream->link);
+		if (link_enc == NULL) {
 
-			/* For MST, multiple streams will share the same link / display
-			 * endpoint. These streams should use the same link encoder
-			 * assigned to that endpoint.
-			 */
-			link_enc = get_link_enc_used_by_link(state, stream->link);
-			if (link_enc == NULL)
-				eng_id = find_first_avail_link_enc(stream->ctx, state);
-			else
-				eng_id =  link_enc->preferred_engine;
-			add_link_enc_assignment(state, stream, eng_id);
+			if (stream->link->ep_type == DISPLAY_ENDPOINT_USB4_DPIA &&
+					stream->link->dpia_preferred_eng_id != ENGINE_ID_UNKNOWN)
+				eng_id_req = stream->link->dpia_preferred_eng_id;
+
+			eng_id = find_first_avail_link_enc(stream->ctx, state, eng_id_req);
 		}
+		else
+			eng_id =  link_enc->preferred_engine;
+
+		add_link_enc_assignment(state, stream, eng_id);
 	}
 
 	link_enc_cfg_validate(dc, state);
@@ -379,6 +401,34 @@ void link_enc_cfg_link_encs_assign(
 			state->res_ctx.link_enc_cfg_ctx.link_enc_assignments[i];
 	}
 
+	/* Log encoder assignments. */
+	for (i = 0; i < MAX_PIPES; i++) {
+		struct link_enc_assignment assignment =
+				dc->current_state->res_ctx.link_enc_cfg_ctx.link_enc_assignments[i];
+
+		if (assignment.valid)
+			DC_LOG_DEBUG("%s: CUR %s(%d) - enc_id(%d)\n",
+					__func__,
+					assignment.ep_id.ep_type == DISPLAY_ENDPOINT_PHY ? "PHY" : "DPIA",
+					assignment.ep_id.ep_type == DISPLAY_ENDPOINT_PHY ?
+							assignment.ep_id.link_id.enum_id :
+							assignment.ep_id.link_id.enum_id - 1,
+					assignment.eng_id);
+	}
+	for (i = 0; i < MAX_PIPES; i++) {
+		struct link_enc_assignment assignment =
+				state->res_ctx.link_enc_cfg_ctx.link_enc_assignments[i];
+
+		if (assignment.valid)
+			DC_LOG_DEBUG("%s: NEW %s(%d) - enc_id(%d)\n",
+					__func__,
+					assignment.ep_id.ep_type == DISPLAY_ENDPOINT_PHY ? "PHY" : "DPIA",
+					assignment.ep_id.ep_type == DISPLAY_ENDPOINT_PHY ?
+							assignment.ep_id.link_id.enum_id :
+							assignment.ep_id.link_id.enum_id - 1,
+					assignment.eng_id);
+	}
+
 	/* Current state mode will be set to steady once this state committed. */
 	state->res_ctx.link_enc_cfg_ctx.mode = LINK_ENC_CFG_STEADY;
 }
@@ -389,10 +439,6 @@ void link_enc_cfg_link_enc_unassign(
 {
 	enum engine_id eng_id = ENGINE_ID_UNKNOWN;
 
-	/* Only DIG link encoders. */
-	if (!is_dig_link_enc_stream(stream))
-		return;
-
 	if (stream->link_enc)
 		eng_id = stream->link_enc->preferred_engine;
 
@@ -444,7 +490,6 @@ struct dc_link *link_enc_cfg_get_link_using_link_enc(
 	if (stream)
 		link = stream->link;
 
-	// dm_output_to_console("%s: No link using DIG(%d).\n", __func__, eng_id);
 	return link;
 }
 
@@ -462,6 +507,8 @@ struct link_encoder *link_enc_cfg_get_link_enc_used_by_link(
 
 	for (i = 0; i < MAX_PIPES; i++) {
 		struct link_enc_assignment assignment = get_assignment(dc, i);
+		if (assignment.eng_id == ENGINE_ID_UNKNOWN)
+			continue;
 
 		if (assignment.valid == true && are_ep_ids_equal(&assignment.ep_id, &ep_id)) {
 			link_enc = link->dc->res_pool->link_encoders[assignment.eng_id - ENGINE_ID_DIGA];
@@ -475,22 +522,23 @@ struct link_encoder *link_enc_cfg_get_link_enc_used_by_link(
 struct link_encoder *link_enc_cfg_get_next_avail_link_enc(struct dc *dc)
 {
 	struct link_encoder *link_enc = NULL;
-	enum engine_id encs_assigned[MAX_DIG_LINK_ENCODERS];
+	enum engine_id encs_assigned[MAX_LINK_ENCODERS];
 	int i;
 
-	for (i = 0; i < MAX_DIG_LINK_ENCODERS; i++)
+	for (i = 0; i < MAX_LINK_ENCODERS; i++)
 		encs_assigned[i] = ENGINE_ID_UNKNOWN;
 
 	/* Add assigned encoders to list. */
 	for (i = 0; i < MAX_PIPES; i++) {
 		struct link_enc_assignment assignment = get_assignment(dc, i);
 
-		if (assignment.valid)
+		if (assignment.valid && assignment.eng_id != ENGINE_ID_UNKNOWN)
 			encs_assigned[assignment.eng_id - ENGINE_ID_DIGA] = assignment.eng_id;
 	}
 
 	for (i = 0; i < dc->res_pool->res_cap->num_dig_link_enc; i++) {
-		if (encs_assigned[i] == ENGINE_ID_UNKNOWN) {
+		if (encs_assigned[i] == ENGINE_ID_UNKNOWN &&
+				dc->res_pool->link_encoders[i] != NULL) {
 			link_enc = dc->res_pool->link_encoders[i];
 			break;
 		}
@@ -499,13 +547,50 @@ struct link_encoder *link_enc_cfg_get_next_avail_link_enc(struct dc *dc)
 	return link_enc;
 }
 
-struct link_encoder *link_enc_cfg_get_link_enc_used_by_stream(
+struct link_encoder *link_enc_cfg_get_link_enc(
+		const struct dc_link *link)
+{
+	struct link_encoder *link_enc = NULL;
+
+	/* Links supporting dynamically assigned link encoder will be assigned next
+	 * available encoder if one not already assigned.
+	 */
+	if (link->is_dig_mapping_flexible &&
+	    link->dc->res_pool->funcs->link_encs_assign) {
+		link_enc = link_enc_cfg_get_link_enc_used_by_link(link->ctx->dc, link);
+		if (link_enc == NULL)
+			link_enc = link_enc_cfg_get_next_avail_link_enc(
+				link->ctx->dc);
+	} else
+		link_enc = link->link_enc;
+
+	return link_enc;
+}
+
+struct link_encoder *link_enc_cfg_get_link_enc_used_by_stream_current(
 		struct dc *dc,
 		const struct dc_stream_state *stream)
 {
-	struct link_encoder *link_enc;
+	struct link_encoder *link_enc = NULL;
+	struct display_endpoint_id ep_id;
+	int i;
+
+	ep_id = (struct display_endpoint_id) {
+		.link_id = stream->link->link_id,
+		.ep_type = stream->link->ep_type};
+
+	for (i = 0; i < MAX_PIPES; i++) {
+		struct link_enc_assignment assignment =
+			dc->current_state->res_ctx.link_enc_cfg_ctx.link_enc_assignments[i];
+
+		if (assignment.eng_id == ENGINE_ID_UNKNOWN)
+			continue;
 
-	link_enc = link_enc_cfg_get_link_enc_used_by_link(dc, stream->link);
+		if (assignment.valid == true && are_ep_ids_equal(&assignment.ep_id, &ep_id)) {
+			link_enc = stream->link->dc->res_pool->link_encoders[assignment.eng_id - ENGINE_ID_DIGA];
+			break;
+		}
+	}
 
 	return link_enc;
 }
@@ -542,8 +627,9 @@ bool link_enc_cfg_validate(struct dc *dc, struct dc_state *state)
 	int i, j;
 	uint8_t valid_count = 0;
 	uint8_t dig_stream_count = 0;
-	int matching_stream_ptrs = 0;
 	int eng_ids_per_ep_id[MAX_PIPES] = {0};
+	int ep_ids_per_eng_id[MAX_PIPES] = {0};
+	int valid_bitmap = 0;
 
 	/* (1) No. valid entries same as stream count. */
 	for (i = 0; i < MAX_PIPES; i++) {
@@ -563,9 +649,7 @@ bool link_enc_cfg_validate(struct dc *dc, struct dc_state *state)
 		struct link_enc_assignment assignment = state->res_ctx.link_enc_cfg_ctx.link_enc_assignments[i];
 
 		if (assignment.valid) {
-			if (assignment.stream == state->streams[i])
-				matching_stream_ptrs++;
-			else
+			if (assignment.stream != state->streams[i])
 				valid_stream_ptrs = false;
 		}
 	}
@@ -578,6 +662,7 @@ bool link_enc_cfg_validate(struct dc *dc, struct dc_state *state)
 			struct display_endpoint_id ep_id_i = assignment_i.ep_id;
 
 			eng_ids_per_ep_id[i]++;
+			ep_ids_per_eng_id[i]++;
 			for (j = 0; j < MAX_PIPES; j++) {
 				struct link_enc_assignment assignment_j =
 					state->res_ctx.link_enc_cfg_ctx.link_enc_assignments[j];
@@ -592,6 +677,10 @@ bool link_enc_cfg_validate(struct dc *dc, struct dc_state *state)
 							assignment_i.eng_id != assignment_j.eng_id) {
 						valid_uniqueness = false;
 						eng_ids_per_ep_id[i]++;
+					} else if (!are_ep_ids_equal(&ep_id_i, &ep_id_j) &&
+							assignment_i.eng_id == assignment_j.eng_id) {
+						valid_uniqueness = false;
+						ep_ids_per_eng_id[i]++;
 					}
 				}
 			}
@@ -625,5 +714,32 @@ bool link_enc_cfg_validate(struct dc *dc, struct dc_state *state)
 	is_valid = valid_entries && valid_stream_ptrs && valid_uniqueness && valid_avail && valid_streams;
 	ASSERT(is_valid);
 
+	if (is_valid == false) {
+		valid_bitmap =
+			(valid_entries & 0x1) |
+			((valid_stream_ptrs & 0x1) << 1) |
+			((valid_uniqueness & 0x1) << 2) |
+			((valid_avail & 0x1) << 3) |
+			((valid_streams & 0x1) << 4);
+		DC_LOG_ERROR("%s: Invalid link encoder assignments - 0x%x\n", __func__, valid_bitmap);
+	}
+
 	return is_valid;
 }
+
+void link_enc_cfg_set_transient_mode(struct dc *dc, struct dc_state *current_state, struct dc_state *new_state)
+{
+	int i = 0;
+	int num_transient_assignments = 0;
+
+	for (i = 0; i < MAX_PIPES; i++) {
+		if (current_state->res_ctx.link_enc_cfg_ctx.transient_assignments[i].valid)
+			num_transient_assignments++;
+	}
+
+	/* Only enter transient mode if the new encoder assignments are valid. */
+	if (new_state->stream_count == num_transient_assignments) {
+		current_state->res_ctx.link_enc_cfg_ctx.mode = LINK_ENC_CFG_TRANSIENT;
+		DC_LOG_DEBUG("%s: current_state(%p) mode(%d)\n", __func__, current_state, LINK_ENC_CFG_TRANSIENT);
+	}
+}
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_link_exports.c b/drivers/gpu/drm/amd/display/dc/core/dc_link_exports.c
new file mode 100644
index 000000000000..9acd30019717
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_link_exports.c
@@ -0,0 +1,530 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+/* FILE POLICY AND INTENDED USAGE:
+ * This file provides single entrance to link functionality declared in dc
+ * public headers. The file is intended to be used as a thin translation layer
+ * that directly calls link internal functions without adding new functional
+ * behavior.
+ *
+ * When exporting a new link related dc function, add function declaration in
+ * dc.h with detail interface documentation, then add function implementation
+ * in this file which calls link functions.
+ */
+#include "link_service.h"
+#include "dce/dce_i2c.h"
+
+struct dc_link *dc_get_link_at_index(struct dc *dc, uint32_t link_index)
+{
+	if (link_index >= MAX_LINKS)
+		return NULL;
+
+	return dc->links[link_index];
+}
+
+void dc_get_edp_links(const struct dc *dc,
+		struct dc_link **edp_links,
+		int *edp_num)
+{
+	int i;
+
+	*edp_num = 0;
+	for (i = 0; i < dc->link_count; i++) {
+		// report any eDP links, even unconnected DDI's
+		if (!dc->links[i])
+			continue;
+		if (dc->links[i]->connector_signal == SIGNAL_TYPE_EDP) {
+			edp_links[*edp_num] = dc->links[i];
+			if (++(*edp_num) == MAX_NUM_EDP)
+				return;
+		}
+	}
+}
+
+bool dc_get_edp_link_panel_inst(const struct dc *dc,
+		const struct dc_link *link,
+		unsigned int *inst_out)
+{
+	struct dc_link *edp_links[MAX_NUM_EDP];
+	int edp_num, i;
+
+	*inst_out = 0;
+	if (link->connector_signal != SIGNAL_TYPE_EDP)
+		return false;
+	dc_get_edp_links(dc, edp_links, &edp_num);
+	for (i = 0; i < edp_num; i++) {
+		if (link == edp_links[i])
+			break;
+		(*inst_out)++;
+	}
+	return true;
+}
+
+bool dc_link_detect(struct dc_link *link, enum dc_detect_reason reason)
+{
+	return link->dc->link_srv->detect_link(link, reason);
+}
+
+bool dc_link_detect_connection_type(struct dc_link *link,
+		enum dc_connection_type *type)
+{
+	return link->dc->link_srv->detect_connection_type(link, type);
+}
+
+const struct dc_link_status *dc_link_get_status(const struct dc_link *link)
+{
+	return link->dc->link_srv->get_status(link);
+}
+
+/* return true if the connected receiver supports the hdcp version */
+bool dc_link_is_hdcp14(struct dc_link *link, enum signal_type signal)
+{
+	return link->dc->link_srv->is_hdcp1x_supported(link, signal);
+}
+
+bool dc_link_is_hdcp22(struct dc_link *link, enum signal_type signal)
+{
+	return link->dc->link_srv->is_hdcp2x_supported(link, signal);
+}
+
+void dc_link_clear_dprx_states(struct dc_link *link)
+{
+	link->dc->link_srv->clear_dprx_states(link);
+}
+
+bool dc_link_reset_cur_dp_mst_topology(struct dc_link *link)
+{
+	return link->dc->link_srv->reset_cur_dp_mst_topology(link);
+}
+
+uint32_t dc_link_bandwidth_kbps(
+	const struct dc_link *link,
+	const struct dc_link_settings *link_settings)
+{
+	return link->dc->link_srv->dp_link_bandwidth_kbps(link, link_settings);
+}
+
+uint32_t dc_link_required_hblank_size_bytes(
+	const struct dc_link *link,
+	struct dp_audio_bandwidth_params *audio_params)
+{
+	return link->dc->link_srv->dp_required_hblank_size_bytes(link,
+			audio_params);
+}
+
+void dc_get_cur_link_res_map(const struct dc *dc, uint32_t *map)
+{
+	dc->link_srv->get_cur_res_map(dc, map);
+}
+
+void dc_restore_link_res_map(const struct dc *dc, uint32_t *map)
+{
+	dc->link_srv->restore_res_map(dc, map);
+}
+
+bool dc_link_update_dsc_config(struct pipe_ctx *pipe_ctx)
+{
+	struct dc_link *link = pipe_ctx->stream->link;
+
+	return link->dc->link_srv->update_dsc_config(pipe_ctx);
+}
+
+struct ddc_service *
+dc_get_oem_i2c_device(struct dc *dc)
+{
+	return dc->res_pool->oem_device;
+}
+
+bool dc_is_oem_i2c_device_present(
+	struct dc *dc,
+	size_t slave_address)
+{
+	if (dc->res_pool->oem_device)
+		return dce_i2c_oem_device_present(
+			dc->res_pool,
+			dc->res_pool->oem_device,
+			slave_address);
+
+	return false;
+}
+
+bool dc_submit_i2c(
+		struct dc *dc,
+		uint32_t link_index,
+		struct i2c_command *cmd)
+{
+
+	struct dc_link *link = dc->links[link_index];
+	struct ddc_service *ddc = link->ddc;
+
+	return dce_i2c_submit_command(
+		dc->res_pool,
+		ddc->ddc_pin,
+		cmd);
+}
+
+bool dc_submit_i2c_oem(
+		struct dc *dc,
+		struct i2c_command *cmd)
+{
+	struct ddc_service *ddc = dc->res_pool->oem_device;
+
+	if (ddc)
+		return dce_i2c_submit_command(
+			dc->res_pool,
+			ddc->ddc_pin,
+			cmd);
+
+	return false;
+}
+
+void dc_link_dp_handle_automated_test(struct dc_link *link)
+{
+	link->dc->link_srv->dp_handle_automated_test(link);
+}
+
+bool dc_link_dp_set_test_pattern(
+	struct dc_link *link,
+	enum dp_test_pattern test_pattern,
+	enum dp_test_pattern_color_space test_pattern_color_space,
+	const struct link_training_settings *p_link_settings,
+	const unsigned char *p_custom_pattern,
+	unsigned int cust_pattern_size)
+{
+	return link->dc->link_srv->dp_set_test_pattern(link, test_pattern,
+			test_pattern_color_space, p_link_settings,
+			p_custom_pattern, cust_pattern_size);
+}
+
+void dc_link_set_drive_settings(struct dc *dc,
+				struct link_training_settings *lt_settings,
+				struct dc_link *link)
+{
+	struct link_resource link_res;
+
+	dc->link_srv->get_cur_link_res(link, &link_res);
+	dc->link_srv->dp_set_drive_settings(link, &link_res, lt_settings);
+}
+
+void dc_link_set_preferred_link_settings(struct dc *dc,
+					 struct dc_link_settings *link_setting,
+					 struct dc_link *link)
+{
+	dc->link_srv->dp_set_preferred_link_settings(dc, link_setting, link);
+}
+
+void dc_link_set_preferred_training_settings(struct dc *dc,
+		struct dc_link_settings *link_setting,
+		struct dc_link_training_overrides *lt_overrides,
+		struct dc_link *link,
+		bool skip_immediate_retrain)
+{
+	dc->link_srv->dp_set_preferred_training_settings(dc, link_setting,
+			lt_overrides, link, skip_immediate_retrain);
+}
+
+bool dc_dp_trace_is_initialized(struct dc_link *link)
+{
+	return link->dc->link_srv->dp_trace_is_initialized(link);
+}
+
+void dc_dp_trace_set_is_logged_flag(struct dc_link *link,
+		bool in_detection,
+		bool is_logged)
+{
+	link->dc->link_srv->dp_trace_set_is_logged_flag(link, in_detection, is_logged);
+}
+
+bool dc_dp_trace_is_logged(struct dc_link *link, bool in_detection)
+{
+	return link->dc->link_srv->dp_trace_is_logged(link, in_detection);
+}
+
+unsigned long long dc_dp_trace_get_lt_end_timestamp(struct dc_link *link,
+		bool in_detection)
+{
+	return link->dc->link_srv->dp_trace_get_lt_end_timestamp(link, in_detection);
+}
+
+const struct dp_trace_lt_counts *dc_dp_trace_get_lt_counts(struct dc_link *link,
+		bool in_detection)
+{
+	return link->dc->link_srv->dp_trace_get_lt_counts(link, in_detection);
+}
+
+unsigned int dc_dp_trace_get_link_loss_count(struct dc_link *link)
+{
+	return link->dc->link_srv->dp_trace_get_link_loss_count(link);
+}
+
+struct dc_sink *dc_link_add_remote_sink(
+		struct dc_link *link,
+		const uint8_t *edid,
+		int len,
+		struct dc_sink_init_data *init_data)
+{
+	return link->dc->link_srv->add_remote_sink(link, edid, len, init_data);
+}
+
+void dc_link_remove_remote_sink(struct dc_link *link, struct dc_sink *sink)
+{
+	link->dc->link_srv->remove_remote_sink(link, sink);
+}
+
+int dc_link_aux_transfer_raw(struct ddc_service *ddc,
+		struct aux_payload *payload,
+		enum aux_return_code_type *operation_result)
+{
+	const struct dc *dc = ddc->link->dc;
+
+	return dc->link_srv->aux_transfer_raw(
+			ddc, payload, operation_result);
+}
+
+uint32_t dc_link_bw_kbps_from_raw_frl_link_rate_data(const struct dc *dc, uint8_t bw)
+{
+	return dc->link_srv->bw_kbps_from_raw_frl_link_rate_data(bw);
+}
+
+bool dc_link_decide_edp_link_settings(struct dc_link *link,
+		struct dc_link_settings *link_setting, uint32_t req_bw)
+{
+	return link->dc->link_srv->edp_decide_link_settings(link, link_setting, req_bw);
+}
+
+
+bool dc_link_dp_get_max_link_enc_cap(const struct dc_link *link,
+		struct dc_link_settings *max_link_enc_cap)
+{
+	return link->dc->link_srv->dp_get_max_link_enc_cap(link, max_link_enc_cap);
+}
+
+enum dp_link_encoding dc_link_dp_mst_decide_link_encoding_format(
+		const struct dc_link *link)
+{
+	return link->dc->link_srv->mst_decide_link_encoding_format(link);
+}
+
+const struct dc_link_settings *dc_link_get_link_cap(const struct dc_link *link)
+{
+	return link->dc->link_srv->dp_get_verified_link_cap(link);
+}
+
+enum dc_link_encoding_format dc_link_get_highest_encoding_format(const struct dc_link *link)
+{
+	if (dc_is_dp_signal(link->connector_signal)) {
+		if (link->dpcd_caps.dongle_type >= DISPLAY_DONGLE_DP_DVI_DONGLE &&
+				link->dpcd_caps.dongle_type <= DISPLAY_DONGLE_DP_HDMI_MISMATCHED_DONGLE)
+			return DC_LINK_ENCODING_HDMI_TMDS;
+		else if (link->dc->link_srv->dp_get_encoding_format(&link->verified_link_cap) ==
+				DP_8b_10b_ENCODING)
+			return DC_LINK_ENCODING_DP_8b_10b;
+		else if (link->dc->link_srv->dp_get_encoding_format(&link->verified_link_cap) ==
+				DP_128b_132b_ENCODING)
+			return DC_LINK_ENCODING_DP_128b_132b;
+	} else if (dc_is_hdmi_signal(link->connector_signal)) {
+	}
+
+	return DC_LINK_ENCODING_UNSPECIFIED;
+}
+
+bool dc_link_is_dp_sink_present(struct dc_link *link)
+{
+	return link->dc->link_srv->dp_is_sink_present(link);
+}
+
+bool dc_link_is_fec_supported(const struct dc_link *link)
+{
+	return link->dc->link_srv->dp_is_fec_supported(link);
+}
+
+void dc_link_overwrite_extended_receiver_cap(
+		struct dc_link *link)
+{
+	link->dc->link_srv->dp_overwrite_extended_receiver_cap(link);
+}
+
+bool dc_link_should_enable_fec(const struct dc_link *link)
+{
+	return link->dc->link_srv->dp_should_enable_fec(link);
+}
+
+void dc_link_dp_dpia_handle_usb4_bandwidth_allocation_for_link(
+		struct dc_link *link, int peak_bw)
+{
+	link->dc->link_srv->dpia_handle_usb4_bandwidth_allocation_for_link(link, peak_bw);
+}
+
+bool dc_link_check_link_loss_status(
+	struct dc_link *link,
+	union hpd_irq_data *hpd_irq_dpcd_data)
+{
+	return link->dc->link_srv->dp_parse_link_loss_status(link, hpd_irq_dpcd_data);
+}
+
+bool dc_link_dp_allow_hpd_rx_irq(const struct dc_link *link)
+{
+	return link->dc->link_srv->dp_should_allow_hpd_rx_irq(link);
+}
+
+void dc_link_dp_handle_link_loss(struct dc_link *link)
+{
+	link->dc->link_srv->dp_handle_link_loss(link);
+}
+
+enum dc_status dc_link_dp_read_hpd_rx_irq_data(
+	struct dc_link *link,
+	union hpd_irq_data *irq_data)
+{
+	return link->dc->link_srv->dp_read_hpd_rx_irq_data(link, irq_data);
+}
+
+bool dc_link_handle_hpd_rx_irq(struct dc_link *link,
+		union hpd_irq_data *out_hpd_irq_dpcd_data, bool *out_link_loss,
+		bool defer_handling, bool *has_left_work)
+{
+	return link->dc->link_srv->dp_handle_hpd_rx_irq(link, out_hpd_irq_dpcd_data,
+			out_link_loss, defer_handling, has_left_work);
+}
+
+void dc_link_dp_receiver_power_ctrl(struct dc_link *link, bool on)
+{
+	link->dc->link_srv->dpcd_write_rx_power_ctrl(link, on);
+}
+
+enum lttpr_mode dc_link_decide_lttpr_mode(struct dc_link *link,
+		struct dc_link_settings *link_setting)
+{
+	return link->dc->link_srv->dp_decide_lttpr_mode(link, link_setting);
+}
+
+void dc_link_edp_panel_backlight_power_on(struct dc_link *link, bool wait_for_hpd)
+{
+	link->dc->link_srv->edp_panel_backlight_power_on(link, wait_for_hpd);
+}
+
+int dc_link_get_backlight_level(const struct dc_link *link)
+{
+	return link->dc->link_srv->edp_get_backlight_level(link);
+}
+
+bool dc_link_get_backlight_level_nits(struct dc_link *link,
+		uint32_t *backlight_millinits_avg,
+		uint32_t *backlight_millinits_peak)
+{
+	return link->dc->link_srv->edp_get_backlight_level_nits(link,
+			backlight_millinits_avg,
+			backlight_millinits_peak);
+}
+
+bool dc_link_set_backlight_level(const struct dc_link *link,
+		struct set_backlight_level_params *backlight_level_params)
+{
+	return link->dc->link_srv->edp_set_backlight_level(link,
+			backlight_level_params);
+}
+
+bool dc_link_set_backlight_level_nits(struct dc_link *link,
+		bool isHDR,
+		uint32_t backlight_millinits,
+		uint32_t transition_time_in_ms)
+{
+	return link->dc->link_srv->edp_set_backlight_level_nits(link, isHDR,
+			backlight_millinits, transition_time_in_ms);
+}
+
+int dc_link_get_target_backlight_pwm(const struct dc_link *link)
+{
+	return link->dc->link_srv->edp_get_target_backlight_pwm(link);
+}
+
+bool dc_link_get_psr_state(const struct dc_link *link, enum dc_psr_state *state)
+{
+	return link->dc->link_srv->edp_get_psr_state(link, state);
+}
+
+bool dc_link_set_psr_allow_active(struct dc_link *link, const bool *allow_active,
+		bool wait, bool force_static, const unsigned int *power_opts)
+{
+	return link->dc->link_srv->edp_set_psr_allow_active(link, allow_active, wait,
+			force_static, power_opts);
+}
+
+bool dc_link_setup_psr(struct dc_link *link,
+		const struct dc_stream_state *stream, struct psr_config *psr_config,
+		struct psr_context *psr_context)
+{
+	return link->dc->link_srv->edp_setup_psr(link, stream, psr_config, psr_context);
+}
+
+bool dc_link_set_replay_allow_active(struct dc_link *link, const bool *allow_active,
+		bool wait, bool force_static, const unsigned int *power_opts)
+{
+	return link->dc->link_srv->edp_set_replay_allow_active(link, allow_active, wait,
+			force_static, power_opts);
+}
+
+bool dc_link_get_replay_state(const struct dc_link *link, uint64_t *state)
+{
+	return link->dc->link_srv->edp_get_replay_state(link, state);
+}
+
+bool dc_link_wait_for_t12(struct dc_link *link)
+{
+	return link->dc->link_srv->edp_wait_for_t12(link);
+}
+
+bool dc_link_get_hpd_state(struct dc_link *link)
+{
+	return link->dc->link_srv->get_hpd_state(link);
+}
+
+void dc_link_enable_hpd(const struct dc_link *link)
+{
+	link->dc->link_srv->enable_hpd(link);
+}
+
+void dc_link_disable_hpd(const struct dc_link *link)
+{
+	link->dc->link_srv->disable_hpd(link);
+}
+
+void dc_link_enable_hpd_filter(struct dc_link *link, bool enable)
+{
+	link->dc->link_srv->enable_hpd_filter(link, enable);
+}
+
+enum dc_status dc_link_validate_dp_tunneling_bandwidth(const struct dc *dc, const struct dc_state *new_ctx)
+{
+	return dc->link_srv->validate_dp_tunnel_bandwidth(dc, new_ctx);
+}
+
+void dc_link_get_alpm_support(struct dc_link *link,
+	bool *auxless_support,
+	bool *auxwake_support)
+{
+	link->dc->link_srv->edp_get_alpm_support(link, auxless_support, auxwake_support);
+}
+
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_link_hwss.c b/drivers/gpu/drm/amd/display/dc/core/dc_link_hwss.c
deleted file mode 100644
index 368e834c6809..000000000000
--- a/drivers/gpu/drm/amd/display/dc/core/dc_link_hwss.c
+++ /dev/null
@@ -1,906 +0,0 @@
-/* Copyright 2015 Advanced Micro Devices, Inc. */
-
-
-#include "dm_services.h"
-#include "dc.h"
-#include "inc/core_types.h"
-#include "include/ddc_service_types.h"
-#include "include/i2caux_interface.h"
-#include "link_hwss.h"
-#include "hw_sequencer.h"
-#include "dc_link_dp.h"
-#include "dc_link_ddc.h"
-#include "dm_helpers.h"
-#include "dpcd_defs.h"
-#include "dsc.h"
-#include "resource.h"
-#include "link_enc_cfg.h"
-#include "clk_mgr.h"
-#include "inc/link_dpcd.h"
-#include "dccg.h"
-
-static uint8_t convert_to_count(uint8_t lttpr_repeater_count)
-{
-	switch (lttpr_repeater_count) {
-	case 0x80: // 1 lttpr repeater
-		return 1;
-	case 0x40: // 2 lttpr repeaters
-		return 2;
-	case 0x20: // 3 lttpr repeaters
-		return 3;
-	case 0x10: // 4 lttpr repeaters
-		return 4;
-	case 0x08: // 5 lttpr repeaters
-		return 5;
-	case 0x04: // 6 lttpr repeaters
-		return 6;
-	case 0x02: // 7 lttpr repeaters
-		return 7;
-	case 0x01: // 8 lttpr repeaters
-		return 8;
-	default:
-		break;
-	}
-	return 0; // invalid value
-}
-
-static inline bool is_immediate_downstream(struct dc_link *link, uint32_t offset)
-{
-	return (convert_to_count(link->dpcd_caps.lttpr_caps.phy_repeater_cnt) == offset);
-}
-
-void dp_receiver_power_ctrl(struct dc_link *link, bool on)
-{
-	uint8_t state;
-
-	state = on ? DP_POWER_STATE_D0 : DP_POWER_STATE_D3;
-
-	if (link->sync_lt_in_progress)
-		return;
-
-	core_link_write_dpcd(link, DP_SET_POWER, &state,
-			sizeof(state));
-}
-
-void dp_source_sequence_trace(struct dc_link *link, uint8_t dp_test_mode)
-{
-	if (link != NULL && link->dc->debug.enable_driver_sequence_debug)
-		core_link_write_dpcd(link, DP_SOURCE_SEQUENCE,
-					&dp_test_mode, sizeof(dp_test_mode));
-}
-
-void dp_enable_link_phy(
-	struct dc_link *link,
-	enum signal_type signal,
-	enum clock_source_id clock_source,
-	const struct dc_link_settings *link_settings)
-{
-	struct link_encoder *link_enc;
-	struct dc  *dc = link->ctx->dc;
-	struct dmcu *dmcu = dc->res_pool->dmcu;
-
-	struct pipe_ctx *pipes =
-			link->dc->current_state->res_ctx.pipe_ctx;
-	struct clock_source *dp_cs =
-			link->dc->res_pool->dp_clock_source;
-	unsigned int i;
-
-	/* Link should always be assigned encoder when en-/disabling. */
-	if (link->is_dig_mapping_flexible && dc->res_pool->funcs->link_encs_assign)
-		link_enc = link_enc_cfg_get_link_enc_used_by_link(dc, link);
-	else
-		link_enc = link->link_enc;
-	ASSERT(link_enc);
-
-	if (link->connector_signal == SIGNAL_TYPE_EDP) {
-		link->dc->hwss.edp_power_control(link, true);
-		link->dc->hwss.edp_wait_for_hpd_ready(link, true);
-	}
-
-	/* If the current pixel clock source is not DTO(happens after
-	 * switching from HDMI passive dongle to DP on the same connector),
-	 * switch the pixel clock source to DTO.
-	 */
-	for (i = 0; i < MAX_PIPES; i++) {
-		if (pipes[i].stream != NULL &&
-			pipes[i].stream->link == link) {
-			if (pipes[i].clock_source != NULL &&
-					pipes[i].clock_source->id != CLOCK_SOURCE_ID_DP_DTO) {
-				pipes[i].clock_source = dp_cs;
-				pipes[i].stream_res.pix_clk_params.requested_pix_clk_100hz =
-						pipes[i].stream->timing.pix_clk_100hz;
-				pipes[i].clock_source->funcs->program_pix_clk(
-							pipes[i].clock_source,
-							&pipes[i].stream_res.pix_clk_params,
-							&pipes[i].pll_settings);
-			}
-		}
-	}
-
-	link->cur_link_settings = *link_settings;
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	if (dp_get_link_encoding_format(link_settings) == DP_128b_132b_ENCODING) {
-		/* TODO - DP2.0 HW: notify link rate change here */
-	} else if (dp_get_link_encoding_format(link_settings) == DP_8b_10b_ENCODING) {
-		if (dc->clk_mgr->funcs->notify_link_rate_change)
-			dc->clk_mgr->funcs->notify_link_rate_change(dc->clk_mgr, link);
-	}
-#else
-	if (dc->clk_mgr->funcs->notify_link_rate_change)
-		dc->clk_mgr->funcs->notify_link_rate_change(dc->clk_mgr, link);
-#endif
-	if (dmcu != NULL && dmcu->funcs->lock_phy)
-		dmcu->funcs->lock_phy(dmcu);
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	if (dp_get_link_encoding_format(link_settings) == DP_128b_132b_ENCODING) {
-		enable_dp_hpo_output(link, link_settings);
-	} else if (dp_get_link_encoding_format(link_settings) == DP_8b_10b_ENCODING) {
-		if (dc_is_dp_sst_signal(signal)) {
-			link_enc->funcs->enable_dp_output(
-							link_enc,
-							link_settings,
-							clock_source);
-		} else {
-			link_enc->funcs->enable_dp_mst_output(
-							link_enc,
-							link_settings,
-							clock_source);
-		}
-	}
-#else
-	if (dc_is_dp_sst_signal(signal)) {
-		link_enc->funcs->enable_dp_output(
-						link_enc,
-						link_settings,
-						clock_source);
-	} else {
-		link_enc->funcs->enable_dp_mst_output(
-						link_enc,
-						link_settings,
-						clock_source);
-	}
-#endif
-	if (dmcu != NULL && dmcu->funcs->unlock_phy)
-		dmcu->funcs->unlock_phy(dmcu);
-
-	dp_source_sequence_trace(link, DPCD_SOURCE_SEQ_AFTER_ENABLE_LINK_PHY);
-	dp_receiver_power_ctrl(link, true);
-}
-
-void edp_add_delay_for_T9(struct dc_link *link)
-{
-	if (link->local_sink &&
-			link->local_sink->edid_caps.panel_patch.extra_delay_backlight_off > 0)
-		udelay(link->local_sink->edid_caps.panel_patch.extra_delay_backlight_off * 1000);
-}
-
-bool edp_receiver_ready_T9(struct dc_link *link)
-{
-	unsigned int tries = 0;
-	unsigned char sinkstatus = 0;
-	unsigned char edpRev = 0;
-	enum dc_status result;
-
-	result = core_link_read_dpcd(link, DP_EDP_DPCD_REV, &edpRev, sizeof(edpRev));
-
-    /* start from eDP version 1.2, SINK_STAUS indicate the sink is ready.*/
-	if (result == DC_OK && edpRev >= DP_EDP_12) {
-		do {
-			sinkstatus = 1;
-			result = core_link_read_dpcd(link, DP_SINK_STATUS, &sinkstatus, sizeof(sinkstatus));
-			if (sinkstatus == 0)
-				break;
-			if (result != DC_OK)
-				break;
-			udelay(100); //MAx T9
-		} while (++tries < 50);
-	}
-
-	return result;
-}
-bool edp_receiver_ready_T7(struct dc_link *link)
-{
-	unsigned char sinkstatus = 0;
-	unsigned char edpRev = 0;
-	enum dc_status result;
-
-	/* use absolute time stamp to constrain max T7*/
-	unsigned long long enter_timestamp = 0;
-	unsigned long long finish_timestamp = 0;
-	unsigned long long time_taken_in_ns = 0;
-
-	result = core_link_read_dpcd(link, DP_EDP_DPCD_REV, &edpRev, sizeof(edpRev));
-
-	if (result == DC_OK && edpRev >= DP_EDP_12) {
-		/* start from eDP version 1.2, SINK_STAUS indicate the sink is ready.*/
-		enter_timestamp = dm_get_timestamp(link->ctx);
-		do {
-			sinkstatus = 0;
-			result = core_link_read_dpcd(link, DP_SINK_STATUS, &sinkstatus, sizeof(sinkstatus));
-			if (sinkstatus == 1)
-				break;
-			if (result != DC_OK)
-				break;
-			udelay(25);
-			finish_timestamp = dm_get_timestamp(link->ctx);
-			time_taken_in_ns = dm_get_elapse_time_in_ns(link->ctx, finish_timestamp, enter_timestamp);
-		} while (time_taken_in_ns < 50 * 1000000); //MAx T7 is 50ms
-	}
-
-	if (link->local_sink &&
-			link->local_sink->edid_caps.panel_patch.extra_t7_ms > 0)
-		udelay(link->local_sink->edid_caps.panel_patch.extra_t7_ms * 1000);
-
-	return result;
-}
-
-void dp_disable_link_phy(struct dc_link *link, enum signal_type signal)
-{
-	struct dc  *dc = link->ctx->dc;
-	struct dmcu *dmcu = dc->res_pool->dmcu;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	struct hpo_dp_link_encoder *hpo_link_enc = link->hpo_dp_link_enc;
-#endif
-	struct link_encoder *link_enc;
-
-	/* Link should always be assigned encoder when en-/disabling. */
-	if (link->is_dig_mapping_flexible && dc->res_pool->funcs->link_encs_assign)
-		link_enc = link_enc_cfg_get_link_enc_used_by_link(dc, link);
-	else
-		link_enc = link->link_enc;
-	ASSERT(link_enc);
-
-	if (!link->wa_flags.dp_keep_receiver_powered)
-		dp_receiver_power_ctrl(link, false);
-
-	if (signal == SIGNAL_TYPE_EDP) {
-		if (link->dc->hwss.edp_backlight_control)
-			link->dc->hwss.edp_backlight_control(link, false);
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-		if (dp_get_link_encoding_format(&link->cur_link_settings) == DP_128b_132b_ENCODING)
-			disable_dp_hpo_output(link, signal);
-		else
-			link_enc->funcs->disable_output(link_enc, signal);
-#else
-		link_enc->funcs->disable_output(link_enc, signal);
-#endif
-		link->dc->hwss.edp_power_control(link, false);
-	} else {
-		if (dmcu != NULL && dmcu->funcs->lock_phy)
-			dmcu->funcs->lock_phy(dmcu);
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-		if (dp_get_link_encoding_format(&link->cur_link_settings) == DP_128b_132b_ENCODING &&
-				hpo_link_enc)
-			disable_dp_hpo_output(link, signal);
-		else
-			link_enc->funcs->disable_output(link_enc, signal);
-#else
-		link_enc->funcs->disable_output(link_enc, signal);
-#endif
-		if (dmcu != NULL && dmcu->funcs->unlock_phy)
-			dmcu->funcs->unlock_phy(dmcu);
-	}
-
-	dp_source_sequence_trace(link, DPCD_SOURCE_SEQ_AFTER_DISABLE_LINK_PHY);
-
-	/* Clear current link setting.*/
-	memset(&link->cur_link_settings, 0,
-			sizeof(link->cur_link_settings));
-
-	if (dc->clk_mgr->funcs->notify_link_rate_change)
-		dc->clk_mgr->funcs->notify_link_rate_change(dc->clk_mgr, link);
-}
-
-void dp_disable_link_phy_mst(struct dc_link *link, enum signal_type signal)
-{
-	/* MST disable link only when no stream use the link */
-	if (link->mst_stream_alloc_table.stream_count > 0)
-		return;
-
-	dp_disable_link_phy(link, signal);
-
-	/* set the sink to SST mode after disabling the link */
-	dp_enable_mst_on_sink(link, false);
-}
-
-bool dp_set_hw_training_pattern(
-	struct dc_link *link,
-	enum dc_dp_training_pattern pattern,
-	uint32_t offset)
-{
-	enum dp_test_pattern test_pattern = DP_TEST_PATTERN_UNSUPPORTED;
-
-	switch (pattern) {
-	case DP_TRAINING_PATTERN_SEQUENCE_1:
-		test_pattern = DP_TEST_PATTERN_TRAINING_PATTERN1;
-		break;
-	case DP_TRAINING_PATTERN_SEQUENCE_2:
-		test_pattern = DP_TEST_PATTERN_TRAINING_PATTERN2;
-		break;
-	case DP_TRAINING_PATTERN_SEQUENCE_3:
-		test_pattern = DP_TEST_PATTERN_TRAINING_PATTERN3;
-		break;
-	case DP_TRAINING_PATTERN_SEQUENCE_4:
-		test_pattern = DP_TEST_PATTERN_TRAINING_PATTERN4;
-		break;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	case DP_128b_132b_TPS1:
-		test_pattern = DP_TEST_PATTERN_128b_132b_TPS1_TRAINING_MODE;
-		break;
-	case DP_128b_132b_TPS2:
-		test_pattern = DP_TEST_PATTERN_128b_132b_TPS2_TRAINING_MODE;
-		break;
-#endif
-	default:
-		break;
-	}
-
-	dp_set_hw_test_pattern(link, test_pattern, NULL, 0);
-
-	return true;
-}
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-#define DC_LOGGER \
-	link->ctx->logger
-#endif
-void dp_set_hw_lane_settings(
-	struct dc_link *link,
-	const struct link_training_settings *link_settings,
-	uint32_t offset)
-{
-	struct link_encoder *encoder = link->link_enc;
-
-	if ((link->lttpr_mode == LTTPR_MODE_NON_TRANSPARENT) && !is_immediate_downstream(link, offset))
-		return;
-
-	/* call Encoder to set lane settings */
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	if (dp_get_link_encoding_format(&link_settings->link_settings) ==
-			DP_128b_132b_ENCODING) {
-		link->hpo_dp_link_enc->funcs->set_ffe(
-				link->hpo_dp_link_enc,
-				&link_settings->link_settings,
-				link_settings->lane_settings[0].FFE_PRESET.raw);
-	} else if (dp_get_link_encoding_format(&link_settings->link_settings)
-			== DP_8b_10b_ENCODING) {
-		encoder->funcs->dp_set_lane_settings(encoder, link_settings);
-	}
-#else
-	encoder->funcs->dp_set_lane_settings(encoder, link_settings);
-#endif
-	memmove(link->cur_lane_setting,
-			link_settings->lane_settings,
-			sizeof(link->cur_lane_setting));
-}
-
-void dp_set_hw_test_pattern(
-	struct dc_link *link,
-	enum dp_test_pattern test_pattern,
-	uint8_t *custom_pattern,
-	uint32_t custom_pattern_size)
-{
-	struct encoder_set_dp_phy_pattern_param pattern_param = {0};
-	struct link_encoder *encoder;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	enum dp_link_encoding link_encoding_format = dp_get_link_encoding_format(&link->cur_link_settings);
-#endif
-
-	/* Access link encoder based on whether it is statically
-	 * or dynamically assigned to a link.
-	 */
-	if (link->is_dig_mapping_flexible &&
-			link->dc->res_pool->funcs->link_encs_assign)
-		encoder = link_enc_cfg_get_link_enc_used_by_link(link->ctx->dc, link);
-	else
-		encoder = link->link_enc;
-
-	pattern_param.dp_phy_pattern = test_pattern;
-	pattern_param.custom_pattern = custom_pattern;
-	pattern_param.custom_pattern_size = custom_pattern_size;
-	pattern_param.dp_panel_mode = dp_get_panel_mode(link);
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	switch (link_encoding_format) {
-	case DP_128b_132b_ENCODING:
-		link->hpo_dp_link_enc->funcs->set_link_test_pattern(
-				link->hpo_dp_link_enc, &pattern_param);
-		break;
-	case DP_8b_10b_ENCODING:
-		ASSERT(encoder);
-		encoder->funcs->dp_set_phy_pattern(encoder, &pattern_param);
-		break;
-	default:
-		DC_LOG_ERROR("%s: Unknown link encoding format.", __func__);
-		break;
-	}
-#else
-	encoder->funcs->dp_set_phy_pattern(encoder, &pattern_param);
-#endif
-	dp_source_sequence_trace(link, DPCD_SOURCE_SEQ_AFTER_SET_SOURCE_PATTERN);
-}
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-#undef DC_LOGGER
-#endif
-
-void dp_retrain_link_dp_test(struct dc_link *link,
-			struct dc_link_settings *link_setting,
-			bool skip_video_pattern)
-{
-	struct pipe_ctx *pipes =
-			&link->dc->current_state->res_ctx.pipe_ctx[0];
-	unsigned int i;
-
-	for (i = 0; i < MAX_PIPES; i++) {
-		if (pipes[i].stream != NULL &&
-			!pipes[i].top_pipe && !pipes[i].prev_odm_pipe &&
-			pipes[i].stream->link != NULL &&
-			pipes[i].stream_res.stream_enc != NULL &&
-			pipes[i].stream->link == link) {
-			udelay(100);
-
-			pipes[i].stream_res.stream_enc->funcs->dp_blank(link,
-					pipes[i].stream_res.stream_enc);
-
-			/* disable any test pattern that might be active */
-			dp_set_hw_test_pattern(link,
-					DP_TEST_PATTERN_VIDEO_MODE, NULL, 0);
-
-			dp_receiver_power_ctrl(link, false);
-
-			link->dc->hwss.disable_stream(&pipes[i]);
-			if ((&pipes[i])->stream_res.audio && !link->dc->debug.az_endpoint_mute_only)
-				(&pipes[i])->stream_res.audio->funcs->az_disable((&pipes[i])->stream_res.audio);
-
-			if (link->link_enc)
-				link->link_enc->funcs->disable_output(
-						link->link_enc,
-						SIGNAL_TYPE_DISPLAY_PORT);
-
-			/* Clear current link setting. */
-			memset(&link->cur_link_settings, 0,
-				sizeof(link->cur_link_settings));
-
-			perform_link_training_with_retries(
-					link_setting,
-					skip_video_pattern,
-					LINK_TRAINING_ATTEMPTS,
-					&pipes[i],
-					SIGNAL_TYPE_DISPLAY_PORT,
-					false);
-
-			link->dc->hwss.enable_stream(&pipes[i]);
-
-			link->dc->hwss.unblank_stream(&pipes[i],
-					link_setting);
-
-			if (pipes[i].stream_res.audio) {
-				/* notify audio driver for
-				 * audio modes of monitor */
-				pipes[i].stream_res.audio->funcs->az_enable(
-						pipes[i].stream_res.audio);
-
-				/* un-mute audio */
-				/* TODO: audio should be per stream rather than
-				 * per link */
-				pipes[i].stream_res.stream_enc->funcs->
-				audio_mute_control(
-					pipes[i].stream_res.stream_enc, false);
-			}
-		}
-	}
-}
-
-#define DC_LOGGER \
-	dsc->ctx->logger
-static void dsc_optc_config_log(struct display_stream_compressor *dsc,
-		struct dsc_optc_config *config)
-{
-	uint32_t precision = 1 << 28;
-	uint32_t bytes_per_pixel_int = config->bytes_per_pixel / precision;
-	uint32_t bytes_per_pixel_mod = config->bytes_per_pixel % precision;
-	uint64_t ll_bytes_per_pix_fraq = bytes_per_pixel_mod;
-
-	/* 7 fractional digits decimal precision for bytes per pixel is enough because DSC
-	 * bits per pixel precision is 1/16th of a pixel, which means bytes per pixel precision is
-	 * 1/16/8 = 1/128 of a byte, or 0.0078125 decimal
-	 */
-	ll_bytes_per_pix_fraq *= 10000000;
-	ll_bytes_per_pix_fraq /= precision;
-
-	DC_LOG_DSC("\tbytes_per_pixel 0x%08x (%d.%07d)",
-			config->bytes_per_pixel, bytes_per_pixel_int, (uint32_t)ll_bytes_per_pix_fraq);
-	DC_LOG_DSC("\tis_pixel_format_444 %d", config->is_pixel_format_444);
-	DC_LOG_DSC("\tslice_width %d", config->slice_width);
-}
-
-bool dp_set_dsc_on_rx(struct pipe_ctx *pipe_ctx, bool enable)
-{
-	struct dc *dc = pipe_ctx->stream->ctx->dc;
-	struct dc_stream_state *stream = pipe_ctx->stream;
-	bool result = false;
-
-	if (dc_is_virtual_signal(stream->signal) || IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment))
-		result = true;
-	else
-		result = dm_helpers_dp_write_dsc_enable(dc->ctx, stream, enable);
-	return result;
-}
-
-/* The stream with these settings can be sent (unblanked) only after DSC was enabled on RX first,
- * i.e. after dp_enable_dsc_on_rx() had been called
- */
-void dp_set_dsc_on_stream(struct pipe_ctx *pipe_ctx, bool enable)
-{
-	struct display_stream_compressor *dsc = pipe_ctx->stream_res.dsc;
-	struct dc *dc = pipe_ctx->stream->ctx->dc;
-	struct dc_stream_state *stream = pipe_ctx->stream;
-	struct pipe_ctx *odm_pipe;
-	int opp_cnt = 1;
-
-	for (odm_pipe = pipe_ctx->next_odm_pipe; odm_pipe; odm_pipe = odm_pipe->next_odm_pipe)
-		opp_cnt++;
-
-	if (enable) {
-		struct dsc_config dsc_cfg;
-		struct dsc_optc_config dsc_optc_cfg;
-		enum optc_dsc_mode optc_dsc_mode;
-
-		/* Enable DSC hw block */
-		dsc_cfg.pic_width = (stream->timing.h_addressable + stream->timing.h_border_left + stream->timing.h_border_right) / opp_cnt;
-		dsc_cfg.pic_height = stream->timing.v_addressable + stream->timing.v_border_top + stream->timing.v_border_bottom;
-		dsc_cfg.pixel_encoding = stream->timing.pixel_encoding;
-		dsc_cfg.color_depth = stream->timing.display_color_depth;
-		dsc_cfg.is_odm = pipe_ctx->next_odm_pipe ? true : false;
-		dsc_cfg.dc_dsc_cfg = stream->timing.dsc_cfg;
-		ASSERT(dsc_cfg.dc_dsc_cfg.num_slices_h % opp_cnt == 0);
-		dsc_cfg.dc_dsc_cfg.num_slices_h /= opp_cnt;
-
-		dsc->funcs->dsc_set_config(dsc, &dsc_cfg, &dsc_optc_cfg);
-		dsc->funcs->dsc_enable(dsc, pipe_ctx->stream_res.opp->inst);
-		for (odm_pipe = pipe_ctx->next_odm_pipe; odm_pipe; odm_pipe = odm_pipe->next_odm_pipe) {
-			struct display_stream_compressor *odm_dsc = odm_pipe->stream_res.dsc;
-
-			odm_dsc->funcs->dsc_set_config(odm_dsc, &dsc_cfg, &dsc_optc_cfg);
-			odm_dsc->funcs->dsc_enable(odm_dsc, odm_pipe->stream_res.opp->inst);
-		}
-		dsc_cfg.dc_dsc_cfg.num_slices_h *= opp_cnt;
-		dsc_cfg.pic_width *= opp_cnt;
-
-		optc_dsc_mode = dsc_optc_cfg.is_pixel_format_444 ? OPTC_DSC_ENABLED_444 : OPTC_DSC_ENABLED_NATIVE_SUBSAMPLED;
-
-		/* Enable DSC in encoder */
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-		if (dc_is_dp_signal(stream->signal) && !IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)
-				&& !is_dp_128b_132b_signal(pipe_ctx)) {
-#else
-		if (dc_is_dp_signal(stream->signal) && !IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) {
-#endif
-			DC_LOG_DSC("Setting stream encoder DSC config for engine %d:", (int)pipe_ctx->stream_res.stream_enc->id);
-			dsc_optc_config_log(dsc, &dsc_optc_cfg);
-			pipe_ctx->stream_res.stream_enc->funcs->dp_set_dsc_config(pipe_ctx->stream_res.stream_enc,
-									optc_dsc_mode,
-									dsc_optc_cfg.bytes_per_pixel,
-									dsc_optc_cfg.slice_width);
-
-			/* PPS SDP is set elsewhere because it has to be done after DIG FE is connected to DIG BE */
-		}
-
-		/* Enable DSC in OPTC */
-		DC_LOG_DSC("Setting optc DSC config for tg instance %d:", pipe_ctx->stream_res.tg->inst);
-		dsc_optc_config_log(dsc, &dsc_optc_cfg);
-		pipe_ctx->stream_res.tg->funcs->set_dsc_config(pipe_ctx->stream_res.tg,
-							optc_dsc_mode,
-							dsc_optc_cfg.bytes_per_pixel,
-							dsc_optc_cfg.slice_width);
-	} else {
-		/* disable DSC in OPTC */
-		pipe_ctx->stream_res.tg->funcs->set_dsc_config(
-				pipe_ctx->stream_res.tg,
-				OPTC_DSC_DISABLED, 0, 0);
-
-		/* disable DSC in stream encoder */
-		if (dc_is_dp_signal(stream->signal)) {
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-			if (is_dp_128b_132b_signal(pipe_ctx))
-				pipe_ctx->stream_res.hpo_dp_stream_enc->funcs->dp_set_dsc_pps_info_packet(
-										pipe_ctx->stream_res.hpo_dp_stream_enc,
-										false,
-										NULL,
-										true);
-			else
-#endif
-				if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) {
-					pipe_ctx->stream_res.stream_enc->funcs->dp_set_dsc_config(
-							pipe_ctx->stream_res.stream_enc,
-							OPTC_DSC_DISABLED, 0, 0);
-					pipe_ctx->stream_res.stream_enc->funcs->dp_set_dsc_pps_info_packet(
-								pipe_ctx->stream_res.stream_enc, false, NULL, true);
-				}
-		}
-
-		/* disable DSC block */
-		pipe_ctx->stream_res.dsc->funcs->dsc_disable(pipe_ctx->stream_res.dsc);
-		for (odm_pipe = pipe_ctx->next_odm_pipe; odm_pipe; odm_pipe = odm_pipe->next_odm_pipe)
-			odm_pipe->stream_res.dsc->funcs->dsc_disable(odm_pipe->stream_res.dsc);
-	}
-}
-
-bool dp_set_dsc_enable(struct pipe_ctx *pipe_ctx, bool enable)
-{
-	struct display_stream_compressor *dsc = pipe_ctx->stream_res.dsc;
-	bool result = false;
-
-	if (!pipe_ctx->stream->timing.flags.DSC)
-		goto out;
-	if (!dsc)
-		goto out;
-
-	if (enable) {
-		{
-			dp_set_dsc_on_stream(pipe_ctx, true);
-			result = true;
-		}
-	} else {
-		dp_set_dsc_on_rx(pipe_ctx, false);
-		dp_set_dsc_on_stream(pipe_ctx, false);
-		result = true;
-	}
-out:
-	return result;
-}
-
-/*
- * For dynamic bpp change case, dsc is programmed with MASTER_UPDATE_LOCK enabled;
- * hence PPS info packet update need to use frame update instead of immediate update.
- * Added parameter immediate_update for this purpose.
- * The decision to use frame update is hard-coded in function dp_update_dsc_config(),
- * which is the only place where a "false" would be passed in for param immediate_update.
- *
- * immediate_update is only applicable when DSC is enabled.
- */
-bool dp_set_dsc_pps_sdp(struct pipe_ctx *pipe_ctx, bool enable, bool immediate_update)
-{
-	struct display_stream_compressor *dsc = pipe_ctx->stream_res.dsc;
-	struct dc_stream_state *stream = pipe_ctx->stream;
-
-	if (!pipe_ctx->stream->timing.flags.DSC || !dsc)
-		return false;
-
-	if (enable) {
-		struct dsc_config dsc_cfg;
-		uint8_t dsc_packed_pps[128];
-
-		memset(&dsc_cfg, 0, sizeof(dsc_cfg));
-		memset(dsc_packed_pps, 0, 128);
-
-		/* Enable DSC hw block */
-		dsc_cfg.pic_width = stream->timing.h_addressable + stream->timing.h_border_left + stream->timing.h_border_right;
-		dsc_cfg.pic_height = stream->timing.v_addressable + stream->timing.v_border_top + stream->timing.v_border_bottom;
-		dsc_cfg.pixel_encoding = stream->timing.pixel_encoding;
-		dsc_cfg.color_depth = stream->timing.display_color_depth;
-		dsc_cfg.is_odm = pipe_ctx->next_odm_pipe ? true : false;
-		dsc_cfg.dc_dsc_cfg = stream->timing.dsc_cfg;
-
-		DC_LOG_DSC(" ");
-		dsc->funcs->dsc_get_packed_pps(dsc, &dsc_cfg, &dsc_packed_pps[0]);
-		if (dc_is_dp_signal(stream->signal)) {
-			DC_LOG_DSC("Setting stream encoder DSC PPS SDP for engine %d\n", (int)pipe_ctx->stream_res.stream_enc->id);
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-			if (is_dp_128b_132b_signal(pipe_ctx))
-				pipe_ctx->stream_res.hpo_dp_stream_enc->funcs->dp_set_dsc_pps_info_packet(
-										pipe_ctx->stream_res.hpo_dp_stream_enc,
-										true,
-										&dsc_packed_pps[0],
-										immediate_update);
-			else
-#endif
-				pipe_ctx->stream_res.stream_enc->funcs->dp_set_dsc_pps_info_packet(
-										pipe_ctx->stream_res.stream_enc,
-										true,
-										&dsc_packed_pps[0],
-										immediate_update);
-		}
-	} else {
-		/* disable DSC PPS in stream encoder */
-		if (dc_is_dp_signal(stream->signal)) {
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-			if (is_dp_128b_132b_signal(pipe_ctx))
-				pipe_ctx->stream_res.hpo_dp_stream_enc->funcs->dp_set_dsc_pps_info_packet(
-										pipe_ctx->stream_res.hpo_dp_stream_enc,
-										false,
-										NULL,
-										true);
-			else
-#endif
-				pipe_ctx->stream_res.stream_enc->funcs->dp_set_dsc_pps_info_packet(
-							pipe_ctx->stream_res.stream_enc, false, NULL, true);
-		}
-	}
-
-	return true;
-}
-
-
-bool dp_update_dsc_config(struct pipe_ctx *pipe_ctx)
-{
-	struct display_stream_compressor *dsc = pipe_ctx->stream_res.dsc;
-
-	if (!pipe_ctx->stream->timing.flags.DSC)
-		return false;
-	if (!dsc)
-		return false;
-
-	dp_set_dsc_on_stream(pipe_ctx, true);
-	dp_set_dsc_pps_sdp(pipe_ctx, true, false);
-	return true;
-}
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-#undef DC_LOGGER
-#define DC_LOGGER \
-	link->ctx->logger
-
-static enum phyd32clk_clock_source get_phyd32clk_src(struct dc_link *link)
-{
-	switch (link->link_enc->transmitter) {
-	case TRANSMITTER_UNIPHY_A:
-		return PHYD32CLKA;
-	case TRANSMITTER_UNIPHY_B:
-		return PHYD32CLKB;
-	case TRANSMITTER_UNIPHY_C:
-		return PHYD32CLKC;
-	case TRANSMITTER_UNIPHY_D:
-		return PHYD32CLKD;
-	case TRANSMITTER_UNIPHY_E:
-		return PHYD32CLKE;
-	default:
-		return PHYD32CLKA;
-	}
-}
-
-void enable_dp_hpo_output(struct dc_link *link, const struct dc_link_settings *link_settings)
-{
-	const struct dc *dc = link->dc;
-	enum phyd32clk_clock_source phyd32clk;
-
-	/* Enable PHY PLL at target bit rate
-	 *   UHBR10 = 10Gbps (SYMCLK32 = 312.5MHz)
-	 *   UBR13.5 = 13.5Gbps (SYMCLK32 = 421.875MHz)
-	 *   UHBR20 = 20Gbps (SYMCLK32 = 625MHz)
-	 */
-	if (IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) {
-		switch (link_settings->link_rate) {
-		case LINK_RATE_UHBR10:
-			dm_set_phyd32clk(dc->ctx, 312500);
-			break;
-		case LINK_RATE_UHBR13_5:
-			dm_set_phyd32clk(dc->ctx, 412875);
-			break;
-		case LINK_RATE_UHBR20:
-			dm_set_phyd32clk(dc->ctx, 625000);
-			break;
-		default:
-			return;
-		}
-	} else {
-		/* DP2.0 HW: call transmitter control to enable PHY */
-		link->hpo_dp_link_enc->funcs->enable_link_phy(
-				link->hpo_dp_link_enc,
-				link_settings,
-				link->link_enc->transmitter);
-	}
-
-	/* DCCG muxing and DTBCLK DTO */
-	if (IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) {
-		dc->res_pool->dccg->funcs->set_physymclk(
-				dc->res_pool->dccg,
-				link->link_enc_hw_inst,
-				PHYSYMCLK_FORCE_SRC_PHYD32CLK,
-				true);
-
-		phyd32clk = get_phyd32clk_src(link);
-		dc->res_pool->dccg->funcs->enable_symclk32_le(
-				dc->res_pool->dccg,
-				link->hpo_dp_link_enc->inst,
-				phyd32clk);
-		link->hpo_dp_link_enc->funcs->link_enable(
-					link->hpo_dp_link_enc,
-					link_settings->lane_count);
-	}
-}
-
-void disable_dp_hpo_output(struct dc_link *link, enum signal_type signal)
-{
-	const struct dc *dc = link->dc;
-
-	link->hpo_dp_link_enc->funcs->link_disable(link->hpo_dp_link_enc);
-
-	if (IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) {
-		dc->res_pool->dccg->funcs->disable_symclk32_le(
-					dc->res_pool->dccg,
-					link->hpo_dp_link_enc->inst);
-
-		dc->res_pool->dccg->funcs->set_physymclk(
-					dc->res_pool->dccg,
-					link->link_enc_hw_inst,
-					PHYSYMCLK_FORCE_SRC_SYMCLK,
-					false);
-
-		dm_set_phyd32clk(dc->ctx, 0);
-	} else {
-		/* DP2.0 HW: call transmitter control to disable PHY */
-		link->hpo_dp_link_enc->funcs->disable_link_phy(
-				link->hpo_dp_link_enc,
-				signal);
-	}
-}
-
-void setup_dp_hpo_stream(struct pipe_ctx *pipe_ctx, bool enable)
-{
-	struct dc_stream_state *stream = pipe_ctx->stream;
-	struct dc *dc = pipe_ctx->stream->ctx->dc;
-	struct pipe_ctx *odm_pipe;
-	int odm_combine_num_segments = 1;
-	enum phyd32clk_clock_source phyd32clk;
-
-	if (enable) {
-		for (odm_pipe = pipe_ctx->next_odm_pipe; odm_pipe; odm_pipe = odm_pipe->next_odm_pipe)
-				odm_combine_num_segments++;
-
-		dc->res_pool->dccg->funcs->set_dpstreamclk(
-				dc->res_pool->dccg,
-				DTBCLK0,
-				pipe_ctx->stream_res.tg->inst);
-
-		phyd32clk = get_phyd32clk_src(stream->link);
-		dc->res_pool->dccg->funcs->enable_symclk32_se(
-				dc->res_pool->dccg,
-				pipe_ctx->stream_res.hpo_dp_stream_enc->inst,
-				phyd32clk);
-
-		dc->res_pool->dccg->funcs->set_dtbclk_dto(
-				dc->res_pool->dccg,
-				pipe_ctx->stream_res.tg->inst,
-				stream->phy_pix_clk,
-				odm_combine_num_segments,
-				&stream->timing);
-	} else {
-		dc->res_pool->dccg->funcs->set_dtbclk_dto(
-				dc->res_pool->dccg,
-				pipe_ctx->stream_res.tg->inst,
-				0,
-				0,
-				&stream->timing);
-		dc->res_pool->dccg->funcs->disable_symclk32_se(
-				dc->res_pool->dccg,
-				pipe_ctx->stream_res.hpo_dp_stream_enc->inst);
-		dc->res_pool->dccg->funcs->set_dpstreamclk(
-				dc->res_pool->dccg,
-				REFCLK,
-				pipe_ctx->stream_res.tg->inst);
-	}
-}
-
-void reset_dp_hpo_stream_encoders_for_link(struct dc_link *link)
-{
-	const struct dc *dc = link->dc;
-	struct dc_state *state = dc->current_state;
-	uint8_t i;
-
-	for (i = 0; i < MAX_PIPES; i++) {
-		if (state->res_ctx.pipe_ctx[i].stream_res.hpo_dp_stream_enc &&
-				state->res_ctx.pipe_ctx[i].stream &&
-				state->res_ctx.pipe_ctx[i].stream->link == link &&
-				!state->res_ctx.pipe_ctx[i].stream->dpms_off) {
-			setup_dp_hpo_stream(&state->res_ctx.pipe_ctx[i], false);
-		}
-	}
-}
-
-#undef DC_LOGGER
-#endif
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_resource.c b/drivers/gpu/drm/amd/display/dc/core/dc_resource.c
index e2d9a46d0e1a..848c267ef11e 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_resource.c
@@ -23,8 +23,6 @@
  *
  */
 
-#include <linux/slab.h>
-
 #include "dm_services.h"
 
 #include "resource.h"
@@ -42,7 +40,17 @@
 #include "virtual/virtual_stream_encoder.h"
 #include "dpcd_defs.h"
 #include "link_enc_cfg.h"
-#include "dc_link_dp.h"
+#include "link_service.h"
+#include "clk_mgr.h"
+#include "dc_state_priv.h"
+#include "dc_stream_priv.h"
+
+#include "virtual/virtual_link_hwss.h"
+#include "link/hwss/link_hwss_dio.h"
+#include "link/hwss/link_hwss_dpia.h"
+#include "link/hwss/link_hwss_hpo_dp.h"
+#include "link/hwss/link_hwss_dio_fixed_vs_pe_retimer.h"
+#include "link/hwss/link_hwss_hpo_fixed_vs_pe_retimer_dp.h"
 
 #if defined(CONFIG_DRM_AMD_DC_SI)
 #include "dce60/dce60_resource.h"
@@ -52,7 +60,6 @@
 #include "dce110/dce110_resource.h"
 #include "dce112/dce112_resource.h"
 #include "dce120/dce120_resource.h"
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 #include "dcn10/dcn10_resource.h"
 #include "dcn20/dcn20_resource.h"
 #include "dcn21/dcn21_resource.h"
@@ -62,13 +69,74 @@
 #include "dcn302/dcn302_resource.h"
 #include "dcn303/dcn303_resource.h"
 #include "dcn31/dcn31_resource.h"
+#include "dcn314/dcn314_resource.h"
+#include "dcn315/dcn315_resource.h"
+#include "dcn316/dcn316_resource.h"
+#include "dcn32/dcn32_resource.h"
+#include "dcn321/dcn321_resource.h"
+#include "dcn35/dcn35_resource.h"
+#include "dcn351/dcn351_resource.h"
+#include "dcn36/dcn36_resource.h"
+#include "dcn401/dcn401_resource.h"
+#if defined(CONFIG_DRM_AMD_DC_FP)
+#include "dc_spl_translate.h"
 #endif
 
+#define VISUAL_CONFIRM_BASE_DEFAULT 3
+#define VISUAL_CONFIRM_BASE_MIN 1
+#define VISUAL_CONFIRM_BASE_MAX 10
+/* we choose 240 because it is a common denominator of common v addressable
+ * such as 2160, 1440, 1200, 960. So we take 1/240 portion of v addressable as
+ * the visual confirm dpp offset height. So visual confirm height can stay
+ * relatively the same independent from timing used.
+ */
+#define VISUAL_CONFIRM_DPP_OFFSET_DENO 240
+
+#define DC_LOGGER \
+	dc->ctx->logger
 #define DC_LOGGER_INIT(logger)
+#include "dml2_0/dml2_wrapper.h"
+
+#define UNABLE_TO_SPLIT -1
+
+static void capture_pipe_topology_data(struct dc *dc, int plane_idx, int slice_idx, int stream_idx,
+									   int dpp_inst, int opp_inst, int tg_inst, bool is_phantom_pipe)
+{
+	struct pipe_topology_snapshot *current_snapshot = &dc->debug_data.topology_history.snapshots[dc->debug_data.topology_history.current_snapshot_index];
+
+	if (current_snapshot->line_count >= MAX_PIPES)
+		return;
+
+	current_snapshot->pipe_log_lines[current_snapshot->line_count].is_phantom_pipe = is_phantom_pipe;
+	current_snapshot->pipe_log_lines[current_snapshot->line_count].plane_idx = plane_idx;
+	current_snapshot->pipe_log_lines[current_snapshot->line_count].slice_idx = slice_idx;
+	current_snapshot->pipe_log_lines[current_snapshot->line_count].stream_idx = stream_idx;
+	current_snapshot->pipe_log_lines[current_snapshot->line_count].dpp_inst = dpp_inst;
+	current_snapshot->pipe_log_lines[current_snapshot->line_count].opp_inst = opp_inst;
+	current_snapshot->pipe_log_lines[current_snapshot->line_count].tg_inst = tg_inst;
+
+	current_snapshot->line_count++;
+}
+
+static void start_new_topology_snapshot(struct dc *dc, struct dc_state *state)
+{
+	// Move to next snapshot slot (circular buffer)
+	dc->debug_data.topology_history.current_snapshot_index = (dc->debug_data.topology_history.current_snapshot_index + 1) % MAX_TOPOLOGY_SNAPSHOTS;
+
+	// Clear the new snapshot
+	struct pipe_topology_snapshot *current_snapshot = &dc->debug_data.topology_history.snapshots[dc->debug_data.topology_history.current_snapshot_index];
+	memset(current_snapshot, 0, sizeof(*current_snapshot));
+
+	// Set metadata
+	current_snapshot->timestamp_us = dm_get_timestamp(dc->ctx);
+	current_snapshot->stream_count = state->stream_count;
+	current_snapshot->phantom_stream_count = state->phantom_stream_count;
+}
 
 enum dce_version resource_parse_asic_id(struct hw_asic_id asic_id)
 {
 	enum dce_version dc_version = DCE_VERSION_UNKNOWN;
+
 	switch (asic_id.chip_family) {
 
 #if defined(CONFIG_DRM_AMD_DC_SI)
@@ -118,7 +186,6 @@ enum dce_version resource_parse_asic_id(struct hw_asic_id asic_id)
 		else
 			dc_version = DCE_VERSION_12_0;
 		break;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	case FAMILY_RV:
 		dc_version = DCN_VERSION_1_0;
 		if (ASICREV_IS_RAVEN2(asic_id.hw_internal_rev))
@@ -131,7 +198,13 @@ enum dce_version resource_parse_asic_id(struct hw_asic_id asic_id)
 
 	case FAMILY_NV:
 		dc_version = DCN_VERSION_2_0;
-		if (asic_id.chip_id == DEVICE_ID_NV_13FE) {
+		if (asic_id.chip_id == DEVICE_ID_NV_13FE ||
+		    asic_id.chip_id == DEVICE_ID_NV_143F ||
+		    asic_id.chip_id == DEVICE_ID_NV_13F9 ||
+		    asic_id.chip_id == DEVICE_ID_NV_13FA ||
+		    asic_id.chip_id == DEVICE_ID_NV_13FB ||
+		    asic_id.chip_id == DEVICE_ID_NV_13FC ||
+		    asic_id.chip_id == DEVICE_ID_NV_13DB) {
 			dc_version = DCN_VERSION_2_01;
 			break;
 		}
@@ -151,8 +224,34 @@ enum dce_version resource_parse_asic_id(struct hw_asic_id asic_id)
 		if (ASICREV_IS_YELLOW_CARP(asic_id.hw_internal_rev))
 			dc_version = DCN_VERSION_3_1;
 		break;
-#endif
-
+	case AMDGPU_FAMILY_GC_10_3_6:
+		if (ASICREV_IS_GC_10_3_6(asic_id.hw_internal_rev))
+			dc_version = DCN_VERSION_3_15;
+		break;
+	case AMDGPU_FAMILY_GC_10_3_7:
+		if (ASICREV_IS_GC_10_3_7(asic_id.hw_internal_rev))
+			dc_version = DCN_VERSION_3_16;
+		break;
+	case AMDGPU_FAMILY_GC_11_0_0:
+		dc_version = DCN_VERSION_3_2;
+		if (ASICREV_IS_GC_11_0_2(asic_id.hw_internal_rev))
+			dc_version = DCN_VERSION_3_21;
+		break;
+	case AMDGPU_FAMILY_GC_11_0_1:
+		dc_version = DCN_VERSION_3_14;
+		break;
+	case AMDGPU_FAMILY_GC_11_5_0:
+		dc_version = DCN_VERSION_3_5;
+		if (ASICREV_IS_GC_11_0_4(asic_id.hw_internal_rev))
+			dc_version = DCN_VERSION_3_51;
+		if (ASICREV_IS_DCN36(asic_id.hw_internal_rev))
+			dc_version = DCN_VERSION_3_6;
+		break;
+	case AMDGPU_FAMILY_GC_12_0_0:
+		if (ASICREV_IS_GC_12_0_1_A0(asic_id.hw_internal_rev) ||
+			ASICREV_IS_GC_12_0_0_A0(asic_id.hw_internal_rev))
+			dc_version = DCN_VERSION_4_01;
+		break;
 	default:
 		dc_version = DCE_VERSION_UNKNOWN;
 		break;
@@ -213,7 +312,7 @@ struct resource_pool *dc_create_resource_pool(struct dc  *dc,
 				init_data->num_virtual_links, dc);
 		break;
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
+#if defined(CONFIG_DRM_AMD_DC_FP)
 	case DCN_VERSION_1_0:
 	case DCN_VERSION_1_01:
 		res_pool = dcn10_create_resource_pool(init_data, dc);
@@ -242,7 +341,34 @@ struct resource_pool *dc_create_resource_pool(struct dc  *dc,
 	case DCN_VERSION_3_1:
 		res_pool = dcn31_create_resource_pool(init_data, dc);
 		break;
-#endif
+	case DCN_VERSION_3_14:
+		res_pool = dcn314_create_resource_pool(init_data, dc);
+		break;
+	case DCN_VERSION_3_15:
+		res_pool = dcn315_create_resource_pool(init_data, dc);
+		break;
+	case DCN_VERSION_3_16:
+		res_pool = dcn316_create_resource_pool(init_data, dc);
+		break;
+	case DCN_VERSION_3_2:
+		res_pool = dcn32_create_resource_pool(init_data, dc);
+		break;
+	case DCN_VERSION_3_21:
+		res_pool = dcn321_create_resource_pool(init_data, dc);
+		break;
+	case DCN_VERSION_3_5:
+		res_pool = dcn35_create_resource_pool(init_data, dc);
+		break;
+	case DCN_VERSION_3_51:
+		res_pool = dcn351_create_resource_pool(init_data, dc);
+		break;
+	case DCN_VERSION_3_6:
+		res_pool = dcn36_create_resource_pool(init_data, dc);
+		break;
+	case DCN_VERSION_4_01:
+		res_pool = dcn401_create_resource_pool(init_data, dc);
+		break;
+#endif /* CONFIG_DRM_AMD_DC_FP */
 	default:
 		break;
 	}
@@ -268,7 +394,7 @@ struct resource_pool *dc_create_resource_pool(struct dc  *dc,
 	return res_pool;
 }
 
-void dc_destroy_resource_pool(struct dc  *dc)
+void dc_destroy_resource_pool(struct dc *dc)
 {
 	if (dc) {
 		if (dc->res_pool)
@@ -354,9 +480,16 @@ bool resource_construct(
 				DC_ERR("DC: failed to create stream_encoder!\n");
 			pool->stream_enc_count++;
 		}
+
+		for (i = 0; i < caps->num_analog_stream_encoder; i++) {
+			pool->stream_enc[caps->num_stream_encoder + i] =
+				create_funcs->create_stream_encoder(ENGINE_ID_DACA + i, ctx);
+			if (pool->stream_enc[caps->num_stream_encoder + i] == NULL)
+				DC_ERR("DC: failed to create analog stream_encoder %d!\n", i);
+			pool->stream_enc_count++;
+		}
 	}
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	pool->hpo_dp_stream_enc_count = 0;
 	if (create_funcs->create_hpo_dp_stream_encoder) {
 		for (i = 0; i < caps->num_hpo_dp_stream_encoder; i++) {
@@ -377,9 +510,7 @@ bool resource_construct(
 			pool->hpo_dp_link_enc_count++;
 		}
 	}
-#endif
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	for (i = 0; i < caps->num_mpc_3dlut; i++) {
 		pool->mpc_lut[i] = dc_create_3dlut_func();
 		if (pool->mpc_lut[i] == NULL)
@@ -388,7 +519,7 @@ bool resource_construct(
 		if (pool->mpc_shaper[i] == NULL)
 			DC_ERR("DC: failed to create MPC shaper!\n");
 	}
-#endif
+
 	dc->caps.dynamic_audio = false;
 	if (pool->audio_count < pool->stream_enc_count) {
 		dc->caps.dynamic_audio = true;
@@ -687,155 +818,363 @@ static inline void get_vp_scan_direction(
 		*flip_horz_scan_dir = !*flip_horz_scan_dir;
 }
 
-int get_num_mpc_splits(struct pipe_ctx *pipe)
+static struct rect intersect_rec(const struct rect *r0, const struct rect *r1)
+{
+	struct rect rec;
+	int r0_x_end = r0->x + r0->width;
+	int r1_x_end = r1->x + r1->width;
+	int r0_y_end = r0->y + r0->height;
+	int r1_y_end = r1->y + r1->height;
+
+	rec.x = r0->x > r1->x ? r0->x : r1->x;
+	rec.width = r0_x_end > r1_x_end ? r1_x_end - rec.x : r0_x_end - rec.x;
+	rec.y = r0->y > r1->y ? r0->y : r1->y;
+	rec.height = r0_y_end > r1_y_end ? r1_y_end - rec.y : r0_y_end - rec.y;
+
+	/* in case that there is no intersection */
+	if (rec.width < 0 || rec.height < 0)
+		memset(&rec, 0, sizeof(rec));
+
+	return rec;
+}
+
+static struct rect shift_rec(const struct rect *rec_in, int x, int y)
 {
-	int mpc_split_count = 0;
-	struct pipe_ctx *other_pipe = pipe->bottom_pipe;
+	struct rect rec_out = *rec_in;
 
-	while (other_pipe && other_pipe->plane_state == pipe->plane_state) {
-		mpc_split_count++;
-		other_pipe = other_pipe->bottom_pipe;
-	}
-	other_pipe = pipe->top_pipe;
-	while (other_pipe && other_pipe->plane_state == pipe->plane_state) {
-		mpc_split_count++;
-		other_pipe = other_pipe->top_pipe;
-	}
+	rec_out.x += x;
+	rec_out.y += y;
 
-	return mpc_split_count;
+	return rec_out;
 }
 
-int get_num_odm_splits(struct pipe_ctx *pipe)
+static struct rect calculate_plane_rec_in_timing_active(
+		struct pipe_ctx *pipe_ctx,
+		const struct rect *rec_in)
 {
-	int odm_split_count = 0;
-	struct pipe_ctx *next_pipe = pipe->next_odm_pipe;
-	while (next_pipe) {
-		odm_split_count++;
-		next_pipe = next_pipe->next_odm_pipe;
+	/*
+	 * The following diagram shows an example where we map a 1920x1200
+	 * desktop to a 2560x1440 timing with a plane rect in the middle
+	 * of the screen. To map a plane rect from Stream Source to Timing
+	 * Active space, we first multiply stream scaling ratios (i.e 2304/1920
+	 * horizontal and 1440/1200 vertical) to the plane's x and y, then
+	 * we add stream destination offsets (i.e 128 horizontal, 0 vertical).
+	 * This will give us a plane rect's position in Timing Active. However
+	 * we have to remove the fractional. The rule is that we find left/right
+	 * and top/bottom positions and round the value to the adjacent integer.
+	 *
+	 * Stream Source Space
+	 * ------------
+	 *        __________________________________________________
+	 *       |Stream Source (1920 x 1200) ^                     |
+	 *       |                            y                     |
+	 *       |         <------- w --------|>                    |
+	 *       |          __________________V                     |
+	 *       |<-- x -->|Plane//////////////| ^                  |
+	 *       |         |(pre scale)////////| |                  |
+	 *       |         |///////////////////| |                  |
+	 *       |         |///////////////////| h                  |
+	 *       |         |///////////////////| |                  |
+	 *       |         |///////////////////| |                  |
+	 *       |         |///////////////////| V                  |
+	 *       |                                                  |
+	 *       |                                                  |
+	 *       |__________________________________________________|
+	 *
+	 *
+	 * Timing Active Space
+	 * ---------------------------------
+	 *
+	 *       Timing Active (2560 x 1440)
+	 *        __________________________________________________
+	 *       |*****|  Stteam Destination (2304 x 1440)    |*****|
+	 *       |*****|                                      |*****|
+	 *       |<128>|                                      |*****|
+	 *       |*****|     __________________               |*****|
+	 *       |*****|    |Plane/////////////|              |*****|
+	 *       |*****|    |(post scale)//////|              |*****|
+	 *       |*****|    |//////////////////|              |*****|
+	 *       |*****|    |//////////////////|              |*****|
+	 *       |*****|    |//////////////////|              |*****|
+	 *       |*****|    |//////////////////|              |*****|
+	 *       |*****|                                      |*****|
+	 *       |*****|                                      |*****|
+	 *       |*****|                                      |*****|
+	 *       |*****|______________________________________|*****|
+	 *
+	 * So the resulting formulas are shown below:
+	 *
+	 * recout_x = 128 + round(plane_x * 2304 / 1920)
+	 * recout_w = 128 + round((plane_x + plane_w) * 2304 / 1920) - recout_x
+	 * recout_y = 0 + round(plane_y * 1440 / 1280)
+	 * recout_h = 0 + round((plane_y + plane_h) * 1440 / 1200) - recout_y
+	 *
+	 * NOTE: fixed point division is not error free. To reduce errors
+	 * introduced by fixed point division, we divide only after
+	 * multiplication is complete.
+	 */
+	const struct dc_stream_state *stream = pipe_ctx->stream;
+	struct rect rec_out = {0};
+	struct fixed31_32 temp;
+
+	temp = dc_fixpt_from_fraction(rec_in->x * (long long)stream->dst.width,
+			stream->src.width);
+	rec_out.x = stream->dst.x + dc_fixpt_round(temp);
+
+	temp = dc_fixpt_from_fraction(
+			(rec_in->x + rec_in->width) * (long long)stream->dst.width,
+			stream->src.width);
+	rec_out.width = stream->dst.x + dc_fixpt_round(temp) - rec_out.x;
+
+	temp = dc_fixpt_from_fraction(rec_in->y * (long long)stream->dst.height,
+			stream->src.height);
+	rec_out.y = stream->dst.y + dc_fixpt_round(temp);
+
+	temp = dc_fixpt_from_fraction(
+			(rec_in->y + rec_in->height) * (long long)stream->dst.height,
+			stream->src.height);
+	rec_out.height = stream->dst.y + dc_fixpt_round(temp) - rec_out.y;
+
+	return rec_out;
+}
+
+static struct rect calculate_mpc_slice_in_timing_active(
+		struct pipe_ctx *pipe_ctx,
+		struct rect *plane_clip_rec)
+{
+	const struct dc_stream_state *stream = pipe_ctx->stream;
+	int mpc_slice_count = resource_get_mpc_slice_count(pipe_ctx);
+	int mpc_slice_idx = resource_get_mpc_slice_index(pipe_ctx);
+	int epimo = mpc_slice_count - plane_clip_rec->width % mpc_slice_count - 1;
+	struct rect mpc_rec;
+
+	mpc_rec.width = plane_clip_rec->width / mpc_slice_count;
+	mpc_rec.x = plane_clip_rec->x + mpc_rec.width * mpc_slice_idx;
+	mpc_rec.height = plane_clip_rec->height;
+	mpc_rec.y = plane_clip_rec->y;
+	ASSERT(mpc_slice_count == 1 ||
+			stream->view_format != VIEW_3D_FORMAT_SIDE_BY_SIDE ||
+			mpc_rec.width % 2 == 0);
+
+	if (stream->view_format == VIEW_3D_FORMAT_SIDE_BY_SIDE)
+		mpc_rec.x -= (mpc_rec.width * mpc_slice_idx);
+
+	/* extra pixels in the division remainder need to go to pipes after
+	 * the extra pixel index minus one(epimo) defined here as:
+	 */
+	if (mpc_slice_idx > epimo) {
+		mpc_rec.x += mpc_slice_idx - epimo - 1;
+		mpc_rec.width += 1;
 	}
-	pipe = pipe->prev_odm_pipe;
-	while (pipe) {
-		odm_split_count++;
-		pipe = pipe->prev_odm_pipe;
+
+	if (stream->view_format == VIEW_3D_FORMAT_TOP_AND_BOTTOM) {
+		ASSERT(mpc_rec.height % 2 == 0);
+		mpc_rec.height /= 2;
 	}
-	return odm_split_count;
+	return mpc_rec;
 }
 
-static void calculate_split_count_and_index(struct pipe_ctx *pipe_ctx, int *split_count, int *split_idx)
+static void calculate_adjust_recout_for_visual_confirm(struct pipe_ctx *pipe_ctx,
+	int *base_offset, int *dpp_offset)
 {
-	*split_count = get_num_odm_splits(pipe_ctx);
-	*split_idx = 0;
-	if (*split_count == 0) {
-		/*Check for mpc split*/
-		struct pipe_ctx *split_pipe = pipe_ctx->top_pipe;
+	struct dc *dc = pipe_ctx->stream->ctx->dc;
+	*base_offset = 0;
+	*dpp_offset = 0;
 
-		*split_count = get_num_mpc_splits(pipe_ctx);
-		while (split_pipe && split_pipe->plane_state == pipe_ctx->plane_state) {
-			(*split_idx)++;
-			split_pipe = split_pipe->top_pipe;
-		}
-	} else {
-		/*Get odm split index*/
-		struct pipe_ctx *split_pipe = pipe_ctx->prev_odm_pipe;
+	if (dc->debug.visual_confirm == VISUAL_CONFIRM_DISABLE || !pipe_ctx->plane_res.dpp)
+		return;
 
-		while (split_pipe) {
-			(*split_idx)++;
-			split_pipe = split_pipe->prev_odm_pipe;
-		}
-	}
+	*dpp_offset = pipe_ctx->stream->timing.v_addressable / VISUAL_CONFIRM_DPP_OFFSET_DENO;
+	*dpp_offset *= pipe_ctx->plane_res.dpp->inst;
+
+	if ((dc->debug.visual_confirm_rect_height >= VISUAL_CONFIRM_BASE_MIN) &&
+			dc->debug.visual_confirm_rect_height <= VISUAL_CONFIRM_BASE_MAX)
+		*base_offset = dc->debug.visual_confirm_rect_height;
+	else
+		*base_offset = VISUAL_CONFIRM_BASE_DEFAULT;
 }
 
-/*
- * This is a preliminary vp size calculation to allow us to check taps support.
- * The result is completely overridden afterwards.
- */
-static void calculate_viewport_size(struct pipe_ctx *pipe_ctx)
+static void reverse_adjust_recout_for_visual_confirm(struct rect *recout,
+		struct pipe_ctx *pipe_ctx)
 {
-	struct scaler_data *data = &pipe_ctx->plane_res.scl_data;
+	int dpp_offset, base_offset;
 
-	data->viewport.width = dc_fixpt_ceil(dc_fixpt_mul_int(data->ratios.horz, data->recout.width));
-	data->viewport.height = dc_fixpt_ceil(dc_fixpt_mul_int(data->ratios.vert, data->recout.height));
-	data->viewport_c.width = dc_fixpt_ceil(dc_fixpt_mul_int(data->ratios.horz_c, data->recout.width));
-	data->viewport_c.height = dc_fixpt_ceil(dc_fixpt_mul_int(data->ratios.vert_c, data->recout.height));
-	if (pipe_ctx->plane_state->rotation == ROTATION_ANGLE_90 ||
-			pipe_ctx->plane_state->rotation == ROTATION_ANGLE_270) {
-		swap(data->viewport.width, data->viewport.height);
-		swap(data->viewport_c.width, data->viewport_c.height);
-	}
+	calculate_adjust_recout_for_visual_confirm(pipe_ctx, &base_offset,
+		&dpp_offset);
+	recout->height += base_offset;
+	recout->height += dpp_offset;
 }
 
-static void calculate_recout(struct pipe_ctx *pipe_ctx)
+static void adjust_recout_for_visual_confirm(struct rect *recout,
+		struct pipe_ctx *pipe_ctx)
 {
-	const struct dc_plane_state *plane_state = pipe_ctx->plane_state;
-	const struct dc_stream_state *stream = pipe_ctx->stream;
-	struct scaler_data *data = &pipe_ctx->plane_res.scl_data;
-	struct rect surf_clip = plane_state->clip_rect;
-	bool split_tb = stream->view_format == VIEW_3D_FORMAT_TOP_AND_BOTTOM;
-	int split_count, split_idx;
+	int dpp_offset, base_offset;
 
-	calculate_split_count_and_index(pipe_ctx, &split_count, &split_idx);
-	if (stream->view_format == VIEW_3D_FORMAT_SIDE_BY_SIDE)
-		split_idx = 0;
+	calculate_adjust_recout_for_visual_confirm(pipe_ctx, &base_offset,
+		&dpp_offset);
+	recout->height -= base_offset;
+	recout->height -= dpp_offset;
+}
 
+/*
+ * The function maps a plane clip from Stream Source Space to ODM Slice Space
+ * and calculates the rec of the overlapping area of MPC slice of the plane
+ * clip, ODM slice associated with the pipe context and stream destination rec.
+ */
+static void calculate_recout(struct pipe_ctx *pipe_ctx)
+{
 	/*
-	 * Only the leftmost ODM pipe should be offset by a nonzero distance
+	 * A plane clip represents the desired plane size and position in Stream
+	 * Source Space. Stream Source is the destination where all planes are
+	 * blended (i.e. positioned, scaled and overlaid). It is a canvas where
+	 * all planes associated with the current stream are drawn together.
+	 * After Stream Source is completed, we will further scale and
+	 * reposition the entire canvas of the stream source to Stream
+	 * Destination in Timing Active Space. This could be due to display
+	 * overscan adjustment where we will need to rescale and reposition all
+	 * the planes so they can fit into a TV with overscan or downscale
+	 * upscale features such as GPU scaling or VSR.
+	 *
+	 * This two step blending is a virtual procedure in software. In
+	 * hardware there is no such thing as Stream Source. all planes are
+	 * blended once in Timing Active Space. Software virtualizes a Stream
+	 * Source space to decouple the math complicity so scaling param
+	 * calculation focuses on one step at a time.
+	 *
+	 * In the following two diagrams, user applied 10% overscan adjustment
+	 * so the Stream Source needs to be scaled down a little before mapping
+	 * to Timing Active Space. As a result the Plane Clip is also scaled
+	 * down by the same ratio, Plane Clip position (i.e. x and y) with
+	 * respect to Stream Source is also scaled down. To map it in Timing
+	 * Active Space additional x and y offsets from Stream Destination are
+	 * added to Plane Clip as well.
+	 *
+	 * Stream Source Space
+	 * ------------
+	 *        __________________________________________________
+	 *       |Stream Source (3840 x 2160) ^                     |
+	 *       |                            y                     |
+	 *       |                            |                     |
+	 *       |          __________________V                     |
+	 *       |<-- x -->|Plane Clip/////////|                    |
+	 *       |         |(pre scale)////////|                    |
+	 *       |         |///////////////////|                    |
+	 *       |         |///////////////////|                    |
+	 *       |         |///////////////////|                    |
+	 *       |         |///////////////////|                    |
+	 *       |         |///////////////////|                    |
+	 *       |                                                  |
+	 *       |                                                  |
+	 *       |__________________________________________________|
+	 *
+	 *
+	 * Timing Active Space (3840 x 2160)
+	 * ---------------------------------
+	 *
+	 *       Timing Active
+	 *        __________________________________________________
+	 *       | y_____________________________________________   |
+	 *       |x |Stream Destination (3456 x 1944)            |  |
+	 *       |  |                                            |  |
+	 *       |  |        __________________                  |  |
+	 *       |  |       |Plane Clip////////|                 |  |
+	 *       |  |       |(post scale)//////|                 |  |
+	 *       |  |       |//////////////////|                 |  |
+	 *       |  |       |//////////////////|                 |  |
+	 *       |  |       |//////////////////|                 |  |
+	 *       |  |       |//////////////////|                 |  |
+	 *       |  |                                            |  |
+	 *       |  |                                            |  |
+	 *       |  |____________________________________________|  |
+	 *       |__________________________________________________|
+	 *
+	 *
+	 * In Timing Active Space a plane clip could be further sliced into
+	 * pieces called MPC slices. Each Pipe Context is responsible for
+	 * processing only one MPC slice so the plane processing workload can be
+	 * distributed to multiple DPP Pipes. MPC slices could be blended
+	 * together to a single ODM slice. Each ODM slice is responsible for
+	 * processing a portion of Timing Active divided horizontally so the
+	 * output pixel processing workload can be distributed to multiple OPP
+	 * pipes. All ODM slices are mapped together in ODM block so all MPC
+	 * slices belong to different ODM slices could be pieced together to
+	 * form a single image in Timing Active. MPC slices must belong to
+	 * single ODM slice. If an MPC slice goes across ODM slice boundary, it
+	 * needs to be divided into two MPC slices one for each ODM slice.
+	 *
+	 * In the following diagram the output pixel processing workload is
+	 * divided horizontally into two ODM slices one for each OPP blend tree.
+	 * OPP0 blend tree is responsible for processing left half of Timing
+	 * Active, while OPP2 blend tree is responsible for processing right
+	 * half.
+	 *
+	 * The plane has two MPC slices. However since the right MPC slice goes
+	 * across ODM boundary, two DPP pipes are needed one for each OPP blend
+	 * tree. (i.e. DPP1 for OPP0 blend tree and DPP2 for OPP2 blend tree).
+	 *
+	 * Assuming that we have a Pipe Context associated with OPP0 and DPP1
+	 * working on processing the plane in the diagram. We want to know the
+	 * width and height of the shaded rectangle and its relative position
+	 * with respect to the ODM slice0. This is called the recout of the pipe
+	 * context.
+	 *
+	 * Planes can be at arbitrary size and position and there could be an
+	 * arbitrary number of MPC and ODM slices. The algorithm needs to take
+	 * all scenarios into account.
+	 *
+	 * Timing Active Space (3840 x 2160)
+	 * ---------------------------------
+	 *
+	 *       Timing Active
+	 *        __________________________________________________
+	 *       |OPP0(ODM slice0)^        |OPP2(ODM slice1)        |
+	 *       |                y        |                        |
+	 *       |                |  <- w ->                        |
+	 *       |           _____V________|____                    |
+	 *       |          |DPP0 ^  |DPP1 |DPP2|                   |
+	 *       |<------ x |-----|->|/////|    |                   |
+	 *       |          |     |  |/////|    |                   |
+	 *       |          |     h  |/////|    |                   |
+	 *       |          |     |  |/////|    |                   |
+	 *       |          |_____V__|/////|____|                   |
+	 *       |                         |                        |
+	 *       |                         |                        |
+	 *       |                         |                        |
+	 *       |_________________________|________________________|
+	 *
+	 *
 	 */
-	if (!pipe_ctx->prev_odm_pipe || split_idx == split_count) {
-		data->recout.x = stream->dst.x;
-		if (stream->src.x < surf_clip.x)
-			data->recout.x += (surf_clip.x - stream->src.x) * stream->dst.width
-						/ stream->src.width;
-	} else
-		data->recout.x = 0;
-
-	if (stream->src.x > surf_clip.x)
-		surf_clip.width -= stream->src.x - surf_clip.x;
-	data->recout.width = surf_clip.width * stream->dst.width / stream->src.width;
-	if (data->recout.width + data->recout.x > stream->dst.x + stream->dst.width)
-		data->recout.width = stream->dst.x + stream->dst.width - data->recout.x;
-
-	data->recout.y = stream->dst.y;
-	if (stream->src.y < surf_clip.y)
-		data->recout.y += (surf_clip.y - stream->src.y) * stream->dst.height
-						/ stream->src.height;
-	else if (stream->src.y > surf_clip.y)
-		surf_clip.height -= stream->src.y - surf_clip.y;
-
-	data->recout.height = surf_clip.height * stream->dst.height / stream->src.height;
-	if (data->recout.height + data->recout.y > stream->dst.y + stream->dst.height)
-		data->recout.height = stream->dst.y + stream->dst.height - data->recout.y;
-
-	/* Handle h & v split */
-	if (split_tb) {
-		ASSERT(data->recout.height % 2 == 0);
-		data->recout.height /= 2;
-	} else if (split_count) {
-		if (!pipe_ctx->next_odm_pipe && !pipe_ctx->prev_odm_pipe) {
-			/* extra pixels in the division remainder need to go to pipes after
-			 * the extra pixel index minus one(epimo) defined here as:
-			 */
-			int epimo = split_count - data->recout.width % (split_count + 1);
-
-			data->recout.x += (data->recout.width / (split_count + 1)) * split_idx;
-			if (split_idx > epimo)
-				data->recout.x += split_idx - epimo - 1;
-			ASSERT(stream->view_format != VIEW_3D_FORMAT_SIDE_BY_SIDE || data->recout.width % 2 == 0);
-			data->recout.width = data->recout.width / (split_count + 1) + (split_idx > epimo ? 1 : 0);
-		} else {
-			/* odm */
-			if (split_idx == split_count) {
-				/* rightmost pipe is the remainder recout */
-				data->recout.width -= data->h_active * split_count - data->recout.x;
-
-				/* ODM combine cases with MPO we can get negative widths */
-				if (data->recout.width < 0)
-					data->recout.width = 0;
-
-				data->recout.x = 0;
-			} else
-				data->recout.width = data->h_active - data->recout.x;
-		}
+	struct rect plane_clip;
+	struct rect mpc_slice_of_plane_clip;
+	struct rect odm_slice_src;
+	struct rect overlapping_area;
+
+	plane_clip = calculate_plane_rec_in_timing_active(pipe_ctx,
+			&pipe_ctx->plane_state->clip_rect);
+	/* guard plane clip from drawing beyond stream dst here */
+	plane_clip = intersect_rec(&plane_clip,
+				&pipe_ctx->stream->dst);
+	mpc_slice_of_plane_clip = calculate_mpc_slice_in_timing_active(
+			pipe_ctx, &plane_clip);
+	odm_slice_src = resource_get_odm_slice_src_rect(pipe_ctx);
+	overlapping_area = intersect_rec(&mpc_slice_of_plane_clip, &odm_slice_src);
+	if (overlapping_area.height > 0 &&
+			overlapping_area.width > 0) {
+		/* shift the overlapping area so it is with respect to current
+		 * ODM slice source's position
+		 */
+		pipe_ctx->plane_res.scl_data.recout = shift_rec(
+				&overlapping_area,
+				-odm_slice_src.x, -odm_slice_src.y);
+		adjust_recout_for_visual_confirm(
+				&pipe_ctx->plane_res.scl_data.recout,
+				pipe_ctx);
+	} else {
+		/* if there is no overlap, zero recout */
+		memset(&pipe_ctx->plane_res.scl_data.recout, 0,
+				sizeof(struct rect));
 	}
+
 }
 
 static void calculate_scaling_ratios(struct pipe_ctx *pipe_ctx)
@@ -957,31 +1296,30 @@ static void calculate_init_and_vp(
 static void calculate_inits_and_viewports(struct pipe_ctx *pipe_ctx)
 {
 	const struct dc_plane_state *plane_state = pipe_ctx->plane_state;
-	const struct dc_stream_state *stream = pipe_ctx->stream;
 	struct scaler_data *data = &pipe_ctx->plane_res.scl_data;
 	struct rect src = plane_state->src_rect;
+	struct rect recout_dst_in_active_timing;
+	struct rect recout_clip_in_active_timing;
+	struct rect recout_clip_in_recout_dst;
+	struct rect overlap_in_active_timing;
+	struct rect odm_slice_src = resource_get_odm_slice_src_rect(pipe_ctx);
 	int vpc_div = (data->format == PIXEL_FORMAT_420BPP8
 				|| data->format == PIXEL_FORMAT_420BPP10) ? 2 : 1;
-	int split_count, split_idx, ro_lb, ro_tb, recout_full_x, recout_full_y;
 	bool orthogonal_rotation, flip_vert_scan_dir, flip_horz_scan_dir;
 
-	calculate_split_count_and_index(pipe_ctx, &split_count, &split_idx);
-	/*
-	 * recout full is what the recout would have been if we didnt clip
-	 * the source plane at all. We only care about left(ro_lb) and top(ro_tb)
-	 * offsets of recout within recout full because those are the directions
-	 * we scan from and therefore the only ones that affect inits.
-	 */
-	recout_full_x = stream->dst.x + (plane_state->dst_rect.x - stream->src.x)
-			* stream->dst.width / stream->src.width;
-	recout_full_y = stream->dst.y + (plane_state->dst_rect.y - stream->src.y)
-			* stream->dst.height / stream->src.height;
-	if (pipe_ctx->prev_odm_pipe && split_idx)
-		ro_lb = data->h_active * split_idx - recout_full_x;
+	recout_clip_in_active_timing = shift_rec(
+			&data->recout, odm_slice_src.x, odm_slice_src.y);
+	recout_dst_in_active_timing = calculate_plane_rec_in_timing_active(
+			pipe_ctx, &plane_state->dst_rect);
+	overlap_in_active_timing = intersect_rec(&recout_clip_in_active_timing,
+			&recout_dst_in_active_timing);
+	if (overlap_in_active_timing.width > 0 &&
+			overlap_in_active_timing.height > 0)
+		recout_clip_in_recout_dst = shift_rec(&overlap_in_active_timing,
+				-recout_dst_in_active_timing.x,
+				-recout_dst_in_active_timing.y);
 	else
-		ro_lb = data->recout.x - recout_full_x;
-	ro_tb = data->recout.y - recout_full_y;
-	ASSERT(ro_lb >= 0 && ro_tb >= 0);
+		memset(&recout_clip_in_recout_dst, 0, sizeof(struct rect));
 
 	/*
 	 * Work in recout rotation since that requires less transformations
@@ -1000,7 +1338,7 @@ static void calculate_inits_and_viewports(struct pipe_ctx *pipe_ctx)
 
 	calculate_init_and_vp(
 			flip_horz_scan_dir,
-			ro_lb,
+			recout_clip_in_recout_dst.x,
 			data->recout.width,
 			src.width,
 			data->taps.h_taps,
@@ -1010,7 +1348,7 @@ static void calculate_inits_and_viewports(struct pipe_ctx *pipe_ctx)
 			&data->viewport.width);
 	calculate_init_and_vp(
 			flip_horz_scan_dir,
-			ro_lb,
+			recout_clip_in_recout_dst.x,
 			data->recout.width,
 			src.width / vpc_div,
 			data->taps.h_taps_c,
@@ -1020,7 +1358,7 @@ static void calculate_inits_and_viewports(struct pipe_ctx *pipe_ctx)
 			&data->viewport_c.width);
 	calculate_init_and_vp(
 			flip_vert_scan_dir,
-			ro_tb,
+			recout_clip_in_recout_dst.y,
 			data->recout.height,
 			src.height,
 			data->taps.v_taps,
@@ -1030,7 +1368,7 @@ static void calculate_inits_and_viewports(struct pipe_ctx *pipe_ctx)
 			&data->viewport.height);
 	calculate_init_and_vp(
 			flip_vert_scan_dir,
-			ro_tb,
+			recout_clip_in_recout_dst.y,
 			data->recout.height,
 			src.height / vpc_div,
 			data->taps.v_taps_c,
@@ -1051,15 +1389,118 @@ static void calculate_inits_and_viewports(struct pipe_ctx *pipe_ctx)
 	data->viewport_c.y += src.y / vpc_div;
 }
 
+static enum controller_dp_test_pattern convert_dp_to_controller_test_pattern(
+				enum dp_test_pattern test_pattern)
+{
+	enum controller_dp_test_pattern controller_test_pattern;
+
+	switch (test_pattern) {
+	case DP_TEST_PATTERN_COLOR_SQUARES:
+		controller_test_pattern =
+				CONTROLLER_DP_TEST_PATTERN_COLORSQUARES;
+	break;
+	case DP_TEST_PATTERN_COLOR_SQUARES_CEA:
+		controller_test_pattern =
+				CONTROLLER_DP_TEST_PATTERN_COLORSQUARES_CEA;
+	break;
+	case DP_TEST_PATTERN_VERTICAL_BARS:
+		controller_test_pattern =
+				CONTROLLER_DP_TEST_PATTERN_VERTICALBARS;
+	break;
+	case DP_TEST_PATTERN_HORIZONTAL_BARS:
+		controller_test_pattern =
+				CONTROLLER_DP_TEST_PATTERN_HORIZONTALBARS;
+	break;
+	case DP_TEST_PATTERN_COLOR_RAMP:
+		controller_test_pattern =
+				CONTROLLER_DP_TEST_PATTERN_COLORRAMP;
+	break;
+	default:
+		controller_test_pattern =
+				CONTROLLER_DP_TEST_PATTERN_VIDEOMODE;
+	break;
+	}
+
+	return controller_test_pattern;
+}
+
+static enum controller_dp_color_space convert_dp_to_controller_color_space(
+		enum dp_test_pattern_color_space color_space)
+{
+	enum controller_dp_color_space controller_color_space;
+
+	switch (color_space) {
+	case DP_TEST_PATTERN_COLOR_SPACE_RGB:
+		controller_color_space = CONTROLLER_DP_COLOR_SPACE_RGB;
+		break;
+	case DP_TEST_PATTERN_COLOR_SPACE_YCBCR601:
+		controller_color_space = CONTROLLER_DP_COLOR_SPACE_YCBCR601;
+		break;
+	case DP_TEST_PATTERN_COLOR_SPACE_YCBCR709:
+		controller_color_space = CONTROLLER_DP_COLOR_SPACE_YCBCR709;
+		break;
+	case DP_TEST_PATTERN_COLOR_SPACE_UNDEFINED:
+	default:
+		controller_color_space = CONTROLLER_DP_COLOR_SPACE_UDEFINED;
+		break;
+	}
+
+	return controller_color_space;
+}
+
+void resource_build_test_pattern_params(struct resource_context *res_ctx,
+				struct pipe_ctx *otg_master)
+{
+	struct pipe_ctx *opp_heads[MAX_PIPES];
+	struct test_pattern_params *params;
+	int odm_cnt;
+	enum controller_dp_test_pattern controller_test_pattern;
+	enum controller_dp_color_space controller_color_space;
+	enum dc_color_depth color_depth = otg_master->stream->timing.display_color_depth;
+	struct rect odm_slice_src;
+	int i;
+
+	controller_test_pattern = convert_dp_to_controller_test_pattern(
+			otg_master->stream->test_pattern.type);
+	controller_color_space = convert_dp_to_controller_color_space(
+			otg_master->stream->test_pattern.color_space);
+
+	if (controller_test_pattern == CONTROLLER_DP_TEST_PATTERN_VIDEOMODE)
+		return;
+
+	odm_cnt = resource_get_opp_heads_for_otg_master(otg_master, res_ctx, opp_heads);
+
+	for (i = 0; i < odm_cnt; i++) {
+		odm_slice_src = resource_get_odm_slice_src_rect(opp_heads[i]);
+		params = &opp_heads[i]->stream_res.test_pattern_params;
+		params->test_pattern = controller_test_pattern;
+		params->color_space = controller_color_space;
+		params->color_depth = color_depth;
+		params->height = odm_slice_src.height;
+		params->offset = odm_slice_src.x;
+		params->width = odm_slice_src.width;
+	}
+}
+
 bool resource_build_scaling_params(struct pipe_ctx *pipe_ctx)
 {
 	const struct dc_plane_state *plane_state = pipe_ctx->plane_state;
 	struct dc_crtc_timing *timing = &pipe_ctx->stream->timing;
+	const struct rect odm_slice_src = resource_get_odm_slice_src_rect(pipe_ctx);
+	struct scaling_taps temp = {0};
 	bool res = false;
+
 	DC_LOGGER_INIT(pipe_ctx->stream->ctx->logger);
 
-	pipe_ctx->plane_res.scl_data.format = convert_pixel_format_to_dalsurface(
-			pipe_ctx->plane_state->format);
+	/* Invalid input */
+	if (!plane_state ||
+			!plane_state->dst_rect.width ||
+			!plane_state->dst_rect.height ||
+			!plane_state->src_rect.width ||
+			!plane_state->src_rect.height) {
+		ASSERT(0);
+		return false;
+	}
 
 	/* Timing borders are part of vactive that we are also supposed to skip in addition
 	 * to any stream dst offset. Since dm logic assumes dst is in addressable
@@ -1070,24 +1511,45 @@ bool resource_build_scaling_params(struct pipe_ctx *pipe_ctx)
 	pipe_ctx->stream->dst.y += timing->v_border_top;
 
 	/* Calculate H and V active size */
-	pipe_ctx->plane_res.scl_data.h_active = timing->h_addressable +
-			timing->h_border_left + timing->h_border_right;
-	pipe_ctx->plane_res.scl_data.v_active = timing->v_addressable +
-		timing->v_border_top + timing->v_border_bottom;
-	if (pipe_ctx->next_odm_pipe || pipe_ctx->prev_odm_pipe)
-		pipe_ctx->plane_res.scl_data.h_active /= get_num_odm_splits(pipe_ctx) + 1;
+	pipe_ctx->plane_res.scl_data.h_active = odm_slice_src.width;
+	pipe_ctx->plane_res.scl_data.v_active = odm_slice_src.height;
+	pipe_ctx->plane_res.scl_data.format = convert_pixel_format_to_dalsurface(
+			pipe_ctx->plane_state->format);
+
+#if defined(CONFIG_DRM_AMD_DC_FP)
+	if ((pipe_ctx->stream->ctx->dc->config.use_spl)	&& (!pipe_ctx->stream->ctx->dc->debug.disable_spl)) {
+		struct spl_in *spl_in = &pipe_ctx->plane_res.spl_in;
+		struct spl_out *spl_out = &pipe_ctx->plane_res.spl_out;
 
+		if (plane_state->ctx->dce_version > DCE_VERSION_MAX)
+			pipe_ctx->plane_res.scl_data.lb_params.depth = LB_PIXEL_DEPTH_36BPP;
+		else
+			pipe_ctx->plane_res.scl_data.lb_params.depth = LB_PIXEL_DEPTH_30BPP;
+
+		pipe_ctx->plane_res.scl_data.lb_params.alpha_en = plane_state->per_pixel_alpha;
+
+		// Convert pipe_ctx to respective input params for SPL
+		translate_SPL_in_params_from_pipe_ctx(pipe_ctx, spl_in);
+		/* Pass visual confirm debug information */
+		calculate_adjust_recout_for_visual_confirm(pipe_ctx,
+			&spl_in->debug.visual_confirm_base_offset,
+			&spl_in->debug.visual_confirm_dpp_offset);
+		// Set SPL output parameters to dscl_prog_data to be used for hw registers
+		spl_out->dscl_prog_data = resource_get_dscl_prog_data(pipe_ctx);
+		// Calculate scaler parameters from SPL
+		res = spl_calculate_scaler_params(spl_in, spl_out);
+		// Convert respective out params from SPL to scaler data
+		translate_SPL_out_params_to_pipe_ctx(pipe_ctx, spl_out);
+
+		/* Ignore scaler failure if pipe context plane is phantom plane */
+		if (!res && plane_state->is_phantom)
+			res = true;
+	} else {
+#endif
 	/* depends on h_active */
 	calculate_recout(pipe_ctx);
 	/* depends on pixel format */
 	calculate_scaling_ratios(pipe_ctx);
-	/* depends on scaling ratios and recout, does not calculate offset yet */
-	calculate_viewport_size(pipe_ctx);
-
-	/* Stopgap for validation of ODM + MPO on one side of screen case */
-	if (pipe_ctx->plane_res.scl_data.viewport.height < 1 ||
-			pipe_ctx->plane_res.scl_data.viewport.width < 1)
-		return false;
 
 	/*
 	 * LB calculations depend on vp size, h/v_active and scaling ratios
@@ -1095,18 +1557,37 @@ bool resource_build_scaling_params(struct pipe_ctx *pipe_ctx)
 	 * on certain displays, such as the Sharp 4k. 36bpp is needed
 	 * to support SURFACE_PIXEL_FORMAT_GRPH_ARGB16161616 and
 	 * SURFACE_PIXEL_FORMAT_GRPH_ABGR16161616 with actual > 10 bpc
-	 * precision on at least DCN display engines. However, at least
-	 * Carrizo with DCE_VERSION_11_0 does not like 36 bpp lb depth,
-	 * so use only 30 bpp on DCE_VERSION_11_0. Testing with DCE 11.2 and 8.3
-	 * did not show such problems, so this seems to be the exception.
+	 * precision on DCN display engines, but apparently not for DCE, as
+	 * far as testing on DCE-11.2 and DCE-8 showed. Various DCE parts have
+	 * problems: Carrizo with DCE_VERSION_11_0 does not like 36 bpp lb depth,
+	 * neither do DCE-8 at 4k resolution, or DCE-11.2 (broken identify pixel
+	 * passthrough). Therefore only use 36 bpp on DCN where it is actually needed.
 	 */
-	if (plane_state->ctx->dce_version > DCE_VERSION_11_0)
+	if (plane_state->ctx->dce_version > DCE_VERSION_MAX)
 		pipe_ctx->plane_res.scl_data.lb_params.depth = LB_PIXEL_DEPTH_36BPP;
 	else
 		pipe_ctx->plane_res.scl_data.lb_params.depth = LB_PIXEL_DEPTH_30BPP;
 
 	pipe_ctx->plane_res.scl_data.lb_params.alpha_en = plane_state->per_pixel_alpha;
 
+	// get TAP value with 100x100 dummy data for max scaling qualify, override
+	// if a new scaling quality required
+	pipe_ctx->plane_res.scl_data.viewport.width = 100;
+	pipe_ctx->plane_res.scl_data.viewport.height = 100;
+	pipe_ctx->plane_res.scl_data.viewport_c.width = 100;
+	pipe_ctx->plane_res.scl_data.viewport_c.height = 100;
+	if (pipe_ctx->plane_res.xfm != NULL)
+		res = pipe_ctx->plane_res.xfm->funcs->transform_get_optimal_number_of_taps(
+				pipe_ctx->plane_res.xfm, &pipe_ctx->plane_res.scl_data, &plane_state->scaling_quality);
+
+	if (pipe_ctx->plane_res.dpp != NULL)
+		res = pipe_ctx->plane_res.dpp->funcs->dpp_get_optimal_number_of_taps(
+				pipe_ctx->plane_res.dpp, &pipe_ctx->plane_res.scl_data, &plane_state->scaling_quality);
+
+	temp = pipe_ctx->plane_res.scl_data.taps;
+
+	calculate_inits_and_viewports(pipe_ctx);
+
 	if (pipe_ctx->plane_res.xfm != NULL)
 		res = pipe_ctx->plane_res.xfm->funcs->transform_get_optimal_number_of_taps(
 				pipe_ctx->plane_res.xfm, &pipe_ctx->plane_res.scl_data, &plane_state->scaling_quality);
@@ -1133,11 +1614,14 @@ bool resource_build_scaling_params(struct pipe_ctx *pipe_ctx)
 					&plane_state->scaling_quality);
 	}
 
-	/*
-	 * Depends on recout, scaling ratios, h_active and taps
-	 * May need to re-check lb size after this in some obscure scenario
-	 */
-	if (res)
+	/* Ignore scaler failure if pipe context plane is phantom plane */
+	if (!res && plane_state->is_phantom)
+		res = true;
+
+	if (res && (pipe_ctx->plane_res.scl_data.taps.v_taps != temp.v_taps ||
+		pipe_ctx->plane_res.scl_data.taps.h_taps != temp.h_taps ||
+		pipe_ctx->plane_res.scl_data.taps.v_taps_c != temp.v_taps_c ||
+		pipe_ctx->plane_res.scl_data.taps.h_taps_c != temp.h_taps_c))
 		calculate_inits_and_viewports(pipe_ctx);
 
 	/*
@@ -1155,18 +1639,14 @@ bool resource_build_scaling_params(struct pipe_ctx *pipe_ctx)
 			pipe_ctx->plane_res.scl_data.recout.x += pipe_ctx->plane_res.scl_data.recout.width;
 	}
 
-	if (!pipe_ctx->stream->ctx->dc->config.enable_windowed_mpo_odm) {
-		if (pipe_ctx->plane_res.scl_data.viewport.height < MIN_VIEWPORT_SIZE ||
-				pipe_ctx->plane_res.scl_data.viewport.width < MIN_VIEWPORT_SIZE)
-			res = false;
-	} else {
-		/* Clamp minimum viewport size */
-		if (pipe_ctx->plane_res.scl_data.viewport.height < MIN_VIEWPORT_SIZE)
-			pipe_ctx->plane_res.scl_data.viewport.height = MIN_VIEWPORT_SIZE;
-		if (pipe_ctx->plane_res.scl_data.viewport.width < MIN_VIEWPORT_SIZE)
-			pipe_ctx->plane_res.scl_data.viewport.width = MIN_VIEWPORT_SIZE;
+	/* Clamp minimum viewport size */
+	if (pipe_ctx->plane_res.scl_data.viewport.height < MIN_VIEWPORT_SIZE)
+		pipe_ctx->plane_res.scl_data.viewport.height = MIN_VIEWPORT_SIZE;
+	if (pipe_ctx->plane_res.scl_data.viewport.width < MIN_VIEWPORT_SIZE)
+		pipe_ctx->plane_res.scl_data.viewport.width = MIN_VIEWPORT_SIZE;
+#ifdef CONFIG_DRM_AMD_DC_FP
 	}
-
+#endif
 	DC_LOG_SCALER("%s pipe %d:\nViewport: height:%d width:%d x:%d y:%d  Recout: height:%d width:%d x:%d y:%d  HACTIVE:%d VACTIVE:%d\n"
 			"src_rect: height:%d width:%d x:%d y:%d  dst_rect: height:%d width:%d x:%d y:%d  clip_rect: height:%d width:%d x:%d y:%d\n",
 			__func__,
@@ -1200,6 +1680,62 @@ bool resource_build_scaling_params(struct pipe_ctx *pipe_ctx)
 	return res;
 }
 
+bool resource_can_pipe_disable_cursor(struct pipe_ctx *pipe_ctx)
+{
+	struct pipe_ctx *test_pipe, *split_pipe;
+	struct rect r1 = pipe_ctx->plane_res.scl_data.recout;
+	int r1_right, r1_bottom;
+	int cur_layer = pipe_ctx->plane_state->layer_index;
+
+	reverse_adjust_recout_for_visual_confirm(&r1, pipe_ctx);
+	r1_right = r1.x + r1.width;
+	r1_bottom = r1.y + r1.height;
+
+	/**
+	 * Disable the cursor if there's another pipe above this with a
+	 * plane that contains this pipe's viewport to prevent double cursor
+	 * and incorrect scaling artifacts.
+	 */
+	for (test_pipe = pipe_ctx->top_pipe; test_pipe;
+	     test_pipe = test_pipe->top_pipe) {
+		struct rect r2;
+		int r2_right, r2_bottom;
+		// Skip invisible layer and pipe-split plane on same layer
+		if (!test_pipe->plane_state ||
+		    !test_pipe->plane_state->visible ||
+		    test_pipe->plane_state->layer_index == cur_layer)
+			continue;
+
+		r2 = test_pipe->plane_res.scl_data.recout;
+		reverse_adjust_recout_for_visual_confirm(&r2, test_pipe);
+		r2_right = r2.x + r2.width;
+		r2_bottom = r2.y + r2.height;
+
+		/**
+		 * There is another half plane on same layer because of
+		 * pipe-split, merge together per same height.
+		 */
+		for (split_pipe = pipe_ctx->top_pipe; split_pipe;
+		     split_pipe = split_pipe->top_pipe)
+			if (split_pipe->plane_state->layer_index == test_pipe->plane_state->layer_index) {
+				struct rect r2_half;
+
+				r2_half = split_pipe->plane_res.scl_data.recout;
+				reverse_adjust_recout_for_visual_confirm(&r2_half, split_pipe);
+				r2.x = min(r2_half.x, r2.x);
+				r2.width = r2.width + r2_half.width;
+				r2_right = r2.x + r2.width;
+				r2_bottom = min(r2_bottom, r2_half.y + r2_half.height);
+				break;
+			}
+
+		if (r1.x >= r2.x && r1.y >= r2.y && r1_right <= r2_right && r1_bottom <= r2_bottom)
+			return true;
+	}
+
+	return false;
+}
+
 
 enum dc_status resource_build_scaling_params_for_context(
 	const struct dc  *dc,
@@ -1217,7 +1753,7 @@ enum dc_status resource_build_scaling_params_for_context(
 	return DC_OK;
 }
 
-struct pipe_ctx *find_idle_secondary_pipe(
+struct pipe_ctx *resource_find_free_secondary_pipe_legacy(
 		struct resource_context *res_ctx,
 		const struct resource_pool *pool,
 		const struct pipe_ctx *primary_pipe)
@@ -1277,28 +1813,715 @@ struct pipe_ctx *find_idle_secondary_pipe(
 	return secondary_pipe;
 }
 
-struct pipe_ctx *resource_get_head_pipe_for_stream(
+int resource_find_free_pipe_used_as_sec_opp_head_by_cur_otg_master(
+		const struct resource_context *cur_res_ctx,
+		struct resource_context *new_res_ctx,
+		const struct pipe_ctx *cur_otg_master)
+{
+	const struct pipe_ctx *cur_sec_opp_head = cur_otg_master->next_odm_pipe;
+	struct pipe_ctx *new_pipe;
+	int free_pipe_idx = FREE_PIPE_INDEX_NOT_FOUND;
+
+	while (cur_sec_opp_head) {
+		new_pipe = &new_res_ctx->pipe_ctx[cur_sec_opp_head->pipe_idx];
+		if (resource_is_pipe_type(new_pipe, FREE_PIPE)) {
+			free_pipe_idx = cur_sec_opp_head->pipe_idx;
+			break;
+		}
+		cur_sec_opp_head = cur_sec_opp_head->next_odm_pipe;
+	}
+
+	return free_pipe_idx;
+}
+
+int resource_find_free_pipe_used_in_cur_mpc_blending_tree(
+		const struct resource_context *cur_res_ctx,
+		struct resource_context *new_res_ctx,
+		const struct pipe_ctx *cur_opp_head)
+{
+	const struct pipe_ctx *cur_sec_dpp = cur_opp_head->bottom_pipe;
+	struct pipe_ctx *new_pipe;
+	int free_pipe_idx = FREE_PIPE_INDEX_NOT_FOUND;
+
+	while (cur_sec_dpp) {
+		/* find a free pipe used in current opp blend tree,
+		 * this is to avoid MPO pipe switching to different opp blending
+		 * tree
+		 */
+		new_pipe = &new_res_ctx->pipe_ctx[cur_sec_dpp->pipe_idx];
+		if (resource_is_pipe_type(new_pipe, FREE_PIPE)) {
+			free_pipe_idx = cur_sec_dpp->pipe_idx;
+			break;
+		}
+		cur_sec_dpp = cur_sec_dpp->bottom_pipe;
+	}
+
+	return free_pipe_idx;
+}
+
+int recource_find_free_pipe_not_used_in_cur_res_ctx(
+		const struct resource_context *cur_res_ctx,
+		struct resource_context *new_res_ctx,
+		const struct resource_pool *pool)
+{
+	int free_pipe_idx = FREE_PIPE_INDEX_NOT_FOUND;
+	const struct pipe_ctx *new_pipe, *cur_pipe;
+	int i;
+
+	for (i = 0; i < pool->pipe_count; i++) {
+		cur_pipe = &cur_res_ctx->pipe_ctx[i];
+		new_pipe = &new_res_ctx->pipe_ctx[i];
+
+		if (resource_is_pipe_type(cur_pipe, FREE_PIPE) &&
+				resource_is_pipe_type(new_pipe, FREE_PIPE)) {
+			free_pipe_idx = i;
+			break;
+		}
+	}
+
+	return free_pipe_idx;
+}
+
+int recource_find_free_pipe_used_as_otg_master_in_cur_res_ctx(
+		const struct resource_context *cur_res_ctx,
+		struct resource_context *new_res_ctx,
+		const struct resource_pool *pool)
+{
+	int free_pipe_idx = FREE_PIPE_INDEX_NOT_FOUND;
+	const struct pipe_ctx *new_pipe, *cur_pipe;
+	int i;
+
+	for (i = 0; i < pool->pipe_count; i++) {
+		cur_pipe = &cur_res_ctx->pipe_ctx[i];
+		new_pipe = &new_res_ctx->pipe_ctx[i];
+
+		if (resource_is_pipe_type(cur_pipe, OTG_MASTER) &&
+				resource_is_pipe_type(new_pipe, FREE_PIPE)) {
+			free_pipe_idx = i;
+			break;
+		}
+	}
+
+	return free_pipe_idx;
+}
+
+int resource_find_free_pipe_used_as_cur_sec_dpp(
+		const struct resource_context *cur_res_ctx,
+		struct resource_context *new_res_ctx,
+		const struct resource_pool *pool)
+{
+	int free_pipe_idx = FREE_PIPE_INDEX_NOT_FOUND;
+	const struct pipe_ctx *new_pipe, *cur_pipe;
+	int i;
+
+	for (i = 0; i < pool->pipe_count; i++) {
+		cur_pipe = &cur_res_ctx->pipe_ctx[i];
+		new_pipe = &new_res_ctx->pipe_ctx[i];
+
+		if (resource_is_pipe_type(cur_pipe, DPP_PIPE) &&
+				!resource_is_pipe_type(cur_pipe, OPP_HEAD) &&
+				resource_is_pipe_type(new_pipe, FREE_PIPE)) {
+			free_pipe_idx = i;
+			break;
+		}
+	}
+
+	return free_pipe_idx;
+}
+
+int resource_find_free_pipe_used_as_cur_sec_dpp_in_mpcc_combine(
+		const struct resource_context *cur_res_ctx,
+		struct resource_context *new_res_ctx,
+		const struct resource_pool *pool)
+{
+	int free_pipe_idx = FREE_PIPE_INDEX_NOT_FOUND;
+	const struct pipe_ctx *new_pipe, *cur_pipe;
+	int i;
+
+	for (i = 0; i < pool->pipe_count; i++) {
+		cur_pipe = &cur_res_ctx->pipe_ctx[i];
+		new_pipe = &new_res_ctx->pipe_ctx[i];
+
+		if (resource_is_pipe_type(cur_pipe, DPP_PIPE) &&
+				!resource_is_pipe_type(cur_pipe, OPP_HEAD) &&
+				resource_get_mpc_slice_index(cur_pipe) > 0 &&
+				resource_is_pipe_type(new_pipe, FREE_PIPE)) {
+			free_pipe_idx = i;
+			break;
+		}
+	}
+
+	return free_pipe_idx;
+}
+
+int resource_find_any_free_pipe(struct resource_context *new_res_ctx,
+		const struct resource_pool *pool)
+{
+	int free_pipe_idx = FREE_PIPE_INDEX_NOT_FOUND;
+	const struct pipe_ctx *new_pipe;
+	int i;
+
+	for (i = 0; i < pool->pipe_count; i++) {
+		new_pipe = &new_res_ctx->pipe_ctx[i];
+
+		if (resource_is_pipe_type(new_pipe, FREE_PIPE)) {
+			free_pipe_idx = i;
+			break;
+		}
+	}
+
+	return free_pipe_idx;
+}
+
+bool resource_is_pipe_type(const struct pipe_ctx *pipe_ctx, enum pipe_type type)
+{
+	switch (type) {
+	case OTG_MASTER:
+		return !pipe_ctx->prev_odm_pipe &&
+				!pipe_ctx->top_pipe &&
+				pipe_ctx->stream;
+	case OPP_HEAD:
+		return !pipe_ctx->top_pipe && pipe_ctx->stream;
+	case DPP_PIPE:
+		return pipe_ctx->plane_state && pipe_ctx->stream;
+	case FREE_PIPE:
+		return !pipe_ctx->plane_state && !pipe_ctx->stream;
+	default:
+		return false;
+	}
+}
+
+struct pipe_ctx *resource_get_otg_master_for_stream(
 		struct resource_context *res_ctx,
-		struct dc_stream_state *stream)
+		const struct dc_stream_state *stream)
 {
 	int i;
 
 	for (i = 0; i < MAX_PIPES; i++) {
-		if (res_ctx->pipe_ctx[i].stream == stream
-				&& !res_ctx->pipe_ctx[i].top_pipe
-				&& !res_ctx->pipe_ctx[i].prev_odm_pipe)
+		if (res_ctx->pipe_ctx[i].stream == stream &&
+				resource_is_pipe_type(&res_ctx->pipe_ctx[i], OTG_MASTER))
 			return &res_ctx->pipe_ctx[i];
 	}
 	return NULL;
 }
 
-static struct pipe_ctx *resource_get_tail_pipe(
+int resource_get_opp_heads_for_otg_master(const struct pipe_ctx *otg_master,
 		struct resource_context *res_ctx,
-		struct pipe_ctx *head_pipe)
+		struct pipe_ctx *opp_heads[MAX_PIPES])
 {
-	struct pipe_ctx *tail_pipe;
+	struct pipe_ctx *opp_head = &res_ctx->pipe_ctx[otg_master->pipe_idx];
+	struct dc *dc = otg_master->stream->ctx->dc;
+	int i = 0;
+
+	DC_LOGGER_INIT(dc->ctx->logger);
 
-	tail_pipe = head_pipe->bottom_pipe;
+	if (!resource_is_pipe_type(otg_master, OTG_MASTER)) {
+		DC_LOG_WARNING("%s called from a non OTG master, something "
+			       "is wrong in the pipe configuration",
+			       __func__);
+		ASSERT(0);
+		return 0;
+	}
+	while (opp_head) {
+		ASSERT(i < MAX_PIPES);
+		opp_heads[i++] = opp_head;
+		opp_head = opp_head->next_odm_pipe;
+	}
+	return i;
+}
+
+int resource_get_dpp_pipes_for_opp_head(const struct pipe_ctx *opp_head,
+		struct resource_context *res_ctx,
+		struct pipe_ctx *dpp_pipes[MAX_PIPES])
+{
+	struct pipe_ctx *pipe = &res_ctx->pipe_ctx[opp_head->pipe_idx];
+	int i = 0;
+
+	if (!resource_is_pipe_type(opp_head, OPP_HEAD)) {
+		ASSERT(0);
+		return 0;
+	}
+	while (pipe && resource_is_pipe_type(pipe, DPP_PIPE)) {
+		ASSERT(i < MAX_PIPES);
+		dpp_pipes[i++] = pipe;
+		pipe = pipe->bottom_pipe;
+	}
+	return i;
+}
+
+int resource_get_dpp_pipes_for_plane(const struct dc_plane_state *plane,
+		struct resource_context *res_ctx,
+		struct pipe_ctx *dpp_pipes[MAX_PIPES])
+{
+	int i = 0, j;
+	struct pipe_ctx *pipe;
+
+	for (j = 0; j < MAX_PIPES; j++) {
+		pipe = &res_ctx->pipe_ctx[j];
+		if (pipe->plane_state == plane && pipe->prev_odm_pipe == NULL) {
+			if (resource_is_pipe_type(pipe, OPP_HEAD) ||
+					pipe->top_pipe->plane_state != plane)
+				break;
+		}
+	}
+
+	if (j < MAX_PIPES) {
+		if (pipe->next_odm_pipe)
+			while (pipe) {
+				dpp_pipes[i++] = pipe;
+				pipe = pipe->next_odm_pipe;
+			}
+		else
+			while (pipe && pipe->plane_state == plane) {
+				dpp_pipes[i++] = pipe;
+				pipe = pipe->bottom_pipe;
+			}
+	}
+	return i;
+}
+
+struct pipe_ctx *resource_get_otg_master(const struct pipe_ctx *pipe_ctx)
+{
+	struct pipe_ctx *otg_master = resource_get_opp_head(pipe_ctx);
+
+	while (otg_master->prev_odm_pipe)
+		otg_master = otg_master->prev_odm_pipe;
+	return otg_master;
+}
+
+struct pipe_ctx *resource_get_opp_head(const struct pipe_ctx *pipe_ctx)
+{
+	struct pipe_ctx *opp_head = (struct pipe_ctx *) pipe_ctx;
+
+	ASSERT(!resource_is_pipe_type(opp_head, FREE_PIPE));
+	while (opp_head->top_pipe)
+		opp_head = opp_head->top_pipe;
+	return opp_head;
+}
+
+struct pipe_ctx *resource_get_primary_dpp_pipe(const struct pipe_ctx *dpp_pipe)
+{
+	struct pipe_ctx *pri_dpp_pipe = (struct pipe_ctx *) dpp_pipe;
+
+	ASSERT(resource_is_pipe_type(dpp_pipe, DPP_PIPE));
+	while (pri_dpp_pipe->prev_odm_pipe)
+		pri_dpp_pipe = pri_dpp_pipe->prev_odm_pipe;
+	while (pri_dpp_pipe->top_pipe &&
+			pri_dpp_pipe->top_pipe->plane_state == pri_dpp_pipe->plane_state)
+		pri_dpp_pipe = pri_dpp_pipe->top_pipe;
+	return pri_dpp_pipe;
+}
+
+
+int resource_get_mpc_slice_index(const struct pipe_ctx *pipe_ctx)
+{
+	struct pipe_ctx *split_pipe = pipe_ctx->top_pipe;
+	int index = 0;
+
+	while (split_pipe && split_pipe->plane_state == pipe_ctx->plane_state) {
+		index++;
+		split_pipe = split_pipe->top_pipe;
+	}
+
+	return index;
+}
+
+int resource_get_mpc_slice_count(const struct pipe_ctx *pipe)
+{
+	int mpc_split_count = 1;
+	const struct pipe_ctx *other_pipe = pipe->bottom_pipe;
+
+	while (other_pipe && other_pipe->plane_state == pipe->plane_state) {
+		mpc_split_count++;
+		other_pipe = other_pipe->bottom_pipe;
+	}
+	other_pipe = pipe->top_pipe;
+	while (other_pipe && other_pipe->plane_state == pipe->plane_state) {
+		mpc_split_count++;
+		other_pipe = other_pipe->top_pipe;
+	}
+
+	return mpc_split_count;
+}
+
+int resource_get_odm_slice_count(const struct pipe_ctx *pipe)
+{
+	int odm_split_count = 1;
+
+	pipe = resource_get_otg_master(pipe);
+
+	while (pipe->next_odm_pipe) {
+		odm_split_count++;
+		pipe = pipe->next_odm_pipe;
+	}
+	return odm_split_count;
+}
+
+int resource_get_odm_slice_index(const struct pipe_ctx *pipe_ctx)
+{
+	int index = 0;
+
+	pipe_ctx = resource_get_opp_head(pipe_ctx);
+	if (!pipe_ctx)
+		return 0;
+
+	while (pipe_ctx->prev_odm_pipe) {
+		index++;
+		pipe_ctx = pipe_ctx->prev_odm_pipe;
+	}
+
+	return index;
+}
+
+int resource_get_odm_slice_dst_width(struct pipe_ctx *otg_master,
+		bool is_last_segment)
+{
+	const struct dc_crtc_timing *timing;
+	int count;
+	int h_active;
+	int width;
+	bool two_pixel_alignment_required = false;
+
+	if (!otg_master || !otg_master->stream)
+		return 0;
+
+	timing = &otg_master->stream->timing;
+	count = resource_get_odm_slice_count(otg_master);
+	h_active = timing->h_addressable +
+			timing->h_border_left +
+			timing->h_border_right +
+			otg_master->dsc_padding_params.dsc_hactive_padding;
+	width = h_active / count;
+
+	if (otg_master->stream_res.tg)
+		two_pixel_alignment_required =
+				otg_master->stream_res.tg->funcs->is_two_pixels_per_container(timing) ||
+				/*
+				 * 422 is sub-sampled horizontally. 1 set of chromas
+				 * (Cb/Cr) is shared for 2 lumas (i.e 2 Y values).
+				 * Therefore even if 422 is still 1 pixel per container,
+				 * ODM segment width still needs to be 2 pixel aligned.
+				 */
+				timing->pixel_encoding == PIXEL_ENCODING_YCBCR422;
+	if ((width % 2) && two_pixel_alignment_required)
+		width++;
+
+	return is_last_segment ?
+			h_active - width * (count - 1) :
+			width;
+}
+
+struct rect resource_get_odm_slice_dst_rect(struct pipe_ctx *pipe_ctx)
+{
+	const struct dc_stream_state *stream = pipe_ctx->stream;
+	bool is_last_odm_slice = pipe_ctx->next_odm_pipe == NULL;
+	struct pipe_ctx *otg_master = resource_get_otg_master(pipe_ctx);
+	int odm_slice_idx = resource_get_odm_slice_index(pipe_ctx);
+	int odm_segment_offset = resource_get_odm_slice_dst_width(otg_master, false);
+	struct rect odm_slice_dst;
+
+	odm_slice_dst.x = odm_segment_offset * odm_slice_idx;
+	odm_slice_dst.width = resource_get_odm_slice_dst_width(otg_master, is_last_odm_slice);
+	odm_slice_dst.y = 0;
+	odm_slice_dst.height = stream->timing.v_addressable +
+			stream->timing.v_border_bottom +
+			stream->timing.v_border_top;
+
+	return odm_slice_dst;
+}
+
+struct rect resource_get_odm_slice_src_rect(struct pipe_ctx *pipe_ctx)
+{
+	struct rect odm_slice_dst;
+	struct rect odm_slice_src;
+	struct pipe_ctx *opp_head = resource_get_opp_head(pipe_ctx);
+	struct output_pixel_processor *opp = opp_head->stream_res.opp;
+	uint32_t left_edge_extra_pixel_count;
+
+	odm_slice_dst = resource_get_odm_slice_dst_rect(opp_head);
+	odm_slice_src = odm_slice_dst;
+
+	if (opp && opp->funcs->opp_get_left_edge_extra_pixel_count)
+		left_edge_extra_pixel_count =
+				opp->funcs->opp_get_left_edge_extra_pixel_count(
+						opp, pipe_ctx->stream->timing.pixel_encoding,
+						resource_is_pipe_type(opp_head, OTG_MASTER));
+	else
+		left_edge_extra_pixel_count = 0;
+
+	odm_slice_src.x -= left_edge_extra_pixel_count;
+	odm_slice_src.width += left_edge_extra_pixel_count;
+
+	return odm_slice_src;
+}
+
+bool resource_is_pipe_topology_changed(const struct dc_state *state_a,
+		const struct dc_state *state_b)
+{
+	int i;
+	const struct pipe_ctx *pipe_a, *pipe_b;
+
+	if (state_a->stream_count != state_b->stream_count)
+		return true;
+
+	for (i = 0; i < MAX_PIPES; i++) {
+		pipe_a = &state_a->res_ctx.pipe_ctx[i];
+		pipe_b = &state_b->res_ctx.pipe_ctx[i];
+
+		if (pipe_a->stream && !pipe_b->stream)
+			return true;
+		else if (!pipe_a->stream && pipe_b->stream)
+			return true;
+
+		if (pipe_a->plane_state && !pipe_b->plane_state)
+			return true;
+		else if (!pipe_a->plane_state && pipe_b->plane_state)
+			return true;
+
+		if (pipe_a->bottom_pipe && pipe_b->bottom_pipe) {
+			if (pipe_a->bottom_pipe->pipe_idx != pipe_b->bottom_pipe->pipe_idx)
+				return true;
+			if ((pipe_a->bottom_pipe->plane_state == pipe_a->plane_state) &&
+					(pipe_b->bottom_pipe->plane_state != pipe_b->plane_state))
+				return true;
+			else if ((pipe_a->bottom_pipe->plane_state != pipe_a->plane_state) &&
+					(pipe_b->bottom_pipe->plane_state == pipe_b->plane_state))
+				return true;
+		} else if (pipe_a->bottom_pipe || pipe_b->bottom_pipe) {
+			return true;
+		}
+
+		if (pipe_a->next_odm_pipe && pipe_b->next_odm_pipe) {
+			if (pipe_a->next_odm_pipe->pipe_idx != pipe_b->next_odm_pipe->pipe_idx)
+				return true;
+		} else if (pipe_a->next_odm_pipe || pipe_b->next_odm_pipe) {
+			return true;
+		}
+	}
+	return false;
+}
+
+bool resource_is_odm_topology_changed(const struct pipe_ctx *otg_master_a,
+		const struct pipe_ctx *otg_master_b)
+{
+	const struct pipe_ctx *opp_head_a = otg_master_a;
+	const struct pipe_ctx *opp_head_b = otg_master_b;
+
+	if (!resource_is_pipe_type(otg_master_a, OTG_MASTER) ||
+			!resource_is_pipe_type(otg_master_b, OTG_MASTER))
+		return true;
+
+	while (opp_head_a && opp_head_b) {
+		if (opp_head_a->stream_res.opp != opp_head_b->stream_res.opp)
+			return true;
+		if ((opp_head_a->next_odm_pipe && !opp_head_b->next_odm_pipe) ||
+				(!opp_head_a->next_odm_pipe && opp_head_b->next_odm_pipe))
+			return true;
+		opp_head_a = opp_head_a->next_odm_pipe;
+		opp_head_b = opp_head_b->next_odm_pipe;
+	}
+
+	return false;
+}
+
+/*
+ * Sample log:
+ *    pipe topology update
+ *  ________________________
+ * | plane0  slice0  stream0|
+ * |DPP0----OPP0----OTG0----| <--- case 0 (OTG master pipe with plane)
+ * | plane1 |       |       |
+ * |DPP1----|       |       | <--- case 5 (DPP pipe not in last slice)
+ * | plane0  slice1 |       |
+ * |DPP2----OPP2----|       | <--- case 2 (OPP head pipe with plane)
+ * | plane1 |               |
+ * |DPP3----|               | <--- case 4 (DPP pipe in last slice)
+ * |         slice0  stream1|
+ * |DPG4----OPP4----OTG4----| <--- case 1 (OTG master pipe without plane)
+ * |         slice1 |       |
+ * |DPG5----OPP5----|       | <--- case 3 (OPP head pipe without plane)
+ * |________________________|
+ */
+
+static void resource_log_pipe(struct dc *dc, struct pipe_ctx *pipe,
+		int stream_idx, int slice_idx, int plane_idx, int slice_count,
+		bool is_primary, bool is_phantom_pipe)
+{
+	DC_LOGGER_INIT(dc->ctx->logger);
+
+	// new format for logging: bit storing code
+	if (slice_idx == 0 && plane_idx == 0 && is_primary) {
+		/* case 0 (OTG master pipe with plane) */
+		DC_LOG_DC(" | plane%d  slice%d  stream%d|",
+				plane_idx, slice_idx, stream_idx);
+		DC_LOG_DC(" |DPP%d----OPP%d----OTG%d----|",
+				pipe->plane_res.dpp->inst,
+				pipe->stream_res.opp->inst,
+				pipe->stream_res.tg->inst);
+		capture_pipe_topology_data(dc, plane_idx, slice_idx, stream_idx,
+				pipe->plane_res.dpp->inst,
+				pipe->stream_res.opp->inst,
+				pipe->stream_res.tg->inst, is_phantom_pipe);
+	} else if (slice_idx == 0 && plane_idx == -1) {
+		/* case 1 (OTG master pipe without plane) */
+		DC_LOG_DC(" |         slice%d  stream%d|",
+				slice_idx, stream_idx);
+		DC_LOG_DC(" |DPG%d----OPP%d----OTG%d----|",
+				pipe->stream_res.opp->inst,
+				pipe->stream_res.opp->inst,
+				pipe->stream_res.tg->inst);
+		capture_pipe_topology_data(dc, 0xF, slice_idx, stream_idx,
+				pipe->plane_res.dpp->inst,
+				pipe->stream_res.opp->inst,
+				pipe->stream_res.tg->inst, is_phantom_pipe);
+	} else if (slice_idx != 0 && plane_idx == 0 && is_primary) {
+		/* case 2 (OPP head pipe with plane) */
+		DC_LOG_DC(" | plane%d  slice%d |       |",
+				plane_idx, slice_idx);
+		DC_LOG_DC(" |DPP%d----OPP%d----|       |",
+				pipe->plane_res.dpp->inst,
+				pipe->stream_res.opp->inst);
+		capture_pipe_topology_data(dc, plane_idx, slice_idx, stream_idx,
+				pipe->plane_res.dpp->inst,
+				pipe->stream_res.opp->inst,
+				pipe->stream_res.tg->inst, is_phantom_pipe);
+	} else if (slice_idx != 0 && plane_idx == -1) {
+		/* case 3 (OPP head pipe without plane) */
+		DC_LOG_DC(" |         slice%d |       |", slice_idx);
+		DC_LOG_DC(" |DPG%d----OPP%d----|       |",
+				pipe->plane_res.dpp->inst,
+				pipe->stream_res.opp->inst);
+		capture_pipe_topology_data(dc, 0xF, slice_idx, stream_idx,
+				pipe->plane_res.dpp->inst,
+				pipe->stream_res.opp->inst,
+				pipe->stream_res.tg->inst, is_phantom_pipe);
+	} else if (slice_idx == slice_count - 1) {
+		/* case 4 (DPP pipe in last slice) */
+		DC_LOG_DC(" | plane%d |               |", plane_idx);
+		DC_LOG_DC(" |DPP%d----|               |",
+				pipe->plane_res.dpp->inst);
+		capture_pipe_topology_data(dc, plane_idx, slice_idx, stream_idx,
+				pipe->plane_res.dpp->inst,
+				pipe->stream_res.opp->inst,
+				pipe->stream_res.tg->inst, is_phantom_pipe);
+	} else {
+		/* case 5 (DPP pipe not in last slice) */
+		DC_LOG_DC(" | plane%d |       |       |", plane_idx);
+		DC_LOG_DC(" |DPP%d----|       |       |",
+				pipe->plane_res.dpp->inst);
+		capture_pipe_topology_data(dc, plane_idx, slice_idx, stream_idx,
+				pipe->plane_res.dpp->inst,
+				pipe->stream_res.opp->inst,
+				pipe->stream_res.tg->inst, is_phantom_pipe);
+	}
+}
+
+static void resource_log_pipe_for_stream(struct dc *dc, struct dc_state *state,
+		struct pipe_ctx *otg_master, int stream_idx, bool is_phantom_pipe)
+{
+	struct pipe_ctx *opp_heads[MAX_PIPES];
+	struct pipe_ctx *dpp_pipes[MAX_PIPES];
+
+	int slice_idx, dpp_idx, plane_idx, slice_count, dpp_count;
+	bool is_primary;
+	DC_LOGGER_INIT(dc->ctx->logger);
+
+	slice_count = resource_get_opp_heads_for_otg_master(otg_master,
+			&state->res_ctx, opp_heads);
+	for (slice_idx = 0; slice_idx < slice_count; slice_idx++) {
+		plane_idx = -1;
+		if (opp_heads[slice_idx]->plane_state) {
+			dpp_count = resource_get_dpp_pipes_for_opp_head(
+					opp_heads[slice_idx],
+					&state->res_ctx,
+					dpp_pipes);
+			for (dpp_idx = 0; dpp_idx < dpp_count; dpp_idx++) {
+				is_primary = !dpp_pipes[dpp_idx]->top_pipe ||
+						dpp_pipes[dpp_idx]->top_pipe->plane_state != dpp_pipes[dpp_idx]->plane_state;
+				if (is_primary)
+					plane_idx++;
+				resource_log_pipe(dc, dpp_pipes[dpp_idx],
+						stream_idx, slice_idx,
+						plane_idx, slice_count,
+						is_primary, is_phantom_pipe);
+			}
+		} else {
+			resource_log_pipe(dc, opp_heads[slice_idx],
+					stream_idx, slice_idx, plane_idx,
+					slice_count, true, is_phantom_pipe);
+		}
+
+	}
+}
+
+static int resource_stream_to_stream_idx(struct dc_state *state,
+		struct dc_stream_state *stream)
+{
+	int i, stream_idx = -1;
+
+	for (i = 0; i < state->stream_count; i++)
+		if (state->streams[i] == stream) {
+			stream_idx = i;
+			break;
+		}
+
+	/* never return negative array index */
+	if (stream_idx == -1) {
+		ASSERT(0);
+		return 0;
+	}
+
+	return stream_idx;
+}
+
+void resource_log_pipe_topology_update(struct dc *dc, struct dc_state *state)
+{
+	struct pipe_ctx *otg_master;
+	int stream_idx, phantom_stream_idx;
+	DC_LOGGER_INIT(dc->ctx->logger);
+	bool is_phantom_pipe = false;
+
+	// Start a new snapshot for this topology update
+	start_new_topology_snapshot(dc, state);
+
+	DC_LOG_DC("    pipe topology update");
+	DC_LOG_DC("  ________________________");
+	for (stream_idx = 0; stream_idx < state->stream_count; stream_idx++) {
+		if (state->streams[stream_idx]->is_phantom)
+			continue;
+
+		otg_master = resource_get_otg_master_for_stream(
+				&state->res_ctx, state->streams[stream_idx]);
+
+		if (!otg_master)
+			continue;
+
+		resource_log_pipe_for_stream(dc, state, otg_master, stream_idx, is_phantom_pipe);
+	}
+	if (state->phantom_stream_count > 0) {
+		is_phantom_pipe = true;
+		DC_LOG_DC(" |    (phantom pipes)     |");
+		for (stream_idx = 0; stream_idx < state->stream_count; stream_idx++) {
+			if (state->stream_status[stream_idx].mall_stream_config.type != SUBVP_MAIN)
+				continue;
+
+			phantom_stream_idx = resource_stream_to_stream_idx(state,
+					state->stream_status[stream_idx].mall_stream_config.paired_stream);
+			otg_master = resource_get_otg_master_for_stream(
+					&state->res_ctx, state->streams[phantom_stream_idx]);
+			if (!otg_master)
+				continue;
+
+			resource_log_pipe_for_stream(dc, state, otg_master, stream_idx, is_phantom_pipe);
+		}
+	}
+	DC_LOG_DC(" |________________________|\n");
+}
+
+static struct pipe_ctx *get_tail_pipe(
+		struct pipe_ctx *head_pipe)
+{
+	struct pipe_ctx *tail_pipe = head_pipe->bottom_pipe;
 
 	while (tail_pipe) {
 		head_pipe = tail_pipe;
@@ -1308,41 +2531,65 @@ static struct pipe_ctx *resource_get_tail_pipe(
 	return head_pipe;
 }
 
-/*
- * A free_pipe for a stream is defined here as a pipe
- * that has no surface attached yet
- */
-static struct pipe_ctx *acquire_free_pipe_for_head(
+static struct pipe_ctx *get_last_opp_head(
+		struct pipe_ctx *opp_head)
+{
+	ASSERT(resource_is_pipe_type(opp_head, OPP_HEAD));
+	while (opp_head->next_odm_pipe)
+		opp_head = opp_head->next_odm_pipe;
+	return opp_head;
+}
+
+static struct pipe_ctx *get_last_dpp_pipe_in_mpcc_combine(
+		struct pipe_ctx *dpp_pipe)
+{
+	ASSERT(resource_is_pipe_type(dpp_pipe, DPP_PIPE));
+	while (dpp_pipe->bottom_pipe &&
+			dpp_pipe->plane_state == dpp_pipe->bottom_pipe->plane_state)
+		dpp_pipe = dpp_pipe->bottom_pipe;
+	return dpp_pipe;
+}
+
+static bool update_pipe_params_after_odm_slice_count_change(
+		struct pipe_ctx *otg_master,
 		struct dc_state *context,
-		const struct resource_pool *pool,
-		struct pipe_ctx *head_pipe)
+		const struct resource_pool *pool)
 {
 	int i;
-	struct resource_context *res_ctx = &context->res_ctx;
+	struct pipe_ctx *pipe;
+	bool result = true;
 
-	if (!head_pipe->plane_state)
-		return head_pipe;
-
-	/* Re-use pipe already acquired for this stream if available*/
-	for (i = pool->pipe_count - 1; i >= 0; i--) {
-		if (res_ctx->pipe_ctx[i].stream == head_pipe->stream &&
-				!res_ctx->pipe_ctx[i].plane_state) {
-			return &res_ctx->pipe_ctx[i];
-		}
+	for (i = 0; i < pool->pipe_count && result; i++) {
+		pipe = &context->res_ctx.pipe_ctx[i];
+		if (pipe->stream == otg_master->stream && pipe->plane_state)
+			result = resource_build_scaling_params(pipe);
 	}
 
-	/*
-	 * At this point we have no re-useable pipe for this stream and we need
-	 * to acquire an idle one to satisfy the request
-	 */
+	if (pool->funcs->build_pipe_pix_clk_params)
+		pool->funcs->build_pipe_pix_clk_params(otg_master);
 
-	if (!pool->funcs->acquire_idle_pipe_for_layer)
-		return NULL;
+	resource_build_test_pattern_params(&context->res_ctx, otg_master);
 
-	return pool->funcs->acquire_idle_pipe_for_layer(context, pool, head_pipe->stream);
+	return result;
+}
+
+static bool update_pipe_params_after_mpc_slice_count_change(
+		const struct dc_plane_state *plane,
+		struct dc_state *context,
+		const struct resource_pool *pool)
+{
+	int i;
+	struct pipe_ctx *pipe;
+	bool result = true;
+
+	for (i = 0; i < pool->pipe_count && result; i++) {
+		pipe = &context->res_ctx.pipe_ctx[i];
+		if (pipe->plane_state == plane)
+			result = resource_build_scaling_params(pipe);
+	}
+	return result;
 }
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 static int acquire_first_split_pipe(
 		struct resource_context *res_ctx,
 		const struct resource_pool *pool,
@@ -1375,117 +2622,529 @@ static int acquire_first_split_pipe(
 			return i;
 		}
 	}
+	return FREE_PIPE_INDEX_NOT_FOUND;
+}
+
+static void update_stream_engine_usage(
+		struct resource_context *res_ctx,
+		const struct resource_pool *pool,
+		struct stream_encoder *stream_enc,
+		bool acquired)
+{
+	int i;
+
+	for (i = 0; i < pool->stream_enc_count; i++) {
+		if (pool->stream_enc[i] == stream_enc)
+			res_ctx->is_stream_enc_acquired[i] = acquired;
+	}
+}
+
+static void update_hpo_dp_stream_engine_usage(
+		struct resource_context *res_ctx,
+		const struct resource_pool *pool,
+		struct hpo_dp_stream_encoder *hpo_dp_stream_enc,
+		bool acquired)
+{
+	int i;
+
+	for (i = 0; i < pool->hpo_dp_stream_enc_count; i++) {
+		if (pool->hpo_dp_stream_enc[i] == hpo_dp_stream_enc)
+			res_ctx->is_hpo_dp_stream_enc_acquired[i] = acquired;
+	}
+}
+
+static inline int find_acquired_hpo_dp_link_enc_for_link(
+		const struct resource_context *res_ctx,
+		const struct dc_link *link)
+{
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(res_ctx->hpo_dp_link_enc_to_link_idx); i++)
+		if (res_ctx->hpo_dp_link_enc_ref_cnts[i] > 0 &&
+				res_ctx->hpo_dp_link_enc_to_link_idx[i] == link->link_index)
+			return i;
+
 	return -1;
 }
-#endif
 
-bool dc_add_plane_to_context(
-		const struct dc *dc,
-		struct dc_stream_state *stream,
-		struct dc_plane_state *plane_state,
-		struct dc_state *context)
+static inline int find_free_hpo_dp_link_enc(const struct resource_context *res_ctx,
+		const struct resource_pool *pool)
 {
 	int i;
-	struct resource_pool *pool = dc->res_pool;
-	struct pipe_ctx *head_pipe, *tail_pipe, *free_pipe;
-	struct dc_stream_status *stream_status = NULL;
 
-	for (i = 0; i < context->stream_count; i++)
-		if (context->streams[i] == stream) {
-			stream_status = &context->stream_status[i];
+	for (i = 0; i < ARRAY_SIZE(res_ctx->hpo_dp_link_enc_ref_cnts); i++)
+		if (res_ctx->hpo_dp_link_enc_ref_cnts[i] == 0)
 			break;
-		}
-	if (stream_status == NULL) {
-		dm_error("Existing stream not found; failed to attach surface!\n");
-		return false;
-	}
 
+	return (i < ARRAY_SIZE(res_ctx->hpo_dp_link_enc_ref_cnts) &&
+			i < pool->hpo_dp_link_enc_count) ? i : -1;
+}
 
-	if (stream_status->plane_count == MAX_SURFACE_NUM) {
-		dm_error("Surface: can not attach plane_state %p! Maximum is: %d\n",
-				plane_state, MAX_SURFACE_NUM);
-		return false;
+static inline void acquire_hpo_dp_link_enc(
+		struct resource_context *res_ctx,
+		unsigned int link_index,
+		int enc_index)
+{
+	res_ctx->hpo_dp_link_enc_to_link_idx[enc_index] = link_index;
+	res_ctx->hpo_dp_link_enc_ref_cnts[enc_index] = 1;
+}
+
+static inline void retain_hpo_dp_link_enc(
+		struct resource_context *res_ctx,
+		int enc_index)
+{
+	res_ctx->hpo_dp_link_enc_ref_cnts[enc_index]++;
+}
+
+static inline void release_hpo_dp_link_enc(
+		struct resource_context *res_ctx,
+		int enc_index)
+{
+	ASSERT(res_ctx->hpo_dp_link_enc_ref_cnts[enc_index] > 0);
+	res_ctx->hpo_dp_link_enc_ref_cnts[enc_index]--;
+}
+
+static bool add_hpo_dp_link_enc_to_ctx(struct resource_context *res_ctx,
+		const struct resource_pool *pool,
+		struct pipe_ctx *pipe_ctx,
+		struct dc_stream_state *stream)
+{
+	int enc_index;
+
+	enc_index = find_acquired_hpo_dp_link_enc_for_link(res_ctx, stream->link);
+
+	if (enc_index >= 0) {
+		retain_hpo_dp_link_enc(res_ctx, enc_index);
+	} else {
+		enc_index = find_free_hpo_dp_link_enc(res_ctx, pool);
+		if (enc_index >= 0)
+			acquire_hpo_dp_link_enc(res_ctx, stream->link->link_index, enc_index);
 	}
 
-	head_pipe = resource_get_head_pipe_for_stream(&context->res_ctx, stream);
+	if (enc_index >= 0)
+		pipe_ctx->link_res.hpo_dp_link_enc = pool->hpo_dp_link_enc[enc_index];
 
-	if (!head_pipe) {
-		dm_error("Head pipe not found for stream_state %p !\n", stream);
-		return false;
+	return pipe_ctx->link_res.hpo_dp_link_enc != NULL;
+}
+
+static void remove_hpo_dp_link_enc_from_ctx(struct resource_context *res_ctx,
+		struct pipe_ctx *pipe_ctx,
+		struct dc_stream_state *stream)
+{
+	int enc_index;
+
+	enc_index = find_acquired_hpo_dp_link_enc_for_link(res_ctx, stream->link);
+
+	if (enc_index >= 0) {
+		release_hpo_dp_link_enc(res_ctx, enc_index);
+		pipe_ctx->link_res.hpo_dp_link_enc = NULL;
 	}
+}
+
+static inline int find_acquired_dio_link_enc_for_link(
+		const struct resource_context *res_ctx,
+		const struct dc_link *link)
+{
+	int i;
 
-	/* retain new surface, but only once per stream */
-	dc_plane_state_retain(plane_state);
+	for (i = 0; i < ARRAY_SIZE(res_ctx->dio_link_enc_ref_cnts); i++)
+		if (res_ctx->dio_link_enc_ref_cnts[i] > 0 &&
+				res_ctx->dio_link_enc_to_link_idx[i] == link->link_index)
+			return i;
 
-	while (head_pipe) {
-		free_pipe = acquire_free_pipe_for_head(context, pool, head_pipe);
+	return -1;
+}
 
-	#if defined(CONFIG_DRM_AMD_DC_DCN)
-		if (!free_pipe) {
-			int pipe_idx = acquire_first_split_pipe(&context->res_ctx, pool, stream);
-			if (pipe_idx >= 0)
-				free_pipe = &context->res_ctx.pipe_ctx[pipe_idx];
-		}
-	#endif
-		if (!free_pipe) {
-			dc_plane_state_release(plane_state);
-			return false;
+static inline int find_fixed_dio_link_enc(const struct dc_link *link)
+{
+	/* the 8b10b dp phy can only use fixed link encoder */
+	return link->eng_id;
+}
+
+static inline int find_free_dio_link_enc(const struct resource_context *res_ctx,
+		const struct dc_link *link, const struct resource_pool *pool, struct dc_stream_state *stream)
+{
+	int i, j = -1;
+	int stream_enc_inst = -1;
+	int enc_count = pool->dig_link_enc_count;
+
+	/* Find stream encoder instance for the stream */
+	if (stream) {
+		for (i = 0; i < pool->pipe_count; i++) {
+			if ((res_ctx->pipe_ctx[i].stream == stream) &&
+				(res_ctx->pipe_ctx[i].stream_res.stream_enc != NULL)) {
+				stream_enc_inst = res_ctx->pipe_ctx[i].stream_res.stream_enc->id;
+				break;
+			}
 		}
+	}
+
+	/* Assign dpia preferred > stream enc instance > available */
+	for (i = 0; i < enc_count; i++) {
+		if (res_ctx->dio_link_enc_ref_cnts[i] == 0) {
+			if (j == -1)
+				j = i;
 
-		free_pipe->plane_state = plane_state;
-
-		if (head_pipe != free_pipe) {
-			tail_pipe = resource_get_tail_pipe(&context->res_ctx, head_pipe);
-			ASSERT(tail_pipe);
-			free_pipe->stream_res.tg = tail_pipe->stream_res.tg;
-			free_pipe->stream_res.abm = tail_pipe->stream_res.abm;
-			free_pipe->stream_res.opp = tail_pipe->stream_res.opp;
-			free_pipe->stream_res.stream_enc = tail_pipe->stream_res.stream_enc;
-			free_pipe->stream_res.audio = tail_pipe->stream_res.audio;
-			free_pipe->clock_source = tail_pipe->clock_source;
-			free_pipe->top_pipe = tail_pipe;
-			tail_pipe->bottom_pipe = free_pipe;
-			if (!free_pipe->next_odm_pipe && tail_pipe->next_odm_pipe && tail_pipe->next_odm_pipe->bottom_pipe) {
-				free_pipe->next_odm_pipe = tail_pipe->next_odm_pipe->bottom_pipe;
-				tail_pipe->next_odm_pipe->bottom_pipe->prev_odm_pipe = free_pipe;
+			if (link->dpia_preferred_eng_id == i) {
+				j = i;
+				break;
 			}
-			if (!free_pipe->prev_odm_pipe && tail_pipe->prev_odm_pipe && tail_pipe->prev_odm_pipe->bottom_pipe) {
-				free_pipe->prev_odm_pipe = tail_pipe->prev_odm_pipe->bottom_pipe;
-				tail_pipe->prev_odm_pipe->bottom_pipe->next_odm_pipe = free_pipe;
+
+			if (stream_enc_inst == i) {
+				j = stream_enc_inst;
 			}
 		}
-		head_pipe = head_pipe->next_odm_pipe;
 	}
-	/* assign new surfaces*/
-	stream_status->plane_states[stream_status->plane_count] = plane_state;
+	return j;
+}
 
-	stream_status->plane_count++;
+static inline void acquire_dio_link_enc(
+		struct resource_context *res_ctx,
+		unsigned int link_index,
+		int enc_index)
+{
+	res_ctx->dio_link_enc_to_link_idx[enc_index] = link_index;
+	res_ctx->dio_link_enc_ref_cnts[enc_index] = 1;
+}
 
-	return true;
+static inline void retain_dio_link_enc(
+		struct resource_context *res_ctx,
+		int enc_index)
+{
+	res_ctx->dio_link_enc_ref_cnts[enc_index]++;
 }
 
-bool dc_remove_plane_from_context(
-		const struct dc *dc,
-		struct dc_stream_state *stream,
+static inline void release_dio_link_enc(
+		struct resource_context *res_ctx,
+		int enc_index)
+{
+	ASSERT(res_ctx->dio_link_enc_ref_cnts[enc_index] > 0);
+	res_ctx->dio_link_enc_ref_cnts[enc_index]--;
+}
+
+static bool is_dio_enc_acquired_by_other_link(const struct dc_link *link,
+		int enc_index,
+		int *link_index)
+{
+	const struct dc *dc  = link->dc;
+	const struct resource_context *res_ctx = &dc->current_state->res_ctx;
+
+	/* pass the link_index that acquired the enc_index */
+	if (res_ctx->dio_link_enc_ref_cnts[enc_index] > 0 &&
+			res_ctx->dio_link_enc_to_link_idx[enc_index] != link->link_index) {
+		*link_index = res_ctx->dio_link_enc_to_link_idx[enc_index];
+		return true;
+	}
+
+	return false;
+}
+
+static void swap_dio_link_enc_to_muxable_ctx(struct dc_state *context,
+		const struct resource_pool *pool,
+		int new_encoder,
+		int old_encoder)
+{
+	struct resource_context *res_ctx = &context->res_ctx;
+	int stream_count = context->stream_count;
+	int i = 0;
+
+	res_ctx->dio_link_enc_ref_cnts[new_encoder] = res_ctx->dio_link_enc_ref_cnts[old_encoder];
+	res_ctx->dio_link_enc_to_link_idx[new_encoder] = res_ctx->dio_link_enc_to_link_idx[old_encoder];
+	res_ctx->dio_link_enc_ref_cnts[old_encoder] = 0;
+
+	for (i = 0; i < stream_count; i++) {
+		struct dc_stream_state *stream = context->streams[i];
+		struct pipe_ctx *pipe_ctx = resource_get_otg_master_for_stream(&context->res_ctx, stream);
+
+		if (pipe_ctx && pipe_ctx->link_res.dio_link_enc == pool->link_encoders[old_encoder])
+			pipe_ctx->link_res.dio_link_enc = pool->link_encoders[new_encoder];
+	}
+}
+
+static bool add_dio_link_enc_to_ctx(const struct dc *dc,
+		struct dc_state *context,
+		const struct resource_pool *pool,
+		struct pipe_ctx *pipe_ctx,
+		struct dc_stream_state *stream)
+{
+	struct resource_context *res_ctx = &context->res_ctx;
+	int enc_index;
+
+	enc_index = find_acquired_dio_link_enc_for_link(res_ctx, stream->link);
+
+	if (enc_index >= 0) {
+		retain_dio_link_enc(res_ctx, enc_index);
+	} else {
+		if (stream->link->is_dig_mapping_flexible)
+			enc_index = find_free_dio_link_enc(res_ctx, stream->link, pool, stream);
+		else {
+			int link_index = 0;
+
+			enc_index = find_fixed_dio_link_enc(stream->link);
+			/* Fixed mapping link can only use its fixed link encoder.
+			 * If the encoder is acquired by other link then get a new free encoder and swap the new
+			 * one into the acquiring link.
+			 */
+			if (enc_index >= 0 && is_dio_enc_acquired_by_other_link(stream->link, enc_index, &link_index)) {
+				int new_enc_index = find_free_dio_link_enc(res_ctx, dc->links[link_index], pool, stream);
+
+				if (new_enc_index >= 0)
+					swap_dio_link_enc_to_muxable_ctx(context, pool, new_enc_index, enc_index);
+				else
+					return false;
+			}
+		}
+
+		if (enc_index >= 0)
+			acquire_dio_link_enc(res_ctx, stream->link->link_index, enc_index);
+	}
+
+	if (enc_index >= 0)
+		pipe_ctx->link_res.dio_link_enc = pool->link_encoders[enc_index];
+
+	return pipe_ctx->link_res.dio_link_enc != NULL;
+}
+
+static void remove_dio_link_enc_from_ctx(struct resource_context *res_ctx,
+		struct pipe_ctx *pipe_ctx,
+		struct dc_stream_state *stream)
+{
+	int enc_index = -1;
+
+	if (stream->link)
+		enc_index = find_acquired_dio_link_enc_for_link(res_ctx, stream->link);
+
+	if (enc_index >= 0) {
+		release_dio_link_enc(res_ctx, enc_index);
+		pipe_ctx->link_res.dio_link_enc = NULL;
+	}
+}
+
+static int get_num_of_free_pipes(const struct resource_pool *pool, const struct dc_state *context)
+{
+	int i;
+	int count = 0;
+
+	for (i = 0; i < pool->pipe_count; i++)
+		if (resource_is_pipe_type(&context->res_ctx.pipe_ctx[i], FREE_PIPE))
+			count++;
+	return count;
+}
+
+enum dc_status resource_add_otg_master_for_stream_output(struct dc_state *new_ctx,
+		const struct resource_pool *pool,
+		struct dc_stream_state *stream)
+{
+	struct dc *dc = stream->ctx->dc;
+
+	return dc->res_pool->funcs->add_stream_to_ctx(dc, new_ctx, stream);
+}
+
+void resource_remove_otg_master_for_stream_output(struct dc_state *context,
+		const struct resource_pool *pool,
+		struct dc_stream_state *stream)
+{
+	struct pipe_ctx *otg_master = resource_get_otg_master_for_stream(
+			&context->res_ctx, stream);
+
+	if (!otg_master)
+		return;
+
+	ASSERT(resource_get_odm_slice_count(otg_master) == 1);
+	ASSERT(otg_master->plane_state == NULL);
+	ASSERT(otg_master->stream_res.stream_enc);
+	update_stream_engine_usage(
+			&context->res_ctx,
+			pool,
+			otg_master->stream_res.stream_enc,
+			false);
+
+	if (stream->ctx->dc->link_srv->dp_is_128b_132b_signal(otg_master)) {
+		update_hpo_dp_stream_engine_usage(
+			&context->res_ctx, pool,
+			otg_master->stream_res.hpo_dp_stream_enc,
+			false);
+		remove_hpo_dp_link_enc_from_ctx(
+				&context->res_ctx, otg_master, stream);
+	}
+
+	if (stream->ctx->dc->config.unify_link_enc_assignment)
+		remove_dio_link_enc_from_ctx(&context->res_ctx, otg_master, stream);
+
+	if (otg_master->stream_res.audio)
+		update_audio_usage(
+			&context->res_ctx,
+			pool,
+			otg_master->stream_res.audio,
+			false);
+
+	resource_unreference_clock_source(&context->res_ctx,
+					  pool,
+					  otg_master->clock_source);
+
+	if (pool->funcs->remove_stream_from_ctx)
+		pool->funcs->remove_stream_from_ctx(
+				stream->ctx->dc, context, stream);
+
+	memset(otg_master, 0, sizeof(*otg_master));
+}
+
+/* For each OPP head of an OTG master, add top plane at plane index 0.
+ *
+ * In the following example, the stream has 2 ODM slices without a top plane.
+ * By adding a plane 0 to OPP heads, we are configuring our hardware to render
+ * plane 0 by using each OPP head's DPP.
+ *
+ *       Inter-pipe Relation (Before Adding Plane)
+ *        __________________________________________________
+ *       |PIPE IDX|   DPP PIPES   | OPP HEADS | OTG MASTER  |
+ *       |        |               | slice 0   |             |
+ *       |   0    |               |blank ----ODM----------- |
+ *       |        |               | slice 1 | |             |
+ *       |   1    |               |blank ---- |             |
+ *       |________|_______________|___________|_____________|
+ *
+ *       Inter-pipe Relation (After Adding Plane)
+ *        __________________________________________________
+ *       |PIPE IDX|   DPP PIPES   | OPP HEADS | OTG MASTER  |
+ *       |        |  plane 0      | slice 0   |             |
+ *       |   0    | -------------------------ODM----------- |
+ *       |        |  plane 0      | slice 1 | |             |
+ *       |   1    | ------------------------- |             |
+ *       |________|_______________|___________|_____________|
+ */
+static bool add_plane_to_opp_head_pipes(struct pipe_ctx *otg_master_pipe,
 		struct dc_plane_state *plane_state,
 		struct dc_state *context)
 {
-	int i;
-	struct dc_stream_status *stream_status = NULL;
-	struct resource_pool *pool = dc->res_pool;
+	struct pipe_ctx *opp_head_pipe = otg_master_pipe;
 
-	for (i = 0; i < context->stream_count; i++)
-		if (context->streams[i] == stream) {
-			stream_status = &context->stream_status[i];
-			break;
+	while (opp_head_pipe) {
+		if (opp_head_pipe->plane_state) {
+			ASSERT(0);
+			return false;
 		}
+		opp_head_pipe->plane_state = plane_state;
+		opp_head_pipe = opp_head_pipe->next_odm_pipe;
+	}
+
+	return true;
+}
+
+/* For each OPP head of an OTG master, acquire a secondary DPP pipe and add
+ * the plane. So the plane is added to all ODM slices associated with the OTG
+ * master pipe in the bottom layer.
+ *
+ * In the following example, the stream has 2 ODM slices and a top plane 0.
+ * By acquiring secondary DPP pipes and adding a plane 1, we are configuring our
+ * hardware to render the plane 1 by acquiring a new pipe for each ODM slice and
+ * render plane 1 using new pipes' DPP in the Z axis below plane 0.
+ *
+ *       Inter-pipe Relation (Before Adding Plane)
+ *        __________________________________________________
+ *       |PIPE IDX|   DPP PIPES   | OPP HEADS | OTG MASTER  |
+ *       |        |  plane 0      | slice 0   |             |
+ *       |   0    | -------------------------ODM----------- |
+ *       |        |  plane 0      | slice 1 | |             |
+ *       |   1    | ------------------------- |             |
+ *       |________|_______________|___________|_____________|
+ *
+ *       Inter-pipe Relation (After Acquiring and Adding Plane)
+ *        __________________________________________________
+ *       |PIPE IDX|   DPP PIPES   | OPP HEADS | OTG MASTER  |
+ *       |        |  plane 0      | slice 0   |             |
+ *       |   0    | -------------MPC---------ODM----------- |
+ *       |        |  plane 1    | |         | |             |
+ *       |   2    | ------------- |         | |             |
+ *       |        |  plane 0      | slice 1 | |             |
+ *       |   1    | -------------MPC--------- |             |
+ *       |        |  plane 1    | |           |             |
+ *       |   3    | ------------- |           |             |
+ *       |________|_______________|___________|_____________|
+ */
+static bool acquire_secondary_dpp_pipes_and_add_plane(
+		struct pipe_ctx *otg_master_pipe,
+		struct dc_plane_state *plane_state,
+		struct dc_state *new_ctx,
+		struct dc_state *cur_ctx,
+		struct resource_pool *pool)
+{
+	struct pipe_ctx *sec_pipe, *tail_pipe;
+	struct pipe_ctx *opp_heads[MAX_PIPES];
+	int opp_head_count;
+	int i;
+
+	if (!pool->funcs->acquire_free_pipe_as_secondary_dpp_pipe) {
+		ASSERT(0);
+		return false;
+	}
 
-	if (stream_status == NULL) {
-		dm_error("Existing stream not found; failed to remove plane.\n");
+	opp_head_count = resource_get_opp_heads_for_otg_master(otg_master_pipe,
+			&new_ctx->res_ctx, opp_heads);
+	if (get_num_of_free_pipes(pool, new_ctx) < opp_head_count)
+		/* not enough free pipes */
 		return false;
+
+	for (i = 0; i < opp_head_count; i++) {
+		sec_pipe = pool->funcs->acquire_free_pipe_as_secondary_dpp_pipe(
+				cur_ctx,
+				new_ctx,
+				pool,
+				opp_heads[i]);
+		ASSERT(sec_pipe);
+		sec_pipe->plane_state = plane_state;
+
+		/* establish pipe relationship */
+		tail_pipe = get_tail_pipe(opp_heads[i]);
+		tail_pipe->bottom_pipe = sec_pipe;
+		sec_pipe->top_pipe = tail_pipe;
+		sec_pipe->bottom_pipe = NULL;
+		if (tail_pipe->prev_odm_pipe) {
+			ASSERT(tail_pipe->prev_odm_pipe->bottom_pipe);
+			sec_pipe->prev_odm_pipe = tail_pipe->prev_odm_pipe->bottom_pipe;
+			tail_pipe->prev_odm_pipe->bottom_pipe->next_odm_pipe = sec_pipe;
+		} else {
+			sec_pipe->prev_odm_pipe = NULL;
+		}
 	}
+	return true;
+}
+
+bool resource_append_dpp_pipes_for_plane_composition(
+		struct dc_state *new_ctx,
+		struct dc_state *cur_ctx,
+		struct resource_pool *pool,
+		struct pipe_ctx *otg_master_pipe,
+		struct dc_plane_state *plane_state)
+{
+	bool success;
+
+	if (otg_master_pipe->plane_state == NULL)
+		success = add_plane_to_opp_head_pipes(otg_master_pipe,
+				plane_state, new_ctx);
+	else
+		success = acquire_secondary_dpp_pipes_and_add_plane(
+				otg_master_pipe, plane_state, new_ctx,
+				cur_ctx, pool);
+	if (success) {
+		/* when appending a plane mpc slice count changes from 0 to 1 */
+		success = update_pipe_params_after_mpc_slice_count_change(
+				plane_state, new_ctx, pool);
+		if (!success)
+			resource_remove_dpp_pipes_for_plane_composition(new_ctx,
+					pool, plane_state);
+	}
+
+	return success;
+}
+
+void resource_remove_dpp_pipes_for_plane_composition(
+		struct dc_state *context,
+		const struct resource_pool *pool,
+		const struct dc_plane_state *plane_state)
+{
+	int i;
 
-	/* release pipe for plane*/
 	for (i = pool->pipe_count - 1; i >= 0; i--) {
 		struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i];
 
@@ -1510,110 +3169,340 @@ bool dc_remove_plane_from_context(
 				memset(pipe_ctx, 0, sizeof(*pipe_ctx));
 		}
 	}
+}
 
+/*
+ * Increase ODM slice count by 1 by acquiring pipes and adding a new ODM slice
+ * at the last index.
+ * return - true if a new ODM slice is added and required pipes are acquired.
+ * false if new_ctx is no longer a valid state after new ODM slice is added.
+ *
+ * This is achieved by duplicating MPC blending tree from previous ODM slice.
+ * In the following example, we have a single MPC tree and 1 ODM slice 0. We
+ * want to add a new odm slice by duplicating the MPC blending tree and add
+ * ODM slice 1.
+ *
+ *       Inter-pipe Relation (Before Acquiring and Adding ODM Slice)
+ *        __________________________________________________
+ *       |PIPE IDX|   DPP PIPES   | OPP HEADS | OTG MASTER  |
+ *       |        |  plane 0      | slice 0   |             |
+ *       |   0    | -------------MPC---------ODM----------- |
+ *       |        |  plane 1    | |           |             |
+ *       |   1    | ------------- |           |             |
+ *       |________|_______________|___________|_____________|
+ *
+ *       Inter-pipe Relation (After Acquiring and Adding ODM Slice)
+ *        __________________________________________________
+ *       |PIPE IDX|   DPP PIPES   | OPP HEADS | OTG MASTER  |
+ *       |        |  plane 0      | slice 0   |             |
+ *       |   0    | -------------MPC---------ODM----------- |
+ *       |        |  plane 1    | |         | |             |
+ *       |   1    | ------------- |         | |             |
+ *       |        |  plane 0      | slice 1 | |             |
+ *       |   2    | -------------MPC--------- |             |
+ *       |        |  plane 1    | |           |             |
+ *       |   3    | ------------- |           |             |
+ *       |________|_______________|___________|_____________|
+ */
+static bool acquire_pipes_and_add_odm_slice(
+		struct pipe_ctx *otg_master_pipe,
+		struct dc_state *new_ctx,
+		const struct dc_state *cur_ctx,
+		const struct resource_pool *pool)
+{
+	struct pipe_ctx *last_opp_head = get_last_opp_head(otg_master_pipe);
+	struct pipe_ctx *new_opp_head;
+	struct pipe_ctx *last_top_dpp_pipe, *last_bottom_dpp_pipe,
+			*new_top_dpp_pipe, *new_bottom_dpp_pipe;
 
-	for (i = 0; i < stream_status->plane_count; i++) {
-		if (stream_status->plane_states[i] == plane_state) {
-
-			dc_plane_state_release(stream_status->plane_states[i]);
-			break;
-		}
-	}
-
-	if (i == stream_status->plane_count) {
-		dm_error("Existing plane_state not found; failed to detach it!\n");
+	if (!pool->funcs->acquire_free_pipe_as_secondary_opp_head) {
+		ASSERT(0);
 		return false;
 	}
+	new_opp_head = pool->funcs->acquire_free_pipe_as_secondary_opp_head(
+					cur_ctx, new_ctx, pool,
+					otg_master_pipe);
+	if (!new_opp_head)
+		return false;
 
-	stream_status->plane_count--;
-
-	/* Start at the plane we've just released, and move all the planes one index forward to "trim" the array */
-	for (; i < stream_status->plane_count; i++)
-		stream_status->plane_states[i] = stream_status->plane_states[i + 1];
+	last_opp_head->next_odm_pipe = new_opp_head;
+	new_opp_head->prev_odm_pipe = last_opp_head;
+	new_opp_head->next_odm_pipe = NULL;
+	new_opp_head->plane_state = last_opp_head->plane_state;
+	last_top_dpp_pipe = last_opp_head;
+	new_top_dpp_pipe = new_opp_head;
+
+	while (last_top_dpp_pipe->bottom_pipe) {
+		last_bottom_dpp_pipe = last_top_dpp_pipe->bottom_pipe;
+		new_bottom_dpp_pipe = pool->funcs->acquire_free_pipe_as_secondary_dpp_pipe(
+				cur_ctx, new_ctx, pool,
+				new_opp_head);
+		if (!new_bottom_dpp_pipe)
+			return false;
 
-	stream_status->plane_states[stream_status->plane_count] = NULL;
+		new_bottom_dpp_pipe->plane_state = last_bottom_dpp_pipe->plane_state;
+		new_top_dpp_pipe->bottom_pipe = new_bottom_dpp_pipe;
+		new_bottom_dpp_pipe->top_pipe = new_top_dpp_pipe;
+		last_bottom_dpp_pipe->next_odm_pipe = new_bottom_dpp_pipe;
+		new_bottom_dpp_pipe->prev_odm_pipe = last_bottom_dpp_pipe;
+		new_bottom_dpp_pipe->next_odm_pipe = NULL;
+		last_top_dpp_pipe = last_bottom_dpp_pipe;
+	}
 
 	return true;
 }
 
-bool dc_rem_all_planes_for_stream(
-		const struct dc *dc,
-		struct dc_stream_state *stream,
-		struct dc_state *context)
+/*
+ * Decrease ODM slice count by 1 by releasing pipes and removing the ODM slice
+ * at the last index.
+ * return - true if the last ODM slice is removed and related pipes are
+ * released. false if there is no removable ODM slice.
+ *
+ * In the following example, we have 2 MPC trees and ODM slice 0 and slice 1.
+ * We want to remove the last ODM i.e slice 1. We are releasing secondary DPP
+ * pipe 3 and OPP head pipe 2.
+ *
+ *       Inter-pipe Relation (Before Releasing and Removing ODM Slice)
+ *        __________________________________________________
+ *       |PIPE IDX|   DPP PIPES   | OPP HEADS | OTG MASTER  |
+ *       |        |  plane 0      | slice 0   |             |
+ *       |   0    | -------------MPC---------ODM----------- |
+ *       |        |  plane 1    | |         | |             |
+ *       |   1    | ------------- |         | |             |
+ *       |        |  plane 0      | slice 1 | |             |
+ *       |   2    | -------------MPC--------- |             |
+ *       |        |  plane 1    | |           |             |
+ *       |   3    | ------------- |           |             |
+ *       |________|_______________|___________|_____________|
+ *
+ *       Inter-pipe Relation (After Releasing and Removing ODM Slice)
+ *        __________________________________________________
+ *       |PIPE IDX|   DPP PIPES   | OPP HEADS | OTG MASTER  |
+ *       |        |  plane 0      | slice 0   |             |
+ *       |   0    | -------------MPC---------ODM----------- |
+ *       |        |  plane 1    | |           |             |
+ *       |   1    | ------------- |           |             |
+ *       |________|_______________|___________|_____________|
+ */
+static bool release_pipes_and_remove_odm_slice(
+		struct pipe_ctx *otg_master_pipe,
+		struct dc_state *context,
+		const struct resource_pool *pool)
 {
-	int i, old_plane_count;
-	struct dc_stream_status *stream_status = NULL;
-	struct dc_plane_state *del_planes[MAX_SURFACE_NUM] = { 0 };
+	struct pipe_ctx *last_opp_head = get_last_opp_head(otg_master_pipe);
+	struct pipe_ctx *tail_pipe = get_tail_pipe(last_opp_head);
 
-	for (i = 0; i < context->stream_count; i++)
-			if (context->streams[i] == stream) {
-				stream_status = &context->stream_status[i];
-				break;
-			}
+	if (!pool->funcs->release_pipe) {
+		ASSERT(0);
+		return false;
+	}
 
-	if (stream_status == NULL) {
-		dm_error("Existing stream %p not found!\n", stream);
+	if (resource_is_pipe_type(last_opp_head, OTG_MASTER))
 		return false;
+
+	while (tail_pipe->top_pipe) {
+		tail_pipe->prev_odm_pipe->next_odm_pipe = NULL;
+		tail_pipe = tail_pipe->top_pipe;
+		pool->funcs->release_pipe(context, tail_pipe->bottom_pipe, pool);
+		tail_pipe->bottom_pipe = NULL;
 	}
+	last_opp_head->prev_odm_pipe->next_odm_pipe = NULL;
+	pool->funcs->release_pipe(context, last_opp_head, pool);
 
-	old_plane_count = stream_status->plane_count;
+	return true;
+}
 
-	for (i = 0; i < old_plane_count; i++)
-		del_planes[i] = stream_status->plane_states[i];
+/*
+ * Increase MPC slice count by 1 by acquiring a new DPP pipe and add it as the
+ * last MPC slice of the plane associated with dpp_pipe.
+ *
+ * return - true if a new MPC slice is added and required pipes are acquired.
+ * false if new_ctx is no longer a valid state after new MPC slice is added.
+ *
+ * In the following example, we add a new MPC slice for plane 0 into the
+ * new_ctx. To do so we pass pipe 0 as dpp_pipe. The function acquires a new DPP
+ * pipe 2 for plane 0 as the bottom most pipe for plane 0.
+ *
+ *       Inter-pipe Relation (Before Acquiring and Adding MPC Slice)
+ *        __________________________________________________
+ *       |PIPE IDX|   DPP PIPES   | OPP HEADS | OTG MASTER  |
+ *       |        |  plane 0      |           |             |
+ *       |   0    | -------------MPC----------------------- |
+ *       |        |  plane 1    | |           |             |
+ *       |   1    | ------------- |           |             |
+ *       |________|_______________|___________|_____________|
+ *
+ *       Inter-pipe Relation (After Acquiring and Adding MPC Slice)
+ *        __________________________________________________
+ *       |PIPE IDX|   DPP PIPES   | OPP HEADS | OTG MASTER  |
+ *       |        |  plane 0      |           |             |
+ *       |   0    | -------------MPC----------------------- |
+ *       |        |  plane 0    | |           |             |
+ *       |   2    | ------------- |           |             |
+ *       |        |  plane 1    | |           |             |
+ *       |   1    | ------------- |           |             |
+ *       |________|_______________|___________|_____________|
+ */
+static bool acquire_dpp_pipe_and_add_mpc_slice(
+		struct pipe_ctx *dpp_pipe,
+		struct dc_state *new_ctx,
+		const struct dc_state *cur_ctx,
+		const struct resource_pool *pool)
+{
+	struct pipe_ctx *last_dpp_pipe =
+			get_last_dpp_pipe_in_mpcc_combine(dpp_pipe);
+	struct pipe_ctx *opp_head = resource_get_opp_head(dpp_pipe);
+	struct pipe_ctx *new_dpp_pipe;
 
-	for (i = 0; i < old_plane_count; i++)
-		if (!dc_remove_plane_from_context(dc, stream, del_planes[i], context))
-			return false;
+	if (!pool->funcs->acquire_free_pipe_as_secondary_dpp_pipe) {
+		ASSERT(0);
+		return false;
+	}
+	new_dpp_pipe = pool->funcs->acquire_free_pipe_as_secondary_dpp_pipe(
+			cur_ctx, new_ctx, pool, opp_head);
+	if (!new_dpp_pipe || resource_get_odm_slice_count(dpp_pipe) > 1)
+		return false;
+
+	new_dpp_pipe->bottom_pipe = last_dpp_pipe->bottom_pipe;
+	if (new_dpp_pipe->bottom_pipe)
+		new_dpp_pipe->bottom_pipe->top_pipe = new_dpp_pipe;
+	new_dpp_pipe->top_pipe = last_dpp_pipe;
+	last_dpp_pipe->bottom_pipe = new_dpp_pipe;
+	new_dpp_pipe->plane_state = last_dpp_pipe->plane_state;
 
 	return true;
 }
 
-static bool add_all_planes_for_stream(
-		const struct dc *dc,
-		struct dc_stream_state *stream,
-		const struct dc_validation_set set[],
-		int set_count,
-		struct dc_state *context)
+/*
+ * Reduce MPC slice count by 1 by releasing the bottom DPP pipe in MPCC combine
+ * with dpp_pipe and removing last MPC slice of the plane associated with
+ * dpp_pipe.
+ *
+ * return - true if the last MPC slice of the plane associated with dpp_pipe is
+ * removed and last DPP pipe in MPCC combine with dpp_pipe is released.
+ * false if there is no removable MPC slice.
+ *
+ * In the following example, we remove an MPC slice for plane 0 from the
+ * context. To do so we pass pipe 0 as dpp_pipe. The function releases pipe 1 as
+ * it is the last pipe for plane 0.
+ *
+ *       Inter-pipe Relation (Before Releasing and Removing MPC Slice)
+ *        __________________________________________________
+ *       |PIPE IDX|   DPP PIPES   | OPP HEADS | OTG MASTER  |
+ *       |        |  plane 0      |           |             |
+ *       |   0    | -------------MPC----------------------- |
+ *       |        |  plane 0    | |           |             |
+ *       |   1    | ------------- |           |             |
+ *       |        |  plane 1    | |           |             |
+ *       |   2    | ------------- |           |             |
+ *       |________|_______________|___________|_____________|
+ *
+ *       Inter-pipe Relation (After Releasing and Removing MPC Slice)
+ *        __________________________________________________
+ *       |PIPE IDX|   DPP PIPES   | OPP HEADS | OTG MASTER  |
+ *       |        |  plane 0      |           |             |
+ *       |   0    | -------------MPC----------------------- |
+ *       |        |  plane 1    | |           |             |
+ *       |   2    | ------------- |           |             |
+ *       |________|_______________|___________|_____________|
+ */
+static bool release_dpp_pipe_and_remove_mpc_slice(
+		struct pipe_ctx *dpp_pipe,
+		struct dc_state *context,
+		const struct resource_pool *pool)
 {
-	int i, j;
-
-	for (i = 0; i < set_count; i++)
-		if (set[i].stream == stream)
-			break;
+	struct pipe_ctx *last_dpp_pipe =
+			get_last_dpp_pipe_in_mpcc_combine(dpp_pipe);
 
-	if (i == set_count) {
-		dm_error("Stream %p not found in set!\n", stream);
+	if (!pool->funcs->release_pipe) {
+		ASSERT(0);
 		return false;
 	}
 
-	for (j = 0; j < set[i].plane_count; j++)
-		if (!dc_add_plane_to_context(dc, stream, set[i].plane_states[j], context))
-			return false;
+	if (resource_is_pipe_type(last_dpp_pipe, OPP_HEAD) ||
+			resource_get_odm_slice_count(dpp_pipe) > 1)
+		return false;
+
+	last_dpp_pipe->top_pipe->bottom_pipe = last_dpp_pipe->bottom_pipe;
+	if (last_dpp_pipe->bottom_pipe)
+		last_dpp_pipe->bottom_pipe->top_pipe = last_dpp_pipe->top_pipe;
+	pool->funcs->release_pipe(context, last_dpp_pipe, pool);
 
 	return true;
 }
 
-bool dc_add_all_planes_for_stream(
-		const struct dc *dc,
-		struct dc_stream_state *stream,
-		struct dc_plane_state * const *plane_states,
-		int plane_count,
-		struct dc_state *context)
+bool resource_update_pipes_for_stream_with_slice_count(
+		struct dc_state *new_ctx,
+		const struct dc_state *cur_ctx,
+		const struct resource_pool *pool,
+		const struct dc_stream_state *stream,
+		int new_slice_count)
 {
-	struct dc_validation_set set;
 	int i;
+	struct pipe_ctx *otg_master = resource_get_otg_master_for_stream(
+			&new_ctx->res_ctx, stream);
+	int cur_slice_count;
+	bool result = true;
+
+	if (!otg_master)
+		return false;
 
-	set.stream = stream;
-	set.plane_count = plane_count;
+	cur_slice_count = resource_get_odm_slice_count(otg_master);
 
-	for (i = 0; i < plane_count; i++)
-		set.plane_states[i] = plane_states[i];
+	if (new_slice_count == cur_slice_count)
+		return result;
 
-	return add_all_planes_for_stream(dc, stream, &set, 1, context);
+	if (new_slice_count > cur_slice_count)
+		for (i = 0; i < new_slice_count - cur_slice_count && result; i++)
+			result = acquire_pipes_and_add_odm_slice(
+					otg_master, new_ctx, cur_ctx, pool);
+	else
+		for (i = 0; i < cur_slice_count - new_slice_count && result; i++)
+			result = release_pipes_and_remove_odm_slice(
+					otg_master, new_ctx, pool);
+	if (result)
+		result = update_pipe_params_after_odm_slice_count_change(
+				otg_master, new_ctx, pool);
+	return result;
 }
 
-static bool is_timing_changed(struct dc_stream_state *cur_stream,
-		struct dc_stream_state *new_stream)
+bool resource_update_pipes_for_plane_with_slice_count(
+		struct dc_state *new_ctx,
+		const struct dc_state *cur_ctx,
+		const struct resource_pool *pool,
+		const struct dc_plane_state *plane,
+		int new_slice_count)
+{
+	int i;
+	int dpp_pipe_count;
+	int cur_slice_count;
+	struct pipe_ctx *dpp_pipes[MAX_PIPES] = {0};
+	bool result = true;
+
+	dpp_pipe_count = resource_get_dpp_pipes_for_plane(plane,
+			&new_ctx->res_ctx, dpp_pipes);
+	ASSERT(dpp_pipe_count > 0);
+	cur_slice_count = resource_get_mpc_slice_count(dpp_pipes[0]);
+
+	if (new_slice_count == cur_slice_count)
+		return result;
+
+	if (new_slice_count > cur_slice_count)
+		for (i = 0; i < new_slice_count - cur_slice_count && result; i++)
+			result = acquire_dpp_pipe_and_add_mpc_slice(
+					dpp_pipes[0], new_ctx, cur_ctx, pool);
+	else
+		for (i = 0; i < cur_slice_count - new_slice_count && result; i++)
+			result = release_dpp_pipe_and_remove_mpc_slice(
+					dpp_pipes[0], new_ctx, pool);
+	if (result)
+		result = update_pipe_params_after_mpc_slice_count_change(
+				dpp_pipes[0]->plane_state, new_ctx, pool);
+	return result;
+}
+
+bool dc_is_timing_changed(struct dc_stream_state *cur_stream,
+		       struct dc_stream_state *new_stream)
 {
 	if (cur_stream == NULL)
 		return true;
@@ -1637,7 +3526,10 @@ static bool are_stream_backends_same(
 	if (stream_a == NULL || stream_b == NULL)
 		return false;
 
-	if (is_timing_changed(stream_a, stream_b))
+	if (dc_is_timing_changed(stream_a, stream_b))
+		return false;
+
+	if (stream_a->signal != stream_b->signal)
 		return false;
 
 	if (stream_a->dpms_off != stream_b->dpms_off)
@@ -1657,6 +3549,8 @@ static bool are_stream_backends_same(
 bool dc_is_stream_unchanged(
 	struct dc_stream_state *old_stream, struct dc_stream_state *stream)
 {
+	if (!old_stream || !stream)
+		return false;
 
 	if (!are_stream_backends_same(old_stream, stream))
 		return false;
@@ -1664,8 +3558,8 @@ bool dc_is_stream_unchanged(
 	if (old_stream->ignore_msa_timing_param != stream->ignore_msa_timing_param)
 		return false;
 
-	// Only Have Audio left to check whether it is same or not. This is a corner case for Tiled sinks
-	if (old_stream->audio_info.mode_count != stream->audio_info.mode_count)
+	/*compare audio info*/
+	if (memcmp(&old_stream->audio_info, &stream->audio_info, sizeof(stream->audio_info)) != 0)
 		return false;
 
 	return true;
@@ -1674,8 +3568,8 @@ bool dc_is_stream_unchanged(
 /*
  * dc_is_stream_scaling_unchanged() - Compare scaling rectangles of two streams.
  */
-bool dc_is_stream_scaling_unchanged(
-	struct dc_stream_state *old_stream, struct dc_stream_state *stream)
+bool dc_is_stream_scaling_unchanged(struct dc_stream_state *old_stream,
+				    struct dc_stream_state *stream)
 {
 	if (old_stream == stream)
 		return true;
@@ -1696,36 +3590,6 @@ bool dc_is_stream_scaling_unchanged(
 	return true;
 }
 
-static void update_stream_engine_usage(
-		struct resource_context *res_ctx,
-		const struct resource_pool *pool,
-		struct stream_encoder *stream_enc,
-		bool acquired)
-{
-	int i;
-
-	for (i = 0; i < pool->stream_enc_count; i++) {
-		if (pool->stream_enc[i] == stream_enc)
-			res_ctx->is_stream_enc_acquired[i] = acquired;
-	}
-}
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-static void update_hpo_dp_stream_engine_usage(
-		struct resource_context *res_ctx,
-		const struct resource_pool *pool,
-		struct hpo_dp_stream_encoder *hpo_dp_stream_enc,
-		bool acquired)
-{
-	int i;
-
-	for (i = 0; i < pool->hpo_dp_stream_enc_count; i++) {
-		if (pool->hpo_dp_stream_enc[i] == hpo_dp_stream_enc)
-			res_ctx->is_hpo_dp_stream_enc_acquired[i] = acquired;
-	}
-}
-#endif
-
 /* TODO: release audio object */
 void update_audio_usage(
 		struct resource_context *res_ctx,
@@ -1740,37 +3604,6 @@ void update_audio_usage(
 	}
 }
 
-static int acquire_first_free_pipe(
-		struct resource_context *res_ctx,
-		const struct resource_pool *pool,
-		struct dc_stream_state *stream)
-{
-	int i;
-
-	for (i = 0; i < pool->pipe_count; i++) {
-		if (!res_ctx->pipe_ctx[i].stream) {
-			struct pipe_ctx *pipe_ctx = &res_ctx->pipe_ctx[i];
-
-			pipe_ctx->stream_res.tg = pool->timing_generators[i];
-			pipe_ctx->plane_res.mi = pool->mis[i];
-			pipe_ctx->plane_res.hubp = pool->hubps[i];
-			pipe_ctx->plane_res.ipp = pool->ipps[i];
-			pipe_ctx->plane_res.xfm = pool->transforms[i];
-			pipe_ctx->plane_res.dpp = pool->dpps[i];
-			pipe_ctx->stream_res.opp = pool->opps[i];
-			if (pool->dpps[i])
-				pipe_ctx->plane_res.mpcc_inst = pool->dpps[i]->inst;
-			pipe_ctx->pipe_idx = i;
-
-
-			pipe_ctx->stream = stream;
-			return i;
-		}
-	}
-	return -1;
-}
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 static struct hpo_dp_stream_encoder *find_first_free_match_hpo_dp_stream_enc_for_link(
 		struct resource_context *res_ctx,
 		const struct resource_pool *pool,
@@ -1788,7 +3621,6 @@ static struct hpo_dp_stream_encoder *find_first_free_match_hpo_dp_stream_enc_for
 
 	return NULL;
 }
-#endif
 
 static struct audio *find_first_free_audio(
 		struct resource_context *res_ctx,
@@ -1798,6 +3630,9 @@ static struct audio *find_first_free_audio(
 {
 	int i, available_audio_count;
 
+	if (id == ENGINE_ID_UNKNOWN)
+		return NULL;
+
 	available_audio_count = pool->audio_count;
 
 	for (i = 0; i < available_audio_count; i++) {
@@ -1819,123 +3654,7 @@ static struct audio *find_first_free_audio(
 			return pool->audios[i];
 		}
 	}
-	return 0;
-}
-
-/*
- * dc_add_stream_to_ctx() - Add a new dc_stream_state to a dc_state.
- */
-enum dc_status dc_add_stream_to_ctx(
-		struct dc *dc,
-		struct dc_state *new_ctx,
-		struct dc_stream_state *stream)
-{
-	enum dc_status res;
-	DC_LOGGER_INIT(dc->ctx->logger);
-
-	if (new_ctx->stream_count >= dc->res_pool->timing_generator_count) {
-		DC_LOG_WARNING("Max streams reached, can't add stream %p !\n", stream);
-		return DC_ERROR_UNEXPECTED;
-	}
-
-	new_ctx->streams[new_ctx->stream_count] = stream;
-	dc_stream_retain(stream);
-	new_ctx->stream_count++;
-
-	res = dc->res_pool->funcs->add_stream_to_ctx(dc, new_ctx, stream);
-	if (res != DC_OK)
-		DC_LOG_WARNING("Adding stream %p to context failed with err %d!\n", stream, res);
-
-	return res;
-}
-
-/*
- * dc_remove_stream_from_ctx() - Remove a stream from a dc_state.
- */
-enum dc_status dc_remove_stream_from_ctx(
-			struct dc *dc,
-			struct dc_state *new_ctx,
-			struct dc_stream_state *stream)
-{
-	int i;
-	struct dc_context *dc_ctx = dc->ctx;
-	struct pipe_ctx *del_pipe = resource_get_head_pipe_for_stream(&new_ctx->res_ctx, stream);
-	struct pipe_ctx *odm_pipe;
-
-	if (!del_pipe) {
-		DC_ERROR("Pipe not found for stream %p !\n", stream);
-		return DC_ERROR_UNEXPECTED;
-	}
-
-	odm_pipe = del_pipe->next_odm_pipe;
-
-	/* Release primary pipe */
-	ASSERT(del_pipe->stream_res.stream_enc);
-	update_stream_engine_usage(
-			&new_ctx->res_ctx,
-				dc->res_pool,
-			del_pipe->stream_res.stream_enc,
-			false);
-	/* Release link encoder from stream in new dc_state. */
-	if (dc->res_pool->funcs->link_enc_unassign)
-		dc->res_pool->funcs->link_enc_unassign(new_ctx, del_pipe->stream);
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	if (is_dp_128b_132b_signal(del_pipe)) {
-		update_hpo_dp_stream_engine_usage(
-			&new_ctx->res_ctx, dc->res_pool,
-			del_pipe->stream_res.hpo_dp_stream_enc,
-			false);
-	}
-#endif
-
-	if (del_pipe->stream_res.audio)
-		update_audio_usage(
-			&new_ctx->res_ctx,
-			dc->res_pool,
-			del_pipe->stream_res.audio,
-			false);
-
-	resource_unreference_clock_source(&new_ctx->res_ctx,
-					  dc->res_pool,
-					  del_pipe->clock_source);
-
-	if (dc->res_pool->funcs->remove_stream_from_ctx)
-		dc->res_pool->funcs->remove_stream_from_ctx(dc, new_ctx, stream);
-
-	while (odm_pipe) {
-		struct pipe_ctx *next_odm_pipe = odm_pipe->next_odm_pipe;
-
-		memset(odm_pipe, 0, sizeof(*odm_pipe));
-		odm_pipe = next_odm_pipe;
-	}
-	memset(del_pipe, 0, sizeof(*del_pipe));
-
-	for (i = 0; i < new_ctx->stream_count; i++)
-		if (new_ctx->streams[i] == stream)
-			break;
-
-	if (new_ctx->streams[i] != stream) {
-		DC_ERROR("Context doesn't have stream %p !\n", stream);
-		return DC_ERROR_UNEXPECTED;
-	}
-
-	dc_stream_release(new_ctx->streams[i]);
-	new_ctx->stream_count--;
-
-	/* Trim back arrays */
-	for (; i < new_ctx->stream_count; i++) {
-		new_ctx->streams[i] = new_ctx->streams[i + 1];
-		new_ctx->stream_status[i] = new_ctx->stream_status[i + 1];
-	}
-
-	new_ctx->streams[new_ctx->stream_count] = NULL;
-	memset(
-			&new_ctx->stream_status[new_ctx->stream_count],
-			0,
-			sizeof(new_ctx->stream_status[0]));
-
-	return DC_OK;
+	return NULL;
 }
 
 static struct dc_stream_state *find_pll_sharable_stream(
@@ -1978,10 +3697,13 @@ static int get_norm_pix_clk(const struct dc_crtc_timing *timing)
 			break;
 		case COLOR_DEPTH_121212:
 			normalized_pix_clk = (pix_clk * 36) / 24;
-		break;
+			break;
+		case COLOR_DEPTH_141414:
+			normalized_pix_clk = (pix_clk * 42) / 24;
+			break;
 		case COLOR_DEPTH_161616:
 			normalized_pix_clk = (pix_clk * 48) / 24;
-		break;
+			break;
 		default:
 			ASSERT(0);
 		break;
@@ -2011,6 +3733,8 @@ static int acquire_resource_from_hw_enabled_state(
 {
 	struct dc_link *link = stream->link;
 	unsigned int i, inst, tg_inst = 0;
+	uint32_t numPipes = 1;
+	uint32_t id_src[4] = {0};
 
 	/* Check for enabled DIG to identify enabled display */
 	if (!link->link_enc->funcs->is_dig_enabled(link->link_enc))
@@ -2040,56 +3764,194 @@ static int acquire_resource_from_hw_enabled_state(
 		struct pipe_ctx *pipe_ctx = &res_ctx->pipe_ctx[tg_inst];
 
 		pipe_ctx->stream_res.tg = pool->timing_generators[tg_inst];
-		pipe_ctx->plane_res.mi = pool->mis[tg_inst];
-		pipe_ctx->plane_res.hubp = pool->hubps[tg_inst];
-		pipe_ctx->plane_res.ipp = pool->ipps[tg_inst];
-		pipe_ctx->plane_res.xfm = pool->transforms[tg_inst];
-		pipe_ctx->plane_res.dpp = pool->dpps[tg_inst];
-		pipe_ctx->stream_res.opp = pool->opps[tg_inst];
+		id_src[0] = tg_inst;
+
+		if (pipe_ctx->stream_res.tg->funcs->get_optc_source)
+			pipe_ctx->stream_res.tg->funcs->get_optc_source(pipe_ctx->stream_res.tg,
+						&numPipes, &id_src[0], &id_src[1]);
+
+		if (id_src[0] == 0xf && id_src[1] == 0xf) {
+			id_src[0] = tg_inst;
+			numPipes = 1;
+		}
 
-		if (pool->dpps[tg_inst]) {
-			pipe_ctx->plane_res.mpcc_inst = pool->dpps[tg_inst]->inst;
+		for (i = 0; i < numPipes; i++) {
+			//Check if src id invalid
+			if (id_src[i] == 0xf)
+				return -1;
 
-			// Read DPP->MPCC->OPP Pipe from HW State
-			if (pool->mpc->funcs->read_mpcc_state) {
-				struct mpcc_state s = {0};
+			pipe_ctx = &res_ctx->pipe_ctx[id_src[i]];
 
-				pool->mpc->funcs->read_mpcc_state(pool->mpc, pipe_ctx->plane_res.mpcc_inst, &s);
+			pipe_ctx->stream_res.tg = pool->timing_generators[tg_inst];
+			pipe_ctx->plane_res.mi = pool->mis[id_src[i]];
+			pipe_ctx->plane_res.hubp = pool->hubps[id_src[i]];
+			pipe_ctx->plane_res.ipp = pool->ipps[id_src[i]];
+			pipe_ctx->plane_res.xfm = pool->transforms[id_src[i]];
+			pipe_ctx->plane_res.dpp = pool->dpps[id_src[i]];
+			pipe_ctx->stream_res.opp = pool->opps[id_src[i]];
 
-				if (s.dpp_id < MAX_MPCC)
-					pool->mpc->mpcc_array[pipe_ctx->plane_res.mpcc_inst].dpp_id = s.dpp_id;
+			if (pool->dpps[id_src[i]]) {
+				pipe_ctx->plane_res.mpcc_inst = pool->dpps[id_src[i]]->inst;
 
-				if (s.bot_mpcc_id < MAX_MPCC)
-					pool->mpc->mpcc_array[pipe_ctx->plane_res.mpcc_inst].mpcc_bot =
-							&pool->mpc->mpcc_array[s.bot_mpcc_id];
+				if (pool->mpc->funcs->read_mpcc_state) {
+					struct mpcc_state s = {0};
 
-				if (s.opp_id < MAX_OPP)
-					pipe_ctx->stream_res.opp->mpc_tree_params.opp_id = s.opp_id;
+					pool->mpc->funcs->read_mpcc_state(pool->mpc, pipe_ctx->plane_res.mpcc_inst, &s);
+
+					if (s.dpp_id < MAX_MPCC)
+						pool->mpc->mpcc_array[pipe_ctx->plane_res.mpcc_inst].dpp_id =
+								s.dpp_id;
+
+					if (s.bot_mpcc_id < MAX_MPCC)
+						pool->mpc->mpcc_array[pipe_ctx->plane_res.mpcc_inst].mpcc_bot =
+								&pool->mpc->mpcc_array[s.bot_mpcc_id];
+
+					if (s.opp_id < MAX_OPP)
+						pipe_ctx->stream_res.opp->mpc_tree_params.opp_id = s.opp_id;
+				}
 			}
+			pipe_ctx->pipe_idx = id_src[i];
+
+			if (id_src[i] >= pool->timing_generator_count) {
+				id_src[i] = pool->timing_generator_count - 1;
+
+				pipe_ctx->stream_res.tg = pool->timing_generators[id_src[i]];
+				pipe_ctx->stream_res.opp = pool->opps[id_src[i]];
+			}
+
+			pipe_ctx->stream = stream;
 		}
-		pipe_ctx->pipe_idx = tg_inst;
 
-		pipe_ctx->stream = stream;
-		return tg_inst;
+		if (numPipes == 2) {
+			stream->apply_boot_odm_mode = dm_odm_combine_policy_2to1;
+			res_ctx->pipe_ctx[id_src[0]].next_odm_pipe = &res_ctx->pipe_ctx[id_src[1]];
+			res_ctx->pipe_ctx[id_src[0]].prev_odm_pipe = NULL;
+			res_ctx->pipe_ctx[id_src[1]].next_odm_pipe = NULL;
+			res_ctx->pipe_ctx[id_src[1]].prev_odm_pipe = &res_ctx->pipe_ctx[id_src[0]];
+		} else
+			stream->apply_boot_odm_mode = dm_odm_combine_mode_disabled;
+
+		return id_src[0];
 	}
 
 	return -1;
 }
 
-static void mark_seamless_boot_stream(
-		const struct dc  *dc,
-		struct dc_stream_state *stream)
+static void mark_seamless_boot_stream(const struct dc  *dc,
+				      struct dc_stream_state *stream)
 {
 	struct dc_bios *dcb = dc->ctx->dc_bios;
 
-	/* TODO: Check Linux */
-	if (dc->config.allow_seamless_boot_optimization &&
-			!dcb->funcs->is_accelerated_mode(dcb)) {
-		if (dc_validate_seamless_boot_timing(dc, stream->sink, &stream->timing))
-			stream->apply_seamless_boot_optimization = true;
+	DC_LOGGER_INIT(dc->ctx->logger);
+
+	if (stream->apply_seamless_boot_optimization)
+		return;
+	if (!dc->config.allow_seamless_boot_optimization)
+		return;
+	if (dcb->funcs->is_accelerated_mode(dcb))
+		return;
+	if (dc_validate_boot_timing(dc, stream->sink, &stream->timing)) {
+		stream->apply_seamless_boot_optimization = true;
+		DC_LOG_DC("Marked stream for seamless boot optimization\n");
 	}
 }
 
+/*
+ * Acquire a pipe as OTG master and assign to the stream in new dc context.
+ * return - true if OTG master pipe is acquired and new dc context is updated.
+ * false if it fails to acquire an OTG master pipe for this stream.
+ *
+ * In the example below, we acquired pipe 0 as OTG master pipe for the stream.
+ * After the function its Inter-pipe Relation is represented by the diagram
+ * below.
+ *
+ *       Inter-pipe Relation
+ *        __________________________________________________
+ *       |PIPE IDX|   DPP PIPES   | OPP HEADS | OTG MASTER  |
+ *       |        |               |           |             |
+ *       |   0    |               |blank ------------------ |
+ *       |________|_______________|___________|_____________|
+ */
+static bool acquire_otg_master_pipe_for_stream(
+		const struct dc_state *cur_ctx,
+		struct dc_state *new_ctx,
+		const struct resource_pool *pool,
+		struct dc_stream_state *stream)
+{
+	/* TODO: Move this function to DCN specific resource file and acquire
+	 * DSC resource here. The reason is that the function should have the
+	 * same level of responsibility as when we acquire secondary OPP head.
+	 * We acquire DSC when we acquire secondary OPP head, so we should
+	 * acquire DSC when we acquire OTG master.
+	 */
+	int pipe_idx;
+	struct pipe_ctx *pipe_ctx = NULL;
+
+	/*
+	 * Upper level code is responsible to optimize unnecessary addition and
+	 * removal for unchanged streams. So unchanged stream will keep the same
+	 * OTG master instance allocated. When current stream is removed and a
+	 * new stream is added, we want to reuse the OTG instance made available
+	 * by the removed stream first. If not found, we try to avoid of using
+	 * any free pipes already used in current context as this could tear
+	 * down exiting ODM/MPC/MPO configuration unnecessarily.
+	 */
+
+	/*
+	 * Try to acquire the same OTG master already in use. This is not
+	 * optimal because resetting an enabled OTG master pipe for a new stream
+	 * requires an extra frame of wait. However there are test automation
+	 * and eDP assumptions that rely on reusing the same OTG master pipe
+	 * during mode change. We have to keep this logic as is for now.
+	 */
+	pipe_idx = recource_find_free_pipe_used_as_otg_master_in_cur_res_ctx(
+			&cur_ctx->res_ctx, &new_ctx->res_ctx, pool);
+	/*
+	 * Try to acquire a pipe not used in current resource context to avoid
+	 * pipe swapping.
+	 */
+	if (pipe_idx == FREE_PIPE_INDEX_NOT_FOUND)
+		pipe_idx = recource_find_free_pipe_not_used_in_cur_res_ctx(
+				&cur_ctx->res_ctx, &new_ctx->res_ctx, pool);
+	/*
+	 * If pipe swapping is unavoidable, try to acquire pipe used as
+	 * secondary DPP pipe in current state as we prioritize to support more
+	 * streams over supporting MPO planes.
+	 */
+	if (pipe_idx == FREE_PIPE_INDEX_NOT_FOUND)
+		pipe_idx = resource_find_free_pipe_used_as_cur_sec_dpp(
+				&cur_ctx->res_ctx, &new_ctx->res_ctx, pool);
+	if (pipe_idx == FREE_PIPE_INDEX_NOT_FOUND)
+		pipe_idx = resource_find_any_free_pipe(&new_ctx->res_ctx, pool);
+	if (pipe_idx != FREE_PIPE_INDEX_NOT_FOUND) {
+		pipe_ctx = &new_ctx->res_ctx.pipe_ctx[pipe_idx];
+		memset(pipe_ctx, 0, sizeof(*pipe_ctx));
+		pipe_ctx->pipe_idx = pipe_idx;
+		pipe_ctx->stream_res.tg = pool->timing_generators[pipe_idx];
+		pipe_ctx->plane_res.mi = pool->mis[pipe_idx];
+		pipe_ctx->plane_res.hubp = pool->hubps[pipe_idx];
+		pipe_ctx->plane_res.ipp = pool->ipps[pipe_idx];
+		pipe_ctx->plane_res.xfm = pool->transforms[pipe_idx];
+		pipe_ctx->plane_res.dpp = pool->dpps[pipe_idx];
+		pipe_ctx->stream_res.opp = pool->opps[pipe_idx];
+		if (pool->dpps[pipe_idx])
+			pipe_ctx->plane_res.mpcc_inst = pool->dpps[pipe_idx]->inst;
+
+		if (pipe_idx >= pool->timing_generator_count && pool->timing_generator_count != 0) {
+			int tg_inst = pool->timing_generator_count - 1;
+
+			pipe_ctx->stream_res.tg = pool->timing_generators[tg_inst];
+			pipe_ctx->stream_res.opp = pool->opps[tg_inst];
+		}
+
+		pipe_ctx->stream = stream;
+	} else {
+		pipe_idx = acquire_first_split_pipe(&new_ctx->res_ctx, pool, stream);
+	}
+
+	return pipe_idx != FREE_PIPE_INDEX_NOT_FOUND;
+}
+
 enum dc_status resource_map_pool_resources(
 		const struct dc  *dc,
 		struct dc_state *context,
@@ -2100,6 +3962,8 @@ enum dc_status resource_map_pool_resources(
 	struct dc_context *dc_ctx = dc->ctx;
 	struct pipe_ctx *pipe_ctx = NULL;
 	int pipe_idx = -1;
+	bool acquired = false;
+	bool is_dio_encoder = true;
 
 	calculate_phy_pix_clks(stream);
 
@@ -2113,22 +3977,20 @@ enum dc_status resource_map_pool_resources(
 		if (pipe_idx < 0)
 			/* hw resource was assigned to other stream */
 			stream->apply_seamless_boot_optimization = false;
+		else
+			acquired = true;
 	}
 
-	if (pipe_idx < 0)
+	if (!acquired)
 		/* acquire new resources */
-		pipe_idx = acquire_first_free_pipe(&context->res_ctx, pool, stream);
+		acquired = acquire_otg_master_pipe_for_stream(dc->current_state,
+				context, pool, stream);
 
-#ifdef CONFIG_DRM_AMD_DC_DCN
-	if (pipe_idx < 0)
-		pipe_idx = acquire_first_split_pipe(&context->res_ctx, pool, stream);
-#endif
+	pipe_ctx = resource_get_otg_master_for_stream(&context->res_ctx, stream);
 
-	if (pipe_idx < 0 || context->res_ctx.pipe_ctx[pipe_idx].stream_res.tg == NULL)
+	if (!pipe_ctx || pipe_ctx->stream_res.tg == NULL)
 		return DC_NO_CONTROLLER_RESOURCE;
 
-	pipe_ctx = &context->res_ctx.pipe_ctx[pipe_idx];
-
 	pipe_ctx->stream_res.stream_enc =
 		dc->res_pool->funcs->find_first_free_match_stream_enc_for_link(
 			&context->res_ctx, pool, stream);
@@ -2141,16 +4003,20 @@ enum dc_status resource_map_pool_resources(
 		pipe_ctx->stream_res.stream_enc,
 		true);
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	/* Allocate DP HPO Stream Encoder based on signal, hw capabilities
 	 * and link settings
 	 */
-	if (dc_is_dp_signal(stream->signal) &&
-			dc->caps.dp_hpo) {
-		struct dc_link_settings link_settings = {0};
+	if (dc_is_dp_signal(stream->signal) ||
+			dc_is_virtual_signal(stream->signal)) {
+		if (!dc->link_srv->dp_decide_link_settings(stream,
+				&pipe_ctx->link_config.dp_link_settings))
+			return DC_FAIL_DP_LINK_BANDWIDTH;
+
+		dc->link_srv->dp_decide_tunnel_settings(stream,
+				&pipe_ctx->link_config.dp_tunnel_settings);
 
-		decide_link_settings(stream, &link_settings);
-		if (dp_get_link_encoding_format(&link_settings) == DP_128b_132b_ENCODING) {
+		if (dc->link_srv->dp_get_encoding_format(
+				&pipe_ctx->link_config.dp_link_settings) == DP_128b_132b_ENCODING) {
 			pipe_ctx->stream_res.hpo_dp_stream_enc =
 					find_first_free_match_hpo_dp_stream_enc_for_link(
 							&context->res_ctx, pool, stream);
@@ -2162,14 +4028,21 @@ enum dc_status resource_map_pool_resources(
 					&context->res_ctx, pool,
 					pipe_ctx->stream_res.hpo_dp_stream_enc,
 					true);
+			if (!add_hpo_dp_link_enc_to_ctx(&context->res_ctx, pool, pipe_ctx, stream))
+				return DC_NO_LINK_ENC_RESOURCE;
 		}
 	}
-#endif
+
+	if (dc->config.unify_link_enc_assignment && is_dio_encoder)
+		if (!add_dio_link_enc_to_ctx(dc, context, pool, pipe_ctx, stream))
+			return DC_NO_LINK_ENC_RESOURCE;
 
 	/* TODO: Add check if ASIC support and EDID audio */
 	if (!stream->converter_disable_audio &&
 	    dc_is_audio_capable_signal(pipe_ctx->stream->signal) &&
-	    stream->audio_info.mode_count && stream->audio_info.flags.all) {
+	    stream->audio_info.mode_count &&
+		(stream->audio_info.flags.all ||
+		(stream->sink && stream->sink->edid_caps.panel_patch.skip_audio_sab_check))) {
 		pipe_ctx->stream_res.audio = find_first_free_audio(
 		&context->res_ctx, pool, pipe_ctx->stream_res.stream_enc->id, dc_ctx->dce_version);
 
@@ -2185,14 +4058,10 @@ enum dc_status resource_map_pool_resources(
 
 	/* Add ABM to the resource if on EDP */
 	if (pipe_ctx->stream && dc_is_embedded_signal(pipe_ctx->stream->signal)) {
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 		if (pool->abm)
 			pipe_ctx->stream_res.abm = pool->abm;
 		else
 			pipe_ctx->stream_res.abm = pool->multiple_abms[pipe_ctx->stream_res.tg->inst];
-#else
-		pipe_ctx->stream_res.abm = pool->abm;
-#endif
 	}
 
 	for (i = 0; i < context->stream_count; i++)
@@ -2209,47 +4078,334 @@ enum dc_status resource_map_pool_resources(
 	return DC_ERROR_UNEXPECTED;
 }
 
-/**
- * dc_resource_state_copy_construct_current() - Creates a new dc_state from existing state
- * Is a shallow copy.  Increments refcounts on existing streams and planes.
- * @dc: copy out of dc->current_state
- * @dst_ctx: copy into this
- */
-void dc_resource_state_copy_construct_current(
-		const struct dc *dc,
-		struct dc_state *dst_ctx)
+bool dc_resource_is_dsc_encoding_supported(const struct dc *dc)
 {
-	dc_resource_state_copy_construct(dc->current_state, dst_ctx);
+	if (dc->res_pool == NULL)
+		return false;
+
+	return dc->res_pool->res_cap->num_dsc > 0;
 }
 
+static bool planes_changed_for_existing_stream(struct dc_state *context,
+					       struct dc_stream_state *stream,
+					       const struct dc_validation_set set[],
+					       int set_count)
+{
+	int i, j;
+	struct dc_stream_status *stream_status = NULL;
+
+	for (i = 0; i < context->stream_count; i++) {
+		if (context->streams[i] == stream) {
+			stream_status = &context->stream_status[i];
+			break;
+		}
+	}
+
+	if (!stream_status) {
+		ASSERT(0);
+		return false;
+	}
+
+	for (i = 0; i < set_count; i++)
+		if (set[i].stream == stream)
+			break;
+
+	if (i == set_count)
+		ASSERT(0);
+
+	if (set[i].plane_count != stream_status->plane_count)
+		return true;
+
+	for (j = 0; j < set[i].plane_count; j++)
+		if (set[i].plane_states[j] != stream_status->plane_states[j])
+			return true;
 
-void dc_resource_state_construct(
+	return false;
+}
+
+static bool add_all_planes_for_stream(
 		const struct dc *dc,
-		struct dc_state *dst_ctx)
+		struct dc_stream_state *stream,
+		const struct dc_validation_set set[],
+		int set_count,
+		struct dc_state *state)
 {
-	dst_ctx->clk_mgr = dc->clk_mgr;
-}
+	int i, j;
 
+	for (i = 0; i < set_count; i++)
+		if (set[i].stream == stream)
+			break;
 
-bool dc_resource_is_dsc_encoding_supported(const struct dc *dc)
+	if (i == set_count) {
+		dm_error("Stream %p not found in set!\n", stream);
+		return false;
+	}
+
+	for (j = 0; j < set[i].plane_count; j++)
+		if (!dc_state_add_plane(dc, stream, set[i].plane_states[j], state))
+			return false;
+
+	return true;
+}
+
+/**
+ * dc_validate_with_context - Validate and update the potential new stream in the context object
+ *
+ * @dc: Used to get the current state status
+ * @set: An array of dc_validation_set with all the current streams reference
+ * @set_count: Total of streams
+ * @context: New context
+ * @validate_mode: identify the validation mode
+ *
+ * This function updates the potential new stream in the context object. It
+ * creates multiple lists for the add, remove, and unchanged streams. In
+ * particular, if the unchanged streams have a plane that changed, it is
+ * necessary to remove all planes from the unchanged streams. In summary, this
+ * function is responsible for validating the new context.
+ *
+ * Return:
+ * In case of success, return DC_OK (1), otherwise, return a DC error.
+ */
+enum dc_status dc_validate_with_context(struct dc *dc,
+					const struct dc_validation_set set[],
+					int set_count,
+					struct dc_state *context,
+					enum dc_validate_mode validate_mode)
 {
-	return dc->res_pool->res_cap->num_dsc > 0;
+	struct dc_stream_state *unchanged_streams[MAX_PIPES] = { 0 };
+	struct dc_stream_state *del_streams[MAX_PIPES] = { 0 };
+	struct dc_stream_state *add_streams[MAX_PIPES] = { 0 };
+	int old_stream_count = context->stream_count;
+	enum dc_status res = DC_ERROR_UNEXPECTED;
+	int unchanged_streams_count = 0;
+	int del_streams_count = 0;
+	int add_streams_count = 0;
+	bool found = false;
+	int i, j, k;
+
+	DC_LOGGER_INIT(dc->ctx->logger);
+
+	/* First build a list of streams to be remove from current context */
+	for (i = 0; i < old_stream_count; i++) {
+		struct dc_stream_state *stream = context->streams[i];
+
+		for (j = 0; j < set_count; j++) {
+			if (stream == set[j].stream) {
+				found = true;
+				break;
+			}
+		}
+
+		if (!found)
+			del_streams[del_streams_count++] = stream;
+
+		found = false;
+	}
+
+	/* Second, build a list of new streams */
+	for (i = 0; i < set_count; i++) {
+		struct dc_stream_state *stream = set[i].stream;
+
+		for (j = 0; j < old_stream_count; j++) {
+			if (stream == context->streams[j]) {
+				found = true;
+				break;
+			}
+		}
+
+		if (!found)
+			add_streams[add_streams_count++] = stream;
+
+		found = false;
+	}
+
+	/* Build a list of unchanged streams which is necessary for handling
+	 * planes change such as added, removed, and updated.
+	 */
+	for (i = 0; i < set_count; i++) {
+		/* Check if stream is part of the delete list */
+		for (j = 0; j < del_streams_count; j++) {
+			if (set[i].stream == del_streams[j]) {
+				found = true;
+				break;
+			}
+		}
+
+		if (!found) {
+			/* Check if stream is part of the add list */
+			for (j = 0; j < add_streams_count; j++) {
+				if (set[i].stream == add_streams[j]) {
+					found = true;
+					break;
+				}
+			}
+		}
+
+		if (!found)
+			unchanged_streams[unchanged_streams_count++] = set[i].stream;
+
+		found = false;
+	}
+
+	/* Remove all planes for unchanged streams if planes changed */
+	for (i = 0; i < unchanged_streams_count; i++) {
+		if (planes_changed_for_existing_stream(context,
+						       unchanged_streams[i],
+						       set,
+						       set_count)) {
+
+			if (!dc_state_rem_all_planes_for_stream(dc,
+							  unchanged_streams[i],
+							  context)) {
+				res = DC_FAIL_DETACH_SURFACES;
+				goto fail;
+			}
+		}
+	}
+
+	/* Remove all planes for removed streams and then remove the streams */
+	for (i = 0; i < del_streams_count; i++) {
+		/* Need to cpy the dwb data from the old stream in order to efc to work */
+		if (del_streams[i]->num_wb_info > 0) {
+			for (j = 0; j < add_streams_count; j++) {
+				if (del_streams[i]->sink == add_streams[j]->sink) {
+					add_streams[j]->num_wb_info = del_streams[i]->num_wb_info;
+					for (k = 0; k < del_streams[i]->num_wb_info; k++)
+						add_streams[j]->writeback_info[k] = del_streams[i]->writeback_info[k];
+				}
+			}
+		}
+
+		if (dc_state_get_stream_subvp_type(context, del_streams[i]) == SUBVP_PHANTOM) {
+			/* remove phantoms specifically */
+			if (!dc_state_rem_all_phantom_planes_for_stream(dc, del_streams[i], context, true)) {
+				res = DC_FAIL_DETACH_SURFACES;
+				goto fail;
+			}
+
+			res = dc_state_remove_phantom_stream(dc, context, del_streams[i]);
+			dc_state_release_phantom_stream(dc, context, del_streams[i]);
+		} else {
+			if (!dc_state_rem_all_planes_for_stream(dc, del_streams[i], context)) {
+				res = DC_FAIL_DETACH_SURFACES;
+				goto fail;
+			}
+
+			res = dc_state_remove_stream(dc, context, del_streams[i]);
+		}
+
+		if (res != DC_OK)
+			goto fail;
+	}
+
+	/* Swap seamless boot stream to pipe 0 (if needed) to ensure pipe_ctx
+	 * matches. This may change in the future if seamless_boot_stream can be
+	 * multiple.
+	 */
+	for (i = 0; i < add_streams_count; i++) {
+		mark_seamless_boot_stream(dc, add_streams[i]);
+		if (add_streams[i]->apply_seamless_boot_optimization && i != 0) {
+			struct dc_stream_state *temp = add_streams[0];
+
+			add_streams[0] = add_streams[i];
+			add_streams[i] = temp;
+			break;
+		}
+	}
+
+	/* Add new streams and then add all planes for the new stream */
+	for (i = 0; i < add_streams_count; i++) {
+		calculate_phy_pix_clks(add_streams[i]);
+		res = dc_state_add_stream(dc, context, add_streams[i]);
+		if (res != DC_OK)
+			goto fail;
+
+		if (!add_all_planes_for_stream(dc, add_streams[i], set, set_count, context)) {
+			res = DC_FAIL_ATTACH_SURFACES;
+			goto fail;
+		}
+	}
+
+	/* Add all planes for unchanged streams if planes changed */
+	for (i = 0; i < unchanged_streams_count; i++) {
+		if (planes_changed_for_existing_stream(context,
+						       unchanged_streams[i],
+						       set,
+						       set_count)) {
+			if (!add_all_planes_for_stream(dc, unchanged_streams[i], set, set_count, context)) {
+				res = DC_FAIL_ATTACH_SURFACES;
+				goto fail;
+			}
+		}
+	}
+
+	/* clear subvp cursor limitations */
+	for (i = 0; i < context->stream_count; i++) {
+		dc_state_set_stream_subvp_cursor_limit(context->streams[i], context, false);
+	}
+
+	res = dc_validate_global_state(dc, context, validate_mode);
+
+	/* calculate pixel rate divider after deciding pxiel clock & odm combine  */
+	if ((dc->hwss.calculate_pix_rate_divider) && (res == DC_OK)) {
+		for (i = 0; i < add_streams_count; i++)
+			dc->hwss.calculate_pix_rate_divider(dc, context, add_streams[i]);
+	}
+
+fail:
+	if (res != DC_OK)
+		DC_LOG_WARNING("%s:resource validation failed, dc_status:%d\n",
+			       __func__,
+			       res);
+
+	return res;
 }
 
+#if defined(CONFIG_DRM_AMD_DC_FP)
+#endif /* CONFIG_DRM_AMD_DC_FP */
+
+/**
+ * calculate_timing_params_for_dsc_with_padding - Calculates timing parameters for DSC with padding.
+ * @pipe_ctx: Pointer to the pipe context structure.
+ *
+ * This function calculates the timing parameters for a given pipe context based on the
+ * display stream compression (DSC) configuration. If the horizontal active pixels (hactive) are less
+ * than the total width of the DSC slices, it sets the dsc_hactive_padding value to the difference. If the
+ * total horizontal timing minus the dsc_hactive_padding value is less than 32, it resets the dsc_hactive_padding
+ * value to 0.
+ */
+static void calculate_timing_params_for_dsc_with_padding(struct pipe_ctx *pipe_ctx)
+{
+	struct dc_stream_state *stream = NULL;
+
+	if (!pipe_ctx)
+		return;
+
+	stream = pipe_ctx->stream;
+	pipe_ctx->dsc_padding_params.dsc_hactive_padding = 0;
+	pipe_ctx->dsc_padding_params.dsc_htotal_padding = 0;
+
+	if (stream)
+		pipe_ctx->dsc_padding_params.dsc_pix_clk_100hz = stream->timing.pix_clk_100hz;
+
+}
 
 /**
- * dc_validate_global_state() - Determine if HW can support a given state
- * Checks HW resource availability and bandwidth requirement.
+ * dc_validate_global_state() - Determine if hardware can support a given state
+ *
  * @dc: dc struct for this driver
  * @new_ctx: state to be validated
- * @fast_validate: set to true if only yes/no to support matters
+ * @validate_mode: identify the validation mode
  *
- * Return: DC_OK if the result can be programmed.  Otherwise, an error code.
+ * Checks hardware resource availability and bandwidth requirement.
+ *
+ * Return:
+ * DC_OK if the result can be programmed. Otherwise, an error code.
  */
 enum dc_status dc_validate_global_state(
 		struct dc *dc,
 		struct dc_state *new_ctx,
-		bool fast_validate)
+		enum dc_validate_mode validate_mode)
 {
 	enum dc_status result = DC_ERROR_UNEXPECTED;
 	int i, j;
@@ -2272,6 +4428,10 @@ enum dc_status dc_validate_global_state(
 			if (pipe_ctx->stream != stream)
 				continue;
 
+			/* Decide whether hblank borrow is needed and save it in pipe_ctx */
+			if (dc->debug.enable_hblank_borrow)
+				calculate_timing_params_for_dsc_with_padding(pipe_ctx);
+
 			if (dc->res_pool->funcs->patch_unknown_plane_state &&
 					pipe_ctx->plane_state &&
 					pipe_ctx->plane_state->tiling_info.gfx9.swizzle == DC_SW_UNKNOWN) {
@@ -2304,18 +4464,7 @@ enum dc_status dc_validate_global_state(
 	result = resource_build_scaling_params_for_context(dc, new_ctx);
 
 	if (result == DC_OK)
-		if (!dc->res_pool->funcs->validate_bandwidth(dc, new_ctx, fast_validate))
-			result = DC_FAIL_BANDWIDTH_VALIDATE;
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	/*
-	 * Only update link encoder to stream assignment after bandwidth validation passed.
-	 * TODO: Split out assignment and validation.
-	 */
-	if (result == DC_OK && dc->res_pool->funcs->link_encs_assign && fast_validate == false)
-		dc->res_pool->funcs->link_encs_assign(
-			dc, new_ctx, new_ctx->streams, new_ctx->stream_count);
-#endif
+		result = dc->res_pool->funcs->validate_bandwidth(dc, new_ctx, validate_mode);
 
 	return result;
 }
@@ -2348,19 +4497,22 @@ static void set_avi_info_frame(
 	uint32_t pixel_encoding = 0;
 	enum scanning_type scan_type = SCANNING_TYPE_NODATA;
 	enum dc_aspect_ratio aspect = ASPECT_RATIO_NO_DATA;
-	bool itc = false;
-	uint8_t itc_value = 0;
-	uint8_t cn0_cn1 = 0;
-	unsigned int cn0_cn1_value = 0;
 	uint8_t *check_sum = NULL;
 	uint8_t byte_index = 0;
 	union hdmi_info_packet hdmi_info;
-	union display_content_support support = {0};
 	unsigned int vic = pipe_ctx->stream->timing.vic;
+	unsigned int rid = pipe_ctx->stream->timing.rid;
+	unsigned int fr_ind = pipe_ctx->stream->timing.fr_index;
 	enum dc_timing_3d_format format;
 
+	if (stream->avi_infopacket.valid) {
+		*info_packet = stream->avi_infopacket;
+		return;
+	}
+
 	memset(&hdmi_info, 0, sizeof(union hdmi_info_packet));
 
+
 	color_space = pipe_ctx->stream->output_color_space;
 	if (color_space == COLOR_SPACE_UNKNOWN)
 		color_space = (stream->timing.pixel_encoding == PIXEL_ENCODING_RGB) ?
@@ -2413,23 +4565,35 @@ static void set_avi_info_frame(
 	hdmi_info.bits.S0_S1 = scan_type;
 
 	/* C0, C1 : Colorimetry */
-	if (color_space == COLOR_SPACE_YCBCR709 ||
-			color_space == COLOR_SPACE_YCBCR709_LIMITED)
+	switch (color_space) {
+	case COLOR_SPACE_YCBCR709:
+	case COLOR_SPACE_YCBCR709_LIMITED:
 		hdmi_info.bits.C0_C1 = COLORIMETRY_ITU709;
-	else if (color_space == COLOR_SPACE_YCBCR601 ||
-			color_space == COLOR_SPACE_YCBCR601_LIMITED)
+		break;
+	case COLOR_SPACE_YCBCR601:
+	case COLOR_SPACE_YCBCR601_LIMITED:
 		hdmi_info.bits.C0_C1 = COLORIMETRY_ITU601;
-	else {
-		hdmi_info.bits.C0_C1 = COLORIMETRY_NO_DATA;
-	}
-	if (color_space == COLOR_SPACE_2020_RGB_FULLRANGE ||
-			color_space == COLOR_SPACE_2020_RGB_LIMITEDRANGE ||
-			color_space == COLOR_SPACE_2020_YCBCR) {
+		break;
+	case COLOR_SPACE_2020_RGB_FULLRANGE:
+	case COLOR_SPACE_2020_RGB_LIMITEDRANGE:
+	case COLOR_SPACE_2020_YCBCR_LIMITED:
 		hdmi_info.bits.EC0_EC2 = COLORIMETRYEX_BT2020RGBYCBCR;
 		hdmi_info.bits.C0_C1   = COLORIMETRY_EXTENDED;
-	} else if (color_space == COLOR_SPACE_ADOBERGB) {
+		break;
+	case COLOR_SPACE_ADOBERGB:
 		hdmi_info.bits.EC0_EC2 = COLORIMETRYEX_ADOBERGB;
 		hdmi_info.bits.C0_C1   = COLORIMETRY_EXTENDED;
+		break;
+	case COLOR_SPACE_SRGB:
+	default:
+		hdmi_info.bits.C0_C1 = COLORIMETRY_NO_DATA;
+		break;
+	}
+
+	if (pixel_encoding && color_space == COLOR_SPACE_2020_YCBCR_LIMITED &&
+			stream->out_transfer_func.tf == TRANSFER_FUNCTION_GAMMA22) {
+		hdmi_info.bits.EC0_EC2 = 0;
+		hdmi_info.bits.C0_C1 = COLORIMETRY_ITU709;
 	}
 
 	/* TODO: un-hardcode aspect ratio */
@@ -2451,49 +4615,27 @@ static void set_avi_info_frame(
 	/* Active Format Aspect ratio - same as Picture Aspect Ratio. */
 	hdmi_info.bits.R0_R3 = ACTIVE_FORMAT_ASPECT_RATIO_SAME_AS_PICTURE;
 
-	/* TODO: un-hardcode cn0_cn1 and itc */
-
-	cn0_cn1 = 0;
-	cn0_cn1_value = 0;
-
-	itc = true;
-	itc_value = 1;
-
-	support = stream->content_support;
-
-	if (itc) {
-		if (!support.bits.valid_content_type) {
-			cn0_cn1_value = 0;
-		} else {
-			if (cn0_cn1 == DISPLAY_CONTENT_TYPE_GRAPHICS) {
-				if (support.bits.graphics_content == 1) {
-					cn0_cn1_value = 0;
-				}
-			} else if (cn0_cn1 == DISPLAY_CONTENT_TYPE_PHOTO) {
-				if (support.bits.photo_content == 1) {
-					cn0_cn1_value = 1;
-				} else {
-					cn0_cn1_value = 0;
-					itc_value = 0;
-				}
-			} else if (cn0_cn1 == DISPLAY_CONTENT_TYPE_CINEMA) {
-				if (support.bits.cinema_content == 1) {
-					cn0_cn1_value = 2;
-				} else {
-					cn0_cn1_value = 0;
-					itc_value = 0;
-				}
-			} else if (cn0_cn1 == DISPLAY_CONTENT_TYPE_GAME) {
-				if (support.bits.game_content == 1) {
-					cn0_cn1_value = 3;
-				} else {
-					cn0_cn1_value = 0;
-					itc_value = 0;
-				}
-			}
-		}
-		hdmi_info.bits.CN0_CN1 = cn0_cn1_value;
-		hdmi_info.bits.ITC = itc_value;
+	switch (stream->content_type) {
+	case DISPLAY_CONTENT_TYPE_NO_DATA:
+		hdmi_info.bits.CN0_CN1 = 0;
+		hdmi_info.bits.ITC = 1;
+		break;
+	case DISPLAY_CONTENT_TYPE_GRAPHICS:
+		hdmi_info.bits.CN0_CN1 = 0;
+		hdmi_info.bits.ITC = 1;
+		break;
+	case DISPLAY_CONTENT_TYPE_PHOTO:
+		hdmi_info.bits.CN0_CN1 = 1;
+		hdmi_info.bits.ITC = 1;
+		break;
+	case DISPLAY_CONTENT_TYPE_CINEMA:
+		hdmi_info.bits.CN0_CN1 = 2;
+		hdmi_info.bits.ITC = 1;
+		break;
+	case DISPLAY_CONTENT_TYPE_GAME:
+		hdmi_info.bits.CN0_CN1 = 3;
+		hdmi_info.bits.ITC = 1;
+		break;
 	}
 
 	if (stream->qs_bit == 1) {
@@ -2510,19 +4652,11 @@ static void set_avi_info_frame(
 
 	/* TODO : We should handle YCC quantization */
 	/* but we do not have matrix calculation */
-	if (stream->qy_bit == 1) {
-		if (color_space == COLOR_SPACE_SRGB ||
-			color_space == COLOR_SPACE_2020_RGB_FULLRANGE)
-			hdmi_info.bits.YQ0_YQ1 = YYC_QUANTIZATION_LIMITED_RANGE;
-		else if (color_space == COLOR_SPACE_SRGB_LIMITED ||
-					color_space == COLOR_SPACE_2020_RGB_LIMITEDRANGE)
-			hdmi_info.bits.YQ0_YQ1 = YYC_QUANTIZATION_LIMITED_RANGE;
-		else
-			hdmi_info.bits.YQ0_YQ1 = YYC_QUANTIZATION_LIMITED_RANGE;
-	} else
-		hdmi_info.bits.YQ0_YQ1 = YYC_QUANTIZATION_LIMITED_RANGE;
+	hdmi_info.bits.YQ0_YQ1 = YYC_QUANTIZATION_LIMITED_RANGE;
 
 	///VIC
+	if (pipe_ctx->stream->timing.hdmi_vic != 0)
+		vic = 0;
 	format = stream->timing.timing_3d_format;
 	/*todo, add 3DStereo support*/
 	if (format != TIMING_3D_FORMAT_NONE) {
@@ -2557,6 +4691,15 @@ static void set_avi_info_frame(
 		hdmi_info.bits.header.length = 14;
 	}
 
+	if (rid != 0 && fr_ind != 0) {
+		hdmi_info.bits.header.version = 4;
+		hdmi_info.bits.header.length = 15;
+
+		hdmi_info.bits.FR0_FR3 = fr_ind & 0xF;
+		hdmi_info.bits.FR4 = (fr_ind >> 4) & 0x1;
+		hdmi_info.bits.RID0_RID5 = rid;
+	}
+
 	/* pixel repetition
 	 * PR0 - PR3 start from 0 whereas pHwPathMode->mode.timing.flags.pixel
 	 * repetition start from 1 */
@@ -2655,58 +4798,69 @@ static void set_vsc_info_packet(
 
 	*info_packet = stream->vsc_infopacket;
 }
-
-void dc_resource_state_destruct(struct dc_state *context)
+static void set_hfvs_info_packet(
+		struct dc_info_packet *info_packet,
+		struct dc_stream_state *stream)
 {
-	int i, j;
-
-	for (i = 0; i < context->stream_count; i++) {
-		for (j = 0; j < context->stream_status[i].plane_count; j++)
-			dc_plane_state_release(
-				context->stream_status[i].plane_states[j]);
+	if (!stream->hfvsif_infopacket.valid)
+		return;
 
-		context->stream_status[i].plane_count = 0;
-		dc_stream_release(context->streams[i]);
-		context->streams[i] = NULL;
-	}
-	context->stream_count = 0;
+	*info_packet = stream->hfvsif_infopacket;
 }
 
-void dc_resource_state_copy_construct(
-		const struct dc_state *src_ctx,
-		struct dc_state *dst_ctx)
+static void adaptive_sync_override_dp_info_packets_sdp_line_num(
+		const struct dc_crtc_timing *timing,
+		struct enc_sdp_line_num *sdp_line_num,
+		unsigned int vstartup_start)
 {
-	int i, j;
-	struct kref refcount = dst_ctx->refcount;
+	uint32_t asic_blank_start = 0;
+	uint32_t asic_blank_end   = 0;
+	uint32_t v_update = 0;
 
-	*dst_ctx = *src_ctx;
+	const struct dc_crtc_timing *tg = timing;
 
-	for (i = 0; i < MAX_PIPES; i++) {
-		struct pipe_ctx *cur_pipe = &dst_ctx->res_ctx.pipe_ctx[i];
+	/* blank_start = frame end - front porch */
+	asic_blank_start = tg->v_total - tg->v_front_porch;
 
-		if (cur_pipe->top_pipe)
-			cur_pipe->top_pipe =  &dst_ctx->res_ctx.pipe_ctx[cur_pipe->top_pipe->pipe_idx];
+	/* blank_end = blank_start - active */
+	asic_blank_end = (asic_blank_start - tg->v_border_bottom -
+						tg->v_addressable - tg->v_border_top);
 
-		if (cur_pipe->bottom_pipe)
-			cur_pipe->bottom_pipe = &dst_ctx->res_ctx.pipe_ctx[cur_pipe->bottom_pipe->pipe_idx];
+	if (vstartup_start > asic_blank_end) {
+		v_update = (tg->v_total - (vstartup_start - asic_blank_end));
+		sdp_line_num->adaptive_sync_line_num_valid = true;
+		sdp_line_num->adaptive_sync_line_num = (tg->v_total - v_update - 1);
+	} else {
+		sdp_line_num->adaptive_sync_line_num_valid = false;
+		sdp_line_num->adaptive_sync_line_num = 0;
+	}
+}
 
-		if (cur_pipe->next_odm_pipe)
-			cur_pipe->next_odm_pipe =  &dst_ctx->res_ctx.pipe_ctx[cur_pipe->next_odm_pipe->pipe_idx];
+static void set_adaptive_sync_info_packet(
+		struct dc_info_packet *info_packet,
+		const struct dc_stream_state *stream,
+		struct encoder_info_frame *info_frame,
+		unsigned int vstartup_start)
+{
+	if (!stream->adaptive_sync_infopacket.valid)
+		return;
 
-		if (cur_pipe->prev_odm_pipe)
-			cur_pipe->prev_odm_pipe = &dst_ctx->res_ctx.pipe_ctx[cur_pipe->prev_odm_pipe->pipe_idx];
-	}
+	adaptive_sync_override_dp_info_packets_sdp_line_num(
+			&stream->timing,
+			&info_frame->sdp_line_num,
+			vstartup_start);
 
-	for (i = 0; i < dst_ctx->stream_count; i++) {
-		dc_stream_retain(dst_ctx->streams[i]);
-		for (j = 0; j < dst_ctx->stream_status[i].plane_count; j++)
-			dc_plane_state_retain(
-				dst_ctx->stream_status[i].plane_states[j]);
-	}
+	*info_packet = stream->adaptive_sync_infopacket;
+}
 
-	/* context refcount should not be overridden */
-	dst_ctx->refcount = refcount;
+static void set_vtem_info_packet(
+		struct dc_info_packet *info_packet,
+		struct dc_stream_state *stream)
+{
+	if (!stream->vtem_infopacket.valid)
+		return;
 
+	*info_packet = stream->vtem_infopacket;
 }
 
 struct clock_source *dc_resource_find_first_free_pll(
@@ -2727,6 +4881,7 @@ void resource_build_info_frame(struct pipe_ctx *pipe_ctx)
 {
 	enum signal_type signal = SIGNAL_TYPE_NONE;
 	struct encoder_info_frame *info = &pipe_ctx->stream_res.encoder_info_frame;
+	unsigned int vstartup_start = 0;
 
 	/* default all packets to invalid */
 	info->avi.valid = false;
@@ -2735,14 +4890,21 @@ void resource_build_info_frame(struct pipe_ctx *pipe_ctx)
 	info->spd.valid = false;
 	info->hdrsmd.valid = false;
 	info->vsc.valid = false;
-
+	info->hfvsif.valid = false;
+	info->vtem.valid = false;
+	info->adaptive_sync.valid = false;
 	signal = pipe_ctx->stream->signal;
 
+	if (pipe_ctx->stream->ctx->dc->res_pool->funcs->get_vstartup_for_pipe)
+		vstartup_start = pipe_ctx->stream->ctx->dc->res_pool->funcs->get_vstartup_for_pipe(pipe_ctx);
+
 	/* HDMi and DP have different info packets*/
 	if (dc_is_hdmi_signal(signal)) {
 		set_avi_info_frame(&info->avi, pipe_ctx);
 
 		set_vendor_info_packet(&info->vendor, pipe_ctx->stream);
+		set_hfvs_info_packet(&info->hfvsif, pipe_ctx->stream);
+		set_vtem_info_packet(&info->vtem, pipe_ctx->stream);
 
 		set_spd_info_packet(&info->spd, pipe_ctx->stream);
 
@@ -2754,6 +4916,10 @@ void resource_build_info_frame(struct pipe_ctx *pipe_ctx)
 		set_spd_info_packet(&info->spd, pipe_ctx->stream);
 
 		set_hdr_static_info_packet(&info->hdrsmd, pipe_ctx->stream);
+		set_adaptive_sync_info_packet(&info->adaptive_sync,
+										pipe_ctx->stream,
+										info,
+										vstartup_start);
 	}
 
 	patch_gamut_packet_checksum(&info->gamut);
@@ -2766,7 +4932,7 @@ enum dc_status resource_map_clock_resources(
 {
 	/* acquire new resources */
 	const struct resource_pool *pool = dc->res_pool;
-	struct pipe_ctx *pipe_ctx = resource_get_head_pipe_for_stream(
+	struct pipe_ctx *pipe_ctx = resource_get_otg_master_for_stream(
 				&context->res_ctx, stream);
 
 	if (!pipe_ctx)
@@ -2828,7 +4994,7 @@ bool pipe_need_reprogram(
 	if (pipe_ctx_old->stream_res.stream_enc != pipe_ctx->stream_res.stream_enc)
 		return true;
 
-	if (is_timing_changed(pipe_ctx_old->stream, pipe_ctx->stream))
+	if (dc_is_timing_changed(pipe_ctx_old->stream, pipe_ctx->stream))
 		return true;
 
 	if (pipe_ctx_old->stream->dpms_off != pipe_ctx->stream->dpms_off)
@@ -2841,21 +5007,23 @@ bool pipe_need_reprogram(
 	if (pipe_ctx_old->stream_res.dsc != pipe_ctx->stream_res.dsc)
 		return true;
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	if (pipe_ctx_old->stream_res.hpo_dp_stream_enc != pipe_ctx->stream_res.hpo_dp_stream_enc)
 		return true;
-#endif
+	if (pipe_ctx_old->link_res.hpo_dp_link_enc != pipe_ctx->link_res.hpo_dp_link_enc)
+		return true;
 
 	/* DIG link encoder resource assignment for stream changed. */
-	if (pipe_ctx_old->stream->ctx->dc->res_pool->funcs->link_encs_assign) {
+	if (pipe_ctx_old->stream->ctx->dc->config.unify_link_enc_assignment) {
+		if (pipe_ctx_old->link_res.dio_link_enc != pipe_ctx->link_res.dio_link_enc)
+			return true;
+	} else if (pipe_ctx_old->stream->ctx->dc->res_pool->funcs->link_encs_assign) {
 		bool need_reprogram = false;
 		struct dc *dc = pipe_ctx_old->stream->ctx->dc;
-		enum link_enc_cfg_mode mode = dc->current_state->res_ctx.link_enc_cfg_ctx.mode;
+		struct link_encoder *link_enc_prev =
+			link_enc_cfg_get_link_enc_used_by_stream_current(dc, pipe_ctx_old->stream);
 
-		dc->current_state->res_ctx.link_enc_cfg_ctx.mode = LINK_ENC_CFG_STEADY;
-		if (link_enc_cfg_get_link_enc_used_by_stream(dc, pipe_ctx_old->stream) != pipe_ctx->stream->link_enc)
+		if (link_enc_prev != pipe_ctx->stream->link_enc)
 			need_reprogram = true;
-		dc->current_state->res_ctx.link_enc_cfg_ctx.mode = mode;
 
 		return need_reprogram;
 	}
@@ -2881,7 +5049,7 @@ void resource_build_bit_depth_reduction_params(struct dc_stream_state *stream,
 			option = DITHER_OPTION_SPATIAL8;
 			break;
 		case COLOR_DEPTH_101010:
-			option = DITHER_OPTION_SPATIAL10;
+			option = DITHER_OPTION_TRUN10;
 			break;
 		default:
 			option = DITHER_OPTION_DISABLE;
@@ -2907,6 +5075,8 @@ void resource_build_bit_depth_reduction_params(struct dc_stream_state *stream,
 			option == DITHER_OPTION_TRUN10_SPATIAL8_FM6) {
 		fmt_bit_depth->flags.TRUNCATE_ENABLED = 1;
 		fmt_bit_depth->flags.TRUNCATE_DEPTH = 2;
+		if (option == DITHER_OPTION_TRUN10)
+			fmt_bit_depth->flags.TRUNCATE_MODE = 1;
 	}
 
 	/* special case - Formatter can only reduce by 4 bits at most.
@@ -2991,6 +5161,9 @@ void resource_build_bit_depth_reduction_params(struct dc_stream_state *stream,
 
 enum dc_status dc_validate_stream(struct dc *dc, struct dc_stream_state *stream)
 {
+	if (dc == NULL || stream == NULL)
+		return DC_ERROR_UNEXPECTED;
+
 	struct dc_link *link = stream->link;
 	struct timing_generator *tg = dc->res_pool->timing_generators[0];
 	enum dc_status res = DC_OK;
@@ -3010,7 +5183,7 @@ enum dc_status dc_validate_stream(struct dc *dc, struct dc_stream_state *stream)
 	/* TODO: validate audio ASIC caps, encoder */
 
 	if (res == DC_OK)
-		res = dc_link_validate_mode_timing(stream,
+		res = dc->link_srv->validate_mode_timing(stream,
 		      link,
 		      &stream->timing);
 
@@ -3051,10 +5224,8 @@ unsigned int resource_pixel_format_to_bpp(enum surface_pixel_format format)
 	case SURFACE_PIXEL_FORMAT_GRPH_ARGB2101010:
 	case SURFACE_PIXEL_FORMAT_GRPH_ABGR2101010:
 	case SURFACE_PIXEL_FORMAT_GRPH_ABGR2101010_XR_BIAS:
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	case SURFACE_PIXEL_FORMAT_GRPH_RGBE:
 	case SURFACE_PIXEL_FORMAT_GRPH_RGBE_ALPHA:
-#endif
 		return 32;
 	case SURFACE_PIXEL_FORMAT_GRPH_ARGB16161616:
 	case SURFACE_PIXEL_FORMAT_GRPH_ABGR16161616:
@@ -3111,22 +5282,434 @@ void get_audio_check(struct audio_info *aud_modes,
 	}
 }
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-struct hpo_dp_link_encoder *resource_get_unused_hpo_dp_link_encoder(
-		const struct resource_pool *pool)
+struct link_encoder *get_temp_dio_link_enc(
+		const struct resource_context *res_ctx,
+		const struct resource_pool *const pool,
+		const struct dc_link *link)
 {
-	uint8_t i;
-	struct hpo_dp_link_encoder *enc = NULL;
+	struct link_encoder *link_enc = NULL;
+	int enc_index;
+
+	if (link->is_dig_mapping_flexible)
+		enc_index = find_acquired_dio_link_enc_for_link(res_ctx, link);
+	else
+		enc_index = link->eng_id;
+
+	if (enc_index < 0)
+		enc_index = find_free_dio_link_enc(res_ctx, link, pool, NULL);
+
+	if (enc_index >= 0)
+		link_enc = pool->link_encoders[enc_index];
 
-	ASSERT(pool->hpo_dp_link_enc_count <= MAX_HPO_DP2_LINK_ENCODERS);
+	return link_enc;
+}
+
+static struct hpo_dp_link_encoder *get_temp_hpo_dp_link_enc(
+		const struct resource_context *res_ctx,
+		const struct resource_pool *const pool,
+		const struct dc_link *link)
+{
+	struct hpo_dp_link_encoder *hpo_dp_link_enc = NULL;
+	int enc_index;
+
+	enc_index = find_acquired_hpo_dp_link_enc_for_link(res_ctx, link);
+
+	if (enc_index < 0)
+		enc_index = find_free_hpo_dp_link_enc(res_ctx, pool);
+
+	if (enc_index >= 0)
+		hpo_dp_link_enc = pool->hpo_dp_link_enc[enc_index];
+
+	return hpo_dp_link_enc;
+}
+
+bool get_temp_dp_link_res(struct dc_link *link,
+		struct link_resource *link_res,
+		struct dc_link_settings *link_settings)
+{
+	const struct dc *dc  = link->dc;
+	const struct resource_context *res_ctx = &dc->current_state->res_ctx;
 
-	for (i = 0; i < pool->hpo_dp_link_enc_count; i++) {
-		if (pool->hpo_dp_link_enc[i]->transmitter == TRANSMITTER_UNKNOWN) {
-			enc = pool->hpo_dp_link_enc[i];
+	memset(link_res, 0, sizeof(*link_res));
+
+	if (dc->link_srv->dp_get_encoding_format(link_settings) == DP_128b_132b_ENCODING) {
+		link_res->hpo_dp_link_enc = get_temp_hpo_dp_link_enc(res_ctx, dc->res_pool, link);
+		if (!link_res->hpo_dp_link_enc)
+			return false;
+	} else if (dc->link_srv->dp_get_encoding_format(link_settings) == DP_8b_10b_ENCODING &&
+				dc->config.unify_link_enc_assignment) {
+		link_res->dio_link_enc = get_temp_dio_link_enc(res_ctx,
+				dc->res_pool, link);
+		if (!link_res->dio_link_enc)
+			return false;
+	}
+
+	return true;
+}
+
+void reset_syncd_pipes_from_disabled_pipes(struct dc *dc,
+		struct dc_state *context)
+{
+	int i, j;
+	struct pipe_ctx *pipe_ctx_old, *pipe_ctx, *pipe_ctx_syncd;
+
+	/* If pipe backend is reset, need to reset pipe syncd status */
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		pipe_ctx_old =	&dc->current_state->res_ctx.pipe_ctx[i];
+		pipe_ctx = &context->res_ctx.pipe_ctx[i];
+
+		if (!resource_is_pipe_type(pipe_ctx_old, OTG_MASTER))
+			continue;
+
+		if (!pipe_ctx->stream ||
+				pipe_need_reprogram(pipe_ctx_old, pipe_ctx)) {
+
+			/* Reset all the syncd pipes from the disabled pipe */
+			for (j = 0; j < dc->res_pool->pipe_count; j++) {
+				pipe_ctx_syncd = &context->res_ctx.pipe_ctx[j];
+				if ((GET_PIPE_SYNCD_FROM_PIPE(pipe_ctx_syncd) == pipe_ctx_old->pipe_idx) ||
+					!IS_PIPE_SYNCD_VALID(pipe_ctx_syncd))
+					SET_PIPE_SYNCD_TO_PIPE(pipe_ctx_syncd, j);
+			}
+		}
+	}
+}
+
+void check_syncd_pipes_for_disabled_master_pipe(struct dc *dc,
+	struct dc_state *context,
+	uint8_t disabled_master_pipe_idx)
+{
+	int i;
+	struct pipe_ctx *pipe_ctx, *pipe_ctx_check;
+
+	pipe_ctx = &context->res_ctx.pipe_ctx[disabled_master_pipe_idx];
+	if ((GET_PIPE_SYNCD_FROM_PIPE(pipe_ctx) != disabled_master_pipe_idx) ||
+		!IS_PIPE_SYNCD_VALID(pipe_ctx))
+		SET_PIPE_SYNCD_TO_PIPE(pipe_ctx, disabled_master_pipe_idx);
+
+	/* for the pipe disabled, check if any slave pipe exists and assert */
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		pipe_ctx_check = &context->res_ctx.pipe_ctx[i];
+
+		if ((GET_PIPE_SYNCD_FROM_PIPE(pipe_ctx_check) == disabled_master_pipe_idx) &&
+		    IS_PIPE_SYNCD_VALID(pipe_ctx_check) && (i != disabled_master_pipe_idx)) {
+			struct pipe_ctx *first_pipe = pipe_ctx_check;
+
+			while (first_pipe->prev_odm_pipe)
+				first_pipe = first_pipe->prev_odm_pipe;
+			/* When ODM combine is enabled, this case is expected. If the disabled pipe
+			 * is part of the ODM tree, then we should not print an error.
+			 * */
+			if (first_pipe->pipe_idx == disabled_master_pipe_idx)
+				continue;
+
+			DC_ERR("DC: Failure: pipe_idx[%d] syncd with disabled master pipe_idx[%d]\n",
+				   i, disabled_master_pipe_idx);
+		}
+	}
+}
+
+void reset_sync_context_for_pipe(const struct dc *dc,
+	struct dc_state *context,
+	uint8_t pipe_idx)
+{
+	int i;
+	struct pipe_ctx *pipe_ctx_reset;
+
+	/* reset the otg sync context for the pipe and its slave pipes if any */
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		pipe_ctx_reset = &context->res_ctx.pipe_ctx[i];
+
+		if (((GET_PIPE_SYNCD_FROM_PIPE(pipe_ctx_reset) == pipe_idx) &&
+			IS_PIPE_SYNCD_VALID(pipe_ctx_reset)) || (i == pipe_idx))
+			SET_PIPE_SYNCD_TO_PIPE(pipe_ctx_reset, i);
+	}
+}
+
+uint8_t resource_transmitter_to_phy_idx(const struct dc *dc, enum transmitter transmitter)
+{
+	/* TODO - get transmitter to phy idx mapping from DMUB */
+	uint8_t phy_idx = transmitter - TRANSMITTER_UNIPHY_A;
+
+	if (dc->ctx->dce_version == DCN_VERSION_3_1 &&
+			dc->ctx->asic_id.hw_internal_rev == YELLOW_CARP_B0) {
+		switch (transmitter) {
+		case TRANSMITTER_UNIPHY_A:
+			phy_idx = 0;
+			break;
+		case TRANSMITTER_UNIPHY_B:
+			phy_idx = 1;
+			break;
+		case TRANSMITTER_UNIPHY_C:
+			phy_idx = 5;
+			break;
+		case TRANSMITTER_UNIPHY_D:
+			phy_idx = 6;
+			break;
+		case TRANSMITTER_UNIPHY_E:
+			phy_idx = 4;
+			break;
+		default:
+			phy_idx = 0;
 			break;
 		}
 	}
 
-	return enc;
+	return phy_idx;
 }
+
+const struct link_hwss *get_link_hwss(const struct dc_link *link,
+		const struct link_resource *link_res)
+{
+	/* Link_hwss is only accessible by getter function instead of accessing
+	 * by pointers in dc with the intent to protect against breaking polymorphism.
+	 */
+	if (can_use_hpo_dp_link_hwss(link, link_res))
+		/* TODO: some assumes that if decided link settings is 128b/132b
+		 * channel coding format hpo_dp_link_enc should be used.
+		 * Others believe that if hpo_dp_link_enc is available in link
+		 * resource then hpo_dp_link_enc must be used. This bound between
+		 * hpo_dp_link_enc != NULL and decided link settings is loosely coupled
+		 * with a premise that both hpo_dp_link_enc pointer and decided link
+		 * settings are determined based on single policy function like
+		 * "decide_link_settings" from upper layer. This "convention"
+		 * cannot be maintained and enforced at current level.
+		 * Therefore a refactor is due so we can enforce a strong bound
+		 * between those two parameters at this level.
+		 *
+		 * To put it simple, we want to make enforcement at low level so that
+		 * we will not return link hwss if caller plans to do 8b/10b
+		 * with an hpo encoder. Or we can return a very dummy one that doesn't
+		 * do work for all functions
+		 */
+		return (requires_fixed_vs_pe_retimer_hpo_link_hwss(link) ?
+				get_hpo_fixed_vs_pe_retimer_dp_link_hwss() : get_hpo_dp_link_hwss());
+	else if (can_use_dpia_link_hwss(link, link_res))
+		return get_dpia_link_hwss();
+	else if (can_use_dio_link_hwss(link, link_res))
+		return (requires_fixed_vs_pe_retimer_dio_link_hwss(link)) ?
+				get_dio_fixed_vs_pe_retimer_link_hwss() : get_dio_link_hwss();
+	else
+		return get_virtual_link_hwss();
+}
+
+bool is_h_timing_divisible_by_2(struct dc_stream_state *stream)
+{
+	bool divisible = false;
+	uint16_t h_blank_start = 0;
+	uint16_t h_blank_end = 0;
+
+	if (stream) {
+		h_blank_start = stream->timing.h_total - stream->timing.h_front_porch;
+		h_blank_end = h_blank_start - stream->timing.h_addressable;
+
+		/* HTOTAL, Hblank start/end, and Hsync start/end all must be
+		 * divisible by 2 in order for the horizontal timing params
+		 * to be considered divisible by 2. Hsync start is always 0.
+		 */
+		divisible = (stream->timing.h_total % 2 == 0) &&
+				(h_blank_start % 2 == 0) &&
+				(h_blank_end % 2 == 0) &&
+				(stream->timing.h_sync_width % 2 == 0);
+	}
+	return divisible;
+}
+
+/* This interface is deprecated for new DCNs. It is replaced by the following
+ * new interfaces. These two interfaces encapsulate pipe selection priority
+ * with DCN specific minimum hardware transition optimization algorithm. With
+ * the new interfaces caller no longer needs to know the implementation detail
+ * of a pipe topology.
+ *
+ * resource_update_pipes_with_odm_slice_count
+ * resource_update_pipes_with_mpc_slice_count
+ *
+ */
+bool dc_resource_acquire_secondary_pipe_for_mpc_odm_legacy(
+		const struct dc *dc,
+		struct dc_state *state,
+		struct pipe_ctx *pri_pipe,
+		struct pipe_ctx *sec_pipe,
+		bool odm)
+{
+	int pipe_idx = sec_pipe->pipe_idx;
+	struct pipe_ctx *sec_top, *sec_bottom, *sec_next, *sec_prev;
+	const struct resource_pool *pool = dc->res_pool;
+
+	sec_top = sec_pipe->top_pipe;
+	sec_bottom = sec_pipe->bottom_pipe;
+	sec_next = sec_pipe->next_odm_pipe;
+	sec_prev = sec_pipe->prev_odm_pipe;
+
+	if (pri_pipe == NULL)
+		return false;
+
+	*sec_pipe = *pri_pipe;
+
+	sec_pipe->top_pipe = sec_top;
+	sec_pipe->bottom_pipe = sec_bottom;
+	sec_pipe->next_odm_pipe = sec_next;
+	sec_pipe->prev_odm_pipe = sec_prev;
+
+	sec_pipe->pipe_idx = pipe_idx;
+	sec_pipe->plane_res.mi = pool->mis[pipe_idx];
+	sec_pipe->plane_res.hubp = pool->hubps[pipe_idx];
+	sec_pipe->plane_res.ipp = pool->ipps[pipe_idx];
+	sec_pipe->plane_res.xfm = pool->transforms[pipe_idx];
+	sec_pipe->plane_res.dpp = pool->dpps[pipe_idx];
+	sec_pipe->plane_res.mpcc_inst = pool->dpps[pipe_idx]->inst;
+	sec_pipe->stream_res.dsc = NULL;
+	if (odm) {
+		if (!sec_pipe->top_pipe)
+			sec_pipe->stream_res.opp = pool->opps[pipe_idx];
+		else
+			sec_pipe->stream_res.opp = sec_pipe->top_pipe->stream_res.opp;
+		if (sec_pipe->stream->timing.flags.DSC == 1) {
+#if defined(CONFIG_DRM_AMD_DC_FP)
+			dcn20_acquire_dsc(dc, &state->res_ctx, &sec_pipe->stream_res.dsc, sec_pipe->stream_res.opp->inst);
 #endif
+			ASSERT(sec_pipe->stream_res.dsc);
+			if (sec_pipe->stream_res.dsc == NULL)
+				return false;
+		}
+#if defined(CONFIG_DRM_AMD_DC_FP)
+		dcn20_build_mapped_resource(dc, state, sec_pipe->stream);
+#endif
+	}
+
+	return true;
+}
+
+enum dc_status update_dp_encoder_resources_for_test_harness(const struct dc *dc,
+		struct dc_state *context,
+		struct pipe_ctx *pipe_ctx)
+{
+	if (dc->link_srv->dp_get_encoding_format(&pipe_ctx->link_config.dp_link_settings) == DP_128b_132b_ENCODING) {
+		if (pipe_ctx->stream_res.hpo_dp_stream_enc == NULL) {
+			pipe_ctx->stream_res.hpo_dp_stream_enc =
+					find_first_free_match_hpo_dp_stream_enc_for_link(
+							&context->res_ctx, dc->res_pool, pipe_ctx->stream);
+
+			if (!pipe_ctx->stream_res.hpo_dp_stream_enc)
+				return DC_NO_STREAM_ENC_RESOURCE;
+
+			update_hpo_dp_stream_engine_usage(
+					&context->res_ctx, dc->res_pool,
+					pipe_ctx->stream_res.hpo_dp_stream_enc,
+					true);
+		}
+
+		if (pipe_ctx->link_res.hpo_dp_link_enc == NULL) {
+			if (!add_hpo_dp_link_enc_to_ctx(&context->res_ctx, dc->res_pool, pipe_ctx, pipe_ctx->stream))
+				return DC_NO_LINK_ENC_RESOURCE;
+		}
+	} else {
+		if (pipe_ctx->stream_res.hpo_dp_stream_enc) {
+			update_hpo_dp_stream_engine_usage(
+					&context->res_ctx, dc->res_pool,
+					pipe_ctx->stream_res.hpo_dp_stream_enc,
+					false);
+			pipe_ctx->stream_res.hpo_dp_stream_enc = NULL;
+		}
+		if (pipe_ctx->link_res.hpo_dp_link_enc)
+			remove_hpo_dp_link_enc_from_ctx(&context->res_ctx, pipe_ctx, pipe_ctx->stream);
+	}
+
+	if (pipe_ctx->link_res.dio_link_enc == NULL && dc->config.unify_link_enc_assignment)
+		if (!add_dio_link_enc_to_ctx(dc, context, dc->res_pool, pipe_ctx, pipe_ctx->stream))
+			return DC_NO_LINK_ENC_RESOURCE;
+
+	return DC_OK;
+}
+
+struct dscl_prog_data *resource_get_dscl_prog_data(struct pipe_ctx *pipe_ctx)
+{
+	return &pipe_ctx->plane_res.scl_data.dscl_prog_data;
+}
+
+static bool resource_allocate_mcache(struct dc_state *context, const struct  dc_mcache_params *mcache_params)
+{
+	if (context->clk_mgr->ctx->dc->res_pool->funcs->program_mcache_pipe_config)
+		context->clk_mgr->ctx->dc->res_pool->funcs->program_mcache_pipe_config(context, mcache_params);
+
+	return true;
+}
+
+void resource_init_common_dml2_callbacks(struct dc *dc, struct dml2_configuration_options *dml2_options)
+{
+	dml2_options->callbacks.dc = dc;
+	dml2_options->callbacks.build_scaling_params = &resource_build_scaling_params;
+	dml2_options->callbacks.build_test_pattern_params = &resource_build_test_pattern_params;
+	dml2_options->callbacks.acquire_secondary_pipe_for_mpc_odm = &dc_resource_acquire_secondary_pipe_for_mpc_odm_legacy;
+	dml2_options->callbacks.update_pipes_for_stream_with_slice_count = &resource_update_pipes_for_stream_with_slice_count;
+	dml2_options->callbacks.update_pipes_for_plane_with_slice_count = &resource_update_pipes_for_plane_with_slice_count;
+	dml2_options->callbacks.get_mpc_slice_index = &resource_get_mpc_slice_index;
+	dml2_options->callbacks.get_mpc_slice_count = &resource_get_mpc_slice_count;
+	dml2_options->callbacks.get_odm_slice_index = &resource_get_odm_slice_index;
+	dml2_options->callbacks.get_odm_slice_count = &resource_get_odm_slice_count;
+	dml2_options->callbacks.get_opp_head = &resource_get_opp_head;
+	dml2_options->callbacks.get_otg_master_for_stream = &resource_get_otg_master_for_stream;
+	dml2_options->callbacks.get_opp_heads_for_otg_master = &resource_get_opp_heads_for_otg_master;
+	dml2_options->callbacks.get_dpp_pipes_for_plane = &resource_get_dpp_pipes_for_plane;
+	dml2_options->callbacks.get_stream_status = &dc_state_get_stream_status;
+	dml2_options->callbacks.get_stream_from_id = &dc_state_get_stream_from_id;
+	dml2_options->callbacks.get_max_flickerless_instant_vtotal_increase = &dc_stream_get_max_flickerless_instant_vtotal_increase;
+	dml2_options->callbacks.allocate_mcache = &resource_allocate_mcache;
+
+	dml2_options->svp_pstate.callbacks.dc = dc;
+	dml2_options->svp_pstate.callbacks.add_phantom_plane = &dc_state_add_phantom_plane;
+	dml2_options->svp_pstate.callbacks.add_phantom_stream = &dc_state_add_phantom_stream;
+	dml2_options->svp_pstate.callbacks.build_scaling_params = &resource_build_scaling_params;
+	dml2_options->svp_pstate.callbacks.create_phantom_plane = &dc_state_create_phantom_plane;
+	dml2_options->svp_pstate.callbacks.remove_phantom_plane = &dc_state_remove_phantom_plane;
+	dml2_options->svp_pstate.callbacks.remove_phantom_stream = &dc_state_remove_phantom_stream;
+	dml2_options->svp_pstate.callbacks.create_phantom_stream = &dc_state_create_phantom_stream;
+	dml2_options->svp_pstate.callbacks.release_phantom_plane = &dc_state_release_phantom_plane;
+	dml2_options->svp_pstate.callbacks.release_phantom_stream = &dc_state_release_phantom_stream;
+	dml2_options->svp_pstate.callbacks.get_pipe_subvp_type = &dc_state_get_pipe_subvp_type;
+	dml2_options->svp_pstate.callbacks.get_stream_subvp_type = &dc_state_get_stream_subvp_type;
+	dml2_options->svp_pstate.callbacks.get_paired_subvp_stream = &dc_state_get_paired_subvp_stream;
+	dml2_options->svp_pstate.callbacks.remove_phantom_streams_and_planes = &dc_state_remove_phantom_streams_and_planes;
+	dml2_options->svp_pstate.callbacks.release_phantom_streams_and_planes = &dc_state_release_phantom_streams_and_planes;
+}
+
+/* Returns number of DET segments allocated for a given OTG_MASTER pipe */
+int resource_calculate_det_for_stream(struct dc_state *state, struct pipe_ctx *otg_master)
+{
+	struct pipe_ctx *opp_heads[MAX_PIPES];
+	struct pipe_ctx *dpp_pipes[MAX_PIPES];
+
+	int dpp_count = 0;
+	int det_segments = 0;
+
+	if (!otg_master->stream)
+		return 0;
+
+	int slice_count = resource_get_opp_heads_for_otg_master(otg_master,
+			&state->res_ctx, opp_heads);
+
+	for (int slice_idx = 0; slice_idx < slice_count; slice_idx++) {
+		if (opp_heads[slice_idx]->plane_state) {
+			dpp_count = resource_get_dpp_pipes_for_opp_head(
+					opp_heads[slice_idx],
+					&state->res_ctx,
+					dpp_pipes);
+			for (int dpp_idx = 0; dpp_idx < dpp_count; dpp_idx++)
+				det_segments += dpp_pipes[dpp_idx]->hubp_regs.det_size;
+		}
+	}
+	return det_segments;
+}
+
+bool resource_is_hpo_acquired(struct dc_state *context)
+{
+	int i;
+
+	for (i = 0; i < MAX_HPO_DP2_ENCODERS; i++) {
+		if (context->res_ctx.is_hpo_dp_stream_enc_acquired[i]) {
+			return true;
+		}
+	}
+
+	return false;
+}
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_sink.c b/drivers/gpu/drm/amd/display/dc/core/dc_sink.c
index a249a0e5edd0..455fa5dd1420 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_sink.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_sink.c
@@ -23,8 +23,6 @@
  *
  */
 
-#include <linux/slab.h>
-
 #include "dm_services.h"
 #include "dm_helpers.h"
 #include "core_types.h"
@@ -33,14 +31,6 @@
  * Private functions
  ******************************************************************************/
 
-static void dc_sink_destruct(struct dc_sink *sink)
-{
-	if (sink->dc_container_id) {
-		kfree(sink->dc_container_id);
-		sink->dc_container_id = NULL;
-	}
-}
-
 static bool dc_sink_construct(struct dc_sink *sink, const struct dc_sink_init_data *init_params)
 {
 
@@ -75,7 +65,7 @@ void dc_sink_retain(struct dc_sink *sink)
 static void dc_sink_free(struct kref *kref)
 {
 	struct dc_sink *sink = container_of(kref, struct dc_sink, refcount);
-	dc_sink_destruct(sink);
+	kfree(sink->dc_container_id);
 	kfree(sink);
 }
 
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_stat.c b/drivers/gpu/drm/amd/display/dc/core/dc_stat.c
index 4b372aa52801..f976ffd6d466 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_stat.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_stat.c
@@ -35,19 +35,15 @@
  */
 
 /**
- *****************************************************************************
- *  Function: dc_stat_get_dmub_notification
+ *  dc_stat_get_dmub_notification
  *
- *  @brief
- *		Calls dmub layer to retrieve dmub notification
+ * Calls dmub layer to retrieve dmub notification
  *
- *  @param
- *		[in] dc: dc structure
- *		[in] notify: dmub notification structure
+ * @dc: dc structure
+ * @notify: dmub notification structure
  *
- *  @return
+ * Returns
  *     None
- *****************************************************************************
  */
 void dc_stat_get_dmub_notification(const struct dc *dc, struct dmub_notification *notify)
 {
@@ -65,26 +61,24 @@ void dc_stat_get_dmub_notification(const struct dc *dc, struct dmub_notification
 	/* For HPD/HPD RX, convert dpia port index into link index */
 	if (notify->type == DMUB_NOTIFICATION_HPD ||
 	    notify->type == DMUB_NOTIFICATION_HPD_IRQ ||
+	    notify->type == DMUB_NOTIFICATION_AUX_REPLY ||
+	    notify->type == DMUB_NOTIFICATION_DPIA_NOTIFICATION ||
 	    notify->type == DMUB_NOTIFICATION_SET_CONFIG_REPLY) {
 		notify->link_index =
-			get_link_index_from_dpia_port_index(dc, notify->link_index);
+			get_link_index_from_dpia_port_index(dc, notify->instance);
 	}
 }
 
 /**
- *****************************************************************************
- *  Function: dc_stat_get_dmub_dataout
+ * dc_stat_get_dmub_dataout
  *
- *  @brief
- *		Calls dmub layer to retrieve dmub gpint dataout
+ * Calls dmub layer to retrieve dmub gpint dataout
  *
- *  @param
- *		[in] dc: dc structure
- *		[in] dataout: dmub gpint dataout
+ * @dc: dc structure
+ * @dataout: dmub gpint dataout
  *
- *  @return
+ * Returns
  *     None
- *****************************************************************************
  */
 void dc_stat_get_dmub_dataout(const struct dc *dc, uint32_t *dataout)
 {
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_state.c b/drivers/gpu/drm/amd/display/dc/core/dc_state.c
new file mode 100644
index 000000000000..2de8ef4a58ec
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_state.c
@@ -0,0 +1,1077 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+#include "dc_types.h"
+#include "core_types.h"
+#include "core_status.h"
+#include "dc_state.h"
+#include "dc_state_priv.h"
+#include "dc_stream_priv.h"
+#include "dc_plane_priv.h"
+
+#include "dm_services.h"
+#include "resource.h"
+#include "link_enc_cfg.h"
+
+#if defined(CONFIG_DRM_AMD_DC_FP)
+#include "dml2_0/dml2_wrapper.h"
+#include "dml2_0/dml2_internal_types.h"
+#endif
+
+#define DC_LOGGER \
+	dc->ctx->logger
+#define DC_LOGGER_INIT(logger)
+
+/* Private dc_state helper functions */
+static bool dc_state_track_phantom_stream(struct dc_state *state,
+		struct dc_stream_state *phantom_stream)
+{
+	if (state->phantom_stream_count >= MAX_PHANTOM_PIPES)
+		return false;
+
+	state->phantom_streams[state->phantom_stream_count++] = phantom_stream;
+
+	return true;
+}
+
+static bool dc_state_untrack_phantom_stream(struct dc_state *state, struct dc_stream_state *phantom_stream)
+{
+	bool res = false;
+	int i;
+
+	/* first find phantom stream in the dc_state */
+	for (i = 0; i < state->phantom_stream_count; i++) {
+		if (state->phantom_streams[i] == phantom_stream) {
+			state->phantom_streams[i] = NULL;
+			res = true;
+			break;
+		}
+	}
+
+	/* failed to find stream in state */
+	if (!res)
+		return res;
+
+	/* trim back phantom streams */
+	state->phantom_stream_count--;
+	for (; i < state->phantom_stream_count; i++)
+		state->phantom_streams[i] = state->phantom_streams[i + 1];
+
+	return res;
+}
+
+static bool dc_state_is_phantom_stream_tracked(struct dc_state *state, struct dc_stream_state *phantom_stream)
+{
+	int i;
+
+	for (i = 0; i < state->phantom_stream_count; i++) {
+		if (state->phantom_streams[i] == phantom_stream)
+			return true;
+	}
+
+	return false;
+}
+
+static bool dc_state_track_phantom_plane(struct dc_state *state,
+		struct dc_plane_state *phantom_plane)
+{
+	if (state->phantom_plane_count >= MAX_PHANTOM_PIPES)
+		return false;
+
+	state->phantom_planes[state->phantom_plane_count++] = phantom_plane;
+
+	return true;
+}
+
+static bool dc_state_untrack_phantom_plane(struct dc_state *state, struct dc_plane_state *phantom_plane)
+{
+	bool res = false;
+	int i;
+
+	/* first find phantom plane in the dc_state */
+	for (i = 0; i < state->phantom_plane_count; i++) {
+		if (state->phantom_planes[i] == phantom_plane) {
+			state->phantom_planes[i] = NULL;
+			res = true;
+			break;
+		}
+	}
+
+	/* failed to find plane in state */
+	if (!res)
+		return res;
+
+	/* trim back phantom planes */
+	state->phantom_plane_count--;
+	for (; i < state->phantom_plane_count; i++)
+		state->phantom_planes[i] = state->phantom_planes[i + 1];
+
+	return res;
+}
+
+static bool dc_state_is_phantom_plane_tracked(struct dc_state *state, struct dc_plane_state *phantom_plane)
+{
+	int i;
+
+	for (i = 0; i < state->phantom_plane_count; i++) {
+		if (state->phantom_planes[i] == phantom_plane)
+			return true;
+	}
+
+	return false;
+}
+
+static void dc_state_copy_internal(struct dc_state *dst_state, struct dc_state *src_state)
+{
+	int i, j;
+
+	memcpy(dst_state, src_state, sizeof(struct dc_state));
+
+	for (i = 0; i < MAX_PIPES; i++) {
+		struct pipe_ctx *cur_pipe = &dst_state->res_ctx.pipe_ctx[i];
+
+		if (cur_pipe->top_pipe)
+			cur_pipe->top_pipe =  &dst_state->res_ctx.pipe_ctx[cur_pipe->top_pipe->pipe_idx];
+
+		if (cur_pipe->bottom_pipe)
+			cur_pipe->bottom_pipe = &dst_state->res_ctx.pipe_ctx[cur_pipe->bottom_pipe->pipe_idx];
+
+		if (cur_pipe->prev_odm_pipe)
+			cur_pipe->prev_odm_pipe =  &dst_state->res_ctx.pipe_ctx[cur_pipe->prev_odm_pipe->pipe_idx];
+
+		if (cur_pipe->next_odm_pipe)
+			cur_pipe->next_odm_pipe = &dst_state->res_ctx.pipe_ctx[cur_pipe->next_odm_pipe->pipe_idx];
+	}
+
+	/* retain phantoms */
+	for (i = 0; i < dst_state->phantom_stream_count; i++)
+		dc_stream_retain(dst_state->phantom_streams[i]);
+
+	for (i = 0; i < dst_state->phantom_plane_count; i++)
+		dc_plane_state_retain(dst_state->phantom_planes[i]);
+
+	/* retain streams and planes */
+	for (i = 0; i < dst_state->stream_count; i++) {
+		dc_stream_retain(dst_state->streams[i]);
+		for (j = 0; j < dst_state->stream_status[i].plane_count; j++)
+			dc_plane_state_retain(
+					dst_state->stream_status[i].plane_states[j]);
+	}
+
+}
+
+static void init_state(struct dc *dc, struct dc_state *state)
+{
+	/* Each context must have their own instance of VBA and in order to
+	 * initialize and obtain IP and SOC the base DML instance from DC is
+	 * initially copied into every context
+	 */
+	memcpy(&state->bw_ctx.dml, &dc->dml, sizeof(struct display_mode_lib));
+}
+
+/* Public dc_state functions */
+struct dc_state *dc_state_create(struct dc *dc, struct dc_state_create_params *params)
+{
+	struct dc_state *state;
+
+	state = kvzalloc(sizeof(struct dc_state), GFP_KERNEL);
+
+	if (!state)
+		return NULL;
+
+	init_state(dc, state);
+	dc_state_construct(dc, state);
+	state->power_source = params ? params->power_source : DC_POWER_SOURCE_AC;
+
+#ifdef CONFIG_DRM_AMD_DC_FP
+	if (dc->debug.using_dml2) {
+		if (!dml2_create(dc, &dc->dml2_options, &state->bw_ctx.dml2)) {
+			dc_state_release(state);
+			return NULL;
+		}
+
+		if (dc->caps.dcmode_power_limits_present && !dml2_create(dc, &dc->dml2_dc_power_options, &state->bw_ctx.dml2_dc_power_source)) {
+			dc_state_release(state);
+			return NULL;
+		}
+	}
+#endif
+
+	kref_init(&state->refcount);
+
+	return state;
+}
+
+void dc_state_copy(struct dc_state *dst_state, struct dc_state *src_state)
+{
+	struct kref refcount = dst_state->refcount;
+#ifdef CONFIG_DRM_AMD_DC_FP
+	struct dml2_context *dst_dml2 = dst_state->bw_ctx.dml2;
+	struct dml2_context *dst_dml2_dc_power_source = dst_state->bw_ctx.dml2_dc_power_source;
+#endif
+
+	dc_state_copy_internal(dst_state, src_state);
+
+#ifdef CONFIG_DRM_AMD_DC_FP
+	dst_state->bw_ctx.dml2 = dst_dml2;
+	if (src_state->bw_ctx.dml2)
+		dml2_copy(dst_state->bw_ctx.dml2, src_state->bw_ctx.dml2);
+
+	dst_state->bw_ctx.dml2_dc_power_source = dst_dml2_dc_power_source;
+	if (src_state->bw_ctx.dml2_dc_power_source)
+		dml2_copy(dst_state->bw_ctx.dml2_dc_power_source, src_state->bw_ctx.dml2_dc_power_source);
+#endif
+
+	/* context refcount should not be overridden */
+	dst_state->refcount = refcount;
+}
+
+struct dc_state *dc_state_create_copy(struct dc_state *src_state)
+{
+	struct dc_state *new_state;
+
+	new_state = kvmalloc(sizeof(struct dc_state),
+			GFP_KERNEL);
+	if (!new_state)
+		return NULL;
+
+	dc_state_copy_internal(new_state, src_state);
+
+#ifdef CONFIG_DRM_AMD_DC_FP
+	new_state->bw_ctx.dml2 = NULL;
+	new_state->bw_ctx.dml2_dc_power_source = NULL;
+
+	if (src_state->bw_ctx.dml2 &&
+			!dml2_create_copy(&new_state->bw_ctx.dml2, src_state->bw_ctx.dml2)) {
+		dc_state_release(new_state);
+		return NULL;
+	}
+
+	if (src_state->bw_ctx.dml2_dc_power_source &&
+			!dml2_create_copy(&new_state->bw_ctx.dml2_dc_power_source, src_state->bw_ctx.dml2_dc_power_source)) {
+		dc_state_release(new_state);
+		return NULL;
+	}
+#endif
+
+	kref_init(&new_state->refcount);
+
+	return new_state;
+}
+
+void dc_state_copy_current(struct dc *dc, struct dc_state *dst_state)
+{
+	dc_state_copy(dst_state, dc->current_state);
+}
+
+struct dc_state *dc_state_create_current_copy(struct dc *dc)
+{
+	return dc_state_create_copy(dc->current_state);
+}
+
+void dc_state_construct(struct dc *dc, struct dc_state *state)
+{
+	state->clk_mgr = dc->clk_mgr;
+
+	/* Initialise DIG link encoder resource tracking variables. */
+	if (dc->res_pool)
+		link_enc_cfg_init(dc, state);
+}
+
+void dc_state_destruct(struct dc_state *state)
+{
+	int i, j;
+
+	for (i = 0; i < state->stream_count; i++) {
+		for (j = 0; j < state->stream_status[i].plane_count; j++)
+			dc_plane_state_release(
+					state->stream_status[i].plane_states[j]);
+
+		state->stream_status[i].plane_count = 0;
+		dc_stream_release(state->streams[i]);
+		state->streams[i] = NULL;
+	}
+	state->stream_count = 0;
+
+	/* release tracked phantoms */
+	for (i = 0; i < state->phantom_stream_count; i++) {
+		dc_stream_release(state->phantom_streams[i]);
+		state->phantom_streams[i] = NULL;
+	}
+	state->phantom_stream_count = 0;
+
+	for (i = 0; i < state->phantom_plane_count; i++) {
+		dc_plane_state_release(state->phantom_planes[i]);
+		state->phantom_planes[i] = NULL;
+	}
+	state->phantom_plane_count = 0;
+
+	state->stream_mask = 0;
+	memset(&state->res_ctx, 0, sizeof(state->res_ctx));
+	memset(&state->pp_display_cfg, 0, sizeof(state->pp_display_cfg));
+	memset(&state->dcn_bw_vars, 0, sizeof(state->dcn_bw_vars));
+	state->clk_mgr = NULL;
+	memset(&state->bw_ctx.bw, 0, sizeof(state->bw_ctx.bw));
+	memset(state->block_sequence, 0, sizeof(state->block_sequence));
+	state->block_sequence_steps = 0;
+	memset(state->dc_dmub_cmd, 0, sizeof(state->dc_dmub_cmd));
+	state->dmub_cmd_count = 0;
+	memset(&state->perf_params, 0, sizeof(state->perf_params));
+}
+
+void dc_state_retain(struct dc_state *state)
+{
+	kref_get(&state->refcount);
+}
+
+static void dc_state_free(struct kref *kref)
+{
+	struct dc_state *state = container_of(kref, struct dc_state, refcount);
+
+	dc_state_destruct(state);
+
+#ifdef CONFIG_DRM_AMD_DC_FP
+	dml2_destroy(state->bw_ctx.dml2);
+	state->bw_ctx.dml2 = 0;
+
+	dml2_destroy(state->bw_ctx.dml2_dc_power_source);
+	state->bw_ctx.dml2_dc_power_source = 0;
+#endif
+
+	kvfree(state);
+}
+
+void dc_state_release(struct dc_state *state)
+{
+	if (state != NULL)
+		kref_put(&state->refcount, dc_state_free);
+}
+/*
+ * dc_state_add_stream() - Add a new dc_stream_state to a dc_state.
+ */
+enum dc_status dc_state_add_stream(
+		const struct dc *dc,
+		struct dc_state *state,
+		struct dc_stream_state *stream)
+{
+	enum dc_status res;
+
+	DC_LOGGER_INIT(dc->ctx->logger);
+
+	if (state->stream_count >= dc->res_pool->timing_generator_count) {
+		DC_LOG_WARNING("Max streams reached, can't add stream %p !\n", stream);
+		return DC_ERROR_UNEXPECTED;
+	}
+
+	state->streams[state->stream_count] = stream;
+	dc_stream_retain(stream);
+	state->stream_count++;
+
+	res = resource_add_otg_master_for_stream_output(
+			state, dc->res_pool, stream);
+	if (res != DC_OK)
+		DC_LOG_WARNING("Adding stream %p to context failed with err %d!\n", stream, res);
+
+	return res;
+}
+
+/*
+ * dc_state_remove_stream() - Remove a stream from a dc_state.
+ */
+enum dc_status dc_state_remove_stream(
+		const struct dc *dc,
+		struct dc_state *state,
+		struct dc_stream_state *stream)
+{
+	int i;
+	struct pipe_ctx *del_pipe = resource_get_otg_master_for_stream(
+			&state->res_ctx, stream);
+
+	if (!del_pipe) {
+		dm_error("Pipe not found for stream %p !\n", stream);
+		return DC_ERROR_UNEXPECTED;
+	}
+
+	resource_update_pipes_for_stream_with_slice_count(state,
+			dc->current_state, dc->res_pool, stream, 1);
+	resource_remove_otg_master_for_stream_output(
+			state, dc->res_pool, stream);
+
+	for (i = 0; i < state->stream_count; i++)
+		if (state->streams[i] == stream)
+			break;
+
+	if (state->streams[i] != stream) {
+		dm_error("Context doesn't have stream %p !\n", stream);
+		return DC_ERROR_UNEXPECTED;
+	}
+
+	dc_stream_release_3dlut_for_stream(dc, stream);
+
+	dc_stream_release(state->streams[i]);
+	state->stream_count--;
+
+	/* Trim back arrays */
+	for (; i < state->stream_count; i++) {
+		state->streams[i] = state->streams[i + 1];
+		state->stream_status[i] = state->stream_status[i + 1];
+	}
+
+	state->streams[state->stream_count] = NULL;
+	memset(
+			&state->stream_status[state->stream_count],
+			0,
+			sizeof(state->stream_status[0]));
+
+	return DC_OK;
+}
+
+static void remove_mpc_combine_for_stream(const struct dc *dc,
+		struct dc_state *new_ctx,
+		const struct dc_state *cur_ctx,
+		struct dc_stream_status *status)
+{
+	int i;
+
+	for (i = 0; i < status->plane_count; i++)
+		resource_update_pipes_for_plane_with_slice_count(
+				new_ctx, cur_ctx, dc->res_pool,
+				status->plane_states[i], 1);
+}
+
+bool dc_state_add_plane(
+		const struct dc *dc,
+		struct dc_stream_state *stream,
+		struct dc_plane_state *plane_state,
+		struct dc_state *state)
+{
+	struct resource_pool *pool = dc->res_pool;
+	struct pipe_ctx *otg_master_pipe;
+	struct dc_stream_status *stream_status = NULL;
+	bool added = false;
+	int odm_slice_count;
+	int i;
+
+	stream_status = dc_state_get_stream_status(state, stream);
+	otg_master_pipe = resource_get_otg_master_for_stream(
+			&state->res_ctx, stream);
+	if (stream_status == NULL) {
+		dm_error("Existing stream not found; failed to attach surface!\n");
+		goto out;
+	} else if (stream_status->plane_count == MAX_SURFACES) {
+		dm_error("Surface: can not attach plane_state %p! Maximum is: %d\n",
+				plane_state, MAX_SURFACES);
+		goto out;
+	} else if (!otg_master_pipe) {
+		goto out;
+	}
+
+	added = resource_append_dpp_pipes_for_plane_composition(state,
+			dc->current_state, pool, otg_master_pipe, plane_state);
+
+	if (!added) {
+		/* try to remove MPC combine to free up pipes */
+		for (i = 0; i < state->stream_count; i++)
+			remove_mpc_combine_for_stream(dc, state,
+					dc->current_state,
+					&state->stream_status[i]);
+		added = resource_append_dpp_pipes_for_plane_composition(state,
+					dc->current_state, pool,
+					otg_master_pipe, plane_state);
+	}
+
+	if (!added) {
+		/* try to decrease ODM slice count gradually to free up pipes */
+		odm_slice_count = resource_get_odm_slice_count(otg_master_pipe);
+		for (i = odm_slice_count - 1; i > 0; i--) {
+			resource_update_pipes_for_stream_with_slice_count(state,
+					dc->current_state, dc->res_pool, stream,
+					i);
+			added = resource_append_dpp_pipes_for_plane_composition(
+					state,
+					dc->current_state, pool,
+					otg_master_pipe, plane_state);
+			if (added)
+				break;
+		}
+	}
+
+	if (added) {
+		stream_status->plane_states[stream_status->plane_count] =
+				plane_state;
+		stream_status->plane_count++;
+		dc_plane_state_retain(plane_state);
+	}
+
+out:
+	return added;
+}
+
+bool dc_state_remove_plane(
+		const struct dc *dc,
+		struct dc_stream_state *stream,
+		struct dc_plane_state *plane_state,
+		struct dc_state *state)
+{
+	int i;
+	struct dc_stream_status *stream_status = NULL;
+	struct resource_pool *pool = dc->res_pool;
+
+	if (!plane_state)
+		return true;
+
+	for (i = 0; i < state->stream_count; i++)
+		if (state->streams[i] == stream) {
+			stream_status = &state->stream_status[i];
+			break;
+		}
+
+	if (stream_status == NULL) {
+		dm_error("Existing stream not found; failed to remove plane.\n");
+		return false;
+	}
+
+	resource_remove_dpp_pipes_for_plane_composition(
+			state, pool, plane_state);
+
+	for (i = 0; i < stream_status->plane_count; i++) {
+		if (stream_status->plane_states[i] == plane_state) {
+			dc_plane_state_release(stream_status->plane_states[i]);
+			break;
+		}
+	}
+
+	if (i == stream_status->plane_count) {
+		dm_error("Existing plane_state not found; failed to detach it!\n");
+		return false;
+	}
+
+	stream_status->plane_count--;
+
+	/* Start at the plane we've just released, and move all the planes one index forward to "trim" the array */
+	for (; i < stream_status->plane_count; i++)
+		stream_status->plane_states[i] = stream_status->plane_states[i + 1];
+
+	stream_status->plane_states[stream_status->plane_count] = NULL;
+
+	return true;
+}
+
+/**
+ * dc_state_rem_all_planes_for_stream - Remove planes attached to the target stream.
+ *
+ * @dc: Current dc state.
+ * @stream: Target stream, which we want to remove the attached plans.
+ * @state: context from which the planes are to be removed.
+ *
+ * Return:
+ * Return true if DC was able to remove all planes from the target
+ * stream, otherwise, return false.
+ */
+bool dc_state_rem_all_planes_for_stream(
+		const struct dc *dc,
+		struct dc_stream_state *stream,
+		struct dc_state *state)
+{
+	int i, old_plane_count;
+	struct dc_stream_status *stream_status = NULL;
+	struct dc_plane_state *del_planes[MAX_SURFACES] = { 0 };
+
+	for (i = 0; i < state->stream_count; i++)
+		if (state->streams[i] == stream) {
+			stream_status = &state->stream_status[i];
+			break;
+		}
+
+	if (stream_status == NULL) {
+		dm_error("Existing stream %p not found!\n", stream);
+		return false;
+	}
+
+	old_plane_count = stream_status->plane_count;
+
+	for (i = 0; i < old_plane_count; i++)
+		del_planes[i] = stream_status->plane_states[i];
+
+	for (i = 0; i < old_plane_count; i++)
+		if (!dc_state_remove_plane(dc, stream, del_planes[i], state))
+			return false;
+
+	return true;
+}
+
+bool dc_state_add_all_planes_for_stream(
+		const struct dc *dc,
+		struct dc_stream_state *stream,
+		struct dc_plane_state * const *plane_states,
+		int plane_count,
+		struct dc_state *state)
+{
+	int i;
+	bool result = true;
+
+	for (i = 0; i < plane_count; i++)
+		if (!dc_state_add_plane(dc, stream, plane_states[i], state)) {
+			result = false;
+			break;
+		}
+
+	return result;
+}
+
+/* Private dc_state functions */
+
+/**
+ * dc_state_get_stream_status - Get stream status from given dc state
+ * @state: DC state to find the stream status in
+ * @stream: The stream to get the stream status for
+ *
+ * The given stream is expected to exist in the given dc state. Otherwise, NULL
+ * will be returned.
+ */
+struct dc_stream_status *dc_state_get_stream_status(
+		struct dc_state *state,
+		const struct dc_stream_state *stream)
+{
+	uint8_t i;
+
+	if (state == NULL)
+		return NULL;
+
+	for (i = 0; i < state->stream_count; i++) {
+		if (stream == state->streams[i])
+			return &state->stream_status[i];
+	}
+
+	return NULL;
+}
+
+enum mall_stream_type dc_state_get_pipe_subvp_type(const struct dc_state *state,
+		const struct pipe_ctx *pipe_ctx)
+{
+	return dc_state_get_stream_subvp_type(state, pipe_ctx->stream);
+}
+
+enum mall_stream_type dc_state_get_stream_subvp_type(const struct dc_state *state,
+		const struct dc_stream_state *stream)
+{
+	int i;
+
+	enum mall_stream_type type = SUBVP_NONE;
+
+	for (i = 0; i < state->stream_count; i++) {
+		if (state->streams[i] == stream) {
+			type = state->stream_status[i].mall_stream_config.type;
+			break;
+		}
+	}
+
+	return type;
+}
+
+struct dc_stream_state *dc_state_get_paired_subvp_stream(const struct dc_state *state,
+		const struct dc_stream_state *stream)
+{
+	int i;
+
+	struct dc_stream_state *paired_stream = NULL;
+
+	for (i = 0; i < state->stream_count; i++) {
+		if (state->streams[i] == stream) {
+			paired_stream = state->stream_status[i].mall_stream_config.paired_stream;
+			break;
+		}
+	}
+
+	return paired_stream;
+}
+
+struct dc_stream_state *dc_state_create_phantom_stream(const struct dc *dc,
+		struct dc_state *state,
+		struct dc_stream_state *main_stream)
+{
+	struct dc_stream_state *phantom_stream;
+
+	DC_LOGGER_INIT(dc->ctx->logger);
+
+	phantom_stream = dc_create_stream_for_sink(main_stream->sink);
+
+	if (!phantom_stream) {
+		DC_LOG_ERROR("Failed to allocate phantom stream.\n");
+		return NULL;
+	}
+
+	/* track phantom stream in dc_state */
+	dc_state_track_phantom_stream(state, phantom_stream);
+
+	phantom_stream->is_phantom = true;
+	phantom_stream->signal = SIGNAL_TYPE_VIRTUAL;
+	phantom_stream->dpms_off = true;
+
+	return phantom_stream;
+}
+
+void dc_state_release_phantom_stream(const struct dc *dc,
+		struct dc_state *state,
+		struct dc_stream_state *phantom_stream)
+{
+	DC_LOGGER_INIT(dc->ctx->logger);
+
+	if (!dc_state_untrack_phantom_stream(state, phantom_stream)) {
+		DC_LOG_ERROR("Failed to free phantom stream %p in dc state %p.\n", phantom_stream, state);
+		return;
+	}
+
+	dc_stream_release(phantom_stream);
+}
+
+struct dc_plane_state *dc_state_create_phantom_plane(const struct dc *dc,
+		struct dc_state *state,
+		struct dc_plane_state *main_plane)
+{
+	struct dc_plane_state *phantom_plane = dc_create_plane_state(dc);
+
+	DC_LOGGER_INIT(dc->ctx->logger);
+
+	if (!phantom_plane) {
+		DC_LOG_ERROR("Failed to allocate phantom plane.\n");
+		return NULL;
+	}
+
+	/* track phantom inside dc_state */
+	dc_state_track_phantom_plane(state, phantom_plane);
+
+	phantom_plane->is_phantom = true;
+
+	return phantom_plane;
+}
+
+void dc_state_release_phantom_plane(const struct dc *dc,
+		struct dc_state *state,
+		struct dc_plane_state *phantom_plane)
+{
+	DC_LOGGER_INIT(dc->ctx->logger);
+
+	if (!dc_state_untrack_phantom_plane(state, phantom_plane)) {
+		DC_LOG_ERROR("Failed to free phantom plane %p in dc state %p.\n", phantom_plane, state);
+		return;
+	}
+
+	dc_plane_state_release(phantom_plane);
+}
+
+/* add phantom streams to context and generate correct meta inside dc_state */
+enum dc_status dc_state_add_phantom_stream(const struct dc *dc,
+		struct dc_state *state,
+		struct dc_stream_state *phantom_stream,
+		struct dc_stream_state *main_stream)
+{
+	struct dc_stream_status *main_stream_status;
+	struct dc_stream_status *phantom_stream_status;
+	enum dc_status res = dc_state_add_stream(dc, state, phantom_stream);
+
+	/* check if stream is tracked */
+	if (res == DC_OK && !dc_state_is_phantom_stream_tracked(state, phantom_stream)) {
+		/* stream must be tracked if added to state */
+		dc_state_track_phantom_stream(state, phantom_stream);
+	}
+
+	/* setup subvp meta */
+	main_stream_status = dc_state_get_stream_status(state, main_stream);
+	if (main_stream_status) {
+		main_stream_status->mall_stream_config.type = SUBVP_MAIN;
+		main_stream_status->mall_stream_config.paired_stream = phantom_stream;
+	}
+
+	phantom_stream_status = dc_state_get_stream_status(state, phantom_stream);
+	if (phantom_stream_status) {
+		phantom_stream_status->mall_stream_config.type = SUBVP_PHANTOM;
+		phantom_stream_status->mall_stream_config.paired_stream = main_stream;
+		phantom_stream_status->mall_stream_config.subvp_limit_cursor_size = false;
+		phantom_stream_status->mall_stream_config.cursor_size_limit_subvp = false;
+	}
+
+	dc_state_set_stream_subvp_cursor_limit(main_stream, state, true);
+
+	return res;
+}
+
+enum dc_status dc_state_remove_phantom_stream(const struct dc *dc,
+		struct dc_state *state,
+		struct dc_stream_state *phantom_stream)
+{
+	struct dc_stream_status *main_stream_status = NULL;
+	struct dc_stream_status *phantom_stream_status;
+
+	/* reset subvp meta */
+	phantom_stream_status = dc_state_get_stream_status(state, phantom_stream);
+	if (phantom_stream_status) {
+		main_stream_status = dc_state_get_stream_status(state, phantom_stream_status->mall_stream_config.paired_stream);
+		phantom_stream_status->mall_stream_config.type = SUBVP_NONE;
+		phantom_stream_status->mall_stream_config.paired_stream = NULL;
+	}
+
+	if (main_stream_status) {
+		main_stream_status->mall_stream_config.type = SUBVP_NONE;
+		main_stream_status->mall_stream_config.paired_stream = NULL;
+	}
+
+	/* remove stream from state */
+	return dc_state_remove_stream(dc, state, phantom_stream);
+}
+
+bool dc_state_add_phantom_plane(
+		const struct dc *dc,
+		struct dc_stream_state *phantom_stream,
+		struct dc_plane_state *phantom_plane,
+		struct dc_state *state)
+{
+	bool res = dc_state_add_plane(dc, phantom_stream, phantom_plane, state);
+
+	/* check if stream is tracked */
+	if (res && !dc_state_is_phantom_plane_tracked(state, phantom_plane)) {
+		/* stream must be tracked if added to state */
+		dc_state_track_phantom_plane(state, phantom_plane);
+	}
+
+	return res;
+}
+
+bool dc_state_remove_phantom_plane(
+		const struct dc *dc,
+		struct dc_stream_state *phantom_stream,
+		struct dc_plane_state *phantom_plane,
+		struct dc_state *state)
+{
+	return dc_state_remove_plane(dc, phantom_stream, phantom_plane, state);
+}
+
+bool dc_state_rem_all_phantom_planes_for_stream(
+		const struct dc *dc,
+		struct dc_stream_state *phantom_stream,
+		struct dc_state *state,
+		bool should_release_planes)
+{
+	int i, old_plane_count;
+	struct dc_stream_status *stream_status = NULL;
+	struct dc_plane_state *del_planes[MAX_SURFACES] = { 0 };
+
+	for (i = 0; i < state->stream_count; i++)
+		if (state->streams[i] == phantom_stream) {
+			stream_status = &state->stream_status[i];
+			break;
+		}
+
+	if (stream_status == NULL) {
+		dm_error("Existing stream %p not found!\n", phantom_stream);
+		return false;
+	}
+
+	old_plane_count = stream_status->plane_count;
+
+	for (i = 0; i < old_plane_count; i++)
+		del_planes[i] = stream_status->plane_states[i];
+
+	for (i = 0; i < old_plane_count; i++) {
+		if (!dc_state_remove_plane(dc, phantom_stream, del_planes[i], state))
+			return false;
+		if (should_release_planes)
+			dc_state_release_phantom_plane(dc, state, del_planes[i]);
+	}
+
+	return true;
+}
+
+bool dc_state_add_all_phantom_planes_for_stream(
+		const struct dc *dc,
+		struct dc_stream_state *phantom_stream,
+		struct dc_plane_state * const *phantom_planes,
+		int plane_count,
+		struct dc_state *state)
+{
+	return dc_state_add_all_planes_for_stream(dc, phantom_stream, phantom_planes, plane_count, state);
+}
+
+bool dc_state_remove_phantom_streams_and_planes(
+	const struct dc *dc,
+	struct dc_state *state)
+{
+	int i;
+	bool removed_phantom = false;
+	struct dc_stream_state *phantom_stream = NULL;
+
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		struct pipe_ctx *pipe = &state->res_ctx.pipe_ctx[i];
+
+		if (pipe->plane_state && pipe->stream && dc_state_get_pipe_subvp_type(state, pipe) == SUBVP_PHANTOM) {
+			phantom_stream = pipe->stream;
+
+			dc_state_rem_all_phantom_planes_for_stream(dc, phantom_stream, state, false);
+			dc_state_remove_phantom_stream(dc, state, phantom_stream);
+			removed_phantom = true;
+		}
+	}
+	return removed_phantom;
+}
+
+void dc_state_release_phantom_streams_and_planes(
+		const struct dc *dc,
+		struct dc_state *state)
+{
+	unsigned int phantom_count;
+	struct dc_stream_state *phantom_streams[MAX_PHANTOM_PIPES];
+	struct dc_plane_state *phantom_planes[MAX_PHANTOM_PIPES];
+	int i;
+
+	phantom_count = state->phantom_stream_count;
+	memcpy(phantom_streams, state->phantom_streams, sizeof(struct dc_stream_state *) * MAX_PHANTOM_PIPES);
+	for (i = 0; i < phantom_count; i++)
+		dc_state_release_phantom_stream(dc, state, phantom_streams[i]);
+
+	phantom_count = state->phantom_plane_count;
+	memcpy(phantom_planes, state->phantom_planes, sizeof(struct dc_plane_state *) * MAX_PHANTOM_PIPES);
+	for (i = 0; i < phantom_count; i++)
+		dc_state_release_phantom_plane(dc, state, phantom_planes[i]);
+}
+
+struct dc_stream_state *dc_state_get_stream_from_id(const struct dc_state *state, unsigned int id)
+{
+	struct dc_stream_state *stream = NULL;
+	int i;
+
+	for (i = 0; i < state->stream_count; i++) {
+		if (state->streams[i] && state->streams[i]->stream_id == id) {
+			stream = state->streams[i];
+			break;
+		}
+	}
+
+	return stream;
+}
+
+bool dc_state_is_fams2_in_use(
+		const struct dc *dc,
+		const struct dc_state *state)
+{
+	bool is_fams2_in_use = false;
+
+	if (state)
+		is_fams2_in_use |= state->bw_ctx.bw.dcn.fams2_global_config.features.bits.enable;
+
+	if (dc->current_state)
+		is_fams2_in_use |= dc->current_state->bw_ctx.bw.dcn.fams2_global_config.features.bits.enable;
+
+	return is_fams2_in_use;
+}
+
+void dc_state_set_stream_subvp_cursor_limit(const struct dc_stream_state *stream,
+		struct dc_state *state,
+		bool limit)
+{
+	struct dc_stream_status *stream_status;
+
+	stream_status = dc_state_get_stream_status(state, stream);
+
+	if (stream_status) {
+		stream_status->mall_stream_config.subvp_limit_cursor_size = limit;
+	}
+}
+
+bool dc_state_get_stream_subvp_cursor_limit(const struct dc_stream_state *stream,
+		struct dc_state *state)
+{
+	bool limit = false;
+
+	struct dc_stream_status *stream_status;
+
+	stream_status = dc_state_get_stream_status(state, stream);
+
+	if (stream_status) {
+		limit = stream_status->mall_stream_config.subvp_limit_cursor_size;
+	}
+
+	return limit;
+}
+
+void dc_state_set_stream_cursor_subvp_limit(const struct dc_stream_state *stream,
+		struct dc_state *state,
+		bool limit)
+{
+	struct dc_stream_status *stream_status;
+
+	stream_status = dc_state_get_stream_status(state, stream);
+
+	if (stream_status) {
+		stream_status->mall_stream_config.cursor_size_limit_subvp = limit;
+	}
+}
+
+bool dc_state_get_stream_cursor_subvp_limit(const struct dc_stream_state *stream,
+		struct dc_state *state)
+{
+	bool limit = false;
+
+	struct dc_stream_status *stream_status;
+
+	stream_status = dc_state_get_stream_status(state, stream);
+
+	if (stream_status) {
+		limit = stream_status->mall_stream_config.cursor_size_limit_subvp;
+	}
+
+	return limit;
+}
+
+bool dc_state_can_clear_stream_cursor_subvp_limit(const struct dc_stream_state *stream,
+		struct dc_state *state)
+{
+	bool can_clear_limit = false;
+
+	struct dc_stream_status *stream_status;
+
+	stream_status = dc_state_get_stream_status(state, stream);
+
+	if (stream_status) {
+		can_clear_limit = dc_state_get_stream_cursor_subvp_limit(stream, state) &&
+				(stream_status->mall_stream_config.type == SUBVP_PHANTOM ||
+				stream->hw_cursor_req ||
+				!stream_status->mall_stream_config.subvp_limit_cursor_size ||
+				!stream->cursor_position.enable ||
+				dc_stream_check_cursor_attributes(stream, state, &stream->cursor_attributes));
+	}
+
+	return can_clear_limit;
+}
+
+bool dc_state_is_subvp_in_use(struct dc_state *state)
+{
+	uint32_t i;
+
+	for (i = 0; i < state->stream_count; i++) {
+		if (dc_state_get_stream_subvp_type(state, state->streams[i]) != SUBVP_NONE)
+			return true;
+	}
+
+	return false;
+}
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_stream.c b/drivers/gpu/drm/amd/display/dc/core/dc_stream.c
index 57cf4cb82370..129cd5f84983 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_stream.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_stream.c
@@ -23,9 +23,6 @@
  *
  */
 
-#include <linux/delay.h>
-#include <linux/slab.h>
-
 #include "dm_services.h"
 #include "basics/dc_common.h"
 #include "dc.h"
@@ -33,8 +30,17 @@
 #include "resource.h"
 #include "ipp.h"
 #include "timing_generator.h"
+#include "dc_dmub_srv.h"
+#include "dc_state_priv.h"
+#include "dc_stream_priv.h"
 
 #define DC_LOGGER dc->ctx->logger
+#ifndef MIN
+#define MIN(X, Y) ((X) < (Y) ? (X) : (Y))
+#endif
+#ifndef MAX
+#define MAX(x, y) ((x > y) ? x : y)
+#endif
 
 /*******************************************************************************
  * Private functions
@@ -56,7 +62,7 @@ void update_stream_signal(struct dc_stream_state *stream, struct dc_sink *sink)
 	}
 }
 
-static bool dc_stream_construct(struct dc_stream_state *stream,
+bool dc_stream_construct(struct dc_stream_state *stream,
 	struct dc_sink *dc_sink_data)
 {
 	uint32_t i = 0;
@@ -73,8 +79,7 @@ static bool dc_stream_construct(struct dc_stream_state *stream,
 
 	/* Copy audio modes */
 	/* TODO - Remove this translation */
-	for (i = 0; i < (dc_sink_data->edid_caps.audio_mode_count); i++)
-	{
+	for (i = 0; i < (dc_sink_data->edid_caps.audio_mode_count); i++) {
 		stream->audio_info.modes[i].channel_count = dc_sink_data->edid_caps.audio_modes[i].channel_count;
 		stream->audio_info.modes[i].format_code = dc_sink_data->edid_caps.audio_modes[i].format_code;
 		stream->audio_info.modes[i].sample_rates.all = dc_sink_data->edid_caps.audio_modes[i].sample_rate;
@@ -117,26 +122,23 @@ static bool dc_stream_construct(struct dc_stream_state *stream,
 
 	update_stream_signal(stream, dc_sink_data);
 
-	stream->out_transfer_func = dc_create_transfer_func();
-	if (stream->out_transfer_func == NULL) {
-		dc_sink_release(dc_sink_data);
-		return false;
-	}
-	stream->out_transfer_func->type = TF_TYPE_BYPASS;
+	stream->out_transfer_func.type = TF_TYPE_BYPASS;
 
-	stream->stream_id = stream->ctx->dc_stream_id_count;
-	stream->ctx->dc_stream_id_count++;
+	dc_stream_assign_stream_id(stream);
 
 	return true;
 }
 
-static void dc_stream_destruct(struct dc_stream_state *stream)
+void dc_stream_destruct(struct dc_stream_state *stream)
 {
 	dc_sink_release(stream->sink);
-	if (stream->out_transfer_func != NULL) {
-		dc_transfer_func_release(stream->out_transfer_func);
-		stream->out_transfer_func = NULL;
-	}
+}
+
+void dc_stream_assign_stream_id(struct dc_stream_state *stream)
+{
+	/* MSB is reserved to indicate phantoms */
+	stream->stream_id = stream->ctx->dc_stream_id_count;
+	stream->ctx->dc_stream_id_count++;
 }
 
 void dc_stream_retain(struct dc_stream_state *stream)
@@ -196,14 +198,11 @@ struct dc_stream_state *dc_copy_stream(const struct dc_stream_state *stream)
 	if (new_stream->sink)
 		dc_sink_retain(new_stream->sink);
 
-	if (new_stream->out_transfer_func)
-		dc_transfer_func_retain(new_stream->out_transfer_func);
-
-	new_stream->stream_id = new_stream->ctx->dc_stream_id_count;
-	new_stream->ctx->dc_stream_id_count++;
+	dc_stream_assign_stream_id(new_stream);
 
 	/* If using dynamic encoder assignment, wait till stream committed to assign encoder. */
-	if (new_stream->ctx->dc->res_pool->funcs->link_encs_assign)
+	if (new_stream->ctx->dc->res_pool->funcs->link_encs_assign &&
+			!new_stream->ctx->dc->config.unify_link_enc_assignment)
 		new_stream->link_enc = NULL;
 
 	kref_init(&new_stream->refcount);
@@ -212,31 +211,6 @@ struct dc_stream_state *dc_copy_stream(const struct dc_stream_state *stream)
 }
 
 /**
- * dc_stream_get_status_from_state - Get stream status from given dc state
- * @state: DC state to find the stream status in
- * @stream: The stream to get the stream status for
- *
- * The given stream is expected to exist in the given dc state. Otherwise, NULL
- * will be returned.
- */
-struct dc_stream_status *dc_stream_get_status_from_state(
-	struct dc_state *state,
-	struct dc_stream_state *stream)
-{
-	uint8_t i;
-
-	if (state == NULL)
-		return NULL;
-
-	for (i = 0; i < state->stream_count; i++) {
-		if (stream == state->streams[i])
-			return &state->stream_status[i];
-	}
-
-	return NULL;
-}
-
-/**
  * dc_stream_get_status() - Get current stream status of the given stream state
  * @stream: The stream to get the stream status for.
  *
@@ -247,17 +221,25 @@ struct dc_stream_status *dc_stream_get_status(
 	struct dc_stream_state *stream)
 {
 	struct dc *dc = stream->ctx->dc;
-	return dc_stream_get_status_from_state(dc->current_state, stream);
+	return dc_state_get_stream_status(dc->current_state, stream);
 }
 
-static void program_cursor_attributes(
+const struct dc_stream_status *dc_stream_get_status_const(
+	const struct dc_stream_state *stream)
+{
+	struct dc *dc = stream->ctx->dc;
+
+	return dc_state_get_stream_status(dc->current_state, stream);
+}
+
+void program_cursor_attributes(
 	struct dc *dc,
-	struct dc_stream_state *stream,
-	const struct dc_cursor_attributes *attributes)
+	struct dc_stream_state *stream)
 {
 	int i;
 	struct resource_context *res_ctx;
 	struct pipe_ctx *pipe_to_program = NULL;
+	bool enable_cursor_offload = dc_dmub_srv_is_cursor_offload_enabled(dc);
 
 	if (!stream)
 		return;
@@ -272,46 +254,47 @@ static void program_cursor_attributes(
 
 		if (!pipe_to_program) {
 			pipe_to_program = pipe_ctx;
-			dc->hwss.cursor_lock(dc, pipe_to_program, true);
+
+			if (enable_cursor_offload && dc->hwss.begin_cursor_offload_update) {
+				dc->hwss.begin_cursor_offload_update(dc, pipe_ctx);
+			} else {
+				dc->hwss.cursor_lock(dc, pipe_to_program, true);
+				if (pipe_to_program->next_odm_pipe)
+					dc->hwss.cursor_lock(dc, pipe_to_program->next_odm_pipe, true);
+			}
 		}
 
 		dc->hwss.set_cursor_attribute(pipe_ctx);
+		if (dc->ctx->dmub_srv)
+			dc_send_update_cursor_info_to_dmu(pipe_ctx, i);
 		if (dc->hwss.set_cursor_sdr_white_level)
 			dc->hwss.set_cursor_sdr_white_level(pipe_ctx);
+		if (enable_cursor_offload && dc->hwss.update_cursor_offload_pipe)
+			dc->hwss.update_cursor_offload_pipe(dc, pipe_ctx);
 	}
 
-	if (pipe_to_program)
-		dc->hwss.cursor_lock(dc, pipe_to_program, false);
-}
-
-#ifndef TRIM_FSFT
-/*
- * dc_optimize_timing_for_fsft() - dc to optimize timing
- */
-bool dc_optimize_timing_for_fsft(
-	struct dc_stream_state *pStream,
-	unsigned int max_input_rate_in_khz)
-{
-	struct dc  *dc;
-
-	dc = pStream->ctx->dc;
-
-	return (dc->hwss.optimize_timing_for_fsft &&
-		dc->hwss.optimize_timing_for_fsft(dc, &pStream->timing, max_input_rate_in_khz));
+	if (pipe_to_program) {
+		if (enable_cursor_offload && dc->hwss.commit_cursor_offload_update) {
+			dc->hwss.commit_cursor_offload_update(dc, pipe_to_program);
+		} else {
+			dc->hwss.cursor_lock(dc, pipe_to_program, false);
+			if (pipe_to_program->next_odm_pipe)
+				dc->hwss.cursor_lock(dc, pipe_to_program->next_odm_pipe, false);
+		}
+	}
 }
-#endif
 
 /*
- * dc_stream_set_cursor_attributes() - Update cursor attributes and set cursor surface address
+ * dc_stream_check_cursor_attributes() - Check validitity of cursor attributes and surface address
  */
-bool dc_stream_set_cursor_attributes(
-	struct dc_stream_state *stream,
+bool dc_stream_check_cursor_attributes(
+	const struct dc_stream_state *stream,
+	struct dc_state *state,
 	const struct dc_cursor_attributes *attributes)
 {
-	struct dc  *dc;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	bool reset_idle_optimizations = false;
-#endif
+	const struct dc *dc;
+
+	unsigned int max_cursor_size;
 
 	if (NULL == stream) {
 		dm_error("DC: dc_stream is NULL!\n");
@@ -328,36 +311,82 @@ bool dc_stream_set_cursor_attributes(
 	}
 
 	dc = stream->ctx->dc;
-	stream->cursor_attributes = *attributes;
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	dc_z10_restore(dc);
-	/* disable idle optimizations while updating cursor */
-	if (dc->idle_optimizations_allowed) {
-		dc_allow_idle_optimizations(dc, false);
-		reset_idle_optimizations = true;
+	/* SubVP is not compatible with HW cursor larger than what can fit in cursor SRAM.
+	 * Therefore, if cursor is greater than this, fallback to SW cursor.
+	 */
+	if (dc->debug.allow_sw_cursor_fallback && dc->res_pool->funcs->get_max_hw_cursor_size) {
+		max_cursor_size = dc->res_pool->funcs->get_max_hw_cursor_size(dc, state, stream);
+		max_cursor_size = max_cursor_size * max_cursor_size * 4;
+
+		if (attributes->height * attributes->width * 4 > max_cursor_size) {
+			return false;
+		}
 	}
 
-#endif
-	program_cursor_attributes(dc, stream, attributes);
+	return true;
+}
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	/* re-enable idle optimizations if necessary */
-	if (reset_idle_optimizations)
-		dc_allow_idle_optimizations(dc, true);
+/*
+ * dc_stream_set_cursor_attributes() - Update cursor attributes and set cursor surface address
+ */
+bool dc_stream_set_cursor_attributes(
+	struct dc_stream_state *stream,
+	const struct dc_cursor_attributes *attributes)
+{
+	bool result = false;
 
-#endif
-	return true;
+	if (!stream)
+		return false;
+
+	if (dc_stream_check_cursor_attributes(stream, stream->ctx->dc->current_state, attributes)) {
+		stream->cursor_attributes = *attributes;
+		result = true;
+	}
+
+	return result;
 }
 
-static void program_cursor_position(
-	struct dc *dc,
+bool dc_stream_program_cursor_attributes(
 	struct dc_stream_state *stream,
-	const struct dc_cursor_position *position)
+	const struct dc_cursor_attributes *attributes)
+{
+	struct dc  *dc;
+	bool reset_idle_optimizations = false;
+
+	if (!stream)
+		return false;
+
+	dc = stream->ctx->dc;
+
+	if (dc_stream_set_cursor_attributes(stream, attributes)) {
+		dc_z10_restore(dc);
+		/* disable idle optimizations while updating cursor */
+		if (dc->idle_optimizations_allowed) {
+			dc_allow_idle_optimizations(dc, false);
+			reset_idle_optimizations = true;
+		}
+
+		program_cursor_attributes(dc, stream);
+
+		/* re-enable idle optimizations if necessary */
+		if (reset_idle_optimizations && !dc->debug.disable_dmub_reallow_idle)
+			dc_allow_idle_optimizations(dc, true);
+
+		return true;
+	}
+
+	return false;
+}
+
+void program_cursor_position(
+	struct dc *dc,
+	struct dc_stream_state *stream)
 {
 	int i;
 	struct resource_context *res_ctx;
 	struct pipe_ctx *pipe_to_program = NULL;
+	bool enable_cursor_offload = dc_dmub_srv_is_cursor_offload_enabled(dc);
 
 	if (!stream)
 		return;
@@ -376,25 +405,33 @@ static void program_cursor_position(
 
 		if (!pipe_to_program) {
 			pipe_to_program = pipe_ctx;
-			dc->hwss.cursor_lock(dc, pipe_to_program, true);
+
+			if (enable_cursor_offload && dc->hwss.begin_cursor_offload_update)
+				dc->hwss.begin_cursor_offload_update(dc, pipe_ctx);
+			else
+				dc->hwss.cursor_lock(dc, pipe_to_program, true);
 		}
 
 		dc->hwss.set_cursor_position(pipe_ctx);
+		if (enable_cursor_offload && dc->hwss.update_cursor_offload_pipe)
+			dc->hwss.update_cursor_offload_pipe(dc, pipe_ctx);
+
+		if (dc->ctx->dmub_srv)
+			dc_send_update_cursor_info_to_dmu(pipe_ctx, i);
 	}
 
-	if (pipe_to_program)
-		dc->hwss.cursor_lock(dc, pipe_to_program, false);
+	if (pipe_to_program) {
+		if (enable_cursor_offload && dc->hwss.commit_cursor_offload_update)
+			dc->hwss.commit_cursor_offload_update(dc, pipe_to_program);
+		else
+			dc->hwss.cursor_lock(dc, pipe_to_program, false);
+	}
 }
 
 bool dc_stream_set_cursor_position(
 	struct dc_stream_state *stream,
 	const struct dc_cursor_position *position)
 {
-	struct dc  *dc;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	bool reset_idle_optimizations = false;
-#endif
-
 	if (NULL == stream) {
 		dm_error("DC: dc_stream is NULL!\n");
 		return false;
@@ -405,27 +442,66 @@ bool dc_stream_set_cursor_position(
 		return false;
 	}
 
+	stream->cursor_position = *position;
+
+
+	return true;
+}
+
+bool dc_stream_program_cursor_position(
+	struct dc_stream_state *stream,
+	const struct dc_cursor_position *position)
+{
+	struct dc *dc;
+	bool reset_idle_optimizations = false;
+	const struct dc_cursor_position *old_position;
+
+	if (!stream)
+		return false;
+
+	old_position = &stream->cursor_position;
 	dc = stream->ctx->dc;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	dc_z10_restore(dc);
 
-	/* disable idle optimizations if enabling cursor */
-	if (dc->idle_optimizations_allowed && !stream->cursor_position.enable && position->enable) {
-		dc_allow_idle_optimizations(dc, false);
-		reset_idle_optimizations = true;
-	}
+	if (dc_stream_set_cursor_position(stream, position)) {
+		dc_z10_restore(dc);
 
-#endif
-	stream->cursor_position = *position;
+		/* disable idle optimizations if enabling cursor */
+		if (dc->idle_optimizations_allowed &&
+		    (!old_position->enable || dc->debug.exit_idle_opt_for_cursor_updates) &&
+		    position->enable) {
+			dc_allow_idle_optimizations(dc, false);
+			reset_idle_optimizations = true;
+		}
 
-	program_cursor_position(dc, stream, position);
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	/* re-enable idle optimizations if necessary */
-	if (reset_idle_optimizations)
-		dc_allow_idle_optimizations(dc, true);
+		program_cursor_position(dc, stream);
+		/* re-enable idle optimizations if necessary */
+		if (reset_idle_optimizations && !dc->debug.disable_dmub_reallow_idle)
+			dc_allow_idle_optimizations(dc, true);
 
-#endif
-	return true;
+		/* apply/update visual confirm */
+		if (dc->debug.visual_confirm == VISUAL_CONFIRM_HW_CURSOR) {
+			/* update software state */
+			int i;
+
+			for (i = 0; i < dc->res_pool->pipe_count; i++) {
+				struct pipe_ctx *pipe_ctx = &dc->current_state->res_ctx.pipe_ctx[i];
+
+				/* adjust visual confirm color for all pipes with current stream */
+				if (stream == pipe_ctx->stream) {
+					get_cursor_visual_confirm_color(pipe_ctx, &(pipe_ctx->visual_confirm_color));
+
+					/* programming hardware */
+					if (pipe_ctx->plane_state)
+						dc->hwss.update_visual_confirm_color(dc, pipe_ctx,
+								pipe_ctx->plane_res.hubp->mpcc_id);
+				}
+			}
+		}
+
+		return true;
+	}
+
+	return false;
 }
 
 bool dc_stream_add_writeback(struct dc *dc,
@@ -451,7 +527,9 @@ bool dc_stream_add_writeback(struct dc *dc,
 		return false;
 	}
 
-	wb_info->dwb_params.out_transfer_func = stream->out_transfer_func;
+	dc_exit_ips_for_hw_access(dc);
+
+	wb_info->dwb_params.out_transfer_func = &stream->out_transfer_func;
 
 	dwb = dc->res_pool->dwbc[wb_info->dwb_pipe_inst];
 	dwb->dwb_is_drc = false;
@@ -468,41 +546,80 @@ bool dc_stream_add_writeback(struct dc *dc,
 	}
 
 	if (!isDrc) {
+		ASSERT(stream->num_wb_info + 1 <= MAX_DWB_PIPES);
 		stream->writeback_info[stream->num_wb_info++] = *wb_info;
 	}
 
 	if (dc->hwss.enable_writeback) {
 		struct dc_stream_status *stream_status = dc_stream_get_status(stream);
 		struct dwbc *dwb = dc->res_pool->dwbc[wb_info->dwb_pipe_inst];
-		dwb->otg_inst = stream_status->primary_otg_inst;
+		if (stream_status)
+			dwb->otg_inst = stream_status->primary_otg_inst;
 	}
-	if (IS_DIAG_DC(dc->ctx->dce_environment)) {
-		if (!dc->hwss.update_bandwidth(dc, dc->current_state)) {
-			dm_error("DC: update_bandwidth failed!\n");
-			return false;
-		}
 
-		/* enable writeback */
-		if (dc->hwss.enable_writeback) {
-			struct dwbc *dwb = dc->res_pool->dwbc[wb_info->dwb_pipe_inst];
+	if (!dc->hwss.update_bandwidth(dc, dc->current_state)) {
+		dm_error("DC: update_bandwidth failed!\n");
+		return false;
+	}
 
-			if (dwb->funcs->is_enabled(dwb)) {
-				/* writeback pipe already enabled, only need to update */
-				dc->hwss.update_writeback(dc, wb_info, dc->current_state);
-			} else {
-				/* Enable writeback pipe from scratch*/
-				dc->hwss.enable_writeback(dc, wb_info, dc->current_state);
-			}
+	/* enable writeback */
+	if (dc->hwss.enable_writeback) {
+		struct dwbc *dwb = dc->res_pool->dwbc[wb_info->dwb_pipe_inst];
+
+		if (dwb->funcs->is_enabled(dwb)) {
+			/* writeback pipe already enabled, only need to update */
+			dc->hwss.update_writeback(dc, wb_info, dc->current_state);
+		} else {
+			/* Enable writeback pipe from scratch*/
+			dc->hwss.enable_writeback(dc, wb_info, dc->current_state);
 		}
 	}
+
+	return true;
+}
+
+bool dc_stream_fc_disable_writeback(struct dc *dc,
+		struct dc_stream_state *stream,
+		uint32_t dwb_pipe_inst)
+{
+	struct dwbc *dwb = dc->res_pool->dwbc[dwb_pipe_inst];
+
+	if (stream == NULL) {
+		dm_error("DC: dc_stream is NULL!\n");
+		return false;
+	}
+
+	if (dwb_pipe_inst >= MAX_DWB_PIPES) {
+		dm_error("DC: writeback pipe is invalid!\n");
+		return false;
+	}
+
+	if (stream->num_wb_info > MAX_DWB_PIPES) {
+		dm_error("DC: num_wb_info is invalid!\n");
+		return false;
+	}
+
+	dc_exit_ips_for_hw_access(dc);
+
+	if (dwb->funcs->set_fc_enable)
+		dwb->funcs->set_fc_enable(dwb, DWB_FRAME_CAPTURE_DISABLE);
+
 	return true;
 }
 
+/**
+ * dc_stream_remove_writeback() - Disables writeback and removes writeback info.
+ * @dc: Display core control structure.
+ * @stream: Display core stream state.
+ * @dwb_pipe_inst: Display writeback pipe.
+ *
+ * Return: returns true on success, false otherwise.
+ */
 bool dc_stream_remove_writeback(struct dc *dc,
 		struct dc_stream_state *stream,
 		uint32_t dwb_pipe_inst)
 {
-	int i = 0, j = 0;
+	unsigned int i, j;
 	if (stream == NULL) {
 		dm_error("DC: dc_stream is NULL!\n");
 		return false;
@@ -513,49 +630,47 @@ bool dc_stream_remove_writeback(struct dc *dc,
 		return false;
 	}
 
-//	stream->writeback_info[dwb_pipe_inst].wb_enabled = false;
-	for (i = 0; i < stream->num_wb_info; i++) {
-		/*dynamic update*/
-		if (stream->writeback_info[i].wb_enabled &&
-			stream->writeback_info[i].dwb_pipe_inst == dwb_pipe_inst) {
-			stream->writeback_info[i].wb_enabled = false;
-		}
+	if (stream->num_wb_info > MAX_DWB_PIPES) {
+		dm_error("DC: num_wb_info is invalid!\n");
+		return false;
 	}
 
 	/* remove writeback info for disabled writeback pipes from stream */
 	for (i = 0, j = 0; i < stream->num_wb_info; i++) {
 		if (stream->writeback_info[i].wb_enabled) {
-			if (i != j)
-				/* trim the array */
-				stream->writeback_info[j] = stream->writeback_info[i];
-			j++;
+
+			if (stream->writeback_info[i].dwb_pipe_inst == dwb_pipe_inst)
+				stream->writeback_info[i].wb_enabled = false;
+
+			/* trim the array */
+			if (j < i) {
+				memcpy(&stream->writeback_info[j], &stream->writeback_info[i],
+						sizeof(struct dc_writeback_info));
+				j++;
+			}
 		}
 	}
 	stream->num_wb_info = j;
 
-	if (IS_DIAG_DC(dc->ctx->dce_environment)) {
-		/* recalculate and apply DML parameters */
-		if (!dc->hwss.update_bandwidth(dc, dc->current_state)) {
-			dm_error("DC: update_bandwidth failed!\n");
-			return false;
-		}
+	/* recalculate and apply DML parameters */
+	if (!dc->hwss.update_bandwidth(dc, dc->current_state)) {
+		dm_error("DC: update_bandwidth failed!\n");
+		return false;
+	}
+
+	dc_exit_ips_for_hw_access(dc);
 
-		/* disable writeback */
-		if (dc->hwss.disable_writeback)
+	/* disable writeback */
+	if (dc->hwss.disable_writeback) {
+		struct dwbc *dwb = dc->res_pool->dwbc[dwb_pipe_inst];
+
+		if (dwb->funcs->is_enabled(dwb))
 			dc->hwss.disable_writeback(dc, dwb_pipe_inst);
 	}
+
 	return true;
 }
 
-bool dc_stream_warmup_writeback(struct dc *dc,
-		int num_dwb,
-		struct dc_writeback_info *wb_info)
-{
-	if (dc->hwss.mmhubbub_warmup)
-		return dc->hwss.mmhubbub_warmup(dc, num_dwb, wb_info);
-	else
-		return false;
-}
 uint32_t dc_stream_get_vblank_counter(const struct dc_stream_state *stream)
 {
 	uint8_t i;
@@ -563,10 +678,12 @@ uint32_t dc_stream_get_vblank_counter(const struct dc_stream_state *stream)
 	struct resource_context *res_ctx =
 		&dc->current_state->res_ctx;
 
+	dc_exit_ips_for_hw_access(dc);
+
 	for (i = 0; i < MAX_PIPES; i++) {
 		struct timing_generator *tg = res_ctx->pipe_ctx[i].stream_res.tg;
 
-		if (res_ctx->pipe_ctx[i].stream != stream)
+		if (res_ctx->pipe_ctx[i].stream != stream || !tg)
 			continue;
 
 		return tg->funcs->get_frame_count(tg);
@@ -591,6 +708,8 @@ bool dc_stream_send_dp_sdp(const struct dc_stream_state *stream,
 	dc = stream->ctx->dc;
 	res_ctx = &dc->current_state->res_ctx;
 
+	dc_exit_ips_for_hw_access(dc);
+
 	for (i = 0; i < MAX_PIPES; i++) {
 		struct pipe_ctx *pipe_ctx = &res_ctx->pipe_ctx[i];
 
@@ -618,14 +737,21 @@ bool dc_stream_get_scanoutpos(const struct dc_stream_state *stream,
 {
 	uint8_t i;
 	bool ret = false;
-	struct dc  *dc = stream->ctx->dc;
-	struct resource_context *res_ctx =
-		&dc->current_state->res_ctx;
+	struct dc  *dc;
+	struct resource_context *res_ctx;
+
+	if (!stream->ctx)
+		return false;
+
+	dc = stream->ctx->dc;
+	res_ctx = &dc->current_state->res_ctx;
+
+	dc_exit_ips_for_hw_access(dc);
 
 	for (i = 0; i < MAX_PIPES; i++) {
 		struct timing_generator *tg = res_ctx->pipe_ctx[i].stream_res.tg;
 
-		if (res_ctx->pipe_ctx[i].stream != stream)
+		if (res_ctx->pipe_ctx[i].stream != stream || !tg)
 			continue;
 
 		tg->funcs->get_scanoutpos(tg,
@@ -658,6 +784,8 @@ bool dc_stream_dmdata_status_done(struct dc *dc, struct dc_stream_state *stream)
 	if (i == MAX_PIPES)
 		return true;
 
+	dc_exit_ips_for_hw_access(dc);
+
 	return dc->hwss.dmdata_status_done(pipe);
 }
 
@@ -692,6 +820,8 @@ bool dc_stream_set_dynamic_metadata(struct dc *dc,
 
 	pipe_ctx->stream->dmdata_address = attr->address;
 
+	dc_exit_ips_for_hw_access(dc);
+
 	dc->hwss.program_dmdata_engine(pipe_ctx);
 
 	if (hubp->funcs->dmdata_set_attributes != NULL &&
@@ -713,6 +843,20 @@ enum dc_status dc_stream_add_dsc_to_resource(struct dc *dc,
 	}
 }
 
+struct pipe_ctx *dc_stream_get_pipe_ctx(struct dc_stream_state *stream)
+{
+	int i = 0;
+
+	for (i = 0; i < MAX_PIPES; i++) {
+		struct pipe_ctx *pipe = &stream->ctx->dc->current_state->res_ctx.pipe_ctx[i];
+
+		if (pipe->stream == stream)
+			return pipe;
+	}
+
+	return NULL;
+}
+
 void dc_stream_log(const struct dc *dc, const struct dc_stream_state *stream)
 {
 	DC_LOG_DC(
@@ -728,14 +872,400 @@ void dc_stream_log(const struct dc *dc, const struct dc_stream_state *stream)
 			stream->dst.height,
 			stream->output_color_space);
 	DC_LOG_DC(
-			"\tpix_clk_khz: %d, h_total: %d, v_total: %d, pixelencoder:%d, displaycolorDepth:%d\n",
+			"\tpix_clk_khz: %d, h_total: %d, v_total: %d, pixel_encoding:%s, color_depth:%s\n",
 			stream->timing.pix_clk_100hz / 10,
 			stream->timing.h_total,
 			stream->timing.v_total,
-			stream->timing.pixel_encoding,
-			stream->timing.display_color_depth);
+			dc_pixel_encoding_to_str(stream->timing.pixel_encoding),
+			dc_color_depth_to_str(stream->timing.display_color_depth));
 	DC_LOG_DC(
 			"\tlink: %d\n",
 			stream->link->link_index);
+
+	DC_LOG_DC(
+			"\tdsc: %d, mst_pbn: %d\n",
+			stream->timing.flags.DSC,
+			stream->timing.dsc_cfg.mst_pbn);
+
+	if (stream->sink) {
+		if (stream->sink->sink_signal != SIGNAL_TYPE_VIRTUAL &&
+			stream->sink->sink_signal != SIGNAL_TYPE_NONE) {
+
+			DC_LOG_DC(
+					"\tsignal: %x dispname: %s manufacturer_id: 0x%x product_id: 0x%x\n",
+					stream->signal,
+					stream->sink->edid_caps.display_name,
+					stream->sink->edid_caps.manufacturer_id,
+					stream->sink->edid_caps.product_id);
+		}
+	}
+}
+
+/*
+*	dc_stream_get_3dlut()
+*	Requirements:
+*	1. Is stream already owns an RMCM instance, return it.
+*	2. If it doesn't and we don't need to allocate, return NULL.
+*	3. If there's a free RMCM instance, assign to stream and return it.
+*	4. If no free RMCM instances, return NULL.
+*/
+
+struct dc_rmcm_3dlut *dc_stream_get_3dlut_for_stream(
+	const struct dc *dc,
+	const struct dc_stream_state *stream,
+	bool allocate_one)
+{
+	unsigned int num_rmcm = dc->caps.color.mpc.num_rmcm_3dluts;
+
+	// see if one is allocated for this stream
+	for (int i = 0; i < num_rmcm; i++) {
+		if (dc->res_pool->rmcm_3dlut[i].isInUse &&
+			dc->res_pool->rmcm_3dlut[i].stream == stream)
+			return &dc->res_pool->rmcm_3dlut[i];
+	}
+
+	//case: not found one, and dont need to allocate
+	if (!allocate_one)
+		return NULL;
+
+	//see if there is an unused 3dlut, allocate
+	for (int i = 0; i < num_rmcm; i++) {
+		if (!dc->res_pool->rmcm_3dlut[i].isInUse) {
+			dc->res_pool->rmcm_3dlut[i].isInUse = true;
+			dc->res_pool->rmcm_3dlut[i].stream = stream;
+			return &dc->res_pool->rmcm_3dlut[i];
+		}
+	}
+
+	//dont have a 3dlut
+	return NULL;
+}
+
+
+void dc_stream_release_3dlut_for_stream(
+	const struct dc *dc,
+	const struct dc_stream_state *stream)
+{
+	struct dc_rmcm_3dlut *rmcm_3dlut =
+		dc_stream_get_3dlut_for_stream(dc, stream, false);
+
+	if (rmcm_3dlut) {
+		rmcm_3dlut->isInUse = false;
+		rmcm_3dlut->stream  = NULL;
+		rmcm_3dlut->protection_bits = 0;
+	}
+}
+
+
+void dc_stream_init_rmcm_3dlut(struct dc *dc)
+{
+	unsigned int num_rmcm = dc->caps.color.mpc.num_rmcm_3dluts;
+
+	for (int i = 0; i < num_rmcm; i++) {
+		dc->res_pool->rmcm_3dlut[i].isInUse = false;
+		dc->res_pool->rmcm_3dlut[i].stream = NULL;
+		dc->res_pool->rmcm_3dlut[i].protection_bits = 0;
+	}
+}
+
+/*
+ * Finds the greatest index in refresh_rate_hz that contains a value <= refresh
+ */
+static int dc_stream_get_nearest_smallest_index(struct dc_stream_state *stream, int refresh)
+{
+	for (int i = 0; i < (LUMINANCE_DATA_TABLE_SIZE - 1); ++i) {
+		if ((stream->lumin_data.refresh_rate_hz[i] <= refresh) && (refresh < stream->lumin_data.refresh_rate_hz[i + 1])) {
+			return i;
+		}
+	}
+	return 9;
+}
+
+/*
+ * Finds a corresponding brightness for a given refresh rate between 2 given indices, where index1 < index2
+ */
+static int dc_stream_get_brightness_millinits_linear_interpolation (struct dc_stream_state *stream,
+								     int index1,
+								     int index2,
+								     int refresh_hz)
+{
+	long long slope = 0;
+	if (stream->lumin_data.refresh_rate_hz[index2] != stream->lumin_data.refresh_rate_hz[index1]) {
+		slope = (stream->lumin_data.luminance_millinits[index2] - stream->lumin_data.luminance_millinits[index1]) /
+			    (stream->lumin_data.refresh_rate_hz[index2] - stream->lumin_data.refresh_rate_hz[index1]);
+	}
+
+	int y_intercept = stream->lumin_data.luminance_millinits[index2] - slope * stream->lumin_data.refresh_rate_hz[index2];
+
+	return (y_intercept + refresh_hz * slope);
+}
+
+/*
+ * Finds a corresponding refresh rate for a given brightness between 2 given indices, where index1 < index2
+ */
+static int dc_stream_get_refresh_hz_linear_interpolation (struct dc_stream_state *stream,
+							   int index1,
+							   int index2,
+							   int brightness_millinits)
+{
+	long long slope = 1;
+	if (stream->lumin_data.refresh_rate_hz[index2] != stream->lumin_data.refresh_rate_hz[index1]) {
+		slope = (stream->lumin_data.luminance_millinits[index2] - stream->lumin_data.luminance_millinits[index1]) /
+				(stream->lumin_data.refresh_rate_hz[index2] - stream->lumin_data.refresh_rate_hz[index1]);
+	}
+
+	int y_intercept = stream->lumin_data.luminance_millinits[index2] - slope * stream->lumin_data.refresh_rate_hz[index2];
+
+	return ((int)div64_s64((brightness_millinits - y_intercept), slope));
+}
+
+/*
+ * Finds the current brightness in millinits given a refresh rate
+ */
+static int dc_stream_get_brightness_millinits_from_refresh (struct dc_stream_state *stream, int refresh_hz)
+{
+	int nearest_smallest_index = dc_stream_get_nearest_smallest_index(stream, refresh_hz);
+	int nearest_smallest_value = stream->lumin_data.refresh_rate_hz[nearest_smallest_index];
+
+	if (nearest_smallest_value == refresh_hz)
+		return stream->lumin_data.luminance_millinits[nearest_smallest_index];
+
+	if (nearest_smallest_index >= 9)
+		return dc_stream_get_brightness_millinits_linear_interpolation(stream, nearest_smallest_index - 1, nearest_smallest_index, refresh_hz);
+
+	if (nearest_smallest_value == stream->lumin_data.refresh_rate_hz[nearest_smallest_index + 1])
+		return stream->lumin_data.luminance_millinits[nearest_smallest_index];
+
+	return dc_stream_get_brightness_millinits_linear_interpolation(stream, nearest_smallest_index, nearest_smallest_index + 1, refresh_hz);
 }
 
+/*
+ * Finds the lowest/highest refresh rate (depending on search_for_max_increase)
+ * that can be achieved from starting_refresh_hz while staying
+ * within flicker criteria
+ */
+static int dc_stream_calculate_flickerless_refresh_rate(struct dc_stream_state *stream,
+							 int current_brightness,
+							 int starting_refresh_hz,
+							 bool is_gaming,
+							 bool search_for_max_increase)
+{
+	int nearest_smallest_index = dc_stream_get_nearest_smallest_index(stream, starting_refresh_hz);
+
+	int flicker_criteria_millinits = is_gaming ?
+					 stream->lumin_data.flicker_criteria_milli_nits_GAMING :
+					 stream->lumin_data.flicker_criteria_milli_nits_STATIC;
+
+	int safe_upper_bound = current_brightness + flicker_criteria_millinits;
+	int safe_lower_bound = current_brightness - flicker_criteria_millinits;
+	int lumin_millinits_temp = 0;
+
+	int offset = -1;
+	if (search_for_max_increase) {
+		offset = 1;
+	}
+
+	/*
+	 * Increments up or down by 1 depending on search_for_max_increase
+	 */
+	for (int i = nearest_smallest_index; (i > 0 && !search_for_max_increase) || (i < (LUMINANCE_DATA_TABLE_SIZE - 1) && search_for_max_increase); i += offset) {
+
+		lumin_millinits_temp = stream->lumin_data.luminance_millinits[i + offset];
+
+		if ((lumin_millinits_temp >= safe_upper_bound) || (lumin_millinits_temp <= safe_lower_bound)) {
+
+			if (stream->lumin_data.refresh_rate_hz[i + offset] == stream->lumin_data.refresh_rate_hz[i])
+				return stream->lumin_data.refresh_rate_hz[i];
+
+			int target_brightness = (stream->lumin_data.luminance_millinits[i + offset] >= (current_brightness + flicker_criteria_millinits)) ?
+											current_brightness + flicker_criteria_millinits :
+											current_brightness - flicker_criteria_millinits;
+
+			int refresh = 0;
+
+			/*
+			 * Need the second input to be < third input for dc_stream_get_refresh_hz_linear_interpolation
+			 */
+			if (search_for_max_increase)
+				refresh = dc_stream_get_refresh_hz_linear_interpolation(stream, i, i + offset, target_brightness);
+			else
+				refresh = dc_stream_get_refresh_hz_linear_interpolation(stream, i + offset, i, target_brightness);
+
+			if (refresh == stream->lumin_data.refresh_rate_hz[i + offset])
+				return stream->lumin_data.refresh_rate_hz[i + offset];
+
+			return refresh;
+		}
+	}
+
+	if (search_for_max_increase)
+		return (int)div64_s64((long long)stream->timing.pix_clk_100hz*100, stream->timing.v_total*(long long)stream->timing.h_total);
+	else
+		return stream->lumin_data.refresh_rate_hz[0];
+}
+
+/*
+ * Gets the max delta luminance within a specified refresh range
+ */
+static int dc_stream_get_max_delta_lumin_millinits(struct dc_stream_state *stream, int hz1, int hz2, bool isGaming)
+{
+	int lower_refresh_brightness = dc_stream_get_brightness_millinits_from_refresh (stream, hz1);
+	int higher_refresh_brightness = dc_stream_get_brightness_millinits_from_refresh (stream, hz2);
+
+	int min = lower_refresh_brightness;
+	int max = higher_refresh_brightness;
+
+	/*
+	 * Static screen, therefore no need to scan through array
+	 */
+	if (!isGaming) {
+		if (lower_refresh_brightness >= higher_refresh_brightness) {
+			return lower_refresh_brightness - higher_refresh_brightness;
+		}
+		return higher_refresh_brightness - lower_refresh_brightness;
+	}
+
+	min = MIN(lower_refresh_brightness, higher_refresh_brightness);
+	max = MAX(lower_refresh_brightness, higher_refresh_brightness);
+
+	int nearest_smallest_index = dc_stream_get_nearest_smallest_index(stream, hz1);
+
+	for (; nearest_smallest_index < (LUMINANCE_DATA_TABLE_SIZE - 1) &&
+			stream->lumin_data.refresh_rate_hz[nearest_smallest_index + 1] <= hz2 ; nearest_smallest_index++) {
+		min = MIN(min, stream->lumin_data.luminance_millinits[nearest_smallest_index + 1]);
+		max = MAX(max, stream->lumin_data.luminance_millinits[nearest_smallest_index + 1]);
+	}
+
+	return (max - min);
+}
+
+/*
+ * Determines the max flickerless instant vtotal delta for a stream.
+ * Determines vtotal increase/decrease based on the bool "increase"
+ */
+static unsigned int dc_stream_get_max_flickerless_instant_vtotal_delta(struct dc_stream_state *stream, bool is_gaming, bool increase)
+{
+	if (stream->timing.v_total * stream->timing.h_total == 0)
+		return 0;
+
+	int current_refresh_hz = (int)div64_s64((long long)stream->timing.pix_clk_100hz*100, stream->timing.v_total*(long long)stream->timing.h_total);
+
+	int safe_refresh_hz = dc_stream_calculate_flickerless_refresh_rate(stream,
+							 dc_stream_get_brightness_millinits_from_refresh(stream, current_refresh_hz),
+							 current_refresh_hz,
+							 is_gaming,
+							 increase);
+
+	int safe_refresh_v_total = (int)div64_s64((long long)stream->timing.pix_clk_100hz*100, safe_refresh_hz*(long long)stream->timing.h_total);
+
+	if (increase)
+		return (((int) stream->timing.v_total - safe_refresh_v_total) >= 0) ? (stream->timing.v_total - safe_refresh_v_total) : 0;
+
+	return ((safe_refresh_v_total - (int) stream->timing.v_total) >= 0) ? (safe_refresh_v_total - stream->timing.v_total) : 0;
+}
+
+/*
+ * Finds the highest refresh rate that can be achieved
+ * from starting_refresh_hz while staying within flicker criteria
+ */
+int dc_stream_calculate_max_flickerless_refresh_rate(struct dc_stream_state *stream, int starting_refresh_hz, bool is_gaming)
+{
+	if (!stream->lumin_data.is_valid)
+		return 0;
+
+	int current_brightness = dc_stream_get_brightness_millinits_from_refresh(stream, starting_refresh_hz);
+
+	return dc_stream_calculate_flickerless_refresh_rate(stream,
+							    current_brightness,
+							    starting_refresh_hz,
+							    is_gaming,
+							    true);
+}
+
+/*
+ * Finds the lowest refresh rate that can be achieved
+ * from starting_refresh_hz while staying within flicker criteria
+ */
+int dc_stream_calculate_min_flickerless_refresh_rate(struct dc_stream_state *stream, int starting_refresh_hz, bool is_gaming)
+{
+	if (!stream->lumin_data.is_valid)
+			return 0;
+
+	int current_brightness = dc_stream_get_brightness_millinits_from_refresh(stream, starting_refresh_hz);
+
+	return dc_stream_calculate_flickerless_refresh_rate(stream,
+							    current_brightness,
+							    starting_refresh_hz,
+							    is_gaming,
+							    false);
+}
+
+/*
+ * Determines if there will be a flicker when moving between 2 refresh rates
+ */
+bool dc_stream_is_refresh_rate_range_flickerless(struct dc_stream_state *stream, int hz1, int hz2, bool is_gaming)
+{
+
+	/*
+	 * Assume that we wont flicker if there is invalid data
+	 */
+	if (!stream->lumin_data.is_valid)
+		return false;
+
+	int dl = dc_stream_get_max_delta_lumin_millinits(stream, hz1, hz2, is_gaming);
+
+	int flicker_criteria_millinits = (is_gaming) ?
+					  stream->lumin_data.flicker_criteria_milli_nits_GAMING :
+					  stream->lumin_data.flicker_criteria_milli_nits_STATIC;
+
+	return (dl <= flicker_criteria_millinits);
+}
+
+/*
+ * Determines the max instant vtotal delta increase that can be applied without
+ * flickering for a given stream
+ */
+unsigned int dc_stream_get_max_flickerless_instant_vtotal_decrease(struct dc_stream_state *stream,
+									  bool is_gaming)
+{
+	if (!stream->lumin_data.is_valid)
+		return 0;
+
+	return dc_stream_get_max_flickerless_instant_vtotal_delta(stream, is_gaming, true);
+}
+
+/*
+ * Determines the max instant vtotal delta decrease that can be applied without
+ * flickering for a given stream
+ */
+unsigned int dc_stream_get_max_flickerless_instant_vtotal_increase(struct dc_stream_state *stream,
+									  bool is_gaming)
+{
+	if (!stream->lumin_data.is_valid)
+		return 0;
+
+	return dc_stream_get_max_flickerless_instant_vtotal_delta(stream, is_gaming, false);
+}
+
+bool dc_stream_is_cursor_limit_pending(struct dc *dc, struct dc_stream_state *stream)
+{
+	bool is_limit_pending = false;
+
+	if (dc->current_state)
+		is_limit_pending = dc_state_get_stream_cursor_subvp_limit(stream, dc->current_state);
+
+	return is_limit_pending;
+}
+
+bool dc_stream_can_clear_cursor_limit(struct dc *dc, struct dc_stream_state *stream)
+{
+	bool can_clear_limit = false;
+
+	if (dc->current_state)
+		can_clear_limit = dc_state_get_stream_cursor_subvp_limit(stream, dc->current_state) &&
+				(stream->hw_cursor_req ||
+				!stream->cursor_position.enable ||
+				dc_stream_check_cursor_attributes(stream, dc->current_state, &stream->cursor_attributes));
+
+	return can_clear_limit;
+}
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_surface.c b/drivers/gpu/drm/amd/display/dc/core/dc_surface.c
index e6b9c6a71841..922f23557f5d 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_surface.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_surface.c
@@ -23,8 +23,6 @@
  *
  */
 
-#include <linux/mm.h>
-
 /* DC interface (public) */
 #include "dm_services.h"
 #include "dc.h"
@@ -34,74 +32,58 @@
 #include "transform.h"
 #include "dpp.h"
 
+#include "dc_plane_priv.h"
+
 /*******************************************************************************
  * Private functions
  ******************************************************************************/
-static void dc_plane_construct(struct dc_context *ctx, struct dc_plane_state *plane_state)
+void dc_plane_construct(struct dc_context *ctx, struct dc_plane_state *plane_state)
 {
 	plane_state->ctx = ctx;
 
-	plane_state->gamma_correction = dc_create_gamma();
-	if (plane_state->gamma_correction != NULL)
-		plane_state->gamma_correction->is_identity = true;
+	plane_state->gamma_correction.is_identity = true;
 
-	plane_state->in_transfer_func = dc_create_transfer_func();
-	if (plane_state->in_transfer_func != NULL) {
-		plane_state->in_transfer_func->type = TF_TYPE_BYPASS;
-	}
-	plane_state->in_shaper_func = dc_create_transfer_func();
-	if (plane_state->in_shaper_func != NULL) {
-		plane_state->in_shaper_func->type = TF_TYPE_BYPASS;
-	}
+	plane_state->in_transfer_func.type = TF_TYPE_BYPASS;
 
-	plane_state->lut3d_func = dc_create_3dlut_func();
+	plane_state->in_shaper_func.type = TF_TYPE_BYPASS;
 
-	plane_state->blend_tf = dc_create_transfer_func();
-	if (plane_state->blend_tf != NULL) {
-		plane_state->blend_tf->type = TF_TYPE_BYPASS;
-	}
+	plane_state->lut3d_func.state.raw = 0;
+
+	plane_state->blend_tf.type = TF_TYPE_BYPASS;
+
+	plane_state->pre_multiplied_alpha = true;
 
 }
 
-static void dc_plane_destruct(struct dc_plane_state *plane_state)
+void dc_plane_destruct(struct dc_plane_state *plane_state)
 {
-	if (plane_state->gamma_correction != NULL) {
-		dc_gamma_release(&plane_state->gamma_correction);
-	}
-	if (plane_state->in_transfer_func != NULL) {
-		dc_transfer_func_release(
-				plane_state->in_transfer_func);
-		plane_state->in_transfer_func = NULL;
-	}
-	if (plane_state->in_shaper_func != NULL) {
-		dc_transfer_func_release(
-				plane_state->in_shaper_func);
-		plane_state->in_shaper_func = NULL;
-	}
-	if (plane_state->lut3d_func != NULL) {
-		dc_3dlut_func_release(
-				plane_state->lut3d_func);
-		plane_state->lut3d_func = NULL;
-	}
-	if (plane_state->blend_tf != NULL) {
-		dc_transfer_func_release(
-				plane_state->blend_tf);
-		plane_state->blend_tf = NULL;
+	// no more pointers to free within dc_plane_state
+}
+
+
+/* dc_state is passed in separately since it may differ from the current dc state accessible from plane_state e.g.
+ * if the driver is doing an update from an old context to a new one and the caller wants the pipe mask for the new
+ * context rather than the existing one
+ */
+uint8_t  dc_plane_get_pipe_mask(struct dc_state *dc_state, const struct dc_plane_state *plane_state)
+{
+	uint8_t pipe_mask = 0;
+	int i;
+
+	for (i = 0; i < plane_state->ctx->dc->res_pool->pipe_count; i++) {
+		struct pipe_ctx *pipe_ctx = &dc_state->res_ctx.pipe_ctx[i];
+
+		if (pipe_ctx->plane_state == plane_state && pipe_ctx->plane_res.hubp)
+			pipe_mask |= 1 << pipe_ctx->plane_res.hubp->inst;
 	}
 
+	return pipe_mask;
 }
 
 /*******************************************************************************
  * Public functions
  ******************************************************************************/
-void enable_surface_flip_reporting(struct dc_plane_state *plane_state,
-		uint32_t controller_id)
-{
-	plane_state->irq_source = controller_id + DC_IRQ_SOURCE_PFLIP1 - 1;
-	/*register_flip_interrupt(surface);*/
-}
-
-struct dc_plane_state *dc_create_plane_state(struct dc *dc)
+struct dc_plane_state *dc_create_plane_state(const struct dc *dc)
 {
 	struct dc_plane_state *plane_state = kvzalloc(sizeof(*plane_state),
 							GFP_KERNEL);
@@ -127,7 +109,8 @@ struct dc_plane_state *dc_create_plane_state(struct dc *dc)
  *****************************************************************************
  */
 const struct dc_plane_status *dc_plane_get_status(
-		const struct dc_plane_state *plane_state)
+		const struct dc_plane_state *plane_state,
+		union dc_plane_status_update_flags flags)
 {
 	const struct dc_plane_status *plane_status;
 	struct dc  *dc;
@@ -154,11 +137,14 @@ const struct dc_plane_status *dc_plane_get_status(
 		if (pipe_ctx->plane_state != plane_state)
 			continue;
 
-		pipe_ctx->plane_state->status.is_flip_pending = false;
+		if (pipe_ctx->plane_state && flags.bits.address)
+			pipe_ctx->plane_state->status.is_flip_pending = false;
 
 		break;
 	}
 
+	dc_exit_ips_for_hw_access(dc);
+
 	for (i = 0; i < dc->res_pool->pipe_count; i++) {
 		struct pipe_ctx *pipe_ctx =
 				&dc->current_state->res_ctx.pipe_ctx[i];
@@ -166,7 +152,8 @@ const struct dc_plane_status *dc_plane_get_status(
 		if (pipe_ctx->plane_state != plane_state)
 			continue;
 
-		dc->hwss.update_pending_status(pipe_ctx);
+		if (flags.bits.address)
+			dc->hwss.update_pending_status(pipe_ctx);
 	}
 
 	return plane_status;
@@ -285,4 +272,41 @@ void dc_3dlut_func_retain(struct dc_3dlut *lut)
 	kref_get(&lut->refcount);
 }
 
+void dc_plane_force_dcc_and_tiling_disable(struct dc_plane_state *plane_state,
+					   bool clear_tiling)
+{
+	struct dc *dc;
+	int i;
+
+	if (!plane_state)
+		return;
+
+	dc = plane_state->ctx->dc;
+
+	if (!dc || !dc->current_state)
+		return;
+
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		struct pipe_ctx *pipe_ctx = &dc->current_state->res_ctx.pipe_ctx[i];
+
+		if (!pipe_ctx)
+			continue;
+
+		if (dc->hwss.clear_surface_dcc_and_tiling)
+			dc->hwss.clear_surface_dcc_and_tiling(pipe_ctx, plane_state, clear_tiling);
+	}
+}
+
+void dc_plane_copy_config(struct dc_plane_state *dst, const struct dc_plane_state *src)
+{
+	struct kref temp_refcount;
+
+	/* backup persistent info */
+	memcpy(&temp_refcount, &dst->refcount, sizeof(struct kref));
 
+	/* copy all configuration information */
+	memcpy(dst, src, sizeof(struct dc_plane_state));
+
+	/* restore persistent info */
+	memcpy(&dst->refcount, &temp_refcount, sizeof(struct kref));
+}
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_vm_helper.c b/drivers/gpu/drm/amd/display/dc/core/dc_vm_helper.c
index cde8ed2560b3..d1e68dc57a2a 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_vm_helper.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_vm_helper.c
@@ -47,9 +47,8 @@ int dc_setup_system_context(struct dc *dc, struct dc_phy_addr_space_config *pa_c
 		 */
 		memcpy(&dc->vm_pa_config, pa_config, sizeof(struct dc_phy_addr_space_config));
 		dc->vm_pa_config.valid = true;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
+		dc->dml2_options.gpuvm_enable = true;
 		dc_z10_save_init(dc);
-#endif
 	}
 
 	return num_vmids;
diff --git a/drivers/gpu/drm/amd/display/dc/dc.h b/drivers/gpu/drm/amd/display/dc/dc.h
index 618e7989176f..29edfa51ea2c 100644
--- a/drivers/gpu/drm/amd/display/dc/dc.h
+++ b/drivers/gpu/drm/amd/display/dc/dc.h
@@ -1,5 +1,5 @@
 /*
- * Copyright 2012-14 Advanced Micro Devices, Inc.
+ * Copyright 2012-2023 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -27,45 +27,71 @@
 #define DC_INTERFACE_H_
 
 #include "dc_types.h"
+#include "dc_state.h"
+#include "dc_plane.h"
 #include "grph_object_defs.h"
 #include "logger_types.h"
-#if defined(CONFIG_DRM_AMD_DC_HDCP)
-#include "hdcp_types.h"
-#endif
+#include "hdcp_msg_types.h"
 #include "gpio_types.h"
 #include "link_service_types.h"
 #include "grph_object_ctrl_defs.h"
 #include <inc/hw/opp.h>
 
-#include "inc/hw_sequencer.h"
+#include "hwss/hw_sequencer.h"
 #include "inc/compressor.h"
 #include "inc/hw/dmcu.h"
 #include "dml/display_mode_lib.h"
 
+#include "dml2_0/dml2_wrapper.h"
+
+#include "dmub/inc/dmub_cmd.h"
+
+#include "sspl/dc_spl_types.h"
+
+struct abm_save_restore;
+
 /* forward declaration */
 struct aux_payload;
 struct set_config_cmd_payload;
 struct dmub_notification;
-
-#define DC_VER "3.2.160"
-
-#define MAX_SURFACES 3
+struct dcn_hubbub_reg_state;
+struct dcn_hubp_reg_state;
+struct dcn_dpp_reg_state;
+struct dcn_mpc_reg_state;
+struct dcn_opp_reg_state;
+struct dcn_dsc_reg_state;
+struct dcn_optc_reg_state;
+struct dcn_dccg_reg_state;
+
+#define DC_VER "3.2.359"
+
+/**
+ * MAX_SURFACES - representative of the upper bound of surfaces that can be piped to a single CRTC
+ */
+#define MAX_SURFACES 4
+/**
+ * MAX_PLANES - representative of the upper bound of planes that are supported by the HW
+ */
 #define MAX_PLANES 6
 #define MAX_STREAMS 6
-#define MAX_SINKS_PER_LINK 4
 #define MIN_VIEWPORT_SIZE 12
 #define MAX_NUM_EDP 2
+#define MAX_SUPPORTED_FORMATS 7
 
-/*******************************************************************************
- * Display Core Interfaces
- ******************************************************************************/
+#define MAX_HOST_ROUTERS_NUM 3
+#define MAX_DPIA_PER_HOST_ROUTER 3
+#define MAX_DPIA_NUM  (MAX_HOST_ROUTERS_NUM * MAX_DPIA_PER_HOST_ROUTER)
+
+/* Display Core Interfaces */
 struct dc_versions {
 	const char *dc_ver;
 	struct dmcu_version dmcu_version;
 };
 
 enum dp_protocol_version {
-	DP_VERSION_1_4,
+	DP_VERSION_1_4 = 0,
+	DP_VERSION_2_1,
+	DP_VERSION_UNKNOWN,
 };
 
 enum dc_plane_type {
@@ -75,10 +101,18 @@ enum dc_plane_type {
 	DC_PLANE_TYPE_DCN_UNIVERSAL,
 };
 
+// Sizes defined as multiples of 64KB
+enum det_size {
+	DET_SIZE_DEFAULT = 0,
+	DET_SIZE_192KB = 3,
+	DET_SIZE_256KB = 4,
+	DET_SIZE_320KB = 5,
+	DET_SIZE_384KB = 6
+};
+
+
 struct dc_plane_cap {
 	enum dc_plane_type type;
-	uint32_t blends_with_above : 1;
-	uint32_t blends_with_below : 1;
 	uint32_t per_pixel_alpha : 1;
 	struct {
 		uint32_t argb8888 : 1;
@@ -108,7 +142,26 @@ struct dc_plane_cap {
 	uint32_t min_height;
 };
 
-// Color management caps (DPP and MPC)
+/**
+ * DOC: color-management-caps
+ *
+ * **Color management caps (DPP and MPC)**
+ *
+ * Modules/color calculates various color operations which are translated to
+ * abstracted HW. DCE 5-12 had almost no important changes, but starting with
+ * DCN1, every new generation comes with fairly major differences in color
+ * pipeline. Therefore, we abstract color pipe capabilities so modules/DM can
+ * decide mapping to HW block based on logical capabilities.
+ */
+
+/**
+ * struct rom_curve_caps - predefined transfer function caps for degamma and regamma
+ * @srgb: RGB color space transfer func
+ * @bt2020: BT.2020 transfer func
+ * @gamma2_2: standard gamma
+ * @pq: perceptual quantizer transfer function
+ * @hlg: hybrid log–gamma transfer function
+ */
 struct rom_curve_caps {
 	uint16_t srgb : 1;
 	uint16_t bt2020 : 1;
@@ -117,41 +170,131 @@ struct rom_curve_caps {
 	uint16_t hlg : 1;
 };
 
+/**
+ * struct dpp_color_caps - color pipeline capabilities for display pipe and
+ * plane blocks
+ *
+ * @dcn_arch: all DCE generations treated the same
+ * @input_lut_shared: shared with DGAM. Input LUT is different than most LUTs,
+ * just plain 256-entry lookup
+ * @icsc: input color space conversion
+ * @dgam_ram: programmable degamma LUT
+ * @post_csc: post color space conversion, before gamut remap
+ * @gamma_corr: degamma correction
+ * @hw_3d_lut: 3D LUT support. It implies a shaper LUT before. It may be shared
+ * with MPC by setting mpc:shared_3d_lut flag
+ * @ogam_ram: programmable out/blend gamma LUT
+ * @ocsc: output color space conversion
+ * @dgam_rom_for_yuv: pre-defined degamma LUT for YUV planes
+ * @dgam_rom_caps: pre-definied curve caps for degamma 1D LUT
+ * @ogam_rom_caps: pre-definied curve caps for regamma 1D LUT
+ *
+ * Note: hdr_mult and gamut remap (CTM) are always available in DPP (in that order)
+ */
 struct dpp_color_caps {
-	uint16_t dcn_arch : 1; // all DCE generations treated the same
-	// input lut is different than most LUTs, just plain 256-entry lookup
-	uint16_t input_lut_shared : 1; // shared with DGAM
+	uint16_t dcn_arch : 1;
+	uint16_t input_lut_shared : 1;
 	uint16_t icsc : 1;
 	uint16_t dgam_ram : 1;
-	uint16_t post_csc : 1; // before gamut remap
+	uint16_t post_csc : 1;
 	uint16_t gamma_corr : 1;
-
-	// hdr_mult and gamut remap always available in DPP (in that order)
-	// 3d lut implies shaper LUT,
-	// it may be shared with MPC - check MPC:shared_3d_lut flag
 	uint16_t hw_3d_lut : 1;
-	uint16_t ogam_ram : 1; // blnd gam
+	uint16_t ogam_ram : 1;
 	uint16_t ocsc : 1;
 	uint16_t dgam_rom_for_yuv : 1;
 	struct rom_curve_caps dgam_rom_caps;
 	struct rom_curve_caps ogam_rom_caps;
 };
 
+/* Below structure is to describe the HW support for mem layout, extend support
+	range to match what OS could handle in the roadmap */
+struct lut3d_caps {
+	uint32_t dma_3d_lut : 1; /*< DMA mode support for 3D LUT */
+	struct {
+		uint32_t swizzle_3d_rgb : 1;
+		uint32_t swizzle_3d_bgr : 1;
+		uint32_t linear_1d : 1;
+	} mem_layout_support;
+	struct {
+		uint32_t unorm_12msb : 1;
+		uint32_t unorm_12lsb : 1;
+		uint32_t float_fp1_5_10 : 1;
+	} mem_format_support;
+	struct {
+		uint32_t order_rgba : 1;
+		uint32_t order_bgra : 1;
+	} mem_pixel_order_support;
+	/*< size options are 9, 17, 33, 45, 65 */
+	struct {
+		uint32_t dim_9 : 1; /* 3D LUT support for 9x9x9 */
+		uint32_t dim_17 : 1; /* 3D LUT support for 17x17x17 */
+		uint32_t dim_33 : 1; /* 3D LUT support for 33x33x33 */
+		uint32_t dim_45 : 1; /* 3D LUT support for 45x45x45 */
+		uint32_t dim_65 : 1; /* 3D LUT support for 65x65x65 */
+	} lut_dim_caps;
+};
+
+/**
+ * struct mpc_color_caps - color pipeline capabilities for multiple pipe and
+ * plane combined blocks
+ *
+ * @gamut_remap: color transformation matrix
+ * @ogam_ram: programmable out gamma LUT
+ * @ocsc: output color space conversion matrix
+ * @num_3dluts: MPC 3D LUT; always assumes a preceding shaper LUT
+ * @num_rmcm_3dluts: number of RMCM 3D LUTS; always assumes a preceding shaper LUT
+ * @shared_3d_lut: shared 3D LUT flag. Can be either DPP or MPC, but single
+ * instance
+ * @ogam_rom_caps: pre-definied curve caps for regamma 1D LUT
+ * @mcm_3d_lut_caps: HW support cap for MCM LUT memory
+ * @rmcm_3d_lut_caps: HW support cap for RMCM LUT memory
+ * @preblend: whether color manager supports preblend with MPC
+ */
 struct mpc_color_caps {
 	uint16_t gamut_remap : 1;
 	uint16_t ogam_ram : 1;
 	uint16_t ocsc : 1;
-	uint16_t num_3dluts : 3; //3d lut always assumes a preceding shaper LUT
-	uint16_t shared_3d_lut:1; //can be in either DPP or MPC, but single instance
-
+	uint16_t num_3dluts : 3;
+	uint16_t num_rmcm_3dluts : 3;
+	uint16_t shared_3d_lut:1;
 	struct rom_curve_caps ogam_rom_caps;
+	struct lut3d_caps mcm_3d_lut_caps;
+	struct lut3d_caps rmcm_3d_lut_caps;
+	bool preblend;
 };
 
+/**
+ * struct dc_color_caps - color pipes capabilities for DPP and MPC hw blocks
+ * @dpp: color pipes caps for DPP
+ * @mpc: color pipes caps for MPC
+ */
 struct dc_color_caps {
 	struct dpp_color_caps dpp;
 	struct mpc_color_caps mpc;
 };
 
+struct dc_dmub_caps {
+	bool psr;
+	bool mclk_sw;
+	bool subvp_psr;
+	bool gecc_enable;
+	uint8_t fams_ver;
+	bool aux_backlight_support;
+};
+
+struct dc_scl_caps {
+	bool sharpener_support;
+};
+
+struct dc_check_config {
+	/**
+	 * max video plane width that can be safely assumed to be always
+	 * supported by single DPP pipe.
+	 */
+	unsigned int max_optimizable_video_width;
+	bool enable_legacy_fast_update;
+};
+
 struct dc_caps {
 	uint32_t max_streams;
 	uint32_t max_links;
@@ -165,6 +308,7 @@ struct dc_caps {
 	uint32_t i2c_speed_in_khz_hdcp;
 	uint32_t dmdata_alloc_size;
 	unsigned int max_cursor_size;
+	unsigned int max_buffered_cursor_size;
 	unsigned int max_video_width;
 	unsigned int min_horizontal_blanking_period;
 	int linear_pitch_alignment;
@@ -178,6 +322,9 @@ struct dc_caps {
 	bool psp_setup_panel_mode;
 	bool extended_aux_timeout_support;
 	bool dmcub_support;
+	bool zstate_support;
+	bool ips_support;
+	bool ips_v2_support;
 	uint32_t num_of_internal_disp;
 	enum dp_protocol_version max_dp_protocol_version;
 	unsigned int mall_size_per_mem_channel;
@@ -185,11 +332,38 @@ struct dc_caps {
 	unsigned int cursor_cache_size;
 	struct dc_plane_cap planes[MAX_PLANES];
 	struct dc_color_caps color;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
+	struct dc_dmub_caps dmub_caps;
 	bool dp_hpo;
-#endif
+	bool dp_hdmi21_pcon_support;
+	bool edp_dsc_support;
 	bool vbios_lttpr_aware;
 	bool vbios_lttpr_enable;
+	bool fused_io_supported;
+	uint32_t max_otg_num;
+	uint32_t max_cab_allocation_bytes;
+	uint32_t cache_line_size;
+	uint32_t cache_num_ways;
+	uint16_t subvp_fw_processing_delay_us;
+	uint8_t subvp_drr_max_vblank_margin_us;
+	uint16_t subvp_prefetch_end_to_mall_start_us;
+	uint8_t subvp_swath_height_margin_lines; // subvp start line must be aligned to 2 x swath height
+	uint16_t subvp_pstate_allow_width_us;
+	uint16_t subvp_vertical_int_margin_us;
+	bool seamless_odm;
+	uint32_t max_v_total;
+	bool vtotal_limited_by_fp2;
+	uint32_t max_disp_clock_khz_at_vmin;
+	uint8_t subvp_drr_vblank_start_margin_us;
+	bool cursor_not_scaled;
+	bool dcmode_power_limits_present;
+	bool sequential_ono;
+	/* Conservative limit for DCC cases which require ODM4:1 to support*/
+	uint32_t dcc_plane_width_limit;
+	struct dc_scl_caps scl_caps;
+	uint8_t num_of_host_routers;
+	uint8_t num_of_dpias_per_host_router;
+	/* limit of the ODM only, could be limited by other factors (like pipe count)*/
+	uint8_t max_odm_combine_factor;
 };
 
 struct dc_bug_wa {
@@ -197,12 +371,24 @@ struct dc_bug_wa {
 	bool dedcn20_305_wa;
 	bool skip_clock_update;
 	bool lt_early_cr_pattern;
+	struct {
+		uint8_t uclk : 1;
+		uint8_t fclk : 1;
+		uint8_t dcfclk : 1;
+		uint8_t dcfclk_ds: 1;
+	} clock_update_disable_mask;
+	bool skip_psr_ips_crtc_disable;
 };
-
 struct dc_dcc_surface_param {
 	struct dc_size surface_size;
 	enum surface_pixel_format format;
-	enum swizzle_mode_values swizzle_mode;
+	unsigned int plane0_pitch;
+	struct dc_size plane1_size;
+	unsigned int plane1_pitch;
+	union {
+		enum swizzle_mode_values swizzle_mode;
+		enum swizzle_mode_addr3_values swizzle_mode_addr3;
+	};
 	enum dc_scan_direction scan;
 };
 
@@ -210,15 +396,16 @@ struct dc_dcc_setting {
 	unsigned int max_compressed_blk_size;
 	unsigned int max_uncompressed_blk_size;
 	bool independent_64b_blks;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	//These bitfields to be used starting with DCN
+	//These bitfields to be used starting with DCN 3.0
 	struct {
-		uint32_t dcc_256_64_64 : 1;//available in ASICs before DCN (the worst compression case)
-		uint32_t dcc_128_128_uncontrained : 1;  //available in ASICs before DCN
-		uint32_t dcc_256_128_128 : 1;		//available starting with DCN
-		uint32_t dcc_256_256_unconstrained : 1;  //available in ASICs before DCN (the best compression case)
+		uint32_t dcc_256_64_64 : 1;//available in ASICs before DCN 3.0 (the worst compression case)
+		uint32_t dcc_128_128_uncontrained : 1;  //available in ASICs before DCN 3.0
+		uint32_t dcc_256_128_128 : 1;		//available starting with DCN 3.0
+		uint32_t dcc_256_256_unconstrained : 1;  //available in ASICs before DCN 3.0 (the best compression case)
+		uint32_t dcc_256_256 : 1;  //available in ASICs starting with DCN 4.0x (the best compression case)
+		uint32_t dcc_256_128 : 1;  //available in ASICs starting with DCN 4.0x
+		uint32_t dcc_256_64 : 1;   //available in ASICs starting with DCN 4.0x (the worst compression case)
 	} dcc_controls;
-#endif
 };
 
 struct dc_surface_dcc_cap {
@@ -280,21 +467,32 @@ enum surface_update_type {
 	UPDATE_TYPE_FULL, /* may need to shuffle resources */
 };
 
+enum dc_lock_descriptor {
+	LOCK_DESCRIPTOR_NONE = 0x0,
+	LOCK_DESCRIPTOR_STREAM = 0x1,
+	LOCK_DESCRIPTOR_LINK = 0x2,
+	LOCK_DESCRIPTOR_GLOBAL = 0x4,
+};
+
+struct surface_update_descriptor {
+	enum surface_update_type update_type;
+	enum dc_lock_descriptor lock_descriptor;
+};
+
 /* Forward declaration*/
 struct dc;
 struct dc_plane_state;
 struct dc_state;
 
-
 struct dc_cap_funcs {
 	bool (*get_dcc_compression_cap)(const struct dc *dc,
 			const struct dc_dcc_surface_param *input,
 			struct dc_surface_dcc_cap *output);
+	bool (*get_subvp_en)(struct dc *dc, struct dc_state *context);
 };
 
 struct link_training_settings;
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 union allow_lttpr_non_transparent_mode {
 	struct {
 		bool DP1_4A : 1;
@@ -302,7 +500,7 @@ union allow_lttpr_non_transparent_mode {
 	} bits;
 	unsigned char raw;
 };
-#endif
+
 /* Structure to hold configuration flags set by dm at dc creation. */
 struct dc_config {
 	bool gpu_vm_support;
@@ -310,28 +508,55 @@ struct dc_config {
 	bool fbc_support;
 	bool disable_fractional_pwm;
 	bool allow_seamless_boot_optimization;
-	bool power_down_display_on_boot;
+	bool seamless_boot_edp_requested;
 	bool edp_not_connected;
 	bool edp_no_power_sequencing;
 	bool force_enum_edp;
 	bool forced_clocks;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	union allow_lttpr_non_transparent_mode allow_lttpr_non_transparent_mode;
-#else
-	bool allow_lttpr_non_transparent_mode;
-#endif
 	bool multi_mon_pp_mclk_switch;
 	bool disable_dmcu;
 	bool enable_4to1MPC;
 	bool enable_windowed_mpo_odm;
-	bool allow_edp_hotplug_detection;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
+	bool forceHBR2CP2520; // Used for switching between test patterns TPS4 and CP2520
+	uint32_t allow_edp_hotplug_detection;
+	bool skip_riommu_prefetch_wa;
 	bool clamp_min_dcfclk;
-#endif
 	uint64_t vblank_alignment_dto_params;
 	uint8_t  vblank_alignment_max_frame_time_diff;
 	bool is_asymmetric_memory;
 	bool is_single_rank_dimm;
+	bool is_vmin_only_asic;
+	bool use_spl;
+	bool prefer_easf;
+	bool use_pipe_ctx_sync_logic;
+	int smart_mux_version;
+	bool ignore_dpref_ss;
+	bool enable_mipi_converter_optimization;
+	bool use_default_clock_table;
+	bool force_bios_enable_lttpr;
+	uint8_t force_bios_fixed_vs;
+	int sdpif_request_limit_words_per_umc;
+	bool dc_mode_clk_limit_support;
+	bool EnableMinDispClkODM;
+	bool enable_auto_dpm_test_logs;
+	unsigned int disable_ips;
+	unsigned int disable_ips_rcg;
+	unsigned int disable_ips_in_vpb;
+	bool disable_ips_in_dpms_off;
+	bool usb4_bw_alloc_support;
+	bool allow_0_dtb_clk;
+	bool use_assr_psp_message;
+	bool support_edp0_on_dp1;
+	unsigned int enable_fpo_flicker_detection;
+	bool disable_hbr_audio_dp2;
+	bool consolidated_dpia_dp_lt;
+	bool set_pipe_unlock_order;
+	bool enable_dpia_pre_training;
+	bool unify_link_enc_assignment;
+	bool enable_cursor_offload;
+	struct spl_sharpness_range dcn_sharpness_range;
+	struct spl_sharpness_range dcn_override_sharpness_range;
 };
 
 enum visual_confirm {
@@ -340,13 +565,37 @@ enum visual_confirm {
 	VISUAL_CONFIRM_HDR = 2,
 	VISUAL_CONFIRM_MPCTREE = 4,
 	VISUAL_CONFIRM_PSR = 5,
+	VISUAL_CONFIRM_SWAPCHAIN = 6,
+	VISUAL_CONFIRM_FAMS = 7,
 	VISUAL_CONFIRM_SWIZZLE = 9,
+	VISUAL_CONFIRM_SMARTMUX_DGPU = 10,
+	VISUAL_CONFIRM_REPLAY = 12,
+	VISUAL_CONFIRM_SUBVP = 14,
+	VISUAL_CONFIRM_MCLK_SWITCH = 16,
+	VISUAL_CONFIRM_FAMS2 = 19,
+	VISUAL_CONFIRM_HW_CURSOR = 20,
+	VISUAL_CONFIRM_VABC = 21,
+	VISUAL_CONFIRM_DCC = 22,
+	VISUAL_CONFIRM_EXPLICIT = 0x80000000,
 };
 
 enum dc_psr_power_opts {
 	psr_power_opt_invalid = 0x0,
 	psr_power_opt_smu_opt_static_screen = 0x1,
 	psr_power_opt_z10_static_screen = 0x10,
+	psr_power_opt_ds_disable_allow = 0x100,
+};
+
+enum dml_hostvm_override_opts {
+	DML_HOSTVM_NO_OVERRIDE = 0x0,
+	DML_HOSTVM_OVERRIDE_FALSE = 0x1,
+	DML_HOSTVM_OVERRIDE_TRUE = 0x2,
+};
+
+enum dc_replay_power_opts {
+	replay_power_opt_invalid		= 0x0,
+	replay_power_opt_smu_opt_static_screen	= 0x1,
+	replay_power_opt_z10_static_screen	= 0x10,
 };
 
 enum dcc_option {
@@ -355,9 +604,38 @@ enum dcc_option {
 	DCC_HALF_REQ_DISALBE = 2,
 };
 
+enum in_game_fams_config {
+	INGAME_FAMS_SINGLE_DISP_ENABLE, // enable in-game fams
+	INGAME_FAMS_DISABLE, // disable in-game fams
+	INGAME_FAMS_MULTI_DISP_ENABLE, //enable in-game fams for multi-display
+	INGAME_FAMS_MULTI_DISP_CLAMPED_ONLY, //enable in-game fams for multi-display only for clamped RR strategies
+};
+
+/**
+ * enum pipe_split_policy - Pipe split strategy supported by DCN
+ *
+ * This enum is used to define the pipe split policy supported by DCN. By
+ * default, DC favors MPC_SPLIT_DYNAMIC.
+ */
 enum pipe_split_policy {
+	/**
+	 * @MPC_SPLIT_DYNAMIC: DC will automatically decide how to split the
+	 * pipe in order to bring the best trade-off between performance and
+	 * power consumption. This is the recommended option.
+	 */
 	MPC_SPLIT_DYNAMIC = 0,
+
+	/**
+	 * @MPC_SPLIT_AVOID: Avoid pipe split, which means that DC will not
+	 * try any sort of split optimization.
+	 */
 	MPC_SPLIT_AVOID = 1,
+
+	/**
+	 * @MPC_SPLIT_AVOID_MULT_DISP: With this option, DC will only try to
+	 * optimize the pipe utilization when using a single display; if the
+	 * user connects to a second display, DC will avoid pipe split.
+	 */
 	MPC_SPLIT_AVOID_MULT_DISP = 2,
 };
 
@@ -379,16 +657,20 @@ enum dcn_pwr_state {
 	DCN_PWR_STATE_LOW_POWER = 3,
 };
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 enum dcn_zstate_support_state {
 	DCN_ZSTATE_SUPPORT_UNKNOWN,
 	DCN_ZSTATE_SUPPORT_ALLOW,
+	DCN_ZSTATE_SUPPORT_ALLOW_Z8_ONLY,
+	DCN_ZSTATE_SUPPORT_ALLOW_Z8_Z10_ONLY,
+	DCN_ZSTATE_SUPPORT_ALLOW_Z10_ONLY,
 	DCN_ZSTATE_SUPPORT_DISALLOW,
 };
-#endif
+
 /*
- * For any clocks that may differ per pipe
- * only the max is stored in this structure
+ * struct dc_clocks - DC pipe clocks
+ *
+ * For any clocks that may differ per pipe only the max is stored in this
+ * structure
  */
 struct dc_clocks {
 	int dispclk_khz;
@@ -403,21 +685,50 @@ struct dc_clocks {
 	int phyclk_khz;
 	int dramclk_khz;
 	bool p_state_change_support;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	enum dcn_zstate_support_state zstate_support;
 	bool dtbclk_en;
-#endif
+	int ref_dtbclk_khz;
+	bool fclk_p_state_change_support;
 	enum dcn_pwr_state pwr_state;
 	/*
 	 * Elements below are not compared for the purposes of
 	 * optimization required
 	 */
 	bool prev_p_state_change_support;
+	bool fclk_prev_p_state_change_support;
+	int num_ways;
+	int host_router_bw_kbps[MAX_HOST_ROUTERS_NUM];
+
+	/*
+	 * @fw_based_mclk_switching
+	 *
+	 * DC has a mechanism that leverage the variable refresh rate to switch
+	 * memory clock in cases that we have a large latency to achieve the
+	 * memory clock change and a short vblank window. DC has some
+	 * requirements to enable this feature, and this field describes if the
+	 * system support or not such a feature.
+	 */
+	bool fw_based_mclk_switching;
+	bool fw_based_mclk_switching_shut_down;
+	int prev_num_ways;
 	enum dtm_pstate dtm_level;
 	int max_supported_dppclk_khz;
 	int max_supported_dispclk_khz;
 	int bw_dppclk_khz; /*a copy of dppclk_khz*/
 	int bw_dispclk_khz;
+	int idle_dramclk_khz;
+	int idle_fclk_khz;
+	int subvp_prefetch_dramclk_khz;
+	int subvp_prefetch_fclk_khz;
+
+	/* Stutter efficiency is technically not clock values
+	 * but stored here so the values are part of the update_clocks call similar to num_ways
+	 * Efficiencies are stored as percentage (0-100)
+	 */
+	struct {
+		uint8_t base_efficiency; //LP1
+		uint8_t low_power_efficiency; //LP2
+	} stutter_efficiency;
 };
 
 struct dc_bw_validation_profile {
@@ -503,13 +814,83 @@ union root_clock_optimization_options {
 	uint32_t u32All;
 };
 
+union fine_grain_clock_gating_enable_options {
+	struct {
+		bool dccg_global_fgcg_rep : 1; /* Global fine grain clock gating of repeaters */
+		bool dchub : 1;	   /* Display controller hub */
+		bool dchubbub : 1;
+		bool dpp : 1;	   /* Display pipes and planes */
+		bool opp : 1;	   /* Output pixel processing */
+		bool optc : 1;	   /* Output pipe timing combiner */
+		bool dio : 1;	   /* Display output */
+		bool dwb : 1;	   /* Display writeback */
+		bool mmhubbub : 1; /* Multimedia hub */
+		bool dmu : 1;	   /* Display core management unit */
+		bool az : 1;	   /* Azalia */
+		bool dchvm : 1;
+		bool dsc : 1;	   /* Display stream compression */
+
+		uint32_t reserved : 19;
+	} bits;
+	uint32_t u32All;
+};
+
+enum pg_hw_pipe_resources {
+	PG_HUBP = 0,
+	PG_DPP,
+	PG_DSC,
+	PG_MPCC,
+	PG_OPP,
+	PG_OPTC,
+	PG_DPSTREAM,
+	PG_HDMISTREAM,
+	PG_PHYSYMCLK,
+	PG_HW_PIPE_RESOURCES_NUM_ELEMENT
+};
+
+enum pg_hw_resources {
+	PG_DCCG = 0,
+	PG_DCIO,
+	PG_DIO,
+	PG_DCHUBBUB,
+	PG_DCHVM,
+	PG_DWB,
+	PG_HPO,
+	PG_DCOH,
+	PG_HW_RESOURCES_NUM_ELEMENT
+};
+
+struct pg_block_update {
+	bool pg_pipe_res_update[PG_HW_PIPE_RESOURCES_NUM_ELEMENT][MAX_PIPES];
+	bool pg_res_update[PG_HW_RESOURCES_NUM_ELEMENT];
+};
+
 union dpia_debug_options {
 	struct {
-		uint32_t disable_dpia:1;
-		uint32_t force_non_lttpr:1;
-		uint32_t extend_aux_rd_interval:1;
-		uint32_t disable_mst_dsc_work_around:1;
-		uint32_t reserved:28;
+		uint32_t disable_dpia:1; /* bit 0 */
+		uint32_t force_non_lttpr:1; /* bit 1 */
+		uint32_t extend_aux_rd_interval:1; /* bit 2 */
+		uint32_t disable_mst_dsc_work_around:1; /* bit 3 */
+		uint32_t enable_force_tbt3_work_around:1; /* bit 4 */
+		uint32_t disable_usb4_pm_support:1; /* bit 5 */
+		uint32_t enable_usb4_bw_zero_alloc_patch:1; /* bit 6 */
+		uint32_t reserved:25;
+	} bits;
+	uint32_t raw;
+};
+
+/* AUX wake work around options
+ * 0: enable/disable work around
+ * 1: use default timeout LINK_AUX_WAKE_TIMEOUT_MS
+ * 15-2: reserved
+ * 31-16: timeout in ms
+ */
+union aux_wake_wa_options {
+	struct {
+		uint32_t enable_wa : 1;
+		uint32_t use_default_timeout : 1;
+		uint32_t rsvd: 14;
+		uint32_t timeout_ms : 16;
 	} bits;
 	uint32_t raw;
 };
@@ -518,6 +899,7 @@ struct dc_debug_data {
 	uint32_t ltFailCount;
 	uint32_t i2cErrorCount;
 	uint32_t auxErrorCount;
+	struct pipe_topology_history topology_history;
 };
 
 struct dc_phy_addr_space_config {
@@ -555,10 +937,13 @@ struct dc_virtual_addr_space_config {
 struct dc_bounding_box_overrides {
 	int sr_exit_time_ns;
 	int sr_enter_plus_exit_time_ns;
+	int sr_exit_z8_time_ns;
+	int sr_enter_plus_exit_z8_time_ns;
 	int urgent_latency_ns;
 	int percent_of_ideal_drambw;
 	int dram_clock_change_latency_ns;
 	int dummy_clock_change_latency_ns;
+	int fclk_clock_change_latency_ns;
 	/* This forces a hard min on the DCFCLK we use
 	 * for DML.  Unlike the debug option for forcing
 	 * DCFCLK, this override affects watermark calculations
@@ -569,15 +954,25 @@ struct dc_bounding_box_overrides {
 struct dc_state;
 struct resource_pool;
 struct dce_hwseq;
+struct link_service;
 
+/*
+ * struct dc_debug_options - DC debug struct
+ *
+ * This struct provides a simple mechanism for developers to change some
+ * configurations, enable/disable features, and activate extra debug options.
+ * This can be very handy to narrow down whether some specific feature is
+ * causing an issue or not.
+ */
 struct dc_debug_options {
 	bool native422_support;
 	bool disable_dsc;
 	enum visual_confirm visual_confirm;
+	int visual_confirm_rect_height;
+
 	bool sanity_checks;
 	bool max_disp_clk;
 	bool surface_trace;
-	bool timing_trace;
 	bool clock_trace;
 	bool validation_trace;
 	bool bandwidth_calcs_trace;
@@ -587,15 +982,28 @@ struct dc_debug_options {
 	bool disable_stutter;
 	bool use_max_lb;
 	enum dcc_option disable_dcc;
+
+	/*
+	 * @pipe_split_policy: Define which pipe split policy is used by the
+	 * display core.
+	 */
 	enum pipe_split_policy pipe_split_policy;
 	bool force_single_disp_pipe_split;
 	bool voltage_align_fclk;
 	bool disable_min_fclk;
 
+	bool hdcp_lc_force_fw_enable;
+	bool hdcp_lc_enable_sw_fallback;
+
 	bool disable_dfs_bypass;
 	bool disable_dpp_power_gate;
 	bool disable_hubp_power_gate;
 	bool disable_dsc_power_gate;
+	bool disable_optc_power_gate;
+	bool disable_hpo_power_gate;
+	bool disable_io_clk_power_gate;
+	bool disable_mem_power_gate;
+	bool disable_dio_power_gate;
 	int dsc_min_slice_height_override;
 	int dsc_bpp_increment_div;
 	bool disable_pplib_wm_range;
@@ -607,6 +1015,8 @@ struct dc_debug_options {
 	int sr_enter_plus_exit_time_dpm0_ns;
 	int sr_exit_time_ns;
 	int sr_enter_plus_exit_time_ns;
+	int sr_exit_z8_time_ns;
+	int sr_enter_plus_exit_z8_time_ns;
 	int urgent_latency_ns;
 	uint32_t underflow_assert_delay_us;
 	int percent_of_ideal_drambw;
@@ -616,11 +1026,8 @@ struct dc_debug_options {
 	bool disable_pplib_clock_request;
 	bool disable_clock_gate;
 	bool disable_mem_low_power;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	bool pstate_enabled;
-#endif
 	bool disable_dmcu;
-	bool disable_psr;
 	bool force_abm_enable;
 	bool disable_stereo_support;
 	bool vsr_support;
@@ -633,22 +1040,21 @@ struct dc_debug_options {
 	bool hdmi20_disable;
 	bool skip_detection_link_training;
 	uint32_t edid_read_retry_times;
-	bool remove_disconnect_edp;
 	unsigned int force_odm_combine; //bit vector based on otg inst
-#if defined(CONFIG_DRM_AMD_DC_DCN)
+	unsigned int seamless_boot_odm_combine;
 	unsigned int force_odm_combine_4to1; //bit vector based on otg inst
+	int minimum_z8_residency_time;
+	int minimum_z10_residency_time;
 	bool disable_z9_mpc;
-#endif
 	unsigned int force_fclk_khz;
 	bool enable_tri_buf;
+	bool ips_disallow_entry;
 	bool dmub_offload_enabled;
 	bool dmcub_emulation;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	bool disable_idle_power_optimizations;
 	unsigned int mall_size_override;
 	unsigned int mall_additional_timer_percent;
 	bool mall_error_as_fatal;
-#endif
 	bool dmub_command_table; /* for testing only */
 	struct dc_bw_validation_profile bw_val_profile;
 	bool disable_fec;
@@ -657,9 +1063,7 @@ struct dc_debug_options {
 	 * watermarks are not affected.
 	 */
 	unsigned int force_min_dcfclk_mhz;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	int dwb_fi_phase;
-#endif
 	bool disable_timing_sync;
 	bool cm_in_bypass;
 	int force_clock_mode;/*every mode change.*/
@@ -669,85 +1073,129 @@ struct dc_debug_options {
 	bool enable_dmcub_surface_flip;
 	bool usbc_combo_phy_reset_wa;
 	bool enable_dram_clock_change_one_display_vactive;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	/* TODO - remove once tested */
 	bool legacy_dp2_lt;
 	bool set_mst_en_for_sst;
-#endif
+	bool disable_uhbr;
+	bool force_dp2_lt_fallback_method;
+	bool ignore_cable_id;
 	union mem_low_power_enable_options enable_mem_low_power;
 	union root_clock_optimization_options root_clock_optimization;
+	union fine_grain_clock_gating_enable_options enable_fine_grain_clock_gating;
 	bool hpo_optimization;
 	bool force_vblank_alignment;
 
 	/* Enable dmub aux for legacy ddc */
 	bool enable_dmub_aux_for_legacy_ddc;
-	bool optimize_edp_link_rate; /* eDP ILR */
+	bool disable_fams;
+	enum in_game_fams_config disable_fams_gaming;
 	/* FEC/PSR1 sequence enable delay in 100us */
 	uint8_t fec_enable_delay_in100us;
 	bool enable_driver_sequence_debug;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
+	enum det_size crb_alloc_policy;
+	int crb_alloc_policy_min_disp_count;
 	bool disable_z10;
-	bool enable_sw_cntl_psr;
+	bool enable_z9_disable_interface;
+	bool psr_skip_crtc_disable;
+	uint32_t ips_skip_crtc_disable_mask;
 	union dpia_debug_options dpia_debug;
-#endif
-};
-
-struct gpu_info_soc_bounding_box_v1_0;
-struct dc {
-	struct dc_debug_options debug;
-	struct dc_versions versions;
-	struct dc_caps caps;
-	struct dc_cap_funcs cap_funcs;
-	struct dc_config config;
-	struct dc_bounding_box_overrides bb_overrides;
-	struct dc_bug_wa work_arounds;
-	struct dc_context *ctx;
-	struct dc_phy_addr_space_config vm_pa_config;
-
-	uint8_t link_count;
-	struct dc_link *links[MAX_PIPES * 2];
-
-	struct dc_state *current_state;
-	struct resource_pool *res_pool;
-
-	struct clk_mgr *clk_mgr;
-
-	/* Display Engine Clock levels */
-	struct dm_pp_clock_levels sclk_lvls;
-
-	/* Inputs into BW and WM calculations. */
-	struct bw_calcs_dceip *bw_dceip;
-	struct bw_calcs_vbios *bw_vbios;
-#ifdef CONFIG_DRM_AMD_DC_DCN
-	struct dcn_soc_bounding_box *dcn_soc;
-	struct dcn_ip_params *dcn_ip;
-	struct display_mode_lib dml;
-#endif
-
-	/* HW functions */
-	struct hw_sequencer_funcs hwss;
-	struct dce_hwseq *hwseq;
-
-	/* Require to optimize clocks and bandwidth for added/removed planes */
-	bool optimized_required;
-	bool wm_optimized_required;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	bool idle_optimizations_allowed;
-#endif
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	bool enable_c20_dtm_b0;
-#endif
-
-	/* Require to maintain clocks and bandwidth for UEFI enabled HW */
-
-	/* FBC compressor */
-	struct compressor *fbc_compressor;
-
-	struct dc_debug_data debug_data;
-	struct dpcd_vendor_signature vendor_signature;
-
-	const char *build_id;
-	struct vm_helper *vm_helper;
+	bool disable_fixed_vs_aux_timeout_wa;
+	uint32_t fixed_vs_aux_delay_config_wa;
+	bool force_disable_subvp;
+	bool force_subvp_mclk_switch;
+	bool allow_sw_cursor_fallback;
+	unsigned int force_subvp_num_ways;
+	unsigned int force_mall_ss_num_ways;
+	bool alloc_extra_way_for_cursor;
+	uint32_t subvp_extra_lines;
+	bool disable_force_pstate_allow_on_hw_release;
+	bool force_usr_allow;
+	/* uses value at boot and disables switch */
+	bool disable_dtb_ref_clk_switch;
+	bool extended_blank_optimization;
+	union aux_wake_wa_options aux_wake_wa;
+	uint32_t mst_start_top_delay;
+	uint8_t psr_power_use_phy_fsm;
+	enum dml_hostvm_override_opts dml_hostvm_override;
+	bool dml_disallow_alternate_prefetch_modes;
+	bool use_legacy_soc_bb_mechanism;
+	bool exit_idle_opt_for_cursor_updates;
+	bool using_dml2;
+	bool enable_single_display_2to1_odm_policy;
+	bool enable_double_buffered_dsc_pg_support;
+	bool enable_dp_dig_pixel_rate_div_policy;
+	bool using_dml21;
+	enum lttpr_mode lttpr_mode_override;
+	unsigned int dsc_delay_factor_wa_x1000;
+	unsigned int min_prefetch_in_strobe_ns;
+	bool disable_unbounded_requesting;
+	bool dig_fifo_off_in_blank;
+	bool override_dispclk_programming;
+	bool otg_crc_db;
+	bool disallow_dispclk_dppclk_ds;
+	bool disable_fpo_optimizations;
+	bool support_eDP1_5;
+	uint32_t fpo_vactive_margin_us;
+	bool disable_fpo_vactive;
+	bool disable_boot_optimizations;
+	bool override_odm_optimization;
+	bool minimize_dispclk_using_odm;
+	bool disable_subvp_high_refresh;
+	bool disable_dp_plus_plus_wa;
+	uint32_t fpo_vactive_min_active_margin_us;
+	uint32_t fpo_vactive_max_blank_us;
+	bool enable_hpo_pg_support;
+	bool disable_dc_mode_overwrite;
+	bool replay_skip_crtc_disabled;
+	bool ignore_pg;/*do nothing, let pmfw control it*/
+	bool psp_disabled_wa;
+	unsigned int ips2_eval_delay_us;
+	unsigned int ips2_entry_delay_us;
+	bool optimize_ips_handshake;
+	bool disable_dmub_reallow_idle;
+	bool disable_timeout;
+	bool disable_extblankadj;
+	bool enable_idle_reg_checks;
+	unsigned int static_screen_wait_frames;
+	uint32_t pwm_freq;
+	bool force_chroma_subsampling_1tap;
+	unsigned int dcc_meta_propagation_delay_us;
+	bool disable_422_left_edge_pixel;
+	bool dml21_force_pstate_method;
+	uint32_t dml21_force_pstate_method_values[MAX_PIPES];
+	uint32_t dml21_disable_pstate_method_mask;
+	union fw_assisted_mclk_switch_version fams_version;
+	union dmub_fams2_global_feature_config fams2_config;
+	unsigned int force_cositing;
+	unsigned int disable_spl;
+	unsigned int force_easf;
+	unsigned int force_sharpness;
+	unsigned int force_sharpness_level;
+	unsigned int force_lls;
+	bool notify_dpia_hr_bw;
+	bool enable_ips_visual_confirm;
+	unsigned int sharpen_policy;
+	unsigned int scale_to_sharpness_policy;
+	unsigned int enable_oled_edp_power_up_opt;
+	bool enable_hblank_borrow;
+	bool force_subvp_df_throttle;
+	uint32_t acpi_transition_bitmasks[MAX_PIPES];
+	bool enable_pg_cntl_debug_logs;
+	unsigned int auxless_alpm_lfps_setup_ns;
+	unsigned int auxless_alpm_lfps_period_ns;
+	unsigned int auxless_alpm_lfps_silence_ns;
+	unsigned int auxless_alpm_lfps_t1t2_us;
+	short auxless_alpm_lfps_t1t2_offset_us;
+	bool disable_stutter_for_wm_program;
+	bool enable_block_sequence_programming;
+};
+
+
+/* Generic structure that can be used to query properties of DC. More fields
+ * can be added as required.
+ */
+struct dc_current_properties {
+	unsigned int cursor_size_limit;
 };
 
 enum frame_buffer_mode {
@@ -765,6 +1213,8 @@ struct dchub_init_data {
 	bool dchub_info_valid;
 };
 
+struct dml2_soc_bb;
+
 struct dc_init_data {
 	struct hw_asic_id asic_id;
 	void *driver; /* ctx */
@@ -786,17 +1236,22 @@ struct dc_init_data {
 	uint64_t log_mask;
 
 	struct dpcd_vendor_signature vendor_signature;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	bool force_smu_not_present;
-#endif
+	/*
+	 * IP offset for run time initializaion of register addresses
+	 *
+	 * DCN3.5+ will fail dc_create() if these fields are null for them. They are
+	 * applicable starting with DCN32/321 and are not used for ASICs upstreamed
+	 * before them.
+	 */
+	uint32_t *dcn_reg_offsets;
+	uint32_t *nbio_reg_offsets;
+	uint32_t *clk_reg_offsets;
+	void *bb_from_dmub;
 };
 
 struct dc_callback_init {
-#ifdef CONFIG_DRM_AMD_DC_HDCP
 	struct cp_psp cp_psp;
-#else
-	uint8_t reserved;
-#endif
 };
 
 struct dc *dc_create(const struct dc_init_data *init_params);
@@ -811,9 +1266,7 @@ void dc_init_callbacks(struct dc *dc,
 void dc_deinit_callbacks(struct dc *dc);
 void dc_destroy(struct dc **dc);
 
-/*******************************************************************************
- * Surface Interfaces
- ******************************************************************************/
+/* Surface Interfaces */
 
 enum {
 	TRANSFER_FUNC_POINTS = 1025
@@ -895,6 +1348,38 @@ union dc_3dlut_state {
 };
 
 
+#define MATRIX_9C__DIM_128_ALIGNED_LEN   16 // 9+8 :  9 * 8 +  7 * 8 = 72  + 56  = 128 % 128 = 0
+#define MATRIX_17C__DIM_128_ALIGNED_LEN  32 //17+15:  17 * 8 + 15 * 8 = 136 + 120 = 256 % 128 = 0
+#define MATRIX_33C__DIM_128_ALIGNED_LEN  64 //17+47:  17 * 8 + 47 * 8 = 136 + 376 = 512 % 128 = 0
+
+struct lut_rgb {
+	uint16_t b;
+	uint16_t g;
+	uint16_t r;
+	uint16_t padding;
+};
+
+//this structure maps directly to how the lut will read it from memory
+struct lut_mem_mapping {
+	union {
+		//NATIVE MODE 1, 2
+		//RGB layout          [b][g][r]      //red  is 128 byte aligned
+		//BGR layout          [r][g][b]      //blue is 128 byte aligned
+		struct lut_rgb rgb_17c[17][17][MATRIX_17C__DIM_128_ALIGNED_LEN];
+		struct lut_rgb rgb_33c[33][33][MATRIX_33C__DIM_128_ALIGNED_LEN];
+
+		//TRANSFORMED
+		uint16_t linear_rgb[(33*33*33*4/128+1)*128];
+	};
+	uint16_t size;
+};
+
+struct dc_rmcm_3dlut {
+	bool isInUse;
+	const struct dc_stream_state *stream;
+	uint8_t protection_bits;
+};
+
 struct dc_3dlut {
 	struct kref refcount;
 	struct tetrahedral_params lut_3d;
@@ -931,7 +1416,6 @@ union surface_update_flags {
 		uint32_t in_transfer_func_change:1;
 		uint32_t input_csc_change:1;
 		uint32_t coeff_reduction_change:1;
-		uint32_t output_tf_change:1;
 		uint32_t pixel_format_change:1;
 		uint32_t plane_size_change:1;
 		uint32_t gamut_remap_change:1;
@@ -944,12 +1428,17 @@ union surface_update_flags {
 		uint32_t clock_change:1;
 		uint32_t stereo_format_change:1;
 		uint32_t lut_3d:1;
+		uint32_t tmz_changed:1;
+		uint32_t mcm_transfer_function_enable_change:1; /* disable or enable MCM transfer func */
 		uint32_t full_update:1;
+		uint32_t sdr_white_level_nits:1;
 	} bits;
 
 	uint32_t raw;
 };
 
+#define DC_REMOVE_PLANE_POINTERS 1
+
 struct dc_plane_state {
 	struct dc_plane_address address;
 	struct dc_plane_flip_time time;
@@ -960,13 +1449,13 @@ struct dc_plane_state {
 	struct rect clip_rect;
 
 	struct plane_size plane_size;
-	union dc_tiling_info tiling_info;
+	struct dc_tiling_info tiling_info;
 
 	struct dc_plane_dcc_param dcc;
 
-	struct dc_gamma *gamma_correction;
-	struct dc_transfer_func *in_transfer_func;
-	struct dc_bias_and_scale *bias_and_scale;
+	struct dc_gamma gamma_correction;
+	struct dc_transfer_func in_transfer_func;
+	struct dc_bias_and_scale bias_and_scale;
 	struct dc_csc_transform input_csc_color_matrix;
 	struct fixed31_32 coeff_reduction_factor;
 	struct fixed31_32 hdr_mult;
@@ -977,19 +1466,18 @@ struct dc_plane_state {
 
 	enum dc_color_space color_space;
 
-	struct dc_3dlut *lut3d_func;
-	struct dc_transfer_func *in_shaper_func;
-	struct dc_transfer_func *blend_tf;
+	struct dc_3dlut lut3d_func;
+	struct dc_transfer_func in_shaper_func;
+	struct dc_transfer_func blend_tf;
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	struct dc_transfer_func *gamcor_tf;
-#endif
 	enum surface_pixel_format format;
 	enum dc_rotation_angle rotation;
 	enum plane_stereo_format stereo_format;
 
 	bool is_tiling_rotated;
 	bool per_pixel_alpha;
+	bool pre_multiplied_alpha;
 	bool global_alpha;
 	int  global_alpha_value;
 	bool visible;
@@ -1008,14 +1496,33 @@ struct dc_plane_state {
 	/* HACK: Workaround for forcing full reprogramming under some conditions */
 	bool force_full_update;
 
+	bool is_phantom; // TODO: Change mall_stream_config into mall_plane_config instead
+
 	/* private to dc_surface.c */
 	enum dc_irq_source irq_source;
 	struct kref refcount;
+	struct tg_color visual_confirm_color;
+
+	bool is_statically_allocated;
+	enum chroma_cositing cositing;
+	enum dc_cm2_shaper_3dlut_setting mcm_shaper_3dlut_setting;
+	bool mcm_lut1d_enable;
+	struct dc_cm2_func_luts mcm_luts;
+	bool lut_bank_a;
+	enum mpcc_movable_cm_location mcm_location;
+	struct dc_csc_transform cursor_csc_color_matrix;
+	bool adaptive_sharpness_en;
+	int adaptive_sharpness_policy;
+	int sharpness_level;
+	enum linear_light_scaling linear_light_scaling;
+	unsigned int sdr_white_level_nits;
+	struct spl_sharpness_range sharpness_range;
+	enum sharpness_range_source sharpness_source;
 };
 
 struct dc_plane_info {
 	struct plane_size plane_size;
-	union dc_tiling_info tiling_info;
+	struct dc_tiling_info tiling_info;
 	struct dc_plane_dcc_param dcc;
 	enum surface_pixel_format format;
 	enum dc_rotation_angle rotation;
@@ -1024,10 +1531,277 @@ struct dc_plane_info {
 	bool horizontal_mirror;
 	bool visible;
 	bool per_pixel_alpha;
+	bool pre_multiplied_alpha;
 	bool global_alpha;
 	int  global_alpha_value;
 	bool input_csc_enabled;
 	int layer_index;
+	enum chroma_cositing cositing;
+};
+
+#include "dc_stream.h"
+
+struct dc_scratch_space {
+	/* used to temporarily backup plane states of a stream during
+	 * dc update. The reason is that plane states are overwritten
+	 * with surface updates in dc update. Once they are overwritten
+	 * current state is no longer valid. We want to temporarily
+	 * store current value in plane states so we can still recover
+	 * a valid current state during dc update.
+	 */
+	struct dc_plane_state plane_states[MAX_SURFACES];
+
+	struct dc_stream_state stream_state;
+};
+
+/*
+ * A link contains one or more sinks and their connected status.
+ * The currently active signal type (HDMI, DP-SST, DP-MST) is also reported.
+ */
+ struct dc_link {
+	struct dc_sink *remote_sinks[MAX_SINKS_PER_LINK];
+	unsigned int sink_count;
+	struct dc_sink *local_sink;
+	unsigned int link_index;
+	enum dc_connection_type type;
+	enum signal_type connector_signal;
+	enum dc_irq_source irq_source_hpd;
+	enum dc_irq_source irq_source_hpd_rx;/* aka DP Short Pulse  */
+	enum dc_irq_source irq_source_read_request;/* Read Request */
+
+	bool is_hpd_filter_disabled;
+	bool dp_ss_off;
+
+	/**
+	 * @link_state_valid:
+	 *
+	 * If there is no link and local sink, this variable should be set to
+	 * false. Otherwise, it should be set to true; usually, the function
+	 * core_link_enable_stream sets this field to true.
+	 */
+	bool link_state_valid;
+	bool aux_access_disabled;
+	bool sync_lt_in_progress;
+	bool skip_stream_reenable;
+	bool is_internal_display;
+	/** @todo Rename. Flag an endpoint as having a programmable mapping to a DIG encoder. */
+	bool is_dig_mapping_flexible;
+	bool hpd_status; /* HPD status of link without physical HPD pin. */
+	bool is_hpd_pending; /* Indicates a new received hpd */
+
+	/* USB4 DPIA links skip verifying link cap, instead performing the fallback method
+	 * for every link training. This is incompatible with DP LL compliance automation,
+	 * which expects the same link settings to be used every retry on a link loss.
+	 * This flag is used to skip the fallback when link loss occurs during automation.
+	 */
+	bool skip_fallback_on_link_loss;
+
+	bool edp_sink_present;
+
+	struct dp_trace dp_trace;
+
+	/* caps is the same as reported_link_cap. link_traing use
+	 * reported_link_cap. Will clean up.  TODO
+	 */
+	struct dc_link_settings reported_link_cap;
+	struct dc_link_settings verified_link_cap;
+	struct dc_link_settings cur_link_settings;
+	struct dc_lane_settings cur_lane_setting[LANE_COUNT_DP_MAX];
+	struct dc_link_settings preferred_link_setting;
+	/* preferred_training_settings are override values that
+	 * come from DM. DM is responsible for the memory
+	 * management of the override pointers.
+	 */
+	struct dc_link_training_overrides preferred_training_settings;
+	struct dp_audio_test_data audio_test_data;
+
+	uint8_t ddc_hw_inst;
+
+	uint8_t hpd_src;
+
+	uint8_t link_enc_hw_inst;
+	/* DIG link encoder ID. Used as index in link encoder resource pool.
+	 * For links with fixed mapping to DIG, this is not changed after dc_link
+	 * object creation.
+	 */
+	enum engine_id eng_id;
+	enum engine_id dpia_preferred_eng_id;
+
+	bool test_pattern_enabled;
+	/* Pending/Current test pattern are only used to perform and track
+	 * FIXED_VS retimer test pattern/lane adjustment override state.
+	 * Pending allows link HWSS to differentiate PHY vs non-PHY pattern,
+	 * to perform specific lane adjust overrides before setting certain
+	 * PHY test patterns. In cases when lane adjust and set test pattern
+	 * calls are not performed atomically (i.e. performing link training),
+	 * pending_test_pattern will be invalid or contain a non-PHY test pattern
+	 * and current_test_pattern will contain required context for any future
+	 * set pattern/set lane adjust to transition between override state(s).
+	 * */
+	enum dp_test_pattern current_test_pattern;
+	enum dp_test_pattern pending_test_pattern;
+
+	union compliance_test_state compliance_test_state;
+
+	void *priv;
+
+	struct ddc_service *ddc;
+
+	enum dp_panel_mode panel_mode;
+	bool aux_mode;
+
+	/* Private to DC core */
+
+	const struct dc *dc;
+
+	struct dc_context *ctx;
+
+	struct panel_cntl *panel_cntl;
+	struct link_encoder *link_enc;
+	struct graphics_object_id link_id;
+	/* Endpoint type distinguishes display endpoints which do not have entries
+	 * in the BIOS connector table from those that do. Helps when tracking link
+	 * encoder to display endpoint assignments.
+	 */
+	enum display_endpoint_type ep_type;
+	union ddi_channel_mapping ddi_channel_mapping;
+	struct connector_device_tag_info device_tag;
+	struct dpcd_caps dpcd_caps;
+	uint32_t dongle_max_pix_clk;
+	unsigned short chip_caps;
+	unsigned int dpcd_sink_count;
+	struct hdcp_caps hdcp_caps;
+	enum edp_revision edp_revision;
+	union dpcd_sink_ext_caps dpcd_sink_ext_caps;
+
+	struct psr_settings psr_settings;
+	struct replay_settings replay_settings;
+
+	/* Drive settings read from integrated info table */
+	struct dc_lane_settings bios_forced_drive_settings;
+
+	/* Vendor specific LTTPR workaround variables */
+	uint8_t vendor_specific_lttpr_link_rate_wa;
+	bool apply_vendor_specific_lttpr_link_rate_wa;
+
+	/* MST record stream using this link */
+	struct link_flags {
+		bool dp_keep_receiver_powered;
+		bool dp_skip_DID2;
+		bool dp_skip_reset_segment;
+		bool dp_skip_fs_144hz;
+		bool dp_mot_reset_segment;
+		/* Some USB4 docks do not handle turning off MST DSC once it has been enabled. */
+		bool dpia_mst_dsc_always_on;
+		/* Forced DPIA into TBT3 compatibility mode. */
+		bool dpia_forced_tbt3_mode;
+		bool dongle_mode_timing_override;
+		bool blank_stream_on_ocs_change;
+		bool read_dpcd204h_on_irq_hpd;
+		bool force_dp_ffe_preset;
+		bool skip_phy_ssc_reduction;
+	} wa_flags;
+	union dc_dp_ffe_preset forced_dp_ffe_preset;
+	struct link_mst_stream_allocation_table mst_stream_alloc_table;
+
+	struct dc_link_status link_status;
+	struct dprx_states dprx_states;
+
+	struct gpio *hpd_gpio;
+	enum dc_link_fec_state fec_state;
+	bool is_dds;
+	bool is_display_mux_present;
+	bool link_powered_externally;	// Used to bypass hardware sequencing delays when panel is powered down forcibly
+
+	struct dc_panel_config panel_config;
+	struct phy_state phy_state;
+	uint32_t phy_transition_bitmask;
+	// BW ALLOCATON USB4 ONLY
+	struct dc_dpia_bw_alloc dpia_bw_alloc_config;
+	bool skip_implict_edp_power_control;
+	enum backlight_control_type backlight_control_type;
+};
+
+struct dc {
+	struct dc_debug_options debug;
+	struct dc_versions versions;
+	struct dc_caps caps;
+	struct dc_check_config check_config;
+	struct dc_cap_funcs cap_funcs;
+	struct dc_config config;
+	struct dc_bounding_box_overrides bb_overrides;
+	struct dc_bug_wa work_arounds;
+	struct dc_context *ctx;
+	struct dc_phy_addr_space_config vm_pa_config;
+
+	uint8_t link_count;
+	struct dc_link *links[MAX_LINKS];
+	uint8_t lowest_dpia_link_index;
+	struct link_service *link_srv;
+
+	struct dc_state *current_state;
+	struct resource_pool *res_pool;
+
+	struct clk_mgr *clk_mgr;
+
+	/* Display Engine Clock levels */
+	struct dm_pp_clock_levels sclk_lvls;
+
+	/* Inputs into BW and WM calculations. */
+	struct bw_calcs_dceip *bw_dceip;
+	struct bw_calcs_vbios *bw_vbios;
+	struct dcn_soc_bounding_box *dcn_soc;
+	struct dcn_ip_params *dcn_ip;
+	struct display_mode_lib dml;
+
+	/* HW functions */
+	struct hw_sequencer_funcs hwss;
+	struct dce_hwseq *hwseq;
+
+	/* Require to optimize clocks and bandwidth for added/removed planes */
+	bool optimized_required;
+	bool idle_optimizations_allowed;
+	bool enable_c20_dtm_b0;
+
+	/* Require to maintain clocks and bandwidth for UEFI enabled HW */
+
+	/* For eDP to know the switching state of SmartMux */
+	bool is_switch_in_progress_orig;
+	bool is_switch_in_progress_dest;
+
+	/* FBC compressor */
+	struct compressor *fbc_compressor;
+
+	struct dc_debug_data debug_data;
+	struct dpcd_vendor_signature vendor_signature;
+
+	const char *build_id;
+	struct vm_helper *vm_helper;
+
+	uint32_t *dcn_reg_offsets;
+	uint32_t *nbio_reg_offsets;
+	uint32_t *clk_reg_offsets;
+
+	/* Scratch memory */
+	struct {
+		struct {
+			/*
+			 * For matching clock_limits table in driver with table
+			 * from PMFW.
+			 */
+			struct _vcs_dpi_voltage_scaling_st clock_limits[DC__VOLTAGE_STATES];
+		} update_bw_bounding_box;
+		struct dc_scratch_space current_state;
+		struct dc_scratch_space new_state;
+		struct dc_stream_state temp_stream; // Used so we don't need to allocate stream on the stack
+		struct dc_link temp_link;
+		bool pipes_to_unlock_first[MAX_PIPES]; /* Any of the pipes indicated here should be unlocked first */
+	} scratch;
+
+	struct dml2_configuration_options dml2_options;
+	struct dml2_configuration_options dml2_dc_power_options;
+	enum dc_acpi_cm_power_state power_state;
+	struct soc_and_ip_translator *soc_and_ip_translator;
 };
 
 struct dc_scaling_info {
@@ -1037,6 +1811,17 @@ struct dc_scaling_info {
 	struct scaling_taps scaling_quality;
 };
 
+struct dc_fast_update {
+	const struct dc_flip_addrs *flip_addr;
+	const struct dc_gamma *gamma;
+	const struct colorspace_transform *gamut_remap_matrix;
+	const struct dc_csc_transform *input_csc_color_matrix;
+	const struct fixed31_32 *coeff_reduction_factor;
+	struct dc_transfer_func *out_transfer_func;
+	struct dc_csc_transform *output_csc_transform;
+	const struct dc_csc_transform *cursor_csc_color_matrix;
+};
+
 struct dc_surface_update {
 	struct dc_plane_state *surface;
 
@@ -1057,18 +1842,44 @@ struct dc_surface_update {
 	const struct dc_3dlut *lut3d_func;
 	const struct dc_transfer_func *blend_tf;
 	const struct colorspace_transform *gamut_remap_matrix;
+	/*
+	 * Color Transformations for pre-blend MCM (Shaper, 3DLUT, 1DLUT)
+	 *
+	 * change cm2_params.component_settings: Full update
+	 * change cm2_params.cm2_luts: Fast update
+	 */
+	const struct dc_cm2_parameters *cm2_params;
+	const struct dc_csc_transform *cursor_csc_color_matrix;
+	unsigned int sdr_white_level_nits;
+	struct dc_bias_and_scale bias_and_scale;
+};
+
+struct dc_underflow_debug_data {
+	struct dcn_hubbub_reg_state *hubbub_reg_state;
+	struct dcn_hubp_reg_state *hubp_reg_state[MAX_PIPES];
+	struct dcn_dpp_reg_state *dpp_reg_state[MAX_PIPES];
+	struct dcn_mpc_reg_state *mpc_reg_state[MAX_PIPES];
+	struct dcn_opp_reg_state *opp_reg_state[MAX_PIPES];
+	struct dcn_dsc_reg_state *dsc_reg_state[MAX_PIPES];
+	struct dcn_optc_reg_state *optc_reg_state[MAX_PIPES];
+	struct dcn_dccg_reg_state *dccg_reg_state[MAX_PIPES];
+};
+
+struct power_features {
+	bool ips;
+	bool rcg;
+	bool replay;
+	bool dds;
+	bool sprs;
+	bool psr;
+	bool fams;
+	bool mpo;
+	bool uclk_p_state;
 };
 
 /*
  * Create a new surface with default parameters;
  */
-struct dc_plane_state *dc_create_plane_state(struct dc *dc);
-const struct dc_plane_status *dc_plane_get_status(
-		const struct dc_plane_state *plane_state);
-
-void dc_plane_state_retain(struct dc_plane_state *plane_state);
-void dc_plane_state_release(struct dc_plane_state *plane_state);
-
 void dc_gamma_retain(struct dc_gamma *dc_gamma);
 void dc_gamma_release(struct dc_gamma **dc_gamma);
 struct dc_gamma *dc_create_gamma(void);
@@ -1080,80 +1891,66 @@ struct dc_transfer_func *dc_create_transfer_func(void);
 struct dc_3dlut *dc_create_3dlut_func(void);
 void dc_3dlut_func_release(struct dc_3dlut *lut);
 void dc_3dlut_func_retain(struct dc_3dlut *lut);
-/*
- * This structure holds a surface address.  There could be multiple addresses
- * in cases such as Stereo 3D, Planar YUV, etc.  Other per-flip attributes such
- * as frame durations and DCC format can also be set.
- */
-struct dc_flip_addrs {
-	struct dc_plane_address address;
-	unsigned int flip_timestamp_in_us;
-	bool flip_immediate;
-	/* TODO: add flip duration for FreeSync */
-	bool triplebuffer_flips;
-};
 
 void dc_post_update_surfaces_to_stream(
 		struct dc *dc);
 
-#include "dc_stream.h"
-
-/*
- * Structure to store surface/stream associations for validation
+/**
+ * struct dc_validation_set - Struct to store surface/stream associations for validation
  */
 struct dc_validation_set {
+	/**
+	 * @stream: Stream state properties
+	 */
 	struct dc_stream_state *stream;
+
+	/**
+	 * @plane_states: Surface state
+	 */
 	struct dc_plane_state *plane_states[MAX_SURFACES];
+
+	/**
+	 * @plane_count: Total of active planes
+	 */
 	uint8_t plane_count;
 };
 
-bool dc_validate_seamless_boot_timing(const struct dc *dc,
+bool dc_validate_boot_timing(const struct dc *dc,
 				const struct dc_sink *sink,
 				struct dc_crtc_timing *crtc_timing);
 
 enum dc_status dc_validate_plane(struct dc *dc, const struct dc_plane_state *plane_state);
 
-void get_clock_requirements_for_state(struct dc_state *state, struct AsicStateEx *info);
+enum dc_status dc_validate_with_context(struct dc *dc,
+					const struct dc_validation_set set[],
+					int set_count,
+					struct dc_state *context,
+					enum dc_validate_mode validate_mode);
 
 bool dc_set_generic_gpio_for_stereo(bool enable,
 		struct gpio_service *gpio_service);
 
-/*
- * fast_validate: we return after determining if we can support the new state,
- * but before we populate the programming info
- */
 enum dc_status dc_validate_global_state(
 		struct dc *dc,
 		struct dc_state *new_ctx,
-		bool fast_validate);
+		enum dc_validate_mode validate_mode);
 
-
-void dc_resource_state_construct(
-		const struct dc *dc,
-		struct dc_state *dst_ctx);
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 bool dc_acquire_release_mpc_3dlut(
 		struct dc *dc, bool acquire,
 		struct dc_stream_state *stream,
 		struct dc_3dlut **lut,
 		struct dc_transfer_func **shaper);
-#endif
-
-void dc_resource_state_copy_construct(
-		const struct dc_state *src_ctx,
-		struct dc_state *dst_ctx);
-
-void dc_resource_state_copy_construct_current(
-		const struct dc *dc,
-		struct dc_state *dst_ctx);
-
-void dc_resource_state_destruct(struct dc_state *context);
 
 bool dc_resource_is_dsc_encoding_supported(const struct dc *dc);
-
+void get_audio_check(struct audio_info *aud_modes,
+	struct audio_check *aud_chk);
+
+bool fast_nonaddr_updates_exist(struct dc_fast_update *fast_update, int surface_count);
+void populate_fast_updates(struct dc_fast_update *fast_update,
+		struct dc_surface_update *srf_updates,
+		int surface_count,
+		struct dc_stream_update *stream_update);
 /*
- * TODO update to make it about validation sets
  * Set up streams and links associated to drive sinks
  * The streams parameter is an absolute set of all active streams.
  *
@@ -1161,118 +1958,591 @@ bool dc_resource_is_dsc_encoding_supported(const struct dc *dc);
  *   Phy, Encoder, Timing Generator are programmed and enabled.
  *   New streams are enabled with blank stream; no memory read.
  */
-bool dc_commit_state(struct dc *dc, struct dc_state *context);
-
-struct dc_state *dc_create_state(struct dc *dc);
-struct dc_state *dc_copy_state(struct dc_state *src_ctx);
-void dc_retain_state(struct dc_state *context);
-void dc_release_state(struct dc_state *context);
-
-/*******************************************************************************
- * Link Interfaces
- ******************************************************************************/
-
-struct dpcd_caps {
-	union dpcd_rev dpcd_rev;
-	union max_lane_count max_ln_count;
-	union max_down_spread max_down_spread;
-	union dprx_feature dprx_feature;
-
-	/* valid only for eDP v1.4 or higher*/
-	uint8_t edp_supported_link_rates_count;
-	enum dc_link_rate edp_supported_link_rates[8];
-
-	/* dongle type (DP converter, CV smart dongle) */
-	enum display_dongle_type dongle_type;
-	/* branch device or sink device */
-	bool is_branch_dev;
-	/* Dongle's downstream count. */
-	union sink_count sink_count;
-	/* If dongle_type == DISPLAY_DONGLE_DP_HDMI_CONVERTER,
-	indicates 'Frame Sequential-to-lllFrame Pack' conversion capability.*/
-	struct dc_dongle_caps dongle_caps;
-
-	uint32_t sink_dev_id;
-	int8_t sink_dev_id_str[6];
-	int8_t sink_hw_revision;
-	int8_t sink_fw_revision[2];
-
-	uint32_t branch_dev_id;
-	int8_t branch_dev_name[6];
-	int8_t branch_hw_revision;
-	int8_t branch_fw_revision[2];
-
-	bool allow_invalid_MSA_timing_param;
-	bool panel_mode_edp;
-	bool dpcd_display_control_capable;
-	bool ext_receiver_cap_field_present;
-	bool dynamic_backlight_capable_edp;
-	union dpcd_fec_capability fec_cap;
-	struct dpcd_dsc_capabilities dsc_caps;
-	struct dc_lttpr_caps lttpr_caps;
-	struct psr_caps psr_caps;
-	struct dpcd_usb4_dp_tunneling_info usb4_dp_tun_info;
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	union dp_128b_132b_supported_link_rates dp_128b_132b_supported_link_rates;
-	union dp_main_line_channel_coding_cap channel_coding_cap;
-	union dp_sink_video_fallback_formats fallback_formats;
-	union dp_fec_capability1 fec_cap1;
-#endif
-};
-
-union dpcd_sink_ext_caps {
-	struct {
-		/* 0 - Sink supports backlight adjust via PWM during SDR/HDR mode
-		 * 1 - Sink supports backlight adjust via AUX during SDR/HDR mode.
-		 */
-		uint8_t sdr_aux_backlight_control : 1;
-		uint8_t hdr_aux_backlight_control : 1;
-		uint8_t reserved_1 : 2;
-		uint8_t oled : 1;
-		uint8_t reserved : 3;
-	} bits;
-	uint8_t raw;
-};
+enum dc_status dc_commit_streams(struct dc *dc, struct dc_commit_streams_params *params);
 
-#if defined(CONFIG_DRM_AMD_DC_HDCP)
-union hdcp_rx_caps {
-	struct {
-		uint8_t version;
-		uint8_t reserved;
-		struct {
-			uint8_t repeater	: 1;
-			uint8_t hdcp_capable	: 1;
-			uint8_t reserved	: 6;
-		} byte0;
-	} fields;
-	uint8_t raw[3];
-};
 
-union hdcp_bcaps {
-	struct {
-		uint8_t HDCP_CAPABLE:1;
-		uint8_t REPEATER:1;
-		uint8_t RESERVED:6;
-	} bits;
-	uint8_t raw;
-};
+struct dc_plane_state *dc_get_surface_for_mpcc(struct dc *dc,
+		struct dc_stream_state *stream,
+		int mpcc_inst);
+
+
+uint32_t dc_get_opp_for_plane(struct dc *dc, struct dc_plane_state *plane);
+
+void dc_set_disable_128b_132b_stream_overhead(bool disable);
 
-struct hdcp_caps {
-	union hdcp_rx_caps rx_caps;
-	union hdcp_bcaps bcaps;
+/* The function returns minimum bandwidth required to drive a given timing
+ * return - minimum required timing bandwidth in kbps.
+ */
+uint32_t dc_bandwidth_in_kbps_from_timing(
+		const struct dc_crtc_timing *timing,
+		const enum dc_link_encoding_format link_encoding);
+
+/* Link Interfaces */
+/* Return an enumerated dc_link.
+ * dc_link order is constant and determined at
+ * boot time.  They cannot be created or destroyed.
+ * Use dc_get_caps() to get number of links.
+ */
+struct dc_link *dc_get_link_at_index(struct dc *dc, uint32_t link_index);
+
+/* Return instance id of the edp link. Inst 0 is primary edp link. */
+bool dc_get_edp_link_panel_inst(const struct dc *dc,
+		const struct dc_link *link,
+		unsigned int *inst_out);
+
+/* Return an array of link pointers to edp links. */
+void dc_get_edp_links(const struct dc *dc,
+		struct dc_link **edp_links,
+		int *edp_num);
+
+void dc_set_edp_power(const struct dc *dc, struct dc_link *edp_link,
+				 bool powerOn);
+
+/* The function initiates detection handshake over the given link. It first
+ * determines if there are display connections over the link. If so it initiates
+ * detection protocols supported by the connected receiver device. The function
+ * contains protocol specific handshake sequences which are sometimes mandatory
+ * to establish a proper connection between TX and RX. So it is always
+ * recommended to call this function as the first link operation upon HPD event
+ * or power up event. Upon completion, the function will update link structure
+ * in place based on latest RX capabilities. The function may also cause dpms
+ * to be reset to off for all currently enabled streams to the link. It is DM's
+ * responsibility to serialize detection and DPMS updates.
+ *
+ * @reason - Indicate which event triggers this detection. dc may customize
+ * detection flow depending on the triggering events.
+ * return false - if detection is not fully completed. This could happen when
+ * there is an unrecoverable error during detection or detection is partially
+ * completed (detection has been delegated to dm mst manager ie.
+ * link->connection_type == dc_connection_mst_branch when returning false).
+ * return true - detection is completed, link has been fully updated with latest
+ * detection result.
+ */
+bool dc_link_detect(struct dc_link *link, enum dc_detect_reason reason);
+
+struct dc_sink_init_data;
+
+/* When link connection type is dc_connection_mst_branch, remote sink can be
+ * added to the link. The interface creates a remote sink and associates it with
+ * current link. The sink will be retained by link until remove remote sink is
+ * called.
+ *
+ * @dc_link - link the remote sink will be added to.
+ * @edid - byte array of EDID raw data.
+ * @len - size of the edid in byte
+ * @init_data -
+ */
+struct dc_sink *dc_link_add_remote_sink(
+		struct dc_link *dc_link,
+		const uint8_t *edid,
+		int len,
+		struct dc_sink_init_data *init_data);
+
+/* Remove remote sink from a link with dc_connection_mst_branch connection type.
+ * @link - link the sink should be removed from
+ * @sink - sink to be removed.
+ */
+void dc_link_remove_remote_sink(
+	struct dc_link *link,
+	struct dc_sink *sink);
+
+/* Enable HPD interrupt handler for a given link */
+void dc_link_enable_hpd(const struct dc_link *link);
+
+/* Disable HPD interrupt handler for a given link */
+void dc_link_disable_hpd(const struct dc_link *link);
+
+/* determine if there is a sink connected to the link
+ *
+ * @type - dc_connection_single if connected, dc_connection_none otherwise.
+ * return - false if an unexpected error occurs, true otherwise.
+ *
+ * NOTE: This function doesn't detect downstream sink connections i.e
+ * dc_connection_mst_branch, dc_connection_sst_branch. In this case, it will
+ * return dc_connection_single if the branch device is connected despite of
+ * downstream sink's connection status.
+ */
+bool dc_link_detect_connection_type(struct dc_link *link,
+		enum dc_connection_type *type);
+
+/* query current hpd pin value
+ * return - true HPD is asserted (HPD high), false otherwise (HPD low)
+ *
+ */
+bool dc_link_get_hpd_state(struct dc_link *link);
+
+/* Getter for cached link status from given link */
+const struct dc_link_status *dc_link_get_status(const struct dc_link *link);
+
+/* enable/disable hardware HPD filter.
+ *
+ * @link - The link the HPD pin is associated with.
+ * @enable = true - enable hardware HPD filter. HPD event will only queued to irq
+ * handler once after no HPD change has been detected within dc default HPD
+ * filtering interval since last HPD event. i.e if display keeps toggling hpd
+ * pulses within default HPD interval, no HPD event will be received until HPD
+ * toggles have stopped. Then HPD event will be queued to irq handler once after
+ * dc default HPD filtering interval since last HPD event.
+ *
+ * @enable = false - disable hardware HPD filter. HPD event will be queued
+ * immediately to irq handler after no HPD change has been detected within
+ * IRQ_HPD (aka HPD short pulse) interval (i.e 2ms).
+ */
+void dc_link_enable_hpd_filter(struct dc_link *link, bool enable);
+
+/* submit i2c read/write payloads through ddc channel
+ * @link_index - index to a link with ddc in i2c mode
+ * @cmd - i2c command structure
+ * return - true if success, false otherwise.
+ */
+bool dc_submit_i2c(
+		struct dc *dc,
+		uint32_t link_index,
+		struct i2c_command *cmd);
+
+/* submit i2c read/write payloads through oem channel
+ * @link_index - index to a link with ddc in i2c mode
+ * @cmd - i2c command structure
+ * return - true if success, false otherwise.
+ */
+bool dc_submit_i2c_oem(
+		struct dc *dc,
+		struct i2c_command *cmd);
+
+enum aux_return_code_type;
+/* Attempt to transfer the given aux payload. This function does not perform
+ * retries or handle error states. The reply is returned in the payload->reply
+ * and the result through operation_result. Returns the number of bytes
+ * transferred,or -1 on a failure.
+ */
+int dc_link_aux_transfer_raw(struct ddc_service *ddc,
+		struct aux_payload *payload,
+		enum aux_return_code_type *operation_result);
+
+struct ddc_service *
+dc_get_oem_i2c_device(struct dc *dc);
+
+bool dc_is_oem_i2c_device_present(
+	struct dc *dc,
+	size_t slave_address
+);
+
+/* return true if the connected receiver supports the hdcp version */
+bool dc_link_is_hdcp14(struct dc_link *link, enum signal_type signal);
+bool dc_link_is_hdcp22(struct dc_link *link, enum signal_type signal);
+
+/* Notify DC about DP RX Interrupt (aka DP IRQ_HPD).
+ *
+ * TODO - When defer_handling is true the function will have a different purpose.
+ * It no longer does complete hpd rx irq handling. We should create a separate
+ * interface specifically for this case.
+ *
+ * Return:
+ * true - Downstream port status changed. DM should call DC to do the
+ * detection.
+ * false - no change in Downstream port status. No further action required
+ * from DM.
+ */
+bool dc_link_handle_hpd_rx_irq(struct dc_link *dc_link,
+		union hpd_irq_data *hpd_irq_dpcd_data, bool *out_link_loss,
+		bool defer_handling, bool *has_left_work);
+/* handle DP specs define test automation sequence*/
+void dc_link_dp_handle_automated_test(struct dc_link *link);
+
+/* handle DP Link loss sequence and try to recover RX link loss with best
+ * effort
+ */
+void dc_link_dp_handle_link_loss(struct dc_link *link);
+
+/* Determine if hpd rx irq should be handled or ignored
+ * return true - hpd rx irq should be handled.
+ * return false - it is safe to ignore hpd rx irq event
+ */
+bool dc_link_dp_allow_hpd_rx_irq(const struct dc_link *link);
+
+/* Determine if link loss is indicated with a given hpd_irq_dpcd_data.
+ * @link - link the hpd irq data associated with
+ * @hpd_irq_dpcd_data - input hpd irq data
+ * return - true if hpd irq data indicates a link lost
+ */
+bool dc_link_check_link_loss_status(struct dc_link *link,
+		union hpd_irq_data *hpd_irq_dpcd_data);
+
+/* Read hpd rx irq data from a given link
+ * @link - link where the hpd irq data should be read from
+ * @irq_data - output hpd irq data
+ * return - DC_OK if hpd irq data is read successfully, otherwise hpd irq data
+ * read has failed.
+ */
+enum dc_status dc_link_dp_read_hpd_rx_irq_data(
+	struct dc_link *link,
+	union hpd_irq_data *irq_data);
+
+/* The function clears recorded DP RX states in the link. DM should call this
+ * function when it is resuming from S3 power state to previously connected links.
+ *
+ * TODO - in the future we should consider to expand link resume interface to
+ * support clearing previous rx states. So we don't have to rely on dm to call
+ * this interface explicitly.
+ */
+void dc_link_clear_dprx_states(struct dc_link *link);
+
+/* Destruct the mst topology of the link and reset the allocated payload table
+ *
+ * NOTE: this should only be called if DM chooses not to call dc_link_detect but
+ * still wants to reset MST topology on an unplug event */
+bool dc_link_reset_cur_dp_mst_topology(struct dc_link *link);
+
+/* The function calculates effective DP link bandwidth when a given link is
+ * using the given link settings.
+ *
+ * return - total effective link bandwidth in kbps.
+ */
+uint32_t dc_link_bandwidth_kbps(
+	const struct dc_link *link,
+	const struct dc_link_settings *link_setting);
+
+struct dp_audio_bandwidth_params {
+	const struct dc_crtc_timing *crtc_timing;
+	enum dp_link_encoding link_encoding;
+	uint32_t channel_count;
+	uint32_t sample_rate_hz;
 };
-#endif
 
-#include "dc_link.h"
+/* The function calculates the minimum size of hblank (in bytes) needed to
+ * support the specified channel count and sample rate combination, given the
+ * link encoding and timing to be used. This calculation is not supported
+ * for 8b/10b SST.
+ *
+ * return - min hblank size in bytes, 0 if 8b/10b SST.
+ */
+uint32_t dc_link_required_hblank_size_bytes(
+	const struct dc_link *link,
+	struct dp_audio_bandwidth_params *audio_params);
+
+/* The function takes a snapshot of current link resource allocation state
+ * @dc: pointer to dc of the dm calling this
+ * @map: a dc link resource snapshot defined internally to dc.
+ *
+ * DM needs to capture a snapshot of current link resource allocation mapping
+ * and store it in its persistent storage.
+ *
+ * Some of the link resource is using first come first serve policy.
+ * The allocation mapping depends on original hotplug order. This information
+ * is lost after driver is loaded next time. The snapshot is used in order to
+ * restore link resource to its previous state so user will get consistent
+ * link capability allocation across reboot.
+ *
+ */
+void dc_get_cur_link_res_map(const struct dc *dc, uint32_t *map);
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-uint32_t dc_get_opp_for_plane(struct dc *dc, struct dc_plane_state *plane);
+/* This function restores link resource allocation state from a snapshot
+ * @dc: pointer to dc of the dm calling this
+ * @map: a dc link resource snapshot defined internally to dc.
+ *
+ * DM needs to call this function after initial link detection on boot and
+ * before first commit streams to restore link resource allocation state
+ * from previous boot session.
+ *
+ * Some of the link resource is using first come first serve policy.
+ * The allocation mapping depends on original hotplug order. This information
+ * is lost after driver is loaded next time. The snapshot is used in order to
+ * restore link resource to its previous state so user will get consistent
+ * link capability allocation across reboot.
+ *
+ */
+void dc_restore_link_res_map(const struct dc *dc, uint32_t *map);
+
+/* TODO: this is not meant to be exposed to DM. Should switch to stream update
+ * interface i.e stream_update->dsc_config
+ */
+bool dc_link_update_dsc_config(struct pipe_ctx *pipe_ctx);
+
+/* translate a raw link rate data to bandwidth in kbps */
+uint32_t dc_link_bw_kbps_from_raw_frl_link_rate_data(const struct dc *dc, uint8_t bw);
+
+/* determine the optimal bandwidth given link and required bw.
+ * @link - current detected link
+ * @req_bw - requested bandwidth in kbps
+ * @link_settings - returned most optimal link settings that can fit the
+ * requested bandwidth
+ * return - false if link can't support requested bandwidth, true if link
+ * settings is found.
+ */
+bool dc_link_decide_edp_link_settings(struct dc_link *link,
+		struct dc_link_settings *link_settings,
+		uint32_t req_bw);
+
+/* return the max dp link settings can be driven by the link without considering
+ * connected RX device and its capability
+ */
+bool dc_link_dp_get_max_link_enc_cap(const struct dc_link *link,
+		struct dc_link_settings *max_link_enc_cap);
+
+/* determine when the link is driving MST mode, what DP link channel coding
+ * format will be used. The decision will remain unchanged until next HPD event.
+ *
+ * @link -  a link with DP RX connection
+ * return - if stream is committed to this link with MST signal type, type of
+ * channel coding format dc will choose.
+ */
+enum dp_link_encoding dc_link_dp_mst_decide_link_encoding_format(
+		const struct dc_link *link);
+
+/* get max dp link settings the link can enable with all things considered. (i.e
+ * TX/RX/Cable capabilities and dp override policies.
+ *
+ * @link - a link with DP RX connection
+ * return - max dp link settings the link can enable.
+ *
+ */
+const struct dc_link_settings *dc_link_get_link_cap(const struct dc_link *link);
+
+/* Get the highest encoding format that the link supports; highest meaning the
+ * encoding format which supports the maximum bandwidth.
+ *
+ * @link - a link with DP RX connection
+ * return - highest encoding format link supports.
+ */
+enum dc_link_encoding_format dc_link_get_highest_encoding_format(const struct dc_link *link);
+
+/* Check if a RX (ex. DP sink, MST hub, passive or active dongle) is connected
+ * to a link with dp connector signal type.
+ * @link - a link with dp connector signal type
+ * return - true if connected, false otherwise
+ */
+bool dc_link_is_dp_sink_present(struct dc_link *link);
+
+/* Force DP lane settings update to main-link video signal and notify the change
+ * to DP RX via DPCD. This is a debug interface used for video signal integrity
+ * tuning purpose. The interface assumes link has already been enabled with DP
+ * signal.
+ *
+ * @lt_settings - a container structure with desired hw_lane_settings
+ */
+void dc_link_set_drive_settings(struct dc *dc,
+				struct link_training_settings *lt_settings,
+				struct dc_link *link);
 
-#endif
-/*******************************************************************************
- * Sink Interfaces - A sink corresponds to a display output device
- ******************************************************************************/
+/* Enable a test pattern in Link or PHY layer in an active link for compliance
+ * test or debugging purpose. The test pattern will remain until next un-plug.
+ *
+ * @link - active link with DP signal output enabled.
+ * @test_pattern - desired test pattern to output.
+ * NOTE: set to DP_TEST_PATTERN_VIDEO_MODE to disable previous test pattern.
+ * @test_pattern_color_space - for video test pattern choose a desired color
+ * space.
+ * @p_link_settings - For PHY pattern choose a desired link settings
+ * @p_custom_pattern - some test pattern will require a custom input to
+ * customize some pattern details. Otherwise keep it to NULL.
+ * @cust_pattern_size - size of the custom pattern input.
+ *
+ */
+bool dc_link_dp_set_test_pattern(
+	struct dc_link *link,
+	enum dp_test_pattern test_pattern,
+	enum dp_test_pattern_color_space test_pattern_color_space,
+	const struct link_training_settings *p_link_settings,
+	const unsigned char *p_custom_pattern,
+	unsigned int cust_pattern_size);
+
+/* Force DP link settings to always use a specific value until reboot to a
+ * specific link. If link has already been enabled, the interface will also
+ * switch to desired link settings immediately. This is a debug interface to
+ * generic dp issue trouble shooting.
+ */
+void dc_link_set_preferred_link_settings(struct dc *dc,
+		struct dc_link_settings *link_setting,
+		struct dc_link *link);
+
+/* Force DP link to customize a specific link training behavior by overriding to
+ * standard DP specs defined protocol. This is a debug interface to trouble shoot
+ * display specific link training issues or apply some display specific
+ * workaround in link training.
+ *
+ * @link_settings - if not NULL, force preferred link settings to the link.
+ * @lt_override - a set of override pointers. If any pointer is none NULL, dc
+ * will apply this particular override in future link training. If NULL is
+ * passed in, dc resets previous overrides.
+ * NOTE: DM must keep the memory from override pointers until DM resets preferred
+ * training settings.
+ */
+void dc_link_set_preferred_training_settings(struct dc *dc,
+		struct dc_link_settings *link_setting,
+		struct dc_link_training_overrides *lt_overrides,
+		struct dc_link *link,
+		bool skip_immediate_retrain);
+
+/* return - true if FEC is supported with connected DP RX, false otherwise */
+bool dc_link_is_fec_supported(const struct dc_link *link);
+
+/* query FEC enablement policy to determine if FEC will be enabled by dc during
+ * link enablement.
+ * return - true if FEC should be enabled, false otherwise.
+ */
+bool dc_link_should_enable_fec(const struct dc_link *link);
+
+/* determine lttpr mode the current link should be enabled with a specific link
+ * settings.
+ */
+enum lttpr_mode dc_link_decide_lttpr_mode(struct dc_link *link,
+		struct dc_link_settings *link_setting);
+
+/* Force DP RX to update its power state.
+ * NOTE: this interface doesn't update dp main-link. Calling this function will
+ * cause DP TX main-link and DP RX power states out of sync. DM has to restore
+ * RX power state back upon finish DM specific execution requiring DP RX in a
+ * specific power state.
+ * @on - true to set DP RX in D0 power state, false to set DP RX in D3 power
+ * state.
+ */
+void dc_link_dp_receiver_power_ctrl(struct dc_link *link, bool on);
+
+/* Force link to read base dp receiver caps from dpcd 000h - 00Fh and overwrite
+ * current value read from extended receiver cap from 02200h - 0220Fh.
+ * Some DP RX has problems of providing accurate DP receiver caps from extended
+ * field, this interface is a workaround to revert link back to use base caps.
+ */
+void dc_link_overwrite_extended_receiver_cap(
+		struct dc_link *link);
+
+void dc_link_edp_panel_backlight_power_on(struct dc_link *link,
+		bool wait_for_hpd);
+
+/* Set backlight level of an embedded panel (eDP, LVDS).
+ * backlight_pwm_u16_16 is unsigned 32 bit with 16 bit integer
+ * and 16 bit fractional, where 1.0 is max backlight value.
+ */
+bool dc_link_set_backlight_level(const struct dc_link *dc_link,
+		struct set_backlight_level_params *backlight_level_params);
+
+/* Set/get nits-based backlight level of an embedded panel (eDP, LVDS). */
+bool dc_link_set_backlight_level_nits(struct dc_link *link,
+		bool isHDR,
+		uint32_t backlight_millinits,
+		uint32_t transition_time_in_ms);
+
+bool dc_link_get_backlight_level_nits(struct dc_link *link,
+		uint32_t *backlight_millinits,
+		uint32_t *backlight_millinits_peak);
+
+int dc_link_get_backlight_level(const struct dc_link *dc_link);
+
+int dc_link_get_target_backlight_pwm(const struct dc_link *link);
+
+bool dc_link_set_psr_allow_active(struct dc_link *dc_link, const bool *enable,
+		bool wait, bool force_static, const unsigned int *power_opts);
+
+bool dc_link_get_psr_state(const struct dc_link *dc_link, enum dc_psr_state *state);
+
+bool dc_link_setup_psr(struct dc_link *dc_link,
+		const struct dc_stream_state *stream, struct psr_config *psr_config,
+		struct psr_context *psr_context);
+
+/*
+ * Communicate with DMUB to allow or disallow Panel Replay on the specified link:
+ *
+ * @link: pointer to the dc_link struct instance
+ * @enable: enable(active) or disable(inactive) replay
+ * @wait: state transition need to wait the active set completed.
+ * @force_static: force disable(inactive) the replay
+ * @power_opts: set power optimazation parameters to DMUB.
+ *
+ * return: allow Replay active will return true, else will return false.
+ */
+bool dc_link_set_replay_allow_active(struct dc_link *dc_link, const bool *enable,
+		bool wait, bool force_static, const unsigned int *power_opts);
+
+bool dc_link_get_replay_state(const struct dc_link *dc_link, uint64_t *state);
+
+/* On eDP links this function call will stall until T12 has elapsed.
+ * If the panel is not in power off state, this function will return
+ * immediately.
+ */
+bool dc_link_wait_for_t12(struct dc_link *link);
+
+/* Determine if dp trace has been initialized to reflect upto date result *
+ * return - true if trace is initialized and has valid data. False dp trace
+ * doesn't have valid result.
+ */
+bool dc_dp_trace_is_initialized(struct dc_link *link);
+
+/* Query a dp trace flag to indicate if the current dp trace data has been
+ * logged before
+ */
+bool dc_dp_trace_is_logged(struct dc_link *link,
+		bool in_detection);
+
+/* Set dp trace flag to indicate whether DM has already logged the current dp
+ * trace data. DM can set is_logged to true upon logging and check
+ * dc_dp_trace_is_logged before logging to avoid logging the same result twice.
+ */
+void dc_dp_trace_set_is_logged_flag(struct dc_link *link,
+		bool in_detection,
+		bool is_logged);
+
+/* Obtain driver time stamp for last dp link training end. The time stamp is
+ * formatted based on dm_get_timestamp DM function.
+ * @in_detection - true to get link training end time stamp of last link
+ * training in detection sequence. false to get link training end time stamp
+ * of last link training in commit (dpms) sequence
+ */
+unsigned long long dc_dp_trace_get_lt_end_timestamp(struct dc_link *link,
+		bool in_detection);
+
+/* Get how many link training attempts dc has done with latest sequence.
+ * @in_detection - true to get link training count of last link
+ * training in detection sequence. false to get link training count of last link
+ * training in commit (dpms) sequence
+ */
+const struct dp_trace_lt_counts *dc_dp_trace_get_lt_counts(struct dc_link *link,
+		bool in_detection);
+
+/* Get how many link loss has happened since last link training attempts */
+unsigned int dc_dp_trace_get_link_loss_count(struct dc_link *link);
+
+/*
+ *  USB4 DPIA BW ALLOCATION PUBLIC FUNCTIONS
+ */
+/*
+ * Send a request from DP-Tx requesting to allocate BW remotely after
+ * allocating it locally. This will get processed by CM and a CB function
+ * will be called.
+ *
+ * @link: pointer to the dc_link struct instance
+ * @req_bw: The requested bw in Kbyte to allocated
+ *
+ * return: none
+ */
+void dc_link_set_usb4_req_bw_req(struct dc_link *link, int req_bw);
+
+/*
+ * Handle the USB4 BW Allocation related functionality here:
+ * Plug => Try to allocate max bw from timing parameters supported by the sink
+ * Unplug => de-allocate bw
+ *
+ * @link: pointer to the dc_link struct instance
+ * @peak_bw: Peak bw used by the link/sink
+ *
+ */
+void dc_link_dp_dpia_handle_usb4_bandwidth_allocation_for_link(
+		struct dc_link *link, int peak_bw);
+
+/*
+ * Calculates the DP tunneling bandwidth required for the stream timing
+ * and aggregates the stream bandwidth for the respective DP tunneling link
+ *
+ * return: dc_status
+ */
+enum dc_status dc_link_validate_dp_tunneling_bandwidth(const struct dc *dc, const struct dc_state *new_ctx);
+
+/*
+ * Get if ALPM is supported by the link
+ */
+void dc_link_get_alpm_support(struct dc_link *link, bool *auxless_support,
+	bool *auxwake_support);
+
+/* Sink Interfaces - A sink corresponds to a display output device */
 
 struct dc_container_id {
 	// 128bit GUID in binary form
@@ -1290,14 +2560,29 @@ struct dc_sink_dsc_caps {
 	// 'true' if these are virtual DPCD's DSC caps (immediately upstream of sink in MST topology),
 	// 'false' if they are sink's DSC caps
 	bool is_virtual_dpcd_dsc;
+	// 'true' if MST topology supports DSC passthrough for sink
+	// 'false' if MST topology does not support DSC passthrough
+	bool is_dsc_passthrough_supported;
 	struct dsc_dec_dpcd_caps dsc_dec_caps;
 };
 
+struct dc_sink_hblank_expansion_caps {
+	// 'true' if these are virtual DPCD's HBlank expansion caps (immediately upstream of sink in MST topology),
+	// 'false' if they are sink's HBlank expansion caps
+	bool is_virtual_dpcd_hblank_expansion;
+	struct hblank_expansion_dpcd_caps dpcd_caps;
+};
+
 struct dc_sink_fec_caps {
 	bool is_rx_fec_supported;
 	bool is_topology_fec_supported;
 };
 
+struct scdc_caps {
+	union hdmi_scdc_manufacturer_OUI_data manufacturer_OUI;
+	union hdmi_scdc_device_id_data device_id;
+};
+
 /*
  * The sink structure contains EDID and other display device properties
  */
@@ -1311,8 +2596,10 @@ struct dc_sink {
 	struct stereo_3d_features features_3d[TIMING_3D_FORMAT_MAX];
 	bool converter_disable_audio;
 
+	struct scdc_caps scdc_caps;
 	struct dc_sink_dsc_caps dsc_caps;
 	struct dc_sink_fec_caps fec_caps;
+	struct dc_sink_hblank_expansion_caps hblank_expansion_caps;
 
 	bool is_vsc_sdp_colorimetry_supported;
 
@@ -1348,9 +2635,7 @@ struct dc_cursor {
 };
 
 
-/*******************************************************************************
- * Interrupt interfaces
- ******************************************************************************/
+/* Interrupt interfaces */
 enum dc_irq_source dc_interrupt_to_irq_source(
 		struct dc *dc,
 		uint32_t src_id,
@@ -1362,9 +2647,7 @@ enum dc_irq_source dc_get_hpd_irq_source_at_index(
 
 void dc_notify_vsync_int_state(struct dc *dc, struct dc_stream_state *stream, bool enable);
 
-/*******************************************************************************
- * Power Interfaces
- ******************************************************************************/
+/* Power Interfaces */
 
 void dc_set_power_state(
 		struct dc *dc,
@@ -1373,7 +2656,6 @@ void dc_resume(struct dc *dc);
 
 void dc_power_down_on_boot(struct dc *dc);
 
-#if defined(CONFIG_DRM_AMD_DC_HDCP)
 /*
  * HDCP Interfaces
  */
@@ -1381,47 +2663,69 @@ enum hdcp_message_status dc_process_hdcp_msg(
 		enum signal_type signal,
 		struct dc_link *link,
 		struct hdcp_protection_message *message_info);
-#endif
 bool dc_is_dmcu_initialized(struct dc *dc);
 
 enum dc_status dc_set_clock(struct dc *dc, enum dc_clock_type clock_type, uint32_t clk_khz, uint32_t stepping);
 void dc_get_clock(struct dc *dc, enum dc_clock_type clock_type, struct dc_clock_config *clock_cfg);
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 
-bool dc_is_plane_eligible_for_idle_optimizations(struct dc *dc, struct dc_plane_state *plane,
-				struct dc_cursor_attributes *cursor_attr);
+bool dc_is_plane_eligible_for_idle_optimizations(struct dc *dc,
+		unsigned int pitch,
+		unsigned int height,
+		enum surface_pixel_format format,
+		struct dc_cursor_attributes *cursor_attr);
 
-void dc_allow_idle_optimizations(struct dc *dc, bool allow);
+#define dc_allow_idle_optimizations(dc, allow) dc_allow_idle_optimizations_internal(dc, allow, __func__)
+#define dc_exit_ips_for_hw_access(dc) dc_exit_ips_for_hw_access_internal(dc, __func__)
 
-/*
- * blank all streams, and set min and max memory clock to
- * lowest and highest DPM level, respectively
- */
+void dc_allow_idle_optimizations_internal(struct dc *dc, bool allow, const char *caller_name);
+void dc_exit_ips_for_hw_access_internal(struct dc *dc, const char *caller_name);
+bool dc_dmub_is_ips_idle_state(struct dc *dc);
+
+/* set min and max memory clock to lowest and highest DPM level, respectively */
 void dc_unlock_memory_clock_frequency(struct dc *dc);
 
-/*
- * set min memory clock to the min required for current mode,
- * max to maxDPM, and unblank streams
- */
+/* set min memory clock to the min required for current mode, max to maxDPM */
 void dc_lock_memory_clock_frequency(struct dc *dc);
 
+/* set soft max for memclk, to be used for AC/DC switching clock limitations */
+void dc_enable_dcmode_clk_limit(struct dc *dc, bool enable);
+
 /* cleanup on driver unload */
 void dc_hardware_release(struct dc *dc);
 
-#endif
+/* disables fw based mclk switch */
+void dc_mclk_switch_using_fw_based_vblank_stretch_shut_down(struct dc *dc);
 
 bool dc_set_psr_allow_active(struct dc *dc, bool enable);
-#if defined(CONFIG_DRM_AMD_DC_DCN)
+
+bool dc_set_replay_allow_active(struct dc *dc, bool active);
+
+bool dc_set_ips_disable(struct dc *dc, unsigned int disable_ips);
+
 void dc_z10_restore(const struct dc *dc);
 void dc_z10_save_init(struct dc *dc);
-#endif
 
+bool dc_is_dmub_outbox_supported(struct dc *dc);
 bool dc_enable_dmub_notifications(struct dc *dc);
 
+bool dc_abm_save_restore(
+		struct dc *dc,
+		struct dc_stream_state *stream,
+		struct abm_save_restore *pData);
+
+void dc_enable_dmub_outbox(struct dc *dc);
+
 bool dc_process_dmub_aux_transfer_async(struct dc *dc,
 				uint32_t link_index,
 				struct aux_payload *payload);
 
+/*
+ * smart power OLED Interfaces
+ */
+bool dc_smart_power_oled_enable(const struct dc_link *link, bool enable, uint16_t peak_nits,
+	uint8_t debug_control, uint16_t fixed_CLL, uint32_t triggerline);
+bool dc_smart_power_oled_get_max_cll(const struct dc_link *link, unsigned int *pCurrent_MaxCLL);
+
 /* Get dc link index from dpia port index */
 uint8_t get_link_index_from_dpia_port_index(const struct dc *dc,
 				uint8_t dpia_port_index);
@@ -1436,14 +2740,544 @@ enum dc_status dc_process_dmub_set_mst_slots(const struct dc *dc,
 				uint8_t mst_alloc_slots,
 				uint8_t *mst_slots_in_use);
 
-/*******************************************************************************
- * DSC Interfaces
- ******************************************************************************/
+void dc_process_dmub_dpia_set_tps_notification(const struct dc *dc, uint32_t link_index, uint8_t tps);
+
+void dc_process_dmub_dpia_hpd_int_enable(const struct dc *dc,
+				uint32_t hpd_int_enable);
+
+void dc_print_dmub_diagnostic_data(const struct dc *dc);
+
+void dc_query_current_properties(struct dc *dc, struct dc_current_properties *properties);
+
+struct dc_power_profile {
+	int power_level; /* Lower is better */
+};
+
+struct dc_power_profile dc_get_power_profile_for_dc_state(const struct dc_state *context);
+
+unsigned int dc_get_det_buffer_size_from_state(const struct dc_state *context);
+
+bool dc_get_host_router_index(const struct dc_link *link, unsigned int *host_router_index);
+
+void dc_log_preos_dmcub_info(const struct dc *dc);
+
+/* DSC Interfaces */
 #include "dc_dsc.h"
 
-/*******************************************************************************
- * Disable acc mode Interfaces
- ******************************************************************************/
+void dc_get_visual_confirm_for_stream(
+	struct dc *dc,
+	struct dc_stream_state *stream_state,
+	struct tg_color *color);
+
+/* Disable acc mode Interfaces */
 void dc_disable_accelerated_mode(struct dc *dc);
 
+bool dc_is_timing_changed(struct dc_stream_state *cur_stream,
+		       struct dc_stream_state *new_stream);
+
+bool dc_is_cursor_limit_pending(struct dc *dc);
+bool dc_can_clear_cursor_limit(const struct dc *dc);
+
+/**
+ * dc_get_underflow_debug_data_for_otg() - Retrieve underflow debug data.
+ *
+ * @dc: Pointer to the display core context.
+ * @primary_otg_inst: Instance index of the primary OTG that underflowed.
+ * @out_data: Pointer to a dc_underflow_debug_data struct to be filled with debug information.
+ *
+ * This function collects and logs underflow-related HW states when underflow happens,
+ * including OTG underflow status, current read positions, frame count, and per-HUBP debug data.
+ * The results are stored in the provided out_data structure for further analysis or logging.
+ */
+void dc_get_underflow_debug_data_for_otg(struct dc *dc, int primary_otg_inst, struct dc_underflow_debug_data *out_data);
+
+void dc_get_power_feature_status(struct dc *dc, int primary_otg_inst, struct power_features *out_data);
+
+/**
+ * Software state variables used to program register fields across the display pipeline
+ */
+struct dc_register_software_state {
+	/* HUBP register programming variables for each pipe */
+	struct {
+		bool valid_plane_state;
+		bool valid_stream;
+		bool min_dc_gfx_version9;
+		uint32_t vtg_sel;                        /* DCHUBP_CNTL->HUBP_VTG_SEL from pipe_ctx->stream_res.tg->inst */
+		uint32_t hubp_clock_enable;              /* HUBP_CLK_CNTL->HUBP_CLOCK_ENABLE from power management */
+		uint32_t surface_pixel_format;           /* DCSURF_SURFACE_CONFIG->SURFACE_PIXEL_FORMAT from plane_state->format */
+		uint32_t rotation_angle;                 /* DCSURF_SURFACE_CONFIG->ROTATION_ANGLE from plane_state->rotation */
+		uint32_t h_mirror_en;                    /* DCSURF_SURFACE_CONFIG->H_MIRROR_EN from plane_state->horizontal_mirror */
+		uint32_t surface_dcc_en;                 /* DCSURF_SURFACE_CONTROL->PRIMARY_SURFACE_DCC_EN from dcc->enable */
+		uint32_t surface_size_width;             /* HUBP_SIZE->SURFACE_SIZE_WIDTH from plane_size.surface_size.width */
+		uint32_t surface_size_height;            /* HUBP_SIZE->SURFACE_SIZE_HEIGHT from plane_size.surface_size.height */
+		uint32_t pri_viewport_width;             /* DCSURF_PRI_VIEWPORT_DIMENSION->PRI_VIEWPORT_WIDTH from scaler_data.viewport.width */
+		uint32_t pri_viewport_height;            /* DCSURF_PRI_VIEWPORT_DIMENSION->PRI_VIEWPORT_HEIGHT from scaler_data.viewport.height */
+		uint32_t pri_viewport_x_start;           /* DCSURF_PRI_VIEWPORT_START->PRI_VIEWPORT_X_START from scaler_data.viewport.x */
+		uint32_t pri_viewport_y_start;           /* DCSURF_PRI_VIEWPORT_START->PRI_VIEWPORT_Y_START from scaler_data.viewport.y */
+		uint32_t cursor_enable;                  /* CURSOR_CONTROL->CURSOR_ENABLE from cursor_attributes.enable */
+		uint32_t cursor_width;                   /* CURSOR_SETTINGS->CURSOR_WIDTH from cursor_position.width */
+		uint32_t cursor_height;                  /* CURSOR_SETTINGS->CURSOR_HEIGHT from cursor_position.height */
+
+		/* Additional DCC configuration */
+		uint32_t surface_dcc_ind_64b_blk;        /* DCSURF_SURFACE_CONTROL->PRIMARY_SURFACE_DCC_IND_64B_BLK from dcc.independent_64b_blks */
+		uint32_t surface_dcc_ind_128b_blk;       /* DCSURF_SURFACE_CONTROL->PRIMARY_SURFACE_DCC_IND_128B_BLK from dcc.independent_128b_blks */
+
+		/* Surface pitch configuration */
+		uint32_t surface_pitch;                  /* DCSURF_SURFACE_PITCH->PITCH from plane_size.surface_pitch */
+		uint32_t meta_pitch;                     /* DCSURF_SURFACE_PITCH->META_PITCH from dcc.meta_pitch */
+		uint32_t chroma_pitch;                   /* DCSURF_SURFACE_PITCH_C->PITCH_C from plane_size.chroma_pitch */
+		uint32_t meta_pitch_c;                   /* DCSURF_SURFACE_PITCH_C->META_PITCH_C from dcc.meta_pitch_c */
+
+		/* Surface addresses */
+		uint32_t primary_surface_address_low;    /* DCSURF_PRIMARY_SURFACE_ADDRESS->PRIMARY_SURFACE_ADDRESS from address.grph.addr.low_part */
+		uint32_t primary_surface_address_high;   /* DCSURF_PRIMARY_SURFACE_ADDRESS_HIGH->PRIMARY_SURFACE_ADDRESS_HIGH from address.grph.addr.high_part */
+		uint32_t primary_meta_surface_address_low;  /* DCSURF_PRIMARY_META_SURFACE_ADDRESS->PRIMARY_META_SURFACE_ADDRESS from address.grph.meta_addr.low_part */
+		uint32_t primary_meta_surface_address_high; /* DCSURF_PRIMARY_META_SURFACE_ADDRESS_HIGH->PRIMARY_META_SURFACE_ADDRESS_HIGH from address.grph.meta_addr.high_part */
+
+		/* TMZ configuration */
+		uint32_t primary_surface_tmz;            /* DCSURF_SURFACE_CONTROL->PRIMARY_SURFACE_TMZ from address.tmz_surface */
+		uint32_t primary_meta_surface_tmz;       /* DCSURF_SURFACE_CONTROL->PRIMARY_META_SURFACE_TMZ from address.tmz_surface */
+
+		/* Tiling configuration */
+		uint32_t sw_mode;                        /* DCSURF_TILING_CONFIG->SW_MODE from tiling_info.gfx9.swizzle */
+		uint32_t num_pipes;                      /* DCSURF_ADDR_CONFIG->NUM_PIPES from tiling_info.gfx9.num_pipes */
+		uint32_t num_banks;                      /* DCSURF_ADDR_CONFIG->NUM_BANKS from tiling_info.gfx9.num_banks */
+		uint32_t pipe_interleave;                /* DCSURF_ADDR_CONFIG->PIPE_INTERLEAVE from tiling_info.gfx9.pipe_interleave */
+		uint32_t num_shader_engines;             /* DCSURF_ADDR_CONFIG->NUM_SE from tiling_info.gfx9.num_shader_engines */
+		uint32_t num_rb_per_se;                  /* DCSURF_ADDR_CONFIG->NUM_RB_PER_SE from tiling_info.gfx9.num_rb_per_se */
+		uint32_t num_pkrs;                       /* DCSURF_ADDR_CONFIG->NUM_PKRS from tiling_info.gfx9.num_pkrs */
+
+		/* DML Request Size Configuration - Luma */
+		uint32_t rq_chunk_size;                  /* DCHUBP_REQ_SIZE_CONFIG->CHUNK_SIZE from rq_regs.rq_regs_l.chunk_size */
+		uint32_t rq_min_chunk_size;              /* DCHUBP_REQ_SIZE_CONFIG->MIN_CHUNK_SIZE from rq_regs.rq_regs_l.min_chunk_size */
+		uint32_t rq_meta_chunk_size;             /* DCHUBP_REQ_SIZE_CONFIG->META_CHUNK_SIZE from rq_regs.rq_regs_l.meta_chunk_size */
+		uint32_t rq_min_meta_chunk_size;         /* DCHUBP_REQ_SIZE_CONFIG->MIN_META_CHUNK_SIZE from rq_regs.rq_regs_l.min_meta_chunk_size */
+		uint32_t rq_dpte_group_size;             /* DCHUBP_REQ_SIZE_CONFIG->DPTE_GROUP_SIZE from rq_regs.rq_regs_l.dpte_group_size */
+		uint32_t rq_mpte_group_size;             /* DCHUBP_REQ_SIZE_CONFIG->MPTE_GROUP_SIZE from rq_regs.rq_regs_l.mpte_group_size */
+		uint32_t rq_swath_height_l;              /* DCHUBP_REQ_SIZE_CONFIG->SWATH_HEIGHT_L from rq_regs.rq_regs_l.swath_height */
+		uint32_t rq_pte_row_height_l;            /* DCHUBP_REQ_SIZE_CONFIG->PTE_ROW_HEIGHT_L from rq_regs.rq_regs_l.pte_row_height */
+
+		/* DML Request Size Configuration - Chroma */
+		uint32_t rq_chunk_size_c;                /* DCHUBP_REQ_SIZE_CONFIG_C->CHUNK_SIZE_C from rq_regs.rq_regs_c.chunk_size */
+		uint32_t rq_min_chunk_size_c;            /* DCHUBP_REQ_SIZE_CONFIG_C->MIN_CHUNK_SIZE_C from rq_regs.rq_regs_c.min_chunk_size */
+		uint32_t rq_meta_chunk_size_c;           /* DCHUBP_REQ_SIZE_CONFIG_C->META_CHUNK_SIZE_C from rq_regs.rq_regs_c.meta_chunk_size */
+		uint32_t rq_min_meta_chunk_size_c;       /* DCHUBP_REQ_SIZE_CONFIG_C->MIN_META_CHUNK_SIZE_C from rq_regs.rq_regs_c.min_meta_chunk_size */
+		uint32_t rq_dpte_group_size_c;           /* DCHUBP_REQ_SIZE_CONFIG_C->DPTE_GROUP_SIZE_C from rq_regs.rq_regs_c.dpte_group_size */
+		uint32_t rq_mpte_group_size_c;           /* DCHUBP_REQ_SIZE_CONFIG_C->MPTE_GROUP_SIZE_C from rq_regs.rq_regs_c.mpte_group_size */
+		uint32_t rq_swath_height_c;              /* DCHUBP_REQ_SIZE_CONFIG_C->SWATH_HEIGHT_C from rq_regs.rq_regs_c.swath_height */
+		uint32_t rq_pte_row_height_c;            /* DCHUBP_REQ_SIZE_CONFIG_C->PTE_ROW_HEIGHT_C from rq_regs.rq_regs_c.pte_row_height */
+
+		/* DML Expansion Modes */
+		uint32_t drq_expansion_mode;             /* DCN_EXPANSION_MODE->DRQ_EXPANSION_MODE from rq_regs.drq_expansion_mode */
+		uint32_t prq_expansion_mode;             /* DCN_EXPANSION_MODE->PRQ_EXPANSION_MODE from rq_regs.prq_expansion_mode */
+		uint32_t mrq_expansion_mode;             /* DCN_EXPANSION_MODE->MRQ_EXPANSION_MODE from rq_regs.mrq_expansion_mode */
+		uint32_t crq_expansion_mode;             /* DCN_EXPANSION_MODE->CRQ_EXPANSION_MODE from rq_regs.crq_expansion_mode */
+
+		/* DML DLG parameters - nominal */
+		uint32_t dst_y_per_vm_vblank;            /* NOM_PARAMETERS_0->DST_Y_PER_VM_VBLANK from dlg_regs.dst_y_per_vm_vblank */
+		uint32_t dst_y_per_row_vblank;           /* NOM_PARAMETERS_0->DST_Y_PER_ROW_VBLANK from dlg_regs.dst_y_per_row_vblank */
+		uint32_t dst_y_per_vm_flip;              /* NOM_PARAMETERS_1->DST_Y_PER_VM_FLIP from dlg_regs.dst_y_per_vm_flip */
+		uint32_t dst_y_per_row_flip;             /* NOM_PARAMETERS_1->DST_Y_PER_ROW_FLIP from dlg_regs.dst_y_per_row_flip */
+
+		/* DML prefetch settings */
+		uint32_t dst_y_prefetch;                 /* PREFETCH_SETTINS->DST_Y_PREFETCH from dlg_regs.dst_y_prefetch */
+		uint32_t vratio_prefetch;                /* PREFETCH_SETTINS->VRATIO_PREFETCH from dlg_regs.vratio_prefetch */
+		uint32_t vratio_prefetch_c;              /* PREFETCH_SETTINS_C->VRATIO_PREFETCH_C from dlg_regs.vratio_prefetch_c */
+
+		/* TTU parameters */
+		uint32_t qos_level_low_wm;               /* TTU_CNTL1->QoSLevelLowWaterMark from ttu_regs.qos_level_low_wm */
+		uint32_t qos_level_high_wm;              /* TTU_CNTL1->QoSLevelHighWaterMark from ttu_regs.qos_level_high_wm */
+		uint32_t qos_level_flip;                 /* TTU_CNTL2->QoS_LEVEL_FLIP_L from ttu_regs.qos_level_flip */
+		uint32_t min_ttu_vblank;                 /* DCN_GLOBAL_TTU_CNTL->MIN_TTU_VBLANK from ttu_regs.min_ttu_vblank */
+	} hubp[MAX_PIPES];
+
+	/* HUBBUB register programming variables */
+	struct {
+		/* Individual DET buffer control per pipe - software state that programs DET registers */
+		uint32_t det0_size;                      /* DCHUBBUB_DET0_CTRL->DET0_SIZE from hubbub->funcs->program_det_size(hubbub, 0, det_buffer_size_kb) */
+		uint32_t det1_size;                      /* DCHUBBUB_DET1_CTRL->DET1_SIZE from hubbub->funcs->program_det_size(hubbub, 1, det_buffer_size_kb) */
+		uint32_t det2_size;                      /* DCHUBBUB_DET2_CTRL->DET2_SIZE from hubbub->funcs->program_det_size(hubbub, 2, det_buffer_size_kb) */
+		uint32_t det3_size;                      /* DCHUBBUB_DET3_CTRL->DET3_SIZE from hubbub->funcs->program_det_size(hubbub, 3, det_buffer_size_kb) */
+
+		/* Compression buffer control - software state that programs COMPBUF registers */
+		uint32_t compbuf_size;                   /* DCHUBBUB_COMPBUF_CTRL->COMPBUF_SIZE from hubbub->funcs->program_compbuf_size(hubbub, compbuf_size_kb, safe_to_increase) */
+		uint32_t compbuf_reserved_space_64b;     /* COMPBUF_RESERVED_SPACE->COMPBUF_RESERVED_SPACE_64B from hubbub2->pixel_chunk_size / 32 */
+		uint32_t compbuf_reserved_space_zs;      /* COMPBUF_RESERVED_SPACE->COMPBUF_RESERVED_SPACE_ZS from hubbub2->pixel_chunk_size / 128 */
+	} hubbub;
+
+	/* DPP register programming variables for each pipe (simplified for available fields) */
+	struct {
+		uint32_t dpp_clock_enable;               /* DPP_CONTROL->DPP_CLOCK_ENABLE from dppclk_enable */
+
+		/* Recout (Rectangle of Interest) configuration */
+		uint32_t recout_start_x;                 /* RECOUT_START->RECOUT_START_X from pipe_ctx->plane_res.scl_data.recout.x */
+		uint32_t recout_start_y;                 /* RECOUT_START->RECOUT_START_Y from pipe_ctx->plane_res.scl_data.recout.y */
+		uint32_t recout_width;                   /* RECOUT_SIZE->RECOUT_WIDTH from pipe_ctx->plane_res.scl_data.recout.width */
+		uint32_t recout_height;                  /* RECOUT_SIZE->RECOUT_HEIGHT from pipe_ctx->plane_res.scl_data.recout.height */
+
+		/* MPC (Multiple Pipe/Plane Combiner) size configuration */
+		uint32_t mpc_width;                      /* MPC_SIZE->MPC_WIDTH from pipe_ctx->plane_res.scl_data.h_active */
+		uint32_t mpc_height;                     /* MPC_SIZE->MPC_HEIGHT from pipe_ctx->plane_res.scl_data.v_active */
+
+		/* DSCL mode configuration */
+		uint32_t dscl_mode;                      /* SCL_MODE->DSCL_MODE from pipe_ctx->plane_res.scl_data.dscl_prog_data.dscl_mode */
+
+		/* Scaler ratios (simplified to integer parts) */
+		uint32_t horz_ratio_int;                 /* SCL_HORZ_FILTER_SCALE_RATIO->SCL_H_SCALE_RATIO integer part from ratios.horz */
+		uint32_t vert_ratio_int;                 /* SCL_VERT_FILTER_SCALE_RATIO->SCL_V_SCALE_RATIO integer part from ratios.vert */
+
+		/* Basic scaler taps */
+		uint32_t h_taps;                         /* SCL_TAP_CONTROL->SCL_H_NUM_TAPS from taps.h_taps */
+		uint32_t v_taps;                         /* SCL_TAP_CONTROL->SCL_V_NUM_TAPS from taps.v_taps */
+	} dpp[MAX_PIPES];
+
+	/* DCCG register programming variables */
+	struct {
+		/* Core Display Clock Control */
+		uint32_t dispclk_khz;                    /* DENTIST_DISPCLK_CNTL->DENTIST_DISPCLK_WDIVIDER from clk_mgr.dispclk_khz */
+		uint32_t dc_mem_global_pwr_req_dis;      /* DC_MEM_GLOBAL_PWR_REQ_CNTL->DC_MEM_GLOBAL_PWR_REQ_DIS from memory power management settings */
+
+		/* DPP Clock Control - 4 fields per pipe */
+		uint32_t dppclk_khz[MAX_PIPES];          /* DPPCLK_CTRL->DPPCLK_R_GATE_DISABLE from dpp_clocks[pipe] */
+		uint32_t dppclk_enable[MAX_PIPES];       /* DPPCLK_CTRL->DPPCLK0_EN,DPPCLK1_EN,DPPCLK2_EN,DPPCLK3_EN from dccg31_update_dpp_dto() */
+		uint32_t dppclk_dto_enable[MAX_PIPES];   /* DPPCLK_DTO_CTRL->DPPCLK_DTO_ENABLE from dccg->dpp_clock_gated[dpp_inst] state */
+		uint32_t dppclk_dto_phase[MAX_PIPES];    /* DPPCLK0_DTO_PARAM->DPPCLK0_DTO_PHASE from phase calculation req_dppclk/ref_dppclk */
+		uint32_t dppclk_dto_modulo[MAX_PIPES];   /* DPPCLK0_DTO_PARAM->DPPCLK0_DTO_MODULO from modulo = 0xff */
+
+		/* DSC Clock Control - 4 fields per DSC resource */
+		uint32_t dscclk_khz[MAX_PIPES]; /* DSCCLK_DTO_CTRL->DSCCLK_DTO_ENABLE from dsc_clocks */
+		uint32_t dscclk_dto_enable[MAX_PIPES]; /* DSCCLK_DTO_CTRL->DSCCLK0_DTO_ENABLE,DSCCLK1_DTO_ENABLE,DSCCLK2_DTO_ENABLE,DSCCLK3_DTO_ENABLE */
+		uint32_t dscclk_dto_phase[MAX_PIPES];  /* DSCCLK0_DTO_PARAM->DSCCLK0_DTO_PHASE from dccg31_enable_dscclk() */
+		uint32_t dscclk_dto_modulo[MAX_PIPES]; /* DSCCLK0_DTO_PARAM->DSCCLK0_DTO_MODULO from dccg31_enable_dscclk() */
+
+		/* Pixel Clock Control - per pipe */
+		uint32_t pixclk_khz[MAX_PIPES];          /* PIXCLK_RESYNC_CNTL->PIXCLK_RESYNC_ENABLE from stream.timing.pix_clk_100hz */
+		uint32_t otg_pixel_rate_div[MAX_PIPES];  /* OTG_PIXEL_RATE_DIV->OTG_PIXEL_RATE_DIV from OTG pixel rate divider control */
+		uint32_t dtbclk_dto_enable[MAX_PIPES];   /* OTG0_PIXEL_RATE_CNTL->DTBCLK_DTO_ENABLE from dccg31_set_dtbclk_dto() */
+		uint32_t pipe_dto_src_sel[MAX_PIPES];    /* OTG0_PIXEL_RATE_CNTL->PIPE_DTO_SRC_SEL from dccg31_set_dtbclk_dto() source selection */
+		uint32_t dtbclk_dto_div[MAX_PIPES];      /* OTG0_PIXEL_RATE_CNTL->DTBCLK_DTO_DIV from dtbdto_div calculation */
+		uint32_t otg_add_pixel[MAX_PIPES];       /* OTG0_PIXEL_RATE_CNTL->OTG_ADD_PIXEL from dccg31_otg_add_pixel() */
+		uint32_t otg_drop_pixel[MAX_PIPES];      /* OTG0_PIXEL_RATE_CNTL->OTG_DROP_PIXEL from dccg31_otg_drop_pixel() */
+
+		/* DTBCLK DTO Control - 4 DTOs */
+		uint32_t dtbclk_dto_modulo[4];           /* DTBCLK_DTO0_MODULO->DTBCLK_DTO0_MODULO from dccg31_set_dtbclk_dto() modulo calculation */
+		uint32_t dtbclk_dto_phase[4];            /* DTBCLK_DTO0_PHASE->DTBCLK_DTO0_PHASE from phase calculation pixclk_khz/ref_dtbclk_khz */
+		uint32_t dtbclk_dto_dbuf_en;             /* DTBCLK_DTO_DBUF_EN->DTBCLK DTO data buffer enable */
+
+		/* DP Stream Clock Control - 4 pipes */
+		uint32_t dpstreamclk_enable[MAX_PIPES];          /* DPSTREAMCLK_CNTL->DPSTREAMCLK_PIPE0_EN,DPSTREAMCLK_PIPE1_EN,DPSTREAMCLK_PIPE2_EN,DPSTREAMCLK_PIPE3_EN */
+		uint32_t dp_dto_modulo[4];               /* DP_DTO0_MODULO->DP_DTO0_MODULO from DP stream DTO programming */
+		uint32_t dp_dto_phase[4];                /* DP_DTO0_PHASE->DP_DTO0_PHASE from DP stream DTO programming */
+		uint32_t dp_dto_dbuf_en;                 /* DP_DTO_DBUF_EN->DP DTO data buffer enable */
+
+		/* PHY Symbol Clock Control - 5 PHYs (A,B,C,D,E) */
+		uint32_t phy_symclk_force_en[5];         /* PHYASYMCLK_CLOCK_CNTL->PHYASYMCLK_FORCE_EN from dccg31_set_physymclk() force_enable */
+		uint32_t phy_symclk_force_src_sel[5];    /* PHYASYMCLK_CLOCK_CNTL->PHYASYMCLK_FORCE_SRC_SEL from dccg31_set_physymclk() clk_src */
+		uint32_t phy_symclk_gate_disable[5];     /* DCCG_GATE_DISABLE_CNTL2->PHYASYMCLK_GATE_DISABLE from debug.root_clock_optimization.bits.physymclk */
+
+		/* SYMCLK32 SE Control - 4 instances */
+		uint32_t symclk32_se_src_sel[4];         /* SYMCLK32_SE_CNTL->SYMCLK32_SE0_SRC_SEL from dccg31_enable_symclk32_se() with get_phy_mux_symclk() mapping */
+		uint32_t symclk32_se_enable[4];          /* SYMCLK32_SE_CNTL->SYMCLK32_SE0_EN from dccg31_enable_symclk32_se() enable */
+		uint32_t symclk32_se_gate_disable[4];    /* DCCG_GATE_DISABLE_CNTL3->SYMCLK32_SE0_GATE_DISABLE from debug.root_clock_optimization.bits.symclk32_se */
+
+		/* SYMCLK32 LE Control - 2 instances */
+		uint32_t symclk32_le_src_sel[2];         /* SYMCLK32_LE_CNTL->SYMCLK32_LE0_SRC_SEL from dccg31_enable_symclk32_le() phyd32clk source */
+		uint32_t symclk32_le_enable[2];          /* SYMCLK32_LE_CNTL->SYMCLK32_LE0_EN from dccg31_enable_symclk32_le() enable */
+		uint32_t symclk32_le_gate_disable[2];    /* DCCG_GATE_DISABLE_CNTL3->SYMCLK32_LE0_GATE_DISABLE from debug.root_clock_optimization.bits.symclk32_le */
+
+		/* DPIA Clock Control */
+		uint32_t dpiaclk_540m_dto_modulo;        /* DPIACLK_540M_DTO_MODULO->DPIA 540MHz DTO modulo */
+		uint32_t dpiaclk_540m_dto_phase;         /* DPIACLK_540M_DTO_PHASE->DPIA 540MHz DTO phase */
+		uint32_t dpiaclk_810m_dto_modulo;        /* DPIACLK_810M_DTO_MODULO->DPIA 810MHz DTO modulo */
+		uint32_t dpiaclk_810m_dto_phase;         /* DPIACLK_810M_DTO_PHASE->DPIA 810MHz DTO phase */
+		uint32_t dpiaclk_dto_cntl;               /* DPIACLK_DTO_CNTL->DPIA clock DTO control */
+		uint32_t dpiasymclk_cntl;                /* DPIASYMCLK_CNTL->DPIA symbol clock control */
+
+		/* Clock Gating Control */
+		uint32_t dccg_gate_disable_cntl;         /* DCCG_GATE_DISABLE_CNTL->Clock gate disable control from dccg31_init() */
+		uint32_t dpstreamclk_gate_disable;       /* DCCG_GATE_DISABLE_CNTL3->DPSTREAMCLK_GATE_DISABLE from debug.root_clock_optimization.bits.dpstream */
+		uint32_t dpstreamclk_root_gate_disable;  /* DCCG_GATE_DISABLE_CNTL3->DPSTREAMCLK_ROOT_GATE_DISABLE from debug.root_clock_optimization.bits.dpstream */
+
+		/* VSync Control */
+		uint32_t vsync_cnt_ctrl;                 /* DCCG_VSYNC_CNT_CTRL->VSync counter control */
+		uint32_t vsync_cnt_int_ctrl;             /* DCCG_VSYNC_CNT_INT_CTRL->VSync counter interrupt control */
+		uint32_t vsync_otg_latch_value[6];       /* DCCG_VSYNC_OTG0_LATCH_VALUE->OTG0 VSync latch value (for OTG0-5) */
+
+		/* Time Base Control */
+		uint32_t microsecond_time_base_div;      /* MICROSECOND_TIME_BASE_DIV->Microsecond time base divider */
+		uint32_t millisecond_time_base_div;      /* MILLISECOND_TIME_BASE_DIV->Millisecond time base divider */
+	} dccg;
+
+	/* DSC essential configuration for underflow analysis */
+	struct {
+		/* DSC active state - critical for bandwidth analysis */
+		uint32_t dsc_clock_enable;               /* DSC enabled - affects bandwidth requirements */
+
+		/* DSC configuration affecting bandwidth and timing */
+		uint32_t dsc_num_slices_h;              /* Horizontal slice count - affects throughput */
+		uint32_t dsc_num_slices_v;              /* Vertical slice count - affects throughput */
+		uint32_t dsc_bits_per_pixel;            /* Compression ratio - affects bandwidth */
+
+		/* OPP integration - affects pipeline flow */
+		uint32_t dscrm_dsc_forward_enable;      /* DSC forwarding to OPP enabled */
+		uint32_t dscrm_dsc_opp_pipe_source;    /* Which OPP receives DSC output */
+	} dsc[MAX_PIPES];
+
+	/* MPC register programming variables */
+	struct {
+		/* MPCC blending tree and mode control */
+		uint32_t mpcc_mode[MAX_PIPES];           /* MPCC_CONTROL->MPCC_MODE from blend_cfg.blend_mode */
+		uint32_t mpcc_alpha_blend_mode[MAX_PIPES]; /* MPCC_CONTROL->MPCC_ALPHA_BLND_MODE from blend_cfg.alpha_mode */
+		uint32_t mpcc_alpha_multiplied_mode[MAX_PIPES]; /* MPCC_CONTROL->MPCC_ALPHA_MULTIPLIED_MODE from blend_cfg.pre_multiplied_alpha */
+		uint32_t mpcc_blnd_active_overlap_only[MAX_PIPES]; /* MPCC_CONTROL->MPCC_BLND_ACTIVE_OVERLAP_ONLY from blend_cfg.overlap_only */
+		uint32_t mpcc_global_alpha[MAX_PIPES];   /* MPCC_CONTROL->MPCC_GLOBAL_ALPHA from blend_cfg.global_alpha */
+		uint32_t mpcc_global_gain[MAX_PIPES];    /* MPCC_CONTROL->MPCC_GLOBAL_GAIN from blend_cfg.global_gain */
+		uint32_t mpcc_bg_bpc[MAX_PIPES];         /* MPCC_CONTROL->MPCC_BG_BPC from background color depth */
+		uint32_t mpcc_bot_gain_mode[MAX_PIPES];  /* MPCC_CONTROL->MPCC_BOT_GAIN_MODE from bottom layer gain control */
+
+		/* MPCC blending tree connections */
+		uint32_t mpcc_bot_sel[MAX_PIPES];        /* MPCC_BOT_SEL->MPCC_BOT_SEL from mpcc_state->bot_sel */
+		uint32_t mpcc_top_sel[MAX_PIPES];        /* MPCC_TOP_SEL->MPCC_TOP_SEL from mpcc_state->dpp_id */
+
+		/* MPCC output gamma control */
+		uint32_t mpcc_ogam_mode[MAX_PIPES];      /* MPCC_OGAM_CONTROL->MPCC_OGAM_MODE from output gamma mode */
+		uint32_t mpcc_ogam_select[MAX_PIPES];    /* MPCC_OGAM_CONTROL->MPCC_OGAM_SELECT from gamma LUT bank selection */
+		uint32_t mpcc_ogam_pwl_disable[MAX_PIPES]; /* MPCC_OGAM_CONTROL->MPCC_OGAM_PWL_DISABLE from PWL control */
+
+		/* MPCC pipe assignment and status */
+		uint32_t mpcc_opp_id[MAX_PIPES];         /* MPCC_OPP_ID->MPCC_OPP_ID from mpcc_state->opp_id */
+		uint32_t mpcc_idle[MAX_PIPES];           /* MPCC_STATUS->MPCC_IDLE from mpcc idle status */
+		uint32_t mpcc_busy[MAX_PIPES];           /* MPCC_STATUS->MPCC_BUSY from mpcc busy status */
+
+		/* MPC output processing */
+		uint32_t mpc_out_csc_mode;               /* MPC_OUT_CSC_COEF->MPC_OUT_CSC_MODE from output_csc */
+		uint32_t mpc_out_gamma_mode;             /* MPC_OUT_GAMMA_LUT->MPC_OUT_GAMMA_MODE from output_gamma */
+	} mpc;
+
+	/* OPP register programming variables for each pipe */
+	struct {
+		/* Display Pattern Generator (DPG) Control - 19 fields from DPG_CONTROL register */
+		uint32_t dpg_enable;                     /* DPG_CONTROL->DPG_EN from test_pattern parameter (enable/disable) */
+
+		/* Format Control (FMT) - 18 fields from FMT_CONTROL register */
+		uint32_t fmt_pixel_encoding;             /* FMT_CONTROL->FMT_PIXEL_ENCODING from clamping->pixel_encoding */
+		uint32_t fmt_subsampling_mode;           /* FMT_CONTROL->FMT_SUBSAMPLING_MODE from force_chroma_subsampling_1tap */
+		uint32_t fmt_cbcr_bit_reduction_bypass;  /* FMT_CONTROL->FMT_CBCR_BIT_REDUCTION_BYPASS from pixel_encoding bypass control */
+		uint32_t fmt_stereosync_override;        /* FMT_CONTROL->FMT_STEREOSYNC_OVERRIDE from stereo timing override */
+		uint32_t fmt_spatial_dither_frame_counter_max; /* FMT_CONTROL->FMT_SPATIAL_DITHER_FRAME_COUNTER_MAX from fmt_bit_depth->flags */
+		uint32_t fmt_spatial_dither_frame_counter_bit_swap; /* FMT_CONTROL->FMT_SPATIAL_DITHER_FRAME_COUNTER_BIT_SWAP from dither control */
+		uint32_t fmt_truncate_enable;            /* FMT_CONTROL->FMT_TRUNCATE_EN from fmt_bit_depth->flags.TRUNCATE_ENABLED */
+		uint32_t fmt_truncate_depth;             /* FMT_CONTROL->FMT_TRUNCATE_DEPTH from fmt_bit_depth->flags.TRUNCATE_DEPTH */
+		uint32_t fmt_truncate_mode;              /* FMT_CONTROL->FMT_TRUNCATE_MODE from fmt_bit_depth->flags.TRUNCATE_MODE */
+		uint32_t fmt_spatial_dither_enable;      /* FMT_CONTROL->FMT_SPATIAL_DITHER_EN from fmt_bit_depth->flags.SPATIAL_DITHER_ENABLED */
+		uint32_t fmt_spatial_dither_mode;        /* FMT_CONTROL->FMT_SPATIAL_DITHER_MODE from fmt_bit_depth->flags.SPATIAL_DITHER_MODE */
+		uint32_t fmt_spatial_dither_depth;       /* FMT_CONTROL->FMT_SPATIAL_DITHER_DEPTH from fmt_bit_depth->flags.SPATIAL_DITHER_DEPTH */
+		uint32_t fmt_temporal_dither_enable;     /* FMT_CONTROL->FMT_TEMPORAL_DITHER_EN from fmt_bit_depth->flags.TEMPORAL_DITHER_ENABLED */
+		uint32_t fmt_clamp_data_enable;          /* FMT_CONTROL->FMT_CLAMP_DATA_EN from clamping->clamping_range enable */
+		uint32_t fmt_clamp_color_format;         /* FMT_CONTROL->FMT_CLAMP_COLOR_FORMAT from clamping->color_format */
+		uint32_t fmt_dynamic_exp_enable;         /* FMT_CONTROL->FMT_DYNAMIC_EXP_EN from color_sp/color_dpth/signal */
+		uint32_t fmt_dynamic_exp_mode;           /* FMT_CONTROL->FMT_DYNAMIC_EXP_MODE from color space mode mapping */
+		uint32_t fmt_bit_depth_control;          /* Legacy field - kept for compatibility */
+
+		/* OPP Pipe Control - 1 field from OPP_PIPE_CONTROL register */
+		uint32_t opp_pipe_clock_enable;          /* OPP_PIPE_CONTROL->OPP_PIPE_CLOCK_EN from enable parameter (bool) */
+
+		/* OPP CRC Control - 3 fields from OPP_PIPE_CRC_CONTROL register */
+		uint32_t opp_crc_enable;                 /* OPP_PIPE_CRC_CONTROL->CRC_EN from CRC enable control */
+		uint32_t opp_crc_select_source;          /* OPP_PIPE_CRC_CONTROL->CRC_SELECT_SOURCE from CRC source selection */
+		uint32_t opp_crc_stereo_cont;            /* OPP_PIPE_CRC_CONTROL->CRC_STEREO_CONT from stereo continuous CRC */
+
+		/* Output Buffer (OPPBUF) Control - 6 fields from OPPBUF_CONTROL register */
+		uint32_t oppbuf_active_width;            /* OPPBUF_CONTROL->OPPBUF_ACTIVE_WIDTH from oppbuf_params->active_width */
+		uint32_t oppbuf_pixel_repetition;        /* OPPBUF_CONTROL->OPPBUF_PIXEL_REPETITION from oppbuf_params->pixel_repetition */
+		uint32_t oppbuf_display_segmentation;    /* OPPBUF_CONTROL->OPPBUF_DISPLAY_SEGMENTATION from oppbuf_params->mso_segmentation */
+		uint32_t oppbuf_overlap_pixel_num;       /* OPPBUF_CONTROL->OPPBUF_OVERLAP_PIXEL_NUM from oppbuf_params->mso_overlap_pixel_num */
+		uint32_t oppbuf_3d_vact_space1_size;     /* OPPBUF_CONTROL->OPPBUF_3D_VACT_SPACE1_SIZE from 3D timing space1_size */
+		uint32_t oppbuf_3d_vact_space2_size;     /* OPPBUF_CONTROL->OPPBUF_3D_VACT_SPACE2_SIZE from 3D timing space2_size */
+
+		/* DSC Forward Config - 3 fields from DSCRM_DSC_FORWARD_CONFIG register */
+		uint32_t dscrm_dsc_forward_enable;       /* DSCRM_DSC_FORWARD_CONFIG->DSCRM_DSC_FORWARD_EN from DSC forward enable control */
+		uint32_t dscrm_dsc_opp_pipe_source;      /* DSCRM_DSC_FORWARD_CONFIG->DSCRM_DSC_OPP_PIPE_SOURCE from opp_pipe parameter */
+		uint32_t dscrm_dsc_forward_enable_status; /* DSCRM_DSC_FORWARD_CONFIG->DSCRM_DSC_FORWARD_EN_STATUS from DSC forward status (read-only) */
+	} opp[MAX_PIPES];
+
+	/* OPTC register programming variables for each pipe */
+	struct {
+		uint32_t otg_master_inst;
+
+		/* OTG_CONTROL register - 5 fields for OTG control */
+		uint32_t otg_master_enable;              /* OTG_CONTROL->OTG_MASTER_EN from timing enable/disable control */
+		uint32_t otg_disable_point_cntl;         /* OTG_CONTROL->OTG_DISABLE_POINT_CNTL from disable timing control */
+		uint32_t otg_start_point_cntl;           /* OTG_CONTROL->OTG_START_POINT_CNTL from start timing control */
+		uint32_t otg_field_number_cntl;          /* OTG_CONTROL->OTG_FIELD_NUMBER_CNTL from interlace field control */
+		uint32_t otg_out_mux;                    /* OTG_CONTROL->OTG_OUT_MUX from output mux selection */
+
+		/* OTG Horizontal Timing - 7 fields */
+		uint32_t otg_h_total;                    /* OTG_H_TOTAL->OTG_H_TOTAL from dc_crtc_timing->h_total */
+		uint32_t otg_h_blank_start;              /* OTG_H_BLANK_START_END->OTG_H_BLANK_START from dc_crtc_timing->h_front_porch */
+		uint32_t otg_h_blank_end;                /* OTG_H_BLANK_START_END->OTG_H_BLANK_END from dc_crtc_timing->h_addressable_video_pixel_width */
+		uint32_t otg_h_sync_start;               /* OTG_H_SYNC_A->OTG_H_SYNC_A_START from dc_crtc_timing->h_sync_width */
+		uint32_t otg_h_sync_end;                 /* OTG_H_SYNC_A->OTG_H_SYNC_A_END from calculated sync end position */
+		uint32_t otg_h_sync_polarity;            /* OTG_H_SYNC_A_CNTL->OTG_H_SYNC_A_POL from dc_crtc_timing->flags.HSYNC_POSITIVE_POLARITY */
+		uint32_t otg_h_timing_div_mode;          /* OTG_H_TIMING_CNTL->OTG_H_TIMING_DIV_MODE from horizontal timing division mode */
+
+		/* OTG Vertical Timing - 7 fields */
+		uint32_t otg_v_total;                    /* OTG_V_TOTAL->OTG_V_TOTAL from dc_crtc_timing->v_total */
+		uint32_t otg_v_blank_start;              /* OTG_V_BLANK_START_END->OTG_V_BLANK_START from dc_crtc_timing->v_front_porch */
+		uint32_t otg_v_blank_end;                /* OTG_V_BLANK_START_END->OTG_V_BLANK_END from dc_crtc_timing->v_addressable_video_line_width */
+		uint32_t otg_v_sync_start;               /* OTG_V_SYNC_A->OTG_V_SYNC_A_START from dc_crtc_timing->v_sync_width */
+		uint32_t otg_v_sync_end;                 /* OTG_V_SYNC_A->OTG_V_SYNC_A_END from calculated sync end position */
+		uint32_t otg_v_sync_polarity;            /* OTG_V_SYNC_A_CNTL->OTG_V_SYNC_A_POL from dc_crtc_timing->flags.VSYNC_POSITIVE_POLARITY */
+		uint32_t otg_v_sync_mode;                /* OTG_V_SYNC_A_CNTL->OTG_V_SYNC_MODE from sync mode selection */
+
+		/* OTG DRR (Dynamic Refresh Rate) Control - 8 fields */
+		uint32_t otg_v_total_max;                /* OTG_V_TOTAL_MAX->OTG_V_TOTAL_MAX from drr_params->vertical_total_max */
+		uint32_t otg_v_total_min;                /* OTG_V_TOTAL_MIN->OTG_V_TOTAL_MIN from drr_params->vertical_total_min */
+		uint32_t otg_v_total_mid;                /* OTG_V_TOTAL_MID->OTG_V_TOTAL_MID from drr_params->vertical_total_mid */
+		uint32_t otg_v_total_max_sel;            /* OTG_V_TOTAL_CONTROL->OTG_V_TOTAL_MAX_SEL from DRR max selection enable */
+		uint32_t otg_v_total_min_sel;            /* OTG_V_TOTAL_CONTROL->OTG_V_TOTAL_MIN_SEL from DRR min selection enable */
+		uint32_t otg_vtotal_mid_replacing_max_en; /* OTG_V_TOTAL_CONTROL->OTG_VTOTAL_MID_REPLACING_MAX_EN from DRR mid-frame enable */
+		uint32_t otg_vtotal_mid_frame_num;       /* OTG_V_TOTAL_CONTROL->OTG_VTOTAL_MID_FRAME_NUM from drr_params->vertical_total_mid_frame_num */
+		uint32_t otg_set_v_total_min_mask;       /* OTG_V_TOTAL_CONTROL->OTG_SET_V_TOTAL_MIN_MASK from DRR trigger mask */
+		uint32_t otg_force_lock_on_event;        /* OTG_V_TOTAL_CONTROL->OTG_FORCE_LOCK_ON_EVENT from DRR force lock control */
+
+		/* OPTC Data Source and ODM - 6 fields */
+		uint32_t optc_seg0_src_sel;              /* OPTC_DATA_SOURCE_SELECT->OPTC_SEG0_SRC_SEL from opp_id[0] ODM segment 0 source */
+		uint32_t optc_seg1_src_sel;              /* OPTC_DATA_SOURCE_SELECT->OPTC_SEG1_SRC_SEL from opp_id[1] ODM segment 1 source */
+		uint32_t optc_seg2_src_sel;              /* OPTC_DATA_SOURCE_SELECT->OPTC_SEG2_SRC_SEL from opp_id[2] ODM segment 2 source */
+		uint32_t optc_seg3_src_sel;              /* OPTC_DATA_SOURCE_SELECT->OPTC_SEG3_SRC_SEL from opp_id[3] ODM segment 3 source */
+		uint32_t optc_num_of_input_segment;      /* OPTC_DATA_SOURCE_SELECT->OPTC_NUM_OF_INPUT_SEGMENT from opp_cnt-1 number of input segments */
+		uint32_t optc_mem_sel;                   /* OPTC_MEMORY_CONFIG->OPTC_MEM_SEL from memory_mask ODM memory selection */
+
+		/* OPTC Data Format and DSC - 4 fields */
+		uint32_t optc_data_format;               /* OPTC_DATA_FORMAT_CONTROL->OPTC_DATA_FORMAT from data format selection */
+		uint32_t optc_dsc_mode;                  /* OPTC_DATA_FORMAT_CONTROL->OPTC_DSC_MODE from dsc_mode parameter */
+		uint32_t optc_dsc_bytes_per_pixel;       /* OPTC_BYTES_PER_PIXEL->OPTC_DSC_BYTES_PER_PIXEL from dsc_bytes_per_pixel parameter */
+		uint32_t optc_segment_width;             /* OPTC_WIDTH_CONTROL->OPTC_SEGMENT_WIDTH from segment_width parameter */
+		uint32_t optc_dsc_slice_width;           /* OPTC_WIDTH_CONTROL->OPTC_DSC_SLICE_WIDTH from dsc_slice_width parameter */
+
+		/* OPTC Clock and Underflow Control - 4 fields */
+		uint32_t optc_input_pix_clk_en;          /* OPTC_INPUT_CLOCK_CONTROL->OPTC_INPUT_PIX_CLK_EN from pixel clock enable */
+		uint32_t optc_underflow_occurred_status; /* OPTC_INPUT_GLOBAL_CONTROL->OPTC_UNDERFLOW_OCCURRED_STATUS from underflow status (read-only) */
+		uint32_t optc_underflow_clear;           /* OPTC_INPUT_GLOBAL_CONTROL->OPTC_UNDERFLOW_CLEAR from underflow clear control */
+		uint32_t otg_clock_enable;               /* OTG_CLOCK_CONTROL->OTG_CLOCK_EN from OTG clock enable */
+		uint32_t otg_clock_gate_dis;             /* OTG_CLOCK_CONTROL->OTG_CLOCK_GATE_DIS from clock gate disable */
+
+		/* OTG Stereo and 3D Control - 6 fields */
+		uint32_t otg_stereo_enable;              /* OTG_STEREO_CONTROL->OTG_STEREO_EN from stereo enable control */
+		uint32_t otg_stereo_sync_output_line_num; /* OTG_STEREO_CONTROL->OTG_STEREO_SYNC_OUTPUT_LINE_NUM from timing->stereo_3d_format line num */
+		uint32_t otg_stereo_sync_output_polarity; /* OTG_STEREO_CONTROL->OTG_STEREO_SYNC_OUTPUT_POLARITY from stereo polarity control */
+		uint32_t otg_3d_structure_en;            /* OTG_3D_STRUCTURE_CONTROL->OTG_3D_STRUCTURE_EN from 3D structure enable */
+		uint32_t otg_3d_structure_v_update_mode; /* OTG_3D_STRUCTURE_CONTROL->OTG_3D_STRUCTURE_V_UPDATE_MODE from 3D vertical update mode */
+		uint32_t otg_3d_structure_stereo_sel_ovr; /* OTG_3D_STRUCTURE_CONTROL->OTG_3D_STRUCTURE_STEREO_SEL_OVR from 3D stereo selection override */
+		uint32_t otg_interlace_enable;           /* OTG_INTERLACE_CONTROL->OTG_INTERLACE_ENABLE from dc_crtc_timing->flags.INTERLACE */
+
+		/* OTG GSL (Global Sync Lock) Control - 5 fields */
+		uint32_t otg_gsl0_en;                    /* OTG_GSL_CONTROL->OTG_GSL0_EN from GSL group 0 enable */
+		uint32_t otg_gsl1_en;                    /* OTG_GSL_CONTROL->OTG_GSL1_EN from GSL group 1 enable */
+		uint32_t otg_gsl2_en;                    /* OTG_GSL_CONTROL->OTG_GSL2_EN from GSL group 2 enable */
+		uint32_t otg_gsl_master_en;              /* OTG_GSL_CONTROL->OTG_GSL_MASTER_EN from GSL master enable */
+		uint32_t otg_gsl_master_mode;            /* OTG_GSL_CONTROL->OTG_GSL_MASTER_MODE from gsl_params->gsl_master mode */
+
+		/* OTG DRR Advanced Control - 4 fields */
+		uint32_t otg_v_total_last_used_by_drr;   /* OTG_DRR_CONTROL->OTG_V_TOTAL_LAST_USED_BY_DRR from last used DRR V_TOTAL (read-only) */
+		uint32_t otg_drr_trigger_window_start_x; /* OTG_DRR_TRIGGER_WINDOW->OTG_DRR_TRIGGER_WINDOW_START_X from window_start parameter */
+		uint32_t otg_drr_trigger_window_end_x;   /* OTG_DRR_TRIGGER_WINDOW->OTG_DRR_TRIGGER_WINDOW_END_X from window_end parameter */
+		uint32_t otg_drr_v_total_change_limit;   /* OTG_DRR_V_TOTAL_CHANGE->OTG_DRR_V_TOTAL_CHANGE_LIMIT from limit parameter */
+
+		/* OTG DSC Position Control - 2 fields */
+		uint32_t otg_dsc_start_position_x;       /* OTG_DSC_START_POSITION->OTG_DSC_START_POSITION_X from DSC start X position */
+		uint32_t otg_dsc_start_position_line_num; /* OTG_DSC_START_POSITION->OTG_DSC_START_POSITION_LINE_NUM from DSC start line number */
+
+		/* OTG Double Buffer Control - 2 fields */
+		uint32_t otg_drr_timing_dbuf_update_mode; /* OTG_DOUBLE_BUFFER_CONTROL->OTG_DRR_TIMING_DBUF_UPDATE_MODE from DRR double buffer mode */
+		uint32_t otg_blank_data_double_buffer_en; /* OTG_DOUBLE_BUFFER_CONTROL->OTG_BLANK_DATA_DOUBLE_BUFFER_EN from blank data double buffer enable */
+
+		/* OTG Vertical Interrupts - 6 fields */
+		uint32_t otg_vertical_interrupt0_int_enable; /* OTG_VERTICAL_INTERRUPT0_CONTROL->OTG_VERTICAL_INTERRUPT0_INT_ENABLE from interrupt 0 enable */
+		uint32_t otg_vertical_interrupt0_line_start; /* OTG_VERTICAL_INTERRUPT0_POSITION->OTG_VERTICAL_INTERRUPT0_LINE_START from start_line parameter */
+		uint32_t otg_vertical_interrupt1_int_enable; /* OTG_VERTICAL_INTERRUPT1_CONTROL->OTG_VERTICAL_INTERRUPT1_INT_ENABLE from interrupt 1 enable */
+		uint32_t otg_vertical_interrupt1_line_start; /* OTG_VERTICAL_INTERRUPT1_POSITION->OTG_VERTICAL_INTERRUPT1_LINE_START from start_line parameter */
+		uint32_t otg_vertical_interrupt2_int_enable; /* OTG_VERTICAL_INTERRUPT2_CONTROL->OTG_VERTICAL_INTERRUPT2_INT_ENABLE from interrupt 2 enable */
+		uint32_t otg_vertical_interrupt2_line_start; /* OTG_VERTICAL_INTERRUPT2_POSITION->OTG_VERTICAL_INTERRUPT2_LINE_START from start_line parameter */
+
+		/* OTG Global Sync Parameters - 6 fields */
+		uint32_t otg_vready_offset;              /* OTG_VREADY_PARAM->OTG_VREADY_OFFSET from vready_offset parameter */
+		uint32_t otg_vstartup_start;             /* OTG_VSTARTUP_PARAM->OTG_VSTARTUP_START from vstartup_start parameter */
+		uint32_t otg_vupdate_offset;             /* OTG_VUPDATE_PARAM->OTG_VUPDATE_OFFSET from vupdate_offset parameter */
+		uint32_t otg_vupdate_width;              /* OTG_VUPDATE_PARAM->OTG_VUPDATE_WIDTH from vupdate_width parameter */
+		uint32_t master_update_lock_vupdate_keepout_start_offset; /* OTG_VUPDATE_KEEPOUT->MASTER_UPDATE_LOCK_VUPDATE_KEEPOUT_START_OFFSET from pstate_keepout start */
+		uint32_t master_update_lock_vupdate_keepout_end_offset;   /* OTG_VUPDATE_KEEPOUT->MASTER_UPDATE_LOCK_VUPDATE_KEEPOUT_END_OFFSET from pstate_keepout end */
+
+		/* OTG Manual Trigger Control - 11 fields */
+		uint32_t otg_triga_source_select;        /* OTG_TRIGA_CNTL->OTG_TRIGA_SOURCE_SELECT from trigger A source selection */
+		uint32_t otg_triga_source_pipe_select;   /* OTG_TRIGA_CNTL->OTG_TRIGA_SOURCE_PIPE_SELECT from trigger A pipe selection */
+		uint32_t otg_triga_rising_edge_detect_cntl; /* OTG_TRIGA_CNTL->OTG_TRIGA_RISING_EDGE_DETECT_CNTL from trigger A rising edge detect */
+		uint32_t otg_triga_falling_edge_detect_cntl; /* OTG_TRIGA_CNTL->OTG_TRIGA_FALLING_EDGE_DETECT_CNTL from trigger A falling edge detect */
+		uint32_t otg_triga_polarity_select;      /* OTG_TRIGA_CNTL->OTG_TRIGA_POLARITY_SELECT from trigger A polarity selection */
+		uint32_t otg_triga_frequency_select;     /* OTG_TRIGA_CNTL->OTG_TRIGA_FREQUENCY_SELECT from trigger A frequency selection */
+		uint32_t otg_triga_delay;                /* OTG_TRIGA_CNTL->OTG_TRIGA_DELAY from trigger A delay */
+		uint32_t otg_triga_clear;                /* OTG_TRIGA_CNTL->OTG_TRIGA_CLEAR from trigger A clear */
+		uint32_t otg_triga_manual_trig;          /* OTG_TRIGA_MANUAL_TRIG->OTG_TRIGA_MANUAL_TRIG from manual trigger A */
+		uint32_t otg_trigb_source_select;        /* OTG_TRIGB_CNTL->OTG_TRIGB_SOURCE_SELECT from trigger B source selection */
+		uint32_t otg_trigb_polarity_select;      /* OTG_TRIGB_CNTL->OTG_TRIGB_POLARITY_SELECT from trigger B polarity selection */
+		uint32_t otg_trigb_manual_trig;          /* OTG_TRIGB_MANUAL_TRIG->OTG_TRIGB_MANUAL_TRIG from manual trigger B */
+
+		/* OTG Static Screen and Update Control - 6 fields */
+		uint32_t otg_static_screen_event_mask;   /* OTG_STATIC_SCREEN_CONTROL->OTG_STATIC_SCREEN_EVENT_MASK from event_triggers parameter */
+		uint32_t otg_static_screen_frame_count;  /* OTG_STATIC_SCREEN_CONTROL->OTG_STATIC_SCREEN_FRAME_COUNT from num_frames parameter */
+		uint32_t master_update_lock;             /* OTG_MASTER_UPDATE_LOCK->MASTER_UPDATE_LOCK from update lock control */
+		uint32_t master_update_mode;             /* OTG_MASTER_UPDATE_MODE->MASTER_UPDATE_MODE from update mode selection */
+		uint32_t otg_force_count_now_mode;       /* OTG_FORCE_COUNT_NOW_CNTL->OTG_FORCE_COUNT_NOW_MODE from force count mode */
+		uint32_t otg_force_count_now_clear;      /* OTG_FORCE_COUNT_NOW_CNTL->OTG_FORCE_COUNT_NOW_CLEAR from force count clear */
+
+		/* VTG Control - 3 fields */
+		uint32_t vtg0_enable;                    /* CONTROL->VTG0_ENABLE from VTG enable control */
+		uint32_t vtg0_fp2;                       /* CONTROL->VTG0_FP2 from VTG front porch 2 */
+		uint32_t vtg0_vcount_init;               /* CONTROL->VTG0_VCOUNT_INIT from VTG vertical count init */
+
+		/* OTG Status (Read-Only) - 12 fields */
+		uint32_t otg_v_blank;                    /* OTG_STATUS->OTG_V_BLANK from vertical blank status (read-only) */
+		uint32_t otg_v_active_disp;              /* OTG_STATUS->OTG_V_ACTIVE_DISP from vertical active display (read-only) */
+		uint32_t otg_frame_count;                /* OTG_STATUS_FRAME_COUNT->OTG_FRAME_COUNT from frame count (read-only) */
+		uint32_t otg_horz_count;                 /* OTG_STATUS_POSITION->OTG_HORZ_COUNT from horizontal position (read-only) */
+		uint32_t otg_vert_count;                 /* OTG_STATUS_POSITION->OTG_VERT_COUNT from vertical position (read-only) */
+		uint32_t otg_horz_count_hv;              /* OTG_STATUS_HV_COUNT->OTG_HORZ_COUNT from horizontal count (read-only) */
+		uint32_t otg_vert_count_nom;             /* OTG_STATUS_HV_COUNT->OTG_VERT_COUNT_NOM from vertical count nominal (read-only) */
+		uint32_t otg_flip_pending;               /* OTG_PIPE_UPDATE_STATUS->OTG_FLIP_PENDING from flip pending status (read-only) */
+		uint32_t otg_dc_reg_update_pending;      /* OTG_PIPE_UPDATE_STATUS->OTG_DC_REG_UPDATE_PENDING from DC register update pending (read-only) */
+		uint32_t otg_cursor_update_pending;      /* OTG_PIPE_UPDATE_STATUS->OTG_CURSOR_UPDATE_PENDING from cursor update pending (read-only) */
+		uint32_t otg_vupdate_keepout_status;     /* OTG_PIPE_UPDATE_STATUS->OTG_VUPDATE_KEEPOUT_STATUS from VUPDATE keepout status (read-only) */
+	} optc[MAX_PIPES];
+
+	/* Metadata */
+	uint32_t active_pipe_count;
+	uint32_t active_stream_count;
+	bool state_valid;
+};
+
+/**
+ * dc_capture_register_software_state() - Capture software state for register programming
+ * @dc: DC context containing current display configuration
+ * @state: Pointer to dc_register_software_state structure to populate
+ *
+ * Extracts all software state variables that are used to program hardware register
+ * fields across the display driver pipeline. This provides a complete snapshot
+ * of the software configuration that drives hardware register programming.
+ *
+ * The function traverses the DC context and extracts values from:
+ * - Stream configurations (timing, format, DSC settings)
+ * - Plane states (surface format, rotation, scaling, cursor)
+ * - Pipe contexts (resource allocation, blending, viewport)
+ * - Clock manager (display clocks, DPP clocks, pixel clocks)
+ * - Resource context (DET buffer allocation, ODM configuration)
+ *
+ * This is essential for underflow debugging as it captures the exact software
+ * state that determines how registers are programmed, allowing analysis of
+ * whether underflow is caused by incorrect register programming or timing issues.
+ *
+ * Return: true if state was successfully captured, false on error
+ */
+bool dc_capture_register_software_state(struct dc *dc, struct dc_register_software_state *state);
+
 #endif /* DC_INTERFACE_H_ */
diff --git a/drivers/gpu/drm/amd/display/dc/dc_bios_types.h b/drivers/gpu/drm/amd/display/dc/dc_bios_types.h
index 67abda44eb1f..40d7a7d83c40 100644
--- a/drivers/gpu/drm/amd/display/dc/dc_bios_types.h
+++ b/drivers/gpu/drm/amd/display/dc/dc_bios_types.h
@@ -91,9 +91,17 @@ struct dc_vbios_funcs {
 		struct device_id id);
 	/* COMMANDS */
 
+	enum bp_result (*select_crtc_source)(
+		struct dc_bios *bios,
+		struct bp_crtc_source_select *bp_params);
 	enum bp_result (*encoder_control)(
 		struct dc_bios *bios,
 		struct bp_encoder_control *cntl);
+	enum bp_result (*dac_load_detection)(
+		struct dc_bios *bios,
+		enum engine_id engine_id,
+		enum dal_device_type device_type,
+		uint32_t enum_id);
 	enum bp_result (*transmitter_control)(
 		struct dc_bios *bios,
 		struct bp_transmitter_control *cntl);
@@ -140,7 +148,8 @@ struct dc_vbios_funcs {
 	enum bp_result (*enable_lvtma_control)(
 		struct dc_bios *bios,
 		uint8_t uc_pwr_on,
-		uint8_t panel_instance);
+		uint8_t pwrseq_instance,
+		uint8_t bypass_panel_control_wait);
 
 	enum bp_result (*get_soc_bb_info)(
 		struct dc_bios *dcb,
@@ -156,9 +165,15 @@ struct dc_vbios_funcs {
 	enum bp_result (*get_lttpr_interop)(
 			struct dc_bios *dcb,
 			uint8_t *dce_caps);
+
+	enum bp_result (*get_connector_speed_cap_info)(
+		struct dc_bios *bios,
+		struct graphics_object_id object_id,
+		struct bp_connector_speed_cap_info *info);
 };
 
 struct bios_registers {
+	uint32_t BIOS_SCRATCH_0;
 	uint32_t BIOS_SCRATCH_3;
 	uint32_t BIOS_SCRATCH_6;
 };
@@ -177,6 +192,7 @@ struct dc_bios {
 	struct dc_firmware_info fw_info;
 	bool fw_info_valid;
 	struct dc_vram_info vram_info;
+	struct bp_soc_bb_info bb_info;
 	struct dc_golden_table golden_table;
 };
 
diff --git a/drivers/gpu/drm/amd/display/dc/dc_ddc_types.h b/drivers/gpu/drm/amd/display/dc/dc_ddc_types.h
index 7769bd099a5a..428e3a9ab65a 100644
--- a/drivers/gpu/drm/amd/display/dc/dc_ddc_types.h
+++ b/drivers/gpu/drm/amd/display/dc/dc_ddc_types.h
@@ -77,6 +77,32 @@ struct aux_reply_transaction_data {
 	uint8_t *data;
 };
 
+struct aux_payload {
+	/* set following flag to read/write I2C data,
+	 * reset it to read/write DPCD data */
+	bool i2c_over_aux;
+	/* set following flag to write data,
+	 * reset it to read data */
+	bool write;
+	bool mot;
+	bool write_status_update;
+
+	uint32_t address;
+	uint32_t length;
+	uint8_t *data;
+	/*
+	 * used to return the reply type of the transaction
+	 * ignored if NULL
+	 */
+	uint8_t *reply;
+	/* expressed in milliseconds
+	 * zero means "use default value"
+	 */
+	uint32_t defer_delay;
+
+};
+#define DEFAULT_AUX_MAX_DATA_SIZE 16
+
 struct i2c_payload {
 	bool write;
 	uint8_t address;
@@ -90,6 +116,8 @@ enum i2c_command_engine {
 	I2C_COMMAND_ENGINE_HW
 };
 
+#define DDC_I2C_COMMAND_ENGINE I2C_COMMAND_ENGINE_SW
+
 struct i2c_command {
 	struct i2c_payload *payloads;
 	uint8_t number_of_payloads;
@@ -150,6 +178,9 @@ enum display_dongle_type {
 	DISPLAY_DONGLE_DP_HDMI_MISMATCHED_DONGLE,
 };
 
+#define DC_MAX_EDID_BUFFER_SIZE 2048
+#define DC_EDID_BLOCK_SIZE 128
+
 struct ddc_service {
 	struct ddc *ddc_pin;
 	struct ddc_flags flags;
diff --git a/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c b/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c
index 360f3199ea6f..7b09af1cb306 100644
--- a/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c
+++ b/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c
@@ -23,13 +23,23 @@
  *
  */
 
+#include "dm_services.h"
 #include "dc.h"
 #include "dc_dmub_srv.h"
 #include "../dmub/dmub_srv.h"
 #include "dm_helpers.h"
+#include "dc_hw_types.h"
+#include "core_types.h"
+#include "../basics/conversion.h"
+#include "cursor_reg_cache.h"
+#include "resource.h"
+#include "clk_mgr.h"
+#include "dc_state_priv.h"
+#include "dc_plane_priv.h"
 
 #define CTX dc_dmub_srv->ctx
 #define DC_LOGGER CTX->logger
+#define GPINT_RETRY_NUM 20
 
 static void dc_dmub_srv_construct(struct dc_dmub_srv *dc_srv, struct dc *dc,
 				  struct dmub_srv *dmub)
@@ -61,129 +71,273 @@ void dc_dmub_srv_destroy(struct dc_dmub_srv **dmub_srv)
 	}
 }
 
-void dc_dmub_srv_cmd_queue(struct dc_dmub_srv *dc_dmub_srv,
-			   union dmub_rb_cmd *cmd)
+bool dc_dmub_srv_wait_for_pending(struct dc_dmub_srv *dc_dmub_srv)
 {
-	struct dmub_srv *dmub = dc_dmub_srv->dmub;
-	struct dc_context *dc_ctx = dc_dmub_srv->ctx;
+	struct dmub_srv *dmub;
+	struct dc_context *dc_ctx;
 	enum dmub_status status;
 
-	status = dmub_srv_cmd_queue(dmub, cmd);
-	if (status == DMUB_STATUS_OK)
-		return;
+	if (!dc_dmub_srv || !dc_dmub_srv->dmub)
+		return false;
 
-	if (status != DMUB_STATUS_QUEUE_FULL)
-		goto error;
+	dc_ctx = dc_dmub_srv->ctx;
+	dmub = dc_dmub_srv->dmub;
 
-	/* Execute and wait for queue to become empty again. */
-	dc_dmub_srv_cmd_execute(dc_dmub_srv);
-	dc_dmub_srv_wait_idle(dc_dmub_srv);
+	do {
+		status = dmub_srv_wait_for_pending(dmub, 100000);
+	} while (dc_dmub_srv->ctx->dc->debug.disable_timeout && status != DMUB_STATUS_OK);
 
-	/* Requeue the command. */
-	status = dmub_srv_cmd_queue(dmub, cmd);
-	if (status == DMUB_STATUS_OK)
-		return;
+	if (status != DMUB_STATUS_OK) {
+		DC_ERROR("Error waiting for DMUB idle: status=%d\n", status);
+		dc_dmub_srv_log_diagnostic_data(dc_dmub_srv);
+	}
 
-error:
-	DC_ERROR("Error queuing DMUB command: status=%d\n", status);
-	dc_dmub_srv_log_diagnostic_data(dc_dmub_srv);
+	return status == DMUB_STATUS_OK;
 }
 
-void dc_dmub_srv_cmd_execute(struct dc_dmub_srv *dc_dmub_srv)
+void dc_dmub_srv_clear_inbox0_ack(struct dc_dmub_srv *dc_dmub_srv)
 {
 	struct dmub_srv *dmub = dc_dmub_srv->dmub;
 	struct dc_context *dc_ctx = dc_dmub_srv->ctx;
-	enum dmub_status status;
+	enum dmub_status status = DMUB_STATUS_OK;
 
-	status = dmub_srv_cmd_execute(dmub);
+	status = dmub_srv_clear_inbox0_ack(dmub);
 	if (status != DMUB_STATUS_OK) {
-		DC_ERROR("Error starting DMUB execution: status=%d\n", status);
+		DC_ERROR("Error clearing INBOX0 ack: status=%d\n", status);
 		dc_dmub_srv_log_diagnostic_data(dc_dmub_srv);
 	}
 }
 
-void dc_dmub_srv_wait_idle(struct dc_dmub_srv *dc_dmub_srv)
+void dc_dmub_srv_wait_for_inbox0_ack(struct dc_dmub_srv *dc_dmub_srv)
 {
 	struct dmub_srv *dmub = dc_dmub_srv->dmub;
 	struct dc_context *dc_ctx = dc_dmub_srv->ctx;
-	enum dmub_status status;
+	enum dmub_status status = DMUB_STATUS_OK;
 
-	status = dmub_srv_wait_for_idle(dmub, 100000);
+	status = dmub_srv_wait_for_inbox0_ack(dmub, 100000);
 	if (status != DMUB_STATUS_OK) {
-		DC_ERROR("Error waiting for DMUB idle: status=%d\n", status);
+		DC_ERROR("Error waiting for INBOX0 HW Lock Ack\n");
 		dc_dmub_srv_log_diagnostic_data(dc_dmub_srv);
 	}
 }
 
-void dc_dmub_srv_send_inbox0_cmd(struct dc_dmub_srv *dmub_srv,
-		union dmub_inbox0_data_register data)
+void dc_dmub_srv_send_inbox0_cmd(struct dc_dmub_srv *dc_dmub_srv,
+				 union dmub_inbox0_data_register data)
 {
-	struct dmub_srv *dmub = dmub_srv->dmub;
-	if (dmub->hw_funcs.send_inbox0_cmd)
-		dmub->hw_funcs.send_inbox0_cmd(dmub, data);
-	// TODO: Add wait command -- poll register for ACK
+	struct dmub_srv *dmub = dc_dmub_srv->dmub;
+	struct dc_context *dc_ctx = dc_dmub_srv->ctx;
+	enum dmub_status status = DMUB_STATUS_OK;
+
+	status = dmub_srv_send_inbox0_cmd(dmub, data);
+	if (status != DMUB_STATUS_OK) {
+		DC_ERROR("Error sending INBOX0 cmd\n");
+		dc_dmub_srv_log_diagnostic_data(dc_dmub_srv);
+	}
+}
+
+static bool dc_dmub_srv_reg_cmd_list_queue_execute(struct dc_dmub_srv *dc_dmub_srv,
+		unsigned int count,
+		union dmub_rb_cmd *cmd_list)
+{
+	struct dc_context *dc_ctx;
+	struct dmub_srv *dmub;
+	enum dmub_status status = DMUB_STATUS_OK;
+	int i;
+
+	if (!dc_dmub_srv || !dc_dmub_srv->dmub)
+		return false;
+
+	dc_ctx = dc_dmub_srv->ctx;
+	dmub = dc_dmub_srv->dmub;
+
+	for (i = 0 ; i < count; i++) {
+		/* confirm no messages pending */
+		do {
+			status = dmub_srv_wait_for_idle(dmub, 100000);
+		} while (dc_dmub_srv->ctx->dc->debug.disable_timeout && status != DMUB_STATUS_OK);
+
+		/* queue command */
+		if (status == DMUB_STATUS_OK)
+			status = dmub_srv_reg_cmd_execute(dmub, &cmd_list[i]);
+
+		/* check for errors */
+		if (status != DMUB_STATUS_OK) {
+			break;
+		}
+	}
+
+	if (status != DMUB_STATUS_OK) {
+		if (status != DMUB_STATUS_POWER_STATE_D3) {
+			DC_ERROR("Error starting DMUB execution: status=%d\n", status);
+			dc_dmub_srv_log_diagnostic_data(dc_dmub_srv);
+		}
+		return false;
+	}
+
+	return true;
 }
 
-bool dc_dmub_srv_cmd_with_reply_data(struct dc_dmub_srv *dc_dmub_srv, union dmub_rb_cmd *cmd)
+static bool dc_dmub_srv_fb_cmd_list_queue_execute(struct dc_dmub_srv *dc_dmub_srv,
+		unsigned int count,
+		union dmub_rb_cmd *cmd_list)
 {
+	struct dc_context *dc_ctx;
 	struct dmub_srv *dmub;
 	enum dmub_status status;
+	int i;
 
 	if (!dc_dmub_srv || !dc_dmub_srv->dmub)
 		return false;
 
+	dc_ctx = dc_dmub_srv->ctx;
 	dmub = dc_dmub_srv->dmub;
 
-	status = dmub_srv_cmd_with_reply_data(dmub, cmd);
+	for (i = 0 ; i < count; i++) {
+		// Queue command
+		if (!cmd_list[i].cmd_common.header.multi_cmd_pending ||
+				dmub_rb_num_free(&dmub->inbox1.rb) >= count - i) {
+			status = dmub_srv_fb_cmd_queue(dmub, &cmd_list[i]);
+		} else {
+			status = DMUB_STATUS_QUEUE_FULL;
+		}
+
+		if (status == DMUB_STATUS_QUEUE_FULL) {
+			/* Execute and wait for queue to become empty again. */
+			status = dmub_srv_fb_cmd_execute(dmub);
+			if (status == DMUB_STATUS_POWER_STATE_D3)
+				return false;
+
+			do {
+					status = dmub_srv_wait_for_inbox_free(dmub, 100000, count - i);
+			} while (dc_dmub_srv->ctx->dc->debug.disable_timeout && status != DMUB_STATUS_OK);
+
+			/* Requeue the command. */
+			status = dmub_srv_fb_cmd_queue(dmub, &cmd_list[i]);
+		}
+
+		if (status != DMUB_STATUS_OK) {
+			if (status != DMUB_STATUS_POWER_STATE_D3) {
+				DC_ERROR("Error queueing DMUB command: status=%d\n", status);
+				dc_dmub_srv_log_diagnostic_data(dc_dmub_srv);
+			}
+			return false;
+		}
+	}
+
+	status = dmub_srv_fb_cmd_execute(dmub);
 	if (status != DMUB_STATUS_OK) {
-		DC_LOG_DEBUG("No reply for DMUB command: status=%d\n", status);
+		if (status != DMUB_STATUS_POWER_STATE_D3) {
+			DC_ERROR("Error starting DMUB execution: status=%d\n", status);
+			dc_dmub_srv_log_diagnostic_data(dc_dmub_srv);
+		}
 		return false;
 	}
 
 	return true;
 }
 
-void dc_dmub_srv_wait_phy_init(struct dc_dmub_srv *dc_dmub_srv)
+bool dc_dmub_srv_cmd_list_queue_execute(struct dc_dmub_srv *dc_dmub_srv,
+		unsigned int count,
+		union dmub_rb_cmd *cmd_list)
 {
-	struct dmub_srv *dmub = dc_dmub_srv->dmub;
-	struct dc_context *dc_ctx = dc_dmub_srv->ctx;
+	bool res = false;
+
+	if (dc_dmub_srv && dc_dmub_srv->dmub) {
+		if (dc_dmub_srv->dmub->inbox_type == DMUB_CMD_INTERFACE_REG) {
+			res = dc_dmub_srv_reg_cmd_list_queue_execute(dc_dmub_srv, count, cmd_list);
+		} else {
+			res = dc_dmub_srv_fb_cmd_list_queue_execute(dc_dmub_srv, count, cmd_list);
+		}
+
+		if (res)
+			res = dmub_srv_update_inbox_status(dc_dmub_srv->dmub) == DMUB_STATUS_OK;
+	}
+
+	return res;
+}
+
+bool dc_dmub_srv_wait_for_idle(struct dc_dmub_srv *dc_dmub_srv,
+		enum dm_dmub_wait_type wait_type,
+		union dmub_rb_cmd *cmd_list)
+{
+	struct dmub_srv *dmub;
 	enum dmub_status status;
 
-	for (;;) {
-		/* Wait up to a second for PHY init. */
-		status = dmub_srv_wait_for_phy_init(dmub, 1000000);
-		if (status == DMUB_STATUS_OK)
-			/* Initialization OK */
-			break;
+	if (!dc_dmub_srv || !dc_dmub_srv->dmub)
+		return false;
 
-		DC_ERROR("DMCUB PHY init failed: status=%d\n", status);
-		ASSERT(0);
+	dmub = dc_dmub_srv->dmub;
 
-		if (status != DMUB_STATUS_TIMEOUT)
-			/*
-			 * Server likely initialized or we don't have
-			 * DMCUB HW support - this won't end.
-			 */
-			break;
+	// Wait for DMUB to process command
+	if (wait_type != DM_DMUB_WAIT_TYPE_NO_WAIT) {
+		do {
+			status = dmub_srv_wait_for_idle(dmub, 100000);
+		} while (dc_dmub_srv->ctx->dc->debug.disable_timeout && status != DMUB_STATUS_OK);
+
+		if (status != DMUB_STATUS_OK) {
+			DC_LOG_DEBUG("No reply for DMUB command: status=%d\n", status);
+			if (!dmub->debug.timeout_info.timeout_occured) {
+				dmub->debug.timeout_info.timeout_occured = true;
+				if (cmd_list)
+					dmub->debug.timeout_info.timeout_cmd = *cmd_list;
+				dmub->debug.timeout_info.timestamp = dm_get_timestamp(dc_dmub_srv->ctx);
+			}
+			dc_dmub_srv_log_diagnostic_data(dc_dmub_srv);
+			return false;
+		}
 
-		/* Continue spinning so we don't hang the ASIC. */
+		// Copy data back from ring buffer into command
+		if (wait_type == DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY && cmd_list) {
+			dmub_srv_cmd_get_response(dc_dmub_srv->dmub, cmd_list);
+		}
 	}
+
+	return true;
 }
 
-bool dc_dmub_srv_notify_stream_mask(struct dc_dmub_srv *dc_dmub_srv,
-				    unsigned int stream_mask)
+bool dc_dmub_srv_cmd_run(struct dc_dmub_srv *dc_dmub_srv, union dmub_rb_cmd *cmd, enum dm_dmub_wait_type wait_type)
+{
+	return dc_dmub_srv_cmd_run_list(dc_dmub_srv, 1, cmd, wait_type);
+}
+
+bool dc_dmub_srv_cmd_run_list(struct dc_dmub_srv *dc_dmub_srv, unsigned int count, union dmub_rb_cmd *cmd_list, enum dm_dmub_wait_type wait_type)
+{
+	if (!dc_dmub_srv_cmd_list_queue_execute(dc_dmub_srv, count, cmd_list))
+		return false;
+
+	return dc_dmub_srv_wait_for_idle(dc_dmub_srv, wait_type, cmd_list);
+}
+
+bool dc_dmub_srv_optimized_init_done(struct dc_dmub_srv *dc_dmub_srv)
 {
 	struct dmub_srv *dmub;
-	const uint32_t timeout = 30;
+	struct dc_context *dc_ctx;
+	union dmub_fw_boot_status boot_status;
+	enum dmub_status status;
 
 	if (!dc_dmub_srv || !dc_dmub_srv->dmub)
 		return false;
 
 	dmub = dc_dmub_srv->dmub;
+	dc_ctx = dc_dmub_srv->ctx;
+
+	status = dmub_srv_get_fw_boot_status(dmub, &boot_status);
+	if (status != DMUB_STATUS_OK) {
+		DC_ERROR("Error querying DMUB boot status: error=%d\n", status);
+		return false;
+	}
+
+	return boot_status.bits.optimized_init_done;
+}
 
-	return dmub_srv_send_gpint_command(
-		       dmub, DMUB_GPINT__IDLE_OPT_NOTIFY_STREAM_MASK,
-		       stream_mask, timeout) == DMUB_STATUS_OK;
+bool dc_dmub_srv_notify_stream_mask(struct dc_dmub_srv *dc_dmub_srv,
+				    unsigned int stream_mask)
+{
+	if (!dc_dmub_srv || !dc_dmub_srv->dmub)
+		return false;
+
+	return dc_wake_and_execute_gpint(dc_dmub_srv->ctx, DMUB_GPINT__IDLE_OPT_NOTIFY_STREAM_MASK,
+					 stream_mask, NULL, DM_DMUB_WAIT_TYPE_WAIT);
 }
 
 bool dc_dmub_srv_is_restore_required(struct dc_dmub_srv *dc_dmub_srv)
@@ -219,93 +373,1998 @@ void dc_dmub_trace_event_control(struct dc *dc, bool enable)
 	dm_helpers_dmub_outbox_interrupt_control(dc->ctx, enable);
 }
 
-bool dc_dmub_srv_get_diagnostic_data(struct dc_dmub_srv *dc_dmub_srv, struct dmub_diagnostic_data *diag_data)
+void dc_dmub_srv_drr_update_cmd(struct dc *dc, uint32_t tg_inst, uint32_t vtotal_min, uint32_t vtotal_max)
 {
-	if (!dc_dmub_srv || !dc_dmub_srv->dmub || !diag_data)
+	union dmub_rb_cmd cmd = { 0 };
+
+	cmd.drr_update.header.type = DMUB_CMD__FW_ASSISTED_MCLK_SWITCH;
+	cmd.drr_update.header.sub_type = DMUB_CMD__FAMS_DRR_UPDATE;
+	cmd.drr_update.dmub_optc_state_req.v_total_max = vtotal_max;
+	cmd.drr_update.dmub_optc_state_req.v_total_min = vtotal_min;
+	cmd.drr_update.dmub_optc_state_req.tg_inst = tg_inst;
+
+	cmd.drr_update.header.payload_bytes = sizeof(cmd.drr_update) - sizeof(cmd.drr_update.header);
+
+	// Send the command to the DMCUB.
+	dc_wake_and_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
+}
+
+void dc_dmub_srv_set_drr_manual_trigger_cmd(struct dc *dc, uint32_t tg_inst)
+{
+	union dmub_rb_cmd cmd = { 0 };
+
+	cmd.drr_update.header.type = DMUB_CMD__FW_ASSISTED_MCLK_SWITCH;
+	cmd.drr_update.header.sub_type = DMUB_CMD__FAMS_SET_MANUAL_TRIGGER;
+	cmd.drr_update.dmub_optc_state_req.tg_inst = tg_inst;
+
+	cmd.drr_update.header.payload_bytes = sizeof(cmd.drr_update) - sizeof(cmd.drr_update.header);
+
+	// Send the command to the DMCUB.
+	dc_wake_and_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
+}
+
+static uint8_t dc_dmub_srv_get_pipes_for_stream(struct dc *dc, struct dc_stream_state *stream)
+{
+	uint8_t pipes = 0;
+	int i = 0;
+
+	for (i = 0; i < MAX_PIPES; i++) {
+		struct pipe_ctx *pipe = &dc->current_state->res_ctx.pipe_ctx[i];
+
+		if (pipe->stream == stream && pipe->stream_res.tg)
+			pipes = i;
+	}
+	return pipes;
+}
+
+static void dc_dmub_srv_populate_fams_pipe_info(struct dc *dc, struct dc_state *context,
+		struct pipe_ctx *head_pipe,
+		struct dmub_cmd_fw_assisted_mclk_switch_pipe_data *fams_pipe_data)
+{
+	int j;
+	int pipe_idx = 0;
+
+	fams_pipe_data->pipe_index[pipe_idx++] = head_pipe->plane_res.hubp->inst;
+	for (j = 0; j < dc->res_pool->pipe_count; j++) {
+		struct pipe_ctx *split_pipe = &context->res_ctx.pipe_ctx[j];
+
+		if (split_pipe->stream == head_pipe->stream && (split_pipe->top_pipe || split_pipe->prev_odm_pipe)) {
+			fams_pipe_data->pipe_index[pipe_idx++] = split_pipe->plane_res.hubp->inst;
+		}
+	}
+	fams_pipe_data->pipe_count = pipe_idx;
+}
+
+bool dc_dmub_srv_p_state_delegate(struct dc *dc, bool should_manage_pstate, struct dc_state *context)
+{
+	union dmub_rb_cmd cmd = { 0 };
+	struct dmub_cmd_fw_assisted_mclk_switch_config *config_data = &cmd.fw_assisted_mclk_switch.config_data;
+	int i = 0, k = 0;
+	int ramp_up_num_steps = 1; // TODO: Ramp is currently disabled. Reenable it.
+	uint8_t visual_confirm_enabled;
+	struct dc_stream_status *stream_status = NULL;
+
+	if (dc == NULL)
 		return false;
-	return dmub_srv_get_diagnostic_data(dc_dmub_srv->dmub, diag_data);
+
+	visual_confirm_enabled = dc->debug.visual_confirm == VISUAL_CONFIRM_FAMS;
+
+	// Format command.
+	cmd.fw_assisted_mclk_switch.header.type = DMUB_CMD__FW_ASSISTED_MCLK_SWITCH;
+	cmd.fw_assisted_mclk_switch.header.sub_type = DMUB_CMD__FAMS_SETUP_FW_CTRL;
+	cmd.fw_assisted_mclk_switch.config_data.fams_enabled = should_manage_pstate;
+	cmd.fw_assisted_mclk_switch.config_data.visual_confirm_enabled = visual_confirm_enabled;
+
+	if (should_manage_pstate) {
+		for (i = 0; i < dc->res_pool->pipe_count; i++) {
+			struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
+
+			if (!pipe->stream)
+				continue;
+
+			/* If FAMS is being used to support P-State and there is a stream
+			 * that does not use FAMS, we are in an FPO + VActive scenario.
+			 * Assign vactive stretch margin in this case.
+			 */
+			stream_status = dc_state_get_stream_status(context, pipe->stream);
+			if (stream_status && !stream_status->fpo_in_use) {
+				cmd.fw_assisted_mclk_switch.config_data.vactive_stretch_margin_us = dc->debug.fpo_vactive_margin_us;
+				break;
+			}
+		}
+	}
+
+	for (i = 0, k = 0; context && i < dc->res_pool->pipe_count; i++) {
+		struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
+
+		if (!resource_is_pipe_type(pipe, OTG_MASTER))
+			continue;
+
+		stream_status = dc_state_get_stream_status(context, pipe->stream);
+		if (stream_status && stream_status->fpo_in_use) {
+			struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
+			uint8_t min_refresh_in_hz = (pipe->stream->timing.min_refresh_in_uhz + 999999) / 1000000;
+
+			config_data->pipe_data[k].pix_clk_100hz = pipe->stream->timing.pix_clk_100hz;
+			config_data->pipe_data[k].min_refresh_in_hz = min_refresh_in_hz;
+			config_data->pipe_data[k].max_ramp_step = ramp_up_num_steps;
+			config_data->pipe_data[k].pipes = dc_dmub_srv_get_pipes_for_stream(dc, pipe->stream);
+			dc_dmub_srv_populate_fams_pipe_info(dc, context, pipe, &config_data->pipe_data[k]);
+			k++;
+		}
+	}
+	cmd.fw_assisted_mclk_switch.header.payload_bytes =
+		sizeof(cmd.fw_assisted_mclk_switch) - sizeof(cmd.fw_assisted_mclk_switch.header);
+
+	// Send the command to the DMCUB.
+	dc_wake_and_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
+
+	return true;
+}
+
+void dc_dmub_srv_query_caps_cmd(struct dc_dmub_srv *dc_dmub_srv)
+{
+	union dmub_rb_cmd cmd = { 0 };
+
+	if (dc_dmub_srv->ctx->dc->debug.dmcub_emulation)
+		return;
+
+	memset(&cmd, 0, sizeof(cmd));
+
+	/* Prepare fw command */
+	cmd.query_feature_caps.header.type = DMUB_CMD__QUERY_FEATURE_CAPS;
+	cmd.query_feature_caps.header.sub_type = 0;
+	cmd.query_feature_caps.header.ret_status = 1;
+	cmd.query_feature_caps.header.payload_bytes = sizeof(struct dmub_cmd_query_feature_caps_data);
+
+	/* If command was processed, copy feature caps to dmub srv */
+	if (dc_wake_and_execute_dmub_cmd(dc_dmub_srv->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY) &&
+	    cmd.query_feature_caps.header.ret_status == 0) {
+		memcpy(&dc_dmub_srv->dmub->feature_caps,
+		       &cmd.query_feature_caps.query_feature_caps_data,
+		       sizeof(struct dmub_feature_caps));
+	}
+}
+
+void dc_dmub_srv_get_visual_confirm_color_cmd(struct dc *dc, struct pipe_ctx *pipe_ctx)
+{
+	union dmub_rb_cmd cmd = { 0 };
+	unsigned int panel_inst = 0;
+
+	if (!dc_get_edp_link_panel_inst(dc, pipe_ctx->stream->link, &panel_inst) &&
+			dc->debug.visual_confirm == VISUAL_CONFIRM_DISABLE)
+		return;
+
+	memset(&cmd, 0, sizeof(cmd));
+
+	// Prepare fw command
+	cmd.visual_confirm_color.header.type = DMUB_CMD__GET_VISUAL_CONFIRM_COLOR;
+	cmd.visual_confirm_color.header.sub_type = 0;
+	cmd.visual_confirm_color.header.ret_status = 1;
+	cmd.visual_confirm_color.header.payload_bytes = sizeof(struct dmub_cmd_visual_confirm_color_data);
+	cmd.visual_confirm_color.visual_confirm_color_data.visual_confirm_color.panel_inst = panel_inst;
+
+	// If command was processed, copy feature caps to dmub srv
+	if (dc_wake_and_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY) &&
+		cmd.visual_confirm_color.header.ret_status == 0) {
+		memcpy(&dc->ctx->dmub_srv->dmub->visual_confirm_color,
+			&cmd.visual_confirm_color.visual_confirm_color_data,
+			sizeof(struct dmub_visual_confirm_color));
+	}
+}
+
+/**
+ * populate_subvp_cmd_drr_info - Helper to populate DRR pipe info for the DMCUB subvp command
+ *
+ * @dc: [in] pointer to dc object
+ * @subvp_pipe: [in] pipe_ctx for the SubVP pipe
+ * @vblank_pipe: [in] pipe_ctx for the DRR pipe
+ * @pipe_data: [in] Pipe data which stores the VBLANK/DRR info
+ * @context: [in] DC state for access to phantom stream
+ *
+ * Populate the DMCUB SubVP command with DRR pipe info. All the information
+ * required for calculating the SubVP + DRR microschedule is populated here.
+ *
+ * High level algorithm:
+ * 1. Get timing for SubVP pipe, phantom pipe, and DRR pipe
+ * 2. Calculate the min and max vtotal which supports SubVP + DRR microschedule
+ * 3. Populate the drr_info with the min and max supported vtotal values
+ */
+static void populate_subvp_cmd_drr_info(struct dc *dc,
+		struct dc_state *context,
+		struct pipe_ctx *subvp_pipe,
+		struct pipe_ctx *vblank_pipe,
+		struct dmub_cmd_fw_assisted_mclk_switch_pipe_data_v2 *pipe_data)
+{
+	struct dc_stream_state *phantom_stream = dc_state_get_paired_subvp_stream(context, subvp_pipe->stream);
+	struct dc_crtc_timing *main_timing = &subvp_pipe->stream->timing;
+	struct dc_crtc_timing *phantom_timing;
+	struct dc_crtc_timing *drr_timing = &vblank_pipe->stream->timing;
+	uint16_t drr_frame_us = 0;
+	uint16_t min_drr_supported_us = 0;
+	uint16_t max_drr_supported_us = 0;
+	uint16_t max_drr_vblank_us = 0;
+	uint16_t max_drr_mallregion_us = 0;
+	uint16_t mall_region_us = 0;
+	uint16_t prefetch_us = 0;
+	uint16_t subvp_active_us = 0;
+	uint16_t drr_active_us = 0;
+	uint16_t min_vtotal_supported = 0;
+	uint16_t max_vtotal_supported = 0;
+
+	if (!phantom_stream)
+		return;
+
+	phantom_timing = &phantom_stream->timing;
+
+	pipe_data->pipe_config.vblank_data.drr_info.drr_in_use = true;
+	pipe_data->pipe_config.vblank_data.drr_info.use_ramping = false; // for now don't use ramping
+	pipe_data->pipe_config.vblank_data.drr_info.drr_window_size_ms = 4; // hardcode 4ms DRR window for now
+
+	drr_frame_us = div64_u64(((uint64_t)drr_timing->v_total * drr_timing->h_total * 1000000),
+			(((uint64_t)drr_timing->pix_clk_100hz * 100)));
+	// P-State allow width and FW delays already included phantom_timing->v_addressable
+	mall_region_us = div64_u64(((uint64_t)phantom_timing->v_addressable * phantom_timing->h_total * 1000000),
+			(((uint64_t)phantom_timing->pix_clk_100hz * 100)));
+	min_drr_supported_us = drr_frame_us + mall_region_us + SUBVP_DRR_MARGIN_US;
+	min_vtotal_supported = div64_u64(((uint64_t)drr_timing->pix_clk_100hz * 100 * min_drr_supported_us),
+			(((uint64_t)drr_timing->h_total * 1000000)));
+
+	prefetch_us = div64_u64(((uint64_t)(phantom_timing->v_total - phantom_timing->v_front_porch) * phantom_timing->h_total * 1000000),
+			(((uint64_t)phantom_timing->pix_clk_100hz * 100) + dc->caps.subvp_prefetch_end_to_mall_start_us));
+	subvp_active_us = div64_u64(((uint64_t)main_timing->v_addressable * main_timing->h_total * 1000000),
+			(((uint64_t)main_timing->pix_clk_100hz * 100)));
+	drr_active_us = div64_u64(((uint64_t)drr_timing->v_addressable * drr_timing->h_total * 1000000),
+			(((uint64_t)drr_timing->pix_clk_100hz * 100)));
+	max_drr_vblank_us = div64_u64((subvp_active_us - prefetch_us -
+			dc->caps.subvp_fw_processing_delay_us - drr_active_us), 2) + drr_active_us;
+	max_drr_mallregion_us = subvp_active_us - prefetch_us - mall_region_us - dc->caps.subvp_fw_processing_delay_us;
+	max_drr_supported_us = max_drr_vblank_us > max_drr_mallregion_us ? max_drr_vblank_us : max_drr_mallregion_us;
+	max_vtotal_supported = div64_u64(((uint64_t)drr_timing->pix_clk_100hz * 100 * max_drr_supported_us),
+			(((uint64_t)drr_timing->h_total * 1000000)));
+
+	/* When calculating the max vtotal supported for SubVP + DRR cases, add
+	 * margin due to possible rounding errors (being off by 1 line in the
+	 * FW calculation can incorrectly push the P-State switch to wait 1 frame
+	 * longer).
+	 */
+	max_vtotal_supported = max_vtotal_supported - dc->caps.subvp_drr_max_vblank_margin_us;
+
+	pipe_data->pipe_config.vblank_data.drr_info.min_vtotal_supported = min_vtotal_supported;
+	pipe_data->pipe_config.vblank_data.drr_info.max_vtotal_supported = max_vtotal_supported;
+	pipe_data->pipe_config.vblank_data.drr_info.drr_vblank_start_margin = dc->caps.subvp_drr_vblank_start_margin_us;
+}
+
+/**
+ * populate_subvp_cmd_vblank_pipe_info - Helper to populate VBLANK pipe info for the DMUB subvp command
+ *
+ * @dc: [in] current dc state
+ * @context: [in] new dc state
+ * @cmd: [in] DMUB cmd to be populated with SubVP info
+ * @vblank_pipe: [in] pipe_ctx for the VBLANK pipe
+ * @cmd_pipe_index: [in] index for the pipe array in DMCUB SubVP cmd
+ *
+ * Populate the DMCUB SubVP command with VBLANK pipe info. All the information
+ * required to calculate the microschedule for SubVP + VBLANK case is stored in
+ * the pipe_data (subvp_data and vblank_data).  Also check if the VBLANK pipe
+ * is a DRR display -- if it is make a call to populate drr_info.
+ */
+static void populate_subvp_cmd_vblank_pipe_info(struct dc *dc,
+		struct dc_state *context,
+		union dmub_rb_cmd *cmd,
+		struct pipe_ctx *vblank_pipe,
+		uint8_t cmd_pipe_index)
+{
+	uint32_t i;
+	struct pipe_ctx *pipe = NULL;
+	struct dmub_cmd_fw_assisted_mclk_switch_pipe_data_v2 *pipe_data =
+			&cmd->fw_assisted_mclk_switch_v2.config_data.pipe_data[cmd_pipe_index];
+
+	// Find the SubVP pipe
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		pipe = &context->res_ctx.pipe_ctx[i];
+
+		// We check for master pipe, but it shouldn't matter since we only need
+		// the pipe for timing info (stream should be same for any pipe splits)
+		if (!resource_is_pipe_type(pipe, OTG_MASTER) ||
+				!resource_is_pipe_type(pipe, DPP_PIPE))
+			continue;
+
+		// Find the SubVP pipe
+		if (dc_state_get_pipe_subvp_type(context, pipe) == SUBVP_MAIN)
+			break;
+	}
+
+	pipe_data->mode = VBLANK;
+	pipe_data->pipe_config.vblank_data.pix_clk_100hz = vblank_pipe->stream->timing.pix_clk_100hz;
+	pipe_data->pipe_config.vblank_data.vblank_start = vblank_pipe->stream->timing.v_total -
+							vblank_pipe->stream->timing.v_front_porch;
+	pipe_data->pipe_config.vblank_data.vtotal = vblank_pipe->stream->timing.v_total;
+	pipe_data->pipe_config.vblank_data.htotal = vblank_pipe->stream->timing.h_total;
+	pipe_data->pipe_config.vblank_data.vblank_pipe_index = vblank_pipe->pipe_idx;
+	pipe_data->pipe_config.vblank_data.vstartup_start = vblank_pipe->pipe_dlg_param.vstartup_start;
+	pipe_data->pipe_config.vblank_data.vblank_end =
+			vblank_pipe->stream->timing.v_total - vblank_pipe->stream->timing.v_front_porch - vblank_pipe->stream->timing.v_addressable;
+
+	if (vblank_pipe->stream->ignore_msa_timing_param &&
+		(vblank_pipe->stream->allow_freesync || vblank_pipe->stream->vrr_active_variable || vblank_pipe->stream->vrr_active_fixed))
+		populate_subvp_cmd_drr_info(dc, context, pipe, vblank_pipe, pipe_data);
+}
+
+/**
+ * update_subvp_prefetch_end_to_mall_start - Helper for SubVP + SubVP case
+ *
+ * @dc: [in] current dc state
+ * @context: [in] new dc state
+ * @cmd: [in] DMUB cmd to be populated with SubVP info
+ * @subvp_pipes: [in] Array of SubVP pipes (should always be length 2)
+ *
+ * For SubVP + SubVP, we use a single vertical interrupt to start the
+ * microschedule for both SubVP pipes. In order for this to work correctly, the
+ * MALL REGION of both SubVP pipes must start at the same time. This function
+ * lengthens the prefetch end to mall start delay of the SubVP pipe that has
+ * the shorter prefetch so that both MALL REGION's will start at the same time.
+ */
+static void update_subvp_prefetch_end_to_mall_start(struct dc *dc,
+		struct dc_state *context,
+		union dmub_rb_cmd *cmd,
+		struct pipe_ctx *subvp_pipes[])
+{
+	uint32_t subvp0_prefetch_us = 0;
+	uint32_t subvp1_prefetch_us = 0;
+	uint32_t prefetch_delta_us = 0;
+	struct dc_stream_state *phantom_stream0 = NULL;
+	struct dc_stream_state *phantom_stream1 = NULL;
+	struct dc_crtc_timing *phantom_timing0 = NULL;
+	struct dc_crtc_timing *phantom_timing1 = NULL;
+	struct dmub_cmd_fw_assisted_mclk_switch_pipe_data_v2 *pipe_data = NULL;
+
+	phantom_stream0 = dc_state_get_paired_subvp_stream(context, subvp_pipes[0]->stream);
+	if (!phantom_stream0)
+		return;
+
+	phantom_stream1 = dc_state_get_paired_subvp_stream(context, subvp_pipes[1]->stream);
+	if (!phantom_stream1)
+		return;
+
+	phantom_timing0 = &phantom_stream0->timing;
+	phantom_timing1 = &phantom_stream1->timing;
+
+	subvp0_prefetch_us = div64_u64(((uint64_t)(phantom_timing0->v_total - phantom_timing0->v_front_porch) *
+			(uint64_t)phantom_timing0->h_total * 1000000),
+			(((uint64_t)phantom_timing0->pix_clk_100hz * 100) + dc->caps.subvp_prefetch_end_to_mall_start_us));
+	subvp1_prefetch_us = div64_u64(((uint64_t)(phantom_timing1->v_total - phantom_timing1->v_front_porch) *
+			(uint64_t)phantom_timing1->h_total * 1000000),
+			(((uint64_t)phantom_timing1->pix_clk_100hz * 100) + dc->caps.subvp_prefetch_end_to_mall_start_us));
+
+	// Whichever SubVP PIPE has the smaller prefetch (including the prefetch end to mall start time)
+	// should increase it's prefetch time to match the other
+	if (subvp0_prefetch_us > subvp1_prefetch_us) {
+		pipe_data = &cmd->fw_assisted_mclk_switch_v2.config_data.pipe_data[1];
+		prefetch_delta_us = subvp0_prefetch_us - subvp1_prefetch_us;
+		pipe_data->pipe_config.subvp_data.prefetch_to_mall_start_lines =
+				div64_u64(((uint64_t)(dc->caps.subvp_prefetch_end_to_mall_start_us + prefetch_delta_us) *
+					((uint64_t)phantom_timing1->pix_clk_100hz * 100) + ((uint64_t)phantom_timing1->h_total * 1000000 - 1)),
+					((uint64_t)phantom_timing1->h_total * 1000000));
+
+	} else if (subvp1_prefetch_us >  subvp0_prefetch_us) {
+		pipe_data = &cmd->fw_assisted_mclk_switch_v2.config_data.pipe_data[0];
+		prefetch_delta_us = subvp1_prefetch_us - subvp0_prefetch_us;
+		pipe_data->pipe_config.subvp_data.prefetch_to_mall_start_lines =
+				div64_u64(((uint64_t)(dc->caps.subvp_prefetch_end_to_mall_start_us + prefetch_delta_us) *
+					((uint64_t)phantom_timing0->pix_clk_100hz * 100) + ((uint64_t)phantom_timing0->h_total * 1000000 - 1)),
+					((uint64_t)phantom_timing0->h_total * 1000000));
+	}
+}
+
+/**
+ * populate_subvp_cmd_pipe_info - Helper to populate the SubVP pipe info for the DMUB subvp command
+ *
+ * @dc: [in] current dc state
+ * @context: [in] new dc state
+ * @cmd: [in] DMUB cmd to be populated with SubVP info
+ * @subvp_pipe: [in] pipe_ctx for the SubVP pipe
+ * @cmd_pipe_index: [in] index for the pipe array in DMCUB SubVP cmd
+ *
+ * Populate the DMCUB SubVP command with SubVP pipe info. All the information
+ * required to calculate the microschedule for the SubVP pipe is stored in the
+ * pipe_data of the DMCUB SubVP command.
+ */
+static void populate_subvp_cmd_pipe_info(struct dc *dc,
+		struct dc_state *context,
+		union dmub_rb_cmd *cmd,
+		struct pipe_ctx *subvp_pipe,
+		uint8_t cmd_pipe_index)
+{
+	uint32_t j;
+	struct dmub_cmd_fw_assisted_mclk_switch_pipe_data_v2 *pipe_data =
+			&cmd->fw_assisted_mclk_switch_v2.config_data.pipe_data[cmd_pipe_index];
+	struct dc_stream_state *phantom_stream = dc_state_get_paired_subvp_stream(context, subvp_pipe->stream);
+	struct dc_crtc_timing *main_timing = &subvp_pipe->stream->timing;
+	struct dc_crtc_timing *phantom_timing;
+	uint32_t out_num_stream, out_den_stream, out_num_plane, out_den_plane, out_num, out_den;
+
+	if (!phantom_stream)
+		return;
+
+	phantom_timing = &phantom_stream->timing;
+
+	pipe_data->mode = SUBVP;
+	pipe_data->pipe_config.subvp_data.pix_clk_100hz = subvp_pipe->stream->timing.pix_clk_100hz;
+	pipe_data->pipe_config.subvp_data.htotal = subvp_pipe->stream->timing.h_total;
+	pipe_data->pipe_config.subvp_data.vtotal = subvp_pipe->stream->timing.v_total;
+	pipe_data->pipe_config.subvp_data.main_vblank_start =
+			main_timing->v_total - main_timing->v_front_porch;
+	pipe_data->pipe_config.subvp_data.main_vblank_end =
+			main_timing->v_total - main_timing->v_front_porch - main_timing->v_addressable;
+	pipe_data->pipe_config.subvp_data.mall_region_lines = phantom_timing->v_addressable;
+	pipe_data->pipe_config.subvp_data.main_pipe_index = subvp_pipe->stream_res.tg->inst;
+	pipe_data->pipe_config.subvp_data.is_drr = subvp_pipe->stream->ignore_msa_timing_param &&
+		(subvp_pipe->stream->allow_freesync || subvp_pipe->stream->vrr_active_variable || subvp_pipe->stream->vrr_active_fixed);
+
+	/* Calculate the scaling factor from the src and dst height.
+	 * e.g. If 3840x2160 being downscaled to 1920x1080, the scaling factor is 1/2.
+	 * Reduce the fraction 1080/2160 = 1/2 for the "scaling factor"
+	 *
+	 * Make sure to combine stream and plane scaling together.
+	 */
+	reduce_fraction(subvp_pipe->stream->src.height, subvp_pipe->stream->dst.height,
+			&out_num_stream, &out_den_stream);
+	reduce_fraction(subvp_pipe->plane_state->src_rect.height, subvp_pipe->plane_state->dst_rect.height,
+			&out_num_plane, &out_den_plane);
+	reduce_fraction(out_num_stream * out_num_plane, out_den_stream * out_den_plane, &out_num, &out_den);
+	pipe_data->pipe_config.subvp_data.scale_factor_numerator = out_num;
+	pipe_data->pipe_config.subvp_data.scale_factor_denominator = out_den;
+
+	// Prefetch lines is equal to VACTIVE + BP + VSYNC
+	pipe_data->pipe_config.subvp_data.prefetch_lines =
+			phantom_timing->v_total - phantom_timing->v_front_porch;
+
+	// Round up
+	pipe_data->pipe_config.subvp_data.prefetch_to_mall_start_lines =
+			div64_u64(((uint64_t)dc->caps.subvp_prefetch_end_to_mall_start_us * ((uint64_t)phantom_timing->pix_clk_100hz * 100) +
+					((uint64_t)phantom_timing->h_total * 1000000 - 1)), ((uint64_t)phantom_timing->h_total * 1000000));
+	pipe_data->pipe_config.subvp_data.processing_delay_lines =
+			div64_u64(((uint64_t)(dc->caps.subvp_fw_processing_delay_us) * ((uint64_t)phantom_timing->pix_clk_100hz * 100) +
+					((uint64_t)phantom_timing->h_total * 1000000 - 1)), ((uint64_t)phantom_timing->h_total * 1000000));
+
+	if (subvp_pipe->bottom_pipe) {
+		pipe_data->pipe_config.subvp_data.main_split_pipe_index = subvp_pipe->bottom_pipe->pipe_idx;
+	} else if (subvp_pipe->next_odm_pipe) {
+		pipe_data->pipe_config.subvp_data.main_split_pipe_index = subvp_pipe->next_odm_pipe->pipe_idx;
+	} else {
+		pipe_data->pipe_config.subvp_data.main_split_pipe_index = 0xF;
+	}
+
+	// Find phantom pipe index based on phantom stream
+	for (j = 0; j < dc->res_pool->pipe_count; j++) {
+		struct pipe_ctx *phantom_pipe = &context->res_ctx.pipe_ctx[j];
+
+		if (resource_is_pipe_type(phantom_pipe, OTG_MASTER) &&
+				phantom_pipe->stream == dc_state_get_paired_subvp_stream(context, subvp_pipe->stream)) {
+			pipe_data->pipe_config.subvp_data.phantom_pipe_index = phantom_pipe->stream_res.tg->inst;
+			if (phantom_pipe->bottom_pipe) {
+				pipe_data->pipe_config.subvp_data.phantom_split_pipe_index = phantom_pipe->bottom_pipe->plane_res.hubp->inst;
+			} else if (phantom_pipe->next_odm_pipe) {
+				pipe_data->pipe_config.subvp_data.phantom_split_pipe_index = phantom_pipe->next_odm_pipe->plane_res.hubp->inst;
+			} else {
+				pipe_data->pipe_config.subvp_data.phantom_split_pipe_index = 0xF;
+			}
+			break;
+		}
+	}
+}
+
+/**
+ * dc_dmub_setup_subvp_dmub_command - Populate the DMCUB SubVP command
+ *
+ * @dc: [in] current dc state
+ * @context: [in] new dc state
+ * @enable: [in] if true enables the pipes population
+ *
+ * This function loops through each pipe and populates the DMUB SubVP CMD info
+ * based on the pipe (e.g. SubVP, VBLANK).
+ */
+void dc_dmub_setup_subvp_dmub_command(struct dc *dc,
+		struct dc_state *context,
+		bool enable)
+{
+	uint8_t cmd_pipe_index = 0;
+	uint32_t i;
+	uint8_t subvp_count = 0;
+	union dmub_rb_cmd cmd;
+	struct pipe_ctx *subvp_pipes[2];
+	uint32_t wm_val_refclk = 0;
+	enum mall_stream_type pipe_mall_type;
+
+	memset(&cmd, 0, sizeof(cmd));
+	// FW command for SUBVP
+	cmd.fw_assisted_mclk_switch_v2.header.type = DMUB_CMD__FW_ASSISTED_MCLK_SWITCH;
+	cmd.fw_assisted_mclk_switch_v2.header.sub_type = DMUB_CMD__HANDLE_SUBVP_CMD;
+	cmd.fw_assisted_mclk_switch_v2.header.payload_bytes =
+			sizeof(cmd.fw_assisted_mclk_switch_v2) - sizeof(cmd.fw_assisted_mclk_switch_v2.header);
+
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
+
+		/* For SubVP pipe count, only count the top most (ODM / MPC) pipe
+		 */
+		if (resource_is_pipe_type(pipe, OTG_MASTER) &&
+				resource_is_pipe_type(pipe, DPP_PIPE) &&
+				dc_state_get_pipe_subvp_type(context, pipe) == SUBVP_MAIN)
+			subvp_pipes[subvp_count++] = pipe;
+	}
+
+	if (enable) {
+		// For each pipe that is a "main" SUBVP pipe, fill in pipe data for DMUB SUBVP cmd
+		for (i = 0; i < dc->res_pool->pipe_count; i++) {
+			struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
+			pipe_mall_type = dc_state_get_pipe_subvp_type(context, pipe);
+
+			if (!pipe->stream)
+				continue;
+
+			/* When populating subvp cmd info, only pass in the top most (ODM / MPC) pipe.
+			 * Any ODM or MPC splits being used in SubVP will be handled internally in
+			 * populate_subvp_cmd_pipe_info
+			 */
+			if (resource_is_pipe_type(pipe, OTG_MASTER) &&
+					resource_is_pipe_type(pipe, DPP_PIPE) &&
+					pipe_mall_type == SUBVP_MAIN) {
+				populate_subvp_cmd_pipe_info(dc, context, &cmd, pipe, cmd_pipe_index++);
+			} else if (resource_is_pipe_type(pipe, OTG_MASTER) &&
+					resource_is_pipe_type(pipe, DPP_PIPE) &&
+					pipe_mall_type == SUBVP_NONE) {
+				// Don't need to check for ActiveDRAMClockChangeMargin < 0, not valid in cases where
+				// we run through DML without calculating "natural" P-state support
+				populate_subvp_cmd_vblank_pipe_info(dc, context, &cmd, pipe, cmd_pipe_index++);
+
+			}
+		}
+		if (subvp_count == 2) {
+			update_subvp_prefetch_end_to_mall_start(dc, context, &cmd, subvp_pipes);
+		}
+		cmd.fw_assisted_mclk_switch_v2.config_data.pstate_allow_width_us = dc->caps.subvp_pstate_allow_width_us;
+		cmd.fw_assisted_mclk_switch_v2.config_data.vertical_int_margin_us = dc->caps.subvp_vertical_int_margin_us;
+
+		// Store the original watermark value for this SubVP config so we can lower it when the
+		// MCLK switch starts
+		wm_val_refclk = context->bw_ctx.bw.dcn.watermarks.a.cstate_pstate.pstate_change_ns *
+				(dc->res_pool->ref_clocks.dchub_ref_clock_inKhz / 1000) / 1000;
+
+		cmd.fw_assisted_mclk_switch_v2.config_data.watermark_a_cache = wm_val_refclk < 0xFFFF ? wm_val_refclk : 0xFFFF;
+	}
+
+	dc_wake_and_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
+}
+
+bool dc_dmub_srv_get_diagnostic_data(struct dc_dmub_srv *dc_dmub_srv)
+{
+	if (!dc_dmub_srv || !dc_dmub_srv->dmub)
+		return false;
+	return dmub_srv_get_diagnostic_data(dc_dmub_srv->dmub);
 }
 
 void dc_dmub_srv_log_diagnostic_data(struct dc_dmub_srv *dc_dmub_srv)
 {
-	struct dmub_diagnostic_data diag_data = {0};
+	uint32_t i;
 
 	if (!dc_dmub_srv || !dc_dmub_srv->dmub) {
 		DC_LOG_ERROR("%s: invalid parameters.", __func__);
 		return;
 	}
 
-	if (!dc_dmub_srv_get_diagnostic_data(dc_dmub_srv, &diag_data)) {
+	DC_LOG_ERROR("%s: DMCUB error - collecting diagnostic data\n", __func__);
+
+	if (!dc_dmub_srv_get_diagnostic_data(dc_dmub_srv)) {
 		DC_LOG_ERROR("%s: dc_dmub_srv_get_diagnostic_data failed.", __func__);
 		return;
 	}
 
-	DC_LOG_DEBUG(
-		"DMCUB STATE\n"
-		"    dmcub_version      : %08x\n"
-		"    scratch  [0]       : %08x\n"
-		"    scratch  [1]       : %08x\n"
-		"    scratch  [2]       : %08x\n"
-		"    scratch  [3]       : %08x\n"
-		"    scratch  [4]       : %08x\n"
-		"    scratch  [5]       : %08x\n"
-		"    scratch  [6]       : %08x\n"
-		"    scratch  [7]       : %08x\n"
-		"    scratch  [8]       : %08x\n"
-		"    scratch  [9]       : %08x\n"
-		"    scratch [10]       : %08x\n"
-		"    scratch [11]       : %08x\n"
-		"    scratch [12]       : %08x\n"
-		"    scratch [13]       : %08x\n"
-		"    scratch [14]       : %08x\n"
-		"    scratch [15]       : %08x\n"
-		"    pc                 : %08x\n"
-		"    unk_fault_addr     : %08x\n"
-		"    inst_fault_addr    : %08x\n"
-		"    data_fault_addr    : %08x\n"
-		"    inbox1_rptr        : %08x\n"
-		"    inbox1_wptr        : %08x\n"
-		"    inbox1_size        : %08x\n"
-		"    inbox0_rptr        : %08x\n"
-		"    inbox0_wptr        : %08x\n"
-		"    inbox0_size        : %08x\n"
-		"    is_enabled         : %d\n"
-		"    is_soft_reset      : %d\n"
-		"    is_secure_reset    : %d\n"
-		"    is_traceport_en    : %d\n"
-		"    is_cw0_en          : %d\n"
-		"    is_cw6_en          : %d\n",
-		diag_data.dmcub_version,
-		diag_data.scratch[0],
-		diag_data.scratch[1],
-		diag_data.scratch[2],
-		diag_data.scratch[3],
-		diag_data.scratch[4],
-		diag_data.scratch[5],
-		diag_data.scratch[6],
-		diag_data.scratch[7],
-		diag_data.scratch[8],
-		diag_data.scratch[9],
-		diag_data.scratch[10],
-		diag_data.scratch[11],
-		diag_data.scratch[12],
-		diag_data.scratch[13],
-		diag_data.scratch[14],
-		diag_data.scratch[15],
-		diag_data.pc,
-		diag_data.undefined_address_fault_addr,
-		diag_data.inst_fetch_fault_addr,
-		diag_data.data_write_fault_addr,
-		diag_data.inbox1_rptr,
-		diag_data.inbox1_wptr,
-		diag_data.inbox1_size,
-		diag_data.inbox0_rptr,
-		diag_data.inbox0_wptr,
-		diag_data.inbox0_size,
-		diag_data.is_dmcub_enabled,
-		diag_data.is_dmcub_soft_reset,
-		diag_data.is_dmcub_secure_reset,
-		diag_data.is_traceport_en,
-		diag_data.is_cw0_enabled,
-		diag_data.is_cw6_enabled);
+	DC_LOG_DEBUG("DMCUB STATE:");
+	DC_LOG_DEBUG("    dmcub_version      : %08x", dc_dmub_srv->dmub->debug.dmcub_version);
+	DC_LOG_DEBUG("    scratch  [0]       : %08x", dc_dmub_srv->dmub->debug.scratch[0]);
+	DC_LOG_DEBUG("    scratch  [1]       : %08x", dc_dmub_srv->dmub->debug.scratch[1]);
+	DC_LOG_DEBUG("    scratch  [2]       : %08x", dc_dmub_srv->dmub->debug.scratch[2]);
+	DC_LOG_DEBUG("    scratch  [3]       : %08x", dc_dmub_srv->dmub->debug.scratch[3]);
+	DC_LOG_DEBUG("    scratch  [4]       : %08x", dc_dmub_srv->dmub->debug.scratch[4]);
+	DC_LOG_DEBUG("    scratch  [5]       : %08x", dc_dmub_srv->dmub->debug.scratch[5]);
+	DC_LOG_DEBUG("    scratch  [6]       : %08x", dc_dmub_srv->dmub->debug.scratch[6]);
+	DC_LOG_DEBUG("    scratch  [7]       : %08x", dc_dmub_srv->dmub->debug.scratch[7]);
+	DC_LOG_DEBUG("    scratch  [8]       : %08x", dc_dmub_srv->dmub->debug.scratch[8]);
+	DC_LOG_DEBUG("    scratch  [9]       : %08x", dc_dmub_srv->dmub->debug.scratch[9]);
+	DC_LOG_DEBUG("    scratch [10]       : %08x", dc_dmub_srv->dmub->debug.scratch[10]);
+	DC_LOG_DEBUG("    scratch [11]       : %08x", dc_dmub_srv->dmub->debug.scratch[11]);
+	DC_LOG_DEBUG("    scratch [12]       : %08x", dc_dmub_srv->dmub->debug.scratch[12]);
+	DC_LOG_DEBUG("    scratch [13]       : %08x", dc_dmub_srv->dmub->debug.scratch[13]);
+	DC_LOG_DEBUG("    scratch [14]       : %08x", dc_dmub_srv->dmub->debug.scratch[14]);
+	DC_LOG_DEBUG("    scratch [15]       : %08x", dc_dmub_srv->dmub->debug.scratch[15]);
+	for (i = 0; i < DMUB_PC_SNAPSHOT_COUNT; i++)
+		DC_LOG_DEBUG("    pc[%d]             : %08x", i, dc_dmub_srv->dmub->debug.pc[i]);
+	DC_LOG_DEBUG("    unk_fault_addr     : %08x", dc_dmub_srv->dmub->debug.undefined_address_fault_addr);
+	DC_LOG_DEBUG("    inst_fault_addr    : %08x", dc_dmub_srv->dmub->debug.inst_fetch_fault_addr);
+	DC_LOG_DEBUG("    data_fault_addr    : %08x", dc_dmub_srv->dmub->debug.data_write_fault_addr);
+	DC_LOG_DEBUG("    inbox1_rptr        : %08x", dc_dmub_srv->dmub->debug.inbox1_rptr);
+	DC_LOG_DEBUG("    inbox1_wptr        : %08x", dc_dmub_srv->dmub->debug.inbox1_wptr);
+	DC_LOG_DEBUG("    inbox1_size        : %08x", dc_dmub_srv->dmub->debug.inbox1_size);
+	DC_LOG_DEBUG("    inbox0_rptr        : %08x", dc_dmub_srv->dmub->debug.inbox0_rptr);
+	DC_LOG_DEBUG("    inbox0_wptr        : %08x", dc_dmub_srv->dmub->debug.inbox0_wptr);
+	DC_LOG_DEBUG("    inbox0_size        : %08x", dc_dmub_srv->dmub->debug.inbox0_size);
+	DC_LOG_DEBUG("    outbox1_rptr       : %08x", dc_dmub_srv->dmub->debug.outbox1_rptr);
+	DC_LOG_DEBUG("    outbox1_wptr       : %08x", dc_dmub_srv->dmub->debug.outbox1_wptr);
+	DC_LOG_DEBUG("    outbox1_size       : %08x", dc_dmub_srv->dmub->debug.outbox1_size);
+	DC_LOG_DEBUG("    is_enabled         : %d", dc_dmub_srv->dmub->debug.is_dmcub_enabled);
+	DC_LOG_DEBUG("    is_soft_reset      : %d", dc_dmub_srv->dmub->debug.is_dmcub_soft_reset);
+	DC_LOG_DEBUG("    is_secure_reset    : %d", dc_dmub_srv->dmub->debug.is_dmcub_secure_reset);
+	DC_LOG_DEBUG("    is_traceport_en    : %d", dc_dmub_srv->dmub->debug.is_traceport_en);
+	DC_LOG_DEBUG("    is_cw0_en          : %d", dc_dmub_srv->dmub->debug.is_cw0_enabled);
+	DC_LOG_DEBUG("    is_cw6_en          : %d", dc_dmub_srv->dmub->debug.is_cw6_enabled);
+}
+
+static bool dc_dmub_should_update_cursor_data(struct pipe_ctx *pipe_ctx)
+{
+	if (pipe_ctx->plane_state != NULL) {
+		if (pipe_ctx->plane_state->address.type == PLN_ADDR_TYPE_VIDEO_PROGRESSIVE ||
+			resource_can_pipe_disable_cursor(pipe_ctx))
+			return false;
+	}
+
+	if ((pipe_ctx->stream->link->psr_settings.psr_version == DC_PSR_VERSION_SU_1 ||
+		pipe_ctx->stream->link->psr_settings.psr_version == DC_PSR_VERSION_1) &&
+		pipe_ctx->stream->ctx->dce_version >= DCN_VERSION_3_1)
+		return true;
+
+	if (pipe_ctx->stream->link->replay_settings.config.replay_supported)
+		return true;
+
+	return false;
+}
+
+static void dc_build_cursor_update_payload0(
+		struct pipe_ctx *pipe_ctx, uint8_t p_idx,
+		struct dmub_cmd_update_cursor_payload0 *payload)
+{
+	struct hubp *hubp = pipe_ctx->plane_res.hubp;
+	unsigned int panel_inst = 0;
+
+	if (!dc_get_edp_link_panel_inst(hubp->ctx->dc,
+		pipe_ctx->stream->link, &panel_inst))
+		return;
+
+	/* Payload: Cursor Rect is built from position & attribute
+	 * x & y are obtained from postion
+	 */
+	payload->cursor_rect.x = hubp->cur_rect.x;
+	payload->cursor_rect.y = hubp->cur_rect.y;
+	/* w & h are obtained from attribute */
+	payload->cursor_rect.width  = hubp->cur_rect.w;
+	payload->cursor_rect.height = hubp->cur_rect.h;
+
+	payload->enable      = hubp->pos.cur_ctl.bits.cur_enable;
+	payload->pipe_idx    = p_idx;
+	payload->cmd_version = DMUB_CMD_PSR_CONTROL_VERSION_1;
+	payload->panel_inst  = panel_inst;
+}
+
+static void dc_build_cursor_position_update_payload0(
+		struct dmub_cmd_update_cursor_payload0 *pl, const uint8_t p_idx,
+		const struct hubp *hubp, const struct dpp *dpp)
+{
+	/* Hubp */
+	pl->position_cfg.pHubp.cur_ctl.raw  = hubp->pos.cur_ctl.raw;
+	pl->position_cfg.pHubp.position.raw = hubp->pos.position.raw;
+	pl->position_cfg.pHubp.hot_spot.raw = hubp->pos.hot_spot.raw;
+	pl->position_cfg.pHubp.dst_offset.raw = hubp->pos.dst_offset.raw;
+
+	/* dpp */
+	pl->position_cfg.pDpp.cur0_ctl.raw = dpp->pos.cur0_ctl.raw;
+	pl->position_cfg.pipe_idx = p_idx;
+}
+
+static void dc_build_cursor_attribute_update_payload1(
+		struct dmub_cursor_attributes_cfg *pl_A, const uint8_t p_idx,
+		const struct hubp *hubp, const struct dpp *dpp)
+{
+	/* Hubp */
+	pl_A->aHubp.SURFACE_ADDR_HIGH = hubp->att.SURFACE_ADDR_HIGH;
+	pl_A->aHubp.SURFACE_ADDR = hubp->att.SURFACE_ADDR;
+	pl_A->aHubp.cur_ctl.raw  = hubp->att.cur_ctl.raw;
+	pl_A->aHubp.size.raw     = hubp->att.size.raw;
+	pl_A->aHubp.settings.raw = hubp->att.settings.raw;
+
+	/* dpp */
+	pl_A->aDpp.cur0_ctl.raw = dpp->att.cur0_ctl.raw;
+}
+
+/**
+ * dc_send_update_cursor_info_to_dmu - Populate the DMCUB Cursor update info command
+ *
+ * @pCtx: [in] pipe context
+ * @pipe_idx: [in] pipe index
+ *
+ * This function would store the cursor related information and pass it into
+ * dmub
+ */
+void dc_send_update_cursor_info_to_dmu(
+		struct pipe_ctx *pCtx, uint8_t pipe_idx)
+{
+	union dmub_rb_cmd cmd[2];
+	union dmub_cmd_update_cursor_info_data *update_cursor_info_0 =
+					&cmd[0].update_cursor_info.update_cursor_info_data;
+
+	memset(cmd, 0, sizeof(cmd));
+
+	if (!dc_dmub_should_update_cursor_data(pCtx))
+		return;
+	/*
+	 * Since we use multi_cmd_pending for dmub command, the 2nd command is
+	 * only assigned to store cursor attributes info.
+	 * 1st command can view as 2 parts, 1st is for PSR/Replay data, the other
+	 * is to store cursor position info.
+	 *
+	 * Command heaer type must be the same type if using  multi_cmd_pending.
+	 * Besides, while process 2nd command in DMU, the sub type is useless.
+	 * So it's meanless to pass the sub type header with different type.
+	 */
+
+	{
+		/* Build Payload#0 Header */
+		cmd[0].update_cursor_info.header.type = DMUB_CMD__UPDATE_CURSOR_INFO;
+		cmd[0].update_cursor_info.header.payload_bytes =
+				sizeof(cmd[0].update_cursor_info.update_cursor_info_data);
+		cmd[0].update_cursor_info.header.multi_cmd_pending = 1; //To combine multi dmu cmd, 1st cmd
+
+		/* Prepare Payload */
+		dc_build_cursor_update_payload0(pCtx, pipe_idx, &update_cursor_info_0->payload0);
+
+		dc_build_cursor_position_update_payload0(&update_cursor_info_0->payload0, pipe_idx,
+				pCtx->plane_res.hubp, pCtx->plane_res.dpp);
+		}
+	{
+		/* Build Payload#1 Header */
+		cmd[1].update_cursor_info.header.type = DMUB_CMD__UPDATE_CURSOR_INFO;
+		cmd[1].update_cursor_info.header.payload_bytes = sizeof(struct cursor_attributes_cfg);
+		cmd[1].update_cursor_info.header.multi_cmd_pending = 0; //Indicate it's the last command.
+
+		dc_build_cursor_attribute_update_payload1(
+				&cmd[1].update_cursor_info.update_cursor_info_data.payload1.attribute_cfg,
+				pipe_idx, pCtx->plane_res.hubp, pCtx->plane_res.dpp);
+
+		/* Combine 2nd cmds update_curosr_info to DMU */
+		dc_wake_and_execute_dmub_cmd_list(pCtx->stream->ctx, 2, cmd, DM_DMUB_WAIT_TYPE_WAIT);
+	}
+}
+
+bool dc_dmub_check_min_version(struct dmub_srv *srv)
+{
+	if (!srv->hw_funcs.is_psrsu_supported)
+		return true;
+	return srv->hw_funcs.is_psrsu_supported(srv);
+}
+
+void dc_dmub_srv_enable_dpia_trace(const struct dc *dc)
+{
+	struct dc_dmub_srv *dc_dmub_srv = dc->ctx->dmub_srv;
+
+	if (!dc_dmub_srv || !dc_dmub_srv->dmub) {
+		DC_LOG_ERROR("%s: invalid parameters.", __func__);
+		return;
+	}
+
+	if (!dc_wake_and_execute_gpint(dc->ctx, DMUB_GPINT__SET_TRACE_BUFFER_MASK_WORD1,
+				       0x0010, NULL, DM_DMUB_WAIT_TYPE_WAIT)) {
+		DC_LOG_ERROR("timeout updating trace buffer mask word\n");
+		return;
+	}
+
+	if (!dc_wake_and_execute_gpint(dc->ctx, DMUB_GPINT__UPDATE_TRACE_BUFFER_MASK,
+				       0x0000, NULL, DM_DMUB_WAIT_TYPE_WAIT)) {
+		DC_LOG_ERROR("timeout updating trace buffer mask word\n");
+		return;
+	}
+
+	DC_LOG_DEBUG("Enabled DPIA trace\n");
+}
+
+void dc_dmub_srv_subvp_save_surf_addr(const struct dc_dmub_srv *dc_dmub_srv, const struct dc_plane_address *addr, uint8_t subvp_index)
+{
+	dmub_srv_subvp_save_surf_addr(dc_dmub_srv->dmub, addr, subvp_index);
+}
+
+void dc_dmub_srv_cursor_offload_init(struct dc *dc)
+{
+	struct dmub_rb_cmd_cursor_offload_init *init;
+	struct dc_dmub_srv *dc_dmub_srv = dc->ctx->dmub_srv;
+	union dmub_rb_cmd cmd;
+
+	if (!dc->config.enable_cursor_offload)
+		return;
+
+	if (!dc_dmub_srv->dmub->meta_info.feature_bits.bits.cursor_offload_v1_support)
+		return;
+
+	if (!dc_dmub_srv->dmub->cursor_offload_fb.gpu_addr || !dc_dmub_srv->dmub->cursor_offload_fb.cpu_addr)
+		return;
+
+	if (!dc_dmub_srv->dmub->cursor_offload_v1)
+		return;
+
+	if (!dc_dmub_srv->dmub->shared_state)
+		return;
+
+	memset(&cmd, 0, sizeof(cmd));
+
+	init = &cmd.cursor_offload_init;
+	init->header.type = DMUB_CMD__CURSOR_OFFLOAD;
+	init->header.sub_type = DMUB_CMD__CURSOR_OFFLOAD_INIT;
+	init->header.payload_bytes = sizeof(init->init_data);
+	init->init_data.state_addr.quad_part = dc_dmub_srv->dmub->cursor_offload_fb.gpu_addr;
+	init->init_data.state_size = dc_dmub_srv->dmub->cursor_offload_fb.size;
+
+	dc_wake_and_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
+
+	dc_dmub_srv->cursor_offload_enabled = true;
+}
+
+void dc_dmub_srv_control_cursor_offload(struct dc *dc, struct dc_state *context,
+					const struct dc_stream_state *stream, bool enable)
+{
+	struct pipe_ctx const *pipe_ctx;
+	struct dmub_rb_cmd_cursor_offload_stream_cntl *cntl;
+	union dmub_rb_cmd cmd;
+
+	if (!dc_dmub_srv_is_cursor_offload_enabled(dc))
+		return;
+
+	if (!stream)
+		return;
+
+	pipe_ctx = resource_get_otg_master_for_stream(&context->res_ctx, stream);
+	if (!pipe_ctx || !pipe_ctx->stream_res.tg || pipe_ctx->stream != stream)
+		return;
+
+	memset(&cmd, 0, sizeof(cmd));
+
+	cntl = &cmd.cursor_offload_stream_ctnl;
+	cntl->header.type = DMUB_CMD__CURSOR_OFFLOAD;
+	cntl->header.sub_type =
+		enable ? DMUB_CMD__CURSOR_OFFLOAD_STREAM_ENABLE : DMUB_CMD__CURSOR_OFFLOAD_STREAM_DISABLE;
+	cntl->header.payload_bytes = sizeof(cntl->data);
+
+	cntl->data.otg_inst = pipe_ctx->stream_res.tg->inst;
+	cntl->data.line_time_in_ns = 1u + (uint32_t)(div64_u64(stream->timing.h_total * 1000000ull,
+							       stream->timing.pix_clk_100hz / 10));
+
+	cntl->data.v_total_max = stream->adjust.v_total_max > stream->timing.v_total ?
+					 stream->adjust.v_total_max :
+					 stream->timing.v_total;
+
+	dc_wake_and_execute_dmub_cmd(dc->ctx, &cmd,
+				     enable ? DM_DMUB_WAIT_TYPE_NO_WAIT : DM_DMUB_WAIT_TYPE_WAIT);
+}
+
+void dc_dmub_srv_program_cursor_now(struct dc *dc, const struct pipe_ctx *pipe)
+{
+	struct dmub_rb_cmd_cursor_offload_stream_cntl *cntl;
+	union dmub_rb_cmd cmd;
+
+	if (!dc_dmub_srv_is_cursor_offload_enabled(dc))
+		return;
+
+	if (!pipe || !pipe->stream || !pipe->stream_res.tg)
+		return;
+
+	memset(&cmd, 0, sizeof(cmd));
+
+	cntl = &cmd.cursor_offload_stream_ctnl;
+	cntl->header.type = DMUB_CMD__CURSOR_OFFLOAD;
+	cntl->header.sub_type = DMUB_CMD__CURSOR_OFFLOAD_STREAM_PROGRAM;
+	cntl->header.payload_bytes = sizeof(cntl->data);
+	cntl->data.otg_inst = pipe->stream_res.tg->inst;
+
+	dc_wake_and_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_NO_WAIT);
+}
+
+bool dc_dmub_srv_is_hw_pwr_up(struct dc_dmub_srv *dc_dmub_srv, bool wait)
+{
+	struct dc_context *dc_ctx;
+	enum dmub_status status;
+
+	if (!dc_dmub_srv || !dc_dmub_srv->dmub)
+		return true;
+
+	if (dc_dmub_srv->ctx->dc->debug.dmcub_emulation)
+		return true;
+
+	dc_ctx = dc_dmub_srv->ctx;
+
+	if (wait) {
+		if (dc_dmub_srv->ctx->dc->debug.disable_timeout) {
+			do {
+				status = dmub_srv_wait_for_hw_pwr_up(dc_dmub_srv->dmub, 500000);
+			} while (status != DMUB_STATUS_OK);
+		} else {
+			status = dmub_srv_wait_for_hw_pwr_up(dc_dmub_srv->dmub, 500000);
+			if (status != DMUB_STATUS_OK) {
+				DC_ERROR("Error querying DMUB hw power up status: error=%d\n", status);
+				return false;
+			}
+		}
+	} else
+		return dmub_srv_is_hw_pwr_up(dc_dmub_srv->dmub);
+
+	return true;
+}
+
+static int count_active_streams(const struct dc *dc)
+{
+	int i, count = 0;
+
+	for (i = 0; i < dc->current_state->stream_count; ++i) {
+		struct dc_stream_state *stream = dc->current_state->streams[i];
+
+		if (stream && (!stream->dpms_off || dc->config.disable_ips_in_dpms_off))
+			count += 1;
+	}
+
+	return count;
+}
+
+static void dc_dmub_srv_notify_idle(const struct dc *dc, bool allow_idle)
+{
+	volatile const struct dmub_shared_state_ips_fw *ips_fw;
+	struct dc_dmub_srv *dc_dmub_srv;
+	union dmub_rb_cmd cmd = {0};
+
+	if (dc->debug.dmcub_emulation)
+		return;
+
+	if (!dc->ctx->dmub_srv || !dc->ctx->dmub_srv->dmub)
+		return;
+
+	dc_dmub_srv = dc->ctx->dmub_srv;
+	ips_fw = &dc_dmub_srv->dmub->shared_state[DMUB_SHARED_SHARE_FEATURE__IPS_FW].data.ips_fw;
+
+	memset(&cmd, 0, sizeof(cmd));
+	cmd.idle_opt_notify_idle.header.type = DMUB_CMD__IDLE_OPT;
+	cmd.idle_opt_notify_idle.header.sub_type = DMUB_CMD__IDLE_OPT_DCN_NOTIFY_IDLE;
+	cmd.idle_opt_notify_idle.header.payload_bytes =
+		sizeof(cmd.idle_opt_notify_idle) -
+		sizeof(cmd.idle_opt_notify_idle.header);
+
+	cmd.idle_opt_notify_idle.cntl_data.driver_idle = allow_idle;
+
+	if (dc->work_arounds.skip_psr_ips_crtc_disable)
+		cmd.idle_opt_notify_idle.cntl_data.skip_otg_disable = true;
+
+	if (allow_idle) {
+		volatile struct dmub_shared_state_ips_driver *ips_driver =
+			&dc_dmub_srv->dmub->shared_state[DMUB_SHARED_SHARE_FEATURE__IPS_DRIVER].data.ips_driver;
+		union dmub_shared_state_ips_driver_signals new_signals;
+
+		DC_LOG_IPS(
+			"%s wait idle (ips1_commit=%u ips2_commit=%u)",
+			__func__,
+			ips_fw->signals.bits.ips1_commit,
+			ips_fw->signals.bits.ips2_commit);
+
+		dc_dmub_srv_wait_for_idle(dc->ctx->dmub_srv, DM_DMUB_WAIT_TYPE_WAIT, NULL);
+
+		memset(&new_signals, 0, sizeof(new_signals));
+
+		new_signals.bits.allow_idle = 1; /* always set */
+
+		if (dc->config.disable_ips == DMUB_IPS_ENABLE ||
+		    dc->config.disable_ips == DMUB_IPS_DISABLE_DYNAMIC) {
+			new_signals.bits.allow_pg = 1;
+			new_signals.bits.allow_ips1 = 1;
+			new_signals.bits.allow_ips2 = 1;
+			new_signals.bits.allow_z10 = 1;
+			// New in IPSv2.0
+			new_signals.bits.allow_ips1z8 = 1;
+		} else if (dc->config.disable_ips == DMUB_IPS_DISABLE_IPS1) {
+			new_signals.bits.allow_ips1 = 1;
+		} else if (dc->config.disable_ips == DMUB_IPS_DISABLE_IPS2) {
+			// IPSv1.0 only
+			new_signals.bits.allow_pg = 1;
+			new_signals.bits.allow_ips1 = 1;
+		} else if (dc->config.disable_ips == DMUB_IPS_DISABLE_IPS2_Z10) {
+			// IPSv1.0 only
+			new_signals.bits.allow_pg = 1;
+			new_signals.bits.allow_ips1 = 1;
+			new_signals.bits.allow_ips2 = 1;
+		} else if (dc->config.disable_ips == DMUB_IPS_RCG_IN_ACTIVE_IPS2_IN_OFF) {
+			/* TODO: Move this logic out to hwseq */
+			if (count_active_streams(dc) == 0) {
+				/* IPS2 - Display off */
+				new_signals.bits.allow_pg = 1;
+				new_signals.bits.allow_ips1 = 1;
+				new_signals.bits.allow_ips2 = 1;
+				new_signals.bits.allow_z10 = 1;
+				// New in IPSv2.0
+				new_signals.bits.allow_ips1z8 = 1;
+			} else {
+				/* RCG only */
+				new_signals.bits.allow_pg = 0;
+				new_signals.bits.allow_ips1 = 1;
+				new_signals.bits.allow_ips2 = 0;
+				new_signals.bits.allow_z10 = 0;
+			}
+		} else if (dc->config.disable_ips == DMUB_IPS_DISABLE_Z8_RETENTION) {
+			new_signals.bits.allow_pg = 1;
+			new_signals.bits.allow_ips1 = 1;
+			new_signals.bits.allow_ips2 = 1;
+			new_signals.bits.allow_z10 = 1;
+		}
+		// Setting RCG allow bits (IPSv2.0)
+		if (dc->config.disable_ips_rcg == DMUB_IPS_RCG_ENABLE) {
+			new_signals.bits.allow_ips0_rcg = 1;
+			new_signals.bits.allow_ips1_rcg = 1;
+		} else if (dc->config.disable_ips_rcg == DMUB_IPS0_RCG_DISABLE) {
+			new_signals.bits.allow_ips1_rcg = 1;
+		} else if (dc->config.disable_ips_rcg == DMUB_IPS1_RCG_DISABLE) {
+			new_signals.bits.allow_ips0_rcg = 1;
+		}
+		// IPS dynamic allow bits (IPSv2 change, vpb use case)
+		if (dc->config.disable_ips_in_vpb == DMUB_IPS_VPB_ENABLE_IPS1_AND_RCG) {
+			new_signals.bits.allow_dynamic_ips1 = 1;
+		} else if (dc->config.disable_ips_in_vpb == DMUB_IPS_VPB_ENABLE_ALL) {
+			new_signals.bits.allow_dynamic_ips1 = 1;
+			new_signals.bits.allow_dynamic_ips1_z8 = 1;
+		}
+		ips_driver->signals = new_signals;
+		dc_dmub_srv->driver_signals = ips_driver->signals;
+	}
+
+	DC_LOG_IPS(
+		"%s send allow_idle=%d (ips1_commit=%u ips2_commit=%u)",
+		__func__,
+		allow_idle,
+		ips_fw->signals.bits.ips1_commit,
+		ips_fw->signals.bits.ips2_commit);
+
+	/* NOTE: This does not use the "wake" interface since this is part of the wake path. */
+	/* We also do not perform a wait since DMCUB could enter idle after the notification. */
+	dm_execute_dmub_cmd(dc->ctx, &cmd, allow_idle ? DM_DMUB_WAIT_TYPE_NO_WAIT : DM_DMUB_WAIT_TYPE_WAIT);
+
+	/* Register access should stop at this point. */
+	if (allow_idle)
+		dc_dmub_srv->needs_idle_wake = true;
+}
+
+static void dc_dmub_srv_exit_low_power_state(const struct dc *dc)
+{
+	struct dc_dmub_srv *dc_dmub_srv;
+	uint32_t rcg_exit_count = 0, ips1_exit_count = 0, ips2_exit_count = 0, ips1z8_exit_count = 0;
+
+	if (dc->debug.dmcub_emulation)
+		return;
+
+	if (!dc->ctx->dmub_srv || !dc->ctx->dmub_srv->dmub)
+		return;
+
+	dc_dmub_srv = dc->ctx->dmub_srv;
+
+	if (dc->clk_mgr->funcs->exit_low_power_state) {
+		volatile const struct dmub_shared_state_ips_fw *ips_fw =
+			&dc_dmub_srv->dmub->shared_state[DMUB_SHARED_SHARE_FEATURE__IPS_FW].data.ips_fw;
+		volatile struct dmub_shared_state_ips_driver *ips_driver =
+			&dc_dmub_srv->dmub->shared_state[DMUB_SHARED_SHARE_FEATURE__IPS_DRIVER].data.ips_driver;
+		union dmub_shared_state_ips_driver_signals prev_driver_signals = ips_driver->signals;
+
+		rcg_exit_count = ips_fw->rcg_exit_count;
+		ips1_exit_count = ips_fw->ips1_exit_count;
+		ips2_exit_count = ips_fw->ips2_exit_count;
+		ips1z8_exit_count = ips_fw->ips1_z8ret_exit_count;
+
+		ips_driver->signals.all = 0;
+		dc_dmub_srv->driver_signals = ips_driver->signals;
+
+		DC_LOG_IPS(
+			"%s (allow ips1=%u ips2=%u) (commit ips1=%u ips2=%u ips1z8=%u) (count rcg=%u ips1=%u ips2=%u ips1_z8=%u)",
+			__func__,
+			ips_driver->signals.bits.allow_ips1,
+			ips_driver->signals.bits.allow_ips2,
+			ips_fw->signals.bits.ips1_commit,
+			ips_fw->signals.bits.ips2_commit,
+			ips_fw->signals.bits.ips1z8_commit,
+			ips_fw->rcg_entry_count,
+			ips_fw->ips1_entry_count,
+			ips_fw->ips2_entry_count,
+			ips_fw->ips1_z8ret_entry_count);
+
+		/* Note: register access has technically not resumed for DCN here, but we
+		 * need to be message PMFW through our standard register interface.
+		 */
+		dc_dmub_srv->needs_idle_wake = false;
+
+		if (!dc->caps.ips_v2_support && ((prev_driver_signals.bits.allow_ips2 || prev_driver_signals.all == 0) &&
+		    (!dc->debug.optimize_ips_handshake ||
+		     ips_fw->signals.bits.ips2_commit || !ips_fw->signals.bits.in_idle))) {
+			DC_LOG_IPS(
+				"wait IPS2 eval (ips1_commit=%u ips2_commit=%u )",
+				ips_fw->signals.bits.ips1_commit,
+				ips_fw->signals.bits.ips2_commit);
+
+			if (!dc->debug.optimize_ips_handshake || !ips_fw->signals.bits.ips2_commit)
+				udelay(dc->debug.ips2_eval_delay_us);
+
+			DC_LOG_IPS(
+				"exit IPS2 #1 (ips1_commit=%u ips2_commit=%u)",
+				ips_fw->signals.bits.ips1_commit,
+				ips_fw->signals.bits.ips2_commit);
+
+			// Tell PMFW to exit low power state
+			dc->clk_mgr->funcs->exit_low_power_state(dc->clk_mgr);
+
+			if (ips_fw->signals.bits.ips2_commit) {
+
+				DC_LOG_IPS(
+					"wait IPS2 entry delay (ips1_commit=%u ips2_commit=%u)",
+					ips_fw->signals.bits.ips1_commit,
+					ips_fw->signals.bits.ips2_commit);
+
+				// Wait for IPS2 entry upper bound
+				udelay(dc->debug.ips2_entry_delay_us);
+
+				DC_LOG_IPS(
+					"exit IPS2 #2 (ips1_commit=%u ips2_commit=%u)",
+					ips_fw->signals.bits.ips1_commit,
+					ips_fw->signals.bits.ips2_commit);
+
+				dc->clk_mgr->funcs->exit_low_power_state(dc->clk_mgr);
+
+				DC_LOG_IPS(
+					"wait IPS2 commit clear (ips1_commit=%u ips2_commit=%u)",
+					ips_fw->signals.bits.ips1_commit,
+					ips_fw->signals.bits.ips2_commit);
+
+				while (ips_fw->signals.bits.ips2_commit)
+					udelay(1);
+
+				DC_LOG_IPS(
+					"wait hw_pwr_up (ips1_commit=%u ips2_commit=%u)",
+					ips_fw->signals.bits.ips1_commit,
+					ips_fw->signals.bits.ips2_commit);
+
+				if (!dc_dmub_srv_is_hw_pwr_up(dc->ctx->dmub_srv, true))
+					ASSERT(0);
+
+				DC_LOG_IPS(
+					"resync inbox1 (ips1_commit=%u ips2_commit=%u)",
+					ips_fw->signals.bits.ips1_commit,
+					ips_fw->signals.bits.ips2_commit);
+
+				dmub_srv_sync_inboxes(dc->ctx->dmub_srv->dmub);
+			}
+		}
+
+		dc_dmub_srv_notify_idle(dc, false);
+		if (prev_driver_signals.bits.allow_ips1 || prev_driver_signals.all == 0) {
+			DC_LOG_IPS(
+				"wait for IPS1 commit clear (ips1_commit=%u ips2_commit=%u ips1z8=%u)",
+				ips_fw->signals.bits.ips1_commit,
+				ips_fw->signals.bits.ips2_commit,
+				ips_fw->signals.bits.ips1z8_commit);
+
+			while (ips_fw->signals.bits.ips1_commit)
+				udelay(1);
+
+			DC_LOG_IPS(
+				"wait for IPS1 commit clear done (ips1_commit=%u ips2_commit=%u ips1z8=%u)",
+				ips_fw->signals.bits.ips1_commit,
+				ips_fw->signals.bits.ips2_commit,
+				ips_fw->signals.bits.ips1z8_commit);
+		}
+	}
+
+	if (!dc_dmub_srv_is_hw_pwr_up(dc->ctx->dmub_srv, true))
+		ASSERT(0);
+
+	DC_LOG_IPS("%s exit (count rcg=%u ips1=%u ips2=%u ips1z8=%u)",
+		__func__,
+		rcg_exit_count,
+		ips1_exit_count,
+		ips2_exit_count,
+		ips1z8_exit_count);
+}
+
+void dc_dmub_srv_set_power_state(struct dc_dmub_srv *dc_dmub_srv, enum dc_acpi_cm_power_state power_state)
+{
+	struct dmub_srv *dmub;
+
+	if (!dc_dmub_srv)
+		return;
+
+	dmub = dc_dmub_srv->dmub;
+
+	if (power_state == DC_ACPI_CM_POWER_STATE_D0)
+		dmub_srv_set_power_state(dmub, DMUB_POWER_STATE_D0);
+	else
+		dmub_srv_set_power_state(dmub, DMUB_POWER_STATE_D3);
+}
+
+void dc_dmub_srv_notify_fw_dc_power_state(struct dc_dmub_srv *dc_dmub_srv,
+					  enum dc_acpi_cm_power_state power_state)
+{
+	union dmub_rb_cmd cmd;
+
+	if (!dc_dmub_srv)
+		return;
+
+	memset(&cmd, 0, sizeof(cmd));
+
+	cmd.idle_opt_set_dc_power_state.header.type = DMUB_CMD__IDLE_OPT;
+	cmd.idle_opt_set_dc_power_state.header.sub_type = DMUB_CMD__IDLE_OPT_SET_DC_POWER_STATE;
+	cmd.idle_opt_set_dc_power_state.header.payload_bytes =
+		sizeof(cmd.idle_opt_set_dc_power_state) - sizeof(cmd.idle_opt_set_dc_power_state.header);
+
+	if (power_state == DC_ACPI_CM_POWER_STATE_D0) {
+		cmd.idle_opt_set_dc_power_state.data.power_state = DMUB_IDLE_OPT_DC_POWER_STATE_D0;
+	} else if (power_state == DC_ACPI_CM_POWER_STATE_D3) {
+		cmd.idle_opt_set_dc_power_state.data.power_state = DMUB_IDLE_OPT_DC_POWER_STATE_D3;
+	} else {
+		cmd.idle_opt_set_dc_power_state.data.power_state = DMUB_IDLE_OPT_DC_POWER_STATE_UNKNOWN;
+	}
+
+	dc_wake_and_execute_dmub_cmd(dc_dmub_srv->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
+}
+
+bool dc_dmub_srv_should_detect(struct dc_dmub_srv *dc_dmub_srv)
+{
+	volatile const struct dmub_shared_state_ips_fw *ips_fw;
+	bool reallow_idle = false, should_detect = false;
+
+	if (!dc_dmub_srv || !dc_dmub_srv->dmub)
+		return false;
+
+	if (dc_dmub_srv->dmub->shared_state &&
+	    dc_dmub_srv->dmub->meta_info.feature_bits.bits.shared_state_link_detection) {
+		ips_fw = &dc_dmub_srv->dmub->shared_state[DMUB_SHARED_SHARE_FEATURE__IPS_FW].data.ips_fw;
+		return ips_fw->signals.bits.detection_required;
+	}
+
+	/* Detection may require reading scratch 0 - exit out of idle prior to the read. */
+	if (dc_dmub_srv->idle_allowed) {
+		dc_dmub_srv_apply_idle_power_optimizations(dc_dmub_srv->ctx->dc, false);
+		reallow_idle = true;
+	}
+
+	should_detect = dmub_srv_should_detect(dc_dmub_srv->dmub);
+
+	/* Re-enter idle if we're not about to immediately redetect links. */
+	if (!should_detect && reallow_idle && dc_dmub_srv->idle_exit_counter == 0 &&
+	    !dc_dmub_srv->ctx->dc->debug.disable_dmub_reallow_idle)
+		dc_dmub_srv_apply_idle_power_optimizations(dc_dmub_srv->ctx->dc, true);
+
+	return should_detect;
+}
+
+void dc_dmub_srv_apply_idle_power_optimizations(const struct dc *dc, bool allow_idle)
+{
+	struct dc_dmub_srv *dc_dmub_srv = dc->ctx->dmub_srv;
+
+	if (!dc_dmub_srv || !dc_dmub_srv->dmub)
+		return;
+
+	allow_idle &= (!dc->debug.ips_disallow_entry);
+
+	if (dc_dmub_srv->idle_allowed == allow_idle)
+		return;
+
+	DC_LOG_IPS("%s state change: old=%d new=%d", __func__, dc_dmub_srv->idle_allowed, allow_idle);
+
+	/*
+	 * Entering a low power state requires a driver notification.
+	 * Powering up the hardware requires notifying PMFW and DMCUB.
+	 * Clearing the driver idle allow requires a DMCUB command.
+	 * DMCUB commands requires the DMCUB to be powered up and restored.
+	 */
+
+	if (!allow_idle) {
+		dc_dmub_srv->idle_exit_counter += 1;
+
+		dc_dmub_srv_exit_low_power_state(dc);
+		/*
+		 * Idle is considered fully exited only after the sequence above
+		 * fully completes. If we have a race of two threads exiting
+		 * at the same time then it's safe to perform the sequence
+		 * twice as long as we're not re-entering.
+		 *
+		 * Infinite command submission is avoided by using the
+		 * dm_execute_dmub_cmd submission instead of the "wake" helpers.
+		 */
+		dc_dmub_srv->idle_allowed = false;
+
+		dc_dmub_srv->idle_exit_counter -= 1;
+		if (dc_dmub_srv->idle_exit_counter < 0) {
+			ASSERT(0);
+			dc_dmub_srv->idle_exit_counter = 0;
+		}
+	} else {
+		/* Consider idle as notified prior to the actual submission to
+		 * prevent multiple entries. */
+		dc_dmub_srv->idle_allowed = true;
+
+		dc_dmub_srv_notify_idle(dc, allow_idle);
+	}
+}
+
+bool dc_wake_and_execute_dmub_cmd(const struct dc_context *ctx, union dmub_rb_cmd *cmd,
+				  enum dm_dmub_wait_type wait_type)
+{
+	return dc_wake_and_execute_dmub_cmd_list(ctx, 1, cmd, wait_type);
+}
+
+bool dc_wake_and_execute_dmub_cmd_list(const struct dc_context *ctx, unsigned int count,
+				       union dmub_rb_cmd *cmd, enum dm_dmub_wait_type wait_type)
+{
+	struct dc_dmub_srv *dc_dmub_srv = ctx->dmub_srv;
+	bool result = false, reallow_idle = false;
+
+	if (!dc_dmub_srv || !dc_dmub_srv->dmub)
+		return false;
+
+	if (count == 0)
+		return true;
+
+	if (dc_dmub_srv->idle_allowed) {
+		dc_dmub_srv_apply_idle_power_optimizations(ctx->dc, false);
+		reallow_idle = true;
+	}
+
+	/*
+	 * These may have different implementations in DM, so ensure
+	 * that we guide it to the expected helper.
+	 */
+	if (count > 1)
+		result = dm_execute_dmub_cmd_list(ctx, count, cmd, wait_type);
+	else
+		result = dm_execute_dmub_cmd(ctx, cmd, wait_type);
+
+	if (result && reallow_idle && dc_dmub_srv->idle_exit_counter == 0 &&
+	    !ctx->dc->debug.disable_dmub_reallow_idle)
+		dc_dmub_srv_apply_idle_power_optimizations(ctx->dc, true);
+
+	return result;
+}
+
+static bool dc_dmub_execute_gpint(const struct dc_context *ctx, enum dmub_gpint_command command_code,
+				  uint16_t param, uint32_t *response, enum dm_dmub_wait_type wait_type)
+{
+	struct dc_dmub_srv *dc_dmub_srv = ctx->dmub_srv;
+	const uint32_t wait_us = wait_type == DM_DMUB_WAIT_TYPE_NO_WAIT ? 0 : 30;
+	enum dmub_status status;
+
+	if (response)
+		*response = 0;
+
+	if (!dc_dmub_srv || !dc_dmub_srv->dmub)
+		return false;
+
+	status = dmub_srv_send_gpint_command(dc_dmub_srv->dmub, command_code, param, wait_us);
+	if (status != DMUB_STATUS_OK) {
+		if (status == DMUB_STATUS_TIMEOUT && wait_type == DM_DMUB_WAIT_TYPE_NO_WAIT)
+			return true;
+
+		return false;
+	}
+
+	if (response && wait_type == DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY)
+		dmub_srv_get_gpint_response(dc_dmub_srv->dmub, response);
+
+	return true;
+}
+
+bool dc_wake_and_execute_gpint(const struct dc_context *ctx, enum dmub_gpint_command command_code,
+			       uint16_t param, uint32_t *response, enum dm_dmub_wait_type wait_type)
+{
+	struct dc_dmub_srv *dc_dmub_srv = ctx->dmub_srv;
+	bool result = false, reallow_idle = false;
+
+	if (!dc_dmub_srv || !dc_dmub_srv->dmub)
+		return false;
+
+	if (dc_dmub_srv->idle_allowed) {
+		dc_dmub_srv_apply_idle_power_optimizations(ctx->dc, false);
+		reallow_idle = true;
+	}
+
+	result = dc_dmub_execute_gpint(ctx, command_code, param, response, wait_type);
+
+	if (result && reallow_idle && dc_dmub_srv->idle_exit_counter == 0 &&
+	    !ctx->dc->debug.disable_dmub_reallow_idle)
+		dc_dmub_srv_apply_idle_power_optimizations(ctx->dc, true);
+
+	return result;
+}
+
+static void dc_dmub_srv_rb_based_fams2_update_config(struct dc *dc,
+		struct dc_state *context,
+		bool enable)
+{
+	uint8_t num_cmds = 1;
+	uint32_t i;
+	union dmub_rb_cmd cmd[2 * MAX_STREAMS + 1];
+	struct dmub_rb_cmd_fams2 *global_cmd = &cmd[0].fams2_config;
+
+	memset(cmd, 0, sizeof(union dmub_rb_cmd) * (2 * MAX_STREAMS + 1));
+	/* fill in generic command header */
+	global_cmd->header.type = DMUB_CMD__FW_ASSISTED_MCLK_SWITCH;
+	global_cmd->header.sub_type = DMUB_CMD__FAMS2_CONFIG;
+	global_cmd->header.payload_bytes =
+			sizeof(struct dmub_rb_cmd_fams2) - sizeof(struct dmub_cmd_header);
+
+	if (enable) {
+		/* send global configuration parameters */
+		memcpy(&global_cmd->config.global, &context->bw_ctx.bw.dcn.fams2_global_config, sizeof(struct dmub_cmd_fams2_global_config));
+
+		/* copy static feature configuration overrides */
+		global_cmd->config.global.features.bits.enable_stall_recovery = dc->debug.fams2_config.bits.enable_stall_recovery;
+		global_cmd->config.global.features.bits.enable_debug = dc->debug.fams2_config.bits.enable_debug;
+		global_cmd->config.global.features.bits.enable_offload_flip = dc->debug.fams2_config.bits.enable_offload_flip;
+
+		/* construct per-stream configs */
+		for (i = 0; i < context->bw_ctx.bw.dcn.fams2_global_config.num_streams; i++) {
+			struct dmub_rb_cmd_fams2 *stream_base_cmd = &cmd[i+1].fams2_config;
+			struct dmub_rb_cmd_fams2 *stream_sub_state_cmd = &cmd[i+1+context->bw_ctx.bw.dcn.fams2_global_config.num_streams].fams2_config;
+
+			/* configure command header */
+			stream_base_cmd->header.type = DMUB_CMD__FW_ASSISTED_MCLK_SWITCH;
+			stream_base_cmd->header.sub_type = DMUB_CMD__FAMS2_CONFIG;
+			stream_base_cmd->header.payload_bytes =
+					sizeof(struct dmub_rb_cmd_fams2) - sizeof(struct dmub_cmd_header);
+			stream_base_cmd->header.multi_cmd_pending = 1;
+			stream_sub_state_cmd->header.type = DMUB_CMD__FW_ASSISTED_MCLK_SWITCH;
+			stream_sub_state_cmd->header.sub_type = DMUB_CMD__FAMS2_CONFIG;
+			stream_sub_state_cmd->header.payload_bytes =
+					sizeof(struct dmub_rb_cmd_fams2) - sizeof(struct dmub_cmd_header);
+			stream_sub_state_cmd->header.multi_cmd_pending = 1;
+			/* copy stream static base state */
+			memcpy(&stream_base_cmd->config,
+					&context->bw_ctx.bw.dcn.fams2_stream_base_params[i],
+					sizeof(union dmub_cmd_fams2_config));
+			/* copy stream static sub state */
+			memcpy(&stream_sub_state_cmd->config,
+					&context->bw_ctx.bw.dcn.fams2_stream_sub_params[i],
+					sizeof(union dmub_cmd_fams2_config));
+		}
+	}
+
+	/* apply feature configuration based on current driver state */
+	global_cmd->config.global.features.bits.enable_visual_confirm = dc->debug.visual_confirm == VISUAL_CONFIRM_FAMS2;
+	global_cmd->config.global.features.bits.enable = enable;
+
+	if (enable && context->bw_ctx.bw.dcn.fams2_global_config.features.bits.enable) {
+		/* set multi pending for global, and unset for last stream cmd */
+		global_cmd->header.multi_cmd_pending = 1;
+		cmd[2 * context->bw_ctx.bw.dcn.fams2_global_config.num_streams].fams2_config.header.multi_cmd_pending = 0;
+		num_cmds += 2 * context->bw_ctx.bw.dcn.fams2_global_config.num_streams;
+	}
+
+	dm_execute_dmub_cmd_list(dc->ctx, num_cmds, cmd, DM_DMUB_WAIT_TYPE_WAIT);
+}
+
+static void dc_dmub_srv_ib_based_fams2_update_config(struct dc *dc,
+		struct dc_state *context,
+		bool enable)
+{
+	struct dmub_fams2_config_v2 *config = (struct dmub_fams2_config_v2 *)dc->ctx->dmub_srv->dmub->ib_mem_gart.cpu_addr;
+	union dmub_rb_cmd cmd;
+	uint32_t i;
+
+	memset(config, 0, sizeof(*config));
+	memset(&cmd, 0, sizeof(cmd));
+
+	cmd.ib_fams2_config.header.type = DMUB_CMD__FW_ASSISTED_MCLK_SWITCH;
+	cmd.ib_fams2_config.header.sub_type = DMUB_CMD__FAMS2_IB_CONFIG;
+
+	cmd.ib_fams2_config.ib_data.src.quad_part = dc->ctx->dmub_srv->dmub->ib_mem_gart.gpu_addr;
+	cmd.ib_fams2_config.ib_data.size = sizeof(*config);
+
+	if (enable && context->bw_ctx.bw.dcn.fams2_global_config.features.bits.enable) {
+		/* copy static feature configuration overrides */
+		config->global.features.bits.enable_stall_recovery = dc->debug.fams2_config.bits.enable_stall_recovery;
+		config->global.features.bits.enable_offload_flip = dc->debug.fams2_config.bits.enable_offload_flip;
+		config->global.features.bits.enable_debug = dc->debug.fams2_config.bits.enable_debug;
+
+		/* send global configuration parameters */
+		memcpy(&config->global, &context->bw_ctx.bw.dcn.fams2_global_config,
+			sizeof(struct dmub_cmd_fams2_global_config));
+
+		/* construct per-stream configs */
+		for (i = 0; i < context->bw_ctx.bw.dcn.fams2_global_config.num_streams; i++) {
+			/* copy stream static base state */
+			memcpy(&config->stream_v1[i].base,
+				&context->bw_ctx.bw.dcn.fams2_stream_base_params[i],
+				sizeof(config->stream_v1[i].base));
+
+			/* copy stream static sub-state */
+			memcpy(&config->stream_v1[i].sub_state,
+				&context->bw_ctx.bw.dcn.fams2_stream_sub_params_v2[i],
+				sizeof(config->stream_v1[i].sub_state));
+		}
+	}
+
+	config->global.features.bits.enable_visual_confirm = dc->debug.visual_confirm == VISUAL_CONFIRM_FAMS2;
+	config->global.features.bits.enable = enable;
+
+	dm_execute_dmub_cmd_list(dc->ctx, 1, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
+}
+
+void dc_dmub_srv_fams2_update_config(struct dc *dc,
+		struct dc_state *context,
+		bool enable)
+{
+	if (dc->debug.fams_version.major == 2)
+		dc_dmub_srv_rb_based_fams2_update_config(dc, context, enable);
+	if (dc->debug.fams_version.major == 3)
+		dc_dmub_srv_ib_based_fams2_update_config(dc, context, enable);
+}
+
+void dc_dmub_srv_fams2_drr_update(struct dc *dc,
+		uint32_t tg_inst,
+		uint32_t vtotal_min,
+		uint32_t vtotal_max,
+		uint32_t vtotal_mid,
+		uint32_t vtotal_mid_frame_num,
+		bool program_manual_trigger)
+{
+	union dmub_rb_cmd cmd = { 0 };
+
+	cmd.fams2_drr_update.header.type = DMUB_CMD__FW_ASSISTED_MCLK_SWITCH;
+	cmd.fams2_drr_update.header.sub_type = DMUB_CMD__FAMS2_DRR_UPDATE;
+	cmd.fams2_drr_update.dmub_optc_state_req.tg_inst = tg_inst;
+	cmd.fams2_drr_update.dmub_optc_state_req.v_total_max = vtotal_max;
+	cmd.fams2_drr_update.dmub_optc_state_req.v_total_min = vtotal_min;
+	cmd.fams2_drr_update.dmub_optc_state_req.v_total_mid = vtotal_mid;
+	cmd.fams2_drr_update.dmub_optc_state_req.v_total_mid_frame_num = vtotal_mid_frame_num;
+	cmd.fams2_drr_update.dmub_optc_state_req.program_manual_trigger = program_manual_trigger;
+
+	cmd.fams2_drr_update.header.payload_bytes =
+			sizeof(cmd.fams2_drr_update) - sizeof(cmd.fams2_drr_update.header);
+
+	dm_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
+}
+
+void dc_dmub_srv_fams2_passthrough_flip(
+		struct dc *dc,
+		struct dc_state *state,
+		struct dc_stream_state *stream,
+		struct dc_surface_update *srf_updates,
+		int surface_count)
+{
+	int plane_index;
+	union dmub_rb_cmd cmds[MAX_PLANES];
+	struct dc_plane_address *address;
+	struct dc_plane_state *plane_state;
+	int num_cmds = 0;
+	struct dc_stream_status *stream_status = dc_stream_get_status(stream);
+
+	if (surface_count <= 0 || stream_status == NULL)
+		return;
+
+	memset(cmds, 0, sizeof(union dmub_rb_cmd) * MAX_PLANES);
+
+	/* build command for each surface update */
+	for (plane_index = 0; plane_index < surface_count; plane_index++) {
+		plane_state = srf_updates[plane_index].surface;
+		address = &plane_state->address;
+
+		/* skip if there is no address update for plane */
+		if (!srf_updates[plane_index].flip_addr)
+			continue;
+
+		/* build command header */
+		cmds[num_cmds].fams2_flip.header.type = DMUB_CMD__FW_ASSISTED_MCLK_SWITCH;
+		cmds[num_cmds].fams2_flip.header.sub_type = DMUB_CMD__FAMS2_FLIP;
+		cmds[num_cmds].fams2_flip.header.payload_bytes =
+				sizeof(struct dmub_rb_cmd_fams2_flip) - sizeof(struct dmub_cmd_header);
+
+		/* for chaining multiple commands, all but last command should set to 1 */
+		cmds[num_cmds].fams2_flip.header.multi_cmd_pending = 1;
+
+		/* set topology info */
+		cmds[num_cmds].fams2_flip.flip_info.pipe_mask = dc_plane_get_pipe_mask(state, plane_state);
+		if (stream_status)
+			cmds[num_cmds].fams2_flip.flip_info.otg_inst = stream_status->primary_otg_inst;
+
+		cmds[num_cmds].fams2_flip.flip_info.config.bits.is_immediate = plane_state->flip_immediate;
+
+		/* build address info for command */
+		switch (address->type) {
+		case PLN_ADDR_TYPE_GRAPHICS:
+			if (address->grph.addr.quad_part == 0) {
+				BREAK_TO_DEBUGGER();
+				break;
+			}
+
+			cmds[num_cmds].fams2_flip.flip_info.addr_info.meta_addr_lo =
+					address->grph.meta_addr.low_part;
+			cmds[num_cmds].fams2_flip.flip_info.addr_info.meta_addr_hi =
+					(uint16_t)address->grph.meta_addr.high_part;
+			cmds[num_cmds].fams2_flip.flip_info.addr_info.surf_addr_lo =
+					address->grph.addr.low_part;
+			cmds[num_cmds].fams2_flip.flip_info.addr_info.surf_addr_hi =
+					(uint16_t)address->grph.addr.high_part;
+			break;
+		case PLN_ADDR_TYPE_VIDEO_PROGRESSIVE:
+			if (address->video_progressive.luma_addr.quad_part == 0 ||
+				address->video_progressive.chroma_addr.quad_part == 0) {
+				BREAK_TO_DEBUGGER();
+				break;
+			}
+
+			cmds[num_cmds].fams2_flip.flip_info.addr_info.meta_addr_lo =
+					address->video_progressive.luma_meta_addr.low_part;
+			cmds[num_cmds].fams2_flip.flip_info.addr_info.meta_addr_hi =
+					(uint16_t)address->video_progressive.luma_meta_addr.high_part;
+			cmds[num_cmds].fams2_flip.flip_info.addr_info.meta_addr_c_lo =
+					address->video_progressive.chroma_meta_addr.low_part;
+			cmds[num_cmds].fams2_flip.flip_info.addr_info.meta_addr_c_hi =
+					(uint16_t)address->video_progressive.chroma_meta_addr.high_part;
+			cmds[num_cmds].fams2_flip.flip_info.addr_info.surf_addr_lo =
+					address->video_progressive.luma_addr.low_part;
+			cmds[num_cmds].fams2_flip.flip_info.addr_info.surf_addr_hi =
+					(uint16_t)address->video_progressive.luma_addr.high_part;
+			cmds[num_cmds].fams2_flip.flip_info.addr_info.surf_addr_c_lo =
+					address->video_progressive.chroma_addr.low_part;
+			cmds[num_cmds].fams2_flip.flip_info.addr_info.surf_addr_c_hi =
+					(uint16_t)address->video_progressive.chroma_addr.high_part;
+			break;
+		default:
+			// Should never be hit
+			BREAK_TO_DEBUGGER();
+			break;
+		}
+
+		num_cmds++;
+	}
+
+	if (num_cmds > 0)  {
+		cmds[num_cmds - 1].fams2_flip.header.multi_cmd_pending = 0;
+		dm_execute_dmub_cmd_list(dc->ctx, num_cmds, cmds, DM_DMUB_WAIT_TYPE_WAIT);
+	}
+}
+
+
+bool dc_dmub_srv_ips_residency_cntl(const struct dc_context *ctx, uint8_t panel_inst, bool start_measurement)
+{
+	union dmub_rb_cmd cmd;
+
+	memset(&cmd, 0, sizeof(cmd));
+
+	cmd.ips_residency_cntl.header.type = DMUB_CMD__IPS;
+	cmd.ips_residency_cntl.header.sub_type = DMUB_CMD__IPS_RESIDENCY_CNTL;
+	cmd.ips_residency_cntl.header.payload_bytes = sizeof(struct dmub_cmd_ips_residency_cntl_data);
+
+	// only panel_inst=0 is supported at the moment
+	cmd.ips_residency_cntl.cntl_data.panel_inst = panel_inst;
+	cmd.ips_residency_cntl.cntl_data.start_measurement = start_measurement;
+
+	if (!dc_wake_and_execute_dmub_cmd(ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY))
+		return false;
+
+	return true;
+}
+
+bool dc_dmub_srv_ips_query_residency_info(const struct dc_context *ctx, uint8_t panel_inst, struct dmub_ips_residency_info *driver_info,
+					  enum ips_residency_mode ips_mode)
+{
+	union dmub_rb_cmd cmd;
+	uint32_t bytes = sizeof(struct dmub_ips_residency_info);
+
+	dmub_flush_buffer_mem(&ctx->dmub_srv->dmub->scratch_mem_fb);
+	memset(&cmd, 0, sizeof(cmd));
+
+	cmd.ips_query_residency_info.header.type = DMUB_CMD__IPS;
+	cmd.ips_query_residency_info.header.sub_type = DMUB_CMD__IPS_QUERY_RESIDENCY_INFO;
+	cmd.ips_query_residency_info.header.payload_bytes = sizeof(struct dmub_cmd_ips_query_residency_info_data);
+
+	cmd.ips_query_residency_info.info_data.dest.quad_part = ctx->dmub_srv->dmub->scratch_mem_fb.gpu_addr;
+	cmd.ips_query_residency_info.info_data.size = bytes;
+	cmd.ips_query_residency_info.info_data.panel_inst = panel_inst;
+	cmd.ips_query_residency_info.info_data.ips_mode = (uint32_t)ips_mode;
+
+	if (!dc_wake_and_execute_dmub_cmd(ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY) ||
+					  cmd.ips_query_residency_info.header.ret_status == 0)
+		return false;
+
+	// copy the result to the output since ret_status != 0 means the command returned data
+	memcpy(driver_info, ctx->dmub_srv->dmub->scratch_mem_fb.cpu_addr, bytes);
+
+	return true;
+}
+
+bool dmub_lsdma_init(struct dc_dmub_srv *dc_dmub_srv)
+{
+	struct dc_context *dc_ctx = dc_dmub_srv->ctx;
+	union dmub_rb_cmd cmd;
+	enum dm_dmub_wait_type wait_type;
+	struct dmub_cmd_lsdma_data *lsdma_data = &cmd.lsdma.lsdma_data;
+	bool result;
+
+	if (!dc_dmub_srv->dmub->feature_caps.lsdma_support_in_dmu)
+		return false;
+
+	memset(&cmd, 0, sizeof(cmd));
+
+	cmd.cmd_common.header.type     = DMUB_CMD__LSDMA;
+	cmd.cmd_common.header.sub_type = DMUB_CMD__LSDMA_INIT_CONFIG;
+	wait_type                      = DM_DMUB_WAIT_TYPE_NO_WAIT;
+
+	lsdma_data->u.init_data.gpu_addr_base.quad_part = dc_ctx->dmub_srv->dmub->lsdma_rb_fb.gpu_addr;
+	lsdma_data->u.init_data.ring_size               = dc_ctx->dmub_srv->dmub->lsdma_rb_fb.size;
+
+	result = dc_wake_and_execute_dmub_cmd(dc_ctx, &cmd, wait_type);
+
+	if (!result)
+		DC_ERROR("LSDMA Init failed in DMUB");
+
+	return result;
+}
+
+bool dmub_lsdma_send_linear_copy_command(
+	struct dc_dmub_srv *dc_dmub_srv,
+	uint64_t src_addr,
+	uint64_t dst_addr,
+	uint32_t count
+)
+{
+	struct dc_context *dc_ctx = dc_dmub_srv->ctx;
+	union dmub_rb_cmd cmd;
+	enum dm_dmub_wait_type wait_type;
+	struct dmub_cmd_lsdma_data *lsdma_data = &cmd.lsdma.lsdma_data;
+	bool result;
+
+	memset(&cmd, 0, sizeof(cmd));
+
+	cmd.cmd_common.header.type     = DMUB_CMD__LSDMA;
+	cmd.cmd_common.header.sub_type = DMUB_CMD__LSDMA_LINEAR_COPY;
+	wait_type                      = DM_DMUB_WAIT_TYPE_NO_WAIT;
+
+	lsdma_data->u.linear_copy_data.count   = count - 1; // LSDMA controller expects bytes to copy -1
+	lsdma_data->u.linear_copy_data.src_lo  = src_addr & 0xFFFFFFFF;
+	lsdma_data->u.linear_copy_data.src_hi  = (src_addr >> 32) & 0xFFFFFFFF;
+	lsdma_data->u.linear_copy_data.dst_lo  = dst_addr & 0xFFFFFFFF;
+	lsdma_data->u.linear_copy_data.dst_hi  = (dst_addr >> 32) & 0xFFFFFFFF;
+
+	result = dc_wake_and_execute_dmub_cmd(dc_ctx, &cmd, wait_type);
+
+	if (!result)
+		DC_ERROR("LSDMA Linear Copy failed in DMUB");
+
+	return result;
+}
+
+bool dmub_lsdma_send_linear_sub_window_copy_command(
+	struct dc_dmub_srv *dc_dmub_srv,
+	struct lsdma_linear_sub_window_copy_params copy_data
+)
+{
+	struct dc_context *dc_ctx = dc_dmub_srv->ctx;
+	union dmub_rb_cmd cmd;
+	enum dm_dmub_wait_type wait_type;
+	struct dmub_cmd_lsdma_data *lsdma_data = &cmd.lsdma.lsdma_data;
+	bool result;
+
+	memset(&cmd, 0, sizeof(cmd));
+
+	cmd.cmd_common.header.type     = DMUB_CMD__LSDMA;
+	cmd.cmd_common.header.sub_type = DMUB_CMD__LSDMA_LINEAR_SUB_WINDOW_COPY;
+	wait_type                      = DM_DMUB_WAIT_TYPE_NO_WAIT;
+
+	lsdma_data->u.linear_sub_window_copy_data.tmz              = copy_data.tmz;
+	lsdma_data->u.linear_sub_window_copy_data.element_size     = copy_data.element_size;
+	lsdma_data->u.linear_sub_window_copy_data.src_lo           = copy_data.src_lo;
+	lsdma_data->u.linear_sub_window_copy_data.src_hi           = copy_data.src_hi;
+	lsdma_data->u.linear_sub_window_copy_data.src_x            = copy_data.src_x;
+	lsdma_data->u.linear_sub_window_copy_data.src_y            = copy_data.src_y;
+	lsdma_data->u.linear_sub_window_copy_data.src_pitch        = copy_data.src_pitch;
+	lsdma_data->u.linear_sub_window_copy_data.src_slice_pitch  = copy_data.src_slice_pitch;
+	lsdma_data->u.linear_sub_window_copy_data.dst_lo           = copy_data.dst_lo;
+	lsdma_data->u.linear_sub_window_copy_data.dst_hi           = copy_data.dst_hi;
+	lsdma_data->u.linear_sub_window_copy_data.dst_x            = copy_data.dst_x;
+	lsdma_data->u.linear_sub_window_copy_data.dst_y            = copy_data.dst_y;
+	lsdma_data->u.linear_sub_window_copy_data.dst_pitch        = copy_data.dst_pitch;
+	lsdma_data->u.linear_sub_window_copy_data.dst_slice_pitch  = copy_data.dst_slice_pitch;
+	lsdma_data->u.linear_sub_window_copy_data.rect_x           = copy_data.rect_x;
+	lsdma_data->u.linear_sub_window_copy_data.rect_y           = copy_data.rect_y;
+	lsdma_data->u.linear_sub_window_copy_data.src_cache_policy = copy_data.src_cache_policy;
+	lsdma_data->u.linear_sub_window_copy_data.dst_cache_policy = copy_data.dst_cache_policy;
+
+	result = dc_wake_and_execute_dmub_cmd(dc_ctx, &cmd, wait_type);
+
+	if (!result)
+		DC_ERROR("LSDMA Linear Sub Window Copy failed in DMUB");
+
+	return result;
+}
+
+bool dmub_lsdma_send_tiled_to_tiled_copy_command(
+	struct dc_dmub_srv *dc_dmub_srv,
+	struct lsdma_send_tiled_to_tiled_copy_command_params params
+)
+{
+	struct dc_context *dc_ctx = dc_dmub_srv->ctx;
+	union dmub_rb_cmd cmd;
+	enum dm_dmub_wait_type wait_type;
+	struct dmub_cmd_lsdma_data *lsdma_data = &cmd.lsdma.lsdma_data;
+	bool result;
+
+	memset(&cmd, 0, sizeof(cmd));
+
+	cmd.cmd_common.header.type     = DMUB_CMD__LSDMA;
+	cmd.cmd_common.header.sub_type = DMUB_CMD__LSDMA_TILED_TO_TILED_COPY;
+	wait_type                      = DM_DMUB_WAIT_TYPE_NO_WAIT;
+
+	lsdma_data->u.tiled_copy_data.src_addr_lo      = params.src_addr & 0xFFFFFFFF;
+	lsdma_data->u.tiled_copy_data.src_addr_hi      = (params.src_addr >> 32) & 0xFFFFFFFF;
+	lsdma_data->u.tiled_copy_data.dst_addr_lo      = params.dst_addr & 0xFFFFFFFF;
+	lsdma_data->u.tiled_copy_data.dst_addr_hi      = (params.dst_addr >> 32) & 0xFFFFFFFF;
+	lsdma_data->u.tiled_copy_data.src_x            = params.src_x;
+	lsdma_data->u.tiled_copy_data.src_y            = params.src_y;
+	lsdma_data->u.tiled_copy_data.dst_x            = params.dst_x;
+	lsdma_data->u.tiled_copy_data.dst_y            = params.dst_y;
+	lsdma_data->u.tiled_copy_data.src_width        = params.src_width;
+	lsdma_data->u.tiled_copy_data.dst_width        = params.dst_width;
+	lsdma_data->u.tiled_copy_data.src_swizzle_mode = params.swizzle_mode;
+	lsdma_data->u.tiled_copy_data.dst_swizzle_mode = params.swizzle_mode;
+	lsdma_data->u.tiled_copy_data.src_element_size = params.element_size;
+	lsdma_data->u.tiled_copy_data.dst_element_size = params.element_size;
+	lsdma_data->u.tiled_copy_data.rect_x           = params.rect_x;
+	lsdma_data->u.tiled_copy_data.rect_y           = params.rect_y;
+	lsdma_data->u.tiled_copy_data.dcc              = params.dcc;
+	lsdma_data->u.tiled_copy_data.tmz              = params.tmz;
+	lsdma_data->u.tiled_copy_data.read_compress    = params.read_compress;
+	lsdma_data->u.tiled_copy_data.write_compress   = params.write_compress;
+	lsdma_data->u.tiled_copy_data.src_height       = params.src_height;
+	lsdma_data->u.tiled_copy_data.dst_height       = params.dst_height;
+	lsdma_data->u.tiled_copy_data.data_format      = params.data_format;
+	lsdma_data->u.tiled_copy_data.max_com          = params.max_com;
+	lsdma_data->u.tiled_copy_data.max_uncom        = params.max_uncom;
+
+	result = dc_wake_and_execute_dmub_cmd(dc_ctx, &cmd, wait_type);
+
+	if (!result)
+		DC_ERROR("LSDMA Tiled to Tiled Copy failed in DMUB");
+
+	return result;
+}
+
+bool dmub_lsdma_send_pio_copy_command(
+	struct dc_dmub_srv *dc_dmub_srv,
+	uint64_t src_addr,
+	uint64_t dst_addr,
+	uint32_t byte_count,
+	uint32_t overlap_disable
+)
+{
+	struct dc_context *dc_ctx = dc_dmub_srv->ctx;
+	union dmub_rb_cmd cmd;
+	enum dm_dmub_wait_type wait_type;
+	struct dmub_cmd_lsdma_data *lsdma_data = &cmd.lsdma.lsdma_data;
+	bool result;
+
+	memset(&cmd, 0, sizeof(cmd));
+
+	cmd.cmd_common.header.type     = DMUB_CMD__LSDMA;
+	cmd.cmd_common.header.sub_type = DMUB_CMD__LSDMA_PIO_COPY;
+	wait_type                      = DM_DMUB_WAIT_TYPE_NO_WAIT;
+
+	lsdma_data->u.pio_copy_data.packet.fields.byte_count      = byte_count;
+	lsdma_data->u.pio_copy_data.packet.fields.overlap_disable = overlap_disable;
+	lsdma_data->u.pio_copy_data.src_lo                        = src_addr & 0xFFFFFFFF;
+	lsdma_data->u.pio_copy_data.src_hi                        = (src_addr >> 32) & 0xFFFFFFFF;
+	lsdma_data->u.pio_copy_data.dst_lo                        = dst_addr & 0xFFFFFFFF;
+	lsdma_data->u.pio_copy_data.dst_hi                        = (dst_addr >> 32) & 0xFFFFFFFF;
+
+	result = dc_wake_and_execute_dmub_cmd(dc_ctx, &cmd, wait_type);
+
+	if (!result)
+		DC_ERROR("LSDMA PIO Copy failed in DMUB");
+
+	return result;
+}
+
+bool dmub_lsdma_send_pio_constfill_command(
+	struct dc_dmub_srv *dc_dmub_srv,
+	uint64_t dst_addr,
+	uint32_t byte_count,
+	uint32_t data
+)
+{
+	struct dc_context *dc_ctx = dc_dmub_srv->ctx;
+	union dmub_rb_cmd cmd;
+	enum dm_dmub_wait_type wait_type;
+	struct dmub_cmd_lsdma_data *lsdma_data = &cmd.lsdma.lsdma_data;
+	bool result;
+
+	memset(&cmd, 0, sizeof(cmd));
+
+	cmd.cmd_common.header.type     = DMUB_CMD__LSDMA;
+	cmd.cmd_common.header.sub_type = DMUB_CMD__LSDMA_PIO_CONSTFILL;
+	wait_type                      = DM_DMUB_WAIT_TYPE_NO_WAIT;
+
+	lsdma_data->u.pio_constfill_data.packet.fields.constant_fill = 1;
+	lsdma_data->u.pio_constfill_data.packet.fields.byte_count    = byte_count;
+	lsdma_data->u.pio_constfill_data.dst_lo                      = dst_addr & 0xFFFFFFFF;
+	lsdma_data->u.pio_constfill_data.dst_hi                      = (dst_addr >> 32) & 0xFFFFFFFF;
+	lsdma_data->u.pio_constfill_data.data                        = data;
+
+	result = dc_wake_and_execute_dmub_cmd(dc_ctx, &cmd, wait_type);
+
+	if (!result)
+		DC_ERROR("LSDMA PIO Constfill failed in DMUB");
+
+	return result;
+}
+
+bool dmub_lsdma_send_poll_reg_write_command(struct dc_dmub_srv *dc_dmub_srv, uint32_t reg_addr, uint32_t reg_data)
+{
+	struct dc_context *dc_ctx = dc_dmub_srv->ctx;
+	union dmub_rb_cmd cmd;
+	enum dm_dmub_wait_type wait_type;
+	struct dmub_cmd_lsdma_data *lsdma_data = &cmd.lsdma.lsdma_data;
+	bool result;
+
+	memset(&cmd, 0, sizeof(cmd));
+
+	cmd.cmd_common.header.type     = DMUB_CMD__LSDMA;
+	cmd.cmd_common.header.sub_type = DMUB_CMD__LSDMA_POLL_REG_WRITE;
+	wait_type                      = DM_DMUB_WAIT_TYPE_NO_WAIT;
+
+	lsdma_data->u.reg_write_data.reg_addr = reg_addr;
+	lsdma_data->u.reg_write_data.reg_data = reg_data;
+
+	result = dc_wake_and_execute_dmub_cmd(dc_ctx, &cmd, wait_type);
+
+	if (!result)
+		DC_ERROR("LSDMA Poll Reg failed in DMUB");
+
+	return result;
+}
+
+bool dc_dmub_srv_is_cursor_offload_enabled(const struct dc *dc)
+{
+	return dc->ctx->dmub_srv && dc->ctx->dmub_srv->cursor_offload_enabled;
+}
+
+void dc_dmub_srv_release_hw(const struct dc *dc)
+{
+	struct dc_dmub_srv *dc_dmub_srv = dc->ctx->dmub_srv;
+	union dmub_rb_cmd cmd = {0};
+
+	if (!dc_dmub_srv || !dc_dmub_srv->dmub)
+		return;
+
+	memset(&cmd, 0, sizeof(cmd));
+	cmd.idle_opt_notify_idle.header.type = DMUB_CMD__IDLE_OPT;
+	cmd.idle_opt_notify_idle.header.sub_type = DMUB_CMD__IDLE_OPT_RELEASE_HW;
+	cmd.idle_opt_notify_idle.header.payload_bytes =
+		sizeof(cmd.idle_opt_notify_idle) -
+		sizeof(cmd.idle_opt_notify_idle.header);
+
+	dm_execute_dmub_cmd(dc->ctx, &cmd,  DM_DMUB_WAIT_TYPE_WAIT);
+}
+
+void dc_dmub_srv_log_preos_dmcub_info(struct dc_dmub_srv *dc_dmub_srv)
+{
+	struct dmub_srv *dmub;
+
+	if (!dc_dmub_srv || !dc_dmub_srv->dmub)
+		return;
+
+	dmub = dc_dmub_srv->dmub;
+
+	if (dmub_srv_get_preos_info(dmub)) {
+		DC_LOG_DEBUG("%s: PreOS DMCUB Info", __func__);
+		DC_LOG_DEBUG("fw_version				: 0x%08x", dmub->preos_info.fw_version);
+		DC_LOG_DEBUG("boot_options				: 0x%08x", dmub->preos_info.boot_options);
+		DC_LOG_DEBUG("boot_status				: 0x%08x", dmub->preos_info.boot_status);
+		DC_LOG_DEBUG("trace_buffer_phy_addr		: 0x%016llx", dmub->preos_info.trace_buffer_phy_addr);
+		DC_LOG_DEBUG("trace_buffer_size_bytes	: 0x%08x", dmub->preos_info.trace_buffer_size);
+		DC_LOG_DEBUG("fb_base					: 0x%016llx", dmub->preos_info.fb_base);
+		DC_LOG_DEBUG("fb_offset					: 0x%016llx", dmub->preos_info.fb_offset);
+	}
 }
diff --git a/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.h b/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.h
index 3e35eee7188c..72e0a41f39f0 100644
--- a/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.h
+++ b/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.h
@@ -26,11 +26,16 @@
 #ifndef _DMUB_DC_SRV_H_
 #define _DMUB_DC_SRV_H_
 
-#include "os_types.h"
+#include "dm_services_types.h"
 #include "dmub/dmub_srv.h"
 
 struct dmub_srv;
 struct dc;
+struct pipe_ctx;
+struct dc_crtc_timing_adjust;
+struct dc_crtc_timing;
+struct dc_state;
+struct dc_surface_update;
 
 struct dc_reg_helper_state {
 	bool gather_in_progress;
@@ -46,21 +51,32 @@ struct dc_dmub_srv {
 
 	struct dc_context *ctx;
 	void *dm;
+
+	int32_t idle_exit_counter;
+	union dmub_shared_state_ips_driver_signals driver_signals;
+	bool idle_allowed;
+	bool needs_idle_wake;
+	bool cursor_offload_enabled;
 };
 
-void dc_dmub_srv_cmd_queue(struct dc_dmub_srv *dc_dmub_srv,
-			   union dmub_rb_cmd *cmd);
+bool dc_dmub_srv_wait_for_pending(struct dc_dmub_srv *dc_dmub_srv);
+
+bool dc_dmub_srv_optimized_init_done(struct dc_dmub_srv *dc_dmub_srv);
 
-void dc_dmub_srv_cmd_execute(struct dc_dmub_srv *dc_dmub_srv);
+bool dc_dmub_srv_cmd_list_queue_execute(struct dc_dmub_srv *dc_dmub_srv,
+		unsigned int count,
+		union dmub_rb_cmd *cmd_list);
 
-void dc_dmub_srv_wait_idle(struct dc_dmub_srv *dc_dmub_srv);
+bool dc_dmub_srv_wait_for_idle(struct dc_dmub_srv *dc_dmub_srv,
+		enum dm_dmub_wait_type wait_type,
+		union dmub_rb_cmd *cmd_list);
 
-void dc_dmub_srv_wait_phy_init(struct dc_dmub_srv *dc_dmub_srv);
+bool dc_dmub_srv_cmd_run(struct dc_dmub_srv *dc_dmub_srv, union dmub_rb_cmd *cmd, enum dm_dmub_wait_type wait_type);
 
-bool dc_dmub_srv_cmd_with_reply_data(struct dc_dmub_srv *dc_dmub_srv, union dmub_rb_cmd *cmd);
+bool dc_dmub_srv_cmd_run_list(struct dc_dmub_srv *dc_dmub_srv, unsigned int count, union dmub_rb_cmd *cmd_list, enum dm_dmub_wait_type wait_type);
 
 bool dc_dmub_srv_notify_stream_mask(struct dc_dmub_srv *dc_dmub_srv,
-				    unsigned int stream_mask);
+				   unsigned int stream_mask);
 
 bool dc_dmub_srv_is_restore_required(struct dc_dmub_srv *dc_dmub_srv);
 
@@ -68,10 +84,294 @@ bool dc_dmub_srv_get_dmub_outbox0_msg(const struct dc *dc, struct dmcub_trace_bu
 
 void dc_dmub_trace_event_control(struct dc *dc, bool enable);
 
+void dc_dmub_srv_drr_update_cmd(struct dc *dc, uint32_t tg_inst, uint32_t vtotal_min, uint32_t vtotal_max);
+
+void dc_dmub_srv_set_drr_manual_trigger_cmd(struct dc *dc, uint32_t tg_inst);
+bool dc_dmub_srv_p_state_delegate(struct dc *dc, bool enable_pstate, struct dc_state *context);
+
+void dc_dmub_srv_query_caps_cmd(struct dc_dmub_srv *dc_dmub_srv);
+void dc_dmub_srv_get_visual_confirm_color_cmd(struct dc *dc, struct pipe_ctx *pipe_ctx);
+void dc_dmub_srv_clear_inbox0_ack(struct dc_dmub_srv *dmub_srv);
+void dc_dmub_srv_wait_for_inbox0_ack(struct dc_dmub_srv *dmub_srv);
 void dc_dmub_srv_send_inbox0_cmd(struct dc_dmub_srv *dmub_srv, union dmub_inbox0_data_register data);
 
-bool dc_dmub_srv_get_diagnostic_data(struct dc_dmub_srv *dc_dmub_srv, struct dmub_diagnostic_data *dmub_oca);
+bool dc_dmub_srv_get_diagnostic_data(struct dc_dmub_srv *dc_dmub_srv);
 
+void dc_dmub_setup_subvp_dmub_command(struct dc *dc, struct dc_state *context, bool enable);
 void dc_dmub_srv_log_diagnostic_data(struct dc_dmub_srv *dc_dmub_srv);
 
+void dc_send_update_cursor_info_to_dmu(struct pipe_ctx *pCtx, uint8_t pipe_idx);
+bool dc_dmub_check_min_version(struct dmub_srv *srv);
+
+void dc_dmub_srv_enable_dpia_trace(const struct dc *dc);
+void dc_dmub_srv_subvp_save_surf_addr(const struct dc_dmub_srv *dc_dmub_srv, const struct dc_plane_address *addr, uint8_t subvp_index);
+
+bool dc_dmub_srv_is_hw_pwr_up(struct dc_dmub_srv *dc_dmub_srv, bool wait);
+
+void dc_dmub_srv_apply_idle_power_optimizations(const struct dc *dc, bool allow_idle);
+
+/**
+ * dc_dmub_srv_set_power_state() - Sets the power state for DMUB service.
+ *
+ * Controls whether messaging the DMCUB or interfacing with it via HW register
+ * interaction is permittable.
+ *
+ * @dc_dmub_srv - The DC DMUB service pointer
+ * @power_state - the DC power state
+ */
+void dc_dmub_srv_set_power_state(struct dc_dmub_srv *dc_dmub_srv, enum dc_acpi_cm_power_state power_state);
+
+/**
+ * dc_dmub_srv_notify_fw_dc_power_state() - Notifies firmware of the DC power state.
+ *
+ * Differs from dc_dmub_srv_set_power_state in that it needs to access HW in order
+ * to message DMCUB of the state transition. Should come after the D0 exit and
+ * before D3 set power state.
+ *
+ * @dc_dmub_srv - The DC DMUB service pointer
+ * @power_state - the DC power state
+ */
+void dc_dmub_srv_notify_fw_dc_power_state(struct dc_dmub_srv *dc_dmub_srv,
+					  enum dc_acpi_cm_power_state power_state);
+
+/**
+ * @dc_dmub_srv_should_detect() - Checks if link detection is required.
+ *
+ * While in idle power states we may need driver to manually redetect in
+ * the case of a missing hotplug. Should be called from a polling timer.
+ *
+ * Return: true if redetection is required.
+ */
+bool dc_dmub_srv_should_detect(struct dc_dmub_srv *dc_dmub_srv);
+
+/**
+ * dc_wake_and_execute_dmub_cmd() - Wrapper for DMUB command execution.
+ *
+ * Refer to dc_wake_and_execute_dmub_cmd_list() for usage and limitations,
+ * This function is a convenience wrapper for a single command execution.
+ *
+ * @ctx: DC context
+ * @cmd: The command to send/receive
+ * @wait_type: The wait behavior for the execution
+ *
+ * Return: true on command submission success, false otherwise
+ */
+bool dc_wake_and_execute_dmub_cmd(const struct dc_context *ctx, union dmub_rb_cmd *cmd,
+				  enum dm_dmub_wait_type wait_type);
+
+/**
+ * dc_wake_and_execute_dmub_cmd_list() - Wrapper for DMUB command list execution.
+ *
+ * If the DMCUB hardware was asleep then it wakes the DMUB before
+ * executing the command and attempts to re-enter if the command
+ * submission was successful.
+ *
+ * This should be the preferred command submission interface provided
+ * the DC lock is acquired.
+ *
+ * Entry/exit out of idle power optimizations would need to be
+ * manually performed otherwise through dc_allow_idle_optimizations().
+ *
+ * @ctx: DC context
+ * @count: Number of commands to send/receive
+ * @cmd: Array of commands to send
+ * @wait_type: The wait behavior for the execution
+ *
+ * Return: true on command submission success, false otherwise
+ */
+bool dc_wake_and_execute_dmub_cmd_list(const struct dc_context *ctx, unsigned int count,
+				       union dmub_rb_cmd *cmd, enum dm_dmub_wait_type wait_type);
+
+/**
+ * dc_wake_and_execute_gpint()
+ *
+ * @ctx: DC context
+ * @command_code: The command ID to send to DMCUB
+ * @param: The parameter to message DMCUB
+ * @response: Optional response out value - may be NULL.
+ * @wait_type: The wait behavior for the execution
+ */
+bool dc_wake_and_execute_gpint(const struct dc_context *ctx, enum dmub_gpint_command command_code,
+			       uint16_t param, uint32_t *response, enum dm_dmub_wait_type wait_type);
+
+void dc_dmub_srv_fams2_update_config(struct dc *dc,
+		struct dc_state *context,
+		bool enable);
+void dc_dmub_srv_fams2_drr_update(struct dc *dc,
+		uint32_t tg_inst,
+		uint32_t vtotal_min,
+		uint32_t vtotal_max,
+		uint32_t vtotal_mid,
+		uint32_t vtotal_mid_frame_num,
+		bool program_manual_trigger);
+void dc_dmub_srv_fams2_passthrough_flip(
+		struct dc *dc,
+		struct dc_state *state,
+		struct dc_stream_state *stream,
+		struct dc_surface_update *srf_updates,
+		int surface_count);
+
+bool dmub_lsdma_init(struct dc_dmub_srv *dc_dmub_srv);
+bool dmub_lsdma_send_linear_copy_command(
+	struct dc_dmub_srv *dc_dmub_srv,
+	uint64_t src_addr,
+	uint64_t dst_addr,
+	uint32_t count);
+
+struct lsdma_linear_sub_window_copy_params {
+	uint32_t src_lo;
+	uint32_t src_hi;
+
+	uint32_t dst_lo;
+	uint32_t dst_hi;
+
+	uint32_t src_x        : 16;
+	uint32_t src_y        : 16;
+
+	uint32_t dst_x        : 16;
+	uint32_t dst_y        : 16;
+
+	uint32_t rect_x       : 16;
+	uint32_t rect_y       : 16;
+
+	uint32_t src_pitch    : 16;
+	uint32_t dst_pitch    : 16;
+
+	uint32_t src_slice_pitch;
+	uint32_t dst_slice_pitch;
+
+	uint32_t tmz              : 1;
+	uint32_t element_size     : 3;
+	uint32_t src_cache_policy : 3;
+	uint32_t dst_cache_policy : 3;
+	uint32_t padding          : 22;
+};
+
+bool dmub_lsdma_send_linear_sub_window_copy_command(
+	struct dc_dmub_srv *dc_dmub_srv,
+	struct lsdma_linear_sub_window_copy_params copy_data
+);
+bool dmub_lsdma_send_pio_copy_command(
+	struct dc_dmub_srv *dc_dmub_srv,
+	uint64_t src_addr,
+	uint64_t dst_addr,
+	uint32_t byte_count,
+	uint32_t overlap_disable);
+bool dmub_lsdma_send_pio_constfill_command(
+	struct dc_dmub_srv *dc_dmub_srv,
+	uint64_t dst_addr,
+	uint32_t byte_count,
+	uint32_t data);
+
+struct lsdma_send_tiled_to_tiled_copy_command_params {
+	uint64_t src_addr;
+	uint64_t dst_addr;
+
+	uint32_t src_x            : 16;
+	uint32_t src_y            : 16;
+
+	uint32_t dst_x            : 16;
+	uint32_t dst_y            : 16;
+
+	uint32_t src_width        : 16;
+	uint32_t dst_width        : 16;
+
+	uint32_t rect_x           : 16;
+	uint32_t rect_y           : 16;
+
+	uint32_t src_height       : 16;
+	uint32_t dst_height       : 16;
+
+	uint32_t data_format      : 6;
+	uint32_t swizzle_mode     : 5;
+	uint32_t element_size     : 3;
+	uint32_t dcc              : 1;
+	uint32_t tmz              : 1;
+	uint32_t read_compress    : 2;
+	uint32_t write_compress   : 2;
+	uint32_t max_com          : 2;
+	uint32_t max_uncom        : 1;
+	uint32_t padding          : 9;
+};
+
+bool dmub_lsdma_send_tiled_to_tiled_copy_command(
+	struct dc_dmub_srv *dc_dmub_srv,
+	struct lsdma_send_tiled_to_tiled_copy_command_params params);
+bool dmub_lsdma_send_poll_reg_write_command(struct dc_dmub_srv *dc_dmub_srv, uint32_t reg_addr, uint32_t reg_data);
+
+/**
+ * struct ips_residency_info - struct containing info from dmub_ips_residency_stats
+ *
+ * @ips_mode: The mode of IPS that the follow stats appertain to
+ * @residency_percent: The percentage of time spent in given IPS mode in millipercent
+ * @entry_counter: The number of entries made in to this IPS state
+ * @total_active_time_us: uint32_t array of length 2 representing time in the given IPS mode
+ *                        in microseconds. Index 0 is lower 32 bits, index 1 is upper 32 bits.
+ * @total_inactive_time_us: uint32_t array of length 2 representing time outside the given IPS mode
+ *                          in microseconds. Index 0 is lower 32 bits, index 1 is upper 32 bits.
+ * @histogram: Histogram of given IPS state durations - bucket definitions in dmub_ips.c
+ */
+struct ips_residency_info {
+	enum ips_residency_mode ips_mode;
+	unsigned int residency_percent;
+	unsigned int entry_counter;
+	unsigned int total_active_time_us[2];
+	unsigned int total_inactive_time_us[2];
+	unsigned int histogram[16];
+};
+
+bool dc_dmub_srv_ips_residency_cntl(const struct dc_context *ctx, uint8_t panel_inst, bool start_measurement);
+
+bool dc_dmub_srv_ips_query_residency_info(const struct dc_context *ctx, uint8_t panel_inst,
+					  struct dmub_ips_residency_info *driver_info,
+					  enum ips_residency_mode ips_mode);
+
+/**
+ * dc_dmub_srv_cursor_offload_init() - Enables or disables cursor offloading for a stream.
+ *
+ * @dc: pointer to DC object
+ */
+void dc_dmub_srv_cursor_offload_init(struct dc *dc);
+
+/**
+ * dc_dmub_srv_control_cursor_offload() - Enables or disables cursor offloading for a stream.
+ *
+ * @dc: pointer to DC object
+ * @context: the DC context to reference for pipe allocations
+ * @stream: the stream to control
+ * @enable: true to enable cursor offload, false to disable
+ */
+void dc_dmub_srv_control_cursor_offload(struct dc *dc, struct dc_state *context,
+					const struct dc_stream_state *stream, bool enable);
+
+/**
+ * dc_dmub_srv_program_cursor_now() - Requests immediate cursor programming for a given pipe.
+ *
+ * @dc: pointer to DC object
+ * @pipe: top-most pipe for a stream.
+ */
+void dc_dmub_srv_program_cursor_now(struct dc *dc, const struct pipe_ctx *pipe);
+
+/**
+ * dc_dmub_srv_is_cursor_offload_enabled() - Checks if cursor offload is supported.
+ *
+ * @dc: pointer to DC object
+ *
+ * Return: true if cursor offload is supported, false otherwise
+ */
+bool dc_dmub_srv_is_cursor_offload_enabled(const struct dc *dc);
+
+/**
+ * dc_dmub_srv_release_hw() - Notifies DMUB service that HW access is no longer required.
+ *
+ * @dc - pointer to DC object
+ */
+void dc_dmub_srv_release_hw(const struct dc *dc);
+
+/**
+ * dc_dmub_srv_log_preos_dmcub_info() - Logs preos dmcub fw info.
+ *
+ * @dc - pointer to DC object
+ */
+void dc_dmub_srv_log_preos_dmcub_info(struct dc_dmub_srv *dc_dmub_srv);
 #endif /* _DMUB_DC_SRV_H_ */
diff --git a/drivers/gpu/drm/amd/display/dc/dc_dp_types.h b/drivers/gpu/drm/amd/display/dc/dc_dp_types.h
index e68e9a86a4d9..79e1696def63 100644
--- a/drivers/gpu/drm/amd/display/dc/dc_dp_types.h
+++ b/drivers/gpu/drm/amd/display/dc/dc_dp_types.h
@@ -27,6 +27,7 @@
 #define DC_DP_TYPES_H
 
 #include "os_types.h"
+#include "dc_ddc_types.h"
 
 enum dc_lane_count {
 	LANE_COUNT_UNKNOWN = 0,
@@ -46,24 +47,21 @@ enum dc_lane_count {
  */
 enum dc_link_rate {
 	LINK_RATE_UNKNOWN = 0,
-	LINK_RATE_LOW = 0x06,		// Rate_1 (RBR)	- 1.62 Gbps/Lane
-	LINK_RATE_RATE_2 = 0x08,	// Rate_2		- 2.16 Gbps/Lane
-	LINK_RATE_RATE_3 = 0x09,	// Rate_3		- 2.43 Gbps/Lane
-	LINK_RATE_HIGH = 0x0A,		// Rate_4 (HBR)	- 2.70 Gbps/Lane
-	LINK_RATE_RBR2 = 0x0C,		// Rate_5 (RBR2)- 3.24 Gbps/Lane
-	LINK_RATE_RATE_6 = 0x10,	// Rate_6		- 4.32 Gbps/Lane
-	LINK_RATE_HIGH2 = 0x14,		// Rate_7 (HBR2)- 5.40 Gbps/Lane
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	LINK_RATE_HIGH3 = 0x1E,		// Rate_8 (HBR3)- 8.10 Gbps/Lane
+	LINK_RATE_LOW = 0x06,		// Rate_1 (RBR)  - 1.62 Gbps/Lane
+	LINK_RATE_RATE_2 = 0x08,	// Rate_2        - 2.16 Gbps/Lane
+	LINK_RATE_RATE_3 = 0x09,	// Rate_3        - 2.43 Gbps/Lane
+	LINK_RATE_HIGH = 0x0A,		// Rate_4 (HBR)  - 2.70 Gbps/Lane
+	LINK_RATE_RBR2 = 0x0C,		// Rate_5 (RBR2) - 3.24 Gbps/Lane
+	LINK_RATE_RATE_6 = 0x10,	// Rate_6        - 4.32 Gbps/Lane
+	LINK_RATE_HIGH2 = 0x14,		// Rate_7 (HBR2) - 5.40 Gbps/Lane
+	LINK_RATE_RATE_8 = 0x19,	// Rate_8        - 6.75 Gbps/Lane
+	LINK_RATE_HIGH3 = 0x1E,		// Rate_9 (HBR3) - 8.10 Gbps/Lane
 	/* Starting from DP2.0 link rate enum directly represents actual
 	 * link rate value in unit of 10 mbps
 	 */
 	LINK_RATE_UHBR10 = 1000,	// UHBR10 - 10.0 Gbps/Lane
 	LINK_RATE_UHBR13_5 = 1350,	// UHBR13.5 - 13.5 Gbps/Lane
-	LINK_RATE_UHBR20 = 2000,	// UHBR10 - 20.0 Gbps/Lane
-#else
-	LINK_RATE_HIGH3 = 0x1E		// Rate_8 (HBR3)- 8.10 Gbps/Lane
-#endif
+	LINK_RATE_UHBR20 = 2000,	// UHBR20 - 20.0 Gbps/Lane
 };
 
 enum dc_link_spread {
@@ -100,7 +98,6 @@ enum dc_post_cursor2 {
 	POST_CURSOR2_MAX_LEVEL = POST_CURSOR2_LEVEL3,
 };
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 enum dc_dp_ffe_preset_level {
 	DP_FFE_PRESET_LEVEL0 = 0,
 	DP_FFE_PRESET_LEVEL1,
@@ -120,7 +117,6 @@ enum dc_dp_ffe_preset_level {
 	DP_FFE_PRESET_LEVEL15,
 	DP_FFE_PRESET_MAX_LEVEL = DP_FFE_PRESET_LEVEL15,
 };
-#endif
 
 enum dc_dp_training_pattern {
 	DP_TRAINING_PATTERN_SEQUENCE_1 = 0,
@@ -128,19 +124,31 @@ enum dc_dp_training_pattern {
 	DP_TRAINING_PATTERN_SEQUENCE_3,
 	DP_TRAINING_PATTERN_SEQUENCE_4,
 	DP_TRAINING_PATTERN_VIDEOIDLE,
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	DP_128b_132b_TPS1,
 	DP_128b_132b_TPS2,
 	DP_128b_132b_TPS2_CDS,
-#endif
 };
 
 enum dp_link_encoding {
 	DP_UNKNOWN_ENCODING = 0,
 	DP_8b_10b_ENCODING = 1,
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	DP_128b_132b_ENCODING = 2,
-#endif
+};
+
+enum dp_test_link_rate {
+	DP_TEST_LINK_RATE_RBR		= 0x06,
+	DP_TEST_LINK_RATE_RATE_2    = 0x08,	// Rate_2        - 2.16 Gbps/Lane
+	DP_TEST_LINK_RATE_RATE_3    = 0x09,	// Rate_3        - 2.43 Gbps/Lane
+	DP_TEST_LINK_RATE_HBR		= 0x0A,
+	DP_TEST_LINK_RATE_RBR2      = 0x0C,	// Rate_5 (RBR2) - 3.24 Gbps/Lane
+	DP_TEST_LINK_RATE_RATE_6    = 0x10,	// Rate_6        - 4.32 Gbps/Lane
+	DP_TEST_LINK_RATE_HBR2		= 0x14,
+	DP_TEST_LINK_RATE_RATE_8    = 0x19,	// Rate_8        - 6.75 Gbps/Lane
+	DP_TEST_LINK_RATE_HBR3		= 0x1E,
+	DP_TEST_LINK_RATE_UHBR10	= 0x01,
+	DP_TEST_LINK_RATE_UHBR20	= 0x02,
+	DP_TEST_LINK_RATE_UHBR13_5_LEGACY	= 0x03, /* For backward compatibility*/
+	DP_TEST_LINK_RATE_UHBR13_5	= 0x04,
 };
 
 struct dc_link_settings {
@@ -149,10 +157,18 @@ struct dc_link_settings {
 	enum dc_link_spread link_spread;
 	bool use_link_rate_set;
 	uint8_t link_rate_set;
-	bool dpcd_source_device_specific_field_support;
 };
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
+struct dc_tunnel_settings {
+	bool should_enable_dp_tunneling;
+	bool should_use_dp_bw_allocation;
+	uint8_t cm_id;
+	uint8_t group_id;
+	uint32_t bw_granularity;
+	uint32_t estimated_bw;
+	uint32_t allocated_bw;
+};
+
 union dc_dp_ffe_preset {
 	struct {
 		uint8_t level		: 4;
@@ -163,24 +179,19 @@ union dc_dp_ffe_preset {
 	} settings;
 	uint8_t raw;
 };
-#endif
 
 struct dc_lane_settings {
 	enum dc_voltage_swing VOLTAGE_SWING;
 	enum dc_pre_emphasis PRE_EMPHASIS;
 	enum dc_post_cursor2 POST_CURSOR2;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	union dc_dp_ffe_preset FFE_PRESET;
-#endif
 };
 
 struct dc_link_training_overrides {
 	enum dc_voltage_swing *voltage_swing;
 	enum dc_pre_emphasis *pre_emphasis;
 	enum dc_post_cursor2 *post_cursor2;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	union dc_dp_ffe_preset *ffe_preset;
-#endif
 
 	uint16_t *cr_pattern_time;
 	uint16_t *eq_pattern_time;
@@ -194,7 +205,6 @@ struct dc_link_training_overrides {
 	bool *fec_enable;
 };
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 union payload_table_update_status {
 	struct {
 		uint8_t  VC_PAYLOAD_TABLE_UPDATED:1;
@@ -202,7 +212,6 @@ union payload_table_update_status {
 	} bits;
 	uint8_t  raw;
 };
-#endif
 
 union dpcd_rev {
 	struct {
@@ -291,32 +300,39 @@ union lane_align_status_updated {
 	struct {
 		uint8_t INTERLANE_ALIGN_DONE:1;
 		uint8_t POST_LT_ADJ_REQ_IN_PROGRESS:1;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 		uint8_t EQ_INTERLANE_ALIGN_DONE_128b_132b:1;
 		uint8_t CDS_INTERLANE_ALIGN_DONE_128b_132b:1;
 		uint8_t LT_FAILED_128b_132b:1;
 		uint8_t RESERVED:1;
-#else
-		uint8_t RESERVED:4;
-#endif
 		uint8_t DOWNSTREAM_PORT_STATUS_CHANGED:1;
 		uint8_t LINK_STATUS_UPDATED:1;
 	} bits;
 	uint8_t raw;
 };
 
+union link_service_irq_vector_esi0 {
+	struct {
+		uint8_t DP_LINK_RX_CAP_CHANGED:1;
+		uint8_t DP_LINK_STATUS_CHANGED:1;
+		uint8_t DP_LINK_STREAM_STATUS_CHANGED:1;
+		uint8_t DP_LINK_HDMI_LINK_STATUS_CHANGED:1;
+		uint8_t DP_LINK_CONNECTED_OFF_ENTRY_REQUESTED:1;
+		uint8_t DP_LINK_TUNNELING_IRQ:1;
+		uint8_t reserved:2;
+	} bits;
+	uint8_t raw;
+};
+
 union lane_adjust {
 	struct {
 		uint8_t VOLTAGE_SWING_LANE:2;
 		uint8_t PRE_EMPHASIS_LANE:2;
 		uint8_t RESERVED:4;
 	} bits;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	struct {
 		uint8_t PRESET_VALUE	:4;
 		uint8_t RESERVED	:4;
 	} tx_ffe;
-#endif
 	uint8_t raw;
 };
 
@@ -346,12 +362,10 @@ union dpcd_training_lane {
 		uint8_t MAX_PRE_EMPHASIS_REACHED:1;
 		uint8_t RESERVED:2;
 	} bits;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	struct {
 		uint8_t PRESET_VALUE	:4;
 		uint8_t RESERVED	:4;
 	} tx_ffe;
-#endif
 	uint8_t raw;
 };
 
@@ -378,7 +392,10 @@ enum dpcd_downstream_port_detailed_type {
 union dwnstream_port_caps_byte2 {
 	struct {
 		uint8_t MAX_BITS_PER_COLOR_COMPONENT:2;
-		uint8_t RESERVED:6;
+		uint8_t MAX_ENCODED_LINK_BW_SUPPORT:3;
+		uint8_t SOURCE_CONTROL_MODE_SUPPORT:1;
+		uint8_t CONCURRENT_LINK_BRING_UP_SEQ_SUPPORT:1;
+		uint8_t RESERVED:1;
 	} bits;
 	uint8_t raw;
 };
@@ -416,6 +433,41 @@ union dwnstream_port_caps_byte3_hdmi {
 	uint8_t raw;
 };
 
+union hdmi_encoded_link_bw {
+	struct {
+		uint8_t FRL_MODE:1; // Bit 0
+		uint8_t BW_9Gbps:1;
+		uint8_t BW_18Gbps:1;
+		uint8_t BW_24Gbps:1;
+		uint8_t BW_32Gbps:1;
+		uint8_t BW_40Gbps:1;
+		uint8_t BW_48Gbps:1;
+		uint8_t FRL_LINK_TRAINING_FINISHED:1; // Bit 7
+	} bits;
+	uint8_t raw;
+};
+
+union hdmi_tx_link_status {
+	struct {
+		uint8_t HDMI_TX_LINK_ACTIVE_STATUS:1;
+		uint8_t HDMI_TX_READY_STATUS:1;
+		uint8_t RESERVED:6;
+	} bits;
+	uint8_t raw;
+};
+
+union autonomous_mode_and_frl_link_status {
+	struct {
+		uint8_t FRL_LT_IN_PROGRESS_STATUS:1;
+		uint8_t FRL_LT_LINK_CONFIG_IN_PROGRESS:3;
+		uint8_t RESERVED:1;
+		uint8_t FALLBACK_POLICY:1;
+		uint8_t FALLBACK_POLICY_VALID:1;
+		uint8_t REGULATED_AUTONOMOUS_MODE_SUPPORTED:1;
+	} bits;
+	uint8_t raw;
+};
+
 /*4-byte structure for detailed capabilities of a down-stream port
 (DP-to-TMDS converter).*/
 union dwnstream_portxcaps {
@@ -454,8 +506,10 @@ union sink_status {
 	uint8_t raw;
 };
 
-/*6-byte structure corresponding to 6 registers (200h-205h)
-read during handling of HPD-IRQ*/
+/* 7-byte structure corresponding to 6 registers (200h-205h)
+ * and LINK_SERVICE_IRQ_ESI0 (2005h) for tunneling IRQ
+ * read during handling of HPD-IRQ
+ */
 union hpd_irq_data {
 	struct {
 		union sink_count sink_cnt;/* 200h */
@@ -463,9 +517,10 @@ union hpd_irq_data {
 		union lane_status lane01_status;/* 202h */
 		union lane_status lane23_status;/* 203h */
 		union lane_align_status_updated lane_status_updated;/* 204h */
-		union sink_status sink_status;
+		union sink_status sink_status;/* 205h */
+		union link_service_irq_vector_esi0 link_service_irq_esi0;/* 2005h */
 	} bytes;
-	uint8_t raw[6];
+	uint8_t raw[7];
 };
 
 union down_stream_port_count {
@@ -494,7 +549,11 @@ union down_spread_ctrl {
 	1 = Main link signal is downspread <= 0.5%
 	with frequency in the range of 30kHz ~ 33kHz*/
 		uint8_t SPREAD_AMP:1;
-		uint8_t RESERVED2:2;/*Bit 6:5 = RESERVED. Read all 0s*/
+		uint8_t RESERVED2:1;/*Bit 5 = RESERVED. Read all 0s*/
+	/* Bit 6 = FIXED_VTOTAL_AS_SDP_EN_IN_PR_ACTIVE.
+	0 = FIXED_VTOTAL_AS_SDP_EN_IN_PR_ACTIVE is not enabled by the Source device (default)
+	1 = FIXED_VTOTAL_AS_SDP_EN_IN_PR_ACTIVE is enabled by Source device */
+		uint8_t FIXED_VTOTAL_AS_SDP_EN_IN_PR_ACTIVE:1;
 	/*Bit 7 = MSA_TIMING_PAR_IGNORE_EN
 	0 = Source device will send valid data for the MSA Timing Params
 	1 = Source device may send invalid data for these MSA Timing Params*/
@@ -552,6 +611,12 @@ struct dpcd_amd_device_id {
 	uint8_t dal_version_byte2;
 };
 
+struct target_luminance_value {
+	uint8_t byte0;
+	uint8_t byte1;
+	uint8_t byte2;
+};
+
 struct dpcd_source_backlight_set {
 	struct  {
 		uint8_t byte0;
@@ -615,7 +680,7 @@ union test_request {
 	uint8_t LINK_TEST_PATTRN             :1;
 	uint8_t EDID_READ                    :1;
 	uint8_t PHY_TEST_PATTERN             :1;
-	uint8_t RESERVED                     :1;
+	uint8_t PHY_TEST_CHANNEL_CODING_TYPE :2;
 	uint8_t AUDIO_TEST_PATTERN           :1;
 	uint8_t TEST_AUDIO_DISABLED_VIDEO    :1;
 	} bits;
@@ -634,18 +699,9 @@ union test_response {
 
 union phy_test_pattern {
 	struct {
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 		/* This field is 7 bits for DP2.0 */
 		uint8_t PATTERN     :7;
 		uint8_t RESERVED    :1;
-#else
-		/* DpcdPhyTestPatterns. This field is 2 bits for DP1.1
-		 * and 3 bits for DP1.2.
-		 */
-		uint8_t PATTERN     :3;
-		/* BY speci, bit7:2 is 0 for DP1.1. */
-		uint8_t RESERVED    :5;
-#endif
 	} bits;
 	uint8_t raw;
 };
@@ -723,14 +779,10 @@ union dpcd_fec_capability {
 		uint8_t UNCORRECTED_BLOCK_ERROR_COUNT_CAPABLE:1;
 		uint8_t CORRECTED_BLOCK_ERROR_COUNT_CAPABLE:1;
 		uint8_t BIT_ERROR_COUNT_CAPABLE:1;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 		uint8_t PARITY_BLOCK_ERROR_COUNT_CAPABLE:1;
 		uint8_t ARITY_BIT_ERROR_COUNT_CAPABLE:1;
 		uint8_t FEC_RUNNING_INDICATOR_SUPPORTED:1;
 		uint8_t FEC_ERROR_REPORTING_POLICY_SUPPORTED:1;
-#else
-		uint8_t RESERVED:4;
-#endif
 	} bits;
 	uint8_t raw;
 };
@@ -852,6 +904,30 @@ struct psr_caps {
 	unsigned char psr_version;
 	unsigned int psr_rfb_setup_time;
 	bool psr_exit_link_training_required;
+	unsigned char edp_revision;
+	unsigned char support_ver;
+	bool su_granularity_required;
+	bool y_coordinate_required;
+	uint8_t su_y_granularity;
+	bool alpm_cap;
+	bool standby_support;
+	uint8_t rate_control_caps;
+	unsigned int psr_power_opt_flag;
+};
+
+union dpcd_dprx_feature_enumeration_list_cont_1 {
+	struct {
+		uint8_t ADAPTIVE_SYNC_SDP_SUPPORT:1;
+		uint8_t AS_SDP_FIRST_HALF_LINE_OR_3840_PIXEL_CYCLE_WINDOW_NOT_SUPPORTED: 1;
+		uint8_t RESERVED0: 2;
+		uint8_t VSC_EXT_SDP_VER1_SUPPORT: 1;
+		uint8_t RESERVED1: 3;
+	} bits;
+	uint8_t raw;
+};
+
+struct adaptive_sync_caps {
+	union dpcd_dprx_feature_enumeration_list_cont_1 dp_adap_sync_caps;
 };
 
 /* Length of router topology ID read from DPCD in bytes. */
@@ -877,86 +953,34 @@ union dpia_info {
 	uint8_t raw;
 };
 
+/* DPCD[0xE0020] USB4_DRIVER_BW_CAPABILITY register. */
+union usb4_driver_bw_cap {
+	struct {
+		uint8_t rsvd :7;
+		uint8_t driver_bw_alloc_support :1;
+	} bits;
+	uint8_t raw;
+};
+
+/* DPCD[0xE0021] DP_IN_ADAPTER_TUNNEL_INFORMATION register. */
+union dpia_tunnel_info {
+	struct {
+		uint8_t group_id :3;
+		uint8_t rsvd :5;
+	} bits;
+	uint8_t raw;
+};
+
 /* DP Tunneling over USB4 */
 struct dpcd_usb4_dp_tunneling_info {
 	union dp_tun_cap_support dp_tun_cap;
 	union dpia_info dpia_info;
+	union usb4_driver_bw_cap driver_bw_cap;
+	union dpia_tunnel_info dpia_tunnel_info;
 	uint8_t usb4_driver_id;
 	uint8_t usb4_topology_id[DPCD_USB4_TOPOLOGY_ID_LEN];
 };
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-#ifndef DP_MAIN_LINK_CHANNEL_CODING_CAP
-#define DP_MAIN_LINK_CHANNEL_CODING_CAP			0x006
-#endif
-#ifndef DP_SINK_VIDEO_FALLBACK_FORMATS
-#define DP_SINK_VIDEO_FALLBACK_FORMATS			0x020
-#endif
-#ifndef DP_FEC_CAPABILITY_1
-#define DP_FEC_CAPABILITY_1				0x091
-#endif
-#ifndef DP_DFP_CAPABILITY_EXTENSION_SUPPORT
-#define DP_DFP_CAPABILITY_EXTENSION_SUPPORT		0x0A3
-#endif
-#ifndef DP_LINK_SQUARE_PATTERN
-#define DP_LINK_SQUARE_PATTERN				0x10F
-#endif
-#ifndef DP_DSC_CONFIGURATION
-#define DP_DSC_CONFIGURATION				0x161
-#endif
-#ifndef DP_PHY_SQUARE_PATTERN
-#define DP_PHY_SQUARE_PATTERN				0x249
-#endif
-#ifndef DP_128b_132b_SUPPORTED_LINK_RATES
-#define DP_128b_132b_SUPPORTED_LINK_RATES		0x2215
-#endif
-#ifndef DP_128b_132b_TRAINING_AUX_RD_INTERVAL
-#define DP_128b_132b_TRAINING_AUX_RD_INTERVAL		0x2216
-#endif
-#ifndef DP_TEST_264BIT_CUSTOM_PATTERN_7_0
-#define DP_TEST_264BIT_CUSTOM_PATTERN_7_0		0X2230
-#endif
-#ifndef DP_TEST_264BIT_CUSTOM_PATTERN_263_256
-#define DP_TEST_264BIT_CUSTOM_PATTERN_263_256		0X2250
-#endif
-#ifndef DP_DSC_SUPPORT_AND_DECODER_COUNT
-#define DP_DSC_SUPPORT_AND_DECODER_COUNT		0x2260
-#endif
-#ifndef DP_DSC_MAX_SLICE_COUNT_AND_AGGREGATION_0
-#define DP_DSC_MAX_SLICE_COUNT_AND_AGGREGATION_0	0x2270
-#endif
-#ifndef DP_DSC_DECODER_0_MAXIMUM_SLICE_COUNT_MASK
-#define DP_DSC_DECODER_0_MAXIMUM_SLICE_COUNT_MASK	(1 << 0)
-#endif
-#ifndef DP_DSC_DECODER_0_AGGREGATION_SUPPORT_MASK
-#define DP_DSC_DECODER_0_AGGREGATION_SUPPORT_MASK	(0b111 << 1)
-#endif
-#ifndef DP_DSC_DECODER_0_AGGREGATION_SUPPORT_SHIFT
-#define DP_DSC_DECODER_0_AGGREGATION_SUPPORT_SHIFT	1
-#endif
-#ifndef DP_DSC_DECODER_COUNT_MASK
-#define DP_DSC_DECODER_COUNT_MASK			(0b111 << 5)
-#endif
-#ifndef DP_DSC_DECODER_COUNT_SHIFT
-#define DP_DSC_DECODER_COUNT_SHIFT			5
-#endif
-#ifndef DP_MAIN_LINK_CHANNEL_CODING_SET
-#define DP_MAIN_LINK_CHANNEL_CODING_SET			0x108
-#endif
-#ifndef DP_MAIN_LINK_CHANNEL_CODING_PHY_REPEATER
-#define DP_MAIN_LINK_CHANNEL_CODING_PHY_REPEATER	0xF0006
-#endif
-#ifndef DP_PHY_REPEATER_128b_132b_RATES
-#define DP_PHY_REPEATER_128b_132b_RATES			0xF0007
-#endif
-#ifndef DP_128b_132b_TRAINING_AUX_RD_INTERVAL_PHY_REPEATER1
-#define DP_128b_132b_TRAINING_AUX_RD_INTERVAL_PHY_REPEATER1	0xF0022
-#endif
-#ifndef DP_INTRA_HOP_AUX_REPLY_INDICATION
-#define DP_INTRA_HOP_AUX_REPLY_INDICATION		(1 << 3)
-#endif
-/* TODO - Use DRM header to replace above once available */
-
 union dp_main_line_channel_coding_cap {
 	struct {
 		uint8_t DP_8b_10b_SUPPORTED	:1;
@@ -987,13 +1011,22 @@ union dp_128b_132b_supported_link_rates {
 union dp_128b_132b_supported_lttpr_link_rates {
 	struct {
 		uint8_t UHBR10	:1;
-		uint8_t UHBR13_5:1;
 		uint8_t UHBR20	:1;
+		uint8_t UHBR13_5:1;
 		uint8_t RESERVED:5;
 	} bits;
 	uint8_t raw;
 };
 
+union dp_alpm_lttpr_cap {
+	struct {
+		uint8_t AUX_LESS_ALPM_SUPPORTED	:1;
+		uint8_t ASSR_SUPPORTED			:1;
+		uint8_t RESERVED			:6;
+	} bits;
+	uint8_t raw;
+};
+
 union dp_sink_video_fallback_formats {
 	struct {
 		uint8_t dp_1024x768_60Hz_24bpp_support	:1;
@@ -1004,6 +1037,29 @@ union dp_sink_video_fallback_formats {
 	uint8_t raw;
 };
 
+union dp_receive_port0_cap {
+	struct {
+		uint8_t RESERVED					:1;
+		uint8_t LOCAL_EDID_PRESENT			:1;
+		uint8_t ASSOCIATED_TO_PRECEDING_PORT:1;
+		uint8_t HBLANK_EXPANSION_CAPABLE	:1;
+		uint8_t BUFFER_SIZE_UNIT			:1;
+		uint8_t BUFFER_SIZE_PER_PORT		:1;
+		uint8_t HBLANK_REDUCTION_CAPABLE	:1;
+		uint8_t RESERVED2:1;
+		uint8_t BUFFER_SIZE:8;
+	} bits;
+	uint8_t raw[2];
+};
+
+union dpcd_max_uncompressed_pixel_rate_cap {
+	struct {
+		uint16_t max_uncompressed_pixel_rate_cap	:15;
+		uint16_t valid			:1;
+	} bits;
+	uint8_t raw[2];
+};
+
 union dp_fec_capability1 {
 	struct {
 		uint8_t AGGREGATED_ERROR_COUNTERS_CAPABLE	:1;
@@ -1012,6 +1068,16 @@ union dp_fec_capability1 {
 	uint8_t raw;
 };
 
+union dp_cable_id {
+	struct {
+		uint8_t UHBR10_20_CAPABILITY	:2;
+		uint8_t UHBR13_5_CAPABILITY	:1;
+		uint8_t CABLE_TYPE		:3;
+		uint8_t RESERVED		:2;
+	} bits;
+	uint8_t raw;
+};
+
 struct dp_color_depth_caps {
 	uint8_t support_6bpc	:1;
 	uint8_t support_8bpc	:1;
@@ -1051,6 +1117,451 @@ union dp_128b_132b_training_aux_rd_interval {
 	} bits;
 	uint8_t raw;
 };
+
+union edp_alpm_caps {
+	struct {
+		uint8_t AUX_WAKE_ALPM_CAP                               :1;
+		uint8_t PM_STATE_2A_SUPPORT                             :1;
+		uint8_t AUX_LESS_ALPM_CAP                               :1;
+		uint8_t AUX_LESS_ALPM_ML_PHY_SLEEP_STATUS_SUPPORTED     :1;
+		uint8_t RESERVED                                        :4;
+	} bits;
+	uint8_t raw;
+};
+
+union edp_psr_dpcd_caps {
+	struct {
+		uint8_t LINK_TRAINING_ON_EXIT_NOT_REQUIRED      :1;
+		uint8_t PSR_SETUP_TIME  :3;
+		uint8_t Y_COORDINATE_REQUIRED   :1;
+		uint8_t SU_GRANULARITY_REQUIRED :1;
+		uint8_t FRAME_SYNC_IS_NOT_NEEDED_FOR_SU :1;
+		uint8_t RESERVED                :1;
+	} bits;
+	uint8_t raw;
+};
+
+struct edp_psr_info {
+	uint8_t psr_version;
+	union edp_psr_dpcd_caps psr_dpcd_caps;
+	uint8_t psr2_su_y_granularity_cap;
+	uint8_t force_psrsu_cap;
+};
+
+struct replay_info {
+	uint8_t pixel_deviation_per_line;
+	uint8_t max_deviation_line;
+};
+
+struct dprx_states {
+	bool cable_id_written;
+};
+
+union dpcd_panel_replay_capability_supported {
+	struct {
+		unsigned char PANEL_REPLAY_SUPPORT          :1;
+		unsigned char SELECTIVE_UPDATE_SUPPORT      :1;
+		unsigned char EARLY_TRANSPORT_SUPPORT       :1;
+		unsigned char RESERVED				        :5;
+	} bits;
+	unsigned char raw;
+};
+
+enum dpcd_downstream_port_max_bpc {
+	DOWN_STREAM_MAX_8BPC = 0,
+	DOWN_STREAM_MAX_10BPC,
+	DOWN_STREAM_MAX_12BPC,
+	DOWN_STREAM_MAX_16BPC
+};
+
+enum link_training_offset {
+	DPRX                = 0,
+	LTTPR_PHY_REPEATER1 = 1,
+	LTTPR_PHY_REPEATER2 = 2,
+	LTTPR_PHY_REPEATER3 = 3,
+	LTTPR_PHY_REPEATER4 = 4,
+	LTTPR_PHY_REPEATER5 = 5,
+	LTTPR_PHY_REPEATER6 = 6,
+	LTTPR_PHY_REPEATER7 = 7,
+	LTTPR_PHY_REPEATER8 = 8
+};
+
+#define MAX_REPEATER_CNT 8
+
+struct dc_lttpr_caps {
+	union dpcd_rev revision;
+	uint8_t mode;
+	uint8_t max_lane_count;
+	uint8_t max_link_rate;
+	uint8_t phy_repeater_cnt;
+	uint8_t max_ext_timeout;
+	union dp_main_link_channel_coding_lttpr_cap main_link_channel_coding;
+	union dp_128b_132b_supported_lttpr_link_rates supported_128b_132b_rates;
+	union dp_alpm_lttpr_cap alpm;
+	uint8_t aux_rd_interval[MAX_REPEATER_CNT - 1];
+	uint8_t lttpr_ieee_oui[3]; // Always read from closest LTTPR to host
+	uint8_t lttpr_device_id[6]; // Always read from closest LTTPR to host
+};
+
+struct dc_dongle_dfp_cap_ext {
+	bool supported;
+	uint16_t max_pixel_rate_in_mps;
+	uint16_t max_video_h_active_width;
+	uint16_t max_video_v_active_height;
+	struct dp_encoding_format_caps encoding_format_caps;
+	struct dp_color_depth_caps rgb_color_depth_caps;
+	struct dp_color_depth_caps ycbcr444_color_depth_caps;
+	struct dp_color_depth_caps ycbcr422_color_depth_caps;
+	struct dp_color_depth_caps ycbcr420_color_depth_caps;
+};
+
+struct dc_dongle_caps {
+	/* dongle type (DP converter, CV smart dongle) */
+	enum display_dongle_type dongle_type;
+	bool extendedCapValid;
+	/* If dongle_type == DISPLAY_DONGLE_DP_HDMI_CONVERTER,
+	indicates 'Frame Sequential-to-lllFrame Pack' conversion capability.*/
+	bool is_dp_hdmi_s3d_converter;
+	bool is_dp_hdmi_ycbcr422_pass_through;
+	bool is_dp_hdmi_ycbcr420_pass_through;
+	bool is_dp_hdmi_ycbcr422_converter;
+	bool is_dp_hdmi_ycbcr420_converter;
+	uint32_t dp_hdmi_max_bpc;
+	uint32_t dp_hdmi_max_pixel_clk_in_khz;
+	uint32_t dp_hdmi_frl_max_link_bw_in_kbps;
+	uint32_t dp_hdmi_regulated_autonomous_mode_support;
+	struct dc_dongle_dfp_cap_ext dfp_cap_ext;
+};
+
+struct dpcd_caps {
+	union dpcd_rev dpcd_rev;
+	union max_lane_count max_ln_count;
+	union max_down_spread max_down_spread;
+	union dprx_feature dprx_feature;
+
+	/* valid only for eDP v1.4 or higher*/
+	uint8_t edp_supported_link_rates_count;
+	enum dc_link_rate edp_supported_link_rates[8];
+
+	/* dongle type (DP converter, CV smart dongle) */
+	enum display_dongle_type dongle_type;
+	bool is_dongle_type_one;
+	/* branch device or sink device */
+	bool is_branch_dev;
+	/* Dongle's downstream count. */
+	union sink_count sink_count;
+	bool is_mst_capable;
+	/* If dongle_type == DISPLAY_DONGLE_DP_HDMI_CONVERTER,
+	indicates 'Frame Sequential-to-lllFrame Pack' conversion capability.*/
+	struct dc_dongle_caps dongle_caps;
+
+	uint32_t sink_dev_id;
+	int8_t sink_dev_id_str[6];
+	int8_t sink_hw_revision;
+	int8_t sink_fw_revision[2];
+
+	uint32_t branch_dev_id;
+	int8_t branch_dev_name[6];
+	int8_t branch_hw_revision;
+	int8_t branch_fw_revision[2];
+	int8_t branch_vendor_specific_data[4];
+
+	bool allow_invalid_MSA_timing_param;
+	bool panel_mode_edp;
+	bool dpcd_display_control_capable;
+	bool ext_receiver_cap_field_present;
+	bool set_power_state_capable_edp;
+	bool dynamic_backlight_capable_edp;
+	union dpcd_fec_capability fec_cap;
+	struct dpcd_dsc_capabilities dsc_caps;
+	struct dc_lttpr_caps lttpr_caps;
+	struct adaptive_sync_caps adaptive_sync_caps;
+	struct dpcd_usb4_dp_tunneling_info usb4_dp_tun_info;
+	union dpcd_max_uncompressed_pixel_rate_cap max_uncompressed_pixel_rate_cap;
+
+	union dp_128b_132b_supported_link_rates dp_128b_132b_supported_link_rates;
+	union dp_main_line_channel_coding_cap channel_coding_cap;
+	union dp_sink_video_fallback_formats fallback_formats;
+	union dp_fec_capability1 fec_cap1;
+	bool panel_luminance_control;
+	union dp_cable_id cable_id;
+	uint8_t edp_rev;
+	union edp_alpm_caps alpm_caps;
+	struct edp_psr_info psr_info;
+
+	struct replay_info pr_info;
+	union dpcd_panel_replay_capability_supported pr_caps_supported;
+	uint16_t edp_oled_emission_rate;
+	union dp_receive_port0_cap receive_port0_cap;
+	/* Indicates the number of SST links supported by MSO (Multi-Stream Output) */
+	uint8_t mso_cap_sst_links_supported;
+	uint8_t dp_edp_general_cap_2;
+};
+
+union dpcd_sink_ext_caps {
+	struct {
+		/* 0 - Sink supports backlight adjust via PWM during SDR/HDR mode
+		 * 1 - Sink supports backlight adjust via AUX during SDR/HDR mode.
+		 */
+		uint8_t sdr_aux_backlight_control : 1;
+		uint8_t hdr_aux_backlight_control : 1;
+		uint8_t reserved_1 : 2;
+		uint8_t oled : 1;
+		uint8_t reserved_2 : 1;
+		uint8_t miniled : 1;
+		uint8_t emission_output : 1;
+	} bits;
+	uint8_t raw;
+};
+
+enum dc_link_fec_state {
+	dc_link_fec_not_ready,
+	dc_link_fec_ready,
+	dc_link_fec_enabled
+};
+
+union dpcd_psr_configuration {
+	struct {
+		unsigned char ENABLE                    : 1;
+		unsigned char TRANSMITTER_ACTIVE_IN_PSR : 1;
+		unsigned char CRC_VERIFICATION          : 1;
+		unsigned char FRAME_CAPTURE_INDICATION  : 1;
+		/* For eDP 1.4, PSR v2*/
+		unsigned char LINE_CAPTURE_INDICATION   : 1;
+		/* For eDP 1.4, PSR v2*/
+		unsigned char IRQ_HPD_WITH_CRC_ERROR    : 1;
+		unsigned char ENABLE_PSR2               : 1;
+		unsigned char EARLY_TRANSPORT_ENABLE    : 1;
+	} bits;
+	unsigned char raw;
+};
+
+union replay_enable_and_configuration {
+	struct {
+		unsigned char FREESYNC_PANEL_REPLAY_MODE              :1;
+		unsigned char TIMING_DESYNC_ERROR_VERIFICATION        :1;
+		unsigned char STATE_TRANSITION_ERROR_DETECTION        :1;
+		unsigned char RESERVED                                :5;
+	} bits;
+	unsigned char raw;
+};
+
+union dpcd_replay_configuration {
+	struct {
+		unsigned char STATE_TRANSITION_ERROR_STATUS    : 1;
+		unsigned char DESYNC_ERROR_STATUS              : 1;
+		unsigned char SINK_DEVICE_REPLAY_STATUS        : 3;
+		unsigned char SINK_FRAME_LOCKED                : 2;
+		unsigned char RESERVED                         : 1;
+	} bits;
+	unsigned char raw;
+};
+
+union panel_replay_enable_and_configuration_1 {
+	struct {
+		unsigned char PANEL_REPLAY_ENABLE                     :1;
+		unsigned char PANEL_REPLAY_CRC_ENABLE                 :1;
+		unsigned char IRQ_HPD_ASSDP_MISSING                   :1;
+		unsigned char IRQ_HPD_VSCSDP_UNCORRECTABLE_ERROR      :1;
+		unsigned char IRQ_HPD_RFB_ERROR                       :1;
+		unsigned char IRQ_HPD_ACTIVE_FRAME_CRC_ERROR          :1;
+		unsigned char PANEL_REPLAY_SELECTIVE_UPDATE_ENABLE    :1;
+		unsigned char PANEL_REPLAY_EARLY_TRANSPORT_ENABLE     :1;
+	} bits;
+	unsigned char raw;
+};
+
+union panel_replay_enable_and_configuration_2 {
+	struct {
+		unsigned char SINK_REFRESH_RATE_UNLOCK_GRANTED        :1;
+		unsigned char RESERVED                                :1;
+		unsigned char SU_Y_GRANULARITY_EXT_VALUE_ENABLED      :1;
+		unsigned char SU_Y_GRANULARITY_EXT_VALUE              :4;
+		unsigned char SU_REGION_SCAN_LINE_CAPTURE_INDICATION  :1;
+	} bits;
+	unsigned char raw;
+};
+
+union dpcd_alpm_configuration {
+	struct {
+		unsigned char ENABLE                    : 1;
+		unsigned char IRQ_HPD_ENABLE            : 1;
+		unsigned char ALPM_MODE_SEL             : 1;
+		unsigned char ACDS_PERIOD_DURATION      : 1;
+		unsigned char RESERVED                  : 4;
+	} bits;
+	unsigned char raw;
+};
+
+union dpcd_sink_active_vtotal_control_mode {
+	struct {
+		unsigned char ENABLE                    : 1;
+		unsigned char RESERVED                  : 7;
+	} bits;
+	unsigned char raw;
+};
+
+union psr_error_status {
+	struct {
+		unsigned char LINK_CRC_ERROR        :1;
+		unsigned char RFB_STORAGE_ERROR     :1;
+		unsigned char VSC_SDP_ERROR         :1;
+		unsigned char RESERVED              :5;
+	} bits;
+	unsigned char raw;
+};
+
+union psr_sink_psr_status {
+	struct {
+	unsigned char SINK_SELF_REFRESH_STATUS  :3;
+	unsigned char RESERVED                  :5;
+	} bits;
+	unsigned char raw;
+};
+
+struct edp_trace_power_timestamps {
+	uint64_t poweroff;
+	uint64_t poweron;
+};
+
+struct dp_trace_lt_counts {
+	unsigned int total;
+	unsigned int fail;
+};
+
+enum link_training_result {
+	LINK_TRAINING_SUCCESS,
+	LINK_TRAINING_CR_FAIL_LANE0,
+	LINK_TRAINING_CR_FAIL_LANE1,
+	LINK_TRAINING_CR_FAIL_LANE23,
+	/* CR DONE bit is cleared during EQ step */
+	LINK_TRAINING_EQ_FAIL_CR,
+	/* CR DONE bit is cleared but LANE0_CR_DONE is set during EQ step */
+	LINK_TRAINING_EQ_FAIL_CR_PARTIAL,
+	/* other failure during EQ step */
+	LINK_TRAINING_EQ_FAIL_EQ,
+	LINK_TRAINING_LQA_FAIL,
+	/* one of the CR,EQ or symbol lock is dropped */
+	LINK_TRAINING_LINK_LOSS,
+	/* Abort link training (because sink unplugged) */
+	LINK_TRAINING_ABORT,
+	DP_128b_132b_LT_FAILED,
+	DP_128b_132b_MAX_LOOP_COUNT_REACHED,
+	DP_128b_132b_CHANNEL_EQ_DONE_TIMEOUT,
+	DP_128b_132b_CDS_DONE_TIMEOUT,
+};
+
+struct dp_trace_lt {
+	struct dp_trace_lt_counts counts;
+	struct dp_trace_timestamps {
+		unsigned long long start;
+		unsigned long long end;
+	} timestamps;
+	enum link_training_result result;
+	bool is_logged;
+};
+
+struct dp_trace {
+	struct dp_trace_lt detect_lt_trace;
+	struct dp_trace_lt commit_lt_trace;
+	unsigned int link_loss_count;
+	bool is_initialized;
+	struct edp_trace_power_timestamps edp_trace_power_timestamps;
+};
+
+/* TODO - This is a temporary location for any new DPCD definitions.
+ * We should move these to drm_dp header.
+ */
+#ifndef DP_LINK_SQUARE_PATTERN
+#define DP_LINK_SQUARE_PATTERN				0x10F
+#endif
+#ifndef DP_CABLE_ATTRIBUTES_UPDATED_BY_DPRX
+#define DP_CABLE_ATTRIBUTES_UPDATED_BY_DPRX		0x2217
+#endif
+#ifndef DP_CABLE_ATTRIBUTES_UPDATED_BY_DPTX
+#define DP_CABLE_ATTRIBUTES_UPDATED_BY_DPTX		0x110
+#endif
+#ifndef DPCD_MAX_UNCOMPRESSED_PIXEL_RATE_CAP
+#define DPCD_MAX_UNCOMPRESSED_PIXEL_RATE_CAP    0x221c
+#endif
+#ifndef DP_LTTPR_ALPM_CAPABILITIES
+#define DP_LTTPR_ALPM_CAPABILITIES              0xF0009
+#endif
+#ifndef DP_REGULATED_AUTONOMOUS_MODE_SUPPORTED_AND_HDMI_LINK_TRAINING_STATUS
+#define DP_REGULATED_AUTONOMOUS_MODE_SUPPORTED_AND_HDMI_LINK_TRAINING_STATUS	0x303C
+#endif
+#ifndef DP_REPEATER_CONFIGURATION_AND_STATUS_SIZE
+#define DP_REPEATER_CONFIGURATION_AND_STATUS_SIZE	0x50
+#endif
+#ifndef DP_TUNNELING_IRQ
+#define DP_TUNNELING_IRQ				(1 << 5)
+#endif
+#ifndef DP_BRANCH_VENDOR_SPECIFIC_START
+#define DP_BRANCH_VENDOR_SPECIFIC_START     0x50C
+#endif
+#ifndef DP_LTTPR_IEEE_OUI
+#define DP_LTTPR_IEEE_OUI 0xF003D
+#endif
+#ifndef DP_LTTPR_DEVICE_ID
+#define DP_LTTPR_DEVICE_ID 0xF0040
+#endif
+/** USB4 DPCD BW Allocation Registers Chapter 10.7 **/
+#ifndef DP_TUNNELING_CAPABILITIES
+#define DP_TUNNELING_CAPABILITIES			0xE000D /* 1.4a */
+#endif
+#ifndef USB4_DRIVER_ID
+#define USB4_DRIVER_ID					0xE000F /* 1.4a */
+#endif
+#ifndef USB4_DRIVER_BW_CAPABILITY
+#define USB4_DRIVER_BW_CAPABILITY			0xE0020 /* 1.4a */
+#endif
+#ifndef DP_IN_ADAPTER_TUNNEL_INFO
+#define DP_IN_ADAPTER_TUNNEL_INFO			0xE0021 /* 1.4a */
+#endif
+#ifndef DP_BW_GRANULALITY
+#define DP_BW_GRANULALITY				0xE0022 /* 1.4a */
+#endif
+#ifndef ESTIMATED_BW
+#define ESTIMATED_BW					0xE0023 /* 1.4a */
+#endif
+#ifndef ALLOCATED_BW
+#define ALLOCATED_BW					0xE0024 /* 1.4a */
+#endif
+#ifndef DP_TUNNELING_STATUS
+#define DP_TUNNELING_STATUS				0xE0025 /* 1.4a */
+#endif
+#ifndef DP_TUNNELING_MAX_LINK_RATE
+#define DP_TUNNELING_MAX_LINK_RATE			0xE0028 /* 1.4a */
+#endif
+#ifndef DP_TUNNELING_MAX_LANE_COUNT
+#define DP_TUNNELING_MAX_LANE_COUNT			0xE0029 /* 1.4a */
+#endif
+#ifndef DPTX_BW_ALLOCATION_MODE_CONTROL
+#define DPTX_BW_ALLOCATION_MODE_CONTROL			0xE0030 /* 1.4a */
+#endif
+#ifndef REQUESTED_BW
+#define REQUESTED_BW					0xE0031 /* 1.4a */
 #endif
+# ifndef DP_TUNNELING_BW_ALLOC_BITS_MASK
+# define DP_TUNNELING_BW_ALLOC_BITS_MASK		(0x0F << 0)
+# endif
+# ifndef DP_TUNNELING_BW_REQUEST_FAILED
+# define DP_TUNNELING_BW_REQUEST_FAILED			(1 << 0)
+# endif
+# ifndef DP_TUNNELING_BW_REQUEST_SUCCEEDED
+# define DP_TUNNELING_BW_REQUEST_SUCCEEDED		(1 << 1)
+# endif
+# ifndef DP_TUNNELING_ESTIMATED_BW_CHANGED
+# define DP_TUNNELING_ESTIMATED_BW_CHANGED		(1 << 2)
+# endif
+# ifndef DP_TUNNELING_BW_ALLOC_CAP_CHANGED
+# define DP_TUNNELING_BW_ALLOC_CAP_CHANGED		(1 << 3)
+# endif
+# ifndef DPTX_BW_ALLOC_UNMASK_IRQ
+# define DPTX_BW_ALLOC_UNMASK_IRQ			(1 << 6)
+# endif
+# ifndef DPTX_BW_ALLOC_MODE_ENABLE
+# define DPTX_BW_ALLOC_MODE_ENABLE			(1 << 7)
+# endif
 
 #endif /* DC_DP_TYPES_H */
diff --git a/drivers/gpu/drm/amd/display/dc/dc_dsc.h b/drivers/gpu/drm/amd/display/dc/dc_dsc.h
index 684713b2cff7..9d18f1c08079 100644
--- a/drivers/gpu/drm/amd/display/dc/dc_dsc.h
+++ b/drivers/gpu/drm/amd/display/dc/dc_dsc.h
@@ -54,6 +54,14 @@ struct dc_dsc_policy {
 	bool enable_dsc_when_not_needed;
 };
 
+struct dc_dsc_config_options {
+	uint32_t dsc_min_slice_height_override;
+	uint32_t max_target_bpp_limit_override_x16;
+	uint32_t slice_height_granularity;
+	uint32_t dsc_force_odm_hslice_override;
+	bool force_dsc_when_not_needed;
+};
+
 bool dc_dsc_parse_dsc_dpcd(const struct dc *dc,
 		const uint8_t *dpcd_dsc_basic_data,
 		const uint8_t *dpcd_dsc_ext_data,
@@ -66,15 +74,16 @@ bool dc_dsc_compute_bandwidth_range(
 		uint32_t max_bpp_x16,
 		const struct dsc_dec_dpcd_caps *dsc_sink_caps,
 		const struct dc_crtc_timing *timing,
+		const enum dc_link_encoding_format link_encoding,
 		struct dc_dsc_bw_range *range);
 
 bool dc_dsc_compute_config(
 		const struct display_stream_compressor *dsc,
 		const struct dsc_dec_dpcd_caps *dsc_sink_caps,
-		uint32_t dsc_min_slice_height_override,
-		uint32_t max_target_bpp_limit_override,
+		const struct dc_dsc_config_options *options,
 		uint32_t target_bandwidth_kbps,
 		const struct dc_crtc_timing *timing,
+		const enum dc_link_encoding_format link_encoding,
 		struct dc_dsc_config *dsc_cfg);
 
 uint32_t dc_dsc_stream_bandwidth_in_kbps(const struct dc_crtc_timing *timing,
@@ -85,6 +94,11 @@ uint32_t dc_dsc_stream_bandwidth_overhead_in_kbps(
 		const int num_slices_h,
 		const bool is_dp);
 
+void dc_dsc_dump_decoder_caps(const struct display_stream_compressor *dsc,
+			      const struct dsc_dec_dpcd_caps *dsc_sink_caps);
+void dc_dsc_dump_encoder_caps(const struct display_stream_compressor *dsc,
+			      const struct dc_crtc_timing *timing);
+
 /* TODO - Hardware/specs limitation should be owned by dc dsc and returned to DM,
  * and DM can choose to OVERRIDE the limitation on CASE BY CASE basis.
  * Hardware/specs limitation should not be writable by DM.
@@ -92,7 +106,8 @@ uint32_t dc_dsc_stream_bandwidth_overhead_in_kbps(
  */
 void dc_dsc_get_policy_for_timing(const struct dc_crtc_timing *timing,
 		uint32_t max_target_bpp_limit_override_x16,
-		struct dc_dsc_policy *policy);
+		struct dc_dsc_policy *policy,
+		const enum dc_link_encoding_format link_encoding);
 
 void dc_dsc_policy_set_max_target_bpp_limit(uint32_t limit);
 
@@ -100,4 +115,6 @@ void dc_dsc_policy_set_enable_dsc_when_not_needed(bool enable);
 
 void dc_dsc_policy_set_disable_dsc_stream_overhead(bool disable);
 
+void dc_dsc_get_default_config_option(const struct dc *dc, struct dc_dsc_config_options *options);
+
 #endif
diff --git a/drivers/gpu/drm/amd/display/dc/dc_fused_io.c b/drivers/gpu/drm/amd/display/dc/dc_fused_io.c
new file mode 100644
index 000000000000..fee69642fb93
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dc_fused_io.c
@@ -0,0 +1,148 @@
+// SPDX-License-Identifier: MIT
+//
+// Copyright 2025 Advanced Micro Devices, Inc.
+
+#include "dc_fused_io.h"
+
+#include "dm_helpers.h"
+#include "gpio.h"
+
+static bool op_i2c_convert(
+		union dmub_rb_cmd *cmd,
+		const struct mod_hdcp_atomic_op_i2c *op,
+		enum dmub_cmd_fused_request_type type,
+		uint32_t ddc_line,
+		bool over_aux
+)
+{
+	struct dmub_cmd_fused_request *req = &cmd->fused_io.request;
+	struct dmub_cmd_fused_request_location_i2c *loc = &req->u.i2c;
+
+	if (!op || op->size > sizeof(req->buffer))
+		return false;
+
+	req->type = type;
+	loc->is_aux = false;
+	loc->ddc_line = ddc_line;
+	loc->over_aux = over_aux;
+	loc->address = op->address;
+	loc->offset = op->offset;
+	loc->length = op->size;
+	memcpy(req->buffer, op->data, op->size);
+
+	return true;
+}
+
+static bool op_aux_convert(
+		union dmub_rb_cmd *cmd,
+		const struct mod_hdcp_atomic_op_aux *op,
+		enum dmub_cmd_fused_request_type type,
+		uint32_t ddc_line
+)
+{
+	struct dmub_cmd_fused_request *req = &cmd->fused_io.request;
+	struct dmub_cmd_fused_request_location_aux *loc = &req->u.aux;
+
+	if (!op || op->size > sizeof(req->buffer))
+		return false;
+
+	req->type = type;
+	loc->is_aux = true;
+	loc->ddc_line = ddc_line;
+	loc->address = op->address;
+	loc->length = op->size;
+	memcpy(req->buffer, op->data, op->size);
+
+	return true;
+}
+
+static bool atomic_write_poll_read(
+		struct dc_link *link,
+		union dmub_rb_cmd commands[3],
+		uint32_t poll_timeout_us,
+		uint8_t poll_mask_msb
+)
+{
+	const uint8_t count = 3;
+	const uint32_t timeout_per_request_us = 10000;
+	const uint32_t timeout_per_aux_transaction_us = 10000;
+	uint64_t timeout_us = 0;
+
+	commands[1].fused_io.request.poll_mask_msb = poll_mask_msb;
+	commands[1].fused_io.request.timeout_us = poll_timeout_us;
+
+	for (uint8_t i = 0; i < count; i++) {
+		struct dmub_rb_cmd_fused_io *io = &commands[i].fused_io;
+
+		io->header.type = DMUB_CMD__FUSED_IO;
+		io->header.sub_type = DMUB_CMD__FUSED_IO_EXECUTE;
+		io->header.multi_cmd_pending = i != count - 1;
+		io->header.payload_bytes = sizeof(commands[i].fused_io) - sizeof(io->header);
+
+		timeout_us += timeout_per_request_us + io->request.timeout_us;
+		if (!io->request.timeout_us && io->request.u.aux.is_aux)
+			timeout_us += timeout_per_aux_transaction_us * (io->request.u.aux.length / 16);
+	}
+
+	if (!dm_helpers_execute_fused_io(link->ctx, link, commands, count, timeout_us))
+		return false;
+
+	return commands[0].fused_io.request.status == FUSED_REQUEST_STATUS_SUCCESS;
+}
+
+bool dm_atomic_write_poll_read_i2c(
+		struct dc_link *link,
+		const struct mod_hdcp_atomic_op_i2c *write,
+		const struct mod_hdcp_atomic_op_i2c *poll,
+		struct mod_hdcp_atomic_op_i2c *read,
+		uint32_t poll_timeout_us,
+		uint8_t poll_mask_msb
+)
+{
+	if (!link)
+		return false;
+
+	const bool over_aux = false;
+	const uint32_t ddc_line = link->ddc->ddc_pin->pin_data->en;
+
+	union dmub_rb_cmd commands[3] = { 0 };
+	const bool converted = op_i2c_convert(&commands[0], write, FUSED_REQUEST_WRITE, ddc_line, over_aux)
+			&& op_i2c_convert(&commands[1], poll, FUSED_REQUEST_POLL, ddc_line, over_aux)
+			&& op_i2c_convert(&commands[2], read, FUSED_REQUEST_READ, ddc_line, over_aux);
+
+	if (!converted)
+		return false;
+
+	const bool result = atomic_write_poll_read(link, commands, poll_timeout_us, poll_mask_msb);
+
+	memcpy(read->data, commands[0].fused_io.request.buffer, read->size);
+	return result;
+}
+
+bool dm_atomic_write_poll_read_aux(
+		struct dc_link *link,
+		const struct mod_hdcp_atomic_op_aux *write,
+		const struct mod_hdcp_atomic_op_aux *poll,
+		struct mod_hdcp_atomic_op_aux *read,
+		uint32_t poll_timeout_us,
+		uint8_t poll_mask_msb
+)
+{
+	if (!link)
+		return false;
+
+	const uint32_t ddc_line = link->ddc->ddc_pin->pin_data->en;
+	union dmub_rb_cmd commands[3] = { 0 };
+	const bool converted = op_aux_convert(&commands[0], write, FUSED_REQUEST_WRITE, ddc_line)
+			&& op_aux_convert(&commands[1], poll, FUSED_REQUEST_POLL, ddc_line)
+			&& op_aux_convert(&commands[2], read, FUSED_REQUEST_READ, ddc_line);
+
+	if (!converted)
+		return false;
+
+	const bool result = atomic_write_poll_read(link, commands, poll_timeout_us, poll_mask_msb);
+
+	memcpy(read->data, commands[0].fused_io.request.buffer, read->size);
+	return result;
+}
+
diff --git a/drivers/gpu/drm/amd/display/dc/dc_fused_io.h b/drivers/gpu/drm/amd/display/dc/dc_fused_io.h
new file mode 100644
index 000000000000..c74917240985
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dc_fused_io.h
@@ -0,0 +1,31 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2025 Advanced Micro Devices, Inc.
+ */
+
+#ifndef __DC_FUSED_IO_H__
+#define __DC_FUSED_IO_H__
+
+#include "dc.h"
+#include "mod_hdcp.h"
+
+bool dm_atomic_write_poll_read_i2c(
+		struct dc_link *link,
+		const struct mod_hdcp_atomic_op_i2c *write,
+		const struct mod_hdcp_atomic_op_i2c *poll,
+		struct mod_hdcp_atomic_op_i2c *read,
+		uint32_t poll_timeout_us,
+		uint8_t poll_mask_msb
+);
+
+bool dm_atomic_write_poll_read_aux(
+		struct dc_link *link,
+		const struct mod_hdcp_atomic_op_aux *write,
+		const struct mod_hdcp_atomic_op_aux *poll,
+		struct mod_hdcp_atomic_op_aux *read,
+		uint32_t poll_timeout_us,
+		uint8_t poll_mask_msb
+);
+
+#endif  // __DC_FUSED_IO_H__
+
diff --git a/drivers/gpu/drm/amd/display/dc/dc_hdmi_types.h b/drivers/gpu/drm/amd/display/dc/dc_hdmi_types.h
new file mode 100644
index 000000000000..b015e80672ec
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dc_hdmi_types.h
@@ -0,0 +1,133 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef DC_HDMI_TYPES_H
+#define DC_HDMI_TYPES_H
+
+#include "os_types.h"
+
+/* Address range from 0x00 to 0x1F.*/
+#define DP_ADAPTOR_TYPE2_SIZE 0x20
+#define DP_ADAPTOR_TYPE2_REG_ID 0x10
+#define DP_ADAPTOR_TYPE2_REG_MAX_TMDS_CLK 0x1D
+/* Identifies adaptor as Dual-mode adaptor */
+#define DP_ADAPTOR_TYPE2_ID 0xA0
+/* MHz*/
+#define DP_ADAPTOR_TYPE2_MAX_TMDS_CLK 600
+/* MHz*/
+#define DP_ADAPTOR_TYPE2_MIN_TMDS_CLK 25
+/* kHZ*/
+#define DP_ADAPTOR_DVI_MAX_TMDS_CLK 165000
+/* kHZ*/
+#define DP_ADAPTOR_HDMI_SAFE_MAX_TMDS_CLK 165000
+
+struct dp_hdmi_dongle_signature_data {
+	int8_t id[15];/* "DP-HDMI ADAPTOR"*/
+	uint8_t eot;/* end of transmition '\x4' */
+};
+
+/* DP-HDMI dongle slave address for retrieving dongle signature*/
+#define DP_HDMI_DONGLE_ADDRESS 0x40
+#define DP_HDMI_DONGLE_SIGNATURE_EOT 0x04
+
+
+/* SCDC Address defines (HDMI 2.0)*/
+#define HDMI_SCDC_WRITE_UPDATE_0_ARRAY 3
+#define HDMI_SCDC_ADDRESS  0x54
+#define HDMI_SCDC_SINK_VERSION 0x01
+#define HDMI_SCDC_SOURCE_VERSION 0x02
+#define HDMI_SCDC_UPDATE_0 0x10
+#define HDMI_SCDC_TMDS_CONFIG 0x20
+#define HDMI_SCDC_SCRAMBLER_STATUS 0x21
+#define HDMI_SCDC_CONFIG_0 0x30
+#define HDMI_SCDC_CONFIG_1 0x31
+#define HDMI_SCDC_SOURCE_TEST_REQ 0x35
+#define HDMI_SCDC_STATUS_FLAGS 0x40
+#define HDMI_SCDC_ERR_DETECT 0x50
+#define HDMI_SCDC_TEST_CONFIG 0xC0
+
+#define HDMI_SCDC_MANUFACTURER_OUI 0xD0
+#define HDMI_SCDC_DEVICE_ID 0xDB
+
+union hdmi_scdc_update_read_data {
+	uint8_t byte[2];
+	struct {
+		uint8_t STATUS_UPDATE:1;
+		uint8_t CED_UPDATE:1;
+		uint8_t RR_TEST:1;
+		uint8_t RESERVED:5;
+		uint8_t RESERVED2:8;
+	} fields;
+};
+
+union hdmi_scdc_status_flags_data {
+	uint8_t byte;
+	struct {
+		uint8_t CLOCK_DETECTED:1;
+		uint8_t CH0_LOCKED:1;
+		uint8_t CH1_LOCKED:1;
+		uint8_t CH2_LOCKED:1;
+		uint8_t RESERVED:4;
+	} fields;
+};
+
+union hdmi_scdc_ced_data {
+	uint8_t byte[11];
+	struct {
+		uint8_t CH0_8LOW:8;
+		uint8_t CH0_7HIGH:7;
+		uint8_t CH0_VALID:1;
+		uint8_t CH1_8LOW:8;
+		uint8_t CH1_7HIGH:7;
+		uint8_t CH1_VALID:1;
+		uint8_t CH2_8LOW:8;
+		uint8_t CH2_7HIGH:7;
+		uint8_t CH2_VALID:1;
+		uint8_t CHECKSUM:8;
+		uint8_t RESERVED:8;
+		uint8_t RESERVED2:8;
+		uint8_t RESERVED3:8;
+		uint8_t RESERVED4:4;
+	} fields;
+};
+
+union hdmi_scdc_manufacturer_OUI_data {
+	uint8_t byte[3];
+	struct {
+		uint8_t Manufacturer_OUI_1:8;
+		uint8_t Manufacturer_OUI_2:8;
+		uint8_t Manufacturer_OUI_3:8;
+	} fields;
+};
+
+union hdmi_scdc_device_id_data {
+	uint8_t byte;
+	struct {
+		uint8_t Hardware_Minor_Rev:4;
+		uint8_t Hardware_Major_Rev:4;
+	} fields;
+};
+
+#endif /* DC_HDMI_TYPES_H */
diff --git a/drivers/gpu/drm/amd/display/dc/dc_helper.c b/drivers/gpu/drm/amd/display/dc/dc_helper.c
index ab6bc5d79012..5a365bd19933 100644
--- a/drivers/gpu/drm/amd/display/dc/dc_helper.c
+++ b/drivers/gpu/drm/amd/display/dc/dc_helper.c
@@ -36,24 +36,21 @@
 #include "dc_dmub_srv.h"
 #include "reg_helper.h"
 
+#define DC_LOGGER \
+	ctx->logger
+
 static inline void submit_dmub_read_modify_write(
 	struct dc_reg_helper_state *offload,
 	const struct dc_context *ctx)
 {
 	struct dmub_rb_cmd_read_modify_write *cmd_buf = &offload->cmd_data.read_modify_write;
-	bool gather = false;
 
 	offload->should_burst_write =
 			(offload->same_addr_count == (DMUB_READ_MODIFY_WRITE_SEQ__MAX - 1));
 	cmd_buf->header.payload_bytes =
 			sizeof(struct dmub_cmd_read_modify_write_sequence) * offload->reg_seq_count;
 
-	gather = ctx->dmub_srv->reg_helper_offload.gather_in_progress;
-	ctx->dmub_srv->reg_helper_offload.gather_in_progress = false;
-
-	dc_dmub_srv_cmd_queue(ctx->dmub_srv, &offload->cmd_data);
-
-	ctx->dmub_srv->reg_helper_offload.gather_in_progress = gather;
+	dc_wake_and_execute_dmub_cmd(ctx, &offload->cmd_data, DM_DMUB_WAIT_TYPE_NO_WAIT);
 
 	memset(cmd_buf, 0, sizeof(*cmd_buf));
 
@@ -66,17 +63,11 @@ static inline void submit_dmub_burst_write(
 	const struct dc_context *ctx)
 {
 	struct dmub_rb_cmd_burst_write *cmd_buf = &offload->cmd_data.burst_write;
-	bool gather = false;
 
 	cmd_buf->header.payload_bytes =
 			sizeof(uint32_t) * offload->reg_seq_count;
 
-	gather = ctx->dmub_srv->reg_helper_offload.gather_in_progress;
-	ctx->dmub_srv->reg_helper_offload.gather_in_progress = false;
-
-	dc_dmub_srv_cmd_queue(ctx->dmub_srv, &offload->cmd_data);
-
-	ctx->dmub_srv->reg_helper_offload.gather_in_progress = gather;
+	dc_wake_and_execute_dmub_cmd(ctx, &offload->cmd_data, DM_DMUB_WAIT_TYPE_NO_WAIT);
 
 	memset(cmd_buf, 0, sizeof(*cmd_buf));
 
@@ -88,17 +79,11 @@ static inline void submit_dmub_reg_wait(
 		const struct dc_context *ctx)
 {
 	struct dmub_rb_cmd_reg_wait *cmd_buf = &offload->cmd_data.reg_wait;
-	bool gather = false;
 
-	gather = ctx->dmub_srv->reg_helper_offload.gather_in_progress;
-	ctx->dmub_srv->reg_helper_offload.gather_in_progress = false;
-
-	dc_dmub_srv_cmd_queue(ctx->dmub_srv, &offload->cmd_data);
+	dc_wake_and_execute_dmub_cmd(ctx, &offload->cmd_data, DM_DMUB_WAIT_TYPE_NO_WAIT);
 
 	memset(cmd_buf, 0, sizeof(*cmd_buf));
 	offload->reg_seq_count = 0;
-
-	ctx->dmub_srv->reg_helper_offload.gather_in_progress = gather;
 }
 
 struct dc_reg_value_masks {
@@ -106,11 +91,6 @@ struct dc_reg_value_masks {
 	uint32_t mask;
 };
 
-struct dc_reg_sequence {
-	uint32_t addr;
-	struct dc_reg_value_masks value_masks;
-};
-
 static inline void set_reg_field_value_masks(
 	struct dc_reg_value_masks *field_value_mask,
 	uint32_t value,
@@ -151,7 +131,6 @@ static void dmub_flush_buffer_execute(
 		const struct dc_context *ctx)
 {
 	submit_dmub_read_modify_write(offload, ctx);
-	dc_dmub_srv_cmd_execute(ctx->dmub_srv);
 }
 
 static void dmub_flush_burst_write_buffer_execute(
@@ -159,7 +138,6 @@ static void dmub_flush_burst_write_buffer_execute(
 		const struct dc_context *ctx)
 {
 	submit_dmub_burst_write(offload, ctx);
-	dc_dmub_srv_cmd_execute(ctx->dmub_srv);
 }
 
 static bool dmub_reg_value_burst_set_pack(const struct dc_context *ctx, uint32_t addr,
@@ -283,7 +261,6 @@ uint32_t generic_reg_set_ex(const struct dc_context *ctx,
 
 	va_end(ap);
 
-
 	/* mmio write directly */
 	reg_val = (reg_val & ~field_value_mask.mask) | field_value_mask.value;
 
@@ -484,8 +461,7 @@ void generic_reg_wait(const struct dc_context *ctx,
 		field_value = get_reg_field_value_ex(reg_val, mask, shift);
 
 		if (field_value == condition_value) {
-			if (i * delay_between_poll_us > 1000 &&
-					!IS_FPGA_MAXIMUS_DC(ctx->dce_environment))
+			if (i * delay_between_poll_us > 1000)
 				DC_LOG_DC("REG_WAIT taking a while: %dms in %s line:%d\n",
 						delay_between_poll_us * i / 1000,
 						func_name, line);
@@ -497,8 +473,7 @@ void generic_reg_wait(const struct dc_context *ctx,
 			delay_between_poll_us, time_out_num_tries,
 			func_name, line);
 
-	if (!IS_FPGA_MAXIMUS_DC(ctx->dce_environment))
-		BREAK_TO_DEBUGGER();
+	BREAK_TO_DEBUGGER();
 }
 
 void generic_write_indirect_reg(const struct dc_context *ctx,
@@ -588,6 +563,66 @@ uint32_t generic_indirect_reg_update_ex(const struct dc_context *ctx,
 	return reg_val;
 }
 
+
+uint32_t generic_indirect_reg_update_ex_sync(const struct dc_context *ctx,
+		uint32_t index, uint32_t reg_val, int n,
+		uint8_t shift1, uint32_t mask1, uint32_t field_value1,
+		...)
+{
+	uint32_t shift, mask, field_value;
+	int i = 1;
+
+	va_list ap;
+
+	va_start(ap, field_value1);
+
+	reg_val = set_reg_field_value_ex(reg_val, field_value1, mask1, shift1);
+
+	while (i < n) {
+		shift = va_arg(ap, uint32_t);
+		mask = va_arg(ap, uint32_t);
+		field_value = va_arg(ap, uint32_t);
+
+		reg_val = set_reg_field_value_ex(reg_val, field_value, mask, shift);
+		i++;
+	}
+
+	dm_write_index_reg(ctx, CGS_IND_REG__PCIE, index, reg_val);
+	va_end(ap);
+
+	return reg_val;
+}
+
+uint32_t generic_indirect_reg_get_sync(const struct dc_context *ctx,
+		uint32_t index, int n,
+		uint8_t shift1, uint32_t mask1, uint32_t *field_value1,
+		...)
+{
+	uint32_t shift, mask, *field_value;
+	uint32_t value = 0;
+	int i = 1;
+
+	va_list ap;
+
+	va_start(ap, field_value1);
+
+	value = dm_read_index_reg(ctx, CGS_IND_REG__PCIE, index);
+	*field_value1 = get_reg_field_value_ex(value, mask1, shift1);
+
+	while (i < n) {
+		shift = va_arg(ap, uint32_t);
+		mask = va_arg(ap, uint32_t);
+		field_value = va_arg(ap, uint32_t *);
+
+		*field_value = get_reg_field_value_ex(value, mask, shift);
+		i++;
+	}
+
+	va_end(ap);
+
+	return value;
+}
+
 void reg_sequence_start_gather(const struct dc_context *ctx)
 {
 	/* if reg sequence is supported and enabled, set flag to
@@ -631,8 +666,6 @@ void reg_sequence_start_execute(const struct dc_context *ctx)
 		default:
 			return;
 		}
-
-		dc_dmub_srv_cmd_execute(ctx->dmub_srv);
 	}
 }
 
@@ -649,6 +682,81 @@ void reg_sequence_wait_done(const struct dc_context *ctx)
 	if (offload &&
 	    ctx->dc->debug.dmub_offload_enabled &&
 	    !ctx->dc->debug.dmcub_emulation) {
-		dc_dmub_srv_wait_idle(ctx->dmub_srv);
+		dc_dmub_srv_wait_for_idle(ctx->dmub_srv, DM_DMUB_WAIT_TYPE_WAIT, NULL);
 	}
 }
+
+char *dce_version_to_string(const int version)
+{
+	switch (version) {
+	case DCE_VERSION_6_0:
+		return "DCE 6.0";
+	case DCE_VERSION_6_1:
+		return "DCE 6.1";
+	case DCE_VERSION_6_4:
+		return "DCE 6.4";
+	case DCE_VERSION_8_0:
+		return "DCE 8.0";
+	case DCE_VERSION_8_1:
+		return "DCE 8.1";
+	case DCE_VERSION_8_3:
+		return "DCE 8.3";
+	case DCE_VERSION_10_0:
+		return "DCE 10.0";
+	case DCE_VERSION_11_0:
+		return "DCE 11.0";
+	case DCE_VERSION_11_2:
+		return "DCE 11.2";
+	case DCE_VERSION_11_22:
+		return "DCE 11.22";
+	case DCE_VERSION_12_0:
+		return "DCE 12.0";
+	case DCE_VERSION_12_1:
+		return "DCE 12.1";
+	case DCN_VERSION_1_0:
+		return "DCN 1.0";
+	case DCN_VERSION_1_01:
+		return "DCN 1.0.1";
+	case DCN_VERSION_2_0:
+		return "DCN 2.0";
+	case DCN_VERSION_2_1:
+		return "DCN 2.1";
+	case DCN_VERSION_2_01:
+		return "DCN 2.0.1";
+	case DCN_VERSION_3_0:
+		return "DCN 3.0";
+	case DCN_VERSION_3_01:
+		return "DCN 3.0.1";
+	case DCN_VERSION_3_02:
+		return "DCN 3.0.2";
+	case DCN_VERSION_3_03:
+		return "DCN 3.0.3";
+	case DCN_VERSION_3_1:
+		return "DCN 3.1.2";
+	case DCN_VERSION_3_14:
+		return "DCN 3.1.4";
+	case DCN_VERSION_3_15:
+		return "DCN 3.1.5";
+	case DCN_VERSION_3_16:
+		return "DCN 3.1.6";
+	case DCN_VERSION_3_2:
+		return "DCN 3.2";
+	case DCN_VERSION_3_21:
+		return "DCN 3.2.1";
+	case DCN_VERSION_3_5:
+		return "DCN 3.5";
+	case DCN_VERSION_3_51:
+		return "DCN 3.5.1";
+	case DCN_VERSION_3_6:
+		return "DCN 3.6";
+	case DCN_VERSION_4_01:
+		return "DCN 4.0.1";
+	default:
+		return "Unknown";
+	}
+}
+
+bool dc_supports_vrr(const enum dce_version v)
+{
+	return v >= DCE_VERSION_8_0;
+}
diff --git a/drivers/gpu/drm/amd/display/dc/dc_hw_types.h b/drivers/gpu/drm/amd/display/dc/dc_hw_types.h
index 52355fe6994c..667852517246 100644
--- a/drivers/gpu/drm/amd/display/dc/dc_hw_types.h
+++ b/drivers/gpu/drm/amd/display/dc/dc_hw_types.h
@@ -60,6 +60,7 @@ union large_integer {
 
 enum dc_plane_addr_type {
 	PLN_ADDR_TYPE_GRAPHICS = 0,
+	PLN_ADDR_TYPE_3DLUT,
 	PLN_ADDR_TYPE_GRPH_STEREO,
 	PLN_ADDR_TYPE_VIDEO_PROGRESSIVE,
 	PLN_ADDR_TYPE_RGBEA
@@ -67,7 +68,7 @@ enum dc_plane_addr_type {
 
 struct dc_plane_address {
 	enum dc_plane_addr_type type;
-	bool tmz_surface;
+	uint8_t tmz_surface;
 	union {
 		struct{
 			PHYSICAL_ADDRESS_LOC addr;
@@ -76,6 +77,10 @@ struct dc_plane_address {
 			union large_integer dcc_const_color;
 		} grph;
 
+		struct {
+			PHYSICAL_ADDRESS_LOC addr;
+		} lut3d;
+
 		/*stereo*/
 		struct {
 			PHYSICAL_ADDRESS_LOC left_addr;
@@ -93,7 +98,6 @@ struct dc_plane_address {
 			PHYSICAL_ADDRESS_LOC right_alpha_addr;
 			PHYSICAL_ADDRESS_LOC right_alpha_meta_addr;
 			union large_integer right_alpha_dcc_const_color;
-
 		} grph_stereo;
 
 		/*video  progressive*/
@@ -201,8 +205,9 @@ enum surface_pixel_format {
 	SURFACE_PIXEL_FORMAT_VIDEO_420_YCrCb,
 	SURFACE_PIXEL_FORMAT_VIDEO_420_10bpc_YCbCr,
 	SURFACE_PIXEL_FORMAT_VIDEO_420_10bpc_YCrCb,
+	SURFACE_PIXEL_FORMAT_SUBSAMPLE_END,
+	SURFACE_PIXEL_FORMAT_VIDEO_ACrYCb2101010 =
 		SURFACE_PIXEL_FORMAT_SUBSAMPLE_END,
-	SURFACE_PIXEL_FORMAT_VIDEO_ACrYCb2101010,
 	SURFACE_PIXEL_FORMAT_VIDEO_CrYCbA1010102,
 	SURFACE_PIXEL_FORMAT_VIDEO_AYCrCb8888,
 	SURFACE_PIXEL_FORMAT_INVALID
@@ -235,6 +240,22 @@ enum pixel_format {
 	PIXEL_FORMAT_UNKNOWN
 };
 
+/*
+ * This structure holds a surface address.  There could be multiple addresses
+ * in cases such as Stereo 3D, Planar YUV, etc.  Other per-flip attributes such
+ * as frame durations and DCC format can also be set.
+ */
+#define DC_MAX_DIRTY_RECTS 3
+struct dc_flip_addrs {
+	struct dc_plane_address address;
+	unsigned long long flip_timestamp_in_us;
+	bool flip_immediate;
+	/* TODO: add flip duration for FreeSync */
+	bool triplebuffer_flips;
+	unsigned int dirty_rect_count;
+	struct rect dirty_rects[DC_MAX_DIRTY_RECTS];
+};
+
 enum tile_split_values {
 	DC_DISPLAY_MICRO_TILING = 0x0,
 	DC_THIN_MICRO_TILING = 0x1,
@@ -246,6 +267,9 @@ enum tripleBuffer_enable {
 	DC_TRIPLEBUFFER_DISABLE = 0x0,
 	DC_TRIPLEBUFFER_ENABLE = 0x1,
 };
+enum tile_split_values_new {
+	DC_SURF_TILE_SPLIT_1KB = 0x4,
+};
 
 /* TODO: These values come from hardware spec. We need to readdress this
  * if they ever change.
@@ -303,86 +327,115 @@ enum swizzle_mode_values {
 	DC_SW_UNKNOWN = DC_SW_MAX
 };
 
-union dc_tiling_info {
-
-	struct {
-		/* Specifies the number of memory banks for tiling
-		 *	purposes.
-		 * Only applies to 2D and 3D tiling modes.
-		 *	POSSIBLE VALUES: 2,4,8,16
-		 */
-		unsigned int num_banks;
-		/* Specifies the number of tiles in the x direction
-		 *	to be incorporated into the same bank.
-		 * Only applies to 2D and 3D tiling modes.
-		 *	POSSIBLE VALUES: 1,2,4,8
-		 */
-		unsigned int bank_width;
-		unsigned int bank_width_c;
-		/* Specifies the number of tiles in the y direction to
-		 *	be incorporated into the same bank.
-		 * Only applies to 2D and 3D tiling modes.
-		 *	POSSIBLE VALUES: 1,2,4,8
-		 */
-		unsigned int bank_height;
-		unsigned int bank_height_c;
-		/* Specifies the macro tile aspect ratio. Only applies
-		 * to 2D and 3D tiling modes.
-		 */
-		unsigned int tile_aspect;
-		unsigned int tile_aspect_c;
-		/* Specifies the number of bytes that will be stored
-		 *	contiguously for each tile.
-		 * If the tile data requires more storage than this
-		 *	amount, it is split into multiple slices.
-		 * This field must not be larger than
-		 *	GB_ADDR_CONFIG.DRAM_ROW_SIZE.
-		 * Only applies to 2D and 3D tiling modes.
-		 * For color render targets, TILE_SPLIT >= 256B.
-		 */
-		enum tile_split_values tile_split;
-		enum tile_split_values tile_split_c;
-		/* Specifies the addressing within a tile.
-		 *	0x0 - DISPLAY_MICRO_TILING
-		 *	0x1 - THIN_MICRO_TILING
-		 *	0x2 - DEPTH_MICRO_TILING
-		 *	0x3 - ROTATED_MICRO_TILING
-		 */
-		enum tile_mode_values tile_mode;
-		enum tile_mode_values tile_mode_c;
-		/* Specifies the number of pipes and how they are
-		 *	interleaved in the surface.
-		 * Refer to memory addressing document for complete
-		 *	details and constraints.
-		 */
-		unsigned int pipe_config;
-		/* Specifies the tiling mode of the surface.
-		 * THIN tiles use an 8x8x1 tile size.
-		 * THICK tiles use an 8x8x4 tile size.
-		 * 2D tiling modes rotate banks for successive Z slices
-		 * 3D tiling modes rotate pipes and banks for Z slices
-		 * Refer to memory addressing document for complete
-		 *	details and constraints.
-		 */
-		enum array_mode_values array_mode;
-	} gfx8;
+// Definition of swizzle modes with addr3 ASICs
+enum swizzle_mode_addr3_values {
+	DC_ADDR3_SW_LINEAR = 0,
+	DC_ADDR3_SW_256B_2D = 1,
+	DC_ADDR3_SW_4KB_2D = 2,
+	DC_ADDR3_SW_64KB_2D = 3,
+	DC_ADDR3_SW_256KB_2D = 4,
+	DC_ADDR3_SW_4KB_3D = 5,
+	DC_ADDR3_SW_64KB_3D = 6,
+	DC_ADDR3_SW_256KB_3D = 7,
+	DC_ADDR3_SW_MAX = 8,
+	DC_ADDR3_SW_UNKNOWN = DC_ADDR3_SW_MAX
+};
 
-	struct {
-		enum swizzle_mode_values swizzle;
-		unsigned int num_pipes;
-		unsigned int max_compressed_frags;
-		unsigned int pipe_interleave;
+enum dc_gfxversion {
+	DcGfxVersion7 = 0,
+	DcGfxVersion8,
+	DcGfxVersion9,
+	DcGfxVersion10,
+	DcGfxVersion11,
+	DcGfxAddr3,
+	DcGfxVersionUnknown
+};
 
-		unsigned int num_banks;
-		unsigned int num_shader_engines;
-		unsigned int num_rb_per_se;
-		bool shaderEnable;
+ struct dc_tiling_info {
+	unsigned int gfxversion;     // Specifies which part of the union to use. Must use DalGfxVersion enum
+	union {
+		struct {
+			/* Specifies the number of memory banks for tiling
+			 *	purposes.
+			 * Only applies to 2D and 3D tiling modes.
+			 *	POSSIBLE VALUES: 2,4,8,16
+			 */
+			unsigned int num_banks;
+			/* Specifies the number of tiles in the x direction
+			 *	to be incorporated into the same bank.
+			 * Only applies to 2D and 3D tiling modes.
+			 *	POSSIBLE VALUES: 1,2,4,8
+			 */
+			unsigned int bank_width;
+			unsigned int bank_width_c;
+			/* Specifies the number of tiles in the y direction to
+			 *	be incorporated into the same bank.
+			 * Only applies to 2D and 3D tiling modes.
+			 *	POSSIBLE VALUES: 1,2,4,8
+			 */
+			unsigned int bank_height;
+			unsigned int bank_height_c;
+			/* Specifies the macro tile aspect ratio. Only applies
+			 * to 2D and 3D tiling modes.
+			 */
+			unsigned int tile_aspect;
+			unsigned int tile_aspect_c;
+			/* Specifies the number of bytes that will be stored
+			 *	contiguously for each tile.
+			 * If the tile data requires more storage than this
+			 *	amount, it is split into multiple slices.
+			 * This field must not be larger than
+			 *	GB_ADDR_CONFIG.DRAM_ROW_SIZE.
+			 * Only applies to 2D and 3D tiling modes.
+			 * For color render targets, TILE_SPLIT >= 256B.
+			 */
+			enum tile_split_values tile_split;
+			enum tile_split_values tile_split_c;
+			/* Specifies the addressing within a tile.
+			 *	0x0 - DISPLAY_MICRO_TILING
+			 *	0x1 - THIN_MICRO_TILING
+			 *	0x2 - DEPTH_MICRO_TILING
+			 *	0x3 - ROTATED_MICRO_TILING
+			 */
+			enum tile_mode_values tile_mode;
+			enum tile_mode_values tile_mode_c;
+			/* Specifies the number of pipes and how they are
+			 *	interleaved in the surface.
+			 * Refer to memory addressing document for complete
+			 *	details and constraints.
+			 */
+			unsigned int pipe_config;
+			/* Specifies the tiling mode of the surface.
+			 * THIN tiles use an 8x8x1 tile size.
+			 * THICK tiles use an 8x8x4 tile size.
+			 * 2D tiling modes rotate banks for successive Z slices
+			 * 3D tiling modes rotate pipes and banks for Z slices
+			 * Refer to memory addressing document for complete
+			 *	details and constraints.
+			 */
+			enum array_mode_values array_mode;
+		} gfx8;
 
-		bool meta_linear;
-		bool rb_aligned;
-		bool pipe_aligned;
-		unsigned int num_pkrs;
-	} gfx9;
+		struct {
+			enum swizzle_mode_values swizzle;
+			unsigned int num_pipes;
+			unsigned int max_compressed_frags;
+			unsigned int pipe_interleave;
+
+			unsigned int num_banks;
+			unsigned int num_shader_engines;
+			unsigned int num_rb_per_se;
+			bool shaderEnable;
+
+			bool meta_linear;
+			bool rb_aligned;
+			bool pipe_aligned;
+			unsigned int num_pkrs;
+		} gfx9;/*gfx9, gfx10 and above*/
+		struct {
+			enum swizzle_mode_addr3_values swizzle;
+		} gfx_addr3;/*gfx with addr3 and above*/
+	};
 };
 
 /* Rotation angle */
@@ -400,19 +453,43 @@ enum dc_scan_direction {
 	SCAN_DIRECTION_VERTICAL = 2,    /* 90, 270 rotation */
 };
 
+/**
+ * struct dc_cursor_position: Hardware cursor data.
+ *
+ * This struct keeps the action information related to the cursor that will be
+ * sent and received from our DC core.
+ */
 struct dc_cursor_position {
+	/**
+	 * @x: It represents the top left abscissa coordinate of the cursor.
+	 */
 	uint32_t x;
+
+	/**
+	 * @y: It is the top ordinate of the cursor coordinate.
+	 */
 	uint32_t y;
 
+	/**
+	 * @x_hotspot: Define the abscissa point where mouse click happens.
+	 */
 	uint32_t x_hotspot;
+
+	/**
+	 * @y_hotspot: Define the ordinate point where mouse click happens.
+	 */
 	uint32_t y_hotspot;
 
-	/*
-	 * This parameter indicates whether HW cursor should be enabled
+	/**
+	 * @enable: This parameter indicates whether hardware cursor should be
+	 * enabled.
 	 */
 	bool enable;
 
-	/* Translate cursor x/y by the source rectangle for each plane. */
+	/**
+	 * @translate_by_source: Translate cursor x/y by the source rectangle
+	 * for each plane.
+	 */
 	bool translate_by_source;
 };
 
@@ -420,10 +497,12 @@ struct dc_cursor_mi_param {
 	unsigned int pixel_clk_khz;
 	unsigned int ref_clk_khz;
 	struct rect viewport;
+	struct rect recout;
 	struct fixed31_32 h_scale_ratio;
 	struct fixed31_32 v_scale_ratio;
 	enum dc_rotation_angle rotation;
 	bool mirror;
+	struct dc_stream_state *stream;
 };
 
 /* IPP related types */
@@ -477,7 +556,9 @@ struct dc_gamma {
 /* Used by both ipp amd opp functions*/
 /* TODO: to be consolidated with enum color_space */
 
-/*
+/**
+ * enum dc_cursor_color_format - DC cursor programming mode
+ *
  * This enum is for programming CURSOR_MODE register field. What this register
  * should be programmed to depends on OS requested cursor shape flags and what
  * we stored in the cursor surface.
@@ -513,17 +594,39 @@ union dc_cursor_attribute_flags {
 };
 
 struct dc_cursor_attributes {
+	/**
+	 * @address: This field represents the framebuffer address associated
+	 * with the cursor. It is important to highlight that this address is
+	 * divided into a high and low parts.
+	 */
 	PHYSICAL_ADDRESS_LOC address;
+
+	/**
+	 * @pitch: Cursor line stride.
+	 */
 	uint32_t pitch;
 
-	/* Width and height should correspond to cursor surface width x heigh */
+	/**
+	 * @width: Width should correspond to cursor surface width.
+	 */
 	uint32_t width;
+	/**
+	 * @heigh: Height should correspond to cursor surface heigh.
+	 */
 	uint32_t height;
 
+	/**
+	 * @color_format: DC cursor programming mode.
+	 */
 	enum dc_cursor_color_format color_format;
-	uint32_t sdr_white_level; // for boosting (SDR) cursor in HDR mode
+	/**
+	 * @sdr_white_level: Boosting (SDR) cursor in HDR mode.
+	 */
+	uint32_t sdr_white_level;
 
-	/* In case we support HW Cursor rotation in the future */
+	/**
+	 * @rotation_angle: In case we support HW Cursor rotation in the future
+	 */
 	enum dc_rotation_angle rotation_angle;
 
 	union dc_cursor_attribute_flags attribute_flags;
@@ -550,7 +653,8 @@ enum dc_color_space {
 	COLOR_SPACE_YCBCR709_LIMITED,
 	COLOR_SPACE_2020_RGB_FULLRANGE,
 	COLOR_SPACE_2020_RGB_LIMITEDRANGE,
-	COLOR_SPACE_2020_YCBCR,
+	COLOR_SPACE_2020_YCBCR_LIMITED,
+	COLOR_SPACE_2020_YCBCR_FULL,
 	COLOR_SPACE_ADOBERGB,
 	COLOR_SPACE_DCIP3,
 	COLOR_SPACE_DISPLAYNATIVE,
@@ -558,6 +662,7 @@ enum dc_color_space {
 	COLOR_SPACE_APPCTRL,
 	COLOR_SPACE_CUSTOMPOINTS,
 	COLOR_SPACE_YCBCR709_BLACK,
+	COLOR_SPACE_2020_YCBCR = COLOR_SPACE_2020_YCBCR_LIMITED,
 };
 
 enum dc_dither_option {
@@ -704,9 +809,6 @@ struct dc_crtc_timing_flags {
 	uint32_t LTE_340MCSC_SCRAMBLE:1;
 
 	uint32_t DSC : 1; /* Use DSC with this timing */
-#ifndef TRIM_FSFT
-	uint32_t FAST_TRANSPORT: 1;
-#endif
 	uint32_t VBLANK_SYNCHRONIZABLE: 1;
 };
 
@@ -732,6 +834,29 @@ enum dc_timing_3d_format {
 	TIMING_3D_FORMAT_MAX,
 };
 
+#define DC_DSC_QP_SET_SIZE 15
+#define DC_DSC_RC_BUF_THRESH_SIZE 14
+struct dc_dsc_rc_params_override {
+	int32_t rc_model_size;
+	int32_t rc_buf_thresh[DC_DSC_RC_BUF_THRESH_SIZE];
+	int32_t rc_minqp[DC_DSC_QP_SET_SIZE];
+	int32_t rc_maxqp[DC_DSC_QP_SET_SIZE];
+	int32_t rc_offset[DC_DSC_QP_SET_SIZE];
+
+	int32_t rc_tgt_offset_hi;
+	int32_t rc_tgt_offset_lo;
+	int32_t rc_edge_factor;
+	int32_t rc_quant_incr_limit0;
+	int32_t rc_quant_incr_limit1;
+
+	int32_t initial_fullness_offset;
+	int32_t initial_delay;
+
+	int32_t flatness_min_qp;
+	int32_t flatness_max_qp;
+	int32_t flatness_det_thresh;
+};
+
 struct dc_dsc_config {
 	uint32_t num_slices_h; /* Number of DSC slices - horizontal */
 	uint32_t num_slices_v; /* Number of DSC slices - vertical */
@@ -741,40 +866,133 @@ struct dc_dsc_config {
 	uint32_t version_minor; /* DSC minor version. Full version is formed as 1.version_minor. */
 	bool ycbcr422_simple; /* Tell DSC engine to convert YCbCr 4:2:2 to 'YCbCr 4:2:2 simple'. */
 	int32_t rc_buffer_size; /* DSC RC buffer block size in bytes */
+	bool is_frl; /* indicate if DSC is applied based on HDMI FRL sink's capability */
 	bool is_dp; /* indicate if DSC is applied based on DP's capability */
+	uint32_t mst_pbn; /* pbn of display on dsc mst hub */
+	const struct dc_dsc_rc_params_override *rc_params_ovrd; /* DM owned memory. If not NULL, apply custom dsc rc params */
 };
+
+/**
+ * struct dc_crtc_timing - Timing parameters used to configure DCN blocks
+ *
+ * DCN provides multiple signals and parameters that can be used to adjust
+ * timing parameters, this struct aggregate multiple of these values for easy
+ * access. In this struct, fields prefixed with h_* are related to horizontal
+ * timing, and v_* to vertical timing. Keep in mind that when we talk about
+ * vertical timings, the values, in general, are described in the number of
+ * lines; on the other hand, the horizontal values are in pixels.
+ */
 struct dc_crtc_timing {
+	/**
+	 * @h_total: The total number of pixels from the rising edge of HSync
+	 * until the rising edge of the current HSync.
+	 */
 	uint32_t h_total;
+
+	/**
+	 * @h_border_left: The black pixels related to the left border
+	 */
 	uint32_t h_border_left;
+
+	/**
+	 * @h_addressable: It is the range of pixels displayed horizontally.
+	 * For example, if the display resolution is 3840@2160, the horizontal
+	 * addressable area is 3840.
+	 */
 	uint32_t h_addressable;
+
+	/**
+	 * @h_border_right: The black pixels related to the right border
+	 */
 	uint32_t h_border_right;
+
+	/**
+	 * @h_front_porch: Period (in pixels) between HBlank start and the
+	 * rising edge of HSync.
+	 */
 	uint32_t h_front_porch;
+
+	/**
+	 * @h_sync_width: HSync duration in pixels.
+	 */
 	uint32_t h_sync_width;
 
+	/**
+	 * @v_total: It is the total number of lines from the rising edge of
+	 * the previous VSync until the rising edge of the current VSync.
+	 *
+	 *          |--------------------------|
+	 *          +-+        V_TOTAL         +-+
+	 *          | |                        | |
+	 * VSync ---+ +--------- // -----------+ +---
+	 */
 	uint32_t v_total;
+
+	/**
+	 * @v_border_top: The black border on the top.
+	 */
 	uint32_t v_border_top;
+
+	/**
+	 * @v_addressable: It is the range of the scanout at which the
+	 * framebuffer is displayed. For example, if the display resolution is
+	 * 3840@2160, the addressable area is 2160 lines, or if the resolution
+	 * is 1920x1080, the addressable area is 1080 lines.
+	 */
 	uint32_t v_addressable;
+
+	/**
+	 * @v_border_bottom: The black border on the bottom.
+	 */
 	uint32_t v_border_bottom;
+
+	/**
+	 * @v_front_porch: Period (in lines) between VBlank start and rising
+	 * edge of VSync.
+	 *                  +-+
+	 * VSync            | |
+	 *        ----------+ +--------...
+	 *          +------------------...
+	 * VBlank   |
+	 *        --+
+	 *          |-------|
+	 *        v_front_porch
+	 */
 	uint32_t v_front_porch;
+
+	/**
+	 * @v_sync_width: VSync signal width in lines.
+	 */
 	uint32_t v_sync_width;
 
+	/**
+	 * @pix_clk_100hz: Pipe pixel precision
+	 *
+	 * This field is used to communicate pixel clocks with 100 Hz accuracy
+	 * from dc_crtc_timing to BIOS command table.
+	 */
 	uint32_t pix_clk_100hz;
 
+	uint32_t min_refresh_in_uhz;
+	uint32_t max_refresh_in_uhz;
+
 	uint32_t vic;
 	uint32_t hdmi_vic;
+	uint32_t rid;
+	uint32_t fr_index;
+	uint32_t frl_uncompressed_video_bandwidth_in_kbps;
 	enum dc_timing_3d_format timing_3d_format;
 	enum dc_color_depth display_color_depth;
 	enum dc_pixel_encoding pixel_encoding;
 	enum dc_aspect_ratio aspect_ratio;
 	enum scanning_type scan_type;
 
-#ifndef TRIM_FSFT
-	uint32_t fast_transport_output_rate_100hz;
-#endif
-
 	struct dc_crtc_timing_flags flags;
 	uint32_t dsc_fixed_bits_per_pixel_x16; /* DSC target bitrate in 1/16 of bpp (e.g. 128 -> 8bpp) */
 	struct dc_dsc_config dsc_cfg;
+
+	/* The number of pixels that HBlank has been expanded by from the original EDID timing. */
+	uint32_t expanded_hblank;
 };
 
 enum trigger_delay {
@@ -799,6 +1017,8 @@ struct dc_crtc_timing_adjust {
 	uint32_t v_total_max;
 	uint32_t v_total_mid;
 	uint32_t v_total_mid_frame_num;
+	uint32_t allow_otg_v_count_halt;
+	uint8_t timing_adjust_pending;
 };
 
 
@@ -875,6 +1095,25 @@ enum cm_gamut_coef_format {
 	CM_GAMUT_REMAP_COEF_FORMAT_S3_12 = 1
 };
 
+enum mpcc_gamut_remap_mode_select {
+	MPCC_GAMUT_REMAP_MODE_SELECT_0 = 0,
+	MPCC_GAMUT_REMAP_MODE_SELECT_1,
+	MPCC_GAMUT_REMAP_MODE_SELECT_2
+};
+
+enum mpcc_gamut_remap_id {
+	MPCC_OGAM_GAMUT_REMAP,
+	MPCC_MCM_FIRST_GAMUT_REMAP,
+	MPCC_MCM_SECOND_GAMUT_REMAP,
+	MPCC_RMCM_GAMUT_REMAP,
+};
+
+enum cursor_matrix_mode {
+	CUR_MATRIX_BYPASS = 0,
+	CUR_MATRIX_SET_A,
+	CUR_MATRIX_SET_B
+};
+
 struct mcif_warmup_params {
 	union large_integer	start_address;
 	unsigned int		address_increment;
@@ -903,5 +1142,19 @@ struct tg_color {
 	uint16_t color_b_cb;
 };
 
+enum symclk_state {
+	SYMCLK_OFF_TX_OFF,
+	SYMCLK_ON_TX_ON,
+	SYMCLK_ON_TX_OFF,
+};
+
+struct phy_state {
+	struct {
+		uint8_t otg		: 1;
+		uint8_t reserved	: 7;
+	} symclk_ref_cnts;
+	enum symclk_state symclk_state;
+};
+
 #endif /* DC_HW_TYPES_H */
 
diff --git a/drivers/gpu/drm/amd/display/dc/dc_link.h b/drivers/gpu/drm/amd/display/dc/dc_link.h
deleted file mode 100644
index fad3d883ed89..000000000000
--- a/drivers/gpu/drm/amd/display/dc/dc_link.h
+++ /dev/null
@@ -1,447 +0,0 @@
-/*
- * Copyright 2012-14 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: AMD
- *
- */
-
-#ifndef DC_LINK_H_
-#define DC_LINK_H_
-
-#include "dc.h"
-#include "dc_types.h"
-#include "grph_object_defs.h"
-
-enum dc_link_fec_state {
-	dc_link_fec_not_ready,
-	dc_link_fec_ready,
-	dc_link_fec_enabled
-};
-
-struct dc_link_status {
-	bool link_active;
-	struct dpcd_caps *dpcd_caps;
-};
-
-/* DP MST stream allocation (payload bandwidth number) */
-struct link_mst_stream_allocation {
-	/* DIG front */
-	const struct stream_encoder *stream_enc;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	/* HPO DP Stream Encoder */
-	const struct hpo_dp_stream_encoder *hpo_dp_stream_enc;
-#endif
-	/* associate DRM payload table with DC stream encoder */
-	uint8_t vcp_id;
-	/* number of slots required for the DP stream in transport packet */
-	uint8_t slot_count;
-};
-
-/* DP MST stream allocation table */
-struct link_mst_stream_allocation_table {
-	/* number of DP video streams */
-	int stream_count;
-	/* array of stream allocations */
-	struct link_mst_stream_allocation stream_allocations[MAX_CONTROLLER_NUM];
-};
-
-struct time_stamp {
-	uint64_t edp_poweroff;
-	uint64_t edp_poweron;
-};
-
-struct link_trace {
-	struct time_stamp time_stamp;
-};
-
-/* PSR feature flags */
-struct psr_settings {
-	bool psr_feature_enabled;		// PSR is supported by sink
-	bool psr_allow_active;			// PSR is currently active
-	enum dc_psr_version psr_version;		// Internal PSR version, determined based on DPCD
-
-	/* These parameters are calculated in Driver,
-	 * based on display timing and Sink capabilities.
-	 * If VBLANK region is too small and Sink takes a long time
-	 * to set up RFB, it may take an extra frame to enter PSR state.
-	 */
-	bool psr_frame_capture_indication_req;
-	unsigned int psr_sdp_transmit_line_num_deadline;
-	unsigned int psr_power_opt;
-};
-
-/*
- * A link contains one or more sinks and their connected status.
- * The currently active signal type (HDMI, DP-SST, DP-MST) is also reported.
- */
-struct dc_link {
-	struct dc_sink *remote_sinks[MAX_SINKS_PER_LINK];
-	unsigned int sink_count;
-	struct dc_sink *local_sink;
-	unsigned int link_index;
-	enum dc_connection_type type;
-	enum signal_type connector_signal;
-	enum dc_irq_source irq_source_hpd;
-	enum dc_irq_source irq_source_hpd_rx;/* aka DP Short Pulse  */
-	bool is_hpd_filter_disabled;
-	bool dp_ss_off;
-	bool link_state_valid;
-	bool aux_access_disabled;
-	bool sync_lt_in_progress;
-	enum lttpr_mode lttpr_mode;
-	bool is_internal_display;
-
-	/* TODO: Rename. Flag an endpoint as having a programmable mapping to a
-	 * DIG encoder. */
-	bool is_dig_mapping_flexible;
-	bool hpd_status; /* HPD status of link without physical HPD pin. */
-
-	bool edp_sink_present;
-
-	/* caps is the same as reported_link_cap. link_traing use
-	 * reported_link_cap. Will clean up.  TODO
-	 */
-	struct dc_link_settings reported_link_cap;
-	struct dc_link_settings verified_link_cap;
-	struct dc_link_settings cur_link_settings;
-	struct dc_lane_settings cur_lane_setting[LANE_COUNT_DP_MAX];
-	struct dc_link_settings preferred_link_setting;
-	/* preferred_training_settings are override values that
-	 * come from DM. DM is responsible for the memory
-	 * management of the override pointers.
-	 */
-	struct dc_link_training_overrides preferred_training_settings;
-	struct dp_audio_test_data audio_test_data;
-
-	uint8_t ddc_hw_inst;
-
-	uint8_t hpd_src;
-
-	uint8_t link_enc_hw_inst;
-	/* DIG link encoder ID. Used as index in link encoder resource pool.
-	 * For links with fixed mapping to DIG, this is not changed after dc_link
-	 * object creation.
-	 */
-	enum engine_id eng_id;
-
-	bool test_pattern_enabled;
-	union compliance_test_state compliance_test_state;
-
-	void *priv;
-
-	struct ddc_service *ddc;
-
-	bool aux_mode;
-
-	/* Private to DC core */
-
-	const struct dc *dc;
-
-	struct dc_context *ctx;
-
-	struct panel_cntl *panel_cntl;
-	struct link_encoder *link_enc;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	struct hpo_dp_link_encoder *hpo_dp_link_enc;
-#endif
-	struct graphics_object_id link_id;
-	/* Endpoint type distinguishes display endpoints which do not have entries
-	 * in the BIOS connector table from those that do. Helps when tracking link
-	 * encoder to display endpoint assignments.
-	 */
-	enum display_endpoint_type ep_type;
-	union ddi_channel_mapping ddi_channel_mapping;
-	struct connector_device_tag_info device_tag;
-	struct dpcd_caps dpcd_caps;
-	uint32_t dongle_max_pix_clk;
-	unsigned short chip_caps;
-	unsigned int dpcd_sink_count;
-#if defined(CONFIG_DRM_AMD_DC_HDCP)
-	struct hdcp_caps hdcp_caps;
-#endif
-	enum edp_revision edp_revision;
-	union dpcd_sink_ext_caps dpcd_sink_ext_caps;
-
-	struct psr_settings psr_settings;
-
-	/* Drive settings read from integrated info table */
-	struct dc_lane_settings bios_forced_drive_settings;
-
-	/* MST record stream using this link */
-	struct link_flags {
-		bool dp_keep_receiver_powered;
-		bool dp_skip_DID2;
-		bool dp_skip_reset_segment;
-		bool dp_mot_reset_segment;
-		/* Some USB4 docks do not handle turning off MST DSC once it has been enabled. */
-		bool dpia_mst_dsc_always_on;
-	} wa_flags;
-	struct link_mst_stream_allocation_table mst_stream_alloc_table;
-
-	struct dc_link_status link_status;
-
-	struct link_trace link_trace;
-	struct gpio *hpd_gpio;
-	enum dc_link_fec_state fec_state;
-};
-
-const struct dc_link_status *dc_link_get_status(const struct dc_link *dc_link);
-
-/**
- * dc_get_link_at_index() - Return an enumerated dc_link.
- *
- * dc_link order is constant and determined at
- * boot time.  They cannot be created or destroyed.
- * Use dc_get_caps() to get number of links.
- */
-static inline struct dc_link *dc_get_link_at_index(struct dc *dc, uint32_t link_index)
-{
-	return dc->links[link_index];
-}
-
-static inline void get_edp_links(const struct dc *dc,
-		struct dc_link **edp_links,
-		int *edp_num)
-{
-	int i;
-
-	*edp_num = 0;
-	for (i = 0; i < dc->link_count; i++) {
-		// report any eDP links, even unconnected DDI's
-		if (!dc->links[i])
-			continue;
-		if (dc->links[i]->connector_signal == SIGNAL_TYPE_EDP) {
-			edp_links[*edp_num] = dc->links[i];
-			if (++(*edp_num) == MAX_NUM_EDP)
-				return;
-		}
-	}
-}
-
-static inline bool dc_get_edp_link_panel_inst(const struct dc *dc,
-		const struct dc_link *link,
-		unsigned int *inst_out)
-{
-	struct dc_link *edp_links[MAX_NUM_EDP];
-	int edp_num;
-
-	if (link->connector_signal != SIGNAL_TYPE_EDP)
-		return false;
-	get_edp_links(dc, edp_links, &edp_num);
-	if ((edp_num > 1) && (link->link_index > edp_links[0]->link_index))
-		*inst_out = 1;
-	else
-		*inst_out = 0;
-	return true;
-}
-
-/* Set backlight level of an embedded panel (eDP, LVDS).
- * backlight_pwm_u16_16 is unsigned 32 bit with 16 bit integer
- * and 16 bit fractional, where 1.0 is max backlight value.
- */
-bool dc_link_set_backlight_level(const struct dc_link *dc_link,
-		uint32_t backlight_pwm_u16_16,
-		uint32_t frame_ramp);
-
-/* Set/get nits-based backlight level of an embedded panel (eDP, LVDS). */
-bool dc_link_set_backlight_level_nits(struct dc_link *link,
-		bool isHDR,
-		uint32_t backlight_millinits,
-		uint32_t transition_time_in_ms);
-
-bool dc_link_get_backlight_level_nits(struct dc_link *link,
-		uint32_t *backlight_millinits,
-		uint32_t *backlight_millinits_peak);
-
-bool dc_link_backlight_enable_aux(struct dc_link *link, bool enable);
-
-bool dc_link_read_default_bl_aux(struct dc_link *link, uint32_t *backlight_millinits);
-bool dc_link_set_default_brightness_aux(struct dc_link *link);
-
-int dc_link_get_backlight_level(const struct dc_link *dc_link);
-
-int dc_link_get_target_backlight_pwm(const struct dc_link *link);
-
-bool dc_link_set_psr_allow_active(struct dc_link *dc_link, const bool *enable,
-		bool wait, bool force_static, const unsigned int *power_opts);
-
-bool dc_link_get_psr_state(const struct dc_link *dc_link, enum dc_psr_state *state);
-
-bool dc_link_setup_psr(struct dc_link *dc_link,
-		const struct dc_stream_state *stream, struct psr_config *psr_config,
-		struct psr_context *psr_context);
-
-void dc_link_get_psr_residency(const struct dc_link *link, uint32_t *residency);
-
-/* Request DC to detect if there is a Panel connected.
- * boot - If this call is during initial boot.
- * Return false for any type of detection failure or MST detection
- * true otherwise. True meaning further action is required (status update
- * and OS notification).
- */
-enum dc_detect_reason {
-	DETECT_REASON_BOOT,
-	DETECT_REASON_HPD,
-	DETECT_REASON_HPDRX,
-	DETECT_REASON_FALLBACK,
-	DETECT_REASON_RETRAIN
-};
-
-bool dc_link_detect(struct dc_link *dc_link, enum dc_detect_reason reason);
-bool dc_link_get_hpd_state(struct dc_link *dc_link);
-enum dc_status dc_link_allocate_mst_payload(struct pipe_ctx *pipe_ctx);
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-enum dc_status dc_link_reduce_mst_payload(struct pipe_ctx *pipe_ctx, uint32_t req_pbn);
-enum dc_status dc_link_increase_mst_payload(struct pipe_ctx *pipe_ctx, uint32_t req_pbn);
-#endif
-
-/* Notify DC about DP RX Interrupt (aka Short Pulse Interrupt).
- * Return:
- * true - Downstream port status changed. DM should call DC to do the
- * detection.
- * false - no change in Downstream port status. No further action required
- * from DM. */
-bool dc_link_handle_hpd_rx_irq(struct dc_link *dc_link,
-		union hpd_irq_data *hpd_irq_dpcd_data, bool *out_link_loss,
-		bool defer_handling, bool *has_left_work);
-
-/*
- * On eDP links this function call will stall until T12 has elapsed.
- * If the panel is not in power off state, this function will return
- * immediately.
- */
-bool dc_link_wait_for_t12(struct dc_link *link);
-
-void dc_link_dp_handle_automated_test(struct dc_link *link);
-void dc_link_dp_handle_link_loss(struct dc_link *link);
-bool dc_link_dp_allow_hpd_rx_irq(const struct dc_link *link);
-
-struct dc_sink_init_data;
-
-struct dc_sink *dc_link_add_remote_sink(
-		struct dc_link *dc_link,
-		const uint8_t *edid,
-		int len,
-		struct dc_sink_init_data *init_data);
-
-void dc_link_remove_remote_sink(
-	struct dc_link *link,
-	struct dc_sink *sink);
-
-/* Used by diagnostics for virtual link at the moment */
-
-void dc_link_dp_set_drive_settings(
-	struct dc_link *link,
-	struct link_training_settings *lt_settings);
-
-bool dc_link_dp_perform_link_training_skip_aux(
-	struct dc_link *link,
-	const struct dc_link_settings *link_setting);
-
-enum link_training_result dc_link_dp_perform_link_training(
-	struct dc_link *link,
-	const struct dc_link_settings *link_settings,
-	bool skip_video_pattern);
-
-bool dc_link_dp_sync_lt_begin(struct dc_link *link);
-
-enum link_training_result dc_link_dp_sync_lt_attempt(
-	struct dc_link *link,
-	struct dc_link_settings *link_setting,
-	struct dc_link_training_overrides *lt_settings);
-
-bool dc_link_dp_sync_lt_end(struct dc_link *link, bool link_down);
-
-void dc_link_dp_enable_hpd(const struct dc_link *link);
-
-void dc_link_dp_disable_hpd(const struct dc_link *link);
-
-bool dc_link_dp_set_test_pattern(
-	struct dc_link *link,
-	enum dp_test_pattern test_pattern,
-	enum dp_test_pattern_color_space test_pattern_color_space,
-	const struct link_training_settings *p_link_settings,
-	const unsigned char *p_custom_pattern,
-	unsigned int cust_pattern_size);
-
-bool dc_link_dp_get_max_link_enc_cap(const struct dc_link *link, struct dc_link_settings *max_link_enc_cap);
-
-void dc_link_enable_hpd_filter(struct dc_link *link, bool enable);
-
-bool dc_link_is_dp_sink_present(struct dc_link *link);
-
-bool dc_link_detect_sink(struct dc_link *link, enum dc_connection_type *type);
-/*
- * DPCD access interfaces
- */
-
-#ifdef CONFIG_DRM_AMD_DC_HDCP
-bool dc_link_is_hdcp14(struct dc_link *link, enum signal_type signal);
-bool dc_link_is_hdcp22(struct dc_link *link, enum signal_type signal);
-#endif
-void dc_link_set_drive_settings(struct dc *dc,
-				struct link_training_settings *lt_settings,
-				const struct dc_link *link);
-void dc_link_set_preferred_link_settings(struct dc *dc,
-					 struct dc_link_settings *link_setting,
-					 struct dc_link *link);
-void dc_link_set_preferred_training_settings(struct dc *dc,
-					struct dc_link_settings *link_setting,
-					struct dc_link_training_overrides *lt_overrides,
-					struct dc_link *link,
-					bool skip_immediate_retrain);
-void dc_link_enable_hpd(const struct dc_link *link);
-void dc_link_disable_hpd(const struct dc_link *link);
-void dc_link_set_test_pattern(struct dc_link *link,
-			enum dp_test_pattern test_pattern,
-			enum dp_test_pattern_color_space test_pattern_color_space,
-			const struct link_training_settings *p_link_settings,
-			const unsigned char *p_custom_pattern,
-			unsigned int cust_pattern_size);
-uint32_t dc_link_bandwidth_kbps(
-	const struct dc_link *link,
-	const struct dc_link_settings *link_setting);
-
-const struct dc_link_settings *dc_link_get_link_cap(
-		const struct dc_link *link);
-
-void dc_link_overwrite_extended_receiver_cap(
-		struct dc_link *link);
-
-bool dc_submit_i2c(
-		struct dc *dc,
-		uint32_t link_index,
-		struct i2c_command *cmd);
-
-bool dc_submit_i2c_oem(
-		struct dc *dc,
-		struct i2c_command *cmd);
-
-uint32_t dc_bandwidth_in_kbps_from_timing(
-	const struct dc_crtc_timing *timing);
-
-bool dc_link_is_fec_supported(const struct dc_link *link);
-bool dc_link_should_enable_fec(const struct dc_link *link);
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-enum dp_link_encoding dc_link_dp_mst_decide_link_encoding_format(const struct dc_link *link);
-#endif
-#endif /* DC_LINK_H_ */
diff --git a/drivers/gpu/drm/amd/display/dc/dc_plane.h b/drivers/gpu/drm/amd/display/dc/dc_plane.h
new file mode 100644
index 000000000000..14feb843e694
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dc_plane.h
@@ -0,0 +1,51 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef _DC_PLANE_H_
+#define _DC_PLANE_H_
+
+#include "dc_hw_types.h"
+
+union dc_plane_status_update_flags {
+	struct {
+		uint32_t address : 1;
+	} bits;
+	uint32_t raw;
+};
+
+struct dc_plane_state *dc_create_plane_state(const struct dc *dc);
+const struct dc_plane_status *dc_plane_get_status(
+		const struct dc_plane_state *plane_state,
+		union dc_plane_status_update_flags flags);
+void dc_plane_state_retain(struct dc_plane_state *plane_state);
+void dc_plane_state_release(struct dc_plane_state *plane_state);
+
+void dc_plane_force_dcc_and_tiling_disable(struct dc_plane_state *plane_state,
+					   bool clear_tiling);
+
+
+void dc_plane_copy_config(struct dc_plane_state *dst, const struct dc_plane_state *src);
+
+#endif /* _DC_PLANE_H_ */
diff --git a/drivers/gpu/drm/amd/display/dc/dc_plane_priv.h b/drivers/gpu/drm/amd/display/dc/dc_plane_priv.h
new file mode 100644
index 000000000000..ab13335f1d01
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dc_plane_priv.h
@@ -0,0 +1,35 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef _DC_PLANE_PRIV_H_
+#define _DC_PLANE_PRIV_H_
+
+#include "dc_plane.h"
+
+void dc_plane_construct(struct dc_context *ctx, struct dc_plane_state *plane_state);
+void dc_plane_destruct(struct dc_plane_state *plane_state);
+uint8_t dc_plane_get_pipe_mask(struct dc_state *dc_state, const struct dc_plane_state *plane_state);
+
+#endif /* _DC_PLANE_PRIV_H_ */
diff --git a/drivers/gpu/drm/amd/display/dc/dc_spl_translate.c b/drivers/gpu/drm/amd/display/dc/dc_spl_translate.c
new file mode 100644
index 000000000000..37d1a79e8241
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dc_spl_translate.c
@@ -0,0 +1,231 @@
+// SPDX-License-Identifier: MIT
+//
+// Copyright 2024 Advanced Micro Devices, Inc.
+
+#include "dc_spl_translate.h"
+#include "dcn20/dcn20_dpp.h"
+#include "dcn32/dcn32_dpp.h"
+#include "dcn401/dcn401_dpp.h"
+
+static struct spl_callbacks dcn2_spl_callbacks = {
+	.spl_calc_lb_num_partitions = dscl2_spl_calc_lb_num_partitions,
+};
+static struct spl_callbacks dcn32_spl_callbacks = {
+	.spl_calc_lb_num_partitions = dscl32_spl_calc_lb_num_partitions,
+};
+static struct spl_callbacks dcn401_spl_callbacks = {
+	.spl_calc_lb_num_partitions = dscl401_spl_calc_lb_num_partitions,
+};
+static void populate_splrect_from_rect(struct spl_rect *spl_rect, const struct rect *rect)
+{
+	spl_rect->x = rect->x;
+	spl_rect->y = rect->y;
+	spl_rect->width = rect->width;
+	spl_rect->height = rect->height;
+}
+static void populate_rect_from_splrect(struct rect *rect, const struct spl_rect *spl_rect)
+{
+	rect->x = spl_rect->x;
+	rect->y = spl_rect->y;
+	rect->width = spl_rect->width;
+	rect->height = spl_rect->height;
+}
+static void populate_spltaps_from_taps(struct spl_taps *spl_scaling_quality,
+		const struct scaling_taps *scaling_quality)
+{
+	spl_scaling_quality->h_taps_c = scaling_quality->h_taps_c;
+	spl_scaling_quality->h_taps = scaling_quality->h_taps;
+	spl_scaling_quality->v_taps_c = scaling_quality->v_taps_c;
+	spl_scaling_quality->v_taps = scaling_quality->v_taps;
+	spl_scaling_quality->integer_scaling = scaling_quality->integer_scaling;
+}
+static void populate_taps_from_spltaps(struct scaling_taps *scaling_quality,
+		const struct spl_taps *spl_scaling_quality)
+{
+	scaling_quality->h_taps_c = spl_scaling_quality->h_taps_c + 1;
+	scaling_quality->h_taps = spl_scaling_quality->h_taps + 1;
+	scaling_quality->v_taps_c = spl_scaling_quality->v_taps_c + 1;
+	scaling_quality->v_taps = spl_scaling_quality->v_taps + 1;
+}
+static void populate_ratios_from_splratios(struct scaling_ratios *ratios,
+		const struct ratio *spl_ratios)
+{
+	ratios->horz = dc_fixpt_from_ux_dy(spl_ratios->h_scale_ratio >> 5, 3, 19);
+	ratios->vert = dc_fixpt_from_ux_dy(spl_ratios->v_scale_ratio >> 5, 3, 19);
+	ratios->horz_c = dc_fixpt_from_ux_dy(spl_ratios->h_scale_ratio_c >> 5, 3, 19);
+	ratios->vert_c = dc_fixpt_from_ux_dy(spl_ratios->v_scale_ratio_c >> 5, 3, 19);
+}
+static void populate_inits_from_splinits(struct scl_inits *inits,
+		const struct init *spl_inits)
+{
+	inits->h = dc_fixpt_from_int_dy(spl_inits->h_filter_init_int, spl_inits->h_filter_init_frac >> 5, 0, 19);
+	inits->v = dc_fixpt_from_int_dy(spl_inits->v_filter_init_int, spl_inits->v_filter_init_frac >> 5, 0, 19);
+	inits->h_c = dc_fixpt_from_int_dy(spl_inits->h_filter_init_int_c, spl_inits->h_filter_init_frac_c >> 5, 0, 19);
+	inits->v_c = dc_fixpt_from_int_dy(spl_inits->v_filter_init_int_c, spl_inits->v_filter_init_frac_c >> 5, 0, 19);
+}
+static void populate_splformat_from_format(enum spl_pixel_format *spl_pixel_format, const enum pixel_format pixel_format)
+{
+	if (pixel_format < PIXEL_FORMAT_INVALID)
+		*spl_pixel_format = (enum spl_pixel_format)pixel_format;
+	else
+		*spl_pixel_format = SPL_PIXEL_FORMAT_INVALID;
+}
+/// @brief Translate SPL input parameters from pipe context
+/// @param pipe_ctx
+/// @param spl_in
+void translate_SPL_in_params_from_pipe_ctx(struct pipe_ctx *pipe_ctx, struct spl_in *spl_in)
+{
+	const struct dc_plane_state *plane_state = pipe_ctx->plane_state;
+	const struct dc_stream_state *stream = pipe_ctx->stream;
+	struct rect odm_slice_src = resource_get_odm_slice_src_rect(pipe_ctx);
+
+	// Assign the function to calculate the number of partitions in the line buffer
+	// This is used to determine the vtap support
+	switch (plane_state->ctx->dce_version)	{
+	case DCN_VERSION_2_0:
+		spl_in->callbacks = dcn2_spl_callbacks;
+		break;
+	case DCN_VERSION_3_2:
+		spl_in->callbacks = dcn32_spl_callbacks;
+		break;
+	case DCN_VERSION_4_01:
+		spl_in->callbacks = dcn401_spl_callbacks;
+		break;
+	default:
+		spl_in->callbacks = dcn2_spl_callbacks;
+	}
+	// Make format field from spl_in point to plane_res scl_data format
+	populate_splformat_from_format(&spl_in->basic_in.format, pipe_ctx->plane_res.scl_data.format);
+	// Make view_format from basic_out point to view_format from stream
+	spl_in->basic_out.view_format = (enum spl_view_3d)stream->view_format;
+	// Populate spl input basic input clip rect from plane state clip rect
+	populate_splrect_from_rect(&spl_in->basic_in.clip_rect, &plane_state->clip_rect);
+	// Populate spl input basic out src rect from stream src rect
+	populate_splrect_from_rect(&spl_in->basic_out.src_rect, &stream->src);
+	// Populate spl input basic out dst rect from stream dst rect
+	populate_splrect_from_rect(&spl_in->basic_out.dst_rect, &stream->dst);
+	// Make spl input basic input info rotation field point to plane state rotation
+	spl_in->basic_in.rotation = (enum spl_rotation_angle)plane_state->rotation;
+	// Populate spl input basic input src rect from plane state src rect
+	populate_splrect_from_rect(&spl_in->basic_in.src_rect, &plane_state->src_rect);
+	// Populate spl input basic input dst rect from plane state dst rect
+	populate_splrect_from_rect(&spl_in->basic_in.dst_rect, &plane_state->dst_rect);
+	// Make spl input basic input info horiz mirror field point to plane state horz mirror
+	spl_in->basic_in.horizontal_mirror = plane_state->horizontal_mirror;
+
+	// Calculate horizontal splits and split index
+	spl_in->basic_in.num_h_slices_recout_width_align.use_recout_width_aligned = false;
+	spl_in->basic_in.num_h_slices_recout_width_align.num_slices_recout_width.mpc_num_h_slices =
+		resource_get_mpc_slice_count(pipe_ctx);
+
+	if (stream->view_format == VIEW_3D_FORMAT_SIDE_BY_SIDE)
+		spl_in->basic_in.mpc_h_slice_index = 0;
+	else
+		spl_in->basic_in.mpc_h_slice_index = resource_get_mpc_slice_index(pipe_ctx);
+
+	populate_splrect_from_rect(&spl_in->basic_out.odm_slice_rect, &odm_slice_src);
+	spl_in->basic_out.odm_combine_factor = 0;
+	spl_in->odm_slice_index = resource_get_odm_slice_index(pipe_ctx);
+	// Make spl input basic out info output_size width point to stream h active
+	spl_in->basic_out.output_size.width =
+		stream->timing.h_addressable + stream->timing.h_border_left + stream->timing.h_border_right + pipe_ctx->dsc_padding_params.dsc_hactive_padding;
+	// Make spl input basic out info output_size height point to v active
+	spl_in->basic_out.output_size.height =
+		stream->timing.v_addressable + stream->timing.v_border_bottom + stream->timing.v_border_top;
+	spl_in->basic_out.max_downscale_src_width =
+			pipe_ctx->stream->ctx->dc->debug.max_downscale_src_width;
+	spl_in->basic_out.always_scale = pipe_ctx->stream->ctx->dc->debug.always_scale;
+	// Make spl input basic output info alpha_en field point to plane res scl_data lb_params alpha_en
+	spl_in->basic_out.alpha_en = pipe_ctx->plane_res.scl_data.lb_params.alpha_en;
+	spl_in->basic_out.use_two_pixels_per_container = pipe_ctx->stream_res.tg->funcs->is_two_pixels_per_container(&stream->timing);
+	// Make spl input basic input info scaling quality field point to plane state scaling_quality
+	populate_spltaps_from_taps(&spl_in->scaling_quality, &plane_state->scaling_quality);
+	// Translate edge adaptive scaler preference
+	spl_in->prefer_easf = pipe_ctx->stream->ctx->dc->config.prefer_easf;
+	spl_in->disable_easf = false;
+	if (pipe_ctx->stream->ctx->dc->debug.force_easf == 1)
+		spl_in->prefer_easf = false;
+	else if (pipe_ctx->stream->ctx->dc->debug.force_easf == 2)
+		spl_in->disable_easf = true;
+	else if (pipe_ctx->stream->ctx->dc->debug.force_easf == 3)
+		spl_in->override_easf = true;
+	/* Translate adaptive sharpening preference */
+	unsigned int sharpness_setting = pipe_ctx->stream->ctx->dc->debug.force_sharpness;
+	unsigned int force_sharpness_level = pipe_ctx->stream->ctx->dc->debug.force_sharpness_level;
+	if (sharpness_setting == SHARPNESS_HW_OFF)
+		spl_in->adaptive_sharpness.enable = false;
+	else if (sharpness_setting == SHARPNESS_ZERO) {
+		spl_in->adaptive_sharpness.enable = true;
+		spl_in->adaptive_sharpness.sharpness_level = 0;
+	} else if (sharpness_setting == SHARPNESS_CUSTOM) {
+		/* SAT: read harpness_range from dc_plane_state */
+		spl_in->adaptive_sharpness.sharpness_range.sdr_rgb_min = plane_state->sharpness_range.sdr_rgb_min;
+		spl_in->adaptive_sharpness.sharpness_range.sdr_rgb_max = plane_state->sharpness_range.sdr_rgb_max;
+		spl_in->adaptive_sharpness.sharpness_range.sdr_rgb_mid = plane_state->sharpness_range.sdr_rgb_mid;
+		spl_in->adaptive_sharpness.sharpness_range.sdr_yuv_min = plane_state->sharpness_range.sdr_yuv_min;
+		spl_in->adaptive_sharpness.sharpness_range.sdr_yuv_max = plane_state->sharpness_range.sdr_yuv_max;
+		spl_in->adaptive_sharpness.sharpness_range.sdr_yuv_mid = plane_state->sharpness_range.sdr_yuv_mid;
+		spl_in->adaptive_sharpness.sharpness_range.hdr_rgb_min = plane_state->sharpness_range.hdr_rgb_min;
+		spl_in->adaptive_sharpness.sharpness_range.hdr_rgb_max = plane_state->sharpness_range.hdr_rgb_max;
+		spl_in->adaptive_sharpness.sharpness_range.hdr_rgb_mid = plane_state->sharpness_range.hdr_rgb_mid;
+
+		if (force_sharpness_level > 0) {
+			if (force_sharpness_level > 10)
+				force_sharpness_level = 10;
+			spl_in->adaptive_sharpness.enable = true;
+			spl_in->adaptive_sharpness.sharpness_level = force_sharpness_level;
+		} else if (!plane_state->adaptive_sharpness_en) {
+			spl_in->adaptive_sharpness.enable = false;
+			spl_in->adaptive_sharpness.sharpness_level = 0;
+		} else {
+			spl_in->adaptive_sharpness.enable = true;
+			spl_in->adaptive_sharpness.sharpness_level = plane_state->sharpness_level;
+		}
+	}
+	// Translate linear light scaling preference
+	if (pipe_ctx->stream->ctx->dc->debug.force_lls > 0)
+		spl_in->lls_pref = pipe_ctx->stream->ctx->dc->debug.force_lls;
+	else
+		spl_in->lls_pref = plane_state->linear_light_scaling;
+	/* Translate chroma subsampling offset ( cositing ) */
+	if (pipe_ctx->stream->ctx->dc->debug.force_cositing)
+		spl_in->basic_in.cositing = pipe_ctx->stream->ctx->dc->debug.force_cositing - 1;
+	else
+		spl_in->basic_in.cositing = plane_state->cositing;
+	/* Translate transfer function */
+	spl_in->basic_in.tf_type = (enum spl_transfer_func_type) plane_state->in_transfer_func.type;
+	spl_in->basic_in.tf_predefined_type = (enum spl_transfer_func_predefined) plane_state->in_transfer_func.tf;
+
+	spl_in->h_active = pipe_ctx->plane_res.scl_data.h_active;
+	spl_in->v_active = pipe_ctx->plane_res.scl_data.v_active;
+
+	spl_in->sharpen_policy = (enum sharpen_policy)plane_state->adaptive_sharpness_policy;
+	spl_in->debug.scale_to_sharpness_policy =
+		(enum scale_to_sharpness_policy)pipe_ctx->stream->ctx->dc->debug.scale_to_sharpness_policy;
+
+	/* Check if it is stream is in fullscreen and if its HDR.
+	 * Use this to determine sharpness levels
+	 */
+	spl_in->is_fullscreen = pipe_ctx->stream->sharpening_required;
+	spl_in->is_hdr_on = dm_helpers_is_hdr_on(pipe_ctx->stream->ctx, pipe_ctx->stream);
+	spl_in->sdr_white_level_nits = plane_state->sdr_white_level_nits;
+}
+
+/// @brief Translate SPL output parameters to pipe context
+/// @param pipe_ctx
+/// @param spl_out
+void translate_SPL_out_params_to_pipe_ctx(struct pipe_ctx *pipe_ctx, struct spl_out *spl_out)
+{
+	// Make scaler data recout point to spl output field recout
+	populate_rect_from_splrect(&pipe_ctx->plane_res.scl_data.recout, &spl_out->dscl_prog_data->recout);
+	// Make scaler data ratios point to spl output field ratios
+	populate_ratios_from_splratios(&pipe_ctx->plane_res.scl_data.ratios, &spl_out->dscl_prog_data->ratios);
+	// Make scaler data viewport point to spl output field viewport
+	populate_rect_from_splrect(&pipe_ctx->plane_res.scl_data.viewport, &spl_out->dscl_prog_data->viewport);
+	// Make scaler data viewport_c point to spl output field viewport_c
+	populate_rect_from_splrect(&pipe_ctx->plane_res.scl_data.viewport_c, &spl_out->dscl_prog_data->viewport_c);
+	// Make scaler data taps point to spl output field scaling taps
+	populate_taps_from_spltaps(&pipe_ctx->plane_res.scl_data.taps, &spl_out->dscl_prog_data->taps);
+	// Make scaler data init point to spl output field init
+	populate_inits_from_splinits(&pipe_ctx->plane_res.scl_data.inits, &spl_out->dscl_prog_data->init);
+}
diff --git a/drivers/gpu/drm/amd/display/dc/dc_spl_translate.h b/drivers/gpu/drm/amd/display/dc/dc_spl_translate.h
new file mode 100644
index 000000000000..eaa5c5373b28
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dc_spl_translate.h
@@ -0,0 +1,23 @@
+// SPDX-License-Identifier: MIT
+//
+// Copyright 2024 Advanced Micro Devices, Inc.
+
+#ifndef __DC_SPL_TRANSLATE_H__
+#define __DC_SPL_TRANSLATE_H__
+#include "dc.h"
+#include "resource.h"
+#include "dm_helpers.h"
+
+/* Map SPL input parameters to pipe context
+ * @pipe_ctx: pipe context
+ * @spl_in: spl input structure
+ */
+void translate_SPL_in_params_from_pipe_ctx(struct pipe_ctx *pipe_ctx, struct spl_in *spl_in);
+
+/* Map SPL output parameters to pipe context
+ * @pipe_ctx: pipe context
+ * @spl_out: spl output structure
+ */
+void translate_SPL_out_params_to_pipe_ctx(struct pipe_ctx *pipe_ctx, struct spl_out *spl_out);
+
+#endif /* __DC_SPL_TRANSLATE_H__ */
diff --git a/drivers/gpu/drm/amd/display/dc/dc_state.h b/drivers/gpu/drm/amd/display/dc/dc_state.h
new file mode 100644
index 000000000000..db1e63a7d460
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dc_state.h
@@ -0,0 +1,77 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef _DC_STATE_H_
+#define _DC_STATE_H_
+
+#include "inc/core_status.h"
+
+struct dc_state *dc_state_create(struct dc *dc, struct dc_state_create_params *params);
+void dc_state_copy(struct dc_state *dst_state, struct dc_state *src_state);
+struct dc_state *dc_state_create_copy(struct dc_state *src_state);
+void dc_state_copy_current(struct dc *dc, struct dc_state *dst_state);
+struct dc_state *dc_state_create_current_copy(struct dc *dc);
+void dc_state_construct(struct dc *dc, struct dc_state *state);
+void dc_state_destruct(struct dc_state *state);
+void dc_state_retain(struct dc_state *state);
+void dc_state_release(struct dc_state *state);
+
+enum dc_status dc_state_add_stream(const struct dc *dc,
+				    struct dc_state *state,
+				    struct dc_stream_state *stream);
+
+enum dc_status dc_state_remove_stream(
+		const struct dc *dc,
+		struct dc_state *state,
+		struct dc_stream_state *stream);
+
+bool dc_state_add_plane(
+		const struct dc *dc,
+		struct dc_stream_state *stream,
+		struct dc_plane_state *plane_state,
+		struct dc_state *state);
+
+bool dc_state_remove_plane(
+		const struct dc *dc,
+		struct dc_stream_state *stream,
+		struct dc_plane_state *plane_state,
+		struct dc_state *state);
+
+bool dc_state_rem_all_planes_for_stream(
+		const struct dc *dc,
+		struct dc_stream_state *stream,
+		struct dc_state *state);
+
+bool dc_state_add_all_planes_for_stream(
+		const struct dc *dc,
+		struct dc_stream_state *stream,
+		struct dc_plane_state * const *plane_states,
+		int plane_count,
+		struct dc_state *state);
+
+struct dc_stream_status *dc_state_get_stream_status(
+	struct dc_state *state,
+	const struct dc_stream_state *stream);
+#endif /* _DC_STATE_H_ */
diff --git a/drivers/gpu/drm/amd/display/dc/dc_state_priv.h b/drivers/gpu/drm/amd/display/dc/dc_state_priv.h
new file mode 100644
index 000000000000..1d9bae56ff6a
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dc_state_priv.h
@@ -0,0 +1,128 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef _DC_STATE_PRIV_H_
+#define _DC_STATE_PRIV_H_
+
+#include "dc_state.h"
+#include "dc_stream.h"
+
+struct dc_stream_state *dc_state_get_stream_from_id(const struct dc_state *state, unsigned int id);
+
+/* Get the type of the provided resource (none, phantom, main) based on the provided
+ * context. If the context is unavailable, determine only if phantom or not.
+ */
+enum mall_stream_type dc_state_get_pipe_subvp_type(const struct dc_state *state,
+		const struct pipe_ctx *pipe_ctx);
+enum mall_stream_type dc_state_get_stream_subvp_type(const struct dc_state *state,
+		const struct dc_stream_state *stream);
+
+/* Gets the phantom stream if main is provided, gets the main if phantom is provided.*/
+struct dc_stream_state *dc_state_get_paired_subvp_stream(const struct dc_state *state,
+		const struct dc_stream_state *stream);
+
+/* allocate's phantom stream or plane and returns pointer to the object */
+struct dc_stream_state *dc_state_create_phantom_stream(const struct dc *dc,
+		struct dc_state *state,
+		struct dc_stream_state *main_stream);
+struct dc_plane_state *dc_state_create_phantom_plane(const struct dc *dc,
+		struct dc_state *state,
+		struct dc_plane_state *main_plane);
+
+/* deallocate's phantom stream or plane */
+void dc_state_release_phantom_stream(const struct dc *dc,
+		struct dc_state *state,
+		struct dc_stream_state *phantom_stream);
+void dc_state_release_phantom_plane(const struct dc *dc,
+		struct dc_state *state,
+		struct dc_plane_state *phantom_plane);
+
+/* add/remove phantom stream to context and generate subvp meta data */
+enum dc_status dc_state_add_phantom_stream(const struct dc *dc,
+		struct dc_state *state,
+		struct dc_stream_state *phantom_stream,
+		struct dc_stream_state *main_stream);
+enum dc_status dc_state_remove_phantom_stream(const struct dc *dc,
+		struct dc_state *state,
+		struct dc_stream_state *phantom_stream);
+
+bool dc_state_add_phantom_plane(
+		const struct dc *dc,
+		struct dc_stream_state *phantom_stream,
+		struct dc_plane_state *phantom_plane,
+		struct dc_state *state);
+
+bool dc_state_remove_phantom_plane(
+		const struct dc *dc,
+		struct dc_stream_state *phantom_stream,
+		struct dc_plane_state *phantom_plane,
+		struct dc_state *state);
+
+bool dc_state_rem_all_phantom_planes_for_stream(
+		const struct dc *dc,
+		struct dc_stream_state *phantom_stream,
+		struct dc_state *state,
+		bool should_release_planes);
+
+bool dc_state_add_all_phantom_planes_for_stream(
+		const struct dc *dc,
+		struct dc_stream_state *phantom_stream,
+		struct dc_plane_state * const *phantom_planes,
+		int plane_count,
+		struct dc_state *state);
+
+bool dc_state_remove_phantom_streams_and_planes(
+		const struct dc *dc,
+		struct dc_state *state);
+
+void dc_state_release_phantom_streams_and_planes(
+		const struct dc *dc,
+		struct dc_state *state);
+
+bool dc_state_is_fams2_in_use(
+		const struct dc *dc,
+		const struct dc_state *state);
+
+
+void dc_state_set_stream_subvp_cursor_limit(const struct dc_stream_state *stream,
+		struct dc_state *state,
+		bool limit);
+
+bool dc_state_get_stream_subvp_cursor_limit(const struct dc_stream_state *stream,
+		struct dc_state *state);
+
+void dc_state_set_stream_cursor_subvp_limit(const struct dc_stream_state *stream,
+		struct dc_state *state,
+		bool limit);
+
+bool dc_state_get_stream_cursor_subvp_limit(const struct dc_stream_state *stream,
+		struct dc_state *state);
+
+bool dc_state_can_clear_stream_cursor_subvp_limit(const struct dc_stream_state *stream,
+		struct dc_state *state);
+
+bool dc_state_is_subvp_in_use(struct dc_state *state);
+
+#endif /* _DC_STATE_PRIV_H_ */
diff --git a/drivers/gpu/drm/amd/display/dc/dc_stream.h b/drivers/gpu/drm/amd/display/dc/dc_stream.h
index e37c4a10bfd5..321cfe92d799 100644
--- a/drivers/gpu/drm/amd/display/dc/dc_stream.h
+++ b/drivers/gpu/drm/amd/display/dc/dc_stream.h
@@ -38,19 +38,30 @@ struct timing_sync_info {
 	bool master;
 };
 
+struct mall_stream_config {
+	/* MALL stream config to indicate if the stream is phantom or not.
+	 * We will use a phantom stream to indicate that the pipe is phantom.
+	 */
+	enum mall_stream_type type;
+	struct dc_stream_state *paired_stream;	// master / slave stream
+	bool subvp_limit_cursor_size; /* stream has/is using subvp limiting hw cursor support */
+	bool cursor_size_limit_subvp; /* stream is using hw cursor config preventing subvp */
+};
+
 struct dc_stream_status {
 	int primary_otg_inst;
 	int stream_enc_inst;
+
+	/**
+	 * @plane_count: Total of planes attached to a single stream
+	 */
 	int plane_count;
 	int audio_inst;
 	struct timing_sync_info timing_sync_info;
-	struct dc_plane_state *plane_states[MAX_SURFACE_NUM];
+	struct dc_plane_state *plane_states[MAX_SURFACES];
 	bool is_abm_supported;
-};
-
-// TODO: References to this needs to be removed..
-struct freesync_context {
-	bool dummy;
+	struct mall_stream_config mall_stream_config;
+	bool fpo_in_use;
 };
 
 enum hubp_dmdata_mode {
@@ -115,12 +126,10 @@ struct periodic_interrupt_config {
 	int lines_offset;
 };
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 struct dc_mst_stream_bw_update {
 	bool is_increase; // is bandwidth reduced or increased
 	uint32_t mst_stream_bw; // new mst bandwidth in kbps
 };
-#endif
 
 union stream_update_flags {
 	struct {
@@ -132,9 +141,11 @@ union stream_update_flags {
 		uint32_t gamut_remap:1;
 		uint32_t wb_update:1;
 		uint32_t dsc_changed : 1;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 		uint32_t mst_bw : 1;
-#endif
+		uint32_t crtc_timing_adjust : 1;
+		uint32_t fams_changed : 1;
+		uint32_t scaler_sharpener : 1;
+		uint32_t sharpening_required : 1;
 	} bits;
 
 	uint32_t raw;
@@ -148,6 +159,30 @@ struct test_pattern {
 	unsigned int cust_pattern_size;
 };
 
+#define SUBVP_DRR_MARGIN_US 100 // 100us for DRR margin (SubVP + DRR)
+
+struct dc_stream_debug_options {
+	char force_odm_combine_segments;
+	/*
+	 * When force_odm_combine_segments is non zero, allow dc to
+	 * temporarily transition to ODM bypass when minimal transition state
+	 * is required to prevent visual glitches showing on the screen
+	 */
+	char allow_transition_for_forced_odm;
+};
+
+#define LUMINANCE_DATA_TABLE_SIZE 10
+
+struct luminance_data {
+	bool is_valid;
+	int refresh_rate_hz[LUMINANCE_DATA_TABLE_SIZE];
+	int luminance_millinits[LUMINANCE_DATA_TABLE_SIZE];
+	int flicker_criteria_milli_nits_GAMING;
+	int flicker_criteria_milli_nits_STATIC;
+	int nominal_refresh_rate;
+	int dm_max_decrease_from_nominal;
+};
+
 struct dc_stream_state {
 	// sink is deprecated, new code should not reference
 	// this pointer
@@ -158,31 +193,33 @@ struct dc_stream_state {
 	 * a stream via the volatile dc_state rather than the static dc_link.
 	 */
 	struct link_encoder *link_enc;
+	struct dc_stream_debug_options debug;
 	struct dc_panel_patch sink_patches;
-	union display_content_support content_support;
 	struct dc_crtc_timing timing;
 	struct dc_crtc_timing_adjust adjust;
 	struct dc_info_packet vrr_infopacket;
 	struct dc_info_packet vsc_infopacket;
 	struct dc_info_packet vsp_infopacket;
-
+	struct dc_info_packet hfvsif_infopacket;
+	struct dc_info_packet vtem_infopacket;
+	struct dc_info_packet adaptive_sync_infopacket;
+	struct dc_info_packet avi_infopacket;
+	uint8_t dsc_packed_pps[128];
 	struct rect src; /* composition area */
 	struct rect dst; /* stream addressable area */
 
-	// TODO: References to this needs to be removed..
-	struct freesync_context freesync_ctx;
-
 	struct audio_info audio_info;
 
 	struct dc_info_packet hdr_static_metadata;
 	PHYSICAL_ADDRESS_LOC dmdata_address;
 	bool   use_dynamic_meta;
 
-	struct dc_transfer_func *out_transfer_func;
+	struct dc_transfer_func out_transfer_func;
 	struct colorspace_transform gamut_remap_matrix;
 	struct dc_csc_transform csc_color_matrix;
 
 	enum dc_color_space output_color_space;
+	enum display_content_type content_type;
 	enum dc_dither_option dither_option;
 
 	enum view_3d_format view_format;
@@ -190,7 +227,21 @@ struct dc_stream_state {
 	bool use_vsc_sdp_for_colorimetry;
 	bool ignore_msa_timing_param;
 
+	/**
+	 * @allow_freesync:
+	 *
+	 * It say if Freesync is enabled or not.
+	 */
+	bool allow_freesync;
+
+	/**
+	 * @vrr_active_variable:
+	 *
+	 * It describes if VRR is in use.
+	 */
+	bool vrr_active_variable;
 	bool freesync_on_desktop;
+	bool vrr_active_fixed;
 
 	bool converter_disable_audio;
 	uint8_t qs_bit;
@@ -203,8 +254,7 @@ struct dc_stream_state {
 	/* DMCU info */
 	unsigned int abm_level;
 
-	struct periodic_interrupt_config periodic_interrupt0;
-	struct periodic_interrupt_config periodic_interrupt1;
+	struct periodic_interrupt_config periodic_interrupt;
 
 	/* from core_stream struct */
 	struct dc_context *ctx;
@@ -221,6 +271,8 @@ struct dc_stream_state {
 
 	struct dc_cursor_attributes cursor_attributes;
 	struct dc_cursor_position cursor_position;
+	bool hw_cursor_req;
+
 	uint32_t sdr_white_level; // for boosting (SDR) cursor in HDR mode
 
 	/* from stream struct */
@@ -249,6 +301,7 @@ struct dc_stream_state {
 
 	bool apply_edp_fast_boot_optimization;
 	bool apply_seamless_boot_optimization;
+	uint32_t apply_boot_odm_mode;
 
 	uint32_t stream_id;
 
@@ -257,6 +310,11 @@ struct dc_stream_state {
 
 	bool has_non_synchronizable_pclk;
 	bool vblank_synchronized;
+	bool is_phantom;
+
+	struct luminance_data lumin_data;
+	bool scaler_sharpener_update;
+	bool sharpening_required;
 };
 
 #define ABM_LEVEL_IMMEDIATE_DISABLE 255
@@ -270,15 +328,21 @@ struct dc_stream_update {
 	struct dc_info_packet *hdr_static_metadata;
 	unsigned int *abm_level;
 
-	struct periodic_interrupt_config *periodic_interrupt0;
-	struct periodic_interrupt_config *periodic_interrupt1;
+	struct periodic_interrupt_config *periodic_interrupt;
 
 	struct dc_info_packet *vrr_infopacket;
 	struct dc_info_packet *vsc_infopacket;
 	struct dc_info_packet *vsp_infopacket;
+	struct dc_info_packet *hfvsif_infopacket;
+	struct dc_info_packet *vtem_infopacket;
+	struct dc_info_packet *adaptive_sync_infopacket;
+	struct dc_info_packet *avi_infopacket;
 
 	bool *dpms_off;
 	bool integer_scaling_update;
+	bool *allow_freesync;
+	bool *vrr_active_variable;
+	bool *vrr_active_fixed;
 
 	struct colorspace_transform *gamut_remap;
 	enum dc_color_space *output_color_space;
@@ -288,13 +352,18 @@ struct dc_stream_update {
 
 	struct dc_writeback_update *wb_update;
 	struct dc_dsc_config *dsc_config;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	struct dc_mst_stream_bw_update *mst_bw_update;
-#endif
 	struct dc_transfer_func *func_shaper;
 	struct dc_3dlut *lut3d_func;
 
 	struct test_pattern *pending_test_pattern;
+	struct dc_crtc_timing_adjust *crtc_timing_adjust;
+
+	struct dc_cursor_attributes *cursor_attributes;
+	struct dc_cursor_position *cursor_position;
+	bool *hw_cursor_req;
+	bool *scaler_sharpener_update;
+	bool *sharpening_required;
 };
 
 bool dc_is_stream_unchanged(
@@ -303,6 +372,9 @@ bool dc_is_stream_scaling_unchanged(
 	struct dc_stream_state *old_stream, struct dc_stream_state *stream);
 
 /*
+ * Setup stream attributes if no stream updates are provided
+ * there will be no impact on the stream parameters
+ *
  * Set up surface attributes and associate to a stream
  * The surfaces parameter is an absolute set of all surface active for the stream.
  * If no surfaces are provided, the stream will be blanked; no memory read.
@@ -311,8 +383,23 @@ bool dc_is_stream_scaling_unchanged(
  * After this call:
  *   Surfaces attributes are programmed and configured to be composed into stream.
  *   This does not trigger a flip.  No surface address is programmed.
+ *
  */
+bool dc_update_planes_and_stream(struct dc *dc,
+		struct dc_surface_update *surface_updates, int surface_count,
+		struct dc_stream_state *dc_stream,
+		struct dc_stream_update *stream_update);
 
+/*
+ * Set up surface attributes and associate to a stream
+ * The surfaces parameter is an absolute set of all surface active for the stream.
+ * If no surfaces are provided, the stream will be blanked; no memory read.
+ * Any flip related attribute changes must be done through this interface.
+ *
+ * After this call:
+ *   Surfaces attributes are programmed and configured to be composed into stream.
+ *   This does not trigger a flip.  No surface address is programmed.
+ */
 void dc_commit_updates_for_stream(struct dc *dc,
 		struct dc_surface_update *srf_updates,
 		int surface_count,
@@ -326,7 +413,6 @@ void dc_stream_log(const struct dc *dc, const struct dc_stream_state *stream);
 
 uint8_t dc_get_current_stream_count(struct dc *dc);
 struct dc_stream_state *dc_get_stream_at_index(struct dc *dc, uint8_t i);
-struct dc_stream_state *dc_stream_find_from_link(const struct dc_link *link);
 
 /*
  * Return the current frame counter.
@@ -350,45 +436,14 @@ bool dc_stream_get_scanoutpos(const struct dc_stream_state *stream,
 				  uint32_t *h_position,
 				  uint32_t *v_position);
 
-enum dc_status dc_add_stream_to_ctx(
-			struct dc *dc,
-		struct dc_state *new_ctx,
-		struct dc_stream_state *stream);
-
-enum dc_status dc_remove_stream_from_ctx(
-		struct dc *dc,
-			struct dc_state *new_ctx,
-			struct dc_stream_state *stream);
-
-
-bool dc_add_plane_to_context(
-		const struct dc *dc,
-		struct dc_stream_state *stream,
-		struct dc_plane_state *plane_state,
-		struct dc_state *context);
-
-bool dc_remove_plane_from_context(
-		const struct dc *dc,
-		struct dc_stream_state *stream,
-		struct dc_plane_state *plane_state,
-		struct dc_state *context);
-
-bool dc_rem_all_planes_for_stream(
-		const struct dc *dc,
-		struct dc_stream_state *stream,
-		struct dc_state *context);
-
-bool dc_add_all_planes_for_stream(
-		const struct dc *dc,
-		struct dc_stream_state *stream,
-		struct dc_plane_state * const *plane_states,
-		int plane_count,
-		struct dc_state *context);
-
 bool dc_stream_add_writeback(struct dc *dc,
 		struct dc_stream_state *stream,
 		struct dc_writeback_info *wb_info);
 
+bool dc_stream_fc_disable_writeback(struct dc *dc,
+		struct dc_stream_state *stream,
+		uint32_t dwb_pipe_inst);
+
 bool dc_stream_remove_writeback(struct dc *dc,
 		struct dc_stream_state *stream,
 		uint32_t dwb_pipe_inst);
@@ -397,10 +452,6 @@ enum dc_status dc_stream_add_dsc_to_resource(struct dc *dc,
 		struct dc_state *state,
 		struct dc_stream_state *stream);
 
-bool dc_stream_warmup_writeback(struct dc *dc,
-		int num_dwb,
-		struct dc_writeback_info *wb_info);
-
 bool dc_stream_dmdata_status_done(struct dc *dc, struct dc_stream_state *stream);
 
 bool dc_stream_set_dynamic_metadata(struct dc *dc,
@@ -410,14 +461,6 @@ bool dc_stream_set_dynamic_metadata(struct dc *dc,
 enum dc_status dc_validate_stream(struct dc *dc, struct dc_stream_state *stream);
 
 /*
- * Set up streams and links associated to drive sinks
- * The streams parameter is an absolute set of all active streams.
- *
- * After this call:
- *   Phy, Encoder, Timing Generator are programmed and enabled.
- *   New streams are enabled with blank stream; no memory read.
- */
-/*
  * Enable stereo when commit_streams is not required,
  * for example, frame alternate.
  */
@@ -430,12 +473,11 @@ void dc_enable_stereo(
 /* Triggers multi-stream synchronization. */
 void dc_trigger_sync(struct dc *dc, struct dc_state *context);
 
-enum surface_update_type dc_check_update_surfaces_for_stream(
-		struct dc *dc,
+struct surface_update_descriptor dc_check_update_surfaces_for_stream(
+		const struct dc_check_config *check_config,
 		struct dc_surface_update *updates,
 		int surface_count,
-		struct dc_stream_update *stream_update,
-		const struct dc_stream_status *stream_status);
+		struct dc_stream_update *stream_update);
 
 /**
  * Create a new default stream for the requested sink
@@ -449,30 +491,43 @@ void update_stream_signal(struct dc_stream_state *stream, struct dc_sink *sink);
 void dc_stream_retain(struct dc_stream_state *dc_stream);
 void dc_stream_release(struct dc_stream_state *dc_stream);
 
-struct dc_stream_status *dc_stream_get_status_from_state(
-	struct dc_state *state,
-	struct dc_stream_state *stream);
-struct dc_stream_status *dc_stream_get_status(
-	struct dc_stream_state *dc_stream);
-
-#ifndef TRIM_FSFT
-bool dc_optimize_timing_for_fsft(
-	struct dc_stream_state *pStream,
-	unsigned int max_input_rate_in_khz);
-#endif
+struct dc_stream_status *dc_stream_get_status(struct dc_stream_state *dc_stream);
+const struct dc_stream_status *dc_stream_get_status_const(const struct dc_stream_state *dc_stream);
 
 /*******************************************************************************
  * Cursor interfaces - To manages the cursor within a stream
  ******************************************************************************/
 /* TODO: Deprecated once we switch to dc_set_cursor_position */
+
+void program_cursor_attributes(
+	struct dc *dc,
+	struct dc_stream_state *stream);
+
+void program_cursor_position(
+	struct dc *dc,
+	struct dc_stream_state *stream);
+
+bool dc_stream_check_cursor_attributes(
+	const struct dc_stream_state *stream,
+	struct dc_state *state,
+	const struct dc_cursor_attributes *attributes);
+
 bool dc_stream_set_cursor_attributes(
 	struct dc_stream_state *stream,
 	const struct dc_cursor_attributes *attributes);
 
+bool dc_stream_program_cursor_attributes(
+	struct dc_stream_state *stream,
+	const struct dc_cursor_attributes *attributes);
+
 bool dc_stream_set_cursor_position(
 	struct dc_stream_state *stream,
 	const struct dc_cursor_position *position);
 
+bool dc_stream_program_cursor_position(
+	struct dc_stream_state *stream,
+	const struct dc_cursor_position *position);
+
 
 bool dc_stream_adjust_vmin_vmax(struct dc *dc,
 				struct dc_stream_state *stream,
@@ -482,27 +537,29 @@ bool dc_stream_get_last_used_drr_vtotal(struct dc *dc,
 		struct dc_stream_state *stream,
 		uint32_t *refresh_rate);
 
-bool dc_stream_get_crtc_position(struct dc *dc,
-				 struct dc_stream_state **stream,
-				 int num_streams,
-				 unsigned int *v_pos,
-				 unsigned int *nom_v_pos);
-
 #if defined(CONFIG_DRM_AMD_SECURE_DISPLAY)
-bool dc_stream_forward_dmcu_crc_window(struct dc *dc, struct dc_stream_state *stream,
-			     struct crc_params *crc_window);
-bool dc_stream_stop_dmcu_crc_win_update(struct dc *dc,
-				 struct dc_stream_state *stream);
+bool dc_stream_forward_crc_window(struct dc_stream_state *stream,
+		struct rect *rect,
+		uint8_t phy_id,
+		bool is_stop);
+
+bool dc_stream_forward_multiple_crc_window(struct dc_stream_state *stream,
+		struct crc_window *window,
+		uint8_t phy_id,
+		bool stop);
 #endif
 
 bool dc_stream_configure_crc(struct dc *dc,
 			     struct dc_stream_state *stream,
 			     struct crc_params *crc_window,
 			     bool enable,
-			     bool continuous);
+			     bool continuous,
+			     uint8_t idx,
+			     bool reset);
 
 bool dc_stream_get_crc(struct dc *dc,
 		       struct dc_stream_state *stream,
+		       uint8_t idx,
 		       uint32_t *r_cr,
 		       uint32_t *g_y,
 		       uint32_t *b_cb);
@@ -524,10 +581,26 @@ bool dc_stream_set_gamut_remap(struct dc *dc,
 bool dc_stream_program_csc_matrix(struct dc *dc,
 				  struct dc_stream_state *stream);
 
-bool dc_stream_get_crtc_position(struct dc *dc,
-				 struct dc_stream_state **stream,
-				 int num_streams,
-				 unsigned int *v_pos,
-				 unsigned int *nom_v_pos);
+struct dc_rmcm_3dlut *dc_stream_get_3dlut_for_stream(
+	const struct dc *dc,
+	const struct dc_stream_state *stream,
+	bool allocate_one);
+
+void dc_stream_release_3dlut_for_stream(
+	const struct dc *dc,
+	const struct dc_stream_state *stream);
+
+void dc_stream_init_rmcm_3dlut(struct dc *dc);
+
+struct pipe_ctx *dc_stream_get_pipe_ctx(struct dc_stream_state *stream);
+
+void dc_dmub_update_dirty_rect(struct dc *dc,
+			       int surface_count,
+			       struct dc_stream_state *stream,
+			       struct dc_surface_update *srf_updates,
+			       struct dc_state *context);
+
+bool dc_stream_is_cursor_limit_pending(struct dc *dc, struct dc_stream_state *stream);
+bool dc_stream_can_clear_cursor_limit(struct dc *dc, struct dc_stream_state *stream);
 
 #endif /* DC_STREAM_H_ */
diff --git a/drivers/gpu/drm/amd/display/dc/dc_stream_priv.h b/drivers/gpu/drm/amd/display/dc/dc_stream_priv.h
new file mode 100644
index 000000000000..ca37eac20986
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dc_stream_priv.h
@@ -0,0 +1,75 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef _DC_STREAM_PRIV_H_
+#define _DC_STREAM_PRIV_H_
+
+#include "dc_stream.h"
+
+bool dc_stream_construct(struct dc_stream_state *stream,
+	struct dc_sink *dc_sink_data);
+void dc_stream_destruct(struct dc_stream_state *stream);
+
+void dc_stream_assign_stream_id(struct dc_stream_state *stream);
+
+/*
+ * Finds the highest refresh rate that can be achieved
+ * from starting_freq while staying within flicker criteria
+ */
+int dc_stream_calculate_max_flickerless_refresh_rate(struct dc_stream_state *stream,
+						      int starting_refresh_hz,
+						      bool is_gaming);
+
+/*
+ * Finds the lowest refresh rate that can be achieved
+ * from starting_freq while staying within flicker criteria
+ */
+int dc_stream_calculate_min_flickerless_refresh_rate(struct dc_stream_state *stream,
+						      int starting_refresh_hz,
+						      bool is_gaming);
+
+/*
+ * Determines if there will be a flicker when moving between 2 refresh rates
+ */
+bool dc_stream_is_refresh_rate_range_flickerless(struct dc_stream_state *stream,
+						  int hz1,
+						  int hz2,
+						  bool is_gaming);
+
+/*
+ * Determines the max instant vtotal delta increase that can be applied without
+ * flickering for a given stream
+ */
+unsigned int dc_stream_get_max_flickerless_instant_vtotal_decrease(struct dc_stream_state *stream,
+									  bool is_gaming);
+
+/*
+ * Determines the max instant vtotal delta decrease that can be applied without
+ * flickering for a given stream
+ */
+unsigned int dc_stream_get_max_flickerless_instant_vtotal_increase(struct dc_stream_state *stream,
+									  bool is_gaming);
+
+#endif // _DC_STREAM_PRIV_H_
diff --git a/drivers/gpu/drm/amd/display/dc/dc_trace.h b/drivers/gpu/drm/amd/display/dc/dc_trace.h
index c711797e5c9e..bbec308a3a5e 100644
--- a/drivers/gpu/drm/amd/display/dc/dc_trace.h
+++ b/drivers/gpu/drm/amd/display/dc/dc_trace.h
@@ -40,3 +40,5 @@
 
 #define TRACE_DCN_FPU(begin, function, line, ref_count) \
 	trace_dcn_fpu(begin, function, line, ref_count)
+#define TRACE_OPTC_LOCK_UNLOCK_STATE(optc, inst, lock) \
+	trace_dcn_optc_lock_unlock_state(optc, inst, lock, __func__, __LINE__)
diff --git a/drivers/gpu/drm/amd/display/dc/dc_types.h b/drivers/gpu/drm/amd/display/dc/dc_types.h
index 388457ffc0a8..f46039f64203 100644
--- a/drivers/gpu/drm/amd/display/dc/dc_types.h
+++ b/drivers/gpu/drm/amd/display/dc/dc_types.h
@@ -32,14 +32,15 @@
 #include "os_types.h"
 #include "fixed31_32.h"
 #include "irq_types.h"
+#include "dc_ddc_types.h"
 #include "dc_dp_types.h"
+#include "dc_hdmi_types.h"
 #include "dc_hw_types.h"
 #include "dal_types.h"
 #include "grph_object_defs.h"
+#include "grph_object_ctrl_defs.h"
 
-#ifdef CONFIG_DRM_AMD_DC_HDCP
 #include "dm_cp_psp.h"
-#endif
 
 /* forward declarations */
 struct dc_plane_state;
@@ -68,13 +69,6 @@ enum dce_environment {
 	DCE_ENV_VIRTUAL_HW
 };
 
-/* Note: use these macro definitions instead of direct comparison! */
-#define IS_FPGA_MAXIMUS_DC(dce_environment) \
-	(dce_environment == DCE_ENV_FPGA_MAXIMUS)
-
-#define IS_DIAG_DC(dce_environment) \
-	(IS_FPGA_MAXIMUS_DC(dce_environment) || (dce_environment == DCE_ENV_DIAG))
-
 struct dc_perf_trace {
 	unsigned long read_count;
 	unsigned long write_count;
@@ -82,13 +76,7 @@ struct dc_perf_trace {
 	unsigned long last_entry_write;
 };
 
-#define DC_MAX_EDID_BUFFER_SIZE 2048
-#define DC_EDID_BLOCK_SIZE 128
-#define MAX_SURFACE_NUM 4
 #define NUM_PIXEL_FORMATS 10
-#define MAX_REPEATER_CNT 8
-
-#include "dc_ddc_types.h"
 
 enum tiling_mode {
 	TILING_MODE_INVALID,
@@ -138,6 +126,7 @@ enum dc_edid_status {
 	EDID_BAD_CHECKSUM,
 	EDID_THE_SAME,
 	EDID_FALL_BACK,
+	EDID_PARTIAL_VALID,
 };
 
 enum act_return_status {
@@ -173,18 +162,6 @@ struct dc_edid {
 
 #define AUDIO_INFO_DISPLAY_NAME_SIZE_IN_CHARS 20
 
-union display_content_support {
-	unsigned int raw;
-	struct {
-		unsigned int valid_content_type :1;
-		unsigned int game_content :1;
-		unsigned int cinema_content :1;
-		unsigned int photo_content :1;
-		unsigned int graphics_content :1;
-		unsigned int reserved :27;
-	} bits;
-};
-
 struct dc_panel_patch {
 	unsigned int dppowerup_delay;
 	unsigned int extra_t12_ms;
@@ -195,9 +172,23 @@ struct dc_panel_patch {
 	unsigned int disable_fec;
 	unsigned int extra_t3_ms;
 	unsigned int max_dsc_target_bpp_limit;
+	unsigned int embedded_tiled_slave;
+	unsigned int disable_fams;
 	unsigned int skip_avmute;
-};
-
+	unsigned int skip_audio_sab_check;
+	unsigned int mst_start_top_delay;
+	unsigned int remove_sink_ext_caps;
+	unsigned int disable_colorimetry;
+	uint8_t blankstream_before_otg_off;
+	bool oled_optimize_display_on;
+	unsigned int force_mst_blocked_discovery;
+	unsigned int wait_after_dpcd_poweroff_ms;
+};
+
+/**
+ * struct dc_edid_caps - Capabilities read from EDID.
+ * @analog: Whether the monitor is analog. Used by DVI-I handling.
+ */
 struct dc_edid_caps {
 	/* sink identification */
 	uint16_t manufacturer_id;
@@ -214,8 +205,6 @@ struct dc_edid_caps {
 	uint32_t audio_latency;
 	uint32_t video_latency;
 
-	union display_content_support content_support;
-
 	uint8_t qs_bit;
 	uint8_t qy_bit;
 
@@ -226,6 +215,9 @@ struct dc_edid_caps {
 
 	bool edid_hdmi;
 	bool hdr_supported;
+	bool rr_capable;
+	bool scdc_present;
+	bool analog;
 
 	struct dc_panel_patch panel_patch;
 };
@@ -276,6 +268,9 @@ enum dc_timing_source {
 	TIMING_SOURCE_EDID_CEA_SVD,
 	TIMING_SOURCE_EDID_CVT_3BYTE,
 	TIMING_SOURCE_EDID_4BYTE,
+	TIMING_SOURCE_EDID_CEA_DISPLAYID_VTDB,
+	TIMING_SOURCE_EDID_CEA_RID,
+	TIMING_SOURCE_EDID_DISPLAYID_TYPE5,
 	TIMING_SOURCE_VBIOS,
 	TIMING_SOURCE_CV,
 	TIMING_SOURCE_TV,
@@ -358,7 +353,8 @@ enum dc_connection_type {
 	dc_connection_none,
 	dc_connection_single,
 	dc_connection_mst_branch,
-	dc_connection_sst_branch
+	dc_connection_sst_branch,
+	dc_connection_dac_load
 };
 
 struct dc_csc_adjustments {
@@ -368,71 +364,6 @@ struct dc_csc_adjustments {
 	struct fixed31_32 hue;
 };
 
-enum dpcd_downstream_port_max_bpc {
-	DOWN_STREAM_MAX_8BPC = 0,
-	DOWN_STREAM_MAX_10BPC,
-	DOWN_STREAM_MAX_12BPC,
-	DOWN_STREAM_MAX_16BPC
-};
-
-
-enum link_training_offset {
-	DPRX                = 0,
-	LTTPR_PHY_REPEATER1 = 1,
-	LTTPR_PHY_REPEATER2 = 2,
-	LTTPR_PHY_REPEATER3 = 3,
-	LTTPR_PHY_REPEATER4 = 4,
-	LTTPR_PHY_REPEATER5 = 5,
-	LTTPR_PHY_REPEATER6 = 6,
-	LTTPR_PHY_REPEATER7 = 7,
-	LTTPR_PHY_REPEATER8 = 8
-};
-
-struct dc_lttpr_caps {
-	union dpcd_rev revision;
-	uint8_t mode;
-	uint8_t max_lane_count;
-	uint8_t max_link_rate;
-	uint8_t phy_repeater_cnt;
-	uint8_t max_ext_timeout;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	union dp_main_link_channel_coding_lttpr_cap main_link_channel_coding;
-	union dp_128b_132b_supported_lttpr_link_rates supported_128b_132b_rates;
-#endif
-	uint8_t aux_rd_interval[MAX_REPEATER_CNT - 1];
-};
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-struct dc_dongle_dfp_cap_ext {
-	bool supported;
-	uint16_t max_pixel_rate_in_mps;
-	uint16_t max_video_h_active_width;
-	uint16_t max_video_v_active_height;
-	struct dp_encoding_format_caps encoding_format_caps;
-	struct dp_color_depth_caps rgb_color_depth_caps;
-	struct dp_color_depth_caps ycbcr444_color_depth_caps;
-	struct dp_color_depth_caps ycbcr422_color_depth_caps;
-	struct dp_color_depth_caps ycbcr420_color_depth_caps;
-};
-#endif
-
-struct dc_dongle_caps {
-	/* dongle type (DP converter, CV smart dongle) */
-	enum display_dongle_type dongle_type;
-	bool extendedCapValid;
-	/* If dongle_type == DISPLAY_DONGLE_DP_HDMI_CONVERTER,
-	indicates 'Frame Sequential-to-lllFrame Pack' conversion capability.*/
-	bool is_dp_hdmi_s3d_converter;
-	bool is_dp_hdmi_ycbcr422_pass_through;
-	bool is_dp_hdmi_ycbcr420_pass_through;
-	bool is_dp_hdmi_ycbcr422_converter;
-	bool is_dp_hdmi_ycbcr420_converter;
-	uint32_t dp_hdmi_max_bpc;
-	uint32_t dp_hdmi_max_pixel_clk_in_khz;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	struct dc_dongle_dfp_cap_ext dfp_cap_ext;
-#endif
-};
 /* Scaling format */
 enum scaling_transformation {
 	SCALING_TRANSFORMATION_UNINITIALIZED,
@@ -505,7 +436,7 @@ struct dc_dwb_params {
 	enum dwb_capture_rate		capture_rate;	/* controls the frame capture rate */
 	struct scaling_taps 		scaler_taps;	/* Scaling taps */
 	enum dwb_subsample_position	subsample_position;
-	struct dc_transfer_func *out_transfer_func;
+	const struct dc_transfer_func *out_transfer_func;
 };
 
 /* audio*/
@@ -639,6 +570,12 @@ struct dc_info_packet_128 {
 	uint8_t sb[128];
 };
 
+struct dc_edid_read_policy {
+	uint32_t max_retry_count;
+	uint32_t delay_time_ms;
+	uint32_t ignore_checksum;
+};
+
 #define DC_PLANE_UPDATE_TIMES_MAX 10
 
 struct dc_plane_flip_time {
@@ -647,6 +584,12 @@ struct dc_plane_flip_time {
 	unsigned int prev_update_time_in_us;
 };
 
+enum dc_alpm_mode {
+	DC_ALPM_AUXWAKE = 0,
+	DC_ALPM_AUXLESS = 1,
+	DC_ALPM_UNSUPPORTED = 0xF,
+};
+
 enum dc_psr_state {
 	PSR_STATE0 = 0x0,
 	PSR_STATE1,
@@ -661,10 +604,19 @@ enum dc_psr_state {
 	PSR_STATE4b,
 	PSR_STATE4c,
 	PSR_STATE4d,
+	PSR_STATE4_FULL_FRAME,
+	PSR_STATE4a_FULL_FRAME,
+	PSR_STATE4b_FULL_FRAME,
+	PSR_STATE4c_FULL_FRAME,
+	PSR_STATE4_FULL_FRAME_POWERUP,
+	PSR_STATE4_FULL_FRAME_HW_LOCK,
 	PSR_STATE5,
 	PSR_STATE5a,
 	PSR_STATE5b,
 	PSR_STATE5c,
+	PSR_STATE_HWLOCK_MGR,
+	PSR_STATE_POLLVUPDATE,
+	PSR_STATE_RELEASE_HWLOCK_MGR_FULL_FRAME,
 	PSR_STATE_INVALID = 0xFF
 };
 
@@ -676,6 +628,14 @@ struct psr_config {
 	unsigned int psr_sdp_transmit_line_num_deadline;
 	bool allow_smu_optimizations;
 	bool allow_multi_disp_optimizations;
+	/* Panel self refresh 2 selective update granularity required */
+	bool su_granularity_required;
+	/* psr2 selective update y granularity capability */
+	uint8_t su_y_granularity;
+	unsigned int line_time_in_us;
+	uint8_t rate_control_caps;
+	uint16_t dsc_slice_height;
+	bool os_request_force_ffu;
 };
 
 union dmcu_psr_level {
@@ -690,7 +650,9 @@ union dmcu_psr_level {
 		unsigned int SKIP_AUTO_STATE_ADVANCE:1;
 		unsigned int DISABLE_PSR_ENTRY_ABORT:1;
 		unsigned int SKIP_SINGLE_OTG_DISABLE:1;
-		unsigned int RESERVED:22;
+		unsigned int DISABLE_ALPM:1;
+		unsigned int ALPM_DEFAULT_PD_MODE:1;
+		unsigned int RESERVED:20;
 	} bits;
 	unsigned int u32all;
 };
@@ -779,6 +741,14 @@ struct psr_context {
 	unsigned int frame_delay;
 	bool allow_smu_optimizations;
 	bool allow_multi_disp_optimizations;
+	/* Panel self refresh 2 selective update granularity required */
+	bool su_granularity_required;
+	/* psr2 selective update y granularity capability */
+	uint8_t su_y_granularity;
+	unsigned int line_time_in_us;
+	uint8_t rate_control_caps;
+	uint16_t dsc_slice_height;
+	bool os_request_force_ffu;
 };
 
 struct colorspace_transform {
@@ -833,6 +803,7 @@ struct dc_context {
 	struct dc *dc;
 
 	void *driver_context; /* e.g. amdgpu_device */
+	struct dal_logger *logger;
 	struct dc_perf_trace *perf_trace;
 	void *cgs_device;
 
@@ -851,10 +822,10 @@ struct dc_context {
 	uint32_t dc_edp_id_count;
 	uint64_t fbc_gpu_addr;
 	struct dc_dmub_srv *dmub_srv;
-#ifdef CONFIG_DRM_AMD_DC_HDCP
 	struct cp_psp cp_psp;
-#endif
-
+	uint32_t *dcn_reg_offsets;
+	uint32_t *nbio_reg_offsets;
+	uint32_t *clk_reg_offsets;
 };
 
 /* DSC DPCD capabilities */
@@ -925,7 +896,15 @@ struct dsc_dec_dpcd_caps {
 	uint32_t branch_overall_throughput_0_mps; /* In MPs */
 	uint32_t branch_overall_throughput_1_mps; /* In MPs */
 	uint32_t branch_max_line_width;
-	bool is_dp;
+	bool is_dp; /* Decoded format */
+};
+
+struct hblank_expansion_dpcd_caps {
+	bool expansion_supported;
+	bool reduction_supported;
+	bool buffer_unit_bytes; /* True: buffer size in bytes. False: buffer size in pixels*/
+	bool buffer_per_port; /* True: buffer size per port. False: buffer size per lane*/
+	uint32_t buffer_size; /* Add 1 to value and multiply by 32 */
 };
 
 struct dc_golden_table {
@@ -948,11 +927,26 @@ enum dc_gpu_mem_alloc_type {
 	DC_MEM_ALLOC_TYPE_AGP
 };
 
+enum dc_link_encoding_format {
+	DC_LINK_ENCODING_UNSPECIFIED = 0,
+	DC_LINK_ENCODING_DP_8b_10b,
+	DC_LINK_ENCODING_DP_128b_132b,
+	DC_LINK_ENCODING_HDMI_TMDS,
+	DC_LINK_ENCODING_HDMI_FRL
+};
+
 enum dc_psr_version {
 	DC_PSR_VERSION_1			= 0,
+	DC_PSR_VERSION_SU_1			= 1,
 	DC_PSR_VERSION_UNSUPPORTED		= 0xFFFFFFFF,
 };
 
+enum dc_replay_version {
+	DC_FREESYNC_REPLAY = 0,
+	DC_VESA_PANEL_REPLAY = 1,
+	DC_REPLAY_VERSION_UNSUPPORTED = 0XFF,
+};
+
 /* Possible values of display_endpoint_id.endpoint */
 enum display_endpoint_type {
 	DISPLAY_ENDPOINT_PHY = 0, /* Physical connector. */
@@ -969,4 +963,470 @@ struct display_endpoint_id {
 	enum display_endpoint_type ep_type;
 };
 
+enum backlight_control_type {
+	BACKLIGHT_CONTROL_PWM = 0,
+	BACKLIGHT_CONTROL_VESA_AUX = 1,
+	BACKLIGHT_CONTROL_AMD_AUX = 2,
+};
+
+#if defined(CONFIG_DRM_AMD_SECURE_DISPLAY)
+#define MAX_CRC_WINDOW_NUM	2
+
+struct otg_phy_mux {
+	uint8_t phy_output_num;
+	uint8_t otg_output_num;
+};
+
+struct crc_window {
+	struct rect rect;
+	bool enable;
+};
+#endif
+
+enum dc_detect_reason {
+	DETECT_REASON_BOOT,
+	DETECT_REASON_RESUMEFROMS3S4,
+	DETECT_REASON_HPD,
+	DETECT_REASON_HPDRX,
+	DETECT_REASON_FALLBACK,
+	DETECT_REASON_RETRAIN,
+	DETECT_REASON_TDR,
+};
+
+struct dc_link_status {
+	bool link_active;
+	struct dpcd_caps *dpcd_caps;
+};
+
+union hdcp_rx_caps {
+	struct {
+		uint8_t version;
+		uint8_t reserved;
+		struct {
+			uint8_t repeater	: 1;
+			uint8_t hdcp_capable	: 1;
+			uint8_t reserved	: 6;
+		} byte0;
+	} fields;
+	uint8_t raw[3];
+};
+
+union hdcp_bcaps {
+	struct {
+		uint8_t HDCP_CAPABLE:1;
+		uint8_t REPEATER:1;
+		uint8_t RESERVED:6;
+	} bits;
+	uint8_t raw;
+};
+
+struct hdcp_caps {
+	union hdcp_rx_caps rx_caps;
+	union hdcp_bcaps bcaps;
+};
+
+/* DP MST stream allocation (payload bandwidth number) */
+struct link_mst_stream_allocation {
+	/* DIG front */
+	const struct stream_encoder *stream_enc;
+	/* HPO DP Stream Encoder */
+	const struct hpo_dp_stream_encoder *hpo_dp_stream_enc;
+	/* associate DRM payload table with DC stream encoder */
+	uint8_t vcp_id;
+	/* number of slots required for the DP stream in transport packet */
+	uint8_t slot_count;
+};
+
+#define MAX_CONTROLLER_NUM 6
+
+/* DP MST stream allocation table */
+struct link_mst_stream_allocation_table {
+	/* number of DP video streams */
+	int stream_count;
+	/* array of stream allocations */
+	struct link_mst_stream_allocation stream_allocations[MAX_CONTROLLER_NUM];
+};
+
+/* PSR feature flags */
+struct psr_settings {
+	bool psr_feature_enabled;		// PSR is supported by sink
+	bool psr_allow_active;			// PSR is currently active
+	enum dc_psr_version psr_version;		// Internal PSR version, determined based on DPCD
+	bool psr_vtotal_control_support;	// Vtotal control is supported by sink
+	unsigned long long psr_dirty_rects_change_timestamp_ns;	// for delay of enabling PSR-SU
+
+	/* These parameters are calculated in Driver,
+	 * based on display timing and Sink capabilities.
+	 * If VBLANK region is too small and Sink takes a long time
+	 * to set up RFB, it may take an extra frame to enter PSR state.
+	 */
+	bool psr_frame_capture_indication_req;
+	unsigned int psr_sdp_transmit_line_num_deadline;
+	uint8_t force_ffu_mode;
+	unsigned int psr_power_opt;
+
+	/**
+	 * Some panels cannot handle idle pattern during PSR entry.
+	 * To power down phy before disable stream to avoid sending
+	 * idle pattern.
+	 */
+	uint8_t power_down_phy_before_disable_stream;
+};
+
+enum replay_coasting_vtotal_type {
+	PR_COASTING_TYPE_NOM = 0,
+	PR_COASTING_TYPE_STATIC,
+	PR_COASTING_TYPE_FULL_SCREEN_VIDEO,
+	PR_COASTING_TYPE_TEST_HARNESS,
+	PR_COASTING_TYPE_NUM,
+};
+
+enum replay_link_off_frame_count_level {
+	PR_LINK_OFF_FRAME_COUNT_FAIL = 0x0,
+	PR_LINK_OFF_FRAME_COUNT_GOOD = 0x2,
+	PR_LINK_OFF_FRAME_COUNT_BEST = 0x6,
+};
+
+/*
+ * This is general Interface for Replay to
+ * set an 32 bit variable to dmub
+ * The Message_type indicates which variable
+ * passed to DMUB.
+ */
+enum replay_FW_Message_type {
+	Replay_Msg_Not_Support = -1,
+	Replay_Set_Timing_Sync_Supported,
+	Replay_Set_Residency_Frameupdate_Timer,
+	Replay_Set_Pseudo_VTotal,
+	Replay_Disabled_Adaptive_Sync_SDP,
+	Replay_Set_Version,
+	Replay_Set_General_Cmd,
+};
+
+union replay_error_status {
+	struct {
+		unsigned int STATE_TRANSITION_ERROR     :1;
+		unsigned int LINK_CRC_ERROR             :1;
+		unsigned int DESYNC_ERROR               :1;
+		unsigned int RESERVED_3                 :1;
+		unsigned int LOW_RR_INCORRECT_VTOTAL    :1;
+		unsigned int NO_DOUBLED_RR              :1;
+		unsigned int RESERVED_6_7               :2;
+	} bits;
+	unsigned char raw;
+};
+
+union replay_low_refresh_rate_enable_options {
+	struct {
+	//BIT[0-3]: Replay Low Hz Support control
+		unsigned int ENABLE_LOW_RR_SUPPORT          :1;
+		unsigned int SKIP_ASIC_CHECK                :1;
+		unsigned int RESERVED_2_3                   :2;
+	//BIT[4-15]: Replay Low Hz Enable Scenarios
+		unsigned int ENABLE_STATIC_SCREEN           :1;
+		unsigned int ENABLE_FULL_SCREEN_VIDEO       :1;
+		unsigned int ENABLE_GENERAL_UI              :1;
+		unsigned int RESERVED_7_15                  :9;
+	//BIT[16-31]: Replay Low Hz Enable Check
+		unsigned int ENABLE_STATIC_FLICKER_CHECK    :1;
+		unsigned int RESERVED_17_31                 :15;
+	} bits;
+	unsigned int raw;
+};
+
+struct replay_config {
+	/* Replay version */
+	enum dc_replay_version replay_version;
+	/* Replay feature is supported */
+	bool replay_supported;
+	/* Replay caps support DPCD & EDID caps*/
+	bool replay_cap_support;
+	/* Power opt flags that are supported */
+	unsigned int replay_power_opt_supported;
+	/* SMU optimization is supported */
+	bool replay_smu_opt_supported;
+	/* Replay enablement option */
+	unsigned int replay_enable_option;
+	/* Replay debug flags */
+	uint32_t debug_flags;
+	/* Replay sync is supported */
+	bool replay_timing_sync_supported;
+	/* Replay Disable desync error check. */
+	bool force_disable_desync_error_check;
+	/* Replay Received Desync Error HPD. */
+	bool received_desync_error_hpd;
+	/* Replay feature is supported long vblank */
+	bool replay_support_fast_resync_in_ultra_sleep_mode;
+	/* Replay error status */
+	union replay_error_status replay_error_status;
+	/* Replay Low Hz enable Options */
+	union replay_low_refresh_rate_enable_options low_rr_enable_options;
+	/* Replay coasting vtotal is within low refresh rate range. */
+	bool low_rr_activated;
+	/* Replay low refresh rate supported*/
+	bool low_rr_supported;
+	/* Replay Video Conferencing Optimization Enabled */
+	bool replay_video_conferencing_optimization_enabled;
+	/* Replay alpm mode */
+	enum dc_alpm_mode alpm_mode;
+	/* Replay full screen only */
+	bool os_request_force_ffu;
+};
+
+/* Replay feature flags*/
+struct replay_settings {
+	/* Replay configuration */
+	struct replay_config config;
+	/* Replay feature is ready for activating */
+	bool replay_feature_enabled;
+	/* Replay is currently active */
+	bool replay_allow_active;
+	/* Replay is currently active */
+	bool replay_allow_long_vblank;
+	/* Power opt flags that are activated currently */
+	unsigned int replay_power_opt_active;
+	/* SMU optimization is enabled */
+	bool replay_smu_opt_enable;
+	/* Current Coasting vtotal */
+	uint32_t coasting_vtotal;
+	/* Coasting vtotal table */
+	uint32_t coasting_vtotal_table[PR_COASTING_TYPE_NUM];
+	/* Defer Update Coasting vtotal table */
+	uint32_t defer_update_coasting_vtotal_table[PR_COASTING_TYPE_NUM];
+	/* Skip frame number table */
+	uint32_t frame_skip_number_table[PR_COASTING_TYPE_NUM];
+	/* Defer skip frame number table */
+	uint32_t defer_frame_skip_number_table[PR_COASTING_TYPE_NUM];
+	/* Maximum link off frame count */
+	uint32_t link_off_frame_count;
+	/* Replay pseudo vtotal for low refresh rate*/
+	uint16_t low_rr_full_screen_video_pseudo_vtotal;
+	/* Replay last pseudo vtotal set to DMUB */
+	uint16_t last_pseudo_vtotal;
+	/* Replay desync error */
+	uint32_t replay_desync_error_fail_count;
+	/* The frame skip number dal send to DMUB */
+	uint16_t frame_skip_number;
+};
+
+/* To split out "global" and "per-panel" config settings.
+ * Add a struct dc_panel_config under dc_link
+ */
+struct dc_panel_config {
+	/* extra panel power sequence parameters */
+	struct pps {
+		unsigned int extra_t3_ms;
+		unsigned int extra_t7_ms;
+		unsigned int extra_delay_backlight_off;
+		unsigned int extra_post_t7_ms;
+		unsigned int extra_pre_t11_ms;
+		unsigned int extra_t12_ms;
+		unsigned int extra_post_OUI_ms;
+	} pps;
+	/* nit brightness */
+	struct nits_brightness {
+		unsigned int peak; /* nits */
+		unsigned int max_avg; /* nits */
+		unsigned int min; /* 1/10000 nits */
+		unsigned int max_nonboost_brightness_millinits;
+		unsigned int min_brightness_millinits;
+	} nits_brightness;
+	/* PSR */
+	struct psr {
+		bool disable_psr;
+		bool disallow_psrsu;
+		bool disallow_replay;
+		bool rc_disable;
+		bool rc_allow_static_screen;
+		bool rc_allow_fullscreen_VPB;
+		bool read_psrcap_again;
+		unsigned int replay_enable_option;
+	} psr;
+	/* ABM */
+	struct varib {
+		unsigned int varibright_feature_enable;
+		unsigned int def_varibright_level;
+		unsigned int abm_config_setting;
+	} varib;
+	/* edp DSC */
+	struct dsc {
+		bool disable_dsc_edp;
+		unsigned int force_dsc_edp_policy;
+	} dsc;
+	/* eDP ILR */
+	struct ilr {
+		bool optimize_edp_link_rate; /* eDP ILR */
+	} ilr;
+};
+
+#define MAX_SINKS_PER_LINK 4
+
+/*
+ *  USB4 DPIA BW ALLOCATION STRUCTS
+ */
+struct dc_dpia_bw_alloc {
+	int remote_sink_req_bw[MAX_SINKS_PER_LINK]; // BW requested by remote sinks
+	int link_verified_bw;  // The Verified BW that link can allocated and use that has been verified already
+	int link_max_bw;       // The Max BW that link can require/support
+	int allocated_bw;      // The Actual Allocated BW for this DPIA
+	int estimated_bw;      // The estimated available BW for this DPIA
+	int bw_granularity;    // BW Granularity
+	int dp_overhead;       // DP overhead in dp tunneling
+	bool bw_alloc_enabled; // The BW Alloc Mode Support is turned ON for all 3:  DP-Tx & Dpia & CM
+	uint8_t nrd_max_lane_count; // Non-reduced max lane count
+	uint8_t nrd_max_link_rate; // Non-reduced max link rate
+};
+
+enum dc_hpd_enable_select {
+	HPD_EN_FOR_ALL_EDP = 0,
+	HPD_EN_FOR_PRIMARY_EDP_ONLY,
+	HPD_EN_FOR_SECONDARY_EDP_ONLY,
+};
+
+enum dc_cm2_shaper_3dlut_setting {
+	DC_CM2_SHAPER_3DLUT_SETTING_BYPASS_ALL,
+	DC_CM2_SHAPER_3DLUT_SETTING_ENABLE_SHAPER,
+	/* Bypassing Shaper will always bypass 3DLUT */
+	DC_CM2_SHAPER_3DLUT_SETTING_ENABLE_SHAPER_3DLUT
+};
+
+enum dc_cm2_gpu_mem_layout {
+	DC_CM2_GPU_MEM_LAYOUT_3D_SWIZZLE_LINEAR_RGB,
+	DC_CM2_GPU_MEM_LAYOUT_3D_SWIZZLE_LINEAR_BGR,
+	DC_CM2_GPU_MEM_LAYOUT_1D_PACKED_LINEAR
+};
+
+enum dc_cm2_gpu_mem_pixel_component_order {
+	DC_CM2_GPU_MEM_PIXEL_COMPONENT_ORDER_RGBA,
+};
+
+enum dc_cm2_gpu_mem_format {
+	DC_CM2_GPU_MEM_FORMAT_16161616_UNORM_12MSB,
+	DC_CM2_GPU_MEM_FORMAT_16161616_UNORM_12LSB,
+	DC_CM2_GPU_MEM_FORMAT_16161616_FLOAT_FP1_5_10
+};
+
+struct dc_cm2_gpu_mem_format_parameters {
+	enum dc_cm2_gpu_mem_format format;
+	union {
+		struct {
+			/* bias & scale for float only */
+			uint16_t bias;
+			uint16_t scale;
+		} float_params;
+	};
+};
+
+enum dc_cm2_gpu_mem_size {
+	DC_CM2_GPU_MEM_SIZE_171717,
+	DC_CM2_GPU_MEM_SIZE_TRANSFORMED,
+};
+
+struct dc_cm2_gpu_mem_parameters {
+	struct dc_plane_address addr;
+	enum dc_cm2_gpu_mem_layout layout;
+	struct dc_cm2_gpu_mem_format_parameters format_params;
+	enum dc_cm2_gpu_mem_pixel_component_order component_order;
+	enum dc_cm2_gpu_mem_size  size;
+	uint16_t bit_depth;
+};
+
+enum dc_cm2_transfer_func_source {
+	DC_CM2_TRANSFER_FUNC_SOURCE_SYSMEM,
+	DC_CM2_TRANSFER_FUNC_SOURCE_VIDMEM
+};
+
+struct dc_cm2_component_settings {
+	enum dc_cm2_shaper_3dlut_setting shaper_3dlut_setting;
+	bool lut1d_enable;
+};
+
+/*
+ * All pointers in this struct must remain valid for as long as the 3DLUTs are used
+ */
+struct dc_cm2_func_luts {
+	const struct dc_transfer_func *shaper;
+	struct {
+		enum dc_cm2_transfer_func_source lut3d_src;
+		union {
+			const struct dc_3dlut *lut3d_func;
+			struct dc_cm2_gpu_mem_parameters gpu_mem_params;
+		};
+		bool rmcm_3dlut_shaper_select;
+		bool mpc_3dlut_enable;
+		bool rmcm_3dlut_enable;
+		bool mpc_mcm_post_blend;
+		uint8_t rmcm_tmz;
+	} lut3d_data;
+	const struct dc_transfer_func *lut1d_func;
+};
+
+struct dc_cm2_parameters {
+	struct dc_cm2_component_settings component_settings;
+	struct dc_cm2_func_luts cm2_luts;
+};
+
+enum mall_stream_type {
+	SUBVP_NONE, // subvp not in use
+	SUBVP_MAIN, // subvp in use, this stream is main stream
+	SUBVP_PHANTOM, // subvp in use, this stream is a phantom stream
+};
+
+enum dc_power_source_type {
+	DC_POWER_SOURCE_AC, // wall power
+	DC_POWER_SOURCE_DC, // battery power
+};
+
+struct dc_state_create_params {
+	enum dc_power_source_type power_source;
+};
+
+struct dc_commit_streams_params {
+	struct dc_stream_state **streams;
+	uint8_t stream_count;
+	enum dc_power_source_type power_source;
+};
+
+struct set_backlight_level_params {
+	/* backlight in pwm */
+	uint32_t backlight_pwm_u16_16;
+	/* brightness ramping */
+	uint32_t frame_ramp;
+	/* backlight control type
+	 * 0: PWM backlight control
+	 * 1: VESA AUX backlight control
+	 * 2: AMD AUX backlight control
+	 */
+	enum backlight_control_type control_type;
+	/* backlight in millinits */
+	uint32_t backlight_millinits;
+	/* transition time in ms */
+	uint32_t transition_time_in_ms;
+	/* minimum luminance in nits */
+	uint32_t min_luminance;
+	/* maximum luminance in nits */
+	uint32_t max_luminance;
+	/* minimum backlight in pwm */
+	uint32_t min_backlight_pwm;
+	/* maximum backlight in pwm */
+	uint32_t max_backlight_pwm;
+	/* AUX HW instance */
+	uint8_t aux_inst;
+};
+
+enum dc_validate_mode {
+	/* validate the mode and program HW */
+	DC_VALIDATE_MODE_AND_PROGRAMMING = 0,
+	/* only validate the mode */
+	DC_VALIDATE_MODE_ONLY = 1,
+	/* validate the mode and get the max state (voltage level) */
+	DC_VALIDATE_MODE_AND_STATE_INDEX = 2,
+};
+
+struct dc_validation_dpia_set {
+	const struct dc_link *link;
+	const struct dc_tunnel_settings *tunnel_settings;
+	uint32_t required_bw;
+};
+
 #endif /* DC_TYPES_H_ */
diff --git a/drivers/gpu/drm/amd/display/dc/dccg/Makefile b/drivers/gpu/drm/amd/display/dc/dccg/Makefile
new file mode 100644
index 000000000000..1d5cf0f8e79d
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dccg/Makefile
@@ -0,0 +1,103 @@
+
+# Copyright 2022 Advanced Micro Devices, Inc.
+#
+# Permission is hereby granted, free of charge, to any person obtaining a
+# copy of this software and associated documentation files (the "Software"),
+# to deal in the Software without restriction, including without limitation
+# the rights to use, copy, modify, merge, publish, distribute, sublicense,
+# and/or sell copies of the Software, and to permit persons to whom the
+# Software is furnished to do so, subject to the following conditions:
+#
+# The above copyright notice and this permission notice shall be included in
+# all copies or substantial portions of the Software.
+#
+# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+# THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+# OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+# ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+# OTHER DEALINGS IN THE SOFTWARE.
+#
+# Makefile for the 'dccg' sub-component of DAL.
+#
+ifdef CONFIG_DRM_AMD_DC_FP
+###############################################################################
+# DCN
+###############################################################################
+
+DCCG_DCN20 = dcn20_dccg.o
+
+AMD_DAL_DCCG_DCN20 = $(addprefix $(AMDDALPATH)/dc/dccg/dcn20/,$(DCCG_DCN20))
+
+AMD_DISPLAY_FILES += $(AMD_DAL_DCCG_DCN20)
+
+###############################################################################
+
+DCCG_DCN201 = dcn201_dccg.o
+
+AMD_DAL_DCCG_DCN201 = $(addprefix $(AMDDALPATH)/dc/dccg/dcn201/,$(DCCG_DCN201))
+
+AMD_DISPLAY_FILES += $(AMD_DAL_DCCG_DCN201)
+
+###############################################################################
+
+DCCG_DCN21 = dcn21_dccg.o
+
+AMD_DAL_DCCG_DCN21 = $(addprefix $(AMDDALPATH)/dc/dccg/dcn21/,$(DCCG_DCN21))
+
+AMD_DISPLAY_FILES += $(AMD_DAL_DCCG_DCN21)
+
+###############################################################################
+DCCG_DCN30 = dcn30_dccg.o
+
+AMD_DAL_DCCG_DCN30 = $(addprefix $(AMDDALPATH)/dc/dccg/dcn30/,$(DCCG_DCN30))
+
+AMD_DISPLAY_FILES += $(AMD_DAL_DCCG_DCN30)
+
+###############################################################################
+DCCG_DCN301 = dcn301_dccg.o
+
+AMD_DAL_DCCG_DCN301 = $(addprefix $(AMDDALPATH)/dc/dccg/dcn301/,$(DCCG_DCN301))
+
+AMD_DISPLAY_FILES += $(AMD_DAL_DCCG_DCN301)
+
+###############################################################################
+
+DCCG_DCN31 = dcn31_dccg.o
+
+AMD_DAL_DCCG_DCN31 = $(addprefix $(AMDDALPATH)/dc/dccg/dcn31/,$(DCCG_DCN31))
+
+AMD_DISPLAY_FILES += $(AMD_DAL_DCCG_DCN31)
+
+###############################################################################
+
+DCCG_DCN314 = dcn314_dccg.o
+
+AMD_DAL_DCCG_DCN314 = $(addprefix $(AMDDALPATH)/dc/dccg/dcn314/,$(DCCG_DCN314))
+
+AMD_DISPLAY_FILES += $(AMD_DAL_DCCG_DCN314)
+
+###############################################################################
+
+DCCG_DCN32 = dcn32_dccg.o
+
+AMD_DAL_DCCG_DCN32 = $(addprefix $(AMDDALPATH)/dc/dccg/dcn32/,$(DCCG_DCN32))
+
+AMD_DISPLAY_FILES += $(AMD_DAL_DCCG_DCN32)
+
+###############################################################################
+
+DCCG_DCN35 = dcn35_dccg.o
+
+AMD_DAL_DCCG_DCN35 = $(addprefix $(AMDDALPATH)/dc/dccg/dcn35/,$(DCCG_DCN35))
+
+AMD_DISPLAY_FILES += $(AMD_DAL_DCCG_DCN35)
+
+###############################################################################
+DCCG_DCN401 = dcn401_dccg.o
+
+AMD_DAL_DCCG_DCN401 = $(addprefix $(AMDDALPATH)/dc/dccg/dcn401/,$(DCCG_DCN401))
+
+AMD_DISPLAY_FILES += $(AMD_DAL_DCCG_DCN401)
+endif
diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_dccg.c b/drivers/gpu/drm/amd/display/dc/dccg/dcn20/dcn20_dccg.c
index 5999b2da3a01..33d8bd91cb01 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_dccg.c
+++ b/drivers/gpu/drm/amd/display/dc/dccg/dcn20/dcn20_dccg.c
@@ -148,7 +148,7 @@ struct dccg *dccg2_create(
 	const struct dccg_shift *dccg_shift,
 	const struct dccg_mask *dccg_mask)
 {
-	struct dcn_dccg *dccg_dcn = kzalloc(sizeof(*dccg_dcn), GFP_ATOMIC);
+	struct dcn_dccg *dccg_dcn = kzalloc(sizeof(*dccg_dcn), GFP_KERNEL);
 	struct dccg *base;
 
 	if (dccg_dcn == NULL) {
diff --git a/drivers/gpu/drm/amd/display/dc/dccg/dcn20/dcn20_dccg.h b/drivers/gpu/drm/amd/display/dc/dccg/dcn20/dcn20_dccg.h
new file mode 100644
index 000000000000..8bdffd9ff31b
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dccg/dcn20/dcn20_dccg.h
@@ -0,0 +1,526 @@
+/*
+ * Copyright 2018 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __DCN20_DCCG_H__
+#define __DCN20_DCCG_H__
+
+#include "dccg.h"
+
+#define DCCG_COMMON_REG_LIST_DCN_BASE() \
+	SR(DPPCLK_DTO_CTRL),\
+	DCCG_SRII(DTO_PARAM, DPPCLK, 0),\
+	DCCG_SRII(DTO_PARAM, DPPCLK, 1),\
+	DCCG_SRII(DTO_PARAM, DPPCLK, 2),\
+	DCCG_SRII(DTO_PARAM, DPPCLK, 3),\
+	SR(REFCLK_CNTL),\
+	DCCG_SRII(PIXEL_RATE_CNTL, OTG, 0),\
+	DCCG_SRII(PIXEL_RATE_CNTL, OTG, 1),\
+	SR(DISPCLK_FREQ_CHANGE_CNTL)
+
+#define DCCG_REG_LIST_DCN2() \
+	DCCG_COMMON_REG_LIST_DCN_BASE(),\
+	DCCG_SRII(DTO_PARAM, DPPCLK, 4),\
+	DCCG_SRII(DTO_PARAM, DPPCLK, 5),\
+	DCCG_SRII(PIXEL_RATE_CNTL, OTG, 2),\
+	DCCG_SRII(PIXEL_RATE_CNTL, OTG, 3),\
+	DCCG_SRII(PIXEL_RATE_CNTL, OTG, 4),\
+	DCCG_SRII(PIXEL_RATE_CNTL, OTG, 5)
+
+#define DCCG_SF(reg_name, field_name, post_fix)\
+	.field_name = reg_name ## __ ## field_name ## post_fix
+
+#define DCCG_SFI(reg_name, field_name, field_prefix, inst, post_fix)\
+	.field_prefix ## _ ## field_name[inst] = reg_name ## __ ## field_prefix ## inst ## _ ## field_name ## post_fix
+
+#define DCCG_SFII(block, reg_name, field_prefix, field_name, inst, post_fix)\
+	.field_prefix ## _ ## field_name[inst] = block ## inst ## _ ## reg_name ## __ ## field_prefix ## inst ## _ ## field_name ## post_fix
+
+#define DCCG_COMMON_MASK_SH_LIST_DCN_COMMON_BASE(mask_sh) \
+	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_ENABLE, DPPCLK, 0, mask_sh),\
+	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_DB_EN, DPPCLK, 0, mask_sh),\
+	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_ENABLE, DPPCLK, 1, mask_sh),\
+	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_DB_EN, DPPCLK, 1, mask_sh),\
+	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_ENABLE, DPPCLK, 2, mask_sh),\
+	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_DB_EN, DPPCLK, 2, mask_sh),\
+	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_ENABLE, DPPCLK, 3, mask_sh),\
+	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_DB_EN, DPPCLK, 3, mask_sh),\
+	DCCG_SF(DPPCLK0_DTO_PARAM, DPPCLK0_DTO_PHASE, mask_sh),\
+	DCCG_SF(DPPCLK0_DTO_PARAM, DPPCLK0_DTO_MODULO, mask_sh),\
+	DCCG_SF(REFCLK_CNTL, REFCLK_CLOCK_EN, mask_sh),\
+	DCCG_SF(REFCLK_CNTL, REFCLK_SRC_SEL, mask_sh),\
+	DCCG_SF(DISPCLK_FREQ_CHANGE_CNTL, DISPCLK_STEP_DELAY, mask_sh),\
+	DCCG_SF(DISPCLK_FREQ_CHANGE_CNTL, DISPCLK_STEP_SIZE, mask_sh),\
+	DCCG_SF(DISPCLK_FREQ_CHANGE_CNTL, DISPCLK_FREQ_RAMP_DONE, mask_sh),\
+	DCCG_SF(DISPCLK_FREQ_CHANGE_CNTL, DISPCLK_MAX_ERRDET_CYCLES, mask_sh),\
+	DCCG_SF(DISPCLK_FREQ_CHANGE_CNTL, DCCG_FIFO_ERRDET_RESET, mask_sh),\
+	DCCG_SF(DISPCLK_FREQ_CHANGE_CNTL, DCCG_FIFO_ERRDET_STATE, mask_sh),\
+	DCCG_SF(DISPCLK_FREQ_CHANGE_CNTL, DCCG_FIFO_ERRDET_OVR_EN, mask_sh),\
+	DCCG_SF(DISPCLK_FREQ_CHANGE_CNTL, DISPCLK_CHG_FWD_CORR_DISABLE, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, ADD_PIXEL, 0, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, ADD_PIXEL, 1, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, DROP_PIXEL, 0, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, DROP_PIXEL, 1, mask_sh)
+
+
+
+
+#define DCCG_MASK_SH_LIST_DCN2(mask_sh) \
+	DCCG_COMMON_MASK_SH_LIST_DCN_COMMON_BASE(mask_sh),\
+	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_ENABLE, DPPCLK, 4, mask_sh),\
+	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_DB_EN, DPPCLK, 4, mask_sh),\
+	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_ENABLE, DPPCLK, 5, mask_sh),\
+	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_DB_EN, DPPCLK, 5, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, ADD_PIXEL, 2, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, ADD_PIXEL, 3, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, ADD_PIXEL, 4, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, ADD_PIXEL, 5, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, DROP_PIXEL, 2, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, DROP_PIXEL, 3, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, DROP_PIXEL, 4, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, DROP_PIXEL, 5, mask_sh)
+
+#define DCCG_MASK_SH_LIST_DCN2_1(mask_sh) \
+	DCCG_COMMON_MASK_SH_LIST_DCN_COMMON_BASE(mask_sh),\
+	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_ENABLE, DPPCLK, 4, mask_sh),\
+	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_DB_EN, DPPCLK, 4, mask_sh),\
+	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_ENABLE, DPPCLK, 5, mask_sh),\
+	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_DB_EN, DPPCLK, 5, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, ADD_PIXEL, 2, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, ADD_PIXEL, 3, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, DROP_PIXEL, 2, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, DROP_PIXEL, 3, mask_sh)
+
+
+#define DCCG_REG_FIELD_LIST(type) \
+	type DPPCLK0_DTO_PHASE;\
+	type DPPCLK0_DTO_MODULO;\
+	type DPPCLK_DTO_ENABLE[6];\
+	type DPPCLK_DTO_DB_EN[6];\
+	type REFCLK_CLOCK_EN;\
+	type REFCLK_SRC_SEL;\
+	type DISPCLK_STEP_DELAY;\
+	type DISPCLK_STEP_SIZE;\
+	type DISPCLK_FREQ_RAMP_DONE;\
+	type DISPCLK_MAX_ERRDET_CYCLES;\
+	type DCCG_FIFO_ERRDET_RESET;\
+	type DCCG_FIFO_ERRDET_STATE;\
+	type DCCG_FIFO_ERRDET_OVR_EN;\
+	type DISPCLK_CHG_FWD_CORR_DISABLE;\
+	type DISPCLK_FREQ_CHANGE_CNTL;\
+	type OTG_ADD_PIXEL[MAX_PIPES];\
+	type OTG_DROP_PIXEL[MAX_PIPES];
+
+#define DCCG3_REG_FIELD_LIST(type) \
+	type HDMICHARCLK0_EN;\
+	type HDMICHARCLK0_SRC_SEL;\
+	type PHYASYMCLK_FORCE_EN;\
+	type PHYASYMCLK_FORCE_SRC_SEL;\
+	type PHYBSYMCLK_FORCE_EN;\
+	type PHYBSYMCLK_FORCE_SRC_SEL;\
+	type PHYCSYMCLK_FORCE_EN;\
+	type PHYCSYMCLK_FORCE_SRC_SEL;
+
+#define DCCG31_REG_FIELD_LIST(type) \
+	type PHYDSYMCLK_FORCE_EN;\
+	type PHYDSYMCLK_FORCE_SRC_SEL;\
+	type PHYESYMCLK_FORCE_EN;\
+	type PHYESYMCLK_FORCE_SRC_SEL;\
+	type DPSTREAMCLK_PIPE0_EN;\
+	type DPSTREAMCLK_PIPE1_EN;\
+	type DPSTREAMCLK_PIPE2_EN;\
+	type DPSTREAMCLK_PIPE3_EN;\
+	type HDMISTREAMCLK0_SRC_SEL;\
+	type HDMISTREAMCLK0_DTO_FORCE_DIS;\
+	type SYMCLK32_SE0_SRC_SEL;\
+	type SYMCLK32_SE1_SRC_SEL;\
+	type SYMCLK32_SE2_SRC_SEL;\
+	type SYMCLK32_SE3_SRC_SEL;\
+	type SYMCLK32_SE0_EN;\
+	type SYMCLK32_SE1_EN;\
+	type SYMCLK32_SE2_EN;\
+	type SYMCLK32_SE3_EN;\
+	type SYMCLK32_LE0_SRC_SEL;\
+	type SYMCLK32_LE1_SRC_SEL;\
+	type SYMCLK32_LE0_EN;\
+	type SYMCLK32_LE1_EN;\
+	type DTBCLK_DTO_ENABLE[MAX_PIPES];\
+	type DTBCLKDTO_ENABLE_STATUS[MAX_PIPES];\
+	type PIPE_DTO_SRC_SEL[MAX_PIPES];\
+	type DTBCLK_DTO_DIV[MAX_PIPES];\
+	type DCCG_AUDIO_DTO_SEL;\
+	type DCCG_AUDIO_DTO0_SOURCE_SEL;\
+	type DENTIST_DISPCLK_CHG_MODE;\
+	type DSCCLK0_DTO_PHASE;\
+	type DSCCLK0_DTO_MODULO;\
+	type DSCCLK1_DTO_PHASE;\
+	type DSCCLK1_DTO_MODULO;\
+	type DSCCLK2_DTO_PHASE;\
+	type DSCCLK2_DTO_MODULO;\
+	type DSCCLK0_DTO_ENABLE;\
+	type DSCCLK1_DTO_ENABLE;\
+	type DSCCLK2_DTO_ENABLE;\
+	type SYMCLK32_ROOT_SE0_GATE_DISABLE;\
+	type SYMCLK32_ROOT_SE1_GATE_DISABLE;\
+	type SYMCLK32_ROOT_SE2_GATE_DISABLE;\
+	type SYMCLK32_ROOT_SE3_GATE_DISABLE;\
+	type SYMCLK32_SE0_GATE_DISABLE;\
+	type SYMCLK32_SE1_GATE_DISABLE;\
+	type SYMCLK32_SE2_GATE_DISABLE;\
+	type SYMCLK32_SE3_GATE_DISABLE;\
+	type SYMCLK32_ROOT_LE0_GATE_DISABLE;\
+	type SYMCLK32_ROOT_LE1_GATE_DISABLE;\
+	type SYMCLK32_LE0_GATE_DISABLE;\
+	type SYMCLK32_LE1_GATE_DISABLE;\
+	type DPSTREAMCLK_ROOT_GATE_DISABLE;\
+	type DPSTREAMCLK_GATE_DISABLE;\
+	type HDMISTREAMCLK0_DTO_PHASE;\
+	type HDMISTREAMCLK0_DTO_MODULO;\
+	type HDMICHARCLK0_GATE_DISABLE;\
+	type HDMICHARCLK0_ROOT_GATE_DISABLE; \
+	type PHYASYMCLK_GATE_DISABLE; \
+	type PHYBSYMCLK_GATE_DISABLE; \
+	type PHYCSYMCLK_GATE_DISABLE; \
+	type PHYDSYMCLK_GATE_DISABLE; \
+	type PHYESYMCLK_GATE_DISABLE;
+
+#define DCCG314_REG_FIELD_LIST(type) \
+	type DSCCLK3_DTO_PHASE;\
+	type DSCCLK3_DTO_MODULO;\
+	type DSCCLK3_DTO_ENABLE;\
+	type DENTIST_DISPCLK_RDIVIDER;\
+	type DENTIST_DISPCLK_WDIVIDER;
+
+#define DCCG32_REG_FIELD_LIST(type) \
+	type DPSTREAMCLK0_EN;\
+	type DPSTREAMCLK1_EN;\
+	type DPSTREAMCLK2_EN;\
+	type DPSTREAMCLK3_EN;\
+	type DPSTREAMCLK0_SRC_SEL;\
+	type DPSTREAMCLK1_SRC_SEL;\
+	type DPSTREAMCLK2_SRC_SEL;\
+	type DPSTREAMCLK3_SRC_SEL;\
+	type HDMISTREAMCLK0_EN;\
+	type OTG0_PIXEL_RATE_DIVK1;\
+	type OTG0_PIXEL_RATE_DIVK2;\
+	type OTG1_PIXEL_RATE_DIVK1;\
+	type OTG1_PIXEL_RATE_DIVK2;\
+	type OTG2_PIXEL_RATE_DIVK1;\
+	type OTG2_PIXEL_RATE_DIVK2;\
+	type OTG3_PIXEL_RATE_DIVK1;\
+	type OTG3_PIXEL_RATE_DIVK2;\
+	type DTBCLK_P0_SRC_SEL;\
+	type DTBCLK_P0_EN;\
+	type DTBCLK_P1_SRC_SEL;\
+	type DTBCLK_P1_EN;\
+	type DTBCLK_P2_SRC_SEL;\
+	type DTBCLK_P2_EN;\
+	type DTBCLK_P3_SRC_SEL;\
+	type DTBCLK_P3_EN;\
+	type DENTIST_DISPCLK_CHG_DONE;
+
+#define DCCG35_REG_FIELD_LIST(type) \
+	type DPPCLK0_EN;\
+	type DPPCLK1_EN;\
+	type DPPCLK2_EN;\
+	type DPPCLK3_EN;\
+	type DSCCLK0_EN;\
+	type DSCCLK1_EN;\
+	type DSCCLK2_EN;\
+	type DSCCLK3_EN;\
+	type DISPCLK_DCCG_GATE_DISABLE;\
+	type DCCG_GLOBAL_FGCG_REP_DIS; \
+	type PHYASYMCLK_EN;\
+	type PHYASYMCLK_SRC_SEL;\
+	type PHYBSYMCLK_EN;\
+	type PHYBSYMCLK_SRC_SEL;\
+	type PHYCSYMCLK_EN;\
+	type PHYCSYMCLK_SRC_SEL;\
+	type PHYDSYMCLK_EN;\
+	type PHYDSYMCLK_SRC_SEL;\
+	type PHYESYMCLK_EN;\
+	type PHYESYMCLK_SRC_SEL;\
+	type PHYASYMCLK_ROOT_GATE_DISABLE;\
+	type PHYBSYMCLK_ROOT_GATE_DISABLE;\
+	type PHYCSYMCLK_ROOT_GATE_DISABLE;\
+	type PHYDSYMCLK_ROOT_GATE_DISABLE;\
+	type PHYESYMCLK_ROOT_GATE_DISABLE;\
+	type HDMISTREAMCLK0_GATE_DISABLE;\
+	type HDMISTREAMCLK1_GATE_DISABLE;\
+	type HDMISTREAMCLK2_GATE_DISABLE;\
+	type HDMISTREAMCLK3_GATE_DISABLE;\
+	type HDMISTREAMCLK4_GATE_DISABLE;\
+	type HDMISTREAMCLK5_GATE_DISABLE;\
+	type SYMCLKA_CLOCK_ENABLE;\
+	type SYMCLKB_CLOCK_ENABLE;\
+	type SYMCLKC_CLOCK_ENABLE;\
+	type SYMCLKD_CLOCK_ENABLE;\
+	type SYMCLKE_CLOCK_ENABLE;\
+	type SYMCLKA_FE_EN;\
+	type SYMCLKB_FE_EN;\
+	type SYMCLKC_FE_EN;\
+	type SYMCLKD_FE_EN;\
+	type SYMCLKE_FE_EN;\
+	type SYMCLKA_SRC_SEL;\
+	type SYMCLKB_SRC_SEL;\
+	type SYMCLKC_SRC_SEL;\
+	type SYMCLKD_SRC_SEL;\
+	type SYMCLKE_SRC_SEL;\
+	type SYMCLKA_FE_SRC_SEL;\
+	type SYMCLKB_FE_SRC_SEL;\
+	type SYMCLKC_FE_SRC_SEL;\
+	type SYMCLKD_FE_SRC_SEL;\
+	type SYMCLKE_FE_SRC_SEL;\
+	type DTBCLK_P0_GATE_DISABLE;\
+	type DTBCLK_P1_GATE_DISABLE;\
+	type DTBCLK_P2_GATE_DISABLE;\
+	type DTBCLK_P3_GATE_DISABLE;\
+	type DSCCLK0_ROOT_GATE_DISABLE;\
+	type DSCCLK1_ROOT_GATE_DISABLE;\
+	type DSCCLK2_ROOT_GATE_DISABLE;\
+	type DSCCLK3_ROOT_GATE_DISABLE;\
+	type SYMCLKA_FE_ROOT_GATE_DISABLE;\
+	type SYMCLKB_FE_ROOT_GATE_DISABLE;\
+	type SYMCLKC_FE_ROOT_GATE_DISABLE;\
+	type SYMCLKD_FE_ROOT_GATE_DISABLE;\
+	type SYMCLKE_FE_ROOT_GATE_DISABLE;\
+	type DPPCLK0_ROOT_GATE_DISABLE;\
+	type DPPCLK1_ROOT_GATE_DISABLE;\
+	type DPPCLK2_ROOT_GATE_DISABLE;\
+	type DPPCLK3_ROOT_GATE_DISABLE;\
+	type HDMISTREAMCLK0_ROOT_GATE_DISABLE;\
+	type SYMCLKA_ROOT_GATE_DISABLE;\
+	type SYMCLKB_ROOT_GATE_DISABLE;\
+	type SYMCLKC_ROOT_GATE_DISABLE;\
+	type SYMCLKD_ROOT_GATE_DISABLE;\
+	type SYMCLKE_ROOT_GATE_DISABLE;\
+	type PHYA_REFCLK_ROOT_GATE_DISABLE;\
+	type PHYB_REFCLK_ROOT_GATE_DISABLE;\
+	type PHYC_REFCLK_ROOT_GATE_DISABLE;\
+	type PHYD_REFCLK_ROOT_GATE_DISABLE;\
+	type PHYE_REFCLK_ROOT_GATE_DISABLE;\
+	type DPSTREAMCLK0_ROOT_GATE_DISABLE;\
+	type DPSTREAMCLK1_ROOT_GATE_DISABLE;\
+	type DPSTREAMCLK2_ROOT_GATE_DISABLE;\
+	type DPSTREAMCLK3_ROOT_GATE_DISABLE;\
+	type DPSTREAMCLK0_GATE_DISABLE;\
+	type DPSTREAMCLK1_GATE_DISABLE;\
+	type DPSTREAMCLK2_GATE_DISABLE;\
+	type DPSTREAMCLK3_GATE_DISABLE;\
+	type SYMCLKA_FE_GATE_DISABLE;\
+	type SYMCLKB_FE_GATE_DISABLE;\
+	type SYMCLKC_FE_GATE_DISABLE;\
+	type SYMCLKD_FE_GATE_DISABLE;\
+	type SYMCLKE_FE_GATE_DISABLE;\
+	type SYMCLKA_GATE_DISABLE;\
+	type SYMCLKB_GATE_DISABLE;\
+	type SYMCLKC_GATE_DISABLE;\
+	type SYMCLKD_GATE_DISABLE;\
+	type SYMCLKE_GATE_DISABLE;\
+
+
+#define DCCG401_REG_FIELD_LIST(type) \
+	type OTG0_TMDS_PIXEL_RATE_DIV;\
+	type DPDTO0_INT;\
+	type OTG1_TMDS_PIXEL_RATE_DIV;\
+	type DPDTO1_INT;\
+	type OTG2_TMDS_PIXEL_RATE_DIV;\
+	type DPDTO2_INT;\
+	type OTG3_TMDS_PIXEL_RATE_DIV;\
+	type DPDTO3_INT;\
+	type SYMCLK32_ROOT_LE2_GATE_DISABLE;\
+	type SYMCLK32_ROOT_LE3_GATE_DISABLE;\
+	type SYMCLK32_LE2_GATE_DISABLE;\
+	type SYMCLK32_LE3_GATE_DISABLE;\
+	type SYMCLK32_LE2_SRC_SEL;\
+	type SYMCLK32_LE3_SRC_SEL;\
+	type SYMCLK32_LE2_EN;\
+	type SYMCLK32_LE3_EN;\
+	type DP_DTO_ENABLE[MAX_PIPES];
+
+struct dccg_shift {
+	DCCG_REG_FIELD_LIST(uint8_t)
+	DCCG3_REG_FIELD_LIST(uint8_t)
+	DCCG31_REG_FIELD_LIST(uint8_t)
+	DCCG314_REG_FIELD_LIST(uint8_t)
+	DCCG32_REG_FIELD_LIST(uint8_t)
+	DCCG35_REG_FIELD_LIST(uint8_t)
+	DCCG401_REG_FIELD_LIST(uint8_t)
+};
+
+struct dccg_mask {
+	DCCG_REG_FIELD_LIST(uint32_t)
+	DCCG3_REG_FIELD_LIST(uint32_t)
+	DCCG31_REG_FIELD_LIST(uint32_t)
+	DCCG314_REG_FIELD_LIST(uint32_t)
+	DCCG32_REG_FIELD_LIST(uint32_t)
+	DCCG35_REG_FIELD_LIST(uint32_t)
+	DCCG401_REG_FIELD_LIST(uint32_t)
+};
+
+#define DCCG_REG_VARIABLE_LIST \
+	uint32_t DPPCLK_DTO_CTRL; \
+	uint32_t DPPCLK_DTO_PARAM[6]; \
+	uint32_t REFCLK_CNTL; \
+	uint32_t DISPCLK_FREQ_CHANGE_CNTL; \
+	uint32_t OTG_PIXEL_RATE_CNTL[MAX_PIPES]; \
+	uint32_t HDMICHARCLK_CLOCK_CNTL[6]; \
+	uint32_t PHYASYMCLK_CLOCK_CNTL; \
+	uint32_t PHYBSYMCLK_CLOCK_CNTL; \
+	uint32_t PHYCSYMCLK_CLOCK_CNTL; \
+	uint32_t PHYDSYMCLK_CLOCK_CNTL; \
+	uint32_t PHYESYMCLK_CLOCK_CNTL; \
+	uint32_t DTBCLK_DTO_MODULO[MAX_PIPES]; \
+	uint32_t DTBCLK_DTO_PHASE[MAX_PIPES]; \
+	uint32_t DCCG_AUDIO_DTBCLK_DTO_MODULO; \
+	uint32_t DCCG_AUDIO_DTBCLK_DTO_PHASE; \
+	uint32_t DCCG_AUDIO_DTO_SOURCE; \
+	uint32_t DPSTREAMCLK_CNTL; \
+	uint32_t HDMISTREAMCLK_CNTL; \
+	uint32_t SYMCLK32_SE_CNTL; \
+	uint32_t SYMCLK32_LE_CNTL; \
+	uint32_t DENTIST_DISPCLK_CNTL; \
+	uint32_t DSCCLK_DTO_CTRL; \
+	uint32_t DSCCLK0_DTO_PARAM; \
+	uint32_t DSCCLK1_DTO_PARAM; \
+	uint32_t DSCCLK2_DTO_PARAM; \
+	uint32_t DSCCLK3_DTO_PARAM; \
+	uint32_t DPSTREAMCLK_ROOT_GATE_DISABLE; \
+	uint32_t DPSTREAMCLK_GATE_DISABLE; \
+	uint32_t DCCG_GATE_DISABLE_CNTL; \
+	uint32_t DCCG_GATE_DISABLE_CNTL2; \
+	uint32_t DCCG_GATE_DISABLE_CNTL3; \
+	uint32_t HDMISTREAMCLK0_DTO_PARAM; \
+	uint32_t DCCG_GATE_DISABLE_CNTL4; \
+	uint32_t OTG_PIXEL_RATE_DIV; \
+	uint32_t DTBCLK_P_CNTL; \
+	uint32_t DPPCLK_CTRL; \
+	uint32_t DCCG_GATE_DISABLE_CNTL5; \
+	uint32_t DCCG_GATE_DISABLE_CNTL6; \
+	uint32_t DCCG_GLOBAL_FGCG_REP_CNTL; \
+	uint32_t SYMCLKA_CLOCK_ENABLE; \
+	uint32_t SYMCLKB_CLOCK_ENABLE; \
+	uint32_t SYMCLKC_CLOCK_ENABLE; \
+	uint32_t SYMCLKD_CLOCK_ENABLE; \
+	uint32_t SYMCLKE_CLOCK_ENABLE; \
+	uint32_t DP_DTO_MODULO[MAX_PIPES]; \
+	uint32_t DP_DTO_PHASE[MAX_PIPES]; \
+	uint32_t DC_MEM_GLOBAL_PWR_REQ_CNTL; \
+	uint32_t DCCG_AUDIO_DTO0_MODULE; \
+	uint32_t DCCG_AUDIO_DTO0_PHASE; \
+	uint32_t DCCG_AUDIO_DTO1_MODULE; \
+	uint32_t DCCG_AUDIO_DTO1_PHASE; \
+	uint32_t DCCG_CAC_STATUS; \
+	uint32_t DCCG_CAC_STATUS2; \
+	uint32_t DCCG_DISP_CNTL_REG; \
+	uint32_t DCCG_DS_CNTL; \
+	uint32_t DCCG_DS_DTO_INCR; \
+	uint32_t DCCG_DS_DTO_MODULO; \
+	uint32_t DCCG_DS_HW_CAL_INTERVAL; \
+	uint32_t DCCG_GTC_CNTL; \
+	uint32_t DCCG_GTC_CURRENT; \
+	uint32_t DCCG_GTC_DTO_INCR; \
+	uint32_t DCCG_GTC_DTO_MODULO; \
+	uint32_t DCCG_PERFMON_CNTL; \
+	uint32_t DCCG_PERFMON_CNTL2; \
+	uint32_t DCCG_SOFT_RESET; \
+	uint32_t DCCG_TEST_CLK_SEL; \
+	uint32_t DCCG_VSYNC_CNT_CTRL; \
+	uint32_t DCCG_VSYNC_CNT_INT_CTRL; \
+	uint32_t DCCG_VSYNC_OTG0_LATCH_VALUE; \
+	uint32_t DCCG_VSYNC_OTG1_LATCH_VALUE; \
+	uint32_t DCCG_VSYNC_OTG2_LATCH_VALUE; \
+	uint32_t DCCG_VSYNC_OTG3_LATCH_VALUE; \
+	uint32_t DCCG_VSYNC_OTG4_LATCH_VALUE; \
+	uint32_t DCCG_VSYNC_OTG5_LATCH_VALUE; \
+	uint32_t DISPCLK_CGTT_BLK_CTRL_REG; \
+	uint32_t DP_DTO_DBUF_EN; \
+	uint32_t DPIACLK_540M_DTO_MODULO; \
+	uint32_t DPIACLK_540M_DTO_PHASE; \
+	uint32_t DPIACLK_810M_DTO_MODULO; \
+	uint32_t DPIACLK_810M_DTO_PHASE; \
+	uint32_t DPIACLK_DTO_CNTL; \
+	uint32_t DPIASYMCLK_CNTL; \
+	uint32_t DPPCLK_CGTT_BLK_CTRL_REG; \
+	uint32_t DPREFCLK_CGTT_BLK_CTRL_REG; \
+	uint32_t DPREFCLK_CNTL; \
+	uint32_t DTBCLK_DTO_DBUF_EN; \
+	uint32_t FORCE_SYMCLK_DISABLE; \
+	uint32_t HDMICHARCLK0_CLOCK_CNTL; \
+	uint32_t MICROSECOND_TIME_BASE_DIV; \
+	uint32_t MILLISECOND_TIME_BASE_DIV; \
+	uint32_t OTG0_PHYPLL_PIXEL_RATE_CNTL; \
+	uint32_t OTG0_PIXEL_RATE_CNTL; \
+	uint32_t OTG1_PHYPLL_PIXEL_RATE_CNTL; \
+	uint32_t OTG1_PIXEL_RATE_CNTL; \
+	uint32_t OTG2_PHYPLL_PIXEL_RATE_CNTL; \
+	uint32_t OTG2_PIXEL_RATE_CNTL; \
+	uint32_t OTG3_PHYPLL_PIXEL_RATE_CNTL; \
+	uint32_t OTG3_PIXEL_RATE_CNTL; \
+	uint32_t PHYPLLA_PIXCLK_RESYNC_CNTL; \
+	uint32_t PHYPLLB_PIXCLK_RESYNC_CNTL; \
+	uint32_t PHYPLLC_PIXCLK_RESYNC_CNTL; \
+	uint32_t PHYPLLD_PIXCLK_RESYNC_CNTL; \
+	uint32_t PHYPLLE_PIXCLK_RESYNC_CNTL; \
+	uint32_t REFCLK_CGTT_BLK_CTRL_REG; \
+	uint32_t SOCCLK_CGTT_BLK_CTRL_REG; \
+	uint32_t SYMCLK_CGTT_BLK_CTRL_REG; \
+	uint32_t SYMCLK_PSP_CNTL
+
+struct dccg_registers {
+	DCCG_REG_VARIABLE_LIST;
+};
+
+struct dcn_dccg {
+	struct dccg base;
+	const struct dccg_registers *regs;
+	const struct dccg_shift *dccg_shift;
+	const struct dccg_mask *dccg_mask;
+};
+
+void dccg2_update_dpp_dto(struct dccg *dccg, int dpp_inst, int req_dppclk);
+
+void dccg2_get_dccg_ref_freq(struct dccg *dccg,
+		unsigned int xtalin_freq_inKhz,
+		unsigned int *dccg_ref_freq_inKhz);
+
+void dccg2_set_fifo_errdet_ovr_en(struct dccg *dccg,
+		bool en);
+void dccg2_otg_add_pixel(struct dccg *dccg,
+		uint32_t otg_inst);
+void dccg2_otg_drop_pixel(struct dccg *dccg,
+		uint32_t otg_inst);
+
+
+void dccg2_init(struct dccg *dccg);
+
+struct dccg *dccg2_create(
+	struct dc_context *ctx,
+	const struct dccg_registers *regs,
+	const struct dccg_shift *dccg_shift,
+	const struct dccg_mask *dccg_mask);
+
+void dcn_dccg_destroy(struct dccg **dccg);
+
+#endif //__DCN20_DCCG_H__
diff --git a/drivers/gpu/drm/amd/display/dc/dcn201/dcn201_dccg.c b/drivers/gpu/drm/amd/display/dc/dccg/dcn201/dcn201_dccg.c
index f5bf04f7da25..9a3402148fde 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn201/dcn201_dccg.c
+++ b/drivers/gpu/drm/amd/display/dc/dccg/dcn201/dcn201_dccg.c
@@ -44,7 +44,8 @@
 #define DC_LOGGER \
 	dccg->ctx->logger
 
-void dccg201_update_dpp_dto(struct dccg *dccg, int dpp_inst, int req_dppclk)
+static void dccg201_update_dpp_dto(struct dccg *dccg, int dpp_inst,
+				   int req_dppclk)
 {
 	/* vbios handles it */
 }
diff --git a/drivers/gpu/drm/amd/display/dc/dcn201/dcn201_dccg.h b/drivers/gpu/drm/amd/display/dc/dccg/dcn201/dcn201_dccg.h
index 80888b0484fb..80888b0484fb 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn201/dcn201_dccg.h
+++ b/drivers/gpu/drm/amd/display/dc/dccg/dcn201/dcn201_dccg.h
diff --git a/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_dccg.c b/drivers/gpu/drm/amd/display/dc/dccg/dcn21/dcn21_dccg.c
index 33fc9aa8621b..d07c04458d31 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_dccg.c
+++ b/drivers/gpu/drm/amd/display/dc/dccg/dcn21/dcn21_dccg.c
@@ -43,7 +43,7 @@
 #define DC_LOGGER \
 	dccg->ctx->logger
 
-void dccg21_update_dpp_dto(struct dccg *dccg, int dpp_inst, int req_dppclk)
+static void dccg21_update_dpp_dto(struct dccg *dccg, int dpp_inst, int req_dppclk)
 {
 	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
 
diff --git a/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_dccg.h b/drivers/gpu/drm/amd/display/dc/dccg/dcn21/dcn21_dccg.h
index e44a37491c1e..b7efa777ec73 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_dccg.h
+++ b/drivers/gpu/drm/amd/display/dc/dccg/dcn21/dcn21_dccg.h
@@ -32,6 +32,5 @@ struct dccg *dccg21_create(
 	const struct dccg_shift *dccg_shift,
 	const struct dccg_mask *dccg_mask);
 
-void dccg21_update_dpp_dto(struct dccg *dccg, int dpp_inst, int req_dppclk);
 
 #endif /* __DCN21_DCCG_H__ */
diff --git a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_dccg.c b/drivers/gpu/drm/amd/display/dc/dccg/dcn30/dcn30_dccg.c
index d445dfefc047..d445dfefc047 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_dccg.c
+++ b/drivers/gpu/drm/amd/display/dc/dccg/dcn30/dcn30_dccg.c
diff --git a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_dccg.h b/drivers/gpu/drm/amd/display/dc/dccg/dcn30/dcn30_dccg.h
index 35a613bb08bf..3f1da7f3a91c 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_dccg.h
+++ b/drivers/gpu/drm/amd/display/dc/dccg/dcn30/dcn30_dccg.h
@@ -29,15 +29,9 @@
 #include "dcn20/dcn20_dccg.h"
 
 
-#define DCCG_REG_LIST_DCN3AG() \
-	DCCG_COMMON_REG_LIST_DCN_BASE(),\
-	SR(PHYASYMCLK_CLOCK_CNTL),\
-	SR(PHYBSYMCLK_CLOCK_CNTL),\
-	SR(PHYCSYMCLK_CLOCK_CNTL)
-
-
 #define DCCG_REG_LIST_DCN30() \
 	DCCG_REG_LIST_DCN2(),\
+	DCCG_SRII(CLOCK_CNTL, HDMICHARCLK, 0),\
 	DCCG_SRII(PIXEL_RATE_CNTL, OTG, 2),\
 	DCCG_SRII(PIXEL_RATE_CNTL, OTG, 3),\
 	DCCG_SRII(PIXEL_RATE_CNTL, OTG, 4),\
@@ -46,19 +40,10 @@
 	SR(PHYBSYMCLK_CLOCK_CNTL),\
 	SR(PHYCSYMCLK_CLOCK_CNTL)
 
-#define DCCG_MASK_SH_LIST_DCN3AG(mask_sh) \
-	DCCG_MASK_SH_LIST_DCN2_1(mask_sh),\
-	DCCG_SF(HDMICHARCLK0_CLOCK_CNTL, HDMICHARCLK0_EN, mask_sh),\
-	DCCG_SF(HDMICHARCLK0_CLOCK_CNTL, HDMICHARCLK0_SRC_SEL, mask_sh),\
-	DCCG_SF(PHYASYMCLK_CLOCK_CNTL, PHYASYMCLK_FORCE_EN, mask_sh),\
-	DCCG_SF(PHYASYMCLK_CLOCK_CNTL, PHYASYMCLK_FORCE_SRC_SEL, mask_sh),\
-	DCCG_SF(PHYBSYMCLK_CLOCK_CNTL, PHYBSYMCLK_FORCE_EN, mask_sh),\
-	DCCG_SF(PHYBSYMCLK_CLOCK_CNTL, PHYBSYMCLK_FORCE_SRC_SEL, mask_sh),\
-	DCCG_SF(PHYCSYMCLK_CLOCK_CNTL, PHYCSYMCLK_FORCE_EN, mask_sh),\
-	DCCG_SF(PHYCSYMCLK_CLOCK_CNTL, PHYCSYMCLK_FORCE_SRC_SEL, mask_sh)
-
 #define DCCG_MASK_SH_LIST_DCN3(mask_sh) \
 	DCCG_MASK_SH_LIST_DCN2(mask_sh),\
+	DCCG_SF(HDMICHARCLK0_CLOCK_CNTL, HDMICHARCLK0_EN, mask_sh),\
+	DCCG_SF(HDMICHARCLK0_CLOCK_CNTL, HDMICHARCLK0_SRC_SEL, mask_sh),\
 	DCCG_SF(PHYASYMCLK_CLOCK_CNTL, PHYASYMCLK_FORCE_EN, mask_sh),\
 	DCCG_SF(PHYASYMCLK_CLOCK_CNTL, PHYASYMCLK_FORCE_SRC_SEL, mask_sh),\
 	DCCG_SF(PHYBSYMCLK_CLOCK_CNTL, PHYBSYMCLK_FORCE_EN, mask_sh),\
diff --git a/drivers/gpu/drm/amd/display/dc/dcn301/dcn301_dccg.c b/drivers/gpu/drm/amd/display/dc/dccg/dcn301/dcn301_dccg.c
index 97e9be87afd9..97e9be87afd9 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn301/dcn301_dccg.c
+++ b/drivers/gpu/drm/amd/display/dc/dccg/dcn301/dcn301_dccg.c
diff --git a/drivers/gpu/drm/amd/display/dc/dcn301/dcn301_dccg.h b/drivers/gpu/drm/amd/display/dc/dccg/dcn301/dcn301_dccg.h
index 73db962dbc03..067e49cb238e 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn301/dcn301_dccg.h
+++ b/drivers/gpu/drm/amd/display/dc/dccg/dcn301/dcn301_dccg.h
@@ -56,10 +56,4 @@ struct dccg *dccg301_create(
 	const struct dccg_shift *dccg_shift,
 	const struct dccg_mask *dccg_mask);
 
-struct dccg *dccg301_create(
-	struct dc_context *ctx,
-	const struct dccg_registers *regs,
-	const struct dccg_shift *dccg_shift,
-	const struct dccg_mask *dccg_mask);
-
 #endif //__DCN301_DCCG_H__
diff --git a/drivers/gpu/drm/amd/display/dc/dcn302/dcn302_dccg.h b/drivers/gpu/drm/amd/display/dc/dccg/dcn302/dcn302_dccg.h
index c884dde1bb25..c884dde1bb25 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn302/dcn302_dccg.h
+++ b/drivers/gpu/drm/amd/display/dc/dccg/dcn302/dcn302_dccg.h
diff --git a/drivers/gpu/drm/amd/display/dc/dccg/dcn303/dcn303_dccg.h b/drivers/gpu/drm/amd/display/dc/dccg/dcn303/dcn303_dccg.h
new file mode 100644
index 000000000000..2e12fb643005
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dccg/dcn303/dcn303_dccg.h
@@ -0,0 +1,64 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright (C) 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ */
+
+#ifndef __DCN303_DCCG_H__
+#define __DCN303_DCCG_H__
+
+#include "dcn30/dcn30_dccg.h"
+
+
+#define DCCG_REG_LIST_DCN3_03() \
+	SR(DPPCLK_DTO_CTRL),\
+	DCCG_SRII(DTO_PARAM, DPPCLK, 0),\
+	DCCG_SRII(DTO_PARAM, DPPCLK, 1),\
+	SR(REFCLK_CNTL),\
+	SR(DISPCLK_FREQ_CHANGE_CNTL),\
+	DCCG_SRII(PIXEL_RATE_CNTL, OTG, 0),\
+	DCCG_SRII(PIXEL_RATE_CNTL, OTG, 1)
+
+
+#define DCCG_MASK_SH_LIST_DCN3_03(mask_sh) \
+		DCCG_SFI(DPPCLK_DTO_CTRL, DTO_ENABLE, DPPCLK, 0, mask_sh),\
+		DCCG_SFI(DPPCLK_DTO_CTRL, DTO_DB_EN, DPPCLK, 0, mask_sh),\
+		DCCG_SFI(DPPCLK_DTO_CTRL, DTO_ENABLE, DPPCLK, 1, mask_sh),\
+		DCCG_SFI(DPPCLK_DTO_CTRL, DTO_DB_EN, DPPCLK, 1, mask_sh),\
+		DCCG_SF(DPPCLK0_DTO_PARAM, DPPCLK0_DTO_PHASE, mask_sh),\
+		DCCG_SF(DPPCLK0_DTO_PARAM, DPPCLK0_DTO_MODULO, mask_sh),\
+		DCCG_SF(REFCLK_CNTL, REFCLK_CLOCK_EN, mask_sh),\
+		DCCG_SF(REFCLK_CNTL, REFCLK_SRC_SEL, mask_sh),\
+		DCCG_SF(DISPCLK_FREQ_CHANGE_CNTL, DISPCLK_STEP_DELAY, mask_sh),\
+		DCCG_SF(DISPCLK_FREQ_CHANGE_CNTL, DISPCLK_STEP_SIZE, mask_sh),\
+		DCCG_SF(DISPCLK_FREQ_CHANGE_CNTL, DISPCLK_FREQ_RAMP_DONE, mask_sh),\
+		DCCG_SF(DISPCLK_FREQ_CHANGE_CNTL, DISPCLK_MAX_ERRDET_CYCLES, mask_sh),\
+		DCCG_SF(DISPCLK_FREQ_CHANGE_CNTL, DCCG_FIFO_ERRDET_RESET, mask_sh),\
+		DCCG_SF(DISPCLK_FREQ_CHANGE_CNTL, DCCG_FIFO_ERRDET_STATE, mask_sh),\
+		DCCG_SF(DISPCLK_FREQ_CHANGE_CNTL, DCCG_FIFO_ERRDET_OVR_EN, mask_sh),\
+		DCCG_SF(DISPCLK_FREQ_CHANGE_CNTL, DISPCLK_CHG_FWD_CORR_DISABLE, mask_sh),\
+		DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, ADD_PIXEL, 0, mask_sh),\
+		DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, ADD_PIXEL, 1, mask_sh),\
+		DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, DROP_PIXEL, 0, mask_sh),\
+		DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, DROP_PIXEL, 1, mask_sh)
+
+#endif //__DCN303_DCCG_H__
diff --git a/drivers/gpu/drm/amd/display/dc/dccg/dcn31/dcn31_dccg.c b/drivers/gpu/drm/amd/display/dc/dccg/dcn31/dcn31_dccg.c
new file mode 100644
index 000000000000..97df04b7e39d
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dccg/dcn31/dcn31_dccg.c
@@ -0,0 +1,878 @@
+/*
+ * Copyright 2018 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#include "reg_helper.h"
+#include "core_types.h"
+#include "dcn31_dccg.h"
+#include "dal_asic_id.h"
+
+#define TO_DCN_DCCG(dccg)\
+	container_of(dccg, struct dcn_dccg, base)
+
+#define REG(reg) \
+	(dccg_dcn->regs->reg)
+
+#undef FN
+#define FN(reg_name, field_name) \
+	dccg_dcn->dccg_shift->field_name, dccg_dcn->dccg_mask->field_name
+
+#define CTX \
+	dccg_dcn->base.ctx
+#define DC_LOGGER \
+	dccg->ctx->logger
+
+void dccg31_update_dpp_dto(struct dccg *dccg, int dpp_inst, int req_dppclk)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	if (dccg->dpp_clock_gated[dpp_inst]) {
+		/*
+		 * Do not update the DPPCLK DTO if the clock is stopped.
+		 * It is treated the same as if the pipe itself were in PG.
+		 */
+		return;
+	}
+
+	if (dccg->ref_dppclk && req_dppclk) {
+		int ref_dppclk = dccg->ref_dppclk;
+		int modulo, phase;
+
+		// phase / modulo = dpp pipe clk / dpp global clk
+		modulo = 0xff;   // use FF at the end
+		phase = ((modulo * req_dppclk) + ref_dppclk - 1) / ref_dppclk;
+
+		if (phase > 0xff) {
+			ASSERT(false);
+			phase = 0xff;
+		}
+
+		REG_SET_2(DPPCLK_DTO_PARAM[dpp_inst], 0,
+				DPPCLK0_DTO_PHASE, phase,
+				DPPCLK0_DTO_MODULO, modulo);
+		REG_UPDATE(DPPCLK_DTO_CTRL,
+				DPPCLK_DTO_ENABLE[dpp_inst], 1);
+	} else {
+		REG_UPDATE(DPPCLK_DTO_CTRL,
+				DPPCLK_DTO_ENABLE[dpp_inst], 0);
+	}
+	dccg->pipe_dppclk_khz[dpp_inst] = req_dppclk;
+}
+
+static enum phyd32clk_clock_source get_phy_mux_symclk(
+		struct dcn_dccg *dccg_dcn,
+		enum phyd32clk_clock_source src)
+{
+	if (dccg_dcn->base.ctx->asic_id.chip_family == FAMILY_YELLOW_CARP &&
+			dccg_dcn->base.ctx->asic_id.hw_internal_rev == YELLOW_CARP_B0) {
+		if (src == PHYD32CLKC)
+			src = PHYD32CLKF;
+		if (src == PHYD32CLKD)
+			src = PHYD32CLKG;
+	}
+	return src;
+}
+
+static void dccg31_enable_dpstreamclk(struct dccg *dccg, int otg_inst)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	/* enabled to select one of the DTBCLKs for pipe */
+	switch (otg_inst) {
+	case 0:
+		REG_UPDATE(DPSTREAMCLK_CNTL,
+				DPSTREAMCLK_PIPE0_EN, 1);
+		break;
+	case 1:
+		REG_UPDATE(DPSTREAMCLK_CNTL,
+				DPSTREAMCLK_PIPE1_EN, 1);
+		break;
+	case 2:
+		REG_UPDATE(DPSTREAMCLK_CNTL,
+				DPSTREAMCLK_PIPE2_EN, 1);
+		break;
+	case 3:
+		REG_UPDATE(DPSTREAMCLK_CNTL,
+				DPSTREAMCLK_PIPE3_EN, 1);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+	if (dccg->ctx->dc->debug.root_clock_optimization.bits.dpstream)
+		REG_UPDATE_2(DCCG_GATE_DISABLE_CNTL3,
+			DPSTREAMCLK_GATE_DISABLE, 1,
+			DPSTREAMCLK_ROOT_GATE_DISABLE, 1);
+}
+
+static void dccg31_disable_dpstreamclk(struct dccg *dccg, int otg_inst)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	if (dccg->ctx->dc->debug.root_clock_optimization.bits.dpstream)
+		REG_UPDATE_2(DCCG_GATE_DISABLE_CNTL3,
+				DPSTREAMCLK_ROOT_GATE_DISABLE, 0,
+				DPSTREAMCLK_GATE_DISABLE, 0);
+
+	switch (otg_inst) {
+	case 0:
+		REG_UPDATE(DPSTREAMCLK_CNTL,
+				DPSTREAMCLK_PIPE0_EN, 0);
+		break;
+	case 1:
+		REG_UPDATE(DPSTREAMCLK_CNTL,
+				DPSTREAMCLK_PIPE1_EN, 0);
+		break;
+	case 2:
+		REG_UPDATE(DPSTREAMCLK_CNTL,
+				DPSTREAMCLK_PIPE2_EN, 0);
+		break;
+	case 3:
+		REG_UPDATE(DPSTREAMCLK_CNTL,
+				DPSTREAMCLK_PIPE3_EN, 0);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+}
+
+void dccg31_set_dpstreamclk(
+		struct dccg *dccg,
+		enum streamclk_source src,
+		int otg_inst,
+		int dp_hpo_inst)
+{
+	if (src == REFCLK)
+		dccg31_disable_dpstreamclk(dccg, otg_inst);
+	else
+		dccg31_enable_dpstreamclk(dccg, otg_inst);
+}
+
+void dccg31_enable_symclk32_se(
+		struct dccg *dccg,
+		int hpo_se_inst,
+		enum phyd32clk_clock_source phyd32clk)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	phyd32clk = get_phy_mux_symclk(dccg_dcn, phyd32clk);
+
+	/* select one of the PHYD32CLKs as the source for symclk32_se */
+	switch (hpo_se_inst) {
+	case 0:
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se)
+			REG_UPDATE_2(DCCG_GATE_DISABLE_CNTL3,
+					SYMCLK32_SE0_GATE_DISABLE, 1,
+					SYMCLK32_ROOT_SE0_GATE_DISABLE, 1);
+		REG_UPDATE_2(SYMCLK32_SE_CNTL,
+				SYMCLK32_SE0_SRC_SEL, phyd32clk,
+				SYMCLK32_SE0_EN, 1);
+		break;
+	case 1:
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se)
+			REG_UPDATE_2(DCCG_GATE_DISABLE_CNTL3,
+					SYMCLK32_SE1_GATE_DISABLE, 1,
+					SYMCLK32_ROOT_SE1_GATE_DISABLE, 1);
+		REG_UPDATE_2(SYMCLK32_SE_CNTL,
+				SYMCLK32_SE1_SRC_SEL, phyd32clk,
+				SYMCLK32_SE1_EN, 1);
+		break;
+	case 2:
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se)
+			REG_UPDATE_2(DCCG_GATE_DISABLE_CNTL3,
+					SYMCLK32_SE2_GATE_DISABLE, 1,
+					SYMCLK32_ROOT_SE2_GATE_DISABLE, 1);
+		REG_UPDATE_2(SYMCLK32_SE_CNTL,
+				SYMCLK32_SE2_SRC_SEL, phyd32clk,
+				SYMCLK32_SE2_EN, 1);
+		break;
+	case 3:
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se)
+			REG_UPDATE_2(DCCG_GATE_DISABLE_CNTL3,
+					SYMCLK32_SE3_GATE_DISABLE, 1,
+					SYMCLK32_ROOT_SE3_GATE_DISABLE, 1);
+		REG_UPDATE_2(SYMCLK32_SE_CNTL,
+				SYMCLK32_SE3_SRC_SEL, phyd32clk,
+				SYMCLK32_SE3_EN, 1);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+}
+
+void dccg31_disable_symclk32_se(
+		struct dccg *dccg,
+		int hpo_se_inst)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	/* set refclk as the source for symclk32_se */
+	switch (hpo_se_inst) {
+	case 0:
+		REG_UPDATE_2(SYMCLK32_SE_CNTL,
+				SYMCLK32_SE0_SRC_SEL, 0,
+				SYMCLK32_SE0_EN, 0);
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se)
+			REG_UPDATE_2(DCCG_GATE_DISABLE_CNTL3,
+					SYMCLK32_SE0_GATE_DISABLE, 0,
+					SYMCLK32_ROOT_SE0_GATE_DISABLE, 0);
+		break;
+	case 1:
+		REG_UPDATE_2(SYMCLK32_SE_CNTL,
+				SYMCLK32_SE1_SRC_SEL, 0,
+				SYMCLK32_SE1_EN, 0);
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se)
+			REG_UPDATE_2(DCCG_GATE_DISABLE_CNTL3,
+					SYMCLK32_SE1_GATE_DISABLE, 0,
+					SYMCLK32_ROOT_SE1_GATE_DISABLE, 0);
+		break;
+	case 2:
+		REG_UPDATE_2(SYMCLK32_SE_CNTL,
+				SYMCLK32_SE2_SRC_SEL, 0,
+				SYMCLK32_SE2_EN, 0);
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se)
+			REG_UPDATE_2(DCCG_GATE_DISABLE_CNTL3,
+					SYMCLK32_SE2_GATE_DISABLE, 0,
+					SYMCLK32_ROOT_SE2_GATE_DISABLE, 0);
+		break;
+	case 3:
+		REG_UPDATE_2(SYMCLK32_SE_CNTL,
+				SYMCLK32_SE3_SRC_SEL, 0,
+				SYMCLK32_SE3_EN, 0);
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se)
+			REG_UPDATE_2(DCCG_GATE_DISABLE_CNTL3,
+					SYMCLK32_SE3_GATE_DISABLE, 0,
+					SYMCLK32_ROOT_SE3_GATE_DISABLE, 0);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+}
+
+void dccg31_enable_symclk32_le(
+		struct dccg *dccg,
+		int hpo_le_inst,
+		enum phyd32clk_clock_source phyd32clk)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	phyd32clk = get_phy_mux_symclk(dccg_dcn, phyd32clk);
+
+	/* select one of the PHYD32CLKs as the source for symclk32_le */
+	switch (hpo_le_inst) {
+	case 0:
+		REG_UPDATE_2(SYMCLK32_LE_CNTL,
+				SYMCLK32_LE0_SRC_SEL, phyd32clk,
+				SYMCLK32_LE0_EN, 1);
+		break;
+	case 1:
+		REG_UPDATE_2(SYMCLK32_LE_CNTL,
+				SYMCLK32_LE1_SRC_SEL, phyd32clk,
+				SYMCLK32_LE1_EN, 1);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+}
+
+void dccg31_disable_symclk32_le(
+		struct dccg *dccg,
+		int hpo_le_inst)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	/* set refclk as the source for symclk32_le */
+	switch (hpo_le_inst) {
+	case 0:
+		REG_UPDATE_2(SYMCLK32_LE_CNTL,
+				SYMCLK32_LE0_SRC_SEL, 0,
+				SYMCLK32_LE0_EN, 0);
+		break;
+	case 1:
+		REG_UPDATE_2(SYMCLK32_LE_CNTL,
+				SYMCLK32_LE1_SRC_SEL, 0,
+				SYMCLK32_LE1_EN, 0);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+}
+
+void dccg31_set_symclk32_le_root_clock_gating(
+		struct dccg *dccg,
+		int hpo_le_inst,
+		bool enable)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	if (!dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_le)
+		return;
+
+	switch (hpo_le_inst) {
+	case 0:
+		REG_UPDATE_2(DCCG_GATE_DISABLE_CNTL3,
+				SYMCLK32_LE0_GATE_DISABLE, enable ? 1 : 0,
+				SYMCLK32_ROOT_LE0_GATE_DISABLE, enable ? 1 : 0);
+		break;
+	case 1:
+		REG_UPDATE_2(DCCG_GATE_DISABLE_CNTL3,
+				SYMCLK32_LE1_GATE_DISABLE, enable ? 1 : 0,
+				SYMCLK32_ROOT_LE1_GATE_DISABLE, enable ? 1 : 0);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+}
+
+void dccg31_disable_dscclk(struct dccg *dccg, int inst)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	if (!dccg->ctx->dc->debug.root_clock_optimization.bits.dsc)
+		return;
+	//DTO must be enabled to generate a 0 Hz clock output
+	switch (inst) {
+	case 0:
+		REG_UPDATE(DSCCLK_DTO_CTRL,
+				DSCCLK0_DTO_ENABLE, 1);
+		REG_UPDATE_2(DSCCLK0_DTO_PARAM,
+				DSCCLK0_DTO_PHASE, 0,
+				DSCCLK0_DTO_MODULO, 1);
+		break;
+	case 1:
+		REG_UPDATE(DSCCLK_DTO_CTRL,
+				DSCCLK1_DTO_ENABLE, 1);
+		REG_UPDATE_2(DSCCLK1_DTO_PARAM,
+				DSCCLK1_DTO_PHASE, 0,
+				DSCCLK1_DTO_MODULO, 1);
+		break;
+	case 2:
+		REG_UPDATE(DSCCLK_DTO_CTRL,
+				DSCCLK2_DTO_ENABLE, 1);
+		REG_UPDATE_2(DSCCLK2_DTO_PARAM,
+				DSCCLK2_DTO_PHASE, 0,
+				DSCCLK2_DTO_MODULO, 1);
+		break;
+	case 3:
+		if (REG(DSCCLK3_DTO_PARAM)) {
+			REG_UPDATE(DSCCLK_DTO_CTRL,
+					DSCCLK3_DTO_ENABLE, 1);
+			REG_UPDATE_2(DSCCLK3_DTO_PARAM,
+					DSCCLK3_DTO_PHASE, 0,
+					DSCCLK3_DTO_MODULO, 1);
+		}
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+}
+
+void dccg31_enable_dscclk(struct dccg *dccg, int inst)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	if (!dccg->ctx->dc->debug.root_clock_optimization.bits.dsc)
+		return;
+	//Disable DTO
+	switch (inst) {
+	case 0:
+		REG_UPDATE_2(DSCCLK0_DTO_PARAM,
+				DSCCLK0_DTO_PHASE, 0,
+				DSCCLK0_DTO_MODULO, 0);
+		REG_UPDATE(DSCCLK_DTO_CTRL,
+				DSCCLK0_DTO_ENABLE, 0);
+		break;
+	case 1:
+		REG_UPDATE_2(DSCCLK1_DTO_PARAM,
+				DSCCLK1_DTO_PHASE, 0,
+				DSCCLK1_DTO_MODULO, 0);
+		REG_UPDATE(DSCCLK_DTO_CTRL,
+				DSCCLK1_DTO_ENABLE, 0);
+		break;
+	case 2:
+		REG_UPDATE_2(DSCCLK2_DTO_PARAM,
+				DSCCLK2_DTO_PHASE, 0,
+				DSCCLK2_DTO_MODULO, 0);
+		REG_UPDATE(DSCCLK_DTO_CTRL,
+				DSCCLK2_DTO_ENABLE, 0);
+		break;
+	case 3:
+		if (REG(DSCCLK3_DTO_PARAM)) {
+			REG_UPDATE(DSCCLK_DTO_CTRL,
+					DSCCLK3_DTO_ENABLE, 0);
+			REG_UPDATE_2(DSCCLK3_DTO_PARAM,
+					DSCCLK3_DTO_PHASE, 0,
+					DSCCLK3_DTO_MODULO, 0);
+		}
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+}
+
+void dccg31_set_physymclk(
+		struct dccg *dccg,
+		int phy_inst,
+		enum physymclk_clock_source clk_src,
+		bool force_enable)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	/* Force PHYSYMCLK on and Select phyd32clk as the source of clock which is output to PHY through DCIO */
+	switch (phy_inst) {
+	case 0:
+		if (force_enable) {
+			REG_UPDATE_2(PHYASYMCLK_CLOCK_CNTL,
+					PHYASYMCLK_FORCE_EN, 1,
+					PHYASYMCLK_FORCE_SRC_SEL, clk_src);
+			if (dccg->ctx->dc->debug.root_clock_optimization.bits.physymclk)
+				REG_UPDATE(DCCG_GATE_DISABLE_CNTL2,
+					PHYASYMCLK_GATE_DISABLE, 1);
+		} else {
+			REG_UPDATE_2(PHYASYMCLK_CLOCK_CNTL,
+					PHYASYMCLK_FORCE_EN, 0,
+					PHYASYMCLK_FORCE_SRC_SEL, 0);
+			if (dccg->ctx->dc->debug.root_clock_optimization.bits.physymclk)
+				REG_UPDATE(DCCG_GATE_DISABLE_CNTL2,
+					PHYASYMCLK_GATE_DISABLE, 0);
+		}
+		break;
+	case 1:
+		if (force_enable) {
+			REG_UPDATE_2(PHYBSYMCLK_CLOCK_CNTL,
+					PHYBSYMCLK_FORCE_EN, 1,
+					PHYBSYMCLK_FORCE_SRC_SEL, clk_src);
+			if (dccg->ctx->dc->debug.root_clock_optimization.bits.physymclk)
+				REG_UPDATE(DCCG_GATE_DISABLE_CNTL2,
+					PHYBSYMCLK_GATE_DISABLE, 1);
+		} else {
+			REG_UPDATE_2(PHYBSYMCLK_CLOCK_CNTL,
+					PHYBSYMCLK_FORCE_EN, 0,
+					PHYBSYMCLK_FORCE_SRC_SEL, 0);
+			if (dccg->ctx->dc->debug.root_clock_optimization.bits.physymclk)
+				REG_UPDATE(DCCG_GATE_DISABLE_CNTL2,
+					PHYBSYMCLK_GATE_DISABLE, 0);
+		}
+		break;
+	case 2:
+		if (force_enable) {
+			REG_UPDATE_2(PHYCSYMCLK_CLOCK_CNTL,
+					PHYCSYMCLK_FORCE_EN, 1,
+					PHYCSYMCLK_FORCE_SRC_SEL, clk_src);
+			if (dccg->ctx->dc->debug.root_clock_optimization.bits.physymclk)
+				REG_UPDATE(DCCG_GATE_DISABLE_CNTL2,
+					PHYCSYMCLK_GATE_DISABLE, 1);
+		} else {
+			REG_UPDATE_2(PHYCSYMCLK_CLOCK_CNTL,
+					PHYCSYMCLK_FORCE_EN, 0,
+					PHYCSYMCLK_FORCE_SRC_SEL, 0);
+			if (dccg->ctx->dc->debug.root_clock_optimization.bits.physymclk)
+				REG_UPDATE(DCCG_GATE_DISABLE_CNTL2,
+					PHYCSYMCLK_GATE_DISABLE, 0);
+		}
+		break;
+	case 3:
+		if (force_enable) {
+			REG_UPDATE_2(PHYDSYMCLK_CLOCK_CNTL,
+					PHYDSYMCLK_FORCE_EN, 1,
+					PHYDSYMCLK_FORCE_SRC_SEL, clk_src);
+			if (dccg->ctx->dc->debug.root_clock_optimization.bits.physymclk)
+				REG_UPDATE(DCCG_GATE_DISABLE_CNTL2,
+					PHYDSYMCLK_GATE_DISABLE, 1);
+		} else {
+			REG_UPDATE_2(PHYDSYMCLK_CLOCK_CNTL,
+					PHYDSYMCLK_FORCE_EN, 0,
+					PHYDSYMCLK_FORCE_SRC_SEL, 0);
+			if (dccg->ctx->dc->debug.root_clock_optimization.bits.physymclk)
+				REG_UPDATE(DCCG_GATE_DISABLE_CNTL2,
+					PHYDSYMCLK_GATE_DISABLE, 0);
+		}
+		break;
+	case 4:
+		if (force_enable) {
+			REG_UPDATE_2(PHYESYMCLK_CLOCK_CNTL,
+					PHYESYMCLK_FORCE_EN, 1,
+					PHYESYMCLK_FORCE_SRC_SEL, clk_src);
+			if (dccg->ctx->dc->debug.root_clock_optimization.bits.physymclk)
+				REG_UPDATE(DCCG_GATE_DISABLE_CNTL2,
+					PHYESYMCLK_GATE_DISABLE, 1);
+		} else {
+			REG_UPDATE_2(PHYESYMCLK_CLOCK_CNTL,
+					PHYESYMCLK_FORCE_EN, 0,
+					PHYESYMCLK_FORCE_SRC_SEL, 0);
+			if (dccg->ctx->dc->debug.root_clock_optimization.bits.physymclk)
+				REG_UPDATE(DCCG_GATE_DISABLE_CNTL2,
+					PHYESYMCLK_GATE_DISABLE, 0);
+		}
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+}
+
+/* Controls the generation of pixel valid for OTG in (OTG -> HPO case) */
+void dccg31_set_dtbclk_dto(
+		struct dccg *dccg,
+		const struct dtbclk_dto_params *params)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+	int req_dtbclk_khz = params->pixclk_khz;
+	uint32_t dtbdto_div;
+
+	/* Mode	                DTBDTO Rate       DTBCLK_DTO<x>_DIV Register
+	 * ODM 4:1 combine      pixel rate/4      2
+	 * ODM 2:1 combine      pixel rate/2      4
+	 * non-DSC 4:2:0 mode   pixel rate/2      4
+	 * DSC native 4:2:0     pixel rate/2      4
+	 * DSC native 4:2:2     pixel rate/2      4
+	 * Other modes          pixel rate        8
+	 */
+	if (params->num_odm_segments == 4) {
+		dtbdto_div = 2;
+		req_dtbclk_khz = params->pixclk_khz / 4;
+	} else if ((params->num_odm_segments == 2) ||
+			(params->timing->pixel_encoding == PIXEL_ENCODING_YCBCR420) ||
+			(params->timing->flags.DSC && params->timing->pixel_encoding == PIXEL_ENCODING_YCBCR422
+					&& !params->timing->dsc_cfg.ycbcr422_simple)) {
+		dtbdto_div = 4;
+		req_dtbclk_khz = params->pixclk_khz / 2;
+	} else
+		dtbdto_div = 8;
+
+	if (params->ref_dtbclk_khz && req_dtbclk_khz) {
+		uint32_t modulo, phase;
+
+		// phase / modulo = dtbclk / dtbclk ref
+		modulo = params->ref_dtbclk_khz * 1000;
+		phase = div_u64((((unsigned long long)modulo * req_dtbclk_khz) + params->ref_dtbclk_khz - 1),
+				params->ref_dtbclk_khz);
+
+		REG_UPDATE(OTG_PIXEL_RATE_CNTL[params->otg_inst],
+				DTBCLK_DTO_DIV[params->otg_inst], dtbdto_div);
+
+		REG_WRITE(DTBCLK_DTO_MODULO[params->otg_inst], modulo);
+		REG_WRITE(DTBCLK_DTO_PHASE[params->otg_inst], phase);
+
+		REG_UPDATE(OTG_PIXEL_RATE_CNTL[params->otg_inst],
+				DTBCLK_DTO_ENABLE[params->otg_inst], 1);
+
+		REG_WAIT(OTG_PIXEL_RATE_CNTL[params->otg_inst],
+				DTBCLKDTO_ENABLE_STATUS[params->otg_inst], 1,
+				1, 100);
+
+		/* The recommended programming sequence to enable DTBCLK DTO to generate
+		 * valid pixel HPO DPSTREAM ENCODER, specifies that DTO source select should
+		 * be set only after DTO is enabled
+		 */
+		REG_UPDATE(OTG_PIXEL_RATE_CNTL[params->otg_inst],
+				PIPE_DTO_SRC_SEL[params->otg_inst], 1);
+	} else {
+		REG_UPDATE_3(OTG_PIXEL_RATE_CNTL[params->otg_inst],
+				DTBCLK_DTO_ENABLE[params->otg_inst], 0,
+				PIPE_DTO_SRC_SEL[params->otg_inst], 0,
+				DTBCLK_DTO_DIV[params->otg_inst], dtbdto_div);
+
+		REG_WRITE(DTBCLK_DTO_MODULO[params->otg_inst], 0);
+		REG_WRITE(DTBCLK_DTO_PHASE[params->otg_inst], 0);
+	}
+}
+
+void dccg31_set_audio_dtbclk_dto(
+		struct dccg *dccg,
+		const struct dtbclk_dto_params *params)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	if (params->ref_dtbclk_khz && params->req_audio_dtbclk_khz) {
+		uint32_t modulo, phase;
+
+		// phase / modulo = dtbclk / dtbclk ref
+		modulo = params->ref_dtbclk_khz * 1000;
+		phase = div_u64((((unsigned long long)modulo * params->req_audio_dtbclk_khz) + params->ref_dtbclk_khz - 1),
+			params->ref_dtbclk_khz);
+
+
+		REG_WRITE(DCCG_AUDIO_DTBCLK_DTO_MODULO, modulo);
+		REG_WRITE(DCCG_AUDIO_DTBCLK_DTO_PHASE, phase);
+
+		//REG_UPDATE(DCCG_AUDIO_DTO_SOURCE,
+		//		DCCG_AUDIO_DTBCLK_DTO_USE_512FBR_DTO, 1);
+
+		REG_UPDATE(DCCG_AUDIO_DTO_SOURCE,
+				DCCG_AUDIO_DTO_SEL, 4);  //  04 - DCCG_AUDIO_DTO_SEL_AUDIO_DTO_DTBCLK
+	} else {
+		REG_WRITE(DCCG_AUDIO_DTBCLK_DTO_PHASE, 0);
+		REG_WRITE(DCCG_AUDIO_DTBCLK_DTO_MODULO, 0);
+
+		REG_UPDATE(DCCG_AUDIO_DTO_SOURCE,
+				DCCG_AUDIO_DTO_SEL, 3);  //  03 - DCCG_AUDIO_DTO_SEL_NO_AUDIO_DTO
+	}
+}
+
+void dccg31_get_dccg_ref_freq(struct dccg *dccg,
+		unsigned int xtalin_freq_inKhz,
+		unsigned int *dccg_ref_freq_inKhz)
+{
+	/*
+	 * Assume refclk is sourced from xtalin
+	 * expect 24MHz
+	 */
+	*dccg_ref_freq_inKhz = xtalin_freq_inKhz;
+	return;
+}
+
+void dccg31_set_dispclk_change_mode(
+	struct dccg *dccg,
+	enum dentist_dispclk_change_mode change_mode)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	REG_UPDATE(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_CHG_MODE,
+		   change_mode == DISPCLK_CHANGE_MODE_RAMPING ? 2 : 0);
+}
+
+void dccg31_init(struct dccg *dccg)
+{
+	/* Set HPO stream encoder to use refclk to avoid case where PHY is
+	 * disabled and SYMCLK32 for HPO SE is sourced from PHYD32CLK which
+	 * will cause DCN to hang.
+	 */
+	dccg31_disable_symclk32_se(dccg, 0);
+	dccg31_disable_symclk32_se(dccg, 1);
+	dccg31_disable_symclk32_se(dccg, 2);
+	dccg31_disable_symclk32_se(dccg, 3);
+
+	dccg31_set_symclk32_le_root_clock_gating(dccg, 0, false);
+	dccg31_set_symclk32_le_root_clock_gating(dccg, 1, false);
+
+	if (dccg->ctx->dc->debug.root_clock_optimization.bits.dpstream) {
+		dccg31_disable_dpstreamclk(dccg, 0);
+		dccg31_disable_dpstreamclk(dccg, 1);
+		dccg31_disable_dpstreamclk(dccg, 2);
+		dccg31_disable_dpstreamclk(dccg, 3);
+	}
+
+	if (dccg->ctx->dc->debug.root_clock_optimization.bits.physymclk) {
+		dccg31_set_physymclk(dccg, 0, PHYSYMCLK_FORCE_SRC_SYMCLK, false);
+		dccg31_set_physymclk(dccg, 1, PHYSYMCLK_FORCE_SRC_SYMCLK, false);
+		dccg31_set_physymclk(dccg, 2, PHYSYMCLK_FORCE_SRC_SYMCLK, false);
+		dccg31_set_physymclk(dccg, 3, PHYSYMCLK_FORCE_SRC_SYMCLK, false);
+		dccg31_set_physymclk(dccg, 4, PHYSYMCLK_FORCE_SRC_SYMCLK, false);
+	}
+}
+
+void dccg31_otg_add_pixel(struct dccg *dccg,
+				 uint32_t otg_inst)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	REG_UPDATE(OTG_PIXEL_RATE_CNTL[otg_inst],
+			OTG_ADD_PIXEL[otg_inst], 1);
+}
+
+void dccg31_otg_drop_pixel(struct dccg *dccg,
+				  uint32_t otg_inst)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	REG_UPDATE(OTG_PIXEL_RATE_CNTL[otg_inst],
+			OTG_DROP_PIXEL[otg_inst], 1);
+}
+
+void dccg31_read_reg_state(struct dccg *dccg, struct dcn_dccg_reg_state *dccg_reg_state)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	dccg_reg_state->dc_mem_global_pwr_req_cntl = REG_READ(DC_MEM_GLOBAL_PWR_REQ_CNTL);
+	dccg_reg_state->dccg_audio_dtbclk_dto_modulo = REG_READ(DCCG_AUDIO_DTBCLK_DTO_MODULO);
+	dccg_reg_state->dccg_audio_dtbclk_dto_phase = REG_READ(DCCG_AUDIO_DTBCLK_DTO_PHASE);
+	dccg_reg_state->dccg_audio_dto_source = REG_READ(DCCG_AUDIO_DTO_SOURCE);
+	dccg_reg_state->dccg_audio_dto0_module = REG_READ(DCCG_AUDIO_DTO0_MODULE);
+	dccg_reg_state->dccg_audio_dto0_phase = REG_READ(DCCG_AUDIO_DTO0_PHASE);
+	dccg_reg_state->dccg_audio_dto1_module = REG_READ(DCCG_AUDIO_DTO1_MODULE);
+	dccg_reg_state->dccg_audio_dto1_phase = REG_READ(DCCG_AUDIO_DTO1_PHASE);
+	dccg_reg_state->dccg_cac_status = REG_READ(DCCG_CAC_STATUS);
+	dccg_reg_state->dccg_cac_status2 = REG_READ(DCCG_CAC_STATUS2);
+	dccg_reg_state->dccg_disp_cntl_reg = REG_READ(DCCG_DISP_CNTL_REG);
+	dccg_reg_state->dccg_ds_cntl = REG_READ(DCCG_DS_CNTL);
+	dccg_reg_state->dccg_ds_dto_incr = REG_READ(DCCG_DS_DTO_INCR);
+	dccg_reg_state->dccg_ds_dto_modulo = REG_READ(DCCG_DS_DTO_MODULO);
+	dccg_reg_state->dccg_ds_hw_cal_interval = REG_READ(DCCG_DS_HW_CAL_INTERVAL);
+	dccg_reg_state->dccg_gate_disable_cntl = REG_READ(DCCG_GATE_DISABLE_CNTL);
+	dccg_reg_state->dccg_gate_disable_cntl2 = REG_READ(DCCG_GATE_DISABLE_CNTL2);
+	dccg_reg_state->dccg_gate_disable_cntl3 = REG_READ(DCCG_GATE_DISABLE_CNTL3);
+	dccg_reg_state->dccg_gate_disable_cntl4 = REG_READ(DCCG_GATE_DISABLE_CNTL4);
+	dccg_reg_state->dccg_gate_disable_cntl5 = REG_READ(DCCG_GATE_DISABLE_CNTL5);
+	dccg_reg_state->dccg_gate_disable_cntl6 = REG_READ(DCCG_GATE_DISABLE_CNTL6);
+	dccg_reg_state->dccg_global_fgcg_rep_cntl = REG_READ(DCCG_GLOBAL_FGCG_REP_CNTL);
+	dccg_reg_state->dccg_gtc_cntl = REG_READ(DCCG_GTC_CNTL);
+	dccg_reg_state->dccg_gtc_current = REG_READ(DCCG_GTC_CURRENT);
+	dccg_reg_state->dccg_gtc_dto_incr = REG_READ(DCCG_GTC_DTO_INCR);
+	dccg_reg_state->dccg_gtc_dto_modulo = REG_READ(DCCG_GTC_DTO_MODULO);
+	dccg_reg_state->dccg_perfmon_cntl = REG_READ(DCCG_PERFMON_CNTL);
+	dccg_reg_state->dccg_perfmon_cntl2 = REG_READ(DCCG_PERFMON_CNTL2);
+	dccg_reg_state->dccg_soft_reset = REG_READ(DCCG_SOFT_RESET);
+	dccg_reg_state->dccg_test_clk_sel = REG_READ(DCCG_TEST_CLK_SEL);
+	dccg_reg_state->dccg_vsync_cnt_ctrl = REG_READ(DCCG_VSYNC_CNT_CTRL);
+	dccg_reg_state->dccg_vsync_cnt_int_ctrl = REG_READ(DCCG_VSYNC_CNT_INT_CTRL);
+	dccg_reg_state->dccg_vsync_otg0_latch_value = REG_READ(DCCG_VSYNC_OTG0_LATCH_VALUE);
+	dccg_reg_state->dccg_vsync_otg1_latch_value = REG_READ(DCCG_VSYNC_OTG1_LATCH_VALUE);
+	dccg_reg_state->dccg_vsync_otg2_latch_value = REG_READ(DCCG_VSYNC_OTG2_LATCH_VALUE);
+	dccg_reg_state->dccg_vsync_otg3_latch_value = REG_READ(DCCG_VSYNC_OTG3_LATCH_VALUE);
+	dccg_reg_state->dccg_vsync_otg4_latch_value = REG_READ(DCCG_VSYNC_OTG4_LATCH_VALUE);
+	dccg_reg_state->dccg_vsync_otg5_latch_value = REG_READ(DCCG_VSYNC_OTG5_LATCH_VALUE);
+	dccg_reg_state->dispclk_cgtt_blk_ctrl_reg = REG_READ(DISPCLK_CGTT_BLK_CTRL_REG);
+	dccg_reg_state->dispclk_freq_change_cntl = REG_READ(DISPCLK_FREQ_CHANGE_CNTL);
+	dccg_reg_state->dp_dto_dbuf_en = REG_READ(DP_DTO_DBUF_EN);
+	dccg_reg_state->dp_dto0_modulo = REG_READ(DP_DTO_MODULO[0]);
+	dccg_reg_state->dp_dto0_phase = REG_READ(DP_DTO_PHASE[0]);
+	dccg_reg_state->dp_dto1_modulo = REG_READ(DP_DTO_MODULO[1]);
+	dccg_reg_state->dp_dto1_phase = REG_READ(DP_DTO_PHASE[1]);
+	dccg_reg_state->dp_dto2_modulo = REG_READ(DP_DTO_MODULO[2]);
+	dccg_reg_state->dp_dto2_phase = REG_READ(DP_DTO_PHASE[2]);
+	dccg_reg_state->dp_dto3_modulo = REG_READ(DP_DTO_MODULO[3]);
+	dccg_reg_state->dp_dto3_phase = REG_READ(DP_DTO_PHASE[3]);
+	dccg_reg_state->dpiaclk_540m_dto_modulo = REG_READ(DPIACLK_540M_DTO_MODULO);
+	dccg_reg_state->dpiaclk_540m_dto_phase = REG_READ(DPIACLK_540M_DTO_PHASE);
+	dccg_reg_state->dpiaclk_810m_dto_modulo = REG_READ(DPIACLK_810M_DTO_MODULO);
+	dccg_reg_state->dpiaclk_810m_dto_phase = REG_READ(DPIACLK_810M_DTO_PHASE);
+	dccg_reg_state->dpiaclk_dto_cntl = REG_READ(DPIACLK_DTO_CNTL);
+	dccg_reg_state->dpiasymclk_cntl = REG_READ(DPIASYMCLK_CNTL);
+	dccg_reg_state->dppclk_cgtt_blk_ctrl_reg = REG_READ(DPPCLK_CGTT_BLK_CTRL_REG);
+	dccg_reg_state->dppclk_ctrl = REG_READ(DPPCLK_CTRL);
+	dccg_reg_state->dppclk_dto_ctrl = REG_READ(DPPCLK_DTO_CTRL);
+	dccg_reg_state->dppclk0_dto_param = REG_READ(DPPCLK_DTO_PARAM[0]);
+	dccg_reg_state->dppclk1_dto_param = REG_READ(DPPCLK_DTO_PARAM[1]);
+	dccg_reg_state->dppclk2_dto_param = REG_READ(DPPCLK_DTO_PARAM[2]);
+	dccg_reg_state->dppclk3_dto_param = REG_READ(DPPCLK_DTO_PARAM[3]);
+	dccg_reg_state->dprefclk_cgtt_blk_ctrl_reg = REG_READ(DPREFCLK_CGTT_BLK_CTRL_REG);
+	dccg_reg_state->dprefclk_cntl = REG_READ(DPREFCLK_CNTL);
+	dccg_reg_state->dpstreamclk_cntl = REG_READ(DPSTREAMCLK_CNTL);
+	dccg_reg_state->dscclk_dto_ctrl = REG_READ(DSCCLK_DTO_CTRL);
+	dccg_reg_state->dscclk0_dto_param = REG_READ(DSCCLK0_DTO_PARAM);
+	dccg_reg_state->dscclk1_dto_param = REG_READ(DSCCLK1_DTO_PARAM);
+	dccg_reg_state->dscclk2_dto_param = REG_READ(DSCCLK2_DTO_PARAM);
+	dccg_reg_state->dscclk3_dto_param = REG_READ(DSCCLK3_DTO_PARAM);
+	dccg_reg_state->dtbclk_dto_dbuf_en = REG_READ(DTBCLK_DTO_DBUF_EN);
+	dccg_reg_state->dtbclk_dto0_modulo = REG_READ(DTBCLK_DTO_MODULO[0]);
+	dccg_reg_state->dtbclk_dto0_phase = REG_READ(DTBCLK_DTO_PHASE[0]);
+	dccg_reg_state->dtbclk_dto1_modulo = REG_READ(DTBCLK_DTO_MODULO[1]);
+	dccg_reg_state->dtbclk_dto1_phase = REG_READ(DTBCLK_DTO_PHASE[1]);
+	dccg_reg_state->dtbclk_dto2_modulo = REG_READ(DTBCLK_DTO_MODULO[2]);
+	dccg_reg_state->dtbclk_dto2_phase = REG_READ(DTBCLK_DTO_PHASE[2]);
+	dccg_reg_state->dtbclk_dto3_modulo = REG_READ(DTBCLK_DTO_MODULO[3]);
+	dccg_reg_state->dtbclk_dto3_phase = REG_READ(DTBCLK_DTO_PHASE[3]);
+	dccg_reg_state->dtbclk_p_cntl = REG_READ(DTBCLK_P_CNTL);
+	dccg_reg_state->force_symclk_disable = REG_READ(FORCE_SYMCLK_DISABLE);
+	dccg_reg_state->hdmicharclk0_clock_cntl = REG_READ(HDMICHARCLK0_CLOCK_CNTL);
+	dccg_reg_state->hdmistreamclk_cntl = REG_READ(HDMISTREAMCLK_CNTL);
+	dccg_reg_state->hdmistreamclk0_dto_param = REG_READ(HDMISTREAMCLK0_DTO_PARAM);
+	dccg_reg_state->microsecond_time_base_div = REG_READ(MICROSECOND_TIME_BASE_DIV);
+	dccg_reg_state->millisecond_time_base_div = REG_READ(MILLISECOND_TIME_BASE_DIV);
+	dccg_reg_state->otg_pixel_rate_div = REG_READ(OTG_PIXEL_RATE_DIV);
+	dccg_reg_state->otg0_phypll_pixel_rate_cntl = REG_READ(OTG0_PHYPLL_PIXEL_RATE_CNTL);
+	dccg_reg_state->otg0_pixel_rate_cntl = REG_READ(OTG0_PIXEL_RATE_CNTL);
+	dccg_reg_state->otg1_phypll_pixel_rate_cntl = REG_READ(OTG1_PHYPLL_PIXEL_RATE_CNTL);
+	dccg_reg_state->otg1_pixel_rate_cntl = REG_READ(OTG1_PIXEL_RATE_CNTL);
+	dccg_reg_state->otg2_phypll_pixel_rate_cntl = REG_READ(OTG2_PHYPLL_PIXEL_RATE_CNTL);
+	dccg_reg_state->otg2_pixel_rate_cntl = REG_READ(OTG2_PIXEL_RATE_CNTL);
+	dccg_reg_state->otg3_phypll_pixel_rate_cntl = REG_READ(OTG3_PHYPLL_PIXEL_RATE_CNTL);
+	dccg_reg_state->otg3_pixel_rate_cntl = REG_READ(OTG3_PIXEL_RATE_CNTL);
+	dccg_reg_state->phyasymclk_clock_cntl = REG_READ(PHYASYMCLK_CLOCK_CNTL);
+	dccg_reg_state->phybsymclk_clock_cntl = REG_READ(PHYBSYMCLK_CLOCK_CNTL);
+	dccg_reg_state->phycsymclk_clock_cntl = REG_READ(PHYCSYMCLK_CLOCK_CNTL);
+	dccg_reg_state->phydsymclk_clock_cntl = REG_READ(PHYDSYMCLK_CLOCK_CNTL);
+	dccg_reg_state->phyesymclk_clock_cntl = REG_READ(PHYESYMCLK_CLOCK_CNTL);
+	dccg_reg_state->phyplla_pixclk_resync_cntl = REG_READ(PHYPLLA_PIXCLK_RESYNC_CNTL);
+	dccg_reg_state->phypllb_pixclk_resync_cntl = REG_READ(PHYPLLB_PIXCLK_RESYNC_CNTL);
+	dccg_reg_state->phypllc_pixclk_resync_cntl = REG_READ(PHYPLLC_PIXCLK_RESYNC_CNTL);
+	dccg_reg_state->phyplld_pixclk_resync_cntl = REG_READ(PHYPLLD_PIXCLK_RESYNC_CNTL);
+	dccg_reg_state->phyplle_pixclk_resync_cntl = REG_READ(PHYPLLE_PIXCLK_RESYNC_CNTL);
+	dccg_reg_state->refclk_cgtt_blk_ctrl_reg = REG_READ(REFCLK_CGTT_BLK_CTRL_REG);
+	dccg_reg_state->socclk_cgtt_blk_ctrl_reg = REG_READ(SOCCLK_CGTT_BLK_CTRL_REG);
+	dccg_reg_state->symclk_cgtt_blk_ctrl_reg = REG_READ(SYMCLK_CGTT_BLK_CTRL_REG);
+	dccg_reg_state->symclk_psp_cntl = REG_READ(SYMCLK_PSP_CNTL);
+	dccg_reg_state->symclk32_le_cntl = REG_READ(SYMCLK32_LE_CNTL);
+	dccg_reg_state->symclk32_se_cntl = REG_READ(SYMCLK32_SE_CNTL);
+	dccg_reg_state->symclka_clock_enable = REG_READ(SYMCLKA_CLOCK_ENABLE);
+	dccg_reg_state->symclkb_clock_enable = REG_READ(SYMCLKB_CLOCK_ENABLE);
+	dccg_reg_state->symclkc_clock_enable = REG_READ(SYMCLKC_CLOCK_ENABLE);
+	dccg_reg_state->symclkd_clock_enable = REG_READ(SYMCLKD_CLOCK_ENABLE);
+	dccg_reg_state->symclke_clock_enable = REG_READ(SYMCLKE_CLOCK_ENABLE);
+}
+
+static const struct dccg_funcs dccg31_funcs = {
+	.update_dpp_dto = dccg31_update_dpp_dto,
+	.get_dccg_ref_freq = dccg31_get_dccg_ref_freq,
+	.dccg_init = dccg31_init,
+	.set_dpstreamclk = dccg31_set_dpstreamclk,
+	.enable_symclk32_se = dccg31_enable_symclk32_se,
+	.disable_symclk32_se = dccg31_disable_symclk32_se,
+	.enable_symclk32_le = dccg31_enable_symclk32_le,
+	.disable_symclk32_le = dccg31_disable_symclk32_le,
+	.set_physymclk = dccg31_set_physymclk,
+	.set_dtbclk_dto = dccg31_set_dtbclk_dto,
+	.set_audio_dtbclk_dto = dccg31_set_audio_dtbclk_dto,
+	.set_fifo_errdet_ovr_en = dccg2_set_fifo_errdet_ovr_en,
+	.otg_add_pixel = dccg31_otg_add_pixel,
+	.otg_drop_pixel = dccg31_otg_drop_pixel,
+	.set_dispclk_change_mode = dccg31_set_dispclk_change_mode,
+	.disable_dsc = dccg31_disable_dscclk,
+	.enable_dsc = dccg31_enable_dscclk,
+	.dccg_read_reg_state = dccg31_read_reg_state,
+};
+
+struct dccg *dccg31_create(
+	struct dc_context *ctx,
+	const struct dccg_registers *regs,
+	const struct dccg_shift *dccg_shift,
+	const struct dccg_mask *dccg_mask)
+{
+	struct dcn_dccg *dccg_dcn = kzalloc(sizeof(*dccg_dcn), GFP_KERNEL);
+	struct dccg *base;
+
+	if (dccg_dcn == NULL) {
+		BREAK_TO_DEBUGGER();
+		return NULL;
+	}
+
+	base = &dccg_dcn->base;
+	base->ctx = ctx;
+	base->funcs = &dccg31_funcs;
+
+	dccg_dcn->regs = regs;
+	dccg_dcn->dccg_shift = dccg_shift;
+	dccg_dcn->dccg_mask = dccg_mask;
+
+	return &dccg_dcn->base;
+}
diff --git a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_dccg.h b/drivers/gpu/drm/amd/display/dc/dccg/dcn31/dcn31_dccg.h
index a013a32bbaf7..bf659920d4cc 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_dccg.h
+++ b/drivers/gpu/drm/amd/display/dc/dccg/dcn31/dcn31_dccg.h
@@ -28,22 +28,20 @@
 
 #include "dcn30/dcn30_dccg.h"
 
-#define DCCG_SFII(block, reg_name, field_prefix, field_name, inst, post_fix)\
-	.field_prefix ## _ ## field_name[inst] = block ## inst ## _ ## reg_name ## __ ## field_prefix ## inst ## _ ## field_name ## post_fix
-
-
 #define DCCG_REG_LIST_DCN31() \
 	SR(DPPCLK_DTO_CTRL),\
 	DCCG_SRII(DTO_PARAM, DPPCLK, 0),\
 	DCCG_SRII(DTO_PARAM, DPPCLK, 1),\
 	DCCG_SRII(DTO_PARAM, DPPCLK, 2),\
 	DCCG_SRII(DTO_PARAM, DPPCLK, 3),\
+	DCCG_SRII(CLOCK_CNTL, HDMICHARCLK, 0),\
 	SR(PHYASYMCLK_CLOCK_CNTL),\
 	SR(PHYBSYMCLK_CLOCK_CNTL),\
 	SR(PHYCSYMCLK_CLOCK_CNTL),\
 	SR(PHYDSYMCLK_CLOCK_CNTL),\
 	SR(PHYESYMCLK_CLOCK_CNTL),\
 	SR(DPSTREAMCLK_CNTL),\
+	SR(HDMISTREAMCLK_CNTL),\
 	SR(SYMCLK32_SE_CNTL),\
 	SR(SYMCLK32_LE_CNTL),\
 	DCCG_SRII(PIXEL_RATE_CNTL, OTG, 0),\
@@ -66,6 +64,7 @@
 	SR(DSCCLK1_DTO_PARAM),\
 	SR(DSCCLK2_DTO_PARAM),\
 	SR(DSCCLK_DTO_CTRL),\
+	SR(DCCG_GATE_DISABLE_CNTL2),\
 	SR(DCCG_GATE_DISABLE_CNTL3),\
 	SR(HDMISTREAMCLK0_DTO_PARAM)
 
@@ -81,6 +80,8 @@
 	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_DB_EN, DPPCLK, 3, mask_sh),\
 	DCCG_SF(DPPCLK0_DTO_PARAM, DPPCLK0_DTO_PHASE, mask_sh),\
 	DCCG_SF(DPPCLK0_DTO_PARAM, DPPCLK0_DTO_MODULO, mask_sh),\
+	DCCG_SF(HDMICHARCLK0_CLOCK_CNTL, HDMICHARCLK0_EN, mask_sh),\
+	DCCG_SF(HDMICHARCLK0_CLOCK_CNTL, HDMICHARCLK0_SRC_SEL, mask_sh),\
 	DCCG_SF(PHYASYMCLK_CLOCK_CNTL, PHYASYMCLK_FORCE_EN, mask_sh),\
 	DCCG_SF(PHYASYMCLK_CLOCK_CNTL, PHYASYMCLK_FORCE_SRC_SEL, mask_sh),\
 	DCCG_SF(PHYBSYMCLK_CLOCK_CNTL, PHYBSYMCLK_FORCE_EN, mask_sh),\
@@ -95,6 +96,8 @@
 	DCCG_SF(DPSTREAMCLK_CNTL, DPSTREAMCLK_PIPE1_EN, mask_sh),\
 	DCCG_SF(DPSTREAMCLK_CNTL, DPSTREAMCLK_PIPE2_EN, mask_sh),\
 	DCCG_SF(DPSTREAMCLK_CNTL, DPSTREAMCLK_PIPE3_EN, mask_sh),\
+	DCCG_SF(HDMISTREAMCLK_CNTL, HDMISTREAMCLK0_SRC_SEL, mask_sh),\
+	DCCG_SF(HDMISTREAMCLK_CNTL, HDMISTREAMCLK0_DTO_FORCE_DIS, mask_sh),\
 	DCCG_SF(SYMCLK32_SE_CNTL, SYMCLK32_SE0_SRC_SEL, mask_sh),\
 	DCCG_SF(SYMCLK32_SE_CNTL, SYMCLK32_SE1_SRC_SEL, mask_sh),\
 	DCCG_SF(SYMCLK32_SE_CNTL, SYMCLK32_SE2_SRC_SEL, mask_sh),\
@@ -123,9 +126,13 @@
 	DCCG_SFII(OTG, PIXEL_RATE_CNTL, DTBCLK_DTO, DIV, 1, mask_sh),\
 	DCCG_SFII(OTG, PIXEL_RATE_CNTL, DTBCLK_DTO, DIV, 2, mask_sh),\
 	DCCG_SFII(OTG, PIXEL_RATE_CNTL, DTBCLK_DTO, DIV, 3, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, ADD_PIXEL, 0, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, ADD_PIXEL, 1, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, ADD_PIXEL, 2, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, ADD_PIXEL, 3, mask_sh),\
 	DCCG_SF(DCCG_AUDIO_DTO_SOURCE, DCCG_AUDIO_DTO_SEL, mask_sh),\
 	DCCG_SF(DCCG_AUDIO_DTO_SOURCE, DCCG_AUDIO_DTO0_SOURCE_SEL, mask_sh),\
-	DCCG_SF(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_CHG_MODE, mask_sh), \
+	DCCG_SF(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_CHG_MODE, mask_sh),\
 	DCCG_SF(DSCCLK0_DTO_PARAM, DSCCLK0_DTO_PHASE, mask_sh),\
 	DCCG_SF(DSCCLK0_DTO_PARAM, DSCCLK0_DTO_MODULO, mask_sh),\
 	DCCG_SF(DSCCLK1_DTO_PARAM, DSCCLK1_DTO_PHASE, mask_sh),\
@@ -135,6 +142,11 @@
 	DCCG_SF(DSCCLK_DTO_CTRL, DSCCLK0_DTO_ENABLE, mask_sh),\
 	DCCG_SF(DSCCLK_DTO_CTRL, DSCCLK1_DTO_ENABLE, mask_sh),\
 	DCCG_SF(DSCCLK_DTO_CTRL, DSCCLK2_DTO_ENABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL2, PHYASYMCLK_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL2, PHYBSYMCLK_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL2, PHYCSYMCLK_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL2, PHYDSYMCLK_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL2, PHYESYMCLK_GATE_DISABLE, mask_sh),\
 	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, DPSTREAMCLK_ROOT_GATE_DISABLE, mask_sh),\
 	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, DPSTREAMCLK_GATE_DISABLE, mask_sh),\
 	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, SYMCLK32_ROOT_SE0_GATE_DISABLE, mask_sh),\
@@ -155,11 +167,6 @@ struct dccg *dccg31_create(
 
 void dccg31_init(struct dccg *dccg);
 
-void dccg31_set_dpstreamclk(
-		struct dccg *dccg,
-		enum hdmistreamclk_source src,
-		int otg_inst);
-
 void dccg31_enable_symclk32_se(
 		struct dccg *dccg,
 		int hpo_se_inst,
@@ -178,6 +185,11 @@ void dccg31_disable_symclk32_le(
 		struct dccg *dccg,
 		int hpo_le_inst);
 
+void dccg31_set_symclk32_le_root_clock_gating(
+		struct dccg *dccg,
+		int hpo_le_inst,
+		bool enable);
+
 void dccg31_set_physymclk(
 		struct dccg *dccg,
 		int phy_inst,
@@ -186,10 +198,44 @@ void dccg31_set_physymclk(
 
 void dccg31_set_audio_dtbclk_dto(
 		struct dccg *dccg,
-		uint32_t req_audio_dtbclk_khz);
+		const struct dtbclk_dto_params *params);
+
+void dccg31_update_dpp_dto(
+	struct dccg *dccg,
+	int dpp_inst,
+	int req_dppclk);
 
-void dccg31_set_hdmistreamclk(
+void dccg31_get_dccg_ref_freq(
+	struct dccg *dccg,
+	unsigned int xtalin_freq_inKhz,
+	unsigned int *dccg_ref_freq_inKhz);
+
+void dccg31_set_dpstreamclk(
+	struct dccg *dccg,
+	enum streamclk_source src,
+	int otg_inst,
+	int dp_hpo_inst);
+
+void dccg31_set_dtbclk_dto(
 		struct dccg *dccg,
-		enum hdmistreamclk_source src);
+		const struct dtbclk_dto_params *params);
+
+void dccg31_otg_add_pixel(
+	struct dccg *dccg,
+	uint32_t otg_inst);
+
+void dccg31_otg_drop_pixel(
+	struct dccg *dccg,
+	uint32_t otg_inst);
+
+void dccg31_set_dispclk_change_mode(
+	struct dccg *dccg,
+	enum dentist_dispclk_change_mode change_mode);
+
+void dccg31_disable_dscclk(struct dccg *dccg, int inst);
+
+void dccg31_enable_dscclk(struct dccg *dccg, int inst);
+
+void dccg31_read_reg_state(struct dccg *dccg, struct dcn_dccg_reg_state *dccg_reg_state);
 
 #endif //__DCN31_DCCG_H__
diff --git a/drivers/gpu/drm/amd/display/dc/dccg/dcn314/dcn314_dccg.c b/drivers/gpu/drm/amd/display/dc/dccg/dcn314/dcn314_dccg.c
new file mode 100644
index 000000000000..ef3db6beba25
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dccg/dcn314/dcn314_dccg.c
@@ -0,0 +1,407 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#include "reg_helper.h"
+#include "core_types.h"
+
+#include "dcn31/dcn31_dccg.h"
+#include "dcn314_dccg.h"
+
+#define TO_DCN_DCCG(dccg)\
+	container_of(dccg, struct dcn_dccg, base)
+
+#define REG(reg) \
+	(dccg_dcn->regs->reg)
+
+#undef FN
+#define FN(reg_name, field_name) \
+	dccg_dcn->dccg_shift->field_name, dccg_dcn->dccg_mask->field_name
+
+#define CTX \
+	dccg_dcn->base.ctx
+#define DC_LOGGER \
+	dccg->ctx->logger
+
+static void dccg314_trigger_dio_fifo_resync(
+	struct dccg *dccg)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+	uint32_t dispclk_rdivider_value = 0;
+
+	REG_GET(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_RDIVIDER, &dispclk_rdivider_value);
+	REG_UPDATE(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_WDIVIDER, dispclk_rdivider_value);
+}
+
+static void dccg314_get_pixel_rate_div(
+		struct dccg *dccg,
+		uint32_t otg_inst,
+		uint32_t *k1,
+		uint32_t *k2)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+	uint32_t val_k1 = PIXEL_RATE_DIV_NA, val_k2 = PIXEL_RATE_DIV_NA;
+
+	*k1 = PIXEL_RATE_DIV_NA;
+	*k2 = PIXEL_RATE_DIV_NA;
+
+	switch (otg_inst) {
+	case 0:
+		REG_GET_2(OTG_PIXEL_RATE_DIV,
+			OTG0_PIXEL_RATE_DIVK1, &val_k1,
+			OTG0_PIXEL_RATE_DIVK2, &val_k2);
+		break;
+	case 1:
+		REG_GET_2(OTG_PIXEL_RATE_DIV,
+			OTG1_PIXEL_RATE_DIVK1, &val_k1,
+			OTG1_PIXEL_RATE_DIVK2, &val_k2);
+		break;
+	case 2:
+		REG_GET_2(OTG_PIXEL_RATE_DIV,
+			OTG2_PIXEL_RATE_DIVK1, &val_k1,
+			OTG2_PIXEL_RATE_DIVK2, &val_k2);
+		break;
+	case 3:
+		REG_GET_2(OTG_PIXEL_RATE_DIV,
+			OTG3_PIXEL_RATE_DIVK1, &val_k1,
+			OTG3_PIXEL_RATE_DIVK2, &val_k2);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+
+	*k1 = val_k1;
+	*k2 = val_k2;
+}
+
+static void dccg314_set_pixel_rate_div(
+		struct dccg *dccg,
+		uint32_t otg_inst,
+		enum pixel_rate_div k1,
+		enum pixel_rate_div k2)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+	uint32_t cur_k1 = PIXEL_RATE_DIV_NA;
+	uint32_t cur_k2 = PIXEL_RATE_DIV_NA;
+
+	// Don't program 0xF into the register field. Not valid since
+	// K1 / K2 field is only 1 / 2 bits wide
+	if (k1 == PIXEL_RATE_DIV_NA || k2 == PIXEL_RATE_DIV_NA) {
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+
+	dccg314_get_pixel_rate_div(dccg, otg_inst, &cur_k1, &cur_k2);
+	if (k1 == cur_k1 && k2 == cur_k2)
+		return;
+
+	switch (otg_inst) {
+	case 0:
+		REG_UPDATE_2(OTG_PIXEL_RATE_DIV,
+				OTG0_PIXEL_RATE_DIVK1, k1,
+				OTG0_PIXEL_RATE_DIVK2, k2);
+		break;
+	case 1:
+		REG_UPDATE_2(OTG_PIXEL_RATE_DIV,
+				OTG1_PIXEL_RATE_DIVK1, k1,
+				OTG1_PIXEL_RATE_DIVK2, k2);
+		break;
+	case 2:
+		REG_UPDATE_2(OTG_PIXEL_RATE_DIV,
+				OTG2_PIXEL_RATE_DIVK1, k1,
+				OTG2_PIXEL_RATE_DIVK2, k2);
+		break;
+	case 3:
+		REG_UPDATE_2(OTG_PIXEL_RATE_DIV,
+				OTG3_PIXEL_RATE_DIVK1, k1,
+				OTG3_PIXEL_RATE_DIVK2, k2);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+}
+
+static void dccg314_set_dtbclk_p_src(
+		struct dccg *dccg,
+		enum streamclk_source src,
+		uint32_t otg_inst)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	uint32_t p_src_sel = 0; /* selects dprefclk */
+
+	if (src == DTBCLK0)
+		p_src_sel = 2;  /* selects dtbclk0 */
+
+	switch (otg_inst) {
+	case 0:
+		if (src == REFCLK)
+			REG_UPDATE(DTBCLK_P_CNTL,
+					DTBCLK_P0_EN, 0);
+		else
+			REG_UPDATE_2(DTBCLK_P_CNTL,
+					DTBCLK_P0_SRC_SEL, p_src_sel,
+					DTBCLK_P0_EN, 1);
+		break;
+	case 1:
+		if (src == REFCLK)
+			REG_UPDATE(DTBCLK_P_CNTL,
+					DTBCLK_P1_EN, 0);
+		else
+			REG_UPDATE_2(DTBCLK_P_CNTL,
+					DTBCLK_P1_SRC_SEL, p_src_sel,
+					DTBCLK_P1_EN, 1);
+		break;
+	case 2:
+		if (src == REFCLK)
+			REG_UPDATE(DTBCLK_P_CNTL,
+					DTBCLK_P2_EN, 0);
+		else
+			REG_UPDATE_2(DTBCLK_P_CNTL,
+					DTBCLK_P2_SRC_SEL, p_src_sel,
+					DTBCLK_P2_EN, 1);
+		break;
+	case 3:
+		if (src == REFCLK)
+			REG_UPDATE(DTBCLK_P_CNTL,
+					DTBCLK_P3_EN, 0);
+		else
+			REG_UPDATE_2(DTBCLK_P_CNTL,
+					DTBCLK_P3_SRC_SEL, p_src_sel,
+					DTBCLK_P3_EN, 1);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+
+}
+
+/* Controls the generation of pixel valid for OTG in (OTG -> HPO case) */
+static void dccg314_set_dtbclk_dto(
+		struct dccg *dccg,
+		const struct dtbclk_dto_params *params)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+	/* DTO Output Rate / Pixel Rate = 1/4 */
+	int req_dtbclk_khz = params->pixclk_khz / 4;
+
+	if (params->ref_dtbclk_khz && req_dtbclk_khz) {
+		uint32_t modulo, phase;
+
+		// phase / modulo = dtbclk / dtbclk ref
+		modulo = params->ref_dtbclk_khz * 1000;
+		phase = req_dtbclk_khz * 1000;
+
+		REG_WRITE(DTBCLK_DTO_MODULO[params->otg_inst], modulo);
+		REG_WRITE(DTBCLK_DTO_PHASE[params->otg_inst], phase);
+
+		REG_UPDATE(OTG_PIXEL_RATE_CNTL[params->otg_inst],
+				DTBCLK_DTO_ENABLE[params->otg_inst], 1);
+
+		REG_WAIT(OTG_PIXEL_RATE_CNTL[params->otg_inst],
+				DTBCLKDTO_ENABLE_STATUS[params->otg_inst], 1,
+				1, 100);
+
+		/* program OTG_PIXEL_RATE_DIV for DIVK1 and DIVK2 fields */
+		dccg314_set_pixel_rate_div(dccg, params->otg_inst, PIXEL_RATE_DIV_BY_1, PIXEL_RATE_DIV_BY_1);
+
+		/* The recommended programming sequence to enable DTBCLK DTO to generate
+		 * valid pixel HPO DPSTREAM ENCODER, specifies that DTO source select should
+		 * be set only after DTO is enabled
+		 */
+		REG_UPDATE(OTG_PIXEL_RATE_CNTL[params->otg_inst],
+				PIPE_DTO_SRC_SEL[params->otg_inst], 2);
+	} else {
+		REG_UPDATE_2(OTG_PIXEL_RATE_CNTL[params->otg_inst],
+				DTBCLK_DTO_ENABLE[params->otg_inst], 0,
+				PIPE_DTO_SRC_SEL[params->otg_inst], 1);
+
+		REG_WRITE(DTBCLK_DTO_MODULO[params->otg_inst], 0);
+		REG_WRITE(DTBCLK_DTO_PHASE[params->otg_inst], 0);
+	}
+}
+
+void dccg314_set_dpstreamclk(
+		struct dccg *dccg,
+		enum streamclk_source src,
+		int otg_inst,
+		int dp_hpo_inst)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	/* set the dtbclk_p source */
+	dccg314_set_dtbclk_p_src(dccg, src, otg_inst);
+
+	/* enabled to select one of the DTBCLKs for pipe */
+	switch (dp_hpo_inst) {
+	case 0:
+		REG_UPDATE_2(DPSTREAMCLK_CNTL,
+					DPSTREAMCLK0_EN, (src == REFCLK) ? 0 : 1,
+					DPSTREAMCLK0_SRC_SEL, otg_inst);
+		break;
+	case 1:
+		REG_UPDATE_2(DPSTREAMCLK_CNTL,
+					DPSTREAMCLK1_EN, (src == REFCLK) ? 0 : 1,
+					DPSTREAMCLK1_SRC_SEL, otg_inst);
+		break;
+	case 2:
+		REG_UPDATE_2(DPSTREAMCLK_CNTL,
+					DPSTREAMCLK2_EN, (src == REFCLK) ? 0 : 1,
+					DPSTREAMCLK2_SRC_SEL, otg_inst);
+		break;
+	case 3:
+		REG_UPDATE_2(DPSTREAMCLK_CNTL,
+					DPSTREAMCLK3_EN, (src == REFCLK) ? 0 : 1,
+					DPSTREAMCLK3_SRC_SEL, otg_inst);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+}
+
+static void dccg314_init(struct dccg *dccg)
+{
+	int otg_inst;
+
+	/* Set HPO stream encoder to use refclk to avoid case where PHY is
+	 * disabled and SYMCLK32 for HPO SE is sourced from PHYD32CLK which
+	 * will cause DCN to hang.
+	 */
+	for (otg_inst = 0; otg_inst < 4; otg_inst++)
+		dccg31_disable_symclk32_se(dccg, otg_inst);
+
+	if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_le)
+		for (otg_inst = 0; otg_inst < 2; otg_inst++)
+			dccg31_disable_symclk32_le(dccg, otg_inst);
+
+	if (dccg->ctx->dc->debug.root_clock_optimization.bits.dpstream)
+		for (otg_inst = 0; otg_inst < 4; otg_inst++)
+			dccg314_set_dpstreamclk(dccg, REFCLK, otg_inst,
+						otg_inst);
+
+	if (dccg->ctx->dc->debug.root_clock_optimization.bits.physymclk)
+		for (otg_inst = 0; otg_inst < 5; otg_inst++)
+			dccg31_set_physymclk(dccg, otg_inst,
+					     PHYSYMCLK_FORCE_SRC_SYMCLK, false);
+}
+
+static void dccg314_set_valid_pixel_rate(
+		struct dccg *dccg,
+		int ref_dtbclk_khz,
+		int otg_inst,
+		int pixclk_khz)
+{
+	struct dtbclk_dto_params dto_params = {0};
+
+	dto_params.ref_dtbclk_khz = ref_dtbclk_khz;
+	dto_params.otg_inst = otg_inst;
+	dto_params.pixclk_khz = pixclk_khz;
+
+	dccg314_set_dtbclk_dto(dccg, &dto_params);
+}
+
+static void dccg314_dpp_root_clock_control(
+		struct dccg *dccg,
+		unsigned int dpp_inst,
+		bool clock_on)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	if (dccg->dpp_clock_gated[dpp_inst] != clock_on)
+		return;
+
+	if (clock_on) {
+		/* turn off the DTO and leave phase/modulo at max */
+		REG_UPDATE(DPPCLK_DTO_CTRL, DPPCLK_DTO_ENABLE[dpp_inst], 0);
+		REG_SET_2(DPPCLK_DTO_PARAM[dpp_inst], 0,
+			  DPPCLK0_DTO_PHASE, 0xFF,
+			  DPPCLK0_DTO_MODULO, 0xFF);
+	} else {
+		/* turn on the DTO to generate a 0hz clock */
+		REG_UPDATE(DPPCLK_DTO_CTRL, DPPCLK_DTO_ENABLE[dpp_inst], 1);
+		REG_SET_2(DPPCLK_DTO_PARAM[dpp_inst], 0,
+			  DPPCLK0_DTO_PHASE, 0,
+			  DPPCLK0_DTO_MODULO, 1);
+	}
+
+	dccg->dpp_clock_gated[dpp_inst] = !clock_on;
+}
+
+static const struct dccg_funcs dccg314_funcs = {
+	.update_dpp_dto = dccg31_update_dpp_dto,
+	.dpp_root_clock_control = dccg314_dpp_root_clock_control,
+	.get_dccg_ref_freq = dccg31_get_dccg_ref_freq,
+	.dccg_init = dccg314_init,
+	.set_dpstreamclk = dccg314_set_dpstreamclk,
+	.enable_symclk32_se = dccg31_enable_symclk32_se,
+	.disable_symclk32_se = dccg31_disable_symclk32_se,
+	.enable_symclk32_le = dccg31_enable_symclk32_le,
+	.disable_symclk32_le = dccg31_disable_symclk32_le,
+	.set_symclk32_le_root_clock_gating = dccg31_set_symclk32_le_root_clock_gating,
+	.set_physymclk = dccg31_set_physymclk,
+	.set_dtbclk_dto = dccg314_set_dtbclk_dto,
+	.set_audio_dtbclk_dto = dccg31_set_audio_dtbclk_dto,
+	.set_fifo_errdet_ovr_en = dccg2_set_fifo_errdet_ovr_en,
+	.otg_add_pixel = dccg31_otg_add_pixel,
+	.otg_drop_pixel = dccg31_otg_drop_pixel,
+	.set_dispclk_change_mode = dccg31_set_dispclk_change_mode,
+	.disable_dsc = dccg31_disable_dscclk,
+	.enable_dsc = dccg31_enable_dscclk,
+	.set_pixel_rate_div = dccg314_set_pixel_rate_div,
+	.get_pixel_rate_div = dccg314_get_pixel_rate_div,
+	.trigger_dio_fifo_resync = dccg314_trigger_dio_fifo_resync,
+	.set_valid_pixel_rate = dccg314_set_valid_pixel_rate,
+	.set_dtbclk_p_src = dccg314_set_dtbclk_p_src,
+	.dccg_read_reg_state = dccg31_read_reg_state
+};
+
+struct dccg *dccg314_create(
+	struct dc_context *ctx,
+	const struct dccg_registers *regs,
+	const struct dccg_shift *dccg_shift,
+	const struct dccg_mask *dccg_mask)
+{
+	struct dcn_dccg *dccg_dcn = kzalloc(sizeof(*dccg_dcn), GFP_KERNEL);
+	struct dccg *base;
+
+	if (dccg_dcn == NULL) {
+		BREAK_TO_DEBUGGER();
+		return NULL;
+	}
+
+	base = &dccg_dcn->base;
+	base->ctx = ctx;
+	base->funcs = &dccg314_funcs;
+
+	dccg_dcn->regs = regs;
+	dccg_dcn->dccg_shift = dccg_shift;
+	dccg_dcn->dccg_mask = dccg_mask;
+
+	return &dccg_dcn->base;
+}
diff --git a/drivers/gpu/drm/amd/display/dc/dccg/dcn314/dcn314_dccg.h b/drivers/gpu/drm/amd/display/dc/dccg/dcn314/dcn314_dccg.h
new file mode 100644
index 000000000000..a609635f35db
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dccg/dcn314/dcn314_dccg.h
@@ -0,0 +1,211 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __DCN314_DCCG_H__
+#define __DCN314_DCCG_H__
+
+#include "dcn31/dcn31_dccg.h"
+
+#define DCCG_SFII(block, reg_name, field_prefix, field_name, inst, post_fix)\
+	.field_prefix ## _ ## field_name[inst] = block ## inst ## _ ## reg_name ## __ ## field_prefix ## inst ## _ ## field_name ## post_fix
+
+
+#define DCCG_REG_LIST_DCN314() \
+	SR(DPPCLK_DTO_CTRL),\
+	DCCG_SRII(DTO_PARAM, DPPCLK, 0),\
+	DCCG_SRII(DTO_PARAM, DPPCLK, 1),\
+	DCCG_SRII(DTO_PARAM, DPPCLK, 2),\
+	DCCG_SRII(DTO_PARAM, DPPCLK, 3),\
+	DCCG_SRII(CLOCK_CNTL, HDMICHARCLK, 0),\
+	SR(PHYASYMCLK_CLOCK_CNTL),\
+	SR(PHYBSYMCLK_CLOCK_CNTL),\
+	SR(PHYCSYMCLK_CLOCK_CNTL),\
+	SR(PHYDSYMCLK_CLOCK_CNTL),\
+	SR(PHYESYMCLK_CLOCK_CNTL),\
+	SR(DPSTREAMCLK_CNTL),\
+	SR(HDMISTREAMCLK_CNTL),\
+	SR(SYMCLK32_SE_CNTL),\
+	SR(SYMCLK32_LE_CNTL),\
+	DCCG_SRII(PIXEL_RATE_CNTL, OTG, 0),\
+	DCCG_SRII(PIXEL_RATE_CNTL, OTG, 1),\
+	DCCG_SRII(PIXEL_RATE_CNTL, OTG, 2),\
+	DCCG_SRII(PIXEL_RATE_CNTL, OTG, 3),\
+	DCCG_SRII(MODULO, DTBCLK_DTO, 0),\
+	DCCG_SRII(MODULO, DTBCLK_DTO, 1),\
+	DCCG_SRII(MODULO, DTBCLK_DTO, 2),\
+	DCCG_SRII(MODULO, DTBCLK_DTO, 3),\
+	DCCG_SRII(PHASE, DTBCLK_DTO, 0),\
+	DCCG_SRII(PHASE, DTBCLK_DTO, 1),\
+	DCCG_SRII(PHASE, DTBCLK_DTO, 2),\
+	DCCG_SRII(PHASE, DTBCLK_DTO, 3),\
+	SR(DCCG_AUDIO_DTBCLK_DTO_MODULO),\
+	SR(DCCG_AUDIO_DTBCLK_DTO_PHASE),\
+	SR(DCCG_AUDIO_DTO_SOURCE),\
+	SR(DENTIST_DISPCLK_CNTL),\
+	SR(DSCCLK0_DTO_PARAM),\
+	SR(DSCCLK1_DTO_PARAM),\
+	SR(DSCCLK2_DTO_PARAM),\
+	SR(DSCCLK3_DTO_PARAM),\
+	SR(DSCCLK_DTO_CTRL),\
+	SR(DCCG_GATE_DISABLE_CNTL2),\
+	SR(DCCG_GATE_DISABLE_CNTL3),\
+	SR(HDMISTREAMCLK0_DTO_PARAM),\
+	SR(OTG_PIXEL_RATE_DIV),\
+	SR(DTBCLK_P_CNTL)
+
+#define DCCG_MASK_SH_LIST_DCN314_COMMON(mask_sh) \
+	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_DB_EN, DPPCLK, 0, mask_sh),\
+	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_DB_EN, DPPCLK, 1, mask_sh),\
+	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_DB_EN, DPPCLK, 2, mask_sh),\
+	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_DB_EN, DPPCLK, 3, mask_sh),\
+	DCCG_SF(DPPCLK0_DTO_PARAM, DPPCLK0_DTO_PHASE, mask_sh),\
+	DCCG_SF(DPPCLK0_DTO_PARAM, DPPCLK0_DTO_MODULO, mask_sh),\
+	DCCG_SF(HDMICHARCLK0_CLOCK_CNTL, HDMICHARCLK0_EN, mask_sh),\
+	DCCG_SF(HDMICHARCLK0_CLOCK_CNTL, HDMICHARCLK0_SRC_SEL, mask_sh),\
+	DCCG_SF(DPSTREAMCLK_CNTL, DPSTREAMCLK0_EN, mask_sh),\
+	DCCG_SF(DPSTREAMCLK_CNTL, DPSTREAMCLK1_EN, mask_sh),\
+	DCCG_SF(DPSTREAMCLK_CNTL, DPSTREAMCLK2_EN, mask_sh),\
+	DCCG_SF(DPSTREAMCLK_CNTL, DPSTREAMCLK3_EN, mask_sh),\
+	DCCG_SF(DPSTREAMCLK_CNTL, DPSTREAMCLK0_SRC_SEL, mask_sh),\
+	DCCG_SF(DPSTREAMCLK_CNTL, DPSTREAMCLK1_SRC_SEL, mask_sh),\
+	DCCG_SF(DPSTREAMCLK_CNTL, DPSTREAMCLK2_SRC_SEL, mask_sh),\
+	DCCG_SF(DPSTREAMCLK_CNTL, DPSTREAMCLK3_SRC_SEL, mask_sh),\
+	DCCG_SF(HDMISTREAMCLK_CNTL, HDMISTREAMCLK0_EN, mask_sh),\
+	DCCG_SF(HDMISTREAMCLK_CNTL, HDMISTREAMCLK0_SRC_SEL, mask_sh),\
+	DCCG_SF(SYMCLK32_SE_CNTL, SYMCLK32_SE0_SRC_SEL, mask_sh),\
+	DCCG_SF(SYMCLK32_SE_CNTL, SYMCLK32_SE1_SRC_SEL, mask_sh),\
+	DCCG_SF(SYMCLK32_SE_CNTL, SYMCLK32_SE2_SRC_SEL, mask_sh),\
+	DCCG_SF(SYMCLK32_SE_CNTL, SYMCLK32_SE3_SRC_SEL, mask_sh),\
+	DCCG_SF(SYMCLK32_SE_CNTL, SYMCLK32_SE0_EN, mask_sh),\
+	DCCG_SF(SYMCLK32_SE_CNTL, SYMCLK32_SE1_EN, mask_sh),\
+	DCCG_SF(SYMCLK32_SE_CNTL, SYMCLK32_SE2_EN, mask_sh),\
+	DCCG_SF(SYMCLK32_SE_CNTL, SYMCLK32_SE3_EN, mask_sh),\
+	DCCG_SF(SYMCLK32_LE_CNTL, SYMCLK32_LE0_SRC_SEL, mask_sh),\
+	DCCG_SF(SYMCLK32_LE_CNTL, SYMCLK32_LE1_SRC_SEL, mask_sh),\
+	DCCG_SF(SYMCLK32_LE_CNTL, SYMCLK32_LE0_EN, mask_sh),\
+	DCCG_SF(SYMCLK32_LE_CNTL, SYMCLK32_LE1_EN, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, DTBCLK_DTO, ENABLE, 0, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, DTBCLK_DTO, ENABLE, 1, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, DTBCLK_DTO, ENABLE, 2, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, DTBCLK_DTO, ENABLE, 3, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, DTBCLKDTO, ENABLE_STATUS, 0, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, DTBCLKDTO, ENABLE_STATUS, 1, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, DTBCLKDTO, ENABLE_STATUS, 2, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, DTBCLKDTO, ENABLE_STATUS, 3, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, PIPE, DTO_SRC_SEL, 0, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, PIPE, DTO_SRC_SEL, 1, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, PIPE, DTO_SRC_SEL, 2, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, PIPE, DTO_SRC_SEL, 3, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, ADD_PIXEL, 0, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, ADD_PIXEL, 1, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, ADD_PIXEL, 2, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, ADD_PIXEL, 3, mask_sh),\
+	DCCG_SF(OTG_PIXEL_RATE_DIV, OTG0_PIXEL_RATE_DIVK1, mask_sh),\
+	DCCG_SF(OTG_PIXEL_RATE_DIV, OTG0_PIXEL_RATE_DIVK2, mask_sh),\
+	DCCG_SF(OTG_PIXEL_RATE_DIV, OTG1_PIXEL_RATE_DIVK1, mask_sh),\
+	DCCG_SF(OTG_PIXEL_RATE_DIV, OTG1_PIXEL_RATE_DIVK2, mask_sh),\
+	DCCG_SF(OTG_PIXEL_RATE_DIV, OTG2_PIXEL_RATE_DIVK1, mask_sh),\
+	DCCG_SF(OTG_PIXEL_RATE_DIV, OTG2_PIXEL_RATE_DIVK2, mask_sh),\
+	DCCG_SF(OTG_PIXEL_RATE_DIV, OTG3_PIXEL_RATE_DIVK1, mask_sh),\
+	DCCG_SF(OTG_PIXEL_RATE_DIV, OTG3_PIXEL_RATE_DIVK2, mask_sh),\
+	DCCG_SF(OTG_PIXEL_RATE_DIV, OTG3_PIXEL_RATE_DIVK2, mask_sh),\
+	DCCG_SF(DTBCLK_P_CNTL, DTBCLK_P0_SRC_SEL, mask_sh),\
+	DCCG_SF(DTBCLK_P_CNTL, DTBCLK_P0_EN, mask_sh),\
+	DCCG_SF(DTBCLK_P_CNTL, DTBCLK_P1_SRC_SEL, mask_sh),\
+	DCCG_SF(DTBCLK_P_CNTL, DTBCLK_P1_EN, mask_sh),\
+	DCCG_SF(DTBCLK_P_CNTL, DTBCLK_P2_SRC_SEL, mask_sh),\
+	DCCG_SF(DTBCLK_P_CNTL, DTBCLK_P2_EN, mask_sh),\
+	DCCG_SF(DTBCLK_P_CNTL, DTBCLK_P3_SRC_SEL, mask_sh),\
+	DCCG_SF(DTBCLK_P_CNTL, DTBCLK_P3_EN, mask_sh),\
+	DCCG_SF(DCCG_AUDIO_DTO_SOURCE, DCCG_AUDIO_DTO0_SOURCE_SEL, mask_sh),\
+	DCCG_SF(DCCG_AUDIO_DTO_SOURCE, DCCG_AUDIO_DTO_SEL, mask_sh),\
+	DCCG_SF(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_CHG_MODE, mask_sh),\
+	DCCG_SF(DSCCLK0_DTO_PARAM, DSCCLK0_DTO_PHASE, mask_sh),\
+	DCCG_SF(DSCCLK0_DTO_PARAM, DSCCLK0_DTO_MODULO, mask_sh),\
+	DCCG_SF(DSCCLK1_DTO_PARAM, DSCCLK1_DTO_PHASE, mask_sh),\
+	DCCG_SF(DSCCLK1_DTO_PARAM, DSCCLK1_DTO_MODULO, mask_sh),\
+	DCCG_SF(DSCCLK2_DTO_PARAM, DSCCLK2_DTO_PHASE, mask_sh),\
+	DCCG_SF(DSCCLK2_DTO_PARAM, DSCCLK2_DTO_MODULO, mask_sh),\
+	DCCG_SF(DSCCLK3_DTO_PARAM, DSCCLK3_DTO_PHASE, mask_sh),\
+	DCCG_SF(DSCCLK3_DTO_PARAM, DSCCLK3_DTO_MODULO, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, SYMCLK32_ROOT_SE0_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, SYMCLK32_ROOT_SE1_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, SYMCLK32_ROOT_SE2_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, SYMCLK32_ROOT_SE3_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, SYMCLK32_ROOT_LE0_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, SYMCLK32_ROOT_LE1_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, SYMCLK32_SE0_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, SYMCLK32_SE1_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, SYMCLK32_SE2_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, SYMCLK32_SE3_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, SYMCLK32_LE0_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, SYMCLK32_LE1_GATE_DISABLE, mask_sh),\
+	DCCG_SF(HDMISTREAMCLK0_DTO_PARAM, HDMISTREAMCLK0_DTO_PHASE, mask_sh),\
+	DCCG_SF(HDMISTREAMCLK0_DTO_PARAM, HDMISTREAMCLK0_DTO_MODULO, mask_sh)
+
+#define DCCG_MASK_SH_LIST_DCN314(mask_sh) \
+	DCCG_MASK_SH_LIST_DCN314_COMMON(mask_sh),\
+	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_ENABLE, DPPCLK, 0, mask_sh),\
+	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_ENABLE, DPPCLK, 1, mask_sh),\
+	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_ENABLE, DPPCLK, 2, mask_sh),\
+	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_ENABLE, DPPCLK, 3, mask_sh),\
+	DCCG_SF(PHYASYMCLK_CLOCK_CNTL, PHYASYMCLK_FORCE_EN, mask_sh),\
+	DCCG_SF(PHYASYMCLK_CLOCK_CNTL, PHYASYMCLK_FORCE_SRC_SEL, mask_sh),\
+	DCCG_SF(PHYBSYMCLK_CLOCK_CNTL, PHYBSYMCLK_FORCE_EN, mask_sh),\
+	DCCG_SF(PHYBSYMCLK_CLOCK_CNTL, PHYBSYMCLK_FORCE_SRC_SEL, mask_sh),\
+	DCCG_SF(PHYCSYMCLK_CLOCK_CNTL, PHYCSYMCLK_FORCE_EN, mask_sh),\
+	DCCG_SF(PHYCSYMCLK_CLOCK_CNTL, PHYCSYMCLK_FORCE_SRC_SEL, mask_sh),\
+	DCCG_SF(PHYDSYMCLK_CLOCK_CNTL, PHYDSYMCLK_FORCE_EN, mask_sh),\
+	DCCG_SF(PHYDSYMCLK_CLOCK_CNTL, PHYDSYMCLK_FORCE_SRC_SEL, mask_sh),\
+	DCCG_SF(PHYESYMCLK_CLOCK_CNTL, PHYESYMCLK_FORCE_EN, mask_sh),\
+	DCCG_SF(PHYESYMCLK_CLOCK_CNTL, PHYESYMCLK_FORCE_SRC_SEL, mask_sh),\
+	DCCG_SF(HDMISTREAMCLK_CNTL, HDMISTREAMCLK0_DTO_FORCE_DIS, mask_sh),\
+	DCCG_SF(DSCCLK_DTO_CTRL, DSCCLK0_DTO_ENABLE, mask_sh),\
+	DCCG_SF(DSCCLK_DTO_CTRL, DSCCLK1_DTO_ENABLE, mask_sh),\
+	DCCG_SF(DSCCLK_DTO_CTRL, DSCCLK2_DTO_ENABLE, mask_sh),\
+	DCCG_SF(DSCCLK_DTO_CTRL, DSCCLK3_DTO_ENABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL2, PHYASYMCLK_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL2, PHYBSYMCLK_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL2, PHYCSYMCLK_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL2, PHYDSYMCLK_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL2, PHYESYMCLK_GATE_DISABLE, mask_sh),\
+	DCCG_SF(HDMISTREAMCLK0_DTO_PARAM, HDMISTREAMCLK0_DTO_MODULO, mask_sh),\
+	DCCG_SF(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_RDIVIDER, mask_sh),\
+	DCCG_SF(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_WDIVIDER, mask_sh)
+
+struct dccg *dccg314_create(
+	struct dc_context *ctx,
+	const struct dccg_registers *regs,
+	const struct dccg_shift *dccg_shift,
+	const struct dccg_mask *dccg_mask);
+
+void dccg314_set_dpstreamclk(
+		struct dccg *dccg,
+		enum streamclk_source src,
+		int otg_inst,
+		int dp_hpo_inst);
+
+#endif //__DCN314_DCCG_H__
diff --git a/drivers/gpu/drm/amd/display/dc/dccg/dcn32/dcn32_dccg.c b/drivers/gpu/drm/amd/display/dc/dccg/dcn32/dcn32_dccg.c
new file mode 100644
index 000000000000..21a6ca5ca192
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dccg/dcn32/dcn32_dccg.c
@@ -0,0 +1,375 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#include "reg_helper.h"
+#include "core_types.h"
+#include "dcn32_dccg.h"
+
+#define TO_DCN_DCCG(dccg)\
+	container_of(dccg, struct dcn_dccg, base)
+
+#define REG(reg) \
+	(dccg_dcn->regs->reg)
+
+#undef FN
+#define FN(reg_name, field_name) \
+	dccg_dcn->dccg_shift->field_name, dccg_dcn->dccg_mask->field_name
+
+#define CTX \
+	dccg_dcn->base.ctx
+#define DC_LOGGER \
+	dccg->ctx->logger
+
+static void dccg32_trigger_dio_fifo_resync(
+	struct dccg *dccg)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+	uint32_t dispclk_rdivider_value = 0;
+
+	REG_GET(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_RDIVIDER, &dispclk_rdivider_value);
+
+	/* Not valid for the WDIVIDER to be set to 0 */
+	if (dispclk_rdivider_value != 0)
+		REG_UPDATE(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_WDIVIDER, dispclk_rdivider_value);
+}
+
+static void dccg32_get_pixel_rate_div(
+		struct dccg *dccg,
+		uint32_t otg_inst,
+		uint32_t *k1,
+		uint32_t *k2)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+	uint32_t val_k1 = PIXEL_RATE_DIV_NA, val_k2 = PIXEL_RATE_DIV_NA;
+
+	*k1 = PIXEL_RATE_DIV_NA;
+	*k2 = PIXEL_RATE_DIV_NA;
+
+	switch (otg_inst) {
+	case 0:
+		REG_GET_2(OTG_PIXEL_RATE_DIV,
+			OTG0_PIXEL_RATE_DIVK1, &val_k1,
+			OTG0_PIXEL_RATE_DIVK2, &val_k2);
+		break;
+	case 1:
+		REG_GET_2(OTG_PIXEL_RATE_DIV,
+			OTG1_PIXEL_RATE_DIVK1, &val_k1,
+			OTG1_PIXEL_RATE_DIVK2, &val_k2);
+		break;
+	case 2:
+		REG_GET_2(OTG_PIXEL_RATE_DIV,
+			OTG2_PIXEL_RATE_DIVK1, &val_k1,
+			OTG2_PIXEL_RATE_DIVK2, &val_k2);
+		break;
+	case 3:
+		REG_GET_2(OTG_PIXEL_RATE_DIV,
+			OTG3_PIXEL_RATE_DIVK1, &val_k1,
+			OTG3_PIXEL_RATE_DIVK2, &val_k2);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+
+	*k1 = val_k1;
+	*k2 = val_k2;
+}
+
+static void dccg32_set_pixel_rate_div(
+		struct dccg *dccg,
+		uint32_t otg_inst,
+		enum pixel_rate_div k1,
+		enum pixel_rate_div k2)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+	uint32_t cur_k1 = PIXEL_RATE_DIV_NA;
+	uint32_t cur_k2 = PIXEL_RATE_DIV_NA;
+
+	// Don't program 0xF into the register field. Not valid since
+	// K1 / K2 field is only 1 / 2 bits wide
+	if (k1 == PIXEL_RATE_DIV_NA || k2 == PIXEL_RATE_DIV_NA) {
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+
+	dccg32_get_pixel_rate_div(dccg, otg_inst, &cur_k1, &cur_k2);
+	if (k1 == cur_k1 && k2 == cur_k2)
+		return;
+
+	switch (otg_inst) {
+	case 0:
+		REG_UPDATE_2(OTG_PIXEL_RATE_DIV,
+				OTG0_PIXEL_RATE_DIVK1, k1,
+				OTG0_PIXEL_RATE_DIVK2, k2);
+		break;
+	case 1:
+		REG_UPDATE_2(OTG_PIXEL_RATE_DIV,
+				OTG1_PIXEL_RATE_DIVK1, k1,
+				OTG1_PIXEL_RATE_DIVK2, k2);
+		break;
+	case 2:
+		REG_UPDATE_2(OTG_PIXEL_RATE_DIV,
+				OTG2_PIXEL_RATE_DIVK1, k1,
+				OTG2_PIXEL_RATE_DIVK2, k2);
+		break;
+	case 3:
+		REG_UPDATE_2(OTG_PIXEL_RATE_DIV,
+				OTG3_PIXEL_RATE_DIVK1, k1,
+				OTG3_PIXEL_RATE_DIVK2, k2);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+}
+
+static void dccg32_set_dtbclk_p_src(
+		struct dccg *dccg,
+		enum streamclk_source src,
+		uint32_t otg_inst)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	uint32_t p_src_sel = 0; /* selects dprefclk */
+	if (src == DTBCLK0)
+		p_src_sel = 2;  /* selects dtbclk0 */
+
+	switch (otg_inst) {
+	case 0:
+		if (src == REFCLK)
+			REG_UPDATE(DTBCLK_P_CNTL,
+					DTBCLK_P0_EN, 0);
+		else
+			REG_UPDATE_2(DTBCLK_P_CNTL,
+					DTBCLK_P0_SRC_SEL, p_src_sel,
+					DTBCLK_P0_EN, 1);
+		break;
+	case 1:
+		if (src == REFCLK)
+			REG_UPDATE(DTBCLK_P_CNTL,
+					DTBCLK_P1_EN, 0);
+		else
+			REG_UPDATE_2(DTBCLK_P_CNTL,
+					DTBCLK_P1_SRC_SEL, p_src_sel,
+					DTBCLK_P1_EN, 1);
+		break;
+	case 2:
+		if (src == REFCLK)
+			REG_UPDATE(DTBCLK_P_CNTL,
+					DTBCLK_P2_EN, 0);
+		else
+			REG_UPDATE_2(DTBCLK_P_CNTL,
+					DTBCLK_P2_SRC_SEL, p_src_sel,
+					DTBCLK_P2_EN, 1);
+		break;
+	case 3:
+		if (src == REFCLK)
+			REG_UPDATE(DTBCLK_P_CNTL,
+					DTBCLK_P3_EN, 0);
+		else
+			REG_UPDATE_2(DTBCLK_P_CNTL,
+					DTBCLK_P3_SRC_SEL, p_src_sel,
+					DTBCLK_P3_EN, 1);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+
+}
+
+/* Controls the generation of pixel valid for OTG in (OTG -> HPO case) */
+static void dccg32_set_dtbclk_dto(
+		struct dccg *dccg,
+		const struct dtbclk_dto_params *params)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+	/* DTO Output Rate / Pixel Rate = 1/4 */
+	int req_dtbclk_khz = params->pixclk_khz / 4;
+
+	if (params->ref_dtbclk_khz && req_dtbclk_khz) {
+		uint32_t modulo, phase;
+
+		// phase / modulo = dtbclk / dtbclk ref
+		modulo = params->ref_dtbclk_khz * 1000;
+		phase = req_dtbclk_khz * 1000;
+
+		REG_WRITE(DTBCLK_DTO_MODULO[params->otg_inst], modulo);
+		REG_WRITE(DTBCLK_DTO_PHASE[params->otg_inst], phase);
+
+		REG_UPDATE(OTG_PIXEL_RATE_CNTL[params->otg_inst],
+				DTBCLK_DTO_ENABLE[params->otg_inst], 1);
+
+		REG_WAIT(OTG_PIXEL_RATE_CNTL[params->otg_inst],
+				DTBCLKDTO_ENABLE_STATUS[params->otg_inst], 1,
+				1, 100);
+
+		/* program OTG_PIXEL_RATE_DIV for DIVK1 and DIVK2 fields */
+		dccg32_set_pixel_rate_div(dccg, params->otg_inst, PIXEL_RATE_DIV_BY_1, PIXEL_RATE_DIV_BY_1);
+
+		/* The recommended programming sequence to enable DTBCLK DTO to generate
+		 * valid pixel HPO DPSTREAM ENCODER, specifies that DTO source select should
+		 * be set only after DTO is enabled
+		 */
+		REG_UPDATE(OTG_PIXEL_RATE_CNTL[params->otg_inst],
+				PIPE_DTO_SRC_SEL[params->otg_inst], 2);
+	} else {
+		REG_UPDATE_2(OTG_PIXEL_RATE_CNTL[params->otg_inst],
+				DTBCLK_DTO_ENABLE[params->otg_inst], 0,
+				PIPE_DTO_SRC_SEL[params->otg_inst], params->is_hdmi ? 0 : 1);
+		REG_WRITE(DTBCLK_DTO_MODULO[params->otg_inst], 0);
+		REG_WRITE(DTBCLK_DTO_PHASE[params->otg_inst], 0);
+	}
+}
+
+static void dccg32_set_valid_pixel_rate(
+		struct dccg *dccg,
+		int ref_dtbclk_khz,
+		int otg_inst,
+		int pixclk_khz)
+{
+	struct dtbclk_dto_params dto_params = {0};
+
+	dto_params.ref_dtbclk_khz = ref_dtbclk_khz;
+	dto_params.otg_inst = otg_inst;
+	dto_params.pixclk_khz = pixclk_khz;
+	dto_params.is_hdmi = true;
+
+	dccg32_set_dtbclk_dto(dccg, &dto_params);
+}
+
+static void dccg32_get_dccg_ref_freq(struct dccg *dccg,
+		unsigned int xtalin_freq_inKhz,
+		unsigned int *dccg_ref_freq_inKhz)
+{
+	/*
+	 * Assume refclk is sourced from xtalin
+	 * expect 100MHz
+	 */
+	*dccg_ref_freq_inKhz = xtalin_freq_inKhz;
+	return;
+}
+
+static void dccg32_set_dpstreamclk(
+		struct dccg *dccg,
+		enum streamclk_source src,
+		int otg_inst,
+		int dp_hpo_inst)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	/* set the dtbclk_p source */
+	/* always program refclk as DTBCLK. No use-case expected to require DPREFCLK as refclk */
+	dccg32_set_dtbclk_p_src(dccg, DTBCLK0, otg_inst);
+
+	/* enabled to select one of the DTBCLKs for pipe */
+	switch (dp_hpo_inst) {
+	case 0:
+		REG_UPDATE_2(DPSTREAMCLK_CNTL,
+			     DPSTREAMCLK0_EN,
+			     (src == REFCLK) ? 0 : 1, DPSTREAMCLK0_SRC_SEL, otg_inst);
+		break;
+	case 1:
+		REG_UPDATE_2(DPSTREAMCLK_CNTL, DPSTREAMCLK1_EN,
+			     (src == REFCLK) ? 0 : 1, DPSTREAMCLK1_SRC_SEL, otg_inst);
+		break;
+	case 2:
+		REG_UPDATE_2(DPSTREAMCLK_CNTL, DPSTREAMCLK2_EN,
+			     (src == REFCLK) ? 0 : 1, DPSTREAMCLK2_SRC_SEL, otg_inst);
+		break;
+	case 3:
+		REG_UPDATE_2(DPSTREAMCLK_CNTL, DPSTREAMCLK3_EN,
+			     (src == REFCLK) ? 0 : 1, DPSTREAMCLK3_SRC_SEL, otg_inst);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+}
+
+static void dccg32_otg_add_pixel(struct dccg *dccg,
+		uint32_t otg_inst)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	REG_UPDATE(OTG_PIXEL_RATE_CNTL[otg_inst],
+			OTG_ADD_PIXEL[otg_inst], 1);
+}
+
+static void dccg32_otg_drop_pixel(struct dccg *dccg,
+		uint32_t otg_inst)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	REG_UPDATE(OTG_PIXEL_RATE_CNTL[otg_inst],
+			OTG_DROP_PIXEL[otg_inst], 1);
+}
+
+static const struct dccg_funcs dccg32_funcs = {
+	.update_dpp_dto = dccg2_update_dpp_dto,
+	.get_dccg_ref_freq = dccg32_get_dccg_ref_freq,
+	.dccg_init = dccg31_init,
+	.set_dpstreamclk = dccg32_set_dpstreamclk,
+	.enable_symclk32_se = dccg31_enable_symclk32_se,
+	.disable_symclk32_se = dccg31_disable_symclk32_se,
+	.enable_symclk32_le = dccg31_enable_symclk32_le,
+	.disable_symclk32_le = dccg31_disable_symclk32_le,
+	.set_physymclk = dccg31_set_physymclk,
+	.set_dtbclk_dto = dccg32_set_dtbclk_dto,
+	.set_valid_pixel_rate = dccg32_set_valid_pixel_rate,
+	.set_fifo_errdet_ovr_en = dccg2_set_fifo_errdet_ovr_en,
+	.set_audio_dtbclk_dto = dccg31_set_audio_dtbclk_dto,
+	.otg_add_pixel = dccg32_otg_add_pixel,
+	.otg_drop_pixel = dccg32_otg_drop_pixel,
+	.set_pixel_rate_div = dccg32_set_pixel_rate_div,
+	.get_pixel_rate_div = dccg32_get_pixel_rate_div,
+	.trigger_dio_fifo_resync = dccg32_trigger_dio_fifo_resync,
+	.set_dtbclk_p_src = dccg32_set_dtbclk_p_src,
+};
+
+struct dccg *dccg32_create(
+	struct dc_context *ctx,
+	const struct dccg_registers *regs,
+	const struct dccg_shift *dccg_shift,
+	const struct dccg_mask *dccg_mask)
+{
+	struct dcn_dccg *dccg_dcn = kzalloc(sizeof(*dccg_dcn), GFP_KERNEL);
+	struct dccg *base;
+
+	if (dccg_dcn == NULL) {
+		BREAK_TO_DEBUGGER();
+		return NULL;
+	}
+
+	base = &dccg_dcn->base;
+	base->ctx = ctx;
+	base->funcs = &dccg32_funcs;
+
+	dccg_dcn->regs = regs;
+	dccg_dcn->dccg_shift = dccg_shift;
+	dccg_dcn->dccg_mask = dccg_mask;
+
+	return &dccg_dcn->base;
+}
diff --git a/drivers/gpu/drm/amd/display/dc/dccg/dcn32/dcn32_dccg.h b/drivers/gpu/drm/amd/display/dc/dccg/dcn32/dcn32_dccg.h
new file mode 100644
index 000000000000..cf5508718122
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dccg/dcn32/dcn32_dccg.h
@@ -0,0 +1,125 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __DCN32_DCCG_H__
+#define __DCN32_DCCG_H__
+
+#include "dcn31/dcn31_dccg.h"
+
+#define DCCG_SFII(block, reg_name, field_prefix, field_name, inst, post_fix)\
+	.field_prefix ## _ ## field_name[inst] = block ## inst ## _ ## reg_name ## __ ## field_prefix ## inst ## _ ## field_name ## post_fix
+
+#define DCCG_MASK_SH_LIST_DCN32(mask_sh) \
+	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_ENABLE, DPPCLK, 0, mask_sh),\
+	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_DB_EN, DPPCLK, 0, mask_sh),\
+	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_ENABLE, DPPCLK, 1, mask_sh),\
+	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_DB_EN, DPPCLK, 1, mask_sh),\
+	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_ENABLE, DPPCLK, 2, mask_sh),\
+	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_DB_EN, DPPCLK, 2, mask_sh),\
+	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_ENABLE, DPPCLK, 3, mask_sh),\
+	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_DB_EN, DPPCLK, 3, mask_sh),\
+	DCCG_SF(DPPCLK0_DTO_PARAM, DPPCLK0_DTO_PHASE, mask_sh),\
+	DCCG_SF(DPPCLK0_DTO_PARAM, DPPCLK0_DTO_MODULO, mask_sh),\
+	DCCG_SF(HDMICHARCLK0_CLOCK_CNTL, HDMICHARCLK0_EN, mask_sh),\
+	DCCG_SF(HDMICHARCLK0_CLOCK_CNTL, HDMICHARCLK0_SRC_SEL, mask_sh),\
+	DCCG_SF(PHYASYMCLK_CLOCK_CNTL, PHYASYMCLK_FORCE_EN, mask_sh),\
+	DCCG_SF(PHYASYMCLK_CLOCK_CNTL, PHYASYMCLK_FORCE_SRC_SEL, mask_sh),\
+	DCCG_SF(PHYBSYMCLK_CLOCK_CNTL, PHYBSYMCLK_FORCE_EN, mask_sh),\
+	DCCG_SF(PHYBSYMCLK_CLOCK_CNTL, PHYBSYMCLK_FORCE_SRC_SEL, mask_sh),\
+	DCCG_SF(PHYCSYMCLK_CLOCK_CNTL, PHYCSYMCLK_FORCE_EN, mask_sh),\
+	DCCG_SF(PHYCSYMCLK_CLOCK_CNTL, PHYCSYMCLK_FORCE_SRC_SEL, mask_sh),\
+	DCCG_SF(PHYDSYMCLK_CLOCK_CNTL, PHYDSYMCLK_FORCE_EN, mask_sh),\
+	DCCG_SF(PHYDSYMCLK_CLOCK_CNTL, PHYDSYMCLK_FORCE_SRC_SEL, mask_sh),\
+	DCCG_SF(PHYESYMCLK_CLOCK_CNTL, PHYESYMCLK_FORCE_EN, mask_sh),\
+	DCCG_SF(PHYESYMCLK_CLOCK_CNTL, PHYESYMCLK_FORCE_SRC_SEL, mask_sh),\
+	DCCG_SF(DPSTREAMCLK_CNTL, DPSTREAMCLK0_EN, mask_sh),\
+	DCCG_SF(DPSTREAMCLK_CNTL, DPSTREAMCLK1_EN, mask_sh),\
+	DCCG_SF(DPSTREAMCLK_CNTL, DPSTREAMCLK2_EN, mask_sh),\
+	DCCG_SF(DPSTREAMCLK_CNTL, DPSTREAMCLK3_EN, mask_sh),\
+	DCCG_SF(DPSTREAMCLK_CNTL, DPSTREAMCLK0_SRC_SEL, mask_sh),\
+	DCCG_SF(DPSTREAMCLK_CNTL, DPSTREAMCLK1_SRC_SEL, mask_sh),\
+	DCCG_SF(DPSTREAMCLK_CNTL, DPSTREAMCLK2_SRC_SEL, mask_sh),\
+	DCCG_SF(DPSTREAMCLK_CNTL, DPSTREAMCLK3_SRC_SEL, mask_sh),\
+	DCCG_SF(HDMISTREAMCLK_CNTL, HDMISTREAMCLK0_EN, mask_sh),\
+	DCCG_SF(HDMISTREAMCLK_CNTL, HDMISTREAMCLK0_DTO_FORCE_DIS, mask_sh),\
+	DCCG_SF(HDMISTREAMCLK_CNTL, HDMISTREAMCLK0_SRC_SEL, mask_sh),\
+	DCCG_SF(SYMCLK32_SE_CNTL, SYMCLK32_SE0_SRC_SEL, mask_sh),\
+	DCCG_SF(SYMCLK32_SE_CNTL, SYMCLK32_SE1_SRC_SEL, mask_sh),\
+	DCCG_SF(SYMCLK32_SE_CNTL, SYMCLK32_SE2_SRC_SEL, mask_sh),\
+	DCCG_SF(SYMCLK32_SE_CNTL, SYMCLK32_SE3_SRC_SEL, mask_sh),\
+	DCCG_SF(SYMCLK32_SE_CNTL, SYMCLK32_SE0_EN, mask_sh),\
+	DCCG_SF(SYMCLK32_SE_CNTL, SYMCLK32_SE1_EN, mask_sh),\
+	DCCG_SF(SYMCLK32_SE_CNTL, SYMCLK32_SE2_EN, mask_sh),\
+	DCCG_SF(SYMCLK32_SE_CNTL, SYMCLK32_SE3_EN, mask_sh),\
+	DCCG_SF(SYMCLK32_LE_CNTL, SYMCLK32_LE0_SRC_SEL, mask_sh),\
+	DCCG_SF(SYMCLK32_LE_CNTL, SYMCLK32_LE1_SRC_SEL, mask_sh),\
+	DCCG_SF(SYMCLK32_LE_CNTL, SYMCLK32_LE0_EN, mask_sh),\
+	DCCG_SF(SYMCLK32_LE_CNTL, SYMCLK32_LE1_EN, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, DTBCLK_DTO, ENABLE, 0, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, DTBCLK_DTO, ENABLE, 1, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, DTBCLK_DTO, ENABLE, 2, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, DTBCLK_DTO, ENABLE, 3, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, DTBCLKDTO, ENABLE_STATUS, 0, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, DTBCLKDTO, ENABLE_STATUS, 1, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, DTBCLKDTO, ENABLE_STATUS, 2, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, DTBCLKDTO, ENABLE_STATUS, 3, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, PIPE, DTO_SRC_SEL, 0, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, PIPE, DTO_SRC_SEL, 1, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, PIPE, DTO_SRC_SEL, 2, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, PIPE, DTO_SRC_SEL, 3, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, ADD_PIXEL, 0, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, ADD_PIXEL, 1, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, ADD_PIXEL, 2, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, ADD_PIXEL, 3, mask_sh),\
+	DCCG_SF(OTG_PIXEL_RATE_DIV, OTG0_PIXEL_RATE_DIVK1, mask_sh),\
+	DCCG_SF(OTG_PIXEL_RATE_DIV, OTG0_PIXEL_RATE_DIVK2, mask_sh),\
+	DCCG_SF(OTG_PIXEL_RATE_DIV, OTG1_PIXEL_RATE_DIVK1, mask_sh),\
+	DCCG_SF(OTG_PIXEL_RATE_DIV, OTG1_PIXEL_RATE_DIVK2, mask_sh),\
+	DCCG_SF(OTG_PIXEL_RATE_DIV, OTG2_PIXEL_RATE_DIVK1, mask_sh),\
+	DCCG_SF(OTG_PIXEL_RATE_DIV, OTG2_PIXEL_RATE_DIVK2, mask_sh),\
+	DCCG_SF(OTG_PIXEL_RATE_DIV, OTG3_PIXEL_RATE_DIVK1, mask_sh),\
+	DCCG_SF(OTG_PIXEL_RATE_DIV, OTG3_PIXEL_RATE_DIVK2, mask_sh),\
+	DCCG_SF(OTG_PIXEL_RATE_DIV, OTG3_PIXEL_RATE_DIVK2, mask_sh),\
+	DCCG_SF(DTBCLK_P_CNTL, DTBCLK_P0_SRC_SEL, mask_sh),\
+	DCCG_SF(DTBCLK_P_CNTL, DTBCLK_P0_EN, mask_sh),\
+	DCCG_SF(DTBCLK_P_CNTL, DTBCLK_P1_SRC_SEL, mask_sh),\
+	DCCG_SF(DTBCLK_P_CNTL, DTBCLK_P1_EN, mask_sh),\
+	DCCG_SF(DTBCLK_P_CNTL, DTBCLK_P2_SRC_SEL, mask_sh),\
+	DCCG_SF(DTBCLK_P_CNTL, DTBCLK_P2_EN, mask_sh),\
+	DCCG_SF(DTBCLK_P_CNTL, DTBCLK_P3_SRC_SEL, mask_sh),\
+	DCCG_SF(DTBCLK_P_CNTL, DTBCLK_P3_EN, mask_sh),\
+	DCCG_SF(DCCG_AUDIO_DTO_SOURCE, DCCG_AUDIO_DTO_SEL, mask_sh),\
+	DCCG_SF(DCCG_AUDIO_DTO_SOURCE, DCCG_AUDIO_DTO0_SOURCE_SEL, mask_sh),\
+	DCCG_SF(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_CHG_DONE, mask_sh),\
+	DCCG_SF(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_RDIVIDER, mask_sh),\
+	DCCG_SF(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_WDIVIDER, mask_sh)
+
+struct dccg *dccg32_create(
+	struct dc_context *ctx,
+	const struct dccg_registers *regs,
+	const struct dccg_shift *dccg_shift,
+	const struct dccg_mask *dccg_mask);
+
+#endif //__DCN32_DCCG_H__
diff --git a/drivers/gpu/drm/amd/display/dc/dccg/dcn35/dcn35_dccg.c b/drivers/gpu/drm/amd/display/dc/dccg/dcn35/dcn35_dccg.c
new file mode 100644
index 000000000000..bd2f528137b2
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dccg/dcn35/dcn35_dccg.c
@@ -0,0 +1,2486 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "reg_helper.h"
+#include "core_types.h"
+#include "resource.h"
+#include "dcn35_dccg.h"
+
+#define TO_DCN_DCCG(dccg)\
+	container_of(dccg, struct dcn_dccg, base)
+
+#define REG(reg) \
+	(dccg_dcn->regs->reg)
+
+#undef FN
+#define FN(reg_name, field_name) \
+	dccg_dcn->dccg_shift->field_name, dccg_dcn->dccg_mask->field_name
+
+#define CTX \
+	dccg_dcn->base.ctx
+#include "logger_types.h"
+#define DC_LOGGER \
+	dccg->ctx->logger
+
+enum symclk_fe_source {
+	SYMCLK_FE_SYMCLK_A = 0,	// Select functional clock from backend symclk A
+	SYMCLK_FE_SYMCLK_B,
+	SYMCLK_FE_SYMCLK_C,
+	SYMCLK_FE_SYMCLK_D,
+	SYMCLK_FE_SYMCLK_E,
+	SYMCLK_FE_REFCLK = 0xFF,	// Arbitrary value to pass refclk selection in software
+};
+
+enum symclk_be_source {
+	SYMCLK_BE_PHYCLK = 0,	// Select phy clk when sym_clk_enable = 1
+	SYMCLK_BE_DPIACLK_810 = 4,
+	SYMCLK_BE_DPIACLK_162 = 5,
+	SYMCLK_BE_DPIACLK_540 = 6,
+	SYMCLK_BE_DPIACLK_270 = 7,
+	SYMCLK_BE_REFCLK = 0xFF,	// Arbitrary value to pass refclk selection in software
+};
+
+enum physymclk_source {
+	PHYSYMCLK_PHYCLK = 0,		// Select symclk as source of clock which is output to PHY through DCIO.
+	PHYSYMCLK_PHYD18CLK,		// Select phyd18clk as the source of clock which is output to PHY through DCIO.
+	PHYSYMCLK_PHYD32CLK,		// Select phyd32clk as the source of clock which is output to PHY through DCIO.
+	PHYSYMCLK_REFCLK = 0xFF,	// Arbitrary value to pass refclk selection in software
+};
+
+enum dtbclk_source {
+	DTBCLK_DPREFCLK = 0,		// Selects source for DTBCLK_P# as DPREFCLK (src sel 0 and 1 are same)
+	DTBCLK_DPREFCLK_0,			// Selects source for DTBCLK_P# as DPREFCLK (src sel 0 and 1 are same)
+	DTBCLK_DTBCLK0,				// Selects source for DTBCLK_P# as DTBCLK0
+	DTBCLK_DTBCLK1,				// Selects source for DTBCLK_P# as DTBCLK0
+	DTBCLK_REFCLK = 0xFF,		// Arbitrary value to pass refclk selection in software
+};
+
+enum dppclk_clock_source {
+	DPP_REFCLK = 0,				// refclk is selected
+	DPP_DCCG_DTO,				// Functional clock selected is DTO tuned DPPCLK
+};
+
+enum dp_stream_clk_source {
+	DP_STREAM_DTBCLK_P0 = 0,	// Selects functional for DP_STREAM_CLK as DTBCLK_P#
+	DP_STREAM_DTBCLK_P1,
+	DP_STREAM_DTBCLK_P2,
+	DP_STREAM_DTBCLK_P3,
+	DP_STREAM_DTBCLK_P4,
+	DP_STREAM_DTBCLK_P5,
+	DP_STREAM_REFCLK = 0xFF,	// Arbitrary value to pass refclk selection in software
+};
+
+enum hdmi_char_clk {
+	HDMI_CHAR_PHYAD18CLK = 0,	// Selects functional for hdmi_char_clk as UNIPHYA PHYD18CLK
+	HDMI_CHAR_PHYBD18CLK,
+	HDMI_CHAR_PHYCD18CLK,
+	HDMI_CHAR_PHYDD18CLK,
+	HDMI_CHAR_PHYED18CLK,
+	HDMI_CHAR_REFCLK = 0xFF,	// Arbitrary value to pass refclk selection in software
+};
+
+enum hdmi_stream_clk_source {
+	HDMI_STREAM_DTBCLK_P0 = 0,	// Selects functional for HDMI_STREAM_CLK as DTBCLK_P#
+	HDMI_STREAM_DTBCLK_P1,
+	HDMI_STREAM_DTBCLK_P2,
+	HDMI_STREAM_DTBCLK_P3,
+	HDMI_STREAM_DTBCLK_P4,
+	HDMI_STREAM_DTBCLK_P5,
+	HDMI_STREAM_REFCLK = 0xFF,	// Arbitrary value to pass refclk selection in software
+};
+
+enum symclk32_se_clk_source {
+	SYMCLK32_SE_PHYAD32CLK = 0,	// Selects functional for SYMCLK32 as UNIPHYA PHYD32CLK
+	SYMCLK32_SE_PHYBD32CLK,
+	SYMCLK32_SE_PHYCD32CLK,
+	SYMCLK32_SE_PHYDD32CLK,
+	SYMCLK32_SE_PHYED32CLK,
+	SYMCLK32_SE_REFCLK = 0xFF,	// Arbitrary value to pass refclk selection in software
+};
+
+enum symclk32_le_clk_source {
+	SYMCLK32_LE_PHYAD32CLK = 0,	// Selects functional for SYMCLK32 as UNIPHYA PHYD32CLK
+	SYMCLK32_LE_PHYBD32CLK,
+	SYMCLK32_LE_PHYCD32CLK,
+	SYMCLK32_LE_PHYDD32CLK,
+	SYMCLK32_LE_PHYED32CLK,
+	SYMCLK32_LE_REFCLK = 0xFF,	// Arbitrary value to pass refclk selection in software
+};
+
+enum dsc_clk_source {
+	DSC_CLK_REF_CLK = 0,			// Ref clock selected for DSC_CLK
+	DSC_DTO_TUNED_CK_GPU_DISCLK_3,	// DTO divided clock selected as functional clock
+};
+
+
+static void dccg35_set_dsc_clk_rcg(struct dccg *dccg, int inst, bool allow_rcg)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	if (!dccg->ctx->dc->debug.root_clock_optimization.bits.dsc && allow_rcg)
+		return;
+
+	switch (inst) {
+	case 0:
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DSCCLK0_ROOT_GATE_DISABLE, allow_rcg ? 0 : 1);
+		break;
+	case 1:
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DSCCLK1_ROOT_GATE_DISABLE, allow_rcg ? 0 : 1);
+		break;
+	case 2:
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DSCCLK2_ROOT_GATE_DISABLE, allow_rcg ? 0 : 1);
+		break;
+	case 3:
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DSCCLK3_ROOT_GATE_DISABLE, allow_rcg ? 0 : 1);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+
+	/* Wait for clock to ramp */
+	if (!allow_rcg)
+		udelay(10);
+}
+
+static void dccg35_set_symclk32_se_rcg(
+	struct dccg *dccg,
+	int inst,
+	bool enable)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	if (!dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se && enable)
+		return;
+
+	/* SYMCLK32_ROOT_SE#_GATE_DISABLE will clock gate in DCCG */
+	/* SYMCLK32_SE#_GATE_DISABLE will clock gate in HPO only */
+	switch (inst) {
+	case 0:
+		REG_UPDATE_2(DCCG_GATE_DISABLE_CNTL3,
+				   SYMCLK32_SE0_GATE_DISABLE, enable ? 0 : 1,
+				   SYMCLK32_ROOT_SE0_GATE_DISABLE, enable ? 0 : 1);
+		break;
+	case 1:
+		REG_UPDATE_2(DCCG_GATE_DISABLE_CNTL3,
+				   SYMCLK32_SE1_GATE_DISABLE, enable ? 0 : 1,
+				   SYMCLK32_ROOT_SE1_GATE_DISABLE, enable ? 0 : 1);
+		break;
+	case 2:
+		REG_UPDATE_2(DCCG_GATE_DISABLE_CNTL3,
+				   SYMCLK32_SE2_GATE_DISABLE, enable ? 0 : 1,
+				   SYMCLK32_ROOT_SE2_GATE_DISABLE, enable ? 0 : 1);
+		break;
+	case 3:
+		REG_UPDATE_2(DCCG_GATE_DISABLE_CNTL3,
+				   SYMCLK32_SE3_GATE_DISABLE, enable ? 0 : 1,
+				   SYMCLK32_ROOT_SE3_GATE_DISABLE, enable ? 0 : 1);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+}
+
+static void dccg35_set_symclk32_le_rcg(
+	struct dccg *dccg,
+	int inst,
+	bool enable)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	if (!dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_le && enable)
+		return;
+
+	switch (inst) {
+	case 0:
+		REG_UPDATE_2(DCCG_GATE_DISABLE_CNTL3,
+				   SYMCLK32_LE0_GATE_DISABLE, enable ? 0 : 1,
+				   SYMCLK32_ROOT_LE0_GATE_DISABLE, enable ? 0 : 1);
+		break;
+	case 1:
+		REG_UPDATE_2(DCCG_GATE_DISABLE_CNTL3,
+				   SYMCLK32_LE1_GATE_DISABLE, enable ? 0 : 1,
+				   SYMCLK32_ROOT_LE1_GATE_DISABLE, enable ? 0 : 1);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+}
+
+static void dccg35_set_physymclk_rcg(
+	struct dccg *dccg,
+	int inst,
+	bool enable)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	if (!dccg->ctx->dc->debug.root_clock_optimization.bits.physymclk && enable)
+		return;
+
+	switch (inst) {
+	case 0:
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL2,
+				PHYASYMCLK_ROOT_GATE_DISABLE, enable ? 0 : 1);
+		break;
+	case 1:
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL2,
+				PHYBSYMCLK_ROOT_GATE_DISABLE, enable ? 0 : 1);
+		break;
+	case 2:
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL2,
+				PHYCSYMCLK_ROOT_GATE_DISABLE, enable ? 0 : 1);
+		break;
+	case 3:
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL2,
+				PHYDSYMCLK_ROOT_GATE_DISABLE, enable ? 0 : 1);
+		break;
+	case 4:
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL2,
+				PHYESYMCLK_ROOT_GATE_DISABLE, enable ? 0 : 1);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+}
+
+static void dccg35_set_symclk_fe_rcg(
+	struct dccg *dccg,
+	int inst,
+	bool enable)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	if (!dccg->ctx->dc->debug.root_clock_optimization.bits.symclk_fe && enable)
+		return;
+
+	switch (inst) {
+	case 0:
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL2,
+				   SYMCLKA_FE_GATE_DISABLE, enable ? 0 : 1);
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL5,
+				   SYMCLKA_FE_ROOT_GATE_DISABLE, enable ? 0 : 1);
+		break;
+	case 1:
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL2,
+				   SYMCLKB_FE_GATE_DISABLE, enable ? 0 : 1);
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL5,
+				   SYMCLKB_FE_ROOT_GATE_DISABLE, enable ? 0 : 1);
+		break;
+	case 2:
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL2,
+				   SYMCLKC_FE_GATE_DISABLE, enable ? 0 : 1);
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL5,
+				   SYMCLKC_FE_ROOT_GATE_DISABLE, enable ? 0 : 1);
+		break;
+	case 3:
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL2,
+				   SYMCLKD_FE_GATE_DISABLE, enable ? 0 : 1);
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL5,
+				   SYMCLKD_FE_ROOT_GATE_DISABLE, enable ? 0 : 1);
+		break;
+	case 4:
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL2,
+				   SYMCLKE_FE_GATE_DISABLE, enable ? 0 : 1);
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL5,
+				   SYMCLKE_FE_ROOT_GATE_DISABLE, enable ? 0 : 1);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+}
+
+static void dccg35_set_symclk_be_rcg(
+	struct dccg *dccg,
+	int inst,
+	bool enable)
+{
+
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	/* TBD add symclk_be in rcg control bits */
+	if (!dccg->ctx->dc->debug.root_clock_optimization.bits.symclk_fe && enable)
+		return;
+
+	switch (inst) {
+	case 0:
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL2,
+				   SYMCLKA_GATE_DISABLE, enable ? 0 : 1);
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL5,
+				   SYMCLKA_ROOT_GATE_DISABLE, enable ? 0 : 1);
+		break;
+	case 1:
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL2,
+				   SYMCLKB_GATE_DISABLE, enable ? 0 : 1);
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL5,
+				   SYMCLKB_ROOT_GATE_DISABLE, enable ? 0 : 1);
+		break;
+	case 2:
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL2,
+				   SYMCLKC_GATE_DISABLE, enable ? 0 : 1);
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL5,
+				   SYMCLKC_ROOT_GATE_DISABLE, enable ? 0 : 1);
+		break;
+	case 3:
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL2,
+				   SYMCLKD_GATE_DISABLE, enable ? 0 : 1);
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL5,
+				   SYMCLKD_ROOT_GATE_DISABLE, enable ? 0 : 1);
+		break;
+	case 4:
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL2,
+				   SYMCLKE_GATE_DISABLE, enable ? 0 : 1);
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL5,
+				   SYMCLKE_ROOT_GATE_DISABLE, enable ? 0 : 1);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+}
+
+static void dccg35_set_dtbclk_p_rcg(struct dccg *dccg, int inst, bool enable)
+{
+
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	if (!dccg->ctx->dc->debug.root_clock_optimization.bits.dpp && enable)
+		return;
+
+	switch (inst) {
+	case 0:
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, DTBCLK_P0_GATE_DISABLE, enable ? 0 : 1);
+		break;
+	case 1:
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, DTBCLK_P1_GATE_DISABLE, enable ? 0 : 1);
+		break;
+	case 2:
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, DTBCLK_P2_GATE_DISABLE, enable ? 0 : 1);
+		break;
+	case 3:
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, DTBCLK_P3_GATE_DISABLE, enable ? 0 : 1);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		break;
+	}
+}
+
+static void dccg35_set_dppclk_rcg(struct dccg *dccg, int inst, bool allow_rcg)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	if (!dccg->ctx->dc->debug.root_clock_optimization.bits.dpp && allow_rcg)
+		return;
+
+	switch (inst) {
+	case 0:
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DPPCLK0_ROOT_GATE_DISABLE, allow_rcg ? 0 : 1);
+		break;
+	case 1:
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DPPCLK1_ROOT_GATE_DISABLE, allow_rcg ? 0 : 1);
+		break;
+	case 2:
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DPPCLK2_ROOT_GATE_DISABLE, allow_rcg ? 0 : 1);
+		break;
+	case 3:
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DPPCLK3_ROOT_GATE_DISABLE, allow_rcg ? 0 : 1);
+		break;
+	default:
+	BREAK_TO_DEBUGGER();
+		break;
+	}
+
+	/* Wait for clock to ramp */
+	if (!allow_rcg)
+		udelay(10);
+}
+
+static void dccg35_set_dpstreamclk_rcg(
+	struct dccg *dccg,
+	int inst,
+	bool enable)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	if (!dccg->ctx->dc->debug.root_clock_optimization.bits.dpstream && enable)
+		return;
+
+	switch (inst) {
+	case 0:
+		REG_UPDATE_2(DCCG_GATE_DISABLE_CNTL5,
+					 DPSTREAMCLK0_GATE_DISABLE, enable ? 0 : 1,
+					 DPSTREAMCLK0_ROOT_GATE_DISABLE, enable ? 0 : 1);
+		break;
+	case 1:
+		REG_UPDATE_2(DCCG_GATE_DISABLE_CNTL5,
+					 DPSTREAMCLK1_GATE_DISABLE, enable ? 0 : 1,
+					 DPSTREAMCLK1_ROOT_GATE_DISABLE, enable ? 0 : 1);
+		break;
+	case 2:
+		REG_UPDATE_2(DCCG_GATE_DISABLE_CNTL5,
+				   DPSTREAMCLK2_GATE_DISABLE, enable ? 0 : 1,
+				   DPSTREAMCLK2_ROOT_GATE_DISABLE, enable ? 0 : 1);
+		break;
+	case 3:
+		REG_UPDATE_2(DCCG_GATE_DISABLE_CNTL5,
+				   DPSTREAMCLK3_GATE_DISABLE, enable ? 0 : 1,
+				   DPSTREAMCLK3_ROOT_GATE_DISABLE, enable ? 0 : 1);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+}
+
+static void dccg35_set_smclk32_se_rcg(
+		struct dccg *dccg,
+		int inst,
+		bool enable)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	if (!dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se && enable)
+		return;
+
+	switch (inst) {
+	case 0:
+		REG_UPDATE_2(DCCG_GATE_DISABLE_CNTL3,
+					 SYMCLK32_SE0_GATE_DISABLE, enable ? 0 : 1,
+					 SYMCLK32_ROOT_SE0_GATE_DISABLE, enable ? 0 : 1);
+		break;
+	case 1:
+		REG_UPDATE_2(DCCG_GATE_DISABLE_CNTL3,
+					 SYMCLK32_SE1_GATE_DISABLE, enable ? 0 : 1,
+					 SYMCLK32_ROOT_SE1_GATE_DISABLE, enable ? 0 : 1);
+		break;
+	case 2:
+		REG_UPDATE_2(DCCG_GATE_DISABLE_CNTL3,
+					 SYMCLK32_SE2_GATE_DISABLE, enable ? 0 : 1,
+					 SYMCLK32_ROOT_SE2_GATE_DISABLE, enable ? 0 : 1);
+		break;
+	case 3:
+		REG_UPDATE_2(DCCG_GATE_DISABLE_CNTL3,
+					 SYMCLK32_SE3_GATE_DISABLE, enable ? 0 : 1,
+					 SYMCLK32_ROOT_SE3_GATE_DISABLE, enable ? 0 : 1);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+}
+
+static void dccg35_set_dsc_clk_src_new(struct dccg *dccg, int inst, enum dsc_clk_source src)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	/* DSCCLK#_EN=0 switches to refclock from functional clock */
+
+	switch (inst) {
+	case 0:
+		REG_UPDATE(DSCCLK_DTO_CTRL, DSCCLK0_EN, src);
+		break;
+	case 1:
+		REG_UPDATE(DSCCLK_DTO_CTRL, DSCCLK1_EN, src);
+		break;
+	case 2:
+		REG_UPDATE(DSCCLK_DTO_CTRL, DSCCLK2_EN, src);
+		break;
+	case 3:
+		REG_UPDATE(DSCCLK_DTO_CTRL, DSCCLK3_EN, src);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+}
+
+static void dccg35_set_symclk32_se_src_new(
+	struct dccg *dccg,
+	int inst,
+	enum symclk32_se_clk_source src
+	)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	switch (inst) {
+	case 0:
+		REG_UPDATE_2(SYMCLK32_SE_CNTL,
+					 SYMCLK32_SE0_SRC_SEL, (src == SYMCLK32_SE_REFCLK) ? 0 : src,
+					 SYMCLK32_SE0_EN,  (src == SYMCLK32_SE_REFCLK) ? 0 : 1);
+		break;
+	case 1:
+		REG_UPDATE_2(SYMCLK32_SE_CNTL,
+					 SYMCLK32_SE1_SRC_SEL, (src == SYMCLK32_SE_REFCLK) ? 0 : src,
+					 SYMCLK32_SE1_EN, (src == SYMCLK32_SE_REFCLK) ? 0 : 1);
+		break;
+	case 2:
+		REG_UPDATE_2(SYMCLK32_SE_CNTL,
+					 SYMCLK32_SE2_SRC_SEL, (src == SYMCLK32_SE_REFCLK) ? 0 : src,
+					 SYMCLK32_SE2_EN, (src == SYMCLK32_SE_REFCLK) ? 0 : 1);
+		break;
+	case 3:
+		REG_UPDATE_2(SYMCLK32_SE_CNTL,
+					 SYMCLK32_SE3_SRC_SEL, (src == SYMCLK32_SE_REFCLK) ? 0 : src,
+					 SYMCLK32_SE3_EN, (src == SYMCLK32_SE_REFCLK) ? 0 : 1);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+}
+
+static int
+dccg35_is_symclk32_se_src_functional_le_new(struct dccg *dccg, int symclk_32_se_inst, int symclk_32_le_inst)
+{
+	uint32_t en;
+	uint32_t src_sel;
+
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	REG_GET_2(SYMCLK32_SE_CNTL, SYMCLK32_SE3_SRC_SEL, &src_sel, SYMCLK32_SE3_EN, &en);
+
+	if (en == 1 && src_sel == symclk_32_le_inst)
+		return 1;
+
+	return 0;
+}
+
+
+static void dccg35_set_symclk32_le_src_new(
+	struct dccg *dccg,
+	int inst,
+	enum symclk32_le_clk_source src)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	switch (inst) {
+	case 0:
+		REG_UPDATE_2(SYMCLK32_LE_CNTL,
+					 SYMCLK32_LE0_SRC_SEL, (src == SYMCLK32_LE_REFCLK) ? 0 : src,
+					 SYMCLK32_LE0_EN, (src == SYMCLK32_LE_REFCLK) ? 0 : 1);
+		break;
+	case 1:
+		REG_UPDATE_2(SYMCLK32_LE_CNTL,
+					 SYMCLK32_LE1_SRC_SEL, (src == SYMCLK32_LE_REFCLK) ? 0 : src,
+					 SYMCLK32_LE1_EN, (src == SYMCLK32_LE_REFCLK) ? 0 : 1);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+}
+
+static void dcn35_set_dppclk_src_new(struct dccg *dccg,
+				 int inst, enum dppclk_clock_source src)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	switch (inst) {
+	case 0:
+		REG_UPDATE(DPPCLK_CTRL, DPPCLK0_EN, src);
+		break;
+	case 1:
+		REG_UPDATE(DPPCLK_CTRL, DPPCLK1_EN, src);
+		break;
+	case 2:
+		REG_UPDATE(DPPCLK_CTRL, DPPCLK2_EN, src);
+		break;
+	case 3:
+		REG_UPDATE(DPPCLK_CTRL, DPPCLK3_EN, src);
+		break;
+	default:
+	BREAK_TO_DEBUGGER();
+		break;
+	}
+}
+
+static void dccg35_set_dtbclk_p_src_new(
+	struct dccg *dccg,
+	enum dtbclk_source src,
+	int inst)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	/* If DTBCLK_P#_EN is 0 refclock is selected as functional clock
+	 * If DTBCLK_P#_EN is 1 functional clock is selected as DTBCLK_P#_SRC_SEL
+	 */
+
+	switch (inst) {
+	case 0:
+		REG_UPDATE_2(DTBCLK_P_CNTL,
+					 DTBCLK_P0_SRC_SEL, (src == DTBCLK_REFCLK) ? 0 : src,
+					 DTBCLK_P0_EN, (src == DTBCLK_REFCLK) ? 0 : 1);
+		break;
+	case 1:
+		REG_UPDATE_2(DTBCLK_P_CNTL,
+					 DTBCLK_P1_SRC_SEL, (src == DTBCLK_REFCLK) ? 0 : src,
+					 DTBCLK_P1_EN, (src == DTBCLK_REFCLK) ? 0 : 1);
+		break;
+	case 2:
+		REG_UPDATE_2(DTBCLK_P_CNTL,
+					 DTBCLK_P2_SRC_SEL, (src == DTBCLK_REFCLK) ? 0 : src,
+					 DTBCLK_P2_EN, (src == DTBCLK_REFCLK) ? 0 : 1);
+		break;
+	case 3:
+		REG_UPDATE_2(DTBCLK_P_CNTL,
+					 DTBCLK_P3_SRC_SEL, (src == DTBCLK_REFCLK) ? 0 : src,
+					 DTBCLK_P3_EN, (src == DTBCLK_REFCLK) ? 0 : 1);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+}
+
+static void dccg35_set_dpstreamclk_src_new(
+	struct dccg *dccg,
+	enum dp_stream_clk_source src,
+	int inst)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	switch (inst) {
+	case 0:
+		REG_UPDATE_2(DPSTREAMCLK_CNTL, DPSTREAMCLK0_EN,
+					 (src == DP_STREAM_REFCLK) ? 0 : 1,
+					 DPSTREAMCLK0_SRC_SEL,
+					 (src == DP_STREAM_REFCLK) ? 0 : src);
+		break;
+	case 1:
+		REG_UPDATE_2(DPSTREAMCLK_CNTL, DPSTREAMCLK1_EN,
+					 (src == DP_STREAM_REFCLK) ? 0 : 1,
+					 DPSTREAMCLK1_SRC_SEL,
+					 (src == DP_STREAM_REFCLK) ? 0 : src);
+
+		break;
+	case 2:
+		REG_UPDATE_2(DPSTREAMCLK_CNTL, DPSTREAMCLK2_EN,
+					 (src == DP_STREAM_REFCLK) ? 0 : 1,
+					 DPSTREAMCLK2_SRC_SEL,
+					 (src == DP_STREAM_REFCLK) ? 0 : src);
+
+		break;
+	case 3:
+		REG_UPDATE_2(DPSTREAMCLK_CNTL, DPSTREAMCLK3_EN,
+					 (src == DP_STREAM_REFCLK) ? 0 : 1,
+					 DPSTREAMCLK3_SRC_SEL,
+					 (src == DP_STREAM_REFCLK) ? 0 : src);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+}
+
+static void dccg35_set_physymclk_src_new(
+	struct dccg *dccg,
+	enum physymclk_source src,
+	int inst)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	switch (inst) {
+	case 0:
+		REG_UPDATE_2(PHYASYMCLK_CLOCK_CNTL, PHYASYMCLK_EN,
+					 (src == PHYSYMCLK_REFCLK) ? 0 : 1,
+					 PHYASYMCLK_SRC_SEL,
+					 (src == PHYSYMCLK_REFCLK) ? 0 : src);
+		break;
+	case 1:
+		REG_UPDATE_2(PHYBSYMCLK_CLOCK_CNTL, PHYBSYMCLK_EN,
+					 (src == PHYSYMCLK_REFCLK) ? 0 : 1,
+					 PHYBSYMCLK_SRC_SEL,
+					 (src == PHYSYMCLK_REFCLK) ? 0 : src);
+		break;
+	case 2:
+		REG_UPDATE_2(PHYCSYMCLK_CLOCK_CNTL, PHYCSYMCLK_EN,
+					 (src == PHYSYMCLK_REFCLK) ? 0 : 1,
+					 PHYCSYMCLK_SRC_SEL,
+					 (src == PHYSYMCLK_REFCLK) ? 0 : src);
+		break;
+	case 3:
+		REG_UPDATE_2(PHYDSYMCLK_CLOCK_CNTL, PHYDSYMCLK_EN,
+					 (src == PHYSYMCLK_REFCLK) ? 0 : 1,
+					 PHYDSYMCLK_SRC_SEL,
+					 (src == PHYSYMCLK_REFCLK) ? 0 : src);
+		break;
+	case 4:
+		REG_UPDATE_2(PHYESYMCLK_CLOCK_CNTL, PHYESYMCLK_EN,
+					 (src == PHYSYMCLK_REFCLK) ? 0 : 1,
+					 PHYESYMCLK_SRC_SEL,
+					 (src == PHYSYMCLK_REFCLK) ? 0 : src);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+}
+
+static void dccg35_set_symclk_be_src_new(
+	struct dccg *dccg,
+	enum symclk_be_source src,
+	int inst)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	switch (inst) {
+	case 0:
+		REG_UPDATE_2(SYMCLKA_CLOCK_ENABLE,
+					 SYMCLKA_CLOCK_ENABLE, (src == SYMCLK_BE_REFCLK) ? 0 : 1,
+					 SYMCLKA_SRC_SEL, (src == SYMCLK_BE_REFCLK) ? 0 : src);
+		break;
+	case 1:
+		REG_UPDATE_2(SYMCLKB_CLOCK_ENABLE,
+					 SYMCLKB_CLOCK_ENABLE, (src == SYMCLK_BE_REFCLK) ? 0 : 1,
+					 SYMCLKB_SRC_SEL, (src == SYMCLK_BE_REFCLK) ? 0 : src);
+		break;
+	case 2:
+		REG_UPDATE_2(SYMCLKC_CLOCK_ENABLE,
+					 SYMCLKC_CLOCK_ENABLE, (src == SYMCLK_BE_REFCLK) ? 0 : 1,
+					 SYMCLKC_SRC_SEL, (src == SYMCLK_BE_REFCLK) ? 0 : src);
+		break;
+	case 3:
+		REG_UPDATE_2(SYMCLKD_CLOCK_ENABLE,
+					 SYMCLKD_CLOCK_ENABLE, (src == SYMCLK_BE_REFCLK) ? 0 : 1,
+					 SYMCLKD_SRC_SEL, (src == SYMCLK_BE_REFCLK) ? 0 : src);
+		break;
+	case 4:
+		REG_UPDATE_2(SYMCLKE_CLOCK_ENABLE,
+					 SYMCLKE_CLOCK_ENABLE, (src == SYMCLK_BE_REFCLK) ? 0 : 1,
+					 SYMCLKE_SRC_SEL, (src == SYMCLK_BE_REFCLK) ? 0 : src);
+		break;
+	}
+}
+
+static int dccg35_is_symclk_fe_src_functional_be(struct dccg *dccg,
+												 int symclk_fe_inst,
+												 int symclk_be_inst)
+{
+
+	uint32_t en = 0;
+	uint32_t src_sel = 0;
+
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	switch (symclk_fe_inst) {
+	case 0:
+		REG_GET_2(SYMCLKA_CLOCK_ENABLE, SYMCLKA_FE_SRC_SEL, &src_sel, SYMCLKA_FE_EN, &en);
+		break;
+	case 1:
+		REG_GET_2(SYMCLKB_CLOCK_ENABLE, SYMCLKB_FE_SRC_SEL, &src_sel, SYMCLKB_FE_EN, &en);
+		break;
+	case 2:
+		REG_GET_2(SYMCLKC_CLOCK_ENABLE, SYMCLKC_FE_SRC_SEL, &src_sel, SYMCLKC_FE_EN, &en);
+		break;
+	case 3:
+		REG_GET_2(SYMCLKD_CLOCK_ENABLE, SYMCLKD_FE_SRC_SEL, &src_sel, SYMCLKD_FE_EN, &en);
+		break;
+	case 4:
+		REG_GET_2(SYMCLKE_CLOCK_ENABLE, SYMCLKE_FE_SRC_SEL, &src_sel, SYMCLKE_FE_EN, &en);
+		break;
+	}
+
+	if (en == 1 && src_sel == symclk_be_inst)
+		return 1;
+
+	return 0;
+}
+
+static void dccg35_set_symclk_fe_src_new(struct dccg *dccg, enum symclk_fe_source src, int inst)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	switch (inst) {
+	case 0:
+		REG_UPDATE_2(SYMCLKA_CLOCK_ENABLE,
+					 SYMCLKA_FE_EN, (src == SYMCLK_FE_REFCLK) ? 0 : 1,
+					 SYMCLKA_FE_SRC_SEL, (src == SYMCLK_FE_REFCLK) ? 0 : src);
+		break;
+	case 1:
+		REG_UPDATE_2(SYMCLKB_CLOCK_ENABLE,
+					 SYMCLKB_FE_EN, (src == SYMCLK_FE_REFCLK) ? 0 : 1,
+					 SYMCLKB_FE_SRC_SEL, (src == SYMCLK_FE_REFCLK) ? 0 : src);
+		break;
+	case 2:
+		REG_UPDATE_2(SYMCLKC_CLOCK_ENABLE,
+					 SYMCLKC_FE_EN, (src == SYMCLK_FE_REFCLK) ? 0 : 1,
+					 SYMCLKC_FE_SRC_SEL, (src == SYMCLK_FE_REFCLK) ? 0 : src);
+		break;
+	case 3:
+		REG_UPDATE_2(SYMCLKD_CLOCK_ENABLE,
+					 SYMCLKD_FE_EN, (src == SYMCLK_FE_REFCLK) ? 0 : 1,
+					 SYMCLKD_FE_SRC_SEL, (src == SYMCLK_FE_REFCLK) ? 0 : src);
+		break;
+	case 4:
+		REG_UPDATE_2(SYMCLKE_CLOCK_ENABLE,
+					 SYMCLKE_FE_EN, (src == SYMCLK_FE_REFCLK) ? 0 : 1,
+					 SYMCLKE_FE_SRC_SEL, (src == SYMCLK_FE_REFCLK) ? 0 : src);
+		break;
+	}
+}
+
+static uint32_t dccg35_is_fe_rcg(struct dccg *dccg, int inst)
+{
+	uint32_t enable = 0;
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	switch (inst) {
+	case 0:
+		REG_GET(DCCG_GATE_DISABLE_CNTL5,
+				SYMCLKA_FE_ROOT_GATE_DISABLE, &enable);
+		break;
+	case 1:
+		REG_GET(DCCG_GATE_DISABLE_CNTL5,
+				SYMCLKB_FE_ROOT_GATE_DISABLE, &enable);
+		break;
+	case 2:
+		REG_GET(DCCG_GATE_DISABLE_CNTL5,
+				SYMCLKC_FE_ROOT_GATE_DISABLE, &enable);
+		break;
+	case 3:
+		REG_GET(DCCG_GATE_DISABLE_CNTL5,
+				SYMCLKD_FE_ROOT_GATE_DISABLE, &enable);
+		break;
+	case 4:
+		REG_GET(DCCG_GATE_DISABLE_CNTL5,
+				SYMCLKE_FE_ROOT_GATE_DISABLE, &enable);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		break;
+	}
+	return enable;
+}
+
+static uint32_t dccg35_is_symclk32_se_rcg(struct dccg *dccg, int inst)
+{
+	uint32_t disable_l1 = 0;
+	uint32_t disable_l2 = 0;
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	switch (inst) {
+	case 0:
+		REG_GET_2(DCCG_GATE_DISABLE_CNTL3,
+				  SYMCLK32_SE0_GATE_DISABLE, &disable_l1,
+				  SYMCLK32_ROOT_SE0_GATE_DISABLE, &disable_l2);
+		break;
+	case 1:
+		REG_GET_2(DCCG_GATE_DISABLE_CNTL3,
+				  SYMCLK32_SE1_GATE_DISABLE, &disable_l1,
+				  SYMCLK32_ROOT_SE1_GATE_DISABLE, &disable_l2);
+		break;
+	case 2:
+		REG_GET_2(DCCG_GATE_DISABLE_CNTL3,
+				  SYMCLK32_SE2_GATE_DISABLE, &disable_l1,
+				  SYMCLK32_ROOT_SE2_GATE_DISABLE, &disable_l2);
+		break;
+	case 3:
+		REG_GET_2(DCCG_GATE_DISABLE_CNTL3,
+				  SYMCLK32_SE3_GATE_DISABLE, &disable_l1,
+				  SYMCLK32_ROOT_SE3_GATE_DISABLE, &disable_l2);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return 0;
+	}
+
+	/* return true if either block level or DCCG level gating is active */
+	return (disable_l1 | disable_l2);
+}
+
+static void dccg35_enable_symclk_fe_new(
+	struct dccg *dccg,
+	int inst,
+	enum symclk_fe_source src)
+{
+	dccg35_set_symclk_fe_rcg(dccg, inst, false);
+	dccg35_set_symclk_fe_src_new(dccg, src, inst);
+}
+
+static void dccg35_disable_symclk_fe_new(
+	struct dccg *dccg,
+	int inst)
+{
+	dccg35_set_symclk_fe_src_new(dccg, SYMCLK_FE_REFCLK, inst);
+	dccg35_set_symclk_fe_rcg(dccg, inst, true);
+}
+
+static void dccg35_enable_symclk_be_new(
+	struct dccg *dccg,
+	int inst,
+	enum symclk_be_source src)
+{
+	dccg35_set_symclk_be_rcg(dccg, inst, false);
+	dccg35_set_symclk_be_src_new(dccg, inst, src);
+}
+
+static void dccg35_disable_symclk_be_new(
+	struct dccg *dccg,
+	int inst)
+{
+	int i;
+
+	/* Switch from functional clock to refclock */
+	dccg35_set_symclk_be_src_new(dccg, inst, SYMCLK_BE_REFCLK);
+
+	/* Check if any other SE connected LE and disable them */
+	for (i = 0; i < 4; i++) {
+		/* Make sure FE is not already in RCG */
+		if (dccg35_is_fe_rcg(dccg, i) == 0) {
+			if (dccg35_is_symclk_fe_src_functional_be(dccg, i, inst))
+				dccg35_disable_symclk_fe_new(dccg, i);
+		}
+	}
+	/* Safe to RCG SYMCLK*/
+	dccg35_set_symclk_be_rcg(dccg, inst, true);
+}
+
+static void dccg35_enable_symclk32_se_new(
+	struct dccg *dccg,
+	int inst,
+	enum symclk32_se_clk_source src)
+{
+	dccg35_set_symclk32_se_rcg(dccg, inst, false);
+	dccg35_set_symclk32_se_src_new(dccg, inst, src);
+}
+
+static void dccg35_disable_symclk32_se_new(
+	struct dccg *dccg,
+	int inst)
+{
+	dccg35_set_symclk32_se_src_new(dccg, SYMCLK32_SE_REFCLK, inst);
+	dccg35_set_symclk32_se_rcg(dccg, inst, true);
+}
+
+static void dccg35_enable_symclk32_le_new(
+	struct dccg *dccg,
+	int inst,
+	enum symclk32_le_clk_source src)
+{
+	dccg35_set_symclk32_le_rcg(dccg, inst, false);
+	dccg35_set_symclk32_le_src_new(dccg, inst, src);
+}
+
+static void dccg35_disable_symclk32_le_new(
+	struct dccg *dccg,
+	int inst)
+{
+	int i;
+
+	/* Switch from functional clock to refclock */
+	dccg35_set_symclk32_le_src_new(dccg, inst, SYMCLK32_LE_REFCLK);
+
+	/* Check if any SE are connected and disable SE as well */
+	for (i = 0; i < 4; i++) {
+		/* Make sure FE is not already in RCG */
+		if (dccg35_is_symclk32_se_rcg(dccg, i) == 0) {
+			/* Disable and SE connected to this LE before RCG */
+			if (dccg35_is_symclk32_se_src_functional_le_new(dccg, i, inst))
+				dccg35_disable_symclk32_se_new(dccg, i);
+		}
+	}
+	/* Safe to RCG SYM32_LE*/
+	dccg35_set_symclk32_le_rcg(dccg, inst, true);
+}
+
+static void dccg35_enable_physymclk_new(struct dccg *dccg,
+					int inst,
+					enum physymclk_source src)
+{
+	dccg35_set_physymclk_rcg(dccg, inst, false);
+	dccg35_set_physymclk_src_new(dccg, src, inst);
+}
+
+static void dccg35_disable_physymclk_new(struct dccg *dccg,
+										 int inst)
+{
+	dccg35_set_physymclk_src_new(dccg, PHYSYMCLK_REFCLK, inst);
+	dccg35_set_physymclk_rcg(dccg, inst, true);
+}
+
+static void dccg35_enable_dpp_clk_new(
+	struct dccg *dccg,
+	int inst,
+	enum dppclk_clock_source src)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+	/* Sanitize inst before use in array de-ref */
+	if (inst < 0) {
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+	dccg35_set_dppclk_rcg(dccg, inst, false);
+	dcn35_set_dppclk_src_new(dccg, inst, src);
+	/* Switch DPP clock to DTO */
+	REG_SET_2(DPPCLK_DTO_PARAM[inst], 0,
+			  DPPCLK0_DTO_PHASE, 0xFF,
+			  DPPCLK0_DTO_MODULO, 0xFF);
+}
+
+
+static void dccg35_disable_dpp_clk_new(
+	struct dccg *dccg,
+	int inst)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+	/* Sanitize inst before use in array de-ref */
+	if (inst < 0) {
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+	dcn35_set_dppclk_src_new(dccg, inst, DPP_REFCLK);
+	REG_SET_2(DPPCLK_DTO_PARAM[inst], 0,
+			  DPPCLK0_DTO_PHASE, 0,
+			  DPPCLK0_DTO_MODULO, 1);
+	dccg35_set_dppclk_rcg(dccg, inst, true);
+}
+
+static void dccg35_disable_dscclk_new(struct dccg *dccg,
+									  int inst)
+{
+	dccg35_set_dsc_clk_src_new(dccg, inst, DSC_CLK_REF_CLK);
+	dccg35_set_dsc_clk_rcg(dccg, inst, true);
+}
+
+static void dccg35_enable_dscclk_new(struct dccg *dccg,
+									 int inst,
+									 enum dsc_clk_source src)
+{
+	dccg35_set_dsc_clk_rcg(dccg, inst, false);
+	dccg35_set_dsc_clk_src_new(dccg, inst, src);
+}
+
+static void dccg35_enable_dtbclk_p_new(struct dccg *dccg,
+									   enum dtbclk_source src,
+									   int inst)
+{
+	dccg35_set_dtbclk_p_rcg(dccg, inst, false);
+	dccg35_set_dtbclk_p_src_new(dccg, src, inst);
+}
+
+static void dccg35_disable_dtbclk_p_new(struct dccg *dccg,
+										int inst)
+{
+	dccg35_set_dtbclk_p_src_new(dccg, DTBCLK_REFCLK, inst);
+	dccg35_set_dtbclk_p_rcg(dccg, inst, true);
+}
+
+static void dccg35_disable_dpstreamclk_new(struct dccg *dccg,
+										  int inst)
+{
+	dccg35_set_dpstreamclk_src_new(dccg, DP_STREAM_REFCLK, inst);
+	dccg35_set_dpstreamclk_rcg(dccg, inst, true);
+}
+
+static void dccg35_enable_dpstreamclk_new(struct dccg *dccg,
+										   enum dp_stream_clk_source src,
+										   int inst)
+{
+	dccg35_set_dpstreamclk_rcg(dccg, inst, false);
+	dccg35_set_dpstreamclk_src_new(dccg, src, inst);
+}
+
+static void dccg35_trigger_dio_fifo_resync(struct dccg *dccg)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+	uint32_t dispclk_rdivider_value = 0;
+
+	REG_GET(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_RDIVIDER, &dispclk_rdivider_value);
+	if (dispclk_rdivider_value != 0)
+		REG_UPDATE(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_WDIVIDER, dispclk_rdivider_value);
+}
+static void dccg35_wait_for_dentist_change_done(
+	struct dccg *dccg)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	uint32_t dentist_dispclk_value = REG_READ(DENTIST_DISPCLK_CNTL);
+
+	REG_WRITE(DENTIST_DISPCLK_CNTL, dentist_dispclk_value);
+	REG_WAIT(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_CHG_DONE, 1, 50, 2000);
+}
+
+static void dcn35_set_dppclk_enable(struct dccg *dccg,
+				 uint32_t dpp_inst, uint32_t enable)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+
+	switch (dpp_inst) {
+	case 0:
+		REG_UPDATE(DPPCLK_CTRL, DPPCLK0_EN, enable);
+		break;
+	case 1:
+		REG_UPDATE(DPPCLK_CTRL, DPPCLK1_EN, enable);
+		break;
+	case 2:
+		REG_UPDATE(DPPCLK_CTRL, DPPCLK2_EN, enable);
+		break;
+	case 3:
+		REG_UPDATE(DPPCLK_CTRL, DPPCLK3_EN, enable);
+		break;
+	default:
+		break;
+	}
+	DC_LOG_DEBUG("%s: dpp_inst(%d) DPPCLK_EN = %d\n", __func__, dpp_inst, enable);
+
+}
+
+static void dccg35_update_dpp_dto(struct dccg *dccg, int dpp_inst,
+				  int req_dppclk)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	if (dccg->dpp_clock_gated[dpp_inst]) {
+		/*
+		 * Do not update the DPPCLK DTO if the clock is stopped.
+		 */
+		return;
+	}
+
+	if (dccg->ref_dppclk && req_dppclk) {
+		int ref_dppclk = dccg->ref_dppclk;
+		int modulo, phase;
+
+		// phase / modulo = dpp pipe clk / dpp global clk
+		modulo = 0xff;   // use FF at the end
+		phase = ((modulo * req_dppclk) + ref_dppclk - 1) / ref_dppclk;
+
+		if (phase > 0xff) {
+			ASSERT(false);
+			phase = 0xff;
+		}
+		dccg35_set_dppclk_rcg(dccg, dpp_inst, false);
+
+		REG_SET_2(DPPCLK_DTO_PARAM[dpp_inst], 0,
+				DPPCLK0_DTO_PHASE, phase,
+				DPPCLK0_DTO_MODULO, modulo);
+
+		dcn35_set_dppclk_enable(dccg, dpp_inst, true);
+	} else {
+		dcn35_set_dppclk_enable(dccg, dpp_inst, false);
+		dccg35_set_dppclk_rcg(dccg, dpp_inst, true);
+	}
+	udelay(10);
+	dccg->pipe_dppclk_khz[dpp_inst] = req_dppclk;
+}
+
+static void dccg35_set_dppclk_root_clock_gating(struct dccg *dccg,
+		 uint32_t dpp_inst, uint32_t disallow_rcg)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	if (!dccg->ctx->dc->debug.root_clock_optimization.bits.dpp && !disallow_rcg)
+		return;
+
+
+	switch (dpp_inst) {
+	case 0:
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DPPCLK0_ROOT_GATE_DISABLE, disallow_rcg);
+		break;
+	case 1:
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DPPCLK1_ROOT_GATE_DISABLE, disallow_rcg);
+		break;
+	case 2:
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DPPCLK2_ROOT_GATE_DISABLE, disallow_rcg);
+		break;
+	case 3:
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DPPCLK3_ROOT_GATE_DISABLE, disallow_rcg);
+		break;
+	default:
+		break;
+	}
+
+	/* Wait for clock to ramp */
+	if (disallow_rcg)
+		udelay(10);
+}
+
+static void dccg35_get_pixel_rate_div(
+		struct dccg *dccg,
+		uint32_t otg_inst,
+		uint32_t *k1,
+		uint32_t *k2)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+	uint32_t val_k1 = PIXEL_RATE_DIV_NA, val_k2 = PIXEL_RATE_DIV_NA;
+
+	*k1 = PIXEL_RATE_DIV_NA;
+	*k2 = PIXEL_RATE_DIV_NA;
+
+	switch (otg_inst) {
+	case 0:
+		REG_GET_2(OTG_PIXEL_RATE_DIV,
+			OTG0_PIXEL_RATE_DIVK1, &val_k1,
+			OTG0_PIXEL_RATE_DIVK2, &val_k2);
+		break;
+	case 1:
+		REG_GET_2(OTG_PIXEL_RATE_DIV,
+			OTG1_PIXEL_RATE_DIVK1, &val_k1,
+			OTG1_PIXEL_RATE_DIVK2, &val_k2);
+		break;
+	case 2:
+		REG_GET_2(OTG_PIXEL_RATE_DIV,
+			OTG2_PIXEL_RATE_DIVK1, &val_k1,
+			OTG2_PIXEL_RATE_DIVK2, &val_k2);
+		break;
+	case 3:
+		REG_GET_2(OTG_PIXEL_RATE_DIV,
+			OTG3_PIXEL_RATE_DIVK1, &val_k1,
+			OTG3_PIXEL_RATE_DIVK2, &val_k2);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+
+	*k1 = val_k1;
+	*k2 = val_k2;
+}
+
+static void dccg35_set_pixel_rate_div(
+		struct dccg *dccg,
+		uint32_t otg_inst,
+		enum pixel_rate_div k1,
+		enum pixel_rate_div k2)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+	uint32_t cur_k1 = PIXEL_RATE_DIV_NA;
+	uint32_t cur_k2 = PIXEL_RATE_DIV_NA;
+
+
+	// Don't program 0xF into the register field. Not valid since
+	// K1 / K2 field is only 1 / 2 bits wide
+	if (k1 == PIXEL_RATE_DIV_NA || k2 == PIXEL_RATE_DIV_NA) {
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+
+	dccg35_get_pixel_rate_div(dccg, otg_inst, &cur_k1, &cur_k2);
+	if (k1 == cur_k1 && k2 == cur_k2)
+		return;
+
+	switch (otg_inst) {
+	case 0:
+		REG_UPDATE_2(OTG_PIXEL_RATE_DIV,
+				OTG0_PIXEL_RATE_DIVK1, k1,
+				OTG0_PIXEL_RATE_DIVK2, k2);
+		break;
+	case 1:
+		REG_UPDATE_2(OTG_PIXEL_RATE_DIV,
+				OTG1_PIXEL_RATE_DIVK1, k1,
+				OTG1_PIXEL_RATE_DIVK2, k2);
+		break;
+	case 2:
+		REG_UPDATE_2(OTG_PIXEL_RATE_DIV,
+				OTG2_PIXEL_RATE_DIVK1, k1,
+				OTG2_PIXEL_RATE_DIVK2, k2);
+		break;
+	case 3:
+		REG_UPDATE_2(OTG_PIXEL_RATE_DIV,
+				OTG3_PIXEL_RATE_DIVK1, k1,
+				OTG3_PIXEL_RATE_DIVK2, k2);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+	if (otg_inst < 4)
+		dccg35_wait_for_dentist_change_done(dccg);
+}
+
+static void dccg35_set_dtbclk_p_src(
+		struct dccg *dccg,
+		enum streamclk_source src,
+		uint32_t otg_inst)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	uint32_t p_src_sel = 0; /* selects dprefclk */
+	if (src == DTBCLK0)
+		p_src_sel = 2;  /* selects dtbclk0 */
+
+	switch (otg_inst) {
+	case 0:
+		if (src == REFCLK)
+			REG_UPDATE(DTBCLK_P_CNTL,
+					DTBCLK_P0_EN, 0);
+		else
+			REG_UPDATE_2(DTBCLK_P_CNTL,
+					DTBCLK_P0_SRC_SEL, p_src_sel,
+					DTBCLK_P0_EN, 1);
+		break;
+	case 1:
+		if (src == REFCLK)
+			REG_UPDATE(DTBCLK_P_CNTL,
+					DTBCLK_P1_EN, 0);
+		else
+			REG_UPDATE_2(DTBCLK_P_CNTL,
+					DTBCLK_P1_SRC_SEL, p_src_sel,
+					DTBCLK_P1_EN, 1);
+		break;
+	case 2:
+		if (src == REFCLK)
+			REG_UPDATE(DTBCLK_P_CNTL,
+					DTBCLK_P2_EN, 0);
+		else
+			REG_UPDATE_2(DTBCLK_P_CNTL,
+					DTBCLK_P2_SRC_SEL, p_src_sel,
+					DTBCLK_P2_EN, 1);
+		break;
+	case 3:
+		if (src == REFCLK)
+			REG_UPDATE(DTBCLK_P_CNTL,
+					DTBCLK_P3_EN, 0);
+		else
+			REG_UPDATE_2(DTBCLK_P_CNTL,
+					DTBCLK_P3_SRC_SEL, p_src_sel,
+					DTBCLK_P3_EN, 1);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+
+}
+
+/* Controls the generation of pixel valid for OTG in (OTG -> HPO case) */
+static void dccg35_set_dtbclk_dto(
+		struct dccg *dccg,
+		const struct dtbclk_dto_params *params)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+	/* DTO Output Rate / Pixel Rate = 1/4 */
+	int req_dtbclk_khz = params->pixclk_khz / 4;
+
+	if (params->ref_dtbclk_khz && req_dtbclk_khz) {
+		uint32_t modulo, phase;
+
+		switch (params->otg_inst) {
+		case 0:
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, DTBCLK_P0_GATE_DISABLE, 1);
+			break;
+		case 1:
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, DTBCLK_P1_GATE_DISABLE, 1);
+			break;
+		case 2:
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, DTBCLK_P2_GATE_DISABLE, 1);
+			break;
+		case 3:
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, DTBCLK_P3_GATE_DISABLE, 1);
+			break;
+		}
+
+		// phase / modulo = dtbclk / dtbclk ref
+		modulo = params->ref_dtbclk_khz * 1000;
+		phase = req_dtbclk_khz * 1000;
+
+		REG_WRITE(DTBCLK_DTO_MODULO[params->otg_inst], modulo);
+		REG_WRITE(DTBCLK_DTO_PHASE[params->otg_inst], phase);
+
+		REG_UPDATE(OTG_PIXEL_RATE_CNTL[params->otg_inst],
+				DTBCLK_DTO_ENABLE[params->otg_inst], 1);
+
+		REG_WAIT(OTG_PIXEL_RATE_CNTL[params->otg_inst],
+				DTBCLKDTO_ENABLE_STATUS[params->otg_inst], 1,
+				1, 100);
+
+		/* program OTG_PIXEL_RATE_DIV for DIVK1 and DIVK2 fields */
+		dccg35_set_pixel_rate_div(dccg, params->otg_inst, PIXEL_RATE_DIV_BY_1, PIXEL_RATE_DIV_BY_1);
+
+		/* The recommended programming sequence to enable DTBCLK DTO to generate
+		 * valid pixel HPO DPSTREAM ENCODER, specifies that DTO source select should
+		 * be set only after DTO is enabled.
+		 * PIPEx_DTO_SRC_SEL should not be programmed during DTBCLK update since OTG may still be on, and the
+		 * programming is handled in program_pix_clk() regardless, so it can be removed from here.
+		 */
+		DC_LOG_DEBUG("%s: OTG%d DTBCLK DTO enabled: pixclk_khz=%d, ref_dtbclk_khz=%d, req_dtbclk_khz=%d, phase=%d, modulo=%d\n",
+				__func__, params->otg_inst, params->pixclk_khz,
+				params->ref_dtbclk_khz, req_dtbclk_khz, phase, modulo);
+
+	} else if (!params->ref_dtbclk_khz && !req_dtbclk_khz) {
+		switch (params->otg_inst) {
+		case 0:
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, DTBCLK_P0_GATE_DISABLE, 0);
+			break;
+		case 1:
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, DTBCLK_P1_GATE_DISABLE, 0);
+			break;
+		case 2:
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, DTBCLK_P2_GATE_DISABLE, 0);
+			break;
+		case 3:
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, DTBCLK_P3_GATE_DISABLE, 0);
+			break;
+		}
+
+		/**
+		 * PIPEx_DTO_SRC_SEL should not be programmed during DTBCLK update since OTG may still be on, and the
+		 * programming is handled in program_pix_clk() regardless, so it can be removed from here.
+		 */
+		REG_UPDATE(OTG_PIXEL_RATE_CNTL[params->otg_inst],
+				DTBCLK_DTO_ENABLE[params->otg_inst], 0);
+
+		REG_WRITE(DTBCLK_DTO_MODULO[params->otg_inst], 0);
+		REG_WRITE(DTBCLK_DTO_PHASE[params->otg_inst], 0);
+
+		DC_LOG_DEBUG("%s: OTG%d DTBCLK DTO disabled\n", __func__, params->otg_inst);
+	}
+}
+
+static void dccg35_set_dpstreamclk(
+		struct dccg *dccg,
+		enum streamclk_source src,
+		int otg_inst,
+		int dp_hpo_inst)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	/* set the dtbclk_p source */
+	dccg35_set_dtbclk_p_src(dccg, src, otg_inst);
+
+	/* enabled to select one of the DTBCLKs for pipe */
+	switch (dp_hpo_inst) {
+	case 0:
+		REG_UPDATE_2(DPSTREAMCLK_CNTL, DPSTREAMCLK0_EN,
+				(src == REFCLK) ? 0 : 1, DPSTREAMCLK0_SRC_SEL, otg_inst);
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.dpstream)
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, DPSTREAMCLK0_ROOT_GATE_DISABLE, (src == REFCLK) ? 0 : 1);
+		break;
+	case 1:
+		REG_UPDATE_2(DPSTREAMCLK_CNTL, DPSTREAMCLK1_EN,
+				(src == REFCLK) ? 0 : 1, DPSTREAMCLK1_SRC_SEL, otg_inst);
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.dpstream)
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, DPSTREAMCLK1_ROOT_GATE_DISABLE, (src == REFCLK) ? 0 : 1);
+		break;
+	case 2:
+		REG_UPDATE_2(DPSTREAMCLK_CNTL, DPSTREAMCLK2_EN,
+				(src == REFCLK) ? 0 : 1, DPSTREAMCLK2_SRC_SEL, otg_inst);
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.dpstream)
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, DPSTREAMCLK2_ROOT_GATE_DISABLE, (src == REFCLK) ? 0 : 1);
+		break;
+	case 3:
+		REG_UPDATE_2(DPSTREAMCLK_CNTL, DPSTREAMCLK3_EN,
+				(src == REFCLK) ? 0 : 1, DPSTREAMCLK3_SRC_SEL, otg_inst);
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.dpstream)
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, DPSTREAMCLK3_ROOT_GATE_DISABLE, (src == REFCLK) ? 0 : 1);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+	DC_LOG_DEBUG("%s: dp_hpo_inst(%d) DPSTREAMCLK_EN = %d, DPSTREAMCLK_SRC_SEL = %d\n",
+			__func__, dp_hpo_inst, (src == REFCLK) ? 0 : 1, otg_inst);
+}
+
+
+static void dccg35_set_dpstreamclk_root_clock_gating(
+		struct dccg *dccg,
+		int dp_hpo_inst,
+		bool enable)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	switch (dp_hpo_inst) {
+	case 0:
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.dpstream) {
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, DPSTREAMCLK0_ROOT_GATE_DISABLE, enable ? 1 : 0);
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, DPSTREAMCLK0_GATE_DISABLE, enable ? 1 : 0);
+		}
+		break;
+	case 1:
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.dpstream) {
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, DPSTREAMCLK1_ROOT_GATE_DISABLE, enable ? 1 : 0);
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, DPSTREAMCLK1_GATE_DISABLE, enable ? 1 : 0);
+		}
+		break;
+	case 2:
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.dpstream) {
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, DPSTREAMCLK2_ROOT_GATE_DISABLE, enable ? 1 : 0);
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, DPSTREAMCLK2_GATE_DISABLE, enable ? 1 : 0);
+		}
+		break;
+	case 3:
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.dpstream) {
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, DPSTREAMCLK3_ROOT_GATE_DISABLE, enable ? 1 : 0);
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, DPSTREAMCLK3_GATE_DISABLE, enable ? 1 : 0);
+		}
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+	DC_LOG_DEBUG("%s: dp_hpo_inst(%d) DPSTREAMCLK_ROOT_GATE_DISABLE = %d\n",
+			__func__, dp_hpo_inst, enable ? 1 : 0);
+}
+
+
+
+static void dccg35_set_physymclk_root_clock_gating(
+		struct dccg *dccg,
+		int phy_inst,
+		bool enable)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	if (!dccg->ctx->dc->debug.root_clock_optimization.bits.physymclk)
+		return;
+
+	switch (phy_inst) {
+	case 0:
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL2,
+				PHYASYMCLK_ROOT_GATE_DISABLE, enable ? 0 : 1);
+		break;
+	case 1:
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL2,
+				PHYBSYMCLK_ROOT_GATE_DISABLE, enable ? 0 : 1);
+		break;
+	case 2:
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL2,
+				PHYCSYMCLK_ROOT_GATE_DISABLE, enable ? 0 : 1);
+		break;
+	case 3:
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL2,
+				PHYDSYMCLK_ROOT_GATE_DISABLE, enable ? 0 : 1);
+		break;
+	case 4:
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL2,
+				PHYESYMCLK_ROOT_GATE_DISABLE, enable ? 0 : 1);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+	DC_LOG_DEBUG("%s: dpp_inst(%d) PHYESYMCLK_ROOT_GATE_DISABLE: %d\n", __func__, phy_inst, enable ? 0 : 1);
+
+}
+
+static void dccg35_set_physymclk(
+		struct dccg *dccg,
+		int phy_inst,
+		enum physymclk_clock_source clk_src,
+		bool force_enable)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	/* Force PHYSYMCLK on and Select phyd32clk as the source of clock which is output to PHY through DCIO */
+	switch (phy_inst) {
+	case 0:
+		if (force_enable) {
+			REG_UPDATE_2(PHYASYMCLK_CLOCK_CNTL,
+					PHYASYMCLK_EN, 1,
+					PHYASYMCLK_SRC_SEL, clk_src);
+		} else {
+			REG_UPDATE_2(PHYASYMCLK_CLOCK_CNTL,
+					PHYASYMCLK_EN, 0,
+					PHYASYMCLK_SRC_SEL, 0);
+		}
+		break;
+	case 1:
+		if (force_enable) {
+			REG_UPDATE_2(PHYBSYMCLK_CLOCK_CNTL,
+					PHYBSYMCLK_EN, 1,
+					PHYBSYMCLK_SRC_SEL, clk_src);
+		} else {
+			REG_UPDATE_2(PHYBSYMCLK_CLOCK_CNTL,
+					PHYBSYMCLK_EN, 0,
+					PHYBSYMCLK_SRC_SEL, 0);
+		}
+		break;
+	case 2:
+		if (force_enable) {
+			REG_UPDATE_2(PHYCSYMCLK_CLOCK_CNTL,
+					PHYCSYMCLK_EN, 1,
+					PHYCSYMCLK_SRC_SEL, clk_src);
+		} else {
+			REG_UPDATE_2(PHYCSYMCLK_CLOCK_CNTL,
+					PHYCSYMCLK_EN, 0,
+					PHYCSYMCLK_SRC_SEL, 0);
+		}
+		break;
+	case 3:
+		if (force_enable) {
+			REG_UPDATE_2(PHYDSYMCLK_CLOCK_CNTL,
+					PHYDSYMCLK_EN, 1,
+					PHYDSYMCLK_SRC_SEL, clk_src);
+		} else {
+			REG_UPDATE_2(PHYDSYMCLK_CLOCK_CNTL,
+					PHYDSYMCLK_EN, 0,
+					PHYDSYMCLK_SRC_SEL, 0);
+		}
+		break;
+	case 4:
+		if (force_enable) {
+			REG_UPDATE_2(PHYESYMCLK_CLOCK_CNTL,
+					PHYESYMCLK_EN, 1,
+					PHYESYMCLK_SRC_SEL, clk_src);
+		} else {
+			REG_UPDATE_2(PHYESYMCLK_CLOCK_CNTL,
+					PHYESYMCLK_EN, 0,
+					PHYESYMCLK_SRC_SEL, 0);
+		}
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+	DC_LOG_DEBUG("%s: phy_inst(%d) PHYxSYMCLK_EN = %d, PHYxSYMCLK_SRC_SEL = %d\n",
+			__func__, phy_inst, force_enable ? 1 : 0, clk_src);
+}
+
+static void dccg35_set_valid_pixel_rate(
+		struct dccg *dccg,
+		int ref_dtbclk_khz,
+		int otg_inst,
+		int pixclk_khz)
+{
+	struct dtbclk_dto_params dto_params = {0};
+
+	dto_params.ref_dtbclk_khz = ref_dtbclk_khz;
+	dto_params.otg_inst = otg_inst;
+	dto_params.pixclk_khz = pixclk_khz;
+	dto_params.is_hdmi = true;
+
+	dccg35_set_dtbclk_dto(dccg, &dto_params);
+}
+
+static void dccg35_dpp_root_clock_control(
+		struct dccg *dccg,
+		unsigned int dpp_inst,
+		bool clock_on)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	if (dccg->dpp_clock_gated[dpp_inst] != clock_on)
+		return;
+
+	if (clock_on) {
+		dccg35_set_dppclk_rcg(dccg, dpp_inst, false);
+
+		/* turn off the DTO and leave phase/modulo at max */
+		dcn35_set_dppclk_enable(dccg, dpp_inst, 1);
+		REG_SET_2(DPPCLK_DTO_PARAM[dpp_inst], 0,
+			  DPPCLK0_DTO_PHASE, 0xFF,
+			  DPPCLK0_DTO_MODULO, 0xFF);
+	} else {
+		dcn35_set_dppclk_enable(dccg, dpp_inst, 0);
+		/* turn on the DTO to generate a 0hz clock */
+		REG_SET_2(DPPCLK_DTO_PARAM[dpp_inst], 0,
+			  DPPCLK0_DTO_PHASE, 0,
+			  DPPCLK0_DTO_MODULO, 1);
+		/*we have this in hwss: disable_plane*/
+		dccg35_set_dppclk_rcg(dccg, dpp_inst, true);
+	}
+
+	// wait for clock to fully ramp
+	udelay(10);
+
+	dccg->dpp_clock_gated[dpp_inst] = !clock_on;
+	DC_LOG_DEBUG("%s: dpp_inst(%d) clock_on = %d\n", __func__, dpp_inst, clock_on);
+}
+
+static void dccg35_disable_symclk32_se(
+		struct dccg *dccg,
+		int hpo_se_inst)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	/* set refclk as the source for symclk32_se */
+	switch (hpo_se_inst) {
+	case 0:
+		REG_UPDATE_2(SYMCLK32_SE_CNTL,
+				SYMCLK32_SE0_SRC_SEL, 0,
+				SYMCLK32_SE0_EN, 0);
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se) {
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL3,
+					SYMCLK32_SE0_GATE_DISABLE, 0);
+//			REG_UPDATE(DCCG_GATE_DISABLE_CNTL3,
+//					SYMCLK32_ROOT_SE0_GATE_DISABLE, 0);
+		}
+		break;
+	case 1:
+		REG_UPDATE_2(SYMCLK32_SE_CNTL,
+				SYMCLK32_SE1_SRC_SEL, 0,
+				SYMCLK32_SE1_EN, 0);
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se) {
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL3,
+					SYMCLK32_SE1_GATE_DISABLE, 0);
+//			REG_UPDATE(DCCG_GATE_DISABLE_CNTL3,
+//					SYMCLK32_ROOT_SE1_GATE_DISABLE, 0);
+		}
+		break;
+	case 2:
+		REG_UPDATE_2(SYMCLK32_SE_CNTL,
+				SYMCLK32_SE2_SRC_SEL, 0,
+				SYMCLK32_SE2_EN, 0);
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se) {
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL3,
+					SYMCLK32_SE2_GATE_DISABLE, 0);
+//			REG_UPDATE(DCCG_GATE_DISABLE_CNTL3,
+//					SYMCLK32_ROOT_SE2_GATE_DISABLE, 0);
+		}
+		break;
+	case 3:
+		REG_UPDATE_2(SYMCLK32_SE_CNTL,
+				SYMCLK32_SE3_SRC_SEL, 0,
+				SYMCLK32_SE3_EN, 0);
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se) {
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL3,
+					SYMCLK32_SE3_GATE_DISABLE, 0);
+//			REG_UPDATE(DCCG_GATE_DISABLE_CNTL3,
+//					SYMCLK32_ROOT_SE3_GATE_DISABLE, 0);
+		}
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+
+}
+
+static void dccg35_init_cb(struct dccg *dccg)
+{
+	(void)dccg;
+	/* Any RCG should be done when driver enter low power mode*/
+}
+void dccg35_init(struct dccg *dccg)
+{
+	int otg_inst;
+	/* Set HPO stream encoder to use refclk to avoid case where PHY is
+	 * disabled and SYMCLK32 for HPO SE is sourced from PHYD32CLK which
+	 * will cause DCN to hang.
+	 */
+	for (otg_inst = 0; otg_inst < 4; otg_inst++)
+		dccg35_disable_symclk32_se(dccg, otg_inst);
+
+	if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_le)
+		for (otg_inst = 0; otg_inst < 2; otg_inst++) {
+			dccg31_disable_symclk32_le(dccg, otg_inst);
+			dccg31_set_symclk32_le_root_clock_gating(dccg, otg_inst, false);
+			DC_LOG_DEBUG("%s: OTG%d SYMCLK32_LE disabled and root clock gating disabled\n",
+					__func__, otg_inst);
+		}
+
+//	if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se)
+//		for (otg_inst = 0; otg_inst < 4; otg_inst++)
+//			dccg35_disable_symclk_se(dccg, otg_inst, otg_inst);
+
+
+	if (dccg->ctx->dc->debug.root_clock_optimization.bits.dpstream)
+		for (otg_inst = 0; otg_inst < 4; otg_inst++) {
+			dccg35_set_dpstreamclk(dccg, REFCLK, otg_inst,
+						otg_inst);
+			dccg35_set_dpstreamclk_root_clock_gating(dccg, otg_inst, false);
+			DC_LOG_DEBUG("%s: OTG%d DPSTREAMCLK disabled and root clock gating disabled\n",
+					__func__, otg_inst);
+		}
+
+/*
+	dccg35_enable_global_fgcg_rep(
+		dccg, dccg->ctx->dc->debug.enable_fine_grain_clock_gating.bits
+			      .dccg_global_fgcg_rep);*/
+}
+
+void dccg35_enable_global_fgcg_rep(struct dccg *dccg, bool value)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	REG_UPDATE(DCCG_GLOBAL_FGCG_REP_CNTL, DCCG_GLOBAL_FGCG_REP_DIS, !value);
+}
+
+static void dccg35_enable_dscclk(struct dccg *dccg, int inst)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	//Disable DTO
+	switch (inst) {
+	case 0:
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DSCCLK0_ROOT_GATE_DISABLE, 1);
+
+		REG_UPDATE_2(DSCCLK0_DTO_PARAM,
+				DSCCLK0_DTO_PHASE, 0,
+				DSCCLK0_DTO_MODULO, 0);
+		REG_UPDATE(DSCCLK_DTO_CTRL,	DSCCLK0_EN, 1);
+		break;
+	case 1:
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DSCCLK1_ROOT_GATE_DISABLE, 1);
+
+		REG_UPDATE_2(DSCCLK1_DTO_PARAM,
+				DSCCLK1_DTO_PHASE, 0,
+				DSCCLK1_DTO_MODULO, 0);
+		REG_UPDATE(DSCCLK_DTO_CTRL, DSCCLK1_EN, 1);
+		break;
+	case 2:
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DSCCLK2_ROOT_GATE_DISABLE, 1);
+
+		REG_UPDATE_2(DSCCLK2_DTO_PARAM,
+				DSCCLK2_DTO_PHASE, 0,
+				DSCCLK2_DTO_MODULO, 0);
+		REG_UPDATE(DSCCLK_DTO_CTRL, DSCCLK2_EN, 1);
+		break;
+	case 3:
+		REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DSCCLK3_ROOT_GATE_DISABLE, 1);
+
+		REG_UPDATE_2(DSCCLK3_DTO_PARAM,
+				DSCCLK3_DTO_PHASE, 0,
+				DSCCLK3_DTO_MODULO, 0);
+		REG_UPDATE(DSCCLK_DTO_CTRL, DSCCLK3_EN, 1);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+
+	/* Wait for clock to ramp */
+	udelay(10);
+}
+
+static void dccg35_disable_dscclk(struct dccg *dccg,
+				int inst)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	switch (inst) {
+	case 0:
+		REG_UPDATE(DSCCLK_DTO_CTRL, DSCCLK0_EN, 0);
+		REG_UPDATE_2(DSCCLK0_DTO_PARAM,
+				DSCCLK0_DTO_PHASE, 0,
+				DSCCLK0_DTO_MODULO, 1);
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.dsc)
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DSCCLK0_ROOT_GATE_DISABLE, 0);
+		break;
+	case 1:
+		REG_UPDATE(DSCCLK_DTO_CTRL, DSCCLK1_EN, 0);
+		REG_UPDATE_2(DSCCLK1_DTO_PARAM,
+				DSCCLK1_DTO_PHASE, 0,
+				DSCCLK1_DTO_MODULO, 1);
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.dsc)
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DSCCLK1_ROOT_GATE_DISABLE, 0);
+		break;
+	case 2:
+		REG_UPDATE(DSCCLK_DTO_CTRL, DSCCLK2_EN, 0);
+		REG_UPDATE_2(DSCCLK2_DTO_PARAM,
+				DSCCLK2_DTO_PHASE, 0,
+				DSCCLK2_DTO_MODULO, 1);
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.dsc)
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DSCCLK2_ROOT_GATE_DISABLE, 0);
+		break;
+	case 3:
+		REG_UPDATE(DSCCLK_DTO_CTRL, DSCCLK3_EN, 0);
+		REG_UPDATE_2(DSCCLK3_DTO_PARAM,
+				DSCCLK3_DTO_PHASE, 0,
+				DSCCLK3_DTO_MODULO, 1);
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.dsc)
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DSCCLK3_ROOT_GATE_DISABLE, 0);
+		break;
+	default:
+		return;
+	}
+
+	/* Wait for clock ramp */
+	udelay(10);
+}
+
+static void dccg35_enable_symclk_se(struct dccg *dccg, uint32_t stream_enc_inst, uint32_t link_enc_inst)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	switch (link_enc_inst) {
+	case 0:
+		REG_UPDATE(SYMCLKA_CLOCK_ENABLE,
+				SYMCLKA_CLOCK_ENABLE, 1);
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se)
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, SYMCLKA_ROOT_GATE_DISABLE, 1);
+		break;
+	case 1:
+		REG_UPDATE(SYMCLKB_CLOCK_ENABLE,
+				SYMCLKB_CLOCK_ENABLE, 1);
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se)
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, SYMCLKB_ROOT_GATE_DISABLE, 1);
+		break;
+	case 2:
+		REG_UPDATE(SYMCLKC_CLOCK_ENABLE,
+				SYMCLKC_CLOCK_ENABLE, 1);
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se)
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, SYMCLKC_ROOT_GATE_DISABLE, 1);
+		break;
+	case 3:
+		REG_UPDATE(SYMCLKD_CLOCK_ENABLE,
+				SYMCLKD_CLOCK_ENABLE, 1);
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se)
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, SYMCLKD_ROOT_GATE_DISABLE, 1);
+		break;
+	case 4:
+		REG_UPDATE(SYMCLKE_CLOCK_ENABLE,
+				SYMCLKE_CLOCK_ENABLE, 1);
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se)
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, SYMCLKE_ROOT_GATE_DISABLE, 1);
+		break;
+	}
+
+	switch (stream_enc_inst) {
+	case 0:
+		REG_UPDATE_2(SYMCLKA_CLOCK_ENABLE,
+				SYMCLKA_FE_EN, 1,
+				SYMCLKA_FE_SRC_SEL, link_enc_inst);
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se)
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, SYMCLKA_FE_ROOT_GATE_DISABLE, 1);
+		break;
+	case 1:
+		REG_UPDATE_2(SYMCLKB_CLOCK_ENABLE,
+				SYMCLKB_FE_EN, 1,
+				SYMCLKB_FE_SRC_SEL, link_enc_inst);
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se)
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, SYMCLKB_FE_ROOT_GATE_DISABLE, 1);
+		break;
+	case 2:
+		REG_UPDATE_2(SYMCLKC_CLOCK_ENABLE,
+				SYMCLKC_FE_EN, 1,
+				SYMCLKC_FE_SRC_SEL, link_enc_inst);
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se)
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, SYMCLKC_FE_ROOT_GATE_DISABLE, 1);
+		break;
+	case 3:
+		REG_UPDATE_2(SYMCLKD_CLOCK_ENABLE,
+				SYMCLKD_FE_EN, 1,
+				SYMCLKD_FE_SRC_SEL, link_enc_inst);
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se)
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, SYMCLKD_FE_ROOT_GATE_DISABLE, 1);
+		break;
+	case 4:
+		REG_UPDATE_2(SYMCLKE_CLOCK_ENABLE,
+				SYMCLKE_FE_EN, 1,
+				SYMCLKE_FE_SRC_SEL, link_enc_inst);
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se)
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, SYMCLKE_FE_ROOT_GATE_DISABLE, 1);
+		break;
+	}
+}
+
+/*get other front end connected to this backend*/
+static uint8_t dccg35_get_number_enabled_symclk_fe_connected_to_be(struct dccg *dccg, uint32_t link_enc_inst)
+{
+	uint8_t num_enabled_symclk_fe = 0;
+	uint32_t fe_clk_en[5] = {0}, be_clk_sel[5] = {0};
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	REG_GET_2(SYMCLKA_CLOCK_ENABLE, SYMCLKA_FE_EN, &fe_clk_en[0],
+		SYMCLKA_FE_SRC_SEL, &be_clk_sel[0]);
+
+	REG_GET_2(SYMCLKB_CLOCK_ENABLE, SYMCLKB_FE_EN, &fe_clk_en[1],
+		SYMCLKB_FE_SRC_SEL, &be_clk_sel[1]);
+
+	REG_GET_2(SYMCLKC_CLOCK_ENABLE, SYMCLKC_FE_EN, &fe_clk_en[2],
+		SYMCLKC_FE_SRC_SEL, &be_clk_sel[2]);
+
+	REG_GET_2(SYMCLKD_CLOCK_ENABLE,	SYMCLKD_FE_EN, &fe_clk_en[3],
+		SYMCLKD_FE_SRC_SEL, &be_clk_sel[3]);
+
+	REG_GET_2(SYMCLKE_CLOCK_ENABLE,	SYMCLKE_FE_EN, &fe_clk_en[4],
+		SYMCLKE_FE_SRC_SEL, &be_clk_sel[4]);
+
+	uint8_t i;
+
+	for (i = 0; i < ARRAY_SIZE(fe_clk_en); i++) {
+		if (fe_clk_en[i] && be_clk_sel[i] == link_enc_inst)
+			num_enabled_symclk_fe++;
+	}
+	return num_enabled_symclk_fe;
+}
+
+static void dccg35_disable_symclk_se(struct dccg *dccg, uint32_t stream_enc_inst, uint32_t link_enc_inst)
+{
+	uint8_t num_enabled_symclk_fe = 0;
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	switch (stream_enc_inst) {
+	case 0:
+		REG_UPDATE_2(SYMCLKA_CLOCK_ENABLE,
+				SYMCLKA_FE_EN, 0,
+				SYMCLKA_FE_SRC_SEL, 0);
+//		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se)
+//			REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, SYMCLKA_FE_ROOT_GATE_DISABLE, 0);
+		break;
+	case 1:
+		REG_UPDATE_2(SYMCLKB_CLOCK_ENABLE,
+				SYMCLKB_FE_EN, 0,
+				SYMCLKB_FE_SRC_SEL, 0);
+//		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se)
+//			REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, SYMCLKB_FE_ROOT_GATE_DISABLE, 0);
+		break;
+	case 2:
+		REG_UPDATE_2(SYMCLKC_CLOCK_ENABLE,
+				SYMCLKC_FE_EN, 0,
+				SYMCLKC_FE_SRC_SEL, 0);
+//		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se)
+//			REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, SYMCLKC_FE_ROOT_GATE_DISABLE, 0);
+		break;
+	case 3:
+		REG_UPDATE_2(SYMCLKD_CLOCK_ENABLE,
+				SYMCLKD_FE_EN, 0,
+				SYMCLKD_FE_SRC_SEL, 0);
+//		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se)
+//			REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, SYMCLKD_FE_ROOT_GATE_DISABLE, 0);
+		break;
+	case 4:
+		REG_UPDATE_2(SYMCLKE_CLOCK_ENABLE,
+				SYMCLKE_FE_EN, 0,
+				SYMCLKE_FE_SRC_SEL, 0);
+//		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se)
+//			REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, SYMCLKE_FE_ROOT_GATE_DISABLE, 0);
+		break;
+	}
+
+	/*check other enabled symclk fe connected to this be */
+	num_enabled_symclk_fe = dccg35_get_number_enabled_symclk_fe_connected_to_be(dccg, link_enc_inst);
+	/*only turn off backend clk if other front end attached to this backend are all off,
+	 for mst, only turn off the backend if this is the last front end*/
+	if (num_enabled_symclk_fe == 0) {
+		switch (link_enc_inst) {
+		case 0:
+			REG_UPDATE(SYMCLKA_CLOCK_ENABLE,
+					SYMCLKA_CLOCK_ENABLE, 0);
+//			if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_le)
+//				REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, SYMCLKA_ROOT_GATE_DISABLE, 0);
+			break;
+		case 1:
+			REG_UPDATE(SYMCLKB_CLOCK_ENABLE,
+					SYMCLKB_CLOCK_ENABLE, 0);
+//			if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_le)
+//				REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, SYMCLKB_ROOT_GATE_DISABLE, 0);
+			break;
+		case 2:
+			REG_UPDATE(SYMCLKC_CLOCK_ENABLE,
+					SYMCLKC_CLOCK_ENABLE, 0);
+//			if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_le)
+//				REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, SYMCLKC_ROOT_GATE_DISABLE, 0);
+			break;
+		case 3:
+			REG_UPDATE(SYMCLKD_CLOCK_ENABLE,
+					SYMCLKD_CLOCK_ENABLE, 0);
+//			if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_le)
+//				REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, SYMCLKD_ROOT_GATE_DISABLE, 0);
+			break;
+		case 4:
+			REG_UPDATE(SYMCLKE_CLOCK_ENABLE,
+					SYMCLKE_CLOCK_ENABLE, 0);
+//			if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_le)
+//				REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, SYMCLKE_ROOT_GATE_DISABLE, 0);
+			break;
+		}
+	}
+}
+
+static void dccg35_set_dpstreamclk_cb(
+		struct dccg *dccg,
+		enum streamclk_source src,
+		int otg_inst,
+		int dp_hpo_inst)
+{
+
+	enum dtbclk_source dtb_clk_src;
+	enum dp_stream_clk_source dp_stream_clk_src;
+
+	switch (src) {
+	case REFCLK:
+		dtb_clk_src = DTBCLK_REFCLK;
+		dp_stream_clk_src = DP_STREAM_REFCLK;
+		break;
+	case DPREFCLK:
+		dtb_clk_src = DTBCLK_DPREFCLK;
+		dp_stream_clk_src = (enum dp_stream_clk_source)otg_inst;
+		break;
+	case DTBCLK0:
+		dtb_clk_src = DTBCLK_DTBCLK0;
+		dp_stream_clk_src = (enum dp_stream_clk_source)otg_inst;
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+
+	if (dtb_clk_src == DTBCLK_REFCLK &&
+		dp_stream_clk_src == DP_STREAM_REFCLK) {
+		dccg35_disable_dtbclk_p_new(dccg, otg_inst);
+		dccg35_disable_dpstreamclk_new(dccg, dp_hpo_inst);
+	} else {
+		dccg35_enable_dtbclk_p_new(dccg, dtb_clk_src, otg_inst);
+		dccg35_enable_dpstreamclk_new(dccg,
+										dp_stream_clk_src,
+										dp_hpo_inst);
+	}
+}
+
+static void dccg35_set_dpstreamclk_root_clock_gating_cb(
+	struct dccg *dccg,
+	int dp_hpo_inst,
+	bool power_on)
+{
+	/* power_on set indicates we need to ungate
+	 * Currently called from optimize_bandwidth and prepare_bandwidth calls
+	 * Since clock source is not passed restore to refclock on ungate
+	 * Instance 0 is implied here since only one streamclock resource
+	 * Redundant as gating when enabled is acheived through set_dpstreamclk
+	 */
+	if (power_on)
+		dccg35_enable_dpstreamclk_new(dccg,
+										DP_STREAM_REFCLK,
+										dp_hpo_inst);
+	else
+		dccg35_disable_dpstreamclk_new(dccg, dp_hpo_inst);
+}
+
+static void dccg35_update_dpp_dto_cb(struct dccg *dccg, int dpp_inst,
+				  int req_dppclk)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	if (dccg->dpp_clock_gated[dpp_inst]) {
+		/*
+		 * Do not update the DPPCLK DTO if the clock is stopped.
+		 */
+		return;
+	}
+
+	if (dccg->ref_dppclk && req_dppclk) {
+		int ref_dppclk = dccg->ref_dppclk;
+		int modulo, phase;
+
+		// phase / modulo = dpp pipe clk / dpp global clk
+		modulo = 0xff;   // use FF at the end
+		phase = ((modulo * req_dppclk) + ref_dppclk - 1) / ref_dppclk;
+
+		if (phase > 0xff) {
+			ASSERT(false);
+			phase = 0xff;
+		}
+
+		/* Enable DPP CLK DTO output */
+		dccg35_enable_dpp_clk_new(dccg, dpp_inst, DPP_DCCG_DTO);
+
+		/* Program DTO */
+		REG_SET_2(DPPCLK_DTO_PARAM[dpp_inst], 0,
+				DPPCLK0_DTO_PHASE, phase,
+				DPPCLK0_DTO_MODULO, modulo);
+	} else
+		dccg35_disable_dpp_clk_new(dccg, dpp_inst);
+
+	dccg->pipe_dppclk_khz[dpp_inst] = req_dppclk;
+}
+
+static void dccg35_dpp_root_clock_control_cb(
+	struct dccg *dccg,
+	unsigned int dpp_inst,
+	bool power_on)
+{
+	if (dccg->dpp_clock_gated[dpp_inst] == power_on)
+		return;
+	/* power_on set indicates we need to ungate
+	 * Currently called from optimize_bandwidth and prepare_bandwidth calls
+	 * Since clock source is not passed restore to refclock on ungate
+	 * Redundant as gating when enabled is acheived through update_dpp_dto
+	 */
+	dccg35_set_dppclk_rcg(dccg, dpp_inst, !power_on);
+
+	dccg->dpp_clock_gated[dpp_inst] = !power_on;
+}
+
+static void dccg35_enable_symclk32_se_cb(
+	struct dccg *dccg,
+	int inst,
+	enum phyd32clk_clock_source phyd32clk)
+{
+	dccg35_enable_symclk32_se_new(dccg, inst, (enum symclk32_se_clk_source)phyd32clk);
+}
+
+static void dccg35_disable_symclk32_se_cb(struct dccg *dccg, int inst)
+{
+	dccg35_disable_symclk32_se_new(dccg, inst);
+}
+
+static void dccg35_enable_symclk32_le_cb(
+			struct dccg *dccg,
+			int inst,
+			enum phyd32clk_clock_source src)
+{
+	dccg35_enable_symclk32_le_new(dccg, inst, (enum symclk32_le_clk_source) src);
+}
+
+static void dccg35_disable_symclk32_le_cb(struct dccg *dccg, int inst)
+{
+	dccg35_disable_symclk32_le_new(dccg, inst);
+}
+
+static void dccg35_set_symclk32_le_root_clock_gating_cb(
+	struct dccg *dccg,
+	int inst,
+	bool power_on)
+{
+	/* power_on set indicates we need to ungate
+	 * Currently called from optimize_bandwidth and prepare_bandwidth calls
+	 * Since clock source is not passed restore to refclock on ungate
+	 * Redundant as gating when enabled is acheived through disable_symclk32_le
+	 */
+	if (power_on)
+		dccg35_enable_symclk32_le_new(dccg, inst, SYMCLK32_LE_REFCLK);
+	else
+		dccg35_disable_symclk32_le_new(dccg, inst);
+}
+
+static void dccg35_set_physymclk_cb(
+	struct dccg *dccg,
+	int inst,
+	enum physymclk_clock_source clk_src,
+	bool force_enable)
+{
+	/* force_enable = 0 indicates we can switch to ref clock */
+	if (force_enable)
+		dccg35_enable_physymclk_new(dccg, inst, (enum physymclk_source)clk_src);
+	else
+		dccg35_disable_physymclk_new(dccg, inst);
+}
+
+static void dccg35_set_physymclk_root_clock_gating_cb(
+	struct dccg *dccg,
+	int inst,
+	bool power_on)
+{
+	/* Redundant RCG already done in disable_physymclk
+	 * power_on = 1 indicates we need to ungate
+	 */
+	if (power_on)
+		dccg35_enable_physymclk_new(dccg, inst, PHYSYMCLK_REFCLK);
+	else
+		dccg35_disable_physymclk_new(dccg, inst);
+}
+
+static void dccg35_set_symclk32_le_root_clock_gating(
+	struct dccg *dccg,
+	int inst,
+	bool power_on)
+{
+	/* power_on set indicates we need to ungate
+	 * Currently called from optimize_bandwidth and prepare_bandwidth calls
+	 * Since clock source is not passed restore to refclock on ungate
+	 * Redundant as gating when enabled is acheived through disable_symclk32_le
+	 */
+	if (power_on)
+		dccg35_enable_symclk32_le_new(dccg, inst, SYMCLK32_LE_REFCLK);
+	else
+		dccg35_disable_symclk32_le_new(dccg, inst);
+}
+
+static void dccg35_set_dtbclk_p_src_cb(
+		struct dccg *dccg,
+		enum streamclk_source src,
+		uint32_t inst)
+{
+	if (src == DTBCLK0)
+		dccg35_enable_dtbclk_p_new(dccg, DTBCLK_DTBCLK0, inst);
+	else
+		dccg35_disable_dtbclk_p_new(dccg, inst);
+}
+
+static void dccg35_set_dtbclk_dto_cb(
+		struct dccg *dccg,
+		const struct dtbclk_dto_params *params)
+{
+	/* set_dtbclk_p_src typ called earlier to switch to DTBCLK
+	 * if params->ref_dtbclk_khz and req_dtbclk_khz are 0 switch to ref-clock
+	 */
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+	/* DTO Output Rate / Pixel Rate = 1/4 */
+	int req_dtbclk_khz = params->pixclk_khz / 4;
+
+	if (params->ref_dtbclk_khz && req_dtbclk_khz) {
+		uint32_t modulo, phase;
+
+		dccg35_enable_dtbclk_p_new(dccg, DTBCLK_DTBCLK0, params->otg_inst);
+
+		// phase / modulo = dtbclk / dtbclk ref
+		modulo = params->ref_dtbclk_khz * 1000;
+		phase = req_dtbclk_khz * 1000;
+
+		REG_WRITE(DTBCLK_DTO_MODULO[params->otg_inst], modulo);
+		REG_WRITE(DTBCLK_DTO_PHASE[params->otg_inst], phase);
+
+		REG_UPDATE(OTG_PIXEL_RATE_CNTL[params->otg_inst],
+				DTBCLK_DTO_ENABLE[params->otg_inst], 1);
+
+		REG_WAIT(OTG_PIXEL_RATE_CNTL[params->otg_inst],
+				DTBCLKDTO_ENABLE_STATUS[params->otg_inst], 1,
+				1, 100);
+
+		/* program OTG_PIXEL_RATE_DIV for DIVK1 and DIVK2 fields */
+		dccg35_set_pixel_rate_div(dccg, params->otg_inst, PIXEL_RATE_DIV_BY_1, PIXEL_RATE_DIV_BY_1);
+
+		/* The recommended programming sequence to enable DTBCLK DTO to generate
+		 * valid pixel HPO DPSTREAM ENCODER, specifies that DTO source select should
+		 * be set only after DTO is enabled
+		 */
+		REG_UPDATE(OTG_PIXEL_RATE_CNTL[params->otg_inst],
+				PIPE_DTO_SRC_SEL[params->otg_inst], 2);
+	} else {
+		dccg35_disable_dtbclk_p_new(dccg, params->otg_inst);
+
+		REG_UPDATE_2(OTG_PIXEL_RATE_CNTL[params->otg_inst],
+					 DTBCLK_DTO_ENABLE[params->otg_inst], 0,
+					 PIPE_DTO_SRC_SEL[params->otg_inst], params->is_hdmi ? 0 : 1);
+
+		REG_WRITE(DTBCLK_DTO_MODULO[params->otg_inst], 0);
+		REG_WRITE(DTBCLK_DTO_PHASE[params->otg_inst], 0);
+	}
+}
+
+static void dccg35_disable_dscclk_cb(struct dccg *dccg,
+									 int inst)
+{
+	dccg35_disable_dscclk_new(dccg, inst);
+}
+
+static void dccg35_enable_dscclk_cb(struct dccg *dccg, int inst)
+{
+	dccg35_enable_dscclk_new(dccg, inst, DSC_DTO_TUNED_CK_GPU_DISCLK_3);
+}
+
+static void dccg35_enable_symclk_se_cb(struct dccg *dccg, uint32_t stream_enc_inst, uint32_t link_enc_inst)
+{
+	/* Switch to functional clock if already not selected */
+	dccg35_enable_symclk_be_new(dccg, SYMCLK_BE_PHYCLK, link_enc_inst);
+
+	dccg35_enable_symclk_fe_new(dccg, stream_enc_inst, (enum symclk_fe_source) link_enc_inst);
+
+}
+
+static void dccg35_disable_symclk_se_cb(
+			struct dccg *dccg,
+			uint32_t stream_enc_inst,
+			uint32_t link_enc_inst)
+{
+	dccg35_disable_symclk_fe_new(dccg, stream_enc_inst);
+
+	/* DMU PHY sequence switches SYMCLK_BE (link_enc_inst) to ref clock once PHY is turned off */
+}
+
+void dccg35_root_gate_disable_control(struct dccg *dccg, uint32_t pipe_idx, uint32_t disable_clock_gating)
+{
+	dccg35_set_dppclk_root_clock_gating(dccg, pipe_idx, disable_clock_gating);
+}
+
+static const struct dccg_funcs dccg35_funcs_new = {
+	.update_dpp_dto = dccg35_update_dpp_dto_cb,
+	.dpp_root_clock_control = dccg35_dpp_root_clock_control_cb,
+	.get_dccg_ref_freq = dccg31_get_dccg_ref_freq,
+	.dccg_init = dccg35_init_cb,
+	.set_dpstreamclk = dccg35_set_dpstreamclk_cb,
+	.set_dpstreamclk_root_clock_gating = dccg35_set_dpstreamclk_root_clock_gating_cb,
+	.enable_symclk32_se = dccg35_enable_symclk32_se_cb,
+	.disable_symclk32_se = dccg35_disable_symclk32_se_cb,
+	.enable_symclk32_le = dccg35_enable_symclk32_le_cb,
+	.disable_symclk32_le = dccg35_disable_symclk32_le_cb,
+	.set_symclk32_le_root_clock_gating = dccg35_set_symclk32_le_root_clock_gating_cb,
+	.set_physymclk = dccg35_set_physymclk_cb,
+	.set_physymclk_root_clock_gating = dccg35_set_physymclk_root_clock_gating_cb,
+	.set_dtbclk_dto = dccg35_set_dtbclk_dto_cb,
+	.set_audio_dtbclk_dto = dccg31_set_audio_dtbclk_dto,
+	.set_fifo_errdet_ovr_en = dccg2_set_fifo_errdet_ovr_en,
+	.otg_add_pixel = dccg31_otg_add_pixel,
+	.otg_drop_pixel = dccg31_otg_drop_pixel,
+	.set_dispclk_change_mode = dccg31_set_dispclk_change_mode,
+	.disable_dsc = dccg35_disable_dscclk_cb,
+	.enable_dsc = dccg35_enable_dscclk_cb,
+	.set_pixel_rate_div = dccg35_set_pixel_rate_div,
+	.get_pixel_rate_div = dccg35_get_pixel_rate_div,
+	.trigger_dio_fifo_resync = dccg35_trigger_dio_fifo_resync,
+	.set_valid_pixel_rate = dccg35_set_valid_pixel_rate,
+	.enable_symclk_se = dccg35_enable_symclk_se_cb,
+	.disable_symclk_se = dccg35_disable_symclk_se_cb,
+	.set_dtbclk_p_src = dccg35_set_dtbclk_p_src_cb,
+};
+
+static const struct dccg_funcs dccg35_funcs = {
+	.update_dpp_dto = dccg35_update_dpp_dto,
+	.dpp_root_clock_control = dccg35_dpp_root_clock_control,
+	.get_dccg_ref_freq = dccg31_get_dccg_ref_freq,
+	.dccg_init = dccg35_init,
+	.set_dpstreamclk = dccg35_set_dpstreamclk,
+	.set_dpstreamclk_root_clock_gating = dccg35_set_dpstreamclk_root_clock_gating,
+	.enable_symclk32_se = dccg31_enable_symclk32_se,
+	.disable_symclk32_se = dccg35_disable_symclk32_se,
+	.enable_symclk32_le = dccg31_enable_symclk32_le,
+	.disable_symclk32_le = dccg31_disable_symclk32_le,
+	.set_symclk32_le_root_clock_gating = dccg31_set_symclk32_le_root_clock_gating,
+	.set_physymclk = dccg35_set_physymclk,
+	.set_physymclk_root_clock_gating = dccg35_set_physymclk_root_clock_gating,
+	.set_dtbclk_dto = dccg35_set_dtbclk_dto,
+	.set_audio_dtbclk_dto = dccg31_set_audio_dtbclk_dto,
+	.set_fifo_errdet_ovr_en = dccg2_set_fifo_errdet_ovr_en,
+	.otg_add_pixel = dccg31_otg_add_pixel,
+	.otg_drop_pixel = dccg31_otg_drop_pixel,
+	.set_dispclk_change_mode = dccg31_set_dispclk_change_mode,
+	.disable_dsc = dccg35_disable_dscclk,
+	.enable_dsc = dccg35_enable_dscclk,
+	.set_pixel_rate_div = dccg35_set_pixel_rate_div,
+	.get_pixel_rate_div = dccg35_get_pixel_rate_div,
+	.trigger_dio_fifo_resync = dccg35_trigger_dio_fifo_resync,
+	.set_valid_pixel_rate = dccg35_set_valid_pixel_rate,
+	.enable_symclk_se = dccg35_enable_symclk_se,
+	.disable_symclk_se = dccg35_disable_symclk_se,
+	.set_dtbclk_p_src = dccg35_set_dtbclk_p_src,
+	.dccg_root_gate_disable_control = dccg35_root_gate_disable_control,
+	.dccg_read_reg_state = dccg31_read_reg_state,
+};
+
+struct dccg *dccg35_create(
+	struct dc_context *ctx,
+	const struct dccg_registers *regs,
+	const struct dccg_shift *dccg_shift,
+	const struct dccg_mask *dccg_mask)
+{
+	struct dcn_dccg *dccg_dcn = kzalloc(sizeof(*dccg_dcn), GFP_KERNEL);
+	struct dccg *base;
+
+	if (dccg_dcn == NULL) {
+		BREAK_TO_DEBUGGER();
+		return NULL;
+	}
+	(void)&dccg35_disable_symclk_be_new;
+	(void)&dccg35_set_symclk32_le_root_clock_gating;
+	(void)&dccg35_set_smclk32_se_rcg;
+	(void)&dccg35_funcs_new;
+
+	base = &dccg_dcn->base;
+	base->ctx = ctx;
+	base->funcs = &dccg35_funcs;
+
+	dccg_dcn->regs = regs;
+	dccg_dcn->dccg_shift = dccg_shift;
+	dccg_dcn->dccg_mask = dccg_mask;
+
+	return &dccg_dcn->base;
+}
diff --git a/drivers/gpu/drm/amd/display/dc/dccg/dcn35/dcn35_dccg.h b/drivers/gpu/drm/amd/display/dc/dccg/dcn35/dcn35_dccg.h
new file mode 100644
index 000000000000..7b9c36456cd9
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dccg/dcn35/dcn35_dccg.h
@@ -0,0 +1,256 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __DCN35_DCCG_H__
+#define __DCN35_DCCG_H__
+
+#include "dcn314/dcn314_dccg.h"
+
+#define DCCG_SFII(block, reg_name, field_prefix, field_name, inst, post_fix)\
+	.field_prefix ## _ ## field_name[inst] = block ## inst ## _ ## reg_name ## __ ## field_prefix ## inst ## _ ## field_name ## post_fix
+
+
+#define DCCG_REG_LIST_DCN35() \
+	DCCG_REG_LIST_DCN314(),\
+	SR(DPPCLK_CTRL),\
+	SR(DCCG_GATE_DISABLE_CNTL4),\
+	SR(DCCG_GATE_DISABLE_CNTL5),\
+	SR(DCCG_GATE_DISABLE_CNTL6),\
+	SR(DCCG_GLOBAL_FGCG_REP_CNTL),\
+	SR(SYMCLKA_CLOCK_ENABLE),\
+	SR(SYMCLKB_CLOCK_ENABLE),\
+	SR(SYMCLKC_CLOCK_ENABLE),\
+	SR(SYMCLKD_CLOCK_ENABLE), \
+	SR(SYMCLKE_CLOCK_ENABLE),\
+	SR(SYMCLK_PSP_CNTL)
+
+#define DCCG_MASK_SH_LIST_DCN35(mask_sh) \
+	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_DB_EN, DPPCLK, 0, mask_sh),\
+	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_DB_EN, DPPCLK, 1, mask_sh),\
+	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_DB_EN, DPPCLK, 2, mask_sh),\
+	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_DB_EN, DPPCLK, 3, mask_sh),\
+	DCCG_SF(DPPCLK_CTRL, DPPCLK0_EN, mask_sh),\
+	DCCG_SF(DPPCLK_CTRL, DPPCLK1_EN, mask_sh),\
+	DCCG_SF(DPPCLK_CTRL, DPPCLK2_EN, mask_sh),\
+	DCCG_SF(DPPCLK_CTRL, DPPCLK3_EN, mask_sh),\
+	DCCG_SF(DPPCLK0_DTO_PARAM, DPPCLK0_DTO_PHASE, mask_sh),\
+	DCCG_SF(DPPCLK0_DTO_PARAM, DPPCLK0_DTO_MODULO, mask_sh),\
+	DCCG_SF(HDMICHARCLK0_CLOCK_CNTL, HDMICHARCLK0_EN, mask_sh),\
+	DCCG_SF(HDMICHARCLK0_CLOCK_CNTL, HDMICHARCLK0_SRC_SEL, mask_sh),\
+	DCCG_SF(PHYASYMCLK_CLOCK_CNTL, PHYASYMCLK_EN, mask_sh),\
+	DCCG_SF(PHYASYMCLK_CLOCK_CNTL, PHYASYMCLK_SRC_SEL, mask_sh),\
+	DCCG_SF(PHYBSYMCLK_CLOCK_CNTL, PHYBSYMCLK_EN, mask_sh),\
+	DCCG_SF(PHYBSYMCLK_CLOCK_CNTL, PHYBSYMCLK_SRC_SEL, mask_sh),\
+	DCCG_SF(PHYCSYMCLK_CLOCK_CNTL, PHYCSYMCLK_EN, mask_sh),\
+	DCCG_SF(PHYCSYMCLK_CLOCK_CNTL, PHYCSYMCLK_SRC_SEL, mask_sh),\
+	DCCG_SF(PHYDSYMCLK_CLOCK_CNTL, PHYDSYMCLK_EN, mask_sh),\
+	DCCG_SF(PHYDSYMCLK_CLOCK_CNTL, PHYDSYMCLK_SRC_SEL, mask_sh),\
+	DCCG_SF(DPSTREAMCLK_CNTL, DPSTREAMCLK0_EN, mask_sh),\
+	DCCG_SF(DPSTREAMCLK_CNTL, DPSTREAMCLK1_EN, mask_sh),\
+	DCCG_SF(DPSTREAMCLK_CNTL, DPSTREAMCLK2_EN, mask_sh),\
+	DCCG_SF(DPSTREAMCLK_CNTL, DPSTREAMCLK3_EN, mask_sh),\
+	DCCG_SF(DPSTREAMCLK_CNTL, DPSTREAMCLK0_SRC_SEL, mask_sh),\
+	DCCG_SF(DPSTREAMCLK_CNTL, DPSTREAMCLK1_SRC_SEL, mask_sh),\
+	DCCG_SF(DPSTREAMCLK_CNTL, DPSTREAMCLK2_SRC_SEL, mask_sh),\
+	DCCG_SF(DPSTREAMCLK_CNTL, DPSTREAMCLK3_SRC_SEL, mask_sh),\
+	DCCG_SF(HDMISTREAMCLK_CNTL, HDMISTREAMCLK0_EN, mask_sh),\
+	DCCG_SF(HDMISTREAMCLK_CNTL, HDMISTREAMCLK0_SRC_SEL, mask_sh),\
+	DCCG_SF(DSCCLK_DTO_CTRL, DSCCLK0_EN, mask_sh),\
+	DCCG_SF(DSCCLK_DTO_CTRL, DSCCLK1_EN, mask_sh),\
+	DCCG_SF(DSCCLK_DTO_CTRL, DSCCLK2_EN, mask_sh),\
+	DCCG_SF(DSCCLK_DTO_CTRL, DSCCLK3_EN, mask_sh),\
+	DCCG_SF(DSCCLK0_DTO_PARAM, DSCCLK0_DTO_PHASE, mask_sh),\
+	DCCG_SF(DSCCLK0_DTO_PARAM, DSCCLK0_DTO_MODULO, mask_sh),\
+	DCCG_SF(DSCCLK1_DTO_PARAM, DSCCLK1_DTO_PHASE, mask_sh),\
+	DCCG_SF(DSCCLK1_DTO_PARAM, DSCCLK1_DTO_MODULO, mask_sh),\
+	DCCG_SF(DSCCLK2_DTO_PARAM, DSCCLK2_DTO_PHASE, mask_sh),\
+	DCCG_SF(DSCCLK2_DTO_PARAM, DSCCLK2_DTO_MODULO, mask_sh),\
+	DCCG_SF(DSCCLK3_DTO_PARAM, DSCCLK3_DTO_PHASE, mask_sh),\
+	DCCG_SF(DSCCLK3_DTO_PARAM, DSCCLK3_DTO_MODULO, mask_sh),\
+	DCCG_SF(SYMCLK32_SE_CNTL, SYMCLK32_SE0_SRC_SEL, mask_sh),\
+	DCCG_SF(SYMCLK32_SE_CNTL, SYMCLK32_SE1_SRC_SEL, mask_sh),\
+	DCCG_SF(SYMCLK32_SE_CNTL, SYMCLK32_SE2_SRC_SEL, mask_sh),\
+	DCCG_SF(SYMCLK32_SE_CNTL, SYMCLK32_SE3_SRC_SEL, mask_sh),\
+	DCCG_SF(SYMCLK32_SE_CNTL, SYMCLK32_SE0_EN, mask_sh),\
+	DCCG_SF(SYMCLK32_SE_CNTL, SYMCLK32_SE1_EN, mask_sh),\
+	DCCG_SF(SYMCLK32_SE_CNTL, SYMCLK32_SE2_EN, mask_sh),\
+	DCCG_SF(SYMCLK32_SE_CNTL, SYMCLK32_SE3_EN, mask_sh),\
+	DCCG_SF(SYMCLK32_LE_CNTL, SYMCLK32_LE0_SRC_SEL, mask_sh),\
+	DCCG_SF(SYMCLK32_LE_CNTL, SYMCLK32_LE1_SRC_SEL, mask_sh),\
+	DCCG_SF(SYMCLK32_LE_CNTL, SYMCLK32_LE0_EN, mask_sh),\
+	DCCG_SF(SYMCLK32_LE_CNTL, SYMCLK32_LE1_EN, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, DTBCLK_DTO, ENABLE, 0, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, DTBCLK_DTO, ENABLE, 1, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, DTBCLK_DTO, ENABLE, 2, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, DTBCLK_DTO, ENABLE, 3, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, DTBCLKDTO, ENABLE_STATUS, 0, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, DTBCLKDTO, ENABLE_STATUS, 1, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, DTBCLKDTO, ENABLE_STATUS, 2, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, DTBCLKDTO, ENABLE_STATUS, 3, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, PIPE, DTO_SRC_SEL, 0, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, PIPE, DTO_SRC_SEL, 1, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, PIPE, DTO_SRC_SEL, 2, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, PIPE, DTO_SRC_SEL, 3, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, ADD_PIXEL, 0, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, ADD_PIXEL, 1, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, ADD_PIXEL, 2, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, ADD_PIXEL, 3, mask_sh),\
+	DCCG_SF(OTG_PIXEL_RATE_DIV, OTG0_PIXEL_RATE_DIVK1, mask_sh),\
+	DCCG_SF(OTG_PIXEL_RATE_DIV, OTG0_PIXEL_RATE_DIVK2, mask_sh),\
+	DCCG_SF(OTG_PIXEL_RATE_DIV, OTG1_PIXEL_RATE_DIVK1, mask_sh),\
+	DCCG_SF(OTG_PIXEL_RATE_DIV, OTG1_PIXEL_RATE_DIVK2, mask_sh),\
+	DCCG_SF(OTG_PIXEL_RATE_DIV, OTG2_PIXEL_RATE_DIVK1, mask_sh),\
+	DCCG_SF(OTG_PIXEL_RATE_DIV, OTG2_PIXEL_RATE_DIVK2, mask_sh),\
+	DCCG_SF(OTG_PIXEL_RATE_DIV, OTG3_PIXEL_RATE_DIVK1, mask_sh),\
+	DCCG_SF(OTG_PIXEL_RATE_DIV, OTG3_PIXEL_RATE_DIVK2, mask_sh),\
+	DCCG_SF(OTG_PIXEL_RATE_DIV, OTG3_PIXEL_RATE_DIVK2, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, DROP_PIXEL, 0, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, DROP_PIXEL, 1, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, DROP_PIXEL, 2, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, DROP_PIXEL, 3, mask_sh),\
+	DCCG_SF(DTBCLK_P_CNTL, DTBCLK_P0_SRC_SEL, mask_sh),\
+	DCCG_SF(DTBCLK_P_CNTL, DTBCLK_P0_EN, mask_sh),\
+	DCCG_SF(DTBCLK_P_CNTL, DTBCLK_P1_SRC_SEL, mask_sh),\
+	DCCG_SF(DTBCLK_P_CNTL, DTBCLK_P1_EN, mask_sh),\
+	DCCG_SF(DTBCLK_P_CNTL, DTBCLK_P2_SRC_SEL, mask_sh),\
+	DCCG_SF(DTBCLK_P_CNTL, DTBCLK_P2_EN, mask_sh),\
+	DCCG_SF(DTBCLK_P_CNTL, DTBCLK_P3_SRC_SEL, mask_sh),\
+	DCCG_SF(DTBCLK_P_CNTL, DTBCLK_P3_EN, mask_sh),\
+	DCCG_SF(PHYASYMCLK_CLOCK_CNTL, PHYASYMCLK_EN, mask_sh),\
+	DCCG_SF(PHYASYMCLK_CLOCK_CNTL, PHYASYMCLK_SRC_SEL, mask_sh),\
+	DCCG_SF(PHYBSYMCLK_CLOCK_CNTL, PHYBSYMCLK_EN, mask_sh),\
+	DCCG_SF(PHYBSYMCLK_CLOCK_CNTL, PHYBSYMCLK_SRC_SEL, mask_sh),\
+	DCCG_SF(PHYCSYMCLK_CLOCK_CNTL, PHYCSYMCLK_EN, mask_sh),\
+	DCCG_SF(PHYCSYMCLK_CLOCK_CNTL, PHYCSYMCLK_SRC_SEL, mask_sh),\
+	DCCG_SF(PHYDSYMCLK_CLOCK_CNTL, PHYDSYMCLK_EN, mask_sh),\
+	DCCG_SF(PHYDSYMCLK_CLOCK_CNTL, PHYDSYMCLK_SRC_SEL, mask_sh),\
+	DCCG_SF(PHYESYMCLK_CLOCK_CNTL, PHYESYMCLK_EN, mask_sh),\
+	DCCG_SF(PHYESYMCLK_CLOCK_CNTL, PHYESYMCLK_SRC_SEL, mask_sh),\
+	DCCG_SF(DCCG_AUDIO_DTO_SOURCE, DCCG_AUDIO_DTO_SEL, mask_sh),\
+	DCCG_SF(DCCG_AUDIO_DTO_SOURCE, DCCG_AUDIO_DTO0_SOURCE_SEL, mask_sh),\
+	DCCG_SF(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_CHG_DONE, mask_sh),\
+	DCCG_SF(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_RDIVIDER, mask_sh),\
+	DCCG_SF(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_WDIVIDER, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL2, PHYASYMCLK_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL2, PHYBSYMCLK_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL2, PHYCSYMCLK_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL2, PHYDSYMCLK_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL2, PHYESYMCLK_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GLOBAL_FGCG_REP_CNTL, DCCG_GLOBAL_FGCG_REP_DIS, mask_sh),\
+	DCCG_SF(SYMCLKA_CLOCK_ENABLE, SYMCLKA_CLOCK_ENABLE, mask_sh),\
+	DCCG_SF(SYMCLKB_CLOCK_ENABLE, SYMCLKB_CLOCK_ENABLE, mask_sh),\
+	DCCG_SF(SYMCLKC_CLOCK_ENABLE, SYMCLKC_CLOCK_ENABLE, mask_sh),\
+	DCCG_SF(SYMCLKD_CLOCK_ENABLE, SYMCLKD_CLOCK_ENABLE, mask_sh),\
+	DCCG_SF(SYMCLKE_CLOCK_ENABLE, SYMCLKE_CLOCK_ENABLE, mask_sh),\
+	DCCG_SF(SYMCLKA_CLOCK_ENABLE, SYMCLKA_FE_EN, mask_sh),\
+	DCCG_SF(SYMCLKB_CLOCK_ENABLE, SYMCLKB_FE_EN, mask_sh),\
+	DCCG_SF(SYMCLKC_CLOCK_ENABLE, SYMCLKC_FE_EN, mask_sh),\
+	DCCG_SF(SYMCLKD_CLOCK_ENABLE, SYMCLKD_FE_EN, mask_sh),\
+	DCCG_SF(SYMCLKE_CLOCK_ENABLE, SYMCLKE_FE_EN, mask_sh),\
+	DCCG_SF(SYMCLKA_CLOCK_ENABLE, SYMCLKA_SRC_SEL, mask_sh),\
+	DCCG_SF(SYMCLKB_CLOCK_ENABLE, SYMCLKB_SRC_SEL, mask_sh),\
+	DCCG_SF(SYMCLKC_CLOCK_ENABLE, SYMCLKC_SRC_SEL, mask_sh),\
+	DCCG_SF(SYMCLKD_CLOCK_ENABLE, SYMCLKD_SRC_SEL, mask_sh),\
+	DCCG_SF(SYMCLKE_CLOCK_ENABLE, SYMCLKE_SRC_SEL, mask_sh),\
+	DCCG_SF(SYMCLKA_CLOCK_ENABLE, SYMCLKA_FE_SRC_SEL, mask_sh),\
+	DCCG_SF(SYMCLKB_CLOCK_ENABLE, SYMCLKB_FE_SRC_SEL, mask_sh),\
+	DCCG_SF(SYMCLKC_CLOCK_ENABLE, SYMCLKC_FE_SRC_SEL, mask_sh),\
+	DCCG_SF(SYMCLKD_CLOCK_ENABLE, SYMCLKD_FE_SRC_SEL, mask_sh),\
+	DCCG_SF(SYMCLKE_CLOCK_ENABLE, SYMCLKE_FE_SRC_SEL, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL5, DTBCLK_P0_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL5, DTBCLK_P1_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL5, DTBCLK_P2_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL5, DTBCLK_P3_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL6, DSCCLK0_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL6, DSCCLK1_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL6, DSCCLK2_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL6, DSCCLK3_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL5, SYMCLKA_FE_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL5, SYMCLKB_FE_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL5, SYMCLKC_FE_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL5, SYMCLKD_FE_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL5, SYMCLKE_FE_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL6, DPPCLK0_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL6, DPPCLK1_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL6, DPPCLK2_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL6, DPPCLK3_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL2, HDMICHARCLK0_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL4, HDMICHARCLK0_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL6, HDMISTREAMCLK0_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL5, SYMCLKA_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL5, SYMCLKB_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL5, SYMCLKC_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL5, SYMCLKD_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL5, SYMCLKE_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, SYMCLK32_ROOT_SE0_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, SYMCLK32_ROOT_SE1_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, SYMCLK32_ROOT_SE2_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, SYMCLK32_ROOT_SE3_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, SYMCLK32_ROOT_LE0_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, SYMCLK32_ROOT_LE1_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, SYMCLK32_SE0_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, SYMCLK32_SE1_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, SYMCLK32_SE2_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, SYMCLK32_SE3_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, SYMCLK32_LE0_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, SYMCLK32_LE1_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL4, PHYA_REFCLK_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL4, PHYB_REFCLK_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL4, PHYC_REFCLK_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL4, PHYD_REFCLK_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL4, PHYE_REFCLK_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL5, DPSTREAMCLK0_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL5, DPSTREAMCLK1_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL5, DPSTREAMCLK2_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL5, DPSTREAMCLK3_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(HDMISTREAMCLK0_DTO_PARAM, HDMISTREAMCLK0_DTO_PHASE, mask_sh),\
+	DCCG_SF(HDMISTREAMCLK0_DTO_PARAM, HDMISTREAMCLK0_DTO_MODULO, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL, DISPCLK_DCCG_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, HDMISTREAMCLK0_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL5, DPSTREAMCLK0_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL5, DPSTREAMCLK1_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL5, DPSTREAMCLK2_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL5, DPSTREAMCLK3_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DISPCLK_FREQ_CHANGE_CNTL, DISPCLK_STEP_DELAY, mask_sh),\
+	DCCG_SF(DISPCLK_FREQ_CHANGE_CNTL, DISPCLK_STEP_SIZE, mask_sh),\
+	DCCG_SF(DISPCLK_FREQ_CHANGE_CNTL, DISPCLK_FREQ_RAMP_DONE, mask_sh),\
+	DCCG_SF(DISPCLK_FREQ_CHANGE_CNTL, DISPCLK_MAX_ERRDET_CYCLES, mask_sh),\
+	DCCG_SF(DISPCLK_FREQ_CHANGE_CNTL, DCCG_FIFO_ERRDET_RESET, mask_sh),\
+	DCCG_SF(DISPCLK_FREQ_CHANGE_CNTL, DCCG_FIFO_ERRDET_STATE, mask_sh),\
+	DCCG_SF(DISPCLK_FREQ_CHANGE_CNTL, DCCG_FIFO_ERRDET_OVR_EN, mask_sh),\
+	DCCG_SF(DISPCLK_FREQ_CHANGE_CNTL, DISPCLK_CHG_FWD_CORR_DISABLE, mask_sh),\
+
+struct dccg *dccg35_create(
+		struct dc_context *ctx,
+		const struct dccg_registers *regs,
+		const struct dccg_shift *dccg_shift,
+		const struct dccg_mask *dccg_mask);
+
+void dccg35_init(struct dccg *dccg);
+
+void dccg35_enable_global_fgcg_rep(struct dccg *dccg, bool value);
+void dccg35_root_gate_disable_control(struct dccg *dccg, uint32_t pipe_idx, uint32_t disable_clock_gating);
+
+
+#endif //__DCN35_DCCG_H__
diff --git a/drivers/gpu/drm/amd/display/dc/dccg/dcn401/dcn401_dccg.c b/drivers/gpu/drm/amd/display/dc/dccg/dcn401/dcn401_dccg.c
new file mode 100644
index 000000000000..663a18ee5162
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dccg/dcn401/dcn401_dccg.c
@@ -0,0 +1,915 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#include "reg_helper.h"
+#include "core_types.h"
+#include "dcn401_dccg.h"
+#include "dcn31/dcn31_dccg.h"
+
+/*
+#include "dmub_common.h"
+#include "dmcub_reg_access_helper.h"
+
+#include "dmub401_common.h"
+#include "dmub401_regs.h"
+#include "dmub401_dccg.h"
+*/
+
+#define TO_DCN_DCCG(dccg)\
+	container_of(dccg, struct dcn_dccg, base)
+
+#define REG(reg) \
+	(dccg_dcn->regs->reg)
+
+#undef FN
+#define FN(reg_name, field_name) \
+	dccg_dcn->dccg_shift->field_name, dccg_dcn->dccg_mask->field_name
+
+#define CTX \
+	dccg_dcn->base.ctx
+#define DC_LOGGER \
+	dccg->ctx->logger
+
+static void dcn401_set_dppclk_enable(struct dccg *dccg,
+				 uint32_t dpp_inst, uint32_t enable)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	switch (dpp_inst) {
+	case 0:
+		REG_UPDATE(DPPCLK_CTRL, DPPCLK0_EN, enable);
+		break;
+	case 1:
+		REG_UPDATE(DPPCLK_CTRL, DPPCLK1_EN, enable);
+		break;
+	case 2:
+		REG_UPDATE(DPPCLK_CTRL, DPPCLK2_EN, enable);
+		break;
+	case 3:
+		REG_UPDATE(DPPCLK_CTRL, DPPCLK3_EN, enable);
+		break;
+	default:
+		break;
+	}
+}
+void dccg401_update_dpp_dto(struct dccg *dccg, int dpp_inst, int req_dppclk)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	if (dccg->ref_dppclk && req_dppclk) {
+		int ref_dppclk = dccg->ref_dppclk;
+		int modulo, phase;
+
+		// phase / modulo = dpp pipe clk / dpp global clk
+		modulo = 0xff;   // use FF at the end
+		phase = ((modulo * req_dppclk) + ref_dppclk - 1) / ref_dppclk;
+
+		if (phase > 0xff) {
+			ASSERT(false);
+			phase = 0xff;
+		}
+
+		REG_SET_2(DPPCLK_DTO_PARAM[dpp_inst], 0,
+				DPPCLK0_DTO_PHASE, phase,
+				DPPCLK0_DTO_MODULO, modulo);
+		dcn401_set_dppclk_enable(dccg, dpp_inst, true);
+	} else {
+		dcn401_set_dppclk_enable(dccg, dpp_inst, false);
+	}
+
+	dccg->pipe_dppclk_khz[dpp_inst] = req_dppclk;
+}
+
+/* This function is a workaround for writing to OTG_PIXEL_RATE_DIV
+ * without the probability of causing a DIG FIFO error.
+ */
+static void dccg401_wait_for_dentist_change_done(
+	struct dccg *dccg)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	uint32_t dentist_dispclk_value = REG_READ(DENTIST_DISPCLK_CNTL);
+
+	REG_WRITE(DENTIST_DISPCLK_CNTL, dentist_dispclk_value);
+	REG_WAIT(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_CHG_DONE, 1, 50, 2000);
+}
+
+void dccg401_get_pixel_rate_div(
+		struct dccg *dccg,
+		uint32_t otg_inst,
+		uint32_t *tmds_div,
+		uint32_t *dp_dto_int)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+	uint32_t val_tmds_div = PIXEL_RATE_DIV_NA;
+
+	switch (otg_inst) {
+	case 0:
+		REG_GET_2(OTG_PIXEL_RATE_DIV,
+			OTG0_TMDS_PIXEL_RATE_DIV, &val_tmds_div,
+			DPDTO0_INT, dp_dto_int);
+		break;
+	case 1:
+		REG_GET_2(OTG_PIXEL_RATE_DIV,
+			OTG1_TMDS_PIXEL_RATE_DIV, &val_tmds_div,
+			DPDTO1_INT, dp_dto_int);
+		break;
+	case 2:
+		REG_GET_2(OTG_PIXEL_RATE_DIV,
+			OTG2_TMDS_PIXEL_RATE_DIV, &val_tmds_div,
+			DPDTO2_INT, dp_dto_int);
+		break;
+	case 3:
+		REG_GET_2(OTG_PIXEL_RATE_DIV,
+			OTG3_TMDS_PIXEL_RATE_DIV, &val_tmds_div,
+			DPDTO3_INT, dp_dto_int);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+
+	*tmds_div = val_tmds_div == 0 ? PIXEL_RATE_DIV_BY_2 : PIXEL_RATE_DIV_BY_4;
+}
+
+void dccg401_set_pixel_rate_div(
+		struct dccg *dccg,
+		uint32_t otg_inst,
+		enum pixel_rate_div tmds_div,
+		enum pixel_rate_div unused)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+	uint32_t cur_tmds_div = PIXEL_RATE_DIV_NA;
+	uint32_t dp_dto_int;
+	uint32_t reg_val;
+
+	// only 2 and 4 are valid on dcn401
+	if (tmds_div != PIXEL_RATE_DIV_BY_2 && tmds_div != PIXEL_RATE_DIV_BY_4) {
+		return;
+	}
+
+	dccg401_get_pixel_rate_div(dccg, otg_inst, &cur_tmds_div, &dp_dto_int);
+	if (tmds_div == cur_tmds_div)
+		return;
+
+	// encode enum to register value
+	reg_val = tmds_div == PIXEL_RATE_DIV_BY_4 ? 1 : 0;
+
+	switch (otg_inst) {
+	case 0:
+		REG_UPDATE(OTG_PIXEL_RATE_DIV,
+				OTG0_TMDS_PIXEL_RATE_DIV, reg_val);
+
+		dccg401_wait_for_dentist_change_done(dccg);
+		break;
+	case 1:
+		REG_UPDATE(OTG_PIXEL_RATE_DIV,
+				OTG1_TMDS_PIXEL_RATE_DIV, reg_val);
+
+		dccg401_wait_for_dentist_change_done(dccg);
+		break;
+	case 2:
+		REG_UPDATE(OTG_PIXEL_RATE_DIV,
+				OTG2_TMDS_PIXEL_RATE_DIV, reg_val);
+
+		dccg401_wait_for_dentist_change_done(dccg);
+		break;
+	case 3:
+		REG_UPDATE(OTG_PIXEL_RATE_DIV,
+				OTG3_TMDS_PIXEL_RATE_DIV, reg_val);
+
+		dccg401_wait_for_dentist_change_done(dccg);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+}
+
+
+void dccg401_set_dtbclk_p_src(
+		struct dccg *dccg,
+		enum streamclk_source src,
+		uint32_t otg_inst)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	uint32_t p_src_sel = 0; /* selects dprefclk */
+	if (src == DTBCLK0)
+		p_src_sel = 2;  /* selects dtbclk0 */
+
+	switch (otg_inst) {
+	case 0:
+		if (src == REFCLK)
+			REG_UPDATE(DTBCLK_P_CNTL,
+					DTBCLK_P0_EN, 0);
+		else
+			REG_UPDATE_2(DTBCLK_P_CNTL,
+					DTBCLK_P0_SRC_SEL, p_src_sel,
+					DTBCLK_P0_EN, 1);
+		break;
+	case 1:
+		if (src == REFCLK)
+			REG_UPDATE(DTBCLK_P_CNTL,
+					DTBCLK_P1_EN, 0);
+		else
+			REG_UPDATE_2(DTBCLK_P_CNTL,
+					DTBCLK_P1_SRC_SEL, p_src_sel,
+					DTBCLK_P1_EN, 1);
+		break;
+	case 2:
+		if (src == REFCLK)
+			REG_UPDATE(DTBCLK_P_CNTL,
+					DTBCLK_P2_EN, 0);
+		else
+			REG_UPDATE_2(DTBCLK_P_CNTL,
+					DTBCLK_P2_SRC_SEL, p_src_sel,
+					DTBCLK_P2_EN, 1);
+		break;
+	case 3:
+		if (src == REFCLK)
+			REG_UPDATE(DTBCLK_P_CNTL,
+					DTBCLK_P3_EN, 0);
+		else
+			REG_UPDATE_2(DTBCLK_P_CNTL,
+					DTBCLK_P3_SRC_SEL, p_src_sel,
+					DTBCLK_P3_EN, 1);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+
+}
+
+void dccg401_set_physymclk(
+		struct dccg *dccg,
+		int phy_inst,
+		enum physymclk_clock_source clk_src,
+		bool force_enable)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	/* Force PHYSYMCLK on and Select phyd32clk as the source of clock which is output to PHY through DCIO */
+	switch (phy_inst) {
+	case 0:
+		if (force_enable) {
+			REG_UPDATE_2(PHYASYMCLK_CLOCK_CNTL,
+					PHYASYMCLK_EN, 1,
+					PHYASYMCLK_SRC_SEL, clk_src);
+			if (dccg->ctx->dc->debug.root_clock_optimization.bits.physymclk)
+				REG_UPDATE(DCCG_GATE_DISABLE_CNTL2,
+					PHYASYMCLK_ROOT_GATE_DISABLE, 1);
+		} else {
+			REG_UPDATE_2(PHYASYMCLK_CLOCK_CNTL,
+					PHYASYMCLK_EN, 0,
+					PHYASYMCLK_SRC_SEL, 0);
+			if (dccg->ctx->dc->debug.root_clock_optimization.bits.physymclk)
+				REG_UPDATE(DCCG_GATE_DISABLE_CNTL2,
+					PHYASYMCLK_ROOT_GATE_DISABLE, 0);
+		}
+		break;
+	case 1:
+		if (force_enable) {
+			REG_UPDATE_2(PHYBSYMCLK_CLOCK_CNTL,
+					PHYBSYMCLK_EN, 1,
+					PHYBSYMCLK_SRC_SEL, clk_src);
+			if (dccg->ctx->dc->debug.root_clock_optimization.bits.physymclk)
+				REG_UPDATE(DCCG_GATE_DISABLE_CNTL2,
+					PHYBSYMCLK_ROOT_GATE_DISABLE, 1);
+		} else {
+			REG_UPDATE_2(PHYBSYMCLK_CLOCK_CNTL,
+					PHYBSYMCLK_EN, 0,
+					PHYBSYMCLK_SRC_SEL, 0);
+			if (dccg->ctx->dc->debug.root_clock_optimization.bits.physymclk)
+				REG_UPDATE(DCCG_GATE_DISABLE_CNTL2,
+					PHYBSYMCLK_ROOT_GATE_DISABLE, 0);
+		}
+		break;
+	case 2:
+		if (force_enable) {
+			REG_UPDATE_2(PHYCSYMCLK_CLOCK_CNTL,
+					PHYCSYMCLK_EN, 1,
+					PHYCSYMCLK_SRC_SEL, clk_src);
+			if (dccg->ctx->dc->debug.root_clock_optimization.bits.physymclk)
+				REG_UPDATE(DCCG_GATE_DISABLE_CNTL2,
+					PHYCSYMCLK_ROOT_GATE_DISABLE, 1);
+		} else {
+			REG_UPDATE_2(PHYCSYMCLK_CLOCK_CNTL,
+					PHYCSYMCLK_EN, 0,
+					PHYCSYMCLK_SRC_SEL, 0);
+			if (dccg->ctx->dc->debug.root_clock_optimization.bits.physymclk)
+				REG_UPDATE(DCCG_GATE_DISABLE_CNTL2,
+					PHYCSYMCLK_ROOT_GATE_DISABLE, 0);
+		}
+		break;
+	case 3:
+		if (force_enable) {
+			REG_UPDATE_2(PHYDSYMCLK_CLOCK_CNTL,
+					PHYDSYMCLK_EN, 1,
+					PHYDSYMCLK_SRC_SEL, clk_src);
+			if (dccg->ctx->dc->debug.root_clock_optimization.bits.physymclk)
+				REG_UPDATE(DCCG_GATE_DISABLE_CNTL2,
+					PHYDSYMCLK_ROOT_GATE_DISABLE, 1);
+		} else {
+			REG_UPDATE_2(PHYDSYMCLK_CLOCK_CNTL,
+					PHYDSYMCLK_EN, 0,
+					PHYDSYMCLK_SRC_SEL, 0);
+			if (dccg->ctx->dc->debug.root_clock_optimization.bits.physymclk)
+				REG_UPDATE(DCCG_GATE_DISABLE_CNTL2,
+					PHYDSYMCLK_ROOT_GATE_DISABLE, 0);
+		}
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+}
+
+void dccg401_get_dccg_ref_freq(struct dccg *dccg,
+		unsigned int xtalin_freq_inKhz,
+		unsigned int *dccg_ref_freq_inKhz)
+{
+	/*
+	 * Assume refclk is sourced from xtalin
+	 * expect 100MHz
+	 */
+	*dccg_ref_freq_inKhz = xtalin_freq_inKhz;
+	return;
+}
+
+static void dccg401_otg_add_pixel(struct dccg *dccg,
+		uint32_t otg_inst)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	REG_UPDATE(OTG_PIXEL_RATE_CNTL[otg_inst],
+			OTG_ADD_PIXEL[otg_inst], 1);
+}
+
+static void dccg401_otg_drop_pixel(struct dccg *dccg,
+		uint32_t otg_inst)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	REG_UPDATE(OTG_PIXEL_RATE_CNTL[otg_inst],
+			OTG_DROP_PIXEL[otg_inst], 1);
+}
+
+void dccg401_enable_symclk32_le(
+		struct dccg *dccg,
+		int hpo_le_inst,
+		enum phyd32clk_clock_source phyd32clk)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	/* select one of the PHYD32CLKs as the source for symclk32_le */
+	switch (hpo_le_inst) {
+	case 0:
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_le)
+			REG_UPDATE_2(DCCG_GATE_DISABLE_CNTL3,
+					SYMCLK32_LE0_GATE_DISABLE, 1,
+					SYMCLK32_ROOT_LE0_GATE_DISABLE, 1);
+		REG_UPDATE_2(SYMCLK32_LE_CNTL,
+				SYMCLK32_LE0_SRC_SEL, phyd32clk,
+				SYMCLK32_LE0_EN, 1);
+		break;
+	case 1:
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_le)
+			REG_UPDATE_2(DCCG_GATE_DISABLE_CNTL3,
+					SYMCLK32_LE1_GATE_DISABLE, 1,
+					SYMCLK32_ROOT_LE1_GATE_DISABLE, 1);
+		REG_UPDATE_2(SYMCLK32_LE_CNTL,
+				SYMCLK32_LE1_SRC_SEL, phyd32clk,
+				SYMCLK32_LE1_EN, 1);
+		break;
+	case 2:
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_le)
+			REG_UPDATE_2(DCCG_GATE_DISABLE_CNTL3,
+					SYMCLK32_LE2_GATE_DISABLE, 1,
+					SYMCLK32_ROOT_LE2_GATE_DISABLE, 1);
+		REG_UPDATE_2(SYMCLK32_LE_CNTL,
+				SYMCLK32_LE2_SRC_SEL, phyd32clk,
+				SYMCLK32_LE2_EN, 1);
+		break;
+	case 3:
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_le)
+			REG_UPDATE_2(DCCG_GATE_DISABLE_CNTL3,
+					SYMCLK32_LE3_GATE_DISABLE, 1,
+					SYMCLK32_ROOT_LE3_GATE_DISABLE, 1);
+		REG_UPDATE_2(SYMCLK32_LE_CNTL,
+				SYMCLK32_LE3_SRC_SEL, phyd32clk,
+				SYMCLK32_LE3_EN, 1);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+}
+
+void dccg401_disable_symclk32_le(
+		struct dccg *dccg,
+		int hpo_le_inst)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	/* set refclk as the source for symclk32_le */
+	switch (hpo_le_inst) {
+	case 0:
+		REG_UPDATE_2(SYMCLK32_LE_CNTL,
+				SYMCLK32_LE0_SRC_SEL, 0,
+				SYMCLK32_LE0_EN, 0);
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_le)
+			REG_UPDATE_2(DCCG_GATE_DISABLE_CNTL3,
+					SYMCLK32_LE0_GATE_DISABLE, 0,
+					SYMCLK32_ROOT_LE0_GATE_DISABLE, 0);
+		break;
+	case 1:
+		REG_UPDATE_2(SYMCLK32_LE_CNTL,
+				SYMCLK32_LE1_SRC_SEL, 0,
+				SYMCLK32_LE1_EN, 0);
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_le)
+			REG_UPDATE_2(DCCG_GATE_DISABLE_CNTL3,
+					SYMCLK32_LE1_GATE_DISABLE, 0,
+					SYMCLK32_ROOT_LE1_GATE_DISABLE, 0);
+		break;
+	case 2:
+		REG_UPDATE_2(SYMCLK32_LE_CNTL,
+				SYMCLK32_LE2_SRC_SEL, 0,
+				SYMCLK32_LE2_EN, 0);
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_le)
+			REG_UPDATE_2(DCCG_GATE_DISABLE_CNTL3,
+					SYMCLK32_LE2_GATE_DISABLE, 0,
+					SYMCLK32_ROOT_LE2_GATE_DISABLE, 0);
+		break;
+	case 3:
+		REG_UPDATE_2(SYMCLK32_LE_CNTL,
+				SYMCLK32_LE3_SRC_SEL, 0,
+				SYMCLK32_LE3_EN, 0);
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_le)
+			REG_UPDATE_2(DCCG_GATE_DISABLE_CNTL3,
+					SYMCLK32_LE3_GATE_DISABLE, 0,
+					SYMCLK32_ROOT_LE3_GATE_DISABLE, 0);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+}
+
+static void dccg401_enable_dpstreamclk(struct dccg *dccg, int otg_inst, int dp_hpo_inst)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	/* enabled to select one of the DTBCLKs for pipe */
+	switch (dp_hpo_inst) {
+	case 0:
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.dpstream)
+			REG_UPDATE_2(DCCG_GATE_DISABLE_CNTL5,
+					DPSTREAMCLK0_ROOT_GATE_DISABLE, 1,
+					DPSTREAMCLK0_GATE_DISABLE, 1);
+		REG_UPDATE_2(DPSTREAMCLK_CNTL,
+				DPSTREAMCLK0_SRC_SEL, otg_inst,
+				DPSTREAMCLK0_EN, 1);
+		break;
+	case 1:
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.dpstream)
+			REG_UPDATE_2(DCCG_GATE_DISABLE_CNTL5,
+					DPSTREAMCLK1_ROOT_GATE_DISABLE, 1,
+					DPSTREAMCLK1_GATE_DISABLE, 1);
+		REG_UPDATE_2(DPSTREAMCLK_CNTL,
+				DPSTREAMCLK1_SRC_SEL, otg_inst,
+				DPSTREAMCLK1_EN, 1);
+		break;
+	case 2:
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.dpstream)
+			REG_UPDATE_2(DCCG_GATE_DISABLE_CNTL5,
+					DPSTREAMCLK2_ROOT_GATE_DISABLE, 1,
+					DPSTREAMCLK2_GATE_DISABLE, 1);
+		REG_UPDATE_2(DPSTREAMCLK_CNTL,
+				DPSTREAMCLK2_SRC_SEL, otg_inst,
+				DPSTREAMCLK2_EN, 1);
+		break;
+	case 3:
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.dpstream)
+			REG_UPDATE_2(DCCG_GATE_DISABLE_CNTL5,
+					DPSTREAMCLK3_ROOT_GATE_DISABLE, 1,
+					DPSTREAMCLK3_GATE_DISABLE, 1);
+		REG_UPDATE_2(DPSTREAMCLK_CNTL,
+				DPSTREAMCLK3_SRC_SEL, otg_inst,
+				DPSTREAMCLK3_EN, 1);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+	if (dccg->ctx->dc->debug.root_clock_optimization.bits.dpstream)
+		REG_UPDATE_2(DCCG_GATE_DISABLE_CNTL3,
+			DPSTREAMCLK_GATE_DISABLE, 1,
+			DPSTREAMCLK_ROOT_GATE_DISABLE, 1);
+}
+
+void dccg401_disable_dpstreamclk(struct dccg *dccg, int dp_hpo_inst)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	switch (dp_hpo_inst) {
+	case 0:
+		REG_UPDATE(DPSTREAMCLK_CNTL,
+				DPSTREAMCLK0_EN, 0);
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.dpstream)
+			REG_UPDATE_2(DCCG_GATE_DISABLE_CNTL5,
+					DPSTREAMCLK0_ROOT_GATE_DISABLE, 0,
+					DPSTREAMCLK0_GATE_DISABLE, 0);
+		break;
+	case 1:
+		REG_UPDATE(DPSTREAMCLK_CNTL,
+				DPSTREAMCLK1_EN, 0);
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.dpstream)
+			REG_UPDATE_2(DCCG_GATE_DISABLE_CNTL5,
+					DPSTREAMCLK1_ROOT_GATE_DISABLE, 0,
+					DPSTREAMCLK1_GATE_DISABLE, 0);
+		break;
+	case 2:
+		REG_UPDATE(DPSTREAMCLK_CNTL,
+				DPSTREAMCLK2_EN, 0);
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.dpstream)
+			REG_UPDATE_2(DCCG_GATE_DISABLE_CNTL5,
+					DPSTREAMCLK2_ROOT_GATE_DISABLE, 0,
+					DPSTREAMCLK2_GATE_DISABLE, 0);
+		break;
+	case 3:
+		REG_UPDATE(DPSTREAMCLK_CNTL,
+				DPSTREAMCLK3_EN, 0);
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.dpstream)
+			REG_UPDATE_2(DCCG_GATE_DISABLE_CNTL5,
+					DPSTREAMCLK3_ROOT_GATE_DISABLE, 0,
+					DPSTREAMCLK3_GATE_DISABLE, 0);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+}
+
+void dccg401_set_dpstreamclk(
+		struct dccg *dccg,
+		enum streamclk_source src,
+		int otg_inst,
+		int dp_hpo_inst)
+{
+	/* enabled to select one of the DTBCLKs for pipe */
+	if (src == REFCLK)
+		dccg401_disable_dpstreamclk(dccg, dp_hpo_inst);
+	else
+		dccg401_enable_dpstreamclk(dccg, otg_inst, dp_hpo_inst);
+}
+
+void dccg401_set_dp_dto(
+		struct dccg *dccg,
+		const struct dp_dto_params *params)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	bool enable = false;
+
+	if (params->otg_inst > 3) {
+		/* dcn401 only has 4 instances */
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+	if (!params->refclk_hz) {
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+
+	if (!dc_is_tmds_signal(params->signal)) {
+		uint64_t dto_integer;
+		uint64_t dto_phase_hz;
+		uint64_t dto_modulo_hz = params->refclk_hz;
+
+		enable = true;
+
+		/* Set DTO values:
+		 * int = target_pix_rate / reference_clock
+		 * phase = target_pix_rate - int * reference_clock,
+		 * modulo = reference_clock */
+		dto_integer = div_u64(params->pixclk_hz, dto_modulo_hz);
+		dto_phase_hz = params->pixclk_hz - dto_integer * dto_modulo_hz;
+
+		if (dto_phase_hz <= 0 && dto_integer <= 0) {
+			/* negative pixel rate should never happen */
+			BREAK_TO_DEBUGGER();
+			return;
+		}
+
+		switch (params->otg_inst) {
+		case 0:
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, DTBCLK_P0_GATE_DISABLE, 1);
+			REG_UPDATE_4(DCCG_GATE_DISABLE_CNTL3,
+					SYMCLK32_SE0_GATE_DISABLE, 1,
+					SYMCLK32_ROOT_SE0_GATE_DISABLE, 1,
+					SYMCLK32_LE0_GATE_DISABLE, 1,
+					SYMCLK32_ROOT_LE0_GATE_DISABLE, 1);
+			break;
+		case 1:
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, DTBCLK_P1_GATE_DISABLE, 1);
+			REG_UPDATE_4(DCCG_GATE_DISABLE_CNTL3,
+					SYMCLK32_SE1_GATE_DISABLE, 1,
+					SYMCLK32_ROOT_SE1_GATE_DISABLE, 1,
+					SYMCLK32_LE1_GATE_DISABLE, 1,
+					SYMCLK32_ROOT_LE1_GATE_DISABLE, 1);
+			break;
+		case 2:
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, DTBCLK_P2_GATE_DISABLE, 1);
+			REG_UPDATE_4(DCCG_GATE_DISABLE_CNTL3,
+					SYMCLK32_SE2_GATE_DISABLE, 1,
+					SYMCLK32_ROOT_SE2_GATE_DISABLE, 1,
+					SYMCLK32_LE2_GATE_DISABLE, 1,
+					SYMCLK32_ROOT_LE2_GATE_DISABLE, 1);
+			break;
+		case 3:
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, DTBCLK_P3_GATE_DISABLE, 1);
+			REG_UPDATE_4(DCCG_GATE_DISABLE_CNTL3,
+					SYMCLK32_SE3_GATE_DISABLE, 1,
+					SYMCLK32_ROOT_SE3_GATE_DISABLE, 1,
+					SYMCLK32_LE3_GATE_DISABLE, 1,
+					SYMCLK32_ROOT_LE3_GATE_DISABLE, 1);
+			break;
+		}
+
+		dccg401_set_dtbclk_p_src(dccg, params->clk_src, params->otg_inst);
+
+		REG_WRITE(DP_DTO_PHASE[params->otg_inst], dto_phase_hz);
+		REG_WRITE(DP_DTO_MODULO[params->otg_inst], dto_modulo_hz);
+
+		switch (params->otg_inst) {
+		case 0:
+			REG_UPDATE(OTG_PIXEL_RATE_DIV,
+					DPDTO0_INT, dto_integer);
+			break;
+		case 1:
+			REG_UPDATE(OTG_PIXEL_RATE_DIV,
+					DPDTO1_INT, dto_integer);
+			break;
+		case 2:
+			REG_UPDATE(OTG_PIXEL_RATE_DIV,
+					DPDTO2_INT, dto_integer);
+			break;
+		case 3:
+			REG_UPDATE(OTG_PIXEL_RATE_DIV,
+					DPDTO3_INT, dto_integer);
+			break;
+		default:
+			BREAK_TO_DEBUGGER();
+			return;
+		}
+	}
+
+	/* Toggle DTO */
+	REG_UPDATE_2(OTG_PIXEL_RATE_CNTL[params->otg_inst],
+			DP_DTO_ENABLE[params->otg_inst], enable,
+			PIPE_DTO_SRC_SEL[params->otg_inst], enable);
+}
+
+void dccg401_init(struct dccg *dccg)
+{
+	/* Set HPO stream encoder to use refclk to avoid case where PHY is
+	 * disabled and SYMCLK32 for HPO SE is sourced from PHYD32CLK which
+	 * will cause DCN to hang.
+	 */
+	dccg31_disable_symclk32_se(dccg, 0);
+	dccg31_disable_symclk32_se(dccg, 1);
+	dccg31_disable_symclk32_se(dccg, 2);
+	dccg31_disable_symclk32_se(dccg, 3);
+
+	if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_le) {
+		dccg401_disable_symclk32_le(dccg, 0);
+		dccg401_disable_symclk32_le(dccg, 1);
+		dccg401_disable_symclk32_le(dccg, 2);
+		dccg401_disable_symclk32_le(dccg, 3);
+	}
+
+	if (dccg->ctx->dc->debug.root_clock_optimization.bits.dpstream) {
+		dccg401_disable_dpstreamclk(dccg, 0);
+		dccg401_disable_dpstreamclk(dccg, 1);
+		dccg401_disable_dpstreamclk(dccg, 2);
+		dccg401_disable_dpstreamclk(dccg, 3);
+	}
+
+	if (dccg->ctx->dc->debug.root_clock_optimization.bits.physymclk) {
+		dccg401_set_physymclk(dccg, 0, PHYSYMCLK_FORCE_SRC_SYMCLK, false);
+		dccg401_set_physymclk(dccg, 1, PHYSYMCLK_FORCE_SRC_SYMCLK, false);
+		dccg401_set_physymclk(dccg, 2, PHYSYMCLK_FORCE_SRC_SYMCLK, false);
+		dccg401_set_physymclk(dccg, 3, PHYSYMCLK_FORCE_SRC_SYMCLK, false);
+	}
+}
+
+void dccg401_set_dto_dscclk(struct dccg *dccg, uint32_t inst, uint32_t num_slices_h)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	switch (inst) {
+	case 0:
+		REG_UPDATE_2(DSCCLK0_DTO_PARAM,
+				DSCCLK0_DTO_PHASE, 1,
+				DSCCLK0_DTO_MODULO, 1);
+		REG_UPDATE(DSCCLK_DTO_CTRL, DSCCLK0_EN, 1);
+
+		break;
+	case 1:
+		REG_UPDATE_2(DSCCLK1_DTO_PARAM,
+				DSCCLK1_DTO_PHASE, 1,
+				DSCCLK1_DTO_MODULO, 1);
+		REG_UPDATE(DSCCLK_DTO_CTRL, DSCCLK1_EN, 1);
+		break;
+	case 2:
+		REG_UPDATE_2(DSCCLK2_DTO_PARAM,
+				DSCCLK2_DTO_PHASE, 1,
+				DSCCLK2_DTO_MODULO, 1);
+		REG_UPDATE(DSCCLK_DTO_CTRL, DSCCLK2_EN, 1);
+		break;
+	case 3:
+		REG_UPDATE_2(DSCCLK3_DTO_PARAM,
+				DSCCLK3_DTO_PHASE, 1,
+				DSCCLK3_DTO_MODULO, 1);
+		REG_UPDATE(DSCCLK_DTO_CTRL, DSCCLK3_EN, 1);
+		break;
+	default:
+		BREAK_TO_DEBUGGER();
+		return;
+	}
+}
+
+void dccg401_set_ref_dscclk(struct dccg *dccg,
+				uint32_t dsc_inst)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	switch (dsc_inst) {
+	case 0:
+		REG_UPDATE(DSCCLK_DTO_CTRL, DSCCLK0_EN, 0);
+		REG_UPDATE_2(DSCCLK0_DTO_PARAM,
+				DSCCLK0_DTO_PHASE, 0,
+				DSCCLK0_DTO_MODULO, 0);
+		break;
+	case 1:
+		REG_UPDATE(DSCCLK_DTO_CTRL, DSCCLK1_EN, 0);
+		REG_UPDATE_2(DSCCLK1_DTO_PARAM,
+				DSCCLK1_DTO_PHASE, 0,
+				DSCCLK1_DTO_MODULO, 0);
+		break;
+	case 2:
+		REG_UPDATE(DSCCLK_DTO_CTRL, DSCCLK2_EN, 0);
+		REG_UPDATE_2(DSCCLK2_DTO_PARAM,
+				DSCCLK2_DTO_PHASE, 0,
+				DSCCLK2_DTO_MODULO, 0);
+		break;
+	case 3:
+		REG_UPDATE(DSCCLK_DTO_CTRL, DSCCLK3_EN, 0);
+		REG_UPDATE_2(DSCCLK3_DTO_PARAM,
+				DSCCLK3_DTO_PHASE, 0,
+				DSCCLK3_DTO_MODULO, 0);
+		break;
+	default:
+		return;
+	}
+}
+
+void dccg401_enable_symclk_se(struct dccg *dccg, uint32_t stream_enc_inst, uint32_t link_enc_inst)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	switch (stream_enc_inst) {
+	case 0:
+		REG_UPDATE_2(SYMCLKA_CLOCK_ENABLE,
+				SYMCLKA_FE_EN, 1,
+				SYMCLKA_FE_SRC_SEL, link_enc_inst);
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se)
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, SYMCLKA_FE_ROOT_GATE_DISABLE, 1);
+		break;
+	case 1:
+		REG_UPDATE_2(SYMCLKB_CLOCK_ENABLE,
+				SYMCLKB_FE_EN, 1,
+				SYMCLKB_FE_SRC_SEL, link_enc_inst);
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se)
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, SYMCLKB_FE_ROOT_GATE_DISABLE, 1);
+		break;
+	case 2:
+		REG_UPDATE_2(SYMCLKC_CLOCK_ENABLE,
+				SYMCLKC_FE_EN, 1,
+				SYMCLKC_FE_SRC_SEL, link_enc_inst);
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se)
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, SYMCLKC_FE_ROOT_GATE_DISABLE, 1);
+		break;
+	case 3:
+		REG_UPDATE_2(SYMCLKD_CLOCK_ENABLE,
+				SYMCLKD_FE_EN, 1,
+				SYMCLKD_FE_SRC_SEL, link_enc_inst);
+		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se)
+			REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, SYMCLKD_FE_ROOT_GATE_DISABLE, 1);
+		break;
+	}
+}
+
+void dccg401_disable_symclk_se(struct dccg *dccg, uint32_t stream_enc_inst, uint32_t link_enc_inst)
+{
+	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
+
+	switch (stream_enc_inst) {
+	case 0:
+		REG_UPDATE_2(SYMCLKA_CLOCK_ENABLE,
+				SYMCLKA_FE_EN, 0,
+				SYMCLKA_FE_SRC_SEL, 0);
+		break;
+	case 1:
+		REG_UPDATE_2(SYMCLKB_CLOCK_ENABLE,
+				SYMCLKB_FE_EN, 0,
+				SYMCLKB_FE_SRC_SEL, 0);
+		break;
+	case 2:
+		REG_UPDATE_2(SYMCLKC_CLOCK_ENABLE,
+				SYMCLKC_FE_EN, 0,
+				SYMCLKC_FE_SRC_SEL, 0);
+		break;
+	case 3:
+		REG_UPDATE_2(SYMCLKD_CLOCK_ENABLE,
+				SYMCLKD_FE_EN, 0,
+				SYMCLKD_FE_SRC_SEL, 0);
+		break;
+	}
+}
+
+static const struct dccg_funcs dccg401_funcs = {
+	.update_dpp_dto = dccg401_update_dpp_dto,
+	.get_dccg_ref_freq = dccg401_get_dccg_ref_freq,
+	.dccg_init = dccg401_init,
+	.set_dpstreamclk = dccg401_set_dpstreamclk,
+	.enable_symclk32_se = dccg31_enable_symclk32_se,
+	.disable_symclk32_se = dccg31_disable_symclk32_se,
+	.enable_symclk32_le = dccg401_enable_symclk32_le,
+	.disable_symclk32_le = dccg401_disable_symclk32_le,
+	.set_physymclk = dccg401_set_physymclk,
+	.set_dtbclk_dto = NULL,
+	.set_dto_dscclk = dccg401_set_dto_dscclk,
+	.set_ref_dscclk = dccg401_set_ref_dscclk,
+	.set_valid_pixel_rate = NULL,
+	.set_fifo_errdet_ovr_en = dccg2_set_fifo_errdet_ovr_en,
+	.set_audio_dtbclk_dto = NULL,
+	.otg_add_pixel = dccg401_otg_add_pixel,
+	.otg_drop_pixel = dccg401_otg_drop_pixel,
+	.set_pixel_rate_div = dccg401_set_pixel_rate_div,
+	.get_pixel_rate_div = dccg401_get_pixel_rate_div,
+	.set_dp_dto = dccg401_set_dp_dto,
+	.enable_symclk_se = dccg401_enable_symclk_se,
+	.disable_symclk_se = dccg401_disable_symclk_se,
+	.set_dtbclk_p_src = dccg401_set_dtbclk_p_src,
+	.dccg_read_reg_state = dccg31_read_reg_state
+};
+
+struct dccg *dccg401_create(
+	struct dc_context *ctx,
+	const struct dccg_registers *regs,
+	const struct dccg_shift *dccg_shift,
+	const struct dccg_mask *dccg_mask)
+{
+	struct dcn_dccg *dccg_dcn = kzalloc(sizeof(*dccg_dcn), GFP_KERNEL);
+	struct dccg *base;
+
+	if (dccg_dcn == NULL) {
+		BREAK_TO_DEBUGGER();
+		return NULL;
+	}
+
+	base = &dccg_dcn->base;
+	base->ctx = ctx;
+	base->funcs = &dccg401_funcs;
+
+	dccg_dcn->regs = regs;
+	dccg_dcn->dccg_shift = dccg_shift;
+	dccg_dcn->dccg_mask = dccg_mask;
+
+	return &dccg_dcn->base;
+}
diff --git a/drivers/gpu/drm/amd/display/dc/dccg/dcn401/dcn401_dccg.h b/drivers/gpu/drm/amd/display/dc/dccg/dcn401/dcn401_dccg.h
new file mode 100644
index 000000000000..5947a35363aa
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dccg/dcn401/dcn401_dccg.h
@@ -0,0 +1,249 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __DCN401_DCCG_H__
+#define __DCN401_DCCG_H__
+
+#include "dcn32/dcn32_dccg.h"
+
+#define DCCG_SFII(block, reg_name, field_prefix, field_name, inst, post_fix)\
+	.field_prefix ## _ ## field_name[inst] = block ## inst ## _ ## reg_name ## __ ## field_prefix ## inst ## _ ## field_name ## post_fix
+
+#define DCCG_MASK_SH_LIST_DCN401(mask_sh) \
+	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_DB_EN, DPPCLK, 0, mask_sh),\
+	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_DB_EN, DPPCLK, 1, mask_sh),\
+	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_DB_EN, DPPCLK, 2, mask_sh),\
+	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_DB_EN, DPPCLK, 3, mask_sh),\
+	DCCG_SF(DPPCLK_CTRL, DPPCLK0_EN, mask_sh),\
+	DCCG_SF(DPPCLK_CTRL, DPPCLK1_EN, mask_sh),\
+	DCCG_SF(DPPCLK_CTRL, DPPCLK2_EN, mask_sh),\
+	DCCG_SF(DPPCLK_CTRL, DPPCLK3_EN, mask_sh),\
+	DCCG_SF(DPPCLK0_DTO_PARAM, DPPCLK0_DTO_PHASE, mask_sh),\
+	DCCG_SF(DPPCLK0_DTO_PARAM, DPPCLK0_DTO_MODULO, mask_sh),\
+	DCCG_SF(HDMICHARCLK0_CLOCK_CNTL, HDMICHARCLK0_EN, mask_sh),\
+	DCCG_SF(HDMICHARCLK0_CLOCK_CNTL, HDMICHARCLK0_SRC_SEL, mask_sh),\
+	DCCG_SF(PHYASYMCLK_CLOCK_CNTL, PHYASYMCLK_EN, mask_sh),\
+	DCCG_SF(PHYASYMCLK_CLOCK_CNTL, PHYASYMCLK_SRC_SEL, mask_sh),\
+	DCCG_SF(PHYBSYMCLK_CLOCK_CNTL, PHYBSYMCLK_EN, mask_sh),\
+	DCCG_SF(PHYBSYMCLK_CLOCK_CNTL, PHYBSYMCLK_SRC_SEL, mask_sh),\
+	DCCG_SF(PHYCSYMCLK_CLOCK_CNTL, PHYCSYMCLK_EN, mask_sh),\
+	DCCG_SF(PHYCSYMCLK_CLOCK_CNTL, PHYCSYMCLK_SRC_SEL, mask_sh),\
+	DCCG_SF(PHYDSYMCLK_CLOCK_CNTL, PHYDSYMCLK_EN, mask_sh),\
+	DCCG_SF(PHYDSYMCLK_CLOCK_CNTL, PHYDSYMCLK_SRC_SEL, mask_sh),\
+	DCCG_SF(DPSTREAMCLK_CNTL, DPSTREAMCLK0_EN, mask_sh),\
+	DCCG_SF(DPSTREAMCLK_CNTL, DPSTREAMCLK1_EN, mask_sh),\
+	DCCG_SF(DPSTREAMCLK_CNTL, DPSTREAMCLK2_EN, mask_sh),\
+	DCCG_SF(DPSTREAMCLK_CNTL, DPSTREAMCLK3_EN, mask_sh),\
+	DCCG_SF(DPSTREAMCLK_CNTL, DPSTREAMCLK0_SRC_SEL, mask_sh),\
+	DCCG_SF(DPSTREAMCLK_CNTL, DPSTREAMCLK1_SRC_SEL, mask_sh),\
+	DCCG_SF(DPSTREAMCLK_CNTL, DPSTREAMCLK2_SRC_SEL, mask_sh),\
+	DCCG_SF(DPSTREAMCLK_CNTL, DPSTREAMCLK3_SRC_SEL, mask_sh),\
+	DCCG_SF(HDMISTREAMCLK_CNTL, HDMISTREAMCLK0_EN, mask_sh),\
+	DCCG_SF(HDMISTREAMCLK_CNTL, HDMISTREAMCLK0_SRC_SEL, mask_sh),\
+	DCCG_SF(SYMCLK32_SE_CNTL, SYMCLK32_SE0_SRC_SEL, mask_sh),\
+	DCCG_SF(SYMCLK32_SE_CNTL, SYMCLK32_SE1_SRC_SEL, mask_sh),\
+	DCCG_SF(SYMCLK32_SE_CNTL, SYMCLK32_SE2_SRC_SEL, mask_sh),\
+	DCCG_SF(SYMCLK32_SE_CNTL, SYMCLK32_SE3_SRC_SEL, mask_sh),\
+	DCCG_SF(SYMCLK32_SE_CNTL, SYMCLK32_SE0_EN, mask_sh),\
+	DCCG_SF(SYMCLK32_SE_CNTL, SYMCLK32_SE1_EN, mask_sh),\
+	DCCG_SF(SYMCLK32_SE_CNTL, SYMCLK32_SE2_EN, mask_sh),\
+	DCCG_SF(SYMCLK32_SE_CNTL, SYMCLK32_SE3_EN, mask_sh),\
+	DCCG_SF(SYMCLK32_LE_CNTL, SYMCLK32_LE0_SRC_SEL, mask_sh),\
+	DCCG_SF(SYMCLK32_LE_CNTL, SYMCLK32_LE1_SRC_SEL, mask_sh),\
+	DCCG_SF(SYMCLK32_LE_CNTL, SYMCLK32_LE0_EN, mask_sh),\
+	DCCG_SF(SYMCLK32_LE_CNTL, SYMCLK32_LE1_EN, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, PIPE, DTO_SRC_SEL, 0, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, PIPE, DTO_SRC_SEL, 1, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, PIPE, DTO_SRC_SEL, 2, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, PIPE, DTO_SRC_SEL, 3, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, ADD_PIXEL, 0, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, ADD_PIXEL, 1, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, ADD_PIXEL, 2, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, ADD_PIXEL, 3, mask_sh),\
+	DCCG_SF(OTG_PIXEL_RATE_DIV, OTG0_TMDS_PIXEL_RATE_DIV, mask_sh),\
+	DCCG_SF(OTG_PIXEL_RATE_DIV, DPDTO0_INT, mask_sh),\
+	DCCG_SF(OTG_PIXEL_RATE_DIV, OTG1_TMDS_PIXEL_RATE_DIV, mask_sh),\
+	DCCG_SF(OTG_PIXEL_RATE_DIV, DPDTO1_INT, mask_sh),\
+	DCCG_SF(OTG_PIXEL_RATE_DIV, OTG2_TMDS_PIXEL_RATE_DIV, mask_sh),\
+	DCCG_SF(OTG_PIXEL_RATE_DIV, DPDTO2_INT, mask_sh),\
+	DCCG_SF(OTG_PIXEL_RATE_DIV, OTG3_TMDS_PIXEL_RATE_DIV, mask_sh),\
+	DCCG_SF(OTG_PIXEL_RATE_DIV, DPDTO3_INT, mask_sh),\
+	DCCG_SF(DTBCLK_P_CNTL, DTBCLK_P0_SRC_SEL, mask_sh),\
+	DCCG_SF(DTBCLK_P_CNTL, DTBCLK_P0_EN, mask_sh),\
+	DCCG_SF(DTBCLK_P_CNTL, DTBCLK_P1_SRC_SEL, mask_sh),\
+	DCCG_SF(DTBCLK_P_CNTL, DTBCLK_P1_EN, mask_sh),\
+	DCCG_SF(DTBCLK_P_CNTL, DTBCLK_P2_SRC_SEL, mask_sh),\
+	DCCG_SF(DTBCLK_P_CNTL, DTBCLK_P2_EN, mask_sh),\
+	DCCG_SF(DTBCLK_P_CNTL, DTBCLK_P3_SRC_SEL, mask_sh),\
+	DCCG_SF(DTBCLK_P_CNTL, DTBCLK_P3_EN, mask_sh),\
+	DCCG_SF(DCCG_AUDIO_DTO_SOURCE, DCCG_AUDIO_DTO_SEL, mask_sh),\
+	DCCG_SF(DCCG_AUDIO_DTO_SOURCE, DCCG_AUDIO_DTO0_SOURCE_SEL, mask_sh),\
+	DCCG_SF(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_CHG_DONE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL2, PHYASYMCLK_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL2, PHYBSYMCLK_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL2, PHYCSYMCLK_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL2, PHYDSYMCLK_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, DP_DTO, ENABLE, 0, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, DP_DTO, ENABLE, 1, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, DP_DTO, ENABLE, 2, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, DP_DTO, ENABLE, 3, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, PIPE, DTO_SRC_SEL, 0, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, PIPE, DTO_SRC_SEL, 1, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, PIPE, DTO_SRC_SEL, 2, mask_sh),\
+	DCCG_SFII(OTG, PIXEL_RATE_CNTL, PIPE, DTO_SRC_SEL, 3, mask_sh),\
+	DCCG_SF(DSCCLK_DTO_CTRL, DSCCLK0_EN, mask_sh),\
+	DCCG_SF(DSCCLK_DTO_CTRL, DSCCLK1_EN, mask_sh),\
+	DCCG_SF(DSCCLK_DTO_CTRL, DSCCLK2_EN, mask_sh),\
+	DCCG_SF(DSCCLK_DTO_CTRL, DSCCLK3_EN, mask_sh),\
+	DCCG_SF(DSCCLK0_DTO_PARAM, DSCCLK0_DTO_PHASE, mask_sh),\
+	DCCG_SF(DSCCLK0_DTO_PARAM, DSCCLK0_DTO_MODULO, mask_sh),\
+	DCCG_SF(DSCCLK1_DTO_PARAM, DSCCLK1_DTO_PHASE, mask_sh),\
+	DCCG_SF(DSCCLK1_DTO_PARAM, DSCCLK1_DTO_MODULO, mask_sh),\
+	DCCG_SF(DSCCLK2_DTO_PARAM, DSCCLK2_DTO_PHASE, mask_sh),\
+	DCCG_SF(DSCCLK2_DTO_PARAM, DSCCLK2_DTO_MODULO, mask_sh),\
+	DCCG_SF(DSCCLK3_DTO_PARAM, DSCCLK3_DTO_PHASE, mask_sh),\
+	DCCG_SF(DSCCLK3_DTO_PARAM, DSCCLK3_DTO_MODULO, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL2, HDMICHARCLK0_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, HDMISTREAMCLK0_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, SYMCLK32_ROOT_SE0_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, SYMCLK32_ROOT_SE1_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, SYMCLK32_ROOT_SE2_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, SYMCLK32_ROOT_SE3_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, SYMCLK32_ROOT_LE0_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, SYMCLK32_ROOT_LE1_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, SYMCLK32_ROOT_LE2_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, SYMCLK32_ROOT_LE3_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, SYMCLK32_SE0_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, SYMCLK32_SE1_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, SYMCLK32_SE2_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, SYMCLK32_SE3_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, SYMCLK32_LE0_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, SYMCLK32_LE1_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, SYMCLK32_LE2_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL3, SYMCLK32_LE3_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL4, HDMICHARCLK0_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL4, PHYA_REFCLK_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL4, PHYB_REFCLK_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL4, PHYC_REFCLK_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL4, PHYD_REFCLK_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL5, DTBCLK_P0_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL5, DTBCLK_P1_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL5, DTBCLK_P2_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL5, DTBCLK_P3_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL5, SYMCLKA_FE_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL5, SYMCLKB_FE_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL5, SYMCLKC_FE_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL5, SYMCLKD_FE_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL5, SYMCLKA_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL5, SYMCLKB_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL5, SYMCLKC_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL5, SYMCLKD_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL5, DPSTREAMCLK0_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL5, DPSTREAMCLK1_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL5, DPSTREAMCLK2_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL5, DPSTREAMCLK3_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL5, DPSTREAMCLK0_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL5, DPSTREAMCLK1_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL5, DPSTREAMCLK2_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL5, DPSTREAMCLK3_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL6, DSCCLK0_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL6, DSCCLK1_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL6, DSCCLK2_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL6, DSCCLK3_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL6, DPPCLK0_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL6, DPPCLK1_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL6, DPPCLK2_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL6, DPPCLK3_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(DCCG_GATE_DISABLE_CNTL6, HDMISTREAMCLK0_ROOT_GATE_DISABLE, mask_sh),\
+	DCCG_SF(SYMCLKA_CLOCK_ENABLE, SYMCLKA_CLOCK_ENABLE, mask_sh),\
+	DCCG_SF(SYMCLKB_CLOCK_ENABLE, SYMCLKB_CLOCK_ENABLE, mask_sh),\
+	DCCG_SF(SYMCLKC_CLOCK_ENABLE, SYMCLKC_CLOCK_ENABLE, mask_sh),\
+	DCCG_SF(SYMCLKD_CLOCK_ENABLE, SYMCLKD_CLOCK_ENABLE, mask_sh),\
+	DCCG_SF(SYMCLKA_CLOCK_ENABLE, SYMCLKA_FE_EN, mask_sh),\
+	DCCG_SF(SYMCLKB_CLOCK_ENABLE, SYMCLKB_FE_EN, mask_sh),\
+	DCCG_SF(SYMCLKC_CLOCK_ENABLE, SYMCLKC_FE_EN, mask_sh),\
+	DCCG_SF(SYMCLKD_CLOCK_ENABLE, SYMCLKD_FE_EN, mask_sh),\
+	DCCG_SF(SYMCLKA_CLOCK_ENABLE, SYMCLKA_FE_SRC_SEL, mask_sh),\
+	DCCG_SF(SYMCLKB_CLOCK_ENABLE, SYMCLKB_FE_SRC_SEL, mask_sh),\
+	DCCG_SF(SYMCLKC_CLOCK_ENABLE, SYMCLKC_FE_SRC_SEL, mask_sh),\
+	DCCG_SF(SYMCLKD_CLOCK_ENABLE, SYMCLKD_FE_SRC_SEL, mask_sh),\
+
+void dccg401_init(struct dccg *dccg);
+
+void dccg401_update_dpp_dto(struct dccg *dccg, int dpp_inst, int req_dppclk);
+void dccg401_get_dccg_ref_freq(struct dccg *dccg,
+		unsigned int xtalin_freq_inKhz,
+		unsigned int *dccg_ref_freq_inKhz);
+void dccg401_set_dpstreamclk(
+		struct dccg *dccg,
+		enum streamclk_source src,
+		int otg_inst,
+		int dp_hpo_inst);
+void dccg401_enable_symclk32_le(
+		struct dccg *dccg,
+		int hpo_le_inst,
+		enum phyd32clk_clock_source phyd32clk);
+void dccg401_disable_symclk32_le(
+		struct dccg *dccg,
+		int hpo_le_inst);
+void dccg401_disable_dpstreamclk(struct dccg *dccg, int dp_hpo_inst);
+void dccg401_set_dto_dscclk(struct dccg *dccg, uint32_t inst, uint32_t num_slices_h);
+void dccg401_set_ref_dscclk(struct dccg *dccg,
+				uint32_t dsc_inst);
+void dccg401_set_src_sel(
+	struct dccg *dccg,
+	const struct dtbclk_dto_params *params);
+void dccg401_set_pixel_rate_div(
+		struct dccg *dccg,
+		uint32_t otg_inst,
+		enum pixel_rate_div tmds_div,
+		enum pixel_rate_div unused);
+void dccg401_get_pixel_rate_div(
+		struct dccg *dccg,
+		uint32_t otg_inst,
+		uint32_t *tmds_div,
+		uint32_t *dp_dto_int);
+void dccg401_set_dp_dto(
+		struct dccg *dccg,
+		const struct dp_dto_params *params);
+void dccg401_enable_symclk_se(struct dccg *dccg, uint32_t stream_enc_inst, uint32_t link_enc_inst);
+void dccg401_disable_symclk_se(struct dccg *dccg, uint32_t stream_enc_inst, uint32_t link_enc_inst);
+void dccg401_set_dtbclk_p_src(
+		struct dccg *dccg,
+		enum streamclk_source src,
+		uint32_t otg_inst);
+struct dccg *dccg401_create(
+	struct dc_context *ctx,
+	const struct dccg_registers *regs,
+	const struct dccg_shift *dccg_shift,
+	const struct dccg_mask *dccg_mask);
+
+void dccg401_set_physymclk(
+	struct dccg *dccg,
+	int phy_inst,
+	enum physymclk_clock_source clk_src,
+	bool force_enable);
+
+#endif //__DCN401_DCCG_H__
diff --git a/drivers/gpu/drm/amd/display/dc/dce/Makefile b/drivers/gpu/drm/amd/display/dc/dce/Makefile
index 0d7db132a20f..986e0e7abbc2 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/Makefile
+++ b/drivers/gpu/drm/amd/display/dc/dce/Makefile
@@ -26,11 +26,11 @@
 #   - register programming through common macros that look up register 
 #     offset/shift/mask stored in dce_hw struct
 
-DCE = dce_audio.o dce_stream_encoder.o dce_link_encoder.o dce_hwseq.o \
+DCE = dce_audio.o dce_stream_encoder.o dce_link_encoder.o \
 dce_mem_input.o dce_clock_source.o dce_scl_filters.o dce_transform.o \
 dce_opp.o dce_dmcu.o dce_abm.o dce_ipp.o dce_aux.o \
-dce_i2c.o dce_i2c_hw.o dce_i2c_sw.o dmub_psr.o dmub_abm.o dce_panel_cntl.o \
-dmub_hw_lock_mgr.o dmub_outbox.o
+dce_i2c.o dce_i2c_hw.o dce_i2c_sw.o dmub_psr.o dmub_abm.o dmub_abm_lcd.o dce_panel_cntl.o \
+dmub_hw_lock_mgr.o dmub_outbox.o dmub_replay.o
 
 AMD_DAL_DCE = $(addprefix $(AMDDALPATH)/dc/dce/,$(DCE))
 
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_abm.c b/drivers/gpu/drm/amd/display/dc/dce/dce_abm.c
index 874b132fe1d7..2dcf394edf22 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_abm.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_abm.c
@@ -135,7 +135,7 @@ static void dmcu_set_backlight_level(
 			0, 1, 80000);
 }
 
-static void dce_abm_init(struct abm *abm, uint32_t backlight)
+static void dce_abm_init(struct abm *abm, uint32_t backlight, uint32_t user_level)
 {
 	struct dce_abm *abm_dce = TO_DCE_ABM(abm);
 
@@ -162,7 +162,7 @@ static void dce_abm_init(struct abm *abm, uint32_t backlight)
 			BL1_PWM_TARGET_ABM_LEVEL, backlight);
 
 	REG_UPDATE(BL1_PWM_USER_LEVEL,
-			BL1_PWM_USER_LEVEL, backlight);
+			BL1_PWM_USER_LEVEL, user_level);
 
 	REG_UPDATE_2(DC_ABM1_LS_MIN_MAX_PIXEL_VALUE_THRES,
 			ABM1_LS_MIN_PIXEL_VALUE_THRES, 0,
@@ -283,7 +283,7 @@ struct abm *dce_abm_create(
 	const struct dce_abm_shift *abm_shift,
 	const struct dce_abm_mask *abm_mask)
 {
-	struct dce_abm *abm_dce = kzalloc(sizeof(*abm_dce), GFP_ATOMIC);
+	struct dce_abm *abm_dce = kzalloc(sizeof(*abm_dce), GFP_KERNEL);
 
 	if (abm_dce == NULL) {
 		BREAK_TO_DEBUGGER();
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_abm.h b/drivers/gpu/drm/amd/display/dc/dce/dce_abm.h
index b699d1b2ba83..3d819fc5654c 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_abm.h
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_abm.h
@@ -168,8 +168,7 @@
 	ABM_SF(DC_ABM1_HGLS_REG_READ_PROGRESS, \
 			ABM1_BL_REG_READ_MISSED_FRAME_CLEAR, mask_sh)
 
-#define ABM_MASK_SH_LIST_DCN10(mask_sh) \
-	ABM_COMMON_MASK_SH_LIST_DCE_COMMON_BASE(mask_sh), \
+#define ABM_MASK_SH_LIST_DCN10_COMMON(mask_sh) \
 	ABM_SF(ABM0_DC_ABM1_HG_MISC_CTRL, \
 			ABM1_HG_NUM_OF_BINS_SEL, mask_sh), \
 	ABM_SF(ABM0_DC_ABM1_HG_MISC_CTRL, \
@@ -199,9 +198,107 @@
 	ABM_SF(ABM0_DC_ABM1_HGLS_REG_READ_PROGRESS, \
 			ABM1_BL_REG_READ_MISSED_FRAME_CLEAR, mask_sh)
 
-#define ABM_MASK_SH_LIST_DCN20(mask_sh) ABM_MASK_SH_LIST_DCE110(mask_sh)
+#define ABM_MASK_SH_LIST_DCN10(mask_sh) \
+	ABM_COMMON_MASK_SH_LIST_DCE_COMMON_BASE(mask_sh), \
+	ABM_MASK_SH_LIST_DCN10_COMMON(mask_sh)
 
+#define ABM_MASK_SH_LIST_DCN20(mask_sh) ABM_MASK_SH_LIST_DCE110(mask_sh)
 #define ABM_MASK_SH_LIST_DCN30(mask_sh) ABM_MASK_SH_LIST_DCN10(mask_sh)
+#define ABM_MASK_SH_LIST_DCN35(mask_sh) ABM_MASK_SH_LIST_DCN10_COMMON(mask_sh)
+
+#define ABM_MASK_SH_LIST_DCN32(mask_sh) \
+	ABM_SF(ABM0_DC_ABM1_HG_MISC_CTRL, \
+			ABM1_HG_NUM_OF_BINS_SEL, mask_sh), \
+	ABM_SF(ABM0_DC_ABM1_HG_MISC_CTRL, \
+			ABM1_HG_VMAX_SEL, mask_sh), \
+	ABM_SF(ABM0_DC_ABM1_HG_MISC_CTRL, \
+			ABM1_HG_BIN_BITWIDTH_SIZE_SEL, mask_sh), \
+	ABM_SF(ABM0_DC_ABM1_IPCSC_COEFF_SEL, \
+			ABM1_IPCSC_COEFF_SEL_R, mask_sh), \
+	ABM_SF(ABM0_DC_ABM1_IPCSC_COEFF_SEL, \
+			ABM1_IPCSC_COEFF_SEL_G, mask_sh), \
+	ABM_SF(ABM0_DC_ABM1_IPCSC_COEFF_SEL, \
+			ABM1_IPCSC_COEFF_SEL_B, mask_sh), \
+	ABM_SF(ABM0_BL1_PWM_CURRENT_ABM_LEVEL, \
+			BL1_PWM_CURRENT_ABM_LEVEL, mask_sh), \
+	ABM_SF(ABM0_BL1_PWM_TARGET_ABM_LEVEL, \
+			BL1_PWM_TARGET_ABM_LEVEL, mask_sh), \
+	ABM_SF(ABM0_BL1_PWM_USER_LEVEL, \
+			BL1_PWM_USER_LEVEL, mask_sh), \
+	ABM_SF(ABM0_DC_ABM1_LS_MIN_MAX_PIXEL_VALUE_THRES, \
+			ABM1_LS_MIN_PIXEL_VALUE_THRES, mask_sh), \
+	ABM_SF(ABM0_DC_ABM1_LS_MIN_MAX_PIXEL_VALUE_THRES, \
+			ABM1_LS_MAX_PIXEL_VALUE_THRES, mask_sh), \
+	ABM_SF(ABM0_DC_ABM1_HGLS_REG_READ_PROGRESS, \
+			ABM1_HG_REG_READ_MISSED_FRAME_CLEAR, mask_sh), \
+	ABM_SF(ABM0_DC_ABM1_HGLS_REG_READ_PROGRESS, \
+			ABM1_LS_REG_READ_MISSED_FRAME_CLEAR, mask_sh), \
+	ABM_SF(ABM0_DC_ABM1_HGLS_REG_READ_PROGRESS, \
+			ABM1_BL_REG_READ_MISSED_FRAME_CLEAR, mask_sh)
+
+#define ABM_MASK_SH_LIST_DCN401(mask_sh) \
+	ABM_SF(ABM0_DC_ABM1_HG_MISC_CTRL, \
+			ABM1_HG_NUM_OF_BINS_SEL, mask_sh), \
+	ABM_SF(ABM0_DC_ABM1_HG_MISC_CTRL, \
+			ABM1_HG_VMAX_SEL, mask_sh), \
+	ABM_SF(ABM0_DC_ABM1_HG_MISC_CTRL, \
+			ABM1_HG_BIN_BITWIDTH_SIZE_SEL, mask_sh), \
+	ABM_SF(ABM0_DC_ABM1_IPCSC_COEFF_SEL, \
+			ABM1_IPCSC_COEFF_SEL_R, mask_sh), \
+	ABM_SF(ABM0_DC_ABM1_IPCSC_COEFF_SEL, \
+			ABM1_IPCSC_COEFF_SEL_G, mask_sh), \
+	ABM_SF(ABM0_DC_ABM1_IPCSC_COEFF_SEL, \
+			ABM1_IPCSC_COEFF_SEL_B, mask_sh), \
+	ABM_SF(ABM0_BL1_PWM_CURRENT_ABM_LEVEL, \
+			BL1_PWM_CURRENT_ABM_LEVEL, mask_sh), \
+	ABM_SF(ABM0_BL1_PWM_TARGET_ABM_LEVEL, \
+			BL1_PWM_TARGET_ABM_LEVEL, mask_sh), \
+	ABM_SF(ABM0_BL1_PWM_USER_LEVEL, \
+			BL1_PWM_USER_LEVEL, mask_sh), \
+	ABM_SF(ABM0_DC_ABM1_LS_MIN_MAX_PIXEL_VALUE_THRES, \
+			ABM1_LS_MIN_PIXEL_VALUE_THRES, mask_sh), \
+	ABM_SF(ABM0_DC_ABM1_LS_MIN_MAX_PIXEL_VALUE_THRES, \
+			ABM1_LS_MAX_PIXEL_VALUE_THRES, mask_sh), \
+	ABM_SF(ABM0_DC_ABM1_HGLS_REG_READ_PROGRESS, \
+			ABM1_HG_REG_READ_MISSED_FRAME_CLEAR, mask_sh), \
+	ABM_SF(ABM0_DC_ABM1_HGLS_REG_READ_PROGRESS, \
+			ABM1_LS_REG_READ_MISSED_FRAME_CLEAR, mask_sh), \
+	ABM_SF(ABM0_DC_ABM1_HGLS_REG_READ_PROGRESS, \
+			ABM1_BL_REG_READ_MISSED_FRAME_CLEAR, mask_sh), \
+	ABM_SF(ABM0_DC_ABM1_ACE_OFFSET_SLOPE_DATA, \
+			ABM1_ACE_SLOPE_DATA, mask_sh), \
+	ABM_SF(ABM0_DC_ABM1_ACE_OFFSET_SLOPE_DATA, \
+			ABM1_ACE_OFFSET_DATA, mask_sh), \
+	ABM_SF(ABM0_DC_ABM1_ACE_PWL_CNTL, \
+			ABM1_ACE_OFFSET_SLOPE_INDEX, mask_sh), \
+	ABM_SF(ABM0_DC_ABM1_ACE_PWL_CNTL, \
+			ABM1_ACE_THRES_INDEX, mask_sh), \
+	ABM_SF(ABM0_DC_ABM1_ACE_PWL_CNTL, \
+			ABM1_ACE_IGNORE_MASTER_LOCK_EN, mask_sh), \
+	ABM_SF(ABM0_DC_ABM1_ACE_PWL_CNTL, \
+			ABM1_ACE_READBACK_DB_REG_VALUE_EN, mask_sh), \
+	ABM_SF(ABM0_DC_ABM1_ACE_PWL_CNTL, \
+			ABM1_ACE_DBUF_REG_UPDATE_PENDING, mask_sh), \
+	ABM_SF(ABM0_DC_ABM1_ACE_PWL_CNTL, \
+			ABM1_ACE_LOCK, mask_sh), \
+	ABM_SF(ABM0_DC_ABM1_ACE_THRES_DATA, \
+			ABM1_ACE_THRES_DATA_1, mask_sh), \
+	ABM_SF(ABM0_DC_ABM1_ACE_THRES_DATA, \
+			ABM1_ACE_THRES_DATA_2, mask_sh), \
+	ABM_SF(ABM0_DC_ABM1_HG_RESULT_DATA, \
+			ABM1_HG_RESULT_DATA, mask_sh), \
+	ABM_SF(ABM0_DC_ABM1_HG_RESULT_INDEX, \
+			ABM1_HG_RESULT_INDEX, mask_sh), \
+	ABM_SF(ABM0_DC_ABM1_HG_BIN_33_40_SHIFT_INDEX, \
+			ABM1_HG_BIN_33_40_SHIFT_INDEX, mask_sh), \
+	ABM_SF(ABM0_DC_ABM1_HG_BIN_33_64_SHIFT_FLAG, \
+			ABM1_HG_BIN_33_64_SHIFT_FLAG, mask_sh), \
+	ABM_SF(ABM0_DC_ABM1_HG_BIN_41_48_SHIFT_INDEX, \
+			ABM1_HG_BIN_41_48_SHIFT_INDEX, mask_sh), \
+	ABM_SF(ABM0_DC_ABM1_HG_BIN_49_56_SHIFT_INDEX, \
+			ABM1_HG_BIN_49_56_SHIFT_INDEX, mask_sh), \
+	ABM_SF(ABM0_DC_ABM1_HG_BIN_57_64_SHIFT_INDEX, \
+			ABM1_HG_BIN_57_64_SHIFT_INDEX, mask_sh)
 
 #define ABM_REG_FIELD_LIST(type) \
 	type ABM1_HG_NUM_OF_BINS_SEL; \
@@ -221,7 +318,24 @@
 	type MASTER_COMM_INTERRUPT; \
 	type MASTER_COMM_CMD_REG_BYTE0; \
 	type MASTER_COMM_CMD_REG_BYTE1; \
-	type MASTER_COMM_CMD_REG_BYTE2
+	type MASTER_COMM_CMD_REG_BYTE2; \
+	type ABM1_HG_BIN_33_40_SHIFT_INDEX; \
+	type ABM1_HG_BIN_33_64_SHIFT_FLAG; \
+	type ABM1_HG_BIN_41_48_SHIFT_INDEX; \
+	type ABM1_HG_BIN_49_56_SHIFT_INDEX; \
+	type ABM1_HG_BIN_57_64_SHIFT_INDEX; \
+	type ABM1_HG_RESULT_DATA; \
+	type ABM1_HG_RESULT_INDEX; \
+	type ABM1_ACE_SLOPE_DATA; \
+	type ABM1_ACE_OFFSET_DATA; \
+	type ABM1_ACE_OFFSET_SLOPE_INDEX; \
+	type ABM1_ACE_THRES_INDEX; \
+	type ABM1_ACE_IGNORE_MASTER_LOCK_EN; \
+	type ABM1_ACE_READBACK_DB_REG_VALUE_EN; \
+	type ABM1_ACE_DBUF_REG_UPDATE_PENDING; \
+	type ABM1_ACE_LOCK; \
+	type ABM1_ACE_THRES_DATA_1; \
+	type ABM1_ACE_THRES_DATA_2
 
 struct dce_abm_shift {
 	ABM_REG_FIELD_LIST(uint8_t);
@@ -243,6 +357,16 @@ struct dce_abm_registers {
 	uint32_t DC_ABM1_LS_MIN_MAX_PIXEL_VALUE_THRES;
 	uint32_t DC_ABM1_HGLS_REG_READ_PROGRESS;
 	uint32_t DC_ABM1_ACE_OFFSET_SLOPE_0;
+	uint32_t DC_ABM1_ACE_OFFSET_SLOPE_DATA;
+	uint32_t DC_ABM1_ACE_PWL_CNTL;
+	uint32_t DC_ABM1_HG_BIN_33_40_SHIFT_INDEX;
+	uint32_t DC_ABM1_HG_BIN_33_64_SHIFT_FLAG;
+	uint32_t DC_ABM1_HG_BIN_41_48_SHIFT_INDEX;
+	uint32_t DC_ABM1_HG_BIN_49_56_SHIFT_INDEX;
+	uint32_t DC_ABM1_HG_BIN_57_64_SHIFT_INDEX;
+	uint32_t DC_ABM1_HG_RESULT_DATA;
+	uint32_t DC_ABM1_HG_RESULT_INDEX;
+	uint32_t DC_ABM1_ACE_THRES_DATA;
 	uint32_t DC_ABM1_ACE_THRES_12;
 	uint32_t MASTER_COMM_CNTL_REG;
 	uint32_t MASTER_COMM_CMD_REG;
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_audio.c b/drivers/gpu/drm/amd/display/dc/dce/dce_audio.c
index 27218ede150a..fcad61c618a1 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_audio.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_audio.c
@@ -23,8 +23,6 @@
  *
  */
 
-#include <linux/slab.h>
-
 #include "reg_helper.h"
 #include "dce_audio.h"
 #include "dce/dce_11_0_d.h"
@@ -67,9 +65,6 @@ static void write_indirect_azalia_reg(struct audio *audio,
 	/* AZALIA_F0_CODEC_ENDPOINT_DATA  endpoint data  */
 	REG_SET(AZALIA_F0_CODEC_ENDPOINT_DATA, 0,
 			AZALIA_ENDPOINT_REG_DATA, reg_data);
-
-	DC_LOG_HW_AUDIO("AUDIO:write_indirect_azalia_reg: index: %u  data: %u\n",
-		reg_index, reg_data);
 }
 
 static uint32_t read_indirect_azalia_reg(struct audio *audio, uint32_t reg_index)
@@ -85,9 +80,6 @@ static uint32_t read_indirect_azalia_reg(struct audio *audio, uint32_t reg_index
 	/* AZALIA_F0_CODEC_ENDPOINT_DATA  endpoint data  */
 	value = REG_READ(AZALIA_F0_CODEC_ENDPOINT_DATA);
 
-	DC_LOG_HW_AUDIO("AUDIO:read_indirect_azalia_reg: index: %u  data: %u\n",
-		reg_index, value);
-
 	return value;
 }
 
@@ -247,27 +239,295 @@ static void check_audio_bandwidth_hdmi(
 		}
 	}
 }
+static struct fixed31_32 get_link_symbol_clk_freq_mhz(enum dc_link_rate link_rate)
+{
+	switch (link_rate) {
+	case LINK_RATE_LOW:
+		return dc_fixpt_from_int(162); /* 162 MHz */
+	case LINK_RATE_HIGH:
+		return dc_fixpt_from_int(270); /* 270 MHz */
+	case LINK_RATE_HIGH2:
+		return dc_fixpt_from_int(540); /* 540 MHz */
+	case LINK_RATE_HIGH3:
+		return dc_fixpt_from_int(810); /* 810 MHz */
+	case LINK_RATE_UHBR10:
+		return dc_fixpt_from_fraction(3125, 10); /* 312.5 MHz */
+	case LINK_RATE_UHBR13_5:
+		return dc_fixpt_from_fraction(421875, 1000); /* 421.875 MHz */
+	case LINK_RATE_UHBR20:
+		return dc_fixpt_from_int(625); /* 625 MHz */
+	default:
+		/* Unexpected case, this requires debug if encountered. */
+		ASSERT(0);
+		return dc_fixpt_from_int(0);
+	}
+}
+
+struct dp_audio_layout_config {
+	uint8_t layouts_per_sample_denom;
+	uint8_t symbols_per_layout;
+	uint8_t max_layouts_per_audio_sdp;
+};
+
+static void get_audio_layout_config(
+	uint32_t channel_count,
+	enum dp_link_encoding encoding,
+	struct dp_audio_layout_config *output)
+{
+	/* Assuming L-PCM audio. Current implementation uses max 1 layout per SDP,
+	 * with each layout being the same size (8ch layout).
+	 */
+	if (encoding == DP_8b_10b_ENCODING) {
+		if (channel_count == 2) {
+			output->layouts_per_sample_denom = 4;
+			output->symbols_per_layout = 40;
+			output->max_layouts_per_audio_sdp = 1;
+		} else if (channel_count == 8 || channel_count == 6) {
+			output->layouts_per_sample_denom = 1;
+			output->symbols_per_layout = 40;
+			output->max_layouts_per_audio_sdp = 1;
+		}
+	} else if (encoding == DP_128b_132b_ENCODING) {
+		if (channel_count == 2) {
+			output->layouts_per_sample_denom = 4;
+			output->symbols_per_layout = 10;
+			output->max_layouts_per_audio_sdp = 1;
+		} else if (channel_count == 8 || channel_count == 6) {
+			output->layouts_per_sample_denom = 1;
+			output->symbols_per_layout = 10;
+			output->max_layouts_per_audio_sdp = 1;
+		}
+	}
+}
 
-/*For DP SST, calculate if specified sample rates can fit into a given timing */
-static void check_audio_bandwidth_dpsst(
+static uint32_t get_av_stream_map_lane_count(
+	enum dp_link_encoding encoding,
+	enum dc_lane_count lane_count,
+	bool is_mst)
+{
+	uint32_t av_stream_map_lane_count = 0;
+
+	if (encoding == DP_8b_10b_ENCODING) {
+		if (!is_mst)
+			av_stream_map_lane_count = lane_count;
+		else
+			av_stream_map_lane_count = 4;
+	} else if (encoding == DP_128b_132b_ENCODING) {
+		av_stream_map_lane_count = 4;
+	}
+
+	ASSERT(av_stream_map_lane_count != 0);
+
+	return av_stream_map_lane_count;
+}
+
+static uint32_t get_audio_sdp_overhead(
+	enum dp_link_encoding encoding,
+	enum dc_lane_count lane_count,
+	bool is_mst)
+{
+	uint32_t audio_sdp_overhead = 0;
+
+	if (encoding == DP_8b_10b_ENCODING) {
+		if (is_mst)
+			audio_sdp_overhead = 16; /* 4 * 2 + 8 */
+		else
+			audio_sdp_overhead = lane_count * 2 + 8;
+	} else if (encoding == DP_128b_132b_ENCODING) {
+		audio_sdp_overhead = 10; /* 4 x 2.5 */
+	}
+
+	ASSERT(audio_sdp_overhead != 0);
+
+	return audio_sdp_overhead;
+}
+
+static uint32_t calculate_required_audio_bw_in_symbols(
 	const struct audio_crtc_info *crtc_info,
+	const struct dp_audio_layout_config *layout_config,
 	uint32_t channel_count,
-	union audio_sample_rates *sample_rates)
+	uint32_t sample_rate_hz,
+	uint32_t av_stream_map_lane_count,
+	uint32_t audio_sdp_overhead)
 {
-	/* do nothing */
+	/* DP spec recommends between 1.05 to 1.1 safety margin to prevent sample under-run */
+	struct fixed31_32 audio_sdp_margin = dc_fixpt_from_fraction(110, 100);
+	struct fixed31_32 horizontal_line_freq_khz = dc_fixpt_from_fraction(
+			crtc_info->requested_pixel_clock_100Hz, (long long)crtc_info->h_total * 10);
+	struct fixed31_32 samples_per_line;
+	struct fixed31_32 layouts_per_line;
+	struct fixed31_32 symbols_per_sdp_max_layout;
+	struct fixed31_32 remainder;
+	uint32_t num_sdp_with_max_layouts;
+	uint32_t required_symbols_per_hblank;
+
+	samples_per_line = dc_fixpt_from_fraction(sample_rate_hz, 1000);
+	samples_per_line = dc_fixpt_div(samples_per_line, horizontal_line_freq_khz);
+	layouts_per_line = dc_fixpt_div_int(samples_per_line, layout_config->layouts_per_sample_denom);
+
+	num_sdp_with_max_layouts = dc_fixpt_floor(
+			dc_fixpt_div_int(layouts_per_line, layout_config->max_layouts_per_audio_sdp));
+	symbols_per_sdp_max_layout = dc_fixpt_from_int(
+			layout_config->max_layouts_per_audio_sdp * layout_config->symbols_per_layout);
+	symbols_per_sdp_max_layout = dc_fixpt_add_int(symbols_per_sdp_max_layout, audio_sdp_overhead);
+	symbols_per_sdp_max_layout = dc_fixpt_mul(symbols_per_sdp_max_layout, audio_sdp_margin);
+	required_symbols_per_hblank = num_sdp_with_max_layouts;
+	required_symbols_per_hblank *= ((dc_fixpt_ceil(symbols_per_sdp_max_layout) + av_stream_map_lane_count) /
+			av_stream_map_lane_count) *	av_stream_map_lane_count;
+
+	if (num_sdp_with_max_layouts !=	dc_fixpt_ceil(
+			dc_fixpt_div_int(layouts_per_line, layout_config->max_layouts_per_audio_sdp))) {
+		remainder = dc_fixpt_sub_int(layouts_per_line,
+				num_sdp_with_max_layouts * layout_config->max_layouts_per_audio_sdp);
+		remainder = dc_fixpt_mul_int(remainder, layout_config->symbols_per_layout);
+		remainder = dc_fixpt_add_int(remainder, audio_sdp_overhead);
+		remainder = dc_fixpt_mul(remainder, audio_sdp_margin);
+		required_symbols_per_hblank += ((dc_fixpt_ceil(remainder) + av_stream_map_lane_count) /
+				av_stream_map_lane_count) * av_stream_map_lane_count;
+	}
+
+	return required_symbols_per_hblank;
 }
 
-/*For DP MST, calculate if specified sample rates can fit into a given timing */
-static void check_audio_bandwidth_dpmst(
+/* Current calculation only applicable for 8b/10b MST and 128b/132b SST/MST.
+ */
+static uint32_t calculate_available_hblank_bw_in_symbols(
 	const struct audio_crtc_info *crtc_info,
+	const struct audio_dp_link_info *dp_link_info)
+{
+	uint64_t hblank = crtc_info->h_total - crtc_info->h_active;
+	struct fixed31_32 hblank_time_msec =
+			dc_fixpt_from_fraction(hblank * 10, crtc_info->requested_pixel_clock_100Hz);
+	struct fixed31_32 lsclkfreq_mhz =
+			get_link_symbol_clk_freq_mhz(dp_link_info->link_rate);
+	struct fixed31_32 average_stream_sym_bw_frac;
+	struct fixed31_32 peak_stream_bw_kbps;
+	struct fixed31_32 bits_per_pixel;
+	struct fixed31_32 link_bw_kbps;
+	struct fixed31_32 available_stream_sym_count;
+	uint32_t available_hblank_bw = 0; /* in stream symbols */
+
+	if (crtc_info->dsc_bits_per_pixel) {
+		bits_per_pixel = dc_fixpt_from_fraction(crtc_info->dsc_bits_per_pixel, 16);
+	} else {
+		switch (crtc_info->color_depth) {
+		case COLOR_DEPTH_666:
+			bits_per_pixel = dc_fixpt_from_int(6);
+			break;
+		case COLOR_DEPTH_888:
+			bits_per_pixel = dc_fixpt_from_int(8);
+			break;
+		case COLOR_DEPTH_101010:
+			bits_per_pixel = dc_fixpt_from_int(10);
+			break;
+		case COLOR_DEPTH_121212:
+			bits_per_pixel = dc_fixpt_from_int(12);
+			break;
+		default:
+			/* Default to commonly supported color depth. */
+			bits_per_pixel = dc_fixpt_from_int(8);
+			break;
+		}
+
+		bits_per_pixel = dc_fixpt_mul_int(bits_per_pixel, 3);
+
+		if (crtc_info->pixel_encoding == PIXEL_ENCODING_YCBCR422) {
+			bits_per_pixel = dc_fixpt_div_int(bits_per_pixel, 3);
+			bits_per_pixel = dc_fixpt_mul_int(bits_per_pixel, 2);
+		} else if (crtc_info->pixel_encoding == PIXEL_ENCODING_YCBCR420) {
+			bits_per_pixel = dc_fixpt_div_int(bits_per_pixel, 2);
+		}
+	}
+
+	/* Use simple stream BW calculation because mainlink overhead is
+	 * accounted for separately in the audio BW calculations.
+	 */
+	peak_stream_bw_kbps = dc_fixpt_from_fraction(crtc_info->requested_pixel_clock_100Hz, 10);
+	peak_stream_bw_kbps = dc_fixpt_mul(peak_stream_bw_kbps, bits_per_pixel);
+	link_bw_kbps = dc_fixpt_from_int(dp_link_info->link_bandwidth_kbps);
+	average_stream_sym_bw_frac = dc_fixpt_div(peak_stream_bw_kbps, link_bw_kbps);
+
+	available_stream_sym_count = dc_fixpt_mul_int(hblank_time_msec, 1000);
+	available_stream_sym_count = dc_fixpt_mul(available_stream_sym_count, lsclkfreq_mhz);
+	available_stream_sym_count = dc_fixpt_mul(available_stream_sym_count, average_stream_sym_bw_frac);
+	available_hblank_bw = dc_fixpt_floor(available_stream_sym_count);
+	available_hblank_bw *= dp_link_info->lane_count;
+	available_hblank_bw -= crtc_info->dsc_num_slices * 4; /* EOC overhead */
+
+	if (available_hblank_bw < dp_link_info->hblank_min_symbol_width)
+		/* Each symbol takes 4 frames */
+		available_hblank_bw = 4 * dp_link_info->hblank_min_symbol_width;
+
+	if (available_hblank_bw < 12)
+		available_hblank_bw = 0;
+	else
+		available_hblank_bw -= 12; /* Main link overhead */
+
+	return available_hblank_bw;
+}
+
+static void check_audio_bandwidth_dp(
+	const struct audio_crtc_info *crtc_info,
+	const struct audio_dp_link_info *dp_link_info,
 	uint32_t channel_count,
 	union audio_sample_rates *sample_rates)
 {
-	/* do nothing  */
+	struct dp_audio_layout_config layout_config = {0};
+	uint32_t available_hblank_bw;
+	uint32_t av_stream_map_lane_count;
+	uint32_t audio_sdp_overhead;
+
+	/* TODO: Add validation for SST 8b/10 case  */
+	if (!dp_link_info->is_mst && dp_link_info->encoding == DP_8b_10b_ENCODING)
+		return;
+
+	available_hblank_bw = calculate_available_hblank_bw_in_symbols(
+			crtc_info, dp_link_info);
+	av_stream_map_lane_count = get_av_stream_map_lane_count(
+			dp_link_info->encoding, dp_link_info->lane_count, dp_link_info->is_mst);
+	audio_sdp_overhead = get_audio_sdp_overhead(
+			dp_link_info->encoding, dp_link_info->lane_count, dp_link_info->is_mst);
+	get_audio_layout_config(
+			channel_count, dp_link_info->encoding, &layout_config);
+
+	if (layout_config.max_layouts_per_audio_sdp == 0 ||
+		layout_config.symbols_per_layout == 0 ||
+		layout_config.layouts_per_sample_denom == 0) {
+		return;
+	}
+	if (available_hblank_bw < calculate_required_audio_bw_in_symbols(
+			crtc_info, &layout_config, channel_count, 192000,
+			av_stream_map_lane_count, audio_sdp_overhead))
+		sample_rates->rate.RATE_192 = 0;
+	if (available_hblank_bw < calculate_required_audio_bw_in_symbols(
+			crtc_info, &layout_config, channel_count, 176400,
+			av_stream_map_lane_count, audio_sdp_overhead))
+		sample_rates->rate.RATE_176_4 = 0;
+	if (available_hblank_bw < calculate_required_audio_bw_in_symbols(
+			crtc_info, &layout_config, channel_count, 96000,
+			av_stream_map_lane_count, audio_sdp_overhead))
+		sample_rates->rate.RATE_96 = 0;
+	if (available_hblank_bw < calculate_required_audio_bw_in_symbols(
+			crtc_info, &layout_config, channel_count, 88200,
+			av_stream_map_lane_count, audio_sdp_overhead))
+		sample_rates->rate.RATE_88_2 = 0;
+	if (available_hblank_bw < calculate_required_audio_bw_in_symbols(
+			crtc_info, &layout_config, channel_count, 48000,
+			av_stream_map_lane_count, audio_sdp_overhead))
+		sample_rates->rate.RATE_48 = 0;
+	if (available_hblank_bw < calculate_required_audio_bw_in_symbols(
+			crtc_info, &layout_config, channel_count, 44100,
+			av_stream_map_lane_count, audio_sdp_overhead))
+		sample_rates->rate.RATE_44_1 = 0;
+	if (available_hblank_bw < calculate_required_audio_bw_in_symbols(
+			crtc_info, &layout_config, channel_count, 32000,
+			av_stream_map_lane_count, audio_sdp_overhead))
+		sample_rates->rate.RATE_32 = 0;
 }
 
 static void check_audio_bandwidth(
 	const struct audio_crtc_info *crtc_info,
+	const struct audio_dp_link_info *dp_link_info,
 	uint32_t channel_count,
 	enum signal_type signal,
 	union audio_sample_rates *sample_rates)
@@ -279,12 +539,9 @@ static void check_audio_bandwidth(
 		break;
 	case SIGNAL_TYPE_EDP:
 	case SIGNAL_TYPE_DISPLAY_PORT:
-		check_audio_bandwidth_dpsst(
-			crtc_info, channel_count, sample_rates);
-		break;
 	case SIGNAL_TYPE_DISPLAY_PORT_MST:
-		check_audio_bandwidth_dpmst(
-			crtc_info, channel_count, sample_rates);
+		check_audio_bandwidth_dp(
+			crtc_info, dp_link_info, channel_count, sample_rates);
 		break;
 	default:
 		break;
@@ -308,7 +565,7 @@ static void set_high_bit_rate_capable(
 	AZ_REG_WRITE(AZALIA_F0_CODEC_PIN_CONTROL_RESPONSE_HBR, value);
 }
 
-/* set video latency in in ms/2+1 */
+/* set video latency in ms/2+1 */
 static void set_video_latency(
 	struct audio *audio,
 	int latency_in_ms)
@@ -328,7 +585,7 @@ static void set_video_latency(
 		value);
 }
 
-/* set audio latency in in ms/2+1 */
+/* set audio latency in ms/2+1 */
 static void set_audio_latency(
 	struct audio *audio,
 	int latency_in_ms)
@@ -373,6 +630,11 @@ void dce_aud_az_enable(struct audio *audio)
 			audio->inst, value);
 }
 
+void dce_aud_az_disable_hbr_audio(struct audio *audio)
+{
+	set_high_bit_rate_capable(audio, false);
+}
+
 void dce_aud_az_disable(struct audio *audio)
 {
 	uint32_t value;
@@ -402,7 +664,8 @@ void dce_aud_az_configure(
 	struct audio *audio,
 	enum signal_type signal,
 	const struct audio_crtc_info *crtc_info,
-	const struct audio_info *audio_info)
+	const struct audio_info *audio_info,
+	const struct audio_dp_link_info *dp_link_info)
 {
 	struct dce_audio *aud = DCE_AUD(audio);
 
@@ -415,6 +678,10 @@ void dce_aud_az_configure(
 	bool is_ac3_supported = false;
 	union audio_sample_rates sample_rate;
 	uint32_t strlen = 0;
+
+	if (signal == SIGNAL_TYPE_VIRTUAL)
+		return;
+
 	value = AZ_REG_READ(AZALIA_F0_CODEC_PIN_CONTROL_HOT_PLUG_CONTROL);
 	set_reg_field_value(value, 1,
 			AZALIA_F0_CODEC_PIN_CONTROL_HOT_PLUG_CONTROL,
@@ -492,6 +759,17 @@ void dce_aud_az_configure(
 
 	AZ_REG_WRITE(AZALIA_F0_CODEC_PIN_CONTROL_CHANNEL_SPEAKER, value);
 
+	/*  ACP Data - Supports AI  */
+	value = AZ_REG_READ(AZALIA_F0_CODEC_PIN_CONTROL_ACP_DATA);
+
+	set_reg_field_value(
+		value,
+		audio_info->flags.info.SUPPORT_AI,
+		AZALIA_F0_CODEC_PIN_CONTROL_ACP_DATA,
+		SUPPORTS_AI);
+
+	AZ_REG_WRITE(AZALIA_F0_CODEC_PIN_CONTROL_ACP_DATA, value);
+
 	/*  Audio Descriptors   */
 	/* pass through all formats */
 	for (format_index = 0; format_index < AUDIO_FORMAT_CODE_COUNT;
@@ -522,6 +800,7 @@ void dce_aud_az_configure(
 
 				check_audio_bandwidth(
 					crtc_info,
+					dp_link_info,
 					channel_count,
 					signal,
 					&sample_rates);
@@ -581,6 +860,7 @@ void dce_aud_az_configure(
 
 	check_audio_bandwidth(
 		crtc_info,
+		dp_link_info,
 		8,
 		signal,
 		&sample_rate);
@@ -775,7 +1055,7 @@ static void get_azalia_clock_info_dp(
 	/*audio_dto_module = dpDtoSourceClockInkhz * 10,000;
 	 *  [khz] ->[100Hz] */
 	azalia_clock_info->audio_dto_module =
-		pll_info->dp_dto_source_clock_in_khz * 10;
+		pll_info->audio_dto_source_clock_in_khz * 10;
 }
 
 void dce_aud_wall_dto_setup(
@@ -863,7 +1143,8 @@ void dce_aud_wall_dto_setup(
 		REG_UPDATE(DCCG_AUDIO_DTO1_PHASE,
 				DCCG_AUDIO_DTO1_PHASE, clock_info.audio_dto_phase);
 
-		REG_UPDATE(DCCG_AUDIO_DTO_SOURCE,
+		if (aud->masks->DCCG_AUDIO_DTO2_USE_512FBR_DTO)
+			REG_UPDATE(DCCG_AUDIO_DTO_SOURCE,
 				DCCG_AUDIO_DTO2_USE_512FBR_DTO, 1);
 
 	}
@@ -1018,6 +1299,7 @@ static const struct audio_funcs funcs = {
 	.az_enable = dce_aud_az_enable,
 	.az_disable = dce_aud_az_disable,
 	.az_configure = dce_aud_az_configure,
+	.az_disable_hbr_audio = dce_aud_az_disable_hbr_audio,
 	.destroy = dce_aud_destroy,
 };
 
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_audio.h b/drivers/gpu/drm/amd/display/dc/dce/dce_audio.h
index 5622d5e32d81..1b7b8b079af4 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_audio.h
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_audio.h
@@ -113,6 +113,7 @@ struct dce_audio_shift {
 	uint8_t DCCG_AUDIO_DTO2_USE_512FBR_DTO;
 	uint32_t DCCG_AUDIO_DTO0_USE_512FBR_DTO;
 	uint32_t DCCG_AUDIO_DTO1_USE_512FBR_DTO;
+	uint32_t CLOCK_GATING_DISABLE;
 };
 
 struct dce_audio_mask {
@@ -132,6 +133,7 @@ struct dce_audio_mask {
 	uint32_t DCCG_AUDIO_DTO2_USE_512FBR_DTO;
 	uint32_t DCCG_AUDIO_DTO0_USE_512FBR_DTO;
 	uint32_t DCCG_AUDIO_DTO1_USE_512FBR_DTO;
+	uint32_t CLOCK_GATING_DISABLE;
 
 };
 
@@ -164,11 +166,13 @@ void dce_aud_hw_init(struct audio *audio);
 
 void dce_aud_az_enable(struct audio *audio);
 void dce_aud_az_disable(struct audio *audio);
+void dce_aud_az_disable_hbr_audio(struct audio *audio);
 
 void dce_aud_az_configure(struct audio *audio,
 	enum signal_type signal,
 	const struct audio_crtc_info *crtc_info,
-	const struct audio_info *audio_info);
+	const struct audio_info *audio_info,
+	const struct audio_dp_link_info *dp_link_info);
 
 void dce_aud_wall_dto_setup(struct audio *audio,
 	enum signal_type signal,
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_aux.c b/drivers/gpu/drm/amd/display/dc/dce/dce_aux.c
index 6d42a9cc9916..673bb87d2c17 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_aux.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_aux.c
@@ -23,9 +23,6 @@
  *
  */
 
-#include <linux/delay.h>
-#include <linux/slab.h>
-
 #include "dm_services.h"
 #include "core_types.h"
 #include "dce_aux.h"
@@ -87,7 +84,8 @@ static void release_engine(
 
 	engine->ddc = NULL;
 
-	REG_UPDATE(AUX_ARB_CONTROL, AUX_SW_DONE_USING_AUX_REG, 1);
+	REG_UPDATE_2(AUX_ARB_CONTROL, AUX_SW_DONE_USING_AUX_REG, 1,
+		AUX_SW_USE_AUX_REG_REQ, 0);
 }
 
 #define SW_CAN_ACCESS_AUX 1
@@ -413,7 +411,8 @@ static bool acquire(
 		return false;
 
 	if (!acquire_engine(engine)) {
-		dal_ddc_close(ddc);
+		engine->ddc = ddc;
+		release_engine(engine);
 		return false;
 	}
 
@@ -564,13 +563,16 @@ int dce_aux_transfer_raw(struct ddc_service *ddc,
 	struct ddc *ddc_pin = ddc->ddc_pin;
 	struct dce_aux *aux_engine;
 	struct aux_request_transaction_data aux_req;
-	struct aux_reply_transaction_data aux_rep;
 	uint8_t returned_bytes = 0;
 	int res = -1;
 	uint32_t status;
 
 	memset(&aux_req, 0, sizeof(aux_req));
-	memset(&aux_rep, 0, sizeof(aux_rep));
+
+	if (ddc_pin == NULL) {
+		*operation_result = AUX_RET_ERROR_ENGINE_ACQUIRE;
+		return -1;
+	}
 
 	aux_engine = ddc->ctx->dc->res_pool->engines[ddc_pin->pin_data->en];
 	if (!acquire(aux_engine, ddc_pin)) {
@@ -723,20 +725,34 @@ bool dce_aux_transfer_with_retries(struct ddc_service *ddc,
 	for (i = 0; i < AUX_MAX_RETRIES; i++) {
 		DC_TRACE_LEVEL_MESSAGE(DAL_TRACE_LEVEL_INFORMATION,
 					LOG_FLAG_I2cAux_DceAux,
-					"dce_aux_transfer_with_retries: link_index=%u: START: retry %d of %d: address=0x%04x length=%u write=%d mot=%d",
+					"dce_aux_transfer_with_retries: link_index=%u: START: retry %d of %d: "
+					"address=0x%04x length=%u write=%d mot=%d is_i2c=%d is_dpia=%d ddc_hw_inst=%d",
 					ddc && ddc->link ? ddc->link->link_index : UINT_MAX,
 					i + 1,
 					(int)AUX_MAX_RETRIES,
 					payload->address,
 					payload->length,
 					(unsigned int) payload->write,
-					(unsigned int) payload->mot);
+					(unsigned int) payload->mot,
+					payload->i2c_over_aux,
+					(ddc->link->ep_type == DISPLAY_ENDPOINT_USB4_DPIA) ? true : false,
+					ddc->link->ddc_hw_inst);
 		if (payload->write)
 			dce_aux_log_payload("  write", payload->data, payload->length, 16);
-		ret = dce_aux_transfer_raw(ddc, payload, &operation_result);
+
+		/* Check whether aux to be processed via dmub or dcn directly */
+		if (ddc->ctx->dc->debug.enable_dmub_aux_for_legacy_ddc
+			|| ddc->ddc_pin == NULL) {
+			ret = dce_aux_transfer_dmub_raw(ddc, payload, &operation_result);
+		} else {
+			ret = dce_aux_transfer_raw(ddc, payload, &operation_result);
+		}
+
 		DC_TRACE_LEVEL_MESSAGE(DAL_TRACE_LEVEL_INFORMATION,
 					LOG_FLAG_I2cAux_DceAux,
-					"dce_aux_transfer_with_retries: link_index=%u: END: retry %d of %d: address=0x%04x length=%u write=%d mot=%d: ret=%d operation_result=%d payload->reply=%u",
+					"dce_aux_transfer_with_retries: link_index=%u: END: retry %d of %d: "
+					"address=0x%04x length=%u write=%d mot=%d: ret=%d operation_result=%d "
+					"payload->reply=%u  is_i2c=%d is_dpia=%d ddc_hw_inst=%d",
 					ddc && ddc->link ? ddc->link->link_index : UINT_MAX,
 					i + 1,
 					(int)AUX_MAX_RETRIES,
@@ -746,7 +762,10 @@ bool dce_aux_transfer_with_retries(struct ddc_service *ddc,
 					(unsigned int) payload->mot,
 					ret,
 					(int)operation_result,
-					(unsigned int) *payload->reply);
+					(unsigned int) *payload->reply,
+					payload->i2c_over_aux,
+					(ddc->link->ep_type == DISPLAY_ENDPOINT_USB4_DPIA) ? true : false,
+					ddc->link->ddc_hw_inst);
 		if (!payload->write)
 			dce_aux_log_payload("  read", payload->data, ret > 0 ? ret : 0, 16);
 
@@ -768,7 +787,7 @@ bool dce_aux_transfer_with_retries(struct ddc_service *ddc,
 									aux_defer_retries,
 									AUX_MAX_RETRIES);
 						goto fail;
-					} else 
+					} else
 						udelay(300);
 				} else if (payload->write && ret > 0) {
 					/* sink requested more time to complete the write via AUX_ACKM */
@@ -788,7 +807,6 @@ bool dce_aux_transfer_with_retries(struct ddc_service *ddc,
 					payload->write_status_update = true;
 					payload->length = 0;
 					udelay(300);
-
 				} else
 					return true;
 			break;
@@ -812,12 +830,6 @@ bool dce_aux_transfer_with_retries(struct ddc_service *ddc,
 								"dce_aux_transfer_with_retries: AUX_RET_SUCCESS: AUX_TRANSACTION_REPLY_I2C_OVER_AUX_DEFER");
 
 				retry_on_defer = true;
-				fallthrough;
-			case AUX_TRANSACTION_REPLY_I2C_OVER_AUX_NACK:
-				if (*payload->reply == AUX_TRANSACTION_REPLY_I2C_OVER_AUX_NACK)
-					DC_TRACE_LEVEL_MESSAGE(DAL_TRACE_LEVEL_INFORMATION,
-								LOG_FLAG_I2cAux_DceAux,
-								"dce_aux_transfer_with_retries: AUX_RET_SUCCESS: AUX_TRANSACTION_REPLY_I2C_OVER_AUX_NACK");
 
 				if (aux_defer_retries >= AUX_MIN_DEFER_RETRIES
 						&& defer_time_in_ms >= AUX_MAX_DEFER_TIMEOUT_MS) {
@@ -836,17 +848,16 @@ bool dce_aux_transfer_with_retries(struct ddc_service *ddc,
 									LOG_FLAG_I2cAux_DceAux,
 									"dce_aux_transfer_with_retries: payload->defer_delay=%u",
 									payload->defer_delay);
-						if (payload->defer_delay > 1) {
-							msleep(payload->defer_delay);
-							defer_time_in_ms += payload->defer_delay;
-						} else if (payload->defer_delay <= 1) {
-							udelay(payload->defer_delay * 1000);
-							defer_time_in_ms += payload->defer_delay;
-						}
+						fsleep(payload->defer_delay * 1000);
+						defer_time_in_ms += payload->defer_delay;
 					}
 				}
 				break;
-
+			case AUX_TRANSACTION_REPLY_I2C_OVER_AUX_NACK:
+				DC_TRACE_LEVEL_MESSAGE(DAL_TRACE_LEVEL_INFORMATION,
+							LOG_FLAG_I2cAux_DceAux,
+							"dce_aux_transfer_with_retries: FAILURE: AUX_TRANSACTION_REPLY_I2C_OVER_AUX_NACK");
+				goto fail;
 			case AUX_TRANSACTION_REPLY_I2C_DEFER:
 				DC_TRACE_LEVEL_MESSAGE(DAL_TRACE_LEVEL_INFORMATION,
 							LOG_FLAG_I2cAux_DceAux,
@@ -878,7 +889,7 @@ bool dce_aux_transfer_with_retries(struct ddc_service *ddc,
 			default:
 				DC_TRACE_LEVEL_MESSAGE(DAL_TRACE_LEVEL_ERROR,
 							LOG_FLAG_Error_I2cAux,
-							"dce_aux_transfer_with_retries: AUX_RET_SUCCESS: FAILURE: AUX_TRANSACTION_REPLY_* unknown, default case.");
+							"dce_aux_transfer_with_retries: AUX_RET_SUCCESS: FAILURE: AUX_TRANSACTION_REPLY_* unknown, default case. Reply: %d", *payload->reply);
 				goto fail;
 			}
 			break;
@@ -942,10 +953,6 @@ bool dce_aux_transfer_with_retries(struct ddc_service *ddc,
 		case AUX_RET_ERROR_ENGINE_ACQUIRE:
 		case AUX_RET_ERROR_UNKNOWN:
 		default:
-			DC_TRACE_LEVEL_MESSAGE(DAL_TRACE_LEVEL_INFORMATION,
-						LOG_FLAG_I2cAux_DceAux,
-						"dce_aux_transfer_with_retries: Failure: operation_result=%d",
-						(int)operation_result);
 			goto fail;
 		}
 	}
@@ -953,14 +960,11 @@ bool dce_aux_transfer_with_retries(struct ddc_service *ddc,
 fail:
 	DC_TRACE_LEVEL_MESSAGE(DAL_TRACE_LEVEL_ERROR,
 				LOG_FLAG_Error_I2cAux,
-				"dce_aux_transfer_with_retries: FAILURE");
+				"%s: Failure: operation_result=%d",
+				__func__,
+				(int)operation_result);
 	if (!payload_reply)
 		payload->reply = NULL;
 
-	DC_TRACE_LEVEL_MESSAGE(DAL_TRACE_LEVEL_ERROR,
-				WPP_BIT_FLAG_DC_ERROR,
-				"AUX transaction failed. Result: %d",
-				operation_result);
-
 	return false;
 }
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_aux.h b/drivers/gpu/drm/amd/display/dc/dce/dce_aux.h
index e69f1899fbf0..c850ed49281f 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_aux.h
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_aux.h
@@ -26,7 +26,7 @@
 #ifndef __DAL_AUX_ENGINE_DCE110_H__
 #define __DAL_AUX_ENGINE_DCE110_H__
 
-#include "i2caux_interface.h"
+#include "gpio_service_interface.h"
 #include "inc/hw/aux_engine.h"
 
 enum aux_return_code_type;
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/dce/dce_clk_mgr.c
index 1435d7bc1f21..e7acd6eec1fd 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_clk_mgr.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_clk_mgr.c
@@ -450,6 +450,8 @@ void dce_clock_read_ss_info(struct dce_clk_mgr *clk_mgr_dce)
 				clk_mgr_dce->dprefclk_ss_percentage =
 						info.spread_spectrum_percentage;
 			}
+			if (clk_mgr_dce->base.ctx->dc->debug.ignore_dpref_ss)
+				clk_mgr_dce->dprefclk_ss_percentage = 0;
 		}
 	}
 }
@@ -638,7 +640,7 @@ static void dce11_pplib_apply_display_requirements(
 	 * on power saving.
 	 *
 	 */
-	pp_display_cfg->min_dcfclock_khz = (context->stream_count > 4)?
+	pp_display_cfg->min_dcfclock_khz = (context->stream_count > 4) ?
 			pp_display_cfg->min_engine_clock_khz : 0;
 
 	pp_display_cfg->min_engine_clock_deep_sleep_khz
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_clock_source.c b/drivers/gpu/drm/amd/display/dc/dce/dce_clock_source.c
index 2c7eb982eabc..b4f5b4a6331a 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_clock_source.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_clock_source.c
@@ -23,8 +23,6 @@
  *
  */
 
-#include <linux/slab.h>
-
 #include "dm_services.h"
 
 
@@ -36,6 +34,7 @@
 
 #include "dce_clock_source.h"
 #include "clk_mgr.h"
+#include "dccg.h"
 
 #include "reg_helper.h"
 
@@ -45,7 +44,10 @@
 #define CTX \
 	clk_src->base.ctx
 
-#define DC_LOGGER_INIT()
+#define DC_LOGGER \
+	calc_pll_cs->ctx->logger
+#define DC_LOGGER_INIT() \
+	struct calc_pll_clock_source *calc_pll_cs = &clk_src->calc_pll
 
 #undef FN
 #define FN(reg_name, field_name) \
@@ -215,7 +217,7 @@ static bool calc_fb_divider_checking_tolerance(
 	actual_calc_clk_100hz = (uint64_t)feedback_divider *
 					calc_pll_cs->fract_fb_divider_factor +
 							fract_feedback_divider;
-	actual_calc_clk_100hz *= calc_pll_cs->ref_freq_khz * 10;
+	actual_calc_clk_100hz *= (uint64_t)calc_pll_cs->ref_freq_khz * 10;
 	actual_calc_clk_100hz =
 		div_u64(actual_calc_clk_100hz,
 			ref_divider * post_divider *
@@ -545,9 +547,11 @@ static void dce112_get_pix_clk_dividers_helper (
 		switch (pix_clk_params->color_depth) {
 		case COLOR_DEPTH_101010:
 			actual_pixel_clock_100hz = (actual_pixel_clock_100hz * 5) >> 2;
+			actual_pixel_clock_100hz -= actual_pixel_clock_100hz % 10;
 			break;
 		case COLOR_DEPTH_121212:
 			actual_pixel_clock_100hz = (actual_pixel_clock_100hz * 6) >> 2;
+			actual_pixel_clock_100hz -= actual_pixel_clock_100hz % 10;
 			break;
 		case COLOR_DEPTH_161616:
 			actual_pixel_clock_100hz = actual_pixel_clock_100hz * 2;
@@ -676,7 +680,7 @@ static bool calculate_ss(
 	 * so have to divided by 100 * 100*/
 	ss_amount = dc_fixpt_mul(
 		fb_div, dc_fixpt_from_fraction(ss_data->percentage,
-					100 * ss_data->percentage_divider));
+					100 * (long long)ss_data->percentage_divider));
 	ds_data->feedback_amount = dc_fixpt_floor(ss_amount);
 
 	ss_nslip_amount = dc_fixpt_sub(ss_amount,
@@ -691,8 +695,8 @@ static bool calculate_ss(
 
 	/* compute SS_STEP_SIZE_DSFRAC */
 	modulation_time = dc_fixpt_from_fraction(
-		pll_settings->reference_freq * 1000,
-		pll_settings->reference_divider * ss_data->modulation_freq_hz);
+		pll_settings->reference_freq * (uint64_t)1000,
+		pll_settings->reference_divider * (uint64_t)ss_data->modulation_freq_hz);
 
 	if (ss_data->flags.CENTER_SPREAD)
 		modulation_time = dc_fixpt_div_int(modulation_time, 4);
@@ -840,6 +844,7 @@ static void dce112_program_pixel_clk_resync(
 static bool dce110_program_pix_clk(
 		struct clock_source *clock_source,
 		struct pixel_clk_params *pix_clk_params,
+		enum dp_link_encoding encoding,
 		struct pll_settings *pll_settings)
 {
 	struct dce110_clk_src *clk_src = TO_DCE110_CLK_SRC(clock_source);
@@ -913,26 +918,12 @@ static bool dce110_program_pix_clk(
 static bool dce112_program_pix_clk(
 		struct clock_source *clock_source,
 		struct pixel_clk_params *pix_clk_params,
+		enum dp_link_encoding encoding,
 		struct pll_settings *pll_settings)
 {
 	struct dce110_clk_src *clk_src = TO_DCE110_CLK_SRC(clock_source);
 	struct bp_pixel_clock_parameters bp_pc_params = {0};
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	if (IS_FPGA_MAXIMUS_DC(clock_source->ctx->dce_environment)) {
-		unsigned int inst = pix_clk_params->controller_id - CONTROLLER_ID_D0;
-		unsigned dp_dto_ref_100hz = 7000000;
-		unsigned clock_100hz = pll_settings->actual_pix_clk_100hz;
-
-		/* Set DTO values: phase = target clock, modulo = reference clock */
-		REG_WRITE(PHASE[inst], clock_100hz);
-		REG_WRITE(MODULO[inst], dp_dto_ref_100hz);
-
-		/* Enable DTO */
-		REG_UPDATE(PIXEL_RATE_CNTL[inst], DP_DTO0_ENABLE, 1);
-		return true;
-	}
-#endif
 	/* First disable SS
 	 * ATOMBIOS will enable by default SS on PLL for DP,
 	 * do not disable it here
@@ -971,6 +962,209 @@ static bool dce112_program_pix_clk(
 	return true;
 }
 
+static bool dcn31_program_pix_clk(
+		struct clock_source *clock_source,
+		struct pixel_clk_params *pix_clk_params,
+		enum dp_link_encoding encoding,
+		struct pll_settings *pll_settings)
+{
+	struct dce110_clk_src *clk_src = TO_DCE110_CLK_SRC(clock_source);
+	unsigned int inst = pix_clk_params->controller_id - CONTROLLER_ID_D0;
+	unsigned int dp_dto_ref_khz = clock_source->ctx->dc->clk_mgr->dprefclk_khz;
+	const struct pixel_rate_range_table_entry *e =
+			look_up_in_video_optimized_rate_tlb(pix_clk_params->requested_pix_clk_100hz / 10);
+	struct bp_pixel_clock_parameters bp_pc_params = {0};
+	enum transmitter_color_depth bp_pc_colour_depth = TRANSMITTER_COLOR_DEPTH_24;
+
+	// Apply ssed(spread spectrum) dpref clock for edp and dp
+	if (clock_source->ctx->dc->clk_mgr->dp_dto_source_clock_in_khz != 0 &&
+		dc_is_dp_signal(pix_clk_params->signal_type) &&
+		encoding == DP_8b_10b_ENCODING)
+		dp_dto_ref_khz = clock_source->ctx->dc->clk_mgr->dp_dto_source_clock_in_khz;
+
+	// For these signal types Driver to program DP_DTO without calling VBIOS Command table
+	if (dc_is_dp_signal(pix_clk_params->signal_type) || dc_is_virtual_signal(pix_clk_params->signal_type)) {
+		if (e) {
+			/* Set DTO values: phase = target clock, modulo = reference clock*/
+			REG_WRITE(PHASE[inst], e->target_pixel_rate_khz * e->mult_factor);
+			REG_WRITE(MODULO[inst], dp_dto_ref_khz * e->div_factor);
+		} else {
+			/* Set DTO values: phase = target clock, modulo = reference clock*/
+			REG_WRITE(PHASE[inst], pll_settings->actual_pix_clk_100hz * 100);
+			REG_WRITE(MODULO[inst], dp_dto_ref_khz * 1000);
+		}
+		/* Enable DTO */
+		if (clk_src->cs_mask->PIPE0_DTO_SRC_SEL)
+			if (encoding == DP_128b_132b_ENCODING)
+				REG_UPDATE_2(PIXEL_RATE_CNTL[inst],
+						DP_DTO0_ENABLE, 1,
+						PIPE0_DTO_SRC_SEL, 2);
+			else
+				REG_UPDATE_2(PIXEL_RATE_CNTL[inst],
+						DP_DTO0_ENABLE, 1,
+						PIPE0_DTO_SRC_SEL, 1);
+		else
+			REG_UPDATE(PIXEL_RATE_CNTL[inst],
+					DP_DTO0_ENABLE, 1);
+	} else {
+
+		if (clk_src->cs_mask->PIPE0_DTO_SRC_SEL)
+			REG_UPDATE(PIXEL_RATE_CNTL[inst],
+					PIPE0_DTO_SRC_SEL, 0);
+
+		/*ATOMBIOS expects pixel rate adjusted by deep color ratio)*/
+		bp_pc_params.controller_id = pix_clk_params->controller_id;
+		bp_pc_params.pll_id = clock_source->id;
+		bp_pc_params.target_pixel_clock_100hz = pll_settings->actual_pix_clk_100hz;
+		bp_pc_params.encoder_object_id = pix_clk_params->encoder_object_id;
+		bp_pc_params.signal_type = pix_clk_params->signal_type;
+
+		// Make sure we send the correct color depth to DMUB for HDMI
+		if (pix_clk_params->signal_type == SIGNAL_TYPE_HDMI_TYPE_A) {
+			switch (pix_clk_params->color_depth) {
+			case COLOR_DEPTH_888:
+				bp_pc_colour_depth = TRANSMITTER_COLOR_DEPTH_24;
+				break;
+			case COLOR_DEPTH_101010:
+				bp_pc_colour_depth = TRANSMITTER_COLOR_DEPTH_30;
+				break;
+			case COLOR_DEPTH_121212:
+				bp_pc_colour_depth = TRANSMITTER_COLOR_DEPTH_36;
+				break;
+			case COLOR_DEPTH_161616:
+				bp_pc_colour_depth = TRANSMITTER_COLOR_DEPTH_48;
+				break;
+			default:
+				bp_pc_colour_depth = TRANSMITTER_COLOR_DEPTH_24;
+				break;
+			}
+			bp_pc_params.color_depth = bp_pc_colour_depth;
+		}
+
+		if (clock_source->id != CLOCK_SOURCE_ID_DP_DTO) {
+			bp_pc_params.flags.SET_GENLOCK_REF_DIV_SRC =
+							pll_settings->use_external_clk;
+			bp_pc_params.flags.SET_XTALIN_REF_SRC =
+							!pll_settings->use_external_clk;
+			if (pix_clk_params->flags.SUPPORT_YCBCR420) {
+				bp_pc_params.flags.SUPPORT_YUV_420 = 1;
+			}
+		}
+		if (clk_src->bios->funcs->set_pixel_clock(
+				clk_src->bios, &bp_pc_params) != BP_RESULT_OK)
+			return false;
+		/* Resync deep color DTO */
+		if (clock_source->id != CLOCK_SOURCE_ID_DP_DTO)
+			dce112_program_pixel_clk_resync(clk_src,
+						pix_clk_params->signal_type,
+						pix_clk_params->color_depth,
+						pix_clk_params->flags.SUPPORT_YCBCR420);
+	}
+
+	return true;
+}
+
+static bool dcn401_program_pix_clk(
+		struct clock_source *clock_source,
+		struct pixel_clk_params *pix_clk_params,
+		enum dp_link_encoding encoding,
+		struct pll_settings *pll_settings)
+{
+	struct dce110_clk_src *clk_src = TO_DCE110_CLK_SRC(clock_source);
+	unsigned int inst = pix_clk_params->controller_id - CONTROLLER_ID_D0;
+	const struct pixel_rate_range_table_entry *e =
+			look_up_in_video_optimized_rate_tlb(pix_clk_params->requested_pix_clk_100hz / 10);
+	struct bp_pixel_clock_parameters bp_pc_params = {0};
+	enum transmitter_color_depth bp_pc_colour_depth = TRANSMITTER_COLOR_DEPTH_24;
+	struct dp_dto_params dto_params = { 0 };
+
+	dto_params.otg_inst = inst;
+	dto_params.signal = pix_clk_params->signal_type;
+
+	// all but TMDS gets Driver to program DP_DTO without calling VBIOS Command table
+	if (!dc_is_tmds_signal(pix_clk_params->signal_type)) {
+		long long dtbclk_p_src_clk_khz;
+
+		dtbclk_p_src_clk_khz = clock_source->ctx->dc->clk_mgr->dprefclk_khz;
+		dto_params.clk_src = DPREFCLK;
+
+		if (e) {
+			dto_params.pixclk_hz = e->target_pixel_rate_khz;
+			dto_params.pixclk_hz *= e->mult_factor;
+			dto_params.refclk_hz = dtbclk_p_src_clk_khz;
+			dto_params.refclk_hz *= e->div_factor;
+		} else {
+			dto_params.pixclk_hz = pix_clk_params->requested_pix_clk_100hz;
+			dto_params.pixclk_hz *= 100;
+			dto_params.refclk_hz = dtbclk_p_src_clk_khz;
+			dto_params.refclk_hz *= 1000;
+		}
+
+		/* enable DP DTO */
+		clock_source->ctx->dc->res_pool->dccg->funcs->set_dp_dto(
+				clock_source->ctx->dc->res_pool->dccg,
+				&dto_params);
+
+	} else {
+		if (pll_settings->actual_pix_clk_100hz > 6000000UL)
+			return false;
+
+		/* disables DP DTO when provided with TMDS signal type */
+		clock_source->ctx->dc->res_pool->dccg->funcs->set_dp_dto(
+				clock_source->ctx->dc->res_pool->dccg,
+				&dto_params);
+
+		/*ATOMBIOS expects pixel rate adjusted by deep color ratio)*/
+		bp_pc_params.controller_id = pix_clk_params->controller_id;
+		bp_pc_params.pll_id = clock_source->id;
+		bp_pc_params.target_pixel_clock_100hz = pll_settings->actual_pix_clk_100hz;
+		bp_pc_params.encoder_object_id = pix_clk_params->encoder_object_id;
+		bp_pc_params.signal_type = pix_clk_params->signal_type;
+
+		// Make sure we send the correct color depth to DMUB for HDMI
+		if (pix_clk_params->signal_type == SIGNAL_TYPE_HDMI_TYPE_A) {
+			switch (pix_clk_params->color_depth) {
+			case COLOR_DEPTH_888:
+				bp_pc_colour_depth = TRANSMITTER_COLOR_DEPTH_24;
+				break;
+			case COLOR_DEPTH_101010:
+				bp_pc_colour_depth = TRANSMITTER_COLOR_DEPTH_30;
+				break;
+			case COLOR_DEPTH_121212:
+				bp_pc_colour_depth = TRANSMITTER_COLOR_DEPTH_36;
+				break;
+			case COLOR_DEPTH_161616:
+				bp_pc_colour_depth = TRANSMITTER_COLOR_DEPTH_48;
+				break;
+			default:
+				bp_pc_colour_depth = TRANSMITTER_COLOR_DEPTH_24;
+				break;
+			}
+			bp_pc_params.color_depth = bp_pc_colour_depth;
+		}
+
+		if (clock_source->id != CLOCK_SOURCE_ID_DP_DTO) {
+			bp_pc_params.flags.SET_GENLOCK_REF_DIV_SRC =
+							pll_settings->use_external_clk;
+			bp_pc_params.flags.SET_XTALIN_REF_SRC =
+							!pll_settings->use_external_clk;
+			if (pix_clk_params->flags.SUPPORT_YCBCR420) {
+				bp_pc_params.flags.SUPPORT_YUV_420 = 1;
+			}
+		}
+		if (clk_src->bios->funcs->set_pixel_clock(
+				clk_src->bios, &bp_pc_params) != BP_RESULT_OK)
+			return false;
+		/* Resync deep color DTO */
+		if (clock_source->id != CLOCK_SOURCE_ID_DP_DTO)
+			dce112_program_pixel_clk_resync(clk_src,
+						pix_clk_params->signal_type,
+						pix_clk_params->color_depth,
+						pix_clk_params->flags.SUPPORT_YCBCR420);
+	}
+
+	return true;
+}
 
 static bool dce110_clock_source_power_down(
 		struct clock_source *clk_src)
@@ -1003,6 +1197,7 @@ static bool get_pixel_clk_frequency_100hz(
 	struct dce110_clk_src *clk_src = TO_DCE110_CLK_SRC(clock_source);
 	unsigned int clock_hz = 0;
 	unsigned int modulo_hz = 0;
+	unsigned int dp_dto_ref_khz = clock_source->ctx->dc->clk_mgr->dprefclk_khz;
 
 	if (clock_source->id == CLOCK_SOURCE_ID_DP_DTO) {
 		clock_hz = REG_READ(PHASE[inst]);
@@ -1013,9 +1208,12 @@ static bool get_pixel_clk_frequency_100hz(
 			 * not be programmed equal to DPREFCLK
 			 */
 			modulo_hz = REG_READ(MODULO[inst]);
-			*pixel_clk_khz = div_u64((uint64_t)clock_hz*
-				clock_source->ctx->dc->clk_mgr->dprefclk_khz*10,
-				modulo_hz);
+			if (modulo_hz)
+				*pixel_clk_khz = div_u64((uint64_t)clock_hz*
+					dp_dto_ref_khz*10,
+					modulo_hz);
+			else
+				*pixel_clk_khz = 0;
 		} else {
 			/* NOTE: There is agreement with VBIOS here that MODULO is
 			 * programmed equal to DPREFCLK, in which case PHASE will be
@@ -1029,13 +1227,13 @@ static bool get_pixel_clk_frequency_100hz(
 	return false;
 }
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 /* this table is use to find *1.001 and /1.001 pixel rates from non-precise pixel rate */
 const struct pixel_rate_range_table_entry video_optimized_pixel_rates[] = {
 	// /1.001 rates
 	{25170, 25180, 25200, 1000, 1001},	//25.2MHz   ->   25.17
 	{59340, 59350, 59400, 1000, 1001},	//59.4Mhz   ->   59.340
 	{74170, 74180, 74250, 1000, 1001},	//74.25Mhz  ->   74.1758
+	{89910, 90000, 90000, 1000, 1001},	//90Mhz     ->   89.91
 	{125870, 125880, 126000, 1000, 1001},	//126Mhz    ->  125.87
 	{148350, 148360, 148500, 1000, 1001},	//148.5Mhz  ->  148.3516
 	{167830, 167840, 168000, 1000, 1001},	//168Mhz    ->  167.83
@@ -1079,17 +1277,17 @@ const struct pixel_rate_range_table_entry *look_up_in_video_optimized_rate_tlb(
 
 	return NULL;
 }
-#endif
 
 static bool dcn20_program_pix_clk(
 		struct clock_source *clock_source,
 		struct pixel_clk_params *pix_clk_params,
+		enum dp_link_encoding encoding,
 		struct pll_settings *pll_settings)
 {
 	struct dce110_clk_src *clk_src = TO_DCE110_CLK_SRC(clock_source);
 	unsigned int inst = pix_clk_params->controller_id - CONTROLLER_ID_D0;
 
-	dce112_program_pix_clk(clock_source, pix_clk_params, pll_settings);
+	dce112_program_pix_clk(clock_source, pix_clk_params, encoding, pll_settings);
 
 	if (clock_source->ctx->dc->hwss.enable_vblanks_synchronization &&
 			clock_source->ctx->dc->config.vblank_alignment_max_frame_time_diff > 0) {
@@ -1126,10 +1324,10 @@ static const struct clock_source_funcs dcn20_clk_src_funcs = {
 	.override_dp_pix_clk = dcn20_override_dp_pix_clk
 };
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 static bool dcn3_program_pix_clk(
 		struct clock_source *clock_source,
 		struct pixel_clk_params *pix_clk_params,
+		enum dp_link_encoding encoding,
 		struct pll_settings *pll_settings)
 {
 	struct dce110_clk_src *clk_src = TO_DCE110_CLK_SRC(clock_source);
@@ -1149,10 +1347,17 @@ static bool dcn3_program_pix_clk(
 			REG_WRITE(PHASE[inst], pll_settings->actual_pix_clk_100hz * 100);
 			REG_WRITE(MODULO[inst], dp_dto_ref_khz * 1000);
 		}
-		REG_UPDATE(PIXEL_RATE_CNTL[inst], DP_DTO0_ENABLE, 1);
+		/* Enable DTO */
+		if (clk_src->cs_mask->PIPE0_DTO_SRC_SEL)
+			REG_UPDATE_2(PIXEL_RATE_CNTL[inst],
+					DP_DTO0_ENABLE, 1,
+					PIPE0_DTO_SRC_SEL, 1);
+		else
+			REG_UPDATE(PIXEL_RATE_CNTL[inst],
+					DP_DTO0_ENABLE, 1);
 	} else
 		// For other signal types(HDMI_TYPE_A, DVI) Driver still to call VBIOS Command table
-		dce112_program_pix_clk(clock_source, pix_clk_params, pll_settings);
+		dce112_program_pix_clk(clock_source, pix_clk_params, encoding, pll_settings);
 
 	return true;
 }
@@ -1162,10 +1367,9 @@ static uint32_t dcn3_get_pix_clk_dividers(
 		struct pixel_clk_params *pix_clk_params,
 		struct pll_settings *pll_settings)
 {
-	unsigned long long actual_pix_clk_100Hz = pix_clk_params->requested_pix_clk_100hz;
-	struct dce110_clk_src *clk_src;
+	unsigned long long actual_pix_clk_100Hz = pix_clk_params ? pix_clk_params->requested_pix_clk_100hz : 0;
+	struct dce110_clk_src *clk_src = TO_DCE110_CLK_SRC(cs);
 
-	clk_src = TO_DCE110_CLK_SRC(cs);
 	DC_LOGGER_INIT();
 
 	if (pix_clk_params == NULL || pll_settings == NULL
@@ -1205,7 +1409,21 @@ static const struct clock_source_funcs dcn3_clk_src_funcs = {
 	.get_pix_clk_dividers = dcn3_get_pix_clk_dividers,
 	.get_pixel_clk_frequency_100hz = get_pixel_clk_frequency_100hz
 };
-#endif
+
+static const struct clock_source_funcs dcn31_clk_src_funcs = {
+	.cs_power_down = dce110_clock_source_power_down,
+	.program_pix_clk = dcn31_program_pix_clk,
+	.get_pix_clk_dividers = dcn3_get_pix_clk_dividers,
+	.get_pixel_clk_frequency_100hz = get_pixel_clk_frequency_100hz
+};
+
+static const struct clock_source_funcs dcn401_clk_src_funcs = {
+	.cs_power_down = dce110_clock_source_power_down,
+	.program_pix_clk = dcn401_program_pix_clk,
+	.get_pix_clk_dividers = dcn3_get_pix_clk_dividers,
+	.get_pixel_clk_frequency_100hz = get_pixel_clk_frequency_100hz
+};
+
 /*****************************************/
 /* Constructor                           */
 /*****************************************/
@@ -1591,7 +1809,6 @@ bool dcn20_clk_src_construct(
 	return ret;
 }
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 bool dcn3_clk_src_construct(
 	struct dce110_clk_src *clk_src,
 	struct dc_context *ctx,
@@ -1607,9 +1824,38 @@ bool dcn3_clk_src_construct(
 
 	return ret;
 }
-#endif
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
+bool dcn31_clk_src_construct(
+	struct dce110_clk_src *clk_src,
+	struct dc_context *ctx,
+	struct dc_bios *bios,
+	enum clock_source_id id,
+	const struct dce110_clk_src_regs *regs,
+	const struct dce110_clk_src_shift *cs_shift,
+	const struct dce110_clk_src_mask *cs_mask)
+{
+	bool ret = dce112_clk_src_construct(clk_src, ctx, bios, id, regs, cs_shift, cs_mask);
+
+	clk_src->base.funcs = &dcn31_clk_src_funcs;
+
+	return ret;
+}
+
+bool dcn401_clk_src_construct(
+	struct dce110_clk_src *clk_src,
+	struct dc_context *ctx,
+	struct dc_bios *bios,
+	enum clock_source_id id,
+	const struct dce110_clk_src_regs *regs,
+	const struct dce110_clk_src_shift *cs_shift,
+	const struct dce110_clk_src_mask *cs_mask)
+{
+	bool ret = dce112_clk_src_construct(clk_src, ctx, bios, id, regs, cs_shift, cs_mask);
+
+	clk_src->base.funcs = &dcn401_clk_src_funcs;
+
+	return ret;
+}
 bool dcn301_clk_src_construct(
 	struct dce110_clk_src *clk_src,
 	struct dc_context *ctx,
@@ -1625,4 +1871,3 @@ bool dcn301_clk_src_construct(
 
 	return ret;
 }
-#endif
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_clock_source.h b/drivers/gpu/drm/amd/display/dc/dce/dce_clock_source.h
index 692fa23ca02b..94128f7a18b1 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_clock_source.h
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_clock_source.h
@@ -100,7 +100,6 @@
 		SRII(PIXEL_RATE_CNTL, OTG, 2),\
 		SRII(PIXEL_RATE_CNTL, OTG, 3)
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 #define CS_COMMON_REG_LIST_DCN3_0(index, pllid) \
 		SRI(PIXCLK_RESYNC_CNTL, PHYPLL, pllid),\
 		SRII(PHASE, DP_DTO, 0),\
@@ -130,9 +129,7 @@
 		SRII(PIXEL_RATE_CNTL, OTG, 1),\
 		SRII(PIXEL_RATE_CNTL, OTG, 2),\
 		SRII(PIXEL_RATE_CNTL, OTG, 3)
-#endif
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 #define CS_COMMON_REG_LIST_DCN3_02(index, pllid) \
 		SRI(PIXCLK_RESYNC_CNTL, PHYPLL, pllid),\
 		SRII(PHASE, DP_DTO, 0),\
@@ -160,14 +157,19 @@
 		SRII(PIXEL_RATE_CNTL, OTG, 0),\
 		SRII(PIXEL_RATE_CNTL, OTG, 1)
 
-#endif
 #define CS_COMMON_MASK_SH_LIST_DCN2_0(mask_sh)\
 	CS_SF(DP_DTO0_PHASE, DP_DTO0_PHASE, mask_sh),\
 	CS_SF(DP_DTO0_MODULO, DP_DTO0_MODULO, mask_sh),\
 	CS_SF(PHYPLLA_PIXCLK_RESYNC_CNTL, PHYPLLA_DCCG_DEEP_COLOR_CNTL, mask_sh),\
 	CS_SF(OTG0_PIXEL_RATE_CNTL, DP_DTO0_ENABLE, mask_sh)
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
+#define CS_COMMON_MASK_SH_LIST_DCN3_1_4(mask_sh)\
+	CS_COMMON_MASK_SH_LIST_DCN2_0(mask_sh),\
+	CS_SF(OTG0_PIXEL_RATE_CNTL, PIPE0_DTO_SRC_SEL, mask_sh),
+
+#define CS_COMMON_MASK_SH_LIST_DCN3_2(mask_sh)\
+	CS_COMMON_MASK_SH_LIST_DCN2_0(mask_sh),\
+	CS_SF(OTG0_PIXEL_RATE_CNTL, PIPE0_DTO_SRC_SEL, mask_sh)
 
 #define CS_COMMON_REG_LIST_DCN1_0(index, pllid) \
 		SRI(PIXCLK_RESYNC_CNTL, PHYPLL, pllid),\
@@ -190,7 +192,6 @@
 	CS_SF(PHYPLLA_PIXCLK_RESYNC_CNTL, PHYPLLA_DCCG_DEEP_COLOR_CNTL, mask_sh),\
 	CS_SF(OTG0_PIXEL_RATE_CNTL, DP_DTO0_ENABLE, mask_sh)
 
-#endif
 
 #define CS_REG_FIELD_LIST(type) \
 	type PLL_REF_DIV_SRC; \
@@ -203,12 +204,17 @@
 	type DP_DTO0_MODULO; \
 	type DP_DTO0_ENABLE;
 
+#define CS_REG_FIELD_LIST_DCN32(type) \
+	type PIPE0_DTO_SRC_SEL;
+
 struct dce110_clk_src_shift {
 	CS_REG_FIELD_LIST(uint8_t)
+	CS_REG_FIELD_LIST_DCN32(uint8_t)
 };
 
 struct dce110_clk_src_mask{
 	CS_REG_FIELD_LIST(uint32_t)
+	CS_REG_FIELD_LIST_DCN32(uint32_t)
 };
 
 struct dce110_clk_src_regs {
@@ -251,7 +257,7 @@ bool dce110_clk_src_construct(
 	struct dce110_clk_src *clk_src,
 	struct dc_context *ctx,
 	struct dc_bios *bios,
-	enum clock_source_id,
+	enum clock_source_id id,
 	const struct dce110_clk_src_regs *regs,
 	const struct dce110_clk_src_shift *cs_shift,
 	const struct dce110_clk_src_mask *cs_mask);
@@ -274,7 +280,6 @@ bool dcn20_clk_src_construct(
 	const struct dce110_clk_src_shift *cs_shift,
 	const struct dce110_clk_src_mask *cs_mask);
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 bool dcn3_clk_src_construct(
 	struct dce110_clk_src *clk_src,
 	struct dc_context *ctx,
@@ -292,8 +297,24 @@ bool dcn301_clk_src_construct(
 	const struct dce110_clk_src_regs *regs,
 	const struct dce110_clk_src_shift *cs_shift,
 	const struct dce110_clk_src_mask *cs_mask);
-#endif
 
+bool dcn31_clk_src_construct(
+	struct dce110_clk_src *clk_src,
+	struct dc_context *ctx,
+	struct dc_bios *bios,
+	enum clock_source_id id,
+	const struct dce110_clk_src_regs *regs,
+	const struct dce110_clk_src_shift *cs_shift,
+	const struct dce110_clk_src_mask *cs_mask);
+
+bool dcn401_clk_src_construct(
+	struct dce110_clk_src *clk_src,
+	struct dc_context *ctx,
+	struct dc_bios *bios,
+	enum clock_source_id id,
+	const struct dce110_clk_src_regs *regs,
+	const struct dce110_clk_src_shift *cs_shift,
+	const struct dce110_clk_src_mask *cs_mask);
 /* this table is use to find *1.001 and /1.001 pixel rates from non-precise pixel rate */
 struct pixel_rate_range_table_entry {
 	unsigned int range_min_khz;
@@ -303,10 +324,8 @@ struct pixel_rate_range_table_entry {
 	unsigned short div_factor;
 };
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 extern const struct pixel_rate_range_table_entry video_optimized_pixel_rates[];
 const struct pixel_rate_range_table_entry *look_up_in_video_optimized_rate_tlb(
 		unsigned int pixel_rate_khz);
-#endif
 
 #endif
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_dmcu.c b/drivers/gpu/drm/amd/display/dc/dce/dce_dmcu.c
index 8cd841320ded..5f8fba45d98d 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_dmcu.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_dmcu.c
@@ -23,9 +23,6 @@
  *
  */
 
-#include <linux/delay.h>
-#include <linux/slab.h>
-
 #include "core_types.h"
 #include "link_encoder.h"
 #include "dce_dmcu.h"
@@ -70,9 +67,7 @@
 //Register access policy version
 #define mmMP0_SMN_C2PMSG_91				0x1609B
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 static const uint32_t abm_gain_stepsize = 0x0060;
-#endif
 
 static bool dce_dmcu_init(struct dmcu *dmcu)
 {
@@ -81,9 +76,9 @@ static bool dce_dmcu_init(struct dmcu *dmcu)
 }
 
 static bool dce_dmcu_load_iram(struct dmcu *dmcu,
-		unsigned int start_offset,
-		const char *src,
-		unsigned int bytes)
+			       unsigned int start_offset,
+			       const char *src,
+			       unsigned int bytes)
 {
 	struct dce_dmcu *dmcu_dce = TO_DCE_DMCU(dmcu);
 	unsigned int count = 0;
@@ -333,7 +328,6 @@ static void dce_get_psr_wait_loop(
 	return;
 }
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 static void dcn10_get_dmcu_version(struct dmcu *dmcu)
 {
 	struct dce_dmcu *dmcu_dce = TO_DCE_DMCU(dmcu);
@@ -592,6 +586,7 @@ static void dcn10_dmcu_set_psr_enable(struct dmcu *dmcu, bool enable, bool wait)
 				if (state == PSR_STATE0)
 					break;
 			}
+			/* must *not* be fsleep - this can be called from high irq levels */
 			udelay(500);
 		}
 
@@ -930,23 +925,23 @@ static bool dcn10_recv_edid_cea_ack(struct dmcu *dmcu, int *offset)
 	return false;
 }
 
-#endif //(CONFIG_DRM_AMD_DC_DCN)
 
 #if defined(CONFIG_DRM_AMD_SECURE_DISPLAY)
 static void dcn10_forward_crc_window(struct dmcu *dmcu,
-					struct crc_region *crc_win,
+					struct rect *rect,
 					struct otg_phy_mux *mux_mapping)
 {
 	struct dce_dmcu *dmcu_dce = TO_DCE_DMCU(dmcu);
 	unsigned int dmcu_max_retry_on_wait_reg_ready = 801;
 	unsigned int dmcu_wait_reg_ready_interval = 100;
 	unsigned int crc_start = 0, crc_end = 0, otg_phy_mux = 0;
+	int x_start, y_start, x_end, y_end;
 
 	/* If microcontroller is not running, do nothing */
 	if (dmcu->dmcu_state != DMCU_RUNNING)
 		return;
 
-	if (!crc_win)
+	if (!rect)
 		return;
 
 	/* waitDMCUReadyForCmd */
@@ -954,9 +949,14 @@ static void dcn10_forward_crc_window(struct dmcu *dmcu,
 				dmcu_wait_reg_ready_interval,
 				dmcu_max_retry_on_wait_reg_ready);
 
+	x_start = rect->x;
+	y_start = rect->y;
+	x_end = x_start + rect->width;
+	y_end = y_start + rect->height;
+
 	/* build up nitification data */
-	crc_start = (((unsigned int) crc_win->x_start) << 16) | crc_win->y_start;
-	crc_end = (((unsigned int) crc_win->x_end) << 16) | crc_win->y_end;
+	crc_start = (((unsigned int) x_start) << 16) | y_start;
+	crc_end = (((unsigned int) x_end) << 16) | y_end;
 	otg_phy_mux =
 		(((unsigned int) mux_mapping->otg_output_num) << 16) | mux_mapping->phy_output_num;
 
@@ -1021,7 +1021,6 @@ static const struct dmcu_funcs dce_funcs = {
 	.is_dmcu_initialized = dce_is_dmcu_initialized
 };
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 static const struct dmcu_funcs dcn10_funcs = {
 	.dmcu_init = dcn10_dmcu_init,
 	.load_iram = dcn10_dmcu_load_iram,
@@ -1065,7 +1064,6 @@ static const struct dmcu_funcs dcn21_funcs = {
 	.lock_phy = dcn20_lock_phy,
 	.unlock_phy = dcn20_unlock_phy
 };
-#endif
 
 static void dce_dmcu_construct(
 	struct dce_dmcu *dmcu_dce,
@@ -1085,7 +1083,6 @@ static void dce_dmcu_construct(
 	dmcu_dce->dmcu_mask = dmcu_mask;
 }
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 static void dcn21_dmcu_construct(
 		struct dce_dmcu *dmcu_dce,
 		struct dc_context *ctx,
@@ -1097,13 +1094,10 @@ static void dcn21_dmcu_construct(
 
 	dce_dmcu_construct(dmcu_dce, ctx, regs, dmcu_shift, dmcu_mask);
 
-	if (!IS_FPGA_MAXIMUS_DC(ctx->dce_environment)) {
-		psp_version = dm_read_reg(ctx, mmMP0_SMN_C2PMSG_58);
-		dmcu_dce->base.auto_load_dmcu = ((psp_version & 0x00FF00FF) > 0x00110029);
-		dmcu_dce->base.psp_version = psp_version;
-	}
+	psp_version = dm_read_reg(ctx, mmMP0_SMN_C2PMSG_58);
+	dmcu_dce->base.auto_load_dmcu = ((psp_version & 0x00FF00FF) > 0x00110029);
+	dmcu_dce->base.psp_version = psp_version;
 }
-#endif
 
 struct dmcu *dce_dmcu_create(
 	struct dc_context *ctx,
@@ -1126,14 +1120,13 @@ struct dmcu *dce_dmcu_create(
 	return &dmcu_dce->base;
 }
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 struct dmcu *dcn10_dmcu_create(
 	struct dc_context *ctx,
 	const struct dce_dmcu_registers *regs,
 	const struct dce_dmcu_shift *dmcu_shift,
 	const struct dce_dmcu_mask *dmcu_mask)
 {
-	struct dce_dmcu *dmcu_dce = kzalloc(sizeof(*dmcu_dce), GFP_ATOMIC);
+	struct dce_dmcu *dmcu_dce = kzalloc(sizeof(*dmcu_dce), GFP_KERNEL);
 
 	if (dmcu_dce == NULL) {
 		BREAK_TO_DEBUGGER();
@@ -1154,7 +1147,7 @@ struct dmcu *dcn20_dmcu_create(
 	const struct dce_dmcu_shift *dmcu_shift,
 	const struct dce_dmcu_mask *dmcu_mask)
 {
-	struct dce_dmcu *dmcu_dce = kzalloc(sizeof(*dmcu_dce), GFP_ATOMIC);
+	struct dce_dmcu *dmcu_dce = kzalloc(sizeof(*dmcu_dce), GFP_KERNEL);
 
 	if (dmcu_dce == NULL) {
 		BREAK_TO_DEBUGGER();
@@ -1175,7 +1168,7 @@ struct dmcu *dcn21_dmcu_create(
 	const struct dce_dmcu_shift *dmcu_shift,
 	const struct dce_dmcu_mask *dmcu_mask)
 {
-	struct dce_dmcu *dmcu_dce = kzalloc(sizeof(*dmcu_dce), GFP_ATOMIC);
+	struct dce_dmcu *dmcu_dce = kzalloc(sizeof(*dmcu_dce), GFP_KERNEL);
 
 	if (dmcu_dce == NULL) {
 		BREAK_TO_DEBUGGER();
@@ -1189,7 +1182,6 @@ struct dmcu *dcn21_dmcu_create(
 
 	return &dmcu_dce->base;
 }
-#endif
 
 void dce_dmcu_destroy(struct dmcu **dmcu)
 {
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_i2c.c b/drivers/gpu/drm/amd/display/dc/dce/dce_i2c.c
index dd41736bb5c4..f5cd2392fc5f 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_i2c.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_i2c.c
@@ -25,6 +25,32 @@
 #include "dce_i2c.h"
 #include "reg_helper.h"
 
+bool dce_i2c_oem_device_present(
+	struct resource_pool *pool,
+	struct ddc_service *ddc,
+	size_t slave_address
+)
+{
+	struct dc *dc = ddc->ctx->dc;
+	struct dc_bios *dcb = dc->ctx->dc_bios;
+	struct graphics_object_id id = {0};
+	struct graphics_object_i2c_info i2c_info;
+
+	if (!dc->ctx->dc_bios->fw_info.oem_i2c_present)
+		return false;
+
+	id.id = dc->ctx->dc_bios->fw_info.oem_i2c_obj_id;
+	id.enum_id = 0;
+	id.type = OBJECT_TYPE_GENERIC;
+	if (dcb->funcs->get_i2c_info(dcb, id, &i2c_info) != BP_RESULT_OK)
+		return false;
+
+	if (i2c_info.i2c_slave_address != slave_address)
+		return false;
+
+	return true;
+}
+
 bool dce_i2c_submit_command(
 	struct resource_pool *pool,
 	struct ddc *ddc,
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_i2c.h b/drivers/gpu/drm/amd/display/dc/dce/dce_i2c.h
index a171c5cd8439..535fd58de450 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_i2c.h
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_i2c.h
@@ -30,6 +30,12 @@
 #include "dce_i2c_hw.h"
 #include "dce_i2c_sw.h"
 
+bool dce_i2c_oem_device_present(
+	struct resource_pool *pool,
+	struct ddc_service *ddc,
+	size_t slave_address
+);
+
 bool dce_i2c_submit_command(
 	struct resource_pool *pool,
 	struct ddc *ddc,
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_i2c_hw.c b/drivers/gpu/drm/amd/display/dc/dce/dce_i2c_hw.c
index 6d1b01c267b7..365dd2e37aea 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_i2c_hw.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_i2c_hw.c
@@ -22,9 +22,6 @@
  * Authors: AMD
  *
  */
-
-#include <linux/delay.h>
-
 #include "resource.h"
 #include "dce_i2c.h"
 #include "dce_i2c_hw.h"
@@ -295,24 +292,51 @@ static void set_speed(
 			     FN(DC_I2C_DDC1_SPEED, DC_I2C_DDC1_THRESHOLD), 2);
 }
 
+static bool acquire_engine(struct dce_i2c_hw *dce_i2c_hw)
+{
+	uint32_t arbitrate = 0;
+
+	REG_GET(DC_I2C_ARBITRATION, DC_I2C_REG_RW_CNTL_STATUS, &arbitrate);
+	switch (arbitrate) {
+	case DC_I2C_STATUS__DC_I2C_STATUS_USED_BY_SW:
+		return true;
+	case DC_I2C_STATUS__DC_I2C_STATUS_USED_BY_HW:
+		return false;
+	case DC_I2C_STATUS__DC_I2C_STATUS_IDLE:
+	default:
+		break;
+	}
+
+	REG_UPDATE(DC_I2C_ARBITRATION, DC_I2C_SW_USE_I2C_REG_REQ, true);
+	REG_GET(DC_I2C_ARBITRATION, DC_I2C_REG_RW_CNTL_STATUS, &arbitrate);
+	if (arbitrate != DC_I2C_STATUS__DC_I2C_STATUS_USED_BY_SW)
+		return false;
+
+	return true;
+}
+
 static bool setup_engine(
 	struct dce_i2c_hw *dce_i2c_hw)
 {
+	// Deassert soft reset to unblock I2C engine registers
+	REG_UPDATE(DC_I2C_CONTROL, DC_I2C_SOFT_RESET, false);
+
 	uint32_t i2c_setup_limit = I2C_SETUP_TIME_LIMIT_DCE;
 	uint32_t  reset_length = 0;
 
-        if (dce_i2c_hw->ctx->dc->debug.enable_mem_low_power.bits.i2c) {
-	     if (dce_i2c_hw->regs->DIO_MEM_PWR_CTRL) {
-		     REG_UPDATE(DIO_MEM_PWR_CTRL, I2C_LIGHT_SLEEP_FORCE, 0);
-		     REG_WAIT(DIO_MEM_PWR_STATUS, I2C_MEM_PWR_STATE, 0, 0, 5);
-		     }
-	     }
+	if (dce_i2c_hw->ctx->dc->debug.enable_mem_low_power.bits.i2c) {
+		if (dce_i2c_hw->regs->DIO_MEM_PWR_CTRL) {
+			REG_UPDATE(DIO_MEM_PWR_CTRL, I2C_LIGHT_SLEEP_FORCE, 0);
+			REG_WAIT(DIO_MEM_PWR_STATUS, I2C_MEM_PWR_STATE, 0, 0, 5);
+		}
+	}
 
-	/* we have checked I2c not used by DMCU, set SW use I2C REQ to 1 to indicate SW using it*/
-	REG_UPDATE(DC_I2C_ARBITRATION, DC_I2C_SW_USE_I2C_REG_REQ, 1);
+	if (dce_i2c_hw->masks->DC_I2C_DDC1_CLK_EN)
+		REG_UPDATE_N(SETUP, 1,
+			     FN(DC_I2C_DDC1_SETUP, DC_I2C_DDC1_CLK_EN), 1);
 
-	/* we have checked I2c not used by DMCU, set SW use I2C REQ to 1 to indicate SW using it*/
-	REG_UPDATE(DC_I2C_ARBITRATION, DC_I2C_SW_USE_I2C_REG_REQ, 1);
+	if (!acquire_engine(dce_i2c_hw))
+		return false;
 
 	/*set SW requested I2c speed to default, if API calls in it will be override later*/
 	set_speed(dce_i2c_hw, dce_i2c_hw->ctx->dc->caps.i2c_speed_in_khz);
@@ -321,9 +345,8 @@ static bool setup_engine(
 		i2c_setup_limit = dce_i2c_hw->setup_limit;
 
 	/* Program pin select */
-	REG_UPDATE_6(DC_I2C_CONTROL,
+	REG_UPDATE_5(DC_I2C_CONTROL,
 		     DC_I2C_GO, 0,
-		     DC_I2C_SOFT_RESET, 0,
 		     DC_I2C_SEND_RESET, 0,
 		     DC_I2C_SW_STATUS_RESET, 1,
 		     DC_I2C_TRANSACTION_COUNT, 0,
@@ -353,6 +376,32 @@ static bool setup_engine(
 	return true;
 }
 
+/**
+ * cntl_stuck_hw_workaround - Workaround for I2C engine stuck state
+ * @dce_i2c_hw: Pointer to dce_i2c_hw structure
+ *
+ * If we boot without an HDMI display, the I2C engine does not get initialized
+ * correctly. One of its symptoms is that SW_USE_I2C does not get cleared after
+ * acquire. After setting SW_DONE_USING_I2C on release, the engine gets
+ * immediately reacquired by SW, preventing DMUB from using it.
+ *
+ * This function checks the I2C arbitration status and applies a release
+ * workaround if necessary.
+ */
+static void cntl_stuck_hw_workaround(struct dce_i2c_hw *dce_i2c_hw)
+{
+	uint32_t arbitrate = 0;
+
+	REG_GET(DC_I2C_ARBITRATION, DC_I2C_REG_RW_CNTL_STATUS, &arbitrate);
+	if (arbitrate != DC_I2C_STATUS__DC_I2C_STATUS_USED_BY_SW)
+		return;
+
+	// Still acquired after release, release again as a workaround
+	REG_UPDATE(DC_I2C_ARBITRATION, DC_I2C_SW_DONE_USING_I2C_REG, true);
+	REG_GET(DC_I2C_ARBITRATION, DC_I2C_REG_RW_CNTL_STATUS, &arbitrate);
+	ASSERT(arbitrate != DC_I2C_STATUS__DC_I2C_STATUS_USED_BY_SW);
+}
+
 static void release_engine(
 	struct dce_i2c_hw *dce_i2c_hw)
 {
@@ -380,9 +429,9 @@ static void release_engine(
 
 	/*for HW HDCP Ri polling failure w/a test*/
 	set_speed(dce_i2c_hw, dce_i2c_hw->ctx->dc->caps.i2c_speed_in_khz_hdcp);
-	/* Release I2C after reset, so HW or DMCU could use it */
-	REG_UPDATE_2(DC_I2C_ARBITRATION, DC_I2C_SW_DONE_USING_I2C_REG, 1,
-		DC_I2C_SW_USE_I2C_REG_REQ, 0);
+	// Release I2C engine so it can be used by HW or DMCU, automatically clears SW_USE_I2C
+	REG_UPDATE(DC_I2C_ARBITRATION, DC_I2C_SW_DONE_USING_I2C_REG, true);
+	cntl_stuck_hw_workaround(dce_i2c_hw);
 
 	if (dce_i2c_hw->ctx->dc->debug.enable_mem_low_power.bits.i2c) {
 		if (dce_i2c_hw->regs->DIO_MEM_PWR_CTRL)
@@ -442,10 +491,9 @@ struct dce_i2c_hw *acquire_i2c_hw_engine(
 	return dce_i2c_hw;
 }
 
-static enum i2c_channel_operation_result dce_i2c_hw_engine_wait_on_operation_result(
-	struct dce_i2c_hw *dce_i2c_hw,
-	uint32_t timeout,
-	enum i2c_channel_operation_result expected_result)
+static enum i2c_channel_operation_result dce_i2c_hw_engine_wait_on_operation_result(struct dce_i2c_hw *dce_i2c_hw,
+										    uint32_t timeout,
+										    enum i2c_channel_operation_result expected_result)
 {
 	enum i2c_channel_operation_result result;
 	uint32_t i = 0;
@@ -509,11 +557,10 @@ static uint32_t get_transaction_timeout_hw(
 	return period_timeout * num_of_clock_stretches;
 }
 
-static bool dce_i2c_hw_engine_submit_payload(
-	struct dce_i2c_hw *dce_i2c_hw,
-	struct i2c_payload *payload,
-	bool middle_of_transaction,
-	uint32_t speed)
+static bool dce_i2c_hw_engine_submit_payload(struct dce_i2c_hw *dce_i2c_hw,
+					     struct i2c_payload *payload,
+					     bool middle_of_transaction,
+					     uint32_t speed)
 {
 
 	struct i2c_request_transaction_data request;
@@ -544,7 +591,7 @@ static bool dce_i2c_hw_engine_submit_payload(
 			DCE_I2C_TRANSACTION_ACTION_I2C_WRITE;
 
 
-	request.address = (uint8_t) ((payload->address << 1) | !payload->write);
+	request.address = (uint8_t) ((payload->address << 1) | (payload->write ? 0 : 1));
 	request.length = payload->length;
 	request.data = payload->data;
 
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_i2c_hw.h b/drivers/gpu/drm/amd/display/dc/dce/dce_i2c_hw.h
index 3f45ecd189a2..a9a16f645994 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_i2c_hw.h
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_i2c_hw.h
@@ -188,6 +188,7 @@ struct dce_i2c_shift {
 	uint8_t DC_I2C_REG_RW_CNTL_STATUS;
 	uint8_t I2C_LIGHT_SLEEP_FORCE;
 	uint8_t I2C_MEM_PWR_STATE;
+	uint8_t DC_I2C_DDC1_CLK_EN;
 };
 
 struct dce_i2c_mask {
@@ -232,6 +233,7 @@ struct dce_i2c_mask {
 	uint32_t DC_I2C_REG_RW_CNTL_STATUS;
 	uint32_t I2C_LIGHT_SLEEP_FORCE;
 	uint32_t I2C_MEM_PWR_STATE;
+	uint32_t DC_I2C_DDC1_CLK_EN;
 };
 
 #define I2C_COMMON_MASK_SH_LIST_DCN2(mask_sh)\
@@ -243,6 +245,14 @@ struct dce_i2c_mask {
 	I2C_SF(DIO_MEM_PWR_CTRL, I2C_LIGHT_SLEEP_FORCE, mask_sh),\
 	I2C_SF(DIO_MEM_PWR_STATUS, I2C_MEM_PWR_STATE, mask_sh)
 
+#define I2C_COMMON_MASK_SH_LIST_DCN35(mask_sh)\
+	I2C_COMMON_MASK_SH_LIST_DCN30(mask_sh),\
+	I2C_SF(DC_I2C_DDC1_SETUP, DC_I2C_DDC1_CLK_EN, mask_sh)
+
+#define I2C_COMMON_MASK_SH_LIST_DCN401(mask_sh)\
+	I2C_COMMON_MASK_SH_LIST_DCN30(mask_sh),\
+	I2C_SF(DC_I2C_DDC1_SETUP, DC_I2C_DDC1_CLK_EN, mask_sh)
+
 struct dce_i2c_registers {
 	uint32_t SETUP;
 	uint32_t SPEED;
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_i2c_sw.c b/drivers/gpu/drm/amd/display/dc/dce/dce_i2c_sw.c
index 6846afd83701..2d73b94c515c 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_i2c_sw.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_i2c_sw.c
@@ -23,8 +23,6 @@
  *
  */
 
-#include <linux/delay.h>
-
 #include "dce_i2c.h"
 #include "dce_i2c_sw.h"
 #include "include/gpio_service_interface.h"
@@ -369,6 +367,7 @@ static bool dce_i2c_sw_engine_acquire_engine(
 
 	return true;
 }
+
 bool dce_i2c_engine_acquire_sw(
 	struct dce_i2c_sw *dce_i2c_sw,
 	struct ddc *ddc_handle)
@@ -394,12 +393,8 @@ bool dce_i2c_engine_acquire_sw(
 	return result;
 }
 
-
-
-
-static void dce_i2c_sw_engine_submit_channel_request(
-	struct dce_i2c_sw *engine,
-	struct i2c_request_transaction_data *req)
+static void dce_i2c_sw_engine_submit_channel_request(struct dce_i2c_sw *engine,
+						     struct i2c_request_transaction_data *req)
 {
 	struct ddc *ddc = engine->ddc;
 	uint16_t clock_delay_div_4 = engine->clock_delay >> 2;
@@ -441,10 +436,9 @@ static void dce_i2c_sw_engine_submit_channel_request(
 		I2C_CHANNEL_OPERATION_FAILED;
 }
 
-static bool dce_i2c_sw_engine_submit_payload(
-	struct dce_i2c_sw *engine,
-	struct i2c_payload *payload,
-	bool middle_of_transaction)
+static bool dce_i2c_sw_engine_submit_payload(struct dce_i2c_sw *engine,
+					     struct i2c_payload *payload,
+					     bool middle_of_transaction)
 {
 	struct i2c_request_transaction_data request;
 
@@ -457,7 +451,7 @@ static bool dce_i2c_sw_engine_submit_payload(
 			DCE_I2C_TRANSACTION_ACTION_I2C_WRITE_MOT :
 			DCE_I2C_TRANSACTION_ACTION_I2C_WRITE;
 
-	request.address = (uint8_t) ((payload->address << 1) | !payload->write);
+	request.address = (uint8_t) ((payload->address << 1) | (payload->write ? 0 : 1));
 	request.length = payload->length;
 	request.data = payload->data;
 
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_ipp.c b/drivers/gpu/drm/amd/display/dc/dce/dce_ipp.c
index 80569a2734eb..34bff9aef66c 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_ipp.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_ipp.c
@@ -23,8 +23,6 @@
  *
  */
 
-#include <linux/slab.h>
-
 #include "dce_ipp.h"
 #include "reg_helper.h"
 #include "dm_services.h"
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_link_encoder.c b/drivers/gpu/drm/amd/display/dc/dce/dce_link_encoder.c
index 1e77ffee71b3..87dbb8d7ed27 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_link_encoder.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_link_encoder.c
@@ -23,16 +23,12 @@
  *
  */
 
-#include <linux/delay.h>
-#include <linux/slab.h>
-
 #include "reg_helper.h"
 
 #include "core_types.h"
 #include "link_encoder.h"
 #include "dce_link_encoder.h"
 #include "stream_encoder.h"
-#include "i2caux_interface.h"
 #include "dc_bios_types.h"
 
 #include "gpio_service_interface.h"
@@ -306,6 +302,10 @@ static void setup_panel_mode(
 	if (ctx->dc->caps.psp_setup_panel_mode)
 		return;
 
+	/* The code below is only applicable to encoders with a digital transmitter. */
+	if (enc110->base.transmitter == TRANSMITTER_UNKNOWN)
+		return;
+
 	ASSERT(REG(DP_DPHY_INTERNAL_CTRL));
 	value = REG_READ(DP_DPHY_INTERNAL_CTRL);
 
@@ -788,8 +788,9 @@ static bool dce110_link_encoder_validate_hdmi_output(
 			crtc_timing->pixel_encoding == PIXEL_ENCODING_YCBCR420)
 		return false;
 
-	if (!enc110->base.features.flags.bits.HDMI_6GB_EN &&
-		adjusted_pix_clk_khz >= 300000)
+	if ((!enc110->base.features.flags.bits.HDMI_6GB_EN ||
+			enc110->base.ctx->dc->debug.hdmi20_disable) &&
+			adjusted_pix_clk_khz >= 300000)
 		return false;
 	if (enc110->base.ctx->dc->debug.hdmi20_disable &&
 		crtc_timing->pixel_encoding == PIXEL_ENCODING_YCBCR420)
@@ -807,6 +808,33 @@ bool dce110_link_encoder_validate_dp_output(
 	return true;
 }
 
+static bool dce110_link_encoder_validate_rgb_output(
+	const struct dce110_link_encoder *enc110,
+	const struct dc_crtc_timing *crtc_timing)
+{
+	/* When the VBIOS doesn't specify any limits, use 400 MHz.
+	 * The value comes from amdgpu_atombios_get_clock_info.
+	 */
+	uint32_t max_pixel_clock_khz = 400000;
+
+	if (enc110->base.ctx->dc_bios->fw_info_valid &&
+	    enc110->base.ctx->dc_bios->fw_info.max_pixel_clock) {
+		max_pixel_clock_khz =
+			enc110->base.ctx->dc_bios->fw_info.max_pixel_clock;
+	}
+
+	if (crtc_timing->pix_clk_100hz > max_pixel_clock_khz * 10)
+		return false;
+
+	if (crtc_timing->display_color_depth != COLOR_DEPTH_888)
+		return false;
+
+	if (crtc_timing->pixel_encoding != PIXEL_ENCODING_RGB)
+		return false;
+
+	return true;
+}
+
 void dce110_link_encoder_construct(
 	struct dce110_link_encoder *enc110,
 	const struct encoder_init_data *init_data,
@@ -827,6 +855,7 @@ void dce110_link_encoder_construct(
 	enc110->base.connector = init_data->connector;
 
 	enc110->base.preferred_engine = ENGINE_ID_UNKNOWN;
+	enc110->base.analog_engine = init_data->analog_engine;
 
 	enc110->base.features = *enc_features;
 
@@ -850,6 +879,11 @@ void dce110_link_encoder_construct(
 		SIGNAL_TYPE_EDP |
 		SIGNAL_TYPE_HDMI_TYPE_A;
 
+	if ((enc110->base.connector.id == CONNECTOR_ID_DUAL_LINK_DVII ||
+	     enc110->base.connector.id == CONNECTOR_ID_SINGLE_LINK_DVII) &&
+		enc110->base.analog_engine != ENGINE_ID_UNKNOWN)
+		enc110->base.output_signals |= SIGNAL_TYPE_RGB;
+
 	/* For DCE 8.0 and 8.1, by design, UNIPHY is hardwired to DIG_BE.
 	 * SW always assign DIG_FE 1:1 mapped to DIG_FE for non-MST UNIPHY.
 	 * SW assign DIG_FE to non-MST UNIPHY first and MST last. So prefer
@@ -888,6 +922,13 @@ void dce110_link_encoder_construct(
 		enc110->base.preferred_engine = ENGINE_ID_DIGG;
 	break;
 	default:
+		if (init_data->analog_engine != ENGINE_ID_UNKNOWN) {
+			/* The connector is analog-only, ie. VGA */
+			enc110->base.preferred_engine = init_data->analog_engine;
+			enc110->base.output_signals = SIGNAL_TYPE_RGB;
+			enc110->base.transmitter = TRANSMITTER_UNKNOWN;
+			break;
+		}
 		ASSERT_CRITICAL(false);
 		enc110->base.preferred_engine = ENGINE_ID_UNKNOWN;
 	}
@@ -899,13 +940,13 @@ void dce110_link_encoder_construct(
 						enc110->base.id, &bp_cap_info);
 
 	/* Override features with DCE-specific values */
-	if (BP_RESULT_OK == result) {
+	if (result == BP_RESULT_OK) {
 		enc110->base.features.flags.bits.IS_HBR2_CAPABLE =
 				bp_cap_info.DP_HBR2_EN;
 		enc110->base.features.flags.bits.IS_HBR3_CAPABLE =
 				bp_cap_info.DP_HBR3_EN;
 		enc110->base.features.flags.bits.HDMI_6GB_EN = bp_cap_info.HDMI_6GB_EN;
-	} else {
+	} else if (result != BP_RESULT_NORECORD) {
 		DC_LOG_WARNING("%s: Failed to get encoder_cap_info from VBIOS with error code %d!\n",
 				__func__,
 				result);
@@ -942,11 +983,13 @@ bool dce110_link_encoder_validate_output_with_stream(
 		is_valid = dce110_link_encoder_validate_dp_output(
 					enc110, &stream->timing);
 	break;
+	case SIGNAL_TYPE_RGB:
+		is_valid = dce110_link_encoder_validate_rgb_output(
+					enc110, &stream->timing);
+	break;
 	case SIGNAL_TYPE_EDP:
 	case SIGNAL_TYPE_LVDS:
-		is_valid =
-			(stream->timing.
-				pixel_encoding == PIXEL_ENCODING_RGB) ? true : false;
+		is_valid = stream->timing.pixel_encoding == PIXEL_ENCODING_RGB;
 	break;
 	case SIGNAL_TYPE_VIRTUAL:
 		is_valid = true;
@@ -974,6 +1017,10 @@ void dce110_link_encoder_hw_init(
 	cntl.coherent = false;
 	cntl.hpd_sel = enc110->base.hpd_source;
 
+	/* The code below is only applicable to encoders with a digital transmitter. */
+	if (enc110->base.transmitter == TRANSMITTER_UNKNOWN)
+		return;
+
 	if (enc110->base.connector.id == CONNECTOR_ID_EDP)
 		cntl.signal = SIGNAL_TYPE_EDP;
 
@@ -1039,6 +1086,8 @@ void dce110_link_encoder_setup(
 		/* DP MST */
 		REG_UPDATE(DIG_BE_CNTL, DIG_MODE, 5);
 		break;
+	case SIGNAL_TYPE_RGB:
+		break;
 	default:
 		ASSERT_CRITICAL(false);
 		/* invalid mode ! */
@@ -1287,6 +1336,24 @@ void dce110_link_encoder_disable_output(
 	struct bp_transmitter_control cntl = { 0 };
 	enum bp_result result;
 
+	switch (enc->analog_engine) {
+	case ENGINE_ID_DACA:
+		REG_UPDATE(DAC_ENABLE, DAC_ENABLE, 0);
+		break;
+	case ENGINE_ID_DACB:
+		/* DACB doesn't seem to be present on DCE6+,
+		 * although there are references to it in the register file.
+		 */
+		DC_LOG_ERROR("%s DACB is unsupported\n", __func__);
+		break;
+	default:
+		break;
+	}
+
+	/* The code below only applies to connectors that support digital signals. */
+	if (enc->transmitter == TRANSMITTER_UNKNOWN)
+		return;
+
 	if (!dce110_is_dig_enabled(enc)) {
 		/* OF_SKIP_POWER_DOWN_INACTIVE_ENCODER */
 		return;
@@ -1324,7 +1391,8 @@ void dce110_link_encoder_disable_output(
 
 void dce110_link_encoder_dp_set_lane_settings(
 	struct link_encoder *enc,
-	const struct link_training_settings *link_settings)
+	const struct dc_link_settings *link_settings,
+	const struct dc_lane_settings lane_settings[LANE_COUNT_DP_MAX])
 {
 	struct dce110_link_encoder *enc110 = TO_DCE110_LINK_ENC(enc);
 	union dpcd_training_lane_set training_lane_set = { { 0 } };
@@ -1339,33 +1407,36 @@ void dce110_link_encoder_dp_set_lane_settings(
 	cntl.action = TRANSMITTER_CONTROL_SET_VOLTAGE_AND_PREEMPASIS;
 	cntl.transmitter = enc110->base.transmitter;
 	cntl.connector_obj_id = enc110->base.connector;
-	cntl.lanes_number = link_settings->link_settings.lane_count;
+	cntl.lanes_number = link_settings->lane_count;
 	cntl.hpd_sel = enc110->base.hpd_source;
-	cntl.pixel_clock = link_settings->link_settings.link_rate *
+	cntl.pixel_clock = link_settings->link_rate *
 						LINK_RATE_REF_FREQ_IN_KHZ;
 
-	for (lane = 0; lane < link_settings->link_settings.lane_count; lane++) {
+	for (lane = 0; lane < link_settings->lane_count; lane++) {
 		/* translate lane settings */
 
 		training_lane_set.bits.VOLTAGE_SWING_SET =
-			link_settings->lane_settings[lane].VOLTAGE_SWING;
+				lane_settings[lane].VOLTAGE_SWING;
 		training_lane_set.bits.PRE_EMPHASIS_SET =
-			link_settings->lane_settings[lane].PRE_EMPHASIS;
+				lane_settings[lane].PRE_EMPHASIS;
 
 		/* post cursor 2 setting only applies to HBR2 link rate */
-		if (link_settings->link_settings.link_rate == LINK_RATE_HIGH2) {
+		if (link_settings->link_rate == LINK_RATE_HIGH2) {
 			/* this is passed to VBIOS
 			 * to program post cursor 2 level */
 
 			training_lane_set.bits.POST_CURSOR2_SET =
-				link_settings->lane_settings[lane].POST_CURSOR2;
+					lane_settings[lane].POST_CURSOR2;
 		}
 
 		cntl.lane_select = lane;
 		cntl.lane_settings = training_lane_set.raw;
 
 		/* call VBIOS table to set voltage swing and pre-emphasis */
-		link_transmitter_control(enc110, &cntl);
+		if (link_transmitter_control(enc110, &cntl) != BP_RESULT_OK) {
+			DC_LOG_ERROR("%s: Failed to execute VBIOS command table!\n", __func__);
+			BREAK_TO_DEBUGGER();
+		}
 	}
 }
 
@@ -1647,7 +1718,7 @@ void dce110_link_encoder_enable_hpd(struct link_encoder *enc)
 	uint32_t hpd_enable = 0;
 	uint32_t value = dm_read_reg(ctx, addr);
 
-	get_reg_field_value(hpd_enable, DC_HPD_CONTROL, DC_HPD_EN);
+	hpd_enable = get_reg_field_value(hpd_enable, DC_HPD_CONTROL, DC_HPD_EN);
 
 	if (hpd_enable == 0)
 		set_reg_field_value(value, 1, DC_HPD_CONTROL, DC_HPD_EN);
@@ -1727,6 +1798,7 @@ void dce60_link_encoder_construct(
 	enc110->base.connector = init_data->connector;
 
 	enc110->base.preferred_engine = ENGINE_ID_UNKNOWN;
+	enc110->base.analog_engine = init_data->analog_engine;
 
 	enc110->base.features = *enc_features;
 
@@ -1750,6 +1822,11 @@ void dce60_link_encoder_construct(
 		SIGNAL_TYPE_EDP |
 		SIGNAL_TYPE_HDMI_TYPE_A;
 
+	if ((enc110->base.connector.id == CONNECTOR_ID_DUAL_LINK_DVII ||
+	     enc110->base.connector.id == CONNECTOR_ID_SINGLE_LINK_DVII) &&
+		enc110->base.analog_engine != ENGINE_ID_UNKNOWN)
+		enc110->base.output_signals |= SIGNAL_TYPE_RGB;
+
 	/* For DCE 8.0 and 8.1, by design, UNIPHY is hardwired to DIG_BE.
 	 * SW always assign DIG_FE 1:1 mapped to DIG_FE for non-MST UNIPHY.
 	 * SW assign DIG_FE to non-MST UNIPHY first and MST last. So prefer
@@ -1788,6 +1865,13 @@ void dce60_link_encoder_construct(
 		enc110->base.preferred_engine = ENGINE_ID_DIGG;
 	break;
 	default:
+		if (init_data->analog_engine != ENGINE_ID_UNKNOWN) {
+			/* The connector is analog-only, ie. VGA */
+			enc110->base.preferred_engine = init_data->analog_engine;
+			enc110->base.output_signals = SIGNAL_TYPE_RGB;
+			enc110->base.transmitter = TRANSMITTER_UNKNOWN;
+			break;
+		}
 		ASSERT_CRITICAL(false);
 		enc110->base.preferred_engine = ENGINE_ID_UNKNOWN;
 	}
@@ -1799,13 +1883,13 @@ void dce60_link_encoder_construct(
 						enc110->base.id, &bp_cap_info);
 
 	/* Override features with DCE-specific values */
-	if (BP_RESULT_OK == result) {
+	if (result == BP_RESULT_OK) {
 		enc110->base.features.flags.bits.IS_HBR2_CAPABLE =
 				bp_cap_info.DP_HBR2_EN;
 		enc110->base.features.flags.bits.IS_HBR3_CAPABLE =
 				bp_cap_info.DP_HBR3_EN;
 		enc110->base.features.flags.bits.HDMI_6GB_EN = bp_cap_info.HDMI_6GB_EN;
-	} else {
+	} else if (result != BP_RESULT_NORECORD) {
 		DC_LOG_WARNING("%s: Failed to get encoder_cap_info from VBIOS with error code %d!\n",
 				__func__,
 				result);
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_link_encoder.h b/drivers/gpu/drm/amd/display/dc/dce/dce_link_encoder.h
index fc6ade824c23..c58b69bc319b 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_link_encoder.h
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_link_encoder.h
@@ -101,18 +101,21 @@
 	SRI(DP_SEC_CNTL, DP, id), \
 	SRI(DP_VID_STREAM_CNTL, DP, id), \
 	SRI(DP_DPHY_FAST_TRAINING, DP, id), \
-	SRI(DP_SEC_CNTL1, DP, id)
+	SRI(DP_SEC_CNTL1, DP, id), \
+	SR(DAC_ENABLE)
 #endif
 
 #define LE_DCE80_REG_LIST(id)\
 	SRI(DP_DPHY_INTERNAL_CTRL, DP, id), \
-	LE_COMMON_REG_LIST_BASE(id)
+	LE_COMMON_REG_LIST_BASE(id), \
+	SR(DAC_ENABLE)
 
 #define LE_DCE100_REG_LIST(id)\
 	LE_COMMON_REG_LIST_BASE(id), \
 	SRI(DP_DPHY_BS_SR_SWAP_CNTL, DP, id), \
 	SRI(DP_DPHY_INTERNAL_CTRL, DP, id), \
-	SR(DCI_MEM_PWR_STATUS)
+	SR(DCI_MEM_PWR_STATUS), \
+	SR(DAC_ENABLE)
 
 #define LE_DCE110_REG_LIST(id)\
 	LE_COMMON_REG_LIST_BASE(id), \
@@ -181,6 +184,9 @@ struct dce110_link_enc_registers {
 	uint32_t DP_DPHY_BS_SR_SWAP_CNTL;
 	uint32_t DP_DPHY_HBR2_PATTERN_CONTROL;
 	uint32_t DP_SEC_CNTL1;
+
+	/* DAC registers */
+	uint32_t DAC_ENABLE;
 };
 
 struct dce110_link_encoder {
@@ -215,10 +221,6 @@ bool dce110_link_encoder_validate_dvi_output(
 	enum signal_type signal,
 	const struct dc_crtc_timing *crtc_timing);
 
-bool dce110_link_encoder_validate_rgb_output(
-	const struct dce110_link_encoder *enc110,
-	const struct dc_crtc_timing *crtc_timing);
-
 bool dce110_link_encoder_validate_dp_output(
 	const struct dce110_link_encoder *enc110,
 	const struct dc_crtc_timing *crtc_timing);
@@ -279,7 +281,8 @@ void dce110_link_encoder_disable_output(
 /* set DP lane settings */
 void dce110_link_encoder_dp_set_lane_settings(
 	struct link_encoder *enc,
-	const struct link_training_settings *link_settings);
+	const struct dc_link_settings *link_settings,
+	const struct dc_lane_settings lane_settings[LANE_COUNT_DP_MAX]);
 
 void dce110_link_encoder_dp_set_phy_pattern(
 	struct link_encoder *enc,
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_mem_input.c b/drivers/gpu/drm/amd/display/dc/dce/dce_mem_input.c
index 4cdd4dacb761..1c2009e38aa1 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_mem_input.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_mem_input.c
@@ -98,7 +98,7 @@ static enum mi_bits_per_pixel get_mi_bpp(
 }
 
 static enum mi_tiling_format get_mi_tiling(
-		union dc_tiling_info *tiling_info)
+		struct dc_tiling_info *tiling_info)
 {
 	switch (tiling_info->gfx8.array_mode) {
 	case DC_ARRAY_1D_TILED_THIN1:
@@ -133,7 +133,7 @@ static bool is_vert_scan(enum dc_rotation_angle rotation)
 static void dce_mi_program_pte_vm(
 		struct mem_input *mi,
 		enum surface_pixel_format format,
-		union dc_tiling_info *tiling_info,
+		struct dc_tiling_info *tiling_info,
 		enum dc_rotation_angle rotation)
 {
 	struct dce_mem_input *dce_mi = TO_DCE_MEM_INPUT(mi);
@@ -430,7 +430,7 @@ static void dce120_mi_program_display_marks(struct mem_input *mi,
 }
 
 static void program_tiling(
-	struct dce_mem_input *dce_mi, const union dc_tiling_info *info)
+	struct dce_mem_input *dce_mi, const struct dc_tiling_info *info)
 {
 	if (dce_mi->masks->GRPH_SW_MODE) { /* GFX9 */
 		REG_UPDATE_6(GRPH_CONTROL,
@@ -481,7 +481,6 @@ static void program_tiling(
 	}
 }
 
-
 static void program_size_and_rotation(
 	struct dce_mem_input *dce_mi,
 	enum dc_rotation_angle rotation,
@@ -627,10 +626,31 @@ static void program_grph_pixel_format(
 			GRPH_PRESCALE_B_SIGN, sign);
 }
 
+static void dce_mi_clear_tiling(
+	struct mem_input *mi)
+{
+	struct dce_mem_input *dce_mi = TO_DCE_MEM_INPUT(mi);
+
+	if (dce_mi->masks->GRPH_SW_MODE) { /* GFX9 */
+		REG_UPDATE(GRPH_CONTROL,
+			   GRPH_SW_MODE, DC_SW_LINEAR);
+	}
+
+	if (dce_mi->masks->GRPH_MICRO_TILE_MODE) { /* GFX8 */
+		REG_UPDATE(GRPH_CONTROL,
+			   GRPH_ARRAY_MODE, DC_SW_LINEAR);
+	}
+
+	if (dce_mi->masks->GRPH_ARRAY_MODE) { /* GFX6 but reuses gfx8 struct */
+		REG_UPDATE(GRPH_CONTROL,
+			   GRPH_ARRAY_MODE, DC_SW_LINEAR);
+	}
+}
+
 static void dce_mi_program_surface_config(
 	struct mem_input *mi,
 	enum surface_pixel_format format,
-	union dc_tiling_info *tiling_info,
+	struct dc_tiling_info *tiling_info,
 	struct plane_size *plane_size,
 	enum dc_rotation_angle rotation,
 	struct dc_plane_dcc_param *dcc,
@@ -642,8 +662,7 @@ static void dce_mi_program_surface_config(
 	program_tiling(dce_mi, tiling_info);
 	program_size_and_rotation(dce_mi, rotation, plane_size);
 
-	if (format >= SURFACE_PIXEL_FORMAT_GRPH_BEGIN &&
-		format < SURFACE_PIXEL_FORMAT_VIDEO_BEGIN)
+	if (format < SURFACE_PIXEL_FORMAT_VIDEO_BEGIN)
 		program_grph_pixel_format(dce_mi, format);
 }
 
@@ -651,7 +670,7 @@ static void dce_mi_program_surface_config(
 static void dce60_mi_program_surface_config(
 	struct mem_input *mi,
 	enum surface_pixel_format format,
-	union dc_tiling_info *tiling_info,
+	struct dc_tiling_info *tiling_info,
 	struct plane_size *plane_size,
 	enum dc_rotation_angle rotation, /* not used in DCE6 */
 	struct dc_plane_dcc_param *dcc,
@@ -663,8 +682,7 @@ static void dce60_mi_program_surface_config(
 	program_tiling(dce_mi, tiling_info);
 	dce60_program_size(dce_mi, rotation, plane_size);
 
-	if (format >= SURFACE_PIXEL_FORMAT_GRPH_BEGIN &&
-		format < SURFACE_PIXEL_FORMAT_VIDEO_BEGIN)
+	if (format < SURFACE_PIXEL_FORMAT_VIDEO_BEGIN)
 		program_grph_pixel_format(dce_mi, format);
 }
 #endif
@@ -886,7 +904,8 @@ static const struct mem_input_funcs dce_mi_funcs = {
 	.mem_input_program_pte_vm = dce_mi_program_pte_vm,
 	.mem_input_program_surface_config =
 			dce_mi_program_surface_config,
-	.mem_input_is_flip_pending = dce_mi_is_flip_pending
+	.mem_input_is_flip_pending = dce_mi_is_flip_pending,
+	.mem_input_clear_tiling = dce_mi_clear_tiling,
 };
 
 #if defined(CONFIG_DRM_AMD_DC_SI)
@@ -899,7 +918,8 @@ static const struct mem_input_funcs dce60_mi_funcs = {
 	.mem_input_program_pte_vm = dce_mi_program_pte_vm,
 	.mem_input_program_surface_config =
 			dce60_mi_program_surface_config,
-	.mem_input_is_flip_pending = dce_mi_is_flip_pending
+	.mem_input_is_flip_pending = dce_mi_is_flip_pending,
+	.mem_input_clear_tiling = dce_mi_clear_tiling,
 };
 #endif
 
@@ -912,7 +932,8 @@ static const struct mem_input_funcs dce112_mi_funcs = {
 	.mem_input_program_pte_vm = dce_mi_program_pte_vm,
 	.mem_input_program_surface_config =
 			dce_mi_program_surface_config,
-	.mem_input_is_flip_pending = dce_mi_is_flip_pending
+	.mem_input_is_flip_pending = dce_mi_is_flip_pending,
+	.mem_input_clear_tiling = dce_mi_clear_tiling,
 };
 
 static const struct mem_input_funcs dce120_mi_funcs = {
@@ -924,7 +945,8 @@ static const struct mem_input_funcs dce120_mi_funcs = {
 	.mem_input_program_pte_vm = dce_mi_program_pte_vm,
 	.mem_input_program_surface_config =
 			dce_mi_program_surface_config,
-	.mem_input_is_flip_pending = dce_mi_is_flip_pending
+	.mem_input_is_flip_pending = dce_mi_is_flip_pending,
+	.mem_input_clear_tiling = dce_mi_clear_tiling,
 };
 
 void dce_mem_input_construct(
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_mem_input.h b/drivers/gpu/drm/amd/display/dc/dce/dce_mem_input.h
index f98400efdd9b..e34e445a4013 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_mem_input.h
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_mem_input.h
@@ -181,6 +181,7 @@ struct dce_mem_input_registers {
 	SFB(blk, GRPH_ENABLE, GRPH_ENABLE, mask_sh),\
 	SFB(blk, GRPH_CONTROL, GRPH_DEPTH, mask_sh),\
 	SFB(blk, GRPH_CONTROL, GRPH_FORMAT, mask_sh),\
+	SFB(blk, GRPH_CONTROL, GRPH_NUM_BANKS, mask_sh),\
 	SFB(blk, GRPH_X_START, GRPH_X_START, mask_sh),\
 	SFB(blk, GRPH_Y_START, GRPH_Y_START, mask_sh),\
 	SFB(blk, GRPH_X_END, GRPH_X_END, mask_sh),\
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_opp.c b/drivers/gpu/drm/amd/display/dc/dce/dce_opp.c
index 895b015b02e8..f342da5a5e50 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_opp.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_opp.c
@@ -23,8 +23,6 @@
  *
  */
 
-#include <linux/slab.h>
-
 #include "dm_services.h"
 #include "basics/conversion.h"
 
@@ -577,7 +575,6 @@ static void dce60_opp_program_clamping_and_pixel_encoding(
 }
 #endif
 
-
 static void program_formatter_420_memory(struct output_pixel_processor *opp)
 {
 	struct dce110_opp *opp110 = TO_DCE110_OPP(opp);
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_opp.h b/drivers/gpu/drm/amd/display/dc/dce/dce_opp.h
index bf1ffc3629c7..3d9be87aae45 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_opp.h
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_opp.h
@@ -111,6 +111,7 @@ enum dce110_opp_reg_type {
 	OPP_SF(FMT_DITHER_RAND_R_SEED, FMT_RAND_R_SEED, mask_sh),\
 	OPP_SF(FMT_DITHER_RAND_G_SEED, FMT_RAND_G_SEED, mask_sh),\
 	OPP_SF(FMT_DITHER_RAND_B_SEED, FMT_RAND_B_SEED, mask_sh),\
+	OPP_SF(FMT_BIT_DEPTH_CONTROL, FMT_TEMPORAL_DITHER_EN, mask_sh),\
 	OPP_SF(FMT_BIT_DEPTH_CONTROL, FMT_TEMPORAL_DITHER_RESET, mask_sh),\
 	OPP_SF(FMT_BIT_DEPTH_CONTROL, FMT_TEMPORAL_DITHER_OFFSET, mask_sh),\
 	OPP_SF(FMT_BIT_DEPTH_CONTROL, FMT_TEMPORAL_DITHER_DEPTH, mask_sh),\
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_panel_cntl.c b/drivers/gpu/drm/amd/display/dc/dce/dce_panel_cntl.c
index e8570060d007..de31fb1b6819 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_panel_cntl.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_panel_cntl.c
@@ -218,7 +218,7 @@ static void dce_driver_set_backlight(struct panel_cntl *panel_cntl,
 	 * contain integer component, lower 16 bits contain fractional component
 	 * of active duty cycle e.g. 0x21BDC0 = 0xEFF0 * 0x24
 	 */
-	active_duty_cycle = backlight_pwm_u16_16 * masked_pwm_period;
+	active_duty_cycle = backlight_pwm_u16_16 * (uint64_t)masked_pwm_period;
 
 	/* 1.3 Calculate 16 bit active duty cycle from integer and fractional
 	 * components shift by bitCount then mask 16 bits and add rounding bit
@@ -290,4 +290,5 @@ void dce_panel_cntl_construct(
 	dce_panel_cntl->base.funcs = &dce_link_panel_cntl_funcs;
 	dce_panel_cntl->base.ctx = init_data->ctx;
 	dce_panel_cntl->base.inst = init_data->inst;
+	dce_panel_cntl->base.pwrseq_inst = 0;
 }
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_stream_encoder.c b/drivers/gpu/drm/amd/display/dc/dce/dce_stream_encoder.c
index 779bc92a2968..574618d5d4a4 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_stream_encoder.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_stream_encoder.c
@@ -23,8 +23,6 @@
  *
  */
 
-#include <linux/delay.h>
-
 #include "dc_bios_types.h"
 #include "dce_stream_encoder.h"
 #include "reg_helper.h"
@@ -33,7 +31,6 @@
 #define DC_LOGGER \
 		enc110->base.ctx->logger
 
-
 #define REG(reg)\
 	(enc110->regs->reg)
 
@@ -136,7 +133,7 @@ static void dce110_update_generic_info_packet(
 			AFMT_GENERIC0_UPDATE, (packet_index == 0),
 			AFMT_GENERIC2_UPDATE, (packet_index == 2));
 	}
-#if defined(CONFIG_DRM_AMD_DC_DCN)
+
 	if (REG(AFMT_VBI_PACKET_CONTROL1)) {
 		switch (packet_index) {
 		case 0:
@@ -175,7 +172,6 @@ static void dce110_update_generic_info_packet(
 			break;
 		}
 	}
-#endif
 }
 
 static void dce110_update_hdmi_info_packet(
@@ -230,7 +226,6 @@ static void dce110_update_hdmi_info_packet(
 				HDMI_GENERIC1_SEND, send,
 				HDMI_GENERIC1_LINE, line);
 		break;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	case 4:
 		if (REG(HDMI_GENERIC_PACKET_CONTROL2))
 			REG_UPDATE_3(HDMI_GENERIC_PACKET_CONTROL2,
@@ -259,7 +254,6 @@ static void dce110_update_hdmi_info_packet(
 					HDMI_GENERIC1_SEND, send,
 					HDMI_GENERIC1_LINE, line);
 		break;
-#endif
 	default:
 		/* invalid HW packet index */
 		DC_LOG_WARNING(
@@ -277,18 +271,15 @@ static void dce110_stream_encoder_dp_set_stream_attribute(
 	bool use_vsc_sdp_for_colorimetry,
 	uint32_t enable_sdp_splitting)
 {
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	uint32_t h_active_start;
 	uint32_t v_active_start;
 	uint32_t misc0 = 0;
 	uint32_t misc1 = 0;
 	uint32_t h_blank;
 	uint32_t h_back_porch;
-	uint8_t synchronous_clock = 0; /* asynchronous mode */
 	uint8_t colorimetry_bpc;
 	uint8_t dynamic_range_rgb = 0; /*full range*/
 	uint8_t dynamic_range_ycbcr = 1; /*bt709*/
-#endif
 
 	struct dce110_stream_encoder *enc110 = DCE110STRENC_FROM_STRENC(enc);
 	struct dc_crtc_timing hw_crtc_timing = *crtc_timing;
@@ -329,10 +320,8 @@ static void dce110_stream_encoder_dp_set_stream_attribute(
 		if (enc110->se_mask->DP_VID_M_DOUBLE_VALUE_EN)
 			REG_UPDATE(DP_VID_TIMING, DP_VID_M_DOUBLE_VALUE_EN, 1);
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 		if (enc110->se_mask->DP_VID_N_MUL)
 			REG_UPDATE(DP_VID_TIMING, DP_VID_N_MUL, 1);
-#endif
 		break;
 	default:
 		REG_UPDATE(DP_PIXEL_FORMAT, DP_PIXEL_ENCODING,
@@ -340,10 +329,8 @@ static void dce110_stream_encoder_dp_set_stream_attribute(
 		break;
 	}
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	if (REG(DP_MSA_MISC))
 		misc1 = REG_READ(DP_MSA_MISC);
-#endif
 
 	/* set color depth */
 
@@ -374,7 +361,6 @@ static void dce110_stream_encoder_dp_set_stream_attribute(
 	/* set dynamic range and YCbCr range */
 
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	switch (hw_crtc_timing.display_color_depth) {
 	case COLOR_DEPTH_666:
 		colorimetry_bpc = 0;
@@ -393,7 +379,6 @@ static void dce110_stream_encoder_dp_set_stream_attribute(
 		break;
 	}
 
-	misc0 = misc0 | synchronous_clock;
 	misc0 = colorimetry_bpc << 5;
 
 	if (REG(DP_MSA_TIMING_PARAM1)) {
@@ -433,7 +418,7 @@ static void dce110_stream_encoder_dp_set_stream_attribute(
 			dynamic_range_rgb = 1; /*limited range*/
 			break;
 		case COLOR_SPACE_2020_RGB_FULLRANGE:
-		case COLOR_SPACE_2020_YCBCR:
+		case COLOR_SPACE_2020_YCBCR_LIMITED:
 		case COLOR_SPACE_XR_RGB:
 		case COLOR_SPACE_MSREF_SCRGB:
 		case COLOR_SPACE_ADOBERGB:
@@ -445,6 +430,7 @@ static void dce110_stream_encoder_dp_set_stream_attribute(
 		case COLOR_SPACE_APPCTRL:
 		case COLOR_SPACE_CUSTOMPOINTS:
 		case COLOR_SPACE_UNKNOWN:
+		default:
 			/* do nothing */
 			break;
 		}
@@ -454,7 +440,6 @@ static void dce110_stream_encoder_dp_set_stream_attribute(
 				DP_DYN_RANGE, dynamic_range_rgb,
 				DP_YCBCR_RANGE, dynamic_range_ycbcr);
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 		if (REG(DP_MSA_COLORIMETRY))
 			REG_SET(DP_MSA_COLORIMETRY, 0, DP_MSA_MISC0, misc0);
 
@@ -468,7 +453,6 @@ static void dce110_stream_encoder_dp_set_stream_attribute(
 			REG_SET_2(DP_MSA_TIMING_PARAM1, 0,
 					DP_MSA_HTOTAL, hw_crtc_timing.h_total,
 					DP_MSA_VTOTAL, hw_crtc_timing.v_total);
-#endif
 
 		/* calcuate from vesa timing parameters
 		 * h_active_start related to leading edge of sync
@@ -489,7 +473,6 @@ static void dce110_stream_encoder_dp_set_stream_attribute(
 				hw_crtc_timing.v_front_porch;
 
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 		/* start at begining of left border */
 		if (REG(DP_MSA_TIMING_PARAM2))
 			REG_SET_2(DP_MSA_TIMING_PARAM2, 0,
@@ -514,9 +497,7 @@ static void dce110_stream_encoder_dp_set_stream_attribute(
 				hw_crtc_timing.h_addressable + hw_crtc_timing.h_border_right,
 				DP_MSA_VHEIGHT, hw_crtc_timing.v_border_top +
 				hw_crtc_timing.v_addressable + hw_crtc_timing.v_border_bottom);
-#endif
 	}
-#endif
 }
 
 static void dce110_stream_encoder_set_stream_attribute_helper(
@@ -650,6 +631,8 @@ static void dce110_stream_encoder_hdmi_set_stream_attribute(
 		HDMI_GC_SEND, 1,
 		HDMI_NULL_SEND, 1);
 
+	REG_UPDATE(HDMI_VBI_PACKET_CONTROL, HDMI_ACP_SEND, 0);
+
 	/* following belongs to audio */
 	REG_UPDATE(HDMI_INFOFRAME_CONTROL0, HDMI_AUDIO_INFO_SEND, 1);
 
@@ -787,7 +770,6 @@ static void dce110_stream_encoder_update_hdmi_info_packets(
 		dce110_update_hdmi_info_packet(enc110, 3, &info_frame->hdrsmd);
 	}
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	if (enc110->se_mask->HDMI_DB_DISABLE) {
 		/* for bring up, disable dp double  TODO */
 		if (REG(HDMI_DB_CONTROL))
@@ -799,7 +781,6 @@ static void dce110_stream_encoder_update_hdmi_info_packets(
 		dce110_update_hdmi_info_packet(enc110, 3, &info_frame->spd);
 		dce110_update_hdmi_info_packet(enc110, 4, &info_frame->hdrsmd);
 	}
-#endif
 }
 
 static void dce110_stream_encoder_stop_hdmi_info_packets(
@@ -825,7 +806,6 @@ static void dce110_stream_encoder_stop_hdmi_info_packets(
 		HDMI_GENERIC1_LINE, 0,
 		HDMI_GENERIC1_SEND, 0);
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	/* stop generic packets 2 & 3 on HDMI */
 	if (REG(HDMI_GENERIC_PACKET_CONTROL2))
 		REG_SET_6(HDMI_GENERIC_PACKET_CONTROL2, 0,
@@ -844,7 +824,6 @@ static void dce110_stream_encoder_stop_hdmi_info_packets(
 			HDMI_GENERIC1_CONT, 0,
 			HDMI_GENERIC1_LINE, 0,
 			HDMI_GENERIC1_SEND, 0);
-#endif
 }
 
 static void dce110_stream_encoder_update_dp_info_packets(
@@ -1045,6 +1024,7 @@ static void dce110_reset_hdmi_stream_attribute(
 	struct stream_encoder *enc)
 {
 	struct dce110_stream_encoder *enc110 = DCE110STRENC_FROM_STRENC(enc);
+
 	if (enc110->se_mask->HDMI_DATA_SCRAMBLE_EN)
 		REG_UPDATE_5(HDMI_CONTROL,
 			HDMI_PACKET_GEN_VERSION, 1,
@@ -1587,3 +1567,17 @@ void dce110_stream_encoder_construct(
 	enc110->se_shift = se_shift;
 	enc110->se_mask = se_mask;
 }
+
+static const struct stream_encoder_funcs dce110_an_str_enc_funcs = {};
+
+void dce110_analog_stream_encoder_construct(
+	struct dce110_stream_encoder *enc110,
+	struct dc_context *ctx,
+	struct dc_bios *bp,
+	enum engine_id eng_id)
+{
+	enc110->base.funcs = &dce110_an_str_enc_funcs;
+	enc110->base.ctx = ctx;
+	enc110->base.id = eng_id;
+	enc110->base.bp = bp;
+}
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_stream_encoder.h b/drivers/gpu/drm/amd/display/dc/dce/dce_stream_encoder.h
index f9cdf2b5242c..068de1392121 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_stream_encoder.h
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_stream_encoder.h
@@ -115,7 +115,7 @@
 #define SE_SF(reg_name, field_name, post_fix)\
 	.field_name = reg_name ## __ ## field_name ## post_fix
 
-#define SE_COMMON_MASK_SH_LIST_DCE_COMMON_BASE(mask_sh)\
+#define SE_COMMON_MASK_SH_LIST_DCE_COMMON(mask_sh)\
 	SE_SF(AFMT_VBI_PACKET_CONTROL, AFMT_GENERIC_INDEX, mask_sh),\
 	SE_SF(AFMT_VBI_PACKET_CONTROL, AFMT_GENERIC0_UPDATE, mask_sh),\
 	SE_SF(AFMT_VBI_PACKET_CONTROL, AFMT_GENERIC2_UPDATE, mask_sh),\
@@ -140,6 +140,7 @@
 	SE_SF(HDMI_VBI_PACKET_CONTROL, HDMI_GC_CONT, mask_sh),\
 	SE_SF(HDMI_VBI_PACKET_CONTROL, HDMI_GC_SEND, mask_sh),\
 	SE_SF(HDMI_VBI_PACKET_CONTROL, HDMI_NULL_SEND, mask_sh),\
+	SE_SF(HDMI_VBI_PACKET_CONTROL, HDMI_ACP_SEND, mask_sh),\
 	SE_SF(HDMI_INFOFRAME_CONTROL0, HDMI_AUDIO_INFO_SEND, mask_sh),\
 	SE_SF(AFMT_INFOFRAME_CONTROL0, AFMT_AUDIO_INFO_UPDATE, mask_sh),\
 	SE_SF(HDMI_INFOFRAME_CONTROL1, HDMI_AUDIO_INFO_LINE, mask_sh),\
@@ -202,10 +203,7 @@
 	SE_SF(AFMT_AUDIO_PACKET_CONTROL, AFMT_AUDIO_SAMPLE_SEND, mask_sh),\
 	SE_SF(DIG_FE_CNTL, DIG_SOURCE_SELECT, mask_sh)
 
-#define SE_COMMON_MASK_SH_LIST_DCE_COMMON(mask_sh)\
-	SE_COMMON_MASK_SH_LIST_DCE_COMMON_BASE(mask_sh)
-
-#define SE_COMMON_MASK_SH_LIST_SOC_BASE(mask_sh)\
+#define SE_COMMON_MASK_SH_LIST_SOC(mask_sh)\
 	SE_SF(DIG0_AFMT_VBI_PACKET_CONTROL, AFMT_GENERIC_INDEX, mask_sh),\
 	SE_SF(DIG0_AFMT_GENERIC_HDR, AFMT_GENERIC_HB0, mask_sh),\
 	SE_SF(DIG0_AFMT_GENERIC_HDR, AFMT_GENERIC_HB1, mask_sh),\
@@ -227,6 +225,7 @@
 	SE_SF(DIG0_HDMI_VBI_PACKET_CONTROL, HDMI_GC_CONT, mask_sh),\
 	SE_SF(DIG0_HDMI_VBI_PACKET_CONTROL, HDMI_GC_SEND, mask_sh),\
 	SE_SF(DIG0_HDMI_VBI_PACKET_CONTROL, HDMI_NULL_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_VBI_PACKET_CONTROL, HDMI_ACP_SEND, mask_sh),\
 	SE_SF(DIG0_HDMI_INFOFRAME_CONTROL0, HDMI_AUDIO_INFO_SEND, mask_sh),\
 	SE_SF(DIG0_AFMT_INFOFRAME_CONTROL0, AFMT_AUDIO_INFO_UPDATE, mask_sh),\
 	SE_SF(DIG0_HDMI_INFOFRAME_CONTROL1, HDMI_AUDIO_INFO_LINE, mask_sh),\
@@ -288,9 +287,6 @@
 	SE_SF(DIG0_DIG_FE_CNTL, DIG_STEREOSYNC_GATE_EN, mask_sh),\
 	SE_SF(DIG0_DIG_FE_CNTL, DIG_SOURCE_SELECT, mask_sh)
 
-#define SE_COMMON_MASK_SH_LIST_SOC(mask_sh)\
-	SE_COMMON_MASK_SH_LIST_SOC_BASE(mask_sh)
-
 #define SE_COMMON_MASK_SH_LIST_DCE80_100(mask_sh)\
 	SE_COMMON_MASK_SH_LIST_DCE_COMMON(mask_sh),\
 	SE_SF(TMDS_CNTL, TMDS_PIXEL_ENCODING, mask_sh),\
@@ -414,6 +410,7 @@ struct dce_stream_encoder_shift {
 	uint8_t HDMI_GC_SEND;
 	uint8_t HDMI_NULL_SEND;
 	uint8_t HDMI_DATA_SCRAMBLE_EN;
+	uint8_t HDMI_ACP_SEND;
 	uint8_t HDMI_AUDIO_INFO_SEND;
 	uint8_t AFMT_AUDIO_INFO_UPDATE;
 	uint8_t HDMI_AUDIO_INFO_LINE;
@@ -545,6 +542,7 @@ struct dce_stream_encoder_mask {
 	uint32_t HDMI_GC_SEND;
 	uint32_t HDMI_NULL_SEND;
 	uint32_t HDMI_DATA_SCRAMBLE_EN;
+	uint32_t HDMI_ACP_SEND;
 	uint32_t HDMI_AUDIO_INFO_SEND;
 	uint32_t AFMT_AUDIO_INFO_UPDATE;
 	uint32_t HDMI_AUDIO_INFO_LINE;
@@ -710,6 +708,11 @@ void dce110_stream_encoder_construct(
 	const struct dce_stream_encoder_shift *se_shift,
 	const struct dce_stream_encoder_mask *se_mask);
 
+void dce110_analog_stream_encoder_construct(
+	struct dce110_stream_encoder *enc110,
+	struct dc_context *ctx,
+	struct dc_bios *bp,
+	enum engine_id eng_id);
 
 void dce110_se_audio_mute_control(
 	struct stream_encoder *enc, bool mute);
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_transform.c b/drivers/gpu/drm/amd/display/dc/dce/dce_transform.c
index d9fd4ec60588..1ab5ae9b5ea5 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_transform.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_transform.c
@@ -154,10 +154,13 @@ static bool dce60_setup_scaling_configuration(
 	REG_SET(SCL_BYPASS_CONTROL, 0, SCL_BYPASS_MODE, 0);
 
 	if (data->taps.h_taps + data->taps.v_taps <= 2) {
-		/* Set bypass */
-
-		/* DCE6 has no SCL_MODE register, skip scale mode programming */
+		/* Disable scaler functionality */
+		REG_WRITE(SCL_SCALER_ENABLE, 0);
 
+		/* Clear registers that can cause glitches even when the scaler is off */
+		REG_WRITE(SCL_TAP_CONTROL, 0);
+		REG_WRITE(SCL_AUTOMATIC_MODE_CONTROL, 0);
+		REG_WRITE(SCL_F_SHARP_CONTROL, 0);
 		return false;
 	}
 
@@ -165,7 +168,7 @@ static bool dce60_setup_scaling_configuration(
 			SCL_H_NUM_OF_TAPS, data->taps.h_taps - 1,
 			SCL_V_NUM_OF_TAPS, data->taps.v_taps - 1);
 
-	/* DCE6 has no SCL_MODE register, skip scale mode programming */
+	REG_WRITE(SCL_SCALER_ENABLE, 1);
 
 	/* DCE6 has no SCL_BOUNDARY_MODE bit, skip replace out of bound pixels */
 
@@ -502,6 +505,8 @@ static void dce60_transform_set_scaler(
 	REG_SET(DC_LB_MEM_SIZE, 0,
 		DC_LB_MEM_SIZE, xfm_dce->lb_memory_size);
 
+	REG_WRITE(SCL_UPDATE, 0x00010000);
+
 	/* Clear SCL_F_SHARP_CONTROL value to 0 */
 	REG_WRITE(SCL_F_SHARP_CONTROL, 0);
 
@@ -527,8 +532,7 @@ static void dce60_transform_set_scaler(
 		if (coeffs_v != xfm_dce->filter_v || coeffs_h != xfm_dce->filter_h) {
 			/* 4. Program vertical filters */
 			if (xfm_dce->filter_v == NULL)
-				REG_SET(SCL_VERT_FILTER_CONTROL, 0,
-						SCL_V_2TAP_HARDCODE_COEF_EN, 0);
+				REG_WRITE(SCL_VERT_FILTER_CONTROL, 0);
 			program_multi_taps_filter(
 					xfm_dce,
 					data->taps.v_taps,
@@ -542,8 +546,7 @@ static void dce60_transform_set_scaler(
 
 			/* 5. Program horizontal filters */
 			if (xfm_dce->filter_h == NULL)
-				REG_SET(SCL_HORZ_FILTER_CONTROL, 0,
-						SCL_H_2TAP_HARDCODE_COEF_EN, 0);
+				REG_WRITE(SCL_HORZ_FILTER_CONTROL, 0);
 			program_multi_taps_filter(
 					xfm_dce,
 					data->taps.h_taps,
@@ -566,6 +569,8 @@ static void dce60_transform_set_scaler(
 	/* DCE6 has no SCL_COEF_UPDATE_COMPLETE bit to flip to new coefficient memory */
 
 	/* DCE6 DATA_FORMAT register does not support ALPHA_EN */
+
+	REG_WRITE(SCL_UPDATE, 0);
 }
 #endif
 
@@ -1009,7 +1014,7 @@ static void dce_transform_set_pixel_storage_depth(
 		color_depth = COLOR_DEPTH_101010;
 		pixel_depth = 0;
 		expan_mode  = 1;
-		BREAK_TO_DEBUGGER();
+		DC_LOG_DC("The pixel depth %d is not valid, set COLOR_DEPTH_101010 instead.", depth);
 		break;
 	}
 
@@ -1023,8 +1028,7 @@ static void dce_transform_set_pixel_storage_depth(
 	if (!(xfm_dce->lb_pixel_depth_supported & depth)) {
 		/*we should use unsupported capabilities
 		 *  unless it is required by w/a*/
-		DC_LOG_WARNING("%s: Capability not supported",
-			__func__);
+		DC_LOG_DC("%s: Capability not supported", __func__);
 	}
 }
 
@@ -1409,7 +1413,7 @@ void dce110_opp_set_csc_default(
 static void program_pwl(struct dce_transform *xfm_dce,
 			const struct pwl_params *params)
 {
-	int retval;
+	uint32_t retval;
 	uint8_t max_tries = 10;
 	uint8_t counter = 0;
 	uint32_t i = 0;
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_transform.h b/drivers/gpu/drm/amd/display/dc/dce/dce_transform.h
index cbce194ec7b8..eb716e8337e2 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_transform.h
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_transform.h
@@ -155,6 +155,9 @@
 	SRI(SCL_COEF_RAM_TAP_DATA, SCL, id), \
 	SRI(VIEWPORT_START, SCL, id), \
 	SRI(VIEWPORT_SIZE, SCL, id), \
+	SRI(SCL_SCALER_ENABLE, SCL, id), \
+	SRI(SCL_HORZ_FILTER_INIT_RGB_LUMA, SCL, id), \
+	SRI(SCL_HORZ_FILTER_INIT_CHROMA, SCL, id), \
 	SRI(SCL_HORZ_FILTER_SCALE_RATIO, SCL, id), \
 	SRI(SCL_VERT_FILTER_SCALE_RATIO, SCL, id), \
 	SRI(SCL_VERT_FILTER_INIT, SCL, id), \
@@ -590,6 +593,7 @@ struct dce_transform_registers {
 	uint32_t SCL_VERT_FILTER_SCALE_RATIO;
 	uint32_t SCL_HORZ_FILTER_INIT;
 #if defined(CONFIG_DRM_AMD_DC_SI)
+	uint32_t SCL_SCALER_ENABLE;
 	uint32_t SCL_HORZ_FILTER_INIT_RGB_LUMA;
 	uint32_t SCL_HORZ_FILTER_INIT_CHROMA;
 #endif
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dmub_abm.c b/drivers/gpu/drm/amd/display/dc/dce/dmub_abm.c
index fb0dec4ed3a6..3b9011ef9b68 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dmub_abm.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dmub_abm.c
@@ -24,212 +24,177 @@
  */
 
 #include "dmub_abm.h"
-#include "dce_abm.h"
+#include "dmub_abm_lcd.h"
 #include "dc.h"
+#include "core_types.h"
+#include "dmub_cmd.h"
 #include "dc_dmub_srv.h"
 #include "dmub/dmub_srv.h"
-#include "core_types.h"
-#include "dm_services.h"
-#include "reg_helper.h"
-#include "fixed31_32.h"
-
-#include "atom.h"
 
 #define TO_DMUB_ABM(abm)\
 	container_of(abm, struct dce_abm, base)
 
-#define REG(reg) \
-	(dce_abm->regs->reg)
+#define ABM_FEATURE_NO_SUPPORT	0
+#define ABM_LCD_SUPPORT			1
 
-#undef FN
-#define FN(reg_name, field_name) \
-	dce_abm->abm_shift->field_name, dce_abm->abm_mask->field_name
+static unsigned int abm_feature_support(struct abm *abm, unsigned int panel_inst)
+{
+	struct dc_context *dc = abm->ctx;
+	struct dc_link *edp_links[MAX_NUM_EDP];
+	int i;
+	int edp_num;
+	unsigned int ret = ABM_FEATURE_NO_SUPPORT;
 
-#define CTX \
-	dce_abm->base.ctx
+	dc_get_edp_links(dc->dc, edp_links, &edp_num);
 
-#define DISABLE_ABM_IMMEDIATELY 255
+	for (i = 0; i < edp_num; i++) {
+		if (panel_inst == i)
+			break;
+	}
 
+	if (i < edp_num) {
+		ret = ABM_LCD_SUPPORT;
+	}
 
+	return ret;
+}
 
-static void dmub_abm_enable_fractional_pwm(struct dc_context *dc)
+static void dmub_abm_init_ex(struct abm *abm, uint32_t backlight, uint32_t user_level)
 {
-	union dmub_rb_cmd cmd;
-	uint32_t fractional_pwm = (dc->dc->config.disable_fractional_pwm == false) ? 1 : 0;
-	uint32_t edp_id_count = dc->dc_edp_id_count;
-	int i;
-	uint8_t panel_mask = 0;
-
-	for (i = 0; i < edp_id_count; i++)
-		panel_mask |= 0x01 << i;
-
-	memset(&cmd, 0, sizeof(cmd));
-	cmd.abm_set_pwm_frac.header.type = DMUB_CMD__ABM;
-	cmd.abm_set_pwm_frac.header.sub_type = DMUB_CMD__ABM_SET_PWM_FRAC;
-	cmd.abm_set_pwm_frac.abm_set_pwm_frac_data.fractional_pwm = fractional_pwm;
-	cmd.abm_set_pwm_frac.abm_set_pwm_frac_data.version = DMUB_CMD_ABM_CONTROL_VERSION_1;
-	cmd.abm_set_pwm_frac.abm_set_pwm_frac_data.panel_mask = panel_mask;
-	cmd.abm_set_pwm_frac.header.payload_bytes = sizeof(struct dmub_cmd_abm_set_pwm_frac_data);
-
-	dc_dmub_srv_cmd_queue(dc->dmub_srv, &cmd);
-	dc_dmub_srv_cmd_execute(dc->dmub_srv);
-	dc_dmub_srv_wait_idle(dc->dmub_srv);
+	dmub_abm_init(abm, backlight, user_level);
 }
 
-static void dmub_abm_init(struct abm *abm, uint32_t backlight)
+static unsigned int dmub_abm_get_current_backlight_ex(struct abm *abm)
 {
-	struct dce_abm *dce_abm = TO_DMUB_ABM(abm);
-
-	REG_WRITE(DC_ABM1_HG_SAMPLE_RATE, 0x3);
-	REG_WRITE(DC_ABM1_HG_SAMPLE_RATE, 0x1);
-	REG_WRITE(DC_ABM1_LS_SAMPLE_RATE, 0x3);
-	REG_WRITE(DC_ABM1_LS_SAMPLE_RATE, 0x1);
-	REG_WRITE(BL1_PWM_BL_UPDATE_SAMPLE_RATE, 0x1);
+	dc_allow_idle_optimizations(abm->ctx->dc, false);
 
-	REG_SET_3(DC_ABM1_HG_MISC_CTRL, 0,
-			ABM1_HG_NUM_OF_BINS_SEL, 0,
-			ABM1_HG_VMAX_SEL, 1,
-			ABM1_HG_BIN_BITWIDTH_SIZE_SEL, 0);
+	return dmub_abm_get_current_backlight(abm);
+}
 
-	REG_SET_3(DC_ABM1_IPCSC_COEFF_SEL, 0,
-			ABM1_IPCSC_COEFF_SEL_R, 2,
-			ABM1_IPCSC_COEFF_SEL_G, 4,
-			ABM1_IPCSC_COEFF_SEL_B, 2);
+static unsigned int dmub_abm_get_target_backlight_ex(struct abm *abm)
+{
+	dc_allow_idle_optimizations(abm->ctx->dc, false);
 
-	REG_UPDATE(BL1_PWM_CURRENT_ABM_LEVEL,
-			BL1_PWM_CURRENT_ABM_LEVEL, backlight);
+	return dmub_abm_get_target_backlight(abm);
+}
 
-	REG_UPDATE(BL1_PWM_TARGET_ABM_LEVEL,
-			BL1_PWM_TARGET_ABM_LEVEL, backlight);
+static bool dmub_abm_set_level_ex(struct abm *abm, uint32_t level)
+{
+	bool ret = false;
+	unsigned int feature_support, i;
+	uint8_t panel_mask0 = 0;
 
-	REG_UPDATE(BL1_PWM_USER_LEVEL,
-			BL1_PWM_USER_LEVEL, backlight);
+	for (i = 0; i < MAX_NUM_EDP; i++) {
+		feature_support = abm_feature_support(abm, i);
 
-	REG_UPDATE_2(DC_ABM1_LS_MIN_MAX_PIXEL_VALUE_THRES,
-			ABM1_LS_MIN_PIXEL_VALUE_THRES, 0,
-			ABM1_LS_MAX_PIXEL_VALUE_THRES, 1000);
+		if (feature_support == ABM_LCD_SUPPORT)
+			panel_mask0 |= (0x01 << i);
+	}
 
-	REG_SET_3(DC_ABM1_HGLS_REG_READ_PROGRESS, 0,
-			ABM1_HG_REG_READ_MISSED_FRAME_CLEAR, 1,
-			ABM1_LS_REG_READ_MISSED_FRAME_CLEAR, 1,
-			ABM1_BL_REG_READ_MISSED_FRAME_CLEAR, 1);
+	if (panel_mask0)
+		ret = dmub_abm_set_level(abm, level, panel_mask0);
 
-	dmub_abm_enable_fractional_pwm(abm->ctx);
+	return ret;
 }
 
-static unsigned int dmub_abm_get_current_backlight(struct abm *abm)
+static bool dmub_abm_init_config_ex(struct abm *abm,
+	const char *src,
+	unsigned int bytes,
+	unsigned int inst)
 {
-	struct dce_abm *dce_abm = TO_DMUB_ABM(abm);
-	unsigned int backlight = REG_READ(BL1_PWM_CURRENT_ABM_LEVEL);
+	unsigned int feature_support;
 
-	/* return backlight in hardware format which is unsigned 17 bits, with
-	 * 1 bit integer and 16 bit fractional
-	 */
-	return backlight;
-}
+	feature_support = abm_feature_support(abm, inst);
 
-static unsigned int dmub_abm_get_target_backlight(struct abm *abm)
-{
-	struct dce_abm *dce_abm = TO_DMUB_ABM(abm);
-	unsigned int backlight = REG_READ(BL1_PWM_TARGET_ABM_LEVEL);
+	if (feature_support == ABM_LCD_SUPPORT)
+		dmub_abm_init_config(abm, src, bytes, inst);
 
-	/* return backlight in hardware format which is unsigned 17 bits, with
-	 * 1 bit integer and 16 bit fractional
-	 */
-	return backlight;
+	return true;
 }
 
-static bool dmub_abm_set_level(struct abm *abm, uint32_t level)
+static bool dmub_abm_set_pause_ex(struct abm *abm, bool pause, unsigned int panel_inst, unsigned int stream_inst)
 {
-	union dmub_rb_cmd cmd;
-	struct dc_context *dc = abm->ctx;
-	struct dc_link *edp_links[MAX_NUM_EDP];
-	int i;
-	int edp_num;
-	uint8_t panel_mask = 0;
-
-	get_edp_links(dc->dc, edp_links, &edp_num);
+	bool ret = false;
+	unsigned int feature_support;
 
-	for (i = 0; i < edp_num; i++) {
-		if (edp_links[i]->link_status.link_active)
-			panel_mask |= (0x01 << i);
-	}
-
-	memset(&cmd, 0, sizeof(cmd));
-	cmd.abm_set_level.header.type = DMUB_CMD__ABM;
-	cmd.abm_set_level.header.sub_type = DMUB_CMD__ABM_SET_LEVEL;
-	cmd.abm_set_level.abm_set_level_data.level = level;
-	cmd.abm_set_level.abm_set_level_data.version = DMUB_CMD_ABM_CONTROL_VERSION_1;
-	cmd.abm_set_level.abm_set_level_data.panel_mask = panel_mask;
-	cmd.abm_set_level.header.payload_bytes = sizeof(struct dmub_cmd_abm_set_level_data);
+	feature_support = abm_feature_support(abm, panel_inst);
 
-	dc_dmub_srv_cmd_queue(dc->dmub_srv, &cmd);
-	dc_dmub_srv_cmd_execute(dc->dmub_srv);
-	dc_dmub_srv_wait_idle(dc->dmub_srv);
+	if (feature_support == ABM_LCD_SUPPORT)
+		ret = dmub_abm_set_pause(abm, pause, panel_inst, stream_inst);
 
-	return true;
+	return ret;
 }
 
-static bool dmub_abm_init_config(struct abm *abm,
-	const char *src,
-	unsigned int bytes,
-	unsigned int inst)
+/*****************************************************************************
+ *  dmub_abm_save_restore_ex() - calls dmub_abm_save_restore for preserving DMUB's
+ *                              Varibright states for LCD only. OLED is TBD
+ *  @abm: used to check get dc context
+ *  @panel_inst: panel instance index
+ *  @pData: contains command to pause/un-pause abm and abm parameters
+ *
+ *
+ ***************************************************************************/
+static bool dmub_abm_save_restore_ex(
+		struct abm *abm,
+		unsigned int panel_inst,
+		struct abm_save_restore *pData)
 {
-	union dmub_rb_cmd cmd;
+	bool ret = false;
+	unsigned int feature_support;
 	struct dc_context *dc = abm->ctx;
-	uint8_t panel_mask = 0x01 << inst;
 
-	// TODO: Optimize by only reading back final 4 bytes
-	dmub_flush_buffer_mem(&dc->dmub_srv->dmub->scratch_mem_fb);
+	feature_support = abm_feature_support(abm, panel_inst);
 
-	// Copy iramtable into cw7
-	memcpy(dc->dmub_srv->dmub->scratch_mem_fb.cpu_addr, (void *)src, bytes);
+	if (feature_support == ABM_LCD_SUPPORT)
+		ret = dmub_abm_save_restore(dc, panel_inst, pData);
 
-	memset(&cmd, 0, sizeof(cmd));
-	// Fw will copy from cw7 to fw_state
-	cmd.abm_init_config.header.type = DMUB_CMD__ABM;
-	cmd.abm_init_config.header.sub_type = DMUB_CMD__ABM_INIT_CONFIG;
-	cmd.abm_init_config.abm_init_config_data.src.quad_part = dc->dmub_srv->dmub->scratch_mem_fb.gpu_addr;
-	cmd.abm_init_config.abm_init_config_data.bytes = bytes;
-	cmd.abm_init_config.abm_init_config_data.version = DMUB_CMD_ABM_CONTROL_VERSION_1;
-	cmd.abm_init_config.abm_init_config_data.panel_mask = panel_mask;
+	return ret;
+}
 
-	cmd.abm_init_config.header.payload_bytes = sizeof(struct dmub_cmd_abm_init_config_data);
+static bool dmub_abm_set_pipe_ex(struct abm *abm,
+		uint32_t otg_inst,
+		uint32_t option,
+		uint32_t panel_inst,
+		uint32_t pwrseq_inst)
+{
+	bool ret = false;
+	unsigned int feature_support;
 
-	dc_dmub_srv_cmd_queue(dc->dmub_srv, &cmd);
-	dc_dmub_srv_cmd_execute(dc->dmub_srv);
-	dc_dmub_srv_wait_idle(dc->dmub_srv);
+	feature_support = abm_feature_support(abm, panel_inst);
 
-	return true;
+	if (feature_support == ABM_LCD_SUPPORT)
+		ret = dmub_abm_set_pipe(abm, otg_inst, option, panel_inst, pwrseq_inst);
+
+	return ret;
 }
 
-static bool dmub_abm_set_pause(struct abm *abm, bool pause, unsigned int panel_inst, unsigned int stream_inst)
+static bool dmub_abm_set_backlight_level_pwm_ex(struct abm *abm,
+		unsigned int backlight_pwm_u16_16,
+		unsigned int frame_ramp,
+		unsigned int controller_id,
+		unsigned int panel_inst)
 {
-	union dmub_rb_cmd cmd;
-	struct dc_context *dc = abm->ctx;
-	uint8_t panel_mask = 0x01 << panel_inst;
+	bool ret = false;
+	unsigned int feature_support;
 
-	memset(&cmd, 0, sizeof(cmd));
-	cmd.abm_pause.header.type = DMUB_CMD__ABM;
-	cmd.abm_pause.header.sub_type = DMUB_CMD__ABM_PAUSE;
-	cmd.abm_pause.abm_pause_data.enable = pause;
-	cmd.abm_pause.abm_pause_data.panel_mask = panel_mask;
-	cmd.abm_set_level.header.payload_bytes = sizeof(struct dmub_cmd_abm_pause_data);
+	feature_support = abm_feature_support(abm, panel_inst);
 
-	dc_dmub_srv_cmd_queue(dc->dmub_srv, &cmd);
-	dc_dmub_srv_cmd_execute(dc->dmub_srv);
-	dc_dmub_srv_wait_idle(dc->dmub_srv);
+	if (feature_support == ABM_LCD_SUPPORT)
+		ret = dmub_abm_set_backlight_level(abm, backlight_pwm_u16_16, frame_ramp, panel_inst);
 
-	return true;
+	return ret;
 }
 
 static const struct abm_funcs abm_funcs = {
-	.abm_init = dmub_abm_init,
-	.set_abm_level = dmub_abm_set_level,
-	.get_current_backlight = dmub_abm_get_current_backlight,
-	.get_target_backlight = dmub_abm_get_target_backlight,
-	.init_abm_config = dmub_abm_init_config,
-	.set_abm_pause = dmub_abm_set_pause,
+	.abm_init = dmub_abm_init_ex,
+	.set_abm_level = dmub_abm_set_level_ex,
+	.get_current_backlight = dmub_abm_get_current_backlight_ex,
+	.get_target_backlight = dmub_abm_get_target_backlight_ex,
+	.init_abm_config = dmub_abm_init_config_ex,
+	.set_abm_pause = dmub_abm_set_pause_ex,
+	.save_restore = dmub_abm_save_restore_ex,
+	.set_pipe_ex = dmub_abm_set_pipe_ex,
+	.set_backlight_level_pwm = dmub_abm_set_backlight_level_pwm_ex,
 };
 
 static void dmub_abm_construct(
@@ -256,16 +221,19 @@ struct abm *dmub_abm_create(
 	const struct dce_abm_shift *abm_shift,
 	const struct dce_abm_mask *abm_mask)
 {
-	struct dce_abm *abm_dce = kzalloc(sizeof(*abm_dce), GFP_KERNEL);
+	if (ctx->dc->caps.dmcub_support) {
+		struct dce_abm *abm_dce = kzalloc(sizeof(*abm_dce), GFP_KERNEL);
 
-	if (abm_dce == NULL) {
-		BREAK_TO_DEBUGGER();
-		return NULL;
-	}
+		if (abm_dce == NULL) {
+			BREAK_TO_DEBUGGER();
+			return NULL;
+		}
 
-	dmub_abm_construct(abm_dce, ctx, regs, abm_shift, abm_mask);
+		dmub_abm_construct(abm_dce, ctx, regs, abm_shift, abm_mask);
 
-	return &abm_dce->base;
+		return &abm_dce->base;
+	}
+	return NULL;
 }
 
 void dmub_abm_destroy(struct abm **abm)
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dmub_abm_lcd.c b/drivers/gpu/drm/amd/display/dc/dce/dmub_abm_lcd.c
new file mode 100644
index 000000000000..a641ae04450c
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dce/dmub_abm_lcd.c
@@ -0,0 +1,318 @@
+/*
+ * Copyright 2019 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#include "dmub_abm.h"
+#include "dmub_abm_lcd.h"
+#include "dce_abm.h"
+#include "dc.h"
+#include "dc_dmub_srv.h"
+#include "dmub/dmub_srv.h"
+#include "core_types.h"
+#include "dm_services.h"
+#include "reg_helper.h"
+#include "fixed31_32.h"
+
+#include "atom.h"
+
+#define TO_DMUB_ABM(abm)\
+	container_of(abm, struct dce_abm, base)
+
+#define REG(reg) \
+	(dce_abm->regs->reg)
+
+#undef FN
+#define FN(reg_name, field_name) \
+	dce_abm->abm_shift->field_name, dce_abm->abm_mask->field_name
+
+#define CTX \
+	dce_abm->base.ctx
+
+#define DISABLE_ABM_IMMEDIATELY 255
+
+
+
+static void dmub_abm_enable_fractional_pwm(struct dc_context *dc)
+{
+	union dmub_rb_cmd cmd;
+	uint32_t fractional_pwm = (dc->dc->config.disable_fractional_pwm == false) ? 1 : 0;
+	uint32_t edp_id_count = dc->dc_edp_id_count;
+	int i;
+	uint8_t panel_mask = 0;
+
+	for (i = 0; i < edp_id_count; i++)
+		panel_mask |= 0x01 << i;
+
+	memset(&cmd, 0, sizeof(cmd));
+	cmd.abm_set_pwm_frac.header.type = DMUB_CMD__ABM;
+	cmd.abm_set_pwm_frac.header.sub_type = DMUB_CMD__ABM_SET_PWM_FRAC;
+	cmd.abm_set_pwm_frac.abm_set_pwm_frac_data.fractional_pwm = fractional_pwm;
+	cmd.abm_set_pwm_frac.abm_set_pwm_frac_data.version = DMUB_CMD_ABM_CONTROL_VERSION_1;
+	cmd.abm_set_pwm_frac.abm_set_pwm_frac_data.panel_mask = panel_mask;
+	cmd.abm_set_pwm_frac.header.payload_bytes = sizeof(struct dmub_cmd_abm_set_pwm_frac_data);
+
+	dc_wake_and_execute_dmub_cmd(dc, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
+}
+
+void dmub_abm_init(struct abm *abm, uint32_t backlight, uint32_t user_level)
+{
+	struct dce_abm *dce_abm = TO_DMUB_ABM(abm);
+
+	REG_WRITE(DC_ABM1_HG_SAMPLE_RATE, 0x3);
+	REG_WRITE(DC_ABM1_HG_SAMPLE_RATE, 0x1);
+	REG_WRITE(DC_ABM1_LS_SAMPLE_RATE, 0x3);
+	REG_WRITE(DC_ABM1_LS_SAMPLE_RATE, 0x1);
+	REG_WRITE(BL1_PWM_BL_UPDATE_SAMPLE_RATE, 0x1);
+
+	REG_SET_3(DC_ABM1_HG_MISC_CTRL, 0,
+			ABM1_HG_NUM_OF_BINS_SEL, 0,
+			ABM1_HG_VMAX_SEL, 1,
+			ABM1_HG_BIN_BITWIDTH_SIZE_SEL, 0);
+
+	REG_SET_3(DC_ABM1_IPCSC_COEFF_SEL, 0,
+			ABM1_IPCSC_COEFF_SEL_R, 2,
+			ABM1_IPCSC_COEFF_SEL_G, 4,
+			ABM1_IPCSC_COEFF_SEL_B, 2);
+
+	REG_UPDATE(BL1_PWM_CURRENT_ABM_LEVEL,
+			BL1_PWM_CURRENT_ABM_LEVEL, backlight);
+
+	REG_UPDATE(BL1_PWM_TARGET_ABM_LEVEL,
+			BL1_PWM_TARGET_ABM_LEVEL, backlight);
+
+	REG_UPDATE(BL1_PWM_USER_LEVEL,
+			BL1_PWM_USER_LEVEL, user_level);
+
+	REG_UPDATE_2(DC_ABM1_LS_MIN_MAX_PIXEL_VALUE_THRES,
+			ABM1_LS_MIN_PIXEL_VALUE_THRES, 0,
+			ABM1_LS_MAX_PIXEL_VALUE_THRES, 1000);
+
+	REG_SET_3(DC_ABM1_HGLS_REG_READ_PROGRESS, 0,
+			ABM1_HG_REG_READ_MISSED_FRAME_CLEAR, 1,
+			ABM1_LS_REG_READ_MISSED_FRAME_CLEAR, 1,
+			ABM1_BL_REG_READ_MISSED_FRAME_CLEAR, 1);
+
+	dmub_abm_enable_fractional_pwm(abm->ctx);
+}
+
+unsigned int dmub_abm_get_current_backlight(struct abm *abm)
+{
+	struct dce_abm *dce_abm = TO_DMUB_ABM(abm);
+	unsigned int backlight = REG_READ(BL1_PWM_CURRENT_ABM_LEVEL);
+
+	/* return backlight in hardware format which is unsigned 17 bits, with
+	 * 1 bit integer and 16 bit fractional
+	 */
+	return backlight;
+}
+
+unsigned int dmub_abm_get_target_backlight(struct abm *abm)
+{
+	struct dce_abm *dce_abm = TO_DMUB_ABM(abm);
+	unsigned int backlight = REG_READ(BL1_PWM_TARGET_ABM_LEVEL);
+
+	/* return backlight in hardware format which is unsigned 17 bits, with
+	 * 1 bit integer and 16 bit fractional
+	 */
+	return backlight;
+}
+
+bool dmub_abm_set_level(struct abm *abm, uint32_t level, uint8_t panel_mask)
+{
+	union dmub_rb_cmd cmd;
+	struct dc_context *dc = abm->ctx;
+
+	memset(&cmd, 0, sizeof(cmd));
+	cmd.abm_set_level.header.type = DMUB_CMD__ABM;
+	cmd.abm_set_level.header.sub_type = DMUB_CMD__ABM_SET_LEVEL;
+	cmd.abm_set_level.abm_set_level_data.level = level;
+	cmd.abm_set_level.abm_set_level_data.version = DMUB_CMD_ABM_CONTROL_VERSION_1;
+	cmd.abm_set_level.abm_set_level_data.panel_mask = panel_mask;
+	cmd.abm_set_level.header.payload_bytes = sizeof(struct dmub_cmd_abm_set_level_data);
+
+	dc_wake_and_execute_dmub_cmd(dc, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
+
+	return true;
+}
+
+void dmub_abm_init_config(struct abm *abm,
+	const char *src,
+	unsigned int bytes,
+	unsigned int inst)
+{
+	union dmub_rb_cmd cmd;
+	struct dc_context *dc = abm->ctx;
+	uint8_t panel_mask = 0x01 << inst;
+
+	// TODO: Optimize by only reading back final 4 bytes
+	dmub_flush_buffer_mem(&dc->dmub_srv->dmub->scratch_mem_fb);
+
+	// Copy iramtable into cw7
+	memcpy(dc->dmub_srv->dmub->scratch_mem_fb.cpu_addr, (void *)src, bytes);
+
+	memset(&cmd, 0, sizeof(cmd));
+	// Fw will copy from cw7 to fw_state
+	cmd.abm_init_config.header.type = DMUB_CMD__ABM;
+	cmd.abm_init_config.header.sub_type = DMUB_CMD__ABM_INIT_CONFIG;
+	cmd.abm_init_config.abm_init_config_data.src.quad_part = dc->dmub_srv->dmub->scratch_mem_fb.gpu_addr;
+	cmd.abm_init_config.abm_init_config_data.bytes = bytes;
+	cmd.abm_init_config.abm_init_config_data.version = DMUB_CMD_ABM_CONTROL_VERSION_1;
+	cmd.abm_init_config.abm_init_config_data.panel_mask = panel_mask;
+
+	cmd.abm_init_config.header.payload_bytes = sizeof(struct dmub_cmd_abm_init_config_data);
+
+	dc_wake_and_execute_dmub_cmd(dc, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
+
+}
+
+bool dmub_abm_set_pause(struct abm *abm, bool pause, unsigned int panel_inst, unsigned int stream_inst)
+{
+	union dmub_rb_cmd cmd;
+	struct dc_context *dc = abm->ctx;
+	uint8_t panel_mask = 0x01 << panel_inst;
+
+	memset(&cmd, 0, sizeof(cmd));
+	cmd.abm_pause.header.type = DMUB_CMD__ABM;
+	cmd.abm_pause.header.sub_type = DMUB_CMD__ABM_PAUSE;
+	cmd.abm_pause.abm_pause_data.enable = pause;
+	cmd.abm_pause.abm_pause_data.panel_mask = panel_mask;
+	cmd.abm_set_level.header.payload_bytes = sizeof(struct dmub_cmd_abm_pause_data);
+
+	dc_wake_and_execute_dmub_cmd(dc, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
+
+	return true;
+}
+
+
+/*****************************************************************************
+ *  dmub_abm_save_restore() - dmub interface for abm save+pause and restore+
+ *                           un-pause
+ *  @dc: dc context
+ *  @panel_inst: panel instance index
+ *  @pData: contains command to pause/un-pause abm and exchange abm parameters
+ *
+ *  When called Pause will get abm data and store in pData, and un-pause will
+ *  set/apply abm data stored in pData.
+ *
+ *****************************************************************************/
+bool dmub_abm_save_restore(
+		struct dc_context *dc,
+		unsigned int panel_inst,
+		struct abm_save_restore *pData)
+{
+	union dmub_rb_cmd cmd;
+	uint8_t panel_mask = 0x01 << panel_inst;
+	unsigned int bytes = sizeof(struct abm_save_restore);
+
+	// TODO: Optimize by only reading back final 4 bytes
+	dmub_flush_buffer_mem(&dc->dmub_srv->dmub->scratch_mem_fb);
+
+	// Copy iramtable into cw7
+	memcpy(dc->dmub_srv->dmub->scratch_mem_fb.cpu_addr, (void *)pData, bytes);
+
+	memset(&cmd, 0, sizeof(cmd));
+	cmd.abm_save_restore.header.type = DMUB_CMD__ABM;
+	cmd.abm_save_restore.header.sub_type = DMUB_CMD__ABM_SAVE_RESTORE;
+
+	cmd.abm_save_restore.abm_init_config_data.src.quad_part = dc->dmub_srv->dmub->scratch_mem_fb.gpu_addr;
+	cmd.abm_save_restore.abm_init_config_data.bytes = bytes;
+	cmd.abm_save_restore.abm_init_config_data.version = DMUB_CMD_ABM_CONTROL_VERSION_1;
+	cmd.abm_save_restore.abm_init_config_data.panel_mask = panel_mask;
+
+	cmd.abm_save_restore.header.payload_bytes =
+			sizeof(struct dmub_rb_cmd_abm_save_restore) - sizeof(struct dmub_cmd_header);
+
+	dc_wake_and_execute_dmub_cmd(dc, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
+
+	// Copy iramtable data into local structure
+	memcpy((void *)pData, dc->dmub_srv->dmub->scratch_mem_fb.cpu_addr, bytes);
+
+	return true;
+}
+
+bool dmub_abm_set_pipe(struct abm *abm,
+		uint32_t otg_inst,
+		uint32_t option,
+		uint32_t panel_inst,
+		uint32_t pwrseq_inst)
+{
+	union dmub_rb_cmd cmd;
+	struct dc_context *dc = abm->ctx;
+	uint8_t ramping_boundary = 0xFF;
+
+	memset(&cmd, 0, sizeof(cmd));
+	cmd.abm_set_pipe.header.type = DMUB_CMD__ABM;
+	cmd.abm_set_pipe.header.sub_type = DMUB_CMD__ABM_SET_PIPE;
+	cmd.abm_set_pipe.abm_set_pipe_data.otg_inst = otg_inst;
+	cmd.abm_set_pipe.abm_set_pipe_data.pwrseq_inst = pwrseq_inst;
+	cmd.abm_set_pipe.abm_set_pipe_data.set_pipe_option = option;
+	cmd.abm_set_pipe.abm_set_pipe_data.panel_inst = panel_inst;
+	cmd.abm_set_pipe.abm_set_pipe_data.ramping_boundary = ramping_boundary;
+	cmd.abm_set_pipe.header.payload_bytes = sizeof(struct dmub_cmd_abm_set_pipe_data);
+
+	dc_wake_and_execute_dmub_cmd(dc, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
+
+	return true;
+}
+
+bool dmub_abm_set_backlight_level(struct abm *abm,
+		unsigned int backlight_pwm_u16_16,
+		unsigned int frame_ramp,
+		unsigned int panel_inst)
+{
+	union dmub_rb_cmd cmd;
+	struct dc_context *dc = abm->ctx;
+
+	memset(&cmd, 0, sizeof(cmd));
+	cmd.abm_set_backlight.header.type = DMUB_CMD__ABM;
+	cmd.abm_set_backlight.header.sub_type = DMUB_CMD__ABM_SET_BACKLIGHT;
+	cmd.abm_set_backlight.abm_set_backlight_data.frame_ramp = frame_ramp;
+	cmd.abm_set_backlight.abm_set_backlight_data.backlight_user_level = backlight_pwm_u16_16;
+	cmd.abm_set_backlight.abm_set_backlight_data.version = DMUB_CMD_ABM_CONTROL_VERSION_1;
+	cmd.abm_set_backlight.abm_set_backlight_data.panel_mask = (0x01 << panel_inst);
+	cmd.abm_set_backlight.header.payload_bytes = sizeof(struct dmub_cmd_abm_set_backlight_data);
+
+	dc_wake_and_execute_dmub_cmd(dc, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
+
+	return true;
+}
+
+bool dmub_abm_set_event(struct abm *abm, unsigned int scaling_enable, unsigned int scaling_strength_map,
+		unsigned int panel_inst)
+{
+	union dmub_rb_cmd cmd;
+	struct dc_context *dc = abm->ctx;
+
+	memset(&cmd, 0, sizeof(cmd));
+	cmd.abm_set_event.header.type = DMUB_CMD__ABM;
+	cmd.abm_set_event.header.sub_type = DMUB_CMD__ABM_SET_EVENT;
+	cmd.abm_set_event.abm_set_event_data.vb_scaling_enable = scaling_enable;
+	cmd.abm_set_event.abm_set_event_data.vb_scaling_strength_mapping = scaling_strength_map;
+	cmd.abm_set_event.abm_set_event_data.panel_mask = (1<<panel_inst);
+	cmd.abm_set_event.header.payload_bytes = sizeof(struct dmub_cmd_abm_set_event_data);
+
+	dc_wake_and_execute_dmub_cmd(dc, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
+
+	return true;
+}
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dmub_abm_lcd.h b/drivers/gpu/drm/amd/display/dc/dce/dmub_abm_lcd.h
new file mode 100644
index 000000000000..13f54f1df780
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dce/dmub_abm_lcd.h
@@ -0,0 +1,54 @@
+/*
+ * Copyright 2019 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __DMUB_ABM_LCD_H__
+#define __DMUB_ABM_LCD_H__
+
+#include "abm.h"
+
+struct abm_save_restore;
+
+void dmub_abm_init(struct abm *abm, uint32_t backlight, uint32_t user_level);
+bool dmub_abm_set_level(struct abm *abm, uint32_t level, uint8_t panel_mask);
+unsigned int dmub_abm_get_current_backlight(struct abm *abm);
+unsigned int dmub_abm_get_target_backlight(struct abm *abm);
+void dmub_abm_init_config(struct abm *abm,
+	const char *src,
+	unsigned int bytes,
+	unsigned int inst);
+
+bool dmub_abm_set_pause(struct abm *abm, bool pause, unsigned int panel_inst, unsigned int stream_inst);
+bool dmub_abm_save_restore(
+		struct dc_context *dc,
+		unsigned int panel_inst,
+		struct abm_save_restore *pData);
+bool dmub_abm_set_pipe(struct abm *abm, uint32_t otg_inst, uint32_t option, uint32_t panel_inst, uint32_t pwrseq_inst);
+bool dmub_abm_set_backlight_level(struct abm *abm,
+		unsigned int backlight_pwm_u16_16,
+		unsigned int frame_ramp,
+		unsigned int panel_inst);
+bool dmub_abm_set_event(struct abm *abm, unsigned int scaling_enable, unsigned int scaling_strength_map,
+		unsigned int panel_inst);
+#endif
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dmub_hw_lock_mgr.c b/drivers/gpu/drm/amd/display/dc/dce/dmub_hw_lock_mgr.c
index 9baf8ca0a920..5bfa2b0d2afd 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dmub_hw_lock_mgr.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dmub_hw_lock_mgr.c
@@ -47,20 +47,63 @@ void dmub_hw_lock_mgr_cmd(struct dc_dmub_srv *dmub_srv,
 	if (!lock)
 		cmd.lock_hw.lock_hw_data.should_release = 1;
 
-	dc_dmub_srv_cmd_queue(dmub_srv, &cmd);
-	dc_dmub_srv_cmd_execute(dmub_srv);
-	dc_dmub_srv_wait_idle(dmub_srv);
+	dc_wake_and_execute_dmub_cmd(dmub_srv->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
 }
 
 void dmub_hw_lock_mgr_inbox0_cmd(struct dc_dmub_srv *dmub_srv,
 		union dmub_inbox0_cmd_lock_hw hw_lock_cmd)
 {
 	union dmub_inbox0_data_register data = { 0 };
+
 	data.inbox0_cmd_lock_hw = hw_lock_cmd;
+	dc_dmub_srv_clear_inbox0_ack(dmub_srv);
 	dc_dmub_srv_send_inbox0_cmd(dmub_srv, data);
+	dc_dmub_srv_wait_for_inbox0_ack(dmub_srv);
 }
 
-bool should_use_dmub_lock(struct dc_link *link)
+bool dmub_hw_lock_mgr_does_link_require_lock(const struct dc *dc, const struct dc_link *link)
 {
+	if (!link)
+		return false;
+
+	if (link->psr_settings.psr_version == DC_PSR_VERSION_SU_1)
+		return true;
+
+	if (link->replay_settings.replay_feature_enabled)
+		return true;
+
+	if (link->psr_settings.psr_version == DC_PSR_VERSION_1) {
+		struct dc_link *edp_links[MAX_NUM_EDP];
+		int edp_num;
+
+		dc_get_edp_links(dc, edp_links, &edp_num);
+		if (edp_num == 1)
+			return true;
+	}
 	return false;
 }
+
+bool dmub_hw_lock_mgr_does_context_require_lock(const struct dc *dc, const struct dc_state *context)
+{
+	if (!context)
+		return false;
+	for (int i = 0; i < context->stream_count; i++) {
+		const struct dc_link *link = context->streams[i]->link;
+
+		if (dmub_hw_lock_mgr_does_link_require_lock(dc, link))
+			return true;
+	}
+	return false;
+}
+
+bool should_use_dmub_inbox1_lock(const struct dc *dc, const struct dc_link *link)
+{
+	/* ASIC doesn't support DMUB */
+	if (!dc->ctx->dmub_srv)
+		return false;
+
+	if (dc->ctx->dce_version >= DCN_VERSION_4_01)
+		return false;
+
+	return dmub_hw_lock_mgr_does_link_require_lock(dc, link);
+}
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dmub_hw_lock_mgr.h b/drivers/gpu/drm/amd/display/dc/dce/dmub_hw_lock_mgr.h
index 5a72b168fb4a..4c80ca8484ad 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dmub_hw_lock_mgr.h
+++ b/drivers/gpu/drm/amd/display/dc/dce/dmub_hw_lock_mgr.h
@@ -37,6 +37,16 @@ void dmub_hw_lock_mgr_cmd(struct dc_dmub_srv *dmub_srv,
 void dmub_hw_lock_mgr_inbox0_cmd(struct dc_dmub_srv *dmub_srv,
 		union dmub_inbox0_cmd_lock_hw hw_lock_cmd);
 
-bool should_use_dmub_lock(struct dc_link *link);
+/**
+ * should_use_dmub_inbox1_lock() - Checks if the DMCUB hardware lock via inbox1 should be used.
+ *
+ * @dc: pointer to DC object
+ * @link: optional pointer to the link object to check for enabled link features
+ *
+ * Return: true if the inbox1 lock should be used, false otherwise
+ */
+bool should_use_dmub_inbox1_lock(const struct dc *dc, const struct dc_link *link);
+bool dmub_hw_lock_mgr_does_link_require_lock(const struct dc *dc, const struct dc_link *link);
+bool dmub_hw_lock_mgr_does_context_require_lock(const struct dc *dc, const struct dc_state *context);
 
 #endif /*_DMUB_HW_LOCK_MGR_H_ */
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dmub_outbox.c b/drivers/gpu/drm/amd/display/dc/dce/dmub_outbox.c
index faad8555ddbb..98a778996e1a 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dmub_outbox.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dmub_outbox.c
@@ -22,20 +22,23 @@
  * Authors: AMD
  */
 
-#include "dmub_outbox.h"
+#include "dc.h"
 #include "dc_dmub_srv.h"
+#include "dmub_outbox.h"
 #include "dmub/inc/dmub_cmd.h"
 
-/**
- *  dmub_enable_outbox_notification - Sends inbox cmd to dmub to enable outbox1
- *                                    messages with interrupt. Dmub sends outbox1
- *                                    message and triggers outbox1 interrupt.
- * @dc: dc structure
+/*
+ *  Function: dmub_enable_outbox_notification
+ *
+ *  @brief
+ *		Sends inbox cmd to dmub for enabling outbox notifications to x86.
+ *
+ *  @param
+ *		[in] dmub_srv: dmub_srv structure
  */
-void dmub_enable_outbox_notification(struct dc *dc)
+void dmub_enable_outbox_notification(struct dc_dmub_srv *dmub_srv)
 {
 	union dmub_rb_cmd cmd;
-	struct dc_context *dc_ctx = dc->ctx;
 
 	memset(&cmd, 0x0, sizeof(cmd));
 	cmd.outbox1_enable.header.type = DMUB_CMD__OUTBOX1_ENABLE;
@@ -45,7 +48,5 @@ void dmub_enable_outbox_notification(struct dc *dc)
 		sizeof(cmd.outbox1_enable.header);
 	cmd.outbox1_enable.enable = true;
 
-	dc_dmub_srv_cmd_queue(dc_ctx->dmub_srv, &cmd);
-	dc_dmub_srv_cmd_execute(dc_ctx->dmub_srv);
-	dc_dmub_srv_wait_idle(dc_ctx->dmub_srv);
+	dc_wake_and_execute_dmub_cmd(dmub_srv->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
 }
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dmub_outbox.h b/drivers/gpu/drm/amd/display/dc/dce/dmub_outbox.h
index 4e0aa0d1a2d5..58ceabb9d497 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dmub_outbox.h
+++ b/drivers/gpu/drm/amd/display/dc/dce/dmub_outbox.h
@@ -26,8 +26,8 @@
 #ifndef _DMUB_OUTBOX_H_
 #define _DMUB_OUTBOX_H_
 
-#include "dc.h"
+struct dc_dmub_srv;
 
-void dmub_enable_outbox_notification(struct dc *dc);
+void dmub_enable_outbox_notification(struct dc_dmub_srv *dmub_srv);
 
 #endif /* _DMUB_OUTBOX_H_ */
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dmub_psr.c b/drivers/gpu/drm/amd/display/dc/dce/dmub_psr.c
index 90eb8eedacf2..87af4fdc04a6 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dmub_psr.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dmub_psr.c
@@ -33,6 +33,10 @@
 
 #define MAX_PIPES 6
 
+static const uint8_t DP_SINK_DEVICE_STR_ID_1[] = {7, 1, 8, 7, 3};
+static const uint8_t DP_SINK_DEVICE_STR_ID_2[] = {7, 1, 8, 7, 5};
+static const uint8_t DP_SINK_DEVICE_STR_ID_3[] = {0x42, 0x61, 0x6c, 0x73, 0x61};
+
 /*
  * Convert dmcub psr state to dmcu psr state.
  */
@@ -74,6 +78,26 @@ static enum dc_psr_state convert_psr_state(uint32_t raw_state)
 		state = PSR_STATE5b;
 	else if (raw_state == 0x53)
 		state = PSR_STATE5c;
+	else if (raw_state == 0x4A)
+		state = PSR_STATE4_FULL_FRAME;
+	else if (raw_state == 0x4B)
+		state = PSR_STATE4a_FULL_FRAME;
+	else if (raw_state == 0x4C)
+		state = PSR_STATE4b_FULL_FRAME;
+	else if (raw_state == 0x4D)
+		state = PSR_STATE4c_FULL_FRAME;
+	else if (raw_state == 0x4E)
+		state = PSR_STATE4_FULL_FRAME_POWERUP;
+	else if (raw_state == 0x4F)
+		state = PSR_STATE4_FULL_FRAME_HW_LOCK;
+	else if (raw_state == 0x60)
+		state = PSR_STATE_HWLOCK_MGR;
+	else if (raw_state == 0x61)
+		state = PSR_STATE_POLLVUPDATE;
+	else if (raw_state == 0x62)
+		state = PSR_STATE_RELEASE_HWLOCK_MGR_FULL_FRAME;
+	else
+		state = PSR_STATE_INVALID;
 
 	return state;
 }
@@ -83,23 +107,18 @@ static enum dc_psr_state convert_psr_state(uint32_t raw_state)
  */
 static void dmub_psr_get_state(struct dmub_psr *dmub, enum dc_psr_state *state, uint8_t panel_inst)
 {
-	struct dmub_srv *srv = dmub->ctx->dmub_srv->dmub;
 	uint32_t raw_state = 0;
 	uint32_t retry_count = 0;
-	enum dmub_status status;
 
 	do {
 		// Send gpint command and wait for ack
-		status = dmub_srv_send_gpint_command(srv, DMUB_GPINT__GET_PSR_STATE, panel_inst, 30);
-
-		if (status == DMUB_STATUS_OK) {
-			// GPINT was executed, get response
-			dmub_srv_get_gpint_response(srv, &raw_state);
+		if (dc_wake_and_execute_gpint(dmub->ctx, DMUB_GPINT__GET_PSR_STATE, panel_inst, &raw_state,
+					      DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY)) {
 			*state = convert_psr_state(raw_state);
-		} else
+		} else {
 			// Return invalid state when GPINT times out
 			*state = PSR_STATE_INVALID;
-
+		}
 	} while (++retry_count <= 1000 && *state == PSR_STATE_INVALID);
 
 	// Assert if max retry hit
@@ -133,18 +152,23 @@ static bool dmub_psr_set_version(struct dmub_psr *dmub, struct dc_stream_state *
 	case DC_PSR_VERSION_1:
 		cmd.psr_set_version.psr_set_version_data.version = PSR_VERSION_1;
 		break;
+	case DC_PSR_VERSION_SU_1:
+		cmd.psr_set_version.psr_set_version_data.version = PSR_VERSION_SU_1;
+		break;
 	case DC_PSR_VERSION_UNSUPPORTED:
 	default:
 		cmd.psr_set_version.psr_set_version_data.version = PSR_VERSION_UNSUPPORTED;
 		break;
 	}
+
+	if (cmd.psr_set_version.psr_set_version_data.version == PSR_VERSION_UNSUPPORTED)
+		return false;
+
 	cmd.psr_set_version.psr_set_version_data.cmd_version = DMUB_CMD_PSR_CONTROL_VERSION_1;
 	cmd.psr_set_version.psr_set_version_data.panel_inst = panel_inst;
 	cmd.psr_set_version.header.payload_bytes = sizeof(struct dmub_cmd_psr_set_version_data);
 
-	dc_dmub_srv_cmd_queue(dc->dmub_srv, &cmd);
-	dc_dmub_srv_cmd_execute(dc->dmub_srv);
-	dc_dmub_srv_wait_idle(dc->dmub_srv);
+	dc_wake_and_execute_dmub_cmd(dc, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
 
 	return true;
 }
@@ -172,9 +196,7 @@ static void dmub_psr_enable(struct dmub_psr *dmub, bool enable, bool wait, uint8
 
 	cmd.psr_enable.header.payload_bytes = 0; // Send header only
 
-	dc_dmub_srv_cmd_queue(dc->dmub_srv, &cmd);
-	dc_dmub_srv_cmd_execute(dc->dmub_srv);
-	dc_dmub_srv_wait_idle(dc->dmub_srv);
+	dc_wake_and_execute_dmub_cmd(dc->dmub_srv->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
 
 	/* Below loops 1000 x 500us = 500 ms.
 	 *  Exit PSR may need to wait 1-2 frames to power up. Timeout after at
@@ -192,6 +214,7 @@ static void dmub_psr_enable(struct dmub_psr *dmub, bool enable, bool wait, uint8
 					break;
 			}
 
+			/* must *not* be fsleep - this can be called from high irq levels */
 			udelay(500);
 		}
 
@@ -222,15 +245,32 @@ static void dmub_psr_set_level(struct dmub_psr *dmub, uint16_t psr_level, uint8_
 	cmd.psr_set_level.psr_set_level_data.psr_level = psr_level;
 	cmd.psr_set_level.psr_set_level_data.cmd_version = DMUB_CMD_PSR_CONTROL_VERSION_1;
 	cmd.psr_set_level.psr_set_level_data.panel_inst = panel_inst;
-	dc_dmub_srv_cmd_queue(dc->dmub_srv, &cmd);
-	dc_dmub_srv_cmd_execute(dc->dmub_srv);
-	dc_dmub_srv_wait_idle(dc->dmub_srv);
+	dc_wake_and_execute_dmub_cmd(dc, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
+}
+
+/*
+ * Set PSR vtotal requirement for FreeSync PSR.
+ */
+static void dmub_psr_set_sink_vtotal_in_psr_active(struct dmub_psr *dmub,
+		uint16_t psr_vtotal_idle, uint16_t psr_vtotal_su)
+{
+	union dmub_rb_cmd cmd;
+	struct dc_context *dc = dmub->ctx;
+
+	memset(&cmd, 0, sizeof(cmd));
+	cmd.psr_set_vtotal.header.type = DMUB_CMD__PSR;
+	cmd.psr_set_vtotal.header.sub_type = DMUB_CMD__SET_SINK_VTOTAL_IN_PSR_ACTIVE;
+	cmd.psr_set_vtotal.header.payload_bytes = sizeof(struct dmub_cmd_psr_set_vtotal_data);
+	cmd.psr_set_vtotal.psr_set_vtotal_data.psr_vtotal_idle = psr_vtotal_idle;
+	cmd.psr_set_vtotal.psr_set_vtotal_data.psr_vtotal_su = psr_vtotal_su;
+
+	dc_wake_and_execute_dmub_cmd(dc, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
 }
 
-/**
+/*
  * Set PSR power optimization flags.
  */
-static void dmub_psr_set_power_opt(struct dmub_psr *dmub, unsigned int power_opt)
+static void dmub_psr_set_power_opt(struct dmub_psr *dmub, unsigned int power_opt, uint8_t panel_inst)
 {
 	union dmub_rb_cmd cmd;
 	struct dc_context *dc = dmub->ctx;
@@ -239,11 +279,11 @@ static void dmub_psr_set_power_opt(struct dmub_psr *dmub, unsigned int power_opt
 	cmd.psr_set_power_opt.header.type = DMUB_CMD__PSR;
 	cmd.psr_set_power_opt.header.sub_type = DMUB_CMD__SET_PSR_POWER_OPT;
 	cmd.psr_set_power_opt.header.payload_bytes = sizeof(struct dmub_cmd_psr_set_power_opt_data);
+	cmd.psr_set_power_opt.psr_set_power_opt_data.cmd_version = DMUB_CMD_PSR_CONTROL_VERSION_1;
 	cmd.psr_set_power_opt.psr_set_power_opt_data.power_opt = power_opt;
+	cmd.psr_set_power_opt.psr_set_power_opt_data.panel_inst = panel_inst;
 
-	dc_dmub_srv_cmd_queue(dc->dmub_srv, &cmd);
-	dc_dmub_srv_cmd_execute(dc->dmub_srv);
-	dc_dmub_srv_wait_idle(dc->dmub_srv);
+	dc_wake_and_execute_dmub_cmd(dc, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
 }
 
 /*
@@ -254,7 +294,7 @@ static bool dmub_psr_copy_settings(struct dmub_psr *dmub,
 		struct psr_context *psr_context,
 		uint8_t panel_inst)
 {
-	union dmub_rb_cmd cmd;
+	union dmub_rb_cmd cmd = { 0 };
 	struct dc_context *dc = dmub->ctx;
 	struct dmub_cmd_psr_copy_settings_data *copy_settings_data
 		= &cmd.psr_copy_settings.psr_copy_settings_data;
@@ -314,6 +354,7 @@ static bool dmub_psr_copy_settings(struct dmub_psr *dmub,
 		copy_settings_data->otg_inst			= 0;
 
 	// Misc
+	copy_settings_data->use_phy_fsm             = link->ctx->dc->debug.psr_power_use_phy_fsm;
 	copy_settings_data->psr_level				= psr_context->psr_level.u32all;
 	copy_settings_data->smu_optimizations_en		= psr_context->allow_smu_optimizations;
 	copy_settings_data->multi_disp_optimizations_en	= psr_context->allow_multi_disp_optimizations;
@@ -323,14 +364,67 @@ static bool dmub_psr_copy_settings(struct dmub_psr *dmub,
 	copy_settings_data->debug.u32All = 0;
 	copy_settings_data->debug.bitfields.visual_confirm	= dc->dc->debug.visual_confirm == VISUAL_CONFIRM_PSR;
 	copy_settings_data->debug.bitfields.use_hw_lock_mgr		= 1;
+	copy_settings_data->debug.bitfields.force_full_frame_update	= 0;
+	copy_settings_data->debug.bitfields.enable_ips_visual_confirm = dc->dc->debug.enable_ips_visual_confirm;
+
+	if (psr_context->su_granularity_required == 0)
+		copy_settings_data->su_y_granularity = 0;
+	else
+		copy_settings_data->su_y_granularity = psr_context->su_y_granularity;
+
+	copy_settings_data->line_capture_indication = 0;
+	copy_settings_data->line_time_in_us = psr_context->line_time_in_us;
+	copy_settings_data->rate_control_caps = psr_context->rate_control_caps;
 	copy_settings_data->fec_enable_status = (link->fec_state == dc_link_fec_enabled);
 	copy_settings_data->fec_enable_delay_in100us = link->dc->debug.fec_enable_delay_in100us;
 	copy_settings_data->cmd_version =  DMUB_CMD_PSR_CONTROL_VERSION_1;
 	copy_settings_data->panel_inst = panel_inst;
+	copy_settings_data->dsc_enable_status = (pipe_ctx->stream->timing.flags.DSC == 1);
 
-	dc_dmub_srv_cmd_queue(dc->dmub_srv, &cmd);
-	dc_dmub_srv_cmd_execute(dc->dmub_srv);
-	dc_dmub_srv_wait_idle(dc->dmub_srv);
+	/**
+	 * WA for PSRSU+DSC on specific TCON, if DSC is enabled, force PSRSU as ffu mode(full frame update)
+	 * Note that PSRSU+DSC is still under development.
+	 */
+	if (copy_settings_data->dsc_enable_status &&
+		link->dpcd_caps.sink_dev_id == DP_DEVICE_ID_38EC11 &&
+		!memcmp(link->dpcd_caps.sink_dev_id_str, DP_SINK_DEVICE_STR_ID_1,
+			sizeof(DP_SINK_DEVICE_STR_ID_1)))
+		link->psr_settings.force_ffu_mode = 1;
+
+	copy_settings_data->force_ffu_mode = link->psr_settings.force_ffu_mode || psr_context->os_request_force_ffu;
+
+	if (((link->dpcd_caps.fec_cap.bits.FEC_CAPABLE &&
+		!link->dc->debug.disable_fec) &&
+		(link->dpcd_caps.dsc_caps.dsc_basic_caps.fields.dsc_support.DSC_SUPPORT &&
+		!link->panel_config.dsc.disable_dsc_edp &&
+		link->dc->caps.edp_dsc_support)) &&
+		link->dpcd_caps.sink_dev_id == DP_DEVICE_ID_38EC11 &&
+		(!memcmp(link->dpcd_caps.sink_dev_id_str, DP_SINK_DEVICE_STR_ID_1,
+			sizeof(DP_SINK_DEVICE_STR_ID_1)) ||
+		!memcmp(link->dpcd_caps.sink_dev_id_str, DP_SINK_DEVICE_STR_ID_2,
+			sizeof(DP_SINK_DEVICE_STR_ID_2))))
+		copy_settings_data->debug.bitfields.force_wakeup_by_tps3 = 1;
+	else
+		copy_settings_data->debug.bitfields.force_wakeup_by_tps3 = 0;
+
+	if (link->psr_settings.psr_version == DC_PSR_VERSION_1 &&
+		link->dpcd_caps.sink_dev_id == DP_DEVICE_ID_0022B9 &&
+		!memcmp(link->dpcd_caps.sink_dev_id_str, DP_SINK_DEVICE_STR_ID_3,
+			sizeof(DP_SINK_DEVICE_STR_ID_3))) {
+		copy_settings_data->poweroff_before_vertical_line = 16;
+	}
+
+	//WA for PSR1 on specific TCON, require frame delay for frame re-lock
+	copy_settings_data->relock_delay_frame_cnt = 0;
+	if (link->dpcd_caps.sink_dev_id == DP_BRANCH_DEVICE_ID_001CF8)
+		copy_settings_data->relock_delay_frame_cnt = 2;
+
+	copy_settings_data->power_down_phy_before_disable_stream =
+		link->psr_settings.power_down_phy_before_disable_stream;
+
+	copy_settings_data->dsc_slice_height = psr_context->dsc_slice_height;
+
+	dc_wake_and_execute_dmub_cmd(dc, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
 
 	return true;
 }
@@ -351,23 +445,22 @@ static void dmub_psr_force_static(struct dmub_psr *dmub, uint8_t panel_inst)
 	cmd.psr_force_static.header.sub_type = DMUB_CMD__PSR_FORCE_STATIC;
 	cmd.psr_enable.header.payload_bytes = 0;
 
-	dc_dmub_srv_cmd_queue(dc->dmub_srv, &cmd);
-	dc_dmub_srv_cmd_execute(dc->dmub_srv);
-	dc_dmub_srv_wait_idle(dc->dmub_srv);
+	dc_wake_and_execute_dmub_cmd(dc, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
 }
 
 /*
  * Get PSR residency from firmware.
  */
-static void dmub_psr_get_residency(struct dmub_psr *dmub, uint32_t *residency, uint8_t panel_inst)
+static void dmub_psr_get_residency(struct dmub_psr *dmub, uint32_t *residency,
+	uint8_t panel_inst, enum psr_residency_mode mode)
 {
-	struct dmub_srv *srv = dmub->ctx->dmub_srv->dmub;
 	uint16_t param = (uint16_t)(panel_inst << 8);
 
-	/* Send gpint command and wait for ack */
-	dmub_srv_send_gpint_command(srv, DMUB_GPINT__PSR_RESIDENCY, param, 30);
+	param |= mode;
 
-	dmub_srv_get_gpint_response(srv, residency);
+	/* Send gpint command and wait for ack */
+	dc_wake_and_execute_gpint(dmub->ctx, DMUB_GPINT__PSR_RESIDENCY, param, residency,
+				  DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY);
 }
 
 static const struct dmub_psr_funcs psr_funcs = {
@@ -377,6 +470,7 @@ static const struct dmub_psr_funcs psr_funcs = {
 	.psr_set_level			= dmub_psr_set_level,
 	.psr_force_static		= dmub_psr_force_static,
 	.psr_get_residency		= dmub_psr_get_residency,
+	.psr_set_sink_vtotal_in_psr_active	= dmub_psr_set_sink_vtotal_in_psr_active,
 	.psr_set_power_opt		= dmub_psr_set_power_opt,
 };
 
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dmub_psr.h b/drivers/gpu/drm/amd/display/dc/dce/dmub_psr.h
index 5dbd479660f1..a6e282d950c3 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dmub_psr.h
+++ b/drivers/gpu/drm/amd/display/dc/dce/dmub_psr.h
@@ -26,8 +26,10 @@
 #ifndef _DMUB_PSR_H_
 #define _DMUB_PSR_H_
 
-#include "os_types.h"
-#include "dc_link.h"
+#include "dc_types.h"
+#include "dmub_cmd.h"
+struct dc_link;
+struct dmub_psr_funcs;
 
 struct dmub_psr {
 	struct dc_context *ctx;
@@ -45,8 +47,10 @@ struct dmub_psr_funcs {
 	uint8_t panel_inst);
 	void (*psr_force_static)(struct dmub_psr *dmub, uint8_t panel_inst);
 	void (*psr_get_residency)(struct dmub_psr *dmub, uint32_t *residency,
-	uint8_t panel_inst);
-	void (*psr_set_power_opt)(struct dmub_psr *dmub, unsigned int power_opt);
+	uint8_t panel_inst,  enum psr_residency_mode mode);
+	void (*psr_set_sink_vtotal_in_psr_active)(struct dmub_psr *dmub,
+	uint16_t psr_vtotal_idle, uint16_t psr_vtotal_su);
+	void (*psr_set_power_opt)(struct dmub_psr *dmub, unsigned int power_opt, uint8_t panel_inst);
 };
 
 struct dmub_psr *dmub_psr_create(struct dc_context *ctx);
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dmub_replay.c b/drivers/gpu/drm/amd/display/dc/dce/dmub_replay.c
new file mode 100644
index 000000000000..cf1372aaff6c
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dce/dmub_replay.c
@@ -0,0 +1,473 @@
+// SPDX-License-Identifier: MIT
+//
+// Copyright 2024 Advanced Micro Devices, Inc.
+
+#include "dc.h"
+#include "link_service.h"
+#include "dc_dmub_srv.h"
+#include "dmub/dmub_srv.h"
+#include "core_types.h"
+#include "dmub_replay.h"
+
+#define DC_TRACE_LEVEL_MESSAGE(...) /* do nothing */
+
+#define MAX_PIPES 6
+
+#define GPINT_RETRY_NUM 20
+
+static const uint8_t DP_SINK_DEVICE_STR_ID_1[] = {7, 1, 8, 7, 3};
+static const uint8_t DP_SINK_DEVICE_STR_ID_2[] = {7, 1, 8, 7, 5};
+
+/*
+ * Get Replay state from firmware.
+ */
+static void dmub_replay_get_state(struct dmub_replay *dmub, enum replay_state *state, uint8_t panel_inst)
+{
+	uint32_t retry_count = 0;
+
+	do {
+		// Send gpint command and wait for ack
+		if (!dc_wake_and_execute_gpint(dmub->ctx, DMUB_GPINT__GET_REPLAY_STATE, panel_inst,
+					       (uint32_t *)state, DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY)) {
+			// Return invalid state when GPINT times out
+			*state = REPLAY_STATE_INVALID;
+		}
+	} while (++retry_count <= 1000 && *state == REPLAY_STATE_INVALID);
+
+	// Assert if max retry hit
+	if (retry_count >= 1000 && *state == REPLAY_STATE_INVALID) {
+		ASSERT(0);
+		/* To-do: Add retry fail log */
+	}
+}
+
+/*
+ * Enable/Disable Replay.
+ */
+static void dmub_replay_enable(struct dmub_replay *dmub, bool enable, bool wait, uint8_t panel_inst)
+{
+	union dmub_rb_cmd cmd;
+	struct dc_context *dc = dmub->ctx;
+	uint32_t retry_count;
+	enum replay_state state = REPLAY_STATE_0;
+
+	memset(&cmd, 0, sizeof(cmd));
+	cmd.replay_enable.header.type = DMUB_CMD__REPLAY;
+	cmd.replay_enable.data.panel_inst = panel_inst;
+
+	cmd.replay_enable.header.sub_type = DMUB_CMD__REPLAY_ENABLE;
+	if (enable)
+		cmd.replay_enable.data.enable = REPLAY_ENABLE;
+	else
+		cmd.replay_enable.data.enable = REPLAY_DISABLE;
+
+	cmd.replay_enable.header.payload_bytes = sizeof(struct dmub_rb_cmd_replay_enable_data);
+
+	dc_wake_and_execute_dmub_cmd(dc, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
+
+	/* Below loops 1000 x 500us = 500 ms.
+	 *  Exit REPLAY may need to wait 1-2 frames to power up. Timeout after at
+	 *  least a few frames. Should never hit the max retry assert below.
+	 */
+	if (wait) {
+		for (retry_count = 0; retry_count <= 1000; retry_count++) {
+			dmub_replay_get_state(dmub, &state, panel_inst);
+
+			if (enable) {
+				if (state != REPLAY_STATE_0)
+					break;
+			} else {
+				if (state == REPLAY_STATE_0)
+					break;
+			}
+
+			/* must *not* be fsleep - this can be called from high irq levels */
+			udelay(500);
+		}
+
+		/* assert if max retry hit */
+		if (retry_count >= 1000)
+			ASSERT(0);
+	}
+}
+
+/*
+ * Set REPLAY power optimization flags.
+ */
+static void dmub_replay_set_power_opt(struct dmub_replay *dmub, unsigned int power_opt, uint8_t panel_inst)
+{
+	union dmub_rb_cmd cmd;
+	struct dc_context *dc = dmub->ctx;
+
+	memset(&cmd, 0, sizeof(cmd));
+	cmd.replay_set_power_opt.header.type = DMUB_CMD__REPLAY;
+	cmd.replay_set_power_opt.header.sub_type = DMUB_CMD__SET_REPLAY_POWER_OPT;
+	cmd.replay_set_power_opt.header.payload_bytes = sizeof(struct dmub_cmd_replay_set_power_opt_data);
+	cmd.replay_set_power_opt.replay_set_power_opt_data.power_opt = power_opt;
+	cmd.replay_set_power_opt.replay_set_power_opt_data.panel_inst = panel_inst;
+
+	dc_wake_and_execute_dmub_cmd(dc, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
+}
+
+/*
+ * Setup Replay by programming phy registers and sending replay hw context values to firmware.
+ */
+static bool dmub_replay_copy_settings(struct dmub_replay *dmub,
+	struct dc_link *link,
+	struct replay_context *replay_context,
+	uint8_t panel_inst)
+{
+	union dmub_rb_cmd cmd;
+	struct dc_context *dc = dmub->ctx;
+	struct dmub_cmd_replay_copy_settings_data *copy_settings_data
+		= &cmd.replay_copy_settings.replay_copy_settings_data;
+	struct pipe_ctx *pipe_ctx = NULL;
+	struct resource_context *res_ctx = &link->ctx->dc->current_state->res_ctx;
+	int i = 0;
+
+	for (i = 0; i < MAX_PIPES; i++) {
+		if (res_ctx &&
+			res_ctx->pipe_ctx[i].stream &&
+			res_ctx->pipe_ctx[i].stream->link &&
+			res_ctx->pipe_ctx[i].stream->link == link &&
+			res_ctx->pipe_ctx[i].stream->link->connector_signal == SIGNAL_TYPE_EDP) {
+			pipe_ctx = &res_ctx->pipe_ctx[i];
+			//TODO: refactor for multi edp support
+			break;
+		}
+	}
+
+	if (!pipe_ctx)
+		return false;
+
+	memset(&cmd, 0, sizeof(cmd));
+	cmd.replay_copy_settings.header.type = DMUB_CMD__REPLAY;
+	cmd.replay_copy_settings.header.sub_type = DMUB_CMD__REPLAY_COPY_SETTINGS;
+	cmd.replay_copy_settings.header.payload_bytes = sizeof(struct dmub_cmd_replay_copy_settings_data);
+
+	// HW insts
+	copy_settings_data->aux_inst				= replay_context->aux_inst;
+	copy_settings_data->digbe_inst				= replay_context->digbe_inst;
+	copy_settings_data->digfe_inst				= replay_context->digfe_inst;
+
+	if (pipe_ctx->plane_res.dpp)
+		copy_settings_data->dpp_inst			= pipe_ctx->plane_res.dpp->inst;
+	else
+		copy_settings_data->dpp_inst			= 0;
+	if (pipe_ctx->stream_res.tg)
+		copy_settings_data->otg_inst			= pipe_ctx->stream_res.tg->inst;
+	else
+		copy_settings_data->otg_inst			= 0;
+
+	copy_settings_data->dpphy_inst				= link->link_enc->transmitter;
+
+	// Misc
+	copy_settings_data->line_time_in_ns			= replay_context->line_time_in_ns;
+	copy_settings_data->panel_inst				= panel_inst;
+	copy_settings_data->debug.u32All			= link->replay_settings.config.debug_flags;
+	copy_settings_data->pixel_deviation_per_line		= link->dpcd_caps.pr_info.pixel_deviation_per_line;
+	copy_settings_data->max_deviation_line			= link->dpcd_caps.pr_info.max_deviation_line;
+	copy_settings_data->smu_optimizations_en		= link->replay_settings.replay_smu_opt_enable;
+	copy_settings_data->replay_timing_sync_supported = link->replay_settings.config.replay_timing_sync_supported;
+	copy_settings_data->replay_support_fast_resync_in_ultra_sleep_mode = link->replay_settings.config.replay_support_fast_resync_in_ultra_sleep_mode;
+
+	copy_settings_data->debug.bitfields.enable_ips_visual_confirm = dc->dc->debug.enable_ips_visual_confirm;
+
+	copy_settings_data->flags.u32All = 0;
+	copy_settings_data->flags.bitfields.fec_enable_status = (link->fec_state == dc_link_fec_enabled);
+	copy_settings_data->flags.bitfields.dsc_enable_status = (pipe_ctx->stream->timing.flags.DSC == 1);
+	// WA for PSRSU+DSC on specific TCON, if DSC is enabled, force PSRSU as ffu mode(full frame update)
+	if (((link->dpcd_caps.fec_cap.bits.FEC_CAPABLE &&
+		!link->dc->debug.disable_fec) &&
+		(link->dpcd_caps.dsc_caps.dsc_basic_caps.fields.dsc_support.DSC_SUPPORT &&
+		!link->panel_config.dsc.disable_dsc_edp &&
+		link->dc->caps.edp_dsc_support)) &&
+		link->dpcd_caps.sink_dev_id == DP_DEVICE_ID_38EC11 &&
+		(!memcmp(link->dpcd_caps.sink_dev_id_str, DP_SINK_DEVICE_STR_ID_1,
+			sizeof(DP_SINK_DEVICE_STR_ID_1)) ||
+		!memcmp(link->dpcd_caps.sink_dev_id_str, DP_SINK_DEVICE_STR_ID_2,
+			sizeof(DP_SINK_DEVICE_STR_ID_2))))
+		copy_settings_data->flags.bitfields.force_wakeup_by_tps3 = 1;
+	else
+		copy_settings_data->flags.bitfields.force_wakeup_by_tps3 = 0;
+
+	copy_settings_data->flags.bitfields.alpm_mode = (enum dmub_alpm_mode)link->replay_settings.config.alpm_mode;
+	if (link->replay_settings.config.alpm_mode == DC_ALPM_AUXLESS) {
+		copy_settings_data->auxless_alpm_data.lfps_setup_ns = dc->dc->debug.auxless_alpm_lfps_setup_ns;
+		copy_settings_data->auxless_alpm_data.lfps_period_ns = dc->dc->debug.auxless_alpm_lfps_period_ns;
+		copy_settings_data->auxless_alpm_data.lfps_silence_ns = dc->dc->debug.auxless_alpm_lfps_silence_ns;
+		copy_settings_data->auxless_alpm_data.lfps_t1_t2_override_us =
+			dc->dc->debug.auxless_alpm_lfps_t1t2_us;
+		copy_settings_data->auxless_alpm_data.lfps_t1_t2_offset_us =
+			dc->dc->debug.auxless_alpm_lfps_t1t2_offset_us;
+		copy_settings_data->auxless_alpm_data.lttpr_count = link->dc->link_srv->dp_get_lttpr_count(link);
+	}
+
+	dc_wake_and_execute_dmub_cmd(dc, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
+
+	return true;
+}
+
+/*
+ * Set coasting vtotal.
+ */
+static void dmub_replay_set_coasting_vtotal(struct dmub_replay *dmub,
+		uint32_t coasting_vtotal,
+		uint8_t panel_inst,
+		uint16_t frame_skip_number)
+{
+	union dmub_rb_cmd cmd;
+	struct dc_context *dc = dmub->ctx;
+	struct dmub_rb_cmd_replay_set_coasting_vtotal *pCmd = NULL;
+
+	pCmd = &(cmd.replay_set_coasting_vtotal);
+
+	memset(&cmd, 0, sizeof(cmd));
+	pCmd->header.type = DMUB_CMD__REPLAY;
+	pCmd->header.sub_type = DMUB_CMD__REPLAY_SET_COASTING_VTOTAL;
+	pCmd->header.payload_bytes = sizeof(struct dmub_cmd_replay_set_coasting_vtotal_data);
+	pCmd->replay_set_coasting_vtotal_data.coasting_vtotal = (coasting_vtotal & 0xFFFF);
+	pCmd->replay_set_coasting_vtotal_data.coasting_vtotal_high = (coasting_vtotal & 0xFFFF0000) >> 16;
+	pCmd->replay_set_coasting_vtotal_data.frame_skip_number = frame_skip_number;
+
+	dc_wake_and_execute_dmub_cmd(dc, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
+}
+
+/*
+ * Get Replay residency from firmware.
+ */
+static void dmub_replay_residency(struct dmub_replay *dmub, uint8_t panel_inst,
+	uint32_t *residency, const bool is_start, enum pr_residency_mode mode)
+{
+	uint16_t param = (uint16_t)(panel_inst << 8);
+	uint32_t i = 0;
+
+	switch (mode) {
+	case PR_RESIDENCY_MODE_PHY:
+		param |= REPLAY_RESIDENCY_FIELD_MODE_PHY;
+		break;
+	case PR_RESIDENCY_MODE_ALPM:
+		param |= REPLAY_RESIDENCY_FIELD_MODE_ALPM;
+		break;
+	case PR_RESIDENCY_MODE_IPS2:
+		param |= REPLAY_RESIDENCY_REVISION_1;
+		param |= REPLAY_RESIDENCY_FIELD_MODE2_IPS;
+		break;
+	case PR_RESIDENCY_MODE_FRAME_CNT:
+		param |= REPLAY_RESIDENCY_REVISION_1;
+		param |= REPLAY_RESIDENCY_FIELD_MODE2_FRAME_CNT;
+		break;
+	case PR_RESIDENCY_MODE_ENABLEMENT_PERIOD:
+		param |= REPLAY_RESIDENCY_REVISION_1;
+		param |= REPLAY_RESIDENCY_FIELD_MODE2_EN_PERIOD;
+		break;
+	default:
+		break;
+	}
+
+	if (is_start)
+		param |= REPLAY_RESIDENCY_ENABLE;
+
+	for (i = 0; i < GPINT_RETRY_NUM; i++) {
+		// Send gpint command and wait for ack
+		if (dc_wake_and_execute_gpint(dmub->ctx, DMUB_GPINT__REPLAY_RESIDENCY, param,
+			residency, DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY))
+			return;
+
+		udelay(100);
+	}
+
+	// it means gpint retry many times
+	*residency = 0;
+}
+
+/*
+ * Set REPLAY power optimization flags and coasting vtotal.
+ */
+static void dmub_replay_set_power_opt_and_coasting_vtotal(struct dmub_replay *dmub,
+		unsigned int power_opt, uint8_t panel_inst, uint32_t coasting_vtotal, uint16_t frame_skip_number)
+{
+	union dmub_rb_cmd cmd;
+	struct dc_context *dc = dmub->ctx;
+	struct dmub_rb_cmd_replay_set_power_opt_and_coasting_vtotal *pCmd = NULL;
+
+	pCmd = &(cmd.replay_set_power_opt_and_coasting_vtotal);
+
+	memset(&cmd, 0, sizeof(cmd));
+	pCmd->header.type = DMUB_CMD__REPLAY;
+	pCmd->header.sub_type = DMUB_CMD__REPLAY_SET_POWER_OPT_AND_COASTING_VTOTAL;
+	pCmd->header.payload_bytes =
+			sizeof(struct dmub_rb_cmd_replay_set_power_opt_and_coasting_vtotal) -
+			sizeof(struct dmub_cmd_header);
+	pCmd->replay_set_power_opt_data.power_opt = power_opt;
+	pCmd->replay_set_power_opt_data.panel_inst = panel_inst;
+	pCmd->replay_set_coasting_vtotal_data.coasting_vtotal = (coasting_vtotal & 0xFFFF);
+	pCmd->replay_set_coasting_vtotal_data.coasting_vtotal_high = (coasting_vtotal & 0xFFFF0000) >> 16;
+	pCmd->replay_set_coasting_vtotal_data.frame_skip_number = frame_skip_number;
+
+	dc_wake_and_execute_dmub_cmd(dc, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
+}
+
+/*
+ * send Replay general cmd to DMUB.
+ */
+static void dmub_replay_send_cmd(struct dmub_replay *dmub,
+		enum replay_FW_Message_type msg, union dmub_replay_cmd_set *cmd_element)
+{
+	union dmub_rb_cmd cmd;
+	struct dc_context *ctx = NULL;
+
+	if (dmub == NULL || cmd_element == NULL)
+		return;
+
+	ctx = dmub->ctx;
+	if (ctx != NULL) {
+
+		if (msg != Replay_Msg_Not_Support) {
+			memset(&cmd, 0, sizeof(cmd));
+			//Header
+			cmd.replay_set_timing_sync.header.type = DMUB_CMD__REPLAY;
+		} else
+			return;
+	} else
+		return;
+
+	switch (msg) {
+	case Replay_Set_Timing_Sync_Supported:
+		//Header
+		cmd.replay_set_timing_sync.header.sub_type =
+			DMUB_CMD__REPLAY_SET_TIMING_SYNC_SUPPORTED;
+		cmd.replay_set_timing_sync.header.payload_bytes =
+			sizeof(struct dmub_rb_cmd_replay_set_timing_sync) -
+			sizeof(struct dmub_cmd_header);
+		//Cmd Body
+		cmd.replay_set_timing_sync.replay_set_timing_sync_data.panel_inst =
+						cmd_element->sync_data.panel_inst;
+		cmd.replay_set_timing_sync.replay_set_timing_sync_data.timing_sync_supported =
+						cmd_element->sync_data.timing_sync_supported;
+		break;
+	case Replay_Set_Residency_Frameupdate_Timer:
+		//Header
+		cmd.replay_set_frameupdate_timer.header.sub_type =
+			DMUB_CMD__REPLAY_SET_RESIDENCY_FRAMEUPDATE_TIMER;
+		cmd.replay_set_frameupdate_timer.header.payload_bytes =
+			sizeof(struct dmub_rb_cmd_replay_set_frameupdate_timer) -
+			sizeof(struct dmub_cmd_header);
+		//Cmd Body
+		cmd.replay_set_frameupdate_timer.data.panel_inst =
+						cmd_element->panel_inst;
+		cmd.replay_set_frameupdate_timer.data.enable =
+						cmd_element->timer_data.enable;
+		cmd.replay_set_frameupdate_timer.data.frameupdate_count =
+						cmd_element->timer_data.frameupdate_count;
+		break;
+	case Replay_Set_Pseudo_VTotal:
+		//Header
+		cmd.replay_set_pseudo_vtotal.header.sub_type =
+			DMUB_CMD__REPLAY_SET_PSEUDO_VTOTAL;
+		cmd.replay_set_pseudo_vtotal.header.payload_bytes =
+			sizeof(struct dmub_rb_cmd_replay_set_pseudo_vtotal) -
+			sizeof(struct dmub_cmd_header);
+		//Cmd Body
+		cmd.replay_set_pseudo_vtotal.data.panel_inst =
+			cmd_element->pseudo_vtotal_data.panel_inst;
+		cmd.replay_set_pseudo_vtotal.data.vtotal =
+			cmd_element->pseudo_vtotal_data.vtotal;
+		break;
+	case Replay_Disabled_Adaptive_Sync_SDP:
+		//Header
+		cmd.replay_disabled_adaptive_sync_sdp.header.sub_type =
+			DMUB_CMD__REPLAY_DISABLED_ADAPTIVE_SYNC_SDP;
+		cmd.replay_disabled_adaptive_sync_sdp.header.payload_bytes =
+			sizeof(struct dmub_rb_cmd_replay_disabled_adaptive_sync_sdp) -
+			sizeof(struct dmub_cmd_header);
+		//Cmd Body
+		cmd.replay_disabled_adaptive_sync_sdp.data.panel_inst =
+			cmd_element->disabled_adaptive_sync_sdp_data.panel_inst;
+		cmd.replay_disabled_adaptive_sync_sdp.data.force_disabled =
+			cmd_element->disabled_adaptive_sync_sdp_data.force_disabled;
+		break;
+	case Replay_Set_Version:
+		//Header
+		cmd.replay_set_version.header.sub_type =
+			DMUB_CMD__REPLAY_SET_VERSION;
+		cmd.replay_set_version.header.payload_bytes =
+			sizeof(struct dmub_rb_cmd_replay_set_version) -
+			sizeof(struct dmub_cmd_header);
+		//Cmd Body
+		cmd.replay_set_version.replay_set_version_data.panel_inst =
+			cmd_element->version_data.panel_inst;
+		cmd.replay_set_version.replay_set_version_data.version =
+			cmd_element->version_data.version;
+		break;
+	case Replay_Set_General_Cmd:
+		//Header
+		cmd.replay_set_general_cmd.header.sub_type =
+			DMUB_CMD__REPLAY_SET_GENERAL_CMD;
+		cmd.replay_set_general_cmd.header.payload_bytes =
+			sizeof(struct dmub_rb_cmd_replay_set_general_cmd) -
+			sizeof(struct dmub_cmd_header);
+		//Cmd Body
+		cmd.replay_set_general_cmd.data.panel_inst =
+			cmd_element->set_general_cmd_data.panel_inst;
+		cmd.replay_set_general_cmd.data.subtype =
+			cmd_element->set_general_cmd_data.subtype;
+		cmd.replay_set_general_cmd.data.param1 =
+			cmd_element->set_general_cmd_data.param1;
+		cmd.replay_set_general_cmd.data.param2 =
+			cmd_element->set_general_cmd_data.param2;
+		break;
+	case Replay_Msg_Not_Support:
+	default:
+		return;
+		break;
+	}
+
+	dc_wake_and_execute_dmub_cmd(ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
+}
+
+static const struct dmub_replay_funcs replay_funcs = {
+	.replay_copy_settings				= dmub_replay_copy_settings,
+	.replay_enable					= dmub_replay_enable,
+	.replay_get_state				= dmub_replay_get_state,
+	.replay_set_power_opt				= dmub_replay_set_power_opt,
+	.replay_set_coasting_vtotal			= dmub_replay_set_coasting_vtotal,
+	.replay_residency				= dmub_replay_residency,
+	.replay_set_power_opt_and_coasting_vtotal	= dmub_replay_set_power_opt_and_coasting_vtotal,
+	.replay_send_cmd				= dmub_replay_send_cmd,
+};
+
+/*
+ * Construct Replay object.
+ */
+static void dmub_replay_construct(struct dmub_replay *replay, struct dc_context *ctx)
+{
+	replay->ctx = ctx;
+	replay->funcs = &replay_funcs;
+}
+
+/*
+ * Allocate and initialize Replay object.
+ */
+struct dmub_replay *dmub_replay_create(struct dc_context *ctx)
+{
+	struct dmub_replay *replay = kzalloc(sizeof(struct dmub_replay), GFP_KERNEL);
+
+	if (replay == NULL) {
+		BREAK_TO_DEBUGGER();
+		return NULL;
+	}
+
+	dmub_replay_construct(replay, ctx);
+
+	return replay;
+}
+
+/*
+ * Deallocate Replay object.
+ */
+void dmub_replay_destroy(struct dmub_replay **dmub)
+{
+	kfree(*dmub);
+	*dmub = NULL;
+}
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dmub_replay.h b/drivers/gpu/drm/amd/display/dc/dce/dmub_replay.h
new file mode 100644
index 000000000000..07c79739a980
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dce/dmub_replay.h
@@ -0,0 +1,42 @@
+// SPDX-License-Identifier: MIT
+//
+// Copyright 2024 Advanced Micro Devices, Inc.
+
+#ifndef _DMUB_REPLAY_H_
+#define _DMUB_REPLAY_H_
+
+#include "dc_types.h"
+#include "dmub_cmd.h"
+struct dc_link;
+struct dmub_replay_funcs;
+
+struct dmub_replay {
+	struct dc_context *ctx;
+	const struct dmub_replay_funcs *funcs;
+};
+
+struct dmub_replay_funcs {
+	void (*replay_get_state)(struct dmub_replay *dmub, enum replay_state *state,
+		uint8_t panel_inst);
+	void (*replay_enable)(struct dmub_replay *dmub, bool enable, bool wait,
+		uint8_t panel_inst);
+	bool (*replay_copy_settings)(struct dmub_replay *dmub, struct dc_link *link,
+		struct replay_context *replay_context, uint8_t panel_inst);
+	void (*replay_set_power_opt)(struct dmub_replay *dmub, unsigned int power_opt,
+		uint8_t panel_inst);
+	void (*replay_send_cmd)(struct dmub_replay *dmub,
+		enum replay_FW_Message_type msg, union dmub_replay_cmd_set *cmd_element);
+	void (*replay_set_coasting_vtotal)(struct dmub_replay *dmub, uint32_t coasting_vtotal,
+		uint8_t panel_inst, uint16_t frame_skip_number);
+	void (*replay_residency)(struct dmub_replay *dmub,
+		uint8_t panel_inst, uint32_t *residency, const bool is_start, const enum pr_residency_mode mode);
+	void (*replay_set_power_opt_and_coasting_vtotal)(struct dmub_replay *dmub,
+		unsigned int power_opt, uint8_t panel_inst, uint32_t coasting_vtotal,
+		uint16_t frame_skip_number);
+};
+
+struct dmub_replay *dmub_replay_create(struct dc_context *ctx);
+void dmub_replay_destroy(struct dmub_replay **dmub);
+
+
+#endif /* _DMUB_REPLAY_H_ */
diff --git a/drivers/gpu/drm/amd/display/dc/dce100/Makefile b/drivers/gpu/drm/amd/display/dc/dce100/Makefile
deleted file mode 100644
index ff20c47f559e..000000000000
--- a/drivers/gpu/drm/amd/display/dc/dce100/Makefile
+++ /dev/null
@@ -1,46 +0,0 @@
-#
-# Copyright 2017 Advanced Micro Devices, Inc.
-#
-# Permission is hereby granted, free of charge, to any person obtaining a
-# copy of this software and associated documentation files (the "Software"),
-# to deal in the Software without restriction, including without limitation
-# the rights to use, copy, modify, merge, publish, distribute, sublicense,
-# and/or sell copies of the Software, and to permit persons to whom the
-# Software is furnished to do so, subject to the following conditions:
-#
-# The above copyright notice and this permission notice shall be included in
-# all copies or substantial portions of the Software.
-#
-# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
-# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
-# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
-# THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
-# OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
-# ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
-# OTHER DEALINGS IN THE SOFTWARE.
-#
-#
-# Makefile for the 'controller' sub-component of DAL.
-# It provides the control and status of HW CRTC block.
-
-CFLAGS_$(AMDDALPATH)/dc/dce100/dce100_resource.o = $(call cc-disable-warning, override-init)
-
-DCE100 = dce100_resource.o dce100_hw_sequencer.o
-
-AMD_DAL_DCE100 = $(addprefix $(AMDDALPATH)/dc/dce100/,$(DCE100))
-
-AMD_DISPLAY_FILES += $(AMD_DAL_DCE100)
-
-
-###############################################################################
-# DCE 10x
-###############################################################################
-ifdef 0#CONFIG_DRM_AMD_DC_DCE11_0
-TG_DCE100 = dce100_resource.o
-
-AMD_DAL_TG_DCE100 = $(addprefix \
-	$(AMDDALPATH)/dc/dce100/,$(TG_DCE100))
-
-AMD_DISPLAY_FILES += $(AMD_DAL_TG_DCE100)
-endif
-
diff --git a/drivers/gpu/drm/amd/display/dc/dce100/dce100_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/dce100/dce100_hw_sequencer.c
deleted file mode 100644
index 753cb8edd996..000000000000
--- a/drivers/gpu/drm/amd/display/dc/dce100/dce100_hw_sequencer.c
+++ /dev/null
@@ -1,142 +0,0 @@
-/*
- * Copyright 2015 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: AMD
- *
- */
-#include "dm_services.h"
-#include "dc.h"
-#include "core_types.h"
-#include "clk_mgr.h"
-#include "dce100_hw_sequencer.h"
-#include "resource.h"
-
-#include "dce110/dce110_hw_sequencer.h"
-
-/* include DCE10 register header files */
-#include "dce/dce_10_0_d.h"
-#include "dce/dce_10_0_sh_mask.h"
-
-struct dce100_hw_seq_reg_offsets {
-	uint32_t blnd;
-	uint32_t crtc;
-};
-
-static const struct dce100_hw_seq_reg_offsets reg_offsets[] = {
-{
-	.crtc = (mmCRTC0_CRTC_GSL_CONTROL - mmCRTC_GSL_CONTROL),
-},
-{
-	.crtc = (mmCRTC1_CRTC_GSL_CONTROL - mmCRTC_GSL_CONTROL),
-},
-{
-	.crtc = (mmCRTC2_CRTC_GSL_CONTROL - mmCRTC_GSL_CONTROL),
-},
-{
-	.crtc = (mmCRTC3_CRTC_GSL_CONTROL - mmCRTC_GSL_CONTROL),
-},
-{
-	.crtc = (mmCRTC4_CRTC_GSL_CONTROL - mmCRTC_GSL_CONTROL),
-},
-{
-	.crtc = (mmCRTC5_CRTC_GSL_CONTROL - mmCRTC_GSL_CONTROL),
-}
-};
-
-#define HW_REG_CRTC(reg, id)\
-	(reg + reg_offsets[id].crtc)
-
-/*******************************************************************************
- * Private definitions
- ******************************************************************************/
-/***************************PIPE_CONTROL***********************************/
-
-bool dce100_enable_display_power_gating(
-	struct dc *dc,
-	uint8_t controller_id,
-	struct dc_bios *dcb,
-	enum pipe_gating_control power_gating)
-{
-	enum bp_result bp_result = BP_RESULT_OK;
-	enum bp_pipe_control_action cntl;
-	struct dc_context *ctx = dc->ctx;
-
-	if (power_gating == PIPE_GATING_CONTROL_INIT)
-		cntl = ASIC_PIPE_INIT;
-	else if (power_gating == PIPE_GATING_CONTROL_ENABLE)
-		cntl = ASIC_PIPE_ENABLE;
-	else
-		cntl = ASIC_PIPE_DISABLE;
-
-	if (!(power_gating == PIPE_GATING_CONTROL_INIT && controller_id != 0)){
-
-		bp_result = dcb->funcs->enable_disp_power_gating(
-						dcb, controller_id + 1, cntl);
-
-		/* Revert MASTER_UPDATE_MODE to 0 because bios sets it 2
-		 * by default when command table is called
-		 */
-		dm_write_reg(ctx,
-			HW_REG_CRTC(mmMASTER_UPDATE_MODE, controller_id),
-			0);
-	}
-
-	if (bp_result == BP_RESULT_OK)
-		return true;
-	else
-		return false;
-}
-
-void dce100_prepare_bandwidth(
-		struct dc *dc,
-		struct dc_state *context)
-{
-	dce110_set_safe_displaymarks(&context->res_ctx, dc->res_pool);
-
-	dc->clk_mgr->funcs->update_clocks(
-			dc->clk_mgr,
-			context,
-			false);
-}
-
-void dce100_optimize_bandwidth(
-		struct dc *dc,
-		struct dc_state *context)
-{
-	dce110_set_safe_displaymarks(&context->res_ctx, dc->res_pool);
-
-	dc->clk_mgr->funcs->update_clocks(
-			dc->clk_mgr,
-			context,
-			true);
-}
-
-/**************************************************************************/
-
-void dce100_hw_sequencer_construct(struct dc *dc)
-{
-	dce110_hw_sequencer_construct(dc);
-
-	dc->hwseq->funcs.enable_display_power_gating = dce100_enable_display_power_gating;
-	dc->hwss.prepare_bandwidth = dce100_prepare_bandwidth;
-	dc->hwss.optimize_bandwidth = dce100_optimize_bandwidth;
-}
-
diff --git a/drivers/gpu/drm/amd/display/dc/dce100/dce100_hw_sequencer.h b/drivers/gpu/drm/amd/display/dc/dce100/dce100_hw_sequencer.h
deleted file mode 100644
index 34518da20009..000000000000
--- a/drivers/gpu/drm/amd/display/dc/dce100/dce100_hw_sequencer.h
+++ /dev/null
@@ -1,50 +0,0 @@
-/*
-* Copyright 2012-15 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: AMD
- *
- */
-
-#ifndef __DC_HWSS_DCE100_H__
-#define __DC_HWSS_DCE100_H__
-
-#include "core_types.h"
-#include "hw_sequencer_private.h"
-
-struct dc;
-struct dc_state;
-
-void dce100_hw_sequencer_construct(struct dc *dc);
-
-void dce100_prepare_bandwidth(
-		struct dc *dc,
-		struct dc_state *context);
-
-void dce100_optimize_bandwidth(
-		struct dc *dc,
-		struct dc_state *context);
-
-bool dce100_enable_display_power_gating(struct dc *dc, uint8_t controller_id,
-					struct dc_bios *dcb,
-					enum pipe_gating_control power_gating);
-
-#endif /* __DC_HWSS_DCE100_H__ */
-
diff --git a/drivers/gpu/drm/amd/display/dc/dce110/Makefile b/drivers/gpu/drm/amd/display/dc/dce110/Makefile
index 84ab48df0c26..c307f040e48f 100644
--- a/drivers/gpu/drm/amd/display/dc/dce110/Makefile
+++ b/drivers/gpu/drm/amd/display/dc/dce110/Makefile
@@ -23,11 +23,11 @@
 # Makefile for the 'controller' sub-component of DAL.
 # It provides the control and status of HW CRTC block.
 
-CFLAGS_$(AMDDALPATH)/dc/dce110/dce110_resource.o = $(call cc-disable-warning, override-init)
+CFLAGS_$(AMDDALPATH)/dc/dce110/dce110_resource.o = -Wno-override-init
 
 DCE110 = dce110_timing_generator.o \
-dce110_compressor.o dce110_hw_sequencer.o dce110_resource.o \
-dce110_opp_regamma_v.o dce110_opp_csc_v.o dce110_timing_generator_v.o \
+dce110_compressor.o dce110_opp_regamma_v.o \
+dce110_opp_csc_v.o dce110_timing_generator_v.o \
 dce110_mem_input_v.o dce110_opp_v.o dce110_transform_v.o
 
 AMD_DAL_DCE110 = $(addprefix $(AMDDALPATH)/dc/dce110/,$(DCE110))
diff --git a/drivers/gpu/drm/amd/display/dc/dce110/dce110_compressor.c b/drivers/gpu/drm/amd/display/dc/dce110/dce110_compressor.c
index 44564a4742b5..59a0961b49da 100644
--- a/drivers/gpu/drm/amd/display/dc/dce110/dce110_compressor.c
+++ b/drivers/gpu/drm/amd/display/dc/dce110/dce110_compressor.c
@@ -23,9 +23,6 @@
  *
  */
 
-#include <linux/delay.h>
-#include <linux/slab.h>
-
 #include "dm_services.h"
 
 #include "dce/dce_11_0_d.h"
@@ -412,19 +409,6 @@ void dce110_compressor_destroy(struct compressor **compressor)
 	*compressor = NULL;
 }
 
-void get_max_support_fbc_buffersize(unsigned int *max_x, unsigned int *max_y)
-{
-	*max_x = FBC_MAX_X;
-	*max_y = FBC_MAX_Y;
-
-	/* if (m_smallLocalFrameBufferMemory == 1)
-	 * {
-	 *	*max_x = FBC_MAX_X_SG;
-	 *	*max_y = FBC_MAX_Y_SG;
-	 * }
-	 */
-}
-
 static const struct compressor_funcs dce110_compressor_funcs = {
 	.power_up_fbc = dce110_compressor_power_up_fbc,
 	.enable_fbc = dce110_compressor_enable_fbc,
diff --git a/drivers/gpu/drm/amd/display/dc/dce110/dce110_compressor.h b/drivers/gpu/drm/amd/display/dc/dce110/dce110_compressor.h
index 26c7335a1cbf..223c57941e92 100644
--- a/drivers/gpu/drm/amd/display/dc/dce110/dce110_compressor.h
+++ b/drivers/gpu/drm/amd/display/dc/dce110/dce110_compressor.h
@@ -75,7 +75,5 @@ void dce110_compressor_program_lpt_control(struct compressor *cp,
 
 bool dce110_compressor_is_lpt_enabled_in_hw(struct compressor *cp);
 
-void get_max_support_fbc_buffersize(unsigned int *max_x, unsigned int *max_y);
-
 #endif
 
diff --git a/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c
deleted file mode 100644
index 24e47df526f6..000000000000
--- a/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c
+++ /dev/null
@@ -1,3130 +0,0 @@
-/*
- * Copyright 2015 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: AMD
- *
- */
-
-#include <linux/delay.h>
-
-#include "dm_services.h"
-#include "dc.h"
-#include "dc_bios_types.h"
-#include "core_types.h"
-#include "core_status.h"
-#include "resource.h"
-#include "dm_helpers.h"
-#include "dce110_timing_generator.h"
-#include "dce/dce_hwseq.h"
-#include "gpio_service_interface.h"
-
-#include "dce110_compressor.h"
-
-#include "bios/bios_parser_helper.h"
-#include "timing_generator.h"
-#include "mem_input.h"
-#include "opp.h"
-#include "ipp.h"
-#include "transform.h"
-#include "stream_encoder.h"
-#include "link_encoder.h"
-#include "link_enc_cfg.h"
-#include "link_hwss.h"
-#include "dc_link_dp.h"
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-#include "dccg.h"
-#endif
-#include "clock_source.h"
-#include "clk_mgr.h"
-#include "abm.h"
-#include "audio.h"
-#include "reg_helper.h"
-#include "panel_cntl.h"
-#include "inc/link_dpcd.h"
-#include "dpcd_defs.h"
-/* include DCE11 register header files */
-#include "dce/dce_11_0_d.h"
-#include "dce/dce_11_0_sh_mask.h"
-#include "custom_float.h"
-
-#include "atomfirmware.h"
-
-#include "dcn10/dcn10_hw_sequencer.h"
-
-#define GAMMA_HW_POINTS_NUM 256
-
-/*
- * All values are in milliseconds;
- * For eDP, after power-up/power/down,
- * 300/500 msec max. delay from LCDVCC to black video generation
- */
-#define PANEL_POWER_UP_TIMEOUT 300
-#define PANEL_POWER_DOWN_TIMEOUT 500
-#define HPD_CHECK_INTERVAL 10
-#define OLED_POST_T7_DELAY 100
-#define OLED_PRE_T11_DELAY 150
-
-#define CTX \
-	hws->ctx
-
-#define DC_LOGGER_INIT()
-
-#define REG(reg)\
-	hws->regs->reg
-
-#undef FN
-#define FN(reg_name, field_name) \
-	hws->shifts->field_name, hws->masks->field_name
-
-struct dce110_hw_seq_reg_offsets {
-	uint32_t crtc;
-};
-
-static const struct dce110_hw_seq_reg_offsets reg_offsets[] = {
-{
-	.crtc = (mmCRTC0_CRTC_GSL_CONTROL - mmCRTC_GSL_CONTROL),
-},
-{
-	.crtc = (mmCRTC1_CRTC_GSL_CONTROL - mmCRTC_GSL_CONTROL),
-},
-{
-	.crtc = (mmCRTC2_CRTC_GSL_CONTROL - mmCRTC_GSL_CONTROL),
-},
-{
-	.crtc = (mmCRTCV_GSL_CONTROL - mmCRTC_GSL_CONTROL),
-}
-};
-
-#define HW_REG_BLND(reg, id)\
-	(reg + reg_offsets[id].blnd)
-
-#define HW_REG_CRTC(reg, id)\
-	(reg + reg_offsets[id].crtc)
-
-#define MAX_WATERMARK 0xFFFF
-#define SAFE_NBP_MARK 0x7FFF
-
-/*******************************************************************************
- * Private definitions
- ******************************************************************************/
-/***************************PIPE_CONTROL***********************************/
-static void dce110_init_pte(struct dc_context *ctx)
-{
-	uint32_t addr;
-	uint32_t value = 0;
-	uint32_t chunk_int = 0;
-	uint32_t chunk_mul = 0;
-
-	addr = mmUNP_DVMM_PTE_CONTROL;
-	value = dm_read_reg(ctx, addr);
-
-	set_reg_field_value(
-		value,
-		0,
-		DVMM_PTE_CONTROL,
-		DVMM_USE_SINGLE_PTE);
-
-	set_reg_field_value(
-		value,
-		1,
-		DVMM_PTE_CONTROL,
-		DVMM_PTE_BUFFER_MODE0);
-
-	set_reg_field_value(
-		value,
-		1,
-		DVMM_PTE_CONTROL,
-		DVMM_PTE_BUFFER_MODE1);
-
-	dm_write_reg(ctx, addr, value);
-
-	addr = mmDVMM_PTE_REQ;
-	value = dm_read_reg(ctx, addr);
-
-	chunk_int = get_reg_field_value(
-		value,
-		DVMM_PTE_REQ,
-		HFLIP_PTEREQ_PER_CHUNK_INT);
-
-	chunk_mul = get_reg_field_value(
-		value,
-		DVMM_PTE_REQ,
-		HFLIP_PTEREQ_PER_CHUNK_MULTIPLIER);
-
-	if (chunk_int != 0x4 || chunk_mul != 0x4) {
-
-		set_reg_field_value(
-			value,
-			255,
-			DVMM_PTE_REQ,
-			MAX_PTEREQ_TO_ISSUE);
-
-		set_reg_field_value(
-			value,
-			4,
-			DVMM_PTE_REQ,
-			HFLIP_PTEREQ_PER_CHUNK_INT);
-
-		set_reg_field_value(
-			value,
-			4,
-			DVMM_PTE_REQ,
-			HFLIP_PTEREQ_PER_CHUNK_MULTIPLIER);
-
-		dm_write_reg(ctx, addr, value);
-	}
-}
-/**************************************************************************/
-
-static void enable_display_pipe_clock_gating(
-	struct dc_context *ctx,
-	bool clock_gating)
-{
-	/*TODO*/
-}
-
-static bool dce110_enable_display_power_gating(
-	struct dc *dc,
-	uint8_t controller_id,
-	struct dc_bios *dcb,
-	enum pipe_gating_control power_gating)
-{
-	enum bp_result bp_result = BP_RESULT_OK;
-	enum bp_pipe_control_action cntl;
-	struct dc_context *ctx = dc->ctx;
-	unsigned int underlay_idx = dc->res_pool->underlay_pipe_index;
-
-	if (IS_FPGA_MAXIMUS_DC(ctx->dce_environment))
-		return true;
-
-	if (power_gating == PIPE_GATING_CONTROL_INIT)
-		cntl = ASIC_PIPE_INIT;
-	else if (power_gating == PIPE_GATING_CONTROL_ENABLE)
-		cntl = ASIC_PIPE_ENABLE;
-	else
-		cntl = ASIC_PIPE_DISABLE;
-
-	if (controller_id == underlay_idx)
-		controller_id = CONTROLLER_ID_UNDERLAY0 - 1;
-
-	if (power_gating != PIPE_GATING_CONTROL_INIT || controller_id == 0){
-
-		bp_result = dcb->funcs->enable_disp_power_gating(
-						dcb, controller_id + 1, cntl);
-
-		/* Revert MASTER_UPDATE_MODE to 0 because bios sets it 2
-		 * by default when command table is called
-		 *
-		 * Bios parser accepts controller_id = 6 as indicative of
-		 * underlay pipe in dce110. But we do not support more
-		 * than 3.
-		 */
-		if (controller_id < CONTROLLER_ID_MAX - 1)
-			dm_write_reg(ctx,
-				HW_REG_CRTC(mmCRTC_MASTER_UPDATE_MODE, controller_id),
-				0);
-	}
-
-	if (power_gating != PIPE_GATING_CONTROL_ENABLE)
-		dce110_init_pte(ctx);
-
-	if (bp_result == BP_RESULT_OK)
-		return true;
-	else
-		return false;
-}
-
-static void build_prescale_params(struct ipp_prescale_params *prescale_params,
-		const struct dc_plane_state *plane_state)
-{
-	prescale_params->mode = IPP_PRESCALE_MODE_FIXED_UNSIGNED;
-
-	switch (plane_state->format) {
-	case SURFACE_PIXEL_FORMAT_GRPH_RGB565:
-		prescale_params->scale = 0x2082;
-		break;
-	case SURFACE_PIXEL_FORMAT_GRPH_ARGB8888:
-	case SURFACE_PIXEL_FORMAT_GRPH_ABGR8888:
-		prescale_params->scale = 0x2020;
-		break;
-	case SURFACE_PIXEL_FORMAT_GRPH_ARGB2101010:
-	case SURFACE_PIXEL_FORMAT_GRPH_ABGR2101010:
-		prescale_params->scale = 0x2008;
-		break;
-	case SURFACE_PIXEL_FORMAT_GRPH_ARGB16161616:
-	case SURFACE_PIXEL_FORMAT_GRPH_ABGR16161616:
-	case SURFACE_PIXEL_FORMAT_GRPH_ABGR16161616F:
-		prescale_params->scale = 0x2000;
-		break;
-	default:
-		ASSERT(false);
-		break;
-	}
-}
-
-static bool
-dce110_set_input_transfer_func(struct dc *dc, struct pipe_ctx *pipe_ctx,
-			       const struct dc_plane_state *plane_state)
-{
-	struct input_pixel_processor *ipp = pipe_ctx->plane_res.ipp;
-	const struct dc_transfer_func *tf = NULL;
-	struct ipp_prescale_params prescale_params = { 0 };
-	bool result = true;
-
-	if (ipp == NULL)
-		return false;
-
-	if (plane_state->in_transfer_func)
-		tf = plane_state->in_transfer_func;
-
-	build_prescale_params(&prescale_params, plane_state);
-	ipp->funcs->ipp_program_prescale(ipp, &prescale_params);
-
-	if (plane_state->gamma_correction &&
-			!plane_state->gamma_correction->is_identity &&
-			dce_use_lut(plane_state->format))
-		ipp->funcs->ipp_program_input_lut(ipp, plane_state->gamma_correction);
-
-	if (tf == NULL) {
-		/* Default case if no input transfer function specified */
-		ipp->funcs->ipp_set_degamma(ipp, IPP_DEGAMMA_MODE_HW_sRGB);
-	} else if (tf->type == TF_TYPE_PREDEFINED) {
-		switch (tf->tf) {
-		case TRANSFER_FUNCTION_SRGB:
-			ipp->funcs->ipp_set_degamma(ipp, IPP_DEGAMMA_MODE_HW_sRGB);
-			break;
-		case TRANSFER_FUNCTION_BT709:
-			ipp->funcs->ipp_set_degamma(ipp, IPP_DEGAMMA_MODE_HW_xvYCC);
-			break;
-		case TRANSFER_FUNCTION_LINEAR:
-			ipp->funcs->ipp_set_degamma(ipp, IPP_DEGAMMA_MODE_BYPASS);
-			break;
-		case TRANSFER_FUNCTION_PQ:
-		default:
-			result = false;
-			break;
-		}
-	} else if (tf->type == TF_TYPE_BYPASS) {
-		ipp->funcs->ipp_set_degamma(ipp, IPP_DEGAMMA_MODE_BYPASS);
-	} else {
-		/*TF_TYPE_DISTRIBUTED_POINTS - Not supported in DCE 11*/
-		result = false;
-	}
-
-	return result;
-}
-
-static bool convert_to_custom_float(struct pwl_result_data *rgb_resulted,
-				    struct curve_points *arr_points,
-				    uint32_t hw_points_num)
-{
-	struct custom_float_format fmt;
-
-	struct pwl_result_data *rgb = rgb_resulted;
-
-	uint32_t i = 0;
-
-	fmt.exponenta_bits = 6;
-	fmt.mantissa_bits = 12;
-	fmt.sign = true;
-
-	if (!convert_to_custom_float_format(arr_points[0].x, &fmt,
-					    &arr_points[0].custom_float_x)) {
-		BREAK_TO_DEBUGGER();
-		return false;
-	}
-
-	if (!convert_to_custom_float_format(arr_points[0].offset, &fmt,
-					    &arr_points[0].custom_float_offset)) {
-		BREAK_TO_DEBUGGER();
-		return false;
-	}
-
-	if (!convert_to_custom_float_format(arr_points[0].slope, &fmt,
-					    &arr_points[0].custom_float_slope)) {
-		BREAK_TO_DEBUGGER();
-		return false;
-	}
-
-	fmt.mantissa_bits = 10;
-	fmt.sign = false;
-
-	if (!convert_to_custom_float_format(arr_points[1].x, &fmt,
-					    &arr_points[1].custom_float_x)) {
-		BREAK_TO_DEBUGGER();
-		return false;
-	}
-
-	if (!convert_to_custom_float_format(arr_points[1].y, &fmt,
-					    &arr_points[1].custom_float_y)) {
-		BREAK_TO_DEBUGGER();
-		return false;
-	}
-
-	if (!convert_to_custom_float_format(arr_points[1].slope, &fmt,
-					    &arr_points[1].custom_float_slope)) {
-		BREAK_TO_DEBUGGER();
-		return false;
-	}
-
-	fmt.mantissa_bits = 12;
-	fmt.sign = true;
-
-	while (i != hw_points_num) {
-		if (!convert_to_custom_float_format(rgb->red, &fmt,
-						    &rgb->red_reg)) {
-			BREAK_TO_DEBUGGER();
-			return false;
-		}
-
-		if (!convert_to_custom_float_format(rgb->green, &fmt,
-						    &rgb->green_reg)) {
-			BREAK_TO_DEBUGGER();
-			return false;
-		}
-
-		if (!convert_to_custom_float_format(rgb->blue, &fmt,
-						    &rgb->blue_reg)) {
-			BREAK_TO_DEBUGGER();
-			return false;
-		}
-
-		if (!convert_to_custom_float_format(rgb->delta_red, &fmt,
-						    &rgb->delta_red_reg)) {
-			BREAK_TO_DEBUGGER();
-			return false;
-		}
-
-		if (!convert_to_custom_float_format(rgb->delta_green, &fmt,
-						    &rgb->delta_green_reg)) {
-			BREAK_TO_DEBUGGER();
-			return false;
-		}
-
-		if (!convert_to_custom_float_format(rgb->delta_blue, &fmt,
-						    &rgb->delta_blue_reg)) {
-			BREAK_TO_DEBUGGER();
-			return false;
-		}
-
-		++rgb;
-		++i;
-	}
-
-	return true;
-}
-
-#define MAX_LOW_POINT      25
-#define NUMBER_REGIONS     16
-#define NUMBER_SW_SEGMENTS 16
-
-static bool
-dce110_translate_regamma_to_hw_format(const struct dc_transfer_func *output_tf,
-				      struct pwl_params *regamma_params)
-{
-	struct curve_points *arr_points;
-	struct pwl_result_data *rgb_resulted;
-	struct pwl_result_data *rgb;
-	struct pwl_result_data *rgb_plus_1;
-	struct fixed31_32 y_r;
-	struct fixed31_32 y_g;
-	struct fixed31_32 y_b;
-	struct fixed31_32 y1_min;
-	struct fixed31_32 y3_max;
-
-	int32_t region_start, region_end;
-	uint32_t i, j, k, seg_distr[NUMBER_REGIONS], increment, start_index, hw_points;
-
-	if (output_tf == NULL || regamma_params == NULL || output_tf->type == TF_TYPE_BYPASS)
-		return false;
-
-	arr_points = regamma_params->arr_points;
-	rgb_resulted = regamma_params->rgb_resulted;
-	hw_points = 0;
-
-	memset(regamma_params, 0, sizeof(struct pwl_params));
-
-	if (output_tf->tf == TRANSFER_FUNCTION_PQ) {
-		/* 16 segments
-		 * segments are from 2^-11 to 2^5
-		 */
-		region_start = -11;
-		region_end = region_start + NUMBER_REGIONS;
-
-		for (i = 0; i < NUMBER_REGIONS; i++)
-			seg_distr[i] = 4;
-
-	} else {
-		/* 10 segments
-		 * segment is from 2^-10 to 2^1
-		 * We include an extra segment for range [2^0, 2^1). This is to
-		 * ensure that colors with normalized values of 1 don't miss the
-		 * LUT.
-		 */
-		region_start = -10;
-		region_end = 1;
-
-		seg_distr[0] = 4;
-		seg_distr[1] = 4;
-		seg_distr[2] = 4;
-		seg_distr[3] = 4;
-		seg_distr[4] = 4;
-		seg_distr[5] = 4;
-		seg_distr[6] = 4;
-		seg_distr[7] = 4;
-		seg_distr[8] = 4;
-		seg_distr[9] = 4;
-		seg_distr[10] = 0;
-		seg_distr[11] = -1;
-		seg_distr[12] = -1;
-		seg_distr[13] = -1;
-		seg_distr[14] = -1;
-		seg_distr[15] = -1;
-	}
-
-	for (k = 0; k < 16; k++) {
-		if (seg_distr[k] != -1)
-			hw_points += (1 << seg_distr[k]);
-	}
-
-	j = 0;
-	for (k = 0; k < (region_end - region_start); k++) {
-		increment = NUMBER_SW_SEGMENTS / (1 << seg_distr[k]);
-		start_index = (region_start + k + MAX_LOW_POINT) *
-				NUMBER_SW_SEGMENTS;
-		for (i = start_index; i < start_index + NUMBER_SW_SEGMENTS;
-				i += increment) {
-			if (j == hw_points - 1)
-				break;
-			rgb_resulted[j].red = output_tf->tf_pts.red[i];
-			rgb_resulted[j].green = output_tf->tf_pts.green[i];
-			rgb_resulted[j].blue = output_tf->tf_pts.blue[i];
-			j++;
-		}
-	}
-
-	/* last point */
-	start_index = (region_end + MAX_LOW_POINT) * NUMBER_SW_SEGMENTS;
-	rgb_resulted[hw_points - 1].red = output_tf->tf_pts.red[start_index];
-	rgb_resulted[hw_points - 1].green = output_tf->tf_pts.green[start_index];
-	rgb_resulted[hw_points - 1].blue = output_tf->tf_pts.blue[start_index];
-
-	arr_points[0].x = dc_fixpt_pow(dc_fixpt_from_int(2),
-					     dc_fixpt_from_int(region_start));
-	arr_points[1].x = dc_fixpt_pow(dc_fixpt_from_int(2),
-					     dc_fixpt_from_int(region_end));
-
-	y_r = rgb_resulted[0].red;
-	y_g = rgb_resulted[0].green;
-	y_b = rgb_resulted[0].blue;
-
-	y1_min = dc_fixpt_min(y_r, dc_fixpt_min(y_g, y_b));
-
-	arr_points[0].y = y1_min;
-	arr_points[0].slope = dc_fixpt_div(arr_points[0].y,
-						 arr_points[0].x);
-
-	y_r = rgb_resulted[hw_points - 1].red;
-	y_g = rgb_resulted[hw_points - 1].green;
-	y_b = rgb_resulted[hw_points - 1].blue;
-
-	/* see comment above, m_arrPoints[1].y should be the Y value for the
-	 * region end (m_numOfHwPoints), not last HW point(m_numOfHwPoints - 1)
-	 */
-	y3_max = dc_fixpt_max(y_r, dc_fixpt_max(y_g, y_b));
-
-	arr_points[1].y = y3_max;
-
-	arr_points[1].slope = dc_fixpt_zero;
-
-	if (output_tf->tf == TRANSFER_FUNCTION_PQ) {
-		/* for PQ, we want to have a straight line from last HW X point,
-		 * and the slope to be such that we hit 1.0 at 10000 nits.
-		 */
-		const struct fixed31_32 end_value = dc_fixpt_from_int(125);
-
-		arr_points[1].slope = dc_fixpt_div(
-				dc_fixpt_sub(dc_fixpt_one, arr_points[1].y),
-				dc_fixpt_sub(end_value, arr_points[1].x));
-	}
-
-	regamma_params->hw_points_num = hw_points;
-
-	k = 0;
-	for (i = 1; i < 16; i++) {
-		if (seg_distr[k] != -1) {
-			regamma_params->arr_curve_points[k].segments_num = seg_distr[k];
-			regamma_params->arr_curve_points[i].offset =
-					regamma_params->arr_curve_points[k].offset + (1 << seg_distr[k]);
-		}
-		k++;
-	}
-
-	if (seg_distr[k] != -1)
-		regamma_params->arr_curve_points[k].segments_num = seg_distr[k];
-
-	rgb = rgb_resulted;
-	rgb_plus_1 = rgb_resulted + 1;
-
-	i = 1;
-
-	while (i != hw_points + 1) {
-		if (dc_fixpt_lt(rgb_plus_1->red, rgb->red))
-			rgb_plus_1->red = rgb->red;
-		if (dc_fixpt_lt(rgb_plus_1->green, rgb->green))
-			rgb_plus_1->green = rgb->green;
-		if (dc_fixpt_lt(rgb_plus_1->blue, rgb->blue))
-			rgb_plus_1->blue = rgb->blue;
-
-		rgb->delta_red = dc_fixpt_sub(rgb_plus_1->red, rgb->red);
-		rgb->delta_green = dc_fixpt_sub(rgb_plus_1->green, rgb->green);
-		rgb->delta_blue = dc_fixpt_sub(rgb_plus_1->blue, rgb->blue);
-
-		++rgb_plus_1;
-		++rgb;
-		++i;
-	}
-
-	convert_to_custom_float(rgb_resulted, arr_points, hw_points);
-
-	return true;
-}
-
-static bool
-dce110_set_output_transfer_func(struct dc *dc, struct pipe_ctx *pipe_ctx,
-				const struct dc_stream_state *stream)
-{
-	struct transform *xfm = pipe_ctx->plane_res.xfm;
-
-	xfm->funcs->opp_power_on_regamma_lut(xfm, true);
-	xfm->regamma_params.hw_points_num = GAMMA_HW_POINTS_NUM;
-
-	if (stream->out_transfer_func &&
-	    stream->out_transfer_func->type == TF_TYPE_PREDEFINED &&
-	    stream->out_transfer_func->tf == TRANSFER_FUNCTION_SRGB) {
-		xfm->funcs->opp_set_regamma_mode(xfm, OPP_REGAMMA_SRGB);
-	} else if (dce110_translate_regamma_to_hw_format(stream->out_transfer_func,
-							 &xfm->regamma_params)) {
-		xfm->funcs->opp_program_regamma_pwl(xfm, &xfm->regamma_params);
-		xfm->funcs->opp_set_regamma_mode(xfm, OPP_REGAMMA_USER);
-	} else {
-		xfm->funcs->opp_set_regamma_mode(xfm, OPP_REGAMMA_BYPASS);
-	}
-
-	xfm->funcs->opp_power_on_regamma_lut(xfm, false);
-
-	return true;
-}
-
-void dce110_update_info_frame(struct pipe_ctx *pipe_ctx)
-{
-	bool is_hdmi_tmds;
-	bool is_dp;
-
-	ASSERT(pipe_ctx->stream);
-
-	if (pipe_ctx->stream_res.stream_enc == NULL)
-		return;  /* this is not root pipe */
-
-	is_hdmi_tmds = dc_is_hdmi_tmds_signal(pipe_ctx->stream->signal);
-	is_dp = dc_is_dp_signal(pipe_ctx->stream->signal);
-
-	if (!is_hdmi_tmds && !is_dp)
-		return;
-
-	if (is_hdmi_tmds)
-		pipe_ctx->stream_res.stream_enc->funcs->update_hdmi_info_packets(
-			pipe_ctx->stream_res.stream_enc,
-			&pipe_ctx->stream_res.encoder_info_frame);
-	else
-		pipe_ctx->stream_res.stream_enc->funcs->update_dp_info_packets(
-			pipe_ctx->stream_res.stream_enc,
-			&pipe_ctx->stream_res.encoder_info_frame);
-}
-
-void dce110_enable_stream(struct pipe_ctx *pipe_ctx)
-{
-	enum dc_lane_count lane_count =
-		pipe_ctx->stream->link->cur_link_settings.lane_count;
-	struct dc_crtc_timing *timing = &pipe_ctx->stream->timing;
-	struct dc_link *link = pipe_ctx->stream->link;
-	const struct dc *dc = link->dc;
-
-	uint32_t active_total_with_borders;
-	uint32_t early_control = 0;
-	struct timing_generator *tg = pipe_ctx->stream_res.tg;
-
-	/* For MST, there are multiply stream go to only one link.
-	 * connect DIG back_end to front_end while enable_stream and
-	 * disconnect them during disable_stream
-	 * BY this, it is logic clean to separate stream and link */
-	link->link_enc->funcs->connect_dig_be_to_fe(link->link_enc,
-						    pipe_ctx->stream_res.stream_enc->id, true);
-
-	dc->hwss.update_info_frame(pipe_ctx);
-
-	/* enable early control to avoid corruption on DP monitor*/
-	active_total_with_borders =
-			timing->h_addressable
-				+ timing->h_border_left
-				+ timing->h_border_right;
-
-	if (lane_count != 0)
-		early_control = active_total_with_borders % lane_count;
-
-	if (early_control == 0)
-		early_control = lane_count;
-
-	tg->funcs->set_early_control(tg, early_control);
-
-	/* enable audio only within mode set */
-	if (pipe_ctx->stream_res.audio != NULL) {
-		if (dc_is_dp_signal(pipe_ctx->stream->signal))
-			pipe_ctx->stream_res.stream_enc->funcs->dp_audio_enable(pipe_ctx->stream_res.stream_enc);
-	}
-
-
-
-
-}
-
-static enum bp_result link_transmitter_control(
-		struct dc_bios *bios,
-	struct bp_transmitter_control *cntl)
-{
-	enum bp_result result;
-
-	result = bios->funcs->transmitter_control(bios, cntl);
-
-	return result;
-}
-
-/*
- * @brief
- * eDP only.
- */
-void dce110_edp_wait_for_hpd_ready(
-		struct dc_link *link,
-		bool power_up)
-{
-	struct dc_context *ctx = link->ctx;
-	struct graphics_object_id connector = link->link_enc->connector;
-	struct gpio *hpd;
-	struct dc_sink *sink = link->local_sink;
-	bool edp_hpd_high = false;
-	uint32_t time_elapsed = 0;
-	uint32_t timeout = power_up ?
-		PANEL_POWER_UP_TIMEOUT : PANEL_POWER_DOWN_TIMEOUT;
-
-	if (dal_graphics_object_id_get_connector_id(connector)
-			!= CONNECTOR_ID_EDP) {
-		BREAK_TO_DEBUGGER();
-		return;
-	}
-
-	if (!power_up)
-		/*
-		 * From KV, we will not HPD low after turning off VCC -
-		 * instead, we will check the SW timer in power_up().
-		 */
-		return;
-
-	/*
-	 * When we power on/off the eDP panel,
-	 * we need to wait until SENSE bit is high/low.
-	 */
-
-	/* obtain HPD */
-	/* TODO what to do with this? */
-	hpd = get_hpd_gpio(ctx->dc_bios, connector, ctx->gpio_service);
-
-	if (!hpd) {
-		BREAK_TO_DEBUGGER();
-		return;
-	}
-
-	if (sink != NULL) {
-		if (sink->edid_caps.panel_patch.extra_t3_ms > 0) {
-			int extra_t3_in_ms = sink->edid_caps.panel_patch.extra_t3_ms;
-
-			msleep(extra_t3_in_ms);
-		}
-	}
-
-	dal_gpio_open(hpd, GPIO_MODE_INTERRUPT);
-
-	/* wait until timeout or panel detected */
-
-	do {
-		uint32_t detected = 0;
-
-		dal_gpio_get_value(hpd, &detected);
-
-		if (!(detected ^ power_up)) {
-			edp_hpd_high = true;
-			break;
-		}
-
-		msleep(HPD_CHECK_INTERVAL);
-
-		time_elapsed += HPD_CHECK_INTERVAL;
-	} while (time_elapsed < timeout);
-
-	dal_gpio_close(hpd);
-
-	dal_gpio_destroy_irq(&hpd);
-
-	if (false == edp_hpd_high) {
-		DC_LOG_ERROR(
-				"%s: wait timed out!\n", __func__);
-	}
-}
-
-void dce110_edp_power_control(
-		struct dc_link *link,
-		bool power_up)
-{
-	struct dc_context *ctx = link->ctx;
-	struct bp_transmitter_control cntl = { 0 };
-	enum bp_result bp_result;
-	uint8_t panel_instance;
-
-
-	if (dal_graphics_object_id_get_connector_id(link->link_enc->connector)
-			!= CONNECTOR_ID_EDP) {
-		BREAK_TO_DEBUGGER();
-		return;
-	}
-
-	if (!link->panel_cntl)
-		return;
-	if (power_up !=
-		link->panel_cntl->funcs->is_panel_powered_on(link->panel_cntl)) {
-
-		unsigned long long current_ts = dm_get_timestamp(ctx);
-		unsigned long long time_since_edp_poweroff_ms =
-				div64_u64(dm_get_elapse_time_in_ns(
-						ctx,
-						current_ts,
-						link->link_trace.time_stamp.edp_poweroff), 1000000);
-		unsigned long long time_since_edp_poweron_ms =
-				div64_u64(dm_get_elapse_time_in_ns(
-						ctx,
-						current_ts,
-						link->link_trace.time_stamp.edp_poweron), 1000000);
-		DC_LOG_HW_RESUME_S3(
-				"%s: transition: power_up=%d current_ts=%llu edp_poweroff=%llu edp_poweron=%llu time_since_edp_poweroff_ms=%llu time_since_edp_poweron_ms=%llu",
-				__func__,
-				power_up,
-				current_ts,
-				link->link_trace.time_stamp.edp_poweroff,
-				link->link_trace.time_stamp.edp_poweron,
-				time_since_edp_poweroff_ms,
-				time_since_edp_poweron_ms);
-
-		/* Send VBIOS command to prompt eDP panel power */
-		if (power_up) {
-			/* edp requires a min of 500ms from LCDVDD off to on */
-			unsigned long long remaining_min_edp_poweroff_time_ms = 500;
-
-			/* add time defined by a patch, if any (usually patch extra_t12_ms is 0) */
-			if (link->local_sink != NULL)
-				remaining_min_edp_poweroff_time_ms +=
-					link->local_sink->edid_caps.panel_patch.extra_t12_ms;
-
-			/* Adjust remaining_min_edp_poweroff_time_ms if this is not the first time. */
-			if (link->link_trace.time_stamp.edp_poweroff != 0) {
-				if (time_since_edp_poweroff_ms < remaining_min_edp_poweroff_time_ms)
-					remaining_min_edp_poweroff_time_ms =
-						remaining_min_edp_poweroff_time_ms - time_since_edp_poweroff_ms;
-				else
-					remaining_min_edp_poweroff_time_ms = 0;
-			}
-
-			if (remaining_min_edp_poweroff_time_ms) {
-				DC_LOG_HW_RESUME_S3(
-						"%s: remaining_min_edp_poweroff_time_ms=%llu: begin wait.\n",
-						__func__, remaining_min_edp_poweroff_time_ms);
-				msleep(remaining_min_edp_poweroff_time_ms);
-				DC_LOG_HW_RESUME_S3(
-						"%s: remaining_min_edp_poweroff_time_ms=%llu: end wait.\n",
-						__func__, remaining_min_edp_poweroff_time_ms);
-				dm_output_to_console("%s: wait %lld ms to power on eDP.\n",
-						__func__, remaining_min_edp_poweroff_time_ms);
-			} else {
-				DC_LOG_HW_RESUME_S3(
-						"%s: remaining_min_edp_poweroff_time_ms=%llu: no wait required.\n",
-						__func__, remaining_min_edp_poweroff_time_ms);
-			}
-		}
-
-		DC_LOG_HW_RESUME_S3(
-				"%s: BEGIN: Panel Power action: %s\n",
-				__func__, (power_up ? "On":"Off"));
-
-		cntl.action = power_up ?
-			TRANSMITTER_CONTROL_POWER_ON :
-			TRANSMITTER_CONTROL_POWER_OFF;
-		cntl.transmitter = link->link_enc->transmitter;
-		cntl.connector_obj_id = link->link_enc->connector;
-		cntl.coherent = false;
-		cntl.lanes_number = LANE_COUNT_FOUR;
-		cntl.hpd_sel = link->link_enc->hpd_source;
-		panel_instance = link->panel_cntl->inst;
-
-		if (ctx->dc->ctx->dmub_srv &&
-				ctx->dc->debug.dmub_command_table) {
-			if (cntl.action == TRANSMITTER_CONTROL_POWER_ON)
-				bp_result = ctx->dc_bios->funcs->enable_lvtma_control(ctx->dc_bios,
-						LVTMA_CONTROL_POWER_ON,
-						panel_instance);
-			else
-				bp_result = ctx->dc_bios->funcs->enable_lvtma_control(ctx->dc_bios,
-						LVTMA_CONTROL_POWER_OFF,
-						panel_instance);
-		}
-
-		bp_result = link_transmitter_control(ctx->dc_bios, &cntl);
-
-		DC_LOG_HW_RESUME_S3(
-				"%s: END: Panel Power action: %s bp_result=%u\n",
-				__func__, (power_up ? "On":"Off"),
-				bp_result);
-
-		if (!power_up)
-			/*save driver power off time stamp*/
-			link->link_trace.time_stamp.edp_poweroff = dm_get_timestamp(ctx);
-		else
-			link->link_trace.time_stamp.edp_poweron = dm_get_timestamp(ctx);
-
-		DC_LOG_HW_RESUME_S3(
-				"%s: updated values: edp_poweroff=%llu edp_poweron=%llu\n",
-				__func__,
-				link->link_trace.time_stamp.edp_poweroff,
-				link->link_trace.time_stamp.edp_poweron);
-
-		if (bp_result != BP_RESULT_OK)
-			DC_LOG_ERROR(
-					"%s: Panel Power bp_result: %d\n",
-					__func__, bp_result);
-	} else {
-		DC_LOG_HW_RESUME_S3(
-				"%s: Skipping Panel Power action: %s\n",
-				__func__, (power_up ? "On":"Off"));
-	}
-}
-
-void dce110_edp_wait_for_T12(
-		struct dc_link *link)
-{
-	struct dc_context *ctx = link->ctx;
-
-	if (dal_graphics_object_id_get_connector_id(link->link_enc->connector)
-			!= CONNECTOR_ID_EDP) {
-		BREAK_TO_DEBUGGER();
-		return;
-	}
-
-	if (!link->panel_cntl)
-		return;
-
-	if (!link->panel_cntl->funcs->is_panel_powered_on(link->panel_cntl) &&
-			link->link_trace.time_stamp.edp_poweroff != 0) {
-		unsigned int t12_duration = 500; // Default T12 as per spec
-		unsigned long long current_ts = dm_get_timestamp(ctx);
-		unsigned long long time_since_edp_poweroff_ms =
-				div64_u64(dm_get_elapse_time_in_ns(
-						ctx,
-						current_ts,
-						link->link_trace.time_stamp.edp_poweroff), 1000000);
-
-		t12_duration += link->local_sink->edid_caps.panel_patch.extra_t12_ms; // Add extra T12
-
-		if (time_since_edp_poweroff_ms < t12_duration)
-			msleep(t12_duration - time_since_edp_poweroff_ms);
-	}
-}
-
-/*todo: cloned in stream enc, fix*/
-/*
- * @brief
- * eDP only. Control the backlight of the eDP panel
- */
-void dce110_edp_backlight_control(
-		struct dc_link *link,
-		bool enable)
-{
-	struct dc_context *ctx = link->ctx;
-	struct bp_transmitter_control cntl = { 0 };
-	uint8_t panel_instance;
-
-	if (dal_graphics_object_id_get_connector_id(link->link_enc->connector)
-		!= CONNECTOR_ID_EDP) {
-		BREAK_TO_DEBUGGER();
-		return;
-	}
-
-	if (link->panel_cntl) {
-		bool is_backlight_on = link->panel_cntl->funcs->is_panel_backlight_on(link->panel_cntl);
-
-		if ((enable && is_backlight_on) || (!enable && !is_backlight_on)) {
-			DC_LOG_HW_RESUME_S3(
-				"%s: panel already powered up/off. Do nothing.\n",
-				__func__);
-			return;
-		}
-	}
-
-	/* Send VBIOS command to control eDP panel backlight */
-
-	DC_LOG_HW_RESUME_S3(
-			"%s: backlight action: %s\n",
-			__func__, (enable ? "On":"Off"));
-
-	cntl.action = enable ?
-		TRANSMITTER_CONTROL_BACKLIGHT_ON :
-		TRANSMITTER_CONTROL_BACKLIGHT_OFF;
-
-	/*cntl.engine_id = ctx->engine;*/
-	cntl.transmitter = link->link_enc->transmitter;
-	cntl.connector_obj_id = link->link_enc->connector;
-	/*todo: unhardcode*/
-	cntl.lanes_number = LANE_COUNT_FOUR;
-	cntl.hpd_sel = link->link_enc->hpd_source;
-	cntl.signal = SIGNAL_TYPE_EDP;
-
-	/* For eDP, the following delays might need to be considered
-	 * after link training completed:
-	 * idle period - min. accounts for required BS-Idle pattern,
-	 * max. allows for source frame synchronization);
-	 * 50 msec max. delay from valid video data from source
-	 * to video on dislpay or backlight enable.
-	 *
-	 * Disable the delay for now.
-	 * Enable it in the future if necessary.
-	 */
-	/* dc_service_sleep_in_milliseconds(50); */
-		/*edp 1.2*/
-	panel_instance = link->panel_cntl->inst;
-
-	if (cntl.action == TRANSMITTER_CONTROL_BACKLIGHT_ON) {
-		if (!link->dc->config.edp_no_power_sequencing)
-		/*
-		 * Sometimes, DP receiver chip power-controlled externally by an
-		 * Embedded Controller could be treated and used as eDP,
-		 * if it drives mobile display. In this case,
-		 * we shouldn't be doing power-sequencing, hence we can skip
-		 * waiting for T7-ready.
-		 */
-			edp_receiver_ready_T7(link);
-		else
-			DC_LOG_DC("edp_receiver_ready_T7 skipped\n");
-	}
-
-	if (ctx->dc->ctx->dmub_srv &&
-			ctx->dc->debug.dmub_command_table) {
-		if (cntl.action == TRANSMITTER_CONTROL_BACKLIGHT_ON)
-			ctx->dc_bios->funcs->enable_lvtma_control(ctx->dc_bios,
-					LVTMA_CONTROL_LCD_BLON,
-					panel_instance);
-		else
-			ctx->dc_bios->funcs->enable_lvtma_control(ctx->dc_bios,
-					LVTMA_CONTROL_LCD_BLOFF,
-					panel_instance);
-	}
-
-	link_transmitter_control(ctx->dc_bios, &cntl);
-
-	if (enable && link->dpcd_sink_ext_caps.bits.oled)
-		msleep(OLED_POST_T7_DELAY);
-
-	if (link->dpcd_sink_ext_caps.bits.oled ||
-		link->dpcd_sink_ext_caps.bits.hdr_aux_backlight_control == 1 ||
-		link->dpcd_sink_ext_caps.bits.sdr_aux_backlight_control == 1)
-		dc_link_backlight_enable_aux(link, enable);
-
-	/*edp 1.2*/
-	if (cntl.action == TRANSMITTER_CONTROL_BACKLIGHT_OFF) {
-		if (!link->dc->config.edp_no_power_sequencing)
-		/*
-		 * Sometimes, DP receiver chip power-controlled externally by an
-		 * Embedded Controller could be treated and used as eDP,
-		 * if it drives mobile display. In this case,
-		 * we shouldn't be doing power-sequencing, hence we can skip
-		 * waiting for T9-ready.
-		 */
-			edp_add_delay_for_T9(link);
-		else
-			DC_LOG_DC("edp_receiver_ready_T9 skipped\n");
-	}
-
-	if (!enable && link->dpcd_sink_ext_caps.bits.oled)
-		msleep(OLED_PRE_T11_DELAY);
-}
-
-void dce110_enable_audio_stream(struct pipe_ctx *pipe_ctx)
-{
-	/* notify audio driver for audio modes of monitor */
-	struct dc *dc;
-	struct clk_mgr *clk_mgr;
-	unsigned int i, num_audio = 1;
-
-	if (!pipe_ctx->stream)
-		return;
-
-	dc = pipe_ctx->stream->ctx->dc;
-	clk_mgr = dc->clk_mgr;
-
-	if (pipe_ctx->stream_res.audio && pipe_ctx->stream_res.audio->enabled == true)
-		return;
-
-	if (pipe_ctx->stream_res.audio) {
-		for (i = 0; i < MAX_PIPES; i++) {
-			/*current_state not updated yet*/
-			if (dc->current_state->res_ctx.pipe_ctx[i].stream_res.audio != NULL)
-				num_audio++;
-		}
-
-		pipe_ctx->stream_res.audio->funcs->az_enable(pipe_ctx->stream_res.audio);
-
-		if (num_audio >= 1 && clk_mgr->funcs->enable_pme_wa)
-			/*this is the first audio. apply the PME w/a in order to wake AZ from D3*/
-			clk_mgr->funcs->enable_pme_wa(clk_mgr);
-		/* un-mute audio */
-		/* TODO: audio should be per stream rather than per link */
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-		if (is_dp_128b_132b_signal(pipe_ctx))
-			pipe_ctx->stream_res.hpo_dp_stream_enc->funcs->audio_mute_control(
-					pipe_ctx->stream_res.hpo_dp_stream_enc, false);
-		else
-			pipe_ctx->stream_res.stream_enc->funcs->audio_mute_control(
-					pipe_ctx->stream_res.stream_enc, false);
-#else
-		pipe_ctx->stream_res.stream_enc->funcs->audio_mute_control(
-				pipe_ctx->stream_res.stream_enc, false);
-#endif
-		if (pipe_ctx->stream_res.audio)
-			pipe_ctx->stream_res.audio->enabled = true;
-	}
-
-	if (dc_is_dp_signal(pipe_ctx->stream->signal))
-		dp_source_sequence_trace(pipe_ctx->stream->link, DPCD_SOURCE_SEQ_AFTER_ENABLE_AUDIO_STREAM);
-}
-
-void dce110_disable_audio_stream(struct pipe_ctx *pipe_ctx)
-{
-	struct dc *dc;
-	struct clk_mgr *clk_mgr;
-
-	if (!pipe_ctx || !pipe_ctx->stream)
-		return;
-
-	dc = pipe_ctx->stream->ctx->dc;
-	clk_mgr = dc->clk_mgr;
-
-	if (pipe_ctx->stream_res.audio && pipe_ctx->stream_res.audio->enabled == false)
-		return;
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	if (is_dp_128b_132b_signal(pipe_ctx))
-		pipe_ctx->stream_res.hpo_dp_stream_enc->funcs->audio_mute_control(
-				pipe_ctx->stream_res.hpo_dp_stream_enc, true);
-	else
-		pipe_ctx->stream_res.stream_enc->funcs->audio_mute_control(
-				pipe_ctx->stream_res.stream_enc, true);
-#else
-	pipe_ctx->stream_res.stream_enc->funcs->audio_mute_control(
-			pipe_ctx->stream_res.stream_enc, true);
-#endif
-	if (pipe_ctx->stream_res.audio) {
-		pipe_ctx->stream_res.audio->enabled = false;
-
-		if (dc_is_dp_signal(pipe_ctx->stream->signal))
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-			if (is_dp_128b_132b_signal(pipe_ctx))
-				pipe_ctx->stream_res.hpo_dp_stream_enc->funcs->dp_audio_disable(
-						pipe_ctx->stream_res.hpo_dp_stream_enc);
-			else
-				pipe_ctx->stream_res.stream_enc->funcs->dp_audio_disable(
-						pipe_ctx->stream_res.stream_enc);
-#else
-			pipe_ctx->stream_res.stream_enc->funcs->dp_audio_disable(
-					pipe_ctx->stream_res.stream_enc);
-#endif
-		else
-			pipe_ctx->stream_res.stream_enc->funcs->hdmi_audio_disable(
-					pipe_ctx->stream_res.stream_enc);
-
-		if (clk_mgr->funcs->enable_pme_wa)
-			/*this is the first audio. apply the PME w/a in order to wake AZ from D3*/
-			clk_mgr->funcs->enable_pme_wa(clk_mgr);
-
-		/* TODO: notify audio driver for if audio modes list changed
-		 * add audio mode list change flag */
-		/* dal_audio_disable_azalia_audio_jack_presence(stream->audio,
-		 * stream->stream_engine_id);
-		 */
-	}
-
-	if (dc_is_dp_signal(pipe_ctx->stream->signal))
-		dp_source_sequence_trace(pipe_ctx->stream->link, DPCD_SOURCE_SEQ_AFTER_DISABLE_AUDIO_STREAM);
-}
-
-void dce110_disable_stream(struct pipe_ctx *pipe_ctx)
-{
-	struct dc_stream_state *stream = pipe_ctx->stream;
-	struct dc_link *link = stream->link;
-	struct dc *dc = pipe_ctx->stream->ctx->dc;
-	struct link_encoder *link_enc = NULL;
-
-	if (dc_is_hdmi_tmds_signal(pipe_ctx->stream->signal)) {
-		pipe_ctx->stream_res.stream_enc->funcs->stop_hdmi_info_packets(
-			pipe_ctx->stream_res.stream_enc);
-		pipe_ctx->stream_res.stream_enc->funcs->hdmi_reset_stream_attribute(
-			pipe_ctx->stream_res.stream_enc);
-	}
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	if (is_dp_128b_132b_signal(pipe_ctx)) {
-		pipe_ctx->stream_res.hpo_dp_stream_enc->funcs->stop_dp_info_packets(
-					pipe_ctx->stream_res.hpo_dp_stream_enc);
-	} else if (dc_is_dp_signal(pipe_ctx->stream->signal))
-#else
-	if (dc_is_dp_signal(pipe_ctx->stream->signal))
-#endif
-		pipe_ctx->stream_res.stream_enc->funcs->stop_dp_info_packets(
-			pipe_ctx->stream_res.stream_enc);
-
-	dc->hwss.disable_audio_stream(pipe_ctx);
-
-	/* Link encoder may have been dynamically assigned to non-physical display endpoint. */
-	if (link->ep_type == DISPLAY_ENDPOINT_PHY)
-		link_enc = link->link_enc;
-	else if (dc->res_pool->funcs->link_encs_assign)
-		link_enc = link_enc_cfg_get_link_enc_used_by_link(link->ctx->dc, link);
-	ASSERT(link_enc);
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	if (is_dp_128b_132b_signal(pipe_ctx)) {
-		pipe_ctx->stream_res.hpo_dp_stream_enc->funcs->disable(
-				pipe_ctx->stream_res.hpo_dp_stream_enc);
-		setup_dp_hpo_stream(pipe_ctx, false);
-	/* TODO - DP2.0 HW: unmap stream from link encoder here */
-	} else {
-		if (link_enc)
-			link_enc->funcs->connect_dig_be_to_fe(
-				link_enc,
-				pipe_ctx->stream_res.stream_enc->id,
-				false);
-	}
-#else
-	if (link_enc)
-		link_enc->funcs->connect_dig_be_to_fe(
-			link->link_enc,
-			pipe_ctx->stream_res.stream_enc->id,
-			false);
-#endif
-	if (dc_is_dp_signal(pipe_ctx->stream->signal))
-		dp_source_sequence_trace(link, DPCD_SOURCE_SEQ_AFTER_DISCONNECT_DIG_FE_BE);
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	if (dc->hwseq->funcs.setup_hpo_hw_control && is_dp_128b_132b_signal(pipe_ctx))
-		dc->hwseq->funcs.setup_hpo_hw_control(dc->hwseq, false);
-#endif
-
-}
-
-void dce110_unblank_stream(struct pipe_ctx *pipe_ctx,
-		struct dc_link_settings *link_settings)
-{
-	struct encoder_unblank_param params = { { 0 } };
-	struct dc_stream_state *stream = pipe_ctx->stream;
-	struct dc_link *link = stream->link;
-	struct dce_hwseq *hws = link->dc->hwseq;
-
-	/* only 3 items below are used by unblank */
-	params.timing = pipe_ctx->stream->timing;
-	params.link_settings.link_rate = link_settings->link_rate;
-
-	if (dc_is_dp_signal(pipe_ctx->stream->signal))
-		pipe_ctx->stream_res.stream_enc->funcs->dp_unblank(link, pipe_ctx->stream_res.stream_enc, &params);
-
-	if (link->local_sink && link->local_sink->sink_signal == SIGNAL_TYPE_EDP) {
-		hws->funcs.edp_backlight_control(link, true);
-	}
-}
-
-void dce110_blank_stream(struct pipe_ctx *pipe_ctx)
-{
-	struct dc_stream_state *stream = pipe_ctx->stream;
-	struct dc_link *link = stream->link;
-	struct dce_hwseq *hws = link->dc->hwseq;
-
-	if (link->local_sink && link->local_sink->sink_signal == SIGNAL_TYPE_EDP) {
-		hws->funcs.edp_backlight_control(link, false);
-		link->dc->hwss.set_abm_immediate_disable(pipe_ctx);
-	}
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	if (is_dp_128b_132b_signal(pipe_ctx)) {
-		/* TODO - DP2.0 HW: Set ODM mode in dp hpo encoder here */
-		pipe_ctx->stream_res.hpo_dp_stream_enc->funcs->dp_blank(
-				pipe_ctx->stream_res.hpo_dp_stream_enc);
-	} else if (dc_is_dp_signal(pipe_ctx->stream->signal)) {
-#else
-	if (dc_is_dp_signal(pipe_ctx->stream->signal)) {
-#endif
-		pipe_ctx->stream_res.stream_enc->funcs->dp_blank(link, pipe_ctx->stream_res.stream_enc);
-
-		if (!dc_is_embedded_signal(pipe_ctx->stream->signal)) {
-			/*
-			 * After output is idle pattern some sinks need time to recognize the stream
-			 * has changed or they enter protection state and hang.
-			 */
-			msleep(60);
-		} else if (pipe_ctx->stream->signal == SIGNAL_TYPE_EDP)
-			edp_receiver_ready_T9(link);
-	}
-
-}
-
-
-void dce110_set_avmute(struct pipe_ctx *pipe_ctx, bool enable)
-{
-	if (pipe_ctx != NULL && pipe_ctx->stream_res.stream_enc != NULL)
-		pipe_ctx->stream_res.stream_enc->funcs->set_avmute(pipe_ctx->stream_res.stream_enc, enable);
-}
-
-static enum audio_dto_source translate_to_dto_source(enum controller_id crtc_id)
-{
-	switch (crtc_id) {
-	case CONTROLLER_ID_D0:
-		return DTO_SOURCE_ID0;
-	case CONTROLLER_ID_D1:
-		return DTO_SOURCE_ID1;
-	case CONTROLLER_ID_D2:
-		return DTO_SOURCE_ID2;
-	case CONTROLLER_ID_D3:
-		return DTO_SOURCE_ID3;
-	case CONTROLLER_ID_D4:
-		return DTO_SOURCE_ID4;
-	case CONTROLLER_ID_D5:
-		return DTO_SOURCE_ID5;
-	default:
-		return DTO_SOURCE_UNKNOWN;
-	}
-}
-
-static void build_audio_output(
-	struct dc_state *state,
-	const struct pipe_ctx *pipe_ctx,
-	struct audio_output *audio_output)
-{
-	const struct dc_stream_state *stream = pipe_ctx->stream;
-	audio_output->engine_id = pipe_ctx->stream_res.stream_enc->id;
-
-	audio_output->signal = pipe_ctx->stream->signal;
-
-	/* audio_crtc_info  */
-
-	audio_output->crtc_info.h_total =
-		stream->timing.h_total;
-
-	/*
-	 * Audio packets are sent during actual CRTC blank physical signal, we
-	 * need to specify actual active signal portion
-	 */
-	audio_output->crtc_info.h_active =
-			stream->timing.h_addressable
-			+ stream->timing.h_border_left
-			+ stream->timing.h_border_right;
-
-	audio_output->crtc_info.v_active =
-			stream->timing.v_addressable
-			+ stream->timing.v_border_top
-			+ stream->timing.v_border_bottom;
-
-	audio_output->crtc_info.pixel_repetition = 1;
-
-	audio_output->crtc_info.interlaced =
-			stream->timing.flags.INTERLACE;
-
-	audio_output->crtc_info.refresh_rate =
-		(stream->timing.pix_clk_100hz*100)/
-		(stream->timing.h_total*stream->timing.v_total);
-
-	audio_output->crtc_info.color_depth =
-		stream->timing.display_color_depth;
-
-	audio_output->crtc_info.requested_pixel_clock_100Hz =
-			pipe_ctx->stream_res.pix_clk_params.requested_pix_clk_100hz;
-
-	audio_output->crtc_info.calculated_pixel_clock_100Hz =
-			pipe_ctx->stream_res.pix_clk_params.requested_pix_clk_100hz;
-
-/*for HDMI, audio ACR is with deep color ratio factor*/
-	if (dc_is_hdmi_tmds_signal(pipe_ctx->stream->signal) &&
-		audio_output->crtc_info.requested_pixel_clock_100Hz ==
-				(stream->timing.pix_clk_100hz)) {
-		if (pipe_ctx->stream_res.pix_clk_params.pixel_encoding == PIXEL_ENCODING_YCBCR420) {
-			audio_output->crtc_info.requested_pixel_clock_100Hz =
-					audio_output->crtc_info.requested_pixel_clock_100Hz/2;
-			audio_output->crtc_info.calculated_pixel_clock_100Hz =
-					pipe_ctx->stream_res.pix_clk_params.requested_pix_clk_100hz/2;
-
-		}
-	}
-
-	if (state->clk_mgr &&
-		(pipe_ctx->stream->signal == SIGNAL_TYPE_DISPLAY_PORT ||
-			pipe_ctx->stream->signal == SIGNAL_TYPE_DISPLAY_PORT_MST)) {
-		audio_output->pll_info.dp_dto_source_clock_in_khz =
-				state->clk_mgr->funcs->get_dp_ref_clk_frequency(
-						state->clk_mgr);
-	}
-
-	audio_output->pll_info.feed_back_divider =
-			pipe_ctx->pll_settings.feedback_divider;
-
-	audio_output->pll_info.dto_source =
-		translate_to_dto_source(
-			pipe_ctx->stream_res.tg->inst + 1);
-
-	/* TODO hard code to enable for now. Need get from stream */
-	audio_output->pll_info.ss_enabled = true;
-
-	audio_output->pll_info.ss_percentage =
-			pipe_ctx->pll_settings.ss_percentage;
-}
-
-static void program_scaler(const struct dc *dc,
-		const struct pipe_ctx *pipe_ctx)
-{
-	struct tg_color color = {0};
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	/* TOFPGA */
-	if (pipe_ctx->plane_res.xfm->funcs->transform_set_pixel_storage_depth == NULL)
-		return;
-#endif
-
-	if (dc->debug.visual_confirm == VISUAL_CONFIRM_SURFACE)
-		get_surface_visual_confirm_color(pipe_ctx, &color);
-	else
-		color_space_to_black_color(dc,
-				pipe_ctx->stream->output_color_space,
-				&color);
-
-	pipe_ctx->plane_res.xfm->funcs->transform_set_pixel_storage_depth(
-		pipe_ctx->plane_res.xfm,
-		pipe_ctx->plane_res.scl_data.lb_params.depth,
-		&pipe_ctx->stream->bit_depth_params);
-
-	if (pipe_ctx->stream_res.tg->funcs->set_overscan_blank_color) {
-		/*
-		 * The way 420 is packed, 2 channels carry Y component, 1 channel
-		 * alternate between Cb and Cr, so both channels need the pixel
-		 * value for Y
-		 */
-		if (pipe_ctx->stream->timing.pixel_encoding == PIXEL_ENCODING_YCBCR420)
-			color.color_r_cr = color.color_g_y;
-
-		pipe_ctx->stream_res.tg->funcs->set_overscan_blank_color(
-				pipe_ctx->stream_res.tg,
-				&color);
-	}
-
-	pipe_ctx->plane_res.xfm->funcs->transform_set_scaler(pipe_ctx->plane_res.xfm,
-		&pipe_ctx->plane_res.scl_data);
-}
-
-static enum dc_status dce110_enable_stream_timing(
-		struct pipe_ctx *pipe_ctx,
-		struct dc_state *context,
-		struct dc *dc)
-{
-	struct dc_stream_state *stream = pipe_ctx->stream;
-	struct pipe_ctx *pipe_ctx_old = &dc->current_state->res_ctx.
-			pipe_ctx[pipe_ctx->pipe_idx];
-	struct tg_color black_color = {0};
-
-	if (!pipe_ctx_old->stream) {
-
-		/* program blank color */
-		color_space_to_black_color(dc,
-				stream->output_color_space, &black_color);
-		pipe_ctx->stream_res.tg->funcs->set_blank_color(
-				pipe_ctx->stream_res.tg,
-				&black_color);
-
-		/*
-		 * Must blank CRTC after disabling power gating and before any
-		 * programming, otherwise CRTC will be hung in bad state
-		 */
-		pipe_ctx->stream_res.tg->funcs->set_blank(pipe_ctx->stream_res.tg, true);
-
-		if (false == pipe_ctx->clock_source->funcs->program_pix_clk(
-				pipe_ctx->clock_source,
-				&pipe_ctx->stream_res.pix_clk_params,
-				&pipe_ctx->pll_settings)) {
-			BREAK_TO_DEBUGGER();
-			return DC_ERROR_UNEXPECTED;
-		}
-
-		pipe_ctx->stream_res.tg->funcs->program_timing(
-				pipe_ctx->stream_res.tg,
-				&stream->timing,
-				0,
-				0,
-				0,
-				0,
-				pipe_ctx->stream->signal,
-				true);
-	}
-
-	if (!pipe_ctx_old->stream) {
-		if (false == pipe_ctx->stream_res.tg->funcs->enable_crtc(
-				pipe_ctx->stream_res.tg)) {
-			BREAK_TO_DEBUGGER();
-			return DC_ERROR_UNEXPECTED;
-		}
-	}
-
-	return DC_OK;
-}
-
-static enum dc_status apply_single_controller_ctx_to_hw(
-		struct pipe_ctx *pipe_ctx,
-		struct dc_state *context,
-		struct dc *dc)
-{
-	struct dc_stream_state *stream = pipe_ctx->stream;
-	struct dc_link *link = stream->link;
-	struct drr_params params = {0};
-	unsigned int event_triggers = 0;
-	struct pipe_ctx *odm_pipe = pipe_ctx->next_odm_pipe;
-	struct dce_hwseq *hws = dc->hwseq;
-
-	if (hws->funcs.disable_stream_gating) {
-		hws->funcs.disable_stream_gating(dc, pipe_ctx);
-	}
-
-	if (pipe_ctx->stream_res.audio != NULL) {
-		struct audio_output audio_output;
-
-		build_audio_output(context, pipe_ctx, &audio_output);
-
-		if (dc_is_dp_signal(pipe_ctx->stream->signal))
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-			if (is_dp_128b_132b_signal(pipe_ctx))
-				pipe_ctx->stream_res.hpo_dp_stream_enc->funcs->dp_audio_setup(
-						pipe_ctx->stream_res.hpo_dp_stream_enc,
-						pipe_ctx->stream_res.audio->inst,
-						&pipe_ctx->stream->audio_info);
-			else
-				pipe_ctx->stream_res.stream_enc->funcs->dp_audio_setup(
-						pipe_ctx->stream_res.stream_enc,
-						pipe_ctx->stream_res.audio->inst,
-						&pipe_ctx->stream->audio_info);
-#else
-			pipe_ctx->stream_res.stream_enc->funcs->dp_audio_setup(
-					pipe_ctx->stream_res.stream_enc,
-					pipe_ctx->stream_res.audio->inst,
-					&pipe_ctx->stream->audio_info);
-#endif
-		else
-			pipe_ctx->stream_res.stream_enc->funcs->hdmi_audio_setup(
-					pipe_ctx->stream_res.stream_enc,
-					pipe_ctx->stream_res.audio->inst,
-					&pipe_ctx->stream->audio_info,
-					&audio_output.crtc_info);
-
-		pipe_ctx->stream_res.audio->funcs->az_configure(
-				pipe_ctx->stream_res.audio,
-				pipe_ctx->stream->signal,
-				&audio_output.crtc_info,
-				&pipe_ctx->stream->audio_info);
-	}
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	/* DCN3.1 FPGA Workaround
-	 * Need to enable HPO DP Stream Encoder before setting OTG master enable.
-	 * To do so, move calling function enable_stream_timing to only be done AFTER calling
-	 * function core_link_enable_stream
-	 */
-	if (!(hws->wa.dp_hpo_and_otg_sequence && is_dp_128b_132b_signal(pipe_ctx)))
-#endif
-		/*  */
-		/* Do not touch stream timing on seamless boot optimization. */
-		if (!pipe_ctx->stream->apply_seamless_boot_optimization)
-			hws->funcs.enable_stream_timing(pipe_ctx, context, dc);
-
-	if (hws->funcs.setup_vupdate_interrupt)
-		hws->funcs.setup_vupdate_interrupt(dc, pipe_ctx);
-
-	params.vertical_total_min = stream->adjust.v_total_min;
-	params.vertical_total_max = stream->adjust.v_total_max;
-	if (pipe_ctx->stream_res.tg->funcs->set_drr)
-		pipe_ctx->stream_res.tg->funcs->set_drr(
-			pipe_ctx->stream_res.tg, &params);
-
-	// DRR should set trigger event to monitor surface update event
-	if (stream->adjust.v_total_min != 0 && stream->adjust.v_total_max != 0)
-		event_triggers = 0x80;
-	/* Event triggers and num frames initialized for DRR, but can be
-	 * later updated for PSR use. Note DRR trigger events are generated
-	 * regardless of whether num frames met.
-	 */
-	if (pipe_ctx->stream_res.tg->funcs->set_static_screen_control)
-		pipe_ctx->stream_res.tg->funcs->set_static_screen_control(
-				pipe_ctx->stream_res.tg, event_triggers, 2);
-
-	if (!dc_is_virtual_signal(pipe_ctx->stream->signal))
-		pipe_ctx->stream_res.stream_enc->funcs->dig_connect_to_otg(
-			pipe_ctx->stream_res.stream_enc,
-			pipe_ctx->stream_res.tg->inst);
-
-	if (dc_is_dp_signal(pipe_ctx->stream->signal))
-		dp_source_sequence_trace(link, DPCD_SOURCE_SEQ_AFTER_CONNECT_DIG_FE_OTG);
-
-	pipe_ctx->stream_res.opp->funcs->opp_set_dyn_expansion(
-			pipe_ctx->stream_res.opp,
-			COLOR_SPACE_YCBCR601,
-			stream->timing.display_color_depth,
-			stream->signal);
-
-	pipe_ctx->stream_res.opp->funcs->opp_program_fmt(
-		pipe_ctx->stream_res.opp,
-		&stream->bit_depth_params,
-		&stream->clamping);
-	while (odm_pipe) {
-		odm_pipe->stream_res.opp->funcs->opp_set_dyn_expansion(
-				odm_pipe->stream_res.opp,
-				COLOR_SPACE_YCBCR601,
-				stream->timing.display_color_depth,
-				stream->signal);
-
-		odm_pipe->stream_res.opp->funcs->opp_program_fmt(
-				odm_pipe->stream_res.opp,
-				&stream->bit_depth_params,
-				&stream->clamping);
-		odm_pipe = odm_pipe->next_odm_pipe;
-	}
-
-	if (!stream->dpms_off)
-		core_link_enable_stream(context, pipe_ctx);
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	/* DCN3.1 FPGA Workaround
-	 * Need to enable HPO DP Stream Encoder before setting OTG master enable.
-	 * To do so, move calling function enable_stream_timing to only be done AFTER calling
-	 * function core_link_enable_stream
-	 */
-	if (hws->wa.dp_hpo_and_otg_sequence && is_dp_128b_132b_signal(pipe_ctx)) {
-		if (!pipe_ctx->stream->apply_seamless_boot_optimization)
-			hws->funcs.enable_stream_timing(pipe_ctx, context, dc);
-	}
-#endif
-
-	pipe_ctx->plane_res.scl_data.lb_params.alpha_en = pipe_ctx->bottom_pipe != 0;
-
-	pipe_ctx->stream->link->psr_settings.psr_feature_enabled = false;
-
-	return DC_OK;
-}
-
-/******************************************************************************/
-
-static void power_down_encoders(struct dc *dc)
-{
-	int i, j;
-
-	for (i = 0; i < dc->link_count; i++) {
-		enum signal_type signal = dc->links[i]->connector_signal;
-
-		if ((signal == SIGNAL_TYPE_EDP) ||
-			(signal == SIGNAL_TYPE_DISPLAY_PORT)) {
-			if (dc->links[i]->link_enc->funcs->get_dig_frontend &&
-				dc->links[i]->link_enc->funcs->is_dig_enabled(dc->links[i]->link_enc)) {
-				unsigned int fe = dc->links[i]->link_enc->funcs->get_dig_frontend(
-									dc->links[i]->link_enc);
-
-				for (j = 0; j < dc->res_pool->stream_enc_count; j++) {
-					if (fe == dc->res_pool->stream_enc[j]->id) {
-						dc->res_pool->stream_enc[j]->funcs->dp_blank(dc->links[i],
-									dc->res_pool->stream_enc[j]);
-						break;
-					}
-				}
-			}
-
-			if (!dc->links[i]->wa_flags.dp_keep_receiver_powered)
-				dp_receiver_power_ctrl(dc->links[i], false);
-		}
-
-		if (signal != SIGNAL_TYPE_EDP)
-			signal = SIGNAL_TYPE_NONE;
-
-		if (dc->links[i]->ep_type == DISPLAY_ENDPOINT_PHY)
-			dc->links[i]->link_enc->funcs->disable_output(
-					dc->links[i]->link_enc, signal);
-
-		dc->links[i]->link_status.link_active = false;
-		memset(&dc->links[i]->cur_link_settings, 0,
-				sizeof(dc->links[i]->cur_link_settings));
-	}
-}
-
-static void power_down_controllers(struct dc *dc)
-{
-	int i;
-
-	for (i = 0; i < dc->res_pool->timing_generator_count; i++) {
-		dc->res_pool->timing_generators[i]->funcs->disable_crtc(
-				dc->res_pool->timing_generators[i]);
-	}
-}
-
-static void power_down_clock_sources(struct dc *dc)
-{
-	int i;
-
-	if (dc->res_pool->dp_clock_source->funcs->cs_power_down(
-		dc->res_pool->dp_clock_source) == false)
-		dm_error("Failed to power down pll! (dp clk src)\n");
-
-	for (i = 0; i < dc->res_pool->clk_src_count; i++) {
-		if (dc->res_pool->clock_sources[i]->funcs->cs_power_down(
-				dc->res_pool->clock_sources[i]) == false)
-			dm_error("Failed to power down pll! (clk src index=%d)\n", i);
-	}
-}
-
-static void power_down_all_hw_blocks(struct dc *dc)
-{
-	power_down_encoders(dc);
-
-	power_down_controllers(dc);
-
-	power_down_clock_sources(dc);
-
-	if (dc->fbc_compressor)
-		dc->fbc_compressor->funcs->disable_fbc(dc->fbc_compressor);
-}
-
-static void disable_vga_and_power_gate_all_controllers(
-		struct dc *dc)
-{
-	int i;
-	struct timing_generator *tg;
-	struct dc_context *ctx = dc->ctx;
-
-	for (i = 0; i < dc->res_pool->timing_generator_count; i++) {
-		tg = dc->res_pool->timing_generators[i];
-
-		if (tg->funcs->disable_vga)
-			tg->funcs->disable_vga(tg);
-	}
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		/* Enable CLOCK gating for each pipe BEFORE controller
-		 * powergating. */
-		enable_display_pipe_clock_gating(ctx,
-				true);
-
-		dc->current_state->res_ctx.pipe_ctx[i].pipe_idx = i;
-		dc->hwss.disable_plane(dc,
-			&dc->current_state->res_ctx.pipe_ctx[i]);
-	}
-}
-
-
-static void get_edp_streams(struct dc_state *context,
-		struct dc_stream_state **edp_streams,
-		int *edp_stream_num)
-{
-	int i;
-
-	*edp_stream_num = 0;
-	for (i = 0; i < context->stream_count; i++) {
-		if (context->streams[i]->signal == SIGNAL_TYPE_EDP) {
-			edp_streams[*edp_stream_num] = context->streams[i];
-			if (++(*edp_stream_num) == MAX_NUM_EDP)
-				return;
-		}
-	}
-}
-
-static void get_edp_links_with_sink(
-		struct dc *dc,
-		struct dc_link **edp_links_with_sink,
-		int *edp_with_sink_num)
-{
-	int i;
-
-	/* check if there is an eDP panel not in use */
-	*edp_with_sink_num = 0;
-	for (i = 0; i < dc->link_count; i++) {
-		if (dc->links[i]->local_sink &&
-			dc->links[i]->local_sink->sink_signal == SIGNAL_TYPE_EDP) {
-			edp_links_with_sink[*edp_with_sink_num] = dc->links[i];
-			if (++(*edp_with_sink_num) == MAX_NUM_EDP)
-				return;
-		}
-	}
-}
-
-/*
- * When ASIC goes from VBIOS/VGA mode to driver/accelerated mode we need:
- *  1. Power down all DC HW blocks
- *  2. Disable VGA engine on all controllers
- *  3. Enable power gating for controller
- *  4. Set acc_mode_change bit (VBIOS will clear this bit when going to FSDOS)
- */
-void dce110_enable_accelerated_mode(struct dc *dc, struct dc_state *context)
-{
-	struct dc_link *edp_links_with_sink[MAX_NUM_EDP];
-	struct dc_link *edp_links[MAX_NUM_EDP];
-	struct dc_stream_state *edp_streams[MAX_NUM_EDP];
-	struct dc_link *edp_link_with_sink = NULL;
-	struct dc_link *edp_link = NULL;
-	struct dc_stream_state *edp_stream = NULL;
-	struct dce_hwseq *hws = dc->hwseq;
-	int edp_with_sink_num;
-	int edp_num;
-	int edp_stream_num;
-	int i;
-	bool can_apply_edp_fast_boot = false;
-	bool can_apply_seamless_boot = false;
-	bool keep_edp_vdd_on = false;
-	DC_LOGGER_INIT();
-
-
-	get_edp_links_with_sink(dc, edp_links_with_sink, &edp_with_sink_num);
-	get_edp_links(dc, edp_links, &edp_num);
-
-	if (hws->funcs.init_pipes)
-		hws->funcs.init_pipes(dc, context);
-
-	get_edp_streams(context, edp_streams, &edp_stream_num);
-
-	// Check fastboot support, disable on DCE8 because of blank screens
-	if (edp_num && dc->ctx->dce_version != DCE_VERSION_8_0 &&
-		    dc->ctx->dce_version != DCE_VERSION_8_1 &&
-		    dc->ctx->dce_version != DCE_VERSION_8_3) {
-		for (i = 0; i < edp_num; i++) {
-			edp_link = edp_links[i];
-			// enable fastboot if backend is enabled on eDP
-			if (edp_link->link_enc->funcs->is_dig_enabled(edp_link->link_enc)) {
-				/* Set optimization flag on eDP stream*/
-				if (edp_stream_num && edp_link->link_status.link_active) {
-					edp_stream = edp_streams[0];
-					can_apply_edp_fast_boot = !is_edp_ilr_optimization_required(edp_stream->link, &edp_stream->timing);
-					edp_stream->apply_edp_fast_boot_optimization = can_apply_edp_fast_boot;
-					if (can_apply_edp_fast_boot)
-						DC_LOG_EVENT_LINK_TRAINING("eDP fast boot disabled to optimize link rate\n");
-
-					break;
-				}
-			}
-		}
-		// We are trying to enable eDP, don't power down VDD
-		if (edp_stream_num)
-			keep_edp_vdd_on = true;
-	}
-
-	// Check seamless boot support
-	for (i = 0; i < context->stream_count; i++) {
-		if (context->streams[i]->apply_seamless_boot_optimization) {
-			can_apply_seamless_boot = true;
-			break;
-		}
-	}
-
-	/* eDP should not have stream in resume from S4 and so even with VBios post
-	 * it should get turned off
-	 */
-	if (edp_with_sink_num)
-		edp_link_with_sink = edp_links_with_sink[0];
-
-	if (!can_apply_edp_fast_boot && !can_apply_seamless_boot) {
-		if (edp_link_with_sink && !keep_edp_vdd_on) {
-			/*turn off backlight before DP_blank and encoder powered down*/
-			hws->funcs.edp_backlight_control(edp_link_with_sink, false);
-		}
-		/*resume from S3, no vbios posting, no need to power down again*/
-		power_down_all_hw_blocks(dc);
-		disable_vga_and_power_gate_all_controllers(dc);
-		if (edp_link_with_sink && !keep_edp_vdd_on)
-			dc->hwss.edp_power_control(edp_link_with_sink, false);
-	}
-	bios_set_scratch_acc_mode_change(dc->ctx->dc_bios, 1);
-}
-
-static uint32_t compute_pstate_blackout_duration(
-	struct bw_fixed blackout_duration,
-	const struct dc_stream_state *stream)
-{
-	uint32_t total_dest_line_time_ns;
-	uint32_t pstate_blackout_duration_ns;
-
-	pstate_blackout_duration_ns = 1000 * blackout_duration.value >> 24;
-
-	total_dest_line_time_ns = 1000000UL *
-		(stream->timing.h_total * 10) /
-		stream->timing.pix_clk_100hz +
-		pstate_blackout_duration_ns;
-
-	return total_dest_line_time_ns;
-}
-
-static void dce110_set_displaymarks(
-	const struct dc *dc,
-	struct dc_state *context)
-{
-	uint8_t i, num_pipes;
-	unsigned int underlay_idx = dc->res_pool->underlay_pipe_index;
-
-	for (i = 0, num_pipes = 0; i < MAX_PIPES; i++) {
-		struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i];
-		uint32_t total_dest_line_time_ns;
-
-		if (pipe_ctx->stream == NULL)
-			continue;
-
-		total_dest_line_time_ns = compute_pstate_blackout_duration(
-			dc->bw_vbios->blackout_duration, pipe_ctx->stream);
-		pipe_ctx->plane_res.mi->funcs->mem_input_program_display_marks(
-			pipe_ctx->plane_res.mi,
-			context->bw_ctx.bw.dce.nbp_state_change_wm_ns[num_pipes],
-			context->bw_ctx.bw.dce.stutter_exit_wm_ns[num_pipes],
-			context->bw_ctx.bw.dce.stutter_entry_wm_ns[num_pipes],
-			context->bw_ctx.bw.dce.urgent_wm_ns[num_pipes],
-			total_dest_line_time_ns);
-		if (i == underlay_idx) {
-			num_pipes++;
-			pipe_ctx->plane_res.mi->funcs->mem_input_program_chroma_display_marks(
-				pipe_ctx->plane_res.mi,
-				context->bw_ctx.bw.dce.nbp_state_change_wm_ns[num_pipes],
-				context->bw_ctx.bw.dce.stutter_exit_wm_ns[num_pipes],
-				context->bw_ctx.bw.dce.urgent_wm_ns[num_pipes],
-				total_dest_line_time_ns);
-		}
-		num_pipes++;
-	}
-}
-
-void dce110_set_safe_displaymarks(
-		struct resource_context *res_ctx,
-		const struct resource_pool *pool)
-{
-	int i;
-	int underlay_idx = pool->underlay_pipe_index;
-	struct dce_watermarks max_marks = {
-		MAX_WATERMARK, MAX_WATERMARK, MAX_WATERMARK, MAX_WATERMARK };
-	struct dce_watermarks nbp_marks = {
-		SAFE_NBP_MARK, SAFE_NBP_MARK, SAFE_NBP_MARK, SAFE_NBP_MARK };
-	struct dce_watermarks min_marks = { 0, 0, 0, 0};
-
-	for (i = 0; i < MAX_PIPES; i++) {
-		if (res_ctx->pipe_ctx[i].stream == NULL || res_ctx->pipe_ctx[i].plane_res.mi == NULL)
-			continue;
-
-		res_ctx->pipe_ctx[i].plane_res.mi->funcs->mem_input_program_display_marks(
-				res_ctx->pipe_ctx[i].plane_res.mi,
-				nbp_marks,
-				max_marks,
-				min_marks,
-				max_marks,
-				MAX_WATERMARK);
-
-		if (i == underlay_idx)
-			res_ctx->pipe_ctx[i].plane_res.mi->funcs->mem_input_program_chroma_display_marks(
-				res_ctx->pipe_ctx[i].plane_res.mi,
-				nbp_marks,
-				max_marks,
-				max_marks,
-				MAX_WATERMARK);
-
-	}
-}
-
-/*******************************************************************************
- * Public functions
- ******************************************************************************/
-
-static void set_drr(struct pipe_ctx **pipe_ctx,
-		int num_pipes, struct dc_crtc_timing_adjust adjust)
-{
-	int i = 0;
-	struct drr_params params = {0};
-	// DRR should set trigger event to monitor surface update event
-	unsigned int event_triggers = 0x80;
-	// Note DRR trigger events are generated regardless of whether num frames met.
-	unsigned int num_frames = 2;
-
-	params.vertical_total_max = adjust.v_total_max;
-	params.vertical_total_min = adjust.v_total_min;
-
-	/* TODO: If multiple pipes are to be supported, you need
-	 * some GSL stuff. Static screen triggers may be programmed differently
-	 * as well.
-	 */
-	for (i = 0; i < num_pipes; i++) {
-		pipe_ctx[i]->stream_res.tg->funcs->set_drr(
-			pipe_ctx[i]->stream_res.tg, &params);
-
-		if (adjust.v_total_max != 0 && adjust.v_total_min != 0)
-			pipe_ctx[i]->stream_res.tg->funcs->set_static_screen_control(
-					pipe_ctx[i]->stream_res.tg,
-					event_triggers, num_frames);
-	}
-}
-
-static void get_position(struct pipe_ctx **pipe_ctx,
-		int num_pipes,
-		struct crtc_position *position)
-{
-	int i = 0;
-
-	/* TODO: handle pipes > 1
-	 */
-	for (i = 0; i < num_pipes; i++)
-		pipe_ctx[i]->stream_res.tg->funcs->get_position(pipe_ctx[i]->stream_res.tg, position);
-}
-
-static void set_static_screen_control(struct pipe_ctx **pipe_ctx,
-		int num_pipes, const struct dc_static_screen_params *params)
-{
-	unsigned int i;
-	unsigned int triggers = 0;
-
-	if (params->triggers.overlay_update)
-		triggers |= 0x100;
-	if (params->triggers.surface_update)
-		triggers |= 0x80;
-	if (params->triggers.cursor_update)
-		triggers |= 0x2;
-	if (params->triggers.force_trigger)
-		triggers |= 0x1;
-
-	if (num_pipes) {
-		struct dc *dc = pipe_ctx[0]->stream->ctx->dc;
-
-		if (dc->fbc_compressor)
-			triggers |= 0x84;
-	}
-
-	for (i = 0; i < num_pipes; i++)
-		pipe_ctx[i]->stream_res.tg->funcs->
-			set_static_screen_control(pipe_ctx[i]->stream_res.tg,
-					triggers, params->num_frames);
-}
-
-/*
- *  Check if FBC can be enabled
- */
-static bool should_enable_fbc(struct dc *dc,
-		struct dc_state *context,
-		uint32_t *pipe_idx)
-{
-	uint32_t i;
-	struct pipe_ctx *pipe_ctx = NULL;
-	struct resource_context *res_ctx = &context->res_ctx;
-	unsigned int underlay_idx = dc->res_pool->underlay_pipe_index;
-
-
-	ASSERT(dc->fbc_compressor);
-
-	/* FBC memory should be allocated */
-	if (!dc->ctx->fbc_gpu_addr)
-		return false;
-
-	/* Only supports single display */
-	if (context->stream_count != 1)
-		return false;
-
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		if (res_ctx->pipe_ctx[i].stream) {
-
-			pipe_ctx = &res_ctx->pipe_ctx[i];
-
-			if (!pipe_ctx)
-				continue;
-
-			/* fbc not applicable on underlay pipe */
-			if (pipe_ctx->pipe_idx != underlay_idx) {
-				*pipe_idx = i;
-				break;
-			}
-		}
-	}
-
-	if (i == dc->res_pool->pipe_count)
-		return false;
-
-	if (!pipe_ctx->stream->link)
-		return false;
-
-	/* Only supports eDP */
-	if (pipe_ctx->stream->link->connector_signal != SIGNAL_TYPE_EDP)
-		return false;
-
-	/* PSR should not be enabled */
-	if (pipe_ctx->stream->link->psr_settings.psr_feature_enabled)
-		return false;
-
-	/* Nothing to compress */
-	if (!pipe_ctx->plane_state)
-		return false;
-
-	/* Only for non-linear tiling */
-	if (pipe_ctx->plane_state->tiling_info.gfx8.array_mode == DC_ARRAY_LINEAR_GENERAL)
-		return false;
-
-	return true;
-}
-
-/*
- *  Enable FBC
- */
-static void enable_fbc(
-		struct dc *dc,
-		struct dc_state *context)
-{
-	uint32_t pipe_idx = 0;
-
-	if (should_enable_fbc(dc, context, &pipe_idx)) {
-		/* Program GRPH COMPRESSED ADDRESS and PITCH */
-		struct compr_addr_and_pitch_params params = {0, 0, 0};
-		struct compressor *compr = dc->fbc_compressor;
-		struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[pipe_idx];
-
-		params.source_view_width = pipe_ctx->stream->timing.h_addressable;
-		params.source_view_height = pipe_ctx->stream->timing.v_addressable;
-		params.inst = pipe_ctx->stream_res.tg->inst;
-		compr->compr_surface_address.quad_part = dc->ctx->fbc_gpu_addr;
-
-		compr->funcs->surface_address_and_pitch(compr, &params);
-		compr->funcs->set_fbc_invalidation_triggers(compr, 1);
-
-		compr->funcs->enable_fbc(compr, &params);
-	}
-}
-
-static void dce110_reset_hw_ctx_wrap(
-		struct dc *dc,
-		struct dc_state *context)
-{
-	int i;
-
-	/* Reset old context */
-	/* look up the targets that have been removed since last commit */
-	for (i = 0; i < MAX_PIPES; i++) {
-		struct pipe_ctx *pipe_ctx_old =
-			&dc->current_state->res_ctx.pipe_ctx[i];
-		struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i];
-
-		/* Note: We need to disable output if clock sources change,
-		 * since bios does optimization and doesn't apply if changing
-		 * PHY when not already disabled.
-		 */
-
-		/* Skip underlay pipe since it will be handled in commit surface*/
-		if (!pipe_ctx_old->stream || pipe_ctx_old->top_pipe)
-			continue;
-
-		if (!pipe_ctx->stream ||
-				pipe_need_reprogram(pipe_ctx_old, pipe_ctx)) {
-			struct clock_source *old_clk = pipe_ctx_old->clock_source;
-
-			/* Disable if new stream is null. O/w, if stream is
-			 * disabled already, no need to disable again.
-			 */
-			if (!pipe_ctx->stream || !pipe_ctx->stream->dpms_off) {
-				core_link_disable_stream(pipe_ctx_old);
-
-				/* free acquired resources*/
-				if (pipe_ctx_old->stream_res.audio) {
-					/*disable az_endpoint*/
-					pipe_ctx_old->stream_res.audio->funcs->
-							az_disable(pipe_ctx_old->stream_res.audio);
-
-					/*free audio*/
-					if (dc->caps.dynamic_audio == true) {
-						/*we have to dynamic arbitrate the audio endpoints*/
-						/*we free the resource, need reset is_audio_acquired*/
-						update_audio_usage(&dc->current_state->res_ctx, dc->res_pool,
-								pipe_ctx_old->stream_res.audio, false);
-						pipe_ctx_old->stream_res.audio = NULL;
-					}
-				}
-			}
-
-			pipe_ctx_old->stream_res.tg->funcs->set_blank(pipe_ctx_old->stream_res.tg, true);
-			if (!hwss_wait_for_blank_complete(pipe_ctx_old->stream_res.tg)) {
-				dm_error("DC: failed to blank crtc!\n");
-				BREAK_TO_DEBUGGER();
-			}
-			pipe_ctx_old->stream_res.tg->funcs->disable_crtc(pipe_ctx_old->stream_res.tg);
-			pipe_ctx_old->plane_res.mi->funcs->free_mem_input(
-					pipe_ctx_old->plane_res.mi, dc->current_state->stream_count);
-
-			if (old_clk && 0 == resource_get_clock_source_reference(&context->res_ctx,
-										dc->res_pool,
-										old_clk))
-				old_clk->funcs->cs_power_down(old_clk);
-
-			dc->hwss.disable_plane(dc, pipe_ctx_old);
-
-			pipe_ctx_old->stream = NULL;
-		}
-	}
-}
-
-static void dce110_setup_audio_dto(
-		struct dc *dc,
-		struct dc_state *context)
-{
-	int i;
-
-	/* program audio wall clock. use HDMI as clock source if HDMI
-	 * audio active. Otherwise, use DP as clock source
-	 * first, loop to find any HDMI audio, if not, loop find DP audio
-	 */
-	/* Setup audio rate clock source */
-	/* Issue:
-	* Audio lag happened on DP monitor when unplug a HDMI monitor
-	*
-	* Cause:
-	* In case of DP and HDMI connected or HDMI only, DCCG_AUDIO_DTO_SEL
-	* is set to either dto0 or dto1, audio should work fine.
-	* In case of DP connected only, DCCG_AUDIO_DTO_SEL should be dto1,
-	* set to dto0 will cause audio lag.
-	*
-	* Solution:
-	* Not optimized audio wall dto setup. When mode set, iterate pipe_ctx,
-	* find first available pipe with audio, setup audio wall DTO per topology
-	* instead of per pipe.
-	*/
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i];
-
-		if (pipe_ctx->stream == NULL)
-			continue;
-
-		if (pipe_ctx->top_pipe)
-			continue;
-		if (pipe_ctx->stream->signal != SIGNAL_TYPE_HDMI_TYPE_A)
-			continue;
-		if (pipe_ctx->stream_res.audio != NULL) {
-			struct audio_output audio_output;
-
-			build_audio_output(context, pipe_ctx, &audio_output);
-
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-			/* For DCN3.1, audio to HPO FRL encoder is using audio DTBCLK DTO */
-			if (dc->res_pool->dccg && dc->res_pool->dccg->funcs->set_audio_dtbclk_dto) {
-				/* disable audio DTBCLK DTO */
-				dc->res_pool->dccg->funcs->set_audio_dtbclk_dto(
-					dc->res_pool->dccg, 0);
-
-				pipe_ctx->stream_res.audio->funcs->wall_dto_setup(
-						pipe_ctx->stream_res.audio,
-						pipe_ctx->stream->signal,
-						&audio_output.crtc_info,
-						&audio_output.pll_info);
-			} else
-				pipe_ctx->stream_res.audio->funcs->wall_dto_setup(
-					pipe_ctx->stream_res.audio,
-					pipe_ctx->stream->signal,
-					&audio_output.crtc_info,
-					&audio_output.pll_info);
-#else
-			pipe_ctx->stream_res.audio->funcs->wall_dto_setup(
-				pipe_ctx->stream_res.audio,
-				pipe_ctx->stream->signal,
-				&audio_output.crtc_info,
-				&audio_output.pll_info);
-#endif
-			break;
-		}
-	}
-
-	/* no HDMI audio is found, try DP audio */
-	if (i == dc->res_pool->pipe_count) {
-		for (i = 0; i < dc->res_pool->pipe_count; i++) {
-			struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i];
-
-			if (pipe_ctx->stream == NULL)
-				continue;
-
-			if (pipe_ctx->top_pipe)
-				continue;
-
-			if (!dc_is_dp_signal(pipe_ctx->stream->signal))
-				continue;
-
-			if (pipe_ctx->stream_res.audio != NULL) {
-				struct audio_output audio_output;
-
-				build_audio_output(context, pipe_ctx, &audio_output);
-
-				pipe_ctx->stream_res.audio->funcs->wall_dto_setup(
-					pipe_ctx->stream_res.audio,
-					pipe_ctx->stream->signal,
-					&audio_output.crtc_info,
-					&audio_output.pll_info);
-				break;
-			}
-		}
-	}
-}
-
-enum dc_status dce110_apply_ctx_to_hw(
-		struct dc *dc,
-		struct dc_state *context)
-{
-	struct dce_hwseq *hws = dc->hwseq;
-	struct dc_bios *dcb = dc->ctx->dc_bios;
-	enum dc_status status;
-	int i;
-
-	/* Reset old context */
-	/* look up the targets that have been removed since last commit */
-	hws->funcs.reset_hw_ctx_wrap(dc, context);
-
-	/* Skip applying if no targets */
-	if (context->stream_count <= 0)
-		return DC_OK;
-
-	/* Apply new context */
-	dcb->funcs->set_scratch_critical_state(dcb, true);
-
-	/* below is for real asic only */
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		struct pipe_ctx *pipe_ctx_old =
-					&dc->current_state->res_ctx.pipe_ctx[i];
-		struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i];
-
-		if (pipe_ctx->stream == NULL || pipe_ctx->top_pipe)
-			continue;
-
-		if (pipe_ctx->stream == pipe_ctx_old->stream) {
-			if (pipe_ctx_old->clock_source != pipe_ctx->clock_source)
-				dce_crtc_switch_to_clk_src(dc->hwseq,
-						pipe_ctx->clock_source, i);
-			continue;
-		}
-
-		hws->funcs.enable_display_power_gating(
-				dc, i, dc->ctx->dc_bios,
-				PIPE_GATING_CONTROL_DISABLE);
-	}
-
-	if (dc->fbc_compressor)
-		dc->fbc_compressor->funcs->disable_fbc(dc->fbc_compressor);
-
-	dce110_setup_audio_dto(dc, context);
-
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		struct pipe_ctx *pipe_ctx_old =
-					&dc->current_state->res_ctx.pipe_ctx[i];
-		struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i];
-
-		if (pipe_ctx->stream == NULL)
-			continue;
-
-		if (pipe_ctx->stream == pipe_ctx_old->stream &&
-			pipe_ctx->stream->link->link_state_valid) {
-			continue;
-		}
-
-		if (pipe_ctx_old->stream && !pipe_need_reprogram(pipe_ctx_old, pipe_ctx))
-			continue;
-
-		if (pipe_ctx->top_pipe || pipe_ctx->prev_odm_pipe)
-			continue;
-
-		status = apply_single_controller_ctx_to_hw(
-				pipe_ctx,
-				context,
-				dc);
-
-		if (DC_OK != status)
-			return status;
-	}
-
-	if (dc->fbc_compressor)
-		enable_fbc(dc, dc->current_state);
-
-	dcb->funcs->set_scratch_critical_state(dcb, false);
-
-	return DC_OK;
-}
-
-/*******************************************************************************
- * Front End programming
- ******************************************************************************/
-static void set_default_colors(struct pipe_ctx *pipe_ctx)
-{
-	struct default_adjustment default_adjust = { 0 };
-
-	default_adjust.force_hw_default = false;
-	default_adjust.in_color_space = pipe_ctx->plane_state->color_space;
-	default_adjust.out_color_space = pipe_ctx->stream->output_color_space;
-	default_adjust.csc_adjust_type = GRAPHICS_CSC_ADJUST_TYPE_SW;
-	default_adjust.surface_pixel_format = pipe_ctx->plane_res.scl_data.format;
-
-	/* display color depth */
-	default_adjust.color_depth =
-		pipe_ctx->stream->timing.display_color_depth;
-
-	/* Lb color depth */
-	default_adjust.lb_color_depth = pipe_ctx->plane_res.scl_data.lb_params.depth;
-
-	pipe_ctx->plane_res.xfm->funcs->opp_set_csc_default(
-					pipe_ctx->plane_res.xfm, &default_adjust);
-}
-
-
-/*******************************************************************************
- * In order to turn on/off specific surface we will program
- * Blender + CRTC
- *
- * In case that we have two surfaces and they have a different visibility
- * we can't turn off the CRTC since it will turn off the entire display
- *
- * |----------------------------------------------- |
- * |bottom pipe|curr pipe  |              |         |
- * |Surface    |Surface    | Blender      |  CRCT   |
- * |visibility |visibility | Configuration|         |
- * |------------------------------------------------|
- * |   off     |    off    | CURRENT_PIPE | blank   |
- * |   off     |    on     | CURRENT_PIPE | unblank |
- * |   on      |    off    | OTHER_PIPE   | unblank |
- * |   on      |    on     | BLENDING     | unblank |
- * -------------------------------------------------|
- *
- ******************************************************************************/
-static void program_surface_visibility(const struct dc *dc,
-		struct pipe_ctx *pipe_ctx)
-{
-	enum blnd_mode blender_mode = BLND_MODE_CURRENT_PIPE;
-	bool blank_target = false;
-
-	if (pipe_ctx->bottom_pipe) {
-
-		/* For now we are supporting only two pipes */
-		ASSERT(pipe_ctx->bottom_pipe->bottom_pipe == NULL);
-
-		if (pipe_ctx->bottom_pipe->plane_state->visible) {
-			if (pipe_ctx->plane_state->visible)
-				blender_mode = BLND_MODE_BLENDING;
-			else
-				blender_mode = BLND_MODE_OTHER_PIPE;
-
-		} else if (!pipe_ctx->plane_state->visible)
-			blank_target = true;
-
-	} else if (!pipe_ctx->plane_state->visible)
-		blank_target = true;
-
-	dce_set_blender_mode(dc->hwseq, pipe_ctx->stream_res.tg->inst, blender_mode);
-	pipe_ctx->stream_res.tg->funcs->set_blank(pipe_ctx->stream_res.tg, blank_target);
-
-}
-
-static void program_gamut_remap(struct pipe_ctx *pipe_ctx)
-{
-	int i = 0;
-	struct xfm_grph_csc_adjustment adjust;
-	memset(&adjust, 0, sizeof(adjust));
-	adjust.gamut_adjust_type = GRAPHICS_GAMUT_ADJUST_TYPE_BYPASS;
-
-
-	if (pipe_ctx->stream->gamut_remap_matrix.enable_remap == true) {
-		adjust.gamut_adjust_type = GRAPHICS_GAMUT_ADJUST_TYPE_SW;
-
-		for (i = 0; i < CSC_TEMPERATURE_MATRIX_SIZE; i++)
-			adjust.temperature_matrix[i] =
-				pipe_ctx->stream->gamut_remap_matrix.matrix[i];
-	}
-
-	pipe_ctx->plane_res.xfm->funcs->transform_set_gamut_remap(pipe_ctx->plane_res.xfm, &adjust);
-}
-static void update_plane_addr(const struct dc *dc,
-		struct pipe_ctx *pipe_ctx)
-{
-	struct dc_plane_state *plane_state = pipe_ctx->plane_state;
-
-	if (plane_state == NULL)
-		return;
-
-	pipe_ctx->plane_res.mi->funcs->mem_input_program_surface_flip_and_addr(
-			pipe_ctx->plane_res.mi,
-			&plane_state->address,
-			plane_state->flip_immediate);
-
-	plane_state->status.requested_address = plane_state->address;
-}
-
-static void dce110_update_pending_status(struct pipe_ctx *pipe_ctx)
-{
-	struct dc_plane_state *plane_state = pipe_ctx->plane_state;
-
-	if (plane_state == NULL)
-		return;
-
-	plane_state->status.is_flip_pending =
-			pipe_ctx->plane_res.mi->funcs->mem_input_is_flip_pending(
-					pipe_ctx->plane_res.mi);
-
-	if (plane_state->status.is_flip_pending && !plane_state->visible)
-		pipe_ctx->plane_res.mi->current_address = pipe_ctx->plane_res.mi->request_address;
-
-	plane_state->status.current_address = pipe_ctx->plane_res.mi->current_address;
-	if (pipe_ctx->plane_res.mi->current_address.type == PLN_ADDR_TYPE_GRPH_STEREO &&
-			pipe_ctx->stream_res.tg->funcs->is_stereo_left_eye) {
-		plane_state->status.is_right_eye =\
-				!pipe_ctx->stream_res.tg->funcs->is_stereo_left_eye(pipe_ctx->stream_res.tg);
-	}
-}
-
-void dce110_power_down(struct dc *dc)
-{
-	power_down_all_hw_blocks(dc);
-	disable_vga_and_power_gate_all_controllers(dc);
-}
-
-static bool wait_for_reset_trigger_to_occur(
-	struct dc_context *dc_ctx,
-	struct timing_generator *tg)
-{
-	bool rc = false;
-
-	/* To avoid endless loop we wait at most
-	 * frames_to_wait_on_triggered_reset frames for the reset to occur. */
-	const uint32_t frames_to_wait_on_triggered_reset = 10;
-	uint32_t i;
-
-	for (i = 0; i < frames_to_wait_on_triggered_reset; i++) {
-
-		if (!tg->funcs->is_counter_moving(tg)) {
-			DC_ERROR("TG counter is not moving!\n");
-			break;
-		}
-
-		if (tg->funcs->did_triggered_reset_occur(tg)) {
-			rc = true;
-			/* usually occurs at i=1 */
-			DC_SYNC_INFO("GSL: reset occurred at wait count: %d\n",
-					i);
-			break;
-		}
-
-		/* Wait for one frame. */
-		tg->funcs->wait_for_state(tg, CRTC_STATE_VACTIVE);
-		tg->funcs->wait_for_state(tg, CRTC_STATE_VBLANK);
-	}
-
-	if (false == rc)
-		DC_ERROR("GSL: Timeout on reset trigger!\n");
-
-	return rc;
-}
-
-/* Enable timing synchronization for a group of Timing Generators. */
-static void dce110_enable_timing_synchronization(
-		struct dc *dc,
-		int group_index,
-		int group_size,
-		struct pipe_ctx *grouped_pipes[])
-{
-	struct dc_context *dc_ctx = dc->ctx;
-	struct dcp_gsl_params gsl_params = { 0 };
-	int i;
-
-	DC_SYNC_INFO("GSL: Setting-up...\n");
-
-	/* Designate a single TG in the group as a master.
-	 * Since HW doesn't care which one, we always assign
-	 * the 1st one in the group. */
-	gsl_params.gsl_group = 0;
-	gsl_params.gsl_master = grouped_pipes[0]->stream_res.tg->inst;
-
-	for (i = 0; i < group_size; i++)
-		grouped_pipes[i]->stream_res.tg->funcs->setup_global_swap_lock(
-					grouped_pipes[i]->stream_res.tg, &gsl_params);
-
-	/* Reset slave controllers on master VSync */
-	DC_SYNC_INFO("GSL: enabling trigger-reset\n");
-
-	for (i = 1 /* skip the master */; i < group_size; i++)
-		grouped_pipes[i]->stream_res.tg->funcs->enable_reset_trigger(
-				grouped_pipes[i]->stream_res.tg,
-				gsl_params.gsl_group);
-
-	for (i = 1 /* skip the master */; i < group_size; i++) {
-		DC_SYNC_INFO("GSL: waiting for reset to occur.\n");
-		wait_for_reset_trigger_to_occur(dc_ctx, grouped_pipes[i]->stream_res.tg);
-		grouped_pipes[i]->stream_res.tg->funcs->disable_reset_trigger(
-				grouped_pipes[i]->stream_res.tg);
-	}
-
-	/* GSL Vblank synchronization is a one time sync mechanism, assumption
-	 * is that the sync'ed displays will not drift out of sync over time*/
-	DC_SYNC_INFO("GSL: Restoring register states.\n");
-	for (i = 0; i < group_size; i++)
-		grouped_pipes[i]->stream_res.tg->funcs->tear_down_global_swap_lock(grouped_pipes[i]->stream_res.tg);
-
-	DC_SYNC_INFO("GSL: Set-up complete.\n");
-}
-
-static void dce110_enable_per_frame_crtc_position_reset(
-		struct dc *dc,
-		int group_size,
-		struct pipe_ctx *grouped_pipes[])
-{
-	struct dc_context *dc_ctx = dc->ctx;
-	struct dcp_gsl_params gsl_params = { 0 };
-	int i;
-
-	gsl_params.gsl_group = 0;
-	gsl_params.gsl_master = 0;
-
-	for (i = 0; i < group_size; i++)
-		grouped_pipes[i]->stream_res.tg->funcs->setup_global_swap_lock(
-					grouped_pipes[i]->stream_res.tg, &gsl_params);
-
-	DC_SYNC_INFO("GSL: enabling trigger-reset\n");
-
-	for (i = 1; i < group_size; i++)
-		grouped_pipes[i]->stream_res.tg->funcs->enable_crtc_reset(
-				grouped_pipes[i]->stream_res.tg,
-				gsl_params.gsl_master,
-				&grouped_pipes[i]->stream->triggered_crtc_reset);
-
-	DC_SYNC_INFO("GSL: waiting for reset to occur.\n");
-	for (i = 1; i < group_size; i++)
-		wait_for_reset_trigger_to_occur(dc_ctx, grouped_pipes[i]->stream_res.tg);
-
-	for (i = 0; i < group_size; i++)
-		grouped_pipes[i]->stream_res.tg->funcs->tear_down_global_swap_lock(grouped_pipes[i]->stream_res.tg);
-
-}
-
-static void init_pipes(struct dc *dc, struct dc_state *context)
-{
-	// Do nothing
-}
-
-static void init_hw(struct dc *dc)
-{
-	int i;
-	struct dc_bios *bp;
-	struct transform *xfm;
-	struct abm *abm;
-	struct dmcu *dmcu;
-	struct dce_hwseq *hws = dc->hwseq;
-	uint32_t backlight = MAX_BACKLIGHT_LEVEL;
-
-	bp = dc->ctx->dc_bios;
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		xfm = dc->res_pool->transforms[i];
-		xfm->funcs->transform_reset(xfm);
-
-		hws->funcs.enable_display_power_gating(
-				dc, i, bp,
-				PIPE_GATING_CONTROL_INIT);
-		hws->funcs.enable_display_power_gating(
-				dc, i, bp,
-				PIPE_GATING_CONTROL_DISABLE);
-		hws->funcs.enable_display_pipe_clock_gating(
-			dc->ctx,
-			true);
-	}
-
-	dce_clock_gating_power_up(dc->hwseq, false);
-	/***************************************/
-
-	for (i = 0; i < dc->link_count; i++) {
-		/****************************************/
-		/* Power up AND update implementation according to the
-		 * required signal (which may be different from the
-		 * default signal on connector). */
-		struct dc_link *link = dc->links[i];
-
-		link->link_enc->funcs->hw_init(link->link_enc);
-	}
-
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		struct timing_generator *tg = dc->res_pool->timing_generators[i];
-
-		tg->funcs->disable_vga(tg);
-
-		/* Blank controller using driver code instead of
-		 * command table. */
-		tg->funcs->set_blank(tg, true);
-		hwss_wait_for_blank_complete(tg);
-	}
-
-	for (i = 0; i < dc->res_pool->audio_count; i++) {
-		struct audio *audio = dc->res_pool->audios[i];
-		audio->funcs->hw_init(audio);
-	}
-
-	for (i = 0; i < dc->link_count; i++) {
-		struct dc_link *link = dc->links[i];
-
-		if (link->panel_cntl)
-			backlight = link->panel_cntl->funcs->hw_init(link->panel_cntl);
-	}
-
-	abm = dc->res_pool->abm;
-	if (abm != NULL)
-		abm->funcs->abm_init(abm, backlight);
-
-	dmcu = dc->res_pool->dmcu;
-	if (dmcu != NULL && abm != NULL)
-		abm->dmcu_is_running = dmcu->funcs->is_dmcu_initialized(dmcu);
-
-	if (dc->fbc_compressor)
-		dc->fbc_compressor->funcs->power_up_fbc(dc->fbc_compressor);
-
-}
-
-
-void dce110_prepare_bandwidth(
-		struct dc *dc,
-		struct dc_state *context)
-{
-	struct clk_mgr *dccg = dc->clk_mgr;
-
-	dce110_set_safe_displaymarks(&context->res_ctx, dc->res_pool);
-
-	dccg->funcs->update_clocks(
-			dccg,
-			context,
-			false);
-}
-
-void dce110_optimize_bandwidth(
-		struct dc *dc,
-		struct dc_state *context)
-{
-	struct clk_mgr *dccg = dc->clk_mgr;
-
-	dce110_set_displaymarks(dc, context);
-
-	dccg->funcs->update_clocks(
-			dccg,
-			context,
-			true);
-}
-
-static void dce110_program_front_end_for_pipe(
-		struct dc *dc, struct pipe_ctx *pipe_ctx)
-{
-	struct mem_input *mi = pipe_ctx->plane_res.mi;
-	struct dc_plane_state *plane_state = pipe_ctx->plane_state;
-	struct xfm_grph_csc_adjustment adjust;
-	struct out_csc_color_matrix tbl_entry;
-	unsigned int i;
-	struct dce_hwseq *hws = dc->hwseq;
-
-	DC_LOGGER_INIT();
-	memset(&tbl_entry, 0, sizeof(tbl_entry));
-
-	memset(&adjust, 0, sizeof(adjust));
-	adjust.gamut_adjust_type = GRAPHICS_GAMUT_ADJUST_TYPE_BYPASS;
-
-	dce_enable_fe_clock(dc->hwseq, mi->inst, true);
-
-	set_default_colors(pipe_ctx);
-	if (pipe_ctx->stream->csc_color_matrix.enable_adjustment
-			== true) {
-		tbl_entry.color_space =
-			pipe_ctx->stream->output_color_space;
-
-		for (i = 0; i < 12; i++)
-			tbl_entry.regval[i] =
-			pipe_ctx->stream->csc_color_matrix.matrix[i];
-
-		pipe_ctx->plane_res.xfm->funcs->opp_set_csc_adjustment
-				(pipe_ctx->plane_res.xfm, &tbl_entry);
-	}
-
-	if (pipe_ctx->stream->gamut_remap_matrix.enable_remap == true) {
-		adjust.gamut_adjust_type = GRAPHICS_GAMUT_ADJUST_TYPE_SW;
-
-		for (i = 0; i < CSC_TEMPERATURE_MATRIX_SIZE; i++)
-			adjust.temperature_matrix[i] =
-				pipe_ctx->stream->gamut_remap_matrix.matrix[i];
-	}
-
-	pipe_ctx->plane_res.xfm->funcs->transform_set_gamut_remap(pipe_ctx->plane_res.xfm, &adjust);
-
-	pipe_ctx->plane_res.scl_data.lb_params.alpha_en = pipe_ctx->bottom_pipe != 0;
-
-	program_scaler(dc, pipe_ctx);
-
-	mi->funcs->mem_input_program_surface_config(
-			mi,
-			plane_state->format,
-			&plane_state->tiling_info,
-			&plane_state->plane_size,
-			plane_state->rotation,
-			NULL,
-			false);
-	if (mi->funcs->set_blank)
-		mi->funcs->set_blank(mi, pipe_ctx->plane_state->visible);
-
-	if (dc->config.gpu_vm_support)
-		mi->funcs->mem_input_program_pte_vm(
-				pipe_ctx->plane_res.mi,
-				plane_state->format,
-				&plane_state->tiling_info,
-				plane_state->rotation);
-
-	/* Moved programming gamma from dc to hwss */
-	if (pipe_ctx->plane_state->update_flags.bits.full_update ||
-			pipe_ctx->plane_state->update_flags.bits.in_transfer_func_change ||
-			pipe_ctx->plane_state->update_flags.bits.gamma_change)
-		hws->funcs.set_input_transfer_func(dc, pipe_ctx, pipe_ctx->plane_state);
-
-	if (pipe_ctx->plane_state->update_flags.bits.full_update)
-		hws->funcs.set_output_transfer_func(dc, pipe_ctx, pipe_ctx->stream);
-
-	DC_LOG_SURFACE(
-			"Pipe:%d %p: addr hi:0x%x, "
-			"addr low:0x%x, "
-			"src: %d, %d, %d,"
-			" %d; dst: %d, %d, %d, %d;"
-			"clip: %d, %d, %d, %d\n",
-			pipe_ctx->pipe_idx,
-			(void *) pipe_ctx->plane_state,
-			pipe_ctx->plane_state->address.grph.addr.high_part,
-			pipe_ctx->plane_state->address.grph.addr.low_part,
-			pipe_ctx->plane_state->src_rect.x,
-			pipe_ctx->plane_state->src_rect.y,
-			pipe_ctx->plane_state->src_rect.width,
-			pipe_ctx->plane_state->src_rect.height,
-			pipe_ctx->plane_state->dst_rect.x,
-			pipe_ctx->plane_state->dst_rect.y,
-			pipe_ctx->plane_state->dst_rect.width,
-			pipe_ctx->plane_state->dst_rect.height,
-			pipe_ctx->plane_state->clip_rect.x,
-			pipe_ctx->plane_state->clip_rect.y,
-			pipe_ctx->plane_state->clip_rect.width,
-			pipe_ctx->plane_state->clip_rect.height);
-
-	DC_LOG_SURFACE(
-			"Pipe %d: width, height, x, y\n"
-			"viewport:%d, %d, %d, %d\n"
-			"recout:  %d, %d, %d, %d\n",
-			pipe_ctx->pipe_idx,
-			pipe_ctx->plane_res.scl_data.viewport.width,
-			pipe_ctx->plane_res.scl_data.viewport.height,
-			pipe_ctx->plane_res.scl_data.viewport.x,
-			pipe_ctx->plane_res.scl_data.viewport.y,
-			pipe_ctx->plane_res.scl_data.recout.width,
-			pipe_ctx->plane_res.scl_data.recout.height,
-			pipe_ctx->plane_res.scl_data.recout.x,
-			pipe_ctx->plane_res.scl_data.recout.y);
-}
-
-static void dce110_apply_ctx_for_surface(
-		struct dc *dc,
-		const struct dc_stream_state *stream,
-		int num_planes,
-		struct dc_state *context)
-{
-	int i;
-
-	if (num_planes == 0)
-		return;
-
-	if (dc->fbc_compressor)
-		dc->fbc_compressor->funcs->disable_fbc(dc->fbc_compressor);
-
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i];
-
-		if (pipe_ctx->stream != stream)
-			continue;
-
-		/* Need to allocate mem before program front end for Fiji */
-		pipe_ctx->plane_res.mi->funcs->allocate_mem_input(
-				pipe_ctx->plane_res.mi,
-				pipe_ctx->stream->timing.h_total,
-				pipe_ctx->stream->timing.v_total,
-				pipe_ctx->stream->timing.pix_clk_100hz / 10,
-				context->stream_count);
-
-		dce110_program_front_end_for_pipe(dc, pipe_ctx);
-
-		dc->hwss.update_plane_addr(dc, pipe_ctx);
-
-		program_surface_visibility(dc, pipe_ctx);
-
-	}
-
-	if (dc->fbc_compressor)
-		enable_fbc(dc, context);
-}
-
-static void dce110_post_unlock_program_front_end(
-		struct dc *dc,
-		struct dc_state *context)
-{
-}
-
-static void dce110_power_down_fe(struct dc *dc, struct pipe_ctx *pipe_ctx)
-{
-	struct dce_hwseq *hws = dc->hwseq;
-	int fe_idx = pipe_ctx->plane_res.mi ?
-		pipe_ctx->plane_res.mi->inst : pipe_ctx->pipe_idx;
-
-	/* Do not power down fe when stream is active on dce*/
-	if (dc->current_state->res_ctx.pipe_ctx[fe_idx].stream)
-		return;
-
-	hws->funcs.enable_display_power_gating(
-		dc, fe_idx, dc->ctx->dc_bios, PIPE_GATING_CONTROL_ENABLE);
-
-	dc->res_pool->transforms[fe_idx]->funcs->transform_reset(
-				dc->res_pool->transforms[fe_idx]);
-}
-
-static void dce110_wait_for_mpcc_disconnect(
-		struct dc *dc,
-		struct resource_pool *res_pool,
-		struct pipe_ctx *pipe_ctx)
-{
-	/* do nothing*/
-}
-
-static void program_output_csc(struct dc *dc,
-		struct pipe_ctx *pipe_ctx,
-		enum dc_color_space colorspace,
-		uint16_t *matrix,
-		int opp_id)
-{
-	int i;
-	struct out_csc_color_matrix tbl_entry;
-
-	if (pipe_ctx->stream->csc_color_matrix.enable_adjustment == true) {
-		enum dc_color_space color_space = pipe_ctx->stream->output_color_space;
-
-		for (i = 0; i < 12; i++)
-			tbl_entry.regval[i] = pipe_ctx->stream->csc_color_matrix.matrix[i];
-
-		tbl_entry.color_space = color_space;
-
-		pipe_ctx->plane_res.xfm->funcs->opp_set_csc_adjustment(
-				pipe_ctx->plane_res.xfm, &tbl_entry);
-	}
-}
-
-static void dce110_set_cursor_position(struct pipe_ctx *pipe_ctx)
-{
-	struct dc_cursor_position pos_cpy = pipe_ctx->stream->cursor_position;
-	struct input_pixel_processor *ipp = pipe_ctx->plane_res.ipp;
-	struct mem_input *mi = pipe_ctx->plane_res.mi;
-	struct dc_cursor_mi_param param = {
-		.pixel_clk_khz = pipe_ctx->stream->timing.pix_clk_100hz / 10,
-		.ref_clk_khz = pipe_ctx->stream->ctx->dc->res_pool->ref_clocks.xtalin_clock_inKhz,
-		.viewport = pipe_ctx->plane_res.scl_data.viewport,
-		.h_scale_ratio = pipe_ctx->plane_res.scl_data.ratios.horz,
-		.v_scale_ratio = pipe_ctx->plane_res.scl_data.ratios.vert,
-		.rotation = pipe_ctx->plane_state->rotation,
-		.mirror = pipe_ctx->plane_state->horizontal_mirror
-	};
-
-	/**
-	 * If the cursor's source viewport is clipped then we need to
-	 * translate the cursor to appear in the correct position on
-	 * the screen.
-	 *
-	 * This translation isn't affected by scaling so it needs to be
-	 * done *after* we adjust the position for the scale factor.
-	 *
-	 * This is only done by opt-in for now since there are still
-	 * some usecases like tiled display that might enable the
-	 * cursor on both streams while expecting dc to clip it.
-	 */
-	if (pos_cpy.translate_by_source) {
-		pos_cpy.x += pipe_ctx->plane_state->src_rect.x;
-		pos_cpy.y += pipe_ctx->plane_state->src_rect.y;
-	}
-
-	if (pipe_ctx->plane_state->address.type
-			== PLN_ADDR_TYPE_VIDEO_PROGRESSIVE)
-		pos_cpy.enable = false;
-
-	if (pipe_ctx->top_pipe && pipe_ctx->plane_state != pipe_ctx->top_pipe->plane_state)
-		pos_cpy.enable = false;
-
-	if (ipp->funcs->ipp_cursor_set_position)
-		ipp->funcs->ipp_cursor_set_position(ipp, &pos_cpy, &param);
-	if (mi->funcs->set_cursor_position)
-		mi->funcs->set_cursor_position(mi, &pos_cpy, &param);
-}
-
-static void dce110_set_cursor_attribute(struct pipe_ctx *pipe_ctx)
-{
-	struct dc_cursor_attributes *attributes = &pipe_ctx->stream->cursor_attributes;
-
-	if (pipe_ctx->plane_res.ipp &&
-	    pipe_ctx->plane_res.ipp->funcs->ipp_cursor_set_attributes)
-		pipe_ctx->plane_res.ipp->funcs->ipp_cursor_set_attributes(
-				pipe_ctx->plane_res.ipp, attributes);
-
-	if (pipe_ctx->plane_res.mi &&
-	    pipe_ctx->plane_res.mi->funcs->set_cursor_attributes)
-		pipe_ctx->plane_res.mi->funcs->set_cursor_attributes(
-				pipe_ctx->plane_res.mi, attributes);
-
-	if (pipe_ctx->plane_res.xfm &&
-	    pipe_ctx->plane_res.xfm->funcs->set_cursor_attributes)
-		pipe_ctx->plane_res.xfm->funcs->set_cursor_attributes(
-				pipe_ctx->plane_res.xfm, attributes);
-}
-
-bool dce110_set_backlight_level(struct pipe_ctx *pipe_ctx,
-		uint32_t backlight_pwm_u16_16,
-		uint32_t frame_ramp)
-{
-	struct dc_link *link = pipe_ctx->stream->link;
-	struct dc  *dc = link->ctx->dc;
-	struct abm *abm = pipe_ctx->stream_res.abm;
-	struct panel_cntl *panel_cntl = link->panel_cntl;
-	struct dmcu *dmcu = dc->res_pool->dmcu;
-	bool fw_set_brightness = true;
-	/* DMCU -1 for all controller id values,
-	 * therefore +1 here
-	 */
-	uint32_t controller_id = pipe_ctx->stream_res.tg->inst + 1;
-
-	if (abm == NULL || panel_cntl == NULL || (abm->funcs->set_backlight_level_pwm == NULL))
-		return false;
-
-	if (dmcu)
-		fw_set_brightness = dmcu->funcs->is_dmcu_initialized(dmcu);
-
-	if (!fw_set_brightness && panel_cntl->funcs->driver_set_backlight)
-		panel_cntl->funcs->driver_set_backlight(panel_cntl, backlight_pwm_u16_16);
-	else
-		abm->funcs->set_backlight_level_pwm(
-				abm,
-				backlight_pwm_u16_16,
-				frame_ramp,
-				controller_id,
-				link->panel_cntl->inst);
-
-	return true;
-}
-
-void dce110_set_abm_immediate_disable(struct pipe_ctx *pipe_ctx)
-{
-	struct abm *abm = pipe_ctx->stream_res.abm;
-	struct panel_cntl *panel_cntl = pipe_ctx->stream->link->panel_cntl;
-
-	if (abm)
-		abm->funcs->set_abm_immediate_disable(abm,
-				pipe_ctx->stream->link->panel_cntl->inst);
-
-	if (panel_cntl)
-		panel_cntl->funcs->store_backlight_level(panel_cntl);
-}
-
-void dce110_set_pipe(struct pipe_ctx *pipe_ctx)
-{
-	struct abm *abm = pipe_ctx->stream_res.abm;
-	struct panel_cntl *panel_cntl = pipe_ctx->stream->link->panel_cntl;
-	uint32_t otg_inst = pipe_ctx->stream_res.tg->inst + 1;
-
-	if (abm && panel_cntl)
-		abm->funcs->set_pipe(abm, otg_inst, panel_cntl->inst);
-}
-
-static const struct hw_sequencer_funcs dce110_funcs = {
-	.program_gamut_remap = program_gamut_remap,
-	.program_output_csc = program_output_csc,
-	.init_hw = init_hw,
-	.apply_ctx_to_hw = dce110_apply_ctx_to_hw,
-	.apply_ctx_for_surface = dce110_apply_ctx_for_surface,
-	.post_unlock_program_front_end = dce110_post_unlock_program_front_end,
-	.update_plane_addr = update_plane_addr,
-	.update_pending_status = dce110_update_pending_status,
-	.enable_accelerated_mode = dce110_enable_accelerated_mode,
-	.enable_timing_synchronization = dce110_enable_timing_synchronization,
-	.enable_per_frame_crtc_position_reset = dce110_enable_per_frame_crtc_position_reset,
-	.update_info_frame = dce110_update_info_frame,
-	.enable_stream = dce110_enable_stream,
-	.disable_stream = dce110_disable_stream,
-	.unblank_stream = dce110_unblank_stream,
-	.blank_stream = dce110_blank_stream,
-	.enable_audio_stream = dce110_enable_audio_stream,
-	.disable_audio_stream = dce110_disable_audio_stream,
-	.disable_plane = dce110_power_down_fe,
-	.pipe_control_lock = dce_pipe_control_lock,
-	.interdependent_update_lock = NULL,
-	.cursor_lock = dce_pipe_control_lock,
-	.prepare_bandwidth = dce110_prepare_bandwidth,
-	.optimize_bandwidth = dce110_optimize_bandwidth,
-	.set_drr = set_drr,
-	.get_position = get_position,
-	.set_static_screen_control = set_static_screen_control,
-	.setup_stereo = NULL,
-	.set_avmute = dce110_set_avmute,
-	.wait_for_mpcc_disconnect = dce110_wait_for_mpcc_disconnect,
-	.edp_backlight_control = dce110_edp_backlight_control,
-	.edp_power_control = dce110_edp_power_control,
-	.edp_wait_for_hpd_ready = dce110_edp_wait_for_hpd_ready,
-	.set_cursor_position = dce110_set_cursor_position,
-	.set_cursor_attribute = dce110_set_cursor_attribute,
-	.set_backlight_level = dce110_set_backlight_level,
-	.set_abm_immediate_disable = dce110_set_abm_immediate_disable,
-	.set_pipe = dce110_set_pipe,
-};
-
-static const struct hwseq_private_funcs dce110_private_funcs = {
-	.init_pipes = init_pipes,
-	.update_plane_addr = update_plane_addr,
-	.set_input_transfer_func = dce110_set_input_transfer_func,
-	.set_output_transfer_func = dce110_set_output_transfer_func,
-	.power_down = dce110_power_down,
-	.enable_display_pipe_clock_gating = enable_display_pipe_clock_gating,
-	.enable_display_power_gating = dce110_enable_display_power_gating,
-	.reset_hw_ctx_wrap = dce110_reset_hw_ctx_wrap,
-	.enable_stream_timing = dce110_enable_stream_timing,
-	.disable_stream_gating = NULL,
-	.enable_stream_gating = NULL,
-	.edp_backlight_control = dce110_edp_backlight_control,
-};
-
-void dce110_hw_sequencer_construct(struct dc *dc)
-{
-	dc->hwss = dce110_funcs;
-	dc->hwseq->funcs = dce110_private_funcs;
-}
-
diff --git a/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.h b/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.h
deleted file mode 100644
index b6f3843d3d05..000000000000
--- a/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.h
+++ /dev/null
@@ -1,95 +0,0 @@
-/*
-* Copyright 2012-15 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: AMD
- *
- */
-
-#ifndef __DC_HWSS_DCE110_H__
-#define __DC_HWSS_DCE110_H__
-
-#include "core_types.h"
-#include "hw_sequencer_private.h"
-
-struct dc;
-struct dc_state;
-struct dm_pp_display_configuration;
-
-void dce110_hw_sequencer_construct(struct dc *dc);
-
-enum dc_status dce110_apply_ctx_to_hw(
-		struct dc *dc,
-		struct dc_state *context);
-
-
-void dce110_enable_stream(struct pipe_ctx *pipe_ctx);
-
-void dce110_disable_stream(struct pipe_ctx *pipe_ctx);
-
-void dce110_unblank_stream(struct pipe_ctx *pipe_ctx,
-		struct dc_link_settings *link_settings);
-
-void dce110_blank_stream(struct pipe_ctx *pipe_ctx);
-
-void dce110_enable_audio_stream(struct pipe_ctx *pipe_ctx);
-void dce110_disable_audio_stream(struct pipe_ctx *pipe_ctx);
-
-void dce110_update_info_frame(struct pipe_ctx *pipe_ctx);
-
-void dce110_set_avmute(struct pipe_ctx *pipe_ctx, bool enable);
-void dce110_enable_accelerated_mode(struct dc *dc, struct dc_state *context);
-
-void dce110_power_down(struct dc *dc);
-
-void dce110_set_safe_displaymarks(
-		struct resource_context *res_ctx,
-		const struct resource_pool *pool);
-
-void dce110_prepare_bandwidth(
-		struct dc *dc,
-		struct dc_state *context);
-
-void dce110_optimize_bandwidth(
-		struct dc *dc,
-		struct dc_state *context);
-
-void dp_receiver_power_ctrl(struct dc_link *link, bool on);
-
-void dce110_edp_power_control(
-		struct dc_link *link,
-		bool power_up);
-
-void dce110_edp_backlight_control(
-	struct dc_link *link,
-	bool enable);
-
-void dce110_edp_wait_for_hpd_ready(
-		struct dc_link *link,
-		bool power_up);
-
-bool dce110_set_backlight_level(struct pipe_ctx *pipe_ctx,
-		uint32_t backlight_pwm_u16_16,
-		uint32_t frame_ramp);
-void dce110_set_abm_immediate_disable(struct pipe_ctx *pipe_ctx);
-void dce110_set_pipe(struct pipe_ctx *pipe_ctx);
-
-#endif /* __DC_HWSS_DCE110_H__ */
-
diff --git a/drivers/gpu/drm/amd/display/dc/dce110/dce110_mem_input_v.c b/drivers/gpu/drm/amd/display/dc/dce110/dce110_mem_input_v.c
index db7557a1c613..2c43c2422638 100644
--- a/drivers/gpu/drm/amd/display/dc/dce110/dce110_mem_input_v.c
+++ b/drivers/gpu/drm/amd/display/dc/dce110/dce110_mem_input_v.c
@@ -76,7 +76,6 @@ UNP_GRPH_PRIMARY_SURFACE_ADDRESS_HIGH_C__GRPH_PRIMARY_SURFACE_ADDRESS_HIGH_C_MAS
 		mmUNP_GRPH_PRIMARY_SURFACE_ADDRESS_HIGH_C,
 		value);
 
-	temp = 0;
 	value = 0;
 	temp = address.low_part >>
 	UNP_GRPH_PRIMARY_SURFACE_ADDRESS_C__GRPH_PRIMARY_SURFACE_ADDRESS_C__SHIFT;
@@ -112,7 +111,6 @@ UNP_GRPH_PRIMARY_SURFACE_ADDRESS_HIGH_L__GRPH_PRIMARY_SURFACE_ADDRESS_HIGH_L_MAS
 		mmUNP_GRPH_PRIMARY_SURFACE_ADDRESS_HIGH_L,
 		value);
 
-	temp = 0;
 	value = 0;
 	temp = address.low_part >>
 	UNP_GRPH_PRIMARY_SURFACE_ADDRESS_L__GRPH_PRIMARY_SURFACE_ADDRESS_L__SHIFT;
@@ -164,7 +162,7 @@ static void enable(struct dce_mem_input *mem_input110)
 
 static void program_tiling(
 	struct dce_mem_input *mem_input110,
-	const union dc_tiling_info *info,
+	const struct dc_tiling_info *info,
 	const enum surface_pixel_format pixel_format)
 {
 	uint32_t value = 0;
@@ -525,7 +523,7 @@ static const unsigned int dvmm_Hw_Setting_Linear[4][9] = {
 
 /* Helper to get table entry from surface info */
 static const unsigned int *get_dvmm_hw_setting(
-		union dc_tiling_info *tiling_info,
+		struct dc_tiling_info *tiling_info,
 		enum surface_pixel_format format,
 		bool chroma)
 {
@@ -565,7 +563,7 @@ static const unsigned int *get_dvmm_hw_setting(
 static void dce_mem_input_v_program_pte_vm(
 		struct mem_input *mem_input,
 		enum surface_pixel_format format,
-		union dc_tiling_info *tiling_info,
+		struct dc_tiling_info *tiling_info,
 		enum dc_rotation_angle rotation)
 {
 	struct dce_mem_input *mem_input110 = TO_DCE_MEM_INPUT(mem_input);
@@ -638,7 +636,7 @@ static void dce_mem_input_v_program_pte_vm(
 static void dce_mem_input_v_program_surface_config(
 	struct mem_input *mem_input,
 	enum surface_pixel_format format,
-	union dc_tiling_info *tiling_info,
+	struct dc_tiling_info *tiling_info,
 	struct plane_size *plane_size,
 	enum dc_rotation_angle rotation,
 	struct dc_plane_dcc_param *dcc,
diff --git a/drivers/gpu/drm/amd/display/dc/dce110/dce110_opp_regamma_v.c b/drivers/gpu/drm/amd/display/dc/dce110/dce110_opp_regamma_v.c
index 34c5e3c7c6d2..9b65b77e8823 100644
--- a/drivers/gpu/drm/amd/display/dc/dce110/dce110_opp_regamma_v.c
+++ b/drivers/gpu/drm/amd/display/dc/dce110/dce110_opp_regamma_v.c
@@ -23,8 +23,6 @@
  *
  */
 
-#include <linux/delay.h>
-
 #include "dm_services.h"
 
 /* include DCE11 register header files */
diff --git a/drivers/gpu/drm/amd/display/dc/dce110/dce110_timing_generator.c b/drivers/gpu/drm/amd/display/dc/dce110/dce110_timing_generator.c
index 27cbb5b42c7e..61b0807693fb 100644
--- a/drivers/gpu/drm/amd/display/dc/dce110/dce110_timing_generator.c
+++ b/drivers/gpu/drm/amd/display/dc/dce110/dce110_timing_generator.c
@@ -288,7 +288,7 @@ bool dce110_timing_generator_program_timing_generator(
 
 	uint32_t vsync_offset = dc_crtc_timing->v_border_bottom +
 			dc_crtc_timing->v_front_porch;
-	uint32_t v_sync_start =dc_crtc_timing->v_addressable + vsync_offset;
+	uint32_t v_sync_start = dc_crtc_timing->v_addressable + vsync_offset;
 
 	uint32_t hsync_offset = dc_crtc_timing->h_border_right +
 			dc_crtc_timing->h_front_porch;
@@ -603,7 +603,7 @@ void dce110_timing_generator_program_blanking(
 {
 	uint32_t vsync_offset = timing->v_border_bottom +
 			timing->v_front_porch;
-	uint32_t v_sync_start =timing->v_addressable + vsync_offset;
+	uint32_t v_sync_start = timing->v_addressable + vsync_offset;
 
 	uint32_t hsync_offset = timing->h_border_right +
 			timing->h_front_porch;
@@ -1955,6 +1955,7 @@ void dce110_tg_program_timing(struct timing_generator *tg,
 	int vstartup_start,
 	int vupdate_offset,
 	int vupdate_width,
+	int pstate_keepout,
 	const enum signal_type signal,
 	bool use_vbios)
 {
@@ -2015,6 +2016,23 @@ bool dce110_tg_validate_timing(struct timing_generator *tg,
 	return dce110_timing_generator_validate_timing(tg, timing, SIGNAL_TYPE_NONE);
 }
 
+/* "Container" vs. "pixel" is a concept within HW blocks, mostly those closer to the back-end. It works like this:
+ *
+ * - In most of the formats (RGB or YCbCr 4:4:4, 4:2:2 uncompressed and DSC 4:2:2 Simple) pixel rate is the same as
+ *   container rate.
+ *
+ * - In 4:2:0 (DSC or uncompressed) there are two pixels per container, hence the target container rate has to be
+ *   halved to maintain the correct pixel rate.
+ *
+ * - Unlike 4:2:2 uncompressed, DSC 4:2:2 Native also has two pixels per container (this happens when DSC is applied
+ *   to it) and has to be treated the same as 4:2:0, i.e. target containter rate has to be halved in this case as well.
+ *
+ */
+bool dce110_is_two_pixels_per_container(const struct dc_crtc_timing *timing)
+{
+	return timing->pixel_encoding == PIXEL_ENCODING_YCBCR420;
+}
+
 void dce110_tg_wait_for_state(struct timing_generator *tg,
 	enum crtc_state state)
 {
@@ -2109,70 +2127,131 @@ bool dce110_configure_crc(struct timing_generator *tg,
 
 	cntl_addr = CRTC_REG(mmCRTC_CRC_CNTL);
 
-	/* First, disable CRC before we configure it. */
-	dm_write_reg(tg->ctx, cntl_addr, 0);
+	if (!params->enable || params->reset)
+		/* First, disable CRC before we configure it. */
+		dm_write_reg(tg->ctx, cntl_addr, 0);
 
 	if (!params->enable)
 		return true;
 
 	/* Program frame boundaries */
-	/* Window A x axis start and end. */
-	value = 0;
-	addr = CRTC_REG(mmCRTC_CRC0_WINDOWA_X_CONTROL);
-	set_reg_field_value(value, params->windowa_x_start,
-			    CRTC_CRC0_WINDOWA_X_CONTROL,
-			    CRTC_CRC0_WINDOWA_X_START);
-	set_reg_field_value(value, params->windowa_x_end,
-			    CRTC_CRC0_WINDOWA_X_CONTROL,
-			    CRTC_CRC0_WINDOWA_X_END);
-	dm_write_reg(tg->ctx, addr, value);
-
-	/* Window A y axis start and end. */
-	value = 0;
-	addr = CRTC_REG(mmCRTC_CRC0_WINDOWA_Y_CONTROL);
-	set_reg_field_value(value, params->windowa_y_start,
-			    CRTC_CRC0_WINDOWA_Y_CONTROL,
-			    CRTC_CRC0_WINDOWA_Y_START);
-	set_reg_field_value(value, params->windowa_y_end,
-			    CRTC_CRC0_WINDOWA_Y_CONTROL,
-			    CRTC_CRC0_WINDOWA_Y_END);
-	dm_write_reg(tg->ctx, addr, value);
-
-	/* Window B x axis start and end. */
-	value = 0;
-	addr = CRTC_REG(mmCRTC_CRC0_WINDOWB_X_CONTROL);
-	set_reg_field_value(value, params->windowb_x_start,
-			    CRTC_CRC0_WINDOWB_X_CONTROL,
-			    CRTC_CRC0_WINDOWB_X_START);
-	set_reg_field_value(value, params->windowb_x_end,
-			    CRTC_CRC0_WINDOWB_X_CONTROL,
-			    CRTC_CRC0_WINDOWB_X_END);
-	dm_write_reg(tg->ctx, addr, value);
-
-	/* Window B y axis start and end. */
-	value = 0;
-	addr = CRTC_REG(mmCRTC_CRC0_WINDOWB_Y_CONTROL);
-	set_reg_field_value(value, params->windowb_y_start,
-			    CRTC_CRC0_WINDOWB_Y_CONTROL,
-			    CRTC_CRC0_WINDOWB_Y_START);
-	set_reg_field_value(value, params->windowb_y_end,
-			    CRTC_CRC0_WINDOWB_Y_CONTROL,
-			    CRTC_CRC0_WINDOWB_Y_END);
-	dm_write_reg(tg->ctx, addr, value);
-
-	/* Set crc mode and selection, and enable. Only using CRC0*/
-	value = 0;
-	set_reg_field_value(value, params->continuous_mode ? 1 : 0,
-			    CRTC_CRC_CNTL, CRTC_CRC_CONT_EN);
-	set_reg_field_value(value, params->selection,
-			    CRTC_CRC_CNTL, CRTC_CRC0_SELECT);
-	set_reg_field_value(value, 1, CRTC_CRC_CNTL, CRTC_CRC_EN);
-	dm_write_reg(tg->ctx, cntl_addr, value);
+	switch (params->crc_eng_inst) {
+	case 0:
+		/* Window A x axis start and end. */
+		value = 0;
+		addr = CRTC_REG(mmCRTC_CRC0_WINDOWA_X_CONTROL);
+		set_reg_field_value(value, params->windowa_x_start,
+				    CRTC_CRC0_WINDOWA_X_CONTROL,
+				    CRTC_CRC0_WINDOWA_X_START);
+		set_reg_field_value(value, params->windowa_x_end,
+				    CRTC_CRC0_WINDOWA_X_CONTROL,
+				    CRTC_CRC0_WINDOWA_X_END);
+		dm_write_reg(tg->ctx, addr, value);
+
+		/* Window A y axis start and end. */
+		value = 0;
+		addr = CRTC_REG(mmCRTC_CRC0_WINDOWA_Y_CONTROL);
+		set_reg_field_value(value, params->windowa_y_start,
+				    CRTC_CRC0_WINDOWA_Y_CONTROL,
+				    CRTC_CRC0_WINDOWA_Y_START);
+		set_reg_field_value(value, params->windowa_y_end,
+				    CRTC_CRC0_WINDOWA_Y_CONTROL,
+				    CRTC_CRC0_WINDOWA_Y_END);
+		dm_write_reg(tg->ctx, addr, value);
+
+		/* Window B x axis start and end. */
+		value = 0;
+		addr = CRTC_REG(mmCRTC_CRC0_WINDOWB_X_CONTROL);
+		set_reg_field_value(value, params->windowb_x_start,
+				    CRTC_CRC0_WINDOWB_X_CONTROL,
+				    CRTC_CRC0_WINDOWB_X_START);
+		set_reg_field_value(value, params->windowb_x_end,
+				    CRTC_CRC0_WINDOWB_X_CONTROL,
+				    CRTC_CRC0_WINDOWB_X_END);
+		dm_write_reg(tg->ctx, addr, value);
+
+		/* Window B y axis start and end. */
+		value = 0;
+		addr = CRTC_REG(mmCRTC_CRC0_WINDOWB_Y_CONTROL);
+		set_reg_field_value(value, params->windowb_y_start,
+				    CRTC_CRC0_WINDOWB_Y_CONTROL,
+				    CRTC_CRC0_WINDOWB_Y_START);
+		set_reg_field_value(value, params->windowb_y_end,
+				    CRTC_CRC0_WINDOWB_Y_CONTROL,
+				    CRTC_CRC0_WINDOWB_Y_END);
+		dm_write_reg(tg->ctx, addr, value);
+
+		/* Set crc mode and selection, and enable.*/
+		value = 0;
+		set_reg_field_value(value, params->continuous_mode ? 1 : 0,
+				    CRTC_CRC_CNTL, CRTC_CRC_CONT_EN);
+		set_reg_field_value(value, params->selection,
+				    CRTC_CRC_CNTL, CRTC_CRC0_SELECT);
+		set_reg_field_value(value, 1, CRTC_CRC_CNTL, CRTC_CRC_EN);
+		dm_write_reg(tg->ctx, cntl_addr, value);
+		break;
+	case 1:
+		/* Window A x axis start and end. */
+		value = 0;
+		addr = CRTC_REG(mmCRTC_CRC1_WINDOWA_X_CONTROL);
+		set_reg_field_value(value, params->windowa_x_start,
+				    CRTC_CRC1_WINDOWA_X_CONTROL,
+				    CRTC_CRC1_WINDOWA_X_START);
+		set_reg_field_value(value, params->windowa_x_end,
+				    CRTC_CRC1_WINDOWA_X_CONTROL,
+				    CRTC_CRC1_WINDOWA_X_END);
+		dm_write_reg(tg->ctx, addr, value);
+
+		/* Window A y axis start and end. */
+		value = 0;
+		addr = CRTC_REG(mmCRTC_CRC1_WINDOWA_Y_CONTROL);
+		set_reg_field_value(value, params->windowa_y_start,
+				    CRTC_CRC1_WINDOWA_Y_CONTROL,
+				    CRTC_CRC1_WINDOWA_Y_START);
+		set_reg_field_value(value, params->windowa_y_end,
+				    CRTC_CRC1_WINDOWA_Y_CONTROL,
+				    CRTC_CRC1_WINDOWA_Y_END);
+		dm_write_reg(tg->ctx, addr, value);
+
+		/* Window B x axis start and end. */
+		value = 0;
+		addr = CRTC_REG(mmCRTC_CRC1_WINDOWB_X_CONTROL);
+		set_reg_field_value(value, params->windowb_x_start,
+				    CRTC_CRC1_WINDOWB_X_CONTROL,
+				    CRTC_CRC1_WINDOWB_X_START);
+		set_reg_field_value(value, params->windowb_x_end,
+				    CRTC_CRC1_WINDOWB_X_CONTROL,
+				    CRTC_CRC1_WINDOWB_X_END);
+		dm_write_reg(tg->ctx, addr, value);
+
+		/* Window B y axis start and end. */
+		value = 0;
+		addr = CRTC_REG(mmCRTC_CRC1_WINDOWB_Y_CONTROL);
+		set_reg_field_value(value, params->windowb_y_start,
+				    CRTC_CRC1_WINDOWB_Y_CONTROL,
+				    CRTC_CRC1_WINDOWB_Y_START);
+		set_reg_field_value(value, params->windowb_y_end,
+				    CRTC_CRC1_WINDOWB_Y_CONTROL,
+				    CRTC_CRC1_WINDOWB_Y_END);
+		dm_write_reg(tg->ctx, addr, value);
+
+		/* Set crc mode and selection, and enable.*/
+		value = 0;
+		set_reg_field_value(value, params->continuous_mode ? 1 : 0,
+				    CRTC_CRC_CNTL, CRTC_CRC_CONT_EN);
+		set_reg_field_value(value, params->selection,
+				    CRTC_CRC_CNTL, CRTC_CRC1_SELECT);
+		set_reg_field_value(value, 1, CRTC_CRC_CNTL, CRTC_CRC_EN);
+		dm_write_reg(tg->ctx, cntl_addr, value);
+		break;
+	default:
+		return false;
+	}
 
 	return true;
 }
 
-bool dce110_get_crc(struct timing_generator *tg,
+bool dce110_get_crc(struct timing_generator *tg, uint8_t idx,
 		    uint32_t *r_cr, uint32_t *g_y, uint32_t *b_cb)
 {
 	uint32_t addr = 0;
@@ -2188,14 +2267,30 @@ bool dce110_get_crc(struct timing_generator *tg,
 	if (!field)
 		return false;
 
-	addr = CRTC_REG(mmCRTC_CRC0_DATA_RG);
-	value = dm_read_reg(tg->ctx, addr);
-	*r_cr = get_reg_field_value(value, CRTC_CRC0_DATA_RG, CRC0_R_CR);
-	*g_y = get_reg_field_value(value, CRTC_CRC0_DATA_RG, CRC0_G_Y);
+	switch (idx) {
+	case 0:
+		addr = CRTC_REG(mmCRTC_CRC0_DATA_RG);
+		value = dm_read_reg(tg->ctx, addr);
+		*r_cr = get_reg_field_value(value, CRTC_CRC0_DATA_RG, CRC0_R_CR);
+		*g_y = get_reg_field_value(value, CRTC_CRC0_DATA_RG, CRC0_G_Y);
 
-	addr = CRTC_REG(mmCRTC_CRC0_DATA_B);
-	value = dm_read_reg(tg->ctx, addr);
-	*b_cb = get_reg_field_value(value, CRTC_CRC0_DATA_B, CRC0_B_CB);
+		addr = CRTC_REG(mmCRTC_CRC0_DATA_B);
+		value = dm_read_reg(tg->ctx, addr);
+		*b_cb = get_reg_field_value(value, CRTC_CRC0_DATA_B, CRC0_B_CB);
+		break;
+	case 1:
+		addr = CRTC_REG(mmCRTC_CRC1_DATA_RG);
+		value = dm_read_reg(tg->ctx, addr);
+		*r_cr = get_reg_field_value(value, CRTC_CRC1_DATA_RG, CRC1_R_CR);
+		*g_y = get_reg_field_value(value, CRTC_CRC1_DATA_RG, CRC1_G_Y);
+
+		addr = CRTC_REG(mmCRTC_CRC1_DATA_B);
+		value = dm_read_reg(tg->ctx, addr);
+		*b_cb = get_reg_field_value(value, CRTC_CRC1_DATA_B, CRC1_B_CB);
+		break;
+	default:
+		return false;
+	}
 
 	return true;
 }
@@ -2239,6 +2334,7 @@ static const struct timing_generator_funcs dce110_tg_funcs = {
 		.is_tg_enabled = dce110_is_tg_enabled,
 		.configure_crc = dce110_configure_crc,
 		.get_crc = dce110_get_crc,
+		.is_two_pixels_per_container = dce110_is_two_pixels_per_container,
 };
 
 void dce110_timing_generator_construct(
diff --git a/drivers/gpu/drm/amd/display/dc/dce110/dce110_timing_generator.h b/drivers/gpu/drm/amd/display/dc/dce110/dce110_timing_generator.h
index d8a5ed7b485d..e4f5cad64f32 100644
--- a/drivers/gpu/drm/amd/display/dc/dce110/dce110_timing_generator.h
+++ b/drivers/gpu/drm/amd/display/dc/dce110/dce110_timing_generator.h
@@ -261,6 +261,7 @@ void dce110_tg_program_timing(struct timing_generator *tg,
 	int vstartup_start,
 	int vupdate_offset,
 	int vupdate_width,
+	int pstate_keepout,
 	const enum signal_type signal,
 	bool use_vbios);
 
@@ -285,7 +286,9 @@ bool dce110_arm_vert_intr(
 bool dce110_configure_crc(struct timing_generator *tg,
 			  const struct crc_params *params);
 
-bool dce110_get_crc(struct timing_generator *tg,
+bool dce110_get_crc(struct timing_generator *tg, uint8_t idx,
 		    uint32_t *r_cr, uint32_t *g_y, uint32_t *b_cb);
 
+bool dce110_is_two_pixels_per_container(const struct dc_crtc_timing *timing);
+
 #endif /* __DC_TIMING_GENERATOR_DCE110_H__ */
diff --git a/drivers/gpu/drm/amd/display/dc/dce110/dce110_timing_generator_v.c b/drivers/gpu/drm/amd/display/dc/dce110/dce110_timing_generator_v.c
index c509384fff54..9837dec837ff 100644
--- a/drivers/gpu/drm/amd/display/dc/dce110/dce110_timing_generator_v.c
+++ b/drivers/gpu/drm/amd/display/dc/dce110/dce110_timing_generator_v.c
@@ -438,6 +438,7 @@ static void dce110_timing_generator_v_program_timing(struct timing_generator *tg
 	int vstartup_start,
 	int vupdate_offset,
 	int vupdate_width,
+	int pstate_keepout,
 	const enum signal_type signal,
 	bool use_vbios)
 {
@@ -682,7 +683,8 @@ static const struct timing_generator_funcs dce110_tg_v_funcs = {
 		.tear_down_global_swap_lock =
 				dce110_timing_generator_v_tear_down_global_swap_lock,
 		.enable_advanced_request =
-				dce110_timing_generator_v_enable_advanced_request
+				dce110_timing_generator_v_enable_advanced_request,
+		.is_two_pixels_per_container = dce110_is_two_pixels_per_container,
 };
 
 void dce110_timing_generator_v_construct(
diff --git a/drivers/gpu/drm/amd/display/dc/dce110/dce110_transform_v.c b/drivers/gpu/drm/amd/display/dc/dce110/dce110_transform_v.c
index 45bca0db5e5e..28d3b2663cd3 100644
--- a/drivers/gpu/drm/amd/display/dc/dce110/dce110_transform_v.c
+++ b/drivers/gpu/drm/amd/display/dc/dce110/dce110_transform_v.c
@@ -23,8 +23,6 @@
  *
  */
 
-#include <linux/delay.h>
-
 #include "dce110_transform_v.h"
 #include "dm_services.h"
 #include "dc.h"
diff --git a/drivers/gpu/drm/amd/display/dc/dce112/Makefile b/drivers/gpu/drm/amd/display/dc/dce112/Makefile
index 9de6501702d2..683866797709 100644
--- a/drivers/gpu/drm/amd/display/dc/dce112/Makefile
+++ b/drivers/gpu/drm/amd/display/dc/dce112/Makefile
@@ -23,10 +23,9 @@
 # Makefile for the 'controller' sub-component of DAL.
 # It provides the control and status of HW CRTC block.
 
-CFLAGS_$(AMDDALPATH)/dc/dce112/dce112_resource.o = $(call cc-disable-warning, override-init)
+CFLAGS_$(AMDDALPATH)/dc/dce112/dce112_resource.o = -Wno-override-init
 
-DCE112 = dce112_compressor.o dce112_hw_sequencer.o \
-dce112_resource.o
+DCE112 = dce112_compressor.o
 
 AMD_DAL_DCE112 = $(addprefix $(AMDDALPATH)/dc/dce112/,$(DCE112))
 
diff --git a/drivers/gpu/drm/amd/display/dc/dce112/dce112_compressor.c b/drivers/gpu/drm/amd/display/dc/dce112/dce112_compressor.c
index 51cb45d8b9ab..faae12cf7968 100644
--- a/drivers/gpu/drm/amd/display/dc/dce112/dce112_compressor.c
+++ b/drivers/gpu/drm/amd/display/dc/dce112/dce112_compressor.c
@@ -23,9 +23,6 @@
  *
  */
 
-#include <linux/delay.h>
-#include <linux/slab.h>
-
 #include "dm_services.h"
 
 #include "dce/dce_11_2_d.h"
diff --git a/drivers/gpu/drm/amd/display/dc/dce112/dce112_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/dce112/dce112_hw_sequencer.c
deleted file mode 100644
index 19873ee1f78d..000000000000
--- a/drivers/gpu/drm/amd/display/dc/dce112/dce112_hw_sequencer.c
+++ /dev/null
@@ -1,163 +0,0 @@
-/*
- * Copyright 2015 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: AMD
- *
- */
-
-#include "dm_services.h"
-#include "dc.h"
-#include "core_types.h"
-#include "dce112_hw_sequencer.h"
-
-#include "dce110/dce110_hw_sequencer.h"
-
-/* include DCE11.2 register header files */
-#include "dce/dce_11_2_d.h"
-#include "dce/dce_11_2_sh_mask.h"
-
-struct dce112_hw_seq_reg_offsets {
-	uint32_t crtc;
-};
-
-
-static const struct dce112_hw_seq_reg_offsets reg_offsets[] = {
-{
-	.crtc = (mmCRTC0_CRTC_GSL_CONTROL - mmCRTC_GSL_CONTROL),
-},
-{
-	.crtc = (mmCRTC1_CRTC_GSL_CONTROL - mmCRTC_GSL_CONTROL),
-},
-{
-	.crtc = (mmCRTC2_CRTC_GSL_CONTROL - mmCRTC_GSL_CONTROL),
-},
-{
-	.crtc = (mmCRTC3_CRTC_GSL_CONTROL - mmCRTC_GSL_CONTROL),
-},
-{
-	.crtc = (mmCRTC4_CRTC_GSL_CONTROL - mmCRTC_GSL_CONTROL),
-},
-{
-	.crtc = (mmCRTC5_CRTC_GSL_CONTROL - mmCRTC_GSL_CONTROL),
-}
-};
-#define HW_REG_CRTC(reg, id)\
-	(reg + reg_offsets[id].crtc)
-
-/*******************************************************************************
- * Private definitions
- ******************************************************************************/
-
-static void dce112_init_pte(struct dc_context *ctx)
-{
-	uint32_t addr;
-	uint32_t value = 0;
-	uint32_t chunk_int = 0;
-	uint32_t chunk_mul = 0;
-
-	addr = mmDVMM_PTE_REQ;
-	value = dm_read_reg(ctx, addr);
-
-	chunk_int = get_reg_field_value(
-		value,
-		DVMM_PTE_REQ,
-		HFLIP_PTEREQ_PER_CHUNK_INT);
-
-	chunk_mul = get_reg_field_value(
-		value,
-		DVMM_PTE_REQ,
-		HFLIP_PTEREQ_PER_CHUNK_MULTIPLIER);
-
-	if (chunk_int != 0x4 || chunk_mul != 0x4) {
-
-		set_reg_field_value(
-			value,
-			255,
-			DVMM_PTE_REQ,
-			MAX_PTEREQ_TO_ISSUE);
-
-		set_reg_field_value(
-			value,
-			4,
-			DVMM_PTE_REQ,
-			HFLIP_PTEREQ_PER_CHUNK_INT);
-
-		set_reg_field_value(
-			value,
-			4,
-			DVMM_PTE_REQ,
-			HFLIP_PTEREQ_PER_CHUNK_MULTIPLIER);
-
-		dm_write_reg(ctx, addr, value);
-	}
-}
-
-static bool dce112_enable_display_power_gating(
-	struct dc *dc,
-	uint8_t controller_id,
-	struct dc_bios *dcb,
-	enum pipe_gating_control power_gating)
-{
-	enum bp_result bp_result = BP_RESULT_OK;
-	enum bp_pipe_control_action cntl;
-	struct dc_context *ctx = dc->ctx;
-
-	if (IS_FPGA_MAXIMUS_DC(ctx->dce_environment))
-		return true;
-
-	if (power_gating == PIPE_GATING_CONTROL_INIT)
-		cntl = ASIC_PIPE_INIT;
-	else if (power_gating == PIPE_GATING_CONTROL_ENABLE)
-		cntl = ASIC_PIPE_ENABLE;
-	else
-		cntl = ASIC_PIPE_DISABLE;
-
-	if (power_gating != PIPE_GATING_CONTROL_INIT || controller_id == 0){
-
-		bp_result = dcb->funcs->enable_disp_power_gating(
-						dcb, controller_id + 1, cntl);
-
-		/* Revert MASTER_UPDATE_MODE to 0 because bios sets it 2
-		 * by default when command table is called
-		 */
-		dm_write_reg(ctx,
-			HW_REG_CRTC(mmCRTC_MASTER_UPDATE_MODE, controller_id),
-			0);
-	}
-
-	if (power_gating != PIPE_GATING_CONTROL_ENABLE)
-		dce112_init_pte(ctx);
-
-	if (bp_result == BP_RESULT_OK)
-		return true;
-	else
-		return false;
-}
-
-void dce112_hw_sequencer_construct(struct dc *dc)
-{
-	/* All registers used by dce11.2 match those in dce11 in offset and
-	 * structure
-	 */
-	dce110_hw_sequencer_construct(dc);
-	dc->hwseq->funcs.enable_display_power_gating = dce112_enable_display_power_gating;
-}
-
diff --git a/drivers/gpu/drm/amd/display/dc/dce120/Makefile b/drivers/gpu/drm/amd/display/dc/dce120/Makefile
index a9cc4b73270b..8f508e662748 100644
--- a/drivers/gpu/drm/amd/display/dc/dce120/Makefile
+++ b/drivers/gpu/drm/amd/display/dc/dce120/Makefile
@@ -24,10 +24,9 @@
 # It provides the control and status of HW CRTC block.
 
 
-CFLAGS_$(AMDDALPATH)/dc/dce120/dce120_resource.o = $(call cc-disable-warning, override-init)
+CFLAGS_$(AMDDALPATH)/dc/dce120/dce120_resource.o = -Wno-override-init
 
-DCE120 = dce120_resource.o dce120_timing_generator.o \
-dce120_hw_sequencer.o
+DCE120 = dce120_timing_generator.o
 
 AMD_DAL_DCE120 = $(addprefix $(AMDDALPATH)/dc/dce120/,$(DCE120))
 
diff --git a/drivers/gpu/drm/amd/display/dc/dce120/dce120_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/dce120/dce120_hw_sequencer.c
deleted file mode 100644
index d4afe6c824d2..000000000000
--- a/drivers/gpu/drm/amd/display/dc/dce120/dce120_hw_sequencer.c
+++ /dev/null
@@ -1,271 +0,0 @@
-/*
- * Copyright 2015 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: AMD
- *
- */
-
-#include "dm_services.h"
-#include "dc.h"
-#include "core_types.h"
-#include "dce120_hw_sequencer.h"
-#include "dce/dce_hwseq.h"
-
-#include "dce110/dce110_hw_sequencer.h"
-
-#include "dce/dce_12_0_offset.h"
-#include "dce/dce_12_0_sh_mask.h"
-#include "soc15_hw_ip.h"
-#include "vega10_ip_offset.h"
-#include "reg_helper.h"
-
-#define CTX \
-	hws->ctx
-#define REG(reg)\
-	hws->regs->reg
-
-#undef FN
-#define FN(reg_name, field_name) \
-	hws->shifts->field_name, hws->masks->field_name
-
-struct dce120_hw_seq_reg_offsets {
-	uint32_t crtc;
-};
-
-#if 0
-static const struct dce120_hw_seq_reg_offsets reg_offsets[] = {
-{
-	.crtc = (mmCRTC0_CRTC_GSL_CONTROL - mmCRTC0_CRTC_GSL_CONTROL),
-},
-{
-	.crtc = (mmCRTC1_CRTC_GSL_CONTROL - mmCRTC0_CRTC_GSL_CONTROL),
-},
-{
-	.crtc = (mmCRTC2_CRTC_GSL_CONTROL - mmCRTC0_CRTC_GSL_CONTROL),
-},
-{
-	.crtc = (mmCRTC3_CRTC_GSL_CONTROL - mmCRTC0_CRTC_GSL_CONTROL),
-},
-{
-	.crtc = (mmCRTC4_CRTC_GSL_CONTROL - mmCRTC0_CRTC_GSL_CONTROL),
-},
-{
-	.crtc = (mmCRTC5_CRTC_GSL_CONTROL - mmCRTC0_CRTC_GSL_CONTROL),
-}
-};
-
-#define HW_REG_CRTC(reg, id)\
-	(reg + reg_offsets[id].crtc)
-
-#define CNTL_ID(controller_id)\
-	controller_id
-/*******************************************************************************
- * Private definitions
- ******************************************************************************/
-static void dce120_init_pte(struct dc_context *ctx, uint8_t controller_id)
-{
-	uint32_t addr;
-	uint32_t value = 0;
-	uint32_t chunk_int = 0;
-	uint32_t chunk_mul = 0;
-/*
-	addr = mmDCP0_DVMM_PTE_CONTROL + controller_id *
-			(mmDCP1_DVMM_PTE_CONTROL- mmDCP0_DVMM_PTE_CONTROL);
-
-	value = dm_read_reg(ctx, addr);
-
-	set_reg_field_value(
-			value, 0, DCP, controller_id,
-			DVMM_PTE_CONTROL,
-			DVMM_USE_SINGLE_PTE);
-
-	set_reg_field_value_soc15(
-			value, 1, DCP, controller_id,
-			DVMM_PTE_CONTROL,
-			DVMM_PTE_BUFFER_MODE0);
-
-	set_reg_field_value_soc15(
-			value, 1, DCP, controller_id,
-			DVMM_PTE_CONTROL,
-			DVMM_PTE_BUFFER_MODE1);
-
-	dm_write_reg(ctx, addr, value);*/
-
-	addr = mmDVMM_PTE_REQ;
-	value = dm_read_reg(ctx, addr);
-
-	chunk_int = get_reg_field_value(
-		value,
-		DVMM_PTE_REQ,
-		HFLIP_PTEREQ_PER_CHUNK_INT);
-
-	chunk_mul = get_reg_field_value(
-		value,
-		DVMM_PTE_REQ,
-		HFLIP_PTEREQ_PER_CHUNK_MULTIPLIER);
-
-	if (chunk_int != 0x4 || chunk_mul != 0x4) {
-
-		set_reg_field_value(
-			value,
-			255,
-			DVMM_PTE_REQ,
-			MAX_PTEREQ_TO_ISSUE);
-
-		set_reg_field_value(
-			value,
-			4,
-			DVMM_PTE_REQ,
-			HFLIP_PTEREQ_PER_CHUNK_INT);
-
-		set_reg_field_value(
-			value,
-			4,
-			DVMM_PTE_REQ,
-			HFLIP_PTEREQ_PER_CHUNK_MULTIPLIER);
-
-		dm_write_reg(ctx, addr, value);
-	}
-}
-#endif
-
-static bool dce120_enable_display_power_gating(
-	struct dc *dc,
-	uint8_t controller_id,
-	struct dc_bios *dcb,
-	enum pipe_gating_control power_gating)
-{
-	/* disable for bringup */
-#if 0
-	enum bp_result bp_result = BP_RESULT_OK;
-	enum bp_pipe_control_action cntl;
-	struct dc_context *ctx = dc->ctx;
-
-	if (IS_FPGA_MAXIMUS_DC(ctx->dce_environment))
-		return true;
-
-	if (power_gating == PIPE_GATING_CONTROL_INIT)
-		cntl = ASIC_PIPE_INIT;
-	else if (power_gating == PIPE_GATING_CONTROL_ENABLE)
-		cntl = ASIC_PIPE_ENABLE;
-	else
-		cntl = ASIC_PIPE_DISABLE;
-
-	if (power_gating != PIPE_GATING_CONTROL_INIT || controller_id == 0) {
-
-		bp_result = dcb->funcs->enable_disp_power_gating(
-						dcb, controller_id + 1, cntl);
-
-		/* Revert MASTER_UPDATE_MODE to 0 because bios sets it 2
-		 * by default when command table is called
-		 */
-		dm_write_reg(ctx,
-			HW_REG_CRTC(mmCRTC0_CRTC_MASTER_UPDATE_MODE, controller_id),
-			0);
-	}
-
-	if (power_gating != PIPE_GATING_CONTROL_ENABLE)
-		dce120_init_pte(ctx, controller_id);
-
-	if (bp_result == BP_RESULT_OK)
-		return true;
-	else
-		return false;
-#endif
-	return false;
-}
-
-static void dce120_update_dchub(
-	struct dce_hwseq *hws,
-	struct dchub_init_data *dh_data)
-{
-	/* TODO: port code from dal2 */
-	switch (dh_data->fb_mode) {
-	case FRAME_BUFFER_MODE_ZFB_ONLY:
-		/*For ZFB case need to put DCHUB FB BASE and TOP upside down to indicate ZFB mode*/
-		REG_UPDATE_2(DCHUB_FB_LOCATION,
-				FB_TOP, 0,
-				FB_BASE, 0x0FFFF);
-
-		REG_UPDATE(DCHUB_AGP_BASE,
-				AGP_BASE, dh_data->zfb_phys_addr_base >> 22);
-
-		REG_UPDATE(DCHUB_AGP_BOT,
-				AGP_BOT, dh_data->zfb_mc_base_addr >> 22);
-
-		REG_UPDATE(DCHUB_AGP_TOP,
-				AGP_TOP, (dh_data->zfb_mc_base_addr + dh_data->zfb_size_in_byte - 1) >> 22);
-		break;
-	case FRAME_BUFFER_MODE_MIXED_ZFB_AND_LOCAL:
-		/*Should not touch FB LOCATION (done by VBIOS on AsicInit table)*/
-		REG_UPDATE(DCHUB_AGP_BASE,
-				AGP_BASE, dh_data->zfb_phys_addr_base >> 22);
-
-		REG_UPDATE(DCHUB_AGP_BOT,
-				AGP_BOT, dh_data->zfb_mc_base_addr >> 22);
-
-		REG_UPDATE(DCHUB_AGP_TOP,
-				AGP_TOP, (dh_data->zfb_mc_base_addr + dh_data->zfb_size_in_byte - 1) >> 22);
-		break;
-	case FRAME_BUFFER_MODE_LOCAL_ONLY:
-		/*Should not touch FB LOCATION (done by VBIOS on AsicInit table)*/
-		REG_UPDATE(DCHUB_AGP_BASE,
-				AGP_BASE, 0);
-
-		REG_UPDATE(DCHUB_AGP_BOT,
-				AGP_BOT, 0x03FFFF);
-
-		REG_UPDATE(DCHUB_AGP_TOP,
-				AGP_TOP, 0);
-		break;
-	default:
-		break;
-	}
-
-	dh_data->dchub_initialzied = true;
-	dh_data->dchub_info_valid = false;
-}
-
-/**
- * dce121_xgmi_enabled() - Check if xGMI is enabled
- * @hws: DCE hardware sequencer object
- *
- * Return true if xGMI is enabled. False otherwise.
- */
-bool dce121_xgmi_enabled(struct dce_hwseq *hws)
-{
-	uint32_t pf_max_region;
-
-	REG_GET(MC_VM_XGMI_LFB_CNTL, PF_MAX_REGION, &pf_max_region);
-	/* PF_MAX_REGION == 0 means xgmi is disabled */
-	return !!pf_max_region;
-}
-
-void dce120_hw_sequencer_construct(struct dc *dc)
-{
-	/* All registers used by dce11.2 match those in dce11 in offset and
-	 * structure
-	 */
-	dce110_hw_sequencer_construct(dc);
-	dc->hwseq->funcs.enable_display_power_gating = dce120_enable_display_power_gating;
-	dc->hwss.update_dchub = dce120_update_dchub;
-}
-
diff --git a/drivers/gpu/drm/amd/display/dc/dce120/dce120_timing_generator.c b/drivers/gpu/drm/amd/display/dc/dce120/dce120_timing_generator.c
index 4af0c70098c4..31c4f44ceaac 100644
--- a/drivers/gpu/drm/amd/display/dc/dce120/dce120_timing_generator.c
+++ b/drivers/gpu/drm/amd/display/dc/dce120/dce120_timing_generator.c
@@ -697,6 +697,7 @@ static void dce120_tg_program_timing(struct timing_generator *tg,
 	int vstartup_start,
 	int vupdate_offset,
 	int vupdate_width,
+	int pstate_keepout,
 	const enum signal_type signal,
 	bool use_vbios)
 {
@@ -1099,45 +1100,79 @@ static bool dce120_configure_crc(struct timing_generator *tg,
 	if (!dce120_is_tg_enabled(tg))
 		return false;
 
-	/* First, disable CRC before we configure it. */
-	dm_write_reg_soc15(tg->ctx, mmCRTC0_CRTC_CRC_CNTL,
-			   tg110->offsets.crtc, 0);
+	if (!params->enable || params->reset)
+		/* First, disable CRC before we configure it. */
+		dm_write_reg_soc15(tg->ctx, mmCRTC0_CRTC_CRC_CNTL,
+				   tg110->offsets.crtc, 0);
 
 	if (!params->enable)
 		return true;
 
 	/* Program frame boundaries */
-	/* Window A x axis start and end. */
-	CRTC_REG_UPDATE_2(CRTC0_CRTC_CRC0_WINDOWA_X_CONTROL,
-			  CRTC_CRC0_WINDOWA_X_START, params->windowa_x_start,
-			  CRTC_CRC0_WINDOWA_X_END, params->windowa_x_end);
-
-	/* Window A y axis start and end. */
-	CRTC_REG_UPDATE_2(CRTC0_CRTC_CRC0_WINDOWA_Y_CONTROL,
-			  CRTC_CRC0_WINDOWA_Y_START, params->windowa_y_start,
-			  CRTC_CRC0_WINDOWA_Y_END, params->windowa_y_end);
-
-	/* Window B x axis start and end. */
-	CRTC_REG_UPDATE_2(CRTC0_CRTC_CRC0_WINDOWB_X_CONTROL,
-			  CRTC_CRC0_WINDOWB_X_START, params->windowb_x_start,
-			  CRTC_CRC0_WINDOWB_X_END, params->windowb_x_end);
-
-	/* Window B y axis start and end. */
-	CRTC_REG_UPDATE_2(CRTC0_CRTC_CRC0_WINDOWB_Y_CONTROL,
-			  CRTC_CRC0_WINDOWB_Y_START, params->windowb_y_start,
-			  CRTC_CRC0_WINDOWB_Y_END, params->windowb_y_end);
-
-	/* Set crc mode and selection, and enable. Only using CRC0*/
-	CRTC_REG_UPDATE_3(CRTC0_CRTC_CRC_CNTL,
-			  CRTC_CRC_EN, params->continuous_mode ? 1 : 0,
-			  CRTC_CRC0_SELECT, params->selection,
-			  CRTC_CRC_EN, 1);
+	switch (params->crc_eng_inst) {
+	case 0:
+		/* Window A x axis start and end. */
+		CRTC_REG_UPDATE_2(CRTC0_CRTC_CRC0_WINDOWA_X_CONTROL,
+				  CRTC_CRC0_WINDOWA_X_START, params->windowa_x_start,
+				  CRTC_CRC0_WINDOWA_X_END, params->windowa_x_end);
+
+		/* Window A y axis start and end. */
+		CRTC_REG_UPDATE_2(CRTC0_CRTC_CRC0_WINDOWA_Y_CONTROL,
+				  CRTC_CRC0_WINDOWA_Y_START, params->windowa_y_start,
+				  CRTC_CRC0_WINDOWA_Y_END, params->windowa_y_end);
+
+		/* Window B x axis start and end. */
+		CRTC_REG_UPDATE_2(CRTC0_CRTC_CRC0_WINDOWB_X_CONTROL,
+				  CRTC_CRC0_WINDOWB_X_START, params->windowb_x_start,
+				  CRTC_CRC0_WINDOWB_X_END, params->windowb_x_end);
+
+		/* Window B y axis start and end. */
+		CRTC_REG_UPDATE_2(CRTC0_CRTC_CRC0_WINDOWB_Y_CONTROL,
+				  CRTC_CRC0_WINDOWB_Y_START, params->windowb_y_start,
+				  CRTC_CRC0_WINDOWB_Y_END, params->windowb_y_end);
+
+		/* Set crc mode and selection, and enable.*/
+		CRTC_REG_UPDATE_3(CRTC0_CRTC_CRC_CNTL,
+				  CRTC_CRC_CONT_EN, params->continuous_mode ? 1 : 0,
+				  CRTC_CRC0_SELECT, params->selection,
+				  CRTC_CRC_EN, 1);
+		break;
+	case 1:
+		/* Window A x axis start and end. */
+		CRTC_REG_UPDATE_2(CRTC0_CRTC_CRC1_WINDOWA_X_CONTROL,
+				  CRTC_CRC1_WINDOWA_X_START, params->windowa_x_start,
+				  CRTC_CRC1_WINDOWA_X_END, params->windowa_x_end);
+
+		/* Window A y axis start and end. */
+		CRTC_REG_UPDATE_2(CRTC0_CRTC_CRC1_WINDOWA_Y_CONTROL,
+				  CRTC_CRC1_WINDOWA_Y_START, params->windowa_y_start,
+				  CRTC_CRC1_WINDOWA_Y_END, params->windowa_y_end);
+
+		/* Window B x axis start and end. */
+		CRTC_REG_UPDATE_2(CRTC0_CRTC_CRC1_WINDOWB_X_CONTROL,
+				  CRTC_CRC1_WINDOWB_X_START, params->windowb_x_start,
+				  CRTC_CRC1_WINDOWB_X_END, params->windowb_x_end);
+
+		/* Window B y axis start and end. */
+		CRTC_REG_UPDATE_2(CRTC0_CRTC_CRC1_WINDOWB_Y_CONTROL,
+				  CRTC_CRC1_WINDOWB_Y_START, params->windowb_y_start,
+				  CRTC_CRC1_WINDOWB_Y_END, params->windowb_y_end);
+
+		/* Set crc mode and selection, and enable */
+		CRTC_REG_UPDATE_3(CRTC0_CRTC_CRC_CNTL,
+				  CRTC_CRC_CONT_EN, params->continuous_mode ? 1 : 0,
+				  CRTC_CRC1_SELECT, params->selection,
+				  CRTC_CRC_EN, 1);
+		break;
+	default:
+		return false;
+	}
 
 	return true;
 }
 
-static bool dce120_get_crc(struct timing_generator *tg, uint32_t *r_cr,
-			   uint32_t *g_y, uint32_t *b_cb)
+static bool dce120_get_crc(struct timing_generator *tg, uint8_t idx,
+			uint32_t *r_cr, uint32_t *g_y, uint32_t *b_cb)
 {
 	struct dce110_timing_generator *tg110 = DCE110TG_FROM_TG(tg);
 	uint32_t value, field;
@@ -1150,14 +1185,30 @@ static bool dce120_get_crc(struct timing_generator *tg, uint32_t *r_cr,
 	if (!field)
 		return false;
 
-	value = dm_read_reg_soc15(tg->ctx, mmCRTC0_CRTC_CRC0_DATA_RG,
-				  tg110->offsets.crtc);
-	*r_cr = get_reg_field_value(value, CRTC0_CRTC_CRC0_DATA_RG, CRC0_R_CR);
-	*g_y = get_reg_field_value(value, CRTC0_CRTC_CRC0_DATA_RG, CRC0_G_Y);
+	switch (idx) {
+	case 0:
+		value = dm_read_reg_soc15(tg->ctx, mmCRTC0_CRTC_CRC0_DATA_RG,
+					  tg110->offsets.crtc);
+		*r_cr = get_reg_field_value(value, CRTC0_CRTC_CRC0_DATA_RG, CRC0_R_CR);
+		*g_y = get_reg_field_value(value, CRTC0_CRTC_CRC0_DATA_RG, CRC0_G_Y);
 
-	value = dm_read_reg_soc15(tg->ctx, mmCRTC0_CRTC_CRC0_DATA_B,
-				  tg110->offsets.crtc);
-	*b_cb = get_reg_field_value(value, CRTC0_CRTC_CRC0_DATA_B, CRC0_B_CB);
+		value = dm_read_reg_soc15(tg->ctx, mmCRTC0_CRTC_CRC0_DATA_B,
+					  tg110->offsets.crtc);
+		*b_cb = get_reg_field_value(value, CRTC0_CRTC_CRC0_DATA_B, CRC0_B_CB);
+		break;
+	case 1:
+		value = dm_read_reg_soc15(tg->ctx, mmCRTC0_CRTC_CRC1_DATA_RG,
+					  tg110->offsets.crtc);
+		*r_cr = get_reg_field_value(value, CRTC0_CRTC_CRC1_DATA_RG, CRC1_R_CR);
+		*g_y = get_reg_field_value(value, CRTC0_CRTC_CRC1_DATA_RG, CRC1_G_Y);
+
+		value = dm_read_reg_soc15(tg->ctx, mmCRTC0_CRTC_CRC1_DATA_B,
+					  tg110->offsets.crtc);
+		*b_cb = get_reg_field_value(value, CRTC0_CRTC_CRC1_DATA_B, CRC1_B_CB);
+		break;
+	default:
+		return false;
+	}
 
 	return true;
 }
@@ -1197,6 +1248,7 @@ static const struct timing_generator_funcs dce120_tg_funcs = {
 		.is_tg_enabled = dce120_is_tg_enabled,
 		.configure_crc = dce120_configure_crc,
 		.get_crc = dce120_get_crc,
+		.is_two_pixels_per_container = dce110_is_two_pixels_per_container,
 };
 
 
diff --git a/drivers/gpu/drm/amd/display/dc/dce60/Makefile b/drivers/gpu/drm/amd/display/dc/dce60/Makefile
index dda596fa1cd7..824f73eb3326 100644
--- a/drivers/gpu/drm/amd/display/dc/dce60/Makefile
+++ b/drivers/gpu/drm/amd/display/dc/dce60/Makefile
@@ -23,10 +23,9 @@
 # Makefile for the 'controller' sub-component of DAL.
 # It provides the control and status of HW CRTC block.
 
-CFLAGS_AMDDALPATH)/dc/dce60/dce60_resource.o = $(call cc-disable-warning, override-init)
+CFLAGS_$(AMDDALPATH)/dc/dce60/dce60_resource.o = -Wno-override-init
 
-DCE60 = dce60_timing_generator.o dce60_hw_sequencer.o \
-	dce60_resource.o
+DCE60 = dce60_timing_generator.o
 
 AMD_DAL_DCE60 = $(addprefix $(AMDDALPATH)/dc/dce60/,$(DCE60))
 
diff --git a/drivers/gpu/drm/amd/display/dc/dce60/dce60_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/dce60/dce60_hw_sequencer.c
deleted file mode 100644
index 920c7ae29d53..000000000000
--- a/drivers/gpu/drm/amd/display/dc/dce60/dce60_hw_sequencer.c
+++ /dev/null
@@ -1,432 +0,0 @@
-/*
- * Copyright 2020 Mauro Rossi <issor.oruam@gmail.com>
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: AMD
- *
- */
-
-#include "dm_services.h"
-#include "dc.h"
-#include "core_types.h"
-#include "dce60_hw_sequencer.h"
-
-#include "dce/dce_hwseq.h"
-#include "dce110/dce110_hw_sequencer.h"
-#include "dce100/dce100_hw_sequencer.h"
-
-/* include DCE6 register header files */
-#include "dce/dce_6_0_d.h"
-#include "dce/dce_6_0_sh_mask.h"
-
-#define DC_LOGGER_INIT()
-
-/*******************************************************************************
- * Private definitions
- ******************************************************************************/
-
-/***************************PIPE_CONTROL***********************************/
-
-/*
- *  Check if FBC can be enabled
- */
-static bool dce60_should_enable_fbc(struct dc *dc,
-		struct dc_state *context,
-		uint32_t *pipe_idx)
-{
-	uint32_t i;
-	struct pipe_ctx *pipe_ctx = NULL;
-	struct resource_context *res_ctx = &context->res_ctx;
-	unsigned int underlay_idx = dc->res_pool->underlay_pipe_index;
-
-
-	ASSERT(dc->fbc_compressor);
-
-	/* FBC memory should be allocated */
-	if (!dc->ctx->fbc_gpu_addr)
-		return false;
-
-	/* Only supports single display */
-	if (context->stream_count != 1)
-		return false;
-
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		if (res_ctx->pipe_ctx[i].stream) {
-
-			pipe_ctx = &res_ctx->pipe_ctx[i];
-
-			if (!pipe_ctx)
-				continue;
-
-			/* fbc not applicable on underlay pipe */
-			if (pipe_ctx->pipe_idx != underlay_idx) {
-				*pipe_idx = i;
-				break;
-			}
-		}
-	}
-
-	if (i == dc->res_pool->pipe_count)
-		return false;
-
-	if (!pipe_ctx->stream->link)
-		return false;
-
-	/* Only supports eDP */
-	if (pipe_ctx->stream->link->connector_signal != SIGNAL_TYPE_EDP)
-		return false;
-
-	/* PSR should not be enabled */
-	if (pipe_ctx->stream->link->psr_settings.psr_feature_enabled)
-		return false;
-
-	/* Nothing to compress */
-	if (!pipe_ctx->plane_state)
-		return false;
-
-	/* Only for non-linear tiling */
-	if (pipe_ctx->plane_state->tiling_info.gfx8.array_mode == DC_ARRAY_LINEAR_GENERAL)
-		return false;
-
-	return true;
-}
-
-/*
- *  Enable FBC
- */
-static void dce60_enable_fbc(
-		struct dc *dc,
-		struct dc_state *context)
-{
-	uint32_t pipe_idx = 0;
-
-	if (dce60_should_enable_fbc(dc, context, &pipe_idx)) {
-		/* Program GRPH COMPRESSED ADDRESS and PITCH */
-		struct compr_addr_and_pitch_params params = {0, 0, 0};
-		struct compressor *compr = dc->fbc_compressor;
-		struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[pipe_idx];
-
-		params.source_view_width = pipe_ctx->stream->timing.h_addressable;
-		params.source_view_height = pipe_ctx->stream->timing.v_addressable;
-		params.inst = pipe_ctx->stream_res.tg->inst;
-		compr->compr_surface_address.quad_part = dc->ctx->fbc_gpu_addr;
-
-		compr->funcs->surface_address_and_pitch(compr, &params);
-		compr->funcs->set_fbc_invalidation_triggers(compr, 1);
-
-		compr->funcs->enable_fbc(compr, &params);
-	}
-}
-
-
-/*******************************************************************************
- * Front End programming
- ******************************************************************************/
-
-static void dce60_set_default_colors(struct pipe_ctx *pipe_ctx)
-{
-	struct default_adjustment default_adjust = { 0 };
-
-	default_adjust.force_hw_default = false;
-	default_adjust.in_color_space = pipe_ctx->plane_state->color_space;
-	default_adjust.out_color_space = pipe_ctx->stream->output_color_space;
-	default_adjust.csc_adjust_type = GRAPHICS_CSC_ADJUST_TYPE_SW;
-	default_adjust.surface_pixel_format = pipe_ctx->plane_res.scl_data.format;
-
-	/* display color depth */
-	default_adjust.color_depth =
-		pipe_ctx->stream->timing.display_color_depth;
-
-	/* Lb color depth */
-	default_adjust.lb_color_depth = pipe_ctx->plane_res.scl_data.lb_params.depth;
-
-	pipe_ctx->plane_res.xfm->funcs->opp_set_csc_default(
-					pipe_ctx->plane_res.xfm, &default_adjust);
-}
-
-/*******************************************************************************
- * In order to turn on surface we will program
- * CRTC
- *
- * DCE6 has no bottom_pipe and no Blender HW
- * We need to set 'blank_target' to false in order to turn on the display
- *
- * |-----------|------------|---------|
- * |curr pipe  | set_blank  |         |
- * |Surface    |blank_target|  CRCT   |
- * |visibility |  argument  |         |
- * |-----------|------------|---------|
- * |    off    |   true     | blank   |
- * |    on     |   false    | unblank |
- * |-----------|------------|---------|
- *
- ******************************************************************************/
-static void dce60_program_surface_visibility(const struct dc *dc,
-		struct pipe_ctx *pipe_ctx)
-{
-	bool blank_target = false;
-
-	/* DCE6 has no bottom_pipe and no Blender HW */
-
-	if (!pipe_ctx->plane_state->visible)
-		blank_target = true;
-
-	/* DCE6 skip dce_set_blender_mode() but then proceed to 'unblank' CRTC */
-	pipe_ctx->stream_res.tg->funcs->set_blank(pipe_ctx->stream_res.tg, blank_target);
-
-}
-
-
-static void dce60_get_surface_visual_confirm_color(const struct pipe_ctx *pipe_ctx,
-		struct tg_color *color)
-{
-	uint32_t color_value = MAX_TG_COLOR_VALUE * (4 - pipe_ctx->stream_res.tg->inst) / 4;
-
-	switch (pipe_ctx->plane_res.scl_data.format) {
-	case PIXEL_FORMAT_ARGB8888:
-		/* set boarder color to red */
-		color->color_r_cr = color_value;
-		break;
-
-	case PIXEL_FORMAT_ARGB2101010:
-		/* set boarder color to blue */
-		color->color_b_cb = color_value;
-		break;
-	case PIXEL_FORMAT_420BPP8:
-		/* set boarder color to green */
-		color->color_g_y = color_value;
-		break;
-	case PIXEL_FORMAT_420BPP10:
-		/* set boarder color to yellow */
-		color->color_g_y = color_value;
-		color->color_r_cr = color_value;
-		break;
-	case PIXEL_FORMAT_FP16:
-		/* set boarder color to white */
-		color->color_r_cr = color_value;
-		color->color_b_cb = color_value;
-		color->color_g_y = color_value;
-		break;
-	default:
-		break;
-	}
-}
-
-static void dce60_program_scaler(const struct dc *dc,
-		const struct pipe_ctx *pipe_ctx)
-{
-	struct tg_color color = {0};
-
-	/* DCE6 skips DCN TOFPGA check for transform_set_pixel_storage_depth == NULL */
-
-	if (dc->debug.visual_confirm == VISUAL_CONFIRM_SURFACE)
-		dce60_get_surface_visual_confirm_color(pipe_ctx, &color);
-	else
-		color_space_to_black_color(dc,
-				pipe_ctx->stream->output_color_space,
-				&color);
-
-	pipe_ctx->plane_res.xfm->funcs->transform_set_pixel_storage_depth(
-		pipe_ctx->plane_res.xfm,
-		pipe_ctx->plane_res.scl_data.lb_params.depth,
-		&pipe_ctx->stream->bit_depth_params);
-
-	if (pipe_ctx->stream_res.tg->funcs->set_overscan_blank_color) {
-		/*
-		 * The way 420 is packed, 2 channels carry Y component, 1 channel
-		 * alternate between Cb and Cr, so both channels need the pixel
-		 * value for Y
-		 */
-		if (pipe_ctx->stream->timing.pixel_encoding == PIXEL_ENCODING_YCBCR420)
-			color.color_r_cr = color.color_g_y;
-
-		pipe_ctx->stream_res.tg->funcs->set_overscan_blank_color(
-				pipe_ctx->stream_res.tg,
-				&color);
-	}
-
-	pipe_ctx->plane_res.xfm->funcs->transform_set_scaler(pipe_ctx->plane_res.xfm,
-		&pipe_ctx->plane_res.scl_data);
-}
-
-static void
-dce60_program_front_end_for_pipe(
-		struct dc *dc, struct pipe_ctx *pipe_ctx)
-{
-	struct mem_input *mi = pipe_ctx->plane_res.mi;
-	struct dc_plane_state *plane_state = pipe_ctx->plane_state;
-	struct xfm_grph_csc_adjustment adjust;
-	struct out_csc_color_matrix tbl_entry;
-	unsigned int i;
-	struct dce_hwseq *hws = dc->hwseq;
-
-	DC_LOGGER_INIT();
-	memset(&tbl_entry, 0, sizeof(tbl_entry));
-
-	memset(&adjust, 0, sizeof(adjust));
-	adjust.gamut_adjust_type = GRAPHICS_GAMUT_ADJUST_TYPE_BYPASS;
-
-	dce_enable_fe_clock(dc->hwseq, mi->inst, true);
-
-	dce60_set_default_colors(pipe_ctx);
-	if (pipe_ctx->stream->csc_color_matrix.enable_adjustment
-			== true) {
-		tbl_entry.color_space =
-			pipe_ctx->stream->output_color_space;
-
-		for (i = 0; i < 12; i++)
-			tbl_entry.regval[i] =
-			pipe_ctx->stream->csc_color_matrix.matrix[i];
-
-		pipe_ctx->plane_res.xfm->funcs->opp_set_csc_adjustment
-				(pipe_ctx->plane_res.xfm, &tbl_entry);
-	}
-
-	if (pipe_ctx->stream->gamut_remap_matrix.enable_remap == true) {
-		adjust.gamut_adjust_type = GRAPHICS_GAMUT_ADJUST_TYPE_SW;
-
-		for (i = 0; i < CSC_TEMPERATURE_MATRIX_SIZE; i++)
-			adjust.temperature_matrix[i] =
-				pipe_ctx->stream->gamut_remap_matrix.matrix[i];
-	}
-
-	pipe_ctx->plane_res.xfm->funcs->transform_set_gamut_remap(pipe_ctx->plane_res.xfm, &adjust);
-
-	pipe_ctx->plane_res.scl_data.lb_params.alpha_en = pipe_ctx->bottom_pipe != 0;
-
-	dce60_program_scaler(dc, pipe_ctx);
-
-	mi->funcs->mem_input_program_surface_config(
-			mi,
-			plane_state->format,
-			&plane_state->tiling_info,
-			&plane_state->plane_size,
-			plane_state->rotation,
-			NULL,
-			false);
-	if (mi->funcs->set_blank)
-		mi->funcs->set_blank(mi, pipe_ctx->plane_state->visible);
-
-	if (dc->config.gpu_vm_support)
-		mi->funcs->mem_input_program_pte_vm(
-				pipe_ctx->plane_res.mi,
-				plane_state->format,
-				&plane_state->tiling_info,
-				plane_state->rotation);
-
-	/* Moved programming gamma from dc to hwss */
-	if (pipe_ctx->plane_state->update_flags.bits.full_update ||
-			pipe_ctx->plane_state->update_flags.bits.in_transfer_func_change ||
-			pipe_ctx->plane_state->update_flags.bits.gamma_change)
-		hws->funcs.set_input_transfer_func(dc, pipe_ctx, pipe_ctx->plane_state);
-
-	if (pipe_ctx->plane_state->update_flags.bits.full_update)
-		hws->funcs.set_output_transfer_func(dc, pipe_ctx, pipe_ctx->stream);
-
-	DC_LOG_SURFACE(
-			"Pipe:%d %p: addr hi:0x%x, "
-			"addr low:0x%x, "
-			"src: %d, %d, %d,"
-			" %d; dst: %d, %d, %d, %d;"
-			"clip: %d, %d, %d, %d\n",
-			pipe_ctx->pipe_idx,
-			(void *) pipe_ctx->plane_state,
-			pipe_ctx->plane_state->address.grph.addr.high_part,
-			pipe_ctx->plane_state->address.grph.addr.low_part,
-			pipe_ctx->plane_state->src_rect.x,
-			pipe_ctx->plane_state->src_rect.y,
-			pipe_ctx->plane_state->src_rect.width,
-			pipe_ctx->plane_state->src_rect.height,
-			pipe_ctx->plane_state->dst_rect.x,
-			pipe_ctx->plane_state->dst_rect.y,
-			pipe_ctx->plane_state->dst_rect.width,
-			pipe_ctx->plane_state->dst_rect.height,
-			pipe_ctx->plane_state->clip_rect.x,
-			pipe_ctx->plane_state->clip_rect.y,
-			pipe_ctx->plane_state->clip_rect.width,
-			pipe_ctx->plane_state->clip_rect.height);
-
-	DC_LOG_SURFACE(
-			"Pipe %d: width, height, x, y\n"
-			"viewport:%d, %d, %d, %d\n"
-			"recout:  %d, %d, %d, %d\n",
-			pipe_ctx->pipe_idx,
-			pipe_ctx->plane_res.scl_data.viewport.width,
-			pipe_ctx->plane_res.scl_data.viewport.height,
-			pipe_ctx->plane_res.scl_data.viewport.x,
-			pipe_ctx->plane_res.scl_data.viewport.y,
-			pipe_ctx->plane_res.scl_data.recout.width,
-			pipe_ctx->plane_res.scl_data.recout.height,
-			pipe_ctx->plane_res.scl_data.recout.x,
-			pipe_ctx->plane_res.scl_data.recout.y);
-}
-
-static void dce60_apply_ctx_for_surface(
-		struct dc *dc,
-		const struct dc_stream_state *stream,
-		int num_planes,
-		struct dc_state *context)
-{
-	int i;
-
-	if (num_planes == 0)
-		return;
-
-	if (dc->fbc_compressor)
-		dc->fbc_compressor->funcs->disable_fbc(dc->fbc_compressor);
-
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i];
-
-		if (pipe_ctx->stream != stream)
-			continue;
-
-		/* Need to allocate mem before program front end for Fiji */
-		pipe_ctx->plane_res.mi->funcs->allocate_mem_input(
-				pipe_ctx->plane_res.mi,
-				pipe_ctx->stream->timing.h_total,
-				pipe_ctx->stream->timing.v_total,
-				pipe_ctx->stream->timing.pix_clk_100hz / 10,
-				context->stream_count);
-
-		dce60_program_front_end_for_pipe(dc, pipe_ctx);
-
-		dc->hwss.update_plane_addr(dc, pipe_ctx);
-
-		dce60_program_surface_visibility(dc, pipe_ctx);
-
-	}
-
-	if (dc->fbc_compressor)
-		dce60_enable_fbc(dc, context);
-}
-
-void dce60_hw_sequencer_construct(struct dc *dc)
-{
-	dce110_hw_sequencer_construct(dc);
-
-	dc->hwseq->funcs.enable_display_power_gating = dce100_enable_display_power_gating;
-	dc->hwss.apply_ctx_for_surface = dce60_apply_ctx_for_surface;
-	dc->hwss.cursor_lock = dce60_pipe_control_lock;
-	dc->hwss.pipe_control_lock = dce60_pipe_control_lock;
-	dc->hwss.prepare_bandwidth = dce100_prepare_bandwidth;
-	dc->hwss.optimize_bandwidth = dce100_optimize_bandwidth;
-}
-
diff --git a/drivers/gpu/drm/amd/display/dc/dce60/dce60_timing_generator.c b/drivers/gpu/drm/amd/display/dc/dce60/dce60_timing_generator.c
index c1a85ee374d9..e691a1cf3356 100644
--- a/drivers/gpu/drm/amd/display/dc/dce60/dce60_timing_generator.c
+++ b/drivers/gpu/drm/amd/display/dc/dce60/dce60_timing_generator.c
@@ -111,13 +111,14 @@ static void program_timing(struct timing_generator *tg,
 	int vstartup_start,
 	int vupdate_offset,
 	int vupdate_width,
+	int pstate_keepout,
 	const enum signal_type signal,
 	bool use_vbios)
 {
 	if (!use_vbios)
 		program_pix_dur(tg, timing->pix_clk_100hz);
 
-	dce110_tg_program_timing(tg, timing, 0, 0, 0, 0, 0, use_vbios);
+	dce110_tg_program_timing(tg, timing, 0, 0, 0, 0, 0, 0, use_vbios);
 }
 
 static void dce60_timing_generator_enable_advanced_request(
@@ -238,6 +239,7 @@ static const struct timing_generator_funcs dce60_tg_funcs = {
 				dce60_timing_generator_enable_advanced_request,
 		.configure_crc = dce60_configure_crc,
 		.get_crc = dce110_get_crc,
+		.is_two_pixels_per_container = dce110_is_two_pixels_per_container,
 };
 
 void dce60_timing_generator_construct(
diff --git a/drivers/gpu/drm/amd/display/dc/dce80/Makefile b/drivers/gpu/drm/amd/display/dc/dce80/Makefile
index 0a9d1a350d8b..fba189d26652 100644
--- a/drivers/gpu/drm/amd/display/dc/dce80/Makefile
+++ b/drivers/gpu/drm/amd/display/dc/dce80/Makefile
@@ -23,10 +23,9 @@
 # Makefile for the 'controller' sub-component of DAL.
 # It provides the control and status of HW CRTC block.
 
-CFLAGS_$(AMDDALPATH)/dc/dce80/dce80_resource.o = $(call cc-disable-warning, override-init)
+CFLAGS_$(AMDDALPATH)/dc/dce80/dce80_resource.o = -Wno-override-init
 
-DCE80 = dce80_timing_generator.o dce80_hw_sequencer.o \
-	dce80_resource.o
+DCE80 = dce80_timing_generator.o
 
 AMD_DAL_DCE80 = $(addprefix $(AMDDALPATH)/dc/dce80/,$(DCE80))
 
diff --git a/drivers/gpu/drm/amd/display/dc/dce80/dce80_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/dce80/dce80_hw_sequencer.c
deleted file mode 100644
index d2ceebdbdf51..000000000000
--- a/drivers/gpu/drm/amd/display/dc/dce80/dce80_hw_sequencer.c
+++ /dev/null
@@ -1,54 +0,0 @@
-/*
- * Copyright 2015 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: AMD
- *
- */
-
-#include "dm_services.h"
-#include "dc.h"
-#include "core_types.h"
-#include "dce80_hw_sequencer.h"
-
-#include "dce/dce_hwseq.h"
-#include "dce110/dce110_hw_sequencer.h"
-#include "dce100/dce100_hw_sequencer.h"
-
-/* include DCE8 register header files */
-#include "dce/dce_8_0_d.h"
-#include "dce/dce_8_0_sh_mask.h"
-
-/*******************************************************************************
- * Private definitions
- ******************************************************************************/
-
-/***************************PIPE_CONTROL***********************************/
-
-void dce80_hw_sequencer_construct(struct dc *dc)
-{
-	dce110_hw_sequencer_construct(dc);
-
-	dc->hwseq->funcs.enable_display_power_gating = dce100_enable_display_power_gating;
-	dc->hwss.pipe_control_lock = dce_pipe_control_lock;
-	dc->hwss.prepare_bandwidth = dce100_prepare_bandwidth;
-	dc->hwss.optimize_bandwidth = dce100_optimize_bandwidth;
-}
-
diff --git a/drivers/gpu/drm/amd/display/dc/dce80/dce80_timing_generator.c b/drivers/gpu/drm/amd/display/dc/dce80/dce80_timing_generator.c
index b8fd43dc010b..88e7a1fc9a30 100644
--- a/drivers/gpu/drm/amd/display/dc/dce80/dce80_timing_generator.c
+++ b/drivers/gpu/drm/amd/display/dc/dce80/dce80_timing_generator.c
@@ -105,19 +105,20 @@ static void program_pix_dur(struct timing_generator *tg, uint32_t pix_clk_100hz)
 	dm_write_reg(tg->ctx, addr, value);
 }
 
-static void program_timing(struct timing_generator *tg,
+static void dce80_timing_generator_program_timing(struct timing_generator *tg,
 	const struct dc_crtc_timing *timing,
 	int vready_offset,
 	int vstartup_start,
 	int vupdate_offset,
 	int vupdate_width,
+	int pstate_keepout,
 	const enum signal_type signal,
 	bool use_vbios)
 {
 	if (!use_vbios)
 		program_pix_dur(tg, timing->pix_clk_100hz);
 
-	dce110_tg_program_timing(tg, timing, 0, 0, 0, 0, 0, use_vbios);
+	dce110_tg_program_timing(tg, timing, 0, 0, 0, 0, 0, 0, use_vbios);
 }
 
 static void dce80_timing_generator_enable_advanced_request(
@@ -184,7 +185,7 @@ static void dce80_timing_generator_enable_advanced_request(
 
 static const struct timing_generator_funcs dce80_tg_funcs = {
 		.validate_timing = dce110_tg_validate_timing,
-		.program_timing = program_timing,
+		.program_timing = dce80_timing_generator_program_timing,
 		.enable_crtc = dce110_timing_generator_enable_crtc,
 		.disable_crtc = dce110_timing_generator_disable_crtc,
 		.is_counter_moving = dce110_timing_generator_is_counter_moving,
@@ -220,6 +221,7 @@ static const struct timing_generator_funcs dce80_tg_funcs = {
 				dce80_timing_generator_enable_advanced_request,
 		.configure_crc = dce110_configure_crc,
 		.get_crc = dce110_get_crc,
+		.is_two_pixels_per_container = dce110_is_two_pixels_per_container,
 };
 
 void dce80_timing_generator_construct(
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/Makefile b/drivers/gpu/drm/amd/display/dc/dcn10/Makefile
index 62ad1a11bff9..e1f6623d4936 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/Makefile
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/Makefile
@@ -22,12 +22,9 @@
 #
 # Makefile for DCN.
 
-DCN10 = dcn10_init.o dcn10_resource.o dcn10_ipp.o dcn10_hw_sequencer.o \
+DCN10 = dcn10_ipp.o \
 		dcn10_hw_sequencer_debug.o \
-		dcn10_dpp.o dcn10_opp.o dcn10_optc.o \
-		dcn10_hubp.o dcn10_mpc.o \
-		dcn10_dpp_dscl.o dcn10_dpp_cm.o dcn10_cm_common.o \
-		dcn10_hubbub.o dcn10_stream_encoder.o dcn10_link_encoder.o
+		dcn10_cm_common.o \
 
 AMD_DAL_DCN10 = $(addprefix $(AMDDALPATH)/dc/dcn10/,$(DCN10))
 
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_cm_common.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_cm_common.c
index 7a00fe525dfb..dcd2cdfe91eb 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_cm_common.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_cm_common.c
@@ -24,7 +24,7 @@
  */
 #include "dc.h"
 #include "reg_helper.h"
-#include "dcn10_dpp.h"
+#include "dcn10/dcn10_dpp.h"
 
 #include "dcn10_cm_common.h"
 #include "custom_float.h"
@@ -62,6 +62,26 @@ void cm_helper_program_color_matrices(
 
 }
 
+void cm_helper_read_color_matrices(struct dc_context *ctx,
+				   uint16_t *regval,
+				   const struct color_matrices_reg *reg)
+{
+	uint32_t cur_csc_reg, regval0, regval1;
+	unsigned int i = 0;
+
+	for (cur_csc_reg = reg->csc_c11_c12;
+	     cur_csc_reg <= reg->csc_c33_c34; cur_csc_reg++) {
+		REG_GET_2(cur_csc_reg,
+				csc_c11, &regval0,
+				csc_c12, &regval1);
+
+		regval[2 * i] = regval0;
+		regval[(2 * i) + 1] = regval1;
+
+		i++;
+	}
+}
+
 void cm_helper_program_xfer_func(
 		struct dc_context *ctx,
 		const struct pwl_params *params,
@@ -308,7 +328,10 @@ bool cm_helper_convert_to_custom_float(
 #define NUMBER_REGIONS     32
 #define NUMBER_SW_SEGMENTS 16
 
-bool cm_helper_translate_curve_to_hw_format(
+#define DC_LOGGER \
+		ctx->logger
+
+bool cm_helper_translate_curve_to_hw_format(struct dc_context *ctx,
 				const struct dc_transfer_func *output_tf,
 				struct pwl_params *lut_params, bool fixpoint)
 {
@@ -342,23 +365,18 @@ bool cm_helper_translate_curve_to_hw_format(
 		region_start = -MAX_LOW_POINT;
 		region_end   = NUMBER_REGIONS - MAX_LOW_POINT;
 	} else {
-		/* 11 segments
-		 * segment is from 2^-10 to 2^1
+		/* 13 segments
+		 * segment is from 2^-12 to 2^0
 		 * There are less than 256 points, for optimization
 		 */
-		seg_distr[0] = 3;
-		seg_distr[1] = 4;
-		seg_distr[2] = 4;
-		seg_distr[3] = 4;
-		seg_distr[4] = 4;
-		seg_distr[5] = 4;
-		seg_distr[6] = 4;
-		seg_distr[7] = 4;
-		seg_distr[8] = 4;
-		seg_distr[9] = 4;
-		seg_distr[10] = 1;
-
-		region_start = -10;
+		const uint8_t SEG_COUNT = 12;
+
+		for (i = 0; i < SEG_COUNT; i++)
+			seg_distr[i] = 4;
+
+		seg_distr[SEG_COUNT] = 1;
+
+		region_start = -SEG_COUNT;
 		region_end = 1;
 	}
 
@@ -379,6 +397,11 @@ bool cm_helper_translate_curve_to_hw_format(
 				i += increment) {
 			if (j == hw_points - 1)
 				break;
+			if (i >= TRANSFER_FUNC_POINTS) {
+				DC_LOG_ERROR("Index out of bounds: i=%d, TRANSFER_FUNC_POINTS=%d\n",
+					     i, TRANSFER_FUNC_POINTS);
+				return false;
+			}
 			rgb_resulted[j].red = output_tf->tf_pts.red[i];
 			rgb_resulted[j].green = output_tf->tf_pts.green[i];
 			rgb_resulted[j].blue = output_tf->tf_pts.blue[i];
@@ -482,10 +505,18 @@ bool cm_helper_translate_curve_to_hw_format(
 		rgb->delta_green = dc_fixpt_sub(rgb_plus_1->green, rgb->green);
 		rgb->delta_blue  = dc_fixpt_sub(rgb_plus_1->blue,  rgb->blue);
 
+
 		if (fixpoint == true) {
-			rgb->delta_red_reg   = dc_fixpt_clamp_u0d10(rgb->delta_red);
-			rgb->delta_green_reg = dc_fixpt_clamp_u0d10(rgb->delta_green);
-			rgb->delta_blue_reg  = dc_fixpt_clamp_u0d10(rgb->delta_blue);
+			uint32_t red_clamp = dc_fixpt_clamp_u0d14(rgb->delta_red);
+			uint32_t green_clamp = dc_fixpt_clamp_u0d14(rgb->delta_green);
+			uint32_t blue_clamp = dc_fixpt_clamp_u0d14(rgb->delta_blue);
+
+			if (red_clamp >> 10 || green_clamp >> 10 || blue_clamp >> 10)
+				DC_LOG_WARNING("Losing delta precision while programming shaper LUT.");
+
+			rgb->delta_red_reg   = red_clamp & 0x3ff;
+			rgb->delta_green_reg = green_clamp & 0x3ff;
+			rgb->delta_blue_reg  = blue_clamp & 0x3ff;
 			rgb->red_reg         = dc_fixpt_clamp_u0d14(rgb->red);
 			rgb->green_reg       = dc_fixpt_clamp_u0d14(rgb->green);
 			rgb->blue_reg        = dc_fixpt_clamp_u0d14(rgb->blue);
@@ -555,6 +586,8 @@ bool cm_helper_translate_curve_to_degamma_hw_format(
 				i += increment) {
 			if (j == hw_points - 1)
 				break;
+			if (i >= TRANSFER_FUNC_POINTS)
+				return false;
 			rgb_resulted[j].red = output_tf->tf_pts.red[i];
 			rgb_resulted[j].green = output_tf->tf_pts.green[i];
 			rgb_resulted[j].blue = output_tf->tf_pts.blue[i];
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_cm_common.h b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_cm_common.h
index 3b8cd7410498..decc50b1ac53 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_cm_common.h
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_cm_common.h
@@ -106,6 +106,7 @@ bool cm_helper_convert_to_custom_float(
 		bool fixpoint);
 
 bool cm_helper_translate_curve_to_hw_format(
+		struct dc_context *ctx,
 		const struct dc_transfer_func *output_tf,
 		struct pwl_params *lut_params, bool fixpoint);
 
@@ -113,5 +114,7 @@ bool cm_helper_translate_curve_to_degamma_hw_format(
 				const struct dc_transfer_func *output_tf,
 				struct pwl_params *lut_params);
 
-
+void cm_helper_read_color_matrices(struct dc_context *ctx,
+				   uint16_t *regval,
+				   const struct color_matrices_reg *reg);
 #endif
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dwb.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dwb.c
index b6391a5ead78..365a3215f6d5 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dwb.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dwb.c
@@ -23,8 +23,6 @@
  *
  */
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-
 #include "reg_helper.h"
 #include "resource.h"
 #include "dwb.h"
@@ -129,6 +127,3 @@ void dcn10_dwbc_construct(struct dcn10_dwbc *dwbc10,
 	dwbc10->dwbc_shift = dwbc_shift;
 	dwbc10->dwbc_mask = dwbc_mask;
 }
-
-
-#endif
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dwb.h b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dwb.h
index d56ea7c8171e..5268c46ae907 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dwb.h
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dwb.h
@@ -24,8 +24,6 @@
 #ifndef __DC_DWBC_DCN10_H__
 #define __DC_DWBC_DCN10_H__
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-
 /* DCN */
 #define BASE_INNER(seg) \
 	DCE_BASE__INST0_SEG ## seg
@@ -267,5 +265,3 @@ void dcn10_dwbc_construct(struct dcn10_dwbc *dwbc10,
 		int inst);
 
 #endif
-
-#endif
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c
deleted file mode 100644
index 04d7bddc915b..000000000000
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c
+++ /dev/null
@@ -1,3775 +0,0 @@
-/*
- * Copyright 2016 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: AMD
- *
- */
-
-#include <linux/delay.h>
-#include "dm_services.h"
-#include "basics/dc_common.h"
-#include "core_types.h"
-#include "resource.h"
-#include "custom_float.h"
-#include "dcn10_hw_sequencer.h"
-#include "dcn10_hw_sequencer_debug.h"
-#include "dce/dce_hwseq.h"
-#include "abm.h"
-#include "dmcu.h"
-#include "dcn10_optc.h"
-#include "dcn10_dpp.h"
-#include "dcn10_mpc.h"
-#include "timing_generator.h"
-#include "opp.h"
-#include "ipp.h"
-#include "mpc.h"
-#include "reg_helper.h"
-#include "dcn10_hubp.h"
-#include "dcn10_hubbub.h"
-#include "dcn10_cm_common.h"
-#include "dc_link_dp.h"
-#include "dccg.h"
-#include "clk_mgr.h"
-#include "link_hwss.h"
-#include "dpcd_defs.h"
-#include "dsc.h"
-#include "dce/dmub_hw_lock_mgr.h"
-#include "dc_trace.h"
-#include "dce/dmub_outbox.h"
-#include "inc/dc_link_dp.h"
-#include "inc/link_dpcd.h"
-
-#define DC_LOGGER_INIT(logger)
-
-#define CTX \
-	hws->ctx
-#define REG(reg)\
-	hws->regs->reg
-
-#undef FN
-#define FN(reg_name, field_name) \
-	hws->shifts->field_name, hws->masks->field_name
-
-/*print is 17 wide, first two characters are spaces*/
-#define DTN_INFO_MICRO_SEC(ref_cycle) \
-	print_microsec(dc_ctx, log_ctx, ref_cycle)
-
-#define GAMMA_HW_POINTS_NUM 256
-
-#define PGFSM_POWER_ON 0
-#define PGFSM_POWER_OFF 2
-
-void print_microsec(struct dc_context *dc_ctx,
-	struct dc_log_buffer_ctx *log_ctx,
-	uint32_t ref_cycle)
-{
-	const uint32_t ref_clk_mhz = dc_ctx->dc->res_pool->ref_clocks.dchub_ref_clock_inKhz / 1000;
-	static const unsigned int frac = 1000;
-	uint32_t us_x10 = (ref_cycle * frac) / ref_clk_mhz;
-
-	DTN_INFO("  %11d.%03d",
-			us_x10 / frac,
-			us_x10 % frac);
-}
-
-void dcn10_lock_all_pipes(struct dc *dc,
-	struct dc_state *context,
-	bool lock)
-{
-	struct pipe_ctx *pipe_ctx;
-	struct timing_generator *tg;
-	int i;
-
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		pipe_ctx = &context->res_ctx.pipe_ctx[i];
-		tg = pipe_ctx->stream_res.tg;
-
-		/*
-		 * Only lock the top pipe's tg to prevent redundant
-		 * (un)locking. Also skip if pipe is disabled.
-		 */
-		if (pipe_ctx->top_pipe ||
-		    !pipe_ctx->stream || !pipe_ctx->plane_state ||
-		    !tg->funcs->is_tg_enabled(tg))
-			continue;
-
-		if (lock)
-			dc->hwss.pipe_control_lock(dc, pipe_ctx, true);
-		else
-			dc->hwss.pipe_control_lock(dc, pipe_ctx, false);
-	}
-}
-
-static void log_mpc_crc(struct dc *dc,
-	struct dc_log_buffer_ctx *log_ctx)
-{
-	struct dc_context *dc_ctx = dc->ctx;
-	struct dce_hwseq *hws = dc->hwseq;
-
-	if (REG(MPC_CRC_RESULT_GB))
-		DTN_INFO("MPC_CRC_RESULT_GB:%d MPC_CRC_RESULT_C:%d MPC_CRC_RESULT_AR:%d\n",
-		REG_READ(MPC_CRC_RESULT_GB), REG_READ(MPC_CRC_RESULT_C), REG_READ(MPC_CRC_RESULT_AR));
-	if (REG(DPP_TOP0_DPP_CRC_VAL_B_A))
-		DTN_INFO("DPP_TOP0_DPP_CRC_VAL_B_A:%d DPP_TOP0_DPP_CRC_VAL_R_G:%d\n",
-		REG_READ(DPP_TOP0_DPP_CRC_VAL_B_A), REG_READ(DPP_TOP0_DPP_CRC_VAL_R_G));
-}
-
-void dcn10_log_hubbub_state(struct dc *dc, struct dc_log_buffer_ctx *log_ctx)
-{
-	struct dc_context *dc_ctx = dc->ctx;
-	struct dcn_hubbub_wm wm;
-	int i;
-
-	memset(&wm, 0, sizeof(struct dcn_hubbub_wm));
-	dc->res_pool->hubbub->funcs->wm_read_state(dc->res_pool->hubbub, &wm);
-
-	DTN_INFO("HUBBUB WM:      data_urgent  pte_meta_urgent"
-			"         sr_enter          sr_exit  dram_clk_change\n");
-
-	for (i = 0; i < 4; i++) {
-		struct dcn_hubbub_wm_set *s;
-
-		s = &wm.sets[i];
-		DTN_INFO("WM_Set[%d]:", s->wm_set);
-		DTN_INFO_MICRO_SEC(s->data_urgent);
-		DTN_INFO_MICRO_SEC(s->pte_meta_urgent);
-		DTN_INFO_MICRO_SEC(s->sr_enter);
-		DTN_INFO_MICRO_SEC(s->sr_exit);
-		DTN_INFO_MICRO_SEC(s->dram_clk_chanage);
-		DTN_INFO("\n");
-	}
-
-	DTN_INFO("\n");
-}
-
-static void dcn10_log_hubp_states(struct dc *dc, void *log_ctx)
-{
-	struct dc_context *dc_ctx = dc->ctx;
-	struct resource_pool *pool = dc->res_pool;
-	int i;
-
-	DTN_INFO(
-		"HUBP:  format  addr_hi  width  height  rot  mir  sw_mode  dcc_en  blank_en  clock_en  ttu_dis  underflow   min_ttu_vblank       qos_low_wm      qos_high_wm\n");
-	for (i = 0; i < pool->pipe_count; i++) {
-		struct hubp *hubp = pool->hubps[i];
-		struct dcn_hubp_state *s = &(TO_DCN10_HUBP(hubp)->state);
-
-		hubp->funcs->hubp_read_state(hubp);
-
-		if (!s->blank_en) {
-			DTN_INFO("[%2d]:  %5xh  %6xh  %5d  %6d  %2xh  %2xh  %6xh  %6d  %8d  %8d  %7d  %8xh",
-					hubp->inst,
-					s->pixel_format,
-					s->inuse_addr_hi,
-					s->viewport_width,
-					s->viewport_height,
-					s->rotation_angle,
-					s->h_mirror_en,
-					s->sw_mode,
-					s->dcc_en,
-					s->blank_en,
-					s->clock_en,
-					s->ttu_disable,
-					s->underflow_status);
-			DTN_INFO_MICRO_SEC(s->min_ttu_vblank);
-			DTN_INFO_MICRO_SEC(s->qos_level_low_wm);
-			DTN_INFO_MICRO_SEC(s->qos_level_high_wm);
-			DTN_INFO("\n");
-		}
-	}
-
-	DTN_INFO("\n=========RQ========\n");
-	DTN_INFO("HUBP:  drq_exp_m  prq_exp_m  mrq_exp_m  crq_exp_m  plane1_ba  L:chunk_s  min_chu_s  meta_ch_s"
-		"  min_m_c_s  dpte_gr_s  mpte_gr_s  swath_hei  pte_row_h  C:chunk_s  min_chu_s  meta_ch_s"
-		"  min_m_c_s  dpte_gr_s  mpte_gr_s  swath_hei  pte_row_h\n");
-	for (i = 0; i < pool->pipe_count; i++) {
-		struct dcn_hubp_state *s = &(TO_DCN10_HUBP(pool->hubps[i])->state);
-		struct _vcs_dpi_display_rq_regs_st *rq_regs = &s->rq_regs;
-
-		if (!s->blank_en)
-			DTN_INFO("[%2d]:  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh\n",
-				pool->hubps[i]->inst, rq_regs->drq_expansion_mode, rq_regs->prq_expansion_mode, rq_regs->mrq_expansion_mode,
-				rq_regs->crq_expansion_mode, rq_regs->plane1_base_address, rq_regs->rq_regs_l.chunk_size,
-				rq_regs->rq_regs_l.min_chunk_size, rq_regs->rq_regs_l.meta_chunk_size,
-				rq_regs->rq_regs_l.min_meta_chunk_size, rq_regs->rq_regs_l.dpte_group_size,
-				rq_regs->rq_regs_l.mpte_group_size, rq_regs->rq_regs_l.swath_height,
-				rq_regs->rq_regs_l.pte_row_height_linear, rq_regs->rq_regs_c.chunk_size, rq_regs->rq_regs_c.min_chunk_size,
-				rq_regs->rq_regs_c.meta_chunk_size, rq_regs->rq_regs_c.min_meta_chunk_size,
-				rq_regs->rq_regs_c.dpte_group_size, rq_regs->rq_regs_c.mpte_group_size,
-				rq_regs->rq_regs_c.swath_height, rq_regs->rq_regs_c.pte_row_height_linear);
-	}
-
-	DTN_INFO("========DLG========\n");
-	DTN_INFO("HUBP:  rc_hbe     dlg_vbe    min_d_y_n  rc_per_ht  rc_x_a_s "
-			"  dst_y_a_s  dst_y_pf   dst_y_vvb  dst_y_rvb  dst_y_vfl  dst_y_rfl  rf_pix_fq"
-			"  vratio_pf  vrat_pf_c  rc_pg_vbl  rc_pg_vbc  rc_mc_vbl  rc_mc_vbc  rc_pg_fll"
-			"  rc_pg_flc  rc_mc_fll  rc_mc_flc  pr_nom_l   pr_nom_c   rc_pg_nl   rc_pg_nc "
-			"  mr_nom_l   mr_nom_c   rc_mc_nl   rc_mc_nc   rc_ld_pl   rc_ld_pc   rc_ld_l  "
-			"  rc_ld_c    cha_cur0   ofst_cur1  cha_cur1   vr_af_vc0  ddrq_limt  x_rt_dlay"
-			"  x_rp_dlay  x_rr_sfl\n");
-	for (i = 0; i < pool->pipe_count; i++) {
-		struct dcn_hubp_state *s = &(TO_DCN10_HUBP(pool->hubps[i])->state);
-		struct _vcs_dpi_display_dlg_regs_st *dlg_regs = &s->dlg_attr;
-
-		if (!s->blank_en)
-			DTN_INFO("[%2d]:  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh"
-				"  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh"
-				"  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh\n",
-				pool->hubps[i]->inst, dlg_regs->refcyc_h_blank_end, dlg_regs->dlg_vblank_end, dlg_regs->min_dst_y_next_start,
-				dlg_regs->refcyc_per_htotal, dlg_regs->refcyc_x_after_scaler, dlg_regs->dst_y_after_scaler,
-				dlg_regs->dst_y_prefetch, dlg_regs->dst_y_per_vm_vblank, dlg_regs->dst_y_per_row_vblank,
-				dlg_regs->dst_y_per_vm_flip, dlg_regs->dst_y_per_row_flip, dlg_regs->ref_freq_to_pix_freq,
-				dlg_regs->vratio_prefetch, dlg_regs->vratio_prefetch_c, dlg_regs->refcyc_per_pte_group_vblank_l,
-				dlg_regs->refcyc_per_pte_group_vblank_c, dlg_regs->refcyc_per_meta_chunk_vblank_l,
-				dlg_regs->refcyc_per_meta_chunk_vblank_c, dlg_regs->refcyc_per_pte_group_flip_l,
-				dlg_regs->refcyc_per_pte_group_flip_c, dlg_regs->refcyc_per_meta_chunk_flip_l,
-				dlg_regs->refcyc_per_meta_chunk_flip_c, dlg_regs->dst_y_per_pte_row_nom_l,
-				dlg_regs->dst_y_per_pte_row_nom_c, dlg_regs->refcyc_per_pte_group_nom_l,
-				dlg_regs->refcyc_per_pte_group_nom_c, dlg_regs->dst_y_per_meta_row_nom_l,
-				dlg_regs->dst_y_per_meta_row_nom_c, dlg_regs->refcyc_per_meta_chunk_nom_l,
-				dlg_regs->refcyc_per_meta_chunk_nom_c, dlg_regs->refcyc_per_line_delivery_pre_l,
-				dlg_regs->refcyc_per_line_delivery_pre_c, dlg_regs->refcyc_per_line_delivery_l,
-				dlg_regs->refcyc_per_line_delivery_c, dlg_regs->chunk_hdl_adjust_cur0, dlg_regs->dst_y_offset_cur1,
-				dlg_regs->chunk_hdl_adjust_cur1, dlg_regs->vready_after_vcount0, dlg_regs->dst_y_delta_drq_limit,
-				dlg_regs->xfc_reg_transfer_delay, dlg_regs->xfc_reg_precharge_delay,
-				dlg_regs->xfc_reg_remote_surface_flip_latency);
-	}
-
-	DTN_INFO("========TTU========\n");
-	DTN_INFO("HUBP:  qos_ll_wm  qos_lh_wm  mn_ttu_vb  qos_l_flp  rc_rd_p_l  rc_rd_l    rc_rd_p_c"
-			"  rc_rd_c    rc_rd_c0   rc_rd_pc0  rc_rd_c1   rc_rd_pc1  qos_lf_l   qos_rds_l"
-			"  qos_lf_c   qos_rds_c  qos_lf_c0  qos_rds_c0 qos_lf_c1  qos_rds_c1\n");
-	for (i = 0; i < pool->pipe_count; i++) {
-		struct dcn_hubp_state *s = &(TO_DCN10_HUBP(pool->hubps[i])->state);
-		struct _vcs_dpi_display_ttu_regs_st *ttu_regs = &s->ttu_attr;
-
-		if (!s->blank_en)
-			DTN_INFO("[%2d]:  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh  %8xh\n",
-				pool->hubps[i]->inst, ttu_regs->qos_level_low_wm, ttu_regs->qos_level_high_wm, ttu_regs->min_ttu_vblank,
-				ttu_regs->qos_level_flip, ttu_regs->refcyc_per_req_delivery_pre_l, ttu_regs->refcyc_per_req_delivery_l,
-				ttu_regs->refcyc_per_req_delivery_pre_c, ttu_regs->refcyc_per_req_delivery_c, ttu_regs->refcyc_per_req_delivery_cur0,
-				ttu_regs->refcyc_per_req_delivery_pre_cur0, ttu_regs->refcyc_per_req_delivery_cur1,
-				ttu_regs->refcyc_per_req_delivery_pre_cur1, ttu_regs->qos_level_fixed_l, ttu_regs->qos_ramp_disable_l,
-				ttu_regs->qos_level_fixed_c, ttu_regs->qos_ramp_disable_c, ttu_regs->qos_level_fixed_cur0,
-				ttu_regs->qos_ramp_disable_cur0, ttu_regs->qos_level_fixed_cur1, ttu_regs->qos_ramp_disable_cur1);
-	}
-	DTN_INFO("\n");
-}
-
-void dcn10_log_hw_state(struct dc *dc,
-	struct dc_log_buffer_ctx *log_ctx)
-{
-	struct dc_context *dc_ctx = dc->ctx;
-	struct resource_pool *pool = dc->res_pool;
-	int i;
-
-	DTN_INFO_BEGIN();
-
-	dcn10_log_hubbub_state(dc, log_ctx);
-
-	dcn10_log_hubp_states(dc, log_ctx);
-
-	DTN_INFO("DPP:    IGAM format  IGAM mode    DGAM mode    RGAM mode"
-			"  GAMUT mode  C11 C12   C13 C14   C21 C22   C23 C24   "
-			"C31 C32   C33 C34\n");
-	for (i = 0; i < pool->pipe_count; i++) {
-		struct dpp *dpp = pool->dpps[i];
-		struct dcn_dpp_state s = {0};
-
-		dpp->funcs->dpp_read_state(dpp, &s);
-
-		if (!s.is_enabled)
-			continue;
-
-		DTN_INFO("[%2d]:  %11xh  %-11s  %-11s  %-11s"
-				"%8x    %08xh %08xh %08xh %08xh %08xh %08xh",
-				dpp->inst,
-				s.igam_input_format,
-				(s.igam_lut_mode == 0) ? "BypassFixed" :
-					((s.igam_lut_mode == 1) ? "BypassFloat" :
-					((s.igam_lut_mode == 2) ? "RAM" :
-					((s.igam_lut_mode == 3) ? "RAM" :
-								 "Unknown"))),
-				(s.dgam_lut_mode == 0) ? "Bypass" :
-					((s.dgam_lut_mode == 1) ? "sRGB" :
-					((s.dgam_lut_mode == 2) ? "Ycc" :
-					((s.dgam_lut_mode == 3) ? "RAM" :
-					((s.dgam_lut_mode == 4) ? "RAM" :
-								 "Unknown")))),
-				(s.rgam_lut_mode == 0) ? "Bypass" :
-					((s.rgam_lut_mode == 1) ? "sRGB" :
-					((s.rgam_lut_mode == 2) ? "Ycc" :
-					((s.rgam_lut_mode == 3) ? "RAM" :
-					((s.rgam_lut_mode == 4) ? "RAM" :
-								 "Unknown")))),
-				s.gamut_remap_mode,
-				s.gamut_remap_c11_c12,
-				s.gamut_remap_c13_c14,
-				s.gamut_remap_c21_c22,
-				s.gamut_remap_c23_c24,
-				s.gamut_remap_c31_c32,
-				s.gamut_remap_c33_c34);
-		DTN_INFO("\n");
-	}
-	DTN_INFO("\n");
-
-	DTN_INFO("MPCC:  OPP  DPP  MPCCBOT  MODE  ALPHA_MODE  PREMULT  OVERLAP_ONLY  IDLE\n");
-	for (i = 0; i < pool->pipe_count; i++) {
-		struct mpcc_state s = {0};
-
-		pool->mpc->funcs->read_mpcc_state(pool->mpc, i, &s);
-		if (s.opp_id != 0xf)
-			DTN_INFO("[%2d]:  %2xh  %2xh  %6xh  %4d  %10d  %7d  %12d  %4d\n",
-				i, s.opp_id, s.dpp_id, s.bot_mpcc_id,
-				s.mode, s.alpha_mode, s.pre_multiplied_alpha, s.overlap_only,
-				s.idle);
-	}
-	DTN_INFO("\n");
-
-	DTN_INFO("OTG:  v_bs  v_be  v_ss  v_se  vpol  vmax  vmin  vmax_sel  vmin_sel  h_bs  h_be  h_ss  h_se  hpol  htot  vtot  underflow blank_en\n");
-
-	for (i = 0; i < pool->timing_generator_count; i++) {
-		struct timing_generator *tg = pool->timing_generators[i];
-		struct dcn_otg_state s = {0};
-		/* Read shared OTG state registers for all DCNx */
-		optc1_read_otg_state(DCN10TG_FROM_TG(tg), &s);
-
-		/*
-		 * For DCN2 and greater, a register on the OPP is used to
-		 * determine if the CRTC is blanked instead of the OTG. So use
-		 * dpg_is_blanked() if exists, otherwise fallback on otg.
-		 *
-		 * TODO: Implement DCN-specific read_otg_state hooks.
-		 */
-		if (pool->opps[i]->funcs->dpg_is_blanked)
-			s.blank_enabled = pool->opps[i]->funcs->dpg_is_blanked(pool->opps[i]);
-		else
-			s.blank_enabled = tg->funcs->is_blanked(tg);
-
-		//only print if OTG master is enabled
-		if ((s.otg_enabled & 1) == 0)
-			continue;
-
-		DTN_INFO("[%d]: %5d %5d %5d %5d %5d %5d %5d %9d %9d %5d %5d %5d %5d %5d %5d %5d  %9d %8d\n",
-				tg->inst,
-				s.v_blank_start,
-				s.v_blank_end,
-				s.v_sync_a_start,
-				s.v_sync_a_end,
-				s.v_sync_a_pol,
-				s.v_total_max,
-				s.v_total_min,
-				s.v_total_max_sel,
-				s.v_total_min_sel,
-				s.h_blank_start,
-				s.h_blank_end,
-				s.h_sync_a_start,
-				s.h_sync_a_end,
-				s.h_sync_a_pol,
-				s.h_total,
-				s.v_total,
-				s.underflow_occurred_status,
-				s.blank_enabled);
-
-		// Clear underflow for debug purposes
-		// We want to keep underflow sticky bit on for the longevity tests outside of test environment.
-		// This function is called only from Windows or Diags test environment, hence it's safe to clear
-		// it from here without affecting the original intent.
-		tg->funcs->clear_optc_underflow(tg);
-	}
-	DTN_INFO("\n");
-
-	// dcn_dsc_state struct field bytes_per_pixel was renamed to bits_per_pixel
-	// TODO: Update golden log header to reflect this name change
-	DTN_INFO("DSC: CLOCK_EN  SLICE_WIDTH  Bytes_pp\n");
-	for (i = 0; i < pool->res_cap->num_dsc; i++) {
-		struct display_stream_compressor *dsc = pool->dscs[i];
-		struct dcn_dsc_state s = {0};
-
-		dsc->funcs->dsc_read_state(dsc, &s);
-		DTN_INFO("[%d]: %-9d %-12d %-10d\n",
-		dsc->inst,
-			s.dsc_clock_en,
-			s.dsc_slice_width,
-			s.dsc_bits_per_pixel);
-		DTN_INFO("\n");
-	}
-	DTN_INFO("\n");
-
-	DTN_INFO("S_ENC: DSC_MODE  SEC_GSP7_LINE_NUM"
-			"  VBID6_LINE_REFERENCE  VBID6_LINE_NUM  SEC_GSP7_ENABLE  SEC_STREAM_ENABLE\n");
-	for (i = 0; i < pool->stream_enc_count; i++) {
-		struct stream_encoder *enc = pool->stream_enc[i];
-		struct enc_state s = {0};
-
-		if (enc->funcs->enc_read_state) {
-			enc->funcs->enc_read_state(enc, &s);
-			DTN_INFO("[%-3d]: %-9d %-18d %-21d %-15d %-16d %-17d\n",
-				enc->id,
-				s.dsc_mode,
-				s.sec_gsp_pps_line_num,
-				s.vbid6_line_reference,
-				s.vbid6_line_num,
-				s.sec_gsp_pps_enable,
-				s.sec_stream_enable);
-			DTN_INFO("\n");
-		}
-	}
-	DTN_INFO("\n");
-
-	DTN_INFO("L_ENC: DPHY_FEC_EN  DPHY_FEC_READY_SHADOW  DPHY_FEC_ACTIVE_STATUS  DP_LINK_TRAINING_COMPLETE\n");
-	for (i = 0; i < dc->link_count; i++) {
-		struct link_encoder *lenc = dc->links[i]->link_enc;
-
-		struct link_enc_state s = {0};
-
-		if (lenc->funcs->read_state) {
-			lenc->funcs->read_state(lenc, &s);
-			DTN_INFO("[%-3d]: %-12d %-22d %-22d %-25d\n",
-				i,
-				s.dphy_fec_en,
-				s.dphy_fec_ready_shadow,
-				s.dphy_fec_active_status,
-				s.dp_link_training_complete);
-			DTN_INFO("\n");
-		}
-	}
-	DTN_INFO("\n");
-
-	DTN_INFO("\nCALCULATED Clocks: dcfclk_khz:%d  dcfclk_deep_sleep_khz:%d  dispclk_khz:%d\n"
-		"dppclk_khz:%d  max_supported_dppclk_khz:%d  fclk_khz:%d  socclk_khz:%d\n\n",
-			dc->current_state->bw_ctx.bw.dcn.clk.dcfclk_khz,
-			dc->current_state->bw_ctx.bw.dcn.clk.dcfclk_deep_sleep_khz,
-			dc->current_state->bw_ctx.bw.dcn.clk.dispclk_khz,
-			dc->current_state->bw_ctx.bw.dcn.clk.dppclk_khz,
-			dc->current_state->bw_ctx.bw.dcn.clk.max_supported_dppclk_khz,
-			dc->current_state->bw_ctx.bw.dcn.clk.fclk_khz,
-			dc->current_state->bw_ctx.bw.dcn.clk.socclk_khz);
-
-	log_mpc_crc(dc, log_ctx);
-
-	{
-		int hpo_dp_link_enc_count = 0;
-
-		if (pool->hpo_dp_stream_enc_count > 0) {
-			DTN_INFO("DP HPO S_ENC:  Enabled  OTG   Format   Depth   Vid   SDP   Compressed  Link\n");
-			for (i = 0; i < pool->hpo_dp_stream_enc_count; i++) {
-				struct hpo_dp_stream_encoder_state hpo_dp_se_state = {0};
-				struct hpo_dp_stream_encoder *hpo_dp_stream_enc = pool->hpo_dp_stream_enc[i];
-
-				if (hpo_dp_stream_enc && hpo_dp_stream_enc->funcs->read_state) {
-					hpo_dp_stream_enc->funcs->read_state(hpo_dp_stream_enc, &hpo_dp_se_state);
-
-					DTN_INFO("[%d]:                 %d    %d   %6s       %d     %d     %d            %d     %d\n",
-							hpo_dp_stream_enc->id - ENGINE_ID_HPO_DP_0,
-							hpo_dp_se_state.stream_enc_enabled,
-							hpo_dp_se_state.otg_inst,
-							(hpo_dp_se_state.pixel_encoding == 0) ? "4:4:4" :
-									((hpo_dp_se_state.pixel_encoding == 1) ? "4:2:2" :
-									(hpo_dp_se_state.pixel_encoding == 2) ? "4:2:0" : "Y-Only"),
-							(hpo_dp_se_state.component_depth == 0) ? 6 :
-									((hpo_dp_se_state.component_depth == 1) ? 8 :
-									(hpo_dp_se_state.component_depth == 2) ? 10 : 12),
-							hpo_dp_se_state.vid_stream_enabled,
-							hpo_dp_se_state.sdp_enabled,
-							hpo_dp_se_state.compressed_format,
-							hpo_dp_se_state.mapped_to_link_enc);
-				}
-			}
-
-			DTN_INFO("\n");
-		}
-
-		/* log DP HPO L_ENC section if any hpo_dp_link_enc exists */
-		for (i = 0; i < dc->link_count; i++)
-			if (dc->links[i]->hpo_dp_link_enc)
-				hpo_dp_link_enc_count++;
-
-		if (hpo_dp_link_enc_count) {
-			DTN_INFO("DP HPO L_ENC:  Enabled  Mode   Lanes   Stream  Slots   VC Rate X    VC Rate Y\n");
-
-			for (i = 0; i < dc->link_count; i++) {
-				struct hpo_dp_link_encoder *hpo_dp_link_enc = dc->links[i]->hpo_dp_link_enc;
-				struct hpo_dp_link_enc_state hpo_dp_le_state = {0};
-
-				if (hpo_dp_link_enc && hpo_dp_link_enc->funcs->read_state) {
-					hpo_dp_link_enc->funcs->read_state(hpo_dp_link_enc, &hpo_dp_le_state);
-					DTN_INFO("[%d]:                 %d  %6s     %d        %d      %d     %d     %d\n",
-							hpo_dp_link_enc->inst,
-							hpo_dp_le_state.link_enc_enabled,
-							(hpo_dp_le_state.link_mode == 0) ? "TPS1" :
-									(hpo_dp_le_state.link_mode == 1) ? "TPS2" :
-									(hpo_dp_le_state.link_mode == 2) ? "ACTIVE" : "TEST",
-							hpo_dp_le_state.lane_count,
-							hpo_dp_le_state.stream_src[0],
-							hpo_dp_le_state.slot_count[0],
-							hpo_dp_le_state.vc_rate_x[0],
-							hpo_dp_le_state.vc_rate_y[0]);
-					DTN_INFO("\n");
-				}
-			}
-
-			DTN_INFO("\n");
-		}
-	}
-
-	DTN_INFO_END();
-}
-
-bool dcn10_did_underflow_occur(struct dc *dc, struct pipe_ctx *pipe_ctx)
-{
-	struct hubp *hubp = pipe_ctx->plane_res.hubp;
-	struct timing_generator *tg = pipe_ctx->stream_res.tg;
-
-	if (tg->funcs->is_optc_underflow_occurred(tg)) {
-		tg->funcs->clear_optc_underflow(tg);
-		return true;
-	}
-
-	if (hubp->funcs->hubp_get_underflow_status(hubp)) {
-		hubp->funcs->hubp_clear_underflow(hubp);
-		return true;
-	}
-	return false;
-}
-
-void dcn10_enable_power_gating_plane(
-	struct dce_hwseq *hws,
-	bool enable)
-{
-	bool force_on = true; /* disable power gating */
-
-	if (enable)
-		force_on = false;
-
-	/* DCHUBP0/1/2/3 */
-	REG_UPDATE(DOMAIN0_PG_CONFIG, DOMAIN0_POWER_FORCEON, force_on);
-	REG_UPDATE(DOMAIN2_PG_CONFIG, DOMAIN2_POWER_FORCEON, force_on);
-	REG_UPDATE(DOMAIN4_PG_CONFIG, DOMAIN4_POWER_FORCEON, force_on);
-	REG_UPDATE(DOMAIN6_PG_CONFIG, DOMAIN6_POWER_FORCEON, force_on);
-
-	/* DPP0/1/2/3 */
-	REG_UPDATE(DOMAIN1_PG_CONFIG, DOMAIN1_POWER_FORCEON, force_on);
-	REG_UPDATE(DOMAIN3_PG_CONFIG, DOMAIN3_POWER_FORCEON, force_on);
-	REG_UPDATE(DOMAIN5_PG_CONFIG, DOMAIN5_POWER_FORCEON, force_on);
-	REG_UPDATE(DOMAIN7_PG_CONFIG, DOMAIN7_POWER_FORCEON, force_on);
-}
-
-void dcn10_disable_vga(
-	struct dce_hwseq *hws)
-{
-	unsigned int in_vga1_mode = 0;
-	unsigned int in_vga2_mode = 0;
-	unsigned int in_vga3_mode = 0;
-	unsigned int in_vga4_mode = 0;
-
-	REG_GET(D1VGA_CONTROL, D1VGA_MODE_ENABLE, &in_vga1_mode);
-	REG_GET(D2VGA_CONTROL, D2VGA_MODE_ENABLE, &in_vga2_mode);
-	REG_GET(D3VGA_CONTROL, D3VGA_MODE_ENABLE, &in_vga3_mode);
-	REG_GET(D4VGA_CONTROL, D4VGA_MODE_ENABLE, &in_vga4_mode);
-
-	if (in_vga1_mode == 0 && in_vga2_mode == 0 &&
-			in_vga3_mode == 0 && in_vga4_mode == 0)
-		return;
-
-	REG_WRITE(D1VGA_CONTROL, 0);
-	REG_WRITE(D2VGA_CONTROL, 0);
-	REG_WRITE(D3VGA_CONTROL, 0);
-	REG_WRITE(D4VGA_CONTROL, 0);
-
-	/* HW Engineer's Notes:
-	 *  During switch from vga->extended, if we set the VGA_TEST_ENABLE and
-	 *  then hit the VGA_TEST_RENDER_START, then the DCHUBP timing gets updated correctly.
-	 *
-	 *  Then vBIOS will have it poll for the VGA_TEST_RENDER_DONE and unset
-	 *  VGA_TEST_ENABLE, to leave it in the same state as before.
-	 */
-	REG_UPDATE(VGA_TEST_CONTROL, VGA_TEST_ENABLE, 1);
-	REG_UPDATE(VGA_TEST_CONTROL, VGA_TEST_RENDER_START, 1);
-}
-
-/**
- * dcn10_dpp_pg_control - DPP power gate control.
- *
- * @hws: dce_hwseq reference.
- * @dpp_inst: DPP instance reference.
- * @power_on: true if we want to enable power gate, false otherwise.
- *
- * Enable or disable power gate in the specific DPP instance.
- */
-void dcn10_dpp_pg_control(
-		struct dce_hwseq *hws,
-		unsigned int dpp_inst,
-		bool power_on)
-{
-	uint32_t power_gate = power_on ? 0 : 1;
-	uint32_t pwr_status = power_on ? PGFSM_POWER_ON : PGFSM_POWER_OFF;
-
-	if (hws->ctx->dc->debug.disable_dpp_power_gate)
-		return;
-	if (REG(DOMAIN1_PG_CONFIG) == 0)
-		return;
-
-	switch (dpp_inst) {
-	case 0: /* DPP0 */
-		REG_UPDATE(DOMAIN1_PG_CONFIG,
-				DOMAIN1_POWER_GATE, power_gate);
-
-		REG_WAIT(DOMAIN1_PG_STATUS,
-				DOMAIN1_PGFSM_PWR_STATUS, pwr_status,
-				1, 1000);
-		break;
-	case 1: /* DPP1 */
-		REG_UPDATE(DOMAIN3_PG_CONFIG,
-				DOMAIN3_POWER_GATE, power_gate);
-
-		REG_WAIT(DOMAIN3_PG_STATUS,
-				DOMAIN3_PGFSM_PWR_STATUS, pwr_status,
-				1, 1000);
-		break;
-	case 2: /* DPP2 */
-		REG_UPDATE(DOMAIN5_PG_CONFIG,
-				DOMAIN5_POWER_GATE, power_gate);
-
-		REG_WAIT(DOMAIN5_PG_STATUS,
-				DOMAIN5_PGFSM_PWR_STATUS, pwr_status,
-				1, 1000);
-		break;
-	case 3: /* DPP3 */
-		REG_UPDATE(DOMAIN7_PG_CONFIG,
-				DOMAIN7_POWER_GATE, power_gate);
-
-		REG_WAIT(DOMAIN7_PG_STATUS,
-				DOMAIN7_PGFSM_PWR_STATUS, pwr_status,
-				1, 1000);
-		break;
-	default:
-		BREAK_TO_DEBUGGER();
-		break;
-	}
-}
-
-/**
- * dcn10_hubp_pg_control - HUBP power gate control.
- *
- * @hws: dce_hwseq reference.
- * @hubp_inst: DPP instance reference.
- * @power_on: true if we want to enable power gate, false otherwise.
- *
- * Enable or disable power gate in the specific HUBP instance.
- */
-void dcn10_hubp_pg_control(
-		struct dce_hwseq *hws,
-		unsigned int hubp_inst,
-		bool power_on)
-{
-	uint32_t power_gate = power_on ? 0 : 1;
-	uint32_t pwr_status = power_on ? PGFSM_POWER_ON : PGFSM_POWER_OFF;
-
-	if (hws->ctx->dc->debug.disable_hubp_power_gate)
-		return;
-	if (REG(DOMAIN0_PG_CONFIG) == 0)
-		return;
-
-	switch (hubp_inst) {
-	case 0: /* DCHUBP0 */
-		REG_UPDATE(DOMAIN0_PG_CONFIG,
-				DOMAIN0_POWER_GATE, power_gate);
-
-		REG_WAIT(DOMAIN0_PG_STATUS,
-				DOMAIN0_PGFSM_PWR_STATUS, pwr_status,
-				1, 1000);
-		break;
-	case 1: /* DCHUBP1 */
-		REG_UPDATE(DOMAIN2_PG_CONFIG,
-				DOMAIN2_POWER_GATE, power_gate);
-
-		REG_WAIT(DOMAIN2_PG_STATUS,
-				DOMAIN2_PGFSM_PWR_STATUS, pwr_status,
-				1, 1000);
-		break;
-	case 2: /* DCHUBP2 */
-		REG_UPDATE(DOMAIN4_PG_CONFIG,
-				DOMAIN4_POWER_GATE, power_gate);
-
-		REG_WAIT(DOMAIN4_PG_STATUS,
-				DOMAIN4_PGFSM_PWR_STATUS, pwr_status,
-				1, 1000);
-		break;
-	case 3: /* DCHUBP3 */
-		REG_UPDATE(DOMAIN6_PG_CONFIG,
-				DOMAIN6_POWER_GATE, power_gate);
-
-		REG_WAIT(DOMAIN6_PG_STATUS,
-				DOMAIN6_PGFSM_PWR_STATUS, pwr_status,
-				1, 1000);
-		break;
-	default:
-		BREAK_TO_DEBUGGER();
-		break;
-	}
-}
-
-static void power_on_plane(
-	struct dce_hwseq *hws,
-	int plane_id)
-{
-	DC_LOGGER_INIT(hws->ctx->logger);
-	if (REG(DC_IP_REQUEST_CNTL)) {
-		REG_SET(DC_IP_REQUEST_CNTL, 0,
-				IP_REQUEST_EN, 1);
-
-		if (hws->funcs.dpp_pg_control)
-			hws->funcs.dpp_pg_control(hws, plane_id, true);
-
-		if (hws->funcs.hubp_pg_control)
-			hws->funcs.hubp_pg_control(hws, plane_id, true);
-
-		REG_SET(DC_IP_REQUEST_CNTL, 0,
-				IP_REQUEST_EN, 0);
-		DC_LOG_DEBUG(
-				"Un-gated front end for pipe %d\n", plane_id);
-	}
-}
-
-static void undo_DEGVIDCN10_253_wa(struct dc *dc)
-{
-	struct dce_hwseq *hws = dc->hwseq;
-	struct hubp *hubp = dc->res_pool->hubps[0];
-
-	if (!hws->wa_state.DEGVIDCN10_253_applied)
-		return;
-
-	hubp->funcs->set_blank(hubp, true);
-
-	REG_SET(DC_IP_REQUEST_CNTL, 0,
-			IP_REQUEST_EN, 1);
-
-	hws->funcs.hubp_pg_control(hws, 0, false);
-	REG_SET(DC_IP_REQUEST_CNTL, 0,
-			IP_REQUEST_EN, 0);
-
-	hws->wa_state.DEGVIDCN10_253_applied = false;
-}
-
-static void apply_DEGVIDCN10_253_wa(struct dc *dc)
-{
-	struct dce_hwseq *hws = dc->hwseq;
-	struct hubp *hubp = dc->res_pool->hubps[0];
-	int i;
-
-	if (dc->debug.disable_stutter)
-		return;
-
-	if (!hws->wa.DEGVIDCN10_253)
-		return;
-
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		if (!dc->res_pool->hubps[i]->power_gated)
-			return;
-	}
-
-	/* all pipe power gated, apply work around to enable stutter. */
-
-	REG_SET(DC_IP_REQUEST_CNTL, 0,
-			IP_REQUEST_EN, 1);
-
-	hws->funcs.hubp_pg_control(hws, 0, true);
-	REG_SET(DC_IP_REQUEST_CNTL, 0,
-			IP_REQUEST_EN, 0);
-
-	hubp->funcs->set_hubp_blank_en(hubp, false);
-	hws->wa_state.DEGVIDCN10_253_applied = true;
-}
-
-void dcn10_bios_golden_init(struct dc *dc)
-{
-	struct dce_hwseq *hws = dc->hwseq;
-	struct dc_bios *bp = dc->ctx->dc_bios;
-	int i;
-	bool allow_self_fresh_force_enable = true;
-
-	if (hws->funcs.s0i3_golden_init_wa && hws->funcs.s0i3_golden_init_wa(dc))
-		return;
-
-	if (dc->res_pool->hubbub->funcs->is_allow_self_refresh_enabled)
-		allow_self_fresh_force_enable =
-				dc->res_pool->hubbub->funcs->is_allow_self_refresh_enabled(dc->res_pool->hubbub);
-
-
-	/* WA for making DF sleep when idle after resume from S0i3.
-	 * DCHUBBUB_ARB_ALLOW_SELF_REFRESH_FORCE_ENABLE is set to 1 by
-	 * command table, if DCHUBBUB_ARB_ALLOW_SELF_REFRESH_FORCE_ENABLE = 0
-	 * before calling command table and it changed to 1 after,
-	 * it should be set back to 0.
-	 */
-
-	/* initialize dcn global */
-	bp->funcs->enable_disp_power_gating(bp,
-			CONTROLLER_ID_D0, ASIC_PIPE_INIT);
-
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		/* initialize dcn per pipe */
-		bp->funcs->enable_disp_power_gating(bp,
-				CONTROLLER_ID_D0 + i, ASIC_PIPE_DISABLE);
-	}
-
-	if (dc->res_pool->hubbub->funcs->allow_self_refresh_control)
-		if (allow_self_fresh_force_enable == false &&
-				dc->res_pool->hubbub->funcs->is_allow_self_refresh_enabled(dc->res_pool->hubbub))
-			dc->res_pool->hubbub->funcs->allow_self_refresh_control(dc->res_pool->hubbub,
-										!dc->res_pool->hubbub->ctx->dc->debug.disable_stutter);
-
-}
-
-static void false_optc_underflow_wa(
-		struct dc *dc,
-		const struct dc_stream_state *stream,
-		struct timing_generator *tg)
-{
-	int i;
-	bool underflow;
-
-	if (!dc->hwseq->wa.false_optc_underflow)
-		return;
-
-	underflow = tg->funcs->is_optc_underflow_occurred(tg);
-
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		struct pipe_ctx *old_pipe_ctx = &dc->current_state->res_ctx.pipe_ctx[i];
-
-		if (old_pipe_ctx->stream != stream)
-			continue;
-
-		dc->hwss.wait_for_mpcc_disconnect(dc, dc->res_pool, old_pipe_ctx);
-	}
-
-	if (tg->funcs->set_blank_data_double_buffer)
-		tg->funcs->set_blank_data_double_buffer(tg, true);
-
-	if (tg->funcs->is_optc_underflow_occurred(tg) && !underflow)
-		tg->funcs->clear_optc_underflow(tg);
-}
-
-enum dc_status dcn10_enable_stream_timing(
-		struct pipe_ctx *pipe_ctx,
-		struct dc_state *context,
-		struct dc *dc)
-{
-	struct dc_stream_state *stream = pipe_ctx->stream;
-	enum dc_color_space color_space;
-	struct tg_color black_color = {0};
-
-	/* by upper caller loop, pipe0 is parent pipe and be called first.
-	 * back end is set up by for pipe0. Other children pipe share back end
-	 * with pipe 0. No program is needed.
-	 */
-	if (pipe_ctx->top_pipe != NULL)
-		return DC_OK;
-
-	/* TODO check if timing_changed, disable stream if timing changed */
-
-	/* HW program guide assume display already disable
-	 * by unplug sequence. OTG assume stop.
-	 */
-	pipe_ctx->stream_res.tg->funcs->enable_optc_clock(pipe_ctx->stream_res.tg, true);
-
-	if (false == pipe_ctx->clock_source->funcs->program_pix_clk(
-			pipe_ctx->clock_source,
-			&pipe_ctx->stream_res.pix_clk_params,
-			&pipe_ctx->pll_settings)) {
-		BREAK_TO_DEBUGGER();
-		return DC_ERROR_UNEXPECTED;
-	}
-
-	pipe_ctx->stream_res.tg->funcs->program_timing(
-			pipe_ctx->stream_res.tg,
-			&stream->timing,
-			pipe_ctx->pipe_dlg_param.vready_offset,
-			pipe_ctx->pipe_dlg_param.vstartup_start,
-			pipe_ctx->pipe_dlg_param.vupdate_offset,
-			pipe_ctx->pipe_dlg_param.vupdate_width,
-			pipe_ctx->stream->signal,
-			true);
-
-#if 0 /* move to after enable_crtc */
-	/* TODO: OPP FMT, ABM. etc. should be done here. */
-	/* or FPGA now. instance 0 only. TODO: move to opp.c */
-
-	inst_offset = reg_offsets[pipe_ctx->stream_res.tg->inst].fmt;
-
-	pipe_ctx->stream_res.opp->funcs->opp_program_fmt(
-				pipe_ctx->stream_res.opp,
-				&stream->bit_depth_params,
-				&stream->clamping);
-#endif
-	/* program otg blank color */
-	color_space = stream->output_color_space;
-	color_space_to_black_color(dc, color_space, &black_color);
-
-	/*
-	 * The way 420 is packed, 2 channels carry Y component, 1 channel
-	 * alternate between Cb and Cr, so both channels need the pixel
-	 * value for Y
-	 */
-	if (stream->timing.pixel_encoding == PIXEL_ENCODING_YCBCR420)
-		black_color.color_r_cr = black_color.color_g_y;
-
-	if (pipe_ctx->stream_res.tg->funcs->set_blank_color)
-		pipe_ctx->stream_res.tg->funcs->set_blank_color(
-				pipe_ctx->stream_res.tg,
-				&black_color);
-
-	if (pipe_ctx->stream_res.tg->funcs->is_blanked &&
-			!pipe_ctx->stream_res.tg->funcs->is_blanked(pipe_ctx->stream_res.tg)) {
-		pipe_ctx->stream_res.tg->funcs->set_blank(pipe_ctx->stream_res.tg, true);
-		hwss_wait_for_blank_complete(pipe_ctx->stream_res.tg);
-		false_optc_underflow_wa(dc, pipe_ctx->stream, pipe_ctx->stream_res.tg);
-	}
-
-	/* VTG is  within DCHUB command block. DCFCLK is always on */
-	if (false == pipe_ctx->stream_res.tg->funcs->enable_crtc(pipe_ctx->stream_res.tg)) {
-		BREAK_TO_DEBUGGER();
-		return DC_ERROR_UNEXPECTED;
-	}
-
-	/* TODO program crtc source select for non-virtual signal*/
-	/* TODO program FMT */
-	/* TODO setup link_enc */
-	/* TODO set stream attributes */
-	/* TODO program audio */
-	/* TODO enable stream if timing changed */
-	/* TODO unblank stream if DP */
-
-	return DC_OK;
-}
-
-static void dcn10_reset_back_end_for_pipe(
-		struct dc *dc,
-		struct pipe_ctx *pipe_ctx,
-		struct dc_state *context)
-{
-	int i;
-	struct dc_link *link;
-	DC_LOGGER_INIT(dc->ctx->logger);
-	if (pipe_ctx->stream_res.stream_enc == NULL) {
-		pipe_ctx->stream = NULL;
-		return;
-	}
-
-	if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) {
-		link = pipe_ctx->stream->link;
-		/* DPMS may already disable or */
-		/* dpms_off status is incorrect due to fastboot
-		 * feature. When system resume from S4 with second
-		 * screen only, the dpms_off would be true but
-		 * VBIOS lit up eDP, so check link status too.
-		 */
-		if (!pipe_ctx->stream->dpms_off || link->link_status.link_active)
-			core_link_disable_stream(pipe_ctx);
-		else if (pipe_ctx->stream_res.audio)
-			dc->hwss.disable_audio_stream(pipe_ctx);
-
-		if (pipe_ctx->stream_res.audio) {
-			/*disable az_endpoint*/
-			pipe_ctx->stream_res.audio->funcs->az_disable(pipe_ctx->stream_res.audio);
-
-			/*free audio*/
-			if (dc->caps.dynamic_audio == true) {
-				/*we have to dynamic arbitrate the audio endpoints*/
-				/*we free the resource, need reset is_audio_acquired*/
-				update_audio_usage(&dc->current_state->res_ctx, dc->res_pool,
-						pipe_ctx->stream_res.audio, false);
-				pipe_ctx->stream_res.audio = NULL;
-			}
-		}
-	}
-
-	/* by upper caller loop, parent pipe: pipe0, will be reset last.
-	 * back end share by all pipes and will be disable only when disable
-	 * parent pipe.
-	 */
-	if (pipe_ctx->top_pipe == NULL) {
-
-		if (pipe_ctx->stream_res.abm)
-			dc->hwss.set_abm_immediate_disable(pipe_ctx);
-
-		pipe_ctx->stream_res.tg->funcs->disable_crtc(pipe_ctx->stream_res.tg);
-
-		pipe_ctx->stream_res.tg->funcs->enable_optc_clock(pipe_ctx->stream_res.tg, false);
-		if (pipe_ctx->stream_res.tg->funcs->set_drr)
-			pipe_ctx->stream_res.tg->funcs->set_drr(
-					pipe_ctx->stream_res.tg, NULL);
-	}
-
-	for (i = 0; i < dc->res_pool->pipe_count; i++)
-		if (&dc->current_state->res_ctx.pipe_ctx[i] == pipe_ctx)
-			break;
-
-	if (i == dc->res_pool->pipe_count)
-		return;
-
-	pipe_ctx->stream = NULL;
-	DC_LOG_DEBUG("Reset back end for pipe %d, tg:%d\n",
-					pipe_ctx->pipe_idx, pipe_ctx->stream_res.tg->inst);
-}
-
-static bool dcn10_hw_wa_force_recovery(struct dc *dc)
-{
-	struct hubp *hubp ;
-	unsigned int i;
-	bool need_recover = true;
-
-	if (!dc->debug.recovery_enabled)
-		return false;
-
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		struct pipe_ctx *pipe_ctx =
-			&dc->current_state->res_ctx.pipe_ctx[i];
-		if (pipe_ctx != NULL) {
-			hubp = pipe_ctx->plane_res.hubp;
-			if (hubp != NULL && hubp->funcs->hubp_get_underflow_status) {
-				if (hubp->funcs->hubp_get_underflow_status(hubp) != 0) {
-					/* one pipe underflow, we will reset all the pipes*/
-					need_recover = true;
-				}
-			}
-		}
-	}
-	if (!need_recover)
-		return false;
-	/*
-	DCHUBP_CNTL:HUBP_BLANK_EN=1
-	DCHUBBUB_SOFT_RESET:DCHUBBUB_GLOBAL_SOFT_RESET=1
-	DCHUBP_CNTL:HUBP_DISABLE=1
-	DCHUBP_CNTL:HUBP_DISABLE=0
-	DCHUBBUB_SOFT_RESET:DCHUBBUB_GLOBAL_SOFT_RESET=0
-	DCSURF_PRIMARY_SURFACE_ADDRESS
-	DCHUBP_CNTL:HUBP_BLANK_EN=0
-	*/
-
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		struct pipe_ctx *pipe_ctx =
-			&dc->current_state->res_ctx.pipe_ctx[i];
-		if (pipe_ctx != NULL) {
-			hubp = pipe_ctx->plane_res.hubp;
-			/*DCHUBP_CNTL:HUBP_BLANK_EN=1*/
-			if (hubp != NULL && hubp->funcs->set_hubp_blank_en)
-				hubp->funcs->set_hubp_blank_en(hubp, true);
-		}
-	}
-	/*DCHUBBUB_SOFT_RESET:DCHUBBUB_GLOBAL_SOFT_RESET=1*/
-	hubbub1_soft_reset(dc->res_pool->hubbub, true);
-
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		struct pipe_ctx *pipe_ctx =
-			&dc->current_state->res_ctx.pipe_ctx[i];
-		if (pipe_ctx != NULL) {
-			hubp = pipe_ctx->plane_res.hubp;
-			/*DCHUBP_CNTL:HUBP_DISABLE=1*/
-			if (hubp != NULL && hubp->funcs->hubp_disable_control)
-				hubp->funcs->hubp_disable_control(hubp, true);
-		}
-	}
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		struct pipe_ctx *pipe_ctx =
-			&dc->current_state->res_ctx.pipe_ctx[i];
-		if (pipe_ctx != NULL) {
-			hubp = pipe_ctx->plane_res.hubp;
-			/*DCHUBP_CNTL:HUBP_DISABLE=0*/
-			if (hubp != NULL && hubp->funcs->hubp_disable_control)
-				hubp->funcs->hubp_disable_control(hubp, true);
-		}
-	}
-	/*DCHUBBUB_SOFT_RESET:DCHUBBUB_GLOBAL_SOFT_RESET=0*/
-	hubbub1_soft_reset(dc->res_pool->hubbub, false);
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		struct pipe_ctx *pipe_ctx =
-			&dc->current_state->res_ctx.pipe_ctx[i];
-		if (pipe_ctx != NULL) {
-			hubp = pipe_ctx->plane_res.hubp;
-			/*DCHUBP_CNTL:HUBP_BLANK_EN=0*/
-			if (hubp != NULL && hubp->funcs->set_hubp_blank_en)
-				hubp->funcs->set_hubp_blank_en(hubp, true);
-		}
-	}
-	return true;
-
-}
-
-void dcn10_verify_allow_pstate_change_high(struct dc *dc)
-{
-	static bool should_log_hw_state; /* prevent hw state log by default */
-
-	if (!hubbub1_verify_allow_pstate_change_high(dc->res_pool->hubbub)) {
-		int i = 0;
-
-		if (should_log_hw_state)
-			dcn10_log_hw_state(dc, NULL);
-
-		TRACE_DC_PIPE_STATE(pipe_ctx, i, MAX_PIPES);
-		BREAK_TO_DEBUGGER();
-		if (dcn10_hw_wa_force_recovery(dc)) {
-		/*check again*/
-			if (!hubbub1_verify_allow_pstate_change_high(dc->res_pool->hubbub))
-				BREAK_TO_DEBUGGER();
-		}
-	}
-}
-
-/* trigger HW to start disconnect plane from stream on the next vsync */
-void dcn10_plane_atomic_disconnect(struct dc *dc, struct pipe_ctx *pipe_ctx)
-{
-	struct dce_hwseq *hws = dc->hwseq;
-	struct hubp *hubp = pipe_ctx->plane_res.hubp;
-	int dpp_id = pipe_ctx->plane_res.dpp->inst;
-	struct mpc *mpc = dc->res_pool->mpc;
-	struct mpc_tree *mpc_tree_params;
-	struct mpcc *mpcc_to_remove = NULL;
-	struct output_pixel_processor *opp = pipe_ctx->stream_res.opp;
-
-	mpc_tree_params = &(opp->mpc_tree_params);
-	mpcc_to_remove = mpc->funcs->get_mpcc_for_dpp(mpc_tree_params, dpp_id);
-
-	/*Already reset*/
-	if (mpcc_to_remove == NULL)
-		return;
-
-	mpc->funcs->remove_mpcc(mpc, mpc_tree_params, mpcc_to_remove);
-	if (opp != NULL)
-		opp->mpcc_disconnect_pending[pipe_ctx->plane_res.mpcc_inst] = true;
-
-	dc->optimized_required = true;
-
-	if (hubp->funcs->hubp_disconnect)
-		hubp->funcs->hubp_disconnect(hubp);
-
-	if (dc->debug.sanity_checks)
-		hws->funcs.verify_allow_pstate_change_high(dc);
-}
-
-/**
- * dcn10_plane_atomic_power_down - Power down plane components.
- *
- * @dc: dc struct reference. used for grab hwseq.
- * @dpp: dpp struct reference.
- * @hubp: hubp struct reference.
- *
- * Keep in mind that this operation requires a power gate configuration;
- * however, requests for switch power gate are precisely controlled to avoid
- * problems. For this reason, power gate request is usually disabled. This
- * function first needs to enable the power gate request before disabling DPP
- * and HUBP. Finally, it disables the power gate request again.
- */
-void dcn10_plane_atomic_power_down(struct dc *dc,
-		struct dpp *dpp,
-		struct hubp *hubp)
-{
-	struct dce_hwseq *hws = dc->hwseq;
-	DC_LOGGER_INIT(dc->ctx->logger);
-
-	if (REG(DC_IP_REQUEST_CNTL)) {
-		REG_SET(DC_IP_REQUEST_CNTL, 0,
-				IP_REQUEST_EN, 1);
-
-		if (hws->funcs.dpp_pg_control)
-			hws->funcs.dpp_pg_control(hws, dpp->inst, false);
-
-		if (hws->funcs.hubp_pg_control)
-			hws->funcs.hubp_pg_control(hws, hubp->inst, false);
-
-		dpp->funcs->dpp_reset(dpp);
-		REG_SET(DC_IP_REQUEST_CNTL, 0,
-				IP_REQUEST_EN, 0);
-		DC_LOG_DEBUG(
-				"Power gated front end %d\n", hubp->inst);
-	}
-}
-
-/* disable HW used by plane.
- * note:  cannot disable until disconnect is complete
- */
-void dcn10_plane_atomic_disable(struct dc *dc, struct pipe_ctx *pipe_ctx)
-{
-	struct dce_hwseq *hws = dc->hwseq;
-	struct hubp *hubp = pipe_ctx->plane_res.hubp;
-	struct dpp *dpp = pipe_ctx->plane_res.dpp;
-	int opp_id = hubp->opp_id;
-
-	dc->hwss.wait_for_mpcc_disconnect(dc, dc->res_pool, pipe_ctx);
-
-	hubp->funcs->hubp_clk_cntl(hubp, false);
-
-	dpp->funcs->dpp_dppclk_control(dpp, false, false);
-
-	if (opp_id != 0xf && pipe_ctx->stream_res.opp->mpc_tree_params.opp_list == NULL)
-		pipe_ctx->stream_res.opp->funcs->opp_pipe_clock_control(
-				pipe_ctx->stream_res.opp,
-				false);
-
-	hubp->power_gated = true;
-	dc->optimized_required = false; /* We're powering off, no need to optimize */
-
-	hws->funcs.plane_atomic_power_down(dc,
-			pipe_ctx->plane_res.dpp,
-			pipe_ctx->plane_res.hubp);
-
-	pipe_ctx->stream = NULL;
-	memset(&pipe_ctx->stream_res, 0, sizeof(pipe_ctx->stream_res));
-	memset(&pipe_ctx->plane_res, 0, sizeof(pipe_ctx->plane_res));
-	pipe_ctx->top_pipe = NULL;
-	pipe_ctx->bottom_pipe = NULL;
-	pipe_ctx->plane_state = NULL;
-}
-
-void dcn10_disable_plane(struct dc *dc, struct pipe_ctx *pipe_ctx)
-{
-	struct dce_hwseq *hws = dc->hwseq;
-	DC_LOGGER_INIT(dc->ctx->logger);
-
-	if (!pipe_ctx->plane_res.hubp || pipe_ctx->plane_res.hubp->power_gated)
-		return;
-
-	hws->funcs.plane_atomic_disable(dc, pipe_ctx);
-
-	apply_DEGVIDCN10_253_wa(dc);
-
-	DC_LOG_DC("Power down front end %d\n",
-					pipe_ctx->pipe_idx);
-}
-
-void dcn10_init_pipes(struct dc *dc, struct dc_state *context)
-{
-	int i;
-	struct dce_hwseq *hws = dc->hwseq;
-	bool can_apply_seamless_boot = false;
-
-	for (i = 0; i < context->stream_count; i++) {
-		if (context->streams[i]->apply_seamless_boot_optimization) {
-			can_apply_seamless_boot = true;
-			break;
-		}
-	}
-
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		struct timing_generator *tg = dc->res_pool->timing_generators[i];
-		struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i];
-
-		/* There is assumption that pipe_ctx is not mapping irregularly
-		 * to non-preferred front end. If pipe_ctx->stream is not NULL,
-		 * we will use the pipe, so don't disable
-		 */
-		if (pipe_ctx->stream != NULL && can_apply_seamless_boot)
-			continue;
-
-		/* Blank controller using driver code instead of
-		 * command table.
-		 */
-		if (tg->funcs->is_tg_enabled(tg)) {
-			if (hws->funcs.init_blank != NULL) {
-				hws->funcs.init_blank(dc, tg);
-				tg->funcs->lock(tg);
-			} else {
-				tg->funcs->lock(tg);
-				tg->funcs->set_blank(tg, true);
-				hwss_wait_for_blank_complete(tg);
-			}
-		}
-	}
-
-	/* num_opp will be equal to number of mpcc */
-	for (i = 0; i < dc->res_pool->res_cap->num_opp; i++) {
-		struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i];
-
-		/* Cannot reset the MPC mux if seamless boot */
-		if (pipe_ctx->stream != NULL && can_apply_seamless_boot)
-			continue;
-
-		dc->res_pool->mpc->funcs->mpc_init_single_inst(
-				dc->res_pool->mpc, i);
-	}
-
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		struct timing_generator *tg = dc->res_pool->timing_generators[i];
-		struct hubp *hubp = dc->res_pool->hubps[i];
-		struct dpp *dpp = dc->res_pool->dpps[i];
-		struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i];
-
-		/* There is assumption that pipe_ctx is not mapping irregularly
-		 * to non-preferred front end. If pipe_ctx->stream is not NULL,
-		 * we will use the pipe, so don't disable
-		 */
-		if (can_apply_seamless_boot &&
-			pipe_ctx->stream != NULL &&
-			pipe_ctx->stream_res.tg->funcs->is_tg_enabled(
-				pipe_ctx->stream_res.tg)) {
-			// Enable double buffering for OTG_BLANK no matter if
-			// seamless boot is enabled or not to suppress global sync
-			// signals when OTG blanked. This is to prevent pipe from
-			// requesting data while in PSR.
-			tg->funcs->tg_init(tg);
-			hubp->power_gated = true;
-			continue;
-		}
-
-		/* Disable on the current state so the new one isn't cleared. */
-		pipe_ctx = &dc->current_state->res_ctx.pipe_ctx[i];
-
-		dpp->funcs->dpp_reset(dpp);
-
-		pipe_ctx->stream_res.tg = tg;
-		pipe_ctx->pipe_idx = i;
-
-		pipe_ctx->plane_res.hubp = hubp;
-		pipe_ctx->plane_res.dpp = dpp;
-		pipe_ctx->plane_res.mpcc_inst = dpp->inst;
-		hubp->mpcc_id = dpp->inst;
-		hubp->opp_id = OPP_ID_INVALID;
-		hubp->power_gated = false;
-
-		dc->res_pool->opps[i]->mpc_tree_params.opp_id = dc->res_pool->opps[i]->inst;
-		dc->res_pool->opps[i]->mpc_tree_params.opp_list = NULL;
-		dc->res_pool->opps[i]->mpcc_disconnect_pending[pipe_ctx->plane_res.mpcc_inst] = true;
-		pipe_ctx->stream_res.opp = dc->res_pool->opps[i];
-
-		hws->funcs.plane_atomic_disconnect(dc, pipe_ctx);
-
-		if (tg->funcs->is_tg_enabled(tg))
-			tg->funcs->unlock(tg);
-
-		dc->hwss.disable_plane(dc, pipe_ctx);
-
-		pipe_ctx->stream_res.tg = NULL;
-		pipe_ctx->plane_res.hubp = NULL;
-
-		tg->funcs->tg_init(tg);
-	}
-}
-
-void dcn10_init_hw(struct dc *dc)
-{
-	int i, j;
-	struct abm *abm = dc->res_pool->abm;
-	struct dmcu *dmcu = dc->res_pool->dmcu;
-	struct dce_hwseq *hws = dc->hwseq;
-	struct dc_bios *dcb = dc->ctx->dc_bios;
-	struct resource_pool *res_pool = dc->res_pool;
-	uint32_t backlight = MAX_BACKLIGHT_LEVEL;
-	bool   is_optimized_init_done = false;
-
-	if (dc->clk_mgr && dc->clk_mgr->funcs->init_clocks)
-		dc->clk_mgr->funcs->init_clocks(dc->clk_mgr);
-
-	/* Align bw context with hw config when system resume. */
-	if (dc->clk_mgr->clks.dispclk_khz != 0 && dc->clk_mgr->clks.dppclk_khz != 0) {
-		dc->current_state->bw_ctx.bw.dcn.clk.dispclk_khz = dc->clk_mgr->clks.dispclk_khz;
-		dc->current_state->bw_ctx.bw.dcn.clk.dppclk_khz = dc->clk_mgr->clks.dppclk_khz;
-	}
-
-	// Initialize the dccg
-	if (dc->res_pool->dccg && dc->res_pool->dccg->funcs->dccg_init)
-		dc->res_pool->dccg->funcs->dccg_init(res_pool->dccg);
-
-	if (IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) {
-
-		REG_WRITE(REFCLK_CNTL, 0);
-		REG_UPDATE(DCHUBBUB_GLOBAL_TIMER_CNTL, DCHUBBUB_GLOBAL_TIMER_ENABLE, 1);
-		REG_WRITE(DIO_MEM_PWR_CTRL, 0);
-
-		if (!dc->debug.disable_clock_gate) {
-			/* enable all DCN clock gating */
-			REG_WRITE(DCCG_GATE_DISABLE_CNTL, 0);
-
-			REG_WRITE(DCCG_GATE_DISABLE_CNTL2, 0);
-
-			REG_UPDATE(DCFCLK_CNTL, DCFCLK_GATE_DIS, 0);
-		}
-
-		//Enable ability to power gate / don't force power on permanently
-		if (hws->funcs.enable_power_gating_plane)
-			hws->funcs.enable_power_gating_plane(hws, true);
-
-		return;
-	}
-
-	if (!dcb->funcs->is_accelerated_mode(dcb))
-		hws->funcs.disable_vga(dc->hwseq);
-
-	hws->funcs.bios_golden_init(dc);
-
-	if (dc->ctx->dc_bios->fw_info_valid) {
-		res_pool->ref_clocks.xtalin_clock_inKhz =
-				dc->ctx->dc_bios->fw_info.pll_info.crystal_frequency;
-
-		if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) {
-			if (res_pool->dccg && res_pool->hubbub) {
-
-				(res_pool->dccg->funcs->get_dccg_ref_freq)(res_pool->dccg,
-						dc->ctx->dc_bios->fw_info.pll_info.crystal_frequency,
-						&res_pool->ref_clocks.dccg_ref_clock_inKhz);
-
-				(res_pool->hubbub->funcs->get_dchub_ref_freq)(res_pool->hubbub,
-						res_pool->ref_clocks.dccg_ref_clock_inKhz,
-						&res_pool->ref_clocks.dchub_ref_clock_inKhz);
-			} else {
-				// Not all ASICs have DCCG sw component
-				res_pool->ref_clocks.dccg_ref_clock_inKhz =
-						res_pool->ref_clocks.xtalin_clock_inKhz;
-				res_pool->ref_clocks.dchub_ref_clock_inKhz =
-						res_pool->ref_clocks.xtalin_clock_inKhz;
-			}
-		}
-	} else
-		ASSERT_CRITICAL(false);
-
-	for (i = 0; i < dc->link_count; i++) {
-		/* Power up AND update implementation according to the
-		 * required signal (which may be different from the
-		 * default signal on connector).
-		 */
-		struct dc_link *link = dc->links[i];
-
-		if (!is_optimized_init_done)
-			link->link_enc->funcs->hw_init(link->link_enc);
-
-		/* Check for enabled DIG to identify enabled display */
-		if (link->link_enc->funcs->is_dig_enabled &&
-			link->link_enc->funcs->is_dig_enabled(link->link_enc))
-			link->link_status.link_active = true;
-	}
-
-	/* Power gate DSCs */
-	if (!is_optimized_init_done) {
-		for (i = 0; i < res_pool->res_cap->num_dsc; i++)
-			if (hws->funcs.dsc_pg_control != NULL)
-				hws->funcs.dsc_pg_control(hws, res_pool->dscs[i]->inst, false);
-	}
-
-	/* Enable outbox notification feature of dmub */
-	if (dc->debug.enable_dmub_aux_for_legacy_ddc)
-		dmub_enable_outbox_notification(dc);
-
-	/* we want to turn off all dp displays before doing detection */
-	if (dc->config.power_down_display_on_boot) {
-		uint8_t dpcd_power_state = '\0';
-		enum dc_status status = DC_ERROR_UNEXPECTED;
-
-		for (i = 0; i < dc->link_count; i++) {
-			if (dc->links[i]->connector_signal != SIGNAL_TYPE_DISPLAY_PORT)
-				continue;
-
-			/* DP 2.0 requires that LTTPR Caps be read first */
-			dp_retrieve_lttpr_cap(dc->links[i]);
-
-			/*
-			 * If any of the displays are lit up turn them off.
-			 * The reason is that some MST hubs cannot be turned off
-			 * completely until we tell them to do so.
-			 * If not turned off, then displays connected to MST hub
-			 * won't light up.
-			 */
-			status = core_link_read_dpcd(dc->links[i], DP_SET_POWER,
-							&dpcd_power_state, sizeof(dpcd_power_state));
-			if (status == DC_OK && dpcd_power_state == DP_POWER_STATE_D0) {
-				/* blank dp stream before power off receiver*/
-				if (dc->links[i]->link_enc->funcs->get_dig_frontend) {
-					unsigned int fe = dc->links[i]->link_enc->funcs->get_dig_frontend(dc->links[i]->link_enc);
-
-					for (j = 0; j < dc->res_pool->stream_enc_count; j++) {
-						if (fe == dc->res_pool->stream_enc[j]->id) {
-							dc->res_pool->stream_enc[j]->funcs->dp_blank(dc->links[i],
-										dc->res_pool->stream_enc[j]);
-							break;
-						}
-					}
-				}
-				dp_receiver_power_ctrl(dc->links[i], false);
-			}
-		}
-	}
-
-	/* If taking control over from VBIOS, we may want to optimize our first
-	 * mode set, so we need to skip powering down pipes until we know which
-	 * pipes we want to use.
-	 * Otherwise, if taking control is not possible, we need to power
-	 * everything down.
-	 */
-	if (dcb->funcs->is_accelerated_mode(dcb) || dc->config.power_down_display_on_boot) {
-		if (!is_optimized_init_done) {
-			hws->funcs.init_pipes(dc, dc->current_state);
-			if (dc->res_pool->hubbub->funcs->allow_self_refresh_control)
-				dc->res_pool->hubbub->funcs->allow_self_refresh_control(dc->res_pool->hubbub,
-						!dc->res_pool->hubbub->ctx->dc->debug.disable_stutter);
-		}
-	}
-
-	if (!is_optimized_init_done) {
-
-		for (i = 0; i < res_pool->audio_count; i++) {
-			struct audio *audio = res_pool->audios[i];
-
-			audio->funcs->hw_init(audio);
-		}
-
-		for (i = 0; i < dc->link_count; i++) {
-			struct dc_link *link = dc->links[i];
-
-			if (link->panel_cntl)
-				backlight = link->panel_cntl->funcs->hw_init(link->panel_cntl);
-		}
-
-		if (abm != NULL)
-			abm->funcs->abm_init(abm, backlight);
-
-		if (dmcu != NULL && !dmcu->auto_load_dmcu)
-			dmcu->funcs->dmcu_init(dmcu);
-	}
-
-	if (abm != NULL && dmcu != NULL)
-		abm->dmcu_is_running = dmcu->funcs->is_dmcu_initialized(dmcu);
-
-	/* power AFMT HDMI memory TODO: may move to dis/en output save power*/
-	if (!is_optimized_init_done)
-		REG_WRITE(DIO_MEM_PWR_CTRL, 0);
-
-	if (!dc->debug.disable_clock_gate) {
-		/* enable all DCN clock gating */
-		REG_WRITE(DCCG_GATE_DISABLE_CNTL, 0);
-
-		REG_WRITE(DCCG_GATE_DISABLE_CNTL2, 0);
-
-		REG_UPDATE(DCFCLK_CNTL, DCFCLK_GATE_DIS, 0);
-	}
-	if (hws->funcs.enable_power_gating_plane)
-		hws->funcs.enable_power_gating_plane(dc->hwseq, true);
-
-	if (dc->clk_mgr->funcs->notify_wm_ranges)
-		dc->clk_mgr->funcs->notify_wm_ranges(dc->clk_mgr);
-}
-
-/* In headless boot cases, DIG may be turned
- * on which causes HW/SW discrepancies.
- * To avoid this, power down hardware on boot
- * if DIG is turned on
- */
-void dcn10_power_down_on_boot(struct dc *dc)
-{
-	struct dc_link *edp_links[MAX_NUM_EDP];
-	struct dc_link *edp_link = NULL;
-	int edp_num;
-	int i = 0;
-
-	get_edp_links(dc, edp_links, &edp_num);
-	if (edp_num)
-		edp_link = edp_links[0];
-
-	if (edp_link && edp_link->link_enc->funcs->is_dig_enabled &&
-			edp_link->link_enc->funcs->is_dig_enabled(edp_link->link_enc) &&
-			dc->hwseq->funcs.edp_backlight_control &&
-			dc->hwss.power_down &&
-			dc->hwss.edp_power_control) {
-		dc->hwseq->funcs.edp_backlight_control(edp_link, false);
-		dc->hwss.power_down(dc);
-		dc->hwss.edp_power_control(edp_link, false);
-	} else {
-		for (i = 0; i < dc->link_count; i++) {
-			struct dc_link *link = dc->links[i];
-
-			if (link->link_enc && link->link_enc->funcs->is_dig_enabled &&
-					link->link_enc->funcs->is_dig_enabled(link->link_enc) &&
-					dc->hwss.power_down) {
-				dc->hwss.power_down(dc);
-				break;
-			}
-
-		}
-	}
-
-	/*
-	 * Call update_clocks with empty context
-	 * to send DISPLAY_OFF
-	 * Otherwise DISPLAY_OFF may not be asserted
-	 */
-	if (dc->clk_mgr->funcs->set_low_power_state)
-		dc->clk_mgr->funcs->set_low_power_state(dc->clk_mgr);
-}
-
-void dcn10_reset_hw_ctx_wrap(
-		struct dc *dc,
-		struct dc_state *context)
-{
-	int i;
-	struct dce_hwseq *hws = dc->hwseq;
-
-	/* Reset Back End*/
-	for (i = dc->res_pool->pipe_count - 1; i >= 0 ; i--) {
-		struct pipe_ctx *pipe_ctx_old =
-			&dc->current_state->res_ctx.pipe_ctx[i];
-		struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i];
-
-		if (!pipe_ctx_old->stream)
-			continue;
-
-		if (pipe_ctx_old->top_pipe)
-			continue;
-
-		if (!pipe_ctx->stream ||
-				pipe_need_reprogram(pipe_ctx_old, pipe_ctx)) {
-			struct clock_source *old_clk = pipe_ctx_old->clock_source;
-
-			dcn10_reset_back_end_for_pipe(dc, pipe_ctx_old, dc->current_state);
-			if (hws->funcs.enable_stream_gating)
-				hws->funcs.enable_stream_gating(dc, pipe_ctx_old);
-			if (old_clk)
-				old_clk->funcs->cs_power_down(old_clk);
-		}
-	}
-}
-
-static bool patch_address_for_sbs_tb_stereo(
-		struct pipe_ctx *pipe_ctx, PHYSICAL_ADDRESS_LOC *addr)
-{
-	struct dc_plane_state *plane_state = pipe_ctx->plane_state;
-	bool sec_split = pipe_ctx->top_pipe &&
-			pipe_ctx->top_pipe->plane_state == pipe_ctx->plane_state;
-	if (sec_split && plane_state->address.type == PLN_ADDR_TYPE_GRPH_STEREO &&
-		(pipe_ctx->stream->timing.timing_3d_format ==
-		 TIMING_3D_FORMAT_SIDE_BY_SIDE ||
-		 pipe_ctx->stream->timing.timing_3d_format ==
-		 TIMING_3D_FORMAT_TOP_AND_BOTTOM)) {
-		*addr = plane_state->address.grph_stereo.left_addr;
-		plane_state->address.grph_stereo.left_addr =
-		plane_state->address.grph_stereo.right_addr;
-		return true;
-	} else {
-		if (pipe_ctx->stream->view_format != VIEW_3D_FORMAT_NONE &&
-			plane_state->address.type != PLN_ADDR_TYPE_GRPH_STEREO) {
-			plane_state->address.type = PLN_ADDR_TYPE_GRPH_STEREO;
-			plane_state->address.grph_stereo.right_addr =
-			plane_state->address.grph_stereo.left_addr;
-			plane_state->address.grph_stereo.right_meta_addr =
-			plane_state->address.grph_stereo.left_meta_addr;
-		}
-	}
-	return false;
-}
-
-void dcn10_update_plane_addr(const struct dc *dc, struct pipe_ctx *pipe_ctx)
-{
-	bool addr_patched = false;
-	PHYSICAL_ADDRESS_LOC addr;
-	struct dc_plane_state *plane_state = pipe_ctx->plane_state;
-
-	if (plane_state == NULL)
-		return;
-
-	addr_patched = patch_address_for_sbs_tb_stereo(pipe_ctx, &addr);
-
-	pipe_ctx->plane_res.hubp->funcs->hubp_program_surface_flip_and_addr(
-			pipe_ctx->plane_res.hubp,
-			&plane_state->address,
-			plane_state->flip_immediate);
-
-	plane_state->status.requested_address = plane_state->address;
-
-	if (plane_state->flip_immediate)
-		plane_state->status.current_address = plane_state->address;
-
-	if (addr_patched)
-		pipe_ctx->plane_state->address.grph_stereo.left_addr = addr;
-}
-
-bool dcn10_set_input_transfer_func(struct dc *dc, struct pipe_ctx *pipe_ctx,
-			const struct dc_plane_state *plane_state)
-{
-	struct dpp *dpp_base = pipe_ctx->plane_res.dpp;
-	const struct dc_transfer_func *tf = NULL;
-	bool result = true;
-
-	if (dpp_base == NULL)
-		return false;
-
-	if (plane_state->in_transfer_func)
-		tf = plane_state->in_transfer_func;
-
-	if (plane_state->gamma_correction &&
-		!dpp_base->ctx->dc->debug.always_use_regamma
-		&& !plane_state->gamma_correction->is_identity
-			&& dce_use_lut(plane_state->format))
-		dpp_base->funcs->dpp_program_input_lut(dpp_base, plane_state->gamma_correction);
-
-	if (tf == NULL)
-		dpp_base->funcs->dpp_set_degamma(dpp_base, IPP_DEGAMMA_MODE_BYPASS);
-	else if (tf->type == TF_TYPE_PREDEFINED) {
-		switch (tf->tf) {
-		case TRANSFER_FUNCTION_SRGB:
-			dpp_base->funcs->dpp_set_degamma(dpp_base, IPP_DEGAMMA_MODE_HW_sRGB);
-			break;
-		case TRANSFER_FUNCTION_BT709:
-			dpp_base->funcs->dpp_set_degamma(dpp_base, IPP_DEGAMMA_MODE_HW_xvYCC);
-			break;
-		case TRANSFER_FUNCTION_LINEAR:
-			dpp_base->funcs->dpp_set_degamma(dpp_base, IPP_DEGAMMA_MODE_BYPASS);
-			break;
-		case TRANSFER_FUNCTION_PQ:
-			dpp_base->funcs->dpp_set_degamma(dpp_base, IPP_DEGAMMA_MODE_USER_PWL);
-			cm_helper_translate_curve_to_degamma_hw_format(tf, &dpp_base->degamma_params);
-			dpp_base->funcs->dpp_program_degamma_pwl(dpp_base, &dpp_base->degamma_params);
-			result = true;
-			break;
-		default:
-			result = false;
-			break;
-		}
-	} else if (tf->type == TF_TYPE_BYPASS) {
-		dpp_base->funcs->dpp_set_degamma(dpp_base, IPP_DEGAMMA_MODE_BYPASS);
-	} else {
-		cm_helper_translate_curve_to_degamma_hw_format(tf,
-					&dpp_base->degamma_params);
-		dpp_base->funcs->dpp_program_degamma_pwl(dpp_base,
-				&dpp_base->degamma_params);
-		result = true;
-	}
-
-	return result;
-}
-
-#define MAX_NUM_HW_POINTS 0x200
-
-static void log_tf(struct dc_context *ctx,
-				struct dc_transfer_func *tf, uint32_t hw_points_num)
-{
-	// DC_LOG_GAMMA is default logging of all hw points
-	// DC_LOG_ALL_GAMMA logs all points, not only hw points
-	// DC_LOG_ALL_TF_POINTS logs all channels of the tf
-	int i = 0;
-
-	DC_LOGGER_INIT(ctx->logger);
-	DC_LOG_GAMMA("Gamma Correction TF");
-	DC_LOG_ALL_GAMMA("Logging all tf points...");
-	DC_LOG_ALL_TF_CHANNELS("Logging all channels...");
-
-	for (i = 0; i < hw_points_num; i++) {
-		DC_LOG_GAMMA("R\t%d\t%llu", i, tf->tf_pts.red[i].value);
-		DC_LOG_ALL_TF_CHANNELS("G\t%d\t%llu", i, tf->tf_pts.green[i].value);
-		DC_LOG_ALL_TF_CHANNELS("B\t%d\t%llu", i, tf->tf_pts.blue[i].value);
-	}
-
-	for (i = hw_points_num; i < MAX_NUM_HW_POINTS; i++) {
-		DC_LOG_ALL_GAMMA("R\t%d\t%llu", i, tf->tf_pts.red[i].value);
-		DC_LOG_ALL_TF_CHANNELS("G\t%d\t%llu", i, tf->tf_pts.green[i].value);
-		DC_LOG_ALL_TF_CHANNELS("B\t%d\t%llu", i, tf->tf_pts.blue[i].value);
-	}
-}
-
-bool dcn10_set_output_transfer_func(struct dc *dc, struct pipe_ctx *pipe_ctx,
-				const struct dc_stream_state *stream)
-{
-	struct dpp *dpp = pipe_ctx->plane_res.dpp;
-
-	if (dpp == NULL)
-		return false;
-
-	dpp->regamma_params.hw_points_num = GAMMA_HW_POINTS_NUM;
-
-	if (stream->out_transfer_func &&
-	    stream->out_transfer_func->type == TF_TYPE_PREDEFINED &&
-	    stream->out_transfer_func->tf == TRANSFER_FUNCTION_SRGB)
-		dpp->funcs->dpp_program_regamma_pwl(dpp, NULL, OPP_REGAMMA_SRGB);
-
-	/* dcn10_translate_regamma_to_hw_format takes 750us, only do it when full
-	 * update.
-	 */
-	else if (cm_helper_translate_curve_to_hw_format(
-			stream->out_transfer_func,
-			&dpp->regamma_params, false)) {
-		dpp->funcs->dpp_program_regamma_pwl(
-				dpp,
-				&dpp->regamma_params, OPP_REGAMMA_USER);
-	} else
-		dpp->funcs->dpp_program_regamma_pwl(dpp, NULL, OPP_REGAMMA_BYPASS);
-
-	if (stream != NULL && stream->ctx != NULL &&
-			stream->out_transfer_func != NULL) {
-		log_tf(stream->ctx,
-				stream->out_transfer_func,
-				dpp->regamma_params.hw_points_num);
-	}
-
-	return true;
-}
-
-void dcn10_pipe_control_lock(
-	struct dc *dc,
-	struct pipe_ctx *pipe,
-	bool lock)
-{
-	struct dce_hwseq *hws = dc->hwseq;
-
-	/* use TG master update lock to lock everything on the TG
-	 * therefore only top pipe need to lock
-	 */
-	if (!pipe || pipe->top_pipe)
-		return;
-
-	if (dc->debug.sanity_checks)
-		hws->funcs.verify_allow_pstate_change_high(dc);
-
-	if (lock)
-		pipe->stream_res.tg->funcs->lock(pipe->stream_res.tg);
-	else
-		pipe->stream_res.tg->funcs->unlock(pipe->stream_res.tg);
-
-	if (dc->debug.sanity_checks)
-		hws->funcs.verify_allow_pstate_change_high(dc);
-}
-
-/**
- * delay_cursor_until_vupdate() - Delay cursor update if too close to VUPDATE.
- *
- * Software keepout workaround to prevent cursor update locking from stalling
- * out cursor updates indefinitely or from old values from being retained in
- * the case where the viewport changes in the same frame as the cursor.
- *
- * The idea is to calculate the remaining time from VPOS to VUPDATE. If it's
- * too close to VUPDATE, then stall out until VUPDATE finishes.
- *
- * TODO: Optimize cursor programming to be once per frame before VUPDATE
- *       to avoid the need for this workaround.
- */
-static void delay_cursor_until_vupdate(struct dc *dc, struct pipe_ctx *pipe_ctx)
-{
-	struct dc_stream_state *stream = pipe_ctx->stream;
-	struct crtc_position position;
-	uint32_t vupdate_start, vupdate_end;
-	unsigned int lines_to_vupdate, us_to_vupdate, vpos;
-	unsigned int us_per_line, us_vupdate;
-
-	if (!dc->hwss.calc_vupdate_position || !dc->hwss.get_position)
-		return;
-
-	if (!pipe_ctx->stream_res.stream_enc || !pipe_ctx->stream_res.tg)
-		return;
-
-	dc->hwss.calc_vupdate_position(dc, pipe_ctx, &vupdate_start,
-				       &vupdate_end);
-
-	dc->hwss.get_position(&pipe_ctx, 1, &position);
-	vpos = position.vertical_count;
-
-	/* Avoid wraparound calculation issues */
-	vupdate_start += stream->timing.v_total;
-	vupdate_end += stream->timing.v_total;
-	vpos += stream->timing.v_total;
-
-	if (vpos <= vupdate_start) {
-		/* VPOS is in VACTIVE or back porch. */
-		lines_to_vupdate = vupdate_start - vpos;
-	} else if (vpos > vupdate_end) {
-		/* VPOS is in the front porch. */
-		return;
-	} else {
-		/* VPOS is in VUPDATE. */
-		lines_to_vupdate = 0;
-	}
-
-	/* Calculate time until VUPDATE in microseconds. */
-	us_per_line =
-		stream->timing.h_total * 10000u / stream->timing.pix_clk_100hz;
-	us_to_vupdate = lines_to_vupdate * us_per_line;
-
-	/* 70 us is a conservative estimate of cursor update time*/
-	if (us_to_vupdate > 70)
-		return;
-
-	/* Stall out until the cursor update completes. */
-	if (vupdate_end < vupdate_start)
-		vupdate_end += stream->timing.v_total;
-	us_vupdate = (vupdate_end - vupdate_start + 1) * us_per_line;
-	udelay(us_to_vupdate + us_vupdate);
-}
-
-void dcn10_cursor_lock(struct dc *dc, struct pipe_ctx *pipe, bool lock)
-{
-	/* cursor lock is per MPCC tree, so only need to lock one pipe per stream */
-	if (!pipe || pipe->top_pipe)
-		return;
-
-	/* Prevent cursor lock from stalling out cursor updates. */
-	if (lock)
-		delay_cursor_until_vupdate(dc, pipe);
-
-	if (pipe->stream && should_use_dmub_lock(pipe->stream->link)) {
-		union dmub_hw_lock_flags hw_locks = { 0 };
-		struct dmub_hw_lock_inst_flags inst_flags = { 0 };
-
-		hw_locks.bits.lock_cursor = 1;
-		inst_flags.opp_inst = pipe->stream_res.opp->inst;
-
-		dmub_hw_lock_mgr_cmd(dc->ctx->dmub_srv,
-					lock,
-					&hw_locks,
-					&inst_flags);
-	} else
-		dc->res_pool->mpc->funcs->cursor_lock(dc->res_pool->mpc,
-				pipe->stream_res.opp->inst, lock);
-}
-
-static bool wait_for_reset_trigger_to_occur(
-	struct dc_context *dc_ctx,
-	struct timing_generator *tg)
-{
-	bool rc = false;
-
-	/* To avoid endless loop we wait at most
-	 * frames_to_wait_on_triggered_reset frames for the reset to occur. */
-	const uint32_t frames_to_wait_on_triggered_reset = 10;
-	int i;
-
-	for (i = 0; i < frames_to_wait_on_triggered_reset; i++) {
-
-		if (!tg->funcs->is_counter_moving(tg)) {
-			DC_ERROR("TG counter is not moving!\n");
-			break;
-		}
-
-		if (tg->funcs->did_triggered_reset_occur(tg)) {
-			rc = true;
-			/* usually occurs at i=1 */
-			DC_SYNC_INFO("GSL: reset occurred at wait count: %d\n",
-					i);
-			break;
-		}
-
-		/* Wait for one frame. */
-		tg->funcs->wait_for_state(tg, CRTC_STATE_VACTIVE);
-		tg->funcs->wait_for_state(tg, CRTC_STATE_VBLANK);
-	}
-
-	if (false == rc)
-		DC_ERROR("GSL: Timeout on reset trigger!\n");
-
-	return rc;
-}
-
-uint64_t reduceSizeAndFraction(
-	uint64_t *numerator,
-	uint64_t *denominator,
-	bool checkUint32Bounary)
-{
-	int i;
-	bool ret = checkUint32Bounary == false;
-	uint64_t max_int32 = 0xffffffff;
-	uint64_t num, denom;
-	static const uint16_t prime_numbers[] = {
-		2, 3, 5, 7, 11, 13, 17, 19, 23, 29, 31, 37, 41, 43,
-		47, 53, 59, 61, 67, 71, 73, 79, 83, 89, 97, 101, 103,
-		107, 109, 113, 127, 131, 137, 139, 149, 151, 157, 163,
-		167, 173, 179, 181, 191, 193, 197, 199, 211, 223, 227,
-		229, 233, 239, 241, 251, 257, 263, 269, 271, 277, 281,
-		283, 293, 307, 311, 313, 317, 331, 337, 347, 349, 353,
-		359, 367, 373, 379, 383, 389, 397, 401, 409, 419, 421,
-		431, 433, 439, 443, 449, 457, 461, 463, 467, 479, 487,
-		491, 499, 503, 509, 521, 523, 541, 547, 557, 563, 569,
-		571, 577, 587, 593, 599, 601, 607, 613, 617, 619, 631,
-		641, 643, 647, 653, 659, 661, 673, 677, 683, 691, 701,
-		709, 719, 727, 733, 739, 743, 751, 757, 761, 769, 773,
-		787, 797, 809, 811, 821, 823, 827, 829, 839, 853, 857,
-		859, 863, 877, 881, 883, 887, 907, 911, 919, 929, 937,
-		941, 947, 953, 967, 971, 977, 983, 991, 997};
-	int count = ARRAY_SIZE(prime_numbers);
-
-	num = *numerator;
-	denom = *denominator;
-	for (i = 0; i < count; i++) {
-		uint32_t num_remainder, denom_remainder;
-		uint64_t num_result, denom_result;
-		if (checkUint32Bounary &&
-			num <= max_int32 && denom <= max_int32) {
-			ret = true;
-			break;
-		}
-		do {
-			num_result = div_u64_rem(num, prime_numbers[i], &num_remainder);
-			denom_result = div_u64_rem(denom, prime_numbers[i], &denom_remainder);
-			if (num_remainder == 0 && denom_remainder == 0) {
-				num = num_result;
-				denom = denom_result;
-			}
-		} while (num_remainder == 0 && denom_remainder == 0);
-	}
-	*numerator = num;
-	*denominator = denom;
-	return ret;
-}
-
-bool is_low_refresh_rate(struct pipe_ctx *pipe)
-{
-	uint32_t master_pipe_refresh_rate =
-		pipe->stream->timing.pix_clk_100hz * 100 /
-		pipe->stream->timing.h_total /
-		pipe->stream->timing.v_total;
-	return master_pipe_refresh_rate <= 30;
-}
-
-uint8_t get_clock_divider(struct pipe_ctx *pipe, bool account_low_refresh_rate)
-{
-	uint32_t clock_divider = 1;
-	uint32_t numpipes = 1;
-
-	if (account_low_refresh_rate && is_low_refresh_rate(pipe))
-		clock_divider *= 2;
-
-	if (pipe->stream_res.pix_clk_params.pixel_encoding == PIXEL_ENCODING_YCBCR420)
-		clock_divider *= 2;
-
-	while (pipe->next_odm_pipe) {
-		pipe = pipe->next_odm_pipe;
-		numpipes++;
-	}
-	clock_divider *= numpipes;
-
-	return clock_divider;
-}
-
-int dcn10_align_pixel_clocks(
-	struct dc *dc,
-	int group_size,
-	struct pipe_ctx *grouped_pipes[])
-{
-	struct dc_context *dc_ctx = dc->ctx;
-	int i, master = -1, embedded = -1;
-	struct dc_crtc_timing hw_crtc_timing[MAX_PIPES] = {0};
-	uint64_t phase[MAX_PIPES];
-	uint64_t modulo[MAX_PIPES];
-	unsigned int pclk;
-
-	uint32_t embedded_pix_clk_100hz;
-	uint16_t embedded_h_total;
-	uint16_t embedded_v_total;
-	bool clamshell_closed = false;
-	uint32_t dp_ref_clk_100hz =
-		dc->res_pool->dp_clock_source->ctx->dc->clk_mgr->dprefclk_khz*10;
-
-	if (dc->config.vblank_alignment_dto_params &&
-		dc->res_pool->dp_clock_source->funcs->override_dp_pix_clk) {
-		clamshell_closed =
-			(dc->config.vblank_alignment_dto_params >> 63);
-		embedded_h_total =
-			(dc->config.vblank_alignment_dto_params >> 32) & 0x7FFF;
-		embedded_v_total =
-			(dc->config.vblank_alignment_dto_params >> 48) & 0x7FFF;
-		embedded_pix_clk_100hz =
-			dc->config.vblank_alignment_dto_params & 0xFFFFFFFF;
-
-		for (i = 0; i < group_size; i++) {
-			grouped_pipes[i]->stream_res.tg->funcs->get_hw_timing(
-					grouped_pipes[i]->stream_res.tg,
-					&hw_crtc_timing[i]);
-			dc->res_pool->dp_clock_source->funcs->get_pixel_clk_frequency_100hz(
-				dc->res_pool->dp_clock_source,
-				grouped_pipes[i]->stream_res.tg->inst,
-				&pclk);
-			hw_crtc_timing[i].pix_clk_100hz = pclk;
-			if (dc_is_embedded_signal(
-					grouped_pipes[i]->stream->signal)) {
-				embedded = i;
-				master = i;
-				phase[i] = embedded_pix_clk_100hz*100;
-				modulo[i] = dp_ref_clk_100hz*100;
-			} else {
-
-				phase[i] = (uint64_t)embedded_pix_clk_100hz*
-					hw_crtc_timing[i].h_total*
-					hw_crtc_timing[i].v_total;
-				phase[i] = div_u64(phase[i], get_clock_divider(grouped_pipes[i], true));
-				modulo[i] = (uint64_t)dp_ref_clk_100hz*
-					embedded_h_total*
-					embedded_v_total;
-
-				if (reduceSizeAndFraction(&phase[i],
-						&modulo[i], true) == false) {
-					/*
-					 * this will help to stop reporting
-					 * this timing synchronizable
-					 */
-					DC_SYNC_INFO("Failed to reduce DTO parameters\n");
-					grouped_pipes[i]->stream->has_non_synchronizable_pclk = true;
-				}
-			}
-		}
-
-		for (i = 0; i < group_size; i++) {
-			if (i != embedded && !grouped_pipes[i]->stream->has_non_synchronizable_pclk) {
-				dc->res_pool->dp_clock_source->funcs->override_dp_pix_clk(
-					dc->res_pool->dp_clock_source,
-					grouped_pipes[i]->stream_res.tg->inst,
-					phase[i], modulo[i]);
-				dc->res_pool->dp_clock_source->funcs->get_pixel_clk_frequency_100hz(
-					dc->res_pool->dp_clock_source,
-					grouped_pipes[i]->stream_res.tg->inst, &pclk);
-					grouped_pipes[i]->stream->timing.pix_clk_100hz =
-						pclk*get_clock_divider(grouped_pipes[i], false);
-				if (master == -1)
-					master = i;
-			}
-		}
-
-	}
-	return master;
-}
-
-void dcn10_enable_vblanks_synchronization(
-	struct dc *dc,
-	int group_index,
-	int group_size,
-	struct pipe_ctx *grouped_pipes[])
-{
-	struct dc_context *dc_ctx = dc->ctx;
-	struct output_pixel_processor *opp;
-	struct timing_generator *tg;
-	int i, width, height, master;
-
-	for (i = 1; i < group_size; i++) {
-		opp = grouped_pipes[i]->stream_res.opp;
-		tg = grouped_pipes[i]->stream_res.tg;
-		tg->funcs->get_otg_active_size(tg, &width, &height);
-		if (opp->funcs->opp_program_dpg_dimensions)
-			opp->funcs->opp_program_dpg_dimensions(opp, width, 2*(height) + 1);
-	}
-
-	for (i = 0; i < group_size; i++) {
-		if (grouped_pipes[i]->stream == NULL)
-			continue;
-		grouped_pipes[i]->stream->vblank_synchronized = false;
-		grouped_pipes[i]->stream->has_non_synchronizable_pclk = false;
-	}
-
-	DC_SYNC_INFO("Aligning DP DTOs\n");
-
-	master = dcn10_align_pixel_clocks(dc, group_size, grouped_pipes);
-
-	DC_SYNC_INFO("Synchronizing VBlanks\n");
-
-	if (master >= 0) {
-		for (i = 0; i < group_size; i++) {
-			if (i != master && !grouped_pipes[i]->stream->has_non_synchronizable_pclk)
-			grouped_pipes[i]->stream_res.tg->funcs->align_vblanks(
-				grouped_pipes[master]->stream_res.tg,
-				grouped_pipes[i]->stream_res.tg,
-				grouped_pipes[master]->stream->timing.pix_clk_100hz,
-				grouped_pipes[i]->stream->timing.pix_clk_100hz,
-				get_clock_divider(grouped_pipes[master], false),
-				get_clock_divider(grouped_pipes[i], false));
-				grouped_pipes[i]->stream->vblank_synchronized = true;
-		}
-		grouped_pipes[master]->stream->vblank_synchronized = true;
-		DC_SYNC_INFO("Sync complete\n");
-	}
-
-	for (i = 1; i < group_size; i++) {
-		opp = grouped_pipes[i]->stream_res.opp;
-		tg = grouped_pipes[i]->stream_res.tg;
-		tg->funcs->get_otg_active_size(tg, &width, &height);
-		if (opp->funcs->opp_program_dpg_dimensions)
-			opp->funcs->opp_program_dpg_dimensions(opp, width, height);
-	}
-}
-
-void dcn10_enable_timing_synchronization(
-	struct dc *dc,
-	int group_index,
-	int group_size,
-	struct pipe_ctx *grouped_pipes[])
-{
-	struct dc_context *dc_ctx = dc->ctx;
-	struct output_pixel_processor *opp;
-	struct timing_generator *tg;
-	int i, width, height;
-
-	DC_SYNC_INFO("Setting up OTG reset trigger\n");
-
-	for (i = 1; i < group_size; i++) {
-		opp = grouped_pipes[i]->stream_res.opp;
-		tg = grouped_pipes[i]->stream_res.tg;
-		tg->funcs->get_otg_active_size(tg, &width, &height);
-		if (opp->funcs->opp_program_dpg_dimensions)
-			opp->funcs->opp_program_dpg_dimensions(opp, width, 2*(height) + 1);
-	}
-
-	for (i = 0; i < group_size; i++) {
-		if (grouped_pipes[i]->stream == NULL)
-			continue;
-		grouped_pipes[i]->stream->vblank_synchronized = false;
-	}
-
-	for (i = 1; i < group_size; i++)
-		grouped_pipes[i]->stream_res.tg->funcs->enable_reset_trigger(
-				grouped_pipes[i]->stream_res.tg,
-				grouped_pipes[0]->stream_res.tg->inst);
-
-	DC_SYNC_INFO("Waiting for trigger\n");
-
-	/* Need to get only check 1 pipe for having reset as all the others are
-	 * synchronized. Look at last pipe programmed to reset.
-	 */
-
-	wait_for_reset_trigger_to_occur(dc_ctx, grouped_pipes[1]->stream_res.tg);
-	for (i = 1; i < group_size; i++)
-		grouped_pipes[i]->stream_res.tg->funcs->disable_reset_trigger(
-				grouped_pipes[i]->stream_res.tg);
-
-	for (i = 1; i < group_size; i++) {
-		opp = grouped_pipes[i]->stream_res.opp;
-		tg = grouped_pipes[i]->stream_res.tg;
-		tg->funcs->get_otg_active_size(tg, &width, &height);
-		if (opp->funcs->opp_program_dpg_dimensions)
-			opp->funcs->opp_program_dpg_dimensions(opp, width, height);
-	}
-
-	DC_SYNC_INFO("Sync complete\n");
-}
-
-void dcn10_enable_per_frame_crtc_position_reset(
-	struct dc *dc,
-	int group_size,
-	struct pipe_ctx *grouped_pipes[])
-{
-	struct dc_context *dc_ctx = dc->ctx;
-	int i;
-
-	DC_SYNC_INFO("Setting up\n");
-	for (i = 0; i < group_size; i++)
-		if (grouped_pipes[i]->stream_res.tg->funcs->enable_crtc_reset)
-			grouped_pipes[i]->stream_res.tg->funcs->enable_crtc_reset(
-					grouped_pipes[i]->stream_res.tg,
-					0,
-					&grouped_pipes[i]->stream->triggered_crtc_reset);
-
-	DC_SYNC_INFO("Waiting for trigger\n");
-
-	for (i = 0; i < group_size; i++)
-		wait_for_reset_trigger_to_occur(dc_ctx, grouped_pipes[i]->stream_res.tg);
-
-	DC_SYNC_INFO("Multi-display sync is complete\n");
-}
-
-static void mmhub_read_vm_system_aperture_settings(struct dcn10_hubp *hubp1,
-		struct vm_system_aperture_param *apt,
-		struct dce_hwseq *hws)
-{
-	PHYSICAL_ADDRESS_LOC physical_page_number;
-	uint32_t logical_addr_low;
-	uint32_t logical_addr_high;
-
-	REG_GET(MC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_MSB,
-			PHYSICAL_PAGE_NUMBER_MSB, &physical_page_number.high_part);
-	REG_GET(MC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB,
-			PHYSICAL_PAGE_NUMBER_LSB, &physical_page_number.low_part);
-
-	REG_GET(MC_VM_SYSTEM_APERTURE_LOW_ADDR,
-			LOGICAL_ADDR, &logical_addr_low);
-
-	REG_GET(MC_VM_SYSTEM_APERTURE_HIGH_ADDR,
-			LOGICAL_ADDR, &logical_addr_high);
-
-	apt->sys_default.quad_part =  physical_page_number.quad_part << 12;
-	apt->sys_low.quad_part =  (int64_t)logical_addr_low << 18;
-	apt->sys_high.quad_part =  (int64_t)logical_addr_high << 18;
-}
-
-/* Temporary read settings, future will get values from kmd directly */
-static void mmhub_read_vm_context0_settings(struct dcn10_hubp *hubp1,
-		struct vm_context0_param *vm0,
-		struct dce_hwseq *hws)
-{
-	PHYSICAL_ADDRESS_LOC fb_base;
-	PHYSICAL_ADDRESS_LOC fb_offset;
-	uint32_t fb_base_value;
-	uint32_t fb_offset_value;
-
-	REG_GET(DCHUBBUB_SDPIF_FB_BASE, SDPIF_FB_BASE, &fb_base_value);
-	REG_GET(DCHUBBUB_SDPIF_FB_OFFSET, SDPIF_FB_OFFSET, &fb_offset_value);
-
-	REG_GET(VM_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32,
-			PAGE_DIRECTORY_ENTRY_HI32, &vm0->pte_base.high_part);
-	REG_GET(VM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32,
-			PAGE_DIRECTORY_ENTRY_LO32, &vm0->pte_base.low_part);
-
-	REG_GET(VM_CONTEXT0_PAGE_TABLE_START_ADDR_HI32,
-			LOGICAL_PAGE_NUMBER_HI4, &vm0->pte_start.high_part);
-	REG_GET(VM_CONTEXT0_PAGE_TABLE_START_ADDR_LO32,
-			LOGICAL_PAGE_NUMBER_LO32, &vm0->pte_start.low_part);
-
-	REG_GET(VM_CONTEXT0_PAGE_TABLE_END_ADDR_HI32,
-			LOGICAL_PAGE_NUMBER_HI4, &vm0->pte_end.high_part);
-	REG_GET(VM_CONTEXT0_PAGE_TABLE_END_ADDR_LO32,
-			LOGICAL_PAGE_NUMBER_LO32, &vm0->pte_end.low_part);
-
-	REG_GET(VM_L2_PROTECTION_FAULT_DEFAULT_ADDR_HI32,
-			PHYSICAL_PAGE_ADDR_HI4, &vm0->fault_default.high_part);
-	REG_GET(VM_L2_PROTECTION_FAULT_DEFAULT_ADDR_LO32,
-			PHYSICAL_PAGE_ADDR_LO32, &vm0->fault_default.low_part);
-
-	/*
-	 * The values in VM_CONTEXT0_PAGE_TABLE_BASE_ADDR is in UMA space.
-	 * Therefore we need to do
-	 * DCN_VM_CONTEXT0_PAGE_TABLE_BASE_ADDR = VM_CONTEXT0_PAGE_TABLE_BASE_ADDR
-	 * - DCHUBBUB_SDPIF_FB_OFFSET + DCHUBBUB_SDPIF_FB_BASE
-	 */
-	fb_base.quad_part = (uint64_t)fb_base_value << 24;
-	fb_offset.quad_part = (uint64_t)fb_offset_value << 24;
-	vm0->pte_base.quad_part += fb_base.quad_part;
-	vm0->pte_base.quad_part -= fb_offset.quad_part;
-}
-
-
-void dcn10_program_pte_vm(struct dce_hwseq *hws, struct hubp *hubp)
-{
-	struct dcn10_hubp *hubp1 = TO_DCN10_HUBP(hubp);
-	struct vm_system_aperture_param apt = {0};
-	struct vm_context0_param vm0 = {0};
-
-	mmhub_read_vm_system_aperture_settings(hubp1, &apt, hws);
-	mmhub_read_vm_context0_settings(hubp1, &vm0, hws);
-
-	hubp->funcs->hubp_set_vm_system_aperture_settings(hubp, &apt);
-	hubp->funcs->hubp_set_vm_context0_settings(hubp, &vm0);
-}
-
-static void dcn10_enable_plane(
-	struct dc *dc,
-	struct pipe_ctx *pipe_ctx,
-	struct dc_state *context)
-{
-	struct dce_hwseq *hws = dc->hwseq;
-
-	if (dc->debug.sanity_checks) {
-		hws->funcs.verify_allow_pstate_change_high(dc);
-	}
-
-	undo_DEGVIDCN10_253_wa(dc);
-
-	power_on_plane(dc->hwseq,
-		pipe_ctx->plane_res.hubp->inst);
-
-	/* enable DCFCLK current DCHUB */
-	pipe_ctx->plane_res.hubp->funcs->hubp_clk_cntl(pipe_ctx->plane_res.hubp, true);
-
-	/* make sure OPP_PIPE_CLOCK_EN = 1 */
-	pipe_ctx->stream_res.opp->funcs->opp_pipe_clock_control(
-			pipe_ctx->stream_res.opp,
-			true);
-
-	if (dc->config.gpu_vm_support)
-		dcn10_program_pte_vm(hws, pipe_ctx->plane_res.hubp);
-
-	if (dc->debug.sanity_checks) {
-		hws->funcs.verify_allow_pstate_change_high(dc);
-	}
-
-	if (!pipe_ctx->top_pipe
-		&& pipe_ctx->plane_state
-		&& pipe_ctx->plane_state->flip_int_enabled
-		&& pipe_ctx->plane_res.hubp->funcs->hubp_set_flip_int)
-			pipe_ctx->plane_res.hubp->funcs->hubp_set_flip_int(pipe_ctx->plane_res.hubp);
-
-}
-
-void dcn10_program_gamut_remap(struct pipe_ctx *pipe_ctx)
-{
-	int i = 0;
-	struct dpp_grph_csc_adjustment adjust;
-	memset(&adjust, 0, sizeof(adjust));
-	adjust.gamut_adjust_type = GRAPHICS_GAMUT_ADJUST_TYPE_BYPASS;
-
-
-	if (pipe_ctx->stream->gamut_remap_matrix.enable_remap == true) {
-		adjust.gamut_adjust_type = GRAPHICS_GAMUT_ADJUST_TYPE_SW;
-		for (i = 0; i < CSC_TEMPERATURE_MATRIX_SIZE; i++)
-			adjust.temperature_matrix[i] =
-				pipe_ctx->stream->gamut_remap_matrix.matrix[i];
-	} else if (pipe_ctx->plane_state &&
-		   pipe_ctx->plane_state->gamut_remap_matrix.enable_remap == true) {
-		adjust.gamut_adjust_type = GRAPHICS_GAMUT_ADJUST_TYPE_SW;
-		for (i = 0; i < CSC_TEMPERATURE_MATRIX_SIZE; i++)
-			adjust.temperature_matrix[i] =
-				pipe_ctx->plane_state->gamut_remap_matrix.matrix[i];
-	}
-
-	pipe_ctx->plane_res.dpp->funcs->dpp_set_gamut_remap(pipe_ctx->plane_res.dpp, &adjust);
-}
-
-
-static bool dcn10_is_rear_mpo_fix_required(struct pipe_ctx *pipe_ctx, enum dc_color_space colorspace)
-{
-	if (pipe_ctx->plane_state && pipe_ctx->plane_state->layer_index > 0 && is_rgb_cspace(colorspace)) {
-		if (pipe_ctx->top_pipe) {
-			struct pipe_ctx *top = pipe_ctx->top_pipe;
-
-			while (top->top_pipe)
-				top = top->top_pipe; // Traverse to top pipe_ctx
-			if (top->plane_state && top->plane_state->layer_index == 0)
-				return true; // Front MPO plane not hidden
-		}
-	}
-	return false;
-}
-
-static void dcn10_set_csc_adjustment_rgb_mpo_fix(struct pipe_ctx *pipe_ctx, uint16_t *matrix)
-{
-	// Override rear plane RGB bias to fix MPO brightness
-	uint16_t rgb_bias = matrix[3];
-
-	matrix[3] = 0;
-	matrix[7] = 0;
-	matrix[11] = 0;
-	pipe_ctx->plane_res.dpp->funcs->dpp_set_csc_adjustment(pipe_ctx->plane_res.dpp, matrix);
-	matrix[3] = rgb_bias;
-	matrix[7] = rgb_bias;
-	matrix[11] = rgb_bias;
-}
-
-void dcn10_program_output_csc(struct dc *dc,
-		struct pipe_ctx *pipe_ctx,
-		enum dc_color_space colorspace,
-		uint16_t *matrix,
-		int opp_id)
-{
-	if (pipe_ctx->stream->csc_color_matrix.enable_adjustment == true) {
-		if (pipe_ctx->plane_res.dpp->funcs->dpp_set_csc_adjustment != NULL) {
-
-			/* MPO is broken with RGB colorspaces when OCSC matrix
-			 * brightness offset >= 0 on DCN1 due to OCSC before MPC
-			 * Blending adds offsets from front + rear to rear plane
-			 *
-			 * Fix is to set RGB bias to 0 on rear plane, top plane
-			 * black value pixels add offset instead of rear + front
-			 */
-
-			int16_t rgb_bias = matrix[3];
-			// matrix[3/7/11] are all the same offset value
-
-			if (rgb_bias > 0 && dcn10_is_rear_mpo_fix_required(pipe_ctx, colorspace)) {
-				dcn10_set_csc_adjustment_rgb_mpo_fix(pipe_ctx, matrix);
-			} else {
-				pipe_ctx->plane_res.dpp->funcs->dpp_set_csc_adjustment(pipe_ctx->plane_res.dpp, matrix);
-			}
-		}
-	} else {
-		if (pipe_ctx->plane_res.dpp->funcs->dpp_set_csc_default != NULL)
-			pipe_ctx->plane_res.dpp->funcs->dpp_set_csc_default(pipe_ctx->plane_res.dpp, colorspace);
-	}
-}
-
-static void dcn10_update_dpp(struct dpp *dpp, struct dc_plane_state *plane_state)
-{
-	struct dc_bias_and_scale bns_params = {0};
-
-	// program the input csc
-	dpp->funcs->dpp_setup(dpp,
-			plane_state->format,
-			EXPANSION_MODE_ZERO,
-			plane_state->input_csc_color_matrix,
-			plane_state->color_space,
-			NULL);
-
-	//set scale and bias registers
-	build_prescale_params(&bns_params, plane_state);
-	if (dpp->funcs->dpp_program_bias_and_scale)
-		dpp->funcs->dpp_program_bias_and_scale(dpp, &bns_params);
-}
-
-void dcn10_update_visual_confirm_color(struct dc *dc, struct pipe_ctx *pipe_ctx, struct tg_color *color, int mpcc_id)
-{
-	struct mpc *mpc = dc->res_pool->mpc;
-
-	if (dc->debug.visual_confirm == VISUAL_CONFIRM_HDR)
-		get_hdr_visual_confirm_color(pipe_ctx, color);
-	else if (dc->debug.visual_confirm == VISUAL_CONFIRM_SURFACE)
-		get_surface_visual_confirm_color(pipe_ctx, color);
-	else if (dc->debug.visual_confirm == VISUAL_CONFIRM_SWIZZLE)
-		get_surface_tile_visual_confirm_color(pipe_ctx, color);
-	else
-		color_space_to_black_color(
-				dc, pipe_ctx->stream->output_color_space, color);
-
-	if (mpc->funcs->set_bg_color)
-		mpc->funcs->set_bg_color(mpc, color, mpcc_id);
-}
-
-void dcn10_update_mpcc(struct dc *dc, struct pipe_ctx *pipe_ctx)
-{
-	struct hubp *hubp = pipe_ctx->plane_res.hubp;
-	struct mpcc_blnd_cfg blnd_cfg = {0};
-	bool per_pixel_alpha = pipe_ctx->plane_state->per_pixel_alpha && pipe_ctx->bottom_pipe;
-	int mpcc_id;
-	struct mpcc *new_mpcc;
-	struct mpc *mpc = dc->res_pool->mpc;
-	struct mpc_tree *mpc_tree_params = &(pipe_ctx->stream_res.opp->mpc_tree_params);
-
-	if (per_pixel_alpha)
-		blnd_cfg.alpha_mode = MPCC_ALPHA_BLEND_MODE_PER_PIXEL_ALPHA;
-	else
-		blnd_cfg.alpha_mode = MPCC_ALPHA_BLEND_MODE_GLOBAL_ALPHA;
-
-	blnd_cfg.overlap_only = false;
-	blnd_cfg.global_gain = 0xff;
-
-	if (pipe_ctx->plane_state->global_alpha)
-		blnd_cfg.global_alpha = pipe_ctx->plane_state->global_alpha_value;
-	else
-		blnd_cfg.global_alpha = 0xff;
-
-	/* DCN1.0 has output CM before MPC which seems to screw with
-	 * pre-multiplied alpha.
-	 */
-	blnd_cfg.pre_multiplied_alpha = is_rgb_cspace(
-			pipe_ctx->stream->output_color_space)
-					&& per_pixel_alpha;
-
-
-	/*
-	 * TODO: remove hack
-	 * Note: currently there is a bug in init_hw such that
-	 * on resume from hibernate, BIOS sets up MPCC0, and
-	 * we do mpcc_remove but the mpcc cannot go to idle
-	 * after remove. This cause us to pick mpcc1 here,
-	 * which causes a pstate hang for yet unknown reason.
-	 */
-	mpcc_id = hubp->inst;
-
-	/* If there is no full update, don't need to touch MPC tree*/
-	if (!pipe_ctx->plane_state->update_flags.bits.full_update) {
-		mpc->funcs->update_blending(mpc, &blnd_cfg, mpcc_id);
-		dc->hwss.update_visual_confirm_color(dc, pipe_ctx, &blnd_cfg.black_color, mpcc_id);
-		return;
-	}
-
-	/* check if this MPCC is already being used */
-	new_mpcc = mpc->funcs->get_mpcc_for_dpp(mpc_tree_params, mpcc_id);
-	/* remove MPCC if being used */
-	if (new_mpcc != NULL)
-		mpc->funcs->remove_mpcc(mpc, mpc_tree_params, new_mpcc);
-	else
-		if (dc->debug.sanity_checks)
-			mpc->funcs->assert_mpcc_idle_before_connect(
-					dc->res_pool->mpc, mpcc_id);
-
-	/* Call MPC to insert new plane */
-	new_mpcc = mpc->funcs->insert_plane(dc->res_pool->mpc,
-			mpc_tree_params,
-			&blnd_cfg,
-			NULL,
-			NULL,
-			hubp->inst,
-			mpcc_id);
-	dc->hwss.update_visual_confirm_color(dc, pipe_ctx, &blnd_cfg.black_color, mpcc_id);
-
-	ASSERT(new_mpcc != NULL);
-
-	hubp->opp_id = pipe_ctx->stream_res.opp->inst;
-	hubp->mpcc_id = mpcc_id;
-}
-
-static void update_scaler(struct pipe_ctx *pipe_ctx)
-{
-	bool per_pixel_alpha =
-			pipe_ctx->plane_state->per_pixel_alpha && pipe_ctx->bottom_pipe;
-
-	pipe_ctx->plane_res.scl_data.lb_params.alpha_en = per_pixel_alpha;
-	pipe_ctx->plane_res.scl_data.lb_params.depth = LB_PIXEL_DEPTH_36BPP;
-	/* scaler configuration */
-	pipe_ctx->plane_res.dpp->funcs->dpp_set_scaler(
-			pipe_ctx->plane_res.dpp, &pipe_ctx->plane_res.scl_data);
-}
-
-static void dcn10_update_dchubp_dpp(
-	struct dc *dc,
-	struct pipe_ctx *pipe_ctx,
-	struct dc_state *context)
-{
-	struct dce_hwseq *hws = dc->hwseq;
-	struct hubp *hubp = pipe_ctx->plane_res.hubp;
-	struct dpp *dpp = pipe_ctx->plane_res.dpp;
-	struct dc_plane_state *plane_state = pipe_ctx->plane_state;
-	struct plane_size size = plane_state->plane_size;
-	unsigned int compat_level = 0;
-	bool should_divided_by_2 = false;
-
-	/* depends on DML calculation, DPP clock value may change dynamically */
-	/* If request max dpp clk is lower than current dispclk, no need to
-	 * divided by 2
-	 */
-	if (plane_state->update_flags.bits.full_update) {
-
-		/* new calculated dispclk, dppclk are stored in
-		 * context->bw_ctx.bw.dcn.clk.dispclk_khz / dppclk_khz. current
-		 * dispclk, dppclk are from dc->clk_mgr->clks.dispclk_khz.
-		 * dcn_validate_bandwidth compute new dispclk, dppclk.
-		 * dispclk will put in use after optimize_bandwidth when
-		 * ramp_up_dispclk_with_dpp is called.
-		 * there are two places for dppclk be put in use. One location
-		 * is the same as the location as dispclk. Another is within
-		 * update_dchubp_dpp which happens between pre_bandwidth and
-		 * optimize_bandwidth.
-		 * dppclk updated within update_dchubp_dpp will cause new
-		 * clock values of dispclk and dppclk not be in use at the same
-		 * time. when clocks are decreased, this may cause dppclk is
-		 * lower than previous configuration and let pipe stuck.
-		 * for example, eDP + external dp,  change resolution of DP from
-		 * 1920x1080x144hz to 1280x960x60hz.
-		 * before change: dispclk = 337889 dppclk = 337889
-		 * change mode, dcn_validate_bandwidth calculate
-		 *                dispclk = 143122 dppclk = 143122
-		 * update_dchubp_dpp be executed before dispclk be updated,
-		 * dispclk = 337889, but dppclk use new value dispclk /2 =
-		 * 168944. this will cause pipe pstate warning issue.
-		 * solution: between pre_bandwidth and optimize_bandwidth, while
-		 * dispclk is going to be decreased, keep dppclk = dispclk
-		 **/
-		if (context->bw_ctx.bw.dcn.clk.dispclk_khz <
-				dc->clk_mgr->clks.dispclk_khz)
-			should_divided_by_2 = false;
-		else
-			should_divided_by_2 =
-					context->bw_ctx.bw.dcn.clk.dppclk_khz <=
-					dc->clk_mgr->clks.dispclk_khz / 2;
-
-		dpp->funcs->dpp_dppclk_control(
-				dpp,
-				should_divided_by_2,
-				true);
-
-		if (dc->res_pool->dccg)
-			dc->res_pool->dccg->funcs->update_dpp_dto(
-					dc->res_pool->dccg,
-					dpp->inst,
-					pipe_ctx->plane_res.bw.dppclk_khz);
-		else
-			dc->clk_mgr->clks.dppclk_khz = should_divided_by_2 ?
-						dc->clk_mgr->clks.dispclk_khz / 2 :
-							dc->clk_mgr->clks.dispclk_khz;
-	}
-
-	/* TODO: Need input parameter to tell current DCHUB pipe tie to which OTG
-	 * VTG is within DCHUBBUB which is commond block share by each pipe HUBP.
-	 * VTG is 1:1 mapping with OTG. Each pipe HUBP will select which VTG
-	 */
-	if (plane_state->update_flags.bits.full_update) {
-		hubp->funcs->hubp_vtg_sel(hubp, pipe_ctx->stream_res.tg->inst);
-
-		hubp->funcs->hubp_setup(
-			hubp,
-			&pipe_ctx->dlg_regs,
-			&pipe_ctx->ttu_regs,
-			&pipe_ctx->rq_regs,
-			&pipe_ctx->pipe_dlg_param);
-		hubp->funcs->hubp_setup_interdependent(
-			hubp,
-			&pipe_ctx->dlg_regs,
-			&pipe_ctx->ttu_regs);
-	}
-
-	size.surface_size = pipe_ctx->plane_res.scl_data.viewport;
-
-	if (plane_state->update_flags.bits.full_update ||
-		plane_state->update_flags.bits.bpp_change)
-		dcn10_update_dpp(dpp, plane_state);
-
-	if (plane_state->update_flags.bits.full_update ||
-		plane_state->update_flags.bits.per_pixel_alpha_change ||
-		plane_state->update_flags.bits.global_alpha_change)
-		hws->funcs.update_mpcc(dc, pipe_ctx);
-
-	if (plane_state->update_flags.bits.full_update ||
-		plane_state->update_flags.bits.per_pixel_alpha_change ||
-		plane_state->update_flags.bits.global_alpha_change ||
-		plane_state->update_flags.bits.scaling_change ||
-		plane_state->update_flags.bits.position_change) {
-		update_scaler(pipe_ctx);
-	}
-
-	if (plane_state->update_flags.bits.full_update ||
-		plane_state->update_flags.bits.scaling_change ||
-		plane_state->update_flags.bits.position_change) {
-		hubp->funcs->mem_program_viewport(
-			hubp,
-			&pipe_ctx->plane_res.scl_data.viewport,
-			&pipe_ctx->plane_res.scl_data.viewport_c);
-	}
-
-	if (pipe_ctx->stream->cursor_attributes.address.quad_part != 0) {
-		dc->hwss.set_cursor_position(pipe_ctx);
-		dc->hwss.set_cursor_attribute(pipe_ctx);
-
-		if (dc->hwss.set_cursor_sdr_white_level)
-			dc->hwss.set_cursor_sdr_white_level(pipe_ctx);
-	}
-
-	if (plane_state->update_flags.bits.full_update) {
-		/*gamut remap*/
-		dc->hwss.program_gamut_remap(pipe_ctx);
-
-		dc->hwss.program_output_csc(dc,
-				pipe_ctx,
-				pipe_ctx->stream->output_color_space,
-				pipe_ctx->stream->csc_color_matrix.matrix,
-				pipe_ctx->stream_res.opp->inst);
-	}
-
-	if (plane_state->update_flags.bits.full_update ||
-		plane_state->update_flags.bits.pixel_format_change ||
-		plane_state->update_flags.bits.horizontal_mirror_change ||
-		plane_state->update_flags.bits.rotation_change ||
-		plane_state->update_flags.bits.swizzle_change ||
-		plane_state->update_flags.bits.dcc_change ||
-		plane_state->update_flags.bits.bpp_change ||
-		plane_state->update_flags.bits.scaling_change ||
-		plane_state->update_flags.bits.plane_size_change) {
-		hubp->funcs->hubp_program_surface_config(
-			hubp,
-			plane_state->format,
-			&plane_state->tiling_info,
-			&size,
-			plane_state->rotation,
-			&plane_state->dcc,
-			plane_state->horizontal_mirror,
-			compat_level);
-	}
-
-	hubp->power_gated = false;
-
-	hws->funcs.update_plane_addr(dc, pipe_ctx);
-
-	if (is_pipe_tree_visible(pipe_ctx))
-		hubp->funcs->set_blank(hubp, false);
-}
-
-void dcn10_blank_pixel_data(
-		struct dc *dc,
-		struct pipe_ctx *pipe_ctx,
-		bool blank)
-{
-	enum dc_color_space color_space;
-	struct tg_color black_color = {0};
-	struct stream_resource *stream_res = &pipe_ctx->stream_res;
-	struct dc_stream_state *stream = pipe_ctx->stream;
-
-	/* program otg blank color */
-	color_space = stream->output_color_space;
-	color_space_to_black_color(dc, color_space, &black_color);
-
-	/*
-	 * The way 420 is packed, 2 channels carry Y component, 1 channel
-	 * alternate between Cb and Cr, so both channels need the pixel
-	 * value for Y
-	 */
-	if (stream->timing.pixel_encoding == PIXEL_ENCODING_YCBCR420)
-		black_color.color_r_cr = black_color.color_g_y;
-
-
-	if (stream_res->tg->funcs->set_blank_color)
-		stream_res->tg->funcs->set_blank_color(
-				stream_res->tg,
-				&black_color);
-
-	if (!blank) {
-		if (stream_res->tg->funcs->set_blank)
-			stream_res->tg->funcs->set_blank(stream_res->tg, blank);
-		if (stream_res->abm) {
-			dc->hwss.set_pipe(pipe_ctx);
-			stream_res->abm->funcs->set_abm_level(stream_res->abm, stream->abm_level);
-		}
-	} else if (blank) {
-		dc->hwss.set_abm_immediate_disable(pipe_ctx);
-		if (stream_res->tg->funcs->set_blank) {
-			stream_res->tg->funcs->wait_for_state(stream_res->tg, CRTC_STATE_VBLANK);
-			stream_res->tg->funcs->set_blank(stream_res->tg, blank);
-		}
-	}
-}
-
-void dcn10_set_hdr_multiplier(struct pipe_ctx *pipe_ctx)
-{
-	struct fixed31_32 multiplier = pipe_ctx->plane_state->hdr_mult;
-	uint32_t hw_mult = 0x1f000; // 1.0 default multiplier
-	struct custom_float_format fmt;
-
-	fmt.exponenta_bits = 6;
-	fmt.mantissa_bits = 12;
-	fmt.sign = true;
-
-
-	if (!dc_fixpt_eq(multiplier, dc_fixpt_from_int(0))) // check != 0
-		convert_to_custom_float_format(multiplier, &fmt, &hw_mult);
-
-	pipe_ctx->plane_res.dpp->funcs->dpp_set_hdr_multiplier(
-			pipe_ctx->plane_res.dpp, hw_mult);
-}
-
-void dcn10_program_pipe(
-		struct dc *dc,
-		struct pipe_ctx *pipe_ctx,
-		struct dc_state *context)
-{
-	struct dce_hwseq *hws = dc->hwseq;
-
-	if (pipe_ctx->top_pipe == NULL) {
-		bool blank = !is_pipe_tree_visible(pipe_ctx);
-
-		pipe_ctx->stream_res.tg->funcs->program_global_sync(
-				pipe_ctx->stream_res.tg,
-				pipe_ctx->pipe_dlg_param.vready_offset,
-				pipe_ctx->pipe_dlg_param.vstartup_start,
-				pipe_ctx->pipe_dlg_param.vupdate_offset,
-				pipe_ctx->pipe_dlg_param.vupdate_width);
-
-		pipe_ctx->stream_res.tg->funcs->set_vtg_params(
-				pipe_ctx->stream_res.tg, &pipe_ctx->stream->timing, true);
-
-		if (hws->funcs.setup_vupdate_interrupt)
-			hws->funcs.setup_vupdate_interrupt(dc, pipe_ctx);
-
-		hws->funcs.blank_pixel_data(dc, pipe_ctx, blank);
-	}
-
-	if (pipe_ctx->plane_state->update_flags.bits.full_update)
-		dcn10_enable_plane(dc, pipe_ctx, context);
-
-	dcn10_update_dchubp_dpp(dc, pipe_ctx, context);
-
-	hws->funcs.set_hdr_multiplier(pipe_ctx);
-
-	if (pipe_ctx->plane_state->update_flags.bits.full_update ||
-			pipe_ctx->plane_state->update_flags.bits.in_transfer_func_change ||
-			pipe_ctx->plane_state->update_flags.bits.gamma_change)
-		hws->funcs.set_input_transfer_func(dc, pipe_ctx, pipe_ctx->plane_state);
-
-	/* dcn10_translate_regamma_to_hw_format takes 750us to finish
-	 * only do gamma programming for full update.
-	 * TODO: This can be further optimized/cleaned up
-	 * Always call this for now since it does memcmp inside before
-	 * doing heavy calculation and programming
-	 */
-	if (pipe_ctx->plane_state->update_flags.bits.full_update)
-		hws->funcs.set_output_transfer_func(dc, pipe_ctx, pipe_ctx->stream);
-}
-
-void dcn10_wait_for_pending_cleared(struct dc *dc,
-		struct dc_state *context)
-{
-		struct pipe_ctx *pipe_ctx;
-		struct timing_generator *tg;
-		int i;
-
-		for (i = 0; i < dc->res_pool->pipe_count; i++) {
-			pipe_ctx = &context->res_ctx.pipe_ctx[i];
-			tg = pipe_ctx->stream_res.tg;
-
-			/*
-			 * Only wait for top pipe's tg penindg bit
-			 * Also skip if pipe is disabled.
-			 */
-			if (pipe_ctx->top_pipe ||
-			    !pipe_ctx->stream || !pipe_ctx->plane_state ||
-			    !tg->funcs->is_tg_enabled(tg))
-				continue;
-
-			/*
-			 * Wait for VBLANK then VACTIVE to ensure we get VUPDATE.
-			 * For some reason waiting for OTG_UPDATE_PENDING cleared
-			 * seems to not trigger the update right away, and if we
-			 * lock again before VUPDATE then we don't get a separated
-			 * operation.
-			 */
-			pipe_ctx->stream_res.tg->funcs->wait_for_state(pipe_ctx->stream_res.tg, CRTC_STATE_VBLANK);
-			pipe_ctx->stream_res.tg->funcs->wait_for_state(pipe_ctx->stream_res.tg, CRTC_STATE_VACTIVE);
-		}
-}
-
-void dcn10_post_unlock_program_front_end(
-		struct dc *dc,
-		struct dc_state *context)
-{
-	int i;
-
-	DC_LOGGER_INIT(dc->ctx->logger);
-
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i];
-
-		if (!pipe_ctx->top_pipe &&
-			!pipe_ctx->prev_odm_pipe &&
-			pipe_ctx->stream) {
-			struct timing_generator *tg = pipe_ctx->stream_res.tg;
-
-			if (context->stream_status[i].plane_count == 0)
-				false_optc_underflow_wa(dc, pipe_ctx->stream, tg);
-		}
-	}
-
-	for (i = 0; i < dc->res_pool->pipe_count; i++)
-		if (context->res_ctx.pipe_ctx[i].update_flags.bits.disable)
-			dc->hwss.disable_plane(dc, &dc->current_state->res_ctx.pipe_ctx[i]);
-
-	for (i = 0; i < dc->res_pool->pipe_count; i++)
-		if (context->res_ctx.pipe_ctx[i].update_flags.bits.disable) {
-			dc->hwss.optimize_bandwidth(dc, context);
-			break;
-		}
-
-	if (dc->hwseq->wa.DEGVIDCN10_254)
-		hubbub1_wm_change_req_wa(dc->res_pool->hubbub);
-}
-
-static void dcn10_stereo_hw_frame_pack_wa(struct dc *dc, struct dc_state *context)
-{
-	uint8_t i;
-
-	for (i = 0; i < context->stream_count; i++) {
-		if (context->streams[i]->timing.timing_3d_format
-				== TIMING_3D_FORMAT_HW_FRAME_PACKING) {
-			/*
-			 * Disable stutter
-			 */
-			hubbub1_allow_self_refresh_control(dc->res_pool->hubbub, false);
-			break;
-		}
-	}
-}
-
-void dcn10_prepare_bandwidth(
-		struct dc *dc,
-		struct dc_state *context)
-{
-	struct dce_hwseq *hws = dc->hwseq;
-	struct hubbub *hubbub = dc->res_pool->hubbub;
-
-	if (dc->debug.sanity_checks)
-		hws->funcs.verify_allow_pstate_change_high(dc);
-
-	if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) {
-		if (context->stream_count == 0)
-			context->bw_ctx.bw.dcn.clk.phyclk_khz = 0;
-
-		dc->clk_mgr->funcs->update_clocks(
-				dc->clk_mgr,
-				context,
-				false);
-	}
-
-	dc->wm_optimized_required = hubbub->funcs->program_watermarks(hubbub,
-			&context->bw_ctx.bw.dcn.watermarks,
-			dc->res_pool->ref_clocks.dchub_ref_clock_inKhz / 1000,
-			true);
-	dcn10_stereo_hw_frame_pack_wa(dc, context);
-
-	if (dc->debug.pplib_wm_report_mode == WM_REPORT_OVERRIDE)
-		dcn_bw_notify_pplib_of_wm_ranges(dc);
-
-	if (dc->debug.sanity_checks)
-		hws->funcs.verify_allow_pstate_change_high(dc);
-}
-
-void dcn10_optimize_bandwidth(
-		struct dc *dc,
-		struct dc_state *context)
-{
-	struct dce_hwseq *hws = dc->hwseq;
-	struct hubbub *hubbub = dc->res_pool->hubbub;
-
-	if (dc->debug.sanity_checks)
-		hws->funcs.verify_allow_pstate_change_high(dc);
-
-	if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) {
-		if (context->stream_count == 0)
-			context->bw_ctx.bw.dcn.clk.phyclk_khz = 0;
-
-		dc->clk_mgr->funcs->update_clocks(
-				dc->clk_mgr,
-				context,
-				true);
-	}
-
-	hubbub->funcs->program_watermarks(hubbub,
-			&context->bw_ctx.bw.dcn.watermarks,
-			dc->res_pool->ref_clocks.dchub_ref_clock_inKhz / 1000,
-			true);
-
-	dcn10_stereo_hw_frame_pack_wa(dc, context);
-
-	if (dc->debug.pplib_wm_report_mode == WM_REPORT_OVERRIDE)
-		dcn_bw_notify_pplib_of_wm_ranges(dc);
-
-	if (dc->debug.sanity_checks)
-		hws->funcs.verify_allow_pstate_change_high(dc);
-}
-
-void dcn10_set_drr(struct pipe_ctx **pipe_ctx,
-		int num_pipes, struct dc_crtc_timing_adjust adjust)
-{
-	int i = 0;
-	struct drr_params params = {0};
-	// DRR set trigger event mapped to OTG_TRIG_A (bit 11) for manual control flow
-	unsigned int event_triggers = 0x800;
-	// Note DRR trigger events are generated regardless of whether num frames met.
-	unsigned int num_frames = 2;
-
-	params.vertical_total_max = adjust.v_total_max;
-	params.vertical_total_min = adjust.v_total_min;
-	params.vertical_total_mid = adjust.v_total_mid;
-	params.vertical_total_mid_frame_num = adjust.v_total_mid_frame_num;
-	/* TODO: If multiple pipes are to be supported, you need
-	 * some GSL stuff. Static screen triggers may be programmed differently
-	 * as well.
-	 */
-	for (i = 0; i < num_pipes; i++) {
-		pipe_ctx[i]->stream_res.tg->funcs->set_drr(
-			pipe_ctx[i]->stream_res.tg, &params);
-		if (adjust.v_total_max != 0 && adjust.v_total_min != 0)
-			pipe_ctx[i]->stream_res.tg->funcs->set_static_screen_control(
-					pipe_ctx[i]->stream_res.tg,
-					event_triggers, num_frames);
-	}
-}
-
-void dcn10_get_position(struct pipe_ctx **pipe_ctx,
-		int num_pipes,
-		struct crtc_position *position)
-{
-	int i = 0;
-
-	/* TODO: handle pipes > 1
-	 */
-	for (i = 0; i < num_pipes; i++)
-		pipe_ctx[i]->stream_res.tg->funcs->get_position(pipe_ctx[i]->stream_res.tg, position);
-}
-
-void dcn10_set_static_screen_control(struct pipe_ctx **pipe_ctx,
-		int num_pipes, const struct dc_static_screen_params *params)
-{
-	unsigned int i;
-	unsigned int triggers = 0;
-
-	if (params->triggers.surface_update)
-		triggers |= 0x80;
-	if (params->triggers.cursor_update)
-		triggers |= 0x2;
-	if (params->triggers.force_trigger)
-		triggers |= 0x1;
-
-	for (i = 0; i < num_pipes; i++)
-		pipe_ctx[i]->stream_res.tg->funcs->
-			set_static_screen_control(pipe_ctx[i]->stream_res.tg,
-					triggers, params->num_frames);
-}
-
-static void dcn10_config_stereo_parameters(
-		struct dc_stream_state *stream, struct crtc_stereo_flags *flags)
-{
-	enum view_3d_format view_format = stream->view_format;
-	enum dc_timing_3d_format timing_3d_format =\
-			stream->timing.timing_3d_format;
-	bool non_stereo_timing = false;
-
-	if (timing_3d_format == TIMING_3D_FORMAT_NONE ||
-		timing_3d_format == TIMING_3D_FORMAT_SIDE_BY_SIDE ||
-		timing_3d_format == TIMING_3D_FORMAT_TOP_AND_BOTTOM)
-		non_stereo_timing = true;
-
-	if (non_stereo_timing == false &&
-		view_format == VIEW_3D_FORMAT_FRAME_SEQUENTIAL) {
-
-		flags->PROGRAM_STEREO         = 1;
-		flags->PROGRAM_POLARITY       = 1;
-		if (timing_3d_format == TIMING_3D_FORMAT_INBAND_FA ||
-			timing_3d_format == TIMING_3D_FORMAT_DP_HDMI_INBAND_FA ||
-			timing_3d_format == TIMING_3D_FORMAT_SIDEBAND_FA) {
-			enum display_dongle_type dongle = \
-					stream->link->ddc->dongle_type;
-			if (dongle == DISPLAY_DONGLE_DP_VGA_CONVERTER ||
-				dongle == DISPLAY_DONGLE_DP_DVI_CONVERTER ||
-				dongle == DISPLAY_DONGLE_DP_HDMI_CONVERTER)
-				flags->DISABLE_STEREO_DP_SYNC = 1;
-		}
-		flags->RIGHT_EYE_POLARITY =\
-				stream->timing.flags.RIGHT_EYE_3D_POLARITY;
-		if (timing_3d_format == TIMING_3D_FORMAT_HW_FRAME_PACKING)
-			flags->FRAME_PACKED = 1;
-	}
-
-	return;
-}
-
-void dcn10_setup_stereo(struct pipe_ctx *pipe_ctx, struct dc *dc)
-{
-	struct crtc_stereo_flags flags = { 0 };
-	struct dc_stream_state *stream = pipe_ctx->stream;
-
-	dcn10_config_stereo_parameters(stream, &flags);
-
-	if (stream->timing.timing_3d_format == TIMING_3D_FORMAT_SIDEBAND_FA) {
-		if (!dc_set_generic_gpio_for_stereo(true, dc->ctx->gpio_service))
-			dc_set_generic_gpio_for_stereo(false, dc->ctx->gpio_service);
-	} else {
-		dc_set_generic_gpio_for_stereo(false, dc->ctx->gpio_service);
-	}
-
-	pipe_ctx->stream_res.opp->funcs->opp_program_stereo(
-		pipe_ctx->stream_res.opp,
-		flags.PROGRAM_STEREO == 1,
-		&stream->timing);
-
-	pipe_ctx->stream_res.tg->funcs->program_stereo(
-		pipe_ctx->stream_res.tg,
-		&stream->timing,
-		&flags);
-
-	return;
-}
-
-static struct hubp *get_hubp_by_inst(struct resource_pool *res_pool, int mpcc_inst)
-{
-	int i;
-
-	for (i = 0; i < res_pool->pipe_count; i++) {
-		if (res_pool->hubps[i]->inst == mpcc_inst)
-			return res_pool->hubps[i];
-	}
-	ASSERT(false);
-	return NULL;
-}
-
-void dcn10_wait_for_mpcc_disconnect(
-		struct dc *dc,
-		struct resource_pool *res_pool,
-		struct pipe_ctx *pipe_ctx)
-{
-	struct dce_hwseq *hws = dc->hwseq;
-	int mpcc_inst;
-
-	if (dc->debug.sanity_checks) {
-		hws->funcs.verify_allow_pstate_change_high(dc);
-	}
-
-	if (!pipe_ctx->stream_res.opp)
-		return;
-
-	for (mpcc_inst = 0; mpcc_inst < MAX_PIPES; mpcc_inst++) {
-		if (pipe_ctx->stream_res.opp->mpcc_disconnect_pending[mpcc_inst]) {
-			struct hubp *hubp = get_hubp_by_inst(res_pool, mpcc_inst);
-
-			res_pool->mpc->funcs->wait_for_idle(res_pool->mpc, mpcc_inst);
-			pipe_ctx->stream_res.opp->mpcc_disconnect_pending[mpcc_inst] = false;
-			hubp->funcs->set_blank(hubp, true);
-		}
-	}
-
-	if (dc->debug.sanity_checks) {
-		hws->funcs.verify_allow_pstate_change_high(dc);
-	}
-
-}
-
-bool dcn10_dummy_display_power_gating(
-	struct dc *dc,
-	uint8_t controller_id,
-	struct dc_bios *dcb,
-	enum pipe_gating_control power_gating)
-{
-	return true;
-}
-
-void dcn10_update_pending_status(struct pipe_ctx *pipe_ctx)
-{
-	struct dc_plane_state *plane_state = pipe_ctx->plane_state;
-	struct timing_generator *tg = pipe_ctx->stream_res.tg;
-	bool flip_pending;
-	struct dc *dc = plane_state->ctx->dc;
-
-	if (plane_state == NULL)
-		return;
-
-	flip_pending = pipe_ctx->plane_res.hubp->funcs->hubp_is_flip_pending(
-					pipe_ctx->plane_res.hubp);
-
-	plane_state->status.is_flip_pending = plane_state->status.is_flip_pending || flip_pending;
-
-	if (!flip_pending)
-		plane_state->status.current_address = plane_state->status.requested_address;
-
-	if (plane_state->status.current_address.type == PLN_ADDR_TYPE_GRPH_STEREO &&
-			tg->funcs->is_stereo_left_eye) {
-		plane_state->status.is_right_eye =
-				!tg->funcs->is_stereo_left_eye(pipe_ctx->stream_res.tg);
-	}
-
-	if (dc->hwseq->wa_state.disallow_self_refresh_during_multi_plane_transition_applied) {
-		struct dce_hwseq *hwseq = dc->hwseq;
-		struct timing_generator *tg = dc->res_pool->timing_generators[0];
-		unsigned int cur_frame = tg->funcs->get_frame_count(tg);
-
-		if (cur_frame != hwseq->wa_state.disallow_self_refresh_during_multi_plane_transition_applied_on_frame) {
-			struct hubbub *hubbub = dc->res_pool->hubbub;
-
-			hubbub->funcs->allow_self_refresh_control(hubbub, !dc->debug.disable_stutter);
-			hwseq->wa_state.disallow_self_refresh_during_multi_plane_transition_applied = false;
-		}
-	}
-}
-
-void dcn10_update_dchub(struct dce_hwseq *hws, struct dchub_init_data *dh_data)
-{
-	struct hubbub *hubbub = hws->ctx->dc->res_pool->hubbub;
-
-	/* In DCN, this programming sequence is owned by the hubbub */
-	hubbub->funcs->update_dchub(hubbub, dh_data);
-}
-
-static bool dcn10_can_pipe_disable_cursor(struct pipe_ctx *pipe_ctx)
-{
-	struct pipe_ctx *test_pipe, *split_pipe;
-	const struct scaler_data *scl_data = &pipe_ctx->plane_res.scl_data;
-	struct rect r1 = scl_data->recout, r2, r2_half;
-	int r1_r = r1.x + r1.width, r1_b = r1.y + r1.height, r2_r, r2_b;
-	int cur_layer = pipe_ctx->plane_state->layer_index;
-
-	/**
-	 * Disable the cursor if there's another pipe above this with a
-	 * plane that contains this pipe's viewport to prevent double cursor
-	 * and incorrect scaling artifacts.
-	 */
-	for (test_pipe = pipe_ctx->top_pipe; test_pipe;
-	     test_pipe = test_pipe->top_pipe) {
-		// Skip invisible layer and pipe-split plane on same layer
-		if (!test_pipe->plane_state->visible || test_pipe->plane_state->layer_index == cur_layer)
-			continue;
-
-		r2 = test_pipe->plane_res.scl_data.recout;
-		r2_r = r2.x + r2.width;
-		r2_b = r2.y + r2.height;
-		split_pipe = test_pipe;
-
-		/**
-		 * There is another half plane on same layer because of
-		 * pipe-split, merge together per same height.
-		 */
-		for (split_pipe = pipe_ctx->top_pipe; split_pipe;
-		     split_pipe = split_pipe->top_pipe)
-			if (split_pipe->plane_state->layer_index == test_pipe->plane_state->layer_index) {
-				r2_half = split_pipe->plane_res.scl_data.recout;
-				r2.x = (r2_half.x < r2.x) ? r2_half.x : r2.x;
-				r2.width = r2.width + r2_half.width;
-				r2_r = r2.x + r2.width;
-				break;
-			}
-
-		if (r1.x >= r2.x && r1.y >= r2.y && r1_r <= r2_r && r1_b <= r2_b)
-			return true;
-	}
-
-	return false;
-}
-
-void dcn10_set_cursor_position(struct pipe_ctx *pipe_ctx)
-{
-	struct dc_cursor_position pos_cpy = pipe_ctx->stream->cursor_position;
-	struct hubp *hubp = pipe_ctx->plane_res.hubp;
-	struct dpp *dpp = pipe_ctx->plane_res.dpp;
-	struct dc_cursor_mi_param param = {
-		.pixel_clk_khz = pipe_ctx->stream->timing.pix_clk_100hz / 10,
-		.ref_clk_khz = pipe_ctx->stream->ctx->dc->res_pool->ref_clocks.dchub_ref_clock_inKhz,
-		.viewport = pipe_ctx->plane_res.scl_data.viewport,
-		.h_scale_ratio = pipe_ctx->plane_res.scl_data.ratios.horz,
-		.v_scale_ratio = pipe_ctx->plane_res.scl_data.ratios.vert,
-		.rotation = pipe_ctx->plane_state->rotation,
-		.mirror = pipe_ctx->plane_state->horizontal_mirror
-	};
-	bool pipe_split_on = (pipe_ctx->top_pipe != NULL) ||
-		(pipe_ctx->bottom_pipe != NULL);
-	bool odm_combine_on = (pipe_ctx->next_odm_pipe != NULL) ||
-		(pipe_ctx->prev_odm_pipe != NULL);
-
-	int x_plane = pipe_ctx->plane_state->dst_rect.x;
-	int y_plane = pipe_ctx->plane_state->dst_rect.y;
-	int x_pos = pos_cpy.x;
-	int y_pos = pos_cpy.y;
-
-	/**
-	 * DC cursor is stream space, HW cursor is plane space and drawn
-	 * as part of the framebuffer.
-	 *
-	 * Cursor position can't be negative, but hotspot can be used to
-	 * shift cursor out of the plane bounds. Hotspot must be smaller
-	 * than the cursor size.
-	 */
-
-	/**
-	 * Translate cursor from stream space to plane space.
-	 *
-	 * If the cursor is scaled then we need to scale the position
-	 * to be in the approximately correct place. We can't do anything
-	 * about the actual size being incorrect, that's a limitation of
-	 * the hardware.
-	 */
-	if (param.rotation == ROTATION_ANGLE_90 || param.rotation == ROTATION_ANGLE_270) {
-		x_pos = (x_pos - x_plane) * pipe_ctx->plane_state->src_rect.height /
-				pipe_ctx->plane_state->dst_rect.width;
-		y_pos = (y_pos - y_plane) * pipe_ctx->plane_state->src_rect.width /
-				pipe_ctx->plane_state->dst_rect.height;
-	} else {
-		x_pos = (x_pos - x_plane) * pipe_ctx->plane_state->src_rect.width /
-				pipe_ctx->plane_state->dst_rect.width;
-		y_pos = (y_pos - y_plane) * pipe_ctx->plane_state->src_rect.height /
-				pipe_ctx->plane_state->dst_rect.height;
-	}
-
-	/**
-	 * If the cursor's source viewport is clipped then we need to
-	 * translate the cursor to appear in the correct position on
-	 * the screen.
-	 *
-	 * This translation isn't affected by scaling so it needs to be
-	 * done *after* we adjust the position for the scale factor.
-	 *
-	 * This is only done by opt-in for now since there are still
-	 * some usecases like tiled display that might enable the
-	 * cursor on both streams while expecting dc to clip it.
-	 */
-	if (pos_cpy.translate_by_source) {
-		x_pos += pipe_ctx->plane_state->src_rect.x;
-		y_pos += pipe_ctx->plane_state->src_rect.y;
-	}
-
-	/**
-	 * If the position is negative then we need to add to the hotspot
-	 * to shift the cursor outside the plane.
-	 */
-
-	if (x_pos < 0) {
-		pos_cpy.x_hotspot -= x_pos;
-		x_pos = 0;
-	}
-
-	if (y_pos < 0) {
-		pos_cpy.y_hotspot -= y_pos;
-		y_pos = 0;
-	}
-
-	pos_cpy.x = (uint32_t)x_pos;
-	pos_cpy.y = (uint32_t)y_pos;
-
-	if (pipe_ctx->plane_state->address.type
-			== PLN_ADDR_TYPE_VIDEO_PROGRESSIVE)
-		pos_cpy.enable = false;
-
-	if (pos_cpy.enable && dcn10_can_pipe_disable_cursor(pipe_ctx))
-		pos_cpy.enable = false;
-
-	// Swap axis and mirror horizontally
-	if (param.rotation == ROTATION_ANGLE_90) {
-		uint32_t temp_x = pos_cpy.x;
-
-		pos_cpy.x = pipe_ctx->plane_res.scl_data.viewport.width -
-				(pos_cpy.y - pipe_ctx->plane_res.scl_data.viewport.x) + pipe_ctx->plane_res.scl_data.viewport.x;
-		pos_cpy.y = temp_x;
-	}
-	// Swap axis and mirror vertically
-	else if (param.rotation == ROTATION_ANGLE_270) {
-		uint32_t temp_y = pos_cpy.y;
-		int viewport_height =
-			pipe_ctx->plane_res.scl_data.viewport.height;
-		int viewport_y =
-			pipe_ctx->plane_res.scl_data.viewport.y;
-
-		/**
-		 * Display groups that are 1xnY, have pos_cpy.x > 2 * viewport.height
-		 * For pipe split cases:
-		 * - apply offset of viewport.y to normalize pos_cpy.x
-		 * - calculate the pos_cpy.y as before
-		 * - shift pos_cpy.y back by same offset to get final value
-		 * - since we iterate through both pipes, use the lower
-		 *   viewport.y for offset
-		 * For non pipe split cases, use the same calculation for
-		 *  pos_cpy.y as the 180 degree rotation case below,
-		 *  but use pos_cpy.x as our input because we are rotating
-		 *  270 degrees
-		 */
-		if (pipe_split_on || odm_combine_on) {
-			int pos_cpy_x_offset;
-			int other_pipe_viewport_y;
-
-			if (pipe_split_on) {
-				if (pipe_ctx->bottom_pipe) {
-					other_pipe_viewport_y =
-						pipe_ctx->bottom_pipe->plane_res.scl_data.viewport.y;
-				} else {
-					other_pipe_viewport_y =
-						pipe_ctx->top_pipe->plane_res.scl_data.viewport.y;
-				}
-			} else {
-				if (pipe_ctx->next_odm_pipe) {
-					other_pipe_viewport_y =
-						pipe_ctx->next_odm_pipe->plane_res.scl_data.viewport.y;
-				} else {
-					other_pipe_viewport_y =
-						pipe_ctx->prev_odm_pipe->plane_res.scl_data.viewport.y;
-				}
-			}
-			pos_cpy_x_offset = (viewport_y > other_pipe_viewport_y) ?
-				other_pipe_viewport_y : viewport_y;
-			pos_cpy.x -= pos_cpy_x_offset;
-			if (pos_cpy.x > viewport_height) {
-				pos_cpy.x = pos_cpy.x - viewport_height;
-				pos_cpy.y = viewport_height - pos_cpy.x;
-			} else {
-				pos_cpy.y = 2 * viewport_height - pos_cpy.x;
-			}
-			pos_cpy.y += pos_cpy_x_offset;
-		} else {
-			pos_cpy.y = (2 * viewport_y) + viewport_height - pos_cpy.x;
-		}
-		pos_cpy.x = temp_y;
-	}
-	// Mirror horizontally and vertically
-	else if (param.rotation == ROTATION_ANGLE_180) {
-		int viewport_width =
-			pipe_ctx->plane_res.scl_data.viewport.width;
-		int viewport_x =
-			pipe_ctx->plane_res.scl_data.viewport.x;
-
-		if (pipe_split_on || odm_combine_on) {
-			if (pos_cpy.x >= viewport_width + viewport_x) {
-				pos_cpy.x = 2 * viewport_width
-						- pos_cpy.x + 2 * viewport_x;
-			} else {
-				uint32_t temp_x = pos_cpy.x;
-
-				pos_cpy.x = 2 * viewport_x - pos_cpy.x;
-				if (temp_x >= viewport_x +
-					(int)hubp->curs_attr.width || pos_cpy.x
-					<= (int)hubp->curs_attr.width +
-					pipe_ctx->plane_state->src_rect.x) {
-					pos_cpy.x = temp_x + viewport_width;
-				}
-			}
-		} else {
-			pos_cpy.x = viewport_width - pos_cpy.x + 2 * viewport_x;
-		}
-
-		/**
-		 * Display groups that are 1xnY, have pos_cpy.y > viewport.height
-		 * Calculation:
-		 *   delta_from_bottom = viewport.y + viewport.height - pos_cpy.y
-		 *   pos_cpy.y_new = viewport.y + delta_from_bottom
-		 * Simplify it as:
-		 *   pos_cpy.y = viewport.y * 2 + viewport.height - pos_cpy.y
-		 */
-		pos_cpy.y = (2 * pipe_ctx->plane_res.scl_data.viewport.y) +
-			pipe_ctx->plane_res.scl_data.viewport.height - pos_cpy.y;
-	}
-
-	hubp->funcs->set_cursor_position(hubp, &pos_cpy, &param);
-	dpp->funcs->set_cursor_position(dpp, &pos_cpy, &param, hubp->curs_attr.width, hubp->curs_attr.height);
-}
-
-void dcn10_set_cursor_attribute(struct pipe_ctx *pipe_ctx)
-{
-	struct dc_cursor_attributes *attributes = &pipe_ctx->stream->cursor_attributes;
-
-	pipe_ctx->plane_res.hubp->funcs->set_cursor_attributes(
-			pipe_ctx->plane_res.hubp, attributes);
-	pipe_ctx->plane_res.dpp->funcs->set_cursor_attributes(
-		pipe_ctx->plane_res.dpp, attributes);
-}
-
-void dcn10_set_cursor_sdr_white_level(struct pipe_ctx *pipe_ctx)
-{
-	uint32_t sdr_white_level = pipe_ctx->stream->cursor_attributes.sdr_white_level;
-	struct fixed31_32 multiplier;
-	struct dpp_cursor_attributes opt_attr = { 0 };
-	uint32_t hw_scale = 0x3c00; // 1.0 default multiplier
-	struct custom_float_format fmt;
-
-	if (!pipe_ctx->plane_res.dpp->funcs->set_optional_cursor_attributes)
-		return;
-
-	fmt.exponenta_bits = 5;
-	fmt.mantissa_bits = 10;
-	fmt.sign = true;
-
-	if (sdr_white_level > 80) {
-		multiplier = dc_fixpt_from_fraction(sdr_white_level, 80);
-		convert_to_custom_float_format(multiplier, &fmt, &hw_scale);
-	}
-
-	opt_attr.scale = hw_scale;
-	opt_attr.bias = 0;
-
-	pipe_ctx->plane_res.dpp->funcs->set_optional_cursor_attributes(
-			pipe_ctx->plane_res.dpp, &opt_attr);
-}
-
-/*
- * apply_front_porch_workaround  TODO FPGA still need?
- *
- * This is a workaround for a bug that has existed since R5xx and has not been
- * fixed keep Front porch at minimum 2 for Interlaced mode or 1 for progressive.
- */
-static void apply_front_porch_workaround(
-	struct dc_crtc_timing *timing)
-{
-	if (timing->flags.INTERLACE == 1) {
-		if (timing->v_front_porch < 2)
-			timing->v_front_porch = 2;
-	} else {
-		if (timing->v_front_porch < 1)
-			timing->v_front_porch = 1;
-	}
-}
-
-int dcn10_get_vupdate_offset_from_vsync(struct pipe_ctx *pipe_ctx)
-{
-	const struct dc_crtc_timing *dc_crtc_timing = &pipe_ctx->stream->timing;
-	struct dc_crtc_timing patched_crtc_timing;
-	int vesa_sync_start;
-	int asic_blank_end;
-	int interlace_factor;
-	int vertical_line_start;
-
-	patched_crtc_timing = *dc_crtc_timing;
-	apply_front_porch_workaround(&patched_crtc_timing);
-
-	interlace_factor = patched_crtc_timing.flags.INTERLACE ? 2 : 1;
-
-	vesa_sync_start = patched_crtc_timing.v_addressable +
-			patched_crtc_timing.v_border_bottom +
-			patched_crtc_timing.v_front_porch;
-
-	asic_blank_end = (patched_crtc_timing.v_total -
-			vesa_sync_start -
-			patched_crtc_timing.v_border_top)
-			* interlace_factor;
-
-	vertical_line_start = asic_blank_end -
-			pipe_ctx->pipe_dlg_param.vstartup_start + 1;
-
-	return vertical_line_start;
-}
-
-void dcn10_calc_vupdate_position(
-		struct dc *dc,
-		struct pipe_ctx *pipe_ctx,
-		uint32_t *start_line,
-		uint32_t *end_line)
-{
-	const struct dc_crtc_timing *dc_crtc_timing = &pipe_ctx->stream->timing;
-	int vline_int_offset_from_vupdate =
-			pipe_ctx->stream->periodic_interrupt0.lines_offset;
-	int vupdate_offset_from_vsync = dc->hwss.get_vupdate_offset_from_vsync(pipe_ctx);
-	int start_position;
-
-	if (vline_int_offset_from_vupdate > 0)
-		vline_int_offset_from_vupdate--;
-	else if (vline_int_offset_from_vupdate < 0)
-		vline_int_offset_from_vupdate++;
-
-	start_position = vline_int_offset_from_vupdate + vupdate_offset_from_vsync;
-
-	if (start_position >= 0)
-		*start_line = start_position;
-	else
-		*start_line = dc_crtc_timing->v_total + start_position - 1;
-
-	*end_line = *start_line + 2;
-
-	if (*end_line >= dc_crtc_timing->v_total)
-		*end_line = 2;
-}
-
-static void dcn10_cal_vline_position(
-		struct dc *dc,
-		struct pipe_ctx *pipe_ctx,
-		enum vline_select vline,
-		uint32_t *start_line,
-		uint32_t *end_line)
-{
-	enum vertical_interrupt_ref_point ref_point = INVALID_POINT;
-
-	if (vline == VLINE0)
-		ref_point = pipe_ctx->stream->periodic_interrupt0.ref_point;
-	else if (vline == VLINE1)
-		ref_point = pipe_ctx->stream->periodic_interrupt1.ref_point;
-
-	switch (ref_point) {
-	case START_V_UPDATE:
-		dcn10_calc_vupdate_position(
-				dc,
-				pipe_ctx,
-				start_line,
-				end_line);
-		break;
-	case START_V_SYNC:
-		// Suppose to do nothing because vsync is 0;
-		break;
-	default:
-		ASSERT(0);
-		break;
-	}
-}
-
-void dcn10_setup_periodic_interrupt(
-		struct dc *dc,
-		struct pipe_ctx *pipe_ctx,
-		enum vline_select vline)
-{
-	struct timing_generator *tg = pipe_ctx->stream_res.tg;
-
-	if (vline == VLINE0) {
-		uint32_t start_line = 0;
-		uint32_t end_line = 0;
-
-		dcn10_cal_vline_position(dc, pipe_ctx, vline, &start_line, &end_line);
-
-		tg->funcs->setup_vertical_interrupt0(tg, start_line, end_line);
-
-	} else if (vline == VLINE1) {
-		pipe_ctx->stream_res.tg->funcs->setup_vertical_interrupt1(
-				tg,
-				pipe_ctx->stream->periodic_interrupt1.lines_offset);
-	}
-}
-
-void dcn10_setup_vupdate_interrupt(struct dc *dc, struct pipe_ctx *pipe_ctx)
-{
-	struct timing_generator *tg = pipe_ctx->stream_res.tg;
-	int start_line = dc->hwss.get_vupdate_offset_from_vsync(pipe_ctx);
-
-	if (start_line < 0) {
-		ASSERT(0);
-		start_line = 0;
-	}
-
-	if (tg->funcs->setup_vertical_interrupt2)
-		tg->funcs->setup_vertical_interrupt2(tg, start_line);
-}
-
-void dcn10_unblank_stream(struct pipe_ctx *pipe_ctx,
-		struct dc_link_settings *link_settings)
-{
-	struct encoder_unblank_param params = {0};
-	struct dc_stream_state *stream = pipe_ctx->stream;
-	struct dc_link *link = stream->link;
-	struct dce_hwseq *hws = link->dc->hwseq;
-
-	/* only 3 items below are used by unblank */
-	params.timing = pipe_ctx->stream->timing;
-
-	params.link_settings.link_rate = link_settings->link_rate;
-
-	if (dc_is_dp_signal(pipe_ctx->stream->signal)) {
-		if (params.timing.pixel_encoding == PIXEL_ENCODING_YCBCR420)
-			params.timing.pix_clk_100hz /= 2;
-		pipe_ctx->stream_res.stream_enc->funcs->dp_unblank(link, pipe_ctx->stream_res.stream_enc, &params);
-	}
-
-	if (link->local_sink && link->local_sink->sink_signal == SIGNAL_TYPE_EDP) {
-		hws->funcs.edp_backlight_control(link, true);
-	}
-}
-
-void dcn10_send_immediate_sdp_message(struct pipe_ctx *pipe_ctx,
-				const uint8_t *custom_sdp_message,
-				unsigned int sdp_message_size)
-{
-	if (dc_is_dp_signal(pipe_ctx->stream->signal)) {
-		pipe_ctx->stream_res.stream_enc->funcs->send_immediate_sdp_message(
-				pipe_ctx->stream_res.stream_enc,
-				custom_sdp_message,
-				sdp_message_size);
-	}
-}
-enum dc_status dcn10_set_clock(struct dc *dc,
-			enum dc_clock_type clock_type,
-			uint32_t clk_khz,
-			uint32_t stepping)
-{
-	struct dc_state *context = dc->current_state;
-	struct dc_clock_config clock_cfg = {0};
-	struct dc_clocks *current_clocks = &context->bw_ctx.bw.dcn.clk;
-
-	if (!dc->clk_mgr || !dc->clk_mgr->funcs->get_clock)
-		return DC_FAIL_UNSUPPORTED_1;
-
-	dc->clk_mgr->funcs->get_clock(dc->clk_mgr,
-		context, clock_type, &clock_cfg);
-
-	if (clk_khz > clock_cfg.max_clock_khz)
-		return DC_FAIL_CLK_EXCEED_MAX;
-
-	if (clk_khz < clock_cfg.min_clock_khz)
-		return DC_FAIL_CLK_BELOW_MIN;
-
-	if (clk_khz < clock_cfg.bw_requirequired_clock_khz)
-		return DC_FAIL_CLK_BELOW_CFG_REQUIRED;
-
-	/*update internal request clock for update clock use*/
-	if (clock_type == DC_CLOCK_TYPE_DISPCLK)
-		current_clocks->dispclk_khz = clk_khz;
-	else if (clock_type == DC_CLOCK_TYPE_DPPCLK)
-		current_clocks->dppclk_khz = clk_khz;
-	else
-		return DC_ERROR_UNEXPECTED;
-
-	if (dc->clk_mgr->funcs->update_clocks)
-				dc->clk_mgr->funcs->update_clocks(dc->clk_mgr,
-				context, true);
-	return DC_OK;
-
-}
-
-void dcn10_get_clock(struct dc *dc,
-			enum dc_clock_type clock_type,
-			struct dc_clock_config *clock_cfg)
-{
-	struct dc_state *context = dc->current_state;
-
-	if (dc->clk_mgr && dc->clk_mgr->funcs->get_clock)
-				dc->clk_mgr->funcs->get_clock(dc->clk_mgr, context, clock_type, clock_cfg);
-
-}
-
-void dcn10_get_dcc_en_bits(struct dc *dc, int *dcc_en_bits)
-{
-	struct resource_pool *pool = dc->res_pool;
-	int i;
-
-	for (i = 0; i < pool->pipe_count; i++) {
-		struct hubp *hubp = pool->hubps[i];
-		struct dcn_hubp_state *s = &(TO_DCN10_HUBP(hubp)->state);
-
-		hubp->funcs->hubp_read_state(hubp);
-
-		if (!s->blank_en)
-			dcc_en_bits[i] = s->dcc_en ? 1 : 0;
-	}
-}
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.h b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.h
deleted file mode 100644
index 9ae07c77fdc0..000000000000
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.h
+++ /dev/null
@@ -1,209 +0,0 @@
-/*
-* Copyright 2016-2020 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: AMD
- *
- */
-
-#ifndef __DC_HWSS_DCN10_H__
-#define __DC_HWSS_DCN10_H__
-
-#include "core_types.h"
-#include "hw_sequencer_private.h"
-
-struct dc;
-
-void dcn10_hw_sequencer_construct(struct dc *dc);
-
-int dcn10_get_vupdate_offset_from_vsync(struct pipe_ctx *pipe_ctx);
-void dcn10_calc_vupdate_position(
-		struct dc *dc,
-		struct pipe_ctx *pipe_ctx,
-		uint32_t *start_line,
-		uint32_t *end_line);
-void dcn10_setup_vupdate_interrupt(struct dc *dc, struct pipe_ctx *pipe_ctx);
-enum dc_status dcn10_enable_stream_timing(
-		struct pipe_ctx *pipe_ctx,
-		struct dc_state *context,
-		struct dc *dc);
-void dcn10_optimize_bandwidth(
-		struct dc *dc,
-		struct dc_state *context);
-void dcn10_prepare_bandwidth(
-		struct dc *dc,
-		struct dc_state *context);
-void dcn10_pipe_control_lock(
-	struct dc *dc,
-	struct pipe_ctx *pipe,
-	bool lock);
-void dcn10_cursor_lock(struct dc *dc, struct pipe_ctx *pipe, bool lock);
-void dcn10_blank_pixel_data(
-		struct dc *dc,
-		struct pipe_ctx *pipe_ctx,
-		bool blank);
-void dcn10_unblank_stream(struct pipe_ctx *pipe_ctx,
-		struct dc_link_settings *link_settings);
-void dcn10_program_output_csc(struct dc *dc,
-		struct pipe_ctx *pipe_ctx,
-		enum dc_color_space colorspace,
-		uint16_t *matrix,
-		int opp_id);
-bool dcn10_set_output_transfer_func(struct dc *dc, struct pipe_ctx *pipe_ctx,
-				const struct dc_stream_state *stream);
-bool dcn10_set_input_transfer_func(struct dc *dc, struct pipe_ctx *pipe_ctx,
-			const struct dc_plane_state *plane_state);
-void dcn10_update_plane_addr(const struct dc *dc, struct pipe_ctx *pipe_ctx);
-void dcn10_update_mpcc(struct dc *dc, struct pipe_ctx *pipe_ctx);
-void dcn10_reset_hw_ctx_wrap(
-		struct dc *dc,
-		struct dc_state *context);
-void dcn10_disable_plane(struct dc *dc, struct pipe_ctx *pipe_ctx);
-void dcn10_lock_all_pipes(
-		struct dc *dc,
-		struct dc_state *context,
-		bool lock);
-void dcn10_post_unlock_program_front_end(
-		struct dc *dc,
-		struct dc_state *context);
-void dcn10_hubp_pg_control(
-		struct dce_hwseq *hws,
-		unsigned int hubp_inst,
-		bool power_on);
-void dcn10_dpp_pg_control(
-		struct dce_hwseq *hws,
-		unsigned int dpp_inst,
-		bool power_on);
-void dcn10_enable_power_gating_plane(
-	struct dce_hwseq *hws,
-	bool enable);
-void dcn10_plane_atomic_disable(struct dc *dc, struct pipe_ctx *pipe_ctx);
-void dcn10_disable_vga(
-	struct dce_hwseq *hws);
-void dcn10_program_pipe(
-		struct dc *dc,
-		struct pipe_ctx *pipe_ctx,
-		struct dc_state *context);
-void dcn10_program_gamut_remap(struct pipe_ctx *pipe_ctx);
-void dcn10_init_hw(struct dc *dc);
-void dcn10_init_pipes(struct dc *dc, struct dc_state *context);
-void dcn10_power_down_on_boot(struct dc *dc);
-enum dc_status dce110_apply_ctx_to_hw(
-		struct dc *dc,
-		struct dc_state *context);
-void dcn10_plane_atomic_disconnect(struct dc *dc, struct pipe_ctx *pipe_ctx);
-void dcn10_update_dchub(struct dce_hwseq *hws, struct dchub_init_data *dh_data);
-void dcn10_update_pending_status(struct pipe_ctx *pipe_ctx);
-void dce110_power_down(struct dc *dc);
-void dce110_enable_accelerated_mode(struct dc *dc, struct dc_state *context);
-void dcn10_enable_timing_synchronization(
-		struct dc *dc,
-		int group_index,
-		int group_size,
-		struct pipe_ctx *grouped_pipes[]);
-void dcn10_enable_vblanks_synchronization(
-		struct dc *dc,
-		int group_index,
-		int group_size,
-		struct pipe_ctx *grouped_pipes[]);
-void dcn10_enable_per_frame_crtc_position_reset(
-		struct dc *dc,
-		int group_size,
-		struct pipe_ctx *grouped_pipes[]);
-void dce110_update_info_frame(struct pipe_ctx *pipe_ctx);
-void dcn10_send_immediate_sdp_message(struct pipe_ctx *pipe_ctx,
-		const uint8_t *custom_sdp_message,
-		unsigned int sdp_message_size);
-void dce110_blank_stream(struct pipe_ctx *pipe_ctx);
-void dce110_enable_audio_stream(struct pipe_ctx *pipe_ctx);
-void dce110_disable_audio_stream(struct pipe_ctx *pipe_ctx);
-bool dcn10_dummy_display_power_gating(
-		struct dc *dc,
-		uint8_t controller_id,
-		struct dc_bios *dcb,
-		enum pipe_gating_control power_gating);
-void dcn10_set_drr(struct pipe_ctx **pipe_ctx,
-		int num_pipes, struct dc_crtc_timing_adjust adjust);
-void dcn10_get_position(struct pipe_ctx **pipe_ctx,
-		int num_pipes,
-		struct crtc_position *position);
-void dcn10_set_static_screen_control(struct pipe_ctx **pipe_ctx,
-		int num_pipes, const struct dc_static_screen_params *params);
-void dcn10_setup_stereo(struct pipe_ctx *pipe_ctx, struct dc *dc);
-void dce110_set_avmute(struct pipe_ctx *pipe_ctx, bool enable);
-void dcn10_log_hw_state(struct dc *dc,
-		struct dc_log_buffer_ctx *log_ctx);
-void dcn10_get_hw_state(struct dc *dc,
-		char *pBuf,
-		unsigned int bufSize,
-		unsigned int mask);
-void dcn10_clear_status_bits(struct dc *dc, unsigned int mask);
-void dcn10_wait_for_mpcc_disconnect(
-		struct dc *dc,
-		struct resource_pool *res_pool,
-		struct pipe_ctx *pipe_ctx);
-void dce110_edp_backlight_control(
-		struct dc_link *link,
-		bool enable);
-void dce110_edp_wait_for_T12(
-		struct dc_link *link);
-void dce110_edp_power_control(
-		struct dc_link *link,
-		bool power_up);
-void dce110_edp_wait_for_hpd_ready(
-		struct dc_link *link,
-		bool power_up);
-void dcn10_set_cursor_position(struct pipe_ctx *pipe_ctx);
-void dcn10_set_cursor_attribute(struct pipe_ctx *pipe_ctx);
-void dcn10_set_cursor_sdr_white_level(struct pipe_ctx *pipe_ctx);
-void dcn10_setup_periodic_interrupt(
-		struct dc *dc,
-		struct pipe_ctx *pipe_ctx,
-		enum vline_select vline);
-enum dc_status dcn10_set_clock(struct dc *dc,
-		enum dc_clock_type clock_type,
-		uint32_t clk_khz,
-		uint32_t stepping);
-void dcn10_get_clock(struct dc *dc,
-		enum dc_clock_type clock_type,
-		struct dc_clock_config *clock_cfg);
-bool dcn10_did_underflow_occur(struct dc *dc, struct pipe_ctx *pipe_ctx);
-void dcn10_bios_golden_init(struct dc *dc);
-void dcn10_plane_atomic_power_down(struct dc *dc,
-		struct dpp *dpp,
-		struct hubp *hubp);
-bool dcn10_disconnect_pipes(
-		struct dc *dc,
-		struct dc_state *context);
-
-void dcn10_wait_for_pending_cleared(struct dc *dc,
-		struct dc_state *context);
-void dcn10_set_hdr_multiplier(struct pipe_ctx *pipe_ctx);
-void dcn10_verify_allow_pstate_change_high(struct dc *dc);
-
-void dcn10_get_dcc_en_bits(struct dc *dc, int *dcc_en_bits);
-
-void dcn10_update_visual_confirm_color(
-		struct dc *dc,
-		struct pipe_ctx *pipe_ctx,
-		struct tg_color *color,
-		int mpcc_id);
-
-#endif /* __DC_HWSS_DCN10_H__ */
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer_debug.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer_debug.c
index e8b6065fffad..baf663b661c8 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer_debug.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer_debug.c
@@ -27,12 +27,12 @@
 #include "core_types.h"
 #include "resource.h"
 #include "custom_float.h"
-#include "dcn10_hw_sequencer.h"
-#include "dce110/dce110_hw_sequencer.h"
+#include "dcn10/dcn10_hwseq.h"
+#include "dce110/dce110_hwseq.h"
 #include "dce/dce_hwseq.h"
 #include "abm.h"
 #include "dmcu.h"
-#include "dcn10_optc.h"
+#include "dcn10/dcn10_optc.h"
 #include "dcn10/dcn10_dpp.h"
 #include "dcn10/dcn10_mpc.h"
 #include "timing_generator.h"
@@ -40,12 +40,13 @@
 #include "ipp.h"
 #include "mpc.h"
 #include "reg_helper.h"
-#include "dcn10_hubp.h"
-#include "dcn10_hubbub.h"
+#include "dcn10/dcn10_hubp.h"
+#include "dcn10/dcn10_hubbub.h"
 #include "dcn10_cm_common.h"
 #include "clk_mgr.h"
 
-unsigned int snprintf_count(char *pBuf, unsigned int bufSize, char *fmt, ...)
+__printf(3, 4)
+unsigned int snprintf_count(char *pbuf, unsigned int bufsize, const char *fmt, ...)
 {
 	int ret_vsnprintf;
 	unsigned int chars_printed;
@@ -53,15 +54,15 @@ unsigned int snprintf_count(char *pBuf, unsigned int bufSize, char *fmt, ...)
 	va_list args;
 	va_start(args, fmt);
 
-	ret_vsnprintf = vsnprintf(pBuf, bufSize, fmt, args);
+	ret_vsnprintf = vsnprintf(pbuf, bufsize, fmt, args);
 
 	va_end(args);
 
 	if (ret_vsnprintf > 0) {
-		if (ret_vsnprintf < bufSize)
+		if (ret_vsnprintf < bufsize)
 			chars_printed = ret_vsnprintf;
 		else
-			chars_printed = bufSize - 1;
+			chars_printed = bufsize - 1;
 	} else
 		chars_printed = 0;
 
@@ -83,7 +84,7 @@ static unsigned int dcn10_get_hubbub_state(struct dc *dc, char *pBuf, unsigned i
 	memset(&wm, 0, sizeof(struct dcn_hubbub_wm));
 	dc->res_pool->hubbub->funcs->wm_read_state(dc->res_pool->hubbub, &wm);
 
-	chars_printed = snprintf_count(pBuf, remaining_buffer, "wm_set_index,data_urgent,pte_meta_urgent,sr_enter,sr_exit,dram_clk_chanage\n");
+	chars_printed = snprintf_count(pBuf, remaining_buffer, "wm_set_index,data_urgent,pte_meta_urgent,sr_enter,sr_exit,dram_clk_change\n");
 	remaining_buffer -= chars_printed;
 	pBuf += chars_printed;
 
@@ -98,7 +99,7 @@ static unsigned int dcn10_get_hubbub_state(struct dc *dc, char *pBuf, unsigned i
 			(s->pte_meta_urgent * frac) / ref_clk_mhz / frac, (s->pte_meta_urgent * frac) / ref_clk_mhz % frac,
 			(s->sr_enter * frac) / ref_clk_mhz / frac, (s->sr_enter * frac) / ref_clk_mhz % frac,
 			(s->sr_exit * frac) / ref_clk_mhz / frac, (s->sr_exit * frac) / ref_clk_mhz % frac,
-			(s->dram_clk_chanage * frac) / ref_clk_mhz / frac, (s->dram_clk_chanage * frac) / ref_clk_mhz % frac);
+			(s->dram_clk_change * frac) / ref_clk_mhz / frac, (s->dram_clk_change * frac) / ref_clk_mhz % frac);
 		remaining_buffer -= chars_printed;
 		pBuf += chars_printed;
 	}
@@ -391,7 +392,7 @@ static unsigned int dcn10_get_mpcc_states(struct dc *dc, char *pBuf, unsigned in
 	remaining_buffer -= chars_printed;
 	pBuf += chars_printed;
 
-	for (i = 0; i < pool->pipe_count; i++) {
+	for (i = 0; i < pool->mpcc_count; i++) {
 		struct mpcc_state s = {0};
 
 		pool->mpc->funcs->read_mpcc_state(pool->mpc, i, &s);
@@ -428,7 +429,9 @@ static unsigned int dcn10_get_otg_states(struct dc *dc, char *pBuf, unsigned int
 		struct dcn_otg_state s = {0};
 		int pix_clk = 0;
 
-		optc1_read_otg_state(DCN10TG_FROM_TG(tg), &s);
+		if (tg->funcs->read_otg_state)
+			tg->funcs->read_otg_state(tg, &s);
+
 		pix_clk = dc->current_state->res_ctx.pipe_ctx[i].stream_res.pix_clk_params.requested_pix_clk_100hz / 10;
 
 		//only print if OTG master is enabled
@@ -494,7 +497,8 @@ static void dcn10_clear_otpc_underflow(struct dc *dc)
 		struct timing_generator *tg = pool->timing_generators[i];
 		struct dcn_otg_state s = {0};
 
-		optc1_read_otg_state(DCN10TG_FROM_TG(tg), &s);
+		if (tg->funcs->read_otg_state)
+			tg->funcs->read_otg_state(tg, &s);
 
 		if (s.otg_enabled & 1)
 			tg->funcs->clear_optc_underflow(tg);
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_ipp.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_ipp.c
index f05371c1fc36..24b68337d76e 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_ipp.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_ipp.c
@@ -23,8 +23,6 @@
  *
  */
 
-#include <linux/slab.h>
-
 #include "dm_services.h"
 #include "dcn10_ipp.h"
 #include "reg_helper.h"
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.h b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.h
deleted file mode 100644
index c50c29984d51..000000000000
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.h
+++ /dev/null
@@ -1,717 +0,0 @@
-/*
- * Copyright 2012-15 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- *  and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: AMD
- *
- */
-
-#ifndef __DC_TIMING_GENERATOR_DCN10_H__
-#define __DC_TIMING_GENERATOR_DCN10_H__
-
-#include "timing_generator.h"
-
-#define DCN10TG_FROM_TG(tg)\
-	container_of(tg, struct optc, base)
-
-#define TG_COMMON_REG_LIST_DCN(inst) \
-	SRI(OTG_VSTARTUP_PARAM, OTG, inst),\
-	SRI(OTG_VUPDATE_PARAM, OTG, inst),\
-	SRI(OTG_VREADY_PARAM, OTG, inst),\
-	SRI(OTG_BLANK_CONTROL, OTG, inst),\
-	SRI(OTG_MASTER_UPDATE_LOCK, OTG, inst),\
-	SRI(OTG_GLOBAL_CONTROL0, OTG, inst),\
-	SRI(OTG_DOUBLE_BUFFER_CONTROL, OTG, inst),\
-	SRI(OTG_H_TOTAL, OTG, inst),\
-	SRI(OTG_H_BLANK_START_END, OTG, inst),\
-	SRI(OTG_H_SYNC_A, OTG, inst),\
-	SRI(OTG_H_SYNC_A_CNTL, OTG, inst),\
-	SRI(OTG_H_TIMING_CNTL, OTG, inst),\
-	SRI(OTG_V_TOTAL, OTG, inst),\
-	SRI(OTG_V_BLANK_START_END, OTG, inst),\
-	SRI(OTG_V_SYNC_A, OTG, inst),\
-	SRI(OTG_V_SYNC_A_CNTL, OTG, inst),\
-	SRI(OTG_INTERLACE_CONTROL, OTG, inst),\
-	SRI(OTG_CONTROL, OTG, inst),\
-	SRI(OTG_STEREO_CONTROL, OTG, inst),\
-	SRI(OTG_3D_STRUCTURE_CONTROL, OTG, inst),\
-	SRI(OTG_STEREO_STATUS, OTG, inst),\
-	SRI(OTG_V_TOTAL_MAX, OTG, inst),\
-	SRI(OTG_V_TOTAL_MID, OTG, inst),\
-	SRI(OTG_V_TOTAL_MIN, OTG, inst),\
-	SRI(OTG_V_TOTAL_CONTROL, OTG, inst),\
-	SRI(OTG_TRIGA_CNTL, OTG, inst),\
-	SRI(OTG_FORCE_COUNT_NOW_CNTL, OTG, inst),\
-	SRI(OTG_STATIC_SCREEN_CONTROL, OTG, inst),\
-	SRI(OTG_STATUS_FRAME_COUNT, OTG, inst),\
-	SRI(OTG_STATUS, OTG, inst),\
-	SRI(OTG_STATUS_POSITION, OTG, inst),\
-	SRI(OTG_NOM_VERT_POSITION, OTG, inst),\
-	SRI(OTG_BLACK_COLOR, OTG, inst),\
-	SRI(OTG_CLOCK_CONTROL, OTG, inst),\
-	SRI(OTG_VERTICAL_INTERRUPT0_CONTROL, OTG, inst),\
-	SRI(OTG_VERTICAL_INTERRUPT0_POSITION, OTG, inst),\
-	SRI(OTG_VERTICAL_INTERRUPT1_CONTROL, OTG, inst),\
-	SRI(OTG_VERTICAL_INTERRUPT1_POSITION, OTG, inst),\
-	SRI(OTG_VERTICAL_INTERRUPT2_CONTROL, OTG, inst),\
-	SRI(OTG_VERTICAL_INTERRUPT2_POSITION, OTG, inst),\
-	SRI(OPTC_INPUT_CLOCK_CONTROL, ODM, inst),\
-	SRI(OPTC_DATA_SOURCE_SELECT, ODM, inst),\
-	SRI(OPTC_INPUT_GLOBAL_CONTROL, ODM, inst),\
-	SRI(CONTROL, VTG, inst),\
-	SRI(OTG_VERT_SYNC_CONTROL, OTG, inst),\
-	SRI(OTG_MASTER_UPDATE_MODE, OTG, inst),\
-	SRI(OTG_GSL_CONTROL, OTG, inst),\
-	SRI(OTG_CRC_CNTL, OTG, inst),\
-	SRI(OTG_CRC0_DATA_RG, OTG, inst),\
-	SRI(OTG_CRC0_DATA_B, OTG, inst),\
-	SRI(OTG_CRC0_WINDOWA_X_CONTROL, OTG, inst),\
-	SRI(OTG_CRC0_WINDOWA_Y_CONTROL, OTG, inst),\
-	SRI(OTG_CRC0_WINDOWB_X_CONTROL, OTG, inst),\
-	SRI(OTG_CRC0_WINDOWB_Y_CONTROL, OTG, inst),\
-	SR(GSL_SOURCE_SELECT),\
-	SRI(OTG_GLOBAL_CONTROL2, OTG, inst),\
-	SRI(OTG_TRIGA_MANUAL_TRIG, OTG, inst)
-
-#define TG_COMMON_REG_LIST_DCN1_0(inst) \
-	TG_COMMON_REG_LIST_DCN(inst),\
-	SRI(OTG_TEST_PATTERN_PARAMETERS, OTG, inst),\
-	SRI(OTG_TEST_PATTERN_CONTROL, OTG, inst),\
-	SRI(OTG_TEST_PATTERN_COLOR, OTG, inst),\
-	SRI(OTG_MANUAL_FLOW_CONTROL, OTG, inst)
-
-
-struct dcn_optc_registers {
-	uint32_t OTG_GLOBAL_CONTROL1;
-	uint32_t OTG_GLOBAL_CONTROL2;
-	uint32_t OTG_VERT_SYNC_CONTROL;
-	uint32_t OTG_MASTER_UPDATE_MODE;
-	uint32_t OTG_GSL_CONTROL;
-	uint32_t OTG_VSTARTUP_PARAM;
-	uint32_t OTG_VUPDATE_PARAM;
-	uint32_t OTG_VREADY_PARAM;
-	uint32_t OTG_BLANK_CONTROL;
-	uint32_t OTG_MASTER_UPDATE_LOCK;
-	uint32_t OTG_GLOBAL_CONTROL0;
-	uint32_t OTG_DOUBLE_BUFFER_CONTROL;
-	uint32_t OTG_H_TOTAL;
-	uint32_t OTG_H_BLANK_START_END;
-	uint32_t OTG_H_SYNC_A;
-	uint32_t OTG_H_SYNC_A_CNTL;
-	uint32_t OTG_H_TIMING_CNTL;
-	uint32_t OTG_V_TOTAL;
-	uint32_t OTG_V_BLANK_START_END;
-	uint32_t OTG_V_SYNC_A;
-	uint32_t OTG_V_SYNC_A_CNTL;
-	uint32_t OTG_INTERLACE_CONTROL;
-	uint32_t OTG_CONTROL;
-	uint32_t OTG_STEREO_CONTROL;
-	uint32_t OTG_3D_STRUCTURE_CONTROL;
-	uint32_t OTG_STEREO_STATUS;
-	uint32_t OTG_V_TOTAL_MAX;
-	uint32_t OTG_V_TOTAL_MID;
-	uint32_t OTG_V_TOTAL_MIN;
-	uint32_t OTG_V_TOTAL_CONTROL;
-	uint32_t OTG_TRIGA_CNTL;
-	uint32_t OTG_TRIGA_MANUAL_TRIG;
-	uint32_t OTG_MANUAL_FLOW_CONTROL;
-	uint32_t OTG_FORCE_COUNT_NOW_CNTL;
-	uint32_t OTG_STATIC_SCREEN_CONTROL;
-	uint32_t OTG_STATUS_FRAME_COUNT;
-	uint32_t OTG_STATUS;
-	uint32_t OTG_STATUS_POSITION;
-	uint32_t OTG_NOM_VERT_POSITION;
-	uint32_t OTG_BLACK_COLOR;
-	uint32_t OTG_TEST_PATTERN_PARAMETERS;
-	uint32_t OTG_TEST_PATTERN_CONTROL;
-	uint32_t OTG_TEST_PATTERN_COLOR;
-	uint32_t OTG_CLOCK_CONTROL;
-	uint32_t OTG_VERTICAL_INTERRUPT0_CONTROL;
-	uint32_t OTG_VERTICAL_INTERRUPT0_POSITION;
-	uint32_t OTG_VERTICAL_INTERRUPT1_CONTROL;
-	uint32_t OTG_VERTICAL_INTERRUPT1_POSITION;
-	uint32_t OTG_VERTICAL_INTERRUPT2_CONTROL;
-	uint32_t OTG_VERTICAL_INTERRUPT2_POSITION;
-	uint32_t OPTC_INPUT_CLOCK_CONTROL;
-	uint32_t OPTC_DATA_SOURCE_SELECT;
-	uint32_t OPTC_MEMORY_CONFIG;
-	uint32_t OPTC_INPUT_GLOBAL_CONTROL;
-	uint32_t CONTROL;
-	uint32_t OTG_GSL_WINDOW_X;
-	uint32_t OTG_GSL_WINDOW_Y;
-	uint32_t OTG_VUPDATE_KEEPOUT;
-	uint32_t OTG_CRC_CNTL;
-	uint32_t OTG_CRC_CNTL2;
-	uint32_t OTG_CRC0_DATA_RG;
-	uint32_t OTG_CRC0_DATA_B;
-	uint32_t OTG_CRC0_WINDOWA_X_CONTROL;
-	uint32_t OTG_CRC0_WINDOWA_Y_CONTROL;
-	uint32_t OTG_CRC0_WINDOWB_X_CONTROL;
-	uint32_t OTG_CRC0_WINDOWB_Y_CONTROL;
-	uint32_t GSL_SOURCE_SELECT;
-	uint32_t DWB_SOURCE_SELECT;
-	uint32_t OTG_DSC_START_POSITION;
-	uint32_t OPTC_DATA_FORMAT_CONTROL;
-	uint32_t OPTC_BYTES_PER_PIXEL;
-	uint32_t OPTC_WIDTH_CONTROL;
-	uint32_t OTG_DRR_CONTROL;
-	uint32_t OTG_BLANK_DATA_COLOR;
-	uint32_t OTG_BLANK_DATA_COLOR_EXT;
-	uint32_t OTG_DRR_TRIGGER_WINDOW;
-	uint32_t OTG_M_CONST_DTO0;
-	uint32_t OTG_M_CONST_DTO1;
-	uint32_t OTG_DRR_V_TOTAL_CHANGE;
-	uint32_t OTG_GLOBAL_CONTROL4;
-};
-
-#define TG_COMMON_MASK_SH_LIST_DCN(mask_sh)\
-	SF(OTG0_OTG_VSTARTUP_PARAM, VSTARTUP_START, mask_sh),\
-	SF(OTG0_OTG_VUPDATE_PARAM, VUPDATE_OFFSET, mask_sh),\
-	SF(OTG0_OTG_VUPDATE_PARAM, VUPDATE_WIDTH, mask_sh),\
-	SF(OTG0_OTG_VREADY_PARAM, VREADY_OFFSET, mask_sh),\
-	SF(OTG0_OTG_BLANK_CONTROL, OTG_BLANK_DATA_EN, mask_sh),\
-	SF(OTG0_OTG_BLANK_CONTROL, OTG_BLANK_DE_MODE, mask_sh),\
-	SF(OTG0_OTG_BLANK_CONTROL, OTG_CURRENT_BLANK_STATE, mask_sh),\
-	SF(OTG0_OTG_MASTER_UPDATE_LOCK, OTG_MASTER_UPDATE_LOCK, mask_sh),\
-	SF(OTG0_OTG_MASTER_UPDATE_LOCK, UPDATE_LOCK_STATUS, mask_sh),\
-	SF(OTG0_OTG_GLOBAL_CONTROL0, OTG_MASTER_UPDATE_LOCK_SEL, mask_sh),\
-	SF(OTG0_OTG_DOUBLE_BUFFER_CONTROL, OTG_UPDATE_PENDING, mask_sh),\
-	SF(OTG0_OTG_DOUBLE_BUFFER_CONTROL, OTG_BLANK_DATA_DOUBLE_BUFFER_EN, mask_sh),\
-	SF(OTG0_OTG_DOUBLE_BUFFER_CONTROL, OTG_RANGE_TIMING_DBUF_UPDATE_MODE, mask_sh),\
-	SF(OTG0_OTG_VUPDATE_KEEPOUT, OTG_MASTER_UPDATE_LOCK_VUPDATE_KEEPOUT_EN, mask_sh), \
-	SF(OTG0_OTG_VUPDATE_KEEPOUT, MASTER_UPDATE_LOCK_VUPDATE_KEEPOUT_START_OFFSET, mask_sh), \
-	SF(OTG0_OTG_VUPDATE_KEEPOUT, MASTER_UPDATE_LOCK_VUPDATE_KEEPOUT_END_OFFSET, mask_sh), \
-	SF(OTG0_OTG_H_TOTAL, OTG_H_TOTAL, mask_sh),\
-	SF(OTG0_OTG_H_BLANK_START_END, OTG_H_BLANK_START, mask_sh),\
-	SF(OTG0_OTG_H_BLANK_START_END, OTG_H_BLANK_END, mask_sh),\
-	SF(OTG0_OTG_H_SYNC_A, OTG_H_SYNC_A_START, mask_sh),\
-	SF(OTG0_OTG_H_SYNC_A, OTG_H_SYNC_A_END, mask_sh),\
-	SF(OTG0_OTG_H_SYNC_A_CNTL, OTG_H_SYNC_A_POL, mask_sh),\
-	SF(OTG0_OTG_H_TIMING_CNTL, OTG_H_TIMING_DIV_BY2, mask_sh),\
-	SF(OTG0_OTG_V_TOTAL, OTG_V_TOTAL, mask_sh),\
-	SF(OTG0_OTG_V_BLANK_START_END, OTG_V_BLANK_START, mask_sh),\
-	SF(OTG0_OTG_V_BLANK_START_END, OTG_V_BLANK_END, mask_sh),\
-	SF(OTG0_OTG_V_SYNC_A, OTG_V_SYNC_A_START, mask_sh),\
-	SF(OTG0_OTG_V_SYNC_A, OTG_V_SYNC_A_END, mask_sh),\
-	SF(OTG0_OTG_V_SYNC_A_CNTL, OTG_V_SYNC_A_POL, mask_sh),\
-	SF(OTG0_OTG_INTERLACE_CONTROL, OTG_INTERLACE_ENABLE, mask_sh),\
-	SF(OTG0_OTG_CONTROL, OTG_MASTER_EN, mask_sh),\
-	SF(OTG0_OTG_CONTROL, OTG_START_POINT_CNTL, mask_sh),\
-	SF(OTG0_OTG_CONTROL, OTG_DISABLE_POINT_CNTL, mask_sh),\
-	SF(OTG0_OTG_CONTROL, OTG_FIELD_NUMBER_CNTL, mask_sh),\
-	SF(OTG0_OTG_CONTROL, OTG_CURRENT_MASTER_EN_STATE, mask_sh),\
-	SF(OTG0_OTG_STEREO_CONTROL, OTG_STEREO_EN, mask_sh),\
-	SF(OTG0_OTG_STEREO_CONTROL, OTG_STEREO_SYNC_OUTPUT_LINE_NUM, mask_sh),\
-	SF(OTG0_OTG_STEREO_CONTROL, OTG_STEREO_SYNC_OUTPUT_POLARITY, mask_sh),\
-	SF(OTG0_OTG_STEREO_CONTROL, OTG_STEREO_EYE_FLAG_POLARITY, mask_sh),\
-	SF(OTG0_OTG_STEREO_CONTROL, OTG_DISABLE_STEREOSYNC_OUTPUT_FOR_DP, mask_sh),\
-	SF(OTG0_OTG_STEREO_CONTROL, OTG_DISABLE_STEREOSYNC_OUTPUT_FOR_DP, mask_sh),\
-	SF(OTG0_OTG_STEREO_STATUS, OTG_STEREO_CURRENT_EYE, mask_sh),\
-	SF(OTG0_OTG_3D_STRUCTURE_CONTROL, OTG_3D_STRUCTURE_EN, mask_sh),\
-	SF(OTG0_OTG_3D_STRUCTURE_CONTROL, OTG_3D_STRUCTURE_V_UPDATE_MODE, mask_sh),\
-	SF(OTG0_OTG_3D_STRUCTURE_CONTROL, OTG_3D_STRUCTURE_STEREO_SEL_OVR, mask_sh),\
-	SF(OTG0_OTG_V_TOTAL_MAX, OTG_V_TOTAL_MAX, mask_sh),\
-	SF(OTG0_OTG_V_TOTAL_MID, OTG_V_TOTAL_MID, mask_sh),\
-	SF(OTG0_OTG_V_TOTAL_MIN, OTG_V_TOTAL_MIN, mask_sh),\
-	SF(OTG0_OTG_V_TOTAL_CONTROL, OTG_V_TOTAL_MIN_SEL, mask_sh),\
-	SF(OTG0_OTG_V_TOTAL_CONTROL, OTG_V_TOTAL_MAX_SEL, mask_sh),\
-	SF(OTG0_OTG_V_TOTAL_CONTROL, OTG_FORCE_LOCK_ON_EVENT, mask_sh),\
-	SF(OTG0_OTG_V_TOTAL_CONTROL, OTG_SET_V_TOTAL_MIN_MASK_EN, mask_sh),\
-	SF(OTG0_OTG_V_TOTAL_CONTROL, OTG_SET_V_TOTAL_MIN_MASK, mask_sh),\
-	SF(OTG0_OTG_V_TOTAL_CONTROL, OTG_VTOTAL_MID_REPLACING_MAX_EN, mask_sh),\
-	SF(OTG0_OTG_V_TOTAL_CONTROL, OTG_VTOTAL_MID_FRAME_NUM, mask_sh),\
-	SF(OTG0_OTG_FORCE_COUNT_NOW_CNTL, OTG_FORCE_COUNT_NOW_CLEAR, mask_sh),\
-	SF(OTG0_OTG_FORCE_COUNT_NOW_CNTL, OTG_FORCE_COUNT_NOW_MODE, mask_sh),\
-	SF(OTG0_OTG_FORCE_COUNT_NOW_CNTL, OTG_FORCE_COUNT_NOW_OCCURRED, mask_sh),\
-	SF(OTG0_OTG_TRIGA_CNTL, OTG_TRIGA_SOURCE_SELECT, mask_sh),\
-	SF(OTG0_OTG_TRIGA_CNTL, OTG_TRIGA_SOURCE_PIPE_SELECT, mask_sh),\
-	SF(OTG0_OTG_TRIGA_CNTL, OTG_TRIGA_RISING_EDGE_DETECT_CNTL, mask_sh),\
-	SF(OTG0_OTG_TRIGA_CNTL, OTG_TRIGA_FALLING_EDGE_DETECT_CNTL, mask_sh),\
-	SF(OTG0_OTG_TRIGA_CNTL, OTG_TRIGA_POLARITY_SELECT, mask_sh),\
-	SF(OTG0_OTG_TRIGA_CNTL, OTG_TRIGA_FREQUENCY_SELECT, mask_sh),\
-	SF(OTG0_OTG_TRIGA_CNTL, OTG_TRIGA_DELAY, mask_sh),\
-	SF(OTG0_OTG_TRIGA_CNTL, OTG_TRIGA_CLEAR, mask_sh),\
-	SF(OTG0_OTG_TRIGA_MANUAL_TRIG, OTG_TRIGA_MANUAL_TRIG, mask_sh),\
-	SF(OTG0_OTG_STATIC_SCREEN_CONTROL, OTG_STATIC_SCREEN_EVENT_MASK, mask_sh),\
-	SF(OTG0_OTG_STATIC_SCREEN_CONTROL, OTG_STATIC_SCREEN_FRAME_COUNT, mask_sh),\
-	SF(OTG0_OTG_STATUS_FRAME_COUNT, OTG_FRAME_COUNT, mask_sh),\
-	SF(OTG0_OTG_STATUS, OTG_V_BLANK, mask_sh),\
-	SF(OTG0_OTG_STATUS, OTG_V_ACTIVE_DISP, mask_sh),\
-	SF(OTG0_OTG_STATUS_POSITION, OTG_HORZ_COUNT, mask_sh),\
-	SF(OTG0_OTG_STATUS_POSITION, OTG_VERT_COUNT, mask_sh),\
-	SF(OTG0_OTG_NOM_VERT_POSITION, OTG_VERT_COUNT_NOM, mask_sh),\
-	SF(OTG0_OTG_BLACK_COLOR, OTG_BLACK_COLOR_B_CB, mask_sh),\
-	SF(OTG0_OTG_BLACK_COLOR, OTG_BLACK_COLOR_G_Y, mask_sh),\
-	SF(OTG0_OTG_BLACK_COLOR, OTG_BLACK_COLOR_R_CR, mask_sh),\
-	SF(OTG0_OTG_CLOCK_CONTROL, OTG_BUSY, mask_sh),\
-	SF(OTG0_OTG_CLOCK_CONTROL, OTG_CLOCK_EN, mask_sh),\
-	SF(OTG0_OTG_CLOCK_CONTROL, OTG_CLOCK_ON, mask_sh),\
-	SF(OTG0_OTG_CLOCK_CONTROL, OTG_CLOCK_GATE_DIS, mask_sh),\
-	SF(OTG0_OTG_VERTICAL_INTERRUPT0_CONTROL, OTG_VERTICAL_INTERRUPT0_INT_ENABLE, mask_sh),\
-	SF(OTG0_OTG_VERTICAL_INTERRUPT0_POSITION, OTG_VERTICAL_INTERRUPT0_LINE_START, mask_sh),\
-	SF(OTG0_OTG_VERTICAL_INTERRUPT0_POSITION, OTG_VERTICAL_INTERRUPT0_LINE_END, mask_sh),\
-	SF(OTG0_OTG_VERTICAL_INTERRUPT1_CONTROL, OTG_VERTICAL_INTERRUPT1_INT_ENABLE, mask_sh),\
-	SF(OTG0_OTG_VERTICAL_INTERRUPT1_POSITION, OTG_VERTICAL_INTERRUPT1_LINE_START, mask_sh),\
-	SF(OTG0_OTG_VERTICAL_INTERRUPT2_CONTROL, OTG_VERTICAL_INTERRUPT2_INT_ENABLE, mask_sh),\
-	SF(OTG0_OTG_VERTICAL_INTERRUPT2_POSITION, OTG_VERTICAL_INTERRUPT2_LINE_START, mask_sh),\
-	SF(ODM0_OPTC_INPUT_CLOCK_CONTROL, OPTC_INPUT_CLK_EN, mask_sh),\
-	SF(ODM0_OPTC_INPUT_CLOCK_CONTROL, OPTC_INPUT_CLK_ON, mask_sh),\
-	SF(ODM0_OPTC_INPUT_CLOCK_CONTROL, OPTC_INPUT_CLK_GATE_DIS, mask_sh),\
-	SF(ODM0_OPTC_INPUT_GLOBAL_CONTROL, OPTC_UNDERFLOW_OCCURRED_STATUS, mask_sh),\
-	SF(ODM0_OPTC_INPUT_GLOBAL_CONTROL, OPTC_UNDERFLOW_CLEAR, mask_sh),\
-	SF(VTG0_CONTROL, VTG0_ENABLE, mask_sh),\
-	SF(VTG0_CONTROL, VTG0_FP2, mask_sh),\
-	SF(VTG0_CONTROL, VTG0_VCOUNT_INIT, mask_sh),\
-	SF(OTG0_OTG_VERT_SYNC_CONTROL, OTG_FORCE_VSYNC_NEXT_LINE_OCCURRED, mask_sh),\
-	SF(OTG0_OTG_VERT_SYNC_CONTROL, OTG_FORCE_VSYNC_NEXT_LINE_CLEAR, mask_sh),\
-	SF(OTG0_OTG_VERT_SYNC_CONTROL, OTG_AUTO_FORCE_VSYNC_MODE, mask_sh),\
-	SF(OTG0_OTG_MASTER_UPDATE_MODE, MASTER_UPDATE_INTERLACED_MODE, mask_sh),\
-	SF(OTG0_OTG_GSL_CONTROL, OTG_GSL0_EN, mask_sh),\
-	SF(OTG0_OTG_GSL_CONTROL, OTG_GSL1_EN, mask_sh),\
-	SF(OTG0_OTG_GSL_CONTROL, OTG_GSL2_EN, mask_sh),\
-	SF(OTG0_OTG_GSL_CONTROL, OTG_GSL_MASTER_EN, mask_sh),\
-	SF(OTG0_OTG_GSL_CONTROL, OTG_GSL_FORCE_DELAY, mask_sh),\
-	SF(OTG0_OTG_GSL_CONTROL, OTG_GSL_CHECK_ALL_FIELDS, mask_sh),\
-	SF(OTG0_OTG_CRC_CNTL, OTG_CRC_CONT_EN, mask_sh),\
-	SF(OTG0_OTG_CRC_CNTL, OTG_CRC0_SELECT, mask_sh),\
-	SF(OTG0_OTG_CRC_CNTL, OTG_CRC_EN, mask_sh),\
-	SF(OTG0_OTG_CRC0_DATA_RG, CRC0_R_CR, mask_sh),\
-	SF(OTG0_OTG_CRC0_DATA_RG, CRC0_G_Y, mask_sh),\
-	SF(OTG0_OTG_CRC0_DATA_B, CRC0_B_CB, mask_sh),\
-	SF(OTG0_OTG_CRC0_WINDOWA_X_CONTROL, OTG_CRC0_WINDOWA_X_START, mask_sh),\
-	SF(OTG0_OTG_CRC0_WINDOWA_X_CONTROL, OTG_CRC0_WINDOWA_X_END, mask_sh),\
-	SF(OTG0_OTG_CRC0_WINDOWA_Y_CONTROL, OTG_CRC0_WINDOWA_Y_START, mask_sh),\
-	SF(OTG0_OTG_CRC0_WINDOWA_Y_CONTROL, OTG_CRC0_WINDOWA_Y_END, mask_sh),\
-	SF(OTG0_OTG_CRC0_WINDOWB_X_CONTROL, OTG_CRC0_WINDOWB_X_START, mask_sh),\
-	SF(OTG0_OTG_CRC0_WINDOWB_X_CONTROL, OTG_CRC0_WINDOWB_X_END, mask_sh),\
-	SF(OTG0_OTG_CRC0_WINDOWB_Y_CONTROL, OTG_CRC0_WINDOWB_Y_START, mask_sh),\
-	SF(OTG0_OTG_CRC0_WINDOWB_Y_CONTROL, OTG_CRC0_WINDOWB_Y_END, mask_sh),\
-	SF(GSL_SOURCE_SELECT, GSL0_READY_SOURCE_SEL, mask_sh),\
-	SF(GSL_SOURCE_SELECT, GSL1_READY_SOURCE_SEL, mask_sh),\
-	SF(GSL_SOURCE_SELECT, GSL2_READY_SOURCE_SEL, mask_sh),\
-	SF(OTG0_OTG_GLOBAL_CONTROL2, MANUAL_FLOW_CONTROL_SEL, mask_sh)
-
-
-
-#define TG_COMMON_MASK_SH_LIST_DCN1_0(mask_sh)\
-	TG_COMMON_MASK_SH_LIST_DCN(mask_sh),\
-	SF(OTG0_OTG_TEST_PATTERN_PARAMETERS, OTG_TEST_PATTERN_INC0, mask_sh),\
-	SF(OTG0_OTG_TEST_PATTERN_PARAMETERS, OTG_TEST_PATTERN_INC1, mask_sh),\
-	SF(OTG0_OTG_TEST_PATTERN_PARAMETERS, OTG_TEST_PATTERN_VRES, mask_sh),\
-	SF(OTG0_OTG_TEST_PATTERN_PARAMETERS, OTG_TEST_PATTERN_HRES, mask_sh),\
-	SF(OTG0_OTG_TEST_PATTERN_PARAMETERS, OTG_TEST_PATTERN_RAMP0_OFFSET, mask_sh),\
-	SF(OTG0_OTG_TEST_PATTERN_CONTROL, OTG_TEST_PATTERN_EN, mask_sh),\
-	SF(OTG0_OTG_TEST_PATTERN_CONTROL, OTG_TEST_PATTERN_MODE, mask_sh),\
-	SF(OTG0_OTG_TEST_PATTERN_CONTROL, OTG_TEST_PATTERN_DYNAMIC_RANGE, mask_sh),\
-	SF(OTG0_OTG_TEST_PATTERN_CONTROL, OTG_TEST_PATTERN_COLOR_FORMAT, mask_sh),\
-	SF(OTG0_OTG_TEST_PATTERN_COLOR, OTG_TEST_PATTERN_MASK, mask_sh),\
-	SF(OTG0_OTG_TEST_PATTERN_COLOR, OTG_TEST_PATTERN_DATA, mask_sh),\
-	SF(ODM0_OPTC_DATA_SOURCE_SELECT, OPTC_SRC_SEL, mask_sh),\
-	SF(OTG0_OTG_MANUAL_FLOW_CONTROL, MANUAL_FLOW_CONTROL, mask_sh),\
-
-#define TG_REG_FIELD_LIST_DCN1_0(type) \
-	type VSTARTUP_START;\
-	type VUPDATE_OFFSET;\
-	type VUPDATE_WIDTH;\
-	type VREADY_OFFSET;\
-	type OTG_BLANK_DATA_EN;\
-	type OTG_BLANK_DE_MODE;\
-	type OTG_CURRENT_BLANK_STATE;\
-	type OTG_MASTER_UPDATE_LOCK;\
-	type UPDATE_LOCK_STATUS;\
-	type OTG_UPDATE_PENDING;\
-	type OTG_MASTER_UPDATE_LOCK_SEL;\
-	type OTG_BLANK_DATA_DOUBLE_BUFFER_EN;\
-	type OTG_H_TOTAL;\
-	type OTG_H_BLANK_START;\
-	type OTG_H_BLANK_END;\
-	type OTG_H_SYNC_A_START;\
-	type OTG_H_SYNC_A_END;\
-	type OTG_H_SYNC_A_POL;\
-	type OTG_H_TIMING_DIV_BY2;\
-	type OTG_V_TOTAL;\
-	type OTG_V_BLANK_START;\
-	type OTG_V_BLANK_END;\
-	type OTG_V_SYNC_A_START;\
-	type OTG_V_SYNC_A_END;\
-	type OTG_V_SYNC_A_POL;\
-	type OTG_INTERLACE_ENABLE;\
-	type OTG_MASTER_EN;\
-	type OTG_START_POINT_CNTL;\
-	type OTG_DISABLE_POINT_CNTL;\
-	type OTG_FIELD_NUMBER_CNTL;\
-	type OTG_CURRENT_MASTER_EN_STATE;\
-	type OTG_STEREO_EN;\
-	type OTG_STEREO_SYNC_OUTPUT_LINE_NUM;\
-	type OTG_STEREO_SYNC_OUTPUT_POLARITY;\
-	type OTG_STEREO_EYE_FLAG_POLARITY;\
-	type OTG_STEREO_CURRENT_EYE;\
-	type OTG_DISABLE_STEREOSYNC_OUTPUT_FOR_DP;\
-	type OTG_3D_STRUCTURE_EN;\
-	type OTG_3D_STRUCTURE_V_UPDATE_MODE;\
-	type OTG_3D_STRUCTURE_STEREO_SEL_OVR;\
-	type OTG_V_TOTAL_MAX;\
-	type OTG_V_TOTAL_MID;\
-	type OTG_V_TOTAL_MIN;\
-	type OTG_V_TOTAL_MIN_SEL;\
-	type OTG_V_TOTAL_MAX_SEL;\
-	type OTG_VTOTAL_MID_REPLACING_MAX_EN;\
-	type OTG_VTOTAL_MID_FRAME_NUM;\
-	type OTG_FORCE_LOCK_ON_EVENT;\
-	type OTG_SET_V_TOTAL_MIN_MASK_EN;\
-	type OTG_SET_V_TOTAL_MIN_MASK;\
-	type OTG_FORCE_COUNT_NOW_CLEAR;\
-	type OTG_FORCE_COUNT_NOW_MODE;\
-	type OTG_FORCE_COUNT_NOW_OCCURRED;\
-	type OTG_TRIGA_SOURCE_SELECT;\
-	type OTG_TRIGA_SOURCE_PIPE_SELECT;\
-	type OTG_TRIGA_RISING_EDGE_DETECT_CNTL;\
-	type OTG_TRIGA_FALLING_EDGE_DETECT_CNTL;\
-	type OTG_TRIGA_POLARITY_SELECT;\
-	type OTG_TRIGA_FREQUENCY_SELECT;\
-	type OTG_TRIGA_DELAY;\
-	type OTG_TRIGA_CLEAR;\
-	type OTG_TRIGA_MANUAL_TRIG;\
-	type OTG_STATIC_SCREEN_EVENT_MASK;\
-	type OTG_STATIC_SCREEN_FRAME_COUNT;\
-	type OTG_FRAME_COUNT;\
-	type OTG_V_BLANK;\
-	type OTG_V_ACTIVE_DISP;\
-	type OTG_HORZ_COUNT;\
-	type OTG_VERT_COUNT;\
-	type OTG_VERT_COUNT_NOM;\
-	type OTG_BLACK_COLOR_B_CB;\
-	type OTG_BLACK_COLOR_G_Y;\
-	type OTG_BLACK_COLOR_R_CR;\
-	type OTG_BLANK_DATA_COLOR_BLUE_CB;\
-	type OTG_BLANK_DATA_COLOR_GREEN_Y;\
-	type OTG_BLANK_DATA_COLOR_RED_CR;\
-	type OTG_BLANK_DATA_COLOR_BLUE_CB_EXT;\
-	type OTG_BLANK_DATA_COLOR_GREEN_Y_EXT;\
-	type OTG_BLANK_DATA_COLOR_RED_CR_EXT;\
-	type OTG_VTOTAL_MID_REPLACING_MIN_EN;\
-	type OTG_TEST_PATTERN_INC0;\
-	type OTG_TEST_PATTERN_INC1;\
-	type OTG_TEST_PATTERN_VRES;\
-	type OTG_TEST_PATTERN_HRES;\
-	type OTG_TEST_PATTERN_RAMP0_OFFSET;\
-	type OTG_TEST_PATTERN_EN;\
-	type OTG_TEST_PATTERN_MODE;\
-	type OTG_TEST_PATTERN_DYNAMIC_RANGE;\
-	type OTG_TEST_PATTERN_COLOR_FORMAT;\
-	type OTG_TEST_PATTERN_MASK;\
-	type OTG_TEST_PATTERN_DATA;\
-	type OTG_BUSY;\
-	type OTG_CLOCK_EN;\
-	type OTG_CLOCK_ON;\
-	type OTG_CLOCK_GATE_DIS;\
-	type OTG_VERTICAL_INTERRUPT0_INT_ENABLE;\
-	type OTG_VERTICAL_INTERRUPT0_LINE_START;\
-	type OTG_VERTICAL_INTERRUPT0_LINE_END;\
-	type OTG_VERTICAL_INTERRUPT1_INT_ENABLE;\
-	type OTG_VERTICAL_INTERRUPT1_LINE_START;\
-	type OTG_VERTICAL_INTERRUPT2_INT_ENABLE;\
-	type OTG_VERTICAL_INTERRUPT2_LINE_START;\
-	type OPTC_INPUT_CLK_EN;\
-	type OPTC_INPUT_CLK_ON;\
-	type OPTC_INPUT_CLK_GATE_DIS;\
-	type OPTC_UNDERFLOW_OCCURRED_STATUS;\
-	type OPTC_UNDERFLOW_CLEAR;\
-	type OPTC_SRC_SEL;\
-	type VTG0_ENABLE;\
-	type VTG0_FP2;\
-	type VTG0_VCOUNT_INIT;\
-	type OTG_FORCE_VSYNC_NEXT_LINE_OCCURRED;\
-	type OTG_FORCE_VSYNC_NEXT_LINE_CLEAR;\
-	type OTG_AUTO_FORCE_VSYNC_MODE;\
-	type MASTER_UPDATE_INTERLACED_MODE;\
-	type OTG_GSL0_EN;\
-	type OTG_GSL1_EN;\
-	type OTG_GSL2_EN;\
-	type OTG_GSL_MASTER_EN;\
-	type OTG_GSL_FORCE_DELAY;\
-	type OTG_GSL_CHECK_ALL_FIELDS;\
-	type OTG_GSL_WINDOW_START_X;\
-	type OTG_GSL_WINDOW_END_X;\
-	type OTG_GSL_WINDOW_START_Y;\
-	type OTG_GSL_WINDOW_END_Y;\
-	type OTG_RANGE_TIMING_DBUF_UPDATE_MODE;\
-	type OTG_GSL_MASTER_MODE;\
-	type OTG_MASTER_UPDATE_LOCK_GSL_EN;\
-	type MASTER_UPDATE_LOCK_VUPDATE_KEEPOUT_START_OFFSET;\
-	type MASTER_UPDATE_LOCK_VUPDATE_KEEPOUT_END_OFFSET;\
-	type OTG_MASTER_UPDATE_LOCK_VUPDATE_KEEPOUT_EN;\
-	type OTG_CRC_CONT_EN;\
-	type OTG_CRC0_SELECT;\
-	type OTG_CRC_EN;\
-	type CRC0_R_CR;\
-	type CRC0_G_Y;\
-	type CRC0_B_CB;\
-	type OTG_CRC0_WINDOWA_X_START;\
-	type OTG_CRC0_WINDOWA_X_END;\
-	type OTG_CRC0_WINDOWA_Y_START;\
-	type OTG_CRC0_WINDOWA_Y_END;\
-	type OTG_CRC0_WINDOWB_X_START;\
-	type OTG_CRC0_WINDOWB_X_END;\
-	type OTG_CRC0_WINDOWB_Y_START;\
-	type OTG_CRC0_WINDOWB_Y_END;\
-	type GSL0_READY_SOURCE_SEL;\
-	type GSL1_READY_SOURCE_SEL;\
-	type GSL2_READY_SOURCE_SEL;\
-	type MANUAL_FLOW_CONTROL;\
-	type MANUAL_FLOW_CONTROL_SEL;
-
-#define TG_REG_FIELD_LIST(type) \
-	TG_REG_FIELD_LIST_DCN1_0(type)\
-	type OTG_V_SYNC_MODE;\
-	type OTG_DRR_TRIGGER_WINDOW_START_X;\
-	type OTG_DRR_TRIGGER_WINDOW_END_X;\
-	type OTG_DRR_V_TOTAL_CHANGE_LIMIT;\
-	type OTG_OUT_MUX;\
-	type OTG_M_CONST_DTO_PHASE;\
-	type OTG_M_CONST_DTO_MODULO;\
-	type MASTER_UPDATE_LOCK_DB_X;\
-	type MASTER_UPDATE_LOCK_DB_Y;\
-	type MASTER_UPDATE_LOCK_DB_EN;\
-	type GLOBAL_UPDATE_LOCK_EN;\
-	type DIG_UPDATE_LOCATION;\
-	type OTG_DSC_START_POSITION_X;\
-	type OTG_DSC_START_POSITION_LINE_NUM;\
-	type OPTC_NUM_OF_INPUT_SEGMENT;\
-	type OPTC_SEG0_SRC_SEL;\
-	type OPTC_SEG1_SRC_SEL;\
-	type OPTC_SEG2_SRC_SEL;\
-	type OPTC_SEG3_SRC_SEL;\
-	type OPTC_MEM_SEL;\
-	type OPTC_DATA_FORMAT;\
-	type OPTC_DSC_MODE;\
-	type OPTC_DSC_BYTES_PER_PIXEL;\
-	type OPTC_DSC_SLICE_WIDTH;\
-	type OPTC_SEGMENT_WIDTH;\
-	type OPTC_DWB0_SOURCE_SELECT;\
-	type OPTC_DWB1_SOURCE_SELECT;\
-	type MASTER_UPDATE_LOCK_DB_START_X;\
-	type MASTER_UPDATE_LOCK_DB_END_X;\
-	type MASTER_UPDATE_LOCK_DB_START_Y;\
-	type MASTER_UPDATE_LOCK_DB_END_Y;\
-	type DIG_UPDATE_POSITION_X;\
-	type DIG_UPDATE_POSITION_Y;\
-	type OTG_H_TIMING_DIV_MODE;\
-	type OTG_DRR_TIMING_DBUF_UPDATE_MODE;\
-	type OTG_CRC_DSC_MODE;\
-	type OTG_CRC_DATA_STREAM_COMBINE_MODE;\
-	type OTG_CRC_DATA_STREAM_SPLIT_MODE;\
-	type OTG_CRC_DATA_FORMAT;\
-	type OTG_V_TOTAL_LAST_USED_BY_DRR;
-
-
-struct dcn_optc_shift {
-	TG_REG_FIELD_LIST(uint8_t)
-};
-
-struct dcn_optc_mask {
-	TG_REG_FIELD_LIST(uint32_t)
-};
-
-struct optc {
-	struct timing_generator base;
-
-	const struct dcn_optc_registers *tg_regs;
-	const struct dcn_optc_shift *tg_shift;
-	const struct dcn_optc_mask *tg_mask;
-
-	int opp_count;
-
-	uint32_t max_h_total;
-	uint32_t max_v_total;
-
-	uint32_t min_h_blank;
-
-	uint32_t min_h_sync_width;
-	uint32_t min_v_sync_width;
-	uint32_t min_v_blank;
-	uint32_t min_v_blank_interlace;
-
-	int vstartup_start;
-	int vupdate_offset;
-	int vupdate_width;
-	int vready_offset;
-	enum signal_type signal;
-};
-
-void dcn10_timing_generator_init(struct optc *optc);
-
-struct dcn_otg_state {
-	uint32_t v_blank_start;
-	uint32_t v_blank_end;
-	uint32_t v_sync_a_pol;
-	uint32_t v_total;
-	uint32_t v_total_max;
-	uint32_t v_total_min;
-	uint32_t v_total_min_sel;
-	uint32_t v_total_max_sel;
-	uint32_t v_sync_a_start;
-	uint32_t v_sync_a_end;
-	uint32_t h_blank_start;
-	uint32_t h_blank_end;
-	uint32_t h_sync_a_start;
-	uint32_t h_sync_a_end;
-	uint32_t h_sync_a_pol;
-	uint32_t h_total;
-	uint32_t underflow_occurred_status;
-	uint32_t otg_enabled;
-	uint32_t blank_enabled;
-	uint32_t vertical_interrupt2_en;
-	uint32_t vertical_interrupt2_line;
-};
-
-void optc1_read_otg_state(struct optc *optc1,
-		struct dcn_otg_state *s);
-
-bool optc1_get_hw_timing(struct timing_generator *tg,
-		struct dc_crtc_timing *hw_crtc_timing);
-
-bool optc1_validate_timing(
-	struct timing_generator *optc,
-	const struct dc_crtc_timing *timing);
-
-void optc1_program_timing(
-	struct timing_generator *optc,
-	const struct dc_crtc_timing *dc_crtc_timing,
-	int vready_offset,
-	int vstartup_start,
-	int vupdate_offset,
-	int vupdate_width,
-	const enum signal_type signal,
-	bool use_vbios);
-
-void optc1_setup_vertical_interrupt0(
-		struct timing_generator *optc,
-		uint32_t start_line,
-		uint32_t end_line);
-void optc1_setup_vertical_interrupt1(
-		struct timing_generator *optc,
-		uint32_t start_line);
-void optc1_setup_vertical_interrupt2(
-		struct timing_generator *optc,
-		uint32_t start_line);
-
-void optc1_program_global_sync(
-		struct timing_generator *optc,
-		int vready_offset,
-		int vstartup_start,
-		int vupdate_offset,
-		int vupdate_width);
-
-bool optc1_disable_crtc(struct timing_generator *optc);
-
-bool optc1_is_counter_moving(struct timing_generator *optc);
-
-void optc1_get_position(struct timing_generator *optc,
-		struct crtc_position *position);
-
-uint32_t optc1_get_vblank_counter(struct timing_generator *optc);
-
-void optc1_get_crtc_scanoutpos(
-	struct timing_generator *optc,
-	uint32_t *v_blank_start,
-	uint32_t *v_blank_end,
-	uint32_t *h_position,
-	uint32_t *v_position);
-
-void optc1_set_early_control(
-	struct timing_generator *optc,
-	uint32_t early_cntl);
-
-void optc1_wait_for_state(struct timing_generator *optc,
-		enum crtc_state state);
-
-void optc1_set_blank(struct timing_generator *optc,
-		bool enable_blanking);
-
-bool optc1_is_blanked(struct timing_generator *optc);
-bool optc1_is_locked(struct timing_generator *optc);
-
-void optc1_program_blank_color(
-		struct timing_generator *optc,
-		const struct tg_color *black_color);
-
-bool optc1_did_triggered_reset_occur(
-	struct timing_generator *optc);
-
-void optc1_enable_reset_trigger(struct timing_generator *optc, int source_tg_inst);
-
-void optc1_disable_reset_trigger(struct timing_generator *optc);
-
-void optc1_lock(struct timing_generator *optc);
-
-void optc1_unlock(struct timing_generator *optc);
-
-void optc1_enable_optc_clock(struct timing_generator *optc, bool enable);
-
-void optc1_set_drr(
-	struct timing_generator *optc,
-	const struct drr_params *params);
-
-void optc1_set_vtotal_min_max(struct timing_generator *optc, int vtotal_min, int vtotal_max);
-
-void optc1_set_static_screen_control(
-	struct timing_generator *optc,
-	uint32_t event_triggers,
-	uint32_t num_frames);
-
-void optc1_program_stereo(struct timing_generator *optc,
-	const struct dc_crtc_timing *timing, struct crtc_stereo_flags *flags);
-
-bool optc1_is_stereo_left_eye(struct timing_generator *optc);
-
-void optc1_clear_optc_underflow(struct timing_generator *optc);
-
-void optc1_tg_init(struct timing_generator *optc);
-
-bool optc1_is_tg_enabled(struct timing_generator *optc);
-
-bool optc1_is_optc_underflow_occurred(struct timing_generator *optc);
-
-void optc1_set_blank_data_double_buffer(struct timing_generator *optc, bool enable);
-
-void optc1_set_timing_double_buffer(struct timing_generator *optc, bool enable);
-
-bool optc1_get_otg_active_size(struct timing_generator *optc,
-		uint32_t *otg_active_width,
-		uint32_t *otg_active_height);
-
-void optc1_enable_crtc_reset(
-		struct timing_generator *optc,
-		int source_tg_inst,
-		struct crtc_trigger_info *crtc_tp);
-
-bool optc1_configure_crc(struct timing_generator *optc,
-			  const struct crc_params *params);
-
-bool optc1_get_crc(struct timing_generator *optc,
-		    uint32_t *r_cr, uint32_t *g_y, uint32_t *b_cb);
-
-bool optc1_is_two_pixels_per_containter(const struct dc_crtc_timing *timing);
-
-void optc1_set_vtg_params(struct timing_generator *optc,
-		const struct dc_crtc_timing *dc_crtc_timing, bool program_fp2);
-
-#endif /* __DC_TIMING_GENERATOR_DCN10_H__ */
diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/Makefile b/drivers/gpu/drm/amd/display/dc/dcn20/Makefile
index 5fcaf78334ff..25ba0d310d46 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn20/Makefile
+++ b/drivers/gpu/drm/amd/display/dc/dcn20/Makefile
@@ -1,38 +1,7 @@
 # SPDX-License-Identifier: MIT
-#
-# Makefile for DCN.
+# Copyright © 2019-2024 Advanced Micro Devices, Inc. All rights reserved.
 
-DCN20 = dcn20_resource.o dcn20_init.o dcn20_hwseq.o dcn20_dpp.o dcn20_dpp_cm.o dcn20_hubp.o \
-		dcn20_mpc.o dcn20_opp.o dcn20_hubbub.o dcn20_optc.o dcn20_mmhubbub.o \
-		dcn20_stream_encoder.o dcn20_link_encoder.o dcn20_dccg.o \
-		dcn20_vmid.o dcn20_dwb.o dcn20_dwb_scl.o
-
-DCN20 += dcn20_dsc.o
-
-ifdef CONFIG_X86
-CFLAGS_$(AMDDALPATH)/dc/dcn20/dcn20_resource.o := -mhard-float -msse
-endif
-
-ifdef CONFIG_PPC64
-CFLAGS_$(AMDDALPATH)/dc/dcn20/dcn20_resource.o := -mhard-float -maltivec
-endif
-
-ifdef CONFIG_CC_IS_GCC
-ifeq ($(call cc-ifversion, -lt, 0701, y), y)
-IS_OLD_GCC = 1
-endif
-endif
-
-ifdef CONFIG_X86
-ifdef IS_OLD_GCC
-# Stack alignment mismatch, proceed with caution.
-# GCC < 7.1 cannot compile code using `double` and -mpreferred-stack-boundary=3
-# (8B stack alignment).
-CFLAGS_$(AMDDALPATH)/dc/dcn20/dcn20_resource.o += -mpreferred-stack-boundary=4
-else
-CFLAGS_$(AMDDALPATH)/dc/dcn20/dcn20_resource.o += -msse2
-endif
-endif
+DCN20 = dcn20_vmid.o dcn20_dwb.o dcn20_dwb_scl.o
 
 AMD_DAL_DCN20 = $(addprefix $(AMDDALPATH)/dc/dcn20/,$(DCN20))
 
diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_dccg.h b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_dccg.h
deleted file mode 100644
index f98aba308028..000000000000
--- a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_dccg.h
+++ /dev/null
@@ -1,273 +0,0 @@
-/*
- * Copyright 2018 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: AMD
- *
- */
-
-#ifndef __DCN20_DCCG_H__
-#define __DCN20_DCCG_H__
-
-#include "dccg.h"
-
-#define DCCG_COMMON_REG_LIST_DCN_BASE() \
-	SR(DPPCLK_DTO_CTRL),\
-	DCCG_SRII(DTO_PARAM, DPPCLK, 0),\
-	DCCG_SRII(DTO_PARAM, DPPCLK, 1),\
-	DCCG_SRII(DTO_PARAM, DPPCLK, 2),\
-	DCCG_SRII(DTO_PARAM, DPPCLK, 3),\
-	SR(REFCLK_CNTL),\
-	DCCG_SRII(PIXEL_RATE_CNTL, OTG, 0),\
-	DCCG_SRII(PIXEL_RATE_CNTL, OTG, 1),\
-	SR(DISPCLK_FREQ_CHANGE_CNTL)
-
-#define DCCG_REG_LIST_DCN2() \
-	DCCG_COMMON_REG_LIST_DCN_BASE(),\
-	DCCG_SRII(DTO_PARAM, DPPCLK, 4),\
-	DCCG_SRII(DTO_PARAM, DPPCLK, 5),\
-	DCCG_SRII(PIXEL_RATE_CNTL, OTG, 2),\
-	DCCG_SRII(PIXEL_RATE_CNTL, OTG, 3),\
-	DCCG_SRII(PIXEL_RATE_CNTL, OTG, 4),\
-	DCCG_SRII(PIXEL_RATE_CNTL, OTG, 5)
-
-#define DCCG_SF(reg_name, field_name, post_fix)\
-	.field_name = reg_name ## __ ## field_name ## post_fix
-
-#define DCCG_SFI(reg_name, field_name, field_prefix, inst, post_fix)\
-	.field_prefix ## _ ## field_name[inst] = reg_name ## __ ## field_prefix ## inst ## _ ## field_name ## post_fix
-
-#define DCCG_SFII(block, reg_name, field_prefix, field_name, inst, post_fix)\
-	.field_prefix ## _ ## field_name[inst] = block ## inst ## _ ## reg_name ## __ ## field_prefix ## inst ## _ ## field_name ## post_fix
-
-#define DCCG_COMMON_MASK_SH_LIST_DCN_COMMON_BASE(mask_sh) \
-	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_ENABLE, DPPCLK, 0, mask_sh),\
-	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_DB_EN, DPPCLK, 0, mask_sh),\
-	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_ENABLE, DPPCLK, 1, mask_sh),\
-	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_DB_EN, DPPCLK, 1, mask_sh),\
-	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_ENABLE, DPPCLK, 2, mask_sh),\
-	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_DB_EN, DPPCLK, 2, mask_sh),\
-	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_ENABLE, DPPCLK, 3, mask_sh),\
-	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_DB_EN, DPPCLK, 3, mask_sh),\
-	DCCG_SF(DPPCLK0_DTO_PARAM, DPPCLK0_DTO_PHASE, mask_sh),\
-	DCCG_SF(DPPCLK0_DTO_PARAM, DPPCLK0_DTO_MODULO, mask_sh),\
-	DCCG_SF(REFCLK_CNTL, REFCLK_CLOCK_EN, mask_sh),\
-	DCCG_SF(REFCLK_CNTL, REFCLK_SRC_SEL, mask_sh),\
-	DCCG_SF(DISPCLK_FREQ_CHANGE_CNTL, DISPCLK_STEP_DELAY, mask_sh),\
-	DCCG_SF(DISPCLK_FREQ_CHANGE_CNTL, DISPCLK_STEP_SIZE, mask_sh),\
-	DCCG_SF(DISPCLK_FREQ_CHANGE_CNTL, DISPCLK_FREQ_RAMP_DONE, mask_sh),\
-	DCCG_SF(DISPCLK_FREQ_CHANGE_CNTL, DISPCLK_MAX_ERRDET_CYCLES, mask_sh),\
-	DCCG_SF(DISPCLK_FREQ_CHANGE_CNTL, DCCG_FIFO_ERRDET_RESET, mask_sh),\
-	DCCG_SF(DISPCLK_FREQ_CHANGE_CNTL, DCCG_FIFO_ERRDET_STATE, mask_sh),\
-	DCCG_SF(DISPCLK_FREQ_CHANGE_CNTL, DCCG_FIFO_ERRDET_OVR_EN, mask_sh),\
-	DCCG_SF(DISPCLK_FREQ_CHANGE_CNTL, DISPCLK_CHG_FWD_CORR_DISABLE, mask_sh),\
-	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, ADD_PIXEL, 0, mask_sh),\
-	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, ADD_PIXEL, 1, mask_sh),\
-	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, DROP_PIXEL, 0, mask_sh),\
-	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, DROP_PIXEL, 1, mask_sh)
-
-
-
-
-#define DCCG_MASK_SH_LIST_DCN2(mask_sh) \
-	DCCG_COMMON_MASK_SH_LIST_DCN_COMMON_BASE(mask_sh),\
-	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_ENABLE, DPPCLK, 4, mask_sh),\
-	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_DB_EN, DPPCLK, 4, mask_sh),\
-	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_ENABLE, DPPCLK, 5, mask_sh),\
-	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_DB_EN, DPPCLK, 5, mask_sh),\
-	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, ADD_PIXEL, 2, mask_sh),\
-	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, ADD_PIXEL, 3, mask_sh),\
-	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, ADD_PIXEL, 4, mask_sh),\
-	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, ADD_PIXEL, 5, mask_sh),\
-	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, DROP_PIXEL, 2, mask_sh),\
-	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, DROP_PIXEL, 3, mask_sh),\
-	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, DROP_PIXEL, 4, mask_sh),\
-	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, DROP_PIXEL, 5, mask_sh)
-
-#define DCCG_MASK_SH_LIST_DCN2_1(mask_sh) \
-	DCCG_COMMON_MASK_SH_LIST_DCN_COMMON_BASE(mask_sh),\
-	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_ENABLE, DPPCLK, 4, mask_sh),\
-	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_DB_EN, DPPCLK, 4, mask_sh),\
-	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_ENABLE, DPPCLK, 5, mask_sh),\
-	DCCG_SFI(DPPCLK_DTO_CTRL, DTO_DB_EN, DPPCLK, 5, mask_sh),\
-	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, ADD_PIXEL, 2, mask_sh),\
-	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, ADD_PIXEL, 3, mask_sh),\
-	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, DROP_PIXEL, 2, mask_sh),\
-	DCCG_SFII(OTG, PIXEL_RATE_CNTL, OTG, DROP_PIXEL, 3, mask_sh)
-
-
-#define DCCG_REG_FIELD_LIST(type) \
-	type DPPCLK0_DTO_PHASE;\
-	type DPPCLK0_DTO_MODULO;\
-	type DPPCLK_DTO_ENABLE[6];\
-	type DPPCLK_DTO_DB_EN[6];\
-	type REFCLK_CLOCK_EN;\
-	type REFCLK_SRC_SEL;\
-	type DISPCLK_STEP_DELAY;\
-	type DISPCLK_STEP_SIZE;\
-	type DISPCLK_FREQ_RAMP_DONE;\
-	type DISPCLK_MAX_ERRDET_CYCLES;\
-	type DCCG_FIFO_ERRDET_RESET;\
-	type DCCG_FIFO_ERRDET_STATE;\
-	type DCCG_FIFO_ERRDET_OVR_EN;\
-	type DISPCLK_CHG_FWD_CORR_DISABLE;\
-	type DISPCLK_FREQ_CHANGE_CNTL;\
-	type OTG_ADD_PIXEL[MAX_PIPES];\
-	type OTG_DROP_PIXEL[MAX_PIPES];
-
-#define DCCG3_REG_FIELD_LIST(type) \
-	type PHYASYMCLK_FORCE_EN;\
-	type PHYASYMCLK_FORCE_SRC_SEL;\
-	type PHYBSYMCLK_FORCE_EN;\
-	type PHYBSYMCLK_FORCE_SRC_SEL;\
-	type PHYCSYMCLK_FORCE_EN;\
-	type PHYCSYMCLK_FORCE_SRC_SEL;
-
-#define DCCG31_REG_FIELD_LIST(type) \
-	type PHYDSYMCLK_FORCE_EN;\
-	type PHYDSYMCLK_FORCE_SRC_SEL;\
-	type PHYESYMCLK_FORCE_EN;\
-	type PHYESYMCLK_FORCE_SRC_SEL;\
-	type DPSTREAMCLK_PIPE0_EN;\
-	type DPSTREAMCLK_PIPE1_EN;\
-	type DPSTREAMCLK_PIPE2_EN;\
-	type DPSTREAMCLK_PIPE3_EN;\
-	type HDMISTREAMCLK0_SRC_SEL;\
-	type HDMISTREAMCLK0_DTO_FORCE_DIS;\
-	type SYMCLK32_SE0_SRC_SEL;\
-	type SYMCLK32_SE1_SRC_SEL;\
-	type SYMCLK32_SE2_SRC_SEL;\
-	type SYMCLK32_SE3_SRC_SEL;\
-	type SYMCLK32_SE0_EN;\
-	type SYMCLK32_SE1_EN;\
-	type SYMCLK32_SE2_EN;\
-	type SYMCLK32_SE3_EN;\
-	type SYMCLK32_LE0_SRC_SEL;\
-	type SYMCLK32_LE1_SRC_SEL;\
-	type SYMCLK32_LE0_EN;\
-	type SYMCLK32_LE1_EN;\
-	type DTBCLK_DTO_ENABLE[MAX_PIPES];\
-	type DTBCLKDTO_ENABLE_STATUS[MAX_PIPES];\
-	type PIPE_DTO_SRC_SEL[MAX_PIPES];\
-	type DTBCLK_DTO_DIV[MAX_PIPES];\
-	type DCCG_AUDIO_DTO_SEL;\
-	type DCCG_AUDIO_DTO0_SOURCE_SEL;\
-	type DENTIST_DISPCLK_CHG_MODE;\
-	type DSCCLK0_DTO_PHASE;\
-	type DSCCLK0_DTO_MODULO;\
-	type DSCCLK1_DTO_PHASE;\
-	type DSCCLK1_DTO_MODULO;\
-	type DSCCLK2_DTO_PHASE;\
-	type DSCCLK2_DTO_MODULO;\
-	type DSCCLK0_DTO_ENABLE;\
-	type DSCCLK1_DTO_ENABLE;\
-	type DSCCLK2_DTO_ENABLE;\
-	type SYMCLK32_ROOT_SE0_GATE_DISABLE;\
-	type SYMCLK32_ROOT_SE1_GATE_DISABLE;\
-	type SYMCLK32_ROOT_SE2_GATE_DISABLE;\
-	type SYMCLK32_ROOT_SE3_GATE_DISABLE;\
-	type SYMCLK32_ROOT_LE0_GATE_DISABLE;\
-	type SYMCLK32_ROOT_LE1_GATE_DISABLE;\
-	type DPSTREAMCLK_ROOT_GATE_DISABLE;\
-	type DPSTREAMCLK_GATE_DISABLE;\
-	type HDMISTREAMCLK0_DTO_PHASE;\
-	type HDMISTREAMCLK0_DTO_MODULO;\
-	type HDMICHARCLK0_GATE_DISABLE;\
-	type HDMICHARCLK0_ROOT_GATE_DISABLE;
-
-
-struct dccg_shift {
-	DCCG_REG_FIELD_LIST(uint8_t)
-	DCCG3_REG_FIELD_LIST(uint8_t)
-	DCCG31_REG_FIELD_LIST(uint8_t)
-};
-
-struct dccg_mask {
-	DCCG_REG_FIELD_LIST(uint32_t)
-	DCCG3_REG_FIELD_LIST(uint32_t)
-	DCCG31_REG_FIELD_LIST(uint32_t)
-};
-
-struct dccg_registers {
-	uint32_t DPPCLK_DTO_CTRL;
-	uint32_t DPPCLK_DTO_PARAM[6];
-	uint32_t REFCLK_CNTL;
-	uint32_t DISPCLK_FREQ_CHANGE_CNTL;
-	uint32_t OTG_PIXEL_RATE_CNTL[MAX_PIPES];
-	uint32_t HDMICHARCLK_CLOCK_CNTL[6];
-	uint32_t PHYASYMCLK_CLOCK_CNTL;
-	uint32_t PHYBSYMCLK_CLOCK_CNTL;
-	uint32_t PHYCSYMCLK_CLOCK_CNTL;
-	uint32_t PHYDSYMCLK_CLOCK_CNTL;
-	uint32_t PHYESYMCLK_CLOCK_CNTL;
-	uint32_t DTBCLK_DTO_MODULO[MAX_PIPES];
-	uint32_t DTBCLK_DTO_PHASE[MAX_PIPES];
-	uint32_t DCCG_AUDIO_DTBCLK_DTO_MODULO;
-	uint32_t DCCG_AUDIO_DTBCLK_DTO_PHASE;
-	uint32_t DCCG_AUDIO_DTO_SOURCE;
-	uint32_t DPSTREAMCLK_CNTL;
-	uint32_t HDMISTREAMCLK_CNTL;
-	uint32_t SYMCLK32_SE_CNTL;
-	uint32_t SYMCLK32_LE_CNTL;
-	uint32_t DENTIST_DISPCLK_CNTL;
-	uint32_t DSCCLK_DTO_CTRL;
-	uint32_t DSCCLK0_DTO_PARAM;
-	uint32_t DSCCLK1_DTO_PARAM;
-	uint32_t DSCCLK2_DTO_PARAM;
-	uint32_t DPSTREAMCLK_ROOT_GATE_DISABLE;
-	uint32_t DPSTREAMCLK_GATE_DISABLE;
-	uint32_t DCCG_GATE_DISABLE_CNTL3;
-	uint32_t HDMISTREAMCLK0_DTO_PARAM;
-	uint32_t DCCG_GATE_DISABLE_CNTL4;
-
-};
-
-struct dcn_dccg {
-	struct dccg base;
-	const struct dccg_registers *regs;
-	const struct dccg_shift *dccg_shift;
-	const struct dccg_mask *dccg_mask;
-};
-
-void dccg2_update_dpp_dto(struct dccg *dccg, int dpp_inst, int req_dppclk);
-
-void dccg2_get_dccg_ref_freq(struct dccg *dccg,
-		unsigned int xtalin_freq_inKhz,
-		unsigned int *dccg_ref_freq_inKhz);
-
-void dccg2_set_fifo_errdet_ovr_en(struct dccg *dccg,
-		bool en);
-void dccg2_otg_add_pixel(struct dccg *dccg,
-		uint32_t otg_inst);
-void dccg2_otg_drop_pixel(struct dccg *dccg,
-		uint32_t otg_inst);
-
-
-void dccg2_init(struct dccg *dccg);
-
-struct dccg *dccg2_create(
-	struct dc_context *ctx,
-	const struct dccg_registers *regs,
-	const struct dccg_shift *dccg_shift,
-	const struct dccg_mask *dccg_mask);
-
-void dcn_dccg_destroy(struct dccg **dccg);
-
-#endif //__DCN20_DCCG_H__
diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_dwb.c b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_dwb.c
index 8d3884b306dd..80779e85e2c5 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_dwb.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_dwb.c
@@ -101,8 +101,8 @@ static bool dwb2_enable(struct dwbc *dwbc, struct dc_dwb_params *params)
 	struct dcn20_dwbc *dwbc20 = TO_DCN20_DWBC(dwbc);
 
 	/* Only chroma scaling (sub-sampling) is supported in DCN2 */
-if ((params->cnv_params.src_width  != params->dest_width) ||
-		(params->cnv_params.src_height != params->dest_height)) {
+	if ((params->cnv_params.src_width  != params->dest_width) ||
+	    (params->cnv_params.src_height != params->dest_height)) {
 
 		DC_LOG_DWB("%s inst = %d, FAILED!LUMA SCALING NOT SUPPORTED", __func__, dwbc20->base.inst);
 		return false;
@@ -299,9 +299,20 @@ void dwb2_set_scaler(struct dwbc *dwbc, struct dc_dwb_params *params)
 		}
 	}
 
+
+	if (dwbc20->dwbc_mask->WBSCL_COEF_RAM_SEL) {
+		/* Swap double buffered coefficient set */
+		uint32_t wbscl_mode = REG_READ(WBSCL_MODE);
+		bool coef_ram_current = get_reg_field_value_ex(
+			wbscl_mode, dwbc20->dwbc_mask->WBSCL_COEF_RAM_SEL_CURRENT,
+			dwbc20->dwbc_shift->WBSCL_COEF_RAM_SEL_CURRENT);
+
+		REG_UPDATE(WBSCL_MODE, WBSCL_COEF_RAM_SEL, !coef_ram_current);
+	}
+
 }
 
-const struct dwbc_funcs dcn20_dwbc_funcs = {
+static const struct dwbc_funcs dcn20_dwbc_funcs = {
 	.get_caps		= dwb2_get_caps,
 	.enable			= dwb2_enable,
 	.disable		= dwb2_disable,
diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_dwb.h b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_dwb.h
index a85ed228dfc2..a9dd9ae23ec9 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_dwb.h
+++ b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_dwb.h
@@ -27,204 +27,177 @@
 #define TO_DCN20_DWBC(dwbc_base) \
 	container_of(dwbc_base, struct dcn20_dwbc, base)
 
-/* DCN */
-#define BASE_INNER(seg) \
-	DCE_BASE__INST0_SEG ## seg
-
-#define BASE(seg) \
-	BASE_INNER(seg)
-
-#define SR(reg_name)\
-		.reg_name = BASE(mm ## reg_name ## _BASE_IDX) +  \
-					mm ## reg_name
-
-#define SRI(reg_name, block, id)\
-	.reg_name = BASE(mm ## block ## id ## _ ## reg_name ## _BASE_IDX) + \
-					mm ## block ## id ## _ ## reg_name
-
-#define SRI2(reg_name, block, id)\
-	.reg_name = BASE(mm ## reg_name ## _BASE_IDX) + \
-					mm ## reg_name
-
-#define SRII(reg_name, block, id)\
-	.reg_name[id] = BASE(mm ## block ## id ## _ ## reg_name ## _BASE_IDX) + \
-					mm ## block ## id ## _ ## reg_name
-
-#define SF(reg_name, field_name, post_fix)\
-	.field_name = reg_name ## __ ## field_name ## post_fix
-
-
 #define DWBC_COMMON_REG_LIST_DCN2_0(inst) \
-	SRI2(WB_ENABLE, CNV, inst),\
-	SRI2(WB_EC_CONFIG, CNV, inst),\
-	SRI2(CNV_MODE, CNV, inst),\
-	SRI2(CNV_WINDOW_START, CNV, inst),\
-	SRI2(CNV_WINDOW_SIZE, CNV, inst),\
-	SRI2(CNV_UPDATE, CNV, inst),\
-	SRI2(CNV_SOURCE_SIZE, CNV, inst),\
-	SRI2(CNV_TEST_CNTL, CNV, inst),\
-	SRI2(CNV_TEST_CRC_RED, CNV, inst),\
-	SRI2(CNV_TEST_CRC_GREEN, CNV, inst),\
-	SRI2(CNV_TEST_CRC_BLUE, CNV, inst),\
-	SRI2(WBSCL_COEF_RAM_SELECT, WBSCL, inst),\
-	SRI2(WBSCL_COEF_RAM_TAP_DATA, WBSCL, inst),\
-	SRI2(WBSCL_MODE, WBSCL, inst),\
-	SRI2(WBSCL_TAP_CONTROL, WBSCL, inst),\
-	SRI2(WBSCL_DEST_SIZE, WBSCL, inst),\
-	SRI2(WBSCL_HORZ_FILTER_SCALE_RATIO, WBSCL, inst),\
-	SRI2(WBSCL_HORZ_FILTER_INIT_Y_RGB, WBSCL, inst),\
-	SRI2(WBSCL_HORZ_FILTER_INIT_CBCR, WBSCL, inst),\
-	SRI2(WBSCL_VERT_FILTER_SCALE_RATIO, WBSCL, inst),\
-	SRI2(WBSCL_VERT_FILTER_INIT_Y_RGB, WBSCL, inst),\
-	SRI2(WBSCL_VERT_FILTER_INIT_CBCR, WBSCL, inst),\
-	SRI2(WBSCL_ROUND_OFFSET, WBSCL, inst),\
-	SRI2(WBSCL_OVERFLOW_STATUS, WBSCL, inst),\
-	SRI2(WBSCL_COEF_RAM_CONFLICT_STATUS, WBSCL, inst),\
-	SRI2(WBSCL_TEST_CNTL, WBSCL, inst),\
-	SRI2(WBSCL_TEST_CRC_RED, WBSCL, inst),\
-	SRI2(WBSCL_TEST_CRC_GREEN, WBSCL, inst),\
-	SRI2(WBSCL_TEST_CRC_BLUE, WBSCL, inst),\
-	SRI2(WBSCL_BACKPRESSURE_CNT_EN, WBSCL, inst),\
-	SRI2(WB_MCIF_BACKPRESSURE_CNT, WBSCL, inst),\
-	SRI2(WBSCL_CLAMP_Y_RGB, WBSCL, inst),\
-	SRI2(WBSCL_CLAMP_CBCR, WBSCL, inst),\
-	SRI2(WBSCL_OUTSIDE_PIX_STRATEGY, WBSCL, inst),\
-	SRI2(WBSCL_OUTSIDE_PIX_STRATEGY_CBCR, WBSCL, inst),\
-	SRI2(WBSCL_DEBUG, WBSCL, inst),\
-	SRI2(WBSCL_TEST_DEBUG_INDEX, WBSCL, inst),\
-	SRI2(WBSCL_TEST_DEBUG_DATA, WBSCL, inst),\
-	SRI2(WB_DEBUG_CTRL, CNV, inst),\
-	SRI2(WB_DBG_MODE, CNV, inst),\
-	SRI2(WB_HW_DEBUG, CNV, inst),\
-	SRI2(CNV_TEST_DEBUG_INDEX, CNV, inst),\
-	SRI2(CNV_TEST_DEBUG_DATA, CNV, inst),\
-	SRI2(WB_SOFT_RESET, CNV, inst),\
-	SRI2(WB_WARM_UP_MODE_CTL1, CNV, inst),\
-	SRI2(WB_WARM_UP_MODE_CTL2, CNV, inst)
+	SRI2_DWB(WB_ENABLE, CNV, inst),\
+	SRI2_DWB(WB_EC_CONFIG, CNV, inst),\
+	SRI2_DWB(CNV_MODE, CNV, inst),\
+	SRI2_DWB(CNV_WINDOW_START, CNV, inst),\
+	SRI2_DWB(CNV_WINDOW_SIZE, CNV, inst),\
+	SRI2_DWB(CNV_UPDATE, CNV, inst),\
+	SRI2_DWB(CNV_SOURCE_SIZE, CNV, inst),\
+	SRI2_DWB(CNV_TEST_CNTL, CNV, inst),\
+	SRI2_DWB(CNV_TEST_CRC_RED, CNV, inst),\
+	SRI2_DWB(CNV_TEST_CRC_GREEN, CNV, inst),\
+	SRI2_DWB(CNV_TEST_CRC_BLUE, CNV, inst),\
+	SRI2_DWB(WBSCL_COEF_RAM_SELECT, WBSCL, inst),\
+	SRI2_DWB(WBSCL_COEF_RAM_TAP_DATA, WBSCL, inst),\
+	SRI2_DWB(WBSCL_MODE, WBSCL, inst),\
+	SRI2_DWB(WBSCL_TAP_CONTROL, WBSCL, inst),\
+	SRI2_DWB(WBSCL_DEST_SIZE, WBSCL, inst),\
+	SRI2_DWB(WBSCL_HORZ_FILTER_SCALE_RATIO, WBSCL, inst),\
+	SRI2_DWB(WBSCL_HORZ_FILTER_INIT_Y_RGB, WBSCL, inst),\
+	SRI2_DWB(WBSCL_HORZ_FILTER_INIT_CBCR, WBSCL, inst),\
+	SRI2_DWB(WBSCL_VERT_FILTER_SCALE_RATIO, WBSCL, inst),\
+	SRI2_DWB(WBSCL_VERT_FILTER_INIT_Y_RGB, WBSCL, inst),\
+	SRI2_DWB(WBSCL_VERT_FILTER_INIT_CBCR, WBSCL, inst),\
+	SRI2_DWB(WBSCL_ROUND_OFFSET, WBSCL, inst),\
+	SRI2_DWB(WBSCL_OVERFLOW_STATUS, WBSCL, inst),\
+	SRI2_DWB(WBSCL_COEF_RAM_CONFLICT_STATUS, WBSCL, inst),\
+	SRI2_DWB(WBSCL_TEST_CNTL, WBSCL, inst),\
+	SRI2_DWB(WBSCL_TEST_CRC_RED, WBSCL, inst),\
+	SRI2_DWB(WBSCL_TEST_CRC_GREEN, WBSCL, inst),\
+	SRI2_DWB(WBSCL_TEST_CRC_BLUE, WBSCL, inst),\
+	SRI2_DWB(WBSCL_BACKPRESSURE_CNT_EN, WBSCL, inst),\
+	SRI2_DWB(WB_MCIF_BACKPRESSURE_CNT, WBSCL, inst),\
+	SRI2_DWB(WBSCL_CLAMP_Y_RGB, WBSCL, inst),\
+	SRI2_DWB(WBSCL_CLAMP_CBCR, WBSCL, inst),\
+	SRI2_DWB(WBSCL_OUTSIDE_PIX_STRATEGY, WBSCL, inst),\
+	SRI2_DWB(WBSCL_OUTSIDE_PIX_STRATEGY_CBCR, WBSCL, inst),\
+	SRI2_DWB(WBSCL_DEBUG, WBSCL, inst),\
+	SRI2_DWB(WBSCL_TEST_DEBUG_INDEX, WBSCL, inst),\
+	SRI2_DWB(WBSCL_TEST_DEBUG_DATA, WBSCL, inst),\
+	SRI2_DWB(WB_DEBUG_CTRL, CNV, inst),\
+	SRI2_DWB(WB_DBG_MODE, CNV, inst),\
+	SRI2_DWB(WB_HW_DEBUG, CNV, inst),\
+	SRI2_DWB(CNV_TEST_DEBUG_INDEX, CNV, inst),\
+	SRI2_DWB(CNV_TEST_DEBUG_DATA, CNV, inst),\
+	SRI2_DWB(WB_SOFT_RESET, CNV, inst),\
+	SRI2_DWB(WB_WARM_UP_MODE_CTL1, CNV, inst),\
+	SRI2_DWB(WB_WARM_UP_MODE_CTL2, CNV, inst)
 
 #define DWBC_COMMON_MASK_SH_LIST_DCN2_0(mask_sh) \
-	SF(WB_ENABLE, WB_ENABLE, mask_sh),\
-	SF(WB_EC_CONFIG, DISPCLK_R_WB_GATE_DIS, mask_sh),\
-	SF(WB_EC_CONFIG, DISPCLK_G_WB_GATE_DIS, mask_sh),\
-	SF(WB_EC_CONFIG, DISPCLK_G_WBSCL_GATE_DIS, mask_sh),\
-	SF(WB_EC_CONFIG, WB_TEST_CLK_SEL, mask_sh),\
-	SF(WB_EC_CONFIG, WB_LB_LS_DIS, mask_sh),\
-	SF(WB_EC_CONFIG, WB_LB_SD_DIS, mask_sh),\
-	SF(WB_EC_CONFIG, WB_LUT_LS_DIS, mask_sh),\
-	SF(WB_EC_CONFIG, WBSCL_LB_MEM_PWR_MODE_SEL, mask_sh),\
-	SF(WB_EC_CONFIG, WBSCL_LB_MEM_PWR_DIS, mask_sh),\
-	SF(WB_EC_CONFIG, WBSCL_LB_MEM_PWR_FORCE, mask_sh),\
-	SF(WB_EC_CONFIG, WBSCL_LB_MEM_PWR_STATE, mask_sh),\
-	SF(WB_EC_CONFIG, WB_RAM_PW_SAVE_MODE, mask_sh),\
-	SF(WB_EC_CONFIG, WBSCL_LUT_MEM_PWR_STATE, mask_sh),\
-	SF(CNV_MODE, CNV_OUT_BPC, mask_sh),\
-	SF(CNV_MODE, CNV_FRAME_CAPTURE_RATE, mask_sh),\
-	SF(CNV_MODE, CNV_WINDOW_CROP_EN, mask_sh),\
-	SF(CNV_MODE, CNV_STEREO_TYPE, mask_sh),\
-	SF(CNV_MODE, CNV_INTERLACED_MODE, mask_sh),\
-	SF(CNV_MODE, CNV_EYE_SELECTION, mask_sh),\
-	SF(CNV_MODE, CNV_STEREO_POLARITY, mask_sh),\
-	SF(CNV_MODE, CNV_INTERLACED_FIELD_ORDER, mask_sh),\
-	SF(CNV_MODE, CNV_STEREO_SPLIT, mask_sh),\
-	SF(CNV_MODE, CNV_NEW_CONTENT, mask_sh),\
-	SF(CNV_MODE, CNV_FRAME_CAPTURE_EN_CURRENT, mask_sh),\
-	SF(CNV_MODE, CNV_FRAME_CAPTURE_EN, mask_sh),\
-	SF(CNV_WINDOW_START, CNV_WINDOW_START_X, mask_sh),\
-	SF(CNV_WINDOW_START, CNV_WINDOW_START_Y, mask_sh),\
-	SF(CNV_WINDOW_SIZE, CNV_WINDOW_WIDTH, mask_sh),\
-	SF(CNV_WINDOW_SIZE, CNV_WINDOW_HEIGHT, mask_sh),\
-	SF(CNV_UPDATE, CNV_UPDATE_PENDING, mask_sh),\
-	SF(CNV_UPDATE, CNV_UPDATE_TAKEN, mask_sh),\
-	SF(CNV_UPDATE, CNV_UPDATE_LOCK, mask_sh),\
-	SF(CNV_SOURCE_SIZE, CNV_SOURCE_WIDTH, mask_sh),\
-	SF(CNV_SOURCE_SIZE, CNV_SOURCE_HEIGHT, mask_sh),\
-	SF(CNV_TEST_CNTL, CNV_TEST_CRC_EN, mask_sh),\
-	SF(CNV_TEST_CNTL, CNV_TEST_CRC_CONT_EN, mask_sh),\
-	SF(CNV_TEST_CRC_RED, CNV_TEST_CRC_RED_MASK, mask_sh),\
-	SF(CNV_TEST_CRC_RED, CNV_TEST_CRC_SIG_RED, mask_sh),\
-	SF(CNV_TEST_CRC_GREEN, CNV_TEST_CRC_GREEN_MASK, mask_sh),\
-	SF(CNV_TEST_CRC_GREEN, CNV_TEST_CRC_SIG_GREEN, mask_sh),\
-	SF(CNV_TEST_CRC_BLUE, CNV_TEST_CRC_BLUE_MASK, mask_sh),\
-	SF(CNV_TEST_CRC_BLUE, CNV_TEST_CRC_SIG_BLUE, mask_sh),\
-	SF(WB_DEBUG_CTRL, WB_DEBUG_EN, mask_sh),\
-	SF(WB_DEBUG_CTRL, WB_DEBUG_SEL, mask_sh),\
-	SF(WB_DBG_MODE, WB_DBG_MODE_EN, mask_sh),\
-	SF(WB_DBG_MODE, WB_DBG_DIN_FMT, mask_sh),\
-	SF(WB_DBG_MODE, WB_DBG_36MODE, mask_sh),\
-	SF(WB_DBG_MODE, WB_DBG_CMAP, mask_sh),\
-	SF(WB_DBG_MODE, WB_DBG_PXLRATE_ERROR, mask_sh),\
-	SF(WB_DBG_MODE, WB_DBG_SOURCE_WIDTH, mask_sh),\
-	SF(WB_HW_DEBUG, WB_HW_DEBUG, mask_sh),\
-	SF(WB_SOFT_RESET, WB_SOFT_RESET, mask_sh),\
-	SF(CNV_TEST_DEBUG_INDEX, CNV_TEST_DEBUG_INDEX, mask_sh),\
-	SF(CNV_TEST_DEBUG_INDEX, CNV_TEST_DEBUG_WRITE_EN, mask_sh),\
-	SF(CNV_TEST_DEBUG_DATA, CNV_TEST_DEBUG_DATA, mask_sh),\
-	SF(WBSCL_COEF_RAM_SELECT, WBSCL_COEF_RAM_TAP_PAIR_IDX, mask_sh),\
-	SF(WBSCL_COEF_RAM_SELECT, WBSCL_COEF_RAM_PHASE, mask_sh),\
-	SF(WBSCL_COEF_RAM_SELECT, WBSCL_COEF_RAM_FILTER_TYPE, mask_sh),\
-	SF(WBSCL_COEF_RAM_TAP_DATA, WBSCL_COEF_RAM_EVEN_TAP_COEF, mask_sh),\
-	SF(WBSCL_COEF_RAM_TAP_DATA, WBSCL_COEF_RAM_EVEN_TAP_COEF_EN, mask_sh),\
-	SF(WBSCL_COEF_RAM_TAP_DATA, WBSCL_COEF_RAM_ODD_TAP_COEF, mask_sh),\
-	SF(WBSCL_COEF_RAM_TAP_DATA, WBSCL_COEF_RAM_ODD_TAP_COEF_EN, mask_sh),\
-	SF(WBSCL_MODE, WBSCL_MODE, mask_sh),\
-	SF(WBSCL_MODE, WBSCL_OUT_BIT_DEPTH, mask_sh),\
-	SF(WBSCL_TAP_CONTROL, WBSCL_V_NUM_OF_TAPS_Y_RGB, mask_sh),\
-	SF(WBSCL_TAP_CONTROL, WBSCL_V_NUM_OF_TAPS_CBCR, mask_sh),\
-	SF(WBSCL_TAP_CONTROL, WBSCL_H_NUM_OF_TAPS_Y_RGB, mask_sh),\
-	SF(WBSCL_TAP_CONTROL, WBSCL_H_NUM_OF_TAPS_CBCR, mask_sh),\
-	SF(WBSCL_DEST_SIZE, WBSCL_DEST_HEIGHT, mask_sh),\
-	SF(WBSCL_DEST_SIZE, WBSCL_DEST_WIDTH, mask_sh),\
-	SF(WBSCL_HORZ_FILTER_SCALE_RATIO, WBSCL_H_SCALE_RATIO, mask_sh),\
-	SF(WBSCL_HORZ_FILTER_INIT_Y_RGB, WBSCL_H_INIT_FRAC_Y_RGB, mask_sh),\
-	SF(WBSCL_HORZ_FILTER_INIT_Y_RGB, WBSCL_H_INIT_INT_Y_RGB, mask_sh),\
-	SF(WBSCL_HORZ_FILTER_INIT_CBCR, WBSCL_H_INIT_FRAC_CBCR, mask_sh),\
-	SF(WBSCL_HORZ_FILTER_INIT_CBCR, WBSCL_H_INIT_INT_CBCR, mask_sh),\
-	SF(WBSCL_VERT_FILTER_SCALE_RATIO, WBSCL_V_SCALE_RATIO, mask_sh),\
-	SF(WBSCL_VERT_FILTER_INIT_Y_RGB, WBSCL_V_INIT_FRAC_Y_RGB, mask_sh),\
-	SF(WBSCL_VERT_FILTER_INIT_Y_RGB, WBSCL_V_INIT_INT_Y_RGB, mask_sh),\
-	SF(WBSCL_VERT_FILTER_INIT_CBCR, WBSCL_V_INIT_FRAC_CBCR, mask_sh),\
-	SF(WBSCL_VERT_FILTER_INIT_CBCR, WBSCL_V_INIT_INT_CBCR, mask_sh),\
-	SF(WBSCL_ROUND_OFFSET, WBSCL_ROUND_OFFSET_Y_RGB, mask_sh),\
-	SF(WBSCL_ROUND_OFFSET, WBSCL_ROUND_OFFSET_CBCR, mask_sh),\
-	SF(WBSCL_OVERFLOW_STATUS, WBSCL_DATA_OVERFLOW_FLAG, mask_sh),\
-	SF(WBSCL_OVERFLOW_STATUS, WBSCL_DATA_OVERFLOW_ACK, mask_sh),\
-	SF(WBSCL_OVERFLOW_STATUS, WBSCL_DATA_OVERFLOW_MASK, mask_sh),\
-	SF(WBSCL_OVERFLOW_STATUS, WBSCL_DATA_OVERFLOW_INT_STATUS, mask_sh),\
-	SF(WBSCL_OVERFLOW_STATUS, WBSCL_DATA_OVERFLOW_INT_TYPE, mask_sh),\
-	SF(WBSCL_COEF_RAM_CONFLICT_STATUS, WBSCL_HOST_CONFLICT_FLAG, mask_sh),\
-	SF(WBSCL_COEF_RAM_CONFLICT_STATUS, WBSCL_HOST_CONFLICT_ACK, mask_sh),\
-	SF(WBSCL_COEF_RAM_CONFLICT_STATUS, WBSCL_HOST_CONFLICT_MASK, mask_sh),\
-	SF(WBSCL_COEF_RAM_CONFLICT_STATUS, WBSCL_HOST_CONFLICT_INT_STATUS, mask_sh),\
-	SF(WBSCL_COEF_RAM_CONFLICT_STATUS, WBSCL_HOST_CONFLICT_INT_TYPE, mask_sh),\
-	SF(WBSCL_TEST_CNTL, WBSCL_TEST_CRC_EN, mask_sh),\
-	SF(WBSCL_TEST_CNTL, WBSCL_TEST_CRC_CONT_EN, mask_sh),\
-	SF(WBSCL_TEST_CRC_RED, WBSCL_TEST_CRC_RED_MASK, mask_sh),\
-	SF(WBSCL_TEST_CRC_RED, WBSCL_TEST_CRC_SIG_RED, mask_sh),\
-	SF(WBSCL_TEST_CRC_GREEN, WBSCL_TEST_CRC_GREEN_MASK, mask_sh),\
-	SF(WBSCL_TEST_CRC_GREEN, WBSCL_TEST_CRC_SIG_GREEN, mask_sh),\
-	SF(WBSCL_TEST_CRC_BLUE, WBSCL_TEST_CRC_BLUE_MASK, mask_sh),\
-	SF(WBSCL_TEST_CRC_BLUE, WBSCL_TEST_CRC_SIG_BLUE, mask_sh),\
-	SF(WBSCL_BACKPRESSURE_CNT_EN, WBSCL_BACKPRESSURE_CNT_EN, mask_sh),\
-	SF(WB_MCIF_BACKPRESSURE_CNT, WB_MCIF_Y_MAX_BACKPRESSURE, mask_sh),\
-	SF(WB_MCIF_BACKPRESSURE_CNT, WB_MCIF_C_MAX_BACKPRESSURE, mask_sh),\
-	SF(WBSCL_CLAMP_Y_RGB, WBSCL_CLAMP_UPPER_Y_RGB, mask_sh),\
-	SF(WBSCL_CLAMP_Y_RGB, WBSCL_CLAMP_LOWER_Y_RGB, mask_sh),\
-	SF(WBSCL_CLAMP_CBCR, WBSCL_CLAMP_UPPER_CBCR, mask_sh),\
-	SF(WBSCL_CLAMP_CBCR, WBSCL_CLAMP_LOWER_CBCR, mask_sh),\
-	SF(WBSCL_OUTSIDE_PIX_STRATEGY, WBSCL_OUTSIDE_PIX_STRATEGY, mask_sh),\
-	SF(WBSCL_OUTSIDE_PIX_STRATEGY, WBSCL_BLACK_COLOR_G_Y, mask_sh),\
-	SF(WBSCL_OUTSIDE_PIX_STRATEGY_CBCR, WBSCL_BLACK_COLOR_B_CB, mask_sh),\
-	SF(WBSCL_OUTSIDE_PIX_STRATEGY_CBCR, WBSCL_BLACK_COLOR_R_CR, mask_sh),\
-	SF(WBSCL_DEBUG, WBSCL_DEBUG, mask_sh),\
-	SF(WBSCL_TEST_DEBUG_INDEX, WBSCL_TEST_DEBUG_INDEX, mask_sh),\
-	SF(WBSCL_TEST_DEBUG_INDEX, WBSCL_TEST_DEBUG_WRITE_EN, mask_sh),\
-	SF(WBSCL_TEST_DEBUG_DATA, WBSCL_TEST_DEBUG_DATA, mask_sh),\
-	SF(WB_WARM_UP_MODE_CTL1, WIDTH_WARMUP, mask_sh),\
-	SF(WB_WARM_UP_MODE_CTL1, HEIGHT_WARMUP, mask_sh),\
-	SF(WB_WARM_UP_MODE_CTL1, GMC_WARM_UP_ENABLE, mask_sh),\
-	SF(WB_WARM_UP_MODE_CTL2, DATA_VALUE_WARMUP, mask_sh),\
-	SF(WB_WARM_UP_MODE_CTL2, MODE_WARMUP, mask_sh),\
-	SF(WB_WARM_UP_MODE_CTL2, DATA_DEPTH_WARMUP, mask_sh)
+	SF_DWB(WB_ENABLE, WB_ENABLE, mask_sh),\
+	SF_DWB(WB_EC_CONFIG, DISPCLK_R_WB_GATE_DIS, mask_sh),\
+	SF_DWB(WB_EC_CONFIG, DISPCLK_G_WB_GATE_DIS, mask_sh),\
+	SF_DWB(WB_EC_CONFIG, DISPCLK_G_WBSCL_GATE_DIS, mask_sh),\
+	SF_DWB(WB_EC_CONFIG, WB_TEST_CLK_SEL, mask_sh),\
+	SF_DWB(WB_EC_CONFIG, WB_LB_LS_DIS, mask_sh),\
+	SF_DWB(WB_EC_CONFIG, WB_LB_SD_DIS, mask_sh),\
+	SF_DWB(WB_EC_CONFIG, WB_LUT_LS_DIS, mask_sh),\
+	SF_DWB(WB_EC_CONFIG, WBSCL_LB_MEM_PWR_MODE_SEL, mask_sh),\
+	SF_DWB(WB_EC_CONFIG, WBSCL_LB_MEM_PWR_DIS, mask_sh),\
+	SF_DWB(WB_EC_CONFIG, WBSCL_LB_MEM_PWR_FORCE, mask_sh),\
+	SF_DWB(WB_EC_CONFIG, WBSCL_LB_MEM_PWR_STATE, mask_sh),\
+	SF_DWB(WB_EC_CONFIG, WB_RAM_PW_SAVE_MODE, mask_sh),\
+	SF_DWB(WB_EC_CONFIG, WBSCL_LUT_MEM_PWR_STATE, mask_sh),\
+	SF_DWB(CNV_MODE, CNV_OUT_BPC, mask_sh),\
+	SF_DWB(CNV_MODE, CNV_FRAME_CAPTURE_RATE, mask_sh),\
+	SF_DWB(CNV_MODE, CNV_WINDOW_CROP_EN, mask_sh),\
+	SF_DWB(CNV_MODE, CNV_STEREO_TYPE, mask_sh),\
+	SF_DWB(CNV_MODE, CNV_INTERLACED_MODE, mask_sh),\
+	SF_DWB(CNV_MODE, CNV_EYE_SELECTION, mask_sh),\
+	SF_DWB(CNV_MODE, CNV_STEREO_POLARITY, mask_sh),\
+	SF_DWB(CNV_MODE, CNV_INTERLACED_FIELD_ORDER, mask_sh),\
+	SF_DWB(CNV_MODE, CNV_STEREO_SPLIT, mask_sh),\
+	SF_DWB(CNV_MODE, CNV_NEW_CONTENT, mask_sh),\
+	SF_DWB(CNV_MODE, CNV_FRAME_CAPTURE_EN_CURRENT, mask_sh),\
+	SF_DWB(CNV_MODE, CNV_FRAME_CAPTURE_EN, mask_sh),\
+	SF_DWB(CNV_WINDOW_START, CNV_WINDOW_START_X, mask_sh),\
+	SF_DWB(CNV_WINDOW_START, CNV_WINDOW_START_Y, mask_sh),\
+	SF_DWB(CNV_WINDOW_SIZE, CNV_WINDOW_WIDTH, mask_sh),\
+	SF_DWB(CNV_WINDOW_SIZE, CNV_WINDOW_HEIGHT, mask_sh),\
+	SF_DWB(CNV_UPDATE, CNV_UPDATE_PENDING, mask_sh),\
+	SF_DWB(CNV_UPDATE, CNV_UPDATE_TAKEN, mask_sh),\
+	SF_DWB(CNV_UPDATE, CNV_UPDATE_LOCK, mask_sh),\
+	SF_DWB(CNV_SOURCE_SIZE, CNV_SOURCE_WIDTH, mask_sh),\
+	SF_DWB(CNV_SOURCE_SIZE, CNV_SOURCE_HEIGHT, mask_sh),\
+	SF_DWB(CNV_TEST_CNTL, CNV_TEST_CRC_EN, mask_sh),\
+	SF_DWB(CNV_TEST_CNTL, CNV_TEST_CRC_CONT_EN, mask_sh),\
+	SF_DWB(CNV_TEST_CRC_RED, CNV_TEST_CRC_RED_MASK, mask_sh),\
+	SF_DWB(CNV_TEST_CRC_RED, CNV_TEST_CRC_SIG_RED, mask_sh),\
+	SF_DWB(CNV_TEST_CRC_GREEN, CNV_TEST_CRC_GREEN_MASK, mask_sh),\
+	SF_DWB(CNV_TEST_CRC_GREEN, CNV_TEST_CRC_SIG_GREEN, mask_sh),\
+	SF_DWB(CNV_TEST_CRC_BLUE, CNV_TEST_CRC_BLUE_MASK, mask_sh),\
+	SF_DWB(CNV_TEST_CRC_BLUE, CNV_TEST_CRC_SIG_BLUE, mask_sh),\
+	SF_DWB(WB_DEBUG_CTRL, WB_DEBUG_EN, mask_sh),\
+	SF_DWB(WB_DEBUG_CTRL, WB_DEBUG_SEL, mask_sh),\
+	SF_DWB(WB_DBG_MODE, WB_DBG_MODE_EN, mask_sh),\
+	SF_DWB(WB_DBG_MODE, WB_DBG_DIN_FMT, mask_sh),\
+	SF_DWB(WB_DBG_MODE, WB_DBG_36MODE, mask_sh),\
+	SF_DWB(WB_DBG_MODE, WB_DBG_CMAP, mask_sh),\
+	SF_DWB(WB_DBG_MODE, WB_DBG_PXLRATE_ERROR, mask_sh),\
+	SF_DWB(WB_DBG_MODE, WB_DBG_SOURCE_WIDTH, mask_sh),\
+	SF_DWB(WB_HW_DEBUG, WB_HW_DEBUG, mask_sh),\
+	SF_DWB(WB_SOFT_RESET, WB_SOFT_RESET, mask_sh),\
+	SF_DWB(CNV_TEST_DEBUG_INDEX, CNV_TEST_DEBUG_INDEX, mask_sh),\
+	SF_DWB(CNV_TEST_DEBUG_INDEX, CNV_TEST_DEBUG_WRITE_EN, mask_sh),\
+	SF_DWB(CNV_TEST_DEBUG_DATA, CNV_TEST_DEBUG_DATA, mask_sh),\
+	SF_DWB(WBSCL_COEF_RAM_SELECT, WBSCL_COEF_RAM_TAP_PAIR_IDX, mask_sh),\
+	SF_DWB(WBSCL_COEF_RAM_SELECT, WBSCL_COEF_RAM_PHASE, mask_sh),\
+	SF_DWB(WBSCL_COEF_RAM_SELECT, WBSCL_COEF_RAM_FILTER_TYPE, mask_sh),\
+	SF_DWB(WBSCL_COEF_RAM_TAP_DATA, WBSCL_COEF_RAM_EVEN_TAP_COEF, mask_sh),\
+	SF_DWB(WBSCL_COEF_RAM_TAP_DATA, WBSCL_COEF_RAM_EVEN_TAP_COEF_EN, mask_sh),\
+	SF_DWB(WBSCL_COEF_RAM_TAP_DATA, WBSCL_COEF_RAM_ODD_TAP_COEF, mask_sh),\
+	SF_DWB(WBSCL_COEF_RAM_TAP_DATA, WBSCL_COEF_RAM_ODD_TAP_COEF_EN, mask_sh),\
+	SF_DWB(WBSCL_MODE, WBSCL_MODE, mask_sh),\
+	SF_DWB(WBSCL_MODE, WBSCL_OUT_BIT_DEPTH, mask_sh),\
+	SF_DWB(WBSCL_TAP_CONTROL, WBSCL_V_NUM_OF_TAPS_Y_RGB, mask_sh),\
+	SF_DWB(WBSCL_TAP_CONTROL, WBSCL_V_NUM_OF_TAPS_CBCR, mask_sh),\
+	SF_DWB(WBSCL_TAP_CONTROL, WBSCL_H_NUM_OF_TAPS_Y_RGB, mask_sh),\
+	SF_DWB(WBSCL_TAP_CONTROL, WBSCL_H_NUM_OF_TAPS_CBCR, mask_sh),\
+	SF_DWB(WBSCL_DEST_SIZE, WBSCL_DEST_HEIGHT, mask_sh),\
+	SF_DWB(WBSCL_DEST_SIZE, WBSCL_DEST_WIDTH, mask_sh),\
+	SF_DWB(WBSCL_HORZ_FILTER_SCALE_RATIO, WBSCL_H_SCALE_RATIO, mask_sh),\
+	SF_DWB(WBSCL_HORZ_FILTER_INIT_Y_RGB, WBSCL_H_INIT_FRAC_Y_RGB, mask_sh),\
+	SF_DWB(WBSCL_HORZ_FILTER_INIT_Y_RGB, WBSCL_H_INIT_INT_Y_RGB, mask_sh),\
+	SF_DWB(WBSCL_HORZ_FILTER_INIT_CBCR, WBSCL_H_INIT_FRAC_CBCR, mask_sh),\
+	SF_DWB(WBSCL_HORZ_FILTER_INIT_CBCR, WBSCL_H_INIT_INT_CBCR, mask_sh),\
+	SF_DWB(WBSCL_VERT_FILTER_SCALE_RATIO, WBSCL_V_SCALE_RATIO, mask_sh),\
+	SF_DWB(WBSCL_VERT_FILTER_INIT_Y_RGB, WBSCL_V_INIT_FRAC_Y_RGB, mask_sh),\
+	SF_DWB(WBSCL_VERT_FILTER_INIT_Y_RGB, WBSCL_V_INIT_INT_Y_RGB, mask_sh),\
+	SF_DWB(WBSCL_VERT_FILTER_INIT_CBCR, WBSCL_V_INIT_FRAC_CBCR, mask_sh),\
+	SF_DWB(WBSCL_VERT_FILTER_INIT_CBCR, WBSCL_V_INIT_INT_CBCR, mask_sh),\
+	SF_DWB(WBSCL_ROUND_OFFSET, WBSCL_ROUND_OFFSET_Y_RGB, mask_sh),\
+	SF_DWB(WBSCL_ROUND_OFFSET, WBSCL_ROUND_OFFSET_CBCR, mask_sh),\
+	SF_DWB(WBSCL_OVERFLOW_STATUS, WBSCL_DATA_OVERFLOW_FLAG, mask_sh),\
+	SF_DWB(WBSCL_OVERFLOW_STATUS, WBSCL_DATA_OVERFLOW_ACK, mask_sh),\
+	SF_DWB(WBSCL_OVERFLOW_STATUS, WBSCL_DATA_OVERFLOW_MASK, mask_sh),\
+	SF_DWB(WBSCL_OVERFLOW_STATUS, WBSCL_DATA_OVERFLOW_INT_STATUS, mask_sh),\
+	SF_DWB(WBSCL_OVERFLOW_STATUS, WBSCL_DATA_OVERFLOW_INT_TYPE, mask_sh),\
+	SF_DWB(WBSCL_COEF_RAM_CONFLICT_STATUS, WBSCL_HOST_CONFLICT_FLAG, mask_sh),\
+	SF_DWB(WBSCL_COEF_RAM_CONFLICT_STATUS, WBSCL_HOST_CONFLICT_ACK, mask_sh),\
+	SF_DWB(WBSCL_COEF_RAM_CONFLICT_STATUS, WBSCL_HOST_CONFLICT_MASK, mask_sh),\
+	SF_DWB(WBSCL_COEF_RAM_CONFLICT_STATUS, WBSCL_HOST_CONFLICT_INT_STATUS, mask_sh),\
+	SF_DWB(WBSCL_COEF_RAM_CONFLICT_STATUS, WBSCL_HOST_CONFLICT_INT_TYPE, mask_sh),\
+	SF_DWB(WBSCL_TEST_CNTL, WBSCL_TEST_CRC_EN, mask_sh),\
+	SF_DWB(WBSCL_TEST_CNTL, WBSCL_TEST_CRC_CONT_EN, mask_sh),\
+	SF_DWB(WBSCL_TEST_CRC_RED, WBSCL_TEST_CRC_RED_MASK, mask_sh),\
+	SF_DWB(WBSCL_TEST_CRC_RED, WBSCL_TEST_CRC_SIG_RED, mask_sh),\
+	SF_DWB(WBSCL_TEST_CRC_GREEN, WBSCL_TEST_CRC_GREEN_MASK, mask_sh),\
+	SF_DWB(WBSCL_TEST_CRC_GREEN, WBSCL_TEST_CRC_SIG_GREEN, mask_sh),\
+	SF_DWB(WBSCL_TEST_CRC_BLUE, WBSCL_TEST_CRC_BLUE_MASK, mask_sh),\
+	SF_DWB(WBSCL_TEST_CRC_BLUE, WBSCL_TEST_CRC_SIG_BLUE, mask_sh),\
+	SF_DWB(WBSCL_BACKPRESSURE_CNT_EN, WBSCL_BACKPRESSURE_CNT_EN, mask_sh),\
+	SF_DWB(WB_MCIF_BACKPRESSURE_CNT, WB_MCIF_Y_MAX_BACKPRESSURE, mask_sh),\
+	SF_DWB(WB_MCIF_BACKPRESSURE_CNT, WB_MCIF_C_MAX_BACKPRESSURE, mask_sh),\
+	SF_DWB(WBSCL_CLAMP_Y_RGB, WBSCL_CLAMP_UPPER_Y_RGB, mask_sh),\
+	SF_DWB(WBSCL_CLAMP_Y_RGB, WBSCL_CLAMP_LOWER_Y_RGB, mask_sh),\
+	SF_DWB(WBSCL_CLAMP_CBCR, WBSCL_CLAMP_UPPER_CBCR, mask_sh),\
+	SF_DWB(WBSCL_CLAMP_CBCR, WBSCL_CLAMP_LOWER_CBCR, mask_sh),\
+	SF_DWB(WBSCL_OUTSIDE_PIX_STRATEGY, WBSCL_OUTSIDE_PIX_STRATEGY, mask_sh),\
+	SF_DWB(WBSCL_OUTSIDE_PIX_STRATEGY, WBSCL_BLACK_COLOR_G_Y, mask_sh),\
+	SF_DWB(WBSCL_OUTSIDE_PIX_STRATEGY_CBCR, WBSCL_BLACK_COLOR_B_CB, mask_sh),\
+	SF_DWB(WBSCL_OUTSIDE_PIX_STRATEGY_CBCR, WBSCL_BLACK_COLOR_R_CR, mask_sh),\
+	SF_DWB(WBSCL_DEBUG, WBSCL_DEBUG, mask_sh),\
+	SF_DWB(WBSCL_TEST_DEBUG_INDEX, WBSCL_TEST_DEBUG_INDEX, mask_sh),\
+	SF_DWB(WBSCL_TEST_DEBUG_INDEX, WBSCL_TEST_DEBUG_WRITE_EN, mask_sh),\
+	SF_DWB(WBSCL_TEST_DEBUG_DATA, WBSCL_TEST_DEBUG_DATA, mask_sh),\
+	SF_DWB(WB_WARM_UP_MODE_CTL1, WIDTH_WARMUP, mask_sh),\
+	SF_DWB(WB_WARM_UP_MODE_CTL1, HEIGHT_WARMUP, mask_sh),\
+	SF_DWB(WB_WARM_UP_MODE_CTL1, GMC_WARM_UP_ENABLE, mask_sh),\
+	SF_DWB(WB_WARM_UP_MODE_CTL2, DATA_VALUE_WARMUP, mask_sh),\
+	SF_DWB(WB_WARM_UP_MODE_CTL2, MODE_WARMUP, mask_sh),\
+	SF_DWB(WB_WARM_UP_MODE_CTL2, DATA_DEPTH_WARMUP, mask_sh)
 
 #define DWBC_REG_FIELD_LIST_DCN2_0(type) \
 	type WB_ENABLE;\
diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_dwb_scl.c b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_dwb_scl.c
index 880954ac0b02..a0d437f0ce2b 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_dwb_scl.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_dwb_scl.c
@@ -527,7 +527,7 @@ static const uint16_t filter_12tap_16p_183[108] = {
 	0, 84, 16328, 16032, 416, 1944, 1944, 416, 16032, 16328, 84, 0,
 };
 
-const uint16_t *wbscl_get_filter_3tap_16p(struct fixed31_32 ratio)
+static const uint16_t *wbscl_get_filter_3tap_16p(struct fixed31_32 ratio)
 {
 	if (ratio.value < dc_fixpt_one.value)
 		return filter_3tap_16p_upscale;
@@ -539,7 +539,7 @@ const uint16_t *wbscl_get_filter_3tap_16p(struct fixed31_32 ratio)
 		return filter_3tap_16p_183;
 }
 
-const uint16_t *wbscl_get_filter_4tap_16p(struct fixed31_32 ratio)
+static const uint16_t *wbscl_get_filter_4tap_16p(struct fixed31_32 ratio)
 {
 	if (ratio.value < dc_fixpt_one.value)
 		return filter_4tap_16p_upscale;
@@ -690,6 +690,9 @@ static void wbscl_set_scaler_filter(
 	int pair;
 	uint16_t odd_coef, even_coef;
 
+	if (!filter)
+		return;
+
 	for (phase = 0; phase < (NUM_PHASES / 2 + 1); phase++) {
 		for (pair = 0; pair < tap_pairs; pair++) {
 			even_coef = filter[phase * taps + 2 * pair];
diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c
deleted file mode 100644
index e6af99ae3d9f..000000000000
--- a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c
+++ /dev/null
@@ -1,2634 +0,0 @@
-/*
- * Copyright 2016 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: AMD
- *
- */
-#include <linux/delay.h>
-
-#include "dm_services.h"
-#include "basics/dc_common.h"
-#include "dm_helpers.h"
-#include "core_types.h"
-#include "resource.h"
-#include "dcn20_resource.h"
-#include "dcn20_hwseq.h"
-#include "dce/dce_hwseq.h"
-#include "dcn20_dsc.h"
-#include "dcn20_optc.h"
-#include "abm.h"
-#include "clk_mgr.h"
-#include "dmcu.h"
-#include "hubp.h"
-#include "timing_generator.h"
-#include "opp.h"
-#include "ipp.h"
-#include "mpc.h"
-#include "mcif_wb.h"
-#include "dchubbub.h"
-#include "reg_helper.h"
-#include "dcn10/dcn10_cm_common.h"
-#include "dc_link_dp.h"
-#include "vm_helper.h"
-#include "dccg.h"
-#include "dc_dmub_srv.h"
-#include "dce/dmub_hw_lock_mgr.h"
-#include "hw_sequencer.h"
-#include "inc/link_dpcd.h"
-#include "dpcd_defs.h"
-#include "inc/link_enc_cfg.h"
-
-#define DC_LOGGER_INIT(logger)
-
-#define CTX \
-	hws->ctx
-#define REG(reg)\
-	hws->regs->reg
-
-#undef FN
-#define FN(reg_name, field_name) \
-	hws->shifts->field_name, hws->masks->field_name
-
-static int find_free_gsl_group(const struct dc *dc)
-{
-	if (dc->res_pool->gsl_groups.gsl_0 == 0)
-		return 1;
-	if (dc->res_pool->gsl_groups.gsl_1 == 0)
-		return 2;
-	if (dc->res_pool->gsl_groups.gsl_2 == 0)
-		return 3;
-
-	return 0;
-}
-
-/* NOTE: This is not a generic setup_gsl function (hence the suffix as_lock)
- * This is only used to lock pipes in pipe splitting case with immediate flip
- * Ordinary MPC/OTG locks suppress VUPDATE which doesn't help with immediate,
- * so we get tearing with freesync since we cannot flip multiple pipes
- * atomically.
- * We use GSL for this:
- * - immediate flip: find first available GSL group if not already assigned
- *                   program gsl with that group, set current OTG as master
- *                   and always us 0x4 = AND of flip_ready from all pipes
- * - vsync flip: disable GSL if used
- *
- * Groups in stream_res are stored as +1 from HW registers, i.e.
- * gsl_0 <=> pipe_ctx->stream_res.gsl_group == 1
- * Using a magic value like -1 would require tracking all inits/resets
- */
-static void dcn20_setup_gsl_group_as_lock(
-		const struct dc *dc,
-		struct pipe_ctx *pipe_ctx,
-		bool enable)
-{
-	struct gsl_params gsl;
-	int group_idx;
-
-	memset(&gsl, 0, sizeof(struct gsl_params));
-
-	if (enable) {
-		/* return if group already assigned since GSL was set up
-		 * for vsync flip, we would unassign so it can't be "left over"
-		 */
-		if (pipe_ctx->stream_res.gsl_group > 0)
-			return;
-
-		group_idx = find_free_gsl_group(dc);
-		ASSERT(group_idx != 0);
-		pipe_ctx->stream_res.gsl_group = group_idx;
-
-		/* set gsl group reg field and mark resource used */
-		switch (group_idx) {
-		case 1:
-			gsl.gsl0_en = 1;
-			dc->res_pool->gsl_groups.gsl_0 = 1;
-			break;
-		case 2:
-			gsl.gsl1_en = 1;
-			dc->res_pool->gsl_groups.gsl_1 = 1;
-			break;
-		case 3:
-			gsl.gsl2_en = 1;
-			dc->res_pool->gsl_groups.gsl_2 = 1;
-			break;
-		default:
-			BREAK_TO_DEBUGGER();
-			return; // invalid case
-		}
-		gsl.gsl_master_en = 1;
-	} else {
-		group_idx = pipe_ctx->stream_res.gsl_group;
-		if (group_idx == 0)
-			return; // if not in use, just return
-
-		pipe_ctx->stream_res.gsl_group = 0;
-
-		/* unset gsl group reg field and mark resource free */
-		switch (group_idx) {
-		case 1:
-			gsl.gsl0_en = 0;
-			dc->res_pool->gsl_groups.gsl_0 = 0;
-			break;
-		case 2:
-			gsl.gsl1_en = 0;
-			dc->res_pool->gsl_groups.gsl_1 = 0;
-			break;
-		case 3:
-			gsl.gsl2_en = 0;
-			dc->res_pool->gsl_groups.gsl_2 = 0;
-			break;
-		default:
-			BREAK_TO_DEBUGGER();
-			return;
-		}
-		gsl.gsl_master_en = 0;
-	}
-
-	/* at this point we want to program whether it's to enable or disable */
-	if (pipe_ctx->stream_res.tg->funcs->set_gsl != NULL &&
-		pipe_ctx->stream_res.tg->funcs->set_gsl_source_select != NULL) {
-		pipe_ctx->stream_res.tg->funcs->set_gsl(
-			pipe_ctx->stream_res.tg,
-			&gsl);
-
-		pipe_ctx->stream_res.tg->funcs->set_gsl_source_select(
-			pipe_ctx->stream_res.tg, group_idx,	enable ? 4 : 0);
-	} else
-		BREAK_TO_DEBUGGER();
-}
-
-void dcn20_set_flip_control_gsl(
-		struct pipe_ctx *pipe_ctx,
-		bool flip_immediate)
-{
-	if (pipe_ctx && pipe_ctx->plane_res.hubp->funcs->hubp_set_flip_control_surface_gsl)
-		pipe_ctx->plane_res.hubp->funcs->hubp_set_flip_control_surface_gsl(
-				pipe_ctx->plane_res.hubp, flip_immediate);
-
-}
-
-void dcn20_enable_power_gating_plane(
-	struct dce_hwseq *hws,
-	bool enable)
-{
-	bool force_on = true; /* disable power gating */
-
-	if (enable)
-		force_on = false;
-
-	/* DCHUBP0/1/2/3/4/5 */
-	REG_UPDATE(DOMAIN0_PG_CONFIG, DOMAIN0_POWER_FORCEON, force_on);
-	REG_UPDATE(DOMAIN2_PG_CONFIG, DOMAIN2_POWER_FORCEON, force_on);
-	REG_UPDATE(DOMAIN4_PG_CONFIG, DOMAIN4_POWER_FORCEON, force_on);
-	REG_UPDATE(DOMAIN6_PG_CONFIG, DOMAIN6_POWER_FORCEON, force_on);
-	if (REG(DOMAIN8_PG_CONFIG))
-		REG_UPDATE(DOMAIN8_PG_CONFIG, DOMAIN8_POWER_FORCEON, force_on);
-	if (REG(DOMAIN10_PG_CONFIG))
-		REG_UPDATE(DOMAIN10_PG_CONFIG, DOMAIN8_POWER_FORCEON, force_on);
-
-	/* DPP0/1/2/3/4/5 */
-	REG_UPDATE(DOMAIN1_PG_CONFIG, DOMAIN1_POWER_FORCEON, force_on);
-	REG_UPDATE(DOMAIN3_PG_CONFIG, DOMAIN3_POWER_FORCEON, force_on);
-	REG_UPDATE(DOMAIN5_PG_CONFIG, DOMAIN5_POWER_FORCEON, force_on);
-	REG_UPDATE(DOMAIN7_PG_CONFIG, DOMAIN7_POWER_FORCEON, force_on);
-	if (REG(DOMAIN9_PG_CONFIG))
-		REG_UPDATE(DOMAIN9_PG_CONFIG, DOMAIN9_POWER_FORCEON, force_on);
-	if (REG(DOMAIN11_PG_CONFIG))
-		REG_UPDATE(DOMAIN11_PG_CONFIG, DOMAIN9_POWER_FORCEON, force_on);
-
-	/* DCS0/1/2/3/4/5 */
-	REG_UPDATE(DOMAIN16_PG_CONFIG, DOMAIN16_POWER_FORCEON, force_on);
-	REG_UPDATE(DOMAIN17_PG_CONFIG, DOMAIN17_POWER_FORCEON, force_on);
-	REG_UPDATE(DOMAIN18_PG_CONFIG, DOMAIN18_POWER_FORCEON, force_on);
-	if (REG(DOMAIN19_PG_CONFIG))
-		REG_UPDATE(DOMAIN19_PG_CONFIG, DOMAIN19_POWER_FORCEON, force_on);
-	if (REG(DOMAIN20_PG_CONFIG))
-		REG_UPDATE(DOMAIN20_PG_CONFIG, DOMAIN20_POWER_FORCEON, force_on);
-	if (REG(DOMAIN21_PG_CONFIG))
-		REG_UPDATE(DOMAIN21_PG_CONFIG, DOMAIN21_POWER_FORCEON, force_on);
-}
-
-void dcn20_dccg_init(struct dce_hwseq *hws)
-{
-	/*
-	 * set MICROSECOND_TIME_BASE_DIV
-	 * 100Mhz refclk -> 0x120264
-	 * 27Mhz refclk -> 0x12021b
-	 * 48Mhz refclk -> 0x120230
-	 *
-	 */
-	REG_WRITE(MICROSECOND_TIME_BASE_DIV, 0x120264);
-
-	/*
-	 * set MILLISECOND_TIME_BASE_DIV
-	 * 100Mhz refclk -> 0x1186a0
-	 * 27Mhz refclk -> 0x106978
-	 * 48Mhz refclk -> 0x10bb80
-	 *
-	 */
-	REG_WRITE(MILLISECOND_TIME_BASE_DIV, 0x1186a0);
-
-	/* This value is dependent on the hardware pipeline delay so set once per SOC */
-	REG_WRITE(DISPCLK_FREQ_CHANGE_CNTL, 0xe01003c);
-}
-
-void dcn20_disable_vga(
-	struct dce_hwseq *hws)
-{
-	REG_WRITE(D1VGA_CONTROL, 0);
-	REG_WRITE(D2VGA_CONTROL, 0);
-	REG_WRITE(D3VGA_CONTROL, 0);
-	REG_WRITE(D4VGA_CONTROL, 0);
-	REG_WRITE(D5VGA_CONTROL, 0);
-	REG_WRITE(D6VGA_CONTROL, 0);
-}
-
-void dcn20_program_triple_buffer(
-	const struct dc *dc,
-	struct pipe_ctx *pipe_ctx,
-	bool enable_triple_buffer)
-{
-	if (pipe_ctx->plane_res.hubp && pipe_ctx->plane_res.hubp->funcs) {
-		pipe_ctx->plane_res.hubp->funcs->hubp_enable_tripleBuffer(
-			pipe_ctx->plane_res.hubp,
-			enable_triple_buffer);
-	}
-}
-
-/* Blank pixel data during initialization */
-void dcn20_init_blank(
-		struct dc *dc,
-		struct timing_generator *tg)
-{
-	struct dce_hwseq *hws = dc->hwseq;
-	enum dc_color_space color_space;
-	struct tg_color black_color = {0};
-	struct output_pixel_processor *opp = NULL;
-	struct output_pixel_processor *bottom_opp = NULL;
-	uint32_t num_opps, opp_id_src0, opp_id_src1;
-	uint32_t otg_active_width, otg_active_height;
-
-	/* program opp dpg blank color */
-	color_space = COLOR_SPACE_SRGB;
-	color_space_to_black_color(dc, color_space, &black_color);
-
-	/* get the OTG active size */
-	tg->funcs->get_otg_active_size(tg,
-			&otg_active_width,
-			&otg_active_height);
-
-	/* get the OPTC source */
-	tg->funcs->get_optc_source(tg, &num_opps, &opp_id_src0, &opp_id_src1);
-
-	if (opp_id_src0 >= dc->res_pool->res_cap->num_opp) {
-		ASSERT(false);
-		return;
-	}
-	opp = dc->res_pool->opps[opp_id_src0];
-
-	if (num_opps == 2) {
-		otg_active_width = otg_active_width / 2;
-
-		if (opp_id_src1 >= dc->res_pool->res_cap->num_opp) {
-			ASSERT(false);
-			return;
-		}
-		bottom_opp = dc->res_pool->opps[opp_id_src1];
-	}
-
-	opp->funcs->opp_set_disp_pattern_generator(
-			opp,
-			CONTROLLER_DP_TEST_PATTERN_SOLID_COLOR,
-			CONTROLLER_DP_COLOR_SPACE_UDEFINED,
-			COLOR_DEPTH_UNDEFINED,
-			&black_color,
-			otg_active_width,
-			otg_active_height,
-			0);
-
-	if (num_opps == 2) {
-		bottom_opp->funcs->opp_set_disp_pattern_generator(
-				bottom_opp,
-				CONTROLLER_DP_TEST_PATTERN_SOLID_COLOR,
-				CONTROLLER_DP_COLOR_SPACE_UDEFINED,
-				COLOR_DEPTH_UNDEFINED,
-				&black_color,
-				otg_active_width,
-				otg_active_height,
-				0);
-	}
-
-	hws->funcs.wait_for_blank_complete(opp);
-}
-
-void dcn20_dsc_pg_control(
-		struct dce_hwseq *hws,
-		unsigned int dsc_inst,
-		bool power_on)
-{
-	uint32_t power_gate = power_on ? 0 : 1;
-	uint32_t pwr_status = power_on ? 0 : 2;
-	uint32_t org_ip_request_cntl = 0;
-
-	if (hws->ctx->dc->debug.disable_dsc_power_gate)
-		return;
-
-	if (REG(DOMAIN16_PG_CONFIG) == 0)
-		return;
-
-	REG_GET(DC_IP_REQUEST_CNTL, IP_REQUEST_EN, &org_ip_request_cntl);
-	if (org_ip_request_cntl == 0)
-		REG_SET(DC_IP_REQUEST_CNTL, 0, IP_REQUEST_EN, 1);
-
-	switch (dsc_inst) {
-	case 0: /* DSC0 */
-		REG_UPDATE(DOMAIN16_PG_CONFIG,
-				DOMAIN16_POWER_GATE, power_gate);
-
-		REG_WAIT(DOMAIN16_PG_STATUS,
-				DOMAIN16_PGFSM_PWR_STATUS, pwr_status,
-				1, 1000);
-		break;
-	case 1: /* DSC1 */
-		REG_UPDATE(DOMAIN17_PG_CONFIG,
-				DOMAIN17_POWER_GATE, power_gate);
-
-		REG_WAIT(DOMAIN17_PG_STATUS,
-				DOMAIN17_PGFSM_PWR_STATUS, pwr_status,
-				1, 1000);
-		break;
-	case 2: /* DSC2 */
-		REG_UPDATE(DOMAIN18_PG_CONFIG,
-				DOMAIN18_POWER_GATE, power_gate);
-
-		REG_WAIT(DOMAIN18_PG_STATUS,
-				DOMAIN18_PGFSM_PWR_STATUS, pwr_status,
-				1, 1000);
-		break;
-	case 3: /* DSC3 */
-		REG_UPDATE(DOMAIN19_PG_CONFIG,
-				DOMAIN19_POWER_GATE, power_gate);
-
-		REG_WAIT(DOMAIN19_PG_STATUS,
-				DOMAIN19_PGFSM_PWR_STATUS, pwr_status,
-				1, 1000);
-		break;
-	case 4: /* DSC4 */
-		REG_UPDATE(DOMAIN20_PG_CONFIG,
-				DOMAIN20_POWER_GATE, power_gate);
-
-		REG_WAIT(DOMAIN20_PG_STATUS,
-				DOMAIN20_PGFSM_PWR_STATUS, pwr_status,
-				1, 1000);
-		break;
-	case 5: /* DSC5 */
-		REG_UPDATE(DOMAIN21_PG_CONFIG,
-				DOMAIN21_POWER_GATE, power_gate);
-
-		REG_WAIT(DOMAIN21_PG_STATUS,
-				DOMAIN21_PGFSM_PWR_STATUS, pwr_status,
-				1, 1000);
-		break;
-	default:
-		BREAK_TO_DEBUGGER();
-		break;
-	}
-
-	if (org_ip_request_cntl == 0)
-		REG_SET(DC_IP_REQUEST_CNTL, 0, IP_REQUEST_EN, 0);
-}
-
-void dcn20_dpp_pg_control(
-		struct dce_hwseq *hws,
-		unsigned int dpp_inst,
-		bool power_on)
-{
-	uint32_t power_gate = power_on ? 0 : 1;
-	uint32_t pwr_status = power_on ? 0 : 2;
-
-	if (hws->ctx->dc->debug.disable_dpp_power_gate)
-		return;
-	if (REG(DOMAIN1_PG_CONFIG) == 0)
-		return;
-
-	switch (dpp_inst) {
-	case 0: /* DPP0 */
-		REG_UPDATE(DOMAIN1_PG_CONFIG,
-				DOMAIN1_POWER_GATE, power_gate);
-
-		REG_WAIT(DOMAIN1_PG_STATUS,
-				DOMAIN1_PGFSM_PWR_STATUS, pwr_status,
-				1, 1000);
-		break;
-	case 1: /* DPP1 */
-		REG_UPDATE(DOMAIN3_PG_CONFIG,
-				DOMAIN3_POWER_GATE, power_gate);
-
-		REG_WAIT(DOMAIN3_PG_STATUS,
-				DOMAIN3_PGFSM_PWR_STATUS, pwr_status,
-				1, 1000);
-		break;
-	case 2: /* DPP2 */
-		REG_UPDATE(DOMAIN5_PG_CONFIG,
-				DOMAIN5_POWER_GATE, power_gate);
-
-		REG_WAIT(DOMAIN5_PG_STATUS,
-				DOMAIN5_PGFSM_PWR_STATUS, pwr_status,
-				1, 1000);
-		break;
-	case 3: /* DPP3 */
-		REG_UPDATE(DOMAIN7_PG_CONFIG,
-				DOMAIN7_POWER_GATE, power_gate);
-
-		REG_WAIT(DOMAIN7_PG_STATUS,
-				DOMAIN7_PGFSM_PWR_STATUS, pwr_status,
-				1, 1000);
-		break;
-	case 4: /* DPP4 */
-		REG_UPDATE(DOMAIN9_PG_CONFIG,
-				DOMAIN9_POWER_GATE, power_gate);
-
-		REG_WAIT(DOMAIN9_PG_STATUS,
-				DOMAIN9_PGFSM_PWR_STATUS, pwr_status,
-				1, 1000);
-		break;
-	case 5: /* DPP5 */
-		/*
-		 * Do not power gate DPP5, should be left at HW default, power on permanently.
-		 * PG on Pipe5 is De-featured, attempting to put it to PG state may result in hard
-		 * reset.
-		 * REG_UPDATE(DOMAIN11_PG_CONFIG,
-		 *		DOMAIN11_POWER_GATE, power_gate);
-		 *
-		 * REG_WAIT(DOMAIN11_PG_STATUS,
-		 *		DOMAIN11_PGFSM_PWR_STATUS, pwr_status,
-		 * 		1, 1000);
-		 */
-		break;
-	default:
-		BREAK_TO_DEBUGGER();
-		break;
-	}
-}
-
-
-void dcn20_hubp_pg_control(
-		struct dce_hwseq *hws,
-		unsigned int hubp_inst,
-		bool power_on)
-{
-	uint32_t power_gate = power_on ? 0 : 1;
-	uint32_t pwr_status = power_on ? 0 : 2;
-
-	if (hws->ctx->dc->debug.disable_hubp_power_gate)
-		return;
-	if (REG(DOMAIN0_PG_CONFIG) == 0)
-		return;
-
-	switch (hubp_inst) {
-	case 0: /* DCHUBP0 */
-		REG_UPDATE(DOMAIN0_PG_CONFIG,
-				DOMAIN0_POWER_GATE, power_gate);
-
-		REG_WAIT(DOMAIN0_PG_STATUS,
-				DOMAIN0_PGFSM_PWR_STATUS, pwr_status,
-				1, 1000);
-		break;
-	case 1: /* DCHUBP1 */
-		REG_UPDATE(DOMAIN2_PG_CONFIG,
-				DOMAIN2_POWER_GATE, power_gate);
-
-		REG_WAIT(DOMAIN2_PG_STATUS,
-				DOMAIN2_PGFSM_PWR_STATUS, pwr_status,
-				1, 1000);
-		break;
-	case 2: /* DCHUBP2 */
-		REG_UPDATE(DOMAIN4_PG_CONFIG,
-				DOMAIN4_POWER_GATE, power_gate);
-
-		REG_WAIT(DOMAIN4_PG_STATUS,
-				DOMAIN4_PGFSM_PWR_STATUS, pwr_status,
-				1, 1000);
-		break;
-	case 3: /* DCHUBP3 */
-		REG_UPDATE(DOMAIN6_PG_CONFIG,
-				DOMAIN6_POWER_GATE, power_gate);
-
-		REG_WAIT(DOMAIN6_PG_STATUS,
-				DOMAIN6_PGFSM_PWR_STATUS, pwr_status,
-				1, 1000);
-		break;
-	case 4: /* DCHUBP4 */
-		REG_UPDATE(DOMAIN8_PG_CONFIG,
-				DOMAIN8_POWER_GATE, power_gate);
-
-		REG_WAIT(DOMAIN8_PG_STATUS,
-				DOMAIN8_PGFSM_PWR_STATUS, pwr_status,
-				1, 1000);
-		break;
-	case 5: /* DCHUBP5 */
-		/*
-		 * Do not power gate DCHUB5, should be left at HW default, power on permanently.
-		 * PG on Pipe5 is De-featured, attempting to put it to PG state may result in hard
-		 * reset.
-		 * REG_UPDATE(DOMAIN10_PG_CONFIG,
-		 *		DOMAIN10_POWER_GATE, power_gate);
-		 *
-		 * REG_WAIT(DOMAIN10_PG_STATUS,
-		 *		DOMAIN10_PGFSM_PWR_STATUS, pwr_status,
-		 *		1, 1000);
-		 */
-		break;
-	default:
-		BREAK_TO_DEBUGGER();
-		break;
-	}
-}
-
-
-/* disable HW used by plane.
- * note:  cannot disable until disconnect is complete
- */
-void dcn20_plane_atomic_disable(struct dc *dc, struct pipe_ctx *pipe_ctx)
-{
-	struct dce_hwseq *hws = dc->hwseq;
-	struct hubp *hubp = pipe_ctx->plane_res.hubp;
-	struct dpp *dpp = pipe_ctx->plane_res.dpp;
-
-	dc->hwss.wait_for_mpcc_disconnect(dc, dc->res_pool, pipe_ctx);
-
-	/* In flip immediate with pipe splitting case GSL is used for
-	 * synchronization so we must disable it when the plane is disabled.
-	 */
-	if (pipe_ctx->stream_res.gsl_group != 0)
-		dcn20_setup_gsl_group_as_lock(dc, pipe_ctx, false);
-
-	dc->hwss.set_flip_control_gsl(pipe_ctx, false);
-
-	hubp->funcs->hubp_clk_cntl(hubp, false);
-
-	dpp->funcs->dpp_dppclk_control(dpp, false, false);
-
-	hubp->power_gated = true;
-
-	hws->funcs.plane_atomic_power_down(dc,
-			pipe_ctx->plane_res.dpp,
-			pipe_ctx->plane_res.hubp);
-
-	pipe_ctx->stream = NULL;
-	memset(&pipe_ctx->stream_res, 0, sizeof(pipe_ctx->stream_res));
-	memset(&pipe_ctx->plane_res, 0, sizeof(pipe_ctx->plane_res));
-	pipe_ctx->top_pipe = NULL;
-	pipe_ctx->bottom_pipe = NULL;
-	pipe_ctx->plane_state = NULL;
-}
-
-
-void dcn20_disable_plane(struct dc *dc, struct pipe_ctx *pipe_ctx)
-{
-	DC_LOGGER_INIT(dc->ctx->logger);
-
-	if (!pipe_ctx->plane_res.hubp || pipe_ctx->plane_res.hubp->power_gated)
-		return;
-
-	dcn20_plane_atomic_disable(dc, pipe_ctx);
-
-	DC_LOG_DC("Power down front end %d\n",
-					pipe_ctx->pipe_idx);
-}
-
-static int calc_mpc_flow_ctrl_cnt(const struct dc_stream_state *stream,
-		int opp_cnt)
-{
-	bool hblank_halved = optc2_is_two_pixels_per_containter(&stream->timing);
-	int flow_ctrl_cnt;
-
-	if (opp_cnt >= 2)
-		hblank_halved = true;
-
-	flow_ctrl_cnt = stream->timing.h_total - stream->timing.h_addressable -
-			stream->timing.h_border_left -
-			stream->timing.h_border_right;
-
-	if (hblank_halved)
-		flow_ctrl_cnt /= 2;
-
-	/* ODM combine 4:1 case */
-	if (opp_cnt == 4)
-		flow_ctrl_cnt /= 2;
-
-	return flow_ctrl_cnt;
-}
-
-enum dc_status dcn20_enable_stream_timing(
-		struct pipe_ctx *pipe_ctx,
-		struct dc_state *context,
-		struct dc *dc)
-{
-	struct dce_hwseq *hws = dc->hwseq;
-	struct dc_stream_state *stream = pipe_ctx->stream;
-	struct drr_params params = {0};
-	unsigned int event_triggers = 0;
-	struct pipe_ctx *odm_pipe;
-	int opp_cnt = 1;
-	int opp_inst[MAX_PIPES] = { pipe_ctx->stream_res.opp->inst };
-	bool interlace = stream->timing.flags.INTERLACE;
-	int i;
-	struct mpc_dwb_flow_control flow_control;
-	struct mpc *mpc = dc->res_pool->mpc;
-	bool rate_control_2x_pclk = (interlace || optc2_is_two_pixels_per_containter(&stream->timing));
-
-	/* by upper caller loop, pipe0 is parent pipe and be called first.
-	 * back end is set up by for pipe0. Other children pipe share back end
-	 * with pipe 0. No program is needed.
-	 */
-	if (pipe_ctx->top_pipe != NULL)
-		return DC_OK;
-
-	/* TODO check if timing_changed, disable stream if timing changed */
-
-	for (odm_pipe = pipe_ctx->next_odm_pipe; odm_pipe; odm_pipe = odm_pipe->next_odm_pipe) {
-		opp_inst[opp_cnt] = odm_pipe->stream_res.opp->inst;
-		opp_cnt++;
-	}
-
-	if (opp_cnt > 1)
-		pipe_ctx->stream_res.tg->funcs->set_odm_combine(
-				pipe_ctx->stream_res.tg,
-				opp_inst, opp_cnt,
-				&pipe_ctx->stream->timing);
-
-	/* HW program guide assume display already disable
-	 * by unplug sequence. OTG assume stop.
-	 */
-	pipe_ctx->stream_res.tg->funcs->enable_optc_clock(pipe_ctx->stream_res.tg, true);
-
-	if (false == pipe_ctx->clock_source->funcs->program_pix_clk(
-			pipe_ctx->clock_source,
-			&pipe_ctx->stream_res.pix_clk_params,
-			&pipe_ctx->pll_settings)) {
-		BREAK_TO_DEBUGGER();
-		return DC_ERROR_UNEXPECTED;
-	}
-
-	if (dc->hwseq->funcs.PLAT_58856_wa && (!dc_is_dp_signal(stream->signal)))
-		dc->hwseq->funcs.PLAT_58856_wa(context, pipe_ctx);
-
-	pipe_ctx->stream_res.tg->funcs->program_timing(
-			pipe_ctx->stream_res.tg,
-			&stream->timing,
-			pipe_ctx->pipe_dlg_param.vready_offset,
-			pipe_ctx->pipe_dlg_param.vstartup_start,
-			pipe_ctx->pipe_dlg_param.vupdate_offset,
-			pipe_ctx->pipe_dlg_param.vupdate_width,
-			pipe_ctx->stream->signal,
-			true);
-
-	rate_control_2x_pclk = rate_control_2x_pclk || opp_cnt > 1;
-	flow_control.flow_ctrl_mode = 0;
-	flow_control.flow_ctrl_cnt0 = 0x80;
-	flow_control.flow_ctrl_cnt1 = calc_mpc_flow_ctrl_cnt(stream, opp_cnt);
-	if (mpc->funcs->set_out_rate_control) {
-		for (i = 0; i < opp_cnt; ++i) {
-			mpc->funcs->set_out_rate_control(
-					mpc, opp_inst[i],
-					true,
-					rate_control_2x_pclk,
-					&flow_control);
-		}
-	}
-
-	for (odm_pipe = pipe_ctx->next_odm_pipe; odm_pipe; odm_pipe = odm_pipe->next_odm_pipe)
-		odm_pipe->stream_res.opp->funcs->opp_pipe_clock_control(
-				odm_pipe->stream_res.opp,
-				true);
-
-	pipe_ctx->stream_res.opp->funcs->opp_pipe_clock_control(
-			pipe_ctx->stream_res.opp,
-			true);
-
-	hws->funcs.blank_pixel_data(dc, pipe_ctx, true);
-
-	/* VTG is  within DCHUB command block. DCFCLK is always on */
-	if (false == pipe_ctx->stream_res.tg->funcs->enable_crtc(pipe_ctx->stream_res.tg)) {
-		BREAK_TO_DEBUGGER();
-		return DC_ERROR_UNEXPECTED;
-	}
-
-	hws->funcs.wait_for_blank_complete(pipe_ctx->stream_res.opp);
-
-	params.vertical_total_min = stream->adjust.v_total_min;
-	params.vertical_total_max = stream->adjust.v_total_max;
-	params.vertical_total_mid = stream->adjust.v_total_mid;
-	params.vertical_total_mid_frame_num = stream->adjust.v_total_mid_frame_num;
-	if (pipe_ctx->stream_res.tg->funcs->set_drr)
-		pipe_ctx->stream_res.tg->funcs->set_drr(
-			pipe_ctx->stream_res.tg, &params);
-
-	// DRR should set trigger event to monitor surface update event
-	if (stream->adjust.v_total_min != 0 && stream->adjust.v_total_max != 0)
-		event_triggers = 0x80;
-	/* Event triggers and num frames initialized for DRR, but can be
-	 * later updated for PSR use. Note DRR trigger events are generated
-	 * regardless of whether num frames met.
-	 */
-	if (pipe_ctx->stream_res.tg->funcs->set_static_screen_control)
-		pipe_ctx->stream_res.tg->funcs->set_static_screen_control(
-				pipe_ctx->stream_res.tg, event_triggers, 2);
-
-	/* TODO program crtc source select for non-virtual signal*/
-	/* TODO program FMT */
-	/* TODO setup link_enc */
-	/* TODO set stream attributes */
-	/* TODO program audio */
-	/* TODO enable stream if timing changed */
-	/* TODO unblank stream if DP */
-
-	return DC_OK;
-}
-
-void dcn20_program_output_csc(struct dc *dc,
-		struct pipe_ctx *pipe_ctx,
-		enum dc_color_space colorspace,
-		uint16_t *matrix,
-		int opp_id)
-{
-	struct mpc *mpc = dc->res_pool->mpc;
-	enum mpc_output_csc_mode ocsc_mode = MPC_OUTPUT_CSC_COEF_A;
-	int mpcc_id = pipe_ctx->plane_res.hubp->inst;
-
-	if (mpc->funcs->power_on_mpc_mem_pwr)
-		mpc->funcs->power_on_mpc_mem_pwr(mpc, mpcc_id, true);
-
-	if (pipe_ctx->stream->csc_color_matrix.enable_adjustment == true) {
-		if (mpc->funcs->set_output_csc != NULL)
-			mpc->funcs->set_output_csc(mpc,
-					opp_id,
-					matrix,
-					ocsc_mode);
-	} else {
-		if (mpc->funcs->set_ocsc_default != NULL)
-			mpc->funcs->set_ocsc_default(mpc,
-					opp_id,
-					colorspace,
-					ocsc_mode);
-	}
-}
-
-bool dcn20_set_output_transfer_func(struct dc *dc, struct pipe_ctx *pipe_ctx,
-				const struct dc_stream_state *stream)
-{
-	int mpcc_id = pipe_ctx->plane_res.hubp->inst;
-	struct mpc *mpc = pipe_ctx->stream_res.opp->ctx->dc->res_pool->mpc;
-	struct pwl_params *params = NULL;
-	/*
-	 * program OGAM only for the top pipe
-	 * if there is a pipe split then fix diagnostic is required:
-	 * how to pass OGAM parameter for stream.
-	 * if programming for all pipes is required then remove condition
-	 * pipe_ctx->top_pipe == NULL ,but then fix the diagnostic.
-	 */
-	if (mpc->funcs->power_on_mpc_mem_pwr)
-		mpc->funcs->power_on_mpc_mem_pwr(mpc, mpcc_id, true);
-	if (pipe_ctx->top_pipe == NULL
-			&& mpc->funcs->set_output_gamma && stream->out_transfer_func) {
-		if (stream->out_transfer_func->type == TF_TYPE_HWPWL)
-			params = &stream->out_transfer_func->pwl;
-		else if (pipe_ctx->stream->out_transfer_func->type ==
-			TF_TYPE_DISTRIBUTED_POINTS &&
-			cm_helper_translate_curve_to_hw_format(
-			stream->out_transfer_func,
-			&mpc->blender_params, false))
-			params = &mpc->blender_params;
-		/*
-		 * there is no ROM
-		 */
-		if (stream->out_transfer_func->type == TF_TYPE_PREDEFINED)
-			BREAK_TO_DEBUGGER();
-	}
-	/*
-	 * if above if is not executed then 'params' equal to 0 and set in bypass
-	 */
-	mpc->funcs->set_output_gamma(mpc, mpcc_id, params);
-
-	return true;
-}
-
-bool dcn20_set_blend_lut(
-	struct pipe_ctx *pipe_ctx, const struct dc_plane_state *plane_state)
-{
-	struct dpp *dpp_base = pipe_ctx->plane_res.dpp;
-	bool result = true;
-	struct pwl_params *blend_lut = NULL;
-
-	if (plane_state->blend_tf) {
-		if (plane_state->blend_tf->type == TF_TYPE_HWPWL)
-			blend_lut = &plane_state->blend_tf->pwl;
-		else if (plane_state->blend_tf->type == TF_TYPE_DISTRIBUTED_POINTS) {
-			cm_helper_translate_curve_to_hw_format(
-					plane_state->blend_tf,
-					&dpp_base->regamma_params, false);
-			blend_lut = &dpp_base->regamma_params;
-		}
-	}
-	result = dpp_base->funcs->dpp_program_blnd_lut(dpp_base, blend_lut);
-
-	return result;
-}
-
-bool dcn20_set_shaper_3dlut(
-	struct pipe_ctx *pipe_ctx, const struct dc_plane_state *plane_state)
-{
-	struct dpp *dpp_base = pipe_ctx->plane_res.dpp;
-	bool result = true;
-	struct pwl_params *shaper_lut = NULL;
-
-	if (plane_state->in_shaper_func) {
-		if (plane_state->in_shaper_func->type == TF_TYPE_HWPWL)
-			shaper_lut = &plane_state->in_shaper_func->pwl;
-		else if (plane_state->in_shaper_func->type == TF_TYPE_DISTRIBUTED_POINTS) {
-			cm_helper_translate_curve_to_hw_format(
-					plane_state->in_shaper_func,
-					&dpp_base->shaper_params, true);
-			shaper_lut = &dpp_base->shaper_params;
-		}
-	}
-
-	result = dpp_base->funcs->dpp_program_shaper_lut(dpp_base, shaper_lut);
-	if (plane_state->lut3d_func &&
-		plane_state->lut3d_func->state.bits.initialized == 1)
-		result = dpp_base->funcs->dpp_program_3dlut(dpp_base,
-								&plane_state->lut3d_func->lut_3d);
-	else
-		result = dpp_base->funcs->dpp_program_3dlut(dpp_base, NULL);
-
-	return result;
-}
-
-bool dcn20_set_input_transfer_func(struct dc *dc,
-				struct pipe_ctx *pipe_ctx,
-				const struct dc_plane_state *plane_state)
-{
-	struct dce_hwseq *hws = dc->hwseq;
-	struct dpp *dpp_base = pipe_ctx->plane_res.dpp;
-	const struct dc_transfer_func *tf = NULL;
-	bool result = true;
-	bool use_degamma_ram = false;
-
-	if (dpp_base == NULL || plane_state == NULL)
-		return false;
-
-	hws->funcs.set_shaper_3dlut(pipe_ctx, plane_state);
-	hws->funcs.set_blend_lut(pipe_ctx, plane_state);
-
-	if (plane_state->in_transfer_func)
-		tf = plane_state->in_transfer_func;
-
-
-	if (tf == NULL) {
-		dpp_base->funcs->dpp_set_degamma(dpp_base,
-				IPP_DEGAMMA_MODE_BYPASS);
-		return true;
-	}
-
-	if (tf->type == TF_TYPE_HWPWL || tf->type == TF_TYPE_DISTRIBUTED_POINTS)
-		use_degamma_ram = true;
-
-	if (use_degamma_ram == true) {
-		if (tf->type == TF_TYPE_HWPWL)
-			dpp_base->funcs->dpp_program_degamma_pwl(dpp_base,
-					&tf->pwl);
-		else if (tf->type == TF_TYPE_DISTRIBUTED_POINTS) {
-			cm_helper_translate_curve_to_degamma_hw_format(tf,
-					&dpp_base->degamma_params);
-			dpp_base->funcs->dpp_program_degamma_pwl(dpp_base,
-				&dpp_base->degamma_params);
-		}
-		return true;
-	}
-	/* handle here the optimized cases when de-gamma ROM could be used.
-	 *
-	 */
-	if (tf->type == TF_TYPE_PREDEFINED) {
-		switch (tf->tf) {
-		case TRANSFER_FUNCTION_SRGB:
-			dpp_base->funcs->dpp_set_degamma(dpp_base,
-					IPP_DEGAMMA_MODE_HW_sRGB);
-			break;
-		case TRANSFER_FUNCTION_BT709:
-			dpp_base->funcs->dpp_set_degamma(dpp_base,
-					IPP_DEGAMMA_MODE_HW_xvYCC);
-			break;
-		case TRANSFER_FUNCTION_LINEAR:
-			dpp_base->funcs->dpp_set_degamma(dpp_base,
-					IPP_DEGAMMA_MODE_BYPASS);
-			break;
-		case TRANSFER_FUNCTION_PQ:
-			dpp_base->funcs->dpp_set_degamma(dpp_base, IPP_DEGAMMA_MODE_USER_PWL);
-			cm_helper_translate_curve_to_degamma_hw_format(tf, &dpp_base->degamma_params);
-			dpp_base->funcs->dpp_program_degamma_pwl(dpp_base, &dpp_base->degamma_params);
-			result = true;
-			break;
-		default:
-			result = false;
-			break;
-		}
-	} else if (tf->type == TF_TYPE_BYPASS)
-		dpp_base->funcs->dpp_set_degamma(dpp_base,
-				IPP_DEGAMMA_MODE_BYPASS);
-	else {
-		/*
-		 * if we are here, we did not handle correctly.
-		 * fix is required for this use case
-		 */
-		BREAK_TO_DEBUGGER();
-		dpp_base->funcs->dpp_set_degamma(dpp_base,
-				IPP_DEGAMMA_MODE_BYPASS);
-	}
-
-	return result;
-}
-
-void dcn20_update_odm(struct dc *dc, struct dc_state *context, struct pipe_ctx *pipe_ctx)
-{
-	struct pipe_ctx *odm_pipe;
-	int opp_cnt = 1;
-	int opp_inst[MAX_PIPES] = { pipe_ctx->stream_res.opp->inst };
-
-	for (odm_pipe = pipe_ctx->next_odm_pipe; odm_pipe; odm_pipe = odm_pipe->next_odm_pipe) {
-		opp_inst[opp_cnt] = odm_pipe->stream_res.opp->inst;
-		opp_cnt++;
-	}
-
-	if (opp_cnt > 1)
-		pipe_ctx->stream_res.tg->funcs->set_odm_combine(
-				pipe_ctx->stream_res.tg,
-				opp_inst, opp_cnt,
-				&pipe_ctx->stream->timing);
-	else
-		pipe_ctx->stream_res.tg->funcs->set_odm_bypass(
-				pipe_ctx->stream_res.tg, &pipe_ctx->stream->timing);
-}
-
-void dcn20_blank_pixel_data(
-		struct dc *dc,
-		struct pipe_ctx *pipe_ctx,
-		bool blank)
-{
-	struct tg_color black_color = {0};
-	struct stream_resource *stream_res = &pipe_ctx->stream_res;
-	struct dc_stream_state *stream = pipe_ctx->stream;
-	enum dc_color_space color_space = stream->output_color_space;
-	enum controller_dp_test_pattern test_pattern = CONTROLLER_DP_TEST_PATTERN_SOLID_COLOR;
-	enum controller_dp_color_space test_pattern_color_space = CONTROLLER_DP_COLOR_SPACE_UDEFINED;
-	struct pipe_ctx *odm_pipe;
-	int odm_cnt = 1;
-
-	int width = stream->timing.h_addressable + stream->timing.h_border_left + stream->timing.h_border_right;
-	int height = stream->timing.v_addressable + stream->timing.v_border_bottom + stream->timing.v_border_top;
-
-	if (stream->link->test_pattern_enabled)
-		return;
-
-	/* get opp dpg blank color */
-	color_space_to_black_color(dc, color_space, &black_color);
-
-	for (odm_pipe = pipe_ctx->next_odm_pipe; odm_pipe; odm_pipe = odm_pipe->next_odm_pipe)
-		odm_cnt++;
-
-	width = width / odm_cnt;
-
-	if (blank) {
-		dc->hwss.set_abm_immediate_disable(pipe_ctx);
-
-		if (dc->debug.visual_confirm != VISUAL_CONFIRM_DISABLE) {
-			test_pattern = CONTROLLER_DP_TEST_PATTERN_COLORSQUARES;
-			test_pattern_color_space = CONTROLLER_DP_COLOR_SPACE_RGB;
-		}
-	} else {
-		test_pattern = CONTROLLER_DP_TEST_PATTERN_VIDEOMODE;
-	}
-
-	dc->hwss.set_disp_pattern_generator(dc,
-			pipe_ctx,
-			test_pattern,
-			test_pattern_color_space,
-			stream->timing.display_color_depth,
-			&black_color,
-			width,
-			height,
-			0);
-
-	for (odm_pipe = pipe_ctx->next_odm_pipe; odm_pipe; odm_pipe = odm_pipe->next_odm_pipe) {
-		dc->hwss.set_disp_pattern_generator(dc,
-				odm_pipe,
-				dc->debug.visual_confirm != VISUAL_CONFIRM_DISABLE && blank ?
-						CONTROLLER_DP_TEST_PATTERN_COLORRAMP : test_pattern,
-				test_pattern_color_space,
-				stream->timing.display_color_depth,
-				&black_color,
-				width,
-				height,
-				0);
-	}
-
-	if (!blank)
-		if (stream_res->abm) {
-			dc->hwss.set_pipe(pipe_ctx);
-			stream_res->abm->funcs->set_abm_level(stream_res->abm, stream->abm_level);
-		}
-}
-
-
-static void dcn20_power_on_plane(
-	struct dce_hwseq *hws,
-	struct pipe_ctx *pipe_ctx)
-{
-	DC_LOGGER_INIT(hws->ctx->logger);
-	if (REG(DC_IP_REQUEST_CNTL)) {
-		REG_SET(DC_IP_REQUEST_CNTL, 0,
-				IP_REQUEST_EN, 1);
-
-		if (hws->funcs.dpp_pg_control)
-			hws->funcs.dpp_pg_control(hws, pipe_ctx->plane_res.dpp->inst, true);
-
-		if (hws->funcs.hubp_pg_control)
-			hws->funcs.hubp_pg_control(hws, pipe_ctx->plane_res.hubp->inst, true);
-
-		REG_SET(DC_IP_REQUEST_CNTL, 0,
-				IP_REQUEST_EN, 0);
-		DC_LOG_DEBUG(
-				"Un-gated front end for pipe %d\n", pipe_ctx->plane_res.hubp->inst);
-	}
-}
-
-void dcn20_enable_plane(
-	struct dc *dc,
-	struct pipe_ctx *pipe_ctx,
-	struct dc_state *context)
-{
-	//if (dc->debug.sanity_checks) {
-	//	dcn10_verify_allow_pstate_change_high(dc);
-	//}
-	dcn20_power_on_plane(dc->hwseq, pipe_ctx);
-
-	/* enable DCFCLK current DCHUB */
-	pipe_ctx->plane_res.hubp->funcs->hubp_clk_cntl(pipe_ctx->plane_res.hubp, true);
-
-	/* initialize HUBP on power up */
-	pipe_ctx->plane_res.hubp->funcs->hubp_init(pipe_ctx->plane_res.hubp);
-
-	/* make sure OPP_PIPE_CLOCK_EN = 1 */
-	pipe_ctx->stream_res.opp->funcs->opp_pipe_clock_control(
-			pipe_ctx->stream_res.opp,
-			true);
-
-/* TODO: enable/disable in dm as per update type.
-	if (plane_state) {
-		DC_LOG_DC(dc->ctx->logger,
-				"Pipe:%d 0x%x: addr hi:0x%x, "
-				"addr low:0x%x, "
-				"src: %d, %d, %d,"
-				" %d; dst: %d, %d, %d, %d;\n",
-				pipe_ctx->pipe_idx,
-				plane_state,
-				plane_state->address.grph.addr.high_part,
-				plane_state->address.grph.addr.low_part,
-				plane_state->src_rect.x,
-				plane_state->src_rect.y,
-				plane_state->src_rect.width,
-				plane_state->src_rect.height,
-				plane_state->dst_rect.x,
-				plane_state->dst_rect.y,
-				plane_state->dst_rect.width,
-				plane_state->dst_rect.height);
-
-		DC_LOG_DC(dc->ctx->logger,
-				"Pipe %d: width, height, x, y         format:%d\n"
-				"viewport:%d, %d, %d, %d\n"
-				"recout:  %d, %d, %d, %d\n",
-				pipe_ctx->pipe_idx,
-				plane_state->format,
-				pipe_ctx->plane_res.scl_data.viewport.width,
-				pipe_ctx->plane_res.scl_data.viewport.height,
-				pipe_ctx->plane_res.scl_data.viewport.x,
-				pipe_ctx->plane_res.scl_data.viewport.y,
-				pipe_ctx->plane_res.scl_data.recout.width,
-				pipe_ctx->plane_res.scl_data.recout.height,
-				pipe_ctx->plane_res.scl_data.recout.x,
-				pipe_ctx->plane_res.scl_data.recout.y);
-		print_rq_dlg_ttu(dc, pipe_ctx);
-	}
-*/
-	if (dc->vm_pa_config.valid) {
-		struct vm_system_aperture_param apt;
-
-		apt.sys_default.quad_part = 0;
-
-		apt.sys_low.quad_part = dc->vm_pa_config.system_aperture.start_addr;
-		apt.sys_high.quad_part = dc->vm_pa_config.system_aperture.end_addr;
-
-		// Program system aperture settings
-		pipe_ctx->plane_res.hubp->funcs->hubp_set_vm_system_aperture_settings(pipe_ctx->plane_res.hubp, &apt);
-	}
-
-	if (!pipe_ctx->top_pipe
-		&& pipe_ctx->plane_state
-		&& pipe_ctx->plane_state->flip_int_enabled
-		&& pipe_ctx->plane_res.hubp->funcs->hubp_set_flip_int)
-			pipe_ctx->plane_res.hubp->funcs->hubp_set_flip_int(pipe_ctx->plane_res.hubp);
-
-//	if (dc->debug.sanity_checks) {
-//		dcn10_verify_allow_pstate_change_high(dc);
-//	}
-}
-
-void dcn20_pipe_control_lock(
-	struct dc *dc,
-	struct pipe_ctx *pipe,
-	bool lock)
-{
-	struct pipe_ctx *temp_pipe;
-	bool flip_immediate = false;
-
-	/* use TG master update lock to lock everything on the TG
-	 * therefore only top pipe need to lock
-	 */
-	if (!pipe || pipe->top_pipe)
-		return;
-
-	if (pipe->plane_state != NULL)
-		flip_immediate = pipe->plane_state->flip_immediate;
-
-	if  (pipe->stream_res.gsl_group > 0) {
-	    temp_pipe = pipe->bottom_pipe;
-	    while (!flip_immediate && temp_pipe) {
-		    if (temp_pipe->plane_state != NULL)
-			    flip_immediate = temp_pipe->plane_state->flip_immediate;
-		    temp_pipe = temp_pipe->bottom_pipe;
-	    }
-	}
-
-	if (flip_immediate && lock) {
-		const int TIMEOUT_FOR_FLIP_PENDING = 100000;
-		int i;
-
-		temp_pipe = pipe;
-		while (temp_pipe) {
-			if (temp_pipe->plane_state && temp_pipe->plane_state->flip_immediate) {
-				for (i = 0; i < TIMEOUT_FOR_FLIP_PENDING; ++i) {
-					if (!temp_pipe->plane_res.hubp->funcs->hubp_is_flip_pending(temp_pipe->plane_res.hubp))
-						break;
-					udelay(1);
-				}
-
-				/* no reason it should take this long for immediate flips */
-				ASSERT(i != TIMEOUT_FOR_FLIP_PENDING);
-			}
-			temp_pipe = temp_pipe->bottom_pipe;
-		}
-	}
-
-	/* In flip immediate and pipe splitting case, we need to use GSL
-	 * for synchronization. Only do setup on locking and on flip type change.
-	 */
-	if (lock && (pipe->bottom_pipe != NULL || !flip_immediate))
-		if ((flip_immediate && pipe->stream_res.gsl_group == 0) ||
-		    (!flip_immediate && pipe->stream_res.gsl_group > 0))
-			dcn20_setup_gsl_group_as_lock(dc, pipe, flip_immediate);
-
-	if (pipe->plane_state != NULL)
-		flip_immediate = pipe->plane_state->flip_immediate;
-
-	temp_pipe = pipe->bottom_pipe;
-	while (flip_immediate && temp_pipe) {
-	    if (temp_pipe->plane_state != NULL)
-		flip_immediate = temp_pipe->plane_state->flip_immediate;
-	    temp_pipe = temp_pipe->bottom_pipe;
-	}
-
-	if (!lock && pipe->stream_res.gsl_group > 0 && pipe->plane_state &&
-		!flip_immediate)
-	    dcn20_setup_gsl_group_as_lock(dc, pipe, false);
-
-	if (pipe->stream && should_use_dmub_lock(pipe->stream->link)) {
-		union dmub_hw_lock_flags hw_locks = { 0 };
-		struct dmub_hw_lock_inst_flags inst_flags = { 0 };
-
-		hw_locks.bits.lock_pipe = 1;
-		inst_flags.otg_inst =  pipe->stream_res.tg->inst;
-
-		if (pipe->plane_state != NULL)
-			hw_locks.bits.triple_buffer_lock = pipe->plane_state->triplebuffer_flips;
-
-		dmub_hw_lock_mgr_cmd(dc->ctx->dmub_srv,
-					lock,
-					&hw_locks,
-					&inst_flags);
-	} else if (pipe->plane_state != NULL && pipe->plane_state->triplebuffer_flips) {
-		if (lock)
-			pipe->stream_res.tg->funcs->triplebuffer_lock(pipe->stream_res.tg);
-		else
-			pipe->stream_res.tg->funcs->triplebuffer_unlock(pipe->stream_res.tg);
-	} else {
-		if (lock)
-			pipe->stream_res.tg->funcs->lock(pipe->stream_res.tg);
-		else
-			pipe->stream_res.tg->funcs->unlock(pipe->stream_res.tg);
-	}
-}
-
-static void dcn20_detect_pipe_changes(struct pipe_ctx *old_pipe, struct pipe_ctx *new_pipe)
-{
-	new_pipe->update_flags.raw = 0;
-
-	/* Exit on unchanged, unused pipe */
-	if (!old_pipe->plane_state && !new_pipe->plane_state)
-		return;
-	/* Detect pipe enable/disable */
-	if (!old_pipe->plane_state && new_pipe->plane_state) {
-		new_pipe->update_flags.bits.enable = 1;
-		new_pipe->update_flags.bits.mpcc = 1;
-		new_pipe->update_flags.bits.dppclk = 1;
-		new_pipe->update_flags.bits.hubp_interdependent = 1;
-		new_pipe->update_flags.bits.hubp_rq_dlg_ttu = 1;
-		new_pipe->update_flags.bits.gamut_remap = 1;
-		new_pipe->update_flags.bits.scaler = 1;
-		new_pipe->update_flags.bits.viewport = 1;
-		new_pipe->update_flags.bits.det_size = 1;
-		if (!new_pipe->top_pipe && !new_pipe->prev_odm_pipe) {
-			new_pipe->update_flags.bits.odm = 1;
-			new_pipe->update_flags.bits.global_sync = 1;
-		}
-		return;
-	}
-	if (old_pipe->plane_state && !new_pipe->plane_state) {
-		new_pipe->update_flags.bits.disable = 1;
-		return;
-	}
-
-	/* Detect plane change */
-	if (old_pipe->plane_state != new_pipe->plane_state) {
-		new_pipe->update_flags.bits.plane_changed = true;
-	}
-
-	/* Detect top pipe only changes */
-	if (!new_pipe->top_pipe && !new_pipe->prev_odm_pipe) {
-		/* Detect odm changes */
-		if ((old_pipe->next_odm_pipe && new_pipe->next_odm_pipe
-			&& old_pipe->next_odm_pipe->pipe_idx != new_pipe->next_odm_pipe->pipe_idx)
-				|| (!old_pipe->next_odm_pipe && new_pipe->next_odm_pipe)
-				|| (old_pipe->next_odm_pipe && !new_pipe->next_odm_pipe)
-				|| old_pipe->stream_res.opp != new_pipe->stream_res.opp)
-			new_pipe->update_flags.bits.odm = 1;
-
-		/* Detect global sync changes */
-		if (old_pipe->pipe_dlg_param.vready_offset != new_pipe->pipe_dlg_param.vready_offset
-				|| old_pipe->pipe_dlg_param.vstartup_start != new_pipe->pipe_dlg_param.vstartup_start
-				|| old_pipe->pipe_dlg_param.vupdate_offset != new_pipe->pipe_dlg_param.vupdate_offset
-				|| old_pipe->pipe_dlg_param.vupdate_width != new_pipe->pipe_dlg_param.vupdate_width)
-			new_pipe->update_flags.bits.global_sync = 1;
-	}
-
-	if (old_pipe->det_buffer_size_kb != new_pipe->det_buffer_size_kb)
-		new_pipe->update_flags.bits.det_size = 1;
-
-	/*
-	 * Detect opp / tg change, only set on change, not on enable
-	 * Assume mpcc inst = pipe index, if not this code needs to be updated
-	 * since mpcc is what is affected by these. In fact all of our sequence
-	 * makes this assumption at the moment with how hubp reset is matched to
-	 * same index mpcc reset.
-	 */
-	if (old_pipe->stream_res.opp != new_pipe->stream_res.opp)
-		new_pipe->update_flags.bits.opp_changed = 1;
-	if (old_pipe->stream_res.tg != new_pipe->stream_res.tg)
-		new_pipe->update_flags.bits.tg_changed = 1;
-
-	/*
-	 * Detect mpcc blending changes, only dpp inst and opp matter here,
-	 * mpccs getting removed/inserted update connected ones during their own
-	 * programming
-	 */
-	if (old_pipe->plane_res.dpp != new_pipe->plane_res.dpp
-			|| old_pipe->stream_res.opp != new_pipe->stream_res.opp)
-		new_pipe->update_flags.bits.mpcc = 1;
-
-	/* Detect dppclk change */
-	if (old_pipe->plane_res.bw.dppclk_khz != new_pipe->plane_res.bw.dppclk_khz)
-		new_pipe->update_flags.bits.dppclk = 1;
-
-	/* Check for scl update */
-	if (memcmp(&old_pipe->plane_res.scl_data, &new_pipe->plane_res.scl_data, sizeof(struct scaler_data)))
-			new_pipe->update_flags.bits.scaler = 1;
-	/* Check for vp update */
-	if (memcmp(&old_pipe->plane_res.scl_data.viewport, &new_pipe->plane_res.scl_data.viewport, sizeof(struct rect))
-			|| memcmp(&old_pipe->plane_res.scl_data.viewport_c,
-				&new_pipe->plane_res.scl_data.viewport_c, sizeof(struct rect)))
-		new_pipe->update_flags.bits.viewport = 1;
-
-	/* Detect dlg/ttu/rq updates */
-	{
-		struct _vcs_dpi_display_dlg_regs_st old_dlg_attr = old_pipe->dlg_regs;
-		struct _vcs_dpi_display_ttu_regs_st old_ttu_attr = old_pipe->ttu_regs;
-		struct _vcs_dpi_display_dlg_regs_st *new_dlg_attr = &new_pipe->dlg_regs;
-		struct _vcs_dpi_display_ttu_regs_st *new_ttu_attr = &new_pipe->ttu_regs;
-
-		/* Detect pipe interdependent updates */
-		if (old_dlg_attr.dst_y_prefetch != new_dlg_attr->dst_y_prefetch ||
-				old_dlg_attr.vratio_prefetch != new_dlg_attr->vratio_prefetch ||
-				old_dlg_attr.vratio_prefetch_c != new_dlg_attr->vratio_prefetch_c ||
-				old_dlg_attr.dst_y_per_vm_vblank != new_dlg_attr->dst_y_per_vm_vblank ||
-				old_dlg_attr.dst_y_per_row_vblank != new_dlg_attr->dst_y_per_row_vblank ||
-				old_dlg_attr.dst_y_per_vm_flip != new_dlg_attr->dst_y_per_vm_flip ||
-				old_dlg_attr.dst_y_per_row_flip != new_dlg_attr->dst_y_per_row_flip ||
-				old_dlg_attr.refcyc_per_meta_chunk_vblank_l != new_dlg_attr->refcyc_per_meta_chunk_vblank_l ||
-				old_dlg_attr.refcyc_per_meta_chunk_vblank_c != new_dlg_attr->refcyc_per_meta_chunk_vblank_c ||
-				old_dlg_attr.refcyc_per_meta_chunk_flip_l != new_dlg_attr->refcyc_per_meta_chunk_flip_l ||
-				old_dlg_attr.refcyc_per_line_delivery_pre_l != new_dlg_attr->refcyc_per_line_delivery_pre_l ||
-				old_dlg_attr.refcyc_per_line_delivery_pre_c != new_dlg_attr->refcyc_per_line_delivery_pre_c ||
-				old_ttu_attr.refcyc_per_req_delivery_pre_l != new_ttu_attr->refcyc_per_req_delivery_pre_l ||
-				old_ttu_attr.refcyc_per_req_delivery_pre_c != new_ttu_attr->refcyc_per_req_delivery_pre_c ||
-				old_ttu_attr.refcyc_per_req_delivery_pre_cur0 != new_ttu_attr->refcyc_per_req_delivery_pre_cur0 ||
-				old_ttu_attr.refcyc_per_req_delivery_pre_cur1 != new_ttu_attr->refcyc_per_req_delivery_pre_cur1 ||
-				old_ttu_attr.min_ttu_vblank != new_ttu_attr->min_ttu_vblank ||
-				old_ttu_attr.qos_level_flip != new_ttu_attr->qos_level_flip) {
-			old_dlg_attr.dst_y_prefetch = new_dlg_attr->dst_y_prefetch;
-			old_dlg_attr.vratio_prefetch = new_dlg_attr->vratio_prefetch;
-			old_dlg_attr.vratio_prefetch_c = new_dlg_attr->vratio_prefetch_c;
-			old_dlg_attr.dst_y_per_vm_vblank = new_dlg_attr->dst_y_per_vm_vblank;
-			old_dlg_attr.dst_y_per_row_vblank = new_dlg_attr->dst_y_per_row_vblank;
-			old_dlg_attr.dst_y_per_vm_flip = new_dlg_attr->dst_y_per_vm_flip;
-			old_dlg_attr.dst_y_per_row_flip = new_dlg_attr->dst_y_per_row_flip;
-			old_dlg_attr.refcyc_per_meta_chunk_vblank_l = new_dlg_attr->refcyc_per_meta_chunk_vblank_l;
-			old_dlg_attr.refcyc_per_meta_chunk_vblank_c = new_dlg_attr->refcyc_per_meta_chunk_vblank_c;
-			old_dlg_attr.refcyc_per_meta_chunk_flip_l = new_dlg_attr->refcyc_per_meta_chunk_flip_l;
-			old_dlg_attr.refcyc_per_line_delivery_pre_l = new_dlg_attr->refcyc_per_line_delivery_pre_l;
-			old_dlg_attr.refcyc_per_line_delivery_pre_c = new_dlg_attr->refcyc_per_line_delivery_pre_c;
-			old_ttu_attr.refcyc_per_req_delivery_pre_l = new_ttu_attr->refcyc_per_req_delivery_pre_l;
-			old_ttu_attr.refcyc_per_req_delivery_pre_c = new_ttu_attr->refcyc_per_req_delivery_pre_c;
-			old_ttu_attr.refcyc_per_req_delivery_pre_cur0 = new_ttu_attr->refcyc_per_req_delivery_pre_cur0;
-			old_ttu_attr.refcyc_per_req_delivery_pre_cur1 = new_ttu_attr->refcyc_per_req_delivery_pre_cur1;
-			old_ttu_attr.min_ttu_vblank = new_ttu_attr->min_ttu_vblank;
-			old_ttu_attr.qos_level_flip = new_ttu_attr->qos_level_flip;
-			new_pipe->update_flags.bits.hubp_interdependent = 1;
-		}
-		/* Detect any other updates to ttu/rq/dlg */
-		if (memcmp(&old_dlg_attr, &new_pipe->dlg_regs, sizeof(old_dlg_attr)) ||
-				memcmp(&old_ttu_attr, &new_pipe->ttu_regs, sizeof(old_ttu_attr)) ||
-				memcmp(&old_pipe->rq_regs, &new_pipe->rq_regs, sizeof(old_pipe->rq_regs)))
-			new_pipe->update_flags.bits.hubp_rq_dlg_ttu = 1;
-	}
-}
-
-static void dcn20_update_dchubp_dpp(
-	struct dc *dc,
-	struct pipe_ctx *pipe_ctx,
-	struct dc_state *context)
-{
-	struct dce_hwseq *hws = dc->hwseq;
-	struct hubp *hubp = pipe_ctx->plane_res.hubp;
-	struct dpp *dpp = pipe_ctx->plane_res.dpp;
-	struct dc_plane_state *plane_state = pipe_ctx->plane_state;
-	bool viewport_changed = false;
-
-	if (pipe_ctx->update_flags.bits.dppclk)
-		dpp->funcs->dpp_dppclk_control(dpp, false, true);
-
-	/* TODO: Need input parameter to tell current DCHUB pipe tie to which OTG
-	 * VTG is within DCHUBBUB which is commond block share by each pipe HUBP.
-	 * VTG is 1:1 mapping with OTG. Each pipe HUBP will select which VTG
-	 */
-	if (pipe_ctx->update_flags.bits.hubp_rq_dlg_ttu) {
-		hubp->funcs->hubp_vtg_sel(hubp, pipe_ctx->stream_res.tg->inst);
-
-		hubp->funcs->hubp_setup(
-			hubp,
-			&pipe_ctx->dlg_regs,
-			&pipe_ctx->ttu_regs,
-			&pipe_ctx->rq_regs,
-			&pipe_ctx->pipe_dlg_param);
-
-		if (hubp->funcs->set_unbounded_requesting)
-			hubp->funcs->set_unbounded_requesting(hubp, pipe_ctx->unbounded_req);
-	}
-	if (pipe_ctx->update_flags.bits.hubp_interdependent)
-		hubp->funcs->hubp_setup_interdependent(
-			hubp,
-			&pipe_ctx->dlg_regs,
-			&pipe_ctx->ttu_regs);
-
-	if (pipe_ctx->update_flags.bits.enable ||
-			pipe_ctx->update_flags.bits.plane_changed ||
-			plane_state->update_flags.bits.bpp_change ||
-			plane_state->update_flags.bits.input_csc_change ||
-			plane_state->update_flags.bits.color_space_change ||
-			plane_state->update_flags.bits.coeff_reduction_change) {
-		struct dc_bias_and_scale bns_params = {0};
-
-		// program the input csc
-		dpp->funcs->dpp_setup(dpp,
-				plane_state->format,
-				EXPANSION_MODE_ZERO,
-				plane_state->input_csc_color_matrix,
-				plane_state->color_space,
-				NULL);
-
-		if (dpp->funcs->dpp_program_bias_and_scale) {
-			//TODO :for CNVC set scale and bias registers if necessary
-			build_prescale_params(&bns_params, plane_state);
-			dpp->funcs->dpp_program_bias_and_scale(dpp, &bns_params);
-		}
-	}
-
-	if (pipe_ctx->update_flags.bits.mpcc
-			|| pipe_ctx->update_flags.bits.plane_changed
-			|| plane_state->update_flags.bits.global_alpha_change
-			|| plane_state->update_flags.bits.per_pixel_alpha_change) {
-		// MPCC inst is equal to pipe index in practice
-		int mpcc_inst = hubp->inst;
-		int opp_inst;
-		int opp_count = dc->res_pool->pipe_count;
-
-		for (opp_inst = 0; opp_inst < opp_count; opp_inst++) {
-			if (dc->res_pool->opps[opp_inst]->mpcc_disconnect_pending[mpcc_inst]) {
-				dc->res_pool->mpc->funcs->wait_for_idle(dc->res_pool->mpc, mpcc_inst);
-				dc->res_pool->opps[opp_inst]->mpcc_disconnect_pending[mpcc_inst] = false;
-				break;
-			}
-		}
-		hws->funcs.update_mpcc(dc, pipe_ctx);
-	}
-
-	if (pipe_ctx->update_flags.bits.scaler ||
-			plane_state->update_flags.bits.scaling_change ||
-			plane_state->update_flags.bits.position_change ||
-			plane_state->update_flags.bits.per_pixel_alpha_change ||
-			pipe_ctx->stream->update_flags.bits.scaling) {
-		pipe_ctx->plane_res.scl_data.lb_params.alpha_en = pipe_ctx->plane_state->per_pixel_alpha;
-		ASSERT(pipe_ctx->plane_res.scl_data.lb_params.depth == LB_PIXEL_DEPTH_36BPP);
-		/* scaler configuration */
-		pipe_ctx->plane_res.dpp->funcs->dpp_set_scaler(
-				pipe_ctx->plane_res.dpp, &pipe_ctx->plane_res.scl_data);
-	}
-
-	if (pipe_ctx->update_flags.bits.viewport ||
-			(context == dc->current_state && plane_state->update_flags.bits.position_change) ||
-			(context == dc->current_state && plane_state->update_flags.bits.scaling_change) ||
-			(context == dc->current_state && pipe_ctx->stream->update_flags.bits.scaling)) {
-
-		hubp->funcs->mem_program_viewport(
-			hubp,
-			&pipe_ctx->plane_res.scl_data.viewport,
-			&pipe_ctx->plane_res.scl_data.viewport_c);
-		viewport_changed = true;
-	}
-
-	/* Any updates are handled in dc interface, just need to apply existing for plane enable */
-	if ((pipe_ctx->update_flags.bits.enable || pipe_ctx->update_flags.bits.opp_changed ||
-			pipe_ctx->update_flags.bits.scaler || viewport_changed == true) &&
-			pipe_ctx->stream->cursor_attributes.address.quad_part != 0) {
-		dc->hwss.set_cursor_position(pipe_ctx);
-		dc->hwss.set_cursor_attribute(pipe_ctx);
-
-		if (dc->hwss.set_cursor_sdr_white_level)
-			dc->hwss.set_cursor_sdr_white_level(pipe_ctx);
-	}
-
-	/* Any updates are handled in dc interface, just need
-	 * to apply existing for plane enable / opp change */
-	if (pipe_ctx->update_flags.bits.enable || pipe_ctx->update_flags.bits.opp_changed
-			|| pipe_ctx->stream->update_flags.bits.gamut_remap
-			|| pipe_ctx->stream->update_flags.bits.out_csc) {
-		/* dpp/cm gamut remap*/
-		dc->hwss.program_gamut_remap(pipe_ctx);
-
-		/*call the dcn2 method which uses mpc csc*/
-		dc->hwss.program_output_csc(dc,
-				pipe_ctx,
-				pipe_ctx->stream->output_color_space,
-				pipe_ctx->stream->csc_color_matrix.matrix,
-				hubp->opp_id);
-	}
-
-	if (pipe_ctx->update_flags.bits.enable ||
-			pipe_ctx->update_flags.bits.plane_changed ||
-			pipe_ctx->update_flags.bits.opp_changed ||
-			plane_state->update_flags.bits.pixel_format_change ||
-			plane_state->update_flags.bits.horizontal_mirror_change ||
-			plane_state->update_flags.bits.rotation_change ||
-			plane_state->update_flags.bits.swizzle_change ||
-			plane_state->update_flags.bits.dcc_change ||
-			plane_state->update_flags.bits.bpp_change ||
-			plane_state->update_flags.bits.scaling_change ||
-			plane_state->update_flags.bits.plane_size_change) {
-		struct plane_size size = plane_state->plane_size;
-
-		size.surface_size = pipe_ctx->plane_res.scl_data.viewport;
-		hubp->funcs->hubp_program_surface_config(
-			hubp,
-			plane_state->format,
-			&plane_state->tiling_info,
-			&size,
-			plane_state->rotation,
-			&plane_state->dcc,
-			plane_state->horizontal_mirror,
-			0);
-		hubp->power_gated = false;
-	}
-
-	if (pipe_ctx->update_flags.bits.enable ||
-		pipe_ctx->update_flags.bits.plane_changed ||
-		plane_state->update_flags.bits.addr_update)
-		hws->funcs.update_plane_addr(dc, pipe_ctx);
-
-
-
-	if (pipe_ctx->update_flags.bits.enable)
-		hubp->funcs->set_blank(hubp, false);
-}
-
-
-static void dcn20_program_pipe(
-		struct dc *dc,
-		struct pipe_ctx *pipe_ctx,
-		struct dc_state *context)
-{
-	struct dce_hwseq *hws = dc->hwseq;
-	/* Only need to unblank on top pipe */
-	if ((pipe_ctx->update_flags.bits.enable || pipe_ctx->stream->update_flags.bits.abm_level)
-			&& !pipe_ctx->top_pipe && !pipe_ctx->prev_odm_pipe)
-		hws->funcs.blank_pixel_data(dc, pipe_ctx, !pipe_ctx->plane_state->visible);
-
-	/* Only update TG on top pipe */
-	if (pipe_ctx->update_flags.bits.global_sync && !pipe_ctx->top_pipe
-			&& !pipe_ctx->prev_odm_pipe) {
-
-		pipe_ctx->stream_res.tg->funcs->program_global_sync(
-				pipe_ctx->stream_res.tg,
-				pipe_ctx->pipe_dlg_param.vready_offset,
-				pipe_ctx->pipe_dlg_param.vstartup_start,
-				pipe_ctx->pipe_dlg_param.vupdate_offset,
-				pipe_ctx->pipe_dlg_param.vupdate_width);
-
-		pipe_ctx->stream_res.tg->funcs->wait_for_state(pipe_ctx->stream_res.tg, CRTC_STATE_VBLANK);
-		pipe_ctx->stream_res.tg->funcs->wait_for_state(pipe_ctx->stream_res.tg, CRTC_STATE_VACTIVE);
-
-		pipe_ctx->stream_res.tg->funcs->set_vtg_params(
-				pipe_ctx->stream_res.tg, &pipe_ctx->stream->timing, true);
-
-		if (hws->funcs.setup_vupdate_interrupt)
-			hws->funcs.setup_vupdate_interrupt(dc, pipe_ctx);
-	}
-
-	if (pipe_ctx->update_flags.bits.odm)
-		hws->funcs.update_odm(dc, context, pipe_ctx);
-
-	if (pipe_ctx->update_flags.bits.enable) {
-		dcn20_enable_plane(dc, pipe_ctx, context);
-		if (dc->res_pool->hubbub->funcs->force_wm_propagate_to_pipes)
-			dc->res_pool->hubbub->funcs->force_wm_propagate_to_pipes(dc->res_pool->hubbub);
-	}
-
-	if (dc->res_pool->hubbub->funcs->program_det_size && pipe_ctx->update_flags.bits.det_size)
-		dc->res_pool->hubbub->funcs->program_det_size(
-			dc->res_pool->hubbub, pipe_ctx->plane_res.hubp->inst, pipe_ctx->det_buffer_size_kb);
-
-	if (pipe_ctx->update_flags.raw || pipe_ctx->plane_state->update_flags.raw || pipe_ctx->stream->update_flags.raw)
-		dcn20_update_dchubp_dpp(dc, pipe_ctx, context);
-
-	if (pipe_ctx->update_flags.bits.enable
-			|| pipe_ctx->plane_state->update_flags.bits.hdr_mult)
-		hws->funcs.set_hdr_multiplier(pipe_ctx);
-
-	if (pipe_ctx->update_flags.bits.enable ||
-			pipe_ctx->plane_state->update_flags.bits.in_transfer_func_change ||
-			pipe_ctx->plane_state->update_flags.bits.gamma_change)
-		hws->funcs.set_input_transfer_func(dc, pipe_ctx, pipe_ctx->plane_state);
-
-	/* dcn10_translate_regamma_to_hw_format takes 750us to finish
-	 * only do gamma programming for powering on, internal memcmp to avoid
-	 * updating on slave planes
-	 */
-	if (pipe_ctx->update_flags.bits.enable || pipe_ctx->stream->update_flags.bits.out_tf)
-		hws->funcs.set_output_transfer_func(dc, pipe_ctx, pipe_ctx->stream);
-
-	/* If the pipe has been enabled or has a different opp, we
-	 * should reprogram the fmt. This deals with cases where
-	 * interation between mpc and odm combine on different streams
-	 * causes a different pipe to be chosen to odm combine with.
-	 */
-	if (pipe_ctx->update_flags.bits.enable
-	    || pipe_ctx->update_flags.bits.opp_changed) {
-
-		pipe_ctx->stream_res.opp->funcs->opp_set_dyn_expansion(
-			pipe_ctx->stream_res.opp,
-			COLOR_SPACE_YCBCR601,
-			pipe_ctx->stream->timing.display_color_depth,
-			pipe_ctx->stream->signal);
-
-		pipe_ctx->stream_res.opp->funcs->opp_program_fmt(
-			pipe_ctx->stream_res.opp,
-			&pipe_ctx->stream->bit_depth_params,
-			&pipe_ctx->stream->clamping);
-	}
-}
-
-void dcn20_program_front_end_for_ctx(
-		struct dc *dc,
-		struct dc_state *context)
-{
-	int i;
-	struct dce_hwseq *hws = dc->hwseq;
-	DC_LOGGER_INIT(dc->ctx->logger);
-
-	/* Carry over GSL groups in case the context is changing. */
-       for (i = 0; i < dc->res_pool->pipe_count; i++) {
-               struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i];
-               struct pipe_ctx *old_pipe_ctx =
-                       &dc->current_state->res_ctx.pipe_ctx[i];
-
-               if (pipe_ctx->stream == old_pipe_ctx->stream)
-                       pipe_ctx->stream_res.gsl_group =
-                               old_pipe_ctx->stream_res.gsl_group;
-       }
-
-	if (dc->hwss.program_triplebuffer != NULL && dc->debug.enable_tri_buf) {
-		for (i = 0; i < dc->res_pool->pipe_count; i++) {
-			struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i];
-
-			if (!pipe_ctx->top_pipe && !pipe_ctx->prev_odm_pipe && pipe_ctx->plane_state) {
-				ASSERT(!pipe_ctx->plane_state->triplebuffer_flips);
-				/*turn off triple buffer for full update*/
-				dc->hwss.program_triplebuffer(
-						dc, pipe_ctx, pipe_ctx->plane_state->triplebuffer_flips);
-			}
-		}
-	}
-
-	/* Set pipe update flags and lock pipes */
-	for (i = 0; i < dc->res_pool->pipe_count; i++)
-		dcn20_detect_pipe_changes(&dc->current_state->res_ctx.pipe_ctx[i],
-				&context->res_ctx.pipe_ctx[i]);
-
-	/* OTG blank before disabling all front ends */
-	for (i = 0; i < dc->res_pool->pipe_count; i++)
-		if (context->res_ctx.pipe_ctx[i].update_flags.bits.disable
-				&& !context->res_ctx.pipe_ctx[i].top_pipe
-				&& !context->res_ctx.pipe_ctx[i].prev_odm_pipe
-				&& context->res_ctx.pipe_ctx[i].stream)
-			hws->funcs.blank_pixel_data(dc, &context->res_ctx.pipe_ctx[i], true);
-
-
-	/* Disconnect mpcc */
-	for (i = 0; i < dc->res_pool->pipe_count; i++)
-		if (context->res_ctx.pipe_ctx[i].update_flags.bits.disable
-				|| context->res_ctx.pipe_ctx[i].update_flags.bits.opp_changed) {
-			struct hubbub *hubbub = dc->res_pool->hubbub;
-
-			if (hubbub->funcs->program_det_size && context->res_ctx.pipe_ctx[i].update_flags.bits.disable)
-				hubbub->funcs->program_det_size(hubbub, dc->current_state->res_ctx.pipe_ctx[i].plane_res.hubp->inst, 0);
-			hws->funcs.plane_atomic_disconnect(dc, &dc->current_state->res_ctx.pipe_ctx[i]);
-			DC_LOG_DC("Reset mpcc for pipe %d\n", dc->current_state->res_ctx.pipe_ctx[i].pipe_idx);
-		}
-
-	/*
-	 * Program all updated pipes, order matters for mpcc setup. Start with
-	 * top pipe and program all pipes that follow in order
-	 */
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
-
-		if (pipe->plane_state && !pipe->top_pipe) {
-			while (pipe) {
-				if (hws->funcs.program_pipe)
-					hws->funcs.program_pipe(dc, pipe, context);
-				else
-					dcn20_program_pipe(dc, pipe, context);
-
-				pipe = pipe->bottom_pipe;
-			}
-		}
-		/* Program secondary blending tree and writeback pipes */
-		pipe = &context->res_ctx.pipe_ctx[i];
-		if (!pipe->top_pipe && !pipe->prev_odm_pipe
-				&& pipe->stream && pipe->stream->num_wb_info > 0
-				&& (pipe->update_flags.raw || (pipe->plane_state && pipe->plane_state->update_flags.raw)
-					|| pipe->stream->update_flags.raw)
-				&& hws->funcs.program_all_writeback_pipes_in_tree)
-			hws->funcs.program_all_writeback_pipes_in_tree(dc, pipe->stream, context);
-	}
-}
-
-void dcn20_post_unlock_program_front_end(
-		struct dc *dc,
-		struct dc_state *context)
-{
-	int i;
-	const unsigned int TIMEOUT_FOR_PIPE_ENABLE_MS = 100;
-	struct dce_hwseq *hwseq = dc->hwseq;
-
-	DC_LOGGER_INIT(dc->ctx->logger);
-
-	for (i = 0; i < dc->res_pool->pipe_count; i++)
-		if (context->res_ctx.pipe_ctx[i].update_flags.bits.disable)
-			dc->hwss.disable_plane(dc, &dc->current_state->res_ctx.pipe_ctx[i]);
-
-	/*
-	 * If we are enabling a pipe, we need to wait for pending clear as this is a critical
-	 * part of the enable operation otherwise, DM may request an immediate flip which
-	 * will cause HW to perform an "immediate enable" (as opposed to "vsync enable") which
-	 * is unsupported on DCN.
-	 */
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
-
-		if (pipe->plane_state && !pipe->top_pipe && pipe->update_flags.bits.enable) {
-			struct hubp *hubp = pipe->plane_res.hubp;
-			int j = 0;
-
-			for (j = 0; j < TIMEOUT_FOR_PIPE_ENABLE_MS*1000
-					&& hubp->funcs->hubp_is_flip_pending(hubp); j++)
-				mdelay(1);
-		}
-	}
-
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
-		struct pipe_ctx *mpcc_pipe;
-
-		if (pipe->vtp_locked) {
-			dc->hwseq->funcs.wait_for_blank_complete(pipe->stream_res.opp);
-			pipe->plane_res.hubp->funcs->set_blank(pipe->plane_res.hubp, true);
-			pipe->vtp_locked = false;
-
-			for (mpcc_pipe = pipe->bottom_pipe; mpcc_pipe; mpcc_pipe = mpcc_pipe->bottom_pipe)
-				mpcc_pipe->plane_res.hubp->funcs->set_blank(mpcc_pipe->plane_res.hubp, true);
-
-			for (i = 0; i < dc->res_pool->pipe_count; i++)
-				if (context->res_ctx.pipe_ctx[i].update_flags.bits.disable)
-					dc->hwss.disable_plane(dc, &dc->current_state->res_ctx.pipe_ctx[i]);
-		}
-	}
-	/* WA to apply WM setting*/
-	if (hwseq->wa.DEGVIDCN21)
-		dc->res_pool->hubbub->funcs->apply_DEDCN21_147_wa(dc->res_pool->hubbub);
-
-
-	/* WA for stutter underflow during MPO transitions when adding 2nd plane */
-	if (hwseq->wa.disallow_self_refresh_during_multi_plane_transition) {
-
-		if (dc->current_state->stream_status[0].plane_count == 1 &&
-				context->stream_status[0].plane_count > 1) {
-
-			struct timing_generator *tg = dc->res_pool->timing_generators[0];
-
-			dc->res_pool->hubbub->funcs->allow_self_refresh_control(dc->res_pool->hubbub, false);
-
-			hwseq->wa_state.disallow_self_refresh_during_multi_plane_transition_applied = true;
-			hwseq->wa_state.disallow_self_refresh_during_multi_plane_transition_applied_on_frame = tg->funcs->get_frame_count(tg);
-		}
-	}
-}
-
-void dcn20_prepare_bandwidth(
-		struct dc *dc,
-		struct dc_state *context)
-{
-	struct hubbub *hubbub = dc->res_pool->hubbub;
-
-	dc->clk_mgr->funcs->update_clocks(
-			dc->clk_mgr,
-			context,
-			false);
-
-	/* program dchubbub watermarks */
-	dc->wm_optimized_required = hubbub->funcs->program_watermarks(hubbub,
-					&context->bw_ctx.bw.dcn.watermarks,
-					dc->res_pool->ref_clocks.dchub_ref_clock_inKhz / 1000,
-					false);
-	/* decrease compbuf size */
-	if (hubbub->funcs->program_compbuf_size)
-		hubbub->funcs->program_compbuf_size(hubbub, context->bw_ctx.bw.dcn.compbuf_size_kb, false);
-}
-
-void dcn20_optimize_bandwidth(
-		struct dc *dc,
-		struct dc_state *context)
-{
-	struct hubbub *hubbub = dc->res_pool->hubbub;
-
-	/* program dchubbub watermarks */
-	hubbub->funcs->program_watermarks(hubbub,
-					&context->bw_ctx.bw.dcn.watermarks,
-					dc->res_pool->ref_clocks.dchub_ref_clock_inKhz / 1000,
-					true);
-
-	dc->clk_mgr->funcs->update_clocks(
-			dc->clk_mgr,
-			context,
-			true);
-	/* increase compbuf size */
-	if (hubbub->funcs->program_compbuf_size)
-		hubbub->funcs->program_compbuf_size(hubbub, context->bw_ctx.bw.dcn.compbuf_size_kb, true);
-}
-
-bool dcn20_update_bandwidth(
-		struct dc *dc,
-		struct dc_state *context)
-{
-	int i;
-	struct dce_hwseq *hws = dc->hwseq;
-
-	/* recalculate DML parameters */
-	if (!dc->res_pool->funcs->validate_bandwidth(dc, context, false))
-		return false;
-
-	/* apply updated bandwidth parameters */
-	dc->hwss.prepare_bandwidth(dc, context);
-
-	/* update hubp configs for all pipes */
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i];
-
-		if (pipe_ctx->plane_state == NULL)
-			continue;
-
-		if (pipe_ctx->top_pipe == NULL) {
-			bool blank = !is_pipe_tree_visible(pipe_ctx);
-
-			pipe_ctx->stream_res.tg->funcs->program_global_sync(
-					pipe_ctx->stream_res.tg,
-					pipe_ctx->pipe_dlg_param.vready_offset,
-					pipe_ctx->pipe_dlg_param.vstartup_start,
-					pipe_ctx->pipe_dlg_param.vupdate_offset,
-					pipe_ctx->pipe_dlg_param.vupdate_width);
-
-			pipe_ctx->stream_res.tg->funcs->set_vtg_params(
-					pipe_ctx->stream_res.tg, &pipe_ctx->stream->timing, false);
-
-			if (pipe_ctx->prev_odm_pipe == NULL)
-				hws->funcs.blank_pixel_data(dc, pipe_ctx, blank);
-
-			if (hws->funcs.setup_vupdate_interrupt)
-				hws->funcs.setup_vupdate_interrupt(dc, pipe_ctx);
-		}
-
-		pipe_ctx->plane_res.hubp->funcs->hubp_setup(
-				pipe_ctx->plane_res.hubp,
-					&pipe_ctx->dlg_regs,
-					&pipe_ctx->ttu_regs,
-					&pipe_ctx->rq_regs,
-					&pipe_ctx->pipe_dlg_param);
-	}
-
-	return true;
-}
-
-void dcn20_enable_writeback(
-		struct dc *dc,
-		struct dc_writeback_info *wb_info,
-		struct dc_state *context)
-{
-	struct dwbc *dwb;
-	struct mcif_wb *mcif_wb;
-	struct timing_generator *optc;
-
-	ASSERT(wb_info->dwb_pipe_inst < MAX_DWB_PIPES);
-	ASSERT(wb_info->wb_enabled);
-	dwb = dc->res_pool->dwbc[wb_info->dwb_pipe_inst];
-	mcif_wb = dc->res_pool->mcif_wb[wb_info->dwb_pipe_inst];
-
-	/* set the OPTC source mux */
-	optc = dc->res_pool->timing_generators[dwb->otg_inst];
-	optc->funcs->set_dwb_source(optc, wb_info->dwb_pipe_inst);
-	/* set MCIF_WB buffer and arbitration configuration */
-	mcif_wb->funcs->config_mcif_buf(mcif_wb, &wb_info->mcif_buf_params, wb_info->dwb_params.dest_height);
-	mcif_wb->funcs->config_mcif_arb(mcif_wb, &context->bw_ctx.bw.dcn.bw_writeback.mcif_wb_arb[wb_info->dwb_pipe_inst]);
-	/* Enable MCIF_WB */
-	mcif_wb->funcs->enable_mcif(mcif_wb);
-	/* Enable DWB */
-	dwb->funcs->enable(dwb, &wb_info->dwb_params);
-	/* TODO: add sequence to enable/disable warmup */
-}
-
-void dcn20_disable_writeback(
-		struct dc *dc,
-		unsigned int dwb_pipe_inst)
-{
-	struct dwbc *dwb;
-	struct mcif_wb *mcif_wb;
-
-	ASSERT(dwb_pipe_inst < MAX_DWB_PIPES);
-	dwb = dc->res_pool->dwbc[dwb_pipe_inst];
-	mcif_wb = dc->res_pool->mcif_wb[dwb_pipe_inst];
-
-	dwb->funcs->disable(dwb);
-	mcif_wb->funcs->disable_mcif(mcif_wb);
-}
-
-bool dcn20_wait_for_blank_complete(
-		struct output_pixel_processor *opp)
-{
-	int counter;
-
-	for (counter = 0; counter < 1000; counter++) {
-		if (opp->funcs->dpg_is_blanked(opp))
-			break;
-
-		udelay(100);
-	}
-
-	if (counter == 1000) {
-		dm_error("DC: failed to blank crtc!\n");
-		return false;
-	}
-
-	return true;
-}
-
-bool dcn20_dmdata_status_done(struct pipe_ctx *pipe_ctx)
-{
-	struct hubp *hubp = pipe_ctx->plane_res.hubp;
-
-	if (!hubp)
-		return false;
-	return hubp->funcs->dmdata_status_done(hubp);
-}
-
-void dcn20_disable_stream_gating(struct dc *dc, struct pipe_ctx *pipe_ctx)
-{
-	struct dce_hwseq *hws = dc->hwseq;
-
-	if (pipe_ctx->stream_res.dsc) {
-		struct pipe_ctx *odm_pipe = pipe_ctx->next_odm_pipe;
-
-		hws->funcs.dsc_pg_control(hws, pipe_ctx->stream_res.dsc->inst, true);
-		while (odm_pipe) {
-			hws->funcs.dsc_pg_control(hws, odm_pipe->stream_res.dsc->inst, true);
-			odm_pipe = odm_pipe->next_odm_pipe;
-		}
-	}
-}
-
-void dcn20_enable_stream_gating(struct dc *dc, struct pipe_ctx *pipe_ctx)
-{
-	struct dce_hwseq *hws = dc->hwseq;
-
-	if (pipe_ctx->stream_res.dsc) {
-		struct pipe_ctx *odm_pipe = pipe_ctx->next_odm_pipe;
-
-		hws->funcs.dsc_pg_control(hws, pipe_ctx->stream_res.dsc->inst, false);
-		while (odm_pipe) {
-			hws->funcs.dsc_pg_control(hws, odm_pipe->stream_res.dsc->inst, false);
-			odm_pipe = odm_pipe->next_odm_pipe;
-		}
-	}
-}
-
-void dcn20_set_dmdata_attributes(struct pipe_ctx *pipe_ctx)
-{
-	struct dc_dmdata_attributes attr = { 0 };
-	struct hubp *hubp = pipe_ctx->plane_res.hubp;
-
-	attr.dmdata_mode = DMDATA_HW_MODE;
-	attr.dmdata_size =
-		dc_is_hdmi_signal(pipe_ctx->stream->signal) ? 32 : 36;
-	attr.address.quad_part =
-			pipe_ctx->stream->dmdata_address.quad_part;
-	attr.dmdata_dl_delta = 0;
-	attr.dmdata_qos_mode = 0;
-	attr.dmdata_qos_level = 0;
-	attr.dmdata_repeat = 1; /* always repeat */
-	attr.dmdata_updated = 1;
-	attr.dmdata_sw_data = NULL;
-
-	hubp->funcs->dmdata_set_attributes(hubp, &attr);
-}
-
-void dcn20_init_vm_ctx(
-		struct dce_hwseq *hws,
-		struct dc *dc,
-		struct dc_virtual_addr_space_config *va_config,
-		int vmid)
-{
-	struct dcn_hubbub_virt_addr_config config;
-
-	if (vmid == 0) {
-		ASSERT(0); /* VMID cannot be 0 for vm context */
-		return;
-	}
-
-	config.page_table_start_addr = va_config->page_table_start_addr;
-	config.page_table_end_addr = va_config->page_table_end_addr;
-	config.page_table_block_size = va_config->page_table_block_size_in_bytes;
-	config.page_table_depth = va_config->page_table_depth;
-	config.page_table_base_addr = va_config->page_table_base_addr;
-
-	dc->res_pool->hubbub->funcs->init_vm_ctx(dc->res_pool->hubbub, &config, vmid);
-}
-
-int dcn20_init_sys_ctx(struct dce_hwseq *hws, struct dc *dc, struct dc_phy_addr_space_config *pa_config)
-{
-	struct dcn_hubbub_phys_addr_config config;
-
-	config.system_aperture.fb_top = pa_config->system_aperture.fb_top;
-	config.system_aperture.fb_offset = pa_config->system_aperture.fb_offset;
-	config.system_aperture.fb_base = pa_config->system_aperture.fb_base;
-	config.system_aperture.agp_top = pa_config->system_aperture.agp_top;
-	config.system_aperture.agp_bot = pa_config->system_aperture.agp_bot;
-	config.system_aperture.agp_base = pa_config->system_aperture.agp_base;
-	config.gart_config.page_table_start_addr = pa_config->gart_config.page_table_start_addr;
-	config.gart_config.page_table_end_addr = pa_config->gart_config.page_table_end_addr;
-	config.gart_config.page_table_base_addr = pa_config->gart_config.page_table_base_addr;
-	config.page_table_default_page_addr = pa_config->page_table_default_page_addr;
-
-	return dc->res_pool->hubbub->funcs->init_dchub_sys_ctx(dc->res_pool->hubbub, &config);
-}
-
-static bool patch_address_for_sbs_tb_stereo(
-		struct pipe_ctx *pipe_ctx, PHYSICAL_ADDRESS_LOC *addr)
-{
-	struct dc_plane_state *plane_state = pipe_ctx->plane_state;
-	bool sec_split = pipe_ctx->top_pipe &&
-			pipe_ctx->top_pipe->plane_state == pipe_ctx->plane_state;
-	if (sec_split && plane_state->address.type == PLN_ADDR_TYPE_GRPH_STEREO &&
-			(pipe_ctx->stream->timing.timing_3d_format ==
-			TIMING_3D_FORMAT_SIDE_BY_SIDE ||
-			pipe_ctx->stream->timing.timing_3d_format ==
-			TIMING_3D_FORMAT_TOP_AND_BOTTOM)) {
-		*addr = plane_state->address.grph_stereo.left_addr;
-		plane_state->address.grph_stereo.left_addr =
-				plane_state->address.grph_stereo.right_addr;
-		return true;
-	}
-
-	if (pipe_ctx->stream->view_format != VIEW_3D_FORMAT_NONE &&
-			plane_state->address.type != PLN_ADDR_TYPE_GRPH_STEREO) {
-		plane_state->address.type = PLN_ADDR_TYPE_GRPH_STEREO;
-		plane_state->address.grph_stereo.right_addr =
-				plane_state->address.grph_stereo.left_addr;
-		plane_state->address.grph_stereo.right_meta_addr =
-				plane_state->address.grph_stereo.left_meta_addr;
-	}
-	return false;
-}
-
-void dcn20_update_plane_addr(const struct dc *dc, struct pipe_ctx *pipe_ctx)
-{
-	bool addr_patched = false;
-	PHYSICAL_ADDRESS_LOC addr;
-	struct dc_plane_state *plane_state = pipe_ctx->plane_state;
-
-	if (plane_state == NULL)
-		return;
-
-	addr_patched = patch_address_for_sbs_tb_stereo(pipe_ctx, &addr);
-
-	// Call Helper to track VMID use
-	vm_helper_mark_vmid_used(dc->vm_helper, plane_state->address.vmid, pipe_ctx->plane_res.hubp->inst);
-
-	pipe_ctx->plane_res.hubp->funcs->hubp_program_surface_flip_and_addr(
-			pipe_ctx->plane_res.hubp,
-			&plane_state->address,
-			plane_state->flip_immediate);
-
-	plane_state->status.requested_address = plane_state->address;
-
-	if (plane_state->flip_immediate)
-		plane_state->status.current_address = plane_state->address;
-
-	if (addr_patched)
-		pipe_ctx->plane_state->address.grph_stereo.left_addr = addr;
-}
-
-void dcn20_unblank_stream(struct pipe_ctx *pipe_ctx,
-		struct dc_link_settings *link_settings)
-{
-	struct encoder_unblank_param params = {0};
-	struct dc_stream_state *stream = pipe_ctx->stream;
-	struct dc_link *link = stream->link;
-	struct dce_hwseq *hws = link->dc->hwseq;
-	struct pipe_ctx *odm_pipe;
-
-	params.opp_cnt = 1;
-	for (odm_pipe = pipe_ctx->next_odm_pipe; odm_pipe; odm_pipe = odm_pipe->next_odm_pipe) {
-		params.opp_cnt++;
-	}
-	/* only 3 items below are used by unblank */
-	params.timing = pipe_ctx->stream->timing;
-
-	params.link_settings.link_rate = link_settings->link_rate;
-
-	if (is_dp_128b_132b_signal(pipe_ctx)) {
-		/* TODO - DP2.0 HW: Set ODM mode in dp hpo encoder here */
-		pipe_ctx->stream_res.hpo_dp_stream_enc->funcs->dp_unblank(
-				pipe_ctx->stream_res.hpo_dp_stream_enc,
-				pipe_ctx->stream_res.tg->inst);
-	} else if (dc_is_dp_signal(pipe_ctx->stream->signal)) {
-		if (optc2_is_two_pixels_per_containter(&stream->timing) || params.opp_cnt > 1)
-			params.timing.pix_clk_100hz /= 2;
-		pipe_ctx->stream_res.stream_enc->funcs->dp_set_odm_combine(
-				pipe_ctx->stream_res.stream_enc, params.opp_cnt > 1);
-		pipe_ctx->stream_res.stream_enc->funcs->dp_unblank(link, pipe_ctx->stream_res.stream_enc, &params);
-	}
-
-	if (link->local_sink && link->local_sink->sink_signal == SIGNAL_TYPE_EDP) {
-		hws->funcs.edp_backlight_control(link, true);
-	}
-}
-
-void dcn20_setup_vupdate_interrupt(struct dc *dc, struct pipe_ctx *pipe_ctx)
-{
-	struct timing_generator *tg = pipe_ctx->stream_res.tg;
-	int start_line = dc->hwss.get_vupdate_offset_from_vsync(pipe_ctx);
-
-	if (start_line < 0)
-		start_line = 0;
-
-	if (tg->funcs->setup_vertical_interrupt2)
-		tg->funcs->setup_vertical_interrupt2(tg, start_line);
-}
-
-static void dcn20_reset_back_end_for_pipe(
-		struct dc *dc,
-		struct pipe_ctx *pipe_ctx,
-		struct dc_state *context)
-{
-	int i;
-	struct dc_link *link;
-	DC_LOGGER_INIT(dc->ctx->logger);
-	if (pipe_ctx->stream_res.stream_enc == NULL) {
-		pipe_ctx->stream = NULL;
-		return;
-	}
-
-	if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) {
-		link = pipe_ctx->stream->link;
-		/* DPMS may already disable or */
-		/* dpms_off status is incorrect due to fastboot
-		 * feature. When system resume from S4 with second
-		 * screen only, the dpms_off would be true but
-		 * VBIOS lit up eDP, so check link status too.
-		 */
-		if (!pipe_ctx->stream->dpms_off || link->link_status.link_active)
-			core_link_disable_stream(pipe_ctx);
-		else if (pipe_ctx->stream_res.audio)
-			dc->hwss.disable_audio_stream(pipe_ctx);
-
-		/* free acquired resources */
-		if (pipe_ctx->stream_res.audio) {
-			/*disable az_endpoint*/
-			pipe_ctx->stream_res.audio->funcs->az_disable(pipe_ctx->stream_res.audio);
-
-			/*free audio*/
-			if (dc->caps.dynamic_audio == true) {
-				/*we have to dynamic arbitrate the audio endpoints*/
-				/*we free the resource, need reset is_audio_acquired*/
-				update_audio_usage(&dc->current_state->res_ctx, dc->res_pool,
-						pipe_ctx->stream_res.audio, false);
-				pipe_ctx->stream_res.audio = NULL;
-			}
-		}
-	}
-	else if (pipe_ctx->stream_res.dsc) {
-		dp_set_dsc_enable(pipe_ctx, false);
-	}
-
-	/* by upper caller loop, parent pipe: pipe0, will be reset last.
-	 * back end share by all pipes and will be disable only when disable
-	 * parent pipe.
-	 */
-	if (pipe_ctx->top_pipe == NULL) {
-
-		dc->hwss.set_abm_immediate_disable(pipe_ctx);
-
-		pipe_ctx->stream_res.tg->funcs->disable_crtc(pipe_ctx->stream_res.tg);
-
-		pipe_ctx->stream_res.tg->funcs->enable_optc_clock(pipe_ctx->stream_res.tg, false);
-		if (pipe_ctx->stream_res.tg->funcs->set_odm_bypass)
-			pipe_ctx->stream_res.tg->funcs->set_odm_bypass(
-					pipe_ctx->stream_res.tg, &pipe_ctx->stream->timing);
-
-		if (pipe_ctx->stream_res.tg->funcs->set_drr)
-			pipe_ctx->stream_res.tg->funcs->set_drr(
-					pipe_ctx->stream_res.tg, NULL);
-	}
-
-	for (i = 0; i < dc->res_pool->pipe_count; i++)
-		if (&dc->current_state->res_ctx.pipe_ctx[i] == pipe_ctx)
-			break;
-
-	if (i == dc->res_pool->pipe_count)
-		return;
-
-	pipe_ctx->stream = NULL;
-	DC_LOG_DEBUG("Reset back end for pipe %d, tg:%d\n",
-					pipe_ctx->pipe_idx, pipe_ctx->stream_res.tg->inst);
-}
-
-void dcn20_reset_hw_ctx_wrap(
-		struct dc *dc,
-		struct dc_state *context)
-{
-	int i;
-	struct dce_hwseq *hws = dc->hwseq;
-
-	/* Reset Back End*/
-	for (i = dc->res_pool->pipe_count - 1; i >= 0 ; i--) {
-		struct pipe_ctx *pipe_ctx_old =
-			&dc->current_state->res_ctx.pipe_ctx[i];
-		struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i];
-
-		if (!pipe_ctx_old->stream)
-			continue;
-
-		if (pipe_ctx_old->top_pipe || pipe_ctx_old->prev_odm_pipe)
-			continue;
-
-		if (!pipe_ctx->stream ||
-				pipe_need_reprogram(pipe_ctx_old, pipe_ctx)) {
-			struct clock_source *old_clk = pipe_ctx_old->clock_source;
-
-			dcn20_reset_back_end_for_pipe(dc, pipe_ctx_old, dc->current_state);
-			if (hws->funcs.enable_stream_gating)
-				hws->funcs.enable_stream_gating(dc, pipe_ctx_old);
-			if (old_clk)
-				old_clk->funcs->cs_power_down(old_clk);
-		}
-	}
-}
-
-void dcn20_update_visual_confirm_color(struct dc *dc, struct pipe_ctx *pipe_ctx, struct tg_color *color, int mpcc_id)
-{
-	struct mpc *mpc = dc->res_pool->mpc;
-
-	// input to MPCC is always RGB, by default leave black_color at 0
-	if (dc->debug.visual_confirm == VISUAL_CONFIRM_HDR)
-		get_hdr_visual_confirm_color(pipe_ctx, color);
-	else if (dc->debug.visual_confirm == VISUAL_CONFIRM_SURFACE)
-		get_surface_visual_confirm_color(pipe_ctx, color);
-	else if (dc->debug.visual_confirm == VISUAL_CONFIRM_MPCTREE)
-		get_mpctree_visual_confirm_color(pipe_ctx, color);
-	else if (dc->debug.visual_confirm == VISUAL_CONFIRM_SWIZZLE)
-		get_surface_tile_visual_confirm_color(pipe_ctx, color);
-
-	if (mpc->funcs->set_bg_color)
-		mpc->funcs->set_bg_color(mpc, color, mpcc_id);
-}
-
-void dcn20_update_mpcc(struct dc *dc, struct pipe_ctx *pipe_ctx)
-{
-	struct hubp *hubp = pipe_ctx->plane_res.hubp;
-	struct mpcc_blnd_cfg blnd_cfg = {0};
-	bool per_pixel_alpha = pipe_ctx->plane_state->per_pixel_alpha;
-	int mpcc_id;
-	struct mpcc *new_mpcc;
-	struct mpc *mpc = dc->res_pool->mpc;
-	struct mpc_tree *mpc_tree_params = &(pipe_ctx->stream_res.opp->mpc_tree_params);
-
-	if (per_pixel_alpha)
-		blnd_cfg.alpha_mode = MPCC_ALPHA_BLEND_MODE_PER_PIXEL_ALPHA;
-	else
-		blnd_cfg.alpha_mode = MPCC_ALPHA_BLEND_MODE_GLOBAL_ALPHA;
-
-	blnd_cfg.overlap_only = false;
-	blnd_cfg.global_gain = 0xff;
-
-	if (pipe_ctx->plane_state->global_alpha)
-		blnd_cfg.global_alpha = pipe_ctx->plane_state->global_alpha_value;
-	else
-		blnd_cfg.global_alpha = 0xff;
-
-	blnd_cfg.background_color_bpc = 4;
-	blnd_cfg.bottom_gain_mode = 0;
-	blnd_cfg.top_gain = 0x1f000;
-	blnd_cfg.bottom_inside_gain = 0x1f000;
-	blnd_cfg.bottom_outside_gain = 0x1f000;
-	blnd_cfg.pre_multiplied_alpha = per_pixel_alpha;
-	if (pipe_ctx->plane_state->format
-			== SURFACE_PIXEL_FORMAT_GRPH_RGBE_ALPHA)
-		blnd_cfg.pre_multiplied_alpha = false;
-
-	/*
-	 * TODO: remove hack
-	 * Note: currently there is a bug in init_hw such that
-	 * on resume from hibernate, BIOS sets up MPCC0, and
-	 * we do mpcc_remove but the mpcc cannot go to idle
-	 * after remove. This cause us to pick mpcc1 here,
-	 * which causes a pstate hang for yet unknown reason.
-	 */
-	mpcc_id = hubp->inst;
-
-	/* If there is no full update, don't need to touch MPC tree*/
-	if (!pipe_ctx->plane_state->update_flags.bits.full_update &&
-		!pipe_ctx->update_flags.bits.mpcc) {
-		mpc->funcs->update_blending(mpc, &blnd_cfg, mpcc_id);
-		dc->hwss.update_visual_confirm_color(dc, pipe_ctx, &blnd_cfg.black_color, mpcc_id);
-		return;
-	}
-
-	/* check if this MPCC is already being used */
-	new_mpcc = mpc->funcs->get_mpcc_for_dpp(mpc_tree_params, mpcc_id);
-	/* remove MPCC if being used */
-	if (new_mpcc != NULL)
-		mpc->funcs->remove_mpcc(mpc, mpc_tree_params, new_mpcc);
-	else
-		if (dc->debug.sanity_checks)
-			mpc->funcs->assert_mpcc_idle_before_connect(
-					dc->res_pool->mpc, mpcc_id);
-
-	/* Call MPC to insert new plane */
-	new_mpcc = mpc->funcs->insert_plane(dc->res_pool->mpc,
-			mpc_tree_params,
-			&blnd_cfg,
-			NULL,
-			NULL,
-			hubp->inst,
-			mpcc_id);
-	dc->hwss.update_visual_confirm_color(dc, pipe_ctx, &blnd_cfg.black_color, mpcc_id);
-
-	ASSERT(new_mpcc != NULL);
-	hubp->opp_id = pipe_ctx->stream_res.opp->inst;
-	hubp->mpcc_id = mpcc_id;
-}
-
-void dcn20_enable_stream(struct pipe_ctx *pipe_ctx)
-{
-	enum dc_lane_count lane_count =
-		pipe_ctx->stream->link->cur_link_settings.lane_count;
-
-	struct dc_crtc_timing *timing = &pipe_ctx->stream->timing;
-	struct dc_link *link = pipe_ctx->stream->link;
-
-	uint32_t active_total_with_borders;
-	uint32_t early_control = 0;
-	struct timing_generator *tg = pipe_ctx->stream_res.tg;
-	struct link_encoder *link_enc;
-
-	if (link->is_dig_mapping_flexible &&
-			link->dc->res_pool->funcs->link_encs_assign)
-		link_enc = link_enc_cfg_get_link_enc_used_by_stream(link->ctx->dc, pipe_ctx->stream);
-	else
-		link_enc = link->link_enc;
-	ASSERT(link_enc);
-
-	/* For MST, there are multiply stream go to only one link.
-	 * connect DIG back_end to front_end while enable_stream and
-	 * disconnect them during disable_stream
-	 * BY this, it is logic clean to separate stream and link
-	 */
-	if (is_dp_128b_132b_signal(pipe_ctx)) {
-		if (pipe_ctx->stream->ctx->dc->hwseq->funcs.setup_hpo_hw_control)
-			pipe_ctx->stream->ctx->dc->hwseq->funcs.setup_hpo_hw_control(
-				pipe_ctx->stream->ctx->dc->hwseq, true);
-		setup_dp_hpo_stream(pipe_ctx, true);
-		pipe_ctx->stream_res.hpo_dp_stream_enc->funcs->enable_stream(
-				pipe_ctx->stream_res.hpo_dp_stream_enc);
-		pipe_ctx->stream_res.hpo_dp_stream_enc->funcs->map_stream_to_link(
-				pipe_ctx->stream_res.hpo_dp_stream_enc,
-				pipe_ctx->stream_res.hpo_dp_stream_enc->inst,
-				link->hpo_dp_link_enc->inst);
-	}
-
-	if (!is_dp_128b_132b_signal(pipe_ctx) && link_enc)
-		link_enc->funcs->connect_dig_be_to_fe(
-			link_enc, pipe_ctx->stream_res.stream_enc->id, true);
-
-	if (dc_is_dp_signal(pipe_ctx->stream->signal))
-		dp_source_sequence_trace(link, DPCD_SOURCE_SEQ_AFTER_CONNECT_DIG_FE_BE);
-
-	if (pipe_ctx->plane_state && pipe_ctx->plane_state->flip_immediate != 1) {
-		if (link->dc->hwss.program_dmdata_engine)
-			link->dc->hwss.program_dmdata_engine(pipe_ctx);
-	}
-
-	link->dc->hwss.update_info_frame(pipe_ctx);
-
-	if (dc_is_dp_signal(pipe_ctx->stream->signal))
-		dp_source_sequence_trace(link, DPCD_SOURCE_SEQ_AFTER_UPDATE_INFO_FRAME);
-
-	/* enable early control to avoid corruption on DP monitor*/
-	active_total_with_borders =
-			timing->h_addressable
-				+ timing->h_border_left
-				+ timing->h_border_right;
-
-	if (lane_count != 0)
-		early_control = active_total_with_borders % lane_count;
-
-	if (early_control == 0)
-		early_control = lane_count;
-
-	tg->funcs->set_early_control(tg, early_control);
-
-	/* enable audio only within mode set */
-	if (pipe_ctx->stream_res.audio != NULL) {
-		if (is_dp_128b_132b_signal(pipe_ctx))
-			pipe_ctx->stream_res.hpo_dp_stream_enc->funcs->dp_audio_enable(pipe_ctx->stream_res.hpo_dp_stream_enc);
-		else if (dc_is_dp_signal(pipe_ctx->stream->signal))
-			pipe_ctx->stream_res.stream_enc->funcs->dp_audio_enable(pipe_ctx->stream_res.stream_enc);
-	}
-}
-
-void dcn20_program_dmdata_engine(struct pipe_ctx *pipe_ctx)
-{
-	struct dc_stream_state    *stream     = pipe_ctx->stream;
-	struct hubp               *hubp       = pipe_ctx->plane_res.hubp;
-	bool                       enable     = false;
-	struct stream_encoder     *stream_enc = pipe_ctx->stream_res.stream_enc;
-	enum dynamic_metadata_mode mode       = dc_is_dp_signal(stream->signal)
-							? dmdata_dp
-							: dmdata_hdmi;
-
-	/* if using dynamic meta, don't set up generic infopackets */
-	if (pipe_ctx->stream->dmdata_address.quad_part != 0) {
-		pipe_ctx->stream_res.encoder_info_frame.hdrsmd.valid = false;
-		enable = true;
-	}
-
-	if (!hubp)
-		return;
-
-	if (!stream_enc || !stream_enc->funcs->set_dynamic_metadata)
-		return;
-
-	stream_enc->funcs->set_dynamic_metadata(stream_enc, enable,
-						hubp->inst, mode);
-}
-
-void dcn20_fpga_init_hw(struct dc *dc)
-{
-	int i, j;
-	struct dce_hwseq *hws = dc->hwseq;
-	struct resource_pool *res_pool = dc->res_pool;
-	struct dc_state  *context = dc->current_state;
-
-	if (dc->clk_mgr && dc->clk_mgr->funcs->init_clocks)
-		dc->clk_mgr->funcs->init_clocks(dc->clk_mgr);
-
-	// Initialize the dccg
-	if (res_pool->dccg->funcs->dccg_init)
-		res_pool->dccg->funcs->dccg_init(res_pool->dccg);
-
-	//Enable ability to power gate / don't force power on permanently
-	hws->funcs.enable_power_gating_plane(hws, true);
-
-	// Specific to FPGA dccg and registers
-	REG_WRITE(RBBMIF_TIMEOUT_DIS, 0xFFFFFFFF);
-	REG_WRITE(RBBMIF_TIMEOUT_DIS_2, 0xFFFFFFFF);
-
-	hws->funcs.dccg_init(hws);
-
-	REG_UPDATE(DCHUBBUB_GLOBAL_TIMER_CNTL, DCHUBBUB_GLOBAL_TIMER_REFDIV, 2);
-	REG_UPDATE(DCHUBBUB_GLOBAL_TIMER_CNTL, DCHUBBUB_GLOBAL_TIMER_ENABLE, 1);
-	if (REG(REFCLK_CNTL))
-		REG_WRITE(REFCLK_CNTL, 0);
-	//
-
-
-	/* Blank pixel data with OPP DPG */
-	for (i = 0; i < dc->res_pool->timing_generator_count; i++) {
-		struct timing_generator *tg = dc->res_pool->timing_generators[i];
-
-		if (tg->funcs->is_tg_enabled(tg))
-			dcn20_init_blank(dc, tg);
-	}
-
-	for (i = 0; i < res_pool->timing_generator_count; i++) {
-		struct timing_generator *tg = dc->res_pool->timing_generators[i];
-
-		if (tg->funcs->is_tg_enabled(tg))
-			tg->funcs->lock(tg);
-	}
-
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		struct dpp *dpp = res_pool->dpps[i];
-
-		dpp->funcs->dpp_reset(dpp);
-	}
-
-	/* Reset all MPCC muxes */
-	res_pool->mpc->funcs->mpc_init(res_pool->mpc);
-
-	/* initialize OPP mpc_tree parameter */
-	for (i = 0; i < dc->res_pool->res_cap->num_opp; i++) {
-		res_pool->opps[i]->mpc_tree_params.opp_id = res_pool->opps[i]->inst;
-		res_pool->opps[i]->mpc_tree_params.opp_list = NULL;
-		for (j = 0; j < MAX_PIPES; j++)
-			res_pool->opps[i]->mpcc_disconnect_pending[j] = false;
-	}
-
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		struct timing_generator *tg = dc->res_pool->timing_generators[i];
-		struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i];
-		struct hubp *hubp = dc->res_pool->hubps[i];
-		struct dpp *dpp = dc->res_pool->dpps[i];
-
-		pipe_ctx->stream_res.tg = tg;
-		pipe_ctx->pipe_idx = i;
-
-		pipe_ctx->plane_res.hubp = hubp;
-		pipe_ctx->plane_res.dpp = dpp;
-		pipe_ctx->plane_res.mpcc_inst = dpp->inst;
-		hubp->mpcc_id = dpp->inst;
-		hubp->opp_id = OPP_ID_INVALID;
-		hubp->power_gated = false;
-		pipe_ctx->stream_res.opp = NULL;
-
-		hubp->funcs->hubp_init(hubp);
-
-		//dc->res_pool->opps[i]->mpc_tree_params.opp_id = dc->res_pool->opps[i]->inst;
-		//dc->res_pool->opps[i]->mpc_tree_params.opp_list = NULL;
-		dc->res_pool->opps[i]->mpcc_disconnect_pending[pipe_ctx->plane_res.mpcc_inst] = true;
-		pipe_ctx->stream_res.opp = dc->res_pool->opps[i];
-		/*to do*/
-		hws->funcs.plane_atomic_disconnect(dc, pipe_ctx);
-	}
-
-	/* initialize DWB pointer to MCIF_WB */
-	for (i = 0; i < res_pool->res_cap->num_dwb; i++)
-		res_pool->dwbc[i]->mcif = res_pool->mcif_wb[i];
-
-	for (i = 0; i < dc->res_pool->timing_generator_count; i++) {
-		struct timing_generator *tg = dc->res_pool->timing_generators[i];
-
-		if (tg->funcs->is_tg_enabled(tg))
-			tg->funcs->unlock(tg);
-	}
-
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i];
-
-		dc->hwss.disable_plane(dc, pipe_ctx);
-
-		pipe_ctx->stream_res.tg = NULL;
-		pipe_ctx->plane_res.hubp = NULL;
-	}
-
-	for (i = 0; i < dc->res_pool->timing_generator_count; i++) {
-		struct timing_generator *tg = dc->res_pool->timing_generators[i];
-
-		tg->funcs->tg_init(tg);
-	}
-
-	if (dc->res_pool->hubbub->funcs->init_crb)
-		dc->res_pool->hubbub->funcs->init_crb(dc->res_pool->hubbub);
-}
-#ifndef TRIM_FSFT
-bool dcn20_optimize_timing_for_fsft(struct dc *dc,
-		struct dc_crtc_timing *timing,
-		unsigned int max_input_rate_in_khz)
-{
-	unsigned int old_v_front_porch;
-	unsigned int old_v_total;
-	unsigned int max_input_rate_in_100hz;
-	unsigned long long new_v_total;
-
-	max_input_rate_in_100hz = max_input_rate_in_khz * 10;
-	if (max_input_rate_in_100hz < timing->pix_clk_100hz)
-		return false;
-
-	old_v_total = timing->v_total;
-	old_v_front_porch = timing->v_front_porch;
-
-	timing->fast_transport_output_rate_100hz = timing->pix_clk_100hz;
-	timing->pix_clk_100hz = max_input_rate_in_100hz;
-
-	new_v_total = div_u64((unsigned long long)old_v_total * max_input_rate_in_100hz, timing->pix_clk_100hz);
-
-	timing->v_total = new_v_total;
-	timing->v_front_porch = old_v_front_porch + (timing->v_total - old_v_total);
-	return true;
-}
-#endif
-
-void dcn20_set_disp_pattern_generator(const struct dc *dc,
-		struct pipe_ctx *pipe_ctx,
-		enum controller_dp_test_pattern test_pattern,
-		enum controller_dp_color_space color_space,
-		enum dc_color_depth color_depth,
-		const struct tg_color *solid_color,
-		int width, int height, int offset)
-{
-	pipe_ctx->stream_res.opp->funcs->opp_set_disp_pattern_generator(pipe_ctx->stream_res.opp, test_pattern,
-			color_space, color_depth, solid_color, width, height, offset);
-}
diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c
deleted file mode 100644
index 3883f918b3bb..000000000000
--- a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c
+++ /dev/null
@@ -1,4106 +0,0 @@
-/*
-* Copyright 2016 Advanced Micro Devices, Inc.
- * Copyright 2019 Raptor Engineering, LLC
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: AMD
- *
- */
-
-#include <linux/slab.h>
-
-#include "dm_services.h"
-#include "dc.h"
-
-#include "dcn20_init.h"
-
-#include "resource.h"
-#include "include/irq_service_interface.h"
-#include "dcn20/dcn20_resource.h"
-
-#include "dml/dcn20/dcn20_fpu.h"
-
-#include "dcn10/dcn10_hubp.h"
-#include "dcn10/dcn10_ipp.h"
-#include "dcn20_hubbub.h"
-#include "dcn20_mpc.h"
-#include "dcn20_hubp.h"
-#include "irq/dcn20/irq_service_dcn20.h"
-#include "dcn20_dpp.h"
-#include "dcn20_optc.h"
-#include "dcn20_hwseq.h"
-#include "dce110/dce110_hw_sequencer.h"
-#include "dcn10/dcn10_resource.h"
-#include "dcn20_opp.h"
-
-#include "dcn20_dsc.h"
-
-#include "dcn20_link_encoder.h"
-#include "dcn20_stream_encoder.h"
-#include "dce/dce_clock_source.h"
-#include "dce/dce_audio.h"
-#include "dce/dce_hwseq.h"
-#include "virtual/virtual_stream_encoder.h"
-#include "dce110/dce110_resource.h"
-#include "dml/display_mode_vba.h"
-#include "dcn20_dccg.h"
-#include "dcn20_vmid.h"
-#include "dc_link_ddc.h"
-#include "dc_link_dp.h"
-#include "dce/dce_panel_cntl.h"
-
-#include "navi10_ip_offset.h"
-
-#include "dcn/dcn_2_0_0_offset.h"
-#include "dcn/dcn_2_0_0_sh_mask.h"
-#include "dpcs/dpcs_2_0_0_offset.h"
-#include "dpcs/dpcs_2_0_0_sh_mask.h"
-
-#include "nbio/nbio_2_3_offset.h"
-
-#include "dcn20/dcn20_dwb.h"
-#include "dcn20/dcn20_mmhubbub.h"
-
-#include "mmhub/mmhub_2_0_0_offset.h"
-#include "mmhub/mmhub_2_0_0_sh_mask.h"
-
-#include "reg_helper.h"
-#include "dce/dce_abm.h"
-#include "dce/dce_dmcu.h"
-#include "dce/dce_aux.h"
-#include "dce/dce_i2c.h"
-#include "vm_helper.h"
-#include "link_enc_cfg.h"
-
-#include "amdgpu_socbb.h"
-
-#define DC_LOGGER_INIT(logger)
-
-struct _vcs_dpi_ip_params_st dcn2_0_ip = {
-	.odm_capable = 1,
-	.gpuvm_enable = 0,
-	.hostvm_enable = 0,
-	.gpuvm_max_page_table_levels = 4,
-	.hostvm_max_page_table_levels = 4,
-	.hostvm_cached_page_table_levels = 0,
-	.pte_group_size_bytes = 2048,
-	.num_dsc = 6,
-	.rob_buffer_size_kbytes = 168,
-	.det_buffer_size_kbytes = 164,
-	.dpte_buffer_size_in_pte_reqs_luma = 84,
-	.pde_proc_buffer_size_64k_reqs = 48,
-	.dpp_output_buffer_pixels = 2560,
-	.opp_output_buffer_lines = 1,
-	.pixel_chunk_size_kbytes = 8,
-	.pte_chunk_size_kbytes = 2,
-	.meta_chunk_size_kbytes = 2,
-	.writeback_chunk_size_kbytes = 2,
-	.line_buffer_size_bits = 789504,
-	.is_line_buffer_bpp_fixed = 0,
-	.line_buffer_fixed_bpp = 0,
-	.dcc_supported = true,
-	.max_line_buffer_lines = 12,
-	.writeback_luma_buffer_size_kbytes = 12,
-	.writeback_chroma_buffer_size_kbytes = 8,
-	.writeback_chroma_line_buffer_width_pixels = 4,
-	.writeback_max_hscl_ratio = 1,
-	.writeback_max_vscl_ratio = 1,
-	.writeback_min_hscl_ratio = 1,
-	.writeback_min_vscl_ratio = 1,
-	.writeback_max_hscl_taps = 12,
-	.writeback_max_vscl_taps = 12,
-	.writeback_line_buffer_luma_buffer_size = 0,
-	.writeback_line_buffer_chroma_buffer_size = 14643,
-	.cursor_buffer_size = 8,
-	.cursor_chunk_size = 2,
-	.max_num_otg = 6,
-	.max_num_dpp = 6,
-	.max_num_wb = 1,
-	.max_dchub_pscl_bw_pix_per_clk = 4,
-	.max_pscl_lb_bw_pix_per_clk = 2,
-	.max_lb_vscl_bw_pix_per_clk = 4,
-	.max_vscl_hscl_bw_pix_per_clk = 4,
-	.max_hscl_ratio = 8,
-	.max_vscl_ratio = 8,
-	.hscl_mults = 4,
-	.vscl_mults = 4,
-	.max_hscl_taps = 8,
-	.max_vscl_taps = 8,
-	.dispclk_ramp_margin_percent = 1,
-	.underscan_factor = 1.10,
-	.min_vblank_lines = 32, //
-	.dppclk_delay_subtotal = 77, //
-	.dppclk_delay_scl_lb_only = 16,
-	.dppclk_delay_scl = 50,
-	.dppclk_delay_cnvc_formatter = 8,
-	.dppclk_delay_cnvc_cursor = 6,
-	.dispclk_delay_subtotal = 87, //
-	.dcfclk_cstate_latency = 10, // SRExitTime
-	.max_inter_dcn_tile_repeaters = 8,
-	.xfc_supported = true,
-	.xfc_fill_bw_overhead_percent = 10.0,
-	.xfc_fill_constant_bytes = 0,
-	.number_of_cursors = 1,
-};
-
-static struct _vcs_dpi_ip_params_st dcn2_0_nv14_ip = {
-	.odm_capable = 1,
-	.gpuvm_enable = 0,
-	.hostvm_enable = 0,
-	.gpuvm_max_page_table_levels = 4,
-	.hostvm_max_page_table_levels = 4,
-	.hostvm_cached_page_table_levels = 0,
-	.num_dsc = 5,
-	.rob_buffer_size_kbytes = 168,
-	.det_buffer_size_kbytes = 164,
-	.dpte_buffer_size_in_pte_reqs_luma = 84,
-	.dpte_buffer_size_in_pte_reqs_chroma = 42,//todo
-	.dpp_output_buffer_pixels = 2560,
-	.opp_output_buffer_lines = 1,
-	.pixel_chunk_size_kbytes = 8,
-	.pte_enable = 1,
-	.max_page_table_levels = 4,
-	.pte_chunk_size_kbytes = 2,
-	.meta_chunk_size_kbytes = 2,
-	.writeback_chunk_size_kbytes = 2,
-	.line_buffer_size_bits = 789504,
-	.is_line_buffer_bpp_fixed = 0,
-	.line_buffer_fixed_bpp = 0,
-	.dcc_supported = true,
-	.max_line_buffer_lines = 12,
-	.writeback_luma_buffer_size_kbytes = 12,
-	.writeback_chroma_buffer_size_kbytes = 8,
-	.writeback_chroma_line_buffer_width_pixels = 4,
-	.writeback_max_hscl_ratio = 1,
-	.writeback_max_vscl_ratio = 1,
-	.writeback_min_hscl_ratio = 1,
-	.writeback_min_vscl_ratio = 1,
-	.writeback_max_hscl_taps = 12,
-	.writeback_max_vscl_taps = 12,
-	.writeback_line_buffer_luma_buffer_size = 0,
-	.writeback_line_buffer_chroma_buffer_size = 14643,
-	.cursor_buffer_size = 8,
-	.cursor_chunk_size = 2,
-	.max_num_otg = 5,
-	.max_num_dpp = 5,
-	.max_num_wb = 1,
-	.max_dchub_pscl_bw_pix_per_clk = 4,
-	.max_pscl_lb_bw_pix_per_clk = 2,
-	.max_lb_vscl_bw_pix_per_clk = 4,
-	.max_vscl_hscl_bw_pix_per_clk = 4,
-	.max_hscl_ratio = 8,
-	.max_vscl_ratio = 8,
-	.hscl_mults = 4,
-	.vscl_mults = 4,
-	.max_hscl_taps = 8,
-	.max_vscl_taps = 8,
-	.dispclk_ramp_margin_percent = 1,
-	.underscan_factor = 1.10,
-	.min_vblank_lines = 32, //
-	.dppclk_delay_subtotal = 77, //
-	.dppclk_delay_scl_lb_only = 16,
-	.dppclk_delay_scl = 50,
-	.dppclk_delay_cnvc_formatter = 8,
-	.dppclk_delay_cnvc_cursor = 6,
-	.dispclk_delay_subtotal = 87, //
-	.dcfclk_cstate_latency = 10, // SRExitTime
-	.max_inter_dcn_tile_repeaters = 8,
-	.xfc_supported = true,
-	.xfc_fill_bw_overhead_percent = 10.0,
-	.xfc_fill_constant_bytes = 0,
-	.ptoi_supported = 0,
-	.number_of_cursors = 1,
-};
-
-static struct _vcs_dpi_soc_bounding_box_st dcn2_0_soc = {
-	/* Defaults that get patched on driver load from firmware. */
-	.clock_limits = {
-			{
-				.state = 0,
-				.dcfclk_mhz = 560.0,
-				.fabricclk_mhz = 560.0,
-				.dispclk_mhz = 513.0,
-				.dppclk_mhz = 513.0,
-				.phyclk_mhz = 540.0,
-				.socclk_mhz = 560.0,
-				.dscclk_mhz = 171.0,
-				.dram_speed_mts = 8960.0,
-			},
-			{
-				.state = 1,
-				.dcfclk_mhz = 694.0,
-				.fabricclk_mhz = 694.0,
-				.dispclk_mhz = 642.0,
-				.dppclk_mhz = 642.0,
-				.phyclk_mhz = 600.0,
-				.socclk_mhz = 694.0,
-				.dscclk_mhz = 214.0,
-				.dram_speed_mts = 11104.0,
-			},
-			{
-				.state = 2,
-				.dcfclk_mhz = 875.0,
-				.fabricclk_mhz = 875.0,
-				.dispclk_mhz = 734.0,
-				.dppclk_mhz = 734.0,
-				.phyclk_mhz = 810.0,
-				.socclk_mhz = 875.0,
-				.dscclk_mhz = 245.0,
-				.dram_speed_mts = 14000.0,
-			},
-			{
-				.state = 3,
-				.dcfclk_mhz = 1000.0,
-				.fabricclk_mhz = 1000.0,
-				.dispclk_mhz = 1100.0,
-				.dppclk_mhz = 1100.0,
-				.phyclk_mhz = 810.0,
-				.socclk_mhz = 1000.0,
-				.dscclk_mhz = 367.0,
-				.dram_speed_mts = 16000.0,
-			},
-			{
-				.state = 4,
-				.dcfclk_mhz = 1200.0,
-				.fabricclk_mhz = 1200.0,
-				.dispclk_mhz = 1284.0,
-				.dppclk_mhz = 1284.0,
-				.phyclk_mhz = 810.0,
-				.socclk_mhz = 1200.0,
-				.dscclk_mhz = 428.0,
-				.dram_speed_mts = 16000.0,
-			},
-			/*Extra state, no dispclk ramping*/
-			{
-				.state = 5,
-				.dcfclk_mhz = 1200.0,
-				.fabricclk_mhz = 1200.0,
-				.dispclk_mhz = 1284.0,
-				.dppclk_mhz = 1284.0,
-				.phyclk_mhz = 810.0,
-				.socclk_mhz = 1200.0,
-				.dscclk_mhz = 428.0,
-				.dram_speed_mts = 16000.0,
-			},
-		},
-	.num_states = 5,
-	.sr_exit_time_us = 8.6,
-	.sr_enter_plus_exit_time_us = 10.9,
-	.urgent_latency_us = 4.0,
-	.urgent_latency_pixel_data_only_us = 4.0,
-	.urgent_latency_pixel_mixed_with_vm_data_us = 4.0,
-	.urgent_latency_vm_data_only_us = 4.0,
-	.urgent_out_of_order_return_per_channel_pixel_only_bytes = 4096,
-	.urgent_out_of_order_return_per_channel_pixel_and_vm_bytes = 4096,
-	.urgent_out_of_order_return_per_channel_vm_only_bytes = 4096,
-	.pct_ideal_dram_sdp_bw_after_urgent_pixel_only = 40.0,
-	.pct_ideal_dram_sdp_bw_after_urgent_pixel_and_vm = 40.0,
-	.pct_ideal_dram_sdp_bw_after_urgent_vm_only = 40.0,
-	.max_avg_sdp_bw_use_normal_percent = 40.0,
-	.max_avg_dram_bw_use_normal_percent = 40.0,
-	.writeback_latency_us = 12.0,
-	.ideal_dram_bw_after_urgent_percent = 40.0,
-	.max_request_size_bytes = 256,
-	.dram_channel_width_bytes = 2,
-	.fabric_datapath_to_dcn_data_return_bytes = 64,
-	.dcn_downspread_percent = 0.5,
-	.downspread_percent = 0.38,
-	.dram_page_open_time_ns = 50.0,
-	.dram_rw_turnaround_time_ns = 17.5,
-	.dram_return_buffer_per_channel_bytes = 8192,
-	.round_trip_ping_latency_dcfclk_cycles = 131,
-	.urgent_out_of_order_return_per_channel_bytes = 256,
-	.channel_interleave_bytes = 256,
-	.num_banks = 8,
-	.num_chans = 16,
-	.vmm_page_size_bytes = 4096,
-	.dram_clock_change_latency_us = 404.0,
-	.dummy_pstate_latency_us = 5.0,
-	.writeback_dram_clock_change_latency_us = 23.0,
-	.return_bus_width_bytes = 64,
-	.dispclk_dppclk_vco_speed_mhz = 3850,
-	.xfc_bus_transport_time_us = 20,
-	.xfc_xbuf_latency_tolerance_us = 4,
-	.use_urgent_burst_bw = 0
-};
-
-static struct _vcs_dpi_soc_bounding_box_st dcn2_0_nv14_soc = {
-	.clock_limits = {
-			{
-				.state = 0,
-				.dcfclk_mhz = 560.0,
-				.fabricclk_mhz = 560.0,
-				.dispclk_mhz = 513.0,
-				.dppclk_mhz = 513.0,
-				.phyclk_mhz = 540.0,
-				.socclk_mhz = 560.0,
-				.dscclk_mhz = 171.0,
-				.dram_speed_mts = 8960.0,
-			},
-			{
-				.state = 1,
-				.dcfclk_mhz = 694.0,
-				.fabricclk_mhz = 694.0,
-				.dispclk_mhz = 642.0,
-				.dppclk_mhz = 642.0,
-				.phyclk_mhz = 600.0,
-				.socclk_mhz = 694.0,
-				.dscclk_mhz = 214.0,
-				.dram_speed_mts = 11104.0,
-			},
-			{
-				.state = 2,
-				.dcfclk_mhz = 875.0,
-				.fabricclk_mhz = 875.0,
-				.dispclk_mhz = 734.0,
-				.dppclk_mhz = 734.0,
-				.phyclk_mhz = 810.0,
-				.socclk_mhz = 875.0,
-				.dscclk_mhz = 245.0,
-				.dram_speed_mts = 14000.0,
-			},
-			{
-				.state = 3,
-				.dcfclk_mhz = 1000.0,
-				.fabricclk_mhz = 1000.0,
-				.dispclk_mhz = 1100.0,
-				.dppclk_mhz = 1100.0,
-				.phyclk_mhz = 810.0,
-				.socclk_mhz = 1000.0,
-				.dscclk_mhz = 367.0,
-				.dram_speed_mts = 16000.0,
-			},
-			{
-				.state = 4,
-				.dcfclk_mhz = 1200.0,
-				.fabricclk_mhz = 1200.0,
-				.dispclk_mhz = 1284.0,
-				.dppclk_mhz = 1284.0,
-				.phyclk_mhz = 810.0,
-				.socclk_mhz = 1200.0,
-				.dscclk_mhz = 428.0,
-				.dram_speed_mts = 16000.0,
-			},
-			/*Extra state, no dispclk ramping*/
-			{
-				.state = 5,
-				.dcfclk_mhz = 1200.0,
-				.fabricclk_mhz = 1200.0,
-				.dispclk_mhz = 1284.0,
-				.dppclk_mhz = 1284.0,
-				.phyclk_mhz = 810.0,
-				.socclk_mhz = 1200.0,
-				.dscclk_mhz = 428.0,
-				.dram_speed_mts = 16000.0,
-			},
-		},
-	.num_states = 5,
-	.sr_exit_time_us = 11.6,
-	.sr_enter_plus_exit_time_us = 13.9,
-	.urgent_latency_us = 4.0,
-	.urgent_latency_pixel_data_only_us = 4.0,
-	.urgent_latency_pixel_mixed_with_vm_data_us = 4.0,
-	.urgent_latency_vm_data_only_us = 4.0,
-	.urgent_out_of_order_return_per_channel_pixel_only_bytes = 4096,
-	.urgent_out_of_order_return_per_channel_pixel_and_vm_bytes = 4096,
-	.urgent_out_of_order_return_per_channel_vm_only_bytes = 4096,
-	.pct_ideal_dram_sdp_bw_after_urgent_pixel_only = 40.0,
-	.pct_ideal_dram_sdp_bw_after_urgent_pixel_and_vm = 40.0,
-	.pct_ideal_dram_sdp_bw_after_urgent_vm_only = 40.0,
-	.max_avg_sdp_bw_use_normal_percent = 40.0,
-	.max_avg_dram_bw_use_normal_percent = 40.0,
-	.writeback_latency_us = 12.0,
-	.ideal_dram_bw_after_urgent_percent = 40.0,
-	.max_request_size_bytes = 256,
-	.dram_channel_width_bytes = 2,
-	.fabric_datapath_to_dcn_data_return_bytes = 64,
-	.dcn_downspread_percent = 0.5,
-	.downspread_percent = 0.38,
-	.dram_page_open_time_ns = 50.0,
-	.dram_rw_turnaround_time_ns = 17.5,
-	.dram_return_buffer_per_channel_bytes = 8192,
-	.round_trip_ping_latency_dcfclk_cycles = 131,
-	.urgent_out_of_order_return_per_channel_bytes = 256,
-	.channel_interleave_bytes = 256,
-	.num_banks = 8,
-	.num_chans = 8,
-	.vmm_page_size_bytes = 4096,
-	.dram_clock_change_latency_us = 404.0,
-	.dummy_pstate_latency_us = 5.0,
-	.writeback_dram_clock_change_latency_us = 23.0,
-	.return_bus_width_bytes = 64,
-	.dispclk_dppclk_vco_speed_mhz = 3850,
-	.xfc_bus_transport_time_us = 20,
-	.xfc_xbuf_latency_tolerance_us = 4,
-	.use_urgent_burst_bw = 0
-};
-
-static struct _vcs_dpi_soc_bounding_box_st dcn2_0_nv12_soc = { 0 };
-
-#ifndef mmDP0_DP_DPHY_INTERNAL_CTRL
-	#define mmDP0_DP_DPHY_INTERNAL_CTRL		0x210f
-	#define mmDP0_DP_DPHY_INTERNAL_CTRL_BASE_IDX	2
-	#define mmDP1_DP_DPHY_INTERNAL_CTRL		0x220f
-	#define mmDP1_DP_DPHY_INTERNAL_CTRL_BASE_IDX	2
-	#define mmDP2_DP_DPHY_INTERNAL_CTRL		0x230f
-	#define mmDP2_DP_DPHY_INTERNAL_CTRL_BASE_IDX	2
-	#define mmDP3_DP_DPHY_INTERNAL_CTRL		0x240f
-	#define mmDP3_DP_DPHY_INTERNAL_CTRL_BASE_IDX	2
-	#define mmDP4_DP_DPHY_INTERNAL_CTRL		0x250f
-	#define mmDP4_DP_DPHY_INTERNAL_CTRL_BASE_IDX	2
-	#define mmDP5_DP_DPHY_INTERNAL_CTRL		0x260f
-	#define mmDP5_DP_DPHY_INTERNAL_CTRL_BASE_IDX	2
-	#define mmDP6_DP_DPHY_INTERNAL_CTRL		0x270f
-	#define mmDP6_DP_DPHY_INTERNAL_CTRL_BASE_IDX	2
-#endif
-
-
-enum dcn20_clk_src_array_id {
-	DCN20_CLK_SRC_PLL0,
-	DCN20_CLK_SRC_PLL1,
-	DCN20_CLK_SRC_PLL2,
-	DCN20_CLK_SRC_PLL3,
-	DCN20_CLK_SRC_PLL4,
-	DCN20_CLK_SRC_PLL5,
-	DCN20_CLK_SRC_TOTAL
-};
-
-/* begin *********************
- * macros to expend register list macro defined in HW object header file */
-
-/* DCN */
-/* TODO awful hack. fixup dcn20_dwb.h */
-#undef BASE_INNER
-#define BASE_INNER(seg) DCN_BASE__INST0_SEG ## seg
-
-#define BASE(seg) BASE_INNER(seg)
-
-#define SR(reg_name)\
-		.reg_name = BASE(mm ## reg_name ## _BASE_IDX) +  \
-					mm ## reg_name
-
-#define SRI(reg_name, block, id)\
-	.reg_name = BASE(mm ## block ## id ## _ ## reg_name ## _BASE_IDX) + \
-					mm ## block ## id ## _ ## reg_name
-
-#define SRIR(var_name, reg_name, block, id)\
-	.var_name = BASE(mm ## block ## id ## _ ## reg_name ## _BASE_IDX) + \
-					mm ## block ## id ## _ ## reg_name
-
-#define SRII(reg_name, block, id)\
-	.reg_name[id] = BASE(mm ## block ## id ## _ ## reg_name ## _BASE_IDX) + \
-					mm ## block ## id ## _ ## reg_name
-
-#define DCCG_SRII(reg_name, block, id)\
-	.block ## _ ## reg_name[id] = BASE(mm ## block ## id ## _ ## reg_name ## _BASE_IDX) + \
-					mm ## block ## id ## _ ## reg_name
-
-#define VUPDATE_SRII(reg_name, block, id)\
-	.reg_name[id] = BASE(mm ## reg_name ## _ ## block ## id ## _BASE_IDX) + \
-					mm ## reg_name ## _ ## block ## id
-
-/* NBIO */
-#define NBIO_BASE_INNER(seg) \
-	NBIO_BASE__INST0_SEG ## seg
-
-#define NBIO_BASE(seg) \
-	NBIO_BASE_INNER(seg)
-
-#define NBIO_SR(reg_name)\
-		.reg_name = NBIO_BASE(mm ## reg_name ## _BASE_IDX) + \
-					mm ## reg_name
-
-/* MMHUB */
-#define MMHUB_BASE_INNER(seg) \
-	MMHUB_BASE__INST0_SEG ## seg
-
-#define MMHUB_BASE(seg) \
-	MMHUB_BASE_INNER(seg)
-
-#define MMHUB_SR(reg_name)\
-		.reg_name = MMHUB_BASE(mmMM ## reg_name ## _BASE_IDX) + \
-					mmMM ## reg_name
-
-static const struct bios_registers bios_regs = {
-		NBIO_SR(BIOS_SCRATCH_3),
-		NBIO_SR(BIOS_SCRATCH_6)
-};
-
-#define clk_src_regs(index, pllid)\
-[index] = {\
-	CS_COMMON_REG_LIST_DCN2_0(index, pllid),\
-}
-
-static const struct dce110_clk_src_regs clk_src_regs[] = {
-	clk_src_regs(0, A),
-	clk_src_regs(1, B),
-	clk_src_regs(2, C),
-	clk_src_regs(3, D),
-	clk_src_regs(4, E),
-	clk_src_regs(5, F)
-};
-
-static const struct dce110_clk_src_shift cs_shift = {
-		CS_COMMON_MASK_SH_LIST_DCN2_0(__SHIFT)
-};
-
-static const struct dce110_clk_src_mask cs_mask = {
-		CS_COMMON_MASK_SH_LIST_DCN2_0(_MASK)
-};
-
-static const struct dce_dmcu_registers dmcu_regs = {
-		DMCU_DCN10_REG_LIST()
-};
-
-static const struct dce_dmcu_shift dmcu_shift = {
-		DMCU_MASK_SH_LIST_DCN10(__SHIFT)
-};
-
-static const struct dce_dmcu_mask dmcu_mask = {
-		DMCU_MASK_SH_LIST_DCN10(_MASK)
-};
-
-static const struct dce_abm_registers abm_regs = {
-		ABM_DCN20_REG_LIST()
-};
-
-static const struct dce_abm_shift abm_shift = {
-		ABM_MASK_SH_LIST_DCN20(__SHIFT)
-};
-
-static const struct dce_abm_mask abm_mask = {
-		ABM_MASK_SH_LIST_DCN20(_MASK)
-};
-
-#define audio_regs(id)\
-[id] = {\
-		AUD_COMMON_REG_LIST(id)\
-}
-
-static const struct dce_audio_registers audio_regs[] = {
-	audio_regs(0),
-	audio_regs(1),
-	audio_regs(2),
-	audio_regs(3),
-	audio_regs(4),
-	audio_regs(5),
-	audio_regs(6),
-};
-
-#define DCE120_AUD_COMMON_MASK_SH_LIST(mask_sh)\
-		SF(AZF0ENDPOINT0_AZALIA_F0_CODEC_ENDPOINT_INDEX, AZALIA_ENDPOINT_REG_INDEX, mask_sh),\
-		SF(AZF0ENDPOINT0_AZALIA_F0_CODEC_ENDPOINT_DATA, AZALIA_ENDPOINT_REG_DATA, mask_sh),\
-		AUD_COMMON_MASK_SH_LIST_BASE(mask_sh)
-
-static const struct dce_audio_shift audio_shift = {
-		DCE120_AUD_COMMON_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dce_audio_mask audio_mask = {
-		DCE120_AUD_COMMON_MASK_SH_LIST(_MASK)
-};
-
-#define stream_enc_regs(id)\
-[id] = {\
-	SE_DCN2_REG_LIST(id)\
-}
-
-static const struct dcn10_stream_enc_registers stream_enc_regs[] = {
-	stream_enc_regs(0),
-	stream_enc_regs(1),
-	stream_enc_regs(2),
-	stream_enc_regs(3),
-	stream_enc_regs(4),
-	stream_enc_regs(5),
-};
-
-static const struct dcn10_stream_encoder_shift se_shift = {
-		SE_COMMON_MASK_SH_LIST_DCN20(__SHIFT)
-};
-
-static const struct dcn10_stream_encoder_mask se_mask = {
-		SE_COMMON_MASK_SH_LIST_DCN20(_MASK)
-};
-
-
-#define aux_regs(id)\
-[id] = {\
-	DCN2_AUX_REG_LIST(id)\
-}
-
-static const struct dcn10_link_enc_aux_registers link_enc_aux_regs[] = {
-		aux_regs(0),
-		aux_regs(1),
-		aux_regs(2),
-		aux_regs(3),
-		aux_regs(4),
-		aux_regs(5)
-};
-
-#define hpd_regs(id)\
-[id] = {\
-	HPD_REG_LIST(id)\
-}
-
-static const struct dcn10_link_enc_hpd_registers link_enc_hpd_regs[] = {
-		hpd_regs(0),
-		hpd_regs(1),
-		hpd_regs(2),
-		hpd_regs(3),
-		hpd_regs(4),
-		hpd_regs(5)
-};
-
-#define link_regs(id, phyid)\
-[id] = {\
-	LE_DCN10_REG_LIST(id), \
-	UNIPHY_DCN2_REG_LIST(phyid), \
-	DPCS_DCN2_REG_LIST(id), \
-	SRI(DP_DPHY_INTERNAL_CTRL, DP, id) \
-}
-
-static const struct dcn10_link_enc_registers link_enc_regs[] = {
-	link_regs(0, A),
-	link_regs(1, B),
-	link_regs(2, C),
-	link_regs(3, D),
-	link_regs(4, E),
-	link_regs(5, F)
-};
-
-static const struct dcn10_link_enc_shift le_shift = {
-	LINK_ENCODER_MASK_SH_LIST_DCN20(__SHIFT),\
-	DPCS_DCN2_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dcn10_link_enc_mask le_mask = {
-	LINK_ENCODER_MASK_SH_LIST_DCN20(_MASK),\
-	DPCS_DCN2_MASK_SH_LIST(_MASK)
-};
-
-static const struct dce_panel_cntl_registers panel_cntl_regs[] = {
-	{ DCN_PANEL_CNTL_REG_LIST() }
-};
-
-static const struct dce_panel_cntl_shift panel_cntl_shift = {
-	DCE_PANEL_CNTL_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dce_panel_cntl_mask panel_cntl_mask = {
-	DCE_PANEL_CNTL_MASK_SH_LIST(_MASK)
-};
-
-#define ipp_regs(id)\
-[id] = {\
-	IPP_REG_LIST_DCN20(id),\
-}
-
-static const struct dcn10_ipp_registers ipp_regs[] = {
-	ipp_regs(0),
-	ipp_regs(1),
-	ipp_regs(2),
-	ipp_regs(3),
-	ipp_regs(4),
-	ipp_regs(5),
-};
-
-static const struct dcn10_ipp_shift ipp_shift = {
-		IPP_MASK_SH_LIST_DCN20(__SHIFT)
-};
-
-static const struct dcn10_ipp_mask ipp_mask = {
-		IPP_MASK_SH_LIST_DCN20(_MASK),
-};
-
-#define opp_regs(id)\
-[id] = {\
-	OPP_REG_LIST_DCN20(id),\
-}
-
-static const struct dcn20_opp_registers opp_regs[] = {
-	opp_regs(0),
-	opp_regs(1),
-	opp_regs(2),
-	opp_regs(3),
-	opp_regs(4),
-	opp_regs(5),
-};
-
-static const struct dcn20_opp_shift opp_shift = {
-		OPP_MASK_SH_LIST_DCN20(__SHIFT)
-};
-
-static const struct dcn20_opp_mask opp_mask = {
-		OPP_MASK_SH_LIST_DCN20(_MASK)
-};
-
-#define aux_engine_regs(id)\
-[id] = {\
-	AUX_COMMON_REG_LIST0(id), \
-	.AUXN_IMPCAL = 0, \
-	.AUXP_IMPCAL = 0, \
-	.AUX_RESET_MASK = DP_AUX0_AUX_CONTROL__AUX_RESET_MASK, \
-}
-
-static const struct dce110_aux_registers aux_engine_regs[] = {
-		aux_engine_regs(0),
-		aux_engine_regs(1),
-		aux_engine_regs(2),
-		aux_engine_regs(3),
-		aux_engine_regs(4),
-		aux_engine_regs(5)
-};
-
-#define tf_regs(id)\
-[id] = {\
-	TF_REG_LIST_DCN20(id),\
-	TF_REG_LIST_DCN20_COMMON_APPEND(id),\
-}
-
-static const struct dcn2_dpp_registers tf_regs[] = {
-	tf_regs(0),
-	tf_regs(1),
-	tf_regs(2),
-	tf_regs(3),
-	tf_regs(4),
-	tf_regs(5),
-};
-
-static const struct dcn2_dpp_shift tf_shift = {
-		TF_REG_LIST_SH_MASK_DCN20(__SHIFT),
-		TF_DEBUG_REG_LIST_SH_DCN20
-};
-
-static const struct dcn2_dpp_mask tf_mask = {
-		TF_REG_LIST_SH_MASK_DCN20(_MASK),
-		TF_DEBUG_REG_LIST_MASK_DCN20
-};
-
-#define dwbc_regs_dcn2(id)\
-[id] = {\
-	DWBC_COMMON_REG_LIST_DCN2_0(id),\
-		}
-
-static const struct dcn20_dwbc_registers dwbc20_regs[] = {
-	dwbc_regs_dcn2(0),
-};
-
-static const struct dcn20_dwbc_shift dwbc20_shift = {
-	DWBC_COMMON_MASK_SH_LIST_DCN2_0(__SHIFT)
-};
-
-static const struct dcn20_dwbc_mask dwbc20_mask = {
-	DWBC_COMMON_MASK_SH_LIST_DCN2_0(_MASK)
-};
-
-#define mcif_wb_regs_dcn2(id)\
-[id] = {\
-	MCIF_WB_COMMON_REG_LIST_DCN2_0(id),\
-		}
-
-static const struct dcn20_mmhubbub_registers mcif_wb20_regs[] = {
-	mcif_wb_regs_dcn2(0),
-};
-
-static const struct dcn20_mmhubbub_shift mcif_wb20_shift = {
-	MCIF_WB_COMMON_MASK_SH_LIST_DCN2_0(__SHIFT)
-};
-
-static const struct dcn20_mmhubbub_mask mcif_wb20_mask = {
-	MCIF_WB_COMMON_MASK_SH_LIST_DCN2_0(_MASK)
-};
-
-static const struct dcn20_mpc_registers mpc_regs = {
-		MPC_REG_LIST_DCN2_0(0),
-		MPC_REG_LIST_DCN2_0(1),
-		MPC_REG_LIST_DCN2_0(2),
-		MPC_REG_LIST_DCN2_0(3),
-		MPC_REG_LIST_DCN2_0(4),
-		MPC_REG_LIST_DCN2_0(5),
-		MPC_OUT_MUX_REG_LIST_DCN2_0(0),
-		MPC_OUT_MUX_REG_LIST_DCN2_0(1),
-		MPC_OUT_MUX_REG_LIST_DCN2_0(2),
-		MPC_OUT_MUX_REG_LIST_DCN2_0(3),
-		MPC_OUT_MUX_REG_LIST_DCN2_0(4),
-		MPC_OUT_MUX_REG_LIST_DCN2_0(5),
-		MPC_DBG_REG_LIST_DCN2_0()
-};
-
-static const struct dcn20_mpc_shift mpc_shift = {
-	MPC_COMMON_MASK_SH_LIST_DCN2_0(__SHIFT),
-	MPC_DEBUG_REG_LIST_SH_DCN20
-};
-
-static const struct dcn20_mpc_mask mpc_mask = {
-	MPC_COMMON_MASK_SH_LIST_DCN2_0(_MASK),
-	MPC_DEBUG_REG_LIST_MASK_DCN20
-};
-
-#define tg_regs(id)\
-[id] = {TG_COMMON_REG_LIST_DCN2_0(id)}
-
-
-static const struct dcn_optc_registers tg_regs[] = {
-	tg_regs(0),
-	tg_regs(1),
-	tg_regs(2),
-	tg_regs(3),
-	tg_regs(4),
-	tg_regs(5)
-};
-
-static const struct dcn_optc_shift tg_shift = {
-	TG_COMMON_MASK_SH_LIST_DCN2_0(__SHIFT)
-};
-
-static const struct dcn_optc_mask tg_mask = {
-	TG_COMMON_MASK_SH_LIST_DCN2_0(_MASK)
-};
-
-#define hubp_regs(id)\
-[id] = {\
-	HUBP_REG_LIST_DCN20(id)\
-}
-
-static const struct dcn_hubp2_registers hubp_regs[] = {
-		hubp_regs(0),
-		hubp_regs(1),
-		hubp_regs(2),
-		hubp_regs(3),
-		hubp_regs(4),
-		hubp_regs(5)
-};
-
-static const struct dcn_hubp2_shift hubp_shift = {
-		HUBP_MASK_SH_LIST_DCN20(__SHIFT)
-};
-
-static const struct dcn_hubp2_mask hubp_mask = {
-		HUBP_MASK_SH_LIST_DCN20(_MASK)
-};
-
-static const struct dcn_hubbub_registers hubbub_reg = {
-		HUBBUB_REG_LIST_DCN20(0)
-};
-
-static const struct dcn_hubbub_shift hubbub_shift = {
-		HUBBUB_MASK_SH_LIST_DCN20(__SHIFT)
-};
-
-static const struct dcn_hubbub_mask hubbub_mask = {
-		HUBBUB_MASK_SH_LIST_DCN20(_MASK)
-};
-
-#define vmid_regs(id)\
-[id] = {\
-		DCN20_VMID_REG_LIST(id)\
-}
-
-static const struct dcn_vmid_registers vmid_regs[] = {
-	vmid_regs(0),
-	vmid_regs(1),
-	vmid_regs(2),
-	vmid_regs(3),
-	vmid_regs(4),
-	vmid_regs(5),
-	vmid_regs(6),
-	vmid_regs(7),
-	vmid_regs(8),
-	vmid_regs(9),
-	vmid_regs(10),
-	vmid_regs(11),
-	vmid_regs(12),
-	vmid_regs(13),
-	vmid_regs(14),
-	vmid_regs(15)
-};
-
-static const struct dcn20_vmid_shift vmid_shifts = {
-		DCN20_VMID_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dcn20_vmid_mask vmid_masks = {
-		DCN20_VMID_MASK_SH_LIST(_MASK)
-};
-
-static const struct dce110_aux_registers_shift aux_shift = {
-		DCN_AUX_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dce110_aux_registers_mask aux_mask = {
-		DCN_AUX_MASK_SH_LIST(_MASK)
-};
-
-static int map_transmitter_id_to_phy_instance(
-	enum transmitter transmitter)
-{
-	switch (transmitter) {
-	case TRANSMITTER_UNIPHY_A:
-		return 0;
-	break;
-	case TRANSMITTER_UNIPHY_B:
-		return 1;
-	break;
-	case TRANSMITTER_UNIPHY_C:
-		return 2;
-	break;
-	case TRANSMITTER_UNIPHY_D:
-		return 3;
-	break;
-	case TRANSMITTER_UNIPHY_E:
-		return 4;
-	break;
-	case TRANSMITTER_UNIPHY_F:
-		return 5;
-	break;
-	default:
-		ASSERT(0);
-		return 0;
-	}
-}
-
-#define dsc_regsDCN20(id)\
-[id] = {\
-	DSC_REG_LIST_DCN20(id)\
-}
-
-static const struct dcn20_dsc_registers dsc_regs[] = {
-	dsc_regsDCN20(0),
-	dsc_regsDCN20(1),
-	dsc_regsDCN20(2),
-	dsc_regsDCN20(3),
-	dsc_regsDCN20(4),
-	dsc_regsDCN20(5)
-};
-
-static const struct dcn20_dsc_shift dsc_shift = {
-	DSC_REG_LIST_SH_MASK_DCN20(__SHIFT)
-};
-
-static const struct dcn20_dsc_mask dsc_mask = {
-	DSC_REG_LIST_SH_MASK_DCN20(_MASK)
-};
-
-static const struct dccg_registers dccg_regs = {
-		DCCG_REG_LIST_DCN2()
-};
-
-static const struct dccg_shift dccg_shift = {
-		DCCG_MASK_SH_LIST_DCN2(__SHIFT)
-};
-
-static const struct dccg_mask dccg_mask = {
-		DCCG_MASK_SH_LIST_DCN2(_MASK)
-};
-
-static const struct resource_caps res_cap_nv10 = {
-		.num_timing_generator = 6,
-		.num_opp = 6,
-		.num_video_plane = 6,
-		.num_audio = 7,
-		.num_stream_encoder = 6,
-		.num_pll = 6,
-		.num_dwb = 1,
-		.num_ddc = 6,
-		.num_vmid = 16,
-		.num_dsc = 6,
-};
-
-static const struct dc_plane_cap plane_cap = {
-	.type = DC_PLANE_TYPE_DCN_UNIVERSAL,
-	.blends_with_above = true,
-	.blends_with_below = true,
-	.per_pixel_alpha = true,
-
-	.pixel_format_support = {
-			.argb8888 = true,
-			.nv12 = true,
-			.fp16 = true,
-			.p010 = true
-	},
-
-	.max_upscale_factor = {
-			.argb8888 = 16000,
-			.nv12 = 16000,
-			.fp16 = 1
-	},
-
-	.max_downscale_factor = {
-			.argb8888 = 250,
-			.nv12 = 250,
-			.fp16 = 1
-	},
-	16,
-	16
-};
-static const struct resource_caps res_cap_nv14 = {
-		.num_timing_generator = 5,
-		.num_opp = 5,
-		.num_video_plane = 5,
-		.num_audio = 6,
-		.num_stream_encoder = 5,
-		.num_pll = 5,
-		.num_dwb = 1,
-		.num_ddc = 5,
-		.num_vmid = 16,
-		.num_dsc = 5,
-};
-
-static const struct dc_debug_options debug_defaults_drv = {
-		.disable_dmcu = false,
-		.force_abm_enable = false,
-		.timing_trace = false,
-		.clock_trace = true,
-		.disable_pplib_clock_request = true,
-		.pipe_split_policy = MPC_SPLIT_AVOID_MULT_DISP,
-		.force_single_disp_pipe_split = false,
-		.disable_dcc = DCC_ENABLE,
-		.vsr_support = true,
-		.performance_trace = false,
-		.max_downscale_src_width = 5120,/*upto 5K*/
-		.disable_pplib_wm_range = false,
-		.scl_reset_length10 = true,
-		.sanity_checks = false,
-		.underflow_assert_delay_us = 0xFFFFFFFF,
-};
-
-static const struct dc_debug_options debug_defaults_diags = {
-		.disable_dmcu = false,
-		.force_abm_enable = false,
-		.timing_trace = true,
-		.clock_trace = true,
-		.disable_dpp_power_gate = true,
-		.disable_hubp_power_gate = true,
-		.disable_clock_gate = true,
-		.disable_pplib_clock_request = true,
-		.disable_pplib_wm_range = true,
-		.disable_stutter = true,
-		.scl_reset_length10 = true,
-		.underflow_assert_delay_us = 0xFFFFFFFF,
-		.enable_tri_buf = true,
-};
-
-void dcn20_dpp_destroy(struct dpp **dpp)
-{
-	kfree(TO_DCN20_DPP(*dpp));
-	*dpp = NULL;
-}
-
-struct dpp *dcn20_dpp_create(
-	struct dc_context *ctx,
-	uint32_t inst)
-{
-	struct dcn20_dpp *dpp =
-		kzalloc(sizeof(struct dcn20_dpp), GFP_ATOMIC);
-
-	if (!dpp)
-		return NULL;
-
-	if (dpp2_construct(dpp, ctx, inst,
-			&tf_regs[inst], &tf_shift, &tf_mask))
-		return &dpp->base;
-
-	BREAK_TO_DEBUGGER();
-	kfree(dpp);
-	return NULL;
-}
-
-struct input_pixel_processor *dcn20_ipp_create(
-	struct dc_context *ctx, uint32_t inst)
-{
-	struct dcn10_ipp *ipp =
-		kzalloc(sizeof(struct dcn10_ipp), GFP_ATOMIC);
-
-	if (!ipp) {
-		BREAK_TO_DEBUGGER();
-		return NULL;
-	}
-
-	dcn20_ipp_construct(ipp, ctx, inst,
-			&ipp_regs[inst], &ipp_shift, &ipp_mask);
-	return &ipp->base;
-}
-
-
-struct output_pixel_processor *dcn20_opp_create(
-	struct dc_context *ctx, uint32_t inst)
-{
-	struct dcn20_opp *opp =
-		kzalloc(sizeof(struct dcn20_opp), GFP_ATOMIC);
-
-	if (!opp) {
-		BREAK_TO_DEBUGGER();
-		return NULL;
-	}
-
-	dcn20_opp_construct(opp, ctx, inst,
-			&opp_regs[inst], &opp_shift, &opp_mask);
-	return &opp->base;
-}
-
-struct dce_aux *dcn20_aux_engine_create(
-	struct dc_context *ctx,
-	uint32_t inst)
-{
-	struct aux_engine_dce110 *aux_engine =
-		kzalloc(sizeof(struct aux_engine_dce110), GFP_ATOMIC);
-
-	if (!aux_engine)
-		return NULL;
-
-	dce110_aux_engine_construct(aux_engine, ctx, inst,
-				    SW_AUX_TIMEOUT_PERIOD_MULTIPLIER * AUX_TIMEOUT_PERIOD,
-				    &aux_engine_regs[inst],
-					&aux_mask,
-					&aux_shift,
-					ctx->dc->caps.extended_aux_timeout_support);
-
-	return &aux_engine->base;
-}
-#define i2c_inst_regs(id) { I2C_HW_ENGINE_COMMON_REG_LIST(id) }
-
-static const struct dce_i2c_registers i2c_hw_regs[] = {
-		i2c_inst_regs(1),
-		i2c_inst_regs(2),
-		i2c_inst_regs(3),
-		i2c_inst_regs(4),
-		i2c_inst_regs(5),
-		i2c_inst_regs(6),
-};
-
-static const struct dce_i2c_shift i2c_shifts = {
-		I2C_COMMON_MASK_SH_LIST_DCN2(__SHIFT)
-};
-
-static const struct dce_i2c_mask i2c_masks = {
-		I2C_COMMON_MASK_SH_LIST_DCN2(_MASK)
-};
-
-struct dce_i2c_hw *dcn20_i2c_hw_create(
-	struct dc_context *ctx,
-	uint32_t inst)
-{
-	struct dce_i2c_hw *dce_i2c_hw =
-		kzalloc(sizeof(struct dce_i2c_hw), GFP_ATOMIC);
-
-	if (!dce_i2c_hw)
-		return NULL;
-
-	dcn2_i2c_hw_construct(dce_i2c_hw, ctx, inst,
-				    &i2c_hw_regs[inst], &i2c_shifts, &i2c_masks);
-
-	return dce_i2c_hw;
-}
-struct mpc *dcn20_mpc_create(struct dc_context *ctx)
-{
-	struct dcn20_mpc *mpc20 = kzalloc(sizeof(struct dcn20_mpc),
-					  GFP_ATOMIC);
-
-	if (!mpc20)
-		return NULL;
-
-	dcn20_mpc_construct(mpc20, ctx,
-			&mpc_regs,
-			&mpc_shift,
-			&mpc_mask,
-			6);
-
-	return &mpc20->base;
-}
-
-struct hubbub *dcn20_hubbub_create(struct dc_context *ctx)
-{
-	int i;
-	struct dcn20_hubbub *hubbub = kzalloc(sizeof(struct dcn20_hubbub),
-					  GFP_ATOMIC);
-
-	if (!hubbub)
-		return NULL;
-
-	hubbub2_construct(hubbub, ctx,
-			&hubbub_reg,
-			&hubbub_shift,
-			&hubbub_mask);
-
-	for (i = 0; i < res_cap_nv10.num_vmid; i++) {
-		struct dcn20_vmid *vmid = &hubbub->vmid[i];
-
-		vmid->ctx = ctx;
-
-		vmid->regs = &vmid_regs[i];
-		vmid->shifts = &vmid_shifts;
-		vmid->masks = &vmid_masks;
-	}
-
-	return &hubbub->base;
-}
-
-struct timing_generator *dcn20_timing_generator_create(
-		struct dc_context *ctx,
-		uint32_t instance)
-{
-	struct optc *tgn10 =
-		kzalloc(sizeof(struct optc), GFP_ATOMIC);
-
-	if (!tgn10)
-		return NULL;
-
-	tgn10->base.inst = instance;
-	tgn10->base.ctx = ctx;
-
-	tgn10->tg_regs = &tg_regs[instance];
-	tgn10->tg_shift = &tg_shift;
-	tgn10->tg_mask = &tg_mask;
-
-	dcn20_timing_generator_init(tgn10);
-
-	return &tgn10->base;
-}
-
-static const struct encoder_feature_support link_enc_feature = {
-		.max_hdmi_deep_color = COLOR_DEPTH_121212,
-		.max_hdmi_pixel_clock = 600000,
-		.hdmi_ycbcr420_supported = true,
-		.dp_ycbcr420_supported = true,
-		.fec_supported = true,
-		.flags.bits.IS_HBR2_CAPABLE = true,
-		.flags.bits.IS_HBR3_CAPABLE = true,
-		.flags.bits.IS_TPS3_CAPABLE = true,
-		.flags.bits.IS_TPS4_CAPABLE = true
-};
-
-struct link_encoder *dcn20_link_encoder_create(
-	const struct encoder_init_data *enc_init_data)
-{
-	struct dcn20_link_encoder *enc20 =
-		kzalloc(sizeof(struct dcn20_link_encoder), GFP_KERNEL);
-	int link_regs_id;
-
-	if (!enc20)
-		return NULL;
-
-	link_regs_id =
-		map_transmitter_id_to_phy_instance(enc_init_data->transmitter);
-
-	dcn20_link_encoder_construct(enc20,
-				      enc_init_data,
-				      &link_enc_feature,
-				      &link_enc_regs[link_regs_id],
-				      &link_enc_aux_regs[enc_init_data->channel - 1],
-				      &link_enc_hpd_regs[enc_init_data->hpd_source],
-				      &le_shift,
-				      &le_mask);
-
-	return &enc20->enc10.base;
-}
-
-static struct panel_cntl *dcn20_panel_cntl_create(const struct panel_cntl_init_data *init_data)
-{
-	struct dce_panel_cntl *panel_cntl =
-		kzalloc(sizeof(struct dce_panel_cntl), GFP_KERNEL);
-
-	if (!panel_cntl)
-		return NULL;
-
-	dce_panel_cntl_construct(panel_cntl,
-			init_data,
-			&panel_cntl_regs[init_data->inst],
-			&panel_cntl_shift,
-			&panel_cntl_mask);
-
-	return &panel_cntl->base;
-}
-
-static struct clock_source *dcn20_clock_source_create(
-	struct dc_context *ctx,
-	struct dc_bios *bios,
-	enum clock_source_id id,
-	const struct dce110_clk_src_regs *regs,
-	bool dp_clk_src)
-{
-	struct dce110_clk_src *clk_src =
-		kzalloc(sizeof(struct dce110_clk_src), GFP_ATOMIC);
-
-	if (!clk_src)
-		return NULL;
-
-	if (dcn20_clk_src_construct(clk_src, ctx, bios, id,
-			regs, &cs_shift, &cs_mask)) {
-		clk_src->base.dp_clk_src = dp_clk_src;
-		return &clk_src->base;
-	}
-
-	kfree(clk_src);
-	BREAK_TO_DEBUGGER();
-	return NULL;
-}
-
-static void read_dce_straps(
-	struct dc_context *ctx,
-	struct resource_straps *straps)
-{
-	generic_reg_get(ctx, mmDC_PINSTRAPS + BASE(mmDC_PINSTRAPS_BASE_IDX),
-		FN(DC_PINSTRAPS, DC_PINSTRAPS_AUDIO), &straps->dc_pinstraps_audio);
-}
-
-static struct audio *dcn20_create_audio(
-		struct dc_context *ctx, unsigned int inst)
-{
-	return dce_audio_create(ctx, inst,
-			&audio_regs[inst], &audio_shift, &audio_mask);
-}
-
-struct stream_encoder *dcn20_stream_encoder_create(
-	enum engine_id eng_id,
-	struct dc_context *ctx)
-{
-	struct dcn10_stream_encoder *enc1 =
-		kzalloc(sizeof(struct dcn10_stream_encoder), GFP_KERNEL);
-
-	if (!enc1)
-		return NULL;
-
-	if (ASICREV_IS_NAVI14_M(ctx->asic_id.hw_internal_rev)) {
-		if (eng_id >= ENGINE_ID_DIGD)
-			eng_id++;
-	}
-
-	dcn20_stream_encoder_construct(enc1, ctx, ctx->dc_bios, eng_id,
-					&stream_enc_regs[eng_id],
-					&se_shift, &se_mask);
-
-	return &enc1->base;
-}
-
-static const struct dce_hwseq_registers hwseq_reg = {
-		HWSEQ_DCN2_REG_LIST()
-};
-
-static const struct dce_hwseq_shift hwseq_shift = {
-		HWSEQ_DCN2_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dce_hwseq_mask hwseq_mask = {
-		HWSEQ_DCN2_MASK_SH_LIST(_MASK)
-};
-
-struct dce_hwseq *dcn20_hwseq_create(
-	struct dc_context *ctx)
-{
-	struct dce_hwseq *hws = kzalloc(sizeof(struct dce_hwseq), GFP_KERNEL);
-
-	if (hws) {
-		hws->ctx = ctx;
-		hws->regs = &hwseq_reg;
-		hws->shifts = &hwseq_shift;
-		hws->masks = &hwseq_mask;
-	}
-	return hws;
-}
-
-static const struct resource_create_funcs res_create_funcs = {
-	.read_dce_straps = read_dce_straps,
-	.create_audio = dcn20_create_audio,
-	.create_stream_encoder = dcn20_stream_encoder_create,
-	.create_hwseq = dcn20_hwseq_create,
-};
-
-static const struct resource_create_funcs res_create_maximus_funcs = {
-	.read_dce_straps = NULL,
-	.create_audio = NULL,
-	.create_stream_encoder = NULL,
-	.create_hwseq = dcn20_hwseq_create,
-};
-
-static void dcn20_pp_smu_destroy(struct pp_smu_funcs **pp_smu);
-
-void dcn20_clock_source_destroy(struct clock_source **clk_src)
-{
-	kfree(TO_DCE110_CLK_SRC(*clk_src));
-	*clk_src = NULL;
-}
-
-
-struct display_stream_compressor *dcn20_dsc_create(
-	struct dc_context *ctx, uint32_t inst)
-{
-	struct dcn20_dsc *dsc =
-		kzalloc(sizeof(struct dcn20_dsc), GFP_ATOMIC);
-
-	if (!dsc) {
-		BREAK_TO_DEBUGGER();
-		return NULL;
-	}
-
-	dsc2_construct(dsc, ctx, inst, &dsc_regs[inst], &dsc_shift, &dsc_mask);
-	return &dsc->base;
-}
-
-void dcn20_dsc_destroy(struct display_stream_compressor **dsc)
-{
-	kfree(container_of(*dsc, struct dcn20_dsc, base));
-	*dsc = NULL;
-}
-
-
-static void dcn20_resource_destruct(struct dcn20_resource_pool *pool)
-{
-	unsigned int i;
-
-	for (i = 0; i < pool->base.stream_enc_count; i++) {
-		if (pool->base.stream_enc[i] != NULL) {
-			kfree(DCN10STRENC_FROM_STRENC(pool->base.stream_enc[i]));
-			pool->base.stream_enc[i] = NULL;
-		}
-	}
-
-	for (i = 0; i < pool->base.res_cap->num_dsc; i++) {
-		if (pool->base.dscs[i] != NULL)
-			dcn20_dsc_destroy(&pool->base.dscs[i]);
-	}
-
-	if (pool->base.mpc != NULL) {
-		kfree(TO_DCN20_MPC(pool->base.mpc));
-		pool->base.mpc = NULL;
-	}
-	if (pool->base.hubbub != NULL) {
-		kfree(pool->base.hubbub);
-		pool->base.hubbub = NULL;
-	}
-	for (i = 0; i < pool->base.pipe_count; i++) {
-		if (pool->base.dpps[i] != NULL)
-			dcn20_dpp_destroy(&pool->base.dpps[i]);
-
-		if (pool->base.ipps[i] != NULL)
-			pool->base.ipps[i]->funcs->ipp_destroy(&pool->base.ipps[i]);
-
-		if (pool->base.hubps[i] != NULL) {
-			kfree(TO_DCN20_HUBP(pool->base.hubps[i]));
-			pool->base.hubps[i] = NULL;
-		}
-
-		if (pool->base.irqs != NULL) {
-			dal_irq_service_destroy(&pool->base.irqs);
-		}
-	}
-
-	for (i = 0; i < pool->base.res_cap->num_ddc; i++) {
-		if (pool->base.engines[i] != NULL)
-			dce110_engine_destroy(&pool->base.engines[i]);
-		if (pool->base.hw_i2cs[i] != NULL) {
-			kfree(pool->base.hw_i2cs[i]);
-			pool->base.hw_i2cs[i] = NULL;
-		}
-		if (pool->base.sw_i2cs[i] != NULL) {
-			kfree(pool->base.sw_i2cs[i]);
-			pool->base.sw_i2cs[i] = NULL;
-		}
-	}
-
-	for (i = 0; i < pool->base.res_cap->num_opp; i++) {
-		if (pool->base.opps[i] != NULL)
-			pool->base.opps[i]->funcs->opp_destroy(&pool->base.opps[i]);
-	}
-
-	for (i = 0; i < pool->base.res_cap->num_timing_generator; i++) {
-		if (pool->base.timing_generators[i] != NULL)	{
-			kfree(DCN10TG_FROM_TG(pool->base.timing_generators[i]));
-			pool->base.timing_generators[i] = NULL;
-		}
-	}
-
-	for (i = 0; i < pool->base.res_cap->num_dwb; i++) {
-		if (pool->base.dwbc[i] != NULL) {
-			kfree(TO_DCN20_DWBC(pool->base.dwbc[i]));
-			pool->base.dwbc[i] = NULL;
-		}
-		if (pool->base.mcif_wb[i] != NULL) {
-			kfree(TO_DCN20_MMHUBBUB(pool->base.mcif_wb[i]));
-			pool->base.mcif_wb[i] = NULL;
-		}
-	}
-
-	for (i = 0; i < pool->base.audio_count; i++) {
-		if (pool->base.audios[i])
-			dce_aud_destroy(&pool->base.audios[i]);
-	}
-
-	for (i = 0; i < pool->base.clk_src_count; i++) {
-		if (pool->base.clock_sources[i] != NULL) {
-			dcn20_clock_source_destroy(&pool->base.clock_sources[i]);
-			pool->base.clock_sources[i] = NULL;
-		}
-	}
-
-	if (pool->base.dp_clock_source != NULL) {
-		dcn20_clock_source_destroy(&pool->base.dp_clock_source);
-		pool->base.dp_clock_source = NULL;
-	}
-
-
-	if (pool->base.abm != NULL)
-		dce_abm_destroy(&pool->base.abm);
-
-	if (pool->base.dmcu != NULL)
-		dce_dmcu_destroy(&pool->base.dmcu);
-
-	if (pool->base.dccg != NULL)
-		dcn_dccg_destroy(&pool->base.dccg);
-
-	if (pool->base.pp_smu != NULL)
-		dcn20_pp_smu_destroy(&pool->base.pp_smu);
-
-	if (pool->base.oem_device != NULL)
-		dal_ddc_service_destroy(&pool->base.oem_device);
-}
-
-struct hubp *dcn20_hubp_create(
-	struct dc_context *ctx,
-	uint32_t inst)
-{
-	struct dcn20_hubp *hubp2 =
-		kzalloc(sizeof(struct dcn20_hubp), GFP_ATOMIC);
-
-	if (!hubp2)
-		return NULL;
-
-	if (hubp2_construct(hubp2, ctx, inst,
-			&hubp_regs[inst], &hubp_shift, &hubp_mask))
-		return &hubp2->base;
-
-	BREAK_TO_DEBUGGER();
-	kfree(hubp2);
-	return NULL;
-}
-
-static void get_pixel_clock_parameters(
-	struct pipe_ctx *pipe_ctx,
-	struct pixel_clk_params *pixel_clk_params)
-{
-	const struct dc_stream_state *stream = pipe_ctx->stream;
-	struct pipe_ctx *odm_pipe;
-	int opp_cnt = 1;
-	struct dc_link *link = stream->link;
-	struct link_encoder *link_enc = NULL;
-
-	for (odm_pipe = pipe_ctx->next_odm_pipe; odm_pipe; odm_pipe = odm_pipe->next_odm_pipe)
-		opp_cnt++;
-
-	pixel_clk_params->requested_pix_clk_100hz = stream->timing.pix_clk_100hz;
-
-	/* Links supporting dynamically assigned link encoder will be assigned next
-	 * available encoder if one not already assigned.
-	 */
-	if (link->is_dig_mapping_flexible &&
-			link->dc->res_pool->funcs->link_encs_assign) {
-		link_enc = link_enc_cfg_get_link_enc_used_by_stream(stream->ctx->dc, stream);
-		if (link_enc == NULL)
-			link_enc = link_enc_cfg_get_next_avail_link_enc(stream->ctx->dc);
-	} else
-		link_enc = stream->link->link_enc;
-	ASSERT(link_enc);
-
-	if (link_enc)
-		pixel_clk_params->encoder_object_id = link_enc->id;
-	pixel_clk_params->signal_type = pipe_ctx->stream->signal;
-	pixel_clk_params->controller_id = pipe_ctx->stream_res.tg->inst + 1;
-	/* TODO: un-hardcode*/
-	/* TODO - DP2.0 HW: calculate requested_sym_clk for UHBR rates */
-	pixel_clk_params->requested_sym_clk = LINK_RATE_LOW *
-		LINK_RATE_REF_FREQ_IN_KHZ;
-	pixel_clk_params->flags.ENABLE_SS = 0;
-	pixel_clk_params->color_depth =
-		stream->timing.display_color_depth;
-	pixel_clk_params->flags.DISPLAY_BLANKED = 1;
-	pixel_clk_params->pixel_encoding = stream->timing.pixel_encoding;
-
-	if (stream->timing.pixel_encoding == PIXEL_ENCODING_YCBCR422)
-		pixel_clk_params->color_depth = COLOR_DEPTH_888;
-
-	if (opp_cnt == 4)
-		pixel_clk_params->requested_pix_clk_100hz /= 4;
-	else if (optc2_is_two_pixels_per_containter(&stream->timing) || opp_cnt == 2)
-		pixel_clk_params->requested_pix_clk_100hz /= 2;
-
-	if (stream->timing.timing_3d_format == TIMING_3D_FORMAT_HW_FRAME_PACKING)
-		pixel_clk_params->requested_pix_clk_100hz *= 2;
-
-}
-
-static void build_clamping_params(struct dc_stream_state *stream)
-{
-	stream->clamping.clamping_level = CLAMPING_FULL_RANGE;
-	stream->clamping.c_depth = stream->timing.display_color_depth;
-	stream->clamping.pixel_encoding = stream->timing.pixel_encoding;
-}
-
-static enum dc_status build_pipe_hw_param(struct pipe_ctx *pipe_ctx)
-{
-
-	get_pixel_clock_parameters(pipe_ctx, &pipe_ctx->stream_res.pix_clk_params);
-
-	pipe_ctx->clock_source->funcs->get_pix_clk_dividers(
-		pipe_ctx->clock_source,
-		&pipe_ctx->stream_res.pix_clk_params,
-		&pipe_ctx->pll_settings);
-
-	pipe_ctx->stream->clamping.pixel_encoding = pipe_ctx->stream->timing.pixel_encoding;
-
-	resource_build_bit_depth_reduction_params(pipe_ctx->stream,
-					&pipe_ctx->stream->bit_depth_params);
-	build_clamping_params(pipe_ctx->stream);
-
-	return DC_OK;
-}
-
-enum dc_status dcn20_build_mapped_resource(const struct dc *dc, struct dc_state *context, struct dc_stream_state *stream)
-{
-	enum dc_status status = DC_OK;
-	struct pipe_ctx *pipe_ctx = resource_get_head_pipe_for_stream(&context->res_ctx, stream);
-
-	if (!pipe_ctx)
-		return DC_ERROR_UNEXPECTED;
-
-
-	status = build_pipe_hw_param(pipe_ctx);
-
-	return status;
-}
-
-
-void dcn20_acquire_dsc(const struct dc *dc,
-			struct resource_context *res_ctx,
-			struct display_stream_compressor **dsc,
-			int pipe_idx)
-{
-	int i;
-	const struct resource_pool *pool = dc->res_pool;
-	struct display_stream_compressor *dsc_old = dc->current_state->res_ctx.pipe_ctx[pipe_idx].stream_res.dsc;
-
-	ASSERT(*dsc == NULL); /* If this ASSERT fails, dsc was not released properly */
-	*dsc = NULL;
-
-	/* Always do 1-to-1 mapping when number of DSCs is same as number of pipes */
-	if (pool->res_cap->num_dsc == pool->res_cap->num_opp) {
-		*dsc = pool->dscs[pipe_idx];
-		res_ctx->is_dsc_acquired[pipe_idx] = true;
-		return;
-	}
-
-	/* Return old DSC to avoid the need for re-programming */
-	if (dsc_old && !res_ctx->is_dsc_acquired[dsc_old->inst]) {
-		*dsc = dsc_old;
-		res_ctx->is_dsc_acquired[dsc_old->inst] = true;
-		return ;
-	}
-
-	/* Find first free DSC */
-	for (i = 0; i < pool->res_cap->num_dsc; i++)
-		if (!res_ctx->is_dsc_acquired[i]) {
-			*dsc = pool->dscs[i];
-			res_ctx->is_dsc_acquired[i] = true;
-			break;
-		}
-}
-
-void dcn20_release_dsc(struct resource_context *res_ctx,
-			const struct resource_pool *pool,
-			struct display_stream_compressor **dsc)
-{
-	int i;
-
-	for (i = 0; i < pool->res_cap->num_dsc; i++)
-		if (pool->dscs[i] == *dsc) {
-			res_ctx->is_dsc_acquired[i] = false;
-			*dsc = NULL;
-			break;
-		}
-}
-
-
-
-enum dc_status dcn20_add_dsc_to_stream_resource(struct dc *dc,
-		struct dc_state *dc_ctx,
-		struct dc_stream_state *dc_stream)
-{
-	enum dc_status result = DC_OK;
-	int i;
-
-	/* Get a DSC if required and available */
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		struct pipe_ctx *pipe_ctx = &dc_ctx->res_ctx.pipe_ctx[i];
-
-		if (pipe_ctx->stream != dc_stream)
-			continue;
-
-		if (pipe_ctx->stream_res.dsc)
-			continue;
-
-		dcn20_acquire_dsc(dc, &dc_ctx->res_ctx, &pipe_ctx->stream_res.dsc, i);
-
-		/* The number of DSCs can be less than the number of pipes */
-		if (!pipe_ctx->stream_res.dsc) {
-			result = DC_NO_DSC_RESOURCE;
-		}
-
-		break;
-	}
-
-	return result;
-}
-
-
-static enum dc_status remove_dsc_from_stream_resource(struct dc *dc,
-		struct dc_state *new_ctx,
-		struct dc_stream_state *dc_stream)
-{
-	struct pipe_ctx *pipe_ctx = NULL;
-	int i;
-
-	for (i = 0; i < MAX_PIPES; i++) {
-		if (new_ctx->res_ctx.pipe_ctx[i].stream == dc_stream && !new_ctx->res_ctx.pipe_ctx[i].top_pipe) {
-			pipe_ctx = &new_ctx->res_ctx.pipe_ctx[i];
-
-			if (pipe_ctx->stream_res.dsc)
-				dcn20_release_dsc(&new_ctx->res_ctx, dc->res_pool, &pipe_ctx->stream_res.dsc);
-		}
-	}
-
-	if (!pipe_ctx)
-		return DC_ERROR_UNEXPECTED;
-	else
-		return DC_OK;
-}
-
-
-enum dc_status dcn20_add_stream_to_ctx(struct dc *dc, struct dc_state *new_ctx, struct dc_stream_state *dc_stream)
-{
-	enum dc_status result = DC_ERROR_UNEXPECTED;
-
-	result = resource_map_pool_resources(dc, new_ctx, dc_stream);
-
-	if (result == DC_OK)
-		result = resource_map_phy_clock_resources(dc, new_ctx, dc_stream);
-
-	/* Get a DSC if required and available */
-	if (result == DC_OK && dc_stream->timing.flags.DSC)
-		result = dcn20_add_dsc_to_stream_resource(dc, new_ctx, dc_stream);
-
-	if (result == DC_OK)
-		result = dcn20_build_mapped_resource(dc, new_ctx, dc_stream);
-
-	return result;
-}
-
-
-enum dc_status dcn20_remove_stream_from_ctx(struct dc *dc, struct dc_state *new_ctx, struct dc_stream_state *dc_stream)
-{
-	enum dc_status result = DC_OK;
-
-	result = remove_dsc_from_stream_resource(dc, new_ctx, dc_stream);
-
-	return result;
-}
-
-
-static void swizzle_to_dml_params(
-		enum swizzle_mode_values swizzle,
-		unsigned int *sw_mode)
-{
-	switch (swizzle) {
-	case DC_SW_LINEAR:
-		*sw_mode = dm_sw_linear;
-		break;
-	case DC_SW_4KB_S:
-		*sw_mode = dm_sw_4kb_s;
-		break;
-	case DC_SW_4KB_S_X:
-		*sw_mode = dm_sw_4kb_s_x;
-		break;
-	case DC_SW_4KB_D:
-		*sw_mode = dm_sw_4kb_d;
-		break;
-	case DC_SW_4KB_D_X:
-		*sw_mode = dm_sw_4kb_d_x;
-		break;
-	case DC_SW_64KB_S:
-		*sw_mode = dm_sw_64kb_s;
-		break;
-	case DC_SW_64KB_S_X:
-		*sw_mode = dm_sw_64kb_s_x;
-		break;
-	case DC_SW_64KB_S_T:
-		*sw_mode = dm_sw_64kb_s_t;
-		break;
-	case DC_SW_64KB_D:
-		*sw_mode = dm_sw_64kb_d;
-		break;
-	case DC_SW_64KB_D_X:
-		*sw_mode = dm_sw_64kb_d_x;
-		break;
-	case DC_SW_64KB_D_T:
-		*sw_mode = dm_sw_64kb_d_t;
-		break;
-	case DC_SW_64KB_R_X:
-		*sw_mode = dm_sw_64kb_r_x;
-		break;
-	case DC_SW_VAR_S:
-		*sw_mode = dm_sw_var_s;
-		break;
-	case DC_SW_VAR_S_X:
-		*sw_mode = dm_sw_var_s_x;
-		break;
-	case DC_SW_VAR_D:
-		*sw_mode = dm_sw_var_d;
-		break;
-	case DC_SW_VAR_D_X:
-		*sw_mode = dm_sw_var_d_x;
-		break;
-	case DC_SW_VAR_R_X:
-		*sw_mode = dm_sw_var_r_x;
-		break;
-	default:
-		ASSERT(0); /* Not supported */
-		break;
-	}
-}
-
-bool dcn20_split_stream_for_odm(
-		const struct dc *dc,
-		struct resource_context *res_ctx,
-		struct pipe_ctx *prev_odm_pipe,
-		struct pipe_ctx *next_odm_pipe)
-{
-	int pipe_idx = next_odm_pipe->pipe_idx;
-	const struct resource_pool *pool = dc->res_pool;
-
-	*next_odm_pipe = *prev_odm_pipe;
-
-	next_odm_pipe->pipe_idx = pipe_idx;
-	next_odm_pipe->plane_res.mi = pool->mis[next_odm_pipe->pipe_idx];
-	next_odm_pipe->plane_res.hubp = pool->hubps[next_odm_pipe->pipe_idx];
-	next_odm_pipe->plane_res.ipp = pool->ipps[next_odm_pipe->pipe_idx];
-	next_odm_pipe->plane_res.xfm = pool->transforms[next_odm_pipe->pipe_idx];
-	next_odm_pipe->plane_res.dpp = pool->dpps[next_odm_pipe->pipe_idx];
-	next_odm_pipe->plane_res.mpcc_inst = pool->dpps[next_odm_pipe->pipe_idx]->inst;
-	next_odm_pipe->stream_res.dsc = NULL;
-	if (prev_odm_pipe->next_odm_pipe && prev_odm_pipe->next_odm_pipe != next_odm_pipe) {
-		next_odm_pipe->next_odm_pipe = prev_odm_pipe->next_odm_pipe;
-		next_odm_pipe->next_odm_pipe->prev_odm_pipe = next_odm_pipe;
-	}
-	if (prev_odm_pipe->top_pipe && prev_odm_pipe->top_pipe->next_odm_pipe) {
-		prev_odm_pipe->top_pipe->next_odm_pipe->bottom_pipe = next_odm_pipe;
-		next_odm_pipe->top_pipe = prev_odm_pipe->top_pipe->next_odm_pipe;
-	}
-	if (prev_odm_pipe->bottom_pipe && prev_odm_pipe->bottom_pipe->next_odm_pipe) {
-		prev_odm_pipe->bottom_pipe->next_odm_pipe->top_pipe = next_odm_pipe;
-		next_odm_pipe->bottom_pipe = prev_odm_pipe->bottom_pipe->next_odm_pipe;
-	}
-	prev_odm_pipe->next_odm_pipe = next_odm_pipe;
-	next_odm_pipe->prev_odm_pipe = prev_odm_pipe;
-
-	if (prev_odm_pipe->plane_state) {
-		struct scaler_data *sd = &prev_odm_pipe->plane_res.scl_data;
-		int new_width;
-
-		/* HACTIVE halved for odm combine */
-		sd->h_active /= 2;
-		/* Calculate new vp and recout for left pipe */
-		/* Need at least 16 pixels width per side */
-		if (sd->recout.x + 16 >= sd->h_active)
-			return false;
-		new_width = sd->h_active - sd->recout.x;
-		sd->viewport.width -= dc_fixpt_floor(dc_fixpt_mul_int(
-				sd->ratios.horz, sd->recout.width - new_width));
-		sd->viewport_c.width -= dc_fixpt_floor(dc_fixpt_mul_int(
-				sd->ratios.horz_c, sd->recout.width - new_width));
-		sd->recout.width = new_width;
-
-		/* Calculate new vp and recout for right pipe */
-		sd = &next_odm_pipe->plane_res.scl_data;
-		/* HACTIVE halved for odm combine */
-		sd->h_active /= 2;
-		/* Need at least 16 pixels width per side */
-		if (new_width <= 16)
-			return false;
-		new_width = sd->recout.width + sd->recout.x - sd->h_active;
-		sd->viewport.width -= dc_fixpt_floor(dc_fixpt_mul_int(
-				sd->ratios.horz, sd->recout.width - new_width));
-		sd->viewport_c.width -= dc_fixpt_floor(dc_fixpt_mul_int(
-				sd->ratios.horz_c, sd->recout.width - new_width));
-		sd->recout.width = new_width;
-		sd->viewport.x += dc_fixpt_floor(dc_fixpt_mul_int(
-				sd->ratios.horz, sd->h_active - sd->recout.x));
-		sd->viewport_c.x += dc_fixpt_floor(dc_fixpt_mul_int(
-				sd->ratios.horz_c, sd->h_active - sd->recout.x));
-		sd->recout.x = 0;
-	}
-	if (!next_odm_pipe->top_pipe)
-		next_odm_pipe->stream_res.opp = pool->opps[next_odm_pipe->pipe_idx];
-	else
-		next_odm_pipe->stream_res.opp = next_odm_pipe->top_pipe->stream_res.opp;
-	if (next_odm_pipe->stream->timing.flags.DSC == 1 && !next_odm_pipe->top_pipe) {
-		dcn20_acquire_dsc(dc, res_ctx, &next_odm_pipe->stream_res.dsc, next_odm_pipe->pipe_idx);
-		ASSERT(next_odm_pipe->stream_res.dsc);
-		if (next_odm_pipe->stream_res.dsc == NULL)
-			return false;
-	}
-
-	return true;
-}
-
-void dcn20_split_stream_for_mpc(
-		struct resource_context *res_ctx,
-		const struct resource_pool *pool,
-		struct pipe_ctx *primary_pipe,
-		struct pipe_ctx *secondary_pipe)
-{
-	int pipe_idx = secondary_pipe->pipe_idx;
-	struct pipe_ctx *sec_bot_pipe = secondary_pipe->bottom_pipe;
-
-	*secondary_pipe = *primary_pipe;
-	secondary_pipe->bottom_pipe = sec_bot_pipe;
-
-	secondary_pipe->pipe_idx = pipe_idx;
-	secondary_pipe->plane_res.mi = pool->mis[secondary_pipe->pipe_idx];
-	secondary_pipe->plane_res.hubp = pool->hubps[secondary_pipe->pipe_idx];
-	secondary_pipe->plane_res.ipp = pool->ipps[secondary_pipe->pipe_idx];
-	secondary_pipe->plane_res.xfm = pool->transforms[secondary_pipe->pipe_idx];
-	secondary_pipe->plane_res.dpp = pool->dpps[secondary_pipe->pipe_idx];
-	secondary_pipe->plane_res.mpcc_inst = pool->dpps[secondary_pipe->pipe_idx]->inst;
-	secondary_pipe->stream_res.dsc = NULL;
-	if (primary_pipe->bottom_pipe && primary_pipe->bottom_pipe != secondary_pipe) {
-		ASSERT(!secondary_pipe->bottom_pipe);
-		secondary_pipe->bottom_pipe = primary_pipe->bottom_pipe;
-		secondary_pipe->bottom_pipe->top_pipe = secondary_pipe;
-	}
-	primary_pipe->bottom_pipe = secondary_pipe;
-	secondary_pipe->top_pipe = primary_pipe;
-
-	ASSERT(primary_pipe->plane_state);
-}
-
-int dcn20_populate_dml_pipes_from_context(
-		struct dc *dc,
-		struct dc_state *context,
-		display_e2e_pipe_params_st *pipes,
-		bool fast_validate)
-{
-	int pipe_cnt, i;
-	bool synchronized_vblank = true;
-	struct resource_context *res_ctx = &context->res_ctx;
-
-	for (i = 0, pipe_cnt = -1; i < dc->res_pool->pipe_count; i++) {
-		if (!res_ctx->pipe_ctx[i].stream)
-			continue;
-
-		if (pipe_cnt < 0) {
-			pipe_cnt = i;
-			continue;
-		}
-
-		if (res_ctx->pipe_ctx[pipe_cnt].stream == res_ctx->pipe_ctx[i].stream)
-			continue;
-
-		if (dc->debug.disable_timing_sync ||
-			(!resource_are_streams_timing_synchronizable(
-				res_ctx->pipe_ctx[pipe_cnt].stream,
-				res_ctx->pipe_ctx[i].stream) &&
-			!resource_are_vblanks_synchronizable(
-				res_ctx->pipe_ctx[pipe_cnt].stream,
-				res_ctx->pipe_ctx[i].stream))) {
-			synchronized_vblank = false;
-			break;
-		}
-	}
-
-	for (i = 0, pipe_cnt = 0; i < dc->res_pool->pipe_count; i++) {
-		struct dc_crtc_timing *timing = &res_ctx->pipe_ctx[i].stream->timing;
-		unsigned int v_total;
-		unsigned int front_porch;
-		int output_bpc;
-		struct audio_check aud_check = {0};
-
-		if (!res_ctx->pipe_ctx[i].stream)
-			continue;
-
-		v_total = timing->v_total;
-		front_porch = timing->v_front_porch;
-
-		/* todo:
-		pipes[pipe_cnt].pipe.src.dynamic_metadata_enable = 0;
-		pipes[pipe_cnt].pipe.src.dcc = 0;
-		pipes[pipe_cnt].pipe.src.vm = 0;*/
-
-		pipes[pipe_cnt].clks_cfg.refclk_mhz = dc->res_pool->ref_clocks.dchub_ref_clock_inKhz / 1000.0;
-
-		pipes[pipe_cnt].dout.dsc_enable = res_ctx->pipe_ctx[i].stream->timing.flags.DSC;
-		/* todo: rotation?*/
-		pipes[pipe_cnt].dout.dsc_slices = res_ctx->pipe_ctx[i].stream->timing.dsc_cfg.num_slices_h;
-		if (res_ctx->pipe_ctx[i].stream->use_dynamic_meta) {
-			pipes[pipe_cnt].pipe.src.dynamic_metadata_enable = true;
-			/* 1/2 vblank */
-			pipes[pipe_cnt].pipe.src.dynamic_metadata_lines_before_active =
-				(v_total - timing->v_addressable
-					- timing->v_border_top - timing->v_border_bottom) / 2;
-			/* 36 bytes dp, 32 hdmi */
-			pipes[pipe_cnt].pipe.src.dynamic_metadata_xmit_bytes =
-				dc_is_dp_signal(res_ctx->pipe_ctx[i].stream->signal) ? 36 : 32;
-		}
-		pipes[pipe_cnt].pipe.src.dcc = false;
-		pipes[pipe_cnt].pipe.src.dcc_rate = 1;
-		pipes[pipe_cnt].pipe.dest.synchronized_vblank_all_planes = synchronized_vblank;
-		pipes[pipe_cnt].pipe.dest.hblank_start = timing->h_total - timing->h_front_porch;
-		pipes[pipe_cnt].pipe.dest.hblank_end = pipes[pipe_cnt].pipe.dest.hblank_start
-				- timing->h_addressable
-				- timing->h_border_left
-				- timing->h_border_right;
-		pipes[pipe_cnt].pipe.dest.vblank_start = v_total - front_porch;
-		pipes[pipe_cnt].pipe.dest.vblank_end = pipes[pipe_cnt].pipe.dest.vblank_start
-				- timing->v_addressable
-				- timing->v_border_top
-				- timing->v_border_bottom;
-		pipes[pipe_cnt].pipe.dest.htotal = timing->h_total;
-		pipes[pipe_cnt].pipe.dest.vtotal = v_total;
-		pipes[pipe_cnt].pipe.dest.hactive =
-			timing->h_addressable + timing->h_border_left + timing->h_border_right;
-		pipes[pipe_cnt].pipe.dest.vactive =
-			timing->v_addressable + timing->v_border_top + timing->v_border_bottom;
-		pipes[pipe_cnt].pipe.dest.interlaced = timing->flags.INTERLACE;
-		pipes[pipe_cnt].pipe.dest.pixel_rate_mhz = timing->pix_clk_100hz/10000.0;
-		if (timing->timing_3d_format == TIMING_3D_FORMAT_HW_FRAME_PACKING)
-			pipes[pipe_cnt].pipe.dest.pixel_rate_mhz *= 2;
-		pipes[pipe_cnt].pipe.dest.otg_inst = res_ctx->pipe_ctx[i].stream_res.tg->inst;
-		pipes[pipe_cnt].dout.dp_lanes = 4;
-		pipes[pipe_cnt].dout.is_virtual = 0;
-		pipes[pipe_cnt].pipe.dest.vtotal_min = res_ctx->pipe_ctx[i].stream->adjust.v_total_min;
-		pipes[pipe_cnt].pipe.dest.vtotal_max = res_ctx->pipe_ctx[i].stream->adjust.v_total_max;
-		switch (get_num_odm_splits(&res_ctx->pipe_ctx[i])) {
-		case 1:
-			pipes[pipe_cnt].pipe.dest.odm_combine = dm_odm_combine_mode_2to1;
-			break;
-		case 3:
-			pipes[pipe_cnt].pipe.dest.odm_combine = dm_odm_combine_mode_4to1;
-			break;
-		default:
-			pipes[pipe_cnt].pipe.dest.odm_combine = dm_odm_combine_mode_disabled;
-		}
-		pipes[pipe_cnt].pipe.src.hsplit_grp = res_ctx->pipe_ctx[i].pipe_idx;
-		if (res_ctx->pipe_ctx[i].top_pipe && res_ctx->pipe_ctx[i].top_pipe->plane_state
-				== res_ctx->pipe_ctx[i].plane_state) {
-			struct pipe_ctx *first_pipe = res_ctx->pipe_ctx[i].top_pipe;
-			int split_idx = 0;
-
-			while (first_pipe->top_pipe && first_pipe->top_pipe->plane_state
-					== res_ctx->pipe_ctx[i].plane_state) {
-				first_pipe = first_pipe->top_pipe;
-				split_idx++;
-			}
-			/* Treat 4to1 mpc combine as an mpo of 2 2-to-1 combines */
-			if (split_idx == 0)
-				pipes[pipe_cnt].pipe.src.hsplit_grp = first_pipe->pipe_idx;
-			else if (split_idx == 1)
-				pipes[pipe_cnt].pipe.src.hsplit_grp = res_ctx->pipe_ctx[i].pipe_idx;
-			else if (split_idx == 2)
-				pipes[pipe_cnt].pipe.src.hsplit_grp = res_ctx->pipe_ctx[i].top_pipe->pipe_idx;
-		} else if (res_ctx->pipe_ctx[i].prev_odm_pipe) {
-			struct pipe_ctx *first_pipe = res_ctx->pipe_ctx[i].prev_odm_pipe;
-
-			while (first_pipe->prev_odm_pipe)
-				first_pipe = first_pipe->prev_odm_pipe;
-			pipes[pipe_cnt].pipe.src.hsplit_grp = first_pipe->pipe_idx;
-		}
-
-		switch (res_ctx->pipe_ctx[i].stream->signal) {
-		case SIGNAL_TYPE_DISPLAY_PORT_MST:
-		case SIGNAL_TYPE_DISPLAY_PORT:
-			pipes[pipe_cnt].dout.output_type = dm_dp;
-			break;
-		case SIGNAL_TYPE_EDP:
-			pipes[pipe_cnt].dout.output_type = dm_edp;
-			break;
-		case SIGNAL_TYPE_HDMI_TYPE_A:
-		case SIGNAL_TYPE_DVI_SINGLE_LINK:
-		case SIGNAL_TYPE_DVI_DUAL_LINK:
-			pipes[pipe_cnt].dout.output_type = dm_hdmi;
-			break;
-		default:
-			/* In case there is no signal, set dp with 4 lanes to allow max config */
-			pipes[pipe_cnt].dout.is_virtual = 1;
-			pipes[pipe_cnt].dout.output_type = dm_dp;
-			pipes[pipe_cnt].dout.dp_lanes = 4;
-		}
-
-		switch (res_ctx->pipe_ctx[i].stream->timing.display_color_depth) {
-		case COLOR_DEPTH_666:
-			output_bpc = 6;
-			break;
-		case COLOR_DEPTH_888:
-			output_bpc = 8;
-			break;
-		case COLOR_DEPTH_101010:
-			output_bpc = 10;
-			break;
-		case COLOR_DEPTH_121212:
-			output_bpc = 12;
-			break;
-		case COLOR_DEPTH_141414:
-			output_bpc = 14;
-			break;
-		case COLOR_DEPTH_161616:
-			output_bpc = 16;
-			break;
-		case COLOR_DEPTH_999:
-			output_bpc = 9;
-			break;
-		case COLOR_DEPTH_111111:
-			output_bpc = 11;
-			break;
-		default:
-			output_bpc = 8;
-			break;
-		}
-
-		switch (res_ctx->pipe_ctx[i].stream->timing.pixel_encoding) {
-		case PIXEL_ENCODING_RGB:
-		case PIXEL_ENCODING_YCBCR444:
-			pipes[pipe_cnt].dout.output_format = dm_444;
-			pipes[pipe_cnt].dout.output_bpp = output_bpc * 3;
-			break;
-		case PIXEL_ENCODING_YCBCR420:
-			pipes[pipe_cnt].dout.output_format = dm_420;
-			pipes[pipe_cnt].dout.output_bpp = (output_bpc * 3.0) / 2;
-			break;
-		case PIXEL_ENCODING_YCBCR422:
-			if (res_ctx->pipe_ctx[i].stream->timing.flags.DSC &&
-			    !res_ctx->pipe_ctx[i].stream->timing.dsc_cfg.ycbcr422_simple)
-				pipes[pipe_cnt].dout.output_format = dm_n422;
-			else
-				pipes[pipe_cnt].dout.output_format = dm_s422;
-			pipes[pipe_cnt].dout.output_bpp = output_bpc * 2;
-			break;
-		default:
-			pipes[pipe_cnt].dout.output_format = dm_444;
-			pipes[pipe_cnt].dout.output_bpp = output_bpc * 3;
-		}
-
-		if (res_ctx->pipe_ctx[i].stream->timing.flags.DSC)
-			pipes[pipe_cnt].dout.output_bpp = res_ctx->pipe_ctx[i].stream->timing.dsc_cfg.bits_per_pixel / 16.0;
-
-		/* todo: default max for now, until there is logic reflecting this in dc*/
-		pipes[pipe_cnt].dout.dsc_input_bpc = 12;
-		/*fill up the audio sample rate (unit in kHz)*/
-		get_audio_check(&res_ctx->pipe_ctx[i].stream->audio_info, &aud_check);
-		pipes[pipe_cnt].dout.max_audio_sample_rate = aud_check.max_audiosample_rate / 1000;
-		/*
-		 * For graphic plane, cursor number is 1, nv12 is 0
-		 * bw calculations due to cursor on/off
-		 */
-		if (res_ctx->pipe_ctx[i].plane_state &&
-				res_ctx->pipe_ctx[i].plane_state->address.type == PLN_ADDR_TYPE_VIDEO_PROGRESSIVE)
-			pipes[pipe_cnt].pipe.src.num_cursors = 0;
-		else
-			pipes[pipe_cnt].pipe.src.num_cursors = dc->dml.ip.number_of_cursors;
-
-		pipes[pipe_cnt].pipe.src.cur0_src_width = 256;
-		pipes[pipe_cnt].pipe.src.cur0_bpp = dm_cur_32bit;
-
-		if (!res_ctx->pipe_ctx[i].plane_state) {
-			pipes[pipe_cnt].pipe.src.is_hsplit = pipes[pipe_cnt].pipe.dest.odm_combine != dm_odm_combine_mode_disabled;
-			pipes[pipe_cnt].pipe.src.source_scan = dm_horz;
-			pipes[pipe_cnt].pipe.src.sw_mode = dm_sw_4kb_s;
-			pipes[pipe_cnt].pipe.src.macro_tile_size = dm_64k_tile;
-			pipes[pipe_cnt].pipe.src.viewport_width = timing->h_addressable;
-			if (pipes[pipe_cnt].pipe.src.viewport_width > 1920)
-				pipes[pipe_cnt].pipe.src.viewport_width = 1920;
-			pipes[pipe_cnt].pipe.src.viewport_height = timing->v_addressable;
-			if (pipes[pipe_cnt].pipe.src.viewport_height > 1080)
-				pipes[pipe_cnt].pipe.src.viewport_height = 1080;
-			pipes[pipe_cnt].pipe.src.surface_height_y = pipes[pipe_cnt].pipe.src.viewport_height;
-			pipes[pipe_cnt].pipe.src.surface_width_y = pipes[pipe_cnt].pipe.src.viewport_width;
-			pipes[pipe_cnt].pipe.src.surface_height_c = pipes[pipe_cnt].pipe.src.viewport_height;
-			pipes[pipe_cnt].pipe.src.surface_width_c = pipes[pipe_cnt].pipe.src.viewport_width;
-			pipes[pipe_cnt].pipe.src.data_pitch = ((pipes[pipe_cnt].pipe.src.viewport_width + 255) / 256) * 256;
-			pipes[pipe_cnt].pipe.src.source_format = dm_444_32;
-			pipes[pipe_cnt].pipe.dest.recout_width = pipes[pipe_cnt].pipe.src.viewport_width; /*vp_width/hratio*/
-			pipes[pipe_cnt].pipe.dest.recout_height = pipes[pipe_cnt].pipe.src.viewport_height; /*vp_height/vratio*/
-			pipes[pipe_cnt].pipe.dest.full_recout_width = pipes[pipe_cnt].pipe.dest.recout_width;  /*when is_hsplit != 1*/
-			pipes[pipe_cnt].pipe.dest.full_recout_height = pipes[pipe_cnt].pipe.dest.recout_height; /*when is_hsplit != 1*/
-			pipes[pipe_cnt].pipe.scale_ratio_depth.lb_depth = dm_lb_16;
-			pipes[pipe_cnt].pipe.scale_ratio_depth.hscl_ratio = 1.0;
-			pipes[pipe_cnt].pipe.scale_ratio_depth.vscl_ratio = 1.0;
-			pipes[pipe_cnt].pipe.scale_ratio_depth.scl_enable = 0; /*Lb only or Full scl*/
-			pipes[pipe_cnt].pipe.scale_taps.htaps = 1;
-			pipes[pipe_cnt].pipe.scale_taps.vtaps = 1;
-			pipes[pipe_cnt].pipe.dest.vtotal_min = v_total;
-			pipes[pipe_cnt].pipe.dest.vtotal_max = v_total;
-
-			if (pipes[pipe_cnt].pipe.dest.odm_combine == dm_odm_combine_mode_2to1) {
-				pipes[pipe_cnt].pipe.src.viewport_width /= 2;
-				pipes[pipe_cnt].pipe.dest.recout_width /= 2;
-			} else if (pipes[pipe_cnt].pipe.dest.odm_combine == dm_odm_combine_mode_4to1) {
-				pipes[pipe_cnt].pipe.src.viewport_width /= 4;
-				pipes[pipe_cnt].pipe.dest.recout_width /= 4;
-			}
-		} else {
-			struct dc_plane_state *pln = res_ctx->pipe_ctx[i].plane_state;
-			struct scaler_data *scl = &res_ctx->pipe_ctx[i].plane_res.scl_data;
-
-			pipes[pipe_cnt].pipe.src.immediate_flip = pln->flip_immediate;
-			pipes[pipe_cnt].pipe.src.is_hsplit = (res_ctx->pipe_ctx[i].bottom_pipe && res_ctx->pipe_ctx[i].bottom_pipe->plane_state == pln)
-					|| (res_ctx->pipe_ctx[i].top_pipe && res_ctx->pipe_ctx[i].top_pipe->plane_state == pln)
-					|| pipes[pipe_cnt].pipe.dest.odm_combine != dm_odm_combine_mode_disabled;
-
-			/* stereo is not split */
-			if (pln->stereo_format == PLANE_STEREO_FORMAT_SIDE_BY_SIDE ||
-			    pln->stereo_format == PLANE_STEREO_FORMAT_TOP_AND_BOTTOM) {
-				pipes[pipe_cnt].pipe.src.is_hsplit = false;
-				pipes[pipe_cnt].pipe.src.hsplit_grp = res_ctx->pipe_ctx[i].pipe_idx;
-			}
-
-			pipes[pipe_cnt].pipe.src.source_scan = pln->rotation == ROTATION_ANGLE_90
-					|| pln->rotation == ROTATION_ANGLE_270 ? dm_vert : dm_horz;
-			pipes[pipe_cnt].pipe.src.viewport_y_y = scl->viewport.y;
-			pipes[pipe_cnt].pipe.src.viewport_y_c = scl->viewport_c.y;
-			pipes[pipe_cnt].pipe.src.viewport_width = scl->viewport.width;
-			pipes[pipe_cnt].pipe.src.viewport_width_c = scl->viewport_c.width;
-			pipes[pipe_cnt].pipe.src.viewport_height = scl->viewport.height;
-			pipes[pipe_cnt].pipe.src.viewport_height_c = scl->viewport_c.height;
-			pipes[pipe_cnt].pipe.src.viewport_width_max = pln->src_rect.width;
-			pipes[pipe_cnt].pipe.src.viewport_height_max = pln->src_rect.height;
-			pipes[pipe_cnt].pipe.src.surface_width_y = pln->plane_size.surface_size.width;
-			pipes[pipe_cnt].pipe.src.surface_height_y = pln->plane_size.surface_size.height;
-			pipes[pipe_cnt].pipe.src.surface_width_c = pln->plane_size.chroma_size.width;
-			pipes[pipe_cnt].pipe.src.surface_height_c = pln->plane_size.chroma_size.height;
-			if (pln->format == SURFACE_PIXEL_FORMAT_GRPH_RGBE_ALPHA
-					|| pln->format >= SURFACE_PIXEL_FORMAT_VIDEO_BEGIN) {
-				pipes[pipe_cnt].pipe.src.data_pitch = pln->plane_size.surface_pitch;
-				pipes[pipe_cnt].pipe.src.data_pitch_c = pln->plane_size.chroma_pitch;
-				pipes[pipe_cnt].pipe.src.meta_pitch = pln->dcc.meta_pitch;
-				pipes[pipe_cnt].pipe.src.meta_pitch_c = pln->dcc.meta_pitch_c;
-			} else {
-				pipes[pipe_cnt].pipe.src.data_pitch = pln->plane_size.surface_pitch;
-				pipes[pipe_cnt].pipe.src.meta_pitch = pln->dcc.meta_pitch;
-			}
-			pipes[pipe_cnt].pipe.src.dcc = pln->dcc.enable;
-			pipes[pipe_cnt].pipe.dest.recout_width = scl->recout.width;
-			pipes[pipe_cnt].pipe.dest.recout_height = scl->recout.height;
-			pipes[pipe_cnt].pipe.dest.full_recout_height = scl->recout.height;
-			pipes[pipe_cnt].pipe.dest.full_recout_width = scl->recout.width;
-			if (pipes[pipe_cnt].pipe.dest.odm_combine == dm_odm_combine_mode_2to1)
-				pipes[pipe_cnt].pipe.dest.full_recout_width *= 2;
-			else if (pipes[pipe_cnt].pipe.dest.odm_combine == dm_odm_combine_mode_4to1)
-				pipes[pipe_cnt].pipe.dest.full_recout_width *= 4;
-			else {
-				struct pipe_ctx *split_pipe = res_ctx->pipe_ctx[i].bottom_pipe;
-
-				while (split_pipe && split_pipe->plane_state == pln) {
-					pipes[pipe_cnt].pipe.dest.full_recout_width += split_pipe->plane_res.scl_data.recout.width;
-					split_pipe = split_pipe->bottom_pipe;
-				}
-				split_pipe = res_ctx->pipe_ctx[i].top_pipe;
-				while (split_pipe && split_pipe->plane_state == pln) {
-					pipes[pipe_cnt].pipe.dest.full_recout_width += split_pipe->plane_res.scl_data.recout.width;
-					split_pipe = split_pipe->top_pipe;
-				}
-			}
-
-			pipes[pipe_cnt].pipe.scale_ratio_depth.lb_depth = dm_lb_16;
-			pipes[pipe_cnt].pipe.scale_ratio_depth.hscl_ratio = (double) scl->ratios.horz.value / (1ULL<<32);
-			pipes[pipe_cnt].pipe.scale_ratio_depth.hscl_ratio_c = (double) scl->ratios.horz_c.value / (1ULL<<32);
-			pipes[pipe_cnt].pipe.scale_ratio_depth.vscl_ratio = (double) scl->ratios.vert.value / (1ULL<<32);
-			pipes[pipe_cnt].pipe.scale_ratio_depth.vscl_ratio_c = (double) scl->ratios.vert_c.value / (1ULL<<32);
-			pipes[pipe_cnt].pipe.scale_ratio_depth.scl_enable =
-					scl->ratios.vert.value != dc_fixpt_one.value
-					|| scl->ratios.horz.value != dc_fixpt_one.value
-					|| scl->ratios.vert_c.value != dc_fixpt_one.value
-					|| scl->ratios.horz_c.value != dc_fixpt_one.value /*Lb only or Full scl*/
-					|| dc->debug.always_scale; /*support always scale*/
-			pipes[pipe_cnt].pipe.scale_taps.htaps = scl->taps.h_taps;
-			pipes[pipe_cnt].pipe.scale_taps.htaps_c = scl->taps.h_taps_c;
-			pipes[pipe_cnt].pipe.scale_taps.vtaps = scl->taps.v_taps;
-			pipes[pipe_cnt].pipe.scale_taps.vtaps_c = scl->taps.v_taps_c;
-
-			pipes[pipe_cnt].pipe.src.macro_tile_size =
-					swizzle_mode_to_macro_tile_size(pln->tiling_info.gfx9.swizzle);
-			swizzle_to_dml_params(pln->tiling_info.gfx9.swizzle,
-					&pipes[pipe_cnt].pipe.src.sw_mode);
-
-			switch (pln->format) {
-			case SURFACE_PIXEL_FORMAT_VIDEO_420_YCbCr:
-			case SURFACE_PIXEL_FORMAT_VIDEO_420_YCrCb:
-				pipes[pipe_cnt].pipe.src.source_format = dm_420_8;
-				break;
-			case SURFACE_PIXEL_FORMAT_VIDEO_420_10bpc_YCbCr:
-			case SURFACE_PIXEL_FORMAT_VIDEO_420_10bpc_YCrCb:
-				pipes[pipe_cnt].pipe.src.source_format = dm_420_10;
-				break;
-			case SURFACE_PIXEL_FORMAT_GRPH_ARGB16161616:
-			case SURFACE_PIXEL_FORMAT_GRPH_ABGR16161616:
-			case SURFACE_PIXEL_FORMAT_GRPH_ARGB16161616F:
-			case SURFACE_PIXEL_FORMAT_GRPH_ABGR16161616F:
-				pipes[pipe_cnt].pipe.src.source_format = dm_444_64;
-				break;
-			case SURFACE_PIXEL_FORMAT_GRPH_ARGB1555:
-			case SURFACE_PIXEL_FORMAT_GRPH_RGB565:
-				pipes[pipe_cnt].pipe.src.source_format = dm_444_16;
-				break;
-			case SURFACE_PIXEL_FORMAT_GRPH_PALETA_256_COLORS:
-				pipes[pipe_cnt].pipe.src.source_format = dm_444_8;
-				break;
-			case SURFACE_PIXEL_FORMAT_GRPH_RGBE_ALPHA:
-				pipes[pipe_cnt].pipe.src.source_format = dm_rgbe_alpha;
-				break;
-			default:
-				pipes[pipe_cnt].pipe.src.source_format = dm_444_32;
-				break;
-			}
-		}
-
-		pipe_cnt++;
-	}
-
-	/* populate writeback information */
-	DC_FP_START();
-	dc->res_pool->funcs->populate_dml_writeback_from_context(dc, res_ctx, pipes);
-	DC_FP_END();
-
-	return pipe_cnt;
-}
-
-unsigned int dcn20_calc_max_scaled_time(
-		unsigned int time_per_pixel,
-		enum mmhubbub_wbif_mode mode,
-		unsigned int urgent_watermark)
-{
-	unsigned int time_per_byte = 0;
-	unsigned int total_y_free_entry = 0x200; /* two memory piece for luma */
-	unsigned int total_c_free_entry = 0x140; /* two memory piece for chroma */
-	unsigned int small_free_entry, max_free_entry;
-	unsigned int buf_lh_capability;
-	unsigned int max_scaled_time;
-
-	if (mode == PACKED_444) /* packed mode */
-		time_per_byte = time_per_pixel/4;
-	else if (mode == PLANAR_420_8BPC)
-		time_per_byte  = time_per_pixel;
-	else if (mode == PLANAR_420_10BPC) /* p010 */
-		time_per_byte  = time_per_pixel * 819/1024;
-
-	if (time_per_byte == 0)
-		time_per_byte = 1;
-
-	small_free_entry  = (total_y_free_entry > total_c_free_entry) ? total_c_free_entry : total_y_free_entry;
-	max_free_entry    = (mode == PACKED_444) ? total_y_free_entry + total_c_free_entry : small_free_entry;
-	buf_lh_capability = max_free_entry*time_per_byte*32/16; /* there is 4bit fraction */
-	max_scaled_time   = buf_lh_capability - urgent_watermark;
-	return max_scaled_time;
-}
-
-void dcn20_set_mcif_arb_params(
-		struct dc *dc,
-		struct dc_state *context,
-		display_e2e_pipe_params_st *pipes,
-		int pipe_cnt)
-{
-	enum mmhubbub_wbif_mode wbif_mode;
-	struct mcif_arb_params *wb_arb_params;
-	int i, j, k, dwb_pipe;
-
-	/* Writeback MCIF_WB arbitration parameters */
-	dwb_pipe = 0;
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-
-		if (!context->res_ctx.pipe_ctx[i].stream)
-			continue;
-
-		for (j = 0; j < MAX_DWB_PIPES; j++) {
-			if (context->res_ctx.pipe_ctx[i].stream->writeback_info[j].wb_enabled == false)
-				continue;
-
-			//wb_arb_params = &context->res_ctx.pipe_ctx[i].stream->writeback_info[j].mcif_arb_params;
-			wb_arb_params = &context->bw_ctx.bw.dcn.bw_writeback.mcif_wb_arb[dwb_pipe];
-
-			if (context->res_ctx.pipe_ctx[i].stream->writeback_info[j].dwb_params.out_format == dwb_scaler_mode_yuv420) {
-				if (context->res_ctx.pipe_ctx[i].stream->writeback_info[j].dwb_params.output_depth == DWB_OUTPUT_PIXEL_DEPTH_8BPC)
-					wbif_mode = PLANAR_420_8BPC;
-				else
-					wbif_mode = PLANAR_420_10BPC;
-			} else
-				wbif_mode = PACKED_444;
-
-			for (k = 0; k < sizeof(wb_arb_params->cli_watermark)/sizeof(wb_arb_params->cli_watermark[0]); k++) {
-				wb_arb_params->cli_watermark[k] = get_wm_writeback_urgent(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-				wb_arb_params->pstate_watermark[k] = get_wm_writeback_dram_clock_change(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-			}
-			wb_arb_params->time_per_pixel = 16.0 * 1000 / (context->res_ctx.pipe_ctx[i].stream->phy_pix_clk / 1000); /* 4 bit fraction, ms */
-			wb_arb_params->slice_lines = 32;
-			wb_arb_params->arbitration_slice = 2;
-			wb_arb_params->max_scaled_time = dcn20_calc_max_scaled_time(wb_arb_params->time_per_pixel,
-				wbif_mode,
-				wb_arb_params->cli_watermark[0]); /* assume 4 watermark sets have the same value */
-
-			dwb_pipe++;
-
-			if (dwb_pipe >= MAX_DWB_PIPES)
-				return;
-		}
-		if (dwb_pipe >= MAX_DWB_PIPES)
-			return;
-	}
-}
-
-bool dcn20_validate_dsc(struct dc *dc, struct dc_state *new_ctx)
-{
-	int i;
-
-	/* Validate DSC config, dsc count validation is already done */
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		struct pipe_ctx *pipe_ctx = &new_ctx->res_ctx.pipe_ctx[i];
-		struct dc_stream_state *stream = pipe_ctx->stream;
-		struct dsc_config dsc_cfg;
-		struct pipe_ctx *odm_pipe;
-		int opp_cnt = 1;
-
-		for (odm_pipe = pipe_ctx->next_odm_pipe; odm_pipe; odm_pipe = odm_pipe->next_odm_pipe)
-			opp_cnt++;
-
-		/* Only need to validate top pipe */
-		if (pipe_ctx->top_pipe || pipe_ctx->prev_odm_pipe || !stream || !stream->timing.flags.DSC)
-			continue;
-
-		dsc_cfg.pic_width = (stream->timing.h_addressable + stream->timing.h_border_left
-				+ stream->timing.h_border_right) / opp_cnt;
-		dsc_cfg.pic_height = stream->timing.v_addressable + stream->timing.v_border_top
-				+ stream->timing.v_border_bottom;
-		dsc_cfg.pixel_encoding = stream->timing.pixel_encoding;
-		dsc_cfg.color_depth = stream->timing.display_color_depth;
-		dsc_cfg.is_odm = pipe_ctx->next_odm_pipe ? true : false;
-		dsc_cfg.dc_dsc_cfg = stream->timing.dsc_cfg;
-		dsc_cfg.dc_dsc_cfg.num_slices_h /= opp_cnt;
-
-		if (!pipe_ctx->stream_res.dsc->funcs->dsc_validate_stream(pipe_ctx->stream_res.dsc, &dsc_cfg))
-			return false;
-	}
-	return true;
-}
-
-struct pipe_ctx *dcn20_find_secondary_pipe(struct dc *dc,
-		struct resource_context *res_ctx,
-		const struct resource_pool *pool,
-		const struct pipe_ctx *primary_pipe)
-{
-	struct pipe_ctx *secondary_pipe = NULL;
-
-	if (dc && primary_pipe) {
-		int j;
-		int preferred_pipe_idx = 0;
-
-		/* first check the prev dc state:
-		 * if this primary pipe has a bottom pipe in prev. state
-		 * and if the bottom pipe is still available (which it should be),
-		 * pick that pipe as secondary
-		 * Same logic applies for ODM pipes
-		 */
-		if (dc->current_state->res_ctx.pipe_ctx[primary_pipe->pipe_idx].next_odm_pipe) {
-			preferred_pipe_idx = dc->current_state->res_ctx.pipe_ctx[primary_pipe->pipe_idx].next_odm_pipe->pipe_idx;
-			if (res_ctx->pipe_ctx[preferred_pipe_idx].stream == NULL) {
-				secondary_pipe = &res_ctx->pipe_ctx[preferred_pipe_idx];
-				secondary_pipe->pipe_idx = preferred_pipe_idx;
-			}
-		}
-		if (secondary_pipe == NULL &&
-				dc->current_state->res_ctx.pipe_ctx[primary_pipe->pipe_idx].bottom_pipe) {
-			preferred_pipe_idx = dc->current_state->res_ctx.pipe_ctx[primary_pipe->pipe_idx].bottom_pipe->pipe_idx;
-			if (res_ctx->pipe_ctx[preferred_pipe_idx].stream == NULL) {
-				secondary_pipe = &res_ctx->pipe_ctx[preferred_pipe_idx];
-				secondary_pipe->pipe_idx = preferred_pipe_idx;
-			}
-		}
-
-		/*
-		 * if this primary pipe does not have a bottom pipe in prev. state
-		 * start backward and find a pipe that did not used to be a bottom pipe in
-		 * prev. dc state. This way we make sure we keep the same assignment as
-		 * last state and will not have to reprogram every pipe
-		 */
-		if (secondary_pipe == NULL) {
-			for (j = dc->res_pool->pipe_count - 1; j >= 0; j--) {
-				if (dc->current_state->res_ctx.pipe_ctx[j].top_pipe == NULL
-						&& dc->current_state->res_ctx.pipe_ctx[j].prev_odm_pipe == NULL) {
-					preferred_pipe_idx = j;
-
-					if (res_ctx->pipe_ctx[preferred_pipe_idx].stream == NULL) {
-						secondary_pipe = &res_ctx->pipe_ctx[preferred_pipe_idx];
-						secondary_pipe->pipe_idx = preferred_pipe_idx;
-						break;
-					}
-				}
-			}
-		}
-		/*
-		 * We should never hit this assert unless assignments are shuffled around
-		 * if this happens we will prob. hit a vsync tdr
-		 */
-		ASSERT(secondary_pipe);
-		/*
-		 * search backwards for the second pipe to keep pipe
-		 * assignment more consistent
-		 */
-		if (secondary_pipe == NULL) {
-			for (j = dc->res_pool->pipe_count - 1; j >= 0; j--) {
-				preferred_pipe_idx = j;
-
-				if (res_ctx->pipe_ctx[preferred_pipe_idx].stream == NULL) {
-					secondary_pipe = &res_ctx->pipe_ctx[preferred_pipe_idx];
-					secondary_pipe->pipe_idx = preferred_pipe_idx;
-					break;
-				}
-			}
-		}
-	}
-
-	return secondary_pipe;
-}
-
-void dcn20_merge_pipes_for_validate(
-		struct dc *dc,
-		struct dc_state *context)
-{
-	int i;
-
-	/* merge previously split odm pipes since mode support needs to make the decision */
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
-		struct pipe_ctx *odm_pipe = pipe->next_odm_pipe;
-
-		if (pipe->prev_odm_pipe)
-			continue;
-
-		pipe->next_odm_pipe = NULL;
-		while (odm_pipe) {
-			struct pipe_ctx *next_odm_pipe = odm_pipe->next_odm_pipe;
-
-			odm_pipe->plane_state = NULL;
-			odm_pipe->stream = NULL;
-			odm_pipe->top_pipe = NULL;
-			odm_pipe->bottom_pipe = NULL;
-			odm_pipe->prev_odm_pipe = NULL;
-			odm_pipe->next_odm_pipe = NULL;
-			if (odm_pipe->stream_res.dsc)
-				dcn20_release_dsc(&context->res_ctx, dc->res_pool, &odm_pipe->stream_res.dsc);
-			/* Clear plane_res and stream_res */
-			memset(&odm_pipe->plane_res, 0, sizeof(odm_pipe->plane_res));
-			memset(&odm_pipe->stream_res, 0, sizeof(odm_pipe->stream_res));
-			odm_pipe = next_odm_pipe;
-		}
-		if (pipe->plane_state)
-			resource_build_scaling_params(pipe);
-	}
-
-	/* merge previously mpc split pipes since mode support needs to make the decision */
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
-		struct pipe_ctx *hsplit_pipe = pipe->bottom_pipe;
-
-		if (!hsplit_pipe || hsplit_pipe->plane_state != pipe->plane_state)
-			continue;
-
-		pipe->bottom_pipe = hsplit_pipe->bottom_pipe;
-		if (hsplit_pipe->bottom_pipe)
-			hsplit_pipe->bottom_pipe->top_pipe = pipe;
-		hsplit_pipe->plane_state = NULL;
-		hsplit_pipe->stream = NULL;
-		hsplit_pipe->top_pipe = NULL;
-		hsplit_pipe->bottom_pipe = NULL;
-
-		/* Clear plane_res and stream_res */
-		memset(&hsplit_pipe->plane_res, 0, sizeof(hsplit_pipe->plane_res));
-		memset(&hsplit_pipe->stream_res, 0, sizeof(hsplit_pipe->stream_res));
-		if (pipe->plane_state)
-			resource_build_scaling_params(pipe);
-	}
-}
-
-int dcn20_validate_apply_pipe_split_flags(
-		struct dc *dc,
-		struct dc_state *context,
-		int vlevel,
-		int *split,
-		bool *merge)
-{
-	int i, pipe_idx, vlevel_split;
-	int plane_count = 0;
-	bool force_split = false;
-	bool avoid_split = dc->debug.pipe_split_policy == MPC_SPLIT_AVOID;
-	struct vba_vars_st *v = &context->bw_ctx.dml.vba;
-	int max_mpc_comb = v->maxMpcComb;
-
-	if (context->stream_count > 1) {
-		if (dc->debug.pipe_split_policy == MPC_SPLIT_AVOID_MULT_DISP)
-			avoid_split = true;
-	} else if (dc->debug.force_single_disp_pipe_split)
-			force_split = true;
-
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
-
-		/**
-		 * Workaround for avoiding pipe-split in cases where we'd split
-		 * planes that are too small, resulting in splits that aren't
-		 * valid for the scaler.
-		 */
-		if (pipe->plane_state &&
-		    (pipe->plane_state->dst_rect.width <= 16 ||
-		     pipe->plane_state->dst_rect.height <= 16 ||
-		     pipe->plane_state->src_rect.width <= 16 ||
-		     pipe->plane_state->src_rect.height <= 16))
-			avoid_split = true;
-
-		/* TODO: fix dc bugs and remove this split threshold thing */
-		if (pipe->stream && !pipe->prev_odm_pipe &&
-				(!pipe->top_pipe || pipe->top_pipe->plane_state != pipe->plane_state))
-			++plane_count;
-	}
-	if (plane_count > dc->res_pool->pipe_count / 2)
-		avoid_split = true;
-
-	/* W/A: Mode timing with borders may not work well with pipe split, avoid for this corner case */
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
-		struct dc_crtc_timing timing;
-
-		if (!pipe->stream)
-			continue;
-		else {
-			timing = pipe->stream->timing;
-			if (timing.h_border_left + timing.h_border_right
-					+ timing.v_border_top + timing.v_border_bottom > 0) {
-				avoid_split = true;
-				break;
-			}
-		}
-	}
-
-	/* Avoid split loop looks for lowest voltage level that allows most unsplit pipes possible */
-	if (avoid_split) {
-		for (i = 0, pipe_idx = 0; i < dc->res_pool->pipe_count; i++) {
-			if (!context->res_ctx.pipe_ctx[i].stream)
-				continue;
-
-			for (vlevel_split = vlevel; vlevel <= context->bw_ctx.dml.soc.num_states; vlevel++)
-				if (v->NoOfDPP[vlevel][0][pipe_idx] == 1 &&
-						v->ModeSupport[vlevel][0])
-					break;
-			/* Impossible to not split this pipe */
-			if (vlevel > context->bw_ctx.dml.soc.num_states)
-				vlevel = vlevel_split;
-			else
-				max_mpc_comb = 0;
-			pipe_idx++;
-		}
-		v->maxMpcComb = max_mpc_comb;
-	}
-
-	/* Split loop sets which pipe should be split based on dml outputs and dc flags */
-	for (i = 0, pipe_idx = 0; i < dc->res_pool->pipe_count; i++) {
-		struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
-		int pipe_plane = v->pipe_plane[pipe_idx];
-		bool split4mpc = context->stream_count == 1 && plane_count == 1
-				&& dc->config.enable_4to1MPC && dc->res_pool->pipe_count >= 4;
-
-		if (!context->res_ctx.pipe_ctx[i].stream)
-			continue;
-
-		if (split4mpc || v->NoOfDPP[vlevel][max_mpc_comb][pipe_plane] == 4)
-			split[i] = 4;
-		else if (force_split || v->NoOfDPP[vlevel][max_mpc_comb][pipe_plane] == 2)
-				split[i] = 2;
-
-		if ((pipe->stream->view_format ==
-				VIEW_3D_FORMAT_SIDE_BY_SIDE ||
-				pipe->stream->view_format ==
-				VIEW_3D_FORMAT_TOP_AND_BOTTOM) &&
-				(pipe->stream->timing.timing_3d_format ==
-				TIMING_3D_FORMAT_TOP_AND_BOTTOM ||
-				 pipe->stream->timing.timing_3d_format ==
-				TIMING_3D_FORMAT_SIDE_BY_SIDE))
-			split[i] = 2;
-		if (dc->debug.force_odm_combine & (1 << pipe->stream_res.tg->inst)) {
-			split[i] = 2;
-			v->ODMCombineEnablePerState[vlevel][pipe_plane] = dm_odm_combine_mode_2to1;
-		}
-		if (dc->debug.force_odm_combine_4to1 & (1 << pipe->stream_res.tg->inst)) {
-			split[i] = 4;
-			v->ODMCombineEnablePerState[vlevel][pipe_plane] = dm_odm_combine_mode_4to1;
-		}
-		/*420 format workaround*/
-		if (pipe->stream->timing.h_addressable > 7680 &&
-				pipe->stream->timing.pixel_encoding == PIXEL_ENCODING_YCBCR420) {
-			split[i] = 4;
-		}
-		v->ODMCombineEnabled[pipe_plane] =
-			v->ODMCombineEnablePerState[vlevel][pipe_plane];
-
-		if (v->ODMCombineEnabled[pipe_plane] == dm_odm_combine_mode_disabled) {
-			if (get_num_mpc_splits(pipe) == 1) {
-				/*If need split for mpc but 2 way split already*/
-				if (split[i] == 4)
-					split[i] = 2; /* 2 -> 4 MPC */
-				else if (split[i] == 2)
-					split[i] = 0; /* 2 -> 2 MPC */
-				else if (pipe->top_pipe && pipe->top_pipe->plane_state == pipe->plane_state)
-					merge[i] = true; /* 2 -> 1 MPC */
-			} else if (get_num_mpc_splits(pipe) == 3) {
-				/*If need split for mpc but 4 way split already*/
-				if (split[i] == 2 && ((pipe->top_pipe && !pipe->top_pipe->top_pipe)
-						|| !pipe->bottom_pipe)) {
-					merge[i] = true; /* 4 -> 2 MPC */
-				} else if (split[i] == 0 && pipe->top_pipe &&
-						pipe->top_pipe->plane_state == pipe->plane_state)
-					merge[i] = true; /* 4 -> 1 MPC */
-				split[i] = 0;
-			} else if (get_num_odm_splits(pipe)) {
-				/* ODM -> MPC transition */
-				if (pipe->prev_odm_pipe) {
-					split[i] = 0;
-					merge[i] = true;
-				}
-			}
-		} else {
-			if (get_num_odm_splits(pipe) == 1) {
-				/*If need split for odm but 2 way split already*/
-				if (split[i] == 4)
-					split[i] = 2; /* 2 -> 4 ODM */
-				else if (split[i] == 2)
-					split[i] = 0; /* 2 -> 2 ODM */
-				else if (pipe->prev_odm_pipe) {
-					ASSERT(0); /* NOT expected yet */
-					merge[i] = true; /* exit ODM */
-				}
-			} else if (get_num_odm_splits(pipe) == 3) {
-				/*If need split for odm but 4 way split already*/
-				if (split[i] == 2 && ((pipe->prev_odm_pipe && !pipe->prev_odm_pipe->prev_odm_pipe)
-						|| !pipe->next_odm_pipe)) {
-					ASSERT(0); /* NOT expected yet */
-					merge[i] = true; /* 4 -> 2 ODM */
-				} else if (split[i] == 0 && pipe->prev_odm_pipe) {
-					ASSERT(0); /* NOT expected yet */
-					merge[i] = true; /* exit ODM */
-				}
-				split[i] = 0;
-			} else if (get_num_mpc_splits(pipe)) {
-				/* MPC -> ODM transition */
-				ASSERT(0); /* NOT expected yet */
-				if (pipe->top_pipe && pipe->top_pipe->plane_state == pipe->plane_state) {
-					split[i] = 0;
-					merge[i] = true;
-				}
-			}
-		}
-
-		/* Adjust dppclk when split is forced, do not bother with dispclk */
-		if (split[i] != 0 && v->NoOfDPP[vlevel][max_mpc_comb][pipe_idx] == 1)
-			v->RequiredDPPCLK[vlevel][max_mpc_comb][pipe_idx] /= 2;
-		pipe_idx++;
-	}
-
-	return vlevel;
-}
-
-bool dcn20_fast_validate_bw(
-		struct dc *dc,
-		struct dc_state *context,
-		display_e2e_pipe_params_st *pipes,
-		int *pipe_cnt_out,
-		int *pipe_split_from,
-		int *vlevel_out,
-		bool fast_validate)
-{
-	bool out = false;
-	int split[MAX_PIPES] = { 0 };
-	int pipe_cnt, i, pipe_idx, vlevel;
-
-	ASSERT(pipes);
-	if (!pipes)
-		return false;
-
-	dcn20_merge_pipes_for_validate(dc, context);
-
-	pipe_cnt = dc->res_pool->funcs->populate_dml_pipes(dc, context, pipes, fast_validate);
-
-	*pipe_cnt_out = pipe_cnt;
-
-	if (!pipe_cnt) {
-		out = true;
-		goto validate_out;
-	}
-
-	vlevel = dml_get_voltage_level(&context->bw_ctx.dml, pipes, pipe_cnt);
-
-	if (vlevel > context->bw_ctx.dml.soc.num_states)
-		goto validate_fail;
-
-	vlevel = dcn20_validate_apply_pipe_split_flags(dc, context, vlevel, split, NULL);
-
-	/*initialize pipe_just_split_from to invalid idx*/
-	for (i = 0; i < MAX_PIPES; i++)
-		pipe_split_from[i] = -1;
-
-	for (i = 0, pipe_idx = -1; i < dc->res_pool->pipe_count; i++) {
-		struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
-		struct pipe_ctx *hsplit_pipe = pipe->bottom_pipe;
-
-		if (!pipe->stream || pipe_split_from[i] >= 0)
-			continue;
-
-		pipe_idx++;
-
-		if (!pipe->top_pipe && !pipe->plane_state && context->bw_ctx.dml.vba.ODMCombineEnabled[pipe_idx]) {
-			hsplit_pipe = dcn20_find_secondary_pipe(dc, &context->res_ctx, dc->res_pool, pipe);
-			ASSERT(hsplit_pipe);
-			if (!dcn20_split_stream_for_odm(
-					dc, &context->res_ctx,
-					pipe, hsplit_pipe))
-				goto validate_fail;
-			pipe_split_from[hsplit_pipe->pipe_idx] = pipe_idx;
-			dcn20_build_mapped_resource(dc, context, pipe->stream);
-		}
-
-		if (!pipe->plane_state)
-			continue;
-		/* Skip 2nd half of already split pipe */
-		if (pipe->top_pipe && pipe->plane_state == pipe->top_pipe->plane_state)
-			continue;
-
-		/* We do not support mpo + odm at the moment */
-		if (hsplit_pipe && hsplit_pipe->plane_state != pipe->plane_state
-				&& context->bw_ctx.dml.vba.ODMCombineEnabled[pipe_idx])
-			goto validate_fail;
-
-		if (split[i] == 2) {
-			if (!hsplit_pipe || hsplit_pipe->plane_state != pipe->plane_state) {
-				/* pipe not split previously needs split */
-				hsplit_pipe = dcn20_find_secondary_pipe(dc, &context->res_ctx, dc->res_pool, pipe);
-				ASSERT(hsplit_pipe);
-				if (!hsplit_pipe) {
-					context->bw_ctx.dml.vba.RequiredDPPCLK[vlevel][context->bw_ctx.dml.vba.maxMpcComb][pipe_idx] *= 2;
-					continue;
-				}
-				if (context->bw_ctx.dml.vba.ODMCombineEnabled[pipe_idx]) {
-					if (!dcn20_split_stream_for_odm(
-							dc, &context->res_ctx,
-							pipe, hsplit_pipe))
-						goto validate_fail;
-					dcn20_build_mapped_resource(dc, context, pipe->stream);
-				} else {
-					dcn20_split_stream_for_mpc(
-							&context->res_ctx, dc->res_pool,
-							pipe, hsplit_pipe);
-					resource_build_scaling_params(pipe);
-					resource_build_scaling_params(hsplit_pipe);
-				}
-				pipe_split_from[hsplit_pipe->pipe_idx] = pipe_idx;
-			}
-		} else if (hsplit_pipe && hsplit_pipe->plane_state == pipe->plane_state) {
-			/* merge should already have been done */
-			ASSERT(0);
-		}
-	}
-	/* Actual dsc count per stream dsc validation*/
-	if (!dcn20_validate_dsc(dc, context)) {
-		context->bw_ctx.dml.vba.ValidationStatus[context->bw_ctx.dml.vba.soc.num_states] =
-				DML_FAIL_DSC_VALIDATION_FAILURE;
-		goto validate_fail;
-	}
-
-	*vlevel_out = vlevel;
-
-	out = true;
-	goto validate_out;
-
-validate_fail:
-	out = false;
-
-validate_out:
-	return out;
-}
-
-static void dcn20_calculate_wm(
-		struct dc *dc, struct dc_state *context,
-		display_e2e_pipe_params_st *pipes,
-		int *out_pipe_cnt,
-		int *pipe_split_from,
-		int vlevel,
-		bool fast_validate)
-{
-	int pipe_cnt, i, pipe_idx;
-
-	for (i = 0, pipe_idx = 0, pipe_cnt = 0; i < dc->res_pool->pipe_count; i++) {
-		if (!context->res_ctx.pipe_ctx[i].stream)
-			continue;
-
-		pipes[pipe_cnt].clks_cfg.refclk_mhz = dc->res_pool->ref_clocks.dchub_ref_clock_inKhz / 1000.0;
-		pipes[pipe_cnt].clks_cfg.dispclk_mhz = context->bw_ctx.dml.vba.RequiredDISPCLK[vlevel][context->bw_ctx.dml.vba.maxMpcComb];
-
-		if (pipe_split_from[i] < 0) {
-			pipes[pipe_cnt].clks_cfg.dppclk_mhz =
-					context->bw_ctx.dml.vba.RequiredDPPCLK[vlevel][context->bw_ctx.dml.vba.maxMpcComb][pipe_idx];
-			if (context->bw_ctx.dml.vba.BlendingAndTiming[pipe_idx] == pipe_idx)
-				pipes[pipe_cnt].pipe.dest.odm_combine =
-						context->bw_ctx.dml.vba.ODMCombineEnabled[pipe_idx];
-			else
-				pipes[pipe_cnt].pipe.dest.odm_combine = 0;
-			pipe_idx++;
-		} else {
-			pipes[pipe_cnt].clks_cfg.dppclk_mhz =
-					context->bw_ctx.dml.vba.RequiredDPPCLK[vlevel][context->bw_ctx.dml.vba.maxMpcComb][pipe_split_from[i]];
-			if (context->bw_ctx.dml.vba.BlendingAndTiming[pipe_split_from[i]] == pipe_split_from[i])
-				pipes[pipe_cnt].pipe.dest.odm_combine =
-						context->bw_ctx.dml.vba.ODMCombineEnabled[pipe_split_from[i]];
-			else
-				pipes[pipe_cnt].pipe.dest.odm_combine = 0;
-		}
-
-		if (dc->config.forced_clocks) {
-			pipes[pipe_cnt].clks_cfg.dispclk_mhz = context->bw_ctx.dml.soc.clock_limits[0].dispclk_mhz;
-			pipes[pipe_cnt].clks_cfg.dppclk_mhz = context->bw_ctx.dml.soc.clock_limits[0].dppclk_mhz;
-		}
-		if (dc->debug.min_disp_clk_khz > pipes[pipe_cnt].clks_cfg.dispclk_mhz * 1000)
-			pipes[pipe_cnt].clks_cfg.dispclk_mhz = dc->debug.min_disp_clk_khz / 1000.0;
-		if (dc->debug.min_dpp_clk_khz > pipes[pipe_cnt].clks_cfg.dppclk_mhz * 1000)
-			pipes[pipe_cnt].clks_cfg.dppclk_mhz = dc->debug.min_dpp_clk_khz / 1000.0;
-
-		pipe_cnt++;
-	}
-
-	if (pipe_cnt != pipe_idx) {
-		if (dc->res_pool->funcs->populate_dml_pipes)
-			pipe_cnt = dc->res_pool->funcs->populate_dml_pipes(dc,
-				context, pipes, fast_validate);
-		else
-			pipe_cnt = dcn20_populate_dml_pipes_from_context(dc,
-				context, pipes, fast_validate);
-	}
-
-	*out_pipe_cnt = pipe_cnt;
-
-	pipes[0].clks_cfg.voltage = vlevel;
-	pipes[0].clks_cfg.dcfclk_mhz = context->bw_ctx.dml.soc.clock_limits[vlevel].dcfclk_mhz;
-	pipes[0].clks_cfg.socclk_mhz = context->bw_ctx.dml.soc.clock_limits[vlevel].socclk_mhz;
-
-	/* only pipe 0 is read for voltage and dcf/soc clocks */
-	if (vlevel < 1) {
-		pipes[0].clks_cfg.voltage = 1;
-		pipes[0].clks_cfg.dcfclk_mhz = context->bw_ctx.dml.soc.clock_limits[1].dcfclk_mhz;
-		pipes[0].clks_cfg.socclk_mhz = context->bw_ctx.dml.soc.clock_limits[1].socclk_mhz;
-	}
-	context->bw_ctx.bw.dcn.watermarks.b.urgent_ns = get_wm_urgent(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.b.cstate_pstate.cstate_enter_plus_exit_ns = get_wm_stutter_enter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.b.cstate_pstate.cstate_exit_ns = get_wm_stutter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.b.cstate_pstate.pstate_change_ns = get_wm_dram_clock_change(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.b.pte_meta_urgent_ns = get_wm_memory_trip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.b.frac_urg_bw_nom = get_fraction_of_urgent_bandwidth(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.b.frac_urg_bw_flip = get_fraction_of_urgent_bandwidth_imm_flip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.b.urgent_latency_ns = get_urgent_latency(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-
-	if (vlevel < 2) {
-		pipes[0].clks_cfg.voltage = 2;
-		pipes[0].clks_cfg.dcfclk_mhz = context->bw_ctx.dml.soc.clock_limits[2].dcfclk_mhz;
-		pipes[0].clks_cfg.socclk_mhz = context->bw_ctx.dml.soc.clock_limits[2].socclk_mhz;
-	}
-	context->bw_ctx.bw.dcn.watermarks.c.urgent_ns = get_wm_urgent(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.c.cstate_pstate.cstate_enter_plus_exit_ns = get_wm_stutter_enter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.c.cstate_pstate.cstate_exit_ns = get_wm_stutter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.c.cstate_pstate.pstate_change_ns = get_wm_dram_clock_change(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.c.pte_meta_urgent_ns = get_wm_memory_trip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.c.frac_urg_bw_nom = get_fraction_of_urgent_bandwidth(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.c.frac_urg_bw_flip = get_fraction_of_urgent_bandwidth_imm_flip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-
-	if (vlevel < 3) {
-		pipes[0].clks_cfg.voltage = 3;
-		pipes[0].clks_cfg.dcfclk_mhz = context->bw_ctx.dml.soc.clock_limits[2].dcfclk_mhz;
-		pipes[0].clks_cfg.socclk_mhz = context->bw_ctx.dml.soc.clock_limits[2].socclk_mhz;
-	}
-	context->bw_ctx.bw.dcn.watermarks.d.urgent_ns = get_wm_urgent(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.d.cstate_pstate.cstate_enter_plus_exit_ns = get_wm_stutter_enter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.d.cstate_pstate.cstate_exit_ns = get_wm_stutter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.d.cstate_pstate.pstate_change_ns = get_wm_dram_clock_change(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.d.pte_meta_urgent_ns = get_wm_memory_trip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.d.frac_urg_bw_nom = get_fraction_of_urgent_bandwidth(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.d.frac_urg_bw_flip = get_fraction_of_urgent_bandwidth_imm_flip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-
-	pipes[0].clks_cfg.voltage = vlevel;
-	pipes[0].clks_cfg.dcfclk_mhz = context->bw_ctx.dml.soc.clock_limits[vlevel].dcfclk_mhz;
-	pipes[0].clks_cfg.socclk_mhz = context->bw_ctx.dml.soc.clock_limits[vlevel].socclk_mhz;
-	context->bw_ctx.bw.dcn.watermarks.a.urgent_ns = get_wm_urgent(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.a.cstate_pstate.cstate_enter_plus_exit_ns = get_wm_stutter_enter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.a.cstate_pstate.cstate_exit_ns = get_wm_stutter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.a.cstate_pstate.pstate_change_ns = get_wm_dram_clock_change(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.a.pte_meta_urgent_ns = get_wm_memory_trip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.a.frac_urg_bw_nom = get_fraction_of_urgent_bandwidth(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.a.frac_urg_bw_flip = get_fraction_of_urgent_bandwidth_imm_flip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-}
-
-static bool is_dtbclk_required(struct dc *dc, struct dc_state *context)
-{
-	int i;
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		if (!context->res_ctx.pipe_ctx[i].stream)
-			continue;
-		if (is_dp_128b_132b_signal(&context->res_ctx.pipe_ctx[i]))
-			return true;
-	}
-	return false;
-}
-
-static enum dcn_zstate_support_state  decide_zstate_support(struct dc *dc, struct dc_state *context)
-{
-	int plane_count;
-	int i;
-
-	plane_count = 0;
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		if (context->res_ctx.pipe_ctx[i].plane_state)
-			plane_count++;
-	}
-
-	/*
-	 * Zstate is allowed in following scenarios:
-	 * 	1. Single eDP with PSR enabled
-	 * 	2. 0 planes (No memory requests)
-	 * 	3. Single eDP without PSR but > 5ms stutter period
-	 */
-	if (plane_count == 0)
-		return DCN_ZSTATE_SUPPORT_ALLOW;
-	else if (context->stream_count == 1 &&  context->streams[0]->signal == SIGNAL_TYPE_EDP) {
-		struct dc_link *link = context->streams[0]->sink->link;
-
-		if ((link->link_index == 0 && link->psr_settings.psr_feature_enabled)
-				|| context->bw_ctx.dml.vba.StutterPeriod > 5000.0)
-			return DCN_ZSTATE_SUPPORT_ALLOW;
-		else
-			return DCN_ZSTATE_SUPPORT_DISALLOW;
-	} else
-		return DCN_ZSTATE_SUPPORT_DISALLOW;
-}
-
-void dcn20_calculate_dlg_params(
-		struct dc *dc, struct dc_state *context,
-		display_e2e_pipe_params_st *pipes,
-		int pipe_cnt,
-		int vlevel)
-{
-	int i, pipe_idx;
-
-	/* Writeback MCIF_WB arbitration parameters */
-	dc->res_pool->funcs->set_mcif_arb_params(dc, context, pipes, pipe_cnt);
-
-	context->bw_ctx.bw.dcn.clk.dispclk_khz = context->bw_ctx.dml.vba.DISPCLK * 1000;
-	context->bw_ctx.bw.dcn.clk.dcfclk_khz = context->bw_ctx.dml.vba.DCFCLK * 1000;
-	context->bw_ctx.bw.dcn.clk.socclk_khz = context->bw_ctx.dml.vba.SOCCLK * 1000;
-	context->bw_ctx.bw.dcn.clk.dramclk_khz = context->bw_ctx.dml.vba.DRAMSpeed * 1000 / 16;
-
-	if (dc->debug.min_dram_clk_khz > context->bw_ctx.bw.dcn.clk.dramclk_khz)
-		context->bw_ctx.bw.dcn.clk.dramclk_khz = dc->debug.min_dram_clk_khz;
-
-	context->bw_ctx.bw.dcn.clk.dcfclk_deep_sleep_khz = context->bw_ctx.dml.vba.DCFCLKDeepSleep * 1000;
-	context->bw_ctx.bw.dcn.clk.fclk_khz = context->bw_ctx.dml.vba.FabricClock * 1000;
-	context->bw_ctx.bw.dcn.clk.p_state_change_support =
-		context->bw_ctx.dml.vba.DRAMClockChangeSupport[vlevel][context->bw_ctx.dml.vba.maxMpcComb]
-							!= dm_dram_clock_change_unsupported;
-	context->bw_ctx.bw.dcn.clk.dppclk_khz = 0;
-
-	context->bw_ctx.bw.dcn.clk.zstate_support = decide_zstate_support(dc, context);
-
-	context->bw_ctx.bw.dcn.clk.dtbclk_en = is_dtbclk_required(dc, context);
-
-	if (context->bw_ctx.bw.dcn.clk.dispclk_khz < dc->debug.min_disp_clk_khz)
-		context->bw_ctx.bw.dcn.clk.dispclk_khz = dc->debug.min_disp_clk_khz;
-
-	for (i = 0, pipe_idx = 0; i < dc->res_pool->pipe_count; i++) {
-		if (!context->res_ctx.pipe_ctx[i].stream)
-			continue;
-		pipes[pipe_idx].pipe.dest.vstartup_start = get_vstartup(&context->bw_ctx.dml, pipes, pipe_cnt, pipe_idx);
-		pipes[pipe_idx].pipe.dest.vupdate_offset = get_vupdate_offset(&context->bw_ctx.dml, pipes, pipe_cnt, pipe_idx);
-		pipes[pipe_idx].pipe.dest.vupdate_width = get_vupdate_width(&context->bw_ctx.dml, pipes, pipe_cnt, pipe_idx);
-		pipes[pipe_idx].pipe.dest.vready_offset = get_vready_offset(&context->bw_ctx.dml, pipes, pipe_cnt, pipe_idx);
-		context->res_ctx.pipe_ctx[i].det_buffer_size_kb = context->bw_ctx.dml.ip.det_buffer_size_kbytes;
-		context->res_ctx.pipe_ctx[i].unbounded_req = pipes[pipe_idx].pipe.src.unbounded_req_mode;
-
-		if (context->bw_ctx.bw.dcn.clk.dppclk_khz < pipes[pipe_idx].clks_cfg.dppclk_mhz * 1000)
-			context->bw_ctx.bw.dcn.clk.dppclk_khz = pipes[pipe_idx].clks_cfg.dppclk_mhz * 1000;
-		context->res_ctx.pipe_ctx[i].plane_res.bw.dppclk_khz =
-						pipes[pipe_idx].clks_cfg.dppclk_mhz * 1000;
-		context->res_ctx.pipe_ctx[i].pipe_dlg_param = pipes[pipe_idx].pipe.dest;
-		pipe_idx++;
-	}
-	/*save a original dppclock copy*/
-	context->bw_ctx.bw.dcn.clk.bw_dppclk_khz = context->bw_ctx.bw.dcn.clk.dppclk_khz;
-	context->bw_ctx.bw.dcn.clk.bw_dispclk_khz = context->bw_ctx.bw.dcn.clk.dispclk_khz;
-	context->bw_ctx.bw.dcn.clk.max_supported_dppclk_khz = context->bw_ctx.dml.soc.clock_limits[vlevel].dppclk_mhz * 1000;
-	context->bw_ctx.bw.dcn.clk.max_supported_dispclk_khz = context->bw_ctx.dml.soc.clock_limits[vlevel].dispclk_mhz * 1000;
-
-	context->bw_ctx.bw.dcn.compbuf_size_kb = context->bw_ctx.dml.ip.config_return_buffer_size_in_kbytes
-						- context->bw_ctx.dml.ip.det_buffer_size_kbytes * pipe_idx;
-
-	for (i = 0, pipe_idx = 0; i < dc->res_pool->pipe_count; i++) {
-		bool cstate_en = context->bw_ctx.dml.vba.PrefetchMode[vlevel][context->bw_ctx.dml.vba.maxMpcComb] != 2;
-
-		if (!context->res_ctx.pipe_ctx[i].stream)
-			continue;
-
-		if (dc->ctx->dce_version == DCN_VERSION_2_01)
-			cstate_en = false;
-
-		context->bw_ctx.dml.funcs.rq_dlg_get_dlg_reg(&context->bw_ctx.dml,
-				&context->res_ctx.pipe_ctx[i].dlg_regs,
-				&context->res_ctx.pipe_ctx[i].ttu_regs,
-				pipes,
-				pipe_cnt,
-				pipe_idx,
-				cstate_en,
-				context->bw_ctx.bw.dcn.clk.p_state_change_support,
-				false, false, true);
-
-		context->bw_ctx.dml.funcs.rq_dlg_get_rq_reg(&context->bw_ctx.dml,
-				&context->res_ctx.pipe_ctx[i].rq_regs,
-				&pipes[pipe_idx].pipe);
-		pipe_idx++;
-	}
-}
-
-static bool dcn20_validate_bandwidth_internal(struct dc *dc, struct dc_state *context,
-		bool fast_validate)
-{
-	bool out = false;
-
-	BW_VAL_TRACE_SETUP();
-
-	int vlevel = 0;
-	int pipe_split_from[MAX_PIPES];
-	int pipe_cnt = 0;
-	display_e2e_pipe_params_st *pipes = kzalloc(dc->res_pool->pipe_count * sizeof(display_e2e_pipe_params_st), GFP_ATOMIC);
-	DC_LOGGER_INIT(dc->ctx->logger);
-
-	BW_VAL_TRACE_COUNT();
-
-	out = dcn20_fast_validate_bw(dc, context, pipes, &pipe_cnt, pipe_split_from, &vlevel, fast_validate);
-
-	if (pipe_cnt == 0)
-		goto validate_out;
-
-	if (!out)
-		goto validate_fail;
-
-	BW_VAL_TRACE_END_VOLTAGE_LEVEL();
-
-	if (fast_validate) {
-		BW_VAL_TRACE_SKIP(fast);
-		goto validate_out;
-	}
-
-	dcn20_calculate_wm(dc, context, pipes, &pipe_cnt, pipe_split_from, vlevel, fast_validate);
-	dcn20_calculate_dlg_params(dc, context, pipes, pipe_cnt, vlevel);
-
-	BW_VAL_TRACE_END_WATERMARKS();
-
-	goto validate_out;
-
-validate_fail:
-	DC_LOG_WARNING("Mode Validation Warning: %s failed validation.\n",
-		dml_get_status_message(context->bw_ctx.dml.vba.ValidationStatus[context->bw_ctx.dml.vba.soc.num_states]));
-
-	BW_VAL_TRACE_SKIP(fail);
-	out = false;
-
-validate_out:
-	kfree(pipes);
-
-	BW_VAL_TRACE_FINISH();
-
-	return out;
-}
-
-/*
- * This must be noinline to ensure anything that deals with FP registers
- * is contained within this call; previously our compiling with hard-float
- * would result in fp instructions being emitted outside of the boundaries
- * of the DC_FP_START/END macros, which makes sense as the compiler has no
- * idea about what is wrapped and what is not
- *
- * This is largely just a workaround to avoid breakage introduced with 5.6,
- * ideally all fp-using code should be moved into its own file, only that
- * should be compiled with hard-float, and all code exported from there
- * should be strictly wrapped with DC_FP_START/END
- */
-static noinline bool dcn20_validate_bandwidth_fp(struct dc *dc,
-		struct dc_state *context, bool fast_validate)
-{
-	bool voltage_supported = false;
-	bool full_pstate_supported = false;
-	bool dummy_pstate_supported = false;
-	double p_state_latency_us;
-
-	p_state_latency_us = context->bw_ctx.dml.soc.dram_clock_change_latency_us;
-	context->bw_ctx.dml.soc.disable_dram_clock_change_vactive_support =
-		dc->debug.disable_dram_clock_change_vactive_support;
-	context->bw_ctx.dml.soc.allow_dram_clock_one_display_vactive =
-		dc->debug.enable_dram_clock_change_one_display_vactive;
-
-	/*Unsafe due to current pipe merge and split logic*/
-	ASSERT(context != dc->current_state);
-
-	if (fast_validate) {
-		return dcn20_validate_bandwidth_internal(dc, context, true);
-	}
-
-	// Best case, we support full UCLK switch latency
-	voltage_supported = dcn20_validate_bandwidth_internal(dc, context, false);
-	full_pstate_supported = context->bw_ctx.bw.dcn.clk.p_state_change_support;
-
-	if (context->bw_ctx.dml.soc.dummy_pstate_latency_us == 0 ||
-		(voltage_supported && full_pstate_supported)) {
-		context->bw_ctx.bw.dcn.clk.p_state_change_support = full_pstate_supported;
-		goto restore_dml_state;
-	}
-
-	// Fallback: Try to only support G6 temperature read latency
-	context->bw_ctx.dml.soc.dram_clock_change_latency_us = context->bw_ctx.dml.soc.dummy_pstate_latency_us;
-
-	voltage_supported = dcn20_validate_bandwidth_internal(dc, context, false);
-	dummy_pstate_supported = context->bw_ctx.bw.dcn.clk.p_state_change_support;
-
-	if (voltage_supported && (dummy_pstate_supported || !(context->stream_count))) {
-		context->bw_ctx.bw.dcn.clk.p_state_change_support = false;
-		goto restore_dml_state;
-	}
-
-	// ERROR: fallback is supposed to always work.
-	ASSERT(false);
-
-restore_dml_state:
-	context->bw_ctx.dml.soc.dram_clock_change_latency_us = p_state_latency_us;
-	return voltage_supported;
-}
-
-bool dcn20_validate_bandwidth(struct dc *dc, struct dc_state *context,
-		bool fast_validate)
-{
-	bool voltage_supported;
-	DC_FP_START();
-	voltage_supported = dcn20_validate_bandwidth_fp(dc, context, fast_validate);
-	DC_FP_END();
-	return voltage_supported;
-}
-
-struct pipe_ctx *dcn20_acquire_idle_pipe_for_layer(
-		struct dc_state *state,
-		const struct resource_pool *pool,
-		struct dc_stream_state *stream)
-{
-	struct resource_context *res_ctx = &state->res_ctx;
-	struct pipe_ctx *head_pipe = resource_get_head_pipe_for_stream(res_ctx, stream);
-	struct pipe_ctx *idle_pipe = find_idle_secondary_pipe(res_ctx, pool, head_pipe);
-
-	if (!head_pipe)
-		ASSERT(0);
-
-	if (!idle_pipe)
-		return NULL;
-
-	idle_pipe->stream = head_pipe->stream;
-	idle_pipe->stream_res.tg = head_pipe->stream_res.tg;
-	idle_pipe->stream_res.opp = head_pipe->stream_res.opp;
-
-	idle_pipe->plane_res.hubp = pool->hubps[idle_pipe->pipe_idx];
-	idle_pipe->plane_res.ipp = pool->ipps[idle_pipe->pipe_idx];
-	idle_pipe->plane_res.dpp = pool->dpps[idle_pipe->pipe_idx];
-	idle_pipe->plane_res.mpcc_inst = pool->dpps[idle_pipe->pipe_idx]->inst;
-
-	return idle_pipe;
-}
-
-bool dcn20_get_dcc_compression_cap(const struct dc *dc,
-		const struct dc_dcc_surface_param *input,
-		struct dc_surface_dcc_cap *output)
-{
-	return dc->res_pool->hubbub->funcs->get_dcc_compression_cap(
-			dc->res_pool->hubbub,
-			input,
-			output);
-}
-
-static void dcn20_destroy_resource_pool(struct resource_pool **pool)
-{
-	struct dcn20_resource_pool *dcn20_pool = TO_DCN20_RES_POOL(*pool);
-
-	dcn20_resource_destruct(dcn20_pool);
-	kfree(dcn20_pool);
-	*pool = NULL;
-}
-
-
-static struct dc_cap_funcs cap_funcs = {
-	.get_dcc_compression_cap = dcn20_get_dcc_compression_cap
-};
-
-
-enum dc_status dcn20_patch_unknown_plane_state(struct dc_plane_state *plane_state)
-{
-	enum surface_pixel_format surf_pix_format = plane_state->format;
-	unsigned int bpp = resource_pixel_format_to_bpp(surf_pix_format);
-
-	enum swizzle_mode_values swizzle = DC_SW_LINEAR;
-
-	if (bpp == 64)
-		swizzle = DC_SW_64KB_D;
-	else
-		swizzle = DC_SW_64KB_S;
-
-	plane_state->tiling_info.gfx9.swizzle = swizzle;
-	return DC_OK;
-}
-
-static const struct resource_funcs dcn20_res_pool_funcs = {
-	.destroy = dcn20_destroy_resource_pool,
-	.link_enc_create = dcn20_link_encoder_create,
-	.panel_cntl_create = dcn20_panel_cntl_create,
-	.validate_bandwidth = dcn20_validate_bandwidth,
-	.acquire_idle_pipe_for_layer = dcn20_acquire_idle_pipe_for_layer,
-	.add_stream_to_ctx = dcn20_add_stream_to_ctx,
-	.add_dsc_to_stream_resource = dcn20_add_dsc_to_stream_resource,
-	.remove_stream_from_ctx = dcn20_remove_stream_from_ctx,
-	.populate_dml_writeback_from_context = dcn20_populate_dml_writeback_from_context,
-	.patch_unknown_plane_state = dcn20_patch_unknown_plane_state,
-	.set_mcif_arb_params = dcn20_set_mcif_arb_params,
-	.populate_dml_pipes = dcn20_populate_dml_pipes_from_context,
-	.find_first_free_match_stream_enc_for_link = dcn10_find_first_free_match_stream_enc_for_link
-};
-
-bool dcn20_dwbc_create(struct dc_context *ctx, struct resource_pool *pool)
-{
-	int i;
-	uint32_t pipe_count = pool->res_cap->num_dwb;
-
-	for (i = 0; i < pipe_count; i++) {
-		struct dcn20_dwbc *dwbc20 = kzalloc(sizeof(struct dcn20_dwbc),
-						    GFP_KERNEL);
-
-		if (!dwbc20) {
-			dm_error("DC: failed to create dwbc20!\n");
-			return false;
-		}
-		dcn20_dwbc_construct(dwbc20, ctx,
-				&dwbc20_regs[i],
-				&dwbc20_shift,
-				&dwbc20_mask,
-				i);
-		pool->dwbc[i] = &dwbc20->base;
-	}
-	return true;
-}
-
-bool dcn20_mmhubbub_create(struct dc_context *ctx, struct resource_pool *pool)
-{
-	int i;
-	uint32_t pipe_count = pool->res_cap->num_dwb;
-
-	ASSERT(pipe_count > 0);
-
-	for (i = 0; i < pipe_count; i++) {
-		struct dcn20_mmhubbub *mcif_wb20 = kzalloc(sizeof(struct dcn20_mmhubbub),
-						    GFP_KERNEL);
-
-		if (!mcif_wb20) {
-			dm_error("DC: failed to create mcif_wb20!\n");
-			return false;
-		}
-
-		dcn20_mmhubbub_construct(mcif_wb20, ctx,
-				&mcif_wb20_regs[i],
-				&mcif_wb20_shift,
-				&mcif_wb20_mask,
-				i);
-
-		pool->mcif_wb[i] = &mcif_wb20->base;
-	}
-	return true;
-}
-
-static struct pp_smu_funcs *dcn20_pp_smu_create(struct dc_context *ctx)
-{
-	struct pp_smu_funcs *pp_smu = kzalloc(sizeof(*pp_smu), GFP_ATOMIC);
-
-	if (!pp_smu)
-		return pp_smu;
-
-	dm_pp_get_funcs(ctx, pp_smu);
-
-	if (pp_smu->ctx.ver != PP_SMU_VER_NV)
-		pp_smu = memset(pp_smu, 0, sizeof(struct pp_smu_funcs));
-
-	return pp_smu;
-}
-
-static void dcn20_pp_smu_destroy(struct pp_smu_funcs **pp_smu)
-{
-	if (pp_smu && *pp_smu) {
-		kfree(*pp_smu);
-		*pp_smu = NULL;
-	}
-}
-
-void dcn20_cap_soc_clocks(
-		struct _vcs_dpi_soc_bounding_box_st *bb,
-		struct pp_smu_nv_clock_table max_clocks)
-{
-	int i;
-
-	// First pass - cap all clocks higher than the reported max
-	for (i = 0; i < bb->num_states; i++) {
-		if ((bb->clock_limits[i].dcfclk_mhz > (max_clocks.dcfClockInKhz / 1000))
-				&& max_clocks.dcfClockInKhz != 0)
-			bb->clock_limits[i].dcfclk_mhz = (max_clocks.dcfClockInKhz / 1000);
-
-		if ((bb->clock_limits[i].dram_speed_mts > (max_clocks.uClockInKhz / 1000) * 16)
-						&& max_clocks.uClockInKhz != 0)
-			bb->clock_limits[i].dram_speed_mts = (max_clocks.uClockInKhz / 1000) * 16;
-
-		if ((bb->clock_limits[i].fabricclk_mhz > (max_clocks.fabricClockInKhz / 1000))
-						&& max_clocks.fabricClockInKhz != 0)
-			bb->clock_limits[i].fabricclk_mhz = (max_clocks.fabricClockInKhz / 1000);
-
-		if ((bb->clock_limits[i].dispclk_mhz > (max_clocks.displayClockInKhz / 1000))
-						&& max_clocks.displayClockInKhz != 0)
-			bb->clock_limits[i].dispclk_mhz = (max_clocks.displayClockInKhz / 1000);
-
-		if ((bb->clock_limits[i].dppclk_mhz > (max_clocks.dppClockInKhz / 1000))
-						&& max_clocks.dppClockInKhz != 0)
-			bb->clock_limits[i].dppclk_mhz = (max_clocks.dppClockInKhz / 1000);
-
-		if ((bb->clock_limits[i].phyclk_mhz > (max_clocks.phyClockInKhz / 1000))
-						&& max_clocks.phyClockInKhz != 0)
-			bb->clock_limits[i].phyclk_mhz = (max_clocks.phyClockInKhz / 1000);
-
-		if ((bb->clock_limits[i].socclk_mhz > (max_clocks.socClockInKhz / 1000))
-						&& max_clocks.socClockInKhz != 0)
-			bb->clock_limits[i].socclk_mhz = (max_clocks.socClockInKhz / 1000);
-
-		if ((bb->clock_limits[i].dscclk_mhz > (max_clocks.dscClockInKhz / 1000))
-						&& max_clocks.dscClockInKhz != 0)
-			bb->clock_limits[i].dscclk_mhz = (max_clocks.dscClockInKhz / 1000);
-	}
-
-	// Second pass - remove all duplicate clock states
-	for (i = bb->num_states - 1; i > 1; i--) {
-		bool duplicate = true;
-
-		if (bb->clock_limits[i-1].dcfclk_mhz != bb->clock_limits[i].dcfclk_mhz)
-			duplicate = false;
-		if (bb->clock_limits[i-1].dispclk_mhz != bb->clock_limits[i].dispclk_mhz)
-			duplicate = false;
-		if (bb->clock_limits[i-1].dppclk_mhz != bb->clock_limits[i].dppclk_mhz)
-			duplicate = false;
-		if (bb->clock_limits[i-1].dram_speed_mts != bb->clock_limits[i].dram_speed_mts)
-			duplicate = false;
-		if (bb->clock_limits[i-1].dscclk_mhz != bb->clock_limits[i].dscclk_mhz)
-			duplicate = false;
-		if (bb->clock_limits[i-1].fabricclk_mhz != bb->clock_limits[i].fabricclk_mhz)
-			duplicate = false;
-		if (bb->clock_limits[i-1].phyclk_mhz != bb->clock_limits[i].phyclk_mhz)
-			duplicate = false;
-		if (bb->clock_limits[i-1].socclk_mhz != bb->clock_limits[i].socclk_mhz)
-			duplicate = false;
-
-		if (duplicate)
-			bb->num_states--;
-	}
-}
-
-void dcn20_update_bounding_box(struct dc *dc, struct _vcs_dpi_soc_bounding_box_st *bb,
-		struct pp_smu_nv_clock_table *max_clocks, unsigned int *uclk_states, unsigned int num_states)
-{
-	struct _vcs_dpi_voltage_scaling_st calculated_states[DC__VOLTAGE_STATES];
-	int i;
-	int num_calculated_states = 0;
-	int min_dcfclk = 0;
-
-	if (num_states == 0)
-		return;
-
-	memset(calculated_states, 0, sizeof(calculated_states));
-
-	if (dc->bb_overrides.min_dcfclk_mhz > 0)
-		min_dcfclk = dc->bb_overrides.min_dcfclk_mhz;
-	else {
-		if (ASICREV_IS_NAVI12_P(dc->ctx->asic_id.hw_internal_rev))
-			min_dcfclk = 310;
-		else
-			// Accounting for SOC/DCF relationship, we can go as high as
-			// 506Mhz in Vmin.
-			min_dcfclk = 506;
-	}
-
-	for (i = 0; i < num_states; i++) {
-		int min_fclk_required_by_uclk;
-		calculated_states[i].state = i;
-		calculated_states[i].dram_speed_mts = uclk_states[i] * 16 / 1000;
-
-		// FCLK:UCLK ratio is 1.08
-		min_fclk_required_by_uclk = div_u64(((unsigned long long)uclk_states[i]) * 1080,
-			1000000);
-
-		calculated_states[i].fabricclk_mhz = (min_fclk_required_by_uclk < min_dcfclk) ?
-				min_dcfclk : min_fclk_required_by_uclk;
-
-		calculated_states[i].socclk_mhz = (calculated_states[i].fabricclk_mhz > max_clocks->socClockInKhz / 1000) ?
-				max_clocks->socClockInKhz / 1000 : calculated_states[i].fabricclk_mhz;
-
-		calculated_states[i].dcfclk_mhz = (calculated_states[i].fabricclk_mhz > max_clocks->dcfClockInKhz / 1000) ?
-				max_clocks->dcfClockInKhz / 1000 : calculated_states[i].fabricclk_mhz;
-
-		calculated_states[i].dispclk_mhz = max_clocks->displayClockInKhz / 1000;
-		calculated_states[i].dppclk_mhz = max_clocks->displayClockInKhz / 1000;
-		calculated_states[i].dscclk_mhz = max_clocks->displayClockInKhz / (1000 * 3);
-
-		calculated_states[i].phyclk_mhz = max_clocks->phyClockInKhz / 1000;
-
-		num_calculated_states++;
-	}
-
-	calculated_states[num_calculated_states - 1].socclk_mhz = max_clocks->socClockInKhz / 1000;
-	calculated_states[num_calculated_states - 1].fabricclk_mhz = max_clocks->socClockInKhz / 1000;
-	calculated_states[num_calculated_states - 1].dcfclk_mhz = max_clocks->dcfClockInKhz / 1000;
-
-	memcpy(bb->clock_limits, calculated_states, sizeof(bb->clock_limits));
-	bb->num_states = num_calculated_states;
-
-	// Duplicate the last state, DML always an extra state identical to max state to work
-	memcpy(&bb->clock_limits[num_calculated_states], &bb->clock_limits[num_calculated_states - 1], sizeof(struct _vcs_dpi_voltage_scaling_st));
-	bb->clock_limits[num_calculated_states].state = bb->num_states;
-}
-
-void dcn20_patch_bounding_box(struct dc *dc, struct _vcs_dpi_soc_bounding_box_st *bb)
-{
-	if ((int)(bb->sr_exit_time_us * 1000) != dc->bb_overrides.sr_exit_time_ns
-			&& dc->bb_overrides.sr_exit_time_ns) {
-		bb->sr_exit_time_us = dc->bb_overrides.sr_exit_time_ns / 1000.0;
-	}
-
-	if ((int)(bb->sr_enter_plus_exit_time_us * 1000)
-				!= dc->bb_overrides.sr_enter_plus_exit_time_ns
-			&& dc->bb_overrides.sr_enter_plus_exit_time_ns) {
-		bb->sr_enter_plus_exit_time_us =
-				dc->bb_overrides.sr_enter_plus_exit_time_ns / 1000.0;
-	}
-
-	if ((int)(bb->urgent_latency_us * 1000) != dc->bb_overrides.urgent_latency_ns
-			&& dc->bb_overrides.urgent_latency_ns) {
-		bb->urgent_latency_us = dc->bb_overrides.urgent_latency_ns / 1000.0;
-	}
-
-	if ((int)(bb->dram_clock_change_latency_us * 1000)
-				!= dc->bb_overrides.dram_clock_change_latency_ns
-			&& dc->bb_overrides.dram_clock_change_latency_ns) {
-		bb->dram_clock_change_latency_us =
-				dc->bb_overrides.dram_clock_change_latency_ns / 1000.0;
-	}
-
-	if ((int)(bb->dummy_pstate_latency_us * 1000)
-				!= dc->bb_overrides.dummy_clock_change_latency_ns
-			&& dc->bb_overrides.dummy_clock_change_latency_ns) {
-		bb->dummy_pstate_latency_us =
-				dc->bb_overrides.dummy_clock_change_latency_ns / 1000.0;
-	}
-}
-
-static struct _vcs_dpi_soc_bounding_box_st *get_asic_rev_soc_bb(
-	uint32_t hw_internal_rev)
-{
-	if (ASICREV_IS_NAVI14_M(hw_internal_rev))
-		return &dcn2_0_nv14_soc;
-
-	if (ASICREV_IS_NAVI12_P(hw_internal_rev))
-		return &dcn2_0_nv12_soc;
-
-	return &dcn2_0_soc;
-}
-
-static struct _vcs_dpi_ip_params_st *get_asic_rev_ip_params(
-	uint32_t hw_internal_rev)
-{
-	/* NV14 */
-	if (ASICREV_IS_NAVI14_M(hw_internal_rev))
-		return &dcn2_0_nv14_ip;
-
-	/* NV12 and NV10 */
-	return &dcn2_0_ip;
-}
-
-static enum dml_project get_dml_project_version(uint32_t hw_internal_rev)
-{
-	return DML_PROJECT_NAVI10v2;
-}
-
-static bool init_soc_bounding_box(struct dc *dc,
-				  struct dcn20_resource_pool *pool)
-{
-	struct _vcs_dpi_soc_bounding_box_st *loaded_bb =
-			get_asic_rev_soc_bb(dc->ctx->asic_id.hw_internal_rev);
-	struct _vcs_dpi_ip_params_st *loaded_ip =
-			get_asic_rev_ip_params(dc->ctx->asic_id.hw_internal_rev);
-
-	DC_LOGGER_INIT(dc->ctx->logger);
-
-	if (pool->base.pp_smu) {
-		struct pp_smu_nv_clock_table max_clocks = {0};
-		unsigned int uclk_states[8] = {0};
-		unsigned int num_states = 0;
-		enum pp_smu_status status;
-		bool clock_limits_available = false;
-		bool uclk_states_available = false;
-
-		if (pool->base.pp_smu->nv_funcs.get_uclk_dpm_states) {
-			status = (pool->base.pp_smu->nv_funcs.get_uclk_dpm_states)
-				(&pool->base.pp_smu->nv_funcs.pp_smu, uclk_states, &num_states);
-
-			uclk_states_available = (status == PP_SMU_RESULT_OK);
-		}
-
-		if (pool->base.pp_smu->nv_funcs.get_maximum_sustainable_clocks) {
-			status = (*pool->base.pp_smu->nv_funcs.get_maximum_sustainable_clocks)
-					(&pool->base.pp_smu->nv_funcs.pp_smu, &max_clocks);
-			/* SMU cannot set DCF clock to anything equal to or higher than SOC clock
-			 */
-			if (max_clocks.dcfClockInKhz >= max_clocks.socClockInKhz)
-				max_clocks.dcfClockInKhz = max_clocks.socClockInKhz - 1000;
-			clock_limits_available = (status == PP_SMU_RESULT_OK);
-		}
-
-		if (clock_limits_available && uclk_states_available && num_states) {
-			DC_FP_START();
-			dcn20_update_bounding_box(dc, loaded_bb, &max_clocks, uclk_states, num_states);
-			DC_FP_END();
-		} else if (clock_limits_available) {
-			DC_FP_START();
-			dcn20_cap_soc_clocks(loaded_bb, max_clocks);
-			DC_FP_END();
-		}
-	}
-
-	loaded_ip->max_num_otg = pool->base.res_cap->num_timing_generator;
-	loaded_ip->max_num_dpp = pool->base.pipe_count;
-	DC_FP_START();
-	dcn20_patch_bounding_box(dc, loaded_bb);
-	DC_FP_END();
-	return true;
-}
-
-static bool dcn20_resource_construct(
-	uint8_t num_virtual_links,
-	struct dc *dc,
-	struct dcn20_resource_pool *pool)
-{
-	int i;
-	struct dc_context *ctx = dc->ctx;
-	struct irq_service_init_data init_data;
-	struct ddc_service_init_data ddc_init_data = {0};
-	struct _vcs_dpi_soc_bounding_box_st *loaded_bb =
-			get_asic_rev_soc_bb(ctx->asic_id.hw_internal_rev);
-	struct _vcs_dpi_ip_params_st *loaded_ip =
-			get_asic_rev_ip_params(ctx->asic_id.hw_internal_rev);
-	enum dml_project dml_project_version =
-			get_dml_project_version(ctx->asic_id.hw_internal_rev);
-
-	ctx->dc_bios->regs = &bios_regs;
-	pool->base.funcs = &dcn20_res_pool_funcs;
-
-	if (ASICREV_IS_NAVI14_M(ctx->asic_id.hw_internal_rev)) {
-		pool->base.res_cap = &res_cap_nv14;
-		pool->base.pipe_count = 5;
-		pool->base.mpcc_count = 5;
-	} else {
-		pool->base.res_cap = &res_cap_nv10;
-		pool->base.pipe_count = 6;
-		pool->base.mpcc_count = 6;
-	}
-	/*************************************************
-	 *  Resource + asic cap harcoding                *
-	 *************************************************/
-	pool->base.underlay_pipe_index = NO_UNDERLAY_PIPE;
-
-	dc->caps.max_downscale_ratio = 200;
-	dc->caps.i2c_speed_in_khz = 100;
-	dc->caps.i2c_speed_in_khz_hdcp = 100; /*1.4 w/a not applied by default*/
-	dc->caps.max_cursor_size = 256;
-	dc->caps.min_horizontal_blanking_period = 80;
-	dc->caps.dmdata_alloc_size = 2048;
-
-	dc->caps.max_slave_planes = 1;
-	dc->caps.max_slave_yuv_planes = 1;
-	dc->caps.max_slave_rgb_planes = 1;
-	dc->caps.post_blend_color_processing = true;
-	dc->caps.force_dp_tps4_for_cp2520 = true;
-	dc->caps.extended_aux_timeout_support = true;
-
-	/* Color pipeline capabilities */
-	dc->caps.color.dpp.dcn_arch = 1;
-	dc->caps.color.dpp.input_lut_shared = 0;
-	dc->caps.color.dpp.icsc = 1;
-	dc->caps.color.dpp.dgam_ram = 1;
-	dc->caps.color.dpp.dgam_rom_caps.srgb = 1;
-	dc->caps.color.dpp.dgam_rom_caps.bt2020 = 1;
-	dc->caps.color.dpp.dgam_rom_caps.gamma2_2 = 0;
-	dc->caps.color.dpp.dgam_rom_caps.pq = 0;
-	dc->caps.color.dpp.dgam_rom_caps.hlg = 0;
-	dc->caps.color.dpp.post_csc = 0;
-	dc->caps.color.dpp.gamma_corr = 0;
-	dc->caps.color.dpp.dgam_rom_for_yuv = 1;
-
-	dc->caps.color.dpp.hw_3d_lut = 1;
-	dc->caps.color.dpp.ogam_ram = 1;
-	// no OGAM ROM on DCN2, only MPC ROM
-	dc->caps.color.dpp.ogam_rom_caps.srgb = 0;
-	dc->caps.color.dpp.ogam_rom_caps.bt2020 = 0;
-	dc->caps.color.dpp.ogam_rom_caps.gamma2_2 = 0;
-	dc->caps.color.dpp.ogam_rom_caps.pq = 0;
-	dc->caps.color.dpp.ogam_rom_caps.hlg = 0;
-	dc->caps.color.dpp.ocsc = 0;
-
-	dc->caps.color.mpc.gamut_remap = 0;
-	dc->caps.color.mpc.num_3dluts = 0;
-	dc->caps.color.mpc.shared_3d_lut = 0;
-	dc->caps.color.mpc.ogam_ram = 1;
-	dc->caps.color.mpc.ogam_rom_caps.srgb = 0;
-	dc->caps.color.mpc.ogam_rom_caps.bt2020 = 0;
-	dc->caps.color.mpc.ogam_rom_caps.gamma2_2 = 0;
-	dc->caps.color.mpc.ogam_rom_caps.pq = 0;
-	dc->caps.color.mpc.ogam_rom_caps.hlg = 0;
-	dc->caps.color.mpc.ocsc = 1;
-
-	if (dc->ctx->dce_environment == DCE_ENV_PRODUCTION_DRV) {
-		dc->debug = debug_defaults_drv;
-	} else if (dc->ctx->dce_environment == DCE_ENV_FPGA_MAXIMUS) {
-		pool->base.pipe_count = 4;
-		pool->base.mpcc_count = pool->base.pipe_count;
-		dc->debug = debug_defaults_diags;
-	} else {
-		dc->debug = debug_defaults_diags;
-	}
-	//dcn2.0x
-	dc->work_arounds.dedcn20_305_wa = true;
-
-	// Init the vm_helper
-	if (dc->vm_helper)
-		vm_helper_init(dc->vm_helper, 16);
-
-	/*************************************************
-	 *  Create resources                             *
-	 *************************************************/
-
-	pool->base.clock_sources[DCN20_CLK_SRC_PLL0] =
-			dcn20_clock_source_create(ctx, ctx->dc_bios,
-				CLOCK_SOURCE_COMBO_PHY_PLL0,
-				&clk_src_regs[0], false);
-	pool->base.clock_sources[DCN20_CLK_SRC_PLL1] =
-			dcn20_clock_source_create(ctx, ctx->dc_bios,
-				CLOCK_SOURCE_COMBO_PHY_PLL1,
-				&clk_src_regs[1], false);
-	pool->base.clock_sources[DCN20_CLK_SRC_PLL2] =
-			dcn20_clock_source_create(ctx, ctx->dc_bios,
-				CLOCK_SOURCE_COMBO_PHY_PLL2,
-				&clk_src_regs[2], false);
-	pool->base.clock_sources[DCN20_CLK_SRC_PLL3] =
-			dcn20_clock_source_create(ctx, ctx->dc_bios,
-				CLOCK_SOURCE_COMBO_PHY_PLL3,
-				&clk_src_regs[3], false);
-	pool->base.clock_sources[DCN20_CLK_SRC_PLL4] =
-			dcn20_clock_source_create(ctx, ctx->dc_bios,
-				CLOCK_SOURCE_COMBO_PHY_PLL4,
-				&clk_src_regs[4], false);
-	pool->base.clock_sources[DCN20_CLK_SRC_PLL5] =
-			dcn20_clock_source_create(ctx, ctx->dc_bios,
-				CLOCK_SOURCE_COMBO_PHY_PLL5,
-				&clk_src_regs[5], false);
-	pool->base.clk_src_count = DCN20_CLK_SRC_TOTAL;
-	/* todo: not reuse phy_pll registers */
-	pool->base.dp_clock_source =
-			dcn20_clock_source_create(ctx, ctx->dc_bios,
-				CLOCK_SOURCE_ID_DP_DTO,
-				&clk_src_regs[0], true);
-
-	for (i = 0; i < pool->base.clk_src_count; i++) {
-		if (pool->base.clock_sources[i] == NULL) {
-			dm_error("DC: failed to create clock sources!\n");
-			BREAK_TO_DEBUGGER();
-			goto create_fail;
-		}
-	}
-
-	pool->base.dccg = dccg2_create(ctx, &dccg_regs, &dccg_shift, &dccg_mask);
-	if (pool->base.dccg == NULL) {
-		dm_error("DC: failed to create dccg!\n");
-		BREAK_TO_DEBUGGER();
-		goto create_fail;
-	}
-
-	pool->base.dmcu = dcn20_dmcu_create(ctx,
-			&dmcu_regs,
-			&dmcu_shift,
-			&dmcu_mask);
-	if (pool->base.dmcu == NULL) {
-		dm_error("DC: failed to create dmcu!\n");
-		BREAK_TO_DEBUGGER();
-		goto create_fail;
-	}
-
-	pool->base.abm = dce_abm_create(ctx,
-			&abm_regs,
-			&abm_shift,
-			&abm_mask);
-	if (pool->base.abm == NULL) {
-		dm_error("DC: failed to create abm!\n");
-		BREAK_TO_DEBUGGER();
-		goto create_fail;
-	}
-
-	pool->base.pp_smu = dcn20_pp_smu_create(ctx);
-
-
-	if (!init_soc_bounding_box(dc, pool)) {
-		dm_error("DC: failed to initialize soc bounding box!\n");
-		BREAK_TO_DEBUGGER();
-		goto create_fail;
-	}
-
-	dml_init_instance(&dc->dml, loaded_bb, loaded_ip, dml_project_version);
-
-	if (!dc->debug.disable_pplib_wm_range) {
-		struct pp_smu_wm_range_sets ranges = {0};
-		int i = 0;
-
-		ranges.num_reader_wm_sets = 0;
-
-		if (loaded_bb->num_states == 1) {
-			ranges.reader_wm_sets[0].wm_inst = i;
-			ranges.reader_wm_sets[0].min_drain_clk_mhz = PP_SMU_WM_SET_RANGE_CLK_UNCONSTRAINED_MIN;
-			ranges.reader_wm_sets[0].max_drain_clk_mhz = PP_SMU_WM_SET_RANGE_CLK_UNCONSTRAINED_MAX;
-			ranges.reader_wm_sets[0].min_fill_clk_mhz = PP_SMU_WM_SET_RANGE_CLK_UNCONSTRAINED_MIN;
-			ranges.reader_wm_sets[0].max_fill_clk_mhz = PP_SMU_WM_SET_RANGE_CLK_UNCONSTRAINED_MAX;
-
-			ranges.num_reader_wm_sets = 1;
-		} else if (loaded_bb->num_states > 1) {
-			for (i = 0; i < 4 && i < loaded_bb->num_states; i++) {
-				ranges.reader_wm_sets[i].wm_inst = i;
-				ranges.reader_wm_sets[i].min_drain_clk_mhz = PP_SMU_WM_SET_RANGE_CLK_UNCONSTRAINED_MIN;
-				ranges.reader_wm_sets[i].max_drain_clk_mhz = PP_SMU_WM_SET_RANGE_CLK_UNCONSTRAINED_MAX;
-				ranges.reader_wm_sets[i].min_fill_clk_mhz = (i > 0) ? (loaded_bb->clock_limits[i - 1].dram_speed_mts / 16) + 1 : 0;
-				ranges.reader_wm_sets[i].max_fill_clk_mhz = loaded_bb->clock_limits[i].dram_speed_mts / 16;
-
-				ranges.num_reader_wm_sets = i + 1;
-			}
-
-			ranges.reader_wm_sets[0].min_fill_clk_mhz = PP_SMU_WM_SET_RANGE_CLK_UNCONSTRAINED_MIN;
-			ranges.reader_wm_sets[ranges.num_reader_wm_sets - 1].max_fill_clk_mhz = PP_SMU_WM_SET_RANGE_CLK_UNCONSTRAINED_MAX;
-		}
-
-		ranges.num_writer_wm_sets = 1;
-
-		ranges.writer_wm_sets[0].wm_inst = 0;
-		ranges.writer_wm_sets[0].min_fill_clk_mhz = PP_SMU_WM_SET_RANGE_CLK_UNCONSTRAINED_MIN;
-		ranges.writer_wm_sets[0].max_fill_clk_mhz = PP_SMU_WM_SET_RANGE_CLK_UNCONSTRAINED_MAX;
-		ranges.writer_wm_sets[0].min_drain_clk_mhz = PP_SMU_WM_SET_RANGE_CLK_UNCONSTRAINED_MIN;
-		ranges.writer_wm_sets[0].max_drain_clk_mhz = PP_SMU_WM_SET_RANGE_CLK_UNCONSTRAINED_MAX;
-
-		/* Notify PP Lib/SMU which Watermarks to use for which clock ranges */
-		if (pool->base.pp_smu->nv_funcs.set_wm_ranges)
-			pool->base.pp_smu->nv_funcs.set_wm_ranges(&pool->base.pp_smu->nv_funcs.pp_smu, &ranges);
-	}
-
-	init_data.ctx = dc->ctx;
-	pool->base.irqs = dal_irq_service_dcn20_create(&init_data);
-	if (!pool->base.irqs)
-		goto create_fail;
-
-	/* mem input -> ipp -> dpp -> opp -> TG */
-	for (i = 0; i < pool->base.pipe_count; i++) {
-		pool->base.hubps[i] = dcn20_hubp_create(ctx, i);
-		if (pool->base.hubps[i] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error(
-				"DC: failed to create memory input!\n");
-			goto create_fail;
-		}
-
-		pool->base.ipps[i] = dcn20_ipp_create(ctx, i);
-		if (pool->base.ipps[i] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error(
-				"DC: failed to create input pixel processor!\n");
-			goto create_fail;
-		}
-
-		pool->base.dpps[i] = dcn20_dpp_create(ctx, i);
-		if (pool->base.dpps[i] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error(
-				"DC: failed to create dpps!\n");
-			goto create_fail;
-		}
-	}
-	for (i = 0; i < pool->base.res_cap->num_ddc; i++) {
-		pool->base.engines[i] = dcn20_aux_engine_create(ctx, i);
-		if (pool->base.engines[i] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error(
-				"DC:failed to create aux engine!!\n");
-			goto create_fail;
-		}
-		pool->base.hw_i2cs[i] = dcn20_i2c_hw_create(ctx, i);
-		if (pool->base.hw_i2cs[i] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error(
-				"DC:failed to create hw i2c!!\n");
-			goto create_fail;
-		}
-		pool->base.sw_i2cs[i] = NULL;
-	}
-
-	for (i = 0; i < pool->base.res_cap->num_opp; i++) {
-		pool->base.opps[i] = dcn20_opp_create(ctx, i);
-		if (pool->base.opps[i] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error(
-				"DC: failed to create output pixel processor!\n");
-			goto create_fail;
-		}
-	}
-
-	for (i = 0; i < pool->base.res_cap->num_timing_generator; i++) {
-		pool->base.timing_generators[i] = dcn20_timing_generator_create(
-				ctx, i);
-		if (pool->base.timing_generators[i] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error("DC: failed to create tg!\n");
-			goto create_fail;
-		}
-	}
-
-	pool->base.timing_generator_count = i;
-
-	pool->base.mpc = dcn20_mpc_create(ctx);
-	if (pool->base.mpc == NULL) {
-		BREAK_TO_DEBUGGER();
-		dm_error("DC: failed to create mpc!\n");
-		goto create_fail;
-	}
-
-	pool->base.hubbub = dcn20_hubbub_create(ctx);
-	if (pool->base.hubbub == NULL) {
-		BREAK_TO_DEBUGGER();
-		dm_error("DC: failed to create hubbub!\n");
-		goto create_fail;
-	}
-
-	for (i = 0; i < pool->base.res_cap->num_dsc; i++) {
-		pool->base.dscs[i] = dcn20_dsc_create(ctx, i);
-		if (pool->base.dscs[i] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error("DC: failed to create display stream compressor %d!\n", i);
-			goto create_fail;
-		}
-	}
-
-	if (!dcn20_dwbc_create(ctx, &pool->base)) {
-		BREAK_TO_DEBUGGER();
-		dm_error("DC: failed to create dwbc!\n");
-		goto create_fail;
-	}
-	if (!dcn20_mmhubbub_create(ctx, &pool->base)) {
-		BREAK_TO_DEBUGGER();
-		dm_error("DC: failed to create mcif_wb!\n");
-		goto create_fail;
-	}
-
-	if (!resource_construct(num_virtual_links, dc, &pool->base,
-			(!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment) ?
-			&res_create_funcs : &res_create_maximus_funcs)))
-			goto create_fail;
-
-	dcn20_hw_sequencer_construct(dc);
-
-	// IF NV12, set PG function pointer to NULL. It's not that
-	// PG isn't supported for NV12, it's that we don't want to
-	// program the registers because that will cause more power
-	// to be consumed. We could have created dcn20_init_hw to get
-	// the same effect by checking ASIC rev, but there was a
-	// request at some point to not check ASIC rev on hw sequencer.
-	if (ASICREV_IS_NAVI12_P(dc->ctx->asic_id.hw_internal_rev)) {
-		dc->hwseq->funcs.enable_power_gating_plane = NULL;
-		dc->debug.disable_dpp_power_gate = true;
-		dc->debug.disable_hubp_power_gate = true;
-	}
-
-
-	dc->caps.max_planes =  pool->base.pipe_count;
-
-	for (i = 0; i < dc->caps.max_planes; ++i)
-		dc->caps.planes[i] = plane_cap;
-
-	dc->cap_funcs = cap_funcs;
-
-	if (dc->ctx->dc_bios->fw_info.oem_i2c_present) {
-		ddc_init_data.ctx = dc->ctx;
-		ddc_init_data.link = NULL;
-		ddc_init_data.id.id = dc->ctx->dc_bios->fw_info.oem_i2c_obj_id;
-		ddc_init_data.id.enum_id = 0;
-		ddc_init_data.id.type = OBJECT_TYPE_GENERIC;
-		pool->base.oem_device = dal_ddc_service_create(&ddc_init_data);
-	} else {
-		pool->base.oem_device = NULL;
-	}
-
-	return true;
-
-create_fail:
-
-	dcn20_resource_destruct(pool);
-
-	return false;
-}
-
-struct resource_pool *dcn20_create_resource_pool(
-		const struct dc_init_data *init_data,
-		struct dc *dc)
-{
-	struct dcn20_resource_pool *pool =
-		kzalloc(sizeof(struct dcn20_resource_pool), GFP_ATOMIC);
-
-	if (!pool)
-		return NULL;
-
-	if (dcn20_resource_construct(init_data->num_virtual_links, dc, pool))
-		return &pool->base;
-
-	BREAK_TO_DEBUGGER();
-	kfree(pool);
-	return NULL;
-}
diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_vmid.c b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_vmid.c
index 96c263223315..5bc3bc60a2ac 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_vmid.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_vmid.c
@@ -38,6 +38,9 @@
 #define FN(reg_name, field_name) \
 	vmid->shifts->field_name, vmid->masks->field_name
 
+#define DC_LOGGER \
+	CTX->logger
+
 static void dcn20_wait_for_vmid_ready(struct dcn20_vmid *vmid)
 {
 	/* According the hardware spec, we need to poll for the lowest
diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_vmid.h b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_vmid.h
index f1ef46e8da5b..e7a1b7fa2cce 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_vmid.h
+++ b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_vmid.h
@@ -28,12 +28,6 @@
 
 #include "vmid.h"
 
-#define BASE_INNER(seg) \
-	DCE_BASE__INST0_SEG ## seg
-
-#define BASE(seg) \
-	BASE_INNER(seg)
-
 #define DCN20_VMID_REG_LIST(id)\
 	SRI(CNTL, DCN_VM_CONTEXT, id),\
 	SRI(PAGE_TABLE_BASE_ADDR_HI32, DCN_VM_CONTEXT, id),\
diff --git a/drivers/gpu/drm/amd/display/dc/dcn201/Makefile b/drivers/gpu/drm/amd/display/dc/dcn201/Makefile
index f68038ceb1b1..c9f4a5a9f522 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn201/Makefile
+++ b/drivers/gpu/drm/amd/display/dc/dcn201/Makefile
@@ -1,36 +1,9 @@
 # SPDX-License-Identifier: MIT
-#
-# Makefile for DCN.
-DCN201 = dcn201_init.o dcn201_resource.o dcn201_hwseq.o \
-	dcn201_hubbub.o\
-	dcn201_mpc.o dcn201_hubp.o dcn201_opp.o dcn201_optc.o dcn201_dpp.o \
-	dcn201_dccg.o dcn201_link_encoder.o
+# Copyright © 2021-2024 Advanced Micro Devices, Inc. All rights reserved.
 
-ifdef CONFIG_X86
-CFLAGS_$(AMDDALPATH)/dc/dcn201/dcn201_resource.o := -mhard-float -msse
-endif
+DCN201 = dcn201_mpc.o dcn201_opp.o \
+	dcn201_link_encoder.o
 
-ifdef CONFIG_PPC64
-CFLAGS_$(AMDDALPATH)/dc/dcn201/dcn201_resource.o := -mhard-float -maltivec
-endif
-
-ifdef CONFIG_CC_IS_GCC
-ifeq ($(call cc-ifversion, -lt, 0701, y), y)
-IS_OLD_GCC = 1
-endif
-CFLAGS_$(AMDDALPATH)/dc/dcn201/dcn201_resource.o += -mhard-float
-endif
-
-ifdef CONFIG_X86
-ifdef IS_OLD_GCC
-# Stack alignment mismatch, proceed with caution.
-# GCC < 7.1 cannot compile code using `double` and -mpreferred-stack-boundary=3
-# (8B stack alignment).
-CFLAGS_$(AMDDALPATH)/dc/dcn201/dcn201_resource.o += -mpreferred-stack-boundary=4
-else
-CFLAGS_$(AMDDALPATH)/dc/dcn201/dcn201_resource.o += -msse2
-endif
-endif
 AMD_DAL_DCN201 = $(addprefix $(AMDDALPATH)/dc/dcn201/,$(DCN201))
 
 AMD_DISPLAY_FILES += $(AMD_DAL_DCN201)
diff --git a/drivers/gpu/drm/amd/display/dc/dcn201/dcn201_link_encoder.c b/drivers/gpu/drm/amd/display/dc/dcn201/dcn201_link_encoder.c
index a65e8f7801db..8d31fa131cd6 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn201/dcn201_link_encoder.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn201/dcn201_link_encoder.c
@@ -29,7 +29,6 @@
 #include "link_encoder.h"
 #include "dcn201_link_encoder.h"
 #include "stream_encoder.h"
-#include "i2caux_interface.h"
 #include "dc_bios_types.h"
 
 #include "gpio_service_interface.h"
@@ -50,8 +49,8 @@
 #define IND_REG(index) \
 	(enc10->link_regs->index)
 
-void dcn201_link_encoder_get_max_link_cap(struct link_encoder *enc,
-	struct dc_link_settings *link_settings)
+static void dcn201_link_encoder_get_max_link_cap(struct link_encoder *enc,
+						 struct dc_link_settings *link_settings)
 {
 	uint32_t value1, value2;
 	struct dcn10_link_encoder *enc10 = TO_DCN10_LINK_ENC(enc);
@@ -66,7 +65,7 @@ void dcn201_link_encoder_get_max_link_cap(struct link_encoder *enc,
 	}
 }
 
-bool dcn201_link_encoder_is_in_alt_mode(struct link_encoder *enc)
+static bool dcn201_link_encoder_is_in_alt_mode(struct link_encoder *enc)
 {
 	uint32_t value;
 	struct dcn10_link_encoder *enc10 = TO_DCN10_LINK_ENC(enc);
diff --git a/drivers/gpu/drm/amd/display/dc/dcn201/dcn201_link_encoder.h b/drivers/gpu/drm/amd/display/dc/dcn201/dcn201_link_encoder.h
index 8b95ef251332..be25e8dc0636 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn201/dcn201_link_encoder.h
+++ b/drivers/gpu/drm/amd/display/dc/dcn201/dcn201_link_encoder.h
@@ -30,6 +30,10 @@
 
 #define DPCS_DCN201_MASK_SH_LIST(mask_sh)\
 	DPCS_MASK_SH_LIST(mask_sh),\
+	LE_SF(DPCSSYS_CR0_RAWLANE0_DIG_PCS_XF_RX_OVRD_IN_2, VCO_LD_VAL_OVRD, mask_sh),\
+	LE_SF(DPCSSYS_CR0_RAWLANE0_DIG_PCS_XF_RX_OVRD_IN_2, VCO_LD_VAL_OVRD_EN, mask_sh),\
+	LE_SF(DPCSSYS_CR0_RAWLANE0_DIG_PCS_XF_RX_OVRD_IN_3, REF_LD_VAL_OVRD, mask_sh),\
+	LE_SF(DPCSSYS_CR0_RAWLANE0_DIG_PCS_XF_RX_OVRD_IN_3, REF_LD_VAL_OVRD_EN, mask_sh),\
 	LE_SF(RDPCSTX0_RDPCSTX_PHY_CNTL2, RDPCS_PHY_DPALT_DISABLE_ACK, mask_sh),\
 	LE_SF(RDPCSTX0_RDPCSTX_PHY_CNTL2, RDPCS_PHY_DPALT_DISABLE, mask_sh),\
 	LE_SF(RDPCSTX0_RDPCSTX_PHY_CNTL2, RDPCS_PHY_DPALT_DP4, mask_sh),\
@@ -44,7 +48,15 @@
 	LE_SF(RDPCSTX0_RDPCSTX_PHY_CNTL11, RDPCS_PHY_DP_REF_CLK_EN, mask_sh)
 
 #define DPCS_DCN201_REG_LIST(id) \
-	DPCS_DCN2_CMN_REG_LIST(id)
+	DPCS_DCN2_CMN_REG_LIST(id), \
+	SRI_IX(RAWLANE0_DIG_PCS_XF_RX_OVRD_IN_2, DPCSSYS_CR, id), \
+	SRI_IX(RAWLANE0_DIG_PCS_XF_RX_OVRD_IN_3, DPCSSYS_CR, id), \
+	SRI_IX(RAWLANE1_DIG_PCS_XF_RX_OVRD_IN_2, DPCSSYS_CR, id), \
+	SRI_IX(RAWLANE1_DIG_PCS_XF_RX_OVRD_IN_3, DPCSSYS_CR, id), \
+	SRI_IX(RAWLANE2_DIG_PCS_XF_RX_OVRD_IN_2, DPCSSYS_CR, id), \
+	SRI_IX(RAWLANE2_DIG_PCS_XF_RX_OVRD_IN_3, DPCSSYS_CR, id), \
+	SRI_IX(RAWLANE3_DIG_PCS_XF_RX_OVRD_IN_2, DPCSSYS_CR, id), \
+	SRI_IX(RAWLANE3_DIG_PCS_XF_RX_OVRD_IN_3, DPCSSYS_CR, id)
 
 void dcn201_link_encoder_construct(
 	struct dcn20_link_encoder *enc20,
diff --git a/drivers/gpu/drm/amd/display/dc/dcn201/dcn201_mpc.c b/drivers/gpu/drm/amd/display/dc/dcn201/dcn201_mpc.c
index 95c4c55f067c..1af03a86ec9b 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn201/dcn201_mpc.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn201/dcn201_mpc.c
@@ -76,7 +76,7 @@ static void mpc201_init_mpcc(struct mpcc *mpcc, int mpcc_inst)
 	mpcc->shared_bottom = false;
 }
 
-const struct mpc_funcs dcn201_mpc_funcs = {
+static const struct mpc_funcs dcn201_mpc_funcs = {
 	.read_mpcc_state = mpc1_read_mpcc_state,
 	.insert_plane = mpc1_insert_plane,
 	.remove_mpcc = mpc1_remove_mpcc,
diff --git a/drivers/gpu/drm/amd/display/dc/dcn201/dcn201_opp.c b/drivers/gpu/drm/amd/display/dc/dcn201/dcn201_opp.c
index 8e77db46a409..e83367a9b6b3 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn201/dcn201_opp.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn201/dcn201_opp.c
@@ -50,9 +50,11 @@ static struct opp_funcs dcn201_opp_funcs = {
 		.opp_set_disp_pattern_generator = opp2_set_disp_pattern_generator,
 		.opp_program_dpg_dimensions = opp2_program_dpg_dimensions,
 		.dpg_is_blanked = opp2_dpg_is_blanked,
+		.dpg_is_pending = opp2_dpg_is_pending,
 		.opp_dpg_set_blank_color = opp2_dpg_set_blank_color,
 		.opp_destroy = opp1_destroy,
 		.opp_program_left_edge_extra_pixel = opp2_program_left_edge_extra_pixel,
+		.opp_get_left_edge_extra_pixel_count = opp2_get_left_edge_extra_pixel_count,
 };
 
 void dcn201_opp_construct(struct dcn201_opp *oppn201,
diff --git a/drivers/gpu/drm/amd/display/dc/dcn201/dcn201_opp.h b/drivers/gpu/drm/amd/display/dc/dcn201/dcn201_opp.h
index aca389ec1779..edb7f9653cb6 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn201/dcn201_opp.h
+++ b/drivers/gpu/drm/amd/display/dc/dcn201/dcn201_opp.h
@@ -42,8 +42,7 @@
 	OPP_MASK_SH_LIST_DCN20(mask_sh)
 
 #define OPP_DCN201_REG_FIELD_LIST(type) \
-	OPP_DCN20_REG_FIELD_LIST(type);
-
+	OPP_DCN20_REG_FIELD_LIST(type)
 struct dcn201_opp_shift {
 	OPP_DCN201_REG_FIELD_LIST(uint8_t);
 };
diff --git a/drivers/gpu/drm/amd/display/dc/dcn21/Makefile b/drivers/gpu/drm/amd/display/dc/dcn21/Makefile
index bb8c95141082..c215f3cc6e44 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn21/Makefile
+++ b/drivers/gpu/drm/amd/display/dc/dcn21/Makefile
@@ -1,34 +1,7 @@
 # SPDX-License-Identifier: MIT
-#
-# Makefile for DCN21.
+# Copyright © 2019-2024 Advanced Micro Devices, Inc. All rights reserved.
 
-DCN21 = dcn21_init.o dcn21_hubp.o dcn21_hubbub.o dcn21_resource.o \
-	 dcn21_hwseq.o dcn21_link_encoder.o dcn21_dccg.o
-
-ifdef CONFIG_X86
-CFLAGS_$(AMDDALPATH)/dc/dcn21/dcn21_resource.o := -mhard-float -msse
-endif
-
-ifdef CONFIG_PPC64
-CFLAGS_$(AMDDALPATH)/dc/dcn21/dcn21_resource.o := -mhard-float -maltivec
-endif
-
-ifdef CONFIG_CC_IS_GCC
-ifeq ($(call cc-ifversion, -lt, 0701, y), y)
-IS_OLD_GCC = 1
-endif
-endif
-
-ifdef CONFIG_X86
-ifdef IS_OLD_GCC
-# Stack alignment mismatch, proceed with caution.
-# GCC < 7.1 cannot compile code using `double` and -mpreferred-stack-boundary=3
-# (8B stack alignment).
-CFLAGS_$(AMDDALPATH)/dc/dcn21/dcn21_resource.o += -mpreferred-stack-boundary=4
-else
-CFLAGS_$(AMDDALPATH)/dc/dcn21/dcn21_resource.o += -msse2
-endif
-endif
+DCN21 = dcn21_link_encoder.o
 
 AMD_DAL_DCN21 = $(addprefix $(AMDDALPATH)/dc/dcn21/,$(DCN21))
 
diff --git a/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_hwseq.c b/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_hwseq.c
deleted file mode 100644
index 69cc192a7e71..000000000000
--- a/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_hwseq.c
+++ /dev/null
@@ -1,246 +0,0 @@
-/*
- * Copyright 2016 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: AMD
- *
- */
-
-#include "dm_services.h"
-#include "dm_helpers.h"
-#include "core_types.h"
-#include "resource.h"
-#include "dce/dce_hwseq.h"
-#include "dce110/dce110_hw_sequencer.h"
-#include "dcn21_hwseq.h"
-#include "vmid.h"
-#include "reg_helper.h"
-#include "hw/clk_mgr.h"
-#include "dc_dmub_srv.h"
-#include "abm.h"
-
-
-#define DC_LOGGER_INIT(logger)
-
-#define CTX \
-	hws->ctx
-#define REG(reg)\
-	hws->regs->reg
-
-#undef FN
-#define FN(reg_name, field_name) \
-	hws->shifts->field_name, hws->masks->field_name
-
-/* Temporary read settings, future will get values from kmd directly */
-static void mmhub_update_page_table_config(struct dcn_hubbub_phys_addr_config *config,
-		struct dce_hwseq *hws)
-{
-	uint32_t page_table_base_hi;
-	uint32_t page_table_base_lo;
-
-	REG_GET(VM_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32,
-			PAGE_DIRECTORY_ENTRY_HI32, &page_table_base_hi);
-	REG_GET(VM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32,
-			PAGE_DIRECTORY_ENTRY_LO32, &page_table_base_lo);
-
-	config->gart_config.page_table_base_addr = ((uint64_t)page_table_base_hi << 32) | page_table_base_lo;
-
-}
-
-int dcn21_init_sys_ctx(struct dce_hwseq *hws, struct dc *dc, struct dc_phy_addr_space_config *pa_config)
-{
-	struct dcn_hubbub_phys_addr_config config;
-
-	config.system_aperture.fb_top = pa_config->system_aperture.fb_top;
-	config.system_aperture.fb_offset = pa_config->system_aperture.fb_offset;
-	config.system_aperture.fb_base = pa_config->system_aperture.fb_base;
-	config.system_aperture.agp_top = pa_config->system_aperture.agp_top;
-	config.system_aperture.agp_bot = pa_config->system_aperture.agp_bot;
-	config.system_aperture.agp_base = pa_config->system_aperture.agp_base;
-	config.gart_config.page_table_start_addr = pa_config->gart_config.page_table_start_addr;
-	config.gart_config.page_table_end_addr = pa_config->gart_config.page_table_end_addr;
-	config.gart_config.page_table_base_addr = pa_config->gart_config.page_table_base_addr;
-
-	mmhub_update_page_table_config(&config, hws);
-
-	return dc->res_pool->hubbub->funcs->init_dchub_sys_ctx(dc->res_pool->hubbub, &config);
-}
-
-// work around for Renoir s0i3, if register is programmed, bypass golden init.
-
-bool dcn21_s0i3_golden_init_wa(struct dc *dc)
-{
-	struct dce_hwseq *hws = dc->hwseq;
-	uint32_t value = 0;
-
-	value = REG_READ(MICROSECOND_TIME_BASE_DIV);
-
-	return value != 0x00120464;
-}
-
-void dcn21_exit_optimized_pwr_state(
-		const struct dc *dc,
-		struct dc_state *context)
-{
-	dc->clk_mgr->funcs->update_clocks(
-			dc->clk_mgr,
-			context,
-			false);
-}
-
-void dcn21_optimize_pwr_state(
-		const struct dc *dc,
-		struct dc_state *context)
-{
-	dc->clk_mgr->funcs->update_clocks(
-			dc->clk_mgr,
-			context,
-			true);
-}
-
-/* If user hotplug a HDMI monitor while in monitor off,
- * OS will do a mode set (with output timing) but keep output off.
- * In this case DAL will ask vbios to power up the pll in the PHY.
- * If user unplug the monitor (while we are on monitor off) or
- * system attempt to enter modern standby (which we will disable PLL),
- * PHY will hang on the next mode set attempt.
- * if enable PLL follow by disable PLL (without executing lane enable/disable),
- * RDPCS_PHY_DP_MPLLB_STATE remains 1,
- * which indicate that PLL disable attempt actually didn't go through.
- * As a workaround, insert PHY lane enable/disable before PLL disable.
- */
-void dcn21_PLAT_58856_wa(struct dc_state *context, struct pipe_ctx *pipe_ctx)
-{
-	if (!pipe_ctx->stream->dpms_off)
-		return;
-
-	pipe_ctx->stream->dpms_off = false;
-	core_link_enable_stream(context, pipe_ctx);
-	core_link_disable_stream(pipe_ctx);
-	pipe_ctx->stream->dpms_off = true;
-}
-
-static bool dmub_abm_set_pipe(struct abm *abm, uint32_t otg_inst, uint32_t option, uint32_t panel_inst)
-{
-	union dmub_rb_cmd cmd;
-	struct dc_context *dc = abm->ctx;
-	uint32_t ramping_boundary = 0xFFFF;
-
-	memset(&cmd, 0, sizeof(cmd));
-	cmd.abm_set_pipe.header.type = DMUB_CMD__ABM;
-	cmd.abm_set_pipe.header.sub_type = DMUB_CMD__ABM_SET_PIPE;
-	cmd.abm_set_pipe.abm_set_pipe_data.otg_inst = otg_inst;
-	cmd.abm_set_pipe.abm_set_pipe_data.set_pipe_option = option;
-	cmd.abm_set_pipe.abm_set_pipe_data.panel_inst = panel_inst;
-	cmd.abm_set_pipe.abm_set_pipe_data.ramping_boundary = ramping_boundary;
-	cmd.abm_set_pipe.header.payload_bytes = sizeof(struct dmub_cmd_abm_set_pipe_data);
-
-	dc_dmub_srv_cmd_queue(dc->dmub_srv, &cmd);
-	dc_dmub_srv_cmd_execute(dc->dmub_srv);
-	dc_dmub_srv_wait_idle(dc->dmub_srv);
-
-	return true;
-}
-
-void dcn21_set_abm_immediate_disable(struct pipe_ctx *pipe_ctx)
-{
-	struct abm *abm = pipe_ctx->stream_res.abm;
-	uint32_t otg_inst = pipe_ctx->stream_res.tg->inst;
-	struct panel_cntl *panel_cntl = pipe_ctx->stream->link->panel_cntl;
-
-	struct dmcu *dmcu = pipe_ctx->stream->ctx->dc->res_pool->dmcu;
-
-	if (dmcu) {
-		dce110_set_abm_immediate_disable(pipe_ctx);
-		return;
-	}
-
-	if (abm && panel_cntl) {
-		dmub_abm_set_pipe(abm, otg_inst, SET_ABM_PIPE_IMMEDIATELY_DISABLE,
-				panel_cntl->inst);
-		panel_cntl->funcs->store_backlight_level(panel_cntl);
-	}
-}
-
-void dcn21_set_pipe(struct pipe_ctx *pipe_ctx)
-{
-	struct abm *abm = pipe_ctx->stream_res.abm;
-	uint32_t otg_inst = pipe_ctx->stream_res.tg->inst;
-	struct panel_cntl *panel_cntl = pipe_ctx->stream->link->panel_cntl;
-	struct dmcu *dmcu = pipe_ctx->stream->ctx->dc->res_pool->dmcu;
-
-	if (dmcu) {
-		dce110_set_pipe(pipe_ctx);
-		return;
-	}
-
-	if (abm && panel_cntl)
-		dmub_abm_set_pipe(abm, otg_inst, SET_ABM_PIPE_NORMAL, panel_cntl->inst);
-}
-
-bool dcn21_set_backlight_level(struct pipe_ctx *pipe_ctx,
-		uint32_t backlight_pwm_u16_16,
-		uint32_t frame_ramp)
-{
-	union dmub_rb_cmd cmd;
-	struct dc_context *dc = pipe_ctx->stream->ctx;
-	struct abm *abm = pipe_ctx->stream_res.abm;
-	uint32_t otg_inst = pipe_ctx->stream_res.tg->inst;
-	struct panel_cntl *panel_cntl = pipe_ctx->stream->link->panel_cntl;
-
-	if (dc->dc->res_pool->dmcu) {
-		dce110_set_backlight_level(pipe_ctx, backlight_pwm_u16_16, frame_ramp);
-		return true;
-	}
-
-	if (abm && panel_cntl)
-		dmub_abm_set_pipe(abm, otg_inst, SET_ABM_PIPE_NORMAL, panel_cntl->inst);
-
-	memset(&cmd, 0, sizeof(cmd));
-	cmd.abm_set_backlight.header.type = DMUB_CMD__ABM;
-	cmd.abm_set_backlight.header.sub_type = DMUB_CMD__ABM_SET_BACKLIGHT;
-	cmd.abm_set_backlight.abm_set_backlight_data.frame_ramp = frame_ramp;
-	cmd.abm_set_backlight.abm_set_backlight_data.backlight_user_level = backlight_pwm_u16_16;
-	cmd.abm_set_backlight.abm_set_backlight_data.version = DMUB_CMD_ABM_CONTROL_VERSION_1;
-	cmd.abm_set_backlight.abm_set_backlight_data.panel_mask = (0x01 << panel_cntl->inst);
-	cmd.abm_set_backlight.header.payload_bytes = sizeof(struct dmub_cmd_abm_set_backlight_data);
-
-	dc_dmub_srv_cmd_queue(dc->dmub_srv, &cmd);
-	dc_dmub_srv_cmd_execute(dc->dmub_srv);
-	dc_dmub_srv_wait_idle(dc->dmub_srv);
-
-	return true;
-}
-
-bool dcn21_is_abm_supported(struct dc *dc,
-		struct dc_state *context, struct dc_stream_state *stream)
-{
-	int i;
-
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i];
-
-		if (pipe_ctx->stream == stream &&
-				(pipe_ctx->prev_odm_pipe == NULL && pipe_ctx->next_odm_pipe == NULL))
-			return true;
-	}
-	return false;
-}
-
diff --git a/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_link_encoder.c b/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_link_encoder.c
index aa46c35b05a2..eb9abb9f9698 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_link_encoder.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_link_encoder.c
@@ -31,7 +31,6 @@
 #include "dcn21_link_encoder.h"
 #include "stream_encoder.h"
 
-#include "i2caux_interface.h"
 #include "dc_bios_types.h"
 
 #include "gpio_service_interface.h"
@@ -203,7 +202,7 @@ static bool update_cfg_data(
 	return true;
 }
 
-bool dcn21_link_encoder_acquire_phy(struct link_encoder *enc)
+static bool dcn21_link_encoder_acquire_phy(struct link_encoder *enc)
 {
 	struct dcn10_link_encoder *enc10 = TO_DCN10_LINK_ENC(enc);
 	int value;
@@ -277,7 +276,7 @@ void dcn21_link_encoder_enable_dp_output(
 
 }
 
-void dcn21_link_encoder_enable_dp_mst_output(
+static void dcn21_link_encoder_enable_dp_mst_output(
 	struct link_encoder *enc,
 	const struct dc_link_settings *link_settings,
 	enum clock_source_id clock_source)
@@ -288,9 +287,8 @@ void dcn21_link_encoder_enable_dp_mst_output(
 	dcn10_link_encoder_enable_dp_mst_output(enc, link_settings, clock_source);
 }
 
-void dcn21_link_encoder_disable_output(
-	struct link_encoder *enc,
-	enum signal_type signal)
+static void dcn21_link_encoder_disable_output(struct link_encoder *enc,
+					      enum signal_type signal)
 {
 	dcn10_link_encoder_disable_output(enc, signal);
 
diff --git a/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_resource.c b/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_resource.c
deleted file mode 100644
index d452a0d1777e..000000000000
--- a/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_resource.c
+++ /dev/null
@@ -1,2289 +0,0 @@
-/*
-* Copyright 2018 Advanced Micro Devices, Inc.
- * Copyright 2019 Raptor Engineering, LLC
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: AMD
- *
- */
-
-#include <linux/slab.h>
-
-#include "dm_services.h"
-#include "dc.h"
-
-#include "dcn21_init.h"
-
-#include "resource.h"
-#include "include/irq_service_interface.h"
-#include "dcn20/dcn20_resource.h"
-
-#include "dml/dcn20/dcn20_fpu.h"
-
-#include "clk_mgr.h"
-#include "dcn10/dcn10_hubp.h"
-#include "dcn10/dcn10_ipp.h"
-#include "dcn20/dcn20_hubbub.h"
-#include "dcn20/dcn20_mpc.h"
-#include "dcn20/dcn20_hubp.h"
-#include "dcn21_hubp.h"
-#include "irq/dcn21/irq_service_dcn21.h"
-#include "dcn20/dcn20_dpp.h"
-#include "dcn20/dcn20_optc.h"
-#include "dcn21/dcn21_hwseq.h"
-#include "dce110/dce110_hw_sequencer.h"
-#include "dcn20/dcn20_opp.h"
-#include "dcn20/dcn20_dsc.h"
-#include "dcn21/dcn21_link_encoder.h"
-#include "dcn20/dcn20_stream_encoder.h"
-#include "dce/dce_clock_source.h"
-#include "dce/dce_audio.h"
-#include "dce/dce_hwseq.h"
-#include "virtual/virtual_stream_encoder.h"
-#include "dml/display_mode_vba.h"
-#include "dcn20/dcn20_dccg.h"
-#include "dcn21/dcn21_dccg.h"
-#include "dcn21_hubbub.h"
-#include "dcn10/dcn10_resource.h"
-#include "dce/dce_panel_cntl.h"
-
-#include "dcn20/dcn20_dwb.h"
-#include "dcn20/dcn20_mmhubbub.h"
-#include "dpcs/dpcs_2_1_0_offset.h"
-#include "dpcs/dpcs_2_1_0_sh_mask.h"
-
-#include "renoir_ip_offset.h"
-#include "dcn/dcn_2_1_0_offset.h"
-#include "dcn/dcn_2_1_0_sh_mask.h"
-
-#include "nbio/nbio_7_0_offset.h"
-
-#include "mmhub/mmhub_2_0_0_offset.h"
-#include "mmhub/mmhub_2_0_0_sh_mask.h"
-
-#include "reg_helper.h"
-#include "dce/dce_abm.h"
-#include "dce/dce_dmcu.h"
-#include "dce/dce_aux.h"
-#include "dce/dce_i2c.h"
-#include "dcn21_resource.h"
-#include "vm_helper.h"
-#include "dcn20/dcn20_vmid.h"
-#include "dce/dmub_psr.h"
-#include "dce/dmub_abm.h"
-
-#define DC_LOGGER_INIT(logger)
-
-
-struct _vcs_dpi_ip_params_st dcn2_1_ip = {
-	.odm_capable = 1,
-	.gpuvm_enable = 1,
-	.hostvm_enable = 1,
-	.gpuvm_max_page_table_levels = 1,
-	.hostvm_max_page_table_levels = 4,
-	.hostvm_cached_page_table_levels = 2,
-	.num_dsc = 3,
-	.rob_buffer_size_kbytes = 168,
-	.det_buffer_size_kbytes = 164,
-	.dpte_buffer_size_in_pte_reqs_luma = 44,
-	.dpte_buffer_size_in_pte_reqs_chroma = 42,//todo
-	.dpp_output_buffer_pixels = 2560,
-	.opp_output_buffer_lines = 1,
-	.pixel_chunk_size_kbytes = 8,
-	.pte_enable = 1,
-	.max_page_table_levels = 4,
-	.pte_chunk_size_kbytes = 2,
-	.meta_chunk_size_kbytes = 2,
-	.min_meta_chunk_size_bytes = 256,
-	.writeback_chunk_size_kbytes = 2,
-	.line_buffer_size_bits = 789504,
-	.is_line_buffer_bpp_fixed = 0,
-	.line_buffer_fixed_bpp = 0,
-	.dcc_supported = true,
-	.max_line_buffer_lines = 12,
-	.writeback_luma_buffer_size_kbytes = 12,
-	.writeback_chroma_buffer_size_kbytes = 8,
-	.writeback_chroma_line_buffer_width_pixels = 4,
-	.writeback_max_hscl_ratio = 1,
-	.writeback_max_vscl_ratio = 1,
-	.writeback_min_hscl_ratio = 1,
-	.writeback_min_vscl_ratio = 1,
-	.writeback_max_hscl_taps = 12,
-	.writeback_max_vscl_taps = 12,
-	.writeback_line_buffer_luma_buffer_size = 0,
-	.writeback_line_buffer_chroma_buffer_size = 14643,
-	.cursor_buffer_size = 8,
-	.cursor_chunk_size = 2,
-	.max_num_otg = 4,
-	.max_num_dpp = 4,
-	.max_num_wb = 1,
-	.max_dchub_pscl_bw_pix_per_clk = 4,
-	.max_pscl_lb_bw_pix_per_clk = 2,
-	.max_lb_vscl_bw_pix_per_clk = 4,
-	.max_vscl_hscl_bw_pix_per_clk = 4,
-	.max_hscl_ratio = 4,
-	.max_vscl_ratio = 4,
-	.hscl_mults = 4,
-	.vscl_mults = 4,
-	.max_hscl_taps = 8,
-	.max_vscl_taps = 8,
-	.dispclk_ramp_margin_percent = 1,
-	.underscan_factor = 1.10,
-	.min_vblank_lines = 32, //
-	.dppclk_delay_subtotal = 77, //
-	.dppclk_delay_scl_lb_only = 16,
-	.dppclk_delay_scl = 50,
-	.dppclk_delay_cnvc_formatter = 8,
-	.dppclk_delay_cnvc_cursor = 6,
-	.dispclk_delay_subtotal = 87, //
-	.dcfclk_cstate_latency = 10, // SRExitTime
-	.max_inter_dcn_tile_repeaters = 8,
-
-	.xfc_supported = false,
-	.xfc_fill_bw_overhead_percent = 10.0,
-	.xfc_fill_constant_bytes = 0,
-	.ptoi_supported = 0,
-	.number_of_cursors = 1,
-};
-
-struct _vcs_dpi_soc_bounding_box_st dcn2_1_soc = {
-	.clock_limits = {
-			{
-				.state = 0,
-				.dcfclk_mhz = 400.0,
-				.fabricclk_mhz = 400.0,
-				.dispclk_mhz = 600.0,
-				.dppclk_mhz = 400.00,
-				.phyclk_mhz = 600.0,
-				.socclk_mhz = 278.0,
-				.dscclk_mhz = 205.67,
-				.dram_speed_mts = 1600.0,
-			},
-			{
-				.state = 1,
-				.dcfclk_mhz = 464.52,
-				.fabricclk_mhz = 800.0,
-				.dispclk_mhz = 654.55,
-				.dppclk_mhz = 626.09,
-				.phyclk_mhz = 600.0,
-				.socclk_mhz = 278.0,
-				.dscclk_mhz = 205.67,
-				.dram_speed_mts = 1600.0,
-			},
-			{
-				.state = 2,
-				.dcfclk_mhz = 514.29,
-				.fabricclk_mhz = 933.0,
-				.dispclk_mhz = 757.89,
-				.dppclk_mhz = 685.71,
-				.phyclk_mhz = 600.0,
-				.socclk_mhz = 278.0,
-				.dscclk_mhz = 287.67,
-				.dram_speed_mts = 1866.0,
-			},
-			{
-				.state = 3,
-				.dcfclk_mhz = 576.00,
-				.fabricclk_mhz = 1067.0,
-				.dispclk_mhz = 847.06,
-				.dppclk_mhz = 757.89,
-				.phyclk_mhz = 600.0,
-				.socclk_mhz = 715.0,
-				.dscclk_mhz = 318.334,
-				.dram_speed_mts = 2134.0,
-			},
-			{
-				.state = 4,
-				.dcfclk_mhz = 626.09,
-				.fabricclk_mhz = 1200.0,
-				.dispclk_mhz = 900.00,
-				.dppclk_mhz = 847.06,
-				.phyclk_mhz = 810.0,
-				.socclk_mhz = 953.0,
-				.dscclk_mhz = 489.0,
-				.dram_speed_mts = 2400.0,
-			},
-			{
-				.state = 5,
-				.dcfclk_mhz = 685.71,
-				.fabricclk_mhz = 1333.0,
-				.dispclk_mhz = 1028.57,
-				.dppclk_mhz = 960.00,
-				.phyclk_mhz = 810.0,
-				.socclk_mhz = 278.0,
-				.dscclk_mhz = 287.67,
-				.dram_speed_mts = 2666.0,
-			},
-			{
-				.state = 6,
-				.dcfclk_mhz = 757.89,
-				.fabricclk_mhz = 1467.0,
-				.dispclk_mhz = 1107.69,
-				.dppclk_mhz = 1028.57,
-				.phyclk_mhz = 810.0,
-				.socclk_mhz = 715.0,
-				.dscclk_mhz = 318.334,
-				.dram_speed_mts = 3200.0,
-			},
-			{
-				.state = 7,
-				.dcfclk_mhz = 847.06,
-				.fabricclk_mhz = 1600.0,
-				.dispclk_mhz = 1395.0,
-				.dppclk_mhz = 1285.00,
-				.phyclk_mhz = 1325.0,
-				.socclk_mhz = 953.0,
-				.dscclk_mhz = 489.0,
-				.dram_speed_mts = 4266.0,
-			},
-			/*Extra state, no dispclk ramping*/
-			{
-				.state = 8,
-				.dcfclk_mhz = 847.06,
-				.fabricclk_mhz = 1600.0,
-				.dispclk_mhz = 1395.0,
-				.dppclk_mhz = 1285.0,
-				.phyclk_mhz = 1325.0,
-				.socclk_mhz = 953.0,
-				.dscclk_mhz = 489.0,
-				.dram_speed_mts = 4266.0,
-			},
-
-		},
-
-	.sr_exit_time_us = 12.5,
-	.sr_enter_plus_exit_time_us = 17.0,
-	.urgent_latency_us = 4.0,
-	.urgent_latency_pixel_data_only_us = 4.0,
-	.urgent_latency_pixel_mixed_with_vm_data_us = 4.0,
-	.urgent_latency_vm_data_only_us = 4.0,
-	.urgent_out_of_order_return_per_channel_pixel_only_bytes = 4096,
-	.urgent_out_of_order_return_per_channel_pixel_and_vm_bytes = 4096,
-	.urgent_out_of_order_return_per_channel_vm_only_bytes = 4096,
-	.pct_ideal_dram_sdp_bw_after_urgent_pixel_only = 80.0,
-	.pct_ideal_dram_sdp_bw_after_urgent_pixel_and_vm = 75.0,
-	.pct_ideal_dram_sdp_bw_after_urgent_vm_only = 40.0,
-	.max_avg_sdp_bw_use_normal_percent = 60.0,
-	.max_avg_dram_bw_use_normal_percent = 100.0,
-	.writeback_latency_us = 12.0,
-	.max_request_size_bytes = 256,
-	.dram_channel_width_bytes = 4,
-	.fabric_datapath_to_dcn_data_return_bytes = 32,
-	.dcn_downspread_percent = 0.5,
-	.downspread_percent = 0.38,
-	.dram_page_open_time_ns = 50.0,
-	.dram_rw_turnaround_time_ns = 17.5,
-	.dram_return_buffer_per_channel_bytes = 8192,
-	.round_trip_ping_latency_dcfclk_cycles = 128,
-	.urgent_out_of_order_return_per_channel_bytes = 4096,
-	.channel_interleave_bytes = 256,
-	.num_banks = 8,
-	.num_chans = 4,
-	.vmm_page_size_bytes = 4096,
-	.dram_clock_change_latency_us = 23.84,
-	.return_bus_width_bytes = 64,
-	.dispclk_dppclk_vco_speed_mhz = 3600,
-	.xfc_bus_transport_time_us = 4,
-	.xfc_xbuf_latency_tolerance_us = 4,
-	.use_urgent_burst_bw = 1,
-	.num_states = 8
-};
-
-#ifndef MAX
-#define MAX(X, Y) ((X) > (Y) ? (X) : (Y))
-#endif
-#ifndef MIN
-#define MIN(X, Y) ((X) < (Y) ? (X) : (Y))
-#endif
-
-/* begin *********************
- * macros to expend register list macro defined in HW object header file */
-
-/* DCN */
-/* TODO awful hack. fixup dcn20_dwb.h */
-#undef BASE_INNER
-#define BASE_INNER(seg) DMU_BASE__INST0_SEG ## seg
-
-#define BASE(seg) BASE_INNER(seg)
-
-#define SR(reg_name)\
-		.reg_name = BASE(mm ## reg_name ## _BASE_IDX) +  \
-					mm ## reg_name
-
-#define SRI(reg_name, block, id)\
-	.reg_name = BASE(mm ## block ## id ## _ ## reg_name ## _BASE_IDX) + \
-					mm ## block ## id ## _ ## reg_name
-
-#define SRIR(var_name, reg_name, block, id)\
-	.var_name = BASE(mm ## block ## id ## _ ## reg_name ## _BASE_IDX) + \
-					mm ## block ## id ## _ ## reg_name
-
-#define SRII(reg_name, block, id)\
-	.reg_name[id] = BASE(mm ## block ## id ## _ ## reg_name ## _BASE_IDX) + \
-					mm ## block ## id ## _ ## reg_name
-
-#define DCCG_SRII(reg_name, block, id)\
-	.block ## _ ## reg_name[id] = BASE(mm ## block ## id ## _ ## reg_name ## _BASE_IDX) + \
-					mm ## block ## id ## _ ## reg_name
-
-#define VUPDATE_SRII(reg_name, block, id)\
-	.reg_name[id] = BASE(mm ## reg_name ## _ ## block ## id ## _BASE_IDX) + \
-					mm ## reg_name ## _ ## block ## id
-
-/* NBIO */
-#define NBIO_BASE_INNER(seg) \
-	NBIF0_BASE__INST0_SEG ## seg
-
-#define NBIO_BASE(seg) \
-	NBIO_BASE_INNER(seg)
-
-#define NBIO_SR(reg_name)\
-		.reg_name = NBIO_BASE(mm ## reg_name ## _BASE_IDX) + \
-					mm ## reg_name
-
-/* MMHUB */
-#define MMHUB_BASE_INNER(seg) \
-	MMHUB_BASE__INST0_SEG ## seg
-
-#define MMHUB_BASE(seg) \
-	MMHUB_BASE_INNER(seg)
-
-#define MMHUB_SR(reg_name)\
-		.reg_name = MMHUB_BASE(mmMM ## reg_name ## _BASE_IDX) + \
-					mmMM ## reg_name
-
-#define clk_src_regs(index, pllid)\
-[index] = {\
-	CS_COMMON_REG_LIST_DCN2_1(index, pllid),\
-}
-
-static const struct dce110_clk_src_regs clk_src_regs[] = {
-	clk_src_regs(0, A),
-	clk_src_regs(1, B),
-	clk_src_regs(2, C),
-	clk_src_regs(3, D),
-	clk_src_regs(4, E),
-};
-
-static const struct dce110_clk_src_shift cs_shift = {
-		CS_COMMON_MASK_SH_LIST_DCN2_0(__SHIFT)
-};
-
-static const struct dce110_clk_src_mask cs_mask = {
-		CS_COMMON_MASK_SH_LIST_DCN2_0(_MASK)
-};
-
-static const struct bios_registers bios_regs = {
-		NBIO_SR(BIOS_SCRATCH_3),
-		NBIO_SR(BIOS_SCRATCH_6)
-};
-
-static const struct dce_dmcu_registers dmcu_regs = {
-		DMCU_DCN20_REG_LIST()
-};
-
-static const struct dce_dmcu_shift dmcu_shift = {
-		DMCU_MASK_SH_LIST_DCN10(__SHIFT)
-};
-
-static const struct dce_dmcu_mask dmcu_mask = {
-		DMCU_MASK_SH_LIST_DCN10(_MASK)
-};
-
-static const struct dce_abm_registers abm_regs = {
-		ABM_DCN20_REG_LIST()
-};
-
-static const struct dce_abm_shift abm_shift = {
-		ABM_MASK_SH_LIST_DCN20(__SHIFT)
-};
-
-static const struct dce_abm_mask abm_mask = {
-		ABM_MASK_SH_LIST_DCN20(_MASK)
-};
-
-#define audio_regs(id)\
-[id] = {\
-		AUD_COMMON_REG_LIST(id)\
-}
-
-static const struct dce_audio_registers audio_regs[] = {
-	audio_regs(0),
-	audio_regs(1),
-	audio_regs(2),
-	audio_regs(3),
-	audio_regs(4),
-	audio_regs(5),
-};
-
-#define DCE120_AUD_COMMON_MASK_SH_LIST(mask_sh)\
-		SF(AZF0ENDPOINT0_AZALIA_F0_CODEC_ENDPOINT_INDEX, AZALIA_ENDPOINT_REG_INDEX, mask_sh),\
-		SF(AZF0ENDPOINT0_AZALIA_F0_CODEC_ENDPOINT_DATA, AZALIA_ENDPOINT_REG_DATA, mask_sh),\
-		AUD_COMMON_MASK_SH_LIST_BASE(mask_sh)
-
-static const struct dce_audio_shift audio_shift = {
-		DCE120_AUD_COMMON_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dce_audio_mask audio_mask = {
-		DCE120_AUD_COMMON_MASK_SH_LIST(_MASK)
-};
-
-static const struct dccg_registers dccg_regs = {
-		DCCG_COMMON_REG_LIST_DCN_BASE()
-};
-
-static const struct dccg_shift dccg_shift = {
-		DCCG_MASK_SH_LIST_DCN2_1(__SHIFT)
-};
-
-static const struct dccg_mask dccg_mask = {
-		DCCG_MASK_SH_LIST_DCN2_1(_MASK)
-};
-
-#define opp_regs(id)\
-[id] = {\
-	OPP_REG_LIST_DCN20(id),\
-}
-
-static const struct dcn20_opp_registers opp_regs[] = {
-	opp_regs(0),
-	opp_regs(1),
-	opp_regs(2),
-	opp_regs(3),
-	opp_regs(4),
-	opp_regs(5),
-};
-
-static const struct dcn20_opp_shift opp_shift = {
-		OPP_MASK_SH_LIST_DCN20(__SHIFT)
-};
-
-static const struct dcn20_opp_mask opp_mask = {
-		OPP_MASK_SH_LIST_DCN20(_MASK)
-};
-
-#define tg_regs(id)\
-[id] = {TG_COMMON_REG_LIST_DCN2_0(id)}
-
-static const struct dcn_optc_registers tg_regs[] = {
-	tg_regs(0),
-	tg_regs(1),
-	tg_regs(2),
-	tg_regs(3)
-};
-
-static const struct dcn_optc_shift tg_shift = {
-	TG_COMMON_MASK_SH_LIST_DCN2_0(__SHIFT)
-};
-
-static const struct dcn_optc_mask tg_mask = {
-	TG_COMMON_MASK_SH_LIST_DCN2_0(_MASK)
-};
-
-static const struct dcn20_mpc_registers mpc_regs = {
-		MPC_REG_LIST_DCN2_0(0),
-		MPC_REG_LIST_DCN2_0(1),
-		MPC_REG_LIST_DCN2_0(2),
-		MPC_REG_LIST_DCN2_0(3),
-		MPC_REG_LIST_DCN2_0(4),
-		MPC_REG_LIST_DCN2_0(5),
-		MPC_OUT_MUX_REG_LIST_DCN2_0(0),
-		MPC_OUT_MUX_REG_LIST_DCN2_0(1),
-		MPC_OUT_MUX_REG_LIST_DCN2_0(2),
-		MPC_OUT_MUX_REG_LIST_DCN2_0(3),
-		MPC_DBG_REG_LIST_DCN2_0()
-};
-
-static const struct dcn20_mpc_shift mpc_shift = {
-	MPC_COMMON_MASK_SH_LIST_DCN2_0(__SHIFT),
-	MPC_DEBUG_REG_LIST_SH_DCN20
-};
-
-static const struct dcn20_mpc_mask mpc_mask = {
-	MPC_COMMON_MASK_SH_LIST_DCN2_0(_MASK),
-	MPC_DEBUG_REG_LIST_MASK_DCN20
-};
-
-#define hubp_regs(id)\
-[id] = {\
-	HUBP_REG_LIST_DCN21(id)\
-}
-
-static const struct dcn_hubp2_registers hubp_regs[] = {
-		hubp_regs(0),
-		hubp_regs(1),
-		hubp_regs(2),
-		hubp_regs(3)
-};
-
-static const struct dcn_hubp2_shift hubp_shift = {
-		HUBP_MASK_SH_LIST_DCN21(__SHIFT)
-};
-
-static const struct dcn_hubp2_mask hubp_mask = {
-		HUBP_MASK_SH_LIST_DCN21(_MASK)
-};
-
-static const struct dcn_hubbub_registers hubbub_reg = {
-		HUBBUB_REG_LIST_DCN21()
-};
-
-static const struct dcn_hubbub_shift hubbub_shift = {
-		HUBBUB_MASK_SH_LIST_DCN21(__SHIFT)
-};
-
-static const struct dcn_hubbub_mask hubbub_mask = {
-		HUBBUB_MASK_SH_LIST_DCN21(_MASK)
-};
-
-
-#define vmid_regs(id)\
-[id] = {\
-		DCN20_VMID_REG_LIST(id)\
-}
-
-static const struct dcn_vmid_registers vmid_regs[] = {
-	vmid_regs(0),
-	vmid_regs(1),
-	vmid_regs(2),
-	vmid_regs(3),
-	vmid_regs(4),
-	vmid_regs(5),
-	vmid_regs(6),
-	vmid_regs(7),
-	vmid_regs(8),
-	vmid_regs(9),
-	vmid_regs(10),
-	vmid_regs(11),
-	vmid_regs(12),
-	vmid_regs(13),
-	vmid_regs(14),
-	vmid_regs(15)
-};
-
-static const struct dcn20_vmid_shift vmid_shifts = {
-		DCN20_VMID_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dcn20_vmid_mask vmid_masks = {
-		DCN20_VMID_MASK_SH_LIST(_MASK)
-};
-
-#define dsc_regsDCN20(id)\
-[id] = {\
-	DSC_REG_LIST_DCN20(id)\
-}
-
-static const struct dcn20_dsc_registers dsc_regs[] = {
-	dsc_regsDCN20(0),
-	dsc_regsDCN20(1),
-	dsc_regsDCN20(2),
-	dsc_regsDCN20(3),
-	dsc_regsDCN20(4),
-	dsc_regsDCN20(5)
-};
-
-static const struct dcn20_dsc_shift dsc_shift = {
-	DSC_REG_LIST_SH_MASK_DCN20(__SHIFT)
-};
-
-static const struct dcn20_dsc_mask dsc_mask = {
-	DSC_REG_LIST_SH_MASK_DCN20(_MASK)
-};
-
-#define ipp_regs(id)\
-[id] = {\
-	IPP_REG_LIST_DCN20(id),\
-}
-
-static const struct dcn10_ipp_registers ipp_regs[] = {
-	ipp_regs(0),
-	ipp_regs(1),
-	ipp_regs(2),
-	ipp_regs(3),
-};
-
-static const struct dcn10_ipp_shift ipp_shift = {
-		IPP_MASK_SH_LIST_DCN20(__SHIFT)
-};
-
-static const struct dcn10_ipp_mask ipp_mask = {
-		IPP_MASK_SH_LIST_DCN20(_MASK),
-};
-
-#define opp_regs(id)\
-[id] = {\
-	OPP_REG_LIST_DCN20(id),\
-}
-
-
-#define aux_engine_regs(id)\
-[id] = {\
-	AUX_COMMON_REG_LIST0(id), \
-	.AUXN_IMPCAL = 0, \
-	.AUXP_IMPCAL = 0, \
-	.AUX_RESET_MASK = DP_AUX0_AUX_CONTROL__AUX_RESET_MASK, \
-}
-
-static const struct dce110_aux_registers aux_engine_regs[] = {
-		aux_engine_regs(0),
-		aux_engine_regs(1),
-		aux_engine_regs(2),
-		aux_engine_regs(3),
-		aux_engine_regs(4),
-};
-
-#define tf_regs(id)\
-[id] = {\
-	TF_REG_LIST_DCN20(id),\
-	TF_REG_LIST_DCN20_COMMON_APPEND(id),\
-}
-
-static const struct dcn2_dpp_registers tf_regs[] = {
-	tf_regs(0),
-	tf_regs(1),
-	tf_regs(2),
-	tf_regs(3),
-};
-
-static const struct dcn2_dpp_shift tf_shift = {
-		TF_REG_LIST_SH_MASK_DCN20(__SHIFT),
-		TF_DEBUG_REG_LIST_SH_DCN20
-};
-
-static const struct dcn2_dpp_mask tf_mask = {
-		TF_REG_LIST_SH_MASK_DCN20(_MASK),
-		TF_DEBUG_REG_LIST_MASK_DCN20
-};
-
-#define stream_enc_regs(id)\
-[id] = {\
-	SE_DCN2_REG_LIST(id)\
-}
-
-static const struct dcn10_stream_enc_registers stream_enc_regs[] = {
-	stream_enc_regs(0),
-	stream_enc_regs(1),
-	stream_enc_regs(2),
-	stream_enc_regs(3),
-	stream_enc_regs(4),
-};
-
-static const struct dce110_aux_registers_shift aux_shift = {
-	DCN_AUX_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dce110_aux_registers_mask aux_mask = {
-	DCN_AUX_MASK_SH_LIST(_MASK)
-};
-
-static const struct dcn10_stream_encoder_shift se_shift = {
-		SE_COMMON_MASK_SH_LIST_DCN20(__SHIFT)
-};
-
-static const struct dcn10_stream_encoder_mask se_mask = {
-		SE_COMMON_MASK_SH_LIST_DCN20(_MASK)
-};
-
-static void dcn21_pp_smu_destroy(struct pp_smu_funcs **pp_smu);
-
-static int dcn21_populate_dml_pipes_from_context(
-		struct dc *dc,
-		struct dc_state *context,
-		display_e2e_pipe_params_st *pipes,
-		bool fast_validate);
-
-static struct input_pixel_processor *dcn21_ipp_create(
-	struct dc_context *ctx, uint32_t inst)
-{
-	struct dcn10_ipp *ipp =
-		kzalloc(sizeof(struct dcn10_ipp), GFP_KERNEL);
-
-	if (!ipp) {
-		BREAK_TO_DEBUGGER();
-		return NULL;
-	}
-
-	dcn20_ipp_construct(ipp, ctx, inst,
-			&ipp_regs[inst], &ipp_shift, &ipp_mask);
-	return &ipp->base;
-}
-
-static struct dpp *dcn21_dpp_create(
-	struct dc_context *ctx,
-	uint32_t inst)
-{
-	struct dcn20_dpp *dpp =
-		kzalloc(sizeof(struct dcn20_dpp), GFP_KERNEL);
-
-	if (!dpp)
-		return NULL;
-
-	if (dpp2_construct(dpp, ctx, inst,
-			&tf_regs[inst], &tf_shift, &tf_mask))
-		return &dpp->base;
-
-	BREAK_TO_DEBUGGER();
-	kfree(dpp);
-	return NULL;
-}
-
-static struct dce_aux *dcn21_aux_engine_create(
-	struct dc_context *ctx,
-	uint32_t inst)
-{
-	struct aux_engine_dce110 *aux_engine =
-		kzalloc(sizeof(struct aux_engine_dce110), GFP_KERNEL);
-
-	if (!aux_engine)
-		return NULL;
-
-	dce110_aux_engine_construct(aux_engine, ctx, inst,
-				    SW_AUX_TIMEOUT_PERIOD_MULTIPLIER * AUX_TIMEOUT_PERIOD,
-				    &aux_engine_regs[inst],
-					&aux_mask,
-					&aux_shift,
-					ctx->dc->caps.extended_aux_timeout_support);
-
-	return &aux_engine->base;
-}
-
-#define i2c_inst_regs(id) { I2C_HW_ENGINE_COMMON_REG_LIST(id) }
-
-static const struct dce_i2c_registers i2c_hw_regs[] = {
-		i2c_inst_regs(1),
-		i2c_inst_regs(2),
-		i2c_inst_regs(3),
-		i2c_inst_regs(4),
-		i2c_inst_regs(5),
-};
-
-static const struct dce_i2c_shift i2c_shifts = {
-		I2C_COMMON_MASK_SH_LIST_DCN2(__SHIFT)
-};
-
-static const struct dce_i2c_mask i2c_masks = {
-		I2C_COMMON_MASK_SH_LIST_DCN2(_MASK)
-};
-
-struct dce_i2c_hw *dcn21_i2c_hw_create(
-	struct dc_context *ctx,
-	uint32_t inst)
-{
-	struct dce_i2c_hw *dce_i2c_hw =
-		kzalloc(sizeof(struct dce_i2c_hw), GFP_KERNEL);
-
-	if (!dce_i2c_hw)
-		return NULL;
-
-	dcn2_i2c_hw_construct(dce_i2c_hw, ctx, inst,
-				    &i2c_hw_regs[inst], &i2c_shifts, &i2c_masks);
-
-	return dce_i2c_hw;
-}
-
-static const struct resource_caps res_cap_rn = {
-		.num_timing_generator = 4,
-		.num_opp = 4,
-		.num_video_plane = 4,
-		.num_audio = 4, // 4 audio endpoints.  4 audio streams
-		.num_stream_encoder = 5,
-		.num_pll = 5,  // maybe 3 because the last two used for USB-c
-		.num_dwb = 1,
-		.num_ddc = 5,
-		.num_vmid = 16,
-		.num_dsc = 3,
-};
-
-#ifdef DIAGS_BUILD
-static const struct resource_caps res_cap_rn_FPGA_4pipe = {
-		.num_timing_generator = 4,
-		.num_opp = 4,
-		.num_video_plane = 4,
-		.num_audio = 7,
-		.num_stream_encoder = 4,
-		.num_pll = 4,
-		.num_dwb = 1,
-		.num_ddc = 4,
-		.num_dsc = 0,
-};
-
-static const struct resource_caps res_cap_rn_FPGA_2pipe_dsc = {
-		.num_timing_generator = 2,
-		.num_opp = 2,
-		.num_video_plane = 2,
-		.num_audio = 7,
-		.num_stream_encoder = 2,
-		.num_pll = 4,
-		.num_dwb = 1,
-		.num_ddc = 4,
-		.num_dsc = 2,
-};
-#endif
-
-static const struct dc_plane_cap plane_cap = {
-	.type = DC_PLANE_TYPE_DCN_UNIVERSAL,
-	.blends_with_above = true,
-	.blends_with_below = true,
-	.per_pixel_alpha = true,
-
-	.pixel_format_support = {
-			.argb8888 = true,
-			.nv12 = true,
-			.fp16 = true,
-			.p010 = true
-	},
-
-	.max_upscale_factor = {
-			.argb8888 = 16000,
-			.nv12 = 16000,
-			.fp16 = 16000
-	},
-
-	.max_downscale_factor = {
-			.argb8888 = 250,
-			.nv12 = 250,
-			.fp16 = 250
-	},
-	64,
-	64
-};
-
-static const struct dc_debug_options debug_defaults_drv = {
-		.disable_dmcu = false,
-		.force_abm_enable = false,
-		.timing_trace = false,
-		.clock_trace = true,
-		.disable_pplib_clock_request = true,
-		.min_disp_clk_khz = 100000,
-		.pipe_split_policy = MPC_SPLIT_AVOID_MULT_DISP,
-		.force_single_disp_pipe_split = false,
-		.disable_dcc = DCC_ENABLE,
-		.vsr_support = true,
-		.performance_trace = false,
-		.max_downscale_src_width = 4096,
-		.disable_pplib_wm_range = false,
-		.scl_reset_length10 = true,
-		.sanity_checks = true,
-		.disable_48mhz_pwrdwn = false,
-		.usbc_combo_phy_reset_wa = true,
-		.dmub_command_table = true,
-		.use_max_lb = true,
-		.optimize_edp_link_rate = true
-};
-
-static const struct dc_debug_options debug_defaults_diags = {
-		.disable_dmcu = false,
-		.force_abm_enable = false,
-		.timing_trace = true,
-		.clock_trace = true,
-		.disable_dpp_power_gate = true,
-		.disable_hubp_power_gate = true,
-		.disable_clock_gate = true,
-		.disable_pplib_clock_request = true,
-		.disable_pplib_wm_range = true,
-		.disable_stutter = true,
-		.disable_48mhz_pwrdwn = true,
-		.disable_psr = true,
-		.enable_tri_buf = true,
-		.use_max_lb = true
-};
-
-enum dcn20_clk_src_array_id {
-	DCN20_CLK_SRC_PLL0,
-	DCN20_CLK_SRC_PLL1,
-	DCN20_CLK_SRC_PLL2,
-	DCN20_CLK_SRC_PLL3,
-	DCN20_CLK_SRC_PLL4,
-	DCN20_CLK_SRC_TOTAL_DCN21
-};
-
-static void dcn21_resource_destruct(struct dcn21_resource_pool *pool)
-{
-	unsigned int i;
-
-	for (i = 0; i < pool->base.stream_enc_count; i++) {
-		if (pool->base.stream_enc[i] != NULL) {
-			kfree(DCN10STRENC_FROM_STRENC(pool->base.stream_enc[i]));
-			pool->base.stream_enc[i] = NULL;
-		}
-	}
-
-	for (i = 0; i < pool->base.res_cap->num_dsc; i++) {
-		if (pool->base.dscs[i] != NULL)
-			dcn20_dsc_destroy(&pool->base.dscs[i]);
-	}
-
-	if (pool->base.mpc != NULL) {
-		kfree(TO_DCN20_MPC(pool->base.mpc));
-		pool->base.mpc = NULL;
-	}
-	if (pool->base.hubbub != NULL) {
-		kfree(pool->base.hubbub);
-		pool->base.hubbub = NULL;
-	}
-	for (i = 0; i < pool->base.pipe_count; i++) {
-		if (pool->base.dpps[i] != NULL)
-			dcn20_dpp_destroy(&pool->base.dpps[i]);
-
-		if (pool->base.ipps[i] != NULL)
-			pool->base.ipps[i]->funcs->ipp_destroy(&pool->base.ipps[i]);
-
-		if (pool->base.hubps[i] != NULL) {
-			kfree(TO_DCN20_HUBP(pool->base.hubps[i]));
-			pool->base.hubps[i] = NULL;
-		}
-
-		if (pool->base.irqs != NULL) {
-			dal_irq_service_destroy(&pool->base.irqs);
-		}
-	}
-
-	for (i = 0; i < pool->base.res_cap->num_ddc; i++) {
-		if (pool->base.engines[i] != NULL)
-			dce110_engine_destroy(&pool->base.engines[i]);
-		if (pool->base.hw_i2cs[i] != NULL) {
-			kfree(pool->base.hw_i2cs[i]);
-			pool->base.hw_i2cs[i] = NULL;
-		}
-		if (pool->base.sw_i2cs[i] != NULL) {
-			kfree(pool->base.sw_i2cs[i]);
-			pool->base.sw_i2cs[i] = NULL;
-		}
-	}
-
-	for (i = 0; i < pool->base.res_cap->num_opp; i++) {
-		if (pool->base.opps[i] != NULL)
-			pool->base.opps[i]->funcs->opp_destroy(&pool->base.opps[i]);
-	}
-
-	for (i = 0; i < pool->base.res_cap->num_timing_generator; i++) {
-		if (pool->base.timing_generators[i] != NULL)	{
-			kfree(DCN10TG_FROM_TG(pool->base.timing_generators[i]));
-			pool->base.timing_generators[i] = NULL;
-		}
-	}
-
-	for (i = 0; i < pool->base.res_cap->num_dwb; i++) {
-		if (pool->base.dwbc[i] != NULL) {
-			kfree(TO_DCN20_DWBC(pool->base.dwbc[i]));
-			pool->base.dwbc[i] = NULL;
-		}
-		if (pool->base.mcif_wb[i] != NULL) {
-			kfree(TO_DCN20_MMHUBBUB(pool->base.mcif_wb[i]));
-			pool->base.mcif_wb[i] = NULL;
-		}
-	}
-
-	for (i = 0; i < pool->base.audio_count; i++) {
-		if (pool->base.audios[i])
-			dce_aud_destroy(&pool->base.audios[i]);
-	}
-
-	for (i = 0; i < pool->base.clk_src_count; i++) {
-		if (pool->base.clock_sources[i] != NULL) {
-			dcn20_clock_source_destroy(&pool->base.clock_sources[i]);
-			pool->base.clock_sources[i] = NULL;
-		}
-	}
-
-	if (pool->base.dp_clock_source != NULL) {
-		dcn20_clock_source_destroy(&pool->base.dp_clock_source);
-		pool->base.dp_clock_source = NULL;
-	}
-
-	if (pool->base.abm != NULL) {
-		if (pool->base.abm->ctx->dc->config.disable_dmcu)
-			dmub_abm_destroy(&pool->base.abm);
-		else
-			dce_abm_destroy(&pool->base.abm);
-	}
-
-	if (pool->base.dmcu != NULL)
-		dce_dmcu_destroy(&pool->base.dmcu);
-
-	if (pool->base.psr != NULL)
-		dmub_psr_destroy(&pool->base.psr);
-
-	if (pool->base.dccg != NULL)
-		dcn_dccg_destroy(&pool->base.dccg);
-
-	if (pool->base.pp_smu != NULL)
-		dcn21_pp_smu_destroy(&pool->base.pp_smu);
-}
-
-
-static void calculate_wm_set_for_vlevel(
-		int vlevel,
-		struct wm_range_table_entry *table_entry,
-		struct dcn_watermarks *wm_set,
-		struct display_mode_lib *dml,
-		display_e2e_pipe_params_st *pipes,
-		int pipe_cnt)
-{
-	double dram_clock_change_latency_cached = dml->soc.dram_clock_change_latency_us;
-
-	ASSERT(vlevel < dml->soc.num_states);
-	/* only pipe 0 is read for voltage and dcf/soc clocks */
-	pipes[0].clks_cfg.voltage = vlevel;
-	pipes[0].clks_cfg.dcfclk_mhz = dml->soc.clock_limits[vlevel].dcfclk_mhz;
-	pipes[0].clks_cfg.socclk_mhz = dml->soc.clock_limits[vlevel].socclk_mhz;
-
-	dml->soc.dram_clock_change_latency_us = table_entry->pstate_latency_us;
-	dml->soc.sr_exit_time_us = table_entry->sr_exit_time_us;
-	dml->soc.sr_enter_plus_exit_time_us = table_entry->sr_enter_plus_exit_time_us;
-
-	wm_set->urgent_ns = get_wm_urgent(dml, pipes, pipe_cnt) * 1000;
-	wm_set->cstate_pstate.cstate_enter_plus_exit_ns = get_wm_stutter_enter_exit(dml, pipes, pipe_cnt) * 1000;
-	wm_set->cstate_pstate.cstate_exit_ns = get_wm_stutter_exit(dml, pipes, pipe_cnt) * 1000;
-	wm_set->cstate_pstate.pstate_change_ns = get_wm_dram_clock_change(dml, pipes, pipe_cnt) * 1000;
-	wm_set->pte_meta_urgent_ns = get_wm_memory_trip(dml, pipes, pipe_cnt) * 1000;
-	wm_set->frac_urg_bw_nom = get_fraction_of_urgent_bandwidth(dml, pipes, pipe_cnt) * 1000;
-	wm_set->frac_urg_bw_flip = get_fraction_of_urgent_bandwidth_imm_flip(dml, pipes, pipe_cnt) * 1000;
-	wm_set->urgent_latency_ns = get_urgent_latency(dml, pipes, pipe_cnt) * 1000;
-	dml->soc.dram_clock_change_latency_us = dram_clock_change_latency_cached;
-
-}
-
-static void patch_bounding_box(struct dc *dc, struct _vcs_dpi_soc_bounding_box_st *bb)
-{
-	int i;
-
-	if (dc->bb_overrides.sr_exit_time_ns) {
-		for (i = 0; i < WM_SET_COUNT; i++) {
-			  dc->clk_mgr->bw_params->wm_table.entries[i].sr_exit_time_us =
-					  dc->bb_overrides.sr_exit_time_ns / 1000.0;
-		}
-	}
-
-	if (dc->bb_overrides.sr_enter_plus_exit_time_ns) {
-		for (i = 0; i < WM_SET_COUNT; i++) {
-			  dc->clk_mgr->bw_params->wm_table.entries[i].sr_enter_plus_exit_time_us =
-					  dc->bb_overrides.sr_enter_plus_exit_time_ns / 1000.0;
-		}
-	}
-
-	if (dc->bb_overrides.urgent_latency_ns) {
-		bb->urgent_latency_us = dc->bb_overrides.urgent_latency_ns / 1000.0;
-	}
-
-	if (dc->bb_overrides.dram_clock_change_latency_ns) {
-		for (i = 0; i < WM_SET_COUNT; i++) {
-			dc->clk_mgr->bw_params->wm_table.entries[i].pstate_latency_us =
-				dc->bb_overrides.dram_clock_change_latency_ns / 1000.0;
-		}
-	}
-}
-
-void dcn21_calculate_wm(
-		struct dc *dc, struct dc_state *context,
-		display_e2e_pipe_params_st *pipes,
-		int *out_pipe_cnt,
-		int *pipe_split_from,
-		int vlevel_req,
-		bool fast_validate)
-{
-	int pipe_cnt, i, pipe_idx;
-	int vlevel, vlevel_max;
-	struct wm_range_table_entry *table_entry;
-	struct clk_bw_params *bw_params = dc->clk_mgr->bw_params;
-
-	ASSERT(bw_params);
-
-	patch_bounding_box(dc, &context->bw_ctx.dml.soc);
-
-	for (i = 0, pipe_idx = 0, pipe_cnt = 0; i < dc->res_pool->pipe_count; i++) {
-			if (!context->res_ctx.pipe_ctx[i].stream)
-				continue;
-
-			pipes[pipe_cnt].clks_cfg.refclk_mhz = dc->res_pool->ref_clocks.dchub_ref_clock_inKhz / 1000.0;
-			pipes[pipe_cnt].clks_cfg.dispclk_mhz = context->bw_ctx.dml.vba.RequiredDISPCLK[vlevel_req][context->bw_ctx.dml.vba.maxMpcComb];
-
-			if (pipe_split_from[i] < 0) {
-				pipes[pipe_cnt].clks_cfg.dppclk_mhz =
-						context->bw_ctx.dml.vba.RequiredDPPCLK[vlevel_req][context->bw_ctx.dml.vba.maxMpcComb][pipe_idx];
-				if (context->bw_ctx.dml.vba.BlendingAndTiming[pipe_idx] == pipe_idx)
-					pipes[pipe_cnt].pipe.dest.odm_combine =
-							context->bw_ctx.dml.vba.ODMCombineEnablePerState[vlevel_req][pipe_idx];
-				else
-					pipes[pipe_cnt].pipe.dest.odm_combine = 0;
-				pipe_idx++;
-			} else {
-				pipes[pipe_cnt].clks_cfg.dppclk_mhz =
-						context->bw_ctx.dml.vba.RequiredDPPCLK[vlevel_req][context->bw_ctx.dml.vba.maxMpcComb][pipe_split_from[i]];
-				if (context->bw_ctx.dml.vba.BlendingAndTiming[pipe_split_from[i]] == pipe_split_from[i])
-					pipes[pipe_cnt].pipe.dest.odm_combine =
-							context->bw_ctx.dml.vba.ODMCombineEnablePerState[vlevel_req][pipe_split_from[i]];
-				else
-					pipes[pipe_cnt].pipe.dest.odm_combine = 0;
-			}
-			pipe_cnt++;
-	}
-
-	if (pipe_cnt != pipe_idx) {
-		if (dc->res_pool->funcs->populate_dml_pipes)
-			pipe_cnt = dc->res_pool->funcs->populate_dml_pipes(dc,
-				context, pipes, fast_validate);
-		else
-			pipe_cnt = dcn21_populate_dml_pipes_from_context(dc,
-				context, pipes, fast_validate);
-	}
-
-	*out_pipe_cnt = pipe_cnt;
-
-	vlevel_max = bw_params->clk_table.num_entries - 1;
-
-
-	/* WM Set D */
-	table_entry = &bw_params->wm_table.entries[WM_D];
-	if (table_entry->wm_type == WM_TYPE_RETRAINING)
-		vlevel = 0;
-	else
-		vlevel = vlevel_max;
-	calculate_wm_set_for_vlevel(vlevel, table_entry, &context->bw_ctx.bw.dcn.watermarks.d,
-						&context->bw_ctx.dml, pipes, pipe_cnt);
-	/* WM Set C */
-	table_entry = &bw_params->wm_table.entries[WM_C];
-	vlevel = MIN(MAX(vlevel_req, 3), vlevel_max);
-	calculate_wm_set_for_vlevel(vlevel, table_entry, &context->bw_ctx.bw.dcn.watermarks.c,
-						&context->bw_ctx.dml, pipes, pipe_cnt);
-	/* WM Set B */
-	table_entry = &bw_params->wm_table.entries[WM_B];
-	vlevel = MIN(MAX(vlevel_req, 2), vlevel_max);
-	calculate_wm_set_for_vlevel(vlevel, table_entry, &context->bw_ctx.bw.dcn.watermarks.b,
-						&context->bw_ctx.dml, pipes, pipe_cnt);
-
-	/* WM Set A */
-	table_entry = &bw_params->wm_table.entries[WM_A];
-	vlevel = MIN(vlevel_req, vlevel_max);
-	calculate_wm_set_for_vlevel(vlevel, table_entry, &context->bw_ctx.bw.dcn.watermarks.a,
-						&context->bw_ctx.dml, pipes, pipe_cnt);
-}
-
-
-static bool dcn21_fast_validate_bw(
-		struct dc *dc,
-		struct dc_state *context,
-		display_e2e_pipe_params_st *pipes,
-		int *pipe_cnt_out,
-		int *pipe_split_from,
-		int *vlevel_out,
-		bool fast_validate)
-{
-	bool out = false;
-	int split[MAX_PIPES] = { 0 };
-	int pipe_cnt, i, pipe_idx, vlevel;
-
-	ASSERT(pipes);
-	if (!pipes)
-		return false;
-
-	dcn20_merge_pipes_for_validate(dc, context);
-
-	pipe_cnt = dc->res_pool->funcs->populate_dml_pipes(dc, context, pipes, fast_validate);
-
-	*pipe_cnt_out = pipe_cnt;
-
-	if (!pipe_cnt) {
-		out = true;
-		goto validate_out;
-	}
-	/*
-	 * DML favors voltage over p-state, but we're more interested in
-	 * supporting p-state over voltage. We can't support p-state in
-	 * prefetch mode > 0 so try capping the prefetch mode to start.
-	 */
-	context->bw_ctx.dml.soc.allow_dram_self_refresh_or_dram_clock_change_in_vblank =
-				dm_allow_self_refresh_and_mclk_switch;
-	vlevel = dml_get_voltage_level(&context->bw_ctx.dml, pipes, pipe_cnt);
-
-	if (vlevel > context->bw_ctx.dml.soc.num_states) {
-		/*
-		 * If mode is unsupported or there's still no p-state support then
-		 * fall back to favoring voltage.
-		 *
-		 * We don't actually support prefetch mode 2, so require that we
-		 * at least support prefetch mode 1.
-		 */
-		context->bw_ctx.dml.soc.allow_dram_self_refresh_or_dram_clock_change_in_vblank =
-					dm_allow_self_refresh;
-		vlevel = dml_get_voltage_level(&context->bw_ctx.dml, pipes, pipe_cnt);
-		if (vlevel > context->bw_ctx.dml.soc.num_states)
-			goto validate_fail;
-	}
-
-	vlevel = dcn20_validate_apply_pipe_split_flags(dc, context, vlevel, split, NULL);
-
-	for (i = 0, pipe_idx = 0; i < dc->res_pool->pipe_count; i++) {
-		struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
-		struct pipe_ctx *mpo_pipe = pipe->bottom_pipe;
-		struct vba_vars_st *vba = &context->bw_ctx.dml.vba;
-
-		if (!pipe->stream)
-			continue;
-
-		/* We only support full screen mpo with ODM */
-		if (vba->ODMCombineEnabled[vba->pipe_plane[pipe_idx]] != dm_odm_combine_mode_disabled
-				&& pipe->plane_state && mpo_pipe
-				&& memcmp(&mpo_pipe->plane_res.scl_data.recout,
-						&pipe->plane_res.scl_data.recout,
-						sizeof(struct rect)) != 0) {
-			ASSERT(mpo_pipe->plane_state != pipe->plane_state);
-			goto validate_fail;
-		}
-		pipe_idx++;
-	}
-
-	/*initialize pipe_just_split_from to invalid idx*/
-	for (i = 0; i < MAX_PIPES; i++)
-		pipe_split_from[i] = -1;
-
-	for (i = 0, pipe_idx = -1; i < dc->res_pool->pipe_count; i++) {
-		struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
-		struct pipe_ctx *hsplit_pipe = pipe->bottom_pipe;
-
-		if (!pipe->stream || pipe_split_from[i] >= 0)
-			continue;
-
-		pipe_idx++;
-
-		if (!pipe->top_pipe && !pipe->plane_state && context->bw_ctx.dml.vba.ODMCombineEnabled[pipe_idx]) {
-			hsplit_pipe = dcn20_find_secondary_pipe(dc, &context->res_ctx, dc->res_pool, pipe);
-			ASSERT(hsplit_pipe);
-			if (!dcn20_split_stream_for_odm(
-					dc, &context->res_ctx,
-					pipe, hsplit_pipe))
-				goto validate_fail;
-			pipe_split_from[hsplit_pipe->pipe_idx] = pipe_idx;
-			dcn20_build_mapped_resource(dc, context, pipe->stream);
-		}
-
-		if (!pipe->plane_state)
-			continue;
-		/* Skip 2nd half of already split pipe */
-		if (pipe->top_pipe && pipe->plane_state == pipe->top_pipe->plane_state)
-			continue;
-
-		if (split[i] == 2) {
-			if (!hsplit_pipe || hsplit_pipe->plane_state != pipe->plane_state) {
-				/* pipe not split previously needs split */
-				hsplit_pipe = dcn20_find_secondary_pipe(dc, &context->res_ctx, dc->res_pool, pipe);
-				ASSERT(hsplit_pipe);
-				if (!hsplit_pipe) {
-					context->bw_ctx.dml.vba.RequiredDPPCLK[vlevel][context->bw_ctx.dml.vba.maxMpcComb][pipe_idx] *= 2;
-					continue;
-				}
-				if (context->bw_ctx.dml.vba.ODMCombineEnabled[pipe_idx]) {
-					if (!dcn20_split_stream_for_odm(
-							dc, &context->res_ctx,
-							pipe, hsplit_pipe))
-						goto validate_fail;
-					dcn20_build_mapped_resource(dc, context, pipe->stream);
-				} else {
-					dcn20_split_stream_for_mpc(
-							&context->res_ctx, dc->res_pool,
-							pipe, hsplit_pipe);
-					resource_build_scaling_params(pipe);
-					resource_build_scaling_params(hsplit_pipe);
-				}
-				pipe_split_from[hsplit_pipe->pipe_idx] = pipe_idx;
-			}
-		} else if (hsplit_pipe && hsplit_pipe->plane_state == pipe->plane_state) {
-			/* merge should already have been done */
-			ASSERT(0);
-		}
-	}
-	/* Actual dsc count per stream dsc validation*/
-	if (!dcn20_validate_dsc(dc, context)) {
-		context->bw_ctx.dml.vba.ValidationStatus[context->bw_ctx.dml.vba.soc.num_states] =
-				DML_FAIL_DSC_VALIDATION_FAILURE;
-		goto validate_fail;
-	}
-
-	*vlevel_out = vlevel;
-
-	out = true;
-	goto validate_out;
-
-validate_fail:
-	out = false;
-
-validate_out:
-	return out;
-}
-
-static noinline bool dcn21_validate_bandwidth_fp(struct dc *dc,
-		struct dc_state *context, bool fast_validate)
-{
-	bool out = false;
-
-	BW_VAL_TRACE_SETUP();
-
-	int vlevel = 0;
-	int pipe_split_from[MAX_PIPES];
-	int pipe_cnt = 0;
-	display_e2e_pipe_params_st *pipes = kzalloc(dc->res_pool->pipe_count * sizeof(display_e2e_pipe_params_st), GFP_ATOMIC);
-	DC_LOGGER_INIT(dc->ctx->logger);
-
-	BW_VAL_TRACE_COUNT();
-
-	/*Unsafe due to current pipe merge and split logic*/
-	ASSERT(context != dc->current_state);
-
-	out = dcn21_fast_validate_bw(dc, context, pipes, &pipe_cnt, pipe_split_from, &vlevel, fast_validate);
-
-	if (pipe_cnt == 0)
-		goto validate_out;
-
-	if (!out)
-		goto validate_fail;
-
-	BW_VAL_TRACE_END_VOLTAGE_LEVEL();
-
-	if (fast_validate) {
-		BW_VAL_TRACE_SKIP(fast);
-		goto validate_out;
-	}
-
-	dcn21_calculate_wm(dc, context, pipes, &pipe_cnt, pipe_split_from, vlevel, fast_validate);
-	dcn20_calculate_dlg_params(dc, context, pipes, pipe_cnt, vlevel);
-
-	BW_VAL_TRACE_END_WATERMARKS();
-
-	goto validate_out;
-
-validate_fail:
-	DC_LOG_WARNING("Mode Validation Warning: %s failed validation.\n",
-		dml_get_status_message(context->bw_ctx.dml.vba.ValidationStatus[context->bw_ctx.dml.vba.soc.num_states]));
-
-	BW_VAL_TRACE_SKIP(fail);
-	out = false;
-
-validate_out:
-	kfree(pipes);
-
-	BW_VAL_TRACE_FINISH();
-
-	return out;
-}
-
-/*
- * Some of the functions further below use the FPU, so we need to wrap this
- * with DC_FP_START()/DC_FP_END(). Use the same approach as for
- * dcn20_validate_bandwidth in dcn20_resource.c.
- */
-bool dcn21_validate_bandwidth(struct dc *dc, struct dc_state *context,
-		bool fast_validate)
-{
-	bool voltage_supported;
-	DC_FP_START();
-	voltage_supported = dcn21_validate_bandwidth_fp(dc, context, fast_validate);
-	DC_FP_END();
-	return voltage_supported;
-}
-
-static void dcn21_destroy_resource_pool(struct resource_pool **pool)
-{
-	struct dcn21_resource_pool *dcn21_pool = TO_DCN21_RES_POOL(*pool);
-
-	dcn21_resource_destruct(dcn21_pool);
-	kfree(dcn21_pool);
-	*pool = NULL;
-}
-
-static struct clock_source *dcn21_clock_source_create(
-		struct dc_context *ctx,
-		struct dc_bios *bios,
-		enum clock_source_id id,
-		const struct dce110_clk_src_regs *regs,
-		bool dp_clk_src)
-{
-	struct dce110_clk_src *clk_src =
-		kzalloc(sizeof(struct dce110_clk_src), GFP_KERNEL);
-
-	if (!clk_src)
-		return NULL;
-
-	if (dcn20_clk_src_construct(clk_src, ctx, bios, id,
-			regs, &cs_shift, &cs_mask)) {
-		clk_src->base.dp_clk_src = dp_clk_src;
-		return &clk_src->base;
-	}
-
-	BREAK_TO_DEBUGGER();
-	return NULL;
-}
-
-static struct hubp *dcn21_hubp_create(
-	struct dc_context *ctx,
-	uint32_t inst)
-{
-	struct dcn21_hubp *hubp21 =
-		kzalloc(sizeof(struct dcn21_hubp), GFP_KERNEL);
-
-	if (!hubp21)
-		return NULL;
-
-	if (hubp21_construct(hubp21, ctx, inst,
-			&hubp_regs[inst], &hubp_shift, &hubp_mask))
-		return &hubp21->base;
-
-	BREAK_TO_DEBUGGER();
-	kfree(hubp21);
-	return NULL;
-}
-
-static struct hubbub *dcn21_hubbub_create(struct dc_context *ctx)
-{
-	int i;
-
-	struct dcn20_hubbub *hubbub = kzalloc(sizeof(struct dcn20_hubbub),
-					  GFP_KERNEL);
-
-	if (!hubbub)
-		return NULL;
-
-	hubbub21_construct(hubbub, ctx,
-			&hubbub_reg,
-			&hubbub_shift,
-			&hubbub_mask);
-
-	for (i = 0; i < res_cap_rn.num_vmid; i++) {
-		struct dcn20_vmid *vmid = &hubbub->vmid[i];
-
-		vmid->ctx = ctx;
-
-		vmid->regs = &vmid_regs[i];
-		vmid->shifts = &vmid_shifts;
-		vmid->masks = &vmid_masks;
-	}
-	hubbub->num_vmid = res_cap_rn.num_vmid;
-
-	return &hubbub->base;
-}
-
-struct output_pixel_processor *dcn21_opp_create(
-	struct dc_context *ctx, uint32_t inst)
-{
-	struct dcn20_opp *opp =
-		kzalloc(sizeof(struct dcn20_opp), GFP_KERNEL);
-
-	if (!opp) {
-		BREAK_TO_DEBUGGER();
-		return NULL;
-	}
-
-	dcn20_opp_construct(opp, ctx, inst,
-			&opp_regs[inst], &opp_shift, &opp_mask);
-	return &opp->base;
-}
-
-struct timing_generator *dcn21_timing_generator_create(
-		struct dc_context *ctx,
-		uint32_t instance)
-{
-	struct optc *tgn10 =
-		kzalloc(sizeof(struct optc), GFP_KERNEL);
-
-	if (!tgn10)
-		return NULL;
-
-	tgn10->base.inst = instance;
-	tgn10->base.ctx = ctx;
-
-	tgn10->tg_regs = &tg_regs[instance];
-	tgn10->tg_shift = &tg_shift;
-	tgn10->tg_mask = &tg_mask;
-
-	dcn20_timing_generator_init(tgn10);
-
-	return &tgn10->base;
-}
-
-struct mpc *dcn21_mpc_create(struct dc_context *ctx)
-{
-	struct dcn20_mpc *mpc20 = kzalloc(sizeof(struct dcn20_mpc),
-					  GFP_KERNEL);
-
-	if (!mpc20)
-		return NULL;
-
-	dcn20_mpc_construct(mpc20, ctx,
-			&mpc_regs,
-			&mpc_shift,
-			&mpc_mask,
-			6);
-
-	return &mpc20->base;
-}
-
-static void read_dce_straps(
-	struct dc_context *ctx,
-	struct resource_straps *straps)
-{
-	generic_reg_get(ctx, mmDC_PINSTRAPS + BASE(mmDC_PINSTRAPS_BASE_IDX),
-		FN(DC_PINSTRAPS, DC_PINSTRAPS_AUDIO), &straps->dc_pinstraps_audio);
-
-}
-
-
-struct display_stream_compressor *dcn21_dsc_create(
-	struct dc_context *ctx, uint32_t inst)
-{
-	struct dcn20_dsc *dsc =
-		kzalloc(sizeof(struct dcn20_dsc), GFP_KERNEL);
-
-	if (!dsc) {
-		BREAK_TO_DEBUGGER();
-		return NULL;
-	}
-
-	dsc2_construct(dsc, ctx, inst, &dsc_regs[inst], &dsc_shift, &dsc_mask);
-	return &dsc->base;
-}
-
-static struct _vcs_dpi_voltage_scaling_st construct_low_pstate_lvl(struct clk_limit_table *clk_table, unsigned int high_voltage_lvl)
-{
-	struct _vcs_dpi_voltage_scaling_st low_pstate_lvl;
-	int i;
-
-	low_pstate_lvl.state = 1;
-	low_pstate_lvl.dcfclk_mhz = clk_table->entries[0].dcfclk_mhz;
-	low_pstate_lvl.fabricclk_mhz = clk_table->entries[0].fclk_mhz;
-	low_pstate_lvl.socclk_mhz = clk_table->entries[0].socclk_mhz;
-	low_pstate_lvl.dram_speed_mts = clk_table->entries[0].memclk_mhz * 2;
-
-	low_pstate_lvl.dispclk_mhz = dcn2_1_soc.clock_limits[high_voltage_lvl].dispclk_mhz;
-	low_pstate_lvl.dppclk_mhz = dcn2_1_soc.clock_limits[high_voltage_lvl].dppclk_mhz;
-	low_pstate_lvl.dram_bw_per_chan_gbps = dcn2_1_soc.clock_limits[high_voltage_lvl].dram_bw_per_chan_gbps;
-	low_pstate_lvl.dscclk_mhz = dcn2_1_soc.clock_limits[high_voltage_lvl].dscclk_mhz;
-	low_pstate_lvl.dtbclk_mhz = dcn2_1_soc.clock_limits[high_voltage_lvl].dtbclk_mhz;
-	low_pstate_lvl.phyclk_d18_mhz = dcn2_1_soc.clock_limits[high_voltage_lvl].phyclk_d18_mhz;
-	low_pstate_lvl.phyclk_mhz = dcn2_1_soc.clock_limits[high_voltage_lvl].phyclk_mhz;
-
-	for (i = clk_table->num_entries; i > 1; i--)
-		clk_table->entries[i] = clk_table->entries[i-1];
-	clk_table->entries[1] = clk_table->entries[0];
-	clk_table->num_entries++;
-
-	return low_pstate_lvl;
-}
-
-static void update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params)
-{
-	struct dcn21_resource_pool *pool = TO_DCN21_RES_POOL(dc->res_pool);
-	struct clk_limit_table *clk_table = &bw_params->clk_table;
-	struct _vcs_dpi_voltage_scaling_st clock_limits[DC__VOLTAGE_STATES];
-	unsigned int i, closest_clk_lvl = 0, k = 0;
-	int j;
-
-	dcn2_1_ip.max_num_otg = pool->base.res_cap->num_timing_generator;
-	dcn2_1_ip.max_num_dpp = pool->base.pipe_count;
-	dcn2_1_soc.num_chans = bw_params->num_channels;
-
-	ASSERT(clk_table->num_entries);
-	/* Copy dcn2_1_soc.clock_limits to clock_limits to avoid copying over null states later */
-	for (i = 0; i < dcn2_1_soc.num_states + 1; i++) {
-		clock_limits[i] = dcn2_1_soc.clock_limits[i];
-	}
-
-	for (i = 0; i < clk_table->num_entries; i++) {
-		/* loop backwards*/
-		for (closest_clk_lvl = 0, j = dcn2_1_soc.num_states - 1; j >= 0; j--) {
-			if ((unsigned int) dcn2_1_soc.clock_limits[j].dcfclk_mhz <= clk_table->entries[i].dcfclk_mhz) {
-				closest_clk_lvl = j;
-				break;
-			}
-		}
-
-		/* clk_table[1] is reserved for min DF PState.  skip here to fill in later. */
-		if (i == 1)
-			k++;
-
-		clock_limits[k].state = k;
-		clock_limits[k].dcfclk_mhz = clk_table->entries[i].dcfclk_mhz;
-		clock_limits[k].fabricclk_mhz = clk_table->entries[i].fclk_mhz;
-		clock_limits[k].socclk_mhz = clk_table->entries[i].socclk_mhz;
-		clock_limits[k].dram_speed_mts = clk_table->entries[i].memclk_mhz * 2;
-
-		clock_limits[k].dispclk_mhz = dcn2_1_soc.clock_limits[closest_clk_lvl].dispclk_mhz;
-		clock_limits[k].dppclk_mhz = dcn2_1_soc.clock_limits[closest_clk_lvl].dppclk_mhz;
-		clock_limits[k].dram_bw_per_chan_gbps = dcn2_1_soc.clock_limits[closest_clk_lvl].dram_bw_per_chan_gbps;
-		clock_limits[k].dscclk_mhz = dcn2_1_soc.clock_limits[closest_clk_lvl].dscclk_mhz;
-		clock_limits[k].dtbclk_mhz = dcn2_1_soc.clock_limits[closest_clk_lvl].dtbclk_mhz;
-		clock_limits[k].phyclk_d18_mhz = dcn2_1_soc.clock_limits[closest_clk_lvl].phyclk_d18_mhz;
-		clock_limits[k].phyclk_mhz = dcn2_1_soc.clock_limits[closest_clk_lvl].phyclk_mhz;
-
-		k++;
-	}
-	for (i = 0; i < clk_table->num_entries + 1; i++)
-		dcn2_1_soc.clock_limits[i] = clock_limits[i];
-	if (clk_table->num_entries) {
-		dcn2_1_soc.num_states = clk_table->num_entries + 1;
-		/* fill in min DF PState */
-		dcn2_1_soc.clock_limits[1] = construct_low_pstate_lvl(clk_table, closest_clk_lvl);
-		/* duplicate last level */
-		dcn2_1_soc.clock_limits[dcn2_1_soc.num_states] = dcn2_1_soc.clock_limits[dcn2_1_soc.num_states - 1];
-		dcn2_1_soc.clock_limits[dcn2_1_soc.num_states].state = dcn2_1_soc.num_states;
-	}
-
-	dml_init_instance(&dc->dml, &dcn2_1_soc, &dcn2_1_ip, DML_PROJECT_DCN21);
-}
-
-static struct pp_smu_funcs *dcn21_pp_smu_create(struct dc_context *ctx)
-{
-	struct pp_smu_funcs *pp_smu = kzalloc(sizeof(*pp_smu), GFP_KERNEL);
-
-	if (!pp_smu)
-		return pp_smu;
-
-	dm_pp_get_funcs(ctx, pp_smu);
-
-	if (pp_smu->ctx.ver != PP_SMU_VER_RN)
-		pp_smu = memset(pp_smu, 0, sizeof(struct pp_smu_funcs));
-
-
-	return pp_smu;
-}
-
-static void dcn21_pp_smu_destroy(struct pp_smu_funcs **pp_smu)
-{
-	if (pp_smu && *pp_smu) {
-		kfree(*pp_smu);
-		*pp_smu = NULL;
-	}
-}
-
-static struct audio *dcn21_create_audio(
-		struct dc_context *ctx, unsigned int inst)
-{
-	return dce_audio_create(ctx, inst,
-			&audio_regs[inst], &audio_shift, &audio_mask);
-}
-
-static struct dc_cap_funcs cap_funcs = {
-	.get_dcc_compression_cap = dcn20_get_dcc_compression_cap
-};
-
-struct stream_encoder *dcn21_stream_encoder_create(
-	enum engine_id eng_id,
-	struct dc_context *ctx)
-{
-	struct dcn10_stream_encoder *enc1 =
-		kzalloc(sizeof(struct dcn10_stream_encoder), GFP_KERNEL);
-
-	if (!enc1)
-		return NULL;
-
-	dcn20_stream_encoder_construct(enc1, ctx, ctx->dc_bios, eng_id,
-					&stream_enc_regs[eng_id],
-					&se_shift, &se_mask);
-
-	return &enc1->base;
-}
-
-static const struct dce_hwseq_registers hwseq_reg = {
-		HWSEQ_DCN21_REG_LIST()
-};
-
-static const struct dce_hwseq_shift hwseq_shift = {
-		HWSEQ_DCN21_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dce_hwseq_mask hwseq_mask = {
-		HWSEQ_DCN21_MASK_SH_LIST(_MASK)
-};
-
-static struct dce_hwseq *dcn21_hwseq_create(
-	struct dc_context *ctx)
-{
-	struct dce_hwseq *hws = kzalloc(sizeof(struct dce_hwseq), GFP_KERNEL);
-
-	if (hws) {
-		hws->ctx = ctx;
-		hws->regs = &hwseq_reg;
-		hws->shifts = &hwseq_shift;
-		hws->masks = &hwseq_mask;
-		hws->wa.DEGVIDCN21 = true;
-		hws->wa.disallow_self_refresh_during_multi_plane_transition = true;
-	}
-	return hws;
-}
-
-static const struct resource_create_funcs res_create_funcs = {
-	.read_dce_straps = read_dce_straps,
-	.create_audio = dcn21_create_audio,
-	.create_stream_encoder = dcn21_stream_encoder_create,
-	.create_hwseq = dcn21_hwseq_create,
-};
-
-static const struct resource_create_funcs res_create_maximus_funcs = {
-	.read_dce_straps = NULL,
-	.create_audio = NULL,
-	.create_stream_encoder = NULL,
-	.create_hwseq = dcn21_hwseq_create,
-};
-
-static const struct encoder_feature_support link_enc_feature = {
-		.max_hdmi_deep_color = COLOR_DEPTH_121212,
-		.max_hdmi_pixel_clock = 600000,
-		.hdmi_ycbcr420_supported = true,
-		.dp_ycbcr420_supported = true,
-		.fec_supported = true,
-		.flags.bits.IS_HBR2_CAPABLE = true,
-		.flags.bits.IS_HBR3_CAPABLE = true,
-		.flags.bits.IS_TPS3_CAPABLE = true,
-		.flags.bits.IS_TPS4_CAPABLE = true
-};
-
-
-#define link_regs(id, phyid)\
-[id] = {\
-	LE_DCN2_REG_LIST(id), \
-	UNIPHY_DCN2_REG_LIST(phyid), \
-	DPCS_DCN21_REG_LIST(id), \
-	SRI(DP_DPHY_INTERNAL_CTRL, DP, id) \
-}
-
-static const struct dcn10_link_enc_registers link_enc_regs[] = {
-	link_regs(0, A),
-	link_regs(1, B),
-	link_regs(2, C),
-	link_regs(3, D),
-	link_regs(4, E),
-};
-
-static const struct dce_panel_cntl_registers panel_cntl_regs[] = {
-	{ DCN_PANEL_CNTL_REG_LIST() }
-};
-
-static const struct dce_panel_cntl_shift panel_cntl_shift = {
-	DCE_PANEL_CNTL_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dce_panel_cntl_mask panel_cntl_mask = {
-	DCE_PANEL_CNTL_MASK_SH_LIST(_MASK)
-};
-
-#define aux_regs(id)\
-[id] = {\
-	DCN2_AUX_REG_LIST(id)\
-}
-
-static const struct dcn10_link_enc_aux_registers link_enc_aux_regs[] = {
-		aux_regs(0),
-		aux_regs(1),
-		aux_regs(2),
-		aux_regs(3),
-		aux_regs(4)
-};
-
-#define hpd_regs(id)\
-[id] = {\
-	HPD_REG_LIST(id)\
-}
-
-static const struct dcn10_link_enc_hpd_registers link_enc_hpd_regs[] = {
-		hpd_regs(0),
-		hpd_regs(1),
-		hpd_regs(2),
-		hpd_regs(3),
-		hpd_regs(4)
-};
-
-static const struct dcn10_link_enc_shift le_shift = {
-	LINK_ENCODER_MASK_SH_LIST_DCN20(__SHIFT),\
-	DPCS_DCN21_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dcn10_link_enc_mask le_mask = {
-	LINK_ENCODER_MASK_SH_LIST_DCN20(_MASK),\
-	DPCS_DCN21_MASK_SH_LIST(_MASK)
-};
-
-static int map_transmitter_id_to_phy_instance(
-	enum transmitter transmitter)
-{
-	switch (transmitter) {
-	case TRANSMITTER_UNIPHY_A:
-		return 0;
-	break;
-	case TRANSMITTER_UNIPHY_B:
-		return 1;
-	break;
-	case TRANSMITTER_UNIPHY_C:
-		return 2;
-	break;
-	case TRANSMITTER_UNIPHY_D:
-		return 3;
-	break;
-	case TRANSMITTER_UNIPHY_E:
-		return 4;
-	break;
-	default:
-		ASSERT(0);
-		return 0;
-	}
-}
-
-static struct link_encoder *dcn21_link_encoder_create(
-	const struct encoder_init_data *enc_init_data)
-{
-	struct dcn21_link_encoder *enc21 =
-		kzalloc(sizeof(struct dcn21_link_encoder), GFP_KERNEL);
-	int link_regs_id;
-
-	if (!enc21)
-		return NULL;
-
-	link_regs_id =
-		map_transmitter_id_to_phy_instance(enc_init_data->transmitter);
-
-	dcn21_link_encoder_construct(enc21,
-				      enc_init_data,
-				      &link_enc_feature,
-				      &link_enc_regs[link_regs_id],
-				      &link_enc_aux_regs[enc_init_data->channel - 1],
-				      &link_enc_hpd_regs[enc_init_data->hpd_source],
-				      &le_shift,
-				      &le_mask);
-
-	return &enc21->enc10.base;
-}
-
-static struct panel_cntl *dcn21_panel_cntl_create(const struct panel_cntl_init_data *init_data)
-{
-	struct dce_panel_cntl *panel_cntl =
-		kzalloc(sizeof(struct dce_panel_cntl), GFP_KERNEL);
-
-	if (!panel_cntl)
-		return NULL;
-
-	dce_panel_cntl_construct(panel_cntl,
-			init_data,
-			&panel_cntl_regs[init_data->inst],
-			&panel_cntl_shift,
-			&panel_cntl_mask);
-
-	return &panel_cntl->base;
-}
-
-#define CTX ctx
-
-#define REG(reg_name) \
-	(DCN_BASE.instance[0].segment[mm ## reg_name ## _BASE_IDX] + mm ## reg_name)
-
-static uint32_t read_pipe_fuses(struct dc_context *ctx)
-{
-	uint32_t value = REG_READ(CC_DC_PIPE_DIS);
-	/* RV1 support max 4 pipes */
-	value = value & 0xf;
-	return value;
-}
-
-static int dcn21_populate_dml_pipes_from_context(
-		struct dc *dc,
-		struct dc_state *context,
-		display_e2e_pipe_params_st *pipes,
-		bool fast_validate)
-{
-	uint32_t pipe_cnt = dcn20_populate_dml_pipes_from_context(dc, context, pipes, fast_validate);
-	int i;
-
-	for (i = 0; i < pipe_cnt; i++) {
-
-		pipes[i].pipe.src.hostvm = dc->res_pool->hubbub->riommu_active;
-		pipes[i].pipe.src.gpuvm = 1;
-	}
-
-	return pipe_cnt;
-}
-
-enum dc_status dcn21_patch_unknown_plane_state(struct dc_plane_state *plane_state)
-{
-	enum dc_status result = DC_OK;
-
-	if (plane_state->ctx->dc->debug.disable_dcc == DCC_ENABLE) {
-		plane_state->dcc.enable = 1;
-		/* align to our worst case block width */
-		plane_state->dcc.meta_pitch = ((plane_state->src_rect.width + 1023) / 1024) * 1024;
-	}
-	result = dcn20_patch_unknown_plane_state(plane_state);
-	return result;
-}
-
-static const struct resource_funcs dcn21_res_pool_funcs = {
-	.destroy = dcn21_destroy_resource_pool,
-	.link_enc_create = dcn21_link_encoder_create,
-	.panel_cntl_create = dcn21_panel_cntl_create,
-	.validate_bandwidth = dcn21_validate_bandwidth,
-	.populate_dml_pipes = dcn21_populate_dml_pipes_from_context,
-	.add_stream_to_ctx = dcn20_add_stream_to_ctx,
-	.add_dsc_to_stream_resource = dcn20_add_dsc_to_stream_resource,
-	.remove_stream_from_ctx = dcn20_remove_stream_from_ctx,
-	.acquire_idle_pipe_for_layer = dcn20_acquire_idle_pipe_for_layer,
-	.populate_dml_writeback_from_context = dcn20_populate_dml_writeback_from_context,
-	.patch_unknown_plane_state = dcn21_patch_unknown_plane_state,
-	.set_mcif_arb_params = dcn20_set_mcif_arb_params,
-	.find_first_free_match_stream_enc_for_link = dcn10_find_first_free_match_stream_enc_for_link,
-	.update_bw_bounding_box = update_bw_bounding_box
-};
-
-static bool dcn21_resource_construct(
-	uint8_t num_virtual_links,
-	struct dc *dc,
-	struct dcn21_resource_pool *pool)
-{
-	int i, j;
-	struct dc_context *ctx = dc->ctx;
-	struct irq_service_init_data init_data;
-	uint32_t pipe_fuses = read_pipe_fuses(ctx);
-	uint32_t num_pipes;
-
-	ctx->dc_bios->regs = &bios_regs;
-
-	pool->base.res_cap = &res_cap_rn;
-#ifdef DIAGS_BUILD
-	if (IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment))
-		//pool->base.res_cap = &res_cap_nv10_FPGA_2pipe_dsc;
-		pool->base.res_cap = &res_cap_rn_FPGA_4pipe;
-#endif
-
-	pool->base.funcs = &dcn21_res_pool_funcs;
-
-	/*************************************************
-	 *  Resource + asic cap harcoding                *
-	 *************************************************/
-	pool->base.underlay_pipe_index = NO_UNDERLAY_PIPE;
-
-	/* max pipe num for ASIC before check pipe fuses */
-	pool->base.pipe_count = pool->base.res_cap->num_timing_generator;
-
-	dc->caps.max_downscale_ratio = 200;
-	dc->caps.i2c_speed_in_khz = 100;
-	dc->caps.i2c_speed_in_khz_hdcp = 5; /*1.4 w/a applied by default*/
-	dc->caps.max_cursor_size = 256;
-	dc->caps.min_horizontal_blanking_period = 80;
-	dc->caps.dmdata_alloc_size = 2048;
-
-	dc->caps.max_slave_planes = 1;
-	dc->caps.max_slave_yuv_planes = 1;
-	dc->caps.max_slave_rgb_planes = 1;
-	dc->caps.post_blend_color_processing = true;
-	dc->caps.force_dp_tps4_for_cp2520 = true;
-	dc->caps.extended_aux_timeout_support = true;
-	dc->caps.dmcub_support = true;
-	dc->caps.is_apu = true;
-
-	/* Color pipeline capabilities */
-	dc->caps.color.dpp.dcn_arch = 1;
-	dc->caps.color.dpp.input_lut_shared = 0;
-	dc->caps.color.dpp.icsc = 1;
-	dc->caps.color.dpp.dgam_ram = 1;
-	dc->caps.color.dpp.dgam_rom_caps.srgb = 1;
-	dc->caps.color.dpp.dgam_rom_caps.bt2020 = 1;
-	dc->caps.color.dpp.dgam_rom_caps.gamma2_2 = 0;
-	dc->caps.color.dpp.dgam_rom_caps.pq = 0;
-	dc->caps.color.dpp.dgam_rom_caps.hlg = 0;
-	dc->caps.color.dpp.post_csc = 0;
-	dc->caps.color.dpp.gamma_corr = 0;
-	dc->caps.color.dpp.dgam_rom_for_yuv = 1;
-
-	dc->caps.color.dpp.hw_3d_lut = 1;
-	dc->caps.color.dpp.ogam_ram = 1;
-	// no OGAM ROM on DCN2
-	dc->caps.color.dpp.ogam_rom_caps.srgb = 0;
-	dc->caps.color.dpp.ogam_rom_caps.bt2020 = 0;
-	dc->caps.color.dpp.ogam_rom_caps.gamma2_2 = 0;
-	dc->caps.color.dpp.ogam_rom_caps.pq = 0;
-	dc->caps.color.dpp.ogam_rom_caps.hlg = 0;
-	dc->caps.color.dpp.ocsc = 0;
-
-	dc->caps.color.mpc.gamut_remap = 0;
-	dc->caps.color.mpc.num_3dluts = 0;
-	dc->caps.color.mpc.shared_3d_lut = 0;
-	dc->caps.color.mpc.ogam_ram = 1;
-	dc->caps.color.mpc.ogam_rom_caps.srgb = 0;
-	dc->caps.color.mpc.ogam_rom_caps.bt2020 = 0;
-	dc->caps.color.mpc.ogam_rom_caps.gamma2_2 = 0;
-	dc->caps.color.mpc.ogam_rom_caps.pq = 0;
-	dc->caps.color.mpc.ogam_rom_caps.hlg = 0;
-	dc->caps.color.mpc.ocsc = 1;
-
-	if (dc->ctx->dce_environment == DCE_ENV_PRODUCTION_DRV)
-		dc->debug = debug_defaults_drv;
-	else if (dc->ctx->dce_environment == DCE_ENV_FPGA_MAXIMUS) {
-		pool->base.pipe_count = 4;
-		dc->debug = debug_defaults_diags;
-	} else
-		dc->debug = debug_defaults_diags;
-
-	// Init the vm_helper
-	if (dc->vm_helper)
-		vm_helper_init(dc->vm_helper, 16);
-
-	/*************************************************
-	 *  Create resources                             *
-	 *************************************************/
-
-	pool->base.clock_sources[DCN20_CLK_SRC_PLL0] =
-			dcn21_clock_source_create(ctx, ctx->dc_bios,
-				CLOCK_SOURCE_COMBO_PHY_PLL0,
-				&clk_src_regs[0], false);
-	pool->base.clock_sources[DCN20_CLK_SRC_PLL1] =
-			dcn21_clock_source_create(ctx, ctx->dc_bios,
-				CLOCK_SOURCE_COMBO_PHY_PLL1,
-				&clk_src_regs[1], false);
-	pool->base.clock_sources[DCN20_CLK_SRC_PLL2] =
-			dcn21_clock_source_create(ctx, ctx->dc_bios,
-				CLOCK_SOURCE_COMBO_PHY_PLL2,
-				&clk_src_regs[2], false);
-	pool->base.clock_sources[DCN20_CLK_SRC_PLL3] =
-			dcn21_clock_source_create(ctx, ctx->dc_bios,
-				CLOCK_SOURCE_COMBO_PHY_PLL3,
-				&clk_src_regs[3], false);
-	pool->base.clock_sources[DCN20_CLK_SRC_PLL4] =
-			dcn21_clock_source_create(ctx, ctx->dc_bios,
-				CLOCK_SOURCE_COMBO_PHY_PLL4,
-				&clk_src_regs[4], false);
-
-	pool->base.clk_src_count = DCN20_CLK_SRC_TOTAL_DCN21;
-
-	/* todo: not reuse phy_pll registers */
-	pool->base.dp_clock_source =
-			dcn21_clock_source_create(ctx, ctx->dc_bios,
-				CLOCK_SOURCE_ID_DP_DTO,
-				&clk_src_regs[0], true);
-
-	for (i = 0; i < pool->base.clk_src_count; i++) {
-		if (pool->base.clock_sources[i] == NULL) {
-			dm_error("DC: failed to create clock sources!\n");
-			BREAK_TO_DEBUGGER();
-			goto create_fail;
-		}
-	}
-
-	pool->base.dccg = dccg21_create(ctx, &dccg_regs, &dccg_shift, &dccg_mask);
-	if (pool->base.dccg == NULL) {
-		dm_error("DC: failed to create dccg!\n");
-		BREAK_TO_DEBUGGER();
-		goto create_fail;
-	}
-
-	if (!dc->config.disable_dmcu) {
-		pool->base.dmcu = dcn21_dmcu_create(ctx,
-				&dmcu_regs,
-				&dmcu_shift,
-				&dmcu_mask);
-		if (pool->base.dmcu == NULL) {
-			dm_error("DC: failed to create dmcu!\n");
-			BREAK_TO_DEBUGGER();
-			goto create_fail;
-		}
-
-		dc->debug.dmub_command_table = false;
-	}
-
-	if (dc->config.disable_dmcu) {
-		pool->base.psr = dmub_psr_create(ctx);
-
-		if (pool->base.psr == NULL) {
-			dm_error("DC: failed to create psr obj!\n");
-			BREAK_TO_DEBUGGER();
-			goto create_fail;
-		}
-	}
-
-	if (dc->config.disable_dmcu)
-		pool->base.abm = dmub_abm_create(ctx,
-			&abm_regs,
-			&abm_shift,
-			&abm_mask);
-	else
-		pool->base.abm = dce_abm_create(ctx,
-			&abm_regs,
-			&abm_shift,
-			&abm_mask);
-
-	pool->base.pp_smu = dcn21_pp_smu_create(ctx);
-
-	num_pipes = dcn2_1_ip.max_num_dpp;
-
-	for (i = 0; i < dcn2_1_ip.max_num_dpp; i++)
-		if (pipe_fuses & 1 << i)
-			num_pipes--;
-	dcn2_1_ip.max_num_dpp = num_pipes;
-	dcn2_1_ip.max_num_otg = num_pipes;
-
-	dml_init_instance(&dc->dml, &dcn2_1_soc, &dcn2_1_ip, DML_PROJECT_DCN21);
-
-	init_data.ctx = dc->ctx;
-	pool->base.irqs = dal_irq_service_dcn21_create(&init_data);
-	if (!pool->base.irqs)
-		goto create_fail;
-
-	j = 0;
-	/* mem input -> ipp -> dpp -> opp -> TG */
-	for (i = 0; i < pool->base.pipe_count; i++) {
-		/* if pipe is disabled, skip instance of HW pipe,
-		 * i.e, skip ASIC register instance
-		 */
-		if ((pipe_fuses & (1 << i)) != 0)
-			continue;
-
-		pool->base.hubps[j] = dcn21_hubp_create(ctx, i);
-		if (pool->base.hubps[j] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error(
-				"DC: failed to create memory input!\n");
-			goto create_fail;
-		}
-
-		pool->base.ipps[j] = dcn21_ipp_create(ctx, i);
-		if (pool->base.ipps[j] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error(
-				"DC: failed to create input pixel processor!\n");
-			goto create_fail;
-		}
-
-		pool->base.dpps[j] = dcn21_dpp_create(ctx, i);
-		if (pool->base.dpps[j] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error(
-				"DC: failed to create dpps!\n");
-			goto create_fail;
-		}
-
-		pool->base.opps[j] = dcn21_opp_create(ctx, i);
-		if (pool->base.opps[j] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error(
-				"DC: failed to create output pixel processor!\n");
-			goto create_fail;
-		}
-
-		pool->base.timing_generators[j] = dcn21_timing_generator_create(
-				ctx, i);
-		if (pool->base.timing_generators[j] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error("DC: failed to create tg!\n");
-			goto create_fail;
-		}
-		j++;
-	}
-
-	for (i = 0; i < pool->base.res_cap->num_ddc; i++) {
-		pool->base.engines[i] = dcn21_aux_engine_create(ctx, i);
-		if (pool->base.engines[i] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error(
-				"DC:failed to create aux engine!!\n");
-			goto create_fail;
-		}
-		pool->base.hw_i2cs[i] = dcn21_i2c_hw_create(ctx, i);
-		if (pool->base.hw_i2cs[i] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error(
-				"DC:failed to create hw i2c!!\n");
-			goto create_fail;
-		}
-		pool->base.sw_i2cs[i] = NULL;
-	}
-
-	pool->base.timing_generator_count = j;
-	pool->base.pipe_count = j;
-	pool->base.mpcc_count = j;
-
-	pool->base.mpc = dcn21_mpc_create(ctx);
-	if (pool->base.mpc == NULL) {
-		BREAK_TO_DEBUGGER();
-		dm_error("DC: failed to create mpc!\n");
-		goto create_fail;
-	}
-
-	pool->base.hubbub = dcn21_hubbub_create(ctx);
-	if (pool->base.hubbub == NULL) {
-		BREAK_TO_DEBUGGER();
-		dm_error("DC: failed to create hubbub!\n");
-		goto create_fail;
-	}
-
-	for (i = 0; i < pool->base.res_cap->num_dsc; i++) {
-		pool->base.dscs[i] = dcn21_dsc_create(ctx, i);
-		if (pool->base.dscs[i] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error("DC: failed to create display stream compressor %d!\n", i);
-			goto create_fail;
-		}
-	}
-
-	if (!dcn20_dwbc_create(ctx, &pool->base)) {
-		BREAK_TO_DEBUGGER();
-		dm_error("DC: failed to create dwbc!\n");
-		goto create_fail;
-	}
-	if (!dcn20_mmhubbub_create(ctx, &pool->base)) {
-		BREAK_TO_DEBUGGER();
-		dm_error("DC: failed to create mcif_wb!\n");
-		goto create_fail;
-	}
-
-	if (!resource_construct(num_virtual_links, dc, &pool->base,
-			(!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment) ?
-			&res_create_funcs : &res_create_maximus_funcs)))
-			goto create_fail;
-
-	dcn21_hw_sequencer_construct(dc);
-
-	dc->caps.max_planes =  pool->base.pipe_count;
-
-	for (i = 0; i < dc->caps.max_planes; ++i)
-		dc->caps.planes[i] = plane_cap;
-
-	dc->cap_funcs = cap_funcs;
-
-	return true;
-
-create_fail:
-
-	dcn21_resource_destruct(pool);
-
-	return false;
-}
-
-struct resource_pool *dcn21_create_resource_pool(
-		const struct dc_init_data *init_data,
-		struct dc *dc)
-{
-	struct dcn21_resource_pool *pool =
-		kzalloc(sizeof(struct dcn21_resource_pool), GFP_KERNEL);
-
-	if (!pool)
-		return NULL;
-
-	if (dcn21_resource_construct(init_data->num_virtual_links, dc, pool))
-		return &pool->base;
-
-	BREAK_TO_DEBUGGER();
-	kfree(pool);
-	return NULL;
-}
diff --git a/drivers/gpu/drm/amd/display/dc/dcn30/Makefile b/drivers/gpu/drm/amd/display/dc/dcn30/Makefile
index dfd77b3cc84d..b17277de0340 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn30/Makefile
+++ b/drivers/gpu/drm/amd/display/dc/dcn30/Makefile
@@ -1,16 +1,16 @@
-# 
+#
 # Copyright 2020 Advanced Micro Devices, Inc.
-# 
+#
 # Permission is hereby granted, free of charge, to any person obtaining a
 # copy of this software and associated documentation files (the "Software"),
 # to deal in the Software without restriction, including without limitation
 # the rights to use, copy, modify, merge, publish, distribute, sublicense,
 # and/or sell copies of the Software, and to permit persons to whom the
 # Software is furnished to do so, subject to the following conditions:
-# 
+#
 # The above copyright notice and this permission notice shall be included in
 # all copies or substantial portions of the Software.
-# 
+#
 # THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
 # IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
 # FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
@@ -18,49 +18,15 @@
 # OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
 # ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
 # OTHER DEALINGS IN THE SOFTWARE.
-# 
+#
 # Authors: AMD
-# 
-# 
+#
+#
 
-
-DCN30 = dcn30_init.o dcn30_hubbub.o dcn30_hubp.o dcn30_dpp.o dcn30_optc.o \
-	dcn30_dccg.o dcn30_hwseq.o dcn30_mpc.o dcn30_vpg.o \
-	dcn30_afmt.o dcn30_dio_stream_encoder.o dcn30_dwb.o \
-	dcn30_dpp_cm.o dcn30_dwb_cm.o dcn30_cm_common.o dcn30_mmhubbub.o \
-	dcn30_dio_link_encoder.o dcn30_resource.o
-
-
-ifdef CONFIG_X86
-CFLAGS_$(AMDDALPATH)/dc/dcn30/dcn30_resource.o := -msse
-CFLAGS_$(AMDDALPATH)/dc/dcn30/dcn30_optc.o := -msse
-endif
-
-ifdef CONFIG_PPC64
-CFLAGS_$(AMDDALPATH)/dc/dcn30/dcn30_resource.o := -mhard-float -maltivec
-CFLAGS_$(AMDDALPATH)/dc/dcn30/dcn30_optc.o := -mhard-float -maltivec
-endif
-
-ifdef CONFIG_CC_IS_GCC
-ifeq ($(call cc-ifversion, -lt, 0701, y), y)
-IS_OLD_GCC = 1
-endif
-CFLAGS_$(AMDDALPATH)/dc/dcn30/dcn30_resource.o += -mhard-float
-CFLAGS_$(AMDDALPATH)/dc/dcn30/dcn30_optc.o += -mhard-float
-endif
-
-ifdef CONFIG_X86
-ifdef IS_OLD_GCC
-# Stack alignment mismatch, proceed with caution.
-# GCC < 7.1 cannot compile code using `double` and -mpreferred-stack-boundary=3
-# (8B stack alignment).
-CFLAGS_$(AMDDALPATH)/dc/dcn30/dcn30_resource.o += -mpreferred-stack-boundary=4
-CFLAGS_$(AMDDALPATH)/dc/dcn30/dcn30_optc.o += -mpreferred-stack-boundary=4
-else
-CFLAGS_$(AMDDALPATH)/dc/dcn30/dcn30_resource.o += -msse2
-CFLAGS_$(AMDDALPATH)/dc/dcn30/dcn30_optc.o += -msse2
-endif
-endif
+DCN30 := dcn30_vpg.o \
+	dcn30_afmt.o \
+	dcn30_cm_common.o \
+	dcn30_mmhubbub.o \
 
 AMD_DAL_DCN30 = $(addprefix $(AMDDALPATH)/dc/dcn30/,$(DCN30))
 
diff --git a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_afmt.c b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_afmt.c
index 95528e5ef89e..55e388c4c98b 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_afmt.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_afmt.c
@@ -123,7 +123,6 @@ void afmt3_se_audio_setup(
 {
 	struct dcn30_afmt *afmt3 = DCN30_AFMT_FROM_AFMT(afmt);
 
-	uint32_t speakers = 0;
 	uint32_t channels = 0;
 
 	ASSERT(audio_info);
@@ -131,7 +130,6 @@ void afmt3_se_audio_setup(
 	if (audio_info == NULL)
 		return;
 
-	speakers = audio_info->flags.info.ALLSPEAKERS;
 	channels = speakers_to_channels(audio_info->flags.speaker_flags).all;
 
 	/* setup the audio stream source select (audio -> dig mapping) */
diff --git a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_cm_common.c b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_cm_common.c
index e0df9b0065f9..0690c346f2c5 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_cm_common.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_cm_common.c
@@ -26,9 +26,9 @@
 #include "dm_services.h"
 #include "core_types.h"
 #include "reg_helper.h"
-#include "dcn30_dpp.h"
+#include "dcn30/dcn30_dpp.h"
 #include "basics/conversion.h"
-#include "dcn30_cm_common.h"
+#include "dcn30/dcn30_cm_common.h"
 #include "custom_float.h"
 
 #define REG(reg) reg
@@ -114,7 +114,6 @@ bool cm3_helper_translate_curve_to_hw_format(
 	struct pwl_result_data *rgb;
 	struct pwl_result_data *rgb_plus_1;
 	struct pwl_result_data *rgb_minus_1;
-	struct fixed31_32 end_value;
 
 	int32_t region_start, region_end;
 	int32_t i;
@@ -141,23 +140,18 @@ bool cm3_helper_translate_curve_to_hw_format(
 		region_start = -MAX_LOW_POINT;
 		region_end   = NUMBER_REGIONS - MAX_LOW_POINT;
 	} else {
-		/* 11 segments
-		 * segment is from 2^-10 to 2^0
+		/* 13 segments
+		 * segment is from 2^-12 to 2^0
 		 * There are less than 256 points, for optimization
 		 */
-		seg_distr[0] = 3;
-		seg_distr[1] = 4;
-		seg_distr[2] = 4;
-		seg_distr[3] = 4;
-		seg_distr[4] = 4;
-		seg_distr[5] = 4;
-		seg_distr[6] = 4;
-		seg_distr[7] = 4;
-		seg_distr[8] = 4;
-		seg_distr[9] = 4;
-		seg_distr[10] = 1;
-
-		region_start = -10;
+		const uint8_t SEG_COUNT = 12;
+
+		for (i = 0; i < SEG_COUNT; i++)
+			seg_distr[i] = 4;
+
+		seg_distr[SEG_COUNT] = 1;
+
+		region_start = -SEG_COUNT;
 		region_end = 1;
 	}
 
@@ -176,8 +170,10 @@ bool cm3_helper_translate_curve_to_hw_format(
 				NUMBER_SW_SEGMENTS;
 		for (i = start_index; i < start_index + NUMBER_SW_SEGMENTS;
 				i += increment) {
-			if (j == hw_points - 1)
+			if (j == hw_points)
 				break;
+			if (i >= TRANSFER_FUNC_POINTS)
+				return false;
 			rgb_resulted[j].red = output_tf->tf_pts.red[i];
 			rgb_resulted[j].green = output_tf->tf_pts.green[i];
 			rgb_resulted[j].blue = output_tf->tf_pts.blue[i];
@@ -187,13 +183,13 @@ bool cm3_helper_translate_curve_to_hw_format(
 
 	/* last point */
 	start_index = (region_end + MAX_LOW_POINT) * NUMBER_SW_SEGMENTS;
-	rgb_resulted[hw_points - 1].red = output_tf->tf_pts.red[start_index];
-	rgb_resulted[hw_points - 1].green = output_tf->tf_pts.green[start_index];
-	rgb_resulted[hw_points - 1].blue = output_tf->tf_pts.blue[start_index];
+	rgb_resulted[hw_points].red = output_tf->tf_pts.red[start_index];
+	rgb_resulted[hw_points].green = output_tf->tf_pts.green[start_index];
+	rgb_resulted[hw_points].blue = output_tf->tf_pts.blue[start_index];
 
-	rgb_resulted[hw_points].red = rgb_resulted[hw_points - 1].red;
-	rgb_resulted[hw_points].green = rgb_resulted[hw_points - 1].green;
-	rgb_resulted[hw_points].blue = rgb_resulted[hw_points - 1].blue;
+	rgb_resulted[hw_points+1].red = rgb_resulted[hw_points].red;
+	rgb_resulted[hw_points+1].green = rgb_resulted[hw_points].green;
+	rgb_resulted[hw_points+1].blue = rgb_resulted[hw_points].blue;
 
 	// All 3 color channels have same x
 	corner_points[0].red.x = dc_fixpt_pow(dc_fixpt_from_int(2),
@@ -220,34 +216,16 @@ bool cm3_helper_translate_curve_to_hw_format(
 	/* see comment above, m_arrPoints[1].y should be the Y value for the
 	 * region end (m_numOfHwPoints), not last HW point(m_numOfHwPoints - 1)
 	 */
-	corner_points[1].red.y = rgb_resulted[hw_points - 1].red;
-	corner_points[1].green.y = rgb_resulted[hw_points - 1].green;
-	corner_points[1].blue.y = rgb_resulted[hw_points - 1].blue;
+	corner_points[1].red.y = rgb_resulted[hw_points].red;
+	corner_points[1].green.y = rgb_resulted[hw_points].green;
+	corner_points[1].blue.y = rgb_resulted[hw_points].blue;
 	corner_points[1].red.slope = dc_fixpt_zero;
 	corner_points[1].green.slope = dc_fixpt_zero;
 	corner_points[1].blue.slope = dc_fixpt_zero;
 
-	if (output_tf->tf == TRANSFER_FUNCTION_PQ || output_tf->tf == TRANSFER_FUNCTION_HLG) {
-		/* for PQ/HLG, we want to have a straight line from last HW X point,
-		 * and the slope to be such that we hit 1.0 at 10000/1000 nits.
-		 */
-
-		if (output_tf->tf == TRANSFER_FUNCTION_PQ)
-			end_value = dc_fixpt_from_int(125);
-		else
-			end_value = dc_fixpt_from_fraction(125, 10);
-
-		corner_points[1].red.slope = dc_fixpt_div(
-			dc_fixpt_sub(dc_fixpt_one, corner_points[1].red.y),
-			dc_fixpt_sub(end_value, corner_points[1].red.x));
-		corner_points[1].green.slope = dc_fixpt_div(
-			dc_fixpt_sub(dc_fixpt_one, corner_points[1].green.y),
-			dc_fixpt_sub(end_value, corner_points[1].green.x));
-		corner_points[1].blue.slope = dc_fixpt_div(
-			dc_fixpt_sub(dc_fixpt_one, corner_points[1].blue.y),
-			dc_fixpt_sub(end_value, corner_points[1].blue.x));
-	}
-	lut_params->hw_points_num = hw_points;
+	// DCN3+ have 257 pts in lieu of no separate slope registers
+	// Prior HW had 256 base+slope pairs
+	lut_params->hw_points_num = hw_points + 1;
 
 	k = 0;
 	for (i = 1; i < MAX_REGIONS_NUMBER; i++) {
@@ -267,187 +245,37 @@ bool cm3_helper_translate_curve_to_hw_format(
 	rgb_plus_1 = rgb_resulted + 1;
 	rgb_minus_1 = rgb;
 
-	i = 1;
-	while (i != hw_points + 1) {
-		if (i >= hw_points - 1) {
-			if (dc_fixpt_lt(rgb_plus_1->red, rgb->red))
-				rgb_plus_1->red = dc_fixpt_add(rgb->red, rgb_minus_1->delta_red);
-			if (dc_fixpt_lt(rgb_plus_1->green, rgb->green))
-				rgb_plus_1->green = dc_fixpt_add(rgb->green, rgb_minus_1->delta_green);
-			if (dc_fixpt_lt(rgb_plus_1->blue, rgb->blue))
-				rgb_plus_1->blue = dc_fixpt_add(rgb->blue, rgb_minus_1->delta_blue);
-		}
-
-		rgb->delta_red   = dc_fixpt_sub(rgb_plus_1->red,   rgb->red);
-		rgb->delta_green = dc_fixpt_sub(rgb_plus_1->green, rgb->green);
-		rgb->delta_blue  = dc_fixpt_sub(rgb_plus_1->blue,  rgb->blue);
+	if (fixpoint == true) {
+		i = 1;
+		while (i != hw_points + 2) {
+			if (i >= hw_points) {
+				if (dc_fixpt_lt(rgb_plus_1->red, rgb->red))
+					rgb_plus_1->red = dc_fixpt_add(rgb->red,
+							rgb_minus_1->delta_red);
+				if (dc_fixpt_lt(rgb_plus_1->green, rgb->green))
+					rgb_plus_1->green = dc_fixpt_add(rgb->green,
+							rgb_minus_1->delta_green);
+				if (dc_fixpt_lt(rgb_plus_1->blue, rgb->blue))
+					rgb_plus_1->blue = dc_fixpt_add(rgb->blue,
+							rgb_minus_1->delta_blue);
+			}
 
-		if (fixpoint == true) {
 			rgb->delta_red_reg   = dc_fixpt_clamp_u0d10(rgb->delta_red);
 			rgb->delta_green_reg = dc_fixpt_clamp_u0d10(rgb->delta_green);
 			rgb->delta_blue_reg  = dc_fixpt_clamp_u0d10(rgb->delta_blue);
 			rgb->red_reg         = dc_fixpt_clamp_u0d14(rgb->red);
 			rgb->green_reg       = dc_fixpt_clamp_u0d14(rgb->green);
 			rgb->blue_reg        = dc_fixpt_clamp_u0d14(rgb->blue);
-		}
-
-		++rgb_plus_1;
-		rgb_minus_1 = rgb;
-		++rgb;
-		++i;
-	}
-	cm3_helper_convert_to_custom_float(rgb_resulted,
-						lut_params->corner_points,
-						hw_points, fixpoint);
-
-	return true;
-}
-
-#define NUM_DEGAMMA_REGIONS    12
-
-
-bool cm3_helper_translate_curve_to_degamma_hw_format(
-				const struct dc_transfer_func *output_tf,
-				struct pwl_params *lut_params)
-{
-	struct curve_points3 *corner_points;
-	struct pwl_result_data *rgb_resulted;
-	struct pwl_result_data *rgb;
-	struct pwl_result_data *rgb_plus_1;
-
-	int32_t region_start, region_end;
-	int32_t i;
-	uint32_t j, k, seg_distr[MAX_REGIONS_NUMBER], increment, start_index, hw_points;
 
-	if (output_tf == NULL || lut_params == NULL || output_tf->type == TF_TYPE_BYPASS)
-		return false;
-
-	corner_points = lut_params->corner_points;
-	rgb_resulted = lut_params->rgb_resulted;
-	hw_points = 0;
-
-	memset(lut_params, 0, sizeof(struct pwl_params));
-	memset(seg_distr, 0, sizeof(seg_distr));
-
-	region_start = -NUM_DEGAMMA_REGIONS;
-	region_end   = 0;
-
-
-	for (i = region_end - region_start; i < MAX_REGIONS_NUMBER ; i++)
-		seg_distr[i] = -1;
-	/* 12 segments
-	 * segments are from 2^-12 to 0
-	 */
-	for (i = 0; i < NUM_DEGAMMA_REGIONS ; i++)
-		seg_distr[i] = 4;
-
-	for (k = 0; k < MAX_REGIONS_NUMBER; k++) {
-		if (seg_distr[k] != -1)
-			hw_points += (1 << seg_distr[k]);
-	}
-
-	j = 0;
-	for (k = 0; k < (region_end - region_start); k++) {
-		increment = NUMBER_SW_SEGMENTS / (1 << seg_distr[k]);
-		start_index = (region_start + k + MAX_LOW_POINT) *
-				NUMBER_SW_SEGMENTS;
-		for (i = start_index; i < start_index + NUMBER_SW_SEGMENTS;
-				i += increment) {
-			if (j == hw_points - 1)
-				break;
-			rgb_resulted[j].red = output_tf->tf_pts.red[i];
-			rgb_resulted[j].green = output_tf->tf_pts.green[i];
-			rgb_resulted[j].blue = output_tf->tf_pts.blue[i];
-			j++;
+			++rgb_plus_1;
+			rgb_minus_1 = rgb;
+			++rgb;
+			++i;
 		}
 	}
-
-	/* last point */
-	start_index = (region_end + MAX_LOW_POINT) * NUMBER_SW_SEGMENTS;
-	rgb_resulted[hw_points - 1].red = output_tf->tf_pts.red[start_index];
-	rgb_resulted[hw_points - 1].green = output_tf->tf_pts.green[start_index];
-	rgb_resulted[hw_points - 1].blue = output_tf->tf_pts.blue[start_index];
-
-	corner_points[0].red.x = dc_fixpt_pow(dc_fixpt_from_int(2),
-					     dc_fixpt_from_int(region_start));
-	corner_points[0].green.x = corner_points[0].red.x;
-	corner_points[0].blue.x = corner_points[0].red.x;
-	corner_points[1].red.x = dc_fixpt_pow(dc_fixpt_from_int(2),
-					     dc_fixpt_from_int(region_end));
-	corner_points[1].green.x = corner_points[1].red.x;
-	corner_points[1].blue.x = corner_points[1].red.x;
-
-	corner_points[0].red.y = rgb_resulted[0].red;
-	corner_points[0].green.y = rgb_resulted[0].green;
-	corner_points[0].blue.y = rgb_resulted[0].blue;
-
-	/* see comment above, m_arrPoints[1].y should be the Y value for the
-	 * region end (m_numOfHwPoints), not last HW point(m_numOfHwPoints - 1)
-	 */
-	corner_points[1].red.y = rgb_resulted[hw_points - 1].red;
-	corner_points[1].green.y = rgb_resulted[hw_points - 1].green;
-	corner_points[1].blue.y = rgb_resulted[hw_points - 1].blue;
-	corner_points[1].red.slope = dc_fixpt_zero;
-	corner_points[1].green.slope = dc_fixpt_zero;
-	corner_points[1].blue.slope = dc_fixpt_zero;
-
-	if (output_tf->tf == TRANSFER_FUNCTION_PQ) {
-		/* for PQ, we want to have a straight line from last HW X point,
-		 * and the slope to be such that we hit 1.0 at 10000 nits.
-		 */
-		const struct fixed31_32 end_value =
-				dc_fixpt_from_int(125);
-
-		corner_points[1].red.slope = dc_fixpt_div(
-			dc_fixpt_sub(dc_fixpt_one, corner_points[1].red.y),
-			dc_fixpt_sub(end_value, corner_points[1].red.x));
-		corner_points[1].green.slope = dc_fixpt_div(
-			dc_fixpt_sub(dc_fixpt_one, corner_points[1].green.y),
-			dc_fixpt_sub(end_value, corner_points[1].green.x));
-		corner_points[1].blue.slope = dc_fixpt_div(
-			dc_fixpt_sub(dc_fixpt_one, corner_points[1].blue.y),
-			dc_fixpt_sub(end_value, corner_points[1].blue.x));
-	}
-
-	lut_params->hw_points_num = hw_points;
-
-	k = 0;
-	for (i = 1; i < MAX_REGIONS_NUMBER; i++) {
-		if (seg_distr[k] != -1) {
-			lut_params->arr_curve_points[k].segments_num =
-					seg_distr[k];
-			lut_params->arr_curve_points[i].offset =
-					lut_params->arr_curve_points[k].offset + (1 << seg_distr[k]);
-		}
-		k++;
-	}
-
-	if (seg_distr[k] != -1)
-		lut_params->arr_curve_points[k].segments_num = seg_distr[k];
-
-	rgb = rgb_resulted;
-	rgb_plus_1 = rgb_resulted + 1;
-
-	i = 1;
-	while (i != hw_points + 1) {
-		if (dc_fixpt_lt(rgb_plus_1->red, rgb->red))
-			rgb_plus_1->red = rgb->red;
-		if (dc_fixpt_lt(rgb_plus_1->green, rgb->green))
-			rgb_plus_1->green = rgb->green;
-		if (dc_fixpt_lt(rgb_plus_1->blue, rgb->blue))
-			rgb_plus_1->blue = rgb->blue;
-
-		rgb->delta_red   = dc_fixpt_sub(rgb_plus_1->red,   rgb->red);
-		rgb->delta_green = dc_fixpt_sub(rgb_plus_1->green, rgb->green);
-		rgb->delta_blue  = dc_fixpt_sub(rgb_plus_1->blue,  rgb->blue);
-
-		++rgb_plus_1;
-		++rgb;
-		++i;
-	}
 	cm3_helper_convert_to_custom_float(rgb_resulted,
 						lut_params->corner_points,
-						hw_points, false);
+						hw_points+1, fixpoint);
 
 	return true;
 }
@@ -603,24 +431,6 @@ bool cm3_helper_convert_to_custom_float(
 			return false;
 		}
 
-		if (!convert_to_custom_float_format(rgb->delta_red, &fmt,
-						    &rgb->delta_red_reg)) {
-			BREAK_TO_DEBUGGER();
-			return false;
-		}
-
-		if (!convert_to_custom_float_format(rgb->delta_green, &fmt,
-						    &rgb->delta_green_reg)) {
-			BREAK_TO_DEBUGGER();
-			return false;
-		}
-
-		if (!convert_to_custom_float_format(rgb->delta_blue, &fmt,
-						    &rgb->delta_blue_reg)) {
-			BREAK_TO_DEBUGGER();
-			return false;
-		}
-
 		++rgb;
 		++i;
 	}
diff --git a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_hubp.c b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_hubp.c
deleted file mode 100644
index eac08926b574..000000000000
--- a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_hubp.c
+++ /dev/null
@@ -1,529 +0,0 @@
-/*
- * Copyright 2020 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: AMD
- *
- */
-
-#include "dcn30_hubp.h"
-
-#include "dm_services.h"
-#include "dce_calcs.h"
-#include "reg_helper.h"
-#include "basics/conversion.h"
-#include "dcn20/dcn20_hubp.h"
-#include "dcn21/dcn21_hubp.h"
-
-#define REG(reg)\
-	hubp2->hubp_regs->reg
-
-#define CTX \
-	hubp2->base.ctx
-
-#undef FN
-#define FN(reg_name, field_name) \
-	hubp2->hubp_shift->field_name, hubp2->hubp_mask->field_name
-
-void hubp3_set_vm_system_aperture_settings(struct hubp *hubp,
-		struct vm_system_aperture_param *apt)
-{
-	struct dcn20_hubp *hubp2 = TO_DCN20_HUBP(hubp);
-
-	PHYSICAL_ADDRESS_LOC mc_vm_apt_default;
-	PHYSICAL_ADDRESS_LOC mc_vm_apt_low;
-	PHYSICAL_ADDRESS_LOC mc_vm_apt_high;
-
-	// The format of default addr is 48:12 of the 48 bit addr
-	mc_vm_apt_default.quad_part = apt->sys_default.quad_part >> 12;
-
-	// The format of high/low are 48:18 of the 48 bit addr
-	mc_vm_apt_low.quad_part = apt->sys_low.quad_part >> 18;
-	mc_vm_apt_high.quad_part = apt->sys_high.quad_part >> 18;
-
-	REG_SET(DCN_VM_SYSTEM_APERTURE_LOW_ADDR, 0,
-			MC_VM_SYSTEM_APERTURE_LOW_ADDR, mc_vm_apt_low.quad_part);
-
-	REG_SET(DCN_VM_SYSTEM_APERTURE_HIGH_ADDR, 0,
-			MC_VM_SYSTEM_APERTURE_HIGH_ADDR, mc_vm_apt_high.quad_part);
-
-	REG_SET_2(DCN_VM_MX_L1_TLB_CNTL, 0,
-			ENABLE_L1_TLB, 1,
-			SYSTEM_ACCESS_MODE, 0x3);
-}
-
-bool hubp3_program_surface_flip_and_addr(
-	struct hubp *hubp,
-	const struct dc_plane_address *address,
-	bool flip_immediate)
-{
-	struct dcn20_hubp *hubp2 = TO_DCN20_HUBP(hubp);
-
-	//program flip type
-	REG_UPDATE(DCSURF_FLIP_CONTROL,
-			SURFACE_FLIP_TYPE, flip_immediate);
-
-	// Program VMID reg
-	if (flip_immediate == 0)
-		REG_UPDATE(VMID_SETTINGS_0,
-			VMID, address->vmid);
-
-	if (address->type == PLN_ADDR_TYPE_GRPH_STEREO) {
-		REG_UPDATE(DCSURF_FLIP_CONTROL, SURFACE_FLIP_MODE_FOR_STEREOSYNC, 0x1);
-		REG_UPDATE(DCSURF_FLIP_CONTROL, SURFACE_FLIP_IN_STEREOSYNC, 0x1);
-
-	} else {
-		// turn off stereo if not in stereo
-		REG_UPDATE(DCSURF_FLIP_CONTROL, SURFACE_FLIP_MODE_FOR_STEREOSYNC, 0x0);
-		REG_UPDATE(DCSURF_FLIP_CONTROL, SURFACE_FLIP_IN_STEREOSYNC, 0x0);
-	}
-
-	/* HW automatically latch rest of address register on write to
-	 * DCSURF_PRIMARY_SURFACE_ADDRESS if SURFACE_UPDATE_LOCK is not used
-	 *
-	 * program high first and then the low addr, order matters!
-	 */
-	switch (address->type) {
-	case PLN_ADDR_TYPE_GRAPHICS:
-		/* DCN1.0 does not support const color
-		 * TODO: program DCHUBBUB_RET_PATH_DCC_CFGx_0/1
-		 * base on address->grph.dcc_const_color
-		 * x = 0, 2, 4, 6 for pipe 0, 1, 2, 3 for rgb and luma
-		 * x = 1, 3, 5, 7 for pipe 0, 1, 2, 3 for chroma
-		 */
-
-		if (address->grph.addr.quad_part == 0)
-			break;
-
-		REG_UPDATE_2(DCSURF_SURFACE_CONTROL,
-				PRIMARY_SURFACE_TMZ, address->tmz_surface,
-				PRIMARY_META_SURFACE_TMZ, address->tmz_surface);
-
-		if (address->grph.meta_addr.quad_part != 0) {
-			REG_SET(DCSURF_PRIMARY_META_SURFACE_ADDRESS_HIGH, 0,
-					PRIMARY_META_SURFACE_ADDRESS_HIGH,
-					address->grph.meta_addr.high_part);
-
-			REG_SET(DCSURF_PRIMARY_META_SURFACE_ADDRESS, 0,
-					PRIMARY_META_SURFACE_ADDRESS,
-					address->grph.meta_addr.low_part);
-		}
-
-		REG_SET(DCSURF_PRIMARY_SURFACE_ADDRESS_HIGH, 0,
-				PRIMARY_SURFACE_ADDRESS_HIGH,
-				address->grph.addr.high_part);
-
-		REG_SET(DCSURF_PRIMARY_SURFACE_ADDRESS, 0,
-				PRIMARY_SURFACE_ADDRESS,
-				address->grph.addr.low_part);
-		break;
-	case PLN_ADDR_TYPE_VIDEO_PROGRESSIVE:
-		if (address->video_progressive.luma_addr.quad_part == 0
-				|| address->video_progressive.chroma_addr.quad_part == 0)
-			break;
-
-		REG_UPDATE_4(DCSURF_SURFACE_CONTROL,
-				PRIMARY_SURFACE_TMZ, address->tmz_surface,
-				PRIMARY_SURFACE_TMZ_C, address->tmz_surface,
-				PRIMARY_META_SURFACE_TMZ, address->tmz_surface,
-				PRIMARY_META_SURFACE_TMZ_C, address->tmz_surface);
-
-		if (address->video_progressive.luma_meta_addr.quad_part != 0) {
-			REG_SET(DCSURF_PRIMARY_META_SURFACE_ADDRESS_HIGH_C, 0,
-					PRIMARY_META_SURFACE_ADDRESS_HIGH_C,
-					address->video_progressive.chroma_meta_addr.high_part);
-
-			REG_SET(DCSURF_PRIMARY_META_SURFACE_ADDRESS_C, 0,
-					PRIMARY_META_SURFACE_ADDRESS_C,
-					address->video_progressive.chroma_meta_addr.low_part);
-
-			REG_SET(DCSURF_PRIMARY_META_SURFACE_ADDRESS_HIGH, 0,
-					PRIMARY_META_SURFACE_ADDRESS_HIGH,
-					address->video_progressive.luma_meta_addr.high_part);
-
-			REG_SET(DCSURF_PRIMARY_META_SURFACE_ADDRESS, 0,
-					PRIMARY_META_SURFACE_ADDRESS,
-					address->video_progressive.luma_meta_addr.low_part);
-		}
-
-		REG_SET(DCSURF_PRIMARY_SURFACE_ADDRESS_HIGH_C, 0,
-				PRIMARY_SURFACE_ADDRESS_HIGH_C,
-				address->video_progressive.chroma_addr.high_part);
-
-		REG_SET(DCSURF_PRIMARY_SURFACE_ADDRESS_C, 0,
-				PRIMARY_SURFACE_ADDRESS_C,
-				address->video_progressive.chroma_addr.low_part);
-
-		REG_SET(DCSURF_PRIMARY_SURFACE_ADDRESS_HIGH, 0,
-				PRIMARY_SURFACE_ADDRESS_HIGH,
-				address->video_progressive.luma_addr.high_part);
-
-		REG_SET(DCSURF_PRIMARY_SURFACE_ADDRESS, 0,
-				PRIMARY_SURFACE_ADDRESS,
-				address->video_progressive.luma_addr.low_part);
-		break;
-	case PLN_ADDR_TYPE_GRPH_STEREO:
-		if (address->grph_stereo.left_addr.quad_part == 0)
-			break;
-		if (address->grph_stereo.right_addr.quad_part == 0)
-			break;
-
-		REG_UPDATE_8(DCSURF_SURFACE_CONTROL,
-				PRIMARY_SURFACE_TMZ, address->tmz_surface,
-				PRIMARY_SURFACE_TMZ_C, address->tmz_surface,
-				PRIMARY_META_SURFACE_TMZ, address->tmz_surface,
-				PRIMARY_META_SURFACE_TMZ_C, address->tmz_surface,
-				SECONDARY_SURFACE_TMZ, address->tmz_surface,
-				SECONDARY_SURFACE_TMZ_C, address->tmz_surface,
-				SECONDARY_META_SURFACE_TMZ, address->tmz_surface,
-				SECONDARY_META_SURFACE_TMZ_C, address->tmz_surface);
-
-		if (address->grph_stereo.right_meta_addr.quad_part != 0) {
-
-			REG_SET(DCSURF_SECONDARY_META_SURFACE_ADDRESS_HIGH_C, 0,
-				SECONDARY_META_SURFACE_ADDRESS_HIGH_C,
-				address->grph_stereo.right_alpha_meta_addr.high_part);
-
-			REG_SET(DCSURF_SECONDARY_META_SURFACE_ADDRESS_C, 0,
-				SECONDARY_META_SURFACE_ADDRESS_C,
-				address->grph_stereo.right_alpha_meta_addr.low_part);
-
-			REG_SET(DCSURF_SECONDARY_META_SURFACE_ADDRESS_HIGH, 0,
-					SECONDARY_META_SURFACE_ADDRESS_HIGH,
-					address->grph_stereo.right_meta_addr.high_part);
-
-			REG_SET(DCSURF_SECONDARY_META_SURFACE_ADDRESS, 0,
-					SECONDARY_META_SURFACE_ADDRESS,
-					address->grph_stereo.right_meta_addr.low_part);
-		}
-		if (address->grph_stereo.left_meta_addr.quad_part != 0) {
-
-			REG_SET(DCSURF_PRIMARY_META_SURFACE_ADDRESS_HIGH_C, 0,
-				PRIMARY_META_SURFACE_ADDRESS_HIGH_C,
-				address->grph_stereo.left_alpha_meta_addr.high_part);
-
-			REG_SET(DCSURF_PRIMARY_META_SURFACE_ADDRESS_C, 0,
-				PRIMARY_META_SURFACE_ADDRESS_C,
-				address->grph_stereo.left_alpha_meta_addr.low_part);
-
-			REG_SET(DCSURF_PRIMARY_META_SURFACE_ADDRESS_HIGH, 0,
-					PRIMARY_META_SURFACE_ADDRESS_HIGH,
-					address->grph_stereo.left_meta_addr.high_part);
-
-			REG_SET(DCSURF_PRIMARY_META_SURFACE_ADDRESS, 0,
-					PRIMARY_META_SURFACE_ADDRESS,
-					address->grph_stereo.left_meta_addr.low_part);
-		}
-
-		REG_SET(DCSURF_SECONDARY_SURFACE_ADDRESS_HIGH_C, 0,
-				SECONDARY_SURFACE_ADDRESS_HIGH_C,
-				address->grph_stereo.right_alpha_addr.high_part);
-
-		REG_SET(DCSURF_SECONDARY_SURFACE_ADDRESS_C, 0,
-				SECONDARY_SURFACE_ADDRESS_C,
-				address->grph_stereo.right_alpha_addr.low_part);
-
-		REG_SET(DCSURF_SECONDARY_SURFACE_ADDRESS_HIGH, 0,
-				SECONDARY_SURFACE_ADDRESS_HIGH,
-				address->grph_stereo.right_addr.high_part);
-
-		REG_SET(DCSURF_SECONDARY_SURFACE_ADDRESS, 0,
-				SECONDARY_SURFACE_ADDRESS,
-				address->grph_stereo.right_addr.low_part);
-
-		REG_SET(DCSURF_PRIMARY_SURFACE_ADDRESS_HIGH_C, 0,
-				PRIMARY_SURFACE_ADDRESS_HIGH_C,
-				address->grph_stereo.left_alpha_addr.high_part);
-
-		REG_SET(DCSURF_PRIMARY_SURFACE_ADDRESS_C, 0,
-				PRIMARY_SURFACE_ADDRESS_C,
-				address->grph_stereo.left_alpha_addr.low_part);
-
-		REG_SET(DCSURF_PRIMARY_SURFACE_ADDRESS_HIGH, 0,
-				PRIMARY_SURFACE_ADDRESS_HIGH,
-				address->grph_stereo.left_addr.high_part);
-
-		REG_SET(DCSURF_PRIMARY_SURFACE_ADDRESS, 0,
-				PRIMARY_SURFACE_ADDRESS,
-				address->grph_stereo.left_addr.low_part);
-		break;
-	case PLN_ADDR_TYPE_RGBEA:
-		if (address->rgbea.addr.quad_part == 0
-				|| address->rgbea.alpha_addr.quad_part == 0)
-			break;
-
-		REG_UPDATE_4(DCSURF_SURFACE_CONTROL,
-				PRIMARY_SURFACE_TMZ, address->tmz_surface,
-				PRIMARY_SURFACE_TMZ_C, address->tmz_surface,
-				PRIMARY_META_SURFACE_TMZ, address->tmz_surface,
-				PRIMARY_META_SURFACE_TMZ_C, address->tmz_surface);
-
-		if (address->rgbea.meta_addr.quad_part != 0) {
-
-			REG_SET(DCSURF_PRIMARY_META_SURFACE_ADDRESS_HIGH_C, 0,
-					PRIMARY_META_SURFACE_ADDRESS_HIGH_C,
-					address->rgbea.alpha_meta_addr.high_part);
-
-			REG_SET(DCSURF_PRIMARY_META_SURFACE_ADDRESS_C, 0,
-					PRIMARY_META_SURFACE_ADDRESS_C,
-					address->rgbea.alpha_meta_addr.low_part);
-
-			REG_SET(DCSURF_PRIMARY_META_SURFACE_ADDRESS_HIGH, 0,
-					PRIMARY_META_SURFACE_ADDRESS_HIGH,
-					address->rgbea.meta_addr.high_part);
-
-			REG_SET(DCSURF_PRIMARY_META_SURFACE_ADDRESS, 0,
-					PRIMARY_META_SURFACE_ADDRESS,
-					address->rgbea.meta_addr.low_part);
-		}
-
-		REG_SET(DCSURF_PRIMARY_SURFACE_ADDRESS_HIGH_C, 0,
-				PRIMARY_SURFACE_ADDRESS_HIGH_C,
-				address->rgbea.alpha_addr.high_part);
-
-		REG_SET(DCSURF_PRIMARY_SURFACE_ADDRESS_C, 0,
-				PRIMARY_SURFACE_ADDRESS_C,
-				address->rgbea.alpha_addr.low_part);
-
-		REG_SET(DCSURF_PRIMARY_SURFACE_ADDRESS_HIGH, 0,
-				PRIMARY_SURFACE_ADDRESS_HIGH,
-				address->rgbea.addr.high_part);
-
-		REG_SET(DCSURF_PRIMARY_SURFACE_ADDRESS, 0,
-				PRIMARY_SURFACE_ADDRESS,
-				address->rgbea.addr.low_part);
-		break;
-	default:
-		BREAK_TO_DEBUGGER();
-		break;
-	}
-
-	hubp->request_address = *address;
-
-	return true;
-}
-
-static void hubp3_program_tiling(
-	struct dcn20_hubp *hubp2,
-	const union dc_tiling_info *info,
-	const enum surface_pixel_format pixel_format)
-{
-	REG_UPDATE_4(DCSURF_ADDR_CONFIG,
-		NUM_PIPES, log_2(info->gfx9.num_pipes),
-		PIPE_INTERLEAVE, info->gfx9.pipe_interleave,
-		MAX_COMPRESSED_FRAGS, log_2(info->gfx9.max_compressed_frags),
-		NUM_PKRS, log_2(info->gfx9.num_pkrs));
-
-	REG_UPDATE_3(DCSURF_TILING_CONFIG,
-			SW_MODE, info->gfx9.swizzle,
-			META_LINEAR, info->gfx9.meta_linear,
-			PIPE_ALIGNED, info->gfx9.pipe_aligned);
-
-}
-
-void hubp3_dcc_control(struct hubp *hubp, bool enable,
-		enum hubp_ind_block_size blk_size)
-{
-	uint32_t dcc_en = enable ? 1 : 0;
-	struct dcn20_hubp *hubp2 = TO_DCN20_HUBP(hubp);
-
-	REG_UPDATE_4(DCSURF_SURFACE_CONTROL,
-			PRIMARY_SURFACE_DCC_EN, dcc_en,
-			PRIMARY_SURFACE_DCC_IND_BLK, blk_size,
-			SECONDARY_SURFACE_DCC_EN, dcc_en,
-			SECONDARY_SURFACE_DCC_IND_BLK, blk_size);
-}
-
-void hubp3_dcc_control_sienna_cichlid(struct hubp *hubp,
-		struct dc_plane_dcc_param *dcc)
-{
-	struct dcn20_hubp *hubp2 = TO_DCN20_HUBP(hubp);
-
-	REG_UPDATE_6(DCSURF_SURFACE_CONTROL,
-		PRIMARY_SURFACE_DCC_EN, dcc->enable,
-		PRIMARY_SURFACE_DCC_IND_BLK, dcc->dcc_ind_blk,
-		PRIMARY_SURFACE_DCC_IND_BLK_C, dcc->dcc_ind_blk_c,
-		SECONDARY_SURFACE_DCC_EN, dcc->enable,
-		SECONDARY_SURFACE_DCC_IND_BLK, dcc->dcc_ind_blk,
-		SECONDARY_SURFACE_DCC_IND_BLK_C, dcc->dcc_ind_blk_c);
-}
-
-void hubp3_dmdata_set_attributes(
-		struct hubp *hubp,
-		const struct dc_dmdata_attributes *attr)
-{
-	struct dcn20_hubp *hubp2 = TO_DCN20_HUBP(hubp);
-
-	/*always HW mode */
-	REG_UPDATE(DMDATA_CNTL,
-			DMDATA_MODE, 1);
-
-	/* for DMDATA flip, need to use SURFACE_UPDATE_LOCK */
-	REG_UPDATE(DCSURF_FLIP_CONTROL, SURFACE_UPDATE_LOCK, 1);
-
-	/* toggle DMDATA_UPDATED and set repeat and size */
-	REG_UPDATE(DMDATA_CNTL,
-			DMDATA_UPDATED, 0);
-	REG_UPDATE_3(DMDATA_CNTL,
-			DMDATA_UPDATED, 1,
-			DMDATA_REPEAT, attr->dmdata_repeat,
-			DMDATA_SIZE, attr->dmdata_size);
-
-	/* set DMDATA address */
-	REG_WRITE(DMDATA_ADDRESS_LOW, attr->address.low_part);
-	REG_UPDATE(DMDATA_ADDRESS_HIGH,
-			DMDATA_ADDRESS_HIGH, attr->address.high_part);
-
-	REG_UPDATE(DCSURF_FLIP_CONTROL, SURFACE_UPDATE_LOCK, 0);
-
-}
-
-
-void hubp3_program_surface_config(
-	struct hubp *hubp,
-	enum surface_pixel_format format,
-	union dc_tiling_info *tiling_info,
-	struct plane_size *plane_size,
-	enum dc_rotation_angle rotation,
-	struct dc_plane_dcc_param *dcc,
-	bool horizontal_mirror,
-	unsigned int compat_level)
-{
-	struct dcn20_hubp *hubp2 = TO_DCN20_HUBP(hubp);
-
-	hubp3_dcc_control_sienna_cichlid(hubp, dcc);
-	hubp3_program_tiling(hubp2, tiling_info, format);
-	hubp2_program_size(hubp, format, plane_size, dcc);
-	hubp2_program_rotation(hubp, rotation, horizontal_mirror);
-	hubp2_program_pixel_format(hubp, format);
-}
-
-static void hubp3_program_deadline(
-		struct hubp *hubp,
-		struct _vcs_dpi_display_dlg_regs_st *dlg_attr,
-		struct _vcs_dpi_display_ttu_regs_st *ttu_attr)
-{
-	struct dcn20_hubp *hubp2 = TO_DCN20_HUBP(hubp);
-
-	hubp2_program_deadline(hubp, dlg_attr, ttu_attr);
-	REG_UPDATE(DCN_DMDATA_VM_CNTL,
-			REFCYC_PER_VM_DMDATA, dlg_attr->refcyc_per_vm_dmdata);
-}
-
-void hubp3_read_state(struct hubp *hubp)
-{
-	struct dcn20_hubp *hubp2 = TO_DCN20_HUBP(hubp);
-	struct dcn_hubp_state *s = &hubp2->state;
-	struct _vcs_dpi_display_rq_regs_st *rq_regs = &s->rq_regs;
-
-	hubp2_read_state_common(hubp);
-
-	REG_GET_7(DCHUBP_REQ_SIZE_CONFIG,
-		CHUNK_SIZE, &rq_regs->rq_regs_l.chunk_size,
-		MIN_CHUNK_SIZE, &rq_regs->rq_regs_l.min_chunk_size,
-		META_CHUNK_SIZE, &rq_regs->rq_regs_l.meta_chunk_size,
-		MIN_META_CHUNK_SIZE, &rq_regs->rq_regs_l.min_meta_chunk_size,
-		DPTE_GROUP_SIZE, &rq_regs->rq_regs_l.dpte_group_size,
-		SWATH_HEIGHT, &rq_regs->rq_regs_l.swath_height,
-		PTE_ROW_HEIGHT_LINEAR, &rq_regs->rq_regs_l.pte_row_height_linear);
-
-	REG_GET_7(DCHUBP_REQ_SIZE_CONFIG_C,
-		CHUNK_SIZE_C, &rq_regs->rq_regs_c.chunk_size,
-		MIN_CHUNK_SIZE_C, &rq_regs->rq_regs_c.min_chunk_size,
-		META_CHUNK_SIZE_C, &rq_regs->rq_regs_c.meta_chunk_size,
-		MIN_META_CHUNK_SIZE_C, &rq_regs->rq_regs_c.min_meta_chunk_size,
-		DPTE_GROUP_SIZE_C, &rq_regs->rq_regs_c.dpte_group_size,
-		SWATH_HEIGHT_C, &rq_regs->rq_regs_c.swath_height,
-		PTE_ROW_HEIGHT_LINEAR_C, &rq_regs->rq_regs_c.pte_row_height_linear);
-
-}
-
-void hubp3_setup(
-		struct hubp *hubp,
-		struct _vcs_dpi_display_dlg_regs_st *dlg_attr,
-		struct _vcs_dpi_display_ttu_regs_st *ttu_attr,
-		struct _vcs_dpi_display_rq_regs_st *rq_regs,
-		struct _vcs_dpi_display_pipe_dest_params_st *pipe_dest)
-{
-	/* otg is locked when this func is called. Register are double buffered.
-	 * disable the requestors is not needed
-	 */
-	hubp2_vready_at_or_After_vsync(hubp, pipe_dest);
-	hubp21_program_requestor(hubp, rq_regs);
-	hubp3_program_deadline(hubp, dlg_attr, ttu_attr);
-}
-
-void hubp3_init(struct hubp *hubp)
-{
-	// DEDCN21-133: Inconsistent row starting line for flip between DPTE and Meta
-	// This is a chicken bit to enable the ECO fix.
-
-	struct dcn20_hubp *hubp2 = TO_DCN20_HUBP(hubp);
-	//hubp[i].HUBPREQ_DEBUG.HUBPREQ_DEBUG[26] = 1;
-	REG_WRITE(HUBPREQ_DEBUG, 1 << 26);
-}
-
-static struct hubp_funcs dcn30_hubp_funcs = {
-	.hubp_enable_tripleBuffer = hubp2_enable_triplebuffer,
-	.hubp_is_triplebuffer_enabled = hubp2_is_triplebuffer_enabled,
-	.hubp_program_surface_flip_and_addr = hubp3_program_surface_flip_and_addr,
-	.hubp_program_surface_config = hubp3_program_surface_config,
-	.hubp_is_flip_pending = hubp2_is_flip_pending,
-	.hubp_setup = hubp3_setup,
-	.hubp_setup_interdependent = hubp2_setup_interdependent,
-	.hubp_set_vm_system_aperture_settings = hubp3_set_vm_system_aperture_settings,
-	.set_blank = hubp2_set_blank,
-	.dcc_control = hubp3_dcc_control,
-	.mem_program_viewport = min_set_viewport,
-	.set_cursor_attributes	= hubp2_cursor_set_attributes,
-	.set_cursor_position	= hubp2_cursor_set_position,
-	.hubp_clk_cntl = hubp2_clk_cntl,
-	.hubp_vtg_sel = hubp2_vtg_sel,
-	.dmdata_set_attributes = hubp3_dmdata_set_attributes,
-	.dmdata_load = hubp2_dmdata_load,
-	.dmdata_status_done = hubp2_dmdata_status_done,
-	.hubp_read_state = hubp3_read_state,
-	.hubp_clear_underflow = hubp2_clear_underflow,
-	.hubp_set_flip_control_surface_gsl = hubp2_set_flip_control_surface_gsl,
-	.hubp_init = hubp3_init,
-	.hubp_in_blank = hubp1_in_blank,
-	.hubp_soft_reset = hubp1_soft_reset,
-	.hubp_set_flip_int = hubp1_set_flip_int,
-};
-
-bool hubp3_construct(
-	struct dcn20_hubp *hubp2,
-	struct dc_context *ctx,
-	uint32_t inst,
-	const struct dcn_hubp2_registers *hubp_regs,
-	const struct dcn_hubp2_shift *hubp_shift,
-	const struct dcn_hubp2_mask *hubp_mask)
-{
-	hubp2->base.funcs = &dcn30_hubp_funcs;
-	hubp2->base.ctx = ctx;
-	hubp2->hubp_regs = hubp_regs;
-	hubp2->hubp_shift = hubp_shift;
-	hubp2->hubp_mask = hubp_mask;
-	hubp2->base.inst = inst;
-	hubp2->base.opp_id = OPP_ID_INVALID;
-	hubp2->base.mpcc_id = 0xf;
-
-	return true;
-}
diff --git a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_hwseq.c b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_hwseq.c
deleted file mode 100644
index df2717116604..000000000000
--- a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_hwseq.c
+++ /dev/null
@@ -1,1011 +0,0 @@
-/*
- * Copyright 2020 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: AMD
- *
- */
-
-
-#include "dm_services.h"
-#include "dm_helpers.h"
-#include "core_types.h"
-#include "resource.h"
-#include "dcn30_hwseq.h"
-#include "dccg.h"
-#include "dce/dce_hwseq.h"
-#include "dcn30_mpc.h"
-#include "dcn30_dpp.h"
-#include "dcn10/dcn10_cm_common.h"
-#include "dcn30_cm_common.h"
-#include "reg_helper.h"
-#include "abm.h"
-#include "clk_mgr.h"
-#include "hubp.h"
-#include "dchubbub.h"
-#include "timing_generator.h"
-#include "opp.h"
-#include "ipp.h"
-#include "mpc.h"
-#include "mcif_wb.h"
-#include "dc_dmub_srv.h"
-#include "link_hwss.h"
-#include "dpcd_defs.h"
-#include "inc/dc_link_dp.h"
-#include "inc/link_dpcd.h"
-
-
-
-
-#define DC_LOGGER_INIT(logger)
-
-#define CTX \
-	hws->ctx
-#define REG(reg)\
-	hws->regs->reg
-#define DC_LOGGER \
-		dc->ctx->logger
-
-
-#undef FN
-#define FN(reg_name, field_name) \
-	hws->shifts->field_name, hws->masks->field_name
-
-bool dcn30_set_blend_lut(
-	struct pipe_ctx *pipe_ctx, const struct dc_plane_state *plane_state)
-{
-	struct dpp *dpp_base = pipe_ctx->plane_res.dpp;
-	bool result = true;
-	struct pwl_params *blend_lut = NULL;
-
-	if (plane_state->blend_tf) {
-		if (plane_state->blend_tf->type == TF_TYPE_HWPWL)
-			blend_lut = &plane_state->blend_tf->pwl;
-		else if (plane_state->blend_tf->type == TF_TYPE_DISTRIBUTED_POINTS) {
-			cm3_helper_translate_curve_to_hw_format(
-					plane_state->blend_tf, &dpp_base->regamma_params, false);
-			blend_lut = &dpp_base->regamma_params;
-		}
-	}
-	result = dpp_base->funcs->dpp_program_blnd_lut(dpp_base, blend_lut);
-
-	return result;
-}
-
-static bool dcn30_set_mpc_shaper_3dlut(
-	struct pipe_ctx *pipe_ctx, const struct dc_stream_state *stream)
-{
-	struct dpp *dpp_base = pipe_ctx->plane_res.dpp;
-	int mpcc_id = pipe_ctx->plane_res.hubp->inst;
-	struct mpc *mpc = pipe_ctx->stream_res.opp->ctx->dc->res_pool->mpc;
-	bool result = false;
-	int acquired_rmu = 0;
-	int mpcc_id_projected = 0;
-
-	const struct pwl_params *shaper_lut = NULL;
-	//get the shaper lut params
-	if (stream->func_shaper) {
-		if (stream->func_shaper->type == TF_TYPE_HWPWL)
-			shaper_lut = &stream->func_shaper->pwl;
-		else if (stream->func_shaper->type == TF_TYPE_DISTRIBUTED_POINTS) {
-			cm_helper_translate_curve_to_hw_format(
-					stream->func_shaper,
-					&dpp_base->shaper_params, true);
-			shaper_lut = &dpp_base->shaper_params;
-		}
-	}
-
-	if (stream->lut3d_func &&
-		stream->lut3d_func->state.bits.initialized == 1 &&
-		stream->lut3d_func->state.bits.rmu_idx_valid == 1) {
-		if (stream->lut3d_func->state.bits.rmu_mux_num == 0)
-			mpcc_id_projected = stream->lut3d_func->state.bits.mpc_rmu0_mux;
-		else if (stream->lut3d_func->state.bits.rmu_mux_num == 1)
-			mpcc_id_projected = stream->lut3d_func->state.bits.mpc_rmu1_mux;
-		else if (stream->lut3d_func->state.bits.rmu_mux_num == 2)
-			mpcc_id_projected = stream->lut3d_func->state.bits.mpc_rmu2_mux;
-		if (mpcc_id_projected != mpcc_id)
-			BREAK_TO_DEBUGGER();
-		/*find the reason why logical layer assigned a differant mpcc_id into acquire_post_bldn_3dlut*/
-		acquired_rmu = mpc->funcs->acquire_rmu(mpc, mpcc_id,
-				stream->lut3d_func->state.bits.rmu_mux_num);
-		if (acquired_rmu != stream->lut3d_func->state.bits.rmu_mux_num)
-			BREAK_TO_DEBUGGER();
-		result = mpc->funcs->program_3dlut(mpc,
-								&stream->lut3d_func->lut_3d,
-								stream->lut3d_func->state.bits.rmu_mux_num);
-		result = mpc->funcs->program_shaper(mpc, shaper_lut,
-				stream->lut3d_func->state.bits.rmu_mux_num);
-	} else
-		/*loop through the available mux and release the requested mpcc_id*/
-		mpc->funcs->release_rmu(mpc, mpcc_id);
-
-
-	return result;
-}
-
-bool dcn30_set_input_transfer_func(struct dc *dc,
-				struct pipe_ctx *pipe_ctx,
-				const struct dc_plane_state *plane_state)
-{
-	struct dce_hwseq *hws = dc->hwseq;
-	struct dpp *dpp_base = pipe_ctx->plane_res.dpp;
-	enum dc_transfer_func_predefined tf;
-	bool result = true;
-	struct pwl_params *params = NULL;
-
-	if (dpp_base == NULL || plane_state == NULL)
-		return false;
-
-	tf = TRANSFER_FUNCTION_UNITY;
-
-	if (plane_state->in_transfer_func &&
-		plane_state->in_transfer_func->type == TF_TYPE_PREDEFINED)
-		tf = plane_state->in_transfer_func->tf;
-
-	dpp_base->funcs->dpp_set_pre_degam(dpp_base, tf);
-
-	if (plane_state->in_transfer_func) {
-		if (plane_state->in_transfer_func->type == TF_TYPE_HWPWL)
-			params = &plane_state->in_transfer_func->pwl;
-		else if (plane_state->in_transfer_func->type == TF_TYPE_DISTRIBUTED_POINTS &&
-			cm3_helper_translate_curve_to_hw_format(plane_state->in_transfer_func,
-					&dpp_base->degamma_params, false))
-			params = &dpp_base->degamma_params;
-	}
-
-	result = dpp_base->funcs->dpp_program_gamcor_lut(dpp_base, params);
-
-	if (pipe_ctx->stream_res.opp && pipe_ctx->stream_res.opp->ctx) {
-		if (dpp_base->funcs->dpp_program_blnd_lut)
-			hws->funcs.set_blend_lut(pipe_ctx, plane_state);
-		if (dpp_base->funcs->dpp_program_shaper_lut &&
-				dpp_base->funcs->dpp_program_3dlut)
-			hws->funcs.set_shaper_3dlut(pipe_ctx, plane_state);
-	}
-
-	return result;
-}
-
-bool dcn30_set_output_transfer_func(struct dc *dc,
-				struct pipe_ctx *pipe_ctx,
-				const struct dc_stream_state *stream)
-{
-	int mpcc_id = pipe_ctx->plane_res.hubp->inst;
-	struct mpc *mpc = pipe_ctx->stream_res.opp->ctx->dc->res_pool->mpc;
-	struct pwl_params *params = NULL;
-	bool ret = false;
-
-	/* program OGAM or 3DLUT only for the top pipe*/
-	if (pipe_ctx->top_pipe == NULL) {
-		/*program rmu shaper and 3dlut in MPC*/
-		ret = dcn30_set_mpc_shaper_3dlut(pipe_ctx, stream);
-		if (ret == false && mpc->funcs->set_output_gamma && stream->out_transfer_func) {
-			if (stream->out_transfer_func->type == TF_TYPE_HWPWL)
-				params = &stream->out_transfer_func->pwl;
-			else if (pipe_ctx->stream->out_transfer_func->type ==
-					TF_TYPE_DISTRIBUTED_POINTS &&
-					cm3_helper_translate_curve_to_hw_format(
-					stream->out_transfer_func,
-					&mpc->blender_params, false))
-				params = &mpc->blender_params;
-			 /* there are no ROM LUTs in OUTGAM */
-			if (stream->out_transfer_func->type == TF_TYPE_PREDEFINED)
-				BREAK_TO_DEBUGGER();
-		}
-	}
-
-	mpc->funcs->set_output_gamma(mpc, mpcc_id, params);
-	return ret;
-}
-
-static void dcn30_set_writeback(
-		struct dc *dc,
-		struct dc_writeback_info *wb_info,
-		struct dc_state *context)
-{
-	struct mcif_wb *mcif_wb;
-	struct mcif_buf_params *mcif_buf_params;
-
-	ASSERT(wb_info->dwb_pipe_inst < MAX_DWB_PIPES);
-	ASSERT(wb_info->wb_enabled);
-	ASSERT(wb_info->mpcc_inst >= 0);
-	ASSERT(wb_info->mpcc_inst < dc->res_pool->mpcc_count);
-	mcif_wb = dc->res_pool->mcif_wb[wb_info->dwb_pipe_inst];
-	mcif_buf_params = &wb_info->mcif_buf_params;
-
-	/* set DWB MPC mux */
-	dc->res_pool->mpc->funcs->set_dwb_mux(dc->res_pool->mpc,
-			wb_info->dwb_pipe_inst, wb_info->mpcc_inst);
-	/* set MCIF_WB buffer and arbitration configuration */
-	mcif_wb->funcs->config_mcif_buf(mcif_wb, mcif_buf_params, wb_info->dwb_params.dest_height);
-	mcif_wb->funcs->config_mcif_arb(mcif_wb, &context->bw_ctx.bw.dcn.bw_writeback.mcif_wb_arb[wb_info->dwb_pipe_inst]);
-}
-
-void dcn30_update_writeback(
-		struct dc *dc,
-		struct dc_writeback_info *wb_info,
-		struct dc_state *context)
-{
-	struct dwbc *dwb;
-	dwb = dc->res_pool->dwbc[wb_info->dwb_pipe_inst];
-	DC_LOG_DWB("%s dwb_pipe_inst = %d, mpcc_inst = %d",\
-		__func__, wb_info->dwb_pipe_inst,\
-		wb_info->mpcc_inst);
-
-	dcn30_set_writeback(dc, wb_info, context);
-
-	/* update DWB */
-	dwb->funcs->update(dwb, &wb_info->dwb_params);
-}
-
-bool dcn30_mmhubbub_warmup(
-	struct dc *dc,
-	unsigned int num_dwb,
-	struct dc_writeback_info *wb_info)
-{
-	struct dwbc *dwb;
-	struct mcif_wb *mcif_wb;
-	struct mcif_warmup_params warmup_params = {0};
-	unsigned int  i, i_buf;
-	/*make sure there is no active DWB eanbled */
-	for (i = 0; i < num_dwb; i++) {
-		dwb = dc->res_pool->dwbc[wb_info[i].dwb_pipe_inst];
-		if (dwb->dwb_is_efc_transition || dwb->dwb_is_drc) {
-			/*can not do warmup while any dwb enabled*/
-			return false;
-		}
-	}
-
-	if (wb_info->mcif_warmup_params.p_vmid == 0)
-		return false;
-
-	/*check whether this is new interface: warmup big buffer once*/
-	if (wb_info->mcif_warmup_params.start_address.quad_part != 0 &&
-		wb_info->mcif_warmup_params.region_size != 0) {
-		/*mmhubbub is shared, so it does not matter which MCIF*/
-		mcif_wb = dc->res_pool->mcif_wb[0];
-		/*warmup a big chunk of VM buffer at once*/
-		warmup_params.start_address.quad_part = wb_info->mcif_warmup_params.start_address.quad_part;
-		warmup_params.address_increment =  wb_info->mcif_warmup_params.region_size;
-		warmup_params.region_size = wb_info->mcif_warmup_params.region_size;
-		warmup_params.p_vmid = wb_info->mcif_warmup_params.p_vmid;
-
-		if (warmup_params.address_increment == 0)
-			warmup_params.address_increment = dc->dml.soc.vmm_page_size_bytes;
-
-		mcif_wb->funcs->warmup_mcif(mcif_wb, &warmup_params);
-		return true;
-	}
-	/*following is the original: warmup each DWB's mcif buffer*/
-	for (i = 0; i < num_dwb; i++) {
-		dwb = dc->res_pool->dwbc[wb_info[i].dwb_pipe_inst];
-		mcif_wb = dc->res_pool->mcif_wb[wb_info[i].dwb_pipe_inst];
-		/*warmup is for VM mode only*/
-		if (wb_info[i].mcif_buf_params.p_vmid == 0)
-			return false;
-
-		/* Warmup MCIF_WB */
-		for (i_buf = 0; i_buf < MCIF_BUF_COUNT; i_buf++) {
-			warmup_params.start_address.quad_part = wb_info[i].mcif_buf_params.luma_address[i_buf];
-			warmup_params.address_increment = dc->dml.soc.vmm_page_size_bytes;
-			warmup_params.region_size = wb_info[i].mcif_buf_params.luma_pitch * wb_info[i].dwb_params.dest_height;
-			warmup_params.p_vmid = wb_info[i].mcif_buf_params.p_vmid;
-			mcif_wb->funcs->warmup_mcif(mcif_wb, &warmup_params);
-		}
-	}
-	return true;
-}
-
-void dcn30_enable_writeback(
-		struct dc *dc,
-		struct dc_writeback_info *wb_info,
-		struct dc_state *context)
-{
-	struct dwbc *dwb;
-	struct mcif_wb *mcif_wb;
-	struct timing_generator *optc;
-
-	dwb = dc->res_pool->dwbc[wb_info->dwb_pipe_inst];
-	mcif_wb = dc->res_pool->mcif_wb[wb_info->dwb_pipe_inst];
-
-	/* set the OPTC source mux */
-	optc = dc->res_pool->timing_generators[dwb->otg_inst];
-	DC_LOG_DWB("%s dwb_pipe_inst = %d, mpcc_inst = %d",\
-		__func__, wb_info->dwb_pipe_inst,\
-		wb_info->mpcc_inst);
-	if (IS_DIAG_DC(dc->ctx->dce_environment)) {
-		/*till diags switch to warmup interface*/
-		dcn30_mmhubbub_warmup(dc, 1, wb_info);
-	}
-	/* Update writeback pipe */
-	dcn30_set_writeback(dc, wb_info, context);
-
-	/* Enable MCIF_WB */
-	mcif_wb->funcs->enable_mcif(mcif_wb);
-	/* Enable DWB */
-	dwb->funcs->enable(dwb, &wb_info->dwb_params);
-}
-
-void dcn30_disable_writeback(
-		struct dc *dc,
-		unsigned int dwb_pipe_inst)
-{
-	struct dwbc *dwb;
-	struct mcif_wb *mcif_wb;
-
-	ASSERT(dwb_pipe_inst < MAX_DWB_PIPES);
-	dwb = dc->res_pool->dwbc[dwb_pipe_inst];
-	mcif_wb = dc->res_pool->mcif_wb[dwb_pipe_inst];
-	DC_LOG_DWB("%s dwb_pipe_inst = %d",\
-		__func__, dwb_pipe_inst);
-
-	/* disable DWB */
-	dwb->funcs->disable(dwb);
-	/* disable MCIF */
-	mcif_wb->funcs->disable_mcif(mcif_wb);
-	/* disable MPC DWB mux */
-	dc->res_pool->mpc->funcs->disable_dwb_mux(dc->res_pool->mpc, dwb_pipe_inst);
-}
-
-void dcn30_program_all_writeback_pipes_in_tree(
-		struct dc *dc,
-		const struct dc_stream_state *stream,
-		struct dc_state *context)
-{
-	struct dc_writeback_info wb_info;
-	struct dwbc *dwb;
-	struct dc_stream_status *stream_status = NULL;
-	int i_wb, i_pipe, i_stream;
-	DC_LOG_DWB("%s", __func__);
-
-	ASSERT(stream);
-	for (i_stream = 0; i_stream < context->stream_count; i_stream++) {
-		if (context->streams[i_stream] == stream) {
-			stream_status = &context->stream_status[i_stream];
-			break;
-		}
-	}
-	ASSERT(stream_status);
-
-	ASSERT(stream->num_wb_info <= dc->res_pool->res_cap->num_dwb);
-	/* For each writeback pipe */
-	for (i_wb = 0; i_wb < stream->num_wb_info; i_wb++) {
-
-		/* copy writeback info to local non-const so mpcc_inst can be set */
-		wb_info = stream->writeback_info[i_wb];
-		if (wb_info.wb_enabled) {
-
-			/* get the MPCC instance for writeback_source_plane */
-			wb_info.mpcc_inst = -1;
-			for (i_pipe = 0; i_pipe < dc->res_pool->pipe_count; i_pipe++) {
-				struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i_pipe];
-
-				if (!pipe_ctx->plane_state)
-					continue;
-
-				if (pipe_ctx->plane_state == wb_info.writeback_source_plane) {
-					wb_info.mpcc_inst = pipe_ctx->plane_res.mpcc_inst;
-					break;
-				}
-			}
-
-			if (wb_info.mpcc_inst == -1) {
-				/* Disable writeback pipe and disconnect from MPCC
-				 * if source plane has been removed
-				 */
-				dc->hwss.disable_writeback(dc, wb_info.dwb_pipe_inst);
-				continue;
-			}
-
-			ASSERT(wb_info.dwb_pipe_inst < dc->res_pool->res_cap->num_dwb);
-			dwb = dc->res_pool->dwbc[wb_info.dwb_pipe_inst];
-			if (dwb->funcs->is_enabled(dwb)) {
-				/* writeback pipe already enabled, only need to update */
-				dc->hwss.update_writeback(dc, &wb_info, context);
-			} else {
-				/* Enable writeback pipe and connect to MPCC */
-				dc->hwss.enable_writeback(dc, &wb_info, context);
-			}
-		} else {
-			/* Disable writeback pipe and disconnect from MPCC */
-			dc->hwss.disable_writeback(dc, wb_info.dwb_pipe_inst);
-		}
-	}
-}
-
-void dcn30_init_hw(struct dc *dc)
-{
-	struct abm **abms = dc->res_pool->multiple_abms;
-	struct dce_hwseq *hws = dc->hwseq;
-	struct dc_bios *dcb = dc->ctx->dc_bios;
-	struct resource_pool *res_pool = dc->res_pool;
-	int i, j;
-	int edp_num;
-	uint32_t backlight = MAX_BACKLIGHT_LEVEL;
-
-	if (dc->clk_mgr && dc->clk_mgr->funcs->init_clocks)
-		dc->clk_mgr->funcs->init_clocks(dc->clk_mgr);
-
-	// Initialize the dccg
-	if (res_pool->dccg->funcs->dccg_init)
-		res_pool->dccg->funcs->dccg_init(res_pool->dccg);
-
-	if (IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) {
-
-		REG_WRITE(REFCLK_CNTL, 0);
-		REG_UPDATE(DCHUBBUB_GLOBAL_TIMER_CNTL, DCHUBBUB_GLOBAL_TIMER_ENABLE, 1);
-		REG_WRITE(DIO_MEM_PWR_CTRL, 0);
-
-		if (!dc->debug.disable_clock_gate) {
-			/* enable all DCN clock gating */
-			REG_WRITE(DCCG_GATE_DISABLE_CNTL, 0);
-
-			REG_WRITE(DCCG_GATE_DISABLE_CNTL2, 0);
-
-			REG_UPDATE(DCFCLK_CNTL, DCFCLK_GATE_DIS, 0);
-		}
-
-		//Enable ability to power gate / don't force power on permanently
-		if (hws->funcs.enable_power_gating_plane)
-			hws->funcs.enable_power_gating_plane(hws, true);
-
-		return;
-	}
-
-	if (!dcb->funcs->is_accelerated_mode(dcb)) {
-		hws->funcs.bios_golden_init(dc);
-		hws->funcs.disable_vga(dc->hwseq);
-	}
-
-	if (dc->debug.enable_mem_low_power.bits.dmcu) {
-		// Force ERAM to shutdown if DMCU is not enabled
-		if (dc->debug.disable_dmcu || dc->config.disable_dmcu) {
-			REG_UPDATE(DMU_MEM_PWR_CNTL, DMCU_ERAM_MEM_PWR_FORCE, 3);
-		}
-	}
-
-	// Set default OPTC memory power states
-	if (dc->debug.enable_mem_low_power.bits.optc) {
-		// Shutdown when unassigned and light sleep in VBLANK
-		REG_SET_2(ODM_MEM_PWR_CTRL3, 0, ODM_MEM_UNASSIGNED_PWR_MODE, 3, ODM_MEM_VBLANK_PWR_MODE, 1);
-	}
-
-	if (dc->ctx->dc_bios->fw_info_valid) {
-		res_pool->ref_clocks.xtalin_clock_inKhz =
-				dc->ctx->dc_bios->fw_info.pll_info.crystal_frequency;
-
-		if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) {
-			if (res_pool->dccg && res_pool->hubbub) {
-
-				(res_pool->dccg->funcs->get_dccg_ref_freq)(res_pool->dccg,
-						dc->ctx->dc_bios->fw_info.pll_info.crystal_frequency,
-						&res_pool->ref_clocks.dccg_ref_clock_inKhz);
-
-				(res_pool->hubbub->funcs->get_dchub_ref_freq)(res_pool->hubbub,
-						res_pool->ref_clocks.dccg_ref_clock_inKhz,
-						&res_pool->ref_clocks.dchub_ref_clock_inKhz);
-			} else {
-				// Not all ASICs have DCCG sw component
-				res_pool->ref_clocks.dccg_ref_clock_inKhz =
-						res_pool->ref_clocks.xtalin_clock_inKhz;
-				res_pool->ref_clocks.dchub_ref_clock_inKhz =
-						res_pool->ref_clocks.xtalin_clock_inKhz;
-			}
-		}
-	} else
-		ASSERT_CRITICAL(false);
-
-	for (i = 0; i < dc->link_count; i++) {
-		/* Power up AND update implementation according to the
-		 * required signal (which may be different from the
-		 * default signal on connector).
-		 */
-		struct dc_link *link = dc->links[i];
-
-		link->link_enc->funcs->hw_init(link->link_enc);
-
-		/* Check for enabled DIG to identify enabled display */
-		if (link->link_enc->funcs->is_dig_enabled &&
-			link->link_enc->funcs->is_dig_enabled(link->link_enc))
-			link->link_status.link_active = true;
-	}
-
-	/* Power gate DSCs */
-	for (i = 0; i < res_pool->res_cap->num_dsc; i++)
-		if (hws->funcs.dsc_pg_control != NULL)
-			hws->funcs.dsc_pg_control(hws, res_pool->dscs[i]->inst, false);
-
-	/* we want to turn off all dp displays before doing detection */
-	if (dc->config.power_down_display_on_boot) {
-		uint8_t dpcd_power_state = '\0';
-		enum dc_status status = DC_ERROR_UNEXPECTED;
-
-		for (i = 0; i < dc->link_count; i++) {
-			if (dc->links[i]->connector_signal != SIGNAL_TYPE_DISPLAY_PORT)
-				continue;
-			/* DP 2.0 states that LTTPR regs must be read first */
-			dp_retrieve_lttpr_cap(dc->links[i]);
-
-			/* if any of the displays are lit up turn them off */
-			status = core_link_read_dpcd(dc->links[i], DP_SET_POWER,
-						     &dpcd_power_state, sizeof(dpcd_power_state));
-			if (status == DC_OK && dpcd_power_state == DP_POWER_STATE_D0) {
-				/* blank dp stream before power off receiver*/
-				if (dc->links[i]->link_enc->funcs->get_dig_frontend) {
-					unsigned int fe;
-
-					fe = dc->links[i]->link_enc->funcs->get_dig_frontend(
-										dc->links[i]->link_enc);
-					if (fe == ENGINE_ID_UNKNOWN)
-						continue;
-
-					for (j = 0; j < dc->res_pool->stream_enc_count; j++) {
-						if (fe == dc->res_pool->stream_enc[j]->id) {
-							dc->res_pool->stream_enc[j]->funcs->dp_blank(dc->links[i],
-										dc->res_pool->stream_enc[j]);
-							break;
-						}
-					}
-				}
-				dp_receiver_power_ctrl(dc->links[i], false);
-			}
-		}
-	}
-
-	/* If taking control over from VBIOS, we may want to optimize our first
-	 * mode set, so we need to skip powering down pipes until we know which
-	 * pipes we want to use.
-	 * Otherwise, if taking control is not possible, we need to power
-	 * everything down.
-	 */
-	if (dcb->funcs->is_accelerated_mode(dcb) || dc->config.power_down_display_on_boot) {
-		hws->funcs.init_pipes(dc, dc->current_state);
-		if (dc->res_pool->hubbub->funcs->allow_self_refresh_control)
-			dc->res_pool->hubbub->funcs->allow_self_refresh_control(dc->res_pool->hubbub,
-					!dc->res_pool->hubbub->ctx->dc->debug.disable_stutter);
-	}
-
-	/* In headless boot cases, DIG may be turned
-	 * on which causes HW/SW discrepancies.
-	 * To avoid this, power down hardware on boot
-	 * if DIG is turned on and seamless boot not enabled
-	 */
-	if (dc->config.power_down_display_on_boot) {
-		struct dc_link *edp_links[MAX_NUM_EDP];
-		struct dc_link *edp_link = NULL;
-
-		get_edp_links(dc, edp_links, &edp_num);
-		if (edp_num)
-			edp_link = edp_links[0];
-		if (edp_link && edp_link->link_enc->funcs->is_dig_enabled &&
-				edp_link->link_enc->funcs->is_dig_enabled(edp_link->link_enc) &&
-				dc->hwss.edp_backlight_control &&
-				dc->hwss.power_down &&
-				dc->hwss.edp_power_control) {
-			dc->hwss.edp_backlight_control(edp_link, false);
-			dc->hwss.power_down(dc);
-			dc->hwss.edp_power_control(edp_link, false);
-		} else {
-			for (i = 0; i < dc->link_count; i++) {
-				struct dc_link *link = dc->links[i];
-
-				if (link->link_enc->funcs->is_dig_enabled &&
-						link->link_enc->funcs->is_dig_enabled(link->link_enc) &&
-						dc->hwss.power_down) {
-					dc->hwss.power_down(dc);
-					break;
-				}
-
-			}
-		}
-	}
-
-	for (i = 0; i < res_pool->audio_count; i++) {
-		struct audio *audio = res_pool->audios[i];
-
-		audio->funcs->hw_init(audio);
-	}
-
-	for (i = 0; i < dc->link_count; i++) {
-		struct dc_link *link = dc->links[i];
-
-		if (link->panel_cntl)
-			backlight = link->panel_cntl->funcs->hw_init(link->panel_cntl);
-	}
-
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		if (abms[i] != NULL)
-			abms[i]->funcs->abm_init(abms[i], backlight);
-	}
-
-	/* power AFMT HDMI memory TODO: may move to dis/en output save power*/
-	REG_WRITE(DIO_MEM_PWR_CTRL, 0);
-
-	if (!dc->debug.disable_clock_gate) {
-		/* enable all DCN clock gating */
-		REG_WRITE(DCCG_GATE_DISABLE_CNTL, 0);
-
-		REG_WRITE(DCCG_GATE_DISABLE_CNTL2, 0);
-
-		REG_UPDATE(DCFCLK_CNTL, DCFCLK_GATE_DIS, 0);
-	}
-	if (hws->funcs.enable_power_gating_plane)
-		hws->funcs.enable_power_gating_plane(dc->hwseq, true);
-
-	if (!dcb->funcs->is_accelerated_mode(dcb) && dc->res_pool->hubbub->funcs->init_watermarks)
-		dc->res_pool->hubbub->funcs->init_watermarks(dc->res_pool->hubbub);
-
-	if (dc->clk_mgr->funcs->notify_wm_ranges)
-		dc->clk_mgr->funcs->notify_wm_ranges(dc->clk_mgr);
-
-	if (dc->clk_mgr->funcs->set_hard_max_memclk)
-		dc->clk_mgr->funcs->set_hard_max_memclk(dc->clk_mgr);
-
-	if (dc->res_pool->hubbub->funcs->force_pstate_change_control)
-		dc->res_pool->hubbub->funcs->force_pstate_change_control(
-				dc->res_pool->hubbub, false, false);
-	if (dc->res_pool->hubbub->funcs->init_crb)
-		dc->res_pool->hubbub->funcs->init_crb(dc->res_pool->hubbub);
-
-}
-
-void dcn30_set_avmute(struct pipe_ctx *pipe_ctx, bool enable)
-{
-	if (pipe_ctx == NULL)
-		return;
-
-	if (dc_is_hdmi_signal(pipe_ctx->stream->signal) && pipe_ctx->stream_res.stream_enc != NULL)
-		pipe_ctx->stream_res.stream_enc->funcs->set_avmute(
-				pipe_ctx->stream_res.stream_enc,
-				enable);
-}
-
-void dcn30_update_info_frame(struct pipe_ctx *pipe_ctx)
-{
-	bool is_hdmi_tmds;
-	bool is_dp;
-
-	ASSERT(pipe_ctx->stream);
-
-	if (pipe_ctx->stream_res.stream_enc == NULL)
-		return;  /* this is not root pipe */
-
-	is_hdmi_tmds = dc_is_hdmi_tmds_signal(pipe_ctx->stream->signal);
-	is_dp = dc_is_dp_signal(pipe_ctx->stream->signal);
-
-	if (!is_hdmi_tmds && !is_dp)
-		return;
-
-	if (is_hdmi_tmds)
-		pipe_ctx->stream_res.stream_enc->funcs->update_hdmi_info_packets(
-			pipe_ctx->stream_res.stream_enc,
-			&pipe_ctx->stream_res.encoder_info_frame);
-	else
-		pipe_ctx->stream_res.stream_enc->funcs->update_dp_info_packets(
-			pipe_ctx->stream_res.stream_enc,
-			&pipe_ctx->stream_res.encoder_info_frame);
-}
-
-void dcn30_program_dmdata_engine(struct pipe_ctx *pipe_ctx)
-{
-	struct dc_stream_state    *stream     = pipe_ctx->stream;
-	struct hubp               *hubp       = pipe_ctx->plane_res.hubp;
-	bool                       enable     = false;
-	struct stream_encoder     *stream_enc = pipe_ctx->stream_res.stream_enc;
-	enum dynamic_metadata_mode mode       = dc_is_dp_signal(stream->signal)
-							? dmdata_dp
-							: dmdata_hdmi;
-
-	/* if using dynamic meta, don't set up generic infopackets */
-	if (pipe_ctx->stream->dmdata_address.quad_part != 0) {
-		pipe_ctx->stream_res.encoder_info_frame.hdrsmd.valid = false;
-		enable = true;
-	}
-
-	if (!hubp)
-		return;
-
-	if (!stream_enc || !stream_enc->funcs->set_dynamic_metadata)
-		return;
-
-	stream_enc->funcs->set_dynamic_metadata(stream_enc, enable,
-							hubp->inst, mode);
-}
-
-bool dcn30_apply_idle_power_optimizations(struct dc *dc, bool enable)
-{
-	union dmub_rb_cmd cmd;
-	uint32_t tmr_delay = 0, tmr_scale = 0;
-	struct dc_cursor_attributes cursor_attr;
-	bool cursor_cache_enable = false;
-	struct dc_stream_state *stream = NULL;
-	struct dc_plane_state *plane = NULL;
-
-	if (!dc->ctx->dmub_srv)
-		return false;
-
-	if (enable) {
-		if (dc->current_state) {
-			int i;
-
-			/* First, check no-memory-requests case */
-			for (i = 0; i < dc->current_state->stream_count; i++) {
-				if (dc->current_state->stream_status[i].plane_count)
-					/* Fail eligibility on a visible stream */
-					break;
-			}
-
-			if (i == dc->current_state->stream_count) {
-				/* Enable no-memory-requests case */
-				memset(&cmd, 0, sizeof(cmd));
-				cmd.mall.header.type = DMUB_CMD__MALL;
-				cmd.mall.header.sub_type = DMUB_CMD__MALL_ACTION_NO_DF_REQ;
-				cmd.mall.header.payload_bytes = sizeof(cmd.mall) - sizeof(cmd.mall.header);
-
-				dc_dmub_srv_cmd_queue(dc->ctx->dmub_srv, &cmd);
-				dc_dmub_srv_cmd_execute(dc->ctx->dmub_srv);
-
-				return true;
-			}
-
-			stream = dc->current_state->streams[0];
-			plane = (stream ? dc->current_state->stream_status[0].plane_states[0] : NULL);
-
-			if (stream && plane) {
-				cursor_cache_enable = stream->cursor_position.enable &&
-						plane->address.grph.cursor_cache_addr.quad_part;
-				cursor_attr = stream->cursor_attributes;
-			}
-
-			/*
-			 * Second, check MALL eligibility
-			 *
-			 * single display only, single surface only, 8 and 16 bit formats only, no VM,
-			 * do not use MALL for displays that support PSR as they use D0i3.2 in DMCUB FW
-			 *
-			 * TODO: When we implement multi-display, PSR displays will be allowed if there is
-			 * a non-PSR display present, since in that case we can't do D0i3.2
-			 */
-			if (dc->current_state->stream_count == 1 &&
-					stream->link->psr_settings.psr_version == DC_PSR_VERSION_UNSUPPORTED &&
-					dc->current_state->stream_status[0].plane_count == 1 &&
-					plane->format <= SURFACE_PIXEL_FORMAT_GRPH_ABGR16161616F &&
-					plane->format >= SURFACE_PIXEL_FORMAT_GRPH_ARGB8888 &&
-					plane->address.page_table_base.quad_part == 0 &&
-					dc->hwss.does_plane_fit_in_mall &&
-					dc->hwss.does_plane_fit_in_mall(dc, plane,
-							cursor_cache_enable ? &cursor_attr : NULL)) {
-				unsigned int v_total = stream->adjust.v_total_max ?
-						stream->adjust.v_total_max : stream->timing.v_total;
-				unsigned int refresh_hz = div_u64((unsigned long long) stream->timing.pix_clk_100hz *
-						100LL, (v_total * stream->timing.h_total));
-
-				/*
-				 * one frame time in microsec:
-				 * Delay_Us = 1000000 / refresh
-				 * dynamic_delay_us = 1000000 / refresh + 2 * stutter_period
-				 *
-				 * one frame time modified by 'additional timer percent' (p):
-				 * Delay_Us_modified = dynamic_delay_us + dynamic_delay_us * p / 100
-				 *                   = dynamic_delay_us * (1 + p / 100)
-				 *                   = (1000000 / refresh + 2 * stutter_period) * (100 + p) / 100
-				 *                   = (1000000 + 2 * stutter_period * refresh) * (100 + p) / (100 * refresh)
-				 *
-				 * formula for timer duration based on parameters, from regspec:
-				 * dynamic_delay_us = 65.28 * (64 + MallFrameCacheTmrDly) * 2^MallFrameCacheTmrScale
-				 *
-				 * dynamic_delay_us / 65.28 = (64 + MallFrameCacheTmrDly) * 2^MallFrameCacheTmrScale
-				 * (dynamic_delay_us / 65.28) / 2^MallFrameCacheTmrScale = 64 + MallFrameCacheTmrDly
-				 * MallFrameCacheTmrDly = ((dynamic_delay_us / 65.28) / 2^MallFrameCacheTmrScale) - 64
-				 *                      = (1000000 + 2 * stutter_period * refresh) * (100 + p) / (100 * refresh) / 65.28 / 2^MallFrameCacheTmrScale - 64
-				 *                      = (1000000 + 2 * stutter_period * refresh) * (100 + p) / (refresh * 6528 * 2^MallFrameCacheTmrScale) - 64
-				 *
-				 * need to round up the result of the division before the subtraction
-				 */
-				unsigned int denom = refresh_hz * 6528;
-				unsigned int stutter_period = dc->current_state->perf_params.stutter_period_us;
-
-				tmr_delay = div_u64(((1000000LL + 2 * stutter_period * refresh_hz) *
-						(100LL + dc->debug.mall_additional_timer_percent) + denom - 1),
-						denom) - 64LL;
-
-				/* In some cases the stutter period is really big (tiny modes) in these
-				 * cases MALL cant be enabled, So skip these cases to avoid a ASSERT()
-				 *
-				 * We can check if stutter_period is more than 1/10th the frame time to
-				 * consider if we can actually meet the range of hysteresis timer
-				 */
-				if (stutter_period > 100000/refresh_hz)
-					return false;
-
-				/* scale should be increased until it fits into 6 bits */
-				while (tmr_delay & ~0x3F) {
-					tmr_scale++;
-
-					if (tmr_scale > 3) {
-						/* Delay exceeds range of hysteresis timer */
-						ASSERT(false);
-						return false;
-					}
-
-					denom *= 2;
-					tmr_delay = div_u64(((1000000LL + 2 * stutter_period * refresh_hz) *
-							(100LL + dc->debug.mall_additional_timer_percent) + denom - 1),
-							denom) - 64LL;
-				}
-
-				/* Copy HW cursor */
-				if (cursor_cache_enable) {
-					memset(&cmd, 0, sizeof(cmd));
-					cmd.mall.header.type = DMUB_CMD__MALL;
-					cmd.mall.header.sub_type = DMUB_CMD__MALL_ACTION_COPY_CURSOR;
-					cmd.mall.header.payload_bytes =
-							sizeof(cmd.mall) - sizeof(cmd.mall.header);
-
-					switch (cursor_attr.color_format) {
-					case CURSOR_MODE_MONO:
-						cmd.mall.cursor_bpp = 2;
-						break;
-					case CURSOR_MODE_COLOR_1BIT_AND:
-					case CURSOR_MODE_COLOR_PRE_MULTIPLIED_ALPHA:
-					case CURSOR_MODE_COLOR_UN_PRE_MULTIPLIED_ALPHA:
-						cmd.mall.cursor_bpp = 32;
-						break;
-
-					case CURSOR_MODE_COLOR_64BIT_FP_PRE_MULTIPLIED:
-					case CURSOR_MODE_COLOR_64BIT_FP_UN_PRE_MULTIPLIED:
-						cmd.mall.cursor_bpp = 64;
-						break;
-					}
-
-					cmd.mall.cursor_copy_src.quad_part = cursor_attr.address.quad_part;
-					cmd.mall.cursor_copy_dst.quad_part =
-							(plane->address.grph.cursor_cache_addr.quad_part + 2047) & ~2047;
-					cmd.mall.cursor_width = cursor_attr.width;
-					cmd.mall.cursor_height = cursor_attr.height;
-					cmd.mall.cursor_pitch = cursor_attr.pitch;
-
-					dc_dmub_srv_cmd_queue(dc->ctx->dmub_srv, &cmd);
-					dc_dmub_srv_cmd_execute(dc->ctx->dmub_srv);
-					dc_dmub_srv_wait_idle(dc->ctx->dmub_srv);
-
-					/* Use copied cursor, and it's okay to not switch back */
-					cursor_attr.address.quad_part = cmd.mall.cursor_copy_dst.quad_part;
-					dc_stream_set_cursor_attributes(stream, &cursor_attr);
-				}
-
-				/* Enable MALL */
-				memset(&cmd, 0, sizeof(cmd));
-				cmd.mall.header.type = DMUB_CMD__MALL;
-				cmd.mall.header.sub_type = DMUB_CMD__MALL_ACTION_ALLOW;
-				cmd.mall.header.payload_bytes = sizeof(cmd.mall) - sizeof(cmd.mall.header);
-				cmd.mall.tmr_delay = tmr_delay;
-				cmd.mall.tmr_scale = tmr_scale;
-				cmd.mall.debug_bits = dc->debug.mall_error_as_fatal;
-
-				dc_dmub_srv_cmd_queue(dc->ctx->dmub_srv, &cmd);
-				dc_dmub_srv_cmd_execute(dc->ctx->dmub_srv);
-
-				return true;
-			}
-		}
-
-		/* No applicable optimizations */
-		return false;
-	}
-
-	/* Disable MALL */
-	memset(&cmd, 0, sizeof(cmd));
-	cmd.mall.header.type = DMUB_CMD__MALL;
-	cmd.mall.header.sub_type = DMUB_CMD__MALL_ACTION_DISALLOW;
-	cmd.mall.header.payload_bytes =
-		sizeof(cmd.mall) - sizeof(cmd.mall.header);
-
-	dc_dmub_srv_cmd_queue(dc->ctx->dmub_srv, &cmd);
-	dc_dmub_srv_cmd_execute(dc->ctx->dmub_srv);
-	dc_dmub_srv_wait_idle(dc->ctx->dmub_srv);
-
-	return true;
-}
-
-bool dcn30_does_plane_fit_in_mall(struct dc *dc, struct dc_plane_state *plane, struct dc_cursor_attributes *cursor_attr)
-{
-	// add meta size?
-	unsigned int surface_size = plane->plane_size.surface_pitch * plane->plane_size.surface_size.height *
-			(plane->format >= SURFACE_PIXEL_FORMAT_GRPH_ARGB16161616 ? 8 : 4);
-	unsigned int mall_size = dc->caps.mall_size_total;
-	unsigned int cursor_size = 0;
-
-	if (dc->debug.mall_size_override)
-		mall_size = 1024 * 1024 * dc->debug.mall_size_override;
-
-	if (cursor_attr) {
-		cursor_size = dc->caps.max_cursor_size * dc->caps.max_cursor_size;
-
-		switch (cursor_attr->color_format) {
-		case CURSOR_MODE_MONO:
-			cursor_size /= 2;
-			break;
-		case CURSOR_MODE_COLOR_1BIT_AND:
-		case CURSOR_MODE_COLOR_PRE_MULTIPLIED_ALPHA:
-		case CURSOR_MODE_COLOR_UN_PRE_MULTIPLIED_ALPHA:
-			cursor_size *= 4;
-			break;
-
-		case CURSOR_MODE_COLOR_64BIT_FP_PRE_MULTIPLIED:
-		case CURSOR_MODE_COLOR_64BIT_FP_UN_PRE_MULTIPLIED:
-			cursor_size *= 8;
-			break;
-		}
-	}
-
-	return (surface_size + cursor_size) < mall_size;
-}
-
-void dcn30_hardware_release(struct dc *dc)
-{
-	/* if pstate unsupported, force it supported */
-	if (!dc->clk_mgr->clks.p_state_change_support &&
-			dc->res_pool->hubbub->funcs->force_pstate_change_control)
-		dc->res_pool->hubbub->funcs->force_pstate_change_control(
-				dc->res_pool->hubbub, true, true);
-}
-
-void dcn30_set_disp_pattern_generator(const struct dc *dc,
-		struct pipe_ctx *pipe_ctx,
-		enum controller_dp_test_pattern test_pattern,
-		enum controller_dp_color_space color_space,
-		enum dc_color_depth color_depth,
-		const struct tg_color *solid_color,
-		int width, int height, int offset)
-{
-	struct stream_resource *stream_res = &pipe_ctx->stream_res;
-	struct pipe_ctx *mpcc_pipe;
-
-	if (test_pattern != CONTROLLER_DP_TEST_PATTERN_VIDEOMODE) {
-		pipe_ctx->vtp_locked = false;
-		/* turning on DPG */
-		stream_res->opp->funcs->opp_set_disp_pattern_generator(stream_res->opp, test_pattern, color_space,
-				color_depth, solid_color, width, height, offset);
-
-		/* Defer hubp blank if tg is locked */
-		if (stream_res->tg->funcs->is_tg_enabled(stream_res->tg)) {
-			if (stream_res->tg->funcs->is_locked(stream_res->tg))
-				pipe_ctx->vtp_locked = true;
-			else {
-				/* Blank HUBP to allow p-state during blank on all timings */
-				pipe_ctx->plane_res.hubp->funcs->set_blank(pipe_ctx->plane_res.hubp, true);
-
-				for (mpcc_pipe = pipe_ctx->bottom_pipe; mpcc_pipe; mpcc_pipe = mpcc_pipe->bottom_pipe)
-					mpcc_pipe->plane_res.hubp->funcs->set_blank(mpcc_pipe->plane_res.hubp, true);
-			}
-		}
-	} else {
-		/* turning off DPG */
-		pipe_ctx->plane_res.hubp->funcs->set_blank(pipe_ctx->plane_res.hubp, false);
-		for (mpcc_pipe = pipe_ctx->bottom_pipe; mpcc_pipe; mpcc_pipe = mpcc_pipe->bottom_pipe)
-			if (mpcc_pipe->plane_res.hubp)
-				mpcc_pipe->plane_res.hubp->funcs->set_blank(mpcc_pipe->plane_res.hubp, false);
-
-		stream_res->opp->funcs->opp_set_disp_pattern_generator(stream_res->opp, test_pattern, color_space,
-				color_depth, solid_color, width, height, offset);
-	}
-}
diff --git a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_mmhubbub.c b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_mmhubbub.c
index 1c4b171c68ad..6f2a0d5d963b 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_mmhubbub.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_mmhubbub.c
@@ -100,7 +100,7 @@ static void mmhubbub3_warmup_mcif(struct mcif_wb *mcif_wb,
 	REG_UPDATE(MMHUBBUB_WARMUP_CONTROL_STATUS, MMHUBBUB_WARMUP_EN, false);
 }
 
-void mmhubbub3_config_mcif_buf(struct mcif_wb *mcif_wb,
+static void mmhubbub3_config_mcif_buf(struct mcif_wb *mcif_wb,
 		struct mcif_buf_params *params,
 		unsigned int dest_height)
 {
@@ -211,7 +211,7 @@ static void mmhubbub3_config_mcif_arb(struct mcif_wb *mcif_wb,
 	REG_UPDATE(MCIF_WB_ARBITRATION_CONTROL, MCIF_WB_CLIENT_ARBITRATION_SLICE,  params->arbitration_slice);
 }
 
-const struct mcif_wb_funcs dcn30_mmhubbub_funcs = {
+static const struct mcif_wb_funcs dcn30_mmhubbub_funcs = {
 	.warmup_mcif		= mmhubbub3_warmup_mcif,
 	.enable_mcif		= mmhubbub2_enable_mcif,
 	.disable_mcif		= mmhubbub2_disable_mcif,
diff --git a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_mmhubbub.h b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_mmhubbub.h
index f2580e65196c..376620a8f02f 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_mmhubbub.h
+++ b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_mmhubbub.h
@@ -31,13 +31,6 @@
 #define TO_DCN30_MMHUBBUB(mcif_wb_base) \
 	container_of(mcif_wb_base, struct dcn30_mmhubbub, base)
 
-/* DCN */
-#define BASE_INNER(seg) \
-	DCE_BASE__INST0_SEG ## seg
-
-#define BASE(seg) \
-	BASE_INNER(seg)
-
 #define MCIF_WB_COMMON_REG_LIST_DCN3_0(inst) \
 	SRI(MCIF_WB_BUFMGR_SW_CONTROL, MCIF_WB, inst),\
 	SRI(MCIF_WB_BUFMGR_STATUS, MCIF_WB, inst),\
@@ -227,11 +220,7 @@
 	SF(MCIF_WB0_MCIF_WB_BUF_3_ADDR_C, MCIF_WB_BUF_3_ADDR_C, mask_sh),\
 	SF(MCIF_WB0_MCIF_WB_BUF_4_ADDR_Y, MCIF_WB_BUF_4_ADDR_Y, mask_sh),\
 	SF(MCIF_WB0_MCIF_WB_BUF_4_ADDR_C, MCIF_WB_BUF_4_ADDR_C, mask_sh),\
-	SF(MCIF_WB0_MCIF_WB_BUFMGR_VCE_CONTROL, MCIF_WB_BUFMGR_VCE_LOCK_IGNORE, mask_sh),\
-	SF(MCIF_WB0_MCIF_WB_BUFMGR_VCE_CONTROL, MCIF_WB_BUFMGR_VCE_INT_EN, mask_sh),\
-	SF(MCIF_WB0_MCIF_WB_BUFMGR_VCE_CONTROL, MCIF_WB_BUFMGR_VCE_INT_ACK, mask_sh),\
 	SF(MCIF_WB0_MCIF_WB_BUFMGR_VCE_CONTROL, MCIF_WB_BUFMGR_VCE_SLICE_INT_EN, mask_sh),\
-	SF(MCIF_WB0_MCIF_WB_BUFMGR_VCE_CONTROL, MCIF_WB_BUFMGR_VCE_LOCK, mask_sh),\
 	SF(MCIF_WB0_MCIF_WB_BUFMGR_VCE_CONTROL, MCIF_WB_BUFMGR_SLICE_SIZE, mask_sh),\
 	SF(MCIF_WB_NB_PSTATE_LATENCY_WATERMARK, NB_PSTATE_CHANGE_REFRESH_WATERMARK, mask_sh),\
 	SF(MCIF_WB_NB_PSTATE_LATENCY_WATERMARK, NB_PSTATE_CHANGE_WATERMARK_MASK, mask_sh),\
@@ -363,11 +352,7 @@
 	SF(MCIF_WB_BUF_3_ADDR_C, MCIF_WB_BUF_3_ADDR_C, mask_sh),\
 	SF(MCIF_WB_BUF_4_ADDR_Y, MCIF_WB_BUF_4_ADDR_Y, mask_sh),\
 	SF(MCIF_WB_BUF_4_ADDR_C, MCIF_WB_BUF_4_ADDR_C, mask_sh),\
-	SF(MCIF_WB_BUFMGR_VCE_CONTROL, MCIF_WB_BUFMGR_VCE_LOCK_IGNORE, mask_sh),\
-	SF(MCIF_WB_BUFMGR_VCE_CONTROL, MCIF_WB_BUFMGR_VCE_INT_EN, mask_sh),\
-	SF(MCIF_WB_BUFMGR_VCE_CONTROL, MCIF_WB_BUFMGR_VCE_INT_ACK, mask_sh),\
 	SF(MCIF_WB_BUFMGR_VCE_CONTROL, MCIF_WB_BUFMGR_VCE_SLICE_INT_EN, mask_sh),\
-	SF(MCIF_WB_BUFMGR_VCE_CONTROL, MCIF_WB_BUFMGR_VCE_LOCK, mask_sh),\
 	SF(MCIF_WB_BUFMGR_VCE_CONTROL, MCIF_WB_BUFMGR_SLICE_SIZE, mask_sh),\
 	SF(MCIF_WB_NB_PSTATE_LATENCY_WATERMARK, NB_PSTATE_CHANGE_REFRESH_WATERMARK, mask_sh),\
 	SF(MCIF_WB_NB_PSTATE_LATENCY_WATERMARK, NB_PSTATE_CHANGE_WATERMARK_MASK, mask_sh),\
diff --git a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_resource.c b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_resource.c
deleted file mode 100644
index 79a66e0c4303..000000000000
--- a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_resource.c
+++ /dev/null
@@ -1,2918 +0,0 @@
-/*
- * Copyright 2020 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: AMD
- *
- */
-
-
-#include "dm_services.h"
-#include "dc.h"
-
-#include "dcn30_init.h"
-
-#include "resource.h"
-#include "include/irq_service_interface.h"
-#include "dcn20/dcn20_resource.h"
-
-#include "dcn30_resource.h"
-
-#include "dcn10/dcn10_ipp.h"
-#include "dcn30/dcn30_hubbub.h"
-#include "dcn30/dcn30_mpc.h"
-#include "dcn30/dcn30_hubp.h"
-#include "irq/dcn30/irq_service_dcn30.h"
-#include "dcn30/dcn30_dpp.h"
-#include "dcn30/dcn30_optc.h"
-#include "dcn20/dcn20_hwseq.h"
-#include "dcn30/dcn30_hwseq.h"
-#include "dce110/dce110_hw_sequencer.h"
-#include "dcn30/dcn30_opp.h"
-#include "dcn20/dcn20_dsc.h"
-#include "dcn30/dcn30_vpg.h"
-#include "dcn30/dcn30_afmt.h"
-#include "dcn30/dcn30_dio_stream_encoder.h"
-#include "dcn30/dcn30_dio_link_encoder.h"
-#include "dce/dce_clock_source.h"
-#include "dce/dce_audio.h"
-#include "dce/dce_hwseq.h"
-#include "clk_mgr.h"
-#include "virtual/virtual_stream_encoder.h"
-#include "dce110/dce110_resource.h"
-#include "dml/display_mode_vba.h"
-#include "dcn30/dcn30_dccg.h"
-#include "dcn10/dcn10_resource.h"
-#include "dc_link_ddc.h"
-#include "dce/dce_panel_cntl.h"
-
-#include "dcn30/dcn30_dwb.h"
-#include "dcn30/dcn30_mmhubbub.h"
-
-#include "sienna_cichlid_ip_offset.h"
-#include "dcn/dcn_3_0_0_offset.h"
-#include "dcn/dcn_3_0_0_sh_mask.h"
-
-#include "nbio/nbio_7_4_offset.h"
-
-#include "dcn/dpcs_3_0_0_offset.h"
-#include "dcn/dpcs_3_0_0_sh_mask.h"
-
-#include "mmhub/mmhub_2_0_0_offset.h"
-#include "mmhub/mmhub_2_0_0_sh_mask.h"
-
-#include "reg_helper.h"
-#include "dce/dmub_abm.h"
-#include "dce/dmub_psr.h"
-#include "dce/dce_aux.h"
-#include "dce/dce_i2c.h"
-
-#include "dml/dcn30/display_mode_vba_30.h"
-#include "vm_helper.h"
-#include "dcn20/dcn20_vmid.h"
-#include "amdgpu_socbb.h"
-
-#define DC_LOGGER_INIT(logger)
-
-struct _vcs_dpi_ip_params_st dcn3_0_ip = {
-	.use_min_dcfclk = 0,
-	.clamp_min_dcfclk = 0,
-	.odm_capable = 1,
-	.gpuvm_enable = 0,
-	.hostvm_enable = 0,
-	.gpuvm_max_page_table_levels = 4,
-	.hostvm_max_page_table_levels = 4,
-	.hostvm_cached_page_table_levels = 0,
-	.pte_group_size_bytes = 2048,
-	.num_dsc = 6,
-	.rob_buffer_size_kbytes = 184,
-	.det_buffer_size_kbytes = 184,
-	.dpte_buffer_size_in_pte_reqs_luma = 84,
-	.pde_proc_buffer_size_64k_reqs = 48,
-	.dpp_output_buffer_pixels = 2560,
-	.opp_output_buffer_lines = 1,
-	.pixel_chunk_size_kbytes = 8,
-	.pte_enable = 1,
-	.max_page_table_levels = 2,
-	.pte_chunk_size_kbytes = 2,  // ?
-	.meta_chunk_size_kbytes = 2,
-	.writeback_chunk_size_kbytes = 8,
-	.line_buffer_size_bits = 789504,
-	.is_line_buffer_bpp_fixed = 0,  // ?
-	.line_buffer_fixed_bpp = 0,     // ?
-	.dcc_supported = true,
-	.writeback_interface_buffer_size_kbytes = 90,
-	.writeback_line_buffer_buffer_size = 0,
-	.max_line_buffer_lines = 12,
-	.writeback_luma_buffer_size_kbytes = 12,  // writeback_line_buffer_buffer_size = 656640
-	.writeback_chroma_buffer_size_kbytes = 8,
-	.writeback_chroma_line_buffer_width_pixels = 4,
-	.writeback_max_hscl_ratio = 1,
-	.writeback_max_vscl_ratio = 1,
-	.writeback_min_hscl_ratio = 1,
-	.writeback_min_vscl_ratio = 1,
-	.writeback_max_hscl_taps = 1,
-	.writeback_max_vscl_taps = 1,
-	.writeback_line_buffer_luma_buffer_size = 0,
-	.writeback_line_buffer_chroma_buffer_size = 14643,
-	.cursor_buffer_size = 8,
-	.cursor_chunk_size = 2,
-	.max_num_otg = 6,
-	.max_num_dpp = 6,
-	.max_num_wb = 1,
-	.max_dchub_pscl_bw_pix_per_clk = 4,
-	.max_pscl_lb_bw_pix_per_clk = 2,
-	.max_lb_vscl_bw_pix_per_clk = 4,
-	.max_vscl_hscl_bw_pix_per_clk = 4,
-	.max_hscl_ratio = 6,
-	.max_vscl_ratio = 6,
-	.hscl_mults = 4,
-	.vscl_mults = 4,
-	.max_hscl_taps = 8,
-	.max_vscl_taps = 8,
-	.dispclk_ramp_margin_percent = 1,
-	.underscan_factor = 1.11,
-	.min_vblank_lines = 32,
-	.dppclk_delay_subtotal = 46,
-	.dynamic_metadata_vm_enabled = true,
-	.dppclk_delay_scl_lb_only = 16,
-	.dppclk_delay_scl = 50,
-	.dppclk_delay_cnvc_formatter = 27,
-	.dppclk_delay_cnvc_cursor = 6,
-	.dispclk_delay_subtotal = 119,
-	.dcfclk_cstate_latency = 5.2, // SRExitTime
-	.max_inter_dcn_tile_repeaters = 8,
-	.odm_combine_4to1_supported = true,
-
-	.xfc_supported = false,
-	.xfc_fill_bw_overhead_percent = 10.0,
-	.xfc_fill_constant_bytes = 0,
-	.gfx7_compat_tiling_supported = 0,
-	.number_of_cursors = 1,
-};
-
-struct _vcs_dpi_soc_bounding_box_st dcn3_0_soc = {
-	.clock_limits = {
-			{
-				.state = 0,
-				.dispclk_mhz = 562.0,
-				.dppclk_mhz = 300.0,
-				.phyclk_mhz = 300.0,
-				.phyclk_d18_mhz = 667.0,
-				.dscclk_mhz = 405.6,
-			},
-		},
-	.min_dcfclk = 500.0, /* TODO: set this to actual min DCFCLK */
-	.num_states = 1,
-	.sr_exit_time_us = 15.5,
-	.sr_enter_plus_exit_time_us = 20,
-	.urgent_latency_us = 4.0,
-	.urgent_latency_pixel_data_only_us = 4.0,
-	.urgent_latency_pixel_mixed_with_vm_data_us = 4.0,
-	.urgent_latency_vm_data_only_us = 4.0,
-	.urgent_out_of_order_return_per_channel_pixel_only_bytes = 4096,
-	.urgent_out_of_order_return_per_channel_pixel_and_vm_bytes = 4096,
-	.urgent_out_of_order_return_per_channel_vm_only_bytes = 4096,
-	.pct_ideal_dram_sdp_bw_after_urgent_pixel_only = 80.0,
-	.pct_ideal_dram_sdp_bw_after_urgent_pixel_and_vm = 60.0,
-	.pct_ideal_dram_sdp_bw_after_urgent_vm_only = 40.0,
-	.max_avg_sdp_bw_use_normal_percent = 60.0,
-	.max_avg_dram_bw_use_normal_percent = 40.0,
-	.writeback_latency_us = 12.0,
-	.max_request_size_bytes = 256,
-	.fabric_datapath_to_dcn_data_return_bytes = 64,
-	.dcn_downspread_percent = 0.5,
-	.downspread_percent = 0.38,
-	.dram_page_open_time_ns = 50.0,
-	.dram_rw_turnaround_time_ns = 17.5,
-	.dram_return_buffer_per_channel_bytes = 8192,
-	.round_trip_ping_latency_dcfclk_cycles = 191,
-	.urgent_out_of_order_return_per_channel_bytes = 4096,
-	.channel_interleave_bytes = 256,
-	.num_banks = 8,
-	.gpuvm_min_page_size_bytes = 4096,
-	.hostvm_min_page_size_bytes = 4096,
-	.dram_clock_change_latency_us = 404,
-	.dummy_pstate_latency_us = 5,
-	.writeback_dram_clock_change_latency_us = 23.0,
-	.return_bus_width_bytes = 64,
-	.dispclk_dppclk_vco_speed_mhz = 3650,
-	.xfc_bus_transport_time_us = 20,      // ?
-	.xfc_xbuf_latency_tolerance_us = 4,  // ?
-	.use_urgent_burst_bw = 1,            // ?
-	.do_urgent_latency_adjustment = true,
-	.urgent_latency_adjustment_fabric_clock_component_us = 1.0,
-	.urgent_latency_adjustment_fabric_clock_reference_mhz = 1000,
-};
-
-enum dcn30_clk_src_array_id {
-	DCN30_CLK_SRC_PLL0,
-	DCN30_CLK_SRC_PLL1,
-	DCN30_CLK_SRC_PLL2,
-	DCN30_CLK_SRC_PLL3,
-	DCN30_CLK_SRC_PLL4,
-	DCN30_CLK_SRC_PLL5,
-	DCN30_CLK_SRC_TOTAL
-};
-
-/* begin *********************
- * macros to expend register list macro defined in HW object header file
- */
-
-/* DCN */
-/* TODO awful hack. fixup dcn20_dwb.h */
-#undef BASE_INNER
-#define BASE_INNER(seg) DCN_BASE__INST0_SEG ## seg
-
-#define BASE(seg) BASE_INNER(seg)
-
-#define SR(reg_name)\
-		.reg_name = BASE(mm ## reg_name ## _BASE_IDX) +  \
-					mm ## reg_name
-
-#define SRI(reg_name, block, id)\
-	.reg_name = BASE(mm ## block ## id ## _ ## reg_name ## _BASE_IDX) + \
-					mm ## block ## id ## _ ## reg_name
-
-#define SRI2(reg_name, block, id)\
-	.reg_name = BASE(mm ## reg_name ## _BASE_IDX) + \
-					mm ## reg_name
-
-#define SRIR(var_name, reg_name, block, id)\
-	.var_name = BASE(mm ## block ## id ## _ ## reg_name ## _BASE_IDX) + \
-					mm ## block ## id ## _ ## reg_name
-
-#define SRII(reg_name, block, id)\
-	.reg_name[id] = BASE(mm ## block ## id ## _ ## reg_name ## _BASE_IDX) + \
-					mm ## block ## id ## _ ## reg_name
-
-#define SRII_MPC_RMU(reg_name, block, id)\
-	.RMU##_##reg_name[id] = BASE(mm ## block ## id ## _ ## reg_name ## _BASE_IDX) + \
-					mm ## block ## id ## _ ## reg_name
-
-#define SRII_DWB(reg_name, temp_name, block, id)\
-	.reg_name[id] = BASE(mm ## block ## id ## _ ## temp_name ## _BASE_IDX) + \
-					mm ## block ## id ## _ ## temp_name
-
-#define DCCG_SRII(reg_name, block, id)\
-	.block ## _ ## reg_name[id] = BASE(mm ## block ## id ## _ ## reg_name ## _BASE_IDX) + \
-					mm ## block ## id ## _ ## reg_name
-
-#define VUPDATE_SRII(reg_name, block, id)\
-	.reg_name[id] = BASE(mm ## reg_name ## _ ## block ## id ## _BASE_IDX) + \
-					mm ## reg_name ## _ ## block ## id
-
-/* NBIO */
-#define NBIO_BASE_INNER(seg) \
-	NBIO_BASE__INST0_SEG ## seg
-
-#define NBIO_BASE(seg) \
-	NBIO_BASE_INNER(seg)
-
-#define NBIO_SR(reg_name)\
-		.reg_name = NBIO_BASE(mm ## reg_name ## _BASE_IDX) + \
-					mm ## reg_name
-
-/* MMHUB */
-#define MMHUB_BASE_INNER(seg) \
-	MMHUB_BASE__INST0_SEG ## seg
-
-#define MMHUB_BASE(seg) \
-	MMHUB_BASE_INNER(seg)
-
-#define MMHUB_SR(reg_name)\
-		.reg_name = MMHUB_BASE(mmMM ## reg_name ## _BASE_IDX) + \
-					mmMM ## reg_name
-
-/* CLOCK */
-#define CLK_BASE_INNER(seg) \
-	CLK_BASE__INST0_SEG ## seg
-
-#define CLK_BASE(seg) \
-	CLK_BASE_INNER(seg)
-
-#define CLK_SRI(reg_name, block, inst)\
-	.reg_name = CLK_BASE(mm ## block ## _ ## inst ## _ ## reg_name ## _BASE_IDX) + \
-					mm ## block ## _ ## inst ## _ ## reg_name
-
-
-static const struct bios_registers bios_regs = {
-		NBIO_SR(BIOS_SCRATCH_3),
-		NBIO_SR(BIOS_SCRATCH_6)
-};
-
-#define clk_src_regs(index, pllid)\
-[index] = {\
-	CS_COMMON_REG_LIST_DCN2_0(index, pllid),\
-}
-
-static const struct dce110_clk_src_regs clk_src_regs[] = {
-	clk_src_regs(0, A),
-	clk_src_regs(1, B),
-	clk_src_regs(2, C),
-	clk_src_regs(3, D),
-	clk_src_regs(4, E),
-	clk_src_regs(5, F)
-};
-
-static const struct dce110_clk_src_shift cs_shift = {
-		CS_COMMON_MASK_SH_LIST_DCN2_0(__SHIFT)
-};
-
-static const struct dce110_clk_src_mask cs_mask = {
-		CS_COMMON_MASK_SH_LIST_DCN2_0(_MASK)
-};
-
-#define abm_regs(id)\
-[id] = {\
-		ABM_DCN30_REG_LIST(id)\
-}
-
-static const struct dce_abm_registers abm_regs[] = {
-		abm_regs(0),
-		abm_regs(1),
-		abm_regs(2),
-		abm_regs(3),
-		abm_regs(4),
-		abm_regs(5),
-};
-
-static const struct dce_abm_shift abm_shift = {
-		ABM_MASK_SH_LIST_DCN30(__SHIFT)
-};
-
-static const struct dce_abm_mask abm_mask = {
-		ABM_MASK_SH_LIST_DCN30(_MASK)
-};
-
-
-
-#define audio_regs(id)\
-[id] = {\
-		AUD_COMMON_REG_LIST(id)\
-}
-
-static const struct dce_audio_registers audio_regs[] = {
-	audio_regs(0),
-	audio_regs(1),
-	audio_regs(2),
-	audio_regs(3),
-	audio_regs(4),
-	audio_regs(5),
-	audio_regs(6)
-};
-
-#define DCE120_AUD_COMMON_MASK_SH_LIST(mask_sh)\
-		SF(AZF0ENDPOINT0_AZALIA_F0_CODEC_ENDPOINT_INDEX, AZALIA_ENDPOINT_REG_INDEX, mask_sh),\
-		SF(AZF0ENDPOINT0_AZALIA_F0_CODEC_ENDPOINT_DATA, AZALIA_ENDPOINT_REG_DATA, mask_sh),\
-		AUD_COMMON_MASK_SH_LIST_BASE(mask_sh)
-
-static const struct dce_audio_shift audio_shift = {
-		DCE120_AUD_COMMON_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dce_audio_mask audio_mask = {
-		DCE120_AUD_COMMON_MASK_SH_LIST(_MASK)
-};
-
-#define vpg_regs(id)\
-[id] = {\
-	VPG_DCN3_REG_LIST(id)\
-}
-
-static const struct dcn30_vpg_registers vpg_regs[] = {
-	vpg_regs(0),
-	vpg_regs(1),
-	vpg_regs(2),
-	vpg_regs(3),
-	vpg_regs(4),
-	vpg_regs(5),
-	vpg_regs(6),
-};
-
-static const struct dcn30_vpg_shift vpg_shift = {
-	DCN3_VPG_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dcn30_vpg_mask vpg_mask = {
-	DCN3_VPG_MASK_SH_LIST(_MASK)
-};
-
-#define afmt_regs(id)\
-[id] = {\
-	AFMT_DCN3_REG_LIST(id)\
-}
-
-static const struct dcn30_afmt_registers afmt_regs[] = {
-	afmt_regs(0),
-	afmt_regs(1),
-	afmt_regs(2),
-	afmt_regs(3),
-	afmt_regs(4),
-	afmt_regs(5),
-	afmt_regs(6),
-};
-
-static const struct dcn30_afmt_shift afmt_shift = {
-	DCN3_AFMT_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dcn30_afmt_mask afmt_mask = {
-	DCN3_AFMT_MASK_SH_LIST(_MASK)
-};
-
-#define stream_enc_regs(id)\
-[id] = {\
-	SE_DCN3_REG_LIST(id)\
-}
-
-static const struct dcn10_stream_enc_registers stream_enc_regs[] = {
-	stream_enc_regs(0),
-	stream_enc_regs(1),
-	stream_enc_regs(2),
-	stream_enc_regs(3),
-	stream_enc_regs(4),
-	stream_enc_regs(5)
-};
-
-static const struct dcn10_stream_encoder_shift se_shift = {
-		SE_COMMON_MASK_SH_LIST_DCN30(__SHIFT)
-};
-
-static const struct dcn10_stream_encoder_mask se_mask = {
-		SE_COMMON_MASK_SH_LIST_DCN30(_MASK)
-};
-
-
-#define aux_regs(id)\
-[id] = {\
-	DCN2_AUX_REG_LIST(id)\
-}
-
-static const struct dcn10_link_enc_aux_registers link_enc_aux_regs[] = {
-		aux_regs(0),
-		aux_regs(1),
-		aux_regs(2),
-		aux_regs(3),
-		aux_regs(4),
-		aux_regs(5)
-};
-
-#define hpd_regs(id)\
-[id] = {\
-	HPD_REG_LIST(id)\
-}
-
-static const struct dcn10_link_enc_hpd_registers link_enc_hpd_regs[] = {
-		hpd_regs(0),
-		hpd_regs(1),
-		hpd_regs(2),
-		hpd_regs(3),
-		hpd_regs(4),
-		hpd_regs(5)
-};
-
-#define link_regs(id, phyid)\
-[id] = {\
-	LE_DCN3_REG_LIST(id), \
-	UNIPHY_DCN2_REG_LIST(phyid), \
-	DPCS_DCN2_REG_LIST(id), \
-	SRI(DP_DPHY_INTERNAL_CTRL, DP, id) \
-}
-
-static const struct dce110_aux_registers_shift aux_shift = {
-	DCN_AUX_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dce110_aux_registers_mask aux_mask = {
-	DCN_AUX_MASK_SH_LIST(_MASK)
-};
-
-static const struct dcn10_link_enc_registers link_enc_regs[] = {
-	link_regs(0, A),
-	link_regs(1, B),
-	link_regs(2, C),
-	link_regs(3, D),
-	link_regs(4, E),
-	link_regs(5, F)
-};
-
-static const struct dcn10_link_enc_shift le_shift = {
-	LINK_ENCODER_MASK_SH_LIST_DCN30(__SHIFT),\
-	DPCS_DCN2_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dcn10_link_enc_mask le_mask = {
-	LINK_ENCODER_MASK_SH_LIST_DCN30(_MASK),\
-	DPCS_DCN2_MASK_SH_LIST(_MASK)
-};
-
-
-static const struct dce_panel_cntl_registers panel_cntl_regs[] = {
-	{ DCN_PANEL_CNTL_REG_LIST() }
-};
-
-static const struct dce_panel_cntl_shift panel_cntl_shift = {
-	DCE_PANEL_CNTL_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dce_panel_cntl_mask panel_cntl_mask = {
-	DCE_PANEL_CNTL_MASK_SH_LIST(_MASK)
-};
-
-#define dpp_regs(id)\
-[id] = {\
-	DPP_REG_LIST_DCN30(id),\
-}
-
-static const struct dcn3_dpp_registers dpp_regs[] = {
-	dpp_regs(0),
-	dpp_regs(1),
-	dpp_regs(2),
-	dpp_regs(3),
-	dpp_regs(4),
-	dpp_regs(5),
-};
-
-static const struct dcn3_dpp_shift tf_shift = {
-		DPP_REG_LIST_SH_MASK_DCN30(__SHIFT)
-};
-
-static const struct dcn3_dpp_mask tf_mask = {
-		DPP_REG_LIST_SH_MASK_DCN30(_MASK)
-};
-
-#define opp_regs(id)\
-[id] = {\
-	OPP_REG_LIST_DCN30(id),\
-}
-
-static const struct dcn20_opp_registers opp_regs[] = {
-	opp_regs(0),
-	opp_regs(1),
-	opp_regs(2),
-	opp_regs(3),
-	opp_regs(4),
-	opp_regs(5)
-};
-
-static const struct dcn20_opp_shift opp_shift = {
-	OPP_MASK_SH_LIST_DCN20(__SHIFT)
-};
-
-static const struct dcn20_opp_mask opp_mask = {
-	OPP_MASK_SH_LIST_DCN20(_MASK)
-};
-
-#define aux_engine_regs(id)\
-[id] = {\
-	AUX_COMMON_REG_LIST0(id), \
-	.AUXN_IMPCAL = 0, \
-	.AUXP_IMPCAL = 0, \
-	.AUX_RESET_MASK = DP_AUX0_AUX_CONTROL__AUX_RESET_MASK, \
-}
-
-static const struct dce110_aux_registers aux_engine_regs[] = {
-		aux_engine_regs(0),
-		aux_engine_regs(1),
-		aux_engine_regs(2),
-		aux_engine_regs(3),
-		aux_engine_regs(4),
-		aux_engine_regs(5)
-};
-
-#define dwbc_regs_dcn3(id)\
-[id] = {\
-	DWBC_COMMON_REG_LIST_DCN30(id),\
-}
-
-static const struct dcn30_dwbc_registers dwbc30_regs[] = {
-	dwbc_regs_dcn3(0),
-};
-
-static const struct dcn30_dwbc_shift dwbc30_shift = {
-	DWBC_COMMON_MASK_SH_LIST_DCN30(__SHIFT)
-};
-
-static const struct dcn30_dwbc_mask dwbc30_mask = {
-	DWBC_COMMON_MASK_SH_LIST_DCN30(_MASK)
-};
-
-#define mcif_wb_regs_dcn3(id)\
-[id] = {\
-	MCIF_WB_COMMON_REG_LIST_DCN30(id),\
-}
-
-static const struct dcn30_mmhubbub_registers mcif_wb30_regs[] = {
-	mcif_wb_regs_dcn3(0)
-};
-
-static const struct dcn30_mmhubbub_shift mcif_wb30_shift = {
-	MCIF_WB_COMMON_MASK_SH_LIST_DCN30(__SHIFT)
-};
-
-static const struct dcn30_mmhubbub_mask mcif_wb30_mask = {
-	MCIF_WB_COMMON_MASK_SH_LIST_DCN30(_MASK)
-};
-
-#define dsc_regsDCN20(id)\
-[id] = {\
-	DSC_REG_LIST_DCN20(id)\
-}
-
-static const struct dcn20_dsc_registers dsc_regs[] = {
-	dsc_regsDCN20(0),
-	dsc_regsDCN20(1),
-	dsc_regsDCN20(2),
-	dsc_regsDCN20(3),
-	dsc_regsDCN20(4),
-	dsc_regsDCN20(5)
-};
-
-static const struct dcn20_dsc_shift dsc_shift = {
-	DSC_REG_LIST_SH_MASK_DCN20(__SHIFT)
-};
-
-static const struct dcn20_dsc_mask dsc_mask = {
-	DSC_REG_LIST_SH_MASK_DCN20(_MASK)
-};
-
-static const struct dcn30_mpc_registers mpc_regs = {
-		MPC_REG_LIST_DCN3_0(0),
-		MPC_REG_LIST_DCN3_0(1),
-		MPC_REG_LIST_DCN3_0(2),
-		MPC_REG_LIST_DCN3_0(3),
-		MPC_REG_LIST_DCN3_0(4),
-		MPC_REG_LIST_DCN3_0(5),
-		MPC_OUT_MUX_REG_LIST_DCN3_0(0),
-		MPC_OUT_MUX_REG_LIST_DCN3_0(1),
-		MPC_OUT_MUX_REG_LIST_DCN3_0(2),
-		MPC_OUT_MUX_REG_LIST_DCN3_0(3),
-		MPC_OUT_MUX_REG_LIST_DCN3_0(4),
-		MPC_OUT_MUX_REG_LIST_DCN3_0(5),
-		MPC_RMU_GLOBAL_REG_LIST_DCN3AG,
-		MPC_RMU_REG_LIST_DCN3AG(0),
-		MPC_RMU_REG_LIST_DCN3AG(1),
-		MPC_RMU_REG_LIST_DCN3AG(2),
-		MPC_DWB_MUX_REG_LIST_DCN3_0(0),
-};
-
-static const struct dcn30_mpc_shift mpc_shift = {
-	MPC_COMMON_MASK_SH_LIST_DCN30(__SHIFT)
-};
-
-static const struct dcn30_mpc_mask mpc_mask = {
-	MPC_COMMON_MASK_SH_LIST_DCN30(_MASK)
-};
-
-#define optc_regs(id)\
-[id] = {OPTC_COMMON_REG_LIST_DCN3_0(id)}
-
-
-static const struct dcn_optc_registers optc_regs[] = {
-	optc_regs(0),
-	optc_regs(1),
-	optc_regs(2),
-	optc_regs(3),
-	optc_regs(4),
-	optc_regs(5)
-};
-
-static const struct dcn_optc_shift optc_shift = {
-	OPTC_COMMON_MASK_SH_LIST_DCN30(__SHIFT)
-};
-
-static const struct dcn_optc_mask optc_mask = {
-	OPTC_COMMON_MASK_SH_LIST_DCN30(_MASK)
-};
-
-#define hubp_regs(id)\
-[id] = {\
-	HUBP_REG_LIST_DCN30(id)\
-}
-
-static const struct dcn_hubp2_registers hubp_regs[] = {
-		hubp_regs(0),
-		hubp_regs(1),
-		hubp_regs(2),
-		hubp_regs(3),
-		hubp_regs(4),
-		hubp_regs(5)
-};
-
-static const struct dcn_hubp2_shift hubp_shift = {
-		HUBP_MASK_SH_LIST_DCN30(__SHIFT)
-};
-
-static const struct dcn_hubp2_mask hubp_mask = {
-		HUBP_MASK_SH_LIST_DCN30(_MASK)
-};
-
-static const struct dcn_hubbub_registers hubbub_reg = {
-		HUBBUB_REG_LIST_DCN30(0)
-};
-
-static const struct dcn_hubbub_shift hubbub_shift = {
-		HUBBUB_MASK_SH_LIST_DCN30(__SHIFT)
-};
-
-static const struct dcn_hubbub_mask hubbub_mask = {
-		HUBBUB_MASK_SH_LIST_DCN30(_MASK)
-};
-
-static const struct dccg_registers dccg_regs = {
-		DCCG_REG_LIST_DCN30()
-};
-
-static const struct dccg_shift dccg_shift = {
-		DCCG_MASK_SH_LIST_DCN3(__SHIFT)
-};
-
-static const struct dccg_mask dccg_mask = {
-		DCCG_MASK_SH_LIST_DCN3(_MASK)
-};
-
-static const struct dce_hwseq_registers hwseq_reg = {
-		HWSEQ_DCN30_REG_LIST()
-};
-
-static const struct dce_hwseq_shift hwseq_shift = {
-		HWSEQ_DCN30_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dce_hwseq_mask hwseq_mask = {
-		HWSEQ_DCN30_MASK_SH_LIST(_MASK)
-};
-#define vmid_regs(id)\
-[id] = {\
-		DCN20_VMID_REG_LIST(id)\
-}
-
-static const struct dcn_vmid_registers vmid_regs[] = {
-	vmid_regs(0),
-	vmid_regs(1),
-	vmid_regs(2),
-	vmid_regs(3),
-	vmid_regs(4),
-	vmid_regs(5),
-	vmid_regs(6),
-	vmid_regs(7),
-	vmid_regs(8),
-	vmid_regs(9),
-	vmid_regs(10),
-	vmid_regs(11),
-	vmid_regs(12),
-	vmid_regs(13),
-	vmid_regs(14),
-	vmid_regs(15)
-};
-
-static const struct dcn20_vmid_shift vmid_shifts = {
-		DCN20_VMID_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dcn20_vmid_mask vmid_masks = {
-		DCN20_VMID_MASK_SH_LIST(_MASK)
-};
-
-static const struct resource_caps res_cap_dcn3 = {
-	.num_timing_generator = 6,
-	.num_opp = 6,
-	.num_video_plane = 6,
-	.num_audio = 6,
-	.num_stream_encoder = 6,
-	.num_pll = 6,
-	.num_dwb = 1,
-	.num_ddc = 6,
-	.num_vmid = 16,
-	.num_mpc_3dlut = 3,
-	.num_dsc = 6,
-};
-
-static const struct dc_plane_cap plane_cap = {
-	.type = DC_PLANE_TYPE_DCN_UNIVERSAL,
-	.blends_with_above = true,
-	.blends_with_below = true,
-	.per_pixel_alpha = true,
-
-	.pixel_format_support = {
-			.argb8888 = true,
-			.nv12 = true,
-			.fp16 = true,
-			.p010 = false,
-			.ayuv = false,
-	},
-
-	.max_upscale_factor = {
-			.argb8888 = 16000,
-			.nv12 = 16000,
-			.fp16 = 16000
-	},
-
-	/* 6:1 downscaling ratio: 1000/6 = 166.666 */
-	.max_downscale_factor = {
-			.argb8888 = 167,
-			.nv12 = 167,
-			.fp16 = 167
-	}
-};
-
-static const struct dc_debug_options debug_defaults_drv = {
-	.disable_dmcu = true, //No DMCU on DCN30
-	.force_abm_enable = false,
-	.timing_trace = false,
-	.clock_trace = true,
-	.disable_pplib_clock_request = true,
-	.pipe_split_policy = MPC_SPLIT_AVOID_MULT_DISP,
-	.force_single_disp_pipe_split = false,
-	.disable_dcc = DCC_ENABLE,
-	.vsr_support = true,
-	.performance_trace = false,
-	.max_downscale_src_width = 7680,/*upto 8K*/
-	.disable_pplib_wm_range = false,
-	.scl_reset_length10 = true,
-	.sanity_checks = false,
-	.underflow_assert_delay_us = 0xFFFFFFFF,
-	.dwb_fi_phase = -1, // -1 = disable,
-	.dmub_command_table = true,
-	.disable_psr = false,
-	.use_max_lb = true
-};
-
-static const struct dc_debug_options debug_defaults_diags = {
-	.disable_dmcu = true, //No dmcu on DCN30
-	.force_abm_enable = false,
-	.timing_trace = true,
-	.clock_trace = true,
-	.disable_dpp_power_gate = true,
-	.disable_hubp_power_gate = true,
-	.disable_clock_gate = true,
-	.disable_pplib_clock_request = true,
-	.disable_pplib_wm_range = true,
-	.disable_stutter = false,
-	.scl_reset_length10 = true,
-	.dwb_fi_phase = -1, // -1 = disable
-	.dmub_command_table = true,
-	.disable_psr = true,
-	.enable_tri_buf = true,
-	.use_max_lb = true
-};
-
-void dcn30_dpp_destroy(struct dpp **dpp)
-{
-	kfree(TO_DCN20_DPP(*dpp));
-	*dpp = NULL;
-}
-
-static struct dpp *dcn30_dpp_create(
-	struct dc_context *ctx,
-	uint32_t inst)
-{
-	struct dcn3_dpp *dpp =
-		kzalloc(sizeof(struct dcn3_dpp), GFP_KERNEL);
-
-	if (!dpp)
-		return NULL;
-
-	if (dpp3_construct(dpp, ctx, inst,
-			&dpp_regs[inst], &tf_shift, &tf_mask))
-		return &dpp->base;
-
-	BREAK_TO_DEBUGGER();
-	kfree(dpp);
-	return NULL;
-}
-
-static struct output_pixel_processor *dcn30_opp_create(
-	struct dc_context *ctx, uint32_t inst)
-{
-	struct dcn20_opp *opp =
-		kzalloc(sizeof(struct dcn20_opp), GFP_KERNEL);
-
-	if (!opp) {
-		BREAK_TO_DEBUGGER();
-		return NULL;
-	}
-
-	dcn20_opp_construct(opp, ctx, inst,
-			&opp_regs[inst], &opp_shift, &opp_mask);
-	return &opp->base;
-}
-
-static struct dce_aux *dcn30_aux_engine_create(
-	struct dc_context *ctx,
-	uint32_t inst)
-{
-	struct aux_engine_dce110 *aux_engine =
-		kzalloc(sizeof(struct aux_engine_dce110), GFP_KERNEL);
-
-	if (!aux_engine)
-		return NULL;
-
-	dce110_aux_engine_construct(aux_engine, ctx, inst,
-				    SW_AUX_TIMEOUT_PERIOD_MULTIPLIER * AUX_TIMEOUT_PERIOD,
-				    &aux_engine_regs[inst],
-					&aux_mask,
-					&aux_shift,
-					ctx->dc->caps.extended_aux_timeout_support);
-
-	return &aux_engine->base;
-}
-
-#define i2c_inst_regs(id) { I2C_HW_ENGINE_COMMON_REG_LIST_DCN30(id) }
-
-static const struct dce_i2c_registers i2c_hw_regs[] = {
-		i2c_inst_regs(1),
-		i2c_inst_regs(2),
-		i2c_inst_regs(3),
-		i2c_inst_regs(4),
-		i2c_inst_regs(5),
-		i2c_inst_regs(6),
-};
-
-static const struct dce_i2c_shift i2c_shifts = {
-		I2C_COMMON_MASK_SH_LIST_DCN30(__SHIFT)
-};
-
-static const struct dce_i2c_mask i2c_masks = {
-		I2C_COMMON_MASK_SH_LIST_DCN30(_MASK)
-};
-
-static struct dce_i2c_hw *dcn30_i2c_hw_create(
-	struct dc_context *ctx,
-	uint32_t inst)
-{
-	struct dce_i2c_hw *dce_i2c_hw =
-		kzalloc(sizeof(struct dce_i2c_hw), GFP_KERNEL);
-
-	if (!dce_i2c_hw)
-		return NULL;
-
-	dcn2_i2c_hw_construct(dce_i2c_hw, ctx, inst,
-				    &i2c_hw_regs[inst], &i2c_shifts, &i2c_masks);
-
-	return dce_i2c_hw;
-}
-
-static struct mpc *dcn30_mpc_create(
-		struct dc_context *ctx,
-		int num_mpcc,
-		int num_rmu)
-{
-	struct dcn30_mpc *mpc30 = kzalloc(sizeof(struct dcn30_mpc),
-					  GFP_KERNEL);
-
-	if (!mpc30)
-		return NULL;
-
-	dcn30_mpc_construct(mpc30, ctx,
-			&mpc_regs,
-			&mpc_shift,
-			&mpc_mask,
-			num_mpcc,
-			num_rmu);
-
-	return &mpc30->base;
-}
-
-struct hubbub *dcn30_hubbub_create(struct dc_context *ctx)
-{
-	int i;
-
-	struct dcn20_hubbub *hubbub3 = kzalloc(sizeof(struct dcn20_hubbub),
-					  GFP_KERNEL);
-
-	if (!hubbub3)
-		return NULL;
-
-	hubbub3_construct(hubbub3, ctx,
-			&hubbub_reg,
-			&hubbub_shift,
-			&hubbub_mask);
-
-
-	for (i = 0; i < res_cap_dcn3.num_vmid; i++) {
-		struct dcn20_vmid *vmid = &hubbub3->vmid[i];
-
-		vmid->ctx = ctx;
-
-		vmid->regs = &vmid_regs[i];
-		vmid->shifts = &vmid_shifts;
-		vmid->masks = &vmid_masks;
-	}
-
-	return &hubbub3->base;
-}
-
-static struct timing_generator *dcn30_timing_generator_create(
-		struct dc_context *ctx,
-		uint32_t instance)
-{
-	struct optc *tgn10 =
-		kzalloc(sizeof(struct optc), GFP_KERNEL);
-
-	if (!tgn10)
-		return NULL;
-
-	tgn10->base.inst = instance;
-	tgn10->base.ctx = ctx;
-
-	tgn10->tg_regs = &optc_regs[instance];
-	tgn10->tg_shift = &optc_shift;
-	tgn10->tg_mask = &optc_mask;
-
-	dcn30_timing_generator_init(tgn10);
-
-	return &tgn10->base;
-}
-
-static const struct encoder_feature_support link_enc_feature = {
-		.max_hdmi_deep_color = COLOR_DEPTH_121212,
-		.max_hdmi_pixel_clock = 600000,
-		.hdmi_ycbcr420_supported = true,
-		.dp_ycbcr420_supported = true,
-		.fec_supported = true,
-		.flags.bits.IS_HBR2_CAPABLE = true,
-		.flags.bits.IS_HBR3_CAPABLE = true,
-		.flags.bits.IS_TPS3_CAPABLE = true,
-		.flags.bits.IS_TPS4_CAPABLE = true
-};
-
-static struct link_encoder *dcn30_link_encoder_create(
-	const struct encoder_init_data *enc_init_data)
-{
-	struct dcn20_link_encoder *enc20 =
-		kzalloc(sizeof(struct dcn20_link_encoder), GFP_KERNEL);
-
-	if (!enc20)
-		return NULL;
-
-	dcn30_link_encoder_construct(enc20,
-			enc_init_data,
-			&link_enc_feature,
-			&link_enc_regs[enc_init_data->transmitter],
-			&link_enc_aux_regs[enc_init_data->channel - 1],
-			&link_enc_hpd_regs[enc_init_data->hpd_source],
-			&le_shift,
-			&le_mask);
-
-	return &enc20->enc10.base;
-}
-
-static struct panel_cntl *dcn30_panel_cntl_create(const struct panel_cntl_init_data *init_data)
-{
-	struct dce_panel_cntl *panel_cntl =
-		kzalloc(sizeof(struct dce_panel_cntl), GFP_KERNEL);
-
-	if (!panel_cntl)
-		return NULL;
-
-	dce_panel_cntl_construct(panel_cntl,
-			init_data,
-			&panel_cntl_regs[init_data->inst],
-			&panel_cntl_shift,
-			&panel_cntl_mask);
-
-	return &panel_cntl->base;
-}
-
-static void read_dce_straps(
-	struct dc_context *ctx,
-	struct resource_straps *straps)
-{
-	generic_reg_get(ctx, mmDC_PINSTRAPS + BASE(mmDC_PINSTRAPS_BASE_IDX),
-		FN(DC_PINSTRAPS, DC_PINSTRAPS_AUDIO), &straps->dc_pinstraps_audio);
-
-}
-
-static struct audio *dcn30_create_audio(
-		struct dc_context *ctx, unsigned int inst)
-{
-	return dce_audio_create(ctx, inst,
-			&audio_regs[inst], &audio_shift, &audio_mask);
-}
-
-static struct vpg *dcn30_vpg_create(
-	struct dc_context *ctx,
-	uint32_t inst)
-{
-	struct dcn30_vpg *vpg3 = kzalloc(sizeof(struct dcn30_vpg), GFP_KERNEL);
-
-	if (!vpg3)
-		return NULL;
-
-	vpg3_construct(vpg3, ctx, inst,
-			&vpg_regs[inst],
-			&vpg_shift,
-			&vpg_mask);
-
-	return &vpg3->base;
-}
-
-static struct afmt *dcn30_afmt_create(
-	struct dc_context *ctx,
-	uint32_t inst)
-{
-	struct dcn30_afmt *afmt3 = kzalloc(sizeof(struct dcn30_afmt), GFP_KERNEL);
-
-	if (!afmt3)
-		return NULL;
-
-	afmt3_construct(afmt3, ctx, inst,
-			&afmt_regs[inst],
-			&afmt_shift,
-			&afmt_mask);
-
-	return &afmt3->base;
-}
-
-struct stream_encoder *dcn30_stream_encoder_create(
-	enum engine_id eng_id,
-	struct dc_context *ctx)
-{
-	struct dcn10_stream_encoder *enc1;
-	struct vpg *vpg;
-	struct afmt *afmt;
-	int vpg_inst;
-	int afmt_inst;
-
-	/* Mapping of VPG, AFMT, DME register blocks to DIO block instance */
-	if (eng_id <= ENGINE_ID_DIGF) {
-		vpg_inst = eng_id;
-		afmt_inst = eng_id;
-	} else
-		return NULL;
-
-	enc1 = kzalloc(sizeof(struct dcn10_stream_encoder), GFP_KERNEL);
-	vpg = dcn30_vpg_create(ctx, vpg_inst);
-	afmt = dcn30_afmt_create(ctx, afmt_inst);
-
-	if (!enc1 || !vpg || !afmt) {
-		kfree(enc1);
-		kfree(vpg);
-		kfree(afmt);
-		return NULL;
-	}
-
-	dcn30_dio_stream_encoder_construct(enc1, ctx, ctx->dc_bios,
-					eng_id, vpg, afmt,
-					&stream_enc_regs[eng_id],
-					&se_shift, &se_mask);
-
-	return &enc1->base;
-}
-
-struct dce_hwseq *dcn30_hwseq_create(
-	struct dc_context *ctx)
-{
-	struct dce_hwseq *hws = kzalloc(sizeof(struct dce_hwseq), GFP_KERNEL);
-
-	if (hws) {
-		hws->ctx = ctx;
-		hws->regs = &hwseq_reg;
-		hws->shifts = &hwseq_shift;
-		hws->masks = &hwseq_mask;
-	}
-	return hws;
-}
-static const struct resource_create_funcs res_create_funcs = {
-	.read_dce_straps = read_dce_straps,
-	.create_audio = dcn30_create_audio,
-	.create_stream_encoder = dcn30_stream_encoder_create,
-	.create_hwseq = dcn30_hwseq_create,
-};
-
-static const struct resource_create_funcs res_create_maximus_funcs = {
-	.read_dce_straps = NULL,
-	.create_audio = NULL,
-	.create_stream_encoder = NULL,
-	.create_hwseq = dcn30_hwseq_create,
-};
-
-static void dcn30_resource_destruct(struct dcn30_resource_pool *pool)
-{
-	unsigned int i;
-
-	for (i = 0; i < pool->base.stream_enc_count; i++) {
-		if (pool->base.stream_enc[i] != NULL) {
-			if (pool->base.stream_enc[i]->vpg != NULL) {
-				kfree(DCN30_VPG_FROM_VPG(pool->base.stream_enc[i]->vpg));
-				pool->base.stream_enc[i]->vpg = NULL;
-			}
-			if (pool->base.stream_enc[i]->afmt != NULL) {
-				kfree(DCN30_AFMT_FROM_AFMT(pool->base.stream_enc[i]->afmt));
-				pool->base.stream_enc[i]->afmt = NULL;
-			}
-			kfree(DCN10STRENC_FROM_STRENC(pool->base.stream_enc[i]));
-			pool->base.stream_enc[i] = NULL;
-		}
-	}
-
-	for (i = 0; i < pool->base.res_cap->num_dsc; i++) {
-		if (pool->base.dscs[i] != NULL)
-			dcn20_dsc_destroy(&pool->base.dscs[i]);
-	}
-
-	if (pool->base.mpc != NULL) {
-		kfree(TO_DCN20_MPC(pool->base.mpc));
-		pool->base.mpc = NULL;
-	}
-	if (pool->base.hubbub != NULL) {
-		kfree(pool->base.hubbub);
-		pool->base.hubbub = NULL;
-	}
-	for (i = 0; i < pool->base.pipe_count; i++) {
-		if (pool->base.dpps[i] != NULL)
-			dcn30_dpp_destroy(&pool->base.dpps[i]);
-
-		if (pool->base.ipps[i] != NULL)
-			pool->base.ipps[i]->funcs->ipp_destroy(&pool->base.ipps[i]);
-
-		if (pool->base.hubps[i] != NULL) {
-			kfree(TO_DCN20_HUBP(pool->base.hubps[i]));
-			pool->base.hubps[i] = NULL;
-		}
-
-		if (pool->base.irqs != NULL) {
-			dal_irq_service_destroy(&pool->base.irqs);
-		}
-	}
-
-	for (i = 0; i < pool->base.res_cap->num_ddc; i++) {
-		if (pool->base.engines[i] != NULL)
-			dce110_engine_destroy(&pool->base.engines[i]);
-		if (pool->base.hw_i2cs[i] != NULL) {
-			kfree(pool->base.hw_i2cs[i]);
-			pool->base.hw_i2cs[i] = NULL;
-		}
-		if (pool->base.sw_i2cs[i] != NULL) {
-			kfree(pool->base.sw_i2cs[i]);
-			pool->base.sw_i2cs[i] = NULL;
-		}
-	}
-
-	for (i = 0; i < pool->base.res_cap->num_opp; i++) {
-		if (pool->base.opps[i] != NULL)
-			pool->base.opps[i]->funcs->opp_destroy(&pool->base.opps[i]);
-	}
-
-	for (i = 0; i < pool->base.res_cap->num_timing_generator; i++) {
-		if (pool->base.timing_generators[i] != NULL)	{
-			kfree(DCN10TG_FROM_TG(pool->base.timing_generators[i]));
-			pool->base.timing_generators[i] = NULL;
-		}
-	}
-
-	for (i = 0; i < pool->base.res_cap->num_dwb; i++) {
-		if (pool->base.dwbc[i] != NULL) {
-			kfree(TO_DCN30_DWBC(pool->base.dwbc[i]));
-			pool->base.dwbc[i] = NULL;
-		}
-		if (pool->base.mcif_wb[i] != NULL) {
-			kfree(TO_DCN30_MMHUBBUB(pool->base.mcif_wb[i]));
-			pool->base.mcif_wb[i] = NULL;
-		}
-	}
-
-	for (i = 0; i < pool->base.audio_count; i++) {
-		if (pool->base.audios[i])
-			dce_aud_destroy(&pool->base.audios[i]);
-	}
-
-	for (i = 0; i < pool->base.clk_src_count; i++) {
-		if (pool->base.clock_sources[i] != NULL) {
-			dcn20_clock_source_destroy(&pool->base.clock_sources[i]);
-			pool->base.clock_sources[i] = NULL;
-		}
-	}
-
-	for (i = 0; i < pool->base.res_cap->num_mpc_3dlut; i++) {
-		if (pool->base.mpc_lut[i] != NULL) {
-			dc_3dlut_func_release(pool->base.mpc_lut[i]);
-			pool->base.mpc_lut[i] = NULL;
-		}
-		if (pool->base.mpc_shaper[i] != NULL) {
-			dc_transfer_func_release(pool->base.mpc_shaper[i]);
-			pool->base.mpc_shaper[i] = NULL;
-		}
-	}
-
-	if (pool->base.dp_clock_source != NULL) {
-		dcn20_clock_source_destroy(&pool->base.dp_clock_source);
-		pool->base.dp_clock_source = NULL;
-	}
-
-	for (i = 0; i < pool->base.pipe_count; i++) {
-		if (pool->base.multiple_abms[i] != NULL)
-			dce_abm_destroy(&pool->base.multiple_abms[i]);
-	}
-
-	if (pool->base.psr != NULL)
-		dmub_psr_destroy(&pool->base.psr);
-
-	if (pool->base.dccg != NULL)
-		dcn_dccg_destroy(&pool->base.dccg);
-
-	if (pool->base.oem_device != NULL)
-		dal_ddc_service_destroy(&pool->base.oem_device);
-}
-
-static struct hubp *dcn30_hubp_create(
-	struct dc_context *ctx,
-	uint32_t inst)
-{
-	struct dcn20_hubp *hubp2 =
-		kzalloc(sizeof(struct dcn20_hubp), GFP_KERNEL);
-
-	if (!hubp2)
-		return NULL;
-
-	if (hubp3_construct(hubp2, ctx, inst,
-			&hubp_regs[inst], &hubp_shift, &hubp_mask))
-		return &hubp2->base;
-
-	BREAK_TO_DEBUGGER();
-	kfree(hubp2);
-	return NULL;
-}
-
-static bool dcn30_dwbc_create(struct dc_context *ctx, struct resource_pool *pool)
-{
-	int i;
-	uint32_t pipe_count = pool->res_cap->num_dwb;
-
-	for (i = 0; i < pipe_count; i++) {
-		struct dcn30_dwbc *dwbc30 = kzalloc(sizeof(struct dcn30_dwbc),
-						    GFP_KERNEL);
-
-		if (!dwbc30) {
-			dm_error("DC: failed to create dwbc30!\n");
-			return false;
-		}
-
-		dcn30_dwbc_construct(dwbc30, ctx,
-				&dwbc30_regs[i],
-				&dwbc30_shift,
-				&dwbc30_mask,
-				i);
-
-		pool->dwbc[i] = &dwbc30->base;
-	}
-	return true;
-}
-
-static bool dcn30_mmhubbub_create(struct dc_context *ctx, struct resource_pool *pool)
-{
-	int i;
-	uint32_t pipe_count = pool->res_cap->num_dwb;
-
-	for (i = 0; i < pipe_count; i++) {
-		struct dcn30_mmhubbub *mcif_wb30 = kzalloc(sizeof(struct dcn30_mmhubbub),
-						    GFP_KERNEL);
-
-		if (!mcif_wb30) {
-			dm_error("DC: failed to create mcif_wb30!\n");
-			return false;
-		}
-
-		dcn30_mmhubbub_construct(mcif_wb30, ctx,
-				&mcif_wb30_regs[i],
-				&mcif_wb30_shift,
-				&mcif_wb30_mask,
-				i);
-
-		pool->mcif_wb[i] = &mcif_wb30->base;
-	}
-	return true;
-}
-
-static struct display_stream_compressor *dcn30_dsc_create(
-	struct dc_context *ctx, uint32_t inst)
-{
-	struct dcn20_dsc *dsc =
-		kzalloc(sizeof(struct dcn20_dsc), GFP_KERNEL);
-
-	if (!dsc) {
-		BREAK_TO_DEBUGGER();
-		return NULL;
-	}
-
-	dsc2_construct(dsc, ctx, inst, &dsc_regs[inst], &dsc_shift, &dsc_mask);
-	return &dsc->base;
-}
-
-enum dc_status dcn30_add_stream_to_ctx(struct dc *dc, struct dc_state *new_ctx, struct dc_stream_state *dc_stream)
-{
-
-	return dcn20_add_stream_to_ctx(dc, new_ctx, dc_stream);
-}
-
-static void dcn30_destroy_resource_pool(struct resource_pool **pool)
-{
-	struct dcn30_resource_pool *dcn30_pool = TO_DCN30_RES_POOL(*pool);
-
-	dcn30_resource_destruct(dcn30_pool);
-	kfree(dcn30_pool);
-	*pool = NULL;
-}
-
-static struct clock_source *dcn30_clock_source_create(
-		struct dc_context *ctx,
-		struct dc_bios *bios,
-		enum clock_source_id id,
-		const struct dce110_clk_src_regs *regs,
-		bool dp_clk_src)
-{
-	struct dce110_clk_src *clk_src =
-		kzalloc(sizeof(struct dce110_clk_src), GFP_KERNEL);
-
-	if (!clk_src)
-		return NULL;
-
-	if (dcn3_clk_src_construct(clk_src, ctx, bios, id,
-			regs, &cs_shift, &cs_mask)) {
-		clk_src->base.dp_clk_src = dp_clk_src;
-		return &clk_src->base;
-	}
-
-	BREAK_TO_DEBUGGER();
-	return NULL;
-}
-
-int dcn30_populate_dml_pipes_from_context(
-	struct dc *dc, struct dc_state *context,
-	display_e2e_pipe_params_st *pipes,
-	bool fast_validate)
-{
-	int i, pipe_cnt;
-	struct resource_context *res_ctx = &context->res_ctx;
-
-	dcn20_populate_dml_pipes_from_context(dc, context, pipes, fast_validate);
-
-	for (i = 0, pipe_cnt = 0; i < dc->res_pool->pipe_count; i++) {
-		if (!res_ctx->pipe_ctx[i].stream)
-			continue;
-
-		pipes[pipe_cnt++].pipe.scale_ratio_depth.lb_depth =
-			dm_lb_16;
-	}
-
-	return pipe_cnt;
-}
-
-void dcn30_populate_dml_writeback_from_context(
-	struct dc *dc, struct resource_context *res_ctx, display_e2e_pipe_params_st *pipes)
-{
-	int pipe_cnt, i, j;
-	double max_calc_writeback_dispclk;
-	double writeback_dispclk;
-	struct writeback_st dout_wb;
-
-	for (i = 0, pipe_cnt = 0; i < dc->res_pool->pipe_count; i++) {
-		struct dc_stream_state *stream = res_ctx->pipe_ctx[i].stream;
-
-		if (!stream)
-			continue;
-		max_calc_writeback_dispclk = 0;
-
-		/* Set writeback information */
-		pipes[pipe_cnt].dout.wb_enable = 0;
-		pipes[pipe_cnt].dout.num_active_wb = 0;
-		for (j = 0; j < stream->num_wb_info; j++) {
-			struct dc_writeback_info *wb_info = &stream->writeback_info[j];
-
-			if (wb_info->wb_enabled && wb_info->writeback_source_plane &&
-					(wb_info->writeback_source_plane == res_ctx->pipe_ctx[i].plane_state)) {
-				pipes[pipe_cnt].dout.wb_enable = 1;
-				pipes[pipe_cnt].dout.num_active_wb++;
-				dout_wb.wb_src_height = wb_info->dwb_params.cnv_params.crop_en ?
-					wb_info->dwb_params.cnv_params.crop_height :
-					wb_info->dwb_params.cnv_params.src_height;
-				dout_wb.wb_src_width = wb_info->dwb_params.cnv_params.crop_en ?
-					wb_info->dwb_params.cnv_params.crop_width :
-					wb_info->dwb_params.cnv_params.src_width;
-				dout_wb.wb_dst_width = wb_info->dwb_params.dest_width;
-				dout_wb.wb_dst_height = wb_info->dwb_params.dest_height;
-
-				/* For IP that doesn't support WB scaling, set h/v taps to 1 to avoid DML validation failure */
-				if (dc->dml.ip.writeback_max_hscl_taps > 1) {
-					dout_wb.wb_htaps_luma = wb_info->dwb_params.scaler_taps.h_taps;
-					dout_wb.wb_vtaps_luma = wb_info->dwb_params.scaler_taps.v_taps;
-				} else {
-					dout_wb.wb_htaps_luma = 1;
-					dout_wb.wb_vtaps_luma = 1;
-				}
-				dout_wb.wb_htaps_chroma = 0;
-				dout_wb.wb_vtaps_chroma = 0;
-				dout_wb.wb_hratio = wb_info->dwb_params.cnv_params.crop_en ?
-					(double)wb_info->dwb_params.cnv_params.crop_width /
-						(double)wb_info->dwb_params.dest_width :
-					(double)wb_info->dwb_params.cnv_params.src_width /
-						(double)wb_info->dwb_params.dest_width;
-				dout_wb.wb_vratio = wb_info->dwb_params.cnv_params.crop_en ?
-					(double)wb_info->dwb_params.cnv_params.crop_height /
-						(double)wb_info->dwb_params.dest_height :
-					(double)wb_info->dwb_params.cnv_params.src_height /
-						(double)wb_info->dwb_params.dest_height;
-				if (wb_info->dwb_params.cnv_params.fc_out_format == DWB_OUT_FORMAT_64BPP_ARGB ||
-					wb_info->dwb_params.cnv_params.fc_out_format == DWB_OUT_FORMAT_64BPP_RGBA)
-					dout_wb.wb_pixel_format = dm_444_64;
-				else
-					dout_wb.wb_pixel_format = dm_444_32;
-
-				/* Workaround for cases where multiple writebacks are connected to same plane
-				 * In which case, need to compute worst case and set the associated writeback parameters
-				 * This workaround is necessary due to DML computation assuming only 1 set of writeback
-				 * parameters per pipe
-				 */
-				writeback_dispclk = dml30_CalculateWriteBackDISPCLK(
-						dout_wb.wb_pixel_format,
-						pipes[pipe_cnt].pipe.dest.pixel_rate_mhz,
-						dout_wb.wb_hratio,
-						dout_wb.wb_vratio,
-						dout_wb.wb_htaps_luma,
-						dout_wb.wb_vtaps_luma,
-						dout_wb.wb_src_width,
-						dout_wb.wb_dst_width,
-						pipes[pipe_cnt].pipe.dest.htotal,
-						dc->current_state->bw_ctx.dml.ip.writeback_line_buffer_buffer_size);
-
-				if (writeback_dispclk > max_calc_writeback_dispclk) {
-					max_calc_writeback_dispclk = writeback_dispclk;
-					pipes[pipe_cnt].dout.wb = dout_wb;
-				}
-			}
-		}
-
-		pipe_cnt++;
-	}
-
-}
-
-unsigned int dcn30_calc_max_scaled_time(
-		unsigned int time_per_pixel,
-		enum mmhubbub_wbif_mode mode,
-		unsigned int urgent_watermark)
-{
-	unsigned int time_per_byte = 0;
-	unsigned int total_free_entry = 0xb40;
-	unsigned int buf_lh_capability;
-	unsigned int max_scaled_time;
-
-	if (mode == PACKED_444) /* packed mode 32 bpp */
-		time_per_byte = time_per_pixel/4;
-	else if (mode == PACKED_444_FP16) /* packed mode 64 bpp */
-		time_per_byte = time_per_pixel/8;
-
-	if (time_per_byte == 0)
-		time_per_byte = 1;
-
-	buf_lh_capability = (total_free_entry*time_per_byte*32) >> 6; /* time_per_byte is in u6.6*/
-	max_scaled_time   = buf_lh_capability - urgent_watermark;
-	return max_scaled_time;
-}
-
-void dcn30_set_mcif_arb_params(
-		struct dc *dc,
-		struct dc_state *context,
-		display_e2e_pipe_params_st *pipes,
-		int pipe_cnt)
-{
-	enum mmhubbub_wbif_mode wbif_mode;
-	struct display_mode_lib *dml = &context->bw_ctx.dml;
-	struct mcif_arb_params *wb_arb_params;
-	int i, j, k, dwb_pipe;
-
-	/* Writeback MCIF_WB arbitration parameters */
-	dwb_pipe = 0;
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-
-		if (!context->res_ctx.pipe_ctx[i].stream)
-			continue;
-
-		for (j = 0; j < MAX_DWB_PIPES; j++) {
-			struct dc_writeback_info *writeback_info = &context->res_ctx.pipe_ctx[i].stream->writeback_info[j];
-
-			if (writeback_info->wb_enabled == false)
-				continue;
-
-			//wb_arb_params = &context->res_ctx.pipe_ctx[i].stream->writeback_info[j].mcif_arb_params;
-			wb_arb_params = &context->bw_ctx.bw.dcn.bw_writeback.mcif_wb_arb[dwb_pipe];
-
-			if (writeback_info->dwb_params.cnv_params.fc_out_format == DWB_OUT_FORMAT_64BPP_ARGB ||
-				writeback_info->dwb_params.cnv_params.fc_out_format == DWB_OUT_FORMAT_64BPP_RGBA)
-				wbif_mode = PACKED_444_FP16;
-			else
-				wbif_mode = PACKED_444;
-
-			for (k = 0; k < sizeof(wb_arb_params->cli_watermark)/sizeof(wb_arb_params->cli_watermark[0]); k++) {
-				wb_arb_params->cli_watermark[k] = get_wm_writeback_urgent(dml, pipes, pipe_cnt) * 1000;
-				wb_arb_params->pstate_watermark[k] = get_wm_writeback_dram_clock_change(dml, pipes, pipe_cnt) * 1000;
-			}
-			wb_arb_params->time_per_pixel = (1000000 << 6) / context->res_ctx.pipe_ctx[i].stream->phy_pix_clk; /* time_per_pixel should be in u6.6 format */
-			wb_arb_params->slice_lines = 32;
-			wb_arb_params->arbitration_slice = 2; /* irrelevant since there is no YUV output */
-			wb_arb_params->max_scaled_time = dcn30_calc_max_scaled_time(wb_arb_params->time_per_pixel,
-					wbif_mode,
-					wb_arb_params->cli_watermark[0]); /* assume 4 watermark sets have the same value */
-			wb_arb_params->dram_speed_change_duration = dml->vba.WritebackAllowDRAMClockChangeEndPosition[j] * pipes[0].clks_cfg.refclk_mhz; /* num_clock_cycles = us * MHz */
-
-			dwb_pipe++;
-
-			if (dwb_pipe >= MAX_DWB_PIPES)
-				return;
-		}
-		if (dwb_pipe >= MAX_DWB_PIPES)
-			return;
-	}
-
-}
-
-static struct dc_cap_funcs cap_funcs = {
-	.get_dcc_compression_cap = dcn20_get_dcc_compression_cap
-};
-
-bool dcn30_acquire_post_bldn_3dlut(
-		struct resource_context *res_ctx,
-		const struct resource_pool *pool,
-		int mpcc_id,
-		struct dc_3dlut **lut,
-		struct dc_transfer_func **shaper)
-{
-	int i;
-	bool ret = false;
-	union dc_3dlut_state *state;
-
-	ASSERT(*lut == NULL && *shaper == NULL);
-	*lut = NULL;
-	*shaper = NULL;
-
-	for (i = 0; i < pool->res_cap->num_mpc_3dlut; i++) {
-		if (!res_ctx->is_mpc_3dlut_acquired[i]) {
-			*lut = pool->mpc_lut[i];
-			*shaper = pool->mpc_shaper[i];
-			state = &pool->mpc_lut[i]->state;
-			res_ctx->is_mpc_3dlut_acquired[i] = true;
-			state->bits.rmu_idx_valid = 1;
-			state->bits.rmu_mux_num = i;
-			if (state->bits.rmu_mux_num == 0)
-				state->bits.mpc_rmu0_mux = mpcc_id;
-			else if (state->bits.rmu_mux_num == 1)
-				state->bits.mpc_rmu1_mux = mpcc_id;
-			else if (state->bits.rmu_mux_num == 2)
-				state->bits.mpc_rmu2_mux = mpcc_id;
-			ret = true;
-			break;
-			}
-		}
-	return ret;
-}
-
-bool dcn30_release_post_bldn_3dlut(
-		struct resource_context *res_ctx,
-		const struct resource_pool *pool,
-		struct dc_3dlut **lut,
-		struct dc_transfer_func **shaper)
-{
-	int i;
-	bool ret = false;
-
-	for (i = 0; i < pool->res_cap->num_mpc_3dlut; i++) {
-		if (pool->mpc_lut[i] == *lut && pool->mpc_shaper[i] == *shaper) {
-			res_ctx->is_mpc_3dlut_acquired[i] = false;
-			pool->mpc_lut[i]->state.raw = 0;
-			*lut = NULL;
-			*shaper = NULL;
-			ret = true;
-			break;
-		}
-	}
-	return ret;
-}
-
-static bool is_soc_bounding_box_valid(struct dc *dc)
-{
-	uint32_t hw_internal_rev = dc->ctx->asic_id.hw_internal_rev;
-
-	if (ASICREV_IS_SIENNA_CICHLID_P(hw_internal_rev))
-		return true;
-
-	return false;
-}
-
-static bool init_soc_bounding_box(struct dc *dc,
-				  struct dcn30_resource_pool *pool)
-{
-	struct _vcs_dpi_soc_bounding_box_st *loaded_bb = &dcn3_0_soc;
-	struct _vcs_dpi_ip_params_st *loaded_ip = &dcn3_0_ip;
-
-	DC_LOGGER_INIT(dc->ctx->logger);
-
-	if (!is_soc_bounding_box_valid(dc)) {
-		DC_LOG_ERROR("%s: not valid soc bounding box\n", __func__);
-		return false;
-	}
-
-	loaded_ip->max_num_otg = pool->base.res_cap->num_timing_generator;
-	loaded_ip->max_num_dpp = pool->base.pipe_count;
-	loaded_ip->clamp_min_dcfclk = dc->config.clamp_min_dcfclk;
-	dcn20_patch_bounding_box(dc, loaded_bb);
-
-	if (dc->ctx->dc_bios->funcs->get_soc_bb_info) {
-		struct bp_soc_bb_info bb_info = {0};
-
-		if (dc->ctx->dc_bios->funcs->get_soc_bb_info(dc->ctx->dc_bios, &bb_info) == BP_RESULT_OK) {
-			if (bb_info.dram_clock_change_latency_100ns > 0)
-				dcn3_0_soc.dram_clock_change_latency_us = bb_info.dram_clock_change_latency_100ns * 10;
-
-			if (bb_info.dram_sr_enter_exit_latency_100ns > 0)
-				dcn3_0_soc.sr_enter_plus_exit_time_us = bb_info.dram_sr_enter_exit_latency_100ns * 10;
-
-			if (bb_info.dram_sr_exit_latency_100ns > 0)
-				dcn3_0_soc.sr_exit_time_us = bb_info.dram_sr_exit_latency_100ns * 10;
-		}
-	}
-
-	return true;
-}
-
-static bool dcn30_split_stream_for_mpc_or_odm(
-		const struct dc *dc,
-		struct resource_context *res_ctx,
-		struct pipe_ctx *pri_pipe,
-		struct pipe_ctx *sec_pipe,
-		bool odm)
-{
-	int pipe_idx = sec_pipe->pipe_idx;
-	const struct resource_pool *pool = dc->res_pool;
-
-	*sec_pipe = *pri_pipe;
-
-	sec_pipe->pipe_idx = pipe_idx;
-	sec_pipe->plane_res.mi = pool->mis[pipe_idx];
-	sec_pipe->plane_res.hubp = pool->hubps[pipe_idx];
-	sec_pipe->plane_res.ipp = pool->ipps[pipe_idx];
-	sec_pipe->plane_res.xfm = pool->transforms[pipe_idx];
-	sec_pipe->plane_res.dpp = pool->dpps[pipe_idx];
-	sec_pipe->plane_res.mpcc_inst = pool->dpps[pipe_idx]->inst;
-	sec_pipe->stream_res.dsc = NULL;
-	if (odm) {
-		if (pri_pipe->next_odm_pipe) {
-			ASSERT(pri_pipe->next_odm_pipe != sec_pipe);
-			sec_pipe->next_odm_pipe = pri_pipe->next_odm_pipe;
-			sec_pipe->next_odm_pipe->prev_odm_pipe = sec_pipe;
-		}
-		if (pri_pipe->top_pipe && pri_pipe->top_pipe->next_odm_pipe) {
-			pri_pipe->top_pipe->next_odm_pipe->bottom_pipe = sec_pipe;
-			sec_pipe->top_pipe = pri_pipe->top_pipe->next_odm_pipe;
-		}
-		if (pri_pipe->bottom_pipe && pri_pipe->bottom_pipe->next_odm_pipe) {
-			pri_pipe->bottom_pipe->next_odm_pipe->top_pipe = sec_pipe;
-			sec_pipe->bottom_pipe = pri_pipe->bottom_pipe->next_odm_pipe;
-		}
-		pri_pipe->next_odm_pipe = sec_pipe;
-		sec_pipe->prev_odm_pipe = pri_pipe;
-
-		if (!sec_pipe->top_pipe)
-			sec_pipe->stream_res.opp = pool->opps[pipe_idx];
-		else
-			sec_pipe->stream_res.opp = sec_pipe->top_pipe->stream_res.opp;
-		if (sec_pipe->stream->timing.flags.DSC == 1) {
-			dcn20_acquire_dsc(dc, res_ctx, &sec_pipe->stream_res.dsc, pipe_idx);
-			ASSERT(sec_pipe->stream_res.dsc);
-			if (sec_pipe->stream_res.dsc == NULL)
-				return false;
-		}
-	} else {
-		if (pri_pipe->bottom_pipe) {
-			ASSERT(pri_pipe->bottom_pipe != sec_pipe);
-			sec_pipe->bottom_pipe = pri_pipe->bottom_pipe;
-			sec_pipe->bottom_pipe->top_pipe = sec_pipe;
-		}
-		pri_pipe->bottom_pipe = sec_pipe;
-		sec_pipe->top_pipe = pri_pipe;
-
-		ASSERT(pri_pipe->plane_state);
-	}
-
-	return true;
-}
-
-static struct pipe_ctx *dcn30_find_split_pipe(
-		struct dc *dc,
-		struct dc_state *context,
-		int old_index)
-{
-	struct pipe_ctx *pipe = NULL;
-	int i;
-
-	if (old_index >= 0 && context->res_ctx.pipe_ctx[old_index].stream == NULL) {
-		pipe = &context->res_ctx.pipe_ctx[old_index];
-		pipe->pipe_idx = old_index;
-	}
-
-	if (!pipe)
-		for (i = dc->res_pool->pipe_count - 1; i >= 0; i--) {
-			if (dc->current_state->res_ctx.pipe_ctx[i].top_pipe == NULL
-					&& dc->current_state->res_ctx.pipe_ctx[i].prev_odm_pipe == NULL) {
-				if (context->res_ctx.pipe_ctx[i].stream == NULL) {
-					pipe = &context->res_ctx.pipe_ctx[i];
-					pipe->pipe_idx = i;
-					break;
-				}
-			}
-		}
-
-	/*
-	 * May need to fix pipes getting tossed from 1 opp to another on flip
-	 * Add for debugging transient underflow during topology updates:
-	 * ASSERT(pipe);
-	 */
-	if (!pipe)
-		for (i = dc->res_pool->pipe_count - 1; i >= 0; i--) {
-			if (context->res_ctx.pipe_ctx[i].stream == NULL) {
-				pipe = &context->res_ctx.pipe_ctx[i];
-				pipe->pipe_idx = i;
-				break;
-			}
-		}
-
-	return pipe;
-}
-
-noinline bool dcn30_internal_validate_bw(
-		struct dc *dc,
-		struct dc_state *context,
-		display_e2e_pipe_params_st *pipes,
-		int *pipe_cnt_out,
-		int *vlevel_out,
-		bool fast_validate)
-{
-	bool out = false;
-	bool repopulate_pipes = false;
-	int split[MAX_PIPES] = { 0 };
-	bool merge[MAX_PIPES] = { false };
-	bool newly_split[MAX_PIPES] = { false };
-	int pipe_cnt, i, pipe_idx, vlevel;
-	struct vba_vars_st *vba = &context->bw_ctx.dml.vba;
-
-	ASSERT(pipes);
-	if (!pipes)
-		return false;
-
-	dc->res_pool->funcs->update_soc_for_wm_a(dc, context);
-	pipe_cnt = dc->res_pool->funcs->populate_dml_pipes(dc, context, pipes, fast_validate);
-
-	DC_FP_START();
-	if (!pipe_cnt) {
-		out = true;
-		goto validate_out;
-	}
-
-	dml_log_pipe_params(&context->bw_ctx.dml, pipes, pipe_cnt);
-
-	if (!fast_validate) {
-		/*
-		 * DML favors voltage over p-state, but we're more interested in
-		 * supporting p-state over voltage. We can't support p-state in
-		 * prefetch mode > 0 so try capping the prefetch mode to start.
-		 */
-		context->bw_ctx.dml.soc.allow_dram_self_refresh_or_dram_clock_change_in_vblank =
-			dm_allow_self_refresh_and_mclk_switch;
-		vlevel = dml_get_voltage_level(&context->bw_ctx.dml, pipes, pipe_cnt);
-		/* This may adjust vlevel and maxMpcComb */
-		if (vlevel < context->bw_ctx.dml.soc.num_states)
-			vlevel = dcn20_validate_apply_pipe_split_flags(dc, context, vlevel, split, merge);
-	}
-	if (fast_validate || vlevel == context->bw_ctx.dml.soc.num_states ||
-			vba->DRAMClockChangeSupport[vlevel][vba->maxMpcComb] == dm_dram_clock_change_unsupported) {
-		/*
-		 * If mode is unsupported or there's still no p-state support then
-		 * fall back to favoring voltage.
-		 *
-		 * We don't actually support prefetch mode 2, so require that we
-		 * at least support prefetch mode 1.
-		 */
-		context->bw_ctx.dml.soc.allow_dram_self_refresh_or_dram_clock_change_in_vblank =
-			dm_allow_self_refresh;
-
-		vlevel = dml_get_voltage_level(&context->bw_ctx.dml, pipes, pipe_cnt);
-		if (vlevel < context->bw_ctx.dml.soc.num_states) {
-			memset(split, 0, sizeof(split));
-			memset(merge, 0, sizeof(merge));
-			vlevel = dcn20_validate_apply_pipe_split_flags(dc, context, vlevel, split, merge);
-		}
-	}
-
-	dml_log_mode_support_params(&context->bw_ctx.dml);
-
-	if (vlevel == context->bw_ctx.dml.soc.num_states)
-		goto validate_fail;
-
-	if (!dc->config.enable_windowed_mpo_odm) {
-		for (i = 0, pipe_idx = 0; i < dc->res_pool->pipe_count; i++) {
-			struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
-			struct pipe_ctx *mpo_pipe = pipe->bottom_pipe;
-
-			if (!pipe->stream)
-				continue;
-
-			/* We only support full screen mpo with ODM */
-			if (vba->ODMCombineEnabled[vba->pipe_plane[pipe_idx]] != dm_odm_combine_mode_disabled
-					&& pipe->plane_state && mpo_pipe
-					&& memcmp(&mpo_pipe->plane_res.scl_data.recout,
-							&pipe->plane_res.scl_data.recout,
-							sizeof(struct rect)) != 0) {
-				ASSERT(mpo_pipe->plane_state != pipe->plane_state);
-				goto validate_fail;
-			}
-			pipe_idx++;
-		}
-	}
-
-	/* merge pipes if necessary */
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
-
-		/*skip pipes that don't need merging*/
-		if (!merge[i])
-			continue;
-
-		/* if ODM merge we ignore mpc tree, mpo pipes will have their own flags */
-		if (pipe->prev_odm_pipe) {
-			/*split off odm pipe*/
-			pipe->prev_odm_pipe->next_odm_pipe = pipe->next_odm_pipe;
-			if (pipe->next_odm_pipe)
-				pipe->next_odm_pipe->prev_odm_pipe = pipe->prev_odm_pipe;
-
-			pipe->bottom_pipe = NULL;
-			pipe->next_odm_pipe = NULL;
-			pipe->plane_state = NULL;
-			pipe->stream = NULL;
-			pipe->top_pipe = NULL;
-			pipe->prev_odm_pipe = NULL;
-			if (pipe->stream_res.dsc)
-				dcn20_release_dsc(&context->res_ctx, dc->res_pool, &pipe->stream_res.dsc);
-			memset(&pipe->plane_res, 0, sizeof(pipe->plane_res));
-			memset(&pipe->stream_res, 0, sizeof(pipe->stream_res));
-			repopulate_pipes = true;
-		} else if (pipe->top_pipe && pipe->top_pipe->plane_state == pipe->plane_state) {
-			struct pipe_ctx *top_pipe = pipe->top_pipe;
-			struct pipe_ctx *bottom_pipe = pipe->bottom_pipe;
-
-			top_pipe->bottom_pipe = bottom_pipe;
-			if (bottom_pipe)
-				bottom_pipe->top_pipe = top_pipe;
-
-			pipe->top_pipe = NULL;
-			pipe->bottom_pipe = NULL;
-			pipe->plane_state = NULL;
-			pipe->stream = NULL;
-			memset(&pipe->plane_res, 0, sizeof(pipe->plane_res));
-			memset(&pipe->stream_res, 0, sizeof(pipe->stream_res));
-			repopulate_pipes = true;
-		} else
-			ASSERT(0); /* Should never try to merge master pipe */
-
-	}
-
-	for (i = 0, pipe_idx = -1; i < dc->res_pool->pipe_count; i++) {
-		struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
-		struct pipe_ctx *old_pipe = &dc->current_state->res_ctx.pipe_ctx[i];
-		struct pipe_ctx *hsplit_pipe = NULL;
-		bool odm;
-		int old_index = -1;
-
-		if (!pipe->stream || newly_split[i])
-			continue;
-
-		pipe_idx++;
-		odm = vba->ODMCombineEnabled[vba->pipe_plane[pipe_idx]] != dm_odm_combine_mode_disabled;
-
-		if (!pipe->plane_state && !odm)
-			continue;
-
-		if (split[i]) {
-			if (odm) {
-				if (split[i] == 4 && old_pipe->next_odm_pipe && old_pipe->next_odm_pipe->next_odm_pipe)
-					old_index = old_pipe->next_odm_pipe->next_odm_pipe->pipe_idx;
-				else if (old_pipe->next_odm_pipe)
-					old_index = old_pipe->next_odm_pipe->pipe_idx;
-			} else {
-				if (split[i] == 4 && old_pipe->bottom_pipe && old_pipe->bottom_pipe->bottom_pipe &&
-						old_pipe->bottom_pipe->bottom_pipe->plane_state == old_pipe->plane_state)
-					old_index = old_pipe->bottom_pipe->bottom_pipe->pipe_idx;
-				else if (old_pipe->bottom_pipe &&
-						old_pipe->bottom_pipe->plane_state == old_pipe->plane_state)
-					old_index = old_pipe->bottom_pipe->pipe_idx;
-			}
-			hsplit_pipe = dcn30_find_split_pipe(dc, context, old_index);
-			ASSERT(hsplit_pipe);
-			if (!hsplit_pipe)
-				goto validate_fail;
-
-			if (!dcn30_split_stream_for_mpc_or_odm(
-					dc, &context->res_ctx,
-					pipe, hsplit_pipe, odm))
-				goto validate_fail;
-
-			newly_split[hsplit_pipe->pipe_idx] = true;
-			repopulate_pipes = true;
-		}
-		if (split[i] == 4) {
-			struct pipe_ctx *pipe_4to1;
-
-			if (odm && old_pipe->next_odm_pipe)
-				old_index = old_pipe->next_odm_pipe->pipe_idx;
-			else if (!odm && old_pipe->bottom_pipe &&
-						old_pipe->bottom_pipe->plane_state == old_pipe->plane_state)
-				old_index = old_pipe->bottom_pipe->pipe_idx;
-			else
-				old_index = -1;
-			pipe_4to1 = dcn30_find_split_pipe(dc, context, old_index);
-			ASSERT(pipe_4to1);
-			if (!pipe_4to1)
-				goto validate_fail;
-			if (!dcn30_split_stream_for_mpc_or_odm(
-					dc, &context->res_ctx,
-					pipe, pipe_4to1, odm))
-				goto validate_fail;
-			newly_split[pipe_4to1->pipe_idx] = true;
-
-			if (odm && old_pipe->next_odm_pipe && old_pipe->next_odm_pipe->next_odm_pipe
-					&& old_pipe->next_odm_pipe->next_odm_pipe->next_odm_pipe)
-				old_index = old_pipe->next_odm_pipe->next_odm_pipe->next_odm_pipe->pipe_idx;
-			else if (!odm && old_pipe->bottom_pipe && old_pipe->bottom_pipe->bottom_pipe &&
-					old_pipe->bottom_pipe->bottom_pipe->bottom_pipe &&
-					old_pipe->bottom_pipe->bottom_pipe->bottom_pipe->plane_state == old_pipe->plane_state)
-				old_index = old_pipe->bottom_pipe->bottom_pipe->bottom_pipe->pipe_idx;
-			else
-				old_index = -1;
-			pipe_4to1 = dcn30_find_split_pipe(dc, context, old_index);
-			ASSERT(pipe_4to1);
-			if (!pipe_4to1)
-				goto validate_fail;
-			if (!dcn30_split_stream_for_mpc_or_odm(
-					dc, &context->res_ctx,
-					hsplit_pipe, pipe_4to1, odm))
-				goto validate_fail;
-			newly_split[pipe_4to1->pipe_idx] = true;
-		}
-		if (odm)
-			dcn20_build_mapped_resource(dc, context, pipe->stream);
-	}
-
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
-
-		if (pipe->plane_state) {
-			if (!resource_build_scaling_params(pipe))
-				goto validate_fail;
-		}
-	}
-
-	/* Actual dsc count per stream dsc validation*/
-	if (!dcn20_validate_dsc(dc, context)) {
-		vba->ValidationStatus[vba->soc.num_states] = DML_FAIL_DSC_VALIDATION_FAILURE;
-		goto validate_fail;
-	}
-
-	if (repopulate_pipes)
-		pipe_cnt = dc->res_pool->funcs->populate_dml_pipes(dc, context, pipes, fast_validate);
-	*vlevel_out = vlevel;
-	*pipe_cnt_out = pipe_cnt;
-
-	out = true;
-	goto validate_out;
-
-validate_fail:
-	out = false;
-
-validate_out:
-	DC_FP_END();
-	return out;
-}
-
-/*
- * This must be noinline to ensure anything that deals with FP registers
- * is contained within this call; previously our compiling with hard-float
- * would result in fp instructions being emitted outside of the boundaries
- * of the DC_FP_START/END macros, which makes sense as the compiler has no
- * idea about what is wrapped and what is not
- *
- * This is largely just a workaround to avoid breakage introduced with 5.6,
- * ideally all fp-using code should be moved into its own file, only that
- * should be compiled with hard-float, and all code exported from there
- * should be strictly wrapped with DC_FP_START/END
- */
-static noinline void dcn30_calculate_wm_and_dlg_fp(
-		struct dc *dc, struct dc_state *context,
-		display_e2e_pipe_params_st *pipes,
-		int pipe_cnt,
-		int vlevel)
-{
-	int maxMpcComb = context->bw_ctx.dml.vba.maxMpcComb;
-	int i, pipe_idx;
-	double dcfclk = context->bw_ctx.dml.vba.DCFCLKState[vlevel][maxMpcComb];
-	bool pstate_en = context->bw_ctx.dml.vba.DRAMClockChangeSupport[vlevel][maxMpcComb] != dm_dram_clock_change_unsupported;
-
-	if (context->bw_ctx.dml.soc.min_dcfclk > dcfclk)
-		dcfclk = context->bw_ctx.dml.soc.min_dcfclk;
-
-	pipes[0].clks_cfg.voltage = vlevel;
-	pipes[0].clks_cfg.dcfclk_mhz = dcfclk;
-	pipes[0].clks_cfg.socclk_mhz = context->bw_ctx.dml.soc.clock_limits[vlevel].socclk_mhz;
-
-	/* Set B:
-	 * DCFCLK: 1GHz or min required above 1GHz
-	 * FCLK/UCLK: Max
-	 */
-	if (dc->clk_mgr->bw_params->wm_table.nv_entries[WM_B].valid) {
-		if (vlevel == 0) {
-			pipes[0].clks_cfg.voltage = 1;
-			pipes[0].clks_cfg.dcfclk_mhz = context->bw_ctx.dml.soc.clock_limits[0].dcfclk_mhz;
-		}
-		context->bw_ctx.dml.soc.dram_clock_change_latency_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_B].dml_input.pstate_latency_us;
-		context->bw_ctx.dml.soc.sr_enter_plus_exit_time_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_B].dml_input.sr_enter_plus_exit_time_us;
-		context->bw_ctx.dml.soc.sr_exit_time_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_B].dml_input.sr_exit_time_us;
-	}
-	context->bw_ctx.bw.dcn.watermarks.b.urgent_ns = get_wm_urgent(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.b.cstate_pstate.cstate_enter_plus_exit_ns = get_wm_stutter_enter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.b.cstate_pstate.cstate_exit_ns = get_wm_stutter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.b.cstate_pstate.pstate_change_ns = get_wm_dram_clock_change(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.b.pte_meta_urgent_ns = get_wm_memory_trip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.b.frac_urg_bw_nom = get_fraction_of_urgent_bandwidth(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.b.frac_urg_bw_flip = get_fraction_of_urgent_bandwidth_imm_flip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.b.urgent_latency_ns = get_urgent_latency(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-
-	pipes[0].clks_cfg.voltage = vlevel;
-	pipes[0].clks_cfg.dcfclk_mhz = dcfclk;
-
-	/* Set D:
-	 * DCFCLK: Min Required
-	 * FCLK(proportional to UCLK): 1GHz or Max
-	 * MALL stutter, sr_enter_exit = 4, sr_exit = 2us
-	 */
-	/*
-	if (dc->clk_mgr->bw_params->wm_table.nv_entries[WM_D].valid) {
-		context->bw_ctx.dml.soc.dram_clock_change_latency_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_D].dml_input.pstate_latency_us;
-		context->bw_ctx.dml.soc.sr_enter_plus_exit_time_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_D].dml_input.sr_enter_plus_exit_time_us;
-		context->bw_ctx.dml.soc.sr_exit_time_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_D].dml_input.sr_exit_time_us;
-	}
-	context->bw_ctx.bw.dcn.watermarks.d.urgent_ns = get_wm_urgent(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.d.cstate_pstate.cstate_enter_plus_exit_ns = get_wm_stutter_enter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.d.cstate_pstate.cstate_exit_ns = get_wm_stutter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.d.cstate_pstate.pstate_change_ns = get_wm_dram_clock_change(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.d.pte_meta_urgent_ns = get_wm_memory_trip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.d.frac_urg_bw_nom = get_fraction_of_urgent_bandwidth(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.d.frac_urg_bw_flip = get_fraction_of_urgent_bandwidth_imm_flip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.d.urgent_latency_ns = get_urgent_latency(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	*/
-
-	/* Set C:
-	 * DCFCLK: Min Required
-	 * FCLK(proportional to UCLK): 1GHz or Max
-	 * pstate latency overridden to 5us
-	 */
-	if (dc->clk_mgr->bw_params->wm_table.nv_entries[WM_C].valid) {
-		unsigned int min_dram_speed_mts = context->bw_ctx.dml.vba.DRAMSpeed;
-		unsigned int min_dram_speed_mts_margin = 160;
-
-		if (context->bw_ctx.dml.vba.DRAMClockChangeSupport[vlevel][context->bw_ctx.dml.vba.maxMpcComb] == dm_dram_clock_change_unsupported)
-			min_dram_speed_mts = dc->clk_mgr->bw_params->clk_table.entries[dc->clk_mgr->bw_params->clk_table.num_entries - 1].memclk_mhz * 16;
-
-		/* find largest table entry that is lower than dram speed, but lower than DPM0 still uses DPM0 */
-		for (i = 3; i > 0; i--)
-			if (min_dram_speed_mts + min_dram_speed_mts_margin > dc->clk_mgr->bw_params->dummy_pstate_table[i].dram_speed_mts)
-				break;
-
-		context->bw_ctx.dml.soc.dram_clock_change_latency_us = dc->clk_mgr->bw_params->dummy_pstate_table[i].dummy_pstate_latency_us;
-		context->bw_ctx.dml.soc.sr_enter_plus_exit_time_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_C].dml_input.sr_enter_plus_exit_time_us;
-		context->bw_ctx.dml.soc.sr_exit_time_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_C].dml_input.sr_exit_time_us;
-	}
-
-	context->bw_ctx.bw.dcn.watermarks.c.urgent_ns = get_wm_urgent(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.c.cstate_pstate.cstate_enter_plus_exit_ns = get_wm_stutter_enter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.c.cstate_pstate.cstate_exit_ns = get_wm_stutter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.c.cstate_pstate.pstate_change_ns = get_wm_dram_clock_change(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.c.pte_meta_urgent_ns = get_wm_memory_trip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.c.frac_urg_bw_nom = get_fraction_of_urgent_bandwidth(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.c.frac_urg_bw_flip = get_fraction_of_urgent_bandwidth_imm_flip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.c.urgent_latency_ns = get_urgent_latency(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-
-	if (!pstate_en) {
-		/* The only difference between A and C is p-state latency, if p-state is not supported we want to
-		 * calculate DLG based on dummy p-state latency, and max out the set A p-state watermark
-		 */
-		context->bw_ctx.bw.dcn.watermarks.a = context->bw_ctx.bw.dcn.watermarks.c;
-		context->bw_ctx.bw.dcn.watermarks.a.cstate_pstate.pstate_change_ns = 0;
-	} else {
-		/* Set A:
-		 * DCFCLK: Min Required
-		 * FCLK(proportional to UCLK): 1GHz or Max
-		 *
-		 * Set A calculated last so that following calculations are based on Set A
-		 */
-		dc->res_pool->funcs->update_soc_for_wm_a(dc, context);
-		context->bw_ctx.bw.dcn.watermarks.a.urgent_ns = get_wm_urgent(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-		context->bw_ctx.bw.dcn.watermarks.a.cstate_pstate.cstate_enter_plus_exit_ns = get_wm_stutter_enter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-		context->bw_ctx.bw.dcn.watermarks.a.cstate_pstate.cstate_exit_ns = get_wm_stutter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-		context->bw_ctx.bw.dcn.watermarks.a.cstate_pstate.pstate_change_ns = get_wm_dram_clock_change(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-		context->bw_ctx.bw.dcn.watermarks.a.pte_meta_urgent_ns = get_wm_memory_trip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-		context->bw_ctx.bw.dcn.watermarks.a.frac_urg_bw_nom = get_fraction_of_urgent_bandwidth(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-		context->bw_ctx.bw.dcn.watermarks.a.frac_urg_bw_flip = get_fraction_of_urgent_bandwidth_imm_flip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-		context->bw_ctx.bw.dcn.watermarks.a.urgent_latency_ns = get_urgent_latency(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	}
-
-	context->perf_params.stutter_period_us = context->bw_ctx.dml.vba.StutterPeriod;
-
-	/* Make set D = set A until set D is enabled */
-	context->bw_ctx.bw.dcn.watermarks.d = context->bw_ctx.bw.dcn.watermarks.a;
-
-	for (i = 0, pipe_idx = 0; i < dc->res_pool->pipe_count; i++) {
-		if (!context->res_ctx.pipe_ctx[i].stream)
-			continue;
-
-		pipes[pipe_idx].clks_cfg.dispclk_mhz = get_dispclk_calculated(&context->bw_ctx.dml, pipes, pipe_cnt);
-		pipes[pipe_idx].clks_cfg.dppclk_mhz = get_dppclk_calculated(&context->bw_ctx.dml, pipes, pipe_cnt, pipe_idx);
-
-		if (dc->config.forced_clocks) {
-			pipes[pipe_idx].clks_cfg.dispclk_mhz = context->bw_ctx.dml.soc.clock_limits[0].dispclk_mhz;
-			pipes[pipe_idx].clks_cfg.dppclk_mhz = context->bw_ctx.dml.soc.clock_limits[0].dppclk_mhz;
-		}
-		if (dc->debug.min_disp_clk_khz > pipes[pipe_idx].clks_cfg.dispclk_mhz * 1000)
-			pipes[pipe_idx].clks_cfg.dispclk_mhz = dc->debug.min_disp_clk_khz / 1000.0;
-		if (dc->debug.min_dpp_clk_khz > pipes[pipe_idx].clks_cfg.dppclk_mhz * 1000)
-			pipes[pipe_idx].clks_cfg.dppclk_mhz = dc->debug.min_dpp_clk_khz / 1000.0;
-
-		pipe_idx++;
-	}
-
-	dcn20_calculate_dlg_params(dc, context, pipes, pipe_cnt, vlevel);
-
-	if (!pstate_en)
-		/* Restore full p-state latency */
-		context->bw_ctx.dml.soc.dram_clock_change_latency_us =
-				dc->clk_mgr->bw_params->wm_table.nv_entries[WM_A].dml_input.pstate_latency_us;
-}
-
-void dcn30_update_soc_for_wm_a(struct dc *dc, struct dc_state *context)
-{
-	if (dc->clk_mgr->bw_params->wm_table.nv_entries[WM_A].valid) {
-		context->bw_ctx.dml.soc.dram_clock_change_latency_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_A].dml_input.pstate_latency_us;
-		context->bw_ctx.dml.soc.sr_enter_plus_exit_time_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_A].dml_input.sr_enter_plus_exit_time_us;
-		context->bw_ctx.dml.soc.sr_exit_time_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_A].dml_input.sr_exit_time_us;
-	}
-}
-
-void dcn30_calculate_wm_and_dlg(
-		struct dc *dc, struct dc_state *context,
-		display_e2e_pipe_params_st *pipes,
-		int pipe_cnt,
-		int vlevel)
-{
-	DC_FP_START();
-	dcn30_calculate_wm_and_dlg_fp(dc, context, pipes, pipe_cnt, vlevel);
-	DC_FP_END();
-}
-
-bool dcn30_validate_bandwidth(struct dc *dc,
-		struct dc_state *context,
-		bool fast_validate)
-{
-	bool out = false;
-
-	BW_VAL_TRACE_SETUP();
-
-	int vlevel = 0;
-	int pipe_cnt = 0;
-	display_e2e_pipe_params_st *pipes = kzalloc(dc->res_pool->pipe_count * sizeof(display_e2e_pipe_params_st), GFP_KERNEL);
-	DC_LOGGER_INIT(dc->ctx->logger);
-
-	BW_VAL_TRACE_COUNT();
-
-	out = dcn30_internal_validate_bw(dc, context, pipes, &pipe_cnt, &vlevel, fast_validate);
-
-	if (pipe_cnt == 0)
-		goto validate_out;
-
-	if (!out)
-		goto validate_fail;
-
-	BW_VAL_TRACE_END_VOLTAGE_LEVEL();
-
-	if (fast_validate) {
-		BW_VAL_TRACE_SKIP(fast);
-		goto validate_out;
-	}
-
-	DC_FP_START();
-	dc->res_pool->funcs->calculate_wm_and_dlg(dc, context, pipes, pipe_cnt, vlevel);
-	DC_FP_END();
-
-	BW_VAL_TRACE_END_WATERMARKS();
-
-	goto validate_out;
-
-validate_fail:
-	DC_LOG_WARNING("Mode Validation Warning: %s failed validation.\n",
-		dml_get_status_message(context->bw_ctx.dml.vba.ValidationStatus[context->bw_ctx.dml.vba.soc.num_states]));
-
-	BW_VAL_TRACE_SKIP(fail);
-	out = false;
-
-validate_out:
-	kfree(pipes);
-
-	BW_VAL_TRACE_FINISH();
-
-	return out;
-}
-
-/*
- * This must be noinline to ensure anything that deals with FP registers
- * is contained within this call; previously our compiling with hard-float
- * would result in fp instructions being emitted outside of the boundaries
- * of the DC_FP_START/END macros, which makes sense as the compiler has no
- * idea about what is wrapped and what is not
- *
- * This is largely just a workaround to avoid breakage introduced with 5.6,
- * ideally all fp-using code should be moved into its own file, only that
- * should be compiled with hard-float, and all code exported from there
- * should be strictly wrapped with DC_FP_START/END
- */
-static noinline void dcn30_get_optimal_dcfclk_fclk_for_uclk(unsigned int uclk_mts,
-		unsigned int *optimal_dcfclk,
-		unsigned int *optimal_fclk)
-{
-       double bw_from_dram, bw_from_dram1, bw_from_dram2;
-
-       bw_from_dram1 = uclk_mts * dcn3_0_soc.num_chans *
-		dcn3_0_soc.dram_channel_width_bytes * (dcn3_0_soc.max_avg_dram_bw_use_normal_percent / 100);
-       bw_from_dram2 = uclk_mts * dcn3_0_soc.num_chans *
-		dcn3_0_soc.dram_channel_width_bytes * (dcn3_0_soc.max_avg_sdp_bw_use_normal_percent / 100);
-
-       bw_from_dram = (bw_from_dram1 < bw_from_dram2) ? bw_from_dram1 : bw_from_dram2;
-
-       if (optimal_fclk)
-               *optimal_fclk = bw_from_dram /
-               (dcn3_0_soc.fabric_datapath_to_dcn_data_return_bytes * (dcn3_0_soc.max_avg_sdp_bw_use_normal_percent / 100));
-
-       if (optimal_dcfclk)
-               *optimal_dcfclk =  bw_from_dram /
-               (dcn3_0_soc.return_bus_width_bytes * (dcn3_0_soc.max_avg_sdp_bw_use_normal_percent / 100));
-}
-
-void dcn30_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params)
-{
-	unsigned int i, j;
-	unsigned int num_states = 0;
-
-	unsigned int dcfclk_mhz[DC__VOLTAGE_STATES] = {0};
-	unsigned int dram_speed_mts[DC__VOLTAGE_STATES] = {0};
-	unsigned int optimal_uclk_for_dcfclk_sta_targets[DC__VOLTAGE_STATES] = {0};
-	unsigned int optimal_dcfclk_for_uclk[DC__VOLTAGE_STATES] = {0};
-
-	unsigned int dcfclk_sta_targets[DC__VOLTAGE_STATES] = {694, 875, 1000, 1200};
-	unsigned int num_dcfclk_sta_targets = 4;
-	unsigned int num_uclk_states;
-
-	if (dc->ctx->dc_bios->vram_info.num_chans)
-		dcn3_0_soc.num_chans = dc->ctx->dc_bios->vram_info.num_chans;
-
-	if (dc->ctx->dc_bios->vram_info.dram_channel_width_bytes)
-		dcn3_0_soc.dram_channel_width_bytes = dc->ctx->dc_bios->vram_info.dram_channel_width_bytes;
-
-	dcn3_0_soc.dispclk_dppclk_vco_speed_mhz = dc->clk_mgr->dentist_vco_freq_khz / 1000.0;
-	dc->dml.soc.dispclk_dppclk_vco_speed_mhz = dc->clk_mgr->dentist_vco_freq_khz / 1000.0;
-
-	if (bw_params->clk_table.entries[0].memclk_mhz) {
-		int max_dcfclk_mhz = 0, max_dispclk_mhz = 0, max_dppclk_mhz = 0, max_phyclk_mhz = 0;
-
-		for (i = 0; i < MAX_NUM_DPM_LVL; i++) {
-			if (bw_params->clk_table.entries[i].dcfclk_mhz > max_dcfclk_mhz)
-				max_dcfclk_mhz = bw_params->clk_table.entries[i].dcfclk_mhz;
-			if (bw_params->clk_table.entries[i].dispclk_mhz > max_dispclk_mhz)
-				max_dispclk_mhz = bw_params->clk_table.entries[i].dispclk_mhz;
-			if (bw_params->clk_table.entries[i].dppclk_mhz > max_dppclk_mhz)
-				max_dppclk_mhz = bw_params->clk_table.entries[i].dppclk_mhz;
-			if (bw_params->clk_table.entries[i].phyclk_mhz > max_phyclk_mhz)
-				max_phyclk_mhz = bw_params->clk_table.entries[i].phyclk_mhz;
-		}
-
-		if (!max_dcfclk_mhz)
-			max_dcfclk_mhz = dcn3_0_soc.clock_limits[0].dcfclk_mhz;
-		if (!max_dispclk_mhz)
-			max_dispclk_mhz = dcn3_0_soc.clock_limits[0].dispclk_mhz;
-		if (!max_dppclk_mhz)
-			max_dppclk_mhz = dcn3_0_soc.clock_limits[0].dppclk_mhz;
-		if (!max_phyclk_mhz)
-			max_phyclk_mhz = dcn3_0_soc.clock_limits[0].phyclk_mhz;
-
-		if (max_dcfclk_mhz > dcfclk_sta_targets[num_dcfclk_sta_targets-1]) {
-			// If max DCFCLK is greater than the max DCFCLK STA target, insert into the DCFCLK STA target array
-			dcfclk_sta_targets[num_dcfclk_sta_targets] = max_dcfclk_mhz;
-			num_dcfclk_sta_targets++;
-		} else if (max_dcfclk_mhz < dcfclk_sta_targets[num_dcfclk_sta_targets-1]) {
-			// If max DCFCLK is less than the max DCFCLK STA target, cap values and remove duplicates
-			for (i = 0; i < num_dcfclk_sta_targets; i++) {
-				if (dcfclk_sta_targets[i] > max_dcfclk_mhz) {
-					dcfclk_sta_targets[i] = max_dcfclk_mhz;
-					break;
-				}
-			}
-			// Update size of array since we "removed" duplicates
-			num_dcfclk_sta_targets = i + 1;
-		}
-
-		num_uclk_states = bw_params->clk_table.num_entries;
-
-		// Calculate optimal dcfclk for each uclk
-		for (i = 0; i < num_uclk_states; i++) {
-			DC_FP_START();
-			dcn30_get_optimal_dcfclk_fclk_for_uclk(bw_params->clk_table.entries[i].memclk_mhz * 16,
-					&optimal_dcfclk_for_uclk[i], NULL);
-			DC_FP_END();
-			if (optimal_dcfclk_for_uclk[i] < bw_params->clk_table.entries[0].dcfclk_mhz) {
-				optimal_dcfclk_for_uclk[i] = bw_params->clk_table.entries[0].dcfclk_mhz;
-			}
-		}
-
-		// Calculate optimal uclk for each dcfclk sta target
-		for (i = 0; i < num_dcfclk_sta_targets; i++) {
-			for (j = 0; j < num_uclk_states; j++) {
-				if (dcfclk_sta_targets[i] < optimal_dcfclk_for_uclk[j]) {
-					optimal_uclk_for_dcfclk_sta_targets[i] =
-							bw_params->clk_table.entries[j].memclk_mhz * 16;
-					break;
-				}
-			}
-		}
-
-		i = 0;
-		j = 0;
-		// create the final dcfclk and uclk table
-		while (i < num_dcfclk_sta_targets && j < num_uclk_states && num_states < DC__VOLTAGE_STATES) {
-			if (dcfclk_sta_targets[i] < optimal_dcfclk_for_uclk[j] && i < num_dcfclk_sta_targets) {
-				dcfclk_mhz[num_states] = dcfclk_sta_targets[i];
-				dram_speed_mts[num_states++] = optimal_uclk_for_dcfclk_sta_targets[i++];
-			} else {
-				if (j < num_uclk_states && optimal_dcfclk_for_uclk[j] <= max_dcfclk_mhz) {
-					dcfclk_mhz[num_states] = optimal_dcfclk_for_uclk[j];
-					dram_speed_mts[num_states++] = bw_params->clk_table.entries[j++].memclk_mhz * 16;
-				} else {
-					j = num_uclk_states;
-				}
-			}
-		}
-
-		while (i < num_dcfclk_sta_targets && num_states < DC__VOLTAGE_STATES) {
-			dcfclk_mhz[num_states] = dcfclk_sta_targets[i];
-			dram_speed_mts[num_states++] = optimal_uclk_for_dcfclk_sta_targets[i++];
-		}
-
-		while (j < num_uclk_states && num_states < DC__VOLTAGE_STATES &&
-				optimal_dcfclk_for_uclk[j] <= max_dcfclk_mhz) {
-			dcfclk_mhz[num_states] = optimal_dcfclk_for_uclk[j];
-			dram_speed_mts[num_states++] = bw_params->clk_table.entries[j++].memclk_mhz * 16;
-		}
-
-		dcn3_0_soc.num_states = num_states;
-		for (i = 0; i < dcn3_0_soc.num_states; i++) {
-			dcn3_0_soc.clock_limits[i].state = i;
-			dcn3_0_soc.clock_limits[i].dcfclk_mhz = dcfclk_mhz[i];
-			dcn3_0_soc.clock_limits[i].fabricclk_mhz = dcfclk_mhz[i];
-			dcn3_0_soc.clock_limits[i].dram_speed_mts = dram_speed_mts[i];
-
-			/* Fill all states with max values of all other clocks */
-			dcn3_0_soc.clock_limits[i].dispclk_mhz = max_dispclk_mhz;
-			dcn3_0_soc.clock_limits[i].dppclk_mhz  = max_dppclk_mhz;
-			dcn3_0_soc.clock_limits[i].phyclk_mhz  = max_phyclk_mhz;
-			dcn3_0_soc.clock_limits[i].dtbclk_mhz = dcn3_0_soc.clock_limits[0].dtbclk_mhz;
-			/* These clocks cannot come from bw_params, always fill from dcn3_0_soc[1] */
-			/* FCLK, PHYCLK_D18, SOCCLK, DSCCLK */
-			dcn3_0_soc.clock_limits[i].phyclk_d18_mhz = dcn3_0_soc.clock_limits[0].phyclk_d18_mhz;
-			dcn3_0_soc.clock_limits[i].socclk_mhz = dcn3_0_soc.clock_limits[0].socclk_mhz;
-			dcn3_0_soc.clock_limits[i].dscclk_mhz = dcn3_0_soc.clock_limits[0].dscclk_mhz;
-		}
-		/* re-init DML with updated bb */
-		dml_init_instance(&dc->dml, &dcn3_0_soc, &dcn3_0_ip, DML_PROJECT_DCN30);
-		if (dc->current_state)
-			dml_init_instance(&dc->current_state->bw_ctx.dml, &dcn3_0_soc, &dcn3_0_ip, DML_PROJECT_DCN30);
-	}
-}
-
-static const struct resource_funcs dcn30_res_pool_funcs = {
-	.destroy = dcn30_destroy_resource_pool,
-	.link_enc_create = dcn30_link_encoder_create,
-	.panel_cntl_create = dcn30_panel_cntl_create,
-	.validate_bandwidth = dcn30_validate_bandwidth,
-	.calculate_wm_and_dlg = dcn30_calculate_wm_and_dlg,
-	.update_soc_for_wm_a = dcn30_update_soc_for_wm_a,
-	.populate_dml_pipes = dcn30_populate_dml_pipes_from_context,
-	.acquire_idle_pipe_for_layer = dcn20_acquire_idle_pipe_for_layer,
-	.add_stream_to_ctx = dcn30_add_stream_to_ctx,
-	.add_dsc_to_stream_resource = dcn20_add_dsc_to_stream_resource,
-	.remove_stream_from_ctx = dcn20_remove_stream_from_ctx,
-	.populate_dml_writeback_from_context = dcn30_populate_dml_writeback_from_context,
-	.set_mcif_arb_params = dcn30_set_mcif_arb_params,
-	.find_first_free_match_stream_enc_for_link = dcn10_find_first_free_match_stream_enc_for_link,
-	.acquire_post_bldn_3dlut = dcn30_acquire_post_bldn_3dlut,
-	.release_post_bldn_3dlut = dcn30_release_post_bldn_3dlut,
-	.update_bw_bounding_box = dcn30_update_bw_bounding_box,
-	.patch_unknown_plane_state = dcn20_patch_unknown_plane_state,
-};
-
-#define CTX ctx
-
-#define REG(reg_name) \
-	(DCN_BASE.instance[0].segment[mm ## reg_name ## _BASE_IDX] + mm ## reg_name)
-
-static uint32_t read_pipe_fuses(struct dc_context *ctx)
-{
-	uint32_t value = REG_READ(CC_DC_PIPE_DIS);
-	/* Support for max 6 pipes */
-	value = value & 0x3f;
-	return value;
-}
-
-static bool dcn30_resource_construct(
-	uint8_t num_virtual_links,
-	struct dc *dc,
-	struct dcn30_resource_pool *pool)
-{
-	int i;
-	struct dc_context *ctx = dc->ctx;
-	struct irq_service_init_data init_data;
-	struct ddc_service_init_data ddc_init_data = {0};
-	uint32_t pipe_fuses = read_pipe_fuses(ctx);
-	uint32_t num_pipes = 0;
-
-	if (!(pipe_fuses == 0 || pipe_fuses == 0x3e)) {
-		BREAK_TO_DEBUGGER();
-		dm_error("DC: Unexpected fuse recipe for navi2x !\n");
-		/* fault to single pipe */
-		pipe_fuses = 0x3e;
-	}
-
-	DC_FP_START();
-
-	ctx->dc_bios->regs = &bios_regs;
-
-	pool->base.res_cap = &res_cap_dcn3;
-
-	pool->base.funcs = &dcn30_res_pool_funcs;
-
-	/*************************************************
-	 *  Resource + asic cap harcoding                *
-	 *************************************************/
-	pool->base.underlay_pipe_index = NO_UNDERLAY_PIPE;
-	pool->base.pipe_count = pool->base.res_cap->num_timing_generator;
-	pool->base.mpcc_count = pool->base.res_cap->num_timing_generator;
-	dc->caps.max_downscale_ratio = 600;
-	dc->caps.i2c_speed_in_khz = 100;
-	dc->caps.i2c_speed_in_khz_hdcp = 100; /*1.4 w/a not applied by default*/
-	dc->caps.max_cursor_size = 256;
-	dc->caps.min_horizontal_blanking_period = 80;
-	dc->caps.dmdata_alloc_size = 2048;
-	dc->caps.mall_size_per_mem_channel = 8;
-	/* total size = mall per channel * num channels * 1024 * 1024 */
-	dc->caps.mall_size_total = dc->caps.mall_size_per_mem_channel * dc->ctx->dc_bios->vram_info.num_chans * 1048576;
-	dc->caps.cursor_cache_size = dc->caps.max_cursor_size * dc->caps.max_cursor_size * 8;
-
-	dc->caps.max_slave_planes = 1;
-	dc->caps.max_slave_yuv_planes = 1;
-	dc->caps.max_slave_rgb_planes = 1;
-	dc->caps.post_blend_color_processing = true;
-	dc->caps.force_dp_tps4_for_cp2520 = true;
-	dc->caps.extended_aux_timeout_support = true;
-	dc->caps.dmcub_support = true;
-
-	/* Color pipeline capabilities */
-	dc->caps.color.dpp.dcn_arch = 1;
-	dc->caps.color.dpp.input_lut_shared = 0;
-	dc->caps.color.dpp.icsc = 1;
-	dc->caps.color.dpp.dgam_ram = 0; // must use gamma_corr
-	dc->caps.color.dpp.dgam_rom_caps.srgb = 1;
-	dc->caps.color.dpp.dgam_rom_caps.bt2020 = 1;
-	dc->caps.color.dpp.dgam_rom_caps.gamma2_2 = 1;
-	dc->caps.color.dpp.dgam_rom_caps.pq = 1;
-	dc->caps.color.dpp.dgam_rom_caps.hlg = 1;
-	dc->caps.color.dpp.post_csc = 1;
-	dc->caps.color.dpp.gamma_corr = 1;
-	dc->caps.color.dpp.dgam_rom_for_yuv = 0;
-
-	dc->caps.color.dpp.hw_3d_lut = 1;
-	dc->caps.color.dpp.ogam_ram = 1;
-	// no OGAM ROM on DCN3
-	dc->caps.color.dpp.ogam_rom_caps.srgb = 0;
-	dc->caps.color.dpp.ogam_rom_caps.bt2020 = 0;
-	dc->caps.color.dpp.ogam_rom_caps.gamma2_2 = 0;
-	dc->caps.color.dpp.ogam_rom_caps.pq = 0;
-	dc->caps.color.dpp.ogam_rom_caps.hlg = 0;
-	dc->caps.color.dpp.ocsc = 0;
-
-	dc->caps.color.mpc.gamut_remap = 1;
-	dc->caps.color.mpc.num_3dluts = pool->base.res_cap->num_mpc_3dlut; //3
-	dc->caps.color.mpc.ogam_ram = 1;
-	dc->caps.color.mpc.ogam_rom_caps.srgb = 0;
-	dc->caps.color.mpc.ogam_rom_caps.bt2020 = 0;
-	dc->caps.color.mpc.ogam_rom_caps.gamma2_2 = 0;
-	dc->caps.color.mpc.ogam_rom_caps.pq = 0;
-	dc->caps.color.mpc.ogam_rom_caps.hlg = 0;
-	dc->caps.color.mpc.ocsc = 1;
-
-	/* read VBIOS LTTPR caps */
-	{
-		if (ctx->dc_bios->funcs->get_lttpr_caps) {
-			enum bp_result bp_query_result;
-			uint8_t is_vbios_lttpr_enable = 0;
-
-			bp_query_result = ctx->dc_bios->funcs->get_lttpr_caps(ctx->dc_bios, &is_vbios_lttpr_enable);
-			dc->caps.vbios_lttpr_enable = (bp_query_result == BP_RESULT_OK) && !!is_vbios_lttpr_enable;
-		}
-
-		if (ctx->dc_bios->funcs->get_lttpr_interop) {
-			enum bp_result bp_query_result;
-			uint8_t is_vbios_interop_enabled = 0;
-
-			bp_query_result = ctx->dc_bios->funcs->get_lttpr_interop(ctx->dc_bios,
-					&is_vbios_interop_enabled);
-			dc->caps.vbios_lttpr_aware = (bp_query_result == BP_RESULT_OK) && !!is_vbios_interop_enabled;
-		}
-	}
-
-	if (dc->ctx->dce_environment == DCE_ENV_PRODUCTION_DRV)
-		dc->debug = debug_defaults_drv;
-	else if (dc->ctx->dce_environment == DCE_ENV_FPGA_MAXIMUS) {
-		dc->debug = debug_defaults_diags;
-	} else
-		dc->debug = debug_defaults_diags;
-	// Init the vm_helper
-	if (dc->vm_helper)
-		vm_helper_init(dc->vm_helper, 16);
-
-	/*************************************************
-	 *  Create resources                             *
-	 *************************************************/
-
-	/* Clock Sources for Pixel Clock*/
-	pool->base.clock_sources[DCN30_CLK_SRC_PLL0] =
-			dcn30_clock_source_create(ctx, ctx->dc_bios,
-				CLOCK_SOURCE_COMBO_PHY_PLL0,
-				&clk_src_regs[0], false);
-	pool->base.clock_sources[DCN30_CLK_SRC_PLL1] =
-			dcn30_clock_source_create(ctx, ctx->dc_bios,
-				CLOCK_SOURCE_COMBO_PHY_PLL1,
-				&clk_src_regs[1], false);
-	pool->base.clock_sources[DCN30_CLK_SRC_PLL2] =
-			dcn30_clock_source_create(ctx, ctx->dc_bios,
-				CLOCK_SOURCE_COMBO_PHY_PLL2,
-				&clk_src_regs[2], false);
-	pool->base.clock_sources[DCN30_CLK_SRC_PLL3] =
-			dcn30_clock_source_create(ctx, ctx->dc_bios,
-				CLOCK_SOURCE_COMBO_PHY_PLL3,
-				&clk_src_regs[3], false);
-	pool->base.clock_sources[DCN30_CLK_SRC_PLL4] =
-			dcn30_clock_source_create(ctx, ctx->dc_bios,
-				CLOCK_SOURCE_COMBO_PHY_PLL4,
-				&clk_src_regs[4], false);
-	pool->base.clock_sources[DCN30_CLK_SRC_PLL5] =
-			dcn30_clock_source_create(ctx, ctx->dc_bios,
-				CLOCK_SOURCE_COMBO_PHY_PLL5,
-				&clk_src_regs[5], false);
-
-	pool->base.clk_src_count = DCN30_CLK_SRC_TOTAL;
-
-	/* todo: not reuse phy_pll registers */
-	pool->base.dp_clock_source =
-			dcn30_clock_source_create(ctx, ctx->dc_bios,
-				CLOCK_SOURCE_ID_DP_DTO,
-				&clk_src_regs[0], true);
-
-	for (i = 0; i < pool->base.clk_src_count; i++) {
-		if (pool->base.clock_sources[i] == NULL) {
-			dm_error("DC: failed to create clock sources!\n");
-			BREAK_TO_DEBUGGER();
-			goto create_fail;
-		}
-	}
-
-	/* DCCG */
-	pool->base.dccg = dccg30_create(ctx, &dccg_regs, &dccg_shift, &dccg_mask);
-	if (pool->base.dccg == NULL) {
-		dm_error("DC: failed to create dccg!\n");
-		BREAK_TO_DEBUGGER();
-		goto create_fail;
-	}
-
-	/* PP Lib and SMU interfaces */
-	init_soc_bounding_box(dc, pool);
-
-	num_pipes = dcn3_0_ip.max_num_dpp;
-
-	for (i = 0; i < dcn3_0_ip.max_num_dpp; i++)
-		if (pipe_fuses & 1 << i)
-			num_pipes--;
-
-	dcn3_0_ip.max_num_dpp = num_pipes;
-	dcn3_0_ip.max_num_otg = num_pipes;
-
-	dml_init_instance(&dc->dml, &dcn3_0_soc, &dcn3_0_ip, DML_PROJECT_DCN30);
-
-	/* IRQ */
-	init_data.ctx = dc->ctx;
-	pool->base.irqs = dal_irq_service_dcn30_create(&init_data);
-	if (!pool->base.irqs)
-		goto create_fail;
-
-	/* HUBBUB */
-	pool->base.hubbub = dcn30_hubbub_create(ctx);
-	if (pool->base.hubbub == NULL) {
-		BREAK_TO_DEBUGGER();
-		dm_error("DC: failed to create hubbub!\n");
-		goto create_fail;
-	}
-
-	/* HUBPs, DPPs, OPPs and TGs */
-	for (i = 0; i < pool->base.pipe_count; i++) {
-		pool->base.hubps[i] = dcn30_hubp_create(ctx, i);
-		if (pool->base.hubps[i] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error(
-				"DC: failed to create hubps!\n");
-			goto create_fail;
-		}
-
-		pool->base.dpps[i] = dcn30_dpp_create(ctx, i);
-		if (pool->base.dpps[i] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error(
-				"DC: failed to create dpps!\n");
-			goto create_fail;
-		}
-	}
-
-	for (i = 0; i < pool->base.res_cap->num_opp; i++) {
-		pool->base.opps[i] = dcn30_opp_create(ctx, i);
-		if (pool->base.opps[i] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error(
-				"DC: failed to create output pixel processor!\n");
-			goto create_fail;
-		}
-	}
-
-	for (i = 0; i < pool->base.res_cap->num_timing_generator; i++) {
-		pool->base.timing_generators[i] = dcn30_timing_generator_create(
-				ctx, i);
-		if (pool->base.timing_generators[i] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error("DC: failed to create tg!\n");
-			goto create_fail;
-		}
-	}
-	pool->base.timing_generator_count = i;
-	/* PSR */
-	pool->base.psr = dmub_psr_create(ctx);
-
-	if (pool->base.psr == NULL) {
-		dm_error("DC: failed to create PSR obj!\n");
-		BREAK_TO_DEBUGGER();
-		goto create_fail;
-	}
-
-	/* ABM */
-	for (i = 0; i < pool->base.res_cap->num_timing_generator; i++) {
-		pool->base.multiple_abms[i] = dmub_abm_create(ctx,
-				&abm_regs[i],
-				&abm_shift,
-				&abm_mask);
-		if (pool->base.multiple_abms[i] == NULL) {
-			dm_error("DC: failed to create abm for pipe %d!\n", i);
-			BREAK_TO_DEBUGGER();
-			goto create_fail;
-		}
-	}
-	/* MPC and DSC */
-	pool->base.mpc = dcn30_mpc_create(ctx, pool->base.mpcc_count, pool->base.res_cap->num_mpc_3dlut);
-	if (pool->base.mpc == NULL) {
-		BREAK_TO_DEBUGGER();
-		dm_error("DC: failed to create mpc!\n");
-		goto create_fail;
-	}
-
-	for (i = 0; i < pool->base.res_cap->num_dsc; i++) {
-		pool->base.dscs[i] = dcn30_dsc_create(ctx, i);
-		if (pool->base.dscs[i] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error("DC: failed to create display stream compressor %d!\n", i);
-			goto create_fail;
-		}
-	}
-
-	/* DWB and MMHUBBUB */
-	if (!dcn30_dwbc_create(ctx, &pool->base)) {
-		BREAK_TO_DEBUGGER();
-		dm_error("DC: failed to create dwbc!\n");
-		goto create_fail;
-	}
-
-	if (!dcn30_mmhubbub_create(ctx, &pool->base)) {
-		BREAK_TO_DEBUGGER();
-		dm_error("DC: failed to create mcif_wb!\n");
-		goto create_fail;
-	}
-
-	/* AUX and I2C */
-	for (i = 0; i < pool->base.res_cap->num_ddc; i++) {
-		pool->base.engines[i] = dcn30_aux_engine_create(ctx, i);
-		if (pool->base.engines[i] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error(
-				"DC:failed to create aux engine!!\n");
-			goto create_fail;
-		}
-		pool->base.hw_i2cs[i] = dcn30_i2c_hw_create(ctx, i);
-		if (pool->base.hw_i2cs[i] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error(
-				"DC:failed to create hw i2c!!\n");
-			goto create_fail;
-		}
-		pool->base.sw_i2cs[i] = NULL;
-	}
-
-	/* Audio, Stream Encoders including DIG and virtual, MPC 3D LUTs */
-	if (!resource_construct(num_virtual_links, dc, &pool->base,
-			(!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment) ?
-			&res_create_funcs : &res_create_maximus_funcs)))
-		goto create_fail;
-
-	/* HW Sequencer and Plane caps */
-	dcn30_hw_sequencer_construct(dc);
-
-	dc->caps.max_planes =  pool->base.pipe_count;
-
-	for (i = 0; i < dc->caps.max_planes; ++i)
-		dc->caps.planes[i] = plane_cap;
-
-	dc->cap_funcs = cap_funcs;
-
-	if (dc->ctx->dc_bios->fw_info.oem_i2c_present) {
-		ddc_init_data.ctx = dc->ctx;
-		ddc_init_data.link = NULL;
-		ddc_init_data.id.id = dc->ctx->dc_bios->fw_info.oem_i2c_obj_id;
-		ddc_init_data.id.enum_id = 0;
-		ddc_init_data.id.type = OBJECT_TYPE_GENERIC;
-		pool->base.oem_device = dal_ddc_service_create(&ddc_init_data);
-	} else {
-		pool->base.oem_device = NULL;
-	}
-
-	DC_FP_END();
-
-	return true;
-
-create_fail:
-
-	DC_FP_END();
-	dcn30_resource_destruct(pool);
-
-	return false;
-}
-
-struct resource_pool *dcn30_create_resource_pool(
-		const struct dc_init_data *init_data,
-		struct dc *dc)
-{
-	struct dcn30_resource_pool *pool =
-		kzalloc(sizeof(struct dcn30_resource_pool), GFP_KERNEL);
-
-	if (!pool)
-		return NULL;
-
-	if (dcn30_resource_construct(init_data->num_virtual_links, dc, pool))
-		return &pool->base;
-
-	BREAK_TO_DEBUGGER();
-	kfree(pool);
-	return NULL;
-}
diff --git a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_vpg.h b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_vpg.h
index ed9a5549c389..466ba20b9c61 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_vpg.h
+++ b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_vpg.h
@@ -26,6 +26,7 @@
 #ifndef __DAL_DCN30_VPG_H__
 #define __DAL_DCN30_VPG_H__
 
+#include "vpg.h"
 
 #define DCN30_VPG_FROM_VPG(vpg)\
 	container_of(vpg, struct dcn30_vpg, base)
@@ -132,28 +133,6 @@ struct dcn30_vpg_mask {
 	VPG_DCN3_REG_FIELD_LIST(uint32_t);
 };
 
-struct vpg;
-
-struct vpg_funcs {
-	void (*update_generic_info_packet)(
-		struct vpg *vpg,
-		uint32_t packet_index,
-		const struct dc_info_packet *info_packet,
-		bool immediate_update);
-
-	void (*vpg_poweron)(
-		struct vpg *vpg);
-
-	void (*vpg_powerdown)(
-		struct vpg *vpg);
-};
-
-struct vpg {
-	const struct vpg_funcs *funcs;
-	struct dc_context *ctx;
-	int inst;
-};
-
 struct dcn30_vpg {
 	struct vpg base;
 	const struct dcn30_vpg_registers *regs;
diff --git a/drivers/gpu/drm/amd/display/dc/dcn301/Makefile b/drivers/gpu/drm/amd/display/dc/dcn301/Makefile
index 7aa628c21973..fb4814ab3f05 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn301/Makefile
+++ b/drivers/gpu/drm/amd/display/dc/dcn301/Makefile
@@ -1,17 +1,9 @@
-#
-# (c) Copyright 2020 Advanced Micro Devices, Inc. All the rights reserved
-#
-#  All rights reserved.  This notice is intended as a precaution against
-#  inadvertent publication and does not imply publication or any waiver
-#  of confidentiality.  The year included in the foregoing notice is the
-#  year of creation of the work.
-#
-#  Authors: AMD
+# SPDX-License-Identifier: MIT
+# Copyright © 2024 Advanced Micro Devices, Inc. All rights reserved.
 #
 # Makefile for dcn30.
 
-DCN301 = dcn301_init.o dcn301_resource.o dcn301_dccg.o \
-		dcn301_dio_link_encoder.o dcn301_hwseq.o dcn301_panel_cntl.o dcn301_hubbub.o
+DCN301 = dcn301_panel_cntl.o
 
 AMD_DAL_DCN301 = $(addprefix $(AMDDALPATH)/dc/dcn301/,$(DCN301))
 
diff --git a/drivers/gpu/drm/amd/display/dc/dcn301/dcn301_panel_cntl.c b/drivers/gpu/drm/amd/display/dc/dcn301/dcn301_panel_cntl.c
index 736bda30abc3..9e96a3ace207 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn301/dcn301_panel_cntl.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn301/dcn301_panel_cntl.c
@@ -93,7 +93,7 @@ static unsigned int dcn301_get_16_bit_backlight_from_pwm(struct panel_cntl *pane
 	return (uint32_t)(current_backlight);
 }
 
-uint32_t dcn301_panel_cntl_hw_init(struct panel_cntl *panel_cntl)
+static uint32_t dcn301_panel_cntl_hw_init(struct panel_cntl *panel_cntl)
 {
 	struct dcn301_panel_cntl *dcn301_panel_cntl = TO_DCN301_PANEL_CNTL(panel_cntl);
 	uint32_t value;
@@ -147,7 +147,7 @@ uint32_t dcn301_panel_cntl_hw_init(struct panel_cntl *panel_cntl)
 	return current_backlight;
 }
 
-void dcn301_panel_cntl_destroy(struct panel_cntl **panel_cntl)
+static void dcn301_panel_cntl_destroy(struct panel_cntl **panel_cntl)
 {
 	struct dcn301_panel_cntl *dcn301_panel_cntl = TO_DCN301_PANEL_CNTL(*panel_cntl);
 
@@ -155,7 +155,7 @@ void dcn301_panel_cntl_destroy(struct panel_cntl **panel_cntl)
 	*panel_cntl = NULL;
 }
 
-bool dcn301_is_panel_backlight_on(struct panel_cntl *panel_cntl)
+static bool dcn301_is_panel_backlight_on(struct panel_cntl *panel_cntl)
 {
 	struct dcn301_panel_cntl *dcn301_panel_cntl = TO_DCN301_PANEL_CNTL(panel_cntl);
 	uint32_t value;
@@ -165,7 +165,7 @@ bool dcn301_is_panel_backlight_on(struct panel_cntl *panel_cntl)
 	return value;
 }
 
-bool dcn301_is_panel_powered_on(struct panel_cntl *panel_cntl)
+static bool dcn301_is_panel_powered_on(struct panel_cntl *panel_cntl)
 {
 	struct dcn301_panel_cntl *dcn301_panel_cntl = TO_DCN301_PANEL_CNTL(panel_cntl);
 	uint32_t pwr_seq_state, dig_on, dig_on_ovrd;
@@ -177,7 +177,7 @@ bool dcn301_is_panel_powered_on(struct panel_cntl *panel_cntl)
 	return (pwr_seq_state == 1) || (dig_on == 1 && dig_on_ovrd == 1);
 }
 
-void dcn301_store_backlight_level(struct panel_cntl *panel_cntl)
+static void dcn301_store_backlight_level(struct panel_cntl *panel_cntl)
 {
 	struct dcn301_panel_cntl *dcn301_panel_cntl = TO_DCN301_PANEL_CNTL(panel_cntl);
 
@@ -215,4 +215,5 @@ void dcn301_panel_cntl_construct(
 	dcn301_panel_cntl->base.funcs = &dcn301_link_panel_cntl_funcs;
 	dcn301_panel_cntl->base.ctx = init_data->ctx;
 	dcn301_panel_cntl->base.inst = init_data->inst;
+	dcn301_panel_cntl->base.pwrseq_inst = 0;
 }
diff --git a/drivers/gpu/drm/amd/display/dc/dcn302/Makefile b/drivers/gpu/drm/amd/display/dc/dcn302/Makefile
deleted file mode 100644
index 101620a8867a..000000000000
--- a/drivers/gpu/drm/amd/display/dc/dcn302/Makefile
+++ /dev/null
@@ -1,43 +0,0 @@
-#
-# (c) Copyright 2020 Advanced Micro Devices, Inc. All the rights reserved
-#
-#  All rights reserved.  This notice is intended as a precaution against
-#  inadvertent publication and does not imply publication or any waiver
-#  of confidentiality.  The year included in the foregoing notice is the
-#  year of creation of the work.
-#
-#  Authors: AMD
-#
-# Makefile for dcn302.
-
-DCN3_02 = dcn302_init.o dcn302_hwseq.o dcn302_resource.o
-
-ifdef CONFIG_X86
-CFLAGS_$(AMDDALPATH)/dc/dcn302/dcn302_resource.o := -msse
-endif
-
-ifdef CONFIG_PPC64
-CFLAGS_$(AMDDALPATH)/dc/dcn302/dcn302_resource.o := -mhard-float -maltivec
-endif
-
-ifdef CONFIG_CC_IS_GCC
-ifeq ($(call cc-ifversion, -lt, 0701, y), y)
-IS_OLD_GCC = 1
-endif
-CFLAGS_$(AMDDALPATH)/dc/dcn302/dcn302_resource.o += -mhard-float
-endif
-
-ifdef CONFIG_X86
-ifdef IS_OLD_GCC
-# Stack alignment mismatch, proceed with caution.
-# GCC < 7.1 cannot compile code using `double` and -mpreferred-stack-boundary=3
-# (8B stack alignment).
-CFLAGS_$(AMDDALPATH)/dc/dcn302/dcn302_resource.o += -mpreferred-stack-boundary=4
-else
-CFLAGS_$(AMDDALPATH)/dc/dcn302/dcn302_resource.o += -msse2
-endif
-endif
-
-AMD_DAL_DCN3_02 = $(addprefix $(AMDDALPATH)/dc/dcn302/,$(DCN3_02))
-
-AMD_DISPLAY_FILES += $(AMD_DAL_DCN3_02)
diff --git a/drivers/gpu/drm/amd/display/dc/dcn302/dcn302_resource.c b/drivers/gpu/drm/amd/display/dc/dcn302/dcn302_resource.c
deleted file mode 100644
index fcf96cf08c76..000000000000
--- a/drivers/gpu/drm/amd/display/dc/dcn302/dcn302_resource.c
+++ /dev/null
@@ -1,1778 +0,0 @@
-/*
- * Copyright 2020 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: AMD
- *
- */
-
-#include "dcn302_init.h"
-#include "dcn302_resource.h"
-#include "dcn302_dccg.h"
-#include "irq/dcn302/irq_service_dcn302.h"
-
-#include "dcn30/dcn30_dio_link_encoder.h"
-#include "dcn30/dcn30_dio_stream_encoder.h"
-#include "dcn30/dcn30_dwb.h"
-#include "dcn30/dcn30_dpp.h"
-#include "dcn30/dcn30_hubbub.h"
-#include "dcn30/dcn30_hubp.h"
-#include "dcn30/dcn30_mmhubbub.h"
-#include "dcn30/dcn30_mpc.h"
-#include "dcn30/dcn30_opp.h"
-#include "dcn30/dcn30_optc.h"
-#include "dcn30/dcn30_resource.h"
-
-#include "dcn20/dcn20_dsc.h"
-#include "dcn20/dcn20_resource.h"
-
-#include "dcn10/dcn10_resource.h"
-
-#include "dce/dce_abm.h"
-#include "dce/dce_audio.h"
-#include "dce/dce_aux.h"
-#include "dce/dce_clock_source.h"
-#include "dce/dce_hwseq.h"
-#include "dce/dce_i2c_hw.h"
-#include "dce/dce_panel_cntl.h"
-#include "dce/dmub_abm.h"
-#include "dce/dmub_psr.h"
-#include "clk_mgr.h"
-
-#include "hw_sequencer_private.h"
-#include "reg_helper.h"
-#include "resource.h"
-#include "vm_helper.h"
-
-#include "dimgrey_cavefish_ip_offset.h"
-#include "dcn/dcn_3_0_2_offset.h"
-#include "dcn/dcn_3_0_2_sh_mask.h"
-#include "dcn/dpcs_3_0_0_offset.h"
-#include "dcn/dpcs_3_0_0_sh_mask.h"
-#include "nbio/nbio_7_4_offset.h"
-#include "amdgpu_socbb.h"
-
-#define DC_LOGGER_INIT(logger)
-
-struct _vcs_dpi_ip_params_st dcn3_02_ip = {
-		.use_min_dcfclk = 0,
-		.clamp_min_dcfclk = 0,
-		.odm_capable = 1,
-		.gpuvm_enable = 1,
-		.hostvm_enable = 0,
-		.gpuvm_max_page_table_levels = 4,
-		.hostvm_max_page_table_levels = 4,
-		.hostvm_cached_page_table_levels = 0,
-		.pte_group_size_bytes = 2048,
-		.num_dsc = 5,
-		.rob_buffer_size_kbytes = 184,
-		.det_buffer_size_kbytes = 184,
-		.dpte_buffer_size_in_pte_reqs_luma = 64,
-		.dpte_buffer_size_in_pte_reqs_chroma = 34,
-		.pde_proc_buffer_size_64k_reqs = 48,
-		.dpp_output_buffer_pixels = 2560,
-		.opp_output_buffer_lines = 1,
-		.pixel_chunk_size_kbytes = 8,
-		.pte_enable = 1,
-		.max_page_table_levels = 2,
-		.pte_chunk_size_kbytes = 2,  // ?
-		.meta_chunk_size_kbytes = 2,
-		.writeback_chunk_size_kbytes = 8,
-		.line_buffer_size_bits = 789504,
-		.is_line_buffer_bpp_fixed = 0,  // ?
-		.line_buffer_fixed_bpp = 0,     // ?
-		.dcc_supported = true,
-		.writeback_interface_buffer_size_kbytes = 90,
-		.writeback_line_buffer_buffer_size = 0,
-		.max_line_buffer_lines = 12,
-		.writeback_luma_buffer_size_kbytes = 12,  // writeback_line_buffer_buffer_size = 656640
-		.writeback_chroma_buffer_size_kbytes = 8,
-		.writeback_chroma_line_buffer_width_pixels = 4,
-		.writeback_max_hscl_ratio = 1,
-		.writeback_max_vscl_ratio = 1,
-		.writeback_min_hscl_ratio = 1,
-		.writeback_min_vscl_ratio = 1,
-		.writeback_max_hscl_taps = 1,
-		.writeback_max_vscl_taps = 1,
-		.writeback_line_buffer_luma_buffer_size = 0,
-		.writeback_line_buffer_chroma_buffer_size = 14643,
-		.cursor_buffer_size = 8,
-		.cursor_chunk_size = 2,
-		.max_num_otg = 5,
-		.max_num_dpp = 5,
-		.max_num_wb = 1,
-		.max_dchub_pscl_bw_pix_per_clk = 4,
-		.max_pscl_lb_bw_pix_per_clk = 2,
-		.max_lb_vscl_bw_pix_per_clk = 4,
-		.max_vscl_hscl_bw_pix_per_clk = 4,
-		.max_hscl_ratio = 6,
-		.max_vscl_ratio = 6,
-		.hscl_mults = 4,
-		.vscl_mults = 4,
-		.max_hscl_taps = 8,
-		.max_vscl_taps = 8,
-		.dispclk_ramp_margin_percent = 1,
-		.underscan_factor = 1.11,
-		.min_vblank_lines = 32,
-		.dppclk_delay_subtotal = 46,
-		.dynamic_metadata_vm_enabled = true,
-		.dppclk_delay_scl_lb_only = 16,
-		.dppclk_delay_scl = 50,
-		.dppclk_delay_cnvc_formatter = 27,
-		.dppclk_delay_cnvc_cursor = 6,
-		.dispclk_delay_subtotal = 119,
-		.dcfclk_cstate_latency = 5.2, // SRExitTime
-		.max_inter_dcn_tile_repeaters = 8,
-		.max_num_hdmi_frl_outputs = 1,
-		.odm_combine_4to1_supported = true,
-
-		.xfc_supported = false,
-		.xfc_fill_bw_overhead_percent = 10.0,
-		.xfc_fill_constant_bytes = 0,
-		.gfx7_compat_tiling_supported = 0,
-		.number_of_cursors = 1,
-};
-
-struct _vcs_dpi_soc_bounding_box_st dcn3_02_soc = {
-		.clock_limits = {
-				{
-						.state = 0,
-						.dispclk_mhz = 562.0,
-						.dppclk_mhz = 300.0,
-						.phyclk_mhz = 300.0,
-						.phyclk_d18_mhz = 667.0,
-						.dscclk_mhz = 405.6,
-				},
-		},
-
-		.min_dcfclk = 500.0, /* TODO: set this to actual min DCFCLK */
-		.num_states = 1,
-		.sr_exit_time_us = 26.5,
-		.sr_enter_plus_exit_time_us = 31,
-		.urgent_latency_us = 4.0,
-		.urgent_latency_pixel_data_only_us = 4.0,
-		.urgent_latency_pixel_mixed_with_vm_data_us = 4.0,
-		.urgent_latency_vm_data_only_us = 4.0,
-		.urgent_out_of_order_return_per_channel_pixel_only_bytes = 4096,
-		.urgent_out_of_order_return_per_channel_pixel_and_vm_bytes = 4096,
-		.urgent_out_of_order_return_per_channel_vm_only_bytes = 4096,
-		.pct_ideal_dram_sdp_bw_after_urgent_pixel_only = 80.0,
-		.pct_ideal_dram_sdp_bw_after_urgent_pixel_and_vm = 60.0,
-		.pct_ideal_dram_sdp_bw_after_urgent_vm_only = 40.0,
-		.max_avg_sdp_bw_use_normal_percent = 60.0,
-		.max_avg_dram_bw_use_normal_percent = 40.0,
-		.writeback_latency_us = 12.0,
-		.max_request_size_bytes = 256,
-		.fabric_datapath_to_dcn_data_return_bytes = 64,
-		.dcn_downspread_percent = 0.5,
-		.downspread_percent = 0.38,
-		.dram_page_open_time_ns = 50.0,
-		.dram_rw_turnaround_time_ns = 17.5,
-		.dram_return_buffer_per_channel_bytes = 8192,
-		.round_trip_ping_latency_dcfclk_cycles = 156,
-		.urgent_out_of_order_return_per_channel_bytes = 4096,
-		.channel_interleave_bytes = 256,
-		.num_banks = 8,
-		.gpuvm_min_page_size_bytes = 4096,
-		.hostvm_min_page_size_bytes = 4096,
-		.dram_clock_change_latency_us = 404,
-		.dummy_pstate_latency_us = 5,
-		.writeback_dram_clock_change_latency_us = 23.0,
-		.return_bus_width_bytes = 64,
-		.dispclk_dppclk_vco_speed_mhz = 3650,
-		.xfc_bus_transport_time_us = 20,      // ?
-		.xfc_xbuf_latency_tolerance_us = 4,  // ?
-		.use_urgent_burst_bw = 1,            // ?
-		.do_urgent_latency_adjustment = true,
-		.urgent_latency_adjustment_fabric_clock_component_us = 1.0,
-		.urgent_latency_adjustment_fabric_clock_reference_mhz = 1000,
-};
-
-static const struct dc_debug_options debug_defaults_drv = {
-		.disable_dmcu = true,
-		.force_abm_enable = false,
-		.timing_trace = false,
-		.clock_trace = true,
-		.disable_pplib_clock_request = true,
-		.pipe_split_policy = MPC_SPLIT_AVOID_MULT_DISP,
-		.force_single_disp_pipe_split = false,
-		.disable_dcc = DCC_ENABLE,
-		.vsr_support = true,
-		.performance_trace = false,
-		.max_downscale_src_width = 7680,/*upto 8K*/
-		.disable_pplib_wm_range = false,
-		.scl_reset_length10 = true,
-		.sanity_checks = false,
-		.underflow_assert_delay_us = 0xFFFFFFFF,
-		.dwb_fi_phase = -1, // -1 = disable,
-		.dmub_command_table = true,
-		.use_max_lb = true
-};
-
-static const struct dc_debug_options debug_defaults_diags = {
-		.disable_dmcu = true,
-		.force_abm_enable = false,
-		.timing_trace = true,
-		.clock_trace = true,
-		.disable_dpp_power_gate = true,
-		.disable_hubp_power_gate = true,
-		.disable_clock_gate = true,
-		.disable_pplib_clock_request = true,
-		.disable_pplib_wm_range = true,
-		.disable_stutter = false,
-		.scl_reset_length10 = true,
-		.dwb_fi_phase = -1, // -1 = disable
-		.dmub_command_table = true,
-		.enable_tri_buf = true,
-		.disable_psr = true,
-		.use_max_lb = true
-};
-
-enum dcn302_clk_src_array_id {
-	DCN302_CLK_SRC_PLL0,
-	DCN302_CLK_SRC_PLL1,
-	DCN302_CLK_SRC_PLL2,
-	DCN302_CLK_SRC_PLL3,
-	DCN302_CLK_SRC_PLL4,
-	DCN302_CLK_SRC_TOTAL
-};
-
-static const struct resource_caps res_cap_dcn302 = {
-		.num_timing_generator = 5,
-		.num_opp = 5,
-		.num_video_plane = 5,
-		.num_audio = 5,
-		.num_stream_encoder = 5,
-		.num_dwb = 1,
-		.num_ddc = 5,
-		.num_vmid = 16,
-		.num_mpc_3dlut = 2,
-		.num_dsc = 5,
-};
-
-static const struct dc_plane_cap plane_cap = {
-		.type = DC_PLANE_TYPE_DCN_UNIVERSAL,
-		.blends_with_above = true,
-		.blends_with_below = true,
-		.per_pixel_alpha = true,
-		.pixel_format_support = {
-				.argb8888 = true,
-				.nv12 = true,
-				.fp16 = true,
-				.p010 = false,
-				.ayuv = false,
-		},
-		.max_upscale_factor = {
-				.argb8888 = 16000,
-				.nv12 = 16000,
-				.fp16 = 16000
-		},
-		/* 6:1 downscaling ratio: 1000/6 = 166.666 */
-		.max_downscale_factor = {
-				.argb8888 = 167,
-				.nv12 = 167,
-				.fp16 = 167
-		},
-		16,
-		16
-};
-
-/* NBIO */
-#define NBIO_BASE_INNER(seg) \
-		NBIO_BASE__INST0_SEG ## seg
-
-#define NBIO_BASE(seg) \
-		NBIO_BASE_INNER(seg)
-
-#define NBIO_SR(reg_name)\
-		.reg_name = NBIO_BASE(mm ## reg_name ## _BASE_IDX) + \
-		mm ## reg_name
-
-/* DCN */
-#undef BASE_INNER
-#define BASE_INNER(seg) DCN_BASE__INST0_SEG ## seg
-
-#define BASE(seg) BASE_INNER(seg)
-
-#define SR(reg_name)\
-		.reg_name = BASE(mm ## reg_name ## _BASE_IDX) + mm ## reg_name
-
-#define SF(reg_name, field_name, post_fix)\
-		.field_name = reg_name ## __ ## field_name ## post_fix
-
-#define SRI(reg_name, block, id)\
-		.reg_name = BASE(mm ## block ## id ## _ ## reg_name ## _BASE_IDX) + mm ## block ## id ## _ ## reg_name
-
-#define SRI2(reg_name, block, id)\
-		.reg_name = BASE(mm ## reg_name ## _BASE_IDX) + mm ## reg_name
-
-#define SRII(reg_name, block, id)\
-		.reg_name[id] = BASE(mm ## block ## id ## _ ## reg_name ## _BASE_IDX) + \
-		mm ## block ## id ## _ ## reg_name
-
-#define DCCG_SRII(reg_name, block, id)\
-		.block ## _ ## reg_name[id] = BASE(mm ## block ## id ## _ ## reg_name ## _BASE_IDX) + \
-		mm ## block ## id ## _ ## reg_name
-
-#define VUPDATE_SRII(reg_name, block, id)\
-		.reg_name[id] = BASE(mm ## reg_name ## _ ## block ## id ## _BASE_IDX) + \
-		mm ## reg_name ## _ ## block ## id
-
-#define SRII_DWB(reg_name, temp_name, block, id)\
-		.reg_name[id] = BASE(mm ## block ## id ## _ ## temp_name ## _BASE_IDX) + \
-		mm ## block ## id ## _ ## temp_name
-
-#define SRII_MPC_RMU(reg_name, block, id)\
-		.RMU##_##reg_name[id] = BASE(mm ## block ## id ## _ ## reg_name ## _BASE_IDX) + \
-		mm ## block ## id ## _ ## reg_name
-
-static const struct dcn_hubbub_registers hubbub_reg = {
-		HUBBUB_REG_LIST_DCN30(0)
-};
-
-static const struct dcn_hubbub_shift hubbub_shift = {
-		HUBBUB_MASK_SH_LIST_DCN30(__SHIFT)
-};
-
-static const struct dcn_hubbub_mask hubbub_mask = {
-		HUBBUB_MASK_SH_LIST_DCN30(_MASK)
-};
-
-#define vmid_regs(id)\
-		[id] = { DCN20_VMID_REG_LIST(id) }
-
-static const struct dcn_vmid_registers vmid_regs[] = {
-		vmid_regs(0),
-		vmid_regs(1),
-		vmid_regs(2),
-		vmid_regs(3),
-		vmid_regs(4),
-		vmid_regs(5),
-		vmid_regs(6),
-		vmid_regs(7),
-		vmid_regs(8),
-		vmid_regs(9),
-		vmid_regs(10),
-		vmid_regs(11),
-		vmid_regs(12),
-		vmid_regs(13),
-		vmid_regs(14),
-		vmid_regs(15)
-};
-
-static const struct dcn20_vmid_shift vmid_shifts = {
-		DCN20_VMID_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dcn20_vmid_mask vmid_masks = {
-		DCN20_VMID_MASK_SH_LIST(_MASK)
-};
-
-static struct hubbub *dcn302_hubbub_create(struct dc_context *ctx)
-{
-	int i;
-
-	struct dcn20_hubbub *hubbub3 = kzalloc(sizeof(struct dcn20_hubbub), GFP_KERNEL);
-
-	if (!hubbub3)
-		return NULL;
-
-	hubbub3_construct(hubbub3, ctx, &hubbub_reg, &hubbub_shift, &hubbub_mask);
-
-	for (i = 0; i < res_cap_dcn302.num_vmid; i++) {
-		struct dcn20_vmid *vmid = &hubbub3->vmid[i];
-
-		vmid->ctx = ctx;
-
-		vmid->regs = &vmid_regs[i];
-		vmid->shifts = &vmid_shifts;
-		vmid->masks = &vmid_masks;
-	}
-
-	return &hubbub3->base;
-}
-
-#define vpg_regs(id)\
-		[id] = { VPG_DCN3_REG_LIST(id) }
-
-static const struct dcn30_vpg_registers vpg_regs[] = {
-		vpg_regs(0),
-		vpg_regs(1),
-		vpg_regs(2),
-		vpg_regs(3),
-		vpg_regs(4),
-		vpg_regs(5)
-};
-
-static const struct dcn30_vpg_shift vpg_shift = {
-		DCN3_VPG_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dcn30_vpg_mask vpg_mask = {
-		DCN3_VPG_MASK_SH_LIST(_MASK)
-};
-
-static struct vpg *dcn302_vpg_create(struct dc_context *ctx, uint32_t inst)
-{
-	struct dcn30_vpg *vpg3 = kzalloc(sizeof(struct dcn30_vpg), GFP_KERNEL);
-
-	if (!vpg3)
-		return NULL;
-
-	vpg3_construct(vpg3, ctx, inst, &vpg_regs[inst], &vpg_shift, &vpg_mask);
-
-	return &vpg3->base;
-}
-
-#define afmt_regs(id)\
-		[id] = { AFMT_DCN3_REG_LIST(id) }
-
-static const struct dcn30_afmt_registers afmt_regs[] = {
-		afmt_regs(0),
-		afmt_regs(1),
-		afmt_regs(2),
-		afmt_regs(3),
-		afmt_regs(4),
-		afmt_regs(5)
-};
-
-static const struct dcn30_afmt_shift afmt_shift = {
-		DCN3_AFMT_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dcn30_afmt_mask afmt_mask = {
-		DCN3_AFMT_MASK_SH_LIST(_MASK)
-};
-
-static struct afmt *dcn302_afmt_create(struct dc_context *ctx, uint32_t inst)
-{
-	struct dcn30_afmt *afmt3 = kzalloc(sizeof(struct dcn30_afmt), GFP_KERNEL);
-
-	if (!afmt3)
-		return NULL;
-
-	afmt3_construct(afmt3, ctx, inst, &afmt_regs[inst], &afmt_shift, &afmt_mask);
-
-	return &afmt3->base;
-}
-
-#define audio_regs(id)\
-		[id] = { AUD_COMMON_REG_LIST(id) }
-
-static const struct dce_audio_registers audio_regs[] = {
-		audio_regs(0),
-		audio_regs(1),
-		audio_regs(2),
-		audio_regs(3),
-		audio_regs(4),
-		audio_regs(5),
-		audio_regs(6)
-};
-
-#define DCE120_AUD_COMMON_MASK_SH_LIST(mask_sh)\
-		SF(AZF0ENDPOINT0_AZALIA_F0_CODEC_ENDPOINT_INDEX, AZALIA_ENDPOINT_REG_INDEX, mask_sh),\
-		SF(AZF0ENDPOINT0_AZALIA_F0_CODEC_ENDPOINT_DATA, AZALIA_ENDPOINT_REG_DATA, mask_sh),\
-		AUD_COMMON_MASK_SH_LIST_BASE(mask_sh)
-
-static const struct dce_audio_shift audio_shift = {
-		DCE120_AUD_COMMON_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dce_audio_mask audio_mask = {
-		DCE120_AUD_COMMON_MASK_SH_LIST(_MASK)
-};
-
-static struct audio *dcn302_create_audio(struct dc_context *ctx, unsigned int inst)
-{
-	return dce_audio_create(ctx, inst, &audio_regs[inst], &audio_shift, &audio_mask);
-}
-
-#define stream_enc_regs(id)\
-		[id] = { SE_DCN3_REG_LIST(id) }
-
-static const struct dcn10_stream_enc_registers stream_enc_regs[] = {
-		stream_enc_regs(0),
-		stream_enc_regs(1),
-		stream_enc_regs(2),
-		stream_enc_regs(3),
-		stream_enc_regs(4)
-};
-
-static const struct dcn10_stream_encoder_shift se_shift = {
-		SE_COMMON_MASK_SH_LIST_DCN30(__SHIFT)
-};
-
-static const struct dcn10_stream_encoder_mask se_mask = {
-		SE_COMMON_MASK_SH_LIST_DCN30(_MASK)
-};
-
-static struct stream_encoder *dcn302_stream_encoder_create(enum engine_id eng_id, struct dc_context *ctx)
-{
-	struct dcn10_stream_encoder *enc1;
-	struct vpg *vpg;
-	struct afmt *afmt;
-	int vpg_inst;
-	int afmt_inst;
-
-	/* Mapping of VPG, AFMT, DME register blocks to DIO block instance */
-	if (eng_id <= ENGINE_ID_DIGE) {
-		vpg_inst = eng_id;
-		afmt_inst = eng_id;
-	} else
-		return NULL;
-
-	enc1 = kzalloc(sizeof(struct dcn10_stream_encoder), GFP_KERNEL);
-	vpg = dcn302_vpg_create(ctx, vpg_inst);
-	afmt = dcn302_afmt_create(ctx, afmt_inst);
-
-	if (!enc1 || !vpg || !afmt) {
-		kfree(enc1);
-		kfree(vpg);
-		kfree(afmt);
-		return NULL;
-	}
-
-	dcn30_dio_stream_encoder_construct(enc1, ctx, ctx->dc_bios, eng_id, vpg, afmt, &stream_enc_regs[eng_id],
-			&se_shift, &se_mask);
-
-	return &enc1->base;
-}
-
-#define clk_src_regs(index, pllid)\
-		[index] = { CS_COMMON_REG_LIST_DCN3_02(index, pllid) }
-
-static const struct dce110_clk_src_regs clk_src_regs[] = {
-		clk_src_regs(0, A),
-		clk_src_regs(1, B),
-		clk_src_regs(2, C),
-		clk_src_regs(3, D),
-		clk_src_regs(4, E)
-};
-
-static const struct dce110_clk_src_shift cs_shift = {
-		CS_COMMON_MASK_SH_LIST_DCN2_0(__SHIFT)
-};
-
-static const struct dce110_clk_src_mask cs_mask = {
-		CS_COMMON_MASK_SH_LIST_DCN2_0(_MASK)
-};
-
-static struct clock_source *dcn302_clock_source_create(struct dc_context *ctx, struct dc_bios *bios,
-		enum clock_source_id id, const struct dce110_clk_src_regs *regs, bool dp_clk_src)
-{
-	struct dce110_clk_src *clk_src = kzalloc(sizeof(struct dce110_clk_src), GFP_KERNEL);
-
-	if (!clk_src)
-		return NULL;
-
-	if (dcn3_clk_src_construct(clk_src, ctx, bios, id, regs, &cs_shift, &cs_mask)) {
-		clk_src->base.dp_clk_src = dp_clk_src;
-		return &clk_src->base;
-	}
-
-	BREAK_TO_DEBUGGER();
-	return NULL;
-}
-
-static const struct dce_hwseq_registers hwseq_reg = {
-		HWSEQ_DCN302_REG_LIST()
-};
-
-static const struct dce_hwseq_shift hwseq_shift = {
-		HWSEQ_DCN302_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dce_hwseq_mask hwseq_mask = {
-		HWSEQ_DCN302_MASK_SH_LIST(_MASK)
-};
-
-static struct dce_hwseq *dcn302_hwseq_create(struct dc_context *ctx)
-{
-	struct dce_hwseq *hws = kzalloc(sizeof(struct dce_hwseq), GFP_KERNEL);
-
-	if (hws) {
-		hws->ctx = ctx;
-		hws->regs = &hwseq_reg;
-		hws->shifts = &hwseq_shift;
-		hws->masks = &hwseq_mask;
-	}
-	return hws;
-}
-
-#define hubp_regs(id)\
-		[id] = { HUBP_REG_LIST_DCN30(id) }
-
-static const struct dcn_hubp2_registers hubp_regs[] = {
-		hubp_regs(0),
-		hubp_regs(1),
-		hubp_regs(2),
-		hubp_regs(3),
-		hubp_regs(4)
-};
-
-static const struct dcn_hubp2_shift hubp_shift = {
-		HUBP_MASK_SH_LIST_DCN30(__SHIFT)
-};
-
-static const struct dcn_hubp2_mask hubp_mask = {
-		HUBP_MASK_SH_LIST_DCN30(_MASK)
-};
-
-static struct hubp *dcn302_hubp_create(struct dc_context *ctx, uint32_t inst)
-{
-	struct dcn20_hubp *hubp2 = kzalloc(sizeof(struct dcn20_hubp), GFP_KERNEL);
-
-	if (!hubp2)
-		return NULL;
-
-	if (hubp3_construct(hubp2, ctx, inst, &hubp_regs[inst], &hubp_shift, &hubp_mask))
-		return &hubp2->base;
-
-	BREAK_TO_DEBUGGER();
-	kfree(hubp2);
-	return NULL;
-}
-
-#define dpp_regs(id)\
-		[id] = { DPP_REG_LIST_DCN30(id) }
-
-static const struct dcn3_dpp_registers dpp_regs[] = {
-		dpp_regs(0),
-		dpp_regs(1),
-		dpp_regs(2),
-		dpp_regs(3),
-		dpp_regs(4)
-};
-
-static const struct dcn3_dpp_shift tf_shift = {
-		DPP_REG_LIST_SH_MASK_DCN30(__SHIFT)
-};
-
-static const struct dcn3_dpp_mask tf_mask = {
-		DPP_REG_LIST_SH_MASK_DCN30(_MASK)
-};
-
-static struct dpp *dcn302_dpp_create(struct dc_context *ctx, uint32_t inst)
-{
-	struct dcn3_dpp *dpp = kzalloc(sizeof(struct dcn3_dpp), GFP_KERNEL);
-
-	if (!dpp)
-		return NULL;
-
-	if (dpp3_construct(dpp, ctx, inst, &dpp_regs[inst], &tf_shift, &tf_mask))
-		return &dpp->base;
-
-	BREAK_TO_DEBUGGER();
-	kfree(dpp);
-	return NULL;
-}
-
-#define opp_regs(id)\
-		[id] = { OPP_REG_LIST_DCN30(id) }
-
-static const struct dcn20_opp_registers opp_regs[] = {
-		opp_regs(0),
-		opp_regs(1),
-		opp_regs(2),
-		opp_regs(3),
-		opp_regs(4)
-};
-
-static const struct dcn20_opp_shift opp_shift = {
-		OPP_MASK_SH_LIST_DCN20(__SHIFT)
-};
-
-static const struct dcn20_opp_mask opp_mask = {
-		OPP_MASK_SH_LIST_DCN20(_MASK)
-};
-
-static struct output_pixel_processor *dcn302_opp_create(struct dc_context *ctx, uint32_t inst)
-{
-	struct dcn20_opp *opp = kzalloc(sizeof(struct dcn20_opp), GFP_KERNEL);
-
-	if (!opp) {
-		BREAK_TO_DEBUGGER();
-		return NULL;
-	}
-
-	dcn20_opp_construct(opp, ctx, inst, &opp_regs[inst], &opp_shift, &opp_mask);
-	return &opp->base;
-}
-
-#define optc_regs(id)\
-		[id] = { OPTC_COMMON_REG_LIST_DCN3_0(id) }
-
-static const struct dcn_optc_registers optc_regs[] = {
-		optc_regs(0),
-		optc_regs(1),
-		optc_regs(2),
-		optc_regs(3),
-		optc_regs(4)
-};
-
-static const struct dcn_optc_shift optc_shift = {
-		OPTC_COMMON_MASK_SH_LIST_DCN30(__SHIFT)
-};
-
-static const struct dcn_optc_mask optc_mask = {
-		OPTC_COMMON_MASK_SH_LIST_DCN30(_MASK)
-};
-
-static struct timing_generator *dcn302_timing_generator_create(struct dc_context *ctx, uint32_t instance)
-{
-	struct optc *tgn10 = kzalloc(sizeof(struct optc), GFP_KERNEL);
-
-	if (!tgn10)
-		return NULL;
-
-	tgn10->base.inst = instance;
-	tgn10->base.ctx = ctx;
-
-	tgn10->tg_regs = &optc_regs[instance];
-	tgn10->tg_shift = &optc_shift;
-	tgn10->tg_mask = &optc_mask;
-
-	dcn30_timing_generator_init(tgn10);
-
-	return &tgn10->base;
-}
-
-static const struct dcn30_mpc_registers mpc_regs = {
-		MPC_REG_LIST_DCN3_0(0),
-		MPC_REG_LIST_DCN3_0(1),
-		MPC_REG_LIST_DCN3_0(2),
-		MPC_REG_LIST_DCN3_0(3),
-		MPC_REG_LIST_DCN3_0(4),
-		MPC_OUT_MUX_REG_LIST_DCN3_0(0),
-		MPC_OUT_MUX_REG_LIST_DCN3_0(1),
-		MPC_OUT_MUX_REG_LIST_DCN3_0(2),
-		MPC_OUT_MUX_REG_LIST_DCN3_0(3),
-		MPC_OUT_MUX_REG_LIST_DCN3_0(4),
-		MPC_RMU_GLOBAL_REG_LIST_DCN3AG,
-		MPC_RMU_REG_LIST_DCN3AG(0),
-		MPC_RMU_REG_LIST_DCN3AG(1),
-		MPC_RMU_REG_LIST_DCN3AG(2),
-		MPC_DWB_MUX_REG_LIST_DCN3_0(0),
-};
-
-static const struct dcn30_mpc_shift mpc_shift = {
-		MPC_COMMON_MASK_SH_LIST_DCN30(__SHIFT)
-};
-
-static const struct dcn30_mpc_mask mpc_mask = {
-		MPC_COMMON_MASK_SH_LIST_DCN30(_MASK)
-};
-
-static struct mpc *dcn302_mpc_create(struct dc_context *ctx, int num_mpcc, int num_rmu)
-{
-	struct dcn30_mpc *mpc30 = kzalloc(sizeof(struct dcn30_mpc), GFP_KERNEL);
-
-	if (!mpc30)
-		return NULL;
-
-	dcn30_mpc_construct(mpc30, ctx, &mpc_regs, &mpc_shift, &mpc_mask, num_mpcc, num_rmu);
-
-	return &mpc30->base;
-}
-
-#define dsc_regsDCN20(id)\
-[id] = { DSC_REG_LIST_DCN20(id) }
-
-static const struct dcn20_dsc_registers dsc_regs[] = {
-		dsc_regsDCN20(0),
-		dsc_regsDCN20(1),
-		dsc_regsDCN20(2),
-		dsc_regsDCN20(3),
-		dsc_regsDCN20(4)
-};
-
-static const struct dcn20_dsc_shift dsc_shift = {
-		DSC_REG_LIST_SH_MASK_DCN20(__SHIFT)
-};
-
-static const struct dcn20_dsc_mask dsc_mask = {
-		DSC_REG_LIST_SH_MASK_DCN20(_MASK)
-};
-
-static struct display_stream_compressor *dcn302_dsc_create(struct dc_context *ctx, uint32_t inst)
-{
-	struct dcn20_dsc *dsc = kzalloc(sizeof(struct dcn20_dsc), GFP_KERNEL);
-
-	if (!dsc) {
-		BREAK_TO_DEBUGGER();
-		return NULL;
-	}
-
-	dsc2_construct(dsc, ctx, inst, &dsc_regs[inst], &dsc_shift, &dsc_mask);
-	return &dsc->base;
-}
-
-#define dwbc_regs_dcn3(id)\
-[id] = { DWBC_COMMON_REG_LIST_DCN30(id) }
-
-static const struct dcn30_dwbc_registers dwbc30_regs[] = {
-		dwbc_regs_dcn3(0)
-};
-
-static const struct dcn30_dwbc_shift dwbc30_shift = {
-		DWBC_COMMON_MASK_SH_LIST_DCN30(__SHIFT)
-};
-
-static const struct dcn30_dwbc_mask dwbc30_mask = {
-		DWBC_COMMON_MASK_SH_LIST_DCN30(_MASK)
-};
-
-static bool dcn302_dwbc_create(struct dc_context *ctx, struct resource_pool *pool)
-{
-	int i;
-	uint32_t pipe_count = pool->res_cap->num_dwb;
-
-	for (i = 0; i < pipe_count; i++) {
-		struct dcn30_dwbc *dwbc30 = kzalloc(sizeof(struct dcn30_dwbc), GFP_KERNEL);
-
-		if (!dwbc30) {
-			dm_error("DC: failed to create dwbc30!\n");
-			return false;
-		}
-
-		dcn30_dwbc_construct(dwbc30, ctx, &dwbc30_regs[i], &dwbc30_shift, &dwbc30_mask, i);
-
-		pool->dwbc[i] = &dwbc30->base;
-	}
-	return true;
-}
-
-#define mcif_wb_regs_dcn3(id)\
-[id] = { MCIF_WB_COMMON_REG_LIST_DCN30(id) }
-
-static const struct dcn30_mmhubbub_registers mcif_wb30_regs[] = {
-		mcif_wb_regs_dcn3(0)
-};
-
-static const struct dcn30_mmhubbub_shift mcif_wb30_shift = {
-		MCIF_WB_COMMON_MASK_SH_LIST_DCN30(__SHIFT)
-};
-
-static const struct dcn30_mmhubbub_mask mcif_wb30_mask = {
-		MCIF_WB_COMMON_MASK_SH_LIST_DCN30(_MASK)
-};
-
-static bool dcn302_mmhubbub_create(struct dc_context *ctx, struct resource_pool *pool)
-{
-	int i;
-	uint32_t pipe_count = pool->res_cap->num_dwb;
-
-	for (i = 0; i < pipe_count; i++) {
-		struct dcn30_mmhubbub *mcif_wb30 = kzalloc(sizeof(struct dcn30_mmhubbub), GFP_KERNEL);
-
-		if (!mcif_wb30) {
-			dm_error("DC: failed to create mcif_wb30!\n");
-			return false;
-		}
-
-		dcn30_mmhubbub_construct(mcif_wb30, ctx, &mcif_wb30_regs[i], &mcif_wb30_shift, &mcif_wb30_mask, i);
-
-		pool->mcif_wb[i] = &mcif_wb30->base;
-	}
-	return true;
-}
-
-#define aux_engine_regs(id)\
-[id] = {\
-		AUX_COMMON_REG_LIST0(id), \
-		.AUXN_IMPCAL = 0, \
-		.AUXP_IMPCAL = 0, \
-		.AUX_RESET_MASK = DP_AUX0_AUX_CONTROL__AUX_RESET_MASK, \
-}
-
-static const struct dce110_aux_registers aux_engine_regs[] = {
-		aux_engine_regs(0),
-		aux_engine_regs(1),
-		aux_engine_regs(2),
-		aux_engine_regs(3),
-		aux_engine_regs(4)
-};
-
-static const struct dce110_aux_registers_shift aux_shift = {
-		DCN_AUX_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dce110_aux_registers_mask aux_mask = {
-		DCN_AUX_MASK_SH_LIST(_MASK)
-};
-
-static struct dce_aux *dcn302_aux_engine_create(struct dc_context *ctx, uint32_t inst)
-{
-	struct aux_engine_dce110 *aux_engine = kzalloc(sizeof(struct aux_engine_dce110), GFP_KERNEL);
-
-	if (!aux_engine)
-		return NULL;
-
-	dce110_aux_engine_construct(aux_engine, ctx, inst, SW_AUX_TIMEOUT_PERIOD_MULTIPLIER * AUX_TIMEOUT_PERIOD,
-			&aux_engine_regs[inst], &aux_mask, &aux_shift, ctx->dc->caps.extended_aux_timeout_support);
-
-	return &aux_engine->base;
-}
-
-#define i2c_inst_regs(id) { I2C_HW_ENGINE_COMMON_REG_LIST(id) }
-
-static const struct dce_i2c_registers i2c_hw_regs[] = {
-		i2c_inst_regs(1),
-		i2c_inst_regs(2),
-		i2c_inst_regs(3),
-		i2c_inst_regs(4),
-		i2c_inst_regs(5)
-};
-
-static const struct dce_i2c_shift i2c_shifts = {
-		I2C_COMMON_MASK_SH_LIST_DCN2(__SHIFT)
-};
-
-static const struct dce_i2c_mask i2c_masks = {
-		I2C_COMMON_MASK_SH_LIST_DCN2(_MASK)
-};
-
-static struct dce_i2c_hw *dcn302_i2c_hw_create(struct dc_context *ctx, uint32_t inst)
-{
-	struct dce_i2c_hw *dce_i2c_hw = kzalloc(sizeof(struct dce_i2c_hw), GFP_KERNEL);
-
-	if (!dce_i2c_hw)
-		return NULL;
-
-	dcn2_i2c_hw_construct(dce_i2c_hw, ctx, inst, &i2c_hw_regs[inst], &i2c_shifts, &i2c_masks);
-
-	return dce_i2c_hw;
-}
-
-static const struct encoder_feature_support link_enc_feature = {
-		.max_hdmi_deep_color = COLOR_DEPTH_121212,
-		.max_hdmi_pixel_clock = 600000,
-		.hdmi_ycbcr420_supported = true,
-		.dp_ycbcr420_supported = true,
-		.fec_supported = true,
-		.flags.bits.IS_HBR2_CAPABLE = true,
-		.flags.bits.IS_HBR3_CAPABLE = true,
-		.flags.bits.IS_TPS3_CAPABLE = true,
-		.flags.bits.IS_TPS4_CAPABLE = true
-};
-
-#define link_regs(id, phyid)\
-		[id] = {\
-				LE_DCN3_REG_LIST(id), \
-				UNIPHY_DCN2_REG_LIST(phyid), \
-				DPCS_DCN2_REG_LIST(id), \
-				SRI(DP_DPHY_INTERNAL_CTRL, DP, id) \
-		}
-
-static const struct dcn10_link_enc_registers link_enc_regs[] = {
-		link_regs(0, A),
-		link_regs(1, B),
-		link_regs(2, C),
-		link_regs(3, D),
-		link_regs(4, E)
-};
-
-static const struct dcn10_link_enc_shift le_shift = {
-		LINK_ENCODER_MASK_SH_LIST_DCN30(__SHIFT),
-		DPCS_DCN2_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dcn10_link_enc_mask le_mask = {
-		LINK_ENCODER_MASK_SH_LIST_DCN30(_MASK),
-		DPCS_DCN2_MASK_SH_LIST(_MASK)
-};
-
-#define aux_regs(id)\
-		[id] = { DCN2_AUX_REG_LIST(id) }
-
-static const struct dcn10_link_enc_aux_registers link_enc_aux_regs[] = {
-		aux_regs(0),
-		aux_regs(1),
-		aux_regs(2),
-		aux_regs(3),
-		aux_regs(4)
-};
-
-#define hpd_regs(id)\
-		[id] = { HPD_REG_LIST(id) }
-
-static const struct dcn10_link_enc_hpd_registers link_enc_hpd_regs[] = {
-		hpd_regs(0),
-		hpd_regs(1),
-		hpd_regs(2),
-		hpd_regs(3),
-		hpd_regs(4)
-};
-
-static struct link_encoder *dcn302_link_encoder_create(const struct encoder_init_data *enc_init_data)
-{
-	struct dcn20_link_encoder *enc20 = kzalloc(sizeof(struct dcn20_link_encoder), GFP_KERNEL);
-
-	if (!enc20)
-		return NULL;
-
-	dcn30_link_encoder_construct(enc20, enc_init_data, &link_enc_feature,
-			&link_enc_regs[enc_init_data->transmitter], &link_enc_aux_regs[enc_init_data->channel - 1],
-			&link_enc_hpd_regs[enc_init_data->hpd_source], &le_shift, &le_mask);
-
-	return &enc20->enc10.base;
-}
-
-static const struct dce_panel_cntl_registers panel_cntl_regs[] = {
-		{ DCN_PANEL_CNTL_REG_LIST() }
-};
-
-static const struct dce_panel_cntl_shift panel_cntl_shift = {
-		DCE_PANEL_CNTL_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dce_panel_cntl_mask panel_cntl_mask = {
-		DCE_PANEL_CNTL_MASK_SH_LIST(_MASK)
-};
-
-static struct panel_cntl *dcn302_panel_cntl_create(const struct panel_cntl_init_data *init_data)
-{
-	struct dce_panel_cntl *panel_cntl = kzalloc(sizeof(struct dce_panel_cntl), GFP_KERNEL);
-
-	if (!panel_cntl)
-		return NULL;
-
-	dce_panel_cntl_construct(panel_cntl, init_data, &panel_cntl_regs[init_data->inst],
-			&panel_cntl_shift, &panel_cntl_mask);
-
-	return &panel_cntl->base;
-}
-
-static void read_dce_straps(struct dc_context *ctx, struct resource_straps *straps)
-{
-	generic_reg_get(ctx, mmDC_PINSTRAPS + BASE(mmDC_PINSTRAPS_BASE_IDX),
-			FN(DC_PINSTRAPS, DC_PINSTRAPS_AUDIO), &straps->dc_pinstraps_audio);
-}
-
-static const struct resource_create_funcs res_create_funcs = {
-		.read_dce_straps = read_dce_straps,
-		.create_audio = dcn302_create_audio,
-		.create_stream_encoder = dcn302_stream_encoder_create,
-		.create_hwseq = dcn302_hwseq_create,
-};
-
-static const struct resource_create_funcs res_create_maximus_funcs = {
-		.read_dce_straps = NULL,
-		.create_audio = NULL,
-		.create_stream_encoder = NULL,
-		.create_hwseq = dcn302_hwseq_create,
-};
-
-static bool is_soc_bounding_box_valid(struct dc *dc)
-{
-	uint32_t hw_internal_rev = dc->ctx->asic_id.hw_internal_rev;
-
-	if (ASICREV_IS_DIMGREY_CAVEFISH_P(hw_internal_rev))
-		return true;
-
-	return false;
-}
-
-static bool init_soc_bounding_box(struct dc *dc,  struct resource_pool *pool)
-{
-	struct _vcs_dpi_soc_bounding_box_st *loaded_bb = &dcn3_02_soc;
-	struct _vcs_dpi_ip_params_st *loaded_ip = &dcn3_02_ip;
-
-	DC_LOGGER_INIT(dc->ctx->logger);
-
-	if (!is_soc_bounding_box_valid(dc)) {
-		DC_LOG_ERROR("%s: not valid soc bounding box\n", __func__);
-		return false;
-	}
-
-	loaded_ip->max_num_otg = pool->pipe_count;
-	loaded_ip->max_num_dpp = pool->pipe_count;
-	loaded_ip->clamp_min_dcfclk = dc->config.clamp_min_dcfclk;
-	dcn20_patch_bounding_box(dc, loaded_bb);
-
-	if (dc->ctx->dc_bios->funcs->get_soc_bb_info) {
-		struct bp_soc_bb_info bb_info = { 0 };
-
-		if (dc->ctx->dc_bios->funcs->get_soc_bb_info(
-			    dc->ctx->dc_bios, &bb_info) == BP_RESULT_OK) {
-			if (bb_info.dram_clock_change_latency_100ns > 0)
-				dcn3_02_soc.dram_clock_change_latency_us =
-					bb_info.dram_clock_change_latency_100ns * 10;
-
-			if (bb_info.dram_sr_enter_exit_latency_100ns > 0)
-				dcn3_02_soc.sr_enter_plus_exit_time_us =
-					bb_info.dram_sr_enter_exit_latency_100ns * 10;
-
-			if (bb_info.dram_sr_exit_latency_100ns > 0)
-				dcn3_02_soc.sr_exit_time_us =
-					bb_info.dram_sr_exit_latency_100ns * 10;
-		}
-	}
-
-	return true;
-}
-
-static void dcn302_resource_destruct(struct resource_pool *pool)
-{
-	unsigned int i;
-
-	for (i = 0; i < pool->stream_enc_count; i++) {
-		if (pool->stream_enc[i] != NULL) {
-			if (pool->stream_enc[i]->vpg != NULL) {
-				kfree(DCN30_VPG_FROM_VPG(pool->stream_enc[i]->vpg));
-				pool->stream_enc[i]->vpg = NULL;
-			}
-			if (pool->stream_enc[i]->afmt != NULL) {
-				kfree(DCN30_AFMT_FROM_AFMT(pool->stream_enc[i]->afmt));
-				pool->stream_enc[i]->afmt = NULL;
-			}
-			kfree(DCN10STRENC_FROM_STRENC(pool->stream_enc[i]));
-			pool->stream_enc[i] = NULL;
-		}
-	}
-
-	for (i = 0; i < pool->res_cap->num_dsc; i++) {
-		if (pool->dscs[i] != NULL)
-			dcn20_dsc_destroy(&pool->dscs[i]);
-	}
-
-	if (pool->mpc != NULL) {
-		kfree(TO_DCN20_MPC(pool->mpc));
-		pool->mpc = NULL;
-	}
-
-	if (pool->hubbub != NULL) {
-		kfree(pool->hubbub);
-		pool->hubbub = NULL;
-	}
-
-	for (i = 0; i < pool->pipe_count; i++) {
-		if (pool->dpps[i] != NULL) {
-			kfree(TO_DCN20_DPP(pool->dpps[i]));
-			pool->dpps[i] = NULL;
-		}
-
-		if (pool->hubps[i] != NULL) {
-			kfree(TO_DCN20_HUBP(pool->hubps[i]));
-			pool->hubps[i] = NULL;
-		}
-
-		if (pool->irqs != NULL)
-			dal_irq_service_destroy(&pool->irqs);
-	}
-
-	for (i = 0; i < pool->res_cap->num_ddc; i++) {
-		if (pool->engines[i] != NULL)
-			dce110_engine_destroy(&pool->engines[i]);
-		if (pool->hw_i2cs[i] != NULL) {
-			kfree(pool->hw_i2cs[i]);
-			pool->hw_i2cs[i] = NULL;
-		}
-		if (pool->sw_i2cs[i] != NULL) {
-			kfree(pool->sw_i2cs[i]);
-			pool->sw_i2cs[i] = NULL;
-		}
-	}
-
-	for (i = 0; i < pool->res_cap->num_opp; i++) {
-		if (pool->opps[i] != NULL)
-			pool->opps[i]->funcs->opp_destroy(&pool->opps[i]);
-	}
-
-	for (i = 0; i < pool->res_cap->num_timing_generator; i++) {
-		if (pool->timing_generators[i] != NULL)	{
-			kfree(DCN10TG_FROM_TG(pool->timing_generators[i]));
-			pool->timing_generators[i] = NULL;
-		}
-	}
-
-	for (i = 0; i < pool->res_cap->num_dwb; i++) {
-		if (pool->dwbc[i] != NULL) {
-			kfree(TO_DCN30_DWBC(pool->dwbc[i]));
-			pool->dwbc[i] = NULL;
-		}
-		if (pool->mcif_wb[i] != NULL) {
-			kfree(TO_DCN30_MMHUBBUB(pool->mcif_wb[i]));
-			pool->mcif_wb[i] = NULL;
-		}
-	}
-
-	for (i = 0; i < pool->audio_count; i++) {
-		if (pool->audios[i])
-			dce_aud_destroy(&pool->audios[i]);
-	}
-
-	for (i = 0; i < pool->clk_src_count; i++) {
-		if (pool->clock_sources[i] != NULL)
-			dcn20_clock_source_destroy(&pool->clock_sources[i]);
-	}
-
-	if (pool->dp_clock_source != NULL)
-		dcn20_clock_source_destroy(&pool->dp_clock_source);
-
-	for (i = 0; i < pool->res_cap->num_mpc_3dlut; i++) {
-		if (pool->mpc_lut[i] != NULL) {
-			dc_3dlut_func_release(pool->mpc_lut[i]);
-			pool->mpc_lut[i] = NULL;
-		}
-		if (pool->mpc_shaper[i] != NULL) {
-			dc_transfer_func_release(pool->mpc_shaper[i]);
-			pool->mpc_shaper[i] = NULL;
-		}
-	}
-
-	for (i = 0; i < pool->pipe_count; i++) {
-		if (pool->multiple_abms[i] != NULL)
-			dce_abm_destroy(&pool->multiple_abms[i]);
-	}
-
-	if (pool->psr != NULL)
-		dmub_psr_destroy(&pool->psr);
-
-	if (pool->dccg != NULL)
-		dcn_dccg_destroy(&pool->dccg);
-}
-
-static void dcn302_destroy_resource_pool(struct resource_pool **pool)
-{
-	dcn302_resource_destruct(*pool);
-	kfree(*pool);
-	*pool = NULL;
-}
-
-static void dcn302_get_optimal_dcfclk_fclk_for_uclk(unsigned int uclk_mts,
-		unsigned int *optimal_dcfclk,
-		unsigned int *optimal_fclk)
-{
-	double bw_from_dram, bw_from_dram1, bw_from_dram2;
-
-	bw_from_dram1 = uclk_mts * dcn3_02_soc.num_chans *
-		dcn3_02_soc.dram_channel_width_bytes * (dcn3_02_soc.max_avg_dram_bw_use_normal_percent / 100);
-	bw_from_dram2 = uclk_mts * dcn3_02_soc.num_chans *
-		dcn3_02_soc.dram_channel_width_bytes * (dcn3_02_soc.max_avg_sdp_bw_use_normal_percent / 100);
-
-	bw_from_dram = (bw_from_dram1 < bw_from_dram2) ? bw_from_dram1 : bw_from_dram2;
-
-	if (optimal_fclk)
-		*optimal_fclk = bw_from_dram /
-		(dcn3_02_soc.fabric_datapath_to_dcn_data_return_bytes * (dcn3_02_soc.max_avg_sdp_bw_use_normal_percent / 100));
-
-	if (optimal_dcfclk)
-		*optimal_dcfclk =  bw_from_dram /
-		(dcn3_02_soc.return_bus_width_bytes * (dcn3_02_soc.max_avg_sdp_bw_use_normal_percent / 100));
-}
-
-void dcn302_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params)
-{
-	unsigned int i, j;
-	unsigned int num_states = 0;
-
-	unsigned int dcfclk_mhz[DC__VOLTAGE_STATES] = {0};
-	unsigned int dram_speed_mts[DC__VOLTAGE_STATES] = {0};
-	unsigned int optimal_uclk_for_dcfclk_sta_targets[DC__VOLTAGE_STATES] = {0};
-	unsigned int optimal_dcfclk_for_uclk[DC__VOLTAGE_STATES] = {0};
-
-	unsigned int dcfclk_sta_targets[DC__VOLTAGE_STATES] = {694, 875, 1000, 1200};
-	unsigned int num_dcfclk_sta_targets = 4;
-	unsigned int num_uclk_states;
-
-
-	if (dc->ctx->dc_bios->vram_info.num_chans)
-		dcn3_02_soc.num_chans = dc->ctx->dc_bios->vram_info.num_chans;
-
-	if (dc->ctx->dc_bios->vram_info.dram_channel_width_bytes)
-		dcn3_02_soc.dram_channel_width_bytes = dc->ctx->dc_bios->vram_info.dram_channel_width_bytes;
-
-	dcn3_02_soc.dispclk_dppclk_vco_speed_mhz = dc->clk_mgr->dentist_vco_freq_khz / 1000.0;
-	dc->dml.soc.dispclk_dppclk_vco_speed_mhz = dc->clk_mgr->dentist_vco_freq_khz / 1000.0;
-
-	if (bw_params->clk_table.entries[0].memclk_mhz) {
-		int max_dcfclk_mhz = 0, max_dispclk_mhz = 0, max_dppclk_mhz = 0, max_phyclk_mhz = 0;
-
-		for (i = 0; i < MAX_NUM_DPM_LVL; i++) {
-			if (bw_params->clk_table.entries[i].dcfclk_mhz > max_dcfclk_mhz)
-				max_dcfclk_mhz = bw_params->clk_table.entries[i].dcfclk_mhz;
-			if (bw_params->clk_table.entries[i].dispclk_mhz > max_dispclk_mhz)
-				max_dispclk_mhz = bw_params->clk_table.entries[i].dispclk_mhz;
-			if (bw_params->clk_table.entries[i].dppclk_mhz > max_dppclk_mhz)
-				max_dppclk_mhz = bw_params->clk_table.entries[i].dppclk_mhz;
-			if (bw_params->clk_table.entries[i].phyclk_mhz > max_phyclk_mhz)
-				max_phyclk_mhz = bw_params->clk_table.entries[i].phyclk_mhz;
-		}
-		if (!max_dcfclk_mhz)
-			max_dcfclk_mhz = dcn3_02_soc.clock_limits[0].dcfclk_mhz;
-		if (!max_dispclk_mhz)
-			max_dispclk_mhz = dcn3_02_soc.clock_limits[0].dispclk_mhz;
-		if (!max_dppclk_mhz)
-			max_dppclk_mhz = dcn3_02_soc.clock_limits[0].dppclk_mhz;
-		if (!max_phyclk_mhz)
-			max_phyclk_mhz = dcn3_02_soc.clock_limits[0].phyclk_mhz;
-
-		if (max_dcfclk_mhz > dcfclk_sta_targets[num_dcfclk_sta_targets-1]) {
-			/* If max DCFCLK is greater than the max DCFCLK STA target, insert into the DCFCLK STA target array */
-			dcfclk_sta_targets[num_dcfclk_sta_targets] = max_dcfclk_mhz;
-			num_dcfclk_sta_targets++;
-		} else if (max_dcfclk_mhz < dcfclk_sta_targets[num_dcfclk_sta_targets-1]) {
-			/* If max DCFCLK is less than the max DCFCLK STA target, cap values and remove duplicates */
-			for (i = 0; i < num_dcfclk_sta_targets; i++) {
-				if (dcfclk_sta_targets[i] > max_dcfclk_mhz) {
-					dcfclk_sta_targets[i] = max_dcfclk_mhz;
-					break;
-				}
-			}
-			/* Update size of array since we "removed" duplicates */
-			num_dcfclk_sta_targets = i + 1;
-		}
-
-		num_uclk_states = bw_params->clk_table.num_entries;
-
-		/* Calculate optimal dcfclk for each uclk */
-		for (i = 0; i < num_uclk_states; i++) {
-			dcn302_get_optimal_dcfclk_fclk_for_uclk(bw_params->clk_table.entries[i].memclk_mhz * 16,
-					&optimal_dcfclk_for_uclk[i], NULL);
-			if (optimal_dcfclk_for_uclk[i] < bw_params->clk_table.entries[0].dcfclk_mhz) {
-				optimal_dcfclk_for_uclk[i] = bw_params->clk_table.entries[0].dcfclk_mhz;
-			}
-		}
-
-		/* Calculate optimal uclk for each dcfclk sta target */
-		for (i = 0; i < num_dcfclk_sta_targets; i++) {
-			for (j = 0; j < num_uclk_states; j++) {
-				if (dcfclk_sta_targets[i] < optimal_dcfclk_for_uclk[j]) {
-					optimal_uclk_for_dcfclk_sta_targets[i] =
-							bw_params->clk_table.entries[j].memclk_mhz * 16;
-					break;
-				}
-			}
-		}
-
-		i = 0;
-		j = 0;
-		/* create the final dcfclk and uclk table */
-		while (i < num_dcfclk_sta_targets && j < num_uclk_states && num_states < DC__VOLTAGE_STATES) {
-			if (dcfclk_sta_targets[i] < optimal_dcfclk_for_uclk[j] && i < num_dcfclk_sta_targets) {
-				dcfclk_mhz[num_states] = dcfclk_sta_targets[i];
-				dram_speed_mts[num_states++] = optimal_uclk_for_dcfclk_sta_targets[i++];
-			} else {
-				if (j < num_uclk_states && optimal_dcfclk_for_uclk[j] <= max_dcfclk_mhz) {
-					dcfclk_mhz[num_states] = optimal_dcfclk_for_uclk[j];
-					dram_speed_mts[num_states++] = bw_params->clk_table.entries[j++].memclk_mhz * 16;
-				} else {
-					j = num_uclk_states;
-				}
-			}
-		}
-
-		while (i < num_dcfclk_sta_targets && num_states < DC__VOLTAGE_STATES) {
-			dcfclk_mhz[num_states] = dcfclk_sta_targets[i];
-			dram_speed_mts[num_states++] = optimal_uclk_for_dcfclk_sta_targets[i++];
-		}
-
-		while (j < num_uclk_states && num_states < DC__VOLTAGE_STATES &&
-				optimal_dcfclk_for_uclk[j] <= max_dcfclk_mhz) {
-			dcfclk_mhz[num_states] = optimal_dcfclk_for_uclk[j];
-			dram_speed_mts[num_states++] = bw_params->clk_table.entries[j++].memclk_mhz * 16;
-		}
-
-		dcn3_02_soc.num_states = num_states;
-		for (i = 0; i < dcn3_02_soc.num_states; i++) {
-			dcn3_02_soc.clock_limits[i].state = i;
-			dcn3_02_soc.clock_limits[i].dcfclk_mhz = dcfclk_mhz[i];
-			dcn3_02_soc.clock_limits[i].fabricclk_mhz = dcfclk_mhz[i];
-			dcn3_02_soc.clock_limits[i].dram_speed_mts = dram_speed_mts[i];
-
-			/* Fill all states with max values of all other clocks */
-			dcn3_02_soc.clock_limits[i].dispclk_mhz = max_dispclk_mhz;
-			dcn3_02_soc.clock_limits[i].dppclk_mhz  = max_dppclk_mhz;
-			dcn3_02_soc.clock_limits[i].phyclk_mhz  = max_phyclk_mhz;
-			/* Populate from bw_params for DTBCLK, SOCCLK */
-			if (!bw_params->clk_table.entries[i].dtbclk_mhz && i > 0)
-				dcn3_02_soc.clock_limits[i].dtbclk_mhz  = dcn3_02_soc.clock_limits[i-1].dtbclk_mhz;
-			else
-				dcn3_02_soc.clock_limits[i].dtbclk_mhz  = bw_params->clk_table.entries[i].dtbclk_mhz;
-			if (!bw_params->clk_table.entries[i].socclk_mhz && i > 0)
-				dcn3_02_soc.clock_limits[i].socclk_mhz = dcn3_02_soc.clock_limits[i-1].socclk_mhz;
-			else
-				dcn3_02_soc.clock_limits[i].socclk_mhz = bw_params->clk_table.entries[i].socclk_mhz;
-			/* These clocks cannot come from bw_params, always fill from dcn3_02_soc[1] */
-			/* FCLK, PHYCLK_D18, DSCCLK */
-			dcn3_02_soc.clock_limits[i].phyclk_d18_mhz = dcn3_02_soc.clock_limits[0].phyclk_d18_mhz;
-			dcn3_02_soc.clock_limits[i].dscclk_mhz = dcn3_02_soc.clock_limits[0].dscclk_mhz;
-		}
-		/* re-init DML with updated bb */
-		dml_init_instance(&dc->dml, &dcn3_02_soc, &dcn3_02_ip, DML_PROJECT_DCN30);
-		if (dc->current_state)
-			dml_init_instance(&dc->current_state->bw_ctx.dml, &dcn3_02_soc, &dcn3_02_ip, DML_PROJECT_DCN30);
-	}
-}
-
-static struct resource_funcs dcn302_res_pool_funcs = {
-		.destroy = dcn302_destroy_resource_pool,
-		.link_enc_create = dcn302_link_encoder_create,
-		.panel_cntl_create = dcn302_panel_cntl_create,
-		.validate_bandwidth = dcn30_validate_bandwidth,
-		.calculate_wm_and_dlg = dcn30_calculate_wm_and_dlg,
-		.update_soc_for_wm_a = dcn30_update_soc_for_wm_a,
-		.populate_dml_pipes = dcn30_populate_dml_pipes_from_context,
-		.acquire_idle_pipe_for_layer = dcn20_acquire_idle_pipe_for_layer,
-		.add_stream_to_ctx = dcn30_add_stream_to_ctx,
-		.add_dsc_to_stream_resource = dcn20_add_dsc_to_stream_resource,
-		.remove_stream_from_ctx = dcn20_remove_stream_from_ctx,
-		.populate_dml_writeback_from_context = dcn30_populate_dml_writeback_from_context,
-		.set_mcif_arb_params = dcn30_set_mcif_arb_params,
-		.find_first_free_match_stream_enc_for_link = dcn10_find_first_free_match_stream_enc_for_link,
-		.acquire_post_bldn_3dlut = dcn30_acquire_post_bldn_3dlut,
-		.release_post_bldn_3dlut = dcn30_release_post_bldn_3dlut,
-		.update_bw_bounding_box = dcn302_update_bw_bounding_box,
-		.patch_unknown_plane_state = dcn20_patch_unknown_plane_state,
-};
-
-static struct dc_cap_funcs cap_funcs = {
-		.get_dcc_compression_cap = dcn20_get_dcc_compression_cap
-};
-
-static const struct bios_registers bios_regs = {
-		NBIO_SR(BIOS_SCRATCH_3),
-		NBIO_SR(BIOS_SCRATCH_6)
-};
-
-static const struct dccg_registers dccg_regs = {
-		DCCG_REG_LIST_DCN3_02()
-};
-
-static const struct dccg_shift dccg_shift = {
-		DCCG_MASK_SH_LIST_DCN3_02(__SHIFT)
-};
-
-static const struct dccg_mask dccg_mask = {
-		DCCG_MASK_SH_LIST_DCN3_02(_MASK)
-};
-
-#define abm_regs(id)\
-		[id] = { ABM_DCN302_REG_LIST(id) }
-
-static const struct dce_abm_registers abm_regs[] = {
-		abm_regs(0),
-		abm_regs(1),
-		abm_regs(2),
-		abm_regs(3),
-		abm_regs(4)
-};
-
-static const struct dce_abm_shift abm_shift = {
-		ABM_MASK_SH_LIST_DCN30(__SHIFT)
-};
-
-static const struct dce_abm_mask abm_mask = {
-		ABM_MASK_SH_LIST_DCN30(_MASK)
-};
-
-static bool dcn302_resource_construct(
-		uint8_t num_virtual_links,
-		struct dc *dc,
-		struct resource_pool *pool)
-{
-	int i;
-	struct dc_context *ctx = dc->ctx;
-	struct irq_service_init_data init_data;
-
-	ctx->dc_bios->regs = &bios_regs;
-
-	pool->res_cap = &res_cap_dcn302;
-
-	pool->funcs = &dcn302_res_pool_funcs;
-
-	/*************************************************
-	 *  Resource + asic cap harcoding                *
-	 *************************************************/
-	pool->underlay_pipe_index = NO_UNDERLAY_PIPE;
-	pool->pipe_count = pool->res_cap->num_timing_generator;
-	pool->mpcc_count = pool->res_cap->num_timing_generator;
-	dc->caps.max_downscale_ratio = 600;
-	dc->caps.i2c_speed_in_khz = 100;
-	dc->caps.i2c_speed_in_khz_hdcp = 5; /*1.4 w/a applied by derfault*/
-	dc->caps.max_cursor_size = 256;
-	dc->caps.min_horizontal_blanking_period = 80;
-	dc->caps.dmdata_alloc_size = 2048;
-	dc->caps.mall_size_per_mem_channel = 4;
-	/* total size = mall per channel * num channels * 1024 * 1024 */
-	dc->caps.mall_size_total = dc->caps.mall_size_per_mem_channel * dc->ctx->dc_bios->vram_info.num_chans * 1048576;
-	dc->caps.cursor_cache_size = dc->caps.max_cursor_size * dc->caps.max_cursor_size * 8;
-	dc->caps.max_slave_planes = 1;
-	dc->caps.max_slave_yuv_planes = 1;
-	dc->caps.max_slave_rgb_planes = 1;
-	dc->caps.post_blend_color_processing = true;
-	dc->caps.force_dp_tps4_for_cp2520 = true;
-	dc->caps.extended_aux_timeout_support = true;
-	dc->caps.dmcub_support = true;
-
-	/* Color pipeline capabilities */
-	dc->caps.color.dpp.dcn_arch = 1;
-	dc->caps.color.dpp.input_lut_shared = 0;
-	dc->caps.color.dpp.icsc = 1;
-	dc->caps.color.dpp.dgam_ram = 0; // must use gamma_corr
-	dc->caps.color.dpp.dgam_rom_caps.srgb = 1;
-	dc->caps.color.dpp.dgam_rom_caps.bt2020 = 1;
-	dc->caps.color.dpp.dgam_rom_caps.gamma2_2 = 1;
-	dc->caps.color.dpp.dgam_rom_caps.pq = 1;
-	dc->caps.color.dpp.dgam_rom_caps.hlg = 1;
-	dc->caps.color.dpp.post_csc = 1;
-	dc->caps.color.dpp.gamma_corr = 1;
-	dc->caps.color.dpp.dgam_rom_for_yuv = 0;
-
-	dc->caps.color.dpp.hw_3d_lut = 1;
-	dc->caps.color.dpp.ogam_ram = 1;
-	// no OGAM ROM on DCN3
-	dc->caps.color.dpp.ogam_rom_caps.srgb = 0;
-	dc->caps.color.dpp.ogam_rom_caps.bt2020 = 0;
-	dc->caps.color.dpp.ogam_rom_caps.gamma2_2 = 0;
-	dc->caps.color.dpp.ogam_rom_caps.pq = 0;
-	dc->caps.color.dpp.ogam_rom_caps.hlg = 0;
-	dc->caps.color.dpp.ocsc = 0;
-
-	dc->caps.color.mpc.gamut_remap = 1;
-	dc->caps.color.mpc.num_3dluts = pool->res_cap->num_mpc_3dlut; //3
-	dc->caps.color.mpc.ogam_ram = 1;
-	dc->caps.color.mpc.ogam_rom_caps.srgb = 0;
-	dc->caps.color.mpc.ogam_rom_caps.bt2020 = 0;
-	dc->caps.color.mpc.ogam_rom_caps.gamma2_2 = 0;
-	dc->caps.color.mpc.ogam_rom_caps.pq = 0;
-	dc->caps.color.mpc.ogam_rom_caps.hlg = 0;
-	dc->caps.color.mpc.ocsc = 1;
-
-	if (dc->ctx->dce_environment == DCE_ENV_PRODUCTION_DRV)
-		dc->debug = debug_defaults_drv;
-	else
-		dc->debug = debug_defaults_diags;
-
-	// Init the vm_helper
-	if (dc->vm_helper)
-		vm_helper_init(dc->vm_helper, 16);
-
-	/*************************************************
-	 *  Create resources                             *
-	 *************************************************/
-
-	/* Clock Sources for Pixel Clock*/
-	pool->clock_sources[DCN302_CLK_SRC_PLL0] =
-			dcn302_clock_source_create(ctx, ctx->dc_bios,
-					CLOCK_SOURCE_COMBO_PHY_PLL0,
-					&clk_src_regs[0], false);
-	pool->clock_sources[DCN302_CLK_SRC_PLL1] =
-			dcn302_clock_source_create(ctx, ctx->dc_bios,
-					CLOCK_SOURCE_COMBO_PHY_PLL1,
-					&clk_src_regs[1], false);
-	pool->clock_sources[DCN302_CLK_SRC_PLL2] =
-			dcn302_clock_source_create(ctx, ctx->dc_bios,
-					CLOCK_SOURCE_COMBO_PHY_PLL2,
-					&clk_src_regs[2], false);
-	pool->clock_sources[DCN302_CLK_SRC_PLL3] =
-			dcn302_clock_source_create(ctx, ctx->dc_bios,
-					CLOCK_SOURCE_COMBO_PHY_PLL3,
-					&clk_src_regs[3], false);
-	pool->clock_sources[DCN302_CLK_SRC_PLL4] =
-			dcn302_clock_source_create(ctx, ctx->dc_bios,
-					CLOCK_SOURCE_COMBO_PHY_PLL4,
-					&clk_src_regs[4], false);
-
-	pool->clk_src_count = DCN302_CLK_SRC_TOTAL;
-
-	/* todo: not reuse phy_pll registers */
-	pool->dp_clock_source =
-			dcn302_clock_source_create(ctx, ctx->dc_bios,
-					CLOCK_SOURCE_ID_DP_DTO,
-					&clk_src_regs[0], true);
-
-	for (i = 0; i < pool->clk_src_count; i++) {
-		if (pool->clock_sources[i] == NULL) {
-			dm_error("DC: failed to create clock sources!\n");
-			BREAK_TO_DEBUGGER();
-			goto create_fail;
-		}
-	}
-
-	/* DCCG */
-	pool->dccg = dccg30_create(ctx, &dccg_regs, &dccg_shift, &dccg_mask);
-	if (pool->dccg == NULL) {
-		dm_error("DC: failed to create dccg!\n");
-		BREAK_TO_DEBUGGER();
-		goto create_fail;
-	}
-
-	/* PP Lib and SMU interfaces */
-	init_soc_bounding_box(dc, pool);
-
-	/* DML */
-	dml_init_instance(&dc->dml, &dcn3_02_soc, &dcn3_02_ip, DML_PROJECT_DCN30);
-
-	/* IRQ */
-	init_data.ctx = dc->ctx;
-	pool->irqs = dal_irq_service_dcn302_create(&init_data);
-	if (!pool->irqs)
-		goto create_fail;
-
-	/* HUBBUB */
-	pool->hubbub = dcn302_hubbub_create(ctx);
-	if (pool->hubbub == NULL) {
-		BREAK_TO_DEBUGGER();
-		dm_error("DC: failed to create hubbub!\n");
-		goto create_fail;
-	}
-
-	/* HUBPs, DPPs, OPPs and TGs */
-	for (i = 0; i < pool->pipe_count; i++) {
-		pool->hubps[i] = dcn302_hubp_create(ctx, i);
-		if (pool->hubps[i] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error("DC: failed to create hubps!\n");
-			goto create_fail;
-		}
-
-		pool->dpps[i] = dcn302_dpp_create(ctx, i);
-		if (pool->dpps[i] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error("DC: failed to create dpps!\n");
-			goto create_fail;
-		}
-	}
-
-	for (i = 0; i < pool->res_cap->num_opp; i++) {
-		pool->opps[i] = dcn302_opp_create(ctx, i);
-		if (pool->opps[i] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error("DC: failed to create output pixel processor!\n");
-			goto create_fail;
-		}
-	}
-
-	for (i = 0; i < pool->res_cap->num_timing_generator; i++) {
-		pool->timing_generators[i] = dcn302_timing_generator_create(ctx, i);
-		if (pool->timing_generators[i] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error("DC: failed to create tg!\n");
-			goto create_fail;
-		}
-	}
-	pool->timing_generator_count = i;
-
-	/* PSR */
-	pool->psr = dmub_psr_create(ctx);
-	if (pool->psr == NULL) {
-		dm_error("DC: failed to create psr!\n");
-		BREAK_TO_DEBUGGER();
-		goto create_fail;
-	}
-
-	/* ABMs */
-	for (i = 0; i < pool->res_cap->num_timing_generator; i++) {
-		pool->multiple_abms[i] = dmub_abm_create(ctx, &abm_regs[i], &abm_shift, &abm_mask);
-		if (pool->multiple_abms[i] == NULL) {
-			dm_error("DC: failed to create abm for pipe %d!\n", i);
-			BREAK_TO_DEBUGGER();
-			goto create_fail;
-		}
-	}
-
-	/* MPC and DSC */
-	pool->mpc = dcn302_mpc_create(ctx, pool->mpcc_count, pool->res_cap->num_mpc_3dlut);
-	if (pool->mpc == NULL) {
-		BREAK_TO_DEBUGGER();
-		dm_error("DC: failed to create mpc!\n");
-		goto create_fail;
-	}
-
-	for (i = 0; i < pool->res_cap->num_dsc; i++) {
-		pool->dscs[i] = dcn302_dsc_create(ctx, i);
-		if (pool->dscs[i] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error("DC: failed to create display stream compressor %d!\n", i);
-			goto create_fail;
-		}
-	}
-
-	/* DWB and MMHUBBUB */
-	if (!dcn302_dwbc_create(ctx, pool)) {
-		BREAK_TO_DEBUGGER();
-		dm_error("DC: failed to create dwbc!\n");
-		goto create_fail;
-	}
-
-	if (!dcn302_mmhubbub_create(ctx, pool)) {
-		BREAK_TO_DEBUGGER();
-		dm_error("DC: failed to create mcif_wb!\n");
-		goto create_fail;
-	}
-
-	/* AUX and I2C */
-	for (i = 0; i < pool->res_cap->num_ddc; i++) {
-		pool->engines[i] = dcn302_aux_engine_create(ctx, i);
-		if (pool->engines[i] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error("DC:failed to create aux engine!!\n");
-			goto create_fail;
-		}
-		pool->hw_i2cs[i] = dcn302_i2c_hw_create(ctx, i);
-		if (pool->hw_i2cs[i] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error("DC:failed to create hw i2c!!\n");
-			goto create_fail;
-		}
-		pool->sw_i2cs[i] = NULL;
-	}
-
-	/* Audio, Stream Encoders including HPO and virtual, MPC 3D LUTs */
-	if (!resource_construct(num_virtual_links, dc, pool,
-			(!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment) ?
-					&res_create_funcs : &res_create_maximus_funcs)))
-		goto create_fail;
-
-	/* HW Sequencer and Plane caps */
-	dcn302_hw_sequencer_construct(dc);
-
-	dc->caps.max_planes =  pool->pipe_count;
-
-	for (i = 0; i < dc->caps.max_planes; ++i)
-		dc->caps.planes[i] = plane_cap;
-
-	dc->cap_funcs = cap_funcs;
-
-	return true;
-
-create_fail:
-
-	dcn302_resource_destruct(pool);
-
-	return false;
-}
-
-struct resource_pool *dcn302_create_resource_pool(const struct dc_init_data *init_data, struct dc *dc)
-{
-	struct resource_pool *pool = kzalloc(sizeof(struct resource_pool), GFP_KERNEL);
-
-	if (!pool)
-		return NULL;
-
-	if (dcn302_resource_construct(init_data->num_virtual_links, dc, pool))
-		return pool;
-
-	BREAK_TO_DEBUGGER();
-	kfree(pool);
-	return NULL;
-}
diff --git a/drivers/gpu/drm/amd/display/dc/dcn303/Makefile b/drivers/gpu/drm/amd/display/dc/dcn303/Makefile
deleted file mode 100644
index 6f7a1f2b49f0..000000000000
--- a/drivers/gpu/drm/amd/display/dc/dcn303/Makefile
+++ /dev/null
@@ -1,39 +0,0 @@
-# SPDX-License-Identifier: MIT
-#
-# Copyright (C) 2021 Advanced Micro Devices, Inc. All the rights reserved
-#
-#  Authors: AMD
-#
-# Makefile for dcn303.
-
-DCN3_03 = dcn303_init.o dcn303_hwseq.o dcn303_resource.o
-
-ifdef CONFIG_X86
-CFLAGS_$(AMDDALPATH)/dc/dcn303/dcn303_resource.o := -msse
-endif
-
-ifdef CONFIG_PPC64
-CFLAGS_$(AMDDALPATH)/dc/dcn303/dcn303_resource.o := -mhard-float -maltivec
-endif
-
-ifdef CONFIG_CC_IS_GCC
-ifeq ($(call cc-ifversion, -lt, 0701, y), y)
-IS_OLD_GCC = 1
-endif
-CFLAGS_$(AMDDALPATH)/dc/dcn303/dcn303_resource.o += -mhard-float
-endif
-
-ifdef CONFIG_X86
-ifdef IS_OLD_GCC
-# Stack alignment mismatch, proceed with caution.
-# GCC < 7.1 cannot compile code using `double` and -mpreferred-stack-boundary=3
-# (8B stack alignment).
-CFLAGS_$(AMDDALPATH)/dc/dcn303/dcn303_resource.o += -mpreferred-stack-boundary=4
-else
-CFLAGS_$(AMDDALPATH)/dc/dcn303/dcn303_resource.o += -msse2
-endif
-endif
-
-AMD_DAL_DCN3_03 = $(addprefix $(AMDDALPATH)/dc/dcn303/,$(DCN3_03))
-
-AMD_DISPLAY_FILES += $(AMD_DAL_DCN3_03)
diff --git a/drivers/gpu/drm/amd/display/dc/dcn303/dcn303_dccg.h b/drivers/gpu/drm/amd/display/dc/dcn303/dcn303_dccg.h
deleted file mode 100644
index a79c54bbc899..000000000000
--- a/drivers/gpu/drm/amd/display/dc/dcn303/dcn303_dccg.h
+++ /dev/null
@@ -1,30 +0,0 @@
-// SPDX-License-Identifier: MIT
-/*
- * Copyright (C) 2021 Advanced Micro Devices, Inc.
- *
- * Authors: AMD
- */
-
-#ifndef __DCN303_DCCG_H__
-#define __DCN303_DCCG_H__
-
-#include "dcn30/dcn30_dccg.h"
-
-
-#define DCCG_REG_LIST_DCN3_03() \
-	SR(DPPCLK_DTO_CTRL),\
-	DCCG_SRII(DTO_PARAM, DPPCLK, 0),\
-	DCCG_SRII(DTO_PARAM, DPPCLK, 1),\
-	SR(REFCLK_CNTL)
-
-#define DCCG_MASK_SH_LIST_DCN3_03(mask_sh) \
-		DCCG_SFI(DPPCLK_DTO_CTRL, DTO_ENABLE, DPPCLK, 0, mask_sh),\
-		DCCG_SFI(DPPCLK_DTO_CTRL, DTO_DB_EN, DPPCLK, 0, mask_sh),\
-		DCCG_SFI(DPPCLK_DTO_CTRL, DTO_ENABLE, DPPCLK, 1, mask_sh),\
-		DCCG_SFI(DPPCLK_DTO_CTRL, DTO_DB_EN, DPPCLK, 1, mask_sh),\
-		DCCG_SF(DPPCLK0_DTO_PARAM, DPPCLK0_DTO_PHASE, mask_sh),\
-		DCCG_SF(DPPCLK0_DTO_PARAM, DPPCLK0_DTO_MODULO, mask_sh),\
-		DCCG_SF(REFCLK_CNTL, REFCLK_CLOCK_EN, mask_sh),\
-		DCCG_SF(REFCLK_CNTL, REFCLK_SRC_SEL, mask_sh)
-
-#endif //__DCN303_DCCG_H__
diff --git a/drivers/gpu/drm/amd/display/dc/dcn303/dcn303_hwseq.c b/drivers/gpu/drm/amd/display/dc/dcn303/dcn303_hwseq.c
deleted file mode 100644
index b48b732aa647..000000000000
--- a/drivers/gpu/drm/amd/display/dc/dcn303/dcn303_hwseq.c
+++ /dev/null
@@ -1,45 +0,0 @@
-// SPDX-License-Identifier: MIT
-/*
- * Copyright (C) 2021 Advanced Micro Devices, Inc.
- *
- * Authors: AMD
- */
-
-#include "dcn303_hwseq.h"
-
-#include "dce/dce_hwseq.h"
-
-#include "reg_helper.h"
-#include "dc.h"
-
-#define DC_LOGGER_INIT(logger)
-
-#define CTX \
-	hws->ctx
-#define REG(reg)\
-	hws->regs->reg
-
-#undef FN
-#define FN(reg_name, field_name) \
-	hws->shifts->field_name, hws->masks->field_name
-
-
-void dcn303_dpp_pg_control(struct dce_hwseq *hws, unsigned int dpp_inst, bool power_on)
-{
-	/*DCN303 removes PG registers*/
-}
-
-void dcn303_hubp_pg_control(struct dce_hwseq *hws, unsigned int hubp_inst, bool power_on)
-{
-	/*DCN303 removes PG registers*/
-}
-
-void dcn303_dsc_pg_control(struct dce_hwseq *hws, unsigned int dsc_inst, bool power_on)
-{
-	/*DCN303 removes PG registers*/
-}
-
-void dcn303_enable_power_gating_plane(struct dce_hwseq *hws, bool enable)
-{
-	/*DCN303 removes PG registers*/
-}
diff --git a/drivers/gpu/drm/amd/display/dc/dcn303/dcn303_hwseq.h b/drivers/gpu/drm/amd/display/dc/dcn303/dcn303_hwseq.h
deleted file mode 100644
index 8b69a3b76c11..000000000000
--- a/drivers/gpu/drm/amd/display/dc/dcn303/dcn303_hwseq.h
+++ /dev/null
@@ -1,18 +0,0 @@
-// SPDX-License-Identifier: MIT
-/*
- * Copyright (C) 2021 Advanced Micro Devices, Inc.
- *
- * Authors: AMD
- */
-
-#ifndef __DC_HWSS_DCN303_H__
-#define __DC_HWSS_DCN303_H__
-
-#include "hw_sequencer_private.h"
-
-void dcn303_dpp_pg_control(struct dce_hwseq *hws, unsigned int dpp_inst, bool power_on);
-void dcn303_hubp_pg_control(struct dce_hwseq *hws, unsigned int hubp_inst, bool power_on);
-void dcn303_dsc_pg_control(struct dce_hwseq *hws, unsigned int dsc_inst, bool power_on);
-void dcn303_enable_power_gating_plane(struct dce_hwseq *hws, bool enable);
-
-#endif /* __DC_HWSS_DCN303_H__ */
diff --git a/drivers/gpu/drm/amd/display/dc/dcn303/dcn303_init.c b/drivers/gpu/drm/amd/display/dc/dcn303/dcn303_init.c
deleted file mode 100644
index aa5dbbade2bd..000000000000
--- a/drivers/gpu/drm/amd/display/dc/dcn303/dcn303_init.c
+++ /dev/null
@@ -1,20 +0,0 @@
-// SPDX-License-Identifier: MIT
-/*
- * Copyright (C) 2021 Advanced Micro Devices, Inc.
- *
- * Authors: AMD
- */
-
-#include "dcn303_hwseq.h"
-#include "dcn30/dcn30_init.h"
-#include "dc.h"
-
-void dcn303_hw_sequencer_construct(struct dc *dc)
-{
-	dcn30_hw_sequencer_construct(dc);
-
-	dc->hwseq->funcs.dpp_pg_control = dcn303_dpp_pg_control;
-	dc->hwseq->funcs.hubp_pg_control = dcn303_hubp_pg_control;
-	dc->hwseq->funcs.dsc_pg_control = dcn303_dsc_pg_control;
-	dc->hwseq->funcs.enable_power_gating_plane = dcn303_enable_power_gating_plane;
-}
diff --git a/drivers/gpu/drm/amd/display/dc/dcn303/dcn303_init.h b/drivers/gpu/drm/amd/display/dc/dcn303/dcn303_init.h
deleted file mode 100644
index 66b1e3604f07..000000000000
--- a/drivers/gpu/drm/amd/display/dc/dcn303/dcn303_init.h
+++ /dev/null
@@ -1,15 +0,0 @@
-// SPDX-License-Identifier: MIT
-/*
- * Copyright (C) 2021 Advanced Micro Devices, Inc.
- *
- * Authors: AMD
- */
-
-#ifndef __DC_DCN303_INIT_H__
-#define __DC_DCN303_INIT_H__
-
-struct dc;
-
-void dcn303_hw_sequencer_construct(struct dc *dc);
-
-#endif /* __DC_DCN303_INIT_H__ */
diff --git a/drivers/gpu/drm/amd/display/dc/dcn303/dcn303_resource.c b/drivers/gpu/drm/amd/display/dc/dcn303/dcn303_resource.c
deleted file mode 100644
index 4a9b64023675..000000000000
--- a/drivers/gpu/drm/amd/display/dc/dcn303/dcn303_resource.c
+++ /dev/null
@@ -1,1720 +0,0 @@
-// SPDX-License-Identifier: MIT
-/*
- * Copyright (C) 2021 Advanced Micro Devices, Inc.
- *
- * Authors: AMD
- */
-
-#include "dcn303_init.h"
-#include "dcn303_resource.h"
-#include "dcn303_dccg.h"
-#include "irq/dcn303/irq_service_dcn303.h"
-
-#include "dcn30/dcn30_dio_link_encoder.h"
-#include "dcn30/dcn30_dio_stream_encoder.h"
-#include "dcn30/dcn30_dpp.h"
-#include "dcn30/dcn30_dwb.h"
-#include "dcn30/dcn30_hubbub.h"
-#include "dcn30/dcn30_hubp.h"
-#include "dcn30/dcn30_mmhubbub.h"
-#include "dcn30/dcn30_mpc.h"
-#include "dcn30/dcn30_opp.h"
-#include "dcn30/dcn30_optc.h"
-#include "dcn30/dcn30_resource.h"
-
-#include "dcn20/dcn20_dsc.h"
-#include "dcn20/dcn20_resource.h"
-
-#include "dcn10/dcn10_resource.h"
-
-#include "dc_link_ddc.h"
-
-#include "dce/dce_abm.h"
-#include "dce/dce_audio.h"
-#include "dce/dce_aux.h"
-#include "dce/dce_clock_source.h"
-#include "dce/dce_hwseq.h"
-#include "dce/dce_i2c_hw.h"
-#include "dce/dce_panel_cntl.h"
-#include "dce/dmub_abm.h"
-#include "dce/dmub_psr.h"
-#include "clk_mgr.h"
-
-#include "hw_sequencer_private.h"
-#include "reg_helper.h"
-#include "resource.h"
-#include "vm_helper.h"
-
-#include "sienna_cichlid_ip_offset.h"
-#include "dcn/dcn_3_0_3_offset.h"
-#include "dcn/dcn_3_0_3_sh_mask.h"
-#include "dcn/dpcs_3_0_3_offset.h"
-#include "dcn/dpcs_3_0_3_sh_mask.h"
-#include "nbio/nbio_2_3_offset.h"
-
-#define DC_LOGGER_INIT(logger)
-
-struct _vcs_dpi_ip_params_st dcn3_03_ip = {
-		.use_min_dcfclk = 0,
-		.clamp_min_dcfclk = 0,
-		.odm_capable = 1,
-		.gpuvm_enable = 1,
-		.hostvm_enable = 0,
-		.gpuvm_max_page_table_levels = 4,
-		.hostvm_max_page_table_levels = 4,
-		.hostvm_cached_page_table_levels = 0,
-		.pte_group_size_bytes = 2048,
-		.num_dsc = 2,
-		.rob_buffer_size_kbytes = 184,
-		.det_buffer_size_kbytes = 184,
-		.dpte_buffer_size_in_pte_reqs_luma = 64,
-		.dpte_buffer_size_in_pte_reqs_chroma = 34,
-		.pde_proc_buffer_size_64k_reqs = 48,
-		.dpp_output_buffer_pixels = 2560,
-		.opp_output_buffer_lines = 1,
-		.pixel_chunk_size_kbytes = 8,
-		.pte_enable = 1,
-		.max_page_table_levels = 2,
-		.pte_chunk_size_kbytes = 2,  // ?
-		.meta_chunk_size_kbytes = 2,
-		.writeback_chunk_size_kbytes = 8,
-		.line_buffer_size_bits = 789504,
-		.is_line_buffer_bpp_fixed = 0,  // ?
-		.line_buffer_fixed_bpp = 0,     // ?
-		.dcc_supported = true,
-		.writeback_interface_buffer_size_kbytes = 90,
-		.writeback_line_buffer_buffer_size = 0,
-		.max_line_buffer_lines = 12,
-		.writeback_luma_buffer_size_kbytes = 12,  // writeback_line_buffer_buffer_size = 656640
-		.writeback_chroma_buffer_size_kbytes = 8,
-		.writeback_chroma_line_buffer_width_pixels = 4,
-		.writeback_max_hscl_ratio = 1,
-		.writeback_max_vscl_ratio = 1,
-		.writeback_min_hscl_ratio = 1,
-		.writeback_min_vscl_ratio = 1,
-		.writeback_max_hscl_taps = 1,
-		.writeback_max_vscl_taps = 1,
-		.writeback_line_buffer_luma_buffer_size = 0,
-		.writeback_line_buffer_chroma_buffer_size = 14643,
-		.cursor_buffer_size = 8,
-		.cursor_chunk_size = 2,
-		.max_num_otg = 2,
-		.max_num_dpp = 2,
-		.max_num_wb = 1,
-		.max_dchub_pscl_bw_pix_per_clk = 4,
-		.max_pscl_lb_bw_pix_per_clk = 2,
-		.max_lb_vscl_bw_pix_per_clk = 4,
-		.max_vscl_hscl_bw_pix_per_clk = 4,
-		.max_hscl_ratio = 6,
-		.max_vscl_ratio = 6,
-		.hscl_mults = 4,
-		.vscl_mults = 4,
-		.max_hscl_taps = 8,
-		.max_vscl_taps = 8,
-		.dispclk_ramp_margin_percent = 1,
-		.underscan_factor = 1.11,
-		.min_vblank_lines = 32,
-		.dppclk_delay_subtotal = 46,
-		.dynamic_metadata_vm_enabled = true,
-		.dppclk_delay_scl_lb_only = 16,
-		.dppclk_delay_scl = 50,
-		.dppclk_delay_cnvc_formatter = 27,
-		.dppclk_delay_cnvc_cursor = 6,
-		.dispclk_delay_subtotal = 119,
-		.dcfclk_cstate_latency = 5.2, // SRExitTime
-		.max_inter_dcn_tile_repeaters = 8,
-		.max_num_hdmi_frl_outputs = 1,
-		.odm_combine_4to1_supported = false,
-		.xfc_supported = false,
-		.xfc_fill_bw_overhead_percent = 10.0,
-		.xfc_fill_constant_bytes = 0,
-		.gfx7_compat_tiling_supported = 0,
-		.number_of_cursors = 1,
-};
-
-struct _vcs_dpi_soc_bounding_box_st dcn3_03_soc = {
-		.clock_limits = {
-				{
-						.state = 0,
-						.dispclk_mhz = 1217.0,
-						.dppclk_mhz = 1217.0,
-						.phyclk_mhz = 810.0,
-						.phyclk_d18_mhz = 667.0,
-						.dscclk_mhz = 405.6,
-				},
-		},
-
-		.min_dcfclk = 500.0, /* TODO: set this to actual min DCFCLK */
-		.num_states = 1,
-		.sr_exit_time_us = 35.5,
-		.sr_enter_plus_exit_time_us = 40,
-		.urgent_latency_us = 4.0,
-		.urgent_latency_pixel_data_only_us = 4.0,
-		.urgent_latency_pixel_mixed_with_vm_data_us = 4.0,
-		.urgent_latency_vm_data_only_us = 4.0,
-		.urgent_out_of_order_return_per_channel_pixel_only_bytes = 4096,
-		.urgent_out_of_order_return_per_channel_pixel_and_vm_bytes = 4096,
-		.urgent_out_of_order_return_per_channel_vm_only_bytes = 4096,
-		.pct_ideal_dram_sdp_bw_after_urgent_pixel_only = 80.0,
-		.pct_ideal_dram_sdp_bw_after_urgent_pixel_and_vm = 60.0,
-		.pct_ideal_dram_sdp_bw_after_urgent_vm_only = 40.0,
-		.max_avg_sdp_bw_use_normal_percent = 60.0,
-		.max_avg_dram_bw_use_normal_percent = 40.0,
-		.writeback_latency_us = 12.0,
-		.max_request_size_bytes = 256,
-		.fabric_datapath_to_dcn_data_return_bytes = 64,
-		.dcn_downspread_percent = 0.5,
-		.downspread_percent = 0.38,
-		.dram_page_open_time_ns = 50.0,
-		.dram_rw_turnaround_time_ns = 17.5,
-		.dram_return_buffer_per_channel_bytes = 8192,
-		.round_trip_ping_latency_dcfclk_cycles = 156,
-		.urgent_out_of_order_return_per_channel_bytes = 4096,
-		.channel_interleave_bytes = 256,
-		.num_banks = 8,
-		.gpuvm_min_page_size_bytes = 4096,
-		.hostvm_min_page_size_bytes = 4096,
-		.dram_clock_change_latency_us = 404,
-		.dummy_pstate_latency_us = 5,
-		.writeback_dram_clock_change_latency_us = 23.0,
-		.return_bus_width_bytes = 64,
-		.dispclk_dppclk_vco_speed_mhz = 3650,
-		.xfc_bus_transport_time_us = 20,      // ?
-		.xfc_xbuf_latency_tolerance_us = 4,  // ?
-		.use_urgent_burst_bw = 1,            // ?
-		.do_urgent_latency_adjustment = true,
-		.urgent_latency_adjustment_fabric_clock_component_us = 1.0,
-		.urgent_latency_adjustment_fabric_clock_reference_mhz = 1000,
-};
-
-static const struct dc_debug_options debug_defaults_drv = {
-		.disable_dmcu = true,
-		.force_abm_enable = false,
-		.timing_trace = false,
-		.clock_trace = true,
-		.disable_pplib_clock_request = true,
-		.pipe_split_policy = MPC_SPLIT_AVOID_MULT_DISP,
-		.force_single_disp_pipe_split = false,
-		.disable_dcc = DCC_ENABLE,
-		.vsr_support = true,
-		.performance_trace = false,
-		.max_downscale_src_width = 7680,/*upto 8K*/
-		.disable_pplib_wm_range = false,
-		.scl_reset_length10 = true,
-		.sanity_checks = false,
-		.underflow_assert_delay_us = 0xFFFFFFFF,
-		.dwb_fi_phase = -1, // -1 = disable,
-		.dmub_command_table = true,
-		.disable_idle_power_optimizations = false,
-};
-
-static const struct dc_debug_options debug_defaults_diags = {
-		.disable_dmcu = true,
-		.force_abm_enable = false,
-		.timing_trace = true,
-		.clock_trace = true,
-		.disable_dpp_power_gate = true,
-		.disable_hubp_power_gate = true,
-		.disable_clock_gate = true,
-		.disable_pplib_clock_request = true,
-		.disable_pplib_wm_range = true,
-		.disable_stutter = false,
-		.scl_reset_length10 = true,
-		.dwb_fi_phase = -1, // -1 = disable
-		.dmub_command_table = true,
-		.enable_tri_buf = true,
-		.disable_psr = true,
-};
-
-enum dcn303_clk_src_array_id {
-	DCN303_CLK_SRC_PLL0,
-	DCN303_CLK_SRC_PLL1,
-	DCN303_CLK_SRC_TOTAL
-};
-
-static const struct resource_caps res_cap_dcn303 = {
-		.num_timing_generator = 2,
-		.num_opp = 2,
-		.num_video_plane = 2,
-		.num_audio = 2,
-		.num_stream_encoder = 2,
-		.num_dwb = 1,
-		.num_ddc = 2,
-		.num_vmid = 16,
-		.num_mpc_3dlut = 1,
-		.num_dsc = 2,
-};
-
-static const struct dc_plane_cap plane_cap = {
-		.type = DC_PLANE_TYPE_DCN_UNIVERSAL,
-		.blends_with_above = true,
-		.blends_with_below = true,
-		.per_pixel_alpha = true,
-		.pixel_format_support = {
-				.argb8888 = true,
-				.nv12 = true,
-				.fp16 = true,
-				.p010 = false,
-				.ayuv = false,
-		},
-		.max_upscale_factor = {
-				.argb8888 = 16000,
-				.nv12 = 16000,
-				.fp16 = 16000
-		},
-		.max_downscale_factor = {
-				.argb8888 = 600,
-				.nv12 = 600,
-				.fp16 = 600
-		},
-		16,
-		16
-};
-
-/* NBIO */
-#define NBIO_BASE_INNER(seg) \
-		NBIO_BASE__INST0_SEG ## seg
-
-#define NBIO_BASE(seg) \
-		NBIO_BASE_INNER(seg)
-
-#define NBIO_SR(reg_name)\
-		.reg_name = NBIO_BASE(mm ## reg_name ## _BASE_IDX) + \
-		mm ## reg_name
-
-/* DCN */
-#undef BASE_INNER
-#define BASE_INNER(seg) DCN_BASE__INST0_SEG ## seg
-
-#define BASE(seg) BASE_INNER(seg)
-
-#define SR(reg_name)\
-		.reg_name = BASE(mm ## reg_name ## _BASE_IDX) + mm ## reg_name
-
-#define SF(reg_name, field_name, post_fix)\
-		.field_name = reg_name ## __ ## field_name ## post_fix
-
-#define SRI(reg_name, block, id)\
-		.reg_name = BASE(mm ## block ## id ## _ ## reg_name ## _BASE_IDX) + mm ## block ## id ## _ ## reg_name
-
-#define SRI2(reg_name, block, id)\
-		.reg_name = BASE(mm ## reg_name ## _BASE_IDX) + mm ## reg_name
-
-#define SRII(reg_name, block, id)\
-		.reg_name[id] = BASE(mm ## block ## id ## _ ## reg_name ## _BASE_IDX) + \
-		mm ## block ## id ## _ ## reg_name
-
-#define DCCG_SRII(reg_name, block, id)\
-		.block ## _ ## reg_name[id] = BASE(mm ## block ## id ## _ ## reg_name ## _BASE_IDX) + \
-		mm ## block ## id ## _ ## reg_name
-
-#define VUPDATE_SRII(reg_name, block, id)\
-		.reg_name[id] = BASE(mm ## reg_name ## _ ## block ## id ## _BASE_IDX) + \
-		mm ## reg_name ## _ ## block ## id
-
-#define SRII_DWB(reg_name, temp_name, block, id)\
-		.reg_name[id] = BASE(mm ## block ## id ## _ ## temp_name ## _BASE_IDX) + \
-		mm ## block ## id ## _ ## temp_name
-
-#define SRII_MPC_RMU(reg_name, block, id)\
-		.RMU##_##reg_name[id] = BASE(mm ## block ## id ## _ ## reg_name ## _BASE_IDX) + \
-		mm ## block ## id ## _ ## reg_name
-
-static const struct dcn_hubbub_registers hubbub_reg = {
-		HUBBUB_REG_LIST_DCN30(0)
-};
-
-static const struct dcn_hubbub_shift hubbub_shift = {
-		HUBBUB_MASK_SH_LIST_DCN30(__SHIFT)
-};
-
-static const struct dcn_hubbub_mask hubbub_mask = {
-		HUBBUB_MASK_SH_LIST_DCN30(_MASK)
-};
-
-#define vmid_regs(id)\
-		[id] = { DCN20_VMID_REG_LIST(id) }
-
-static const struct dcn_vmid_registers vmid_regs[] = {
-		vmid_regs(0),
-		vmid_regs(1),
-		vmid_regs(2),
-		vmid_regs(3),
-		vmid_regs(4),
-		vmid_regs(5),
-		vmid_regs(6),
-		vmid_regs(7),
-		vmid_regs(8),
-		vmid_regs(9),
-		vmid_regs(10),
-		vmid_regs(11),
-		vmid_regs(12),
-		vmid_regs(13),
-		vmid_regs(14),
-		vmid_regs(15)
-};
-
-static const struct dcn20_vmid_shift vmid_shifts = {
-		DCN20_VMID_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dcn20_vmid_mask vmid_masks = {
-		DCN20_VMID_MASK_SH_LIST(_MASK)
-};
-
-static struct hubbub *dcn303_hubbub_create(struct dc_context *ctx)
-{
-	int i;
-
-	struct dcn20_hubbub *hubbub3 = kzalloc(sizeof(struct dcn20_hubbub), GFP_KERNEL);
-
-	if (!hubbub3)
-		return NULL;
-
-	hubbub3_construct(hubbub3, ctx, &hubbub_reg, &hubbub_shift, &hubbub_mask);
-
-	for (i = 0; i < res_cap_dcn303.num_vmid; i++) {
-		struct dcn20_vmid *vmid = &hubbub3->vmid[i];
-
-		vmid->ctx = ctx;
-
-		vmid->regs = &vmid_regs[i];
-		vmid->shifts = &vmid_shifts;
-		vmid->masks = &vmid_masks;
-	}
-
-	return &hubbub3->base;
-}
-
-#define vpg_regs(id)\
-		[id] = { VPG_DCN3_REG_LIST(id) }
-
-static const struct dcn30_vpg_registers vpg_regs[] = {
-		vpg_regs(0),
-		vpg_regs(1),
-		vpg_regs(2)
-};
-
-static const struct dcn30_vpg_shift vpg_shift = {
-		DCN3_VPG_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dcn30_vpg_mask vpg_mask = {
-		DCN3_VPG_MASK_SH_LIST(_MASK)
-};
-
-static struct vpg *dcn303_vpg_create(struct dc_context *ctx, uint32_t inst)
-{
-	struct dcn30_vpg *vpg3 = kzalloc(sizeof(struct dcn30_vpg), GFP_KERNEL);
-
-	if (!vpg3)
-		return NULL;
-
-	vpg3_construct(vpg3, ctx, inst, &vpg_regs[inst], &vpg_shift, &vpg_mask);
-
-	return &vpg3->base;
-}
-
-#define afmt_regs(id)\
-		[id] = { AFMT_DCN3_REG_LIST(id) }
-
-static const struct dcn30_afmt_registers afmt_regs[] = {
-		afmt_regs(0),
-		afmt_regs(1),
-		afmt_regs(2)
-};
-
-static const struct dcn30_afmt_shift afmt_shift = {
-		DCN3_AFMT_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dcn30_afmt_mask afmt_mask = {
-		DCN3_AFMT_MASK_SH_LIST(_MASK)
-};
-
-static struct afmt *dcn303_afmt_create(struct dc_context *ctx, uint32_t inst)
-{
-	struct dcn30_afmt *afmt3 = kzalloc(sizeof(struct dcn30_afmt), GFP_KERNEL);
-
-	if (!afmt3)
-		return NULL;
-
-	afmt3_construct(afmt3, ctx, inst, &afmt_regs[inst], &afmt_shift, &afmt_mask);
-
-	return &afmt3->base;
-}
-
-#define audio_regs(id)\
-		[id] = { AUD_COMMON_REG_LIST(id) }
-
-static const struct dce_audio_registers audio_regs[] = {
-		audio_regs(0),
-		audio_regs(1),
-		audio_regs(2),
-		audio_regs(3),
-		audio_regs(4),
-		audio_regs(5),
-		audio_regs(6)
-};
-
-#define DCE120_AUD_COMMON_MASK_SH_LIST(mask_sh)\
-		SF(AZF0ENDPOINT0_AZALIA_F0_CODEC_ENDPOINT_INDEX, AZALIA_ENDPOINT_REG_INDEX, mask_sh),\
-		SF(AZF0ENDPOINT0_AZALIA_F0_CODEC_ENDPOINT_DATA, AZALIA_ENDPOINT_REG_DATA, mask_sh),\
-		AUD_COMMON_MASK_SH_LIST_BASE(mask_sh)
-
-static const struct dce_audio_shift audio_shift = {
-		DCE120_AUD_COMMON_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dce_audio_mask audio_mask = {
-		DCE120_AUD_COMMON_MASK_SH_LIST(_MASK)
-};
-
-static struct audio *dcn303_create_audio(struct dc_context *ctx, unsigned int inst)
-{
-	return dce_audio_create(ctx, inst, &audio_regs[inst], &audio_shift, &audio_mask);
-}
-
-#define stream_enc_regs(id)\
-		[id] = { SE_DCN3_REG_LIST(id) }
-
-static const struct dcn10_stream_enc_registers stream_enc_regs[] = {
-		stream_enc_regs(0),
-		stream_enc_regs(1)
-};
-
-static const struct dcn10_stream_encoder_shift se_shift = {
-		SE_COMMON_MASK_SH_LIST_DCN30(__SHIFT)
-};
-
-static const struct dcn10_stream_encoder_mask se_mask = {
-		SE_COMMON_MASK_SH_LIST_DCN30(_MASK)
-};
-
-static struct stream_encoder *dcn303_stream_encoder_create(enum engine_id eng_id, struct dc_context *ctx)
-{
-	struct dcn10_stream_encoder *enc1;
-	struct vpg *vpg;
-	struct afmt *afmt;
-	int vpg_inst;
-	int afmt_inst;
-
-	/* Mapping of VPG, AFMT, DME register blocks to DIO block instance */
-	if (eng_id <= ENGINE_ID_DIGE) {
-		vpg_inst = eng_id;
-		afmt_inst = eng_id;
-	} else
-		return NULL;
-
-	enc1 = kzalloc(sizeof(struct dcn10_stream_encoder), GFP_KERNEL);
-	vpg = dcn303_vpg_create(ctx, vpg_inst);
-	afmt = dcn303_afmt_create(ctx, afmt_inst);
-
-	if (!enc1 || !vpg || !afmt) {
-		kfree(enc1);
-		kfree(vpg);
-		kfree(afmt);
-		return NULL;
-	}
-
-	dcn30_dio_stream_encoder_construct(enc1, ctx, ctx->dc_bios, eng_id, vpg, afmt, &stream_enc_regs[eng_id],
-			&se_shift, &se_mask);
-
-	return &enc1->base;
-}
-
-#define clk_src_regs(index, pllid)\
-		[index] = { CS_COMMON_REG_LIST_DCN3_03(index, pllid) }
-
-static const struct dce110_clk_src_regs clk_src_regs[] = {
-		clk_src_regs(0, A),
-		clk_src_regs(1, B)
-};
-
-static const struct dce110_clk_src_shift cs_shift = {
-		CS_COMMON_MASK_SH_LIST_DCN2_0(__SHIFT)
-};
-
-static const struct dce110_clk_src_mask cs_mask = {
-		CS_COMMON_MASK_SH_LIST_DCN2_0(_MASK)
-};
-
-static struct clock_source *dcn303_clock_source_create(struct dc_context *ctx, struct dc_bios *bios,
-		enum clock_source_id id, const struct dce110_clk_src_regs *regs, bool dp_clk_src)
-{
-	struct dce110_clk_src *clk_src = kzalloc(sizeof(struct dce110_clk_src), GFP_KERNEL);
-
-	if (!clk_src)
-		return NULL;
-
-	if (dcn3_clk_src_construct(clk_src, ctx, bios, id, regs, &cs_shift, &cs_mask)) {
-		clk_src->base.dp_clk_src = dp_clk_src;
-		return &clk_src->base;
-	}
-
-	BREAK_TO_DEBUGGER();
-	return NULL;
-}
-
-static const struct dce_hwseq_registers hwseq_reg = {
-		HWSEQ_DCN303_REG_LIST()
-};
-
-static const struct dce_hwseq_shift hwseq_shift = {
-		HWSEQ_DCN303_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dce_hwseq_mask hwseq_mask = {
-		HWSEQ_DCN303_MASK_SH_LIST(_MASK)
-};
-
-static struct dce_hwseq *dcn303_hwseq_create(struct dc_context *ctx)
-{
-	struct dce_hwseq *hws = kzalloc(sizeof(struct dce_hwseq), GFP_KERNEL);
-
-	if (hws) {
-		hws->ctx = ctx;
-		hws->regs = &hwseq_reg;
-		hws->shifts = &hwseq_shift;
-		hws->masks = &hwseq_mask;
-	}
-	return hws;
-}
-
-#define hubp_regs(id)\
-		[id] = { HUBP_REG_LIST_DCN30(id) }
-
-static const struct dcn_hubp2_registers hubp_regs[] = {
-		hubp_regs(0),
-		hubp_regs(1)
-};
-
-static const struct dcn_hubp2_shift hubp_shift = {
-		HUBP_MASK_SH_LIST_DCN30(__SHIFT)
-};
-
-static const struct dcn_hubp2_mask hubp_mask = {
-		HUBP_MASK_SH_LIST_DCN30(_MASK)
-};
-
-static struct hubp *dcn303_hubp_create(struct dc_context *ctx, uint32_t inst)
-{
-	struct dcn20_hubp *hubp2 = kzalloc(sizeof(struct dcn20_hubp), GFP_KERNEL);
-
-	if (!hubp2)
-		return NULL;
-
-	if (hubp3_construct(hubp2, ctx, inst, &hubp_regs[inst], &hubp_shift, &hubp_mask))
-		return &hubp2->base;
-
-	BREAK_TO_DEBUGGER();
-	kfree(hubp2);
-	return NULL;
-}
-
-#define dpp_regs(id)\
-		[id] = { DPP_REG_LIST_DCN30(id) }
-
-static const struct dcn3_dpp_registers dpp_regs[] = {
-		dpp_regs(0),
-		dpp_regs(1)
-};
-
-static const struct dcn3_dpp_shift tf_shift = {
-		DPP_REG_LIST_SH_MASK_DCN30(__SHIFT)
-};
-
-static const struct dcn3_dpp_mask tf_mask = {
-		DPP_REG_LIST_SH_MASK_DCN30(_MASK)
-};
-
-static struct dpp *dcn303_dpp_create(struct dc_context *ctx, uint32_t inst)
-{
-	struct dcn3_dpp *dpp = kzalloc(sizeof(struct dcn3_dpp), GFP_KERNEL);
-
-	if (!dpp)
-		return NULL;
-
-	if (dpp3_construct(dpp, ctx, inst, &dpp_regs[inst], &tf_shift, &tf_mask))
-		return &dpp->base;
-
-	BREAK_TO_DEBUGGER();
-	kfree(dpp);
-	return NULL;
-}
-
-#define opp_regs(id)\
-		[id] = { OPP_REG_LIST_DCN30(id) }
-
-static const struct dcn20_opp_registers opp_regs[] = {
-		opp_regs(0),
-		opp_regs(1)
-};
-
-static const struct dcn20_opp_shift opp_shift = {
-		OPP_MASK_SH_LIST_DCN20(__SHIFT)
-};
-
-static const struct dcn20_opp_mask opp_mask = {
-		OPP_MASK_SH_LIST_DCN20(_MASK)
-};
-
-static struct output_pixel_processor *dcn303_opp_create(struct dc_context *ctx, uint32_t inst)
-{
-	struct dcn20_opp *opp = kzalloc(sizeof(struct dcn20_opp), GFP_KERNEL);
-
-	if (!opp) {
-		BREAK_TO_DEBUGGER();
-		return NULL;
-	}
-
-	dcn20_opp_construct(opp, ctx, inst, &opp_regs[inst], &opp_shift, &opp_mask);
-	return &opp->base;
-}
-
-#define optc_regs(id)\
-		[id] = { OPTC_COMMON_REG_LIST_DCN3_0(id) }
-
-static const struct dcn_optc_registers optc_regs[] = {
-		optc_regs(0),
-		optc_regs(1)
-};
-
-static const struct dcn_optc_shift optc_shift = {
-		OPTC_COMMON_MASK_SH_LIST_DCN30(__SHIFT)
-};
-
-static const struct dcn_optc_mask optc_mask = {
-		OPTC_COMMON_MASK_SH_LIST_DCN30(_MASK)
-};
-
-static struct timing_generator *dcn303_timing_generator_create(struct dc_context *ctx, uint32_t instance)
-{
-	struct optc *tgn10 = kzalloc(sizeof(struct optc), GFP_KERNEL);
-
-	if (!tgn10)
-		return NULL;
-
-	tgn10->base.inst = instance;
-	tgn10->base.ctx = ctx;
-
-	tgn10->tg_regs = &optc_regs[instance];
-	tgn10->tg_shift = &optc_shift;
-	tgn10->tg_mask = &optc_mask;
-
-	dcn30_timing_generator_init(tgn10);
-
-	return &tgn10->base;
-}
-
-static const struct dcn30_mpc_registers mpc_regs = {
-		MPC_REG_LIST_DCN3_0(0),
-		MPC_REG_LIST_DCN3_0(1),
-		MPC_OUT_MUX_REG_LIST_DCN3_0(0),
-		MPC_OUT_MUX_REG_LIST_DCN3_0(1),
-		MPC_RMU_GLOBAL_REG_LIST_DCN3AG,
-		MPC_RMU_REG_LIST_DCN3AG(0),
-		MPC_DWB_MUX_REG_LIST_DCN3_0(0),
-};
-
-static const struct dcn30_mpc_shift mpc_shift = {
-		MPC_COMMON_MASK_SH_LIST_DCN303(__SHIFT)
-};
-
-static const struct dcn30_mpc_mask mpc_mask = {
-		MPC_COMMON_MASK_SH_LIST_DCN303(_MASK)
-};
-
-static struct mpc *dcn303_mpc_create(struct dc_context *ctx, int num_mpcc, int num_rmu)
-{
-	struct dcn30_mpc *mpc30 = kzalloc(sizeof(struct dcn30_mpc), GFP_KERNEL);
-
-	if (!mpc30)
-		return NULL;
-
-	dcn30_mpc_construct(mpc30, ctx, &mpc_regs, &mpc_shift, &mpc_mask, num_mpcc, num_rmu);
-
-	return &mpc30->base;
-}
-
-#define dsc_regsDCN20(id)\
-[id] = { DSC_REG_LIST_DCN20(id) }
-
-static const struct dcn20_dsc_registers dsc_regs[] = {
-		dsc_regsDCN20(0),
-		dsc_regsDCN20(1)
-};
-
-static const struct dcn20_dsc_shift dsc_shift = {
-		DSC_REG_LIST_SH_MASK_DCN20(__SHIFT)
-};
-
-static const struct dcn20_dsc_mask dsc_mask = {
-		DSC_REG_LIST_SH_MASK_DCN20(_MASK)
-};
-
-static struct display_stream_compressor *dcn303_dsc_create(struct dc_context *ctx, uint32_t inst)
-{
-	struct dcn20_dsc *dsc = kzalloc(sizeof(struct dcn20_dsc), GFP_KERNEL);
-
-	if (!dsc) {
-		BREAK_TO_DEBUGGER();
-		return NULL;
-	}
-
-	dsc2_construct(dsc, ctx, inst, &dsc_regs[inst], &dsc_shift, &dsc_mask);
-	return &dsc->base;
-}
-
-#define dwbc_regs_dcn3(id)\
-[id] = { DWBC_COMMON_REG_LIST_DCN30(id) }
-
-static const struct dcn30_dwbc_registers dwbc30_regs[] = {
-		dwbc_regs_dcn3(0)
-};
-
-static const struct dcn30_dwbc_shift dwbc30_shift = {
-		DWBC_COMMON_MASK_SH_LIST_DCN30(__SHIFT)
-};
-
-static const struct dcn30_dwbc_mask dwbc30_mask = {
-		DWBC_COMMON_MASK_SH_LIST_DCN30(_MASK)
-};
-
-static bool dcn303_dwbc_create(struct dc_context *ctx, struct resource_pool *pool)
-{
-	int i;
-	uint32_t pipe_count = pool->res_cap->num_dwb;
-
-	for (i = 0; i < pipe_count; i++) {
-		struct dcn30_dwbc *dwbc30 = kzalloc(sizeof(struct dcn30_dwbc), GFP_KERNEL);
-
-		if (!dwbc30) {
-			dm_error("DC: failed to create dwbc30!\n");
-			return false;
-		}
-
-		dcn30_dwbc_construct(dwbc30, ctx, &dwbc30_regs[i], &dwbc30_shift, &dwbc30_mask, i);
-
-		pool->dwbc[i] = &dwbc30->base;
-	}
-	return true;
-}
-
-#define mcif_wb_regs_dcn3(id)\
-[id] = { MCIF_WB_COMMON_REG_LIST_DCN30(id) }
-
-static const struct dcn30_mmhubbub_registers mcif_wb30_regs[] = {
-		mcif_wb_regs_dcn3(0)
-};
-
-static const struct dcn30_mmhubbub_shift mcif_wb30_shift = {
-		MCIF_WB_COMMON_MASK_SH_LIST_DCN30(__SHIFT)
-};
-
-static const struct dcn30_mmhubbub_mask mcif_wb30_mask = {
-		MCIF_WB_COMMON_MASK_SH_LIST_DCN30(_MASK)
-};
-
-static bool dcn303_mmhubbub_create(struct dc_context *ctx, struct resource_pool *pool)
-{
-	int i;
-	uint32_t pipe_count = pool->res_cap->num_dwb;
-
-	for (i = 0; i < pipe_count; i++) {
-		struct dcn30_mmhubbub *mcif_wb30 = kzalloc(sizeof(struct dcn30_mmhubbub), GFP_KERNEL);
-
-		if (!mcif_wb30) {
-			dm_error("DC: failed to create mcif_wb30!\n");
-			return false;
-		}
-
-		dcn30_mmhubbub_construct(mcif_wb30, ctx, &mcif_wb30_regs[i], &mcif_wb30_shift, &mcif_wb30_mask, i);
-
-		pool->mcif_wb[i] = &mcif_wb30->base;
-	}
-	return true;
-}
-
-#define aux_engine_regs(id)\
-[id] = {\
-		AUX_COMMON_REG_LIST0(id), \
-		.AUXN_IMPCAL = 0, \
-		.AUXP_IMPCAL = 0, \
-		.AUX_RESET_MASK = DP_AUX0_AUX_CONTROL__AUX_RESET_MASK, \
-}
-
-static const struct dce110_aux_registers aux_engine_regs[] = {
-		aux_engine_regs(0),
-		aux_engine_regs(1)
-};
-
-static const struct dce110_aux_registers_shift aux_shift = {
-		DCN_AUX_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dce110_aux_registers_mask aux_mask = {
-		DCN_AUX_MASK_SH_LIST(_MASK)
-};
-
-static struct dce_aux *dcn303_aux_engine_create(struct dc_context *ctx, uint32_t inst)
-{
-	struct aux_engine_dce110 *aux_engine = kzalloc(sizeof(struct aux_engine_dce110), GFP_KERNEL);
-
-	if (!aux_engine)
-		return NULL;
-
-	dce110_aux_engine_construct(aux_engine, ctx, inst, SW_AUX_TIMEOUT_PERIOD_MULTIPLIER * AUX_TIMEOUT_PERIOD,
-			&aux_engine_regs[inst], &aux_mask, &aux_shift, ctx->dc->caps.extended_aux_timeout_support);
-
-	return &aux_engine->base;
-}
-
-#define i2c_inst_regs(id) { I2C_HW_ENGINE_COMMON_REG_LIST(id) }
-
-static const struct dce_i2c_registers i2c_hw_regs[] = {
-		i2c_inst_regs(1),
-		i2c_inst_regs(2)
-};
-
-static const struct dce_i2c_shift i2c_shifts = {
-		I2C_COMMON_MASK_SH_LIST_DCN2(__SHIFT)
-};
-
-static const struct dce_i2c_mask i2c_masks = {
-		I2C_COMMON_MASK_SH_LIST_DCN2(_MASK)
-};
-
-static struct dce_i2c_hw *dcn303_i2c_hw_create(struct dc_context *ctx, uint32_t inst)
-{
-	struct dce_i2c_hw *dce_i2c_hw = kzalloc(sizeof(struct dce_i2c_hw), GFP_KERNEL);
-
-	if (!dce_i2c_hw)
-		return NULL;
-
-	dcn2_i2c_hw_construct(dce_i2c_hw, ctx, inst, &i2c_hw_regs[inst], &i2c_shifts, &i2c_masks);
-
-	return dce_i2c_hw;
-}
-
-static const struct encoder_feature_support link_enc_feature = {
-		.max_hdmi_deep_color = COLOR_DEPTH_121212,
-		.max_hdmi_pixel_clock = 600000,
-		.hdmi_ycbcr420_supported = true,
-		.dp_ycbcr420_supported = true,
-		.fec_supported = true,
-		.flags.bits.IS_HBR2_CAPABLE = true,
-		.flags.bits.IS_HBR3_CAPABLE = true,
-		.flags.bits.IS_TPS3_CAPABLE = true,
-		.flags.bits.IS_TPS4_CAPABLE = true
-};
-
-#define link_regs(id, phyid)\
-		[id] = {\
-				LE_DCN3_REG_LIST(id), \
-				UNIPHY_DCN2_REG_LIST(phyid), \
-				SRI(DP_DPHY_INTERNAL_CTRL, DP, id) \
-		}
-
-static const struct dcn10_link_enc_registers link_enc_regs[] = {
-		link_regs(0, A),
-		link_regs(1, B)
-};
-
-static const struct dcn10_link_enc_shift le_shift = {
-		LINK_ENCODER_MASK_SH_LIST_DCN30(__SHIFT),
-		DPCS_DCN2_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dcn10_link_enc_mask le_mask = {
-		LINK_ENCODER_MASK_SH_LIST_DCN30(_MASK),
-		DPCS_DCN2_MASK_SH_LIST(_MASK)
-};
-
-#define aux_regs(id)\
-		[id] = { DCN2_AUX_REG_LIST(id) }
-
-static const struct dcn10_link_enc_aux_registers link_enc_aux_regs[] = {
-		aux_regs(0),
-		aux_regs(1)
-};
-
-#define hpd_regs(id)\
-		[id] = { HPD_REG_LIST(id) }
-
-static const struct dcn10_link_enc_hpd_registers link_enc_hpd_regs[] = {
-		hpd_regs(0),
-		hpd_regs(1)
-};
-
-static struct link_encoder *dcn303_link_encoder_create(const struct encoder_init_data *enc_init_data)
-{
-	struct dcn20_link_encoder *enc20 = kzalloc(sizeof(struct dcn20_link_encoder), GFP_KERNEL);
-
-	if (!enc20)
-		return NULL;
-
-	dcn30_link_encoder_construct(enc20, enc_init_data, &link_enc_feature,
-			&link_enc_regs[enc_init_data->transmitter], &link_enc_aux_regs[enc_init_data->channel - 1],
-			&link_enc_hpd_regs[enc_init_data->hpd_source], &le_shift, &le_mask);
-
-	return &enc20->enc10.base;
-}
-
-static const struct dce_panel_cntl_registers panel_cntl_regs[] = {
-		{ DCN_PANEL_CNTL_REG_LIST() }
-};
-
-static const struct dce_panel_cntl_shift panel_cntl_shift = {
-		DCE_PANEL_CNTL_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dce_panel_cntl_mask panel_cntl_mask = {
-		DCE_PANEL_CNTL_MASK_SH_LIST(_MASK)
-};
-
-static struct panel_cntl *dcn303_panel_cntl_create(const struct panel_cntl_init_data *init_data)
-{
-	struct dce_panel_cntl *panel_cntl = kzalloc(sizeof(struct dce_panel_cntl), GFP_KERNEL);
-
-	if (!panel_cntl)
-		return NULL;
-
-	dce_panel_cntl_construct(panel_cntl, init_data, &panel_cntl_regs[init_data->inst],
-			&panel_cntl_shift, &panel_cntl_mask);
-
-	return &panel_cntl->base;
-}
-
-static void read_dce_straps(struct dc_context *ctx, struct resource_straps *straps)
-{
-	generic_reg_get(ctx, mmDC_PINSTRAPS + BASE(mmDC_PINSTRAPS_BASE_IDX),
-			FN(DC_PINSTRAPS, DC_PINSTRAPS_AUDIO), &straps->dc_pinstraps_audio);
-}
-
-static const struct resource_create_funcs res_create_funcs = {
-		.read_dce_straps = read_dce_straps,
-		.create_audio = dcn303_create_audio,
-		.create_stream_encoder = dcn303_stream_encoder_create,
-		.create_hwseq = dcn303_hwseq_create,
-};
-
-static const struct resource_create_funcs res_create_maximus_funcs = {
-		.read_dce_straps = NULL,
-		.create_audio = NULL,
-		.create_stream_encoder = NULL,
-		.create_hwseq = dcn303_hwseq_create,
-};
-
-static bool is_soc_bounding_box_valid(struct dc *dc)
-{
-	uint32_t hw_internal_rev = dc->ctx->asic_id.hw_internal_rev;
-
-	if (ASICREV_IS_BEIGE_GOBY_P(hw_internal_rev))
-		return true;
-
-	return false;
-}
-
-static bool init_soc_bounding_box(struct dc *dc,  struct resource_pool *pool)
-{
-	struct _vcs_dpi_soc_bounding_box_st *loaded_bb = &dcn3_03_soc;
-	struct _vcs_dpi_ip_params_st *loaded_ip = &dcn3_03_ip;
-
-	DC_LOGGER_INIT(dc->ctx->logger);
-
-	if (!is_soc_bounding_box_valid(dc)) {
-		DC_LOG_ERROR("%s: not valid soc bounding box/n", __func__);
-		return false;
-	}
-
-	loaded_ip->max_num_otg = pool->pipe_count;
-	loaded_ip->max_num_dpp = pool->pipe_count;
-	loaded_ip->clamp_min_dcfclk = dc->config.clamp_min_dcfclk;
-	dcn20_patch_bounding_box(dc, loaded_bb);
-
-	if (dc->ctx->dc_bios->funcs->get_soc_bb_info) {
-		struct bp_soc_bb_info bb_info = { 0 };
-
-		if (dc->ctx->dc_bios->funcs->get_soc_bb_info(
-			    dc->ctx->dc_bios, &bb_info) == BP_RESULT_OK) {
-			if (bb_info.dram_clock_change_latency_100ns > 0)
-				dcn3_03_soc.dram_clock_change_latency_us =
-					bb_info.dram_clock_change_latency_100ns * 10;
-
-			if (bb_info.dram_sr_enter_exit_latency_100ns > 0)
-				dcn3_03_soc.sr_enter_plus_exit_time_us =
-					bb_info.dram_sr_enter_exit_latency_100ns * 10;
-
-			if (bb_info.dram_sr_exit_latency_100ns > 0)
-				dcn3_03_soc.sr_exit_time_us =
-					bb_info.dram_sr_exit_latency_100ns * 10;
-		}
-	}
-
-	return true;
-}
-
-static void dcn303_resource_destruct(struct resource_pool *pool)
-{
-	unsigned int i;
-
-	for (i = 0; i < pool->stream_enc_count; i++) {
-		if (pool->stream_enc[i] != NULL) {
-			if (pool->stream_enc[i]->vpg != NULL) {
-				kfree(DCN30_VPG_FROM_VPG(pool->stream_enc[i]->vpg));
-				pool->stream_enc[i]->vpg = NULL;
-			}
-			if (pool->stream_enc[i]->afmt != NULL) {
-				kfree(DCN30_AFMT_FROM_AFMT(pool->stream_enc[i]->afmt));
-				pool->stream_enc[i]->afmt = NULL;
-			}
-			kfree(DCN10STRENC_FROM_STRENC(pool->stream_enc[i]));
-			pool->stream_enc[i] = NULL;
-		}
-	}
-
-	for (i = 0; i < pool->res_cap->num_dsc; i++) {
-		if (pool->dscs[i] != NULL)
-			dcn20_dsc_destroy(&pool->dscs[i]);
-	}
-
-	if (pool->mpc != NULL) {
-		kfree(TO_DCN20_MPC(pool->mpc));
-		pool->mpc = NULL;
-	}
-
-	if (pool->hubbub != NULL) {
-		kfree(pool->hubbub);
-		pool->hubbub = NULL;
-	}
-
-	for (i = 0; i < pool->pipe_count; i++) {
-		if (pool->dpps[i] != NULL) {
-			kfree(TO_DCN20_DPP(pool->dpps[i]));
-			pool->dpps[i] = NULL;
-		}
-
-		if (pool->hubps[i] != NULL) {
-			kfree(TO_DCN20_HUBP(pool->hubps[i]));
-			pool->hubps[i] = NULL;
-		}
-
-		if (pool->irqs != NULL)
-			dal_irq_service_destroy(&pool->irqs);
-	}
-
-	for (i = 0; i < pool->res_cap->num_ddc; i++) {
-		if (pool->engines[i] != NULL)
-			dce110_engine_destroy(&pool->engines[i]);
-		if (pool->hw_i2cs[i] != NULL) {
-			kfree(pool->hw_i2cs[i]);
-			pool->hw_i2cs[i] = NULL;
-		}
-		if (pool->sw_i2cs[i] != NULL) {
-			kfree(pool->sw_i2cs[i]);
-			pool->sw_i2cs[i] = NULL;
-		}
-	}
-
-	for (i = 0; i < pool->res_cap->num_opp; i++) {
-		if (pool->opps[i] != NULL)
-			pool->opps[i]->funcs->opp_destroy(&pool->opps[i]);
-	}
-
-	for (i = 0; i < pool->res_cap->num_timing_generator; i++) {
-		if (pool->timing_generators[i] != NULL)	{
-			kfree(DCN10TG_FROM_TG(pool->timing_generators[i]));
-			pool->timing_generators[i] = NULL;
-		}
-	}
-
-	for (i = 0; i < pool->res_cap->num_dwb; i++) {
-		if (pool->dwbc[i] != NULL) {
-			kfree(TO_DCN30_DWBC(pool->dwbc[i]));
-			pool->dwbc[i] = NULL;
-		}
-		if (pool->mcif_wb[i] != NULL) {
-			kfree(TO_DCN30_MMHUBBUB(pool->mcif_wb[i]));
-			pool->mcif_wb[i] = NULL;
-		}
-	}
-
-	for (i = 0; i < pool->audio_count; i++) {
-		if (pool->audios[i])
-			dce_aud_destroy(&pool->audios[i]);
-	}
-
-	for (i = 0; i < pool->clk_src_count; i++) {
-		if (pool->clock_sources[i] != NULL)
-			dcn20_clock_source_destroy(&pool->clock_sources[i]);
-	}
-
-	if (pool->dp_clock_source != NULL)
-		dcn20_clock_source_destroy(&pool->dp_clock_source);
-
-	for (i = 0; i < pool->res_cap->num_mpc_3dlut; i++) {
-		if (pool->mpc_lut[i] != NULL) {
-			dc_3dlut_func_release(pool->mpc_lut[i]);
-			pool->mpc_lut[i] = NULL;
-		}
-		if (pool->mpc_shaper[i] != NULL) {
-			dc_transfer_func_release(pool->mpc_shaper[i]);
-			pool->mpc_shaper[i] = NULL;
-		}
-	}
-
-	for (i = 0; i < pool->pipe_count; i++) {
-		if (pool->multiple_abms[i] != NULL)
-			dce_abm_destroy(&pool->multiple_abms[i]);
-	}
-
-	if (pool->psr != NULL)
-		dmub_psr_destroy(&pool->psr);
-
-	if (pool->dccg != NULL)
-		dcn_dccg_destroy(&pool->dccg);
-
-	if (pool->oem_device != NULL)
-		dal_ddc_service_destroy(&pool->oem_device);
-}
-
-static void dcn303_destroy_resource_pool(struct resource_pool **pool)
-{
-	dcn303_resource_destruct(*pool);
-	kfree(*pool);
-	*pool = NULL;
-}
-
-static void dcn303_get_optimal_dcfclk_fclk_for_uclk(unsigned int uclk_mts,
-		unsigned int *optimal_dcfclk,
-		unsigned int *optimal_fclk)
-{
-	double bw_from_dram, bw_from_dram1, bw_from_dram2;
-
-	bw_from_dram1 = uclk_mts * dcn3_03_soc.num_chans *
-		dcn3_03_soc.dram_channel_width_bytes * (dcn3_03_soc.max_avg_dram_bw_use_normal_percent / 100);
-	bw_from_dram2 = uclk_mts * dcn3_03_soc.num_chans *
-		dcn3_03_soc.dram_channel_width_bytes * (dcn3_03_soc.max_avg_sdp_bw_use_normal_percent / 100);
-
-	bw_from_dram = (bw_from_dram1 < bw_from_dram2) ? bw_from_dram1 : bw_from_dram2;
-
-	if (optimal_fclk)
-		*optimal_fclk = bw_from_dram /
-		(dcn3_03_soc.fabric_datapath_to_dcn_data_return_bytes *
-				(dcn3_03_soc.max_avg_sdp_bw_use_normal_percent / 100));
-
-	if (optimal_dcfclk)
-		*optimal_dcfclk =  bw_from_dram /
-		(dcn3_03_soc.return_bus_width_bytes * (dcn3_03_soc.max_avg_sdp_bw_use_normal_percent / 100));
-}
-
-void dcn303_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params)
-{
-	unsigned int i, j;
-	unsigned int num_states = 0;
-
-	unsigned int dcfclk_mhz[DC__VOLTAGE_STATES] = {0};
-	unsigned int dram_speed_mts[DC__VOLTAGE_STATES] = {0};
-	unsigned int optimal_uclk_for_dcfclk_sta_targets[DC__VOLTAGE_STATES] = {0};
-	unsigned int optimal_dcfclk_for_uclk[DC__VOLTAGE_STATES] = {0};
-
-	unsigned int dcfclk_sta_targets[DC__VOLTAGE_STATES] = {694, 875, 1000, 1200};
-	unsigned int num_dcfclk_sta_targets = 4;
-	unsigned int num_uclk_states;
-
-
-	if (dc->ctx->dc_bios->vram_info.num_chans)
-		dcn3_03_soc.num_chans = dc->ctx->dc_bios->vram_info.num_chans;
-
-	if (dc->ctx->dc_bios->vram_info.dram_channel_width_bytes)
-		dcn3_03_soc.dram_channel_width_bytes = dc->ctx->dc_bios->vram_info.dram_channel_width_bytes;
-
-	dcn3_03_soc.dispclk_dppclk_vco_speed_mhz = dc->clk_mgr->dentist_vco_freq_khz / 1000.0;
-	dc->dml.soc.dispclk_dppclk_vco_speed_mhz = dc->clk_mgr->dentist_vco_freq_khz / 1000.0;
-
-	if (bw_params->clk_table.entries[0].memclk_mhz) {
-		int max_dcfclk_mhz = 0, max_dispclk_mhz = 0, max_dppclk_mhz = 0, max_phyclk_mhz = 0;
-
-		for (i = 0; i < MAX_NUM_DPM_LVL; i++) {
-			if (bw_params->clk_table.entries[i].dcfclk_mhz > max_dcfclk_mhz)
-				max_dcfclk_mhz = bw_params->clk_table.entries[i].dcfclk_mhz;
-			if (bw_params->clk_table.entries[i].dispclk_mhz > max_dispclk_mhz)
-				max_dispclk_mhz = bw_params->clk_table.entries[i].dispclk_mhz;
-			if (bw_params->clk_table.entries[i].dppclk_mhz > max_dppclk_mhz)
-				max_dppclk_mhz = bw_params->clk_table.entries[i].dppclk_mhz;
-			if (bw_params->clk_table.entries[i].phyclk_mhz > max_phyclk_mhz)
-				max_phyclk_mhz = bw_params->clk_table.entries[i].phyclk_mhz;
-		}
-		if (!max_dcfclk_mhz)
-			max_dcfclk_mhz = dcn3_03_soc.clock_limits[0].dcfclk_mhz;
-		if (!max_dispclk_mhz)
-			max_dispclk_mhz = dcn3_03_soc.clock_limits[0].dispclk_mhz;
-		if (!max_dppclk_mhz)
-			max_dppclk_mhz = dcn3_03_soc.clock_limits[0].dppclk_mhz;
-		if (!max_phyclk_mhz)
-			max_phyclk_mhz = dcn3_03_soc.clock_limits[0].phyclk_mhz;
-
-		if (max_dcfclk_mhz > dcfclk_sta_targets[num_dcfclk_sta_targets-1]) {
-			dcfclk_sta_targets[num_dcfclk_sta_targets] = max_dcfclk_mhz;
-			num_dcfclk_sta_targets++;
-		} else if (max_dcfclk_mhz < dcfclk_sta_targets[num_dcfclk_sta_targets-1]) {
-			for (i = 0; i < num_dcfclk_sta_targets; i++) {
-				if (dcfclk_sta_targets[i] > max_dcfclk_mhz) {
-					dcfclk_sta_targets[i] = max_dcfclk_mhz;
-					break;
-				}
-			}
-			/* Update size of array since we "removed" duplicates */
-			num_dcfclk_sta_targets = i + 1;
-		}
-
-		num_uclk_states = bw_params->clk_table.num_entries;
-
-		/* Calculate optimal dcfclk for each uclk */
-		for (i = 0; i < num_uclk_states; i++) {
-			dcn303_get_optimal_dcfclk_fclk_for_uclk(bw_params->clk_table.entries[i].memclk_mhz * 16,
-					&optimal_dcfclk_for_uclk[i], NULL);
-			if (optimal_dcfclk_for_uclk[i] < bw_params->clk_table.entries[0].dcfclk_mhz)
-				optimal_dcfclk_for_uclk[i] = bw_params->clk_table.entries[0].dcfclk_mhz;
-		}
-
-		/* Calculate optimal uclk for each dcfclk sta target */
-		for (i = 0; i < num_dcfclk_sta_targets; i++) {
-			for (j = 0; j < num_uclk_states; j++) {
-				if (dcfclk_sta_targets[i] < optimal_dcfclk_for_uclk[j]) {
-					optimal_uclk_for_dcfclk_sta_targets[i] =
-							bw_params->clk_table.entries[j].memclk_mhz * 16;
-					break;
-				}
-			}
-		}
-
-		i = 0;
-		j = 0;
-		/* create the final dcfclk and uclk table */
-		while (i < num_dcfclk_sta_targets && j < num_uclk_states && num_states < DC__VOLTAGE_STATES) {
-			if (dcfclk_sta_targets[i] < optimal_dcfclk_for_uclk[j] && i < num_dcfclk_sta_targets) {
-				dcfclk_mhz[num_states] = dcfclk_sta_targets[i];
-				dram_speed_mts[num_states++] = optimal_uclk_for_dcfclk_sta_targets[i++];
-			} else {
-				if (j < num_uclk_states && optimal_dcfclk_for_uclk[j] <= max_dcfclk_mhz) {
-					dcfclk_mhz[num_states] = optimal_dcfclk_for_uclk[j];
-					dram_speed_mts[num_states++] =
-							bw_params->clk_table.entries[j++].memclk_mhz * 16;
-				} else {
-					j = num_uclk_states;
-				}
-			}
-		}
-
-		while (i < num_dcfclk_sta_targets && num_states < DC__VOLTAGE_STATES) {
-			dcfclk_mhz[num_states] = dcfclk_sta_targets[i];
-			dram_speed_mts[num_states++] = optimal_uclk_for_dcfclk_sta_targets[i++];
-		}
-
-		while (j < num_uclk_states && num_states < DC__VOLTAGE_STATES &&
-				optimal_dcfclk_for_uclk[j] <= max_dcfclk_mhz) {
-			dcfclk_mhz[num_states] = optimal_dcfclk_for_uclk[j];
-			dram_speed_mts[num_states++] = bw_params->clk_table.entries[j++].memclk_mhz * 16;
-		}
-
-		dcn3_03_soc.num_states = num_states;
-		for (i = 0; i < dcn3_03_soc.num_states; i++) {
-			dcn3_03_soc.clock_limits[i].state = i;
-			dcn3_03_soc.clock_limits[i].dcfclk_mhz = dcfclk_mhz[i];
-			dcn3_03_soc.clock_limits[i].fabricclk_mhz = dcfclk_mhz[i];
-			dcn3_03_soc.clock_limits[i].dram_speed_mts = dram_speed_mts[i];
-
-			/* Fill all states with max values of all other clocks */
-			dcn3_03_soc.clock_limits[i].dispclk_mhz = max_dispclk_mhz;
-			dcn3_03_soc.clock_limits[i].dppclk_mhz  = max_dppclk_mhz;
-			dcn3_03_soc.clock_limits[i].phyclk_mhz  = max_phyclk_mhz;
-			/* Populate from bw_params for DTBCLK, SOCCLK */
-			if (!bw_params->clk_table.entries[i].dtbclk_mhz && i > 0)
-				dcn3_03_soc.clock_limits[i].dtbclk_mhz = dcn3_03_soc.clock_limits[i-1].dtbclk_mhz;
-			else
-				dcn3_03_soc.clock_limits[i].dtbclk_mhz = bw_params->clk_table.entries[i].dtbclk_mhz;
-			if (!bw_params->clk_table.entries[i].socclk_mhz && i > 0)
-				dcn3_03_soc.clock_limits[i].socclk_mhz = dcn3_03_soc.clock_limits[i-1].socclk_mhz;
-			else
-				dcn3_03_soc.clock_limits[i].socclk_mhz = bw_params->clk_table.entries[i].socclk_mhz;
-			/* These clocks cannot come from bw_params, always fill from dcn3_03_soc[1] */
-			/* FCLK, PHYCLK_D18, DSCCLK */
-			dcn3_03_soc.clock_limits[i].phyclk_d18_mhz = dcn3_03_soc.clock_limits[0].phyclk_d18_mhz;
-			dcn3_03_soc.clock_limits[i].dscclk_mhz = dcn3_03_soc.clock_limits[0].dscclk_mhz;
-		}
-
-		// WA: patch strobe modes to compensate for DCN303 BW issue
-		if (dcn3_03_soc.num_chans <= 4) {
-			for (i = 0; i < dcn3_03_soc.num_states; i++) {
-				if (dcn3_03_soc.clock_limits[i].dram_speed_mts > 1700)
-					break;
-
-				if (dcn3_03_soc.clock_limits[i].dram_speed_mts >= 1500) {
-					dcn3_03_soc.clock_limits[i].dcfclk_mhz = 100;
-					dcn3_03_soc.clock_limits[i].fabricclk_mhz = 100;
-				}
-			}
-		}
-
-		/* re-init DML with updated bb */
-		dml_init_instance(&dc->dml, &dcn3_03_soc, &dcn3_03_ip, DML_PROJECT_DCN30);
-		if (dc->current_state)
-			dml_init_instance(&dc->current_state->bw_ctx.dml, &dcn3_03_soc, &dcn3_03_ip, DML_PROJECT_DCN30);
-	}
-}
-
-static struct resource_funcs dcn303_res_pool_funcs = {
-		.destroy = dcn303_destroy_resource_pool,
-		.link_enc_create = dcn303_link_encoder_create,
-		.panel_cntl_create = dcn303_panel_cntl_create,
-		.validate_bandwidth = dcn30_validate_bandwidth,
-		.calculate_wm_and_dlg = dcn30_calculate_wm_and_dlg,
-		.update_soc_for_wm_a = dcn30_update_soc_for_wm_a,
-		.populate_dml_pipes = dcn30_populate_dml_pipes_from_context,
-		.acquire_idle_pipe_for_layer = dcn20_acquire_idle_pipe_for_layer,
-		.add_stream_to_ctx = dcn30_add_stream_to_ctx,
-		.add_dsc_to_stream_resource = dcn20_add_dsc_to_stream_resource,
-		.remove_stream_from_ctx = dcn20_remove_stream_from_ctx,
-		.populate_dml_writeback_from_context = dcn30_populate_dml_writeback_from_context,
-		.set_mcif_arb_params = dcn30_set_mcif_arb_params,
-		.find_first_free_match_stream_enc_for_link = dcn10_find_first_free_match_stream_enc_for_link,
-		.acquire_post_bldn_3dlut = dcn30_acquire_post_bldn_3dlut,
-		.release_post_bldn_3dlut = dcn30_release_post_bldn_3dlut,
-		.update_bw_bounding_box = dcn303_update_bw_bounding_box,
-		.patch_unknown_plane_state = dcn20_patch_unknown_plane_state,
-};
-
-static struct dc_cap_funcs cap_funcs = {
-		.get_dcc_compression_cap = dcn20_get_dcc_compression_cap
-};
-
-static const struct bios_registers bios_regs = {
-		NBIO_SR(BIOS_SCRATCH_3),
-		NBIO_SR(BIOS_SCRATCH_6)
-};
-
-static const struct dccg_registers dccg_regs = {
-		DCCG_REG_LIST_DCN3_03()
-};
-
-static const struct dccg_shift dccg_shift = {
-		DCCG_MASK_SH_LIST_DCN3_03(__SHIFT)
-};
-
-static const struct dccg_mask dccg_mask = {
-		DCCG_MASK_SH_LIST_DCN3_03(_MASK)
-};
-
-#define abm_regs(id)\
-		[id] = { ABM_DCN302_REG_LIST(id) }
-
-static const struct dce_abm_registers abm_regs[] = {
-		abm_regs(0),
-		abm_regs(1)
-};
-
-static const struct dce_abm_shift abm_shift = {
-		ABM_MASK_SH_LIST_DCN30(__SHIFT)
-};
-
-static const struct dce_abm_mask abm_mask = {
-		ABM_MASK_SH_LIST_DCN30(_MASK)
-};
-
-static bool dcn303_resource_construct(
-		uint8_t num_virtual_links,
-		struct dc *dc,
-		struct resource_pool *pool)
-{
-	int i;
-	struct dc_context *ctx = dc->ctx;
-	struct irq_service_init_data init_data;
-	struct ddc_service_init_data ddc_init_data;
-
-	ctx->dc_bios->regs = &bios_regs;
-
-	pool->res_cap = &res_cap_dcn303;
-
-	pool->funcs = &dcn303_res_pool_funcs;
-
-	/*************************************************
-	 *  Resource + asic cap harcoding                *
-	 *************************************************/
-	pool->underlay_pipe_index = NO_UNDERLAY_PIPE;
-	pool->pipe_count = pool->res_cap->num_timing_generator;
-	pool->mpcc_count = pool->res_cap->num_timing_generator;
-	dc->caps.max_downscale_ratio = 600;
-	dc->caps.i2c_speed_in_khz = 100;
-	dc->caps.i2c_speed_in_khz_hdcp = 5; /*1.4 w/a applied by derfault*/
-	dc->caps.max_cursor_size = 256;
-	dc->caps.min_horizontal_blanking_period = 80;
-	dc->caps.dmdata_alloc_size = 2048;
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	dc->caps.mall_size_per_mem_channel = 4;
-	/* total size = mall per channel * num channels * 1024 * 1024 */
-	dc->caps.mall_size_total = dc->caps.mall_size_per_mem_channel *
-				   dc->ctx->dc_bios->vram_info.num_chans *
-				   1024 * 1024;
-	dc->caps.cursor_cache_size =
-		dc->caps.max_cursor_size * dc->caps.max_cursor_size * 8;
-#endif
-	dc->caps.max_slave_planes = 1;
-	dc->caps.post_blend_color_processing = true;
-	dc->caps.force_dp_tps4_for_cp2520 = true;
-	dc->caps.extended_aux_timeout_support = true;
-	dc->caps.dmcub_support = true;
-
-	/* Color pipeline capabilities */
-	dc->caps.color.dpp.dcn_arch = 1;
-	dc->caps.color.dpp.input_lut_shared = 0;
-	dc->caps.color.dpp.icsc = 1;
-	dc->caps.color.dpp.dgam_ram = 0; // must use gamma_corr
-	dc->caps.color.dpp.dgam_rom_caps.srgb = 1;
-	dc->caps.color.dpp.dgam_rom_caps.bt2020 = 1;
-	dc->caps.color.dpp.dgam_rom_caps.gamma2_2 = 1;
-	dc->caps.color.dpp.dgam_rom_caps.pq = 1;
-	dc->caps.color.dpp.dgam_rom_caps.hlg = 1;
-	dc->caps.color.dpp.post_csc = 1;
-	dc->caps.color.dpp.gamma_corr = 1;
-	dc->caps.color.dpp.dgam_rom_for_yuv = 0;
-
-	dc->caps.color.dpp.hw_3d_lut = 1;
-	dc->caps.color.dpp.ogam_ram = 1;
-	// no OGAM ROM on DCN3
-	dc->caps.color.dpp.ogam_rom_caps.srgb = 0;
-	dc->caps.color.dpp.ogam_rom_caps.bt2020 = 0;
-	dc->caps.color.dpp.ogam_rom_caps.gamma2_2 = 0;
-	dc->caps.color.dpp.ogam_rom_caps.pq = 0;
-	dc->caps.color.dpp.ogam_rom_caps.hlg = 0;
-	dc->caps.color.dpp.ocsc = 0;
-
-	dc->caps.color.mpc.gamut_remap = 1;
-	dc->caps.color.mpc.num_3dluts = pool->res_cap->num_mpc_3dlut; //3
-	dc->caps.color.mpc.ogam_ram = 1;
-	dc->caps.color.mpc.ogam_rom_caps.srgb = 0;
-	dc->caps.color.mpc.ogam_rom_caps.bt2020 = 0;
-	dc->caps.color.mpc.ogam_rom_caps.gamma2_2 = 0;
-	dc->caps.color.mpc.ogam_rom_caps.pq = 0;
-	dc->caps.color.mpc.ogam_rom_caps.hlg = 0;
-	dc->caps.color.mpc.ocsc = 1;
-
-	if (dc->ctx->dce_environment == DCE_ENV_PRODUCTION_DRV)
-		dc->debug = debug_defaults_drv;
-	else
-		dc->debug = debug_defaults_diags;
-
-	// Init the vm_helper
-	if (dc->vm_helper)
-		vm_helper_init(dc->vm_helper, 16);
-
-	/*************************************************
-	 *  Create resources                             *
-	 *************************************************/
-
-	/* Clock Sources for Pixel Clock*/
-	pool->clock_sources[DCN303_CLK_SRC_PLL0] =
-			dcn303_clock_source_create(ctx, ctx->dc_bios,
-					CLOCK_SOURCE_COMBO_PHY_PLL0,
-					&clk_src_regs[0], false);
-	pool->clock_sources[DCN303_CLK_SRC_PLL1] =
-			dcn303_clock_source_create(ctx, ctx->dc_bios,
-					CLOCK_SOURCE_COMBO_PHY_PLL1,
-					&clk_src_regs[1], false);
-
-	pool->clk_src_count = DCN303_CLK_SRC_TOTAL;
-
-	/* todo: not reuse phy_pll registers */
-	pool->dp_clock_source =
-			dcn303_clock_source_create(ctx, ctx->dc_bios,
-					CLOCK_SOURCE_ID_DP_DTO,
-					&clk_src_regs[0], true);
-
-	for (i = 0; i < pool->clk_src_count; i++) {
-		if (pool->clock_sources[i] == NULL) {
-			dm_error("DC: failed to create clock sources!\n");
-			BREAK_TO_DEBUGGER();
-			goto create_fail;
-		}
-	}
-
-	/* DCCG */
-	pool->dccg = dccg30_create(ctx, &dccg_regs, &dccg_shift, &dccg_mask);
-	if (pool->dccg == NULL) {
-		dm_error("DC: failed to create dccg!\n");
-		BREAK_TO_DEBUGGER();
-		goto create_fail;
-	}
-
-	/* PP Lib and SMU interfaces */
-	init_soc_bounding_box(dc, pool);
-
-	/* DML */
-	dml_init_instance(&dc->dml, &dcn3_03_soc, &dcn3_03_ip, DML_PROJECT_DCN30);
-
-	/* IRQ */
-	init_data.ctx = dc->ctx;
-	pool->irqs = dal_irq_service_dcn303_create(&init_data);
-	if (!pool->irqs)
-		goto create_fail;
-
-	/* HUBBUB */
-	pool->hubbub = dcn303_hubbub_create(ctx);
-	if (pool->hubbub == NULL) {
-		BREAK_TO_DEBUGGER();
-		dm_error("DC: failed to create hubbub!\n");
-		goto create_fail;
-	}
-
-	/* HUBPs, DPPs, OPPs and TGs */
-	for (i = 0; i < pool->pipe_count; i++) {
-		pool->hubps[i] = dcn303_hubp_create(ctx, i);
-		if (pool->hubps[i] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error("DC: failed to create hubps!\n");
-			goto create_fail;
-		}
-
-		pool->dpps[i] = dcn303_dpp_create(ctx, i);
-		if (pool->dpps[i] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error("DC: failed to create dpps!\n");
-			goto create_fail;
-		}
-	}
-
-	for (i = 0; i < pool->res_cap->num_opp; i++) {
-		pool->opps[i] = dcn303_opp_create(ctx, i);
-		if (pool->opps[i] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error("DC: failed to create output pixel processor!\n");
-			goto create_fail;
-		}
-	}
-
-	for (i = 0; i < pool->res_cap->num_timing_generator; i++) {
-		pool->timing_generators[i] = dcn303_timing_generator_create(ctx, i);
-		if (pool->timing_generators[i] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error("DC: failed to create tg!\n");
-			goto create_fail;
-		}
-	}
-	pool->timing_generator_count = i;
-
-	/* PSR */
-	pool->psr = dmub_psr_create(ctx);
-	if (pool->psr == NULL) {
-		dm_error("DC: failed to create psr!\n");
-		BREAK_TO_DEBUGGER();
-		goto create_fail;
-	}
-
-	/* ABM */
-	for (i = 0; i < pool->res_cap->num_timing_generator; i++) {
-		pool->multiple_abms[i] = dmub_abm_create(ctx, &abm_regs[i], &abm_shift, &abm_mask);
-		if (pool->multiple_abms[i] == NULL) {
-			dm_error("DC: failed to create abm for pipe %d!\n", i);
-			BREAK_TO_DEBUGGER();
-			goto create_fail;
-		}
-	}
-
-	/* MPC and DSC */
-	pool->mpc = dcn303_mpc_create(ctx, pool->mpcc_count, pool->res_cap->num_mpc_3dlut);
-	if (pool->mpc == NULL) {
-		BREAK_TO_DEBUGGER();
-		dm_error("DC: failed to create mpc!\n");
-		goto create_fail;
-	}
-
-	for (i = 0; i < pool->res_cap->num_dsc; i++) {
-		pool->dscs[i] = dcn303_dsc_create(ctx, i);
-		if (pool->dscs[i] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error("DC: failed to create display stream compressor %d!\n", i);
-			goto create_fail;
-		}
-	}
-
-	/* DWB and MMHUBBUB */
-	if (!dcn303_dwbc_create(ctx, pool)) {
-		BREAK_TO_DEBUGGER();
-		dm_error("DC: failed to create dwbc!\n");
-		goto create_fail;
-	}
-
-	if (!dcn303_mmhubbub_create(ctx, pool)) {
-		BREAK_TO_DEBUGGER();
-		dm_error("DC: failed to create mcif_wb!\n");
-		goto create_fail;
-	}
-
-	/* AUX and I2C */
-	for (i = 0; i < pool->res_cap->num_ddc; i++) {
-		pool->engines[i] = dcn303_aux_engine_create(ctx, i);
-		if (pool->engines[i] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error("DC:failed to create aux engine!!\n");
-			goto create_fail;
-		}
-		pool->hw_i2cs[i] = dcn303_i2c_hw_create(ctx, i);
-		if (pool->hw_i2cs[i] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error("DC:failed to create hw i2c!!\n");
-			goto create_fail;
-		}
-		pool->sw_i2cs[i] = NULL;
-	}
-
-	/* Audio, Stream Encoders including HPO and virtual, MPC 3D LUTs */
-	if (!resource_construct(num_virtual_links, dc, pool,
-			(!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment) ?
-					&res_create_funcs : &res_create_maximus_funcs)))
-		goto create_fail;
-
-	/* HW Sequencer and Plane caps */
-	dcn303_hw_sequencer_construct(dc);
-
-	dc->caps.max_planes =  pool->pipe_count;
-
-	for (i = 0; i < dc->caps.max_planes; ++i)
-		dc->caps.planes[i] = plane_cap;
-
-	dc->cap_funcs = cap_funcs;
-
-	if (dc->ctx->dc_bios->fw_info.oem_i2c_present) {
-		ddc_init_data.ctx = dc->ctx;
-		ddc_init_data.link = NULL;
-		ddc_init_data.id.id = dc->ctx->dc_bios->fw_info.oem_i2c_obj_id;
-		ddc_init_data.id.enum_id = 0;
-		ddc_init_data.id.type = OBJECT_TYPE_GENERIC;
-		pool->oem_device = dal_ddc_service_create(&ddc_init_data);
-	} else {
-		pool->oem_device = NULL;
-	}
-
-	return true;
-
-create_fail:
-
-	dcn303_resource_destruct(pool);
-
-	return false;
-}
-
-struct resource_pool *dcn303_create_resource_pool(const struct dc_init_data *init_data, struct dc *dc)
-{
-	struct resource_pool *pool = kzalloc(sizeof(struct resource_pool), GFP_KERNEL);
-
-	if (!pool)
-		return NULL;
-
-	if (dcn303_resource_construct(init_data->num_virtual_links, dc, pool))
-		return pool;
-
-	BREAK_TO_DEBUGGER();
-	kfree(pool);
-	return NULL;
-}
diff --git a/drivers/gpu/drm/amd/display/dc/dcn303/dcn303_resource.h b/drivers/gpu/drm/amd/display/dc/dcn303/dcn303_resource.h
deleted file mode 100644
index 5b590c169763..000000000000
--- a/drivers/gpu/drm/amd/display/dc/dcn303/dcn303_resource.h
+++ /dev/null
@@ -1,17 +0,0 @@
-// SPDX-License-Identifier: MIT
-/*
- * Copyright (C) 2021 Advanced Micro Devices, Inc.
- *
- * Authors: AMD
- */
-
-#ifndef _DCN303_RESOURCE_H_
-#define _DCN303_RESOURCE_H_
-
-#include "core_types.h"
-
-struct resource_pool *dcn303_create_resource_pool(const struct dc_init_data *init_data, struct dc *dc);
-
-void dcn303_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params);
-
-#endif /* _DCN303_RESOURCE_H_ */
diff --git a/drivers/gpu/drm/amd/display/dc/dcn31/Makefile b/drivers/gpu/drm/amd/display/dc/dcn31/Makefile
index d20e3b8ccc30..d510e4652c18 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn31/Makefile
+++ b/drivers/gpu/drm/amd/display/dc/dcn31/Makefile
@@ -1,46 +1,13 @@
+# SPDX-License-Identifier: MIT
+# Copyright © 2024 Advanced Micro Devices, Inc. All rights reserved.
 #
-# (c) Copyright 2020 Advanced Micro Devices, Inc. All the rights reserved
-#
-#  All rights reserved.  This notice is intended as a precaution against
-#  inadvertent publication and does not imply publication or any waiver
-#  of confidentiality.  The year included in the foregoing notice is the
-#  year of creation of the work.
-#
-#  Authors: AMD
 #
 # Makefile for dcn31.
 
-DCN31 = dcn31_resource.o dcn31_hubbub.o dcn31_hwseq.o dcn31_init.o dcn31_hubp.o \
-	dcn31_dccg.o dcn31_optc.o dcn31_dio_link_encoder.o dcn31_panel_cntl.o \
-	dcn31_apg.o dcn31_hpo_dp_stream_encoder.o dcn31_hpo_dp_link_encoder.o \
+DCN31 = dcn31_panel_cntl.o \
+	dcn31_apg.o \
 	dcn31_afmt.o dcn31_vpg.o
 
-ifdef CONFIG_X86
-CFLAGS_$(AMDDALPATH)/dc/dcn31/dcn31_resource.o := -msse
-endif
-
-ifdef CONFIG_PPC64
-CFLAGS_$(AMDDALPATH)/dc/dcn31/dcn31_resource.o := -mhard-float -maltivec
-endif
-
-ifdef CONFIG_CC_IS_GCC
-ifeq ($(call cc-ifversion, -lt, 0701, y), y)
-IS_OLD_GCC = 1
-endif
-CFLAGS_$(AMDDALPATH)/dc/dcn31/dcn31_resource.o += -mhard-float
-endif
-
-ifdef CONFIG_X86
-ifdef IS_OLD_GCC
-# Stack alignment mismatch, proceed with caution.
-# GCC < 7.1 cannot compile code using `double` and -mpreferred-stack-boundary=3
-# (8B stack alignment).
-CFLAGS_$(AMDDALPATH)/dc/dcn31/dcn31_resource.o += -mpreferred-stack-boundary=4
-else
-CFLAGS_$(AMDDALPATH)/dc/dcn31/dcn31_resource.o += -msse2
-endif
-endif
-
 AMD_DAL_DCN31 = $(addprefix $(AMDDALPATH)/dc/dcn31/,$(DCN31))
 
 AMD_DISPLAY_FILES += $(AMD_DAL_DCN31)
diff --git a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_apg.c b/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_apg.c
index de5e18c2a3ac..05aac3e444b4 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_apg.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_apg.c
@@ -72,40 +72,6 @@ static void apg31_disable(
 	REG_UPDATE(APG_CONTROL2, APG_ENABLE, 0);
 }
 
-static union audio_cea_channels speakers_to_channels(
-	struct audio_speaker_flags speaker_flags)
-{
-	union audio_cea_channels cea_channels = {0};
-
-	/* these are one to one */
-	cea_channels.channels.FL = speaker_flags.FL_FR;
-	cea_channels.channels.FR = speaker_flags.FL_FR;
-	cea_channels.channels.LFE = speaker_flags.LFE;
-	cea_channels.channels.FC = speaker_flags.FC;
-
-	/* if Rear Left and Right exist move RC speaker to channel 7
-	 * otherwise to channel 5
-	 */
-	if (speaker_flags.RL_RR) {
-		cea_channels.channels.RL_RC = speaker_flags.RL_RR;
-		cea_channels.channels.RR = speaker_flags.RL_RR;
-		cea_channels.channels.RC_RLC_FLC = speaker_flags.RC;
-	} else {
-		cea_channels.channels.RL_RC = speaker_flags.RC;
-	}
-
-	/* FRONT Left Right Center and REAR Left Right Center are exclusive */
-	if (speaker_flags.FLC_FRC) {
-		cea_channels.channels.RC_RLC_FLC = speaker_flags.FLC_FRC;
-		cea_channels.channels.RRC_FRC = speaker_flags.FLC_FRC;
-	} else {
-		cea_channels.channels.RC_RLC_FLC = speaker_flags.RLC_RRC;
-		cea_channels.channels.RRC_FRC = speaker_flags.RLC_RRC;
-	}
-
-	return cea_channels;
-}
-
 static void apg31_se_audio_setup(
 	struct apg *apg,
 	unsigned int az_inst,
@@ -113,44 +79,24 @@ static void apg31_se_audio_setup(
 {
 	struct dcn31_apg *apg31 = DCN31_APG_FROM_APG(apg);
 
-	uint32_t speakers = 0;
-	uint32_t channels = 0;
-
 	ASSERT(audio_info);
 	/* This should not happen.it does so we don't get BSOD*/
 	if (audio_info == NULL)
 		return;
 
-	speakers = audio_info->flags.info.ALLSPEAKERS;
-	channels = speakers_to_channels(audio_info->flags.speaker_flags).all;
-
 	/* DisplayPort only allows for one audio stream with stream ID 0 */
 	REG_UPDATE(APG_CONTROL2, APG_DP_AUDIO_STREAM_ID, 0);
 
 	/* When running in "pair mode", pairs of audio channels have their own enable
 	 * this is for really old audio drivers */
 	REG_UPDATE(APG_DBG_GEN_CONTROL, APG_DBG_AUDIO_CHANNEL_ENABLE, 0xFF);
-	// REG_UPDATE(APG_DBG_GEN_CONTROL, APG_DBG_AUDIO_CHANNEL_ENABLE, channels);
 
 	/* Disable forced mem power off */
 	REG_UPDATE(APG_MEM_PWR, APG_MEM_PWR_FORCE, 0);
-
-	apg31_enable(apg);
-}
-
-static void apg31_audio_mute_control(
-	struct apg *apg,
-	bool mute)
-{
-	if (mute)
-		apg31_disable(apg);
-	else
-		apg31_enable(apg);
 }
 
 static struct apg_funcs dcn31_apg_funcs = {
 	.se_audio_setup			= apg31_se_audio_setup,
-	.audio_mute_control		= apg31_audio_mute_control,
 	.enable_apg			= apg31_enable,
 	.disable_apg			= apg31_disable,
 };
diff --git a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_apg.h b/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_apg.h
index 24f568e120d8..1b81f6773c53 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_apg.h
+++ b/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_apg.h
@@ -84,10 +84,6 @@ struct apg_funcs {
 		unsigned int az_inst,
 		struct audio_info *audio_info);
 
-	void (*audio_mute_control)(
-		struct apg *apg,
-		bool mute);
-
 	void (*enable_apg)(
 		struct apg *apg);
 
diff --git a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_dccg.c b/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_dccg.c
deleted file mode 100644
index 815481a3ef54..000000000000
--- a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_dccg.c
+++ /dev/null
@@ -1,660 +0,0 @@
-/*
- * Copyright 2018 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: AMD
- *
- */
-
-#include "reg_helper.h"
-#include "core_types.h"
-#include "dcn31_dccg.h"
-#include "dal_asic_id.h"
-
-#define TO_DCN_DCCG(dccg)\
-	container_of(dccg, struct dcn_dccg, base)
-
-#define REG(reg) \
-	(dccg_dcn->regs->reg)
-
-#undef FN
-#define FN(reg_name, field_name) \
-	dccg_dcn->dccg_shift->field_name, dccg_dcn->dccg_mask->field_name
-
-#define CTX \
-	dccg_dcn->base.ctx
-#define DC_LOGGER \
-	dccg->ctx->logger
-
-static void dccg31_update_dpp_dto(struct dccg *dccg, int dpp_inst, int req_dppclk)
-{
-	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
-
-	if (dccg->ref_dppclk && req_dppclk) {
-		int ref_dppclk = dccg->ref_dppclk;
-		int modulo, phase;
-
-		// phase / modulo = dpp pipe clk / dpp global clk
-		modulo = 0xff;   // use FF at the end
-		phase = ((modulo * req_dppclk) + ref_dppclk - 1) / ref_dppclk;
-
-		if (phase > 0xff) {
-			ASSERT(false);
-			phase = 0xff;
-		}
-
-		REG_SET_2(DPPCLK_DTO_PARAM[dpp_inst], 0,
-				DPPCLK0_DTO_PHASE, phase,
-				DPPCLK0_DTO_MODULO, modulo);
-		REG_UPDATE(DPPCLK_DTO_CTRL,
-				DPPCLK_DTO_ENABLE[dpp_inst], 1);
-	} else {
-		//DTO must be enabled to generate a 0Hz clock output
-		if (dccg->ctx->dc->debug.root_clock_optimization.bits.dpp) {
-			REG_UPDATE(DPPCLK_DTO_CTRL,
-					DPPCLK_DTO_ENABLE[dpp_inst], 1);
-			REG_SET_2(DPPCLK_DTO_PARAM[dpp_inst], 0,
-					DPPCLK0_DTO_PHASE, 0,
-					DPPCLK0_DTO_MODULO, 1);
-		} else {
-			REG_UPDATE(DPPCLK_DTO_CTRL,
-					DPPCLK_DTO_ENABLE[dpp_inst], 0);
-		}
-	}
-	dccg->pipe_dppclk_khz[dpp_inst] = req_dppclk;
-}
-
-static enum phyd32clk_clock_source get_phy_mux_symclk(
-		struct dcn_dccg *dccg_dcn,
-		enum phyd32clk_clock_source src)
-{
-	if (dccg_dcn->base.ctx->asic_id.hw_internal_rev == YELLOW_CARP_B0) {
-		if (src == PHYD32CLKC)
-			src = PHYD32CLKF;
-		if (src == PHYD32CLKD)
-			src = PHYD32CLKG;
-	}
-	return src;
-}
-
-static void dccg31_enable_dpstreamclk(struct dccg *dccg, int otg_inst)
-{
-	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
-
-	/* enabled to select one of the DTBCLKs for pipe */
-	switch (otg_inst) {
-	case 0:
-		REG_UPDATE(DPSTREAMCLK_CNTL,
-				DPSTREAMCLK_PIPE0_EN, 1);
-		break;
-	case 1:
-		REG_UPDATE(DPSTREAMCLK_CNTL,
-				DPSTREAMCLK_PIPE1_EN, 1);
-		break;
-	case 2:
-		REG_UPDATE(DPSTREAMCLK_CNTL,
-				DPSTREAMCLK_PIPE2_EN, 1);
-		break;
-	case 3:
-		REG_UPDATE(DPSTREAMCLK_CNTL,
-				DPSTREAMCLK_PIPE3_EN, 1);
-		break;
-	default:
-		BREAK_TO_DEBUGGER();
-		return;
-	}
-	if (dccg->ctx->dc->debug.root_clock_optimization.bits.dpstream)
-		REG_UPDATE(DCCG_GATE_DISABLE_CNTL3,
-			DPSTREAMCLK_ROOT_GATE_DISABLE, 1);
-}
-
-static void dccg31_disable_dpstreamclk(struct dccg *dccg, int otg_inst)
-{
-	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
-
-	if (dccg->ctx->dc->debug.root_clock_optimization.bits.dpstream)
-		REG_UPDATE(DCCG_GATE_DISABLE_CNTL3,
-				DPSTREAMCLK_ROOT_GATE_DISABLE, 0);
-
-	switch (otg_inst) {
-	case 0:
-		REG_UPDATE(DPSTREAMCLK_CNTL,
-				DPSTREAMCLK_PIPE0_EN, 0);
-		break;
-	case 1:
-		REG_UPDATE(DPSTREAMCLK_CNTL,
-				DPSTREAMCLK_PIPE1_EN, 0);
-		break;
-	case 2:
-		REG_UPDATE(DPSTREAMCLK_CNTL,
-				DPSTREAMCLK_PIPE2_EN, 0);
-		break;
-	case 3:
-		REG_UPDATE(DPSTREAMCLK_CNTL,
-				DPSTREAMCLK_PIPE3_EN, 0);
-		break;
-	default:
-		BREAK_TO_DEBUGGER();
-		return;
-	}
-}
-
-void dccg31_set_dpstreamclk(
-		struct dccg *dccg,
-		enum hdmistreamclk_source src,
-		int otg_inst)
-{
-	if (src == REFCLK)
-		dccg31_disable_dpstreamclk(dccg, otg_inst);
-	else
-		dccg31_enable_dpstreamclk(dccg, otg_inst);
-}
-
-void dccg31_enable_symclk32_se(
-		struct dccg *dccg,
-		int hpo_se_inst,
-		enum phyd32clk_clock_source phyd32clk)
-{
-	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
-
-	phyd32clk = get_phy_mux_symclk(dccg_dcn, phyd32clk);
-
-	/* select one of the PHYD32CLKs as the source for symclk32_se */
-	switch (hpo_se_inst) {
-	case 0:
-		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se)
-			REG_UPDATE(DCCG_GATE_DISABLE_CNTL3,
-					SYMCLK32_ROOT_SE0_GATE_DISABLE, 1);
-		REG_UPDATE_2(SYMCLK32_SE_CNTL,
-				SYMCLK32_SE0_SRC_SEL, phyd32clk,
-				SYMCLK32_SE0_EN, 1);
-		break;
-	case 1:
-		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se)
-			REG_UPDATE(DCCG_GATE_DISABLE_CNTL3,
-					SYMCLK32_ROOT_SE1_GATE_DISABLE, 1);
-		REG_UPDATE_2(SYMCLK32_SE_CNTL,
-				SYMCLK32_SE1_SRC_SEL, phyd32clk,
-				SYMCLK32_SE1_EN, 1);
-		break;
-	case 2:
-		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se)
-			REG_UPDATE(DCCG_GATE_DISABLE_CNTL3,
-					SYMCLK32_ROOT_SE2_GATE_DISABLE, 1);
-		REG_UPDATE_2(SYMCLK32_SE_CNTL,
-				SYMCLK32_SE2_SRC_SEL, phyd32clk,
-				SYMCLK32_SE2_EN, 1);
-		break;
-	case 3:
-		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se)
-			REG_UPDATE(DCCG_GATE_DISABLE_CNTL3,
-					SYMCLK32_ROOT_SE3_GATE_DISABLE, 1);
-		REG_UPDATE_2(SYMCLK32_SE_CNTL,
-				SYMCLK32_SE3_SRC_SEL, phyd32clk,
-				SYMCLK32_SE3_EN, 1);
-		break;
-	default:
-		BREAK_TO_DEBUGGER();
-		return;
-	}
-}
-
-void dccg31_disable_symclk32_se(
-		struct dccg *dccg,
-		int hpo_se_inst)
-{
-	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
-
-	/* set refclk as the source for symclk32_se */
-	switch (hpo_se_inst) {
-	case 0:
-		REG_UPDATE_2(SYMCLK32_SE_CNTL,
-				SYMCLK32_SE0_SRC_SEL, 0,
-				SYMCLK32_SE0_EN, 0);
-		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se)
-			REG_UPDATE(DCCG_GATE_DISABLE_CNTL3,
-					SYMCLK32_ROOT_SE0_GATE_DISABLE, 0);
-		break;
-	case 1:
-		REG_UPDATE_2(SYMCLK32_SE_CNTL,
-				SYMCLK32_SE1_SRC_SEL, 0,
-				SYMCLK32_SE1_EN, 0);
-		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se)
-			REG_UPDATE(DCCG_GATE_DISABLE_CNTL3,
-					SYMCLK32_ROOT_SE1_GATE_DISABLE, 0);
-		break;
-	case 2:
-		REG_UPDATE_2(SYMCLK32_SE_CNTL,
-				SYMCLK32_SE2_SRC_SEL, 0,
-				SYMCLK32_SE2_EN, 0);
-		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se)
-			REG_UPDATE(DCCG_GATE_DISABLE_CNTL3,
-					SYMCLK32_ROOT_SE2_GATE_DISABLE, 0);
-		break;
-	case 3:
-		REG_UPDATE_2(SYMCLK32_SE_CNTL,
-				SYMCLK32_SE3_SRC_SEL, 0,
-				SYMCLK32_SE3_EN, 0);
-		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se)
-			REG_UPDATE(DCCG_GATE_DISABLE_CNTL3,
-					SYMCLK32_ROOT_SE3_GATE_DISABLE, 0);
-		break;
-	default:
-		BREAK_TO_DEBUGGER();
-		return;
-	}
-}
-
-void dccg31_enable_symclk32_le(
-		struct dccg *dccg,
-		int hpo_le_inst,
-		enum phyd32clk_clock_source phyd32clk)
-{
-	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
-
-	phyd32clk = get_phy_mux_symclk(dccg_dcn, phyd32clk);
-
-	/* select one of the PHYD32CLKs as the source for symclk32_le */
-	switch (hpo_le_inst) {
-	case 0:
-		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_le)
-			REG_UPDATE(DCCG_GATE_DISABLE_CNTL3,
-					SYMCLK32_ROOT_LE0_GATE_DISABLE, 1);
-		REG_UPDATE_2(SYMCLK32_LE_CNTL,
-				SYMCLK32_LE0_SRC_SEL, phyd32clk,
-				SYMCLK32_LE0_EN, 1);
-		break;
-	case 1:
-		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_le)
-			REG_UPDATE(DCCG_GATE_DISABLE_CNTL3,
-					SYMCLK32_ROOT_LE1_GATE_DISABLE, 1);
-		REG_UPDATE_2(SYMCLK32_LE_CNTL,
-				SYMCLK32_LE1_SRC_SEL, phyd32clk,
-				SYMCLK32_LE1_EN, 1);
-		break;
-	default:
-		BREAK_TO_DEBUGGER();
-		return;
-	}
-}
-
-void dccg31_disable_symclk32_le(
-		struct dccg *dccg,
-		int hpo_le_inst)
-{
-	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
-
-	/* set refclk as the source for symclk32_le */
-	switch (hpo_le_inst) {
-	case 0:
-		REG_UPDATE_2(SYMCLK32_LE_CNTL,
-				SYMCLK32_LE0_SRC_SEL, 0,
-				SYMCLK32_LE0_EN, 0);
-		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_le)
-			REG_UPDATE(DCCG_GATE_DISABLE_CNTL3,
-					SYMCLK32_ROOT_LE0_GATE_DISABLE, 0);
-		break;
-	case 1:
-		REG_UPDATE_2(SYMCLK32_LE_CNTL,
-				SYMCLK32_LE1_SRC_SEL, 0,
-				SYMCLK32_LE1_EN, 0);
-		if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_le)
-			REG_UPDATE(DCCG_GATE_DISABLE_CNTL3,
-					SYMCLK32_ROOT_LE1_GATE_DISABLE, 0);
-		break;
-	default:
-		BREAK_TO_DEBUGGER();
-		return;
-	}
-}
-
-static void dccg31_disable_dscclk(struct dccg *dccg, int inst)
-{
-	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
-
-	if (!dccg->ctx->dc->debug.root_clock_optimization.bits.dsc)
-		return;
-	//DTO must be enabled to generate a 0 Hz clock output
-	switch (inst) {
-	case 0:
-		REG_UPDATE(DSCCLK_DTO_CTRL,
-				DSCCLK0_DTO_ENABLE, 1);
-		REG_UPDATE_2(DSCCLK0_DTO_PARAM,
-				DSCCLK0_DTO_PHASE, 0,
-				DSCCLK0_DTO_MODULO, 1);
-		break;
-	case 1:
-		REG_UPDATE(DSCCLK_DTO_CTRL,
-				DSCCLK1_DTO_ENABLE, 1);
-		REG_UPDATE_2(DSCCLK1_DTO_PARAM,
-				DSCCLK1_DTO_PHASE, 0,
-				DSCCLK1_DTO_MODULO, 1);
-		break;
-	case 2:
-		REG_UPDATE(DSCCLK_DTO_CTRL,
-				DSCCLK2_DTO_ENABLE, 1);
-		REG_UPDATE_2(DSCCLK2_DTO_PARAM,
-				DSCCLK2_DTO_PHASE, 0,
-				DSCCLK2_DTO_MODULO, 1);
-		break;
-	default:
-		BREAK_TO_DEBUGGER();
-		return;
-	}
-}
-
-static void dccg31_enable_dscclk(struct dccg *dccg, int inst)
-{
-	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
-
-	if (!dccg->ctx->dc->debug.root_clock_optimization.bits.dsc)
-		return;
-	//Disable DTO
-	switch (inst) {
-	case 0:
-		REG_UPDATE_2(DSCCLK0_DTO_PARAM,
-				DSCCLK0_DTO_PHASE, 0,
-				DSCCLK0_DTO_MODULO, 0);
-		REG_UPDATE(DSCCLK_DTO_CTRL,
-				DSCCLK0_DTO_ENABLE, 0);
-		break;
-	case 1:
-		REG_UPDATE_2(DSCCLK1_DTO_PARAM,
-				DSCCLK1_DTO_PHASE, 0,
-				DSCCLK1_DTO_MODULO, 0);
-		REG_UPDATE(DSCCLK_DTO_CTRL,
-				DSCCLK1_DTO_ENABLE, 0);
-		break;
-	case 2:
-		REG_UPDATE_2(DSCCLK2_DTO_PARAM,
-				DSCCLK2_DTO_PHASE, 0,
-				DSCCLK2_DTO_MODULO, 0);
-		REG_UPDATE(DSCCLK_DTO_CTRL,
-				DSCCLK2_DTO_ENABLE, 0);
-		break;
-	default:
-		BREAK_TO_DEBUGGER();
-		return;
-	}
-}
-
-void dccg31_set_physymclk(
-		struct dccg *dccg,
-		int phy_inst,
-		enum physymclk_clock_source clk_src,
-		bool force_enable)
-{
-	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
-
-	/* Force PHYSYMCLK on and Select phyd32clk as the source of clock which is output to PHY through DCIO */
-	switch (phy_inst) {
-	case 0:
-		if (force_enable)
-			REG_UPDATE_2(PHYASYMCLK_CLOCK_CNTL,
-					PHYASYMCLK_FORCE_EN, 1,
-					PHYASYMCLK_FORCE_SRC_SEL, clk_src);
-		else
-			REG_UPDATE_2(PHYASYMCLK_CLOCK_CNTL,
-					PHYASYMCLK_FORCE_EN, 0,
-					PHYASYMCLK_FORCE_SRC_SEL, 0);
-		break;
-	case 1:
-		if (force_enable)
-			REG_UPDATE_2(PHYBSYMCLK_CLOCK_CNTL,
-					PHYBSYMCLK_FORCE_EN, 1,
-					PHYBSYMCLK_FORCE_SRC_SEL, clk_src);
-		else
-			REG_UPDATE_2(PHYBSYMCLK_CLOCK_CNTL,
-					PHYBSYMCLK_FORCE_EN, 0,
-					PHYBSYMCLK_FORCE_SRC_SEL, 0);
-		break;
-	case 2:
-		if (force_enable)
-			REG_UPDATE_2(PHYCSYMCLK_CLOCK_CNTL,
-					PHYCSYMCLK_FORCE_EN, 1,
-					PHYCSYMCLK_FORCE_SRC_SEL, clk_src);
-		else
-			REG_UPDATE_2(PHYCSYMCLK_CLOCK_CNTL,
-					PHYCSYMCLK_FORCE_EN, 0,
-					PHYCSYMCLK_FORCE_SRC_SEL, 0);
-		break;
-	case 3:
-		if (force_enable)
-			REG_UPDATE_2(PHYDSYMCLK_CLOCK_CNTL,
-					PHYDSYMCLK_FORCE_EN, 1,
-					PHYDSYMCLK_FORCE_SRC_SEL, clk_src);
-		else
-			REG_UPDATE_2(PHYDSYMCLK_CLOCK_CNTL,
-					PHYDSYMCLK_FORCE_EN, 0,
-					PHYDSYMCLK_FORCE_SRC_SEL, 0);
-		break;
-	case 4:
-		if (force_enable)
-			REG_UPDATE_2(PHYESYMCLK_CLOCK_CNTL,
-					PHYESYMCLK_FORCE_EN, 1,
-					PHYESYMCLK_FORCE_SRC_SEL, clk_src);
-		else
-			REG_UPDATE_2(PHYESYMCLK_CLOCK_CNTL,
-					PHYESYMCLK_FORCE_EN, 0,
-					PHYESYMCLK_FORCE_SRC_SEL, 0);
-		break;
-	default:
-		BREAK_TO_DEBUGGER();
-		return;
-	}
-}
-
-/* Controls the generation of pixel valid for OTG in (OTG -> HPO case) */
-void dccg31_set_dtbclk_dto(
-		struct dccg *dccg,
-		int dtbclk_inst,
-		int req_dtbclk_khz,
-		int num_odm_segments,
-		const struct dc_crtc_timing *timing)
-{
-	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
-	uint32_t dtbdto_div;
-
-	/* Mode	                DTBDTO Rate       DTBCLK_DTO<x>_DIV Register
-	 * ODM 4:1 combine      pixel rate/4      2
-	 * ODM 2:1 combine      pixel rate/2      4
-	 * non-DSC 4:2:0 mode   pixel rate/2      4
-	 * DSC native 4:2:0     pixel rate/2      4
-	 * DSC native 4:2:2     pixel rate/2      4
-	 * Other modes          pixel rate        8
-	 */
-	if (num_odm_segments == 4) {
-		dtbdto_div = 2;
-		req_dtbclk_khz = req_dtbclk_khz / 4;
-	} else if ((num_odm_segments == 2) ||
-			(timing->pixel_encoding == PIXEL_ENCODING_YCBCR420) ||
-			(timing->flags.DSC && timing->pixel_encoding == PIXEL_ENCODING_YCBCR422
-					&& !timing->dsc_cfg.ycbcr422_simple)) {
-		dtbdto_div = 4;
-		req_dtbclk_khz = req_dtbclk_khz / 2;
-	} else
-		dtbdto_div = 8;
-
-	if (dccg->ref_dtbclk_khz && req_dtbclk_khz) {
-		uint32_t modulo, phase;
-
-		// phase / modulo = dtbclk / dtbclk ref
-		modulo = dccg->ref_dtbclk_khz * 1000;
-		phase = div_u64((((unsigned long long)modulo * req_dtbclk_khz) + dccg->ref_dtbclk_khz - 1),
-			dccg->ref_dtbclk_khz);
-
-		REG_UPDATE(OTG_PIXEL_RATE_CNTL[dtbclk_inst],
-				DTBCLK_DTO_DIV[dtbclk_inst], dtbdto_div);
-
-		REG_WRITE(DTBCLK_DTO_MODULO[dtbclk_inst], modulo);
-		REG_WRITE(DTBCLK_DTO_PHASE[dtbclk_inst], phase);
-
-		REG_UPDATE(OTG_PIXEL_RATE_CNTL[dtbclk_inst],
-				DTBCLK_DTO_ENABLE[dtbclk_inst], 1);
-
-		REG_WAIT(OTG_PIXEL_RATE_CNTL[dtbclk_inst],
-				DTBCLKDTO_ENABLE_STATUS[dtbclk_inst], 1,
-				1, 100);
-
-		/* The recommended programming sequence to enable DTBCLK DTO to generate
-		 * valid pixel HPO DPSTREAM ENCODER, specifies that DTO source select should
-		 * be set only after DTO is enabled
-		 */
-		REG_UPDATE(OTG_PIXEL_RATE_CNTL[dtbclk_inst],
-				PIPE_DTO_SRC_SEL[dtbclk_inst], 1);
-
-		dccg->dtbclk_khz[dtbclk_inst] = req_dtbclk_khz;
-	} else {
-		REG_UPDATE_3(OTG_PIXEL_RATE_CNTL[dtbclk_inst],
-				DTBCLK_DTO_ENABLE[dtbclk_inst], 0,
-				PIPE_DTO_SRC_SEL[dtbclk_inst], 0,
-				DTBCLK_DTO_DIV[dtbclk_inst], dtbdto_div);
-
-		REG_WRITE(DTBCLK_DTO_MODULO[dtbclk_inst], 0);
-		REG_WRITE(DTBCLK_DTO_PHASE[dtbclk_inst], 0);
-
-		dccg->dtbclk_khz[dtbclk_inst] = 0;
-	}
-}
-
-void dccg31_set_audio_dtbclk_dto(
-		struct dccg *dccg,
-		uint32_t req_audio_dtbclk_khz)
-{
-	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
-
-	if (dccg->ref_dtbclk_khz && req_audio_dtbclk_khz) {
-		uint32_t modulo, phase;
-
-		// phase / modulo = dtbclk / dtbclk ref
-		modulo = dccg->ref_dtbclk_khz * 1000;
-		phase = div_u64((((unsigned long long)modulo * req_audio_dtbclk_khz) + dccg->ref_dtbclk_khz - 1),
-			dccg->ref_dtbclk_khz);
-
-
-		REG_WRITE(DCCG_AUDIO_DTBCLK_DTO_MODULO, modulo);
-		REG_WRITE(DCCG_AUDIO_DTBCLK_DTO_PHASE, phase);
-
-		//REG_UPDATE(DCCG_AUDIO_DTO_SOURCE,
-		//		DCCG_AUDIO_DTBCLK_DTO_USE_512FBR_DTO, 1);
-
-		REG_UPDATE(DCCG_AUDIO_DTO_SOURCE,
-				DCCG_AUDIO_DTO_SEL, 4);  //  04 - DCCG_AUDIO_DTO_SEL_AUDIO_DTO_DTBCLK
-
-		dccg->audio_dtbclk_khz = req_audio_dtbclk_khz;
-	} else {
-		REG_WRITE(DCCG_AUDIO_DTBCLK_DTO_PHASE, 0);
-		REG_WRITE(DCCG_AUDIO_DTBCLK_DTO_MODULO, 0);
-
-		REG_UPDATE(DCCG_AUDIO_DTO_SOURCE,
-				DCCG_AUDIO_DTO_SEL, 3);  //  03 - DCCG_AUDIO_DTO_SEL_NO_AUDIO_DTO
-
-		dccg->audio_dtbclk_khz = 0;
-	}
-}
-
-static void dccg31_get_dccg_ref_freq(struct dccg *dccg,
-		unsigned int xtalin_freq_inKhz,
-		unsigned int *dccg_ref_freq_inKhz)
-{
-	/*
-	 * Assume refclk is sourced from xtalin
-	 * expect 24MHz
-	 */
-	*dccg_ref_freq_inKhz = xtalin_freq_inKhz;
-	return;
-}
-
-static void dccg31_set_dispclk_change_mode(
-	struct dccg *dccg,
-	enum dentist_dispclk_change_mode change_mode)
-{
-	struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg);
-
-	REG_UPDATE(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_CHG_MODE,
-		   change_mode == DISPCLK_CHANGE_MODE_RAMPING ? 2 : 0);
-}
-
-void dccg31_init(struct dccg *dccg)
-{
-	/* Set HPO stream encoder to use refclk to avoid case where PHY is
-	 * disabled and SYMCLK32 for HPO SE is sourced from PHYD32CLK which
-	 * will cause DCN to hang.
-	 */
-	dccg31_disable_symclk32_se(dccg, 0);
-	dccg31_disable_symclk32_se(dccg, 1);
-	dccg31_disable_symclk32_se(dccg, 2);
-	dccg31_disable_symclk32_se(dccg, 3);
-
-	if (dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_le) {
-		dccg31_disable_symclk32_le(dccg, 0);
-		dccg31_disable_symclk32_le(dccg, 1);
-	}
-
-	if (dccg->ctx->dc->debug.root_clock_optimization.bits.dpstream) {
-		dccg31_disable_dpstreamclk(dccg, 0);
-		dccg31_disable_dpstreamclk(dccg, 1);
-		dccg31_disable_dpstreamclk(dccg, 2);
-		dccg31_disable_dpstreamclk(dccg, 3);
-	}
-
-}
-
-static const struct dccg_funcs dccg31_funcs = {
-	.update_dpp_dto = dccg31_update_dpp_dto,
-	.get_dccg_ref_freq = dccg31_get_dccg_ref_freq,
-	.dccg_init = dccg31_init,
-	.set_dpstreamclk = dccg31_set_dpstreamclk,
-	.enable_symclk32_se = dccg31_enable_symclk32_se,
-	.disable_symclk32_se = dccg31_disable_symclk32_se,
-	.enable_symclk32_le = dccg31_enable_symclk32_le,
-	.disable_symclk32_le = dccg31_disable_symclk32_le,
-	.set_physymclk = dccg31_set_physymclk,
-	.set_dtbclk_dto = dccg31_set_dtbclk_dto,
-	.set_audio_dtbclk_dto = dccg31_set_audio_dtbclk_dto,
-	.set_dispclk_change_mode = dccg31_set_dispclk_change_mode,
-	.disable_dsc = dccg31_disable_dscclk,
-	.enable_dsc = dccg31_enable_dscclk,
-};
-
-struct dccg *dccg31_create(
-	struct dc_context *ctx,
-	const struct dccg_registers *regs,
-	const struct dccg_shift *dccg_shift,
-	const struct dccg_mask *dccg_mask)
-{
-	struct dcn_dccg *dccg_dcn = kzalloc(sizeof(*dccg_dcn), GFP_KERNEL);
-	struct dccg *base;
-
-	if (dccg_dcn == NULL) {
-		BREAK_TO_DEBUGGER();
-		return NULL;
-	}
-
-	base = &dccg_dcn->base;
-	base->ctx = ctx;
-	base->funcs = &dccg31_funcs;
-
-	dccg_dcn->regs = regs;
-	dccg_dcn->dccg_shift = dccg_shift;
-	dccg_dcn->dccg_mask = dccg_mask;
-
-	return &dccg_dcn->base;
-}
diff --git a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_hwseq.c b/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_hwseq.c
deleted file mode 100644
index 4d4ac4ceb1e8..000000000000
--- a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_hwseq.c
+++ /dev/null
@@ -1,619 +0,0 @@
-/*
- * Copyright 2016 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: AMD
- *
- */
-
-
-#include "dm_services.h"
-#include "dm_helpers.h"
-#include "core_types.h"
-#include "resource.h"
-#include "dccg.h"
-#include "dce/dce_hwseq.h"
-#include "clk_mgr.h"
-#include "reg_helper.h"
-#include "abm.h"
-#include "hubp.h"
-#include "dchubbub.h"
-#include "timing_generator.h"
-#include "opp.h"
-#include "ipp.h"
-#include "mpc.h"
-#include "mcif_wb.h"
-#include "dc_dmub_srv.h"
-#include "dcn31_hwseq.h"
-#include "link_hwss.h"
-#include "dpcd_defs.h"
-#include "dce/dmub_outbox.h"
-#include "dc_link_dp.h"
-#include "inc/link_dpcd.h"
-#include "dcn10/dcn10_hw_sequencer.h"
-#include "inc/link_enc_cfg.h"
-#include "dcn30/dcn30_vpg.h"
-#include "dce/dce_i2c_hw.h"
-
-#define DC_LOGGER_INIT(logger)
-
-#define CTX \
-	hws->ctx
-#define REG(reg)\
-	hws->regs->reg
-#define DC_LOGGER \
-		dc->ctx->logger
-
-
-#undef FN
-#define FN(reg_name, field_name) \
-	hws->shifts->field_name, hws->masks->field_name
-
-static void enable_memory_low_power(struct dc *dc)
-{
-	struct dce_hwseq *hws = dc->hwseq;
-	int i;
-
-	if (dc->debug.enable_mem_low_power.bits.dmcu) {
-		// Force ERAM to shutdown if DMCU is not enabled
-		if (dc->debug.disable_dmcu || dc->config.disable_dmcu) {
-			REG_UPDATE(DMU_MEM_PWR_CNTL, DMCU_ERAM_MEM_PWR_FORCE, 3);
-		}
-	}
-
-	// Set default OPTC memory power states
-	if (dc->debug.enable_mem_low_power.bits.optc) {
-		// Shutdown when unassigned and light sleep in VBLANK
-		REG_SET_2(ODM_MEM_PWR_CTRL3, 0, ODM_MEM_UNASSIGNED_PWR_MODE, 3, ODM_MEM_VBLANK_PWR_MODE, 1);
-	}
-
-	if (dc->debug.enable_mem_low_power.bits.vga) {
-		// Power down VGA memory
-		REG_UPDATE(MMHUBBUB_MEM_PWR_CNTL, VGA_MEM_PWR_FORCE, 1);
-	}
-
-	if (dc->debug.enable_mem_low_power.bits.mpc)
-		dc->res_pool->mpc->funcs->set_mpc_mem_lp_mode(dc->res_pool->mpc);
-
-
-	if (dc->debug.enable_mem_low_power.bits.vpg && dc->res_pool->stream_enc[0]->vpg->funcs->vpg_powerdown) {
-		// Power down VPGs
-		for (i = 0; i < dc->res_pool->stream_enc_count; i++)
-			dc->res_pool->stream_enc[i]->vpg->funcs->vpg_powerdown(dc->res_pool->stream_enc[i]->vpg);
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-		for (i = 0; i < dc->res_pool->hpo_dp_stream_enc_count; i++)
-			dc->res_pool->hpo_dp_stream_enc[i]->vpg->funcs->vpg_powerdown(dc->res_pool->hpo_dp_stream_enc[i]->vpg);
-#endif
-	}
-
-}
-
-void dcn31_init_hw(struct dc *dc)
-{
-	struct abm **abms = dc->res_pool->multiple_abms;
-	struct dce_hwseq *hws = dc->hwseq;
-	struct dc_bios *dcb = dc->ctx->dc_bios;
-	struct resource_pool *res_pool = dc->res_pool;
-	uint32_t backlight = MAX_BACKLIGHT_LEVEL;
-	int i, j;
-
-	if (dc->clk_mgr && dc->clk_mgr->funcs->init_clocks)
-		dc->clk_mgr->funcs->init_clocks(dc->clk_mgr);
-
-	if (IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) {
-
-		REG_WRITE(REFCLK_CNTL, 0);
-		REG_UPDATE(DCHUBBUB_GLOBAL_TIMER_CNTL, DCHUBBUB_GLOBAL_TIMER_ENABLE, 1);
-		REG_WRITE(DIO_MEM_PWR_CTRL, 0);
-
-		if (!dc->debug.disable_clock_gate) {
-			/* enable all DCN clock gating */
-			REG_WRITE(DCCG_GATE_DISABLE_CNTL, 0);
-
-			REG_WRITE(DCCG_GATE_DISABLE_CNTL2, 0);
-
-			REG_UPDATE(DCFCLK_CNTL, DCFCLK_GATE_DIS, 0);
-		}
-
-		//Enable ability to power gate / don't force power on permanently
-		if (hws->funcs.enable_power_gating_plane)
-			hws->funcs.enable_power_gating_plane(hws, true);
-
-		return;
-	}
-
-	if (!dcb->funcs->is_accelerated_mode(dcb)) {
-		hws->funcs.bios_golden_init(dc);
-		hws->funcs.disable_vga(dc->hwseq);
-	}
-	// Initialize the dccg
-	if (res_pool->dccg->funcs->dccg_init)
-		res_pool->dccg->funcs->dccg_init(res_pool->dccg);
-
-	enable_memory_low_power(dc);
-
-	if (dc->ctx->dc_bios->fw_info_valid) {
-		res_pool->ref_clocks.xtalin_clock_inKhz =
-				dc->ctx->dc_bios->fw_info.pll_info.crystal_frequency;
-
-		if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) {
-			if (res_pool->dccg && res_pool->hubbub) {
-
-				(res_pool->dccg->funcs->get_dccg_ref_freq)(res_pool->dccg,
-						dc->ctx->dc_bios->fw_info.pll_info.crystal_frequency,
-						&res_pool->ref_clocks.dccg_ref_clock_inKhz);
-
-				(res_pool->hubbub->funcs->get_dchub_ref_freq)(res_pool->hubbub,
-						res_pool->ref_clocks.dccg_ref_clock_inKhz,
-						&res_pool->ref_clocks.dchub_ref_clock_inKhz);
-			} else {
-				// Not all ASICs have DCCG sw component
-				res_pool->ref_clocks.dccg_ref_clock_inKhz =
-						res_pool->ref_clocks.xtalin_clock_inKhz;
-				res_pool->ref_clocks.dchub_ref_clock_inKhz =
-						res_pool->ref_clocks.xtalin_clock_inKhz;
-			}
-		}
-	} else
-		ASSERT_CRITICAL(false);
-
-	for (i = 0; i < dc->link_count; i++) {
-		/* Power up AND update implementation according to the
-		 * required signal (which may be different from the
-		 * default signal on connector).
-		 */
-		struct dc_link *link = dc->links[i];
-
-		if (link->ep_type != DISPLAY_ENDPOINT_PHY)
-			continue;
-
-		link->link_enc->funcs->hw_init(link->link_enc);
-
-		/* Check for enabled DIG to identify enabled display */
-		if (link->link_enc->funcs->is_dig_enabled &&
-			link->link_enc->funcs->is_dig_enabled(link->link_enc))
-			link->link_status.link_active = true;
-	}
-
-	/* Power gate DSCs */
-	for (i = 0; i < res_pool->res_cap->num_dsc; i++)
-		if (hws->funcs.dsc_pg_control != NULL)
-			hws->funcs.dsc_pg_control(hws, res_pool->dscs[i]->inst, false);
-
-	/* Enables outbox notifications for usb4 dpia */
-	if (dc->res_pool->usb4_dpia_count)
-		dmub_enable_outbox_notification(dc);
-
-	/* we want to turn off all dp displays before doing detection */
-	if (dc->config.power_down_display_on_boot) {
-		uint8_t dpcd_power_state = '\0';
-		enum dc_status status = DC_ERROR_UNEXPECTED;
-
-		for (i = 0; i < dc->link_count; i++) {
-			if (dc->links[i]->connector_signal != SIGNAL_TYPE_DISPLAY_PORT)
-				continue;
-
-			/* if any of the displays are lit up turn them off */
-			status = core_link_read_dpcd(dc->links[i], DP_SET_POWER,
-						     &dpcd_power_state, sizeof(dpcd_power_state));
-			if (status == DC_OK && dpcd_power_state == DP_POWER_STATE_D0) {
-				/* blank dp stream before power off receiver*/
-				if (dc->links[i]->ep_type == DISPLAY_ENDPOINT_PHY &&
-						dc->links[i]->link_enc->funcs->get_dig_frontend) {
-					unsigned int fe;
-
-					fe = dc->links[i]->link_enc->funcs->get_dig_frontend(
-										dc->links[i]->link_enc);
-					if (fe == ENGINE_ID_UNKNOWN)
-						continue;
-
-					for (j = 0; j < dc->res_pool->stream_enc_count; j++) {
-						if (fe == dc->res_pool->stream_enc[j]->id) {
-							dc->res_pool->stream_enc[j]->funcs->dp_blank(dc->links[i],
-										dc->res_pool->stream_enc[j]);
-							break;
-						}
-					}
-				}
-				dp_receiver_power_ctrl(dc->links[i], false);
-			}
-		}
-	}
-
-	/* If taking control over from VBIOS, we may want to optimize our first
-	 * mode set, so we need to skip powering down pipes until we know which
-	 * pipes we want to use.
-	 * Otherwise, if taking control is not possible, we need to power
-	 * everything down.
-	 */
-	if (dcb->funcs->is_accelerated_mode(dcb) || dc->config.power_down_display_on_boot) {
-		hws->funcs.init_pipes(dc, dc->current_state);
-		if (dc->res_pool->hubbub->funcs->allow_self_refresh_control)
-			dc->res_pool->hubbub->funcs->allow_self_refresh_control(dc->res_pool->hubbub,
-					!dc->res_pool->hubbub->ctx->dc->debug.disable_stutter);
-	}
-
-	for (i = 0; i < res_pool->audio_count; i++) {
-		struct audio *audio = res_pool->audios[i];
-
-		audio->funcs->hw_init(audio);
-	}
-
-	for (i = 0; i < dc->link_count; i++) {
-		struct dc_link *link = dc->links[i];
-
-		if (link->panel_cntl)
-			backlight = link->panel_cntl->funcs->hw_init(link->panel_cntl);
-	}
-
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		if (abms[i] != NULL)
-			abms[i]->funcs->abm_init(abms[i], backlight);
-	}
-
-	/* power AFMT HDMI memory TODO: may move to dis/en output save power*/
-	REG_WRITE(DIO_MEM_PWR_CTRL, 0);
-
-	// Set i2c to light sleep until engine is setup
-	if (dc->debug.enable_mem_low_power.bits.i2c)
-		REG_UPDATE(DIO_MEM_PWR_CTRL, I2C_LIGHT_SLEEP_FORCE, 1);
-
-	if (hws->funcs.setup_hpo_hw_control)
-		hws->funcs.setup_hpo_hw_control(hws, false);
-
-	if (!dc->debug.disable_clock_gate) {
-		/* enable all DCN clock gating */
-		REG_WRITE(DCCG_GATE_DISABLE_CNTL, 0);
-
-		REG_WRITE(DCCG_GATE_DISABLE_CNTL2, 0);
-
-		REG_UPDATE(DCFCLK_CNTL, DCFCLK_GATE_DIS, 0);
-	}
-	if (hws->funcs.enable_power_gating_plane)
-		hws->funcs.enable_power_gating_plane(dc->hwseq, true);
-
-	if (!dcb->funcs->is_accelerated_mode(dcb) && dc->res_pool->hubbub->funcs->init_watermarks)
-		dc->res_pool->hubbub->funcs->init_watermarks(dc->res_pool->hubbub);
-
-	if (dc->clk_mgr->funcs->notify_wm_ranges)
-		dc->clk_mgr->funcs->notify_wm_ranges(dc->clk_mgr);
-
-	if (dc->clk_mgr->funcs->set_hard_max_memclk)
-		dc->clk_mgr->funcs->set_hard_max_memclk(dc->clk_mgr);
-
-	if (dc->res_pool->hubbub->funcs->force_pstate_change_control)
-		dc->res_pool->hubbub->funcs->force_pstate_change_control(
-				dc->res_pool->hubbub, false, false);
-#if defined(CONFIG_DRM_AMD_DC_DCN)
-	if (dc->res_pool->hubbub->funcs->init_crb)
-		dc->res_pool->hubbub->funcs->init_crb(dc->res_pool->hubbub);
-#endif
-}
-
-void dcn31_dsc_pg_control(
-		struct dce_hwseq *hws,
-		unsigned int dsc_inst,
-		bool power_on)
-{
-	uint32_t power_gate = power_on ? 0 : 1;
-	uint32_t pwr_status = power_on ? 0 : 2;
-	uint32_t org_ip_request_cntl = 0;
-
-	if (hws->ctx->dc->debug.disable_dsc_power_gate)
-		return;
-
-	if (hws->ctx->dc->debug.root_clock_optimization.bits.dsc &&
-		hws->ctx->dc->res_pool->dccg->funcs->enable_dsc &&
-		power_on)
-		hws->ctx->dc->res_pool->dccg->funcs->enable_dsc(
-			hws->ctx->dc->res_pool->dccg, dsc_inst);
-
-	REG_GET(DC_IP_REQUEST_CNTL, IP_REQUEST_EN, &org_ip_request_cntl);
-	if (org_ip_request_cntl == 0)
-		REG_SET(DC_IP_REQUEST_CNTL, 0, IP_REQUEST_EN, 1);
-
-	switch (dsc_inst) {
-	case 0: /* DSC0 */
-		REG_UPDATE(DOMAIN16_PG_CONFIG,
-				DOMAIN_POWER_GATE, power_gate);
-
-		REG_WAIT(DOMAIN16_PG_STATUS,
-				DOMAIN_PGFSM_PWR_STATUS, pwr_status,
-				1, 1000);
-		break;
-	case 1: /* DSC1 */
-		REG_UPDATE(DOMAIN17_PG_CONFIG,
-				DOMAIN_POWER_GATE, power_gate);
-
-		REG_WAIT(DOMAIN17_PG_STATUS,
-				DOMAIN_PGFSM_PWR_STATUS, pwr_status,
-				1, 1000);
-		break;
-	case 2: /* DSC2 */
-		REG_UPDATE(DOMAIN18_PG_CONFIG,
-				DOMAIN_POWER_GATE, power_gate);
-
-		REG_WAIT(DOMAIN18_PG_STATUS,
-				DOMAIN_PGFSM_PWR_STATUS, pwr_status,
-				1, 1000);
-		break;
-	default:
-		BREAK_TO_DEBUGGER();
-		break;
-	}
-
-	if (org_ip_request_cntl == 0)
-		REG_SET(DC_IP_REQUEST_CNTL, 0, IP_REQUEST_EN, 0);
-
-	if (hws->ctx->dc->debug.root_clock_optimization.bits.dsc) {
-		if (hws->ctx->dc->res_pool->dccg->funcs->disable_dsc && !power_on)
-			hws->ctx->dc->res_pool->dccg->funcs->disable_dsc(
-				hws->ctx->dc->res_pool->dccg, dsc_inst);
-	}
-
-}
-
-
-void dcn31_enable_power_gating_plane(
-	struct dce_hwseq *hws,
-	bool enable)
-{
-	bool force_on = true; /* disable power gating */
-
-	if (enable)
-		force_on = false;
-
-	/* DCHUBP0/1/2/3/4/5 */
-	REG_UPDATE(DOMAIN0_PG_CONFIG, DOMAIN_POWER_FORCEON, force_on);
-	REG_UPDATE(DOMAIN2_PG_CONFIG, DOMAIN_POWER_FORCEON, force_on);
-
-	/* DPP0/1/2/3/4/5 */
-	REG_UPDATE(DOMAIN1_PG_CONFIG, DOMAIN_POWER_FORCEON, force_on);
-	REG_UPDATE(DOMAIN3_PG_CONFIG, DOMAIN_POWER_FORCEON, force_on);
-
-	/* DCS0/1/2/3/4/5 */
-	REG_UPDATE(DOMAIN16_PG_CONFIG, DOMAIN_POWER_FORCEON, force_on);
-	REG_UPDATE(DOMAIN17_PG_CONFIG, DOMAIN_POWER_FORCEON, force_on);
-	REG_UPDATE(DOMAIN18_PG_CONFIG, DOMAIN_POWER_FORCEON, force_on);
-}
-
-void dcn31_update_info_frame(struct pipe_ctx *pipe_ctx)
-{
-	bool is_hdmi_tmds;
-	bool is_dp;
-
-	ASSERT(pipe_ctx->stream);
-
-	if (pipe_ctx->stream_res.stream_enc == NULL)
-		return;  /* this is not root pipe */
-
-	is_hdmi_tmds = dc_is_hdmi_tmds_signal(pipe_ctx->stream->signal);
-	is_dp = dc_is_dp_signal(pipe_ctx->stream->signal);
-
-	if (!is_hdmi_tmds && !is_dp)
-		return;
-
-	if (is_hdmi_tmds)
-		pipe_ctx->stream_res.stream_enc->funcs->update_hdmi_info_packets(
-			pipe_ctx->stream_res.stream_enc,
-			&pipe_ctx->stream_res.encoder_info_frame);
-	else {
-		pipe_ctx->stream_res.stream_enc->funcs->update_dp_info_packets(
-			pipe_ctx->stream_res.stream_enc,
-			&pipe_ctx->stream_res.encoder_info_frame);
-	}
-}
-void dcn31_z10_save_init(struct dc *dc)
-{
-	union dmub_rb_cmd cmd;
-
-	memset(&cmd, 0, sizeof(cmd));
-	cmd.dcn_restore.header.type = DMUB_CMD__IDLE_OPT;
-	cmd.dcn_restore.header.sub_type = DMUB_CMD__IDLE_OPT_DCN_SAVE_INIT;
-
-	dc_dmub_srv_cmd_queue(dc->ctx->dmub_srv, &cmd);
-	dc_dmub_srv_cmd_execute(dc->ctx->dmub_srv);
-	dc_dmub_srv_wait_idle(dc->ctx->dmub_srv);
-}
-
-void dcn31_z10_restore(const struct dc *dc)
-{
-	union dmub_rb_cmd cmd;
-
-	/*
-	 * DMUB notifies whether restore is required.
-	 * Optimization to avoid sending commands when not required.
-	 */
-	if (!dc_dmub_srv_is_restore_required(dc->ctx->dmub_srv))
-		return;
-
-	memset(&cmd, 0, sizeof(cmd));
-	cmd.dcn_restore.header.type = DMUB_CMD__IDLE_OPT;
-	cmd.dcn_restore.header.sub_type = DMUB_CMD__IDLE_OPT_DCN_RESTORE;
-
-	dc_dmub_srv_cmd_queue(dc->ctx->dmub_srv, &cmd);
-	dc_dmub_srv_cmd_execute(dc->ctx->dmub_srv);
-	dc_dmub_srv_wait_idle(dc->ctx->dmub_srv);
-}
-
-void dcn31_hubp_pg_control(struct dce_hwseq *hws, unsigned int hubp_inst, bool power_on)
-{
-	uint32_t power_gate = power_on ? 0 : 1;
-	uint32_t pwr_status = power_on ? 0 : 2;
-
-	if (hws->ctx->dc->debug.disable_hubp_power_gate)
-		return;
-
-	if (REG(DOMAIN0_PG_CONFIG) == 0)
-		return;
-
-	switch (hubp_inst) {
-	case 0:
-		REG_SET(DOMAIN0_PG_CONFIG, 0, DOMAIN_POWER_GATE, power_gate);
-		REG_WAIT(DOMAIN0_PG_STATUS, DOMAIN_PGFSM_PWR_STATUS, pwr_status, 1, 1000);
-		break;
-	case 1:
-		REG_SET(DOMAIN1_PG_CONFIG, 0, DOMAIN_POWER_GATE, power_gate);
-		REG_WAIT(DOMAIN1_PG_STATUS, DOMAIN_PGFSM_PWR_STATUS, pwr_status, 1, 1000);
-		break;
-	case 2:
-		REG_SET(DOMAIN2_PG_CONFIG, 0, DOMAIN_POWER_GATE, power_gate);
-		REG_WAIT(DOMAIN2_PG_STATUS, DOMAIN_PGFSM_PWR_STATUS, pwr_status, 1, 1000);
-		break;
-	case 3:
-		REG_SET(DOMAIN3_PG_CONFIG, 0, DOMAIN_POWER_GATE, power_gate);
-		REG_WAIT(DOMAIN3_PG_STATUS, DOMAIN_PGFSM_PWR_STATUS, pwr_status, 1, 1000);
-		break;
-	default:
-		BREAK_TO_DEBUGGER();
-		break;
-	}
-}
-
-int dcn31_init_sys_ctx(struct dce_hwseq *hws, struct dc *dc, struct dc_phy_addr_space_config *pa_config)
-{
-	struct dcn_hubbub_phys_addr_config config;
-
-	config.system_aperture.fb_top = pa_config->system_aperture.fb_top;
-	config.system_aperture.fb_offset = pa_config->system_aperture.fb_offset;
-	config.system_aperture.fb_base = pa_config->system_aperture.fb_base;
-	config.system_aperture.agp_top = pa_config->system_aperture.agp_top;
-	config.system_aperture.agp_bot = pa_config->system_aperture.agp_bot;
-	config.system_aperture.agp_base = pa_config->system_aperture.agp_base;
-	config.gart_config.page_table_start_addr = pa_config->gart_config.page_table_start_addr;
-	config.gart_config.page_table_end_addr = pa_config->gart_config.page_table_end_addr;
-
-	if (pa_config->gart_config.base_addr_is_mc_addr) {
-		/* Convert from MC address to offset into FB */
-		config.gart_config.page_table_base_addr = pa_config->gart_config.page_table_base_addr -
-				pa_config->system_aperture.fb_base +
-				pa_config->system_aperture.fb_offset;
-	} else
-		config.gart_config.page_table_base_addr = pa_config->gart_config.page_table_base_addr;
-
-	return dc->res_pool->hubbub->funcs->init_dchub_sys_ctx(dc->res_pool->hubbub, &config);
-}
-
-static void dcn31_reset_back_end_for_pipe(
-		struct dc *dc,
-		struct pipe_ctx *pipe_ctx,
-		struct dc_state *context)
-{
-	struct dc_link *link;
-
-	DC_LOGGER_INIT(dc->ctx->logger);
-	if (pipe_ctx->stream_res.stream_enc == NULL) {
-		pipe_ctx->stream = NULL;
-		return;
-	}
-	ASSERT(!pipe_ctx->top_pipe);
-
-	dc->hwss.set_abm_immediate_disable(pipe_ctx);
-
-	pipe_ctx->stream_res.tg->funcs->set_dsc_config(
-			pipe_ctx->stream_res.tg,
-			OPTC_DSC_DISABLED, 0, 0);
-	pipe_ctx->stream_res.tg->funcs->disable_crtc(pipe_ctx->stream_res.tg);
-
-	pipe_ctx->stream_res.tg->funcs->enable_optc_clock(pipe_ctx->stream_res.tg, false);
-	if (pipe_ctx->stream_res.tg->funcs->set_odm_bypass)
-		pipe_ctx->stream_res.tg->funcs->set_odm_bypass(
-				pipe_ctx->stream_res.tg, &pipe_ctx->stream->timing);
-
-	if (pipe_ctx->stream_res.tg->funcs->set_drr)
-		pipe_ctx->stream_res.tg->funcs->set_drr(
-				pipe_ctx->stream_res.tg, NULL);
-
-	if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) {
-		link = pipe_ctx->stream->link;
-		/* DPMS may already disable or */
-		/* dpms_off status is incorrect due to fastboot
-		 * feature. When system resume from S4 with second
-		 * screen only, the dpms_off would be true but
-		 * VBIOS lit up eDP, so check link status too.
-		 */
-		if (!pipe_ctx->stream->dpms_off || link->link_status.link_active)
-			core_link_disable_stream(pipe_ctx);
-		else if (pipe_ctx->stream_res.audio)
-			dc->hwss.disable_audio_stream(pipe_ctx);
-
-		/* free acquired resources */
-		if (pipe_ctx->stream_res.audio) {
-			/*disable az_endpoint*/
-			pipe_ctx->stream_res.audio->funcs->az_disable(pipe_ctx->stream_res.audio);
-
-			/*free audio*/
-			if (dc->caps.dynamic_audio == true) {
-				/*we have to dynamic arbitrate the audio endpoints*/
-				/*we free the resource, need reset is_audio_acquired*/
-				update_audio_usage(&dc->current_state->res_ctx, dc->res_pool,
-						pipe_ctx->stream_res.audio, false);
-				pipe_ctx->stream_res.audio = NULL;
-			}
-		}
-	} else if (pipe_ctx->stream_res.dsc) {
-			dp_set_dsc_enable(pipe_ctx, false);
-	}
-
-	pipe_ctx->stream = NULL;
-	DC_LOG_DEBUG("Reset back end for pipe %d, tg:%d\n",
-					pipe_ctx->pipe_idx, pipe_ctx->stream_res.tg->inst);
-}
-
-void dcn31_reset_hw_ctx_wrap(
-		struct dc *dc,
-		struct dc_state *context)
-{
-	int i;
-	struct dce_hwseq *hws = dc->hwseq;
-
-	/* Reset Back End*/
-	for (i = dc->res_pool->pipe_count - 1; i >= 0 ; i--) {
-		struct pipe_ctx *pipe_ctx_old =
-			&dc->current_state->res_ctx.pipe_ctx[i];
-		struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i];
-
-		if (!pipe_ctx_old->stream)
-			continue;
-
-		if (pipe_ctx_old->top_pipe || pipe_ctx_old->prev_odm_pipe)
-			continue;
-
-		if (!pipe_ctx->stream ||
-				pipe_need_reprogram(pipe_ctx_old, pipe_ctx)) {
-			struct clock_source *old_clk = pipe_ctx_old->clock_source;
-
-			dcn31_reset_back_end_for_pipe(dc, pipe_ctx_old, dc->current_state);
-			if (hws->funcs.enable_stream_gating)
-				hws->funcs.enable_stream_gating(dc, pipe_ctx_old);
-			if (old_clk)
-				old_clk->funcs->cs_power_down(old_clk);
-		}
-	}
-
-	/* New dc_state in the process of being applied to hardware. */
-	dc->current_state->res_ctx.link_enc_cfg_ctx.mode = LINK_ENC_CFG_TRANSIENT;
-}
-
-void dcn31_setup_hpo_hw_control(const struct dce_hwseq *hws, bool enable)
-{
-	if (hws->ctx->dc->debug.hpo_optimization)
-		REG_UPDATE(HPO_TOP_HW_CONTROL, HPO_IO_EN, !!enable);
-}
diff --git a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_optc.c b/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_optc.c
deleted file mode 100644
index a4b1d98f0007..000000000000
--- a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_optc.c
+++ /dev/null
@@ -1,287 +0,0 @@
-/*
- * Copyright 2012-15 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: AMD
- *
- */
-
-#include "dcn31_optc.h"
-
-#include "dcn30/dcn30_optc.h"
-#include "reg_helper.h"
-#include "dc.h"
-#include "dcn_calc_math.h"
-
-#define REG(reg)\
-	optc1->tg_regs->reg
-
-#define CTX \
-	optc1->base.ctx
-
-#undef FN
-#define FN(reg_name, field_name) \
-	optc1->tg_shift->field_name, optc1->tg_mask->field_name
-
-static void optc31_set_odm_combine(struct timing_generator *optc, int *opp_id, int opp_cnt,
-		struct dc_crtc_timing *timing)
-{
-	struct optc *optc1 = DCN10TG_FROM_TG(optc);
-	int mpcc_hactive = (timing->h_addressable + timing->h_border_left + timing->h_border_right)
-			/ opp_cnt;
-	uint32_t memory_mask = 0;
-	int mem_count_per_opp = (mpcc_hactive + 2559) / 2560;
-
-	/* Assume less than 6 pipes */
-	if (opp_cnt == 4) {
-		if (mem_count_per_opp == 1)
-			memory_mask = 0xf;
-		else {
-			ASSERT(mem_count_per_opp == 2);
-			memory_mask = 0xff;
-		}
-	} else if (mem_count_per_opp == 1)
-		memory_mask = 0x1 << (opp_id[0] * 2) | 0x1 << (opp_id[1] * 2);
-	else if (mem_count_per_opp == 2)
-		memory_mask = 0x3 << (opp_id[0] * 2) | 0x3 << (opp_id[1] * 2);
-	else if (mem_count_per_opp == 3)
-		memory_mask = 0x77;
-	else if (mem_count_per_opp == 4)
-		memory_mask = 0xff;
-
-	if (REG(OPTC_MEMORY_CONFIG))
-		REG_SET(OPTC_MEMORY_CONFIG, 0,
-			OPTC_MEM_SEL, memory_mask);
-
-	if (opp_cnt == 2) {
-		REG_SET_3(OPTC_DATA_SOURCE_SELECT, 0,
-				OPTC_NUM_OF_INPUT_SEGMENT, 1,
-				OPTC_SEG0_SRC_SEL, opp_id[0],
-				OPTC_SEG1_SRC_SEL, opp_id[1]);
-	} else if (opp_cnt == 4) {
-		REG_SET_5(OPTC_DATA_SOURCE_SELECT, 0,
-				OPTC_NUM_OF_INPUT_SEGMENT, 3,
-				OPTC_SEG0_SRC_SEL, opp_id[0],
-				OPTC_SEG1_SRC_SEL, opp_id[1],
-				OPTC_SEG2_SRC_SEL, opp_id[2],
-				OPTC_SEG3_SRC_SEL, opp_id[3]);
-	}
-
-	REG_UPDATE(OPTC_WIDTH_CONTROL,
-			OPTC_SEGMENT_WIDTH, mpcc_hactive);
-
-	REG_SET(OTG_H_TIMING_CNTL, 0, OTG_H_TIMING_DIV_MODE, opp_cnt - 1);
-	optc1->opp_count = opp_cnt;
-}
-
-/**
- * Enable CRTC
- * Enable CRTC - call ASIC Control Object to enable Timing generator.
- */
-static bool optc31_enable_crtc(struct timing_generator *optc)
-{
-	struct optc *optc1 = DCN10TG_FROM_TG(optc);
-
-	/* opp instance for OTG, 1 to 1 mapping and odm will adjust */
-	REG_UPDATE(OPTC_DATA_SOURCE_SELECT,
-			OPTC_SEG0_SRC_SEL, optc->inst);
-
-	/* VTG enable first is for HW workaround */
-	REG_UPDATE(CONTROL,
-			VTG0_ENABLE, 1);
-
-	REG_SEQ_START();
-
-	/* Enable CRTC */
-	REG_UPDATE_2(OTG_CONTROL,
-			OTG_DISABLE_POINT_CNTL, 2,
-			OTG_MASTER_EN, 1);
-
-	REG_SEQ_SUBMIT();
-	REG_SEQ_WAIT_DONE();
-
-	return true;
-}
-
-/* disable_crtc - call ASIC Control Object to disable Timing generator. */
-static bool optc31_disable_crtc(struct timing_generator *optc)
-{
-	struct optc *optc1 = DCN10TG_FROM_TG(optc);
-
-	/* disable otg request until end of the first line
-	 * in the vertical blank region
-	 */
-	REG_UPDATE(OTG_CONTROL,
-			OTG_MASTER_EN, 0);
-
-	REG_UPDATE(CONTROL,
-			VTG0_ENABLE, 0);
-
-	/* CRTC disabled, so disable  clock. */
-	REG_WAIT(OTG_CLOCK_CONTROL,
-			OTG_BUSY, 0,
-			1, 100000);
-
-	return true;
-}
-
-static bool optc31_immediate_disable_crtc(struct timing_generator *optc)
-{
-	struct optc *optc1 = DCN10TG_FROM_TG(optc);
-
-	REG_UPDATE_2(OTG_CONTROL,
-			OTG_DISABLE_POINT_CNTL, 0,
-			OTG_MASTER_EN, 0);
-
-	REG_UPDATE(CONTROL,
-			VTG0_ENABLE, 0);
-
-	/* CRTC disabled, so disable  clock. */
-	REG_WAIT(OTG_CLOCK_CONTROL,
-			OTG_BUSY, 0,
-			1, 100000);
-
-	return true;
-}
-
-static void optc31_set_drr(
-	struct timing_generator *optc,
-	const struct drr_params *params)
-{
-	struct optc *optc1 = DCN10TG_FROM_TG(optc);
-
-	if (params != NULL &&
-		params->vertical_total_max > 0 &&
-		params->vertical_total_min > 0) {
-
-		if (params->vertical_total_mid != 0) {
-
-			REG_SET(OTG_V_TOTAL_MID, 0,
-				OTG_V_TOTAL_MID, params->vertical_total_mid - 1);
-
-			REG_UPDATE_2(OTG_V_TOTAL_CONTROL,
-					OTG_VTOTAL_MID_REPLACING_MAX_EN, 1,
-					OTG_VTOTAL_MID_FRAME_NUM,
-					(uint8_t)params->vertical_total_mid_frame_num);
-
-		}
-
-		optc->funcs->set_vtotal_min_max(optc, params->vertical_total_min - 1, params->vertical_total_max - 1);
-
-		/*
-		 * MIN_MASK_EN is gone and MASK is now always enabled.
-		 *
-		 * To get it to it work with manual trigger we need to make sure
-		 * we program the correct bit.
-		 */
-		REG_UPDATE_4(OTG_V_TOTAL_CONTROL,
-				OTG_V_TOTAL_MIN_SEL, 1,
-				OTG_V_TOTAL_MAX_SEL, 1,
-				OTG_FORCE_LOCK_ON_EVENT, 0,
-				OTG_SET_V_TOTAL_MIN_MASK, (1 << 1)); /* TRIGA */
-
-		// Setup manual flow control for EOF via TRIG_A
-		optc->funcs->setup_manual_trigger(optc);
-
-	} else {
-		REG_UPDATE_4(OTG_V_TOTAL_CONTROL,
-				OTG_SET_V_TOTAL_MIN_MASK, 0,
-				OTG_V_TOTAL_MIN_SEL, 0,
-				OTG_V_TOTAL_MAX_SEL, 0,
-				OTG_FORCE_LOCK_ON_EVENT, 0);
-
-		optc->funcs->set_vtotal_min_max(optc, 0, 0);
-	}
-}
-
-static struct timing_generator_funcs dcn31_tg_funcs = {
-		.validate_timing = optc1_validate_timing,
-		.program_timing = optc1_program_timing,
-		.setup_vertical_interrupt0 = optc1_setup_vertical_interrupt0,
-		.setup_vertical_interrupt1 = optc1_setup_vertical_interrupt1,
-		.setup_vertical_interrupt2 = optc1_setup_vertical_interrupt2,
-		.program_global_sync = optc1_program_global_sync,
-		.enable_crtc = optc31_enable_crtc,
-		.disable_crtc = optc31_disable_crtc,
-		.immediate_disable_crtc = optc31_immediate_disable_crtc,
-		/* used by enable_timing_synchronization. Not need for FPGA */
-		.is_counter_moving = optc1_is_counter_moving,
-		.get_position = optc1_get_position,
-		.get_frame_count = optc1_get_vblank_counter,
-		.get_scanoutpos = optc1_get_crtc_scanoutpos,
-		.get_otg_active_size = optc1_get_otg_active_size,
-		.set_early_control = optc1_set_early_control,
-		/* used by enable_timing_synchronization. Not need for FPGA */
-		.wait_for_state = optc1_wait_for_state,
-		.set_blank_color = optc3_program_blank_color,
-		.did_triggered_reset_occur = optc1_did_triggered_reset_occur,
-		.triplebuffer_lock = optc3_triplebuffer_lock,
-		.triplebuffer_unlock = optc2_triplebuffer_unlock,
-		.enable_reset_trigger = optc1_enable_reset_trigger,
-		.enable_crtc_reset = optc1_enable_crtc_reset,
-		.disable_reset_trigger = optc1_disable_reset_trigger,
-		.lock = optc3_lock,
-		.is_locked = optc1_is_locked,
-		.unlock = optc1_unlock,
-		.lock_doublebuffer_enable = optc3_lock_doublebuffer_enable,
-		.lock_doublebuffer_disable = optc3_lock_doublebuffer_disable,
-		.enable_optc_clock = optc1_enable_optc_clock,
-		.set_drr = optc31_set_drr,
-		.set_vtotal_min_max = optc1_set_vtotal_min_max,
-		.set_static_screen_control = optc1_set_static_screen_control,
-		.program_stereo = optc1_program_stereo,
-		.is_stereo_left_eye = optc1_is_stereo_left_eye,
-		.tg_init = optc3_tg_init,
-		.is_tg_enabled = optc1_is_tg_enabled,
-		.is_optc_underflow_occurred = optc1_is_optc_underflow_occurred,
-		.clear_optc_underflow = optc1_clear_optc_underflow,
-		.setup_global_swap_lock = NULL,
-		.get_crc = optc1_get_crc,
-		.configure_crc = optc2_configure_crc,
-		.set_dsc_config = optc3_set_dsc_config,
-		.set_dwb_source = NULL,
-		.set_odm_bypass = optc3_set_odm_bypass,
-		.set_odm_combine = optc31_set_odm_combine,
-		.get_optc_source = optc2_get_optc_source,
-		.set_out_mux = optc3_set_out_mux,
-		.set_drr_trigger_window = optc3_set_drr_trigger_window,
-		.set_vtotal_change_limit = optc3_set_vtotal_change_limit,
-		.set_gsl = optc2_set_gsl,
-		.set_gsl_source_select = optc2_set_gsl_source_select,
-		.set_vtg_params = optc1_set_vtg_params,
-		.program_manual_trigger = optc2_program_manual_trigger,
-		.setup_manual_trigger = optc2_setup_manual_trigger,
-		.get_hw_timing = optc1_get_hw_timing,
-};
-
-void dcn31_timing_generator_init(struct optc *optc1)
-{
-	optc1->base.funcs = &dcn31_tg_funcs;
-
-	optc1->max_h_total = optc1->tg_mask->OTG_H_TOTAL + 1;
-	optc1->max_v_total = optc1->tg_mask->OTG_V_TOTAL + 1;
-
-	optc1->min_h_blank = 32;
-	optc1->min_v_blank = 3;
-	optc1->min_v_blank_interlace = 5;
-	optc1->min_h_sync_width = 4;
-	optc1->min_v_sync_width = 1;
-}
-
diff --git a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_panel_cntl.c b/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_panel_cntl.c
index 3b3721386571..f9961a6446f3 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_panel_cntl.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_panel_cntl.c
@@ -50,9 +50,9 @@ static bool dcn31_query_backlight_info(struct panel_cntl *panel_cntl, union dmub
 	cmd->panel_cntl.header.type = DMUB_CMD__PANEL_CNTL;
 	cmd->panel_cntl.header.sub_type = DMUB_CMD__PANEL_CNTL_QUERY_BACKLIGHT_INFO;
 	cmd->panel_cntl.header.payload_bytes = sizeof(cmd->panel_cntl.data);
-	cmd->panel_cntl.data.inst = dcn31_panel_cntl->base.inst;
+	cmd->panel_cntl.data.pwrseq_inst = dcn31_panel_cntl->base.pwrseq_inst;
 
-	return dc_dmub_srv_cmd_with_reply_data(dc_dmub_srv, cmd);
+	return dc_wake_and_execute_dmub_cmd(dc_dmub_srv->ctx, cmd, DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY);
 }
 
 static uint32_t dcn31_get_16_bit_backlight_from_pwm(struct panel_cntl *panel_cntl)
@@ -65,11 +65,12 @@ static uint32_t dcn31_get_16_bit_backlight_from_pwm(struct panel_cntl *panel_cnt
 	return cmd.panel_cntl.data.current_backlight;
 }
 
-uint32_t dcn31_panel_cntl_hw_init(struct panel_cntl *panel_cntl)
+static uint32_t dcn31_panel_cntl_hw_init(struct panel_cntl *panel_cntl)
 {
 	struct dcn31_panel_cntl *dcn31_panel_cntl = TO_DCN31_PANEL_CNTL(panel_cntl);
 	struct dc_dmub_srv *dc_dmub_srv = panel_cntl->ctx->dmub_srv;
 	union dmub_rb_cmd cmd;
+	uint32_t freq_to_set = panel_cntl->ctx->dc->debug.pwm_freq;
 
 	if (!dc_dmub_srv)
 		return 0;
@@ -78,13 +79,14 @@ uint32_t dcn31_panel_cntl_hw_init(struct panel_cntl *panel_cntl)
 	cmd.panel_cntl.header.type = DMUB_CMD__PANEL_CNTL;
 	cmd.panel_cntl.header.sub_type = DMUB_CMD__PANEL_CNTL_HW_INIT;
 	cmd.panel_cntl.header.payload_bytes = sizeof(cmd.panel_cntl.data);
-	cmd.panel_cntl.data.inst = dcn31_panel_cntl->base.inst;
+	cmd.panel_cntl.data.pwrseq_inst = dcn31_panel_cntl->base.pwrseq_inst;
 	cmd.panel_cntl.data.bl_pwm_cntl = panel_cntl->stored_backlight_registers.BL_PWM_CNTL;
 	cmd.panel_cntl.data.bl_pwm_period_cntl = panel_cntl->stored_backlight_registers.BL_PWM_PERIOD_CNTL;
 	cmd.panel_cntl.data.bl_pwm_ref_div1 =
 		panel_cntl->stored_backlight_registers.LVTMA_PWRSEQ_REF_DIV_BL_PWM_REF_DIV;
-
-	if (!dc_dmub_srv_cmd_with_reply_data(dc_dmub_srv, &cmd))
+	cmd.panel_cntl.data.bl_pwm_ref_div2 =
+		panel_cntl->stored_backlight_registers.PANEL_PWRSEQ_REF_DIV2;
+	if (!dc_wake_and_execute_dmub_cmd(dc_dmub_srv->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY))
 		return 0;
 
 	panel_cntl->stored_backlight_registers.BL_PWM_CNTL = cmd.panel_cntl.data.bl_pwm_cntl;
@@ -92,11 +94,26 @@ uint32_t dcn31_panel_cntl_hw_init(struct panel_cntl *panel_cntl)
 	panel_cntl->stored_backlight_registers.BL_PWM_PERIOD_CNTL = cmd.panel_cntl.data.bl_pwm_period_cntl;
 	panel_cntl->stored_backlight_registers.LVTMA_PWRSEQ_REF_DIV_BL_PWM_REF_DIV =
 		cmd.panel_cntl.data.bl_pwm_ref_div1;
-
+	panel_cntl->stored_backlight_registers.PANEL_PWRSEQ_REF_DIV2 =
+		cmd.panel_cntl.data.bl_pwm_ref_div2;
+
+	if (freq_to_set >= MIN_DEBUG_FREQ_HZ && freq_to_set <= MAX_DEBUG_FREQ_HZ) {
+		uint32_t xtal = panel_cntl->ctx->dc->res_pool->ref_clocks.dccg_ref_clock_inKhz;
+
+		memset(&cmd, 0, sizeof(cmd));
+		cmd.panel_cntl.header.type = DMUB_CMD__PANEL_CNTL;
+		cmd.panel_cntl.header.sub_type = DMUB_CMD__PANEL_DEBUG_PWM_FREQ;
+		cmd.panel_cntl.header.payload_bytes = sizeof(cmd.panel_cntl.data);
+		cmd.panel_cntl.data.pwrseq_inst = dcn31_panel_cntl->base.pwrseq_inst;
+		cmd.panel_cntl.data.bl_pwm_cntl = xtal;
+		cmd.panel_cntl.data.bl_pwm_period_cntl = freq_to_set;
+		if (!dc_wake_and_execute_dmub_cmd(dc_dmub_srv->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY))
+			return 0;
+	}
 	return cmd.panel_cntl.data.current_backlight;
 }
 
-void dcn31_panel_cntl_destroy(struct panel_cntl **panel_cntl)
+static void dcn31_panel_cntl_destroy(struct panel_cntl **panel_cntl)
 {
 	struct dcn31_panel_cntl *dcn31_panel_cntl = TO_DCN31_PANEL_CNTL(*panel_cntl);
 
@@ -104,7 +121,7 @@ void dcn31_panel_cntl_destroy(struct panel_cntl **panel_cntl)
 	*panel_cntl = NULL;
 }
 
-bool dcn31_is_panel_backlight_on(struct panel_cntl *panel_cntl)
+static bool dcn31_is_panel_backlight_on(struct panel_cntl *panel_cntl)
 {
 	union dmub_rb_cmd cmd;
 
@@ -114,7 +131,7 @@ bool dcn31_is_panel_backlight_on(struct panel_cntl *panel_cntl)
 	return cmd.panel_cntl.data.is_backlight_on;
 }
 
-bool dcn31_is_panel_powered_on(struct panel_cntl *panel_cntl)
+static bool dcn31_is_panel_powered_on(struct panel_cntl *panel_cntl)
 {
 	union dmub_rb_cmd cmd;
 
@@ -124,7 +141,7 @@ bool dcn31_is_panel_powered_on(struct panel_cntl *panel_cntl)
 	return cmd.panel_cntl.data.is_powered_on;
 }
 
-void dcn31_store_backlight_level(struct panel_cntl *panel_cntl)
+static void dcn31_store_backlight_level(struct panel_cntl *panel_cntl)
 {
 	union dmub_rb_cmd cmd;
 
@@ -151,7 +168,33 @@ void dcn31_panel_cntl_construct(
 	struct dcn31_panel_cntl *dcn31_panel_cntl,
 	const struct panel_cntl_init_data *init_data)
 {
+
 	dcn31_panel_cntl->base.funcs = &dcn31_link_panel_cntl_funcs;
 	dcn31_panel_cntl->base.ctx = init_data->ctx;
 	dcn31_panel_cntl->base.inst = init_data->inst;
+
+	if (dcn31_panel_cntl->base.ctx->dc->config.support_edp0_on_dp1) {
+		//If supported, power sequencer mapping shall follow the DIG instance
+		uint8_t pwrseq_inst = 0xF;
+
+		switch (init_data->eng_id) {
+		case ENGINE_ID_DIGA:
+			pwrseq_inst = 0;
+			break;
+		case ENGINE_ID_DIGB:
+			pwrseq_inst = 1;
+			break;
+		default:
+			DC_LOG_WARNING("Unsupported pwrseq engine id: %d!\n", init_data->eng_id);
+			ASSERT(false);
+			break;
+		}
+
+		dcn31_panel_cntl->base.pwrseq_inst = pwrseq_inst;
+	} else {
+		/* If not supported, pwrseq will be assigned in order,
+		 * so first pwrseq will be assigned to first panel instance (legacy behavior)
+		 */
+		dcn31_panel_cntl->base.pwrseq_inst = dcn31_panel_cntl->base.inst;
+	}
 }
diff --git a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_panel_cntl.h b/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_panel_cntl.h
index d33ccd6ef8c3..8cf0259e211e 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_panel_cntl.h
+++ b/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_panel_cntl.h
@@ -29,6 +29,9 @@
 #include "panel_cntl.h"
 #include "dce/dce_panel_cntl.h"
 
+#define MIN_DEBUG_FREQ_HZ 200
+#define MAX_DEBUG_FREQ_HZ 6250
+
 struct dcn31_panel_cntl {
 	struct panel_cntl base;
 };
diff --git a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_resource.c b/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_resource.c
deleted file mode 100644
index 18896294ae12..000000000000
--- a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_resource.c
+++ /dev/null
@@ -1,2493 +0,0 @@
-/*
- * Copyright 2019 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: AMD
- *
- */
-
-
-#include "dm_services.h"
-#include "dc.h"
-
-#include "dcn31/dcn31_init.h"
-
-#include "resource.h"
-#include "include/irq_service_interface.h"
-#include "dcn31_resource.h"
-
-#include "dcn20/dcn20_resource.h"
-#include "dcn30/dcn30_resource.h"
-
-#include "dcn10/dcn10_ipp.h"
-#include "dcn30/dcn30_hubbub.h"
-#include "dcn31/dcn31_hubbub.h"
-#include "dcn30/dcn30_mpc.h"
-#include "dcn31/dcn31_hubp.h"
-#include "irq/dcn31/irq_service_dcn31.h"
-#include "dcn30/dcn30_dpp.h"
-#include "dcn31/dcn31_optc.h"
-#include "dcn20/dcn20_hwseq.h"
-#include "dcn30/dcn30_hwseq.h"
-#include "dce110/dce110_hw_sequencer.h"
-#include "dcn30/dcn30_opp.h"
-#include "dcn20/dcn20_dsc.h"
-#include "dcn30/dcn30_vpg.h"
-#include "dcn30/dcn30_afmt.h"
-#include "dcn30/dcn30_dio_stream_encoder.h"
-#include "dcn31/dcn31_hpo_dp_stream_encoder.h"
-#include "dcn31/dcn31_hpo_dp_link_encoder.h"
-#include "dcn31/dcn31_apg.h"
-#include "dcn31/dcn31_dio_link_encoder.h"
-#include "dcn31/dcn31_vpg.h"
-#include "dcn31/dcn31_afmt.h"
-#include "dce/dce_clock_source.h"
-#include "dce/dce_audio.h"
-#include "dce/dce_hwseq.h"
-#include "clk_mgr.h"
-#include "virtual/virtual_stream_encoder.h"
-#include "dce110/dce110_resource.h"
-#include "dml/display_mode_vba.h"
-#include "dcn31/dcn31_dccg.h"
-#include "dcn10/dcn10_resource.h"
-#include "dcn31_panel_cntl.h"
-
-#include "dcn30/dcn30_dwb.h"
-#include "dcn30/dcn30_mmhubbub.h"
-
-// TODO: change include headers /amd/include/asic_reg after upstream
-#include "yellow_carp_offset.h"
-#include "dcn/dcn_3_1_2_offset.h"
-#include "dcn/dcn_3_1_2_sh_mask.h"
-#include "nbio/nbio_7_2_0_offset.h"
-#include "dpcs/dpcs_4_2_0_offset.h"
-#include "dpcs/dpcs_4_2_0_sh_mask.h"
-#include "mmhub/mmhub_2_3_0_offset.h"
-#include "mmhub/mmhub_2_3_0_sh_mask.h"
-
-
-#define regDCHUBBUB_DEBUG_CTRL_0                                              0x04d6
-#define regDCHUBBUB_DEBUG_CTRL_0_BASE_IDX                                     2
-#define DCHUBBUB_DEBUG_CTRL_0__DET_DEPTH__SHIFT                               0x10
-#define DCHUBBUB_DEBUG_CTRL_0__DET_DEPTH_MASK                                 0x01FF0000L
-
-#include "reg_helper.h"
-#include "dce/dmub_abm.h"
-#include "dce/dmub_psr.h"
-#include "dce/dce_aux.h"
-#include "dce/dce_i2c.h"
-
-#include "dml/dcn30/display_mode_vba_30.h"
-#include "vm_helper.h"
-#include "dcn20/dcn20_vmid.h"
-
-#include "link_enc_cfg.h"
-
-#define DC_LOGGER_INIT(logger)
-
-#define DCN3_1_DEFAULT_DET_SIZE 384
-
-struct _vcs_dpi_ip_params_st dcn3_1_ip = {
-	.gpuvm_enable = 1,
-	.gpuvm_max_page_table_levels = 1,
-	.hostvm_enable = 1,
-	.hostvm_max_page_table_levels = 2,
-	.rob_buffer_size_kbytes = 64,
-	.det_buffer_size_kbytes = DCN3_1_DEFAULT_DET_SIZE,
-	.config_return_buffer_size_in_kbytes = 1792,
-	.compressed_buffer_segment_size_in_kbytes = 64,
-	.meta_fifo_size_in_kentries = 32,
-	.zero_size_buffer_entries = 512,
-	.compbuf_reserved_space_64b = 256,
-	.compbuf_reserved_space_zs = 64,
-	.dpp_output_buffer_pixels = 2560,
-	.opp_output_buffer_lines = 1,
-	.pixel_chunk_size_kbytes = 8,
-	.meta_chunk_size_kbytes = 2,
-	.min_meta_chunk_size_bytes = 256,
-	.writeback_chunk_size_kbytes = 8,
-	.ptoi_supported = false,
-	.num_dsc = 3,
-	.maximum_dsc_bits_per_component = 10,
-	.dsc422_native_support = false,
-	.is_line_buffer_bpp_fixed = true,
-	.line_buffer_fixed_bpp = 48,
-	.line_buffer_size_bits = 789504,
-	.max_line_buffer_lines = 12,
-	.writeback_interface_buffer_size_kbytes = 90,
-	.max_num_dpp = 4,
-	.max_num_otg = 4,
-	.max_num_hdmi_frl_outputs = 1,
-	.max_num_wb = 1,
-	.max_dchub_pscl_bw_pix_per_clk = 4,
-	.max_pscl_lb_bw_pix_per_clk = 2,
-	.max_lb_vscl_bw_pix_per_clk = 4,
-	.max_vscl_hscl_bw_pix_per_clk = 4,
-	.max_hscl_ratio = 6,
-	.max_vscl_ratio = 6,
-	.max_hscl_taps = 8,
-	.max_vscl_taps = 8,
-	.dpte_buffer_size_in_pte_reqs_luma = 64,
-	.dpte_buffer_size_in_pte_reqs_chroma = 34,
-	.dispclk_ramp_margin_percent = 1,
-	.max_inter_dcn_tile_repeaters = 8,
-	.cursor_buffer_size = 16,
-	.cursor_chunk_size = 2,
-	.writeback_line_buffer_buffer_size = 0,
-	.writeback_min_hscl_ratio = 1,
-	.writeback_min_vscl_ratio = 1,
-	.writeback_max_hscl_ratio = 1,
-	.writeback_max_vscl_ratio = 1,
-	.writeback_max_hscl_taps = 1,
-	.writeback_max_vscl_taps = 1,
-	.dppclk_delay_subtotal = 46,
-	.dppclk_delay_scl = 50,
-	.dppclk_delay_scl_lb_only = 16,
-	.dppclk_delay_cnvc_formatter = 27,
-	.dppclk_delay_cnvc_cursor = 6,
-	.dispclk_delay_subtotal = 119,
-	.dynamic_metadata_vm_enabled = false,
-	.odm_combine_4to1_supported = false,
-	.dcc_supported = true,
-};
-
-struct _vcs_dpi_soc_bounding_box_st dcn3_1_soc = {
-		/*TODO: correct dispclk/dppclk voltage level determination*/
-	.clock_limits = {
-		{
-			.state = 0,
-			.dispclk_mhz = 1200.0,
-			.dppclk_mhz = 1200.0,
-			.phyclk_mhz = 600.0,
-			.phyclk_d18_mhz = 667.0,
-			.dscclk_mhz = 186.0,
-			.dtbclk_mhz = 625.0,
-		},
-		{
-			.state = 1,
-			.dispclk_mhz = 1200.0,
-			.dppclk_mhz = 1200.0,
-			.phyclk_mhz = 810.0,
-			.phyclk_d18_mhz = 667.0,
-			.dscclk_mhz = 209.0,
-			.dtbclk_mhz = 625.0,
-		},
-		{
-			.state = 2,
-			.dispclk_mhz = 1200.0,
-			.dppclk_mhz = 1200.0,
-			.phyclk_mhz = 810.0,
-			.phyclk_d18_mhz = 667.0,
-			.dscclk_mhz = 209.0,
-			.dtbclk_mhz = 625.0,
-		},
-		{
-			.state = 3,
-			.dispclk_mhz = 1200.0,
-			.dppclk_mhz = 1200.0,
-			.phyclk_mhz = 810.0,
-			.phyclk_d18_mhz = 667.0,
-			.dscclk_mhz = 371.0,
-			.dtbclk_mhz = 625.0,
-		},
-		{
-			.state = 4,
-			.dispclk_mhz = 1200.0,
-			.dppclk_mhz = 1200.0,
-			.phyclk_mhz = 810.0,
-			.phyclk_d18_mhz = 667.0,
-			.dscclk_mhz = 417.0,
-			.dtbclk_mhz = 625.0,
-		},
-	},
-	.num_states = 5,
-	.sr_exit_time_us = 9.0,
-	.sr_enter_plus_exit_time_us = 11.0,
-	.sr_exit_z8_time_us = 442.0,
-	.sr_enter_plus_exit_z8_time_us = 560.0,
-	.writeback_latency_us = 12.0,
-	.dram_channel_width_bytes = 4,
-	.round_trip_ping_latency_dcfclk_cycles = 106,
-	.urgent_latency_pixel_data_only_us = 4.0,
-	.urgent_latency_pixel_mixed_with_vm_data_us = 4.0,
-	.urgent_latency_vm_data_only_us = 4.0,
-	.urgent_out_of_order_return_per_channel_pixel_only_bytes = 4096,
-	.urgent_out_of_order_return_per_channel_pixel_and_vm_bytes = 4096,
-	.urgent_out_of_order_return_per_channel_vm_only_bytes = 4096,
-	.pct_ideal_sdp_bw_after_urgent = 80.0,
-	.pct_ideal_dram_sdp_bw_after_urgent_pixel_only = 65.0,
-	.pct_ideal_dram_sdp_bw_after_urgent_pixel_and_vm = 60.0,
-	.pct_ideal_dram_sdp_bw_after_urgent_vm_only = 30.0,
-	.max_avg_sdp_bw_use_normal_percent = 60.0,
-	.max_avg_dram_bw_use_normal_percent = 60.0,
-	.fabric_datapath_to_dcn_data_return_bytes = 32,
-	.return_bus_width_bytes = 64,
-	.downspread_percent = 0.38,
-	.dcn_downspread_percent = 0.5,
-	.gpuvm_min_page_size_bytes = 4096,
-	.hostvm_min_page_size_bytes = 4096,
-	.do_urgent_latency_adjustment = false,
-	.urgent_latency_adjustment_fabric_clock_component_us = 0,
-	.urgent_latency_adjustment_fabric_clock_reference_mhz = 0,
-};
-
-enum dcn31_clk_src_array_id {
-	DCN31_CLK_SRC_PLL0,
-	DCN31_CLK_SRC_PLL1,
-	DCN31_CLK_SRC_PLL2,
-	DCN31_CLK_SRC_PLL3,
-	DCN31_CLK_SRC_PLL4,
-	DCN30_CLK_SRC_TOTAL
-};
-
-/* begin *********************
- * macros to expend register list macro defined in HW object header file
- */
-
-/* DCN */
-/* TODO awful hack. fixup dcn20_dwb.h */
-#undef BASE_INNER
-#define BASE_INNER(seg) DCN_BASE__INST0_SEG ## seg
-
-#define BASE(seg) BASE_INNER(seg)
-
-#define SR(reg_name)\
-		.reg_name = BASE(reg ## reg_name ## _BASE_IDX) +  \
-					reg ## reg_name
-
-#define SRI(reg_name, block, id)\
-	.reg_name = BASE(reg ## block ## id ## _ ## reg_name ## _BASE_IDX) + \
-					reg ## block ## id ## _ ## reg_name
-
-#define SRI2(reg_name, block, id)\
-	.reg_name = BASE(reg ## reg_name ## _BASE_IDX) + \
-					reg ## reg_name
-
-#define SRIR(var_name, reg_name, block, id)\
-	.var_name = BASE(reg ## block ## id ## _ ## reg_name ## _BASE_IDX) + \
-					reg ## block ## id ## _ ## reg_name
-
-#define SRII(reg_name, block, id)\
-	.reg_name[id] = BASE(reg ## block ## id ## _ ## reg_name ## _BASE_IDX) + \
-					reg ## block ## id ## _ ## reg_name
-
-#define SRII_MPC_RMU(reg_name, block, id)\
-	.RMU##_##reg_name[id] = BASE(reg ## block ## id ## _ ## reg_name ## _BASE_IDX) + \
-					reg ## block ## id ## _ ## reg_name
-
-#define SRII_DWB(reg_name, temp_name, block, id)\
-	.reg_name[id] = BASE(reg ## block ## id ## _ ## temp_name ## _BASE_IDX) + \
-					reg ## block ## id ## _ ## temp_name
-
-#define DCCG_SRII(reg_name, block, id)\
-	.block ## _ ## reg_name[id] = BASE(reg ## block ## id ## _ ## reg_name ## _BASE_IDX) + \
-					reg ## block ## id ## _ ## reg_name
-
-#define VUPDATE_SRII(reg_name, block, id)\
-	.reg_name[id] = BASE(reg ## reg_name ## _ ## block ## id ## _BASE_IDX) + \
-					reg ## reg_name ## _ ## block ## id
-
-/* NBIO */
-#define NBIO_BASE_INNER(seg) \
-	NBIO_BASE__INST0_SEG ## seg
-
-#define NBIO_BASE(seg) \
-	NBIO_BASE_INNER(seg)
-
-#define NBIO_SR(reg_name)\
-		.reg_name = NBIO_BASE(regBIF_BX1_ ## reg_name ## _BASE_IDX) + \
-					regBIF_BX1_ ## reg_name
-
-/* MMHUB */
-#define MMHUB_BASE_INNER(seg) \
-	MMHUB_BASE__INST0_SEG ## seg
-
-#define MMHUB_BASE(seg) \
-	MMHUB_BASE_INNER(seg)
-
-#define MMHUB_SR(reg_name)\
-		.reg_name = MMHUB_BASE(mm ## reg_name ## _BASE_IDX) + \
-					mm ## reg_name
-
-/* CLOCK */
-#define CLK_BASE_INNER(seg) \
-	CLK_BASE__INST0_SEG ## seg
-
-#define CLK_BASE(seg) \
-	CLK_BASE_INNER(seg)
-
-#define CLK_SRI(reg_name, block, inst)\
-	.reg_name = CLK_BASE(reg ## block ## _ ## inst ## _ ## reg_name ## _BASE_IDX) + \
-					reg ## block ## _ ## inst ## _ ## reg_name
-
-
-static const struct bios_registers bios_regs = {
-		NBIO_SR(BIOS_SCRATCH_3),
-		NBIO_SR(BIOS_SCRATCH_6)
-};
-
-#define clk_src_regs(index, pllid)\
-[index] = {\
-	CS_COMMON_REG_LIST_DCN3_0(index, pllid),\
-}
-
-static const struct dce110_clk_src_regs clk_src_regs[] = {
-	clk_src_regs(0, A),
-	clk_src_regs(1, B),
-	clk_src_regs(2, C),
-	clk_src_regs(3, D),
-	clk_src_regs(4, E)
-};
-
-static const struct dce110_clk_src_shift cs_shift = {
-		CS_COMMON_MASK_SH_LIST_DCN2_0(__SHIFT)
-};
-
-static const struct dce110_clk_src_mask cs_mask = {
-		CS_COMMON_MASK_SH_LIST_DCN2_0(_MASK)
-};
-
-#define abm_regs(id)\
-[id] = {\
-		ABM_DCN302_REG_LIST(id)\
-}
-
-static const struct dce_abm_registers abm_regs[] = {
-		abm_regs(0),
-		abm_regs(1),
-		abm_regs(2),
-		abm_regs(3),
-};
-
-static const struct dce_abm_shift abm_shift = {
-		ABM_MASK_SH_LIST_DCN30(__SHIFT)
-};
-
-static const struct dce_abm_mask abm_mask = {
-		ABM_MASK_SH_LIST_DCN30(_MASK)
-};
-
-#define audio_regs(id)\
-[id] = {\
-		AUD_COMMON_REG_LIST(id)\
-}
-
-static const struct dce_audio_registers audio_regs[] = {
-	audio_regs(0),
-	audio_regs(1),
-	audio_regs(2),
-	audio_regs(3),
-	audio_regs(4),
-	audio_regs(5),
-	audio_regs(6)
-};
-
-#define DCE120_AUD_COMMON_MASK_SH_LIST(mask_sh)\
-		SF(AZF0ENDPOINT0_AZALIA_F0_CODEC_ENDPOINT_INDEX, AZALIA_ENDPOINT_REG_INDEX, mask_sh),\
-		SF(AZF0ENDPOINT0_AZALIA_F0_CODEC_ENDPOINT_DATA, AZALIA_ENDPOINT_REG_DATA, mask_sh),\
-		AUD_COMMON_MASK_SH_LIST_BASE(mask_sh)
-
-static const struct dce_audio_shift audio_shift = {
-		DCE120_AUD_COMMON_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dce_audio_mask audio_mask = {
-		DCE120_AUD_COMMON_MASK_SH_LIST(_MASK)
-};
-
-#define vpg_regs(id)\
-[id] = {\
-	VPG_DCN31_REG_LIST(id)\
-}
-
-static const struct dcn31_vpg_registers vpg_regs[] = {
-	vpg_regs(0),
-	vpg_regs(1),
-	vpg_regs(2),
-	vpg_regs(3),
-	vpg_regs(4),
-	vpg_regs(5),
-	vpg_regs(6),
-	vpg_regs(7),
-	vpg_regs(8),
-	vpg_regs(9),
-};
-
-static const struct dcn31_vpg_shift vpg_shift = {
-	DCN31_VPG_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dcn31_vpg_mask vpg_mask = {
-	DCN31_VPG_MASK_SH_LIST(_MASK)
-};
-
-#define afmt_regs(id)\
-[id] = {\
-	AFMT_DCN31_REG_LIST(id)\
-}
-
-static const struct dcn31_afmt_registers afmt_regs[] = {
-	afmt_regs(0),
-	afmt_regs(1),
-	afmt_regs(2),
-	afmt_regs(3),
-	afmt_regs(4),
-	afmt_regs(5)
-};
-
-static const struct dcn31_afmt_shift afmt_shift = {
-	DCN31_AFMT_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dcn31_afmt_mask afmt_mask = {
-	DCN31_AFMT_MASK_SH_LIST(_MASK)
-};
-
-#define apg_regs(id)\
-[id] = {\
-	APG_DCN31_REG_LIST(id)\
-}
-
-static const struct dcn31_apg_registers apg_regs[] = {
-	apg_regs(0),
-	apg_regs(1),
-	apg_regs(2),
-	apg_regs(3)
-};
-
-static const struct dcn31_apg_shift apg_shift = {
-	DCN31_APG_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dcn31_apg_mask apg_mask = {
-		DCN31_APG_MASK_SH_LIST(_MASK)
-};
-
-#define stream_enc_regs(id)\
-[id] = {\
-	SE_DCN3_REG_LIST(id)\
-}
-
-static const struct dcn10_stream_enc_registers stream_enc_regs[] = {
-	stream_enc_regs(0),
-	stream_enc_regs(1),
-	stream_enc_regs(2),
-	stream_enc_regs(3),
-	stream_enc_regs(4)
-};
-
-static const struct dcn10_stream_encoder_shift se_shift = {
-		SE_COMMON_MASK_SH_LIST_DCN30(__SHIFT)
-};
-
-static const struct dcn10_stream_encoder_mask se_mask = {
-		SE_COMMON_MASK_SH_LIST_DCN30(_MASK)
-};
-
-
-#define aux_regs(id)\
-[id] = {\
-	DCN2_AUX_REG_LIST(id)\
-}
-
-static const struct dcn10_link_enc_aux_registers link_enc_aux_regs[] = {
-		aux_regs(0),
-		aux_regs(1),
-		aux_regs(2),
-		aux_regs(3),
-		aux_regs(4)
-};
-
-#define hpd_regs(id)\
-[id] = {\
-	HPD_REG_LIST(id)\
-}
-
-static const struct dcn10_link_enc_hpd_registers link_enc_hpd_regs[] = {
-		hpd_regs(0),
-		hpd_regs(1),
-		hpd_regs(2),
-		hpd_regs(3),
-		hpd_regs(4)
-};
-
-#define link_regs(id, phyid)\
-[id] = {\
-	LE_DCN31_REG_LIST(id), \
-	UNIPHY_DCN2_REG_LIST(phyid), \
-	DPCS_DCN31_REG_LIST(id), \
-}
-
-static const struct dce110_aux_registers_shift aux_shift = {
-	DCN_AUX_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dce110_aux_registers_mask aux_mask = {
-	DCN_AUX_MASK_SH_LIST(_MASK)
-};
-
-static const struct dcn10_link_enc_registers link_enc_regs[] = {
-	link_regs(0, A),
-	link_regs(1, B),
-	link_regs(2, C),
-	link_regs(3, D),
-	link_regs(4, E)
-};
-
-static const struct dcn10_link_enc_shift le_shift = {
-	LINK_ENCODER_MASK_SH_LIST_DCN31(__SHIFT), \
-	DPCS_DCN31_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dcn10_link_enc_mask le_mask = {
-	LINK_ENCODER_MASK_SH_LIST_DCN31(_MASK), \
-	DPCS_DCN31_MASK_SH_LIST(_MASK)
-};
-
-#define hpo_dp_stream_encoder_reg_list(id)\
-[id] = {\
-	DCN3_1_HPO_DP_STREAM_ENC_REG_LIST(id)\
-}
-
-static const struct dcn31_hpo_dp_stream_encoder_registers hpo_dp_stream_enc_regs[] = {
-	hpo_dp_stream_encoder_reg_list(0),
-	hpo_dp_stream_encoder_reg_list(1),
-	hpo_dp_stream_encoder_reg_list(2),
-	hpo_dp_stream_encoder_reg_list(3),
-};
-
-static const struct dcn31_hpo_dp_stream_encoder_shift hpo_dp_se_shift = {
-	DCN3_1_HPO_DP_STREAM_ENC_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dcn31_hpo_dp_stream_encoder_mask hpo_dp_se_mask = {
-	DCN3_1_HPO_DP_STREAM_ENC_MASK_SH_LIST(_MASK)
-};
-
-#define hpo_dp_link_encoder_reg_list(id)\
-[id] = {\
-	DCN3_1_HPO_DP_LINK_ENC_REG_LIST(id),\
-	DCN3_1_RDPCSTX_REG_LIST(0),\
-	DCN3_1_RDPCSTX_REG_LIST(1),\
-	DCN3_1_RDPCSTX_REG_LIST(2),\
-	DCN3_1_RDPCSTX_REG_LIST(3),\
-	DCN3_1_RDPCSTX_REG_LIST(4)\
-}
-
-static const struct dcn31_hpo_dp_link_encoder_registers hpo_dp_link_enc_regs[] = {
-	hpo_dp_link_encoder_reg_list(0),
-	hpo_dp_link_encoder_reg_list(1),
-};
-
-static const struct dcn31_hpo_dp_link_encoder_shift hpo_dp_le_shift = {
-	DCN3_1_HPO_DP_LINK_ENC_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dcn31_hpo_dp_link_encoder_mask hpo_dp_le_mask = {
-	DCN3_1_HPO_DP_LINK_ENC_MASK_SH_LIST(_MASK)
-};
-
-#define dpp_regs(id)\
-[id] = {\
-	DPP_REG_LIST_DCN30(id),\
-}
-
-static const struct dcn3_dpp_registers dpp_regs[] = {
-	dpp_regs(0),
-	dpp_regs(1),
-	dpp_regs(2),
-	dpp_regs(3)
-};
-
-static const struct dcn3_dpp_shift tf_shift = {
-		DPP_REG_LIST_SH_MASK_DCN30(__SHIFT)
-};
-
-static const struct dcn3_dpp_mask tf_mask = {
-		DPP_REG_LIST_SH_MASK_DCN30(_MASK)
-};
-
-#define opp_regs(id)\
-[id] = {\
-	OPP_REG_LIST_DCN30(id),\
-}
-
-static const struct dcn20_opp_registers opp_regs[] = {
-	opp_regs(0),
-	opp_regs(1),
-	opp_regs(2),
-	opp_regs(3)
-};
-
-static const struct dcn20_opp_shift opp_shift = {
-	OPP_MASK_SH_LIST_DCN20(__SHIFT)
-};
-
-static const struct dcn20_opp_mask opp_mask = {
-	OPP_MASK_SH_LIST_DCN20(_MASK)
-};
-
-#define aux_engine_regs(id)\
-[id] = {\
-	AUX_COMMON_REG_LIST0(id), \
-	.AUXN_IMPCAL = 0, \
-	.AUXP_IMPCAL = 0, \
-	.AUX_RESET_MASK = DP_AUX0_AUX_CONTROL__AUX_RESET_MASK, \
-}
-
-static const struct dce110_aux_registers aux_engine_regs[] = {
-		aux_engine_regs(0),
-		aux_engine_regs(1),
-		aux_engine_regs(2),
-		aux_engine_regs(3),
-		aux_engine_regs(4)
-};
-
-#define dwbc_regs_dcn3(id)\
-[id] = {\
-	DWBC_COMMON_REG_LIST_DCN30(id),\
-}
-
-static const struct dcn30_dwbc_registers dwbc30_regs[] = {
-	dwbc_regs_dcn3(0),
-};
-
-static const struct dcn30_dwbc_shift dwbc30_shift = {
-	DWBC_COMMON_MASK_SH_LIST_DCN30(__SHIFT)
-};
-
-static const struct dcn30_dwbc_mask dwbc30_mask = {
-	DWBC_COMMON_MASK_SH_LIST_DCN30(_MASK)
-};
-
-#define mcif_wb_regs_dcn3(id)\
-[id] = {\
-	MCIF_WB_COMMON_REG_LIST_DCN30(id),\
-}
-
-static const struct dcn30_mmhubbub_registers mcif_wb30_regs[] = {
-	mcif_wb_regs_dcn3(0)
-};
-
-static const struct dcn30_mmhubbub_shift mcif_wb30_shift = {
-	MCIF_WB_COMMON_MASK_SH_LIST_DCN30(__SHIFT)
-};
-
-static const struct dcn30_mmhubbub_mask mcif_wb30_mask = {
-	MCIF_WB_COMMON_MASK_SH_LIST_DCN30(_MASK)
-};
-
-#define dsc_regsDCN20(id)\
-[id] = {\
-	DSC_REG_LIST_DCN20(id)\
-}
-
-static const struct dcn20_dsc_registers dsc_regs[] = {
-	dsc_regsDCN20(0),
-	dsc_regsDCN20(1),
-	dsc_regsDCN20(2)
-};
-
-static const struct dcn20_dsc_shift dsc_shift = {
-	DSC_REG_LIST_SH_MASK_DCN20(__SHIFT)
-};
-
-static const struct dcn20_dsc_mask dsc_mask = {
-	DSC_REG_LIST_SH_MASK_DCN20(_MASK)
-};
-
-static const struct dcn30_mpc_registers mpc_regs = {
-		MPC_REG_LIST_DCN3_0(0),
-		MPC_REG_LIST_DCN3_0(1),
-		MPC_REG_LIST_DCN3_0(2),
-		MPC_REG_LIST_DCN3_0(3),
-		MPC_OUT_MUX_REG_LIST_DCN3_0(0),
-		MPC_OUT_MUX_REG_LIST_DCN3_0(1),
-		MPC_OUT_MUX_REG_LIST_DCN3_0(2),
-		MPC_OUT_MUX_REG_LIST_DCN3_0(3),
-		MPC_RMU_GLOBAL_REG_LIST_DCN3AG,
-		MPC_RMU_REG_LIST_DCN3AG(0),
-		MPC_RMU_REG_LIST_DCN3AG(1),
-		//MPC_RMU_REG_LIST_DCN3AG(2),
-		MPC_DWB_MUX_REG_LIST_DCN3_0(0),
-};
-
-static const struct dcn30_mpc_shift mpc_shift = {
-	MPC_COMMON_MASK_SH_LIST_DCN30(__SHIFT)
-};
-
-static const struct dcn30_mpc_mask mpc_mask = {
-	MPC_COMMON_MASK_SH_LIST_DCN30(_MASK)
-};
-
-#define optc_regs(id)\
-[id] = {OPTC_COMMON_REG_LIST_DCN3_1(id)}
-
-static const struct dcn_optc_registers optc_regs[] = {
-	optc_regs(0),
-	optc_regs(1),
-	optc_regs(2),
-	optc_regs(3)
-};
-
-static const struct dcn_optc_shift optc_shift = {
-	OPTC_COMMON_MASK_SH_LIST_DCN3_1(__SHIFT)
-};
-
-static const struct dcn_optc_mask optc_mask = {
-	OPTC_COMMON_MASK_SH_LIST_DCN3_1(_MASK)
-};
-
-#define hubp_regs(id)\
-[id] = {\
-	HUBP_REG_LIST_DCN30(id)\
-}
-
-static const struct dcn_hubp2_registers hubp_regs[] = {
-		hubp_regs(0),
-		hubp_regs(1),
-		hubp_regs(2),
-		hubp_regs(3)
-};
-
-
-static const struct dcn_hubp2_shift hubp_shift = {
-		HUBP_MASK_SH_LIST_DCN31(__SHIFT)
-};
-
-static const struct dcn_hubp2_mask hubp_mask = {
-		HUBP_MASK_SH_LIST_DCN31(_MASK)
-};
-static const struct dcn_hubbub_registers hubbub_reg = {
-		HUBBUB_REG_LIST_DCN31(0)
-};
-
-static const struct dcn_hubbub_shift hubbub_shift = {
-		HUBBUB_MASK_SH_LIST_DCN31(__SHIFT)
-};
-
-static const struct dcn_hubbub_mask hubbub_mask = {
-		HUBBUB_MASK_SH_LIST_DCN31(_MASK)
-};
-
-static const struct dccg_registers dccg_regs = {
-		DCCG_REG_LIST_DCN31()
-};
-
-static const struct dccg_shift dccg_shift = {
-		DCCG_MASK_SH_LIST_DCN31(__SHIFT)
-};
-
-static const struct dccg_mask dccg_mask = {
-		DCCG_MASK_SH_LIST_DCN31(_MASK)
-};
-
-
-#define SRII2(reg_name_pre, reg_name_post, id)\
-	.reg_name_pre ## _ ##  reg_name_post[id] = BASE(reg ## reg_name_pre \
-			## id ## _ ## reg_name_post ## _BASE_IDX) + \
-			reg ## reg_name_pre ## id ## _ ## reg_name_post
-
-
-#define HWSEQ_DCN31_REG_LIST()\
-	SR(DCHUBBUB_GLOBAL_TIMER_CNTL), \
-	SR(DCHUBBUB_ARB_HOSTVM_CNTL), \
-	SR(DIO_MEM_PWR_CTRL), \
-	SR(ODM_MEM_PWR_CTRL3), \
-	SR(DMU_MEM_PWR_CNTL), \
-	SR(MMHUBBUB_MEM_PWR_CNTL), \
-	SR(DCCG_GATE_DISABLE_CNTL), \
-	SR(DCCG_GATE_DISABLE_CNTL2), \
-	SR(DCFCLK_CNTL),\
-	SR(DC_MEM_GLOBAL_PWR_REQ_CNTL), \
-	SRII(PIXEL_RATE_CNTL, OTG, 0), \
-	SRII(PIXEL_RATE_CNTL, OTG, 1),\
-	SRII(PIXEL_RATE_CNTL, OTG, 2),\
-	SRII(PIXEL_RATE_CNTL, OTG, 3),\
-	SRII(PHYPLL_PIXEL_RATE_CNTL, OTG, 0),\
-	SRII(PHYPLL_PIXEL_RATE_CNTL, OTG, 1),\
-	SRII(PHYPLL_PIXEL_RATE_CNTL, OTG, 2),\
-	SRII(PHYPLL_PIXEL_RATE_CNTL, OTG, 3),\
-	SR(MICROSECOND_TIME_BASE_DIV), \
-	SR(MILLISECOND_TIME_BASE_DIV), \
-	SR(DISPCLK_FREQ_CHANGE_CNTL), \
-	SR(RBBMIF_TIMEOUT_DIS), \
-	SR(RBBMIF_TIMEOUT_DIS_2), \
-	SR(DCHUBBUB_CRC_CTRL), \
-	SR(DPP_TOP0_DPP_CRC_CTRL), \
-	SR(DPP_TOP0_DPP_CRC_VAL_B_A), \
-	SR(DPP_TOP0_DPP_CRC_VAL_R_G), \
-	SR(MPC_CRC_CTRL), \
-	SR(MPC_CRC_RESULT_GB), \
-	SR(MPC_CRC_RESULT_C), \
-	SR(MPC_CRC_RESULT_AR), \
-	SR(DOMAIN0_PG_CONFIG), \
-	SR(DOMAIN1_PG_CONFIG), \
-	SR(DOMAIN2_PG_CONFIG), \
-	SR(DOMAIN3_PG_CONFIG), \
-	SR(DOMAIN16_PG_CONFIG), \
-	SR(DOMAIN17_PG_CONFIG), \
-	SR(DOMAIN18_PG_CONFIG), \
-	SR(DOMAIN0_PG_STATUS), \
-	SR(DOMAIN1_PG_STATUS), \
-	SR(DOMAIN2_PG_STATUS), \
-	SR(DOMAIN3_PG_STATUS), \
-	SR(DOMAIN16_PG_STATUS), \
-	SR(DOMAIN17_PG_STATUS), \
-	SR(DOMAIN18_PG_STATUS), \
-	SR(D1VGA_CONTROL), \
-	SR(D2VGA_CONTROL), \
-	SR(D3VGA_CONTROL), \
-	SR(D4VGA_CONTROL), \
-	SR(D5VGA_CONTROL), \
-	SR(D6VGA_CONTROL), \
-	SR(DC_IP_REQUEST_CNTL), \
-	SR(AZALIA_AUDIO_DTO), \
-	SR(AZALIA_CONTROLLER_CLOCK_GATING), \
-	SR(HPO_TOP_HW_CONTROL)
-
-static const struct dce_hwseq_registers hwseq_reg = {
-		HWSEQ_DCN31_REG_LIST()
-};
-
-#define HWSEQ_DCN31_MASK_SH_LIST(mask_sh)\
-	HWSEQ_DCN_MASK_SH_LIST(mask_sh), \
-	HWS_SF(, DCHUBBUB_GLOBAL_TIMER_CNTL, DCHUBBUB_GLOBAL_TIMER_REFDIV, mask_sh), \
-	HWS_SF(, DCHUBBUB_ARB_HOSTVM_CNTL, DISABLE_HOSTVM_FORCE_ALLOW_PSTATE, mask_sh), \
-	HWS_SF(, DOMAIN0_PG_CONFIG, DOMAIN_POWER_FORCEON, mask_sh), \
-	HWS_SF(, DOMAIN0_PG_CONFIG, DOMAIN_POWER_GATE, mask_sh), \
-	HWS_SF(, DOMAIN1_PG_CONFIG, DOMAIN_POWER_FORCEON, mask_sh), \
-	HWS_SF(, DOMAIN1_PG_CONFIG, DOMAIN_POWER_GATE, mask_sh), \
-	HWS_SF(, DOMAIN2_PG_CONFIG, DOMAIN_POWER_FORCEON, mask_sh), \
-	HWS_SF(, DOMAIN2_PG_CONFIG, DOMAIN_POWER_GATE, mask_sh), \
-	HWS_SF(, DOMAIN3_PG_CONFIG, DOMAIN_POWER_FORCEON, mask_sh), \
-	HWS_SF(, DOMAIN3_PG_CONFIG, DOMAIN_POWER_GATE, mask_sh), \
-	HWS_SF(, DOMAIN16_PG_CONFIG, DOMAIN_POWER_FORCEON, mask_sh), \
-	HWS_SF(, DOMAIN16_PG_CONFIG, DOMAIN_POWER_GATE, mask_sh), \
-	HWS_SF(, DOMAIN17_PG_CONFIG, DOMAIN_POWER_FORCEON, mask_sh), \
-	HWS_SF(, DOMAIN17_PG_CONFIG, DOMAIN_POWER_GATE, mask_sh), \
-	HWS_SF(, DOMAIN18_PG_CONFIG, DOMAIN_POWER_FORCEON, mask_sh), \
-	HWS_SF(, DOMAIN18_PG_CONFIG, DOMAIN_POWER_GATE, mask_sh), \
-	HWS_SF(, DOMAIN0_PG_STATUS, DOMAIN_PGFSM_PWR_STATUS, mask_sh), \
-	HWS_SF(, DOMAIN1_PG_STATUS, DOMAIN_PGFSM_PWR_STATUS, mask_sh), \
-	HWS_SF(, DOMAIN2_PG_STATUS, DOMAIN_PGFSM_PWR_STATUS, mask_sh), \
-	HWS_SF(, DOMAIN3_PG_STATUS, DOMAIN_PGFSM_PWR_STATUS, mask_sh), \
-	HWS_SF(, DOMAIN16_PG_STATUS, DOMAIN_PGFSM_PWR_STATUS, mask_sh), \
-	HWS_SF(, DOMAIN17_PG_STATUS, DOMAIN_PGFSM_PWR_STATUS, mask_sh), \
-	HWS_SF(, DOMAIN18_PG_STATUS, DOMAIN_PGFSM_PWR_STATUS, mask_sh), \
-	HWS_SF(, DC_IP_REQUEST_CNTL, IP_REQUEST_EN, mask_sh), \
-	HWS_SF(, AZALIA_AUDIO_DTO, AZALIA_AUDIO_DTO_MODULE, mask_sh), \
-	HWS_SF(, HPO_TOP_CLOCK_CONTROL, HPO_HDMISTREAMCLK_G_GATE_DIS, mask_sh), \
-	HWS_SF(, DMU_MEM_PWR_CNTL, DMCU_ERAM_MEM_PWR_FORCE, mask_sh), \
-	HWS_SF(, ODM_MEM_PWR_CTRL3, ODM_MEM_UNASSIGNED_PWR_MODE, mask_sh), \
-	HWS_SF(, ODM_MEM_PWR_CTRL3, ODM_MEM_VBLANK_PWR_MODE, mask_sh), \
-	HWS_SF(, MMHUBBUB_MEM_PWR_CNTL, VGA_MEM_PWR_FORCE, mask_sh), \
-	HWS_SF(, DIO_MEM_PWR_CTRL, I2C_LIGHT_SLEEP_FORCE, mask_sh), \
-	HWS_SF(, HPO_TOP_HW_CONTROL, HPO_IO_EN, mask_sh)
-
-static const struct dce_hwseq_shift hwseq_shift = {
-		HWSEQ_DCN31_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dce_hwseq_mask hwseq_mask = {
-		HWSEQ_DCN31_MASK_SH_LIST(_MASK)
-};
-#define vmid_regs(id)\
-[id] = {\
-		DCN20_VMID_REG_LIST(id)\
-}
-
-static const struct dcn_vmid_registers vmid_regs[] = {
-	vmid_regs(0),
-	vmid_regs(1),
-	vmid_regs(2),
-	vmid_regs(3),
-	vmid_regs(4),
-	vmid_regs(5),
-	vmid_regs(6),
-	vmid_regs(7),
-	vmid_regs(8),
-	vmid_regs(9),
-	vmid_regs(10),
-	vmid_regs(11),
-	vmid_regs(12),
-	vmid_regs(13),
-	vmid_regs(14),
-	vmid_regs(15)
-};
-
-static const struct dcn20_vmid_shift vmid_shifts = {
-		DCN20_VMID_MASK_SH_LIST(__SHIFT)
-};
-
-static const struct dcn20_vmid_mask vmid_masks = {
-		DCN20_VMID_MASK_SH_LIST(_MASK)
-};
-
-static const struct resource_caps res_cap_dcn31 = {
-	.num_timing_generator = 4,
-	.num_opp = 4,
-	.num_video_plane = 4,
-	.num_audio = 5,
-	.num_stream_encoder = 5,
-	.num_dig_link_enc = 5,
-	.num_hpo_dp_stream_encoder = 4,
-	.num_hpo_dp_link_encoder = 2,
-	.num_pll = 5,
-	.num_dwb = 1,
-	.num_ddc = 5,
-	.num_vmid = 16,
-	.num_mpc_3dlut = 2,
-	.num_dsc = 3,
-};
-
-static const struct dc_plane_cap plane_cap = {
-	.type = DC_PLANE_TYPE_DCN_UNIVERSAL,
-	.blends_with_above = true,
-	.blends_with_below = true,
-	.per_pixel_alpha = true,
-
-	.pixel_format_support = {
-			.argb8888 = true,
-			.nv12 = true,
-			.fp16 = true,
-			.p010 = false,
-			.ayuv = false,
-	},
-
-	.max_upscale_factor = {
-			.argb8888 = 16000,
-			.nv12 = 16000,
-			.fp16 = 16000
-	},
-
-	// 6:1 downscaling ratio: 1000/6 = 166.666
-	.max_downscale_factor = {
-			.argb8888 = 167,
-			.nv12 = 167,
-			.fp16 = 167
-	},
-	64,
-	64
-};
-
-static const struct dc_debug_options debug_defaults_drv = {
-	.disable_dmcu = true,
-	.force_abm_enable = false,
-	.timing_trace = false,
-	.clock_trace = true,
-	.disable_pplib_clock_request = false,
-	.pipe_split_policy = MPC_SPLIT_AVOID,
-	.force_single_disp_pipe_split = false,
-	.disable_dcc = DCC_ENABLE,
-	.vsr_support = true,
-	.performance_trace = false,
-	.max_downscale_src_width = 4096,/*upto true 4K*/
-	.disable_pplib_wm_range = false,
-	.scl_reset_length10 = true,
-	.sanity_checks = false,
-	.underflow_assert_delay_us = 0xFFFFFFFF,
-	.dwb_fi_phase = -1, // -1 = disable,
-	.dmub_command_table = true,
-	.pstate_enabled = true,
-	.use_max_lb = true,
-	.enable_mem_low_power = {
-		.bits = {
-			.vga = true,
-			.i2c = true,
-			.dmcu = false, // This is previously known to cause hang on S3 cycles if enabled
-			.dscl = true,
-			.cm = true,
-			.mpc = true,
-			.optc = true,
-			.vpg = true,
-			.afmt = true,
-		}
-	},
-	.optimize_edp_link_rate = true,
-	.enable_sw_cntl_psr = true,
-};
-
-static const struct dc_debug_options debug_defaults_diags = {
-	.disable_dmcu = true,
-	.force_abm_enable = false,
-	.timing_trace = true,
-	.clock_trace = true,
-	.disable_dpp_power_gate = true,
-	.disable_hubp_power_gate = true,
-	.disable_clock_gate = true,
-	.disable_pplib_clock_request = true,
-	.disable_pplib_wm_range = true,
-	.disable_stutter = false,
-	.scl_reset_length10 = true,
-	.dwb_fi_phase = -1, // -1 = disable
-	.dmub_command_table = true,
-	.enable_tri_buf = true,
-	.use_max_lb = true
-};
-
-static void dcn31_dpp_destroy(struct dpp **dpp)
-{
-	kfree(TO_DCN20_DPP(*dpp));
-	*dpp = NULL;
-}
-
-static struct dpp *dcn31_dpp_create(
-	struct dc_context *ctx,
-	uint32_t inst)
-{
-	struct dcn3_dpp *dpp =
-		kzalloc(sizeof(struct dcn3_dpp), GFP_KERNEL);
-
-	if (!dpp)
-		return NULL;
-
-	if (dpp3_construct(dpp, ctx, inst,
-			&dpp_regs[inst], &tf_shift, &tf_mask))
-		return &dpp->base;
-
-	BREAK_TO_DEBUGGER();
-	kfree(dpp);
-	return NULL;
-}
-
-static struct output_pixel_processor *dcn31_opp_create(
-	struct dc_context *ctx, uint32_t inst)
-{
-	struct dcn20_opp *opp =
-		kzalloc(sizeof(struct dcn20_opp), GFP_KERNEL);
-
-	if (!opp) {
-		BREAK_TO_DEBUGGER();
-		return NULL;
-	}
-
-	dcn20_opp_construct(opp, ctx, inst,
-			&opp_regs[inst], &opp_shift, &opp_mask);
-	return &opp->base;
-}
-
-static struct dce_aux *dcn31_aux_engine_create(
-	struct dc_context *ctx,
-	uint32_t inst)
-{
-	struct aux_engine_dce110 *aux_engine =
-		kzalloc(sizeof(struct aux_engine_dce110), GFP_KERNEL);
-
-	if (!aux_engine)
-		return NULL;
-
-	dce110_aux_engine_construct(aux_engine, ctx, inst,
-				    SW_AUX_TIMEOUT_PERIOD_MULTIPLIER * AUX_TIMEOUT_PERIOD,
-				    &aux_engine_regs[inst],
-					&aux_mask,
-					&aux_shift,
-					ctx->dc->caps.extended_aux_timeout_support);
-
-	return &aux_engine->base;
-}
-#define i2c_inst_regs(id) { I2C_HW_ENGINE_COMMON_REG_LIST_DCN30(id) }
-
-static const struct dce_i2c_registers i2c_hw_regs[] = {
-		i2c_inst_regs(1),
-		i2c_inst_regs(2),
-		i2c_inst_regs(3),
-		i2c_inst_regs(4),
-		i2c_inst_regs(5),
-};
-
-static const struct dce_i2c_shift i2c_shifts = {
-		I2C_COMMON_MASK_SH_LIST_DCN30(__SHIFT)
-};
-
-static const struct dce_i2c_mask i2c_masks = {
-		I2C_COMMON_MASK_SH_LIST_DCN30(_MASK)
-};
-
-static struct dce_i2c_hw *dcn31_i2c_hw_create(
-	struct dc_context *ctx,
-	uint32_t inst)
-{
-	struct dce_i2c_hw *dce_i2c_hw =
-		kzalloc(sizeof(struct dce_i2c_hw), GFP_KERNEL);
-
-	if (!dce_i2c_hw)
-		return NULL;
-
-	dcn2_i2c_hw_construct(dce_i2c_hw, ctx, inst,
-				    &i2c_hw_regs[inst], &i2c_shifts, &i2c_masks);
-
-	return dce_i2c_hw;
-}
-static struct mpc *dcn31_mpc_create(
-		struct dc_context *ctx,
-		int num_mpcc,
-		int num_rmu)
-{
-	struct dcn30_mpc *mpc30 = kzalloc(sizeof(struct dcn30_mpc),
-					  GFP_KERNEL);
-
-	if (!mpc30)
-		return NULL;
-
-	dcn30_mpc_construct(mpc30, ctx,
-			&mpc_regs,
-			&mpc_shift,
-			&mpc_mask,
-			num_mpcc,
-			num_rmu);
-
-	return &mpc30->base;
-}
-
-static struct hubbub *dcn31_hubbub_create(struct dc_context *ctx)
-{
-	int i;
-
-	struct dcn20_hubbub *hubbub3 = kzalloc(sizeof(struct dcn20_hubbub),
-					  GFP_KERNEL);
-
-	if (!hubbub3)
-		return NULL;
-
-	hubbub31_construct(hubbub3, ctx,
-			&hubbub_reg,
-			&hubbub_shift,
-			&hubbub_mask,
-			dcn3_1_ip.det_buffer_size_kbytes,
-			dcn3_1_ip.pixel_chunk_size_kbytes,
-			dcn3_1_ip.config_return_buffer_size_in_kbytes);
-
-
-	for (i = 0; i < res_cap_dcn31.num_vmid; i++) {
-		struct dcn20_vmid *vmid = &hubbub3->vmid[i];
-
-		vmid->ctx = ctx;
-
-		vmid->regs = &vmid_regs[i];
-		vmid->shifts = &vmid_shifts;
-		vmid->masks = &vmid_masks;
-	}
-
-	return &hubbub3->base;
-}
-
-static struct timing_generator *dcn31_timing_generator_create(
-		struct dc_context *ctx,
-		uint32_t instance)
-{
-	struct optc *tgn10 =
-		kzalloc(sizeof(struct optc), GFP_KERNEL);
-
-	if (!tgn10)
-		return NULL;
-
-	tgn10->base.inst = instance;
-	tgn10->base.ctx = ctx;
-
-	tgn10->tg_regs = &optc_regs[instance];
-	tgn10->tg_shift = &optc_shift;
-	tgn10->tg_mask = &optc_mask;
-
-	dcn31_timing_generator_init(tgn10);
-
-	return &tgn10->base;
-}
-
-static const struct encoder_feature_support link_enc_feature = {
-		.max_hdmi_deep_color = COLOR_DEPTH_121212,
-		.max_hdmi_pixel_clock = 600000,
-		.hdmi_ycbcr420_supported = true,
-		.dp_ycbcr420_supported = true,
-		.fec_supported = true,
-		.flags.bits.IS_HBR2_CAPABLE = true,
-		.flags.bits.IS_HBR3_CAPABLE = true,
-		.flags.bits.IS_TPS3_CAPABLE = true,
-		.flags.bits.IS_TPS4_CAPABLE = true
-};
-
-static struct link_encoder *dcn31_link_encoder_create(
-	const struct encoder_init_data *enc_init_data)
-{
-	struct dcn20_link_encoder *enc20 =
-		kzalloc(sizeof(struct dcn20_link_encoder), GFP_KERNEL);
-
-	if (!enc20)
-		return NULL;
-
-	dcn31_link_encoder_construct(enc20,
-			enc_init_data,
-			&link_enc_feature,
-			&link_enc_regs[enc_init_data->transmitter],
-			&link_enc_aux_regs[enc_init_data->channel - 1],
-			&link_enc_hpd_regs[enc_init_data->hpd_source],
-			&le_shift,
-			&le_mask);
-
-	return &enc20->enc10.base;
-}
-
-/* Create a minimal link encoder object not associated with a particular
- * physical connector.
- * resource_funcs.link_enc_create_minimal
- */
-static struct link_encoder *dcn31_link_enc_create_minimal(
-		struct dc_context *ctx, enum engine_id eng_id)
-{
-	struct dcn20_link_encoder *enc20;
-
-	if ((eng_id - ENGINE_ID_DIGA) > ctx->dc->res_pool->res_cap->num_dig_link_enc)
-		return NULL;
-
-	enc20 = kzalloc(sizeof(struct dcn20_link_encoder), GFP_KERNEL);
-	if (!enc20)
-		return NULL;
-
-	dcn31_link_encoder_construct_minimal(
-			enc20,
-			ctx,
-			&link_enc_feature,
-			&link_enc_regs[eng_id - ENGINE_ID_DIGA],
-			eng_id);
-
-	return &enc20->enc10.base;
-}
-
-struct panel_cntl *dcn31_panel_cntl_create(const struct panel_cntl_init_data *init_data)
-{
-	struct dcn31_panel_cntl *panel_cntl =
-		kzalloc(sizeof(struct dcn31_panel_cntl), GFP_KERNEL);
-
-	if (!panel_cntl)
-		return NULL;
-
-	dcn31_panel_cntl_construct(panel_cntl, init_data);
-
-	return &panel_cntl->base;
-}
-
-static void read_dce_straps(
-	struct dc_context *ctx,
-	struct resource_straps *straps)
-{
-	generic_reg_get(ctx, regDC_PINSTRAPS + BASE(regDC_PINSTRAPS_BASE_IDX),
-		FN(DC_PINSTRAPS, DC_PINSTRAPS_AUDIO), &straps->dc_pinstraps_audio);
-
-}
-
-static struct audio *dcn31_create_audio(
-		struct dc_context *ctx, unsigned int inst)
-{
-	return dce_audio_create(ctx, inst,
-			&audio_regs[inst], &audio_shift, &audio_mask);
-}
-
-static struct vpg *dcn31_vpg_create(
-	struct dc_context *ctx,
-	uint32_t inst)
-{
-	struct dcn31_vpg *vpg31 = kzalloc(sizeof(struct dcn31_vpg), GFP_KERNEL);
-
-	if (!vpg31)
-		return NULL;
-
-	vpg31_construct(vpg31, ctx, inst,
-			&vpg_regs[inst],
-			&vpg_shift,
-			&vpg_mask);
-
-	return &vpg31->base;
-}
-
-static struct afmt *dcn31_afmt_create(
-	struct dc_context *ctx,
-	uint32_t inst)
-{
-	struct dcn31_afmt *afmt31 = kzalloc(sizeof(struct dcn31_afmt), GFP_KERNEL);
-
-	if (!afmt31)
-		return NULL;
-
-	afmt31_construct(afmt31, ctx, inst,
-			&afmt_regs[inst],
-			&afmt_shift,
-			&afmt_mask);
-
-	// Light sleep by default, no need to power down here
-
-	return &afmt31->base;
-}
-
-static struct apg *dcn31_apg_create(
-	struct dc_context *ctx,
-	uint32_t inst)
-{
-	struct dcn31_apg *apg31 = kzalloc(sizeof(struct dcn31_apg), GFP_KERNEL);
-
-	if (!apg31)
-		return NULL;
-
-	apg31_construct(apg31, ctx, inst,
-			&apg_regs[inst],
-			&apg_shift,
-			&apg_mask);
-
-	return &apg31->base;
-}
-
-static struct stream_encoder *dcn31_stream_encoder_create(
-	enum engine_id eng_id,
-	struct dc_context *ctx)
-{
-	struct dcn10_stream_encoder *enc1;
-	struct vpg *vpg;
-	struct afmt *afmt;
-	int vpg_inst;
-	int afmt_inst;
-
-	/* Mapping of VPG, AFMT, DME register blocks to DIO block instance */
-	if (eng_id <= ENGINE_ID_DIGF) {
-		vpg_inst = eng_id;
-		afmt_inst = eng_id;
-	} else
-		return NULL;
-
-	enc1 = kzalloc(sizeof(struct dcn10_stream_encoder), GFP_KERNEL);
-	vpg = dcn31_vpg_create(ctx, vpg_inst);
-	afmt = dcn31_afmt_create(ctx, afmt_inst);
-
-	if (!enc1 || !vpg || !afmt) {
-		kfree(enc1);
-		kfree(vpg);
-		kfree(afmt);
-		return NULL;
-	}
-
-	if (ctx->asic_id.chip_family == FAMILY_YELLOW_CARP &&
-			ctx->asic_id.hw_internal_rev == YELLOW_CARP_B0) {
-		if ((eng_id == ENGINE_ID_DIGC) || (eng_id == ENGINE_ID_DIGD))
-			eng_id = eng_id + 3; // For B0 only. C->F, D->G.
-	}
-
-	dcn30_dio_stream_encoder_construct(enc1, ctx, ctx->dc_bios,
-					eng_id, vpg, afmt,
-					&stream_enc_regs[eng_id],
-					&se_shift, &se_mask);
-
-	return &enc1->base;
-}
-
-static struct hpo_dp_stream_encoder *dcn31_hpo_dp_stream_encoder_create(
-	enum engine_id eng_id,
-	struct dc_context *ctx)
-{
-	struct dcn31_hpo_dp_stream_encoder *hpo_dp_enc31;
-	struct vpg *vpg;
-	struct apg *apg;
-	uint32_t hpo_dp_inst;
-	uint32_t vpg_inst;
-	uint32_t apg_inst;
-
-	ASSERT((eng_id >= ENGINE_ID_HPO_DP_0) && (eng_id <= ENGINE_ID_HPO_DP_3));
-	hpo_dp_inst = eng_id - ENGINE_ID_HPO_DP_0;
-
-	/* Mapping of VPG register blocks to HPO DP block instance:
-	 * VPG[6] -> HPO_DP[0]
-	 * VPG[7] -> HPO_DP[1]
-	 * VPG[8] -> HPO_DP[2]
-	 * VPG[9] -> HPO_DP[3]
-	 */
-	vpg_inst = hpo_dp_inst + 6;
-
-	/* Mapping of APG register blocks to HPO DP block instance:
-	 * APG[0] -> HPO_DP[0]
-	 * APG[1] -> HPO_DP[1]
-	 * APG[2] -> HPO_DP[2]
-	 * APG[3] -> HPO_DP[3]
-	 */
-	apg_inst = hpo_dp_inst;
-
-	/* allocate HPO stream encoder and create VPG sub-block */
-	hpo_dp_enc31 = kzalloc(sizeof(struct dcn31_hpo_dp_stream_encoder), GFP_KERNEL);
-	vpg = dcn31_vpg_create(ctx, vpg_inst);
-	apg = dcn31_apg_create(ctx, apg_inst);
-
-	if (!hpo_dp_enc31 || !vpg || !apg) {
-		kfree(hpo_dp_enc31);
-		kfree(vpg);
-		kfree(apg);
-		return NULL;
-	}
-
-	dcn31_hpo_dp_stream_encoder_construct(hpo_dp_enc31, ctx, ctx->dc_bios,
-					hpo_dp_inst, eng_id, vpg, apg,
-					&hpo_dp_stream_enc_regs[hpo_dp_inst],
-					&hpo_dp_se_shift, &hpo_dp_se_mask);
-
-	return &hpo_dp_enc31->base;
-}
-
-static struct hpo_dp_link_encoder *dcn31_hpo_dp_link_encoder_create(
-	uint8_t inst,
-	struct dc_context *ctx)
-{
-	struct dcn31_hpo_dp_link_encoder *hpo_dp_enc31;
-
-	/* allocate HPO link encoder */
-	hpo_dp_enc31 = kzalloc(sizeof(struct dcn31_hpo_dp_link_encoder), GFP_KERNEL);
-
-	hpo_dp_link_encoder31_construct(hpo_dp_enc31, ctx, inst,
-					&hpo_dp_link_enc_regs[inst],
-					&hpo_dp_le_shift, &hpo_dp_le_mask);
-
-	return &hpo_dp_enc31->base;
-}
-
-static struct dce_hwseq *dcn31_hwseq_create(
-	struct dc_context *ctx)
-{
-	struct dce_hwseq *hws = kzalloc(sizeof(struct dce_hwseq), GFP_KERNEL);
-
-	if (hws) {
-		hws->ctx = ctx;
-		hws->regs = &hwseq_reg;
-		hws->shifts = &hwseq_shift;
-		hws->masks = &hwseq_mask;
-		/* DCN3.1 FPGA Workaround
-		 * Need to enable HPO DP Stream Encoder before setting OTG master enable.
-		 * To do so, move calling function enable_stream_timing to only be done AFTER calling
-		 * function core_link_enable_stream
-		 */
-		if (IS_FPGA_MAXIMUS_DC(ctx->dce_environment))
-			hws->wa.dp_hpo_and_otg_sequence = true;
-	}
-	return hws;
-}
-static const struct resource_create_funcs res_create_funcs = {
-	.read_dce_straps = read_dce_straps,
-	.create_audio = dcn31_create_audio,
-	.create_stream_encoder = dcn31_stream_encoder_create,
-	.create_hpo_dp_stream_encoder = dcn31_hpo_dp_stream_encoder_create,
-	.create_hpo_dp_link_encoder = dcn31_hpo_dp_link_encoder_create,
-	.create_hwseq = dcn31_hwseq_create,
-};
-
-static const struct resource_create_funcs res_create_maximus_funcs = {
-	.read_dce_straps = NULL,
-	.create_audio = NULL,
-	.create_stream_encoder = NULL,
-	.create_hpo_dp_stream_encoder = dcn31_hpo_dp_stream_encoder_create,
-	.create_hpo_dp_link_encoder = dcn31_hpo_dp_link_encoder_create,
-	.create_hwseq = dcn31_hwseq_create,
-};
-
-static void dcn31_resource_destruct(struct dcn31_resource_pool *pool)
-{
-	unsigned int i;
-
-	for (i = 0; i < pool->base.stream_enc_count; i++) {
-		if (pool->base.stream_enc[i] != NULL) {
-			if (pool->base.stream_enc[i]->vpg != NULL) {
-				kfree(DCN30_VPG_FROM_VPG(pool->base.stream_enc[i]->vpg));
-				pool->base.stream_enc[i]->vpg = NULL;
-			}
-			if (pool->base.stream_enc[i]->afmt != NULL) {
-				kfree(DCN30_AFMT_FROM_AFMT(pool->base.stream_enc[i]->afmt));
-				pool->base.stream_enc[i]->afmt = NULL;
-			}
-			kfree(DCN10STRENC_FROM_STRENC(pool->base.stream_enc[i]));
-			pool->base.stream_enc[i] = NULL;
-		}
-	}
-
-	for (i = 0; i < pool->base.hpo_dp_stream_enc_count; i++) {
-		if (pool->base.hpo_dp_stream_enc[i] != NULL) {
-			if (pool->base.hpo_dp_stream_enc[i]->vpg != NULL) {
-				kfree(DCN30_VPG_FROM_VPG(pool->base.hpo_dp_stream_enc[i]->vpg));
-				pool->base.hpo_dp_stream_enc[i]->vpg = NULL;
-			}
-			if (pool->base.hpo_dp_stream_enc[i]->apg != NULL) {
-				kfree(DCN31_APG_FROM_APG(pool->base.hpo_dp_stream_enc[i]->apg));
-				pool->base.hpo_dp_stream_enc[i]->apg = NULL;
-			}
-			kfree(DCN3_1_HPO_DP_STREAM_ENC_FROM_HPO_STREAM_ENC(pool->base.hpo_dp_stream_enc[i]));
-			pool->base.hpo_dp_stream_enc[i] = NULL;
-		}
-	}
-
-	for (i = 0; i < pool->base.hpo_dp_link_enc_count; i++) {
-		if (pool->base.hpo_dp_link_enc[i] != NULL) {
-			kfree(DCN3_1_HPO_DP_LINK_ENC_FROM_HPO_LINK_ENC(pool->base.hpo_dp_link_enc[i]));
-			pool->base.hpo_dp_link_enc[i] = NULL;
-		}
-	}
-
-	for (i = 0; i < pool->base.res_cap->num_dsc; i++) {
-		if (pool->base.dscs[i] != NULL)
-			dcn20_dsc_destroy(&pool->base.dscs[i]);
-	}
-
-	if (pool->base.mpc != NULL) {
-		kfree(TO_DCN20_MPC(pool->base.mpc));
-		pool->base.mpc = NULL;
-	}
-	if (pool->base.hubbub != NULL) {
-		kfree(pool->base.hubbub);
-		pool->base.hubbub = NULL;
-	}
-	for (i = 0; i < pool->base.pipe_count; i++) {
-		if (pool->base.dpps[i] != NULL)
-			dcn31_dpp_destroy(&pool->base.dpps[i]);
-
-		if (pool->base.ipps[i] != NULL)
-			pool->base.ipps[i]->funcs->ipp_destroy(&pool->base.ipps[i]);
-
-		if (pool->base.hubps[i] != NULL) {
-			kfree(TO_DCN20_HUBP(pool->base.hubps[i]));
-			pool->base.hubps[i] = NULL;
-		}
-
-		if (pool->base.irqs != NULL) {
-			dal_irq_service_destroy(&pool->base.irqs);
-		}
-	}
-
-	for (i = 0; i < pool->base.res_cap->num_ddc; i++) {
-		if (pool->base.engines[i] != NULL)
-			dce110_engine_destroy(&pool->base.engines[i]);
-		if (pool->base.hw_i2cs[i] != NULL) {
-			kfree(pool->base.hw_i2cs[i]);
-			pool->base.hw_i2cs[i] = NULL;
-		}
-		if (pool->base.sw_i2cs[i] != NULL) {
-			kfree(pool->base.sw_i2cs[i]);
-			pool->base.sw_i2cs[i] = NULL;
-		}
-	}
-
-	for (i = 0; i < pool->base.res_cap->num_opp; i++) {
-		if (pool->base.opps[i] != NULL)
-			pool->base.opps[i]->funcs->opp_destroy(&pool->base.opps[i]);
-	}
-
-	for (i = 0; i < pool->base.res_cap->num_timing_generator; i++) {
-		if (pool->base.timing_generators[i] != NULL)	{
-			kfree(DCN10TG_FROM_TG(pool->base.timing_generators[i]));
-			pool->base.timing_generators[i] = NULL;
-		}
-	}
-
-	for (i = 0; i < pool->base.res_cap->num_dwb; i++) {
-		if (pool->base.dwbc[i] != NULL) {
-			kfree(TO_DCN30_DWBC(pool->base.dwbc[i]));
-			pool->base.dwbc[i] = NULL;
-		}
-		if (pool->base.mcif_wb[i] != NULL) {
-			kfree(TO_DCN30_MMHUBBUB(pool->base.mcif_wb[i]));
-			pool->base.mcif_wb[i] = NULL;
-		}
-	}
-
-	for (i = 0; i < pool->base.audio_count; i++) {
-		if (pool->base.audios[i])
-			dce_aud_destroy(&pool->base.audios[i]);
-	}
-
-	for (i = 0; i < pool->base.clk_src_count; i++) {
-		if (pool->base.clock_sources[i] != NULL) {
-			dcn20_clock_source_destroy(&pool->base.clock_sources[i]);
-			pool->base.clock_sources[i] = NULL;
-		}
-	}
-
-	for (i = 0; i < pool->base.res_cap->num_mpc_3dlut; i++) {
-		if (pool->base.mpc_lut[i] != NULL) {
-			dc_3dlut_func_release(pool->base.mpc_lut[i]);
-			pool->base.mpc_lut[i] = NULL;
-		}
-		if (pool->base.mpc_shaper[i] != NULL) {
-			dc_transfer_func_release(pool->base.mpc_shaper[i]);
-			pool->base.mpc_shaper[i] = NULL;
-		}
-	}
-
-	if (pool->base.dp_clock_source != NULL) {
-		dcn20_clock_source_destroy(&pool->base.dp_clock_source);
-		pool->base.dp_clock_source = NULL;
-	}
-
-	for (i = 0; i < pool->base.res_cap->num_timing_generator; i++) {
-		if (pool->base.multiple_abms[i] != NULL)
-			dce_abm_destroy(&pool->base.multiple_abms[i]);
-	}
-
-	if (pool->base.psr != NULL)
-		dmub_psr_destroy(&pool->base.psr);
-
-	if (pool->base.dccg != NULL)
-		dcn_dccg_destroy(&pool->base.dccg);
-}
-
-static struct hubp *dcn31_hubp_create(
-	struct dc_context *ctx,
-	uint32_t inst)
-{
-	struct dcn20_hubp *hubp2 =
-		kzalloc(sizeof(struct dcn20_hubp), GFP_KERNEL);
-
-	if (!hubp2)
-		return NULL;
-
-	if (hubp31_construct(hubp2, ctx, inst,
-			&hubp_regs[inst], &hubp_shift, &hubp_mask))
-		return &hubp2->base;
-
-	BREAK_TO_DEBUGGER();
-	kfree(hubp2);
-	return NULL;
-}
-
-static bool dcn31_dwbc_create(struct dc_context *ctx, struct resource_pool *pool)
-{
-	int i;
-	uint32_t pipe_count = pool->res_cap->num_dwb;
-
-	for (i = 0; i < pipe_count; i++) {
-		struct dcn30_dwbc *dwbc30 = kzalloc(sizeof(struct dcn30_dwbc),
-						    GFP_KERNEL);
-
-		if (!dwbc30) {
-			dm_error("DC: failed to create dwbc30!\n");
-			return false;
-		}
-
-		dcn30_dwbc_construct(dwbc30, ctx,
-				&dwbc30_regs[i],
-				&dwbc30_shift,
-				&dwbc30_mask,
-				i);
-
-		pool->dwbc[i] = &dwbc30->base;
-	}
-	return true;
-}
-
-static bool dcn31_mmhubbub_create(struct dc_context *ctx, struct resource_pool *pool)
-{
-	int i;
-	uint32_t pipe_count = pool->res_cap->num_dwb;
-
-	for (i = 0; i < pipe_count; i++) {
-		struct dcn30_mmhubbub *mcif_wb30 = kzalloc(sizeof(struct dcn30_mmhubbub),
-						    GFP_KERNEL);
-
-		if (!mcif_wb30) {
-			dm_error("DC: failed to create mcif_wb30!\n");
-			return false;
-		}
-
-		dcn30_mmhubbub_construct(mcif_wb30, ctx,
-				&mcif_wb30_regs[i],
-				&mcif_wb30_shift,
-				&mcif_wb30_mask,
-				i);
-
-		pool->mcif_wb[i] = &mcif_wb30->base;
-	}
-	return true;
-}
-
-static struct display_stream_compressor *dcn31_dsc_create(
-	struct dc_context *ctx, uint32_t inst)
-{
-	struct dcn20_dsc *dsc =
-		kzalloc(sizeof(struct dcn20_dsc), GFP_KERNEL);
-
-	if (!dsc) {
-		BREAK_TO_DEBUGGER();
-		return NULL;
-	}
-
-	dsc2_construct(dsc, ctx, inst, &dsc_regs[inst], &dsc_shift, &dsc_mask);
-	return &dsc->base;
-}
-
-static void dcn31_destroy_resource_pool(struct resource_pool **pool)
-{
-	struct dcn31_resource_pool *dcn31_pool = TO_DCN31_RES_POOL(*pool);
-
-	dcn31_resource_destruct(dcn31_pool);
-	kfree(dcn31_pool);
-	*pool = NULL;
-}
-
-static struct clock_source *dcn31_clock_source_create(
-		struct dc_context *ctx,
-		struct dc_bios *bios,
-		enum clock_source_id id,
-		const struct dce110_clk_src_regs *regs,
-		bool dp_clk_src)
-{
-	struct dce110_clk_src *clk_src =
-		kzalloc(sizeof(struct dce110_clk_src), GFP_KERNEL);
-
-	if (!clk_src)
-		return NULL;
-
-	if (dcn3_clk_src_construct(clk_src, ctx, bios, id,
-			regs, &cs_shift, &cs_mask)) {
-		clk_src->base.dp_clk_src = dp_clk_src;
-		return &clk_src->base;
-	}
-
-	BREAK_TO_DEBUGGER();
-	return NULL;
-}
-
-static bool is_dual_plane(enum surface_pixel_format format)
-{
-	return format >= SURFACE_PIXEL_FORMAT_VIDEO_BEGIN || format == SURFACE_PIXEL_FORMAT_GRPH_RGBE_ALPHA;
-}
-
-static int dcn31_populate_dml_pipes_from_context(
-	struct dc *dc, struct dc_state *context,
-	display_e2e_pipe_params_st *pipes,
-	bool fast_validate)
-{
-	int i, pipe_cnt;
-	struct resource_context *res_ctx = &context->res_ctx;
-	struct pipe_ctx *pipe;
-
-	dcn20_populate_dml_pipes_from_context(dc, context, pipes, fast_validate);
-
-	for (i = 0, pipe_cnt = 0; i < dc->res_pool->pipe_count; i++) {
-		struct dc_crtc_timing *timing;
-
-		if (!res_ctx->pipe_ctx[i].stream)
-			continue;
-		pipe = &res_ctx->pipe_ctx[i];
-		timing = &pipe->stream->timing;
-
-		/*
-		 * Immediate flip can be set dynamically after enabling the plane.
-		 * We need to require support for immediate flip or underflow can be
-		 * intermittently experienced depending on peak b/w requirements.
-		 */
-		pipes[pipe_cnt].pipe.src.immediate_flip = true;
-
-		pipes[pipe_cnt].pipe.src.unbounded_req_mode = false;
-		pipes[pipe_cnt].pipe.src.gpuvm = true;
-		pipes[pipe_cnt].pipe.src.dcc_fraction_of_zs_req_luma = 0;
-		pipes[pipe_cnt].pipe.src.dcc_fraction_of_zs_req_chroma = 0;
-		pipes[pipe_cnt].pipe.dest.vfront_porch = timing->v_front_porch;
-		pipes[pipe_cnt].pipe.src.dcc_rate = 3;
-		pipes[pipe_cnt].dout.dsc_input_bpc = 0;
-
-		if (pipes[pipe_cnt].dout.dsc_enable) {
-			switch (timing->display_color_depth) {
-			case COLOR_DEPTH_888:
-				pipes[pipe_cnt].dout.dsc_input_bpc = 8;
-				break;
-			case COLOR_DEPTH_101010:
-				pipes[pipe_cnt].dout.dsc_input_bpc = 10;
-				break;
-			case COLOR_DEPTH_121212:
-				pipes[pipe_cnt].dout.dsc_input_bpc = 12;
-				break;
-			default:
-				ASSERT(0);
-				break;
-			}
-		}
-
-		pipe_cnt++;
-	}
-	context->bw_ctx.dml.ip.det_buffer_size_kbytes = DCN3_1_DEFAULT_DET_SIZE;
-	dc->config.enable_4to1MPC = false;
-	if (pipe_cnt == 1 && pipe->plane_state && !dc->debug.disable_z9_mpc) {
-		if (is_dual_plane(pipe->plane_state->format)
-				&& pipe->plane_state->src_rect.width <= 1920 && pipe->plane_state->src_rect.height <= 1080) {
-			dc->config.enable_4to1MPC = true;
-		} else if (!is_dual_plane(pipe->plane_state->format)) {
-			context->bw_ctx.dml.ip.det_buffer_size_kbytes = 192;
-			pipes[0].pipe.src.unbounded_req_mode = true;
-		}
-	}
-
-	return pipe_cnt;
-}
-
-void dcn31_update_soc_for_wm_a(struct dc *dc, struct dc_state *context)
-{
-	if (dc->clk_mgr->bw_params->wm_table.entries[WM_A].valid) {
-		context->bw_ctx.dml.soc.dram_clock_change_latency_us = dc->clk_mgr->bw_params->wm_table.entries[WM_A].pstate_latency_us;
-		context->bw_ctx.dml.soc.sr_enter_plus_exit_time_us = dc->clk_mgr->bw_params->wm_table.entries[WM_A].sr_enter_plus_exit_time_us;
-		context->bw_ctx.dml.soc.sr_exit_time_us = dc->clk_mgr->bw_params->wm_table.entries[WM_A].sr_exit_time_us;
-	}
-}
-
-static void dcn31_calculate_wm_and_dlg_fp(
-		struct dc *dc, struct dc_state *context,
-		display_e2e_pipe_params_st *pipes,
-		int pipe_cnt,
-		int vlevel)
-{
-	int i, pipe_idx;
-	double dcfclk = context->bw_ctx.dml.vba.DCFCLKState[vlevel][context->bw_ctx.dml.vba.maxMpcComb];
-
-	if (context->bw_ctx.dml.soc.min_dcfclk > dcfclk)
-		dcfclk = context->bw_ctx.dml.soc.min_dcfclk;
-
-	/* We don't recalculate clocks for 0 pipe configs, which can block
-	 * S0i3 as high clocks will block low power states
-	 * Override any clocks that can block S0i3 to min here
-	 */
-	if (pipe_cnt == 0) {
-		context->bw_ctx.bw.dcn.clk.dcfclk_khz = dcfclk; // always should be vlevel 0
-		return;
-	}
-
-	pipes[0].clks_cfg.voltage = vlevel;
-	pipes[0].clks_cfg.dcfclk_mhz = dcfclk;
-	pipes[0].clks_cfg.socclk_mhz = context->bw_ctx.dml.soc.clock_limits[vlevel].socclk_mhz;
-
-#if 0 // TODO
-	/* Set B:
-	 * TODO
-	 */
-	if (dc->clk_mgr->bw_params->wm_table.nv_entries[WM_B].valid) {
-		if (vlevel == 0) {
-			pipes[0].clks_cfg.voltage = 1;
-			pipes[0].clks_cfg.dcfclk_mhz = context->bw_ctx.dml.soc.clock_limits[0].dcfclk_mhz;
-		}
-		context->bw_ctx.dml.soc.dram_clock_change_latency_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_B].dml_input.pstate_latency_us;
-		context->bw_ctx.dml.soc.sr_enter_plus_exit_time_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_B].dml_input.sr_enter_plus_exit_time_us;
-		context->bw_ctx.dml.soc.sr_exit_time_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_B].dml_input.sr_exit_time_us;
-	}
-	context->bw_ctx.bw.dcn.watermarks.b.urgent_ns = get_wm_urgent(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.b.cstate_pstate.cstate_enter_plus_exit_ns = get_wm_stutter_enter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.b.cstate_pstate.cstate_exit_ns = get_wm_stutter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.b.cstate_pstate.cstate_enter_plus_exit_z8_ns = get_wm_z8_stutter_enter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.b.cstate_pstate.cstate_exit_z8_ns = get_wm_z8_stutter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.b.cstate_pstate.pstate_change_ns = get_wm_dram_clock_change(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.b.pte_meta_urgent_ns = get_wm_memory_trip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.b.frac_urg_bw_nom = get_fraction_of_urgent_bandwidth(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.b.frac_urg_bw_flip = get_fraction_of_urgent_bandwidth_imm_flip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.b.urgent_latency_ns = get_urgent_latency(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-
-	pipes[0].clks_cfg.voltage = vlevel;
-	pipes[0].clks_cfg.dcfclk_mhz = dcfclk;
-
-	/* Set C:
-	 * TODO
-	 */
-	if (dc->clk_mgr->bw_params->wm_table.nv_entries[WM_C].valid) {
-		context->bw_ctx.dml.soc.dram_clock_change_latency_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_C].dml_input.pstate_latency_us;
-		context->bw_ctx.dml.soc.sr_enter_plus_exit_time_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_C].dml_input.sr_enter_plus_exit_time_us;
-		context->bw_ctx.dml.soc.sr_exit_time_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_C].dml_input.sr_exit_time_us;
-	}
-	context->bw_ctx.bw.dcn.watermarks.c.urgent_ns = get_wm_urgent(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.c.cstate_pstate.cstate_enter_plus_exit_ns = get_wm_stutter_enter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.c.cstate_pstate.cstate_exit_ns = get_wm_stutter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.c.cstate_pstate.cstate_enter_plus_exit_z8_ns = get_wm_z8_stutter_enter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.c.cstate_pstate.cstate_exit_z8_ns = get_wm_z8_stutter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.c.cstate_pstate.pstate_change_ns = get_wm_dram_clock_change(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.c.pte_meta_urgent_ns = get_wm_memory_trip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.c.frac_urg_bw_nom = get_fraction_of_urgent_bandwidth(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.c.frac_urg_bw_flip = get_fraction_of_urgent_bandwidth_imm_flip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.c.urgent_latency_ns = get_urgent_latency(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-
-	/* Set D:
-	 * TODO
-	 */
-	if (dc->clk_mgr->bw_params->wm_table.nv_entries[WM_D].valid) {
-		context->bw_ctx.dml.soc.dram_clock_change_latency_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_D].dml_input.pstate_latency_us;
-		context->bw_ctx.dml.soc.sr_enter_plus_exit_time_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_D].dml_input.sr_enter_plus_exit_time_us;
-		context->bw_ctx.dml.soc.sr_exit_time_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_D].dml_input.sr_exit_time_us;
-	}
-	context->bw_ctx.bw.dcn.watermarks.d.urgent_ns = get_wm_urgent(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.d.cstate_pstate.cstate_enter_plus_exit_ns = get_wm_stutter_enter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.d.cstate_pstate.cstate_exit_ns = get_wm_stutter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.d.cstate_pstate.pstate_change_ns = get_wm_dram_clock_change(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.d.cstate_pstate.cstate_enter_plus_exit_z8_ns = get_wm_z8_stutter_enter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.d.cstate_pstate.cstate_exit_z8_ns = get_wm_z8_stutter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.d.pte_meta_urgent_ns = get_wm_memory_trip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.d.frac_urg_bw_nom = get_fraction_of_urgent_bandwidth(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.d.frac_urg_bw_flip = get_fraction_of_urgent_bandwidth_imm_flip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.d.urgent_latency_ns = get_urgent_latency(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-#endif
-
-	/* Set A:
-	 * All clocks min required
-	 *
-	 * Set A calculated last so that following calculations are based on Set A
-	 */
-	dc->res_pool->funcs->update_soc_for_wm_a(dc, context);
-	context->bw_ctx.bw.dcn.watermarks.a.urgent_ns = get_wm_urgent(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.a.cstate_pstate.cstate_enter_plus_exit_ns = get_wm_stutter_enter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.a.cstate_pstate.cstate_exit_ns = get_wm_stutter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.a.cstate_pstate.pstate_change_ns = get_wm_dram_clock_change(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.a.cstate_pstate.cstate_enter_plus_exit_z8_ns = get_wm_z8_stutter_enter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.a.cstate_pstate.cstate_exit_z8_ns = get_wm_z8_stutter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.a.pte_meta_urgent_ns = get_wm_memory_trip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.a.frac_urg_bw_nom = get_fraction_of_urgent_bandwidth(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.a.frac_urg_bw_flip = get_fraction_of_urgent_bandwidth_imm_flip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.a.urgent_latency_ns = get_urgent_latency(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	/* TODO: remove: */
-	context->bw_ctx.bw.dcn.watermarks.b = context->bw_ctx.bw.dcn.watermarks.a;
-	context->bw_ctx.bw.dcn.watermarks.c = context->bw_ctx.bw.dcn.watermarks.a;
-	context->bw_ctx.bw.dcn.watermarks.d = context->bw_ctx.bw.dcn.watermarks.a;
-	/* end remove*/
-
-	for (i = 0, pipe_idx = 0; i < dc->res_pool->pipe_count; i++) {
-		if (!context->res_ctx.pipe_ctx[i].stream)
-			continue;
-
-		pipes[pipe_idx].clks_cfg.dispclk_mhz = get_dispclk_calculated(&context->bw_ctx.dml, pipes, pipe_cnt);
-		pipes[pipe_idx].clks_cfg.dppclk_mhz = get_dppclk_calculated(&context->bw_ctx.dml, pipes, pipe_cnt, pipe_idx);
-
-		if (dc->config.forced_clocks) {
-			pipes[pipe_idx].clks_cfg.dispclk_mhz = context->bw_ctx.dml.soc.clock_limits[0].dispclk_mhz;
-			pipes[pipe_idx].clks_cfg.dppclk_mhz = context->bw_ctx.dml.soc.clock_limits[0].dppclk_mhz;
-		}
-		if (dc->debug.min_disp_clk_khz > pipes[pipe_idx].clks_cfg.dispclk_mhz * 1000)
-			pipes[pipe_idx].clks_cfg.dispclk_mhz = dc->debug.min_disp_clk_khz / 1000.0;
-		if (dc->debug.min_dpp_clk_khz > pipes[pipe_idx].clks_cfg.dppclk_mhz * 1000)
-			pipes[pipe_idx].clks_cfg.dppclk_mhz = dc->debug.min_dpp_clk_khz / 1000.0;
-
-		pipe_idx++;
-	}
-
-	dcn20_calculate_dlg_params(dc, context, pipes, pipe_cnt, vlevel);
-}
-
-void dcn31_calculate_wm_and_dlg(
-		struct dc *dc, struct dc_state *context,
-		display_e2e_pipe_params_st *pipes,
-		int pipe_cnt,
-		int vlevel)
-{
-	DC_FP_START();
-	dcn31_calculate_wm_and_dlg_fp(dc, context, pipes, pipe_cnt, vlevel);
-	DC_FP_END();
-}
-
-bool dcn31_validate_bandwidth(struct dc *dc,
-		struct dc_state *context,
-		bool fast_validate)
-{
-	bool out = false;
-
-	BW_VAL_TRACE_SETUP();
-
-	int vlevel = 0;
-	int pipe_cnt = 0;
-	display_e2e_pipe_params_st *pipes = kzalloc(dc->res_pool->pipe_count * sizeof(display_e2e_pipe_params_st), GFP_KERNEL);
-	DC_LOGGER_INIT(dc->ctx->logger);
-
-	BW_VAL_TRACE_COUNT();
-
-	out = dcn30_internal_validate_bw(dc, context, pipes, &pipe_cnt, &vlevel, fast_validate);
-
-	// Disable fast_validate to set min dcfclk in alculate_wm_and_dlg
-	if (pipe_cnt == 0)
-		fast_validate = false;
-
-	if (!out)
-		goto validate_fail;
-
-	BW_VAL_TRACE_END_VOLTAGE_LEVEL();
-
-	if (fast_validate) {
-		BW_VAL_TRACE_SKIP(fast);
-		goto validate_out;
-	}
-
-	dc->res_pool->funcs->calculate_wm_and_dlg(dc, context, pipes, pipe_cnt, vlevel);
-
-	BW_VAL_TRACE_END_WATERMARKS();
-
-	goto validate_out;
-
-validate_fail:
-	DC_LOG_WARNING("Mode Validation Warning: %s failed validation.\n",
-		dml_get_status_message(context->bw_ctx.dml.vba.ValidationStatus[context->bw_ctx.dml.vba.soc.num_states]));
-
-	BW_VAL_TRACE_SKIP(fail);
-	out = false;
-
-validate_out:
-	kfree(pipes);
-
-	BW_VAL_TRACE_FINISH();
-
-	return out;
-}
-
-static struct dc_cap_funcs cap_funcs = {
-	.get_dcc_compression_cap = dcn20_get_dcc_compression_cap
-};
-
-static void dcn31_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params)
-{
-	struct clk_limit_table *clk_table = &bw_params->clk_table;
-	struct _vcs_dpi_voltage_scaling_st clock_limits[DC__VOLTAGE_STATES];
-	unsigned int i, closest_clk_lvl;
-	int j;
-
-	// Default clock levels are used for diags, which may lead to overclocking.
-	if (!IS_DIAG_DC(dc->ctx->dce_environment)) {
-		int max_dispclk_mhz = 0, max_dppclk_mhz = 0;
-
-		dcn3_1_ip.max_num_otg = dc->res_pool->res_cap->num_timing_generator;
-		dcn3_1_ip.max_num_dpp = dc->res_pool->pipe_count;
-		dcn3_1_soc.num_chans = bw_params->num_channels;
-
-		ASSERT(clk_table->num_entries);
-
-		/* Prepass to find max clocks independent of voltage level. */
-		for (i = 0; i < clk_table->num_entries; ++i) {
-			if (clk_table->entries[i].dispclk_mhz > max_dispclk_mhz)
-				max_dispclk_mhz = clk_table->entries[i].dispclk_mhz;
-			if (clk_table->entries[i].dppclk_mhz > max_dppclk_mhz)
-				max_dppclk_mhz = clk_table->entries[i].dppclk_mhz;
-		}
-
-		for (i = 0; i < clk_table->num_entries; i++) {
-			/* loop backwards*/
-			for (closest_clk_lvl = 0, j = dcn3_1_soc.num_states - 1; j >= 0; j--) {
-				if ((unsigned int) dcn3_1_soc.clock_limits[j].dcfclk_mhz <= clk_table->entries[i].dcfclk_mhz) {
-					closest_clk_lvl = j;
-					break;
-				}
-			}
-
-			clock_limits[i].state = i;
-
-			/* Clocks dependent on voltage level. */
-			clock_limits[i].dcfclk_mhz = clk_table->entries[i].dcfclk_mhz;
-			clock_limits[i].fabricclk_mhz = clk_table->entries[i].fclk_mhz;
-			clock_limits[i].socclk_mhz = clk_table->entries[i].socclk_mhz;
-			clock_limits[i].dram_speed_mts = clk_table->entries[i].memclk_mhz * 2 * clk_table->entries[i].wck_ratio;
-
-			/* Clocks independent of voltage level. */
-			clock_limits[i].dispclk_mhz = max_dispclk_mhz ? max_dispclk_mhz :
-				dcn3_1_soc.clock_limits[closest_clk_lvl].dispclk_mhz;
-
-			clock_limits[i].dppclk_mhz = max_dppclk_mhz ? max_dppclk_mhz :
-				dcn3_1_soc.clock_limits[closest_clk_lvl].dppclk_mhz;
-
-			clock_limits[i].dram_bw_per_chan_gbps = dcn3_1_soc.clock_limits[closest_clk_lvl].dram_bw_per_chan_gbps;
-			clock_limits[i].dscclk_mhz = dcn3_1_soc.clock_limits[closest_clk_lvl].dscclk_mhz;
-			clock_limits[i].dtbclk_mhz = dcn3_1_soc.clock_limits[closest_clk_lvl].dtbclk_mhz;
-			clock_limits[i].phyclk_d18_mhz = dcn3_1_soc.clock_limits[closest_clk_lvl].phyclk_d18_mhz;
-			clock_limits[i].phyclk_mhz = dcn3_1_soc.clock_limits[closest_clk_lvl].phyclk_mhz;
-		}
-		for (i = 0; i < clk_table->num_entries; i++)
-			dcn3_1_soc.clock_limits[i] = clock_limits[i];
-		if (clk_table->num_entries) {
-			dcn3_1_soc.num_states = clk_table->num_entries;
-		}
-	}
-
-	dcn3_1_soc.dispclk_dppclk_vco_speed_mhz = dc->clk_mgr->dentist_vco_freq_khz / 1000.0;
-	dc->dml.soc.dispclk_dppclk_vco_speed_mhz = dc->clk_mgr->dentist_vco_freq_khz / 1000.0;
-
-	if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment))
-		dml_init_instance(&dc->dml, &dcn3_1_soc, &dcn3_1_ip, DML_PROJECT_DCN31);
-	else
-		dml_init_instance(&dc->dml, &dcn3_1_soc, &dcn3_1_ip, DML_PROJECT_DCN31_FPGA);
-}
-
-static struct resource_funcs dcn31_res_pool_funcs = {
-	.destroy = dcn31_destroy_resource_pool,
-	.link_enc_create = dcn31_link_encoder_create,
-	.link_enc_create_minimal = dcn31_link_enc_create_minimal,
-	.link_encs_assign = link_enc_cfg_link_encs_assign,
-	.link_enc_unassign = link_enc_cfg_link_enc_unassign,
-	.panel_cntl_create = dcn31_panel_cntl_create,
-	.validate_bandwidth = dcn31_validate_bandwidth,
-	.calculate_wm_and_dlg = dcn31_calculate_wm_and_dlg,
-	.update_soc_for_wm_a = dcn31_update_soc_for_wm_a,
-	.populate_dml_pipes = dcn31_populate_dml_pipes_from_context,
-	.acquire_idle_pipe_for_layer = dcn20_acquire_idle_pipe_for_layer,
-	.add_stream_to_ctx = dcn30_add_stream_to_ctx,
-	.add_dsc_to_stream_resource = dcn20_add_dsc_to_stream_resource,
-	.remove_stream_from_ctx = dcn20_remove_stream_from_ctx,
-	.populate_dml_writeback_from_context = dcn30_populate_dml_writeback_from_context,
-	.set_mcif_arb_params = dcn30_set_mcif_arb_params,
-	.find_first_free_match_stream_enc_for_link = dcn10_find_first_free_match_stream_enc_for_link,
-	.acquire_post_bldn_3dlut = dcn30_acquire_post_bldn_3dlut,
-	.release_post_bldn_3dlut = dcn30_release_post_bldn_3dlut,
-	.update_bw_bounding_box = dcn31_update_bw_bounding_box,
-	.patch_unknown_plane_state = dcn20_patch_unknown_plane_state,
-};
-
-static struct clock_source *dcn30_clock_source_create(
-		struct dc_context *ctx,
-		struct dc_bios *bios,
-		enum clock_source_id id,
-		const struct dce110_clk_src_regs *regs,
-		bool dp_clk_src)
-{
-	struct dce110_clk_src *clk_src =
-		kzalloc(sizeof(struct dce110_clk_src), GFP_KERNEL);
-
-	if (!clk_src)
-		return NULL;
-
-	if (dcn3_clk_src_construct(clk_src, ctx, bios, id,
-			regs, &cs_shift, &cs_mask)) {
-		clk_src->base.dp_clk_src = dp_clk_src;
-		return &clk_src->base;
-	}
-
-	BREAK_TO_DEBUGGER();
-	return NULL;
-}
-
-static bool dcn31_resource_construct(
-	uint8_t num_virtual_links,
-	struct dc *dc,
-	struct dcn31_resource_pool *pool)
-{
-	int i;
-	struct dc_context *ctx = dc->ctx;
-	struct irq_service_init_data init_data;
-
-	DC_FP_START();
-
-	ctx->dc_bios->regs = &bios_regs;
-
-	pool->base.res_cap = &res_cap_dcn31;
-
-	pool->base.funcs = &dcn31_res_pool_funcs;
-
-	/*************************************************
-	 *  Resource + asic cap harcoding                *
-	 *************************************************/
-	pool->base.underlay_pipe_index = NO_UNDERLAY_PIPE;
-	pool->base.pipe_count = pool->base.res_cap->num_timing_generator;
-	pool->base.mpcc_count = pool->base.res_cap->num_timing_generator;
-	dc->caps.max_downscale_ratio = 600;
-	dc->caps.i2c_speed_in_khz = 100;
-	dc->caps.i2c_speed_in_khz_hdcp = 5; /*1.4 w/a applied by default*/
-	dc->caps.max_cursor_size = 256;
-	dc->caps.min_horizontal_blanking_period = 80;
-	dc->caps.dmdata_alloc_size = 2048;
-
-	dc->caps.max_slave_planes = 1;
-	dc->caps.max_slave_yuv_planes = 1;
-	dc->caps.max_slave_rgb_planes = 1;
-	dc->caps.post_blend_color_processing = true;
-	dc->caps.force_dp_tps4_for_cp2520 = true;
-	dc->caps.dp_hpo = true;
-	dc->caps.extended_aux_timeout_support = true;
-	dc->caps.dmcub_support = true;
-	dc->caps.is_apu = true;
-
-	/* Color pipeline capabilities */
-	dc->caps.color.dpp.dcn_arch = 1;
-	dc->caps.color.dpp.input_lut_shared = 0;
-	dc->caps.color.dpp.icsc = 1;
-	dc->caps.color.dpp.dgam_ram = 0; // must use gamma_corr
-	dc->caps.color.dpp.dgam_rom_caps.srgb = 1;
-	dc->caps.color.dpp.dgam_rom_caps.bt2020 = 1;
-	dc->caps.color.dpp.dgam_rom_caps.gamma2_2 = 1;
-	dc->caps.color.dpp.dgam_rom_caps.pq = 1;
-	dc->caps.color.dpp.dgam_rom_caps.hlg = 1;
-	dc->caps.color.dpp.post_csc = 1;
-	dc->caps.color.dpp.gamma_corr = 1;
-	dc->caps.color.dpp.dgam_rom_for_yuv = 0;
-
-	dc->caps.color.dpp.hw_3d_lut = 1;
-	dc->caps.color.dpp.ogam_ram = 1;
-	// no OGAM ROM on DCN301
-	dc->caps.color.dpp.ogam_rom_caps.srgb = 0;
-	dc->caps.color.dpp.ogam_rom_caps.bt2020 = 0;
-	dc->caps.color.dpp.ogam_rom_caps.gamma2_2 = 0;
-	dc->caps.color.dpp.ogam_rom_caps.pq = 0;
-	dc->caps.color.dpp.ogam_rom_caps.hlg = 0;
-	dc->caps.color.dpp.ocsc = 0;
-
-	dc->caps.color.mpc.gamut_remap = 1;
-	dc->caps.color.mpc.num_3dluts = pool->base.res_cap->num_mpc_3dlut; //2
-	dc->caps.color.mpc.ogam_ram = 1;
-	dc->caps.color.mpc.ogam_rom_caps.srgb = 0;
-	dc->caps.color.mpc.ogam_rom_caps.bt2020 = 0;
-	dc->caps.color.mpc.ogam_rom_caps.gamma2_2 = 0;
-	dc->caps.color.mpc.ogam_rom_caps.pq = 0;
-	dc->caps.color.mpc.ogam_rom_caps.hlg = 0;
-	dc->caps.color.mpc.ocsc = 1;
-
-	/* read VBIOS LTTPR caps */
-	{
-		if (ctx->dc_bios->funcs->get_lttpr_caps) {
-			enum bp_result bp_query_result;
-			uint8_t is_vbios_lttpr_enable = 0;
-
-			bp_query_result = ctx->dc_bios->funcs->get_lttpr_caps(ctx->dc_bios, &is_vbios_lttpr_enable);
-			dc->caps.vbios_lttpr_enable = (bp_query_result == BP_RESULT_OK) && !!is_vbios_lttpr_enable;
-		}
-
-		/* interop bit is implicit */
-		{
-			dc->caps.vbios_lttpr_aware = true;
-		}
-	}
-
-	if (dc->ctx->dce_environment == DCE_ENV_PRODUCTION_DRV)
-		dc->debug = debug_defaults_drv;
-	else if (dc->ctx->dce_environment == DCE_ENV_FPGA_MAXIMUS) {
-		dc->debug = debug_defaults_diags;
-	} else
-		dc->debug = debug_defaults_diags;
-	// Init the vm_helper
-	if (dc->vm_helper)
-		vm_helper_init(dc->vm_helper, 16);
-
-	/*************************************************
-	 *  Create resources                             *
-	 *************************************************/
-
-	/* Clock Sources for Pixel Clock*/
-	pool->base.clock_sources[DCN31_CLK_SRC_PLL0] =
-			dcn30_clock_source_create(ctx, ctx->dc_bios,
-				CLOCK_SOURCE_COMBO_PHY_PLL0,
-				&clk_src_regs[0], false);
-	pool->base.clock_sources[DCN31_CLK_SRC_PLL1] =
-			dcn30_clock_source_create(ctx, ctx->dc_bios,
-				CLOCK_SOURCE_COMBO_PHY_PLL1,
-				&clk_src_regs[1], false);
-	pool->base.clock_sources[DCN31_CLK_SRC_PLL2] =
-			dcn30_clock_source_create(ctx, ctx->dc_bios,
-				CLOCK_SOURCE_COMBO_PHY_PLL2,
-				&clk_src_regs[2], false);
-	pool->base.clock_sources[DCN31_CLK_SRC_PLL3] =
-			dcn30_clock_source_create(ctx, ctx->dc_bios,
-				CLOCK_SOURCE_COMBO_PHY_PLL3,
-				&clk_src_regs[3], false);
-	pool->base.clock_sources[DCN31_CLK_SRC_PLL4] =
-			dcn30_clock_source_create(ctx, ctx->dc_bios,
-				CLOCK_SOURCE_COMBO_PHY_PLL4,
-				&clk_src_regs[4], false);
-
-	pool->base.clk_src_count = DCN30_CLK_SRC_TOTAL;
-
-	/* todo: not reuse phy_pll registers */
-	pool->base.dp_clock_source =
-			dcn31_clock_source_create(ctx, ctx->dc_bios,
-				CLOCK_SOURCE_ID_DP_DTO,
-				&clk_src_regs[0], true);
-
-	for (i = 0; i < pool->base.clk_src_count; i++) {
-		if (pool->base.clock_sources[i] == NULL) {
-			dm_error("DC: failed to create clock sources!\n");
-			BREAK_TO_DEBUGGER();
-			goto create_fail;
-		}
-	}
-
-	/* TODO: DCCG */
-	pool->base.dccg = dccg31_create(ctx, &dccg_regs, &dccg_shift, &dccg_mask);
-	if (pool->base.dccg == NULL) {
-		dm_error("DC: failed to create dccg!\n");
-		BREAK_TO_DEBUGGER();
-		goto create_fail;
-	}
-
-	/* TODO: IRQ */
-	init_data.ctx = dc->ctx;
-	pool->base.irqs = dal_irq_service_dcn31_create(&init_data);
-	if (!pool->base.irqs)
-		goto create_fail;
-
-	/* HUBBUB */
-	pool->base.hubbub = dcn31_hubbub_create(ctx);
-	if (pool->base.hubbub == NULL) {
-		BREAK_TO_DEBUGGER();
-		dm_error("DC: failed to create hubbub!\n");
-		goto create_fail;
-	}
-
-	/* HUBPs, DPPs, OPPs and TGs */
-	for (i = 0; i < pool->base.pipe_count; i++) {
-		pool->base.hubps[i] = dcn31_hubp_create(ctx, i);
-		if (pool->base.hubps[i] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error(
-				"DC: failed to create hubps!\n");
-			goto create_fail;
-		}
-
-		pool->base.dpps[i] = dcn31_dpp_create(ctx, i);
-		if (pool->base.dpps[i] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error(
-				"DC: failed to create dpps!\n");
-			goto create_fail;
-		}
-	}
-
-	for (i = 0; i < pool->base.res_cap->num_opp; i++) {
-		pool->base.opps[i] = dcn31_opp_create(ctx, i);
-		if (pool->base.opps[i] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error(
-				"DC: failed to create output pixel processor!\n");
-			goto create_fail;
-		}
-	}
-
-	for (i = 0; i < pool->base.res_cap->num_timing_generator; i++) {
-		pool->base.timing_generators[i] = dcn31_timing_generator_create(
-				ctx, i);
-		if (pool->base.timing_generators[i] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error("DC: failed to create tg!\n");
-			goto create_fail;
-		}
-	}
-	pool->base.timing_generator_count = i;
-
-	/* PSR */
-	pool->base.psr = dmub_psr_create(ctx);
-	if (pool->base.psr == NULL) {
-		dm_error("DC: failed to create psr obj!\n");
-		BREAK_TO_DEBUGGER();
-		goto create_fail;
-	}
-
-	/* ABM */
-	for (i = 0; i < pool->base.res_cap->num_timing_generator; i++) {
-		pool->base.multiple_abms[i] = dmub_abm_create(ctx,
-				&abm_regs[i],
-				&abm_shift,
-				&abm_mask);
-		if (pool->base.multiple_abms[i] == NULL) {
-			dm_error("DC: failed to create abm for pipe %d!\n", i);
-			BREAK_TO_DEBUGGER();
-			goto create_fail;
-		}
-	}
-
-	/* MPC and DSC */
-	pool->base.mpc = dcn31_mpc_create(ctx, pool->base.mpcc_count, pool->base.res_cap->num_mpc_3dlut);
-	if (pool->base.mpc == NULL) {
-		BREAK_TO_DEBUGGER();
-		dm_error("DC: failed to create mpc!\n");
-		goto create_fail;
-	}
-
-	for (i = 0; i < pool->base.res_cap->num_dsc; i++) {
-		pool->base.dscs[i] = dcn31_dsc_create(ctx, i);
-		if (pool->base.dscs[i] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error("DC: failed to create display stream compressor %d!\n", i);
-			goto create_fail;
-		}
-	}
-
-	/* DWB and MMHUBBUB */
-	if (!dcn31_dwbc_create(ctx, &pool->base)) {
-		BREAK_TO_DEBUGGER();
-		dm_error("DC: failed to create dwbc!\n");
-		goto create_fail;
-	}
-
-	if (!dcn31_mmhubbub_create(ctx, &pool->base)) {
-		BREAK_TO_DEBUGGER();
-		dm_error("DC: failed to create mcif_wb!\n");
-		goto create_fail;
-	}
-
-	/* AUX and I2C */
-	for (i = 0; i < pool->base.res_cap->num_ddc; i++) {
-		pool->base.engines[i] = dcn31_aux_engine_create(ctx, i);
-		if (pool->base.engines[i] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error(
-				"DC:failed to create aux engine!!\n");
-			goto create_fail;
-		}
-		pool->base.hw_i2cs[i] = dcn31_i2c_hw_create(ctx, i);
-		if (pool->base.hw_i2cs[i] == NULL) {
-			BREAK_TO_DEBUGGER();
-			dm_error(
-				"DC:failed to create hw i2c!!\n");
-			goto create_fail;
-		}
-		pool->base.sw_i2cs[i] = NULL;
-	}
-
-	if (dc->ctx->asic_id.chip_family == FAMILY_YELLOW_CARP &&
-	    dc->ctx->asic_id.hw_internal_rev == YELLOW_CARP_B0 &&
-	    !dc->debug.dpia_debug.bits.disable_dpia) {
-		/* YELLOW CARP B0 has 4 DPIA's */
-		pool->base.usb4_dpia_count = 4;
-	}
-
-	/* Audio, Stream Encoders including HPO and virtual, MPC 3D LUTs */
-	if (!resource_construct(num_virtual_links, dc, &pool->base,
-			(!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment) ?
-			&res_create_funcs : &res_create_maximus_funcs)))
-			goto create_fail;
-
-	/* HW Sequencer and Plane caps */
-	dcn31_hw_sequencer_construct(dc);
-
-	dc->caps.max_planes =  pool->base.pipe_count;
-
-	for (i = 0; i < dc->caps.max_planes; ++i)
-		dc->caps.planes[i] = plane_cap;
-
-	dc->cap_funcs = cap_funcs;
-
-	dc->dcn_ip->max_num_dpp = dcn3_1_ip.max_num_dpp;
-
-	DC_FP_END();
-
-	return true;
-
-create_fail:
-
-	DC_FP_END();
-	dcn31_resource_destruct(pool);
-
-	return false;
-}
-
-struct resource_pool *dcn31_create_resource_pool(
-		const struct dc_init_data *init_data,
-		struct dc *dc)
-{
-	struct dcn31_resource_pool *pool =
-		kzalloc(sizeof(struct dcn31_resource_pool), GFP_KERNEL);
-
-	if (!pool)
-		return NULL;
-
-	if (dcn31_resource_construct(init_data->num_virtual_links, dc, pool))
-		return &pool->base;
-
-	BREAK_TO_DEBUGGER();
-	kfree(pool);
-	return NULL;
-}
diff --git a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_resource.h b/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_resource.h
deleted file mode 100644
index 416fe7a721d8..000000000000
--- a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_resource.h
+++ /dev/null
@@ -1,52 +0,0 @@
-/*
- * Copyright 2020 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: AMD
- *
- */
-
-#ifndef _DCN31_RESOURCE_H_
-#define _DCN31_RESOURCE_H_
-
-#include "core_types.h"
-
-#define TO_DCN31_RES_POOL(pool)\
-	container_of(pool, struct dcn31_resource_pool, base)
-
-struct dcn31_resource_pool {
-	struct resource_pool base;
-};
-
-bool dcn31_validate_bandwidth(struct dc *dc,
-		struct dc_state *context,
-		bool fast_validate);
-void dcn31_calculate_wm_and_dlg(
-		struct dc *dc, struct dc_state *context,
-		display_e2e_pipe_params_st *pipes,
-		int pipe_cnt,
-		int vlevel);
-void dcn31_update_soc_for_wm_a(struct dc *dc, struct dc_state *context);
-
-struct resource_pool *dcn31_create_resource_pool(
-		const struct dc_init_data *init_data,
-		struct dc *dc);
-
-#endif /* _DCN31_RESOURCE_H_ */
diff --git a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_vpg.c b/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_vpg.c
index f1deb1c3c363..cfb923d85630 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_vpg.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_vpg.c
@@ -63,7 +63,12 @@ void vpg31_poweron(struct vpg *vpg)
 {
 	struct dcn31_vpg *vpg31 = DCN31_VPG_FROM_VPG(vpg);
 
-	if (vpg->ctx->dc->debug.enable_mem_low_power.bits.vpg == false)
+	uint32_t vpg_gsp_mem_pwr_state;
+
+	REG_GET(VPG_MEM_PWR, VPG_GSP_MEM_PWR_STATE, &vpg_gsp_mem_pwr_state);
+
+	if (vpg->ctx->dc->debug.enable_mem_low_power.bits.vpg == false &&
+			vpg_gsp_mem_pwr_state == 0)
 		return;
 
 	REG_UPDATE_2(VPG_MEM_PWR, VPG_GSP_MEM_LIGHT_SLEEP_DIS, 1, VPG_GSP_LIGHT_SLEEP_FORCE, 0);
diff --git a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_vpg.h b/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_vpg.h
index 0e76eabce441..609e58dbd056 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_vpg.h
+++ b/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_vpg.h
@@ -26,6 +26,7 @@
 #ifndef __DAL_DCN31_VPG_H__
 #define __DAL_DCN31_VPG_H__
 
+#include "vpg.h"
 
 #define DCN31_VPG_FROM_VPG(vpg)\
 	container_of(vpg, struct dcn31_vpg, base)
diff --git a/drivers/gpu/drm/amd/display/dc/dio/Makefile b/drivers/gpu/drm/amd/display/dc/dio/Makefile
new file mode 100644
index 000000000000..0dfd480976f7
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dio/Makefile
@@ -0,0 +1,117 @@
+#
+# Copyright 2020 Advanced Micro Devices, Inc.
+#
+# Permission is hereby granted, free of charge, to any person obtaining a
+# copy of this software and associated documentation files (the "Software"),
+# to deal in the Software without restriction, including without limitation
+# the rights to use, copy, modify, merge, publish, distribute, sublicense,
+# and/or sell copies of the Software, and to permit persons to whom the
+# Software is furnished to do so, subject to the following conditions:
+#
+# The above copyright notice and this permission notice shall be included in
+# all copies or substantial portions of the Software.
+#
+# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+# THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+# OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+# ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+# OTHER DEALINGS IN THE SOFTWARE.
+#
+# Authors: AMD
+#
+#
+
+ifdef CONFIG_DRM_AMD_DC_FP
+###############################################################################
+# DCN10
+###############################################################################
+DIO_DCN10 = dcn10_link_encoder.o dcn10_stream_encoder.o
+
+AMD_DAL_DIO_DCN10 = $(addprefix $(AMDDALPATH)/dc/dio/dcn10/,$(DIO_DCN10))
+
+AMD_DISPLAY_FILES += $(AMD_DAL_DIO_DCN10)
+
+###############################################################################
+# DCN20
+###############################################################################
+DIO_DCN20 = dcn20_link_encoder.o dcn20_stream_encoder.o
+
+AMD_DAL_DIO_DCN20 = $(addprefix $(AMDDALPATH)/dc/dio/dcn20/,$(DIO_DCN20))
+
+AMD_DISPLAY_FILES += $(AMD_DAL_DIO_DCN20)
+
+###############################################################################
+# DCN30
+###############################################################################
+DIO_DCN30 = dcn30_dio_link_encoder.o dcn30_dio_stream_encoder.o
+
+AMD_DAL_DIO_DCN30 = $(addprefix $(AMDDALPATH)/dc/dio/dcn30/,$(DIO_DCN30))
+
+AMD_DISPLAY_FILES += $(AMD_DAL_DIO_DCN30)
+
+###############################################################################
+# DCN301
+###############################################################################
+DIO_DCN301 = dcn301_dio_link_encoder.o
+
+AMD_DAL_DIO_DCN301 = $(addprefix $(AMDDALPATH)/dc/dio/dcn301/,$(DIO_DCN301))
+
+AMD_DISPLAY_FILES += $(AMD_DAL_DIO_DCN301)
+
+###############################################################################
+# DCN31
+###############################################################################
+DIO_DCN31 = dcn31_dio_link_encoder.o
+
+AMD_DAL_DIO_DCN31 = $(addprefix $(AMDDALPATH)/dc/dio/dcn31/,$(DIO_DCN31))
+
+AMD_DISPLAY_FILES += $(AMD_DAL_DIO_DCN31)
+
+###############################################################################
+# DCN314
+###############################################################################
+DIO_DCN314 = dcn314_dio_stream_encoder.o
+
+AMD_DAL_DIO_DCN314 = $(addprefix $(AMDDALPATH)/dc/dio/dcn314/,$(DIO_DCN314))
+
+AMD_DISPLAY_FILES += $(AMD_DAL_DIO_DCN314)
+
+###############################################################################
+# DCN32
+###############################################################################
+DIO_DCN32 = dcn32_dio_link_encoder.o dcn32_dio_stream_encoder.o
+
+AMD_DAL_DIO_DCN32 = $(addprefix $(AMDDALPATH)/dc/dio/dcn32/,$(DIO_DCN32))
+
+AMD_DISPLAY_FILES += $(AMD_DAL_DIO_DCN32)
+
+###############################################################################
+# DCN35
+###############################################################################
+DIO_DCN35 = dcn35_dio_link_encoder.o dcn35_dio_stream_encoder.o
+
+AMD_DAL_DIO_DCN35 = $(addprefix $(AMDDALPATH)/dc/dio/dcn35/,$(DIO_DCN35))
+
+AMD_DISPLAY_FILES += $(AMD_DAL_DIO_DCN35)
+
+###############################################################################
+# DCN321
+###############################################################################
+DIO_DCN321 = dcn321_dio_link_encoder.o
+
+AMD_DAL_DIO_DCN321 = $(addprefix $(AMDDALPATH)/dc/dio/dcn321/,$(DIO_DCN321))
+
+AMD_DISPLAY_FILES += $(AMD_DAL_DIO_DCN321)
+
+
+###############################################################################
+# DCN401
+###############################################################################
+DIO_DCN401 = dcn401_dio_link_encoder.o dcn401_dio_stream_encoder.o
+
+AMD_DAL_DIO_DCN401 = $(addprefix $(AMDDALPATH)/dc/dio/dcn401/,$(DIO_DCN401))
+
+AMD_DISPLAY_FILES += $(AMD_DAL_DIO_DCN401)
+endif
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_link_encoder.c b/drivers/gpu/drm/amd/display/dc/dio/dcn10/dcn10_link_encoder.c
index 2dc4b4e4ba02..1c1228116487 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_link_encoder.c
+++ b/drivers/gpu/drm/amd/display/dc/dio/dcn10/dcn10_link_encoder.c
@@ -23,16 +23,12 @@
  *
  */
 
-#include <linux/delay.h>
-#include <linux/slab.h>
-
 #include "reg_helper.h"
 
 #include "core_types.h"
 #include "link_encoder.h"
 #include "dcn10_link_encoder.h"
 #include "stream_encoder.h"
-#include "i2caux_interface.h"
 #include "dc_bios_types.h"
 
 #include "gpio_service_interface.h"
@@ -646,8 +642,9 @@ static bool dcn10_link_encoder_validate_hdmi_output(
 			crtc_timing->pixel_encoding == PIXEL_ENCODING_YCBCR420)
 		return false;
 
-	if (!enc10->base.features.flags.bits.HDMI_6GB_EN &&
-		adjusted_pix_clk_100hz >= 3000000)
+	if ((!enc10->base.features.flags.bits.HDMI_6GB_EN ||
+			enc10->base.ctx->dc->debug.hdmi20_disable) &&
+			adjusted_pix_clk_100hz >= 3000000)
 		return false;
 	if (enc10->base.ctx->dc->debug.hdmi20_disable &&
 		crtc_timing->pixel_encoding == PIXEL_ENCODING_YCBCR420)
@@ -815,7 +812,7 @@ bool dcn10_link_encoder_validate_output_with_stream(
 					enc10, &stream->timing);
 	break;
 	case SIGNAL_TYPE_EDP:
-		is_valid = (stream->timing.pixel_encoding == PIXEL_ENCODING_RGB) ? true : false;
+		is_valid = stream->timing.pixel_encoding == PIXEL_ENCODING_RGB;
 	break;
 	case SIGNAL_TYPE_VIRTUAL:
 		is_valid = true;
@@ -1059,7 +1056,7 @@ void dcn10_link_encoder_disable_output(
 	struct bp_transmitter_control cntl = { 0 };
 	enum bp_result result;
 
-	if (!dcn10_is_dig_enabled(enc)) {
+	if (enc->funcs->is_dig_enabled && !enc->funcs->is_dig_enabled(enc)) {
 		/* OF_SKIP_POWER_DOWN_INACTIVE_ENCODER */
 	/*in DP_Alt_No_Connect case, we turn off the dig already,
 	after excuation the PHY w/a sequence, not allow touch PHY any more*/
@@ -1100,12 +1097,14 @@ void dcn10_link_encoder_disable_output(
 
 void dcn10_link_encoder_dp_set_lane_settings(
 	struct link_encoder *enc,
-	const struct link_training_settings *link_settings)
+	const struct dc_link_settings *link_settings,
+	const struct dc_lane_settings lane_settings[LANE_COUNT_DP_MAX])
 {
 	struct dcn10_link_encoder *enc10 = TO_DCN10_LINK_ENC(enc);
 	union dpcd_training_lane_set training_lane_set = { { 0 } };
 	int32_t lane = 0;
 	struct bp_transmitter_control cntl = { 0 };
+	enum bp_result result;
 
 	if (!link_settings) {
 		BREAK_TO_DEBUGGER();
@@ -1115,33 +1114,37 @@ void dcn10_link_encoder_dp_set_lane_settings(
 	cntl.action = TRANSMITTER_CONTROL_SET_VOLTAGE_AND_PREEMPASIS;
 	cntl.transmitter = enc10->base.transmitter;
 	cntl.connector_obj_id = enc10->base.connector;
-	cntl.lanes_number = link_settings->link_settings.lane_count;
+	cntl.lanes_number = link_settings->lane_count;
 	cntl.hpd_sel = enc10->base.hpd_source;
-	cntl.pixel_clock = link_settings->link_settings.link_rate *
-						LINK_RATE_REF_FREQ_IN_KHZ;
+	cntl.pixel_clock = link_settings->link_rate * LINK_RATE_REF_FREQ_IN_KHZ;
 
-	for (lane = 0; lane < link_settings->link_settings.lane_count; lane++) {
+	for (lane = 0; lane < link_settings->lane_count; lane++) {
 		/* translate lane settings */
 
 		training_lane_set.bits.VOLTAGE_SWING_SET =
-			link_settings->lane_settings[lane].VOLTAGE_SWING;
+				lane_settings[lane].VOLTAGE_SWING;
 		training_lane_set.bits.PRE_EMPHASIS_SET =
-			link_settings->lane_settings[lane].PRE_EMPHASIS;
+				lane_settings[lane].PRE_EMPHASIS;
 
 		/* post cursor 2 setting only applies to HBR2 link rate */
-		if (link_settings->link_settings.link_rate == LINK_RATE_HIGH2) {
+		if (link_settings->link_rate == LINK_RATE_HIGH2) {
 			/* this is passed to VBIOS
 			 * to program post cursor 2 level
 			 */
 			training_lane_set.bits.POST_CURSOR2_SET =
-				link_settings->lane_settings[lane].POST_CURSOR2;
+					lane_settings[lane].POST_CURSOR2;
 		}
 
 		cntl.lane_select = lane;
 		cntl.lane_settings = training_lane_set.raw;
 
 		/* call VBIOS table to set voltage swing and pre-emphasis */
-		link_transmitter_control(enc10, &cntl);
+		result = link_transmitter_control(enc10, &cntl);
+
+		if (result != BP_RESULT_OK) {
+			DC_LOG_ERROR("%s: Failed to execute VBIOS command table!\n", __func__);
+			BREAK_TO_DEBUGGER();
+		}
 	}
 }
 
@@ -1222,7 +1225,6 @@ void dcn10_link_encoder_update_mst_stream_allocation_table(
 	const struct link_mst_stream_allocation_table *table)
 {
 	struct dcn10_link_encoder *enc10 = TO_DCN10_LINK_ENC(enc);
-	uint32_t value0 = 0;
 	uint32_t value1 = 0;
 	uint32_t value2 = 0;
 	uint32_t slots = 0;
@@ -1324,7 +1326,7 @@ void dcn10_link_encoder_update_mst_stream_allocation_table(
 	do {
 		udelay(10);
 
-		value0 = REG_READ(DP_MSE_SAT_UPDATE);
+		REG_READ(DP_MSE_SAT_UPDATE);
 
 		REG_GET(DP_MSE_SAT_UPDATE,
 				DP_MSE_SAT_UPDATE, &value1);
@@ -1443,7 +1445,6 @@ enum signal_type dcn10_get_dig_mode(
 	default:
 		return SIGNAL_TYPE_NONE;
 	}
-	return SIGNAL_TYPE_NONE;
 }
 
 void dcn10_link_encoder_get_max_link_cap(struct link_encoder *enc,
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_link_encoder.h b/drivers/gpu/drm/amd/display/dc/dio/dcn10/dcn10_link_encoder.h
index c337588231ff..b7a89c39f445 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_link_encoder.h
+++ b/drivers/gpu/drm/amd/display/dc/dio/dcn10/dcn10_link_encoder.h
@@ -167,6 +167,8 @@ struct dcn10_link_enc_registers {
 	uint32_t DIO_LINKD_CNTL;
 	uint32_t DIO_LINKE_CNTL;
 	uint32_t DIO_LINKF_CNTL;
+	uint32_t DIO_CLK_CNTL;
+	uint32_t DIG_BE_CLK_CNTL;
 };
 
 #define LE_SF(reg_name, field_name, post_fix)\
@@ -472,11 +474,41 @@ struct dcn10_link_enc_registers {
 	type HPO_DP_ENC_SEL;\
 	type HPO_HDMI_ENC_SEL
 
+#define DCN35_LINK_ENCODER_REG_FIELD_LIST(type) \
+	type DIG_BE_ENABLE;\
+	type DIG_RB_SWITCH_EN;\
+	type DIG_BE_MODE;\
+	type DIG_BE_CLK_EN;\
+	type DIG_BE_SOFT_RESET;\
+	type HDCP_SOFT_RESET;\
+	type DIG_BE_SYMCLK_G_CLOCK_ON;\
+	type DIG_BE_SYMCLK_G_HDCP_CLOCK_ON;\
+	type DIG_BE_SYMCLK_G_TMDS_CLOCK_ON;\
+	type DISPCLK_R_GATE_DIS;\
+	type DISPCLK_G_GATE_DIS;\
+	type REFCLK_R_GATE_DIS;\
+	type REFCLK_G_GATE_DIS;\
+	type SOCCLK_G_GATE_DIS;\
+	type SYMCLK_FE_R_GATE_DIS;\
+	type SYMCLK_FE_G_GATE_DIS;\
+	type SYMCLK_R_GATE_DIS;\
+	type SYMCLK_G_GATE_DIS;\
+	type DIO_FGCG_REP_DIS;\
+	type DISPCLK_G_HDCP_GATE_DIS;\
+	type SYMCLKA_G_HDCP_GATE_DIS;\
+	type SYMCLKB_G_HDCP_GATE_DIS;\
+	type SYMCLKC_G_HDCP_GATE_DIS;\
+	type SYMCLKD_G_HDCP_GATE_DIS;\
+	type SYMCLKE_G_HDCP_GATE_DIS;\
+	type SYMCLKF_G_HDCP_GATE_DIS;\
+	type SYMCLKG_G_HDCP_GATE_DIS
+
 struct dcn10_link_enc_shift {
 	DCN_LINK_ENCODER_REG_FIELD_LIST(uint8_t);
 	DCN20_LINK_ENCODER_REG_FIELD_LIST(uint8_t);
 	DCN30_LINK_ENCODER_REG_FIELD_LIST(uint8_t);
 	DCN31_LINK_ENCODER_REG_FIELD_LIST(uint8_t);
+	DCN35_LINK_ENCODER_REG_FIELD_LIST(uint8_t);
 };
 
 struct dcn10_link_enc_mask {
@@ -484,6 +516,7 @@ struct dcn10_link_enc_mask {
 	DCN20_LINK_ENCODER_REG_FIELD_LIST(uint32_t);
 	DCN30_LINK_ENCODER_REG_FIELD_LIST(uint32_t);
 	DCN31_LINK_ENCODER_REG_FIELD_LIST(uint32_t);
+	DCN35_LINK_ENCODER_REG_FIELD_LIST(uint32_t);
 };
 
 struct dcn10_link_encoder {
@@ -581,7 +614,8 @@ void dcn10_link_encoder_disable_output(
 /* set DP lane settings */
 void dcn10_link_encoder_dp_set_lane_settings(
 	struct link_encoder *enc,
-	const struct link_training_settings *link_settings);
+	const struct dc_link_settings *link_settings,
+	const struct dc_lane_settings lane_settings[LANE_COUNT_DP_MAX]);
 
 void dcn10_link_encoder_dp_set_phy_pattern(
 	struct link_encoder *enc,
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_stream_encoder.c b/drivers/gpu/drm/amd/display/dc/dio/dcn10/dcn10_stream_encoder.c
index b0c08ee6bc2c..d928b4dcf6b8 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_stream_encoder.c
+++ b/drivers/gpu/drm/amd/display/dc/dio/dcn10/dcn10_stream_encoder.c
@@ -23,20 +23,18 @@
  *
  */
 
-#include <linux/delay.h>
-
+#include "dm_services.h"
 #include "dc_bios_types.h"
 #include "dcn10_stream_encoder.h"
 #include "reg_helper.h"
 #include "hw_shared.h"
-#include "inc/link_dpcd.h"
+#include "link_service.h"
 #include "dpcd_defs.h"
 #include "dcn30/dcn30_afmt.h"
 
 #define DC_LOGGER \
 		enc1->base.ctx->logger
 
-
 #define REG(reg)\
 	(enc1->regs->reg)
 
@@ -61,7 +59,6 @@ void enc1_update_generic_info_packet(
 	uint32_t packet_index,
 	const struct dc_info_packet *info_packet)
 {
-	uint32_t regval;
 	/* TODOFPGA Figure out a proper number for max_retries polling for lock
 	 * use 50 for now.
 	 */
@@ -90,7 +87,6 @@ void enc1_update_generic_info_packet(
 	REG_UPDATE(AFMT_VBI_PACKET_CONTROL, AFMT_GENERIC_CONFLICT_CLR, 1);
 
 	/* choose which generic packet to use */
-	regval = REG_READ(AFMT_VBI_PACKET_CONTROL);
 	REG_UPDATE(AFMT_VBI_PACKET_CONTROL,
 			AFMT_GENERIC_INDEX, packet_index);
 
@@ -259,10 +255,7 @@ void enc1_stream_encoder_dp_set_stream_attribute(
 	uint32_t misc1 = 0;
 	uint32_t h_blank;
 	uint32_t h_back_porch;
-	uint8_t synchronous_clock = 0; /* asynchronous mode */
 	uint8_t colorimetry_bpc;
-	uint8_t dynamic_range_rgb = 0; /*full range*/
-	uint8_t dynamic_range_ycbcr = 1; /*bt709*/
 	uint8_t dp_pixel_encoding = 0;
 	uint8_t dp_component_depth = 0;
 
@@ -368,24 +361,20 @@ void enc1_stream_encoder_dp_set_stream_attribute(
 		break;
 	}
 
-	misc0 = misc0 | synchronous_clock;
 	misc0 = colorimetry_bpc << 5;
 
 	switch (output_color_space) {
 	case COLOR_SPACE_SRGB:
 		misc1 = misc1 & ~0x80; /* bit7 = 0*/
-		dynamic_range_rgb = 0; /*full range*/
 		break;
 	case COLOR_SPACE_SRGB_LIMITED:
 		misc0 = misc0 | 0x8; /* bit3=1 */
 		misc1 = misc1 & ~0x80; /* bit7 = 0*/
-		dynamic_range_rgb = 1; /*limited range*/
 		break;
 	case COLOR_SPACE_YCBCR601:
 	case COLOR_SPACE_YCBCR601_LIMITED:
 		misc0 = misc0 | 0x8; /* bit3=1, bit4=0 */
 		misc1 = misc1 & ~0x80; /* bit7 = 0*/
-		dynamic_range_ycbcr = 0; /*bt601*/
 		if (hw_crtc_timing.pixel_encoding == PIXEL_ENCODING_YCBCR422)
 			misc0 = misc0 | 0x2; /* bit2=0, bit1=1 */
 		else if (hw_crtc_timing.pixel_encoding == PIXEL_ENCODING_YCBCR444)
@@ -395,17 +384,14 @@ void enc1_stream_encoder_dp_set_stream_attribute(
 	case COLOR_SPACE_YCBCR709_LIMITED:
 		misc0 = misc0 | 0x18; /* bit3=1, bit4=1 */
 		misc1 = misc1 & ~0x80; /* bit7 = 0*/
-		dynamic_range_ycbcr = 1; /*bt709*/
 		if (hw_crtc_timing.pixel_encoding == PIXEL_ENCODING_YCBCR422)
 			misc0 = misc0 | 0x2; /* bit2=0, bit1=1 */
 		else if (hw_crtc_timing.pixel_encoding == PIXEL_ENCODING_YCBCR444)
 			misc0 = misc0 | 0x4; /* bit2=1, bit1=0 */
 		break;
 	case COLOR_SPACE_2020_RGB_LIMITEDRANGE:
-		dynamic_range_rgb = 1; /*limited range*/
-		break;
 	case COLOR_SPACE_2020_RGB_FULLRANGE:
-	case COLOR_SPACE_2020_YCBCR:
+	case COLOR_SPACE_2020_YCBCR_LIMITED:
 	case COLOR_SPACE_XR_RGB:
 	case COLOR_SPACE_MSREF_SCRGB:
 	case COLOR_SPACE_ADOBERGB:
@@ -418,6 +404,7 @@ void enc1_stream_encoder_dp_set_stream_attribute(
 	case COLOR_SPACE_CUSTOMPOINTS:
 	case COLOR_SPACE_UNKNOWN:
 	case COLOR_SPACE_YCBCR709_BLACK:
+	default:
 		/* do nothing */
 		break;
 	}
@@ -596,6 +583,8 @@ void enc1_stream_encoder_hdmi_set_stream_attribute(
 		HDMI_GC_SEND, 1,
 		HDMI_NULL_SEND, 1);
 
+	REG_UPDATE(HDMI_VBI_PACKET_CONTROL, HDMI_ACP_SEND, 0);
+
 	/* following belongs to audio */
 	REG_UPDATE(HDMI_INFOFRAME_CONTROL0, HDMI_AUDIO_INFO_SEND, 1);
 
@@ -674,11 +663,13 @@ static void enc1_stream_encoder_update_hdmi_info_packets(
 	/* for bring up, disable dp double  TODO */
 	REG_UPDATE(HDMI_DB_CONTROL, HDMI_DB_DISABLE, 1);
 
+	/*Always add mandatory packets first followed by optional ones*/
 	enc1_update_hdmi_info_packet(enc1, 0, &info_frame->avi);
-	enc1_update_hdmi_info_packet(enc1, 1, &info_frame->vendor);
+	enc1_update_hdmi_info_packet(enc1, 1, &info_frame->hfvsif);
 	enc1_update_hdmi_info_packet(enc1, 2, &info_frame->gamut);
-	enc1_update_hdmi_info_packet(enc1, 3, &info_frame->spd);
-	enc1_update_hdmi_info_packet(enc1, 4, &info_frame->hdrsmd);
+	enc1_update_hdmi_info_packet(enc1, 3, &info_frame->vendor);
+	enc1_update_hdmi_info_packet(enc1, 4, &info_frame->spd);
+	enc1_update_hdmi_info_packet(enc1, 5, &info_frame->hdrsmd);
 }
 
 static void enc1_stream_encoder_stop_hdmi_info_packets(
@@ -761,12 +752,19 @@ void enc1_stream_encoder_update_dp_info_packets(
 	 * use other packetIndex (such as 5,6) for other info packet
 	 */
 
+	if (info_frame->adaptive_sync.valid)
+		enc1_update_generic_info_packet(
+				enc1,
+				5,  /* packetIndex */
+				&info_frame->adaptive_sync);
+
 	/* enable/disable transmission of packet(s).
 	 * If enabled, packet transmission begins on the next frame
 	 */
 	REG_UPDATE(DP_SEC_CNTL, DP_SEC_GSP0_ENABLE, info_frame->vsc.valid);
 	REG_UPDATE(DP_SEC_CNTL, DP_SEC_GSP2_ENABLE, info_frame->spd.valid);
 	REG_UPDATE(DP_SEC_CNTL, DP_SEC_GSP3_ENABLE, info_frame->hdrsmd.valid);
+	REG_UPDATE(DP_SEC_CNTL, DP_SEC_GSP5_ENABLE, info_frame->adaptive_sync.valid);
 
 	/* This bit is the master enable bit.
 	 * When enabling secondary stream engine,
@@ -934,7 +932,7 @@ void enc1_stream_encoder_dp_blank(
 	/* disable DP stream */
 	REG_UPDATE(DP_VID_STREAM_CNTL, DP_VID_STREAM_ENABLE, 0);
 
-	dp_source_sequence_trace(link, DPCD_SOURCE_SEQ_AFTER_DISABLE_DP_VID_STREAM);
+	link->dc->link_srv->dp_trace_source_sequence(link, DPCD_SOURCE_SEQ_AFTER_DISABLE_DP_VID_STREAM);
 
 	/* the encoder stops sending the video stream
 	 * at the start of the vertical blanking.
@@ -953,7 +951,7 @@ void enc1_stream_encoder_dp_blank(
 
 	REG_UPDATE(DP_STEER_FIFO, DP_STEER_FIFO_RESET, true);
 
-	dp_source_sequence_trace(link, DPCD_SOURCE_SEQ_AFTER_FIFO_STEER_RESET);
+	link->dc->link_srv->dp_trace_source_sequence(link, DPCD_SOURCE_SEQ_AFTER_FIFO_STEER_RESET);
 }
 
 /* output video stream to link encoder */
@@ -1026,7 +1024,8 @@ void enc1_stream_encoder_dp_unblank(
 
 	REG_UPDATE(DP_VID_STREAM_CNTL, DP_VID_STREAM_ENABLE, true);
 
-	dp_source_sequence_trace(link, DPCD_SOURCE_SEQ_AFTER_ENABLE_DP_VID_STREAM);
+	link->dc->link_srv->dp_trace_source_sequence(link,
+			DPCD_SOURCE_SEQ_AFTER_ENABLE_DP_VID_STREAM);
 }
 
 void enc1_stream_encoder_set_avmute(
@@ -1471,10 +1470,9 @@ void enc1_se_hdmi_audio_setup(
 void enc1_se_hdmi_audio_disable(
 	struct stream_encoder *enc)
 {
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 	if (enc->afmt && enc->afmt->funcs->afmt_powerdown)
 		enc->afmt->funcs->afmt_powerdown(enc->afmt);
-#endif
+
 	enc1_se_enable_audio_clock(enc, false);
 }
 
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_stream_encoder.h b/drivers/gpu/drm/amd/display/dc/dio/dcn10/dcn10_stream_encoder.h
index 687d7e4bf7ca..54a6a4ebd636 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_stream_encoder.h
+++ b/drivers/gpu/drm/amd/display/dc/dio/dcn10/dcn10_stream_encoder.h
@@ -73,6 +73,7 @@
 	SRI(HDMI_ACR_48_1, DIG, id),\
 	SRI(DP_DB_CNTL, DP, id), \
 	SRI(DP_MSA_MISC, DP, id), \
+	SRI(DP_MSA_VBID_MISC, DP, id), \
 	SRI(DP_MSA_COLORIMETRY, DP, id), \
 	SRI(DP_MSA_TIMING_PARAM1, DP, id), \
 	SRI(DP_MSA_TIMING_PARAM2, DP, id), \
@@ -92,6 +93,8 @@
 	SRI(DP_VID_STREAM_CNTL, DP, id), \
 	SRI(DP_VID_TIMING, DP, id), \
 	SRI(DP_SEC_AUD_N, DP, id), \
+	SRI(DP_SEC_AUD_N_READBACK, DP, id), \
+	SRI(DP_SEC_AUD_M_READBACK, DP, id), \
 	SRI(DP_SEC_TIMESTAMP, DP, id), \
 	SRI(DIG_CLOCK_PATTERN, DIG, id)
 
@@ -124,7 +127,6 @@ struct dcn10_stream_enc_registers {
 	uint32_t AFMT_60958_1;
 	uint32_t AFMT_60958_2;
 	uint32_t DIG_FE_CNTL;
-	uint32_t DIG_FE_CNTL2;
 	uint32_t DIG_FIFO_STATUS;
 	uint32_t DP_MSE_RATE_CNTL;
 	uint32_t DP_MSE_RATE_UPDATE;
@@ -140,6 +142,8 @@ struct dcn10_stream_enc_registers {
 	uint32_t DP_VID_STREAM_CNTL;
 	uint32_t DP_VID_TIMING;
 	uint32_t DP_SEC_AUD_N;
+	uint32_t DP_SEC_AUD_N_READBACK;
+	uint32_t DP_SEC_AUD_M_READBACK;
 	uint32_t DP_SEC_TIMESTAMP;
 	uint32_t HDMI_CONTROL;
 	uint32_t HDMI_GC;
@@ -182,13 +186,17 @@ struct dcn10_stream_enc_registers {
 	uint32_t HDMI_GENERIC_PACKET_CONTROL9;
 	uint32_t HDMI_GENERIC_PACKET_CONTROL10;
 	uint32_t DIG_CLOCK_PATTERN;
+	uint32_t DIG_FIFO_CTRL0;
+	uint32_t DIG_FE_CLK_CNTL;
+	uint32_t DIG_FE_EN_CNTL;
+	uint32_t STREAM_MAPPER_CONTROL;
 };
 
 
 #define SE_SF(reg_name, field_name, post_fix)\
 	.field_name = reg_name ## __ ## field_name ## post_fix
 
-#define SE_COMMON_MASK_SH_LIST_SOC_BASE(mask_sh)\
+#define SE_COMMON_MASK_SH_LIST_SOC(mask_sh)\
 	SE_SF(DIG0_AFMT_VBI_PACKET_CONTROL, AFMT_GENERIC_INDEX, mask_sh),\
 	SE_SF(DIG0_AFMT_GENERIC_HDR, AFMT_GENERIC_HB0, mask_sh),\
 	SE_SF(DIG0_AFMT_GENERIC_HDR, AFMT_GENERIC_HB1, mask_sh),\
@@ -205,6 +213,7 @@ struct dcn10_stream_enc_registers {
 	SE_SF(DIG0_HDMI_VBI_PACKET_CONTROL, HDMI_GC_CONT, mask_sh),\
 	SE_SF(DIG0_HDMI_VBI_PACKET_CONTROL, HDMI_GC_SEND, mask_sh),\
 	SE_SF(DIG0_HDMI_VBI_PACKET_CONTROL, HDMI_NULL_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_VBI_PACKET_CONTROL, HDMI_ACP_SEND, mask_sh),\
 	SE_SF(DIG0_HDMI_INFOFRAME_CONTROL0, HDMI_AUDIO_INFO_SEND, mask_sh),\
 	SE_SF(DIG0_AFMT_INFOFRAME_CONTROL0, AFMT_AUDIO_INFO_UPDATE, mask_sh),\
 	SE_SF(DIG0_HDMI_INFOFRAME_CONTROL1, HDMI_AUDIO_INFO_LINE, mask_sh),\
@@ -256,6 +265,8 @@ struct dcn10_stream_enc_registers {
 	SE_SF(DIG0_AFMT_60958_2, AFMT_60958_CS_CHANNEL_NUMBER_6, mask_sh),\
 	SE_SF(DIG0_AFMT_60958_2, AFMT_60958_CS_CHANNEL_NUMBER_7, mask_sh),\
 	SE_SF(DP0_DP_SEC_AUD_N, DP_SEC_AUD_N, mask_sh),\
+	SE_SF(DP0_DP_SEC_AUD_N_READBACK, DP_SEC_AUD_N_READBACK, mask_sh),\
+	SE_SF(DP0_DP_SEC_AUD_M_READBACK, DP_SEC_AUD_M_READBACK, mask_sh),\
 	SE_SF(DP0_DP_SEC_TIMESTAMP, DP_SEC_TIMESTAMP_MODE, mask_sh),\
 	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_ASP_ENABLE, mask_sh),\
 	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_ATP_ENABLE, mask_sh),\
@@ -331,9 +342,6 @@ struct dcn10_stream_enc_registers {
 	SE_SF(DIG0_DIG_FE_CNTL, DIG_SOURCE_SELECT, mask_sh),\
 	SE_SF(DIG0_DIG_CLOCK_PATTERN, DIG_CLOCK_PATTERN, mask_sh)
 
-#define SE_COMMON_MASK_SH_LIST_SOC(mask_sh)\
-	SE_COMMON_MASK_SH_LIST_SOC_BASE(mask_sh)
-
 #define SE_COMMON_MASK_SH_LIST_DCN10(mask_sh)\
 	SE_COMMON_MASK_SH_LIST_SOC(mask_sh),\
 	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC0_CONT, mask_sh),\
@@ -473,6 +481,8 @@ struct dcn10_stream_enc_registers {
 	type AFMT_60958_CS_CHANNEL_NUMBER_6;\
 	type AFMT_60958_CS_CHANNEL_NUMBER_7;\
 	type DP_SEC_AUD_N;\
+	type DP_SEC_AUD_N_READBACK;\
+	type DP_SEC_AUD_M_READBACK;\
 	type DP_SEC_TIMESTAMP_MODE;\
 	type DP_SEC_ASP_ENABLE;\
 	type DP_SEC_ATP_ENABLE;\
@@ -559,16 +569,53 @@ struct dcn10_stream_enc_registers {
 	type DP_SEC_GSP11_ENABLE;\
 	type DP_SEC_GSP11_LINE_NUM
 
+#define SE_REG_FIELD_LIST_DCN3_1_COMMON(type) \
+	type DIG_FIFO_OUTPUT_PIXEL_MODE;\
+	type DP_PIXEL_PER_CYCLE_PROCESSING_MODE;\
+	type DIG_SYMCLK_FE_ON;\
+	type DIG_FIFO_READ_START_LEVEL;\
+	type DIG_FIFO_ENABLE;\
+	type DIG_FIFO_RESET;\
+	type DIG_FIFO_RESET_DONE;\
+	type PIXEL_ENCODING_TYPE;\
+	type UNCOMPRESSED_PIXEL_FORMAT;\
+	type UNCOMPRESSED_COMPONENT_DEPTH
+
+#define SE_REG_FIELD_LIST_DCN3_5_COMMON(type) \
+	type DIG_FE_CLK_EN;\
+	type DIG_FE_MODE;\
+	type DIG_FE_SOFT_RESET;\
+	type DIG_FE_ENABLE;\
+	type DIG_FE_SYMCLK_FE_G_CLOCK_ON;\
+	type DIG_FE_DISPCLK_G_CLOCK_ON;\
+	type DIG_FE_SYMCLK_FE_G_AFMT_CLOCK_ON;\
+	type DIG_FE_SYMCLK_FE_G_TMDS_CLOCK_ON;\
+	type DIG_FE_SOCCLK_G_AFMT_CLOCK_ON;\
+	type DIG_STREAM_LINK_TARGET
+
+#define SE_REG_FIELD_LIST_DCN4_01_COMMON(type) \
+	type COMPRESSED_PIXEL_FORMAT;\
+	type DP_VID_N_INTERVAL;\
+	type DIG_FIFO_OUTPUT_PIXEL_PER_CYCLE;\
+	type DP_STEER_FIFO_ENABLE
 struct dcn10_stream_encoder_shift {
 	SE_REG_FIELD_LIST_DCN1_0(uint8_t);
+	uint8_t HDMI_ACP_SEND;
 	SE_REG_FIELD_LIST_DCN2_0(uint8_t);
 	SE_REG_FIELD_LIST_DCN3_0(uint8_t);
+	SE_REG_FIELD_LIST_DCN3_1_COMMON(uint8_t);
+	SE_REG_FIELD_LIST_DCN3_5_COMMON(uint8_t);
+	SE_REG_FIELD_LIST_DCN4_01_COMMON(uint32_t);
 };
 
 struct dcn10_stream_encoder_mask {
 	SE_REG_FIELD_LIST_DCN1_0(uint32_t);
+	uint32_t HDMI_ACP_SEND;
 	SE_REG_FIELD_LIST_DCN2_0(uint32_t);
 	SE_REG_FIELD_LIST_DCN3_0(uint32_t);
+	SE_REG_FIELD_LIST_DCN3_1_COMMON(uint32_t);
+	SE_REG_FIELD_LIST_DCN3_5_COMMON(uint32_t);
+	SE_REG_FIELD_LIST_DCN4_01_COMMON(uint32_t);
 };
 
 struct dcn10_stream_encoder {
diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_link_encoder.c b/drivers/gpu/drm/amd/display/dc/dio/dcn20/dcn20_link_encoder.c
index 2f9bfaeaba8d..51a57dae1811 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_link_encoder.c
+++ b/drivers/gpu/drm/amd/display/dc/dio/dcn20/dcn20_link_encoder.c
@@ -29,7 +29,6 @@
 #include "link_encoder.h"
 #include "dcn20_link_encoder.h"
 #include "stream_encoder.h"
-#include "i2caux_interface.h"
 #include "dc_bios_types.h"
 
 #include "gpio_service_interface.h"
diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_link_encoder.h b/drivers/gpu/drm/amd/display/dc/dio/dcn20/dcn20_link_encoder.h
index b2b266953d18..762c579fcb44 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_link_encoder.h
+++ b/drivers/gpu/drm/amd/display/dc/dio/dcn20/dcn20_link_encoder.h
@@ -147,7 +147,8 @@
 	LE_SF(DPCSTX0_DPCSTX_TX_CNTL, DPCS_TX_DATA_SWAP, mask_sh),\
 	LE_SF(DPCSTX0_DPCSTX_TX_CNTL, DPCS_TX_DATA_ORDER_INVERT, mask_sh),\
 	LE_SF(DPCSTX0_DPCSTX_TX_CNTL, DPCS_TX_FIFO_EN, mask_sh),\
-	LE_SF(DPCSTX0_DPCSTX_TX_CNTL, DPCS_TX_FIFO_RD_START_DELAY, mask_sh)
+	LE_SF(DPCSTX0_DPCSTX_TX_CNTL, DPCS_TX_FIFO_RD_START_DELAY, mask_sh),\
+	LE_SF(DPCSTX0_DPCSTX_DEBUG_CONFIG, DPCS_DBG_CBUS_DIS, mask_sh)
 
 #define DPCS_DCN2_MASK_SH_LIST(mask_sh)\
 	DPCS_MASK_SH_LIST(mask_sh),\
@@ -231,6 +232,8 @@
 	SRI(RDPCSTX_PHY_FUSE3, RDPCSTX, id), \
 	SRI(DPCSTX_TX_CLOCK_CNTL, DPCSTX, id), \
 	SRI(DPCSTX_TX_CNTL, DPCSTX, id), \
+	SRI(DPCSTX_DEBUG_CONFIG, DPCSTX, id), \
+	SRI(RDPCSTX_DEBUG_CONFIG, RDPCSTX, id), \
 	SR(RDPCSTX0_RDPCSTX_SCRATCH)
 
 
@@ -279,7 +282,6 @@ struct mpll_cfg {
 	uint32_t tx_peaking_lvl;
 	uint32_t ctr_reqs_pll;
 
-
 };
 
 struct dpcssys_phy_seq_cfg {
diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_stream_encoder.c b/drivers/gpu/drm/amd/display/dc/dio/dcn20/dcn20_stream_encoder.c
index aab25ca8343a..bec0b4aaeb2b 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_stream_encoder.c
+++ b/drivers/gpu/drm/amd/display/dc/dio/dcn20/dcn20_stream_encoder.c
@@ -29,13 +29,12 @@
 #include "dcn20_stream_encoder.h"
 #include "reg_helper.h"
 #include "hw_shared.h"
-#include "inc/link_dpcd.h"
+#include "link_service.h"
 #include "dpcd_defs.h"
 
 #define DC_LOGGER \
 		enc1->base.ctx->logger
 
-
 #define REG(reg)\
 	(enc1->regs->reg)
 
@@ -159,6 +158,7 @@ static void enc2_stream_encoder_update_hdmi_info_packets(
 	enc2_update_hdmi_info_packet(enc1, 3, &info_frame->vendor);
 	enc2_update_hdmi_info_packet(enc1, 4, &info_frame->spd);
 	enc2_update_hdmi_info_packet(enc1, 5, &info_frame->hdrsmd);
+	enc2_update_hdmi_info_packet(enc1, 6, &info_frame->vtem);
 }
 
 static void enc2_stream_encoder_stop_hdmi_info_packets(
@@ -423,6 +423,22 @@ void enc2_set_dynamic_metadata(struct stream_encoder *enc,
 	}
 }
 
+static void enc2_stream_encoder_update_dp_info_packets_sdp_line_num(
+		struct stream_encoder *enc,
+		struct encoder_info_frame *info_frame)
+{
+	struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc);
+
+	if (info_frame->adaptive_sync.valid == true &&
+		info_frame->sdp_line_num.adaptive_sync_line_num_valid == true) {
+		//00: REFER_TO_DP_SOF, 01: REFER_TO_OTG_SOF
+		REG_UPDATE(DP_SEC_CNTL1, DP_SEC_GSP5_LINE_REFERENCE, 1);
+
+		REG_UPDATE(DP_SEC_CNTL5, DP_SEC_GSP5_LINE_NUM,
+					info_frame->sdp_line_num.adaptive_sync_line_num);
+	}
+}
+
 static void enc2_stream_encoder_update_dp_info_packets(
 	struct stream_encoder *enc,
 	const struct encoder_info_frame *info_frame)
@@ -530,7 +546,8 @@ void enc2_stream_encoder_dp_unblank(
 
 	REG_UPDATE(DP_VID_STREAM_CNTL, DP_VID_STREAM_ENABLE, true);
 
-	dp_source_sequence_trace(link, DPCD_SOURCE_SEQ_AFTER_ENABLE_DP_VID_STREAM);
+	link->dc->link_srv->dp_trace_source_sequence(link,
+			DPCD_SOURCE_SEQ_AFTER_ENABLE_DP_VID_STREAM);
 }
 
 static void enc2_dp_set_odm_combine(
@@ -587,6 +604,8 @@ static const struct stream_encoder_funcs dcn20_str_enc_funcs = {
 		enc2_stream_encoder_update_hdmi_info_packets,
 	.stop_hdmi_info_packets =
 		enc2_stream_encoder_stop_hdmi_info_packets,
+	.update_dp_info_packets_sdp_line_num =
+		enc2_stream_encoder_update_dp_info_packets_sdp_line_num,
 	.update_dp_info_packets =
 		enc2_stream_encoder_update_dp_info_packets,
 	.send_immediate_sdp_message =
diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_stream_encoder.h b/drivers/gpu/drm/amd/display/dc/dio/dcn20/dcn20_stream_encoder.h
index baa1e539f341..baa1e539f341 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_stream_encoder.h
+++ b/drivers/gpu/drm/amd/display/dc/dio/dcn20/dcn20_stream_encoder.h
diff --git a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_dio_link_encoder.c b/drivers/gpu/drm/amd/display/dc/dio/dcn30/dcn30_dio_link_encoder.c
index 6f3c2fb60790..b8e31b5ea114 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_dio_link_encoder.c
+++ b/drivers/gpu/drm/amd/display/dc/dio/dcn30/dcn30_dio_link_encoder.c
@@ -29,10 +29,7 @@
 #include "link_encoder.h"
 #include "dcn30_dio_link_encoder.h"
 #include "stream_encoder.h"
-#include "i2caux_interface.h"
 #include "dc_bios_types.h"
-/* #include "dcn3ag/dcn3ag_phy_fw.h" */
-
 #include "gpio_service_interface.h"
 
 #define CTX \
diff --git a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_dio_link_encoder.h b/drivers/gpu/drm/amd/display/dc/dio/dcn30/dcn30_dio_link_encoder.h
index f2d90f2b8bf1..5b6177c2ae98 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_dio_link_encoder.h
+++ b/drivers/gpu/drm/amd/display/dc/dio/dcn30/dcn30_dio_link_encoder.h
@@ -55,7 +55,8 @@
 	SRI(DP_DPHY_HBR2_PATTERN_CONTROL, DP, id)
 
 #define LINK_ENCODER_MASK_SH_LIST_DCN30(mask_sh) \
-	LINK_ENCODER_MASK_SH_LIST_DCN20(mask_sh)
+	LINK_ENCODER_MASK_SH_LIST_DCN20(mask_sh),\
+	LE_SF(DIG0_TMDS_DCBALANCER_CONTROL, TMDS_SYNC_DCBAL_EN, mask_sh)
 
 #define DPCS_DCN3_MASK_SH_LIST(mask_sh)\
 	DPCS_DCN2_MASK_SH_LIST(mask_sh),\
diff --git a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_dio_stream_encoder.c b/drivers/gpu/drm/amd/display/dc/dio/dcn30/dcn30_dio_stream_encoder.c
index ebd9c35c914f..e93be7b6d9b0 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_dio_stream_encoder.c
+++ b/drivers/gpu/drm/amd/display/dc/dio/dcn30/dcn30_dio_stream_encoder.c
@@ -28,14 +28,11 @@
 #include "dcn30_dio_stream_encoder.h"
 #include "reg_helper.h"
 #include "hw_shared.h"
-#include "core_types.h"
-#include <linux/delay.h>
-
+#include "dc.h"
 
 #define DC_LOGGER \
 		enc1->base.ctx->logger
 
-
 #define REG(reg)\
 	(enc1->regs->reg)
 
@@ -50,23 +47,7 @@
 	enc1->base.ctx
 
 
-void convert_dc_info_packet_to_128(
-	const struct dc_info_packet *info_packet,
-	struct dc_info_packet_128 *info_packet_128)
-{
-	unsigned int i;
-
-	info_packet_128->hb0 = info_packet->hb0;
-	info_packet_128->hb1 = info_packet->hb1;
-	info_packet_128->hb2 = info_packet->hb2;
-	info_packet_128->hb3 = info_packet->hb3;
-
-	for (i = 0; i < 32; i++) {
-		info_packet_128->sb[i] = info_packet->sb[i];
-	}
-
-}
-static void enc3_update_hdmi_info_packet(
+void enc3_update_hdmi_info_packet(
 	struct dcn10_stream_encoder *enc1,
 	uint32_t packet_index,
 	const struct dc_info_packet *info_packet)
@@ -211,7 +192,7 @@ static void enc3_update_hdmi_info_packet(
 	}
 }
 
-static void enc3_stream_encoder_update_hdmi_info_packets(
+void enc3_stream_encoder_update_hdmi_info_packets(
 	struct stream_encoder *enc,
 	const struct encoder_info_frame *info_frame)
 {
@@ -228,9 +209,10 @@ static void enc3_stream_encoder_update_hdmi_info_packets(
 	enc3_update_hdmi_info_packet(enc1, 1, &info_frame->vendor);
 	enc3_update_hdmi_info_packet(enc1, 3, &info_frame->spd);
 	enc3_update_hdmi_info_packet(enc1, 4, &info_frame->hdrsmd);
+	enc3_update_hdmi_info_packet(enc1, 6, &info_frame->vtem);
 }
 
-static void enc3_stream_encoder_stop_hdmi_info_packets(
+void enc3_stream_encoder_stop_hdmi_info_packets(
 	struct stream_encoder *enc)
 {
 	struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc);
@@ -334,7 +316,7 @@ static void enc3_dp_set_dsc_config(struct stream_encoder *enc,
 }
 
 
-static void enc3_dp_set_dsc_pps_info_packet(struct stream_encoder *enc,
+void enc3_dp_set_dsc_pps_info_packet(struct stream_encoder *enc,
 					bool enable,
 					uint8_t *dsc_packed_pps,
 					bool immediate_update)
@@ -420,7 +402,23 @@ static void enc3_read_state(struct stream_encoder *enc, struct enc_state *s)
 	}
 }
 
-static void enc3_stream_encoder_update_dp_info_packets(
+void enc3_stream_encoder_update_dp_info_packets_sdp_line_num(
+		struct stream_encoder *enc,
+		struct encoder_info_frame *info_frame)
+{
+	struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc);
+
+	if (info_frame->adaptive_sync.valid == true &&
+		info_frame->sdp_line_num.adaptive_sync_line_num_valid == true) {
+		//00: REFER_TO_DP_SOF, 01: REFER_TO_OTG_SOF
+		REG_UPDATE(DP_SEC_CNTL1, DP_SEC_GSP5_LINE_REFERENCE, 1);
+
+		REG_UPDATE(DP_SEC_CNTL5, DP_SEC_GSP5_LINE_NUM,
+					info_frame->sdp_line_num.adaptive_sync_line_num);
+	}
+}
+
+void enc3_stream_encoder_update_dp_info_packets(
 	struct stream_encoder *enc,
 	const struct encoder_info_frame *info_frame)
 {
@@ -435,6 +433,36 @@ static void enc3_stream_encoder_update_dp_info_packets(
 				&info_frame->vsc,
 				true);
 	}
+	/* TODO: VSC SDP at packetIndex 1 should be retricted only if PSR-SU on.
+	 * There should have another Infopacket type (e.g. vsc_psrsu) for PSR_SU.
+	 * In addition, currently the driver check the valid bit then update and
+	 * send the corresponding Infopacket. For PSR-SU, the SDP only be sent
+	 * while entering PSR-SU mode. So we need another parameter(e.g. send)
+	 * in dc_info_packet to indicate which infopacket should be enabled by
+	 * default here.
+	 */
+	if (info_frame->vsc.valid) {
+		enc->vpg->funcs->update_generic_info_packet(
+				enc->vpg,
+				1,  /* packetIndex */
+				&info_frame->vsc,
+				true);
+	}
+	/* TODO: VSC SDP at packetIndex 1 should be restricted only if PSR-SU on.
+	 * There should have another Infopacket type (e.g. vsc_psrsu) for PSR_SU.
+	 * In addition, currently the driver check the valid bit then update and
+	 * send the corresponding Infopacket. For PSR-SU, the SDP only be sent
+	 * while entering PSR-SU mode. So we need another parameter(e.g. send)
+	 * in dc_info_packet to indicate which infopacket should be enabled by
+	 * default here.
+	 */
+	if (info_frame->vsc.valid) {
+		enc->vpg->funcs->update_generic_info_packet(
+				enc->vpg,
+				1,  /* packetIndex */
+				&info_frame->vsc,
+				true);
+	}
 	if (info_frame->spd.valid) {
 		enc->vpg->funcs->update_generic_info_packet(
 				enc->vpg,
@@ -453,12 +481,20 @@ static void enc3_stream_encoder_update_dp_info_packets(
 	 * use other packetIndex (such as 5,6) for other info packet
 	 */
 
+	if (info_frame->adaptive_sync.valid)
+		enc->vpg->funcs->update_generic_info_packet(
+				enc->vpg,
+				5,  /* packetIndex */
+				&info_frame->adaptive_sync,
+				true);
+
 	/* enable/disable transmission of packet(s).
 	 * If enabled, packet transmission begins on the next frame
 	 */
 	REG_UPDATE(DP_SEC_CNTL, DP_SEC_GSP0_ENABLE, info_frame->vsc.valid);
 	REG_UPDATE(DP_SEC_CNTL, DP_SEC_GSP2_ENABLE, info_frame->spd.valid);
 	REG_UPDATE(DP_SEC_CNTL, DP_SEC_GSP3_ENABLE, info_frame->hdrsmd.valid);
+	REG_UPDATE(DP_SEC_CNTL, DP_SEC_GSP5_ENABLE, info_frame->adaptive_sync.valid);
 
 	/* This bit is the master enable bit.
 	 * When enabling secondary stream engine,
@@ -489,7 +525,7 @@ static void enc3_dp_set_odm_combine(
 }
 
 /* setup stream encoder in dvi mode */
-void enc3_stream_encoder_dvi_set_stream_attribute(
+static void enc3_stream_encoder_dvi_set_stream_attribute(
 	struct stream_encoder *enc,
 	struct dc_crtc_timing *crtc_timing,
 	bool is_dual_link)
@@ -652,6 +688,9 @@ static void enc3_stream_encoder_hdmi_set_stream_attribute(
 		HDMI_GC_SEND, 1,
 		HDMI_NULL_SEND, 1);
 
+	/* Disable Audio Content Protection packet transmission */
+	REG_UPDATE(HDMI_VBI_PACKET_CONTROL, HDMI_ACP_SEND, 0);
+
 	/* following belongs to audio */
 	/* Enable Audio InfoFrame packet transmission. */
 	REG_UPDATE(HDMI_INFOFRAME_CONTROL0, HDMI_AUDIO_INFO_SEND, 1);
@@ -668,7 +707,7 @@ static void enc3_stream_encoder_hdmi_set_stream_attribute(
 	REG_UPDATE(HDMI_GC, HDMI_GC_AVMUTE, 0);
 }
 
-static void enc3_audio_mute_control(
+void enc3_audio_mute_control(
 	struct stream_encoder *enc,
 	bool mute)
 {
@@ -676,7 +715,7 @@ static void enc3_audio_mute_control(
 	enc->afmt->funcs->audio_mute_control(enc->afmt, mute);
 }
 
-static void enc3_se_dp_audio_setup(
+void enc3_se_dp_audio_setup(
 	struct stream_encoder *enc,
 	unsigned int az_inst,
 	struct audio_info *info)
@@ -707,7 +746,7 @@ static void enc3_se_setup_dp_audio(
 	enc->afmt->funcs->setup_dp_audio(enc->afmt);
 }
 
-static void enc3_se_dp_audio_enable(
+void enc3_se_dp_audio_enable(
 	struct stream_encoder *enc)
 {
 	enc1_se_enable_audio_clock(enc, true);
@@ -773,7 +812,7 @@ static void enc3_se_setup_hdmi_audio(
 	 */
 }
 
-static void enc3_se_hdmi_audio_setup(
+void enc3_se_hdmi_audio_setup(
 	struct stream_encoder *enc,
 	unsigned int az_inst,
 	struct audio_info *info,
@@ -801,6 +840,8 @@ static const struct stream_encoder_funcs dcn30_str_enc_funcs = {
 		enc3_stream_encoder_update_hdmi_info_packets,
 	.stop_hdmi_info_packets =
 		enc3_stream_encoder_stop_hdmi_info_packets,
+	.update_dp_info_packets_sdp_line_num =
+		enc3_stream_encoder_update_dp_info_packets_sdp_line_num,
 	.update_dp_info_packets =
 		enc3_stream_encoder_update_dp_info_packets,
 	.stop_dp_info_packets =
diff --git a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_dio_stream_encoder.h b/drivers/gpu/drm/amd/display/dc/dio/dcn30/dcn30_dio_stream_encoder.h
index e2c264ecb20f..830ce7e47035 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_dio_stream_encoder.h
+++ b/drivers/gpu/drm/amd/display/dc/dio/dcn30/dcn30_dio_stream_encoder.h
@@ -95,6 +95,8 @@
 	SRI(DP_VID_STREAM_CNTL, DP, id), \
 	SRI(DP_VID_TIMING, DP, id), \
 	SRI(DP_SEC_AUD_N, DP, id), \
+	SRI(DP_SEC_AUD_N_READBACK, DP, id), \
+	SRI(DP_SEC_AUD_M_READBACK, DP, id), \
 	SRI(DP_SEC_TIMESTAMP, DP, id), \
 	SRI(DP_DSC_CNTL, DP, id), \
 	SRI(DP_DSC_BYTES_PER_PIXEL, DP, id), \
@@ -110,7 +112,7 @@
 	SRI(DIG_CLOCK_PATTERN, DIG, id)
 
 
-#define SE_COMMON_MASK_SH_LIST_DCN30_BASE(mask_sh)\
+#define SE_COMMON_MASK_SH_LIST_DCN30(mask_sh)\
 	SE_SF(DP0_DP_PIXEL_FORMAT, DP_PIXEL_ENCODING, mask_sh),\
 	SE_SF(DP0_DP_PIXEL_FORMAT, DP_COMPONENT_DEPTH, mask_sh),\
 	SE_SF(DIG0_HDMI_CONTROL, HDMI_PACKET_GEN_VERSION, mask_sh),\
@@ -122,6 +124,7 @@
 	SE_SF(DIG0_HDMI_VBI_PACKET_CONTROL, HDMI_GC_CONT, mask_sh),\
 	SE_SF(DIG0_HDMI_VBI_PACKET_CONTROL, HDMI_GC_SEND, mask_sh),\
 	SE_SF(DIG0_HDMI_VBI_PACKET_CONTROL, HDMI_NULL_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_VBI_PACKET_CONTROL, HDMI_ACP_SEND, mask_sh),\
 	SE_SF(DIG0_HDMI_INFOFRAME_CONTROL0, HDMI_AUDIO_INFO_SEND, mask_sh),\
 	SE_SF(DIG0_HDMI_INFOFRAME_CONTROL1, HDMI_AUDIO_INFO_LINE, mask_sh),\
 	SE_SF(DIG0_HDMI_GC, HDMI_GC_AVMUTE, mask_sh),\
@@ -157,6 +160,8 @@
 	SE_SF(DIG0_HDMI_ACR_48_0, HDMI_ACR_CTS_48, mask_sh),\
 	SE_SF(DIG0_HDMI_ACR_48_1, HDMI_ACR_N_48, mask_sh),\
 	SE_SF(DP0_DP_SEC_AUD_N, DP_SEC_AUD_N, mask_sh),\
+	SE_SF(DP0_DP_SEC_AUD_N_READBACK, DP_SEC_AUD_N_READBACK, mask_sh),\
+	SE_SF(DP0_DP_SEC_AUD_M_READBACK, DP_SEC_AUD_M_READBACK, mask_sh),\
 	SE_SF(DP0_DP_SEC_TIMESTAMP, DP_SEC_TIMESTAMP_MODE, mask_sh),\
 	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_ASP_ENABLE, mask_sh),\
 	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_ATP_ENABLE, mask_sh),\
@@ -269,9 +274,6 @@
 	SE_SF(DP0_DP_SEC_FRAMING4, DP_SST_SDP_SPLITTING, mask_sh),\
 	SE_SF(DIG0_DIG_CLOCK_PATTERN, DIG_CLOCK_PATTERN, mask_sh)
 
-#define SE_COMMON_MASK_SH_LIST_DCN30(mask_sh)\
-	SE_COMMON_MASK_SH_LIST_DCN30_BASE(mask_sh)
-
 void dcn30_dio_stream_encoder_construct(
 	struct dcn10_stream_encoder *enc1,
 	struct dc_context *ctx,
@@ -283,4 +285,47 @@ void dcn30_dio_stream_encoder_construct(
 	const struct dcn10_stream_encoder_shift *se_shift,
 	const struct dcn10_stream_encoder_mask *se_mask);
 
+void enc3_stream_encoder_update_hdmi_info_packets(
+	struct stream_encoder *enc,
+	const struct encoder_info_frame *info_frame);
+
+void enc3_stream_encoder_stop_hdmi_info_packets(
+	struct stream_encoder *enc);
+
+void enc3_stream_encoder_update_dp_info_packets_sdp_line_num(
+		struct stream_encoder *enc,
+		struct encoder_info_frame *info_frame);
+
+void enc3_stream_encoder_update_dp_info_packets(
+	struct stream_encoder *enc,
+	const struct encoder_info_frame *info_frame);
+
+void enc3_audio_mute_control(
+	struct stream_encoder *enc,
+	bool mute);
+
+void enc3_se_dp_audio_setup(
+	struct stream_encoder *enc,
+	unsigned int az_inst,
+	struct audio_info *info);
+
+void enc3_se_dp_audio_enable(
+	struct stream_encoder *enc);
+
+void enc3_se_hdmi_audio_setup(
+	struct stream_encoder *enc,
+	unsigned int az_inst,
+	struct audio_info *info,
+	struct audio_crtc_info *audio_crtc_info);
+
+void enc3_dp_set_dsc_pps_info_packet(
+	struct stream_encoder *enc,
+	bool enable,
+	uint8_t *dsc_packed_pps,
+	bool immediate_update);
+void enc3_update_hdmi_info_packet(
+	struct dcn10_stream_encoder *enc1,
+	uint32_t packet_index,
+	const struct dc_info_packet *info_packet);
+
 #endif /* __DC_DIO_STREAM_ENCODER_DCN30_H__ */
diff --git a/drivers/gpu/drm/amd/display/dc/dcn301/dcn301_dio_link_encoder.c b/drivers/gpu/drm/amd/display/dc/dio/dcn301/dcn301_dio_link_encoder.c
index c9fbaed23965..1b39a6e8a1ac 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn301/dcn301_dio_link_encoder.c
+++ b/drivers/gpu/drm/amd/display/dc/dio/dcn301/dcn301_dio_link_encoder.c
@@ -29,7 +29,6 @@
 #include "link_encoder.h"
 #include "dcn301_dio_link_encoder.h"
 #include "stream_encoder.h"
-#include "i2caux_interface.h"
 #include "dc_bios_types.h"
 #include "gpio_service_interface.h"
 
diff --git a/drivers/gpu/drm/amd/display/dc/dcn301/dcn301_dio_link_encoder.h b/drivers/gpu/drm/amd/display/dc/dio/dcn301/dcn301_dio_link_encoder.h
index 49f8d91d4951..49f8d91d4951 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn301/dcn301_dio_link_encoder.h
+++ b/drivers/gpu/drm/amd/display/dc/dio/dcn301/dcn301_dio_link_encoder.h
diff --git a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_dio_link_encoder.c b/drivers/gpu/drm/amd/display/dc/dio/dcn31/dcn31_dio_link_encoder.c
index ee6f13bef377..84cc2ddc52fe 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_dio_link_encoder.c
+++ b/drivers/gpu/drm/amd/display/dc/dio/dcn31/dcn31_dio_link_encoder.c
@@ -30,7 +30,6 @@
 #include "link_encoder.h"
 #include "dcn31_dio_link_encoder.h"
 #include "stream_encoder.h"
-#include "i2caux_interface.h"
 #include "dc_bios_types.h"
 
 #include "gpio_service_interface.h"
@@ -38,6 +37,7 @@
 #include "link_enc_cfg.h"
 #include "dc_dmub_srv.h"
 #include "dal_asic_id.h"
+#include "link_service.h"
 
 #define CTX \
 	enc10->base.ctx
@@ -67,6 +67,70 @@
 #define MIN(X, Y) ((X) < (Y) ? (X) : (Y))
 #endif
 
+static uint8_t phy_id_from_transmitter(enum transmitter t)
+{
+	uint8_t phy_id;
+
+	switch (t) {
+	case TRANSMITTER_UNIPHY_A:
+		phy_id = 0;
+		break;
+	case TRANSMITTER_UNIPHY_B:
+		phy_id = 1;
+		break;
+	case TRANSMITTER_UNIPHY_C:
+		phy_id = 2;
+		break;
+	case TRANSMITTER_UNIPHY_D:
+		phy_id = 3;
+		break;
+	case TRANSMITTER_UNIPHY_E:
+		phy_id = 4;
+		break;
+	case TRANSMITTER_UNIPHY_F:
+		phy_id = 5;
+		break;
+	case TRANSMITTER_UNIPHY_G:
+		phy_id = 6;
+		break;
+	default:
+		phy_id = 0;
+		break;
+	}
+	return phy_id;
+}
+
+static bool has_query_dp_alt(struct link_encoder *enc)
+{
+	struct dc_dmub_srv *dc_dmub_srv = enc->ctx->dmub_srv;
+
+	if (enc->ctx->dce_version >= DCN_VERSION_3_15)
+		return true;
+
+	/* Supports development firmware and firmware >= 4.0.11 */
+	return dc_dmub_srv &&
+	       !(dc_dmub_srv->dmub->fw_version >= DMUB_FW_VERSION(4, 0, 0) &&
+		 dc_dmub_srv->dmub->fw_version <= DMUB_FW_VERSION(4, 0, 10));
+}
+
+static bool query_dp_alt_from_dmub(struct link_encoder *enc,
+				   union dmub_rb_cmd *cmd)
+{
+	struct dcn10_link_encoder *enc10 = TO_DCN10_LINK_ENC(enc);
+
+	memset(cmd, 0, sizeof(*cmd));
+	cmd->query_dp_alt.header.type = DMUB_CMD__VBIOS;
+	cmd->query_dp_alt.header.sub_type =
+		DMUB_CMD__VBIOS_TRANSMITTER_QUERY_DP_ALT;
+	cmd->query_dp_alt.header.payload_bytes = sizeof(cmd->query_dp_alt.data);
+	cmd->query_dp_alt.data.phy_id = phy_id_from_transmitter(enc10->base.transmitter);
+
+	if (!dc_wake_and_execute_dmub_cmd(enc->ctx, cmd, DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY))
+		return false;
+
+	return true;
+}
+
 void dcn31_link_encoder_set_dio_phy_mux(
 	struct link_encoder *enc,
 	enum encoder_type_select sel,
@@ -168,9 +232,7 @@ void enc31_hw_init(struct link_encoder *enc)
 	AUX_RX_PHASE_DETECT_LEN,  [21,20] = 0x3 default is 3
 	AUX_RX_DETECTION_THRESHOLD [30:28] = 1
 */
-	AUX_REG_WRITE(AUX_DPHY_RX_CONTROL0, 0x103d1110);
-
-	AUX_REG_WRITE(AUX_DPHY_TX_CONTROL, 0x21c7a);
+	// dmub will read AUX_DPHY_RX_CONTROL0/AUX_DPHY_TX_CONTROL from vbios table in dp_aux_init
 
 	//AUX_DPHY_TX_REF_CONTROL'AUX_TX_REF_DIV HW default is 0x32;
 	// Set AUX_TX_REF_DIV Divider to generate 2 MHz reference from refclk
@@ -178,18 +240,9 @@ void enc31_hw_init(struct link_encoder *enc)
 	// 100MHz -> 0x32
 	// 48MHz -> 0x18
 
-#ifdef CLEANUP_FIXME
-	/*from display_init*/
-	REG_WRITE(RDPCSTX_DEBUG_CONFIG, 0);
-#endif
-
 	// Set TMDS_CTL0 to 1.  This is a legacy setting.
 	REG_UPDATE(TMDS_CTL_BITS, TMDS_CTL0, 1);
 
-	/*HW default is 5*/
-	REG_UPDATE(RDPCSTX_CNTL,
-			RDPCS_TX_FIFO_RD_START_DELAY, 4);
-
 	dcn10_aux_initialize(enc10);
 }
 
@@ -371,7 +424,6 @@ static bool link_dpia_control(struct dc_context *dc_ctx,
 	struct dmub_cmd_dig_dpia_control_data *dpia_control)
 {
 	union dmub_rb_cmd cmd;
-	struct dc_dmub_srv *dmub = dc_ctx->dmub_srv;
 
 	memset(&cmd, 0, sizeof(cmd));
 
@@ -384,9 +436,7 @@ static bool link_dpia_control(struct dc_context *dc_ctx,
 
 	cmd.dig1_dpia_control.dpia_control = *dpia_control;
 
-	dc_dmub_srv_cmd_queue(dmub, &cmd);
-	dc_dmub_srv_cmd_execute(dmub);
-	dc_dmub_srv_wait_idle(dmub);
+	dc_wake_and_execute_dmub_cmd(dc_ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
 
 	return true;
 }
@@ -407,6 +457,7 @@ void dcn31_link_encoder_enable_dp_output(
 	/* Enable transmitter and encoder. */
 	if (!link_enc_cfg_is_transmitter_mappable(enc->ctx->dc, enc)) {
 
+		DC_LOG_DEBUG("%s: enc_id(%d)\n", __func__, enc->preferred_engine);
 		dcn20_link_encoder_enable_dp_output(enc, link_settings, clock_source);
 
 	} else {
@@ -431,13 +482,14 @@ void dcn31_link_encoder_enable_dp_output(
 
 		if (link) {
 			dpia_control.dpia_id = link->ddc_hw_inst;
-			dpia_control.fec_rdy = dc_link_should_enable_fec(link);
+			dpia_control.fec_rdy = link->dc->link_srv->dp_should_enable_fec(link);
 		} else {
 			DC_LOG_ERROR("%s: Failed to execute DPIA enable DMUB command.\n", __func__);
 			BREAK_TO_DEBUGGER();
 			return;
 		}
 
+		DC_LOG_DEBUG("%s: DPIA(%d) - enc_id(%d)\n", __func__, dpia_control.dpia_id, dpia_control.enc_id);
 		link_dpia_control(enc->ctx, &dpia_control);
 	}
 }
@@ -452,6 +504,7 @@ void dcn31_link_encoder_enable_dp_mst_output(
 	/* Enable transmitter and encoder. */
 	if (!link_enc_cfg_is_transmitter_mappable(enc->ctx->dc, enc)) {
 
+		DC_LOG_DEBUG("%s: enc_id(%d)\n", __func__, enc->preferred_engine);
 		dcn10_link_encoder_enable_dp_mst_output(enc, link_settings, clock_source);
 
 	} else {
@@ -476,13 +529,14 @@ void dcn31_link_encoder_enable_dp_mst_output(
 
 		if (link) {
 			dpia_control.dpia_id = link->ddc_hw_inst;
-			dpia_control.fec_rdy = dc_link_should_enable_fec(link);
+			dpia_control.fec_rdy = link->dc->link_srv->dp_should_enable_fec(link);
 		} else {
 			DC_LOG_ERROR("%s: Failed to execute DPIA enable DMUB command.\n", __func__);
 			BREAK_TO_DEBUGGER();
 			return;
 		}
 
+		DC_LOG_DEBUG("%s: DPIA(%d) - enc_id(%d)\n", __func__, dpia_control.dpia_id, dpia_control.enc_id);
 		link_dpia_control(enc->ctx, &dpia_control);
 	}
 }
@@ -496,6 +550,7 @@ void dcn31_link_encoder_disable_output(
 	/* Disable transmitter and encoder. */
 	if (!link_enc_cfg_is_transmitter_mappable(enc->ctx->dc, enc)) {
 
+		DC_LOG_DEBUG("%s: enc_id(%d)\n", __func__, enc->preferred_engine);
 		dcn10_link_encoder_disable_output(enc, signal);
 
 	} else {
@@ -503,7 +558,7 @@ void dcn31_link_encoder_disable_output(
 		struct dmub_cmd_dig_dpia_control_data dpia_control = { 0 };
 		struct dc_link *link;
 
-		if (!dcn10_is_dig_enabled(enc))
+		if (enc->funcs->is_dig_enabled && !enc->funcs->is_dig_enabled(enc))
 			return;
 
 		link = link_enc_cfg_get_link_using_link_enc(enc->ctx->dc, enc->preferred_engine);
@@ -527,6 +582,7 @@ void dcn31_link_encoder_disable_output(
 			return;
 		}
 
+		DC_LOG_DEBUG("%s: DPIA(%d) - enc_id(%d)\n", __func__, dpia_control.dpia_id, dpia_control.enc_id);
 		link_dpia_control(enc->ctx, &dpia_control);
 
 		link_encoder_disable(enc10);
@@ -536,57 +592,91 @@ void dcn31_link_encoder_disable_output(
 bool dcn31_link_encoder_is_in_alt_mode(struct link_encoder *enc)
 {
 	struct dcn10_link_encoder *enc10 = TO_DCN10_LINK_ENC(enc);
+	union dmub_rb_cmd cmd;
 	uint32_t dp_alt_mode_disable;
-	bool is_usb_c_alt_mode = false;
 
-	if (enc->features.flags.bits.DP_IS_USB_C) {
-		if (enc->ctx->asic_id.hw_internal_rev != YELLOW_CARP_B0) {
-			// [Note] no need to check hw_internal_rev once phy mux selection is ready
-			REG_GET(RDPCSTX_PHY_CNTL6, RDPCS_PHY_DPALT_DISABLE, &dp_alt_mode_disable);
-		} else {
+	/* Only applicable to USB-C PHY. */
+	if (!enc->features.flags.bits.DP_IS_USB_C)
+		return false;
+
+	/*
+	 * Use the new interface from DMCUB if available.
+	 * Avoids hanging the RDCPSPIPE if DMCUB wasn't already running.
+	 */
+	if (has_query_dp_alt(enc)) {
+		if (!query_dp_alt_from_dmub(enc, &cmd))
+			return false;
+
+		return (cmd.query_dp_alt.data.is_dp_alt_disable == 0);
+	}
+
+	/* Legacy path, avoid if possible. */
+	if (enc->ctx->asic_id.hw_internal_rev != YELLOW_CARP_B0) {
+		REG_GET(RDPCSTX_PHY_CNTL6, RDPCS_PHY_DPALT_DISABLE,
+			&dp_alt_mode_disable);
+	} else {
 		/*
 		 * B0 phys use a new set of registers to check whether alt mode is disabled.
 		 * if value == 1 alt mode is disabled, otherwise it is enabled.
 		 */
-			if ((enc10->base.transmitter == TRANSMITTER_UNIPHY_A)
-					|| (enc10->base.transmitter == TRANSMITTER_UNIPHY_B)
-					|| (enc10->base.transmitter == TRANSMITTER_UNIPHY_E)) {
-				REG_GET(RDPCSTX_PHY_CNTL6, RDPCS_PHY_DPALT_DISABLE, &dp_alt_mode_disable);
-			} else {
-			// [Note] need to change TRANSMITTER_UNIPHY_C/D to F/G once phy mux selection is ready
-				REG_GET(RDPCSPIPE_PHY_CNTL6, RDPCS_PHY_DPALT_DISABLE, &dp_alt_mode_disable);
-			}
+		if ((enc10->base.transmitter == TRANSMITTER_UNIPHY_A) ||
+		    (enc10->base.transmitter == TRANSMITTER_UNIPHY_B) ||
+		    (enc10->base.transmitter == TRANSMITTER_UNIPHY_E)) {
+			REG_GET(RDPCSTX_PHY_CNTL6, RDPCS_PHY_DPALT_DISABLE,
+				&dp_alt_mode_disable);
+		} else {
+			REG_GET(RDPCSPIPE_PHY_CNTL6, RDPCS_PHY_DPALT_DISABLE,
+				&dp_alt_mode_disable);
 		}
-
-		is_usb_c_alt_mode = (dp_alt_mode_disable == 0);
 	}
 
-	return is_usb_c_alt_mode;
+	return (dp_alt_mode_disable == 0);
 }
 
-void dcn31_link_encoder_get_max_link_cap(struct link_encoder *enc,
-										 struct dc_link_settings *link_settings)
+void dcn31_link_encoder_get_max_link_cap(struct link_encoder *enc, struct dc_link_settings *link_settings)
 {
 	struct dcn10_link_encoder *enc10 = TO_DCN10_LINK_ENC(enc);
+	union dmub_rb_cmd cmd;
 	uint32_t is_in_usb_c_dp4_mode = 0;
 
 	dcn10_link_encoder_get_max_link_cap(enc, link_settings);
 
-	/* in usb c dp2 mode, max lane count is 2 */
-	if (enc->funcs->is_in_alt_mode && enc->funcs->is_in_alt_mode(enc)) {
-		if (enc->ctx->asic_id.hw_internal_rev != YELLOW_CARP_B0) {
-			// [Note] no need to check hw_internal_rev once phy mux selection is ready
-			REG_GET(RDPCSTX_PHY_CNTL6, RDPCS_PHY_DPALT_DP4, &is_in_usb_c_dp4_mode);
+	/* Take the link cap directly if not USB */
+	if (!enc->features.flags.bits.DP_IS_USB_C)
+		return;
+
+	/*
+	 * Use the new interface from DMCUB if available.
+	 * Avoids hanging the RDCPSPIPE if DMCUB wasn't already running.
+	 */
+	if (has_query_dp_alt(enc)) {
+		if (!query_dp_alt_from_dmub(enc, &cmd))
+			return;
+
+		if (cmd.query_dp_alt.data.is_dp_alt_disable == 0 &&
+				cmd.query_dp_alt.data.is_usb &&
+				cmd.query_dp_alt.data.is_dp4 == 0)
+			link_settings->lane_count = MIN(LANE_COUNT_TWO, link_settings->lane_count);
+
+		return;
+	}
+
+	/* Legacy path, avoid if possible. */
+	if (enc->ctx->asic_id.hw_internal_rev != YELLOW_CARP_B0) {
+		REG_GET(RDPCSTX_PHY_CNTL6, RDPCS_PHY_DPALT_DP4,
+			&is_in_usb_c_dp4_mode);
+	} else {
+		if ((enc10->base.transmitter == TRANSMITTER_UNIPHY_A) ||
+		    (enc10->base.transmitter == TRANSMITTER_UNIPHY_B) ||
+		    (enc10->base.transmitter == TRANSMITTER_UNIPHY_E)) {
+			REG_GET(RDPCSTX_PHY_CNTL6, RDPCS_PHY_DPALT_DP4,
+				&is_in_usb_c_dp4_mode);
 		} else {
-			if ((enc10->base.transmitter == TRANSMITTER_UNIPHY_A)
-					|| (enc10->base.transmitter == TRANSMITTER_UNIPHY_B)
-					|| (enc10->base.transmitter == TRANSMITTER_UNIPHY_E)) {
-				REG_GET(RDPCSTX_PHY_CNTL6, RDPCS_PHY_DPALT_DP4, &is_in_usb_c_dp4_mode);
-			} else {
-				REG_GET(RDPCSPIPE_PHY_CNTL6, RDPCS_PHY_DPALT_DP4, &is_in_usb_c_dp4_mode);
-			}
+			REG_GET(RDPCSPIPE_PHY_CNTL6, RDPCS_PHY_DPALT_DP4,
+				&is_in_usb_c_dp4_mode);
 		}
-		if (!is_in_usb_c_dp4_mode)
-			link_settings->lane_count = MIN(LANE_COUNT_TWO, link_settings->lane_count);
 	}
+
+	if (!is_in_usb_c_dp4_mode)
+		link_settings->lane_count = MIN(LANE_COUNT_TWO, link_settings->lane_count);
 }
diff --git a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_dio_link_encoder.h b/drivers/gpu/drm/amd/display/dc/dio/dcn31/dcn31_dio_link_encoder.h
index 3454f1e7c1f1..ee78ba80797c 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_dio_link_encoder.h
+++ b/drivers/gpu/drm/amd/display/dc/dio/dcn31/dcn31_dio_link_encoder.h
@@ -89,6 +89,7 @@
 	SRI(RDPCSTX_PHY_FUSE1, RDPCSTX, id), \
 	SRI(RDPCSTX_PHY_FUSE2, RDPCSTX, id), \
 	SRI(RDPCSTX_PHY_FUSE3, RDPCSTX, id), \
+	SRI(RDPCSTX_DEBUG_CONFIG, RDPCSTX, id), \
 	SR(RDPCSTX0_RDPCSTX_SCRATCH), \
 	SRI(RDPCSTX_PHY_RX_LD_VAL, RDPCSTX, id),\
 	SRI(RDPCSTX_DMCU_DPALT_DIS_BLOCK_REG, RDPCSTX, id)
@@ -198,6 +199,35 @@
 	LE_SF(RDPCSTX0_RDPCSTX_PHY_FUSE3, RDPCS_PHY_DP_TX3_EQ_PRE, mask_sh),\
 	LE_SF(RDPCSTX0_RDPCSTX_PHY_FUSE3, RDPCS_PHY_DP_TX3_EQ_POST, mask_sh)
 
+#define DPCS_DCN314_REG_LIST(id) \
+	SRI(TMDS_CTL_BITS, DIG, id), \
+	SRI(RDPCSTX_PHY_CNTL3, RDPCSTX, id), \
+	SRI(RDPCSTX_PHY_CNTL4, RDPCSTX, id), \
+	SRI(RDPCSTX_PHY_CNTL5, RDPCSTX, id), \
+	SRI(RDPCSTX_PHY_CNTL7, RDPCSTX, id), \
+	SRI(RDPCSTX_PHY_CNTL8, RDPCSTX, id), \
+	SRI(RDPCSTX_PHY_CNTL9, RDPCSTX, id), \
+	SRI(RDPCSTX_PHY_CNTL10, RDPCSTX, id), \
+	SRI(RDPCSTX_PHY_CNTL11, RDPCSTX, id), \
+	SRI(RDPCSTX_PHY_CNTL12, RDPCSTX, id), \
+	SRI(RDPCSTX_PHY_CNTL13, RDPCSTX, id), \
+	SRI(RDPCSTX_PHY_CNTL14, RDPCSTX, id), \
+	SRI(RDPCSTX_CNTL, RDPCSTX, id), \
+	SRI(RDPCSTX_CLOCK_CNTL, RDPCSTX, id), \
+	SRI(RDPCSTX_INTERRUPT_CONTROL, RDPCSTX, id), \
+	SRI(RDPCSTX_PHY_CNTL0, RDPCSTX, id), \
+	SRI(RDPCSTX_PHY_CNTL2, RDPCSTX, id), \
+	SRI(RDPCS_TX_CR_ADDR, RDPCSTX, id), \
+	SRI(RDPCS_TX_CR_DATA, RDPCSTX, id), \
+	SRI(RDPCSTX_PHY_FUSE0, RDPCSTX, id), \
+	SRI(RDPCSTX_PHY_FUSE1, RDPCSTX, id), \
+	SRI(RDPCSTX_PHY_FUSE2, RDPCSTX, id), \
+	SRI(RDPCSTX_PHY_FUSE3, RDPCSTX, id), \
+	SRI(RDPCSTX_DEBUG_CONFIG, RDPCSTX, id), \
+	SR(RDPCSTX0_RDPCSTX_SCRATCH), \
+	SRI(RDPCSTX_PHY_RX_LD_VAL, RDPCSTX, id),\
+	SRI(RDPCSTX_DMCU_DPALT_DIS_BLOCK_REG, RDPCSTX, id)
+
 void dcn31_link_encoder_construct(
 	struct dcn20_link_encoder *enc20,
 	const struct encoder_init_data *init_data,
@@ -255,4 +285,6 @@ bool dcn31_link_encoder_is_in_alt_mode(
 void dcn31_link_encoder_get_max_link_cap(struct link_encoder *enc,
 	struct dc_link_settings *link_settings);
 
+void enc31_hw_init(struct link_encoder *enc);
+
 #endif /* __DC_LINK_ENCODER__DCN31_H__ */
diff --git a/drivers/gpu/drm/amd/display/dc/dio/dcn314/dcn314_dio_stream_encoder.c b/drivers/gpu/drm/amd/display/dc/dio/dcn314/dcn314_dio_stream_encoder.c
new file mode 100644
index 000000000000..3e85e9c3d2cb
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dio/dcn314/dcn314_dio_stream_encoder.c
@@ -0,0 +1,506 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ *  and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+
+#include "dc_bios_types.h"
+#include "dcn30/dcn30_dio_stream_encoder.h"
+#include "dcn314_dio_stream_encoder.h"
+#include "reg_helper.h"
+#include "hw_shared.h"
+#include "link_service.h"
+#include "dpcd_defs.h"
+
+#define DC_LOGGER \
+		enc1->base.ctx->logger
+
+#define REG(reg)\
+	(enc1->regs->reg)
+
+#undef FN
+#define FN(reg_name, field_name) \
+	enc1->se_shift->field_name, enc1->se_mask->field_name
+
+#define VBI_LINE_0 0
+#define HDMI_CLOCK_CHANNEL_RATE_MORE_340M 340000
+
+#define CTX \
+	enc1->base.ctx
+
+void enc314_reset_fifo(struct stream_encoder *enc, bool reset)
+{
+	struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc);
+	uint32_t reset_val = reset ? 1 : 0;
+	uint32_t is_symclk_on;
+
+	REG_UPDATE(DIG_FIFO_CTRL0, DIG_FIFO_RESET, reset_val);
+	REG_GET(DIG_FE_CNTL, DIG_SYMCLK_FE_ON, &is_symclk_on);
+
+	if (is_symclk_on)
+		REG_WAIT(DIG_FIFO_CTRL0, DIG_FIFO_RESET_DONE, reset_val, 10, 5000);
+	else
+		udelay(10);
+}
+
+void enc314_enable_fifo(struct stream_encoder *enc)
+{
+	struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc);
+
+	REG_UPDATE(DIG_FIFO_CTRL0, DIG_FIFO_READ_START_LEVEL, 0x7);
+
+	enc314_reset_fifo(enc, true);
+	enc314_reset_fifo(enc, false);
+
+	REG_UPDATE(DIG_FIFO_CTRL0, DIG_FIFO_ENABLE, 1);
+}
+
+void enc314_disable_fifo(struct stream_encoder *enc)
+{
+	struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc);
+
+	REG_UPDATE(DIG_FIFO_CTRL0, DIG_FIFO_ENABLE, 0);
+}
+
+static bool enc314_is_fifo_enabled(struct stream_encoder *enc)
+{
+	struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc);
+	uint32_t reset_val;
+
+	REG_GET(DIG_FIFO_CTRL0, DIG_FIFO_ENABLE, &reset_val);
+	return (reset_val != 0);
+}
+
+void enc314_dp_set_odm_combine(
+	struct stream_encoder *enc,
+	bool odm_combine)
+{
+	struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc);
+
+	REG_UPDATE(DP_PIXEL_FORMAT, DP_PIXEL_PER_CYCLE_PROCESSING_MODE, odm_combine);
+}
+
+/* setup stream encoder in dvi mode */
+void enc314_stream_encoder_dvi_set_stream_attribute(
+	struct stream_encoder *enc,
+	struct dc_crtc_timing *crtc_timing,
+	bool is_dual_link)
+{
+	struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc);
+
+	if (!enc->ctx->dc->debug.avoid_vbios_exec_table) {
+		struct bp_encoder_control cntl = {0};
+
+		cntl.action = ENCODER_CONTROL_SETUP;
+		cntl.engine_id = enc1->base.id;
+		cntl.signal = is_dual_link ?
+			SIGNAL_TYPE_DVI_DUAL_LINK : SIGNAL_TYPE_DVI_SINGLE_LINK;
+		cntl.enable_dp_audio = false;
+		cntl.pixel_clock = crtc_timing->pix_clk_100hz / 10;
+		cntl.lanes_number = (is_dual_link) ? LANE_COUNT_EIGHT : LANE_COUNT_FOUR;
+
+		if (enc1->base.bp->funcs->encoder_control(
+				enc1->base.bp, &cntl) != BP_RESULT_OK)
+			return;
+
+	} else {
+
+		//Set pattern for clock channel, default vlue 0x63 does not work
+		REG_UPDATE(DIG_CLOCK_PATTERN, DIG_CLOCK_PATTERN, 0x1F);
+
+		//DIG_BE_TMDS_DVI_MODE : TMDS-DVI mode is already set in link_encoder_setup
+
+		//DIG_SOURCE_SELECT is already set in dig_connect_to_otg
+
+		enc314_enable_fifo(enc);
+	}
+
+	ASSERT(crtc_timing->pixel_encoding == PIXEL_ENCODING_RGB);
+	ASSERT(crtc_timing->display_color_depth == COLOR_DEPTH_888);
+	enc1_stream_encoder_set_stream_attribute_helper(enc1, crtc_timing);
+}
+
+/* setup stream encoder in hdmi mode */
+void enc314_stream_encoder_hdmi_set_stream_attribute(
+	struct stream_encoder *enc,
+	struct dc_crtc_timing *crtc_timing,
+	int actual_pix_clk_khz,
+	bool enable_audio)
+{
+	struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc);
+
+	if (!enc->ctx->dc->debug.avoid_vbios_exec_table) {
+		struct bp_encoder_control cntl = {0};
+
+		cntl.action = ENCODER_CONTROL_SETUP;
+		cntl.engine_id = enc1->base.id;
+		cntl.signal = SIGNAL_TYPE_HDMI_TYPE_A;
+		cntl.enable_dp_audio = enable_audio;
+		cntl.pixel_clock = actual_pix_clk_khz;
+		cntl.lanes_number = LANE_COUNT_FOUR;
+
+		if (enc1->base.bp->funcs->encoder_control(
+				enc1->base.bp, &cntl) != BP_RESULT_OK)
+			return;
+
+	} else {
+
+		//Set pattern for clock channel, default vlue 0x63 does not work
+		REG_UPDATE(DIG_CLOCK_PATTERN, DIG_CLOCK_PATTERN, 0x1F);
+
+		//DIG_BE_TMDS_HDMI_MODE : TMDS-HDMI mode is already set in link_encoder_setup
+
+		//DIG_SOURCE_SELECT is already set in dig_connect_to_otg
+
+		enc314_enable_fifo(enc);
+	}
+
+	/* Configure pixel encoding */
+	enc1_stream_encoder_set_stream_attribute_helper(enc1, crtc_timing);
+
+	/* setup HDMI engine */
+	REG_UPDATE_6(HDMI_CONTROL,
+		HDMI_PACKET_GEN_VERSION, 1,
+		HDMI_KEEPOUT_MODE, 1,
+		HDMI_DEEP_COLOR_ENABLE, 0,
+		HDMI_DATA_SCRAMBLE_EN, 0,
+		HDMI_NO_EXTRA_NULL_PACKET_FILLED, 1,
+		HDMI_CLOCK_CHANNEL_RATE, 0);
+
+	/* Configure color depth */
+	switch (crtc_timing->display_color_depth) {
+	case COLOR_DEPTH_888:
+		REG_UPDATE(HDMI_CONTROL, HDMI_DEEP_COLOR_DEPTH, 0);
+		break;
+	case COLOR_DEPTH_101010:
+		if (crtc_timing->pixel_encoding == PIXEL_ENCODING_YCBCR422) {
+			REG_UPDATE_2(HDMI_CONTROL,
+					HDMI_DEEP_COLOR_DEPTH, 1,
+					HDMI_DEEP_COLOR_ENABLE, 0);
+		} else {
+			REG_UPDATE_2(HDMI_CONTROL,
+					HDMI_DEEP_COLOR_DEPTH, 1,
+					HDMI_DEEP_COLOR_ENABLE, 1);
+			}
+		break;
+	case COLOR_DEPTH_121212:
+		if (crtc_timing->pixel_encoding == PIXEL_ENCODING_YCBCR422) {
+			REG_UPDATE_2(HDMI_CONTROL,
+					HDMI_DEEP_COLOR_DEPTH, 2,
+					HDMI_DEEP_COLOR_ENABLE, 0);
+		} else {
+			REG_UPDATE_2(HDMI_CONTROL,
+					HDMI_DEEP_COLOR_DEPTH, 2,
+					HDMI_DEEP_COLOR_ENABLE, 1);
+			}
+		break;
+	case COLOR_DEPTH_161616:
+		REG_UPDATE_2(HDMI_CONTROL,
+				HDMI_DEEP_COLOR_DEPTH, 3,
+				HDMI_DEEP_COLOR_ENABLE, 1);
+		break;
+	default:
+		break;
+	}
+
+	if (actual_pix_clk_khz >= HDMI_CLOCK_CHANNEL_RATE_MORE_340M) {
+		/* enable HDMI data scrambler
+		 * HDMI_CLOCK_CHANNEL_RATE_MORE_340M
+		 * Clock channel frequency is 1/4 of character rate.
+		 */
+		REG_UPDATE_2(HDMI_CONTROL,
+			HDMI_DATA_SCRAMBLE_EN, 1,
+			HDMI_CLOCK_CHANNEL_RATE, 1);
+	} else if (crtc_timing->flags.LTE_340MCSC_SCRAMBLE) {
+
+		/* TODO: New feature for DCE11, still need to implement */
+
+		/* enable HDMI data scrambler
+		 * HDMI_CLOCK_CHANNEL_FREQ_EQUAL_TO_CHAR_RATE
+		 * Clock channel frequency is the same
+		 * as character rate
+		 */
+		REG_UPDATE_2(HDMI_CONTROL,
+			HDMI_DATA_SCRAMBLE_EN, 1,
+			HDMI_CLOCK_CHANNEL_RATE, 0);
+	}
+
+
+	/* Enable transmission of General Control packet on every frame */
+	REG_UPDATE_3(HDMI_VBI_PACKET_CONTROL,
+		HDMI_GC_CONT, 1,
+		HDMI_GC_SEND, 1,
+		HDMI_NULL_SEND, 1);
+
+	/* Disable Audio Content Protection packet transmission */
+	REG_UPDATE(HDMI_VBI_PACKET_CONTROL, HDMI_ACP_SEND, 0);
+
+	/* following belongs to audio */
+	/* Enable Audio InfoFrame packet transmission. */
+	REG_UPDATE(HDMI_INFOFRAME_CONTROL0, HDMI_AUDIO_INFO_SEND, 1);
+
+	/* update double-buffered AUDIO_INFO registers immediately */
+	ASSERT(enc->afmt);
+	enc->afmt->funcs->audio_info_immediate_update(enc->afmt);
+
+	/* Select line number on which to send Audio InfoFrame packets */
+	REG_UPDATE(HDMI_INFOFRAME_CONTROL1, HDMI_AUDIO_INFO_LINE,
+				VBI_LINE_0 + 2);
+
+	/* set HDMI GC AVMUTE */
+	REG_UPDATE(HDMI_GC, HDMI_GC_AVMUTE, 0);
+}
+
+
+
+static bool is_two_pixels_per_containter(const struct dc_crtc_timing *timing)
+{
+	bool two_pix = timing->pixel_encoding == PIXEL_ENCODING_YCBCR420;
+
+	two_pix = two_pix || (timing->flags.DSC && timing->pixel_encoding == PIXEL_ENCODING_YCBCR422
+			&& !timing->dsc_cfg.ycbcr422_simple);
+	return two_pix;
+}
+
+void enc314_stream_encoder_dp_blank(
+	struct dc_link *link,
+	struct stream_encoder *enc)
+{
+	enc1_stream_encoder_dp_blank(link, enc);
+
+	/* Disable FIFO after the DP vid stream is disabled to avoid corruption. */
+	if (enc->ctx->dc->debug.dig_fifo_off_in_blank)
+		enc314_disable_fifo(enc);
+}
+
+void enc314_stream_encoder_dp_unblank(
+		struct dc_link *link,
+		struct stream_encoder *enc,
+		const struct encoder_unblank_param *param)
+{
+	struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc);
+
+	if (param->link_settings.link_rate != LINK_RATE_UNKNOWN) {
+		uint32_t n_vid = 0x8000;
+		uint32_t m_vid;
+		uint32_t n_multiply = 0;
+		uint32_t pix_per_cycle = 0;
+		uint64_t m_vid_l = n_vid;
+
+		/* YCbCr 4:2:0 : Computed VID_M will be 2X the input rate */
+		if (is_two_pixels_per_containter(&param->timing) || param->opp_cnt > 1) {
+			/*this logic should be the same in get_pixel_clock_parameters() */
+			n_multiply = 1;
+			pix_per_cycle = 1;
+		}
+		/* M / N = Fstream / Flink
+		 * m_vid / n_vid = pixel rate / link rate
+		 */
+
+		m_vid_l *= param->timing.pix_clk_100hz / 10;
+		m_vid_l = div_u64(m_vid_l,
+			param->link_settings.link_rate
+				* LINK_RATE_REF_FREQ_IN_KHZ);
+
+		m_vid = (uint32_t) m_vid_l;
+
+		/* enable auto measurement */
+
+		REG_UPDATE(DP_VID_TIMING, DP_VID_M_N_GEN_EN, 0);
+
+		/* auto measurement need 1 full 0x8000 symbol cycle to kick in,
+		 * therefore program initial value for Mvid and Nvid
+		 */
+
+		REG_UPDATE(DP_VID_N, DP_VID_N, n_vid);
+
+		REG_UPDATE(DP_VID_M, DP_VID_M, m_vid);
+
+		REG_UPDATE_2(DP_VID_TIMING,
+				DP_VID_M_N_GEN_EN, 1,
+				DP_VID_N_MUL, n_multiply);
+
+		REG_UPDATE(DP_PIXEL_FORMAT,
+				DP_PIXEL_PER_CYCLE_PROCESSING_MODE,
+				pix_per_cycle);
+	}
+
+	/* make sure stream is disabled before resetting steer fifo */
+	REG_UPDATE(DP_VID_STREAM_CNTL, DP_VID_STREAM_ENABLE, false);
+	REG_WAIT(DP_VID_STREAM_CNTL, DP_VID_STREAM_STATUS, 0, 10, 5000);
+
+	/* DIG_START is removed from the register spec */
+
+	/* switch DP encoder to CRTC data, but reset it the fifo first. It may happen
+	 * that it overflows during mode transition, and sometimes doesn't recover.
+	 */
+	REG_UPDATE(DP_STEER_FIFO, DP_STEER_FIFO_RESET, 1);
+	udelay(10);
+
+	REG_UPDATE(DP_STEER_FIFO, DP_STEER_FIFO_RESET, 0);
+
+	/* wait 100us for DIG/DP logic to prime
+	 * (i.e. a few video lines)
+	 */
+	udelay(100);
+
+	/* the hardware would start sending video at the start of the next DP
+	 * frame (i.e. rising edge of the vblank).
+	 * NOTE: We used to program DP_VID_STREAM_DIS_DEFER = 2 here, but this
+	 * register has no effect on enable transition! HW always guarantees
+	 * VID_STREAM enable at start of next frame, and this is not
+	 * programmable
+	 */
+
+	REG_UPDATE(DP_VID_STREAM_CNTL, DP_VID_STREAM_ENABLE, true);
+
+	/*
+	 * DIG Resync FIFO now needs to be explicitly enabled.
+	 * This should come after DP_VID_STREAM_ENABLE per HW docs.
+	 */
+	enc314_enable_fifo(enc);
+
+	link->dc->link_srv->dp_trace_source_sequence(link, DPCD_SOURCE_SEQ_AFTER_ENABLE_DP_VID_STREAM);
+}
+
+/* Set DSC-related configuration.
+ *   dsc_mode: 0 disables DSC, other values enable DSC in specified format
+ *   sc_bytes_per_pixel: DP_DSC_BYTES_PER_PIXEL removed in DCN32
+ *   dsc_slice_width: DP_DSC_SLICE_WIDTH removed in DCN32
+ */
+void enc314_dp_set_dsc_config(struct stream_encoder *enc,
+					enum optc_dsc_mode dsc_mode,
+					uint32_t dsc_bytes_per_pixel,
+					uint32_t dsc_slice_width)
+{
+	struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc);
+
+	REG_UPDATE(DP_DSC_CNTL,	DP_DSC_MODE, dsc_mode == OPTC_DSC_DISABLED ? 0 : 1);
+}
+
+/* this function read dsc related register fields to be logged later in dcn10_log_hw_state
+ * into a dcn_dsc_state struct.
+ */
+void enc314_read_state(struct stream_encoder *enc, struct enc_state *s)
+{
+	struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc);
+
+	//if dsc is enabled, continue to read
+	REG_GET(DP_DSC_CNTL, DP_DSC_MODE, &s->dsc_mode);
+	if (s->dsc_mode) {
+		REG_GET(DP_GSP11_CNTL, DP_SEC_GSP11_LINE_NUM, &s->sec_gsp_pps_line_num);
+
+		REG_GET(DP_MSA_VBID_MISC, DP_VBID6_LINE_REFERENCE, &s->vbid6_line_reference);
+		REG_GET(DP_MSA_VBID_MISC, DP_VBID6_LINE_NUM, &s->vbid6_line_num);
+
+		REG_GET(DP_GSP11_CNTL, DP_SEC_GSP11_ENABLE, &s->sec_gsp_pps_enable);
+		REG_GET(DP_SEC_CNTL, DP_SEC_STREAM_ENABLE, &s->sec_stream_enable);
+	}
+}
+
+void enc314_set_dig_input_mode(struct stream_encoder *enc, unsigned int pix_per_container)
+{
+	struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc);
+
+	// The naming of this field is confusing, what it means is the output mode of otg, which
+	// is the input mode of the dig
+	REG_UPDATE(DIG_FIFO_CTRL0, DIG_FIFO_OUTPUT_PIXEL_MODE, pix_per_container == 2 ? 0x1 : 0x0);
+}
+
+
+static const struct stream_encoder_funcs dcn314_str_enc_funcs = {
+	.dp_set_odm_combine =
+		enc314_dp_set_odm_combine,
+	.dp_set_stream_attribute =
+		enc2_stream_encoder_dp_set_stream_attribute,
+	.hdmi_set_stream_attribute =
+		enc314_stream_encoder_hdmi_set_stream_attribute,
+	.dvi_set_stream_attribute =
+		enc314_stream_encoder_dvi_set_stream_attribute,
+	.set_throttled_vcp_size =
+		enc1_stream_encoder_set_throttled_vcp_size,
+	.update_hdmi_info_packets =
+		enc3_stream_encoder_update_hdmi_info_packets,
+	.stop_hdmi_info_packets =
+		enc3_stream_encoder_stop_hdmi_info_packets,
+	.update_dp_info_packets_sdp_line_num =
+		enc3_stream_encoder_update_dp_info_packets_sdp_line_num,
+	.update_dp_info_packets =
+		enc3_stream_encoder_update_dp_info_packets,
+	.stop_dp_info_packets =
+		enc1_stream_encoder_stop_dp_info_packets,
+	.dp_blank =
+		enc314_stream_encoder_dp_blank,
+	.dp_unblank =
+		enc314_stream_encoder_dp_unblank,
+	.audio_mute_control = enc3_audio_mute_control,
+
+	.dp_audio_setup = enc3_se_dp_audio_setup,
+	.dp_audio_enable = enc3_se_dp_audio_enable,
+	.dp_audio_disable = enc1_se_dp_audio_disable,
+
+	.hdmi_audio_setup = enc3_se_hdmi_audio_setup,
+	.hdmi_audio_disable = enc1_se_hdmi_audio_disable,
+	.setup_stereo_sync  = enc1_setup_stereo_sync,
+	.set_avmute = enc1_stream_encoder_set_avmute,
+	.dig_connect_to_otg = enc1_dig_connect_to_otg,
+	.dig_source_otg = enc1_dig_source_otg,
+
+	.dp_get_pixel_format  = enc1_stream_encoder_dp_get_pixel_format,
+
+	.enc_read_state = enc314_read_state,
+	.dp_set_dsc_config = enc314_dp_set_dsc_config,
+	.dp_set_dsc_pps_info_packet = enc3_dp_set_dsc_pps_info_packet,
+	.set_dynamic_metadata = enc2_set_dynamic_metadata,
+	.hdmi_reset_stream_attribute = enc1_reset_hdmi_stream_attribute,
+
+	.enable_fifo = enc314_enable_fifo,
+	.disable_fifo = enc314_disable_fifo,
+	.is_fifo_enabled = enc314_is_fifo_enabled,
+	.set_input_mode = enc314_set_dig_input_mode,
+};
+
+void dcn314_dio_stream_encoder_construct(
+	struct dcn10_stream_encoder *enc1,
+	struct dc_context *ctx,
+	struct dc_bios *bp,
+	enum engine_id eng_id,
+	struct vpg *vpg,
+	struct afmt *afmt,
+	const struct dcn10_stream_enc_registers *regs,
+	const struct dcn10_stream_encoder_shift *se_shift,
+	const struct dcn10_stream_encoder_mask *se_mask)
+{
+	enc1->base.funcs = &dcn314_str_enc_funcs;
+	enc1->base.ctx = ctx;
+	enc1->base.id = eng_id;
+	enc1->base.bp = bp;
+	enc1->base.vpg = vpg;
+	enc1->base.afmt = afmt;
+	enc1->regs = regs;
+	enc1->se_shift = se_shift;
+	enc1->se_mask = se_mask;
+	enc1->base.stream_enc_inst = vpg->inst;
+}
+
diff --git a/drivers/gpu/drm/amd/display/dc/dio/dcn314/dcn314_dio_stream_encoder.h b/drivers/gpu/drm/amd/display/dc/dio/dcn314/dcn314_dio_stream_encoder.h
new file mode 100644
index 000000000000..86548be591be
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dio/dcn314/dcn314_dio_stream_encoder.h
@@ -0,0 +1,355 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ *  and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __DC_DIO_STREAM_ENCODER_DCN314_H__
+#define __DC_DIO_STREAM_ENCODER_DCN314_H__
+
+#include "dcn30/dcn30_vpg.h"
+#include "dcn30/dcn30_afmt.h"
+#include "stream_encoder.h"
+#include "dcn20/dcn20_stream_encoder.h"
+
+/* Register bit field name change */
+#define RDPCSTX0_RDPCSTX_CLOCK_CNTL__RDPCS_SYMCLK_DIV2_GATE_DIS__SHIFT                                        0x8
+#define RDPCSTX0_RDPCSTX_CLOCK_CNTL__RDPCS_SYMCLK_DIV2_EN__SHIFT                                              0x9
+#define RDPCSTX0_RDPCSTX_CLOCK_CNTL__RDPCS_SYMCLK_DIV2_CLOCK_ON__SHIFT                                        0xa
+#define DPCSTX0_DPCSTX_TX_CNTL__DPCS_TX_DATA_SWAP__SHIFT                                                      0xe
+#define DPCSTX0_DPCSTX_TX_CNTL__DPCS_TX_DATA_ORDER_INVERT__SHIFT                                              0xf
+
+#define RDPCSTX0_RDPCSTX_CLOCK_CNTL__RDPCS_SYMCLK_DIV2_GATE_DIS_MASK                                          0x00000100L
+#define RDPCSTX0_RDPCSTX_CLOCK_CNTL__RDPCS_SYMCLK_DIV2_EN_MASK                                                0x00000200L
+#define RDPCSTX0_RDPCSTX_CLOCK_CNTL__RDPCS_SYMCLK_DIV2_CLOCK_ON_MASK                                          0x00000400L
+#define DPCSTX0_DPCSTX_TX_CNTL__DPCS_TX_DATA_SWAP_MASK                                                        0x00004000L
+#define DPCSTX0_DPCSTX_TX_CNTL__DPCS_TX_DATA_ORDER_INVERT_MASK                                                0x00008000L
+
+
+#define SE_DCN314_REG_LIST(id)\
+	SRI(AFMT_CNTL, DIG, id), \
+	SRI(DIG_FE_CNTL, DIG, id), \
+	SRI(HDMI_CONTROL, DIG, id), \
+	SRI(HDMI_DB_CONTROL, DIG, id), \
+	SRI(HDMI_GC, DIG, id), \
+	SRI(HDMI_GENERIC_PACKET_CONTROL0, DIG, id), \
+	SRI(HDMI_GENERIC_PACKET_CONTROL1, DIG, id), \
+	SRI(HDMI_GENERIC_PACKET_CONTROL2, DIG, id), \
+	SRI(HDMI_GENERIC_PACKET_CONTROL3, DIG, id), \
+	SRI(HDMI_GENERIC_PACKET_CONTROL4, DIG, id), \
+	SRI(HDMI_GENERIC_PACKET_CONTROL5, DIG, id), \
+	SRI(HDMI_GENERIC_PACKET_CONTROL6, DIG, id), \
+	SRI(HDMI_GENERIC_PACKET_CONTROL7, DIG, id), \
+	SRI(HDMI_GENERIC_PACKET_CONTROL8, DIG, id), \
+	SRI(HDMI_GENERIC_PACKET_CONTROL9, DIG, id), \
+	SRI(HDMI_GENERIC_PACKET_CONTROL10, DIG, id), \
+	SRI(HDMI_INFOFRAME_CONTROL0, DIG, id), \
+	SRI(HDMI_INFOFRAME_CONTROL1, DIG, id), \
+	SRI(HDMI_VBI_PACKET_CONTROL, DIG, id), \
+	SRI(HDMI_AUDIO_PACKET_CONTROL, DIG, id),\
+	SRI(HDMI_ACR_PACKET_CONTROL, DIG, id),\
+	SRI(HDMI_ACR_32_0, DIG, id),\
+	SRI(HDMI_ACR_32_1, DIG, id),\
+	SRI(HDMI_ACR_44_0, DIG, id),\
+	SRI(HDMI_ACR_44_1, DIG, id),\
+	SRI(HDMI_ACR_48_0, DIG, id),\
+	SRI(HDMI_ACR_48_1, DIG, id),\
+	SRI(DP_DB_CNTL, DP, id), \
+	SRI(DP_MSA_MISC, DP, id), \
+	SRI(DP_MSA_VBID_MISC, DP, id), \
+	SRI(DP_MSA_COLORIMETRY, DP, id), \
+	SRI(DP_MSA_TIMING_PARAM1, DP, id), \
+	SRI(DP_MSA_TIMING_PARAM2, DP, id), \
+	SRI(DP_MSA_TIMING_PARAM3, DP, id), \
+	SRI(DP_MSA_TIMING_PARAM4, DP, id), \
+	SRI(DP_MSE_RATE_CNTL, DP, id), \
+	SRI(DP_MSE_RATE_UPDATE, DP, id), \
+	SRI(DP_PIXEL_FORMAT, DP, id), \
+	SRI(DP_SEC_CNTL, DP, id), \
+	SRI(DP_SEC_CNTL1, DP, id), \
+	SRI(DP_SEC_CNTL2, DP, id), \
+	SRI(DP_SEC_CNTL5, DP, id), \
+	SRI(DP_SEC_CNTL6, DP, id), \
+	SRI(DP_STEER_FIFO, DP, id), \
+	SRI(DP_VID_M, DP, id), \
+	SRI(DP_VID_N, DP, id), \
+	SRI(DP_VID_STREAM_CNTL, DP, id), \
+	SRI(DP_VID_TIMING, DP, id), \
+	SRI(DP_SEC_AUD_N, DP, id), \
+	SRI(DP_SEC_TIMESTAMP, DP, id), \
+	SRI(DP_DSC_CNTL, DP, id), \
+	SRI(DP_SEC_METADATA_TRANSMISSION, DP, id), \
+	SRI(HDMI_METADATA_PACKET_CONTROL, DIG, id), \
+	SRI(DP_SEC_FRAMING4, DP, id), \
+	SRI(DP_GSP11_CNTL, DP, id), \
+	SRI(DME_CONTROL, DME, id),\
+	SRI(DP_SEC_METADATA_TRANSMISSION, DP, id), \
+	SRI(HDMI_METADATA_PACKET_CONTROL, DIG, id), \
+	SRI(DIG_FE_CNTL, DIG, id), \
+	SRI(DIG_CLOCK_PATTERN, DIG, id), \
+	SRI(DIG_FIFO_CTRL0, DIG, id)
+
+
+#define SE_COMMON_MASK_SH_LIST_DCN314(mask_sh)\
+	SE_SF(DP0_DP_PIXEL_FORMAT, DP_PIXEL_ENCODING, mask_sh),\
+	SE_SF(DP0_DP_PIXEL_FORMAT, DP_COMPONENT_DEPTH, mask_sh),\
+	SE_SF(DP0_DP_PIXEL_FORMAT, DP_PIXEL_PER_CYCLE_PROCESSING_MODE, mask_sh),\
+	SE_SF(DIG0_HDMI_CONTROL, HDMI_PACKET_GEN_VERSION, mask_sh),\
+	SE_SF(DIG0_HDMI_CONTROL, HDMI_KEEPOUT_MODE, mask_sh),\
+	SE_SF(DIG0_HDMI_CONTROL, HDMI_DEEP_COLOR_ENABLE, mask_sh),\
+	SE_SF(DIG0_HDMI_CONTROL, HDMI_DEEP_COLOR_DEPTH, mask_sh),\
+	SE_SF(DIG0_HDMI_CONTROL, HDMI_DATA_SCRAMBLE_EN, mask_sh),\
+	SE_SF(DIG0_HDMI_CONTROL, HDMI_NO_EXTRA_NULL_PACKET_FILLED, mask_sh),\
+	SE_SF(DIG0_HDMI_VBI_PACKET_CONTROL, HDMI_GC_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_VBI_PACKET_CONTROL, HDMI_GC_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_VBI_PACKET_CONTROL, HDMI_NULL_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_VBI_PACKET_CONTROL, HDMI_ACP_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_INFOFRAME_CONTROL0, HDMI_AUDIO_INFO_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_INFOFRAME_CONTROL1, HDMI_AUDIO_INFO_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GC, HDMI_GC_AVMUTE, mask_sh),\
+	SE_SF(DP0_DP_MSE_RATE_CNTL, DP_MSE_RATE_X, mask_sh),\
+	SE_SF(DP0_DP_MSE_RATE_CNTL, DP_MSE_RATE_Y, mask_sh),\
+	SE_SF(DP0_DP_MSE_RATE_UPDATE, DP_MSE_RATE_UPDATE_PENDING, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_GSP0_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_STREAM_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_GSP1_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_GSP2_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_GSP3_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_MPG_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL1, DP_SEC_GSP5_LINE_REFERENCE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL2, DP_SEC_GSP4_SEND, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL2, DP_SEC_GSP4_SEND_PENDING, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL4, DP_SEC_GSP4_LINE_NUM, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL5, DP_SEC_GSP5_LINE_NUM, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL2, DP_SEC_GSP4_SEND_ANY_LINE, mask_sh),\
+	SE_SF(DP0_DP_VID_STREAM_CNTL, DP_VID_STREAM_DIS_DEFER, mask_sh),\
+	SE_SF(DP0_DP_VID_STREAM_CNTL, DP_VID_STREAM_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_VID_STREAM_CNTL, DP_VID_STREAM_STATUS, mask_sh),\
+	SE_SF(DP0_DP_STEER_FIFO, DP_STEER_FIFO_RESET, mask_sh),\
+	SE_SF(DP0_DP_VID_TIMING, DP_VID_M_N_GEN_EN, mask_sh),\
+	SE_SF(DP0_DP_VID_N, DP_VID_N, mask_sh),\
+	SE_SF(DP0_DP_VID_M, DP_VID_M, mask_sh),\
+	SE_SF(DIG0_HDMI_AUDIO_PACKET_CONTROL, HDMI_AUDIO_DELAY_EN, mask_sh),\
+	SE_SF(DIG0_HDMI_ACR_PACKET_CONTROL, HDMI_ACR_AUTO_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_ACR_PACKET_CONTROL, HDMI_ACR_SOURCE, mask_sh),\
+	SE_SF(DIG0_HDMI_ACR_PACKET_CONTROL, HDMI_ACR_AUDIO_PRIORITY, mask_sh),\
+	SE_SF(DIG0_HDMI_ACR_32_0, HDMI_ACR_CTS_32, mask_sh),\
+	SE_SF(DIG0_HDMI_ACR_32_1, HDMI_ACR_N_32, mask_sh),\
+	SE_SF(DIG0_HDMI_ACR_44_0, HDMI_ACR_CTS_44, mask_sh),\
+	SE_SF(DIG0_HDMI_ACR_44_1, HDMI_ACR_N_44, mask_sh),\
+	SE_SF(DIG0_HDMI_ACR_48_0, HDMI_ACR_CTS_48, mask_sh),\
+	SE_SF(DIG0_HDMI_ACR_48_1, HDMI_ACR_N_48, mask_sh),\
+	SE_SF(DP0_DP_SEC_AUD_N, DP_SEC_AUD_N, mask_sh),\
+	SE_SF(DP0_DP_SEC_TIMESTAMP, DP_SEC_TIMESTAMP_MODE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_ASP_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_ATP_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_AIP_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_ACM_ENABLE, mask_sh),\
+	SE_SF(DIG0_AFMT_CNTL, AFMT_AUDIO_CLOCK_EN, mask_sh),\
+	SE_SF(DIG0_HDMI_CONTROL, HDMI_CLOCK_CHANNEL_RATE, mask_sh),\
+	SE_SF(DIG0_DIG_FE_CNTL, TMDS_PIXEL_ENCODING, mask_sh),\
+	SE_SF(DIG0_DIG_FE_CNTL, TMDS_COLOR_FORMAT, mask_sh),\
+	SE_SF(DIG0_DIG_FE_CNTL, DIG_STEREOSYNC_SELECT, mask_sh),\
+	SE_SF(DIG0_DIG_FE_CNTL, DIG_STEREOSYNC_GATE_EN, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_GSP4_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_GSP5_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_GSP6_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_GSP7_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL2, DP_SEC_GSP7_SEND, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL6, DP_SEC_GSP7_LINE_NUM, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL2, DP_SEC_GSP11_PPS, mask_sh),\
+	SE_SF(DP0_DP_GSP11_CNTL, DP_SEC_GSP11_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_GSP11_CNTL, DP_SEC_GSP11_LINE_NUM, mask_sh),\
+	SE_SF(DP0_DP_DB_CNTL, DP_DB_DISABLE, mask_sh),\
+	SE_SF(DP0_DP_MSA_COLORIMETRY, DP_MSA_MISC0, mask_sh),\
+	SE_SF(DP0_DP_MSA_TIMING_PARAM1, DP_MSA_HTOTAL, mask_sh),\
+	SE_SF(DP0_DP_MSA_TIMING_PARAM1, DP_MSA_VTOTAL, mask_sh),\
+	SE_SF(DP0_DP_MSA_TIMING_PARAM2, DP_MSA_HSTART, mask_sh),\
+	SE_SF(DP0_DP_MSA_TIMING_PARAM2, DP_MSA_VSTART, mask_sh),\
+	SE_SF(DP0_DP_MSA_TIMING_PARAM3, DP_MSA_HSYNCWIDTH, mask_sh),\
+	SE_SF(DP0_DP_MSA_TIMING_PARAM3, DP_MSA_HSYNCPOLARITY, mask_sh),\
+	SE_SF(DP0_DP_MSA_TIMING_PARAM3, DP_MSA_VSYNCWIDTH, mask_sh),\
+	SE_SF(DP0_DP_MSA_TIMING_PARAM3, DP_MSA_VSYNCPOLARITY, mask_sh),\
+	SE_SF(DP0_DP_MSA_TIMING_PARAM4, DP_MSA_HWIDTH, mask_sh),\
+	SE_SF(DP0_DP_MSA_TIMING_PARAM4, DP_MSA_VHEIGHT, mask_sh),\
+	SE_SF(DIG0_HDMI_DB_CONTROL, HDMI_DB_DISABLE, mask_sh),\
+	SE_SF(DP0_DP_VID_TIMING, DP_VID_N_MUL, mask_sh),\
+	SE_SF(DIG0_DIG_FE_CNTL, DIG_SOURCE_SELECT, mask_sh), \
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC0_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC0_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC1_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC1_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC2_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC2_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC3_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC3_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC4_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC4_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC5_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC5_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC6_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC6_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC7_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC7_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC8_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC8_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC9_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC9_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC10_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC10_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC11_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC11_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC12_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC12_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC13_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC13_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC14_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC14_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL1, HDMI_GENERIC0_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL1, HDMI_GENERIC1_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL2, HDMI_GENERIC2_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL2, HDMI_GENERIC3_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL3, HDMI_GENERIC4_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL3, HDMI_GENERIC5_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL4, HDMI_GENERIC6_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL4, HDMI_GENERIC7_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL7, HDMI_GENERIC8_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL7, HDMI_GENERIC9_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL8, HDMI_GENERIC10_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL8, HDMI_GENERIC11_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL9, HDMI_GENERIC12_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL9, HDMI_GENERIC13_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL10, HDMI_GENERIC14_LINE, mask_sh),\
+	SE_SF(DP0_DP_DSC_CNTL, DP_DSC_MODE, mask_sh),\
+	SE_SF(DP0_DP_MSA_VBID_MISC, DP_VBID6_LINE_REFERENCE, mask_sh),\
+	SE_SF(DP0_DP_MSA_VBID_MISC, DP_VBID6_LINE_NUM, mask_sh),\
+	SE_SF(DME0_DME_CONTROL, METADATA_ENGINE_EN, mask_sh),\
+	SE_SF(DME0_DME_CONTROL, METADATA_HUBP_REQUESTOR_ID, mask_sh),\
+	SE_SF(DME0_DME_CONTROL, METADATA_STREAM_TYPE, mask_sh),\
+	SE_SF(DP0_DP_SEC_METADATA_TRANSMISSION, DP_SEC_METADATA_PACKET_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_METADATA_TRANSMISSION, DP_SEC_METADATA_PACKET_LINE_REFERENCE, mask_sh),\
+	SE_SF(DP0_DP_SEC_METADATA_TRANSMISSION, DP_SEC_METADATA_PACKET_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_METADATA_PACKET_CONTROL, HDMI_METADATA_PACKET_ENABLE, mask_sh),\
+	SE_SF(DIG0_HDMI_METADATA_PACKET_CONTROL, HDMI_METADATA_PACKET_LINE_REFERENCE, mask_sh),\
+	SE_SF(DIG0_HDMI_METADATA_PACKET_CONTROL, HDMI_METADATA_PACKET_LINE, mask_sh),\
+	SE_SF(DIG0_DIG_FE_CNTL, DOLBY_VISION_EN, mask_sh),\
+	SE_SF(DIG0_DIG_FE_CNTL, DIG_SYMCLK_FE_ON, mask_sh),\
+	SE_SF(DP0_DP_SEC_FRAMING4, DP_SST_SDP_SPLITTING, mask_sh),\
+	SE_SF(DIG0_DIG_CLOCK_PATTERN, DIG_CLOCK_PATTERN, mask_sh),\
+	SE_SF(DIG0_DIG_FIFO_CTRL0, DIG_FIFO_OUTPUT_PIXEL_MODE, mask_sh),\
+	SE_SF(DIG0_DIG_FIFO_CTRL0, DIG_FIFO_READ_START_LEVEL, mask_sh),\
+	SE_SF(DIG0_DIG_FIFO_CTRL0, DIG_FIFO_ENABLE, mask_sh),\
+	SE_SF(DIG0_DIG_FIFO_CTRL0, DIG_FIFO_RESET, mask_sh),\
+	SE_SF(DIG0_DIG_FIFO_CTRL0, DIG_FIFO_RESET_DONE, mask_sh)
+
+void dcn314_dio_stream_encoder_construct(
+	struct dcn10_stream_encoder *enc1,
+	struct dc_context *ctx,
+	struct dc_bios *bp,
+	enum engine_id eng_id,
+	struct vpg *vpg,
+	struct afmt *afmt,
+	const struct dcn10_stream_enc_registers *regs,
+	const struct dcn10_stream_encoder_shift *se_shift,
+	const struct dcn10_stream_encoder_mask *se_mask);
+
+void enc3_stream_encoder_update_hdmi_info_packets(
+	struct stream_encoder *enc,
+	const struct encoder_info_frame *info_frame);
+
+void enc3_stream_encoder_stop_hdmi_info_packets(
+	struct stream_encoder *enc);
+
+void enc3_stream_encoder_update_dp_info_packets_sdp_line_num(
+		struct stream_encoder *enc,
+		struct encoder_info_frame *info_frame);
+
+void enc3_stream_encoder_update_dp_info_packets(
+	struct stream_encoder *enc,
+	const struct encoder_info_frame *info_frame);
+
+void enc3_audio_mute_control(
+	struct stream_encoder *enc,
+	bool mute);
+
+void enc3_se_dp_audio_setup(
+	struct stream_encoder *enc,
+	unsigned int az_inst,
+	struct audio_info *info);
+
+void enc3_se_dp_audio_enable(
+	struct stream_encoder *enc);
+
+void enc3_se_hdmi_audio_setup(
+	struct stream_encoder *enc,
+	unsigned int az_inst,
+	struct audio_info *info,
+	struct audio_crtc_info *audio_crtc_info);
+
+void enc3_dp_set_dsc_pps_info_packet(
+	struct stream_encoder *enc,
+	bool enable,
+	uint8_t *dsc_packed_pps,
+	bool immediate_update);
+
+void enc314_stream_encoder_dvi_set_stream_attribute(
+	struct stream_encoder *enc,
+	struct dc_crtc_timing *crtc_timing,
+	bool is_dual_link);
+
+void enc314_stream_encoder_hdmi_set_stream_attribute(
+	struct stream_encoder *enc,
+	struct dc_crtc_timing *crtc_timing,
+	int actual_pix_clk_khz,
+	bool enable_audio);
+
+void enc314_stream_encoder_dp_blank(
+	struct dc_link *link,
+	struct stream_encoder *enc);
+
+void enc314_stream_encoder_dp_unblank(
+		struct dc_link *link,
+		struct stream_encoder *enc,
+		const struct encoder_unblank_param *param);
+
+void enc314_reset_fifo(struct stream_encoder *enc, bool reset);
+
+void enc314_enable_fifo(struct stream_encoder *enc);
+
+void enc314_disable_fifo(struct stream_encoder *enc);
+
+void enc314_set_dig_input_mode(struct stream_encoder *enc, unsigned int pix_per_container);
+
+void enc314_read_state(struct stream_encoder *enc, struct enc_state *s);
+
+void enc314_dp_set_odm_combine(
+	struct stream_encoder *enc,
+	bool odm_combine);
+
+void enc314_dp_set_dsc_config(
+	struct stream_encoder *enc,
+	enum optc_dsc_mode dsc_mode,
+	uint32_t dsc_bytes_per_pixel,
+	uint32_t dsc_slice_width);
+
+#endif /* __DC_DIO_STREAM_ENCODER_DCN314_H__ */
diff --git a/drivers/gpu/drm/amd/display/dc/dio/dcn32/dcn32_dio_link_encoder.c b/drivers/gpu/drm/amd/display/dc/dio/dcn32/dcn32_dio_link_encoder.c
new file mode 100644
index 000000000000..06907e8a4eda
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dio/dcn32/dcn32_dio_link_encoder.c
@@ -0,0 +1,328 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+
+#include "reg_helper.h"
+
+#include "core_types.h"
+#include "link_encoder.h"
+#include "dcn31/dcn31_dio_link_encoder.h"
+#include "dcn32_dio_link_encoder.h"
+#include "stream_encoder.h"
+#include "dc_bios_types.h"
+#include "link_enc_cfg.h"
+
+#include "dc_dmub_srv.h"
+#include "gpio_service_interface.h"
+
+#ifndef MIN
+#define MIN(X, Y) ((X) < (Y) ? (X) : (Y))
+#endif
+
+#define CTX \
+	enc10->base.ctx
+#define DC_LOGGER \
+	enc10->base.ctx->logger
+
+#define REG(reg)\
+	(enc10->link_regs->reg)
+
+#undef FN
+#define FN(reg_name, field_name) \
+	enc10->link_shift->field_name, enc10->link_mask->field_name
+
+#define AUX_REG(reg)\
+	(enc10->aux_regs->reg)
+
+#define AUX_REG_READ(reg_name) \
+		dm_read_reg(CTX, AUX_REG(reg_name))
+
+#define AUX_REG_WRITE(reg_name, val) \
+			dm_write_reg(CTX, AUX_REG(reg_name), val)
+
+static uint8_t phy_id_from_transmitter(enum transmitter t)
+{
+	uint8_t phy_id;
+
+	switch (t) {
+	case TRANSMITTER_UNIPHY_A:
+		phy_id = 0;
+		break;
+	case TRANSMITTER_UNIPHY_B:
+		phy_id = 1;
+		break;
+	case TRANSMITTER_UNIPHY_C:
+		phy_id = 2;
+		break;
+	case TRANSMITTER_UNIPHY_D:
+		phy_id = 3;
+		break;
+	case TRANSMITTER_UNIPHY_E:
+		phy_id = 4;
+		break;
+	case TRANSMITTER_UNIPHY_F:
+		phy_id = 5;
+		break;
+	case TRANSMITTER_UNIPHY_G:
+		phy_id = 6;
+		break;
+	default:
+		phy_id = 0;
+		break;
+	}
+	return phy_id;
+}
+
+void enc32_hw_init(struct link_encoder *enc)
+{
+	struct dcn10_link_encoder *enc10 = TO_DCN10_LINK_ENC(enc);
+
+/*
+	00 - DP_AUX_DPHY_RX_DETECTION_THRESHOLD__1to2 : 1/2
+	01 - DP_AUX_DPHY_RX_DETECTION_THRESHOLD__3to4 : 3/4
+	02 - DP_AUX_DPHY_RX_DETECTION_THRESHOLD__7to8 : 7/8
+	03 - DP_AUX_DPHY_RX_DETECTION_THRESHOLD__15to16 : 15/16
+	04 - DP_AUX_DPHY_RX_DETECTION_THRESHOLD__31to32 : 31/32
+	05 - DP_AUX_DPHY_RX_DETECTION_THRESHOLD__63to64 : 63/64
+	06 - DP_AUX_DPHY_RX_DETECTION_THRESHOLD__127to128 : 127/128
+	07 - DP_AUX_DPHY_RX_DETECTION_THRESHOLD__255to256 : 255/256
+*/
+
+/*
+	AUX_REG_UPDATE_5(AUX_DPHY_RX_CONTROL0,
+	AUX_RX_START_WINDOW = 1 [6:4]
+	AUX_RX_RECEIVE_WINDOW = 1 default is 2 [10:8]
+	AUX_RX_HALF_SYM_DETECT_LEN  = 1 [13:12] default is 1
+	AUX_RX_TRANSITION_FILTER_EN = 1 [16] default is 1
+	AUX_RX_ALLOW_BELOW_THRESHOLD_PHASE_DETECT [17] is 0  default is 0
+	AUX_RX_ALLOW_BELOW_THRESHOLD_START [18] is 1  default is 1
+	AUX_RX_ALLOW_BELOW_THRESHOLD_STOP [19] is 1  default is 1
+	AUX_RX_PHASE_DETECT_LEN,  [21,20] = 0x3 default is 3
+	AUX_RX_DETECTION_THRESHOLD [30:28] = 1
+*/
+	AUX_REG_WRITE(AUX_DPHY_RX_CONTROL0, 0x103d1110);
+
+	AUX_REG_WRITE(AUX_DPHY_TX_CONTROL, 0x21c7a);
+
+	//AUX_DPHY_TX_REF_CONTROL'AUX_TX_REF_DIV HW default is 0x32;
+	// Set AUX_TX_REF_DIV Divider to generate 2 MHz reference from refclk
+	// 27MHz -> 0xd
+	// 100MHz -> 0x32
+	// 48MHz -> 0x18
+
+	// Set TMDS_CTL0 to 1.  This is a legacy setting.
+	REG_UPDATE(TMDS_CTL_BITS, TMDS_CTL0, 1);
+
+	dcn10_aux_initialize(enc10);
+}
+
+
+void dcn32_link_encoder_enable_dp_output(
+	struct link_encoder *enc,
+	const struct dc_link_settings *link_settings,
+	enum clock_source_id clock_source)
+{
+	if (!enc->ctx->dc->debug.avoid_vbios_exec_table) {
+		dcn10_link_encoder_enable_dp_output(enc, link_settings, clock_source);
+		return;
+	}
+}
+
+static bool query_dp_alt_from_dmub(struct link_encoder *enc,
+	union dmub_rb_cmd *cmd)
+{
+	struct dcn10_link_encoder *enc10 = TO_DCN10_LINK_ENC(enc);
+
+	memset(cmd, 0, sizeof(*cmd));
+	cmd->query_dp_alt.header.type = DMUB_CMD__VBIOS;
+	cmd->query_dp_alt.header.sub_type =
+		DMUB_CMD__VBIOS_TRANSMITTER_QUERY_DP_ALT;
+	cmd->query_dp_alt.header.payload_bytes = sizeof(cmd->query_dp_alt.data);
+	cmd->query_dp_alt.data.phy_id = phy_id_from_transmitter(enc10->base.transmitter);
+
+	if (!dc_wake_and_execute_dmub_cmd(enc->ctx, cmd, DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY))
+		return false;
+
+	return true;
+}
+
+bool dcn32_link_encoder_is_in_alt_mode(struct link_encoder *enc)
+{
+	union dmub_rb_cmd cmd;
+
+	if (!query_dp_alt_from_dmub(enc, &cmd))
+		return false;
+
+	return (cmd.query_dp_alt.data.is_dp_alt_disable == 0);
+}
+
+void dcn32_link_encoder_get_max_link_cap(struct link_encoder *enc,
+	struct dc_link_settings *link_settings)
+{
+	union dmub_rb_cmd cmd;
+
+	dcn10_link_encoder_get_max_link_cap(enc, link_settings);
+
+	if (!query_dp_alt_from_dmub(enc, &cmd))
+		return;
+
+	if (cmd.query_dp_alt.data.is_usb &&
+			cmd.query_dp_alt.data.is_dp4 == 0)
+		link_settings->lane_count = MIN(LANE_COUNT_TWO, link_settings->lane_count);
+}
+
+
+static const struct link_encoder_funcs dcn32_link_enc_funcs = {
+	.read_state = link_enc2_read_state,
+	.validate_output_with_stream =
+			dcn30_link_encoder_validate_output_with_stream,
+	.hw_init = enc32_hw_init,
+	.setup = dcn10_link_encoder_setup,
+	.enable_tmds_output = dcn10_link_encoder_enable_tmds_output,
+	.enable_dp_output = dcn32_link_encoder_enable_dp_output,
+	.enable_dp_mst_output = dcn10_link_encoder_enable_dp_mst_output,
+	.disable_output = dcn10_link_encoder_disable_output,
+	.dp_set_lane_settings = dcn10_link_encoder_dp_set_lane_settings,
+	.dp_set_phy_pattern = dcn10_link_encoder_dp_set_phy_pattern,
+	.update_mst_stream_allocation_table =
+		dcn10_link_encoder_update_mst_stream_allocation_table,
+	.psr_program_dp_dphy_fast_training =
+			dcn10_psr_program_dp_dphy_fast_training,
+	.psr_program_secondary_packet = dcn10_psr_program_secondary_packet,
+	.connect_dig_be_to_fe = dcn10_link_encoder_connect_dig_be_to_fe,
+	.enable_hpd = dcn10_link_encoder_enable_hpd,
+	.disable_hpd = dcn10_link_encoder_disable_hpd,
+	.is_dig_enabled = dcn10_is_dig_enabled,
+	.destroy = dcn10_link_encoder_destroy,
+	.fec_set_enable = enc2_fec_set_enable,
+	.fec_set_ready = enc2_fec_set_ready,
+	.fec_is_active = enc2_fec_is_active,
+	.get_dig_frontend = dcn10_get_dig_frontend,
+	.get_dig_mode = dcn10_get_dig_mode,
+	.is_in_alt_mode = dcn32_link_encoder_is_in_alt_mode,
+	.get_max_link_cap = dcn32_link_encoder_get_max_link_cap,
+	.set_dio_phy_mux = dcn31_link_encoder_set_dio_phy_mux,
+};
+
+void dcn32_link_encoder_construct(
+	struct dcn20_link_encoder *enc20,
+	const struct encoder_init_data *init_data,
+	const struct encoder_feature_support *enc_features,
+	const struct dcn10_link_enc_registers *link_regs,
+	const struct dcn10_link_enc_aux_registers *aux_regs,
+	const struct dcn10_link_enc_hpd_registers *hpd_regs,
+	const struct dcn10_link_enc_shift *link_shift,
+	const struct dcn10_link_enc_mask *link_mask)
+{
+	struct bp_connector_speed_cap_info bp_cap_info = {0};
+	const struct dc_vbios_funcs *bp_funcs = init_data->ctx->dc_bios->funcs;
+	enum bp_result result = BP_RESULT_OK;
+	struct dcn10_link_encoder *enc10 = &enc20->enc10;
+
+	enc10->base.funcs = &dcn32_link_enc_funcs;
+	enc10->base.ctx = init_data->ctx;
+	enc10->base.id = init_data->encoder;
+
+	enc10->base.hpd_source = init_data->hpd_source;
+	enc10->base.connector = init_data->connector;
+
+	enc10->base.preferred_engine = ENGINE_ID_UNKNOWN;
+
+	enc10->base.features = *enc_features;
+
+	enc10->base.transmitter = init_data->transmitter;
+
+	/* set the flag to indicate whether driver poll the I2C data pin
+	 * while doing the DP sink detect
+	 */
+
+/*	if (dal_adapter_service_is_feature_supported(as,
+		FEATURE_DP_SINK_DETECT_POLL_DATA_PIN))
+		enc10->base.features.flags.bits.
+			DP_SINK_DETECT_POLL_DATA_PIN = true;*/
+
+	enc10->base.output_signals =
+		SIGNAL_TYPE_DVI_SINGLE_LINK |
+		SIGNAL_TYPE_DVI_DUAL_LINK |
+		SIGNAL_TYPE_LVDS |
+		SIGNAL_TYPE_DISPLAY_PORT |
+		SIGNAL_TYPE_DISPLAY_PORT_MST |
+		SIGNAL_TYPE_EDP |
+		SIGNAL_TYPE_HDMI_TYPE_A;
+
+	enc10->link_regs = link_regs;
+	enc10->aux_regs = aux_regs;
+	enc10->hpd_regs = hpd_regs;
+	enc10->link_shift = link_shift;
+	enc10->link_mask = link_mask;
+
+	switch (enc10->base.transmitter) {
+	case TRANSMITTER_UNIPHY_A:
+		enc10->base.preferred_engine = ENGINE_ID_DIGA;
+	break;
+	case TRANSMITTER_UNIPHY_B:
+		enc10->base.preferred_engine = ENGINE_ID_DIGB;
+	break;
+	case TRANSMITTER_UNIPHY_C:
+		enc10->base.preferred_engine = ENGINE_ID_DIGC;
+	break;
+	case TRANSMITTER_UNIPHY_D:
+		enc10->base.preferred_engine = ENGINE_ID_DIGD;
+	break;
+	case TRANSMITTER_UNIPHY_E:
+		enc10->base.preferred_engine = ENGINE_ID_DIGE;
+	break;
+	default:
+		ASSERT_CRITICAL(false);
+		enc10->base.preferred_engine = ENGINE_ID_UNKNOWN;
+	}
+
+	/* default to one to mirror Windows behavior */
+	enc10->base.features.flags.bits.HDMI_6GB_EN = 1;
+
+	if (bp_funcs->get_connector_speed_cap_info)
+		result = bp_funcs->get_connector_speed_cap_info(enc10->base.ctx->dc_bios,
+						enc10->base.connector, &bp_cap_info);
+
+	/* Override features with DCE-specific values */
+	if (result == BP_RESULT_OK) {
+		enc10->base.features.flags.bits.IS_HBR2_CAPABLE =
+				bp_cap_info.DP_HBR2_EN;
+		enc10->base.features.flags.bits.IS_HBR3_CAPABLE =
+				bp_cap_info.DP_HBR3_EN;
+		enc10->base.features.flags.bits.HDMI_6GB_EN = bp_cap_info.HDMI_6GB_EN;
+		enc10->base.features.flags.bits.IS_DP2_CAPABLE = 1;
+		enc10->base.features.flags.bits.IS_UHBR10_CAPABLE = bp_cap_info.DP_UHBR10_EN;
+		enc10->base.features.flags.bits.IS_UHBR13_5_CAPABLE = bp_cap_info.DP_UHBR13_5_EN;
+		enc10->base.features.flags.bits.IS_UHBR20_CAPABLE = bp_cap_info.DP_UHBR20_EN;
+	} else {
+		DC_LOG_WARNING("%s: Failed to get encoder_cap_info from VBIOS with error code %d!\n",
+				__func__,
+				result);
+	}
+	if (enc10->base.ctx->dc->debug.hdmi20_disable) {
+		enc10->base.features.flags.bits.HDMI_6GB_EN = 0;
+	}
+}
diff --git a/drivers/gpu/drm/amd/display/dc/dio/dcn32/dcn32_dio_link_encoder.h b/drivers/gpu/drm/amd/display/dc/dio/dcn32/dcn32_dio_link_encoder.h
new file mode 100644
index 000000000000..35d23d9db45e
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dio/dcn32/dcn32_dio_link_encoder.h
@@ -0,0 +1,53 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ *  and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __DC_LINK_ENCODER__DCN32_H__
+#define __DC_LINK_ENCODER__DCN32_H__
+
+#include "dcn30/dcn30_dio_link_encoder.h"
+
+void dcn32_link_encoder_construct(
+	struct dcn20_link_encoder *enc20,
+	const struct encoder_init_data *init_data,
+	const struct encoder_feature_support *enc_features,
+	const struct dcn10_link_enc_registers *link_regs,
+	const struct dcn10_link_enc_aux_registers *aux_regs,
+	const struct dcn10_link_enc_hpd_registers *hpd_regs,
+	const struct dcn10_link_enc_shift *link_shift,
+	const struct dcn10_link_enc_mask *link_mask);
+
+void enc32_hw_init(struct link_encoder *enc);
+
+void dcn32_link_encoder_enable_dp_output(
+	struct link_encoder *enc,
+	const struct dc_link_settings *link_settings,
+	enum clock_source_id clock_source);
+
+bool dcn32_link_encoder_is_in_alt_mode(struct link_encoder *enc);
+
+void dcn32_link_encoder_get_max_link_cap(struct link_encoder *enc,
+	struct dc_link_settings *link_settings);
+
+#endif /* __DC_LINK_ENCODER__DCN32_H__ */
diff --git a/drivers/gpu/drm/amd/display/dc/dio/dcn32/dcn32_dio_stream_encoder.c b/drivers/gpu/drm/amd/display/dc/dio/dcn32/dcn32_dio_stream_encoder.c
new file mode 100644
index 000000000000..3523d1cdc1a3
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dio/dcn32/dcn32_dio_stream_encoder.c
@@ -0,0 +1,493 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ *  and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+
+#include "dc_bios_types.h"
+#include "dcn30/dcn30_dio_stream_encoder.h"
+#include "dcn32_dio_stream_encoder.h"
+#include "reg_helper.h"
+#include "hw_shared.h"
+#include "link_service.h"
+#include "dpcd_defs.h"
+
+#define DC_LOGGER \
+		enc1->base.ctx->logger
+
+#define REG(reg)\
+	(enc1->regs->reg)
+
+#undef FN
+#define FN(reg_name, field_name) \
+	enc1->se_shift->field_name, enc1->se_mask->field_name
+
+#define VBI_LINE_0 0
+#define HDMI_CLOCK_CHANNEL_RATE_MORE_340M 340000
+
+#define CTX \
+	enc1->base.ctx
+
+
+
+static void enc32_dp_set_odm_combine(
+	struct stream_encoder *enc,
+	bool two_pixel_per_cyle)
+{
+	struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc);
+
+	REG_UPDATE(DP_PIXEL_FORMAT, DP_PIXEL_PER_CYCLE_PROCESSING_MODE, two_pixel_per_cyle ? 1 : 0);
+}
+
+/* setup stream encoder in dvi mode */
+static void enc32_stream_encoder_dvi_set_stream_attribute(
+	struct stream_encoder *enc,
+	struct dc_crtc_timing *crtc_timing,
+	bool is_dual_link)
+{
+	struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc);
+
+	if (!enc->ctx->dc->debug.avoid_vbios_exec_table) {
+		struct bp_encoder_control cntl = {0};
+
+		cntl.action = ENCODER_CONTROL_SETUP;
+		cntl.engine_id = enc1->base.id;
+		cntl.signal = is_dual_link ?
+			SIGNAL_TYPE_DVI_DUAL_LINK : SIGNAL_TYPE_DVI_SINGLE_LINK;
+		cntl.enable_dp_audio = false;
+		cntl.pixel_clock = crtc_timing->pix_clk_100hz / 10;
+		cntl.lanes_number = (is_dual_link) ? LANE_COUNT_EIGHT : LANE_COUNT_FOUR;
+
+		if (enc1->base.bp->funcs->encoder_control(
+				enc1->base.bp, &cntl) != BP_RESULT_OK)
+			return;
+
+	} else {
+
+		//Set pattern for clock channel, default vlue 0x63 does not work
+		REG_UPDATE(DIG_CLOCK_PATTERN, DIG_CLOCK_PATTERN, 0x1F);
+
+		//DIG_BE_TMDS_DVI_MODE : TMDS-DVI mode is already set in link_encoder_setup
+
+		//DIG_SOURCE_SELECT is already set in dig_connect_to_otg
+
+		/* DIG_START is removed from the register spec */
+	}
+
+	ASSERT(crtc_timing->pixel_encoding == PIXEL_ENCODING_RGB);
+	ASSERT(crtc_timing->display_color_depth == COLOR_DEPTH_888);
+	enc1_stream_encoder_set_stream_attribute_helper(enc1, crtc_timing);
+}
+
+/* setup stream encoder in hdmi mode */
+static void enc32_stream_encoder_hdmi_set_stream_attribute(
+	struct stream_encoder *enc,
+	struct dc_crtc_timing *crtc_timing,
+	int actual_pix_clk_khz,
+	bool enable_audio)
+{
+	struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc);
+
+	if (!enc->ctx->dc->debug.avoid_vbios_exec_table) {
+		struct bp_encoder_control cntl = {0};
+
+		cntl.action = ENCODER_CONTROL_SETUP;
+		cntl.engine_id = enc1->base.id;
+		cntl.signal = SIGNAL_TYPE_HDMI_TYPE_A;
+		cntl.enable_dp_audio = enable_audio;
+		cntl.pixel_clock = actual_pix_clk_khz;
+		cntl.lanes_number = LANE_COUNT_FOUR;
+
+		if (enc1->base.bp->funcs->encoder_control(
+				enc1->base.bp, &cntl) != BP_RESULT_OK)
+			return;
+
+	} else {
+
+		//Set pattern for clock channel, default vlue 0x63 does not work
+		REG_UPDATE(DIG_CLOCK_PATTERN, DIG_CLOCK_PATTERN, 0x1F);
+
+		//DIG_BE_TMDS_HDMI_MODE : TMDS-HDMI mode is already set in link_encoder_setup
+
+		//DIG_SOURCE_SELECT is already set in dig_connect_to_otg
+
+		/* DIG_START is removed from the register spec */
+	}
+
+	/* Configure pixel encoding */
+	enc1_stream_encoder_set_stream_attribute_helper(enc1, crtc_timing);
+
+	/* setup HDMI engine */
+	REG_UPDATE_6(HDMI_CONTROL,
+		HDMI_PACKET_GEN_VERSION, 1,
+		HDMI_KEEPOUT_MODE, 1,
+		HDMI_DEEP_COLOR_ENABLE, 0,
+		HDMI_DATA_SCRAMBLE_EN, 0,
+		HDMI_NO_EXTRA_NULL_PACKET_FILLED, 1,
+		HDMI_CLOCK_CHANNEL_RATE, 0);
+
+	/* Configure color depth */
+	switch (crtc_timing->display_color_depth) {
+	case COLOR_DEPTH_888:
+		REG_UPDATE(HDMI_CONTROL, HDMI_DEEP_COLOR_DEPTH, 0);
+		break;
+	case COLOR_DEPTH_101010:
+		if (crtc_timing->pixel_encoding == PIXEL_ENCODING_YCBCR422) {
+			REG_UPDATE_2(HDMI_CONTROL,
+					HDMI_DEEP_COLOR_DEPTH, 1,
+					HDMI_DEEP_COLOR_ENABLE, 0);
+		} else {
+			REG_UPDATE_2(HDMI_CONTROL,
+					HDMI_DEEP_COLOR_DEPTH, 1,
+					HDMI_DEEP_COLOR_ENABLE, 1);
+			}
+		break;
+	case COLOR_DEPTH_121212:
+		if (crtc_timing->pixel_encoding == PIXEL_ENCODING_YCBCR422) {
+			REG_UPDATE_2(HDMI_CONTROL,
+					HDMI_DEEP_COLOR_DEPTH, 2,
+					HDMI_DEEP_COLOR_ENABLE, 0);
+		} else {
+			REG_UPDATE_2(HDMI_CONTROL,
+					HDMI_DEEP_COLOR_DEPTH, 2,
+					HDMI_DEEP_COLOR_ENABLE, 1);
+			}
+		break;
+	case COLOR_DEPTH_161616:
+		REG_UPDATE_2(HDMI_CONTROL,
+				HDMI_DEEP_COLOR_DEPTH, 3,
+				HDMI_DEEP_COLOR_ENABLE, 1);
+		break;
+	default:
+		break;
+	}
+
+	if (actual_pix_clk_khz >= HDMI_CLOCK_CHANNEL_RATE_MORE_340M) {
+		/* enable HDMI data scrambler
+		 * HDMI_CLOCK_CHANNEL_RATE_MORE_340M
+		 * Clock channel frequency is 1/4 of character rate.
+		 */
+		REG_UPDATE_2(HDMI_CONTROL,
+			HDMI_DATA_SCRAMBLE_EN, 1,
+			HDMI_CLOCK_CHANNEL_RATE, 1);
+	} else if (crtc_timing->flags.LTE_340MCSC_SCRAMBLE) {
+
+		/* TODO: New feature for DCE11, still need to implement */
+
+		/* enable HDMI data scrambler
+		 * HDMI_CLOCK_CHANNEL_FREQ_EQUAL_TO_CHAR_RATE
+		 * Clock channel frequency is the same
+		 * as character rate
+		 */
+		REG_UPDATE_2(HDMI_CONTROL,
+			HDMI_DATA_SCRAMBLE_EN, 1,
+			HDMI_CLOCK_CHANNEL_RATE, 0);
+	}
+
+
+	/* Enable transmission of General Control packet on every frame */
+	REG_UPDATE_3(HDMI_VBI_PACKET_CONTROL,
+		HDMI_GC_CONT, 1,
+		HDMI_GC_SEND, 1,
+		HDMI_NULL_SEND, 1);
+
+	/* Disable Audio Content Protection packet transmission */
+	REG_UPDATE(HDMI_VBI_PACKET_CONTROL, HDMI_ACP_SEND, 0);
+
+	/* following belongs to audio */
+	/* Enable Audio InfoFrame packet transmission. */
+	REG_UPDATE(HDMI_INFOFRAME_CONTROL0, HDMI_AUDIO_INFO_SEND, 1);
+
+	/* update double-buffered AUDIO_INFO registers immediately */
+	ASSERT(enc->afmt);
+	enc->afmt->funcs->audio_info_immediate_update(enc->afmt);
+
+	/* Select line number on which to send Audio InfoFrame packets */
+	REG_UPDATE(HDMI_INFOFRAME_CONTROL1, HDMI_AUDIO_INFO_LINE,
+				VBI_LINE_0 + 2);
+
+	/* set HDMI GC AVMUTE */
+	REG_UPDATE(HDMI_GC, HDMI_GC_AVMUTE, 0);
+}
+
+
+
+static bool is_two_pixels_per_containter(const struct dc_crtc_timing *timing)
+{
+	bool two_pix = timing->pixel_encoding == PIXEL_ENCODING_YCBCR420;
+
+	two_pix = two_pix || (timing->flags.DSC && timing->pixel_encoding == PIXEL_ENCODING_YCBCR422
+			&& !timing->dsc_cfg.ycbcr422_simple);
+	return two_pix;
+}
+
+void enc32_stream_encoder_dp_unblank(
+	struct dc_link *link,
+	struct stream_encoder *enc,
+	const struct encoder_unblank_param *param)
+{
+	struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc);
+
+	if (param->link_settings.link_rate != LINK_RATE_UNKNOWN) {
+		uint32_t n_vid = 0x8000;
+		uint32_t m_vid;
+		uint32_t n_multiply = 0;
+		uint32_t pix_per_cycle = 0;
+		uint64_t m_vid_l = n_vid;
+
+		/* YCbCr 4:2:0 : Computed VID_M will be 2X the input rate */
+		if (is_two_pixels_per_containter(&param->timing) || param->opp_cnt > 1
+			|| param->pix_per_cycle > 1) {
+			/*this logic should be the same in get_pixel_clock_parameters() */
+			n_multiply = 1;
+			pix_per_cycle = 1;
+		}
+		/* M / N = Fstream / Flink
+		 * m_vid / n_vid = pixel rate / link rate
+		 */
+
+		m_vid_l *= param->timing.pix_clk_100hz / 10;
+		m_vid_l = div_u64(m_vid_l,
+			param->link_settings.link_rate
+				* LINK_RATE_REF_FREQ_IN_KHZ);
+
+		m_vid = (uint32_t) m_vid_l;
+
+		/* enable auto measurement */
+
+		REG_UPDATE(DP_VID_TIMING, DP_VID_M_N_GEN_EN, 0);
+
+		/* auto measurement need 1 full 0x8000 symbol cycle to kick in,
+		 * therefore program initial value for Mvid and Nvid
+		 */
+
+		REG_UPDATE(DP_VID_N, DP_VID_N, n_vid);
+
+		REG_UPDATE(DP_VID_M, DP_VID_M, m_vid);
+
+		REG_UPDATE_2(DP_VID_TIMING,
+				DP_VID_M_N_GEN_EN, 1,
+				DP_VID_N_MUL, n_multiply);
+
+		REG_UPDATE(DP_PIXEL_FORMAT,
+				DP_PIXEL_PER_CYCLE_PROCESSING_MODE,
+				pix_per_cycle);
+	}
+
+	/* make sure stream is disabled before resetting steer fifo */
+	REG_UPDATE(DP_VID_STREAM_CNTL, DP_VID_STREAM_ENABLE, false);
+	REG_WAIT(DP_VID_STREAM_CNTL, DP_VID_STREAM_STATUS, 0, 10, 5000);
+
+	/* DIG_START is removed from the register spec */
+
+	/* switch DP encoder to CRTC data, but reset it the fifo first. It may happen
+	 * that it overflows during mode transition, and sometimes doesn't recover.
+	 */
+	REG_UPDATE(DP_STEER_FIFO, DP_STEER_FIFO_RESET, 1);
+	udelay(10);
+
+	REG_UPDATE(DP_STEER_FIFO, DP_STEER_FIFO_RESET, 0);
+
+	/* DIG Resync FIFO now needs to be explicitly enabled
+	 */
+	// TODO: Confirm if we need to wait for DIG_SYMCLK_FE_ON
+	REG_WAIT(DIG_FE_CNTL, DIG_SYMCLK_FE_ON, 1, 10, 5000);
+
+	/* read start level = 0 will bring underflow / overflow and DIG_FIFO_ERROR = 1
+	 * so set it to 1/2 full = 7 before reset as suggested by hardware team.
+	 */
+	REG_UPDATE(DIG_FIFO_CTRL0, DIG_FIFO_READ_START_LEVEL, 0x7);
+
+	REG_UPDATE(DIG_FIFO_CTRL0, DIG_FIFO_RESET, 1);
+
+	REG_WAIT(DIG_FIFO_CTRL0, DIG_FIFO_RESET_DONE, 1, 10, 5000);
+
+	REG_UPDATE(DIG_FIFO_CTRL0, DIG_FIFO_RESET, 0);
+
+	REG_WAIT(DIG_FIFO_CTRL0, DIG_FIFO_RESET_DONE, 0, 10, 5000);
+
+	REG_UPDATE(DIG_FIFO_CTRL0, DIG_FIFO_ENABLE, 1);
+
+	/* wait 100us for DIG/DP logic to prime
+	 * (i.e. a few video lines)
+	 */
+	udelay(100);
+
+	/* the hardware would start sending video at the start of the next DP
+	 * frame (i.e. rising edge of the vblank).
+	 * NOTE: We used to program DP_VID_STREAM_DIS_DEFER = 2 here, but this
+	 * register has no effect on enable transition! HW always guarantees
+	 * VID_STREAM enable at start of next frame, and this is not
+	 * programmable
+	 */
+
+	REG_UPDATE(DP_VID_STREAM_CNTL, DP_VID_STREAM_ENABLE, true);
+
+	link->dc->link_srv->dp_trace_source_sequence(link, DPCD_SOURCE_SEQ_AFTER_ENABLE_DP_VID_STREAM);
+}
+
+/* Set DSC-related configuration.
+ *   dsc_mode: 0 disables DSC, other values enable DSC in specified format
+ *   sc_bytes_per_pixel: DP_DSC_BYTES_PER_PIXEL removed in DCN32
+ *   dsc_slice_width: DP_DSC_SLICE_WIDTH removed in DCN32
+ */
+static void enc32_dp_set_dsc_config(struct stream_encoder *enc,
+					enum optc_dsc_mode dsc_mode,
+					uint32_t dsc_bytes_per_pixel,
+					uint32_t dsc_slice_width)
+{
+	struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc);
+
+	REG_UPDATE(DP_DSC_CNTL,	DP_DSC_MODE, dsc_mode == OPTC_DSC_DISABLED ? 0 : 1);
+}
+
+/* this function read dsc related register fields to be logged later in dcn10_log_hw_state
+ * into a dcn_dsc_state struct.
+ */
+static void enc32_read_state(struct stream_encoder *enc, struct enc_state *s)
+{
+	struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc);
+
+	//if dsc is enabled, continue to read
+	REG_GET(DP_DSC_CNTL, DP_DSC_MODE, &s->dsc_mode);
+	if (s->dsc_mode) {
+		REG_GET(DP_GSP11_CNTL, DP_SEC_GSP11_LINE_NUM, &s->sec_gsp_pps_line_num);
+
+		REG_GET(DP_MSA_VBID_MISC, DP_VBID6_LINE_REFERENCE, &s->vbid6_line_reference);
+		REG_GET(DP_MSA_VBID_MISC, DP_VBID6_LINE_NUM, &s->vbid6_line_num);
+
+		REG_GET(DP_GSP11_CNTL, DP_SEC_GSP11_ENABLE, &s->sec_gsp_pps_enable);
+		REG_GET(DP_SEC_CNTL, DP_SEC_STREAM_ENABLE, &s->sec_stream_enable);
+	}
+}
+
+static void enc32_set_dig_input_mode(struct stream_encoder *enc, unsigned int pix_per_container)
+{
+	struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc);
+
+	/* The naming of this field is confusing, what it means is the output mode of otg, which
+	 * is the input mode of the dig
+	 */
+	REG_UPDATE(DIG_FIFO_CTRL0, DIG_FIFO_OUTPUT_PIXEL_MODE, pix_per_container == 2 ? 0x1 : 0x0);
+}
+
+static void enc32_reset_fifo(struct stream_encoder *enc, bool reset)
+{
+	struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc);
+	uint32_t reset_val = reset ? 1 : 0;
+	uint32_t is_symclk_on;
+
+	REG_UPDATE(DIG_FIFO_CTRL0, DIG_FIFO_RESET, reset_val);
+	REG_GET(DIG_FE_CNTL, DIG_SYMCLK_FE_ON, &is_symclk_on);
+
+	if (is_symclk_on)
+		REG_WAIT(DIG_FIFO_CTRL0, DIG_FIFO_RESET_DONE, reset_val, 10, 5000);
+	else
+		udelay(10);
+}
+
+void enc32_enable_fifo(struct stream_encoder *enc)
+{
+	struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc);
+
+	REG_UPDATE(DIG_FIFO_CTRL0, DIG_FIFO_READ_START_LEVEL, 0x7);
+
+	enc32_reset_fifo(enc, true);
+	enc32_reset_fifo(enc, false);
+
+	REG_UPDATE(DIG_FIFO_CTRL0, DIG_FIFO_ENABLE, 1);
+}
+
+static const struct stream_encoder_funcs dcn32_str_enc_funcs = {
+	.dp_set_odm_combine =
+		enc32_dp_set_odm_combine,
+	.dp_set_stream_attribute =
+		enc2_stream_encoder_dp_set_stream_attribute,
+	.hdmi_set_stream_attribute =
+		enc32_stream_encoder_hdmi_set_stream_attribute,
+	.dvi_set_stream_attribute =
+		enc32_stream_encoder_dvi_set_stream_attribute,
+	.set_throttled_vcp_size =
+		enc1_stream_encoder_set_throttled_vcp_size,
+	.update_hdmi_info_packets =
+		enc3_stream_encoder_update_hdmi_info_packets,
+	.stop_hdmi_info_packets =
+		enc3_stream_encoder_stop_hdmi_info_packets,
+	.update_dp_info_packets_sdp_line_num =
+		enc3_stream_encoder_update_dp_info_packets_sdp_line_num,
+	.update_dp_info_packets =
+		enc3_stream_encoder_update_dp_info_packets,
+	.stop_dp_info_packets =
+		enc1_stream_encoder_stop_dp_info_packets,
+	.dp_blank =
+		enc1_stream_encoder_dp_blank,
+	.dp_unblank =
+		enc32_stream_encoder_dp_unblank,
+	.audio_mute_control = enc3_audio_mute_control,
+
+	.dp_audio_setup = enc3_se_dp_audio_setup,
+	.dp_audio_enable = enc3_se_dp_audio_enable,
+	.dp_audio_disable = enc1_se_dp_audio_disable,
+
+	.hdmi_audio_setup = enc3_se_hdmi_audio_setup,
+	.hdmi_audio_disable = enc1_se_hdmi_audio_disable,
+	.setup_stereo_sync  = enc1_setup_stereo_sync,
+	.set_avmute = enc1_stream_encoder_set_avmute,
+	.dig_connect_to_otg = enc1_dig_connect_to_otg,
+	.dig_source_otg = enc1_dig_source_otg,
+
+	.dp_get_pixel_format  = enc1_stream_encoder_dp_get_pixel_format,
+
+	.enc_read_state = enc32_read_state,
+	.dp_set_dsc_config = enc32_dp_set_dsc_config,
+	.dp_set_dsc_pps_info_packet = enc3_dp_set_dsc_pps_info_packet,
+	.set_dynamic_metadata = enc2_set_dynamic_metadata,
+	.hdmi_reset_stream_attribute = enc1_reset_hdmi_stream_attribute,
+
+	.set_input_mode = enc32_set_dig_input_mode,
+	.enable_fifo = enc32_enable_fifo,
+};
+
+void dcn32_dio_stream_encoder_construct(
+	struct dcn10_stream_encoder *enc1,
+	struct dc_context *ctx,
+	struct dc_bios *bp,
+	enum engine_id eng_id,
+	struct vpg *vpg,
+	struct afmt *afmt,
+	const struct dcn10_stream_enc_registers *regs,
+	const struct dcn10_stream_encoder_shift *se_shift,
+	const struct dcn10_stream_encoder_mask *se_mask)
+{
+	enc1->base.funcs = &dcn32_str_enc_funcs;
+	enc1->base.ctx = ctx;
+	enc1->base.id = eng_id;
+	enc1->base.bp = bp;
+	enc1->base.vpg = vpg;
+	enc1->base.afmt = afmt;
+	enc1->regs = regs;
+	enc1->se_shift = se_shift;
+	enc1->se_mask = se_mask;
+	enc1->base.stream_enc_inst = vpg->inst;
+}
+
diff --git a/drivers/gpu/drm/amd/display/dc/dio/dcn32/dcn32_dio_stream_encoder.h b/drivers/gpu/drm/amd/display/dc/dio/dcn32/dcn32_dio_stream_encoder.h
new file mode 100644
index 000000000000..ca53d39561d2
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dio/dcn32/dcn32_dio_stream_encoder.h
@@ -0,0 +1,206 @@
+/*
+ * Copyright 2021 - Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ *  and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __DC_DIO_STREAM_ENCODER_DCN32_H__
+#define __DC_DIO_STREAM_ENCODER_DCN32_H__
+
+#include "dcn30/dcn30_vpg.h"
+#include "dcn30/dcn30_afmt.h"
+#include "stream_encoder.h"
+#include "dcn20/dcn20_stream_encoder.h"
+
+#define SE_COMMON_MASK_SH_LIST_DCN32(mask_sh)\
+	SE_SF(DP0_DP_PIXEL_FORMAT, DP_PIXEL_ENCODING, mask_sh),\
+	SE_SF(DP0_DP_PIXEL_FORMAT, DP_COMPONENT_DEPTH, mask_sh),\
+	SE_SF(DP0_DP_PIXEL_FORMAT, DP_PIXEL_PER_CYCLE_PROCESSING_MODE, mask_sh),\
+	SE_SF(DIG0_HDMI_CONTROL, HDMI_PACKET_GEN_VERSION, mask_sh),\
+	SE_SF(DIG0_HDMI_CONTROL, HDMI_KEEPOUT_MODE, mask_sh),\
+	SE_SF(DIG0_HDMI_CONTROL, HDMI_DEEP_COLOR_ENABLE, mask_sh),\
+	SE_SF(DIG0_HDMI_CONTROL, HDMI_DEEP_COLOR_DEPTH, mask_sh),\
+	SE_SF(DIG0_HDMI_CONTROL, HDMI_DATA_SCRAMBLE_EN, mask_sh),\
+	SE_SF(DIG0_HDMI_CONTROL, HDMI_NO_EXTRA_NULL_PACKET_FILLED, mask_sh),\
+	SE_SF(DIG0_HDMI_VBI_PACKET_CONTROL, HDMI_GC_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_VBI_PACKET_CONTROL, HDMI_GC_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_VBI_PACKET_CONTROL, HDMI_NULL_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_VBI_PACKET_CONTROL, HDMI_ACP_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_INFOFRAME_CONTROL0, HDMI_AUDIO_INFO_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_INFOFRAME_CONTROL1, HDMI_AUDIO_INFO_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GC, HDMI_GC_AVMUTE, mask_sh),\
+	SE_SF(DP0_DP_MSE_RATE_CNTL, DP_MSE_RATE_X, mask_sh),\
+	SE_SF(DP0_DP_MSE_RATE_CNTL, DP_MSE_RATE_Y, mask_sh),\
+	SE_SF(DP0_DP_MSE_RATE_UPDATE, DP_MSE_RATE_UPDATE_PENDING, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_GSP0_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_STREAM_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_GSP1_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_GSP2_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_GSP3_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_MPG_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL1, DP_SEC_GSP5_LINE_REFERENCE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL2, DP_SEC_GSP4_SEND, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL2, DP_SEC_GSP4_SEND_PENDING, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL4, DP_SEC_GSP4_LINE_NUM, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL5, DP_SEC_GSP5_LINE_NUM, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL2, DP_SEC_GSP4_SEND_ANY_LINE, mask_sh),\
+	SE_SF(DP0_DP_VID_STREAM_CNTL, DP_VID_STREAM_DIS_DEFER, mask_sh),\
+	SE_SF(DP0_DP_VID_STREAM_CNTL, DP_VID_STREAM_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_VID_STREAM_CNTL, DP_VID_STREAM_STATUS, mask_sh),\
+	SE_SF(DP0_DP_STEER_FIFO, DP_STEER_FIFO_RESET, mask_sh),\
+	SE_SF(DP0_DP_VID_TIMING, DP_VID_M_N_GEN_EN, mask_sh),\
+	SE_SF(DP0_DP_VID_N, DP_VID_N, mask_sh),\
+	SE_SF(DP0_DP_VID_M, DP_VID_M, mask_sh),\
+	SE_SF(DIG0_HDMI_AUDIO_PACKET_CONTROL, HDMI_AUDIO_DELAY_EN, mask_sh),\
+	SE_SF(DIG0_HDMI_ACR_PACKET_CONTROL, HDMI_ACR_AUTO_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_ACR_PACKET_CONTROL, HDMI_ACR_SOURCE, mask_sh),\
+	SE_SF(DIG0_HDMI_ACR_PACKET_CONTROL, HDMI_ACR_AUDIO_PRIORITY, mask_sh),\
+	SE_SF(DIG0_HDMI_ACR_32_0, HDMI_ACR_CTS_32, mask_sh),\
+	SE_SF(DIG0_HDMI_ACR_32_1, HDMI_ACR_N_32, mask_sh),\
+	SE_SF(DIG0_HDMI_ACR_44_0, HDMI_ACR_CTS_44, mask_sh),\
+	SE_SF(DIG0_HDMI_ACR_44_1, HDMI_ACR_N_44, mask_sh),\
+	SE_SF(DIG0_HDMI_ACR_48_0, HDMI_ACR_CTS_48, mask_sh),\
+	SE_SF(DIG0_HDMI_ACR_48_1, HDMI_ACR_N_48, mask_sh),\
+	SE_SF(DP0_DP_SEC_AUD_N, DP_SEC_AUD_N, mask_sh),\
+	SE_SF(DP0_DP_SEC_TIMESTAMP, DP_SEC_TIMESTAMP_MODE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_ASP_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_ATP_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_AIP_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_ACM_ENABLE, mask_sh),\
+	SE_SF(DIG0_AFMT_CNTL, AFMT_AUDIO_CLOCK_EN, mask_sh),\
+	SE_SF(DIG0_HDMI_CONTROL, HDMI_CLOCK_CHANNEL_RATE, mask_sh),\
+	SE_SF(DIG0_DIG_FE_CNTL, TMDS_PIXEL_ENCODING, mask_sh),\
+	SE_SF(DIG0_DIG_FE_CNTL, TMDS_COLOR_FORMAT, mask_sh),\
+	SE_SF(DIG0_DIG_FE_CNTL, DIG_STEREOSYNC_SELECT, mask_sh),\
+	SE_SF(DIG0_DIG_FE_CNTL, DIG_STEREOSYNC_GATE_EN, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_GSP4_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_GSP5_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_GSP6_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_GSP7_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL2, DP_SEC_GSP7_SEND, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL6, DP_SEC_GSP7_LINE_NUM, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL2, DP_SEC_GSP11_PPS, mask_sh),\
+	SE_SF(DP0_DP_GSP11_CNTL, DP_SEC_GSP11_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_GSP11_CNTL, DP_SEC_GSP11_LINE_NUM, mask_sh),\
+	SE_SF(DP0_DP_DB_CNTL, DP_DB_DISABLE, mask_sh),\
+	SE_SF(DP0_DP_MSA_COLORIMETRY, DP_MSA_MISC0, mask_sh),\
+	SE_SF(DP0_DP_MSA_TIMING_PARAM1, DP_MSA_HTOTAL, mask_sh),\
+	SE_SF(DP0_DP_MSA_TIMING_PARAM1, DP_MSA_VTOTAL, mask_sh),\
+	SE_SF(DP0_DP_MSA_TIMING_PARAM2, DP_MSA_HSTART, mask_sh),\
+	SE_SF(DP0_DP_MSA_TIMING_PARAM2, DP_MSA_VSTART, mask_sh),\
+	SE_SF(DP0_DP_MSA_TIMING_PARAM3, DP_MSA_HSYNCWIDTH, mask_sh),\
+	SE_SF(DP0_DP_MSA_TIMING_PARAM3, DP_MSA_HSYNCPOLARITY, mask_sh),\
+	SE_SF(DP0_DP_MSA_TIMING_PARAM3, DP_MSA_VSYNCWIDTH, mask_sh),\
+	SE_SF(DP0_DP_MSA_TIMING_PARAM3, DP_MSA_VSYNCPOLARITY, mask_sh),\
+	SE_SF(DP0_DP_MSA_TIMING_PARAM4, DP_MSA_HWIDTH, mask_sh),\
+	SE_SF(DP0_DP_MSA_TIMING_PARAM4, DP_MSA_VHEIGHT, mask_sh),\
+	SE_SF(DIG0_HDMI_DB_CONTROL, HDMI_DB_DISABLE, mask_sh),\
+	SE_SF(DP0_DP_VID_TIMING, DP_VID_N_MUL, mask_sh),\
+	SE_SF(DIG0_DIG_FE_CNTL, DIG_SOURCE_SELECT, mask_sh), \
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC0_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC0_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC1_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC1_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC2_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC2_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC3_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC3_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC4_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC4_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC5_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC5_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC6_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC6_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC7_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC7_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC8_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC8_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC9_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC9_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC10_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC10_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC11_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC11_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC12_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC12_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC13_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC13_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC14_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC14_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL1, HDMI_GENERIC0_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL1, HDMI_GENERIC1_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL2, HDMI_GENERIC2_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL2, HDMI_GENERIC3_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL3, HDMI_GENERIC4_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL3, HDMI_GENERIC5_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL4, HDMI_GENERIC6_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL4, HDMI_GENERIC7_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL7, HDMI_GENERIC8_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL7, HDMI_GENERIC9_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL8, HDMI_GENERIC10_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL8, HDMI_GENERIC11_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL9, HDMI_GENERIC12_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL9, HDMI_GENERIC13_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL10, HDMI_GENERIC14_LINE, mask_sh),\
+	SE_SF(DP0_DP_DSC_CNTL, DP_DSC_MODE, mask_sh),\
+	SE_SF(DP0_DP_MSA_VBID_MISC, DP_VBID6_LINE_REFERENCE, mask_sh),\
+	SE_SF(DP0_DP_MSA_VBID_MISC, DP_VBID6_LINE_NUM, mask_sh),\
+	SE_SF(DME0_DME_CONTROL, METADATA_ENGINE_EN, mask_sh),\
+	SE_SF(DME0_DME_CONTROL, METADATA_HUBP_REQUESTOR_ID, mask_sh),\
+	SE_SF(DME0_DME_CONTROL, METADATA_STREAM_TYPE, mask_sh),\
+	SE_SF(DP0_DP_SEC_METADATA_TRANSMISSION, DP_SEC_METADATA_PACKET_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_METADATA_TRANSMISSION, DP_SEC_METADATA_PACKET_LINE_REFERENCE, mask_sh),\
+	SE_SF(DP0_DP_SEC_METADATA_TRANSMISSION, DP_SEC_METADATA_PACKET_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_METADATA_PACKET_CONTROL, HDMI_METADATA_PACKET_ENABLE, mask_sh),\
+	SE_SF(DIG0_HDMI_METADATA_PACKET_CONTROL, HDMI_METADATA_PACKET_LINE_REFERENCE, mask_sh),\
+	SE_SF(DIG0_HDMI_METADATA_PACKET_CONTROL, HDMI_METADATA_PACKET_LINE, mask_sh),\
+	SE_SF(DIG0_DIG_FE_CNTL, DOLBY_VISION_EN, mask_sh),\
+	SE_SF(DIG0_DIG_FE_CNTL, DIG_SYMCLK_FE_ON, mask_sh),\
+	SE_SF(DP0_DP_SEC_FRAMING4, DP_SST_SDP_SPLITTING, mask_sh),\
+	SE_SF(DIG0_DIG_CLOCK_PATTERN, DIG_CLOCK_PATTERN, mask_sh),\
+	SE_SF(DIG0_DIG_FIFO_CTRL0, DIG_FIFO_OUTPUT_PIXEL_MODE, mask_sh),\
+	SE_SF(DIG0_DIG_FIFO_CTRL0, DIG_FIFO_READ_START_LEVEL, mask_sh),\
+	SE_SF(DIG0_DIG_FIFO_CTRL0, DIG_FIFO_ENABLE, mask_sh),\
+	SE_SF(DIG0_DIG_FIFO_CTRL0, DIG_FIFO_RESET, mask_sh),\
+	SE_SF(DIG0_DIG_FIFO_CTRL0, DIG_FIFO_RESET_DONE, mask_sh)
+
+
+void dcn32_dio_stream_encoder_construct(
+	struct dcn10_stream_encoder *enc1,
+	struct dc_context *ctx,
+	struct dc_bios *bp,
+	enum engine_id eng_id,
+	struct vpg *vpg,
+	struct afmt *afmt,
+	const struct dcn10_stream_enc_registers *regs,
+	const struct dcn10_stream_encoder_shift *se_shift,
+	const struct dcn10_stream_encoder_mask *se_mask);
+
+
+void enc32_enable_fifo(struct stream_encoder *enc);
+
+void enc32_stream_encoder_dp_unblank(
+		struct dc_link *link,
+		struct stream_encoder *enc,
+		const struct encoder_unblank_param *param);
+
+#endif /* __DC_DIO_STREAM_ENCODER_DCN32_H__ */
diff --git a/drivers/gpu/drm/amd/display/dc/dio/dcn321/dcn321_dio_link_encoder.c b/drivers/gpu/drm/amd/display/dc/dio/dcn321/dcn321_dio_link_encoder.c
new file mode 100644
index 000000000000..2ed382a8e79c
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dio/dcn321/dcn321_dio_link_encoder.c
@@ -0,0 +1,190 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#include "reg_helper.h"
+
+#include "core_types.h"
+#include "link_encoder.h"
+#include "dcn321_dio_link_encoder.h"
+#include "dcn31/dcn31_dio_link_encoder.h"
+#include "stream_encoder.h"
+#include "dc_bios_types.h"
+
+#include "gpio_service_interface.h"
+
+#ifndef MIN
+#define MIN(X, Y) ((X) < (Y) ? (X) : (Y))
+#endif
+
+#define CTX \
+	enc10->base.ctx
+#define DC_LOGGER \
+	enc10->base.ctx->logger
+
+#define REG(reg)\
+	(enc10->link_regs->reg)
+
+#undef FN
+#define FN(reg_name, field_name) \
+	enc10->link_shift->field_name, enc10->link_mask->field_name
+
+#define AUX_REG(reg)\
+	(enc10->aux_regs->reg)
+
+#define AUX_REG_READ(reg_name) \
+		dm_read_reg(CTX, AUX_REG(reg_name))
+
+#define AUX_REG_WRITE(reg_name, val) \
+			dm_write_reg(CTX, AUX_REG(reg_name), val)
+
+static const struct link_encoder_funcs dcn321_link_enc_funcs = {
+	.read_state = link_enc2_read_state,
+	.validate_output_with_stream =
+			dcn30_link_encoder_validate_output_with_stream,
+	.hw_init = enc32_hw_init,
+	.setup = dcn10_link_encoder_setup,
+	.enable_tmds_output = dcn10_link_encoder_enable_tmds_output,
+	.enable_dp_output = dcn32_link_encoder_enable_dp_output,
+	.enable_dp_mst_output = dcn10_link_encoder_enable_dp_mst_output,
+	.disable_output = dcn10_link_encoder_disable_output,
+	.dp_set_lane_settings = dcn10_link_encoder_dp_set_lane_settings,
+	.dp_set_phy_pattern = dcn10_link_encoder_dp_set_phy_pattern,
+	.update_mst_stream_allocation_table =
+		dcn10_link_encoder_update_mst_stream_allocation_table,
+	.psr_program_dp_dphy_fast_training =
+			dcn10_psr_program_dp_dphy_fast_training,
+	.psr_program_secondary_packet = dcn10_psr_program_secondary_packet,
+	.connect_dig_be_to_fe = dcn10_link_encoder_connect_dig_be_to_fe,
+	.enable_hpd = dcn10_link_encoder_enable_hpd,
+	.disable_hpd = dcn10_link_encoder_disable_hpd,
+	.is_dig_enabled = dcn10_is_dig_enabled,
+	.destroy = dcn10_link_encoder_destroy,
+	.fec_set_enable = enc2_fec_set_enable,
+	.fec_set_ready = enc2_fec_set_ready,
+	.fec_is_active = enc2_fec_is_active,
+	.get_dig_frontend = dcn10_get_dig_frontend,
+	.get_dig_mode = dcn10_get_dig_mode,
+	.is_in_alt_mode = dcn20_link_encoder_is_in_alt_mode,
+	.get_max_link_cap = dcn20_link_encoder_get_max_link_cap,
+	.set_dio_phy_mux = dcn31_link_encoder_set_dio_phy_mux,
+};
+
+void dcn321_link_encoder_construct(
+	struct dcn20_link_encoder *enc20,
+	const struct encoder_init_data *init_data,
+	const struct encoder_feature_support *enc_features,
+	const struct dcn10_link_enc_registers *link_regs,
+	const struct dcn10_link_enc_aux_registers *aux_regs,
+	const struct dcn10_link_enc_hpd_registers *hpd_regs,
+	const struct dcn10_link_enc_shift *link_shift,
+	const struct dcn10_link_enc_mask *link_mask)
+{
+	struct bp_connector_speed_cap_info bp_cap_info = {0};
+	const struct dc_vbios_funcs *bp_funcs = init_data->ctx->dc_bios->funcs;
+	enum bp_result result = BP_RESULT_OK;
+	struct dcn10_link_encoder *enc10 = &enc20->enc10;
+
+	enc10->base.funcs = &dcn321_link_enc_funcs;
+	enc10->base.ctx = init_data->ctx;
+	enc10->base.id = init_data->encoder;
+
+	enc10->base.hpd_source = init_data->hpd_source;
+	enc10->base.connector = init_data->connector;
+
+	if (enc10->base.connector.id == CONNECTOR_ID_USBC)
+		enc10->base.features.flags.bits.DP_IS_USB_C = 1;
+
+	enc10->base.preferred_engine = ENGINE_ID_UNKNOWN;
+
+	enc10->base.features = *enc_features;
+
+	enc10->base.transmitter = init_data->transmitter;
+
+	/* set the flag to indicate whether driver poll the I2C data pin
+	 * while doing the DP sink detect
+	 */
+
+	enc10->base.output_signals =
+		SIGNAL_TYPE_DVI_SINGLE_LINK |
+		SIGNAL_TYPE_DVI_DUAL_LINK |
+		SIGNAL_TYPE_LVDS |
+		SIGNAL_TYPE_DISPLAY_PORT |
+		SIGNAL_TYPE_DISPLAY_PORT_MST |
+		SIGNAL_TYPE_EDP |
+		SIGNAL_TYPE_HDMI_TYPE_A;
+
+	enc10->link_regs = link_regs;
+	enc10->aux_regs = aux_regs;
+	enc10->hpd_regs = hpd_regs;
+	enc10->link_shift = link_shift;
+	enc10->link_mask = link_mask;
+
+	switch (enc10->base.transmitter) {
+	case TRANSMITTER_UNIPHY_A:
+		enc10->base.preferred_engine = ENGINE_ID_DIGA;
+	break;
+	case TRANSMITTER_UNIPHY_B:
+		enc10->base.preferred_engine = ENGINE_ID_DIGB;
+	break;
+	case TRANSMITTER_UNIPHY_C:
+		enc10->base.preferred_engine = ENGINE_ID_DIGC;
+	break;
+	case TRANSMITTER_UNIPHY_D:
+		enc10->base.preferred_engine = ENGINE_ID_DIGD;
+	break;
+	case TRANSMITTER_UNIPHY_E:
+		enc10->base.preferred_engine = ENGINE_ID_DIGE;
+	break;
+	default:
+		ASSERT_CRITICAL(false);
+		enc10->base.preferred_engine = ENGINE_ID_UNKNOWN;
+	}
+
+	/* default to one to mirror Windows behavior */
+	enc10->base.features.flags.bits.HDMI_6GB_EN = 1;
+
+	if (bp_funcs->get_connector_speed_cap_info)
+		result = bp_funcs->get_connector_speed_cap_info(enc10->base.ctx->dc_bios,
+						enc10->base.connector, &bp_cap_info);
+
+	/* Override features with DCE-specific values */
+	if (result == BP_RESULT_OK) {
+		enc10->base.features.flags.bits.IS_HBR2_CAPABLE =
+				bp_cap_info.DP_HBR2_EN;
+		enc10->base.features.flags.bits.IS_HBR3_CAPABLE =
+				bp_cap_info.DP_HBR3_EN;
+		enc10->base.features.flags.bits.HDMI_6GB_EN = bp_cap_info.HDMI_6GB_EN;
+		enc10->base.features.flags.bits.IS_DP2_CAPABLE = 1;
+		enc10->base.features.flags.bits.IS_UHBR10_CAPABLE = bp_cap_info.DP_UHBR10_EN;
+		enc10->base.features.flags.bits.IS_UHBR13_5_CAPABLE = bp_cap_info.DP_UHBR13_5_EN;
+		enc10->base.features.flags.bits.IS_UHBR20_CAPABLE = bp_cap_info.DP_UHBR20_EN;
+	} else {
+		DC_LOG_WARNING("%s: Failed to get encoder_cap_info from VBIOS with error code %d!\n",
+				__func__,
+				result);
+	}
+	if (enc10->base.ctx->dc->debug.hdmi20_disable)
+		enc10->base.features.flags.bits.HDMI_6GB_EN = 0;
+}
diff --git a/drivers/gpu/drm/amd/display/dc/dio/dcn321/dcn321_dio_link_encoder.h b/drivers/gpu/drm/amd/display/dc/dio/dcn321/dcn321_dio_link_encoder.h
new file mode 100644
index 000000000000..2205f39b0a24
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dio/dcn321/dcn321_dio_link_encoder.h
@@ -0,0 +1,42 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ *  and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __DC_LINK_ENCODER__DCN321_H__
+#define __DC_LINK_ENCODER__DCN321_H__
+
+#include "dcn32/dcn32_dio_link_encoder.h"
+
+void dcn321_link_encoder_construct(
+	struct dcn20_link_encoder *enc20,
+	const struct encoder_init_data *init_data,
+	const struct encoder_feature_support *enc_features,
+	const struct dcn10_link_enc_registers *link_regs,
+	const struct dcn10_link_enc_aux_registers *aux_regs,
+	const struct dcn10_link_enc_hpd_registers *hpd_regs,
+	const struct dcn10_link_enc_shift *link_shift,
+	const struct dcn10_link_enc_mask *link_mask);
+
+
+#endif /* __DC_LINK_ENCODER__DCN321_H__ */
diff --git a/drivers/gpu/drm/amd/display/dc/dio/dcn35/dcn35_dio_link_encoder.c b/drivers/gpu/drm/amd/display/dc/dio/dcn35/dcn35_dio_link_encoder.c
new file mode 100644
index 000000000000..9972911330b6
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dio/dcn35/dcn35_dio_link_encoder.c
@@ -0,0 +1,391 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "reg_helper.h"
+
+#include "core_types.h"
+#include "link_encoder.h"
+#include "dcn31/dcn31_dio_link_encoder.h"
+#include "dcn35_dio_link_encoder.h"
+#include "dc_dmub_srv.h"
+#define CTX \
+	enc10->base.ctx
+#define DC_LOGGER \
+	enc10->base.ctx->logger
+
+#define REG(reg)\
+	(enc10->link_regs->reg)
+
+#undef FN
+#define FN(reg_name, field_name) \
+	enc10->link_shift->field_name, enc10->link_mask->field_name
+/*
+ * @brief
+ * Trigger Source Select
+ * ASIC-dependent, actual values for register programming
+ */
+#define DCN35_DIG_FE_SOURCE_SELECT_INVALID 0x0
+#define DCN35_DIG_FE_SOURCE_SELECT_DIGA 0x1
+#define DCN35_DIG_FE_SOURCE_SELECT_DIGB 0x2
+#define DCN35_DIG_FE_SOURCE_SELECT_DIGC 0x4
+#define DCN35_DIG_FE_SOURCE_SELECT_DIGD 0x08
+#define DCN35_DIG_FE_SOURCE_SELECT_DIGE 0x10
+
+
+bool dcn35_is_dig_enabled(struct link_encoder *enc)
+{
+	uint32_t enabled;
+	struct dcn10_link_encoder *enc10 = TO_DCN10_LINK_ENC(enc);
+
+	REG_GET(DIG_BE_CLK_CNTL, DIG_BE_CLK_EN, &enabled);
+	return (enabled == 1);
+}
+
+enum signal_type dcn35_get_dig_mode(
+	struct link_encoder *enc)
+{
+	uint32_t value;
+	struct dcn10_link_encoder *enc10 = TO_DCN10_LINK_ENC(enc);
+
+	REG_GET(DIG_BE_CLK_CNTL, DIG_BE_MODE, &value);
+	switch (value) {
+	case 0:
+		return SIGNAL_TYPE_DISPLAY_PORT;
+	case 2:
+		return SIGNAL_TYPE_DVI_SINGLE_LINK;
+	case 3:
+		return SIGNAL_TYPE_HDMI_TYPE_A;
+	case 5:
+		return SIGNAL_TYPE_DISPLAY_PORT_MST;
+	default:
+		return SIGNAL_TYPE_NONE;
+	}
+}
+
+void dcn35_link_encoder_setup(
+	struct link_encoder *enc,
+	enum signal_type signal)
+{
+	struct dcn10_link_encoder *enc10 = TO_DCN10_LINK_ENC(enc);
+
+	switch (signal) {
+	case SIGNAL_TYPE_EDP:
+	case SIGNAL_TYPE_DISPLAY_PORT:
+		/* DP SST */
+		REG_UPDATE(DIG_BE_CLK_CNTL, DIG_BE_MODE, 0);
+		break;
+	case SIGNAL_TYPE_DVI_SINGLE_LINK:
+	case SIGNAL_TYPE_DVI_DUAL_LINK:
+		/* TMDS-DVI */
+		REG_UPDATE(DIG_BE_CLK_CNTL, DIG_BE_MODE, 2);
+		break;
+	case SIGNAL_TYPE_HDMI_TYPE_A:
+		/* TMDS-HDMI */
+		REG_UPDATE(DIG_BE_CLK_CNTL, DIG_BE_MODE, 3);
+		break;
+	case SIGNAL_TYPE_DISPLAY_PORT_MST:
+		/* DP MST */
+		REG_UPDATE(DIG_BE_CLK_CNTL, DIG_BE_MODE, 5);
+		break;
+	default:
+		ASSERT_CRITICAL(false);
+		/* invalid mode ! */
+		break;
+	}
+	REG_UPDATE(DIG_BE_CLK_CNTL, DIG_BE_CLK_EN, 1);
+
+}
+
+void dcn35_link_encoder_init(struct link_encoder *enc)
+{
+	enc31_hw_init(enc);
+	dcn35_link_encoder_set_fgcg(enc, enc->ctx->dc->debug.enable_fine_grain_clock_gating.bits.dio);
+}
+
+void dcn35_link_encoder_set_fgcg(struct link_encoder *enc, bool enable)
+{
+	struct dcn10_link_encoder *enc10 = TO_DCN10_LINK_ENC(enc);
+
+	REG_UPDATE(DIO_CLK_CNTL, DIO_FGCG_REP_DIS, !enable);
+}
+
+static const struct link_encoder_funcs dcn35_link_enc_funcs = {
+	.read_state = link_enc2_read_state,
+	.validate_output_with_stream =
+			dcn30_link_encoder_validate_output_with_stream,
+	.hw_init = dcn35_link_encoder_init,
+	.setup = dcn35_link_encoder_setup,
+	.enable_tmds_output = dcn10_link_encoder_enable_tmds_output,
+	.enable_dp_output = dcn35_link_encoder_enable_dp_output,
+	.enable_dp_mst_output = dcn35_link_encoder_enable_dp_mst_output,
+	.disable_output = dcn35_link_encoder_disable_output,
+	.dp_set_lane_settings = dcn10_link_encoder_dp_set_lane_settings,
+	.dp_set_phy_pattern = dcn10_link_encoder_dp_set_phy_pattern,
+	.update_mst_stream_allocation_table =
+		dcn10_link_encoder_update_mst_stream_allocation_table,
+	.psr_program_dp_dphy_fast_training =
+			dcn10_psr_program_dp_dphy_fast_training,
+	.psr_program_secondary_packet = dcn10_psr_program_secondary_packet,
+	.connect_dig_be_to_fe = dcn10_link_encoder_connect_dig_be_to_fe,
+	.enable_hpd = dcn10_link_encoder_enable_hpd,
+	.disable_hpd = dcn10_link_encoder_disable_hpd,
+	.is_dig_enabled = dcn35_is_dig_enabled,
+	.destroy = dcn10_link_encoder_destroy,
+	.fec_set_enable = enc2_fec_set_enable,
+	.fec_set_ready = enc2_fec_set_ready,
+	.fec_is_active = enc2_fec_is_active,
+	.get_dig_frontend = dcn10_get_dig_frontend,
+	.get_dig_mode = dcn35_get_dig_mode,
+	.is_in_alt_mode = dcn31_link_encoder_is_in_alt_mode,
+	.get_max_link_cap = dcn31_link_encoder_get_max_link_cap,
+	.set_dio_phy_mux = dcn31_link_encoder_set_dio_phy_mux,
+	.enable_dpia_output = dcn35_link_encoder_enable_dpia_output,
+	.disable_dpia_output = dcn35_link_encoder_disable_dpia_output,
+};
+
+void dcn35_link_encoder_construct(
+	struct dcn20_link_encoder *enc20,
+	const struct encoder_init_data *init_data,
+	const struct encoder_feature_support *enc_features,
+	const struct dcn10_link_enc_registers *link_regs,
+	const struct dcn10_link_enc_aux_registers *aux_regs,
+	const struct dcn10_link_enc_hpd_registers *hpd_regs,
+	const struct dcn10_link_enc_shift *link_shift,
+	const struct dcn10_link_enc_mask *link_mask)
+{
+	struct bp_connector_speed_cap_info bp_cap_info = {0};
+	const struct dc_vbios_funcs *bp_funcs = init_data->ctx->dc_bios->funcs;
+	enum bp_result result = BP_RESULT_OK;
+	struct dcn10_link_encoder *enc10 = &enc20->enc10;
+
+	enc10->base.funcs = &dcn35_link_enc_funcs;
+	enc10->base.ctx = init_data->ctx;
+	enc10->base.id = init_data->encoder;
+
+	enc10->base.hpd_source = init_data->hpd_source;
+	enc10->base.connector = init_data->connector;
+
+	enc10->base.preferred_engine = ENGINE_ID_UNKNOWN;
+
+	enc10->base.features = *enc_features;
+
+	if (enc10->base.connector.id == CONNECTOR_ID_USBC)
+		enc10->base.features.flags.bits.DP_IS_USB_C = 1;
+
+	enc10->base.transmitter = init_data->transmitter;
+
+	/* set the flag to indicate whether driver poll the I2C data pin
+	 * while doing the DP sink detect
+	 */
+
+/*	if (dal_adapter_service_is_feature_supported(as,
+ *		FEATURE_DP_SINK_DETECT_POLL_DATA_PIN))
+ *		enc10->base.features.flags.bits.
+ *			DP_SINK_DETECT_POLL_DATA_PIN = true;
+ */
+
+	enc10->base.output_signals =
+		SIGNAL_TYPE_DVI_SINGLE_LINK |
+		SIGNAL_TYPE_DVI_DUAL_LINK |
+		SIGNAL_TYPE_LVDS |
+		SIGNAL_TYPE_DISPLAY_PORT |
+		SIGNAL_TYPE_DISPLAY_PORT_MST |
+		SIGNAL_TYPE_EDP |
+		SIGNAL_TYPE_HDMI_TYPE_A;
+
+	enc10->link_regs = link_regs;
+	enc10->aux_regs = aux_regs;
+	enc10->hpd_regs = hpd_regs;
+	enc10->link_shift = link_shift;
+	enc10->link_mask = link_mask;
+
+	switch (enc10->base.transmitter) {
+	case TRANSMITTER_UNIPHY_A:
+		enc10->base.preferred_engine = ENGINE_ID_DIGA;
+	break;
+	case TRANSMITTER_UNIPHY_B:
+		enc10->base.preferred_engine = ENGINE_ID_DIGB;
+	break;
+	case TRANSMITTER_UNIPHY_C:
+		enc10->base.preferred_engine = ENGINE_ID_DIGC;
+	break;
+	case TRANSMITTER_UNIPHY_D:
+		enc10->base.preferred_engine = ENGINE_ID_DIGD;
+	break;
+	case TRANSMITTER_UNIPHY_E:
+		enc10->base.preferred_engine = ENGINE_ID_DIGE;
+	break;
+	default:
+		ASSERT_CRITICAL(false);
+		enc10->base.preferred_engine = ENGINE_ID_UNKNOWN;
+	}
+
+	enc10->base.features.flags.bits.HDMI_6GB_EN = 1;
+
+	if (bp_funcs->get_connector_speed_cap_info)
+		result = bp_funcs->get_connector_speed_cap_info(enc10->base.ctx->dc_bios,
+						enc10->base.connector, &bp_cap_info);
+
+	/* Override features with DCE-specific values */
+	if (result == BP_RESULT_OK) {
+		enc10->base.features.flags.bits.IS_HBR2_CAPABLE =
+				bp_cap_info.DP_HBR2_EN;
+		enc10->base.features.flags.bits.IS_HBR3_CAPABLE =
+				bp_cap_info.DP_HBR3_EN;
+		enc10->base.features.flags.bits.HDMI_6GB_EN = bp_cap_info.HDMI_6GB_EN;
+		enc10->base.features.flags.bits.IS_DP2_CAPABLE = 1;
+		enc10->base.features.flags.bits.IS_UHBR10_CAPABLE = bp_cap_info.DP_UHBR10_EN;
+		enc10->base.features.flags.bits.IS_UHBR13_5_CAPABLE = bp_cap_info.DP_UHBR13_5_EN;
+		enc10->base.features.flags.bits.IS_UHBR20_CAPABLE = bp_cap_info.DP_UHBR20_EN;
+
+	} else {
+		DC_LOG_WARNING("%s: Failed to get encoder_cap_info from VBIOS with error code %d!\n",
+				__func__,
+				result);
+	}
+	if (enc10->base.ctx->dc->debug.hdmi20_disable)
+		enc10->base.features.flags.bits.HDMI_6GB_EN = 0;
+
+}
+
+/* DPIA equivalent of link_transmitter_control. */
+static bool link_dpia_control(struct dc_context *dc_ctx,
+	struct dmub_cmd_dig_dpia_control_data *dpia_control)
+{
+	union dmub_rb_cmd cmd;
+
+	memset(&cmd, 0, sizeof(cmd));
+
+	cmd.dig1_dpia_control.header.type = DMUB_CMD__DPIA;
+	cmd.dig1_dpia_control.header.sub_type =
+			DMUB_CMD__DPIA_DIG1_DPIA_CONTROL;
+	cmd.dig1_dpia_control.header.payload_bytes =
+		sizeof(cmd.dig1_dpia_control) -
+		sizeof(cmd.dig1_dpia_control.header);
+
+	cmd.dig1_dpia_control.dpia_control = *dpia_control;
+
+	dc_wake_and_execute_dmub_cmd(dc_ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
+
+	return true;
+}
+
+static void link_encoder_disable(struct dcn10_link_encoder *enc10)
+{
+	/* reset training complete */
+	REG_UPDATE(DP_LINK_CNTL, DP_LINK_TRAINING_COMPLETE, 0);
+}
+
+void dcn35_link_encoder_enable_dp_output(
+	struct link_encoder *enc,
+	const struct dc_link_settings *link_settings,
+	enum clock_source_id clock_source)
+{
+	struct dcn10_link_encoder *enc10 = TO_DCN10_LINK_ENC(enc);
+
+	if (!enc->ctx->dc->config.unify_link_enc_assignment)
+		dcn31_link_encoder_enable_dp_output(enc, link_settings, clock_source);
+	else {
+		DC_LOG_DEBUG("%s: enc_id(%d)\n", __func__, enc->preferred_engine);
+		dcn20_link_encoder_enable_dp_output(enc, link_settings, clock_source);
+	}
+}
+
+void dcn35_link_encoder_enable_dp_mst_output(
+	struct link_encoder *enc,
+	const struct dc_link_settings *link_settings,
+	enum clock_source_id clock_source)
+{
+	struct dcn10_link_encoder *enc10 = TO_DCN10_LINK_ENC(enc);
+
+	if (!enc->ctx->dc->config.unify_link_enc_assignment)
+		dcn31_link_encoder_enable_dp_mst_output(enc, link_settings, clock_source);
+	else {
+		DC_LOG_DEBUG("%s: enc_id(%d)\n", __func__, enc->preferred_engine);
+		dcn10_link_encoder_enable_dp_mst_output(enc, link_settings, clock_source);
+	}
+}
+
+void dcn35_link_encoder_disable_output(
+	struct link_encoder *enc,
+	enum signal_type signal)
+{
+	struct dcn10_link_encoder *enc10 = TO_DCN10_LINK_ENC(enc);
+
+	if (!enc->ctx->dc->config.unify_link_enc_assignment)
+		dcn31_link_encoder_disable_output(enc, signal);
+	else {
+		DC_LOG_DEBUG("%s: enc_id(%d)\n", __func__, enc->preferred_engine);
+		dcn10_link_encoder_disable_output(enc, signal);
+	}
+}
+
+void dcn35_link_encoder_enable_dpia_output(
+	struct link_encoder *enc,
+	const struct dc_link_settings *link_settings,
+	uint8_t dpia_id,
+	uint8_t digmode,
+	uint8_t fec_rdy)
+{
+	struct dcn10_link_encoder *enc10 = TO_DCN10_LINK_ENC(enc);
+	struct dmub_cmd_dig_dpia_control_data dpia_control = { 0 };
+
+	enc1_configure_encoder(enc10, link_settings);
+
+	dpia_control.action = (uint8_t)TRANSMITTER_CONTROL_ENABLE;
+	dpia_control.enc_id = enc->preferred_engine;
+	dpia_control.mode_laneset.digmode = digmode;
+	dpia_control.lanenum = (uint8_t)link_settings->lane_count;
+	dpia_control.symclk_10khz = link_settings->link_rate *
+			LINK_RATE_REF_FREQ_IN_KHZ / 10;
+	/* DIG_BE_CNTL.DIG_HPD_SELECT set to 5 (hpdsel - 1) to indicate HPD pin unused by DPIA. */
+	dpia_control.hpdsel = 6;
+	dpia_control.dpia_id = dpia_id;
+	dpia_control.fec_rdy = fec_rdy;
+
+	DC_LOG_DEBUG("%s: DPIA(%d) - enc_id(%d)\n", __func__, dpia_control.dpia_id, dpia_control.enc_id);
+	link_dpia_control(enc->ctx, &dpia_control);
+}
+
+void dcn35_link_encoder_disable_dpia_output(
+	struct link_encoder *enc,
+	uint8_t dpia_id,
+	uint8_t digmode)
+{
+	struct dcn10_link_encoder *enc10 = TO_DCN10_LINK_ENC(enc);
+	struct dmub_cmd_dig_dpia_control_data dpia_control = { 0 };
+
+	if (enc->funcs->is_dig_enabled && !enc->funcs->is_dig_enabled(enc))
+		return;
+
+	dpia_control.action = (uint8_t)TRANSMITTER_CONTROL_DISABLE;
+	dpia_control.enc_id = enc->preferred_engine;
+	dpia_control.mode_laneset.digmode = digmode;
+	dpia_control.dpia_id = dpia_id;
+
+	DC_LOG_DEBUG("%s: DPIA(%d) - enc_id(%d)\n", __func__, dpia_control.dpia_id, dpia_control.enc_id);
+	link_dpia_control(enc->ctx, &dpia_control);
+
+	link_encoder_disable(enc10);
+}
diff --git a/drivers/gpu/drm/amd/display/dc/dio/dcn35/dcn35_dio_link_encoder.h b/drivers/gpu/drm/amd/display/dc/dio/dcn35/dcn35_dio_link_encoder.h
new file mode 100644
index 000000000000..5712e6553fab
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dio/dcn35/dcn35_dio_link_encoder.h
@@ -0,0 +1,188 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __DC_LINK_ENCODER__DCN35_H__
+#define __DC_LINK_ENCODER__DCN35_H__
+
+#include "dcn32/dcn32_dio_link_encoder.h"
+#include "dcn30/dcn30_dio_link_encoder.h"
+#include "dcn31/dcn31_dio_link_encoder.h"
+
+#define LINK_ENCODER_MASK_SH_LIST_DCN35(mask_sh) \
+	LE_SF(DIG0_DIG_BE_EN_CNTL, DIG_BE_ENABLE, mask_sh),\
+	LE_SF(DIG0_DIG_BE_CNTL, DIG_RB_SWITCH_EN, mask_sh),\
+	LE_SF(DIG0_DIG_BE_CNTL, DIG_HPD_SELECT, mask_sh),\
+	LE_SF(DIG0_DIG_BE_CNTL, DIG_FE_SOURCE_SELECT, mask_sh),\
+	LE_SF(DIG0_DIG_BE_CLK_CNTL, DIG_BE_MODE, mask_sh),\
+	LE_SF(DIG0_DIG_BE_CLK_CNTL, DIG_BE_CLK_EN, mask_sh),\
+	LE_SF(DIG0_DIG_BE_CLK_CNTL, DIG_BE_SOFT_RESET, mask_sh),\
+	LE_SF(DIG0_DIG_BE_CLK_CNTL, HDCP_SOFT_RESET, mask_sh),\
+	LE_SF(DIG0_DIG_BE_CLK_CNTL, DIG_BE_SYMCLK_G_CLOCK_ON, mask_sh),\
+	LE_SF(DIG0_DIG_BE_CLK_CNTL, DIG_BE_SYMCLK_G_HDCP_CLOCK_ON, mask_sh),\
+	LE_SF(DIG0_DIG_BE_CLK_CNTL, DIG_BE_SYMCLK_G_TMDS_CLOCK_ON, mask_sh),\
+	LE_SF(DIG0_DIG_CLOCK_PATTERN, DIG_CLOCK_PATTERN, mask_sh),\
+	LE_SF(DIG0_TMDS_CTL_BITS, TMDS_CTL0, mask_sh), \
+	LE_SF(DP0_DP_DPHY_CNTL, DPHY_BYPASS, mask_sh),\
+	LE_SF(DP0_DP_DPHY_CNTL, DPHY_ATEST_SEL_LANE0, mask_sh),\
+	LE_SF(DP0_DP_DPHY_CNTL, DPHY_ATEST_SEL_LANE1, mask_sh),\
+	LE_SF(DP0_DP_DPHY_CNTL, DPHY_ATEST_SEL_LANE2, mask_sh),\
+	LE_SF(DP0_DP_DPHY_CNTL, DPHY_ATEST_SEL_LANE3, mask_sh),\
+	LE_SF(DP0_DP_DPHY_PRBS_CNTL, DPHY_PRBS_EN, mask_sh),\
+	LE_SF(DP0_DP_DPHY_PRBS_CNTL, DPHY_PRBS_SEL, mask_sh),\
+	LE_SF(DP0_DP_DPHY_SYM0, DPHY_SYM1, mask_sh),\
+	LE_SF(DP0_DP_DPHY_SYM0, DPHY_SYM2, mask_sh),\
+	LE_SF(DP0_DP_DPHY_SYM0, DPHY_SYM3, mask_sh),\
+	LE_SF(DP0_DP_DPHY_SYM1, DPHY_SYM4, mask_sh),\
+	LE_SF(DP0_DP_DPHY_SYM1, DPHY_SYM5, mask_sh),\
+	LE_SF(DP0_DP_DPHY_SYM1, DPHY_SYM6, mask_sh),\
+	LE_SF(DP0_DP_DPHY_SYM2, DPHY_SYM7, mask_sh),\
+	LE_SF(DP0_DP_DPHY_SYM2, DPHY_SYM8, mask_sh),\
+	LE_SF(DP0_DP_DPHY_SCRAM_CNTL, DPHY_SCRAMBLER_BS_COUNT, mask_sh),\
+	LE_SF(DP0_DP_DPHY_SCRAM_CNTL, DPHY_SCRAMBLER_ADVANCE, mask_sh),\
+	LE_SF(DP0_DP_DPHY_FAST_TRAINING, DPHY_RX_FAST_TRAINING_CAPABLE, mask_sh),\
+	LE_SF(DP0_DP_DPHY_BS_SR_SWAP_CNTL, DPHY_LOAD_BS_COUNT, mask_sh),\
+	LE_SF(DP0_DP_DPHY_TRAINING_PATTERN_SEL, DPHY_TRAINING_PATTERN_SEL, mask_sh),\
+	LE_SF(DP0_DP_DPHY_HBR2_PATTERN_CONTROL, DP_DPHY_HBR2_PATTERN_CONTROL, mask_sh),\
+	LE_SF(DP0_DP_LINK_CNTL, DP_LINK_TRAINING_COMPLETE, mask_sh),\
+	LE_SF(DP0_DP_LINK_FRAMING_CNTL, DP_IDLE_BS_INTERVAL, mask_sh),\
+	LE_SF(DP0_DP_LINK_FRAMING_CNTL, DP_VBID_DISABLE, mask_sh),\
+	LE_SF(DP0_DP_LINK_FRAMING_CNTL, DP_VID_ENHANCED_FRAME_MODE, mask_sh),\
+	LE_SF(DP0_DP_VID_STREAM_CNTL, DP_VID_STREAM_ENABLE, mask_sh),\
+	LE_SF(DP0_DP_CONFIG, DP_UDI_LANES, mask_sh),\
+	LE_SF(DP0_DP_SEC_CNTL1, DP_SEC_GSP0_LINE_NUM, mask_sh),\
+	LE_SF(DP0_DP_SEC_CNTL1, DP_SEC_GSP0_PRIORITY, mask_sh),\
+	LE_SF(DP0_DP_MSE_SAT0, DP_MSE_SAT_SRC0, mask_sh),\
+	LE_SF(DP0_DP_MSE_SAT0, DP_MSE_SAT_SRC1, mask_sh),\
+	LE_SF(DP0_DP_MSE_SAT0, DP_MSE_SAT_SLOT_COUNT0, mask_sh),\
+	LE_SF(DP0_DP_MSE_SAT0, DP_MSE_SAT_SLOT_COUNT1, mask_sh),\
+	LE_SF(DP0_DP_MSE_SAT1, DP_MSE_SAT_SRC2, mask_sh),\
+	LE_SF(DP0_DP_MSE_SAT1, DP_MSE_SAT_SRC3, mask_sh),\
+	LE_SF(DP0_DP_MSE_SAT1, DP_MSE_SAT_SLOT_COUNT2, mask_sh),\
+	LE_SF(DP0_DP_MSE_SAT1, DP_MSE_SAT_SLOT_COUNT3, mask_sh),\
+	LE_SF(DP0_DP_MSE_SAT_UPDATE, DP_MSE_SAT_UPDATE, mask_sh),\
+	LE_SF(DP0_DP_MSE_SAT_UPDATE, DP_MSE_16_MTP_KEEPOUT, mask_sh),\
+	LE_SF(DP_AUX0_AUX_CONTROL, AUX_HPD_SEL, mask_sh),\
+	LE_SF(DP_AUX0_AUX_CONTROL, AUX_LS_READ_EN, mask_sh),\
+	LE_SF(DP_AUX0_AUX_DPHY_RX_CONTROL0, AUX_RX_RECEIVE_WINDOW, mask_sh),\
+	LE_SF(HPD0_DC_HPD_CONTROL, DC_HPD_EN, mask_sh),\
+	LE_SF(DP0_DP_DPHY_CNTL, DPHY_FEC_EN, mask_sh),\
+	LE_SF(DP0_DP_DPHY_CNTL, DPHY_FEC_READY_SHADOW, mask_sh),\
+	LE_SF(DP0_DP_DPHY_CNTL, DPHY_FEC_ACTIVE_STATUS, mask_sh),\
+	LE_SF(DIG0_TMDS_CTL_BITS, TMDS_CTL0, mask_sh), \
+	LE_SF(DP_AUX0_AUX_DPHY_RX_CONTROL0, AUX_RX_START_WINDOW, mask_sh),\
+	LE_SF(DP_AUX0_AUX_DPHY_RX_CONTROL0, AUX_RX_HALF_SYM_DETECT_LEN, mask_sh),\
+	LE_SF(DP_AUX0_AUX_DPHY_RX_CONTROL0, AUX_RX_TRANSITION_FILTER_EN, mask_sh),\
+	LE_SF(DP_AUX0_AUX_DPHY_RX_CONTROL0, AUX_RX_ALLOW_BELOW_THRESHOLD_PHASE_DETECT, mask_sh),\
+	LE_SF(DP_AUX0_AUX_DPHY_RX_CONTROL0, AUX_RX_ALLOW_BELOW_THRESHOLD_START, mask_sh),\
+	LE_SF(DP_AUX0_AUX_DPHY_RX_CONTROL0, AUX_RX_ALLOW_BELOW_THRESHOLD_STOP, mask_sh),\
+	LE_SF(DP_AUX0_AUX_DPHY_RX_CONTROL0, AUX_RX_PHASE_DETECT_LEN, mask_sh),\
+	LE_SF(DP_AUX0_AUX_DPHY_RX_CONTROL0, AUX_RX_DETECTION_THRESHOLD, mask_sh), \
+	LE_SF(DP_AUX0_AUX_DPHY_TX_CONTROL, AUX_TX_PRECHARGE_LEN, mask_sh),\
+	LE_SF(DP_AUX0_AUX_DPHY_TX_CONTROL, AUX_TX_PRECHARGE_SYMBOLS, mask_sh),\
+	LE_SF(DP_AUX0_AUX_DPHY_TX_CONTROL, AUX_MODE_DET_CHECK_DELAY, mask_sh),\
+	LE_SF(DP_AUX0_AUX_DPHY_RX_CONTROL1, AUX_RX_PRECHARGE_SKIP, mask_sh),\
+	LE_SF(DP_AUX0_AUX_DPHY_RX_CONTROL1, AUX_RX_TIMEOUT_LEN, mask_sh),\
+	LE_SF(DP_AUX0_AUX_DPHY_RX_CONTROL1, AUX_RX_TIMEOUT_LEN_MUL, mask_sh),\
+	LE_SF(DIO_LINKA_CNTL, ENC_TYPE_SEL, mask_sh),\
+	LE_SF(DIO_LINKA_CNTL, HPO_DP_ENC_SEL, mask_sh),\
+	LE_SF(DIO_LINKA_CNTL, HPO_HDMI_ENC_SEL, mask_sh),\
+	LE_SF(DIO_CLK_CNTL, DISPCLK_R_GATE_DIS, mask_sh),\
+	LE_SF(DIO_CLK_CNTL, DISPCLK_G_GATE_DIS, mask_sh),\
+	LE_SF(DIO_CLK_CNTL, REFCLK_R_GATE_DIS, mask_sh),\
+	LE_SF(DIO_CLK_CNTL, REFCLK_G_GATE_DIS, mask_sh),\
+	LE_SF(DIO_CLK_CNTL, SOCCLK_G_GATE_DIS, mask_sh),\
+	LE_SF(DIO_CLK_CNTL, SYMCLK_FE_R_GATE_DIS, mask_sh),\
+	LE_SF(DIO_CLK_CNTL, SYMCLK_FE_G_GATE_DIS, mask_sh),\
+	LE_SF(DIO_CLK_CNTL, SYMCLK_R_GATE_DIS, mask_sh),\
+	LE_SF(DIO_CLK_CNTL, SYMCLK_G_GATE_DIS, mask_sh),\
+	LE_SF(DIO_CLK_CNTL, DIO_FGCG_REP_DIS, mask_sh),\
+	LE_SF(DIO_CLK_CNTL, DISPCLK_G_HDCP_GATE_DIS, mask_sh),\
+	LE_SF(DIO_CLK_CNTL, SYMCLKA_G_HDCP_GATE_DIS, mask_sh),\
+	LE_SF(DIO_CLK_CNTL, SYMCLKB_G_HDCP_GATE_DIS, mask_sh),\
+	LE_SF(DIO_CLK_CNTL, SYMCLKC_G_HDCP_GATE_DIS, mask_sh),\
+	LE_SF(DIO_CLK_CNTL, SYMCLKD_G_HDCP_GATE_DIS, mask_sh),\
+	LE_SF(DIO_CLK_CNTL, SYMCLKE_G_HDCP_GATE_DIS, mask_sh),\
+	LE_SF(DIO_CLK_CNTL, SYMCLKF_G_HDCP_GATE_DIS, mask_sh),\
+	LE_SF(DIO_CLK_CNTL, SYMCLKG_G_HDCP_GATE_DIS, mask_sh)
+
+
+void dcn35_link_encoder_construct(
+	struct dcn20_link_encoder *enc20,
+	const struct encoder_init_data *init_data,
+	const struct encoder_feature_support *enc_features,
+	const struct dcn10_link_enc_registers *link_regs,
+	const struct dcn10_link_enc_aux_registers *aux_regs,
+	const struct dcn10_link_enc_hpd_registers *hpd_regs,
+	const struct dcn10_link_enc_shift *link_shift,
+	const struct dcn10_link_enc_mask *link_mask);
+
+void dcn35_link_encoder_init(struct link_encoder *enc);
+void dcn35_link_encoder_set_fgcg(struct link_encoder *enc, bool enabled);
+bool dcn35_is_dig_enabled(struct link_encoder *enc);
+
+enum signal_type dcn35_get_dig_mode(struct link_encoder *enc);
+void dcn35_link_encoder_setup(struct link_encoder *enc, enum signal_type signal);
+
+/*
+ * Enable DP transmitter and its encoder.
+ */
+void dcn35_link_encoder_enable_dp_output(
+	struct link_encoder *enc,
+	const struct dc_link_settings *link_settings,
+	enum clock_source_id clock_source);
+
+/*
+ * Enable DP transmitter and its encoder in MST mode.
+ */
+void dcn35_link_encoder_enable_dp_mst_output(
+	struct link_encoder *enc,
+	const struct dc_link_settings *link_settings,
+	enum clock_source_id clock_source);
+
+/*
+ * Disable transmitter and its encoder.
+ */
+void dcn35_link_encoder_disable_output(
+	struct link_encoder *enc,
+	enum signal_type signal);
+
+/*
+ * Enable DP transmitter and its encoder for dpia port.
+ */
+void dcn35_link_encoder_enable_dpia_output(
+	struct link_encoder *enc,
+	const struct dc_link_settings *link_settings,
+	uint8_t dpia_id,
+	uint8_t digmode,
+	uint8_t fec_rdy);
+
+/*
+ * Disable transmitter and its encoder for dpia port.
+ */
+void dcn35_link_encoder_disable_dpia_output(
+	struct link_encoder *enc,
+	uint8_t dpia_id,
+	uint8_t digmode);
+
+#endif /* __DC_LINK_ENCODER__DCN35_H__ */
diff --git a/drivers/gpu/drm/amd/display/dc/dio/dcn35/dcn35_dio_stream_encoder.c b/drivers/gpu/drm/amd/display/dc/dio/dcn35/dcn35_dio_stream_encoder.c
new file mode 100644
index 000000000000..fd5d1dbf9dc6
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dio/dcn35/dcn35_dio_stream_encoder.c
@@ -0,0 +1,520 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "dc_bios_types.h"
+#include "dcn30/dcn30_dio_stream_encoder.h"
+#include "dcn314/dcn314_dio_stream_encoder.h"
+#include "dcn32/dcn32_dio_stream_encoder.h"
+#include "dcn35_dio_stream_encoder.h"
+#include "reg_helper.h"
+#include "hw_shared.h"
+#include "link_service.h"
+#include "dpcd_defs.h"
+
+#define DC_LOGGER \
+		enc1->base.ctx->logger
+
+#define REG(reg)\
+	(enc1->regs->reg)
+
+#undef FN
+#define FN(reg_name, field_name) \
+	enc1->se_shift->field_name, enc1->se_mask->field_name
+
+#define VBI_LINE_0 0
+#define HDMI_CLOCK_CHANNEL_RATE_MORE_340M 340000
+
+#define CTX \
+	enc1->base.ctx
+/* setup stream encoder in dvi mode */
+static void enc35_stream_encoder_dvi_set_stream_attribute(
+	struct stream_encoder *enc,
+	struct dc_crtc_timing *crtc_timing,
+	bool is_dual_link)
+{
+	struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc);
+
+	if (!enc->ctx->dc->debug.avoid_vbios_exec_table) {
+		struct bp_encoder_control cntl = {0};
+
+		cntl.action = ENCODER_CONTROL_SETUP;
+		cntl.engine_id = enc1->base.id;
+		cntl.signal = is_dual_link ?
+			SIGNAL_TYPE_DVI_DUAL_LINK : SIGNAL_TYPE_DVI_SINGLE_LINK;
+		cntl.enable_dp_audio = false;
+		cntl.pixel_clock = crtc_timing->pix_clk_100hz / 10;
+		cntl.lanes_number = (is_dual_link) ? LANE_COUNT_EIGHT : LANE_COUNT_FOUR;
+
+		if (enc1->base.bp->funcs->encoder_control(
+				enc1->base.bp, &cntl) != BP_RESULT_OK)
+			return;
+
+	} else {
+
+		//Set pattern for clock channel, default vlue 0x63 does not work
+		REG_UPDATE(DIG_CLOCK_PATTERN, DIG_CLOCK_PATTERN, 0x1F);
+
+		//DIG_BE_TMDS_DVI_MODE : TMDS-DVI mode is already set in link_encoder_setup
+
+		//DIG_SOURCE_SELECT is already set in dig_connect_to_otg
+
+		/* DIG_START is removed from the register spec */
+	}
+
+	ASSERT(crtc_timing->pixel_encoding == PIXEL_ENCODING_RGB);
+	ASSERT(crtc_timing->display_color_depth == COLOR_DEPTH_888);
+	enc1_stream_encoder_set_stream_attribute_helper(enc1, crtc_timing);
+}
+/* setup stream encoder in hdmi mode */
+static void enc35_stream_encoder_hdmi_set_stream_attribute(
+	struct stream_encoder *enc,
+	struct dc_crtc_timing *crtc_timing,
+	int actual_pix_clk_khz,
+	bool enable_audio)
+{
+	struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc);
+
+	if (!enc->ctx->dc->debug.avoid_vbios_exec_table) {
+		struct bp_encoder_control cntl = {0};
+
+		cntl.action = ENCODER_CONTROL_SETUP;
+		cntl.engine_id = enc1->base.id;
+		cntl.signal = SIGNAL_TYPE_HDMI_TYPE_A;
+		cntl.enable_dp_audio = enable_audio;
+		cntl.pixel_clock = actual_pix_clk_khz;
+		cntl.lanes_number = LANE_COUNT_FOUR;
+
+		if (enc1->base.bp->funcs->encoder_control(
+				enc1->base.bp, &cntl) != BP_RESULT_OK)
+			return;
+
+	} else {
+
+		//Set pattern for clock channel, default vlue 0x63 does not work
+		REG_UPDATE(DIG_CLOCK_PATTERN, DIG_CLOCK_PATTERN, 0x1F);
+
+		//DIG_BE_TMDS_HDMI_MODE : TMDS-HDMI mode is already set in link_encoder_setup
+
+		//DIG_SOURCE_SELECT is already set in dig_connect_to_otg
+
+		/* DIG_START is removed from the register spec */
+		enc314_enable_fifo(enc);
+	}
+
+	/* Configure pixel encoding */
+	enc1_stream_encoder_set_stream_attribute_helper(enc1, crtc_timing);
+
+	/* setup HDMI engine */
+	REG_UPDATE_6(HDMI_CONTROL,
+		HDMI_PACKET_GEN_VERSION, 1,
+		HDMI_KEEPOUT_MODE, 1,
+		HDMI_DEEP_COLOR_ENABLE, 0,
+		HDMI_DATA_SCRAMBLE_EN, 0,
+		HDMI_NO_EXTRA_NULL_PACKET_FILLED, 1,
+		HDMI_CLOCK_CHANNEL_RATE, 0);
+
+	/* Configure color depth */
+	switch (crtc_timing->display_color_depth) {
+	case COLOR_DEPTH_888:
+		REG_UPDATE(HDMI_CONTROL, HDMI_DEEP_COLOR_DEPTH, 0);
+		break;
+	case COLOR_DEPTH_101010:
+		if (crtc_timing->pixel_encoding == PIXEL_ENCODING_YCBCR422) {
+			REG_UPDATE_2(HDMI_CONTROL,
+					HDMI_DEEP_COLOR_DEPTH, 1,
+					HDMI_DEEP_COLOR_ENABLE, 0);
+		} else {
+			REG_UPDATE_2(HDMI_CONTROL,
+					HDMI_DEEP_COLOR_DEPTH, 1,
+					HDMI_DEEP_COLOR_ENABLE, 1);
+			}
+		break;
+	case COLOR_DEPTH_121212:
+		if (crtc_timing->pixel_encoding == PIXEL_ENCODING_YCBCR422) {
+			REG_UPDATE_2(HDMI_CONTROL,
+					HDMI_DEEP_COLOR_DEPTH, 2,
+					HDMI_DEEP_COLOR_ENABLE, 0);
+		} else {
+			REG_UPDATE_2(HDMI_CONTROL,
+					HDMI_DEEP_COLOR_DEPTH, 2,
+					HDMI_DEEP_COLOR_ENABLE, 1);
+			}
+		break;
+	case COLOR_DEPTH_161616:
+		REG_UPDATE_2(HDMI_CONTROL,
+				HDMI_DEEP_COLOR_DEPTH, 3,
+				HDMI_DEEP_COLOR_ENABLE, 1);
+		break;
+	default:
+		break;
+	}
+
+	if (actual_pix_clk_khz >= HDMI_CLOCK_CHANNEL_RATE_MORE_340M) {
+		/* enable HDMI data scrambler
+		 * HDMI_CLOCK_CHANNEL_RATE_MORE_340M
+		 * Clock channel frequency is 1/4 of character rate.
+		 */
+		REG_UPDATE_2(HDMI_CONTROL,
+			HDMI_DATA_SCRAMBLE_EN, 1,
+			HDMI_CLOCK_CHANNEL_RATE, 1);
+	} else if (crtc_timing->flags.LTE_340MCSC_SCRAMBLE) {
+
+		/* TODO: New feature for DCE11, still need to implement */
+
+		/* enable HDMI data scrambler
+		 * HDMI_CLOCK_CHANNEL_FREQ_EQUAL_TO_CHAR_RATE
+		 * Clock channel frequency is the same
+		 * as character rate
+		 */
+		REG_UPDATE_2(HDMI_CONTROL,
+			HDMI_DATA_SCRAMBLE_EN, 1,
+			HDMI_CLOCK_CHANNEL_RATE, 0);
+	}
+
+
+	/* Enable transmission of General Control packet on every frame */
+	REG_UPDATE_3(HDMI_VBI_PACKET_CONTROL,
+		HDMI_GC_CONT, 1,
+		HDMI_GC_SEND, 1,
+		HDMI_NULL_SEND, 1);
+
+	/* Disable Audio Content Protection packet transmission */
+	REG_UPDATE(HDMI_VBI_PACKET_CONTROL, HDMI_ACP_SEND, 0);
+
+	/* following belongs to audio */
+	/* Enable Audio InfoFrame packet transmission. */
+	REG_UPDATE(HDMI_INFOFRAME_CONTROL0, HDMI_AUDIO_INFO_SEND, 1);
+
+	/* update double-buffered AUDIO_INFO registers immediately */
+	ASSERT(enc->afmt);
+	enc->afmt->funcs->audio_info_immediate_update(enc->afmt);
+
+	/* Select line number on which to send Audio InfoFrame packets */
+	REG_UPDATE(HDMI_INFOFRAME_CONTROL1, HDMI_AUDIO_INFO_LINE,
+				VBI_LINE_0 + 2);
+
+	/* set HDMI GC AVMUTE */
+	REG_UPDATE(HDMI_GC, HDMI_GC_AVMUTE, 0);
+	switch (crtc_timing->pixel_encoding) {
+	case PIXEL_ENCODING_YCBCR422:
+		REG_UPDATE(HDMI_CONTROL, TMDS_PIXEL_ENCODING, 1);
+	break;
+	default:
+		REG_UPDATE(HDMI_CONTROL, TMDS_PIXEL_ENCODING, 0);
+	break;
+	}
+	REG_UPDATE(HDMI_CONTROL, TMDS_COLOR_FORMAT, 0);
+}
+
+
+
+static void enc35_stream_encoder_enable(
+	struct stream_encoder *enc,
+	enum signal_type signal,
+	bool enable)
+{
+	struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc);
+
+	if (enable) {
+		switch (signal) {
+		case SIGNAL_TYPE_DVI_SINGLE_LINK:
+		case SIGNAL_TYPE_DVI_DUAL_LINK:
+			/* TMDS-DVI */
+			REG_UPDATE(DIG_FE_CLK_CNTL, DIG_FE_MODE, 2);
+			break;
+		case SIGNAL_TYPE_HDMI_TYPE_A:
+			/* TMDS-HDMI */
+			REG_UPDATE(DIG_FE_CLK_CNTL, DIG_FE_MODE, 3);
+			break;
+		case SIGNAL_TYPE_DISPLAY_PORT_MST:
+			/* DP MST */
+			REG_UPDATE(DIG_FE_CLK_CNTL, DIG_FE_MODE, 5);
+			break;
+		case SIGNAL_TYPE_EDP:
+		case SIGNAL_TYPE_DISPLAY_PORT:
+		case SIGNAL_TYPE_VIRTUAL:
+			/* DP SST */
+			REG_UPDATE(DIG_FE_CLK_CNTL, DIG_FE_MODE, 0);
+			break;
+		default:
+			/* invalid mode ! */
+			ASSERT_CRITICAL(false);
+		}
+	}
+}
+
+static bool is_two_pixels_per_containter(const struct dc_crtc_timing *timing)
+{
+	bool two_pix = timing->pixel_encoding == PIXEL_ENCODING_YCBCR420;
+
+	two_pix = two_pix || (timing->flags.DSC && timing->pixel_encoding == PIXEL_ENCODING_YCBCR422
+			&& !timing->dsc_cfg.ycbcr422_simple);
+	return two_pix;
+}
+
+static void enc35_stream_encoder_dp_unblank(
+		struct dc_link *link,
+		struct stream_encoder *enc,
+		const struct encoder_unblank_param *param)
+{
+	struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc);
+
+	if (param->link_settings.link_rate != LINK_RATE_UNKNOWN) {
+		uint32_t n_vid = 0x8000;
+		uint32_t m_vid;
+		uint32_t n_multiply = 0;
+		uint32_t pix_per_cycle = 0;
+		uint64_t m_vid_l = n_vid;
+
+		/* YCbCr 4:2:0 : Computed VID_M will be 2X the input rate */
+		if (is_two_pixels_per_containter(&param->timing) || param->opp_cnt > 1
+			|| param->pix_per_cycle > 1) {
+			/*this logic should be the same in get_pixel_clock_parameters() */
+			n_multiply = 1;
+			pix_per_cycle = 1;
+		}
+		/* M / N = Fstream / Flink
+		 * m_vid / n_vid = pixel rate / link rate
+		 */
+
+		m_vid_l *= param->timing.pix_clk_100hz / 10;
+		m_vid_l = div_u64(m_vid_l,
+			param->link_settings.link_rate
+				* LINK_RATE_REF_FREQ_IN_KHZ);
+
+		m_vid = (uint32_t) m_vid_l;
+
+		/* enable auto measurement */
+
+		REG_UPDATE(DP_VID_TIMING, DP_VID_M_N_GEN_EN, 0);
+
+		/* auto measurement need 1 full 0x8000 symbol cycle to kick in,
+		 * therefore program initial value for Mvid and Nvid
+		 */
+
+		REG_UPDATE(DP_VID_N, DP_VID_N, n_vid);
+
+		REG_UPDATE(DP_VID_M, DP_VID_M, m_vid);
+
+		REG_UPDATE_2(DP_VID_TIMING,
+				DP_VID_M_N_GEN_EN, 1,
+				DP_VID_N_MUL, n_multiply);
+
+		REG_UPDATE(DP_PIXEL_FORMAT,
+				DP_PIXEL_PER_CYCLE_PROCESSING_MODE,
+				pix_per_cycle);
+	}
+
+	/* make sure stream is disabled before resetting steer fifo */
+	REG_UPDATE(DP_VID_STREAM_CNTL, DP_VID_STREAM_ENABLE, false);
+	REG_WAIT(DP_VID_STREAM_CNTL, DP_VID_STREAM_STATUS, 0, 10, 5000);
+
+	/* DIG_START is removed from the register spec */
+
+	/* switch DP encoder to CRTC data, but reset it the fifo first. It may happen
+	 * that it overflows during mode transition, and sometimes doesn't recover.
+	 */
+	REG_UPDATE(DP_STEER_FIFO, DP_STEER_FIFO_RESET, 1);
+	udelay(10);
+
+	REG_UPDATE(DP_STEER_FIFO, DP_STEER_FIFO_RESET, 0);
+
+	/* wait 100us for DIG/DP logic to prime
+	 * (i.e. a few video lines)
+	 */
+	udelay(100);
+
+	/* the hardware would start sending video at the start of the next DP
+	 * frame (i.e. rising edge of the vblank).
+	 * NOTE: We used to program DP_VID_STREAM_DIS_DEFER = 2 here, but this
+	 * register has no effect on enable transition! HW always makes sure
+	 * VID_STREAM enable at start of next frame, and this is not
+	 * programmable
+	 */
+
+	REG_UPDATE(DP_VID_STREAM_CNTL, DP_VID_STREAM_ENABLE, true);
+
+	/*
+	 * DIG Resync FIFO now needs to be explicitly enabled.
+	 * This should come after DP_VID_STREAM_ENABLE per HW docs.
+	 */
+	enc314_enable_fifo(enc);
+
+	link->dc->link_srv->dp_trace_source_sequence(link, DPCD_SOURCE_SEQ_AFTER_ENABLE_DP_VID_STREAM);
+}
+
+static void enc35_stream_encoder_map_to_link(
+		struct stream_encoder *enc,
+		uint32_t stream_enc_inst,
+		uint32_t link_enc_inst)
+{
+	struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc);
+
+	ASSERT(stream_enc_inst < 5 && link_enc_inst < 5);
+	REG_UPDATE(STREAM_MAPPER_CONTROL,
+				DIG_STREAM_LINK_TARGET, link_enc_inst);
+}
+
+static void enc35_reset_fifo(struct stream_encoder *enc, bool reset)
+{
+	struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc);
+	uint32_t reset_val = reset ? 1 : 0;
+	uint32_t is_symclk_on;
+
+	REG_UPDATE(DIG_FIFO_CTRL0, DIG_FIFO_RESET, reset_val);
+	REG_GET(DIG_FE_CLK_CNTL, DIG_FE_SYMCLK_FE_G_CLOCK_ON, &is_symclk_on);
+
+	if (is_symclk_on)
+		REG_WAIT(DIG_FIFO_CTRL0, DIG_FIFO_RESET_DONE, reset_val, 10, 5000);
+	else
+		udelay(10);
+}
+
+static bool enc35_is_fifo_enabled(struct stream_encoder *enc)
+{
+	struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc);
+	uint32_t reset_val;
+
+	REG_GET(DIG_FIFO_CTRL0, DIG_FIFO_ENABLE, &reset_val);
+	return reset_val != 0;
+}
+void enc35_disable_fifo(struct stream_encoder *enc)
+{
+	struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc);
+
+	REG_UPDATE(DIG_FIFO_CTRL0, DIG_FIFO_ENABLE, 0);
+	REG_UPDATE(DIG_FE_EN_CNTL, DIG_FE_ENABLE, 0);
+	REG_UPDATE(DIG_FE_CLK_CNTL, DIG_FE_CLK_EN, 0);
+}
+
+void enc35_enable_fifo(struct stream_encoder *enc)
+{
+	struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc);
+
+	REG_UPDATE(DIG_FIFO_CTRL0, DIG_FIFO_READ_START_LEVEL, 0x7);
+	REG_UPDATE(DIG_FE_CLK_CNTL, DIG_FE_CLK_EN, 1);
+	REG_UPDATE(DIG_FE_EN_CNTL, DIG_FE_ENABLE, 1);
+
+	enc35_reset_fifo(enc, true);
+	enc35_reset_fifo(enc, false);
+
+	REG_UPDATE(DIG_FIFO_CTRL0, DIG_FIFO_ENABLE, 1);
+}
+
+static uint32_t enc35_get_pixels_per_cycle(struct stream_encoder *enc)
+{
+	struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc);
+	uint32_t value;
+
+	REG_GET(DIG_FIFO_CTRL0, DIG_FIFO_OUTPUT_PIXEL_MODE, &value);
+
+	switch (value) {
+	case 0:
+		return 1;
+	case 1:
+		return 2;
+	default:
+		ASSERT_CRITICAL(false);
+		return 1;
+	}
+}
+
+static const struct stream_encoder_funcs dcn35_str_enc_funcs = {
+	.dp_set_odm_combine =
+		enc314_dp_set_odm_combine,
+	.dp_set_stream_attribute =
+		enc2_stream_encoder_dp_set_stream_attribute,
+	.hdmi_set_stream_attribute =
+		enc35_stream_encoder_hdmi_set_stream_attribute,
+	.dvi_set_stream_attribute =
+		enc35_stream_encoder_dvi_set_stream_attribute,
+	.set_throttled_vcp_size =
+		enc1_stream_encoder_set_throttled_vcp_size,
+	.update_hdmi_info_packets =
+		enc3_stream_encoder_update_hdmi_info_packets,
+	.stop_hdmi_info_packets =
+		enc3_stream_encoder_stop_hdmi_info_packets,
+	.update_dp_info_packets_sdp_line_num =
+		enc3_stream_encoder_update_dp_info_packets_sdp_line_num,
+	.update_dp_info_packets =
+		enc3_stream_encoder_update_dp_info_packets,
+	.stop_dp_info_packets =
+		enc1_stream_encoder_stop_dp_info_packets,
+	.dp_blank =
+		enc314_stream_encoder_dp_blank,
+	.dp_unblank =
+		enc35_stream_encoder_dp_unblank,
+	.audio_mute_control = enc3_audio_mute_control,
+
+	.dp_audio_setup = enc3_se_dp_audio_setup,
+	.dp_audio_enable = enc3_se_dp_audio_enable,
+	.dp_audio_disable = enc1_se_dp_audio_disable,
+
+	.hdmi_audio_setup = enc3_se_hdmi_audio_setup,
+	.hdmi_audio_disable = enc1_se_hdmi_audio_disable,
+	.setup_stereo_sync  = enc1_setup_stereo_sync,
+	.set_avmute = enc1_stream_encoder_set_avmute,
+	.dig_connect_to_otg = enc1_dig_connect_to_otg,
+	.dig_source_otg = enc1_dig_source_otg,
+
+	.dp_get_pixel_format  = enc1_stream_encoder_dp_get_pixel_format,
+
+	.enc_read_state = enc314_read_state,
+	.dp_set_dsc_config = enc314_dp_set_dsc_config,
+	.dp_set_dsc_pps_info_packet = enc3_dp_set_dsc_pps_info_packet,
+	.set_dynamic_metadata = enc2_set_dynamic_metadata,
+	.hdmi_reset_stream_attribute = enc1_reset_hdmi_stream_attribute,
+	.enable_stream = enc35_stream_encoder_enable,
+
+	.set_input_mode = enc314_set_dig_input_mode,
+	.enable_fifo = enc35_enable_fifo,
+	.disable_fifo = enc35_disable_fifo,
+	.is_fifo_enabled = enc35_is_fifo_enabled,
+	.map_stream_to_link = enc35_stream_encoder_map_to_link,
+	.get_pixels_per_cycle = enc35_get_pixels_per_cycle,
+};
+
+void dcn35_dio_stream_encoder_construct(
+	struct dcn10_stream_encoder *enc1,
+	struct dc_context *ctx,
+	struct dc_bios *bp,
+	enum engine_id eng_id,
+	struct vpg *vpg,
+	struct afmt *afmt,
+	const struct dcn10_stream_enc_registers *regs,
+	const struct dcn10_stream_encoder_shift *se_shift,
+	const struct dcn10_stream_encoder_mask *se_mask)
+{
+	enc1->base.funcs = &dcn35_str_enc_funcs;
+	enc1->base.ctx = ctx;
+	enc1->base.id = eng_id;
+	enc1->base.bp = bp;
+	enc1->base.vpg = vpg;
+	enc1->base.afmt = afmt;
+	enc1->regs = regs;
+	enc1->se_shift = se_shift;
+	enc1->se_mask = se_mask;
+	enc1->base.stream_enc_inst = vpg->inst;
+}
+
diff --git a/drivers/gpu/drm/amd/display/dc/dio/dcn35/dcn35_dio_stream_encoder.h b/drivers/gpu/drm/amd/display/dc/dio/dcn35/dcn35_dio_stream_encoder.h
new file mode 100644
index 000000000000..ddb33fdfb4ee
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dio/dcn35/dcn35_dio_stream_encoder.h
@@ -0,0 +1,332 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __DC_DIO_STREAM_ENCODER_DCN35_H__
+#define __DC_DIO_STREAM_ENCODER_DCN35_H__
+
+#include "dcn30/dcn30_vpg.h"
+#include "dcn30/dcn30_afmt.h"
+#include "stream_encoder.h"
+#include "dcn20/dcn20_stream_encoder.h"
+
+/* Register bit field name change */
+#define RDPCSTX0_RDPCSTX_CLOCK_CNTL__RDPCS_SYMCLK_DIV2_GATE_DIS__SHIFT                                        0x8
+#define RDPCSTX0_RDPCSTX_CLOCK_CNTL__RDPCS_SYMCLK_DIV2_EN__SHIFT                                              0x9
+#define RDPCSTX0_RDPCSTX_CLOCK_CNTL__RDPCS_SYMCLK_DIV2_CLOCK_ON__SHIFT                                        0xa
+#define DPCSTX0_DPCSTX_TX_CNTL__DPCS_TX_DATA_SWAP__SHIFT                                                      0xe
+#define DPCSTX0_DPCSTX_TX_CNTL__DPCS_TX_DATA_ORDER_INVERT__SHIFT                                              0xf
+
+#define RDPCSTX0_RDPCSTX_CLOCK_CNTL__RDPCS_SYMCLK_DIV2_GATE_DIS_MASK                                          0x00000100L
+#define RDPCSTX0_RDPCSTX_CLOCK_CNTL__RDPCS_SYMCLK_DIV2_EN_MASK                                                0x00000200L
+#define RDPCSTX0_RDPCSTX_CLOCK_CNTL__RDPCS_SYMCLK_DIV2_CLOCK_ON_MASK                                          0x00000400L
+#define DPCSTX0_DPCSTX_TX_CNTL__DPCS_TX_DATA_SWAP_MASK                                                        0x00004000L
+#define DPCSTX0_DPCSTX_TX_CNTL__DPCS_TX_DATA_ORDER_INVERT_MASK                                                0x00008000L
+
+
+#define SE_DCN35_REG_LIST(id)\
+	SRI(AFMT_CNTL, DIG, id), \
+	SRI(DIG_FE_CNTL, DIG, id), \
+	SRI(HDMI_CONTROL, DIG, id), \
+	SRI(HDMI_DB_CONTROL, DIG, id), \
+	SRI(HDMI_GC, DIG, id), \
+	SRI(HDMI_GENERIC_PACKET_CONTROL0, DIG, id), \
+	SRI(HDMI_GENERIC_PACKET_CONTROL1, DIG, id), \
+	SRI(HDMI_GENERIC_PACKET_CONTROL2, DIG, id), \
+	SRI(HDMI_GENERIC_PACKET_CONTROL3, DIG, id), \
+	SRI(HDMI_GENERIC_PACKET_CONTROL4, DIG, id), \
+	SRI(HDMI_GENERIC_PACKET_CONTROL5, DIG, id), \
+	SRI(HDMI_GENERIC_PACKET_CONTROL6, DIG, id), \
+	SRI(HDMI_GENERIC_PACKET_CONTROL7, DIG, id), \
+	SRI(HDMI_GENERIC_PACKET_CONTROL8, DIG, id), \
+	SRI(HDMI_GENERIC_PACKET_CONTROL9, DIG, id), \
+	SRI(HDMI_GENERIC_PACKET_CONTROL10, DIG, id), \
+	SRI(HDMI_INFOFRAME_CONTROL0, DIG, id), \
+	SRI(HDMI_INFOFRAME_CONTROL1, DIG, id), \
+	SRI(HDMI_VBI_PACKET_CONTROL, DIG, id), \
+	SRI(HDMI_AUDIO_PACKET_CONTROL, DIG, id),\
+	SRI(HDMI_ACR_PACKET_CONTROL, DIG, id),\
+	SRI(HDMI_ACR_32_0, DIG, id),\
+	SRI(HDMI_ACR_32_1, DIG, id),\
+	SRI(HDMI_ACR_44_0, DIG, id),\
+	SRI(HDMI_ACR_44_1, DIG, id),\
+	SRI(HDMI_ACR_48_0, DIG, id),\
+	SRI(HDMI_ACR_48_1, DIG, id),\
+	SRI(DP_DB_CNTL, DP, id), \
+	SRI(DP_MSA_MISC, DP, id), \
+	SRI(DP_MSA_VBID_MISC, DP, id), \
+	SRI(DP_MSA_COLORIMETRY, DP, id), \
+	SRI(DP_MSA_TIMING_PARAM1, DP, id), \
+	SRI(DP_MSA_TIMING_PARAM2, DP, id), \
+	SRI(DP_MSA_TIMING_PARAM3, DP, id), \
+	SRI(DP_MSA_TIMING_PARAM4, DP, id), \
+	SRI(DP_MSE_RATE_CNTL, DP, id), \
+	SRI(DP_MSE_RATE_UPDATE, DP, id), \
+	SRI(DP_PIXEL_FORMAT, DP, id), \
+	SRI(DP_SEC_CNTL, DP, id), \
+	SRI(DP_SEC_CNTL1, DP, id), \
+	SRI(DP_SEC_CNTL2, DP, id), \
+	SRI(DP_SEC_CNTL5, DP, id), \
+	SRI(DP_SEC_CNTL6, DP, id), \
+	SRI(DP_STEER_FIFO, DP, id), \
+	SRI(DP_VID_M, DP, id), \
+	SRI(DP_VID_N, DP, id), \
+	SRI(DP_VID_STREAM_CNTL, DP, id), \
+	SRI(DP_VID_TIMING, DP, id), \
+	SRI(DP_SEC_AUD_N, DP, id), \
+	SRI(DP_SEC_TIMESTAMP, DP, id), \
+	SRI(DP_DSC_CNTL, DP, id), \
+	SRI(DP_SEC_METADATA_TRANSMISSION, DP, id), \
+	SRI(HDMI_METADATA_PACKET_CONTROL, DIG, id), \
+	SRI(DP_SEC_FRAMING4, DP, id), \
+	SRI(DP_GSP11_CNTL, DP, id), \
+	SRI(DME_CONTROL, DME, id),\
+	SRI(DP_SEC_METADATA_TRANSMISSION, DP, id), \
+	SRI(HDMI_METADATA_PACKET_CONTROL, DIG, id), \
+	SRI(DIG_FE_CNTL, DIG, id), \
+	SRI(DIG_FE_EN_CNTL, DIG, id), \
+	SRI(DIG_FE_CLK_CNTL, DIG, id), \
+	SRI(DIG_CLOCK_PATTERN, DIG, id), \
+	SRI(DIG_FIFO_CTRL0, DIG, id),\
+	SRI(STREAM_MAPPER_CONTROL, DIG, id)
+
+
+#define SE_COMMON_MASK_SH_LIST_DCN35(mask_sh)\
+	SE_SF(DP0_DP_PIXEL_FORMAT, DP_PIXEL_ENCODING, mask_sh),\
+	SE_SF(DP0_DP_PIXEL_FORMAT, DP_COMPONENT_DEPTH, mask_sh),\
+	SE_SF(DP0_DP_PIXEL_FORMAT, DP_PIXEL_PER_CYCLE_PROCESSING_MODE, mask_sh),\
+	SE_SF(DIG0_HDMI_CONTROL, HDMI_PACKET_GEN_VERSION, mask_sh),\
+	SE_SF(DIG0_HDMI_CONTROL, HDMI_KEEPOUT_MODE, mask_sh),\
+	SE_SF(DIG0_HDMI_CONTROL, HDMI_DEEP_COLOR_ENABLE, mask_sh),\
+	SE_SF(DIG0_HDMI_CONTROL, HDMI_DEEP_COLOR_DEPTH, mask_sh),\
+	SE_SF(DIG0_HDMI_CONTROL, HDMI_DATA_SCRAMBLE_EN, mask_sh),\
+	SE_SF(DIG0_HDMI_CONTROL, HDMI_NO_EXTRA_NULL_PACKET_FILLED, mask_sh),\
+	SE_SF(DIG0_HDMI_VBI_PACKET_CONTROL, HDMI_GC_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_VBI_PACKET_CONTROL, HDMI_GC_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_VBI_PACKET_CONTROL, HDMI_NULL_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_VBI_PACKET_CONTROL, HDMI_ACP_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_INFOFRAME_CONTROL0, HDMI_AUDIO_INFO_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_INFOFRAME_CONTROL1, HDMI_AUDIO_INFO_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GC, HDMI_GC_AVMUTE, mask_sh),\
+	SE_SF(DP0_DP_MSE_RATE_CNTL, DP_MSE_RATE_X, mask_sh),\
+	SE_SF(DP0_DP_MSE_RATE_CNTL, DP_MSE_RATE_Y, mask_sh),\
+	SE_SF(DP0_DP_MSE_RATE_UPDATE, DP_MSE_RATE_UPDATE_PENDING, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_GSP0_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_STREAM_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_GSP1_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_GSP2_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_GSP3_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_MPG_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL1, DP_SEC_GSP5_LINE_REFERENCE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL2, DP_SEC_GSP4_SEND, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL2, DP_SEC_GSP4_SEND_PENDING, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL4, DP_SEC_GSP4_LINE_NUM, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL5, DP_SEC_GSP5_LINE_NUM, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL2, DP_SEC_GSP4_SEND_ANY_LINE, mask_sh),\
+	SE_SF(DP0_DP_VID_STREAM_CNTL, DP_VID_STREAM_DIS_DEFER, mask_sh),\
+	SE_SF(DP0_DP_VID_STREAM_CNTL, DP_VID_STREAM_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_VID_STREAM_CNTL, DP_VID_STREAM_STATUS, mask_sh),\
+	SE_SF(DP0_DP_STEER_FIFO, DP_STEER_FIFO_RESET, mask_sh),\
+	SE_SF(DP0_DP_VID_TIMING, DP_VID_M_N_GEN_EN, mask_sh),\
+	SE_SF(DP0_DP_VID_N, DP_VID_N, mask_sh),\
+	SE_SF(DP0_DP_VID_M, DP_VID_M, mask_sh),\
+	SE_SF(DIG0_HDMI_AUDIO_PACKET_CONTROL, HDMI_AUDIO_DELAY_EN, mask_sh),\
+	SE_SF(DIG0_HDMI_ACR_PACKET_CONTROL, HDMI_ACR_AUTO_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_ACR_PACKET_CONTROL, HDMI_ACR_SOURCE, mask_sh),\
+	SE_SF(DIG0_HDMI_ACR_PACKET_CONTROL, HDMI_ACR_AUDIO_PRIORITY, mask_sh),\
+	SE_SF(DIG0_HDMI_ACR_32_0, HDMI_ACR_CTS_32, mask_sh),\
+	SE_SF(DIG0_HDMI_ACR_32_1, HDMI_ACR_N_32, mask_sh),\
+	SE_SF(DIG0_HDMI_ACR_44_0, HDMI_ACR_CTS_44, mask_sh),\
+	SE_SF(DIG0_HDMI_ACR_44_1, HDMI_ACR_N_44, mask_sh),\
+	SE_SF(DIG0_HDMI_ACR_48_0, HDMI_ACR_CTS_48, mask_sh),\
+	SE_SF(DIG0_HDMI_ACR_48_1, HDMI_ACR_N_48, mask_sh),\
+	SE_SF(DP0_DP_SEC_AUD_N, DP_SEC_AUD_N, mask_sh),\
+	SE_SF(DP0_DP_SEC_TIMESTAMP, DP_SEC_TIMESTAMP_MODE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_ASP_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_ATP_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_AIP_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_ACM_ENABLE, mask_sh),\
+	SE_SF(DIG0_AFMT_CNTL, AFMT_AUDIO_CLOCK_EN, mask_sh),\
+	SE_SF(DIG0_HDMI_CONTROL, HDMI_CLOCK_CHANNEL_RATE, mask_sh),\
+	SE_SF(DIG0_HDMI_CONTROL, TMDS_PIXEL_ENCODING, mask_sh),\
+	SE_SF(DIG0_HDMI_CONTROL, TMDS_COLOR_FORMAT, mask_sh),\
+	SE_SF(DIG0_DIG_FE_CNTL, DIG_STEREOSYNC_SELECT, mask_sh),\
+	SE_SF(DIG0_DIG_FE_CNTL, DIG_STEREOSYNC_GATE_EN, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_GSP4_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_GSP5_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_GSP6_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_GSP7_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL2, DP_SEC_GSP7_SEND, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL6, DP_SEC_GSP7_LINE_NUM, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL2, DP_SEC_GSP11_PPS, mask_sh),\
+	SE_SF(DP0_DP_GSP11_CNTL, DP_SEC_GSP11_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_GSP11_CNTL, DP_SEC_GSP11_LINE_NUM, mask_sh),\
+	SE_SF(DP0_DP_DB_CNTL, DP_DB_DISABLE, mask_sh),\
+	SE_SF(DP0_DP_MSA_COLORIMETRY, DP_MSA_MISC0, mask_sh),\
+	SE_SF(DP0_DP_MSA_TIMING_PARAM1, DP_MSA_HTOTAL, mask_sh),\
+	SE_SF(DP0_DP_MSA_TIMING_PARAM1, DP_MSA_VTOTAL, mask_sh),\
+	SE_SF(DP0_DP_MSA_TIMING_PARAM2, DP_MSA_HSTART, mask_sh),\
+	SE_SF(DP0_DP_MSA_TIMING_PARAM2, DP_MSA_VSTART, mask_sh),\
+	SE_SF(DP0_DP_MSA_TIMING_PARAM3, DP_MSA_HSYNCWIDTH, mask_sh),\
+	SE_SF(DP0_DP_MSA_TIMING_PARAM3, DP_MSA_HSYNCPOLARITY, mask_sh),\
+	SE_SF(DP0_DP_MSA_TIMING_PARAM3, DP_MSA_VSYNCWIDTH, mask_sh),\
+	SE_SF(DP0_DP_MSA_TIMING_PARAM3, DP_MSA_VSYNCPOLARITY, mask_sh),\
+	SE_SF(DP0_DP_MSA_TIMING_PARAM4, DP_MSA_HWIDTH, mask_sh),\
+	SE_SF(DP0_DP_MSA_TIMING_PARAM4, DP_MSA_VHEIGHT, mask_sh),\
+	SE_SF(DIG0_HDMI_DB_CONTROL, HDMI_DB_DISABLE, mask_sh),\
+	SE_SF(DP0_DP_VID_TIMING, DP_VID_N_MUL, mask_sh),\
+	SE_SF(DIG0_DIG_FE_CNTL, DIG_SOURCE_SELECT, mask_sh), \
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC0_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC0_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC1_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC1_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC2_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC2_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC3_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC3_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC4_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC4_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC5_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC5_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC6_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC6_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC7_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC7_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC8_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC8_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC9_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC9_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC10_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC10_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC11_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC11_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC12_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC12_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC13_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC13_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC14_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC14_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL1, HDMI_GENERIC0_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL1, HDMI_GENERIC1_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL2, HDMI_GENERIC2_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL2, HDMI_GENERIC3_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL3, HDMI_GENERIC4_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL3, HDMI_GENERIC5_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL4, HDMI_GENERIC6_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL4, HDMI_GENERIC7_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL7, HDMI_GENERIC8_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL7, HDMI_GENERIC9_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL8, HDMI_GENERIC10_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL8, HDMI_GENERIC11_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL9, HDMI_GENERIC12_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL9, HDMI_GENERIC13_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL10, HDMI_GENERIC14_LINE, mask_sh),\
+	SE_SF(DP0_DP_DSC_CNTL, DP_DSC_MODE, mask_sh),\
+	SE_SF(DP0_DP_MSA_VBID_MISC, DP_VBID6_LINE_REFERENCE, mask_sh),\
+	SE_SF(DP0_DP_MSA_VBID_MISC, DP_VBID6_LINE_NUM, mask_sh),\
+	SE_SF(DME0_DME_CONTROL, METADATA_ENGINE_EN, mask_sh),\
+	SE_SF(DME0_DME_CONTROL, METADATA_HUBP_REQUESTOR_ID, mask_sh),\
+	SE_SF(DME0_DME_CONTROL, METADATA_STREAM_TYPE, mask_sh),\
+	SE_SF(DP0_DP_SEC_METADATA_TRANSMISSION, DP_SEC_METADATA_PACKET_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_METADATA_TRANSMISSION, DP_SEC_METADATA_PACKET_LINE_REFERENCE, mask_sh),\
+	SE_SF(DP0_DP_SEC_METADATA_TRANSMISSION, DP_SEC_METADATA_PACKET_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_METADATA_PACKET_CONTROL, HDMI_METADATA_PACKET_ENABLE, mask_sh),\
+	SE_SF(DIG0_HDMI_METADATA_PACKET_CONTROL, HDMI_METADATA_PACKET_LINE_REFERENCE, mask_sh),\
+	SE_SF(DIG0_HDMI_METADATA_PACKET_CONTROL, HDMI_METADATA_PACKET_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_CONTROL, DOLBY_VISION_EN, mask_sh),\
+	SE_SF(DIG0_DIG_FE_EN_CNTL, DIG_FE_ENABLE, mask_sh),\
+	SE_SF(DIG0_DIG_FE_CLK_CNTL, DIG_FE_MODE, mask_sh),\
+	SE_SF(DIG0_DIG_FE_CLK_CNTL, DIG_FE_CLK_EN, mask_sh),\
+	SE_SF(DIG0_DIG_FE_CLK_CNTL, DIG_FE_SOFT_RESET, mask_sh),\
+	SE_SF(DIG0_DIG_FE_CLK_CNTL, DIG_FE_DISPCLK_G_CLOCK_ON, mask_sh),\
+	SE_SF(DIG0_DIG_FE_CLK_CNTL, DIG_FE_SYMCLK_FE_G_CLOCK_ON, mask_sh),\
+	SE_SF(DIG0_DIG_FE_CLK_CNTL, DIG_FE_SYMCLK_FE_G_AFMT_CLOCK_ON, mask_sh),\
+	SE_SF(DIG0_DIG_FE_CLK_CNTL, DIG_FE_SYMCLK_FE_G_TMDS_CLOCK_ON, mask_sh),\
+	SE_SF(DIG0_DIG_FE_CLK_CNTL, DIG_FE_SOCCLK_G_AFMT_CLOCK_ON, mask_sh),\
+	SE_SF(DP0_DP_SEC_FRAMING4, DP_SST_SDP_SPLITTING, mask_sh),\
+	SE_SF(DIG0_DIG_CLOCK_PATTERN, DIG_CLOCK_PATTERN, mask_sh),\
+	SE_SF(DIG0_DIG_FIFO_CTRL0, DIG_FIFO_OUTPUT_PIXEL_MODE, mask_sh),\
+	SE_SF(DIG0_DIG_FIFO_CTRL0, DIG_FIFO_READ_START_LEVEL, mask_sh),\
+	SE_SF(DIG0_DIG_FIFO_CTRL0, DIG_FIFO_ENABLE, mask_sh),\
+	SE_SF(DIG0_DIG_FIFO_CTRL0, DIG_FIFO_RESET, mask_sh),\
+	SE_SF(DIG0_DIG_FIFO_CTRL0, DIG_FIFO_RESET_DONE, mask_sh),\
+	SE_SF(DIG0_STREAM_MAPPER_CONTROL, DIG_STREAM_LINK_TARGET, mask_sh),
+
+void dcn35_dio_stream_encoder_construct(
+	struct dcn10_stream_encoder *enc1,
+	struct dc_context *ctx,
+	struct dc_bios *bp,
+	enum engine_id eng_id,
+	struct vpg *vpg,
+	struct afmt *afmt,
+	const struct dcn10_stream_enc_registers *regs,
+	const struct dcn10_stream_encoder_shift *se_shift,
+	const struct dcn10_stream_encoder_mask *se_mask);
+
+void enc3_stream_encoder_update_hdmi_info_packets(
+	struct stream_encoder *enc,
+	const struct encoder_info_frame *info_frame);
+
+void enc3_stream_encoder_stop_hdmi_info_packets(
+	struct stream_encoder *enc);
+
+void enc3_stream_encoder_update_dp_info_packets_sdp_line_num(
+		struct stream_encoder *enc,
+		struct encoder_info_frame *info_frame);
+
+void enc3_stream_encoder_update_dp_info_packets(
+	struct stream_encoder *enc,
+	const struct encoder_info_frame *info_frame);
+
+void enc3_audio_mute_control(
+	struct stream_encoder *enc,
+	bool mute);
+
+void enc3_se_dp_audio_setup(
+	struct stream_encoder *enc,
+	unsigned int az_inst,
+	struct audio_info *info);
+
+void enc3_se_dp_audio_enable(
+	struct stream_encoder *enc);
+
+void enc3_se_hdmi_audio_setup(
+	struct stream_encoder *enc,
+	unsigned int az_inst,
+	struct audio_info *info,
+	struct audio_crtc_info *audio_crtc_info);
+
+void enc3_dp_set_dsc_pps_info_packet(
+	struct stream_encoder *enc,
+	bool enable,
+	uint8_t *dsc_packed_pps,
+	bool immediate_update);
+
+void enc35_disable_fifo(
+	struct stream_encoder *enc);
+
+void enc35_enable_fifo(
+	struct stream_encoder *enc);
+
+
+#endif /* __DC_DIO_STREAM_ENCODER_DCN35_H__ */
diff --git a/drivers/gpu/drm/amd/display/dc/dio/dcn401/dcn401_dio_link_encoder.c b/drivers/gpu/drm/amd/display/dc/dio/dcn401/dcn401_dio_link_encoder.c
new file mode 100644
index 000000000000..7e558ca195ef
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dio/dcn401/dcn401_dio_link_encoder.c
@@ -0,0 +1,322 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+
+#include "reg_helper.h"
+
+#include "core_types.h"
+#include "link_encoder.h"
+#include "dcn31/dcn31_dio_link_encoder.h"
+#include "dcn32/dcn32_dio_link_encoder.h"
+#include "dcn401_dio_link_encoder.h"
+#include "stream_encoder.h"
+#include "dc_bios_types.h"
+
+#include "gpio_service_interface.h"
+
+#ifndef MIN
+#define MIN(X, Y) ((X) < (Y) ? (X) : (Y))
+#endif
+
+#define CTX \
+	enc10->base.ctx
+#define DC_LOGGER \
+	enc10->base.ctx->logger
+
+#define REG(reg)\
+	(enc10->link_regs->reg)
+
+#undef FN
+#define FN(reg_name, field_name) \
+	enc10->link_shift->field_name, enc10->link_mask->field_name
+
+#define AUX_REG(reg)\
+	(enc10->aux_regs->reg)
+
+#define AUX_REG_READ(reg_name) \
+		dm_read_reg(CTX, AUX_REG(reg_name))
+
+#define AUX_REG_WRITE(reg_name, val) \
+			dm_write_reg(CTX, AUX_REG(reg_name), val)
+
+#ifndef MIN
+#define MIN(X, Y) ((X) < (Y) ? (X) : (Y))
+#endif
+
+void enc401_hw_init(struct link_encoder *enc)
+{
+	struct dcn10_link_encoder *enc10 = TO_DCN10_LINK_ENC(enc);
+
+/*
+	00 - DP_AUX_DPHY_RX_DETECTION_THRESHOLD__1to2 : 1/2
+	01 - DP_AUX_DPHY_RX_DETECTION_THRESHOLD__3to4 : 3/4
+	02 - DP_AUX_DPHY_RX_DETECTION_THRESHOLD__7to8 : 7/8
+	03 - DP_AUX_DPHY_RX_DETECTION_THRESHOLD__15to16 : 15/16
+	04 - DP_AUX_DPHY_RX_DETECTION_THRESHOLD__31to32 : 31/32
+	05 - DP_AUX_DPHY_RX_DETECTION_THRESHOLD__63to64 : 63/64
+	06 - DP_AUX_DPHY_RX_DETECTION_THRESHOLD__127to128 : 127/128
+	07 - DP_AUX_DPHY_RX_DETECTION_THRESHOLD__255to256 : 255/256
+*/
+
+/*
+	AUX_REG_UPDATE_5(AUX_DPHY_RX_CONTROL0,
+	AUX_RX_START_WINDOW = 1 [6:4]
+	AUX_RX_RECEIVE_WINDOW = 1 default is 2 [10:8]
+	AUX_RX_HALF_SYM_DETECT_LEN  = 1 [13:12] default is 1
+	AUX_RX_TRANSITION_FILTER_EN = 1 [16] default is 1
+	AUX_RX_ALLOW_BELOW_THRESHOLD_PHASE_DETECT [17] is 0  default is 0
+	AUX_RX_ALLOW_BELOW_THRESHOLD_START [18] is 1  default is 1
+	AUX_RX_ALLOW_BELOW_THRESHOLD_STOP [19] is 1  default is 1
+	AUX_RX_PHASE_DETECT_LEN,  [21,20] = 0x3 default is 3
+	AUX_RX_DETECTION_THRESHOLD [30:28] = 1
+*/
+	AUX_REG_WRITE(AUX_DPHY_RX_CONTROL0, 0x103d1110);
+
+	AUX_REG_WRITE(AUX_DPHY_TX_CONTROL, 0x21c7a);
+
+	//AUX_DPHY_TX_REF_CONTROL'AUX_TX_REF_DIV HW default is 0x32;
+	// Set AUX_TX_REF_DIV Divider to generate 2 MHz reference from refclk
+	// 27MHz -> 0xd
+	// 100MHz -> 0x32
+	// 48MHz -> 0x18
+
+	// Set TMDS_CTL0 to 1.  This is a legacy setting.
+	REG_UPDATE(TMDS_CTL_BITS, TMDS_CTL0, 1);
+
+	dcn10_aux_initialize(enc10);
+}
+
+
+void dcn401_link_encoder_enable_dp_output(
+	struct link_encoder *enc,
+	const struct dc_link_settings *link_settings,
+	enum clock_source_id clock_source)
+{
+	if (!enc->ctx->dc->debug.avoid_vbios_exec_table) {
+		dcn10_link_encoder_enable_dp_output(enc, link_settings, clock_source);
+		return;
+	}
+}
+
+void dcn401_link_encoder_setup(
+	struct link_encoder *enc,
+	enum signal_type signal)
+{
+	struct dcn10_link_encoder *enc10 = TO_DCN10_LINK_ENC(enc);
+
+	switch (signal) {
+	case SIGNAL_TYPE_EDP:
+	case SIGNAL_TYPE_DISPLAY_PORT:
+		/* DP SST */
+		REG_UPDATE(DIG_BE_CLK_CNTL, DIG_BE_MODE, 0);
+		break;
+	case SIGNAL_TYPE_DVI_SINGLE_LINK:
+	case SIGNAL_TYPE_DVI_DUAL_LINK:
+		/* TMDS-DVI */
+		REG_UPDATE(DIG_BE_CLK_CNTL, DIG_BE_MODE, 2);
+		break;
+	case SIGNAL_TYPE_HDMI_TYPE_A:
+		/* TMDS-HDMI */
+		REG_UPDATE(DIG_BE_CLK_CNTL, DIG_BE_MODE, 3);
+		break;
+	case SIGNAL_TYPE_DISPLAY_PORT_MST:
+		/* DP MST */
+		REG_UPDATE(DIG_BE_CLK_CNTL, DIG_BE_MODE, 5);
+		break;
+	default:
+		ASSERT_CRITICAL(false);
+		/* invalid mode ! */
+		break;
+	}
+	REG_UPDATE(DIG_BE_CLK_CNTL, DIG_BE_CLK_EN, 1);
+	REG_UPDATE(DIG_BE_EN_CNTL, DIG_BE_ENABLE, 1);
+}
+
+bool dcn401_is_dig_enabled(struct link_encoder *enc)
+{
+	uint32_t clk_enabled;
+	uint32_t dig_enabled;
+	struct dcn10_link_encoder *enc10 = TO_DCN10_LINK_ENC(enc);
+
+	REG_GET(DIG_BE_CLK_CNTL, DIG_BE_CLK_EN, &clk_enabled);
+	REG_GET(DIG_BE_EN_CNTL, DIG_BE_ENABLE, &dig_enabled);
+	return (clk_enabled == 1 && dig_enabled == 1);
+}
+
+enum signal_type dcn401_get_dig_mode(
+	struct link_encoder *enc)
+{
+	struct dcn10_link_encoder *enc10 = TO_DCN10_LINK_ENC(enc);
+	uint32_t value;
+	REG_GET(DIG_BE_CLK_CNTL, DIG_BE_MODE, &value);
+	switch (value) {
+	case 0:
+		return SIGNAL_TYPE_DISPLAY_PORT;
+	case 2:
+		return SIGNAL_TYPE_DVI_SINGLE_LINK;
+	case 3:
+		return SIGNAL_TYPE_HDMI_TYPE_A;
+	case 5:
+		return SIGNAL_TYPE_DISPLAY_PORT_MST;
+	default:
+		return SIGNAL_TYPE_NONE;
+	}
+}
+
+static const struct link_encoder_funcs dcn401_link_enc_funcs = {
+	.read_state = link_enc2_read_state,
+	.validate_output_with_stream =
+			dcn30_link_encoder_validate_output_with_stream,
+	.hw_init = enc401_hw_init,
+	.setup = dcn401_link_encoder_setup,
+	.enable_tmds_output = dcn10_link_encoder_enable_tmds_output,
+	.enable_dp_output = dcn401_link_encoder_enable_dp_output,
+	.enable_dp_mst_output = dcn10_link_encoder_enable_dp_mst_output,
+	.disable_output = dcn10_link_encoder_disable_output,
+	.dp_set_lane_settings = dcn10_link_encoder_dp_set_lane_settings,
+	.dp_set_phy_pattern = dcn10_link_encoder_dp_set_phy_pattern,
+	.update_mst_stream_allocation_table =
+		dcn10_link_encoder_update_mst_stream_allocation_table,
+	.psr_program_dp_dphy_fast_training =
+			dcn10_psr_program_dp_dphy_fast_training,
+	.psr_program_secondary_packet = dcn10_psr_program_secondary_packet,
+	.connect_dig_be_to_fe = dcn10_link_encoder_connect_dig_be_to_fe,
+	.enable_hpd = dcn10_link_encoder_enable_hpd,
+	.disable_hpd = dcn10_link_encoder_disable_hpd,
+	.is_dig_enabled = dcn401_is_dig_enabled,
+	.destroy = dcn10_link_encoder_destroy,
+	.fec_set_enable = enc2_fec_set_enable,
+	.fec_set_ready = enc2_fec_set_ready,
+	.fec_is_active = enc2_fec_is_active,
+	.get_dig_frontend = dcn10_get_dig_frontend,
+	.get_dig_mode = dcn401_get_dig_mode,
+	.is_in_alt_mode = dcn32_link_encoder_is_in_alt_mode,
+	.get_max_link_cap = dcn32_link_encoder_get_max_link_cap,
+	.set_dio_phy_mux = dcn31_link_encoder_set_dio_phy_mux,
+};
+
+void dcn401_link_encoder_construct(
+	struct dcn20_link_encoder *enc20,
+	const struct encoder_init_data *init_data,
+	const struct encoder_feature_support *enc_features,
+	const struct dcn10_link_enc_registers *link_regs,
+	const struct dcn10_link_enc_aux_registers *aux_regs,
+	const struct dcn10_link_enc_hpd_registers *hpd_regs,
+	const struct dcn10_link_enc_shift *link_shift,
+	const struct dcn10_link_enc_mask *link_mask)
+{
+	struct bp_connector_speed_cap_info bp_cap_info = {0};
+	const struct dc_vbios_funcs *bp_funcs = init_data->ctx->dc_bios->funcs;
+	enum bp_result result = BP_RESULT_OK;
+	struct dcn10_link_encoder *enc10 = &enc20->enc10;
+
+	enc10->base.funcs = &dcn401_link_enc_funcs;
+	enc10->base.ctx = init_data->ctx;
+	enc10->base.id = init_data->encoder;
+
+	enc10->base.hpd_source = init_data->hpd_source;
+	enc10->base.connector = init_data->connector;
+
+
+	enc10->base.preferred_engine = ENGINE_ID_UNKNOWN;
+
+	enc10->base.features = *enc_features;
+	if (enc10->base.connector.id == CONNECTOR_ID_USBC)
+		enc10->base.features.flags.bits.DP_IS_USB_C = 1;
+
+	enc10->base.transmitter = init_data->transmitter;
+
+	/* set the flag to indicate whether driver poll the I2C data pin
+	 * while doing the DP sink detect
+	 */
+
+/*	if (dal_adapter_service_is_feature_supported(as,
+		FEATURE_DP_SINK_DETECT_POLL_DATA_PIN))
+		enc10->base.features.flags.bits.
+			DP_SINK_DETECT_POLL_DATA_PIN = true;*/
+
+	enc10->base.output_signals =
+		SIGNAL_TYPE_DVI_SINGLE_LINK |
+		SIGNAL_TYPE_DVI_DUAL_LINK |
+		SIGNAL_TYPE_LVDS |
+		SIGNAL_TYPE_DISPLAY_PORT |
+		SIGNAL_TYPE_DISPLAY_PORT_MST |
+		SIGNAL_TYPE_EDP |
+		SIGNAL_TYPE_HDMI_TYPE_A;
+
+	enc10->link_regs = link_regs;
+	enc10->aux_regs = aux_regs;
+	enc10->hpd_regs = hpd_regs;
+	enc10->link_shift = link_shift;
+	enc10->link_mask = link_mask;
+
+	switch (enc10->base.transmitter) {
+	case TRANSMITTER_UNIPHY_A:
+		enc10->base.preferred_engine = ENGINE_ID_DIGA;
+	break;
+	case TRANSMITTER_UNIPHY_B:
+		enc10->base.preferred_engine = ENGINE_ID_DIGB;
+	break;
+	case TRANSMITTER_UNIPHY_C:
+		enc10->base.preferred_engine = ENGINE_ID_DIGC;
+	break;
+	case TRANSMITTER_UNIPHY_D:
+		enc10->base.preferred_engine = ENGINE_ID_DIGD;
+	break;
+	case TRANSMITTER_UNIPHY_E:
+		enc10->base.preferred_engine = ENGINE_ID_DIGE;
+	break;
+	default:
+		ASSERT_CRITICAL(false);
+		enc10->base.preferred_engine = ENGINE_ID_UNKNOWN;
+	}
+
+	/* default to one to mirror Windows behavior */
+	enc10->base.features.flags.bits.HDMI_6GB_EN = 1;
+
+	if (bp_funcs->get_connector_speed_cap_info)
+		result = bp_funcs->get_connector_speed_cap_info(enc10->base.ctx->dc_bios,
+						enc10->base.connector, &bp_cap_info);
+
+	/* Override features with DCE-specific values */
+	if (result == BP_RESULT_OK) {
+		enc10->base.features.flags.bits.IS_HBR2_CAPABLE =
+				bp_cap_info.DP_HBR2_EN;
+		enc10->base.features.flags.bits.IS_HBR3_CAPABLE =
+				bp_cap_info.DP_HBR3_EN;
+		enc10->base.features.flags.bits.HDMI_6GB_EN = bp_cap_info.HDMI_6GB_EN;
+		enc10->base.features.flags.bits.IS_DP2_CAPABLE = 1;
+		enc10->base.features.flags.bits.IS_UHBR10_CAPABLE = bp_cap_info.DP_UHBR10_EN;
+		enc10->base.features.flags.bits.IS_UHBR13_5_CAPABLE = bp_cap_info.DP_UHBR13_5_EN;
+		enc10->base.features.flags.bits.IS_UHBR20_CAPABLE = bp_cap_info.DP_UHBR20_EN;
+	} else {
+		DC_LOG_WARNING("%s: Failed to get encoder_cap_info from VBIOS with error code %d!\n",
+				__func__,
+				result);
+	}
+	if (enc10->base.ctx->dc->debug.hdmi20_disable) {
+		enc10->base.features.flags.bits.HDMI_6GB_EN = 0;
+	}
+}
diff --git a/drivers/gpu/drm/amd/display/dc/dio/dcn401/dcn401_dio_link_encoder.h b/drivers/gpu/drm/amd/display/dc/dio/dcn401/dcn401_dio_link_encoder.h
new file mode 100644
index 000000000000..6baab8302b81
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dio/dcn401/dcn401_dio_link_encoder.h
@@ -0,0 +1,134 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ *  and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __DC_LINK_ENCODER__DCN401_H__
+#define __DC_LINK_ENCODER__DCN401_H__
+
+#include "dcn30/dcn30_dio_link_encoder.h"
+
+#define LINK_ENCODER_MASK_SH_LIST_DCN401(mask_sh) \
+	LE_SF(DIG0_DIG_BE_EN_CNTL, DIG_BE_ENABLE, mask_sh),\
+	LE_SF(DIG0_DIG_BE_CNTL, DIG_RB_SWITCH_EN, mask_sh),\
+	LE_SF(DIG0_DIG_BE_CNTL, DIG_HPD_SELECT, mask_sh),\
+	LE_SF(DIG0_DIG_BE_CNTL, DIG_FE_SOURCE_SELECT, mask_sh),\
+	LE_SF(DIG0_DIG_BE_CLK_CNTL, DIG_BE_MODE, mask_sh),\
+	LE_SF(DIG0_DIG_BE_CLK_CNTL, DIG_BE_CLK_EN, mask_sh),\
+	LE_SF(DIG0_DIG_BE_CLK_CNTL, DIG_BE_SOFT_RESET, mask_sh),\
+	LE_SF(DIG0_DIG_BE_CLK_CNTL, HDCP_SOFT_RESET, mask_sh),\
+	LE_SF(DIG0_DIG_BE_CLK_CNTL, DIG_BE_SYMCLK_G_CLOCK_ON, mask_sh),\
+	LE_SF(DIG0_DIG_BE_CLK_CNTL, DIG_BE_SYMCLK_G_HDCP_CLOCK_ON, mask_sh),\
+	LE_SF(DIG0_DIG_BE_CLK_CNTL, DIG_BE_SYMCLK_G_TMDS_CLOCK_ON, mask_sh),\
+	LE_SF(DIG0_DIG_CLOCK_PATTERN, DIG_CLOCK_PATTERN, mask_sh),\
+	LE_SF(DIG0_TMDS_CTL_BITS, TMDS_CTL0, mask_sh), \
+	LE_SF(DP0_DP_DPHY_CNTL, DPHY_BYPASS, mask_sh),\
+	LE_SF(DP0_DP_DPHY_CNTL, DPHY_ATEST_SEL_LANE0, mask_sh),\
+	LE_SF(DP0_DP_DPHY_CNTL, DPHY_ATEST_SEL_LANE1, mask_sh),\
+	LE_SF(DP0_DP_DPHY_CNTL, DPHY_ATEST_SEL_LANE2, mask_sh),\
+	LE_SF(DP0_DP_DPHY_CNTL, DPHY_ATEST_SEL_LANE3, mask_sh),\
+	LE_SF(DP0_DP_DPHY_PRBS_CNTL, DPHY_PRBS_EN, mask_sh),\
+	LE_SF(DP0_DP_DPHY_PRBS_CNTL, DPHY_PRBS_SEL, mask_sh),\
+	LE_SF(DP0_DP_DPHY_SYM0, DPHY_SYM1, mask_sh),\
+	LE_SF(DP0_DP_DPHY_SYM0, DPHY_SYM2, mask_sh),\
+	LE_SF(DP0_DP_DPHY_SYM0, DPHY_SYM3, mask_sh),\
+	LE_SF(DP0_DP_DPHY_SYM1, DPHY_SYM4, mask_sh),\
+	LE_SF(DP0_DP_DPHY_SYM1, DPHY_SYM5, mask_sh),\
+	LE_SF(DP0_DP_DPHY_SYM1, DPHY_SYM6, mask_sh),\
+	LE_SF(DP0_DP_DPHY_SYM2, DPHY_SYM7, mask_sh),\
+	LE_SF(DP0_DP_DPHY_SYM2, DPHY_SYM8, mask_sh),\
+	LE_SF(DP0_DP_DPHY_SCRAM_CNTL, DPHY_SCRAMBLER_BS_COUNT, mask_sh),\
+	LE_SF(DP0_DP_DPHY_SCRAM_CNTL, DPHY_SCRAMBLER_ADVANCE, mask_sh),\
+	LE_SF(DP0_DP_DPHY_FAST_TRAINING, DPHY_RX_FAST_TRAINING_CAPABLE, mask_sh),\
+	LE_SF(DP0_DP_DPHY_BS_SR_SWAP_CNTL, DPHY_LOAD_BS_COUNT, mask_sh),\
+	LE_SF(DP0_DP_DPHY_TRAINING_PATTERN_SEL, DPHY_TRAINING_PATTERN_SEL, mask_sh),\
+	LE_SF(DP0_DP_DPHY_HBR2_PATTERN_CONTROL, DP_DPHY_HBR2_PATTERN_CONTROL, mask_sh),\
+	LE_SF(DP0_DP_LINK_CNTL, DP_LINK_TRAINING_COMPLETE, mask_sh),\
+	LE_SF(DP0_DP_LINK_FRAMING_CNTL, DP_IDLE_BS_INTERVAL, mask_sh),\
+	LE_SF(DP0_DP_LINK_FRAMING_CNTL, DP_VBID_DISABLE, mask_sh),\
+	LE_SF(DP0_DP_LINK_FRAMING_CNTL, DP_VID_ENHANCED_FRAME_MODE, mask_sh),\
+	LE_SF(DP0_DP_VID_STREAM_CNTL, DP_VID_STREAM_ENABLE, mask_sh),\
+	LE_SF(DP0_DP_CONFIG, DP_UDI_LANES, mask_sh),\
+	LE_SF(DP0_DP_SEC_CNTL1, DP_SEC_GSP0_LINE_NUM, mask_sh),\
+	LE_SF(DP0_DP_SEC_CNTL1, DP_SEC_GSP0_PRIORITY, mask_sh),\
+	LE_SF(DP0_DP_MSE_SAT0, DP_MSE_SAT_SRC0, mask_sh),\
+	LE_SF(DP0_DP_MSE_SAT0, DP_MSE_SAT_SRC1, mask_sh),\
+	LE_SF(DP0_DP_MSE_SAT0, DP_MSE_SAT_SLOT_COUNT0, mask_sh),\
+	LE_SF(DP0_DP_MSE_SAT0, DP_MSE_SAT_SLOT_COUNT1, mask_sh),\
+	LE_SF(DP0_DP_MSE_SAT1, DP_MSE_SAT_SRC2, mask_sh),\
+	LE_SF(DP0_DP_MSE_SAT1, DP_MSE_SAT_SRC3, mask_sh),\
+	LE_SF(DP0_DP_MSE_SAT1, DP_MSE_SAT_SLOT_COUNT2, mask_sh),\
+	LE_SF(DP0_DP_MSE_SAT1, DP_MSE_SAT_SLOT_COUNT3, mask_sh),\
+	LE_SF(DP0_DP_MSE_SAT_UPDATE, DP_MSE_SAT_UPDATE, mask_sh),\
+	LE_SF(DP0_DP_MSE_SAT_UPDATE, DP_MSE_16_MTP_KEEPOUT, mask_sh),\
+	LE_SF(DP_AUX0_AUX_CONTROL, AUX_HPD_SEL, mask_sh),\
+	LE_SF(DP_AUX0_AUX_CONTROL, AUX_LS_READ_EN, mask_sh),\
+	LE_SF(DP_AUX0_AUX_DPHY_RX_CONTROL0, AUX_RX_RECEIVE_WINDOW, mask_sh),\
+	LE_SF(HPD0_DC_HPD_CONTROL, DC_HPD_EN, mask_sh),\
+	LE_SF(DP0_DP_DPHY_CNTL, DPHY_FEC_EN, mask_sh),\
+	LE_SF(DP0_DP_DPHY_CNTL, DPHY_FEC_READY_SHADOW, mask_sh),\
+	LE_SF(DP0_DP_DPHY_CNTL, DPHY_FEC_ACTIVE_STATUS, mask_sh),\
+	LE_SF(DIG0_TMDS_CTL_BITS, TMDS_CTL0, mask_sh), \
+	LE_SF(DP_AUX0_AUX_DPHY_RX_CONTROL0, AUX_RX_START_WINDOW, mask_sh),\
+	LE_SF(DP_AUX0_AUX_DPHY_RX_CONTROL0, AUX_RX_HALF_SYM_DETECT_LEN, mask_sh),\
+	LE_SF(DP_AUX0_AUX_DPHY_RX_CONTROL0, AUX_RX_TRANSITION_FILTER_EN, mask_sh),\
+	LE_SF(DP_AUX0_AUX_DPHY_RX_CONTROL0, AUX_RX_ALLOW_BELOW_THRESHOLD_PHASE_DETECT, mask_sh),\
+	LE_SF(DP_AUX0_AUX_DPHY_RX_CONTROL0, AUX_RX_ALLOW_BELOW_THRESHOLD_START, mask_sh),\
+	LE_SF(DP_AUX0_AUX_DPHY_RX_CONTROL0, AUX_RX_ALLOW_BELOW_THRESHOLD_STOP, mask_sh),\
+	LE_SF(DP_AUX0_AUX_DPHY_RX_CONTROL0, AUX_RX_PHASE_DETECT_LEN, mask_sh),\
+	LE_SF(DP_AUX0_AUX_DPHY_RX_CONTROL0, AUX_RX_DETECTION_THRESHOLD, mask_sh), \
+	LE_SF(DP_AUX0_AUX_DPHY_TX_CONTROL, AUX_TX_PRECHARGE_LEN, mask_sh),\
+	LE_SF(DP_AUX0_AUX_DPHY_TX_CONTROL, AUX_TX_PRECHARGE_SYMBOLS, mask_sh),\
+	LE_SF(DP_AUX0_AUX_DPHY_TX_CONTROL, AUX_MODE_DET_CHECK_DELAY, mask_sh),\
+	LE_SF(DP_AUX0_AUX_DPHY_RX_CONTROL1, AUX_RX_PRECHARGE_SKIP, mask_sh),\
+	LE_SF(DP_AUX0_AUX_DPHY_RX_CONTROL1, AUX_RX_TIMEOUT_LEN, mask_sh),\
+	LE_SF(DP_AUX0_AUX_DPHY_RX_CONTROL1, AUX_RX_TIMEOUT_LEN_MUL, mask_sh)
+
+void dcn401_link_encoder_construct(
+	struct dcn20_link_encoder *enc20,
+	const struct encoder_init_data *init_data,
+	const struct encoder_feature_support *enc_features,
+	const struct dcn10_link_enc_registers *link_regs,
+	const struct dcn10_link_enc_aux_registers *aux_regs,
+	const struct dcn10_link_enc_hpd_registers *hpd_regs,
+	const struct dcn10_link_enc_shift *link_shift,
+	const struct dcn10_link_enc_mask *link_mask);
+
+void enc401_hw_init(struct link_encoder *enc);
+
+void dcn401_link_encoder_enable_dp_output(
+	struct link_encoder *enc,
+	const struct dc_link_settings *link_settings,
+	enum clock_source_id clock_source);
+
+void dcn401_link_encoder_setup(
+	struct link_encoder *enc,
+	enum signal_type signal);
+
+enum signal_type dcn401_get_dig_mode(
+	struct link_encoder *enc);
+
+bool dcn401_is_dig_enabled(struct link_encoder *enc);
+
+enum signal_type dcn401_get_dig_mode(struct link_encoder *enc);
+#endif /* __DC_LINK_ENCODER__DCN401_H__ */
diff --git a/drivers/gpu/drm/amd/display/dc/dio/dcn401/dcn401_dio_stream_encoder.c b/drivers/gpu/drm/amd/display/dc/dio/dcn401/dcn401_dio_stream_encoder.c
new file mode 100644
index 000000000000..99aab70ef3e1
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dio/dcn401/dcn401_dio_stream_encoder.c
@@ -0,0 +1,856 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ *  and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+
+#include "dc_bios_types.h"
+#include "dcn30/dcn30_dio_stream_encoder.h"
+#include "dcn32/dcn32_dio_stream_encoder.h"
+#include "dcn35/dcn35_dio_stream_encoder.h"
+
+#include "dcn401_dio_stream_encoder.h"
+#include "reg_helper.h"
+#include "hw_shared.h"
+#include "link_service.h"
+#include "dpcd_defs.h"
+
+#define DC_LOGGER \
+		enc1->base.ctx->logger
+
+#define REG(reg)\
+	(enc1->regs->reg)
+
+#undef FN
+#define FN(reg_name, field_name) \
+	enc1->se_shift->field_name, enc1->se_mask->field_name
+
+#define VBI_LINE_0 0
+#define HDMI_CLOCK_CHANNEL_RATE_MORE_340M 340000
+
+#define CTX \
+	enc1->base.ctx
+
+
+
+static void enc401_dp_set_odm_combine(
+	struct stream_encoder *enc,
+	bool odm_combine)
+{
+}
+
+/* setup stream encoder in dvi mode */
+void enc401_stream_encoder_dvi_set_stream_attribute(
+	struct stream_encoder *enc,
+	struct dc_crtc_timing *crtc_timing,
+	bool is_dual_link)
+{
+	struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc);
+
+	if (!enc->ctx->dc->debug.avoid_vbios_exec_table) {
+		struct bp_encoder_control cntl = {0};
+
+		cntl.action = ENCODER_CONTROL_SETUP;
+		cntl.engine_id = enc1->base.id;
+		cntl.signal = is_dual_link ?
+			SIGNAL_TYPE_DVI_DUAL_LINK : SIGNAL_TYPE_DVI_SINGLE_LINK;
+		cntl.enable_dp_audio = false;
+		cntl.pixel_clock = crtc_timing->pix_clk_100hz / 10;
+		cntl.lanes_number = (is_dual_link) ? LANE_COUNT_EIGHT : LANE_COUNT_FOUR;
+
+		if (enc1->base.bp->funcs->encoder_control(
+				enc1->base.bp, &cntl) != BP_RESULT_OK)
+			return;
+
+	} else {
+
+		//Set pattern for clock channel, default vlue 0x63 does not work
+		REG_UPDATE(DIG_CLOCK_PATTERN, DIG_CLOCK_PATTERN, 0x1F);
+
+		//DIG_BE_TMDS_DVI_MODE : TMDS-DVI mode is already set in link_encoder_setup
+
+		//DIG_SOURCE_SELECT is already set in dig_connect_to_otg
+
+		/* DIG_START is removed from the register spec */
+	}
+
+	ASSERT(crtc_timing->pixel_encoding == PIXEL_ENCODING_RGB);
+	ASSERT(crtc_timing->display_color_depth == COLOR_DEPTH_888);
+	enc401_stream_encoder_set_stream_attribute_helper(enc1, crtc_timing);
+}
+
+/* setup stream encoder in hdmi mode */
+void enc401_stream_encoder_hdmi_set_stream_attribute(
+	struct stream_encoder *enc,
+	struct dc_crtc_timing *crtc_timing,
+	int actual_pix_clk_khz,
+	bool enable_audio)
+{
+	struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc);
+
+	if (!enc->ctx->dc->debug.avoid_vbios_exec_table) {
+		struct bp_encoder_control cntl = {0};
+
+		cntl.action = ENCODER_CONTROL_SETUP;
+		cntl.engine_id = enc1->base.id;
+		cntl.signal = SIGNAL_TYPE_HDMI_TYPE_A;
+		cntl.enable_dp_audio = enable_audio;
+		cntl.pixel_clock = actual_pix_clk_khz;
+		cntl.lanes_number = LANE_COUNT_FOUR;
+
+		if (enc1->base.bp->funcs->encoder_control(
+				enc1->base.bp, &cntl) != BP_RESULT_OK)
+			return;
+
+	} else {
+
+		//Set pattern for clock channel, default vlue 0x63 does not work
+		REG_UPDATE(DIG_CLOCK_PATTERN, DIG_CLOCK_PATTERN, 0x1F);
+
+		//DIG_BE_TMDS_HDMI_MODE : TMDS-HDMI mode is already set in link_encoder_setup
+
+		//DIG_SOURCE_SELECT is already set in dig_connect_to_otg
+
+		/* DIG_START is removed from the register spec */
+	}
+
+	/* Configure pixel encoding */
+	enc401_stream_encoder_set_stream_attribute_helper(enc1, crtc_timing);
+
+	/* setup HDMI engine */
+	REG_UPDATE_6(HDMI_CONTROL,
+		HDMI_PACKET_GEN_VERSION, 1,
+		HDMI_KEEPOUT_MODE, 1,
+		HDMI_DEEP_COLOR_ENABLE, 0,
+		HDMI_DATA_SCRAMBLE_EN, 0,
+		HDMI_NO_EXTRA_NULL_PACKET_FILLED, 1,
+		HDMI_CLOCK_CHANNEL_RATE, 0);
+
+	/* Configure color depth */
+	switch (crtc_timing->display_color_depth) {
+	case COLOR_DEPTH_888:
+		REG_UPDATE(HDMI_CONTROL, HDMI_DEEP_COLOR_DEPTH, 0);
+		break;
+	case COLOR_DEPTH_101010:
+		if (crtc_timing->pixel_encoding == PIXEL_ENCODING_YCBCR422) {
+			REG_UPDATE_2(HDMI_CONTROL,
+					HDMI_DEEP_COLOR_DEPTH, 1,
+					HDMI_DEEP_COLOR_ENABLE, 0);
+		} else {
+			REG_UPDATE_2(HDMI_CONTROL,
+					HDMI_DEEP_COLOR_DEPTH, 1,
+					HDMI_DEEP_COLOR_ENABLE, 1);
+			}
+		break;
+	case COLOR_DEPTH_121212:
+		if (crtc_timing->pixel_encoding == PIXEL_ENCODING_YCBCR422) {
+			REG_UPDATE_2(HDMI_CONTROL,
+					HDMI_DEEP_COLOR_DEPTH, 2,
+					HDMI_DEEP_COLOR_ENABLE, 0);
+		} else {
+			REG_UPDATE_2(HDMI_CONTROL,
+					HDMI_DEEP_COLOR_DEPTH, 2,
+					HDMI_DEEP_COLOR_ENABLE, 1);
+			}
+		break;
+	case COLOR_DEPTH_161616:
+		REG_UPDATE_2(HDMI_CONTROL,
+				HDMI_DEEP_COLOR_DEPTH, 3,
+				HDMI_DEEP_COLOR_ENABLE, 1);
+		break;
+	default:
+		break;
+	}
+
+	if (actual_pix_clk_khz >= HDMI_CLOCK_CHANNEL_RATE_MORE_340M) {
+		/* enable HDMI data scrambler
+		 * HDMI_CLOCK_CHANNEL_RATE_MORE_340M
+		 * Clock channel frequency is 1/4 of character rate.
+		 */
+		REG_UPDATE_2(HDMI_CONTROL,
+			HDMI_DATA_SCRAMBLE_EN, 1,
+			HDMI_CLOCK_CHANNEL_RATE, 1);
+	} else if (crtc_timing->flags.LTE_340MCSC_SCRAMBLE) {
+
+		/* TODO: New feature for DCE11, still need to implement */
+
+		/* enable HDMI data scrambler
+		 * HDMI_CLOCK_CHANNEL_FREQ_EQUAL_TO_CHAR_RATE
+		 * Clock channel frequency is the same
+		 * as character rate
+		 */
+		REG_UPDATE_2(HDMI_CONTROL,
+			HDMI_DATA_SCRAMBLE_EN, 1,
+			HDMI_CLOCK_CHANNEL_RATE, 0);
+	}
+
+
+	/* Enable transmission of General Control packet on every frame */
+	REG_UPDATE_3(HDMI_VBI_PACKET_CONTROL,
+		HDMI_GC_CONT, 1,
+		HDMI_GC_SEND, 1,
+		HDMI_NULL_SEND, 1);
+
+	/* Disable Audio Content Protection packet transmission */
+	REG_UPDATE(HDMI_VBI_PACKET_CONTROL, HDMI_ACP_SEND, 0);
+	/* following belongs to audio */
+	/* Enable Audio InfoFrame packet transmission. */
+	REG_UPDATE(HDMI_INFOFRAME_CONTROL0, HDMI_AUDIO_INFO_SEND, 1);
+
+	/* update double-buffered AUDIO_INFO registers immediately */
+	ASSERT(enc->afmt);
+	enc->afmt->funcs->audio_info_immediate_update(enc->afmt);
+
+	/* Select line number on which to send Audio InfoFrame packets */
+	REG_UPDATE(HDMI_INFOFRAME_CONTROL1, HDMI_AUDIO_INFO_LINE,
+				VBI_LINE_0 + 2);
+
+	/* set HDMI GC AVMUTE */
+	REG_UPDATE(HDMI_GC, HDMI_GC_AVMUTE, 0);
+}
+
+void enc401_set_dig_input_mode(struct stream_encoder *enc, unsigned int pix_per_container)
+{
+	struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc);
+
+	// The naming of this field is confusing, what it means is the output mode of otg, which
+	// is the input mode of the dig
+	switch (pix_per_container)	{
+	case 2:
+		REG_UPDATE(DIG_FIFO_CTRL0, DIG_FIFO_OUTPUT_PIXEL_PER_CYCLE, 0x1);
+		break;
+	case 4:
+		REG_UPDATE(DIG_FIFO_CTRL0, DIG_FIFO_OUTPUT_PIXEL_PER_CYCLE, 0x2);
+		break;
+	case 8:
+		REG_UPDATE(DIG_FIFO_CTRL0, DIG_FIFO_OUTPUT_PIXEL_PER_CYCLE, 0x3);
+		break;
+	default:
+		REG_UPDATE(DIG_FIFO_CTRL0, DIG_FIFO_OUTPUT_PIXEL_PER_CYCLE, 0x0);
+		break;
+	}
+}
+
+static bool is_two_pixels_per_containter(const struct dc_crtc_timing *timing)
+{
+	bool two_pix = timing->pixel_encoding == PIXEL_ENCODING_YCBCR420;
+
+	two_pix = two_pix || (timing->flags.DSC && timing->pixel_encoding == PIXEL_ENCODING_YCBCR422
+			&& !timing->dsc_cfg.ycbcr422_simple);
+	return two_pix;
+}
+
+void enc401_stream_encoder_dp_unblank(
+		struct dc_link *link,
+		struct stream_encoder *enc,
+		const struct encoder_unblank_param *param)
+{
+	struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc);
+
+	if (param->link_settings.link_rate != LINK_RATE_UNKNOWN) {
+		uint32_t n_vid = 0x8000;
+		uint32_t m_vid;
+		uint32_t pix_per_container = 1;
+		uint64_t m_vid_l = n_vid;
+
+		/* YCbCr 4:2:0 or YCbCr4:2:2 simple + DSC: Computed VID_M will be 2X the input rate */
+		if (is_two_pixels_per_containter(&param->timing)) {
+			pix_per_container = 2;
+		}
+
+		/* M / N = Fstream / Flink
+		 * m_vid / n_vid = pixel rate / link rate
+		 */
+		m_vid_l *= param->timing.pix_clk_100hz / pix_per_container / 10;
+		m_vid_l = div_u64(m_vid_l,
+			param->link_settings.link_rate
+				* LINK_RATE_REF_FREQ_IN_KHZ);
+
+		m_vid = (uint32_t) m_vid_l;
+
+		/* enable auto measurement */
+
+		REG_UPDATE(DP_VID_TIMING, DP_VID_M_N_GEN_EN, 0);
+
+		/* auto measurement need 1 full 0x8000 symbol cycle to kick in,
+		 * therefore program initial value for Mvid and Nvid
+		 */
+
+		REG_UPDATE(DP_VID_N, DP_VID_N, n_vid);
+
+		REG_UPDATE(DP_VID_M, DP_VID_M, m_vid);
+
+		/* reduce jitter based on read rate */
+		switch (param->pix_per_cycle)	{
+		case 2:
+			REG_UPDATE(DP_VID_TIMING, DP_VID_N_INTERVAL, 0x1);
+			break;
+		case 4:
+			REG_UPDATE(DP_VID_TIMING, DP_VID_N_INTERVAL, 0x2);
+			break;
+		case 8:
+			REG_UPDATE(DP_VID_TIMING, DP_VID_N_INTERVAL, 0x3);
+			break;
+		default:
+			REG_UPDATE(DP_VID_TIMING, DP_VID_N_INTERVAL, 0x0);
+			break;
+		}
+
+		REG_UPDATE(DP_VID_TIMING, DP_VID_M_N_GEN_EN, 1);
+	}
+
+	/* make sure stream is disabled before resetting steer fifo */
+	REG_UPDATE(DP_VID_STREAM_CNTL, DP_VID_STREAM_ENABLE, false);
+	REG_WAIT(DP_VID_STREAM_CNTL, DP_VID_STREAM_STATUS, 0, 10, 5000);
+
+	/* DIG_START is removed from the register spec */
+
+	/* switch DP encoder to CRTC data, but reset it the fifo first. It may happen
+	 * that it overflows during mode transition, and sometimes doesn't recover.
+	 */
+	REG_UPDATE(DP_STEER_FIFO, DP_STEER_FIFO_RESET, 1);
+	udelay(10);
+
+	REG_UPDATE(DP_STEER_FIFO, DP_STEER_FIFO_RESET, 0);
+
+	REG_UPDATE(DP_STEER_FIFO, DP_STEER_FIFO_ENABLE, 1);
+
+	REG_UPDATE_2(DP_VID_STREAM_CNTL, DP_VID_STREAM_ENABLE, 1, DP_VID_STREAM_DIS_DEFER, 2);
+	udelay(200);
+
+	/* DIG Resync FIFO now needs to be explicitly enabled
+	 */
+	/* read start level = 0 will bring underflow / overflow and DIG_FIFO_ERROR = 1
+	 * so set it to 1/2 full = 7 before reset as suggested by hardware team.
+	 */
+	REG_UPDATE(DIG_FIFO_CTRL0, DIG_FIFO_READ_START_LEVEL, 0x7);
+
+	REG_UPDATE(DIG_FIFO_CTRL0, DIG_FIFO_RESET, 1);
+
+	REG_WAIT(DIG_FIFO_CTRL0, DIG_FIFO_RESET_DONE, 1, 10, 5000);
+
+	REG_UPDATE(DIG_FIFO_CTRL0, DIG_FIFO_RESET, 0);
+
+	REG_WAIT(DIG_FIFO_CTRL0, DIG_FIFO_RESET_DONE, 0, 10, 5000);
+
+	REG_UPDATE(DIG_FIFO_CTRL0, DIG_FIFO_ENABLE, 1);
+
+	/* wait 100us for DIG/DP logic to prime
+	 * (i.e. a few video lines)
+	 */
+	udelay(100);
+
+	/* the hardware would start sending video at the start of the next DP
+	 * frame (i.e. rising edge of the vblank).
+	 * NOTE: We used to program DP_VID_STREAM_DIS_DEFER = 2 here, but this
+	 * register has no effect on enable transition! HW always guarantees
+	 * VID_STREAM enable at start of next frame, and this is not
+	 * programmable
+	 */
+
+	REG_UPDATE(DP_VID_STREAM_CNTL, DP_VID_STREAM_ENABLE, true);
+
+	link->dc->link_srv->dp_trace_source_sequence(link, DPCD_SOURCE_SEQ_AFTER_ENABLE_DP_VID_STREAM);
+}
+
+/* this function read dsc related register fields to be logged later in dcn10_log_hw_state
+ * into a dcn_dsc_state struct.
+ */
+void enc401_read_state(struct stream_encoder *enc, struct enc_state *s)
+{
+	struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc);
+
+	//if dsc is enabled, continue to read
+	REG_GET(DP_PIXEL_FORMAT, PIXEL_ENCODING_TYPE, &s->dsc_mode);
+
+	if (s->dsc_mode) {
+		REG_GET(DP_GSP11_CNTL, DP_SEC_GSP11_LINE_NUM, &s->sec_gsp_pps_line_num);
+
+		REG_GET(DP_MSA_VBID_MISC, DP_VBID6_LINE_REFERENCE, &s->vbid6_line_reference);
+		REG_GET(DP_MSA_VBID_MISC, DP_VBID6_LINE_NUM, &s->vbid6_line_num);
+
+		REG_GET(DP_GSP11_CNTL, DP_SEC_GSP11_ENABLE, &s->sec_gsp_pps_enable);
+		REG_GET(DP_SEC_CNTL, DP_SEC_STREAM_ENABLE, &s->sec_stream_enable);
+	}
+}
+
+void enc401_stream_encoder_enable(
+	struct stream_encoder *enc,
+	enum signal_type signal,
+	bool enable)
+{
+	struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc);
+	if (enable) {
+		switch (signal) {
+		case SIGNAL_TYPE_DVI_SINGLE_LINK:
+		case SIGNAL_TYPE_DVI_DUAL_LINK:
+			/* TMDS-DVI */
+			REG_UPDATE(DIG_FE_CLK_CNTL, DIG_FE_MODE, 2);
+			break;
+		case SIGNAL_TYPE_HDMI_TYPE_A:
+			/* TMDS-HDMI */
+			REG_UPDATE(DIG_FE_CLK_CNTL, DIG_FE_MODE, 3);
+			break;
+		case SIGNAL_TYPE_DISPLAY_PORT_MST:
+			/* DP MST */
+			REG_UPDATE(DIG_FE_CLK_CNTL, DIG_FE_MODE, 5);
+			break;
+		case SIGNAL_TYPE_EDP:
+		case SIGNAL_TYPE_DISPLAY_PORT:
+		case SIGNAL_TYPE_VIRTUAL:
+			/* DP SST */
+			REG_UPDATE(DIG_FE_CLK_CNTL, DIG_FE_MODE, 0);
+			break;
+		default:
+			/* invalid mode ! */
+			ASSERT_CRITICAL(false);
+		}
+
+		REG_UPDATE(DIG_FE_CLK_CNTL, DIG_FE_CLK_EN, 1);
+		REG_UPDATE(DIG_FE_EN_CNTL, DIG_FE_ENABLE, 1);
+	} else {
+		REG_UPDATE(DIG_FE_EN_CNTL, DIG_FE_ENABLE, 0);
+		REG_UPDATE(DIG_FE_CLK_CNTL, DIG_FE_CLK_EN, 0);
+	}
+}
+
+void enc401_stream_encoder_dp_set_stream_attribute(
+	struct stream_encoder *enc,
+	struct dc_crtc_timing *crtc_timing,
+	enum dc_color_space output_color_space,
+	bool use_vsc_sdp_for_colorimetry,
+	uint32_t enable_sdp_splitting)
+{
+	uint32_t h_active_start;
+	uint32_t v_active_start;
+	uint32_t misc0 = 0;
+	uint32_t misc1 = 0;
+	uint32_t h_blank;
+	uint32_t h_back_porch;
+	uint8_t colorimetry_bpc;
+	uint8_t dp_pixel_encoding = 0;
+	uint8_t dp_component_depth = 0;
+	uint8_t dp_translate_pixel_enc = 0;
+	// Fix set but not used warnings
+	//uint8_t dp_pixel_encoding_type = 0;
+	uint8_t dp_compressed_pixel_format = 0;
+
+	struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc);
+	struct dc_crtc_timing hw_crtc_timing = *crtc_timing;
+
+	if (hw_crtc_timing.flags.INTERLACE) {
+		/*the input timing is in VESA spec format with Interlace flag =1*/
+		hw_crtc_timing.v_total /= 2;
+		hw_crtc_timing.v_border_top /= 2;
+		hw_crtc_timing.v_addressable /= 2;
+		hw_crtc_timing.v_border_bottom /= 2;
+		hw_crtc_timing.v_front_porch /= 2;
+		hw_crtc_timing.v_sync_width /= 2;
+	}
+
+
+	/* set pixel encoding */
+	switch (hw_crtc_timing.pixel_encoding) {
+	case PIXEL_ENCODING_YCBCR422:
+		dp_pixel_encoding = DP_PIXEL_ENCODING_TYPE_YCBCR422;
+		break;
+	case PIXEL_ENCODING_YCBCR444:
+		dp_pixel_encoding = DP_PIXEL_ENCODING_TYPE_YCBCR444;
+
+		if (hw_crtc_timing.flags.Y_ONLY)
+			if (hw_crtc_timing.display_color_depth != COLOR_DEPTH_666)
+				/* HW testing only, no use case yet.
+				 * Color depth of Y-only could be
+				 * 8, 10, 12, 16 bits
+				 */
+				dp_pixel_encoding = DP_PIXEL_ENCODING_TYPE_Y_ONLY;
+
+		/* Note: DP_MSA_MISC1 bit 7 is the indicator
+		 * of Y-only mode.
+		 * This bit is set in HW if register
+		 * DP_PIXEL_ENCODING is programmed to 0x4
+		 */
+		break;
+	case PIXEL_ENCODING_YCBCR420:
+		dp_pixel_encoding = DP_PIXEL_ENCODING_TYPE_YCBCR420;
+		break;
+	default:
+		dp_pixel_encoding = DP_PIXEL_ENCODING_TYPE_RGB444;
+		break;
+	}
+
+	misc1 = REG_READ(DP_MSA_MISC);
+	/* For YCbCr420 and BT2020 Colorimetry Formats, VSC SDP shall be used.
+	 * When MISC1, bit 6, is Set to 1, a Source device uses a VSC SDP to indicate the
+	 * Pixel Encoding/Colorimetry Format and that a Sink device shall ignore MISC1, bit 7,
+	 * and MISC0, bits 7:1 (MISC1, bit 7, and MISC0, bits 7:1, become "don't care").
+	 */
+	if (use_vsc_sdp_for_colorimetry)
+		misc1 = misc1 | 0x40;
+	else
+		misc1 = misc1 & ~0x40;
+
+	/* set color depth */
+	switch (hw_crtc_timing.display_color_depth) {
+	case COLOR_DEPTH_666:
+		dp_component_depth = DP_COMPONENT_PIXEL_DEPTH_6BPC;
+		break;
+	case COLOR_DEPTH_888:
+		dp_component_depth = DP_COMPONENT_PIXEL_DEPTH_8BPC;
+		break;
+	case COLOR_DEPTH_101010:
+		dp_component_depth = DP_COMPONENT_PIXEL_DEPTH_10BPC;
+		break;
+	case COLOR_DEPTH_121212:
+		dp_component_depth = DP_COMPONENT_PIXEL_DEPTH_12BPC;
+		break;
+	case COLOR_DEPTH_161616:
+		dp_component_depth = DP_COMPONENT_PIXEL_DEPTH_16BPC;
+		break;
+	default:
+		dp_component_depth = DP_COMPONENT_PIXEL_DEPTH_6BPC;
+		break;
+	}
+
+	if (hw_crtc_timing.flags.DSC) {
+		// Fix set but not used error
+		//dp_pixel_encoding_type = 1;
+		switch (hw_crtc_timing.pixel_encoding) {
+		case PIXEL_ENCODING_YCBCR444:
+			dp_compressed_pixel_format = 0;
+			break;
+		case PIXEL_ENCODING_YCBCR422:
+			dp_compressed_pixel_format = 1;
+			if (hw_crtc_timing.dsc_cfg.ycbcr422_simple)
+				dp_compressed_pixel_format = 0;
+			break;
+		case PIXEL_ENCODING_YCBCR420:
+			dp_compressed_pixel_format = 1;
+			break;
+		default:
+			dp_compressed_pixel_format = 0;
+			break;
+		}
+	} else {
+		// Fix set but not used error
+		//dp_pixel_encoding_type = 0;
+		switch (dp_pixel_encoding) {
+		case DP_PIXEL_ENCODING_TYPE_RGB444:
+			dp_translate_pixel_enc = 0;
+			break;
+		case DP_PIXEL_ENCODING_TYPE_YCBCR422:
+			dp_translate_pixel_enc = 1;
+			break;
+		case DP_PIXEL_ENCODING_TYPE_YCBCR444:
+			dp_translate_pixel_enc = 0;
+			break;
+		case DP_PIXEL_ENCODING_TYPE_Y_ONLY:
+			dp_translate_pixel_enc = 3;
+			break;
+		case DP_PIXEL_ENCODING_TYPE_YCBCR420:
+			dp_translate_pixel_enc = 2;
+			break;
+		default:
+			ASSERT(0);
+			break;
+		}
+	}
+	/* Set DP pixel encoding and component depth */
+	REG_UPDATE_4(DP_PIXEL_FORMAT,
+			PIXEL_ENCODING_TYPE, hw_crtc_timing.flags.DSC ? 1 : 0,
+			UNCOMPRESSED_PIXEL_FORMAT, dp_translate_pixel_enc,
+			UNCOMPRESSED_COMPONENT_DEPTH, dp_component_depth,
+			COMPRESSED_PIXEL_FORMAT, dp_compressed_pixel_format);
+
+	/* set dynamic range and YCbCr range */
+
+	switch (hw_crtc_timing.display_color_depth) {
+	case COLOR_DEPTH_666:
+		colorimetry_bpc = 0;
+		break;
+	case COLOR_DEPTH_888:
+		colorimetry_bpc = 1;
+		break;
+	case COLOR_DEPTH_101010:
+		colorimetry_bpc = 2;
+		break;
+	case COLOR_DEPTH_121212:
+		colorimetry_bpc = 3;
+		break;
+	default:
+		colorimetry_bpc = 0;
+		break;
+	}
+
+	misc0 = colorimetry_bpc << 5;
+
+	switch (output_color_space) {
+	case COLOR_SPACE_SRGB:
+		misc1 = misc1 & ~0x80; /* bit7 = 0*/
+		break;
+	case COLOR_SPACE_SRGB_LIMITED:
+		misc0 = misc0 | 0x8; /* bit3=1 */
+		misc1 = misc1 & ~0x80; /* bit7 = 0*/
+		break;
+	case COLOR_SPACE_YCBCR601:
+	case COLOR_SPACE_YCBCR601_LIMITED:
+		misc0 = misc0 | 0x8; /* bit3=1, bit4=0 */
+		misc1 = misc1 & ~0x80; /* bit7 = 0*/
+		if (hw_crtc_timing.pixel_encoding == PIXEL_ENCODING_YCBCR422)
+			misc0 = misc0 | 0x2; /* bit2=0, bit1=1 */
+		else if (hw_crtc_timing.pixel_encoding == PIXEL_ENCODING_YCBCR444)
+			misc0 = misc0 | 0x4; /* bit2=1, bit1=0 */
+		break;
+	case COLOR_SPACE_YCBCR709:
+	case COLOR_SPACE_YCBCR709_LIMITED:
+		misc0 = misc0 | 0x18; /* bit3=1, bit4=1 */
+		misc1 = misc1 & ~0x80; /* bit7 = 0*/
+		if (hw_crtc_timing.pixel_encoding == PIXEL_ENCODING_YCBCR422)
+			misc0 = misc0 | 0x2; /* bit2=0, bit1=1 */
+		else if (hw_crtc_timing.pixel_encoding == PIXEL_ENCODING_YCBCR444)
+			misc0 = misc0 | 0x4; /* bit2=1, bit1=0 */
+		break;
+	case COLOR_SPACE_2020_RGB_LIMITEDRANGE:
+	case COLOR_SPACE_2020_RGB_FULLRANGE:
+	case COLOR_SPACE_2020_YCBCR_LIMITED:
+	case COLOR_SPACE_XR_RGB:
+	case COLOR_SPACE_MSREF_SCRGB:
+	case COLOR_SPACE_ADOBERGB:
+	case COLOR_SPACE_DCIP3:
+	case COLOR_SPACE_XV_YCC_709:
+	case COLOR_SPACE_XV_YCC_601:
+	case COLOR_SPACE_DISPLAYNATIVE:
+	case COLOR_SPACE_DOLBYVISION:
+	case COLOR_SPACE_APPCTRL:
+	case COLOR_SPACE_CUSTOMPOINTS:
+	case COLOR_SPACE_UNKNOWN:
+	case COLOR_SPACE_YCBCR709_BLACK:
+	default:
+		/* do nothing */
+		break;
+	}
+
+	REG_SET(DP_MSA_COLORIMETRY, 0, DP_MSA_MISC0, misc0);
+	REG_WRITE(DP_MSA_MISC, misc1);   /* MSA_MISC1 */
+
+	/* dcn new register
+	 * dc_crtc_timing is vesa dmt struct. data from edid
+	 */
+	REG_SET_2(DP_MSA_TIMING_PARAM1, 0,
+			DP_MSA_HTOTAL, hw_crtc_timing.h_total,
+			DP_MSA_VTOTAL, hw_crtc_timing.v_total);
+
+	/* calculate from vesa timing parameters
+	 * h_active_start related to leading edge of sync
+	 */
+
+	h_blank = hw_crtc_timing.h_total - hw_crtc_timing.h_border_left -
+			hw_crtc_timing.h_addressable - hw_crtc_timing.h_border_right;
+
+	h_back_porch = h_blank - hw_crtc_timing.h_front_porch -
+			hw_crtc_timing.h_sync_width;
+
+	/* start at beginning of left border */
+	h_active_start = hw_crtc_timing.h_sync_width + h_back_porch;
+
+
+	v_active_start = hw_crtc_timing.v_total - hw_crtc_timing.v_border_top -
+			hw_crtc_timing.v_addressable - hw_crtc_timing.v_border_bottom -
+			hw_crtc_timing.v_front_porch;
+
+
+	/* start at beginning of left border */
+	REG_SET_2(DP_MSA_TIMING_PARAM2, 0,
+		DP_MSA_HSTART, h_active_start,
+		DP_MSA_VSTART, v_active_start);
+
+	REG_SET_4(DP_MSA_TIMING_PARAM3, 0,
+			DP_MSA_HSYNCWIDTH,
+			hw_crtc_timing.h_sync_width,
+			DP_MSA_HSYNCPOLARITY,
+			!hw_crtc_timing.flags.HSYNC_POSITIVE_POLARITY,
+			DP_MSA_VSYNCWIDTH,
+			hw_crtc_timing.v_sync_width,
+			DP_MSA_VSYNCPOLARITY,
+			!hw_crtc_timing.flags.VSYNC_POSITIVE_POLARITY);
+
+	/* HWDITH include border or overscan */
+	REG_SET_2(DP_MSA_TIMING_PARAM4, 0,
+		DP_MSA_HWIDTH, hw_crtc_timing.h_border_left +
+		hw_crtc_timing.h_addressable + hw_crtc_timing.h_border_right,
+		DP_MSA_VHEIGHT, hw_crtc_timing.v_border_top +
+		hw_crtc_timing.v_addressable + hw_crtc_timing.v_border_bottom);
+
+	REG_UPDATE(DP_SEC_FRAMING4,
+		DP_SST_SDP_SPLITTING, enable_sdp_splitting);
+}
+
+void enc401_stream_encoder_map_to_link(
+		struct stream_encoder *enc,
+		uint32_t stream_enc_inst,
+		uint32_t link_enc_inst)
+{
+	struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc);
+
+	REG_UPDATE(STREAM_MAPPER_CONTROL,
+				DIG_STREAM_LINK_TARGET, link_enc_inst);
+}
+
+static const struct stream_encoder_funcs dcn401_str_enc_funcs = {
+	.dp_set_odm_combine =
+		enc401_dp_set_odm_combine,
+	.dp_set_stream_attribute =
+		enc401_stream_encoder_dp_set_stream_attribute,
+	.hdmi_set_stream_attribute =
+		enc401_stream_encoder_hdmi_set_stream_attribute,
+	.dvi_set_stream_attribute =
+		enc401_stream_encoder_dvi_set_stream_attribute,
+	.set_throttled_vcp_size =
+		enc1_stream_encoder_set_throttled_vcp_size,
+	.update_hdmi_info_packets =
+		enc3_stream_encoder_update_hdmi_info_packets,
+	.stop_hdmi_info_packets =
+		enc3_stream_encoder_stop_hdmi_info_packets,
+	.update_dp_info_packets_sdp_line_num =
+		enc3_stream_encoder_update_dp_info_packets_sdp_line_num,
+	.update_dp_info_packets =
+		enc3_stream_encoder_update_dp_info_packets,
+	.stop_dp_info_packets =
+		enc1_stream_encoder_stop_dp_info_packets,
+	.dp_blank =
+		enc1_stream_encoder_dp_blank,
+	.dp_unblank =
+		enc401_stream_encoder_dp_unblank,
+	.audio_mute_control = enc3_audio_mute_control,
+
+	.dp_audio_setup = enc3_se_dp_audio_setup,
+	.dp_audio_enable = enc3_se_dp_audio_enable,
+	.dp_audio_disable = enc1_se_dp_audio_disable,
+
+	.hdmi_audio_setup = enc3_se_hdmi_audio_setup,
+	.hdmi_audio_disable = enc1_se_hdmi_audio_disable,
+	.setup_stereo_sync  = enc1_setup_stereo_sync,
+	.set_avmute = enc1_stream_encoder_set_avmute,
+	.dig_connect_to_otg = enc1_dig_connect_to_otg,
+	.dig_source_otg = enc1_dig_source_otg,
+
+	.dp_get_pixel_format  = enc1_stream_encoder_dp_get_pixel_format,
+
+	.enc_read_state = enc401_read_state,
+	.dp_set_dsc_config = NULL,
+	.dp_set_dsc_pps_info_packet = enc3_dp_set_dsc_pps_info_packet,
+	.set_dynamic_metadata = enc401_set_dynamic_metadata,
+	.hdmi_reset_stream_attribute = enc1_reset_hdmi_stream_attribute,
+	.enable_stream = enc401_stream_encoder_enable,
+
+	.set_input_mode = enc401_set_dig_input_mode,
+	.enable_fifo = enc35_enable_fifo,
+	.disable_fifo = enc35_disable_fifo,
+	.map_stream_to_link = enc401_stream_encoder_map_to_link,
+};
+
+void dcn401_dio_stream_encoder_construct(
+	struct dcn10_stream_encoder *enc1,
+	struct dc_context *ctx,
+	struct dc_bios *bp,
+	enum engine_id eng_id,
+	struct vpg *vpg,
+	struct afmt *afmt,
+	const struct dcn10_stream_enc_registers *regs,
+	const struct dcn10_stream_encoder_shift *se_shift,
+	const struct dcn10_stream_encoder_mask *se_mask)
+{
+	enc1->base.funcs = &dcn401_str_enc_funcs;
+	enc1->base.ctx = ctx;
+	enc1->base.id = eng_id;
+	enc1->base.bp = bp;
+	enc1->base.vpg = vpg;
+	enc1->base.afmt = afmt;
+	enc1->regs = regs;
+	enc1->se_shift = se_shift;
+	enc1->se_mask = se_mask;
+	enc1->base.stream_enc_inst = vpg->inst;
+}
+
+void enc401_set_dynamic_metadata(struct stream_encoder *enc,
+		bool enable_dme,
+		uint32_t hubp_requestor_id,
+		enum dynamic_metadata_mode dmdata_mode)
+{
+	struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc);
+
+	if (enable_dme) {
+		REG_UPDATE_2(DME_CONTROL,
+				METADATA_HUBP_REQUESTOR_ID, hubp_requestor_id,
+				METADATA_STREAM_TYPE, (dmdata_mode == dmdata_dolby_vision) ? 1 : 0);
+
+		/* Use default line reference DP_SOF for bringup.
+		 * Should use OTG_SOF for DRR cases
+		 */
+		if (dmdata_mode == dmdata_dp)
+			REG_UPDATE_3(DP_SEC_METADATA_TRANSMISSION,
+					DP_SEC_METADATA_PACKET_ENABLE, 1,
+					DP_SEC_METADATA_PACKET_LINE_REFERENCE, 0,
+					DP_SEC_METADATA_PACKET_LINE, 20);
+		else {
+			REG_UPDATE_3(HDMI_METADATA_PACKET_CONTROL,
+					HDMI_METADATA_PACKET_ENABLE, 1,
+					HDMI_METADATA_PACKET_LINE_REFERENCE, 0,
+					HDMI_METADATA_PACKET_LINE, 2);
+
+			if (dmdata_mode == dmdata_dolby_vision)
+				REG_UPDATE(HDMI_CONTROL,
+						DOLBY_VISION_EN, 1);
+		}
+
+		REG_UPDATE(DME_CONTROL,
+				METADATA_ENGINE_EN, 1);
+	} else {
+		REG_UPDATE(DME_CONTROL,
+				METADATA_ENGINE_EN, 0);
+
+		if (dmdata_mode == dmdata_dp)
+			REG_UPDATE(DP_SEC_METADATA_TRANSMISSION,
+					DP_SEC_METADATA_PACKET_ENABLE, 0);
+		else {
+			REG_UPDATE(HDMI_METADATA_PACKET_CONTROL,
+					HDMI_METADATA_PACKET_ENABLE, 0);
+			REG_UPDATE(HDMI_CONTROL,
+					DOLBY_VISION_EN, 0);
+		}
+	}
+}
+void enc401_stream_encoder_set_stream_attribute_helper(
+		struct dcn10_stream_encoder *enc1,
+		struct dc_crtc_timing *crtc_timing)
+{
+	switch (crtc_timing->pixel_encoding) {
+	case PIXEL_ENCODING_YCBCR422:
+		REG_UPDATE(HDMI_CONTROL, TMDS_PIXEL_ENCODING, 1);
+		break;
+	default:
+		REG_UPDATE(HDMI_CONTROL, TMDS_PIXEL_ENCODING, 0);
+		break;
+	}
+	REG_UPDATE(HDMI_CONTROL, TMDS_COLOR_FORMAT, 0);
+}
diff --git a/drivers/gpu/drm/amd/display/dc/dio/dcn401/dcn401_dio_stream_encoder.h b/drivers/gpu/drm/amd/display/dc/dio/dcn401/dcn401_dio_stream_encoder.h
new file mode 100644
index 000000000000..d6b00cd246b1
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dio/dcn401/dcn401_dio_stream_encoder.h
@@ -0,0 +1,240 @@
+/*
+ * Copyright 2021 - Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ *  and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __DC_DIO_STREAM_ENCODER_DCN401_H__
+#define __DC_DIO_STREAM_ENCODER_DCN401_H__
+
+#include "dcn30/dcn30_vpg.h"
+#include "dcn30/dcn30_afmt.h"
+#include "stream_encoder.h"
+#include "dcn20/dcn20_stream_encoder.h"
+
+#define SE_COMMON_MASK_SH_LIST_DCN401(mask_sh)\
+	SE_SF(DP0_DP_PIXEL_FORMAT, PIXEL_ENCODING_TYPE, mask_sh),\
+	SE_SF(DP0_DP_PIXEL_FORMAT, UNCOMPRESSED_PIXEL_FORMAT, mask_sh),\
+	SE_SF(DP0_DP_PIXEL_FORMAT, UNCOMPRESSED_COMPONENT_DEPTH, mask_sh),\
+	SE_SF(DP0_DP_PIXEL_FORMAT, COMPRESSED_PIXEL_FORMAT, mask_sh),\
+	SE_SF(DIG0_HDMI_CONTROL, HDMI_PACKET_GEN_VERSION, mask_sh),\
+	SE_SF(DIG0_HDMI_CONTROL, HDMI_KEEPOUT_MODE, mask_sh),\
+	SE_SF(DIG0_HDMI_CONTROL, HDMI_DEEP_COLOR_ENABLE, mask_sh),\
+	SE_SF(DIG0_HDMI_CONTROL, HDMI_DEEP_COLOR_DEPTH, mask_sh),\
+	SE_SF(DIG0_HDMI_CONTROL, HDMI_DATA_SCRAMBLE_EN, mask_sh),\
+	SE_SF(DIG0_HDMI_CONTROL, HDMI_NO_EXTRA_NULL_PACKET_FILLED, mask_sh),\
+	SE_SF(DIG0_HDMI_VBI_PACKET_CONTROL, HDMI_GC_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_VBI_PACKET_CONTROL, HDMI_GC_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_VBI_PACKET_CONTROL, HDMI_NULL_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_VBI_PACKET_CONTROL, HDMI_ACP_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_INFOFRAME_CONTROL0, HDMI_AUDIO_INFO_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_INFOFRAME_CONTROL1, HDMI_AUDIO_INFO_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GC, HDMI_GC_AVMUTE, mask_sh),\
+	SE_SF(DP0_DP_MSE_RATE_CNTL, DP_MSE_RATE_X, mask_sh),\
+	SE_SF(DP0_DP_MSE_RATE_CNTL, DP_MSE_RATE_Y, mask_sh),\
+	SE_SF(DP0_DP_MSE_RATE_UPDATE, DP_MSE_RATE_UPDATE_PENDING, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_GSP0_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_STREAM_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_GSP1_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_GSP2_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_GSP3_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_MPG_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL1, DP_SEC_GSP5_LINE_REFERENCE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL2, DP_SEC_GSP4_SEND, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL2, DP_SEC_GSP4_SEND_PENDING, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL4, DP_SEC_GSP4_LINE_NUM, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL5, DP_SEC_GSP5_LINE_NUM, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL2, DP_SEC_GSP4_SEND_ANY_LINE, mask_sh),\
+	SE_SF(DP0_DP_VID_STREAM_CNTL, DP_VID_STREAM_DIS_DEFER, mask_sh),\
+	SE_SF(DP0_DP_VID_STREAM_CNTL, DP_VID_STREAM_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_VID_STREAM_CNTL, DP_VID_STREAM_STATUS, mask_sh),\
+	SE_SF(DP0_DP_STEER_FIFO, DP_STEER_FIFO_RESET, mask_sh),\
+	SE_SF(DP0_DP_STEER_FIFO, DP_STEER_FIFO_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_VID_TIMING, DP_VID_M_N_GEN_EN, mask_sh),\
+	SE_SF(DP0_DP_VID_N, DP_VID_N, mask_sh),\
+	SE_SF(DP0_DP_VID_M, DP_VID_M, mask_sh),\
+	SE_SF(DIG0_HDMI_AUDIO_PACKET_CONTROL, HDMI_AUDIO_DELAY_EN, mask_sh),\
+	SE_SF(DIG0_HDMI_ACR_PACKET_CONTROL, HDMI_ACR_AUTO_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_ACR_PACKET_CONTROL, HDMI_ACR_SOURCE, mask_sh),\
+	SE_SF(DIG0_HDMI_ACR_PACKET_CONTROL, HDMI_ACR_AUDIO_PRIORITY, mask_sh),\
+	SE_SF(DIG0_HDMI_ACR_32_0, HDMI_ACR_CTS_32, mask_sh),\
+	SE_SF(DIG0_HDMI_ACR_32_1, HDMI_ACR_N_32, mask_sh),\
+	SE_SF(DIG0_HDMI_ACR_44_0, HDMI_ACR_CTS_44, mask_sh),\
+	SE_SF(DIG0_HDMI_ACR_44_1, HDMI_ACR_N_44, mask_sh),\
+	SE_SF(DIG0_HDMI_ACR_48_0, HDMI_ACR_CTS_48, mask_sh),\
+	SE_SF(DIG0_HDMI_ACR_48_1, HDMI_ACR_N_48, mask_sh),\
+	SE_SF(DP0_DP_SEC_AUD_N, DP_SEC_AUD_N, mask_sh),\
+	SE_SF(DP0_DP_SEC_TIMESTAMP, DP_SEC_TIMESTAMP_MODE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_ASP_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_ATP_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_AIP_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_ACM_ENABLE, mask_sh),\
+	SE_SF(DIG0_AFMT_CNTL, AFMT_AUDIO_CLOCK_EN, mask_sh),\
+	SE_SF(DIG0_HDMI_CONTROL, HDMI_CLOCK_CHANNEL_RATE, mask_sh),\
+	SE_SF(DIG1_HDMI_CONTROL, TMDS_PIXEL_ENCODING, mask_sh),\
+	SE_SF(DIG1_HDMI_CONTROL, TMDS_COLOR_FORMAT, mask_sh),\
+	SE_SF(DIG0_DIG_FE_CNTL, DIG_STEREOSYNC_SELECT, mask_sh),\
+	SE_SF(DIG0_DIG_FE_CNTL, DIG_STEREOSYNC_GATE_EN, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_GSP4_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_GSP5_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_GSP6_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL, DP_SEC_GSP7_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL2, DP_SEC_GSP7_SEND, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL6, DP_SEC_GSP7_LINE_NUM, mask_sh),\
+	SE_SF(DP0_DP_SEC_CNTL2, DP_SEC_GSP11_PPS, mask_sh),\
+	SE_SF(DP0_DP_GSP11_CNTL, DP_SEC_GSP11_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_GSP11_CNTL, DP_SEC_GSP11_LINE_NUM, mask_sh),\
+	SE_SF(DP0_DP_DB_CNTL, DP_DB_DISABLE, mask_sh),\
+	SE_SF(DP0_DP_MSA_COLORIMETRY, DP_MSA_MISC0, mask_sh),\
+	SE_SF(DP0_DP_MSA_TIMING_PARAM1, DP_MSA_HTOTAL, mask_sh),\
+	SE_SF(DP0_DP_MSA_TIMING_PARAM1, DP_MSA_VTOTAL, mask_sh),\
+	SE_SF(DP0_DP_MSA_TIMING_PARAM2, DP_MSA_HSTART, mask_sh),\
+	SE_SF(DP0_DP_MSA_TIMING_PARAM2, DP_MSA_VSTART, mask_sh),\
+	SE_SF(DP0_DP_MSA_TIMING_PARAM3, DP_MSA_HSYNCWIDTH, mask_sh),\
+	SE_SF(DP0_DP_MSA_TIMING_PARAM3, DP_MSA_HSYNCPOLARITY, mask_sh),\
+	SE_SF(DP0_DP_MSA_TIMING_PARAM3, DP_MSA_VSYNCWIDTH, mask_sh),\
+	SE_SF(DP0_DP_MSA_TIMING_PARAM3, DP_MSA_VSYNCPOLARITY, mask_sh),\
+	SE_SF(DP0_DP_MSA_TIMING_PARAM4, DP_MSA_HWIDTH, mask_sh),\
+	SE_SF(DP0_DP_MSA_TIMING_PARAM4, DP_MSA_VHEIGHT, mask_sh),\
+	SE_SF(DIG0_HDMI_DB_CONTROL, HDMI_DB_DISABLE, mask_sh),\
+	SE_SF(DP0_DP_VID_TIMING, DP_VID_N_INTERVAL, mask_sh),\
+	SE_SF(DIG0_DIG_FE_CNTL, DIG_SOURCE_SELECT, mask_sh), \
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC0_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC0_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC1_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC1_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC2_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC2_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC3_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC3_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC4_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC4_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC5_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC5_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC6_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC6_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC7_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL0, HDMI_GENERIC7_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC8_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC8_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC9_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC9_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC10_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC10_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC11_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC11_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC12_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC12_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC13_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC13_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC14_CONT, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL6, HDMI_GENERIC14_SEND, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL1, HDMI_GENERIC0_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL1, HDMI_GENERIC1_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL2, HDMI_GENERIC2_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL2, HDMI_GENERIC3_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL3, HDMI_GENERIC4_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL3, HDMI_GENERIC5_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL4, HDMI_GENERIC6_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL4, HDMI_GENERIC7_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL7, HDMI_GENERIC8_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL7, HDMI_GENERIC9_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL8, HDMI_GENERIC10_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL8, HDMI_GENERIC11_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL9, HDMI_GENERIC12_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL9, HDMI_GENERIC13_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_GENERIC_PACKET_CONTROL10, HDMI_GENERIC14_LINE, mask_sh),\
+	SE_SF(DP0_DP_MSA_VBID_MISC, DP_VBID6_LINE_REFERENCE, mask_sh),\
+	SE_SF(DP0_DP_MSA_VBID_MISC, DP_VBID6_LINE_NUM, mask_sh),\
+	SE_SF(DME0_DME_CONTROL, METADATA_ENGINE_EN, mask_sh),\
+	SE_SF(DME0_DME_CONTROL, METADATA_HUBP_REQUESTOR_ID, mask_sh),\
+	SE_SF(DME0_DME_CONTROL, METADATA_STREAM_TYPE, mask_sh),\
+	SE_SF(DP0_DP_SEC_METADATA_TRANSMISSION, DP_SEC_METADATA_PACKET_ENABLE, mask_sh),\
+	SE_SF(DP0_DP_SEC_METADATA_TRANSMISSION, DP_SEC_METADATA_PACKET_LINE_REFERENCE, mask_sh),\
+	SE_SF(DP0_DP_SEC_METADATA_TRANSMISSION, DP_SEC_METADATA_PACKET_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_METADATA_PACKET_CONTROL, HDMI_METADATA_PACKET_ENABLE, mask_sh),\
+	SE_SF(DIG0_HDMI_METADATA_PACKET_CONTROL, HDMI_METADATA_PACKET_LINE_REFERENCE, mask_sh),\
+	SE_SF(DIG0_HDMI_METADATA_PACKET_CONTROL, HDMI_METADATA_PACKET_LINE, mask_sh),\
+	SE_SF(DIG0_HDMI_CONTROL, DOLBY_VISION_EN, mask_sh),\
+	SE_SF(DIG0_DIG_FE_EN_CNTL, DIG_FE_ENABLE, mask_sh),\
+	SE_SF(DIG0_DIG_FE_CLK_CNTL, DIG_FE_MODE, mask_sh),\
+	SE_SF(DIG0_DIG_FE_CLK_CNTL, DIG_FE_CLK_EN, mask_sh),\
+	SE_SF(DIG0_DIG_FE_CLK_CNTL, DIG_FE_SOFT_RESET, mask_sh),\
+	SE_SF(DIG0_DIG_FE_CNTL, DIG_STEREOSYNC_GATE_EN, mask_sh),\
+	SE_SF(DP0_DP_SEC_FRAMING4, DP_SST_SDP_SPLITTING, mask_sh),\
+	SE_SF(DIG0_DIG_CLOCK_PATTERN, DIG_CLOCK_PATTERN, mask_sh),\
+	SE_SF(DIG0_DIG_FIFO_CTRL0, DIG_FIFO_OUTPUT_PIXEL_PER_CYCLE, mask_sh),\
+	SE_SF(DIG0_DIG_FIFO_CTRL0, DIG_FIFO_READ_START_LEVEL, mask_sh),\
+	SE_SF(DIG0_DIG_FIFO_CTRL0, DIG_FIFO_ENABLE, mask_sh),\
+	SE_SF(DIG0_DIG_FIFO_CTRL0, DIG_FIFO_RESET, mask_sh),\
+	SE_SF(DIG0_DIG_FIFO_CTRL0, DIG_FIFO_RESET_DONE, mask_sh),\
+	SE_SF(DIG0_STREAM_MAPPER_CONTROL, DIG_STREAM_LINK_TARGET, mask_sh),
+
+
+void dcn401_dio_stream_encoder_construct(
+	struct dcn10_stream_encoder *enc1,
+	struct dc_context *ctx,
+	struct dc_bios *bp,
+	enum engine_id eng_id,
+	struct vpg *vpg,
+	struct afmt *afmt,
+	const struct dcn10_stream_enc_registers *regs,
+	const struct dcn10_stream_encoder_shift *se_shift,
+	const struct dcn10_stream_encoder_mask *se_mask);
+
+void enc401_set_dynamic_metadata(struct stream_encoder *enc,
+		bool enable_dme,
+		uint32_t hubp_requestor_id,
+		enum dynamic_metadata_mode dmdata_mode);
+void enc401_stream_encoder_set_stream_attribute_helper(
+		struct dcn10_stream_encoder *enc1,
+		struct dc_crtc_timing *crtc_timing);
+void enc401_stream_encoder_dp_set_stream_attribute(
+	struct stream_encoder *enc,
+	struct dc_crtc_timing *crtc_timing,
+	enum dc_color_space output_color_space,
+	bool use_vsc_sdp_for_colorimetry,
+	uint32_t enable_sdp_splitting);
+void enc401_stream_encoder_dvi_set_stream_attribute(
+	struct stream_encoder *enc,
+	struct dc_crtc_timing *crtc_timing,
+	bool is_dual_link);
+void enc401_stream_encoder_dp_unblank(
+		struct dc_link *link,
+		struct stream_encoder *enc,
+		const struct encoder_unblank_param *param);
+void enc401_stream_encoder_enable(
+	struct stream_encoder *enc,
+	enum signal_type signal,
+	bool enable);
+void enc401_set_dig_input_mode(struct stream_encoder *enc, unsigned int pix_per_container);
+void enc401_stream_encoder_map_to_link(
+		struct stream_encoder *enc,
+		uint32_t stream_enc_inst,
+		uint32_t link_enc_inst);
+void enc401_read_state(struct stream_encoder *enc, struct enc_state *s);
+void enc401_stream_encoder_hdmi_set_stream_attribute(
+	struct stream_encoder *enc,
+	struct dc_crtc_timing *crtc_timing,
+	int actual_pix_clk_khz,
+	bool enable_audio);
+#endif /* __DC_DIO_STREAM_ENCODER_DCN401_H__ */
diff --git a/drivers/gpu/drm/amd/display/dc/dm_cp_psp.h b/drivers/gpu/drm/amd/display/dc/dm_cp_psp.h
index 511f9e1159c7..f4d3f04ec857 100644
--- a/drivers/gpu/drm/amd/display/dc/dm_cp_psp.h
+++ b/drivers/gpu/drm/amd/display/dc/dm_cp_psp.h
@@ -26,6 +26,9 @@
 #ifndef DM_CP_PSP_IF__H
 #define DM_CP_PSP_IF__H
 
+/*
+ * Interface to CPLIB/PSP to enable ASSR
+ */
 struct dc_link;
 
 struct cp_psp_stream_config {
@@ -34,12 +37,12 @@ struct cp_psp_stream_config {
 	uint8_t dig_fe;
 	uint8_t link_enc_idx;
 	uint8_t stream_enc_idx;
-	uint8_t phy_idx;
 	uint8_t dio_output_idx;
-	uint8_t dio_output_type;
+	uint8_t phy_idx;
 	uint8_t assr_enabled;
 	uint8_t mst_enabled;
 	uint8_t dp2_enabled;
+	uint8_t usb4_enabled;
 	void *dm_stream_ctx;
 	bool dpms_off;
 };
diff --git a/drivers/gpu/drm/amd/display/dc/dm_helpers.h b/drivers/gpu/drm/amd/display/dc/dm_helpers.h
index 0fe66b080a03..9d160b39e8c5 100644
--- a/drivers/gpu/drm/amd/display/dc/dm_helpers.h
+++ b/drivers/gpu/drm/amd/display/dc/dm_helpers.h
@@ -33,9 +33,10 @@
 #include "dc_types.h"
 #include "dc.h"
 
-struct dp_mst_stream_allocation_table;
+struct dc_dp_mst_stream_allocation_table;
 struct aux_payload;
 enum aux_return_code_type;
+enum set_config_status;
 
 /*
  * Allocate memory accessible by the GPU
@@ -59,7 +60,7 @@ void dm_helpers_free_gpu_mem(
 		void *pvMem);
 
 enum dc_edid_status dm_helpers_parse_edid_caps(
-	struct dc_context *ctx,
+	struct dc_link *link,
 	const struct dc_edid *edid,
 	struct dc_edid_caps *edid_caps);
 
@@ -77,7 +78,7 @@ void dm_helpers_dp_update_branch_info(
 bool dm_helpers_dp_mst_write_payload_allocation_table(
 		struct dc_context *ctx,
 		const struct dc_stream_state *stream,
-		struct dp_mst_stream_allocation_table *proposed_table,
+		struct dc_dp_mst_stream_allocation_table *proposed_table,
 		bool enable);
 
 /*
@@ -103,19 +104,30 @@ enum act_return_status dm_helpers_dp_mst_poll_for_allocation_change_trigger(
 /*
  * Sends ALLOCATE_PAYLOAD message.
  */
-bool dm_helpers_dp_mst_send_payload_allocation(
+void dm_helpers_dp_mst_send_payload_allocation(
 		struct dc_context *ctx,
-		const struct dc_stream_state *stream,
-		bool enable);
+		const struct dc_stream_state *stream);
+
+/*
+ * Update mst manager relevant variables
+ */
+void dm_helpers_dp_mst_update_mst_mgr_for_deallocation(
+		struct dc_context *ctx,
+		const struct dc_stream_state *stream);
 
 bool dm_helpers_dp_mst_start_top_mgr(
 		struct dc_context *ctx,
 		const struct dc_link *link,
 		bool boot);
 
-void dm_helpers_dp_mst_stop_top_mgr(
+bool dm_helpers_dp_mst_stop_top_mgr(
+		struct dc_context *ctx,
+		struct dc_link *link);
+
+void dm_helpers_dp_mst_update_branch_bandwidth(
 		struct dc_context *ctx,
 		struct dc_link *link);
+
 /**
  * OS specific aux read callback.
  */
@@ -141,11 +153,24 @@ bool dm_helpers_submit_i2c(
 		const struct dc_link *link,
 		struct i2c_command *cmd);
 
+bool dm_helpers_execute_fused_io(
+		struct dc_context *ctx,
+		struct dc_link *link,
+		union dmub_rb_cmd *commands,
+		uint8_t count,
+		uint32_t timeout_us
+);
+
 bool dm_helpers_dp_write_dsc_enable(
 		struct dc_context *ctx,
 		const struct dc_stream_state *stream,
 		bool enable
 );
+
+bool dm_helpers_dp_write_hblank_reduction(
+		struct dc_context *ctx,
+		const struct dc_stream_state *stream);
+
 bool dm_helpers_is_dp_sink_present(
 		struct dc_link *link);
 
@@ -156,13 +181,17 @@ enum dc_edid_status dm_helpers_read_local_edid(
 		struct dc_link *link,
 		struct dc_sink *sink);
 
+bool dm_helpers_dp_handle_test_pattern_request(
+		struct dc_context *ctx,
+		const struct dc_link *link,
+		union link_test_pattern dpcd_test_pattern,
+		union test_misc dpcd_test_params);
+
 void dm_set_dcn_clocks(
 		struct dc_context *ctx,
 		struct dc_clocks *clks);
 
-#if defined(CONFIG_DRM_AMD_DC_DCN)
 void dm_helpers_enable_periodic_detection(struct dc_context *ctx, bool enable);
-#endif
 
 void dm_set_phyd32clk(struct dc_context *ctx, int freq_khz);
 
@@ -170,18 +199,31 @@ bool dm_helpers_dmub_outbox_interrupt_control(struct dc_context *ctx, bool enabl
 
 void dm_helpers_smu_timeout(struct dc_context *ctx, unsigned int msg_id, unsigned int param, unsigned int timeout_us);
 
-// 0x1 = Result_OK, 0xFE = Result_UnkmownCmd
+// 0x1 = Result_OK, 0xFE = Result_UnkmownCmd, 0x0 = Status_Busy
 #define IS_SMU_TIMEOUT(result) \
-	(!(result == 0x1 || result == 0xFE))
-
+	(result == 0x0)
+void dm_helpers_init_panel_settings(
+	struct dc_context *ctx,
+	struct dc_panel_config *config,
+	struct dc_sink *sink);
+void dm_helpers_override_panel_settings(
+	struct dc_context *ctx,
+	struct dc_panel_config *config);
 int dm_helper_dmub_aux_transfer_sync(
 		struct dc_context *ctx,
 		const struct dc_link *link,
 		struct aux_payload *payload,
 		enum aux_return_code_type *operation_result);
-enum set_config_status;
+
 int dm_helpers_dmub_set_config_sync(struct dc_context *ctx,
 		const struct dc_link *link,
 		struct set_config_cmd_payload *payload,
 		enum set_config_status *operation_result);
+enum adaptive_sync_type dm_get_adaptive_sync_support_type(struct dc_link *link);
+
+enum dc_edid_status dm_helpers_get_sbios_edid(struct dc_link *link, struct dc_edid *edid);
+
+bool dm_helpers_is_fullscreen(struct dc_context *ctx, struct dc_stream_state *stream);
+bool dm_helpers_is_hdr_on(struct dc_context *ctx, struct dc_stream_state *stream);
+
 #endif /* __DM_HELPERS__ */
diff --git a/drivers/gpu/drm/amd/display/dc/dm_pp_smu.h b/drivers/gpu/drm/amd/display/dc/dm_pp_smu.h
index 4440d08743aa..b0e17a594ec3 100644
--- a/drivers/gpu/drm/amd/display/dc/dm_pp_smu.h
+++ b/drivers/gpu/drm/amd/display/dc/dm_pp_smu.h
@@ -40,8 +40,9 @@ enum pp_smu_ver {
 	PP_SMU_UNSUPPORTED,
 	PP_SMU_VER_RV,
 	PP_SMU_VER_NV,
-	PP_SMU_VER_RN,
 
+	PP_SMU_VER_RN,
+	PP_SMU_VER_VG,
 	PP_SMU_VER_MAX
 };
 
@@ -247,6 +248,7 @@ struct pp_smu_funcs_nv {
 #define PP_SMU_NUM_MEMCLK_DPM_LEVELS  4
 #define PP_SMU_NUM_DCLK_DPM_LEVELS    8
 #define PP_SMU_NUM_VCLK_DPM_LEVELS    8
+#define PP_SMU_NUM_VPECLK_DPM_LEVELS  8
 
 struct dpm_clock {
   uint32_t  Freq;    // In MHz
@@ -262,6 +264,7 @@ struct dpm_clocks {
 	struct dpm_clock MemClocks[PP_SMU_NUM_MEMCLK_DPM_LEVELS];
 	struct dpm_clock VClocks[PP_SMU_NUM_VCLK_DPM_LEVELS];
 	struct dpm_clock DClocks[PP_SMU_NUM_DCLK_DPM_LEVELS];
+	struct dpm_clock VPEClocks[PP_SMU_NUM_VPECLK_DPM_LEVELS];
 };
 
 
diff --git a/drivers/gpu/drm/amd/display/dc/dm_services.h b/drivers/gpu/drm/amd/display/dc/dm_services.h
index 9a3f2a44f882..fbbf9c757b3c 100644
--- a/drivers/gpu/drm/amd/display/dc/dm_services.h
+++ b/drivers/gpu/drm/amd/display/dc/dm_services.h
@@ -40,6 +40,7 @@
 
 struct dmub_srv;
 struct dc_dmub_srv;
+union dmub_rb_cmd;
 
 irq_handler_idx dm_register_interrupt(
 	struct dc_context *ctx,
@@ -142,7 +143,7 @@ void generic_reg_wait(const struct dc_context *ctx,
 	unsigned int delay_between_poll_us, unsigned int time_out_num_tries,
 	const char *func_name, int line);
 
-unsigned int snprintf_count(char *pBuf, unsigned int bufSize, char *fmt, ...);
+unsigned int snprintf_count(char *pBuf, unsigned int bufSize, const char *fmt, ...);
 
 /* These macros need to be used with soc15 registers in order to retrieve
  * the actual offset.
@@ -274,6 +275,30 @@ void dm_perf_trace_timestamp(const char *func_name, unsigned int line, struct dc
 #define PERF_TRACE_CTX(__CTX)	dm_perf_trace_timestamp(__func__, __LINE__, __CTX)
 
 /*
+ * SMU message tracing
+ */
+void dm_trace_smu_enter(uint32_t msg_id, uint32_t param_in, unsigned int delay, struct dc_context *ctx);
+void dm_trace_smu_exit(bool success, uint32_t response, struct dc_context *ctx);
+
+#define TRACE_SMU_MSG_DELAY(msg_id, param_in, delay, ctx)	dm_trace_smu_enter(msg_id, param_in, delay, ctx)
+#define TRACE_SMU_MSG(msg_id, param_in, ctx)	dm_trace_smu_enter(msg_id, param_in, 0, ctx)
+#define TRACE_SMU_MSG_ENTER(msg_id, param_in, ctx)	dm_trace_smu_enter(msg_id, param_in, 0, ctx)
+#define TRACE_SMU_MSG_EXIT(success, response, ctx)	dm_trace_smu_exit(success, response, ctx)
+
+/*
+ * DMUB Interfaces
+ */
+bool dm_execute_dmub_cmd(const struct dc_context *ctx, union dmub_rb_cmd *cmd, enum dm_dmub_wait_type wait_type);
+bool dm_execute_dmub_cmd_list(const struct dc_context *ctx, unsigned int count, union dmub_rb_cmd *cmd, enum dm_dmub_wait_type wait_type);
+
+/*
+ * ACPI Interfaces
+ */
+void dm_acpi_process_phy_transition_interlock(
+	const struct dc_context *ctx,
+	struct dm_process_phy_transition_init_params process_phy_transition_init_params);
+
+/*
  * Debug and verification hooks
  */
 
@@ -285,4 +310,8 @@ void dm_dtn_log_append_v(struct dc_context *ctx,
 void dm_dtn_log_end(struct dc_context *ctx,
 	struct dc_log_buffer_ctx *log_ctx);
 
+char *dce_version_to_string(const int version);
+
+bool dc_supports_vrr(const enum dce_version v);
+
 #endif /* __DM_SERVICES_H__ */
diff --git a/drivers/gpu/drm/amd/display/dc/dm_services_types.h b/drivers/gpu/drm/amd/display/dc/dm_services_types.h
index b52ba6ffabe1..3b093b8699ab 100644
--- a/drivers/gpu/drm/amd/display/dc/dm_services_types.h
+++ b/drivers/gpu/drm/amd/display/dc/dm_services_types.h
@@ -127,7 +127,7 @@ struct dm_pp_single_disp_config {
 	uint32_t src_height;
 	uint32_t src_width;
 	uint32_t v_refresh;
-	uint32_t sym_clock; /* HDMI only */
+	uint32_t pixel_clock; /* Pixel clock in KHz (for HDMI only: normalized) */
 	struct dc_link_settings link_settings; /* DP only */
 };
 
@@ -269,4 +269,36 @@ struct dtn_min_clk_info {
 	uint32_t min_memory_clock_khz;
 };
 
+enum dm_dmub_wait_type {
+	DM_DMUB_WAIT_TYPE_NO_WAIT,
+	DM_DMUB_WAIT_TYPE_WAIT,
+	DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY,
+};
+
+enum dm_acpi_transition_link_type {
+	hdmi_tmds,
+	hdmi_frl,
+	dp_8b_10b,
+	dp_128b_132b,
+	none
+};
+
+struct dm_process_phy_transition_init_params {
+	uint32_t phy_id;
+	uint8_t action;
+	uint32_t sym_clock_10khz;
+	enum signal_type signal;
+	enum dc_lane_count display_port_lanes_count;
+	enum dc_link_rate display_port_link_rate;
+	uint32_t transition_bitmask;
+	uint8_t hdmi_frl_num_lanes;
+};
+
+struct dm_process_phy_transition_input_params {
+	uint32_t phy_id;
+	uint32_t transition_id;
+	uint32_t phy_configuration;
+	uint32_t data_rate;
+};
+
 #endif
diff --git a/drivers/gpu/drm/amd/display/dc/dml/Makefile b/drivers/gpu/drm/amd/display/dc/dml/Makefile
index eee6672bd32d..b357683b4255 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/Makefile
+++ b/drivers/gpu/drm/amd/display/dc/dml/Makefile
@@ -25,56 +25,55 @@
 # It provides the general basic services required by other DAL
 # subcomponents.
 
-ifdef CONFIG_X86
-dml_ccflags := -mhard-float -msse
-endif
-
-ifdef CONFIG_PPC64
-dml_ccflags := -mhard-float -maltivec
-endif
-
-ifdef CONFIG_CC_IS_GCC
-ifeq ($(call cc-ifversion, -lt, 0701, y), y)
-IS_OLD_GCC = 1
-endif
-endif
-
-ifdef CONFIG_X86
-ifdef IS_OLD_GCC
-# Stack alignment mismatch, proceed with caution.
-# GCC < 7.1 cannot compile code using `double` and -mpreferred-stack-boundary=3
-# (8B stack alignment).
-dml_ccflags += -mpreferred-stack-boundary=4
-else
-dml_ccflags += -msse2
-endif
-endif
+dml_ccflags := $(CC_FLAGS_FPU)
+dml_rcflags := $(CC_FLAGS_NO_FPU)
 
 ifneq ($(CONFIG_FRAME_WARN),0)
-frame_warn_flag := -Wframe-larger-than=2048
+    ifeq ($(filter y,$(CONFIG_KASAN)$(CONFIG_KCSAN)),y)
+        frame_warn_limit := 3072
+    else
+        frame_warn_limit := 2048
+    endif
+
+    ifeq ($(call test-lt, $(CONFIG_FRAME_WARN), $(frame_warn_limit)),y)
+        frame_warn_flag := -Wframe-larger-than=$(frame_warn_limit)
+    endif
 endif
 
 CFLAGS_$(AMDDALPATH)/dc/dml/display_mode_lib.o := $(dml_ccflags)
-
-ifdef CONFIG_DRM_AMD_DC_DCN
 CFLAGS_$(AMDDALPATH)/dc/dml/display_mode_vba.o := $(dml_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml/dcn10/dcn10_fpu.o := $(dml_ccflags)
 CFLAGS_$(AMDDALPATH)/dc/dml/dcn20/dcn20_fpu.o := $(dml_ccflags)
-CFLAGS_$(AMDDALPATH)/dc/dml/dcn20/display_mode_vba_20.o := $(dml_ccflags)
-CFLAGS_$(AMDDALPATH)/dc/dml/dcn20/display_mode_vba_20.o := $(dml_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml/dcn20/display_mode_vba_20.o := $(dml_ccflags) $(frame_warn_flag)
 CFLAGS_$(AMDDALPATH)/dc/dml/dcn20/display_rq_dlg_calc_20.o := $(dml_ccflags)
-CFLAGS_$(AMDDALPATH)/dc/dml/dcn20/display_mode_vba_20v2.o := $(dml_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml/dcn20/display_mode_vba_20v2.o := $(dml_ccflags) $(frame_warn_flag)
 CFLAGS_$(AMDDALPATH)/dc/dml/dcn20/display_rq_dlg_calc_20v2.o := $(dml_ccflags)
-CFLAGS_$(AMDDALPATH)/dc/dml/dcn21/display_mode_vba_21.o := $(dml_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml/dcn21/display_mode_vba_21.o := $(dml_ccflags) $(frame_warn_flag)
 CFLAGS_$(AMDDALPATH)/dc/dml/dcn21/display_rq_dlg_calc_21.o := $(dml_ccflags)
 CFLAGS_$(AMDDALPATH)/dc/dml/dcn30/display_mode_vba_30.o := $(dml_ccflags) $(frame_warn_flag)
 CFLAGS_$(AMDDALPATH)/dc/dml/dcn30/display_rq_dlg_calc_30.o := $(dml_ccflags)
 CFLAGS_$(AMDDALPATH)/dc/dml/dcn31/display_mode_vba_31.o := $(dml_ccflags) $(frame_warn_flag)
 CFLAGS_$(AMDDALPATH)/dc/dml/dcn31/display_rq_dlg_calc_31.o := $(dml_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml/dcn314/display_mode_vba_314.o := $(dml_ccflags) $(frame_warn_flag)
+CFLAGS_$(AMDDALPATH)/dc/dml/dcn314/display_rq_dlg_calc_314.o := $(dml_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml/dcn314/dcn314_fpu.o := $(dml_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml/dcn30/dcn30_fpu.o := $(dml_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml/dcn32/dcn32_fpu.o := $(dml_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml/dcn32/display_mode_vba_32.o := $(dml_ccflags) $(frame_warn_flag)
+CFLAGS_$(AMDDALPATH)/dc/dml/dcn32/display_rq_dlg_calc_32.o := $(dml_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml/dcn32/display_mode_vba_util_32.o := $(dml_ccflags) $(frame_warn_flag)
+CFLAGS_$(AMDDALPATH)/dc/dml/dcn321/dcn321_fpu.o := $(dml_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml/dcn35/dcn35_fpu.o := $(dml_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml/dcn351/dcn351_fpu.o := $(dml_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml/dcn31/dcn31_fpu.o := $(dml_ccflags)
 CFLAGS_$(AMDDALPATH)/dc/dml/dcn301/dcn301_fpu.o := $(dml_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml/dcn302/dcn302_fpu.o := $(dml_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml/dcn303/dcn303_fpu.o := $(dml_ccflags)
 CFLAGS_$(AMDDALPATH)/dc/dml/dsc/rc_calc_fpu.o := $(dml_ccflags)
-CFLAGS_$(AMDDALPATH)/dc/dml/display_mode_lib.o := $(dml_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml/calcs/dcn_calcs.o := $(dml_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml/calcs/dcn_calc_auto.o := $(dml_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml/calcs/dcn_calc_math.o := $(dml_ccflags) -Wno-tautological-compare
 CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml/display_mode_vba.o := $(dml_rcflags)
-CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml/dcn2x/dcn2x.o := $(dml_rcflags)
 CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml/dcn20/display_mode_vba_20.o := $(dml_rcflags)
 CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml/dcn20/display_rq_dlg_calc_20.o := $(dml_rcflags)
 CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml/dcn20/display_mode_vba_20v2.o := $(dml_rcflags)
@@ -85,26 +84,58 @@ CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml/dcn30/display_mode_vba_30.o := $(dml_rcflags)
 CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml/dcn30/display_rq_dlg_calc_30.o := $(dml_rcflags)
 CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml/dcn31/display_mode_vba_31.o := $(dml_rcflags)
 CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml/dcn31/display_rq_dlg_calc_31.o := $(dml_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml/dcn32/display_mode_vba_32.o := $(dml_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml/dcn32/display_rq_dlg_calc_32.o := $(dml_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml/dcn32/display_mode_vba_util_32.o := $(dml_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml/dcn314/display_mode_vba_314.o := $(dml_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml/dcn314/display_rq_dlg_calc_314.o := $(dml_rcflags)
 CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml/dcn301/dcn301_fpu.o := $(dml_rcflags)
 CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml/display_mode_lib.o := $(dml_rcflags)
 CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml/dsc/rc_calc_fpu.o  := $(dml_rcflags)
-endif
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml/dcn10/dcn10_fpu.o := $(dml_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml/dcn20/dcn20_fpu.o := $(dml_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml/dcn314/display_mode_vba_314.o := $(dml_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml/dcn314/display_rq_dlg_calc_314.o := $(dml_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml/dcn314/dcn314_fpu.o := $(dml_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml/dcn30/dcn30_fpu.o := $(dml_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml/dcn32/dcn32_fpu.o := $(dml_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml/dcn321/dcn321_fpu.o := $(dml_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml/dcn35/dcn35_fpu.o := $(dml_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml/dcn351/dcn351_fpu.o := $(dml_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml/dcn31/dcn31_fpu.o := $(dml_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml/dcn302/dcn302_fpu.o := $(dml_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml/dcn303/dcn303_fpu.o := $(dml_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml/calcs/dcn_calc_math.o := $(dml_rcflags)
 CFLAGS_$(AMDDALPATH)/dc/dml/dml1_display_rq_dlg_calc.o := $(dml_ccflags)
 CFLAGS_$(AMDDALPATH)/dc/dml/display_rq_dlg_helpers.o := $(dml_ccflags)
 CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml/dml1_display_rq_dlg_calc.o := $(dml_rcflags)
 CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml/display_rq_dlg_helpers.o := $(dml_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml/calcs/dcn_calcs.o := $(dml_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml/calcs/dcn_calc_auto.o := $(dml_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml/calcs/dcn_calc_math.o := $(dml_rcflags)
 
-DML = display_mode_lib.o display_rq_dlg_helpers.o dml1_display_rq_dlg_calc.o \
-
-ifdef CONFIG_DRM_AMD_DC_DCN
+ifdef CONFIG_DRM_AMD_DC_FP
+DML += display_mode_lib.o display_rq_dlg_helpers.o dml1_display_rq_dlg_calc.o
+DML += dcn10/dcn10_fpu.o
 DML += dcn20/dcn20_fpu.o
 DML += display_mode_vba.o dcn20/display_rq_dlg_calc_20.o dcn20/display_mode_vba_20.o
 DML += dcn20/display_rq_dlg_calc_20v2.o dcn20/display_mode_vba_20v2.o
 DML += dcn21/display_rq_dlg_calc_21.o dcn21/display_mode_vba_21.o
-DML += dcn30/display_mode_vba_30.o dcn30/display_rq_dlg_calc_30.o
+DML += dcn30/dcn30_fpu.o dcn30/display_mode_vba_30.o dcn30/display_rq_dlg_calc_30.o
 DML += dcn31/display_mode_vba_31.o dcn31/display_rq_dlg_calc_31.o
+DML += dcn314/display_mode_vba_314.o dcn314/display_rq_dlg_calc_314.o
+DML += dcn32/display_mode_vba_32.o dcn32/display_rq_dlg_calc_32.o dcn32/display_mode_vba_util_32.o
+DML += dcn31/dcn31_fpu.o
+DML += dcn32/dcn32_fpu.o
+DML += dcn321/dcn321_fpu.o
 DML += dcn301/dcn301_fpu.o
+DML += dcn302/dcn302_fpu.o
+DML += dcn303/dcn303_fpu.o
+DML += dcn314/dcn314_fpu.o
+DML += dcn35/dcn35_fpu.o
+DML += dcn351/dcn351_fpu.o
 DML += dsc/rc_calc_fpu.o
+DML += calcs/dcn_calcs.o calcs/dcn_calc_math.o calcs/dcn_calc_auto.o
 endif
 
 AMD_DAL_DML = $(addprefix $(AMDDALPATH)/dc/dml/,$(DML))
diff --git a/drivers/gpu/drm/amd/display/dc/calcs/dcn_calc_auto.c b/drivers/gpu/drm/amd/display/dc/dml/calcs/dcn_calc_auto.c
index 41284e263325..288d22a16cf2 100644
--- a/drivers/gpu/drm/amd/display/dc/calcs/dcn_calc_auto.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/calcs/dcn_calc_auto.c
@@ -526,10 +526,10 @@ void mode_support_and_system_configuration(struct dcn_bw_internal_vars *v)
 				}
 				if (v->max_swath_height_c[k] > 0.0) {
 					v->swath_width_granularity_c = 256.0 /dcn_bw_ceil2(v->byte_per_pixel_in_detc[k], 2.0) / v->max_swath_height_c[k];
-				}
-				v->rounded_up_max_swath_size_bytes_c = (dcn_bw_ceil2(v->swath_width_yper_state[i][j][k] / 2.0 - 1.0, v->swath_width_granularity_c) + v->swath_width_granularity_c) * v->byte_per_pixel_in_detc[k] * v->max_swath_height_c[k];
-				if (v->source_pixel_format[k] == dcn_bw_yuv420_sub_10) {
-					v->rounded_up_max_swath_size_bytes_c =dcn_bw_ceil2(v->rounded_up_max_swath_size_bytes_c, 256.0) + 256;
+					v->rounded_up_max_swath_size_bytes_c = (dcn_bw_ceil2(v->swath_width_yper_state[i][j][k] / 2.0 - 1.0, v->swath_width_granularity_c) + v->swath_width_granularity_c) * v->byte_per_pixel_in_detc[k] * v->max_swath_height_c[k];
+					if (v->source_pixel_format[k] == dcn_bw_yuv420_sub_10) {
+						v->rounded_up_max_swath_size_bytes_c = dcn_bw_ceil2(v->rounded_up_max_swath_size_bytes_c, 256.0) + 256;
+					}
 				}
 				if (v->rounded_up_max_swath_size_bytes_y + v->rounded_up_max_swath_size_bytes_c <= v->det_buffer_size_in_kbyte * 1024.0 / 2.0) {
 					v->swath_height_yper_state[i][j][k] = v->max_swath_height_y[k];
@@ -552,14 +552,14 @@ void mode_support_and_system_configuration(struct dcn_bw_internal_vars *v)
 					v->lines_in_det_chroma = v->det_buffer_size_in_kbyte * 1024.0 / 3.0 / v->byte_per_pixel_in_dety[k] / (v->swath_width_yper_state[i][j][k] / 2.0);
 				}
 				v->effective_lb_latency_hiding_source_lines_luma =dcn_bw_min2(v->max_line_buffer_lines,dcn_bw_floor2(v->line_buffer_size / v->lb_bit_per_pixel[k] / (v->swath_width_yper_state[i][j][k] /dcn_bw_max2(v->h_ratio[k], 1.0)), 1.0)) - (v->vtaps[k] - 1.0);
-				v->effective_lb_latency_hiding_source_lines_chroma =dcn_bw_min2(v->max_line_buffer_lines,dcn_bw_floor2(v->line_buffer_size / v->lb_bit_per_pixel[k] / (v->swath_width_yper_state[i][j][k] / 2.0 /dcn_bw_max2(v->h_ratio[k] / 2.0, 1.0)), 1.0)) - (v->vta_pschroma[k] - 1.0);
 				v->effective_detlb_lines_luma =dcn_bw_floor2(v->lines_in_det_luma +dcn_bw_min2(v->lines_in_det_luma * v->required_dispclk[i][j] * v->byte_per_pixel_in_dety[k] * v->pscl_factor[k] / v->return_bw_per_state[i], v->effective_lb_latency_hiding_source_lines_luma), v->swath_height_yper_state[i][j][k]);
-				v->effective_detlb_lines_chroma =dcn_bw_floor2(v->lines_in_det_chroma +dcn_bw_min2(v->lines_in_det_chroma * v->required_dispclk[i][j] * v->byte_per_pixel_in_detc[k] * v->pscl_factor_chroma[k] / v->return_bw_per_state[i], v->effective_lb_latency_hiding_source_lines_chroma), v->swath_height_cper_state[i][j][k]);
 				if (v->byte_per_pixel_in_detc[k] == 0.0) {
 					v->urgent_latency_support_us_per_state[i][j][k] = v->effective_detlb_lines_luma * (v->htotal[k] / v->pixel_clock[k]) / v->v_ratio[k] - v->effective_detlb_lines_luma * v->swath_width_yper_state[i][j][k] *dcn_bw_ceil2(v->byte_per_pixel_in_dety[k], 1.0) / (v->return_bw_per_state[i] / v->no_of_dpp[i][j][k]);
 				}
 				else {
-					v->urgent_latency_support_us_per_state[i][j][k] =dcn_bw_min2(v->effective_detlb_lines_luma * (v->htotal[k] / v->pixel_clock[k]) / v->v_ratio[k] - v->effective_detlb_lines_luma * v->swath_width_yper_state[i][j][k] *dcn_bw_ceil2(v->byte_per_pixel_in_dety[k], 1.0) / (v->return_bw_per_state[i] / v->no_of_dpp[i][j][k]), v->effective_detlb_lines_chroma * (v->htotal[k] / v->pixel_clock[k]) / (v->v_ratio[k] / 2.0) - v->effective_detlb_lines_chroma * v->swath_width_yper_state[i][j][k] / 2.0 *dcn_bw_ceil2(v->byte_per_pixel_in_detc[k], 2.0) / (v->return_bw_per_state[i] / v->no_of_dpp[i][j][k]));
+					v->effective_lb_latency_hiding_source_lines_chroma = dcn_bw_min2(v->max_line_buffer_lines, dcn_bw_floor2(v->line_buffer_size / v->lb_bit_per_pixel[k] / (v->swath_width_yper_state[i][j][k] / 2.0 / dcn_bw_max2(v->h_ratio[k] / 2.0, 1.0)), 1.0)) - (v->vta_pschroma[k] - 1.0);
+					v->effective_detlb_lines_chroma = dcn_bw_floor2(v->lines_in_det_chroma + dcn_bw_min2(v->lines_in_det_chroma * v->required_dispclk[i][j] * v->byte_per_pixel_in_detc[k] * v->pscl_factor_chroma[k] / v->return_bw_per_state[i], v->effective_lb_latency_hiding_source_lines_chroma), v->swath_height_cper_state[i][j][k]);
+					v->urgent_latency_support_us_per_state[i][j][k] = dcn_bw_min2(v->effective_detlb_lines_luma * (v->htotal[k] / v->pixel_clock[k]) / v->v_ratio[k] - v->effective_detlb_lines_luma * v->swath_width_yper_state[i][j][k] * dcn_bw_ceil2(v->byte_per_pixel_in_dety[k], 1.0) / (v->return_bw_per_state[i] / v->no_of_dpp[i][j][k]), v->effective_detlb_lines_chroma * (v->htotal[k] / v->pixel_clock[k]) / (v->v_ratio[k] / 2.0) - v->effective_detlb_lines_chroma * v->swath_width_yper_state[i][j][k] / 2.0 * dcn_bw_ceil2(v->byte_per_pixel_in_detc[k], 2.0) / (v->return_bw_per_state[i] / v->no_of_dpp[i][j][k]));
 				}
 			}
 		}
@@ -1146,10 +1146,10 @@ void display_pipe_configuration(struct dcn_bw_internal_vars *v)
 		}
 		if (v->maximum_swath_height_c > 0.0) {
 			v->swath_width_granularity_c = 256.0 /dcn_bw_ceil2(v->byte_per_pix_detc, 2.0) / v->maximum_swath_height_c;
-		}
-		v->rounded_up_max_swath_size_bytes_c = (dcn_bw_ceil2(v->swath_width / 2.0 - 1.0, v->swath_width_granularity_c) + v->swath_width_granularity_c) * v->byte_per_pix_detc * v->maximum_swath_height_c;
-		if (v->source_pixel_format[k] == dcn_bw_yuv420_sub_10) {
-			v->rounded_up_max_swath_size_bytes_c =dcn_bw_ceil2(v->rounded_up_max_swath_size_bytes_c, 256.0) + 256;
+			v->rounded_up_max_swath_size_bytes_c = (dcn_bw_ceil2(v->swath_width / 2.0 - 1.0, v->swath_width_granularity_c) + v->swath_width_granularity_c) * v->byte_per_pix_detc * v->maximum_swath_height_c;
+			if (v->source_pixel_format[k] == dcn_bw_yuv420_sub_10) {
+				v->rounded_up_max_swath_size_bytes_c = dcn_bw_ceil2(v->rounded_up_max_swath_size_bytes_c, 256.0) + 256;
+			}
 		}
 		if (v->rounded_up_max_swath_size_bytes_y + v->rounded_up_max_swath_size_bytes_c <= v->det_buffer_size_in_kbyte * 1024.0 / 2.0) {
 			v->swath_height_y[k] = v->maximum_swath_height_y;
diff --git a/drivers/gpu/drm/amd/display/dc/calcs/dcn_calc_auto.h b/drivers/gpu/drm/amd/display/dc/dml/calcs/dcn_calc_auto.h
index ce35de79a6c7..ce35de79a6c7 100644
--- a/drivers/gpu/drm/amd/display/dc/calcs/dcn_calc_auto.h
+++ b/drivers/gpu/drm/amd/display/dc/dml/calcs/dcn_calc_auto.h
diff --git a/drivers/gpu/drm/amd/display/dc/calcs/dcn_calc_math.c b/drivers/gpu/drm/amd/display/dc/dml/calcs/dcn_calc_math.c
index 07d18e78de49..cac72413a097 100644
--- a/drivers/gpu/drm/amd/display/dc/calcs/dcn_calc_math.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/calcs/dcn_calc_math.c
@@ -23,6 +23,7 @@
  *
  */
 
+#include "os_types.h"
 #include "dcn_calc_math.h"
 
 #define isNaN(number) ((number) != (number))
@@ -69,8 +70,8 @@ float dcn_bw_max2(const float arg1, const float arg2)
 
 float dcn_bw_floor2(const float arg, const float significance)
 {
-	if (significance == 0)
-		return 0;
+	ASSERT(significance != 0);
+
 	return ((int) (arg / significance)) * significance;
 }
 float dcn_bw_floor(const float arg)
@@ -80,17 +81,14 @@ float dcn_bw_floor(const float arg)
 
 float dcn_bw_ceil(const float arg)
 {
-	float flr = dcn_bw_floor2(arg, 1);
-
-	return flr + 0.00001 >= arg ? arg : flr + 1;
+	return (int) (arg + 0.99999);
 }
 
 float dcn_bw_ceil2(const float arg, const float significance)
 {
-	float flr = dcn_bw_floor2(arg, significance);
-	if (significance == 0)
-		return 0;
-	return flr + 0.00001 >= arg ? arg : flr + significance;
+	ASSERT(significance != 0);
+
+	return ((int) (arg / significance + 0.99999)) * significance;
 }
 
 float dcn_bw_max3(float v1, float v2, float v3)
diff --git a/drivers/gpu/drm/amd/display/dc/calcs/dcn_calcs.c b/drivers/gpu/drm/amd/display/dc/dml/calcs/dcn_calcs.c
index 6b248cd2a461..74962791302f 100644
--- a/drivers/gpu/drm/amd/display/dc/calcs/dcn_calcs.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/calcs/dcn_calcs.c
@@ -30,7 +30,7 @@
 #include "dcn_calc_auto.h"
 #include "dal_asic_id.h"
 #include "resource.h"
-#include "dcn10/dcn10_resource.h"
+#include "resource/dcn10/dcn10_resource.h"
 #include "dcn10/dcn10_hubbub.h"
 #include "dml/dml1_display_rq_dlg_calc.h"
 
@@ -503,7 +503,6 @@ static void dcn_bw_calc_rq_dlg_ttu(
 	//input[in_idx].dout.output_standard;
 
 	/*todo: soc->sr_enter_plus_exit_time??*/
-	dlg_sys_param->t_srx_delay_us = dc->dcn_ip->dcfclk_cstate_latency / v->dcf_clk_deep_sleep;
 
 	dml1_rq_dlg_get_rq_params(dml, rq_param, &input->pipe.src);
 	dml1_extract_rq_regs(dml, rq_regs, rq_param);
@@ -640,7 +639,6 @@ static bool dcn_bw_apply_registry_override(struct dc *dc)
 {
 	bool updated = false;
 
-	DC_FP_START();
 	if ((int)(dc->dcn_soc->sr_exit_time * 1000) != dc->debug.sr_exit_time_ns
 			&& dc->debug.sr_exit_time_ns) {
 		updated = true;
@@ -676,7 +674,6 @@ static bool dcn_bw_apply_registry_override(struct dc *dc)
 		dc->dcn_soc->dram_clock_change_latency =
 				dc->debug.dram_clock_change_latency_ns / 1000.0;
 	}
-	DC_FP_END();
 
 	return updated;
 }
@@ -693,7 +690,7 @@ static void hack_disable_optional_pipe_split(struct dcn_bw_internal_vars *v)
 static void hack_force_pipe_split(struct dcn_bw_internal_vars *v,
 		unsigned int pixel_rate_100hz)
 {
-	float pixel_rate_mhz = pixel_rate_100hz / 10000;
+	float pixel_rate_mhz = pixel_rate_100hz / 10000.0;
 
 	/*
 	 * force enabling pipe split by lower dpp clock for DPM0 to just
@@ -739,34 +736,19 @@ static void hack_bounding_box(struct dcn_bw_internal_vars *v,
 		hack_force_pipe_split(v, context->streams[0]->timing.pix_clk_100hz);
 }
 
-unsigned int get_highest_allowed_voltage_level(uint32_t chip_family, uint32_t hw_internal_rev, uint32_t pci_revision_id)
+static unsigned int get_highest_allowed_voltage_level(bool is_vmin_only_asic)
 {
 	/* for low power RV2 variants, the highest voltage level we want is 0 */
-	if ((chip_family == FAMILY_RV) &&
-	     ASICREV_IS_RAVEN2(hw_internal_rev))
-		switch (pci_revision_id) {
-		case PRID_DALI_DE:
-		case PRID_DALI_DF:
-		case PRID_DALI_E3:
-		case PRID_DALI_E4:
-		case PRID_POLLOCK_94:
-		case PRID_POLLOCK_95:
-		case PRID_POLLOCK_E9:
-		case PRID_POLLOCK_EA:
-		case PRID_POLLOCK_EB:
-			return 0;
-		default:
-			break;
-		}
-
-	/* we are ok with all levels */
-	return 4;
+	if (is_vmin_only_asic)
+		return 0;
+	else	/* we are ok with all levels */
+		return 4;
 }
 
 bool dcn_validate_bandwidth(
 		struct dc *dc,
 		struct dc_state *context,
-		bool fast_validate)
+		enum dc_validate_mode validate_mode)
 {
 	/*
 	 * we want a breakdown of the various stages of validation, which the
@@ -789,7 +771,6 @@ bool dcn_validate_bandwidth(
 		dcn_bw_sync_calcs_and_dml(dc);
 
 	memset(v, 0, sizeof(*v));
-	DC_FP_START();
 
 	v->sr_exit_time = dc->dcn_soc->sr_exit_time;
 	v->sr_enter_plus_exit_time = dc->dcn_soc->sr_enter_plus_exit_time;
@@ -1138,7 +1119,7 @@ bool dcn_validate_bandwidth(
 
 	BW_VAL_TRACE_END_VOLTAGE_LEVEL();
 
-	if (v->voltage_level != number_of_states_plus_one && !fast_validate) {
+	if (v->voltage_level != number_of_states_plus_one && validate_mode == DC_VALIDATE_MODE_AND_PROGRAMMING) {
 		float bw_consumed = v->total_bandwidth_consumed_gbyte_per_second;
 
 		if (bw_consumed < v->fabric_and_dram_bandwidth_vmin0p65)
@@ -1277,7 +1258,7 @@ bool dcn_validate_bandwidth(
 						hsplit_pipe->pipe_dlg_param.vblank_end = pipe->pipe_dlg_param.vblank_end;
 					} else {
 						/* pipe not split previously needs split */
-						hsplit_pipe = find_idle_secondary_pipe(&context->res_ctx, pool, pipe);
+						hsplit_pipe = resource_find_free_secondary_pipe_legacy(&context->res_ctx, pool, pipe);
 						ASSERT(hsplit_pipe);
 						split_stream_across_pipes(&context->res_ctx, pool, pipe, hsplit_pipe);
 					}
@@ -1305,7 +1286,7 @@ bool dcn_validate_bandwidth(
 		}
 	} else if (v->voltage_level == number_of_states_plus_one) {
 		BW_VAL_TRACE_SKIP(fail);
-	} else if (fast_validate) {
+	} else if (validate_mode != DC_VALIDATE_MODE_AND_PROGRAMMING) {
 		BW_VAL_TRACE_SKIP(fast);
 	}
 
@@ -1322,235 +1303,75 @@ bool dcn_validate_bandwidth(
 	bw_limit = dc->dcn_soc->percent_disp_bw_limit * v->fabric_and_dram_bandwidth_vmax0p9;
 	bw_limit_pass = (v->total_data_read_bandwidth / 1000.0) < bw_limit;
 
-	DC_FP_END();
-
 	PERFORMANCE_TRACE_END();
 	BW_VAL_TRACE_FINISH();
 
-	if (bw_limit_pass && v->voltage_level <= get_highest_allowed_voltage_level(
-							dc->ctx->asic_id.chip_family,
-							dc->ctx->asic_id.hw_internal_rev,
-							dc->ctx->asic_id.pci_revision_id))
+	if (bw_limit_pass && v->voltage_level <= get_highest_allowed_voltage_level(dc->config.is_vmin_only_asic))
 		return true;
 	else
 		return false;
 }
 
-static unsigned int dcn_find_normalized_clock_vdd_Level(
-	const struct dc *dc,
-	enum dm_pp_clock_type clocks_type,
-	int clocks_in_khz)
-{
-	int vdd_level = dcn_bw_v_min0p65;
-
-	if (clocks_in_khz == 0)/*todo some clock not in the considerations*/
-		return vdd_level;
-
-	switch (clocks_type) {
-	case DM_PP_CLOCK_TYPE_DISPLAY_CLK:
-		if (clocks_in_khz > dc->dcn_soc->max_dispclk_vmax0p9*1000) {
-			vdd_level = dcn_bw_v_max0p91;
-			BREAK_TO_DEBUGGER();
-		} else if (clocks_in_khz > dc->dcn_soc->max_dispclk_vnom0p8*1000) {
-			vdd_level = dcn_bw_v_max0p9;
-		} else if (clocks_in_khz > dc->dcn_soc->max_dispclk_vmid0p72*1000) {
-			vdd_level = dcn_bw_v_nom0p8;
-		} else if (clocks_in_khz > dc->dcn_soc->max_dispclk_vmin0p65*1000) {
-			vdd_level = dcn_bw_v_mid0p72;
-		} else
-			vdd_level = dcn_bw_v_min0p65;
-		break;
-	case DM_PP_CLOCK_TYPE_DISPLAYPHYCLK:
-		if (clocks_in_khz > dc->dcn_soc->phyclkv_max0p9*1000) {
-			vdd_level = dcn_bw_v_max0p91;
-			BREAK_TO_DEBUGGER();
-		} else if (clocks_in_khz > dc->dcn_soc->phyclkv_nom0p8*1000) {
-			vdd_level = dcn_bw_v_max0p9;
-		} else if (clocks_in_khz > dc->dcn_soc->phyclkv_mid0p72*1000) {
-			vdd_level = dcn_bw_v_nom0p8;
-		} else if (clocks_in_khz > dc->dcn_soc->phyclkv_min0p65*1000) {
-			vdd_level = dcn_bw_v_mid0p72;
-		} else
-			vdd_level = dcn_bw_v_min0p65;
-		break;
-
-	case DM_PP_CLOCK_TYPE_DPPCLK:
-		if (clocks_in_khz > dc->dcn_soc->max_dppclk_vmax0p9*1000) {
-			vdd_level = dcn_bw_v_max0p91;
-			BREAK_TO_DEBUGGER();
-		} else if (clocks_in_khz > dc->dcn_soc->max_dppclk_vnom0p8*1000) {
-			vdd_level = dcn_bw_v_max0p9;
-		} else if (clocks_in_khz > dc->dcn_soc->max_dppclk_vmid0p72*1000) {
-			vdd_level = dcn_bw_v_nom0p8;
-		} else if (clocks_in_khz > dc->dcn_soc->max_dppclk_vmin0p65*1000) {
-			vdd_level = dcn_bw_v_mid0p72;
-		} else
-			vdd_level = dcn_bw_v_min0p65;
-		break;
-
-	case DM_PP_CLOCK_TYPE_MEMORY_CLK:
-		{
-			unsigned factor = (ddr4_dram_factor_single_Channel * dc->dcn_soc->number_of_channels);
-
-			if (clocks_in_khz > dc->dcn_soc->fabric_and_dram_bandwidth_vmax0p9*1000000/factor) {
-				vdd_level = dcn_bw_v_max0p91;
-				BREAK_TO_DEBUGGER();
-			} else if (clocks_in_khz > dc->dcn_soc->fabric_and_dram_bandwidth_vnom0p8*1000000/factor) {
-				vdd_level = dcn_bw_v_max0p9;
-			} else if (clocks_in_khz > dc->dcn_soc->fabric_and_dram_bandwidth_vmid0p72*1000000/factor) {
-				vdd_level = dcn_bw_v_nom0p8;
-			} else if (clocks_in_khz > dc->dcn_soc->fabric_and_dram_bandwidth_vmin0p65*1000000/factor) {
-				vdd_level = dcn_bw_v_mid0p72;
-			} else
-				vdd_level = dcn_bw_v_min0p65;
-		}
-		break;
-
-	case DM_PP_CLOCK_TYPE_DCFCLK:
-		if (clocks_in_khz > dc->dcn_soc->dcfclkv_max0p9*1000) {
-			vdd_level = dcn_bw_v_max0p91;
-			BREAK_TO_DEBUGGER();
-		} else if (clocks_in_khz > dc->dcn_soc->dcfclkv_nom0p8*1000) {
-			vdd_level = dcn_bw_v_max0p9;
-		} else if (clocks_in_khz > dc->dcn_soc->dcfclkv_mid0p72*1000) {
-			vdd_level = dcn_bw_v_nom0p8;
-		} else if (clocks_in_khz > dc->dcn_soc->dcfclkv_min0p65*1000) {
-			vdd_level = dcn_bw_v_mid0p72;
-		} else
-			vdd_level = dcn_bw_v_min0p65;
-		break;
-
-	default:
-		 break;
-	}
-	return vdd_level;
-}
-
-unsigned int dcn_find_dcfclk_suits_all(
-	const struct dc *dc,
-	struct dc_clocks *clocks)
+void dcn_bw_update_from_pplib_fclks(
+	struct dc *dc,
+	struct dm_pp_clock_levels_with_voltage *fclks)
 {
-	unsigned vdd_level, vdd_level_temp;
-	unsigned dcf_clk;
-
-	/*find a common supported voltage level*/
-	vdd_level = dcn_find_normalized_clock_vdd_Level(
-		dc, DM_PP_CLOCK_TYPE_DISPLAY_CLK, clocks->dispclk_khz);
-	vdd_level_temp = dcn_find_normalized_clock_vdd_Level(
-		dc, DM_PP_CLOCK_TYPE_DISPLAYPHYCLK, clocks->phyclk_khz);
-
-	vdd_level = dcn_bw_max(vdd_level, vdd_level_temp);
-	vdd_level_temp = dcn_find_normalized_clock_vdd_Level(
-		dc, DM_PP_CLOCK_TYPE_DPPCLK, clocks->dppclk_khz);
-	vdd_level = dcn_bw_max(vdd_level, vdd_level_temp);
-
-	vdd_level_temp = dcn_find_normalized_clock_vdd_Level(
-		dc, DM_PP_CLOCK_TYPE_MEMORY_CLK, clocks->fclk_khz);
-	vdd_level = dcn_bw_max(vdd_level, vdd_level_temp);
-	vdd_level_temp = dcn_find_normalized_clock_vdd_Level(
-		dc, DM_PP_CLOCK_TYPE_DCFCLK, clocks->dcfclk_khz);
-
-	/*find that level conresponding dcfclk*/
-	vdd_level = dcn_bw_max(vdd_level, vdd_level_temp);
-	if (vdd_level == dcn_bw_v_max0p91) {
-		BREAK_TO_DEBUGGER();
-		dcf_clk = dc->dcn_soc->dcfclkv_max0p9*1000;
-	} else if (vdd_level == dcn_bw_v_max0p9)
-		dcf_clk =  dc->dcn_soc->dcfclkv_max0p9*1000;
-	else if (vdd_level == dcn_bw_v_nom0p8)
-		dcf_clk =  dc->dcn_soc->dcfclkv_nom0p8*1000;
-	else if (vdd_level == dcn_bw_v_mid0p72)
-		dcf_clk =  dc->dcn_soc->dcfclkv_mid0p72*1000;
-	else
-		dcf_clk =  dc->dcn_soc->dcfclkv_min0p65*1000;
+	unsigned vmin0p65_idx, vmid0p72_idx, vnom0p8_idx, vmax0p9_idx;
 
-	DC_LOG_BANDWIDTH_CALCS("\tdcf_clk for voltage = %d\n", dcf_clk);
-	return dcf_clk;
+	ASSERT(fclks->num_levels);
+
+	vmin0p65_idx = 0;
+	vmid0p72_idx = fclks->num_levels > 2 ? fclks->num_levels - 3 : 0;
+	vnom0p8_idx = fclks->num_levels > 1 ? fclks->num_levels - 2 : 0;
+	vmax0p9_idx = fclks->num_levels > 0 ? fclks->num_levels - 1 : 0;
+
+	dc->dcn_soc->fabric_and_dram_bandwidth_vmin0p65 =
+		32 * (fclks->data[vmin0p65_idx].clocks_in_khz / 1000.0) / 1000.0;
+	dc->dcn_soc->fabric_and_dram_bandwidth_vmid0p72 =
+		dc->dcn_soc->number_of_channels *
+		(fclks->data[vmid0p72_idx].clocks_in_khz / 1000.0)
+		* ddr4_dram_factor_single_Channel / 1000.0;
+	dc->dcn_soc->fabric_and_dram_bandwidth_vnom0p8 =
+		dc->dcn_soc->number_of_channels *
+		(fclks->data[vnom0p8_idx].clocks_in_khz / 1000.0)
+		* ddr4_dram_factor_single_Channel / 1000.0;
+	dc->dcn_soc->fabric_and_dram_bandwidth_vmax0p9 =
+		dc->dcn_soc->number_of_channels *
+		(fclks->data[vmax0p9_idx].clocks_in_khz / 1000.0)
+		* ddr4_dram_factor_single_Channel / 1000.0;
 }
 
-static bool verify_clock_values(struct dm_pp_clock_levels_with_voltage *clks)
+void dcn_bw_update_from_pplib_dcfclks(
+	struct dc *dc,
+	struct dm_pp_clock_levels_with_voltage *dcfclks)
 {
-	int i;
-
-	if (clks->num_levels == 0)
-		return false;
-
-	for (i = 0; i < clks->num_levels; i++)
-		/* Ensure that the result is sane */
-		if (clks->data[i].clocks_in_khz == 0)
-			return false;
-
-	return true;
+	if (dcfclks->num_levels >= 3) {
+		dc->dcn_soc->dcfclkv_min0p65 = dcfclks->data[0].clocks_in_khz / 1000.0;
+		dc->dcn_soc->dcfclkv_mid0p72 = dcfclks->data[dcfclks->num_levels - 3].clocks_in_khz / 1000.0;
+		dc->dcn_soc->dcfclkv_nom0p8 = dcfclks->data[dcfclks->num_levels - 2].clocks_in_khz / 1000.0;
+		dc->dcn_soc->dcfclkv_max0p9 = dcfclks->data[dcfclks->num_levels - 1].clocks_in_khz / 1000.0;
+	}
 }
 
-void dcn_bw_update_from_pplib(struct dc *dc)
+void dcn_get_soc_clks(
+	struct dc *dc,
+	int *min_fclk_khz,
+	int *min_dcfclk_khz,
+	int *socclk_khz)
 {
-	struct dc_context *ctx = dc->ctx;
-	struct dm_pp_clock_levels_with_voltage fclks = {0}, dcfclks = {0};
-	bool res;
-	unsigned vmin0p65_idx, vmid0p72_idx, vnom0p8_idx, vmax0p9_idx;
-
-	/* TODO: This is not the proper way to obtain fabric_and_dram_bandwidth, should be min(fclk, memclk) */
-	res = dm_pp_get_clock_levels_by_type_with_voltage(
-			ctx, DM_PP_CLOCK_TYPE_FCLK, &fclks);
-
-	DC_FP_START();
-
-	if (res)
-		res = verify_clock_values(&fclks);
-
-	if (res) {
-		ASSERT(fclks.num_levels);
-
-		vmin0p65_idx = 0;
-		vmid0p72_idx = fclks.num_levels -
-			(fclks.num_levels > 2 ? 3 : (fclks.num_levels > 1 ? 2 : 1));
-		vnom0p8_idx = fclks.num_levels - (fclks.num_levels > 1 ? 2 : 1);
-		vmax0p9_idx = fclks.num_levels - 1;
-
-		dc->dcn_soc->fabric_and_dram_bandwidth_vmin0p65 =
-			32 * (fclks.data[vmin0p65_idx].clocks_in_khz / 1000.0) / 1000.0;
-		dc->dcn_soc->fabric_and_dram_bandwidth_vmid0p72 =
-			dc->dcn_soc->number_of_channels *
-			(fclks.data[vmid0p72_idx].clocks_in_khz / 1000.0)
-			* ddr4_dram_factor_single_Channel / 1000.0;
-		dc->dcn_soc->fabric_and_dram_bandwidth_vnom0p8 =
-			dc->dcn_soc->number_of_channels *
-			(fclks.data[vnom0p8_idx].clocks_in_khz / 1000.0)
-			* ddr4_dram_factor_single_Channel / 1000.0;
-		dc->dcn_soc->fabric_and_dram_bandwidth_vmax0p9 =
-			dc->dcn_soc->number_of_channels *
-			(fclks.data[vmax0p9_idx].clocks_in_khz / 1000.0)
-			* ddr4_dram_factor_single_Channel / 1000.0;
-	} else
-		BREAK_TO_DEBUGGER();
-
-	DC_FP_END();
-
-	res = dm_pp_get_clock_levels_by_type_with_voltage(
-			ctx, DM_PP_CLOCK_TYPE_DCFCLK, &dcfclks);
-
-	DC_FP_START();
-
-	if (res)
-		res = verify_clock_values(&dcfclks);
-
-	if (res && dcfclks.num_levels >= 3) {
-		dc->dcn_soc->dcfclkv_min0p65 = dcfclks.data[0].clocks_in_khz / 1000.0;
-		dc->dcn_soc->dcfclkv_mid0p72 = dcfclks.data[dcfclks.num_levels - 3].clocks_in_khz / 1000.0;
-		dc->dcn_soc->dcfclkv_nom0p8 = dcfclks.data[dcfclks.num_levels - 2].clocks_in_khz / 1000.0;
-		dc->dcn_soc->dcfclkv_max0p9 = dcfclks.data[dcfclks.num_levels - 1].clocks_in_khz / 1000.0;
-	} else
-		BREAK_TO_DEBUGGER();
-
-	DC_FP_END();
+	*min_fclk_khz = dc->dcn_soc->fabric_and_dram_bandwidth_vmin0p65 * 1000000 / 32;
+	*min_dcfclk_khz = dc->dcn_soc->dcfclkv_min0p65 * 1000;
+	*socclk_khz = dc->dcn_soc->socclk * 1000;
 }
 
-void dcn_bw_notify_pplib_of_wm_ranges(struct dc *dc)
+void dcn_bw_notify_pplib_of_wm_ranges(
+	struct dc *dc,
+	int min_fclk_khz,
+	int min_dcfclk_khz,
+	int socclk_khz)
 {
 	struct pp_smu_funcs_rv *pp = NULL;
 	struct pp_smu_wm_range_sets ranges = {0};
-	int min_fclk_khz, min_dcfclk_khz, socclk_khz;
 	const int overdrive = 5000000; /* 5 GHz to cover Overdrive */
 
 	if (dc->res_pool->pp_smu)
@@ -1558,12 +1379,6 @@ void dcn_bw_notify_pplib_of_wm_ranges(struct dc *dc)
 	if (!pp || !pp->set_wm_ranges)
 		return;
 
-	DC_FP_START();
-	min_fclk_khz = dc->dcn_soc->fabric_and_dram_bandwidth_vmin0p65 * 1000000 / 32;
-	min_dcfclk_khz = dc->dcn_soc->dcfclkv_min0p65 * 1000;
-	socclk_khz = dc->dcn_soc->socclk * 1000;
-	DC_FP_END();
-
 	/* Now notify PPLib/SMU about which Watermarks sets they should select
 	 * depending on DPM state they are in. And update BW MGR GFX Engine and
 	 * Memory clock member variables for Watermarks calculations for each
@@ -1613,7 +1428,6 @@ void dcn_bw_notify_pplib_of_wm_ranges(struct dc *dc)
 
 void dcn_bw_sync_calcs_and_dml(struct dc *dc)
 {
-	DC_FP_START();
 	DC_LOG_BANDWIDTH_CALCS("sr_exit_time: %f ns\n"
 			"sr_enter_plus_exit_time: %f ns\n"
 			"urgent_latency: %f ns\n"
@@ -1802,5 +1616,4 @@ void dcn_bw_sync_calcs_and_dml(struct dc *dc)
 	dc->dml.ip.bug_forcing_LC_req_same_size_fixed =
 		dc->dcn_ip->bug_forcing_luma_and_chroma_request_to_same_size_fixed == dcn_bw_yes;
 	dc->dml.ip.dcfclk_cstate_latency = dc->dcn_ip->dcfclk_cstate_latency;
-	DC_FP_END();
 }
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dc_features.h b/drivers/gpu/drm/amd/display/dc/dml/dc_features.h
index 2a1983324629..6e669a2c5b2d 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dc_features.h
+++ b/drivers/gpu/drm/amd/display/dc/dml/dc_features.h
@@ -29,7 +29,14 @@
 #define DC__PRESENT 1
 #define DC__PRESENT__1 1
 #define DC__NUM_DPP 4
-#define DC__VOLTAGE_STATES 9
+
+/**
+ * @DC__VOLTAGE_STATES:
+ *
+ * Define the maximum amount of states supported by the ASIC. Every ASIC has a
+ * specific number of states; this macro defines the maximum number of states.
+ */
+#define DC__VOLTAGE_STATES 40
 #define DC__NUM_DPP__4 1
 #define DC__NUM_DPP__0_PRESENT 1
 #define DC__NUM_DPP__1_PRESENT 1
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn10/dcn10_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn10/dcn10_fpu.c
new file mode 100644
index 000000000000..c5e84190c17a
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn10/dcn10_fpu.c
@@ -0,0 +1,159 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#include "dcn10/dcn10_resource.h"
+
+#include "dcn10_fpu.h"
+#include "resource.h"
+#include "amdgpu_dm/dc_fpu.h"
+
+/**
+ * DOC: DCN10 FPU manipulation Overview
+ *
+ * The DCN architecture relies on FPU operations, which require special
+ * compilation flags and the use of kernel_fpu_begin/end functions; ideally, we
+ * want to avoid spreading FPU access across multiple files. With this idea in
+ * mind, this file aims to centralize DCN10 functions that require FPU access
+ * in a single place. Code in this file follows the following code pattern:
+ *
+ * 1. Functions that use FPU operations should be isolated in static functions.
+ * 2. The FPU functions should have the noinline attribute to ensure anything
+ *    that deals with FP register is contained within this call.
+ * 3. All function that needs to be accessed outside this file requires a
+ *    public interface that not uses any FPU reference.
+ * 4. Developers **must not** use DC_FP_START/END in this file, but they need
+ *    to ensure that the caller invokes it before access any function available
+ *    in this file. For this reason, public functions in this file must invoke
+ *    dc_assert_fp_enabled();
+ *
+ * Let's expand a little bit more the idea in the code pattern. To fully
+ * isolate FPU operations in a single place, we must avoid situations where
+ * compilers spill FP values to registers due to FP enable in a specific C
+ * file. Note that even if we isolate all FPU functions in a single file and
+ * call its interface from other files, the compiler might enable the use of
+ * FPU before we call DC_FP_START. Nevertheless, it is the programmer's
+ * responsibility to invoke DC_FP_START/END in the correct place. To highlight
+ * situations where developers forgot to use the FP protection before calling
+ * the DC FPU interface functions, we introduce a helper that checks if the
+ * function is invoked under FP protection. If not, it will trigger a kernel
+ * warning.
+ */
+
+struct _vcs_dpi_ip_params_st dcn1_0_ip = {
+	.rob_buffer_size_kbytes = 64,
+	.det_buffer_size_kbytes = 164,
+	.dpte_buffer_size_in_pte_reqs_luma = 42,
+	.dpp_output_buffer_pixels = 2560,
+	.opp_output_buffer_lines = 1,
+	.pixel_chunk_size_kbytes = 8,
+	.pte_enable = 1,
+	.pte_chunk_size_kbytes = 2,
+	.meta_chunk_size_kbytes = 2,
+	.writeback_chunk_size_kbytes = 2,
+	.line_buffer_size_bits = 589824,
+	.max_line_buffer_lines = 12,
+	.IsLineBufferBppFixed = 0,
+	.LineBufferFixedBpp = -1,
+	.writeback_luma_buffer_size_kbytes = 12,
+	.writeback_chroma_buffer_size_kbytes = 8,
+	.max_num_dpp = 4,
+	.max_num_wb = 2,
+	.max_dchub_pscl_bw_pix_per_clk = 4,
+	.max_pscl_lb_bw_pix_per_clk = 2,
+	.max_lb_vscl_bw_pix_per_clk = 4,
+	.max_vscl_hscl_bw_pix_per_clk = 4,
+	.max_hscl_ratio = 4,
+	.max_vscl_ratio = 4,
+	.hscl_mults = 4,
+	.vscl_mults = 4,
+	.max_hscl_taps = 8,
+	.max_vscl_taps = 8,
+	.dispclk_ramp_margin_percent = 1,
+	.underscan_factor = 1.10,
+	.min_vblank_lines = 14,
+	.dppclk_delay_subtotal = 90,
+	.dispclk_delay_subtotal = 42,
+	.dcfclk_cstate_latency = 10,
+	.max_inter_dcn_tile_repeaters = 8,
+	.can_vstartup_lines_exceed_vsync_plus_back_porch_lines_minus_one = 0,
+	.bug_forcing_LC_req_same_size_fixed = 0,
+};
+
+struct _vcs_dpi_soc_bounding_box_st dcn1_0_soc = {
+	.sr_exit_time_us = 9.0,
+	.sr_enter_plus_exit_time_us = 11.0,
+	.urgent_latency_us = 4.0,
+	.writeback_latency_us = 12.0,
+	.ideal_dram_bw_after_urgent_percent = 80.0,
+	.max_request_size_bytes = 256,
+	.downspread_percent = 0.5,
+	.dram_page_open_time_ns = 50.0,
+	.dram_rw_turnaround_time_ns = 17.5,
+	.dram_return_buffer_per_channel_bytes = 8192,
+	.round_trip_ping_latency_dcfclk_cycles = 128,
+	.urgent_out_of_order_return_per_channel_bytes = 256,
+	.channel_interleave_bytes = 256,
+	.num_banks = 8,
+	.num_chans = 2,
+	.vmm_page_size_bytes = 4096,
+	.dram_clock_change_latency_us = 17.0,
+	.writeback_dram_clock_change_latency_us = 23.0,
+	.return_bus_width_bytes = 64,
+};
+
+void dcn10_resource_construct_fp(struct dc *dc)
+{
+	dc_assert_fp_enabled();
+	if (dc->ctx->dce_version == DCN_VERSION_1_01) {
+		struct dcn_soc_bounding_box *dcn_soc = dc->dcn_soc;
+		struct dcn_ip_params *dcn_ip = dc->dcn_ip;
+		struct display_mode_lib *dml = &dc->dml;
+
+		dml->ip.max_num_dpp = 3;
+		/* TODO how to handle 23.84? */
+		dcn_soc->dram_clock_change_latency = 23;
+		dcn_ip->max_num_dpp = 3;
+	}
+	if (ASICREV_IS_RV1_F0(dc->ctx->asic_id.hw_internal_rev)) {
+		dc->dcn_soc->urgent_latency = 3;
+		dc->debug.disable_dmcu = true;
+		dc->dcn_soc->fabric_and_dram_bandwidth_vmax0p9 = 41.60f;
+	}
+
+	dc->dcn_soc->number_of_channels = dc->ctx->asic_id.vram_width / ddr4_dram_width;
+	ASSERT(dc->dcn_soc->number_of_channels < 3);
+	if (dc->dcn_soc->number_of_channels == 0)/*old sbios bug*/
+		dc->dcn_soc->number_of_channels = 2;
+
+	if (dc->dcn_soc->number_of_channels == 1) {
+		dc->dcn_soc->fabric_and_dram_bandwidth_vmax0p9 = 19.2f;
+		dc->dcn_soc->fabric_and_dram_bandwidth_vnom0p8 = 17.066f;
+		dc->dcn_soc->fabric_and_dram_bandwidth_vmid0p72 = 14.933f;
+		dc->dcn_soc->fabric_and_dram_bandwidth_vmin0p65 = 12.8f;
+		if (ASICREV_IS_RV1_F0(dc->ctx->asic_id.hw_internal_rev))
+			dc->dcn_soc->fabric_and_dram_bandwidth_vmax0p9 = 20.80f;
+	}
+}
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn10/dcn10_fpu.h b/drivers/gpu/drm/amd/display/dc/dml/dcn10/dcn10_fpu.h
new file mode 100644
index 000000000000..1bf6b12f5663
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn10/dcn10_fpu.h
@@ -0,0 +1,32 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __DCN10_FPU_H__
+#define __DCN10_FPU_H__
+
+void dcn10_resource_construct_fp(struct dc *dc);
+
+#endif /* __DCN10_FPU_H__ */
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn20/dcn20_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn20/dcn20_fpu.c
index d590dc917363..7aaf13bbd4e4 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dcn20/dcn20_fpu.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn20/dcn20_fpu.c
@@ -25,8 +25,28 @@
  */
 
 #include "resource.h"
-
+#include "clk_mgr.h"
+#include "dchubbub.h"
+#include "dcn20/dcn20_resource.h"
+#include "dcn21/dcn21_resource.h"
+#include "clk_mgr/dcn21/rn_clk_mgr.h"
+#include "link_service.h"
 #include "dcn20_fpu.h"
+#include "dc_state_priv.h"
+
+#define DC_LOGGER \
+	dc->ctx->logger
+#define DC_LOGGER_INIT(logger)
+
+#ifndef MAX
+#define MAX(X, Y) ((X) > (Y) ? (X) : (Y))
+#endif
+#ifndef MIN
+#define MIN(X, Y) ((X) < (Y) ? (X) : (Y))
+#endif
+
+/* Constant */
+#define LPDDR_MEM_RETRAIN_LATENCY 4.977 /* Number obtained from LPDDR4 Training Counter Requirement doc */
 
 /**
  * DOC: DCN2x FPU manipulation Overview
@@ -61,6 +81,911 @@
  * warning.
  */
 
+struct _vcs_dpi_ip_params_st dcn2_0_ip = {
+	.odm_capable = 1,
+	.gpuvm_enable = 0,
+	.hostvm_enable = 0,
+	.gpuvm_max_page_table_levels = 4,
+	.hostvm_max_page_table_levels = 4,
+	.hostvm_cached_page_table_levels = 0,
+	.pte_group_size_bytes = 2048,
+	.num_dsc = 6,
+	.rob_buffer_size_kbytes = 168,
+	.det_buffer_size_kbytes = 164,
+	.dpte_buffer_size_in_pte_reqs_luma = 84,
+	.pde_proc_buffer_size_64k_reqs = 48,
+	.dpp_output_buffer_pixels = 2560,
+	.opp_output_buffer_lines = 1,
+	.pixel_chunk_size_kbytes = 8,
+	.pte_chunk_size_kbytes = 2,
+	.meta_chunk_size_kbytes = 2,
+	.writeback_chunk_size_kbytes = 2,
+	.line_buffer_size_bits = 789504,
+	.is_line_buffer_bpp_fixed = 0,
+	.line_buffer_fixed_bpp = 0,
+	.dcc_supported = true,
+	.max_line_buffer_lines = 12,
+	.writeback_luma_buffer_size_kbytes = 12,
+	.writeback_chroma_buffer_size_kbytes = 8,
+	.writeback_chroma_line_buffer_width_pixels = 4,
+	.writeback_max_hscl_ratio = 1,
+	.writeback_max_vscl_ratio = 1,
+	.writeback_min_hscl_ratio = 1,
+	.writeback_min_vscl_ratio = 1,
+	.writeback_max_hscl_taps = 12,
+	.writeback_max_vscl_taps = 12,
+	.writeback_line_buffer_luma_buffer_size = 0,
+	.writeback_line_buffer_chroma_buffer_size = 14643,
+	.cursor_buffer_size = 8,
+	.cursor_chunk_size = 2,
+	.max_num_otg = 6,
+	.max_num_dpp = 6,
+	.max_num_wb = 1,
+	.max_dchub_pscl_bw_pix_per_clk = 4,
+	.max_pscl_lb_bw_pix_per_clk = 2,
+	.max_lb_vscl_bw_pix_per_clk = 4,
+	.max_vscl_hscl_bw_pix_per_clk = 4,
+	.max_hscl_ratio = 8,
+	.max_vscl_ratio = 8,
+	.hscl_mults = 4,
+	.vscl_mults = 4,
+	.max_hscl_taps = 8,
+	.max_vscl_taps = 8,
+	.dispclk_ramp_margin_percent = 1,
+	.underscan_factor = 1.10,
+	.min_vblank_lines = 32, //
+	.dppclk_delay_subtotal = 77, //
+	.dppclk_delay_scl_lb_only = 16,
+	.dppclk_delay_scl = 50,
+	.dppclk_delay_cnvc_formatter = 8,
+	.dppclk_delay_cnvc_cursor = 6,
+	.dispclk_delay_subtotal = 87, //
+	.dcfclk_cstate_latency = 10, // SRExitTime
+	.max_inter_dcn_tile_repeaters = 8,
+	.xfc_supported = true,
+	.xfc_fill_bw_overhead_percent = 10.0,
+	.xfc_fill_constant_bytes = 0,
+	.number_of_cursors = 1,
+};
+
+struct _vcs_dpi_ip_params_st dcn2_0_nv14_ip = {
+	.odm_capable = 1,
+	.gpuvm_enable = 0,
+	.hostvm_enable = 0,
+	.gpuvm_max_page_table_levels = 4,
+	.hostvm_max_page_table_levels = 4,
+	.hostvm_cached_page_table_levels = 0,
+	.num_dsc = 5,
+	.rob_buffer_size_kbytes = 168,
+	.det_buffer_size_kbytes = 164,
+	.dpte_buffer_size_in_pte_reqs_luma = 84,
+	.dpte_buffer_size_in_pte_reqs_chroma = 42,//todo
+	.dpp_output_buffer_pixels = 2560,
+	.opp_output_buffer_lines = 1,
+	.pixel_chunk_size_kbytes = 8,
+	.pte_enable = 1,
+	.max_page_table_levels = 4,
+	.pte_chunk_size_kbytes = 2,
+	.meta_chunk_size_kbytes = 2,
+	.writeback_chunk_size_kbytes = 2,
+	.line_buffer_size_bits = 789504,
+	.is_line_buffer_bpp_fixed = 0,
+	.line_buffer_fixed_bpp = 0,
+	.dcc_supported = true,
+	.max_line_buffer_lines = 12,
+	.writeback_luma_buffer_size_kbytes = 12,
+	.writeback_chroma_buffer_size_kbytes = 8,
+	.writeback_chroma_line_buffer_width_pixels = 4,
+	.writeback_max_hscl_ratio = 1,
+	.writeback_max_vscl_ratio = 1,
+	.writeback_min_hscl_ratio = 1,
+	.writeback_min_vscl_ratio = 1,
+	.writeback_max_hscl_taps = 12,
+	.writeback_max_vscl_taps = 12,
+	.writeback_line_buffer_luma_buffer_size = 0,
+	.writeback_line_buffer_chroma_buffer_size = 14643,
+	.cursor_buffer_size = 8,
+	.cursor_chunk_size = 2,
+	.max_num_otg = 5,
+	.max_num_dpp = 5,
+	.max_num_wb = 1,
+	.max_dchub_pscl_bw_pix_per_clk = 4,
+	.max_pscl_lb_bw_pix_per_clk = 2,
+	.max_lb_vscl_bw_pix_per_clk = 4,
+	.max_vscl_hscl_bw_pix_per_clk = 4,
+	.max_hscl_ratio = 8,
+	.max_vscl_ratio = 8,
+	.hscl_mults = 4,
+	.vscl_mults = 4,
+	.max_hscl_taps = 8,
+	.max_vscl_taps = 8,
+	.dispclk_ramp_margin_percent = 1,
+	.underscan_factor = 1.10,
+	.min_vblank_lines = 32, //
+	.dppclk_delay_subtotal = 77, //
+	.dppclk_delay_scl_lb_only = 16,
+	.dppclk_delay_scl = 50,
+	.dppclk_delay_cnvc_formatter = 8,
+	.dppclk_delay_cnvc_cursor = 6,
+	.dispclk_delay_subtotal = 87, //
+	.dcfclk_cstate_latency = 10, // SRExitTime
+	.max_inter_dcn_tile_repeaters = 8,
+	.xfc_supported = true,
+	.xfc_fill_bw_overhead_percent = 10.0,
+	.xfc_fill_constant_bytes = 0,
+	.ptoi_supported = 0,
+	.number_of_cursors = 1,
+};
+
+struct _vcs_dpi_soc_bounding_box_st dcn2_0_soc = {
+	/* Defaults that get patched on driver load from firmware. */
+	.clock_limits = {
+			{
+				.state = 0,
+				.dcfclk_mhz = 560.0,
+				.fabricclk_mhz = 560.0,
+				.dispclk_mhz = 513.0,
+				.dppclk_mhz = 513.0,
+				.phyclk_mhz = 540.0,
+				.socclk_mhz = 560.0,
+				.dscclk_mhz = 171.0,
+				.dram_speed_mts = 8960.0,
+			},
+			{
+				.state = 1,
+				.dcfclk_mhz = 694.0,
+				.fabricclk_mhz = 694.0,
+				.dispclk_mhz = 642.0,
+				.dppclk_mhz = 642.0,
+				.phyclk_mhz = 600.0,
+				.socclk_mhz = 694.0,
+				.dscclk_mhz = 214.0,
+				.dram_speed_mts = 11104.0,
+			},
+			{
+				.state = 2,
+				.dcfclk_mhz = 875.0,
+				.fabricclk_mhz = 875.0,
+				.dispclk_mhz = 734.0,
+				.dppclk_mhz = 734.0,
+				.phyclk_mhz = 810.0,
+				.socclk_mhz = 875.0,
+				.dscclk_mhz = 245.0,
+				.dram_speed_mts = 14000.0,
+			},
+			{
+				.state = 3,
+				.dcfclk_mhz = 1000.0,
+				.fabricclk_mhz = 1000.0,
+				.dispclk_mhz = 1100.0,
+				.dppclk_mhz = 1100.0,
+				.phyclk_mhz = 810.0,
+				.socclk_mhz = 1000.0,
+				.dscclk_mhz = 367.0,
+				.dram_speed_mts = 16000.0,
+			},
+			{
+				.state = 4,
+				.dcfclk_mhz = 1200.0,
+				.fabricclk_mhz = 1200.0,
+				.dispclk_mhz = 1284.0,
+				.dppclk_mhz = 1284.0,
+				.phyclk_mhz = 810.0,
+				.socclk_mhz = 1200.0,
+				.dscclk_mhz = 428.0,
+				.dram_speed_mts = 16000.0,
+			},
+			/*Extra state, no dispclk ramping*/
+			{
+				.state = 5,
+				.dcfclk_mhz = 1200.0,
+				.fabricclk_mhz = 1200.0,
+				.dispclk_mhz = 1284.0,
+				.dppclk_mhz = 1284.0,
+				.phyclk_mhz = 810.0,
+				.socclk_mhz = 1200.0,
+				.dscclk_mhz = 428.0,
+				.dram_speed_mts = 16000.0,
+			},
+		},
+	.num_states = 5,
+	.sr_exit_time_us = 8.6,
+	.sr_enter_plus_exit_time_us = 10.9,
+	.urgent_latency_us = 4.0,
+	.urgent_latency_pixel_data_only_us = 4.0,
+	.urgent_latency_pixel_mixed_with_vm_data_us = 4.0,
+	.urgent_latency_vm_data_only_us = 4.0,
+	.urgent_out_of_order_return_per_channel_pixel_only_bytes = 4096,
+	.urgent_out_of_order_return_per_channel_pixel_and_vm_bytes = 4096,
+	.urgent_out_of_order_return_per_channel_vm_only_bytes = 4096,
+	.pct_ideal_dram_sdp_bw_after_urgent_pixel_only = 40.0,
+	.pct_ideal_dram_sdp_bw_after_urgent_pixel_and_vm = 40.0,
+	.pct_ideal_dram_sdp_bw_after_urgent_vm_only = 40.0,
+	.max_avg_sdp_bw_use_normal_percent = 40.0,
+	.max_avg_dram_bw_use_normal_percent = 40.0,
+	.writeback_latency_us = 12.0,
+	.ideal_dram_bw_after_urgent_percent = 40.0,
+	.max_request_size_bytes = 256,
+	.dram_channel_width_bytes = 2,
+	.fabric_datapath_to_dcn_data_return_bytes = 64,
+	.dcn_downspread_percent = 0.5,
+	.downspread_percent = 0.38,
+	.dram_page_open_time_ns = 50.0,
+	.dram_rw_turnaround_time_ns = 17.5,
+	.dram_return_buffer_per_channel_bytes = 8192,
+	.round_trip_ping_latency_dcfclk_cycles = 131,
+	.urgent_out_of_order_return_per_channel_bytes = 256,
+	.channel_interleave_bytes = 256,
+	.num_banks = 8,
+	.num_chans = 16,
+	.vmm_page_size_bytes = 4096,
+	.dram_clock_change_latency_us = 404.0,
+	.dummy_pstate_latency_us = 5.0,
+	.writeback_dram_clock_change_latency_us = 23.0,
+	.return_bus_width_bytes = 64,
+	.dispclk_dppclk_vco_speed_mhz = 3850,
+	.xfc_bus_transport_time_us = 20,
+	.xfc_xbuf_latency_tolerance_us = 4,
+	.use_urgent_burst_bw = 0
+};
+
+struct _vcs_dpi_soc_bounding_box_st dcn2_0_nv14_soc = {
+	.clock_limits = {
+			{
+				.state = 0,
+				.dcfclk_mhz = 560.0,
+				.fabricclk_mhz = 560.0,
+				.dispclk_mhz = 513.0,
+				.dppclk_mhz = 513.0,
+				.phyclk_mhz = 540.0,
+				.socclk_mhz = 560.0,
+				.dscclk_mhz = 171.0,
+				.dram_speed_mts = 8960.0,
+			},
+			{
+				.state = 1,
+				.dcfclk_mhz = 694.0,
+				.fabricclk_mhz = 694.0,
+				.dispclk_mhz = 642.0,
+				.dppclk_mhz = 642.0,
+				.phyclk_mhz = 600.0,
+				.socclk_mhz = 694.0,
+				.dscclk_mhz = 214.0,
+				.dram_speed_mts = 11104.0,
+			},
+			{
+				.state = 2,
+				.dcfclk_mhz = 875.0,
+				.fabricclk_mhz = 875.0,
+				.dispclk_mhz = 734.0,
+				.dppclk_mhz = 734.0,
+				.phyclk_mhz = 810.0,
+				.socclk_mhz = 875.0,
+				.dscclk_mhz = 245.0,
+				.dram_speed_mts = 14000.0,
+			},
+			{
+				.state = 3,
+				.dcfclk_mhz = 1000.0,
+				.fabricclk_mhz = 1000.0,
+				.dispclk_mhz = 1100.0,
+				.dppclk_mhz = 1100.0,
+				.phyclk_mhz = 810.0,
+				.socclk_mhz = 1000.0,
+				.dscclk_mhz = 367.0,
+				.dram_speed_mts = 16000.0,
+			},
+			{
+				.state = 4,
+				.dcfclk_mhz = 1200.0,
+				.fabricclk_mhz = 1200.0,
+				.dispclk_mhz = 1284.0,
+				.dppclk_mhz = 1284.0,
+				.phyclk_mhz = 810.0,
+				.socclk_mhz = 1200.0,
+				.dscclk_mhz = 428.0,
+				.dram_speed_mts = 16000.0,
+			},
+			/*Extra state, no dispclk ramping*/
+			{
+				.state = 5,
+				.dcfclk_mhz = 1200.0,
+				.fabricclk_mhz = 1200.0,
+				.dispclk_mhz = 1284.0,
+				.dppclk_mhz = 1284.0,
+				.phyclk_mhz = 810.0,
+				.socclk_mhz = 1200.0,
+				.dscclk_mhz = 428.0,
+				.dram_speed_mts = 16000.0,
+			},
+		},
+	.num_states = 5,
+	.sr_exit_time_us = 11.6,
+	.sr_enter_plus_exit_time_us = 13.9,
+	.urgent_latency_us = 4.0,
+	.urgent_latency_pixel_data_only_us = 4.0,
+	.urgent_latency_pixel_mixed_with_vm_data_us = 4.0,
+	.urgent_latency_vm_data_only_us = 4.0,
+	.urgent_out_of_order_return_per_channel_pixel_only_bytes = 4096,
+	.urgent_out_of_order_return_per_channel_pixel_and_vm_bytes = 4096,
+	.urgent_out_of_order_return_per_channel_vm_only_bytes = 4096,
+	.pct_ideal_dram_sdp_bw_after_urgent_pixel_only = 40.0,
+	.pct_ideal_dram_sdp_bw_after_urgent_pixel_and_vm = 40.0,
+	.pct_ideal_dram_sdp_bw_after_urgent_vm_only = 40.0,
+	.max_avg_sdp_bw_use_normal_percent = 40.0,
+	.max_avg_dram_bw_use_normal_percent = 40.0,
+	.writeback_latency_us = 12.0,
+	.ideal_dram_bw_after_urgent_percent = 40.0,
+	.max_request_size_bytes = 256,
+	.dram_channel_width_bytes = 2,
+	.fabric_datapath_to_dcn_data_return_bytes = 64,
+	.dcn_downspread_percent = 0.5,
+	.downspread_percent = 0.38,
+	.dram_page_open_time_ns = 50.0,
+	.dram_rw_turnaround_time_ns = 17.5,
+	.dram_return_buffer_per_channel_bytes = 8192,
+	.round_trip_ping_latency_dcfclk_cycles = 131,
+	.urgent_out_of_order_return_per_channel_bytes = 256,
+	.channel_interleave_bytes = 256,
+	.num_banks = 8,
+	.num_chans = 8,
+	.vmm_page_size_bytes = 4096,
+	.dram_clock_change_latency_us = 404.0,
+	.dummy_pstate_latency_us = 5.0,
+	.writeback_dram_clock_change_latency_us = 23.0,
+	.return_bus_width_bytes = 64,
+	.dispclk_dppclk_vco_speed_mhz = 3850,
+	.xfc_bus_transport_time_us = 20,
+	.xfc_xbuf_latency_tolerance_us = 4,
+	.use_urgent_burst_bw = 0
+};
+
+struct _vcs_dpi_soc_bounding_box_st dcn2_0_nv12_soc = {
+	.clock_limits = {
+		{
+			.state = 0,
+			.dcfclk_mhz = 560.0,
+			.fabricclk_mhz = 560.0,
+			.dispclk_mhz = 513.0,
+			.dppclk_mhz = 513.0,
+			.phyclk_mhz = 540.0,
+			.socclk_mhz = 560.0,
+			.dscclk_mhz = 171.0,
+			.dram_speed_mts = 1069.0,
+		},
+		{
+			.state = 1,
+			.dcfclk_mhz = 694.0,
+			.fabricclk_mhz = 694.0,
+			.dispclk_mhz = 642.0,
+			.dppclk_mhz = 642.0,
+			.phyclk_mhz = 600.0,
+			.socclk_mhz = 694.0,
+			.dscclk_mhz = 214.0,
+			.dram_speed_mts = 1324.0,
+		},
+		{
+			.state = 2,
+			.dcfclk_mhz = 875.0,
+			.fabricclk_mhz = 875.0,
+			.dispclk_mhz = 734.0,
+			.dppclk_mhz = 734.0,
+			.phyclk_mhz = 810.0,
+			.socclk_mhz = 875.0,
+			.dscclk_mhz = 245.0,
+			.dram_speed_mts = 1670.0,
+		},
+		{
+			.state = 3,
+			.dcfclk_mhz = 1000.0,
+			.fabricclk_mhz = 1000.0,
+			.dispclk_mhz = 1100.0,
+			.dppclk_mhz = 1100.0,
+			.phyclk_mhz = 810.0,
+			.socclk_mhz = 1000.0,
+			.dscclk_mhz = 367.0,
+			.dram_speed_mts = 2000.0,
+		},
+		{
+			.state = 4,
+			.dcfclk_mhz = 1200.0,
+			.fabricclk_mhz = 1200.0,
+			.dispclk_mhz = 1284.0,
+			.dppclk_mhz = 1284.0,
+			.phyclk_mhz = 810.0,
+			.socclk_mhz = 1200.0,
+			.dscclk_mhz = 428.0,
+			.dram_speed_mts = 2000.0,
+		},
+		{
+			.state = 5,
+			.dcfclk_mhz = 1200.0,
+			.fabricclk_mhz = 1200.0,
+			.dispclk_mhz = 1284.0,
+			.dppclk_mhz = 1284.0,
+			.phyclk_mhz = 810.0,
+			.socclk_mhz = 1200.0,
+			.dscclk_mhz = 428.0,
+			.dram_speed_mts = 2000.0,
+		},
+	},
+
+	.num_states = 5,
+	.sr_exit_time_us = 1.9,
+	.sr_enter_plus_exit_time_us = 4.4,
+	.urgent_latency_us = 3.0,
+	.urgent_latency_pixel_data_only_us = 4.0,
+	.urgent_latency_pixel_mixed_with_vm_data_us = 4.0,
+	.urgent_latency_vm_data_only_us = 4.0,
+	.urgent_out_of_order_return_per_channel_pixel_only_bytes = 4096,
+	.urgent_out_of_order_return_per_channel_pixel_and_vm_bytes = 4096,
+	.urgent_out_of_order_return_per_channel_vm_only_bytes = 4096,
+	.pct_ideal_dram_sdp_bw_after_urgent_pixel_only = 40.0,
+	.pct_ideal_dram_sdp_bw_after_urgent_pixel_and_vm = 40.0,
+	.pct_ideal_dram_sdp_bw_after_urgent_vm_only = 40.0,
+	.max_avg_sdp_bw_use_normal_percent = 40.0,
+	.max_avg_dram_bw_use_normal_percent = 40.0,
+	.writeback_latency_us = 12.0,
+	.ideal_dram_bw_after_urgent_percent = 40.0,
+	.max_request_size_bytes = 256,
+	.dram_channel_width_bytes = 16,
+	.fabric_datapath_to_dcn_data_return_bytes = 64,
+	.dcn_downspread_percent = 0.5,
+	.downspread_percent = 0.5,
+	.dram_page_open_time_ns = 50.0,
+	.dram_rw_turnaround_time_ns = 17.5,
+	.dram_return_buffer_per_channel_bytes = 8192,
+	.round_trip_ping_latency_dcfclk_cycles = 131,
+	.urgent_out_of_order_return_per_channel_bytes = 4096,
+	.channel_interleave_bytes = 256,
+	.num_banks = 8,
+	.num_chans = 16,
+	.vmm_page_size_bytes = 4096,
+	.dram_clock_change_latency_us = 45.0,
+	.writeback_dram_clock_change_latency_us = 23.0,
+	.return_bus_width_bytes = 64,
+	.dispclk_dppclk_vco_speed_mhz = 3850,
+	.xfc_bus_transport_time_us = 20,
+	.xfc_xbuf_latency_tolerance_us = 50,
+	.use_urgent_burst_bw = 0,
+};
+
+struct _vcs_dpi_ip_params_st dcn2_1_ip = {
+	.odm_capable = 1,
+	.gpuvm_enable = 1,
+	.hostvm_enable = 1,
+	.gpuvm_max_page_table_levels = 1,
+	.hostvm_max_page_table_levels = 4,
+	.hostvm_cached_page_table_levels = 2,
+	.num_dsc = 3,
+	.rob_buffer_size_kbytes = 168,
+	.det_buffer_size_kbytes = 164,
+	.dpte_buffer_size_in_pte_reqs_luma = 44,
+	.dpte_buffer_size_in_pte_reqs_chroma = 42,//todo
+	.dpp_output_buffer_pixels = 2560,
+	.opp_output_buffer_lines = 1,
+	.pixel_chunk_size_kbytes = 8,
+	.pte_enable = 1,
+	.max_page_table_levels = 4,
+	.pte_chunk_size_kbytes = 2,
+	.meta_chunk_size_kbytes = 2,
+	.min_meta_chunk_size_bytes = 256,
+	.writeback_chunk_size_kbytes = 2,
+	.line_buffer_size_bits = 789504,
+	.is_line_buffer_bpp_fixed = 0,
+	.line_buffer_fixed_bpp = 0,
+	.dcc_supported = true,
+	.max_line_buffer_lines = 12,
+	.writeback_luma_buffer_size_kbytes = 12,
+	.writeback_chroma_buffer_size_kbytes = 8,
+	.writeback_chroma_line_buffer_width_pixels = 4,
+	.writeback_max_hscl_ratio = 1,
+	.writeback_max_vscl_ratio = 1,
+	.writeback_min_hscl_ratio = 1,
+	.writeback_min_vscl_ratio = 1,
+	.writeback_max_hscl_taps = 12,
+	.writeback_max_vscl_taps = 12,
+	.writeback_line_buffer_luma_buffer_size = 0,
+	.writeback_line_buffer_chroma_buffer_size = 14643,
+	.cursor_buffer_size = 8,
+	.cursor_chunk_size = 2,
+	.max_num_otg = 4,
+	.max_num_dpp = 4,
+	.max_num_wb = 1,
+	.max_dchub_pscl_bw_pix_per_clk = 4,
+	.max_pscl_lb_bw_pix_per_clk = 2,
+	.max_lb_vscl_bw_pix_per_clk = 4,
+	.max_vscl_hscl_bw_pix_per_clk = 4,
+	.max_hscl_ratio = 4,
+	.max_vscl_ratio = 4,
+	.hscl_mults = 4,
+	.vscl_mults = 4,
+	.max_hscl_taps = 8,
+	.max_vscl_taps = 8,
+	.dispclk_ramp_margin_percent = 1,
+	.underscan_factor = 1.10,
+	.min_vblank_lines = 32, //
+	.dppclk_delay_subtotal = 77, //
+	.dppclk_delay_scl_lb_only = 16,
+	.dppclk_delay_scl = 50,
+	.dppclk_delay_cnvc_formatter = 8,
+	.dppclk_delay_cnvc_cursor = 6,
+	.dispclk_delay_subtotal = 87, //
+	.dcfclk_cstate_latency = 10, // SRExitTime
+	.max_inter_dcn_tile_repeaters = 8,
+
+	.xfc_supported = false,
+	.xfc_fill_bw_overhead_percent = 10.0,
+	.xfc_fill_constant_bytes = 0,
+	.ptoi_supported = 0,
+	.number_of_cursors = 1,
+};
+
+struct _vcs_dpi_soc_bounding_box_st dcn2_1_soc = {
+	.clock_limits = {
+			{
+				.state = 0,
+				.dcfclk_mhz = 400.0,
+				.fabricclk_mhz = 400.0,
+				.dispclk_mhz = 600.0,
+				.dppclk_mhz = 400.00,
+				.phyclk_mhz = 600.0,
+				.socclk_mhz = 278.0,
+				.dscclk_mhz = 205.67,
+				.dram_speed_mts = 1600.0,
+			},
+			{
+				.state = 1,
+				.dcfclk_mhz = 464.52,
+				.fabricclk_mhz = 800.0,
+				.dispclk_mhz = 654.55,
+				.dppclk_mhz = 626.09,
+				.phyclk_mhz = 600.0,
+				.socclk_mhz = 278.0,
+				.dscclk_mhz = 205.67,
+				.dram_speed_mts = 1600.0,
+			},
+			{
+				.state = 2,
+				.dcfclk_mhz = 514.29,
+				.fabricclk_mhz = 933.0,
+				.dispclk_mhz = 757.89,
+				.dppclk_mhz = 685.71,
+				.phyclk_mhz = 600.0,
+				.socclk_mhz = 278.0,
+				.dscclk_mhz = 287.67,
+				.dram_speed_mts = 1866.0,
+			},
+			{
+				.state = 3,
+				.dcfclk_mhz = 576.00,
+				.fabricclk_mhz = 1067.0,
+				.dispclk_mhz = 847.06,
+				.dppclk_mhz = 757.89,
+				.phyclk_mhz = 600.0,
+				.socclk_mhz = 715.0,
+				.dscclk_mhz = 318.334,
+				.dram_speed_mts = 2134.0,
+			},
+			{
+				.state = 4,
+				.dcfclk_mhz = 626.09,
+				.fabricclk_mhz = 1200.0,
+				.dispclk_mhz = 900.00,
+				.dppclk_mhz = 847.06,
+				.phyclk_mhz = 810.0,
+				.socclk_mhz = 953.0,
+				.dscclk_mhz = 300.0,
+				.dram_speed_mts = 2400.0,
+			},
+			{
+				.state = 5,
+				.dcfclk_mhz = 685.71,
+				.fabricclk_mhz = 1333.0,
+				.dispclk_mhz = 1028.57,
+				.dppclk_mhz = 960.00,
+				.phyclk_mhz = 810.0,
+				.socclk_mhz = 278.0,
+				.dscclk_mhz = 342.86,
+				.dram_speed_mts = 2666.0,
+			},
+			{
+				.state = 6,
+				.dcfclk_mhz = 757.89,
+				.fabricclk_mhz = 1467.0,
+				.dispclk_mhz = 1107.69,
+				.dppclk_mhz = 1028.57,
+				.phyclk_mhz = 810.0,
+				.socclk_mhz = 715.0,
+				.dscclk_mhz = 369.23,
+				.dram_speed_mts = 3200.0,
+			},
+			{
+				.state = 7,
+				.dcfclk_mhz = 847.06,
+				.fabricclk_mhz = 1600.0,
+				.dispclk_mhz = 1395.0,
+				.dppclk_mhz = 1285.00,
+				.phyclk_mhz = 1325.0,
+				.socclk_mhz = 953.0,
+				.dscclk_mhz = 489.0,
+				.dram_speed_mts = 4266.0,
+			},
+			/*Extra state, no dispclk ramping*/
+			{
+				.state = 8,
+				.dcfclk_mhz = 847.06,
+				.fabricclk_mhz = 1600.0,
+				.dispclk_mhz = 1395.0,
+				.dppclk_mhz = 1285.0,
+				.phyclk_mhz = 1325.0,
+				.socclk_mhz = 953.0,
+				.dscclk_mhz = 489.0,
+				.dram_speed_mts = 4266.0,
+			},
+
+		},
+
+	.sr_exit_time_us = 12.5,
+	.sr_enter_plus_exit_time_us = 17.0,
+	.urgent_latency_us = 4.0,
+	.urgent_latency_pixel_data_only_us = 4.0,
+	.urgent_latency_pixel_mixed_with_vm_data_us = 4.0,
+	.urgent_latency_vm_data_only_us = 4.0,
+	.urgent_out_of_order_return_per_channel_pixel_only_bytes = 4096,
+	.urgent_out_of_order_return_per_channel_pixel_and_vm_bytes = 4096,
+	.urgent_out_of_order_return_per_channel_vm_only_bytes = 4096,
+	.pct_ideal_dram_sdp_bw_after_urgent_pixel_only = 80.0,
+	.pct_ideal_dram_sdp_bw_after_urgent_pixel_and_vm = 75.0,
+	.pct_ideal_dram_sdp_bw_after_urgent_vm_only = 40.0,
+	.max_avg_sdp_bw_use_normal_percent = 60.0,
+	.max_avg_dram_bw_use_normal_percent = 100.0,
+	.writeback_latency_us = 12.0,
+	.max_request_size_bytes = 256,
+	.dram_channel_width_bytes = 4,
+	.fabric_datapath_to_dcn_data_return_bytes = 32,
+	.dcn_downspread_percent = 0.5,
+	.downspread_percent = 0.38,
+	.dram_page_open_time_ns = 50.0,
+	.dram_rw_turnaround_time_ns = 17.5,
+	.dram_return_buffer_per_channel_bytes = 8192,
+	.round_trip_ping_latency_dcfclk_cycles = 128,
+	.urgent_out_of_order_return_per_channel_bytes = 4096,
+	.channel_interleave_bytes = 256,
+	.num_banks = 8,
+	.num_chans = 4,
+	.vmm_page_size_bytes = 4096,
+	.dram_clock_change_latency_us = 23.84,
+	.return_bus_width_bytes = 64,
+	.dispclk_dppclk_vco_speed_mhz = 3600,
+	.xfc_bus_transport_time_us = 4,
+	.xfc_xbuf_latency_tolerance_us = 4,
+	.use_urgent_burst_bw = 1,
+	.num_states = 8
+};
+
+struct wm_table ddr4_wm_table_gs = {
+	.entries = {
+		{
+			.wm_inst = WM_A,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.72,
+			.sr_exit_time_us = 7.09,
+			.sr_enter_plus_exit_time_us = 8.14,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_B,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.72,
+			.sr_exit_time_us = 10.12,
+			.sr_enter_plus_exit_time_us = 11.48,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_C,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.72,
+			.sr_exit_time_us = 10.12,
+			.sr_enter_plus_exit_time_us = 11.48,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_D,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.72,
+			.sr_exit_time_us = 10.12,
+			.sr_enter_plus_exit_time_us = 11.48,
+			.valid = true,
+		},
+	}
+};
+
+struct wm_table lpddr4_wm_table_gs = {
+	.entries = {
+		{
+			.wm_inst = WM_A,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.65333,
+			.sr_exit_time_us = 5.32,
+			.sr_enter_plus_exit_time_us = 6.38,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_B,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.65333,
+			.sr_exit_time_us = 9.82,
+			.sr_enter_plus_exit_time_us = 11.196,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_C,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.65333,
+			.sr_exit_time_us = 9.89,
+			.sr_enter_plus_exit_time_us = 11.24,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_D,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.65333,
+			.sr_exit_time_us = 9.748,
+			.sr_enter_plus_exit_time_us = 11.102,
+			.valid = true,
+		},
+	}
+};
+
+struct wm_table lpddr4_wm_table_with_disabled_ppt = {
+	.entries = {
+		{
+			.wm_inst = WM_A,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.65333,
+			.sr_exit_time_us = 8.32,
+			.sr_enter_plus_exit_time_us = 9.38,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_B,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.65333,
+			.sr_exit_time_us = 9.82,
+			.sr_enter_plus_exit_time_us = 11.196,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_C,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.65333,
+			.sr_exit_time_us = 9.89,
+			.sr_enter_plus_exit_time_us = 11.24,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_D,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.65333,
+			.sr_exit_time_us = 9.748,
+			.sr_enter_plus_exit_time_us = 11.102,
+			.valid = true,
+		},
+	}
+};
+
+struct wm_table ddr4_wm_table_rn = {
+	.entries = {
+		{
+			.wm_inst = WM_A,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.72,
+			.sr_exit_time_us = 11.90,
+			.sr_enter_plus_exit_time_us = 12.80,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_B,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.72,
+			.sr_exit_time_us = 13.18,
+			.sr_enter_plus_exit_time_us = 14.30,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_C,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.72,
+			.sr_exit_time_us = 13.18,
+			.sr_enter_plus_exit_time_us = 14.30,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_D,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.72,
+			.sr_exit_time_us = 13.18,
+			.sr_enter_plus_exit_time_us = 14.30,
+			.valid = true,
+		},
+	}
+};
+
+struct wm_table ddr4_1R_wm_table_rn = {
+	.entries = {
+		{
+			.wm_inst = WM_A,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.72,
+			.sr_exit_time_us = 13.90,
+			.sr_enter_plus_exit_time_us = 14.80,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_B,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.72,
+			.sr_exit_time_us = 13.90,
+			.sr_enter_plus_exit_time_us = 14.80,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_C,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.72,
+			.sr_exit_time_us = 13.90,
+			.sr_enter_plus_exit_time_us = 14.80,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_D,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.72,
+			.sr_exit_time_us = 13.90,
+			.sr_enter_plus_exit_time_us = 14.80,
+			.valid = true,
+		},
+	}
+};
+
+struct wm_table lpddr4_wm_table_rn = {
+	.entries = {
+		{
+			.wm_inst = WM_A,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.65333,
+			.sr_exit_time_us = 7.32,
+			.sr_enter_plus_exit_time_us = 8.38,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_B,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.65333,
+			.sr_exit_time_us = 9.82,
+			.sr_enter_plus_exit_time_us = 11.196,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_C,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.65333,
+			.sr_exit_time_us = 9.89,
+			.sr_enter_plus_exit_time_us = 11.24,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_D,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.65333,
+			.sr_exit_time_us = 9.748,
+			.sr_enter_plus_exit_time_us = 11.102,
+			.valid = true,
+		},
+	}
+};
+
 void dcn20_populate_dml_writeback_from_context(struct dc *dc,
 					       struct resource_context *res_ctx,
 					       display_e2e_pipe_params_st *pipes)
@@ -100,3 +1025,1535 @@ void dcn20_populate_dml_writeback_from_context(struct dc *dc,
 		pipe_cnt++;
 	}
 }
+
+void dcn20_fpu_set_wb_arb_params(struct mcif_arb_params *wb_arb_params,
+				 struct dc_state *context,
+				 display_e2e_pipe_params_st *pipes,
+				 int pipe_cnt, int i)
+{
+	int k;
+
+	dc_assert_fp_enabled();
+
+	for (k = 0; k < sizeof(wb_arb_params->cli_watermark)/sizeof(wb_arb_params->cli_watermark[0]); k++) {
+		wb_arb_params->cli_watermark[k] = get_wm_writeback_urgent(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+		wb_arb_params->pstate_watermark[k] = get_wm_writeback_dram_clock_change(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	}
+	wb_arb_params->time_per_pixel = 16.0 * 1000 / (context->res_ctx.pipe_ctx[i].stream->phy_pix_clk / 1000); /* 4 bit fraction, ms */
+}
+
+static bool is_dtbclk_required(struct dc *dc, struct dc_state *context)
+{
+	int i;
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		if (!context->res_ctx.pipe_ctx[i].stream)
+			continue;
+		if (dc->link_srv->dp_is_128b_132b_signal(&context->res_ctx.pipe_ctx[i]))
+			return true;
+	}
+	return false;
+}
+
+static enum dcn_zstate_support_state  decide_zstate_support(struct dc *dc, struct dc_state *context)
+{
+	int plane_count;
+	int i;
+
+	plane_count = 0;
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		if (context->res_ctx.pipe_ctx[i].plane_state)
+			plane_count++;
+	}
+
+	/*
+	 * Z9 and Z10 allowed cases:
+	 * 	1. 0 Planes enabled
+	 * 	2. single eDP, on link 0, 1 plane and stutter period > 5ms
+	 * Z10 only cases:
+	 * 	1. single eDP, on link 0, 1 plane and stutter period >= 5ms
+	 * Z8 cases:
+	 * 	1. stutter period sufficient
+	 * Zstate not allowed cases:
+	 * 	1. Everything else
+	 */
+	if (plane_count == 0)
+		return DCN_ZSTATE_SUPPORT_ALLOW;
+	else if (context->stream_count == 1 &&  context->streams[0]->signal == SIGNAL_TYPE_EDP) {
+		struct dc_link *link = context->streams[0]->sink->link;
+		struct dc_stream_status *stream_status = &context->stream_status[0];
+		int minmum_z8_residency = dc->debug.minimum_z8_residency_time > 0 ? dc->debug.minimum_z8_residency_time : 1000;
+		bool allow_z8 = context->bw_ctx.dml.vba.StutterPeriod > (double)minmum_z8_residency;
+		bool is_pwrseq0 = (link && link->link_index == 0);
+		bool is_psr = (link && (link->psr_settings.psr_version == DC_PSR_VERSION_1 ||
+						link->psr_settings.psr_version == DC_PSR_VERSION_SU_1) && !link->panel_config.psr.disable_psr);
+		bool is_replay = link && link->replay_settings.replay_feature_enabled;
+
+		/* Don't support multi-plane configurations */
+		if (stream_status->plane_count > 1)
+			return DCN_ZSTATE_SUPPORT_DISALLOW;
+
+		if (is_pwrseq0 && context->bw_ctx.dml.vba.StutterPeriod > 5000.0)
+			return DCN_ZSTATE_SUPPORT_ALLOW;
+		else if (is_pwrseq0 && (is_psr || is_replay))
+			return DCN_ZSTATE_SUPPORT_ALLOW_Z8_Z10_ONLY;
+		else
+			return allow_z8 ? DCN_ZSTATE_SUPPORT_ALLOW_Z8_ONLY : DCN_ZSTATE_SUPPORT_DISALLOW;
+	} else {
+		return DCN_ZSTATE_SUPPORT_DISALLOW;
+	}
+}
+
+static void dcn20_adjust_freesync_v_startup(
+		const struct dc_crtc_timing *dc_crtc_timing, int *vstartup_start)
+{
+	struct dc_crtc_timing patched_crtc_timing;
+	uint32_t asic_blank_end   = 0;
+	uint32_t asic_blank_start = 0;
+	uint32_t newVstartup	  = 0;
+
+	patched_crtc_timing = *dc_crtc_timing;
+
+	if (patched_crtc_timing.flags.INTERLACE == 1) {
+		if (patched_crtc_timing.v_front_porch < 2)
+			patched_crtc_timing.v_front_porch = 2;
+	} else {
+		if (patched_crtc_timing.v_front_porch < 1)
+			patched_crtc_timing.v_front_porch = 1;
+	}
+
+	/* blank_start = frame end - front porch */
+	asic_blank_start = patched_crtc_timing.v_total -
+					patched_crtc_timing.v_front_porch;
+
+	/* blank_end = blank_start - active */
+	asic_blank_end = asic_blank_start -
+					patched_crtc_timing.v_border_bottom -
+					patched_crtc_timing.v_addressable -
+					patched_crtc_timing.v_border_top;
+
+	/* The newVStartUp is 1 line before vsync point */
+	newVstartup = asic_blank_end + 1;
+
+	*vstartup_start = ((newVstartup > *vstartup_start) ? newVstartup : *vstartup_start);
+}
+
+void dcn20_calculate_dlg_params(struct dc *dc,
+				struct dc_state *context,
+				display_e2e_pipe_params_st *pipes,
+				int pipe_cnt,
+				int vlevel)
+{
+	int i, pipe_idx, active_hubp_count = 0;
+
+	dc_assert_fp_enabled();
+
+	/* Writeback MCIF_WB arbitration parameters */
+	dc->res_pool->funcs->set_mcif_arb_params(dc, context, pipes, pipe_cnt);
+
+	context->bw_ctx.bw.dcn.clk.dispclk_khz = context->bw_ctx.dml.vba.DISPCLK * 1000;
+	context->bw_ctx.bw.dcn.clk.dcfclk_khz = context->bw_ctx.dml.vba.DCFCLK * 1000;
+	context->bw_ctx.bw.dcn.clk.socclk_khz = context->bw_ctx.dml.vba.SOCCLK * 1000;
+	context->bw_ctx.bw.dcn.clk.dramclk_khz = context->bw_ctx.dml.vba.DRAMSpeed * 1000 / 16;
+
+	if (dc->debug.min_dram_clk_khz > context->bw_ctx.bw.dcn.clk.dramclk_khz)
+		context->bw_ctx.bw.dcn.clk.dramclk_khz = dc->debug.min_dram_clk_khz;
+
+	context->bw_ctx.bw.dcn.clk.dcfclk_deep_sleep_khz = context->bw_ctx.dml.vba.DCFCLKDeepSleep * 1000;
+	context->bw_ctx.bw.dcn.clk.fclk_khz = context->bw_ctx.dml.vba.FabricClock * 1000;
+	context->bw_ctx.bw.dcn.clk.p_state_change_support =
+		context->bw_ctx.dml.vba.DRAMClockChangeSupport[vlevel][context->bw_ctx.dml.vba.maxMpcComb]
+							!= dm_dram_clock_change_unsupported;
+
+	/* Pstate change might not be supported by hardware, but it might be
+	 * possible with firmware driven vertical blank stretching.
+	 */
+	context->bw_ctx.bw.dcn.clk.p_state_change_support |= context->bw_ctx.bw.dcn.clk.fw_based_mclk_switching;
+
+	context->bw_ctx.bw.dcn.clk.dppclk_khz = 0;
+
+	context->bw_ctx.bw.dcn.clk.dtbclk_en = is_dtbclk_required(dc, context);
+
+	if (context->bw_ctx.bw.dcn.clk.dispclk_khz < dc->debug.min_disp_clk_khz)
+		context->bw_ctx.bw.dcn.clk.dispclk_khz = dc->debug.min_disp_clk_khz;
+
+	for (i = 0, pipe_idx = 0; i < dc->res_pool->pipe_count; i++) {
+		if (!context->res_ctx.pipe_ctx[i].stream)
+			continue;
+		if (context->res_ctx.pipe_ctx[i].plane_state)
+			active_hubp_count++;
+		pipes[pipe_idx].pipe.dest.vstartup_start = get_vstartup(&context->bw_ctx.dml, pipes, pipe_cnt, pipe_idx);
+		pipes[pipe_idx].pipe.dest.vupdate_offset = get_vupdate_offset(&context->bw_ctx.dml, pipes, pipe_cnt, pipe_idx);
+		pipes[pipe_idx].pipe.dest.vupdate_width = get_vupdate_width(&context->bw_ctx.dml, pipes, pipe_cnt, pipe_idx);
+		pipes[pipe_idx].pipe.dest.vready_offset = get_vready_offset(&context->bw_ctx.dml, pipes, pipe_cnt, pipe_idx);
+
+		if (dc_state_get_pipe_subvp_type(context, &context->res_ctx.pipe_ctx[i]) == SUBVP_PHANTOM) {
+			// Phantom pipe requires that DET_SIZE = 0 and no unbounded requests
+			context->res_ctx.pipe_ctx[i].det_buffer_size_kb = 0;
+			context->res_ctx.pipe_ctx[i].unbounded_req = false;
+		} else {
+			context->res_ctx.pipe_ctx[i].det_buffer_size_kb = context->bw_ctx.dml.ip.det_buffer_size_kbytes;
+			context->res_ctx.pipe_ctx[i].unbounded_req = pipes[pipe_idx].pipe.src.unbounded_req_mode;
+		}
+
+		if (context->bw_ctx.bw.dcn.clk.dppclk_khz < pipes[pipe_idx].clks_cfg.dppclk_mhz * 1000)
+			context->bw_ctx.bw.dcn.clk.dppclk_khz = pipes[pipe_idx].clks_cfg.dppclk_mhz * 1000;
+		context->res_ctx.pipe_ctx[i].plane_res.bw.dppclk_khz =
+						pipes[pipe_idx].clks_cfg.dppclk_mhz * 1000;
+		context->res_ctx.pipe_ctx[i].pipe_dlg_param = pipes[pipe_idx].pipe.dest;
+		if (dc->ctx->dce_version < DCN_VERSION_3_1 &&
+		    context->res_ctx.pipe_ctx[i].stream->adaptive_sync_infopacket.valid)
+			dcn20_adjust_freesync_v_startup(
+				&context->res_ctx.pipe_ctx[i].stream->timing,
+				&context->res_ctx.pipe_ctx[i].pipe_dlg_param.vstartup_start);
+
+		pipe_idx++;
+	}
+	/* If DCN isn't making memory requests we can allow pstate change */
+	if (!active_hubp_count) {
+		context->bw_ctx.bw.dcn.clk.p_state_change_support = true;
+	}
+	/*save a original dppclock copy*/
+	context->bw_ctx.bw.dcn.clk.bw_dppclk_khz = context->bw_ctx.bw.dcn.clk.dppclk_khz;
+	context->bw_ctx.bw.dcn.clk.bw_dispclk_khz = context->bw_ctx.bw.dcn.clk.dispclk_khz;
+	context->bw_ctx.bw.dcn.clk.max_supported_dppclk_khz = context->bw_ctx.dml.soc.clock_limits[vlevel].dppclk_mhz * 1000;
+	context->bw_ctx.bw.dcn.clk.max_supported_dispclk_khz = context->bw_ctx.dml.soc.clock_limits[vlevel].dispclk_mhz * 1000;
+
+	context->bw_ctx.bw.dcn.compbuf_size_kb = context->bw_ctx.dml.ip.config_return_buffer_size_in_kbytes
+						- context->bw_ctx.dml.ip.det_buffer_size_kbytes * pipe_idx;
+
+	for (i = 0, pipe_idx = 0; i < dc->res_pool->pipe_count; i++) {
+		bool cstate_en = context->bw_ctx.dml.vba.PrefetchMode[vlevel][context->bw_ctx.dml.vba.maxMpcComb] != 2;
+
+		if (!context->res_ctx.pipe_ctx[i].stream)
+			continue;
+
+		/* cstate disabled on 201 */
+		if (dc->ctx->dce_version == DCN_VERSION_2_01)
+			cstate_en = false;
+
+		context->bw_ctx.dml.funcs.rq_dlg_get_dlg_reg(&context->bw_ctx.dml,
+				&context->res_ctx.pipe_ctx[i].dlg_regs,
+				&context->res_ctx.pipe_ctx[i].ttu_regs,
+				pipes,
+				pipe_cnt,
+				pipe_idx,
+				cstate_en,
+				context->bw_ctx.bw.dcn.clk.p_state_change_support,
+				false, false, true);
+
+		context->bw_ctx.dml.funcs.rq_dlg_get_rq_reg(&context->bw_ctx.dml,
+				&context->res_ctx.pipe_ctx[i].rq_regs,
+				&pipes[pipe_idx].pipe);
+		pipe_idx++;
+	}
+	context->bw_ctx.bw.dcn.clk.zstate_support = decide_zstate_support(dc, context);
+}
+
+static void swizzle_to_dml_params(
+		enum swizzle_mode_values swizzle,
+		unsigned int *sw_mode)
+{
+	switch (swizzle) {
+	case DC_SW_LINEAR:
+		*sw_mode = dm_sw_linear;
+		break;
+	case DC_SW_4KB_S:
+		*sw_mode = dm_sw_4kb_s;
+		break;
+	case DC_SW_4KB_S_X:
+		*sw_mode = dm_sw_4kb_s_x;
+		break;
+	case DC_SW_4KB_D:
+		*sw_mode = dm_sw_4kb_d;
+		break;
+	case DC_SW_4KB_D_X:
+		*sw_mode = dm_sw_4kb_d_x;
+		break;
+	case DC_SW_64KB_S:
+		*sw_mode = dm_sw_64kb_s;
+		break;
+	case DC_SW_64KB_S_X:
+		*sw_mode = dm_sw_64kb_s_x;
+		break;
+	case DC_SW_64KB_S_T:
+		*sw_mode = dm_sw_64kb_s_t;
+		break;
+	case DC_SW_64KB_D:
+		*sw_mode = dm_sw_64kb_d;
+		break;
+	case DC_SW_64KB_D_X:
+		*sw_mode = dm_sw_64kb_d_x;
+		break;
+	case DC_SW_64KB_D_T:
+		*sw_mode = dm_sw_64kb_d_t;
+		break;
+	case DC_SW_64KB_R_X:
+		*sw_mode = dm_sw_64kb_r_x;
+		break;
+	case DC_SW_VAR_S:
+		*sw_mode = dm_sw_var_s;
+		break;
+	case DC_SW_VAR_S_X:
+		*sw_mode = dm_sw_var_s_x;
+		break;
+	case DC_SW_VAR_D:
+		*sw_mode = dm_sw_var_d;
+		break;
+	case DC_SW_VAR_D_X:
+		*sw_mode = dm_sw_var_d_x;
+		break;
+	case DC_SW_VAR_R_X:
+		*sw_mode = dm_sw_var_r_x;
+		break;
+	default:
+		ASSERT(0); /* Not supported */
+		break;
+	}
+}
+
+int dcn20_populate_dml_pipes_from_context(struct dc *dc,
+					  struct dc_state *context,
+					  display_e2e_pipe_params_st *pipes,
+					  enum dc_validate_mode validate_mode)
+{
+	int pipe_cnt, i;
+	bool synchronized_vblank = true;
+	struct resource_context *res_ctx = &context->res_ctx;
+
+	dc_assert_fp_enabled();
+
+	for (i = 0, pipe_cnt = -1; i < dc->res_pool->pipe_count; i++) {
+		if (!res_ctx->pipe_ctx[i].stream)
+			continue;
+
+		if (pipe_cnt < 0) {
+			pipe_cnt = i;
+			continue;
+		}
+
+		if (res_ctx->pipe_ctx[pipe_cnt].stream == res_ctx->pipe_ctx[i].stream)
+			continue;
+
+		if (dc->debug.disable_timing_sync ||
+			(!resource_are_streams_timing_synchronizable(
+				res_ctx->pipe_ctx[pipe_cnt].stream,
+				res_ctx->pipe_ctx[i].stream) &&
+			!resource_are_vblanks_synchronizable(
+				res_ctx->pipe_ctx[pipe_cnt].stream,
+				res_ctx->pipe_ctx[i].stream))) {
+			synchronized_vblank = false;
+			break;
+		}
+	}
+
+	for (i = 0, pipe_cnt = 0; i < dc->res_pool->pipe_count; i++) {
+		struct dc_crtc_timing *timing = &res_ctx->pipe_ctx[i].stream->timing;
+		unsigned int v_total;
+		unsigned int front_porch;
+		int output_bpc;
+		struct audio_check aud_check = {0};
+
+		if (!res_ctx->pipe_ctx[i].stream)
+			continue;
+
+		v_total = timing->v_total;
+		front_porch = timing->v_front_porch;
+
+		/* todo:
+		pipes[pipe_cnt].pipe.src.dynamic_metadata_enable = 0;
+		pipes[pipe_cnt].pipe.src.dcc = 0;
+		pipes[pipe_cnt].pipe.src.vm = 0;*/
+
+		pipes[pipe_cnt].clks_cfg.refclk_mhz = dc->res_pool->ref_clocks.dchub_ref_clock_inKhz / 1000.0;
+
+		pipes[pipe_cnt].pipe.dest.use_maximum_vstartup = dc->ctx->dce_version == DCN_VERSION_2_01;
+
+		pipes[pipe_cnt].dout.dsc_enable = res_ctx->pipe_ctx[i].stream->timing.flags.DSC;
+		/* todo: rotation?*/
+		pipes[pipe_cnt].dout.dsc_slices = res_ctx->pipe_ctx[i].stream->timing.dsc_cfg.num_slices_h;
+		if (res_ctx->pipe_ctx[i].stream->use_dynamic_meta) {
+			pipes[pipe_cnt].pipe.src.dynamic_metadata_enable = true;
+			/* 1/2 vblank */
+			pipes[pipe_cnt].pipe.src.dynamic_metadata_lines_before_active =
+				(v_total - timing->v_addressable
+					- timing->v_border_top - timing->v_border_bottom) / 2;
+			/* 36 bytes dp, 32 hdmi */
+			pipes[pipe_cnt].pipe.src.dynamic_metadata_xmit_bytes =
+				dc_is_dp_signal(res_ctx->pipe_ctx[i].stream->signal) ? 36 : 32;
+		}
+		pipes[pipe_cnt].pipe.src.dcc = false;
+		pipes[pipe_cnt].pipe.src.dcc_rate = 1;
+		pipes[pipe_cnt].pipe.dest.synchronized_vblank_all_planes = synchronized_vblank;
+		pipes[pipe_cnt].pipe.dest.synchronize_timings = synchronized_vblank;
+		pipes[pipe_cnt].pipe.dest.hblank_start = timing->h_total - timing->h_front_porch;
+		pipes[pipe_cnt].pipe.dest.hblank_end = pipes[pipe_cnt].pipe.dest.hblank_start
+				- timing->h_addressable
+				- timing->h_border_left
+				- timing->h_border_right;
+		pipes[pipe_cnt].pipe.dest.vblank_start = v_total - front_porch;
+		pipes[pipe_cnt].pipe.dest.vblank_end = pipes[pipe_cnt].pipe.dest.vblank_start
+				- timing->v_addressable
+				- timing->v_border_top
+				- timing->v_border_bottom;
+		pipes[pipe_cnt].pipe.dest.htotal = timing->h_total;
+		pipes[pipe_cnt].pipe.dest.vtotal = v_total;
+		pipes[pipe_cnt].pipe.dest.hactive =
+			timing->h_addressable + timing->h_border_left + timing->h_border_right;
+		pipes[pipe_cnt].pipe.dest.vactive =
+			timing->v_addressable + timing->v_border_top + timing->v_border_bottom;
+		pipes[pipe_cnt].pipe.dest.interlaced = timing->flags.INTERLACE;
+		pipes[pipe_cnt].pipe.dest.pixel_rate_mhz = timing->pix_clk_100hz/10000.0;
+		if (timing->timing_3d_format == TIMING_3D_FORMAT_HW_FRAME_PACKING)
+			pipes[pipe_cnt].pipe.dest.pixel_rate_mhz *= 2;
+		pipes[pipe_cnt].pipe.dest.otg_inst = res_ctx->pipe_ctx[i].stream_res.tg->inst;
+		pipes[pipe_cnt].dout.dp_lanes = 4;
+		pipes[pipe_cnt].dout.dp_rate = dm_dp_rate_na;
+		pipes[pipe_cnt].dout.is_virtual = 0;
+		pipes[pipe_cnt].pipe.dest.vtotal_min = res_ctx->pipe_ctx[i].stream->adjust.v_total_min;
+		pipes[pipe_cnt].pipe.dest.vtotal_max = res_ctx->pipe_ctx[i].stream->adjust.v_total_max;
+		switch (resource_get_odm_slice_count(&res_ctx->pipe_ctx[i])) {
+		case 2:
+			pipes[pipe_cnt].pipe.dest.odm_combine = dm_odm_combine_mode_2to1;
+			break;
+		case 4:
+			pipes[pipe_cnt].pipe.dest.odm_combine = dm_odm_combine_mode_4to1;
+			break;
+		default:
+			pipes[pipe_cnt].pipe.dest.odm_combine = dm_odm_combine_mode_disabled;
+		}
+		pipes[pipe_cnt].pipe.src.hsplit_grp = res_ctx->pipe_ctx[i].pipe_idx;
+		if (res_ctx->pipe_ctx[i].top_pipe && res_ctx->pipe_ctx[i].top_pipe->plane_state
+				== res_ctx->pipe_ctx[i].plane_state) {
+			struct pipe_ctx *first_pipe = res_ctx->pipe_ctx[i].top_pipe;
+			int split_idx = 0;
+
+			while (first_pipe->top_pipe && first_pipe->top_pipe->plane_state
+					== res_ctx->pipe_ctx[i].plane_state) {
+				first_pipe = first_pipe->top_pipe;
+				split_idx++;
+			}
+			/* Treat 4to1 mpc combine as an mpo of 2 2-to-1 combines */
+			if (split_idx == 0)
+				pipes[pipe_cnt].pipe.src.hsplit_grp = first_pipe->pipe_idx;
+			else if (split_idx == 1)
+				pipes[pipe_cnt].pipe.src.hsplit_grp = res_ctx->pipe_ctx[i].pipe_idx;
+			else if (split_idx == 2)
+				pipes[pipe_cnt].pipe.src.hsplit_grp = res_ctx->pipe_ctx[i].top_pipe->pipe_idx;
+		} else if (res_ctx->pipe_ctx[i].prev_odm_pipe) {
+			struct pipe_ctx *first_pipe = res_ctx->pipe_ctx[i].prev_odm_pipe;
+
+			while (first_pipe->prev_odm_pipe)
+				first_pipe = first_pipe->prev_odm_pipe;
+			pipes[pipe_cnt].pipe.src.hsplit_grp = first_pipe->pipe_idx;
+		}
+
+		switch (res_ctx->pipe_ctx[i].stream->signal) {
+		case SIGNAL_TYPE_DISPLAY_PORT_MST:
+		case SIGNAL_TYPE_DISPLAY_PORT:
+			pipes[pipe_cnt].dout.output_type = dm_dp;
+			if (dc->link_srv->dp_is_128b_132b_signal(&res_ctx->pipe_ctx[i]))
+				pipes[pipe_cnt].dout.output_type = dm_dp2p0;
+			break;
+		case SIGNAL_TYPE_EDP:
+			pipes[pipe_cnt].dout.output_type = dm_edp;
+			break;
+		case SIGNAL_TYPE_HDMI_TYPE_A:
+		case SIGNAL_TYPE_DVI_SINGLE_LINK:
+		case SIGNAL_TYPE_DVI_DUAL_LINK:
+			pipes[pipe_cnt].dout.output_type = dm_hdmi;
+			break;
+		default:
+			/* In case there is no signal, set dp with 4 lanes to allow max config */
+			pipes[pipe_cnt].dout.is_virtual = 1;
+			pipes[pipe_cnt].dout.output_type = dm_dp;
+			pipes[pipe_cnt].dout.dp_lanes = 4;
+		}
+
+		switch (res_ctx->pipe_ctx[i].stream->timing.display_color_depth) {
+		case COLOR_DEPTH_666:
+			output_bpc = 6;
+			break;
+		case COLOR_DEPTH_888:
+			output_bpc = 8;
+			break;
+		case COLOR_DEPTH_101010:
+			output_bpc = 10;
+			break;
+		case COLOR_DEPTH_121212:
+			output_bpc = 12;
+			break;
+		case COLOR_DEPTH_141414:
+			output_bpc = 14;
+			break;
+		case COLOR_DEPTH_161616:
+			output_bpc = 16;
+			break;
+		case COLOR_DEPTH_999:
+			output_bpc = 9;
+			break;
+		case COLOR_DEPTH_111111:
+			output_bpc = 11;
+			break;
+		default:
+			output_bpc = 8;
+			break;
+		}
+
+		switch (res_ctx->pipe_ctx[i].stream->timing.pixel_encoding) {
+		case PIXEL_ENCODING_RGB:
+		case PIXEL_ENCODING_YCBCR444:
+			pipes[pipe_cnt].dout.output_format = dm_444;
+			pipes[pipe_cnt].dout.output_bpp = output_bpc * 3;
+			break;
+		case PIXEL_ENCODING_YCBCR420:
+			pipes[pipe_cnt].dout.output_format = dm_420;
+			pipes[pipe_cnt].dout.output_bpp = (output_bpc * 3.0) / 2;
+			break;
+		case PIXEL_ENCODING_YCBCR422:
+			if (res_ctx->pipe_ctx[i].stream->timing.flags.DSC &&
+			    !res_ctx->pipe_ctx[i].stream->timing.dsc_cfg.ycbcr422_simple)
+				pipes[pipe_cnt].dout.output_format = dm_n422;
+			else
+				pipes[pipe_cnt].dout.output_format = dm_s422;
+			pipes[pipe_cnt].dout.output_bpp = output_bpc * 2;
+			break;
+		default:
+			pipes[pipe_cnt].dout.output_format = dm_444;
+			pipes[pipe_cnt].dout.output_bpp = output_bpc * 3;
+		}
+
+		if (res_ctx->pipe_ctx[i].stream->timing.flags.DSC)
+			pipes[pipe_cnt].dout.output_bpp = res_ctx->pipe_ctx[i].stream->timing.dsc_cfg.bits_per_pixel / 16.0;
+
+		/* todo: default max for now, until there is logic reflecting this in dc*/
+		pipes[pipe_cnt].dout.dsc_input_bpc = 12;
+		/*fill up the audio sample rate (unit in kHz)*/
+		get_audio_check(&res_ctx->pipe_ctx[i].stream->audio_info, &aud_check);
+		pipes[pipe_cnt].dout.max_audio_sample_rate = aud_check.max_audiosample_rate / 1000;
+		/*
+		 * For graphic plane, cursor number is 1, nv12 is 0
+		 * bw calculations due to cursor on/off
+		 */
+		if (res_ctx->pipe_ctx[i].plane_state &&
+				(res_ctx->pipe_ctx[i].plane_state->address.type == PLN_ADDR_TYPE_VIDEO_PROGRESSIVE ||
+				dc_state_get_pipe_subvp_type(context, &res_ctx->pipe_ctx[i]) == SUBVP_PHANTOM))
+			pipes[pipe_cnt].pipe.src.num_cursors = 0;
+		else
+			pipes[pipe_cnt].pipe.src.num_cursors = dc->dml.ip.number_of_cursors;
+
+		pipes[pipe_cnt].pipe.src.cur0_src_width = 256;
+		pipes[pipe_cnt].pipe.src.cur0_bpp = dm_cur_32bit;
+
+		if (!res_ctx->pipe_ctx[i].plane_state) {
+			pipes[pipe_cnt].pipe.src.is_hsplit = pipes[pipe_cnt].pipe.dest.odm_combine != dm_odm_combine_mode_disabled;
+			pipes[pipe_cnt].pipe.src.source_scan = dm_horz;
+			pipes[pipe_cnt].pipe.src.source_rotation = dm_rotation_0;
+			pipes[pipe_cnt].pipe.src.sw_mode = dm_sw_4kb_s;
+			pipes[pipe_cnt].pipe.src.macro_tile_size = dm_64k_tile;
+			pipes[pipe_cnt].pipe.src.viewport_width = timing->h_addressable;
+			if (pipes[pipe_cnt].pipe.src.viewport_width > 1920)
+				pipes[pipe_cnt].pipe.src.viewport_width = 1920;
+			pipes[pipe_cnt].pipe.src.viewport_height = timing->v_addressable;
+			if (pipes[pipe_cnt].pipe.src.viewport_height > 1080)
+				pipes[pipe_cnt].pipe.src.viewport_height = 1080;
+			pipes[pipe_cnt].pipe.src.surface_height_y = pipes[pipe_cnt].pipe.src.viewport_height;
+			pipes[pipe_cnt].pipe.src.surface_width_y = pipes[pipe_cnt].pipe.src.viewport_width;
+			pipes[pipe_cnt].pipe.src.surface_height_c = pipes[pipe_cnt].pipe.src.viewport_height;
+			pipes[pipe_cnt].pipe.src.surface_width_c = pipes[pipe_cnt].pipe.src.viewport_width;
+			pipes[pipe_cnt].pipe.src.data_pitch = ((pipes[pipe_cnt].pipe.src.viewport_width + 255) / 256) * 256;
+			pipes[pipe_cnt].pipe.src.source_format = dm_444_32;
+			pipes[pipe_cnt].pipe.src.cur0_src_width = 0;
+			pipes[pipe_cnt].pipe.src.cur1_src_width = 0;
+			pipes[pipe_cnt].pipe.dest.recout_width = pipes[pipe_cnt].pipe.src.viewport_width; /*vp_width/hratio*/
+			pipes[pipe_cnt].pipe.dest.recout_height = pipes[pipe_cnt].pipe.src.viewport_height; /*vp_height/vratio*/
+			pipes[pipe_cnt].pipe.dest.full_recout_width = pipes[pipe_cnt].pipe.dest.recout_width;  /*when is_hsplit != 1*/
+			pipes[pipe_cnt].pipe.dest.full_recout_height = pipes[pipe_cnt].pipe.dest.recout_height; /*when is_hsplit != 1*/
+			pipes[pipe_cnt].pipe.scale_ratio_depth.lb_depth = dm_lb_16;
+			pipes[pipe_cnt].pipe.scale_ratio_depth.hscl_ratio = 1.0;
+			pipes[pipe_cnt].pipe.scale_ratio_depth.vscl_ratio = 1.0;
+			pipes[pipe_cnt].pipe.scale_ratio_depth.scl_enable = 0; /*Lb only or Full scl*/
+			pipes[pipe_cnt].pipe.scale_taps.htaps = 1;
+			pipes[pipe_cnt].pipe.scale_taps.vtaps = 1;
+			pipes[pipe_cnt].pipe.dest.vtotal_min = v_total;
+			pipes[pipe_cnt].pipe.dest.vtotal_max = v_total;
+
+			if (pipes[pipe_cnt].pipe.dest.odm_combine == dm_odm_combine_mode_2to1) {
+				pipes[pipe_cnt].pipe.src.viewport_width /= 2;
+				pipes[pipe_cnt].pipe.dest.recout_width /= 2;
+			} else if (pipes[pipe_cnt].pipe.dest.odm_combine == dm_odm_combine_mode_4to1) {
+				pipes[pipe_cnt].pipe.src.viewport_width /= 4;
+				pipes[pipe_cnt].pipe.dest.recout_width /= 4;
+			}
+		} else {
+			struct dc_plane_state *pln = res_ctx->pipe_ctx[i].plane_state;
+			struct scaler_data *scl = &res_ctx->pipe_ctx[i].plane_res.scl_data;
+
+			pipes[pipe_cnt].pipe.src.immediate_flip = pln->flip_immediate;
+			pipes[pipe_cnt].pipe.src.is_hsplit = (res_ctx->pipe_ctx[i].bottom_pipe && res_ctx->pipe_ctx[i].bottom_pipe->plane_state == pln)
+					|| (res_ctx->pipe_ctx[i].top_pipe && res_ctx->pipe_ctx[i].top_pipe->plane_state == pln)
+					|| pipes[pipe_cnt].pipe.dest.odm_combine != dm_odm_combine_mode_disabled;
+
+			/* stereo is not split */
+			if (pln->stereo_format == PLANE_STEREO_FORMAT_SIDE_BY_SIDE ||
+			    pln->stereo_format == PLANE_STEREO_FORMAT_TOP_AND_BOTTOM) {
+				pipes[pipe_cnt].pipe.src.is_hsplit = false;
+				pipes[pipe_cnt].pipe.src.hsplit_grp = res_ctx->pipe_ctx[i].pipe_idx;
+			}
+
+			pipes[pipe_cnt].pipe.src.source_scan = pln->rotation == ROTATION_ANGLE_90
+					|| pln->rotation == ROTATION_ANGLE_270 ? dm_vert : dm_horz;
+			switch (pln->rotation) {
+			case ROTATION_ANGLE_0:
+				pipes[pipe_cnt].pipe.src.source_rotation = dm_rotation_0;
+				break;
+			case ROTATION_ANGLE_90:
+				pipes[pipe_cnt].pipe.src.source_rotation = dm_rotation_90;
+				break;
+			case ROTATION_ANGLE_180:
+				pipes[pipe_cnt].pipe.src.source_rotation = dm_rotation_180;
+				break;
+			case ROTATION_ANGLE_270:
+				pipes[pipe_cnt].pipe.src.source_rotation = dm_rotation_270;
+				break;
+			default:
+				break;
+			}
+
+			pipes[pipe_cnt].pipe.src.viewport_y_y = scl->viewport.y;
+			pipes[pipe_cnt].pipe.src.viewport_y_c = scl->viewport_c.y;
+			pipes[pipe_cnt].pipe.src.viewport_x_y = scl->viewport.x;
+			pipes[pipe_cnt].pipe.src.viewport_x_c = scl->viewport_c.x;
+			pipes[pipe_cnt].pipe.src.viewport_width = scl->viewport.width;
+			pipes[pipe_cnt].pipe.src.viewport_width_c = scl->viewport_c.width;
+			pipes[pipe_cnt].pipe.src.viewport_height = scl->viewport.height;
+			pipes[pipe_cnt].pipe.src.viewport_height_c = scl->viewport_c.height;
+			pipes[pipe_cnt].pipe.src.viewport_width_max = pln->src_rect.width;
+			pipes[pipe_cnt].pipe.src.viewport_height_max = pln->src_rect.height;
+			pipes[pipe_cnt].pipe.src.surface_width_y = pln->plane_size.surface_size.width;
+			pipes[pipe_cnt].pipe.src.surface_height_y = pln->plane_size.surface_size.height;
+			pipes[pipe_cnt].pipe.src.surface_width_c = pln->plane_size.chroma_size.width;
+			pipes[pipe_cnt].pipe.src.surface_height_c = pln->plane_size.chroma_size.height;
+			if (pln->format == SURFACE_PIXEL_FORMAT_GRPH_RGBE_ALPHA
+					|| pln->format >= SURFACE_PIXEL_FORMAT_VIDEO_BEGIN) {
+				pipes[pipe_cnt].pipe.src.data_pitch = pln->plane_size.surface_pitch;
+				pipes[pipe_cnt].pipe.src.data_pitch_c = pln->plane_size.chroma_pitch;
+				pipes[pipe_cnt].pipe.src.meta_pitch = pln->dcc.meta_pitch;
+				pipes[pipe_cnt].pipe.src.meta_pitch_c = pln->dcc.meta_pitch_c;
+			} else {
+				pipes[pipe_cnt].pipe.src.data_pitch = pln->plane_size.surface_pitch;
+				pipes[pipe_cnt].pipe.src.meta_pitch = pln->dcc.meta_pitch;
+			}
+			pipes[pipe_cnt].pipe.src.dcc = pln->dcc.enable;
+			pipes[pipe_cnt].pipe.dest.recout_width = scl->recout.width;
+			pipes[pipe_cnt].pipe.dest.recout_height = scl->recout.height;
+			pipes[pipe_cnt].pipe.dest.full_recout_height = scl->recout.height;
+			pipes[pipe_cnt].pipe.dest.full_recout_width = scl->recout.width;
+			if (pipes[pipe_cnt].pipe.dest.odm_combine == dm_odm_combine_mode_2to1)
+				pipes[pipe_cnt].pipe.dest.full_recout_width *= 2;
+			else if (pipes[pipe_cnt].pipe.dest.odm_combine == dm_odm_combine_mode_4to1)
+				pipes[pipe_cnt].pipe.dest.full_recout_width *= 4;
+			else {
+				struct pipe_ctx *split_pipe = res_ctx->pipe_ctx[i].bottom_pipe;
+
+				while (split_pipe && split_pipe->plane_state == pln) {
+					pipes[pipe_cnt].pipe.dest.full_recout_width += split_pipe->plane_res.scl_data.recout.width;
+					split_pipe = split_pipe->bottom_pipe;
+				}
+				split_pipe = res_ctx->pipe_ctx[i].top_pipe;
+				while (split_pipe && split_pipe->plane_state == pln) {
+					pipes[pipe_cnt].pipe.dest.full_recout_width += split_pipe->plane_res.scl_data.recout.width;
+					split_pipe = split_pipe->top_pipe;
+				}
+			}
+
+			pipes[pipe_cnt].pipe.scale_ratio_depth.lb_depth = dm_lb_16;
+			pipes[pipe_cnt].pipe.scale_ratio_depth.hscl_ratio = (double) scl->ratios.horz.value / (1ULL<<32);
+			pipes[pipe_cnt].pipe.scale_ratio_depth.hscl_ratio_c = (double) scl->ratios.horz_c.value / (1ULL<<32);
+			pipes[pipe_cnt].pipe.scale_ratio_depth.vscl_ratio = (double) scl->ratios.vert.value / (1ULL<<32);
+			pipes[pipe_cnt].pipe.scale_ratio_depth.vscl_ratio_c = (double) scl->ratios.vert_c.value / (1ULL<<32);
+			pipes[pipe_cnt].pipe.scale_ratio_depth.scl_enable =
+					scl->ratios.vert.value != dc_fixpt_one.value
+					|| scl->ratios.horz.value != dc_fixpt_one.value
+					|| scl->ratios.vert_c.value != dc_fixpt_one.value
+					|| scl->ratios.horz_c.value != dc_fixpt_one.value /*Lb only or Full scl*/
+					|| dc->debug.always_scale; /*support always scale*/
+			pipes[pipe_cnt].pipe.scale_taps.htaps = scl->taps.h_taps;
+			pipes[pipe_cnt].pipe.scale_taps.htaps_c = scl->taps.h_taps_c;
+			pipes[pipe_cnt].pipe.scale_taps.vtaps = scl->taps.v_taps;
+			pipes[pipe_cnt].pipe.scale_taps.vtaps_c = scl->taps.v_taps_c;
+
+			pipes[pipe_cnt].pipe.src.macro_tile_size =
+					swizzle_mode_to_macro_tile_size(pln->tiling_info.gfx9.swizzle);
+			swizzle_to_dml_params(pln->tiling_info.gfx9.swizzle,
+					&pipes[pipe_cnt].pipe.src.sw_mode);
+
+			switch (pln->format) {
+			case SURFACE_PIXEL_FORMAT_VIDEO_420_YCbCr:
+			case SURFACE_PIXEL_FORMAT_VIDEO_420_YCrCb:
+				pipes[pipe_cnt].pipe.src.source_format = dm_420_8;
+				break;
+			case SURFACE_PIXEL_FORMAT_VIDEO_420_10bpc_YCbCr:
+			case SURFACE_PIXEL_FORMAT_VIDEO_420_10bpc_YCrCb:
+				pipes[pipe_cnt].pipe.src.source_format = dm_420_10;
+				break;
+			case SURFACE_PIXEL_FORMAT_GRPH_ARGB16161616:
+			case SURFACE_PIXEL_FORMAT_GRPH_ABGR16161616:
+			case SURFACE_PIXEL_FORMAT_GRPH_ARGB16161616F:
+			case SURFACE_PIXEL_FORMAT_GRPH_ABGR16161616F:
+				pipes[pipe_cnt].pipe.src.source_format = dm_444_64;
+				break;
+			case SURFACE_PIXEL_FORMAT_GRPH_ARGB1555:
+			case SURFACE_PIXEL_FORMAT_GRPH_RGB565:
+				pipes[pipe_cnt].pipe.src.source_format = dm_444_16;
+				break;
+			case SURFACE_PIXEL_FORMAT_GRPH_PALETA_256_COLORS:
+				pipes[pipe_cnt].pipe.src.source_format = dm_444_8;
+				break;
+			case SURFACE_PIXEL_FORMAT_GRPH_RGBE_ALPHA:
+				pipes[pipe_cnt].pipe.src.source_format = dm_rgbe_alpha;
+				break;
+			default:
+				pipes[pipe_cnt].pipe.src.source_format = dm_444_32;
+				break;
+			}
+		}
+
+		pipe_cnt++;
+	}
+
+	/* populate writeback information */
+	dc->res_pool->funcs->populate_dml_writeback_from_context(dc, res_ctx, pipes);
+
+	return pipe_cnt;
+}
+
+void dcn20_calculate_wm(struct dc *dc, struct dc_state *context,
+			display_e2e_pipe_params_st *pipes,
+			int *out_pipe_cnt,
+			int *pipe_split_from,
+			int vlevel,
+			enum dc_validate_mode validate_mode)
+{
+	int pipe_cnt, i, pipe_idx;
+
+	dc_assert_fp_enabled();
+
+	for (i = 0, pipe_idx = 0, pipe_cnt = 0; i < dc->res_pool->pipe_count; i++) {
+		if (!context->res_ctx.pipe_ctx[i].stream)
+			continue;
+
+		pipes[pipe_cnt].clks_cfg.refclk_mhz = dc->res_pool->ref_clocks.dchub_ref_clock_inKhz / 1000.0;
+		pipes[pipe_cnt].clks_cfg.dispclk_mhz = context->bw_ctx.dml.vba.RequiredDISPCLK[vlevel][context->bw_ctx.dml.vba.maxMpcComb];
+
+		if (pipe_split_from[i] < 0) {
+			pipes[pipe_cnt].clks_cfg.dppclk_mhz =
+					context->bw_ctx.dml.vba.RequiredDPPCLK[vlevel][context->bw_ctx.dml.vba.maxMpcComb][pipe_idx];
+			if (context->bw_ctx.dml.vba.BlendingAndTiming[pipe_idx] == pipe_idx)
+				pipes[pipe_cnt].pipe.dest.odm_combine =
+						context->bw_ctx.dml.vba.ODMCombineEnabled[pipe_idx];
+			else
+				pipes[pipe_cnt].pipe.dest.odm_combine = 0;
+			pipe_idx++;
+		} else {
+			pipes[pipe_cnt].clks_cfg.dppclk_mhz =
+					context->bw_ctx.dml.vba.RequiredDPPCLK[vlevel][context->bw_ctx.dml.vba.maxMpcComb][pipe_split_from[i]];
+			if (context->bw_ctx.dml.vba.BlendingAndTiming[pipe_split_from[i]] == pipe_split_from[i])
+				pipes[pipe_cnt].pipe.dest.odm_combine =
+						context->bw_ctx.dml.vba.ODMCombineEnabled[pipe_split_from[i]];
+			else
+				pipes[pipe_cnt].pipe.dest.odm_combine = 0;
+		}
+
+		if (dc->config.forced_clocks) {
+			pipes[pipe_cnt].clks_cfg.dispclk_mhz = context->bw_ctx.dml.soc.clock_limits[0].dispclk_mhz;
+			pipes[pipe_cnt].clks_cfg.dppclk_mhz = context->bw_ctx.dml.soc.clock_limits[0].dppclk_mhz;
+		}
+		if (dc->debug.min_disp_clk_khz > pipes[pipe_cnt].clks_cfg.dispclk_mhz * 1000)
+			pipes[pipe_cnt].clks_cfg.dispclk_mhz = dc->debug.min_disp_clk_khz / 1000.0;
+		if (dc->debug.min_dpp_clk_khz > pipes[pipe_cnt].clks_cfg.dppclk_mhz * 1000)
+			pipes[pipe_cnt].clks_cfg.dppclk_mhz = dc->debug.min_dpp_clk_khz / 1000.0;
+
+		pipe_cnt++;
+	}
+
+	if (pipe_cnt != pipe_idx) {
+		if (dc->res_pool->funcs->populate_dml_pipes)
+			pipe_cnt = dc->res_pool->funcs->populate_dml_pipes(dc,
+				context, pipes, validate_mode);
+		else
+			pipe_cnt = dcn20_populate_dml_pipes_from_context(dc,
+				context, pipes, validate_mode);
+	}
+
+	*out_pipe_cnt = pipe_cnt;
+
+	pipes[0].clks_cfg.voltage = vlevel;
+	pipes[0].clks_cfg.dcfclk_mhz = context->bw_ctx.dml.soc.clock_limits[vlevel].dcfclk_mhz;
+	pipes[0].clks_cfg.socclk_mhz = context->bw_ctx.dml.soc.clock_limits[vlevel].socclk_mhz;
+
+	/* only pipe 0 is read for voltage and dcf/soc clocks */
+	if (vlevel < 1) {
+		pipes[0].clks_cfg.voltage = 1;
+		pipes[0].clks_cfg.dcfclk_mhz = context->bw_ctx.dml.soc.clock_limits[1].dcfclk_mhz;
+		pipes[0].clks_cfg.socclk_mhz = context->bw_ctx.dml.soc.clock_limits[1].socclk_mhz;
+	}
+	context->bw_ctx.bw.dcn.watermarks.b.urgent_ns = get_wm_urgent(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.b.cstate_pstate.cstate_enter_plus_exit_ns = get_wm_stutter_enter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.b.cstate_pstate.cstate_exit_ns = get_wm_stutter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.b.cstate_pstate.pstate_change_ns = get_wm_dram_clock_change(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.b.pte_meta_urgent_ns = get_wm_memory_trip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.b.frac_urg_bw_nom = get_fraction_of_urgent_bandwidth(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.b.frac_urg_bw_flip = get_fraction_of_urgent_bandwidth_imm_flip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.b.urgent_latency_ns = get_urgent_latency(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+
+	if (vlevel < 2) {
+		pipes[0].clks_cfg.voltage = 2;
+		pipes[0].clks_cfg.dcfclk_mhz = context->bw_ctx.dml.soc.clock_limits[2].dcfclk_mhz;
+		pipes[0].clks_cfg.socclk_mhz = context->bw_ctx.dml.soc.clock_limits[2].socclk_mhz;
+	}
+	context->bw_ctx.bw.dcn.watermarks.c.urgent_ns = get_wm_urgent(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.c.cstate_pstate.cstate_enter_plus_exit_ns = get_wm_stutter_enter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.c.cstate_pstate.cstate_exit_ns = get_wm_stutter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.c.cstate_pstate.pstate_change_ns = get_wm_dram_clock_change(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.c.pte_meta_urgent_ns = get_wm_memory_trip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.c.frac_urg_bw_nom = get_fraction_of_urgent_bandwidth(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.c.frac_urg_bw_flip = get_fraction_of_urgent_bandwidth_imm_flip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+
+	if (vlevel < 3) {
+		pipes[0].clks_cfg.voltage = 3;
+		pipes[0].clks_cfg.dcfclk_mhz = context->bw_ctx.dml.soc.clock_limits[2].dcfclk_mhz;
+		pipes[0].clks_cfg.socclk_mhz = context->bw_ctx.dml.soc.clock_limits[2].socclk_mhz;
+	}
+	context->bw_ctx.bw.dcn.watermarks.d.urgent_ns = get_wm_urgent(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.d.cstate_pstate.cstate_enter_plus_exit_ns = get_wm_stutter_enter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.d.cstate_pstate.cstate_exit_ns = get_wm_stutter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.d.cstate_pstate.pstate_change_ns = get_wm_dram_clock_change(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.d.pte_meta_urgent_ns = get_wm_memory_trip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.d.frac_urg_bw_nom = get_fraction_of_urgent_bandwidth(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.d.frac_urg_bw_flip = get_fraction_of_urgent_bandwidth_imm_flip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+
+	pipes[0].clks_cfg.voltage = vlevel;
+	pipes[0].clks_cfg.dcfclk_mhz = context->bw_ctx.dml.soc.clock_limits[vlevel].dcfclk_mhz;
+	pipes[0].clks_cfg.socclk_mhz = context->bw_ctx.dml.soc.clock_limits[vlevel].socclk_mhz;
+	context->bw_ctx.bw.dcn.watermarks.a.urgent_ns = get_wm_urgent(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.a.cstate_pstate.cstate_enter_plus_exit_ns = get_wm_stutter_enter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.a.cstate_pstate.cstate_exit_ns = get_wm_stutter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.a.cstate_pstate.pstate_change_ns = get_wm_dram_clock_change(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.a.pte_meta_urgent_ns = get_wm_memory_trip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.a.frac_urg_bw_nom = get_fraction_of_urgent_bandwidth(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.a.frac_urg_bw_flip = get_fraction_of_urgent_bandwidth_imm_flip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+}
+
+void dcn20_update_bounding_box(struct dc *dc,
+			       struct _vcs_dpi_soc_bounding_box_st *bb,
+			       struct pp_smu_nv_clock_table *max_clocks,
+			       unsigned int *uclk_states,
+			       unsigned int num_states)
+{
+	int num_calculated_states = 0;
+	int min_dcfclk = 0;
+	int i;
+
+	dc_assert_fp_enabled();
+
+	if (num_states == 0)
+		return;
+
+	memset(bb->clock_limits, 0, sizeof(bb->clock_limits));
+
+	if (dc->bb_overrides.min_dcfclk_mhz > 0) {
+		min_dcfclk = dc->bb_overrides.min_dcfclk_mhz;
+	} else {
+		if (ASICREV_IS_NAVI12_P(dc->ctx->asic_id.hw_internal_rev))
+			min_dcfclk = 310;
+		else
+			// Accounting for SOC/DCF relationship, we can go as high as
+			// 506Mhz in Vmin.
+			min_dcfclk = 506;
+	}
+
+	for (i = 0; i < num_states; i++) {
+		int min_fclk_required_by_uclk;
+		bb->clock_limits[i].state = i;
+		bb->clock_limits[i].dram_speed_mts = uclk_states[i] * 16 / 1000;
+
+		// FCLK:UCLK ratio is 1.08
+		min_fclk_required_by_uclk = div_u64(((unsigned long long)uclk_states[i]) * 1080,
+			1000000);
+
+		bb->clock_limits[i].fabricclk_mhz = (min_fclk_required_by_uclk < min_dcfclk) ?
+				min_dcfclk : min_fclk_required_by_uclk;
+
+		bb->clock_limits[i].socclk_mhz = (bb->clock_limits[i].fabricclk_mhz > max_clocks->socClockInKhz / 1000.0) ?
+				max_clocks->socClockInKhz / 1000 : bb->clock_limits[i].fabricclk_mhz;
+
+		bb->clock_limits[i].dcfclk_mhz = (bb->clock_limits[i].fabricclk_mhz > max_clocks->dcfClockInKhz / 1000.0) ?
+				max_clocks->dcfClockInKhz / 1000 : bb->clock_limits[i].fabricclk_mhz;
+
+		bb->clock_limits[i].dispclk_mhz = max_clocks->displayClockInKhz / 1000;
+		bb->clock_limits[i].dppclk_mhz = max_clocks->displayClockInKhz / 1000;
+		bb->clock_limits[i].dscclk_mhz = max_clocks->displayClockInKhz / (1000 * 3);
+
+		bb->clock_limits[i].phyclk_mhz = max_clocks->phyClockInKhz / 1000;
+
+		num_calculated_states++;
+	}
+
+	bb->clock_limits[num_calculated_states - 1].socclk_mhz = max_clocks->socClockInKhz / 1000;
+	bb->clock_limits[num_calculated_states - 1].fabricclk_mhz = max_clocks->socClockInKhz / 1000;
+	bb->clock_limits[num_calculated_states - 1].dcfclk_mhz = max_clocks->dcfClockInKhz / 1000;
+
+	bb->num_states = num_calculated_states;
+
+	// Duplicate the last state, DML always an extra state identical to max state to work
+	memcpy(&bb->clock_limits[num_calculated_states], &bb->clock_limits[num_calculated_states - 1], sizeof(struct _vcs_dpi_voltage_scaling_st));
+	bb->clock_limits[num_calculated_states].state = bb->num_states;
+}
+
+void dcn20_cap_soc_clocks(struct _vcs_dpi_soc_bounding_box_st *bb,
+			  struct pp_smu_nv_clock_table max_clocks)
+{
+	int i;
+
+	dc_assert_fp_enabled();
+
+	// First pass - cap all clocks higher than the reported max
+	for (i = 0; i < bb->num_states; i++) {
+		if ((bb->clock_limits[i].dcfclk_mhz > (max_clocks.dcfClockInKhz / 1000.0))
+				&& max_clocks.dcfClockInKhz != 0)
+			bb->clock_limits[i].dcfclk_mhz = (max_clocks.dcfClockInKhz / 1000);
+
+		if ((bb->clock_limits[i].dram_speed_mts > (max_clocks.uClockInKhz / 1000.0) * 16)
+						&& max_clocks.uClockInKhz != 0)
+			bb->clock_limits[i].dram_speed_mts = (max_clocks.uClockInKhz / 1000) * 16;
+
+		if ((bb->clock_limits[i].fabricclk_mhz > (max_clocks.fabricClockInKhz / 1000.0))
+						&& max_clocks.fabricClockInKhz != 0)
+			bb->clock_limits[i].fabricclk_mhz = (max_clocks.fabricClockInKhz / 1000);
+
+		if ((bb->clock_limits[i].dispclk_mhz > (max_clocks.displayClockInKhz / 1000.0))
+						&& max_clocks.displayClockInKhz != 0)
+			bb->clock_limits[i].dispclk_mhz = (max_clocks.displayClockInKhz / 1000);
+
+		if ((bb->clock_limits[i].dppclk_mhz > (max_clocks.dppClockInKhz / 1000.0))
+						&& max_clocks.dppClockInKhz != 0)
+			bb->clock_limits[i].dppclk_mhz = (max_clocks.dppClockInKhz / 1000);
+
+		if ((bb->clock_limits[i].phyclk_mhz > (max_clocks.phyClockInKhz / 1000.0))
+						&& max_clocks.phyClockInKhz != 0)
+			bb->clock_limits[i].phyclk_mhz = (max_clocks.phyClockInKhz / 1000);
+
+		if ((bb->clock_limits[i].socclk_mhz > (max_clocks.socClockInKhz / 1000.0))
+						&& max_clocks.socClockInKhz != 0)
+			bb->clock_limits[i].socclk_mhz = (max_clocks.socClockInKhz / 1000);
+
+		if ((bb->clock_limits[i].dscclk_mhz > (max_clocks.dscClockInKhz / 1000.0))
+						&& max_clocks.dscClockInKhz != 0)
+			bb->clock_limits[i].dscclk_mhz = (max_clocks.dscClockInKhz / 1000);
+	}
+
+	// Second pass - remove all duplicate clock states
+	for (i = bb->num_states - 1; i > 1; i--) {
+		bool duplicate = true;
+
+		if (bb->clock_limits[i-1].dcfclk_mhz != bb->clock_limits[i].dcfclk_mhz)
+			duplicate = false;
+		if (bb->clock_limits[i-1].dispclk_mhz != bb->clock_limits[i].dispclk_mhz)
+			duplicate = false;
+		if (bb->clock_limits[i-1].dppclk_mhz != bb->clock_limits[i].dppclk_mhz)
+			duplicate = false;
+		if (bb->clock_limits[i-1].dram_speed_mts != bb->clock_limits[i].dram_speed_mts)
+			duplicate = false;
+		if (bb->clock_limits[i-1].dscclk_mhz != bb->clock_limits[i].dscclk_mhz)
+			duplicate = false;
+		if (bb->clock_limits[i-1].fabricclk_mhz != bb->clock_limits[i].fabricclk_mhz)
+			duplicate = false;
+		if (bb->clock_limits[i-1].phyclk_mhz != bb->clock_limits[i].phyclk_mhz)
+			duplicate = false;
+		if (bb->clock_limits[i-1].socclk_mhz != bb->clock_limits[i].socclk_mhz)
+			duplicate = false;
+
+		if (duplicate)
+			bb->num_states--;
+	}
+}
+
+void dcn20_patch_bounding_box(struct dc *dc, struct _vcs_dpi_soc_bounding_box_st *bb)
+{
+	dc_assert_fp_enabled();
+
+	if ((int)(bb->sr_exit_time_us * 1000) != dc->bb_overrides.sr_exit_time_ns
+			&& dc->bb_overrides.sr_exit_time_ns) {
+		bb->sr_exit_time_us = dc->bb_overrides.sr_exit_time_ns / 1000.0;
+	}
+
+	if ((int)(bb->sr_enter_plus_exit_time_us * 1000)
+				!= dc->bb_overrides.sr_enter_plus_exit_time_ns
+			&& dc->bb_overrides.sr_enter_plus_exit_time_ns) {
+		bb->sr_enter_plus_exit_time_us =
+				dc->bb_overrides.sr_enter_plus_exit_time_ns / 1000.0;
+	}
+
+	if ((int)(bb->sr_exit_z8_time_us * 1000)
+				!= dc->bb_overrides.sr_exit_z8_time_ns
+			&& dc->bb_overrides.sr_exit_z8_time_ns) {
+		bb->sr_exit_z8_time_us = dc->bb_overrides.sr_exit_z8_time_ns / 1000.0;
+	}
+
+	if ((int)(bb->sr_enter_plus_exit_z8_time_us * 1000)
+				!= dc->bb_overrides.sr_enter_plus_exit_z8_time_ns
+			&& dc->bb_overrides.sr_enter_plus_exit_z8_time_ns) {
+		bb->sr_enter_plus_exit_z8_time_us = dc->bb_overrides.sr_enter_plus_exit_z8_time_ns / 1000.0;
+	}
+	if ((int)(bb->urgent_latency_us * 1000) != dc->bb_overrides.urgent_latency_ns
+			&& dc->bb_overrides.urgent_latency_ns) {
+		bb->urgent_latency_us = dc->bb_overrides.urgent_latency_ns / 1000.0;
+	}
+
+	if ((int)(bb->dram_clock_change_latency_us * 1000)
+				!= dc->bb_overrides.dram_clock_change_latency_ns
+			&& dc->bb_overrides.dram_clock_change_latency_ns) {
+		bb->dram_clock_change_latency_us =
+				dc->bb_overrides.dram_clock_change_latency_ns / 1000.0;
+	}
+
+	if ((int)(bb->dummy_pstate_latency_us * 1000)
+				!= dc->bb_overrides.dummy_clock_change_latency_ns
+			&& dc->bb_overrides.dummy_clock_change_latency_ns) {
+		bb->dummy_pstate_latency_us =
+				dc->bb_overrides.dummy_clock_change_latency_ns / 1000.0;
+	}
+}
+
+static bool dcn20_validate_bandwidth_internal(struct dc *dc, struct dc_state *context,
+		enum dc_validate_mode validate_mode, display_e2e_pipe_params_st *pipes)
+{
+	bool out = false;
+
+	BW_VAL_TRACE_SETUP();
+
+	int vlevel = 0;
+	int pipe_split_from[MAX_PIPES];
+	int pipe_cnt = 0;
+	DC_LOGGER_INIT(dc->ctx->logger);
+
+	BW_VAL_TRACE_COUNT();
+
+	out = dcn20_fast_validate_bw(dc, context, pipes, &pipe_cnt, pipe_split_from, &vlevel, validate_mode);
+
+	if (pipe_cnt == 0)
+		goto validate_out;
+
+	if (!out)
+		goto validate_fail;
+
+	BW_VAL_TRACE_END_VOLTAGE_LEVEL();
+
+	if (validate_mode != DC_VALIDATE_MODE_AND_PROGRAMMING) {
+		BW_VAL_TRACE_SKIP(fast);
+		goto validate_out;
+	}
+
+	dcn20_calculate_wm(dc, context, pipes, &pipe_cnt, pipe_split_from, vlevel, validate_mode);
+	dcn20_calculate_dlg_params(dc, context, pipes, pipe_cnt, vlevel);
+
+	BW_VAL_TRACE_END_WATERMARKS();
+
+	goto validate_out;
+
+validate_fail:
+	DC_LOG_WARNING("Mode Validation Warning: %s failed validation.\n",
+		dml_get_status_message(context->bw_ctx.dml.vba.ValidationStatus[context->bw_ctx.dml.vba.soc.num_states]));
+
+	BW_VAL_TRACE_SKIP(fail);
+	out = false;
+
+validate_out:
+
+	BW_VAL_TRACE_FINISH();
+
+	return out;
+}
+
+bool dcn20_validate_bandwidth_fp(struct dc *dc, struct dc_state *context,
+				 enum dc_validate_mode validate_mode, display_e2e_pipe_params_st *pipes)
+{
+	bool voltage_supported = false;
+	bool full_pstate_supported = false;
+	bool dummy_pstate_supported = false;
+	double p_state_latency_us;
+
+	dc_assert_fp_enabled();
+
+	p_state_latency_us = context->bw_ctx.dml.soc.dram_clock_change_latency_us;
+	context->bw_ctx.dml.soc.disable_dram_clock_change_vactive_support =
+		dc->debug.disable_dram_clock_change_vactive_support;
+	context->bw_ctx.dml.soc.allow_dram_clock_one_display_vactive =
+		dc->debug.enable_dram_clock_change_one_display_vactive;
+
+	/*Unsafe due to current pipe merge and split logic*/
+	ASSERT(context != dc->current_state);
+
+	if (validate_mode != DC_VALIDATE_MODE_AND_PROGRAMMING)
+		return dcn20_validate_bandwidth_internal(dc, context, validate_mode, pipes);
+
+	// Best case, we support full UCLK switch latency
+	voltage_supported = dcn20_validate_bandwidth_internal(dc, context, DC_VALIDATE_MODE_AND_PROGRAMMING, pipes);
+	full_pstate_supported = context->bw_ctx.bw.dcn.clk.p_state_change_support;
+
+	if (context->bw_ctx.dml.soc.dummy_pstate_latency_us == 0 ||
+		(voltage_supported && full_pstate_supported)) {
+		context->bw_ctx.bw.dcn.clk.p_state_change_support = full_pstate_supported;
+		goto restore_dml_state;
+	}
+
+	// Fallback: Try to only support G6 temperature read latency
+	context->bw_ctx.dml.soc.dram_clock_change_latency_us = context->bw_ctx.dml.soc.dummy_pstate_latency_us;
+
+	memset(pipes, 0, dc->res_pool->pipe_count * sizeof(display_e2e_pipe_params_st));
+	voltage_supported = dcn20_validate_bandwidth_internal(dc, context, DC_VALIDATE_MODE_AND_PROGRAMMING, pipes);
+	dummy_pstate_supported = context->bw_ctx.bw.dcn.clk.p_state_change_support;
+
+	if (voltage_supported && (dummy_pstate_supported || !(context->stream_count))) {
+		context->bw_ctx.bw.dcn.clk.p_state_change_support = false;
+		goto restore_dml_state;
+	}
+
+	// ERROR: fallback is supposed to always work.
+	ASSERT(false);
+
+restore_dml_state:
+	context->bw_ctx.dml.soc.dram_clock_change_latency_us = p_state_latency_us;
+	return voltage_supported;
+}
+
+void dcn20_fpu_set_wm_ranges(int i,
+			     struct pp_smu_wm_range_sets *ranges,
+			     struct _vcs_dpi_soc_bounding_box_st *loaded_bb)
+{
+	dc_assert_fp_enabled();
+
+	ranges->reader_wm_sets[i].min_fill_clk_mhz = (i > 0) ? (loaded_bb->clock_limits[i - 1].dram_speed_mts / 16) + 1 : 0;
+	ranges->reader_wm_sets[i].max_fill_clk_mhz = loaded_bb->clock_limits[i].dram_speed_mts / 16;
+}
+
+void dcn20_fpu_adjust_dppclk(struct vba_vars_st *v,
+			     int vlevel,
+			     int max_mpc_comb,
+			     int pipe_idx,
+			     bool is_validating_bw)
+{
+	dc_assert_fp_enabled();
+
+	if (is_validating_bw)
+		v->RequiredDPPCLK[vlevel][max_mpc_comb][pipe_idx] *= 2;
+	else
+		v->RequiredDPPCLK[vlevel][max_mpc_comb][pipe_idx] /= 2;
+}
+
+int dcn21_populate_dml_pipes_from_context(struct dc *dc,
+					  struct dc_state *context,
+					  display_e2e_pipe_params_st *pipes,
+					  enum dc_validate_mode validate_mode)
+{
+	uint32_t pipe_cnt;
+	int i;
+
+	dc_assert_fp_enabled();
+
+	pipe_cnt = dcn20_populate_dml_pipes_from_context(dc, context, pipes, validate_mode);
+
+	for (i = 0; i < pipe_cnt; i++) {
+
+		pipes[i].pipe.src.hostvm = dc->res_pool->hubbub->riommu_active;
+		pipes[i].pipe.src.gpuvm = 1;
+	}
+
+	return pipe_cnt;
+}
+
+static void patch_bounding_box(struct dc *dc, struct _vcs_dpi_soc_bounding_box_st *bb)
+{
+	int i;
+
+	if (dc->bb_overrides.sr_exit_time_ns) {
+		for (i = 0; i < WM_SET_COUNT; i++) {
+			  dc->clk_mgr->bw_params->wm_table.entries[i].sr_exit_time_us =
+					  dc->bb_overrides.sr_exit_time_ns / 1000.0;
+		}
+	}
+
+	if (dc->bb_overrides.sr_enter_plus_exit_time_ns) {
+		for (i = 0; i < WM_SET_COUNT; i++) {
+			  dc->clk_mgr->bw_params->wm_table.entries[i].sr_enter_plus_exit_time_us =
+					  dc->bb_overrides.sr_enter_plus_exit_time_ns / 1000.0;
+		}
+	}
+
+	if (dc->bb_overrides.urgent_latency_ns) {
+		bb->urgent_latency_us = dc->bb_overrides.urgent_latency_ns / 1000.0;
+	}
+
+	if (dc->bb_overrides.dram_clock_change_latency_ns) {
+		for (i = 0; i < WM_SET_COUNT; i++) {
+			dc->clk_mgr->bw_params->wm_table.entries[i].pstate_latency_us =
+				dc->bb_overrides.dram_clock_change_latency_ns / 1000.0;
+		}
+	}
+}
+
+static void calculate_wm_set_for_vlevel(int vlevel,
+					struct wm_range_table_entry *table_entry,
+					struct dcn_watermarks *wm_set,
+					struct display_mode_lib *dml,
+					display_e2e_pipe_params_st *pipes,
+					int pipe_cnt)
+{
+	double dram_clock_change_latency_cached = dml->soc.dram_clock_change_latency_us;
+
+	ASSERT(vlevel < dml->soc.num_states);
+	/* only pipe 0 is read for voltage and dcf/soc clocks */
+	pipes[0].clks_cfg.voltage = vlevel;
+	pipes[0].clks_cfg.dcfclk_mhz = dml->soc.clock_limits[vlevel].dcfclk_mhz;
+	pipes[0].clks_cfg.socclk_mhz = dml->soc.clock_limits[vlevel].socclk_mhz;
+
+	dml->soc.dram_clock_change_latency_us = table_entry->pstate_latency_us;
+	dml->soc.sr_exit_time_us = table_entry->sr_exit_time_us;
+	dml->soc.sr_enter_plus_exit_time_us = table_entry->sr_enter_plus_exit_time_us;
+
+	wm_set->urgent_ns = get_wm_urgent(dml, pipes, pipe_cnt) * 1000;
+	wm_set->cstate_pstate.cstate_enter_plus_exit_ns = get_wm_stutter_enter_exit(dml, pipes, pipe_cnt) * 1000;
+	wm_set->cstate_pstate.cstate_exit_ns = get_wm_stutter_exit(dml, pipes, pipe_cnt) * 1000;
+	wm_set->cstate_pstate.pstate_change_ns = get_wm_dram_clock_change(dml, pipes, pipe_cnt) * 1000;
+	wm_set->pte_meta_urgent_ns = get_wm_memory_trip(dml, pipes, pipe_cnt) * 1000;
+	wm_set->frac_urg_bw_nom = get_fraction_of_urgent_bandwidth(dml, pipes, pipe_cnt) * 1000;
+	wm_set->frac_urg_bw_flip = get_fraction_of_urgent_bandwidth_imm_flip(dml, pipes, pipe_cnt) * 1000;
+	wm_set->urgent_latency_ns = get_urgent_latency(dml, pipes, pipe_cnt) * 1000;
+	dml->soc.dram_clock_change_latency_us = dram_clock_change_latency_cached;
+}
+
+static void dcn21_calculate_wm(struct dc *dc, struct dc_state *context,
+			display_e2e_pipe_params_st *pipes,
+			int *out_pipe_cnt,
+			int *pipe_split_from,
+			int vlevel_req,
+			enum dc_validate_mode validate_mode)
+{
+	int pipe_cnt, i, pipe_idx;
+	int vlevel, vlevel_max;
+	struct wm_range_table_entry *table_entry;
+	struct clk_bw_params *bw_params = dc->clk_mgr->bw_params;
+
+	ASSERT(bw_params);
+
+	patch_bounding_box(dc, &context->bw_ctx.dml.soc);
+
+	for (i = 0, pipe_idx = 0, pipe_cnt = 0; i < dc->res_pool->pipe_count; i++) {
+			if (!context->res_ctx.pipe_ctx[i].stream)
+				continue;
+
+			pipes[pipe_cnt].clks_cfg.refclk_mhz = dc->res_pool->ref_clocks.dchub_ref_clock_inKhz / 1000.0;
+			pipes[pipe_cnt].clks_cfg.dispclk_mhz = context->bw_ctx.dml.vba.RequiredDISPCLK[vlevel_req][context->bw_ctx.dml.vba.maxMpcComb];
+
+			if (pipe_split_from[i] < 0) {
+				pipes[pipe_cnt].clks_cfg.dppclk_mhz =
+						context->bw_ctx.dml.vba.RequiredDPPCLK[vlevel_req][context->bw_ctx.dml.vba.maxMpcComb][pipe_idx];
+				if (context->bw_ctx.dml.vba.BlendingAndTiming[pipe_idx] == pipe_idx)
+					pipes[pipe_cnt].pipe.dest.odm_combine =
+							context->bw_ctx.dml.vba.ODMCombineEnablePerState[vlevel_req][pipe_idx];
+				else
+					pipes[pipe_cnt].pipe.dest.odm_combine = 0;
+				pipe_idx++;
+			} else {
+				pipes[pipe_cnt].clks_cfg.dppclk_mhz =
+						context->bw_ctx.dml.vba.RequiredDPPCLK[vlevel_req][context->bw_ctx.dml.vba.maxMpcComb][pipe_split_from[i]];
+				if (context->bw_ctx.dml.vba.BlendingAndTiming[pipe_split_from[i]] == pipe_split_from[i])
+					pipes[pipe_cnt].pipe.dest.odm_combine =
+							context->bw_ctx.dml.vba.ODMCombineEnablePerState[vlevel_req][pipe_split_from[i]];
+				else
+					pipes[pipe_cnt].pipe.dest.odm_combine = 0;
+			}
+			pipe_cnt++;
+	}
+
+	if (pipe_cnt != pipe_idx) {
+		if (dc->res_pool->funcs->populate_dml_pipes)
+			pipe_cnt = dc->res_pool->funcs->populate_dml_pipes(dc,
+				context, pipes, validate_mode);
+		else
+			pipe_cnt = dcn21_populate_dml_pipes_from_context(dc,
+				context, pipes, validate_mode);
+	}
+
+	*out_pipe_cnt = pipe_cnt;
+
+	vlevel_max = bw_params->clk_table.num_entries - 1;
+
+
+	/* WM Set D */
+	table_entry = &bw_params->wm_table.entries[WM_D];
+	if (table_entry->wm_type == WM_TYPE_RETRAINING)
+		vlevel = 0;
+	else
+		vlevel = vlevel_max;
+	calculate_wm_set_for_vlevel(vlevel, table_entry, &context->bw_ctx.bw.dcn.watermarks.d,
+						&context->bw_ctx.dml, pipes, pipe_cnt);
+	/* WM Set C */
+	table_entry = &bw_params->wm_table.entries[WM_C];
+	vlevel = MIN(MAX(vlevel_req, 3), vlevel_max);
+	calculate_wm_set_for_vlevel(vlevel, table_entry, &context->bw_ctx.bw.dcn.watermarks.c,
+						&context->bw_ctx.dml, pipes, pipe_cnt);
+	/* WM Set B */
+	table_entry = &bw_params->wm_table.entries[WM_B];
+	vlevel = MIN(MAX(vlevel_req, 2), vlevel_max);
+	calculate_wm_set_for_vlevel(vlevel, table_entry, &context->bw_ctx.bw.dcn.watermarks.b,
+						&context->bw_ctx.dml, pipes, pipe_cnt);
+
+	/* WM Set A */
+	table_entry = &bw_params->wm_table.entries[WM_A];
+	vlevel = MIN(vlevel_req, vlevel_max);
+	calculate_wm_set_for_vlevel(vlevel, table_entry, &context->bw_ctx.bw.dcn.watermarks.a,
+						&context->bw_ctx.dml, pipes, pipe_cnt);
+}
+
+bool dcn21_validate_bandwidth_fp(struct dc *dc, struct dc_state *context,
+				 enum dc_validate_mode validate_mode, display_e2e_pipe_params_st *pipes)
+{
+	bool out = false;
+
+	BW_VAL_TRACE_SETUP();
+
+	int vlevel = 0;
+	int pipe_split_from[MAX_PIPES];
+	int pipe_cnt = 0;
+	DC_LOGGER_INIT(dc->ctx->logger);
+
+	BW_VAL_TRACE_COUNT();
+
+	dc_assert_fp_enabled();
+
+	/*Unsafe due to current pipe merge and split logic*/
+	ASSERT(context != dc->current_state);
+
+	out = dcn21_fast_validate_bw(dc, context, pipes, &pipe_cnt, pipe_split_from, &vlevel, validate_mode);
+
+	if (pipe_cnt == 0)
+		goto validate_out;
+
+	if (!out)
+		goto validate_fail;
+
+	BW_VAL_TRACE_END_VOLTAGE_LEVEL();
+
+	if (validate_mode != DC_VALIDATE_MODE_AND_PROGRAMMING) {
+		BW_VAL_TRACE_SKIP(fast);
+		goto validate_out;
+	}
+
+	dcn21_calculate_wm(dc, context, pipes, &pipe_cnt, pipe_split_from, vlevel, validate_mode);
+	dcn20_calculate_dlg_params(dc, context, pipes, pipe_cnt, vlevel);
+
+	BW_VAL_TRACE_END_WATERMARKS();
+
+	goto validate_out;
+
+validate_fail:
+	DC_LOG_WARNING("Mode Validation Warning: %s failed validation.\n",
+			dml_get_status_message(context->bw_ctx.dml.vba.ValidationStatus[context->bw_ctx.dml.vba.soc.num_states]));
+
+	BW_VAL_TRACE_SKIP(fail);
+	out = false;
+
+validate_out:
+
+	BW_VAL_TRACE_FINISH();
+
+	return out;
+}
+
+static struct _vcs_dpi_voltage_scaling_st construct_low_pstate_lvl(struct clk_limit_table *clk_table, unsigned int high_voltage_lvl)
+{
+	struct _vcs_dpi_voltage_scaling_st low_pstate_lvl = {0};
+	int i;
+
+	low_pstate_lvl.state = 1;
+	low_pstate_lvl.dcfclk_mhz = clk_table->entries[0].dcfclk_mhz;
+	low_pstate_lvl.fabricclk_mhz = clk_table->entries[0].fclk_mhz;
+	low_pstate_lvl.socclk_mhz = clk_table->entries[0].socclk_mhz;
+	low_pstate_lvl.dram_speed_mts = clk_table->entries[0].memclk_mhz * 2;
+
+	low_pstate_lvl.dispclk_mhz = dcn2_1_soc.clock_limits[high_voltage_lvl].dispclk_mhz;
+	low_pstate_lvl.dppclk_mhz = dcn2_1_soc.clock_limits[high_voltage_lvl].dppclk_mhz;
+	low_pstate_lvl.dram_bw_per_chan_gbps = dcn2_1_soc.clock_limits[high_voltage_lvl].dram_bw_per_chan_gbps;
+	low_pstate_lvl.dscclk_mhz = dcn2_1_soc.clock_limits[high_voltage_lvl].dscclk_mhz;
+	low_pstate_lvl.dtbclk_mhz = dcn2_1_soc.clock_limits[high_voltage_lvl].dtbclk_mhz;
+	low_pstate_lvl.phyclk_d18_mhz = dcn2_1_soc.clock_limits[high_voltage_lvl].phyclk_d18_mhz;
+	low_pstate_lvl.phyclk_mhz = dcn2_1_soc.clock_limits[high_voltage_lvl].phyclk_mhz;
+
+	for (i = clk_table->num_entries; i > 1; i--)
+		clk_table->entries[i] = clk_table->entries[i-1];
+	clk_table->entries[1] = clk_table->entries[0];
+	clk_table->num_entries++;
+
+	return low_pstate_lvl;
+}
+
+void dcn21_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params)
+{
+	struct _vcs_dpi_voltage_scaling_st *s = dc->scratch.update_bw_bounding_box.clock_limits;
+	struct dcn21_resource_pool *pool = TO_DCN21_RES_POOL(dc->res_pool);
+	struct clk_limit_table *clk_table = &bw_params->clk_table;
+	unsigned int i, closest_clk_lvl = 0, k = 0;
+	int j;
+
+	dc_assert_fp_enabled();
+
+	dcn2_1_ip.max_num_otg = pool->base.res_cap->num_timing_generator;
+	dcn2_1_ip.max_num_dpp = pool->base.pipe_count;
+	dcn2_1_soc.num_chans = bw_params->num_channels;
+
+	ASSERT(clk_table->num_entries);
+	/* Copy dcn2_1_soc.clock_limits to clock_limits to avoid copying over null states later */
+	memcpy(s, dcn2_1_soc.clock_limits, sizeof(dcn2_1_soc.clock_limits));
+
+	for (i = 0; i < clk_table->num_entries; i++) {
+		/* loop backwards*/
+		for (closest_clk_lvl = 0, j = dcn2_1_soc.num_states - 1; j >= 0; j--) {
+			if ((unsigned int) dcn2_1_soc.clock_limits[j].dcfclk_mhz <= clk_table->entries[i].dcfclk_mhz) {
+				closest_clk_lvl = j;
+				break;
+			}
+		}
+
+		/* clk_table[1] is reserved for min DF PState.  skip here to fill in later. */
+		if (i == 1)
+			k++;
+
+		s[k].state = k;
+		s[k].dcfclk_mhz = clk_table->entries[i].dcfclk_mhz;
+		s[k].fabricclk_mhz = clk_table->entries[i].fclk_mhz;
+		s[k].socclk_mhz = clk_table->entries[i].socclk_mhz;
+		s[k].dram_speed_mts = clk_table->entries[i].memclk_mhz * 2;
+
+		s[k].dispclk_mhz = dcn2_1_soc.clock_limits[closest_clk_lvl].dispclk_mhz;
+		s[k].dppclk_mhz = dcn2_1_soc.clock_limits[closest_clk_lvl].dppclk_mhz;
+		s[k].dram_bw_per_chan_gbps =
+			dcn2_1_soc.clock_limits[closest_clk_lvl].dram_bw_per_chan_gbps;
+		s[k].dscclk_mhz = dcn2_1_soc.clock_limits[closest_clk_lvl].dscclk_mhz;
+		s[k].dtbclk_mhz = dcn2_1_soc.clock_limits[closest_clk_lvl].dtbclk_mhz;
+		s[k].phyclk_d18_mhz = dcn2_1_soc.clock_limits[closest_clk_lvl].phyclk_d18_mhz;
+		s[k].phyclk_mhz = dcn2_1_soc.clock_limits[closest_clk_lvl].phyclk_mhz;
+
+		k++;
+	}
+
+	memcpy(&dcn2_1_soc.clock_limits, s, sizeof(dcn2_1_soc.clock_limits));
+
+	if (clk_table->num_entries) {
+		dcn2_1_soc.num_states = clk_table->num_entries + 1;
+		/* fill in min DF PState */
+		dcn2_1_soc.clock_limits[1] = construct_low_pstate_lvl(clk_table, closest_clk_lvl);
+		/* duplicate last level */
+		dcn2_1_soc.clock_limits[dcn2_1_soc.num_states] = dcn2_1_soc.clock_limits[dcn2_1_soc.num_states - 1];
+		dcn2_1_soc.clock_limits[dcn2_1_soc.num_states].state = dcn2_1_soc.num_states;
+	}
+
+	dml_init_instance(&dc->dml, &dcn2_1_soc, &dcn2_1_ip, DML_PROJECT_DCN21);
+}
+
+void dcn21_clk_mgr_set_bw_params_wm_table(struct clk_bw_params *bw_params)
+{
+	dc_assert_fp_enabled();
+
+	bw_params->wm_table.entries[WM_D].pstate_latency_us = LPDDR_MEM_RETRAIN_LATENCY;
+	bw_params->wm_table.entries[WM_D].wm_inst = WM_D;
+	bw_params->wm_table.entries[WM_D].wm_type = WM_TYPE_RETRAINING;
+	bw_params->wm_table.entries[WM_D].valid = true;
+}
+
+void dcn201_populate_dml_writeback_from_context_fpu(struct dc *dc,
+						    struct resource_context *res_ctx,
+						    display_e2e_pipe_params_st *pipes)
+{
+	int pipe_cnt, i, j;
+	double max_calc_writeback_dispclk;
+	double writeback_dispclk;
+	struct writeback_st dout_wb = {0};
+
+	dc_assert_fp_enabled();
+
+	for (i = 0, pipe_cnt = 0; i < dc->res_pool->pipe_count; i++) {
+		struct dc_stream_state *stream = res_ctx->pipe_ctx[i].stream;
+
+		if (!stream)
+			continue;
+		max_calc_writeback_dispclk = 0;
+
+		/* Set writeback information */
+		pipes[pipe_cnt].dout.wb_enable = 0;
+		pipes[pipe_cnt].dout.num_active_wb = 0;
+		for (j = 0; j < stream->num_wb_info; j++) {
+			struct dc_writeback_info *wb_info = &stream->writeback_info[j];
+
+			if (wb_info->wb_enabled && wb_info->writeback_source_plane &&
+					(wb_info->writeback_source_plane == res_ctx->pipe_ctx[i].plane_state)) {
+				pipes[pipe_cnt].dout.wb_enable = 1;
+				pipes[pipe_cnt].dout.num_active_wb++;
+				dout_wb.wb_src_height = wb_info->dwb_params.cnv_params.crop_en ?
+					wb_info->dwb_params.cnv_params.crop_height :
+					wb_info->dwb_params.cnv_params.src_height;
+				dout_wb.wb_src_width = wb_info->dwb_params.cnv_params.crop_en ?
+					wb_info->dwb_params.cnv_params.crop_width :
+					wb_info->dwb_params.cnv_params.src_width;
+				dout_wb.wb_dst_width = wb_info->dwb_params.dest_width;
+				dout_wb.wb_dst_height = wb_info->dwb_params.dest_height;
+				dout_wb.wb_htaps_luma = wb_info->dwb_params.scaler_taps.h_taps;
+				dout_wb.wb_vtaps_luma = wb_info->dwb_params.scaler_taps.v_taps;
+				dout_wb.wb_htaps_chroma = wb_info->dwb_params.scaler_taps.h_taps_c;
+				dout_wb.wb_vtaps_chroma = wb_info->dwb_params.scaler_taps.v_taps_c;
+				dout_wb.wb_hratio = wb_info->dwb_params.cnv_params.crop_en ?
+					(double)wb_info->dwb_params.cnv_params.crop_width /
+						(double)wb_info->dwb_params.dest_width :
+					(double)wb_info->dwb_params.cnv_params.src_width /
+						(double)wb_info->dwb_params.dest_width;
+				dout_wb.wb_vratio = wb_info->dwb_params.cnv_params.crop_en ?
+					(double)wb_info->dwb_params.cnv_params.crop_height /
+						(double)wb_info->dwb_params.dest_height :
+					(double)wb_info->dwb_params.cnv_params.src_height /
+						(double)wb_info->dwb_params.dest_height;
+				if (wb_info->dwb_params.out_format == dwb_scaler_mode_yuv420) {
+					if (wb_info->dwb_params.output_depth == DWB_OUTPUT_PIXEL_DEPTH_8BPC)
+						dout_wb.wb_pixel_format = dm_420_8;
+					else
+						dout_wb.wb_pixel_format = dm_420_10;
+				} else
+					dout_wb.wb_pixel_format = dm_444_32;
+
+				/* Workaround for cases where multiple writebacks are connected to same plane
+				 * In which case, need to compute worst case and set the associated writeback parameters
+				 * This workaround is necessary due to DML computation assuming only 1 set of writeback
+				 * parameters per pipe */
+				writeback_dispclk = CalculateWriteBackDISPCLK(
+						dout_wb.wb_pixel_format,
+						pipes[pipe_cnt].pipe.dest.pixel_rate_mhz,
+						dout_wb.wb_hratio,
+						dout_wb.wb_vratio,
+						dout_wb.wb_htaps_luma,
+						dout_wb.wb_vtaps_luma,
+						dout_wb.wb_htaps_chroma,
+						dout_wb.wb_vtaps_chroma,
+						dout_wb.wb_dst_width,
+						pipes[pipe_cnt].pipe.dest.htotal,
+						2);
+
+				if (writeback_dispclk > max_calc_writeback_dispclk) {
+					max_calc_writeback_dispclk = writeback_dispclk;
+					pipes[pipe_cnt].dout.wb = dout_wb;
+				}
+			}
+		}
+
+		pipe_cnt++;
+	}
+
+}
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn20/dcn20_fpu.h b/drivers/gpu/drm/amd/display/dc/dml/dcn20/dcn20_fpu.h
index 36f26126d574..aed00039ca62 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dcn20/dcn20_fpu.h
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn20/dcn20_fpu.h
@@ -23,6 +23,7 @@
  * Authors: AMD
  *
  */
+#include "core_types.h"
 
 #ifndef __DCN20_FPU_H__
 #define __DCN20_FPU_H__
@@ -31,4 +32,58 @@ void dcn20_populate_dml_writeback_from_context(struct dc *dc,
 					       struct resource_context *res_ctx,
 					       display_e2e_pipe_params_st *pipes);
 
+void dcn20_fpu_set_wb_arb_params(struct mcif_arb_params *wb_arb_params,
+				 struct dc_state *context,
+				 display_e2e_pipe_params_st *pipes,
+				 int pipe_cnt, int i);
+void dcn20_calculate_dlg_params(struct dc *dc,
+				struct dc_state *context,
+				display_e2e_pipe_params_st *pipes,
+				int pipe_cnt,
+				int vlevel);
+int dcn20_populate_dml_pipes_from_context(struct dc *dc,
+					  struct dc_state *context,
+					  display_e2e_pipe_params_st *pipes,
+					  enum dc_validate_mode validate_mode);
+void dcn20_calculate_wm(struct dc *dc,
+			struct dc_state *context,
+			display_e2e_pipe_params_st *pipes,
+			int *out_pipe_cnt,
+			int *pipe_split_from,
+			int vlevel,
+			enum dc_validate_mode validate_mode);
+void dcn20_cap_soc_clocks(struct _vcs_dpi_soc_bounding_box_st *bb,
+			  struct pp_smu_nv_clock_table max_clocks);
+void dcn20_update_bounding_box(struct dc *dc,
+			       struct _vcs_dpi_soc_bounding_box_st *bb,
+			       struct pp_smu_nv_clock_table *max_clocks,
+			       unsigned int *uclk_states,
+			       unsigned int num_states);
+void dcn20_patch_bounding_box(struct dc *dc,
+			      struct _vcs_dpi_soc_bounding_box_st *bb);
+bool dcn20_validate_bandwidth_fp(struct dc *dc, struct dc_state *context,
+				 enum dc_validate_mode validate_mode, display_e2e_pipe_params_st *pipes);
+void dcn20_fpu_set_wm_ranges(int i,
+			     struct pp_smu_wm_range_sets *ranges,
+			     struct _vcs_dpi_soc_bounding_box_st *loaded_bb);
+void dcn20_fpu_adjust_dppclk(struct vba_vars_st *v,
+			     int vlevel,
+			     int max_mpc_comb,
+			     int pipe_idx,
+			     bool is_validating_bw);
+
+int dcn21_populate_dml_pipes_from_context(struct dc *dc,
+					  struct dc_state *context,
+					  display_e2e_pipe_params_st *pipes,
+					  enum dc_validate_mode validate_mode);
+bool dcn21_validate_bandwidth_fp(struct dc *dc, struct dc_state *context, enum
+				 dc_validate_mode, display_e2e_pipe_params_st *pipes);
+void dcn21_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params);
+
+void dcn21_clk_mgr_set_bw_params_wm_table(struct clk_bw_params *bw_params);
+
+void dcn201_populate_dml_writeback_from_context_fpu(struct dc *dc,
+						struct resource_context *res_ctx,
+						display_e2e_pipe_params_st *pipes);
+
 #endif /* __DCN20_FPU_H__ */
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn20/display_mode_vba_20.c b/drivers/gpu/drm/amd/display/dc/dml/dcn20/display_mode_vba_20.c
index d3b5b6fedf04..0c8c4a080c50 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dcn20/display_mode_vba_20.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn20/display_mode_vba_20.c
@@ -785,12 +785,9 @@ static bool CalculatePrefetchSchedule(
 
 	if (MyError) {
 		*PrefetchBandwidth = 0;
-		TimeForFetchingMetaPTE = 0;
-		TimeForFetchingRowInVBlank = 0;
 		*DestinationLinesToRequestVMInVBlank = 0;
 		*DestinationLinesToRequestRowInVBlank = 0;
 		*DestinationLinesForPrefetch = 0;
-		LinesToRequestPrefetchPixelData = 0;
 		*VRatioPrefetchY = 0;
 		*VRatioPrefetchC = 0;
 		*RequiredPrefetchPixDataBW = 0;
@@ -1017,7 +1014,7 @@ static unsigned int CalculateVMAndRowBytes(
 			if (ScanDirection == dm_horz)
 				FractionOfPTEReturnDrop = 0;
 			else
-				FractionOfPTEReturnDrop = 7 / 8;
+				FractionOfPTEReturnDrop = 7.0 / 8;
 		} else if (VMMPageSize == 4096 && MacroTileSizeBytes > 4096) {
 			PixelPTEReqHeight = 16 * BlockHeight256Bytes;
 			PixelPTEReqWidth = 16 * BlockWidth256Bytes;
@@ -3231,22 +3228,22 @@ static unsigned int TruncToValidBPP(
 			if (Format == dm_420) {
 				if (DecimalBPP < 6)
 					return BPP_INVALID;
-				else if (DecimalBPP >= 1.5 * DSCInputBitPerComponent - 1 / 16)
-					return 1.5 * DSCInputBitPerComponent - 1 / 16;
+				else if (DecimalBPP >= 1.5 * DSCInputBitPerComponent - 1.0 / 16)
+					return 1.5 * DSCInputBitPerComponent - 1.0 / 16;
 				else
 					return dml_floor(16 * DecimalBPP, 1) / 16;
 			} else if (Format == dm_n422) {
 				if (DecimalBPP < 7)
 					return BPP_INVALID;
-				else if (DecimalBPP >= 2 * DSCInputBitPerComponent - 1 / 16)
-					return 2 * DSCInputBitPerComponent - 1 / 16;
+				else if (DecimalBPP >= 2 * DSCInputBitPerComponent - 1.0 / 16)
+					return 2 * DSCInputBitPerComponent - 1.0 / 16;
 				else
 					return dml_floor(16 * DecimalBPP, 1) / 16;
 			} else {
 				if (DecimalBPP < 8)
 					return BPP_INVALID;
-				else if (DecimalBPP >= 3 * DSCInputBitPerComponent - 1 / 16)
-					return 3 * DSCInputBitPerComponent - 1 / 16;
+				else if (DecimalBPP >= 3 * DSCInputBitPerComponent - 1.0 / 16)
+					return 3 * DSCInputBitPerComponent - 1.0 / 16;
 				else
 					return dml_floor(16 * DecimalBPP, 1) / 16;
 			}
@@ -3897,14 +3894,14 @@ void dml20_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 					mode_lib->vba.PlaneRequiredDISPCLKWithODMCombine = mode_lib->vba.PixelClock[k] / 2
 							* (1 + mode_lib->vba.DISPCLKDPPCLKDSCCLKDownSpreading / 100.0);
 
-				locals->ODMCombineEnablePerState[i][k] = false;
+				locals->ODMCombineEnablePerState[i][k] = dm_odm_combine_mode_disabled;
 				mode_lib->vba.PlaneRequiredDISPCLK = mode_lib->vba.PlaneRequiredDISPCLKWithoutODMCombine;
 				if (mode_lib->vba.ODMCapability) {
 					if (locals->PlaneRequiredDISPCLKWithoutODMCombine > mode_lib->vba.MaxDispclkRoundedDownToDFSGranularity) {
-						locals->ODMCombineEnablePerState[i][k] = true;
+						locals->ODMCombineEnablePerState[i][k] = dm_odm_combine_mode_2to1;
 						mode_lib->vba.PlaneRequiredDISPCLK = mode_lib->vba.PlaneRequiredDISPCLKWithODMCombine;
 					} else if (locals->HActive[k] > DCN20_MAX_420_IMAGE_WIDTH && locals->OutputFormat[k] == dm_420) {
-						locals->ODMCombineEnablePerState[i][k] = true;
+						locals->ODMCombineEnablePerState[i][k] = dm_odm_combine_mode_2to1;
 						mode_lib->vba.PlaneRequiredDISPCLK = mode_lib->vba.PlaneRequiredDISPCLKWithODMCombine;
 					}
 				}
@@ -3957,7 +3954,7 @@ void dml20_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 				locals->RequiredDISPCLK[i][j] = 0.0;
 				locals->DISPCLK_DPPCLK_Support[i][j] = true;
 				for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) {
-					locals->ODMCombineEnablePerState[i][k] = false;
+					locals->ODMCombineEnablePerState[i][k] = dm_odm_combine_mode_disabled;
 					if (locals->SwathWidthYSingleDPP[k] <= locals->MaximumSwathWidth[k]) {
 						locals->NoOfDPP[i][j][k] = 1;
 						locals->RequiredDPPCLK[i][j][k] = locals->MinDPPCLKUsingSingleDPP[k]
@@ -4322,7 +4319,7 @@ void dml20_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 					locals->RoundedUpMaxSwathSizeBytesC = 0;
 				}
 
-				if (locals->RoundedUpMaxSwathSizeBytesY + locals->RoundedUpMaxSwathSizeBytesC <= locals->DETBufferSizeInKByte[0] * 1024 / 2) {
+				if (locals->RoundedUpMaxSwathSizeBytesY + locals->RoundedUpMaxSwathSizeBytesC <= locals->DETBufferSizeInKByte[0] * 1024.0 / 2) {
 					locals->SwathHeightYPerState[i][j][k] = locals->MaxSwathHeightY[k];
 					locals->SwathHeightCPerState[i][j][k] = locals->MaxSwathHeightC[k];
 				} else {
@@ -4356,12 +4353,16 @@ void dml20_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 						locals->PSCL_FACTOR[k] / locals->ReturnBWPerState[i][0],
 						locals->EffectiveLBLatencyHidingSourceLinesLuma),
 						locals->SwathHeightYPerState[i][j][k]);
-
-				locals->EffectiveDETLBLinesChroma = dml_floor(locals->LinesInDETChroma + dml_min(
-						locals->LinesInDETChroma * locals->RequiredDISPCLK[i][j] * locals->BytePerPixelInDETC[k] *
-						locals->PSCL_FACTOR_CHROMA[k] / locals->ReturnBWPerState[i][0],
-						locals->EffectiveLBLatencyHidingSourceLinesChroma),
-						locals->SwathHeightCPerState[i][j][k]);
+				if (locals->LinesInDETChroma) {
+					locals->EffectiveDETLBLinesChroma = dml_floor(locals->LinesInDETChroma +
+						    dml_min(locals->LinesInDETChroma * locals->RequiredDISPCLK[i][j] *
+						    locals->BytePerPixelInDETC[k] *
+							locals->PSCL_FACTOR_CHROMA[k] / locals->ReturnBWPerState[i][0],
+							locals->EffectiveLBLatencyHidingSourceLinesChroma),
+							locals->SwathHeightCPerState[i][j][k]);
+				} else {
+					locals->EffectiveDETLBLinesChroma = 0;
+				}
 
 				if (locals->BytePerPixelInDETC[k] == 0) {
 					locals->UrgentLatencySupportUsPerState[i][j][k] = locals->EffectiveDETLBLinesLuma * (locals->HTotal[k] / locals->PixelClock[k])
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn20/display_mode_vba_20v2.c b/drivers/gpu/drm/amd/display/dc/dml/dcn20/display_mode_vba_20v2.c
index 63bbdf8b8678..c935903b68e1 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dcn20/display_mode_vba_20v2.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn20/display_mode_vba_20v2.c
@@ -845,12 +845,9 @@ static bool CalculatePrefetchSchedule(
 
 	if (MyError) {
 		*PrefetchBandwidth = 0;
-		TimeForFetchingMetaPTE = 0;
-		TimeForFetchingRowInVBlank = 0;
 		*DestinationLinesToRequestVMInVBlank = 0;
 		*DestinationLinesToRequestRowInVBlank = 0;
 		*DestinationLinesForPrefetch = 0;
-		LinesToRequestPrefetchPixelData = 0;
 		*VRatioPrefetchY = 0;
 		*VRatioPrefetchC = 0;
 		*RequiredPrefetchPixDataBW = 0;
@@ -1077,7 +1074,7 @@ static unsigned int CalculateVMAndRowBytes(
 			if (ScanDirection == dm_horz)
 				FractionOfPTEReturnDrop = 0;
 			else
-				FractionOfPTEReturnDrop = 7 / 8;
+				FractionOfPTEReturnDrop = 7.0 / 8;
 		} else if (VMMPageSize == 4096 && MacroTileSizeBytes > 4096) {
 			PixelPTEReqHeight = 16 * BlockHeight256Bytes;
 			PixelPTEReqWidth = 16 * BlockWidth256Bytes;
@@ -4008,17 +4005,17 @@ void dml20v2_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode
 					mode_lib->vba.PlaneRequiredDISPCLKWithODMCombine = mode_lib->vba.PixelClock[k] / 2
 							* (1 + mode_lib->vba.DISPCLKDPPCLKDSCCLKDownSpreading / 100.0);
 
-				locals->ODMCombineEnablePerState[i][k] = false;
+				locals->ODMCombineEnablePerState[i][k] = dm_odm_combine_mode_disabled;
 				mode_lib->vba.PlaneRequiredDISPCLK = mode_lib->vba.PlaneRequiredDISPCLKWithoutODMCombine;
 				if (mode_lib->vba.ODMCapability) {
 					if (locals->PlaneRequiredDISPCLKWithoutODMCombine > MaxMaxDispclkRoundedDown) {
-						locals->ODMCombineEnablePerState[i][k] = true;
+						locals->ODMCombineEnablePerState[i][k] = dm_odm_combine_mode_2to1;
 						mode_lib->vba.PlaneRequiredDISPCLK = mode_lib->vba.PlaneRequiredDISPCLKWithODMCombine;
 					} else if (locals->DSCEnabled[k] && (locals->HActive[k] > DCN20_MAX_DSC_IMAGE_WIDTH)) {
-						locals->ODMCombineEnablePerState[i][k] = true;
+						locals->ODMCombineEnablePerState[i][k] = dm_odm_combine_mode_2to1;
 						mode_lib->vba.PlaneRequiredDISPCLK = mode_lib->vba.PlaneRequiredDISPCLKWithODMCombine;
 					} else if (locals->HActive[k] > DCN20_MAX_420_IMAGE_WIDTH && locals->OutputFormat[k] == dm_420) {
-						locals->ODMCombineEnablePerState[i][k] = true;
+						locals->ODMCombineEnablePerState[i][k] = dm_odm_combine_mode_2to1;
 						mode_lib->vba.PlaneRequiredDISPCLK = mode_lib->vba.PlaneRequiredDISPCLKWithODMCombine;
 					}
 				}
@@ -4071,7 +4068,7 @@ void dml20v2_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode
 				locals->RequiredDISPCLK[i][j] = 0.0;
 				locals->DISPCLK_DPPCLK_Support[i][j] = true;
 				for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) {
-					locals->ODMCombineEnablePerState[i][k] = false;
+					locals->ODMCombineEnablePerState[i][k] = dm_odm_combine_mode_disabled;
 					if (locals->SwathWidthYSingleDPP[k] <= locals->MaximumSwathWidth[k]) {
 						locals->NoOfDPP[i][j][k] = 1;
 						locals->RequiredDPPCLK[i][j][k] = locals->MinDPPCLKUsingSingleDPP[k]
@@ -4443,7 +4440,7 @@ void dml20v2_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode
 					locals->RoundedUpMaxSwathSizeBytesC = 0;
 				}
 
-				if (locals->RoundedUpMaxSwathSizeBytesY + locals->RoundedUpMaxSwathSizeBytesC <= locals->DETBufferSizeInKByte[0] * 1024 / 2) {
+				if (locals->RoundedUpMaxSwathSizeBytesY + locals->RoundedUpMaxSwathSizeBytesC <= locals->DETBufferSizeInKByte[0] * 1024.0 / 2) {
 					locals->SwathHeightYPerState[i][j][k] = locals->MaxSwathHeightY[k];
 					locals->SwathHeightCPerState[i][j][k] = locals->MaxSwathHeightC[k];
 				} else {
@@ -4478,17 +4475,17 @@ void dml20v2_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode
 						locals->EffectiveLBLatencyHidingSourceLinesLuma),
 						locals->SwathHeightYPerState[i][j][k]);
 
-				locals->EffectiveDETLBLinesChroma = dml_floor(locals->LinesInDETChroma + dml_min(
-						locals->LinesInDETChroma * locals->RequiredDISPCLK[i][j] * locals->BytePerPixelInDETC[k] *
-						locals->PSCL_FACTOR_CHROMA[k] / locals->ReturnBWPerState[i][0],
-						locals->EffectiveLBLatencyHidingSourceLinesChroma),
-						locals->SwathHeightCPerState[i][j][k]);
 
 				if (locals->BytePerPixelInDETC[k] == 0) {
 					locals->UrgentLatencySupportUsPerState[i][j][k] = locals->EffectiveDETLBLinesLuma * (locals->HTotal[k] / locals->PixelClock[k])
 							/ locals->VRatio[k] - locals->EffectiveDETLBLinesLuma * locals->SwathWidthYPerState[i][j][k] *
 								dml_ceil(locals->BytePerPixelInDETY[k], 1) / (locals->ReturnBWPerState[i][0] / locals->NoOfDPP[i][j][k]);
 				} else {
+					locals->EffectiveDETLBLinesChroma = dml_floor(locals->LinesInDETChroma + dml_min(
+							locals->LinesInDETChroma * locals->RequiredDISPCLK[i][j] * locals->BytePerPixelInDETC[k] *
+							locals->PSCL_FACTOR_CHROMA[k] / locals->ReturnBWPerState[i][0],
+							locals->EffectiveLBLatencyHidingSourceLinesChroma),
+							locals->SwathHeightCPerState[i][j][k]);
 					locals->UrgentLatencySupportUsPerState[i][j][k] = dml_min(
 						locals->EffectiveDETLBLinesLuma * (locals->HTotal[k] / locals->PixelClock[k])
 						/ locals->VRatio[k] - locals->EffectiveDETLBLinesLuma * locals->SwathWidthYPerState[i][j][k] *
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn20/display_rq_dlg_calc_20.c b/drivers/gpu/drm/amd/display/dc/dml/dcn20/display_rq_dlg_calc_20.c
index 246071c72f6b..9c58ff1069d6 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dcn20/display_rq_dlg_calc_20.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn20/display_rq_dlg_calc_20.c
@@ -78,7 +78,7 @@ static void calculate_ttu_cursor(struct display_mode_lib *mode_lib,
 
 static unsigned int get_bytes_per_element(enum source_format_class source_format, bool is_chroma)
 {
-	unsigned int ret_val = 0;
+	unsigned int ret_val = 1;
 
 	if (source_format == dm_444_16) {
 		if (!is_chroma)
@@ -313,9 +313,6 @@ static void handle_det_buf_split(struct display_mode_lib *mode_lib,
 
 		if (swath_height_c > 0)
 			log2_swath_height_c = dml_log2(swath_height_c);
-
-		if (req128_c && log2_swath_height_c > 0)
-			log2_swath_height_c -= 1;
 	}
 
 	rq_param->dlg.rq_l.swath_height = 1 << log2_swath_height_l;
@@ -446,8 +443,6 @@ static void get_meta_and_pte_attr(struct display_mode_lib *mode_lib,
 	blk_bytes = surf_linear ?
 			256 : get_blk_size_bytes((enum source_macro_tile_size) macro_tile_size);
 	log2_blk_bytes = dml_log2((double) blk_bytes);
-	log2_blk_height = 0;
-	log2_blk_width = 0;
 
 	// remember log rule
 	// "+" in log is multiply
@@ -494,8 +489,6 @@ static void get_meta_and_pte_attr(struct display_mode_lib *mode_lib,
 			- log2_meta_req_height;
 	meta_req_width = 1 << log2_meta_req_width;
 	meta_req_height = 1 << log2_meta_req_height;
-	log2_meta_row_height = 0;
-	meta_row_width_ub = 0;
 
 	// the dimensions of a meta row are meta_row_width x meta_row_height in elements.
 	// calculate upper bound of the meta_row_width
@@ -653,7 +646,7 @@ static void get_meta_and_pte_attr(struct display_mode_lib *mode_lib,
 
 	// the dpte_group_bytes is reduced for the specific case of vertical
 	// access of a tile surface that has dpte request of 8x1 ptes.
-	if (!surf_linear & (log2_dpte_req_height_ptes == 0) & surf_vert) //reduced, in this case, will have page fault within a group
+	if (!surf_linear && (log2_dpte_req_height_ptes == 0) && surf_vert) //reduced, in this case, will have page fault within a group
 		rq_sizing_param->dpte_group_bytes = 512;
 	else
 		//full size
@@ -688,12 +681,11 @@ static void get_surf_rq_param(struct display_mode_lib *mode_lib,
 		const display_pipe_source_params_st *pipe_src_param,
 		bool is_chroma)
 {
-	bool mode_422 = false;
 	unsigned int vp_width = 0;
 	unsigned int vp_height = 0;
 	unsigned int data_pitch = 0;
 	unsigned int meta_pitch = 0;
-	unsigned int ppe = mode_422 ? 2 : 1;
+	unsigned int ppe = 1;
 
 	// TODO check if ppe apply for both luma and chroma in 422 case
 	if (is_chroma) {
@@ -825,7 +817,6 @@ static void dml20_rq_dlg_get_dlg_params(struct display_mode_lib *mode_lib,
 	double min_dst_y_ttu_vblank;
 	unsigned int dlg_vblank_start;
 	bool dual_plane;
-	bool mode_422;
 	unsigned int access_dir;
 	unsigned int vp_height_l;
 	unsigned int vp_width_l;
@@ -971,7 +962,6 @@ static void dml20_rq_dlg_get_dlg_params(struct display_mode_lib *mode_lib,
 	// Source
 //             dcc_en              = src.dcc;
 	dual_plane = is_dual_plane((enum source_format_class)(src->source_format));
-	mode_422 = false; // TODO
 	access_dir = (src->source_scan == dm_vert); // vp access direction: horizontal or vertical accessed
 //      bytes_per_element_l = get_bytes_per_element(source_format_class(src.source_format), 0);
 //      bytes_per_element_c = get_bytes_per_element(source_format_class(src.source_format), 1);
@@ -1148,18 +1138,8 @@ static void dml20_rq_dlg_get_dlg_params(struct display_mode_lib *mode_lib,
 	dpte_row_height_l = rq_dlg_param->rq_l.dpte_row_height;
 	dpte_row_height_c = rq_dlg_param->rq_c.dpte_row_height;
 
-	if (mode_422) {
-		swath_width_pixels_ub_l = swath_width_ub_l * 2;  // *2 for 2 pixel per element
-		swath_width_pixels_ub_c = swath_width_ub_c * 2;
-	} else {
-		swath_width_pixels_ub_l = swath_width_ub_l * 1;
-		swath_width_pixels_ub_c = swath_width_ub_c * 1;
-	}
-
-	hscale_pixel_rate_l = 0.;
-	hscale_pixel_rate_c = 0.;
-	min_hratio_fact_l = 1.0;
-	min_hratio_fact_c = 1.0;
+	swath_width_pixels_ub_l = swath_width_ub_l;
+	swath_width_pixels_ub_c = swath_width_ub_c;
 
 	if (htaps_l <= 1)
 		min_hratio_fact_l = 2.0;
@@ -1576,8 +1556,6 @@ void dml20_rq_dlg_get_dlg_reg(struct display_mode_lib *mode_lib,
 	dlg_sys_param.total_flip_bytes = get_total_immediate_flip_bytes(mode_lib,
 			e2e_pipe_param,
 			num_pipes);
-	dlg_sys_param.t_srx_delay_us = mode_lib->ip.dcfclk_cstate_latency
-			/ dlg_sys_param.deepsleep_dcfclk_mhz; // TODO: Deprecated
 
 	print__dlg_sys_params_st(mode_lib, &dlg_sys_param);
 
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn20/display_rq_dlg_calc_20v2.c b/drivers/gpu/drm/amd/display/dc/dml/dcn20/display_rq_dlg_calc_20v2.c
index 015e7f2c0b16..570e6e39eb45 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dcn20/display_rq_dlg_calc_20v2.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn20/display_rq_dlg_calc_20v2.c
@@ -78,7 +78,7 @@ static void calculate_ttu_cursor(struct display_mode_lib *mode_lib,
 
 static unsigned int get_bytes_per_element(enum source_format_class source_format, bool is_chroma)
 {
-	unsigned int ret_val = 0;
+	unsigned int ret_val = 1;
 
 	if (source_format == dm_444_16) {
 		if (!is_chroma)
@@ -313,9 +313,6 @@ static void handle_det_buf_split(struct display_mode_lib *mode_lib,
 
 		if (swath_height_c > 0)
 			log2_swath_height_c = dml_log2(swath_height_c);
-
-		if (req128_c && log2_swath_height_c > 0)
-			log2_swath_height_c -= 1;
 	}
 
 	rq_param->dlg.rq_l.swath_height = 1 << log2_swath_height_l;
@@ -446,8 +443,6 @@ static void get_meta_and_pte_attr(struct display_mode_lib *mode_lib,
 	blk_bytes = surf_linear ?
 			256 : get_blk_size_bytes((enum source_macro_tile_size) macro_tile_size);
 	log2_blk_bytes = dml_log2((double) blk_bytes);
-	log2_blk_height = 0;
-	log2_blk_width = 0;
 
 	// remember log rule
 	// "+" in log is multiply
@@ -494,8 +489,6 @@ static void get_meta_and_pte_attr(struct display_mode_lib *mode_lib,
 			- log2_meta_req_height;
 	meta_req_width = 1 << log2_meta_req_width;
 	meta_req_height = 1 << log2_meta_req_height;
-	log2_meta_row_height = 0;
-	meta_row_width_ub = 0;
 
 	// the dimensions of a meta row are meta_row_width x meta_row_height in elements.
 	// calculate upper bound of the meta_row_width
@@ -653,7 +646,7 @@ static void get_meta_and_pte_attr(struct display_mode_lib *mode_lib,
 
 	// the dpte_group_bytes is reduced for the specific case of vertical
 	// access of a tile surface that has dpte request of 8x1 ptes.
-	if (!surf_linear & (log2_dpte_req_height_ptes == 0) & surf_vert) //reduced, in this case, will have page fault within a group
+	if (!surf_linear && (log2_dpte_req_height_ptes == 0) && surf_vert) //reduced, in this case, will have page fault within a group
 		rq_sizing_param->dpte_group_bytes = 512;
 	else
 		//full size
@@ -688,12 +681,11 @@ static void get_surf_rq_param(struct display_mode_lib *mode_lib,
 		const display_pipe_source_params_st *pipe_src_param,
 		bool is_chroma)
 {
-	bool mode_422 = false;
 	unsigned int vp_width = 0;
 	unsigned int vp_height = 0;
 	unsigned int data_pitch = 0;
 	unsigned int meta_pitch = 0;
-	unsigned int ppe = mode_422 ? 2 : 1;
+	unsigned int ppe = 1;
 
 	// TODO check if ppe apply for both luma and chroma in 422 case
 	if (is_chroma) {
@@ -825,7 +817,6 @@ static void dml20v2_rq_dlg_get_dlg_params(struct display_mode_lib *mode_lib,
 	double min_dst_y_ttu_vblank;
 	unsigned int dlg_vblank_start;
 	bool dual_plane;
-	bool mode_422;
 	unsigned int access_dir;
 	unsigned int vp_height_l;
 	unsigned int vp_width_l;
@@ -972,7 +963,6 @@ static void dml20v2_rq_dlg_get_dlg_params(struct display_mode_lib *mode_lib,
 	// Source
 //             dcc_en              = src.dcc;
 	dual_plane = is_dual_plane((enum source_format_class)(src->source_format));
-	mode_422 = false; // TODO
 	access_dir = (src->source_scan == dm_vert); // vp access direction: horizontal or vertical accessed
 //      bytes_per_element_l = get_bytes_per_element(source_format_class(src.source_format), 0);
 //      bytes_per_element_c = get_bytes_per_element(source_format_class(src.source_format), 1);
@@ -1149,18 +1139,8 @@ static void dml20v2_rq_dlg_get_dlg_params(struct display_mode_lib *mode_lib,
 	dpte_row_height_l = rq_dlg_param->rq_l.dpte_row_height;
 	dpte_row_height_c = rq_dlg_param->rq_c.dpte_row_height;
 
-	if (mode_422) {
-		swath_width_pixels_ub_l = swath_width_ub_l * 2;  // *2 for 2 pixel per element
-		swath_width_pixels_ub_c = swath_width_ub_c * 2;
-	} else {
-		swath_width_pixels_ub_l = swath_width_ub_l * 1;
-		swath_width_pixels_ub_c = swath_width_ub_c * 1;
-	}
-
-	hscale_pixel_rate_l = 0.;
-	hscale_pixel_rate_c = 0.;
-	min_hratio_fact_l = 1.0;
-	min_hratio_fact_c = 1.0;
+	swath_width_pixels_ub_l = swath_width_ub_l;
+	swath_width_pixels_ub_c = swath_width_ub_c;
 
 	if (htaps_l <= 1)
 		min_hratio_fact_l = 2.0;
@@ -1577,8 +1557,6 @@ void dml20v2_rq_dlg_get_dlg_reg(struct display_mode_lib *mode_lib,
 	dlg_sys_param.total_flip_bytes = get_total_immediate_flip_bytes(mode_lib,
 			e2e_pipe_param,
 			num_pipes);
-	dlg_sys_param.t_srx_delay_us = mode_lib->ip.dcfclk_cstate_latency
-			/ dlg_sys_param.deepsleep_dcfclk_mhz; // TODO: Deprecated
 
 	print__dlg_sys_params_st(mode_lib, &dlg_sys_param);
 
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn21/display_mode_vba_21.c b/drivers/gpu/drm/amd/display/dc/dml/dcn21/display_mode_vba_21.c
index 8a7485e21d53..cd8cca651419 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dcn21/display_mode_vba_21.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn21/display_mode_vba_21.c
@@ -806,10 +806,12 @@ static bool CalculatePrefetchSchedule(
 
 	if (myPipe->SourceScan == dm_horz) {
 		*swath_width_luma_ub = dml_ceil(SwathWidthY - 1, myPipe->BlockWidth256BytesY) + myPipe->BlockWidth256BytesY;
-		*swath_width_chroma_ub = dml_ceil(SwathWidthY / 2 - 1, myPipe->BlockWidth256BytesC) + myPipe->BlockWidth256BytesC;
+		if (myPipe->BlockWidth256BytesC > 0)
+			*swath_width_chroma_ub = dml_ceil(SwathWidthY / 2 - 1, myPipe->BlockWidth256BytesC) + myPipe->BlockWidth256BytesC;
 	} else {
 		*swath_width_luma_ub = dml_ceil(SwathWidthY - 1, myPipe->BlockHeight256BytesY) + myPipe->BlockHeight256BytesY;
-		*swath_width_chroma_ub = dml_ceil(SwathWidthY / 2 - 1, myPipe->BlockHeight256BytesC) + myPipe->BlockHeight256BytesC;
+		if (myPipe->BlockHeight256BytesC > 0)
+			*swath_width_chroma_ub = dml_ceil(SwathWidthY / 2 - 1, myPipe->BlockHeight256BytesC) + myPipe->BlockHeight256BytesC;
 	}
 
 	prefetch_bw_oto = (PrefetchSourceLinesY * *swath_width_luma_ub * dml_ceil(BytePerPixelDETY, 1) + PrefetchSourceLinesC * *swath_width_chroma_ub * dml_ceil(BytePerPixelDETC, 2)) / Tsw_oto;
@@ -1047,12 +1049,9 @@ static bool CalculatePrefetchSchedule(
 
 	if (MyError) {
 		*PrefetchBandwidth = 0;
-		TimeForFetchingMetaPTE = 0;
-		TimeForFetchingRowInVBlank = 0;
 		*DestinationLinesToRequestVMInVBlank = 0;
 		*DestinationLinesToRequestRowInVBlank = 0;
 		*DestinationLinesForPrefetch = 0;
-		LinesToRequestPrefetchPixelData = 0;
 		*VRatioPrefetchY = 0;
 		*VRatioPrefetchC = 0;
 		*RequiredPrefetchPixDataBWLuma = 0;
@@ -1397,7 +1396,7 @@ static unsigned int CalculateVMAndRowBytes(
 			if (ScanDirection == dm_horz)
 				FractionOfPTEReturnDrop = 0;
 			else
-				FractionOfPTEReturnDrop = 7 / 8;
+				FractionOfPTEReturnDrop = 7.0 / 8;
 		} else if (VMMPageSize == 4096 && MacroTileSizeBytes > 4096) {
 			PixelPTEReqHeightPTEs = 16;
 			*PixelPTEReqHeight = 16 * BlockHeight256Bytes;
@@ -2634,7 +2633,7 @@ static void DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerforman
 					&mode_lib->vba.SrcActiveDrainRate,
 					&mode_lib->vba.TInitXFill,
 					&mode_lib->vba.TslvChk);
-					locals->XFCRemoteSurfaceFlipLatency[k] =
+			locals->XFCRemoteSurfaceFlipLatency[k] =
 					dml_floor(
 							mode_lib->vba.XFCRemoteSurfaceFlipDelay
 									/ (mode_lib->vba.HTotal[k]
@@ -3192,7 +3191,7 @@ static void CalculateFlipSchedule(
 	unsigned int HostVMDynamicLevels;
 	double TimeForFetchingMetaPTEImmediateFlip;
 	double TimeForFetchingRowInVBlankImmediateFlip;
-	double ImmediateFlipBW;
+	double ImmediateFlipBW = 1.0;
 	double HostVMInefficiencyFactor;
 	double VRatioClamped;
 
@@ -4100,17 +4099,17 @@ void dml21_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 					mode_lib->vba.PlaneRequiredDISPCLKWithODMCombine = mode_lib->vba.PixelClock[k] / 2
 							* (1 + mode_lib->vba.DISPCLKDPPCLKDSCCLKDownSpreading / 100.0);
 
-				locals->ODMCombineEnablePerState[i][k] = false;
+				locals->ODMCombineEnablePerState[i][k] = dm_odm_combine_mode_disabled;
 				mode_lib->vba.PlaneRequiredDISPCLK = mode_lib->vba.PlaneRequiredDISPCLKWithoutODMCombine;
 				if (mode_lib->vba.ODMCapability) {
 					if (locals->PlaneRequiredDISPCLKWithoutODMCombine > MaxMaxDispclkRoundedDown) {
-						locals->ODMCombineEnablePerState[i][k] = true;
+						locals->ODMCombineEnablePerState[i][k] = dm_odm_combine_mode_2to1;
 						mode_lib->vba.PlaneRequiredDISPCLK = mode_lib->vba.PlaneRequiredDISPCLKWithODMCombine;
 					} else if (locals->DSCEnabled[k] && (locals->HActive[k] > DCN21_MAX_DSC_IMAGE_WIDTH)) {
-						locals->ODMCombineEnablePerState[i][k] = true;
+						locals->ODMCombineEnablePerState[i][k] = dm_odm_combine_mode_2to1;
 						mode_lib->vba.PlaneRequiredDISPCLK = mode_lib->vba.PlaneRequiredDISPCLKWithODMCombine;
 					} else if (locals->HActive[k] > DCN21_MAX_420_IMAGE_WIDTH && locals->OutputFormat[k] == dm_420) {
-						locals->ODMCombineEnablePerState[i][k] = true;
+						locals->ODMCombineEnablePerState[i][k] = dm_odm_combine_mode_2to1;
 						mode_lib->vba.PlaneRequiredDISPCLK = mode_lib->vba.PlaneRequiredDISPCLKWithODMCombine;
 					}
 				}
@@ -4163,7 +4162,7 @@ void dml21_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 				locals->RequiredDISPCLK[i][j] = 0.0;
 				locals->DISPCLK_DPPCLK_Support[i][j] = true;
 				for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) {
-					locals->ODMCombineEnablePerState[i][k] = false;
+					locals->ODMCombineEnablePerState[i][k] = dm_odm_combine_mode_disabled;
 					if (locals->SwathWidthYSingleDPP[k] <= locals->MaximumSwathWidth[k]) {
 						locals->NoOfDPP[i][j][k] = 1;
 						locals->RequiredDPPCLK[i][j][k] = locals->MinDPPCLKUsingSingleDPP[k]
@@ -5228,7 +5227,7 @@ void dml21_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 			mode_lib->vba.ODMCombineEnabled[k] =
 					locals->ODMCombineEnablePerState[mode_lib->vba.VoltageLevel][k];
 		} else {
-			mode_lib->vba.ODMCombineEnabled[k] = false;
+			mode_lib->vba.ODMCombineEnabled[k] = dm_odm_combine_mode_disabled;
 		}
 		mode_lib->vba.DSCEnabled[k] =
 				locals->RequiresDSC[mode_lib->vba.VoltageLevel][k];
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn21/display_rq_dlg_calc_21.c b/drivers/gpu/drm/amd/display/dc/dml/dcn21/display_rq_dlg_calc_21.c
index 46c433c0bcb0..f549da082c01 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dcn21/display_rq_dlg_calc_21.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn21/display_rq_dlg_calc_21.c
@@ -53,7 +53,7 @@ static void calculate_ttu_cursor(
 
 static unsigned int get_bytes_per_element(enum source_format_class source_format, bool is_chroma)
 {
-	unsigned int ret_val = 0;
+	unsigned int ret_val = 1;
 
 	if (source_format == dm_444_16) {
 		if (!is_chroma)
@@ -297,9 +297,6 @@ static void handle_det_buf_split(
 
 		if (swath_height_c > 0)
 			log2_swath_height_c = dml_log2(swath_height_c);
-
-		if (req128_c && log2_swath_height_c > 0)
-			log2_swath_height_c -= 1;
 	}
 
 	rq_param->dlg.rq_l.swath_height = 1 << log2_swath_height_l;
@@ -438,8 +435,6 @@ static void get_meta_and_pte_attr(
 	blk_bytes = surf_linear ?
 			256 : get_blk_size_bytes((enum source_macro_tile_size) macro_tile_size);
 	log2_blk_bytes = dml_log2((double) blk_bytes);
-	log2_blk_height = 0;
-	log2_blk_width = 0;
 
 	// remember log rule
 	// "+" in log is multiply
@@ -488,8 +483,6 @@ static void get_meta_and_pte_attr(
 			- log2_meta_req_height;
 	meta_req_width = 1 << log2_meta_req_width;
 	meta_req_height = 1 << log2_meta_req_height;
-	log2_meta_row_height = 0;
-	meta_row_width_ub = 0;
 
 	// the dimensions of a meta row are meta_row_width x meta_row_height in elements.
 	// calculate upper bound of the meta_row_width
@@ -659,7 +652,7 @@ static void get_meta_and_pte_attr(
 	if (hostvm_enable)
 		rq_sizing_param->dpte_group_bytes = 512;
 	else {
-		if (!surf_linear & (log2_dpte_req_height_ptes == 0) & surf_vert) //reduced, in this case, will have page fault within a group
+		if (!surf_linear && (log2_dpte_req_height_ptes == 0) && surf_vert) //reduced, in this case, will have page fault within a group
 			rq_sizing_param->dpte_group_bytes = 512;
 		else
 			//full size
@@ -697,12 +690,11 @@ static void get_surf_rq_param(
 		const display_pipe_params_st *pipe_param,
 		bool is_chroma)
 {
-	bool mode_422 = false;
 	unsigned int vp_width = 0;
 	unsigned int vp_height = 0;
 	unsigned int data_pitch = 0;
 	unsigned int meta_pitch = 0;
-	unsigned int ppe = mode_422 ? 2 : 1;
+	unsigned int ppe = 1;
 
 	// FIXME check if ppe apply for both luma and chroma in 422 case
 	if (is_chroma) {
@@ -871,7 +863,6 @@ static void dml_rq_dlg_get_dlg_params(
 	double min_dst_y_ttu_vblank;
 	unsigned int dlg_vblank_start;
 	bool dual_plane;
-	bool mode_422;
 	unsigned int access_dir;
 	unsigned int vp_height_l;
 	unsigned int vp_width_l;
@@ -1023,7 +1014,6 @@ static void dml_rq_dlg_get_dlg_params(
 	// Source
 	//             dcc_en              = src.dcc;
 	dual_plane = is_dual_plane((enum source_format_class) (src->source_format));
-	mode_422 = false; // FIXME
 	access_dir = (src->source_scan == dm_vert); // vp access direction: horizontal or vertical accessed
 						    //      bytes_per_element_l = get_bytes_per_element(source_format_class(src.source_format), 0);
 						    //      bytes_per_element_c = get_bytes_per_element(source_format_class(src.source_format), 1);
@@ -1200,18 +1190,8 @@ static void dml_rq_dlg_get_dlg_params(
 	dpte_row_height_l = rq_dlg_param->rq_l.dpte_row_height;
 	dpte_row_height_c = rq_dlg_param->rq_c.dpte_row_height;
 
-	if (mode_422) {
-		swath_width_pixels_ub_l = swath_width_ub_l * 2;  // *2 for 2 pixel per element
-		swath_width_pixels_ub_c = swath_width_ub_c * 2;
-	} else {
-		swath_width_pixels_ub_l = swath_width_ub_l * 1;
-		swath_width_pixels_ub_c = swath_width_ub_c * 1;
-	}
-
-	hscale_pixel_rate_l = 0.;
-	hscale_pixel_rate_c = 0.;
-	min_hratio_fact_l = 1.0;
-	min_hratio_fact_c = 1.0;
+	swath_width_pixels_ub_l = swath_width_ub_l;
+	swath_width_pixels_ub_c = swath_width_ub_c;
 
 	if (hratio_l <= 1)
 		min_hratio_fact_l = 2.0;
@@ -1688,8 +1668,6 @@ void dml21_rq_dlg_get_dlg_reg(
 			mode_lib,
 			e2e_pipe_param,
 			num_pipes);
-	dlg_sys_param.t_srx_delay_us = mode_lib->ip.dcfclk_cstate_latency
-			/ dlg_sys_param.deepsleep_dcfclk_mhz; // TODO: Deprecated
 
 	print__dlg_sys_params_st(mode_lib, &dlg_sys_param);
 
@@ -1711,14 +1689,6 @@ void dml21_rq_dlg_get_dlg_reg(
 	dml_print("DML_DLG: Calculation for pipe[%d] end\n", pipe_idx);
 }
 
-void dml_rq_dlg_get_arb_params(struct display_mode_lib *mode_lib, display_arb_params_st *arb_param)
-{
-	memset(arb_param, 0, sizeof(*arb_param));
-	arb_param->max_req_outstanding = 256;
-	arb_param->min_req_outstanding = 68;
-	arb_param->sat_level_us = 60;
-}
-
 static void calculate_ttu_cursor(
 		struct display_mode_lib *mode_lib,
 		double *refcyc_per_req_delivery_pre_cur,
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn30/dcn30_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn30/dcn30_fpu.c
new file mode 100644
index 000000000000..e5f5c0663750
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn30/dcn30_fpu.c
@@ -0,0 +1,739 @@
+/*
+ * Copyright 2020-2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+#include "resource.h"
+#include "clk_mgr.h"
+#include "reg_helper.h"
+#include "dcn_calc_math.h"
+#include "dcn20/dcn20_resource.h"
+#include "dcn30/dcn30_resource.h"
+
+#include "clk_mgr/dcn30/dcn30_smu11_driver_if.h"
+#include "display_mode_vba_30.h"
+#include "dcn30_fpu.h"
+
+#define REG(reg)\
+	optc1->tg_regs->reg
+
+#define CTX \
+	optc1->base.ctx
+
+#undef FN
+#define FN(reg_name, field_name) \
+	optc1->tg_shift->field_name, optc1->tg_mask->field_name
+
+
+struct _vcs_dpi_ip_params_st dcn3_0_ip = {
+	.use_min_dcfclk = 0,
+	.clamp_min_dcfclk = 0,
+	.odm_capable = 1,
+	.gpuvm_enable = 0,
+	.hostvm_enable = 0,
+	.gpuvm_max_page_table_levels = 4,
+	.hostvm_max_page_table_levels = 4,
+	.hostvm_cached_page_table_levels = 0,
+	.pte_group_size_bytes = 2048,
+	.num_dsc = 6,
+	.rob_buffer_size_kbytes = 184,
+	.det_buffer_size_kbytes = 184,
+	.dpte_buffer_size_in_pte_reqs_luma = 84,
+	.pde_proc_buffer_size_64k_reqs = 48,
+	.dpp_output_buffer_pixels = 2560,
+	.opp_output_buffer_lines = 1,
+	.pixel_chunk_size_kbytes = 8,
+	.pte_enable = 1,
+	.max_page_table_levels = 2,
+	.pte_chunk_size_kbytes = 2,  // ?
+	.meta_chunk_size_kbytes = 2,
+	.writeback_chunk_size_kbytes = 8,
+	.line_buffer_size_bits = 789504,
+	.is_line_buffer_bpp_fixed = 0,  // ?
+	.line_buffer_fixed_bpp = 0,     // ?
+	.dcc_supported = true,
+	.writeback_interface_buffer_size_kbytes = 90,
+	.writeback_line_buffer_buffer_size = 0,
+	.max_line_buffer_lines = 12,
+	.writeback_luma_buffer_size_kbytes = 12,  // writeback_line_buffer_buffer_size = 656640
+	.writeback_chroma_buffer_size_kbytes = 8,
+	.writeback_chroma_line_buffer_width_pixels = 4,
+	.writeback_max_hscl_ratio = 1,
+	.writeback_max_vscl_ratio = 1,
+	.writeback_min_hscl_ratio = 1,
+	.writeback_min_vscl_ratio = 1,
+	.writeback_max_hscl_taps = 1,
+	.writeback_max_vscl_taps = 1,
+	.writeback_line_buffer_luma_buffer_size = 0,
+	.writeback_line_buffer_chroma_buffer_size = 14643,
+	.cursor_buffer_size = 8,
+	.cursor_chunk_size = 2,
+	.max_num_otg = 6,
+	.max_num_dpp = 6,
+	.max_num_wb = 1,
+	.max_dchub_pscl_bw_pix_per_clk = 4,
+	.max_pscl_lb_bw_pix_per_clk = 2,
+	.max_lb_vscl_bw_pix_per_clk = 4,
+	.max_vscl_hscl_bw_pix_per_clk = 4,
+	.max_hscl_ratio = 6,
+	.max_vscl_ratio = 6,
+	.hscl_mults = 4,
+	.vscl_mults = 4,
+	.max_hscl_taps = 8,
+	.max_vscl_taps = 8,
+	.dispclk_ramp_margin_percent = 1,
+	.underscan_factor = 1.11,
+	.min_vblank_lines = 32,
+	.dppclk_delay_subtotal = 46,
+	.dynamic_metadata_vm_enabled = true,
+	.dppclk_delay_scl_lb_only = 16,
+	.dppclk_delay_scl = 50,
+	.dppclk_delay_cnvc_formatter = 27,
+	.dppclk_delay_cnvc_cursor = 6,
+	.dispclk_delay_subtotal = 119,
+	.dcfclk_cstate_latency = 5.2, // SRExitTime
+	.max_inter_dcn_tile_repeaters = 8,
+	.max_num_hdmi_frl_outputs = 1,
+	.odm_combine_4to1_supported = true,
+
+	.xfc_supported = false,
+	.xfc_fill_bw_overhead_percent = 10.0,
+	.xfc_fill_constant_bytes = 0,
+	.gfx7_compat_tiling_supported = 0,
+	.number_of_cursors = 1,
+};
+
+struct _vcs_dpi_soc_bounding_box_st dcn3_0_soc = {
+	.clock_limits = {
+			{
+				.state = 0,
+				.dispclk_mhz = 562.0,
+				.dppclk_mhz = 300.0,
+				.phyclk_mhz = 300.0,
+				.phyclk_d18_mhz = 667.0,
+				.dscclk_mhz = 405.6,
+			},
+		},
+
+	.min_dcfclk = 500.0, /* TODO: set this to actual min DCFCLK */
+	.num_states = 1,
+	.sr_exit_time_us = 15.5,
+	.sr_enter_plus_exit_time_us = 20,
+	.urgent_latency_us = 4.0,
+	.urgent_latency_pixel_data_only_us = 4.0,
+	.urgent_latency_pixel_mixed_with_vm_data_us = 4.0,
+	.urgent_latency_vm_data_only_us = 4.0,
+	.urgent_out_of_order_return_per_channel_pixel_only_bytes = 4096,
+	.urgent_out_of_order_return_per_channel_pixel_and_vm_bytes = 4096,
+	.urgent_out_of_order_return_per_channel_vm_only_bytes = 4096,
+	.pct_ideal_dram_sdp_bw_after_urgent_pixel_only = 80.0,
+	.pct_ideal_dram_sdp_bw_after_urgent_pixel_and_vm = 60.0,
+	.pct_ideal_dram_sdp_bw_after_urgent_vm_only = 40.0,
+	.max_avg_sdp_bw_use_normal_percent = 60.0,
+	.max_avg_dram_bw_use_normal_percent = 40.0,
+	.writeback_latency_us = 12.0,
+	.max_request_size_bytes = 256,
+	.fabric_datapath_to_dcn_data_return_bytes = 64,
+	.dcn_downspread_percent = 0.5,
+	.downspread_percent = 0.38,
+	.dram_page_open_time_ns = 50.0,
+	.dram_rw_turnaround_time_ns = 17.5,
+	.dram_return_buffer_per_channel_bytes = 8192,
+	.round_trip_ping_latency_dcfclk_cycles = 191,
+	.urgent_out_of_order_return_per_channel_bytes = 4096,
+	.channel_interleave_bytes = 256,
+	.num_banks = 8,
+	.gpuvm_min_page_size_bytes = 4096,
+	.hostvm_min_page_size_bytes = 4096,
+	.dram_clock_change_latency_us = 404,
+	.dummy_pstate_latency_us = 5,
+	.writeback_dram_clock_change_latency_us = 23.0,
+	.return_bus_width_bytes = 64,
+	.dispclk_dppclk_vco_speed_mhz = 3650,
+	.xfc_bus_transport_time_us = 20,      // ?
+	.xfc_xbuf_latency_tolerance_us = 4,  // ?
+	.use_urgent_burst_bw = 1,            // ?
+	.do_urgent_latency_adjustment = true,
+	.urgent_latency_adjustment_fabric_clock_component_us = 1.0,
+	.urgent_latency_adjustment_fabric_clock_reference_mhz = 1000,
+};
+
+
+void dcn30_fpu_populate_dml_writeback_from_context(
+		struct dc *dc, struct resource_context *res_ctx, display_e2e_pipe_params_st *pipes)
+{
+	int pipe_cnt, i, j;
+	double max_calc_writeback_dispclk;
+	double writeback_dispclk;
+	struct writeback_st dout_wb = {0};
+
+	dc_assert_fp_enabled();
+
+	for (i = 0, pipe_cnt = 0; i < dc->res_pool->pipe_count; i++) {
+		struct dc_stream_state *stream = res_ctx->pipe_ctx[i].stream;
+
+		if (!stream)
+			continue;
+		max_calc_writeback_dispclk = 0;
+
+		/* Set writeback information */
+		pipes[pipe_cnt].dout.wb_enable = 0;
+		pipes[pipe_cnt].dout.num_active_wb = 0;
+		for (j = 0; j < stream->num_wb_info; j++) {
+			struct dc_writeback_info *wb_info = &stream->writeback_info[j];
+
+			if (wb_info->wb_enabled && wb_info->writeback_source_plane &&
+					(wb_info->writeback_source_plane == res_ctx->pipe_ctx[i].plane_state)) {
+				pipes[pipe_cnt].dout.wb_enable = 1;
+				pipes[pipe_cnt].dout.num_active_wb++;
+				dout_wb.wb_src_height = wb_info->dwb_params.cnv_params.crop_en ?
+					wb_info->dwb_params.cnv_params.crop_height :
+					wb_info->dwb_params.cnv_params.src_height;
+				dout_wb.wb_src_width = wb_info->dwb_params.cnv_params.crop_en ?
+					wb_info->dwb_params.cnv_params.crop_width :
+					wb_info->dwb_params.cnv_params.src_width;
+				dout_wb.wb_dst_width = wb_info->dwb_params.dest_width;
+				dout_wb.wb_dst_height = wb_info->dwb_params.dest_height;
+
+				/* For IP that doesn't support WB scaling, set h/v taps to 1 to avoid DML validation failure */
+				if (dc->dml.ip.writeback_max_hscl_taps > 1) {
+					dout_wb.wb_htaps_luma = wb_info->dwb_params.scaler_taps.h_taps;
+					dout_wb.wb_vtaps_luma = wb_info->dwb_params.scaler_taps.v_taps;
+				} else {
+					dout_wb.wb_htaps_luma = 1;
+					dout_wb.wb_vtaps_luma = 1;
+				}
+				dout_wb.wb_htaps_chroma = 0;
+				dout_wb.wb_vtaps_chroma = 0;
+				dout_wb.wb_hratio = wb_info->dwb_params.cnv_params.crop_en ?
+					(double)wb_info->dwb_params.cnv_params.crop_width /
+						(double)wb_info->dwb_params.dest_width :
+					(double)wb_info->dwb_params.cnv_params.src_width /
+						(double)wb_info->dwb_params.dest_width;
+				dout_wb.wb_vratio = wb_info->dwb_params.cnv_params.crop_en ?
+					(double)wb_info->dwb_params.cnv_params.crop_height /
+						(double)wb_info->dwb_params.dest_height :
+					(double)wb_info->dwb_params.cnv_params.src_height /
+						(double)wb_info->dwb_params.dest_height;
+				if (wb_info->dwb_params.cnv_params.fc_out_format == DWB_OUT_FORMAT_64BPP_ARGB ||
+					wb_info->dwb_params.cnv_params.fc_out_format == DWB_OUT_FORMAT_64BPP_RGBA)
+					dout_wb.wb_pixel_format = dm_444_64;
+				else
+					dout_wb.wb_pixel_format = dm_444_32;
+
+				/* Workaround for cases where multiple writebacks are connected to same plane
+				 * In which case, need to compute worst case and set the associated writeback parameters
+				 * This workaround is necessary due to DML computation assuming only 1 set of writeback
+				 * parameters per pipe
+				 */
+				writeback_dispclk = dml30_CalculateWriteBackDISPCLK(
+						dout_wb.wb_pixel_format,
+						pipes[pipe_cnt].pipe.dest.pixel_rate_mhz,
+						dout_wb.wb_hratio,
+						dout_wb.wb_vratio,
+						dout_wb.wb_htaps_luma,
+						dout_wb.wb_vtaps_luma,
+						dout_wb.wb_src_width,
+						dout_wb.wb_dst_width,
+						pipes[pipe_cnt].pipe.dest.htotal,
+						dc->current_state->bw_ctx.dml.ip.writeback_line_buffer_buffer_size);
+
+				if (writeback_dispclk > max_calc_writeback_dispclk) {
+					max_calc_writeback_dispclk = writeback_dispclk;
+					pipes[pipe_cnt].dout.wb = dout_wb;
+				}
+			}
+		}
+
+		pipe_cnt++;
+	}
+}
+
+void dcn30_fpu_set_mcif_arb_params(struct mcif_arb_params *wb_arb_params,
+	struct display_mode_lib *dml,
+	display_e2e_pipe_params_st *pipes,
+	int pipe_cnt,
+	int cur_pipe)
+{
+	int i;
+
+	dc_assert_fp_enabled();
+
+	for (i = 0; i < ARRAY_SIZE(wb_arb_params->cli_watermark); i++) {
+		wb_arb_params->cli_watermark[i] = get_wm_writeback_urgent(dml, pipes, pipe_cnt) * 1000;
+		wb_arb_params->pstate_watermark[i] = get_wm_writeback_dram_clock_change(dml, pipes, pipe_cnt) * 1000;
+	}
+
+	wb_arb_params->dram_speed_change_duration = dml->vba.WritebackAllowDRAMClockChangeEndPosition[cur_pipe] * pipes[0].clks_cfg.refclk_mhz; /* num_clock_cycles = us * MHz */
+}
+
+void dcn30_fpu_update_soc_for_wm_a(struct dc *dc, struct dc_state *context)
+{
+
+	dc_assert_fp_enabled();
+
+	if (dc->clk_mgr->bw_params->wm_table.nv_entries[WM_A].valid) {
+		if (!context->bw_ctx.bw.dcn.clk.fw_based_mclk_switching ||
+				context->bw_ctx.dml.soc.dram_clock_change_latency_us == 0)
+			context->bw_ctx.dml.soc.dram_clock_change_latency_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_A].dml_input.pstate_latency_us;
+		context->bw_ctx.dml.soc.sr_enter_plus_exit_time_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_A].dml_input.sr_enter_plus_exit_time_us;
+		context->bw_ctx.dml.soc.sr_exit_time_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_A].dml_input.sr_exit_time_us;
+	}
+}
+
+void dcn30_fpu_calculate_wm_and_dlg(
+		struct dc *dc, struct dc_state *context,
+		display_e2e_pipe_params_st *pipes,
+		int pipe_cnt,
+		int vlevel)
+{
+	int maxMpcComb = context->bw_ctx.dml.vba.maxMpcComb;
+	int i, pipe_idx;
+	double dcfclk = context->bw_ctx.dml.vba.DCFCLKState[vlevel][maxMpcComb];
+	bool pstate_en = context->bw_ctx.dml.vba.DRAMClockChangeSupport[vlevel][maxMpcComb] != dm_dram_clock_change_unsupported;
+	unsigned int dummy_latency_index = 0;
+	struct dc_stream_status *stream_status = NULL;
+
+	dc_assert_fp_enabled();
+
+	context->bw_ctx.bw.dcn.clk.fw_based_mclk_switching = false;
+	for (i = 0; i < context->stream_count; i++) {
+		stream_status = NULL;
+		if (context->streams[i])
+			stream_status = dc_state_get_stream_status(context, context->streams[i]);
+		if (stream_status)
+			stream_status->fpo_in_use = false;
+	}
+
+	if (!pstate_en) {
+		/* only when the mclk switch can not be natural, is the fw based vblank stretch attempted */
+		context->bw_ctx.bw.dcn.clk.fw_based_mclk_switching =
+			dcn30_can_support_mclk_switch_using_fw_based_vblank_stretch(dc, context);
+
+		if (context->bw_ctx.bw.dcn.clk.fw_based_mclk_switching) {
+			dummy_latency_index = dcn30_find_dummy_latency_index_for_fw_based_mclk_switch(dc,
+				context, pipes, pipe_cnt, vlevel);
+
+			/* After calling dcn30_find_dummy_latency_index_for_fw_based_mclk_switch
+			 * we reinstate the original dram_clock_change_latency_us on the context
+			 * and all variables that may have changed up to this point, except the
+			 * newly found dummy_latency_index
+			 */
+			context->bw_ctx.dml.soc.dram_clock_change_latency_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_A].dml_input.pstate_latency_us;
+			dcn30_internal_validate_bw(dc, context, pipes, &pipe_cnt, &vlevel,
+				DC_VALIDATE_MODE_AND_PROGRAMMING, true);
+			maxMpcComb = context->bw_ctx.dml.vba.maxMpcComb;
+			dcfclk = context->bw_ctx.dml.vba.DCFCLKState[vlevel][context->bw_ctx.dml.vba.maxMpcComb];
+			pstate_en = context->bw_ctx.dml.vba.DRAMClockChangeSupport[vlevel][maxMpcComb] != dm_dram_clock_change_unsupported;
+		}
+	}
+
+	if (context->bw_ctx.dml.soc.min_dcfclk > dcfclk)
+		dcfclk = context->bw_ctx.dml.soc.min_dcfclk;
+
+	pipes[0].clks_cfg.voltage = vlevel;
+	pipes[0].clks_cfg.dcfclk_mhz = dcfclk;
+	pipes[0].clks_cfg.socclk_mhz = context->bw_ctx.dml.soc.clock_limits[vlevel].socclk_mhz;
+
+	/* Set B:
+	 * DCFCLK: 1GHz or min required above 1GHz
+	 * FCLK/UCLK: Max
+	 */
+	if (dc->clk_mgr->bw_params->wm_table.nv_entries[WM_B].valid) {
+		if (vlevel == 0) {
+			pipes[0].clks_cfg.voltage = 1;
+			pipes[0].clks_cfg.dcfclk_mhz = context->bw_ctx.dml.soc.clock_limits[0].dcfclk_mhz;
+		}
+		context->bw_ctx.dml.soc.dram_clock_change_latency_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_B].dml_input.pstate_latency_us;
+		context->bw_ctx.dml.soc.sr_enter_plus_exit_time_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_B].dml_input.sr_enter_plus_exit_time_us;
+		context->bw_ctx.dml.soc.sr_exit_time_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_B].dml_input.sr_exit_time_us;
+	}
+	context->bw_ctx.bw.dcn.watermarks.b.urgent_ns = get_wm_urgent(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.b.cstate_pstate.cstate_enter_plus_exit_ns = get_wm_stutter_enter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.b.cstate_pstate.cstate_exit_ns = get_wm_stutter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.b.cstate_pstate.pstate_change_ns = get_wm_dram_clock_change(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.b.pte_meta_urgent_ns = get_wm_memory_trip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.b.frac_urg_bw_nom = get_fraction_of_urgent_bandwidth(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.b.frac_urg_bw_flip = get_fraction_of_urgent_bandwidth_imm_flip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.b.urgent_latency_ns = get_urgent_latency(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+
+	pipes[0].clks_cfg.voltage = vlevel;
+	pipes[0].clks_cfg.dcfclk_mhz = dcfclk;
+
+	/* Set D:
+	 * DCFCLK: Min Required
+	 * FCLK(proportional to UCLK): 1GHz or Max
+	 * MALL stutter, sr_enter_exit = 4, sr_exit = 2us
+	 */
+	/*
+	if (dc->clk_mgr->bw_params->wm_table.nv_entries[WM_D].valid) {
+		context->bw_ctx.dml.soc.dram_clock_change_latency_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_D].dml_input.pstate_latency_us;
+		context->bw_ctx.dml.soc.sr_enter_plus_exit_time_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_D].dml_input.sr_enter_plus_exit_time_us;
+		context->bw_ctx.dml.soc.sr_exit_time_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_D].dml_input.sr_exit_time_us;
+	}
+	context->bw_ctx.bw.dcn.watermarks.d.urgent_ns = get_wm_urgent(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.d.cstate_pstate.cstate_enter_plus_exit_ns = get_wm_stutter_enter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.d.cstate_pstate.cstate_exit_ns = get_wm_stutter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.d.cstate_pstate.pstate_change_ns = get_wm_dram_clock_change(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.d.pte_meta_urgent_ns = get_wm_memory_trip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.d.frac_urg_bw_nom = get_fraction_of_urgent_bandwidth(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.d.frac_urg_bw_flip = get_fraction_of_urgent_bandwidth_imm_flip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.d.urgent_latency_ns = get_urgent_latency(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	*/
+
+	/* Set C:
+	 * DCFCLK: Min Required
+	 * FCLK(proportional to UCLK): 1GHz or Max
+	 * pstate latency overridden to 5us
+	 */
+	if (dc->clk_mgr->bw_params->wm_table.nv_entries[WM_C].valid) {
+		unsigned int min_dram_speed_mts = context->bw_ctx.dml.vba.DRAMSpeed;
+		unsigned int min_dram_speed_mts_margin = 160;
+
+		context->bw_ctx.dml.soc.dram_clock_change_latency_us =
+			dc->clk_mgr->bw_params->dummy_pstate_table[0].dummy_pstate_latency_us;
+
+		if (context->bw_ctx.dml.vba.DRAMClockChangeSupport[vlevel][maxMpcComb] ==
+			dm_dram_clock_change_unsupported) {
+			int min_dram_speed_mts_offset = dc->clk_mgr->bw_params->clk_table.num_entries - 1;
+
+			min_dram_speed_mts =
+				dc->clk_mgr->bw_params->clk_table.entries[min_dram_speed_mts_offset].memclk_mhz * 16;
+		}
+
+		if (!context->bw_ctx.bw.dcn.clk.fw_based_mclk_switching) {
+			/* find largest table entry that is lower than dram speed,
+			 * but lower than DPM0 still uses DPM0
+			 */
+			for (dummy_latency_index = 3; dummy_latency_index > 0; dummy_latency_index--)
+				if (min_dram_speed_mts + min_dram_speed_mts_margin >
+					dc->clk_mgr->bw_params->dummy_pstate_table[dummy_latency_index].dram_speed_mts)
+					break;
+		}
+
+		context->bw_ctx.dml.soc.dram_clock_change_latency_us =
+			dc->clk_mgr->bw_params->dummy_pstate_table[dummy_latency_index].dummy_pstate_latency_us;
+
+		context->bw_ctx.dml.soc.sr_enter_plus_exit_time_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_C].dml_input.sr_enter_plus_exit_time_us;
+		context->bw_ctx.dml.soc.sr_exit_time_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_C].dml_input.sr_exit_time_us;
+	}
+
+	context->bw_ctx.bw.dcn.watermarks.c.urgent_ns = get_wm_urgent(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.c.cstate_pstate.cstate_enter_plus_exit_ns = get_wm_stutter_enter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.c.cstate_pstate.cstate_exit_ns = get_wm_stutter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.c.cstate_pstate.pstate_change_ns = get_wm_dram_clock_change(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.c.pte_meta_urgent_ns = get_wm_memory_trip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.c.frac_urg_bw_nom = get_fraction_of_urgent_bandwidth(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.c.frac_urg_bw_flip = get_fraction_of_urgent_bandwidth_imm_flip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.c.urgent_latency_ns = get_urgent_latency(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+
+	if (!pstate_en) {
+		/* The only difference between A and C is p-state latency, if p-state is not supported we want to
+		 * calculate DLG based on dummy p-state latency, and max out the set A p-state watermark
+		 */
+		context->bw_ctx.bw.dcn.watermarks.a = context->bw_ctx.bw.dcn.watermarks.c;
+		context->bw_ctx.bw.dcn.watermarks.a.cstate_pstate.pstate_change_ns = 0;
+	} else {
+		/* Set A:
+		 * DCFCLK: Min Required
+		 * FCLK(proportional to UCLK): 1GHz or Max
+		 *
+		 * Set A calculated last so that following calculations are based on Set A
+		 */
+		dc->res_pool->funcs->update_soc_for_wm_a(dc, context);
+		context->bw_ctx.bw.dcn.watermarks.a.urgent_ns = get_wm_urgent(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+		context->bw_ctx.bw.dcn.watermarks.a.cstate_pstate.cstate_enter_plus_exit_ns = get_wm_stutter_enter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+		context->bw_ctx.bw.dcn.watermarks.a.cstate_pstate.cstate_exit_ns = get_wm_stutter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+		context->bw_ctx.bw.dcn.watermarks.a.cstate_pstate.pstate_change_ns = get_wm_dram_clock_change(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+		context->bw_ctx.bw.dcn.watermarks.a.pte_meta_urgent_ns = get_wm_memory_trip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+		context->bw_ctx.bw.dcn.watermarks.a.frac_urg_bw_nom = get_fraction_of_urgent_bandwidth(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+		context->bw_ctx.bw.dcn.watermarks.a.frac_urg_bw_flip = get_fraction_of_urgent_bandwidth_imm_flip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+		context->bw_ctx.bw.dcn.watermarks.a.urgent_latency_ns = get_urgent_latency(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	}
+
+	context->perf_params.stutter_period_us = context->bw_ctx.dml.vba.StutterPeriod;
+
+	/* Make set D = set A until set D is enabled */
+	context->bw_ctx.bw.dcn.watermarks.d = context->bw_ctx.bw.dcn.watermarks.a;
+
+	for (i = 0, pipe_idx = 0; i < dc->res_pool->pipe_count; i++) {
+		if (!context->res_ctx.pipe_ctx[i].stream)
+			continue;
+
+		pipes[pipe_idx].clks_cfg.dispclk_mhz = get_dispclk_calculated(&context->bw_ctx.dml, pipes, pipe_cnt);
+		pipes[pipe_idx].clks_cfg.dppclk_mhz = get_dppclk_calculated(&context->bw_ctx.dml, pipes, pipe_cnt, pipe_idx);
+
+		if (dc->config.forced_clocks) {
+			pipes[pipe_idx].clks_cfg.dispclk_mhz = context->bw_ctx.dml.soc.clock_limits[0].dispclk_mhz;
+			pipes[pipe_idx].clks_cfg.dppclk_mhz = context->bw_ctx.dml.soc.clock_limits[0].dppclk_mhz;
+		}
+		if (dc->debug.min_disp_clk_khz > pipes[pipe_idx].clks_cfg.dispclk_mhz * 1000)
+			pipes[pipe_idx].clks_cfg.dispclk_mhz = dc->debug.min_disp_clk_khz / 1000.0;
+		if (dc->debug.min_dpp_clk_khz > pipes[pipe_idx].clks_cfg.dppclk_mhz * 1000)
+			pipes[pipe_idx].clks_cfg.dppclk_mhz = dc->debug.min_dpp_clk_khz / 1000.0;
+
+		pipe_idx++;
+	}
+
+	// WA: restrict FPO to use first non-strobe mode (NV24 BW issue)
+	if (context->bw_ctx.bw.dcn.clk.fw_based_mclk_switching &&
+			dc->dml.soc.num_chans <= 4 &&
+			context->bw_ctx.dml.vba.DRAMSpeed <= 1700 &&
+			context->bw_ctx.dml.vba.DRAMSpeed >= 1500) {
+
+		for (i = 0; i < dc->dml.soc.num_states; i++) {
+			if (dc->dml.soc.clock_limits[i].dram_speed_mts > 1700) {
+				context->bw_ctx.dml.vba.DRAMSpeed = dc->dml.soc.clock_limits[i].dram_speed_mts;
+				break;
+			}
+		}
+	}
+
+	dcn20_calculate_dlg_params(dc, context, pipes, pipe_cnt, vlevel);
+
+	if (!pstate_en)
+		/* Restore full p-state latency */
+		context->bw_ctx.dml.soc.dram_clock_change_latency_us =
+				dc->clk_mgr->bw_params->wm_table.nv_entries[WM_A].dml_input.pstate_latency_us;
+
+	if (context->bw_ctx.bw.dcn.clk.fw_based_mclk_switching)
+		dcn30_setup_mclk_switch_using_fw_based_vblank_stretch(dc, context);
+}
+
+void dcn30_fpu_update_dram_channel_width_bytes(struct dc *dc)
+{
+	dc_assert_fp_enabled();
+
+	if (dc->ctx->dc_bios->vram_info.dram_channel_width_bytes)
+		dcn3_0_soc.dram_channel_width_bytes = dc->ctx->dc_bios->vram_info.dram_channel_width_bytes;
+}
+
+void dcn30_fpu_update_max_clk(struct dc_bounding_box_max_clk *dcn30_bb_max_clk)
+{
+		dc_assert_fp_enabled();
+
+		if (!dcn30_bb_max_clk->max_dcfclk_mhz)
+			dcn30_bb_max_clk->max_dcfclk_mhz = dcn3_0_soc.clock_limits[0].dcfclk_mhz;
+		if (!dcn30_bb_max_clk->max_dispclk_mhz)
+			dcn30_bb_max_clk->max_dispclk_mhz = dcn3_0_soc.clock_limits[0].dispclk_mhz;
+		if (!dcn30_bb_max_clk->max_dppclk_mhz)
+			dcn30_bb_max_clk->max_dppclk_mhz = dcn3_0_soc.clock_limits[0].dppclk_mhz;
+		if (!dcn30_bb_max_clk->max_phyclk_mhz)
+			dcn30_bb_max_clk->max_phyclk_mhz = dcn3_0_soc.clock_limits[0].phyclk_mhz;
+}
+
+void dcn30_fpu_get_optimal_dcfclk_fclk_for_uclk(unsigned int uclk_mts,
+		unsigned int *optimal_dcfclk,
+		unsigned int *optimal_fclk)
+{
+	double bw_from_dram, bw_from_dram1, bw_from_dram2;
+
+	dc_assert_fp_enabled();
+
+	bw_from_dram1 = uclk_mts * dcn3_0_soc.num_chans *
+		dcn3_0_soc.dram_channel_width_bytes * (dcn3_0_soc.max_avg_dram_bw_use_normal_percent / 100);
+	bw_from_dram2 = uclk_mts * dcn3_0_soc.num_chans *
+		dcn3_0_soc.dram_channel_width_bytes * (dcn3_0_soc.max_avg_sdp_bw_use_normal_percent / 100);
+
+	bw_from_dram = (bw_from_dram1 < bw_from_dram2) ? bw_from_dram1 : bw_from_dram2;
+
+	if (optimal_fclk)
+		*optimal_fclk = bw_from_dram /
+		(dcn3_0_soc.fabric_datapath_to_dcn_data_return_bytes * (dcn3_0_soc.max_avg_sdp_bw_use_normal_percent / 100));
+
+	if (optimal_dcfclk)
+		*optimal_dcfclk =  bw_from_dram /
+		(dcn3_0_soc.return_bus_width_bytes * (dcn3_0_soc.max_avg_sdp_bw_use_normal_percent / 100));
+}
+
+void dcn30_fpu_update_bw_bounding_box(struct dc *dc,
+	struct clk_bw_params *bw_params,
+	struct dc_bounding_box_max_clk *dcn30_bb_max_clk,
+	unsigned int *dcfclk_mhz,
+	unsigned int *dram_speed_mts)
+{
+	unsigned int i;
+
+	dc_assert_fp_enabled();
+
+	dcn3_0_soc.dispclk_dppclk_vco_speed_mhz = dc->clk_mgr->dentist_vco_freq_khz / 1000.0;
+	dc->dml.soc.dispclk_dppclk_vco_speed_mhz = dc->clk_mgr->dentist_vco_freq_khz / 1000.0;
+
+	for (i = 0; i < dcn3_0_soc.num_states; i++) {
+		dcn3_0_soc.clock_limits[i].state = i;
+		dcn3_0_soc.clock_limits[i].dcfclk_mhz = dcfclk_mhz[i];
+		dcn3_0_soc.clock_limits[i].fabricclk_mhz = dcfclk_mhz[i];
+		dcn3_0_soc.clock_limits[i].dram_speed_mts = dram_speed_mts[i];
+
+		/* Fill all states with max values of all other clocks */
+		dcn3_0_soc.clock_limits[i].dispclk_mhz = dcn30_bb_max_clk->max_dispclk_mhz;
+		dcn3_0_soc.clock_limits[i].dppclk_mhz  = dcn30_bb_max_clk->max_dppclk_mhz;
+		dcn3_0_soc.clock_limits[i].phyclk_mhz  = dcn30_bb_max_clk->max_phyclk_mhz;
+		dcn3_0_soc.clock_limits[i].dtbclk_mhz = dcn3_0_soc.clock_limits[0].dtbclk_mhz;
+		/* These clocks cannot come from bw_params, always fill from dcn3_0_soc[1] */
+		/* FCLK, PHYCLK_D18, SOCCLK, DSCCLK */
+		dcn3_0_soc.clock_limits[i].phyclk_d18_mhz = dcn3_0_soc.clock_limits[0].phyclk_d18_mhz;
+		dcn3_0_soc.clock_limits[i].socclk_mhz = dcn3_0_soc.clock_limits[0].socclk_mhz;
+		dcn3_0_soc.clock_limits[i].dscclk_mhz = dcn3_0_soc.clock_limits[0].dscclk_mhz;
+	}
+	/* re-init DML with updated bb */
+	dml_init_instance(&dc->dml, &dcn3_0_soc, &dcn3_0_ip, DML_PROJECT_DCN30);
+	if (dc->current_state)
+		dml_init_instance(&dc->current_state->bw_ctx.dml, &dcn3_0_soc, &dcn3_0_ip, DML_PROJECT_DCN30);
+
+}
+
+/**
+ * dcn30_find_dummy_latency_index_for_fw_based_mclk_switch() - Finds
+ * dummy_latency_index when MCLK switching using firmware based vblank stretch
+ * is enabled. This function will iterate through the table of dummy pstate
+ * latencies until the lowest value that allows
+ * dm_allow_self_refresh_and_mclk_switch to happen is found
+ *
+ * @dc: Current DC state
+ * @context: new dc state
+ * @pipes: DML pipe params
+ * @pipe_cnt: number of DML pipes
+ * @vlevel: Voltage level calculated by DML
+ *
+ * Return: lowest dummy_latency_index value
+ */
+int dcn30_find_dummy_latency_index_for_fw_based_mclk_switch(struct dc *dc,
+							    struct dc_state *context,
+							    display_e2e_pipe_params_st *pipes,
+							    int pipe_cnt,
+							    int vlevel)
+{
+	const int max_latency_table_entries = 4;
+	int dummy_latency_index = 0;
+
+	dc_assert_fp_enabled();
+
+	while (dummy_latency_index < max_latency_table_entries) {
+		context->bw_ctx.dml.soc.dram_clock_change_latency_us =
+				dc->clk_mgr->bw_params->dummy_pstate_table[dummy_latency_index].dummy_pstate_latency_us;
+		dcn30_internal_validate_bw(dc, context, pipes, &pipe_cnt, &vlevel,
+			DC_VALIDATE_MODE_AND_PROGRAMMING, true);
+
+		if (context->bw_ctx.dml.soc.allow_dram_self_refresh_or_dram_clock_change_in_vblank ==
+			dm_allow_self_refresh_and_mclk_switch)
+			break;
+
+		dummy_latency_index++;
+	}
+
+	if (dummy_latency_index == max_latency_table_entries) {
+		ASSERT(dummy_latency_index != max_latency_table_entries);
+		/* If the execution gets here, it means dummy p_states are
+		 * not possible. This should never happen and would mean
+		 * something is severely wrong.
+		 * Here we reset dummy_latency_index to 3, because it is
+		 * better to have underflows than system crashes.
+		 */
+		dummy_latency_index = 3;
+	}
+
+	return dummy_latency_index;
+}
+
+void dcn3_fpu_build_wm_range_table(struct clk_mgr *base)
+{
+	/* defaults */
+	double pstate_latency_us = base->ctx->dc->dml.soc.dram_clock_change_latency_us;
+	double sr_exit_time_us = base->ctx->dc->dml.soc.sr_exit_time_us;
+	double sr_enter_plus_exit_time_us = base->ctx->dc->dml.soc.sr_enter_plus_exit_time_us;
+	uint16_t min_uclk_mhz = base->bw_params->clk_table.entries[0].memclk_mhz;
+
+	dc_assert_fp_enabled();
+
+	/* Set A - Normal - default values*/
+	base->bw_params->wm_table.nv_entries[WM_A].valid = true;
+	base->bw_params->wm_table.nv_entries[WM_A].dml_input.pstate_latency_us = pstate_latency_us;
+	base->bw_params->wm_table.nv_entries[WM_A].dml_input.sr_exit_time_us = sr_exit_time_us;
+	base->bw_params->wm_table.nv_entries[WM_A].dml_input.sr_enter_plus_exit_time_us = sr_enter_plus_exit_time_us;
+	base->bw_params->wm_table.nv_entries[WM_A].pmfw_breakdown.wm_type = WATERMARKS_CLOCK_RANGE;
+	base->bw_params->wm_table.nv_entries[WM_A].pmfw_breakdown.min_dcfclk = 0;
+	base->bw_params->wm_table.nv_entries[WM_A].pmfw_breakdown.max_dcfclk = 0xFFFF;
+	base->bw_params->wm_table.nv_entries[WM_A].pmfw_breakdown.min_uclk = min_uclk_mhz;
+	base->bw_params->wm_table.nv_entries[WM_A].pmfw_breakdown.max_uclk = 0xFFFF;
+
+	/* Set B - Performance - higher minimum clocks */
+//	base->bw_params->wm_table.nv_entries[WM_B].valid = true;
+//	base->bw_params->wm_table.nv_entries[WM_B].dml_input.pstate_latency_us = pstate_latency_us;
+//	base->bw_params->wm_table.nv_entries[WM_B].dml_input.sr_exit_time_us = sr_exit_time_us;
+//	base->bw_params->wm_table.nv_entries[WM_B].dml_input.sr_enter_plus_exit_time_us = sr_enter_plus_exit_time_us;
+//	base->bw_params->wm_table.nv_entries[WM_B].pmfw_breakdown.wm_type = WATERMARKS_CLOCK_RANGE;
+//	base->bw_params->wm_table.nv_entries[WM_B].pmfw_breakdown.min_dcfclk = TUNED VALUE;
+//	base->bw_params->wm_table.nv_entries[WM_B].pmfw_breakdown.max_dcfclk = 0xFFFF;
+//	base->bw_params->wm_table.nv_entries[WM_B].pmfw_breakdown.min_uclk = TUNED VALUE;
+//	base->bw_params->wm_table.nv_entries[WM_B].pmfw_breakdown.max_uclk = 0xFFFF;
+
+	/* Set C - Dummy P-State - P-State latency set to "dummy p-state" value */
+	base->bw_params->wm_table.nv_entries[WM_C].valid = true;
+	base->bw_params->wm_table.nv_entries[WM_C].dml_input.pstate_latency_us = 0;
+	base->bw_params->wm_table.nv_entries[WM_C].dml_input.sr_exit_time_us = sr_exit_time_us;
+	base->bw_params->wm_table.nv_entries[WM_C].dml_input.sr_enter_plus_exit_time_us = sr_enter_plus_exit_time_us;
+	base->bw_params->wm_table.nv_entries[WM_C].pmfw_breakdown.wm_type = WATERMARKS_DUMMY_PSTATE;
+	base->bw_params->wm_table.nv_entries[WM_C].pmfw_breakdown.min_dcfclk = 0;
+	base->bw_params->wm_table.nv_entries[WM_C].pmfw_breakdown.max_dcfclk = 0xFFFF;
+	base->bw_params->wm_table.nv_entries[WM_C].pmfw_breakdown.min_uclk = min_uclk_mhz;
+	base->bw_params->wm_table.nv_entries[WM_C].pmfw_breakdown.max_uclk = 0xFFFF;
+	base->bw_params->dummy_pstate_table[0].dram_speed_mts = 1600;
+	base->bw_params->dummy_pstate_table[0].dummy_pstate_latency_us = 38;
+	base->bw_params->dummy_pstate_table[1].dram_speed_mts = 8000;
+	base->bw_params->dummy_pstate_table[1].dummy_pstate_latency_us = 9;
+	base->bw_params->dummy_pstate_table[2].dram_speed_mts = 10000;
+	base->bw_params->dummy_pstate_table[2].dummy_pstate_latency_us = 8;
+	base->bw_params->dummy_pstate_table[3].dram_speed_mts = 16000;
+	base->bw_params->dummy_pstate_table[3].dummy_pstate_latency_us = 5;
+
+	/* Set D - MALL - SR enter and exit times adjusted for MALL */
+	base->bw_params->wm_table.nv_entries[WM_D].valid = true;
+	base->bw_params->wm_table.nv_entries[WM_D].dml_input.pstate_latency_us = pstate_latency_us;
+	base->bw_params->wm_table.nv_entries[WM_D].dml_input.sr_exit_time_us = 2;
+	base->bw_params->wm_table.nv_entries[WM_D].dml_input.sr_enter_plus_exit_time_us = 4;
+	base->bw_params->wm_table.nv_entries[WM_D].pmfw_breakdown.wm_type = WATERMARKS_MALL;
+	base->bw_params->wm_table.nv_entries[WM_D].pmfw_breakdown.min_dcfclk = 0;
+	base->bw_params->wm_table.nv_entries[WM_D].pmfw_breakdown.max_dcfclk = 0xFFFF;
+	base->bw_params->wm_table.nv_entries[WM_D].pmfw_breakdown.min_uclk = min_uclk_mhz;
+	base->bw_params->wm_table.nv_entries[WM_D].pmfw_breakdown.max_uclk = 0xFFFF;
+}
+
+void patch_dcn30_soc_bounding_box(struct dc *dc, struct _vcs_dpi_soc_bounding_box_st *dcn3_0_ip)
+{
+	dc_assert_fp_enabled();
+
+	if (dc->ctx->dc_bios->funcs->get_soc_bb_info) {
+		struct bp_soc_bb_info bb_info = {0};
+
+		if (dc->ctx->dc_bios->funcs->get_soc_bb_info(dc->ctx->dc_bios, &bb_info) == BP_RESULT_OK) {
+			if (bb_info.dram_clock_change_latency_100ns > 0)
+				dcn3_0_soc.dram_clock_change_latency_us = bb_info.dram_clock_change_latency_100ns * 10;
+
+			if (bb_info.dram_sr_enter_exit_latency_100ns > 0)
+				dcn3_0_soc.sr_enter_plus_exit_time_us = bb_info.dram_sr_enter_exit_latency_100ns * 10;
+
+			if (bb_info.dram_sr_exit_latency_100ns > 0)
+				dcn3_0_soc.sr_exit_time_us = bb_info.dram_sr_exit_latency_100ns * 10;
+		}
+	}
+}
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn30/dcn30_fpu.h b/drivers/gpu/drm/amd/display/dc/dml/dcn30/dcn30_fpu.h
new file mode 100644
index 000000000000..e3b6ad6a8784
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn30/dcn30_fpu.h
@@ -0,0 +1,73 @@
+/*
+ * Copyright 2020-2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __DCN30_FPU_H__
+#define __DCN30_FPU_H__
+
+#include "core_types.h"
+#include "dcn20/dcn20_optc.h"
+
+void dcn30_fpu_populate_dml_writeback_from_context(
+		struct dc *dc, struct resource_context *res_ctx, display_e2e_pipe_params_st *pipes);
+
+void dcn30_fpu_set_mcif_arb_params(struct mcif_arb_params *wb_arb_params,
+	struct display_mode_lib *dml,
+	display_e2e_pipe_params_st *pipes,
+	int pipe_cnt,
+	int cur_pipe);
+
+void dcn30_fpu_update_soc_for_wm_a(struct dc *dc, struct dc_state *context);
+
+void dcn30_fpu_calculate_wm_and_dlg(
+		struct dc *dc, struct dc_state *context,
+		display_e2e_pipe_params_st *pipes,
+		int pipe_cnt,
+		int vlevel);
+
+void dcn30_fpu_update_dram_channel_width_bytes(struct dc *dc);
+
+void dcn30_fpu_update_max_clk(struct dc_bounding_box_max_clk *dcn30_bb_max_clk);
+
+void dcn30_fpu_get_optimal_dcfclk_fclk_for_uclk(unsigned int uclk_mts,
+		unsigned int *optimal_dcfclk,
+		unsigned int *optimal_fclk);
+
+void dcn30_fpu_update_bw_bounding_box(struct dc *dc,
+	struct clk_bw_params *bw_params,
+	struct dc_bounding_box_max_clk *dcn30_bb_max_clk,
+	unsigned int *dcfclk_mhz,
+	unsigned int *dram_speed_mts);
+
+int dcn30_find_dummy_latency_index_for_fw_based_mclk_switch(struct dc *dc,
+							    struct dc_state *context,
+							    display_e2e_pipe_params_st *pipes,
+							    int pipe_cnt,
+							    int vlevel);
+
+void dcn3_fpu_build_wm_range_table(struct clk_mgr *base);
+
+void patch_dcn30_soc_bounding_box(struct dc *dc, struct _vcs_dpi_soc_bounding_box_st *dcn3_0_ip);
+
+#endif /* __DCN30_FPU_H__*/
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn30/display_mode_vba_30.c b/drivers/gpu/drm/amd/display/dc/dml/dcn30/display_mode_vba_30.c
index f47d82da115c..8d24763938ea 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dcn30/display_mode_vba_30.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn30/display_mode_vba_30.c
@@ -23,9 +23,7 @@
  *
  */
 
-#ifdef CONFIG_DRM_AMD_DC_DCN
 #include "dc.h"
-#include "dc_link.h"
 #include "../display_mode_lib.h"
 #include "display_mode_vba_30.h"
 #include "../dml_inline_defs.h"
@@ -283,10 +281,10 @@ static void CalculateDynamicMetadataParameters(
 		double DISPCLK,
 		double DCFClkDeepSleep,
 		double PixelClock,
-		long HTotal,
-		long VBlank,
-		long DynamicMetadataTransmittedBytes,
-		long DynamicMetadataLinesBeforeActiveRequired,
+		unsigned int HTotal,
+		unsigned int VBlank,
+		unsigned int DynamicMetadataTransmittedBytes,
+		int DynamicMetadataLinesBeforeActiveRequired,
 		int InterlaceEnable,
 		bool ProgressiveToInterlaceUnitInOPP,
 		double *Tsetup,
@@ -396,64 +394,10 @@ static void CalculateUrgentBurstFactor(
 
 static void UseMinimumDCFCLK(
 		struct display_mode_lib *mode_lib,
-		int MaxInterDCNTileRepeaters,
+		struct vba_vars_st *v,
 		int MaxPrefetchMode,
-		double FinalDRAMClockChangeLatency,
-		double SREnterPlusExitTime,
-		int ReturnBusWidth,
-		int RoundTripPingLatencyCycles,
-		int ReorderingBytes,
-		int PixelChunkSizeInKByte,
-		int MetaChunkSize,
-		bool GPUVMEnable,
-		int GPUVMMaxPageTableLevels,
-		bool HostVMEnable,
-		int NumberOfActivePlanes,
-		double HostVMMinPageSize,
-		int HostVMMaxNonCachedPageTableLevels,
-		bool DynamicMetadataVMEnabled,
-		enum immediate_flip_requirement ImmediateFlipRequirement,
-		bool ProgressiveToInterlaceUnitInOPP,
-		double MaxAveragePercentOfIdealSDPPortBWDisplayCanUseInNormalSystemOperation,
-		double PercentOfIdealDRAMFabricAndSDPPortBWReceivedAfterUrgLatencyPixelMixedWithVMData,
-		double PercentOfIdealDRAMFabricAndSDPPortBWReceivedAfterUrgLatencyVMDataOnly,
-		double PercentOfIdealDRAMFabricAndSDPPortBWReceivedAfterUrgLatencyPixelDataOnly,
-		int VTotal[],
-		int VActive[],
-		int DynamicMetadataTransmittedBytes[],
-		int DynamicMetadataLinesBeforeActiveRequired[],
-		bool Interlace[],
-		double RequiredDPPCLK[][2][DC__NUM_DPP__MAX],
-		double RequiredDISPCLK[][2],
-		double UrgLatency[],
-		unsigned int NoOfDPP[][2][DC__NUM_DPP__MAX],
-		double ProjectedDCFCLKDeepSleep[][2],
-		double MaximumVStartup[][2][DC__NUM_DPP__MAX],
-		double TotalVActivePixelBandwidth[][2],
-		double TotalVActiveCursorBandwidth[][2],
-		double TotalMetaRowBandwidth[][2],
-		double TotalDPTERowBandwidth[][2],
-		unsigned int TotalNumberOfActiveDPP[][2],
-		unsigned int TotalNumberOfDCCActiveDPP[][2],
-		int dpte_group_bytes[],
-		double PrefetchLinesY[][2][DC__NUM_DPP__MAX],
-		double PrefetchLinesC[][2][DC__NUM_DPP__MAX],
-		int swath_width_luma_ub_all_states[][2][DC__NUM_DPP__MAX],
-		int swath_width_chroma_ub_all_states[][2][DC__NUM_DPP__MAX],
-		int BytePerPixelY[],
-		int BytePerPixelC[],
-		int HTotal[],
-		double PixelClock[],
-		double PDEAndMetaPTEBytesPerFrame[][2][DC__NUM_DPP__MAX],
-		double DPTEBytesPerRow[][2][DC__NUM_DPP__MAX],
-		double MetaRowBytes[][2][DC__NUM_DPP__MAX],
-		bool DynamicMetadataEnable[],
-		double VActivePixelBandwidth[][2][DC__NUM_DPP__MAX],
-		double VActiveCursorBandwidth[][2][DC__NUM_DPP__MAX],
-		double ReadBandwidthLuma[],
-		double ReadBandwidthChroma[],
-		double DCFCLKPerState[],
-		double DCFCLKState[][2]);
+		int ReorderingBytes);
+
 static void CalculatePixelDeliveryTimes(
 		unsigned int NumberOfActivePlanes,
 		double VRatio[],
@@ -712,18 +656,6 @@ static double CalculateUrgentLatency(
 		double UrgentLatencyAdjustmentFabricClockReference,
 		double FabricClockSingle);
 
-static bool CalculateBytePerPixelAnd256BBlockSizes(
-		enum source_format_class SourcePixelFormat,
-		enum dm_swizzle_mode SurfaceTiling,
-		unsigned int *BytePerPixelY,
-		unsigned int *BytePerPixelC,
-		double       *BytePerPixelDETY,
-		double       *BytePerPixelDETC,
-		unsigned int *BlockHeight256BytesY,
-		unsigned int *BlockHeight256BytesC,
-		unsigned int *BlockWidth256BytesY,
-		unsigned int *BlockWidth256BytesC);
-
 void dml30_recalculate(struct display_mode_lib *mode_lib)
 {
 	ModeSupportAndSystemConfiguration(mode_lib);
@@ -852,8 +784,7 @@ static unsigned int dscComputeDelay(enum output_format_class pixelFormat, enum o
 		Delay = Delay + 1;
 		//   sft
 		Delay = Delay + 1;
-	}
-	else {
+	} else {
 		//   sfr
 		Delay = Delay + 2;
 		//   dsccif
@@ -1071,6 +1002,7 @@ static bool CalculatePrefetchSchedule(
 
 	dst_y_prefetch_equ = VStartup - (Tsetup + dml_max(TWait + TCalc, *Tdmdl)) / LineTime
 			- (*DSTYAfterScaler + *DSTXAfterScaler / myPipe->HTotal);
+	dst_y_prefetch_equ = dml_min(dst_y_prefetch_equ, 63.75); // limit to the reg limit of U6.2 for DST_Y_PREFETCH
 
 	Lsw_oto = dml_max(PrefetchSourceLinesY, PrefetchSourceLinesC);
 	Tsw_oto = Lsw_oto * LineTime;
@@ -1349,12 +1281,9 @@ static bool CalculatePrefetchSchedule(
 
 	if (MyError) {
 		*PrefetchBandwidth = 0;
-		TimeForFetchingMetaPTE = 0;
-		TimeForFetchingRowInVBlank = 0;
 		*DestinationLinesToRequestVMInVBlank = 0;
 		*DestinationLinesToRequestRowInVBlank = 0;
 		*DestinationLinesForPrefetch = 0;
-		LinesToRequestPrefetchPixelData = 0;
 		*VRatioPrefetchY = 0;
 		*VRatioPrefetchC = 0;
 		*RequiredPrefetchPixDataBWLuma = 0;
@@ -1844,15 +1773,6 @@ static unsigned int CalculateVMAndRowBytes(
 		*PixelPTEReqWidth = 32768.0 / BytePerPixel;
 		*PTERequestSize = 64;
 		FractionOfPTEReturnDrop = 0;
-	} else if (MacroTileSizeBytes == 4096) {
-		PixelPTEReqHeightPTEs = 1;
-		*PixelPTEReqHeight = MacroTileHeight;
-		*PixelPTEReqWidth = 8 * *MacroTileWidth;
-		*PTERequestSize = 64;
-		if (ScanDirection != dm_vert)
-			FractionOfPTEReturnDrop = 0;
-		else
-			FractionOfPTEReturnDrop = 7 / 8;
 	} else if (GPUVMMinPageSize == 4 && MacroTileSizeBytes > 4096) {
 		PixelPTEReqHeightPTEs = 16;
 		*PixelPTEReqHeight = 16 * BlockHeight256Bytes;
@@ -1868,7 +1788,10 @@ static unsigned int CalculateVMAndRowBytes(
 	}
 
 	if (SurfaceTiling == dm_sw_linear) {
-		*dpte_row_height = dml_min(128, 1 << (unsigned int) dml_floor(dml_log2(PTEBufferSizeInRequests * *PixelPTEReqWidth / Pitch), 1));
+		if (PTEBufferSizeInRequests == 0)
+			*dpte_row_height = 1;
+		else
+			*dpte_row_height = dml_min(128, 1 << (unsigned int) dml_floor(dml_log2(PTEBufferSizeInRequests * *PixelPTEReqWidth / Pitch), 1));
 		*dpte_row_width_ub = (dml_ceil(((double) SwathWidth - 1) / *PixelPTEReqWidth, 1) + 1) * *PixelPTEReqWidth;
 		*PixelPTEBytesPerRow = *dpte_row_width_ub / *PixelPTEReqWidth * *PTERequestSize;
 	} else if (ScanDirection != dm_vert) {
@@ -2095,7 +2018,7 @@ static void DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerforman
 	DTRACE("   return_bus_bw      = %f", v->ReturnBW);
 
 	for (k = 0; k < v->NumberOfActivePlanes; ++k) {
-		CalculateBytePerPixelAnd256BBlockSizes(
+		dml30_CalculateBytePerPixelAnd256BBlockSizes(
 				v->SourcePixelFormat[k],
 				v->SurfaceTiling[k],
 				&v->BytePerPixelY[k],
@@ -3049,40 +2972,12 @@ static void DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerforman
 
 	{
 		//Maximum Bandwidth Used
-		double TotalWRBandwidth = 0;
-		double MaxPerPlaneVActiveWRBandwidth = 0;
-		double WRBandwidth = 0;
-		double MaxUsedBW = 0;
-		for (k = 0; k < v->NumberOfActivePlanes; ++k) {
-			if (v->WritebackEnable[k] == true
-					&& v->WritebackPixelFormat[k] == dm_444_32) {
-				WRBandwidth = v->WritebackDestinationWidth[k] * v->WritebackDestinationHeight[k]
-						/ (v->HTotal[k] * v->WritebackSourceHeight[k] / v->PixelClock[k]) * 4;
-			} else if (v->WritebackEnable[k] == true) {
-				WRBandwidth = v->WritebackDestinationWidth[k] * v->WritebackDestinationHeight[k]
-						/ (v->HTotal[k] * v->WritebackSourceHeight[k] / v->PixelClock[k]) * 8;
-			}
-			TotalWRBandwidth = TotalWRBandwidth + WRBandwidth;
-			MaxPerPlaneVActiveWRBandwidth = dml_max(MaxPerPlaneVActiveWRBandwidth, WRBandwidth);
-		}
-
 		v->TotalDataReadBandwidth = 0;
 		for (k = 0; k < v->NumberOfActivePlanes; ++k) {
 			v->TotalDataReadBandwidth = v->TotalDataReadBandwidth
 					+ v->ReadBandwidthPlaneLuma[k]
 					+ v->ReadBandwidthPlaneChroma[k];
 		}
-
-		{
-			double MaxPerPlaneVActiveRDBandwidth = 0;
-			for (k = 0; k < v->NumberOfActivePlanes; ++k) {
-				MaxPerPlaneVActiveRDBandwidth = dml_max(MaxPerPlaneVActiveRDBandwidth,
-						v->ReadBandwidthPlaneLuma[k] + v->ReadBandwidthPlaneChroma[k]);
-
-			}
-		}
-
-		MaxUsedBW = MaxTotalRDBandwidth + TotalWRBandwidth;
 	}
 
 	// VStartup Margin
@@ -3165,7 +3060,7 @@ static void DisplayPipeConfiguration(struct display_mode_lib *mode_lib)
 
 	for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) {
 
-		CalculateBytePerPixelAnd256BBlockSizes(
+		dml30_CalculateBytePerPixelAnd256BBlockSizes(
 				mode_lib->vba.SourcePixelFormat[k],
 				mode_lib->vba.SurfaceTiling[k],
 				&BytePerPixY[k],
@@ -3218,7 +3113,7 @@ static void DisplayPipeConfiguration(struct display_mode_lib *mode_lib)
 			&dummysinglestring);
 }
 
-static bool CalculateBytePerPixelAnd256BBlockSizes(
+void dml30_CalculateBytePerPixelAnd256BBlockSizes(
 		enum source_format_class SourcePixelFormat,
 		enum dm_swizzle_mode SurfaceTiling,
 		unsigned int *BytePerPixelY,
@@ -3305,7 +3200,6 @@ static bool CalculateBytePerPixelAnd256BBlockSizes(
 		*BlockWidth256BytesY = 256U / *BytePerPixelY / *BlockHeight256BytesY;
 		*BlockWidth256BytesC = 256U / *BytePerPixelC / *BlockHeight256BytesC;
 	}
-	return true;
 }
 
 static double CalculateTWait(
@@ -3372,8 +3266,8 @@ static double CalculateWriteBackDelay(
 
 
 static void CalculateDynamicMetadataParameters(int MaxInterDCNTileRepeaters, double DPPCLK, double DISPCLK,
-		double DCFClkDeepSleep, double PixelClock, long HTotal, long VBlank, long DynamicMetadataTransmittedBytes,
-		long DynamicMetadataLinesBeforeActiveRequired, int InterlaceEnable, bool ProgressiveToInterlaceUnitInOPP,
+		double DCFClkDeepSleep, double PixelClock, unsigned int HTotal, unsigned int VBlank, unsigned int DynamicMetadataTransmittedBytes,
+		int DynamicMetadataLinesBeforeActiveRequired, int InterlaceEnable, bool ProgressiveToInterlaceUnitInOPP,
 		double *Tsetup, double *Tdmbf, double *Tdmec, double *Tdmsks)
 {
 	double TotalRepeaterDelayTime = 0;
@@ -3583,8 +3477,7 @@ static double TruncToValidBPP(
 		if (Format == dm_n422) {
 			MinDSCBPP = 7;
 			MaxDSCBPP = 2 * DSCInputBitPerComponent - 1.0 / 16.0;
-		}
-		else {
+		} else {
 			MinDSCBPP = 8;
 			MaxDSCBPP = 3 * DSCInputBitPerComponent - 1.0 / 16.0;
 		}
@@ -3631,14 +3524,13 @@ static double TruncToValidBPP(
 			return DesiredBPP;
 		}
 	}
-	return BPP_INVALID;
 }
 
 void dml30_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_lib)
 {
 	struct vba_vars_st *v = &mode_lib->vba;
 	int MinPrefetchMode, MaxPrefetchMode;
-	int i;
+	int i, start_state;
 	unsigned int j, k, m;
 	bool   EnoughWritebackUnits = true;
 	bool   WritebackModeSupport = true;
@@ -3649,6 +3541,11 @@ void dml30_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 
 	/*MODE SUPPORT, VOLTAGE STATE AND SOC CONFIGURATION*/
 
+	if (mode_lib->validate_max_state)
+		start_state = v->soc.num_states - 1;
+	else
+		start_state = 0;
+
 	CalculateMinAndMaxPrefetchMode(
 		mode_lib->vba.AllowDRAMSelfRefreshOrDRAMClockChangeInVblank,
 		&MinPrefetchMode, &MaxPrefetchMode);
@@ -3709,7 +3606,7 @@ void dml30_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 	/*Bandwidth Support Check*/
 
 	for (k = 0; k <= v->NumberOfActivePlanes - 1; k++) {
-		CalculateBytePerPixelAnd256BBlockSizes(
+		dml30_CalculateBytePerPixelAnd256BBlockSizes(
 				v->SourcePixelFormat[k],
 				v->SurfaceTiling[k],
 				&v->BytePerPixelY[k],
@@ -3947,7 +3844,7 @@ void dml30_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 			v->SingleDPPViewportSizeSupportPerPlane,
 			&v->ViewportSizeSupport[0][0]);
 
-	for (i = 0; i < v->soc.num_states; i++) {
+	for (i = start_state; i < v->soc.num_states; i++) {
 		for (j = 0; j < 2; j++) {
 			v->MaxDispclkRoundedDownToDFSGranularity = RoundToDFSGranularityDown(v->MaxDispclk[i], v->DISPCLKDPPCLKVCOSpeed);
 			v->MaxDppclkRoundedDownToDFSGranularity = RoundToDFSGranularityDown(v->MaxDppclk[i], v->DISPCLKDPPCLKVCOSpeed);
@@ -4103,7 +4000,7 @@ void dml30_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 
 	/*Total Available Pipes Support Check*/
 
-	for (i = 0; i < v->soc.num_states; i++) {
+	for (i = start_state; i < v->soc.num_states; i++) {
 		for (j = 0; j < 2; j++) {
 			if (v->TotalNumberOfActiveDPP[i][j] <= v->MaxNumDPP) {
 				v->TotalAvailablePipesSupport[i][j] = true;
@@ -4142,7 +4039,7 @@ void dml30_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 		}
 	}
 
-	for (i = 0; i < v->soc.num_states; i++) {
+	for (i = start_state; i < v->soc.num_states; i++) {
 		for (k = 0; k <= v->NumberOfActivePlanes - 1; k++) {
 			v->RequiresDSC[i][k] = false;
 			v->RequiresFEC[i][k] = false;
@@ -4270,7 +4167,7 @@ void dml30_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 			}
 		}
 	}
-	for (i = 0; i < v->soc.num_states; i++) {
+	for (i = start_state; i < v->soc.num_states; i++) {
 		v->DIOSupport[i] = true;
 		for (k = 0; k <= v->NumberOfActivePlanes - 1; k++) {
 			if (!v->skip_dio_check[k] && v->BlendingAndTiming[k] == k && (v->Output[k] == dm_dp || v->Output[k] == dm_edp || v->Output[k] == dm_hdmi)
@@ -4281,7 +4178,7 @@ void dml30_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 		}
 	}
 
-	for (i = 0; i < v->soc.num_states; ++i) {
+	for (i = start_state; i < v->soc.num_states; ++i) {
 		v->ODMCombine4To1SupportCheckOK[i] = true;
 		for (k = 0; k < v->NumberOfActivePlanes; ++k) {
 			if (v->BlendingAndTiming[k] == k && v->ODMCombineEnablePerState[i][k] == dm_odm_combine_mode_4to1
@@ -4293,7 +4190,7 @@ void dml30_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 
 	/* Skip dscclk validation: as long as dispclk is supported, dscclk is also implicitly supported */
 
-	for (i = 0; i < v->soc.num_states; i++) {
+	for (i = start_state; i < v->soc.num_states; i++) {
 		v->NotEnoughDSCUnits[i] = false;
 		v->TotalDSCUnitsRequired = 0.0;
 		for (k = 0; k <= v->NumberOfActivePlanes - 1; k++) {
@@ -4313,7 +4210,7 @@ void dml30_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 	}
 	/*DSC Delay per state*/
 
-	for (i = 0; i < v->soc.num_states; i++) {
+	for (i = start_state; i < v->soc.num_states; i++) {
 		for (k = 0; k <= v->NumberOfActivePlanes - 1; k++) {
 			if (v->OutputBppPerState[i][k] == BPP_INVALID) {
 				v->BPP = 0.0;
@@ -4364,7 +4261,7 @@ void dml30_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 
 	//Calculate Swath, DET Configuration, DCFCLKDeepSleep
 	//
-	for (i = 0; i < mode_lib->soc.num_states; ++i) {
+	for (i = start_state; i < mode_lib->soc.num_states; ++i) {
 		for (j = 0; j <= 1; ++j) {
 			for (k = 0; k < v->NumberOfActivePlanes; ++k) {
 				v->RequiredDPPCLKThisState[k] = v->RequiredDPPCLK[i][j][k];
@@ -4429,7 +4326,7 @@ void dml30_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 		v->cursor_bw[k] = v->NumberOfCursors[k] * v->CursorWidth[k][0] * v->CursorBPP[k][0] / 8.0 / (v->HTotal[k] / v->PixelClock[k]) * v->VRatio[k];
 	}
 
-	for (i = 0; i < v->soc.num_states; i++) {
+	for (i = start_state; i < v->soc.num_states; i++) {
 		for (j = 0; j < 2; j++) {
 			for (k = 0; k <= v->NumberOfActivePlanes - 1; k++) {
 				v->swath_width_luma_ub_this_state[k] = v->swath_width_luma_ub_all_states[i][j][k];
@@ -4667,7 +4564,7 @@ void dml30_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 
 	//Calculate Return BW
 
-	for (i = 0; i < mode_lib->soc.num_states; ++i) {
+	for (i = start_state; i < mode_lib->soc.num_states; ++i) {
 		for (j = 0; j <= 1; ++j) {
 			for (k = 0; k <= v->NumberOfActivePlanes - 1; k++) {
 				if (v->BlendingAndTiming[k] == k) {
@@ -4726,77 +4623,18 @@ void dml30_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 					v->UrgentOutOfOrderReturnPerChannelVMDataOnly);
 	v->FinalDRAMClockChangeLatency = (v->DRAMClockChangeLatencyOverride > 0 ? v->DRAMClockChangeLatencyOverride : v->DRAMClockChangeLatency);
 
-	for (i = 0; i < mode_lib->soc.num_states; ++i) {
+	for (i = start_state; i < mode_lib->soc.num_states; ++i) {
 		for (j = 0; j <= 1; ++j) {
 			v->DCFCLKState[i][j] = v->DCFCLKPerState[i];
 		}
 	}
 
 	if (v->UseMinimumRequiredDCFCLK == true) {
-		UseMinimumDCFCLK(
-				mode_lib,
-				v->MaxInterDCNTileRepeaters,
-				MaxPrefetchMode,
-				v->FinalDRAMClockChangeLatency,
-				v->SREnterPlusExitTime,
-				v->ReturnBusWidth,
-				v->RoundTripPingLatencyCycles,
-				ReorderingBytes,
-				v->PixelChunkSizeInKByte,
-				v->MetaChunkSize,
-				v->GPUVMEnable,
-				v->GPUVMMaxPageTableLevels,
-				v->HostVMEnable,
-				v->NumberOfActivePlanes,
-				v->HostVMMinPageSize,
-				v->HostVMMaxNonCachedPageTableLevels,
-				v->DynamicMetadataVMEnabled,
-				v->ImmediateFlipRequirement[0],
-				v->ProgressiveToInterlaceUnitInOPP,
-				v->MaxAveragePercentOfIdealSDPPortBWDisplayCanUseInNormalSystemOperation,
-				v->PercentOfIdealDRAMFabricAndSDPPortBWReceivedAfterUrgLatencyPixelMixedWithVMData,
-				v->PercentOfIdealDRAMFabricAndSDPPortBWReceivedAfterUrgLatencyVMDataOnly,
-				v->PercentOfIdealDRAMFabricAndSDPPortBWReceivedAfterUrgLatencyPixelDataOnly,
-				v->VTotal,
-				v->VActive,
-				v->DynamicMetadataTransmittedBytes,
-				v->DynamicMetadataLinesBeforeActiveRequired,
-				v->Interlace,
-				v->RequiredDPPCLK,
-				v->RequiredDISPCLK,
-				v->UrgLatency,
-				v->NoOfDPP,
-				v->ProjectedDCFCLKDeepSleep,
-				v->MaximumVStartup,
-				v->TotalVActivePixelBandwidth,
-				v->TotalVActiveCursorBandwidth,
-				v->TotalMetaRowBandwidth,
-				v->TotalDPTERowBandwidth,
-				v->TotalNumberOfActiveDPP,
-				v->TotalNumberOfDCCActiveDPP,
-				v->dpte_group_bytes,
-				v->PrefetchLinesY,
-				v->PrefetchLinesC,
-				v->swath_width_luma_ub_all_states,
-				v->swath_width_chroma_ub_all_states,
-				v->BytePerPixelY,
-				v->BytePerPixelC,
-				v->HTotal,
-				v->PixelClock,
-				v->PDEAndMetaPTEBytesPerFrame,
-				v->DPTEBytesPerRow,
-				v->MetaRowBytes,
-				v->DynamicMetadataEnable,
-				v->VActivePixelBandwidth,
-				v->VActiveCursorBandwidth,
-				v->ReadBandwidthLuma,
-				v->ReadBandwidthChroma,
-				v->DCFCLKPerState,
-				v->DCFCLKState);
+		UseMinimumDCFCLK(mode_lib, v, MaxPrefetchMode, ReorderingBytes);
 
 		if (v->ClampMinDCFCLK) {
 			/* Clamp calculated values to actual minimum */
-			for (i = 0; i < mode_lib->soc.num_states; ++i) {
+			for (i = start_state; i < mode_lib->soc.num_states; ++i) {
 				for (j = 0; j <= 1; ++j) {
 					if (v->DCFCLKState[i][j] < mode_lib->soc.min_dcfclk) {
 						v->DCFCLKState[i][j] = mode_lib->soc.min_dcfclk;
@@ -4806,7 +4644,7 @@ void dml30_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 		}
 	}
 
-	for (i = 0; i < mode_lib->soc.num_states; ++i) {
+	for (i = start_state; i < mode_lib->soc.num_states; ++i) {
 		for (j = 0; j <= 1; ++j) {
 			v->IdealSDPPortBandwidthPerState[i][j] = dml_min3(
 					v->ReturnBusWidth * v->DCFCLKState[i][j],
@@ -4824,7 +4662,7 @@ void dml30_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 
 	//Re-ordering Buffer Support Check
 
-	for (i = 0; i < mode_lib->soc.num_states; ++i) {
+	for (i = start_state; i < mode_lib->soc.num_states; ++i) {
 		for (j = 0; j <= 1; ++j) {
 			if ((v->ROBBufferSizeInKByte - v->PixelChunkSizeInKByte) * 1024 / v->ReturnBWPerState[i][j]
 					> (v->RoundTripPingLatencyCycles + 32) / v->DCFCLKState[i][j] + ReorderingBytes / v->ReturnBWPerState[i][j]) {
@@ -4842,7 +4680,7 @@ void dml30_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 		MaxTotalVActiveRDBandwidth = MaxTotalVActiveRDBandwidth + v->ReadBandwidthLuma[k] + v->ReadBandwidthChroma[k];
 	}
 
-	for (i = 0; i < mode_lib->soc.num_states; ++i) {
+	for (i = start_state; i < mode_lib->soc.num_states; ++i) {
 		for (j = 0; j <= 1; ++j) {
 			v->MaxTotalVerticalActiveAvailableBandwidth[i][j] = dml_min(
 					v->IdealSDPPortBandwidthPerState[i][j] * v->MaxAveragePercentOfIdealSDPPortBWDisplayCanUseInNormalSystemOperation / 100,
@@ -4858,7 +4696,7 @@ void dml30_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 
 	//Prefetch Check
 
-	for (i = 0; i < mode_lib->soc.num_states; ++i) {
+	for (i = start_state; i < mode_lib->soc.num_states; ++i) {
 		for (j = 0; j <= 1; ++j) {
 			int NextPrefetchModeState = MinPrefetchMode;
 
@@ -5005,7 +4843,7 @@ void dml30_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 							v->SwathHeightYThisState[k],
 							v->SwathHeightCThisState[k],
 							v->HTotal[k] / v->PixelClock[k],
-							v->UrgentLatency,
+							v->UrgLatency[i],
 							v->CursorBufferSize,
 							v->CursorWidth[k][0],
 							v->CursorBPP[k][0],
@@ -5017,7 +4855,7 @@ void dml30_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 							v->DETBufferSizeCThisState[k],
 							&v->UrgentBurstFactorCursorPre[k],
 							&v->UrgentBurstFactorLumaPre[k],
-							&v->UrgentBurstFactorChroma[k],
+							&v->UrgentBurstFactorChromaPre[k],
 							&v->NoUrgentLatencyHidingPre[k]);
 				}
 
@@ -5092,8 +4930,8 @@ void dml30_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 					}
 					v->TotImmediateFlipBytes = 0.0;
 					for (k = 0; k <= v->NumberOfActivePlanes - 1; k++) {
-						v->TotImmediateFlipBytes = v->TotImmediateFlipBytes + v->NoOfDPP[i][j][k] * v->PDEAndMetaPTEBytesPerFrame[i][j][k]
-								+ v->MetaRowBytes[i][j][k] + v->DPTEBytesPerRow[i][j][k];
+						v->TotImmediateFlipBytes = v->TotImmediateFlipBytes + v->NoOfDPP[i][j][k] * (v->PDEAndMetaPTEBytesPerFrame[i][j][k]
+								+ v->MetaRowBytes[i][j][k] + v->DPTEBytesPerRow[i][j][k]);
 					}
 
 					for (k = 0; k <= v->NumberOfActivePlanes - 1; k++) {
@@ -5230,7 +5068,7 @@ void dml30_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 
 	/*PTE Buffer Size Check*/
 
-	for (i = 0; i < v->soc.num_states; i++) {
+	for (i = start_state; i < v->soc.num_states; i++) {
 		for (j = 0; j < 2; j++) {
 			v->PTEBufferSizeNotExceeded[i][j] = true;
 			for (k = 0; k <= v->NumberOfActivePlanes - 1; k++) {
@@ -5291,7 +5129,7 @@ void dml30_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 	}
 	/*Mode Support, Voltage State and SOC Configuration*/
 
-	for (i = v->soc.num_states - 1; i >= 0; i--) {
+	for (i = v->soc.num_states - 1; i >= start_state; i--) {
 		for (j = 0; j < 2; j++) {
 			if (v->ScaleRatioAndTapsSupport == 1 && v->SourceFormatPixelAndScanSupport == 1 && v->ViewportSizeSupport[i][j] == 1
 					&& v->DIOSupport[i] == 1 && v->ODMCombine4To1SupportCheckOK[i] == 1
@@ -5313,7 +5151,7 @@ void dml30_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 	}
 	{
 		unsigned int MaximumMPCCombine = 0;
-		for (i = v->soc.num_states; i >= 0; i--) {
+		for (i = v->soc.num_states; i >= start_state; i--) {
 			if (i == v->soc.num_states || v->ModeSupport[i][0] == true || v->ModeSupport[i][1] == true) {
 				v->VoltageLevel = i;
 				v->ModeIsSupported = v->ModeSupport[i][0] == true || v->ModeSupport[i][1] == true;
@@ -6476,10 +6314,6 @@ static void CalculateSwathWidth(
 
 	for (k = 0; k < NumberOfActivePlanes; ++k) {
 		enum odm_combine_mode MainPlaneODMCombine = 0;
-		surface_width_ub_l = dml_ceil(SurfaceWidthY[k], Read256BytesBlockWidthY[k]);
-		surface_height_ub_l = dml_ceil(SurfaceHeightY[k], Read256BytesBlockHeightY[k]);
-		surface_width_ub_c = dml_ceil(SurfaceWidthC[k], Read256BytesBlockWidthC[k]);
-		surface_height_ub_c = dml_ceil(SurfaceHeightC[k], Read256BytesBlockHeightC[k]);
 
 		if (SourceScan[k] != dm_vert) {
 			SwathWidthSingleDPPY[k] = ViewportWidth[k];
@@ -6519,8 +6353,6 @@ static void CalculateSwathWidth(
 
 		surface_width_ub_l  = dml_ceil(SurfaceWidthY[k], Read256BytesBlockWidthY[k]);
 		surface_height_ub_l = dml_ceil(SurfaceHeightY[k], Read256BytesBlockHeightY[k]);
-		surface_width_ub_c  = dml_ceil(SurfaceWidthC[k], Read256BytesBlockWidthC[k]);
-		surface_height_ub_c = dml_ceil(SurfaceHeightC[k], Read256BytesBlockHeightC[k]);
 
 		if (SourceScan[k] != dm_vert) {
 			MaximumSwathHeightY[k] = Read256BytesBlockHeightY[k];
@@ -6528,6 +6360,7 @@ static void CalculateSwathWidth(
 			swath_width_luma_ub[k] = dml_min(surface_width_ub_l, (long) dml_ceil(SwathWidthY[k] - 1,
 					Read256BytesBlockWidthY[k]) + Read256BytesBlockWidthY[k]);
 			if (BytePerPixC[k] > 0) {
+				surface_width_ub_c  = dml_ceil(SurfaceWidthC[k], Read256BytesBlockWidthC[k]);
 				swath_width_chroma_ub[k] = dml_min(surface_width_ub_c, (long) dml_ceil(SwathWidthC[k] - 1,
 						Read256BytesBlockWidthC[k]) + Read256BytesBlockWidthC[k]);
 			} else {
@@ -6539,6 +6372,7 @@ static void CalculateSwathWidth(
 			swath_width_luma_ub[k] = dml_min(surface_height_ub_l, (long) dml_ceil(SwathWidthY[k] - 1,
 					Read256BytesBlockHeightY[k]) + Read256BytesBlockHeightY[k]);
 			if (BytePerPixC[k] > 0) {
+				surface_height_ub_c = dml_ceil(SurfaceHeightC[k], Read256BytesBlockHeightC[k]);
 				swath_width_chroma_ub[k] = dml_min(surface_height_ub_c, (long) dml_ceil(SwathWidthC[k] - 1,
 						Read256BytesBlockHeightC[k]) + Read256BytesBlockHeightC[k]);
 			} else {
@@ -6651,77 +6485,21 @@ static double CalculateUrgentLatency(
 	return ret;
 }
 
-
-static void UseMinimumDCFCLK(
+static noinline_for_stack void UseMinimumDCFCLK(
 		struct display_mode_lib *mode_lib,
-		int MaxInterDCNTileRepeaters,
+		struct vba_vars_st *v,
 		int MaxPrefetchMode,
-		double FinalDRAMClockChangeLatency,
-		double SREnterPlusExitTime,
-		int ReturnBusWidth,
-		int RoundTripPingLatencyCycles,
-		int ReorderingBytes,
-		int PixelChunkSizeInKByte,
-		int MetaChunkSize,
-		bool GPUVMEnable,
-		int GPUVMMaxPageTableLevels,
-		bool HostVMEnable,
-		int NumberOfActivePlanes,
-		double HostVMMinPageSize,
-		int HostVMMaxNonCachedPageTableLevels,
-		bool DynamicMetadataVMEnabled,
-		enum immediate_flip_requirement ImmediateFlipRequirement,
-		bool ProgressiveToInterlaceUnitInOPP,
-		double MaxAveragePercentOfIdealSDPPortBWDisplayCanUseInNormalSystemOperation,
-		double PercentOfIdealDRAMFabricAndSDPPortBWReceivedAfterUrgLatencyPixelMixedWithVMData,
-		double PercentOfIdealDRAMFabricAndSDPPortBWReceivedAfterUrgLatencyVMDataOnly,
-		double PercentOfIdealDRAMFabricAndSDPPortBWReceivedAfterUrgLatencyPixelDataOnly,
-		int VTotal[],
-		int VActive[],
-		int DynamicMetadataTransmittedBytes[],
-		int DynamicMetadataLinesBeforeActiveRequired[],
-		bool Interlace[],
-		double RequiredDPPCLK[][2][DC__NUM_DPP__MAX],
-		double RequiredDISPCLK[][2],
-		double UrgLatency[],
-		unsigned int NoOfDPP[][2][DC__NUM_DPP__MAX],
-		double ProjectedDCFCLKDeepSleep[][2],
-		double MaximumVStartup[][2][DC__NUM_DPP__MAX],
-		double TotalVActivePixelBandwidth[][2],
-		double TotalVActiveCursorBandwidth[][2],
-		double TotalMetaRowBandwidth[][2],
-		double TotalDPTERowBandwidth[][2],
-		unsigned int TotalNumberOfActiveDPP[][2],
-		unsigned int TotalNumberOfDCCActiveDPP[][2],
-		int dpte_group_bytes[],
-		double PrefetchLinesY[][2][DC__NUM_DPP__MAX],
-		double PrefetchLinesC[][2][DC__NUM_DPP__MAX],
-		int swath_width_luma_ub_all_states[][2][DC__NUM_DPP__MAX],
-		int swath_width_chroma_ub_all_states[][2][DC__NUM_DPP__MAX],
-		int BytePerPixelY[],
-		int BytePerPixelC[],
-		int HTotal[],
-		double PixelClock[],
-		double PDEAndMetaPTEBytesPerFrame[][2][DC__NUM_DPP__MAX],
-		double DPTEBytesPerRow[][2][DC__NUM_DPP__MAX],
-		double MetaRowBytes[][2][DC__NUM_DPP__MAX],
-		bool DynamicMetadataEnable[],
-		double VActivePixelBandwidth[][2][DC__NUM_DPP__MAX],
-		double VActiveCursorBandwidth[][2][DC__NUM_DPP__MAX],
-		double ReadBandwidthLuma[],
-		double ReadBandwidthChroma[],
-		double DCFCLKPerState[],
-		double DCFCLKState[][2])
+		int ReorderingBytes)
 {
 	double   NormalEfficiency = 0;
 	double   PTEEfficiency = 0;
 	double   TotalMaxPrefetchFlipDPTERowBandwidth[DC__VOLTAGE_STATES][2] = { { 0 } };
 	unsigned int i, j, k;
 
-	NormalEfficiency =  (HostVMEnable == true ? PercentOfIdealDRAMFabricAndSDPPortBWReceivedAfterUrgLatencyPixelMixedWithVMData
-			: PercentOfIdealDRAMFabricAndSDPPortBWReceivedAfterUrgLatencyPixelDataOnly) / 100.0;
-	PTEEfficiency =  (HostVMEnable == true ? PercentOfIdealDRAMFabricAndSDPPortBWReceivedAfterUrgLatencyVMDataOnly
-			/ PercentOfIdealDRAMFabricAndSDPPortBWReceivedAfterUrgLatencyPixelMixedWithVMData : 1.0);
+	NormalEfficiency =  (v->HostVMEnable == true ? v->PercentOfIdealDRAMFabricAndSDPPortBWReceivedAfterUrgLatencyPixelMixedWithVMData
+			: v->PercentOfIdealDRAMFabricAndSDPPortBWReceivedAfterUrgLatencyPixelDataOnly) / 100.0;
+	PTEEfficiency =  (v->HostVMEnable == true ? v->PercentOfIdealDRAMFabricAndSDPPortBWReceivedAfterUrgLatencyVMDataOnly
+			/ v->PercentOfIdealDRAMFabricAndSDPPortBWReceivedAfterUrgLatencyPixelMixedWithVMData : 1.0);
 	for (i = 0; i < mode_lib->soc.num_states; ++i) {
 		for (j = 0; j <= 1; ++j) {
 			double PixelDCFCLKCyclesRequiredInPrefetch[DC__NUM_DPP__MAX] = { 0 };
@@ -6739,58 +6517,58 @@ static void UseMinimumDCFCLK(
 			double MinimumTvmPlus2Tr0 = 0;
 
 			TotalMaxPrefetchFlipDPTERowBandwidth[i][j] = 0;
-			for (k = 0; k < NumberOfActivePlanes; ++k) {
+			for (k = 0; k < v->NumberOfActivePlanes; ++k) {
 				TotalMaxPrefetchFlipDPTERowBandwidth[i][j] = TotalMaxPrefetchFlipDPTERowBandwidth[i][j]
-					+ NoOfDPP[i][j][k] * DPTEBytesPerRow[i][j][k] / (15.75 * HTotal[k] / PixelClock[k]);
+					+ v->NoOfDPP[i][j][k] * v->DPTEBytesPerRow[i][j][k] / (15.75 * v->HTotal[k] / v->PixelClock[k]);
 			}
 
-			for (k = 0; k <= NumberOfActivePlanes - 1; ++k) {
-				NoOfDPPState[k] = NoOfDPP[i][j][k];
+			for (k = 0; k <= v->NumberOfActivePlanes - 1; ++k) {
+				NoOfDPPState[k] = v->NoOfDPP[i][j][k];
 			}
 
-			MinimumTWait = CalculateTWait(MaxPrefetchMode, FinalDRAMClockChangeLatency, UrgLatency[i], SREnterPlusExitTime);
-			NonDPTEBandwidth = TotalVActivePixelBandwidth[i][j] + TotalVActiveCursorBandwidth[i][j] + TotalMetaRowBandwidth[i][j];
-			DPTEBandwidth =  (HostVMEnable == true || ImmediateFlipRequirement == dm_immediate_flip_required) ?
-					TotalMaxPrefetchFlipDPTERowBandwidth[i][j] : TotalDPTERowBandwidth[i][j];
-			DCFCLKRequiredForAverageBandwidth = dml_max3(ProjectedDCFCLKDeepSleep[i][j],
-					(NonDPTEBandwidth + TotalDPTERowBandwidth[i][j]) / ReturnBusWidth / (MaxAveragePercentOfIdealSDPPortBWDisplayCanUseInNormalSystemOperation / 100),
-					(NonDPTEBandwidth + DPTEBandwidth / PTEEfficiency) / NormalEfficiency / ReturnBusWidth);
-
-			ExtraLatencyBytes = CalculateExtraLatencyBytes(ReorderingBytes, TotalNumberOfActiveDPP[i][j], PixelChunkSizeInKByte, TotalNumberOfDCCActiveDPP[i][j],
-					MetaChunkSize, GPUVMEnable, HostVMEnable, NumberOfActivePlanes, NoOfDPPState, dpte_group_bytes,
-					PercentOfIdealDRAMFabricAndSDPPortBWReceivedAfterUrgLatencyPixelMixedWithVMData, PercentOfIdealDRAMFabricAndSDPPortBWReceivedAfterUrgLatencyVMDataOnly,
-					HostVMMinPageSize, HostVMMaxNonCachedPageTableLevels);
-			ExtraLatencyCycles = RoundTripPingLatencyCycles + 32 + ExtraLatencyBytes / NormalEfficiency / ReturnBusWidth;
-			for (k = 0; k < NumberOfActivePlanes; ++k) {
+			MinimumTWait = CalculateTWait(MaxPrefetchMode, v->FinalDRAMClockChangeLatency, v->UrgLatency[i], v->SREnterPlusExitTime);
+			NonDPTEBandwidth = v->TotalVActivePixelBandwidth[i][j] + v->TotalVActiveCursorBandwidth[i][j] + v->TotalMetaRowBandwidth[i][j];
+			DPTEBandwidth =  (v->HostVMEnable == true || v->ImmediateFlipRequirement[0] == dm_immediate_flip_required) ?
+					TotalMaxPrefetchFlipDPTERowBandwidth[i][j] : v->TotalDPTERowBandwidth[i][j];
+			DCFCLKRequiredForAverageBandwidth = dml_max3(v->ProjectedDCFCLKDeepSleep[i][j],
+					(NonDPTEBandwidth + v->TotalDPTERowBandwidth[i][j]) / v->ReturnBusWidth / (v->MaxAveragePercentOfIdealSDPPortBWDisplayCanUseInNormalSystemOperation / 100),
+					(NonDPTEBandwidth + DPTEBandwidth / PTEEfficiency) / NormalEfficiency / v->ReturnBusWidth);
+
+			ExtraLatencyBytes = CalculateExtraLatencyBytes(ReorderingBytes, v->TotalNumberOfActiveDPP[i][j], v->PixelChunkSizeInKByte, v->TotalNumberOfDCCActiveDPP[i][j],
+					v->MetaChunkSize, v->GPUVMEnable, v->HostVMEnable, v->NumberOfActivePlanes, NoOfDPPState, v->dpte_group_bytes,
+					v->PercentOfIdealDRAMFabricAndSDPPortBWReceivedAfterUrgLatencyPixelMixedWithVMData, v->PercentOfIdealDRAMFabricAndSDPPortBWReceivedAfterUrgLatencyVMDataOnly,
+					v->HostVMMinPageSize, v->HostVMMaxNonCachedPageTableLevels);
+			ExtraLatencyCycles = v->RoundTripPingLatencyCycles + 32 + ExtraLatencyBytes / NormalEfficiency / v->ReturnBusWidth;
+			for (k = 0; k < v->NumberOfActivePlanes; ++k) {
 				double DCFCLKCyclesRequiredInPrefetch = { 0 };
 				double ExpectedPrefetchBWAcceleration = { 0 };
 				double PrefetchTime = { 0 };
 
-				PixelDCFCLKCyclesRequiredInPrefetch[k] = (PrefetchLinesY[i][j][k] * swath_width_luma_ub_all_states[i][j][k] * BytePerPixelY[k]
-					+ PrefetchLinesC[i][j][k] * swath_width_chroma_ub_all_states[i][j][k] * BytePerPixelC[k]) / NormalEfficiency / ReturnBusWidth;
-				DCFCLKCyclesRequiredInPrefetch = 2 * ExtraLatencyCycles / NoOfDPPState[k] + PDEAndMetaPTEBytesPerFrame[i][j][k] / PTEEfficiency
-					/ NormalEfficiency / ReturnBusWidth *  (GPUVMMaxPageTableLevels > 2 ? 1 : 0) + 2 * DPTEBytesPerRow[i][j][k] / PTEEfficiency
-					/ NormalEfficiency / ReturnBusWidth + 2 * MetaRowBytes[i][j][k] / NormalEfficiency / ReturnBusWidth + PixelDCFCLKCyclesRequiredInPrefetch[k];
-				PrefetchPixelLinesTime[k] = dml_max(PrefetchLinesY[i][j][k], PrefetchLinesC[i][j][k]) * HTotal[k] / PixelClock[k];
-				ExpectedPrefetchBWAcceleration = (VActivePixelBandwidth[i][j][k] + VActiveCursorBandwidth[i][j][k]) / (ReadBandwidthLuma[k] + ReadBandwidthChroma[k]);
-				DynamicMetadataVMExtraLatency[k] = (GPUVMEnable == true && DynamicMetadataEnable[k] == true && DynamicMetadataVMEnabled == true) ?
-						UrgLatency[i] * GPUVMMaxPageTableLevels *  (HostVMEnable == true ? HostVMMaxNonCachedPageTableLevels + 1 : 1) : 0;
-				PrefetchTime = (MaximumVStartup[i][j][k] - 1) * HTotal[k] / PixelClock[k] - MinimumTWait - UrgLatency[i] * ((GPUVMMaxPageTableLevels <= 2 ? GPUVMMaxPageTableLevels
-						: GPUVMMaxPageTableLevels - 2) * (HostVMEnable == true ? HostVMMaxNonCachedPageTableLevels + 1 : 1) - 1) - DynamicMetadataVMExtraLatency[k];
+				PixelDCFCLKCyclesRequiredInPrefetch[k] = (v->PrefetchLinesY[i][j][k] * v->swath_width_luma_ub_all_states[i][j][k] * v->BytePerPixelY[k]
+					+ v->PrefetchLinesC[i][j][k] * v->swath_width_chroma_ub_all_states[i][j][k] * v->BytePerPixelC[k]) / NormalEfficiency / v->ReturnBusWidth;
+				DCFCLKCyclesRequiredInPrefetch = 2 * ExtraLatencyCycles / NoOfDPPState[k] + v->PDEAndMetaPTEBytesPerFrame[i][j][k] / PTEEfficiency
+					/ NormalEfficiency / v->ReturnBusWidth *  (v->GPUVMMaxPageTableLevels > 2 ? 1 : 0) + 2 * v->DPTEBytesPerRow[i][j][k] / PTEEfficiency
+					/ NormalEfficiency / v->ReturnBusWidth + 2 * v->MetaRowBytes[i][j][k] / NormalEfficiency / v->ReturnBusWidth + PixelDCFCLKCyclesRequiredInPrefetch[k];
+				PrefetchPixelLinesTime[k] = dml_max(v->PrefetchLinesY[i][j][k], v->PrefetchLinesC[i][j][k]) * v->HTotal[k] / v->PixelClock[k];
+				ExpectedPrefetchBWAcceleration = (v->VActivePixelBandwidth[i][j][k] + v->VActiveCursorBandwidth[i][j][k]) / (v->ReadBandwidthLuma[k] + v->ReadBandwidthChroma[k]);
+				DynamicMetadataVMExtraLatency[k] = (v->GPUVMEnable == true && v->DynamicMetadataEnable[k] == true && v->DynamicMetadataVMEnabled == true) ?
+						v->UrgLatency[i] * v->GPUVMMaxPageTableLevels *  (v->HostVMEnable == true ? v->HostVMMaxNonCachedPageTableLevels + 1 : 1) : 0;
+				PrefetchTime = (v->MaximumVStartup[i][j][k] - 1) * v->HTotal[k] / v->PixelClock[k] - MinimumTWait - v->UrgLatency[i] * ((v->GPUVMMaxPageTableLevels <= 2 ? v->GPUVMMaxPageTableLevels
+						: v->GPUVMMaxPageTableLevels - 2) * (v->HostVMEnable == true ? v->HostVMMaxNonCachedPageTableLevels + 1 : 1) - 1) - DynamicMetadataVMExtraLatency[k];
 
 				if (PrefetchTime > 0) {
 					double ExpectedVRatioPrefetch = { 0 };
 					ExpectedVRatioPrefetch = PrefetchPixelLinesTime[k] / (PrefetchTime * PixelDCFCLKCyclesRequiredInPrefetch[k] / DCFCLKCyclesRequiredInPrefetch);
 					DCFCLKRequiredForPeakBandwidthPerPlane[k] = NoOfDPPState[k] * PixelDCFCLKCyclesRequiredInPrefetch[k] / PrefetchPixelLinesTime[k]
 						* dml_max(1.0, ExpectedVRatioPrefetch) * dml_max(1.0, ExpectedVRatioPrefetch / 4) * ExpectedPrefetchBWAcceleration;
-					if (HostVMEnable == true || ImmediateFlipRequirement == dm_immediate_flip_required) {
+					if (v->HostVMEnable == true || v->ImmediateFlipRequirement[0] == dm_immediate_flip_required) {
 						DCFCLKRequiredForPeakBandwidthPerPlane[k] = DCFCLKRequiredForPeakBandwidthPerPlane[k]
-							+ NoOfDPPState[k] * DPTEBandwidth / PTEEfficiency / NormalEfficiency / ReturnBusWidth;
+							+ NoOfDPPState[k] * DPTEBandwidth / PTEEfficiency / NormalEfficiency / v->ReturnBusWidth;
 					}
 				} else {
-					DCFCLKRequiredForPeakBandwidthPerPlane[k] = DCFCLKPerState[i];
+					DCFCLKRequiredForPeakBandwidthPerPlane[k] = v->DCFCLKPerState[i];
 				}
-				if (DynamicMetadataEnable[k] == true) {
+				if (v->DynamicMetadataEnable[k] == true) {
 					double TsetupPipe = { 0 };
 					double TdmbfPipe = { 0 };
 					double TdmsksPipe = { 0 };
@@ -6798,52 +6576,51 @@ static void UseMinimumDCFCLK(
 					double AllowedTimeForUrgentExtraLatency = { 0 };
 
 					CalculateDynamicMetadataParameters(
-							MaxInterDCNTileRepeaters,
-							RequiredDPPCLK[i][j][k],
-							RequiredDISPCLK[i][j],
-							ProjectedDCFCLKDeepSleep[i][j],
-							PixelClock[k],
-							HTotal[k],
-							VTotal[k] - VActive[k],
-							DynamicMetadataTransmittedBytes[k],
-							DynamicMetadataLinesBeforeActiveRequired[k],
-							Interlace[k],
-							ProgressiveToInterlaceUnitInOPP,
+							v->MaxInterDCNTileRepeaters,
+							v->RequiredDPPCLK[i][j][k],
+							v->RequiredDISPCLK[i][j],
+							v->ProjectedDCFCLKDeepSleep[i][j],
+							v->PixelClock[k],
+							v->HTotal[k],
+							v->VTotal[k] - v->VActive[k],
+							v->DynamicMetadataTransmittedBytes[k],
+							v->DynamicMetadataLinesBeforeActiveRequired[k],
+							v->Interlace[k],
+							v->ProgressiveToInterlaceUnitInOPP,
 							&TsetupPipe,
 							&TdmbfPipe,
 							&TdmecPipe,
 							&TdmsksPipe);
-					AllowedTimeForUrgentExtraLatency = MaximumVStartup[i][j][k] * HTotal[k] / PixelClock[k] - MinimumTWait - TsetupPipe
+					AllowedTimeForUrgentExtraLatency = v->MaximumVStartup[i][j][k] * v->HTotal[k] / v->PixelClock[k] - MinimumTWait - TsetupPipe
 							- TdmbfPipe - TdmecPipe - TdmsksPipe - DynamicMetadataVMExtraLatency[k];
 					if (AllowedTimeForUrgentExtraLatency > 0) {
 						DCFCLKRequiredForPeakBandwidthPerPlane[k] = dml_max(DCFCLKRequiredForPeakBandwidthPerPlane[k],
 								ExtraLatencyCycles / AllowedTimeForUrgentExtraLatency);
 					} else {
-						DCFCLKRequiredForPeakBandwidthPerPlane[k] = DCFCLKPerState[i];
+						DCFCLKRequiredForPeakBandwidthPerPlane[k] = v->DCFCLKPerState[i];
 					}
 				}
 			}
 			DCFCLKRequiredForPeakBandwidth = 0;
-			for (k = 0; k <= NumberOfActivePlanes - 1; ++k) {
+			for (k = 0; k <= v->NumberOfActivePlanes - 1; ++k) {
 				DCFCLKRequiredForPeakBandwidth = DCFCLKRequiredForPeakBandwidth + DCFCLKRequiredForPeakBandwidthPerPlane[k];
 			}
-			MinimumTvmPlus2Tr0 = UrgLatency[i] * (GPUVMEnable == true ? (HostVMEnable == true ?
-					(GPUVMMaxPageTableLevels + 2) * (HostVMMaxNonCachedPageTableLevels + 1) - 1 : GPUVMMaxPageTableLevels + 1) : 0);
-			for (k = 0; k < NumberOfActivePlanes; ++k) {
+			MinimumTvmPlus2Tr0 = v->UrgLatency[i] * (v->GPUVMEnable == true ? (v->HostVMEnable == true ?
+					(v->GPUVMMaxPageTableLevels + 2) * (v->HostVMMaxNonCachedPageTableLevels + 1) - 1 : v->GPUVMMaxPageTableLevels + 1) : 0);
+			for (k = 0; k < v->NumberOfActivePlanes; ++k) {
 				double MaximumTvmPlus2Tr0PlusTsw = { 0 };
-				MaximumTvmPlus2Tr0PlusTsw = (MaximumVStartup[i][j][k] - 2) * HTotal[k] / PixelClock[k] - MinimumTWait - DynamicMetadataVMExtraLatency[k];
+				MaximumTvmPlus2Tr0PlusTsw = (v->MaximumVStartup[i][j][k] - 2) * v->HTotal[k] / v->PixelClock[k] - MinimumTWait - DynamicMetadataVMExtraLatency[k];
 				if (MaximumTvmPlus2Tr0PlusTsw <= MinimumTvmPlus2Tr0 + PrefetchPixelLinesTime[k] / 4) {
-					DCFCLKRequiredForPeakBandwidth = DCFCLKPerState[i];
+					DCFCLKRequiredForPeakBandwidth = v->DCFCLKPerState[i];
 				} else {
 					DCFCLKRequiredForPeakBandwidth = dml_max3(DCFCLKRequiredForPeakBandwidth, 2 * ExtraLatencyCycles
 							/ (MaximumTvmPlus2Tr0PlusTsw - MinimumTvmPlus2Tr0 - PrefetchPixelLinesTime[k] / 4),
 						(2 * ExtraLatencyCycles + PixelDCFCLKCyclesRequiredInPrefetch[k]) / (MaximumTvmPlus2Tr0PlusTsw - MinimumTvmPlus2Tr0));
 				}
 			}
-			DCFCLKState[i][j] = dml_min(DCFCLKPerState[i], 1.05 * (1 + mode_lib->vba.PercentMarginOverMinimumRequiredDCFCLK / 100)
+			v->DCFCLKState[i][j] = dml_min(v->DCFCLKPerState[i], 1.05 * (1 + mode_lib->vba.PercentMarginOverMinimumRequiredDCFCLK / 100)
 					* dml_max(DCFCLKRequiredForAverageBandwidth, DCFCLKRequiredForPeakBandwidth));
 		}
 	}
 }
 
-#endif /* CONFIG_DRM_AMD_DC_DCN */
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn30/display_mode_vba_30.h b/drivers/gpu/drm/amd/display/dc/dml/dcn30/display_mode_vba_30.h
index 4e249eaabfdb..daaf0883b84d 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dcn30/display_mode_vba_30.h
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn30/display_mode_vba_30.h
@@ -39,5 +39,16 @@ double dml30_CalculateWriteBackDISPCLK(
 		long   WritebackDestinationWidth,
 		unsigned int HTotal,
 		unsigned int WritebackLineBufferSize);
+void dml30_CalculateBytePerPixelAnd256BBlockSizes(
+		enum source_format_class SourcePixelFormat,
+		enum dm_swizzle_mode SurfaceTiling,
+		unsigned int *BytePerPixelY,
+		unsigned int *BytePerPixelC,
+		double       *BytePerPixelDETY,
+		double       *BytePerPixelDETC,
+		unsigned int *BlockHeight256BytesY,
+		unsigned int *BlockHeight256BytesC,
+		unsigned int *BlockWidth256BytesY,
+		unsigned int *BlockWidth256BytesC);
 
 #endif /* __DML30_DISPLAY_MODE_VBA_H__ */
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn30/display_rq_dlg_calc_30.c b/drivers/gpu/drm/amd/display/dc/dml/dcn30/display_rq_dlg_calc_30.c
index aef854270054..4fb37df54d59 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dcn30/display_rq_dlg_calc_30.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn30/display_rq_dlg_calc_30.c
@@ -23,12 +23,11 @@
  *
  */
 
-#ifdef CONFIG_DRM_AMD_DC_DCN
-
 #include "../display_mode_lib.h"
 #include "../display_mode_vba.h"
 #include "../dml_inline_defs.h"
 #include "display_rq_dlg_calc_30.h"
+#include "display_mode_vba_30.h"
 
 static bool is_dual_plane(enum source_format_class source_format)
 {
@@ -275,96 +274,6 @@ static void handle_det_buf_split(struct display_mode_lib *mode_lib,
 		full_swath_bytes_packed_c);
 }
 
-static bool CalculateBytePerPixelAnd256BBlockSizes(
-		enum source_format_class SourcePixelFormat,
-		enum dm_swizzle_mode SurfaceTiling,
-		unsigned int *BytePerPixelY,
-		unsigned int *BytePerPixelC,
-		double       *BytePerPixelDETY,
-		double       *BytePerPixelDETC,
-		unsigned int *BlockHeight256BytesY,
-		unsigned int *BlockHeight256BytesC,
-		unsigned int *BlockWidth256BytesY,
-		unsigned int *BlockWidth256BytesC)
-{
-	if (SourcePixelFormat == dm_444_64) {
-		*BytePerPixelDETY = 8;
-		*BytePerPixelDETC = 0;
-		*BytePerPixelY = 8;
-		*BytePerPixelC = 0;
-	} else if (SourcePixelFormat == dm_444_32 || SourcePixelFormat == dm_rgbe) {
-		*BytePerPixelDETY = 4;
-		*BytePerPixelDETC = 0;
-		*BytePerPixelY = 4;
-		*BytePerPixelC = 0;
-	} else if (SourcePixelFormat == dm_444_16) {
-		*BytePerPixelDETY = 2;
-		*BytePerPixelDETC = 0;
-		*BytePerPixelY = 2;
-		*BytePerPixelC = 0;
-	} else if (SourcePixelFormat == dm_444_8) {
-		*BytePerPixelDETY = 1;
-		*BytePerPixelDETC = 0;
-		*BytePerPixelY = 1;
-		*BytePerPixelC = 0;
-	} else if (SourcePixelFormat == dm_rgbe_alpha) {
-		*BytePerPixelDETY = 4;
-		*BytePerPixelDETC = 1;
-		*BytePerPixelY = 4;
-		*BytePerPixelC = 1;
-	} else if (SourcePixelFormat == dm_420_8) {
-		*BytePerPixelDETY = 1;
-		*BytePerPixelDETC = 2;
-		*BytePerPixelY = 1;
-		*BytePerPixelC = 2;
-	} else if (SourcePixelFormat == dm_420_12) {
-		*BytePerPixelDETY = 2;
-		*BytePerPixelDETC = 4;
-		*BytePerPixelY = 2;
-		*BytePerPixelC = 4;
-	} else {
-		*BytePerPixelDETY = 4.0 / 3;
-		*BytePerPixelDETC = 8.0 / 3;
-		*BytePerPixelY = 2;
-		*BytePerPixelC = 4;
-	}
-
-	if ((SourcePixelFormat == dm_444_64 || SourcePixelFormat == dm_444_32
-			|| SourcePixelFormat == dm_444_16 || SourcePixelFormat == dm_444_8
-			|| SourcePixelFormat == dm_mono_16 || SourcePixelFormat == dm_mono_8
-			|| SourcePixelFormat == dm_rgbe)) {
-		if (SurfaceTiling == dm_sw_linear) {
-			*BlockHeight256BytesY = 1;
-		} else if (SourcePixelFormat == dm_444_64) {
-			*BlockHeight256BytesY = 4;
-		} else if (SourcePixelFormat == dm_444_8) {
-			*BlockHeight256BytesY = 16;
-		} else {
-			*BlockHeight256BytesY = 8;
-		}
-		*BlockWidth256BytesY = 256U / *BytePerPixelY / *BlockHeight256BytesY;
-		*BlockHeight256BytesC = 0;
-		*BlockWidth256BytesC = 0;
-	} else {
-		if (SurfaceTiling == dm_sw_linear) {
-			*BlockHeight256BytesY = 1;
-			*BlockHeight256BytesC = 1;
-		} else if (SourcePixelFormat == dm_rgbe_alpha) {
-			*BlockHeight256BytesY = 8;
-			*BlockHeight256BytesC = 16;
-		} else if (SourcePixelFormat == dm_420_8) {
-			*BlockHeight256BytesY = 16;
-			*BlockHeight256BytesC = 8;
-		} else {
-			*BlockHeight256BytesY = 8;
-			*BlockHeight256BytesC = 8;
-		}
-		*BlockWidth256BytesY = 256U / *BytePerPixelY / *BlockHeight256BytesY;
-		*BlockWidth256BytesC = 256U / *BytePerPixelC / *BlockHeight256BytesC;
-	}
-	return true;
-}
-
 static void get_meta_and_pte_attr(struct display_mode_lib *mode_lib,
 	display_data_rq_dlg_params_st *rq_dlg_param,
 	display_data_rq_misc_params_st *rq_misc_param,
@@ -450,7 +359,7 @@ static void get_meta_and_pte_attr(struct display_mode_lib *mode_lib,
 	double byte_per_pixel_det_y = 0;
 	double byte_per_pixel_det_c = 0;
 
-	CalculateBytePerPixelAnd256BBlockSizes((enum source_format_class)(source_format),
+	dml30_CalculateBytePerPixelAnd256BBlockSizes((enum source_format_class)(source_format),
 		(enum dm_swizzle_mode)(tiling),
 		&bytes_per_element_y,
 		&bytes_per_element_c,
@@ -483,8 +392,6 @@ static void get_meta_and_pte_attr(struct display_mode_lib *mode_lib,
 	blk_bytes = surf_linear ?
 		256 : get_blk_size_bytes((enum source_macro_tile_size) macro_tile_size);
 	log2_blk_bytes = dml_log2((double)blk_bytes);
-	log2_blk_height = 0;
-	log2_blk_width = 0;
 
 	// remember log rule
 	// "+" in log is multiply
@@ -555,8 +462,6 @@ static void get_meta_and_pte_attr(struct display_mode_lib *mode_lib,
 		- log2_meta_req_height;
 	meta_req_width = 1 << log2_meta_req_width;
 	meta_req_height = 1 << log2_meta_req_height;
-	log2_meta_row_height = 0;
-	meta_row_width_ub = 0;
 
 	// the dimensions of a meta row are meta_row_width x meta_row_height in elements.
 	// calculate upper bound of the meta_row_width
@@ -715,7 +620,7 @@ static void get_meta_and_pte_attr(struct display_mode_lib *mode_lib,
 	if (hostvm_enable)
 		rq_sizing_param->dpte_group_bytes = 512;
 	else {
-		if (!surf_linear & (log2_dpte_req_height_ptes == 0) & surf_vert) //reduced, in this case, will have page fault within a group
+		if (!surf_linear && (log2_dpte_req_height_ptes == 0) && surf_vert) //reduced, in this case, will have page fault within a group
 			rq_sizing_param->dpte_group_bytes = 512;
 		else
 			rq_sizing_param->dpte_group_bytes = 2048;
@@ -751,13 +656,12 @@ static void get_surf_rq_param(struct display_mode_lib *mode_lib,
 	bool is_chroma,
 	bool is_alpha)
 {
-	bool mode_422 = 0;
 	unsigned int vp_width = 0;
 	unsigned int vp_height = 0;
 	unsigned int data_pitch = 0;
 	unsigned int meta_pitch = 0;
 	unsigned int surface_height = 0;
-	unsigned int ppe = mode_422 ? 2 : 1;
+	unsigned int ppe = 1;
 
 	// FIXME check if ppe apply for both luma and chroma in 422 case
 	if (is_chroma | is_alpha) {
@@ -1025,7 +929,6 @@ static void dml_rq_dlg_get_dlg_params(struct display_mode_lib *mode_lib,
 	double min_dst_y_ttu_vblank = 0;
 	unsigned int dlg_vblank_start = 0;
 	bool dual_plane = false;
-	bool mode_422 = false;
 	unsigned int access_dir = 0;
 	unsigned int vp_height_l = 0;
 	unsigned int vp_width_l = 0;
@@ -1071,7 +974,7 @@ static void dml_rq_dlg_get_dlg_params(struct display_mode_lib *mode_lib,
 
 	unsigned int vstartup_start = 0;
 	unsigned int dst_x_after_scaler = 0;
-	unsigned int dst_y_after_scaler = 0;
+	int dst_y_after_scaler = 0;
 	double line_wait = 0;
 	double dst_y_prefetch = 0;
 	double dst_y_per_vm_vblank = 0;
@@ -1174,7 +1077,6 @@ static void dml_rq_dlg_get_dlg_params(struct display_mode_lib *mode_lib,
 	// Source
 	//			 dcc_en			  = src.dcc;
 	dual_plane = is_dual_plane((enum source_format_class)(src->source_format));
-	mode_422 = false; // TODO
 	access_dir = (src->source_scan == dm_vert); // vp access direction: horizontal or vertical accessed
 	vp_height_l = src->viewport_height;
 	vp_width_l = src->viewport_width;
@@ -1262,6 +1164,8 @@ static void dml_rq_dlg_get_dlg_params(struct display_mode_lib *mode_lib,
 
 	dst_x_after_scaler = get_dst_x_after_scaler(mode_lib, e2e_pipe_param, num_pipes, pipe_idx);
 	dst_y_after_scaler = get_dst_y_after_scaler(mode_lib, e2e_pipe_param, num_pipes, pipe_idx);
+	if (dst_y_after_scaler < 0)
+		dst_y_after_scaler = 0;
 
 	// do some adjustment on the dst_after scaler to account for odm combine mode
 	dml_print("DML_DLG: %s: input dst_x_after_scaler                     = %d\n",
@@ -1390,18 +1294,8 @@ static void dml_rq_dlg_get_dlg_params(struct display_mode_lib *mode_lib,
 	dpte_row_height_l = rq_dlg_param.rq_l.dpte_row_height;
 	dpte_row_height_c = rq_dlg_param.rq_c.dpte_row_height;
 
-	if (mode_422) {
-		swath_width_pixels_ub_l = swath_width_ub_l * 2;  // *2 for 2 pixel per element
-		swath_width_pixels_ub_c = swath_width_ub_c * 2;
-	} else {
-		swath_width_pixels_ub_l = swath_width_ub_l * 1;
-		swath_width_pixels_ub_c = swath_width_ub_c * 1;
-	}
-
-	hscale_pixel_rate_l = 0.;
-	hscale_pixel_rate_c = 0.;
-	min_hratio_fact_l = 1.0;
-	min_hratio_fact_c = 1.0;
+	swath_width_pixels_ub_l = swath_width_ub_l;
+	swath_width_pixels_ub_c = swath_width_ub_c;
 
 	if (hratio_l <= 1)
 		min_hratio_fact_l = 2.0;
@@ -1668,6 +1562,7 @@ static void dml_rq_dlg_get_dlg_params(struct display_mode_lib *mode_lib,
 	dml_print("DML_DLG: %s: disp_dlg_regs->dst_y_per_row_vblank = 0x%x\n", __func__, disp_dlg_regs->dst_y_per_row_vblank);
 	dml_print("DML_DLG: %s: disp_dlg_regs->dst_y_per_vm_flip    = 0x%x\n", __func__, disp_dlg_regs->dst_y_per_vm_flip);
 	dml_print("DML_DLG: %s: disp_dlg_regs->dst_y_per_row_flip   = 0x%x\n", __func__, disp_dlg_regs->dst_y_per_row_flip);
+
 	disp_dlg_regs->refcyc_per_pte_group_vblank_l =
 		(unsigned int)(dst_y_per_row_vblank * (double)htotal
 			* ref_freq_to_pix_freq / (double)dpte_groups_per_row_ub_l);
@@ -1858,8 +1753,6 @@ void dml30_rq_dlg_get_dlg_reg(struct display_mode_lib *mode_lib,
 	dlg_sys_param.total_flip_bytes = get_total_immediate_flip_bytes(mode_lib,
 		e2e_pipe_param,
 		num_pipes);
-	dlg_sys_param.t_srx_delay_us = mode_lib->ip.dcfclk_cstate_latency
-		/ dlg_sys_param.deepsleep_dcfclk_mhz; // TODO: Deprecated
 
 	print__dlg_sys_params_st(mode_lib, &dlg_sys_param);
 
@@ -1883,4 +1776,3 @@ void dml30_rq_dlg_get_dlg_reg(struct display_mode_lib *mode_lib,
 	dml_print("DML_DLG: Calculation for pipe[%d] end\n", pipe_idx);
 }
 
-#endif
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn301/dcn301_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn301/dcn301_fpu.c
index 94c32832a0e7..1aaa77265eed 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dcn301/dcn301_fpu.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn301/dcn301_fpu.c
@@ -26,6 +26,7 @@
 #include "clk_mgr.h"
 #include "dcn20/dcn20_resource.h"
 #include "dcn301/dcn301_resource.h"
+#include "clk_mgr/dcn301/vg_clk_mgr.h"
 
 #include "dml/dcn20/dcn20_fpu.h"
 #include "dcn301_fpu.h"
@@ -214,6 +215,80 @@ struct _vcs_dpi_soc_bounding_box_st dcn3_01_soc = {
 	.urgent_latency_adjustment_fabric_clock_reference_mhz = 0,
 };
 
+struct wm_table ddr4_wm_table = {
+	.entries = {
+		{
+			.wm_inst = WM_A,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.72,
+			.sr_exit_time_us = 6.09,
+			.sr_enter_plus_exit_time_us = 7.14,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_B,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.72,
+			.sr_exit_time_us = 10.12,
+			.sr_enter_plus_exit_time_us = 11.48,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_C,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.72,
+			.sr_exit_time_us = 10.12,
+			.sr_enter_plus_exit_time_us = 11.48,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_D,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.72,
+			.sr_exit_time_us = 10.12,
+			.sr_enter_plus_exit_time_us = 11.48,
+			.valid = true,
+		},
+	}
+};
+
+struct wm_table lpddr5_wm_table = {
+	.entries = {
+		{
+			.wm_inst = WM_A,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.65333,
+			.sr_exit_time_us = 13.5,
+			.sr_enter_plus_exit_time_us = 16.5,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_B,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.65333,
+			.sr_exit_time_us = 13.5,
+			.sr_enter_plus_exit_time_us = 16.5,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_C,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.65333,
+			.sr_exit_time_us = 13.5,
+			.sr_enter_plus_exit_time_us = 16.5,
+			.valid = true,
+		},
+		{
+			.wm_inst = WM_D,
+			.wm_type = WM_TYPE_PSTATE_CHG,
+			.pstate_latency_us = 11.65333,
+			.sr_exit_time_us = 13.5,
+			.sr_enter_plus_exit_time_us = 16.5,
+			.valid = true,
+		},
+	}
+};
+
 static void calculate_wm_set_for_vlevel(int vlevel,
 		struct wm_range_table_entry *table_entry,
 		struct dcn_watermarks *wm_set,
@@ -245,61 +320,82 @@ static void calculate_wm_set_for_vlevel(int vlevel,
 
 }
 
-void dcn301_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params)
+void dcn301_fpu_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params)
 {
+	struct _vcs_dpi_voltage_scaling_st *s = dc->scratch.update_bw_bounding_box.clock_limits;
 	struct dcn301_resource_pool *pool = TO_DCN301_RES_POOL(dc->res_pool);
 	struct clk_limit_table *clk_table = &bw_params->clk_table;
-	struct _vcs_dpi_voltage_scaling_st clock_limits[DC__VOLTAGE_STATES];
 	unsigned int i, closest_clk_lvl;
-	int j;
+	int j = 0, max_dispclk_mhz = 0, max_dppclk_mhz = 0;
 
 	dc_assert_fp_enabled();
 
+	memcpy(s, dcn3_01_soc.clock_limits, sizeof(dcn3_01_soc.clock_limits));
+
 	/* Default clock levels are used for diags, which may lead to overclocking. */
-	if (!IS_DIAG_DC(dc->ctx->dce_environment)) {
-		dcn3_01_ip.max_num_otg = pool->base.res_cap->num_timing_generator;
-		dcn3_01_ip.max_num_dpp = pool->base.pipe_count;
-		dcn3_01_soc.num_chans = bw_params->num_channels;
-
-		ASSERT(clk_table->num_entries);
-		for (i = 0; i < clk_table->num_entries; i++) {
-			/* loop backwards*/
-			for (closest_clk_lvl = 0, j = dcn3_01_soc.num_states - 1; j >= 0; j--) {
-				if ((unsigned int) dcn3_01_soc.clock_limits[j].dcfclk_mhz <= clk_table->entries[i].dcfclk_mhz) {
-					closest_clk_lvl = j;
-					break;
-				}
-			}
+	dcn3_01_ip.max_num_otg = pool->base.res_cap->num_timing_generator;
+	dcn3_01_ip.max_num_dpp = pool->base.pipe_count;
+	dcn3_01_soc.num_chans = bw_params->num_channels;
+
+	ASSERT(clk_table->num_entries);
+
+	/* Prepass to find max clocks independent of voltage level. */
+	for (i = 0; i < clk_table->num_entries; ++i) {
+		if (clk_table->entries[i].dispclk_mhz > max_dispclk_mhz)
+			max_dispclk_mhz = clk_table->entries[i].dispclk_mhz;
+		if (clk_table->entries[i].dppclk_mhz > max_dppclk_mhz)
+			max_dppclk_mhz = clk_table->entries[i].dppclk_mhz;
+	}
 
-			clock_limits[i].state = i;
-			clock_limits[i].dcfclk_mhz = clk_table->entries[i].dcfclk_mhz;
-			clock_limits[i].fabricclk_mhz = clk_table->entries[i].fclk_mhz;
-			clock_limits[i].socclk_mhz = clk_table->entries[i].socclk_mhz;
-			clock_limits[i].dram_speed_mts = clk_table->entries[i].memclk_mhz * 2;
-
-			clock_limits[i].dispclk_mhz = dcn3_01_soc.clock_limits[closest_clk_lvl].dispclk_mhz;
-			clock_limits[i].dppclk_mhz = dcn3_01_soc.clock_limits[closest_clk_lvl].dppclk_mhz;
-			clock_limits[i].dram_bw_per_chan_gbps = dcn3_01_soc.clock_limits[closest_clk_lvl].dram_bw_per_chan_gbps;
-			clock_limits[i].dscclk_mhz = dcn3_01_soc.clock_limits[closest_clk_lvl].dscclk_mhz;
-			clock_limits[i].dtbclk_mhz = dcn3_01_soc.clock_limits[closest_clk_lvl].dtbclk_mhz;
-			clock_limits[i].phyclk_d18_mhz = dcn3_01_soc.clock_limits[closest_clk_lvl].phyclk_d18_mhz;
-			clock_limits[i].phyclk_mhz = dcn3_01_soc.clock_limits[closest_clk_lvl].phyclk_mhz;
+	for (i = 0; i < clk_table->num_entries; i++) {
+		/* loop backwards*/
+		for (closest_clk_lvl = 0, j = dcn3_01_soc.num_states - 1; j >= 0; j--) {
+			if ((unsigned int) dcn3_01_soc.clock_limits[j].dcfclk_mhz <= clk_table->entries[i].dcfclk_mhz) {
+				closest_clk_lvl = j;
+				break;
+			}
 		}
 
-		for (i = 0; i < clk_table->num_entries; i++)
-			dcn3_01_soc.clock_limits[i] = clock_limits[i];
+		s[i].state = i;
+		s[i].dcfclk_mhz = clk_table->entries[i].dcfclk_mhz;
+		s[i].fabricclk_mhz = clk_table->entries[i].fclk_mhz;
+		s[i].socclk_mhz = clk_table->entries[i].socclk_mhz;
+		s[i].dram_speed_mts = clk_table->entries[i].memclk_mhz * 2;
+
+		/* Clocks independent of voltage level. */
+		s[i].dispclk_mhz = max_dispclk_mhz ? max_dispclk_mhz :
+			dcn3_01_soc.clock_limits[closest_clk_lvl].dispclk_mhz;
+
+		s[i].dppclk_mhz = max_dppclk_mhz ? max_dppclk_mhz :
+			dcn3_01_soc.clock_limits[closest_clk_lvl].dppclk_mhz;
+
+		s[i].dram_bw_per_chan_gbps =
+			dcn3_01_soc.clock_limits[closest_clk_lvl].dram_bw_per_chan_gbps;
+		s[i].dscclk_mhz = dcn3_01_soc.clock_limits[closest_clk_lvl].dscclk_mhz;
+		s[i].dtbclk_mhz = dcn3_01_soc.clock_limits[closest_clk_lvl].dtbclk_mhz;
+		s[i].phyclk_d18_mhz =
+			dcn3_01_soc.clock_limits[closest_clk_lvl].phyclk_d18_mhz;
+		s[i].phyclk_mhz = dcn3_01_soc.clock_limits[closest_clk_lvl].phyclk_mhz;
+	}
 
-		if (clk_table->num_entries) {
-			dcn3_01_soc.num_states = clk_table->num_entries;
-			/* duplicate last level */
-			dcn3_01_soc.clock_limits[dcn3_01_soc.num_states] = dcn3_01_soc.clock_limits[dcn3_01_soc.num_states - 1];
-			dcn3_01_soc.clock_limits[dcn3_01_soc.num_states].state = dcn3_01_soc.num_states;
-		}
+	if (clk_table->num_entries) {
+		dcn3_01_soc.num_states = clk_table->num_entries;
+		/* duplicate last level */
+		s[dcn3_01_soc.num_states] =
+			dcn3_01_soc.clock_limits[dcn3_01_soc.num_states - 1];
+		s[dcn3_01_soc.num_states].state = dcn3_01_soc.num_states;
 	}
 
+	memcpy(dcn3_01_soc.clock_limits, s, sizeof(dcn3_01_soc.clock_limits));
+
 	dcn3_01_soc.dispclk_dppclk_vco_speed_mhz = dc->clk_mgr->dentist_vco_freq_khz / 1000.0;
 	dc->dml.soc.dispclk_dppclk_vco_speed_mhz = dc->clk_mgr->dentist_vco_freq_khz / 1000.0;
 
+	if ((int)(dcn3_01_soc.dram_clock_change_latency_us * 1000)
+				!= dc->debug.dram_clock_change_latency_ns
+			&& dc->debug.dram_clock_change_latency_ns) {
+		dcn3_01_soc.dram_clock_change_latency_us = dc->debug.dram_clock_change_latency_ns / 1000.0;
+	}
 	dml_init_instance(&dc->dml, &dcn3_01_soc, &dcn3_01_ip, DML_PROJECT_DCN30);
 }
 
@@ -327,7 +423,7 @@ void dcn301_fpu_init_soc_bounding_box(struct bp_soc_bb_info bb_info)
 		dcn3_01_soc.sr_exit_time_us = bb_info.dram_sr_exit_latency_100ns * 10;
 }
 
-void dcn301_calculate_wm_and_dlg(struct dc *dc,
+void dcn301_fpu_calculate_wm_and_dlg(struct dc *dc,
 		struct dc_state *context,
 		display_e2e_pipe_params_st *pipes,
 		int pipe_cnt,
@@ -353,12 +449,12 @@ void dcn301_calculate_wm_and_dlg(struct dc *dc,
 						&context->bw_ctx.dml, pipes, pipe_cnt);
 	/* WM Set C */
 	table_entry = &bw_params->wm_table.entries[WM_C];
-	vlevel = min(max(vlevel_req, 2), vlevel_max);
+	vlevel = clamp(vlevel_req, 2, vlevel_max);
 	calculate_wm_set_for_vlevel(vlevel, table_entry, &context->bw_ctx.bw.dcn.watermarks.c,
 						&context->bw_ctx.dml, pipes, pipe_cnt);
 	/* WM Set B */
 	table_entry = &bw_params->wm_table.entries[WM_B];
-	vlevel = min(max(vlevel_req, 1), vlevel_max);
+	vlevel = clamp(vlevel_req, 1, vlevel_max);
 	calculate_wm_set_for_vlevel(vlevel, table_entry, &context->bw_ctx.bw.dcn.watermarks.b,
 						&context->bw_ctx.dml, pipes, pipe_cnt);
 
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn301/dcn301_fpu.h b/drivers/gpu/drm/amd/display/dc/dml/dcn301/dcn301_fpu.h
index fc7065d17842..3e103e23dc6f 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dcn301/dcn301_fpu.h
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn301/dcn301_fpu.h
@@ -26,15 +26,14 @@
 #ifndef __DCN301_FPU_H__
 #define __DCN301_FPU_H__
 
-void dcn301_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params);
+void dcn301_fpu_init_soc_bounding_box(struct bp_soc_bb_info bb_info);
+void dcn301_fpu_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params);
 
 void dcn301_fpu_set_wm_ranges(int i,
 	struct pp_smu_wm_range_sets *ranges,
 	struct _vcs_dpi_soc_bounding_box_st *loaded_bb);
 
-void dcn301_fpu_init_soc_bounding_box(struct bp_soc_bb_info bb_info);
-
-void dcn301_calculate_wm_and_dlg(struct dc *dc,
+void dcn301_fpu_calculate_wm_and_dlg(struct dc *dc,
 		struct dc_state *context,
 		display_e2e_pipe_params_st *pipes,
 		int pipe_cnt,
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn302/dcn302_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn302/dcn302_fpu.c
new file mode 100644
index 000000000000..8d7c59ec701d
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn302/dcn302_fpu.c
@@ -0,0 +1,367 @@
+/*
+ * Copyright 2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#include "resource.h"
+#include "clk_mgr.h"
+#include "dcn20/dcn20_resource.h"
+#include "dcn302/dcn302_resource.h"
+
+#include "dml/dcn20/dcn20_fpu.h"
+#include "dcn302_fpu.h"
+
+struct _vcs_dpi_ip_params_st dcn3_02_ip = {
+		.use_min_dcfclk = 0,
+		.clamp_min_dcfclk = 0,
+		.odm_capable = 1,
+		.gpuvm_enable = 1,
+		.hostvm_enable = 0,
+		.gpuvm_max_page_table_levels = 4,
+		.hostvm_max_page_table_levels = 4,
+		.hostvm_cached_page_table_levels = 0,
+		.pte_group_size_bytes = 2048,
+		.num_dsc = 5,
+		.rob_buffer_size_kbytes = 184,
+		.det_buffer_size_kbytes = 184,
+		.dpte_buffer_size_in_pte_reqs_luma = 64,
+		.dpte_buffer_size_in_pte_reqs_chroma = 34,
+		.pde_proc_buffer_size_64k_reqs = 48,
+		.dpp_output_buffer_pixels = 2560,
+		.opp_output_buffer_lines = 1,
+		.pixel_chunk_size_kbytes = 8,
+		.pte_enable = 1,
+		.max_page_table_levels = 2,
+		.pte_chunk_size_kbytes = 2,  // ?
+		.meta_chunk_size_kbytes = 2,
+		.writeback_chunk_size_kbytes = 8,
+		.line_buffer_size_bits = 789504,
+		.is_line_buffer_bpp_fixed = 0,  // ?
+		.line_buffer_fixed_bpp = 0,     // ?
+		.dcc_supported = true,
+		.writeback_interface_buffer_size_kbytes = 90,
+		.writeback_line_buffer_buffer_size = 0,
+		.max_line_buffer_lines = 12,
+		.writeback_luma_buffer_size_kbytes = 12,  // writeback_line_buffer_buffer_size = 656640
+		.writeback_chroma_buffer_size_kbytes = 8,
+		.writeback_chroma_line_buffer_width_pixels = 4,
+		.writeback_max_hscl_ratio = 1,
+		.writeback_max_vscl_ratio = 1,
+		.writeback_min_hscl_ratio = 1,
+		.writeback_min_vscl_ratio = 1,
+		.writeback_max_hscl_taps = 1,
+		.writeback_max_vscl_taps = 1,
+		.writeback_line_buffer_luma_buffer_size = 0,
+		.writeback_line_buffer_chroma_buffer_size = 14643,
+		.cursor_buffer_size = 8,
+		.cursor_chunk_size = 2,
+		.max_num_otg = 5,
+		.max_num_dpp = 5,
+		.max_num_wb = 1,
+		.max_dchub_pscl_bw_pix_per_clk = 4,
+		.max_pscl_lb_bw_pix_per_clk = 2,
+		.max_lb_vscl_bw_pix_per_clk = 4,
+		.max_vscl_hscl_bw_pix_per_clk = 4,
+		.max_hscl_ratio = 6,
+		.max_vscl_ratio = 6,
+		.hscl_mults = 4,
+		.vscl_mults = 4,
+		.max_hscl_taps = 8,
+		.max_vscl_taps = 8,
+		.dispclk_ramp_margin_percent = 1,
+		.underscan_factor = 1.11,
+		.min_vblank_lines = 32,
+		.dppclk_delay_subtotal = 46,
+		.dynamic_metadata_vm_enabled = true,
+		.dppclk_delay_scl_lb_only = 16,
+		.dppclk_delay_scl = 50,
+		.dppclk_delay_cnvc_formatter = 27,
+		.dppclk_delay_cnvc_cursor = 6,
+		.dispclk_delay_subtotal = 119,
+		.dcfclk_cstate_latency = 5.2, // SRExitTime
+		.max_inter_dcn_tile_repeaters = 8,
+		.max_num_hdmi_frl_outputs = 1,
+		.odm_combine_4to1_supported = true,
+
+		.xfc_supported = false,
+		.xfc_fill_bw_overhead_percent = 10.0,
+		.xfc_fill_constant_bytes = 0,
+		.gfx7_compat_tiling_supported = 0,
+		.number_of_cursors = 1,
+};
+
+struct _vcs_dpi_soc_bounding_box_st dcn3_02_soc = {
+		.clock_limits = {
+				{
+						.state = 0,
+						.dispclk_mhz = 562.0,
+						.dppclk_mhz = 300.0,
+						.phyclk_mhz = 300.0,
+						.phyclk_d18_mhz = 667.0,
+						.dscclk_mhz = 405.6,
+				},
+		},
+
+		.min_dcfclk = 500.0, /* TODO: set this to actual min DCFCLK */
+		.num_states = 1,
+		.sr_exit_time_us = 26.5,
+		.sr_enter_plus_exit_time_us = 31,
+		.urgent_latency_us = 4.0,
+		.urgent_latency_pixel_data_only_us = 4.0,
+		.urgent_latency_pixel_mixed_with_vm_data_us = 4.0,
+		.urgent_latency_vm_data_only_us = 4.0,
+		.urgent_out_of_order_return_per_channel_pixel_only_bytes = 4096,
+		.urgent_out_of_order_return_per_channel_pixel_and_vm_bytes = 4096,
+		.urgent_out_of_order_return_per_channel_vm_only_bytes = 4096,
+		.pct_ideal_dram_sdp_bw_after_urgent_pixel_only = 80.0,
+		.pct_ideal_dram_sdp_bw_after_urgent_pixel_and_vm = 60.0,
+		.pct_ideal_dram_sdp_bw_after_urgent_vm_only = 40.0,
+		.max_avg_sdp_bw_use_normal_percent = 60.0,
+		.max_avg_dram_bw_use_normal_percent = 40.0,
+		.writeback_latency_us = 12.0,
+		.max_request_size_bytes = 256,
+		.fabric_datapath_to_dcn_data_return_bytes = 64,
+		.dcn_downspread_percent = 0.5,
+		.downspread_percent = 0.38,
+		.dram_page_open_time_ns = 50.0,
+		.dram_rw_turnaround_time_ns = 17.5,
+		.dram_return_buffer_per_channel_bytes = 8192,
+		.round_trip_ping_latency_dcfclk_cycles = 156,
+		.urgent_out_of_order_return_per_channel_bytes = 4096,
+		.channel_interleave_bytes = 256,
+		.num_banks = 8,
+		.gpuvm_min_page_size_bytes = 4096,
+		.hostvm_min_page_size_bytes = 4096,
+		.dram_clock_change_latency_us = 404,
+		.dummy_pstate_latency_us = 5,
+		.writeback_dram_clock_change_latency_us = 23.0,
+		.return_bus_width_bytes = 64,
+		.dispclk_dppclk_vco_speed_mhz = 3650,
+		.xfc_bus_transport_time_us = 20,      // ?
+		.xfc_xbuf_latency_tolerance_us = 4,  // ?
+		.use_urgent_burst_bw = 1,            // ?
+		.do_urgent_latency_adjustment = true,
+		.urgent_latency_adjustment_fabric_clock_component_us = 1.0,
+		.urgent_latency_adjustment_fabric_clock_reference_mhz = 1000,
+};
+
+static void dcn302_get_optimal_dcfclk_fclk_for_uclk(unsigned int uclk_mts,
+	unsigned int *optimal_dcfclk,
+	unsigned int *optimal_fclk)
+{
+
+			double bw_from_dram, bw_from_dram1, bw_from_dram2;
+
+			bw_from_dram1 = uclk_mts * dcn3_02_soc.num_chans *
+				dcn3_02_soc.dram_channel_width_bytes *
+				(dcn3_02_soc.max_avg_dram_bw_use_normal_percent / 100);
+			bw_from_dram2 = uclk_mts * dcn3_02_soc.num_chans *
+				dcn3_02_soc.dram_channel_width_bytes *
+				(dcn3_02_soc.max_avg_sdp_bw_use_normal_percent / 100);
+
+			bw_from_dram = (bw_from_dram1 < bw_from_dram2) ? bw_from_dram1 : bw_from_dram2;
+
+			if (optimal_fclk)
+				*optimal_fclk = bw_from_dram /
+				(dcn3_02_soc.fabric_datapath_to_dcn_data_return_bytes *
+				 (dcn3_02_soc.max_avg_sdp_bw_use_normal_percent / 100));
+
+			if (optimal_dcfclk)
+				*optimal_dcfclk =  bw_from_dram /
+				(dcn3_02_soc.return_bus_width_bytes *
+				 (dcn3_02_soc.max_avg_sdp_bw_use_normal_percent / 100));
+}
+
+void dcn302_fpu_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params)
+{
+	unsigned int i, j;
+	unsigned int num_states = 0;
+
+	unsigned int dcfclk_mhz[DC__VOLTAGE_STATES] = {0};
+	unsigned int dram_speed_mts[DC__VOLTAGE_STATES] = {0};
+	unsigned int optimal_uclk_for_dcfclk_sta_targets[DC__VOLTAGE_STATES] = {0};
+	unsigned int optimal_dcfclk_for_uclk[DC__VOLTAGE_STATES] = {0};
+
+	unsigned int dcfclk_sta_targets[DC__VOLTAGE_STATES] = {694, 875, 1000, 1200};
+	unsigned int num_dcfclk_sta_targets = 4;
+	unsigned int num_uclk_states;
+
+	dc_assert_fp_enabled();
+
+	if (dc->ctx->dc_bios->vram_info.num_chans)
+		dcn3_02_soc.num_chans = dc->ctx->dc_bios->vram_info.num_chans;
+
+	if (dc->ctx->dc_bios->vram_info.dram_channel_width_bytes)
+		dcn3_02_soc.dram_channel_width_bytes = dc->ctx->dc_bios->vram_info.dram_channel_width_bytes;
+
+	dcn3_02_soc.dispclk_dppclk_vco_speed_mhz = dc->clk_mgr->dentist_vco_freq_khz / 1000.0;
+	dc->dml.soc.dispclk_dppclk_vco_speed_mhz = dc->clk_mgr->dentist_vco_freq_khz / 1000.0;
+
+	if (bw_params->clk_table.entries[0].memclk_mhz) {
+		int max_dcfclk_mhz = 0, max_dispclk_mhz = 0, max_dppclk_mhz = 0, max_phyclk_mhz = 0;
+
+		for (i = 0; i < MAX_NUM_DPM_LVL; i++) {
+			if (bw_params->clk_table.entries[i].dcfclk_mhz > max_dcfclk_mhz)
+				max_dcfclk_mhz = bw_params->clk_table.entries[i].dcfclk_mhz;
+			if (bw_params->clk_table.entries[i].dispclk_mhz > max_dispclk_mhz)
+				max_dispclk_mhz = bw_params->clk_table.entries[i].dispclk_mhz;
+			if (bw_params->clk_table.entries[i].dppclk_mhz > max_dppclk_mhz)
+				max_dppclk_mhz = bw_params->clk_table.entries[i].dppclk_mhz;
+			if (bw_params->clk_table.entries[i].phyclk_mhz > max_phyclk_mhz)
+				max_phyclk_mhz = bw_params->clk_table.entries[i].phyclk_mhz;
+		}
+		if (!max_dcfclk_mhz)
+			max_dcfclk_mhz = dcn3_02_soc.clock_limits[0].dcfclk_mhz;
+		if (!max_dispclk_mhz)
+			max_dispclk_mhz = dcn3_02_soc.clock_limits[0].dispclk_mhz;
+		if (!max_dppclk_mhz)
+			max_dppclk_mhz = dcn3_02_soc.clock_limits[0].dppclk_mhz;
+		if (!max_phyclk_mhz)
+			max_phyclk_mhz = dcn3_02_soc.clock_limits[0].phyclk_mhz;
+
+		if (max_dcfclk_mhz > dcfclk_sta_targets[num_dcfclk_sta_targets-1]) {
+			/* If max DCFCLK is greater than the max DCFCLK STA target, insert into the DCFCLK STA target array */
+			dcfclk_sta_targets[num_dcfclk_sta_targets] = max_dcfclk_mhz;
+			num_dcfclk_sta_targets++;
+		} else if (max_dcfclk_mhz < dcfclk_sta_targets[num_dcfclk_sta_targets-1]) {
+			/* If max DCFCLK is less than the max DCFCLK STA target, cap values and remove duplicates */
+			for (i = 0; i < num_dcfclk_sta_targets; i++) {
+				if (dcfclk_sta_targets[i] > max_dcfclk_mhz) {
+					dcfclk_sta_targets[i] = max_dcfclk_mhz;
+					break;
+				}
+			}
+			/* Update size of array since we "removed" duplicates */
+			num_dcfclk_sta_targets = i + 1;
+		}
+
+		num_uclk_states = bw_params->clk_table.num_entries;
+
+		/* Calculate optimal dcfclk for each uclk */
+		for (i = 0; i < num_uclk_states; i++) {
+			dcn302_get_optimal_dcfclk_fclk_for_uclk(bw_params->clk_table.entries[i].memclk_mhz * 16,
+					&optimal_dcfclk_for_uclk[i], NULL);
+			if (optimal_dcfclk_for_uclk[i] < bw_params->clk_table.entries[0].dcfclk_mhz)
+				optimal_dcfclk_for_uclk[i] = bw_params->clk_table.entries[0].dcfclk_mhz;
+		}
+
+		/* Calculate optimal uclk for each dcfclk sta target */
+		for (i = 0; i < num_dcfclk_sta_targets; i++) {
+			for (j = 0; j < num_uclk_states; j++) {
+				if (dcfclk_sta_targets[i] < optimal_dcfclk_for_uclk[j]) {
+					optimal_uclk_for_dcfclk_sta_targets[i] =
+							bw_params->clk_table.entries[j].memclk_mhz * 16;
+					break;
+				}
+			}
+		}
+
+		i = 0;
+		j = 0;
+		/* create the final dcfclk and uclk table */
+		while (i < num_dcfclk_sta_targets && j < num_uclk_states && num_states < DC__VOLTAGE_STATES) {
+			if (dcfclk_sta_targets[i] < optimal_dcfclk_for_uclk[j]) {
+				dcfclk_mhz[num_states] = dcfclk_sta_targets[i];
+				dram_speed_mts[num_states++] = optimal_uclk_for_dcfclk_sta_targets[i++];
+			} else {
+				if (j < num_uclk_states && optimal_dcfclk_for_uclk[j] <= max_dcfclk_mhz) {
+					dcfclk_mhz[num_states] = optimal_dcfclk_for_uclk[j];
+					dram_speed_mts[num_states++] = bw_params->clk_table.entries[j++].memclk_mhz * 16;
+				} else {
+					j = num_uclk_states;
+				}
+			}
+		}
+
+		while (i < num_dcfclk_sta_targets && num_states < DC__VOLTAGE_STATES) {
+			dcfclk_mhz[num_states] = dcfclk_sta_targets[i];
+			dram_speed_mts[num_states++] = optimal_uclk_for_dcfclk_sta_targets[i++];
+		}
+
+		while (j < num_uclk_states && num_states < DC__VOLTAGE_STATES &&
+				optimal_dcfclk_for_uclk[j] <= max_dcfclk_mhz) {
+			dcfclk_mhz[num_states] = optimal_dcfclk_for_uclk[j];
+			dram_speed_mts[num_states++] = bw_params->clk_table.entries[j++].memclk_mhz * 16;
+		}
+
+		/* bw_params->clk_table.entries[MAX_NUM_DPM_LVL].
+		 * MAX_NUM_DPM_LVL is 8.
+		 * dcn3_02_soc.clock_limits[DC__VOLTAGE_STATES].
+		 * DC__VOLTAGE_STATES is 40.
+		 */
+		if (num_states > MAX_NUM_DPM_LVL) {
+			ASSERT(0);
+			return;
+		}
+
+		dcn3_02_soc.num_states = num_states;
+		for (i = 0; i < dcn3_02_soc.num_states; i++) {
+			dcn3_02_soc.clock_limits[i].state = i;
+			dcn3_02_soc.clock_limits[i].dcfclk_mhz = dcfclk_mhz[i];
+			dcn3_02_soc.clock_limits[i].fabricclk_mhz = dcfclk_mhz[i];
+			dcn3_02_soc.clock_limits[i].dram_speed_mts = dram_speed_mts[i];
+
+			/* Fill all states with max values of all other clocks */
+			dcn3_02_soc.clock_limits[i].dispclk_mhz = max_dispclk_mhz;
+			dcn3_02_soc.clock_limits[i].dppclk_mhz  = max_dppclk_mhz;
+			dcn3_02_soc.clock_limits[i].phyclk_mhz  = max_phyclk_mhz;
+			/* Populate from bw_params for DTBCLK, SOCCLK */
+			if (!bw_params->clk_table.entries[i].dtbclk_mhz && i > 0)
+				dcn3_02_soc.clock_limits[i].dtbclk_mhz  = dcn3_02_soc.clock_limits[i-1].dtbclk_mhz;
+			else
+				dcn3_02_soc.clock_limits[i].dtbclk_mhz  = bw_params->clk_table.entries[i].dtbclk_mhz;
+			if (!bw_params->clk_table.entries[i].socclk_mhz && i > 0)
+				dcn3_02_soc.clock_limits[i].socclk_mhz = dcn3_02_soc.clock_limits[i-1].socclk_mhz;
+			else
+				dcn3_02_soc.clock_limits[i].socclk_mhz = bw_params->clk_table.entries[i].socclk_mhz;
+			/* These clocks cannot come from bw_params, always fill from dcn3_02_soc[1] */
+			/* FCLK, PHYCLK_D18, DSCCLK */
+			dcn3_02_soc.clock_limits[i].phyclk_d18_mhz = dcn3_02_soc.clock_limits[0].phyclk_d18_mhz;
+			dcn3_02_soc.clock_limits[i].dscclk_mhz = dcn3_02_soc.clock_limits[0].dscclk_mhz;
+		}
+		/* re-init DML with updated bb */
+		dml_init_instance(&dc->dml, &dcn3_02_soc, &dcn3_02_ip, DML_PROJECT_DCN30);
+		if (dc->current_state)
+			dml_init_instance(&dc->current_state->bw_ctx.dml, &dcn3_02_soc, &dcn3_02_ip, DML_PROJECT_DCN30);
+	}
+}
+
+void dcn302_fpu_init_soc_bounding_box(struct bp_soc_bb_info bb_info)
+{
+
+			dc_assert_fp_enabled();
+
+			if (bb_info.dram_clock_change_latency_100ns > 0)
+				dcn3_02_soc.dram_clock_change_latency_us =
+					bb_info.dram_clock_change_latency_100ns * 10;
+
+			if (bb_info.dram_sr_enter_exit_latency_100ns > 0)
+				dcn3_02_soc.sr_enter_plus_exit_time_us =
+					bb_info.dram_sr_enter_exit_latency_100ns * 10;
+
+			if (bb_info.dram_sr_exit_latency_100ns > 0)
+				dcn3_02_soc.sr_exit_time_us =
+					bb_info.dram_sr_exit_latency_100ns * 10;
+}
+
+
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn302/dcn302_fpu.h b/drivers/gpu/drm/amd/display/dc/dml/dcn302/dcn302_fpu.h
new file mode 100644
index 000000000000..548305d96cee
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn302/dcn302_fpu.h
@@ -0,0 +1,32 @@
+/*
+ * Copyright 2019-2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __DCN302_FPU_H__
+#define __DCN302_FPU_H__
+
+void dcn302_fpu_init_soc_bounding_box(struct bp_soc_bb_info bb_info);
+void dcn302_fpu_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params);
+
+#endif /* __DCN302_FPU_H__*/
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn303/dcn303_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn303/dcn303_fpu.c
new file mode 100644
index 000000000000..b5d3fd4c3694
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn303/dcn303_fpu.c
@@ -0,0 +1,380 @@
+/*
+ * Copyright 2019-2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+#include "resource.h"
+#include "clk_mgr.h"
+#include "dcn20/dcn20_resource.h"
+#include "dcn303/dcn303_resource.h"
+
+#include "dml/dcn20/dcn20_fpu.h"
+#include "dcn303_fpu.h"
+
+struct _vcs_dpi_ip_params_st dcn3_03_ip = {
+		.use_min_dcfclk = 0,
+		.clamp_min_dcfclk = 0,
+		.odm_capable = 1,
+		.gpuvm_enable = 1,
+		.hostvm_enable = 0,
+		.gpuvm_max_page_table_levels = 4,
+		.hostvm_max_page_table_levels = 4,
+		.hostvm_cached_page_table_levels = 0,
+		.pte_group_size_bytes = 2048,
+		.num_dsc = 2,
+		.rob_buffer_size_kbytes = 184,
+		.det_buffer_size_kbytes = 184,
+		.dpte_buffer_size_in_pte_reqs_luma = 64,
+		.dpte_buffer_size_in_pte_reqs_chroma = 34,
+		.pde_proc_buffer_size_64k_reqs = 48,
+		.dpp_output_buffer_pixels = 2560,
+		.opp_output_buffer_lines = 1,
+		.pixel_chunk_size_kbytes = 8,
+		.pte_enable = 1,
+		.max_page_table_levels = 2,
+		.pte_chunk_size_kbytes = 2,  // ?
+		.meta_chunk_size_kbytes = 2,
+		.writeback_chunk_size_kbytes = 8,
+		.line_buffer_size_bits = 789504,
+		.is_line_buffer_bpp_fixed = 0,  // ?
+		.line_buffer_fixed_bpp = 0,     // ?
+		.dcc_supported = true,
+		.writeback_interface_buffer_size_kbytes = 90,
+		.writeback_line_buffer_buffer_size = 0,
+		.max_line_buffer_lines = 12,
+		.writeback_luma_buffer_size_kbytes = 12,  // writeback_line_buffer_buffer_size = 656640
+		.writeback_chroma_buffer_size_kbytes = 8,
+		.writeback_chroma_line_buffer_width_pixels = 4,
+		.writeback_max_hscl_ratio = 1,
+		.writeback_max_vscl_ratio = 1,
+		.writeback_min_hscl_ratio = 1,
+		.writeback_min_vscl_ratio = 1,
+		.writeback_max_hscl_taps = 1,
+		.writeback_max_vscl_taps = 1,
+		.writeback_line_buffer_luma_buffer_size = 0,
+		.writeback_line_buffer_chroma_buffer_size = 14643,
+		.cursor_buffer_size = 8,
+		.cursor_chunk_size = 2,
+		.max_num_otg = 2,
+		.max_num_dpp = 2,
+		.max_num_wb = 1,
+		.max_dchub_pscl_bw_pix_per_clk = 4,
+		.max_pscl_lb_bw_pix_per_clk = 2,
+		.max_lb_vscl_bw_pix_per_clk = 4,
+		.max_vscl_hscl_bw_pix_per_clk = 4,
+		.max_hscl_ratio = 6,
+		.max_vscl_ratio = 6,
+		.hscl_mults = 4,
+		.vscl_mults = 4,
+		.max_hscl_taps = 8,
+		.max_vscl_taps = 8,
+		.dispclk_ramp_margin_percent = 1,
+		.underscan_factor = 1.11,
+		.min_vblank_lines = 32,
+		.dppclk_delay_subtotal = 46,
+		.dynamic_metadata_vm_enabled = true,
+		.dppclk_delay_scl_lb_only = 16,
+		.dppclk_delay_scl = 50,
+		.dppclk_delay_cnvc_formatter = 27,
+		.dppclk_delay_cnvc_cursor = 6,
+		.dispclk_delay_subtotal = 119,
+		.dcfclk_cstate_latency = 5.2, // SRExitTime
+		.max_inter_dcn_tile_repeaters = 8,
+		.max_num_hdmi_frl_outputs = 1,
+		.odm_combine_4to1_supported = false,
+
+		.xfc_supported = false,
+		.xfc_fill_bw_overhead_percent = 10.0,
+		.xfc_fill_constant_bytes = 0,
+		.gfx7_compat_tiling_supported = 0,
+		.number_of_cursors = 1,
+};
+
+struct _vcs_dpi_soc_bounding_box_st dcn3_03_soc = {
+		.clock_limits = {
+				{
+						.state = 0,
+						.dispclk_mhz = 562.0,
+						.dppclk_mhz = 300.0,
+						.phyclk_mhz = 300.0,
+						.phyclk_d18_mhz = 667.0,
+						.dscclk_mhz = 405.6,
+				},
+		},
+
+		.min_dcfclk = 500.0, /* TODO: set this to actual min DCFCLK */
+		.num_states = 1,
+		.sr_exit_time_us = 35.5,
+		.sr_enter_plus_exit_time_us = 40,
+		.urgent_latency_us = 4.0,
+		.urgent_latency_pixel_data_only_us = 4.0,
+		.urgent_latency_pixel_mixed_with_vm_data_us = 4.0,
+		.urgent_latency_vm_data_only_us = 4.0,
+		.urgent_out_of_order_return_per_channel_pixel_only_bytes = 4096,
+		.urgent_out_of_order_return_per_channel_pixel_and_vm_bytes = 4096,
+		.urgent_out_of_order_return_per_channel_vm_only_bytes = 4096,
+		.pct_ideal_dram_sdp_bw_after_urgent_pixel_only = 80.0,
+		.pct_ideal_dram_sdp_bw_after_urgent_pixel_and_vm = 60.0,
+		.pct_ideal_dram_sdp_bw_after_urgent_vm_only = 40.0,
+		.max_avg_sdp_bw_use_normal_percent = 60.0,
+		.max_avg_dram_bw_use_normal_percent = 40.0,
+		.writeback_latency_us = 12.0,
+		.max_request_size_bytes = 256,
+		.fabric_datapath_to_dcn_data_return_bytes = 64,
+		.dcn_downspread_percent = 0.5,
+		.downspread_percent = 0.38,
+		.dram_page_open_time_ns = 50.0,
+		.dram_rw_turnaround_time_ns = 17.5,
+		.dram_return_buffer_per_channel_bytes = 8192,
+		.round_trip_ping_latency_dcfclk_cycles = 156,
+		.urgent_out_of_order_return_per_channel_bytes = 4096,
+		.channel_interleave_bytes = 256,
+		.num_banks = 8,
+		.gpuvm_min_page_size_bytes = 4096,
+		.hostvm_min_page_size_bytes = 4096,
+		.dram_clock_change_latency_us = 404,
+		.dummy_pstate_latency_us = 5,
+		.writeback_dram_clock_change_latency_us = 23.0,
+		.return_bus_width_bytes = 64,
+		.dispclk_dppclk_vco_speed_mhz = 3650,
+		.xfc_bus_transport_time_us = 20,      // ?
+		.xfc_xbuf_latency_tolerance_us = 4,  // ?
+		.use_urgent_burst_bw = 1,            // ?
+		.do_urgent_latency_adjustment = true,
+		.urgent_latency_adjustment_fabric_clock_component_us = 1.0,
+		.urgent_latency_adjustment_fabric_clock_reference_mhz = 1000,
+};
+
+static void dcn303_get_optimal_dcfclk_fclk_for_uclk(unsigned int uclk_mts,
+		unsigned int *optimal_dcfclk,
+		unsigned int *optimal_fclk)
+{
+	double bw_from_dram, bw_from_dram1, bw_from_dram2;
+
+	bw_from_dram1 = uclk_mts * dcn3_03_soc.num_chans *
+		dcn3_03_soc.dram_channel_width_bytes * (dcn3_03_soc.max_avg_dram_bw_use_normal_percent / 100);
+	bw_from_dram2 = uclk_mts * dcn3_03_soc.num_chans *
+		dcn3_03_soc.dram_channel_width_bytes * (dcn3_03_soc.max_avg_sdp_bw_use_normal_percent / 100);
+
+	bw_from_dram = (bw_from_dram1 < bw_from_dram2) ? bw_from_dram1 : bw_from_dram2;
+
+	if (optimal_fclk)
+		*optimal_fclk = bw_from_dram /
+		(dcn3_03_soc.fabric_datapath_to_dcn_data_return_bytes *
+				(dcn3_03_soc.max_avg_sdp_bw_use_normal_percent / 100));
+
+	if (optimal_dcfclk)
+		*optimal_dcfclk =  bw_from_dram /
+		(dcn3_03_soc.return_bus_width_bytes * (dcn3_03_soc.max_avg_sdp_bw_use_normal_percent / 100));
+}
+
+
+void dcn303_fpu_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params)
+{
+	unsigned int i, j;
+	unsigned int num_states = 0;
+
+	unsigned int dcfclk_mhz[DC__VOLTAGE_STATES] = {0};
+	unsigned int dram_speed_mts[DC__VOLTAGE_STATES] = {0};
+	unsigned int optimal_uclk_for_dcfclk_sta_targets[DC__VOLTAGE_STATES] = {0};
+	unsigned int optimal_dcfclk_for_uclk[DC__VOLTAGE_STATES] = {0};
+
+	unsigned int dcfclk_sta_targets[DC__VOLTAGE_STATES] = {694, 875, 1000, 1200};
+	unsigned int num_dcfclk_sta_targets = 4;
+	unsigned int num_uclk_states;
+
+	dc_assert_fp_enabled();
+
+	if (dc->ctx->dc_bios->vram_info.num_chans)
+		dcn3_03_soc.num_chans = dc->ctx->dc_bios->vram_info.num_chans;
+
+	if (dc->ctx->dc_bios->vram_info.dram_channel_width_bytes)
+		dcn3_03_soc.dram_channel_width_bytes = dc->ctx->dc_bios->vram_info.dram_channel_width_bytes;
+
+	dcn3_03_soc.dispclk_dppclk_vco_speed_mhz = dc->clk_mgr->dentist_vco_freq_khz / 1000.0;
+	dc->dml.soc.dispclk_dppclk_vco_speed_mhz = dc->clk_mgr->dentist_vco_freq_khz / 1000.0;
+
+	if (bw_params->clk_table.entries[0].memclk_mhz) {
+		int max_dcfclk_mhz = 0, max_dispclk_mhz = 0, max_dppclk_mhz = 0, max_phyclk_mhz = 0;
+
+		for (i = 0; i < MAX_NUM_DPM_LVL; i++) {
+			if (bw_params->clk_table.entries[i].dcfclk_mhz > max_dcfclk_mhz)
+				max_dcfclk_mhz = bw_params->clk_table.entries[i].dcfclk_mhz;
+			if (bw_params->clk_table.entries[i].dispclk_mhz > max_dispclk_mhz)
+				max_dispclk_mhz = bw_params->clk_table.entries[i].dispclk_mhz;
+			if (bw_params->clk_table.entries[i].dppclk_mhz > max_dppclk_mhz)
+				max_dppclk_mhz = bw_params->clk_table.entries[i].dppclk_mhz;
+			if (bw_params->clk_table.entries[i].phyclk_mhz > max_phyclk_mhz)
+				max_phyclk_mhz = bw_params->clk_table.entries[i].phyclk_mhz;
+		}
+		if (!max_dcfclk_mhz)
+			max_dcfclk_mhz = dcn3_03_soc.clock_limits[0].dcfclk_mhz;
+		if (!max_dispclk_mhz)
+			max_dispclk_mhz = dcn3_03_soc.clock_limits[0].dispclk_mhz;
+		if (!max_dppclk_mhz)
+			max_dppclk_mhz = dcn3_03_soc.clock_limits[0].dppclk_mhz;
+		if (!max_phyclk_mhz)
+			max_phyclk_mhz = dcn3_03_soc.clock_limits[0].phyclk_mhz;
+
+		if (max_dcfclk_mhz > dcfclk_sta_targets[num_dcfclk_sta_targets-1]) {
+			dcfclk_sta_targets[num_dcfclk_sta_targets] = max_dcfclk_mhz;
+			num_dcfclk_sta_targets++;
+		} else if (max_dcfclk_mhz < dcfclk_sta_targets[num_dcfclk_sta_targets-1]) {
+			for (i = 0; i < num_dcfclk_sta_targets; i++) {
+				if (dcfclk_sta_targets[i] > max_dcfclk_mhz) {
+					dcfclk_sta_targets[i] = max_dcfclk_mhz;
+					break;
+				}
+			}
+			/* Update size of array since we "removed" duplicates */
+			num_dcfclk_sta_targets = i + 1;
+		}
+
+		num_uclk_states = bw_params->clk_table.num_entries;
+
+		/* Calculate optimal dcfclk for each uclk */
+		for (i = 0; i < num_uclk_states; i++) {
+			dcn303_get_optimal_dcfclk_fclk_for_uclk(bw_params->clk_table.entries[i].memclk_mhz * 16,
+					&optimal_dcfclk_for_uclk[i], NULL);
+			if (optimal_dcfclk_for_uclk[i] < bw_params->clk_table.entries[0].dcfclk_mhz)
+				optimal_dcfclk_for_uclk[i] = bw_params->clk_table.entries[0].dcfclk_mhz;
+		}
+
+		/* Calculate optimal uclk for each dcfclk sta target */
+		for (i = 0; i < num_dcfclk_sta_targets; i++) {
+			for (j = 0; j < num_uclk_states; j++) {
+				if (dcfclk_sta_targets[i] < optimal_dcfclk_for_uclk[j]) {
+					optimal_uclk_for_dcfclk_sta_targets[i] =
+							bw_params->clk_table.entries[j].memclk_mhz * 16;
+					break;
+				} else {
+					/* condition where (dcfclk_sta_targets[i] >= optimal_dcfclk_for_uclk[j]):
+					 * This is required for dcn303 because it just so happens that the memory
+					 * bandwidth is low enough such that all the optimal DCFCLK for each UCLK
+					 * is lower than the smallest DCFCLK STA target. In this case we need to
+					 * populate the optimal UCLK for each DCFCLK STA target to be the max UCLK.
+					 */
+					if (j == num_uclk_states - 1) {
+						optimal_uclk_for_dcfclk_sta_targets[i] =
+								bw_params->clk_table.entries[j].memclk_mhz * 16;
+					}
+				}
+			}
+		}
+
+		i = 0;
+		j = 0;
+		/* create the final dcfclk and uclk table */
+		while (i < num_dcfclk_sta_targets && j < num_uclk_states && num_states < DC__VOLTAGE_STATES) {
+			if (dcfclk_sta_targets[i] < optimal_dcfclk_for_uclk[j]) {
+				dcfclk_mhz[num_states] = dcfclk_sta_targets[i];
+				dram_speed_mts[num_states++] = optimal_uclk_for_dcfclk_sta_targets[i++];
+			} else {
+				if (j < num_uclk_states && optimal_dcfclk_for_uclk[j] <= max_dcfclk_mhz) {
+					dcfclk_mhz[num_states] = optimal_dcfclk_for_uclk[j];
+					dram_speed_mts[num_states++] =
+							bw_params->clk_table.entries[j++].memclk_mhz * 16;
+				} else {
+					j = num_uclk_states;
+				}
+			}
+		}
+
+		while (i < num_dcfclk_sta_targets && num_states < DC__VOLTAGE_STATES) {
+			dcfclk_mhz[num_states] = dcfclk_sta_targets[i];
+			dram_speed_mts[num_states++] = optimal_uclk_for_dcfclk_sta_targets[i++];
+		}
+
+		while (j < num_uclk_states && num_states < DC__VOLTAGE_STATES &&
+				optimal_dcfclk_for_uclk[j] <= max_dcfclk_mhz) {
+			dcfclk_mhz[num_states] = optimal_dcfclk_for_uclk[j];
+			dram_speed_mts[num_states++] = bw_params->clk_table.entries[j++].memclk_mhz * 16;
+		}
+
+		/* bw_params->clk_table.entries[MAX_NUM_DPM_LVL].
+		 * MAX_NUM_DPM_LVL is 8.
+		 * dcn3_02_soc.clock_limits[DC__VOLTAGE_STATES].
+		 * DC__VOLTAGE_STATES is 40.
+		 */
+		if (num_states > MAX_NUM_DPM_LVL) {
+			ASSERT(0);
+			return;
+		}
+
+		dcn3_03_soc.num_states = num_states;
+		for (i = 0; i < dcn3_03_soc.num_states; i++) {
+			dcn3_03_soc.clock_limits[i].state = i;
+			dcn3_03_soc.clock_limits[i].dcfclk_mhz = dcfclk_mhz[i];
+			dcn3_03_soc.clock_limits[i].fabricclk_mhz = dcfclk_mhz[i];
+			dcn3_03_soc.clock_limits[i].dram_speed_mts = dram_speed_mts[i];
+
+			/* Fill all states with max values of all other clocks */
+			dcn3_03_soc.clock_limits[i].dispclk_mhz = max_dispclk_mhz;
+			dcn3_03_soc.clock_limits[i].dppclk_mhz  = max_dppclk_mhz;
+			dcn3_03_soc.clock_limits[i].phyclk_mhz  = max_phyclk_mhz;
+			/* Populate from bw_params for DTBCLK, SOCCLK */
+			if (!bw_params->clk_table.entries[i].dtbclk_mhz && i > 0)
+				dcn3_03_soc.clock_limits[i].dtbclk_mhz = dcn3_03_soc.clock_limits[i-1].dtbclk_mhz;
+			else
+				dcn3_03_soc.clock_limits[i].dtbclk_mhz = bw_params->clk_table.entries[i].dtbclk_mhz;
+			if (!bw_params->clk_table.entries[i].socclk_mhz && i > 0)
+				dcn3_03_soc.clock_limits[i].socclk_mhz = dcn3_03_soc.clock_limits[i-1].socclk_mhz;
+			else
+				dcn3_03_soc.clock_limits[i].socclk_mhz = bw_params->clk_table.entries[i].socclk_mhz;
+			/* These clocks cannot come from bw_params, always fill from dcn3_03_soc[1] */
+			/* FCLK, PHYCLK_D18, DSCCLK */
+			dcn3_03_soc.clock_limits[i].phyclk_d18_mhz = dcn3_03_soc.clock_limits[0].phyclk_d18_mhz;
+			dcn3_03_soc.clock_limits[i].dscclk_mhz = dcn3_03_soc.clock_limits[0].dscclk_mhz;
+		}
+
+		if (dcn3_03_soc.num_chans <= 4) {
+			for (i = 0; i < dcn3_03_soc.num_states; i++) {
+				if (dcn3_03_soc.clock_limits[i].dram_speed_mts > 1700)
+					break;
+
+				if (dcn3_03_soc.clock_limits[i].dram_speed_mts >= 1500) {
+					dcn3_03_soc.clock_limits[i].dcfclk_mhz = 100;
+					dcn3_03_soc.clock_limits[i].fabricclk_mhz = 100;
+				}
+			}
+		}
+
+		/* re-init DML with updated bb */
+		dml_init_instance(&dc->dml, &dcn3_03_soc, &dcn3_03_ip, DML_PROJECT_DCN30);
+		if (dc->current_state)
+			dml_init_instance(&dc->current_state->bw_ctx.dml, &dcn3_03_soc, &dcn3_03_ip, DML_PROJECT_DCN30);
+	}
+}
+
+void dcn303_fpu_init_soc_bounding_box(struct bp_soc_bb_info bb_info)
+{
+	dc_assert_fp_enabled();
+
+	if (bb_info.dram_clock_change_latency_100ns > 0)
+		dcn3_03_soc.dram_clock_change_latency_us = bb_info.dram_clock_change_latency_100ns * 10;
+
+	if (bb_info.dram_sr_enter_exit_latency_100ns > 0)
+		dcn3_03_soc.sr_enter_plus_exit_time_us = bb_info.dram_sr_enter_exit_latency_100ns * 10;
+
+	if (bb_info.dram_sr_exit_latency_100ns > 0)
+		dcn3_03_soc.sr_exit_time_us = bb_info.dram_sr_exit_latency_100ns * 10;
+}
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn303/dcn303_fpu.h b/drivers/gpu/drm/amd/display/dc/dml/dcn303/dcn303_fpu.h
new file mode 100644
index 000000000000..92ec833fa528
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn303/dcn303_fpu.h
@@ -0,0 +1,32 @@
+/*
+ * Copyright 2019-2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __DCN303_FPU_H__
+#define __DCN303_FPU_H__
+
+void dcn303_fpu_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params);
+void dcn303_fpu_init_soc_bounding_box(struct bp_soc_bb_info bb_info);
+
+#endif /* __DCN303_FPU_H__*/
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn31/dcn31_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn31/dcn31_fpu.c
new file mode 100644
index 000000000000..1a28061bb9ff
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn31/dcn31_fpu.c
@@ -0,0 +1,827 @@
+/*
+ * Copyright 2019-2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#include "resource.h"
+#include "clk_mgr.h"
+#include "dcn31/dcn31_resource.h"
+#include "dcn315/dcn315_resource.h"
+#include "dcn316/dcn316_resource.h"
+
+#include "dml/dcn20/dcn20_fpu.h"
+#include "dcn31_fpu.h"
+
+/**
+ * DOC: DCN31x FPU manipulation Overview
+ *
+ * The DCN architecture relies on FPU operations, which require special
+ * compilation flags and the use of kernel_fpu_begin/end functions; ideally, we
+ * want to avoid spreading FPU access across multiple files. With this idea in
+ * mind, this file aims to centralize all DCN3.1.x functions that require FPU
+ * access in a single place. Code in this file follows the following code
+ * pattern:
+ *
+ * 1. Functions that use FPU operations should be isolated in static functions.
+ * 2. The FPU functions should have the noinline attribute to ensure anything
+ *    that deals with FP register is contained within this call.
+ * 3. All function that needs to be accessed outside this file requires a
+ *    public interface that not uses any FPU reference.
+ * 4. Developers **must not** use DC_FP_START/END in this file, but they need
+ *    to ensure that the caller invokes it before access any function available
+ *    in this file. For this reason, public functions in this file must invoke
+ *    dc_assert_fp_enabled();
+ */
+
+struct _vcs_dpi_ip_params_st dcn3_1_ip = {
+	.gpuvm_enable = 1,
+	.gpuvm_max_page_table_levels = 1,
+	.hostvm_enable = 1,
+	.hostvm_max_page_table_levels = 2,
+	.rob_buffer_size_kbytes = 64,
+	.det_buffer_size_kbytes = DCN3_1_DEFAULT_DET_SIZE,
+	.config_return_buffer_size_in_kbytes = 1792,
+	.compressed_buffer_segment_size_in_kbytes = 64,
+	.meta_fifo_size_in_kentries = 32,
+	.zero_size_buffer_entries = 512,
+	.compbuf_reserved_space_64b = 256,
+	.compbuf_reserved_space_zs = 64,
+	.dpp_output_buffer_pixels = 2560,
+	.opp_output_buffer_lines = 1,
+	.pixel_chunk_size_kbytes = 8,
+	.meta_chunk_size_kbytes = 2,
+	.min_meta_chunk_size_bytes = 256,
+	.writeback_chunk_size_kbytes = 8,
+	.ptoi_supported = false,
+	.num_dsc = 3,
+	.maximum_dsc_bits_per_component = 10,
+	.dsc422_native_support = false,
+	.is_line_buffer_bpp_fixed = true,
+	.line_buffer_fixed_bpp = 48,
+	.line_buffer_size_bits = 789504,
+	.max_line_buffer_lines = 12,
+	.writeback_interface_buffer_size_kbytes = 90,
+	.max_num_dpp = 4,
+	.max_num_otg = 4,
+	.max_num_hdmi_frl_outputs = 1,
+	.max_num_wb = 1,
+	.max_dchub_pscl_bw_pix_per_clk = 4,
+	.max_pscl_lb_bw_pix_per_clk = 2,
+	.max_lb_vscl_bw_pix_per_clk = 4,
+	.max_vscl_hscl_bw_pix_per_clk = 4,
+	.max_hscl_ratio = 6,
+	.max_vscl_ratio = 6,
+	.max_hscl_taps = 8,
+	.max_vscl_taps = 8,
+	.dpte_buffer_size_in_pte_reqs_luma = 64,
+	.dpte_buffer_size_in_pte_reqs_chroma = 34,
+	.dispclk_ramp_margin_percent = 1,
+	.max_inter_dcn_tile_repeaters = 8,
+	.cursor_buffer_size = 16,
+	.cursor_chunk_size = 2,
+	.writeback_line_buffer_buffer_size = 0,
+	.writeback_min_hscl_ratio = 1,
+	.writeback_min_vscl_ratio = 1,
+	.writeback_max_hscl_ratio = 1,
+	.writeback_max_vscl_ratio = 1,
+	.writeback_max_hscl_taps = 1,
+	.writeback_max_vscl_taps = 1,
+	.dppclk_delay_subtotal = 46,
+	.dppclk_delay_scl = 50,
+	.dppclk_delay_scl_lb_only = 16,
+	.dppclk_delay_cnvc_formatter = 27,
+	.dppclk_delay_cnvc_cursor = 6,
+	.dispclk_delay_subtotal = 119,
+	.dynamic_metadata_vm_enabled = false,
+	.odm_combine_4to1_supported = false,
+	.dcc_supported = true,
+};
+
+static struct _vcs_dpi_soc_bounding_box_st dcn3_1_soc = {
+		/*TODO: correct dispclk/dppclk voltage level determination*/
+	.clock_limits = {
+		{
+			.state = 0,
+			.dispclk_mhz = 1200.0,
+			.dppclk_mhz = 1200.0,
+			.phyclk_mhz = 600.0,
+			.phyclk_d18_mhz = 667.0,
+			.dscclk_mhz = 186.0,
+			.dtbclk_mhz = 625.0,
+		},
+		{
+			.state = 1,
+			.dispclk_mhz = 1200.0,
+			.dppclk_mhz = 1200.0,
+			.phyclk_mhz = 810.0,
+			.phyclk_d18_mhz = 667.0,
+			.dscclk_mhz = 209.0,
+			.dtbclk_mhz = 625.0,
+		},
+		{
+			.state = 2,
+			.dispclk_mhz = 1200.0,
+			.dppclk_mhz = 1200.0,
+			.phyclk_mhz = 810.0,
+			.phyclk_d18_mhz = 667.0,
+			.dscclk_mhz = 209.0,
+			.dtbclk_mhz = 625.0,
+		},
+		{
+			.state = 3,
+			.dispclk_mhz = 1200.0,
+			.dppclk_mhz = 1200.0,
+			.phyclk_mhz = 810.0,
+			.phyclk_d18_mhz = 667.0,
+			.dscclk_mhz = 371.0,
+			.dtbclk_mhz = 625.0,
+		},
+		{
+			.state = 4,
+			.dispclk_mhz = 1200.0,
+			.dppclk_mhz = 1200.0,
+			.phyclk_mhz = 810.0,
+			.phyclk_d18_mhz = 667.0,
+			.dscclk_mhz = 417.0,
+			.dtbclk_mhz = 625.0,
+		},
+	},
+	.num_states = 5,
+	.sr_exit_time_us = 9.0,
+	.sr_enter_plus_exit_time_us = 11.0,
+	.sr_exit_z8_time_us = 442.0,
+	.sr_enter_plus_exit_z8_time_us = 560.0,
+	.writeback_latency_us = 12.0,
+	.dram_channel_width_bytes = 4,
+	.round_trip_ping_latency_dcfclk_cycles = 106,
+	.urgent_latency_pixel_data_only_us = 4.0,
+	.urgent_latency_pixel_mixed_with_vm_data_us = 4.0,
+	.urgent_latency_vm_data_only_us = 4.0,
+	.urgent_out_of_order_return_per_channel_pixel_only_bytes = 4096,
+	.urgent_out_of_order_return_per_channel_pixel_and_vm_bytes = 4096,
+	.urgent_out_of_order_return_per_channel_vm_only_bytes = 4096,
+	.pct_ideal_sdp_bw_after_urgent = 80.0,
+	.pct_ideal_dram_sdp_bw_after_urgent_pixel_only = 65.0,
+	.pct_ideal_dram_sdp_bw_after_urgent_pixel_and_vm = 60.0,
+	.pct_ideal_dram_sdp_bw_after_urgent_vm_only = 30.0,
+	.max_avg_sdp_bw_use_normal_percent = 60.0,
+	.max_avg_dram_bw_use_normal_percent = 60.0,
+	.fabric_datapath_to_dcn_data_return_bytes = 32,
+	.return_bus_width_bytes = 64,
+	.downspread_percent = 0.38,
+	.dcn_downspread_percent = 0.5,
+	.gpuvm_min_page_size_bytes = 4096,
+	.hostvm_min_page_size_bytes = 4096,
+	.do_urgent_latency_adjustment = false,
+	.urgent_latency_adjustment_fabric_clock_component_us = 0,
+	.urgent_latency_adjustment_fabric_clock_reference_mhz = 0,
+};
+
+struct _vcs_dpi_ip_params_st dcn3_15_ip = {
+	.gpuvm_enable = 1,
+	.gpuvm_max_page_table_levels = 1,
+	.hostvm_enable = 1,
+	.hostvm_max_page_table_levels = 2,
+	.rob_buffer_size_kbytes = 64,
+	.det_buffer_size_kbytes = DCN3_15_DEFAULT_DET_SIZE,
+	.min_comp_buffer_size_kbytes = 64,
+	.config_return_buffer_size_in_kbytes = 1024,
+	.compressed_buffer_segment_size_in_kbytes = 64,
+	.meta_fifo_size_in_kentries = 32,
+	.zero_size_buffer_entries = 512,
+	.compbuf_reserved_space_64b = 256,
+	.compbuf_reserved_space_zs = 64,
+	.dpp_output_buffer_pixels = 2560,
+	.opp_output_buffer_lines = 1,
+	.pixel_chunk_size_kbytes = 8,
+	.meta_chunk_size_kbytes = 2,
+	.min_meta_chunk_size_bytes = 256,
+	.writeback_chunk_size_kbytes = 8,
+	.ptoi_supported = false,
+	.num_dsc = 3,
+	.maximum_dsc_bits_per_component = 10,
+	.dsc422_native_support = false,
+	.is_line_buffer_bpp_fixed = true,
+	.line_buffer_fixed_bpp = 48,
+	.line_buffer_size_bits = 789504,
+	.max_line_buffer_lines = 12,
+	.writeback_interface_buffer_size_kbytes = 90,
+	.max_num_dpp = 4,
+	.max_num_otg = 4,
+	.max_num_hdmi_frl_outputs = 1,
+	.max_num_wb = 1,
+	.max_dchub_pscl_bw_pix_per_clk = 4,
+	.max_pscl_lb_bw_pix_per_clk = 2,
+	.max_lb_vscl_bw_pix_per_clk = 4,
+	.max_vscl_hscl_bw_pix_per_clk = 4,
+	.max_hscl_ratio = 6,
+	.max_vscl_ratio = 6,
+	.max_hscl_taps = 8,
+	.max_vscl_taps = 8,
+	.dpte_buffer_size_in_pte_reqs_luma = 64,
+	.dpte_buffer_size_in_pte_reqs_chroma = 34,
+	.dispclk_ramp_margin_percent = 1,
+	.max_inter_dcn_tile_repeaters = 9,
+	.cursor_buffer_size = 16,
+	.cursor_chunk_size = 2,
+	.writeback_line_buffer_buffer_size = 0,
+	.writeback_min_hscl_ratio = 1,
+	.writeback_min_vscl_ratio = 1,
+	.writeback_max_hscl_ratio = 1,
+	.writeback_max_vscl_ratio = 1,
+	.writeback_max_hscl_taps = 1,
+	.writeback_max_vscl_taps = 1,
+	.dppclk_delay_subtotal = 46,
+	.dppclk_delay_scl = 50,
+	.dppclk_delay_scl_lb_only = 16,
+	.dppclk_delay_cnvc_formatter = 27,
+	.dppclk_delay_cnvc_cursor = 6,
+	.dispclk_delay_subtotal = 119,
+	.dynamic_metadata_vm_enabled = false,
+	.odm_combine_4to1_supported = false,
+	.dcc_supported = true,
+};
+
+static struct _vcs_dpi_soc_bounding_box_st dcn3_15_soc = {
+	.sr_exit_time_us = 9.0,
+	.sr_enter_plus_exit_time_us = 11.0,
+	.sr_exit_z8_time_us = 50.0,
+	.sr_enter_plus_exit_z8_time_us = 50.0,
+	.writeback_latency_us = 12.0,
+	.dram_channel_width_bytes = 4,
+	.round_trip_ping_latency_dcfclk_cycles = 106,
+	.urgent_latency_pixel_data_only_us = 4.0,
+	.urgent_latency_pixel_mixed_with_vm_data_us = 4.0,
+	.urgent_latency_vm_data_only_us = 4.0,
+	.urgent_out_of_order_return_per_channel_pixel_only_bytes = 4096,
+	.urgent_out_of_order_return_per_channel_pixel_and_vm_bytes = 4096,
+	.urgent_out_of_order_return_per_channel_vm_only_bytes = 4096,
+	.pct_ideal_sdp_bw_after_urgent = 80.0,
+	.pct_ideal_dram_sdp_bw_after_urgent_pixel_only = 65.0,
+	.pct_ideal_dram_sdp_bw_after_urgent_pixel_and_vm = 60.0,
+	.pct_ideal_dram_sdp_bw_after_urgent_vm_only = 30.0,
+	.max_avg_sdp_bw_use_normal_percent = 60.0,
+	.max_avg_dram_bw_use_normal_percent = 60.0,
+	.fabric_datapath_to_dcn_data_return_bytes = 32,
+	.return_bus_width_bytes = 64,
+	.downspread_percent = 0.38,
+	.dcn_downspread_percent = 0.38,
+	.gpuvm_min_page_size_bytes = 4096,
+	.hostvm_min_page_size_bytes = 4096,
+	.do_urgent_latency_adjustment = false,
+	.urgent_latency_adjustment_fabric_clock_component_us = 0,
+	.urgent_latency_adjustment_fabric_clock_reference_mhz = 0,
+	.dispclk_dppclk_vco_speed_mhz = 2400.0,
+	.num_chans = 4,
+	.dummy_pstate_latency_us = 10.0
+};
+
+struct _vcs_dpi_ip_params_st dcn3_16_ip = {
+	.gpuvm_enable = 1,
+	.gpuvm_max_page_table_levels = 1,
+	.hostvm_enable = 1,
+	.hostvm_max_page_table_levels = 2,
+	.rob_buffer_size_kbytes = 64,
+	.det_buffer_size_kbytes = DCN3_16_DEFAULT_DET_SIZE,
+	.min_comp_buffer_size_kbytes = 64,
+	.config_return_buffer_size_in_kbytes = 1024,
+	.compressed_buffer_segment_size_in_kbytes = 64,
+	.meta_fifo_size_in_kentries = 32,
+	.zero_size_buffer_entries = 512,
+	.compbuf_reserved_space_64b = 256,
+	.compbuf_reserved_space_zs = 64,
+	.dpp_output_buffer_pixels = 2560,
+	.opp_output_buffer_lines = 1,
+	.pixel_chunk_size_kbytes = 8,
+	.meta_chunk_size_kbytes = 2,
+	.min_meta_chunk_size_bytes = 256,
+	.writeback_chunk_size_kbytes = 8,
+	.ptoi_supported = false,
+	.num_dsc = 3,
+	.maximum_dsc_bits_per_component = 10,
+	.dsc422_native_support = false,
+	.is_line_buffer_bpp_fixed = true,
+	.line_buffer_fixed_bpp = 48,
+	.line_buffer_size_bits = 789504,
+	.max_line_buffer_lines = 12,
+	.writeback_interface_buffer_size_kbytes = 90,
+	.max_num_dpp = 4,
+	.max_num_otg = 4,
+	.max_num_hdmi_frl_outputs = 1,
+	.max_num_wb = 1,
+	.max_dchub_pscl_bw_pix_per_clk = 4,
+	.max_pscl_lb_bw_pix_per_clk = 2,
+	.max_lb_vscl_bw_pix_per_clk = 4,
+	.max_vscl_hscl_bw_pix_per_clk = 4,
+	.max_hscl_ratio = 6,
+	.max_vscl_ratio = 6,
+	.max_hscl_taps = 8,
+	.max_vscl_taps = 8,
+	.dpte_buffer_size_in_pte_reqs_luma = 64,
+	.dpte_buffer_size_in_pte_reqs_chroma = 34,
+	.dispclk_ramp_margin_percent = 1,
+	.max_inter_dcn_tile_repeaters = 8,
+	.cursor_buffer_size = 16,
+	.cursor_chunk_size = 2,
+	.writeback_line_buffer_buffer_size = 0,
+	.writeback_min_hscl_ratio = 1,
+	.writeback_min_vscl_ratio = 1,
+	.writeback_max_hscl_ratio = 1,
+	.writeback_max_vscl_ratio = 1,
+	.writeback_max_hscl_taps = 1,
+	.writeback_max_vscl_taps = 1,
+	.dppclk_delay_subtotal = 46,
+	.dppclk_delay_scl = 50,
+	.dppclk_delay_scl_lb_only = 16,
+	.dppclk_delay_cnvc_formatter = 27,
+	.dppclk_delay_cnvc_cursor = 6,
+	.dispclk_delay_subtotal = 119,
+	.dynamic_metadata_vm_enabled = false,
+	.odm_combine_4to1_supported = false,
+	.dcc_supported = true,
+};
+
+static struct _vcs_dpi_soc_bounding_box_st dcn3_16_soc = {
+		/*TODO: correct dispclk/dppclk voltage level determination*/
+	.clock_limits = {
+		{
+			.state = 0,
+			.dispclk_mhz = 556.0,
+			.dppclk_mhz = 556.0,
+			.phyclk_mhz = 600.0,
+			.phyclk_d18_mhz = 445.0,
+			.dscclk_mhz = 186.0,
+			.dtbclk_mhz = 625.0,
+		},
+		{
+			.state = 1,
+			.dispclk_mhz = 625.0,
+			.dppclk_mhz = 625.0,
+			.phyclk_mhz = 810.0,
+			.phyclk_d18_mhz = 667.0,
+			.dscclk_mhz = 209.0,
+			.dtbclk_mhz = 625.0,
+		},
+		{
+			.state = 2,
+			.dispclk_mhz = 625.0,
+			.dppclk_mhz = 625.0,
+			.phyclk_mhz = 810.0,
+			.phyclk_d18_mhz = 667.0,
+			.dscclk_mhz = 209.0,
+			.dtbclk_mhz = 625.0,
+		},
+		{
+			.state = 3,
+			.dispclk_mhz = 1112.0,
+			.dppclk_mhz = 1112.0,
+			.phyclk_mhz = 810.0,
+			.phyclk_d18_mhz = 667.0,
+			.dscclk_mhz = 371.0,
+			.dtbclk_mhz = 625.0,
+		},
+		{
+			.state = 4,
+			.dispclk_mhz = 1250.0,
+			.dppclk_mhz = 1250.0,
+			.phyclk_mhz = 810.0,
+			.phyclk_d18_mhz = 667.0,
+			.dscclk_mhz = 417.0,
+			.dtbclk_mhz = 625.0,
+		},
+	},
+	.num_states = 5,
+	.sr_exit_time_us = 9.0,
+	.sr_enter_plus_exit_time_us = 11.0,
+	.sr_exit_z8_time_us = 442.0,
+	.sr_enter_plus_exit_z8_time_us = 560.0,
+	.writeback_latency_us = 12.0,
+	.dram_channel_width_bytes = 4,
+	.round_trip_ping_latency_dcfclk_cycles = 106,
+	.urgent_latency_pixel_data_only_us = 4.0,
+	.urgent_latency_pixel_mixed_with_vm_data_us = 4.0,
+	.urgent_latency_vm_data_only_us = 4.0,
+	.urgent_out_of_order_return_per_channel_pixel_only_bytes = 4096,
+	.urgent_out_of_order_return_per_channel_pixel_and_vm_bytes = 4096,
+	.urgent_out_of_order_return_per_channel_vm_only_bytes = 4096,
+	.pct_ideal_sdp_bw_after_urgent = 80.0,
+	.pct_ideal_dram_sdp_bw_after_urgent_pixel_only = 65.0,
+	.pct_ideal_dram_sdp_bw_after_urgent_pixel_and_vm = 60.0,
+	.pct_ideal_dram_sdp_bw_after_urgent_vm_only = 30.0,
+	.max_avg_sdp_bw_use_normal_percent = 60.0,
+	.max_avg_dram_bw_use_normal_percent = 60.0,
+	.fabric_datapath_to_dcn_data_return_bytes = 32,
+	.return_bus_width_bytes = 64,
+	.downspread_percent = 0.38,
+	.dcn_downspread_percent = 0.5,
+	.gpuvm_min_page_size_bytes = 4096,
+	.hostvm_min_page_size_bytes = 4096,
+	.do_urgent_latency_adjustment = false,
+	.urgent_latency_adjustment_fabric_clock_component_us = 0,
+	.urgent_latency_adjustment_fabric_clock_reference_mhz = 0,
+	.dispclk_dppclk_vco_speed_mhz = 2500.0,
+};
+
+void dcn31_zero_pipe_dcc_fraction(display_e2e_pipe_params_st *pipes,
+				  int pipe_cnt)
+{
+	dc_assert_fp_enabled();
+
+	pipes[pipe_cnt].pipe.src.dcc_fraction_of_zs_req_luma = 0;
+	pipes[pipe_cnt].pipe.src.dcc_fraction_of_zs_req_chroma = 0;
+}
+
+void dcn31_update_soc_for_wm_a(struct dc *dc, struct dc_state *context)
+{
+	dc_assert_fp_enabled();
+
+	if (dc->clk_mgr->bw_params->wm_table.entries[WM_A].valid) {
+		context->bw_ctx.dml.soc.dram_clock_change_latency_us = dc->clk_mgr->bw_params->wm_table.entries[WM_A].pstate_latency_us;
+		context->bw_ctx.dml.soc.sr_enter_plus_exit_time_us = dc->clk_mgr->bw_params->wm_table.entries[WM_A].sr_enter_plus_exit_time_us;
+		context->bw_ctx.dml.soc.sr_exit_time_us = dc->clk_mgr->bw_params->wm_table.entries[WM_A].sr_exit_time_us;
+	}
+}
+
+void dcn315_update_soc_for_wm_a(struct dc *dc, struct dc_state *context)
+{
+	dc_assert_fp_enabled();
+
+	if (dc->clk_mgr->bw_params->wm_table.entries[WM_A].valid) {
+		/* For 315 pstate change is only supported if possible in vactive */
+		if (context->bw_ctx.dml.vba.DRAMClockChangeSupport[context->bw_ctx.dml.vba.VoltageLevel][context->bw_ctx.dml.vba.maxMpcComb] != dm_dram_clock_change_vactive)
+			context->bw_ctx.dml.soc.dram_clock_change_latency_us = context->bw_ctx.dml.soc.dummy_pstate_latency_us;
+		else
+			context->bw_ctx.dml.soc.dram_clock_change_latency_us = dc->clk_mgr->bw_params->wm_table.entries[WM_A].pstate_latency_us;
+		context->bw_ctx.dml.soc.sr_enter_plus_exit_time_us =
+				dc->clk_mgr->bw_params->wm_table.entries[WM_A].sr_enter_plus_exit_time_us;
+		context->bw_ctx.dml.soc.sr_exit_time_us =
+				dc->clk_mgr->bw_params->wm_table.entries[WM_A].sr_exit_time_us;
+	}
+}
+
+void dcn31_calculate_wm_and_dlg_fp(
+		struct dc *dc, struct dc_state *context,
+		display_e2e_pipe_params_st *pipes,
+		int pipe_cnt,
+		int vlevel)
+{
+	int i, pipe_idx, total_det = 0, active_hubp_count = 0;
+	double dcfclk = context->bw_ctx.dml.vba.DCFCLKState[vlevel][context->bw_ctx.dml.vba.maxMpcComb];
+	uint32_t cstate_enter_plus_exit_z8_ns;
+
+	dc_assert_fp_enabled();
+
+	if (context->bw_ctx.dml.soc.min_dcfclk > dcfclk)
+		dcfclk = context->bw_ctx.dml.soc.min_dcfclk;
+
+	/* We don't recalculate clocks for 0 pipe configs, which can block
+	 * S0i3 as high clocks will block low power states
+	 * Override any clocks that can block S0i3 to min here
+	 */
+	if (pipe_cnt == 0) {
+		context->bw_ctx.bw.dcn.clk.dcfclk_khz = dcfclk; // always should be vlevel 0
+		return;
+	}
+
+	pipes[0].clks_cfg.voltage = vlevel;
+	pipes[0].clks_cfg.dcfclk_mhz = dcfclk;
+	pipes[0].clks_cfg.socclk_mhz = context->bw_ctx.dml.soc.clock_limits[vlevel].socclk_mhz;
+
+	cstate_enter_plus_exit_z8_ns =
+		get_wm_z8_stutter_enter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+
+	if (get_stutter_period(&context->bw_ctx.dml, pipes, pipe_cnt) < dc->debug.minimum_z8_residency_time &&
+			cstate_enter_plus_exit_z8_ns < dc->debug.minimum_z8_residency_time * 1000)
+		cstate_enter_plus_exit_z8_ns = dc->debug.minimum_z8_residency_time * 1000;
+
+	/* Set A:
+	 * All clocks min required
+	 *
+	 * Set A calculated last so that following calculations are based on Set A
+	 */
+	dc->res_pool->funcs->update_soc_for_wm_a(dc, context);
+	context->bw_ctx.bw.dcn.watermarks.a.urgent_ns = get_wm_urgent(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.a.cstate_pstate.cstate_enter_plus_exit_ns = get_wm_stutter_enter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.a.cstate_pstate.cstate_exit_ns = get_wm_stutter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.a.cstate_pstate.pstate_change_ns = get_wm_dram_clock_change(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.a.cstate_pstate.cstate_enter_plus_exit_z8_ns = cstate_enter_plus_exit_z8_ns;
+	context->bw_ctx.bw.dcn.watermarks.a.cstate_pstate.cstate_exit_z8_ns = get_wm_z8_stutter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.a.pte_meta_urgent_ns = get_wm_memory_trip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.a.frac_urg_bw_nom = get_fraction_of_urgent_bandwidth(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.a.frac_urg_bw_flip = get_fraction_of_urgent_bandwidth_imm_flip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.a.urgent_latency_ns = get_urgent_latency(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.b = context->bw_ctx.bw.dcn.watermarks.a;
+	context->bw_ctx.bw.dcn.watermarks.c = context->bw_ctx.bw.dcn.watermarks.a;
+	context->bw_ctx.bw.dcn.watermarks.d = context->bw_ctx.bw.dcn.watermarks.a;
+
+	for (i = 0, pipe_idx = 0; i < dc->res_pool->pipe_count; i++) {
+		if (!context->res_ctx.pipe_ctx[i].stream)
+			continue;
+
+		if (context->res_ctx.pipe_ctx[i].plane_state)
+			active_hubp_count++;
+
+		pipes[pipe_idx].clks_cfg.dispclk_mhz = get_dispclk_calculated(&context->bw_ctx.dml, pipes, pipe_cnt);
+		pipes[pipe_idx].clks_cfg.dppclk_mhz = get_dppclk_calculated(&context->bw_ctx.dml, pipes, pipe_cnt, pipe_idx);
+
+		if (dc->config.forced_clocks || dc->debug.max_disp_clk) {
+			pipes[pipe_idx].clks_cfg.dispclk_mhz = context->bw_ctx.dml.soc.clock_limits[0].dispclk_mhz;
+			pipes[pipe_idx].clks_cfg.dppclk_mhz = context->bw_ctx.dml.soc.clock_limits[0].dppclk_mhz;
+		}
+		if (dc->debug.min_disp_clk_khz > pipes[pipe_idx].clks_cfg.dispclk_mhz * 1000)
+			pipes[pipe_idx].clks_cfg.dispclk_mhz = dc->debug.min_disp_clk_khz / 1000.0;
+		if (dc->debug.min_dpp_clk_khz > pipes[pipe_idx].clks_cfg.dppclk_mhz * 1000)
+			pipes[pipe_idx].clks_cfg.dppclk_mhz = dc->debug.min_dpp_clk_khz / 1000.0;
+
+		pipe_idx++;
+	}
+
+	dcn20_calculate_dlg_params(dc, context, pipes, pipe_cnt, vlevel);
+	/* For 31x apu pstate change is only supported if possible in vactive*/
+	context->bw_ctx.bw.dcn.clk.p_state_change_support =
+			context->bw_ctx.dml.vba.DRAMClockChangeSupport[vlevel][context->bw_ctx.dml.vba.maxMpcComb] == dm_dram_clock_change_vactive;
+	/* If DCN isn't making memory requests we can allow pstate change and lower clocks */
+	if (!active_hubp_count) {
+		context->bw_ctx.bw.dcn.clk.socclk_khz = 0;
+		context->bw_ctx.bw.dcn.clk.dppclk_khz = 0;
+		context->bw_ctx.bw.dcn.clk.dcfclk_khz = 0;
+		context->bw_ctx.bw.dcn.clk.dcfclk_deep_sleep_khz = 0;
+		context->bw_ctx.bw.dcn.clk.dramclk_khz = 0;
+		context->bw_ctx.bw.dcn.clk.fclk_khz = 0;
+		context->bw_ctx.bw.dcn.clk.p_state_change_support = true;
+		for (i = 0; i < dc->res_pool->pipe_count; i++)
+			if (context->res_ctx.pipe_ctx[i].stream)
+				context->res_ctx.pipe_ctx[i].plane_res.bw.dppclk_khz = 0;
+	}
+	for (i = 0, pipe_idx = 0; i < dc->res_pool->pipe_count; i++) {
+		if (!context->res_ctx.pipe_ctx[i].stream)
+			continue;
+
+		context->res_ctx.pipe_ctx[i].det_buffer_size_kb =
+				get_det_buffer_size_kbytes(&context->bw_ctx.dml, pipes, pipe_cnt, pipe_idx);
+		if (context->res_ctx.pipe_ctx[i].det_buffer_size_kb > 384)
+			context->res_ctx.pipe_ctx[i].det_buffer_size_kb /= 2;
+		total_det += context->res_ctx.pipe_ctx[i].det_buffer_size_kb;
+		pipe_idx++;
+	}
+	context->bw_ctx.bw.dcn.compbuf_size_kb = context->bw_ctx.dml.ip.config_return_buffer_size_in_kbytes - total_det;
+}
+
+void dcn31_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params)
+{
+	struct _vcs_dpi_voltage_scaling_st *s = dc->scratch.update_bw_bounding_box.clock_limits;
+	struct clk_limit_table *clk_table = &bw_params->clk_table;
+	unsigned int i, closest_clk_lvl;
+	int max_dispclk_mhz = 0, max_dppclk_mhz = 0;
+	int j;
+
+	dc_assert_fp_enabled();
+
+	memcpy(s, dcn3_1_soc.clock_limits, sizeof(dcn3_1_soc.clock_limits));
+
+	// Default clock levels are used for diags, which may lead to overclocking.
+	dcn3_1_ip.max_num_otg = dc->res_pool->res_cap->num_timing_generator;
+	dcn3_1_ip.max_num_dpp = dc->res_pool->pipe_count;
+	dcn3_1_soc.num_chans = bw_params->num_channels;
+
+	ASSERT(clk_table->num_entries);
+
+	/* Prepass to find max clocks independent of voltage level. */
+	for (i = 0; i < clk_table->num_entries; ++i) {
+		if (clk_table->entries[i].dispclk_mhz > max_dispclk_mhz)
+			max_dispclk_mhz = clk_table->entries[i].dispclk_mhz;
+		if (clk_table->entries[i].dppclk_mhz > max_dppclk_mhz)
+			max_dppclk_mhz = clk_table->entries[i].dppclk_mhz;
+	}
+
+	for (i = 0; i < clk_table->num_entries; i++) {
+		/* loop backwards*/
+		for (closest_clk_lvl = 0, j = dcn3_1_soc.num_states - 1; j >= 0; j--) {
+			if ((unsigned int) dcn3_1_soc.clock_limits[j].dcfclk_mhz <= clk_table->entries[i].dcfclk_mhz) {
+				closest_clk_lvl = j;
+				break;
+			}
+		}
+
+		s[i].state = i;
+
+		/* Clocks dependent on voltage level. */
+		s[i].dcfclk_mhz = clk_table->entries[i].dcfclk_mhz;
+		s[i].fabricclk_mhz = clk_table->entries[i].fclk_mhz;
+		s[i].socclk_mhz = clk_table->entries[i].socclk_mhz;
+		s[i].dram_speed_mts = clk_table->entries[i].memclk_mhz *
+			2 * clk_table->entries[i].wck_ratio;
+
+		/* Clocks independent of voltage level. */
+		s[i].dispclk_mhz = max_dispclk_mhz ? max_dispclk_mhz :
+			dcn3_1_soc.clock_limits[closest_clk_lvl].dispclk_mhz;
+
+		s[i].dppclk_mhz = max_dppclk_mhz ? max_dppclk_mhz :
+			dcn3_1_soc.clock_limits[closest_clk_lvl].dppclk_mhz;
+
+		s[i].dram_bw_per_chan_gbps =
+			dcn3_1_soc.clock_limits[closest_clk_lvl].dram_bw_per_chan_gbps;
+		s[i].dscclk_mhz = dcn3_1_soc.clock_limits[closest_clk_lvl].dscclk_mhz;
+		s[i].dtbclk_mhz = dcn3_1_soc.clock_limits[closest_clk_lvl].dtbclk_mhz;
+		s[i].phyclk_d18_mhz =
+			dcn3_1_soc.clock_limits[closest_clk_lvl].phyclk_d18_mhz;
+		s[i].phyclk_mhz = dcn3_1_soc.clock_limits[closest_clk_lvl].phyclk_mhz;
+	}
+
+	if (clk_table->num_entries)
+		dcn3_1_soc.num_states = clk_table->num_entries;
+
+	memcpy(dcn3_1_soc.clock_limits, s, sizeof(dcn3_1_soc.clock_limits));
+
+	dcn3_1_soc.dispclk_dppclk_vco_speed_mhz = dc->clk_mgr->dentist_vco_freq_khz / 1000.0;
+	dc->dml.soc.dispclk_dppclk_vco_speed_mhz = dc->clk_mgr->dentist_vco_freq_khz / 1000.0;
+
+	if ((int)(dcn3_1_soc.dram_clock_change_latency_us * 1000)
+				!= dc->debug.dram_clock_change_latency_ns
+			&& dc->debug.dram_clock_change_latency_ns) {
+		dcn3_1_soc.dram_clock_change_latency_us = dc->debug.dram_clock_change_latency_ns / 1000;
+	}
+
+	dml_init_instance(&dc->dml, &dcn3_1_soc, &dcn3_1_ip, DML_PROJECT_DCN31);
+}
+
+void dcn315_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params)
+{
+	struct clk_limit_table *clk_table = &bw_params->clk_table;
+	int i, max_dispclk_mhz = 0, max_dppclk_mhz = 0;
+
+	dc_assert_fp_enabled();
+
+	dcn3_15_ip.max_num_otg = dc->res_pool->res_cap->num_timing_generator;
+	dcn3_15_ip.max_num_dpp = dc->res_pool->pipe_count;
+
+	if (bw_params->num_channels > 0)
+		dcn3_15_soc.num_chans = bw_params->num_channels;
+	if (bw_params->dram_channel_width_bytes > 0)
+		dcn3_15_soc.dram_channel_width_bytes = bw_params->dram_channel_width_bytes;
+
+	ASSERT(clk_table->num_entries);
+
+	/* Setup soc to always use max dispclk/dppclk to avoid odm-to-lower-voltage */
+	for (i = 0; i < clk_table->num_entries; ++i) {
+		if (clk_table->entries[i].dispclk_mhz > max_dispclk_mhz)
+			max_dispclk_mhz = clk_table->entries[i].dispclk_mhz;
+		if (clk_table->entries[i].dppclk_mhz > max_dppclk_mhz)
+			max_dppclk_mhz = clk_table->entries[i].dppclk_mhz;
+	}
+
+	for (i = 0; i < clk_table->num_entries; i++) {
+		dcn3_15_soc.clock_limits[i].state = i;
+
+		/* Clocks dependent on voltage level. */
+		dcn3_15_soc.clock_limits[i].dcfclk_mhz = clk_table->entries[i].dcfclk_mhz;
+		dcn3_15_soc.clock_limits[i].fabricclk_mhz = clk_table->entries[i].fclk_mhz;
+		dcn3_15_soc.clock_limits[i].socclk_mhz = clk_table->entries[i].socclk_mhz;
+		dcn3_15_soc.clock_limits[i].dram_speed_mts = clk_table->entries[i].memclk_mhz * 2 * clk_table->entries[i].wck_ratio;
+
+		/* These aren't actually read from smu, but rather set in clk_mgr defaults */
+		dcn3_15_soc.clock_limits[i].dtbclk_mhz = clk_table->entries[i].dtbclk_mhz;
+		dcn3_15_soc.clock_limits[i].phyclk_d18_mhz = clk_table->entries[i].phyclk_d18_mhz;
+		dcn3_15_soc.clock_limits[i].phyclk_mhz = clk_table->entries[i].phyclk_mhz;
+
+		/* Clocks independent of voltage level. */
+		dcn3_15_soc.clock_limits[i].dispclk_mhz = max_dispclk_mhz;
+		dcn3_15_soc.clock_limits[i].dppclk_mhz = max_dppclk_mhz;
+		dcn3_15_soc.clock_limits[i].dscclk_mhz = max_dispclk_mhz / 3.0;
+	}
+	dcn3_15_soc.num_states = clk_table->num_entries;
+
+
+	/* Set vco to max_dispclk * 2 to make sure the highest dispclk is always available for dml calcs,
+	 * no impact outside of dml validation
+	 */
+	dcn3_15_soc.dispclk_dppclk_vco_speed_mhz = max_dispclk_mhz * 2;
+
+	if ((int)(dcn3_15_soc.dram_clock_change_latency_us * 1000)
+				!= dc->debug.dram_clock_change_latency_ns
+			&& dc->debug.dram_clock_change_latency_ns) {
+		dcn3_15_soc.dram_clock_change_latency_us = dc->debug.dram_clock_change_latency_ns / 1000;
+	}
+
+	dml_init_instance(&dc->dml, &dcn3_15_soc, &dcn3_15_ip, DML_PROJECT_DCN315);
+}
+
+void dcn316_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params)
+{
+	struct _vcs_dpi_voltage_scaling_st *s = dc->scratch.update_bw_bounding_box.clock_limits;
+	struct clk_limit_table *clk_table = &bw_params->clk_table;
+	unsigned int i, closest_clk_lvl;
+	int max_dispclk_mhz = 0, max_dppclk_mhz = 0;
+	int j;
+
+	dc_assert_fp_enabled();
+
+	memcpy(s, dcn3_16_soc.clock_limits, sizeof(dcn3_16_soc.clock_limits));
+
+	// Default clock levels are used for diags, which may lead to overclocking.
+	dcn3_16_ip.max_num_otg = dc->res_pool->res_cap->num_timing_generator;
+	dcn3_16_ip.max_num_dpp = dc->res_pool->pipe_count;
+	dcn3_16_soc.num_chans = bw_params->num_channels;
+
+	ASSERT(clk_table->num_entries);
+
+	/* Prepass to find max clocks independent of voltage level. */
+	for (i = 0; i < clk_table->num_entries; ++i) {
+		if (clk_table->entries[i].dispclk_mhz > max_dispclk_mhz)
+			max_dispclk_mhz = clk_table->entries[i].dispclk_mhz;
+		if (clk_table->entries[i].dppclk_mhz > max_dppclk_mhz)
+			max_dppclk_mhz = clk_table->entries[i].dppclk_mhz;
+	}
+
+	for (i = 0; i < clk_table->num_entries; i++) {
+		/* loop backwards*/
+		for (closest_clk_lvl = 0, j = dcn3_16_soc.num_states - 1; j >= 0; j--) {
+			if ((unsigned int) dcn3_16_soc.clock_limits[j].dcfclk_mhz <=
+			    clk_table->entries[i].dcfclk_mhz) {
+				closest_clk_lvl = j;
+				break;
+			}
+		}
+
+		s[i].state = i;
+
+		/* Clocks dependent on voltage level. */
+		s[i].dcfclk_mhz = clk_table->entries[i].dcfclk_mhz;
+		s[i].fabricclk_mhz = clk_table->entries[i].fclk_mhz;
+		s[i].socclk_mhz = clk_table->entries[i].socclk_mhz;
+		s[i].dram_speed_mts = clk_table->entries[i].memclk_mhz *
+			2 * clk_table->entries[i].wck_ratio;
+
+		/* Clocks independent of voltage level. */
+		s[i].dispclk_mhz = max_dispclk_mhz ? max_dispclk_mhz :
+			dcn3_16_soc.clock_limits[closest_clk_lvl].dispclk_mhz;
+
+		s[i].dppclk_mhz = max_dppclk_mhz ? max_dppclk_mhz :
+			dcn3_16_soc.clock_limits[closest_clk_lvl].dppclk_mhz;
+
+		s[i].dram_bw_per_chan_gbps =
+			dcn3_16_soc.clock_limits[closest_clk_lvl].dram_bw_per_chan_gbps;
+		s[i].dscclk_mhz = dcn3_16_soc.clock_limits[closest_clk_lvl].dscclk_mhz;
+		s[i].dtbclk_mhz = dcn3_16_soc.clock_limits[closest_clk_lvl].dtbclk_mhz;
+		s[i].phyclk_d18_mhz =
+			dcn3_16_soc.clock_limits[closest_clk_lvl].phyclk_d18_mhz;
+		s[i].phyclk_mhz = dcn3_16_soc.clock_limits[closest_clk_lvl].phyclk_mhz;
+	}
+
+	if (clk_table->num_entries)
+		dcn3_16_soc.num_states = clk_table->num_entries;
+
+	memcpy(dcn3_16_soc.clock_limits, s, sizeof(dcn3_16_soc.clock_limits));
+
+	if (max_dispclk_mhz) {
+		dcn3_16_soc.dispclk_dppclk_vco_speed_mhz = max_dispclk_mhz * 2;
+		dc->dml.soc.dispclk_dppclk_vco_speed_mhz = max_dispclk_mhz * 2;
+	}
+	if ((int)(dcn3_16_soc.dram_clock_change_latency_us * 1000)
+				!= dc->debug.dram_clock_change_latency_ns
+			&& dc->debug.dram_clock_change_latency_ns) {
+		dcn3_16_soc.dram_clock_change_latency_us = dc->debug.dram_clock_change_latency_ns / 1000;
+	}
+
+	dml_init_instance(&dc->dml, &dcn3_16_soc, &dcn3_16_ip, DML_PROJECT_DCN31);
+}
+
+int dcn_get_max_non_odm_pix_rate_100hz(struct _vcs_dpi_soc_bounding_box_st *soc)
+{
+	dc_assert_fp_enabled();
+
+	return soc->clock_limits[0].dispclk_mhz * 10000.0 / (1.0 + soc->dcn_downspread_percent / 100.0);
+}
+
+int dcn_get_approx_det_segs_required_for_pstate(
+		struct _vcs_dpi_soc_bounding_box_st *soc,
+		int pix_clk_100hz, int bpp, int seg_size_kb)
+{
+	dc_assert_fp_enabled();
+
+	/* Roughly calculate required crb to hide latency. In practice there is slightly
+	 * more buffer available for latency hiding
+	 */
+	return (int)(soc->dram_clock_change_latency_us * pix_clk_100hz * bpp
+					/ 10240000 + seg_size_kb - 1) /	seg_size_kb;
+}
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn31/dcn31_fpu.h b/drivers/gpu/drm/amd/display/dc/dml/dcn31/dcn31_fpu.h
new file mode 100644
index 000000000000..dfcc5d50071e
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn31/dcn31_fpu.h
@@ -0,0 +1,59 @@
+/*
+ * Copyright 2019-2021 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __DCN31_FPU_H__
+#define __DCN31_FPU_H__
+
+#define DCN3_1_DEFAULT_DET_SIZE 384
+#define DCN3_15_DEFAULT_DET_SIZE 192
+#define DCN3_15_MIN_COMPBUF_SIZE_KB 128
+#define DCN3_16_DEFAULT_DET_SIZE 192
+#define DCN3_16_MIN_COMPBUF_SIZE_KB 128
+
+void dcn31_zero_pipe_dcc_fraction(display_e2e_pipe_params_st *pipes,
+				  int pipe_cnt);
+
+void dcn31_update_soc_for_wm_a(struct dc *dc, struct dc_state *context);
+void dcn315_update_soc_for_wm_a(struct dc *dc, struct dc_state *context);
+
+void dcn31_calculate_wm_and_dlg_fp(
+		struct dc *dc, struct dc_state *context,
+		display_e2e_pipe_params_st *pipes,
+		int pipe_cnt,
+		int vlevel);
+
+void dcn31_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params);
+void dcn315_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params);
+void dcn316_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params);
+int dcn_get_max_non_odm_pix_rate_100hz(struct _vcs_dpi_soc_bounding_box_st *soc);
+int dcn_get_approx_det_segs_required_for_pstate(
+		struct _vcs_dpi_soc_bounding_box_st *soc,
+		int pix_clk_100hz, int bpp, int seg_size_kb);
+
+int dcn31x_populate_dml_pipes_from_context(struct dc *dc,
+					  struct dc_state *context,
+					  display_e2e_pipe_params_st *pipes,
+					  enum dc_validate_mode validate_mode);
+#endif /* __DCN31_FPU_H__*/
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c b/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c
index 7e937bdcea00..ed59c77bc6f6 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c
@@ -24,8 +24,8 @@
  */
 
 #include "dc.h"
-#include "dc_link.h"
 #include "../display_mode_lib.h"
+#include "../dcn30/display_mode_vba_30.h"
 #include "display_mode_vba_31.h"
 #include "../dml_inline_defs.h"
 
@@ -42,6 +42,8 @@
 #define BPP_BLENDED_PIPE 0xffffffff
 #define DCN31_MAX_DSC_IMAGE_WIDTH 5184
 #define DCN31_MAX_FMT_420_BUFFER_WIDTH 4096
+#define DCN3_15_MIN_COMPBUF_SIZE_KB 128
+#define DCN3_15_MAX_DET_SIZE 384
 
 // For DML-C changes that hasn't been propagated to VBA yet
 //#define __DML_VBA_ALLOW_DELTA__
@@ -64,6 +66,8 @@ typedef struct {
 	double DCFCLKDeepSleep;
 	unsigned int DPPPerPlane;
 	bool ScalerEnabled;
+	double VRatio;
+	double VRatioChroma;
 	enum scan_direction_class SourceScan;
 	unsigned int BlockWidth256BytesY;
 	unsigned int BlockHeight256BytesY;
@@ -84,17 +88,6 @@ typedef struct {
 #define BPP_INVALID 0
 #define BPP_BLENDED_PIPE 0xffffffff
 
-static bool CalculateBytePerPixelAnd256BBlockSizes(
-		enum source_format_class SourcePixelFormat,
-		enum dm_swizzle_mode SurfaceTiling,
-		unsigned int *BytePerPixelY,
-		unsigned int *BytePerPixelC,
-		double *BytePerPixelDETY,
-		double *BytePerPixelDETC,
-		unsigned int *BlockHeight256BytesY,
-		unsigned int *BlockHeight256BytesC,
-		unsigned int *BlockWidth256BytesY,
-		unsigned int *BlockWidth256BytesC);
 static void DisplayPipeConfiguration(struct display_mode_lib *mode_lib);
 static void DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation(struct display_mode_lib *mode_lib);
 static unsigned int dscceComputeDelay(
@@ -259,33 +252,13 @@ static void CalculateRowBandwidth(
 
 static void CalculateFlipSchedule(
 		struct display_mode_lib *mode_lib,
+		unsigned int k,
 		double HostVMInefficiencyFactor,
 		double UrgentExtraLatency,
 		double UrgentLatency,
-		unsigned int GPUVMMaxPageTableLevels,
-		bool HostVMEnable,
-		unsigned int HostVMMaxNonCachedPageTableLevels,
-		bool GPUVMEnable,
-		double HostVMMinPageSize,
 		double PDEAndMetaPTEBytesPerFrame,
 		double MetaRowBytes,
-		double DPTEBytesPerRow,
-		double BandwidthAvailableForImmediateFlip,
-		unsigned int TotImmediateFlipBytes,
-		enum source_format_class SourcePixelFormat,
-		double LineTime,
-		double VRatio,
-		double VRatioChroma,
-		double Tno_bw,
-		bool DCCEnable,
-		unsigned int dpte_row_height,
-		unsigned int meta_row_height,
-		unsigned int dpte_row_height_chroma,
-		unsigned int meta_row_height_chroma,
-		double *DestinationLinesToRequestVMInImmediateFlip,
-		double *DestinationLinesToRequestRowInImmediateFlip,
-		double *final_flip_bw,
-		bool *ImmediateFlipSupportedForPipe);
+		double DPTEBytesPerRow);
 static double CalculateWriteBackDelay(
 		enum source_format_class WritebackPixelFormat,
 		double WritebackHRatio,
@@ -319,64 +292,28 @@ static void CalculateVupdateAndDynamicMetadataParameters(
 static void CalculateWatermarksAndDRAMSpeedChangeSupport(
 		struct display_mode_lib *mode_lib,
 		unsigned int PrefetchMode,
-		unsigned int NumberOfActivePlanes,
-		unsigned int MaxLineBufferLines,
-		unsigned int LineBufferSize,
-		unsigned int WritebackInterfaceBufferSize,
 		double DCFCLK,
 		double ReturnBW,
-		bool SynchronizedVBlank,
-		unsigned int dpte_group_bytes[],
-		unsigned int MetaChunkSize,
 		double UrgentLatency,
 		double ExtraLatency,
-		double WritebackLatency,
-		double WritebackChunkSize,
 		double SOCCLK,
-		double DRAMClockChangeLatency,
-		double SRExitTime,
-		double SREnterPlusExitTime,
-		double SRExitZ8Time,
-		double SREnterPlusExitZ8Time,
 		double DCFCLKDeepSleep,
 		unsigned int DETBufferSizeY[],
 		unsigned int DETBufferSizeC[],
 		unsigned int SwathHeightY[],
 		unsigned int SwathHeightC[],
-		unsigned int LBBitPerPixel[],
 		double SwathWidthY[],
 		double SwathWidthC[],
-		double HRatio[],
-		double HRatioChroma[],
-		unsigned int vtaps[],
-		unsigned int VTAPsChroma[],
-		double VRatio[],
-		double VRatioChroma[],
-		unsigned int HTotal[],
-		double PixelClock[],
-		unsigned int BlendingAndTiming[],
 		unsigned int DPPPerPlane[],
 		double BytePerPixelDETY[],
 		double BytePerPixelDETC[],
-		double DSTXAfterScaler[],
-		double DSTYAfterScaler[],
-		bool WritebackEnable[],
-		enum source_format_class WritebackPixelFormat[],
-		double WritebackDestinationWidth[],
-		double WritebackDestinationHeight[],
-		double WritebackSourceHeight[],
 		bool UnboundedRequestEnabled,
 		int unsigned CompressedBufferSizeInkByte,
 		enum clock_change_support *DRAMClockChangeSupport,
-		double *UrgentWatermark,
-		double *WritebackUrgentWatermark,
-		double *DRAMClockChangeWatermark,
-		double *WritebackDRAMClockChangeWatermark,
 		double *StutterExitWatermark,
 		double *StutterEnterPlusExitWatermark,
 		double *Z8StutterExitWatermark,
-		double *Z8StutterEnterPlusExitWatermark,
-		double *MinActiveDRAMClockChangeLatencySupported);
+		double *Z8StutterEnterPlusExitWatermark);
 
 static void CalculateDCFCLKDeepSleep(
 		struct display_mode_lib *mode_lib,
@@ -422,62 +359,8 @@ static void CalculateUrgentBurstFactor(
 
 static void UseMinimumDCFCLK(
 		struct display_mode_lib *mode_lib,
-		int MaxInterDCNTileRepeaters,
 		int MaxPrefetchMode,
-		double FinalDRAMClockChangeLatency,
-		double SREnterPlusExitTime,
-		int ReturnBusWidth,
-		int RoundTripPingLatencyCycles,
-		int ReorderingBytes,
-		int PixelChunkSizeInKByte,
-		int MetaChunkSize,
-		bool GPUVMEnable,
-		int GPUVMMaxPageTableLevels,
-		bool HostVMEnable,
-		int NumberOfActivePlanes,
-		double HostVMMinPageSize,
-		int HostVMMaxNonCachedPageTableLevels,
-		bool DynamicMetadataVMEnabled,
-		enum immediate_flip_requirement ImmediateFlipRequirement,
-		bool ProgressiveToInterlaceUnitInOPP,
-		double MaxAveragePercentOfIdealFabricAndSDPPortBWDisplayCanUseInNormalSystemOperation,
-		double PercentOfIdealFabricAndSDPPortBWReceivedAfterUrgLatency,
-		int VTotal[],
-		int VActive[],
-		int DynamicMetadataTransmittedBytes[],
-		int DynamicMetadataLinesBeforeActiveRequired[],
-		bool Interlace[],
-		double RequiredDPPCLK[][2][DC__NUM_DPP__MAX],
-		double RequiredDISPCLK[][2],
-		double UrgLatency[],
-		unsigned int NoOfDPP[][2][DC__NUM_DPP__MAX],
-		double ProjectedDCFCLKDeepSleep[][2],
-		double MaximumVStartup[][2][DC__NUM_DPP__MAX],
-		double TotalVActivePixelBandwidth[][2],
-		double TotalVActiveCursorBandwidth[][2],
-		double TotalMetaRowBandwidth[][2],
-		double TotalDPTERowBandwidth[][2],
-		unsigned int TotalNumberOfActiveDPP[][2],
-		unsigned int TotalNumberOfDCCActiveDPP[][2],
-		int dpte_group_bytes[],
-		double PrefetchLinesY[][2][DC__NUM_DPP__MAX],
-		double PrefetchLinesC[][2][DC__NUM_DPP__MAX],
-		int swath_width_luma_ub_all_states[][2][DC__NUM_DPP__MAX],
-		int swath_width_chroma_ub_all_states[][2][DC__NUM_DPP__MAX],
-		int BytePerPixelY[],
-		int BytePerPixelC[],
-		int HTotal[],
-		double PixelClock[],
-		double PDEAndMetaPTEBytesPerFrame[][2][DC__NUM_DPP__MAX],
-		double DPTEBytesPerRow[][2][DC__NUM_DPP__MAX],
-		double MetaRowBytes[][2][DC__NUM_DPP__MAX],
-		bool DynamicMetadataEnable[],
-		double VActivePixelBandwidth[][2][DC__NUM_DPP__MAX],
-		double VActiveCursorBandwidth[][2][DC__NUM_DPP__MAX],
-		double ReadBandwidthLuma[],
-		double ReadBandwidthChroma[],
-		double DCFCLKPerState[],
-		double DCFCLKState[][2]);
+		int ReorderingBytes);
 
 static void CalculatePixelDeliveryTimes(
 		unsigned int NumberOfActivePlanes,
@@ -649,7 +532,8 @@ static void CalculateStutterEfficiency(
 static void CalculateSwathAndDETConfiguration(
 		bool ForceSingleDPP,
 		int NumberOfActivePlanes,
-		unsigned int DETBufferSizeInKByte,
+		bool DETSharedByAllDPP,
+		unsigned int DETBufferSizeInKByte[],
 		double MaximumSwathWidthLuma[],
 		double MaximumSwathWidthChroma[],
 		enum scan_direction_class SourceScan[],
@@ -994,8 +878,11 @@ static bool CalculatePrefetchSchedule(
 	double DSTTotalPixelsAfterScaler;
 	double LineTime;
 	double dst_y_prefetch_equ;
+#ifdef __DML_VBA_DEBUG__
 	double Tsw_oto;
+#endif
 	double prefetch_bw_oto;
+	double prefetch_bw_pr;
 	double Tvm_oto;
 	double Tr0_oto;
 	double Tvm_oto_lines;
@@ -1025,6 +912,7 @@ static bool CalculatePrefetchSchedule(
 	double min_Lsw;
 	double Tsw_est1 = 0;
 	double Tsw_est3 = 0;
+	double  max_Tsw = 0;
 
 	if (GPUVMEnable == true && HostVMEnable == true) {
 		HostVMDynamicLevelsTrips = HostVMMaxNonCachedPageTableLevels;
@@ -1165,15 +1053,19 @@ static bool CalculatePrefetchSchedule(
 		bytes_pp = myPipe->BytePerPixelY + myPipe->BytePerPixelC / 4;
 	else
 		bytes_pp = myPipe->BytePerPixelY + myPipe->BytePerPixelC;
-
+	/*rev 99*/
+	prefetch_bw_pr = bytes_pp * myPipe->PixelClock / (double) myPipe->DPPPerPlane;
+	prefetch_bw_pr = dml_min(1, myPipe->VRatio) * prefetch_bw_pr;
+	max_Tsw = dml_max(PrefetchSourceLinesY, PrefetchSourceLinesC) * LineTime;
 	prefetch_sw_bytes = PrefetchSourceLinesY * swath_width_luma_ub * myPipe->BytePerPixelY + PrefetchSourceLinesC * swath_width_chroma_ub * myPipe->BytePerPixelC;
-	prefetch_bw_oto = dml_max(bytes_pp * myPipe->PixelClock / myPipe->DPPPerPlane, prefetch_sw_bytes / (dml_max(PrefetchSourceLinesY, PrefetchSourceLinesC) * LineTime));
+	prefetch_bw_oto = dml_max(prefetch_bw_pr, prefetch_sw_bytes / max_Tsw);
 
-	min_Lsw = dml_max(PrefetchSourceLinesY, PrefetchSourceLinesC) / max_vratio_pre;
+	min_Lsw = dml_max(1, dml_max(PrefetchSourceLinesY, PrefetchSourceLinesC) / max_vratio_pre);
 	Lsw_oto = dml_ceil(4 * dml_max(prefetch_sw_bytes / prefetch_bw_oto / LineTime, min_Lsw), 1) / 4;
+#ifdef __DML_VBA_DEBUG__
 	Tsw_oto = Lsw_oto * LineTime;
+#endif
 
-	prefetch_bw_oto = (PrefetchSourceLinesY * swath_width_luma_ub * myPipe->BytePerPixelY + PrefetchSourceLinesC * swath_width_chroma_ub * myPipe->BytePerPixelC) / Tsw_oto;
 
 #ifdef __DML_VBA_DEBUG__
 	dml_print("DML: HTotal: %d\n", myPipe->HTotal);
@@ -1213,6 +1105,7 @@ static bool CalculatePrefetchSchedule(
 	Tr0_oto_lines = dml_ceil(4.0 * Tr0_oto / LineTime, 1) / 4.0;
 	dst_y_prefetch_oto = Tvm_oto_lines + 2 * Tr0_oto_lines + Lsw_oto;
 	dst_y_prefetch_equ =  VStartup - (*TSetup + dml_max(TWait + TCalc, *Tdmdl)) / LineTime - (*DSTYAfterScaler + *DSTXAfterScaler / myPipe->HTotal);
+	dst_y_prefetch_equ = dml_min(dst_y_prefetch_equ, 63.75); // limit to the reg limit of U6.2 for DST_Y_PREFETCH
 	dst_y_prefetch_equ = dml_floor(4.0 * (dst_y_prefetch_equ + 0.125), 1) / 4.0;
 	Tpre_rounded = dst_y_prefetch_equ * LineTime;
 
@@ -1443,7 +1336,7 @@ static bool CalculatePrefetchSchedule(
 			dml_print("DML::%s: SwathHeightC = %d\n", __func__, SwathHeightC);
 			dml_print("DML::%s: VInitPreFillC = %f\n", __func__, VInitPreFillC);
 #endif
-			if ((SwathHeightC > 4)) {
+			if ((SwathHeightC > 4) || VInitPreFillC > 3) {
 				if (LinesToRequestPrefetchPixelData > (VInitPreFillC - 3.0) / 2.0) {
 					*VRatioPrefetchC = dml_max(
 							*VRatioPrefetchC,
@@ -1552,12 +1445,9 @@ static bool CalculatePrefetchSchedule(
 
 	if (MyError) {
 		*PrefetchBandwidth = 0;
-		TimeForFetchingMetaPTE = 0;
-		TimeForFetchingRowInVBlank = 0;
 		*DestinationLinesToRequestVMInVBlank = 0;
 		*DestinationLinesToRequestRowInVBlank = 0;
 		*DestinationLinesForPrefetch = 0;
-		LinesToRequestPrefetchPixelData = 0;
 		*VRatioPrefetchY = 0;
 		*VRatioPrefetchC = 0;
 		*RequiredPrefetchPixDataBWLuma = 0;
@@ -2032,15 +1922,6 @@ static unsigned int CalculateVMAndRowBytes(
 		*PixelPTEReqWidth = 32768.0 / BytePerPixel;
 		*PTERequestSize = 64;
 		FractionOfPTEReturnDrop = 0;
-	} else if (MacroTileSizeBytes == 4096) {
-		PixelPTEReqHeightPTEs = 1;
-		*PixelPTEReqHeight = MacroTileHeight;
-		*PixelPTEReqWidth = 8 * *MacroTileWidth;
-		*PTERequestSize = 64;
-		if (ScanDirection != dm_vert)
-			FractionOfPTEReturnDrop = 0;
-		else
-			FractionOfPTEReturnDrop = 7 / 8;
 	} else if (GPUVMMinPageSize == 4 && MacroTileSizeBytes > 4096) {
 		PixelPTEReqHeightPTEs = 16;
 		*PixelPTEReqHeight = 16 * BlockHeight256Bytes;
@@ -2267,7 +2148,7 @@ static void DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerforman
 	DTRACE("   return_bus_bw      = %f", v->ReturnBW);
 
 	for (k = 0; k < v->NumberOfActivePlanes; ++k) {
-		CalculateBytePerPixelAnd256BBlockSizes(
+		dml30_CalculateBytePerPixelAnd256BBlockSizes(
 				v->SourcePixelFormat[k],
 				v->SurfaceTiling[k],
 				&v->BytePerPixelY[k],
@@ -2717,6 +2598,8 @@ static void DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerforman
 			myPipe.DCFCLKDeepSleep = v->DCFCLKDeepSleep;
 			myPipe.DPPPerPlane = v->DPPPerPlane[k];
 			myPipe.ScalerEnabled = v->ScalerEnabled[k];
+			myPipe.VRatio = v->VRatio[k];
+			myPipe.VRatioChroma = v->VRatioChroma[k];
 			myPipe.SourceScan = v->SourceScan[k];
 			myPipe.BlockWidth256BytesY = v->BlockWidth256BytesY[k];
 			myPipe.BlockHeight256BytesY = v->BlockHeight256BytesY[k];
@@ -2959,33 +2842,13 @@ static void DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerforman
 			for (k = 0; k < v->NumberOfActivePlanes; ++k) {
 				CalculateFlipSchedule(
 						mode_lib,
+						k,
 						HostVMInefficiencyFactor,
 						v->UrgentExtraLatency,
 						v->UrgentLatency,
-						v->GPUVMMaxPageTableLevels,
-						v->HostVMEnable,
-						v->HostVMMaxNonCachedPageTableLevels,
-						v->GPUVMEnable,
-						v->HostVMMinPageSize,
 						v->PDEAndMetaPTEBytesFrame[k],
 						v->MetaRowByte[k],
-						v->PixelPTEBytesPerRow[k],
-						v->BandwidthAvailableForImmediateFlip,
-						v->TotImmediateFlipBytes,
-						v->SourcePixelFormat[k],
-						v->HTotal[k] / v->PixelClock[k],
-						v->VRatio[k],
-						v->VRatioChroma[k],
-						v->Tno_bw[k],
-						v->DCCEnable[k],
-						v->dpte_row_height[k],
-						v->meta_row_height[k],
-						v->dpte_row_height_chroma[k],
-						v->meta_row_height_chroma[k],
-						&v->DestinationLinesToRequestVMInImmediateFlip[k],
-						&v->DestinationLinesToRequestRowInImmediateFlip[k],
-						&v->final_flip_bw[k],
-						&v->ImmediateFlipSupportedForPipe[k]);
+						v->PixelPTEBytesPerRow[k]);
 			}
 
 			v->total_dcn_read_bw_with_flip = 0.0;
@@ -3041,7 +2904,7 @@ static void DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerforman
 				v->ImmediateFlipSupported)) ? true : false;
 #ifdef __DML_VBA_DEBUG__
 		dml_print("DML::%s: PrefetchModeSupported %d\n", __func__, v->PrefetchModeSupported);
-		dml_print("DML::%s: ImmediateFlipRequirement %d\n", __func__, v->ImmediateFlipRequirement == dm_immediate_flip_required);
+		dml_print("DML::%s: ImmediateFlipRequirement[0] %d\n", __func__, v->ImmediateFlipRequirement[0] == dm_immediate_flip_required);
 		dml_print("DML::%s: ImmediateFlipSupported %d\n", __func__, v->ImmediateFlipSupported);
 		dml_print("DML::%s: ImmediateFlipSupport %d\n", __func__, v->ImmediateFlipSupport);
 		dml_print("DML::%s: HostVMEnable %d\n", __func__, v->HostVMEnable);
@@ -3072,64 +2935,28 @@ static void DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerforman
 		CalculateWatermarksAndDRAMSpeedChangeSupport(
 				mode_lib,
 				PrefetchMode,
-				v->NumberOfActivePlanes,
-				v->MaxLineBufferLines,
-				v->LineBufferSize,
-				v->WritebackInterfaceBufferSize,
 				v->DCFCLK,
 				v->ReturnBW,
-				v->SynchronizedVBlank,
-				v->dpte_group_bytes,
-				v->MetaChunkSize,
 				v->UrgentLatency,
 				v->UrgentExtraLatency,
-				v->WritebackLatency,
-				v->WritebackChunkSize,
 				v->SOCCLK,
-				v->DRAMClockChangeLatency,
-				v->SRExitTime,
-				v->SREnterPlusExitTime,
-				v->SRExitZ8Time,
-				v->SREnterPlusExitZ8Time,
 				v->DCFCLKDeepSleep,
 				v->DETBufferSizeY,
 				v->DETBufferSizeC,
 				v->SwathHeightY,
 				v->SwathHeightC,
-				v->LBBitPerPixel,
 				v->SwathWidthY,
 				v->SwathWidthC,
-				v->HRatio,
-				v->HRatioChroma,
-				v->vtaps,
-				v->VTAPsChroma,
-				v->VRatio,
-				v->VRatioChroma,
-				v->HTotal,
-				v->PixelClock,
-				v->BlendingAndTiming,
 				v->DPPPerPlane,
 				v->BytePerPixelDETY,
 				v->BytePerPixelDETC,
-				v->DSTXAfterScaler,
-				v->DSTYAfterScaler,
-				v->WritebackEnable,
-				v->WritebackPixelFormat,
-				v->WritebackDestinationWidth,
-				v->WritebackDestinationHeight,
-				v->WritebackSourceHeight,
 				v->UnboundedRequestEnabled,
 				v->CompressedBufferSizeInkByte,
 				&DRAMClockChangeSupport,
-				&v->UrgentWatermark,
-				&v->WritebackUrgentWatermark,
-				&v->DRAMClockChangeWatermark,
-				&v->WritebackDRAMClockChangeWatermark,
 				&v->StutterExitWatermark,
 				&v->StutterEnterPlusExitWatermark,
 				&v->Z8StutterExitWatermark,
-				&v->Z8StutterEnterPlusExitWatermark,
-				&v->MinActiveDRAMClockChangeLatencySupported);
+				&v->Z8StutterEnterPlusExitWatermark);
 
 		for (k = 0; k < v->NumberOfActivePlanes; ++k) {
 			if (v->WritebackEnable[k] == true) {
@@ -3281,7 +3108,7 @@ static void DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerforman
 				v->SurfaceWidthC[k],
 				v->SurfaceHeightY[k],
 				v->SurfaceHeightC[k],
-				v->DETBufferSizeInKByte[0] * 1024,
+				v->DETBufferSizeInKByte[k] * 1024,
 				v->BlockHeight256BytesY[k],
 				v->BlockHeight256BytesC[k],
 				v->SurfaceTiling[k],
@@ -3460,7 +3287,7 @@ static void DisplayPipeConfiguration(struct display_mode_lib *mode_lib)
 
 	for (k = 0; k < v->NumberOfActivePlanes; ++k) {
 
-		CalculateBytePerPixelAnd256BBlockSizes(
+		dml30_CalculateBytePerPixelAnd256BBlockSizes(
 				v->SourcePixelFormat[k],
 				v->SurfaceTiling[k],
 				&BytePerPixY[k],
@@ -3476,7 +3303,8 @@ static void DisplayPipeConfiguration(struct display_mode_lib *mode_lib)
 	CalculateSwathAndDETConfiguration(
 			false,
 			v->NumberOfActivePlanes,
-			v->DETBufferSizeInKByte[0],
+			mode_lib->project == DML_PROJECT_DCN315 && v->DETSizeOverride[0],
+			v->DETBufferSizeInKByte,
 			dummy1,
 			dummy2,
 			v->SourceScan,
@@ -3514,94 +3342,6 @@ static void DisplayPipeConfiguration(struct display_mode_lib *mode_lib)
 			&dummysinglestring);
 }
 
-static bool CalculateBytePerPixelAnd256BBlockSizes(
-		enum source_format_class SourcePixelFormat,
-		enum dm_swizzle_mode SurfaceTiling,
-		unsigned int *BytePerPixelY,
-		unsigned int *BytePerPixelC,
-		double *BytePerPixelDETY,
-		double *BytePerPixelDETC,
-		unsigned int *BlockHeight256BytesY,
-		unsigned int *BlockHeight256BytesC,
-		unsigned int *BlockWidth256BytesY,
-		unsigned int *BlockWidth256BytesC)
-{
-	if (SourcePixelFormat == dm_444_64) {
-		*BytePerPixelDETY = 8;
-		*BytePerPixelDETC = 0;
-		*BytePerPixelY = 8;
-		*BytePerPixelC = 0;
-	} else if (SourcePixelFormat == dm_444_32 || SourcePixelFormat == dm_rgbe) {
-		*BytePerPixelDETY = 4;
-		*BytePerPixelDETC = 0;
-		*BytePerPixelY = 4;
-		*BytePerPixelC = 0;
-	} else if (SourcePixelFormat == dm_444_16) {
-		*BytePerPixelDETY = 2;
-		*BytePerPixelDETC = 0;
-		*BytePerPixelY = 2;
-		*BytePerPixelC = 0;
-	} else if (SourcePixelFormat == dm_444_8) {
-		*BytePerPixelDETY = 1;
-		*BytePerPixelDETC = 0;
-		*BytePerPixelY = 1;
-		*BytePerPixelC = 0;
-	} else if (SourcePixelFormat == dm_rgbe_alpha) {
-		*BytePerPixelDETY = 4;
-		*BytePerPixelDETC = 1;
-		*BytePerPixelY = 4;
-		*BytePerPixelC = 1;
-	} else if (SourcePixelFormat == dm_420_8) {
-		*BytePerPixelDETY = 1;
-		*BytePerPixelDETC = 2;
-		*BytePerPixelY = 1;
-		*BytePerPixelC = 2;
-	} else if (SourcePixelFormat == dm_420_12) {
-		*BytePerPixelDETY = 2;
-		*BytePerPixelDETC = 4;
-		*BytePerPixelY = 2;
-		*BytePerPixelC = 4;
-	} else {
-		*BytePerPixelDETY = 4.0 / 3;
-		*BytePerPixelDETC = 8.0 / 3;
-		*BytePerPixelY = 2;
-		*BytePerPixelC = 4;
-	}
-
-	if ((SourcePixelFormat == dm_444_64 || SourcePixelFormat == dm_444_32 || SourcePixelFormat == dm_444_16 || SourcePixelFormat == dm_444_8 || SourcePixelFormat == dm_mono_16
-			|| SourcePixelFormat == dm_mono_8 || SourcePixelFormat == dm_rgbe)) {
-		if (SurfaceTiling == dm_sw_linear) {
-			*BlockHeight256BytesY = 1;
-		} else if (SourcePixelFormat == dm_444_64) {
-			*BlockHeight256BytesY = 4;
-		} else if (SourcePixelFormat == dm_444_8) {
-			*BlockHeight256BytesY = 16;
-		} else {
-			*BlockHeight256BytesY = 8;
-		}
-		*BlockWidth256BytesY = 256U / *BytePerPixelY / *BlockHeight256BytesY;
-		*BlockHeight256BytesC = 0;
-		*BlockWidth256BytesC = 0;
-	} else {
-		if (SurfaceTiling == dm_sw_linear) {
-			*BlockHeight256BytesY = 1;
-			*BlockHeight256BytesC = 1;
-		} else if (SourcePixelFormat == dm_rgbe_alpha) {
-			*BlockHeight256BytesY = 8;
-			*BlockHeight256BytesC = 16;
-		} else if (SourcePixelFormat == dm_420_8) {
-			*BlockHeight256BytesY = 16;
-			*BlockHeight256BytesC = 8;
-		} else {
-			*BlockHeight256BytesY = 8;
-			*BlockHeight256BytesC = 8;
-		}
-		*BlockWidth256BytesY = 256U / *BytePerPixelY / *BlockHeight256BytesY;
-		*BlockWidth256BytesC = 256U / *BytePerPixelC / *BlockHeight256BytesC;
-	}
-	return true;
-}
-
 static double CalculateTWait(unsigned int PrefetchMode, double DRAMClockChangeLatency, double UrgentLatency, double SREnterPlusExitTime)
 {
 	if (PrefetchMode == 0) {
@@ -3741,61 +3481,43 @@ static void CalculateRowBandwidth(
 
 static void CalculateFlipSchedule(
 		struct display_mode_lib *mode_lib,
+		unsigned int k,
 		double HostVMInefficiencyFactor,
 		double UrgentExtraLatency,
 		double UrgentLatency,
-		unsigned int GPUVMMaxPageTableLevels,
-		bool HostVMEnable,
-		unsigned int HostVMMaxNonCachedPageTableLevels,
-		bool GPUVMEnable,
-		double HostVMMinPageSize,
 		double PDEAndMetaPTEBytesPerFrame,
 		double MetaRowBytes,
-		double DPTEBytesPerRow,
-		double BandwidthAvailableForImmediateFlip,
-		unsigned int TotImmediateFlipBytes,
-		enum source_format_class SourcePixelFormat,
-		double LineTime,
-		double VRatio,
-		double VRatioChroma,
-		double Tno_bw,
-		bool DCCEnable,
-		unsigned int dpte_row_height,
-		unsigned int meta_row_height,
-		unsigned int dpte_row_height_chroma,
-		unsigned int meta_row_height_chroma,
-		double *DestinationLinesToRequestVMInImmediateFlip,
-		double *DestinationLinesToRequestRowInImmediateFlip,
-		double *final_flip_bw,
-		bool *ImmediateFlipSupportedForPipe)
+		double DPTEBytesPerRow)
 {
+	struct vba_vars_st *v = &mode_lib->vba;
 	double min_row_time = 0.0;
 	unsigned int HostVMDynamicLevelsTrips;
 	double TimeForFetchingMetaPTEImmediateFlip;
 	double TimeForFetchingRowInVBlankImmediateFlip;
-	double ImmediateFlipBW;
+	double ImmediateFlipBW = 1.0;
+	double LineTime = v->HTotal[k] / v->PixelClock[k];
 
-	if (GPUVMEnable == true && HostVMEnable == true) {
-		HostVMDynamicLevelsTrips = HostVMMaxNonCachedPageTableLevels;
+	if (v->GPUVMEnable == true && v->HostVMEnable == true) {
+		HostVMDynamicLevelsTrips = v->HostVMMaxNonCachedPageTableLevels;
 	} else {
 		HostVMDynamicLevelsTrips = 0;
 	}
 
-	if (GPUVMEnable == true || DCCEnable == true) {
-		ImmediateFlipBW = (PDEAndMetaPTEBytesPerFrame + MetaRowBytes + DPTEBytesPerRow) * BandwidthAvailableForImmediateFlip / TotImmediateFlipBytes;
+	if (v->GPUVMEnable == true || v->DCCEnable[k] == true) {
+		ImmediateFlipBW = (PDEAndMetaPTEBytesPerFrame + MetaRowBytes + DPTEBytesPerRow) * v->BandwidthAvailableForImmediateFlip / v->TotImmediateFlipBytes;
 	}
 
-	if (GPUVMEnable == true) {
+	if (v->GPUVMEnable == true) {
 		TimeForFetchingMetaPTEImmediateFlip = dml_max3(
-				Tno_bw + PDEAndMetaPTEBytesPerFrame * HostVMInefficiencyFactor / ImmediateFlipBW,
-				UrgentExtraLatency + UrgentLatency * (GPUVMMaxPageTableLevels * (HostVMDynamicLevelsTrips + 1) - 1),
+				v->Tno_bw[k] + PDEAndMetaPTEBytesPerFrame * HostVMInefficiencyFactor / ImmediateFlipBW,
+				UrgentExtraLatency + UrgentLatency * (v->GPUVMMaxPageTableLevels * (HostVMDynamicLevelsTrips + 1) - 1),
 				LineTime / 4.0);
 	} else {
 		TimeForFetchingMetaPTEImmediateFlip = 0;
 	}
 
-	*DestinationLinesToRequestVMInImmediateFlip = dml_ceil(4.0 * (TimeForFetchingMetaPTEImmediateFlip / LineTime), 1) / 4.0;
-	if ((GPUVMEnable == true || DCCEnable == true)) {
+	v->DestinationLinesToRequestVMInImmediateFlip[k] = dml_ceil(4.0 * (TimeForFetchingMetaPTEImmediateFlip / LineTime), 1) / 4.0;
+	if ((v->GPUVMEnable == true || v->DCCEnable[k] == true)) {
 		TimeForFetchingRowInVBlankImmediateFlip = dml_max3(
 				(MetaRowBytes + DPTEBytesPerRow * HostVMInefficiencyFactor) / ImmediateFlipBW,
 				UrgentLatency * (HostVMDynamicLevelsTrips + 1),
@@ -3804,54 +3526,54 @@ static void CalculateFlipSchedule(
 		TimeForFetchingRowInVBlankImmediateFlip = 0;
 	}
 
-	*DestinationLinesToRequestRowInImmediateFlip = dml_ceil(4.0 * (TimeForFetchingRowInVBlankImmediateFlip / LineTime), 1) / 4.0;
+	v->DestinationLinesToRequestRowInImmediateFlip[k] = dml_ceil(4.0 * (TimeForFetchingRowInVBlankImmediateFlip / LineTime), 1) / 4.0;
 
-	if (GPUVMEnable == true) {
-		*final_flip_bw = dml_max(
-				PDEAndMetaPTEBytesPerFrame * HostVMInefficiencyFactor / (*DestinationLinesToRequestVMInImmediateFlip * LineTime),
-				(MetaRowBytes + DPTEBytesPerRow * HostVMInefficiencyFactor) / (*DestinationLinesToRequestRowInImmediateFlip * LineTime));
-	} else if ((GPUVMEnable == true || DCCEnable == true)) {
-		*final_flip_bw = (MetaRowBytes + DPTEBytesPerRow * HostVMInefficiencyFactor) / (*DestinationLinesToRequestRowInImmediateFlip * LineTime);
+	if (v->GPUVMEnable == true) {
+		v->final_flip_bw[k] = dml_max(
+				PDEAndMetaPTEBytesPerFrame * HostVMInefficiencyFactor / (v->DestinationLinesToRequestVMInImmediateFlip[k] * LineTime),
+				(MetaRowBytes + DPTEBytesPerRow * HostVMInefficiencyFactor) / (v->DestinationLinesToRequestRowInImmediateFlip[k] * LineTime));
+	} else if ((v->GPUVMEnable == true || v->DCCEnable[k] == true)) {
+		v->final_flip_bw[k] = (MetaRowBytes + DPTEBytesPerRow * HostVMInefficiencyFactor) / (v->DestinationLinesToRequestRowInImmediateFlip[k] * LineTime);
 	} else {
-		*final_flip_bw = 0;
+		v->final_flip_bw[k] = 0;
 	}
 
-	if (SourcePixelFormat == dm_420_8 || SourcePixelFormat == dm_420_10 || SourcePixelFormat == dm_rgbe_alpha) {
-		if (GPUVMEnable == true && DCCEnable != true) {
-			min_row_time = dml_min(dpte_row_height * LineTime / VRatio, dpte_row_height_chroma * LineTime / VRatioChroma);
-		} else if (GPUVMEnable != true && DCCEnable == true) {
-			min_row_time = dml_min(meta_row_height * LineTime / VRatio, meta_row_height_chroma * LineTime / VRatioChroma);
+	if (v->SourcePixelFormat[k] == dm_420_8 || v->SourcePixelFormat[k] == dm_420_10 || v->SourcePixelFormat[k] == dm_rgbe_alpha) {
+		if (v->GPUVMEnable == true && v->DCCEnable[k] != true) {
+			min_row_time = dml_min(v->dpte_row_height[k] * LineTime / v->VRatio[k], v->dpte_row_height_chroma[k] * LineTime / v->VRatioChroma[k]);
+		} else if (v->GPUVMEnable != true && v->DCCEnable[k] == true) {
+			min_row_time = dml_min(v->meta_row_height[k] * LineTime / v->VRatio[k], v->meta_row_height_chroma[k] * LineTime / v->VRatioChroma[k]);
 		} else {
 			min_row_time = dml_min4(
-					dpte_row_height * LineTime / VRatio,
-					meta_row_height * LineTime / VRatio,
-					dpte_row_height_chroma * LineTime / VRatioChroma,
-					meta_row_height_chroma * LineTime / VRatioChroma);
+					v->dpte_row_height[k] * LineTime / v->VRatio[k],
+					v->meta_row_height[k] * LineTime / v->VRatio[k],
+					v->dpte_row_height_chroma[k] * LineTime / v->VRatioChroma[k],
+					v->meta_row_height_chroma[k] * LineTime / v->VRatioChroma[k]);
 		}
 	} else {
-		if (GPUVMEnable == true && DCCEnable != true) {
-			min_row_time = dpte_row_height * LineTime / VRatio;
-		} else if (GPUVMEnable != true && DCCEnable == true) {
-			min_row_time = meta_row_height * LineTime / VRatio;
+		if (v->GPUVMEnable == true && v->DCCEnable[k] != true) {
+			min_row_time = v->dpte_row_height[k] * LineTime / v->VRatio[k];
+		} else if (v->GPUVMEnable != true && v->DCCEnable[k] == true) {
+			min_row_time = v->meta_row_height[k] * LineTime / v->VRatio[k];
 		} else {
-			min_row_time = dml_min(dpte_row_height * LineTime / VRatio, meta_row_height * LineTime / VRatio);
+			min_row_time = dml_min(v->dpte_row_height[k] * LineTime / v->VRatio[k], v->meta_row_height[k] * LineTime / v->VRatio[k]);
 		}
 	}
 
-	if (*DestinationLinesToRequestVMInImmediateFlip >= 32 || *DestinationLinesToRequestRowInImmediateFlip >= 16
+	if (v->DestinationLinesToRequestVMInImmediateFlip[k] >= 32 || v->DestinationLinesToRequestRowInImmediateFlip[k] >= 16
 			|| TimeForFetchingMetaPTEImmediateFlip + 2 * TimeForFetchingRowInVBlankImmediateFlip > min_row_time) {
-		*ImmediateFlipSupportedForPipe = false;
+		v->ImmediateFlipSupportedForPipe[k] = false;
 	} else {
-		*ImmediateFlipSupportedForPipe = true;
+		v->ImmediateFlipSupportedForPipe[k] = true;
 	}
 
 #ifdef __DML_VBA_DEBUG__
-	dml_print("DML::%s: DestinationLinesToRequestVMInImmediateFlip = %f\n", __func__, *DestinationLinesToRequestVMInImmediateFlip);
-	dml_print("DML::%s: DestinationLinesToRequestRowInImmediateFlip = %f\n", __func__, *DestinationLinesToRequestRowInImmediateFlip);
+	dml_print("DML::%s: DestinationLinesToRequestVMInImmediateFlip = %f\n", __func__, v->DestinationLinesToRequestVMInImmediateFlip[k]);
+	dml_print("DML::%s: DestinationLinesToRequestRowInImmediateFlip = %f\n", __func__, v->DestinationLinesToRequestRowInImmediateFlip[k]);
 	dml_print("DML::%s: TimeForFetchingMetaPTEImmediateFlip = %f\n", __func__, TimeForFetchingMetaPTEImmediateFlip);
 	dml_print("DML::%s: TimeForFetchingRowInVBlankImmediateFlip = %f\n", __func__, TimeForFetchingRowInVBlankImmediateFlip);
 	dml_print("DML::%s: min_row_time = %f\n", __func__, min_row_time);
-	dml_print("DML::%s: ImmediateFlipSupportedForPipe = %d\n", __func__, *ImmediateFlipSupportedForPipe);
+	dml_print("DML::%s: ImmediateFlipSupportedForPipe = %d\n", __func__, v->ImmediateFlipSupportedForPipe[k]);
 #endif
 
 }
@@ -3884,7 +3606,7 @@ static double TruncToValidBPP(
 		NonDSCBPP1 = 15;
 		NonDSCBPP2 = 18;
 		MinDSCBPP = 6;
-		MaxDSCBPP = 1.5 * DSCInputBitPerComponent - 1 / 16;
+		MaxDSCBPP = 1.5 * DSCInputBitPerComponent - 1.0 / 16;
 	} else if (Format == dm_444) {
 		NonDSCBPP0 = 24;
 		NonDSCBPP1 = 30;
@@ -3946,9 +3668,120 @@ static double TruncToValidBPP(
 			return DesiredBPP;
 		}
 	}
-	return BPP_INVALID;
 }
 
+static noinline void CalculatePrefetchSchedulePerPlane(
+		struct display_mode_lib *mode_lib,
+		double HostVMInefficiencyFactor,
+		int i,
+		unsigned j,
+		unsigned k)
+{
+	struct vba_vars_st *v = &mode_lib->vba;
+	Pipe myPipe;
+
+	myPipe.DPPCLK = v->RequiredDPPCLK[i][j][k];
+	myPipe.DISPCLK = v->RequiredDISPCLK[i][j];
+	myPipe.PixelClock = v->PixelClock[k];
+	myPipe.DCFCLKDeepSleep = v->ProjectedDCFCLKDeepSleep[i][j];
+	myPipe.DPPPerPlane = v->NoOfDPP[i][j][k];
+	myPipe.ScalerEnabled = v->ScalerEnabled[k];
+	myPipe.VRatio = mode_lib->vba.VRatio[k];
+	myPipe.VRatioChroma = mode_lib->vba.VRatioChroma[k];
+
+	myPipe.SourceScan = v->SourceScan[k];
+	myPipe.BlockWidth256BytesY = v->Read256BlockWidthY[k];
+	myPipe.BlockHeight256BytesY = v->Read256BlockHeightY[k];
+	myPipe.BlockWidth256BytesC = v->Read256BlockWidthC[k];
+	myPipe.BlockHeight256BytesC = v->Read256BlockHeightC[k];
+	myPipe.InterlaceEnable = v->Interlace[k];
+	myPipe.NumberOfCursors = v->NumberOfCursors[k];
+	myPipe.VBlank = v->VTotal[k] - v->VActive[k];
+	myPipe.HTotal = v->HTotal[k];
+	myPipe.DCCEnable = v->DCCEnable[k];
+	myPipe.ODMCombineIsEnabled = v->ODMCombineEnablePerState[i][k] == dm_odm_combine_mode_4to1
+		|| v->ODMCombineEnablePerState[i][k] == dm_odm_combine_mode_2to1;
+	myPipe.SourcePixelFormat = v->SourcePixelFormat[k];
+	myPipe.BytePerPixelY = v->BytePerPixelY[k];
+	myPipe.BytePerPixelC = v->BytePerPixelC[k];
+	myPipe.ProgressiveToInterlaceUnitInOPP = v->ProgressiveToInterlaceUnitInOPP;
+	v->NoTimeForPrefetch[i][j][k] = CalculatePrefetchSchedule(
+		mode_lib,
+		HostVMInefficiencyFactor,
+		&myPipe,
+		v->DSCDelayPerState[i][k],
+		v->DPPCLKDelaySubtotal + v->DPPCLKDelayCNVCFormater,
+		v->DPPCLKDelaySCL,
+		v->DPPCLKDelaySCLLBOnly,
+		v->DPPCLKDelayCNVCCursor,
+		v->DISPCLKDelaySubtotal,
+		v->SwathWidthYThisState[k] / v->HRatio[k],
+		v->OutputFormat[k],
+		v->MaxInterDCNTileRepeaters,
+		dml_min(v->MaxVStartup, v->MaximumVStartup[i][j][k]),
+		v->MaximumVStartup[i][j][k],
+		v->GPUVMMaxPageTableLevels,
+		v->GPUVMEnable,
+		v->HostVMEnable,
+		v->HostVMMaxNonCachedPageTableLevels,
+		v->HostVMMinPageSize,
+		v->DynamicMetadataEnable[k],
+		v->DynamicMetadataVMEnabled,
+		v->DynamicMetadataLinesBeforeActiveRequired[k],
+		v->DynamicMetadataTransmittedBytes[k],
+		v->UrgLatency[i],
+		v->ExtraLatency,
+		v->TimeCalc,
+		v->PDEAndMetaPTEBytesPerFrame[i][j][k],
+		v->MetaRowBytes[i][j][k],
+		v->DPTEBytesPerRow[i][j][k],
+		v->PrefetchLinesY[i][j][k],
+		v->SwathWidthYThisState[k],
+		v->PrefillY[k],
+		v->MaxNumSwY[k],
+		v->PrefetchLinesC[i][j][k],
+		v->SwathWidthCThisState[k],
+		v->PrefillC[k],
+		v->MaxNumSwC[k],
+		v->swath_width_luma_ub_this_state[k],
+		v->swath_width_chroma_ub_this_state[k],
+		v->SwathHeightYThisState[k],
+		v->SwathHeightCThisState[k],
+		v->TWait,
+		&v->DSTXAfterScaler[k],
+		&v->DSTYAfterScaler[k],
+		&v->LineTimesForPrefetch[k],
+		&v->PrefetchBW[k],
+		&v->LinesForMetaPTE[k],
+		&v->LinesForMetaAndDPTERow[k],
+		&v->VRatioPreY[i][j][k],
+		&v->VRatioPreC[i][j][k],
+		&v->RequiredPrefetchPixelDataBWLuma[i][j][k],
+		&v->RequiredPrefetchPixelDataBWChroma[i][j][k],
+		&v->NoTimeForDynamicMetadata[i][j][k],
+		&v->Tno_bw[k],
+		&v->prefetch_vmrow_bw[k],
+		&v->dummy7[k],
+		&v->dummy8[k],
+		&v->dummy13[k],
+		&v->VUpdateOffsetPix[k],
+		&v->VUpdateWidthPix[k],
+		&v->VReadyOffsetPix[k]);
+}
+
+static void PatchDETBufferSizeInKByte(unsigned int NumberOfActivePlanes, int NoOfDPPThisState[], unsigned int config_return_buffer_size_in_kbytes, unsigned int DETBufferSizeInKByte[])
+{
+	int i, total_pipes = 0;
+	for (i = 0; i < NumberOfActivePlanes; i++)
+		total_pipes += NoOfDPPThisState[i];
+	DETBufferSizeInKByte[0] = ((config_return_buffer_size_in_kbytes - DCN3_15_MIN_COMPBUF_SIZE_KB) / 64 / total_pipes) * 64;
+	if (DETBufferSizeInKByte[0] > DCN3_15_MAX_DET_SIZE)
+		DETBufferSizeInKByte[0] = DCN3_15_MAX_DET_SIZE;
+	for (i = 1; i < NumberOfActivePlanes; i++)
+		DETBufferSizeInKByte[i] = DETBufferSizeInKByte[0];
+}
+
+
 void dml31_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_lib)
 {
 	struct vba_vars_st *v = &mode_lib->vba;
@@ -4012,7 +3845,7 @@ void dml31_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 	/*Bandwidth Support Check*/
 
 	for (k = 0; k < v->NumberOfActivePlanes; k++) {
-		CalculateBytePerPixelAnd256BBlockSizes(
+		dml30_CalculateBytePerPixelAnd256BBlockSizes(
 				v->SourcePixelFormat[k],
 				v->SurfaceTiling[k],
 				&v->BytePerPixelY[k],
@@ -4185,7 +4018,8 @@ void dml31_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 	CalculateSwathAndDETConfiguration(
 			true,
 			v->NumberOfActivePlanes,
-			v->DETBufferSizeInKByte[0],
+			mode_lib->project == DML_PROJECT_DCN315 && v->DETSizeOverride[0],
+			v->DETBufferSizeInKByte,
 			v->MaximumSwathWidthLuma,
 			v->MaximumSwathWidthChroma,
 			v->SourceScan,
@@ -4256,6 +4090,7 @@ void dml31_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 
 				if (v->ODMCombinePolicy == dm_odm_combine_policy_none
 						|| !(v->Output[k] == dm_dp ||
+						     v->Output[k] == dm_dp2p0 ||
 						     v->Output[k] == dm_edp)) {
 					v->ODMCombineEnablePerState[i][k] = dm_odm_combine_mode_disabled;
 					v->PlaneRequiredDISPCLK = v->PlaneRequiredDISPCLKWithoutODMCombine;
@@ -4288,7 +4123,9 @@ void dml31_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 				}
 				if (v->OutputFormat[k] == dm_420 && v->HActive[k] > DCN31_MAX_FMT_420_BUFFER_WIDTH
 						&& v->ODMCombineEnablePerState[i][k] != dm_odm_combine_mode_4to1) {
-					if (v->HActive[k] / 2 > DCN31_MAX_FMT_420_BUFFER_WIDTH) {
+					if (v->Output[k] == dm_hdmi) {
+						FMTBufferExceeded = true;
+					} else if (v->HActive[k] / 2 > DCN31_MAX_FMT_420_BUFFER_WIDTH) {
 						v->ODMCombineEnablePerState[i][k] = dm_odm_combine_mode_4to1;
 						v->PlaneRequiredDISPCLK = v->PlaneRequiredDISPCLKWithODMCombine4To1;
 
@@ -4324,6 +4161,10 @@ void dml31_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 						|| (v->PlaneRequiredDISPCLK > v->MaxDispclkRoundedDownToDFSGranularity)) {
 					v->DISPCLK_DPPCLK_Support[i][j] = false;
 				}
+				if (mode_lib->project == DML_PROJECT_DCN315 && v->DETSizeOverride[k] > DCN3_15_MAX_DET_SIZE && v->NoOfDPP[i][j][k] < 2) {
+					v->MPCCombine[i][j][k] = true;
+					v->NoOfDPP[i][j][k] = 2;
+				}
 			}
 			v->TotalNumberOfActiveDPP[i][j] = 0;
 			v->TotalNumberOfSingleDPPPlanes[i][j] = 0;
@@ -4465,11 +4306,11 @@ void dml31_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 							v->AudioSampleRate[k],
 							v->AudioSampleLayout[k],
 							v->ODMCombineEnablePerState[i][k]);
-				} else if (v->Output[k] == dm_dp || v->Output[k] == dm_edp) {
+				} else if (v->Output[k] == dm_dp || v->Output[k] == dm_edp || v->Output[k] == dm_dp2p0) {
 					if (v->DSCEnable[k] == true) {
 						v->RequiresDSC[i][k] = true;
 						v->LinkDSCEnable = true;
-						if (v->Output[k] == dm_dp) {
+						if (v->Output[k] == dm_dp || v->Output[k] == dm_dp2p0) {
 							v->RequiresFEC[i][k] = true;
 						} else {
 							v->RequiresFEC[i][k] = false;
@@ -4477,107 +4318,201 @@ void dml31_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 					} else {
 						v->RequiresDSC[i][k] = false;
 						v->LinkDSCEnable = false;
-						v->RequiresFEC[i][k] = false;
-					}
-
-					v->Outbpp = BPP_INVALID;
-					if (v->PHYCLKPerState[i] >= 270.0) {
-						v->Outbpp = TruncToValidBPP(
-								(1.0 - v->Downspreading / 100.0) * 2700,
-								v->OutputLinkDPLanes[k],
-								v->HTotal[k],
-								v->HActive[k],
-								v->PixelClockBackEnd[k],
-								v->ForcedOutputLinkBPP[k],
-								v->LinkDSCEnable,
-								v->Output[k],
-								v->OutputFormat[k],
-								v->DSCInputBitPerComponent[k],
-								v->NumberOfDSCSlices[k],
-								v->AudioSampleRate[k],
-								v->AudioSampleLayout[k],
-								v->ODMCombineEnablePerState[i][k]);
-						v->OutputBppPerState[i][k] = v->Outbpp;
-						// TODO: Need some other way to handle this nonsense
-						// v->OutputTypeAndRatePerState[i][k] = v->Output[k] & " HBR"
-					}
-					if (v->Outbpp == BPP_INVALID && v->PHYCLKPerState[i] >= 540.0) {
-						v->Outbpp = TruncToValidBPP(
-								(1.0 - v->Downspreading / 100.0) * 5400,
-								v->OutputLinkDPLanes[k],
-								v->HTotal[k],
-								v->HActive[k],
-								v->PixelClockBackEnd[k],
-								v->ForcedOutputLinkBPP[k],
-								v->LinkDSCEnable,
-								v->Output[k],
-								v->OutputFormat[k],
-								v->DSCInputBitPerComponent[k],
-								v->NumberOfDSCSlices[k],
-								v->AudioSampleRate[k],
-								v->AudioSampleLayout[k],
-								v->ODMCombineEnablePerState[i][k]);
-						v->OutputBppPerState[i][k] = v->Outbpp;
-						// TODO: Need some other way to handle this nonsense
-						// v->OutputTypeAndRatePerState[i][k] = v->Output[k] & " HBR2"
-					}
-					if (v->Outbpp == BPP_INVALID && v->PHYCLKPerState[i] >= 810.0) {
-						v->Outbpp = TruncToValidBPP(
-								(1.0 - v->Downspreading / 100.0) * 8100,
-								v->OutputLinkDPLanes[k],
-								v->HTotal[k],
-								v->HActive[k],
-								v->PixelClockBackEnd[k],
-								v->ForcedOutputLinkBPP[k],
-								v->LinkDSCEnable,
-								v->Output[k],
-								v->OutputFormat[k],
-								v->DSCInputBitPerComponent[k],
-								v->NumberOfDSCSlices[k],
-								v->AudioSampleRate[k],
-								v->AudioSampleLayout[k],
-								v->ODMCombineEnablePerState[i][k]);
-						v->OutputBppPerState[i][k] = v->Outbpp;
-						// TODO: Need some other way to handle this nonsense
-						// v->OutputTypeAndRatePerState[i][k] = v->Output[k] & " HBR3"
-					}
-					if (v->Outbpp == BPP_INVALID && v->PHYCLKD18PerState[i] >= 10000.0 / 18) {
-						v->Outbpp = TruncToValidBPP(
-								(1.0 - v->Downspreading / 100.0) * 10000,
-								4,
-								v->HTotal[k],
-								v->HActive[k],
-								v->PixelClockBackEnd[k],
-								v->ForcedOutputLinkBPP[k],
-								v->LinkDSCEnable,
-								v->Output[k],
-								v->OutputFormat[k],
-								v->DSCInputBitPerComponent[k],
-								v->NumberOfDSCSlices[k],
-								v->AudioSampleRate[k],
-								v->AudioSampleLayout[k],
-								v->ODMCombineEnablePerState[i][k]);
-						v->OutputBppPerState[i][k] = v->Outbpp;
-						//v->OutputTypeAndRatePerState[i][k] = v->Output[k] & "10x4";
+						if (v->Output[k] == dm_dp2p0) {
+							v->RequiresFEC[i][k] = true;
+						} else {
+							v->RequiresFEC[i][k] = false;
+						}
 					}
-					if (v->Outbpp == BPP_INVALID && v->PHYCLKD18PerState[i] >= 12000.0 / 18) {
-						v->Outbpp = TruncToValidBPP(
-								12000,
-								4,
-								v->HTotal[k],
-								v->HActive[k],
-								v->PixelClockBackEnd[k],
-								v->ForcedOutputLinkBPP[k],
-								v->LinkDSCEnable,
-								v->Output[k],
-								v->OutputFormat[k],
-								v->DSCInputBitPerComponent[k],
-								v->NumberOfDSCSlices[k],
-								v->AudioSampleRate[k],
-								v->AudioSampleLayout[k],
-								v->ODMCombineEnablePerState[i][k]);
-						v->OutputBppPerState[i][k] = v->Outbpp;
-						//v->OutputTypeAndRatePerState[i][k] = v->Output[k] & "12x4";
+					if (v->Output[k] == dm_dp2p0) {
+						v->Outbpp = BPP_INVALID;
+						if ((v->OutputLinkDPRate[k] == dm_dp_rate_na || v->OutputLinkDPRate[k] == dm_dp_rate_uhbr10) &&
+							v->PHYCLKD18PerState[k] >= 10000.0 / 18.0) {
+							v->Outbpp = TruncToValidBPP(
+									(1.0 - v->Downspreading / 100.0) * 10000,
+									v->OutputLinkDPLanes[k],
+									v->HTotal[k],
+									v->HActive[k],
+									v->PixelClockBackEnd[k],
+									v->ForcedOutputLinkBPP[k],
+									v->LinkDSCEnable,
+									v->Output[k],
+									v->OutputFormat[k],
+									v->DSCInputBitPerComponent[k],
+									v->NumberOfDSCSlices[k],
+									v->AudioSampleRate[k],
+									v->AudioSampleLayout[k],
+									v->ODMCombineEnablePerState[i][k]);
+							if (v->Outbpp == BPP_INVALID && v->PHYCLKD18PerState[k] < 13500.0 / 18.0 &&
+								v->DSCEnable[k] == true && v->ForcedOutputLinkBPP[k] == 0) {
+								v->RequiresDSC[i][k] = true;
+								v->LinkDSCEnable = true;
+								v->Outbpp = TruncToValidBPP(
+										(1.0 - v->Downspreading / 100.0) * 10000,
+										v->OutputLinkDPLanes[k],
+										v->HTotal[k],
+										v->HActive[k],
+										v->PixelClockBackEnd[k],
+										v->ForcedOutputLinkBPP[k],
+										v->LinkDSCEnable,
+										v->Output[k],
+										v->OutputFormat[k],
+										v->DSCInputBitPerComponent[k],
+										v->NumberOfDSCSlices[k],
+										v->AudioSampleRate[k],
+										v->AudioSampleLayout[k],
+										v->ODMCombineEnablePerState[i][k]);
+							}
+							v->OutputBppPerState[i][k] = v->Outbpp;
+							// TODO: Need some other way to handle this nonsense
+							// v->OutputTypeAndRatePerState[i][k] = v->Output[k] & " UHBR10"
+						}
+						if (v->Outbpp == BPP_INVALID &&
+							(v->OutputLinkDPRate[k] == dm_dp_rate_na || v->OutputLinkDPRate[k] == dm_dp_rate_uhbr13p5) &&
+							v->PHYCLKD18PerState[k] >= 13500.0 / 18.0) {
+							v->Outbpp = TruncToValidBPP(
+									(1.0 - v->Downspreading / 100.0) * 13500,
+									v->OutputLinkDPLanes[k],
+									v->HTotal[k],
+									v->HActive[k],
+									v->PixelClockBackEnd[k],
+									v->ForcedOutputLinkBPP[k],
+									v->LinkDSCEnable,
+									v->Output[k],
+									v->OutputFormat[k],
+									v->DSCInputBitPerComponent[k],
+									v->NumberOfDSCSlices[k],
+									v->AudioSampleRate[k],
+									v->AudioSampleLayout[k],
+									v->ODMCombineEnablePerState[i][k]);
+							if (v->Outbpp == BPP_INVALID && v->PHYCLKD18PerState[k] < 20000.0 / 18.0 &&
+								v->DSCEnable[k] == true && v->ForcedOutputLinkBPP[k] == 0) {
+								v->RequiresDSC[i][k] = true;
+								v->LinkDSCEnable = true;
+								v->Outbpp = TruncToValidBPP(
+										(1.0 - v->Downspreading / 100.0) * 13500,
+										v->OutputLinkDPLanes[k],
+										v->HTotal[k],
+										v->HActive[k],
+										v->PixelClockBackEnd[k],
+										v->ForcedOutputLinkBPP[k],
+										v->LinkDSCEnable,
+										v->Output[k],
+										v->OutputFormat[k],
+										v->DSCInputBitPerComponent[k],
+										v->NumberOfDSCSlices[k],
+										v->AudioSampleRate[k],
+										v->AudioSampleLayout[k],
+										v->ODMCombineEnablePerState[i][k]);
+							}
+							v->OutputBppPerState[i][k] = v->Outbpp;
+							// TODO: Need some other way to handle this nonsense
+							// v->OutputTypeAndRatePerState[i][k] = v->Output[k] & " UHBR13p5"
+						}
+						if (v->Outbpp == BPP_INVALID &&
+							(v->OutputLinkDPRate[k] == dm_dp_rate_na || v->OutputLinkDPRate[k] == dm_dp_rate_uhbr20) &&
+							v->PHYCLKD18PerState[k] >= 20000.0 / 18.0) {
+							v->Outbpp = TruncToValidBPP(
+									(1.0 - v->Downspreading / 100.0) * 20000,
+									v->OutputLinkDPLanes[k],
+									v->HTotal[k],
+									v->HActive[k],
+									v->PixelClockBackEnd[k],
+									v->ForcedOutputLinkBPP[k],
+									v->LinkDSCEnable,
+									v->Output[k],
+									v->OutputFormat[k],
+									v->DSCInputBitPerComponent[k],
+									v->NumberOfDSCSlices[k],
+									v->AudioSampleRate[k],
+									v->AudioSampleLayout[k],
+									v->ODMCombineEnablePerState[i][k]);
+							if (v->Outbpp == BPP_INVALID && v->DSCEnable[k] == true &&
+								v->ForcedOutputLinkBPP[k] == 0) {
+								v->RequiresDSC[i][k] = true;
+								v->LinkDSCEnable = true;
+								v->Outbpp = TruncToValidBPP(
+										(1.0 - v->Downspreading / 100.0) * 20000,
+										v->OutputLinkDPLanes[k],
+										v->HTotal[k],
+										v->HActive[k],
+										v->PixelClockBackEnd[k],
+										v->ForcedOutputLinkBPP[k],
+										v->LinkDSCEnable,
+										v->Output[k],
+										v->OutputFormat[k],
+										v->DSCInputBitPerComponent[k],
+										v->NumberOfDSCSlices[k],
+										v->AudioSampleRate[k],
+										v->AudioSampleLayout[k],
+										v->ODMCombineEnablePerState[i][k]);
+							}
+							v->OutputBppPerState[i][k] = v->Outbpp;
+							// TODO: Need some other way to handle this nonsense
+							// v->OutputTypeAndRatePerState[i][k] = v->Output[k] & " UHBR20"
+						}
+					} else {
+						v->Outbpp = BPP_INVALID;
+						if (v->PHYCLKPerState[i] >= 270.0) {
+							v->Outbpp = TruncToValidBPP(
+									(1.0 - v->Downspreading / 100.0) * 2700,
+									v->OutputLinkDPLanes[k],
+									v->HTotal[k],
+									v->HActive[k],
+									v->PixelClockBackEnd[k],
+									v->ForcedOutputLinkBPP[k],
+									v->LinkDSCEnable,
+									v->Output[k],
+									v->OutputFormat[k],
+									v->DSCInputBitPerComponent[k],
+									v->NumberOfDSCSlices[k],
+									v->AudioSampleRate[k],
+									v->AudioSampleLayout[k],
+									v->ODMCombineEnablePerState[i][k]);
+							v->OutputBppPerState[i][k] = v->Outbpp;
+							// TODO: Need some other way to handle this nonsense
+							// v->OutputTypeAndRatePerState[i][k] = v->Output[k] & " HBR"
+						}
+						if (v->Outbpp == BPP_INVALID && v->PHYCLKPerState[i] >= 540.0) {
+							v->Outbpp = TruncToValidBPP(
+									(1.0 - v->Downspreading / 100.0) * 5400,
+									v->OutputLinkDPLanes[k],
+									v->HTotal[k],
+									v->HActive[k],
+									v->PixelClockBackEnd[k],
+									v->ForcedOutputLinkBPP[k],
+									v->LinkDSCEnable,
+									v->Output[k],
+									v->OutputFormat[k],
+									v->DSCInputBitPerComponent[k],
+									v->NumberOfDSCSlices[k],
+									v->AudioSampleRate[k],
+									v->AudioSampleLayout[k],
+									v->ODMCombineEnablePerState[i][k]);
+							v->OutputBppPerState[i][k] = v->Outbpp;
+							// TODO: Need some other way to handle this nonsense
+							// v->OutputTypeAndRatePerState[i][k] = v->Output[k] & " HBR2"
+						}
+						if (v->Outbpp == BPP_INVALID && v->PHYCLKPerState[i] >= 810.0) {
+							v->Outbpp = TruncToValidBPP(
+									(1.0 - v->Downspreading / 100.0) * 8100,
+									v->OutputLinkDPLanes[k],
+									v->HTotal[k],
+									v->HActive[k],
+									v->PixelClockBackEnd[k],
+									v->ForcedOutputLinkBPP[k],
+									v->LinkDSCEnable,
+									v->Output[k],
+									v->OutputFormat[k],
+									v->DSCInputBitPerComponent[k],
+									v->NumberOfDSCSlices[k],
+									v->AudioSampleRate[k],
+									v->AudioSampleLayout[k],
+									v->ODMCombineEnablePerState[i][k]);
+							v->OutputBppPerState[i][k] = v->Outbpp;
+							// TODO: Need some other way to handle this nonsense
+							// v->OutputTypeAndRatePerState[i][k] = v->Output[k] & " HBR3"
+						}
 					}
 				}
 			} else {
@@ -4706,10 +4641,13 @@ void dml31_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 				v->ODMCombineEnableThisState[k] = v->ODMCombineEnablePerState[i][k];
 			}
 
+			if (v->NumberOfActivePlanes > 1 && mode_lib->project == DML_PROJECT_DCN315 && !v->DETSizeOverride[0])
+				PatchDETBufferSizeInKByte(v->NumberOfActivePlanes, v->NoOfDPPThisState, v->ip.config_return_buffer_size_in_kbytes, v->DETBufferSizeInKByte);
 			CalculateSwathAndDETConfiguration(
 					false,
 					v->NumberOfActivePlanes,
-					v->DETBufferSizeInKByte[0],
+					mode_lib->project == DML_PROJECT_DCN315 && v->DETSizeOverride[0],
+					v->DETBufferSizeInKByte,
 					v->MaximumSwathWidthLuma,
 					v->MaximumSwathWidthChroma,
 					v->SourceScan,
@@ -4945,6 +4883,17 @@ void dml31_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 						&v->meta_row_bandwidth[i][j][k],
 						&v->dpte_row_bandwidth[i][j][k]);
 			}
+			/*DCCMetaBufferSizeSupport(i, j) = True
+			For k = 0 To NumberOfActivePlanes - 1
+			If MetaRowBytes(i, j, k) > 24064 Then
+			DCCMetaBufferSizeSupport(i, j) = False
+			End If
+			Next k*/
+			v->DCCMetaBufferSizeSupport[i][j] = true;
+			for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+				if (v->MetaRowBytes[i][j][k] > 24064)
+					v->DCCMetaBufferSizeSupport[i][j] = false;
+			}
 			v->UrgLatency[i] = CalculateUrgentLatency(
 					v->UrgentLatencyPixelDataOnly,
 					v->UrgentLatencyPixelMixedWithVMData,
@@ -5079,66 +5028,8 @@ void dml31_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 		}
 	}
 
-	if (v->UseMinimumRequiredDCFCLK == true) {
-		UseMinimumDCFCLK(
-				mode_lib,
-				v->MaxInterDCNTileRepeaters,
-				MaxPrefetchMode,
-				v->DRAMClockChangeLatency,
-				v->SREnterPlusExitTime,
-				v->ReturnBusWidth,
-				v->RoundTripPingLatencyCycles,
-				ReorderingBytes,
-				v->PixelChunkSizeInKByte,
-				v->MetaChunkSize,
-				v->GPUVMEnable,
-				v->GPUVMMaxPageTableLevels,
-				v->HostVMEnable,
-				v->NumberOfActivePlanes,
-				v->HostVMMinPageSize,
-				v->HostVMMaxNonCachedPageTableLevels,
-				v->DynamicMetadataVMEnabled,
-				v->ImmediateFlipRequirement[0],
-				v->ProgressiveToInterlaceUnitInOPP,
-				v->MaxAveragePercentOfIdealFabricAndSDPPortBWDisplayCanUseInNormalSystemOperation,
-				v->PercentOfIdealFabricAndSDPPortBWReceivedAfterUrgLatency,
-				v->VTotal,
-				v->VActive,
-				v->DynamicMetadataTransmittedBytes,
-				v->DynamicMetadataLinesBeforeActiveRequired,
-				v->Interlace,
-				v->RequiredDPPCLK,
-				v->RequiredDISPCLK,
-				v->UrgLatency,
-				v->NoOfDPP,
-				v->ProjectedDCFCLKDeepSleep,
-				v->MaximumVStartup,
-				v->TotalVActivePixelBandwidth,
-				v->TotalVActiveCursorBandwidth,
-				v->TotalMetaRowBandwidth,
-				v->TotalDPTERowBandwidth,
-				v->TotalNumberOfActiveDPP,
-				v->TotalNumberOfDCCActiveDPP,
-				v->dpte_group_bytes,
-				v->PrefetchLinesY,
-				v->PrefetchLinesC,
-				v->swath_width_luma_ub_all_states,
-				v->swath_width_chroma_ub_all_states,
-				v->BytePerPixelY,
-				v->BytePerPixelC,
-				v->HTotal,
-				v->PixelClock,
-				v->PDEAndMetaPTEBytesPerFrame,
-				v->DPTEBytesPerRow,
-				v->MetaRowBytes,
-				v->DynamicMetadataEnable,
-				v->VActivePixelBandwidth,
-				v->VActiveCursorBandwidth,
-				v->ReadBandwidthLuma,
-				v->ReadBandwidthChroma,
-				v->DCFCLKPerState,
-				v->DCFCLKState);
-	}
+	if (v->UseMinimumRequiredDCFCLK == true)
+		UseMinimumDCFCLK(mode_lib, MaxPrefetchMode, ReorderingBytes);
 
 	for (i = 0; i < v->soc.num_states; ++i) {
 		for (j = 0; j <= 1; ++j) {
@@ -5276,92 +5167,9 @@ void dml31_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 						v->SREnterPlusExitTime);
 
 				for (k = 0; k < v->NumberOfActivePlanes; k++) {
-					Pipe myPipe;
-
-					myPipe.DPPCLK = v->RequiredDPPCLK[i][j][k];
-					myPipe.DISPCLK = v->RequiredDISPCLK[i][j];
-					myPipe.PixelClock = v->PixelClock[k];
-					myPipe.DCFCLKDeepSleep = v->ProjectedDCFCLKDeepSleep[i][j];
-					myPipe.DPPPerPlane = v->NoOfDPP[i][j][k];
-					myPipe.ScalerEnabled = v->ScalerEnabled[k];
-					myPipe.SourceScan = v->SourceScan[k];
-					myPipe.BlockWidth256BytesY = v->Read256BlockWidthY[k];
-					myPipe.BlockHeight256BytesY = v->Read256BlockHeightY[k];
-					myPipe.BlockWidth256BytesC = v->Read256BlockWidthC[k];
-					myPipe.BlockHeight256BytesC = v->Read256BlockHeightC[k];
-					myPipe.InterlaceEnable = v->Interlace[k];
-					myPipe.NumberOfCursors = v->NumberOfCursors[k];
-					myPipe.VBlank = v->VTotal[k] - v->VActive[k];
-					myPipe.HTotal = v->HTotal[k];
-					myPipe.DCCEnable = v->DCCEnable[k];
-					myPipe.ODMCombineIsEnabled = v->ODMCombineEnablePerState[i][k] == dm_odm_combine_mode_4to1
-							|| v->ODMCombineEnablePerState[i][k] == dm_odm_combine_mode_2to1;
-					myPipe.SourcePixelFormat = v->SourcePixelFormat[k];
-					myPipe.BytePerPixelY = v->BytePerPixelY[k];
-					myPipe.BytePerPixelC = v->BytePerPixelC[k];
-					myPipe.ProgressiveToInterlaceUnitInOPP = v->ProgressiveToInterlaceUnitInOPP;
-					v->NoTimeForPrefetch[i][j][k] = CalculatePrefetchSchedule(
-							mode_lib,
-							HostVMInefficiencyFactor,
-							&myPipe,
-							v->DSCDelayPerState[i][k],
-							v->DPPCLKDelaySubtotal + v->DPPCLKDelayCNVCFormater,
-							v->DPPCLKDelaySCL,
-							v->DPPCLKDelaySCLLBOnly,
-							v->DPPCLKDelayCNVCCursor,
-							v->DISPCLKDelaySubtotal,
-							v->SwathWidthYThisState[k] / v->HRatio[k],
-							v->OutputFormat[k],
-							v->MaxInterDCNTileRepeaters,
-							dml_min(v->MaxVStartup, v->MaximumVStartup[i][j][k]),
-							v->MaximumVStartup[i][j][k],
-							v->GPUVMMaxPageTableLevels,
-							v->GPUVMEnable,
-							v->HostVMEnable,
-							v->HostVMMaxNonCachedPageTableLevels,
-							v->HostVMMinPageSize,
-							v->DynamicMetadataEnable[k],
-							v->DynamicMetadataVMEnabled,
-							v->DynamicMetadataLinesBeforeActiveRequired[k],
-							v->DynamicMetadataTransmittedBytes[k],
-							v->UrgLatency[i],
-							v->ExtraLatency,
-							v->TimeCalc,
-							v->PDEAndMetaPTEBytesPerFrame[i][j][k],
-							v->MetaRowBytes[i][j][k],
-							v->DPTEBytesPerRow[i][j][k],
-							v->PrefetchLinesY[i][j][k],
-							v->SwathWidthYThisState[k],
-							v->PrefillY[k],
-							v->MaxNumSwY[k],
-							v->PrefetchLinesC[i][j][k],
-							v->SwathWidthCThisState[k],
-							v->PrefillC[k],
-							v->MaxNumSwC[k],
-							v->swath_width_luma_ub_this_state[k],
-							v->swath_width_chroma_ub_this_state[k],
-							v->SwathHeightYThisState[k],
-							v->SwathHeightCThisState[k],
-							v->TWait,
-							&v->DSTXAfterScaler[k],
-							&v->DSTYAfterScaler[k],
-							&v->LineTimesForPrefetch[k],
-							&v->PrefetchBW[k],
-							&v->LinesForMetaPTE[k],
-							&v->LinesForMetaAndDPTERow[k],
-							&v->VRatioPreY[i][j][k],
-							&v->VRatioPreC[i][j][k],
-							&v->RequiredPrefetchPixelDataBWLuma[i][j][k],
-							&v->RequiredPrefetchPixelDataBWChroma[i][j][k],
-							&v->NoTimeForDynamicMetadata[i][j][k],
-							&v->Tno_bw[k],
-							&v->prefetch_vmrow_bw[k],
-							&v->dummy7[k],
-							&v->dummy8[k],
-							&v->dummy13[k],
-							&v->VUpdateOffsetPix[k],
-							&v->VUpdateWidthPix[k],
-							&v->VReadyOffsetPix[k]);
+					CalculatePrefetchSchedulePerPlane(mode_lib,
+									  HostVMInefficiencyFactor,
+									  i, j,	k);
 				}
 
 				for (k = 0; k < v->NumberOfActivePlanes; k++) {
@@ -5371,7 +5179,7 @@ void dml31_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 							v->SwathHeightYThisState[k],
 							v->SwathHeightCThisState[k],
 							v->HTotal[k] / v->PixelClock[k],
-							v->UrgentLatency,
+							v->UrgLatency[i],
 							v->CursorBufferSize,
 							v->CursorWidth[k][0],
 							v->CursorBPP[k][0],
@@ -5383,7 +5191,7 @@ void dml31_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 							v->DETBufferSizeCThisState[k],
 							&v->UrgentBurstFactorCursorPre[k],
 							&v->UrgentBurstFactorLumaPre[k],
-							&v->UrgentBurstFactorChroma[k],
+							&v->UrgentBurstFactorChromaPre[k],
 							&v->NotUrgentLatencyHidingPre[k]);
 				}
 
@@ -5466,40 +5274,20 @@ void dml31_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 					v->TotImmediateFlipBytes = 0.0;
 					for (k = 0; k < v->NumberOfActivePlanes; k++) {
 						v->TotImmediateFlipBytes = v->TotImmediateFlipBytes
-								+ v->NoOfDPP[i][j][k] * v->PDEAndMetaPTEBytesPerFrame[i][j][k] + v->MetaRowBytes[i][j][k]
-								+ v->DPTEBytesPerRow[i][j][k];
+								+ v->NoOfDPP[i][j][k] * (v->PDEAndMetaPTEBytesPerFrame[i][j][k] + v->MetaRowBytes[i][j][k]
+								+ v->DPTEBytesPerRow[i][j][k]);
 					}
 
 					for (k = 0; k < v->NumberOfActivePlanes; k++) {
 						CalculateFlipSchedule(
 								mode_lib,
+								k,
 								HostVMInefficiencyFactor,
 								v->ExtraLatency,
 								v->UrgLatency[i],
-								v->GPUVMMaxPageTableLevels,
-								v->HostVMEnable,
-								v->HostVMMaxNonCachedPageTableLevels,
-								v->GPUVMEnable,
-								v->HostVMMinPageSize,
 								v->PDEAndMetaPTEBytesPerFrame[i][j][k],
 								v->MetaRowBytes[i][j][k],
-								v->DPTEBytesPerRow[i][j][k],
-								v->BandwidthAvailableForImmediateFlip,
-								v->TotImmediateFlipBytes,
-								v->SourcePixelFormat[k],
-								v->HTotal[k] / v->PixelClock[k],
-								v->VRatio[k],
-								v->VRatioChroma[k],
-								v->Tno_bw[k],
-								v->DCCEnable[k],
-								v->dpte_row_height[k],
-								v->meta_row_height[k],
-								v->dpte_row_height_chroma[k],
-								v->meta_row_height_chroma[k],
-								&v->DestinationLinesToRequestVMInImmediateFlip[k],
-								&v->DestinationLinesToRequestRowInImmediateFlip[k],
-								&v->final_flip_bw[k],
-								&v->ImmediateFlipSupportedForPipe[k]);
+								v->DPTEBytesPerRow[i][j][k]);
 					}
 					v->total_dcn_read_bw_with_flip = 0.0;
 					for (k = 0; k < v->NumberOfActivePlanes; k++) {
@@ -5557,64 +5345,28 @@ void dml31_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 			CalculateWatermarksAndDRAMSpeedChangeSupport(
 					mode_lib,
 					v->PrefetchModePerState[i][j],
-					v->NumberOfActivePlanes,
-					v->MaxLineBufferLines,
-					v->LineBufferSize,
-					v->WritebackInterfaceBufferSize,
 					v->DCFCLKState[i][j],
 					v->ReturnBWPerState[i][j],
-					v->SynchronizedVBlank,
-					v->dpte_group_bytes,
-					v->MetaChunkSize,
 					v->UrgLatency[i],
 					v->ExtraLatency,
-					v->WritebackLatency,
-					v->WritebackChunkSize,
 					v->SOCCLKPerState[i],
-					v->DRAMClockChangeLatency,
-					v->SRExitTime,
-					v->SREnterPlusExitTime,
-					v->SRExitZ8Time,
-					v->SREnterPlusExitZ8Time,
 					v->ProjectedDCFCLKDeepSleep[i][j],
 					v->DETBufferSizeYThisState,
 					v->DETBufferSizeCThisState,
 					v->SwathHeightYThisState,
 					v->SwathHeightCThisState,
-					v->LBBitPerPixel,
 					v->SwathWidthYThisState,
 					v->SwathWidthCThisState,
-					v->HRatio,
-					v->HRatioChroma,
-					v->vtaps,
-					v->VTAPsChroma,
-					v->VRatio,
-					v->VRatioChroma,
-					v->HTotal,
-					v->PixelClock,
-					v->BlendingAndTiming,
 					v->NoOfDPPThisState,
 					v->BytePerPixelInDETY,
 					v->BytePerPixelInDETC,
-					v->DSTXAfterScaler,
-					v->DSTYAfterScaler,
-					v->WritebackEnable,
-					v->WritebackPixelFormat,
-					v->WritebackDestinationWidth,
-					v->WritebackDestinationHeight,
-					v->WritebackSourceHeight,
 					UnboundedRequestEnabledThisState,
 					CompressedBufferSizeInkByteThisState,
 					&v->DRAMClockChangeSupport[i][j],
-					&v->UrgentWatermark,
-					&v->WritebackUrgentWatermark,
-					&v->DRAMClockChangeWatermark,
-					&v->WritebackDRAMClockChangeWatermark,
 					&dummy,
 					&dummy,
 					&dummy,
-					&dummy,
-					&v->MinActiveDRAMClockChangeLatencySupported);
+					&dummy);
 		}
 	}
 
@@ -5705,6 +5457,58 @@ void dml31_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 				v->ModeSupport[i][j] = true;
 			} else {
 				v->ModeSupport[i][j] = false;
+#ifdef __DML_VBA_DEBUG__
+				if (v->ScaleRatioAndTapsSupport == false)
+					dml_print("DML SUPPORT:     ScaleRatioAndTapsSupport failed");
+				if (v->SourceFormatPixelAndScanSupport == false)
+					dml_print("DML SUPPORT:     SourceFormatPixelAndScanSupport failed");
+				if (v->ViewportSizeSupport[i][j] == false)
+					dml_print("DML SUPPORT:     ViewportSizeSupport failed");
+				if (v->LinkCapacitySupport[i] == false)
+					dml_print("DML SUPPORT:     LinkCapacitySupport failed");
+				if (v->ODMCombine4To1SupportCheckOK[i] == false)
+					dml_print("DML SUPPORT:     DSC422NativeNotSupported failed");
+				if (v->NotEnoughDSCUnits[i] == true)
+					dml_print("DML SUPPORT:     NotEnoughDSCUnits");
+				if (v->DTBCLKRequiredMoreThanSupported[i] == true)
+					dml_print("DML SUPPORT:     DTBCLKRequiredMoreThanSupported");
+				if (v->ROBSupport[i][j] == false)
+					dml_print("DML SUPPORT:     ROBSupport failed");
+				if (v->DISPCLK_DPPCLK_Support[i][j] == false)
+					dml_print("DML SUPPORT:     DISPCLK_DPPCLK_Support failed");
+				if (v->TotalAvailablePipesSupport[i][j] == false)
+					dml_print("DML SUPPORT:     DSC422NativeNotSupported failed");
+				if (EnoughWritebackUnits == false)
+					dml_print("DML SUPPORT:     DSC422NativeNotSupported failed");
+				if (v->WritebackLatencySupport == false)
+					dml_print("DML SUPPORT:     WritebackLatencySupport failed");
+				if (v->WritebackScaleRatioAndTapsSupport == false)
+					dml_print("DML SUPPORT:     DSC422NativeNotSupported ");
+				if (v->CursorSupport == false)
+					dml_print("DML SUPPORT:     DSC422NativeNotSupported failed");
+				if (v->PitchSupport == false)
+					dml_print("DML SUPPORT:     PitchSupport failed");
+				if (ViewportExceedsSurface == true)
+					dml_print("DML SUPPORT:     ViewportExceedsSurface failed");
+				if (v->PrefetchSupported[i][j] == false)
+					dml_print("DML SUPPORT:     PrefetchSupported failed");
+				if (v->DynamicMetadataSupported[i][j] == false)
+					dml_print("DML SUPPORT:     DSC422NativeNotSupported failed");
+				if (v->TotalVerticalActiveBandwidthSupport[i][j] == false)
+					dml_print("DML SUPPORT:     TotalVerticalActiveBandwidthSupport failed");
+				if (v->VRatioInPrefetchSupported[i][j] == false)
+					dml_print("DML SUPPORT:     VRatioInPrefetchSupported failed");
+				if (v->PTEBufferSizeNotExceeded[i][j] == false)
+					dml_print("DML SUPPORT:     PTEBufferSizeNotExceeded failed");
+				if (v->NonsupportedDSCInputBPC == true)
+					dml_print("DML SUPPORT:     NonsupportedDSCInputBPC failed");
+				if (!((v->HostVMEnable == false
+					&& v->ImmediateFlipRequirement[0] != dm_immediate_flip_required)
+							|| v->ImmediateFlipSupportedForState[i][j] == true))
+					dml_print("DML SUPPORT:     ImmediateFlipRequirement failed");
+				if (FMTBufferExceeded == true)
+					dml_print("DML SUPPORT:     FMTBufferExceeded failed");
+#endif
 			}
 		}
 	}
@@ -5739,64 +5543,28 @@ void dml31_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l
 static void CalculateWatermarksAndDRAMSpeedChangeSupport(
 		struct display_mode_lib *mode_lib,
 		unsigned int PrefetchMode,
-		unsigned int NumberOfActivePlanes,
-		unsigned int MaxLineBufferLines,
-		unsigned int LineBufferSize,
-		unsigned int WritebackInterfaceBufferSize,
 		double DCFCLK,
 		double ReturnBW,
-		bool SynchronizedVBlank,
-		unsigned int dpte_group_bytes[],
-		unsigned int MetaChunkSize,
 		double UrgentLatency,
 		double ExtraLatency,
-		double WritebackLatency,
-		double WritebackChunkSize,
 		double SOCCLK,
-		double DRAMClockChangeLatency,
-		double SRExitTime,
-		double SREnterPlusExitTime,
-		double SRExitZ8Time,
-		double SREnterPlusExitZ8Time,
 		double DCFCLKDeepSleep,
 		unsigned int DETBufferSizeY[],
 		unsigned int DETBufferSizeC[],
 		unsigned int SwathHeightY[],
 		unsigned int SwathHeightC[],
-		unsigned int LBBitPerPixel[],
 		double SwathWidthY[],
 		double SwathWidthC[],
-		double HRatio[],
-		double HRatioChroma[],
-		unsigned int vtaps[],
-		unsigned int VTAPsChroma[],
-		double VRatio[],
-		double VRatioChroma[],
-		unsigned int HTotal[],
-		double PixelClock[],
-		unsigned int BlendingAndTiming[],
 		unsigned int DPPPerPlane[],
 		double BytePerPixelDETY[],
 		double BytePerPixelDETC[],
-		double DSTXAfterScaler[],
-		double DSTYAfterScaler[],
-		bool WritebackEnable[],
-		enum source_format_class WritebackPixelFormat[],
-		double WritebackDestinationWidth[],
-		double WritebackDestinationHeight[],
-		double WritebackSourceHeight[],
 		bool UnboundedRequestEnabled,
 		int unsigned CompressedBufferSizeInkByte,
 		enum clock_change_support *DRAMClockChangeSupport,
-		double *UrgentWatermark,
-		double *WritebackUrgentWatermark,
-		double *DRAMClockChangeWatermark,
-		double *WritebackDRAMClockChangeWatermark,
 		double *StutterExitWatermark,
 		double *StutterEnterPlusExitWatermark,
 		double *Z8StutterExitWatermark,
-		double *Z8StutterEnterPlusExitWatermark,
-		double *MinActiveDRAMClockChangeLatencySupported)
+		double *Z8StutterEnterPlusExitWatermark)
 {
 	struct vba_vars_st *v = &mode_lib->vba;
 	double EffectiveLBLatencyHidingY;
@@ -5816,103 +5584,103 @@ static void CalculateWatermarksAndDRAMSpeedChangeSupport(
 	double TotalPixelBW = 0.0;
 	int k, j;
 
-	*UrgentWatermark = UrgentLatency + ExtraLatency;
+	v->UrgentWatermark = UrgentLatency + ExtraLatency;
 
 #ifdef __DML_VBA_DEBUG__
 	dml_print("DML::%s: UrgentLatency = %f\n", __func__, UrgentLatency);
 	dml_print("DML::%s: ExtraLatency = %f\n", __func__, ExtraLatency);
-	dml_print("DML::%s: UrgentWatermark = %f\n", __func__, *UrgentWatermark);
+	dml_print("DML::%s: UrgentWatermark = %f\n", __func__, v->UrgentWatermark);
 #endif
 
-	*DRAMClockChangeWatermark = DRAMClockChangeLatency + *UrgentWatermark;
+	v->DRAMClockChangeWatermark = v->DRAMClockChangeLatency + v->UrgentWatermark;
 
 #ifdef __DML_VBA_DEBUG__
-	dml_print("DML::%s: DRAMClockChangeLatency = %f\n", __func__, DRAMClockChangeLatency);
-	dml_print("DML::%s: DRAMClockChangeWatermark = %f\n", __func__, *DRAMClockChangeWatermark);
+	dml_print("DML::%s: v->DRAMClockChangeLatency = %f\n", __func__, v->DRAMClockChangeLatency);
+	dml_print("DML::%s: DRAMClockChangeWatermark = %f\n", __func__, v->DRAMClockChangeWatermark);
 #endif
 
 	v->TotalActiveWriteback = 0;
-	for (k = 0; k < NumberOfActivePlanes; ++k) {
-		if (WritebackEnable[k] == true) {
+	for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+		if (v->WritebackEnable[k] == true) {
 			v->TotalActiveWriteback = v->TotalActiveWriteback + 1;
 		}
 	}
 
 	if (v->TotalActiveWriteback <= 1) {
-		*WritebackUrgentWatermark = WritebackLatency;
+		v->WritebackUrgentWatermark = v->WritebackLatency;
 	} else {
-		*WritebackUrgentWatermark = WritebackLatency + WritebackChunkSize * 1024.0 / 32.0 / SOCCLK;
+		v->WritebackUrgentWatermark = v->WritebackLatency + v->WritebackChunkSize * 1024.0 / 32.0 / SOCCLK;
 	}
 
 	if (v->TotalActiveWriteback <= 1) {
-		*WritebackDRAMClockChangeWatermark = DRAMClockChangeLatency + WritebackLatency;
+		v->WritebackDRAMClockChangeWatermark = v->DRAMClockChangeLatency + v->WritebackLatency;
 	} else {
-		*WritebackDRAMClockChangeWatermark = DRAMClockChangeLatency + WritebackLatency + WritebackChunkSize * 1024.0 / 32.0 / SOCCLK;
+		v->WritebackDRAMClockChangeWatermark = v->DRAMClockChangeLatency + v->WritebackLatency + v->WritebackChunkSize * 1024.0 / 32.0 / SOCCLK;
 	}
 
-	for (k = 0; k < NumberOfActivePlanes; ++k) {
+	for (k = 0; k < v->NumberOfActivePlanes; ++k) {
 		TotalPixelBW = TotalPixelBW
-				+ DPPPerPlane[k] * (SwathWidthY[k] * BytePerPixelDETY[k] * VRatio[k] + SwathWidthC[k] * BytePerPixelDETC[k] * VRatioChroma[k])
-						/ (HTotal[k] / PixelClock[k]);
+				+ DPPPerPlane[k] * (SwathWidthY[k] * BytePerPixelDETY[k] * v->VRatio[k] + SwathWidthC[k] * BytePerPixelDETC[k] * v->VRatioChroma[k])
+						/ (v->HTotal[k] / v->PixelClock[k]);
 	}
 
-	for (k = 0; k < NumberOfActivePlanes; ++k) {
+	for (k = 0; k < v->NumberOfActivePlanes; ++k) {
 		double EffectiveDETBufferSizeY = DETBufferSizeY[k];
 
 		v->LBLatencyHidingSourceLinesY = dml_min(
-				(double) MaxLineBufferLines,
-				dml_floor(LineBufferSize / LBBitPerPixel[k] / (SwathWidthY[k] / dml_max(HRatio[k], 1.0)), 1)) - (vtaps[k] - 1);
+				(double) v->MaxLineBufferLines,
+				dml_floor(v->LineBufferSize / v->LBBitPerPixel[k] / (SwathWidthY[k] / dml_max(v->HRatio[k], 1.0)), 1)) - (v->vtaps[k] - 1);
 
 		v->LBLatencyHidingSourceLinesC = dml_min(
-				(double) MaxLineBufferLines,
-				dml_floor(LineBufferSize / LBBitPerPixel[k] / (SwathWidthC[k] / dml_max(HRatioChroma[k], 1.0)), 1)) - (VTAPsChroma[k] - 1);
+				(double) v->MaxLineBufferLines,
+				dml_floor(v->LineBufferSize / v->LBBitPerPixel[k] / (SwathWidthC[k] / dml_max(v->HRatioChroma[k], 1.0)), 1)) - (v->VTAPsChroma[k] - 1);
 
-		EffectiveLBLatencyHidingY = v->LBLatencyHidingSourceLinesY / VRatio[k] * (HTotal[k] / PixelClock[k]);
+		EffectiveLBLatencyHidingY = v->LBLatencyHidingSourceLinesY / v->VRatio[k] * (v->HTotal[k] / v->PixelClock[k]);
 
-		EffectiveLBLatencyHidingC = v->LBLatencyHidingSourceLinesC / VRatioChroma[k] * (HTotal[k] / PixelClock[k]);
+		EffectiveLBLatencyHidingC = v->LBLatencyHidingSourceLinesC / v->VRatioChroma[k] * (v->HTotal[k] / v->PixelClock[k]);
 
 		if (UnboundedRequestEnabled) {
 			EffectiveDETBufferSizeY = EffectiveDETBufferSizeY
-					+ CompressedBufferSizeInkByte * 1024 * SwathWidthY[k] * BytePerPixelDETY[k] * VRatio[k] / (HTotal[k] / PixelClock[k]) / TotalPixelBW;
+					+ CompressedBufferSizeInkByte * 1024 * SwathWidthY[k] * BytePerPixelDETY[k] * v->VRatio[k] / (v->HTotal[k] / v->PixelClock[k]) / TotalPixelBW;
 		}
 
 		LinesInDETY[k] = (double) EffectiveDETBufferSizeY / BytePerPixelDETY[k] / SwathWidthY[k];
 		LinesInDETYRoundedDownToSwath[k] = dml_floor(LinesInDETY[k], SwathHeightY[k]);
-		FullDETBufferingTimeY = LinesInDETYRoundedDownToSwath[k] * (HTotal[k] / PixelClock[k]) / VRatio[k];
+		FullDETBufferingTimeY = LinesInDETYRoundedDownToSwath[k] * (v->HTotal[k] / v->PixelClock[k]) / v->VRatio[k];
 		if (BytePerPixelDETC[k] > 0) {
 			LinesInDETC = v->DETBufferSizeC[k] / BytePerPixelDETC[k] / SwathWidthC[k];
 			LinesInDETCRoundedDownToSwath = dml_floor(LinesInDETC, SwathHeightC[k]);
-			FullDETBufferingTimeC = LinesInDETCRoundedDownToSwath * (HTotal[k] / PixelClock[k]) / VRatioChroma[k];
+			FullDETBufferingTimeC = LinesInDETCRoundedDownToSwath * (v->HTotal[k] / v->PixelClock[k]) / v->VRatioChroma[k];
 		} else {
 			LinesInDETC = 0;
 			FullDETBufferingTimeC = 999999;
 		}
 
 		ActiveDRAMClockChangeLatencyMarginY = EffectiveLBLatencyHidingY + FullDETBufferingTimeY
-				- ((double) DSTXAfterScaler[k] / HTotal[k] + DSTYAfterScaler[k]) * HTotal[k] / PixelClock[k] - *UrgentWatermark - *DRAMClockChangeWatermark;
+				- ((double) v->DSTXAfterScaler[k] / v->HTotal[k] + v->DSTYAfterScaler[k]) * v->HTotal[k] / v->PixelClock[k] - v->UrgentWatermark - v->DRAMClockChangeWatermark;
 
-		if (NumberOfActivePlanes > 1) {
+		if (v->NumberOfActivePlanes > 1) {
 			ActiveDRAMClockChangeLatencyMarginY = ActiveDRAMClockChangeLatencyMarginY
-					- (1 - 1.0 / NumberOfActivePlanes) * SwathHeightY[k] * HTotal[k] / PixelClock[k] / VRatio[k];
+					- (1 - 1.0 / v->NumberOfActivePlanes) * SwathHeightY[k] * v->HTotal[k] / v->PixelClock[k] / v->VRatio[k];
 		}
 
 		if (BytePerPixelDETC[k] > 0) {
 			ActiveDRAMClockChangeLatencyMarginC = EffectiveLBLatencyHidingC + FullDETBufferingTimeC
-					- ((double) DSTXAfterScaler[k] / HTotal[k] + DSTYAfterScaler[k]) * HTotal[k] / PixelClock[k] - *UrgentWatermark - *DRAMClockChangeWatermark;
+					- ((double) v->DSTXAfterScaler[k] / v->HTotal[k] + v->DSTYAfterScaler[k]) * v->HTotal[k] / v->PixelClock[k] - v->UrgentWatermark - v->DRAMClockChangeWatermark;
 
-			if (NumberOfActivePlanes > 1) {
+			if (v->NumberOfActivePlanes > 1) {
 				ActiveDRAMClockChangeLatencyMarginC = ActiveDRAMClockChangeLatencyMarginC
-						- (1 - 1.0 / NumberOfActivePlanes) * SwathHeightC[k] * HTotal[k] / PixelClock[k] / VRatioChroma[k];
+						- (1 - 1.0 / v->NumberOfActivePlanes) * SwathHeightC[k] * v->HTotal[k] / v->PixelClock[k] / v->VRatioChroma[k];
 			}
 			v->ActiveDRAMClockChangeLatencyMargin[k] = dml_min(ActiveDRAMClockChangeLatencyMarginY, ActiveDRAMClockChangeLatencyMarginC);
 		} else {
 			v->ActiveDRAMClockChangeLatencyMargin[k] = ActiveDRAMClockChangeLatencyMarginY;
 		}
 
-		if (WritebackEnable[k] == true) {
-			WritebackDRAMClockChangeLatencyHiding = WritebackInterfaceBufferSize * 1024
-					/ (WritebackDestinationWidth[k] * WritebackDestinationHeight[k] / (WritebackSourceHeight[k] * HTotal[k] / PixelClock[k]) * 4);
-			if (WritebackPixelFormat[k] == dm_444_64) {
+		if (v->WritebackEnable[k] == true) {
+			WritebackDRAMClockChangeLatencyHiding = v->WritebackInterfaceBufferSize * 1024
+					/ (v->WritebackDestinationWidth[k] * v->WritebackDestinationHeight[k] / (v->WritebackSourceHeight[k] * v->HTotal[k] / v->PixelClock[k]) * 4);
+			if (v->WritebackPixelFormat[k] == dm_444_64) {
 				WritebackDRAMClockChangeLatencyHiding = WritebackDRAMClockChangeLatencyHiding / 2;
 			}
 			WritebackDRAMClockChangeLatencyMargin = WritebackDRAMClockChangeLatencyHiding - v->WritebackDRAMClockChangeWatermark;
@@ -5922,14 +5690,14 @@ static void CalculateWatermarksAndDRAMSpeedChangeSupport(
 
 	v->MinActiveDRAMClockChangeMargin = 999999;
 	PlaneWithMinActiveDRAMClockChangeMargin = 0;
-	for (k = 0; k < NumberOfActivePlanes; ++k) {
+	for (k = 0; k < v->NumberOfActivePlanes; ++k) {
 		if (v->ActiveDRAMClockChangeLatencyMargin[k] < v->MinActiveDRAMClockChangeMargin) {
 			v->MinActiveDRAMClockChangeMargin = v->ActiveDRAMClockChangeLatencyMargin[k];
-			if (BlendingAndTiming[k] == k) {
+			if (v->BlendingAndTiming[k] == k) {
 				PlaneWithMinActiveDRAMClockChangeMargin = k;
 			} else {
-				for (j = 0; j < NumberOfActivePlanes; ++j) {
-					if (BlendingAndTiming[k] == j) {
+				for (j = 0; j < v->NumberOfActivePlanes; ++j) {
+					if (v->BlendingAndTiming[k] == j) {
 						PlaneWithMinActiveDRAMClockChangeMargin = j;
 					}
 				}
@@ -5937,11 +5705,11 @@ static void CalculateWatermarksAndDRAMSpeedChangeSupport(
 		}
 	}
 
-	*MinActiveDRAMClockChangeLatencySupported = v->MinActiveDRAMClockChangeMargin + DRAMClockChangeLatency;
+	v->MinActiveDRAMClockChangeLatencySupported = v->MinActiveDRAMClockChangeMargin + v->DRAMClockChangeLatency ;
 
 	SecondMinActiveDRAMClockChangeMarginOneDisplayInVBLank = 999999;
-	for (k = 0; k < NumberOfActivePlanes; ++k) {
-		if (!((k == PlaneWithMinActiveDRAMClockChangeMargin) && (BlendingAndTiming[k] == k)) && !(BlendingAndTiming[k] == PlaneWithMinActiveDRAMClockChangeMargin)
+	for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+		if (!((k == PlaneWithMinActiveDRAMClockChangeMargin) && (v->BlendingAndTiming[k] == k)) && !(v->BlendingAndTiming[k] == PlaneWithMinActiveDRAMClockChangeMargin)
 				&& v->ActiveDRAMClockChangeLatencyMargin[k] < SecondMinActiveDRAMClockChangeMarginOneDisplayInVBLank) {
 			SecondMinActiveDRAMClockChangeMarginOneDisplayInVBLank = v->ActiveDRAMClockChangeLatencyMargin[k];
 		}
@@ -5949,25 +5717,25 @@ static void CalculateWatermarksAndDRAMSpeedChangeSupport(
 
 	v->TotalNumberOfActiveOTG = 0;
 
-	for (k = 0; k < NumberOfActivePlanes; ++k) {
-		if (BlendingAndTiming[k] == k) {
+	for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+		if (v->BlendingAndTiming[k] == k) {
 			v->TotalNumberOfActiveOTG = v->TotalNumberOfActiveOTG + 1;
 		}
 	}
 
 	if (v->MinActiveDRAMClockChangeMargin > 0 && PrefetchMode == 0) {
 		*DRAMClockChangeSupport = dm_dram_clock_change_vactive;
-	} else if ((SynchronizedVBlank == true || v->TotalNumberOfActiveOTG == 1
+	} else if ((v->SynchronizedVBlank == true || v->TotalNumberOfActiveOTG == 1
 			|| SecondMinActiveDRAMClockChangeMarginOneDisplayInVBLank > 0) && PrefetchMode == 0) {
 		*DRAMClockChangeSupport = dm_dram_clock_change_vblank;
 	} else {
 		*DRAMClockChangeSupport = dm_dram_clock_change_unsupported;
 	}
 
-	*StutterExitWatermark = SRExitTime + ExtraLatency + 10 / DCFCLKDeepSleep;
-	*StutterEnterPlusExitWatermark = (SREnterPlusExitTime + ExtraLatency + 10 / DCFCLKDeepSleep);
-	*Z8StutterExitWatermark = SRExitZ8Time + ExtraLatency + 10 / DCFCLKDeepSleep;
-	*Z8StutterEnterPlusExitWatermark = SREnterPlusExitZ8Time + ExtraLatency + 10 / DCFCLKDeepSleep;
+	*StutterExitWatermark = v->SRExitTime + ExtraLatency + 10 / DCFCLKDeepSleep;
+	*StutterEnterPlusExitWatermark = (v->SREnterPlusExitTime + ExtraLatency + 10 / DCFCLKDeepSleep);
+	*Z8StutterExitWatermark = v->SRExitZ8Time + ExtraLatency + 10 / DCFCLKDeepSleep;
+	*Z8StutterEnterPlusExitWatermark = v->SREnterPlusExitZ8Time + ExtraLatency + 10 / DCFCLKDeepSleep;
 
 #ifdef __DML_VBA_DEBUG__
 	dml_print("DML::%s: StutterExitWatermark = %f\n", __func__, *StutterExitWatermark);
@@ -6843,7 +6611,8 @@ static void CalculateStutterEfficiency(
 static void CalculateSwathAndDETConfiguration(
 		bool ForceSingleDPP,
 		int NumberOfActivePlanes,
-		unsigned int DETBufferSizeInKByte,
+		bool DETSharedByAllDPP,
+		unsigned int DETBufferSizeInKByteA[],
 		double MaximumSwathWidthLuma[],
 		double MaximumSwathWidthChroma[],
 		enum scan_direction_class SourceScan[],
@@ -6927,6 +6696,10 @@ static void CalculateSwathAndDETConfiguration(
 
 	*ViewportSizeSupport = true;
 	for (k = 0; k < NumberOfActivePlanes; ++k) {
+		unsigned int DETBufferSizeInKByte = DETBufferSizeInKByteA[k];
+
+		if (DETSharedByAllDPP && DPPPerPlane[k])
+			DETBufferSizeInKByte /= DPPPerPlane[k];
 		if ((SourcePixelFormat[k] == dm_444_64 || SourcePixelFormat[k] == dm_444_32 || SourcePixelFormat[k] == dm_444_16 || SourcePixelFormat[k] == dm_mono_16
 				|| SourcePixelFormat[k] == dm_mono_8 || SourcePixelFormat[k] == dm_rgbe)) {
 			if (SurfaceTiling[k] == dm_sw_linear
@@ -7106,8 +6879,6 @@ static void CalculateSwathWidth(
 		{
 		int surface_width_ub_l = dml_ceil(SurfaceWidthY[k], Read256BytesBlockWidthY[k]);
 		int surface_height_ub_l = dml_ceil(SurfaceHeightY[k], Read256BytesBlockHeightY[k]);
-		int surface_width_ub_c = dml_ceil(SurfaceWidthC[k], Read256BytesBlockWidthC[k]);
-		int surface_height_ub_c = dml_ceil(SurfaceHeightC[k], Read256BytesBlockHeightC[k]);
 
 #ifdef __DML_VBA_DEBUG__
 		dml_print("DML::%s: k=%d surface_width_ub_l=%0d\n", __func__, k, surface_width_ub_l);
@@ -7118,6 +6889,8 @@ static void CalculateSwathWidth(
 			MaximumSwathHeightC[k] = Read256BytesBlockHeightC[k];
 			swath_width_luma_ub[k] = dml_min(surface_width_ub_l, (int) dml_ceil(SwathWidthY[k] - 1, Read256BytesBlockWidthY[k]) + Read256BytesBlockWidthY[k]);
 			if (BytePerPixC[k] > 0) {
+				int surface_width_ub_c = dml_ceil(SurfaceWidthC[k], Read256BytesBlockWidthC[k]);
+
 				swath_width_chroma_ub[k] = dml_min(
 						surface_width_ub_c,
 						(int) dml_ceil(SwathWidthC[k] - 1, Read256BytesBlockWidthC[k]) + Read256BytesBlockWidthC[k]);
@@ -7129,6 +6902,8 @@ static void CalculateSwathWidth(
 			MaximumSwathHeightC[k] = Read256BytesBlockWidthC[k];
 			swath_width_luma_ub[k] = dml_min(surface_height_ub_l, (int) dml_ceil(SwathWidthY[k] - 1, Read256BytesBlockHeightY[k]) + Read256BytesBlockHeightY[k]);
 			if (BytePerPixC[k] > 0) {
+				int surface_height_ub_c = dml_ceil(SurfaceHeightC[k], Read256BytesBlockHeightC[k]);
+
 				swath_width_chroma_ub[k] = dml_min(
 						surface_height_ub_c,
 						(int) dml_ceil(SwathWidthC[k] - 1, Read256BytesBlockHeightC[k]) + Read256BytesBlockHeightC[k]);
@@ -7247,71 +7022,17 @@ static double CalculateUrgentLatency(
 	return ret;
 }
 
-static void UseMinimumDCFCLK(
+static noinline_for_stack void UseMinimumDCFCLK(
 		struct display_mode_lib *mode_lib,
-		int MaxInterDCNTileRepeaters,
 		int MaxPrefetchMode,
-		double FinalDRAMClockChangeLatency,
-		double SREnterPlusExitTime,
-		int ReturnBusWidth,
-		int RoundTripPingLatencyCycles,
-		int ReorderingBytes,
-		int PixelChunkSizeInKByte,
-		int MetaChunkSize,
-		bool GPUVMEnable,
-		int GPUVMMaxPageTableLevels,
-		bool HostVMEnable,
-		int NumberOfActivePlanes,
-		double HostVMMinPageSize,
-		int HostVMMaxNonCachedPageTableLevels,
-		bool DynamicMetadataVMEnabled,
-		enum immediate_flip_requirement ImmediateFlipRequirement,
-		bool ProgressiveToInterlaceUnitInOPP,
-		double MaxAveragePercentOfIdealFabricAndSDPPortBWDisplayCanUseInNormalSystemOperation,
-		double PercentOfIdealFabricAndSDPPortBWReceivedAfterUrgLatency,
-		int VTotal[],
-		int VActive[],
-		int DynamicMetadataTransmittedBytes[],
-		int DynamicMetadataLinesBeforeActiveRequired[],
-		bool Interlace[],
-		double RequiredDPPCLK[][2][DC__NUM_DPP__MAX],
-		double RequiredDISPCLK[][2],
-		double UrgLatency[],
-		unsigned int NoOfDPP[][2][DC__NUM_DPP__MAX],
-		double ProjectedDCFCLKDeepSleep[][2],
-		double MaximumVStartup[][2][DC__NUM_DPP__MAX],
-		double TotalVActivePixelBandwidth[][2],
-		double TotalVActiveCursorBandwidth[][2],
-		double TotalMetaRowBandwidth[][2],
-		double TotalDPTERowBandwidth[][2],
-		unsigned int TotalNumberOfActiveDPP[][2],
-		unsigned int TotalNumberOfDCCActiveDPP[][2],
-		int dpte_group_bytes[],
-		double PrefetchLinesY[][2][DC__NUM_DPP__MAX],
-		double PrefetchLinesC[][2][DC__NUM_DPP__MAX],
-		int swath_width_luma_ub_all_states[][2][DC__NUM_DPP__MAX],
-		int swath_width_chroma_ub_all_states[][2][DC__NUM_DPP__MAX],
-		int BytePerPixelY[],
-		int BytePerPixelC[],
-		int HTotal[],
-		double PixelClock[],
-		double PDEAndMetaPTEBytesPerFrame[][2][DC__NUM_DPP__MAX],
-		double DPTEBytesPerRow[][2][DC__NUM_DPP__MAX],
-		double MetaRowBytes[][2][DC__NUM_DPP__MAX],
-		bool DynamicMetadataEnable[],
-		double VActivePixelBandwidth[][2][DC__NUM_DPP__MAX],
-		double VActiveCursorBandwidth[][2][DC__NUM_DPP__MAX],
-		double ReadBandwidthLuma[],
-		double ReadBandwidthChroma[],
-		double DCFCLKPerState[],
-		double DCFCLKState[][2])
+		int ReorderingBytes)
 {
 	struct vba_vars_st *v = &mode_lib->vba;
 	int dummy1, i, j, k;
 	double NormalEfficiency,  dummy2, dummy3;
 	double TotalMaxPrefetchFlipDPTERowBandwidth[DC__VOLTAGE_STATES][2];
 
-	NormalEfficiency = PercentOfIdealFabricAndSDPPortBWReceivedAfterUrgLatency / 100.0;
+	NormalEfficiency = v->PercentOfIdealFabricAndSDPPortBWReceivedAfterUrgLatency / 100.0;
 	for (i = 0; i < v->soc.num_states; ++i) {
 		for (j = 0; j <= 1; ++j) {
 			double PixelDCFCLKCyclesRequiredInPrefetch[DC__NUM_DPP__MAX];
@@ -7329,61 +7050,61 @@ static void UseMinimumDCFCLK(
 			double MinimumTvmPlus2Tr0;
 
 			TotalMaxPrefetchFlipDPTERowBandwidth[i][j] = 0;
-			for (k = 0; k < NumberOfActivePlanes; ++k) {
+			for (k = 0; k < v->NumberOfActivePlanes; ++k) {
 				TotalMaxPrefetchFlipDPTERowBandwidth[i][j] = TotalMaxPrefetchFlipDPTERowBandwidth[i][j]
-						+ NoOfDPP[i][j][k] * DPTEBytesPerRow[i][j][k] / (15.75 * HTotal[k] / PixelClock[k]);
+						+ v->NoOfDPP[i][j][k] * v->DPTEBytesPerRow[i][j][k] / (15.75 * v->HTotal[k] / v->PixelClock[k]);
 			}
 
-			for (k = 0; k <= NumberOfActivePlanes - 1; ++k) {
-				NoOfDPPState[k] = NoOfDPP[i][j][k];
+			for (k = 0; k <= v->NumberOfActivePlanes - 1; ++k) {
+				NoOfDPPState[k] = v->NoOfDPP[i][j][k];
 			}
 
-			MinimumTWait = CalculateTWait(MaxPrefetchMode, FinalDRAMClockChangeLatency, UrgLatency[i], SREnterPlusExitTime);
-			NonDPTEBandwidth = TotalVActivePixelBandwidth[i][j] + TotalVActiveCursorBandwidth[i][j] + TotalMetaRowBandwidth[i][j];
-			DPTEBandwidth = (HostVMEnable == true || ImmediateFlipRequirement == dm_immediate_flip_required) ?
-					TotalMaxPrefetchFlipDPTERowBandwidth[i][j] : TotalDPTERowBandwidth[i][j];
+			MinimumTWait = CalculateTWait(MaxPrefetchMode, v->FinalDRAMClockChangeLatency, v->UrgLatency[i], v->SREnterPlusExitTime);
+			NonDPTEBandwidth = v->TotalVActivePixelBandwidth[i][j] + v->TotalVActiveCursorBandwidth[i][j] + v->TotalMetaRowBandwidth[i][j];
+			DPTEBandwidth = (v->HostVMEnable == true || v->ImmediateFlipRequirement[0] == dm_immediate_flip_required) ?
+					TotalMaxPrefetchFlipDPTERowBandwidth[i][j] : v->TotalDPTERowBandwidth[i][j];
 			DCFCLKRequiredForAverageBandwidth = dml_max3(
-					ProjectedDCFCLKDeepSleep[i][j],
-					(NonDPTEBandwidth + TotalDPTERowBandwidth[i][j]) / ReturnBusWidth
-							/ (MaxAveragePercentOfIdealFabricAndSDPPortBWDisplayCanUseInNormalSystemOperation / 100),
-					(NonDPTEBandwidth + DPTEBandwidth / NormalEfficiency) / NormalEfficiency / ReturnBusWidth);
+					v->ProjectedDCFCLKDeepSleep[i][j],
+					(NonDPTEBandwidth + v->TotalDPTERowBandwidth[i][j]) / v->ReturnBusWidth
+							/ (v->MaxAveragePercentOfIdealFabricAndSDPPortBWDisplayCanUseInNormalSystemOperation / 100),
+					(NonDPTEBandwidth + DPTEBandwidth / NormalEfficiency) / NormalEfficiency / v->ReturnBusWidth);
 
 			ExtraLatencyBytes = CalculateExtraLatencyBytes(
 					ReorderingBytes,
-					TotalNumberOfActiveDPP[i][j],
-					PixelChunkSizeInKByte,
-					TotalNumberOfDCCActiveDPP[i][j],
-					MetaChunkSize,
-					GPUVMEnable,
-					HostVMEnable,
-					NumberOfActivePlanes,
+					v->TotalNumberOfActiveDPP[i][j],
+					v->PixelChunkSizeInKByte,
+					v->TotalNumberOfDCCActiveDPP[i][j],
+					v->MetaChunkSize,
+					v->GPUVMEnable,
+					v->HostVMEnable,
+					v->NumberOfActivePlanes,
 					NoOfDPPState,
-					dpte_group_bytes,
+					v->dpte_group_bytes,
 					1,
-					HostVMMinPageSize,
-					HostVMMaxNonCachedPageTableLevels);
-			ExtraLatencyCycles = RoundTripPingLatencyCycles + __DML_ARB_TO_RET_DELAY__ + ExtraLatencyBytes / NormalEfficiency / ReturnBusWidth;
-			for (k = 0; k < NumberOfActivePlanes; ++k) {
+					v->HostVMMinPageSize,
+					v->HostVMMaxNonCachedPageTableLevels);
+			ExtraLatencyCycles = v->RoundTripPingLatencyCycles + __DML_ARB_TO_RET_DELAY__ + ExtraLatencyBytes / NormalEfficiency / v->ReturnBusWidth;
+			for (k = 0; k < v->NumberOfActivePlanes; ++k) {
 				double DCFCLKCyclesRequiredInPrefetch;
 				double ExpectedPrefetchBWAcceleration;
 				double PrefetchTime;
 
-				PixelDCFCLKCyclesRequiredInPrefetch[k] = (PrefetchLinesY[i][j][k] * swath_width_luma_ub_all_states[i][j][k] * BytePerPixelY[k]
-						+ PrefetchLinesC[i][j][k] * swath_width_chroma_ub_all_states[i][j][k] * BytePerPixelC[k]) / NormalEfficiency / ReturnBusWidth;
+				PixelDCFCLKCyclesRequiredInPrefetch[k] = (v->PrefetchLinesY[i][j][k] * v->swath_width_luma_ub_all_states[i][j][k] * v->BytePerPixelY[k]
+						+ v->PrefetchLinesC[i][j][k] * v->swath_width_chroma_ub_all_states[i][j][k] * v->BytePerPixelC[k]) / NormalEfficiency / v->ReturnBusWidth;
 				DCFCLKCyclesRequiredInPrefetch = 2 * ExtraLatencyCycles / NoOfDPPState[k]
-						+ PDEAndMetaPTEBytesPerFrame[i][j][k] / NormalEfficiency / NormalEfficiency / ReturnBusWidth * (GPUVMMaxPageTableLevels > 2 ? 1 : 0)
-						+ 2 * DPTEBytesPerRow[i][j][k] / NormalEfficiency / NormalEfficiency / ReturnBusWidth
-						+ 2 * MetaRowBytes[i][j][k] / NormalEfficiency / ReturnBusWidth + PixelDCFCLKCyclesRequiredInPrefetch[k];
-				PrefetchPixelLinesTime[k] = dml_max(PrefetchLinesY[i][j][k], PrefetchLinesC[i][j][k]) * HTotal[k] / PixelClock[k];
-				ExpectedPrefetchBWAcceleration = (VActivePixelBandwidth[i][j][k] + VActiveCursorBandwidth[i][j][k])
-						/ (ReadBandwidthLuma[k] + ReadBandwidthChroma[k]);
+						+ v->PDEAndMetaPTEBytesPerFrame[i][j][k] / NormalEfficiency / NormalEfficiency / v->ReturnBusWidth * (v->GPUVMMaxPageTableLevels > 2 ? 1 : 0)
+						+ 2 * v->DPTEBytesPerRow[i][j][k] / NormalEfficiency / NormalEfficiency / v->ReturnBusWidth
+						+ 2 * v->MetaRowBytes[i][j][k] / NormalEfficiency / v->ReturnBusWidth + PixelDCFCLKCyclesRequiredInPrefetch[k];
+				PrefetchPixelLinesTime[k] = dml_max(v->PrefetchLinesY[i][j][k], v->PrefetchLinesC[i][j][k]) * v->HTotal[k] / v->PixelClock[k];
+				ExpectedPrefetchBWAcceleration = (v->VActivePixelBandwidth[i][j][k] + v->VActiveCursorBandwidth[i][j][k])
+						/ (v->ReadBandwidthLuma[k] + v->ReadBandwidthChroma[k]);
 				DynamicMetadataVMExtraLatency[k] =
-						(GPUVMEnable == true && DynamicMetadataEnable[k] == true && DynamicMetadataVMEnabled == true) ?
-								UrgLatency[i] * GPUVMMaxPageTableLevels * (HostVMEnable == true ? HostVMMaxNonCachedPageTableLevels + 1 : 1) : 0;
-				PrefetchTime = (MaximumVStartup[i][j][k] - 1) * HTotal[k] / PixelClock[k] - MinimumTWait
-						- UrgLatency[i]
-								* ((GPUVMMaxPageTableLevels <= 2 ? GPUVMMaxPageTableLevels : GPUVMMaxPageTableLevels - 2)
-										* (HostVMEnable == true ? HostVMMaxNonCachedPageTableLevels + 1 : 1) - 1)
+						(v->GPUVMEnable == true && v->DynamicMetadataEnable[k] == true && v->DynamicMetadataVMEnabled == true) ?
+								v->UrgLatency[i] * v->GPUVMMaxPageTableLevels * (v->HostVMEnable == true ? v->HostVMMaxNonCachedPageTableLevels + 1 : 1) : 0;
+				PrefetchTime = (v->MaximumVStartup[i][j][k] - 1) * v->HTotal[k] / v->PixelClock[k] - MinimumTWait
+						- v->UrgLatency[i]
+								* ((v->GPUVMMaxPageTableLevels <= 2 ? v->GPUVMMaxPageTableLevels : v->GPUVMMaxPageTableLevels - 2)
+										* (v->HostVMEnable == true ? v->HostVMMaxNonCachedPageTableLevels + 1 : 1) - 1)
 						- DynamicMetadataVMExtraLatency[k];
 
 				if (PrefetchTime > 0) {
@@ -7392,14 +7113,14 @@ static void UseMinimumDCFCLK(
 							/ (PrefetchTime * PixelDCFCLKCyclesRequiredInPrefetch[k] / DCFCLKCyclesRequiredInPrefetch);
 					DCFCLKRequiredForPeakBandwidthPerPlane[k] = NoOfDPPState[k] * PixelDCFCLKCyclesRequiredInPrefetch[k] / PrefetchPixelLinesTime[k]
 							* dml_max(1.0, ExpectedVRatioPrefetch) * dml_max(1.0, ExpectedVRatioPrefetch / 4) * ExpectedPrefetchBWAcceleration;
-					if (HostVMEnable == true || ImmediateFlipRequirement == dm_immediate_flip_required) {
+					if (v->HostVMEnable == true || v->ImmediateFlipRequirement[0] == dm_immediate_flip_required) {
 						DCFCLKRequiredForPeakBandwidthPerPlane[k] = DCFCLKRequiredForPeakBandwidthPerPlane[k]
-								+ NoOfDPPState[k] * DPTEBandwidth / NormalEfficiency / NormalEfficiency / ReturnBusWidth;
+								+ NoOfDPPState[k] * DPTEBandwidth / NormalEfficiency / NormalEfficiency / v->ReturnBusWidth;
 					}
 				} else {
-					DCFCLKRequiredForPeakBandwidthPerPlane[k] = DCFCLKPerState[i];
+					DCFCLKRequiredForPeakBandwidthPerPlane[k] = v->DCFCLKPerState[i];
 				}
-				if (DynamicMetadataEnable[k] == true) {
+				if (v->DynamicMetadataEnable[k] == true) {
 					double TSetupPipe;
 					double TdmbfPipe;
 					double TdmsksPipe;
@@ -7407,17 +7128,17 @@ static void UseMinimumDCFCLK(
 					double AllowedTimeForUrgentExtraLatency;
 
 					CalculateVupdateAndDynamicMetadataParameters(
-							MaxInterDCNTileRepeaters,
-							RequiredDPPCLK[i][j][k],
-							RequiredDISPCLK[i][j],
-							ProjectedDCFCLKDeepSleep[i][j],
-							PixelClock[k],
-							HTotal[k],
-							VTotal[k] - VActive[k],
-							DynamicMetadataTransmittedBytes[k],
-							DynamicMetadataLinesBeforeActiveRequired[k],
-							Interlace[k],
-							ProgressiveToInterlaceUnitInOPP,
+							v->MaxInterDCNTileRepeaters,
+							v->RequiredDPPCLK[i][j][k],
+							v->RequiredDISPCLK[i][j],
+							v->ProjectedDCFCLKDeepSleep[i][j],
+							v->PixelClock[k],
+							v->HTotal[k],
+							v->VTotal[k] - v->VActive[k],
+							v->DynamicMetadataTransmittedBytes[k],
+							v->DynamicMetadataLinesBeforeActiveRequired[k],
+							v->Interlace[k],
+							v->ProgressiveToInterlaceUnitInOPP,
 							&TSetupPipe,
 							&TdmbfPipe,
 							&TdmecPipe,
@@ -7425,31 +7146,31 @@ static void UseMinimumDCFCLK(
 							&dummy1,
 							&dummy2,
 							&dummy3);
-					AllowedTimeForUrgentExtraLatency = MaximumVStartup[i][j][k] * HTotal[k] / PixelClock[k] - MinimumTWait - TSetupPipe - TdmbfPipe - TdmecPipe
+					AllowedTimeForUrgentExtraLatency = v->MaximumVStartup[i][j][k] * v->HTotal[k] / v->PixelClock[k] - MinimumTWait - TSetupPipe - TdmbfPipe - TdmecPipe
 							- TdmsksPipe - DynamicMetadataVMExtraLatency[k];
 					if (AllowedTimeForUrgentExtraLatency > 0) {
 						DCFCLKRequiredForPeakBandwidthPerPlane[k] = dml_max(
 								DCFCLKRequiredForPeakBandwidthPerPlane[k],
 								ExtraLatencyCycles / AllowedTimeForUrgentExtraLatency);
 					} else {
-						DCFCLKRequiredForPeakBandwidthPerPlane[k] = DCFCLKPerState[i];
+						DCFCLKRequiredForPeakBandwidthPerPlane[k] = v->DCFCLKPerState[i];
 					}
 				}
 			}
 			DCFCLKRequiredForPeakBandwidth = 0;
-			for (k = 0; k <= NumberOfActivePlanes - 1; ++k) {
+			for (k = 0; k <= v->NumberOfActivePlanes - 1; ++k) {
 				DCFCLKRequiredForPeakBandwidth = DCFCLKRequiredForPeakBandwidth + DCFCLKRequiredForPeakBandwidthPerPlane[k];
 			}
-			MinimumTvmPlus2Tr0 = UrgLatency[i]
-					* (GPUVMEnable == true ?
-							(HostVMEnable == true ?
-									(GPUVMMaxPageTableLevels + 2) * (HostVMMaxNonCachedPageTableLevels + 1) - 1 : GPUVMMaxPageTableLevels + 1) :
+			MinimumTvmPlus2Tr0 = v->UrgLatency[i]
+					* (v->GPUVMEnable == true ?
+							(v->HostVMEnable == true ?
+									(v->GPUVMMaxPageTableLevels + 2) * (v->HostVMMaxNonCachedPageTableLevels + 1) - 1 : v->GPUVMMaxPageTableLevels + 1) :
 							0);
-			for (k = 0; k < NumberOfActivePlanes; ++k) {
+			for (k = 0; k < v->NumberOfActivePlanes; ++k) {
 				double MaximumTvmPlus2Tr0PlusTsw;
-				MaximumTvmPlus2Tr0PlusTsw = (MaximumVStartup[i][j][k] - 2) * HTotal[k] / PixelClock[k] - MinimumTWait - DynamicMetadataVMExtraLatency[k];
+				MaximumTvmPlus2Tr0PlusTsw = (v->MaximumVStartup[i][j][k] - 2) * v->HTotal[k] / v->PixelClock[k] - MinimumTWait - DynamicMetadataVMExtraLatency[k];
 				if (MaximumTvmPlus2Tr0PlusTsw <= MinimumTvmPlus2Tr0 + PrefetchPixelLinesTime[k] / 4) {
-					DCFCLKRequiredForPeakBandwidth = DCFCLKPerState[i];
+					DCFCLKRequiredForPeakBandwidth = v->DCFCLKPerState[i];
 				} else {
 					DCFCLKRequiredForPeakBandwidth = dml_max3(
 							DCFCLKRequiredForPeakBandwidth,
@@ -7457,7 +7178,7 @@ static void UseMinimumDCFCLK(
 							(2 * ExtraLatencyCycles + PixelDCFCLKCyclesRequiredInPrefetch[k]) / (MaximumTvmPlus2Tr0PlusTsw - MinimumTvmPlus2Tr0));
 				}
 			}
-			DCFCLKState[i][j] = dml_min(DCFCLKPerState[i], 1.05 * dml_max(DCFCLKRequiredForAverageBandwidth, DCFCLKRequiredForPeakBandwidth));
+			v->DCFCLKState[i][j] = dml_min(v->DCFCLKPerState[i], 1.05 * dml_max(DCFCLKRequiredForAverageBandwidth, DCFCLKRequiredForPeakBandwidth));
 		}
 	}
 }
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_rq_dlg_calc_31.c b/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_rq_dlg_calc_31.c
index e0fecf127bd5..bfeb01477f0c 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_rq_dlg_calc_31.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_rq_dlg_calc_31.c
@@ -27,94 +27,7 @@
 #include "../display_mode_vba.h"
 #include "../dml_inline_defs.h"
 #include "display_rq_dlg_calc_31.h"
-
-static bool CalculateBytePerPixelAnd256BBlockSizes(
-		enum source_format_class SourcePixelFormat,
-		enum dm_swizzle_mode SurfaceTiling,
-		unsigned int *BytePerPixelY,
-		unsigned int *BytePerPixelC,
-		double *BytePerPixelDETY,
-		double *BytePerPixelDETC,
-		unsigned int *BlockHeight256BytesY,
-		unsigned int *BlockHeight256BytesC,
-		unsigned int *BlockWidth256BytesY,
-		unsigned int *BlockWidth256BytesC)
-{
-	if (SourcePixelFormat == dm_444_64) {
-		*BytePerPixelDETY = 8;
-		*BytePerPixelDETC = 0;
-		*BytePerPixelY = 8;
-		*BytePerPixelC = 0;
-	} else if (SourcePixelFormat == dm_444_32 || SourcePixelFormat == dm_rgbe) {
-		*BytePerPixelDETY = 4;
-		*BytePerPixelDETC = 0;
-		*BytePerPixelY = 4;
-		*BytePerPixelC = 0;
-	} else if (SourcePixelFormat == dm_444_16) {
-		*BytePerPixelDETY = 2;
-		*BytePerPixelDETC = 0;
-		*BytePerPixelY = 2;
-		*BytePerPixelC = 0;
-	} else if (SourcePixelFormat == dm_444_8) {
-		*BytePerPixelDETY = 1;
-		*BytePerPixelDETC = 0;
-		*BytePerPixelY = 1;
-		*BytePerPixelC = 0;
-	} else if (SourcePixelFormat == dm_rgbe_alpha) {
-		*BytePerPixelDETY = 4;
-		*BytePerPixelDETC = 1;
-		*BytePerPixelY = 4;
-		*BytePerPixelC = 1;
-	} else if (SourcePixelFormat == dm_420_8) {
-		*BytePerPixelDETY = 1;
-		*BytePerPixelDETC = 2;
-		*BytePerPixelY = 1;
-		*BytePerPixelC = 2;
-	} else if (SourcePixelFormat == dm_420_12) {
-		*BytePerPixelDETY = 2;
-		*BytePerPixelDETC = 4;
-		*BytePerPixelY = 2;
-		*BytePerPixelC = 4;
-	} else {
-		*BytePerPixelDETY = 4.0 / 3;
-		*BytePerPixelDETC = 8.0 / 3;
-		*BytePerPixelY = 2;
-		*BytePerPixelC = 4;
-	}
-
-	if ((SourcePixelFormat == dm_444_64 || SourcePixelFormat == dm_444_32 || SourcePixelFormat == dm_444_16 || SourcePixelFormat == dm_444_8 || SourcePixelFormat == dm_mono_16
-			|| SourcePixelFormat == dm_mono_8 || SourcePixelFormat == dm_rgbe)) {
-		if (SurfaceTiling == dm_sw_linear) {
-			*BlockHeight256BytesY = 1;
-		} else if (SourcePixelFormat == dm_444_64) {
-			*BlockHeight256BytesY = 4;
-		} else if (SourcePixelFormat == dm_444_8) {
-			*BlockHeight256BytesY = 16;
-		} else {
-			*BlockHeight256BytesY = 8;
-		}
-		*BlockWidth256BytesY = 256U / *BytePerPixelY / *BlockHeight256BytesY;
-		*BlockHeight256BytesC = 0;
-		*BlockWidth256BytesC = 0;
-	} else {
-		if (SurfaceTiling == dm_sw_linear) {
-			*BlockHeight256BytesY = 1;
-			*BlockHeight256BytesC = 1;
-		} else if (SourcePixelFormat == dm_rgbe_alpha) {
-			*BlockHeight256BytesY = 8;
-			*BlockHeight256BytesC = 16;
-		} else if (SourcePixelFormat == dm_420_8) {
-			*BlockHeight256BytesY = 16;
-			*BlockHeight256BytesC = 8;
-		} else {
-			*BlockHeight256BytesY = 8;
-			*BlockHeight256BytesC = 8;
-		}
-		*BlockWidth256BytesY = 256U / *BytePerPixelY / *BlockHeight256BytesY;
-		*BlockWidth256BytesC = 256U / *BytePerPixelC / *BlockHeight256BytesC;
-	}
-	return true;
-}
+#include "../dcn30/display_mode_vba_30.h"
 
 static bool is_dual_plane(enum source_format_class source_format)
 {
@@ -467,7 +380,7 @@ static void get_meta_and_pte_attr(
 	double byte_per_pixel_det_y;
 	double byte_per_pixel_det_c;
 
-	CalculateBytePerPixelAnd256BBlockSizes(
+	dml30_CalculateBytePerPixelAnd256BBlockSizes(
 			(enum source_format_class) (source_format),
 			(enum dm_swizzle_mode) (tiling),
 			&bytes_per_element_y,
@@ -500,8 +413,6 @@ static void get_meta_and_pte_attr(
 	log2_blk256_height = dml_log2((double) blk256_height);
 	blk_bytes = surf_linear ? 256 : get_blk_size_bytes((enum source_macro_tile_size) macro_tile_size);
 	log2_blk_bytes = dml_log2((double) blk_bytes);
-	log2_blk_height = 0;
-	log2_blk_width = 0;
 
 	// remember log rule
 	// "+" in log is multiply
@@ -568,8 +479,6 @@ static void get_meta_and_pte_attr(
 	log2_meta_req_width = log2_meta_req_bytes + 8 - log2_bytes_per_element - log2_meta_req_height;
 	meta_req_width = 1 << log2_meta_req_width;
 	meta_req_height = 1 << log2_meta_req_height;
-	log2_meta_row_height = 0;
-	meta_row_width_ub = 0;
 
 	// the dimensions of a meta row are meta_row_width x meta_row_height in elements.
 	// calculate upper bound of the meta_row_width
@@ -706,7 +615,7 @@ static void get_meta_and_pte_attr(
 	if (hostvm_enable)
 		rq_sizing_param->dpte_group_bytes = 512;
 	else {
-		if (!surf_linear & (log2_dpte_req_height_ptes == 0) & surf_vert) //reduced, in this case, will have page fault within a group
+		if (!surf_linear && (log2_dpte_req_height_ptes == 0) && surf_vert) //reduced, in this case, will have page fault within a group
 			rq_sizing_param->dpte_group_bytes = 512;
 		else
 			rq_sizing_param->dpte_group_bytes = 2048;
@@ -742,13 +651,12 @@ static void get_surf_rq_param(
 		bool is_chroma,
 		bool is_alpha)
 {
-	bool mode_422 = 0;
 	unsigned int vp_width = 0;
 	unsigned int vp_height = 0;
 	unsigned int data_pitch = 0;
 	unsigned int meta_pitch = 0;
 	unsigned int surface_height = 0;
-	unsigned int ppe = mode_422 ? 2 : 1;
+	unsigned int ppe = 1;
 
 	// FIXME check if ppe apply for both luma and chroma in 422 case
 	if (is_chroma | is_alpha) {
@@ -953,7 +861,6 @@ static void dml_rq_dlg_get_dlg_params(
 {
 	const display_pipe_source_params_st *src = &e2e_pipe_param[pipe_idx].pipe.src;
 	const display_pipe_dest_params_st *dst = &e2e_pipe_param[pipe_idx].pipe.dest;
-	const display_output_params_st *dout = &e2e_pipe_param[pipe_idx].dout;
 	const display_clocks_and_cfg_st *clks = &e2e_pipe_param[pipe_idx].clks_cfg;
 	const scaler_ratio_depth_st *scl = &e2e_pipe_param[pipe_idx].pipe.scale_ratio_depth;
 	const scaler_taps_st *taps = &e2e_pipe_param[pipe_idx].pipe.scale_taps;
@@ -976,7 +883,6 @@ static void dml_rq_dlg_get_dlg_params(
 	double min_ttu_vblank;
 	unsigned int dlg_vblank_start;
 	bool dual_plane;
-	bool mode_422;
 	unsigned int access_dir;
 	unsigned int vp_height_l;
 	unsigned int vp_width_l;
@@ -990,7 +896,6 @@ static void dml_rq_dlg_get_dlg_params(
 	double hratio_c;
 	double vratio_l;
 	double vratio_c;
-	bool scl_enable;
 
 	unsigned int swath_width_ub_l;
 	unsigned int dpte_groups_per_row_ub_l;
@@ -1003,9 +908,6 @@ static void dml_rq_dlg_get_dlg_params(
 	unsigned int vupdate_width;
 	unsigned int vready_offset;
 
-	unsigned int dppclk_delay_subtotal;
-	unsigned int dispclk_delay_subtotal;
-
 	unsigned int vstartup_start;
 	unsigned int dst_x_after_scaler;
 	unsigned int dst_y_after_scaler;
@@ -1078,8 +980,8 @@ static void dml_rq_dlg_get_dlg_params(
 	min_ttu_vblank = get_min_ttu_vblank_in_us(mode_lib, e2e_pipe_param, num_pipes, pipe_idx);	// From VBA
 
 	dlg_vblank_start = interlaced ? (vblank_start / 2) : vblank_start;
-
 	disp_dlg_regs->min_dst_y_next_start = (unsigned int) (((double) dlg_vblank_start) * dml_pow(2, 2));
+	disp_dlg_regs->min_dst_y_next_start_us = 0;
 	ASSERT(disp_dlg_regs->min_dst_y_next_start < (unsigned int)dml_pow(2, 18));
 
 	dml_print("DML_DLG: %s: min_ttu_vblank (us)         = %3.2f\n", __func__, min_ttu_vblank);
@@ -1096,7 +998,6 @@ static void dml_rq_dlg_get_dlg_params(
 	// Prefetch Calc
 	// Source
 	dual_plane = is_dual_plane((enum source_format_class) (src->source_format));
-	mode_422 = 0;
 	access_dir = (src->source_scan == dm_vert);	// vp access direction: horizontal or vertical accessed
 	vp_height_l = src->viewport_height;
 	vp_width_l = src->viewport_width;
@@ -1110,7 +1011,6 @@ static void dml_rq_dlg_get_dlg_params(
 	hratio_c = scl->hscl_ratio_c;
 	vratio_l = scl->vscl_ratio;
 	vratio_c = scl->vscl_ratio_c;
-	scl_enable = scl->scl_enable;
 
 	swath_width_ub_l = rq_dlg_param->rq_l.swath_width_ub;
 	dpte_groups_per_row_ub_l = rq_dlg_param->rq_l.dpte_groups_per_row_ub;
@@ -1123,21 +1023,6 @@ static void dml_rq_dlg_get_dlg_params(
 	vupdate_width = dst->vupdate_width;
 	vready_offset = dst->vready_offset;
 
-	dppclk_delay_subtotal = mode_lib->ip.dppclk_delay_subtotal;
-	dispclk_delay_subtotal = mode_lib->ip.dispclk_delay_subtotal;
-
-	if (scl_enable)
-		dppclk_delay_subtotal += mode_lib->ip.dppclk_delay_scl;
-	else
-		dppclk_delay_subtotal += mode_lib->ip.dppclk_delay_scl_lb_only;
-
-	dppclk_delay_subtotal += mode_lib->ip.dppclk_delay_cnvc_formatter + src->num_cursors * mode_lib->ip.dppclk_delay_cnvc_cursor;
-
-	if (dout->dsc_enable) {
-		double dsc_delay = get_dsc_delay(mode_lib, e2e_pipe_param, num_pipes, pipe_idx); // FROM VBA
-		dispclk_delay_subtotal += dsc_delay;
-	}
-
 	vstartup_start = dst->vstartup_start;
 	if (interlaced) {
 		if (vstartup_start / 2.0 - (double) (vready_offset + vupdate_width + vupdate_offset) / htotal <= vblank_end / 2.0)
@@ -1250,18 +1135,8 @@ static void dml_rq_dlg_get_dlg_params(
 	dpte_row_height_l = rq_dlg_param->rq_l.dpte_row_height;
 	dpte_row_height_c = rq_dlg_param->rq_c.dpte_row_height;
 
-	if (mode_422) {
-		swath_width_pixels_ub_l = swath_width_ub_l * 2;  // *2 for 2 pixel per element
-		swath_width_pixels_ub_c = swath_width_ub_c * 2;
-	} else {
-		swath_width_pixels_ub_l = swath_width_ub_l * 1;
-		swath_width_pixels_ub_c = swath_width_ub_c * 1;
-	}
-
-	hscale_pixel_rate_l = 0.;
-	hscale_pixel_rate_c = 0.;
-	min_hratio_fact_l = 1.0;
-	min_hratio_fact_c = 1.0;
+	swath_width_pixels_ub_l = swath_width_ub_l;
+	swath_width_pixels_ub_c = swath_width_ub_c;
 
 	if (hratio_l <= 1)
 		min_hratio_fact_l = 2.0;
@@ -1540,14 +1415,6 @@ static void dml_rq_dlg_get_dlg_params(
 	dml_print("DML_DLG: %s: disp_dlg_regs->dst_y_per_vm_flip    = 0x%x\n", __func__, disp_dlg_regs->dst_y_per_vm_flip);
 	dml_print("DML_DLG: %s: disp_dlg_regs->dst_y_per_row_flip   = 0x%x\n", __func__, disp_dlg_regs->dst_y_per_row_flip);
 
-	// hack for FPGA
-	if (mode_lib->project == DML_PROJECT_DCN31_FPGA) {
-		if (disp_dlg_regs->vratio_prefetch >= (unsigned int) dml_pow(2, 22)) {
-			disp_dlg_regs->vratio_prefetch = (unsigned int) dml_pow(2, 22) - 1;
-			dml_print("vratio_prefetch exceed the max value, the register field is [21:0]\n");
-		}
-	}
-
 	disp_dlg_regs->refcyc_per_pte_group_vblank_l = (unsigned int) (dst_y_per_row_vblank * (double) htotal * ref_freq_to_pix_freq / (double) dpte_groups_per_row_ub_l);
 	ASSERT(disp_dlg_regs->refcyc_per_pte_group_vblank_l < (unsigned int)dml_pow(2, 13));
 
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn314/dcn314_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn314/dcn314_fpu.c
new file mode 100644
index 000000000000..df9d50b9b57c
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn314/dcn314_fpu.c
@@ -0,0 +1,432 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#include "clk_mgr.h"
+#include "resource.h"
+#include "dcn31/dcn31_hubbub.h"
+#include "dcn314_fpu.h"
+#include "dml/dcn20/dcn20_fpu.h"
+#include "dml/dcn31/dcn31_fpu.h"
+#include "dml/display_mode_vba.h"
+#include "dml/dml_inline_defs.h"
+
+struct _vcs_dpi_ip_params_st dcn3_14_ip = {
+	.VBlankNomDefaultUS = 668,
+	.gpuvm_enable = 1,
+	.gpuvm_max_page_table_levels = 1,
+	.hostvm_enable = 1,
+	.hostvm_max_page_table_levels = 2,
+	.rob_buffer_size_kbytes = 64,
+	.det_buffer_size_kbytes = DCN3_14_DEFAULT_DET_SIZE,
+	.config_return_buffer_size_in_kbytes = 1792,
+	.compressed_buffer_segment_size_in_kbytes = 64,
+	.meta_fifo_size_in_kentries = 32,
+	.zero_size_buffer_entries = 512,
+	.compbuf_reserved_space_64b = 256,
+	.compbuf_reserved_space_zs = 64,
+	.dpp_output_buffer_pixels = 2560,
+	.opp_output_buffer_lines = 1,
+	.pixel_chunk_size_kbytes = 8,
+	.meta_chunk_size_kbytes = 2,
+	.min_meta_chunk_size_bytes = 256,
+	.writeback_chunk_size_kbytes = 8,
+	.ptoi_supported = false,
+	.num_dsc = 4,
+	.maximum_dsc_bits_per_component = 10,
+	.dsc422_native_support = false,
+	.is_line_buffer_bpp_fixed = true,
+	.line_buffer_fixed_bpp = 48,
+	.line_buffer_size_bits = 789504,
+	.max_line_buffer_lines = 12,
+	.writeback_interface_buffer_size_kbytes = 90,
+	.max_num_dpp = 4,
+	.max_num_otg = 4,
+	.max_num_hdmi_frl_outputs = 1,
+	.max_num_wb = 1,
+	.max_dchub_pscl_bw_pix_per_clk = 4,
+	.max_pscl_lb_bw_pix_per_clk = 2,
+	.max_lb_vscl_bw_pix_per_clk = 4,
+	.max_vscl_hscl_bw_pix_per_clk = 4,
+	.max_hscl_ratio = 6,
+	.max_vscl_ratio = 6,
+	.max_hscl_taps = 8,
+	.max_vscl_taps = 8,
+	.dpte_buffer_size_in_pte_reqs_luma = 64,
+	.dpte_buffer_size_in_pte_reqs_chroma = 34,
+	.dispclk_ramp_margin_percent = 1,
+	.max_inter_dcn_tile_repeaters = 8,
+	.cursor_buffer_size = 16,
+	.cursor_chunk_size = 2,
+	.writeback_line_buffer_buffer_size = 0,
+	.writeback_min_hscl_ratio = 1,
+	.writeback_min_vscl_ratio = 1,
+	.writeback_max_hscl_ratio = 1,
+	.writeback_max_vscl_ratio = 1,
+	.writeback_max_hscl_taps = 1,
+	.writeback_max_vscl_taps = 1,
+	.dppclk_delay_subtotal = 46,
+	.dppclk_delay_scl = 50,
+	.dppclk_delay_scl_lb_only = 16,
+	.dppclk_delay_cnvc_formatter = 27,
+	.dppclk_delay_cnvc_cursor = 6,
+	.dispclk_delay_subtotal = 119,
+	.dynamic_metadata_vm_enabled = false,
+	.odm_combine_4to1_supported = false,
+	.dcc_supported = true,
+};
+
+static struct _vcs_dpi_soc_bounding_box_st dcn3_14_soc = {
+		/*TODO: correct dispclk/dppclk voltage level determination*/
+	.clock_limits = {
+		{
+			.state = 0,
+			.dispclk_mhz = 1200.0,
+			.dppclk_mhz = 1200.0,
+			.phyclk_mhz = 600.0,
+			.phyclk_d18_mhz = 667.0,
+			.dscclk_mhz = 186.0,
+			.dtbclk_mhz = 600.0,
+		},
+		{
+			.state = 1,
+			.dispclk_mhz = 1200.0,
+			.dppclk_mhz = 1200.0,
+			.phyclk_mhz = 810.0,
+			.phyclk_d18_mhz = 667.0,
+			.dscclk_mhz = 209.0,
+			.dtbclk_mhz = 600.0,
+		},
+		{
+			.state = 2,
+			.dispclk_mhz = 1200.0,
+			.dppclk_mhz = 1200.0,
+			.phyclk_mhz = 810.0,
+			.phyclk_d18_mhz = 667.0,
+			.dscclk_mhz = 209.0,
+			.dtbclk_mhz = 600.0,
+		},
+		{
+			.state = 3,
+			.dispclk_mhz = 1200.0,
+			.dppclk_mhz = 1200.0,
+			.phyclk_mhz = 810.0,
+			.phyclk_d18_mhz = 667.0,
+			.dscclk_mhz = 371.0,
+			.dtbclk_mhz = 600.0,
+		},
+		{
+			.state = 4,
+			.dispclk_mhz = 1200.0,
+			.dppclk_mhz = 1200.0,
+			.phyclk_mhz = 810.0,
+			.phyclk_d18_mhz = 667.0,
+			.dscclk_mhz = 417.0,
+			.dtbclk_mhz = 600.0,
+		},
+	},
+	.num_states = 5,
+	.sr_exit_time_us = 16.5,
+	.sr_enter_plus_exit_time_us = 18.5,
+	.sr_exit_z8_time_us = 268.0,
+	.sr_enter_plus_exit_z8_time_us = 393.0,
+	.writeback_latency_us = 12.0,
+	.dram_channel_width_bytes = 4,
+	.round_trip_ping_latency_dcfclk_cycles = 106,
+	.urgent_latency_pixel_data_only_us = 4.0,
+	.urgent_latency_pixel_mixed_with_vm_data_us = 4.0,
+	.urgent_latency_vm_data_only_us = 4.0,
+	.urgent_out_of_order_return_per_channel_pixel_only_bytes = 4096,
+	.urgent_out_of_order_return_per_channel_pixel_and_vm_bytes = 4096,
+	.urgent_out_of_order_return_per_channel_vm_only_bytes = 4096,
+	.pct_ideal_sdp_bw_after_urgent = 80.0,
+	.pct_ideal_dram_sdp_bw_after_urgent_pixel_only = 65.0,
+	.pct_ideal_dram_sdp_bw_after_urgent_pixel_and_vm = 60.0,
+	.pct_ideal_dram_sdp_bw_after_urgent_vm_only = 30.0,
+	.max_avg_sdp_bw_use_normal_percent = 60.0,
+	.max_avg_dram_bw_use_normal_percent = 60.0,
+	.fabric_datapath_to_dcn_data_return_bytes = 32,
+	.return_bus_width_bytes = 64,
+	.downspread_percent = 0.38,
+	.dcn_downspread_percent = 0.5,
+	.gpuvm_min_page_size_bytes = 4096,
+	.hostvm_min_page_size_bytes = 4096,
+	.do_urgent_latency_adjustment = false,
+	.urgent_latency_adjustment_fabric_clock_component_us = 0,
+	.urgent_latency_adjustment_fabric_clock_reference_mhz = 0,
+};
+
+
+void dcn314_update_bw_bounding_box_fpu(struct dc *dc, struct clk_bw_params *bw_params)
+{
+	struct clk_limit_table *clk_table = &bw_params->clk_table;
+	struct _vcs_dpi_voltage_scaling_st *clock_limits =
+		dcn3_14_soc.clock_limits;
+	unsigned int i, closest_clk_lvl;
+	int max_dispclk_mhz = 0, max_dppclk_mhz = 0;
+	int j;
+
+	dc_assert_fp_enabled();
+
+	// Default clock levels are used for diags, which may lead to overclocking.
+	if (dc->config.use_default_clock_table == false) {
+		dcn3_14_ip.max_num_otg = dc->res_pool->res_cap->num_timing_generator;
+		dcn3_14_ip.max_num_dpp = dc->res_pool->pipe_count;
+
+		if (bw_params->dram_channel_width_bytes > 0)
+			dcn3_14_soc.dram_channel_width_bytes = bw_params->dram_channel_width_bytes;
+
+		if (bw_params->num_channels > 0)
+			dcn3_14_soc.num_chans = bw_params->num_channels;
+
+		ASSERT(dcn3_14_soc.num_chans);
+		ASSERT(clk_table->num_entries);
+
+		/* Prepass to find max clocks independent of voltage level. */
+		for (i = 0; i < clk_table->num_entries; ++i) {
+			if (clk_table->entries[i].dispclk_mhz > max_dispclk_mhz)
+				max_dispclk_mhz = clk_table->entries[i].dispclk_mhz;
+			if (clk_table->entries[i].dppclk_mhz > max_dppclk_mhz)
+				max_dppclk_mhz = clk_table->entries[i].dppclk_mhz;
+		}
+
+		for (i = 0; i < clk_table->num_entries; i++) {
+			/* loop backwards*/
+			for (closest_clk_lvl = 0, j = dcn3_14_soc.num_states - 1; j >= 0; j--) {
+				if ((unsigned int) dcn3_14_soc.clock_limits[j].dcfclk_mhz <= clk_table->entries[i].dcfclk_mhz) {
+					closest_clk_lvl = j;
+					break;
+				}
+			}
+			if (clk_table->num_entries == 1) {
+				/*smu gives one DPM level, let's take the highest one*/
+				closest_clk_lvl = dcn3_14_soc.num_states - 1;
+			}
+
+			clock_limits[i].state = i;
+
+			/* Clocks dependent on voltage level. */
+			clock_limits[i].dcfclk_mhz = clk_table->entries[i].dcfclk_mhz;
+			if (clk_table->num_entries == 1 &&
+				clock_limits[i].dcfclk_mhz < dcn3_14_soc.clock_limits[closest_clk_lvl].dcfclk_mhz) {
+				/*SMU fix not released yet*/
+				clock_limits[i].dcfclk_mhz = dcn3_14_soc.clock_limits[closest_clk_lvl].dcfclk_mhz;
+			}
+			clock_limits[i].fabricclk_mhz = clk_table->entries[i].fclk_mhz;
+			clock_limits[i].socclk_mhz = clk_table->entries[i].socclk_mhz;
+
+			if (clk_table->entries[i].memclk_mhz && clk_table->entries[i].wck_ratio)
+				clock_limits[i].dram_speed_mts = clk_table->entries[i].memclk_mhz * 2 * clk_table->entries[i].wck_ratio;
+
+			/* Clocks independent of voltage level. */
+			clock_limits[i].dispclk_mhz = max_dispclk_mhz ? max_dispclk_mhz :
+				dcn3_14_soc.clock_limits[closest_clk_lvl].dispclk_mhz;
+
+			clock_limits[i].dppclk_mhz = max_dppclk_mhz ? max_dppclk_mhz :
+				dcn3_14_soc.clock_limits[closest_clk_lvl].dppclk_mhz;
+
+			clock_limits[i].dram_bw_per_chan_gbps = dcn3_14_soc.clock_limits[closest_clk_lvl].dram_bw_per_chan_gbps;
+			clock_limits[i].dscclk_mhz = dcn3_14_soc.clock_limits[closest_clk_lvl].dscclk_mhz;
+			clock_limits[i].dtbclk_mhz = dcn3_14_soc.clock_limits[closest_clk_lvl].dtbclk_mhz;
+			clock_limits[i].phyclk_d18_mhz = dcn3_14_soc.clock_limits[closest_clk_lvl].phyclk_d18_mhz;
+			clock_limits[i].phyclk_mhz = dcn3_14_soc.clock_limits[closest_clk_lvl].phyclk_mhz;
+		}
+		for (i = 0; i < clk_table->num_entries; i++)
+			dcn3_14_soc.clock_limits[i] = clock_limits[i];
+		if (clk_table->num_entries) {
+			dcn3_14_soc.num_states = clk_table->num_entries;
+		}
+	}
+
+	if (max_dispclk_mhz) {
+		dcn3_14_soc.dispclk_dppclk_vco_speed_mhz = max_dispclk_mhz * 2;
+		dc->dml.soc.dispclk_dppclk_vco_speed_mhz = max_dispclk_mhz * 2;
+	}
+
+	dcn20_patch_bounding_box(dc, &dcn3_14_soc);
+	dml_init_instance(&dc->dml, &dcn3_14_soc, &dcn3_14_ip, DML_PROJECT_DCN314);
+}
+
+static bool is_dual_plane(enum surface_pixel_format format)
+{
+	return format >= SURFACE_PIXEL_FORMAT_VIDEO_BEGIN || format == SURFACE_PIXEL_FORMAT_GRPH_RGBE_ALPHA;
+}
+
+/*
+ * micro_sec_to_vert_lines () - converts time to number of vertical lines for a given timing
+ *
+ * @param: num_us: number of microseconds
+ * @return: number of vertical lines. If exact number of vertical lines is not found then
+ *          it will round up to next number of lines to guarantee num_us
+ */
+static unsigned int micro_sec_to_vert_lines(unsigned int num_us, struct dc_crtc_timing *timing)
+{
+	unsigned int num_lines = 0;
+	unsigned int lines_time_in_ns = 1000.0 *
+			(((float)timing->h_total * 1000.0) /
+			 ((float)timing->pix_clk_100hz / 10.0));
+
+	num_lines = dml_ceil(1000.0 * num_us / lines_time_in_ns, 1.0);
+
+	return num_lines;
+}
+
+static unsigned int get_vertical_back_porch(struct dc_crtc_timing *timing)
+{
+	unsigned int v_active = 0, v_blank = 0, v_back_porch = 0;
+
+	v_active = timing->v_border_top + timing->v_addressable + timing->v_border_bottom;
+	v_blank = timing->v_total - v_active;
+	v_back_porch = v_blank - timing->v_front_porch - timing->v_sync_width;
+
+	return v_back_porch;
+}
+
+int dcn314_populate_dml_pipes_from_context_fpu(struct dc *dc, struct dc_state *context,
+					       display_e2e_pipe_params_st *pipes,
+					       enum dc_validate_mode validate_mode)
+{
+	int i, pipe_cnt;
+	struct resource_context *res_ctx = &context->res_ctx;
+	struct pipe_ctx *pipe = 0;
+	bool upscaled = false;
+	const unsigned int max_allowed_vblank_nom = 1023;
+
+	dc_assert_fp_enabled();
+
+	dcn31x_populate_dml_pipes_from_context(dc, context, pipes, validate_mode);
+
+	for (i = 0, pipe_cnt = 0; i < dc->res_pool->pipe_count; i++) {
+		struct dc_crtc_timing *timing;
+		unsigned int num_lines = 0;
+		unsigned int v_back_porch = 0;
+
+		if (!res_ctx->pipe_ctx[i].stream)
+			continue;
+		pipe = &res_ctx->pipe_ctx[i];
+		timing = &pipe->stream->timing;
+
+		num_lines = micro_sec_to_vert_lines(dcn3_14_ip.VBlankNomDefaultUS, timing);
+
+		if (pipe->stream->adjust.v_total_min != 0)
+			pipes[pipe_cnt].pipe.dest.vtotal = pipe->stream->adjust.v_total_min;
+		else
+			pipes[pipe_cnt].pipe.dest.vtotal = timing->v_total;
+
+		v_back_porch  = get_vertical_back_porch(timing);
+
+		pipes[pipe_cnt].pipe.dest.vblank_nom = timing->v_total - pipes[pipe_cnt].pipe.dest.vactive;
+		pipes[pipe_cnt].pipe.dest.vblank_nom = min(pipes[pipe_cnt].pipe.dest.vblank_nom, num_lines);
+		// vblank_nom should not smaller than (VSync (timing->v_sync_width + v_back_porch) + 2)
+		// + 2 is because
+		// 1 -> VStartup_start should be 1 line before VSync
+		// 1 -> always reserve 1 line between start of vblank to vstartup signal
+		pipes[pipe_cnt].pipe.dest.vblank_nom =
+			max(pipes[pipe_cnt].pipe.dest.vblank_nom, timing->v_sync_width + v_back_porch + 2);
+		pipes[pipe_cnt].pipe.dest.vblank_nom = min(pipes[pipe_cnt].pipe.dest.vblank_nom, max_allowed_vblank_nom);
+
+		if (pipe->plane_state &&
+				(pipe->plane_state->src_rect.height < pipe->plane_state->dst_rect.height ||
+				pipe->plane_state->src_rect.width < pipe->plane_state->dst_rect.width))
+			upscaled = true;
+
+		/* Apply HostVM policy - either based on hypervisor globally enabled, or rIOMMU active */
+		if (dc->debug.dml_hostvm_override == DML_HOSTVM_NO_OVERRIDE)
+			pipes[i].pipe.src.hostvm = dc->vm_pa_config.is_hvm_enabled || dc->res_pool->hubbub->riommu_active;
+
+		/*
+		 * Immediate flip can be set dynamically after enabling the plane.
+		 * We need to require support for immediate flip or underflow can be
+		 * intermittently experienced depending on peak b/w requirements.
+		 */
+		pipes[pipe_cnt].pipe.src.immediate_flip = true;
+
+		pipes[pipe_cnt].pipe.src.unbounded_req_mode = false;
+		pipes[pipe_cnt].pipe.src.dcc_fraction_of_zs_req_luma = 0;
+		pipes[pipe_cnt].pipe.src.dcc_fraction_of_zs_req_chroma = 0;
+		pipes[pipe_cnt].pipe.dest.vfront_porch = timing->v_front_porch;
+		pipes[pipe_cnt].pipe.src.dcc_rate = 3;
+		pipes[pipe_cnt].dout.dsc_input_bpc = 0;
+
+		if (pipes[pipe_cnt].dout.dsc_enable) {
+			switch (timing->display_color_depth) {
+			case COLOR_DEPTH_888:
+				pipes[pipe_cnt].dout.dsc_input_bpc = 8;
+				break;
+			case COLOR_DEPTH_101010:
+				pipes[pipe_cnt].dout.dsc_input_bpc = 10;
+				break;
+			case COLOR_DEPTH_121212:
+				pipes[pipe_cnt].dout.dsc_input_bpc = 12;
+				break;
+			default:
+				ASSERT(0);
+				break;
+			}
+		}
+
+		pipe_cnt++;
+	}
+	context->bw_ctx.dml.ip.det_buffer_size_kbytes = DCN3_14_DEFAULT_DET_SIZE;
+
+	dc->config.enable_4to1MPC = false;
+	if (pipe_cnt == 1 && pipe->plane_state
+		&& pipe->plane_state->rotation == ROTATION_ANGLE_0 && !dc->debug.disable_z9_mpc) {
+		if (is_dual_plane(pipe->plane_state->format)
+				&& pipe->plane_state->src_rect.width <= 1920 && pipe->plane_state->src_rect.height <= 1080) {
+			dc->config.enable_4to1MPC = true;
+		} else if (!is_dual_plane(pipe->plane_state->format) && pipe->plane_state->src_rect.width <= 5120) {
+			/* Limit to 5k max to avoid forced pipe split when there is not enough detile for swath */
+			context->bw_ctx.dml.ip.det_buffer_size_kbytes = 192;
+			pipes[0].pipe.src.unbounded_req_mode = true;
+		}
+	} else if (context->stream_count >= dc->debug.crb_alloc_policy_min_disp_count
+			&& dc->debug.crb_alloc_policy > DET_SIZE_DEFAULT) {
+		context->bw_ctx.dml.ip.det_buffer_size_kbytes = dc->debug.crb_alloc_policy * 64;
+	} else if (context->stream_count >= 3 && upscaled) {
+		context->bw_ctx.dml.ip.det_buffer_size_kbytes = 192;
+	}
+
+	if (dc->debug.force_odm_combine_4to1)
+		context->bw_ctx.dml.ip.odm_combine_4to1_supported = true;
+
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
+
+		if (!pipe->stream)
+			continue;
+
+		if (pipe->stream->signal == SIGNAL_TYPE_EDP && dc->debug.seamless_boot_odm_combine &&
+				pipe->stream->apply_seamless_boot_optimization) {
+
+			if (pipe->stream->apply_boot_odm_mode == dm_odm_combine_policy_2to1) {
+				context->bw_ctx.dml.vba.ODMCombinePolicy = dm_odm_combine_policy_2to1;
+				break;
+			}
+		}
+	}
+
+	return pipe_cnt;
+}
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn314/dcn314_fpu.h b/drivers/gpu/drm/amd/display/dc/dml/dcn314/dcn314_fpu.h
new file mode 100644
index 000000000000..362ac79184ea
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn314/dcn314_fpu.h
@@ -0,0 +1,40 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __DCN314_FPU_H__
+#define __DCN314_FPU_H__
+
+#define DCN3_14_DEFAULT_DET_SIZE 384
+#define DCN3_14_MAX_DET_SIZE 384
+#define DCN3_14_MIN_COMPBUF_SIZE_KB 128
+#define DCN3_14_CRB_SEGMENT_SIZE_KB 64
+
+void dcn314_update_bw_bounding_box_fpu(struct dc *dc, struct clk_bw_params *bw_params);
+int dcn314_populate_dml_pipes_from_context_fpu(struct dc *dc, struct dc_state *context,
+					       display_e2e_pipe_params_st *pipes,
+					       enum dc_validate_mode validate_mode);
+
+#endif
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn314/display_mode_vba_314.c b/drivers/gpu/drm/amd/display/dc/dml/dcn314/display_mode_vba_314.c
new file mode 100644
index 000000000000..9f3938a50240
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn314/display_mode_vba_314.c
@@ -0,0 +1,7343 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#include "dc.h"
+#include "../display_mode_lib.h"
+#include "display_mode_vba_314.h"
+#include "../dml_inline_defs.h"
+
+/*
+ * NOTE:
+ *   This file is gcc-parsable HW gospel, coming straight from HW engineers.
+ *
+ * It doesn't adhere to Linux kernel style and sometimes will do things in odd
+ * ways. Unless there is something clearly wrong with it the code should
+ * remain as-is as it provides us with a guarantee from HW that it is correct.
+ */
+
+#define BPP_INVALID 0
+#define BPP_BLENDED_PIPE 0xffffffff
+#define DCN314_MAX_DSC_IMAGE_WIDTH 5184
+#define DCN314_MAX_FMT_420_BUFFER_WIDTH 4096
+
+// For DML-C changes that hasn't been propagated to VBA yet
+//#define __DML_VBA_ALLOW_DELTA__
+
+// Move these to ip parameters/constant
+
+// At which vstartup the DML start to try if the mode can be supported
+#define __DML_VBA_MIN_VSTARTUP__    9
+
+// Delay in DCFCLK from ARB to DET (1st num is ARB to SDPIF, 2nd number is SDPIF to DET)
+#define __DML_ARB_TO_RET_DELAY__    (7 + 95)
+
+// fudge factor for min dcfclk calclation
+#define __DML_MIN_DCFCLK_FACTOR__   1.15
+
+typedef struct {
+	double DPPCLK;
+	double DISPCLK;
+	double PixelClock;
+	double DCFCLKDeepSleep;
+	unsigned int DPPPerPlane;
+	bool ScalerEnabled;
+	double VRatio;
+	double VRatioChroma;
+	enum scan_direction_class SourceScan;
+	unsigned int BlockWidth256BytesY;
+	unsigned int BlockHeight256BytesY;
+	unsigned int BlockWidth256BytesC;
+	unsigned int BlockHeight256BytesC;
+	unsigned int InterlaceEnable;
+	unsigned int NumberOfCursors;
+	unsigned int VBlank;
+	unsigned int HTotal;
+	unsigned int DCCEnable;
+	bool ODMCombineIsEnabled;
+	enum source_format_class SourcePixelFormat;
+	int BytePerPixelY;
+	int BytePerPixelC;
+	bool ProgressiveToInterlaceUnitInOPP;
+} Pipe;
+
+#define BPP_INVALID 0
+#define BPP_BLENDED_PIPE 0xffffffff
+
+static bool CalculateBytePerPixelAnd256BBlockSizes(
+		enum source_format_class SourcePixelFormat,
+		enum dm_swizzle_mode SurfaceTiling,
+		unsigned int *BytePerPixelY,
+		unsigned int *BytePerPixelC,
+		double *BytePerPixelDETY,
+		double *BytePerPixelDETC,
+		unsigned int *BlockHeight256BytesY,
+		unsigned int *BlockHeight256BytesC,
+		unsigned int *BlockWidth256BytesY,
+		unsigned int *BlockWidth256BytesC);
+static void DisplayPipeConfiguration(struct display_mode_lib *mode_lib);
+static void DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation(struct display_mode_lib *mode_lib);
+static unsigned int dscceComputeDelay(
+		unsigned int bpc,
+		double BPP,
+		unsigned int sliceWidth,
+		unsigned int numSlices,
+		enum output_format_class pixelFormat,
+		enum output_encoder_class Output);
+static unsigned int dscComputeDelay(enum output_format_class pixelFormat, enum output_encoder_class Output);
+static bool CalculatePrefetchSchedule(
+		struct display_mode_lib *mode_lib,
+		double HostVMInefficiencyFactor,
+		Pipe *myPipe,
+		unsigned int DSCDelay,
+		double DPPCLKDelaySubtotalPlusCNVCFormater,
+		double DPPCLKDelaySCL,
+		double DPPCLKDelaySCLLBOnly,
+		double DPPCLKDelayCNVCCursor,
+		double DISPCLKDelaySubtotal,
+		unsigned int DPP_RECOUT_WIDTH,
+		enum output_format_class OutputFormat,
+		unsigned int MaxInterDCNTileRepeaters,
+		unsigned int VStartup,
+		unsigned int MaxVStartup,
+		unsigned int GPUVMPageTableLevels,
+		bool GPUVMEnable,
+		bool HostVMEnable,
+		unsigned int HostVMMaxNonCachedPageTableLevels,
+		double HostVMMinPageSize,
+		bool DynamicMetadataEnable,
+		bool DynamicMetadataVMEnabled,
+		int DynamicMetadataLinesBeforeActiveRequired,
+		unsigned int DynamicMetadataTransmittedBytes,
+		double UrgentLatency,
+		double UrgentExtraLatency,
+		double TCalc,
+		unsigned int PDEAndMetaPTEBytesFrame,
+		unsigned int MetaRowByte,
+		unsigned int PixelPTEBytesPerRow,
+		double PrefetchSourceLinesY,
+		unsigned int SwathWidthY,
+		double VInitPreFillY,
+		unsigned int MaxNumSwathY,
+		double PrefetchSourceLinesC,
+		unsigned int SwathWidthC,
+		double VInitPreFillC,
+		unsigned int MaxNumSwathC,
+		int swath_width_luma_ub,
+		int swath_width_chroma_ub,
+		unsigned int SwathHeightY,
+		unsigned int SwathHeightC,
+		double TWait,
+		double *DSTXAfterScaler,
+		double *DSTYAfterScaler,
+		double *DestinationLinesForPrefetch,
+		double *PrefetchBandwidth,
+		double *DestinationLinesToRequestVMInVBlank,
+		double *DestinationLinesToRequestRowInVBlank,
+		double *VRatioPrefetchY,
+		double *VRatioPrefetchC,
+		double *RequiredPrefetchPixDataBWLuma,
+		double *RequiredPrefetchPixDataBWChroma,
+		bool *NotEnoughTimeForDynamicMetadata,
+		double *Tno_bw,
+		double *prefetch_vmrow_bw,
+		double *Tdmdl_vm,
+		double *Tdmdl,
+		double *TSetup,
+		int *VUpdateOffsetPix,
+		double *VUpdateWidthPix,
+		double *VReadyOffsetPix);
+static double RoundToDFSGranularityUp(double Clock, double VCOSpeed);
+static double RoundToDFSGranularityDown(double Clock, double VCOSpeed);
+static void CalculateDCCConfiguration(
+		bool DCCEnabled,
+		bool DCCProgrammingAssumesScanDirectionUnknown,
+		enum source_format_class SourcePixelFormat,
+		unsigned int SurfaceWidthLuma,
+		unsigned int SurfaceWidthChroma,
+		unsigned int SurfaceHeightLuma,
+		unsigned int SurfaceHeightChroma,
+		double DETBufferSize,
+		unsigned int RequestHeight256ByteLuma,
+		unsigned int RequestHeight256ByteChroma,
+		enum dm_swizzle_mode TilingFormat,
+		unsigned int BytePerPixelY,
+		unsigned int BytePerPixelC,
+		double BytePerPixelDETY,
+		double BytePerPixelDETC,
+		enum scan_direction_class ScanOrientation,
+		unsigned int *MaxUncompressedBlockLuma,
+		unsigned int *MaxUncompressedBlockChroma,
+		unsigned int *MaxCompressedBlockLuma,
+		unsigned int *MaxCompressedBlockChroma,
+		unsigned int *IndependentBlockLuma,
+		unsigned int *IndependentBlockChroma);
+static double CalculatePrefetchSourceLines(
+		struct display_mode_lib *mode_lib,
+		double VRatio,
+		double vtaps,
+		bool Interlace,
+		bool ProgressiveToInterlaceUnitInOPP,
+		unsigned int SwathHeight,
+		unsigned int ViewportYStart,
+		double *VInitPreFill,
+		unsigned int *MaxNumSwath);
+static unsigned int CalculateVMAndRowBytes(
+		struct display_mode_lib *mode_lib,
+		bool DCCEnable,
+		unsigned int BlockHeight256Bytes,
+		unsigned int BlockWidth256Bytes,
+		enum source_format_class SourcePixelFormat,
+		unsigned int SurfaceTiling,
+		unsigned int BytePerPixel,
+		enum scan_direction_class ScanDirection,
+		unsigned int SwathWidth,
+		unsigned int ViewportHeight,
+		bool GPUVMEnable,
+		bool HostVMEnable,
+		unsigned int HostVMMaxNonCachedPageTableLevels,
+		unsigned int GPUVMMinPageSize,
+		unsigned int HostVMMinPageSize,
+		unsigned int PTEBufferSizeInRequests,
+		unsigned int Pitch,
+		unsigned int DCCMetaPitch,
+		unsigned int *MacroTileWidth,
+		unsigned int *MetaRowByte,
+		unsigned int *PixelPTEBytesPerRow,
+		bool *PTEBufferSizeNotExceeded,
+		int *dpte_row_width_ub,
+		unsigned int *dpte_row_height,
+		unsigned int *MetaRequestWidth,
+		unsigned int *MetaRequestHeight,
+		unsigned int *meta_row_width,
+		unsigned int *meta_row_height,
+		int *vm_group_bytes,
+		unsigned int *dpte_group_bytes,
+		unsigned int *PixelPTEReqWidth,
+		unsigned int *PixelPTEReqHeight,
+		unsigned int *PTERequestSize,
+		int *DPDE0BytesFrame,
+		int *MetaPTEBytesFrame);
+static double CalculateTWait(unsigned int PrefetchMode, double DRAMClockChangeLatency, double UrgentLatency, double SREnterPlusExitTime);
+static void CalculateRowBandwidth(
+		bool GPUVMEnable,
+		enum source_format_class SourcePixelFormat,
+		double VRatio,
+		double VRatioChroma,
+		bool DCCEnable,
+		double LineTime,
+		unsigned int MetaRowByteLuma,
+		unsigned int MetaRowByteChroma,
+		unsigned int meta_row_height_luma,
+		unsigned int meta_row_height_chroma,
+		unsigned int PixelPTEBytesPerRowLuma,
+		unsigned int PixelPTEBytesPerRowChroma,
+		unsigned int dpte_row_height_luma,
+		unsigned int dpte_row_height_chroma,
+		double *meta_row_bw,
+		double *dpte_row_bw);
+
+static void CalculateFlipSchedule(
+		struct display_mode_lib *mode_lib,
+		unsigned int k,
+		double HostVMInefficiencyFactor,
+		double UrgentExtraLatency,
+		double UrgentLatency,
+		double PDEAndMetaPTEBytesPerFrame,
+		double MetaRowBytes,
+		double DPTEBytesPerRow);
+static double CalculateWriteBackDelay(
+		enum source_format_class WritebackPixelFormat,
+		double WritebackHRatio,
+		double WritebackVRatio,
+		unsigned int WritebackVTaps,
+		int WritebackDestinationWidth,
+		int WritebackDestinationHeight,
+		int WritebackSourceHeight,
+		unsigned int HTotal);
+
+static void CalculateVupdateAndDynamicMetadataParameters(
+		int MaxInterDCNTileRepeaters,
+		double DPPCLK,
+		double DISPCLK,
+		double DCFClkDeepSleep,
+		double PixelClock,
+		int HTotal,
+		int VBlank,
+		int DynamicMetadataTransmittedBytes,
+		int DynamicMetadataLinesBeforeActiveRequired,
+		int InterlaceEnable,
+		bool ProgressiveToInterlaceUnitInOPP,
+		double *TSetup,
+		double *Tdmbf,
+		double *Tdmec,
+		double *Tdmsks,
+		int *VUpdateOffsetPix,
+		double *VUpdateWidthPix,
+		double *VReadyOffsetPix);
+
+static void CalculateWatermarksAndDRAMSpeedChangeSupport(
+		struct display_mode_lib *mode_lib,
+		unsigned int PrefetchMode,
+		double DCFCLK,
+		double ReturnBW,
+		double UrgentLatency,
+		double ExtraLatency,
+		double SOCCLK,
+		double DCFCLKDeepSleep,
+		unsigned int DETBufferSizeY[],
+		unsigned int DETBufferSizeC[],
+		unsigned int SwathHeightY[],
+		unsigned int SwathHeightC[],
+		double SwathWidthY[],
+		double SwathWidthC[],
+		unsigned int DPPPerPlane[],
+		double BytePerPixelDETY[],
+		double BytePerPixelDETC[],
+		bool UnboundedRequestEnabled,
+		unsigned int CompressedBufferSizeInkByte,
+		enum clock_change_support *DRAMClockChangeSupport,
+		double *StutterExitWatermark,
+		double *StutterEnterPlusExitWatermark,
+		double *Z8StutterExitWatermark,
+		double *Z8StutterEnterPlusExitWatermark);
+
+static void CalculateDCFCLKDeepSleep(
+		struct display_mode_lib *mode_lib,
+		unsigned int NumberOfActivePlanes,
+		int BytePerPixelY[],
+		int BytePerPixelC[],
+		double VRatio[],
+		double VRatioChroma[],
+		double SwathWidthY[],
+		double SwathWidthC[],
+		unsigned int DPPPerPlane[],
+		double HRatio[],
+		double HRatioChroma[],
+		double PixelClock[],
+		double PSCL_THROUGHPUT[],
+		double PSCL_THROUGHPUT_CHROMA[],
+		double DPPCLK[],
+		double ReadBandwidthLuma[],
+		double ReadBandwidthChroma[],
+		int ReturnBusWidth,
+		double *DCFCLKDeepSleep);
+
+static void CalculateUrgentBurstFactor(
+		int swath_width_luma_ub,
+		int swath_width_chroma_ub,
+		unsigned int SwathHeightY,
+		unsigned int SwathHeightC,
+		double LineTime,
+		double UrgentLatency,
+		double CursorBufferSize,
+		unsigned int CursorWidth,
+		unsigned int CursorBPP,
+		double VRatio,
+		double VRatioC,
+		double BytePerPixelInDETY,
+		double BytePerPixelInDETC,
+		double DETBufferSizeY,
+		double DETBufferSizeC,
+		double *UrgentBurstFactorCursor,
+		double *UrgentBurstFactorLuma,
+		double *UrgentBurstFactorChroma,
+		bool *NotEnoughUrgentLatencyHiding);
+
+static void UseMinimumDCFCLK(
+		struct display_mode_lib *mode_lib,
+		int MaxPrefetchMode,
+		int ReorderingBytes);
+
+static void CalculatePixelDeliveryTimes(
+		unsigned int NumberOfActivePlanes,
+		double VRatio[],
+		double VRatioChroma[],
+		double VRatioPrefetchY[],
+		double VRatioPrefetchC[],
+		unsigned int swath_width_luma_ub[],
+		unsigned int swath_width_chroma_ub[],
+		unsigned int DPPPerPlane[],
+		double HRatio[],
+		double HRatioChroma[],
+		double PixelClock[],
+		double PSCL_THROUGHPUT[],
+		double PSCL_THROUGHPUT_CHROMA[],
+		double DPPCLK[],
+		int BytePerPixelC[],
+		enum scan_direction_class SourceScan[],
+		unsigned int NumberOfCursors[],
+		unsigned int CursorWidth[][DC__NUM_CURSOR__MAX],
+		unsigned int CursorBPP[][DC__NUM_CURSOR__MAX],
+		unsigned int BlockWidth256BytesY[],
+		unsigned int BlockHeight256BytesY[],
+		unsigned int BlockWidth256BytesC[],
+		unsigned int BlockHeight256BytesC[],
+		double DisplayPipeLineDeliveryTimeLuma[],
+		double DisplayPipeLineDeliveryTimeChroma[],
+		double DisplayPipeLineDeliveryTimeLumaPrefetch[],
+		double DisplayPipeLineDeliveryTimeChromaPrefetch[],
+		double DisplayPipeRequestDeliveryTimeLuma[],
+		double DisplayPipeRequestDeliveryTimeChroma[],
+		double DisplayPipeRequestDeliveryTimeLumaPrefetch[],
+		double DisplayPipeRequestDeliveryTimeChromaPrefetch[],
+		double CursorRequestDeliveryTime[],
+		double CursorRequestDeliveryTimePrefetch[]);
+
+static void CalculateMetaAndPTETimes(
+		int NumberOfActivePlanes,
+		bool GPUVMEnable,
+		int MetaChunkSize,
+		int MinMetaChunkSizeBytes,
+		int HTotal[],
+		double VRatio[],
+		double VRatioChroma[],
+		double DestinationLinesToRequestRowInVBlank[],
+		double DestinationLinesToRequestRowInImmediateFlip[],
+		bool DCCEnable[],
+		double PixelClock[],
+		int BytePerPixelY[],
+		int BytePerPixelC[],
+		enum scan_direction_class SourceScan[],
+		int dpte_row_height[],
+		int dpte_row_height_chroma[],
+		int meta_row_width[],
+		int meta_row_width_chroma[],
+		int meta_row_height[],
+		int meta_row_height_chroma[],
+		int meta_req_width[],
+		int meta_req_width_chroma[],
+		int meta_req_height[],
+		int meta_req_height_chroma[],
+		int dpte_group_bytes[],
+		int PTERequestSizeY[],
+		int PTERequestSizeC[],
+		int PixelPTEReqWidthY[],
+		int PixelPTEReqHeightY[],
+		int PixelPTEReqWidthC[],
+		int PixelPTEReqHeightC[],
+		int dpte_row_width_luma_ub[],
+		int dpte_row_width_chroma_ub[],
+		double DST_Y_PER_PTE_ROW_NOM_L[],
+		double DST_Y_PER_PTE_ROW_NOM_C[],
+		double DST_Y_PER_META_ROW_NOM_L[],
+		double DST_Y_PER_META_ROW_NOM_C[],
+		double TimePerMetaChunkNominal[],
+		double TimePerChromaMetaChunkNominal[],
+		double TimePerMetaChunkVBlank[],
+		double TimePerChromaMetaChunkVBlank[],
+		double TimePerMetaChunkFlip[],
+		double TimePerChromaMetaChunkFlip[],
+		double time_per_pte_group_nom_luma[],
+		double time_per_pte_group_vblank_luma[],
+		double time_per_pte_group_flip_luma[],
+		double time_per_pte_group_nom_chroma[],
+		double time_per_pte_group_vblank_chroma[],
+		double time_per_pte_group_flip_chroma[]);
+
+static void CalculateVMGroupAndRequestTimes(
+		unsigned int NumberOfActivePlanes,
+		bool GPUVMEnable,
+		unsigned int GPUVMMaxPageTableLevels,
+		unsigned int HTotal[],
+		int BytePerPixelC[],
+		double DestinationLinesToRequestVMInVBlank[],
+		double DestinationLinesToRequestVMInImmediateFlip[],
+		bool DCCEnable[],
+		double PixelClock[],
+		int dpte_row_width_luma_ub[],
+		int dpte_row_width_chroma_ub[],
+		int vm_group_bytes[],
+		unsigned int dpde0_bytes_per_frame_ub_l[],
+		unsigned int dpde0_bytes_per_frame_ub_c[],
+		int meta_pte_bytes_per_frame_ub_l[],
+		int meta_pte_bytes_per_frame_ub_c[],
+		double TimePerVMGroupVBlank[],
+		double TimePerVMGroupFlip[],
+		double TimePerVMRequestVBlank[],
+		double TimePerVMRequestFlip[]);
+
+static void CalculateStutterEfficiency(
+		struct display_mode_lib *mode_lib,
+		int CompressedBufferSizeInkByte,
+		bool UnboundedRequestEnabled,
+		int ConfigReturnBufferSizeInKByte,
+		int MetaFIFOSizeInKEntries,
+		int ZeroSizeBufferEntries,
+		int NumberOfActivePlanes,
+		int ROBBufferSizeInKByte,
+		double TotalDataReadBandwidth,
+		double DCFCLK,
+		double ReturnBW,
+		double COMPBUF_RESERVED_SPACE_64B,
+		double COMPBUF_RESERVED_SPACE_ZS,
+		double SRExitTime,
+		double SRExitZ8Time,
+		bool SynchronizedVBlank,
+		double Z8StutterEnterPlusExitWatermark,
+		double StutterEnterPlusExitWatermark,
+		bool ProgressiveToInterlaceUnitInOPP,
+		bool Interlace[],
+		double MinTTUVBlank[],
+		int DPPPerPlane[],
+		unsigned int DETBufferSizeY[],
+		int BytePerPixelY[],
+		double BytePerPixelDETY[],
+		double SwathWidthY[],
+		int SwathHeightY[],
+		int SwathHeightC[],
+		double NetDCCRateLuma[],
+		double NetDCCRateChroma[],
+		double DCCFractionOfZeroSizeRequestsLuma[],
+		double DCCFractionOfZeroSizeRequestsChroma[],
+		int HTotal[],
+		int VTotal[],
+		double PixelClock[],
+		double VRatio[],
+		enum scan_direction_class SourceScan[],
+		int BlockHeight256BytesY[],
+		int BlockWidth256BytesY[],
+		int BlockHeight256BytesC[],
+		int BlockWidth256BytesC[],
+		int DCCYMaxUncompressedBlock[],
+		int DCCCMaxUncompressedBlock[],
+		int VActive[],
+		bool DCCEnable[],
+		bool WritebackEnable[],
+		double ReadBandwidthPlaneLuma[],
+		double ReadBandwidthPlaneChroma[],
+		double meta_row_bw[],
+		double dpte_row_bw[],
+		double *StutterEfficiencyNotIncludingVBlank,
+		double *StutterEfficiency,
+		int *NumberOfStutterBurstsPerFrame,
+		double *Z8StutterEfficiencyNotIncludingVBlank,
+		double *Z8StutterEfficiency,
+		int *Z8NumberOfStutterBurstsPerFrame,
+		double *StutterPeriod);
+
+static void CalculateSwathAndDETConfiguration(
+		bool ForceSingleDPP,
+		int NumberOfActivePlanes,
+		unsigned int DETBufferSizeInKByte,
+		double MaximumSwathWidthLuma[],
+		double MaximumSwathWidthChroma[],
+		enum scan_direction_class SourceScan[],
+		enum source_format_class SourcePixelFormat[],
+		enum dm_swizzle_mode SurfaceTiling[],
+		int ViewportWidth[],
+		int ViewportHeight[],
+		int SurfaceWidthY[],
+		int SurfaceWidthC[],
+		int SurfaceHeightY[],
+		int SurfaceHeightC[],
+		int Read256BytesBlockHeightY[],
+		int Read256BytesBlockHeightC[],
+		int Read256BytesBlockWidthY[],
+		int Read256BytesBlockWidthC[],
+		enum odm_combine_mode ODMCombineEnabled[],
+		int BlendingAndTiming[],
+		int BytePerPixY[],
+		int BytePerPixC[],
+		double BytePerPixDETY[],
+		double BytePerPixDETC[],
+		int HActive[],
+		double HRatio[],
+		double HRatioChroma[],
+		int DPPPerPlane[],
+		int swath_width_luma_ub[],
+		int swath_width_chroma_ub[],
+		double SwathWidth[],
+		double SwathWidthChroma[],
+		int SwathHeightY[],
+		int SwathHeightC[],
+		unsigned int DETBufferSizeY[],
+		unsigned int DETBufferSizeC[],
+		bool ViewportSizeSupportPerPlane[],
+		bool *ViewportSizeSupport);
+static void CalculateSwathWidth(
+		bool ForceSingleDPP,
+		int NumberOfActivePlanes,
+		enum source_format_class SourcePixelFormat[],
+		enum scan_direction_class SourceScan[],
+		int ViewportWidth[],
+		int ViewportHeight[],
+		int SurfaceWidthY[],
+		int SurfaceWidthC[],
+		int SurfaceHeightY[],
+		int SurfaceHeightC[],
+		enum odm_combine_mode ODMCombineEnabled[],
+		int BytePerPixY[],
+		int BytePerPixC[],
+		int Read256BytesBlockHeightY[],
+		int Read256BytesBlockHeightC[],
+		int Read256BytesBlockWidthY[],
+		int Read256BytesBlockWidthC[],
+		int BlendingAndTiming[],
+		int HActive[],
+		double HRatio[],
+		int DPPPerPlane[],
+		double SwathWidthSingleDPPY[],
+		double SwathWidthSingleDPPC[],
+		double SwathWidthY[],
+		double SwathWidthC[],
+		int MaximumSwathHeightY[],
+		int MaximumSwathHeightC[],
+		int swath_width_luma_ub[],
+		int swath_width_chroma_ub[]);
+
+static double CalculateExtraLatency(
+		int RoundTripPingLatencyCycles,
+		int ReorderingBytes,
+		double DCFCLK,
+		int TotalNumberOfActiveDPP,
+		int PixelChunkSizeInKByte,
+		int TotalNumberOfDCCActiveDPP,
+		int MetaChunkSize,
+		double ReturnBW,
+		bool GPUVMEnable,
+		bool HostVMEnable,
+		int NumberOfActivePlanes,
+		int NumberOfDPP[],
+		int dpte_group_bytes[],
+		double HostVMInefficiencyFactor,
+		double HostVMMinPageSize,
+		int HostVMMaxNonCachedPageTableLevels);
+
+static double CalculateExtraLatencyBytes(
+		int ReorderingBytes,
+		int TotalNumberOfActiveDPP,
+		int PixelChunkSizeInKByte,
+		int TotalNumberOfDCCActiveDPP,
+		int MetaChunkSize,
+		bool GPUVMEnable,
+		bool HostVMEnable,
+		int NumberOfActivePlanes,
+		int NumberOfDPP[],
+		int dpte_group_bytes[],
+		double HostVMInefficiencyFactor,
+		double HostVMMinPageSize,
+		int HostVMMaxNonCachedPageTableLevels);
+
+static double CalculateUrgentLatency(
+		double UrgentLatencyPixelDataOnly,
+		double UrgentLatencyPixelMixedWithVMData,
+		double UrgentLatencyVMDataOnly,
+		bool DoUrgentLatencyAdjustment,
+		double UrgentLatencyAdjustmentFabricClockComponent,
+		double UrgentLatencyAdjustmentFabricClockReference,
+		double FabricClockSingle);
+
+static void CalculateUnboundedRequestAndCompressedBufferSize(
+		unsigned int DETBufferSizeInKByte,
+		int ConfigReturnBufferSizeInKByte,
+		enum unbounded_requesting_policy UseUnboundedRequestingFinal,
+		int TotalActiveDPP,
+		bool NoChromaPlanes,
+		int MaxNumDPP,
+		int CompressedBufferSegmentSizeInkByteFinal,
+		enum output_encoder_class *Output,
+		bool *UnboundedRequestEnabled,
+		int *CompressedBufferSizeInkByte);
+
+static bool UnboundedRequest(enum unbounded_requesting_policy UseUnboundedRequestingFinal, int TotalNumberOfActiveDPP, bool NoChroma, enum output_encoder_class Output);
+static unsigned int CalculateMaxVStartup(
+		unsigned int VTotal,
+		unsigned int VActive,
+		unsigned int VBlankNom,
+		unsigned int HTotal,
+		double PixelClock,
+		bool ProgressiveTointerlaceUnitinOPP,
+		bool Interlace,
+		unsigned int VBlankNomDefaultUS,
+		double WritebackDelayTime);
+
+void dml314_recalculate(struct display_mode_lib *mode_lib)
+{
+	ModeSupportAndSystemConfiguration(mode_lib);
+	PixelClockAdjustmentForProgressiveToInterlaceUnit(mode_lib);
+	DisplayPipeConfiguration(mode_lib);
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: Calling DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation\n", __func__);
+#endif
+	DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation(mode_lib);
+}
+
+static unsigned int dscceComputeDelay(
+		unsigned int bpc,
+		double BPP,
+		unsigned int sliceWidth,
+		unsigned int numSlices,
+		enum output_format_class pixelFormat,
+		enum output_encoder_class Output)
+{
+	// valid bpc         = source bits per component in the set of {8, 10, 12}
+	// valid bpp         = increments of 1/16 of a bit
+	//                    min = 6/7/8 in N420/N422/444, respectively
+	//                    max = such that compression is 1:1
+	//valid sliceWidth  = number of pixels per slice line, must be less than or equal to 5184/numSlices (or 4096/numSlices in 420 mode)
+	//valid numSlices   = number of slices in the horiziontal direction per DSC engine in the set of {1, 2, 3, 4}
+	//valid pixelFormat = pixel/color format in the set of {:N444_RGB, :S422, :N422, :N420}
+
+	// fixed value
+	unsigned int rcModelSize = 8192;
+
+	// N422/N420 operate at 2 pixels per clock
+	unsigned int pixelsPerClock = 0, lstall, D, initalXmitDelay, w, s, ix, wx, P, l0, a, ax, L, Delay, pixels;
+
+	if (pixelFormat == dm_420)
+		pixelsPerClock = 2;
+	else if (pixelFormat == dm_444)
+		pixelsPerClock = 1;
+	else if (pixelFormat == dm_n422)
+		pixelsPerClock = 2;
+	// #all other modes operate at 1 pixel per clock
+	else
+		pixelsPerClock = 1;
+
+	//initial transmit delay as per PPS
+	initalXmitDelay = dml_round(rcModelSize / 2.0 / BPP / pixelsPerClock);
+
+	//compute ssm delay
+	if (bpc == 8)
+		D = 81;
+	else if (bpc == 10)
+		D = 89;
+	else
+		D = 113;
+
+	//divide by pixel per cycle to compute slice width as seen by DSC
+	w = sliceWidth / pixelsPerClock;
+
+	//422 mode has an additional cycle of delay
+	if (pixelFormat == dm_420 || pixelFormat == dm_444 || pixelFormat == dm_n422)
+		s = 0;
+	else
+		s = 1;
+
+	//main calculation for the dscce
+	ix = initalXmitDelay + 45;
+	wx = (w + 2) / 3;
+	P = 3 * wx - w;
+	l0 = ix / w;
+	a = ix + P * l0;
+	ax = (a + 2) / 3 + D + 6 + 1;
+	L = (ax + wx - 1) / wx;
+	if ((ix % w) == 0 && P != 0)
+		lstall = 1;
+	else
+		lstall = 0;
+	Delay = L * wx * (numSlices - 1) + ax + s + lstall + 22;
+
+	//dsc processes 3 pixel containers per cycle and a container can contain 1 or 2 pixels
+	pixels = Delay * 3 * pixelsPerClock;
+	return pixels;
+}
+
+static unsigned int dscComputeDelay(enum output_format_class pixelFormat, enum output_encoder_class Output)
+{
+	unsigned int Delay = 0;
+
+	if (pixelFormat == dm_420) {
+		//   sfr
+		Delay = Delay + 2;
+		//   dsccif
+		Delay = Delay + 0;
+		//   dscc - input deserializer
+		Delay = Delay + 3;
+		//   dscc gets pixels every other cycle
+		Delay = Delay + 2;
+		//   dscc - input cdc fifo
+		Delay = Delay + 12;
+		//   dscc gets pixels every other cycle
+		Delay = Delay + 13;
+		//   dscc - cdc uncertainty
+		Delay = Delay + 2;
+		//   dscc - output cdc fifo
+		Delay = Delay + 7;
+		//   dscc gets pixels every other cycle
+		Delay = Delay + 3;
+		//   dscc - cdc uncertainty
+		Delay = Delay + 2;
+		//   dscc - output serializer
+		Delay = Delay + 1;
+		//   sft
+		Delay = Delay + 1;
+	} else if (pixelFormat == dm_n422) {
+		//   sfr
+		Delay = Delay + 2;
+		//   dsccif
+		Delay = Delay + 1;
+		//   dscc - input deserializer
+		Delay = Delay + 5;
+		//  dscc - input cdc fifo
+		Delay = Delay + 25;
+		//   dscc - cdc uncertainty
+		Delay = Delay + 2;
+		//   dscc - output cdc fifo
+		Delay = Delay + 10;
+		//   dscc - cdc uncertainty
+		Delay = Delay + 2;
+		//   dscc - output serializer
+		Delay = Delay + 1;
+		//   sft
+		Delay = Delay + 1;
+	} else {
+		//   sfr
+		Delay = Delay + 2;
+		//   dsccif
+		Delay = Delay + 0;
+		//   dscc - input deserializer
+		Delay = Delay + 3;
+		//   dscc - input cdc fifo
+		Delay = Delay + 12;
+		//   dscc - cdc uncertainty
+		Delay = Delay + 2;
+		//   dscc - output cdc fifo
+		Delay = Delay + 7;
+		//   dscc - output serializer
+		Delay = Delay + 1;
+		//   dscc - cdc uncertainty
+		Delay = Delay + 2;
+		//   sft
+		Delay = Delay + 1;
+	}
+
+	return Delay;
+}
+
+static bool CalculatePrefetchSchedule(
+		struct display_mode_lib *mode_lib,
+		double HostVMInefficiencyFactor,
+		Pipe *myPipe,
+		unsigned int DSCDelay,
+		double DPPCLKDelaySubtotalPlusCNVCFormater,
+		double DPPCLKDelaySCL,
+		double DPPCLKDelaySCLLBOnly,
+		double DPPCLKDelayCNVCCursor,
+		double DISPCLKDelaySubtotal,
+		unsigned int DPP_RECOUT_WIDTH,
+		enum output_format_class OutputFormat,
+		unsigned int MaxInterDCNTileRepeaters,
+		unsigned int VStartup,
+		unsigned int MaxVStartup,
+		unsigned int GPUVMPageTableLevels,
+		bool GPUVMEnable,
+		bool HostVMEnable,
+		unsigned int HostVMMaxNonCachedPageTableLevels,
+		double HostVMMinPageSize,
+		bool DynamicMetadataEnable,
+		bool DynamicMetadataVMEnabled,
+		int DynamicMetadataLinesBeforeActiveRequired,
+		unsigned int DynamicMetadataTransmittedBytes,
+		double UrgentLatency,
+		double UrgentExtraLatency,
+		double TCalc,
+		unsigned int PDEAndMetaPTEBytesFrame,
+		unsigned int MetaRowByte,
+		unsigned int PixelPTEBytesPerRow,
+		double PrefetchSourceLinesY,
+		unsigned int SwathWidthY,
+		double VInitPreFillY,
+		unsigned int MaxNumSwathY,
+		double PrefetchSourceLinesC,
+		unsigned int SwathWidthC,
+		double VInitPreFillC,
+		unsigned int MaxNumSwathC,
+		int swath_width_luma_ub,
+		int swath_width_chroma_ub,
+		unsigned int SwathHeightY,
+		unsigned int SwathHeightC,
+		double TWait,
+		double *DSTXAfterScaler,
+		double *DSTYAfterScaler,
+		double *DestinationLinesForPrefetch,
+		double *PrefetchBandwidth,
+		double *DestinationLinesToRequestVMInVBlank,
+		double *DestinationLinesToRequestRowInVBlank,
+		double *VRatioPrefetchY,
+		double *VRatioPrefetchC,
+		double *RequiredPrefetchPixDataBWLuma,
+		double *RequiredPrefetchPixDataBWChroma,
+		bool *NotEnoughTimeForDynamicMetadata,
+		double *Tno_bw,
+		double *prefetch_vmrow_bw,
+		double *Tdmdl_vm,
+		double *Tdmdl,
+		double *TSetup,
+		int *VUpdateOffsetPix,
+		double *VUpdateWidthPix,
+		double *VReadyOffsetPix)
+{
+	bool MyError = false;
+	unsigned int DPPCycles, DISPCLKCycles;
+	double DSTTotalPixelsAfterScaler;
+	double LineTime;
+	double dst_y_prefetch_equ;
+#ifdef __DML_VBA_DEBUG__
+	double Tsw_oto;
+#endif
+	double prefetch_bw_oto;
+	double prefetch_bw_pr;
+	double Tvm_oto;
+	double Tr0_oto;
+	double Tvm_oto_lines;
+	double Tr0_oto_lines;
+	double dst_y_prefetch_oto;
+	double TimeForFetchingMetaPTE = 0;
+	double TimeForFetchingRowInVBlank = 0;
+	double LinesToRequestPrefetchPixelData = 0;
+	unsigned int HostVMDynamicLevelsTrips;
+	double trip_to_mem;
+	double Tvm_trips;
+	double Tr0_trips;
+	double Tvm_trips_rounded;
+	double Tr0_trips_rounded;
+	double Lsw_oto;
+	double Tpre_rounded;
+	double prefetch_bw_equ;
+	double Tvm_equ;
+	double Tr0_equ;
+	double Tdmbf;
+	double Tdmec;
+	double Tdmsks;
+	double prefetch_sw_bytes;
+	double bytes_pp;
+	double dep_bytes;
+	int max_vratio_pre = 4;
+	double min_Lsw;
+	double Tsw_est1 = 0;
+	double Tsw_est3 = 0;
+	double  max_Tsw = 0;
+
+	if (GPUVMEnable == true && HostVMEnable == true) {
+		HostVMDynamicLevelsTrips = HostVMMaxNonCachedPageTableLevels;
+	} else {
+		HostVMDynamicLevelsTrips = 0;
+	}
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: GPUVMEnable=%d HostVMEnable=%d HostVMInefficiencyFactor=%f\n", __func__, GPUVMEnable, HostVMEnable, HostVMInefficiencyFactor);
+#endif
+	CalculateVupdateAndDynamicMetadataParameters(
+			MaxInterDCNTileRepeaters,
+			myPipe->DPPCLK,
+			myPipe->DISPCLK,
+			myPipe->DCFCLKDeepSleep,
+			myPipe->PixelClock,
+			myPipe->HTotal,
+			myPipe->VBlank,
+			DynamicMetadataTransmittedBytes,
+			DynamicMetadataLinesBeforeActiveRequired,
+			myPipe->InterlaceEnable,
+			myPipe->ProgressiveToInterlaceUnitInOPP,
+			TSetup,
+			&Tdmbf,
+			&Tdmec,
+			&Tdmsks,
+			VUpdateOffsetPix,
+			VUpdateWidthPix,
+			VReadyOffsetPix);
+
+	LineTime = myPipe->HTotal / myPipe->PixelClock;
+	trip_to_mem = UrgentLatency;
+	Tvm_trips = UrgentExtraLatency + trip_to_mem * (GPUVMPageTableLevels * (HostVMDynamicLevelsTrips + 1) - 1);
+
+#ifdef __DML_VBA_ALLOW_DELTA__
+	if (DynamicMetadataVMEnabled == true && GPUVMEnable == true) {
+#else
+	if (DynamicMetadataVMEnabled == true) {
+#endif
+		*Tdmdl = TWait + Tvm_trips + trip_to_mem;
+	} else {
+		*Tdmdl = TWait + UrgentExtraLatency;
+	}
+
+#ifdef __DML_VBA_ALLOW_DELTA__
+	if (DynamicMetadataEnable == false) {
+		*Tdmdl = 0.0;
+	}
+#endif
+
+	if (DynamicMetadataEnable == true) {
+		if (VStartup * LineTime < *TSetup + *Tdmdl + Tdmbf + Tdmec + Tdmsks) {
+			*NotEnoughTimeForDynamicMetadata = true;
+			dml_print("DML::%s: Not Enough Time for Dynamic Meta!\n", __func__);
+			dml_print("DML::%s: Tdmbf: %fus - time for dmd transfer from dchub to dio output buffer\n", __func__, Tdmbf);
+			dml_print("DML::%s: Tdmec: %fus - time dio takes to transfer dmd\n", __func__, Tdmec);
+			dml_print("DML::%s: Tdmsks: %fus - time before active dmd must complete transmission at dio\n", __func__, Tdmsks);
+			dml_print("DML::%s: Tdmdl: %fus - time for fabric to become ready and fetch dmd\n", __func__, *Tdmdl);
+		} else {
+			*NotEnoughTimeForDynamicMetadata = false;
+		}
+	} else {
+		*NotEnoughTimeForDynamicMetadata = false;
+	}
+
+	*Tdmdl_vm = (DynamicMetadataEnable == true && DynamicMetadataVMEnabled == true && GPUVMEnable == true ? TWait + Tvm_trips : 0);
+
+	if (myPipe->ScalerEnabled)
+		DPPCycles = DPPCLKDelaySubtotalPlusCNVCFormater + DPPCLKDelaySCL;
+	else
+		DPPCycles = DPPCLKDelaySubtotalPlusCNVCFormater + DPPCLKDelaySCLLBOnly;
+
+	DPPCycles = DPPCycles + myPipe->NumberOfCursors * DPPCLKDelayCNVCCursor;
+
+	DISPCLKCycles = DISPCLKDelaySubtotal;
+
+	if (myPipe->DPPCLK == 0.0 || myPipe->DISPCLK == 0.0)
+		return true;
+
+	*DSTXAfterScaler = DPPCycles * myPipe->PixelClock / myPipe->DPPCLK + DISPCLKCycles * myPipe->PixelClock / myPipe->DISPCLK + DSCDelay;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: DPPCycles: %d\n", __func__, DPPCycles);
+	dml_print("DML::%s: PixelClock: %f\n", __func__, myPipe->PixelClock);
+	dml_print("DML::%s: DPPCLK: %f\n", __func__, myPipe->DPPCLK);
+	dml_print("DML::%s: DISPCLKCycles: %d\n", __func__, DISPCLKCycles);
+	dml_print("DML::%s: DISPCLK: %f\n", __func__, myPipe->DISPCLK);
+	dml_print("DML::%s: DSCDelay: %d\n", __func__, DSCDelay);
+	dml_print("DML::%s: DSTXAfterScaler: %d\n", __func__, *DSTXAfterScaler);
+	dml_print("DML::%s: ODMCombineIsEnabled: %d\n", __func__, myPipe->ODMCombineIsEnabled);
+#endif
+
+	*DSTXAfterScaler = *DSTXAfterScaler + ((myPipe->ODMCombineIsEnabled) ? 18 : 0) + (myPipe->DPPPerPlane - 1) * DPP_RECOUT_WIDTH;
+
+	if (OutputFormat == dm_420 || (myPipe->InterlaceEnable && myPipe->ProgressiveToInterlaceUnitInOPP))
+		*DSTYAfterScaler = 1;
+	else
+		*DSTYAfterScaler = 0;
+
+	DSTTotalPixelsAfterScaler = *DSTYAfterScaler * myPipe->HTotal + *DSTXAfterScaler;
+	*DSTYAfterScaler = dml_floor(DSTTotalPixelsAfterScaler / myPipe->HTotal, 1);
+	*DSTXAfterScaler = DSTTotalPixelsAfterScaler - ((double) (*DSTYAfterScaler * myPipe->HTotal));
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: DSTXAfterScaler: %d (final)\n", __func__, *DSTXAfterScaler);
+#endif
+
+	MyError = false;
+
+	Tr0_trips = trip_to_mem * (HostVMDynamicLevelsTrips + 1);
+	Tvm_trips_rounded = dml_ceil(4.0 * Tvm_trips / LineTime, 1) / 4 * LineTime;
+	Tr0_trips_rounded = dml_ceil(4.0 * Tr0_trips / LineTime, 1) / 4 * LineTime;
+
+#ifdef __DML_VBA_ALLOW_DELTA__
+	if (!myPipe->DCCEnable) {
+		Tr0_trips = 0.0;
+		Tr0_trips_rounded = 0.0;
+	}
+#endif
+
+	if (!GPUVMEnable) {
+		Tvm_trips = 0.0;
+		Tvm_trips_rounded = 0.0;
+	}
+
+	if (GPUVMEnable) {
+		if (GPUVMPageTableLevels >= 3) {
+			*Tno_bw = UrgentExtraLatency + trip_to_mem * ((GPUVMPageTableLevels - 2) - 1);
+		} else {
+			*Tno_bw = 0;
+		}
+	} else if (!myPipe->DCCEnable) {
+		*Tno_bw = LineTime;
+	} else {
+		*Tno_bw = LineTime / 4;
+	}
+
+	if (myPipe->SourcePixelFormat == dm_420_8 || myPipe->SourcePixelFormat == dm_420_10 || myPipe->SourcePixelFormat == dm_420_12)
+		bytes_pp = myPipe->BytePerPixelY + myPipe->BytePerPixelC / 4;
+	else
+		bytes_pp = myPipe->BytePerPixelY + myPipe->BytePerPixelC;
+	/*rev 99*/
+	prefetch_bw_pr = bytes_pp * myPipe->PixelClock / (double) myPipe->DPPPerPlane;
+	prefetch_bw_pr = dml_min(1, myPipe->VRatio) * prefetch_bw_pr;
+	max_Tsw = dml_max(PrefetchSourceLinesY, PrefetchSourceLinesC) * LineTime;
+	prefetch_sw_bytes = PrefetchSourceLinesY * swath_width_luma_ub * myPipe->BytePerPixelY + PrefetchSourceLinesC * swath_width_chroma_ub * myPipe->BytePerPixelC;
+	prefetch_bw_oto = dml_max(prefetch_bw_pr, prefetch_sw_bytes / max_Tsw);
+
+	min_Lsw = dml_max(1, dml_max(PrefetchSourceLinesY, PrefetchSourceLinesC) / max_vratio_pre);
+	Lsw_oto = dml_ceil(4 * dml_max(prefetch_sw_bytes / prefetch_bw_oto / LineTime, min_Lsw), 1) / 4;
+#ifdef __DML_VBA_DEBUG__
+	Tsw_oto = Lsw_oto * LineTime;
+#endif
+
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML: HTotal: %d\n", myPipe->HTotal);
+	dml_print("DML: prefetch_bw_oto: %f\n", prefetch_bw_oto);
+	dml_print("DML: PrefetchSourceLinesY: %f\n", PrefetchSourceLinesY);
+	dml_print("DML: swath_width_luma_ub: %d\n", swath_width_luma_ub);
+	dml_print("DML: BytePerPixelY: %d\n", myPipe->BytePerPixelY);
+	dml_print("DML: Tsw_oto: %f\n", Tsw_oto);
+#endif
+
+	if (GPUVMEnable == true)
+		Tvm_oto = dml_max3(*Tno_bw + PDEAndMetaPTEBytesFrame * HostVMInefficiencyFactor / prefetch_bw_oto, Tvm_trips, LineTime / 4.0);
+	else
+		Tvm_oto = LineTime / 4.0;
+
+	if ((GPUVMEnable == true || myPipe->DCCEnable == true)) {
+		Tr0_oto = dml_max4((MetaRowByte + PixelPTEBytesPerRow * HostVMInefficiencyFactor) / prefetch_bw_oto, Tr0_trips, // PREVIOUS_ERROR (missing this term)
+				LineTime - Tvm_oto,
+				LineTime / 4);
+	} else {
+		Tr0_oto = (LineTime - Tvm_oto) / 2.0;
+	}
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: Tvm_trips = %f\n", __func__, Tvm_trips);
+	dml_print("DML::%s: Tr0_trips = %f\n", __func__, Tr0_trips);
+	dml_print("DML::%s: PDEAndMetaPTEBytesFrame = %d\n", __func__, MetaRowByte);
+	dml_print("DML::%s: MetaRowByte = %d\n", __func__, MetaRowByte);
+	dml_print("DML::%s: PixelPTEBytesPerRow = %d\n", __func__, PixelPTEBytesPerRow);
+	dml_print("DML::%s: HostVMInefficiencyFactor = %f\n", __func__, HostVMInefficiencyFactor);
+	dml_print("DML::%s: prefetch_bw_oto = %f\n", __func__, prefetch_bw_oto);
+	dml_print("DML::%s: Tr0_oto = %f\n", __func__, Tr0_oto);
+	dml_print("DML::%s: Tvm_oto = %f\n", __func__, Tvm_oto);
+#endif
+
+	Tvm_oto_lines = dml_ceil(4.0 * Tvm_oto / LineTime, 1) / 4.0;
+	Tr0_oto_lines = dml_ceil(4.0 * Tr0_oto / LineTime, 1) / 4.0;
+	dst_y_prefetch_oto = Tvm_oto_lines + 2 * Tr0_oto_lines + Lsw_oto;
+	dst_y_prefetch_equ =  VStartup - (*TSetup + dml_max(TWait + TCalc, *Tdmdl)) / LineTime - (*DSTYAfterScaler + *DSTXAfterScaler / myPipe->HTotal);
+	dst_y_prefetch_equ = dml_min(dst_y_prefetch_equ, 63.75); // limit to the reg limit of U6.2 for DST_Y_PREFETCH
+	dst_y_prefetch_equ = dml_floor(4.0 * (dst_y_prefetch_equ + 0.125), 1) / 4.0;
+	Tpre_rounded = dst_y_prefetch_equ * LineTime;
+
+	dep_bytes = dml_max(PDEAndMetaPTEBytesFrame * HostVMInefficiencyFactor, MetaRowByte + PixelPTEBytesPerRow * HostVMInefficiencyFactor);
+
+	if (prefetch_sw_bytes < dep_bytes)
+		prefetch_sw_bytes = 2 * dep_bytes;
+
+	dml_print("DML: dst_y_prefetch_oto: %f\n", dst_y_prefetch_oto);
+	dml_print("DML: Tvm_oto_lines: %f\n", Tvm_oto_lines);
+	dml_print("DML: Tr0_oto_lines: %f\n", Tr0_oto_lines);
+	dml_print("DML: Lsw_oto: %f\n", Lsw_oto);
+	dml_print("DML: LineTime: %f\n", LineTime);
+	dml_print("DML: dst_y_prefetch_equ: %f (after round)\n", dst_y_prefetch_equ);
+
+	dml_print("DML: LineTime: %f\n", LineTime);
+	dml_print("DML: VStartup: %d\n", VStartup);
+	dml_print("DML: Tvstartup: %fus - time between vstartup and first pixel of active\n", VStartup * LineTime);
+	dml_print("DML: TSetup: %fus - time from vstartup to vready\n", *TSetup);
+	dml_print("DML: TCalc: %fus - time for calculations in dchub starting at vready\n", TCalc);
+	dml_print("DML: TWait: %fus - time for fabric to become ready max(pstate exit,cstate enter/exit, urgent latency) after TCalc\n", TWait);
+	dml_print("DML: Tdmbf: %fus - time for dmd transfer from dchub to dio output buffer\n", Tdmbf);
+	dml_print("DML: Tdmec: %fus - time dio takes to transfer dmd\n", Tdmec);
+	dml_print("DML: Tdmsks: %fus - time before active dmd must complete transmission at dio\n", Tdmsks);
+	dml_print("DML: Tdmdl_vm: %fus - time for vm stages of dmd\n", *Tdmdl_vm);
+	dml_print("DML: Tdmdl: %fus - time for fabric to become ready and fetch dmd\n", *Tdmdl);
+	dml_print("DML: DSTXAfterScaler: %f pixels - number of pixel clocks pipeline and buffer delay after scaler\n", *DSTXAfterScaler);
+	dml_print("DML: DSTYAfterScaler: %f lines - number of lines of pipeline and buffer delay after scaler\n", *DSTYAfterScaler);
+
+	*PrefetchBandwidth = 0;
+	*DestinationLinesToRequestVMInVBlank = 0;
+	*DestinationLinesToRequestRowInVBlank = 0;
+	*VRatioPrefetchY = 0;
+	*VRatioPrefetchC = 0;
+	*RequiredPrefetchPixDataBWLuma = 0;
+	if (dst_y_prefetch_equ > 1) {
+		double PrefetchBandwidth1;
+		double PrefetchBandwidth2;
+		double PrefetchBandwidth3;
+		double PrefetchBandwidth4;
+
+		if (Tpre_rounded - *Tno_bw > 0) {
+			PrefetchBandwidth1 = (PDEAndMetaPTEBytesFrame * HostVMInefficiencyFactor + 2 * MetaRowByte + 2 * PixelPTEBytesPerRow * HostVMInefficiencyFactor
+					+ prefetch_sw_bytes) / (Tpre_rounded - *Tno_bw);
+			Tsw_est1 = prefetch_sw_bytes / PrefetchBandwidth1;
+		} else {
+			PrefetchBandwidth1 = 0;
+		}
+
+		if (VStartup == MaxVStartup && Tsw_est1 / LineTime < min_Lsw && Tpre_rounded - min_Lsw * LineTime - 0.75 * LineTime - *Tno_bw > 0) {
+			PrefetchBandwidth1 = (PDEAndMetaPTEBytesFrame * HostVMInefficiencyFactor + 2 * MetaRowByte + 2 * PixelPTEBytesPerRow * HostVMInefficiencyFactor)
+					/ (Tpre_rounded - min_Lsw * LineTime - 0.75 * LineTime - *Tno_bw);
+		}
+
+		if (Tpre_rounded - *Tno_bw - 2 * Tr0_trips_rounded > 0)
+			PrefetchBandwidth2 = (PDEAndMetaPTEBytesFrame * HostVMInefficiencyFactor + prefetch_sw_bytes) / (Tpre_rounded - *Tno_bw - 2 * Tr0_trips_rounded);
+		else
+			PrefetchBandwidth2 = 0;
+
+		if (Tpre_rounded - Tvm_trips_rounded > 0) {
+			PrefetchBandwidth3 = (2 * MetaRowByte + 2 * PixelPTEBytesPerRow * HostVMInefficiencyFactor
+					+ prefetch_sw_bytes) / (Tpre_rounded - Tvm_trips_rounded);
+			Tsw_est3 = prefetch_sw_bytes / PrefetchBandwidth3;
+		} else {
+			PrefetchBandwidth3 = 0;
+		}
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: Tpre_rounded: %f\n", __func__, Tpre_rounded);
+		dml_print("DML::%s: Tvm_trips_rounded: %f\n", __func__, Tvm_trips_rounded);
+		dml_print("DML::%s: PrefetchBandwidth3: %f\n", __func__, PrefetchBandwidth3);
+#endif
+		if (VStartup == MaxVStartup && Tsw_est3 / LineTime < min_Lsw && Tpre_rounded - min_Lsw * LineTime - 0.75 * LineTime - Tvm_trips_rounded > 0) {
+			PrefetchBandwidth3 = (2 * MetaRowByte + 2 * PixelPTEBytesPerRow * HostVMInefficiencyFactor)
+					/ (Tpre_rounded - min_Lsw * LineTime - 0.75 * LineTime - Tvm_trips_rounded);
+		}
+
+		if (Tpre_rounded - Tvm_trips_rounded - 2 * Tr0_trips_rounded > 0)
+			PrefetchBandwidth4 = prefetch_sw_bytes / (Tpre_rounded - Tvm_trips_rounded - 2 * Tr0_trips_rounded);
+		else
+			PrefetchBandwidth4 = 0;
+
+		{
+			bool Case1OK;
+			bool Case2OK;
+			bool Case3OK;
+
+			if (PrefetchBandwidth1 > 0) {
+				if (*Tno_bw + PDEAndMetaPTEBytesFrame * HostVMInefficiencyFactor / PrefetchBandwidth1 >= Tvm_trips_rounded
+						&& (MetaRowByte + PixelPTEBytesPerRow * HostVMInefficiencyFactor) / PrefetchBandwidth1 >= Tr0_trips_rounded) {
+					Case1OK = true;
+				} else {
+					Case1OK = false;
+				}
+			} else {
+				Case1OK = false;
+			}
+
+			if (PrefetchBandwidth2 > 0) {
+				if (*Tno_bw + PDEAndMetaPTEBytesFrame * HostVMInefficiencyFactor / PrefetchBandwidth2 >= Tvm_trips_rounded
+						&& (MetaRowByte + PixelPTEBytesPerRow * HostVMInefficiencyFactor) / PrefetchBandwidth2 < Tr0_trips_rounded) {
+					Case2OK = true;
+				} else {
+					Case2OK = false;
+				}
+			} else {
+				Case2OK = false;
+			}
+
+			if (PrefetchBandwidth3 > 0) {
+				if (*Tno_bw + PDEAndMetaPTEBytesFrame * HostVMInefficiencyFactor / PrefetchBandwidth3 < Tvm_trips_rounded
+						&& (MetaRowByte + PixelPTEBytesPerRow * HostVMInefficiencyFactor) / PrefetchBandwidth3 >= Tr0_trips_rounded) {
+					Case3OK = true;
+				} else {
+					Case3OK = false;
+				}
+			} else {
+				Case3OK = false;
+			}
+
+			if (Case1OK) {
+				prefetch_bw_equ = PrefetchBandwidth1;
+			} else if (Case2OK) {
+				prefetch_bw_equ = PrefetchBandwidth2;
+			} else if (Case3OK) {
+				prefetch_bw_equ = PrefetchBandwidth3;
+			} else {
+				prefetch_bw_equ = PrefetchBandwidth4;
+			}
+
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: Case1OK: %d\n", __func__, Case1OK);
+			dml_print("DML::%s: Case2OK: %d\n", __func__, Case2OK);
+			dml_print("DML::%s: Case3OK: %d\n", __func__, Case3OK);
+			dml_print("DML::%s: prefetch_bw_equ: %f\n", __func__, prefetch_bw_equ);
+#endif
+
+			if (prefetch_bw_equ > 0) {
+				if (GPUVMEnable == true) {
+					Tvm_equ = dml_max3(*Tno_bw + PDEAndMetaPTEBytesFrame * HostVMInefficiencyFactor / prefetch_bw_equ, Tvm_trips, LineTime / 4);
+				} else {
+					Tvm_equ = LineTime / 4;
+				}
+
+				if ((GPUVMEnable == true || myPipe->DCCEnable == true)) {
+					Tr0_equ = dml_max4(
+							(MetaRowByte + PixelPTEBytesPerRow * HostVMInefficiencyFactor) / prefetch_bw_equ,
+							Tr0_trips,
+							(LineTime - Tvm_equ) / 2,
+							LineTime / 4);
+				} else {
+					Tr0_equ = (LineTime - Tvm_equ) / 2;
+				}
+			} else {
+				Tvm_equ = 0;
+				Tr0_equ = 0;
+				dml_print("DML: prefetch_bw_equ equals 0! %s:%d\n", __FILE__, __LINE__);
+			}
+		}
+
+		if (dst_y_prefetch_oto < dst_y_prefetch_equ) {
+			*DestinationLinesForPrefetch = dst_y_prefetch_oto;
+			TimeForFetchingMetaPTE = Tvm_oto;
+			TimeForFetchingRowInVBlank = Tr0_oto;
+			*PrefetchBandwidth = prefetch_bw_oto;
+		} else {
+			*DestinationLinesForPrefetch = dst_y_prefetch_equ;
+			TimeForFetchingMetaPTE = Tvm_equ;
+			TimeForFetchingRowInVBlank = Tr0_equ;
+			*PrefetchBandwidth = prefetch_bw_equ;
+		}
+
+		*DestinationLinesToRequestVMInVBlank = dml_ceil(4.0 * TimeForFetchingMetaPTE / LineTime, 1.0) / 4.0;
+
+		*DestinationLinesToRequestRowInVBlank = dml_ceil(4.0 * TimeForFetchingRowInVBlank / LineTime, 1.0) / 4.0;
+
+#ifdef __DML_VBA_ALLOW_DELTA__
+		LinesToRequestPrefetchPixelData = *DestinationLinesForPrefetch
+		// See note above dated 5/30/2018
+		//                      - ((NumberOfCursors > 0 || GPUVMEnable || DCCEnable) ?
+				- ((GPUVMEnable || myPipe->DCCEnable) ? (*DestinationLinesToRequestVMInVBlank + 2 * *DestinationLinesToRequestRowInVBlank) : 0.0); // TODO: Did someone else add this??
+#else
+		LinesToRequestPrefetchPixelData = *DestinationLinesForPrefetch - *DestinationLinesToRequestVMInVBlank - 2 * *DestinationLinesToRequestRowInVBlank;
+#endif
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: DestinationLinesForPrefetch = %f\n", __func__, *DestinationLinesForPrefetch);
+		dml_print("DML::%s: DestinationLinesToRequestVMInVBlank = %f\n", __func__, *DestinationLinesToRequestVMInVBlank);
+		dml_print("DML::%s: TimeForFetchingRowInVBlank = %f\n", __func__, TimeForFetchingRowInVBlank);
+		dml_print("DML::%s: LineTime = %f\n", __func__, LineTime);
+		dml_print("DML::%s: DestinationLinesToRequestRowInVBlank = %f\n", __func__, *DestinationLinesToRequestRowInVBlank);
+		dml_print("DML::%s: PrefetchSourceLinesY = %f\n", __func__, PrefetchSourceLinesY);
+		dml_print("DML::%s: LinesToRequestPrefetchPixelData = %f\n", __func__, LinesToRequestPrefetchPixelData);
+#endif
+
+		if (LinesToRequestPrefetchPixelData > 0 && prefetch_bw_equ > 0) {
+
+			*VRatioPrefetchY = (double) PrefetchSourceLinesY / LinesToRequestPrefetchPixelData;
+			*VRatioPrefetchY = dml_max(*VRatioPrefetchY, 1.0);
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: VRatioPrefetchY = %f\n", __func__, *VRatioPrefetchY);
+			dml_print("DML::%s: SwathHeightY = %d\n", __func__, SwathHeightY);
+			dml_print("DML::%s: VInitPreFillY = %f\n", __func__, VInitPreFillY);
+#endif
+			if ((SwathHeightY > 4) && (VInitPreFillY > 3)) {
+				if (LinesToRequestPrefetchPixelData > (VInitPreFillY - 3.0) / 2.0) {
+					*VRatioPrefetchY = dml_max(
+							(double) PrefetchSourceLinesY / LinesToRequestPrefetchPixelData,
+							(double) MaxNumSwathY * SwathHeightY / (LinesToRequestPrefetchPixelData - (VInitPreFillY - 3.0) / 2.0));
+					*VRatioPrefetchY = dml_max(*VRatioPrefetchY, 1.0);
+				} else {
+					MyError = true;
+					dml_print("DML: MyErr set %s:%d\n", __FILE__, __LINE__);
+					*VRatioPrefetchY = 0;
+				}
+#ifdef __DML_VBA_DEBUG__
+				dml_print("DML::%s: VRatioPrefetchY = %f\n", __func__, *VRatioPrefetchY);
+				dml_print("DML::%s: PrefetchSourceLinesY = %f\n", __func__, PrefetchSourceLinesY);
+				dml_print("DML::%s: MaxNumSwathY = %d\n", __func__, MaxNumSwathY);
+#endif
+			}
+
+			*VRatioPrefetchC = (double) PrefetchSourceLinesC / LinesToRequestPrefetchPixelData;
+			*VRatioPrefetchC = dml_max(*VRatioPrefetchC, 1.0);
+
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: VRatioPrefetchC = %f\n", __func__, *VRatioPrefetchC);
+			dml_print("DML::%s: SwathHeightC = %d\n", __func__, SwathHeightC);
+			dml_print("DML::%s: VInitPreFillC = %f\n", __func__, VInitPreFillC);
+#endif
+			if ((SwathHeightC > 4) || VInitPreFillC > 3) {
+				if (LinesToRequestPrefetchPixelData > (VInitPreFillC - 3.0) / 2.0) {
+					*VRatioPrefetchC = dml_max(
+							*VRatioPrefetchC,
+							(double) MaxNumSwathC * SwathHeightC / (LinesToRequestPrefetchPixelData - (VInitPreFillC - 3.0) / 2.0));
+					*VRatioPrefetchC = dml_max(*VRatioPrefetchC, 1.0);
+				} else {
+					MyError = true;
+					dml_print("DML: MyErr set %s:%d\n", __FILE__, __LINE__);
+					*VRatioPrefetchC = 0;
+				}
+#ifdef __DML_VBA_DEBUG__
+				dml_print("DML::%s: VRatioPrefetchC = %f\n", __func__, *VRatioPrefetchC);
+				dml_print("DML::%s: PrefetchSourceLinesC = %f\n", __func__, PrefetchSourceLinesC);
+				dml_print("DML::%s: MaxNumSwathC = %d\n", __func__, MaxNumSwathC);
+#endif
+			}
+
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: BytePerPixelY = %d\n", __func__, myPipe->BytePerPixelY);
+			dml_print("DML::%s: swath_width_luma_ub = %d\n", __func__, swath_width_luma_ub);
+			dml_print("DML::%s: LineTime = %f\n", __func__, LineTime);
+#endif
+
+			*RequiredPrefetchPixDataBWLuma = (double) PrefetchSourceLinesY / LinesToRequestPrefetchPixelData * myPipe->BytePerPixelY * swath_width_luma_ub / LineTime;
+
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: RequiredPrefetchPixDataBWLuma = %f\n", __func__, *RequiredPrefetchPixDataBWLuma);
+#endif
+
+			*RequiredPrefetchPixDataBWChroma = (double) PrefetchSourceLinesC / LinesToRequestPrefetchPixelData * myPipe->BytePerPixelC * swath_width_chroma_ub
+					/ LineTime;
+		} else {
+			MyError = true;
+			dml_print("DML: MyErr set %s:%d\n", __FILE__, __LINE__);
+			dml_print("DML: LinesToRequestPrefetchPixelData: %f, should be > 0\n", LinesToRequestPrefetchPixelData);
+			*VRatioPrefetchY = 0;
+			*VRatioPrefetchC = 0;
+			*RequiredPrefetchPixDataBWLuma = 0;
+			*RequiredPrefetchPixDataBWChroma = 0;
+		}
+
+		dml_print(
+				"DML: Tpre: %fus - sum of time to request meta pte, 2 x data pte + meta data, swaths\n",
+				(double) LinesToRequestPrefetchPixelData * LineTime + 2.0 * TimeForFetchingRowInVBlank + TimeForFetchingMetaPTE);
+		dml_print("DML:  Tvm: %fus - time to fetch page tables for meta surface\n", TimeForFetchingMetaPTE);
+		dml_print("DML:  Tr0: %fus - time to fetch first row of data pagetables and first row of meta data (done in parallel)\n", TimeForFetchingRowInVBlank);
+		dml_print(
+				"DML:  Tsw: %fus = time to fetch enough pixel data and cursor data to feed the scalers init position and detile\n",
+				(double) LinesToRequestPrefetchPixelData * LineTime);
+		dml_print("DML: To: %fus - time for propagation from scaler to optc\n", (*DSTYAfterScaler + ((double) (*DSTXAfterScaler) / (double) myPipe->HTotal)) * LineTime);
+		dml_print("DML: Tvstartup - TSetup - Tcalc - Twait - Tpre - To > 0\n");
+		dml_print(
+				"DML: Tslack(pre): %fus - time left over in schedule\n",
+				VStartup * LineTime - TimeForFetchingMetaPTE - 2 * TimeForFetchingRowInVBlank
+						- (*DSTYAfterScaler + ((double) (*DSTXAfterScaler) / (double) myPipe->HTotal)) * LineTime - TWait - TCalc - *TSetup);
+		dml_print("DML: row_bytes = dpte_row_bytes (per_pipe) = PixelPTEBytesPerRow = : %d\n", PixelPTEBytesPerRow);
+
+	} else {
+		MyError = true;
+		dml_print("DML: MyErr set %s:%d\n", __FILE__, __LINE__);
+	}
+
+	{
+		double prefetch_vm_bw;
+		double prefetch_row_bw;
+
+		if (PDEAndMetaPTEBytesFrame == 0) {
+			prefetch_vm_bw = 0;
+		} else if (*DestinationLinesToRequestVMInVBlank > 0) {
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: PDEAndMetaPTEBytesFrame = %d\n", __func__, PDEAndMetaPTEBytesFrame);
+			dml_print("DML::%s: HostVMInefficiencyFactor = %f\n", __func__, HostVMInefficiencyFactor);
+			dml_print("DML::%s: DestinationLinesToRequestVMInVBlank = %f\n", __func__, *DestinationLinesToRequestVMInVBlank);
+			dml_print("DML::%s: LineTime = %f\n", __func__, LineTime);
+#endif
+			prefetch_vm_bw = PDEAndMetaPTEBytesFrame * HostVMInefficiencyFactor / (*DestinationLinesToRequestVMInVBlank * LineTime);
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: prefetch_vm_bw = %f\n", __func__, prefetch_vm_bw);
+#endif
+		} else {
+			prefetch_vm_bw = 0;
+			MyError = true;
+			dml_print("DML: MyErr set %s:%d\n", __FILE__, __LINE__);
+		}
+
+		if (MetaRowByte + PixelPTEBytesPerRow == 0) {
+			prefetch_row_bw = 0;
+		} else if (*DestinationLinesToRequestRowInVBlank > 0) {
+			prefetch_row_bw = (MetaRowByte + PixelPTEBytesPerRow * HostVMInefficiencyFactor) / (*DestinationLinesToRequestRowInVBlank * LineTime);
+
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: MetaRowByte = %d\n", __func__, MetaRowByte);
+			dml_print("DML::%s: PixelPTEBytesPerRow = %d\n", __func__, PixelPTEBytesPerRow);
+			dml_print("DML::%s: DestinationLinesToRequestRowInVBlank = %f\n", __func__, *DestinationLinesToRequestRowInVBlank);
+			dml_print("DML::%s: prefetch_row_bw = %f\n", __func__, prefetch_row_bw);
+#endif
+		} else {
+			prefetch_row_bw = 0;
+			MyError = true;
+			dml_print("DML: MyErr set %s:%d\n", __FILE__, __LINE__);
+		}
+
+		*prefetch_vmrow_bw = dml_max(prefetch_vm_bw, prefetch_row_bw);
+	}
+
+	if (MyError) {
+		*PrefetchBandwidth = 0;
+		*DestinationLinesToRequestVMInVBlank = 0;
+		*DestinationLinesToRequestRowInVBlank = 0;
+		*DestinationLinesForPrefetch = 0;
+		*VRatioPrefetchY = 0;
+		*VRatioPrefetchC = 0;
+		*RequiredPrefetchPixDataBWLuma = 0;
+		*RequiredPrefetchPixDataBWChroma = 0;
+	}
+
+	return MyError;
+}
+
+static double RoundToDFSGranularityUp(double Clock, double VCOSpeed)
+{
+	return VCOSpeed * 4 / dml_floor(VCOSpeed * 4 / Clock, 1);
+}
+
+static double RoundToDFSGranularityDown(double Clock, double VCOSpeed)
+{
+	return VCOSpeed * 4 / dml_ceil(VCOSpeed * 4.0 / Clock, 1);
+}
+
+static void CalculateDCCConfiguration(
+		bool DCCEnabled,
+		bool DCCProgrammingAssumesScanDirectionUnknown,
+		enum source_format_class SourcePixelFormat,
+		unsigned int SurfaceWidthLuma,
+		unsigned int SurfaceWidthChroma,
+		unsigned int SurfaceHeightLuma,
+		unsigned int SurfaceHeightChroma,
+		double DETBufferSize,
+		unsigned int RequestHeight256ByteLuma,
+		unsigned int RequestHeight256ByteChroma,
+		enum dm_swizzle_mode TilingFormat,
+		unsigned int BytePerPixelY,
+		unsigned int BytePerPixelC,
+		double BytePerPixelDETY,
+		double BytePerPixelDETC,
+		enum scan_direction_class ScanOrientation,
+		unsigned int *MaxUncompressedBlockLuma,
+		unsigned int *MaxUncompressedBlockChroma,
+		unsigned int *MaxCompressedBlockLuma,
+		unsigned int *MaxCompressedBlockChroma,
+		unsigned int *IndependentBlockLuma,
+		unsigned int *IndependentBlockChroma)
+{
+	int yuv420;
+	int horz_div_l;
+	int horz_div_c;
+	int vert_div_l;
+	int vert_div_c;
+
+	int swath_buf_size;
+	double detile_buf_vp_horz_limit;
+	double detile_buf_vp_vert_limit;
+
+	int MAS_vp_horz_limit;
+	int MAS_vp_vert_limit;
+	int max_vp_horz_width;
+	int max_vp_vert_height;
+	int eff_surf_width_l;
+	int eff_surf_width_c;
+	int eff_surf_height_l;
+	int eff_surf_height_c;
+
+	int full_swath_bytes_horz_wc_l;
+	int full_swath_bytes_horz_wc_c;
+	int full_swath_bytes_vert_wc_l;
+	int full_swath_bytes_vert_wc_c;
+	int req128_horz_wc_l;
+	int req128_horz_wc_c;
+	int req128_vert_wc_l;
+	int req128_vert_wc_c;
+	int segment_order_horz_contiguous_luma;
+	int segment_order_horz_contiguous_chroma;
+	int segment_order_vert_contiguous_luma;
+	int segment_order_vert_contiguous_chroma;
+
+	typedef enum {
+		REQ_256Bytes, REQ_128BytesNonContiguous, REQ_128BytesContiguous, REQ_NA
+	} RequestType;
+	RequestType RequestLuma;
+	RequestType RequestChroma;
+
+	yuv420 = ((SourcePixelFormat == dm_420_8 || SourcePixelFormat == dm_420_10 || SourcePixelFormat == dm_420_12) ? 1 : 0);
+	horz_div_l = 1;
+	horz_div_c = 1;
+	vert_div_l = 1;
+	vert_div_c = 1;
+
+	if (BytePerPixelY == 1)
+		vert_div_l = 0;
+	if (BytePerPixelC == 1)
+		vert_div_c = 0;
+	if (BytePerPixelY == 8 && (TilingFormat == dm_sw_64kb_s || TilingFormat == dm_sw_64kb_s_t || TilingFormat == dm_sw_64kb_s_x))
+		horz_div_l = 0;
+	if (BytePerPixelC == 8 && (TilingFormat == dm_sw_64kb_s || TilingFormat == dm_sw_64kb_s_t || TilingFormat == dm_sw_64kb_s_x))
+		horz_div_c = 0;
+
+	if (BytePerPixelC == 0) {
+		swath_buf_size = DETBufferSize / 2 - 2 * 256;
+		detile_buf_vp_horz_limit = (double) swath_buf_size / ((double) RequestHeight256ByteLuma * BytePerPixelY / (1 + horz_div_l));
+		detile_buf_vp_vert_limit = (double) swath_buf_size / (256.0 / RequestHeight256ByteLuma / (1 + vert_div_l));
+	} else {
+		swath_buf_size = DETBufferSize / 2 - 2 * 2 * 256;
+		detile_buf_vp_horz_limit = (double) swath_buf_size
+				/ ((double) RequestHeight256ByteLuma * BytePerPixelY / (1 + horz_div_l)
+						+ (double) RequestHeight256ByteChroma * BytePerPixelC / (1 + horz_div_c) / (1 + yuv420));
+		detile_buf_vp_vert_limit = (double) swath_buf_size
+				/ (256.0 / RequestHeight256ByteLuma / (1 + vert_div_l) + 256.0 / RequestHeight256ByteChroma / (1 + vert_div_c) / (1 + yuv420));
+	}
+
+	if (SourcePixelFormat == dm_420_10) {
+		detile_buf_vp_horz_limit = 1.5 * detile_buf_vp_horz_limit;
+		detile_buf_vp_vert_limit = 1.5 * detile_buf_vp_vert_limit;
+	}
+
+	detile_buf_vp_horz_limit = dml_floor(detile_buf_vp_horz_limit - 1, 16);
+	detile_buf_vp_vert_limit = dml_floor(detile_buf_vp_vert_limit - 1, 16);
+
+	MAS_vp_horz_limit = SourcePixelFormat == dm_rgbe_alpha ? 3840 : 5760;
+	MAS_vp_vert_limit = (BytePerPixelC > 0 ? 2880 : 5760);
+	max_vp_horz_width = dml_min((double) MAS_vp_horz_limit, detile_buf_vp_horz_limit);
+	max_vp_vert_height = dml_min((double) MAS_vp_vert_limit, detile_buf_vp_vert_limit);
+	eff_surf_width_l = (SurfaceWidthLuma > max_vp_horz_width ? max_vp_horz_width : SurfaceWidthLuma);
+	eff_surf_width_c = eff_surf_width_l / (1 + yuv420);
+	eff_surf_height_l = (SurfaceHeightLuma > max_vp_vert_height ? max_vp_vert_height : SurfaceHeightLuma);
+	eff_surf_height_c = eff_surf_height_l / (1 + yuv420);
+
+	full_swath_bytes_horz_wc_l = eff_surf_width_l * RequestHeight256ByteLuma * BytePerPixelY;
+	full_swath_bytes_vert_wc_l = eff_surf_height_l * 256 / RequestHeight256ByteLuma;
+	if (BytePerPixelC > 0) {
+		full_swath_bytes_horz_wc_c = eff_surf_width_c * RequestHeight256ByteChroma * BytePerPixelC;
+		full_swath_bytes_vert_wc_c = eff_surf_height_c * 256 / RequestHeight256ByteChroma;
+	} else {
+		full_swath_bytes_horz_wc_c = 0;
+		full_swath_bytes_vert_wc_c = 0;
+	}
+
+	if (SourcePixelFormat == dm_420_10) {
+		full_swath_bytes_horz_wc_l = dml_ceil(full_swath_bytes_horz_wc_l * 2 / 3, 256);
+		full_swath_bytes_horz_wc_c = dml_ceil(full_swath_bytes_horz_wc_c * 2 / 3, 256);
+		full_swath_bytes_vert_wc_l = dml_ceil(full_swath_bytes_vert_wc_l * 2 / 3, 256);
+		full_swath_bytes_vert_wc_c = dml_ceil(full_swath_bytes_vert_wc_c * 2 / 3, 256);
+	}
+
+	if (2 * full_swath_bytes_horz_wc_l + 2 * full_swath_bytes_horz_wc_c <= DETBufferSize) {
+		req128_horz_wc_l = 0;
+		req128_horz_wc_c = 0;
+	} else if (full_swath_bytes_horz_wc_l < 1.5 * full_swath_bytes_horz_wc_c && 2 * full_swath_bytes_horz_wc_l + full_swath_bytes_horz_wc_c <= DETBufferSize) {
+		req128_horz_wc_l = 0;
+		req128_horz_wc_c = 1;
+	} else if (full_swath_bytes_horz_wc_l >= 1.5 * full_swath_bytes_horz_wc_c && full_swath_bytes_horz_wc_l + 2 * full_swath_bytes_horz_wc_c <= DETBufferSize) {
+		req128_horz_wc_l = 1;
+		req128_horz_wc_c = 0;
+	} else {
+		req128_horz_wc_l = 1;
+		req128_horz_wc_c = 1;
+	}
+
+	if (2 * full_swath_bytes_vert_wc_l + 2 * full_swath_bytes_vert_wc_c <= DETBufferSize) {
+		req128_vert_wc_l = 0;
+		req128_vert_wc_c = 0;
+	} else if (full_swath_bytes_vert_wc_l < 1.5 * full_swath_bytes_vert_wc_c && 2 * full_swath_bytes_vert_wc_l + full_swath_bytes_vert_wc_c <= DETBufferSize) {
+		req128_vert_wc_l = 0;
+		req128_vert_wc_c = 1;
+	} else if (full_swath_bytes_vert_wc_l >= 1.5 * full_swath_bytes_vert_wc_c && full_swath_bytes_vert_wc_l + 2 * full_swath_bytes_vert_wc_c <= DETBufferSize) {
+		req128_vert_wc_l = 1;
+		req128_vert_wc_c = 0;
+	} else {
+		req128_vert_wc_l = 1;
+		req128_vert_wc_c = 1;
+	}
+
+	if (BytePerPixelY == 2 || (BytePerPixelY == 4 && TilingFormat != dm_sw_64kb_r_x)) {
+		segment_order_horz_contiguous_luma = 0;
+	} else {
+		segment_order_horz_contiguous_luma = 1;
+	}
+	if ((BytePerPixelY == 8 && (TilingFormat == dm_sw_64kb_d || TilingFormat == dm_sw_64kb_d_x || TilingFormat == dm_sw_64kb_d_t || TilingFormat == dm_sw_64kb_r_x))
+			|| (BytePerPixelY == 4 && TilingFormat == dm_sw_64kb_r_x)) {
+		segment_order_vert_contiguous_luma = 0;
+	} else {
+		segment_order_vert_contiguous_luma = 1;
+	}
+	if (BytePerPixelC == 2 || (BytePerPixelC == 4 && TilingFormat != dm_sw_64kb_r_x)) {
+		segment_order_horz_contiguous_chroma = 0;
+	} else {
+		segment_order_horz_contiguous_chroma = 1;
+	}
+	if ((BytePerPixelC == 8 && (TilingFormat == dm_sw_64kb_d || TilingFormat == dm_sw_64kb_d_x || TilingFormat == dm_sw_64kb_d_t || TilingFormat == dm_sw_64kb_r_x))
+			|| (BytePerPixelC == 4 && TilingFormat == dm_sw_64kb_r_x)) {
+		segment_order_vert_contiguous_chroma = 0;
+	} else {
+		segment_order_vert_contiguous_chroma = 1;
+	}
+
+	if (DCCProgrammingAssumesScanDirectionUnknown == true) {
+		if (req128_horz_wc_l == 0 && req128_vert_wc_l == 0) {
+			RequestLuma = REQ_256Bytes;
+		} else if ((req128_horz_wc_l == 1 && segment_order_horz_contiguous_luma == 0) || (req128_vert_wc_l == 1 && segment_order_vert_contiguous_luma == 0)) {
+			RequestLuma = REQ_128BytesNonContiguous;
+		} else {
+			RequestLuma = REQ_128BytesContiguous;
+		}
+		if (req128_horz_wc_c == 0 && req128_vert_wc_c == 0) {
+			RequestChroma = REQ_256Bytes;
+		} else if ((req128_horz_wc_c == 1 && segment_order_horz_contiguous_chroma == 0) || (req128_vert_wc_c == 1 && segment_order_vert_contiguous_chroma == 0)) {
+			RequestChroma = REQ_128BytesNonContiguous;
+		} else {
+			RequestChroma = REQ_128BytesContiguous;
+		}
+	} else if (ScanOrientation != dm_vert) {
+		if (req128_horz_wc_l == 0) {
+			RequestLuma = REQ_256Bytes;
+		} else if (segment_order_horz_contiguous_luma == 0) {
+			RequestLuma = REQ_128BytesNonContiguous;
+		} else {
+			RequestLuma = REQ_128BytesContiguous;
+		}
+		if (req128_horz_wc_c == 0) {
+			RequestChroma = REQ_256Bytes;
+		} else if (segment_order_horz_contiguous_chroma == 0) {
+			RequestChroma = REQ_128BytesNonContiguous;
+		} else {
+			RequestChroma = REQ_128BytesContiguous;
+		}
+	} else {
+		if (req128_vert_wc_l == 0) {
+			RequestLuma = REQ_256Bytes;
+		} else if (segment_order_vert_contiguous_luma == 0) {
+			RequestLuma = REQ_128BytesNonContiguous;
+		} else {
+			RequestLuma = REQ_128BytesContiguous;
+		}
+		if (req128_vert_wc_c == 0) {
+			RequestChroma = REQ_256Bytes;
+		} else if (segment_order_vert_contiguous_chroma == 0) {
+			RequestChroma = REQ_128BytesNonContiguous;
+		} else {
+			RequestChroma = REQ_128BytesContiguous;
+		}
+	}
+
+	if (RequestLuma == REQ_256Bytes) {
+		*MaxUncompressedBlockLuma = 256;
+		*MaxCompressedBlockLuma = 256;
+		*IndependentBlockLuma = 0;
+	} else if (RequestLuma == REQ_128BytesContiguous) {
+		*MaxUncompressedBlockLuma = 256;
+		*MaxCompressedBlockLuma = 128;
+		*IndependentBlockLuma = 128;
+	} else {
+		*MaxUncompressedBlockLuma = 256;
+		*MaxCompressedBlockLuma = 64;
+		*IndependentBlockLuma = 64;
+	}
+
+	if (RequestChroma == REQ_256Bytes) {
+		*MaxUncompressedBlockChroma = 256;
+		*MaxCompressedBlockChroma = 256;
+		*IndependentBlockChroma = 0;
+	} else if (RequestChroma == REQ_128BytesContiguous) {
+		*MaxUncompressedBlockChroma = 256;
+		*MaxCompressedBlockChroma = 128;
+		*IndependentBlockChroma = 128;
+	} else {
+		*MaxUncompressedBlockChroma = 256;
+		*MaxCompressedBlockChroma = 64;
+		*IndependentBlockChroma = 64;
+	}
+
+	if (DCCEnabled != true || BytePerPixelC == 0) {
+		*MaxUncompressedBlockChroma = 0;
+		*MaxCompressedBlockChroma = 0;
+		*IndependentBlockChroma = 0;
+	}
+
+	if (DCCEnabled != true) {
+		*MaxUncompressedBlockLuma = 0;
+		*MaxCompressedBlockLuma = 0;
+		*IndependentBlockLuma = 0;
+	}
+}
+
+static double CalculatePrefetchSourceLines(
+		struct display_mode_lib *mode_lib,
+		double VRatio,
+		double vtaps,
+		bool Interlace,
+		bool ProgressiveToInterlaceUnitInOPP,
+		unsigned int SwathHeight,
+		unsigned int ViewportYStart,
+		double *VInitPreFill,
+		unsigned int *MaxNumSwath)
+{
+	struct vba_vars_st *v = &mode_lib->vba;
+	unsigned int MaxPartialSwath;
+
+	if (ProgressiveToInterlaceUnitInOPP)
+		*VInitPreFill = dml_floor((VRatio + vtaps + 1) / 2.0, 1);
+	else
+		*VInitPreFill = dml_floor((VRatio + vtaps + 1 + Interlace * 0.5 * VRatio) / 2.0, 1);
+
+	if (!v->IgnoreViewportPositioning) {
+
+		*MaxNumSwath = dml_ceil((*VInitPreFill - 1.0) / SwathHeight, 1) + 1.0;
+
+		if (*VInitPreFill > 1.0)
+			MaxPartialSwath = (unsigned int) (*VInitPreFill - 2) % SwathHeight;
+		else
+			MaxPartialSwath = (unsigned int) (*VInitPreFill + SwathHeight - 2) % SwathHeight;
+		MaxPartialSwath = dml_max(1U, MaxPartialSwath);
+
+	} else {
+
+		if (ViewportYStart != 0)
+			dml_print("WARNING DML: using viewport y position of 0 even though actual viewport y position is non-zero in prefetch source lines calculation\n");
+
+		*MaxNumSwath = dml_ceil(*VInitPreFill / SwathHeight, 1);
+
+		if (*VInitPreFill > 1.0)
+			MaxPartialSwath = (unsigned int) (*VInitPreFill - 1) % SwathHeight;
+		else
+			MaxPartialSwath = (unsigned int) (*VInitPreFill + SwathHeight - 1) % SwathHeight;
+	}
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: VRatio = %f\n", __func__, VRatio);
+	dml_print("DML::%s: vtaps = %f\n", __func__, vtaps);
+	dml_print("DML::%s: VInitPreFill = %f\n", __func__, *VInitPreFill);
+	dml_print("DML::%s: ProgressiveToInterlaceUnitInOPP = %d\n", __func__, ProgressiveToInterlaceUnitInOPP);
+	dml_print("DML::%s: IgnoreViewportPositioning = %d\n", __func__, v->IgnoreViewportPositioning);
+	dml_print("DML::%s: SwathHeight = %d\n", __func__, SwathHeight);
+	dml_print("DML::%s: MaxPartialSwath = %d\n", __func__, MaxPartialSwath);
+	dml_print("DML::%s: MaxNumSwath = %d\n", __func__, *MaxNumSwath);
+	dml_print("DML::%s: Prefetch source lines = %d\n", __func__, *MaxNumSwath * SwathHeight + MaxPartialSwath);
+#endif
+	return *MaxNumSwath * SwathHeight + MaxPartialSwath;
+}
+
+static unsigned int CalculateVMAndRowBytes(
+		struct display_mode_lib *mode_lib,
+		bool DCCEnable,
+		unsigned int BlockHeight256Bytes,
+		unsigned int BlockWidth256Bytes,
+		enum source_format_class SourcePixelFormat,
+		unsigned int SurfaceTiling,
+		unsigned int BytePerPixel,
+		enum scan_direction_class ScanDirection,
+		unsigned int SwathWidth,
+		unsigned int ViewportHeight,
+		bool GPUVMEnable,
+		bool HostVMEnable,
+		unsigned int HostVMMaxNonCachedPageTableLevels,
+		unsigned int GPUVMMinPageSize,
+		unsigned int HostVMMinPageSize,
+		unsigned int PTEBufferSizeInRequests,
+		unsigned int Pitch,
+		unsigned int DCCMetaPitch,
+		unsigned int *MacroTileWidth,
+		unsigned int *MetaRowByte,
+		unsigned int *PixelPTEBytesPerRow,
+		bool *PTEBufferSizeNotExceeded,
+		int *dpte_row_width_ub,
+		unsigned int *dpte_row_height,
+		unsigned int *MetaRequestWidth,
+		unsigned int *MetaRequestHeight,
+		unsigned int *meta_row_width,
+		unsigned int *meta_row_height,
+		int *vm_group_bytes,
+		unsigned int *dpte_group_bytes,
+		unsigned int *PixelPTEReqWidth,
+		unsigned int *PixelPTEReqHeight,
+		unsigned int *PTERequestSize,
+		int *DPDE0BytesFrame,
+		int *MetaPTEBytesFrame)
+{
+	struct vba_vars_st *v = &mode_lib->vba;
+	unsigned int MPDEBytesFrame;
+	unsigned int DCCMetaSurfaceBytes;
+	unsigned int MacroTileSizeBytes;
+	unsigned int MacroTileHeight;
+	unsigned int ExtraDPDEBytesFrame;
+	unsigned int PDEAndMetaPTEBytesFrame;
+	unsigned int PixelPTEReqHeightPTEs = 0;
+	unsigned int HostVMDynamicLevels = 0;
+	double FractionOfPTEReturnDrop;
+
+	if (GPUVMEnable == true && HostVMEnable == true) {
+		if (HostVMMinPageSize < 2048) {
+			HostVMDynamicLevels = HostVMMaxNonCachedPageTableLevels;
+		} else if (HostVMMinPageSize >= 2048 && HostVMMinPageSize < 1048576) {
+			HostVMDynamicLevels = dml_max(0, (int) HostVMMaxNonCachedPageTableLevels - 1);
+		} else {
+			HostVMDynamicLevels = dml_max(0, (int) HostVMMaxNonCachedPageTableLevels - 2);
+		}
+	}
+
+	*MetaRequestHeight = 8 * BlockHeight256Bytes;
+	*MetaRequestWidth = 8 * BlockWidth256Bytes;
+	if (ScanDirection != dm_vert) {
+		*meta_row_height = *MetaRequestHeight;
+		*meta_row_width = dml_ceil((double) SwathWidth - 1, *MetaRequestWidth) + *MetaRequestWidth;
+		*MetaRowByte = *meta_row_width * *MetaRequestHeight * BytePerPixel / 256.0;
+	} else {
+		*meta_row_height = *MetaRequestWidth;
+		*meta_row_width = dml_ceil((double) SwathWidth - 1, *MetaRequestHeight) + *MetaRequestHeight;
+		*MetaRowByte = *meta_row_width * *MetaRequestWidth * BytePerPixel / 256.0;
+	}
+	DCCMetaSurfaceBytes = DCCMetaPitch * (dml_ceil(ViewportHeight - 1, 64 * BlockHeight256Bytes) + 64 * BlockHeight256Bytes) * BytePerPixel / 256;
+	if (GPUVMEnable == true) {
+		*MetaPTEBytesFrame = (dml_ceil((double) (DCCMetaSurfaceBytes - 4.0 * 1024.0) / (8 * 4.0 * 1024), 1) + 1) * 64;
+		MPDEBytesFrame = 128 * (v->GPUVMMaxPageTableLevels - 1);
+	} else {
+		*MetaPTEBytesFrame = 0;
+		MPDEBytesFrame = 0;
+	}
+
+	if (DCCEnable != true) {
+		*MetaPTEBytesFrame = 0;
+		MPDEBytesFrame = 0;
+		*MetaRowByte = 0;
+	}
+
+	if (SurfaceTiling == dm_sw_linear) {
+		MacroTileSizeBytes = 256;
+		MacroTileHeight = BlockHeight256Bytes;
+	} else {
+		MacroTileSizeBytes = 65536;
+		MacroTileHeight = 16 * BlockHeight256Bytes;
+	}
+	*MacroTileWidth = MacroTileSizeBytes / BytePerPixel / MacroTileHeight;
+
+	if (GPUVMEnable == true && v->GPUVMMaxPageTableLevels > 1) {
+		if (ScanDirection != dm_vert) {
+			*DPDE0BytesFrame = 64
+					* (dml_ceil(
+							((Pitch * (dml_ceil(ViewportHeight - 1, MacroTileHeight) + MacroTileHeight) * BytePerPixel) - MacroTileSizeBytes)
+									/ (8 * 2097152),
+							1) + 1);
+		} else {
+			*DPDE0BytesFrame = 64
+					* (dml_ceil(
+							((Pitch * (dml_ceil((double) SwathWidth - 1, MacroTileHeight) + MacroTileHeight) * BytePerPixel) - MacroTileSizeBytes)
+									/ (8 * 2097152),
+							1) + 1);
+		}
+		ExtraDPDEBytesFrame = 128 * (v->GPUVMMaxPageTableLevels - 2);
+	} else {
+		*DPDE0BytesFrame = 0;
+		ExtraDPDEBytesFrame = 0;
+	}
+
+	PDEAndMetaPTEBytesFrame = *MetaPTEBytesFrame + MPDEBytesFrame + *DPDE0BytesFrame + ExtraDPDEBytesFrame;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: MetaPTEBytesFrame = %d\n", __func__, *MetaPTEBytesFrame);
+	dml_print("DML::%s: MPDEBytesFrame = %d\n", __func__, MPDEBytesFrame);
+	dml_print("DML::%s: DPDE0BytesFrame = %d\n", __func__, *DPDE0BytesFrame);
+	dml_print("DML::%s: ExtraDPDEBytesFrame= %d\n", __func__, ExtraDPDEBytesFrame);
+	dml_print("DML::%s: PDEAndMetaPTEBytesFrame = %d\n", __func__, PDEAndMetaPTEBytesFrame);
+#endif
+
+	if (HostVMEnable == true) {
+		PDEAndMetaPTEBytesFrame = PDEAndMetaPTEBytesFrame * (1 + 8 * HostVMDynamicLevels);
+	}
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: PDEAndMetaPTEBytesFrame = %d\n", __func__, PDEAndMetaPTEBytesFrame);
+#endif
+
+	if (SurfaceTiling == dm_sw_linear) {
+		PixelPTEReqHeightPTEs = 1;
+		*PixelPTEReqHeight = 1;
+		*PixelPTEReqWidth = 32768.0 / BytePerPixel;
+		*PTERequestSize = 64;
+		FractionOfPTEReturnDrop = 0;
+	} else if (GPUVMMinPageSize == 4 && MacroTileSizeBytes > 4096) {
+		PixelPTEReqHeightPTEs = 16;
+		*PixelPTEReqHeight = 16 * BlockHeight256Bytes;
+		*PixelPTEReqWidth = 16 * BlockWidth256Bytes;
+		*PTERequestSize = 128;
+		FractionOfPTEReturnDrop = 0;
+	} else {
+		PixelPTEReqHeightPTEs = 1;
+		*PixelPTEReqHeight = MacroTileHeight;
+		*PixelPTEReqWidth = 8 * *MacroTileWidth;
+		*PTERequestSize = 64;
+		FractionOfPTEReturnDrop = 0;
+	}
+
+	if (SurfaceTiling == dm_sw_linear) {
+		*dpte_row_height = dml_min(128, 1 << (unsigned int) dml_floor(dml_log2(PTEBufferSizeInRequests * *PixelPTEReqWidth / Pitch), 1));
+		*dpte_row_width_ub = (dml_ceil((double)(Pitch * *dpte_row_height - 1) / *PixelPTEReqWidth, 1) + 1) * *PixelPTEReqWidth;
+		*PixelPTEBytesPerRow = *dpte_row_width_ub / *PixelPTEReqWidth * *PTERequestSize;
+	} else if (ScanDirection != dm_vert) {
+		*dpte_row_height = *PixelPTEReqHeight;
+		*dpte_row_width_ub = (dml_ceil((double) (SwathWidth - 1) / *PixelPTEReqWidth, 1) + 1) * *PixelPTEReqWidth;
+		*PixelPTEBytesPerRow = *dpte_row_width_ub / *PixelPTEReqWidth * *PTERequestSize;
+	} else {
+		*dpte_row_height = dml_min(*PixelPTEReqWidth, *MacroTileWidth);
+		*dpte_row_width_ub = (dml_ceil((double) (SwathWidth - 1) / *PixelPTEReqHeight, 1) + 1) * *PixelPTEReqHeight;
+		*PixelPTEBytesPerRow = *dpte_row_width_ub / *PixelPTEReqHeight * *PTERequestSize;
+	}
+
+	if (*PixelPTEBytesPerRow * (1 - FractionOfPTEReturnDrop) <= 64 * PTEBufferSizeInRequests) {
+		*PTEBufferSizeNotExceeded = true;
+	} else {
+		*PTEBufferSizeNotExceeded = false;
+	}
+
+	if (GPUVMEnable != true) {
+		*PixelPTEBytesPerRow = 0;
+		*PTEBufferSizeNotExceeded = true;
+	}
+
+	dml_print("DML: vm_bytes = meta_pte_bytes_per_frame (per_pipe) = MetaPTEBytesFrame = : %i\n", *MetaPTEBytesFrame);
+
+	if (HostVMEnable == true) {
+		*PixelPTEBytesPerRow = *PixelPTEBytesPerRow * (1 + 8 * HostVMDynamicLevels);
+	}
+
+	if (HostVMEnable == true) {
+		*vm_group_bytes = 512;
+		*dpte_group_bytes = 512;
+	} else if (GPUVMEnable == true) {
+		*vm_group_bytes = 2048;
+		if (SurfaceTiling != dm_sw_linear && PixelPTEReqHeightPTEs == 1 && ScanDirection == dm_vert) {
+			*dpte_group_bytes = 512;
+		} else {
+			*dpte_group_bytes = 2048;
+		}
+	} else {
+		*vm_group_bytes = 0;
+		*dpte_group_bytes = 0;
+	}
+	return PDEAndMetaPTEBytesFrame;
+}
+
+static void DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation(struct display_mode_lib *mode_lib)
+{
+	struct vba_vars_st *v = &mode_lib->vba;
+	unsigned int j, k;
+	double HostVMInefficiencyFactor = 1.0;
+	bool NoChromaPlanes = true;
+	int ReorderBytes;
+	double VMDataOnlyReturnBW;
+	double MaxTotalRDBandwidth = 0;
+	int PrefetchMode = v->PrefetchModePerState[v->VoltageLevel][v->maxMpcComb];
+
+	v->WritebackDISPCLK = 0.0;
+	v->DISPCLKWithRamping = 0;
+	v->DISPCLKWithoutRamping = 0;
+	v->GlobalDPPCLK = 0.0;
+	/* DAL custom code: need to update ReturnBW in case min dcfclk is overridden */
+	{
+	double IdealFabricAndSDPPortBandwidthPerState = dml_min(
+			v->ReturnBusWidth * v->DCFCLKState[v->VoltageLevel][v->maxMpcComb],
+			v->FabricClockPerState[v->VoltageLevel] * v->FabricDatapathToDCNDataReturn);
+	double IdealDRAMBandwidthPerState = v->DRAMSpeedPerState[v->VoltageLevel] * v->NumberOfChannels * v->DRAMChannelWidth;
+
+	if (v->HostVMEnable != true) {
+		v->ReturnBW = dml_min(
+				IdealFabricAndSDPPortBandwidthPerState * v->PercentOfIdealFabricAndSDPPortBWReceivedAfterUrgLatency / 100.0,
+				IdealDRAMBandwidthPerState * v->PercentOfIdealDRAMBWReceivedAfterUrgLatencyPixelDataOnly / 100.0);
+	} else {
+		v->ReturnBW = dml_min(
+				IdealFabricAndSDPPortBandwidthPerState * v->PercentOfIdealFabricAndSDPPortBWReceivedAfterUrgLatency / 100.0,
+				IdealDRAMBandwidthPerState * v->PercentOfIdealDRAMBWReceivedAfterUrgLatencyPixelMixedWithVMData / 100.0);
+	}
+	}
+	/* End DAL custom code */
+
+	// DISPCLK and DPPCLK Calculation
+	//
+	for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+		if (v->WritebackEnable[k]) {
+			v->WritebackDISPCLK = dml_max(
+					v->WritebackDISPCLK,
+					dml314_CalculateWriteBackDISPCLK(
+							v->WritebackPixelFormat[k],
+							v->PixelClock[k],
+							v->WritebackHRatio[k],
+							v->WritebackVRatio[k],
+							v->WritebackHTaps[k],
+							v->WritebackVTaps[k],
+							v->WritebackSourceWidth[k],
+							v->WritebackDestinationWidth[k],
+							v->HTotal[k],
+							v->WritebackLineBufferSize));
+		}
+	}
+
+	for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+		if (v->HRatio[k] > 1) {
+			v->PSCL_THROUGHPUT_LUMA[k] = dml_min(
+					v->MaxDCHUBToPSCLThroughput,
+					v->MaxPSCLToLBThroughput * v->HRatio[k] / dml_ceil(v->htaps[k] / 6.0, 1));
+		} else {
+			v->PSCL_THROUGHPUT_LUMA[k] = dml_min(v->MaxDCHUBToPSCLThroughput, v->MaxPSCLToLBThroughput);
+		}
+
+		v->DPPCLKUsingSingleDPPLuma = v->PixelClock[k]
+				* dml_max(
+						v->vtaps[k] / 6.0 * dml_min(1.0, v->HRatio[k]),
+						dml_max(v->HRatio[k] * v->VRatio[k] / v->PSCL_THROUGHPUT_LUMA[k], 1.0));
+
+		if ((v->htaps[k] > 6 || v->vtaps[k] > 6) && v->DPPCLKUsingSingleDPPLuma < 2 * v->PixelClock[k]) {
+			v->DPPCLKUsingSingleDPPLuma = 2 * v->PixelClock[k];
+		}
+
+		if ((v->SourcePixelFormat[k] != dm_420_8 && v->SourcePixelFormat[k] != dm_420_10 && v->SourcePixelFormat[k] != dm_420_12
+				&& v->SourcePixelFormat[k] != dm_rgbe_alpha)) {
+			v->PSCL_THROUGHPUT_CHROMA[k] = 0.0;
+			v->DPPCLKUsingSingleDPP[k] = v->DPPCLKUsingSingleDPPLuma;
+		} else {
+			if (v->HRatioChroma[k] > 1) {
+				v->PSCL_THROUGHPUT_CHROMA[k] = dml_min(
+						v->MaxDCHUBToPSCLThroughput,
+						v->MaxPSCLToLBThroughput * v->HRatioChroma[k] / dml_ceil(v->HTAPsChroma[k] / 6.0, 1.0));
+			} else {
+				v->PSCL_THROUGHPUT_CHROMA[k] = dml_min(v->MaxDCHUBToPSCLThroughput, v->MaxPSCLToLBThroughput);
+			}
+			v->DPPCLKUsingSingleDPPChroma = v->PixelClock[k]
+					* dml_max3(
+							v->VTAPsChroma[k] / 6.0 * dml_min(1.0, v->HRatioChroma[k]),
+							v->HRatioChroma[k] * v->VRatioChroma[k] / v->PSCL_THROUGHPUT_CHROMA[k],
+							1.0);
+
+			if ((v->HTAPsChroma[k] > 6 || v->VTAPsChroma[k] > 6) && v->DPPCLKUsingSingleDPPChroma < 2 * v->PixelClock[k]) {
+				v->DPPCLKUsingSingleDPPChroma = 2 * v->PixelClock[k];
+			}
+
+			v->DPPCLKUsingSingleDPP[k] = dml_max(v->DPPCLKUsingSingleDPPLuma, v->DPPCLKUsingSingleDPPChroma);
+		}
+	}
+
+	for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+		if (v->BlendingAndTiming[k] != k)
+			continue;
+		if (v->ODMCombineEnabled[k] == dm_odm_combine_mode_4to1) {
+			v->DISPCLKWithRamping = dml_max(
+					v->DISPCLKWithRamping,
+					v->PixelClock[k] / 4 * (1 + v->DISPCLKDPPCLKDSCCLKDownSpreading / 100)
+							* (1 + v->DISPCLKRampingMargin / 100));
+			v->DISPCLKWithoutRamping = dml_max(
+					v->DISPCLKWithoutRamping,
+					v->PixelClock[k] / 4 * (1 + v->DISPCLKDPPCLKDSCCLKDownSpreading / 100));
+		} else if (v->ODMCombineEnabled[k] == dm_odm_combine_mode_2to1) {
+			v->DISPCLKWithRamping = dml_max(
+					v->DISPCLKWithRamping,
+					v->PixelClock[k] / 2 * (1 + v->DISPCLKDPPCLKDSCCLKDownSpreading / 100)
+							* (1 + v->DISPCLKRampingMargin / 100));
+			v->DISPCLKWithoutRamping = dml_max(
+					v->DISPCLKWithoutRamping,
+					v->PixelClock[k] / 2 * (1 + v->DISPCLKDPPCLKDSCCLKDownSpreading / 100));
+		} else {
+			v->DISPCLKWithRamping = dml_max(
+					v->DISPCLKWithRamping,
+					v->PixelClock[k] * (1 + v->DISPCLKDPPCLKDSCCLKDownSpreading / 100) * (1 + v->DISPCLKRampingMargin / 100));
+			v->DISPCLKWithoutRamping = dml_max(
+					v->DISPCLKWithoutRamping,
+					v->PixelClock[k] * (1 + v->DISPCLKDPPCLKDSCCLKDownSpreading / 100));
+		}
+	}
+
+	v->DISPCLKWithRamping = dml_max(v->DISPCLKWithRamping, v->WritebackDISPCLK);
+	v->DISPCLKWithoutRamping = dml_max(v->DISPCLKWithoutRamping, v->WritebackDISPCLK);
+
+	ASSERT(v->DISPCLKDPPCLKVCOSpeed != 0);
+	v->DISPCLKWithRampingRoundedToDFSGranularity = RoundToDFSGranularityUp(v->DISPCLKWithRamping, v->DISPCLKDPPCLKVCOSpeed);
+	v->DISPCLKWithoutRampingRoundedToDFSGranularity = RoundToDFSGranularityUp(v->DISPCLKWithoutRamping, v->DISPCLKDPPCLKVCOSpeed);
+	v->MaxDispclkRoundedToDFSGranularity = RoundToDFSGranularityDown(
+			v->soc.clock_limits[v->soc.num_states - 1].dispclk_mhz,
+			v->DISPCLKDPPCLKVCOSpeed);
+	if (v->DISPCLKWithoutRampingRoundedToDFSGranularity > v->MaxDispclkRoundedToDFSGranularity) {
+		v->DISPCLK_calculated = v->DISPCLKWithoutRampingRoundedToDFSGranularity;
+	} else if (v->DISPCLKWithRampingRoundedToDFSGranularity > v->MaxDispclkRoundedToDFSGranularity) {
+		v->DISPCLK_calculated = v->MaxDispclkRoundedToDFSGranularity;
+	} else {
+		v->DISPCLK_calculated = v->DISPCLKWithRampingRoundedToDFSGranularity;
+	}
+	v->DISPCLK = v->DISPCLK_calculated;
+	DTRACE("   dispclk_mhz (calculated) = %f", v->DISPCLK_calculated);
+
+	for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+		v->DPPCLK_calculated[k] = v->DPPCLKUsingSingleDPP[k] / v->DPPPerPlane[k] * (1 + v->DISPCLKDPPCLKDSCCLKDownSpreading / 100);
+		v->GlobalDPPCLK = dml_max(v->GlobalDPPCLK, v->DPPCLK_calculated[k]);
+	}
+	v->GlobalDPPCLK = RoundToDFSGranularityUp(v->GlobalDPPCLK, v->DISPCLKDPPCLKVCOSpeed);
+	for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+		v->DPPCLK_calculated[k] = v->GlobalDPPCLK / 255 * dml_ceil(v->DPPCLK_calculated[k] * 255.0 / v->GlobalDPPCLK, 1);
+		DTRACE("   dppclk_mhz[%i] (calculated) = %f", k, v->DPPCLK_calculated[k]);
+	}
+
+	for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+		v->DPPCLK[k] = v->DPPCLK_calculated[k];
+	}
+
+	// Urgent and B P-State/DRAM Clock Change Watermark
+	DTRACE("   dcfclk_mhz         = %f", v->DCFCLK);
+	DTRACE("   return_bus_bw      = %f", v->ReturnBW);
+
+	for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+		CalculateBytePerPixelAnd256BBlockSizes(
+				v->SourcePixelFormat[k],
+				v->SurfaceTiling[k],
+				&v->BytePerPixelY[k],
+				&v->BytePerPixelC[k],
+				&v->BytePerPixelDETY[k],
+				&v->BytePerPixelDETC[k],
+				&v->BlockHeight256BytesY[k],
+				&v->BlockHeight256BytesC[k],
+				&v->BlockWidth256BytesY[k],
+				&v->BlockWidth256BytesC[k]);
+	}
+
+	CalculateSwathWidth(
+			false,
+			v->NumberOfActivePlanes,
+			v->SourcePixelFormat,
+			v->SourceScan,
+			v->ViewportWidth,
+			v->ViewportHeight,
+			v->SurfaceWidthY,
+			v->SurfaceWidthC,
+			v->SurfaceHeightY,
+			v->SurfaceHeightC,
+			v->ODMCombineEnabled,
+			v->BytePerPixelY,
+			v->BytePerPixelC,
+			v->BlockHeight256BytesY,
+			v->BlockHeight256BytesC,
+			v->BlockWidth256BytesY,
+			v->BlockWidth256BytesC,
+			v->BlendingAndTiming,
+			v->HActive,
+			v->HRatio,
+			v->DPPPerPlane,
+			v->SwathWidthSingleDPPY,
+			v->SwathWidthSingleDPPC,
+			v->SwathWidthY,
+			v->SwathWidthC,
+			v->dummyinteger3,
+			v->dummyinteger4,
+			v->swath_width_luma_ub,
+			v->swath_width_chroma_ub);
+
+	for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+		v->ReadBandwidthPlaneLuma[k] = v->SwathWidthSingleDPPY[k] * v->BytePerPixelY[k] / (v->HTotal[k] / v->PixelClock[k])
+				* v->VRatio[k];
+		v->ReadBandwidthPlaneChroma[k] = v->SwathWidthSingleDPPC[k] * v->BytePerPixelC[k] / (v->HTotal[k] / v->PixelClock[k])
+				* v->VRatioChroma[k];
+		DTRACE("   read_bw[%i] = %fBps", k, v->ReadBandwidthPlaneLuma[k] + v->ReadBandwidthPlaneChroma[k]);
+	}
+
+	// DCFCLK Deep Sleep
+	CalculateDCFCLKDeepSleep(
+			mode_lib,
+			v->NumberOfActivePlanes,
+			v->BytePerPixelY,
+			v->BytePerPixelC,
+			v->VRatio,
+			v->VRatioChroma,
+			v->SwathWidthY,
+			v->SwathWidthC,
+			v->DPPPerPlane,
+			v->HRatio,
+			v->HRatioChroma,
+			v->PixelClock,
+			v->PSCL_THROUGHPUT_LUMA,
+			v->PSCL_THROUGHPUT_CHROMA,
+			v->DPPCLK,
+			v->ReadBandwidthPlaneLuma,
+			v->ReadBandwidthPlaneChroma,
+			v->ReturnBusWidth,
+			&v->DCFCLKDeepSleep);
+
+	// DSCCLK
+	for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+		if ((v->BlendingAndTiming[k] != k) || !v->DSCEnabled[k]) {
+			v->DSCCLK_calculated[k] = 0.0;
+		} else {
+			if (v->OutputFormat[k] == dm_420)
+				v->DSCFormatFactor = 2;
+			else if (v->OutputFormat[k] == dm_444)
+				v->DSCFormatFactor = 1;
+			else if (v->OutputFormat[k] == dm_n422)
+				v->DSCFormatFactor = 2;
+			else
+				v->DSCFormatFactor = 1;
+			if (v->ODMCombineEnabled[k] == dm_odm_combine_mode_4to1)
+				v->DSCCLK_calculated[k] = v->PixelClockBackEnd[k] / 12 / v->DSCFormatFactor
+						/ (1 - v->DISPCLKDPPCLKDSCCLKDownSpreading / 100);
+			else if (v->ODMCombineEnabled[k] == dm_odm_combine_mode_2to1)
+				v->DSCCLK_calculated[k] = v->PixelClockBackEnd[k] / 6 / v->DSCFormatFactor
+						/ (1 - v->DISPCLKDPPCLKDSCCLKDownSpreading / 100);
+			else
+				v->DSCCLK_calculated[k] = v->PixelClockBackEnd[k] / 3 / v->DSCFormatFactor
+						/ (1 - v->DISPCLKDPPCLKDSCCLKDownSpreading / 100);
+		}
+	}
+
+	// DSC Delay
+	for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+		double BPP = v->OutputBpp[k];
+
+		if (v->DSCEnabled[k] && BPP != 0) {
+			if (v->ODMCombineEnabled[k] == dm_odm_combine_mode_disabled) {
+				v->DSCDelay[k] = dscceComputeDelay(
+						v->DSCInputBitPerComponent[k],
+						BPP,
+						dml_ceil((double) v->HActive[k] / v->NumberOfDSCSlices[k], 1),
+						v->NumberOfDSCSlices[k],
+						v->OutputFormat[k],
+						v->Output[k]) + dscComputeDelay(v->OutputFormat[k], v->Output[k]);
+			} else if (v->ODMCombineEnabled[k] == dm_odm_combine_mode_2to1) {
+				v->DSCDelay[k] = 2
+						* (dscceComputeDelay(
+								v->DSCInputBitPerComponent[k],
+								BPP,
+								dml_ceil((double) v->HActive[k] / v->NumberOfDSCSlices[k], 1),
+								v->NumberOfDSCSlices[k] / 2.0,
+								v->OutputFormat[k],
+								v->Output[k]) + dscComputeDelay(v->OutputFormat[k], v->Output[k]));
+			} else {
+				v->DSCDelay[k] = 4
+						* (dscceComputeDelay(
+								v->DSCInputBitPerComponent[k],
+								BPP,
+								dml_ceil((double) v->HActive[k] / v->NumberOfDSCSlices[k], 1),
+								v->NumberOfDSCSlices[k] / 4.0,
+								v->OutputFormat[k],
+								v->Output[k]) + dscComputeDelay(v->OutputFormat[k], v->Output[k]));
+			}
+			v->DSCDelay[k] = v->DSCDelay[k] + (v->HTotal[k] - v->HActive[k]) * dml_ceil((double) v->DSCDelay[k] / v->HActive[k], 1);
+			v->DSCDelay[k] = v->DSCDelay[k] * v->PixelClock[k] / v->PixelClockBackEnd[k];
+		} else {
+			v->DSCDelay[k] = 0;
+		}
+	}
+
+	for (k = 0; k < v->NumberOfActivePlanes; ++k)
+		for (j = 0; j < v->NumberOfActivePlanes; ++j) // NumberOfPlanes
+			if (j != k && v->BlendingAndTiming[k] == j && v->DSCEnabled[j])
+				v->DSCDelay[k] = v->DSCDelay[j];
+
+	// Prefetch
+	for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+		unsigned int PDEAndMetaPTEBytesFrameY;
+		unsigned int PixelPTEBytesPerRowY;
+		unsigned int MetaRowByteY;
+		unsigned int MetaRowByteC;
+		unsigned int PDEAndMetaPTEBytesFrameC;
+		unsigned int PixelPTEBytesPerRowC;
+		bool PTEBufferSizeNotExceededY;
+		bool PTEBufferSizeNotExceededC;
+
+		if (v->SourcePixelFormat[k] == dm_420_8 || v->SourcePixelFormat[k] == dm_420_10 || v->SourcePixelFormat[k] == dm_420_12
+				|| v->SourcePixelFormat[k] == dm_rgbe_alpha) {
+			if ((v->SourcePixelFormat[k] == dm_420_10 || v->SourcePixelFormat[k] == dm_420_12) && v->SourceScan[k] != dm_vert) {
+				v->PTEBufferSizeInRequestsForLuma = (v->PTEBufferSizeInRequestsLuma + v->PTEBufferSizeInRequestsChroma) / 2;
+				v->PTEBufferSizeInRequestsForChroma = v->PTEBufferSizeInRequestsForLuma;
+			} else {
+				v->PTEBufferSizeInRequestsForLuma = v->PTEBufferSizeInRequestsLuma;
+				v->PTEBufferSizeInRequestsForChroma = v->PTEBufferSizeInRequestsChroma;
+			}
+
+			PDEAndMetaPTEBytesFrameC = CalculateVMAndRowBytes(
+					mode_lib,
+					v->DCCEnable[k],
+					v->BlockHeight256BytesC[k],
+					v->BlockWidth256BytesC[k],
+					v->SourcePixelFormat[k],
+					v->SurfaceTiling[k],
+					v->BytePerPixelC[k],
+					v->SourceScan[k],
+					v->SwathWidthC[k],
+					v->ViewportHeightChroma[k],
+					v->GPUVMEnable,
+					v->HostVMEnable,
+					v->HostVMMaxNonCachedPageTableLevels,
+					v->GPUVMMinPageSize,
+					v->HostVMMinPageSize,
+					v->PTEBufferSizeInRequestsForChroma,
+					v->PitchC[k],
+					v->DCCMetaPitchC[k],
+					&v->MacroTileWidthC[k],
+					&MetaRowByteC,
+					&PixelPTEBytesPerRowC,
+					&PTEBufferSizeNotExceededC,
+					&v->dpte_row_width_chroma_ub[k],
+					&v->dpte_row_height_chroma[k],
+					&v->meta_req_width_chroma[k],
+					&v->meta_req_height_chroma[k],
+					&v->meta_row_width_chroma[k],
+					&v->meta_row_height_chroma[k],
+					&v->dummyinteger1,
+					&v->dummyinteger2,
+					&v->PixelPTEReqWidthC[k],
+					&v->PixelPTEReqHeightC[k],
+					&v->PTERequestSizeC[k],
+					&v->dpde0_bytes_per_frame_ub_c[k],
+					&v->meta_pte_bytes_per_frame_ub_c[k]);
+
+			v->PrefetchSourceLinesC[k] = CalculatePrefetchSourceLines(
+					mode_lib,
+					v->VRatioChroma[k],
+					v->VTAPsChroma[k],
+					v->Interlace[k],
+					v->ProgressiveToInterlaceUnitInOPP,
+					v->SwathHeightC[k],
+					v->ViewportYStartC[k],
+					&v->VInitPreFillC[k],
+					&v->MaxNumSwathC[k]);
+		} else {
+			v->PTEBufferSizeInRequestsForLuma = v->PTEBufferSizeInRequestsLuma + v->PTEBufferSizeInRequestsChroma;
+			v->PTEBufferSizeInRequestsForChroma = 0;
+			PixelPTEBytesPerRowC = 0;
+			PDEAndMetaPTEBytesFrameC = 0;
+			MetaRowByteC = 0;
+			v->MaxNumSwathC[k] = 0;
+			v->PrefetchSourceLinesC[k] = 0;
+		}
+
+		PDEAndMetaPTEBytesFrameY = CalculateVMAndRowBytes(
+				mode_lib,
+				v->DCCEnable[k],
+				v->BlockHeight256BytesY[k],
+				v->BlockWidth256BytesY[k],
+				v->SourcePixelFormat[k],
+				v->SurfaceTiling[k],
+				v->BytePerPixelY[k],
+				v->SourceScan[k],
+				v->SwathWidthY[k],
+				v->ViewportHeight[k],
+				v->GPUVMEnable,
+				v->HostVMEnable,
+				v->HostVMMaxNonCachedPageTableLevels,
+				v->GPUVMMinPageSize,
+				v->HostVMMinPageSize,
+				v->PTEBufferSizeInRequestsForLuma,
+				v->PitchY[k],
+				v->DCCMetaPitchY[k],
+				&v->MacroTileWidthY[k],
+				&MetaRowByteY,
+				&PixelPTEBytesPerRowY,
+				&PTEBufferSizeNotExceededY,
+				&v->dpte_row_width_luma_ub[k],
+				&v->dpte_row_height[k],
+				&v->meta_req_width[k],
+				&v->meta_req_height[k],
+				&v->meta_row_width[k],
+				&v->meta_row_height[k],
+				&v->vm_group_bytes[k],
+				&v->dpte_group_bytes[k],
+				&v->PixelPTEReqWidthY[k],
+				&v->PixelPTEReqHeightY[k],
+				&v->PTERequestSizeY[k],
+				&v->dpde0_bytes_per_frame_ub_l[k],
+				&v->meta_pte_bytes_per_frame_ub_l[k]);
+
+		v->PrefetchSourceLinesY[k] = CalculatePrefetchSourceLines(
+				mode_lib,
+				v->VRatio[k],
+				v->vtaps[k],
+				v->Interlace[k],
+				v->ProgressiveToInterlaceUnitInOPP,
+				v->SwathHeightY[k],
+				v->ViewportYStartY[k],
+				&v->VInitPreFillY[k],
+				&v->MaxNumSwathY[k]);
+		v->PixelPTEBytesPerRow[k] = PixelPTEBytesPerRowY + PixelPTEBytesPerRowC;
+		v->PDEAndMetaPTEBytesFrame[k] = PDEAndMetaPTEBytesFrameY + PDEAndMetaPTEBytesFrameC;
+		v->MetaRowByte[k] = MetaRowByteY + MetaRowByteC;
+
+		CalculateRowBandwidth(
+				v->GPUVMEnable,
+				v->SourcePixelFormat[k],
+				v->VRatio[k],
+				v->VRatioChroma[k],
+				v->DCCEnable[k],
+				v->HTotal[k] / v->PixelClock[k],
+				MetaRowByteY,
+				MetaRowByteC,
+				v->meta_row_height[k],
+				v->meta_row_height_chroma[k],
+				PixelPTEBytesPerRowY,
+				PixelPTEBytesPerRowC,
+				v->dpte_row_height[k],
+				v->dpte_row_height_chroma[k],
+				&v->meta_row_bw[k],
+				&v->dpte_row_bw[k]);
+	}
+
+	v->TotalDCCActiveDPP = 0;
+	v->TotalActiveDPP = 0;
+	for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+		v->TotalActiveDPP = v->TotalActiveDPP + v->DPPPerPlane[k];
+		if (v->DCCEnable[k])
+			v->TotalDCCActiveDPP = v->TotalDCCActiveDPP + v->DPPPerPlane[k];
+		if (v->SourcePixelFormat[k] == dm_420_8 || v->SourcePixelFormat[k] == dm_420_10 || v->SourcePixelFormat[k] == dm_420_12
+				|| v->SourcePixelFormat[k] == dm_rgbe_alpha)
+			NoChromaPlanes = false;
+	}
+
+	ReorderBytes = v->NumberOfChannels
+			* dml_max3(
+					v->UrgentOutOfOrderReturnPerChannelPixelDataOnly,
+					v->UrgentOutOfOrderReturnPerChannelPixelMixedWithVMData,
+					v->UrgentOutOfOrderReturnPerChannelVMDataOnly);
+
+	VMDataOnlyReturnBW = dml_min(
+			dml_min(v->ReturnBusWidth * v->DCFCLK, v->FabricClock * v->FabricDatapathToDCNDataReturn)
+					* v->PercentOfIdealFabricAndSDPPortBWReceivedAfterUrgLatency / 100.0,
+			v->DRAMSpeed * v->NumberOfChannels * v->DRAMChannelWidth
+					* v->PercentOfIdealDRAMBWReceivedAfterUrgLatencyVMDataOnly / 100.0);
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: v->ReturnBusWidth = %f\n", __func__, v->ReturnBusWidth);
+	dml_print("DML::%s: v->DCFCLK = %f\n", __func__, v->DCFCLK);
+	dml_print("DML::%s: v->FabricClock = %f\n", __func__, v->FabricClock);
+	dml_print("DML::%s: v->FabricDatapathToDCNDataReturn = %f\n", __func__, v->FabricDatapathToDCNDataReturn);
+	dml_print("DML::%s: v->PercentOfIdealFabricAndSDPPortBWReceivedAfterUrgLatency = %f\n", __func__, v->PercentOfIdealFabricAndSDPPortBWReceivedAfterUrgLatency);
+	dml_print("DML::%s: v->DRAMSpeed = %f\n", __func__, v->DRAMSpeed);
+	dml_print("DML::%s: v->NumberOfChannels = %f\n", __func__, v->NumberOfChannels);
+	dml_print("DML::%s: v->DRAMChannelWidth = %f\n", __func__, v->DRAMChannelWidth);
+	dml_print("DML::%s: v->PercentOfIdealDRAMBWReceivedAfterUrgLatencyVMDataOnly = %f\n", __func__, v->PercentOfIdealDRAMBWReceivedAfterUrgLatencyVMDataOnly);
+	dml_print("DML::%s: VMDataOnlyReturnBW = %f\n", __func__, VMDataOnlyReturnBW);
+	dml_print("DML::%s: ReturnBW = %f\n", __func__, v->ReturnBW);
+#endif
+
+	if (v->GPUVMEnable && v->HostVMEnable)
+		HostVMInefficiencyFactor = v->ReturnBW / VMDataOnlyReturnBW;
+
+	v->UrgentExtraLatency = CalculateExtraLatency(
+			v->RoundTripPingLatencyCycles,
+			ReorderBytes,
+			v->DCFCLK,
+			v->TotalActiveDPP,
+			v->PixelChunkSizeInKByte,
+			v->TotalDCCActiveDPP,
+			v->MetaChunkSize,
+			v->ReturnBW,
+			v->GPUVMEnable,
+			v->HostVMEnable,
+			v->NumberOfActivePlanes,
+			v->DPPPerPlane,
+			v->dpte_group_bytes,
+			HostVMInefficiencyFactor,
+			v->HostVMMinPageSize,
+			v->HostVMMaxNonCachedPageTableLevels);
+
+	v->TCalc = 24.0 / v->DCFCLKDeepSleep;
+
+	for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+		if (v->BlendingAndTiming[k] == k) {
+			if (v->WritebackEnable[k] == true) {
+				v->WritebackDelay[v->VoltageLevel][k] = v->WritebackLatency
+						+ CalculateWriteBackDelay(
+								v->WritebackPixelFormat[k],
+								v->WritebackHRatio[k],
+								v->WritebackVRatio[k],
+								v->WritebackVTaps[k],
+								v->WritebackDestinationWidth[k],
+								v->WritebackDestinationHeight[k],
+								v->WritebackSourceHeight[k],
+								v->HTotal[k]) / v->DISPCLK;
+			} else
+				v->WritebackDelay[v->VoltageLevel][k] = 0;
+			for (j = 0; j < v->NumberOfActivePlanes; ++j) {
+				if (v->BlendingAndTiming[j] == k && v->WritebackEnable[j] == true) {
+					v->WritebackDelay[v->VoltageLevel][k] = dml_max(
+							v->WritebackDelay[v->VoltageLevel][k],
+							v->WritebackLatency
+									+ CalculateWriteBackDelay(
+											v->WritebackPixelFormat[j],
+											v->WritebackHRatio[j],
+											v->WritebackVRatio[j],
+											v->WritebackVTaps[j],
+											v->WritebackDestinationWidth[j],
+											v->WritebackDestinationHeight[j],
+											v->WritebackSourceHeight[j],
+											v->HTotal[k]) / v->DISPCLK);
+				}
+			}
+		}
+	}
+
+	for (k = 0; k < v->NumberOfActivePlanes; ++k)
+		for (j = 0; j < v->NumberOfActivePlanes; ++j)
+			if (v->BlendingAndTiming[k] == j)
+				v->WritebackDelay[v->VoltageLevel][k] = v->WritebackDelay[v->VoltageLevel][j];
+
+	for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+		v->MaxVStartupLines[k] =
+			CalculateMaxVStartup(
+					v->VTotal[k],
+					v->VActive[k],
+					v->VBlankNom[k],
+					v->HTotal[k],
+					v->PixelClock[k],
+					v->ProgressiveToInterlaceUnitInOPP,
+					v->Interlace[k],
+					v->ip.VBlankNomDefaultUS,
+					v->WritebackDelay[v->VoltageLevel][k]);
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%d MaxVStartupLines = %d\n", __func__, k, v->MaxVStartupLines[k]);
+		dml_print("DML::%s: k=%d VoltageLevel = %d\n", __func__, k, v->VoltageLevel);
+		dml_print("DML::%s: k=%d WritebackDelay = %f\n", __func__, k, v->WritebackDelay[v->VoltageLevel][k]);
+#endif
+	}
+
+	v->MaximumMaxVStartupLines = 0;
+	for (k = 0; k < v->NumberOfActivePlanes; ++k)
+		v->MaximumMaxVStartupLines = dml_max(v->MaximumMaxVStartupLines, v->MaxVStartupLines[k]);
+
+	// VBA_DELTA
+	// We don't really care to iterate between the various prefetch modes
+	//v->PrefetchERROR = CalculateMinAndMaxPrefetchMode(v->AllowDRAMSelfRefreshOrDRAMClockChangeInVblank, &v->MinPrefetchMode, &v->MaxPrefetchMode);
+
+	v->UrgentLatency = CalculateUrgentLatency(
+			v->UrgentLatencyPixelDataOnly,
+			v->UrgentLatencyPixelMixedWithVMData,
+			v->UrgentLatencyVMDataOnly,
+			v->DoUrgentLatencyAdjustment,
+			v->UrgentLatencyAdjustmentFabricClockComponent,
+			v->UrgentLatencyAdjustmentFabricClockReference,
+			v->FabricClock);
+
+	v->FractionOfUrgentBandwidth = 0.0;
+	v->FractionOfUrgentBandwidthImmediateFlip = 0.0;
+
+	v->VStartupLines = __DML_VBA_MIN_VSTARTUP__;
+
+	do {
+		double MaxTotalRDBandwidthNoUrgentBurst = 0.0;
+		bool DestinationLineTimesForPrefetchLessThan2 = false;
+		bool VRatioPrefetchMoreThan4 = false;
+		double TWait = CalculateTWait(PrefetchMode, v->DRAMClockChangeLatency, v->UrgentLatency, v->SREnterPlusExitTime);
+
+		MaxTotalRDBandwidth = 0;
+
+		dml_print("DML::%s: Start loop: VStartup = %d\n", __func__, v->VStartupLines);
+
+		for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+			Pipe myPipe;
+
+			myPipe.DPPCLK = v->DPPCLK[k];
+			myPipe.DISPCLK = v->DISPCLK;
+			myPipe.PixelClock = v->PixelClock[k];
+			myPipe.DCFCLKDeepSleep = v->DCFCLKDeepSleep;
+			myPipe.DPPPerPlane = v->DPPPerPlane[k];
+			myPipe.ScalerEnabled = v->ScalerEnabled[k];
+			myPipe.VRatio = v->VRatio[k];
+			myPipe.VRatioChroma = v->VRatioChroma[k];
+			myPipe.SourceScan = v->SourceScan[k];
+			myPipe.BlockWidth256BytesY = v->BlockWidth256BytesY[k];
+			myPipe.BlockHeight256BytesY = v->BlockHeight256BytesY[k];
+			myPipe.BlockWidth256BytesC = v->BlockWidth256BytesC[k];
+			myPipe.BlockHeight256BytesC = v->BlockHeight256BytesC[k];
+			myPipe.InterlaceEnable = v->Interlace[k];
+			myPipe.NumberOfCursors = v->NumberOfCursors[k];
+			myPipe.VBlank = v->VTotal[k] - v->VActive[k];
+			myPipe.HTotal = v->HTotal[k];
+			myPipe.DCCEnable = v->DCCEnable[k];
+			myPipe.ODMCombineIsEnabled = v->ODMCombineEnabled[k] == dm_odm_combine_mode_4to1
+					|| v->ODMCombineEnabled[k] == dm_odm_combine_mode_2to1;
+			myPipe.SourcePixelFormat = v->SourcePixelFormat[k];
+			myPipe.BytePerPixelY = v->BytePerPixelY[k];
+			myPipe.BytePerPixelC = v->BytePerPixelC[k];
+			myPipe.ProgressiveToInterlaceUnitInOPP = v->ProgressiveToInterlaceUnitInOPP;
+			v->ErrorResult[k] = CalculatePrefetchSchedule(
+					mode_lib,
+					HostVMInefficiencyFactor,
+					&myPipe,
+					v->DSCDelay[k],
+					v->DPPCLKDelaySubtotal + v->DPPCLKDelayCNVCFormater,
+					v->DPPCLKDelaySCL,
+					v->DPPCLKDelaySCLLBOnly,
+					v->DPPCLKDelayCNVCCursor,
+					v->DISPCLKDelaySubtotal,
+					(unsigned int) (v->SwathWidthY[k] / v->HRatio[k]),
+					v->OutputFormat[k],
+					v->MaxInterDCNTileRepeaters,
+					dml_min(v->VStartupLines, v->MaxVStartupLines[k]),
+					v->MaxVStartupLines[k],
+					v->GPUVMMaxPageTableLevels,
+					v->GPUVMEnable,
+					v->HostVMEnable,
+					v->HostVMMaxNonCachedPageTableLevels,
+					v->HostVMMinPageSize,
+					v->DynamicMetadataEnable[k],
+					v->DynamicMetadataVMEnabled,
+					v->DynamicMetadataLinesBeforeActiveRequired[k],
+					v->DynamicMetadataTransmittedBytes[k],
+					v->UrgentLatency,
+					v->UrgentExtraLatency,
+					v->TCalc,
+					v->PDEAndMetaPTEBytesFrame[k],
+					v->MetaRowByte[k],
+					v->PixelPTEBytesPerRow[k],
+					v->PrefetchSourceLinesY[k],
+					v->SwathWidthY[k],
+					v->VInitPreFillY[k],
+					v->MaxNumSwathY[k],
+					v->PrefetchSourceLinesC[k],
+					v->SwathWidthC[k],
+					v->VInitPreFillC[k],
+					v->MaxNumSwathC[k],
+					v->swath_width_luma_ub[k],
+					v->swath_width_chroma_ub[k],
+					v->SwathHeightY[k],
+					v->SwathHeightC[k],
+					TWait,
+					&v->DSTXAfterScaler[k],
+					&v->DSTYAfterScaler[k],
+					&v->DestinationLinesForPrefetch[k],
+					&v->PrefetchBandwidth[k],
+					&v->DestinationLinesToRequestVMInVBlank[k],
+					&v->DestinationLinesToRequestRowInVBlank[k],
+					&v->VRatioPrefetchY[k],
+					&v->VRatioPrefetchC[k],
+					&v->RequiredPrefetchPixDataBWLuma[k],
+					&v->RequiredPrefetchPixDataBWChroma[k],
+					&v->NotEnoughTimeForDynamicMetadata[k],
+					&v->Tno_bw[k],
+					&v->prefetch_vmrow_bw[k],
+					&v->Tdmdl_vm[k],
+					&v->Tdmdl[k],
+					&v->TSetup[k],
+					&v->VUpdateOffsetPix[k],
+					&v->VUpdateWidthPix[k],
+					&v->VReadyOffsetPix[k]);
+
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: k=%0d Prefetch cal result=%0d\n", __func__, k, v->ErrorResult[k]);
+#endif
+			v->VStartup[k] = dml_min(v->VStartupLines, v->MaxVStartupLines[k]);
+		}
+
+		v->NoEnoughUrgentLatencyHiding = false;
+		v->NoEnoughUrgentLatencyHidingPre = false;
+
+		for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+			v->cursor_bw[k] = v->NumberOfCursors[k] * v->CursorWidth[k][0] * v->CursorBPP[k][0] / 8.0
+					/ (v->HTotal[k] / v->PixelClock[k]) * v->VRatio[k];
+			v->cursor_bw_pre[k] = v->NumberOfCursors[k] * v->CursorWidth[k][0] * v->CursorBPP[k][0] / 8.0
+					/ (v->HTotal[k] / v->PixelClock[k]) * v->VRatioPrefetchY[k];
+
+			CalculateUrgentBurstFactor(
+					v->swath_width_luma_ub[k],
+					v->swath_width_chroma_ub[k],
+					v->SwathHeightY[k],
+					v->SwathHeightC[k],
+					v->HTotal[k] / v->PixelClock[k],
+					v->UrgentLatency,
+					v->CursorBufferSize,
+					v->CursorWidth[k][0],
+					v->CursorBPP[k][0],
+					v->VRatio[k],
+					v->VRatioChroma[k],
+					v->BytePerPixelDETY[k],
+					v->BytePerPixelDETC[k],
+					v->DETBufferSizeY[k],
+					v->DETBufferSizeC[k],
+					&v->UrgBurstFactorCursor[k],
+					&v->UrgBurstFactorLuma[k],
+					&v->UrgBurstFactorChroma[k],
+					&v->NoUrgentLatencyHiding[k]);
+
+			CalculateUrgentBurstFactor(
+					v->swath_width_luma_ub[k],
+					v->swath_width_chroma_ub[k],
+					v->SwathHeightY[k],
+					v->SwathHeightC[k],
+					v->HTotal[k] / v->PixelClock[k],
+					v->UrgentLatency,
+					v->CursorBufferSize,
+					v->CursorWidth[k][0],
+					v->CursorBPP[k][0],
+					v->VRatioPrefetchY[k],
+					v->VRatioPrefetchC[k],
+					v->BytePerPixelDETY[k],
+					v->BytePerPixelDETC[k],
+					v->DETBufferSizeY[k],
+					v->DETBufferSizeC[k],
+					&v->UrgBurstFactorCursorPre[k],
+					&v->UrgBurstFactorLumaPre[k],
+					&v->UrgBurstFactorChromaPre[k],
+					&v->NoUrgentLatencyHidingPre[k]);
+
+			MaxTotalRDBandwidth = MaxTotalRDBandwidth
+					+ dml_max3(
+							v->DPPPerPlane[k] * v->prefetch_vmrow_bw[k],
+							v->ReadBandwidthPlaneLuma[k] * v->UrgBurstFactorLuma[k]
+									+ v->ReadBandwidthPlaneChroma[k] * v->UrgBurstFactorChroma[k]
+									+ v->cursor_bw[k] * v->UrgBurstFactorCursor[k]
+									+ v->DPPPerPlane[k] * (v->meta_row_bw[k] + v->dpte_row_bw[k]),
+							v->DPPPerPlane[k]
+									* (v->RequiredPrefetchPixDataBWLuma[k] * v->UrgBurstFactorLumaPre[k]
+											+ v->RequiredPrefetchPixDataBWChroma[k] * v->UrgBurstFactorChromaPre[k])
+									+ v->cursor_bw_pre[k] * v->UrgBurstFactorCursorPre[k]);
+
+			MaxTotalRDBandwidthNoUrgentBurst = MaxTotalRDBandwidthNoUrgentBurst
+					+ dml_max3(
+							v->DPPPerPlane[k] * v->prefetch_vmrow_bw[k],
+							v->ReadBandwidthPlaneLuma[k] + v->ReadBandwidthPlaneChroma[k] + v->cursor_bw[k]
+									+ v->DPPPerPlane[k] * (v->meta_row_bw[k] + v->dpte_row_bw[k]),
+							v->DPPPerPlane[k] * (v->RequiredPrefetchPixDataBWLuma[k] + v->RequiredPrefetchPixDataBWChroma[k])
+									+ v->cursor_bw_pre[k]);
+
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: k=%0d DPPPerPlane=%d\n", __func__, k, v->DPPPerPlane[k]);
+			dml_print("DML::%s: k=%0d UrgBurstFactorLuma=%f\n", __func__, k, v->UrgBurstFactorLuma[k]);
+			dml_print("DML::%s: k=%0d UrgBurstFactorChroma=%f\n", __func__, k, v->UrgBurstFactorChroma[k]);
+			dml_print("DML::%s: k=%0d UrgBurstFactorLumaPre=%f\n", __func__, k, v->UrgBurstFactorLumaPre[k]);
+			dml_print("DML::%s: k=%0d UrgBurstFactorChromaPre=%f\n", __func__, k, v->UrgBurstFactorChromaPre[k]);
+
+			dml_print("DML::%s: k=%0d VRatioPrefetchY=%f\n", __func__, k, v->VRatioPrefetchY[k]);
+			dml_print("DML::%s: k=%0d VRatioY=%f\n", __func__, k, v->VRatio[k]);
+
+			dml_print("DML::%s: k=%0d prefetch_vmrow_bw=%f\n", __func__, k, v->prefetch_vmrow_bw[k]);
+			dml_print("DML::%s: k=%0d ReadBandwidthPlaneLuma=%f\n", __func__, k, v->ReadBandwidthPlaneLuma[k]);
+			dml_print("DML::%s: k=%0d ReadBandwidthPlaneChroma=%f\n", __func__, k, v->ReadBandwidthPlaneChroma[k]);
+			dml_print("DML::%s: k=%0d cursor_bw=%f\n", __func__, k, v->cursor_bw[k]);
+			dml_print("DML::%s: k=%0d meta_row_bw=%f\n", __func__, k, v->meta_row_bw[k]);
+			dml_print("DML::%s: k=%0d dpte_row_bw=%f\n", __func__, k, v->dpte_row_bw[k]);
+			dml_print("DML::%s: k=%0d RequiredPrefetchPixDataBWLuma=%f\n", __func__, k, v->RequiredPrefetchPixDataBWLuma[k]);
+			dml_print("DML::%s: k=%0d RequiredPrefetchPixDataBWChroma=%f\n", __func__, k, v->RequiredPrefetchPixDataBWChroma[k]);
+			dml_print("DML::%s: k=%0d cursor_bw_pre=%f\n", __func__, k, v->cursor_bw_pre[k]);
+			dml_print("DML::%s: k=%0d MaxTotalRDBandwidthNoUrgentBurst=%f\n", __func__, k, MaxTotalRDBandwidthNoUrgentBurst);
+#endif
+
+			if (v->DestinationLinesForPrefetch[k] < 2)
+				DestinationLineTimesForPrefetchLessThan2 = true;
+
+			if (v->VRatioPrefetchY[k] > 4 || v->VRatioPrefetchC[k] > 4)
+				VRatioPrefetchMoreThan4 = true;
+
+			if (v->NoUrgentLatencyHiding[k] == true)
+				v->NoEnoughUrgentLatencyHiding = true;
+
+			if (v->NoUrgentLatencyHidingPre[k] == true)
+				v->NoEnoughUrgentLatencyHidingPre = true;
+		}
+
+		v->FractionOfUrgentBandwidth = MaxTotalRDBandwidthNoUrgentBurst / v->ReturnBW;
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: MaxTotalRDBandwidthNoUrgentBurst=%f\n", __func__, MaxTotalRDBandwidthNoUrgentBurst);
+		dml_print("DML::%s: ReturnBW=%f\n", __func__, v->ReturnBW);
+		dml_print("DML::%s: FractionOfUrgentBandwidth=%f\n", __func__, v->FractionOfUrgentBandwidth);
+#endif
+
+		if (MaxTotalRDBandwidth <= v->ReturnBW && v->NoEnoughUrgentLatencyHiding == 0 && v->NoEnoughUrgentLatencyHidingPre == 0
+				&& !VRatioPrefetchMoreThan4 && !DestinationLineTimesForPrefetchLessThan2)
+			v->PrefetchModeSupported = true;
+		else {
+			v->PrefetchModeSupported = false;
+			dml_print("DML::%s: ***failed***. Bandwidth violation. Results are NOT valid\n", __func__);
+			dml_print("DML::%s: MaxTotalRDBandwidth:%f AvailReturnBandwidth:%f\n", __func__, MaxTotalRDBandwidth, v->ReturnBW);
+			dml_print("DML::%s: VRatioPrefetch %s more than 4\n", __func__, (VRatioPrefetchMoreThan4) ? "is" : "is not");
+			dml_print("DML::%s: DestinationLines for Prefetch %s less than 2\n", __func__, (DestinationLineTimesForPrefetchLessThan2) ? "is" : "is not");
+		}
+
+		// PREVIOUS_ERROR
+		// This error result check was done after the PrefetchModeSupported. So we will
+		// still try to calculate flip schedule even prefetch mode not supported
+		for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+			if (v->ErrorResult[k] == true || v->NotEnoughTimeForDynamicMetadata[k] == true) {
+				v->PrefetchModeSupported = false;
+				dml_print("DML::%s: ***failed***. Prefetch schedule violation. Results are NOT valid\n", __func__);
+			}
+		}
+
+		if (v->PrefetchModeSupported == true && v->ImmediateFlipSupport == true) {
+			v->BandwidthAvailableForImmediateFlip = v->ReturnBW;
+			for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+				v->BandwidthAvailableForImmediateFlip = v->BandwidthAvailableForImmediateFlip
+						- dml_max(
+								v->ReadBandwidthPlaneLuma[k] * v->UrgBurstFactorLuma[k]
+										+ v->ReadBandwidthPlaneChroma[k] * v->UrgBurstFactorChroma[k]
+										+ v->cursor_bw[k] * v->UrgBurstFactorCursor[k],
+								v->DPPPerPlane[k]
+										* (v->RequiredPrefetchPixDataBWLuma[k] * v->UrgBurstFactorLumaPre[k]
+												+ v->RequiredPrefetchPixDataBWChroma[k] * v->UrgBurstFactorChromaPre[k])
+										+ v->cursor_bw_pre[k] * v->UrgBurstFactorCursorPre[k]);
+			}
+
+			v->TotImmediateFlipBytes = 0;
+			for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+				v->TotImmediateFlipBytes = v->TotImmediateFlipBytes
+						+ v->DPPPerPlane[k] * (v->PDEAndMetaPTEBytesFrame[k] + v->MetaRowByte[k] + v->PixelPTEBytesPerRow[k]);
+			}
+			for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+				CalculateFlipSchedule(
+						mode_lib,
+						k,
+						HostVMInefficiencyFactor,
+						v->UrgentExtraLatency,
+						v->UrgentLatency,
+						v->PDEAndMetaPTEBytesFrame[k],
+						v->MetaRowByte[k],
+						v->PixelPTEBytesPerRow[k]);
+			}
+
+			v->total_dcn_read_bw_with_flip = 0.0;
+			v->total_dcn_read_bw_with_flip_no_urgent_burst = 0.0;
+			for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+				v->total_dcn_read_bw_with_flip = v->total_dcn_read_bw_with_flip
+						+ dml_max3(
+								v->DPPPerPlane[k] * v->prefetch_vmrow_bw[k],
+								v->DPPPerPlane[k] * v->final_flip_bw[k]
+										+ v->ReadBandwidthLuma[k] * v->UrgBurstFactorLuma[k]
+										+ v->ReadBandwidthChroma[k] * v->UrgBurstFactorChroma[k]
+										+ v->cursor_bw[k] * v->UrgBurstFactorCursor[k],
+								v->DPPPerPlane[k]
+										* (v->final_flip_bw[k]
+												+ v->RequiredPrefetchPixDataBWLuma[k] * v->UrgBurstFactorLumaPre[k]
+												+ v->RequiredPrefetchPixDataBWChroma[k] * v->UrgBurstFactorChromaPre[k])
+										+ v->cursor_bw_pre[k] * v->UrgBurstFactorCursorPre[k]);
+				v->total_dcn_read_bw_with_flip_no_urgent_burst = v->total_dcn_read_bw_with_flip_no_urgent_burst
+						+ dml_max3(
+								v->DPPPerPlane[k] * v->prefetch_vmrow_bw[k],
+								v->DPPPerPlane[k] * v->final_flip_bw[k] + v->ReadBandwidthPlaneLuma[k]
+										+ v->ReadBandwidthPlaneChroma[k] + v->cursor_bw[k],
+								v->DPPPerPlane[k]
+										* (v->final_flip_bw[k] + v->RequiredPrefetchPixDataBWLuma[k]
+												+ v->RequiredPrefetchPixDataBWChroma[k]) + v->cursor_bw_pre[k]);
+			}
+			v->FractionOfUrgentBandwidthImmediateFlip = v->total_dcn_read_bw_with_flip_no_urgent_burst / v->ReturnBW;
+
+			v->ImmediateFlipSupported = true;
+			if (v->total_dcn_read_bw_with_flip > v->ReturnBW) {
+#ifdef __DML_VBA_DEBUG__
+				dml_print("DML::%s: total_dcn_read_bw_with_flip %f (bw w/ flip too high!)\n", __func__, v->total_dcn_read_bw_with_flip);
+#endif
+				v->ImmediateFlipSupported = false;
+				v->total_dcn_read_bw_with_flip = MaxTotalRDBandwidth;
+			}
+			for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+				if (v->ImmediateFlipSupportedForPipe[k] == false) {
+#ifdef __DML_VBA_DEBUG__
+					dml_print("DML::%s: Pipe %0d not supporting iflip\n", __func__, k);
+#endif
+					v->ImmediateFlipSupported = false;
+				}
+			}
+		} else {
+			v->ImmediateFlipSupported = false;
+		}
+
+		v->PrefetchAndImmediateFlipSupported =
+				(v->PrefetchModeSupported == true && ((!v->ImmediateFlipSupport && !v->HostVMEnable
+				&& v->ImmediateFlipRequirement[0] != dm_immediate_flip_required) ||
+				v->ImmediateFlipSupported)) ? true : false;
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: PrefetchModeSupported %d\n", __func__, v->PrefetchModeSupported);
+		dml_print("DML::%s: ImmediateFlipRequirement %d\n", __func__, v->ImmediateFlipRequirement == dm_immediate_flip_required);
+		dml_print("DML::%s: ImmediateFlipSupported %d\n", __func__, v->ImmediateFlipSupported);
+		dml_print("DML::%s: ImmediateFlipSupport %d\n", __func__, v->ImmediateFlipSupport);
+		dml_print("DML::%s: HostVMEnable %d\n", __func__, v->HostVMEnable);
+		dml_print("DML::%s: PrefetchAndImmediateFlipSupported %d\n", __func__, v->PrefetchAndImmediateFlipSupported);
+#endif
+		dml_print("DML::%s: Done loop: Vstartup=%d, Max Vstartup is %d\n", __func__, v->VStartupLines, v->MaximumMaxVStartupLines);
+
+		v->VStartupLines = v->VStartupLines + 1;
+	} while (!v->PrefetchAndImmediateFlipSupported && v->VStartupLines <= v->MaximumMaxVStartupLines);
+	ASSERT(v->PrefetchAndImmediateFlipSupported);
+
+	// Unbounded Request Enabled
+	CalculateUnboundedRequestAndCompressedBufferSize(
+			v->DETBufferSizeInKByte[0],
+			v->ConfigReturnBufferSizeInKByte,
+			v->UseUnboundedRequesting,
+			v->TotalActiveDPP,
+			NoChromaPlanes,
+			v->MaxNumDPP,
+			v->CompressedBufferSegmentSizeInkByte,
+			v->Output,
+			&v->UnboundedRequestEnabled,
+			&v->CompressedBufferSizeInkByte);
+
+	//Watermarks and NB P-State/DRAM Clock Change Support
+	{
+		enum clock_change_support DRAMClockChangeSupport; // dummy
+
+		CalculateWatermarksAndDRAMSpeedChangeSupport(
+				mode_lib,
+				PrefetchMode,
+				v->DCFCLK,
+				v->ReturnBW,
+				v->UrgentLatency,
+				v->UrgentExtraLatency,
+				v->SOCCLK,
+				v->DCFCLKDeepSleep,
+				v->DETBufferSizeY,
+				v->DETBufferSizeC,
+				v->SwathHeightY,
+				v->SwathHeightC,
+				v->SwathWidthY,
+				v->SwathWidthC,
+				v->DPPPerPlane,
+				v->BytePerPixelDETY,
+				v->BytePerPixelDETC,
+				v->UnboundedRequestEnabled,
+				v->CompressedBufferSizeInkByte,
+				&DRAMClockChangeSupport,
+				&v->StutterExitWatermark,
+				&v->StutterEnterPlusExitWatermark,
+				&v->Z8StutterExitWatermark,
+				&v->Z8StutterEnterPlusExitWatermark);
+
+		for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+			if (v->WritebackEnable[k] == true) {
+				v->WritebackAllowDRAMClockChangeEndPosition[k] = dml_max(
+						0,
+						v->VStartup[k] * v->HTotal[k] / v->PixelClock[k] - v->WritebackDRAMClockChangeWatermark);
+			} else {
+				v->WritebackAllowDRAMClockChangeEndPosition[k] = 0;
+			}
+		}
+	}
+
+	//Display Pipeline Delivery Time in Prefetch, Groups
+	CalculatePixelDeliveryTimes(
+			v->NumberOfActivePlanes,
+			v->VRatio,
+			v->VRatioChroma,
+			v->VRatioPrefetchY,
+			v->VRatioPrefetchC,
+			v->swath_width_luma_ub,
+			v->swath_width_chroma_ub,
+			v->DPPPerPlane,
+			v->HRatio,
+			v->HRatioChroma,
+			v->PixelClock,
+			v->PSCL_THROUGHPUT_LUMA,
+			v->PSCL_THROUGHPUT_CHROMA,
+			v->DPPCLK,
+			v->BytePerPixelC,
+			v->SourceScan,
+			v->NumberOfCursors,
+			v->CursorWidth,
+			v->CursorBPP,
+			v->BlockWidth256BytesY,
+			v->BlockHeight256BytesY,
+			v->BlockWidth256BytesC,
+			v->BlockHeight256BytesC,
+			v->DisplayPipeLineDeliveryTimeLuma,
+			v->DisplayPipeLineDeliveryTimeChroma,
+			v->DisplayPipeLineDeliveryTimeLumaPrefetch,
+			v->DisplayPipeLineDeliveryTimeChromaPrefetch,
+			v->DisplayPipeRequestDeliveryTimeLuma,
+			v->DisplayPipeRequestDeliveryTimeChroma,
+			v->DisplayPipeRequestDeliveryTimeLumaPrefetch,
+			v->DisplayPipeRequestDeliveryTimeChromaPrefetch,
+			v->CursorRequestDeliveryTime,
+			v->CursorRequestDeliveryTimePrefetch);
+
+	CalculateMetaAndPTETimes(
+			v->NumberOfActivePlanes,
+			v->GPUVMEnable,
+			v->MetaChunkSize,
+			v->MinMetaChunkSizeBytes,
+			v->HTotal,
+			v->VRatio,
+			v->VRatioChroma,
+			v->DestinationLinesToRequestRowInVBlank,
+			v->DestinationLinesToRequestRowInImmediateFlip,
+			v->DCCEnable,
+			v->PixelClock,
+			v->BytePerPixelY,
+			v->BytePerPixelC,
+			v->SourceScan,
+			v->dpte_row_height,
+			v->dpte_row_height_chroma,
+			v->meta_row_width,
+			v->meta_row_width_chroma,
+			v->meta_row_height,
+			v->meta_row_height_chroma,
+			v->meta_req_width,
+			v->meta_req_width_chroma,
+			v->meta_req_height,
+			v->meta_req_height_chroma,
+			v->dpte_group_bytes,
+			v->PTERequestSizeY,
+			v->PTERequestSizeC,
+			v->PixelPTEReqWidthY,
+			v->PixelPTEReqHeightY,
+			v->PixelPTEReqWidthC,
+			v->PixelPTEReqHeightC,
+			v->dpte_row_width_luma_ub,
+			v->dpte_row_width_chroma_ub,
+			v->DST_Y_PER_PTE_ROW_NOM_L,
+			v->DST_Y_PER_PTE_ROW_NOM_C,
+			v->DST_Y_PER_META_ROW_NOM_L,
+			v->DST_Y_PER_META_ROW_NOM_C,
+			v->TimePerMetaChunkNominal,
+			v->TimePerChromaMetaChunkNominal,
+			v->TimePerMetaChunkVBlank,
+			v->TimePerChromaMetaChunkVBlank,
+			v->TimePerMetaChunkFlip,
+			v->TimePerChromaMetaChunkFlip,
+			v->time_per_pte_group_nom_luma,
+			v->time_per_pte_group_vblank_luma,
+			v->time_per_pte_group_flip_luma,
+			v->time_per_pte_group_nom_chroma,
+			v->time_per_pte_group_vblank_chroma,
+			v->time_per_pte_group_flip_chroma);
+
+	CalculateVMGroupAndRequestTimes(
+			v->NumberOfActivePlanes,
+			v->GPUVMEnable,
+			v->GPUVMMaxPageTableLevels,
+			v->HTotal,
+			v->BytePerPixelC,
+			v->DestinationLinesToRequestVMInVBlank,
+			v->DestinationLinesToRequestVMInImmediateFlip,
+			v->DCCEnable,
+			v->PixelClock,
+			v->dpte_row_width_luma_ub,
+			v->dpte_row_width_chroma_ub,
+			v->vm_group_bytes,
+			v->dpde0_bytes_per_frame_ub_l,
+			v->dpde0_bytes_per_frame_ub_c,
+			v->meta_pte_bytes_per_frame_ub_l,
+			v->meta_pte_bytes_per_frame_ub_c,
+			v->TimePerVMGroupVBlank,
+			v->TimePerVMGroupFlip,
+			v->TimePerVMRequestVBlank,
+			v->TimePerVMRequestFlip);
+
+	// Min TTUVBlank
+	for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+		if (PrefetchMode == 0) {
+			v->AllowDRAMClockChangeDuringVBlank[k] = true;
+			v->AllowDRAMSelfRefreshDuringVBlank[k] = true;
+			v->MinTTUVBlank[k] = dml_max(
+					v->DRAMClockChangeWatermark,
+					dml_max(v->StutterEnterPlusExitWatermark, v->UrgentWatermark));
+		} else if (PrefetchMode == 1) {
+			v->AllowDRAMClockChangeDuringVBlank[k] = false;
+			v->AllowDRAMSelfRefreshDuringVBlank[k] = true;
+			v->MinTTUVBlank[k] = dml_max(v->StutterEnterPlusExitWatermark, v->UrgentWatermark);
+		} else {
+			v->AllowDRAMClockChangeDuringVBlank[k] = false;
+			v->AllowDRAMSelfRefreshDuringVBlank[k] = false;
+			v->MinTTUVBlank[k] = v->UrgentWatermark;
+		}
+		if (!v->DynamicMetadataEnable[k])
+			v->MinTTUVBlank[k] = v->TCalc + v->MinTTUVBlank[k];
+	}
+
+	// DCC Configuration
+	v->ActiveDPPs = 0;
+	for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+		CalculateDCCConfiguration(v->DCCEnable[k], false, // We should always know the direction DCCProgrammingAssumesScanDirectionUnknown,
+				v->SourcePixelFormat[k],
+				v->SurfaceWidthY[k],
+				v->SurfaceWidthC[k],
+				v->SurfaceHeightY[k],
+				v->SurfaceHeightC[k],
+				v->DETBufferSizeInKByte[0] * 1024,
+				v->BlockHeight256BytesY[k],
+				v->BlockHeight256BytesC[k],
+				v->SurfaceTiling[k],
+				v->BytePerPixelY[k],
+				v->BytePerPixelC[k],
+				v->BytePerPixelDETY[k],
+				v->BytePerPixelDETC[k],
+				v->SourceScan[k],
+				&v->DCCYMaxUncompressedBlock[k],
+				&v->DCCCMaxUncompressedBlock[k],
+				&v->DCCYMaxCompressedBlock[k],
+				&v->DCCCMaxCompressedBlock[k],
+				&v->DCCYIndependentBlock[k],
+				&v->DCCCIndependentBlock[k]);
+	}
+
+	// VStartup Adjustment
+	for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+		bool isInterlaceTiming;
+		double Tvstartup_margin = (v->MaxVStartupLines[k] - v->VStartup[k]) * v->HTotal[k] / v->PixelClock[k];
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%d, MinTTUVBlank = %f (before margin)\n", __func__, k, v->MinTTUVBlank[k]);
+#endif
+
+		v->MinTTUVBlank[k] = v->MinTTUVBlank[k] + Tvstartup_margin;
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%d, Tvstartup_margin = %f\n", __func__, k, Tvstartup_margin);
+		dml_print("DML::%s: k=%d, MaxVStartupLines = %d\n", __func__, k, v->MaxVStartupLines[k]);
+		dml_print("DML::%s: k=%d, VStartup = %d\n", __func__, k, v->VStartup[k]);
+		dml_print("DML::%s: k=%d, MinTTUVBlank = %f\n", __func__, k, v->MinTTUVBlank[k]);
+#endif
+
+		v->Tdmdl[k] = v->Tdmdl[k] + Tvstartup_margin;
+		if (v->DynamicMetadataEnable[k] && v->DynamicMetadataVMEnabled) {
+			v->Tdmdl_vm[k] = v->Tdmdl_vm[k] + Tvstartup_margin;
+		}
+
+		isInterlaceTiming = (v->Interlace[k] && !v->ProgressiveToInterlaceUnitInOPP);
+		v->VStartup[k] = (isInterlaceTiming ? (2 * v->MaxVStartupLines[k]) : v->MaxVStartupLines[k]);
+		if (v->Interlace[k] && !v->ProgressiveToInterlaceUnitInOPP) {
+			v->MIN_DST_Y_NEXT_START[k] = dml_floor((v->VTotal[k] - v->VFrontPorch[k] + v->VTotal[k] - v->VActive[k] - v->VStartup[k]) / 2.0, 1.0);
+		} else {
+			v->MIN_DST_Y_NEXT_START[k] = v->VTotal[k] - v->VFrontPorch[k] + v->VTotal[k] - v->VActive[k] - v->VStartup[k];
+		}
+		v->MIN_DST_Y_NEXT_START[k] += dml_floor(4.0 * v->TSetup[k] / ((double)v->HTotal[k] / v->PixelClock[k]), 1.0) / 4.0;
+		if (((v->VUpdateOffsetPix[k] + v->VUpdateWidthPix[k] + v->VReadyOffsetPix[k]) / v->HTotal[k])
+				<= (isInterlaceTiming ?
+						dml_floor((v->VTotal[k] - v->VActive[k] - v->VFrontPorch[k] - v->VStartup[k]) / 2.0, 1.0) :
+						(int) (v->VTotal[k] - v->VActive[k] - v->VFrontPorch[k] - v->VStartup[k]))) {
+			v->VREADY_AT_OR_AFTER_VSYNC[k] = true;
+		} else {
+			v->VREADY_AT_OR_AFTER_VSYNC[k] = false;
+		}
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%d, VStartup = %d (max)\n", __func__, k, v->VStartup[k]);
+		dml_print("DML::%s: k=%d, VUpdateOffsetPix = %d\n", __func__, k, v->VUpdateOffsetPix[k]);
+		dml_print("DML::%s: k=%d, VUpdateWidthPix = %d\n", __func__, k, v->VUpdateWidthPix[k]);
+		dml_print("DML::%s: k=%d, VReadyOffsetPix = %d\n", __func__, k, v->VReadyOffsetPix[k]);
+		dml_print("DML::%s: k=%d, HTotal = %d\n", __func__, k, v->HTotal[k]);
+		dml_print("DML::%s: k=%d, VTotal = %d\n", __func__, k, v->VTotal[k]);
+		dml_print("DML::%s: k=%d, VActive = %d\n", __func__, k, v->VActive[k]);
+		dml_print("DML::%s: k=%d, VFrontPorch = %d\n", __func__, k, v->VFrontPorch[k]);
+		dml_print("DML::%s: k=%d, VStartup = %d\n", __func__, k, v->VStartup[k]);
+		dml_print("DML::%s: k=%d, MIN_DST_Y_NEXT_START = %f\n", __func__, k, v->MIN_DST_Y_NEXT_START[k]);
+		dml_print("DML::%s: k=%d, VREADY_AT_OR_AFTER_VSYNC = %d\n", __func__, k, v->VREADY_AT_OR_AFTER_VSYNC[k]);
+#endif
+	}
+
+	{
+		//Maximum Bandwidth Used
+		double TotalWRBandwidth = 0;
+		double MaxPerPlaneVActiveWRBandwidth = 0;
+		double WRBandwidth = 0;
+
+		for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+			if (v->WritebackEnable[k] == true && v->WritebackPixelFormat[k] == dm_444_32) {
+				WRBandwidth = v->WritebackDestinationWidth[k] * v->WritebackDestinationHeight[k]
+						/ (v->HTotal[k] * v->WritebackSourceHeight[k] / v->PixelClock[k]) * 4;
+			} else if (v->WritebackEnable[k] == true) {
+				WRBandwidth = v->WritebackDestinationWidth[k] * v->WritebackDestinationHeight[k]
+						/ (v->HTotal[k] * v->WritebackSourceHeight[k] / v->PixelClock[k]) * 8;
+			}
+			TotalWRBandwidth = TotalWRBandwidth + WRBandwidth;
+			MaxPerPlaneVActiveWRBandwidth = dml_max(MaxPerPlaneVActiveWRBandwidth, WRBandwidth);
+		}
+
+		v->TotalDataReadBandwidth = 0;
+		for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+			v->TotalDataReadBandwidth = v->TotalDataReadBandwidth + v->ReadBandwidthPlaneLuma[k] + v->ReadBandwidthPlaneChroma[k];
+		}
+	}
+	// Stutter Efficiency
+	CalculateStutterEfficiency(
+			mode_lib,
+			v->CompressedBufferSizeInkByte,
+			v->UnboundedRequestEnabled,
+			v->ConfigReturnBufferSizeInKByte,
+			v->MetaFIFOSizeInKEntries,
+			v->ZeroSizeBufferEntries,
+			v->NumberOfActivePlanes,
+			v->ROBBufferSizeInKByte,
+			v->TotalDataReadBandwidth,
+			v->DCFCLK,
+			v->ReturnBW,
+			v->COMPBUF_RESERVED_SPACE_64B,
+			v->COMPBUF_RESERVED_SPACE_ZS,
+			v->SRExitTime,
+			v->SRExitZ8Time,
+			v->SynchronizedVBlank,
+			v->StutterEnterPlusExitWatermark,
+			v->Z8StutterEnterPlusExitWatermark,
+			v->ProgressiveToInterlaceUnitInOPP,
+			v->Interlace,
+			v->MinTTUVBlank,
+			v->DPPPerPlane,
+			v->DETBufferSizeY,
+			v->BytePerPixelY,
+			v->BytePerPixelDETY,
+			v->SwathWidthY,
+			v->SwathHeightY,
+			v->SwathHeightC,
+			v->DCCRateLuma,
+			v->DCCRateChroma,
+			v->DCCFractionOfZeroSizeRequestsLuma,
+			v->DCCFractionOfZeroSizeRequestsChroma,
+			v->HTotal,
+			v->VTotal,
+			v->PixelClock,
+			v->VRatio,
+			v->SourceScan,
+			v->BlockHeight256BytesY,
+			v->BlockWidth256BytesY,
+			v->BlockHeight256BytesC,
+			v->BlockWidth256BytesC,
+			v->DCCYMaxUncompressedBlock,
+			v->DCCCMaxUncompressedBlock,
+			v->VActive,
+			v->DCCEnable,
+			v->WritebackEnable,
+			v->ReadBandwidthPlaneLuma,
+			v->ReadBandwidthPlaneChroma,
+			v->meta_row_bw,
+			v->dpte_row_bw,
+			&v->StutterEfficiencyNotIncludingVBlank,
+			&v->StutterEfficiency,
+			&v->NumberOfStutterBurstsPerFrame,
+			&v->Z8StutterEfficiencyNotIncludingVBlank,
+			&v->Z8StutterEfficiency,
+			&v->Z8NumberOfStutterBurstsPerFrame,
+			&v->StutterPeriod);
+}
+
+static void DisplayPipeConfiguration(struct display_mode_lib *mode_lib)
+{
+	struct vba_vars_st *v = &mode_lib->vba;
+	// Display Pipe Configuration
+	double BytePerPixDETY[DC__NUM_DPP__MAX];
+	double BytePerPixDETC[DC__NUM_DPP__MAX];
+	int BytePerPixY[DC__NUM_DPP__MAX];
+	int BytePerPixC[DC__NUM_DPP__MAX];
+	int Read256BytesBlockHeightY[DC__NUM_DPP__MAX];
+	int Read256BytesBlockHeightC[DC__NUM_DPP__MAX];
+	int Read256BytesBlockWidthY[DC__NUM_DPP__MAX];
+	int Read256BytesBlockWidthC[DC__NUM_DPP__MAX];
+	double dummy1[DC__NUM_DPP__MAX];
+	double dummy2[DC__NUM_DPP__MAX];
+	double dummy3[DC__NUM_DPP__MAX];
+	double dummy4[DC__NUM_DPP__MAX];
+	int dummy5[DC__NUM_DPP__MAX];
+	int dummy6[DC__NUM_DPP__MAX];
+	bool dummy7[DC__NUM_DPP__MAX];
+	bool dummysinglestring;
+
+	unsigned int k;
+
+	for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+
+		CalculateBytePerPixelAnd256BBlockSizes(
+				v->SourcePixelFormat[k],
+				v->SurfaceTiling[k],
+				&BytePerPixY[k],
+				&BytePerPixC[k],
+				&BytePerPixDETY[k],
+				&BytePerPixDETC[k],
+				&Read256BytesBlockHeightY[k],
+				&Read256BytesBlockHeightC[k],
+				&Read256BytesBlockWidthY[k],
+				&Read256BytesBlockWidthC[k]);
+	}
+
+	CalculateSwathAndDETConfiguration(
+			false,
+			v->NumberOfActivePlanes,
+			v->DETBufferSizeInKByte[0],
+			dummy1,
+			dummy2,
+			v->SourceScan,
+			v->SourcePixelFormat,
+			v->SurfaceTiling,
+			v->ViewportWidth,
+			v->ViewportHeight,
+			v->SurfaceWidthY,
+			v->SurfaceWidthC,
+			v->SurfaceHeightY,
+			v->SurfaceHeightC,
+			Read256BytesBlockHeightY,
+			Read256BytesBlockHeightC,
+			Read256BytesBlockWidthY,
+			Read256BytesBlockWidthC,
+			v->ODMCombineEnabled,
+			v->BlendingAndTiming,
+			BytePerPixY,
+			BytePerPixC,
+			BytePerPixDETY,
+			BytePerPixDETC,
+			v->HActive,
+			v->HRatio,
+			v->HRatioChroma,
+			v->DPPPerPlane,
+			dummy5,
+			dummy6,
+			dummy3,
+			dummy4,
+			v->SwathHeightY,
+			v->SwathHeightC,
+			v->DETBufferSizeY,
+			v->DETBufferSizeC,
+			dummy7,
+			&dummysinglestring);
+}
+
+static bool CalculateBytePerPixelAnd256BBlockSizes(
+		enum source_format_class SourcePixelFormat,
+		enum dm_swizzle_mode SurfaceTiling,
+		unsigned int *BytePerPixelY,
+		unsigned int *BytePerPixelC,
+		double *BytePerPixelDETY,
+		double *BytePerPixelDETC,
+		unsigned int *BlockHeight256BytesY,
+		unsigned int *BlockHeight256BytesC,
+		unsigned int *BlockWidth256BytesY,
+		unsigned int *BlockWidth256BytesC)
+{
+	if (SourcePixelFormat == dm_444_64) {
+		*BytePerPixelDETY = 8;
+		*BytePerPixelDETC = 0;
+		*BytePerPixelY = 8;
+		*BytePerPixelC = 0;
+	} else if (SourcePixelFormat == dm_444_32 || SourcePixelFormat == dm_rgbe) {
+		*BytePerPixelDETY = 4;
+		*BytePerPixelDETC = 0;
+		*BytePerPixelY = 4;
+		*BytePerPixelC = 0;
+	} else if (SourcePixelFormat == dm_444_16) {
+		*BytePerPixelDETY = 2;
+		*BytePerPixelDETC = 0;
+		*BytePerPixelY = 2;
+		*BytePerPixelC = 0;
+	} else if (SourcePixelFormat == dm_444_8) {
+		*BytePerPixelDETY = 1;
+		*BytePerPixelDETC = 0;
+		*BytePerPixelY = 1;
+		*BytePerPixelC = 0;
+	} else if (SourcePixelFormat == dm_rgbe_alpha) {
+		*BytePerPixelDETY = 4;
+		*BytePerPixelDETC = 1;
+		*BytePerPixelY = 4;
+		*BytePerPixelC = 1;
+	} else if (SourcePixelFormat == dm_420_8) {
+		*BytePerPixelDETY = 1;
+		*BytePerPixelDETC = 2;
+		*BytePerPixelY = 1;
+		*BytePerPixelC = 2;
+	} else if (SourcePixelFormat == dm_420_12) {
+		*BytePerPixelDETY = 2;
+		*BytePerPixelDETC = 4;
+		*BytePerPixelY = 2;
+		*BytePerPixelC = 4;
+	} else {
+		*BytePerPixelDETY = 4.0 / 3;
+		*BytePerPixelDETC = 8.0 / 3;
+		*BytePerPixelY = 2;
+		*BytePerPixelC = 4;
+	}
+
+	if ((SourcePixelFormat == dm_444_64 || SourcePixelFormat == dm_444_32 || SourcePixelFormat == dm_444_16 || SourcePixelFormat == dm_444_8 || SourcePixelFormat == dm_mono_16
+			|| SourcePixelFormat == dm_mono_8 || SourcePixelFormat == dm_rgbe)) {
+		if (SurfaceTiling == dm_sw_linear) {
+			*BlockHeight256BytesY = 1;
+		} else if (SourcePixelFormat == dm_444_64) {
+			*BlockHeight256BytesY = 4;
+		} else if (SourcePixelFormat == dm_444_8) {
+			*BlockHeight256BytesY = 16;
+		} else {
+			*BlockHeight256BytesY = 8;
+		}
+		*BlockWidth256BytesY = 256U / *BytePerPixelY / *BlockHeight256BytesY;
+		*BlockHeight256BytesC = 0;
+		*BlockWidth256BytesC = 0;
+	} else {
+		if (SurfaceTiling == dm_sw_linear) {
+			*BlockHeight256BytesY = 1;
+			*BlockHeight256BytesC = 1;
+		} else if (SourcePixelFormat == dm_rgbe_alpha) {
+			*BlockHeight256BytesY = 8;
+			*BlockHeight256BytesC = 16;
+		} else if (SourcePixelFormat == dm_420_8) {
+			*BlockHeight256BytesY = 16;
+			*BlockHeight256BytesC = 8;
+		} else {
+			*BlockHeight256BytesY = 8;
+			*BlockHeight256BytesC = 8;
+		}
+		*BlockWidth256BytesY = 256U / *BytePerPixelY / *BlockHeight256BytesY;
+		*BlockWidth256BytesC = 256U / *BytePerPixelC / *BlockHeight256BytesC;
+	}
+	return true;
+}
+
+static double CalculateTWait(unsigned int PrefetchMode, double DRAMClockChangeLatency, double UrgentLatency, double SREnterPlusExitTime)
+{
+	if (PrefetchMode == 0) {
+		return dml_max(DRAMClockChangeLatency + UrgentLatency, dml_max(SREnterPlusExitTime, UrgentLatency));
+	} else if (PrefetchMode == 1) {
+		return dml_max(SREnterPlusExitTime, UrgentLatency);
+	} else {
+		return UrgentLatency;
+	}
+}
+
+double dml314_CalculateWriteBackDISPCLK(
+		enum source_format_class WritebackPixelFormat,
+		double PixelClock,
+		double WritebackHRatio,
+		double WritebackVRatio,
+		unsigned int WritebackHTaps,
+		unsigned int WritebackVTaps,
+		long WritebackSourceWidth,
+		long WritebackDestinationWidth,
+		unsigned int HTotal,
+		unsigned int WritebackLineBufferSize)
+{
+	double DISPCLK_H, DISPCLK_V, DISPCLK_HB;
+
+	DISPCLK_H = PixelClock * dml_ceil(WritebackHTaps / 8.0, 1) / WritebackHRatio;
+	DISPCLK_V = PixelClock * (WritebackVTaps * dml_ceil(WritebackDestinationWidth / 6.0, 1) + 8.0) / HTotal;
+	DISPCLK_HB = PixelClock * WritebackVTaps * (WritebackDestinationWidth * WritebackVTaps - WritebackLineBufferSize / 57.0) / 6.0 / WritebackSourceWidth;
+	return dml_max3(DISPCLK_H, DISPCLK_V, DISPCLK_HB);
+}
+
+static double CalculateWriteBackDelay(
+		enum source_format_class WritebackPixelFormat,
+		double WritebackHRatio,
+		double WritebackVRatio,
+		unsigned int WritebackVTaps,
+		int WritebackDestinationWidth,
+		int WritebackDestinationHeight,
+		int WritebackSourceHeight,
+		unsigned int HTotal)
+{
+	double CalculateWriteBackDelay;
+	double Line_length;
+	double Output_lines_last_notclamped;
+	double WritebackVInit;
+
+	WritebackVInit = (WritebackVRatio + WritebackVTaps + 1) / 2;
+	Line_length = dml_max((double) WritebackDestinationWidth, dml_ceil(WritebackDestinationWidth / 6.0, 1) * WritebackVTaps);
+	Output_lines_last_notclamped = WritebackDestinationHeight - 1 - dml_ceil((WritebackSourceHeight - WritebackVInit) / WritebackVRatio, 1);
+	if (Output_lines_last_notclamped < 0) {
+		CalculateWriteBackDelay = 0;
+	} else {
+		CalculateWriteBackDelay = Output_lines_last_notclamped * Line_length + (HTotal - WritebackDestinationWidth) + 80;
+	}
+	return CalculateWriteBackDelay;
+}
+
+static void CalculateVupdateAndDynamicMetadataParameters(
+		int MaxInterDCNTileRepeaters,
+		double DPPCLK,
+		double DISPCLK,
+		double DCFClkDeepSleep,
+		double PixelClock,
+		int HTotal,
+		int VBlank,
+		int DynamicMetadataTransmittedBytes,
+		int DynamicMetadataLinesBeforeActiveRequired,
+		int InterlaceEnable,
+		bool ProgressiveToInterlaceUnitInOPP,
+		double *TSetup,
+		double *Tdmbf,
+		double *Tdmec,
+		double *Tdmsks,
+		int *VUpdateOffsetPix,
+		double *VUpdateWidthPix,
+		double *VReadyOffsetPix)
+{
+	double TotalRepeaterDelayTime;
+
+	TotalRepeaterDelayTime = MaxInterDCNTileRepeaters * (2 / DPPCLK + 3 / DISPCLK);
+	*VUpdateWidthPix = dml_ceil((14.0 / DCFClkDeepSleep + 12.0 / DPPCLK + TotalRepeaterDelayTime) * PixelClock, 1.0);
+	*VReadyOffsetPix = dml_ceil(dml_max(150.0 / DPPCLK, TotalRepeaterDelayTime + 20.0 / DCFClkDeepSleep + 10.0 / DPPCLK) * PixelClock, 1.0);
+	*VUpdateOffsetPix = dml_ceil(HTotal / 4.0, 1);
+	*TSetup = (*VUpdateOffsetPix + *VUpdateWidthPix + *VReadyOffsetPix) / PixelClock;
+	*Tdmbf = DynamicMetadataTransmittedBytes / 4.0 / DISPCLK;
+	*Tdmec = HTotal / PixelClock;
+	if (DynamicMetadataLinesBeforeActiveRequired == 0) {
+		*Tdmsks = VBlank * HTotal / PixelClock / 2.0;
+	} else {
+		*Tdmsks = DynamicMetadataLinesBeforeActiveRequired * HTotal / PixelClock;
+	}
+	if (InterlaceEnable == 1 && ProgressiveToInterlaceUnitInOPP == false) {
+		*Tdmsks = *Tdmsks / 2;
+	}
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: VUpdateWidthPix = %d\n", __func__, *VUpdateWidthPix);
+	dml_print("DML::%s: VReadyOffsetPix = %d\n", __func__, *VReadyOffsetPix);
+	dml_print("DML::%s: VUpdateOffsetPix = %d\n", __func__, *VUpdateOffsetPix);
+#endif
+}
+
+static void CalculateRowBandwidth(
+		bool GPUVMEnable,
+		enum source_format_class SourcePixelFormat,
+		double VRatio,
+		double VRatioChroma,
+		bool DCCEnable,
+		double LineTime,
+		unsigned int MetaRowByteLuma,
+		unsigned int MetaRowByteChroma,
+		unsigned int meta_row_height_luma,
+		unsigned int meta_row_height_chroma,
+		unsigned int PixelPTEBytesPerRowLuma,
+		unsigned int PixelPTEBytesPerRowChroma,
+		unsigned int dpte_row_height_luma,
+		unsigned int dpte_row_height_chroma,
+		double *meta_row_bw,
+		double *dpte_row_bw)
+{
+	if (DCCEnable != true) {
+		*meta_row_bw = 0;
+	} else if (SourcePixelFormat == dm_420_8 || SourcePixelFormat == dm_420_10 || SourcePixelFormat == dm_420_12 || SourcePixelFormat == dm_rgbe_alpha) {
+		*meta_row_bw = VRatio * MetaRowByteLuma / (meta_row_height_luma * LineTime) + VRatioChroma * MetaRowByteChroma / (meta_row_height_chroma * LineTime);
+	} else {
+		*meta_row_bw = VRatio * MetaRowByteLuma / (meta_row_height_luma * LineTime);
+	}
+
+	if (GPUVMEnable != true) {
+		*dpte_row_bw = 0;
+	} else if (SourcePixelFormat == dm_420_8 || SourcePixelFormat == dm_420_10 || SourcePixelFormat == dm_420_12 || SourcePixelFormat == dm_rgbe_alpha) {
+		*dpte_row_bw = VRatio * PixelPTEBytesPerRowLuma / (dpte_row_height_luma * LineTime)
+				+ VRatioChroma * PixelPTEBytesPerRowChroma / (dpte_row_height_chroma * LineTime);
+	} else {
+		*dpte_row_bw = VRatio * PixelPTEBytesPerRowLuma / (dpte_row_height_luma * LineTime);
+	}
+}
+
+static void CalculateFlipSchedule(
+		struct display_mode_lib *mode_lib,
+		unsigned int k,
+		double HostVMInefficiencyFactor,
+		double UrgentExtraLatency,
+		double UrgentLatency,
+		double PDEAndMetaPTEBytesPerFrame,
+		double MetaRowBytes,
+		double DPTEBytesPerRow)
+{
+	struct vba_vars_st *v = &mode_lib->vba;
+	double min_row_time = 0.0;
+	unsigned int HostVMDynamicLevelsTrips;
+	double TimeForFetchingMetaPTEImmediateFlip;
+	double TimeForFetchingRowInVBlankImmediateFlip;
+	double ImmediateFlipBW = 1.0;
+	double LineTime = v->HTotal[k] / v->PixelClock[k];
+
+	if (v->GPUVMEnable == true && v->HostVMEnable == true) {
+		HostVMDynamicLevelsTrips = v->HostVMMaxNonCachedPageTableLevels;
+	} else {
+		HostVMDynamicLevelsTrips = 0;
+	}
+
+	if (v->GPUVMEnable == true || v->DCCEnable[k] == true) {
+		ImmediateFlipBW = (PDEAndMetaPTEBytesPerFrame + MetaRowBytes + DPTEBytesPerRow) * v->BandwidthAvailableForImmediateFlip / v->TotImmediateFlipBytes;
+	}
+
+	if (v->GPUVMEnable == true) {
+		TimeForFetchingMetaPTEImmediateFlip = dml_max3(
+				v->Tno_bw[k] + PDEAndMetaPTEBytesPerFrame * HostVMInefficiencyFactor / ImmediateFlipBW,
+				UrgentExtraLatency + UrgentLatency * (v->GPUVMMaxPageTableLevels * (HostVMDynamicLevelsTrips + 1) - 1),
+				LineTime / 4.0);
+	} else {
+		TimeForFetchingMetaPTEImmediateFlip = 0;
+	}
+
+	v->DestinationLinesToRequestVMInImmediateFlip[k] = dml_ceil(4.0 * (TimeForFetchingMetaPTEImmediateFlip / LineTime), 1) / 4.0;
+	if ((v->GPUVMEnable == true || v->DCCEnable[k] == true)) {
+		TimeForFetchingRowInVBlankImmediateFlip = dml_max3(
+				(MetaRowBytes + DPTEBytesPerRow * HostVMInefficiencyFactor) / ImmediateFlipBW,
+				UrgentLatency * (HostVMDynamicLevelsTrips + 1),
+				LineTime / 4);
+	} else {
+		TimeForFetchingRowInVBlankImmediateFlip = 0;
+	}
+
+	v->DestinationLinesToRequestRowInImmediateFlip[k] = dml_ceil(4.0 * (TimeForFetchingRowInVBlankImmediateFlip / LineTime), 1) / 4.0;
+
+	if (v->GPUVMEnable == true) {
+		v->final_flip_bw[k] = dml_max(
+				PDEAndMetaPTEBytesPerFrame * HostVMInefficiencyFactor / (v->DestinationLinesToRequestVMInImmediateFlip[k] * LineTime),
+				(MetaRowBytes + DPTEBytesPerRow * HostVMInefficiencyFactor) / (v->DestinationLinesToRequestRowInImmediateFlip[k] * LineTime));
+	} else if ((v->GPUVMEnable == true || v->DCCEnable[k] == true)) {
+		v->final_flip_bw[k] = (MetaRowBytes + DPTEBytesPerRow * HostVMInefficiencyFactor) / (v->DestinationLinesToRequestRowInImmediateFlip[k] * LineTime);
+	} else {
+		v->final_flip_bw[k] = 0;
+	}
+
+	if (v->SourcePixelFormat[k] == dm_420_8 || v->SourcePixelFormat[k] == dm_420_10 || v->SourcePixelFormat[k] == dm_rgbe_alpha) {
+		if (v->GPUVMEnable == true && v->DCCEnable[k] != true) {
+			min_row_time = dml_min(v->dpte_row_height[k] * LineTime / v->VRatio[k], v->dpte_row_height_chroma[k] * LineTime / v->VRatioChroma[k]);
+		} else if (v->GPUVMEnable != true && v->DCCEnable[k] == true) {
+			min_row_time = dml_min(v->meta_row_height[k] * LineTime / v->VRatio[k], v->meta_row_height_chroma[k] * LineTime / v->VRatioChroma[k]);
+		} else {
+			min_row_time = dml_min4(
+					v->dpte_row_height[k] * LineTime / v->VRatio[k],
+					v->meta_row_height[k] * LineTime / v->VRatio[k],
+					v->dpte_row_height_chroma[k] * LineTime / v->VRatioChroma[k],
+					v->meta_row_height_chroma[k] * LineTime / v->VRatioChroma[k]);
+		}
+	} else {
+		if (v->GPUVMEnable == true && v->DCCEnable[k] != true) {
+			min_row_time = v->dpte_row_height[k] * LineTime / v->VRatio[k];
+		} else if (v->GPUVMEnable != true && v->DCCEnable[k] == true) {
+			min_row_time = v->meta_row_height[k] * LineTime / v->VRatio[k];
+		} else {
+			min_row_time = dml_min(v->dpte_row_height[k] * LineTime / v->VRatio[k], v->meta_row_height[k] * LineTime / v->VRatio[k]);
+		}
+	}
+
+	if (v->DestinationLinesToRequestVMInImmediateFlip[k] >= 32 || v->DestinationLinesToRequestRowInImmediateFlip[k] >= 16
+			|| TimeForFetchingMetaPTEImmediateFlip + 2 * TimeForFetchingRowInVBlankImmediateFlip > min_row_time) {
+		v->ImmediateFlipSupportedForPipe[k] = false;
+	} else {
+		v->ImmediateFlipSupportedForPipe[k] = true;
+	}
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: DestinationLinesToRequestVMInImmediateFlip = %f\n", __func__, v->DestinationLinesToRequestVMInImmediateFlip[k]);
+	dml_print("DML::%s: DestinationLinesToRequestRowInImmediateFlip = %f\n", __func__, v->DestinationLinesToRequestRowInImmediateFlip[k]);
+	dml_print("DML::%s: TimeForFetchingMetaPTEImmediateFlip = %f\n", __func__, TimeForFetchingMetaPTEImmediateFlip);
+	dml_print("DML::%s: TimeForFetchingRowInVBlankImmediateFlip = %f\n", __func__, TimeForFetchingRowInVBlankImmediateFlip);
+	dml_print("DML::%s: min_row_time = %f\n", __func__, min_row_time);
+	dml_print("DML::%s: ImmediateFlipSupportedForPipe = %d\n", __func__, v->ImmediateFlipSupportedForPipe[k]);
+#endif
+
+}
+
+static double TruncToValidBPP(
+		double LinkBitRate,
+		int Lanes,
+		int HTotal,
+		int HActive,
+		double PixelClock,
+		double DesiredBPP,
+		bool DSCEnable,
+		enum output_encoder_class Output,
+		enum output_format_class Format,
+		unsigned int DSCInputBitPerComponent,
+		int DSCSlices,
+		int AudioRate,
+		int AudioLayout,
+		enum odm_combine_mode ODMCombine)
+{
+	double MaxLinkBPP;
+	int MinDSCBPP;
+	double MaxDSCBPP;
+	int NonDSCBPP0;
+	int NonDSCBPP1;
+	int NonDSCBPP2;
+
+	if (Format == dm_420) {
+		NonDSCBPP0 = 12;
+		NonDSCBPP1 = 15;
+		NonDSCBPP2 = 18;
+		MinDSCBPP = 6;
+		MaxDSCBPP = 1.5 * DSCInputBitPerComponent - 1.0 / 16;
+	} else if (Format == dm_444) {
+		NonDSCBPP0 = 24;
+		NonDSCBPP1 = 30;
+		NonDSCBPP2 = 36;
+		MinDSCBPP = 8;
+		MaxDSCBPP = 3 * DSCInputBitPerComponent - 1.0 / 16;
+	} else {
+
+		NonDSCBPP0 = 16;
+		NonDSCBPP1 = 20;
+		NonDSCBPP2 = 24;
+
+		if (Format == dm_n422) {
+			MinDSCBPP = 7;
+			MaxDSCBPP = 2 * DSCInputBitPerComponent - 1.0 / 16.0;
+		} else {
+			MinDSCBPP = 8;
+			MaxDSCBPP = 3 * DSCInputBitPerComponent - 1.0 / 16.0;
+		}
+	}
+
+	if (DSCEnable && Output == dm_dp) {
+		MaxLinkBPP = LinkBitRate / 10 * 8 * Lanes / PixelClock * (1 - 2.4 / 100);
+	} else {
+		MaxLinkBPP = LinkBitRate / 10 * 8 * Lanes / PixelClock;
+	}
+
+	if (ODMCombine == dm_odm_combine_mode_4to1 && MaxLinkBPP > 16) {
+		MaxLinkBPP = 16;
+	} else if (ODMCombine == dm_odm_combine_mode_2to1 && MaxLinkBPP > 32) {
+		MaxLinkBPP = 32;
+	}
+
+	if (DesiredBPP == 0) {
+		if (DSCEnable) {
+			if (MaxLinkBPP < MinDSCBPP) {
+				return BPP_INVALID;
+			} else if (MaxLinkBPP >= MaxDSCBPP) {
+				return MaxDSCBPP;
+			} else {
+				return dml_floor(16.0 * MaxLinkBPP, 1.0) / 16.0;
+			}
+		} else {
+			if (MaxLinkBPP >= NonDSCBPP2) {
+				return NonDSCBPP2;
+			} else if (MaxLinkBPP >= NonDSCBPP1) {
+				return NonDSCBPP1;
+			} else if (MaxLinkBPP >= NonDSCBPP0) {
+				return 16.0;
+			} else {
+				return BPP_INVALID;
+			}
+		}
+	} else {
+		if (!((DSCEnable == false && (DesiredBPP == NonDSCBPP2 || DesiredBPP == NonDSCBPP1 || DesiredBPP <= NonDSCBPP0))
+				|| (DSCEnable && DesiredBPP >= MinDSCBPP && DesiredBPP <= MaxDSCBPP))) {
+			return BPP_INVALID;
+		} else {
+			return DesiredBPP;
+		}
+	}
+}
+
+static noinline void CalculatePrefetchSchedulePerPlane(
+		struct display_mode_lib *mode_lib,
+		double HostVMInefficiencyFactor,
+		int i,
+		unsigned int j,
+		unsigned int k)
+{
+	struct vba_vars_st *v = &mode_lib->vba;
+	Pipe myPipe;
+
+	myPipe.DPPCLK = v->RequiredDPPCLK[i][j][k];
+	myPipe.DISPCLK = v->RequiredDISPCLK[i][j];
+	myPipe.PixelClock = v->PixelClock[k];
+	myPipe.DCFCLKDeepSleep = v->ProjectedDCFCLKDeepSleep[i][j];
+	myPipe.DPPPerPlane = v->NoOfDPP[i][j][k];
+	myPipe.ScalerEnabled = v->ScalerEnabled[k];
+	myPipe.VRatio = mode_lib->vba.VRatio[k];
+	myPipe.VRatioChroma = mode_lib->vba.VRatioChroma[k];
+
+	myPipe.SourceScan = v->SourceScan[k];
+	myPipe.BlockWidth256BytesY = v->Read256BlockWidthY[k];
+	myPipe.BlockHeight256BytesY = v->Read256BlockHeightY[k];
+	myPipe.BlockWidth256BytesC = v->Read256BlockWidthC[k];
+	myPipe.BlockHeight256BytesC = v->Read256BlockHeightC[k];
+	myPipe.InterlaceEnable = v->Interlace[k];
+	myPipe.NumberOfCursors = v->NumberOfCursors[k];
+	myPipe.VBlank = v->VTotal[k] - v->VActive[k];
+	myPipe.HTotal = v->HTotal[k];
+	myPipe.DCCEnable = v->DCCEnable[k];
+	myPipe.ODMCombineIsEnabled = v->ODMCombineEnablePerState[i][k] == dm_odm_combine_mode_4to1
+		|| v->ODMCombineEnablePerState[i][k] == dm_odm_combine_mode_2to1;
+	myPipe.SourcePixelFormat = v->SourcePixelFormat[k];
+	myPipe.BytePerPixelY = v->BytePerPixelY[k];
+	myPipe.BytePerPixelC = v->BytePerPixelC[k];
+	myPipe.ProgressiveToInterlaceUnitInOPP = v->ProgressiveToInterlaceUnitInOPP;
+	v->NoTimeForPrefetch[i][j][k] = CalculatePrefetchSchedule(
+		mode_lib,
+		HostVMInefficiencyFactor,
+		&myPipe,
+		v->DSCDelayPerState[i][k],
+		v->DPPCLKDelaySubtotal + v->DPPCLKDelayCNVCFormater,
+		v->DPPCLKDelaySCL,
+		v->DPPCLKDelaySCLLBOnly,
+		v->DPPCLKDelayCNVCCursor,
+		v->DISPCLKDelaySubtotal,
+		v->SwathWidthYThisState[k] / v->HRatio[k],
+		v->OutputFormat[k],
+		v->MaxInterDCNTileRepeaters,
+		dml_min(v->MaxVStartup, v->MaximumVStartup[i][j][k]),
+		v->MaximumVStartup[i][j][k],
+		v->GPUVMMaxPageTableLevels,
+		v->GPUVMEnable,
+		v->HostVMEnable,
+		v->HostVMMaxNonCachedPageTableLevels,
+		v->HostVMMinPageSize,
+		v->DynamicMetadataEnable[k],
+		v->DynamicMetadataVMEnabled,
+		v->DynamicMetadataLinesBeforeActiveRequired[k],
+		v->DynamicMetadataTransmittedBytes[k],
+		v->UrgLatency[i],
+		v->ExtraLatency,
+		v->TimeCalc,
+		v->PDEAndMetaPTEBytesPerFrame[i][j][k],
+		v->MetaRowBytes[i][j][k],
+		v->DPTEBytesPerRow[i][j][k],
+		v->PrefetchLinesY[i][j][k],
+		v->SwathWidthYThisState[k],
+		v->PrefillY[k],
+		v->MaxNumSwY[k],
+		v->PrefetchLinesC[i][j][k],
+		v->SwathWidthCThisState[k],
+		v->PrefillC[k],
+		v->MaxNumSwC[k],
+		v->swath_width_luma_ub_this_state[k],
+		v->swath_width_chroma_ub_this_state[k],
+		v->SwathHeightYThisState[k],
+		v->SwathHeightCThisState[k],
+		v->TWait,
+		&v->DSTXAfterScaler[k],
+		&v->DSTYAfterScaler[k],
+		&v->LineTimesForPrefetch[k],
+		&v->PrefetchBW[k],
+		&v->LinesForMetaPTE[k],
+		&v->LinesForMetaAndDPTERow[k],
+		&v->VRatioPreY[i][j][k],
+		&v->VRatioPreC[i][j][k],
+		&v->RequiredPrefetchPixelDataBWLuma[i][j][k],
+		&v->RequiredPrefetchPixelDataBWChroma[i][j][k],
+		&v->NoTimeForDynamicMetadata[i][j][k],
+		&v->Tno_bw[k],
+		&v->prefetch_vmrow_bw[k],
+		&v->dummy7[k],
+		&v->dummy8[k],
+		&v->dummy13[k],
+		&v->VUpdateOffsetPix[k],
+		&v->VUpdateWidthPix[k],
+		&v->VReadyOffsetPix[k]);
+}
+
+void dml314_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_lib)
+{
+	struct vba_vars_st *v = &mode_lib->vba;
+
+	int i, j;
+	unsigned int k, m;
+	int ReorderingBytes;
+	int MinPrefetchMode = 0, MaxPrefetchMode = 2;
+	bool NoChroma = true;
+	bool EnoughWritebackUnits = true;
+	bool P2IWith420 = false;
+	bool DSCOnlyIfNecessaryWithBPP = false;
+	bool DSC422NativeNotSupported = false;
+	double MaxTotalVActiveRDBandwidth;
+	bool ViewportExceedsSurface = false;
+	bool FMTBufferExceeded = false;
+
+	/*MODE SUPPORT, VOLTAGE STATE AND SOC CONFIGURATION*/
+
+	CalculateMinAndMaxPrefetchMode(
+		mode_lib->vba.AllowDRAMSelfRefreshOrDRAMClockChangeInVblank,
+		&MinPrefetchMode, &MaxPrefetchMode);
+
+	/*Scale Ratio, taps Support Check*/
+
+	v->ScaleRatioAndTapsSupport = true;
+	for (k = 0; k < v->NumberOfActivePlanes; k++) {
+		if (v->ScalerEnabled[k] == false
+				&& ((v->SourcePixelFormat[k] != dm_444_64 && v->SourcePixelFormat[k] != dm_444_32
+						&& v->SourcePixelFormat[k] != dm_444_16 && v->SourcePixelFormat[k] != dm_mono_16
+						&& v->SourcePixelFormat[k] != dm_mono_8 && v->SourcePixelFormat[k] != dm_rgbe
+						&& v->SourcePixelFormat[k] != dm_rgbe_alpha) || v->HRatio[k] != 1.0 || v->htaps[k] != 1.0
+						|| v->VRatio[k] != 1.0 || v->vtaps[k] != 1.0)) {
+			v->ScaleRatioAndTapsSupport = false;
+		} else if (v->vtaps[k] < 1.0 || v->vtaps[k] > 8.0 || v->htaps[k] < 1.0 || v->htaps[k] > 8.0
+				|| (v->htaps[k] > 1.0 && (v->htaps[k] % 2) == 1) || v->HRatio[k] > v->MaxHSCLRatio
+				|| v->VRatio[k] > v->MaxVSCLRatio || v->HRatio[k] > v->htaps[k]
+				|| v->VRatio[k] > v->vtaps[k]
+				|| (v->SourcePixelFormat[k] != dm_444_64 && v->SourcePixelFormat[k] != dm_444_32
+						&& v->SourcePixelFormat[k] != dm_444_16 && v->SourcePixelFormat[k] != dm_mono_16
+						&& v->SourcePixelFormat[k] != dm_mono_8 && v->SourcePixelFormat[k] != dm_rgbe
+						&& (v->VTAPsChroma[k] < 1 || v->VTAPsChroma[k] > 8 || v->HTAPsChroma[k] < 1
+								|| v->HTAPsChroma[k] > 8 || (v->HTAPsChroma[k] > 1 && v->HTAPsChroma[k] % 2 == 1)
+								|| v->HRatioChroma[k] > v->MaxHSCLRatio
+								|| v->VRatioChroma[k] > v->MaxVSCLRatio
+								|| v->HRatioChroma[k] > v->HTAPsChroma[k]
+								|| v->VRatioChroma[k] > v->VTAPsChroma[k]))) {
+			v->ScaleRatioAndTapsSupport = false;
+		}
+	}
+	/*Source Format, Pixel Format and Scan Support Check*/
+
+	v->SourceFormatPixelAndScanSupport = true;
+	for (k = 0; k < v->NumberOfActivePlanes; k++) {
+		if (v->SurfaceTiling[k] == dm_sw_linear && (!(v->SourceScan[k] != dm_vert) || v->DCCEnable[k] == true)) {
+			v->SourceFormatPixelAndScanSupport = false;
+		}
+	}
+	/*Bandwidth Support Check*/
+
+	for (k = 0; k < v->NumberOfActivePlanes; k++) {
+		CalculateBytePerPixelAnd256BBlockSizes(
+				v->SourcePixelFormat[k],
+				v->SurfaceTiling[k],
+				&v->BytePerPixelY[k],
+				&v->BytePerPixelC[k],
+				&v->BytePerPixelInDETY[k],
+				&v->BytePerPixelInDETC[k],
+				&v->Read256BlockHeightY[k],
+				&v->Read256BlockHeightC[k],
+				&v->Read256BlockWidthY[k],
+				&v->Read256BlockWidthC[k]);
+	}
+	for (k = 0; k < v->NumberOfActivePlanes; k++) {
+		if (v->SourceScan[k] != dm_vert) {
+			v->SwathWidthYSingleDPP[k] = v->ViewportWidth[k];
+			v->SwathWidthCSingleDPP[k] = v->ViewportWidthChroma[k];
+		} else {
+			v->SwathWidthYSingleDPP[k] = v->ViewportHeight[k];
+			v->SwathWidthCSingleDPP[k] = v->ViewportHeightChroma[k];
+		}
+	}
+	for (k = 0; k < v->NumberOfActivePlanes; k++) {
+		v->ReadBandwidthLuma[k] = v->SwathWidthYSingleDPP[k] * dml_ceil(v->BytePerPixelInDETY[k], 1.0)
+				/ (v->HTotal[k] / v->PixelClock[k]) * v->VRatio[k];
+		v->ReadBandwidthChroma[k] = v->SwathWidthYSingleDPP[k] / 2 * dml_ceil(v->BytePerPixelInDETC[k], 2.0)
+				/ (v->HTotal[k] / v->PixelClock[k]) * v->VRatio[k] / 2.0;
+	}
+	for (k = 0; k < v->NumberOfActivePlanes; k++) {
+		if (v->WritebackEnable[k] == true && v->WritebackPixelFormat[k] == dm_444_64) {
+			v->WriteBandwidth[k] = v->WritebackDestinationWidth[k] * v->WritebackDestinationHeight[k]
+					/ (v->WritebackSourceHeight[k] * v->HTotal[k] / v->PixelClock[k]) * 8.0;
+		} else if (v->WritebackEnable[k] == true) {
+			v->WriteBandwidth[k] = v->WritebackDestinationWidth[k] * v->WritebackDestinationHeight[k]
+					/ (v->WritebackSourceHeight[k] * v->HTotal[k] / v->PixelClock[k]) * 4.0;
+		} else {
+			v->WriteBandwidth[k] = 0.0;
+		}
+	}
+
+	/*Writeback Latency support check*/
+
+	v->WritebackLatencySupport = true;
+	for (k = 0; k < v->NumberOfActivePlanes; k++) {
+		if (v->WritebackEnable[k] == true && (v->WriteBandwidth[k] > v->WritebackInterfaceBufferSize * 1024 / v->WritebackLatency)) {
+			v->WritebackLatencySupport = false;
+		}
+	}
+
+	/*Writeback Mode Support Check*/
+
+	v->TotalNumberOfActiveWriteback = 0;
+	for (k = 0; k < v->NumberOfActivePlanes; k++) {
+		if (v->WritebackEnable[k] == true) {
+			v->TotalNumberOfActiveWriteback = v->TotalNumberOfActiveWriteback + 1;
+		}
+	}
+
+	if (v->TotalNumberOfActiveWriteback > v->MaxNumWriteback) {
+		EnoughWritebackUnits = false;
+	}
+
+	/*Writeback Scale Ratio and Taps Support Check*/
+
+	v->WritebackScaleRatioAndTapsSupport = true;
+	for (k = 0; k < v->NumberOfActivePlanes; k++) {
+		if (v->WritebackEnable[k] == true) {
+			if (v->WritebackHRatio[k] > v->WritebackMaxHSCLRatio || v->WritebackVRatio[k] > v->WritebackMaxVSCLRatio
+					|| v->WritebackHRatio[k] < v->WritebackMinHSCLRatio
+					|| v->WritebackVRatio[k] < v->WritebackMinVSCLRatio
+					|| v->WritebackHTaps[k] > v->WritebackMaxHSCLTaps
+					|| v->WritebackVTaps[k] > v->WritebackMaxVSCLTaps
+					|| v->WritebackHRatio[k] > v->WritebackHTaps[k] || v->WritebackVRatio[k] > v->WritebackVTaps[k]
+					|| (v->WritebackHTaps[k] > 2.0 && ((v->WritebackHTaps[k] % 2) == 1))) {
+				v->WritebackScaleRatioAndTapsSupport = false;
+			}
+			if (2.0 * v->WritebackDestinationWidth[k] * (v->WritebackVTaps[k] - 1) * 57 > v->WritebackLineBufferSize) {
+				v->WritebackScaleRatioAndTapsSupport = false;
+			}
+		}
+	}
+	/*Maximum DISPCLK/DPPCLK Support check*/
+
+	v->WritebackRequiredDISPCLK = 0.0;
+	for (k = 0; k < v->NumberOfActivePlanes; k++) {
+		if (v->WritebackEnable[k] == true) {
+			v->WritebackRequiredDISPCLK = dml_max(
+					v->WritebackRequiredDISPCLK,
+					dml314_CalculateWriteBackDISPCLK(
+							v->WritebackPixelFormat[k],
+							v->PixelClock[k],
+							v->WritebackHRatio[k],
+							v->WritebackVRatio[k],
+							v->WritebackHTaps[k],
+							v->WritebackVTaps[k],
+							v->WritebackSourceWidth[k],
+							v->WritebackDestinationWidth[k],
+							v->HTotal[k],
+							v->WritebackLineBufferSize));
+		}
+	}
+	for (k = 0; k < v->NumberOfActivePlanes; k++) {
+		if (v->HRatio[k] > 1.0) {
+			v->PSCL_FACTOR[k] = dml_min(
+					v->MaxDCHUBToPSCLThroughput,
+					v->MaxPSCLToLBThroughput * v->HRatio[k] / dml_ceil(v->htaps[k] / 6.0, 1.0));
+		} else {
+			v->PSCL_FACTOR[k] = dml_min(v->MaxDCHUBToPSCLThroughput, v->MaxPSCLToLBThroughput);
+		}
+		if (v->BytePerPixelC[k] == 0.0) {
+			v->PSCL_FACTOR_CHROMA[k] = 0.0;
+			v->MinDPPCLKUsingSingleDPP[k] = v->PixelClock[k]
+					* dml_max3(
+							v->vtaps[k] / 6.0 * dml_min(1.0, v->HRatio[k]),
+							v->HRatio[k] * v->VRatio[k] / v->PSCL_FACTOR[k],
+							1.0);
+			if ((v->htaps[k] > 6.0 || v->vtaps[k] > 6.0) && v->MinDPPCLKUsingSingleDPP[k] < 2.0 * v->PixelClock[k]) {
+				v->MinDPPCLKUsingSingleDPP[k] = 2.0 * v->PixelClock[k];
+			}
+		} else {
+			if (v->HRatioChroma[k] > 1.0) {
+				v->PSCL_FACTOR_CHROMA[k] = dml_min(
+						v->MaxDCHUBToPSCLThroughput,
+						v->MaxPSCLToLBThroughput * v->HRatioChroma[k] / dml_ceil(v->HTAPsChroma[k] / 6.0, 1.0));
+			} else {
+				v->PSCL_FACTOR_CHROMA[k] = dml_min(v->MaxDCHUBToPSCLThroughput, v->MaxPSCLToLBThroughput);
+			}
+			v->MinDPPCLKUsingSingleDPP[k] = v->PixelClock[k]
+					* dml_max5(
+							v->vtaps[k] / 6.0 * dml_min(1.0, v->HRatio[k]),
+							v->HRatio[k] * v->VRatio[k] / v->PSCL_FACTOR[k],
+							v->VTAPsChroma[k] / 6.0 * dml_min(1.0, v->HRatioChroma[k]),
+							v->HRatioChroma[k] * v->VRatioChroma[k] / v->PSCL_FACTOR_CHROMA[k],
+							1.0);
+			if ((v->htaps[k] > 6.0 || v->vtaps[k] > 6.0 || v->HTAPsChroma[k] > 6.0 || v->VTAPsChroma[k] > 6.0)
+					&& v->MinDPPCLKUsingSingleDPP[k] < 2.0 * v->PixelClock[k]) {
+				v->MinDPPCLKUsingSingleDPP[k] = 2.0 * v->PixelClock[k];
+			}
+		}
+	}
+	for (k = 0; k < v->NumberOfActivePlanes; k++) {
+		int MaximumSwathWidthSupportLuma;
+		int MaximumSwathWidthSupportChroma;
+
+		if (v->SurfaceTiling[k] == dm_sw_linear) {
+			MaximumSwathWidthSupportLuma = 8192.0;
+		} else if (v->SourceScan[k] == dm_vert && v->BytePerPixelC[k] > 0) {
+			MaximumSwathWidthSupportLuma = 2880.0;
+		} else if (v->SourcePixelFormat[k] == dm_rgbe_alpha) {
+			MaximumSwathWidthSupportLuma = 3840.0;
+		} else {
+			MaximumSwathWidthSupportLuma = 5760.0;
+		}
+
+		if (v->SourcePixelFormat[k] == dm_420_8 || v->SourcePixelFormat[k] == dm_420_10 || v->SourcePixelFormat[k] == dm_420_12) {
+			MaximumSwathWidthSupportChroma = MaximumSwathWidthSupportLuma / 2.0;
+		} else {
+			MaximumSwathWidthSupportChroma = MaximumSwathWidthSupportLuma;
+		}
+		v->MaximumSwathWidthInLineBufferLuma = v->LineBufferSize * dml_max(v->HRatio[k], 1.0) / v->LBBitPerPixel[k]
+				/ (v->vtaps[k] + dml_max(dml_ceil(v->VRatio[k], 1.0) - 2, 0.0));
+		if (v->BytePerPixelC[k] == 0.0) {
+			v->MaximumSwathWidthInLineBufferChroma = 0;
+		} else {
+			v->MaximumSwathWidthInLineBufferChroma = v->LineBufferSize * dml_max(v->HRatioChroma[k], 1.0) / v->LBBitPerPixel[k]
+					/ (v->VTAPsChroma[k] + dml_max(dml_ceil(v->VRatioChroma[k], 1.0) - 2, 0.0));
+		}
+		v->MaximumSwathWidthLuma[k] = dml_min(MaximumSwathWidthSupportLuma, v->MaximumSwathWidthInLineBufferLuma);
+		v->MaximumSwathWidthChroma[k] = dml_min(MaximumSwathWidthSupportChroma, v->MaximumSwathWidthInLineBufferChroma);
+	}
+
+	CalculateSwathAndDETConfiguration(
+			true,
+			v->NumberOfActivePlanes,
+			v->DETBufferSizeInKByte[0],
+			v->MaximumSwathWidthLuma,
+			v->MaximumSwathWidthChroma,
+			v->SourceScan,
+			v->SourcePixelFormat,
+			v->SurfaceTiling,
+			v->ViewportWidth,
+			v->ViewportHeight,
+			v->SurfaceWidthY,
+			v->SurfaceWidthC,
+			v->SurfaceHeightY,
+			v->SurfaceHeightC,
+			v->Read256BlockHeightY,
+			v->Read256BlockHeightC,
+			v->Read256BlockWidthY,
+			v->Read256BlockWidthC,
+			v->odm_combine_dummy,
+			v->BlendingAndTiming,
+			v->BytePerPixelY,
+			v->BytePerPixelC,
+			v->BytePerPixelInDETY,
+			v->BytePerPixelInDETC,
+			v->HActive,
+			v->HRatio,
+			v->HRatioChroma,
+			v->NoOfDPPThisState,
+			v->swath_width_luma_ub_this_state,
+			v->swath_width_chroma_ub_this_state,
+			v->SwathWidthYThisState,
+			v->SwathWidthCThisState,
+			v->SwathHeightYThisState,
+			v->SwathHeightCThisState,
+			v->DETBufferSizeYThisState,
+			v->DETBufferSizeCThisState,
+			v->SingleDPPViewportSizeSupportPerPlane,
+			&v->ViewportSizeSupport[0][0]);
+
+	for (i = 0; i < v->soc.num_states; i++) {
+		for (j = 0; j < 2; j++) {
+			v->MaxDispclkRoundedDownToDFSGranularity = RoundToDFSGranularityDown(v->MaxDispclk[i], v->DISPCLKDPPCLKVCOSpeed);
+			v->MaxDppclkRoundedDownToDFSGranularity = RoundToDFSGranularityDown(v->MaxDppclk[i], v->DISPCLKDPPCLKVCOSpeed);
+			v->RequiredDISPCLK[i][j] = 0.0;
+			v->DISPCLK_DPPCLK_Support[i][j] = true;
+			for (k = 0; k < v->NumberOfActivePlanes; k++) {
+				v->PlaneRequiredDISPCLKWithoutODMCombine = v->PixelClock[k] * (1.0 + v->DISPCLKDPPCLKDSCCLKDownSpreading / 100.0)
+						* (1.0 + v->DISPCLKRampingMargin / 100.0);
+				if ((v->PlaneRequiredDISPCLKWithoutODMCombine >= v->MaxDispclk[i]
+						&& v->MaxDispclk[i] == v->MaxDispclk[v->soc.num_states - 1]
+						&& v->MaxDppclk[i] == v->MaxDppclk[v->soc.num_states - 1])) {
+					v->PlaneRequiredDISPCLKWithoutODMCombine = v->PixelClock[k]
+							* (1 + v->DISPCLKDPPCLKDSCCLKDownSpreading / 100.0);
+				}
+				v->PlaneRequiredDISPCLKWithODMCombine2To1 = v->PixelClock[k] / 2 * (1 + v->DISPCLKDPPCLKDSCCLKDownSpreading / 100.0)
+						* (1 + v->DISPCLKRampingMargin / 100.0);
+				if ((v->PlaneRequiredDISPCLKWithODMCombine2To1 >= v->MaxDispclk[i]
+						&& v->MaxDispclk[i] == v->MaxDispclk[v->soc.num_states - 1]
+						&& v->MaxDppclk[i] == v->MaxDppclk[v->soc.num_states - 1])) {
+					v->PlaneRequiredDISPCLKWithODMCombine2To1 = v->PixelClock[k] / 2
+							* (1 + v->DISPCLKDPPCLKDSCCLKDownSpreading / 100.0);
+				}
+				v->PlaneRequiredDISPCLKWithODMCombine4To1 = v->PixelClock[k] / 4 * (1 + v->DISPCLKDPPCLKDSCCLKDownSpreading / 100.0)
+						* (1 + v->DISPCLKRampingMargin / 100.0);
+				if ((v->PlaneRequiredDISPCLKWithODMCombine4To1 >= v->MaxDispclk[i]
+						&& v->MaxDispclk[i] == v->MaxDispclk[v->soc.num_states - 1]
+						&& v->MaxDppclk[i] == v->MaxDppclk[v->soc.num_states - 1])) {
+					v->PlaneRequiredDISPCLKWithODMCombine4To1 = v->PixelClock[k] / 4
+							* (1 + v->DISPCLKDPPCLKDSCCLKDownSpreading / 100.0);
+				}
+
+				if (v->ODMCombinePolicy == dm_odm_combine_policy_none
+						|| !(v->Output[k] == dm_dp ||
+						     v->Output[k] == dm_dp2p0 ||
+						     v->Output[k] == dm_edp)) {
+					v->ODMCombineEnablePerState[i][k] = dm_odm_combine_mode_disabled;
+					v->PlaneRequiredDISPCLK = v->PlaneRequiredDISPCLKWithoutODMCombine;
+
+					if (v->HActive[k] / 2 > DCN314_MAX_FMT_420_BUFFER_WIDTH)
+						FMTBufferExceeded = true;
+				} else if (v->ODMCombinePolicy == dm_odm_combine_policy_2to1) {
+					v->ODMCombineEnablePerState[i][k] = dm_odm_combine_mode_2to1;
+					v->PlaneRequiredDISPCLK = v->PlaneRequiredDISPCLKWithODMCombine2To1;
+				} else if (v->ODMCombinePolicy == dm_odm_combine_policy_4to1
+						|| v->PlaneRequiredDISPCLKWithODMCombine2To1 > v->MaxDispclkRoundedDownToDFSGranularity) {
+					v->ODMCombineEnablePerState[i][k] = dm_odm_combine_mode_4to1;
+					v->PlaneRequiredDISPCLK = v->PlaneRequiredDISPCLKWithODMCombine4To1;
+				} else if (v->PlaneRequiredDISPCLKWithoutODMCombine > v->MaxDispclkRoundedDownToDFSGranularity) {
+					v->ODMCombineEnablePerState[i][k] = dm_odm_combine_mode_2to1;
+					v->PlaneRequiredDISPCLK = v->PlaneRequiredDISPCLKWithODMCombine2To1;
+				} else {
+					v->ODMCombineEnablePerState[i][k] = dm_odm_combine_mode_disabled;
+					v->PlaneRequiredDISPCLK = v->PlaneRequiredDISPCLKWithoutODMCombine;
+				}
+				if (v->DSCEnabled[k] && v->HActive[k] > DCN314_MAX_DSC_IMAGE_WIDTH
+						&& v->ODMCombineEnablePerState[i][k] != dm_odm_combine_mode_4to1) {
+					if (v->HActive[k] / 2 > DCN314_MAX_DSC_IMAGE_WIDTH) {
+						v->ODMCombineEnablePerState[i][k] = dm_odm_combine_mode_4to1;
+						v->PlaneRequiredDISPCLK = v->PlaneRequiredDISPCLKWithODMCombine4To1;
+					} else {
+						v->ODMCombineEnablePerState[i][k] = dm_odm_combine_mode_2to1;
+						v->PlaneRequiredDISPCLK = v->PlaneRequiredDISPCLKWithODMCombine2To1;
+					}
+				}
+				if (v->OutputFormat[k] == dm_420 && v->HActive[k] > DCN314_MAX_FMT_420_BUFFER_WIDTH
+						&& v->ODMCombineEnablePerState[i][k] != dm_odm_combine_mode_4to1) {
+					if (v->Output[k] == dm_hdmi) {
+						FMTBufferExceeded = true;
+					} else if (v->HActive[k] / 2 > DCN314_MAX_FMT_420_BUFFER_WIDTH) {
+						v->ODMCombineEnablePerState[i][k] = dm_odm_combine_mode_4to1;
+						v->PlaneRequiredDISPCLK = v->PlaneRequiredDISPCLKWithODMCombine4To1;
+
+						if (v->HActive[k] / 4 > DCN314_MAX_FMT_420_BUFFER_WIDTH)
+							FMTBufferExceeded = true;
+					} else {
+						v->ODMCombineEnablePerState[i][k] = dm_odm_combine_mode_2to1;
+						v->PlaneRequiredDISPCLK = v->PlaneRequiredDISPCLKWithODMCombine2To1;
+					}
+				}
+				if (v->ODMCombineEnablePerState[i][k] == dm_odm_combine_mode_4to1) {
+					v->MPCCombine[i][j][k] = false;
+					v->NoOfDPP[i][j][k] = 4;
+					v->RequiredDPPCLK[i][j][k] = v->MinDPPCLKUsingSingleDPP[k] * (1 + v->DISPCLKDPPCLKDSCCLKDownSpreading / 100.0) / 4;
+				} else if (v->ODMCombineEnablePerState[i][k] == dm_odm_combine_mode_2to1) {
+					v->MPCCombine[i][j][k] = false;
+					v->NoOfDPP[i][j][k] = 2;
+					v->RequiredDPPCLK[i][j][k] = v->MinDPPCLKUsingSingleDPP[k] * (1 + v->DISPCLKDPPCLKDSCCLKDownSpreading / 100.0) / 2;
+				} else if ((v->WhenToDoMPCCombine == dm_mpc_never
+						|| (v->MinDPPCLKUsingSingleDPP[k] * (1 + v->DISPCLKDPPCLKDSCCLKDownSpreading / 100.0)
+								<= v->MaxDppclkRoundedDownToDFSGranularity && v->SingleDPPViewportSizeSupportPerPlane[k] == true))) {
+					v->MPCCombine[i][j][k] = false;
+					v->NoOfDPP[i][j][k] = 1;
+					v->RequiredDPPCLK[i][j][k] = v->MinDPPCLKUsingSingleDPP[k] * (1.0 + v->DISPCLKDPPCLKDSCCLKDownSpreading / 100.0);
+				} else {
+					v->MPCCombine[i][j][k] = true;
+					v->NoOfDPP[i][j][k] = 2;
+					v->RequiredDPPCLK[i][j][k] = v->MinDPPCLKUsingSingleDPP[k] * (1.0 + v->DISPCLKDPPCLKDSCCLKDownSpreading / 100.0) / 2.0;
+				}
+				v->RequiredDISPCLK[i][j] = dml_max(v->RequiredDISPCLK[i][j], v->PlaneRequiredDISPCLK);
+				if ((v->MinDPPCLKUsingSingleDPP[k] / v->NoOfDPP[i][j][k] * (1.0 + v->DISPCLKDPPCLKDSCCLKDownSpreading / 100.0)
+						> v->MaxDppclkRoundedDownToDFSGranularity)
+						|| (v->PlaneRequiredDISPCLK > v->MaxDispclkRoundedDownToDFSGranularity)) {
+					v->DISPCLK_DPPCLK_Support[i][j] = false;
+				}
+			}
+			v->TotalNumberOfActiveDPP[i][j] = 0;
+			v->TotalNumberOfSingleDPPPlanes[i][j] = 0;
+			for (k = 0; k < v->NumberOfActivePlanes; k++) {
+				v->TotalNumberOfActiveDPP[i][j] = v->TotalNumberOfActiveDPP[i][j] + v->NoOfDPP[i][j][k];
+				if (v->NoOfDPP[i][j][k] == 1)
+					v->TotalNumberOfSingleDPPPlanes[i][j] = v->TotalNumberOfSingleDPPPlanes[i][j] + 1;
+				if (v->SourcePixelFormat[k] == dm_420_8 || v->SourcePixelFormat[k] == dm_420_10
+						|| v->SourcePixelFormat[k] == dm_420_12 || v->SourcePixelFormat[k] == dm_rgbe_alpha)
+					NoChroma = false;
+			}
+
+			// UPTO
+			if (j == 1 && v->WhenToDoMPCCombine != dm_mpc_never
+					&& !UnboundedRequest(v->UseUnboundedRequesting, v->TotalNumberOfActiveDPP[i][j], NoChroma, v->Output[0])) {
+				while (!(v->TotalNumberOfActiveDPP[i][j] >= v->MaxNumDPP || v->TotalNumberOfSingleDPPPlanes[i][j] == 0)) {
+					double BWOfNonSplitPlaneOfMaximumBandwidth;
+					unsigned int NumberOfNonSplitPlaneOfMaximumBandwidth;
+
+					BWOfNonSplitPlaneOfMaximumBandwidth = 0;
+					NumberOfNonSplitPlaneOfMaximumBandwidth = 0;
+					for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+						if (v->ReadBandwidthLuma[k] + v->ReadBandwidthChroma[k] > BWOfNonSplitPlaneOfMaximumBandwidth
+								&& v->ODMCombineEnablePerState[i][k] == dm_odm_combine_mode_disabled && v->MPCCombine[i][j][k] == false) {
+							BWOfNonSplitPlaneOfMaximumBandwidth = v->ReadBandwidthLuma[k] + v->ReadBandwidthChroma[k];
+							NumberOfNonSplitPlaneOfMaximumBandwidth = k;
+						}
+					}
+					v->MPCCombine[i][j][NumberOfNonSplitPlaneOfMaximumBandwidth] = true;
+					v->NoOfDPP[i][j][NumberOfNonSplitPlaneOfMaximumBandwidth] = 2;
+					v->RequiredDPPCLK[i][j][NumberOfNonSplitPlaneOfMaximumBandwidth] =
+							v->MinDPPCLKUsingSingleDPP[NumberOfNonSplitPlaneOfMaximumBandwidth]
+									* (1 + v->DISPCLKDPPCLKDSCCLKDownSpreading / 100) / 2;
+					v->TotalNumberOfActiveDPP[i][j] = v->TotalNumberOfActiveDPP[i][j] + 1;
+					v->TotalNumberOfSingleDPPPlanes[i][j] = v->TotalNumberOfSingleDPPPlanes[i][j] - 1;
+				}
+			}
+			if (v->TotalNumberOfActiveDPP[i][j] > v->MaxNumDPP) {
+				v->RequiredDISPCLK[i][j] = 0.0;
+				v->DISPCLK_DPPCLK_Support[i][j] = true;
+				for (k = 0; k < v->NumberOfActivePlanes; k++) {
+					v->ODMCombineEnablePerState[i][k] = dm_odm_combine_mode_disabled;
+					if (v->SingleDPPViewportSizeSupportPerPlane[k] == false && v->WhenToDoMPCCombine != dm_mpc_never) {
+						v->MPCCombine[i][j][k] = true;
+						v->NoOfDPP[i][j][k] = 2;
+						v->RequiredDPPCLK[i][j][k] = v->MinDPPCLKUsingSingleDPP[k]
+								* (1.0 + v->DISPCLKDPPCLKDSCCLKDownSpreading / 100.0) / 2.0;
+					} else {
+						v->MPCCombine[i][j][k] = false;
+						v->NoOfDPP[i][j][k] = 1;
+						v->RequiredDPPCLK[i][j][k] = v->MinDPPCLKUsingSingleDPP[k]
+								* (1.0 + v->DISPCLKDPPCLKDSCCLKDownSpreading / 100.0);
+					}
+					if (!(v->MaxDispclk[i] == v->MaxDispclk[v->soc.num_states - 1]
+							&& v->MaxDppclk[i] == v->MaxDppclk[v->soc.num_states - 1])) {
+						v->PlaneRequiredDISPCLK = v->PixelClock[k] * (1.0 + v->DISPCLKDPPCLKDSCCLKDownSpreading / 100.0)
+								* (1.0 + v->DISPCLKRampingMargin / 100.0);
+					} else {
+						v->PlaneRequiredDISPCLK = v->PixelClock[k] * (1.0 + v->DISPCLKDPPCLKDSCCLKDownSpreading / 100.0);
+					}
+					v->RequiredDISPCLK[i][j] = dml_max(v->RequiredDISPCLK[i][j], v->PlaneRequiredDISPCLK);
+					if ((v->MinDPPCLKUsingSingleDPP[k] / v->NoOfDPP[i][j][k] * (1.0 + v->DISPCLKDPPCLKDSCCLKDownSpreading / 100.0)
+							> v->MaxDppclkRoundedDownToDFSGranularity)
+							|| (v->PlaneRequiredDISPCLK > v->MaxDispclkRoundedDownToDFSGranularity)) {
+						v->DISPCLK_DPPCLK_Support[i][j] = false;
+					}
+				}
+				v->TotalNumberOfActiveDPP[i][j] = 0.0;
+				for (k = 0; k < v->NumberOfActivePlanes; k++) {
+					v->TotalNumberOfActiveDPP[i][j] = v->TotalNumberOfActiveDPP[i][j] + v->NoOfDPP[i][j][k];
+				}
+			}
+			v->RequiredDISPCLK[i][j] = dml_max(v->RequiredDISPCLK[i][j], v->WritebackRequiredDISPCLK);
+			if (v->MaxDispclkRoundedDownToDFSGranularity < v->WritebackRequiredDISPCLK) {
+				v->DISPCLK_DPPCLK_Support[i][j] = false;
+			}
+		}
+	}
+
+	/*Total Available Pipes Support Check*/
+
+	for (i = 0; i < v->soc.num_states; i++) {
+		for (j = 0; j < 2; j++) {
+			if (v->TotalNumberOfActiveDPP[i][j] <= v->MaxNumDPP) {
+				v->TotalAvailablePipesSupport[i][j] = true;
+			} else {
+				v->TotalAvailablePipesSupport[i][j] = false;
+			}
+		}
+	}
+	/*Display IO and DSC Support Check*/
+
+	v->NonsupportedDSCInputBPC = false;
+	for (k = 0; k < v->NumberOfActivePlanes; k++) {
+		if (!(v->DSCInputBitPerComponent[k] == 12.0 || v->DSCInputBitPerComponent[k] == 10.0 || v->DSCInputBitPerComponent[k] == 8.0)
+				|| v->DSCInputBitPerComponent[k] > v->MaximumDSCBitsPerComponent) {
+			v->NonsupportedDSCInputBPC = true;
+		}
+	}
+
+	/*Number Of DSC Slices*/
+	for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+		if (v->BlendingAndTiming[k] == k) {
+			if (v->PixelClockBackEnd[k] > 3200) {
+				v->NumberOfDSCSlices[k] = dml_ceil(v->PixelClockBackEnd[k] / 400.0, 4.0);
+			} else if (v->PixelClockBackEnd[k] > 1360) {
+				v->NumberOfDSCSlices[k] = 8;
+			} else if (v->PixelClockBackEnd[k] > 680) {
+				v->NumberOfDSCSlices[k] = 4;
+			} else if (v->PixelClockBackEnd[k] > 340) {
+				v->NumberOfDSCSlices[k] = 2;
+			} else {
+				v->NumberOfDSCSlices[k] = 1;
+			}
+		} else {
+			v->NumberOfDSCSlices[k] = 0;
+		}
+	}
+
+	for (i = 0; i < v->soc.num_states; i++) {
+		for (k = 0; k < v->NumberOfActivePlanes; k++) {
+			v->RequiresDSC[i][k] = false;
+			v->RequiresFEC[i][k] = false;
+			if (v->BlendingAndTiming[k] == k) {
+				if (v->Output[k] == dm_hdmi) {
+					v->RequiresDSC[i][k] = false;
+					v->RequiresFEC[i][k] = false;
+					v->OutputBppPerState[i][k] = TruncToValidBPP(
+							dml_min(600.0, v->PHYCLKPerState[i]) * 10,
+							3,
+							v->HTotal[k],
+							v->HActive[k],
+							v->PixelClockBackEnd[k],
+							v->ForcedOutputLinkBPP[k],
+							false,
+							v->Output[k],
+							v->OutputFormat[k],
+							v->DSCInputBitPerComponent[k],
+							v->NumberOfDSCSlices[k],
+							v->AudioSampleRate[k],
+							v->AudioSampleLayout[k],
+							v->ODMCombineEnablePerState[i][k]);
+				} else if (v->Output[k] == dm_dp || v->Output[k] == dm_edp || v->Output[k] == dm_dp2p0) {
+					if (v->DSCEnable[k] == true) {
+						v->RequiresDSC[i][k] = true;
+						v->LinkDSCEnable = true;
+						if (v->Output[k] == dm_dp || v->Output[k] == dm_dp2p0) {
+							v->RequiresFEC[i][k] = true;
+						} else {
+							v->RequiresFEC[i][k] = false;
+						}
+					} else {
+						v->RequiresDSC[i][k] = false;
+						v->LinkDSCEnable = false;
+						if (v->Output[k] == dm_dp2p0) {
+							v->RequiresFEC[i][k] = true;
+						} else {
+							v->RequiresFEC[i][k] = false;
+						}
+					}
+					if (v->Output[k] == dm_dp2p0) {
+						v->Outbpp = BPP_INVALID;
+						if ((v->OutputLinkDPRate[k] == dm_dp_rate_na || v->OutputLinkDPRate[k] == dm_dp_rate_uhbr10) &&
+							v->PHYCLKD18PerState[k] >= 10000.0 / 18.0) {
+							v->Outbpp = TruncToValidBPP(
+									(1.0 - v->Downspreading / 100.0) * 10000,
+									v->OutputLinkDPLanes[k],
+									v->HTotal[k],
+									v->HActive[k],
+									v->PixelClockBackEnd[k],
+									v->ForcedOutputLinkBPP[k],
+									v->LinkDSCEnable,
+									v->Output[k],
+									v->OutputFormat[k],
+									v->DSCInputBitPerComponent[k],
+									v->NumberOfDSCSlices[k],
+									v->AudioSampleRate[k],
+									v->AudioSampleLayout[k],
+									v->ODMCombineEnablePerState[i][k]);
+							if (v->Outbpp == BPP_INVALID && v->PHYCLKD18PerState[k] < 13500.0 / 18.0 &&
+								v->DSCEnable[k] == true && v->ForcedOutputLinkBPP[k] == 0) {
+								v->RequiresDSC[i][k] = true;
+								v->LinkDSCEnable = true;
+								v->Outbpp = TruncToValidBPP(
+										(1.0 - v->Downspreading / 100.0) * 10000,
+										v->OutputLinkDPLanes[k],
+										v->HTotal[k],
+										v->HActive[k],
+										v->PixelClockBackEnd[k],
+										v->ForcedOutputLinkBPP[k],
+										v->LinkDSCEnable,
+										v->Output[k],
+										v->OutputFormat[k],
+										v->DSCInputBitPerComponent[k],
+										v->NumberOfDSCSlices[k],
+										v->AudioSampleRate[k],
+										v->AudioSampleLayout[k],
+										v->ODMCombineEnablePerState[i][k]);
+							}
+							v->OutputBppPerState[i][k] = v->Outbpp;
+							// TODO: Need some other way to handle this nonsense
+							// v->OutputTypeAndRatePerState[i][k] = v->Output[k] & " UHBR10"
+						}
+						if (v->Outbpp == BPP_INVALID &&
+							(v->OutputLinkDPRate[k] == dm_dp_rate_na || v->OutputLinkDPRate[k] == dm_dp_rate_uhbr13p5) &&
+							v->PHYCLKD18PerState[k] >= 13500.0 / 18.0) {
+							v->Outbpp = TruncToValidBPP(
+									(1.0 - v->Downspreading / 100.0) * 13500,
+									v->OutputLinkDPLanes[k],
+									v->HTotal[k],
+									v->HActive[k],
+									v->PixelClockBackEnd[k],
+									v->ForcedOutputLinkBPP[k],
+									v->LinkDSCEnable,
+									v->Output[k],
+									v->OutputFormat[k],
+									v->DSCInputBitPerComponent[k],
+									v->NumberOfDSCSlices[k],
+									v->AudioSampleRate[k],
+									v->AudioSampleLayout[k],
+									v->ODMCombineEnablePerState[i][k]);
+							if (v->Outbpp == BPP_INVALID && v->PHYCLKD18PerState[k] < 20000.0 / 18.0 &&
+								v->DSCEnable[k] == true && v->ForcedOutputLinkBPP[k] == 0) {
+								v->RequiresDSC[i][k] = true;
+								v->LinkDSCEnable = true;
+								v->Outbpp = TruncToValidBPP(
+										(1.0 - v->Downspreading / 100.0) * 13500,
+										v->OutputLinkDPLanes[k],
+										v->HTotal[k],
+										v->HActive[k],
+										v->PixelClockBackEnd[k],
+										v->ForcedOutputLinkBPP[k],
+										v->LinkDSCEnable,
+										v->Output[k],
+										v->OutputFormat[k],
+										v->DSCInputBitPerComponent[k],
+										v->NumberOfDSCSlices[k],
+										v->AudioSampleRate[k],
+										v->AudioSampleLayout[k],
+										v->ODMCombineEnablePerState[i][k]);
+							}
+							v->OutputBppPerState[i][k] = v->Outbpp;
+							// TODO: Need some other way to handle this nonsense
+							// v->OutputTypeAndRatePerState[i][k] = v->Output[k] & " UHBR13p5"
+						}
+						if (v->Outbpp == BPP_INVALID &&
+							(v->OutputLinkDPRate[k] == dm_dp_rate_na || v->OutputLinkDPRate[k] == dm_dp_rate_uhbr20) &&
+							v->PHYCLKD18PerState[k] >= 20000.0 / 18.0) {
+							v->Outbpp = TruncToValidBPP(
+									(1.0 - v->Downspreading / 100.0) * 20000,
+									v->OutputLinkDPLanes[k],
+									v->HTotal[k],
+									v->HActive[k],
+									v->PixelClockBackEnd[k],
+									v->ForcedOutputLinkBPP[k],
+									v->LinkDSCEnable,
+									v->Output[k],
+									v->OutputFormat[k],
+									v->DSCInputBitPerComponent[k],
+									v->NumberOfDSCSlices[k],
+									v->AudioSampleRate[k],
+									v->AudioSampleLayout[k],
+									v->ODMCombineEnablePerState[i][k]);
+							if (v->Outbpp == BPP_INVALID && v->DSCEnable[k] == true &&
+								v->ForcedOutputLinkBPP[k] == 0) {
+								v->RequiresDSC[i][k] = true;
+								v->LinkDSCEnable = true;
+								v->Outbpp = TruncToValidBPP(
+										(1.0 - v->Downspreading / 100.0) * 20000,
+										v->OutputLinkDPLanes[k],
+										v->HTotal[k],
+										v->HActive[k],
+										v->PixelClockBackEnd[k],
+										v->ForcedOutputLinkBPP[k],
+										v->LinkDSCEnable,
+										v->Output[k],
+										v->OutputFormat[k],
+										v->DSCInputBitPerComponent[k],
+										v->NumberOfDSCSlices[k],
+										v->AudioSampleRate[k],
+										v->AudioSampleLayout[k],
+										v->ODMCombineEnablePerState[i][k]);
+							}
+							v->OutputBppPerState[i][k] = v->Outbpp;
+							// TODO: Need some other way to handle this nonsense
+							// v->OutputTypeAndRatePerState[i][k] = v->Output[k] & " UHBR20"
+						}
+					} else {
+						v->Outbpp = BPP_INVALID;
+						if (v->PHYCLKPerState[i] >= 270.0) {
+							v->Outbpp = TruncToValidBPP(
+									(1.0 - v->Downspreading / 100.0) * 2700,
+									v->OutputLinkDPLanes[k],
+									v->HTotal[k],
+									v->HActive[k],
+									v->PixelClockBackEnd[k],
+									v->ForcedOutputLinkBPP[k],
+									v->LinkDSCEnable,
+									v->Output[k],
+									v->OutputFormat[k],
+									v->DSCInputBitPerComponent[k],
+									v->NumberOfDSCSlices[k],
+									v->AudioSampleRate[k],
+									v->AudioSampleLayout[k],
+									v->ODMCombineEnablePerState[i][k]);
+							v->OutputBppPerState[i][k] = v->Outbpp;
+							// TODO: Need some other way to handle this nonsense
+							// v->OutputTypeAndRatePerState[i][k] = v->Output[k] & " HBR"
+						}
+						if (v->Outbpp == BPP_INVALID && v->PHYCLKPerState[i] >= 540.0) {
+							v->Outbpp = TruncToValidBPP(
+									(1.0 - v->Downspreading / 100.0) * 5400,
+									v->OutputLinkDPLanes[k],
+									v->HTotal[k],
+									v->HActive[k],
+									v->PixelClockBackEnd[k],
+									v->ForcedOutputLinkBPP[k],
+									v->LinkDSCEnable,
+									v->Output[k],
+									v->OutputFormat[k],
+									v->DSCInputBitPerComponent[k],
+									v->NumberOfDSCSlices[k],
+									v->AudioSampleRate[k],
+									v->AudioSampleLayout[k],
+									v->ODMCombineEnablePerState[i][k]);
+							v->OutputBppPerState[i][k] = v->Outbpp;
+							// TODO: Need some other way to handle this nonsense
+							// v->OutputTypeAndRatePerState[i][k] = v->Output[k] & " HBR2"
+						}
+						if (v->Outbpp == BPP_INVALID && v->PHYCLKPerState[i] >= 810.0) {
+							v->Outbpp = TruncToValidBPP(
+									(1.0 - v->Downspreading / 100.0) * 8100,
+									v->OutputLinkDPLanes[k],
+									v->HTotal[k],
+									v->HActive[k],
+									v->PixelClockBackEnd[k],
+									v->ForcedOutputLinkBPP[k],
+									v->LinkDSCEnable,
+									v->Output[k],
+									v->OutputFormat[k],
+									v->DSCInputBitPerComponent[k],
+									v->NumberOfDSCSlices[k],
+									v->AudioSampleRate[k],
+									v->AudioSampleLayout[k],
+									v->ODMCombineEnablePerState[i][k]);
+							v->OutputBppPerState[i][k] = v->Outbpp;
+							// TODO: Need some other way to handle this nonsense
+							// v->OutputTypeAndRatePerState[i][k] = v->Output[k] & " HBR3"
+						}
+					}
+				}
+			} else {
+				v->OutputBppPerState[i][k] = 0;
+			}
+		}
+	}
+
+	for (i = 0; i < v->soc.num_states; i++) {
+		v->LinkCapacitySupport[i] = true;
+		for (k = 0; k < v->NumberOfActivePlanes; k++) {
+			if (v->BlendingAndTiming[k] == k
+					&& (v->Output[k] == dm_dp ||
+					    v->Output[k] == dm_edp ||
+					    v->Output[k] == dm_hdmi) && v->OutputBppPerState[i][k] == 0) {
+				v->LinkCapacitySupport[i] = false;
+			}
+		}
+	}
+
+	// UPTO 2172
+	for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+		if (v->BlendingAndTiming[k] == k
+				&& (v->Output[k] == dm_dp ||
+				    v->Output[k] == dm_edp ||
+				    v->Output[k] == dm_hdmi)) {
+			if (v->OutputFormat[k] == dm_420 && v->Interlace[k] == 1 && v->ProgressiveToInterlaceUnitInOPP == true) {
+				P2IWith420 = true;
+			}
+			if (v->DSCEnable[k] == true && v->OutputFormat[k] == dm_n422
+					&& !v->DSC422NativeSupport) {
+				DSC422NativeNotSupported = true;
+			}
+		}
+	}
+
+
+	for (i = 0; i < v->soc.num_states; ++i) {
+		v->ODMCombine4To1SupportCheckOK[i] = true;
+		for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+			if (v->BlendingAndTiming[k] == k && v->ODMCombineEnablePerState[i][k] == dm_odm_combine_mode_4to1
+					&& (v->ODMCombine4To1Supported == false || v->Output[k] == dm_dp || v->Output[k] == dm_edp
+							|| v->Output[k] == dm_hdmi)) {
+				v->ODMCombine4To1SupportCheckOK[i] = false;
+			}
+		}
+	}
+
+	/* Skip dscclk validation: as long as dispclk is supported, dscclk is also implicitly supported */
+
+	for (i = 0; i < v->soc.num_states; i++) {
+		v->NotEnoughDSCUnits[i] = false;
+		v->TotalDSCUnitsRequired = 0.0;
+		for (k = 0; k < v->NumberOfActivePlanes; k++) {
+			if (v->RequiresDSC[i][k] == true) {
+				if (v->ODMCombineEnablePerState[i][k] == dm_odm_combine_mode_4to1) {
+					v->TotalDSCUnitsRequired = v->TotalDSCUnitsRequired + 4.0;
+				} else if (v->ODMCombineEnablePerState[i][k] == dm_odm_combine_mode_2to1) {
+					v->TotalDSCUnitsRequired = v->TotalDSCUnitsRequired + 2.0;
+				} else {
+					v->TotalDSCUnitsRequired = v->TotalDSCUnitsRequired + 1.0;
+				}
+			}
+		}
+		if (v->TotalDSCUnitsRequired > v->NumberOfDSC) {
+			v->NotEnoughDSCUnits[i] = true;
+		}
+	}
+	/*DSC Delay per state*/
+
+	for (i = 0; i < v->soc.num_states; i++) {
+		for (k = 0; k < v->NumberOfActivePlanes; k++) {
+			if (v->OutputBppPerState[i][k] == BPP_INVALID) {
+				v->BPP = 0.0;
+			} else {
+				v->BPP = v->OutputBppPerState[i][k];
+			}
+			if (v->RequiresDSC[i][k] == true && v->BPP != 0.0) {
+				if (v->ODMCombineEnablePerState[i][k] == dm_odm_combine_mode_disabled) {
+					v->DSCDelayPerState[i][k] = dscceComputeDelay(
+							v->DSCInputBitPerComponent[k],
+							v->BPP,
+							dml_ceil(1.0 * v->HActive[k] / v->NumberOfDSCSlices[k], 1.0),
+							v->NumberOfDSCSlices[k],
+							v->OutputFormat[k],
+							v->Output[k]) + dscComputeDelay(v->OutputFormat[k], v->Output[k]);
+				} else if (v->ODMCombineEnablePerState[i][k] == dm_odm_combine_mode_2to1) {
+					v->DSCDelayPerState[i][k] = 2.0
+							* (dscceComputeDelay(
+									v->DSCInputBitPerComponent[k],
+									v->BPP,
+									dml_ceil(1.0 * v->HActive[k] / v->NumberOfDSCSlices[k], 1.0),
+									v->NumberOfDSCSlices[k] / 2,
+									v->OutputFormat[k],
+									v->Output[k]) + dscComputeDelay(v->OutputFormat[k], v->Output[k]));
+				} else {
+					v->DSCDelayPerState[i][k] = 4.0
+							* (dscceComputeDelay(
+									v->DSCInputBitPerComponent[k],
+									v->BPP,
+									dml_ceil(1.0 * v->HActive[k] / v->NumberOfDSCSlices[k], 1.0),
+									v->NumberOfDSCSlices[k] / 4,
+									v->OutputFormat[k],
+									v->Output[k]) + dscComputeDelay(v->OutputFormat[k], v->Output[k]));
+				}
+				v->DSCDelayPerState[i][k] = v->DSCDelayPerState[i][k] + (v->HTotal[k] - v->HActive[k]) * dml_ceil((double) v->DSCDelayPerState[i][k] / v->HActive[k], 1.0);
+				v->DSCDelayPerState[i][k] = v->DSCDelayPerState[i][k] * v->PixelClock[k] / v->PixelClockBackEnd[k];
+			} else {
+				v->DSCDelayPerState[i][k] = 0.0;
+			}
+		}
+		for (k = 0; k < v->NumberOfActivePlanes; k++) {
+			for (m = 0; m < v->NumberOfActivePlanes; m++) {
+				if (v->BlendingAndTiming[k] == m && v->RequiresDSC[i][m] == true) {
+					v->DSCDelayPerState[i][k] = v->DSCDelayPerState[i][m];
+				}
+			}
+		}
+	}
+
+	//Calculate Swath, DET Configuration, DCFCLKDeepSleep
+	//
+	for (i = 0; i < v->soc.num_states; ++i) {
+		for (j = 0; j <= 1; ++j) {
+			for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+				v->RequiredDPPCLKThisState[k] = v->RequiredDPPCLK[i][j][k];
+				v->NoOfDPPThisState[k] = v->NoOfDPP[i][j][k];
+				v->ODMCombineEnableThisState[k] = v->ODMCombineEnablePerState[i][k];
+			}
+
+			CalculateSwathAndDETConfiguration(
+					false,
+					v->NumberOfActivePlanes,
+					v->DETBufferSizeInKByte[0],
+					v->MaximumSwathWidthLuma,
+					v->MaximumSwathWidthChroma,
+					v->SourceScan,
+					v->SourcePixelFormat,
+					v->SurfaceTiling,
+					v->ViewportWidth,
+					v->ViewportHeight,
+					v->SurfaceWidthY,
+					v->SurfaceWidthC,
+					v->SurfaceHeightY,
+					v->SurfaceHeightC,
+					v->Read256BlockHeightY,
+					v->Read256BlockHeightC,
+					v->Read256BlockWidthY,
+					v->Read256BlockWidthC,
+					v->ODMCombineEnableThisState,
+					v->BlendingAndTiming,
+					v->BytePerPixelY,
+					v->BytePerPixelC,
+					v->BytePerPixelInDETY,
+					v->BytePerPixelInDETC,
+					v->HActive,
+					v->HRatio,
+					v->HRatioChroma,
+					v->NoOfDPPThisState,
+					v->swath_width_luma_ub_this_state,
+					v->swath_width_chroma_ub_this_state,
+					v->SwathWidthYThisState,
+					v->SwathWidthCThisState,
+					v->SwathHeightYThisState,
+					v->SwathHeightCThisState,
+					v->DETBufferSizeYThisState,
+					v->DETBufferSizeCThisState,
+					v->dummystring,
+					&v->ViewportSizeSupport[i][j]);
+
+			CalculateDCFCLKDeepSleep(
+					mode_lib,
+					v->NumberOfActivePlanes,
+					v->BytePerPixelY,
+					v->BytePerPixelC,
+					v->VRatio,
+					v->VRatioChroma,
+					v->SwathWidthYThisState,
+					v->SwathWidthCThisState,
+					v->NoOfDPPThisState,
+					v->HRatio,
+					v->HRatioChroma,
+					v->PixelClock,
+					v->PSCL_FACTOR,
+					v->PSCL_FACTOR_CHROMA,
+					v->RequiredDPPCLKThisState,
+					v->ReadBandwidthLuma,
+					v->ReadBandwidthChroma,
+					v->ReturnBusWidth,
+					&v->ProjectedDCFCLKDeepSleep[i][j]);
+
+			for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+				v->swath_width_luma_ub_all_states[i][j][k] = v->swath_width_luma_ub_this_state[k];
+				v->swath_width_chroma_ub_all_states[i][j][k] = v->swath_width_chroma_ub_this_state[k];
+				v->SwathWidthYAllStates[i][j][k] = v->SwathWidthYThisState[k];
+				v->SwathWidthCAllStates[i][j][k] = v->SwathWidthCThisState[k];
+				v->SwathHeightYAllStates[i][j][k] = v->SwathHeightYThisState[k];
+				v->SwathHeightCAllStates[i][j][k] = v->SwathHeightCThisState[k];
+				v->DETBufferSizeYAllStates[i][j][k] = v->DETBufferSizeYThisState[k];
+				v->DETBufferSizeCAllStates[i][j][k] = v->DETBufferSizeCThisState[k];
+			}
+		}
+	}
+
+	for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+		v->cursor_bw[k] = v->NumberOfCursors[k] * v->CursorWidth[k][0] * v->CursorBPP[k][0] / 8.0
+				/ (v->HTotal[k] / v->PixelClock[k]) * v->VRatio[k];
+	}
+
+	for (i = 0; i < v->soc.num_states; i++) {
+		for (j = 0; j < 2; j++) {
+			bool NotUrgentLatencyHiding[DC__NUM_DPP__MAX];
+
+			for (k = 0; k < v->NumberOfActivePlanes; k++) {
+				v->swath_width_luma_ub_this_state[k] = v->swath_width_luma_ub_all_states[i][j][k];
+				v->swath_width_chroma_ub_this_state[k] = v->swath_width_chroma_ub_all_states[i][j][k];
+				v->SwathWidthYThisState[k] = v->SwathWidthYAllStates[i][j][k];
+				v->SwathWidthCThisState[k] = v->SwathWidthCAllStates[i][j][k];
+				v->SwathHeightYThisState[k] = v->SwathHeightYAllStates[i][j][k];
+				v->SwathHeightCThisState[k] = v->SwathHeightCAllStates[i][j][k];
+				v->DETBufferSizeYThisState[k] = v->DETBufferSizeYAllStates[i][j][k];
+				v->DETBufferSizeCThisState[k] = v->DETBufferSizeCAllStates[i][j][k];
+			}
+
+			v->TotalNumberOfDCCActiveDPP[i][j] = 0;
+			for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+				if (v->DCCEnable[k] == true) {
+					v->TotalNumberOfDCCActiveDPP[i][j] = v->TotalNumberOfDCCActiveDPP[i][j] + v->NoOfDPP[i][j][k];
+				}
+			}
+
+			for (k = 0; k < v->NumberOfActivePlanes; k++) {
+				if (v->SourcePixelFormat[k] == dm_420_8 || v->SourcePixelFormat[k] == dm_420_10
+						|| v->SourcePixelFormat[k] == dm_420_12 || v->SourcePixelFormat[k] == dm_rgbe_alpha) {
+
+					if ((v->SourcePixelFormat[k] == dm_420_10 || v->SourcePixelFormat[k] == dm_420_12)
+							&& v->SourceScan[k] != dm_vert) {
+						v->PTEBufferSizeInRequestsForLuma = (v->PTEBufferSizeInRequestsLuma + v->PTEBufferSizeInRequestsChroma)
+								/ 2;
+						v->PTEBufferSizeInRequestsForChroma = v->PTEBufferSizeInRequestsForLuma;
+					} else {
+						v->PTEBufferSizeInRequestsForLuma = v->PTEBufferSizeInRequestsLuma;
+						v->PTEBufferSizeInRequestsForChroma = v->PTEBufferSizeInRequestsChroma;
+					}
+
+					v->PDEAndMetaPTEBytesPerFrameC = CalculateVMAndRowBytes(
+							mode_lib,
+							v->DCCEnable[k],
+							v->Read256BlockHeightC[k],
+							v->Read256BlockWidthC[k],
+							v->SourcePixelFormat[k],
+							v->SurfaceTiling[k],
+							v->BytePerPixelC[k],
+							v->SourceScan[k],
+							v->SwathWidthCThisState[k],
+							v->ViewportHeightChroma[k],
+							v->GPUVMEnable,
+							v->HostVMEnable,
+							v->HostVMMaxNonCachedPageTableLevels,
+							v->GPUVMMinPageSize,
+							v->HostVMMinPageSize,
+							v->PTEBufferSizeInRequestsForChroma,
+							v->PitchC[k],
+							0.0,
+							&v->MacroTileWidthC[k],
+							&v->MetaRowBytesC,
+							&v->DPTEBytesPerRowC,
+							&v->PTEBufferSizeNotExceededC[i][j][k],
+							&v->dummyinteger7,
+							&v->dpte_row_height_chroma[k],
+							&v->dummyinteger28,
+							&v->dummyinteger26,
+							&v->dummyinteger23,
+							&v->meta_row_height_chroma[k],
+							&v->dummyinteger8,
+							&v->dummyinteger9,
+							&v->dummyinteger19,
+							&v->dummyinteger20,
+							&v->dummyinteger17,
+							&v->dummyinteger10,
+							&v->dummyinteger11);
+
+					v->PrefetchLinesC[i][j][k] = CalculatePrefetchSourceLines(
+							mode_lib,
+							v->VRatioChroma[k],
+							v->VTAPsChroma[k],
+							v->Interlace[k],
+							v->ProgressiveToInterlaceUnitInOPP,
+							v->SwathHeightCThisState[k],
+							v->ViewportYStartC[k],
+							&v->PrefillC[k],
+							&v->MaxNumSwC[k]);
+				} else {
+					v->PTEBufferSizeInRequestsForLuma = v->PTEBufferSizeInRequestsLuma + v->PTEBufferSizeInRequestsChroma;
+					v->PTEBufferSizeInRequestsForChroma = 0;
+					v->PDEAndMetaPTEBytesPerFrameC = 0.0;
+					v->MetaRowBytesC = 0.0;
+					v->DPTEBytesPerRowC = 0.0;
+					v->PrefetchLinesC[i][j][k] = 0.0;
+					v->PTEBufferSizeNotExceededC[i][j][k] = true;
+				}
+				v->PDEAndMetaPTEBytesPerFrameY = CalculateVMAndRowBytes(
+						mode_lib,
+						v->DCCEnable[k],
+						v->Read256BlockHeightY[k],
+						v->Read256BlockWidthY[k],
+						v->SourcePixelFormat[k],
+						v->SurfaceTiling[k],
+						v->BytePerPixelY[k],
+						v->SourceScan[k],
+						v->SwathWidthYThisState[k],
+						v->ViewportHeight[k],
+						v->GPUVMEnable,
+						v->HostVMEnable,
+						v->HostVMMaxNonCachedPageTableLevels,
+						v->GPUVMMinPageSize,
+						v->HostVMMinPageSize,
+						v->PTEBufferSizeInRequestsForLuma,
+						v->PitchY[k],
+						v->DCCMetaPitchY[k],
+						&v->MacroTileWidthY[k],
+						&v->MetaRowBytesY,
+						&v->DPTEBytesPerRowY,
+						&v->PTEBufferSizeNotExceededY[i][j][k],
+						&v->dummyinteger7,
+						&v->dpte_row_height[k],
+						&v->dummyinteger29,
+						&v->dummyinteger27,
+						&v->dummyinteger24,
+						&v->meta_row_height[k],
+						&v->dummyinteger25,
+						&v->dpte_group_bytes[k],
+						&v->dummyinteger21,
+						&v->dummyinteger22,
+						&v->dummyinteger18,
+						&v->dummyinteger5,
+						&v->dummyinteger6);
+				v->PrefetchLinesY[i][j][k] = CalculatePrefetchSourceLines(
+						mode_lib,
+						v->VRatio[k],
+						v->vtaps[k],
+						v->Interlace[k],
+						v->ProgressiveToInterlaceUnitInOPP,
+						v->SwathHeightYThisState[k],
+						v->ViewportYStartY[k],
+						&v->PrefillY[k],
+						&v->MaxNumSwY[k]);
+				v->PDEAndMetaPTEBytesPerFrame[i][j][k] = v->PDEAndMetaPTEBytesPerFrameY + v->PDEAndMetaPTEBytesPerFrameC;
+				v->MetaRowBytes[i][j][k] = v->MetaRowBytesY + v->MetaRowBytesC;
+				v->DPTEBytesPerRow[i][j][k] = v->DPTEBytesPerRowY + v->DPTEBytesPerRowC;
+
+				CalculateRowBandwidth(
+						v->GPUVMEnable,
+						v->SourcePixelFormat[k],
+						v->VRatio[k],
+						v->VRatioChroma[k],
+						v->DCCEnable[k],
+						v->HTotal[k] / v->PixelClock[k],
+						v->MetaRowBytesY,
+						v->MetaRowBytesC,
+						v->meta_row_height[k],
+						v->meta_row_height_chroma[k],
+						v->DPTEBytesPerRowY,
+						v->DPTEBytesPerRowC,
+						v->dpte_row_height[k],
+						v->dpte_row_height_chroma[k],
+						&v->meta_row_bandwidth[i][j][k],
+						&v->dpte_row_bandwidth[i][j][k]);
+			}
+			/*
+			 * DCCMetaBufferSizeSupport(i, j) = True
+			 * For k = 0 To NumberOfActivePlanes - 1
+			 * If MetaRowBytes(i, j, k) > 24064 Then
+			 * DCCMetaBufferSizeSupport(i, j) = False
+			 * End If
+			 * Next k
+			 */
+			v->DCCMetaBufferSizeSupport[i][j] = true;
+			for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+				if (v->MetaRowBytes[i][j][k] > 24064)
+					v->DCCMetaBufferSizeSupport[i][j] = false;
+			}
+			v->UrgLatency[i] = CalculateUrgentLatency(
+					v->UrgentLatencyPixelDataOnly,
+					v->UrgentLatencyPixelMixedWithVMData,
+					v->UrgentLatencyVMDataOnly,
+					v->DoUrgentLatencyAdjustment,
+					v->UrgentLatencyAdjustmentFabricClockComponent,
+					v->UrgentLatencyAdjustmentFabricClockReference,
+					v->FabricClockPerState[i]);
+
+			for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+				CalculateUrgentBurstFactor(
+						v->swath_width_luma_ub_this_state[k],
+						v->swath_width_chroma_ub_this_state[k],
+						v->SwathHeightYThisState[k],
+						v->SwathHeightCThisState[k],
+						v->HTotal[k] / v->PixelClock[k],
+						v->UrgLatency[i],
+						v->CursorBufferSize,
+						v->CursorWidth[k][0],
+						v->CursorBPP[k][0],
+						v->VRatio[k],
+						v->VRatioChroma[k],
+						v->BytePerPixelInDETY[k],
+						v->BytePerPixelInDETC[k],
+						v->DETBufferSizeYThisState[k],
+						v->DETBufferSizeCThisState[k],
+						&v->UrgentBurstFactorCursor[k],
+						&v->UrgentBurstFactorLuma[k],
+						&v->UrgentBurstFactorChroma[k],
+						&NotUrgentLatencyHiding[k]);
+			}
+
+			v->NotEnoughUrgentLatencyHidingA[i][j] = false;
+			for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+				if (NotUrgentLatencyHiding[k]) {
+					v->NotEnoughUrgentLatencyHidingA[i][j] = true;
+				}
+			}
+
+			for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+				v->VActivePixelBandwidth[i][j][k] = v->ReadBandwidthLuma[k] * v->UrgentBurstFactorLuma[k]
+						+ v->ReadBandwidthChroma[k] * v->UrgentBurstFactorChroma[k];
+				v->VActiveCursorBandwidth[i][j][k] = v->cursor_bw[k] * v->UrgentBurstFactorCursor[k];
+			}
+
+			v->TotalVActivePixelBandwidth[i][j] = 0;
+			v->TotalVActiveCursorBandwidth[i][j] = 0;
+			v->TotalMetaRowBandwidth[i][j] = 0;
+			v->TotalDPTERowBandwidth[i][j] = 0;
+			for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+				v->TotalVActivePixelBandwidth[i][j] = v->TotalVActivePixelBandwidth[i][j] + v->VActivePixelBandwidth[i][j][k];
+				v->TotalVActiveCursorBandwidth[i][j] = v->TotalVActiveCursorBandwidth[i][j] + v->VActiveCursorBandwidth[i][j][k];
+				v->TotalMetaRowBandwidth[i][j] = v->TotalMetaRowBandwidth[i][j] + v->NoOfDPP[i][j][k] * v->meta_row_bandwidth[i][j][k];
+				v->TotalDPTERowBandwidth[i][j] = v->TotalDPTERowBandwidth[i][j] + v->NoOfDPP[i][j][k] * v->dpte_row_bandwidth[i][j][k];
+			}
+		}
+	}
+
+	//Calculate Return BW
+	for (i = 0; i < v->soc.num_states; ++i) {
+		for (j = 0; j <= 1; ++j) {
+			for (k = 0; k < v->NumberOfActivePlanes; k++) {
+				if (v->BlendingAndTiming[k] == k) {
+					if (v->WritebackEnable[k] == true) {
+						v->WritebackDelayTime[k] = v->WritebackLatency
+								+ CalculateWriteBackDelay(
+										v->WritebackPixelFormat[k],
+										v->WritebackHRatio[k],
+										v->WritebackVRatio[k],
+										v->WritebackVTaps[k],
+										v->WritebackDestinationWidth[k],
+										v->WritebackDestinationHeight[k],
+										v->WritebackSourceHeight[k],
+										v->HTotal[k]) / v->RequiredDISPCLK[i][j];
+					} else {
+						v->WritebackDelayTime[k] = 0.0;
+					}
+					for (m = 0; m < v->NumberOfActivePlanes; m++) {
+						if (v->BlendingAndTiming[m] == k && v->WritebackEnable[m] == true) {
+							v->WritebackDelayTime[k] = dml_max(
+									v->WritebackDelayTime[k],
+									v->WritebackLatency
+											+ CalculateWriteBackDelay(
+													v->WritebackPixelFormat[m],
+													v->WritebackHRatio[m],
+													v->WritebackVRatio[m],
+													v->WritebackVTaps[m],
+													v->WritebackDestinationWidth[m],
+													v->WritebackDestinationHeight[m],
+													v->WritebackSourceHeight[m],
+													v->HTotal[m]) / v->RequiredDISPCLK[i][j]);
+						}
+					}
+				}
+			}
+			for (k = 0; k < v->NumberOfActivePlanes; k++) {
+				for (m = 0; m < v->NumberOfActivePlanes; m++) {
+					if (v->BlendingAndTiming[k] == m) {
+						v->WritebackDelayTime[k] = v->WritebackDelayTime[m];
+					}
+				}
+			}
+			v->MaxMaxVStartup[i][j] = 0;
+			for (k = 0; k < v->NumberOfActivePlanes; k++) {
+				v->MaximumVStartup[i][j][k] =
+					CalculateMaxVStartup(
+						v->VTotal[k],
+						v->VActive[k],
+						v->VBlankNom[k],
+						v->HTotal[k],
+						v->PixelClock[k],
+						v->ProgressiveToInterlaceUnitInOPP,
+						v->Interlace[k],
+						v->ip.VBlankNomDefaultUS,
+						v->WritebackDelayTime[k]);
+				v->MaxMaxVStartup[i][j] = dml_max(v->MaxMaxVStartup[i][j], v->MaximumVStartup[i][j][k]);
+				}
+		}
+	}
+
+	ReorderingBytes = v->NumberOfChannels
+			* dml_max3(
+					v->UrgentOutOfOrderReturnPerChannelPixelDataOnly,
+					v->UrgentOutOfOrderReturnPerChannelPixelMixedWithVMData,
+					v->UrgentOutOfOrderReturnPerChannelVMDataOnly);
+
+	for (i = 0; i < v->soc.num_states; ++i) {
+		for (j = 0; j <= 1; ++j) {
+			v->DCFCLKState[i][j] = v->DCFCLKPerState[i];
+		}
+	}
+
+	if (v->UseMinimumRequiredDCFCLK == true)
+		UseMinimumDCFCLK(mode_lib, MaxPrefetchMode, ReorderingBytes);
+
+	for (i = 0; i < v->soc.num_states; ++i) {
+		for (j = 0; j <= 1; ++j) {
+			double IdealFabricAndSDPPortBandwidthPerState = dml_min(
+					v->ReturnBusWidth * v->DCFCLKState[i][j],
+					v->FabricClockPerState[i] * v->FabricDatapathToDCNDataReturn);
+			double IdealDRAMBandwidthPerState = v->DRAMSpeedPerState[i] * v->NumberOfChannels * v->DRAMChannelWidth;
+			double PixelDataOnlyReturnBWPerState = dml_min(
+					IdealFabricAndSDPPortBandwidthPerState * v->PercentOfIdealFabricAndSDPPortBWReceivedAfterUrgLatency / 100.0,
+					IdealDRAMBandwidthPerState * v->PercentOfIdealDRAMBWReceivedAfterUrgLatencyPixelDataOnly / 100.0);
+			double PixelMixedWithVMDataReturnBWPerState = dml_min(
+					IdealFabricAndSDPPortBandwidthPerState * v->PercentOfIdealFabricAndSDPPortBWReceivedAfterUrgLatency / 100.0,
+					IdealDRAMBandwidthPerState * v->PercentOfIdealDRAMBWReceivedAfterUrgLatencyPixelMixedWithVMData / 100.0);
+
+			if (v->HostVMEnable != true) {
+				v->ReturnBWPerState[i][j] = PixelDataOnlyReturnBWPerState;
+			} else {
+				v->ReturnBWPerState[i][j] = PixelMixedWithVMDataReturnBWPerState;
+			}
+		}
+	}
+
+	//Re-ordering Buffer Support Check
+	for (i = 0; i < v->soc.num_states; ++i) {
+		for (j = 0; j <= 1; ++j) {
+			if ((v->ROBBufferSizeInKByte - v->PixelChunkSizeInKByte) * 1024 / v->ReturnBWPerState[i][j]
+					> (v->RoundTripPingLatencyCycles + __DML_ARB_TO_RET_DELAY__) / v->DCFCLKState[i][j] + ReorderingBytes / v->ReturnBWPerState[i][j]) {
+				v->ROBSupport[i][j] = true;
+			} else {
+				v->ROBSupport[i][j] = false;
+			}
+		}
+	}
+
+	//Vertical Active BW support check
+
+	MaxTotalVActiveRDBandwidth = 0;
+	for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+		MaxTotalVActiveRDBandwidth = MaxTotalVActiveRDBandwidth + v->ReadBandwidthLuma[k] + v->ReadBandwidthChroma[k];
+	}
+
+	for (i = 0; i < v->soc.num_states; ++i) {
+		for (j = 0; j <= 1; ++j) {
+			v->MaxTotalVerticalActiveAvailableBandwidth[i][j] = dml_min(
+					dml_min(
+							v->ReturnBusWidth * v->DCFCLKState[i][j],
+							v->FabricClockPerState[i] * v->FabricDatapathToDCNDataReturn)
+							* v->MaxAveragePercentOfIdealFabricAndSDPPortBWDisplayCanUseInNormalSystemOperation / 100,
+					v->DRAMSpeedPerState[i] * v->NumberOfChannels * v->DRAMChannelWidth
+							* v->MaxAveragePercentOfIdealDRAMBWDisplayCanUseInNormalSystemOperation / 100);
+
+			if (MaxTotalVActiveRDBandwidth <= v->MaxTotalVerticalActiveAvailableBandwidth[i][j]) {
+				v->TotalVerticalActiveBandwidthSupport[i][j] = true;
+			} else {
+				v->TotalVerticalActiveBandwidthSupport[i][j] = false;
+			}
+		}
+	}
+
+	v->UrgentLatency = CalculateUrgentLatency(
+			v->UrgentLatencyPixelDataOnly,
+			v->UrgentLatencyPixelMixedWithVMData,
+			v->UrgentLatencyVMDataOnly,
+			v->DoUrgentLatencyAdjustment,
+			v->UrgentLatencyAdjustmentFabricClockComponent,
+			v->UrgentLatencyAdjustmentFabricClockReference,
+			v->FabricClock);
+	//Prefetch Check
+	for (i = 0; i < v->soc.num_states; ++i) {
+		for (j = 0; j <= 1; ++j) {
+			double VMDataOnlyReturnBWPerState;
+			double HostVMInefficiencyFactor = 1;
+			int NextPrefetchModeState = MinPrefetchMode;
+			bool UnboundedRequestEnabledThisState = false;
+			int CompressedBufferSizeInkByteThisState = 0;
+			double dummy;
+
+			v->TimeCalc = 24 / v->ProjectedDCFCLKDeepSleep[i][j];
+
+			v->BandwidthWithoutPrefetchSupported[i][j] = true;
+			if (v->TotalVActivePixelBandwidth[i][j] + v->TotalVActiveCursorBandwidth[i][j] + v->TotalMetaRowBandwidth[i][j]
+					+ v->TotalDPTERowBandwidth[i][j] > v->ReturnBWPerState[i][j] || v->NotEnoughUrgentLatencyHidingA[i][j]) {
+				v->BandwidthWithoutPrefetchSupported[i][j] = false;
+			}
+
+			for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+				v->NoOfDPPThisState[k] = v->NoOfDPP[i][j][k];
+				v->swath_width_luma_ub_this_state[k] = v->swath_width_luma_ub_all_states[i][j][k];
+				v->swath_width_chroma_ub_this_state[k] = v->swath_width_chroma_ub_all_states[i][j][k];
+				v->SwathWidthYThisState[k] = v->SwathWidthYAllStates[i][j][k];
+				v->SwathWidthCThisState[k] = v->SwathWidthCAllStates[i][j][k];
+				v->SwathHeightYThisState[k] = v->SwathHeightYAllStates[i][j][k];
+				v->SwathHeightCThisState[k] = v->SwathHeightCAllStates[i][j][k];
+				v->DETBufferSizeYThisState[k] = v->DETBufferSizeYAllStates[i][j][k];
+				v->DETBufferSizeCThisState[k] = v->DETBufferSizeCAllStates[i][j][k];
+			}
+
+			VMDataOnlyReturnBWPerState = dml_min(
+					dml_min(
+							v->ReturnBusWidth * v->DCFCLKState[i][j],
+							v->FabricClockPerState[i] * v->FabricDatapathToDCNDataReturn)
+							* v->PercentOfIdealFabricAndSDPPortBWReceivedAfterUrgLatency / 100.0,
+					v->DRAMSpeedPerState[i] * v->NumberOfChannels * v->DRAMChannelWidth
+							* v->PercentOfIdealDRAMBWReceivedAfterUrgLatencyVMDataOnly / 100.0);
+			if (v->GPUVMEnable && v->HostVMEnable)
+				HostVMInefficiencyFactor = v->ReturnBWPerState[i][j] / VMDataOnlyReturnBWPerState;
+
+			v->ExtraLatency = CalculateExtraLatency(
+					v->RoundTripPingLatencyCycles,
+					ReorderingBytes,
+					v->DCFCLKState[i][j],
+					v->TotalNumberOfActiveDPP[i][j],
+					v->PixelChunkSizeInKByte,
+					v->TotalNumberOfDCCActiveDPP[i][j],
+					v->MetaChunkSize,
+					v->ReturnBWPerState[i][j],
+					v->GPUVMEnable,
+					v->HostVMEnable,
+					v->NumberOfActivePlanes,
+					v->NoOfDPPThisState,
+					v->dpte_group_bytes,
+					HostVMInefficiencyFactor,
+					v->HostVMMinPageSize,
+					v->HostVMMaxNonCachedPageTableLevels);
+
+			v->NextMaxVStartup = v->MaxMaxVStartup[i][j];
+			do {
+				v->PrefetchModePerState[i][j] = NextPrefetchModeState;
+				v->MaxVStartup = v->NextMaxVStartup;
+
+				v->TWait = CalculateTWait(
+						v->PrefetchModePerState[i][j],
+						v->DRAMClockChangeLatency,
+						v->UrgLatency[i],
+						v->SREnterPlusExitTime);
+
+				for (k = 0; k < v->NumberOfActivePlanes; k++) {
+					CalculatePrefetchSchedulePerPlane(mode_lib,
+									  HostVMInefficiencyFactor,
+									  i, j,	k);
+				}
+
+				for (k = 0; k < v->NumberOfActivePlanes; k++) {
+					CalculateUrgentBurstFactor(
+							v->swath_width_luma_ub_this_state[k],
+							v->swath_width_chroma_ub_this_state[k],
+							v->SwathHeightYThisState[k],
+							v->SwathHeightCThisState[k],
+							v->HTotal[k] / v->PixelClock[k],
+							v->UrgLatency[i],
+							v->CursorBufferSize,
+							v->CursorWidth[k][0],
+							v->CursorBPP[k][0],
+							v->VRatioPreY[i][j][k],
+							v->VRatioPreC[i][j][k],
+							v->BytePerPixelInDETY[k],
+							v->BytePerPixelInDETC[k],
+							v->DETBufferSizeYThisState[k],
+							v->DETBufferSizeCThisState[k],
+							&v->UrgentBurstFactorCursorPre[k],
+							&v->UrgentBurstFactorLumaPre[k],
+							&v->UrgentBurstFactorChromaPre[k],
+							&v->NotUrgentLatencyHidingPre[k]);
+				}
+
+				v->MaximumReadBandwidthWithPrefetch = 0.0;
+				for (k = 0; k < v->NumberOfActivePlanes; k++) {
+					v->cursor_bw_pre[k] = v->NumberOfCursors[k] * v->CursorWidth[k][0] * v->CursorBPP[k][0] / 8.0
+							/ (v->HTotal[k] / v->PixelClock[k]) * v->VRatioPreY[i][j][k];
+
+					v->MaximumReadBandwidthWithPrefetch =
+							v->MaximumReadBandwidthWithPrefetch
+									+ dml_max3(
+											v->VActivePixelBandwidth[i][j][k]
+													+ v->VActiveCursorBandwidth[i][j][k]
+													+ v->NoOfDPP[i][j][k]
+															* (v->meta_row_bandwidth[i][j][k]
+																	+ v->dpte_row_bandwidth[i][j][k]),
+											v->NoOfDPP[i][j][k] * v->prefetch_vmrow_bw[k],
+											v->NoOfDPP[i][j][k]
+													* (v->RequiredPrefetchPixelDataBWLuma[i][j][k]
+															* v->UrgentBurstFactorLumaPre[k]
+															+ v->RequiredPrefetchPixelDataBWChroma[i][j][k]
+																	* v->UrgentBurstFactorChromaPre[k])
+													+ v->cursor_bw_pre[k] * v->UrgentBurstFactorCursorPre[k]);
+				}
+
+				v->NotEnoughUrgentLatencyHidingPre = false;
+				for (k = 0; k < v->NumberOfActivePlanes; k++) {
+					if (v->NotUrgentLatencyHidingPre[k] == true) {
+						v->NotEnoughUrgentLatencyHidingPre = true;
+					}
+				}
+
+				v->PrefetchSupported[i][j] = true;
+				if (v->BandwidthWithoutPrefetchSupported[i][j] == false || v->MaximumReadBandwidthWithPrefetch > v->ReturnBWPerState[i][j]
+						|| v->NotEnoughUrgentLatencyHidingPre == 1) {
+					v->PrefetchSupported[i][j] = false;
+				}
+				for (k = 0; k < v->NumberOfActivePlanes; k++) {
+					if (v->LineTimesForPrefetch[k] < 2.0 || v->LinesForMetaPTE[k] >= 32.0 || v->LinesForMetaAndDPTERow[k] >= 16.0
+							|| v->NoTimeForPrefetch[i][j][k] == true) {
+						v->PrefetchSupported[i][j] = false;
+					}
+				}
+
+				v->DynamicMetadataSupported[i][j] = true;
+				for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+					if (v->NoTimeForDynamicMetadata[i][j][k] == true) {
+						v->DynamicMetadataSupported[i][j] = false;
+					}
+				}
+
+				v->VRatioInPrefetchSupported[i][j] = true;
+				for (k = 0; k < v->NumberOfActivePlanes; k++) {
+					if (v->VRatioPreY[i][j][k] > 4.0 || v->VRatioPreC[i][j][k] > 4.0 || v->NoTimeForPrefetch[i][j][k] == true) {
+						v->VRatioInPrefetchSupported[i][j] = false;
+					}
+				}
+				v->AnyLinesForVMOrRowTooLarge = false;
+				for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+					if (v->LinesForMetaAndDPTERow[k] >= 16 || v->LinesForMetaPTE[k] >= 32) {
+						v->AnyLinesForVMOrRowTooLarge = true;
+					}
+				}
+
+				v->NextPrefetchMode = v->NextPrefetchMode + 1;
+
+				if (v->PrefetchSupported[i][j] == true && v->VRatioInPrefetchSupported[i][j] == true) {
+					v->BandwidthAvailableForImmediateFlip = v->ReturnBWPerState[i][j];
+					for (k = 0; k < v->NumberOfActivePlanes; k++) {
+						v->BandwidthAvailableForImmediateFlip = v->BandwidthAvailableForImmediateFlip
+								- dml_max(
+										v->VActivePixelBandwidth[i][j][k] + v->VActiveCursorBandwidth[i][j][k],
+										v->NoOfDPP[i][j][k]
+												* (v->RequiredPrefetchPixelDataBWLuma[i][j][k]
+														* v->UrgentBurstFactorLumaPre[k]
+														+ v->RequiredPrefetchPixelDataBWChroma[i][j][k]
+																* v->UrgentBurstFactorChromaPre[k])
+												+ v->cursor_bw_pre[k] * v->UrgentBurstFactorCursorPre[k]);
+					}
+					v->TotImmediateFlipBytes = 0.0;
+					for (k = 0; k < v->NumberOfActivePlanes; k++) {
+						v->TotImmediateFlipBytes = v->TotImmediateFlipBytes
+								+ v->NoOfDPP[i][j][k] * (v->PDEAndMetaPTEBytesPerFrame[i][j][k] + v->MetaRowBytes[i][j][k]
+								+ v->DPTEBytesPerRow[i][j][k]);
+					}
+
+					for (k = 0; k < v->NumberOfActivePlanes; k++) {
+						CalculateFlipSchedule(
+								mode_lib,
+								k,
+								HostVMInefficiencyFactor,
+								v->ExtraLatency,
+								v->UrgLatency[i],
+								v->PDEAndMetaPTEBytesPerFrame[i][j][k],
+								v->MetaRowBytes[i][j][k],
+								v->DPTEBytesPerRow[i][j][k]);
+					}
+					v->total_dcn_read_bw_with_flip = 0.0;
+					for (k = 0; k < v->NumberOfActivePlanes; k++) {
+						v->total_dcn_read_bw_with_flip = v->total_dcn_read_bw_with_flip
+								+ dml_max3(
+										v->NoOfDPP[i][j][k] * v->prefetch_vmrow_bw[k],
+										v->NoOfDPP[i][j][k] * v->final_flip_bw[k] + v->VActivePixelBandwidth[i][j][k]
+												+ v->VActiveCursorBandwidth[i][j][k],
+										v->NoOfDPP[i][j][k]
+												* (v->final_flip_bw[k]
+														+ v->RequiredPrefetchPixelDataBWLuma[i][j][k]
+																* v->UrgentBurstFactorLumaPre[k]
+														+ v->RequiredPrefetchPixelDataBWChroma[i][j][k]
+																* v->UrgentBurstFactorChromaPre[k])
+												+ v->cursor_bw_pre[k] * v->UrgentBurstFactorCursorPre[k]);
+					}
+					v->ImmediateFlipSupportedForState[i][j] = true;
+					if (v->total_dcn_read_bw_with_flip > v->ReturnBWPerState[i][j]) {
+						v->ImmediateFlipSupportedForState[i][j] = false;
+					}
+					for (k = 0; k < v->NumberOfActivePlanes; k++) {
+						if (v->ImmediateFlipSupportedForPipe[k] == false) {
+							v->ImmediateFlipSupportedForState[i][j] = false;
+						}
+					}
+				} else {
+					v->ImmediateFlipSupportedForState[i][j] = false;
+				}
+
+				if (v->MaxVStartup <= __DML_VBA_MIN_VSTARTUP__ || v->AnyLinesForVMOrRowTooLarge == false) {
+					v->NextMaxVStartup = v->MaxMaxVStartup[i][j];
+					NextPrefetchModeState = NextPrefetchModeState + 1;
+				} else {
+					v->NextMaxVStartup = v->NextMaxVStartup - 1;
+				}
+				v->NextPrefetchMode = v->NextPrefetchMode + 1;
+			} while (!((v->PrefetchSupported[i][j] == true && v->DynamicMetadataSupported[i][j] == true && v->VRatioInPrefetchSupported[i][j] == true
+					&& ((v->HostVMEnable == false &&
+							v->ImmediateFlipRequirement[0] != dm_immediate_flip_required)
+							|| v->ImmediateFlipSupportedForState[i][j] == true))
+					|| (v->NextMaxVStartup == v->MaxMaxVStartup[i][j] && NextPrefetchModeState > MaxPrefetchMode)));
+
+			CalculateUnboundedRequestAndCompressedBufferSize(
+					v->DETBufferSizeInKByte[0],
+					v->ConfigReturnBufferSizeInKByte,
+					v->UseUnboundedRequesting,
+					v->TotalNumberOfActiveDPP[i][j],
+					NoChroma,
+					v->MaxNumDPP,
+					v->CompressedBufferSegmentSizeInkByte,
+					v->Output,
+					&UnboundedRequestEnabledThisState,
+					&CompressedBufferSizeInkByteThisState);
+
+			CalculateWatermarksAndDRAMSpeedChangeSupport(
+					mode_lib,
+					v->PrefetchModePerState[i][j],
+					v->DCFCLKState[i][j],
+					v->ReturnBWPerState[i][j],
+					v->UrgLatency[i],
+					v->ExtraLatency,
+					v->SOCCLKPerState[i],
+					v->ProjectedDCFCLKDeepSleep[i][j],
+					v->DETBufferSizeYThisState,
+					v->DETBufferSizeCThisState,
+					v->SwathHeightYThisState,
+					v->SwathHeightCThisState,
+					v->SwathWidthYThisState,
+					v->SwathWidthCThisState,
+					v->NoOfDPPThisState,
+					v->BytePerPixelInDETY,
+					v->BytePerPixelInDETC,
+					UnboundedRequestEnabledThisState,
+					CompressedBufferSizeInkByteThisState,
+					&v->DRAMClockChangeSupport[i][j],
+					&dummy,
+					&dummy,
+					&dummy,
+					&dummy);
+		}
+	}
+
+	/*PTE Buffer Size Check*/
+	for (i = 0; i < v->soc.num_states; i++) {
+		for (j = 0; j < 2; j++) {
+			v->PTEBufferSizeNotExceeded[i][j] = true;
+			for (k = 0; k < v->NumberOfActivePlanes; k++) {
+				if (v->PTEBufferSizeNotExceededY[i][j][k] == false || v->PTEBufferSizeNotExceededC[i][j][k] == false) {
+					v->PTEBufferSizeNotExceeded[i][j] = false;
+				}
+			}
+		}
+	}
+
+	/*Cursor Support Check*/
+	v->CursorSupport = true;
+	for (k = 0; k < v->NumberOfActivePlanes; k++) {
+		if (v->CursorWidth[k][0] > 0.0) {
+			if (v->CursorBPP[k][0] == 64 && v->Cursor64BppSupport == false) {
+				v->CursorSupport = false;
+			}
+		}
+	}
+
+	/*Valid Pitch Check*/
+	v->PitchSupport = true;
+	for (k = 0; k < v->NumberOfActivePlanes; k++) {
+		v->AlignedYPitch[k] = dml_ceil(dml_max(v->PitchY[k], v->SurfaceWidthY[k]), v->MacroTileWidthY[k]);
+		if (v->DCCEnable[k] == true) {
+			v->AlignedDCCMetaPitchY[k] = dml_ceil(dml_max(v->DCCMetaPitchY[k], v->SurfaceWidthY[k]), 64.0 * v->Read256BlockWidthY[k]);
+		} else {
+			v->AlignedDCCMetaPitchY[k] = v->DCCMetaPitchY[k];
+		}
+		if (v->SourcePixelFormat[k] != dm_444_64 && v->SourcePixelFormat[k] != dm_444_32 && v->SourcePixelFormat[k] != dm_444_16
+				&& v->SourcePixelFormat[k] != dm_mono_16 && v->SourcePixelFormat[k] != dm_rgbe
+				&& v->SourcePixelFormat[k] != dm_mono_8) {
+			v->AlignedCPitch[k] = dml_ceil(dml_max(v->PitchC[k], v->SurfaceWidthC[k]), v->MacroTileWidthC[k]);
+			if (v->DCCEnable[k] == true) {
+				v->AlignedDCCMetaPitchC[k] = dml_ceil(
+						dml_max(v->DCCMetaPitchC[k], v->SurfaceWidthC[k]),
+						64.0 * v->Read256BlockWidthC[k]);
+			} else {
+				v->AlignedDCCMetaPitchC[k] = v->DCCMetaPitchC[k];
+			}
+		} else {
+			v->AlignedCPitch[k] = v->PitchC[k];
+			v->AlignedDCCMetaPitchC[k] = v->DCCMetaPitchC[k];
+		}
+		if (v->AlignedYPitch[k] > v->PitchY[k] || v->AlignedCPitch[k] > v->PitchC[k]
+				|| v->AlignedDCCMetaPitchY[k] > v->DCCMetaPitchY[k] || v->AlignedDCCMetaPitchC[k] > v->DCCMetaPitchC[k]) {
+			v->PitchSupport = false;
+		}
+	}
+
+	for (k = 0; k < v->NumberOfActivePlanes; k++) {
+		if (v->ViewportWidth[k] > v->SurfaceWidthY[k] || v->ViewportHeight[k] > v->SurfaceHeightY[k]) {
+			ViewportExceedsSurface = true;
+			if (v->SourcePixelFormat[k] != dm_444_64 && v->SourcePixelFormat[k] != dm_444_32
+					&& v->SourcePixelFormat[k] != dm_444_16 && v->SourcePixelFormat[k] != dm_444_8
+					&& v->SourcePixelFormat[k] != dm_rgbe) {
+				if (v->ViewportWidthChroma[k] > v->SurfaceWidthC[k]
+						|| v->ViewportHeightChroma[k] > v->SurfaceHeightC[k]) {
+					ViewportExceedsSurface = true;
+				}
+			}
+		}
+	}
+
+	/*Mode Support, Voltage State and SOC Configuration*/
+	for (i = v->soc.num_states - 1; i >= 0; i--) {
+		for (j = 0; j < 2; j++) {
+			if (v->ScaleRatioAndTapsSupport == true && v->SourceFormatPixelAndScanSupport == true && v->ViewportSizeSupport[i][j] == true
+					&& v->LinkCapacitySupport[i] == true && !P2IWith420 && !DSCOnlyIfNecessaryWithBPP
+					&& !DSC422NativeNotSupported && v->ODMCombine4To1SupportCheckOK[i] == true && v->NotEnoughDSCUnits[i] == false
+					&& v->DTBCLKRequiredMoreThanSupported[i] == false
+					&& v->ROBSupport[i][j] == true && v->DISPCLK_DPPCLK_Support[i][j] == true
+					&& v->TotalAvailablePipesSupport[i][j] == true && EnoughWritebackUnits == true
+					&& v->WritebackLatencySupport == true && v->WritebackScaleRatioAndTapsSupport == true
+					&& v->CursorSupport == true && v->PitchSupport == true && ViewportExceedsSurface == false
+					&& v->PrefetchSupported[i][j] == true && v->DynamicMetadataSupported[i][j] == true
+					&& v->TotalVerticalActiveBandwidthSupport[i][j] == true && v->VRatioInPrefetchSupported[i][j] == true
+					&& v->PTEBufferSizeNotExceeded[i][j] == true && v->NonsupportedDSCInputBPC == false
+					&& ((v->HostVMEnable == false
+					&& v->ImmediateFlipRequirement[0] != dm_immediate_flip_required)
+							|| v->ImmediateFlipSupportedForState[i][j] == true)
+					&& FMTBufferExceeded == false) {
+				v->ModeSupport[i][j] = true;
+			} else {
+				v->ModeSupport[i][j] = false;
+			}
+		}
+	}
+	for (i = v->soc.num_states; i >= 0; i--) {
+		for (j = 0; j < 2; j++) {
+			enum dm_validation_status status = DML_VALIDATION_OK;
+
+			if (!v->ScaleRatioAndTapsSupport) {
+				status = DML_FAIL_SCALE_RATIO_TAP;
+			} else if (!v->SourceFormatPixelAndScanSupport) {
+				status = DML_FAIL_SOURCE_PIXEL_FORMAT;
+			} else if (!v->ViewportSizeSupport[i][j]) {
+				status = DML_FAIL_VIEWPORT_SIZE;
+			} else if (P2IWith420) {
+				status = DML_FAIL_P2I_WITH_420;
+			} else if (DSCOnlyIfNecessaryWithBPP) {
+				status = DML_FAIL_DSC_ONLY_IF_NECESSARY_WITH_BPP;
+			} else if (DSC422NativeNotSupported) {
+				status = DML_FAIL_NOT_DSC422_NATIVE;
+			} else if (!v->ODMCombine4To1SupportCheckOK[i]) {
+				status = DML_FAIL_ODM_COMBINE4TO1;
+			} else if (v->NotEnoughDSCUnits[i]) {
+				status = DML_FAIL_NOT_ENOUGH_DSC;
+			} else if (!v->ROBSupport[i][j]) {
+				status = DML_FAIL_REORDERING_BUFFER;
+			} else if (!v->DISPCLK_DPPCLK_Support[i][j]) {
+				status = DML_FAIL_DISPCLK_DPPCLK;
+			} else if (!v->TotalAvailablePipesSupport[i][j]) {
+				status = DML_FAIL_TOTAL_AVAILABLE_PIPES;
+			} else if (!EnoughWritebackUnits) {
+				status = DML_FAIL_ENOUGH_WRITEBACK_UNITS;
+			} else if (!v->WritebackLatencySupport) {
+				status = DML_FAIL_WRITEBACK_LATENCY;
+			} else if (!v->WritebackScaleRatioAndTapsSupport) {
+				status = DML_FAIL_WRITEBACK_SCALE_RATIO_TAP;
+			} else if (!v->CursorSupport) {
+				status = DML_FAIL_CURSOR_SUPPORT;
+			} else if (!v->PitchSupport) {
+				status = DML_FAIL_PITCH_SUPPORT;
+			} else if (ViewportExceedsSurface) {
+				status = DML_FAIL_VIEWPORT_EXCEEDS_SURFACE;
+			} else if (!v->PrefetchSupported[i][j]) {
+				status = DML_FAIL_PREFETCH_SUPPORT;
+			} else if (!v->DynamicMetadataSupported[i][j]) {
+				status = DML_FAIL_DYNAMIC_METADATA;
+			} else if (!v->TotalVerticalActiveBandwidthSupport[i][j]) {
+				status = DML_FAIL_TOTAL_V_ACTIVE_BW;
+			} else if (!v->VRatioInPrefetchSupported[i][j]) {
+				status = DML_FAIL_V_RATIO_PREFETCH;
+			} else if (!v->PTEBufferSizeNotExceeded[i][j]) {
+				status = DML_FAIL_PTE_BUFFER_SIZE;
+			} else if (v->NonsupportedDSCInputBPC) {
+				status = DML_FAIL_DSC_INPUT_BPC;
+			} else if ((v->HostVMEnable
+					&& !v->ImmediateFlipSupportedForState[i][j])) {
+				status = DML_FAIL_HOST_VM_IMMEDIATE_FLIP;
+			} else if (FMTBufferExceeded) {
+				status = DML_FAIL_FMT_BUFFER_EXCEEDED;
+			}
+			mode_lib->vba.ValidationStatus[i] = status;
+		}
+	}
+
+	{
+		unsigned int MaximumMPCCombine = 0;
+
+		for (i = v->soc.num_states; i >= 0; i--) {
+			if (i == v->soc.num_states || v->ModeSupport[i][0] == true || v->ModeSupport[i][1] == true) {
+				v->VoltageLevel = i;
+				v->ModeIsSupported = v->ModeSupport[i][0] == true || v->ModeSupport[i][1] == true;
+				if (v->ModeSupport[i][0] == true) {
+					MaximumMPCCombine = 0;
+				} else {
+					MaximumMPCCombine = 1;
+				}
+			}
+		}
+		v->ImmediateFlipSupport = v->ImmediateFlipSupportedForState[v->VoltageLevel][MaximumMPCCombine];
+		for (k = 0; k <= v->NumberOfActivePlanes - 1; k++) {
+			v->MPCCombineEnable[k] = v->MPCCombine[v->VoltageLevel][MaximumMPCCombine][k];
+			v->DPPPerPlane[k] = v->NoOfDPP[v->VoltageLevel][MaximumMPCCombine][k];
+		}
+		v->DCFCLK = v->DCFCLKState[v->VoltageLevel][MaximumMPCCombine];
+		v->DRAMSpeed = v->DRAMSpeedPerState[v->VoltageLevel];
+		v->FabricClock = v->FabricClockPerState[v->VoltageLevel];
+		v->SOCCLK = v->SOCCLKPerState[v->VoltageLevel];
+		v->ReturnBW = v->ReturnBWPerState[v->VoltageLevel][MaximumMPCCombine];
+		v->maxMpcComb = MaximumMPCCombine;
+	}
+}
+
+static void CalculateWatermarksAndDRAMSpeedChangeSupport(
+		struct display_mode_lib *mode_lib,
+		unsigned int PrefetchMode,
+		double DCFCLK,
+		double ReturnBW,
+		double UrgentLatency,
+		double ExtraLatency,
+		double SOCCLK,
+		double DCFCLKDeepSleep,
+		unsigned int DETBufferSizeY[],
+		unsigned int DETBufferSizeC[],
+		unsigned int SwathHeightY[],
+		unsigned int SwathHeightC[],
+		double SwathWidthY[],
+		double SwathWidthC[],
+		unsigned int DPPPerPlane[],
+		double BytePerPixelDETY[],
+		double BytePerPixelDETC[],
+		bool UnboundedRequestEnabled,
+		unsigned int CompressedBufferSizeInkByte,
+		enum clock_change_support *DRAMClockChangeSupport,
+		double *StutterExitWatermark,
+		double *StutterEnterPlusExitWatermark,
+		double *Z8StutterExitWatermark,
+		double *Z8StutterEnterPlusExitWatermark)
+{
+	struct vba_vars_st *v = &mode_lib->vba;
+	double EffectiveLBLatencyHidingY;
+	double EffectiveLBLatencyHidingC;
+	double LinesInDETY[DC__NUM_DPP__MAX];
+	double LinesInDETC;
+	unsigned int LinesInDETYRoundedDownToSwath[DC__NUM_DPP__MAX];
+	unsigned int LinesInDETCRoundedDownToSwath;
+	double FullDETBufferingTimeY;
+	double FullDETBufferingTimeC;
+	double ActiveDRAMClockChangeLatencyMarginY;
+	double ActiveDRAMClockChangeLatencyMarginC;
+	double WritebackDRAMClockChangeLatencyMargin;
+	double PlaneWithMinActiveDRAMClockChangeMargin;
+	double SecondMinActiveDRAMClockChangeMarginOneDisplayInVBLank;
+	double WritebackDRAMClockChangeLatencyHiding;
+	double TotalPixelBW = 0.0;
+	int k, j;
+
+	v->UrgentWatermark = UrgentLatency + ExtraLatency;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: UrgentLatency = %f\n", __func__, UrgentLatency);
+	dml_print("DML::%s: ExtraLatency = %f\n", __func__, ExtraLatency);
+	dml_print("DML::%s: UrgentWatermark = %f\n", __func__, v->UrgentWatermark);
+#endif
+
+	v->DRAMClockChangeWatermark = v->DRAMClockChangeLatency + v->UrgentWatermark;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: v->DRAMClockChangeLatency = %f\n", __func__, v->DRAMClockChangeLatency);
+	dml_print("DML::%s: DRAMClockChangeWatermark = %f\n", __func__, v->DRAMClockChangeWatermark);
+#endif
+
+	v->TotalActiveWriteback = 0;
+	for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+		if (v->WritebackEnable[k] == true) {
+			v->TotalActiveWriteback = v->TotalActiveWriteback + 1;
+		}
+	}
+
+	if (v->TotalActiveWriteback <= 1) {
+		v->WritebackUrgentWatermark = v->WritebackLatency;
+	} else {
+		v->WritebackUrgentWatermark = v->WritebackLatency + v->WritebackChunkSize * 1024.0 / 32.0 / SOCCLK;
+	}
+
+	if (v->TotalActiveWriteback <= 1) {
+		v->WritebackDRAMClockChangeWatermark = v->DRAMClockChangeLatency + v->WritebackLatency;
+	} else {
+		v->WritebackDRAMClockChangeWatermark = v->DRAMClockChangeLatency + v->WritebackLatency + v->WritebackChunkSize * 1024.0 / 32.0 / SOCCLK;
+	}
+
+	for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+		TotalPixelBW = TotalPixelBW
+				+ DPPPerPlane[k] * (SwathWidthY[k] * BytePerPixelDETY[k] * v->VRatio[k] + SwathWidthC[k] * BytePerPixelDETC[k] * v->VRatioChroma[k])
+						/ (v->HTotal[k] / v->PixelClock[k]);
+	}
+
+	for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+		double EffectiveDETBufferSizeY = DETBufferSizeY[k];
+
+		v->LBLatencyHidingSourceLinesY = dml_min(
+				(double) v->MaxLineBufferLines,
+				dml_floor(v->LineBufferSize / v->LBBitPerPixel[k] / (SwathWidthY[k] / dml_max(v->HRatio[k], 1.0)), 1)) - (v->vtaps[k] - 1);
+
+		v->LBLatencyHidingSourceLinesC = dml_min(
+				(double) v->MaxLineBufferLines,
+				dml_floor(v->LineBufferSize / v->LBBitPerPixel[k] / (SwathWidthC[k] / dml_max(v->HRatioChroma[k], 1.0)), 1)) - (v->VTAPsChroma[k] - 1);
+
+		EffectiveLBLatencyHidingY = v->LBLatencyHidingSourceLinesY / v->VRatio[k] * (v->HTotal[k] / v->PixelClock[k]);
+
+		EffectiveLBLatencyHidingC = v->LBLatencyHidingSourceLinesC / v->VRatioChroma[k] * (v->HTotal[k] / v->PixelClock[k]);
+
+		if (UnboundedRequestEnabled) {
+			EffectiveDETBufferSizeY = EffectiveDETBufferSizeY
+					+ CompressedBufferSizeInkByte * 1024 * SwathWidthY[k] * BytePerPixelDETY[k] * v->VRatio[k] / (v->HTotal[k] / v->PixelClock[k]) / TotalPixelBW;
+		}
+
+		LinesInDETY[k] = (double) EffectiveDETBufferSizeY / BytePerPixelDETY[k] / SwathWidthY[k];
+		LinesInDETYRoundedDownToSwath[k] = dml_floor(LinesInDETY[k], SwathHeightY[k]);
+		FullDETBufferingTimeY = LinesInDETYRoundedDownToSwath[k] * (v->HTotal[k] / v->PixelClock[k]) / v->VRatio[k];
+		if (BytePerPixelDETC[k] > 0) {
+			LinesInDETC = v->DETBufferSizeC[k] / BytePerPixelDETC[k] / SwathWidthC[k];
+			LinesInDETCRoundedDownToSwath = dml_floor(LinesInDETC, SwathHeightC[k]);
+			FullDETBufferingTimeC = LinesInDETCRoundedDownToSwath * (v->HTotal[k] / v->PixelClock[k]) / v->VRatioChroma[k];
+		} else {
+			LinesInDETC = 0;
+			FullDETBufferingTimeC = 999999;
+		}
+
+		ActiveDRAMClockChangeLatencyMarginY = EffectiveLBLatencyHidingY + FullDETBufferingTimeY
+				- ((double) v->DSTXAfterScaler[k] / v->HTotal[k] + v->DSTYAfterScaler[k]) * v->HTotal[k] / v->PixelClock[k] - v->UrgentWatermark - v->DRAMClockChangeWatermark;
+
+		if (v->NumberOfActivePlanes > 1) {
+			ActiveDRAMClockChangeLatencyMarginY = ActiveDRAMClockChangeLatencyMarginY
+					- (1 - 1.0 / v->NumberOfActivePlanes) * SwathHeightY[k] * v->HTotal[k] / v->PixelClock[k] / v->VRatio[k];
+		}
+
+		if (BytePerPixelDETC[k] > 0) {
+			ActiveDRAMClockChangeLatencyMarginC = EffectiveLBLatencyHidingC + FullDETBufferingTimeC
+					- ((double) v->DSTXAfterScaler[k] / v->HTotal[k] + v->DSTYAfterScaler[k]) * v->HTotal[k] / v->PixelClock[k] - v->UrgentWatermark - v->DRAMClockChangeWatermark;
+
+			if (v->NumberOfActivePlanes > 1) {
+				ActiveDRAMClockChangeLatencyMarginC = ActiveDRAMClockChangeLatencyMarginC
+						- (1 - 1.0 / v->NumberOfActivePlanes) * SwathHeightC[k] * v->HTotal[k] / v->PixelClock[k] / v->VRatioChroma[k];
+			}
+			v->ActiveDRAMClockChangeLatencyMargin[k] = dml_min(ActiveDRAMClockChangeLatencyMarginY, ActiveDRAMClockChangeLatencyMarginC);
+		} else {
+			v->ActiveDRAMClockChangeLatencyMargin[k] = ActiveDRAMClockChangeLatencyMarginY;
+		}
+
+		if (v->WritebackEnable[k] == true) {
+			WritebackDRAMClockChangeLatencyHiding = v->WritebackInterfaceBufferSize * 1024
+					/ (v->WritebackDestinationWidth[k] * v->WritebackDestinationHeight[k] / (v->WritebackSourceHeight[k] * v->HTotal[k] / v->PixelClock[k]) * 4);
+			if (v->WritebackPixelFormat[k] == dm_444_64) {
+				WritebackDRAMClockChangeLatencyHiding = WritebackDRAMClockChangeLatencyHiding / 2;
+			}
+			WritebackDRAMClockChangeLatencyMargin = WritebackDRAMClockChangeLatencyHiding - v->WritebackDRAMClockChangeWatermark;
+			v->ActiveDRAMClockChangeLatencyMargin[k] = dml_min(v->ActiveDRAMClockChangeLatencyMargin[k], WritebackDRAMClockChangeLatencyMargin);
+		}
+	}
+
+	v->MinActiveDRAMClockChangeMargin = 999999;
+	PlaneWithMinActiveDRAMClockChangeMargin = 0;
+	for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+		if (v->ActiveDRAMClockChangeLatencyMargin[k] < v->MinActiveDRAMClockChangeMargin) {
+			v->MinActiveDRAMClockChangeMargin = v->ActiveDRAMClockChangeLatencyMargin[k];
+			if (v->BlendingAndTiming[k] == k) {
+				PlaneWithMinActiveDRAMClockChangeMargin = k;
+			} else {
+				for (j = 0; j < v->NumberOfActivePlanes; ++j) {
+					if (v->BlendingAndTiming[k] == j) {
+						PlaneWithMinActiveDRAMClockChangeMargin = j;
+					}
+				}
+			}
+		}
+	}
+
+	v->MinActiveDRAMClockChangeLatencySupported = v->MinActiveDRAMClockChangeMargin + v->DRAMClockChangeLatency ;
+
+	SecondMinActiveDRAMClockChangeMarginOneDisplayInVBLank = 999999;
+	for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+		if (!((k == PlaneWithMinActiveDRAMClockChangeMargin) && (v->BlendingAndTiming[k] == k)) && !(v->BlendingAndTiming[k] == PlaneWithMinActiveDRAMClockChangeMargin)
+				&& v->ActiveDRAMClockChangeLatencyMargin[k] < SecondMinActiveDRAMClockChangeMarginOneDisplayInVBLank) {
+			SecondMinActiveDRAMClockChangeMarginOneDisplayInVBLank = v->ActiveDRAMClockChangeLatencyMargin[k];
+		}
+	}
+
+	v->TotalNumberOfActiveOTG = 0;
+
+	for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+		if (v->BlendingAndTiming[k] == k) {
+			v->TotalNumberOfActiveOTG = v->TotalNumberOfActiveOTG + 1;
+		}
+	}
+
+	if (v->MinActiveDRAMClockChangeMargin > 0 && PrefetchMode == 0) {
+		*DRAMClockChangeSupport = dm_dram_clock_change_vactive;
+	} else if ((v->SynchronizedVBlank == true || v->TotalNumberOfActiveOTG == 1
+			|| SecondMinActiveDRAMClockChangeMarginOneDisplayInVBLank > 0) && PrefetchMode == 0) {
+		*DRAMClockChangeSupport = dm_dram_clock_change_vblank;
+	} else {
+		*DRAMClockChangeSupport = dm_dram_clock_change_unsupported;
+	}
+
+	*StutterExitWatermark = v->SRExitTime + ExtraLatency + 10 / DCFCLKDeepSleep;
+	*StutterEnterPlusExitWatermark = (v->SREnterPlusExitTime + ExtraLatency + 10 / DCFCLKDeepSleep);
+	*Z8StutterExitWatermark = v->SRExitZ8Time + ExtraLatency + 10 / DCFCLKDeepSleep;
+	*Z8StutterEnterPlusExitWatermark = v->SREnterPlusExitZ8Time + ExtraLatency + 10 / DCFCLKDeepSleep;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: StutterExitWatermark = %f\n", __func__, *StutterExitWatermark);
+	dml_print("DML::%s: StutterEnterPlusExitWatermark = %f\n", __func__, *StutterEnterPlusExitWatermark);
+	dml_print("DML::%s: Z8StutterExitWatermark = %f\n", __func__, *Z8StutterExitWatermark);
+	dml_print("DML::%s: Z8StutterEnterPlusExitWatermark = %f\n", __func__, *Z8StutterEnterPlusExitWatermark);
+#endif
+}
+
+static void CalculateDCFCLKDeepSleep(
+		struct display_mode_lib *mode_lib,
+		unsigned int NumberOfActivePlanes,
+		int BytePerPixelY[],
+		int BytePerPixelC[],
+		double VRatio[],
+		double VRatioChroma[],
+		double SwathWidthY[],
+		double SwathWidthC[],
+		unsigned int DPPPerPlane[],
+		double HRatio[],
+		double HRatioChroma[],
+		double PixelClock[],
+		double PSCL_THROUGHPUT[],
+		double PSCL_THROUGHPUT_CHROMA[],
+		double DPPCLK[],
+		double ReadBandwidthLuma[],
+		double ReadBandwidthChroma[],
+		int ReturnBusWidth,
+		double *DCFCLKDeepSleep)
+{
+	struct vba_vars_st *v = &mode_lib->vba;
+	double DisplayPipeLineDeliveryTimeLuma;
+	double DisplayPipeLineDeliveryTimeChroma;
+	double ReadBandwidth = 0.0;
+	int k;
+
+	for (k = 0; k < NumberOfActivePlanes; ++k) {
+
+		if (VRatio[k] <= 1) {
+			DisplayPipeLineDeliveryTimeLuma = SwathWidthY[k] * DPPPerPlane[k] / HRatio[k] / PixelClock[k];
+		} else {
+			DisplayPipeLineDeliveryTimeLuma = SwathWidthY[k] / PSCL_THROUGHPUT[k] / DPPCLK[k];
+		}
+		if (BytePerPixelC[k] == 0) {
+			DisplayPipeLineDeliveryTimeChroma = 0;
+		} else {
+			if (VRatioChroma[k] <= 1) {
+				DisplayPipeLineDeliveryTimeChroma = SwathWidthC[k] * DPPPerPlane[k] / HRatioChroma[k] / PixelClock[k];
+			} else {
+				DisplayPipeLineDeliveryTimeChroma = SwathWidthC[k] / PSCL_THROUGHPUT_CHROMA[k] / DPPCLK[k];
+			}
+		}
+
+		if (BytePerPixelC[k] > 0) {
+			v->DCFCLKDeepSleepPerPlane[k] = dml_max(__DML_MIN_DCFCLK_FACTOR__ * SwathWidthY[k] * BytePerPixelY[k] / 32.0 / DisplayPipeLineDeliveryTimeLuma,
+			__DML_MIN_DCFCLK_FACTOR__ * SwathWidthC[k] * BytePerPixelC[k] / 32.0 / DisplayPipeLineDeliveryTimeChroma);
+		} else {
+			v->DCFCLKDeepSleepPerPlane[k] = __DML_MIN_DCFCLK_FACTOR__ * SwathWidthY[k] * BytePerPixelY[k] / 64.0 / DisplayPipeLineDeliveryTimeLuma;
+		}
+		v->DCFCLKDeepSleepPerPlane[k] = dml_max(v->DCFCLKDeepSleepPerPlane[k], PixelClock[k] / 16);
+
+	}
+
+	for (k = 0; k < NumberOfActivePlanes; ++k) {
+		ReadBandwidth = ReadBandwidth + ReadBandwidthLuma[k] + ReadBandwidthChroma[k];
+	}
+
+	*DCFCLKDeepSleep = dml_max(8.0, __DML_MIN_DCFCLK_FACTOR__ * ReadBandwidth / ReturnBusWidth);
+
+	for (k = 0; k < NumberOfActivePlanes; ++k) {
+		*DCFCLKDeepSleep = dml_max(*DCFCLKDeepSleep, v->DCFCLKDeepSleepPerPlane[k]);
+	}
+}
+
+static void CalculateUrgentBurstFactor(
+		int swath_width_luma_ub,
+		int swath_width_chroma_ub,
+		unsigned int SwathHeightY,
+		unsigned int SwathHeightC,
+		double LineTime,
+		double UrgentLatency,
+		double CursorBufferSize,
+		unsigned int CursorWidth,
+		unsigned int CursorBPP,
+		double VRatio,
+		double VRatioC,
+		double BytePerPixelInDETY,
+		double BytePerPixelInDETC,
+		double DETBufferSizeY,
+		double DETBufferSizeC,
+		double *UrgentBurstFactorCursor,
+		double *UrgentBurstFactorLuma,
+		double *UrgentBurstFactorChroma,
+		bool *NotEnoughUrgentLatencyHiding)
+{
+	double LinesInDETLuma;
+	double LinesInDETChroma;
+	unsigned int LinesInCursorBuffer;
+	double CursorBufferSizeInTime;
+	double DETBufferSizeInTimeLuma;
+	double DETBufferSizeInTimeChroma;
+
+	*NotEnoughUrgentLatencyHiding = 0;
+
+	if (CursorWidth > 0) {
+		LinesInCursorBuffer = 1 << (unsigned int) dml_floor(dml_log2(CursorBufferSize * 1024.0 / (CursorWidth * CursorBPP / 8.0)), 1.0);
+		if (VRatio > 0) {
+			CursorBufferSizeInTime = LinesInCursorBuffer * LineTime / VRatio;
+			if (CursorBufferSizeInTime - UrgentLatency <= 0) {
+				*NotEnoughUrgentLatencyHiding = 1;
+				*UrgentBurstFactorCursor = 0;
+			} else {
+				*UrgentBurstFactorCursor = CursorBufferSizeInTime / (CursorBufferSizeInTime - UrgentLatency);
+			}
+		} else {
+			*UrgentBurstFactorCursor = 1;
+		}
+	}
+
+	LinesInDETLuma = DETBufferSizeY / BytePerPixelInDETY / swath_width_luma_ub;
+	if (VRatio > 0) {
+		DETBufferSizeInTimeLuma = dml_floor(LinesInDETLuma, SwathHeightY) * LineTime / VRatio;
+		if (DETBufferSizeInTimeLuma - UrgentLatency <= 0) {
+			*NotEnoughUrgentLatencyHiding = 1;
+			*UrgentBurstFactorLuma = 0;
+		} else {
+			*UrgentBurstFactorLuma = DETBufferSizeInTimeLuma / (DETBufferSizeInTimeLuma - UrgentLatency);
+		}
+	} else {
+		*UrgentBurstFactorLuma = 1;
+	}
+
+	if (BytePerPixelInDETC > 0) {
+		LinesInDETChroma = DETBufferSizeC / BytePerPixelInDETC / swath_width_chroma_ub;
+		if (VRatio > 0) {
+			DETBufferSizeInTimeChroma = dml_floor(LinesInDETChroma, SwathHeightC) * LineTime / VRatio;
+			if (DETBufferSizeInTimeChroma - UrgentLatency <= 0) {
+				*NotEnoughUrgentLatencyHiding = 1;
+				*UrgentBurstFactorChroma = 0;
+			} else {
+				*UrgentBurstFactorChroma = DETBufferSizeInTimeChroma / (DETBufferSizeInTimeChroma - UrgentLatency);
+			}
+		} else {
+			*UrgentBurstFactorChroma = 1;
+		}
+	}
+}
+
+static void CalculatePixelDeliveryTimes(
+		unsigned int NumberOfActivePlanes,
+		double VRatio[],
+		double VRatioChroma[],
+		double VRatioPrefetchY[],
+		double VRatioPrefetchC[],
+		unsigned int swath_width_luma_ub[],
+		unsigned int swath_width_chroma_ub[],
+		unsigned int DPPPerPlane[],
+		double HRatio[],
+		double HRatioChroma[],
+		double PixelClock[],
+		double PSCL_THROUGHPUT[],
+		double PSCL_THROUGHPUT_CHROMA[],
+		double DPPCLK[],
+		int BytePerPixelC[],
+		enum scan_direction_class SourceScan[],
+		unsigned int NumberOfCursors[],
+		unsigned int CursorWidth[][DC__NUM_CURSOR__MAX],
+		unsigned int CursorBPP[][DC__NUM_CURSOR__MAX],
+		unsigned int BlockWidth256BytesY[],
+		unsigned int BlockHeight256BytesY[],
+		unsigned int BlockWidth256BytesC[],
+		unsigned int BlockHeight256BytesC[],
+		double DisplayPipeLineDeliveryTimeLuma[],
+		double DisplayPipeLineDeliveryTimeChroma[],
+		double DisplayPipeLineDeliveryTimeLumaPrefetch[],
+		double DisplayPipeLineDeliveryTimeChromaPrefetch[],
+		double DisplayPipeRequestDeliveryTimeLuma[],
+		double DisplayPipeRequestDeliveryTimeChroma[],
+		double DisplayPipeRequestDeliveryTimeLumaPrefetch[],
+		double DisplayPipeRequestDeliveryTimeChromaPrefetch[],
+		double CursorRequestDeliveryTime[],
+		double CursorRequestDeliveryTimePrefetch[])
+{
+	double req_per_swath_ub;
+	int k;
+
+	for (k = 0; k < NumberOfActivePlanes; ++k) {
+		if (VRatio[k] <= 1) {
+			DisplayPipeLineDeliveryTimeLuma[k] = swath_width_luma_ub[k] * DPPPerPlane[k] / HRatio[k] / PixelClock[k];
+		} else {
+			DisplayPipeLineDeliveryTimeLuma[k] = swath_width_luma_ub[k] / PSCL_THROUGHPUT[k] / DPPCLK[k];
+		}
+
+		if (BytePerPixelC[k] == 0) {
+			DisplayPipeLineDeliveryTimeChroma[k] = 0;
+		} else {
+			if (VRatioChroma[k] <= 1) {
+				DisplayPipeLineDeliveryTimeChroma[k] = swath_width_chroma_ub[k] * DPPPerPlane[k] / HRatioChroma[k] / PixelClock[k];
+			} else {
+				DisplayPipeLineDeliveryTimeChroma[k] = swath_width_chroma_ub[k] / PSCL_THROUGHPUT_CHROMA[k] / DPPCLK[k];
+			}
+		}
+
+		if (VRatioPrefetchY[k] <= 1) {
+			DisplayPipeLineDeliveryTimeLumaPrefetch[k] = swath_width_luma_ub[k] * DPPPerPlane[k] / HRatio[k] / PixelClock[k];
+		} else {
+			DisplayPipeLineDeliveryTimeLumaPrefetch[k] = swath_width_luma_ub[k] / PSCL_THROUGHPUT[k] / DPPCLK[k];
+		}
+
+		if (BytePerPixelC[k] == 0) {
+			DisplayPipeLineDeliveryTimeChromaPrefetch[k] = 0;
+		} else {
+			if (VRatioPrefetchC[k] <= 1) {
+				DisplayPipeLineDeliveryTimeChromaPrefetch[k] = swath_width_chroma_ub[k] * DPPPerPlane[k] / HRatioChroma[k] / PixelClock[k];
+			} else {
+				DisplayPipeLineDeliveryTimeChromaPrefetch[k] = swath_width_chroma_ub[k] / PSCL_THROUGHPUT_CHROMA[k] / DPPCLK[k];
+			}
+		}
+	}
+
+	for (k = 0; k < NumberOfActivePlanes; ++k) {
+		if (SourceScan[k] != dm_vert) {
+			req_per_swath_ub = swath_width_luma_ub[k] / BlockWidth256BytesY[k];
+		} else {
+			req_per_swath_ub = swath_width_luma_ub[k] / BlockHeight256BytesY[k];
+		}
+		DisplayPipeRequestDeliveryTimeLuma[k] = DisplayPipeLineDeliveryTimeLuma[k] / req_per_swath_ub;
+		DisplayPipeRequestDeliveryTimeLumaPrefetch[k] = DisplayPipeLineDeliveryTimeLumaPrefetch[k] / req_per_swath_ub;
+		if (BytePerPixelC[k] == 0) {
+			DisplayPipeRequestDeliveryTimeChroma[k] = 0;
+			DisplayPipeRequestDeliveryTimeChromaPrefetch[k] = 0;
+		} else {
+			if (SourceScan[k] != dm_vert) {
+				req_per_swath_ub = swath_width_chroma_ub[k] / BlockWidth256BytesC[k];
+			} else {
+				req_per_swath_ub = swath_width_chroma_ub[k] / BlockHeight256BytesC[k];
+			}
+			DisplayPipeRequestDeliveryTimeChroma[k] = DisplayPipeLineDeliveryTimeChroma[k] / req_per_swath_ub;
+			DisplayPipeRequestDeliveryTimeChromaPrefetch[k] = DisplayPipeLineDeliveryTimeChromaPrefetch[k] / req_per_swath_ub;
+		}
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%d : HRatio = %f\n", __func__, k, HRatio[k]);
+		dml_print("DML::%s: k=%d : VRatio = %f\n", __func__, k, VRatio[k]);
+		dml_print("DML::%s: k=%d : HRatioChroma = %f\n", __func__, k, HRatioChroma[k]);
+		dml_print("DML::%s: k=%d : VRatioChroma = %f\n", __func__, k, VRatioChroma[k]);
+		dml_print("DML::%s: k=%d : DisplayPipeLineDeliveryTimeLuma = %f\n", __func__, k, DisplayPipeLineDeliveryTimeLuma[k]);
+		dml_print("DML::%s: k=%d : DisplayPipeLineDeliveryTimeLumaPrefetch = %f\n", __func__, k, DisplayPipeLineDeliveryTimeLumaPrefetch[k]);
+		dml_print("DML::%s: k=%d : DisplayPipeLineDeliveryTimeChroma = %f\n", __func__, k, DisplayPipeLineDeliveryTimeChroma[k]);
+		dml_print("DML::%s: k=%d : DisplayPipeLineDeliveryTimeChromaPrefetch = %f\n", __func__, k, DisplayPipeLineDeliveryTimeChromaPrefetch[k]);
+		dml_print("DML::%s: k=%d : DisplayPipeRequestDeliveryTimeLuma = %f\n", __func__, k, DisplayPipeRequestDeliveryTimeLuma[k]);
+		dml_print("DML::%s: k=%d : DisplayPipeRequestDeliveryTimeLumaPrefetch = %f\n", __func__, k, DisplayPipeRequestDeliveryTimeLumaPrefetch[k]);
+		dml_print("DML::%s: k=%d : DisplayPipeRequestDeliveryTimeChroma = %f\n", __func__, k, DisplayPipeRequestDeliveryTimeChroma[k]);
+		dml_print("DML::%s: k=%d : DisplayPipeRequestDeliveryTimeChromaPrefetch = %f\n", __func__, k, DisplayPipeRequestDeliveryTimeChromaPrefetch[k]);
+#endif
+	}
+
+	for (k = 0; k < NumberOfActivePlanes; ++k) {
+		int cursor_req_per_width;
+
+		cursor_req_per_width = dml_ceil(CursorWidth[k][0] * CursorBPP[k][0] / 256 / 8, 1);
+		if (NumberOfCursors[k] > 0) {
+			if (VRatio[k] <= 1) {
+				CursorRequestDeliveryTime[k] = CursorWidth[k][0] / HRatio[k] / PixelClock[k] / cursor_req_per_width;
+			} else {
+				CursorRequestDeliveryTime[k] = CursorWidth[k][0] / PSCL_THROUGHPUT[k] / DPPCLK[k] / cursor_req_per_width;
+			}
+			if (VRatioPrefetchY[k] <= 1) {
+				CursorRequestDeliveryTimePrefetch[k] = CursorWidth[k][0] / HRatio[k] / PixelClock[k] / cursor_req_per_width;
+			} else {
+				CursorRequestDeliveryTimePrefetch[k] = CursorWidth[k][0] / PSCL_THROUGHPUT[k] / DPPCLK[k] / cursor_req_per_width;
+			}
+		} else {
+			CursorRequestDeliveryTime[k] = 0;
+			CursorRequestDeliveryTimePrefetch[k] = 0;
+		}
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%d : NumberOfCursors = %d\n", __func__, k, NumberOfCursors[k]);
+		dml_print("DML::%s: k=%d : CursorRequestDeliveryTime = %f\n", __func__, k, CursorRequestDeliveryTime[k]);
+		dml_print("DML::%s: k=%d : CursorRequestDeliveryTimePrefetch = %f\n", __func__, k, CursorRequestDeliveryTimePrefetch[k]);
+#endif
+	}
+}
+
+static void CalculateMetaAndPTETimes(
+		int NumberOfActivePlanes,
+		bool GPUVMEnable,
+		int MetaChunkSize,
+		int MinMetaChunkSizeBytes,
+		int HTotal[],
+		double VRatio[],
+		double VRatioChroma[],
+		double DestinationLinesToRequestRowInVBlank[],
+		double DestinationLinesToRequestRowInImmediateFlip[],
+		bool DCCEnable[],
+		double PixelClock[],
+		int BytePerPixelY[],
+		int BytePerPixelC[],
+		enum scan_direction_class SourceScan[],
+		int dpte_row_height[],
+		int dpte_row_height_chroma[],
+		int meta_row_width[],
+		int meta_row_width_chroma[],
+		int meta_row_height[],
+		int meta_row_height_chroma[],
+		int meta_req_width[],
+		int meta_req_width_chroma[],
+		int meta_req_height[],
+		int meta_req_height_chroma[],
+		int dpte_group_bytes[],
+		int PTERequestSizeY[],
+		int PTERequestSizeC[],
+		int PixelPTEReqWidthY[],
+		int PixelPTEReqHeightY[],
+		int PixelPTEReqWidthC[],
+		int PixelPTEReqHeightC[],
+		int dpte_row_width_luma_ub[],
+		int dpte_row_width_chroma_ub[],
+		double DST_Y_PER_PTE_ROW_NOM_L[],
+		double DST_Y_PER_PTE_ROW_NOM_C[],
+		double DST_Y_PER_META_ROW_NOM_L[],
+		double DST_Y_PER_META_ROW_NOM_C[],
+		double TimePerMetaChunkNominal[],
+		double TimePerChromaMetaChunkNominal[],
+		double TimePerMetaChunkVBlank[],
+		double TimePerChromaMetaChunkVBlank[],
+		double TimePerMetaChunkFlip[],
+		double TimePerChromaMetaChunkFlip[],
+		double time_per_pte_group_nom_luma[],
+		double time_per_pte_group_vblank_luma[],
+		double time_per_pte_group_flip_luma[],
+		double time_per_pte_group_nom_chroma[],
+		double time_per_pte_group_vblank_chroma[],
+		double time_per_pte_group_flip_chroma[])
+{
+	unsigned int meta_chunk_width;
+	unsigned int min_meta_chunk_width;
+	unsigned int meta_chunk_per_row_int;
+	unsigned int meta_row_remainder;
+	unsigned int meta_chunk_threshold;
+	unsigned int meta_chunks_per_row_ub;
+	unsigned int meta_chunk_width_chroma;
+	unsigned int min_meta_chunk_width_chroma;
+	unsigned int meta_chunk_per_row_int_chroma;
+	unsigned int meta_row_remainder_chroma;
+	unsigned int meta_chunk_threshold_chroma;
+	unsigned int meta_chunks_per_row_ub_chroma;
+	unsigned int dpte_group_width_luma;
+	unsigned int dpte_groups_per_row_luma_ub;
+	unsigned int dpte_group_width_chroma;
+	unsigned int dpte_groups_per_row_chroma_ub;
+	int k;
+
+	for (k = 0; k < NumberOfActivePlanes; ++k) {
+		DST_Y_PER_PTE_ROW_NOM_L[k] = dpte_row_height[k] / VRatio[k];
+		if (BytePerPixelC[k] == 0) {
+			DST_Y_PER_PTE_ROW_NOM_C[k] = 0;
+		} else {
+			DST_Y_PER_PTE_ROW_NOM_C[k] = dpte_row_height_chroma[k] / VRatioChroma[k];
+		}
+		DST_Y_PER_META_ROW_NOM_L[k] = meta_row_height[k] / VRatio[k];
+		if (BytePerPixelC[k] == 0) {
+			DST_Y_PER_META_ROW_NOM_C[k] = 0;
+		} else {
+			DST_Y_PER_META_ROW_NOM_C[k] = meta_row_height_chroma[k] / VRatioChroma[k];
+		}
+	}
+
+	for (k = 0; k < NumberOfActivePlanes; ++k) {
+		if (DCCEnable[k] == true) {
+			meta_chunk_width = MetaChunkSize * 1024 * 256 / BytePerPixelY[k] / meta_row_height[k];
+			min_meta_chunk_width = MinMetaChunkSizeBytes * 256 / BytePerPixelY[k] / meta_row_height[k];
+			meta_chunk_per_row_int = meta_row_width[k] / meta_chunk_width;
+			meta_row_remainder = meta_row_width[k] % meta_chunk_width;
+			if (SourceScan[k] != dm_vert) {
+				meta_chunk_threshold = 2 * min_meta_chunk_width - meta_req_width[k];
+			} else {
+				meta_chunk_threshold = 2 * min_meta_chunk_width - meta_req_height[k];
+			}
+			if (meta_row_remainder <= meta_chunk_threshold) {
+				meta_chunks_per_row_ub = meta_chunk_per_row_int + 1;
+			} else {
+				meta_chunks_per_row_ub = meta_chunk_per_row_int + 2;
+			}
+			TimePerMetaChunkNominal[k] = meta_row_height[k] / VRatio[k] * HTotal[k] / PixelClock[k] / meta_chunks_per_row_ub;
+			TimePerMetaChunkVBlank[k] = DestinationLinesToRequestRowInVBlank[k] * HTotal[k] / PixelClock[k] / meta_chunks_per_row_ub;
+			TimePerMetaChunkFlip[k] = DestinationLinesToRequestRowInImmediateFlip[k] * HTotal[k] / PixelClock[k] / meta_chunks_per_row_ub;
+			if (BytePerPixelC[k] == 0) {
+				TimePerChromaMetaChunkNominal[k] = 0;
+				TimePerChromaMetaChunkVBlank[k] = 0;
+				TimePerChromaMetaChunkFlip[k] = 0;
+			} else {
+				meta_chunk_width_chroma = MetaChunkSize * 1024 * 256 / BytePerPixelC[k] / meta_row_height_chroma[k];
+				min_meta_chunk_width_chroma = MinMetaChunkSizeBytes * 256 / BytePerPixelC[k] / meta_row_height_chroma[k];
+				meta_chunk_per_row_int_chroma = (double) meta_row_width_chroma[k] / meta_chunk_width_chroma;
+				meta_row_remainder_chroma = meta_row_width_chroma[k] % meta_chunk_width_chroma;
+				if (SourceScan[k] != dm_vert) {
+					meta_chunk_threshold_chroma = 2 * min_meta_chunk_width_chroma - meta_req_width_chroma[k];
+				} else {
+					meta_chunk_threshold_chroma = 2 * min_meta_chunk_width_chroma - meta_req_height_chroma[k];
+				}
+				if (meta_row_remainder_chroma <= meta_chunk_threshold_chroma) {
+					meta_chunks_per_row_ub_chroma = meta_chunk_per_row_int_chroma + 1;
+				} else {
+					meta_chunks_per_row_ub_chroma = meta_chunk_per_row_int_chroma + 2;
+				}
+				TimePerChromaMetaChunkNominal[k] = meta_row_height_chroma[k] / VRatioChroma[k] * HTotal[k] / PixelClock[k] / meta_chunks_per_row_ub_chroma;
+				TimePerChromaMetaChunkVBlank[k] = DestinationLinesToRequestRowInVBlank[k] * HTotal[k] / PixelClock[k] / meta_chunks_per_row_ub_chroma;
+				TimePerChromaMetaChunkFlip[k] = DestinationLinesToRequestRowInImmediateFlip[k] * HTotal[k] / PixelClock[k] / meta_chunks_per_row_ub_chroma;
+			}
+		} else {
+			TimePerMetaChunkNominal[k] = 0;
+			TimePerMetaChunkVBlank[k] = 0;
+			TimePerMetaChunkFlip[k] = 0;
+			TimePerChromaMetaChunkNominal[k] = 0;
+			TimePerChromaMetaChunkVBlank[k] = 0;
+			TimePerChromaMetaChunkFlip[k] = 0;
+		}
+	}
+
+	for (k = 0; k < NumberOfActivePlanes; ++k) {
+		if (GPUVMEnable == true) {
+			if (SourceScan[k] != dm_vert) {
+				dpte_group_width_luma = dpte_group_bytes[k] / PTERequestSizeY[k] * PixelPTEReqWidthY[k];
+			} else {
+				dpte_group_width_luma = dpte_group_bytes[k] / PTERequestSizeY[k] * PixelPTEReqHeightY[k];
+			}
+			dpte_groups_per_row_luma_ub = dml_ceil(1.0 * dpte_row_width_luma_ub[k] / dpte_group_width_luma, 1);
+			time_per_pte_group_nom_luma[k] = DST_Y_PER_PTE_ROW_NOM_L[k] * HTotal[k] / PixelClock[k] / dpte_groups_per_row_luma_ub;
+			time_per_pte_group_vblank_luma[k] = DestinationLinesToRequestRowInVBlank[k] * HTotal[k] / PixelClock[k] / dpte_groups_per_row_luma_ub;
+			time_per_pte_group_flip_luma[k] = DestinationLinesToRequestRowInImmediateFlip[k] * HTotal[k] / PixelClock[k] / dpte_groups_per_row_luma_ub;
+			if (BytePerPixelC[k] == 0) {
+				time_per_pte_group_nom_chroma[k] = 0;
+				time_per_pte_group_vblank_chroma[k] = 0;
+				time_per_pte_group_flip_chroma[k] = 0;
+			} else {
+				if (SourceScan[k] != dm_vert) {
+					dpte_group_width_chroma = dpte_group_bytes[k] / PTERequestSizeC[k] * PixelPTEReqWidthC[k];
+				} else {
+					dpte_group_width_chroma = dpte_group_bytes[k] / PTERequestSizeC[k] * PixelPTEReqHeightC[k];
+				}
+				dpte_groups_per_row_chroma_ub = dml_ceil(1.0 * dpte_row_width_chroma_ub[k] / dpte_group_width_chroma, 1);
+				time_per_pte_group_nom_chroma[k] = DST_Y_PER_PTE_ROW_NOM_C[k] * HTotal[k] / PixelClock[k] / dpte_groups_per_row_chroma_ub;
+				time_per_pte_group_vblank_chroma[k] = DestinationLinesToRequestRowInVBlank[k] * HTotal[k] / PixelClock[k] / dpte_groups_per_row_chroma_ub;
+				time_per_pte_group_flip_chroma[k] = DestinationLinesToRequestRowInImmediateFlip[k] * HTotal[k] / PixelClock[k] / dpte_groups_per_row_chroma_ub;
+			}
+		} else {
+			time_per_pte_group_nom_luma[k] = 0;
+			time_per_pte_group_vblank_luma[k] = 0;
+			time_per_pte_group_flip_luma[k] = 0;
+			time_per_pte_group_nom_chroma[k] = 0;
+			time_per_pte_group_vblank_chroma[k] = 0;
+			time_per_pte_group_flip_chroma[k] = 0;
+		}
+	}
+}
+
+static void CalculateVMGroupAndRequestTimes(
+		unsigned int NumberOfActivePlanes,
+		bool GPUVMEnable,
+		unsigned int GPUVMMaxPageTableLevels,
+		unsigned int HTotal[],
+		int BytePerPixelC[],
+		double DestinationLinesToRequestVMInVBlank[],
+		double DestinationLinesToRequestVMInImmediateFlip[],
+		bool DCCEnable[],
+		double PixelClock[],
+		int dpte_row_width_luma_ub[],
+		int dpte_row_width_chroma_ub[],
+		int vm_group_bytes[],
+		unsigned int dpde0_bytes_per_frame_ub_l[],
+		unsigned int dpde0_bytes_per_frame_ub_c[],
+		int meta_pte_bytes_per_frame_ub_l[],
+		int meta_pte_bytes_per_frame_ub_c[],
+		double TimePerVMGroupVBlank[],
+		double TimePerVMGroupFlip[],
+		double TimePerVMRequestVBlank[],
+		double TimePerVMRequestFlip[])
+{
+	int num_group_per_lower_vm_stage;
+	int num_req_per_lower_vm_stage;
+	int k;
+
+	for (k = 0; k < NumberOfActivePlanes; ++k) {
+		if (GPUVMEnable == true && (DCCEnable[k] == true || GPUVMMaxPageTableLevels > 1)) {
+			if (DCCEnable[k] == false) {
+				if (BytePerPixelC[k] > 0) {
+					num_group_per_lower_vm_stage = dml_ceil((double) (dpde0_bytes_per_frame_ub_l[k]) / (double) (vm_group_bytes[k]), 1)
+							+ dml_ceil((double) (dpde0_bytes_per_frame_ub_c[k]) / (double) (vm_group_bytes[k]), 1);
+				} else {
+					num_group_per_lower_vm_stage = dml_ceil((double) (dpde0_bytes_per_frame_ub_l[k]) / (double) (vm_group_bytes[k]), 1);
+				}
+			} else {
+				if (GPUVMMaxPageTableLevels == 1) {
+					if (BytePerPixelC[k] > 0) {
+						num_group_per_lower_vm_stage = dml_ceil((double) (meta_pte_bytes_per_frame_ub_l[k]) / (double) (vm_group_bytes[k]), 1)
+								+ dml_ceil((double) (meta_pte_bytes_per_frame_ub_c[k]) / (double) (vm_group_bytes[k]), 1);
+					} else {
+						num_group_per_lower_vm_stage = dml_ceil((double) (meta_pte_bytes_per_frame_ub_l[k]) / (double) (vm_group_bytes[k]), 1);
+					}
+				} else {
+					if (BytePerPixelC[k] > 0) {
+						num_group_per_lower_vm_stage = 2 + dml_ceil((double) (dpde0_bytes_per_frame_ub_l[k]) / (double) (vm_group_bytes[k]), 1)
+								+ dml_ceil((double) (dpde0_bytes_per_frame_ub_c[k]) / (double) (vm_group_bytes[k]), 1)
+								+ dml_ceil((double) (meta_pte_bytes_per_frame_ub_l[k]) / (double) (vm_group_bytes[k]), 1)
+								+ dml_ceil((double) (meta_pte_bytes_per_frame_ub_c[k]) / (double) (vm_group_bytes[k]), 1);
+					} else {
+						num_group_per_lower_vm_stage = 1 + dml_ceil((double) (dpde0_bytes_per_frame_ub_l[k]) / (double) (vm_group_bytes[k]), 1)
+								+ dml_ceil((double) (meta_pte_bytes_per_frame_ub_l[k]) / (double) (vm_group_bytes[k]), 1);
+					}
+				}
+			}
+
+			if (DCCEnable[k] == false) {
+				if (BytePerPixelC[k] > 0) {
+					num_req_per_lower_vm_stage = dpde0_bytes_per_frame_ub_l[k] / 64 + dpde0_bytes_per_frame_ub_c[k] / 64;
+				} else {
+					num_req_per_lower_vm_stage = dpde0_bytes_per_frame_ub_l[k] / 64;
+				}
+			} else {
+				if (GPUVMMaxPageTableLevels == 1) {
+					if (BytePerPixelC[k] > 0) {
+						num_req_per_lower_vm_stage = meta_pte_bytes_per_frame_ub_l[k] / 64 + meta_pte_bytes_per_frame_ub_c[k] / 64;
+					} else {
+						num_req_per_lower_vm_stage = meta_pte_bytes_per_frame_ub_l[k] / 64;
+					}
+				} else {
+					if (BytePerPixelC[k] > 0) {
+						num_req_per_lower_vm_stage = dpde0_bytes_per_frame_ub_l[k] / 64 + dpde0_bytes_per_frame_ub_c[k] / 64
+								+ meta_pte_bytes_per_frame_ub_l[k] / 64 + meta_pte_bytes_per_frame_ub_c[k] / 64;
+					} else {
+						num_req_per_lower_vm_stage = dpde0_bytes_per_frame_ub_l[k] / 64 + meta_pte_bytes_per_frame_ub_l[k] / 64;
+					}
+				}
+			}
+
+			TimePerVMGroupVBlank[k] = DestinationLinesToRequestVMInVBlank[k] * HTotal[k] / PixelClock[k] / num_group_per_lower_vm_stage;
+			TimePerVMGroupFlip[k] = DestinationLinesToRequestVMInImmediateFlip[k] * HTotal[k] / PixelClock[k] / num_group_per_lower_vm_stage;
+			TimePerVMRequestVBlank[k] = DestinationLinesToRequestVMInVBlank[k] * HTotal[k] / PixelClock[k] / num_req_per_lower_vm_stage;
+			TimePerVMRequestFlip[k] = DestinationLinesToRequestVMInImmediateFlip[k] * HTotal[k] / PixelClock[k] / num_req_per_lower_vm_stage;
+
+			if (GPUVMMaxPageTableLevels > 2) {
+				TimePerVMGroupVBlank[k] = TimePerVMGroupVBlank[k] / 2;
+				TimePerVMGroupFlip[k] = TimePerVMGroupFlip[k] / 2;
+				TimePerVMRequestVBlank[k] = TimePerVMRequestVBlank[k] / 2;
+				TimePerVMRequestFlip[k] = TimePerVMRequestFlip[k] / 2;
+			}
+
+		} else {
+			TimePerVMGroupVBlank[k] = 0;
+			TimePerVMGroupFlip[k] = 0;
+			TimePerVMRequestVBlank[k] = 0;
+			TimePerVMRequestFlip[k] = 0;
+		}
+	}
+}
+
+static void CalculateStutterEfficiency(
+		struct display_mode_lib *mode_lib,
+		int CompressedBufferSizeInkByte,
+		bool UnboundedRequestEnabled,
+		int ConfigReturnBufferSizeInKByte,
+		int MetaFIFOSizeInKEntries,
+		int ZeroSizeBufferEntries,
+		int NumberOfActivePlanes,
+		int ROBBufferSizeInKByte,
+		double TotalDataReadBandwidth,
+		double DCFCLK,
+		double ReturnBW,
+		double COMPBUF_RESERVED_SPACE_64B,
+		double COMPBUF_RESERVED_SPACE_ZS,
+		double SRExitTime,
+		double SRExitZ8Time,
+		bool SynchronizedVBlank,
+		double Z8StutterEnterPlusExitWatermark,
+		double StutterEnterPlusExitWatermark,
+		bool ProgressiveToInterlaceUnitInOPP,
+		bool Interlace[],
+		double MinTTUVBlank[],
+		int DPPPerPlane[],
+		unsigned int DETBufferSizeY[],
+		int BytePerPixelY[],
+		double BytePerPixelDETY[],
+		double SwathWidthY[],
+		int SwathHeightY[],
+		int SwathHeightC[],
+		double NetDCCRateLuma[],
+		double NetDCCRateChroma[],
+		double DCCFractionOfZeroSizeRequestsLuma[],
+		double DCCFractionOfZeroSizeRequestsChroma[],
+		int HTotal[],
+		int VTotal[],
+		double PixelClock[],
+		double VRatio[],
+		enum scan_direction_class SourceScan[],
+		int BlockHeight256BytesY[],
+		int BlockWidth256BytesY[],
+		int BlockHeight256BytesC[],
+		int BlockWidth256BytesC[],
+		int DCCYMaxUncompressedBlock[],
+		int DCCCMaxUncompressedBlock[],
+		int VActive[],
+		bool DCCEnable[],
+		bool WritebackEnable[],
+		double ReadBandwidthPlaneLuma[],
+		double ReadBandwidthPlaneChroma[],
+		double meta_row_bw[],
+		double dpte_row_bw[],
+		double *StutterEfficiencyNotIncludingVBlank,
+		double *StutterEfficiency,
+		int *NumberOfStutterBurstsPerFrame,
+		double *Z8StutterEfficiencyNotIncludingVBlank,
+		double *Z8StutterEfficiency,
+		int *Z8NumberOfStutterBurstsPerFrame,
+		double *StutterPeriod)
+{
+	struct vba_vars_st *v = &mode_lib->vba;
+
+	double DETBufferingTimeY;
+	double SwathWidthYCriticalPlane = 0;
+	double VActiveTimeCriticalPlane = 0;
+	double FrameTimeCriticalPlane = 0;
+	int BytePerPixelYCriticalPlane = 0;
+	double LinesToFinishSwathTransferStutterCriticalPlane = 0;
+	double MinTTUVBlankCriticalPlane = 0;
+	double TotalCompressedReadBandwidth;
+	double TotalRowReadBandwidth;
+	double AverageDCCCompressionRate;
+	double EffectiveCompressedBufferSize;
+	double PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer;
+	double StutterBurstTime;
+	int TotalActiveWriteback;
+	double LinesInDETY;
+	double LinesInDETYRoundedDownToSwath;
+	double MaximumEffectiveCompressionLuma;
+	double MaximumEffectiveCompressionChroma;
+	double TotalZeroSizeRequestReadBandwidth;
+	double TotalZeroSizeCompressedReadBandwidth;
+	double AverageDCCZeroSizeFraction;
+	double AverageZeroSizeCompressionRate;
+	int TotalNumberOfActiveOTG = 0;
+	double LastStutterPeriod = 0.0;
+	double LastZ8StutterPeriod = 0.0;
+	int k;
+
+	TotalZeroSizeRequestReadBandwidth = 0;
+	TotalZeroSizeCompressedReadBandwidth = 0;
+	TotalRowReadBandwidth = 0;
+	TotalCompressedReadBandwidth = 0;
+
+	for (k = 0; k < NumberOfActivePlanes; ++k) {
+		if (DCCEnable[k] == true) {
+			if ((SourceScan[k] == dm_vert && BlockWidth256BytesY[k] > SwathHeightY[k]) || (SourceScan[k] != dm_vert && BlockHeight256BytesY[k] > SwathHeightY[k])
+					|| DCCYMaxUncompressedBlock[k] < 256) {
+				MaximumEffectiveCompressionLuma = 2;
+			} else {
+				MaximumEffectiveCompressionLuma = 4;
+			}
+			TotalCompressedReadBandwidth = TotalCompressedReadBandwidth + ReadBandwidthPlaneLuma[k] / dml_min(NetDCCRateLuma[k], MaximumEffectiveCompressionLuma);
+			TotalZeroSizeRequestReadBandwidth = TotalZeroSizeRequestReadBandwidth + ReadBandwidthPlaneLuma[k] * DCCFractionOfZeroSizeRequestsLuma[k];
+			TotalZeroSizeCompressedReadBandwidth = TotalZeroSizeCompressedReadBandwidth
+					+ ReadBandwidthPlaneLuma[k] * DCCFractionOfZeroSizeRequestsLuma[k] / MaximumEffectiveCompressionLuma;
+			if (ReadBandwidthPlaneChroma[k] > 0) {
+				if ((SourceScan[k] == dm_vert && BlockWidth256BytesC[k] > SwathHeightC[k])
+						|| (SourceScan[k] != dm_vert && BlockHeight256BytesC[k] > SwathHeightC[k]) || DCCCMaxUncompressedBlock[k] < 256) {
+					MaximumEffectiveCompressionChroma = 2;
+				} else {
+					MaximumEffectiveCompressionChroma = 4;
+				}
+				TotalCompressedReadBandwidth = TotalCompressedReadBandwidth
+						+ ReadBandwidthPlaneChroma[k] / dml_min(NetDCCRateChroma[k], MaximumEffectiveCompressionChroma);
+				TotalZeroSizeRequestReadBandwidth = TotalZeroSizeRequestReadBandwidth + ReadBandwidthPlaneChroma[k] * DCCFractionOfZeroSizeRequestsChroma[k];
+				TotalZeroSizeCompressedReadBandwidth = TotalZeroSizeCompressedReadBandwidth
+						+ ReadBandwidthPlaneChroma[k] * DCCFractionOfZeroSizeRequestsChroma[k] / MaximumEffectiveCompressionChroma;
+			}
+		} else {
+			TotalCompressedReadBandwidth = TotalCompressedReadBandwidth + ReadBandwidthPlaneLuma[k] + ReadBandwidthPlaneChroma[k];
+		}
+		TotalRowReadBandwidth = TotalRowReadBandwidth + DPPPerPlane[k] * (meta_row_bw[k] + dpte_row_bw[k]);
+	}
+
+	AverageDCCCompressionRate = TotalDataReadBandwidth / TotalCompressedReadBandwidth;
+	AverageDCCZeroSizeFraction = TotalZeroSizeRequestReadBandwidth / TotalDataReadBandwidth;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: TotalCompressedReadBandwidth = %f\n", __func__, TotalCompressedReadBandwidth);
+	dml_print("DML::%s: TotalZeroSizeRequestReadBandwidth = %f\n", __func__, TotalZeroSizeRequestReadBandwidth);
+	dml_print("DML::%s: TotalZeroSizeCompressedReadBandwidth = %f\n", __func__, TotalZeroSizeCompressedReadBandwidth);
+	dml_print("DML::%s: MaximumEffectiveCompressionLuma = %f\n", __func__, MaximumEffectiveCompressionLuma);
+	dml_print("DML::%s: MaximumEffectiveCompressionChroma = %f\n", __func__, MaximumEffectiveCompressionChroma);
+	dml_print("DML::%s: AverageDCCCompressionRate = %f\n", __func__, AverageDCCCompressionRate);
+	dml_print("DML::%s: AverageDCCZeroSizeFraction = %f\n", __func__, AverageDCCZeroSizeFraction);
+	dml_print("DML::%s: CompressedBufferSizeInkByte = %d\n", __func__, CompressedBufferSizeInkByte);
+#endif
+
+	if (AverageDCCZeroSizeFraction == 1) {
+		AverageZeroSizeCompressionRate = TotalZeroSizeRequestReadBandwidth / TotalZeroSizeCompressedReadBandwidth;
+		EffectiveCompressedBufferSize = MetaFIFOSizeInKEntries * 1024 * 64 * AverageZeroSizeCompressionRate + (ZeroSizeBufferEntries - COMPBUF_RESERVED_SPACE_ZS) * 64 * AverageZeroSizeCompressionRate;
+	} else if (AverageDCCZeroSizeFraction > 0) {
+		AverageZeroSizeCompressionRate = TotalZeroSizeRequestReadBandwidth / TotalZeroSizeCompressedReadBandwidth;
+		EffectiveCompressedBufferSize = dml_min(
+				CompressedBufferSizeInkByte * 1024 * AverageDCCCompressionRate,
+				MetaFIFOSizeInKEntries * 1024 * 64 / (AverageDCCZeroSizeFraction / AverageZeroSizeCompressionRate + 1 / AverageDCCCompressionRate))
+					+ dml_min((ROBBufferSizeInKByte * 1024 - COMPBUF_RESERVED_SPACE_64B * 64) * AverageDCCCompressionRate,
+						(ZeroSizeBufferEntries - COMPBUF_RESERVED_SPACE_ZS) * 64 / (AverageDCCZeroSizeFraction / AverageZeroSizeCompressionRate));
+		dml_print("DML::%s: min 1 = %f\n", __func__, CompressedBufferSizeInkByte * 1024 * AverageDCCCompressionRate);
+		dml_print(
+				"DML::%s: min 2 = %f\n",
+				__func__,
+				MetaFIFOSizeInKEntries * 1024 * 64 / (AverageDCCZeroSizeFraction / AverageZeroSizeCompressionRate + 1 / AverageDCCCompressionRate));
+		dml_print("DML::%s: min 3 = %f\n", __func__, ROBBufferSizeInKByte * 1024 * AverageDCCCompressionRate);
+		dml_print("DML::%s: min 4 = %f\n", __func__, ZeroSizeBufferEntries * 64 / (AverageDCCZeroSizeFraction / AverageZeroSizeCompressionRate));
+	} else {
+		EffectiveCompressedBufferSize = dml_min(
+				CompressedBufferSizeInkByte * 1024 * AverageDCCCompressionRate,
+				MetaFIFOSizeInKEntries * 1024 * 64 * AverageDCCCompressionRate) + (ROBBufferSizeInKByte * 1024 - COMPBUF_RESERVED_SPACE_64B * 64) * AverageDCCCompressionRate;
+		dml_print("DML::%s: min 1 = %f\n", __func__, CompressedBufferSizeInkByte * 1024 * AverageDCCCompressionRate);
+		dml_print("DML::%s: min 2 = %f\n", __func__, MetaFIFOSizeInKEntries * 1024 * 64 * AverageDCCCompressionRate);
+	}
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: MetaFIFOSizeInKEntries = %d\n", __func__, MetaFIFOSizeInKEntries);
+	dml_print("DML::%s: AverageZeroSizeCompressionRate = %f\n", __func__, AverageZeroSizeCompressionRate);
+	dml_print("DML::%s: EffectiveCompressedBufferSize = %f\n", __func__, EffectiveCompressedBufferSize);
+#endif
+
+	*StutterPeriod = 0;
+	for (k = 0; k < NumberOfActivePlanes; ++k) {
+		LinesInDETY = (DETBufferSizeY[k] + (UnboundedRequestEnabled == true ? EffectiveCompressedBufferSize : 0) * ReadBandwidthPlaneLuma[k] / TotalDataReadBandwidth)
+				/ BytePerPixelDETY[k] / SwathWidthY[k];
+		LinesInDETYRoundedDownToSwath = dml_floor(LinesInDETY, SwathHeightY[k]);
+		DETBufferingTimeY = LinesInDETYRoundedDownToSwath * (HTotal[k] / PixelClock[k]) / VRatio[k];
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%0d DETBufferSizeY = %f\n", __func__, k, DETBufferSizeY[k]);
+		dml_print("DML::%s: k=%0d BytePerPixelDETY = %f\n", __func__, k, BytePerPixelDETY[k]);
+		dml_print("DML::%s: k=%0d SwathWidthY = %f\n", __func__, k, SwathWidthY[k]);
+		dml_print("DML::%s: k=%0d ReadBandwidthPlaneLuma = %f\n", __func__, k, ReadBandwidthPlaneLuma[k]);
+		dml_print("DML::%s: k=%0d TotalDataReadBandwidth = %f\n", __func__, k, TotalDataReadBandwidth);
+		dml_print("DML::%s: k=%0d LinesInDETY = %f\n", __func__, k, LinesInDETY);
+		dml_print("DML::%s: k=%0d LinesInDETYRoundedDownToSwath = %f\n", __func__, k, LinesInDETYRoundedDownToSwath);
+		dml_print("DML::%s: k=%0d HTotal = %d\n", __func__, k, HTotal[k]);
+		dml_print("DML::%s: k=%0d PixelClock = %f\n", __func__, k, PixelClock[k]);
+		dml_print("DML::%s: k=%0d VRatio = %f\n", __func__, k, VRatio[k]);
+		dml_print("DML::%s: k=%0d DETBufferingTimeY = %f\n", __func__, k, DETBufferingTimeY);
+		dml_print("DML::%s: k=%0d PixelClock = %f\n", __func__, k, PixelClock[k]);
+#endif
+
+		if (k == 0 || DETBufferingTimeY < *StutterPeriod) {
+			bool isInterlaceTiming = Interlace[k] && !ProgressiveToInterlaceUnitInOPP;
+
+			*StutterPeriod = DETBufferingTimeY;
+			FrameTimeCriticalPlane = (isInterlaceTiming ? dml_floor(VTotal[k] / 2.0, 1.0) : VTotal[k]) * HTotal[k] / PixelClock[k];
+			VActiveTimeCriticalPlane = (isInterlaceTiming ? dml_floor(VActive[k] / 2.0, 1.0) : VActive[k]) * HTotal[k] / PixelClock[k];
+			BytePerPixelYCriticalPlane = BytePerPixelY[k];
+			SwathWidthYCriticalPlane = SwathWidthY[k];
+			LinesToFinishSwathTransferStutterCriticalPlane = SwathHeightY[k] - (LinesInDETY - LinesInDETYRoundedDownToSwath);
+			MinTTUVBlankCriticalPlane = MinTTUVBlank[k];
+
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: StutterPeriod = %f\n", __func__, *StutterPeriod);
+			dml_print("DML::%s: MinTTUVBlankCriticalPlane = %f\n", __func__, MinTTUVBlankCriticalPlane);
+			dml_print("DML::%s: FrameTimeCriticalPlane = %f\n", __func__, FrameTimeCriticalPlane);
+			dml_print("DML::%s: VActiveTimeCriticalPlane = %f\n", __func__, VActiveTimeCriticalPlane);
+			dml_print("DML::%s: BytePerPixelYCriticalPlane = %d\n", __func__, BytePerPixelYCriticalPlane);
+			dml_print("DML::%s: SwathWidthYCriticalPlane = %f\n", __func__, SwathWidthYCriticalPlane);
+			dml_print("DML::%s: LinesToFinishSwathTransferStutterCriticalPlane = %f\n", __func__, LinesToFinishSwathTransferStutterCriticalPlane);
+#endif
+		}
+	}
+
+	PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer = dml_min(*StutterPeriod * TotalDataReadBandwidth, EffectiveCompressedBufferSize);
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: ROBBufferSizeInKByte = %d\n", __func__, ROBBufferSizeInKByte);
+	dml_print("DML::%s: AverageDCCCompressionRate = %f\n", __func__, AverageDCCCompressionRate);
+	dml_print("DML::%s: StutterPeriod * TotalDataReadBandwidth = %f\n", __func__, *StutterPeriod * TotalDataReadBandwidth);
+	dml_print("DML::%s: ROBBufferSizeInKByte * 1024 * AverageDCCCompressionRate + EffectiveCompressedBufferSize = %f\n", __func__, ROBBufferSizeInKByte * 1024 * AverageDCCCompressionRate + EffectiveCompressedBufferSize);
+	dml_print("DML::%s: EffectiveCompressedBufferSize = %f\n", __func__, EffectiveCompressedBufferSize);
+	dml_print("DML::%s: PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer = %f\n", __func__, PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer);
+	dml_print("DML::%s: ReturnBW = %f\n", __func__, ReturnBW);
+	dml_print("DML::%s: TotalDataReadBandwidth = %f\n", __func__, TotalDataReadBandwidth);
+	dml_print("DML::%s: TotalRowReadBandwidth = %f\n", __func__, TotalRowReadBandwidth);
+	dml_print("DML::%s: DCFCLK = %f\n", __func__, DCFCLK);
+#endif
+
+	StutterBurstTime = PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer / AverageDCCCompressionRate / ReturnBW
+			+ (*StutterPeriod * TotalDataReadBandwidth - PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer) / (DCFCLK * 64)
+			+ *StutterPeriod * TotalRowReadBandwidth / ReturnBW;
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: Part 1 = %f\n", __func__, PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer / AverageDCCCompressionRate / ReturnBW);
+	dml_print("DML::%s: StutterPeriod * TotalDataReadBandwidth = %f\n", __func__, (*StutterPeriod * TotalDataReadBandwidth));
+	dml_print("DML::%s: Part 2 = %f\n", __func__, (*StutterPeriod * TotalDataReadBandwidth - PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer) / (DCFCLK * 64));
+	dml_print("DML::%s: Part 3 = %f\n", __func__, *StutterPeriod * TotalRowReadBandwidth / ReturnBW);
+	dml_print("DML::%s: StutterBurstTime = %f\n", __func__, StutterBurstTime);
+#endif
+	StutterBurstTime = dml_max(StutterBurstTime, LinesToFinishSwathTransferStutterCriticalPlane * BytePerPixelYCriticalPlane * SwathWidthYCriticalPlane / ReturnBW);
+
+	dml_print(
+			"DML::%s: Time to finish residue swath=%f\n",
+			__func__,
+			LinesToFinishSwathTransferStutterCriticalPlane * BytePerPixelYCriticalPlane * SwathWidthYCriticalPlane / ReturnBW);
+
+	TotalActiveWriteback = 0;
+	for (k = 0; k < NumberOfActivePlanes; ++k) {
+		if (WritebackEnable[k]) {
+			TotalActiveWriteback = TotalActiveWriteback + 1;
+		}
+	}
+
+	if (TotalActiveWriteback == 0) {
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: SRExitTime = %f\n", __func__, SRExitTime);
+		dml_print("DML::%s: SRExitZ8Time = %f\n", __func__, SRExitZ8Time);
+		dml_print("DML::%s: StutterBurstTime = %f (final)\n", __func__, StutterBurstTime);
+		dml_print("DML::%s: StutterPeriod = %f\n", __func__, *StutterPeriod);
+#endif
+		*StutterEfficiencyNotIncludingVBlank = dml_max(0., 1 - (SRExitTime + StutterBurstTime) / *StutterPeriod) * 100;
+		*Z8StutterEfficiencyNotIncludingVBlank = dml_max(0., 1 - (SRExitZ8Time + StutterBurstTime) / *StutterPeriod) * 100;
+		*NumberOfStutterBurstsPerFrame = (*StutterEfficiencyNotIncludingVBlank > 0 ? dml_ceil(VActiveTimeCriticalPlane / *StutterPeriod, 1) : 0);
+		*Z8NumberOfStutterBurstsPerFrame = (*Z8StutterEfficiencyNotIncludingVBlank > 0 ? dml_ceil(VActiveTimeCriticalPlane / *StutterPeriod, 1) : 0);
+	} else {
+		*StutterEfficiencyNotIncludingVBlank = 0.;
+		*Z8StutterEfficiencyNotIncludingVBlank = 0.;
+		*NumberOfStutterBurstsPerFrame = 0;
+		*Z8NumberOfStutterBurstsPerFrame = 0;
+	}
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: VActiveTimeCriticalPlane = %f\n", __func__, VActiveTimeCriticalPlane);
+	dml_print("DML::%s: StutterEfficiencyNotIncludingVBlank = %f\n", __func__, *StutterEfficiencyNotIncludingVBlank);
+	dml_print("DML::%s: Z8StutterEfficiencyNotIncludingVBlank = %f\n", __func__, *Z8StutterEfficiencyNotIncludingVBlank);
+	dml_print("DML::%s: NumberOfStutterBurstsPerFrame = %d\n", __func__, *NumberOfStutterBurstsPerFrame);
+	dml_print("DML::%s: Z8NumberOfStutterBurstsPerFrame = %d\n", __func__, *Z8NumberOfStutterBurstsPerFrame);
+#endif
+
+	for (k = 0; k < NumberOfActivePlanes; ++k) {
+		if (v->BlendingAndTiming[k] == k) {
+			TotalNumberOfActiveOTG = TotalNumberOfActiveOTG + 1;
+		}
+	}
+
+	if (*StutterEfficiencyNotIncludingVBlank > 0) {
+		LastStutterPeriod = VActiveTimeCriticalPlane - (*NumberOfStutterBurstsPerFrame - 1) * *StutterPeriod;
+
+		if ((SynchronizedVBlank || TotalNumberOfActiveOTG == 1) && LastStutterPeriod + MinTTUVBlankCriticalPlane > StutterEnterPlusExitWatermark) {
+			*StutterEfficiency = (1 - (*NumberOfStutterBurstsPerFrame * SRExitTime + StutterBurstTime * VActiveTimeCriticalPlane
+					/ *StutterPeriod) / FrameTimeCriticalPlane) * 100;
+		} else {
+			*StutterEfficiency = *StutterEfficiencyNotIncludingVBlank;
+		}
+	} else {
+		*StutterEfficiency = 0;
+	}
+
+	if (*Z8StutterEfficiencyNotIncludingVBlank > 0) {
+		LastZ8StutterPeriod = VActiveTimeCriticalPlane - (*NumberOfStutterBurstsPerFrame - 1) * *StutterPeriod;
+		if ((SynchronizedVBlank || TotalNumberOfActiveOTG == 1) && LastZ8StutterPeriod + MinTTUVBlankCriticalPlane > Z8StutterEnterPlusExitWatermark) {
+			*Z8StutterEfficiency = (1 - (*NumberOfStutterBurstsPerFrame * SRExitZ8Time + StutterBurstTime * VActiveTimeCriticalPlane
+					/ *StutterPeriod) / FrameTimeCriticalPlane) * 100;
+		} else {
+			*Z8StutterEfficiency = *Z8StutterEfficiencyNotIncludingVBlank;
+		}
+	} else {
+		*Z8StutterEfficiency = 0.;
+	}
+
+	dml_print("DML::%s: LastZ8StutterPeriod = %f\n", __func__, LastZ8StutterPeriod);
+	dml_print("DML::%s: Z8StutterEnterPlusExitWatermark = %f\n", __func__, Z8StutterEnterPlusExitWatermark);
+	dml_print("DML::%s: StutterBurstTime = %f\n", __func__, StutterBurstTime);
+	dml_print("DML::%s: StutterPeriod = %f\n", __func__, *StutterPeriod);
+	dml_print("DML::%s: StutterEfficiency = %f\n", __func__, *StutterEfficiency);
+	dml_print("DML::%s: Z8StutterEfficiency = %f\n", __func__, *Z8StutterEfficiency);
+	dml_print("DML::%s: StutterEfficiencyNotIncludingVBlank = %f\n", __func__, *StutterEfficiencyNotIncludingVBlank);
+	dml_print("DML::%s: Z8NumberOfStutterBurstsPerFrame = %d\n", __func__, *Z8NumberOfStutterBurstsPerFrame);
+}
+
+static void CalculateSwathAndDETConfiguration(
+		bool ForceSingleDPP,
+		int NumberOfActivePlanes,
+		unsigned int DETBufferSizeInKByte,
+		double MaximumSwathWidthLuma[],
+		double MaximumSwathWidthChroma[],
+		enum scan_direction_class SourceScan[],
+		enum source_format_class SourcePixelFormat[],
+		enum dm_swizzle_mode SurfaceTiling[],
+		int ViewportWidth[],
+		int ViewportHeight[],
+		int SurfaceWidthY[],
+		int SurfaceWidthC[],
+		int SurfaceHeightY[],
+		int SurfaceHeightC[],
+		int Read256BytesBlockHeightY[],
+		int Read256BytesBlockHeightC[],
+		int Read256BytesBlockWidthY[],
+		int Read256BytesBlockWidthC[],
+		enum odm_combine_mode ODMCombineEnabled[],
+		int BlendingAndTiming[],
+		int BytePerPixY[],
+		int BytePerPixC[],
+		double BytePerPixDETY[],
+		double BytePerPixDETC[],
+		int HActive[],
+		double HRatio[],
+		double HRatioChroma[],
+		int DPPPerPlane[],
+		int swath_width_luma_ub[],
+		int swath_width_chroma_ub[],
+		double SwathWidth[],
+		double SwathWidthChroma[],
+		int SwathHeightY[],
+		int SwathHeightC[],
+		unsigned int DETBufferSizeY[],
+		unsigned int DETBufferSizeC[],
+		bool ViewportSizeSupportPerPlane[],
+		bool *ViewportSizeSupport)
+{
+	int MaximumSwathHeightY[DC__NUM_DPP__MAX];
+	int MaximumSwathHeightC[DC__NUM_DPP__MAX];
+	int MinimumSwathHeightY;
+	int MinimumSwathHeightC;
+	int RoundedUpMaxSwathSizeBytesY;
+	int RoundedUpMaxSwathSizeBytesC;
+	int RoundedUpMinSwathSizeBytesY;
+	int RoundedUpMinSwathSizeBytesC;
+	int RoundedUpSwathSizeBytesY;
+	int RoundedUpSwathSizeBytesC;
+	double SwathWidthSingleDPP[DC__NUM_DPP__MAX];
+	double SwathWidthSingleDPPChroma[DC__NUM_DPP__MAX];
+	int k;
+
+	CalculateSwathWidth(
+			ForceSingleDPP,
+			NumberOfActivePlanes,
+			SourcePixelFormat,
+			SourceScan,
+			ViewportWidth,
+			ViewportHeight,
+			SurfaceWidthY,
+			SurfaceWidthC,
+			SurfaceHeightY,
+			SurfaceHeightC,
+			ODMCombineEnabled,
+			BytePerPixY,
+			BytePerPixC,
+			Read256BytesBlockHeightY,
+			Read256BytesBlockHeightC,
+			Read256BytesBlockWidthY,
+			Read256BytesBlockWidthC,
+			BlendingAndTiming,
+			HActive,
+			HRatio,
+			DPPPerPlane,
+			SwathWidthSingleDPP,
+			SwathWidthSingleDPPChroma,
+			SwathWidth,
+			SwathWidthChroma,
+			MaximumSwathHeightY,
+			MaximumSwathHeightC,
+			swath_width_luma_ub,
+			swath_width_chroma_ub);
+
+	*ViewportSizeSupport = true;
+	for (k = 0; k < NumberOfActivePlanes; ++k) {
+		if ((SourcePixelFormat[k] == dm_444_64 || SourcePixelFormat[k] == dm_444_32 || SourcePixelFormat[k] == dm_444_16 || SourcePixelFormat[k] == dm_mono_16
+				|| SourcePixelFormat[k] == dm_mono_8 || SourcePixelFormat[k] == dm_rgbe)) {
+			if (SurfaceTiling[k] == dm_sw_linear
+					|| (SourcePixelFormat[k] == dm_444_64
+							&& (SurfaceTiling[k] == dm_sw_64kb_s || SurfaceTiling[k] == dm_sw_64kb_s_t || SurfaceTiling[k] == dm_sw_64kb_s_x)
+							&& SourceScan[k] != dm_vert)) {
+				MinimumSwathHeightY = MaximumSwathHeightY[k];
+			} else if (SourcePixelFormat[k] == dm_444_8 && SourceScan[k] == dm_vert) {
+				MinimumSwathHeightY = MaximumSwathHeightY[k];
+			} else {
+				MinimumSwathHeightY = MaximumSwathHeightY[k] / 2;
+			}
+			MinimumSwathHeightC = MaximumSwathHeightC[k];
+		} else {
+			if (SurfaceTiling[k] == dm_sw_linear) {
+				MinimumSwathHeightY = MaximumSwathHeightY[k];
+				MinimumSwathHeightC = MaximumSwathHeightC[k];
+			} else if (SourcePixelFormat[k] == dm_rgbe_alpha && SourceScan[k] == dm_vert) {
+				MinimumSwathHeightY = MaximumSwathHeightY[k] / 2;
+				MinimumSwathHeightC = MaximumSwathHeightC[k];
+			} else if (SourcePixelFormat[k] == dm_rgbe_alpha) {
+				MinimumSwathHeightY = MaximumSwathHeightY[k] / 2;
+				MinimumSwathHeightC = MaximumSwathHeightC[k] / 2;
+			} else if (SourcePixelFormat[k] == dm_420_8 && SourceScan[k] == dm_vert) {
+				MinimumSwathHeightY = MaximumSwathHeightY[k];
+				MinimumSwathHeightC = MaximumSwathHeightC[k] / 2;
+			} else {
+				MinimumSwathHeightC = MaximumSwathHeightC[k] / 2;
+				MinimumSwathHeightY = MaximumSwathHeightY[k] / 2;
+			}
+		}
+
+		RoundedUpMaxSwathSizeBytesY = swath_width_luma_ub[k] * BytePerPixDETY[k] * MaximumSwathHeightY[k];
+		RoundedUpMinSwathSizeBytesY = swath_width_luma_ub[k] * BytePerPixDETY[k] * MinimumSwathHeightY;
+		if (SourcePixelFormat[k] == dm_420_10) {
+			RoundedUpMaxSwathSizeBytesY = dml_ceil((double) RoundedUpMaxSwathSizeBytesY, 256);
+			RoundedUpMinSwathSizeBytesY = dml_ceil((double) RoundedUpMinSwathSizeBytesY, 256);
+		}
+		RoundedUpMaxSwathSizeBytesC = swath_width_chroma_ub[k] * BytePerPixDETC[k] * MaximumSwathHeightC[k];
+		RoundedUpMinSwathSizeBytesC = swath_width_chroma_ub[k] * BytePerPixDETC[k] * MinimumSwathHeightC;
+		if (SourcePixelFormat[k] == dm_420_10) {
+			RoundedUpMaxSwathSizeBytesC = dml_ceil(RoundedUpMaxSwathSizeBytesC, 256);
+			RoundedUpMinSwathSizeBytesC = dml_ceil(RoundedUpMinSwathSizeBytesC, 256);
+		}
+
+		if (RoundedUpMaxSwathSizeBytesY + RoundedUpMaxSwathSizeBytesC <= DETBufferSizeInKByte * 1024 / 2) {
+			SwathHeightY[k] = MaximumSwathHeightY[k];
+			SwathHeightC[k] = MaximumSwathHeightC[k];
+			RoundedUpSwathSizeBytesY = RoundedUpMaxSwathSizeBytesY;
+			RoundedUpSwathSizeBytesC = RoundedUpMaxSwathSizeBytesC;
+		} else if (RoundedUpMaxSwathSizeBytesY >= 1.5 * RoundedUpMaxSwathSizeBytesC
+				&& RoundedUpMinSwathSizeBytesY + RoundedUpMaxSwathSizeBytesC <= DETBufferSizeInKByte * 1024 / 2) {
+			SwathHeightY[k] = MinimumSwathHeightY;
+			SwathHeightC[k] = MaximumSwathHeightC[k];
+			RoundedUpSwathSizeBytesY = RoundedUpMinSwathSizeBytesY;
+			RoundedUpSwathSizeBytesC = RoundedUpMaxSwathSizeBytesC;
+		} else if (RoundedUpMaxSwathSizeBytesY < 1.5 * RoundedUpMaxSwathSizeBytesC
+				&& RoundedUpMaxSwathSizeBytesY + RoundedUpMinSwathSizeBytesC <= DETBufferSizeInKByte * 1024 / 2) {
+			SwathHeightY[k] = MaximumSwathHeightY[k];
+			SwathHeightC[k] = MinimumSwathHeightC;
+			RoundedUpSwathSizeBytesY = RoundedUpMaxSwathSizeBytesY;
+			RoundedUpSwathSizeBytesC = RoundedUpMinSwathSizeBytesC;
+		} else {
+			SwathHeightY[k] = MinimumSwathHeightY;
+			SwathHeightC[k] = MinimumSwathHeightC;
+			RoundedUpSwathSizeBytesY = RoundedUpMinSwathSizeBytesY;
+			RoundedUpSwathSizeBytesC = RoundedUpMinSwathSizeBytesC;
+		}
+		{
+		double actDETBufferSizeInKByte = dml_ceil(DETBufferSizeInKByte, 64);
+
+		if (SwathHeightC[k] == 0) {
+			DETBufferSizeY[k] = actDETBufferSizeInKByte * 1024;
+			DETBufferSizeC[k] = 0;
+		} else if (RoundedUpSwathSizeBytesY <= 1.5 * RoundedUpSwathSizeBytesC) {
+			DETBufferSizeY[k] = actDETBufferSizeInKByte * 1024 / 2;
+			DETBufferSizeC[k] = actDETBufferSizeInKByte * 1024 / 2;
+		} else {
+			DETBufferSizeY[k] = dml_floor(actDETBufferSizeInKByte * 1024 * 2 / 3, 1024);
+			DETBufferSizeC[k] = actDETBufferSizeInKByte * 1024 / 3;
+		}
+
+		if (RoundedUpMinSwathSizeBytesY + RoundedUpMinSwathSizeBytesC > actDETBufferSizeInKByte * 1024 / 2 || SwathWidth[k] > MaximumSwathWidthLuma[k]
+				|| (SwathHeightC[k] > 0 && SwathWidthChroma[k] > MaximumSwathWidthChroma[k])) {
+			*ViewportSizeSupport = false;
+			ViewportSizeSupportPerPlane[k] = false;
+		} else {
+			ViewportSizeSupportPerPlane[k] = true;
+		}
+		}
+	}
+}
+
+static void CalculateSwathWidth(
+		bool ForceSingleDPP,
+		int NumberOfActivePlanes,
+		enum source_format_class SourcePixelFormat[],
+		enum scan_direction_class SourceScan[],
+		int ViewportWidth[],
+		int ViewportHeight[],
+		int SurfaceWidthY[],
+		int SurfaceWidthC[],
+		int SurfaceHeightY[],
+		int SurfaceHeightC[],
+		enum odm_combine_mode ODMCombineEnabled[],
+		int BytePerPixY[],
+		int BytePerPixC[],
+		int Read256BytesBlockHeightY[],
+		int Read256BytesBlockHeightC[],
+		int Read256BytesBlockWidthY[],
+		int Read256BytesBlockWidthC[],
+		int BlendingAndTiming[],
+		int HActive[],
+		double HRatio[],
+		int DPPPerPlane[],
+		double SwathWidthSingleDPPY[],
+		double SwathWidthSingleDPPC[],
+		double SwathWidthY[],
+		double SwathWidthC[],
+		int MaximumSwathHeightY[],
+		int MaximumSwathHeightC[],
+		int swath_width_luma_ub[],
+		int swath_width_chroma_ub[])
+{
+	enum odm_combine_mode MainPlaneODMCombine;
+	int j, k;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: NumberOfActivePlanes = %d\n", __func__, NumberOfActivePlanes);
+#endif
+
+	for (k = 0; k < NumberOfActivePlanes; ++k) {
+		if (SourceScan[k] != dm_vert) {
+			SwathWidthSingleDPPY[k] = ViewportWidth[k];
+		} else {
+			SwathWidthSingleDPPY[k] = ViewportHeight[k];
+		}
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%d ViewportWidth=%d\n", __func__, k, ViewportWidth[k]);
+		dml_print("DML::%s: k=%d ViewportHeight=%d\n", __func__, k, ViewportHeight[k]);
+#endif
+
+		MainPlaneODMCombine = ODMCombineEnabled[k];
+		for (j = 0; j < NumberOfActivePlanes; ++j) {
+			if (BlendingAndTiming[k] == j) {
+				MainPlaneODMCombine = ODMCombineEnabled[j];
+			}
+		}
+
+		if (MainPlaneODMCombine == dm_odm_combine_mode_4to1)
+			SwathWidthY[k] = dml_min(SwathWidthSingleDPPY[k], dml_round(HActive[k] / 4.0 * HRatio[k]));
+		else if (MainPlaneODMCombine == dm_odm_combine_mode_2to1)
+			SwathWidthY[k] = dml_min(SwathWidthSingleDPPY[k], dml_round(HActive[k] / 2.0 * HRatio[k]));
+		else if (DPPPerPlane[k] == 2)
+			SwathWidthY[k] = SwathWidthSingleDPPY[k] / 2;
+		else
+			SwathWidthY[k] = SwathWidthSingleDPPY[k];
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%d SwathWidthSingleDPPY=%f\n", __func__, k, SwathWidthSingleDPPY[k]);
+		dml_print("DML::%s: k=%d SwathWidthY=%f\n", __func__, k, SwathWidthY[k]);
+#endif
+
+		if (SourcePixelFormat[k] == dm_420_8 || SourcePixelFormat[k] == dm_420_10 || SourcePixelFormat[k] == dm_420_12) {
+			SwathWidthC[k] = SwathWidthY[k] / 2;
+			SwathWidthSingleDPPC[k] = SwathWidthSingleDPPY[k] / 2;
+		} else {
+			SwathWidthC[k] = SwathWidthY[k];
+			SwathWidthSingleDPPC[k] = SwathWidthSingleDPPY[k];
+		}
+
+		if (ForceSingleDPP == true) {
+			SwathWidthY[k] = SwathWidthSingleDPPY[k];
+			SwathWidthC[k] = SwathWidthSingleDPPC[k];
+		}
+		{
+		int surface_width_ub_l = dml_ceil(SurfaceWidthY[k], Read256BytesBlockWidthY[k]);
+		int surface_height_ub_l = dml_ceil(SurfaceHeightY[k], Read256BytesBlockHeightY[k]);
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%d surface_width_ub_l=%0d\n", __func__, k, surface_width_ub_l);
+#endif
+
+		if (SourceScan[k] != dm_vert) {
+			MaximumSwathHeightY[k] = Read256BytesBlockHeightY[k];
+			MaximumSwathHeightC[k] = Read256BytesBlockHeightC[k];
+			swath_width_luma_ub[k] = dml_min(surface_width_ub_l, (int) dml_ceil(SwathWidthY[k] - 1, Read256BytesBlockWidthY[k]) + Read256BytesBlockWidthY[k]);
+			if (BytePerPixC[k] > 0) {
+				int surface_width_ub_c = dml_ceil(SurfaceWidthC[k], Read256BytesBlockWidthC[k]);
+
+				swath_width_chroma_ub[k] = dml_min(
+						surface_width_ub_c,
+						(int) dml_ceil(SwathWidthC[k] - 1, Read256BytesBlockWidthC[k]) + Read256BytesBlockWidthC[k]);
+			} else {
+				swath_width_chroma_ub[k] = 0;
+			}
+		} else {
+			MaximumSwathHeightY[k] = Read256BytesBlockWidthY[k];
+			MaximumSwathHeightC[k] = Read256BytesBlockWidthC[k];
+			swath_width_luma_ub[k] = dml_min(surface_height_ub_l, (int) dml_ceil(SwathWidthY[k] - 1, Read256BytesBlockHeightY[k]) + Read256BytesBlockHeightY[k]);
+			if (BytePerPixC[k] > 0) {
+				int surface_height_ub_c = dml_ceil(SurfaceHeightC[k], Read256BytesBlockHeightC[k]);
+
+				swath_width_chroma_ub[k] = dml_min(
+						surface_height_ub_c,
+						(int) dml_ceil(SwathWidthC[k] - 1, Read256BytesBlockHeightC[k]) + Read256BytesBlockHeightC[k]);
+			} else {
+				swath_width_chroma_ub[k] = 0;
+			}
+		}
+		}
+	}
+}
+
+static double CalculateExtraLatency(
+		int RoundTripPingLatencyCycles,
+		int ReorderingBytes,
+		double DCFCLK,
+		int TotalNumberOfActiveDPP,
+		int PixelChunkSizeInKByte,
+		int TotalNumberOfDCCActiveDPP,
+		int MetaChunkSize,
+		double ReturnBW,
+		bool GPUVMEnable,
+		bool HostVMEnable,
+		int NumberOfActivePlanes,
+		int NumberOfDPP[],
+		int dpte_group_bytes[],
+		double HostVMInefficiencyFactor,
+		double HostVMMinPageSize,
+		int HostVMMaxNonCachedPageTableLevels)
+{
+	double ExtraLatencyBytes;
+	double ExtraLatency;
+
+	ExtraLatencyBytes = CalculateExtraLatencyBytes(
+			ReorderingBytes,
+			TotalNumberOfActiveDPP,
+			PixelChunkSizeInKByte,
+			TotalNumberOfDCCActiveDPP,
+			MetaChunkSize,
+			GPUVMEnable,
+			HostVMEnable,
+			NumberOfActivePlanes,
+			NumberOfDPP,
+			dpte_group_bytes,
+			HostVMInefficiencyFactor,
+			HostVMMinPageSize,
+			HostVMMaxNonCachedPageTableLevels);
+
+	ExtraLatency = (RoundTripPingLatencyCycles + __DML_ARB_TO_RET_DELAY__) / DCFCLK + ExtraLatencyBytes / ReturnBW;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: RoundTripPingLatencyCycles=%d\n", __func__, RoundTripPingLatencyCycles);
+	dml_print("DML::%s: DCFCLK=%f\n", __func__, DCFCLK);
+	dml_print("DML::%s: ExtraLatencyBytes=%f\n", __func__, ExtraLatencyBytes);
+	dml_print("DML::%s: ReturnBW=%f\n", __func__, ReturnBW);
+	dml_print("DML::%s: ExtraLatency=%f\n", __func__, ExtraLatency);
+#endif
+
+	return ExtraLatency;
+}
+
+static double CalculateExtraLatencyBytes(
+		int ReorderingBytes,
+		int TotalNumberOfActiveDPP,
+		int PixelChunkSizeInKByte,
+		int TotalNumberOfDCCActiveDPP,
+		int MetaChunkSize,
+		bool GPUVMEnable,
+		bool HostVMEnable,
+		int NumberOfActivePlanes,
+		int NumberOfDPP[],
+		int dpte_group_bytes[],
+		double HostVMInefficiencyFactor,
+		double HostVMMinPageSize,
+		int HostVMMaxNonCachedPageTableLevels)
+{
+	double ret;
+	int HostVMDynamicLevels = 0, k;
+
+	if (GPUVMEnable == true && HostVMEnable == true) {
+		if (HostVMMinPageSize < 2048)
+			HostVMDynamicLevels = HostVMMaxNonCachedPageTableLevels;
+		else if (HostVMMinPageSize >= 2048 && HostVMMinPageSize < 1048576)
+			HostVMDynamicLevels = dml_max(0, (int) HostVMMaxNonCachedPageTableLevels - 1);
+		else
+			HostVMDynamicLevels = dml_max(0, (int) HostVMMaxNonCachedPageTableLevels - 2);
+	} else {
+		HostVMDynamicLevels = 0;
+	}
+
+	ret = ReorderingBytes + (TotalNumberOfActiveDPP * PixelChunkSizeInKByte + TotalNumberOfDCCActiveDPP * MetaChunkSize) * 1024.0;
+
+	if (GPUVMEnable == true) {
+		for (k = 0; k < NumberOfActivePlanes; ++k)
+			ret = ret + NumberOfDPP[k] * dpte_group_bytes[k] * (1 + 8 * HostVMDynamicLevels) * HostVMInefficiencyFactor;
+	}
+	return ret;
+}
+
+static double CalculateUrgentLatency(
+		double UrgentLatencyPixelDataOnly,
+		double UrgentLatencyPixelMixedWithVMData,
+		double UrgentLatencyVMDataOnly,
+		bool DoUrgentLatencyAdjustment,
+		double UrgentLatencyAdjustmentFabricClockComponent,
+		double UrgentLatencyAdjustmentFabricClockReference,
+		double FabricClock)
+{
+	double ret;
+
+	ret = dml_max3(UrgentLatencyPixelDataOnly, UrgentLatencyPixelMixedWithVMData, UrgentLatencyVMDataOnly);
+	if (DoUrgentLatencyAdjustment == true)
+		ret = ret + UrgentLatencyAdjustmentFabricClockComponent * (UrgentLatencyAdjustmentFabricClockReference / FabricClock - 1);
+	return ret;
+}
+
+static noinline_for_stack void UseMinimumDCFCLK(
+		struct display_mode_lib *mode_lib,
+		int MaxPrefetchMode,
+		int ReorderingBytes)
+{
+	struct vba_vars_st *v = &mode_lib->vba;
+	int dummy1, i, j, k;
+	double NormalEfficiency,  dummy2, dummy3;
+	double TotalMaxPrefetchFlipDPTERowBandwidth[DC__VOLTAGE_STATES][2];
+
+	NormalEfficiency = v->PercentOfIdealFabricAndSDPPortBWReceivedAfterUrgLatency / 100.0;
+	for (i = 0; i < v->soc.num_states; ++i) {
+		for (j = 0; j <= 1; ++j) {
+			double PixelDCFCLKCyclesRequiredInPrefetch[DC__NUM_DPP__MAX];
+			double PrefetchPixelLinesTime[DC__NUM_DPP__MAX];
+			double DCFCLKRequiredForPeakBandwidthPerPlane[DC__NUM_DPP__MAX];
+			double DynamicMetadataVMExtraLatency[DC__NUM_DPP__MAX];
+			double MinimumTWait;
+			double NonDPTEBandwidth;
+			double DPTEBandwidth;
+			double DCFCLKRequiredForAverageBandwidth;
+			double ExtraLatencyBytes;
+			double ExtraLatencyCycles;
+			double DCFCLKRequiredForPeakBandwidth;
+			int NoOfDPPState[DC__NUM_DPP__MAX];
+			double MinimumTvmPlus2Tr0;
+
+			TotalMaxPrefetchFlipDPTERowBandwidth[i][j] = 0;
+			for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+				TotalMaxPrefetchFlipDPTERowBandwidth[i][j] = TotalMaxPrefetchFlipDPTERowBandwidth[i][j]
+						+ v->NoOfDPP[i][j][k] * v->DPTEBytesPerRow[i][j][k] / (15.75 * v->HTotal[k] / v->PixelClock[k]);
+			}
+
+			for (k = 0; k <= v->NumberOfActivePlanes - 1; ++k)
+				NoOfDPPState[k] = v->NoOfDPP[i][j][k];
+
+			MinimumTWait = CalculateTWait(MaxPrefetchMode, v->FinalDRAMClockChangeLatency, v->UrgLatency[i], v->SREnterPlusExitTime);
+			NonDPTEBandwidth = v->TotalVActivePixelBandwidth[i][j] + v->TotalVActiveCursorBandwidth[i][j] + v->TotalMetaRowBandwidth[i][j];
+			DPTEBandwidth = (v->HostVMEnable == true || v->ImmediateFlipRequirement[0] == dm_immediate_flip_required) ?
+					TotalMaxPrefetchFlipDPTERowBandwidth[i][j] : v->TotalDPTERowBandwidth[i][j];
+			DCFCLKRequiredForAverageBandwidth = dml_max3(
+					v->ProjectedDCFCLKDeepSleep[i][j],
+					(NonDPTEBandwidth + v->TotalDPTERowBandwidth[i][j]) / v->ReturnBusWidth
+							/ (v->MaxAveragePercentOfIdealFabricAndSDPPortBWDisplayCanUseInNormalSystemOperation / 100),
+					(NonDPTEBandwidth + DPTEBandwidth / NormalEfficiency) / NormalEfficiency / v->ReturnBusWidth);
+
+			ExtraLatencyBytes = CalculateExtraLatencyBytes(
+					ReorderingBytes,
+					v->TotalNumberOfActiveDPP[i][j],
+					v->PixelChunkSizeInKByte,
+					v->TotalNumberOfDCCActiveDPP[i][j],
+					v->MetaChunkSize,
+					v->GPUVMEnable,
+					v->HostVMEnable,
+					v->NumberOfActivePlanes,
+					NoOfDPPState,
+					v->dpte_group_bytes,
+					1,
+					v->HostVMMinPageSize,
+					v->HostVMMaxNonCachedPageTableLevels);
+			ExtraLatencyCycles = v->RoundTripPingLatencyCycles + __DML_ARB_TO_RET_DELAY__ + ExtraLatencyBytes / NormalEfficiency / v->ReturnBusWidth;
+			for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+				double DCFCLKCyclesRequiredInPrefetch;
+				double ExpectedPrefetchBWAcceleration;
+				double PrefetchTime;
+
+				PixelDCFCLKCyclesRequiredInPrefetch[k] = (v->PrefetchLinesY[i][j][k] * v->swath_width_luma_ub_all_states[i][j][k] * v->BytePerPixelY[k]
+						+ v->PrefetchLinesC[i][j][k] * v->swath_width_chroma_ub_all_states[i][j][k] * v->BytePerPixelC[k]) / NormalEfficiency / v->ReturnBusWidth;
+				DCFCLKCyclesRequiredInPrefetch = 2 * ExtraLatencyCycles / NoOfDPPState[k]
+						+ v->PDEAndMetaPTEBytesPerFrame[i][j][k] / NormalEfficiency / NormalEfficiency / v->ReturnBusWidth * (v->GPUVMMaxPageTableLevels > 2 ? 1 : 0)
+						+ 2 * v->DPTEBytesPerRow[i][j][k] / NormalEfficiency / NormalEfficiency / v->ReturnBusWidth
+						+ 2 * v->MetaRowBytes[i][j][k] / NormalEfficiency / v->ReturnBusWidth + PixelDCFCLKCyclesRequiredInPrefetch[k];
+				PrefetchPixelLinesTime[k] = dml_max(v->PrefetchLinesY[i][j][k], v->PrefetchLinesC[i][j][k]) * v->HTotal[k] / v->PixelClock[k];
+				ExpectedPrefetchBWAcceleration = (v->VActivePixelBandwidth[i][j][k] + v->VActiveCursorBandwidth[i][j][k])
+						/ (v->ReadBandwidthLuma[k] + v->ReadBandwidthChroma[k]);
+				DynamicMetadataVMExtraLatency[k] =
+						(v->GPUVMEnable == true && v->DynamicMetadataEnable[k] == true && v->DynamicMetadataVMEnabled == true) ?
+								v->UrgLatency[i] * v->GPUVMMaxPageTableLevels * (v->HostVMEnable == true ? v->HostVMMaxNonCachedPageTableLevels + 1 : 1) : 0;
+				PrefetchTime = (v->MaximumVStartup[i][j][k] - 1) * v->HTotal[k] / v->PixelClock[k] - MinimumTWait
+						- v->UrgLatency[i]
+								* ((v->GPUVMMaxPageTableLevels <= 2 ? v->GPUVMMaxPageTableLevels : v->GPUVMMaxPageTableLevels - 2)
+										* (v->HostVMEnable == true ? v->HostVMMaxNonCachedPageTableLevels + 1 : 1) - 1)
+						- DynamicMetadataVMExtraLatency[k];
+
+				if (PrefetchTime > 0) {
+					double ExpectedVRatioPrefetch;
+
+					ExpectedVRatioPrefetch = PrefetchPixelLinesTime[k]
+							/ (PrefetchTime * PixelDCFCLKCyclesRequiredInPrefetch[k] / DCFCLKCyclesRequiredInPrefetch);
+					DCFCLKRequiredForPeakBandwidthPerPlane[k] = NoOfDPPState[k] * PixelDCFCLKCyclesRequiredInPrefetch[k] / PrefetchPixelLinesTime[k]
+							* dml_max(1.0, ExpectedVRatioPrefetch) * dml_max(1.0, ExpectedVRatioPrefetch / 4) * ExpectedPrefetchBWAcceleration;
+					if (v->HostVMEnable == true || v->ImmediateFlipRequirement[0] == dm_immediate_flip_required) {
+						DCFCLKRequiredForPeakBandwidthPerPlane[k] = DCFCLKRequiredForPeakBandwidthPerPlane[k]
+								+ NoOfDPPState[k] * DPTEBandwidth / NormalEfficiency / NormalEfficiency / v->ReturnBusWidth;
+					}
+				} else {
+					DCFCLKRequiredForPeakBandwidthPerPlane[k] = v->DCFCLKPerState[i];
+				}
+				if (v->DynamicMetadataEnable[k] == true) {
+					double TSetupPipe;
+					double TdmbfPipe;
+					double TdmsksPipe;
+					double TdmecPipe;
+					double AllowedTimeForUrgentExtraLatency;
+
+					CalculateVupdateAndDynamicMetadataParameters(
+							v->MaxInterDCNTileRepeaters,
+							v->RequiredDPPCLK[i][j][k],
+							v->RequiredDISPCLK[i][j],
+							v->ProjectedDCFCLKDeepSleep[i][j],
+							v->PixelClock[k],
+							v->HTotal[k],
+							v->VTotal[k] - v->VActive[k],
+							v->DynamicMetadataTransmittedBytes[k],
+							v->DynamicMetadataLinesBeforeActiveRequired[k],
+							v->Interlace[k],
+							v->ProgressiveToInterlaceUnitInOPP,
+							&TSetupPipe,
+							&TdmbfPipe,
+							&TdmecPipe,
+							&TdmsksPipe,
+							&dummy1,
+							&dummy2,
+							&dummy3);
+					AllowedTimeForUrgentExtraLatency = v->MaximumVStartup[i][j][k] * v->HTotal[k] / v->PixelClock[k] - MinimumTWait - TSetupPipe - TdmbfPipe - TdmecPipe
+							- TdmsksPipe - DynamicMetadataVMExtraLatency[k];
+					if (AllowedTimeForUrgentExtraLatency > 0) {
+						DCFCLKRequiredForPeakBandwidthPerPlane[k] = dml_max(
+								DCFCLKRequiredForPeakBandwidthPerPlane[k],
+								ExtraLatencyCycles / AllowedTimeForUrgentExtraLatency);
+					} else {
+						DCFCLKRequiredForPeakBandwidthPerPlane[k] = v->DCFCLKPerState[i];
+					}
+				}
+			}
+			DCFCLKRequiredForPeakBandwidth = 0;
+			for (k = 0; k <= v->NumberOfActivePlanes - 1; ++k)
+				DCFCLKRequiredForPeakBandwidth = DCFCLKRequiredForPeakBandwidth + DCFCLKRequiredForPeakBandwidthPerPlane[k];
+
+			MinimumTvmPlus2Tr0 = v->UrgLatency[i]
+					* (v->GPUVMEnable == true ?
+							(v->HostVMEnable == true ?
+									(v->GPUVMMaxPageTableLevels + 2) * (v->HostVMMaxNonCachedPageTableLevels + 1) - 1 : v->GPUVMMaxPageTableLevels + 1) :
+							0);
+			for (k = 0; k < v->NumberOfActivePlanes; ++k) {
+				double MaximumTvmPlus2Tr0PlusTsw;
+
+				MaximumTvmPlus2Tr0PlusTsw = (v->MaximumVStartup[i][j][k] - 2) * v->HTotal[k] / v->PixelClock[k] - MinimumTWait - DynamicMetadataVMExtraLatency[k];
+				if (MaximumTvmPlus2Tr0PlusTsw <= MinimumTvmPlus2Tr0 + PrefetchPixelLinesTime[k] / 4) {
+					DCFCLKRequiredForPeakBandwidth = v->DCFCLKPerState[i];
+				} else {
+					DCFCLKRequiredForPeakBandwidth = dml_max3(
+							DCFCLKRequiredForPeakBandwidth,
+							2 * ExtraLatencyCycles / (MaximumTvmPlus2Tr0PlusTsw - MinimumTvmPlus2Tr0 - PrefetchPixelLinesTime[k] / 4),
+							(2 * ExtraLatencyCycles + PixelDCFCLKCyclesRequiredInPrefetch[k]) / (MaximumTvmPlus2Tr0PlusTsw - MinimumTvmPlus2Tr0));
+				}
+			}
+			v->DCFCLKState[i][j] = dml_min(v->DCFCLKPerState[i], 1.05 * dml_max(DCFCLKRequiredForAverageBandwidth, DCFCLKRequiredForPeakBandwidth));
+		}
+	}
+}
+
+static void CalculateUnboundedRequestAndCompressedBufferSize(
+		unsigned int DETBufferSizeInKByte,
+		int ConfigReturnBufferSizeInKByte,
+		enum unbounded_requesting_policy UseUnboundedRequestingFinal,
+		int TotalActiveDPP,
+		bool NoChromaPlanes,
+		int MaxNumDPP,
+		int CompressedBufferSegmentSizeInkByteFinal,
+		enum output_encoder_class *Output,
+		bool *UnboundedRequestEnabled,
+		int *CompressedBufferSizeInkByte)
+{
+	double actDETBufferSizeInKByte = dml_ceil(DETBufferSizeInKByte, 64);
+
+	*UnboundedRequestEnabled = UnboundedRequest(UseUnboundedRequestingFinal, TotalActiveDPP, NoChromaPlanes, Output[0]);
+	*CompressedBufferSizeInkByte = (
+			*UnboundedRequestEnabled == true ?
+					ConfigReturnBufferSizeInKByte - TotalActiveDPP * actDETBufferSizeInKByte :
+					ConfigReturnBufferSizeInKByte - MaxNumDPP * actDETBufferSizeInKByte);
+	*CompressedBufferSizeInkByte = *CompressedBufferSizeInkByte * CompressedBufferSegmentSizeInkByteFinal / 64;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: TotalActiveDPP = %d\n", __func__, TotalActiveDPP);
+	dml_print("DML::%s: DETBufferSizeInKByte = %d\n", __func__, DETBufferSizeInKByte);
+	dml_print("DML::%s: ConfigReturnBufferSizeInKByte = %d\n", __func__, ConfigReturnBufferSizeInKByte);
+	dml_print("DML::%s: UseUnboundedRequestingFinal = %d\n", __func__, UseUnboundedRequestingFinal);
+	dml_print("DML::%s: actDETBufferSizeInKByte = %f\n", __func__, actDETBufferSizeInKByte);
+	dml_print("DML::%s: UnboundedRequestEnabled = %d\n", __func__, *UnboundedRequestEnabled);
+	dml_print("DML::%s: CompressedBufferSizeInkByte = %d\n", __func__, *CompressedBufferSizeInkByte);
+#endif
+}
+
+static bool UnboundedRequest(enum unbounded_requesting_policy UseUnboundedRequestingFinal, int TotalNumberOfActiveDPP, bool NoChroma, enum output_encoder_class Output)
+{
+	bool ret_val = false;
+
+	ret_val = (UseUnboundedRequestingFinal != dm_unbounded_requesting_disable && TotalNumberOfActiveDPP == 1 && NoChroma);
+	if (UseUnboundedRequestingFinal == dm_unbounded_requesting_edp_only && Output != dm_edp)
+		ret_val = false;
+	return ret_val;
+}
+
+static unsigned int CalculateMaxVStartup(
+		unsigned int VTotal,
+		unsigned int VActive,
+		unsigned int VBlankNom,
+		unsigned int HTotal,
+		double PixelClock,
+		bool ProgressiveTointerlaceUnitinOPP,
+		bool Interlace,
+		unsigned int VBlankNomDefaultUS,
+		double WritebackDelayTime)
+{
+	unsigned int MaxVStartup = 0;
+	unsigned int vblank_size = 0;
+	double line_time_us = HTotal / PixelClock;
+	unsigned int vblank_actual = VTotal - VActive;
+	unsigned int vblank_nom_default_in_line = dml_floor(VBlankNomDefaultUS / line_time_us, 1.0);
+	unsigned int vblank_nom_input = VBlankNom; //dml_min(VBlankNom, vblank_nom_default_in_line);
+	unsigned int vblank_avail = vblank_nom_input == 0 ? vblank_nom_default_in_line : vblank_nom_input;
+
+	vblank_size = (unsigned int) dml_min(vblank_actual, vblank_avail);
+	if (Interlace && !ProgressiveTointerlaceUnitinOPP)
+		MaxVStartup = dml_floor(vblank_size / 2.0, 1.0);
+	else
+		MaxVStartup = vblank_size - dml_max(1.0, dml_ceil(WritebackDelayTime / line_time_us, 1.0));
+	if (MaxVStartup > 1023)
+		MaxVStartup = 1023;
+	return MaxVStartup;
+}
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn314/display_mode_vba_314.h b/drivers/gpu/drm/amd/display/dc/dml/dcn314/display_mode_vba_314.h
new file mode 100644
index 000000000000..a8199ab7d26a
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn314/display_mode_vba_314.h
@@ -0,0 +1,44 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __DML314_DISPLAY_MODE_VBA_H__
+#define __DML314_DISPLAY_MODE_VBA_H__
+
+void dml314_recalculate(struct display_mode_lib *mode_lib);
+void dml314_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_lib);
+double dml314_CalculateWriteBackDISPCLK(
+		enum source_format_class WritebackPixelFormat,
+		double PixelClock,
+		double WritebackHRatio,
+		double WritebackVRatio,
+		unsigned int WritebackHTaps,
+		unsigned int WritebackVTaps,
+		long   WritebackSourceWidth,
+		long   WritebackDestinationWidth,
+		unsigned int HTotal,
+		unsigned int WritebackLineBufferSize);
+
+#endif /* __DML314_DISPLAY_MODE_VBA_H__ */
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn314/display_rq_dlg_calc_314.c b/drivers/gpu/drm/amd/display/dc/dml/dcn314/display_rq_dlg_calc_314.c
new file mode 100644
index 000000000000..04df263ff65e
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn314/display_rq_dlg_calc_314.c
@@ -0,0 +1,1678 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#include "../display_mode_lib.h"
+#include "../display_mode_vba.h"
+#include "../dml_inline_defs.h"
+#include "display_rq_dlg_calc_314.h"
+
+static bool CalculateBytePerPixelAnd256BBlockSizes(
+		enum source_format_class SourcePixelFormat,
+		enum dm_swizzle_mode SurfaceTiling,
+		unsigned int *BytePerPixelY,
+		unsigned int *BytePerPixelC,
+		double *BytePerPixelDETY,
+		double *BytePerPixelDETC,
+		unsigned int *BlockHeight256BytesY,
+		unsigned int *BlockHeight256BytesC,
+		unsigned int *BlockWidth256BytesY,
+		unsigned int *BlockWidth256BytesC)
+{
+	if (SourcePixelFormat == dm_444_64) {
+		*BytePerPixelDETY = 8;
+		*BytePerPixelDETC = 0;
+		*BytePerPixelY = 8;
+		*BytePerPixelC = 0;
+	} else if (SourcePixelFormat == dm_444_32 || SourcePixelFormat == dm_rgbe) {
+		*BytePerPixelDETY = 4;
+		*BytePerPixelDETC = 0;
+		*BytePerPixelY = 4;
+		*BytePerPixelC = 0;
+	} else if (SourcePixelFormat == dm_444_16) {
+		*BytePerPixelDETY = 2;
+		*BytePerPixelDETC = 0;
+		*BytePerPixelY = 2;
+		*BytePerPixelC = 0;
+	} else if (SourcePixelFormat == dm_444_8) {
+		*BytePerPixelDETY = 1;
+		*BytePerPixelDETC = 0;
+		*BytePerPixelY = 1;
+		*BytePerPixelC = 0;
+	} else if (SourcePixelFormat == dm_rgbe_alpha) {
+		*BytePerPixelDETY = 4;
+		*BytePerPixelDETC = 1;
+		*BytePerPixelY = 4;
+		*BytePerPixelC = 1;
+	} else if (SourcePixelFormat == dm_420_8) {
+		*BytePerPixelDETY = 1;
+		*BytePerPixelDETC = 2;
+		*BytePerPixelY = 1;
+		*BytePerPixelC = 2;
+	} else if (SourcePixelFormat == dm_420_12) {
+		*BytePerPixelDETY = 2;
+		*BytePerPixelDETC = 4;
+		*BytePerPixelY = 2;
+		*BytePerPixelC = 4;
+	} else {
+		*BytePerPixelDETY = 4.0 / 3;
+		*BytePerPixelDETC = 8.0 / 3;
+		*BytePerPixelY = 2;
+		*BytePerPixelC = 4;
+	}
+
+	if ((SourcePixelFormat == dm_444_64 || SourcePixelFormat == dm_444_32 || SourcePixelFormat == dm_444_16 || SourcePixelFormat == dm_444_8 || SourcePixelFormat == dm_mono_16
+			|| SourcePixelFormat == dm_mono_8 || SourcePixelFormat == dm_rgbe)) {
+		if (SurfaceTiling == dm_sw_linear)
+			*BlockHeight256BytesY = 1;
+		else if (SourcePixelFormat == dm_444_64)
+			*BlockHeight256BytesY = 4;
+		else if (SourcePixelFormat == dm_444_8)
+			*BlockHeight256BytesY = 16;
+		else
+			*BlockHeight256BytesY = 8;
+
+		*BlockWidth256BytesY = 256U / *BytePerPixelY / *BlockHeight256BytesY;
+		*BlockHeight256BytesC = 0;
+		*BlockWidth256BytesC = 0;
+	} else {
+		if (SurfaceTiling == dm_sw_linear) {
+			*BlockHeight256BytesY = 1;
+			*BlockHeight256BytesC = 1;
+		} else if (SourcePixelFormat == dm_rgbe_alpha) {
+			*BlockHeight256BytesY = 8;
+			*BlockHeight256BytesC = 16;
+		} else if (SourcePixelFormat == dm_420_8) {
+			*BlockHeight256BytesY = 16;
+			*BlockHeight256BytesC = 8;
+		} else {
+			*BlockHeight256BytesY = 8;
+			*BlockHeight256BytesC = 8;
+		}
+		*BlockWidth256BytesY = 256U / *BytePerPixelY / *BlockHeight256BytesY;
+		*BlockWidth256BytesC = 256U / *BytePerPixelC / *BlockHeight256BytesC;
+	}
+	return true;
+}
+
+static bool is_dual_plane(enum source_format_class source_format)
+{
+	bool ret_val = 0;
+
+	if ((source_format == dm_420_12) || (source_format == dm_420_8) || (source_format == dm_420_10) || (source_format == dm_rgbe_alpha))
+		ret_val = 1;
+
+	return ret_val;
+}
+
+static double get_refcyc_per_delivery(
+		struct display_mode_lib *mode_lib,
+		double refclk_freq_in_mhz,
+		double pclk_freq_in_mhz,
+		unsigned int odm_combine,
+		unsigned int recout_width,
+		unsigned int hactive,
+		double vratio,
+		double hscale_pixel_rate,
+		unsigned int delivery_width,
+		unsigned int req_per_swath_ub)
+{
+	double refcyc_per_delivery = 0.0;
+
+	if (vratio <= 1.0) {
+		if (odm_combine)
+			refcyc_per_delivery = (double) refclk_freq_in_mhz * (double) ((unsigned int) odm_combine * 2)
+					* dml_min((double) recout_width, (double) hactive / ((unsigned int) odm_combine * 2)) / pclk_freq_in_mhz / (double) req_per_swath_ub;
+		else
+			refcyc_per_delivery = (double) refclk_freq_in_mhz * (double) recout_width / pclk_freq_in_mhz / (double) req_per_swath_ub;
+	} else {
+		refcyc_per_delivery = (double) refclk_freq_in_mhz * (double) delivery_width / (double) hscale_pixel_rate / (double) req_per_swath_ub;
+	}
+
+#ifdef __DML_RQ_DLG_CALC_DEBUG__
+	dml_print("DML_DLG: %s: refclk_freq_in_mhz = %3.2f\n", __func__, refclk_freq_in_mhz);
+	dml_print("DML_DLG: %s: pclk_freq_in_mhz   = %3.2f\n", __func__, pclk_freq_in_mhz);
+	dml_print("DML_DLG: %s: recout_width       = %d\n", __func__, recout_width);
+	dml_print("DML_DLG: %s: vratio             = %3.2f\n", __func__, vratio);
+	dml_print("DML_DLG: %s: req_per_swath_ub   = %d\n", __func__, req_per_swath_ub);
+	dml_print("DML_DLG: %s: hscale_pixel_rate  = %3.2f\n", __func__, hscale_pixel_rate);
+	dml_print("DML_DLG: %s: delivery_width     = %d\n", __func__, delivery_width);
+	dml_print("DML_DLG: %s: refcyc_per_delivery= %3.2f\n", __func__, refcyc_per_delivery);
+#endif
+
+	return refcyc_per_delivery;
+
+}
+
+static unsigned int get_blk_size_bytes(const enum source_macro_tile_size tile_size)
+{
+	if (tile_size == dm_256k_tile)
+		return (256 * 1024);
+	else if (tile_size == dm_64k_tile)
+		return (64 * 1024);
+	else
+		return (4 * 1024);
+}
+
+static void extract_rq_sizing_regs(struct display_mode_lib *mode_lib, display_data_rq_regs_st *rq_regs, const display_data_rq_sizing_params_st *rq_sizing)
+{
+	print__data_rq_sizing_params_st(mode_lib, rq_sizing);
+
+	rq_regs->chunk_size = dml_log2(rq_sizing->chunk_bytes) - 10;
+
+	if (rq_sizing->min_chunk_bytes == 0)
+		rq_regs->min_chunk_size = 0;
+	else
+		rq_regs->min_chunk_size = dml_log2(rq_sizing->min_chunk_bytes) - 8 + 1;
+
+	rq_regs->meta_chunk_size = dml_log2(rq_sizing->meta_chunk_bytes) - 10;
+	if (rq_sizing->min_meta_chunk_bytes == 0)
+		rq_regs->min_meta_chunk_size = 0;
+	else
+		rq_regs->min_meta_chunk_size = dml_log2(rq_sizing->min_meta_chunk_bytes) - 6 + 1;
+
+	rq_regs->dpte_group_size = dml_log2(rq_sizing->dpte_group_bytes) - 6;
+	rq_regs->mpte_group_size = dml_log2(rq_sizing->mpte_group_bytes) - 6;
+}
+
+static void extract_rq_regs(struct display_mode_lib *mode_lib, display_rq_regs_st *rq_regs, const display_rq_params_st *rq_param)
+{
+	unsigned int detile_buf_size_in_bytes = mode_lib->ip.det_buffer_size_kbytes * 1024;
+	unsigned int detile_buf_plane1_addr = 0;
+
+	extract_rq_sizing_regs(mode_lib, &(rq_regs->rq_regs_l), &rq_param->sizing.rq_l);
+
+	rq_regs->rq_regs_l.pte_row_height_linear = dml_floor(dml_log2(rq_param->dlg.rq_l.dpte_row_height), 1) - 3;
+
+	if (rq_param->yuv420) {
+		extract_rq_sizing_regs(mode_lib, &(rq_regs->rq_regs_c), &rq_param->sizing.rq_c);
+		rq_regs->rq_regs_c.pte_row_height_linear = dml_floor(dml_log2(rq_param->dlg.rq_c.dpte_row_height), 1) - 3;
+	}
+
+	rq_regs->rq_regs_l.swath_height = dml_log2(rq_param->dlg.rq_l.swath_height);
+	rq_regs->rq_regs_c.swath_height = dml_log2(rq_param->dlg.rq_c.swath_height);
+
+	// FIXME: take the max between luma, chroma chunk size?
+	// okay for now, as we are setting chunk_bytes to 8kb anyways
+	if (rq_param->sizing.rq_l.chunk_bytes >= 32 * 1024 || (rq_param->yuv420 && rq_param->sizing.rq_c.chunk_bytes >= 32 * 1024)) { //32kb
+		rq_regs->drq_expansion_mode = 0;
+	} else {
+		rq_regs->drq_expansion_mode = 2;
+	}
+	rq_regs->prq_expansion_mode = 1;
+	rq_regs->mrq_expansion_mode = 1;
+	rq_regs->crq_expansion_mode = 1;
+
+	// Note: detile_buf_plane1_addr is in unit of 1KB
+	if (rq_param->yuv420) {
+		if ((double) rq_param->misc.rq_l.stored_swath_bytes / (double) rq_param->misc.rq_c.stored_swath_bytes <= 1.5) {
+			detile_buf_plane1_addr = (detile_buf_size_in_bytes / 2.0 / 1024.0); // half to chroma
+#ifdef __DML_RQ_DLG_CALC_DEBUG__
+					dml_print("DML_DLG: %s: detile_buf_plane1_addr = %0d (1/2 to chroma)\n", __func__, detile_buf_plane1_addr);
+#endif
+		} else {
+			detile_buf_plane1_addr = dml_round_to_multiple((unsigned int) ((2.0 * detile_buf_size_in_bytes) / 3.0), 1024, 0) / 1024.0; // 2/3 to luma
+#ifdef __DML_RQ_DLG_CALC_DEBUG__
+					dml_print("DML_DLG: %s: detile_buf_plane1_addr = %0d (1/3 chroma)\n", __func__, detile_buf_plane1_addr);
+#endif
+		}
+	}
+	rq_regs->plane1_base_address = detile_buf_plane1_addr;
+
+#ifdef __DML_RQ_DLG_CALC_DEBUG__
+	dml_print("DML_DLG: %s: detile_buf_size_in_bytes = %0d\n", __func__, detile_buf_size_in_bytes);
+	dml_print("DML_DLG: %s: detile_buf_plane1_addr = %0d\n", __func__, detile_buf_plane1_addr);
+	dml_print("DML_DLG: %s: plane1_base_address = %0d\n", __func__, rq_regs->plane1_base_address);
+	dml_print("DML_DLG: %s: rq_l.stored_swath_bytes = %0d\n", __func__, rq_param->misc.rq_l.stored_swath_bytes);
+	dml_print("DML_DLG: %s: rq_c.stored_swath_bytes = %0d\n", __func__, rq_param->misc.rq_c.stored_swath_bytes);
+	dml_print("DML_DLG: %s: rq_l.swath_height = %0d\n", __func__, rq_param->dlg.rq_l.swath_height);
+	dml_print("DML_DLG: %s: rq_c.swath_height = %0d\n", __func__, rq_param->dlg.rq_c.swath_height);
+#endif
+}
+
+static void handle_det_buf_split(struct display_mode_lib *mode_lib, display_rq_params_st *rq_param, const display_pipe_source_params_st *pipe_src_param)
+{
+	unsigned int total_swath_bytes = 0;
+	unsigned int swath_bytes_l = 0;
+	unsigned int swath_bytes_c = 0;
+	unsigned int full_swath_bytes_packed_l = 0;
+	unsigned int full_swath_bytes_packed_c = 0;
+	bool req128_l = 0;
+	bool req128_c = 0;
+	bool surf_linear = (pipe_src_param->sw_mode == dm_sw_linear);
+	bool surf_vert = (pipe_src_param->source_scan == dm_vert);
+	unsigned int log2_swath_height_l = 0;
+	unsigned int log2_swath_height_c = 0;
+	unsigned int detile_buf_size_in_bytes = mode_lib->ip.det_buffer_size_kbytes * 1024;
+
+	full_swath_bytes_packed_l = rq_param->misc.rq_l.full_swath_bytes;
+	full_swath_bytes_packed_c = rq_param->misc.rq_c.full_swath_bytes;
+
+#ifdef __DML_RQ_DLG_CALC_DEBUG__
+	dml_print("DML_DLG: %s: full_swath_bytes_packed_l = %0d\n", __func__, full_swath_bytes_packed_l);
+	dml_print("DML_DLG: %s: full_swath_bytes_packed_c = %0d\n", __func__, full_swath_bytes_packed_c);
+#endif
+
+	if (rq_param->yuv420_10bpc) {
+		full_swath_bytes_packed_l = dml_round_to_multiple(rq_param->misc.rq_l.full_swath_bytes * 2.0 / 3.0, 256, 1) + 256;
+		full_swath_bytes_packed_c = dml_round_to_multiple(rq_param->misc.rq_c.full_swath_bytes * 2.0 / 3.0, 256, 1) + 256;
+#ifdef __DML_RQ_DLG_CALC_DEBUG__
+		dml_print("DML_DLG: %s: full_swath_bytes_packed_l = %0d (3-2 packing)\n", __func__, full_swath_bytes_packed_l);
+		dml_print("DML_DLG: %s: full_swath_bytes_packed_c = %0d (3-2 packing)\n", __func__, full_swath_bytes_packed_c);
+#endif
+	}
+
+	if (rq_param->yuv420)
+		total_swath_bytes = 2 * full_swath_bytes_packed_l + 2 * full_swath_bytes_packed_c;
+	else
+		total_swath_bytes = 2 * full_swath_bytes_packed_l;
+
+#ifdef __DML_RQ_DLG_CALC_DEBUG__
+	dml_print("DML_DLG: %s: total_swath_bytes = %0d\n", __func__, total_swath_bytes);
+	dml_print("DML_DLG: %s: detile_buf_size_in_bytes = %0d\n", __func__, detile_buf_size_in_bytes);
+#endif
+
+	if (total_swath_bytes <= detile_buf_size_in_bytes) { //full 256b request
+		req128_l = 0;
+		req128_c = 0;
+		swath_bytes_l = full_swath_bytes_packed_l;
+		swath_bytes_c = full_swath_bytes_packed_c;
+	} else if (!rq_param->yuv420) {
+		req128_l = 1;
+		req128_c = 0;
+		swath_bytes_c = full_swath_bytes_packed_c;
+		swath_bytes_l = full_swath_bytes_packed_l / 2;
+	} else if ((double) full_swath_bytes_packed_l / (double) full_swath_bytes_packed_c < 1.5) {
+		req128_l = 0;
+		req128_c = 1;
+		swath_bytes_l = full_swath_bytes_packed_l;
+		swath_bytes_c = full_swath_bytes_packed_c / 2;
+
+		total_swath_bytes = 2 * swath_bytes_l + 2 * swath_bytes_c;
+
+		if (total_swath_bytes > detile_buf_size_in_bytes) {
+			req128_l = 1;
+			swath_bytes_l = full_swath_bytes_packed_l / 2;
+		}
+	} else {
+		req128_l = 1;
+		req128_c = 0;
+		swath_bytes_l = full_swath_bytes_packed_l / 2;
+		swath_bytes_c = full_swath_bytes_packed_c;
+
+		total_swath_bytes = 2 * swath_bytes_l + 2 * swath_bytes_c;
+
+		if (total_swath_bytes > detile_buf_size_in_bytes) {
+			req128_c = 1;
+			swath_bytes_c = full_swath_bytes_packed_c / 2;
+		}
+	}
+
+	if (rq_param->yuv420)
+		total_swath_bytes = 2 * swath_bytes_l + 2 * swath_bytes_c;
+	else
+		total_swath_bytes = 2 * swath_bytes_l;
+
+	rq_param->misc.rq_l.stored_swath_bytes = swath_bytes_l;
+	rq_param->misc.rq_c.stored_swath_bytes = swath_bytes_c;
+
+#ifdef __DML_RQ_DLG_CALC_DEBUG__
+	dml_print("DML_DLG: %s: total_swath_bytes = %0d\n", __func__, total_swath_bytes);
+	dml_print("DML_DLG: %s: rq_l.stored_swath_bytes = %0d\n", __func__, rq_param->misc.rq_l.stored_swath_bytes);
+	dml_print("DML_DLG: %s: rq_c.stored_swath_bytes = %0d\n", __func__, rq_param->misc.rq_c.stored_swath_bytes);
+#endif
+	if (surf_linear) {
+		log2_swath_height_l = 0;
+		log2_swath_height_c = 0;
+	} else {
+		unsigned int swath_height_l;
+		unsigned int swath_height_c;
+
+		if (!surf_vert) {
+			swath_height_l = rq_param->misc.rq_l.blk256_height;
+			swath_height_c = rq_param->misc.rq_c.blk256_height;
+		} else {
+			swath_height_l = rq_param->misc.rq_l.blk256_width;
+			swath_height_c = rq_param->misc.rq_c.blk256_width;
+		}
+
+		if (swath_height_l > 0)
+			log2_swath_height_l = dml_log2(swath_height_l);
+
+		if (req128_l && log2_swath_height_l > 0)
+			log2_swath_height_l -= 1;
+
+		if (swath_height_c > 0)
+			log2_swath_height_c = dml_log2(swath_height_c);
+
+		if (req128_c && log2_swath_height_c > 0)
+			log2_swath_height_c -= 1;
+	}
+
+	rq_param->dlg.rq_l.swath_height = 1 << log2_swath_height_l;
+	rq_param->dlg.rq_c.swath_height = 1 << log2_swath_height_c;
+
+#ifdef __DML_RQ_DLG_CALC_DEBUG__
+	dml_print("DML_DLG: %s: req128_l = %0d\n", __func__, req128_l);
+	dml_print("DML_DLG: %s: req128_c = %0d\n", __func__, req128_c);
+	dml_print("DML_DLG: %s: full_swath_bytes_packed_l = %0d\n", __func__, full_swath_bytes_packed_l);
+	dml_print("DML_DLG: %s: full_swath_bytes_packed_c = %0d\n", __func__, full_swath_bytes_packed_c);
+	dml_print("DML_DLG: %s: swath_height luma = %0d\n", __func__, rq_param->dlg.rq_l.swath_height);
+	dml_print("DML_DLG: %s: swath_height chroma = %0d\n", __func__, rq_param->dlg.rq_c.swath_height);
+#endif
+}
+
+static void get_meta_and_pte_attr(
+		struct display_mode_lib *mode_lib,
+		display_data_rq_dlg_params_st *rq_dlg_param,
+		display_data_rq_misc_params_st *rq_misc_param,
+		display_data_rq_sizing_params_st *rq_sizing_param,
+		unsigned int vp_width,
+		unsigned int vp_height,
+		unsigned int data_pitch,
+		unsigned int meta_pitch,
+		unsigned int source_format,
+		unsigned int tiling,
+		unsigned int macro_tile_size,
+		unsigned int source_scan,
+		unsigned int hostvm_enable,
+		unsigned int is_chroma,
+		unsigned int surface_height)
+{
+	bool surf_linear = (tiling == dm_sw_linear);
+	bool surf_vert = (source_scan == dm_vert);
+
+	unsigned int bytes_per_element;
+	unsigned int bytes_per_element_y;
+	unsigned int bytes_per_element_c;
+
+	unsigned int blk256_width = 0;
+	unsigned int blk256_height = 0;
+
+	unsigned int blk256_width_y = 0;
+	unsigned int blk256_height_y = 0;
+	unsigned int blk256_width_c = 0;
+	unsigned int blk256_height_c = 0;
+	unsigned int log2_bytes_per_element;
+	unsigned int log2_blk256_width;
+	unsigned int log2_blk256_height;
+	unsigned int blk_bytes;
+	unsigned int log2_blk_bytes;
+	unsigned int log2_blk_height;
+	unsigned int log2_blk_width;
+	unsigned int log2_meta_req_bytes;
+	unsigned int log2_meta_req_height;
+	unsigned int log2_meta_req_width;
+	unsigned int meta_req_width;
+	unsigned int meta_req_height;
+	unsigned int log2_meta_row_height;
+	unsigned int meta_row_width_ub;
+	unsigned int log2_meta_chunk_bytes;
+	unsigned int log2_meta_chunk_height;
+
+	//full sized meta chunk width in unit of data elements
+	unsigned int log2_meta_chunk_width;
+	unsigned int log2_min_meta_chunk_bytes;
+	unsigned int min_meta_chunk_width;
+	unsigned int meta_chunk_width;
+	unsigned int meta_chunk_per_row_int;
+	unsigned int meta_row_remainder;
+	unsigned int meta_chunk_threshold;
+	unsigned int meta_blk_height;
+	unsigned int meta_surface_bytes;
+	unsigned int vmpg_bytes;
+	unsigned int meta_pte_req_per_frame_ub;
+	unsigned int meta_pte_bytes_per_frame_ub;
+	const unsigned int log2_vmpg_bytes = dml_log2(mode_lib->soc.gpuvm_min_page_size_bytes);
+	const bool dual_plane_en = is_dual_plane((enum source_format_class) (source_format));
+	const unsigned int dpte_buf_in_pte_reqs =
+			dual_plane_en ? (is_chroma ? mode_lib->ip.dpte_buffer_size_in_pte_reqs_chroma : mode_lib->ip.dpte_buffer_size_in_pte_reqs_luma) : (mode_lib->ip.dpte_buffer_size_in_pte_reqs_luma
+							+ mode_lib->ip.dpte_buffer_size_in_pte_reqs_chroma);
+
+	unsigned int log2_vmpg_height = 0;
+	unsigned int log2_vmpg_width = 0;
+	unsigned int log2_dpte_req_height_ptes = 0;
+	unsigned int log2_dpte_req_height = 0;
+	unsigned int log2_dpte_req_width = 0;
+	unsigned int log2_dpte_row_height_linear = 0;
+	unsigned int log2_dpte_row_height = 0;
+	unsigned int log2_dpte_group_width = 0;
+	unsigned int dpte_row_width_ub = 0;
+	unsigned int dpte_req_height = 0;
+	unsigned int dpte_req_width = 0;
+	unsigned int dpte_group_width = 0;
+	unsigned int log2_dpte_group_bytes = 0;
+	unsigned int log2_dpte_group_length = 0;
+	double byte_per_pixel_det_y;
+	double byte_per_pixel_det_c;
+
+	CalculateBytePerPixelAnd256BBlockSizes(
+			(enum source_format_class) (source_format),
+			(enum dm_swizzle_mode) (tiling),
+			&bytes_per_element_y,
+			&bytes_per_element_c,
+			&byte_per_pixel_det_y,
+			&byte_per_pixel_det_c,
+			&blk256_height_y,
+			&blk256_height_c,
+			&blk256_width_y,
+			&blk256_width_c);
+
+	if (!is_chroma) {
+		blk256_width = blk256_width_y;
+		blk256_height = blk256_height_y;
+		bytes_per_element = bytes_per_element_y;
+	} else {
+		blk256_width = blk256_width_c;
+		blk256_height = blk256_height_c;
+		bytes_per_element = bytes_per_element_c;
+	}
+
+	log2_bytes_per_element = dml_log2(bytes_per_element);
+
+	dml_print("DML_DLG: %s: surf_linear        = %d\n", __func__, surf_linear);
+	dml_print("DML_DLG: %s: surf_vert          = %d\n", __func__, surf_vert);
+	dml_print("DML_DLG: %s: blk256_width       = %d\n", __func__, blk256_width);
+	dml_print("DML_DLG: %s: blk256_height      = %d\n", __func__, blk256_height);
+
+	log2_blk256_width = dml_log2((double) blk256_width);
+	log2_blk256_height = dml_log2((double) blk256_height);
+	blk_bytes = surf_linear ? 256 : get_blk_size_bytes((enum source_macro_tile_size) macro_tile_size);
+	log2_blk_bytes = dml_log2((double) blk_bytes);
+
+	// remember log rule
+	// "+" in log is multiply
+	// "-" in log is divide
+	// "/2" is like square root
+	// blk is vertical biased
+	if (tiling != dm_sw_linear)
+		log2_blk_height = log2_blk256_height + dml_ceil((double) (log2_blk_bytes - 8) / 2.0, 1);
+	else
+		log2_blk_height = 0;	// blk height of 1
+
+	log2_blk_width = log2_blk_bytes - log2_bytes_per_element - log2_blk_height;
+
+	if (!surf_vert) {
+		unsigned int temp;
+
+		temp = dml_round_to_multiple(vp_width - 1, blk256_width, 1) + blk256_width;
+		if (data_pitch < blk256_width) {
+			dml_print("WARNING: DML_DLG: %s: swath_size calculation ignoring data_pitch=%u < blk256_width=%u\n", __func__, data_pitch, blk256_width);
+		} else {
+			if (temp > data_pitch) {
+				if (data_pitch >= vp_width)
+					temp = data_pitch;
+				else
+					dml_print("WARNING: DML_DLG: %s: swath_size calculation ignoring data_pitch=%u < vp_width=%u\n", __func__, data_pitch, vp_width);
+			}
+		}
+		rq_dlg_param->swath_width_ub = temp;
+		rq_dlg_param->req_per_swath_ub = temp >> log2_blk256_width;
+	} else {
+		unsigned int temp;
+
+		temp = dml_round_to_multiple(vp_height - 1, blk256_height, 1) + blk256_height;
+		if (surface_height < blk256_height) {
+			dml_print("WARNING: DML_DLG: %s swath_size calculation ignored surface_height=%u < blk256_height=%u\n", __func__, surface_height, blk256_height);
+		} else {
+			if (temp > surface_height) {
+				if (surface_height >= vp_height)
+					temp = surface_height;
+				else
+					dml_print("WARNING: DML_DLG: %s swath_size calculation ignored surface_height=%u < vp_height=%u\n", __func__, surface_height, vp_height);
+			}
+		}
+		rq_dlg_param->swath_width_ub = temp;
+		rq_dlg_param->req_per_swath_ub = temp >> log2_blk256_height;
+	}
+
+	if (!surf_vert)
+		rq_misc_param->full_swath_bytes = rq_dlg_param->swath_width_ub * blk256_height * bytes_per_element;
+	else
+		rq_misc_param->full_swath_bytes = rq_dlg_param->swath_width_ub * blk256_width * bytes_per_element;
+
+	rq_misc_param->blk256_height = blk256_height;
+	rq_misc_param->blk256_width = blk256_width;
+
+	// -------
+	// meta
+	// -------
+	log2_meta_req_bytes = 6;	// meta request is 64b and is 8x8byte meta element
+
+	// each 64b meta request for dcn is 8x8 meta elements and
+	// a meta element covers one 256b block of the data surface.
+	log2_meta_req_height = log2_blk256_height + 3;	// meta req is 8x8 byte, each byte represent 1 blk256
+	log2_meta_req_width = log2_meta_req_bytes + 8 - log2_bytes_per_element - log2_meta_req_height;
+	meta_req_width = 1 << log2_meta_req_width;
+	meta_req_height = 1 << log2_meta_req_height;
+
+	// the dimensions of a meta row are meta_row_width x meta_row_height in elements.
+	// calculate upper bound of the meta_row_width
+	if (!surf_vert) {
+		log2_meta_row_height = log2_meta_req_height;
+		meta_row_width_ub = dml_round_to_multiple(vp_width - 1, meta_req_width, 1) + meta_req_width;
+		rq_dlg_param->meta_req_per_row_ub = meta_row_width_ub / meta_req_width;
+	} else {
+		log2_meta_row_height = log2_meta_req_width;
+		meta_row_width_ub = dml_round_to_multiple(vp_height - 1, meta_req_height, 1) + meta_req_height;
+		rq_dlg_param->meta_req_per_row_ub = meta_row_width_ub / meta_req_height;
+	}
+	rq_dlg_param->meta_bytes_per_row_ub = rq_dlg_param->meta_req_per_row_ub * 64;
+
+	rq_dlg_param->meta_row_height = 1 << log2_meta_row_height;
+
+	log2_meta_chunk_bytes = dml_log2(rq_sizing_param->meta_chunk_bytes);
+	log2_meta_chunk_height = log2_meta_row_height;
+
+	//full sized meta chunk width in unit of data elements
+	log2_meta_chunk_width = log2_meta_chunk_bytes + 8 - log2_bytes_per_element - log2_meta_chunk_height;
+	log2_min_meta_chunk_bytes = dml_log2(rq_sizing_param->min_meta_chunk_bytes);
+	min_meta_chunk_width = 1 << (log2_min_meta_chunk_bytes + 8 - log2_bytes_per_element - log2_meta_chunk_height);
+	meta_chunk_width = 1 << log2_meta_chunk_width;
+	meta_chunk_per_row_int = (unsigned int) (meta_row_width_ub / meta_chunk_width);
+	meta_row_remainder = meta_row_width_ub % meta_chunk_width;
+	meta_chunk_threshold = 0;
+	meta_blk_height = blk256_height * 64;
+	meta_surface_bytes = meta_pitch * (dml_round_to_multiple(vp_height - 1, meta_blk_height, 1) + meta_blk_height) * bytes_per_element / 256;
+	vmpg_bytes = mode_lib->soc.gpuvm_min_page_size_bytes;
+	meta_pte_req_per_frame_ub = (dml_round_to_multiple(meta_surface_bytes - vmpg_bytes, 8 * vmpg_bytes, 1) + 8 * vmpg_bytes) / (8 * vmpg_bytes);
+	meta_pte_bytes_per_frame_ub = meta_pte_req_per_frame_ub * 64;	//64B mpte request
+	rq_dlg_param->meta_pte_bytes_per_frame_ub = meta_pte_bytes_per_frame_ub;
+
+	dml_print("DML_DLG: %s: meta_blk_height             = %d\n", __func__, meta_blk_height);
+	dml_print("DML_DLG: %s: meta_surface_bytes          = %d\n", __func__, meta_surface_bytes);
+	dml_print("DML_DLG: %s: meta_pte_req_per_frame_ub   = %d\n", __func__, meta_pte_req_per_frame_ub);
+	dml_print("DML_DLG: %s: meta_pte_bytes_per_frame_ub = %d\n", __func__, meta_pte_bytes_per_frame_ub);
+
+	if (!surf_vert)
+		meta_chunk_threshold = 2 * min_meta_chunk_width - meta_req_width;
+	else
+		meta_chunk_threshold = 2 * min_meta_chunk_width - meta_req_height;
+
+	if (meta_row_remainder <= meta_chunk_threshold)
+		rq_dlg_param->meta_chunks_per_row_ub = meta_chunk_per_row_int + 1;
+	else
+		rq_dlg_param->meta_chunks_per_row_ub = meta_chunk_per_row_int + 2;
+
+	// ------
+	// dpte
+	// ------
+	if (surf_linear)
+		log2_vmpg_height = 0;   // one line high
+	else
+		log2_vmpg_height = (log2_vmpg_bytes - 8) / 2 + log2_blk256_height;
+
+	log2_vmpg_width = log2_vmpg_bytes - log2_bytes_per_element - log2_vmpg_height;
+
+	// only 3 possible shapes for dpte request in dimensions of ptes: 8x1, 4x2, 2x4.
+	if (surf_linear) { //one 64B PTE request returns 8 PTEs
+		log2_dpte_req_height_ptes = 0;
+		log2_dpte_req_width = log2_vmpg_width + 3;
+		log2_dpte_req_height = 0;
+	} else if (log2_blk_bytes == 12) { //4KB tile means 4kB page size
+		//one 64B req gives 8x1 PTEs for 4KB tile
+		log2_dpte_req_height_ptes = 0;
+		log2_dpte_req_width = log2_blk_width + 3;
+		log2_dpte_req_height = log2_blk_height + 0;
+	} else if ((log2_blk_bytes >= 16) && (log2_vmpg_bytes == 12)) { // tile block >= 64KB
+		//two 64B reqs of 2x4 PTEs give 16 PTEs to cover 64KB
+		log2_dpte_req_height_ptes = 4;
+		log2_dpte_req_width = log2_blk256_width + 4;		// log2_64KB_width
+		log2_dpte_req_height = log2_blk256_height + 4;		// log2_64KB_height
+	} else { //64KB page size and must 64KB tile block
+		 //one 64B req gives 8x1 PTEs for 64KB tile
+		log2_dpte_req_height_ptes = 0;
+		log2_dpte_req_width = log2_blk_width + 3;
+		log2_dpte_req_height = log2_blk_height + 0;
+	}
+
+	// The dpte request dimensions in data elements is dpte_req_width x dpte_req_height
+	// log2_vmpg_width is how much 1 pte represent, now calculating how much a 64b pte req represent
+	// That depends on the pte shape (i.e. 8x1, 4x2, 2x4)
+	//log2_dpte_req_height    = log2_vmpg_height + log2_dpte_req_height_ptes;
+	//log2_dpte_req_width     = log2_vmpg_width + log2_dpte_req_width_ptes;
+	dpte_req_height = 1 << log2_dpte_req_height;
+	dpte_req_width = 1 << log2_dpte_req_width;
+
+	// calculate pitch dpte row buffer can hold
+	// round the result down to a power of two.
+	if (surf_linear) {
+		unsigned int dpte_row_height;
+
+		log2_dpte_row_height_linear = dml_floor(dml_log2(dpte_buf_in_pte_reqs * dpte_req_width / data_pitch), 1);
+
+		dml_print("DML_DLG: %s: is_chroma                   = %d\n", __func__, is_chroma);
+		dml_print("DML_DLG: %s: dpte_buf_in_pte_reqs        = %d\n", __func__, dpte_buf_in_pte_reqs);
+		dml_print("DML_DLG: %s: log2_dpte_row_height_linear = %d\n", __func__, log2_dpte_row_height_linear);
+
+		ASSERT(log2_dpte_row_height_linear >= 3);
+
+		if (log2_dpte_row_height_linear > 7)
+			log2_dpte_row_height_linear = 7;
+
+		log2_dpte_row_height = log2_dpte_row_height_linear;
+		// For linear, the dpte row is pitch dependent and the pte requests wrap at the pitch boundary.
+		// the dpte_row_width_ub is the upper bound of data_pitch*dpte_row_height in elements with this unique buffering.
+		dpte_row_height = 1 << log2_dpte_row_height;
+		dpte_row_width_ub = dml_round_to_multiple(data_pitch * dpte_row_height - 1, dpte_req_width, 1) + dpte_req_width;
+		rq_dlg_param->dpte_req_per_row_ub = dpte_row_width_ub / dpte_req_width;
+	} else {
+		// the upper bound of the dpte_row_width without dependency on viewport position follows.
+		// for tiled mode, row height is the same as req height and row store up to vp size upper bound
+		if (!surf_vert) {
+			log2_dpte_row_height = log2_dpte_req_height;
+			dpte_row_width_ub = dml_round_to_multiple(vp_width - 1, dpte_req_width, 1) + dpte_req_width;
+			rq_dlg_param->dpte_req_per_row_ub = dpte_row_width_ub / dpte_req_width;
+		} else {
+			log2_dpte_row_height = (log2_blk_width < log2_dpte_req_width) ? log2_blk_width : log2_dpte_req_width;
+			dpte_row_width_ub = dml_round_to_multiple(vp_height - 1, dpte_req_height, 1) + dpte_req_height;
+			rq_dlg_param->dpte_req_per_row_ub = dpte_row_width_ub / dpte_req_height;
+		}
+	}
+	if (log2_blk_bytes >= 16 && log2_vmpg_bytes == 12) // tile block >= 64KB
+		rq_dlg_param->dpte_bytes_per_row_ub = rq_dlg_param->dpte_req_per_row_ub * 128; //2*64B dpte request
+	else
+		rq_dlg_param->dpte_bytes_per_row_ub = rq_dlg_param->dpte_req_per_row_ub * 64; //64B dpte request
+
+	rq_dlg_param->dpte_row_height = 1 << log2_dpte_row_height;
+
+	// the dpte_group_bytes is reduced for the specific case of vertical
+	// access of a tile surface that has dpte request of 8x1 ptes.
+	if (hostvm_enable)
+		rq_sizing_param->dpte_group_bytes = 512;
+	else {
+		if (!surf_linear && (log2_dpte_req_height_ptes == 0) && surf_vert) //reduced, in this case, will have page fault within a group
+			rq_sizing_param->dpte_group_bytes = 512;
+		else
+			rq_sizing_param->dpte_group_bytes = 2048;
+	}
+
+	//since pte request size is 64byte, the number of data pte requests per full sized group is as follows.
+	log2_dpte_group_bytes = dml_log2(rq_sizing_param->dpte_group_bytes);
+	log2_dpte_group_length = log2_dpte_group_bytes - 6; //length in 64b requests
+
+	// full sized data pte group width in elements
+	if (!surf_vert)
+		log2_dpte_group_width = log2_dpte_group_length + log2_dpte_req_width;
+	else
+		log2_dpte_group_width = log2_dpte_group_length + log2_dpte_req_height;
+
+	//But if the tile block >=64KB and the page size is 4KB, then each dPTE request is 2*64B
+	if ((log2_blk_bytes >= 16) && (log2_vmpg_bytes == 12)) // tile block >= 64KB
+		log2_dpte_group_width = log2_dpte_group_width - 1;
+
+	dpte_group_width = 1 << log2_dpte_group_width;
+
+	// since dpte groups are only aligned to dpte_req_width and not dpte_group_width,
+	// the upper bound for the dpte groups per row is as follows.
+	rq_dlg_param->dpte_groups_per_row_ub = dml_ceil((double) dpte_row_width_ub / dpte_group_width, 1);
+}
+
+static void get_surf_rq_param(
+		struct display_mode_lib *mode_lib,
+		display_data_rq_sizing_params_st *rq_sizing_param,
+		display_data_rq_dlg_params_st *rq_dlg_param,
+		display_data_rq_misc_params_st *rq_misc_param,
+		const display_pipe_params_st *pipe_param,
+		bool is_chroma,
+		bool is_alpha)
+{
+	unsigned int vp_width = 0;
+	unsigned int vp_height = 0;
+	unsigned int data_pitch = 0;
+	unsigned int meta_pitch = 0;
+	unsigned int surface_height = 0;
+	unsigned int ppe = 1;
+
+	// FIXME check if ppe apply for both luma and chroma in 422 case
+	if (is_chroma | is_alpha) {
+		vp_width = pipe_param->src.viewport_width_c / ppe;
+		vp_height = pipe_param->src.viewport_height_c;
+		data_pitch = pipe_param->src.data_pitch_c;
+		meta_pitch = pipe_param->src.meta_pitch_c;
+		surface_height = pipe_param->src.surface_height_y / 2.0;
+	} else {
+		vp_width = pipe_param->src.viewport_width / ppe;
+		vp_height = pipe_param->src.viewport_height;
+		data_pitch = pipe_param->src.data_pitch;
+		meta_pitch = pipe_param->src.meta_pitch;
+		surface_height = pipe_param->src.surface_height_y;
+	}
+
+	if (pipe_param->dest.odm_combine) {
+		unsigned int access_dir;
+		unsigned int full_src_vp_width;
+		unsigned int hactive_odm;
+		unsigned int src_hactive_odm;
+
+		access_dir = (pipe_param->src.source_scan == dm_vert); // vp access direction: horizontal or vertical accessed
+		hactive_odm = pipe_param->dest.hactive / ((unsigned int) pipe_param->dest.odm_combine * 2);
+		if (is_chroma) {
+			full_src_vp_width = pipe_param->scale_ratio_depth.hscl_ratio_c * pipe_param->dest.full_recout_width;
+			src_hactive_odm = pipe_param->scale_ratio_depth.hscl_ratio_c * hactive_odm;
+		} else {
+			full_src_vp_width = pipe_param->scale_ratio_depth.hscl_ratio * pipe_param->dest.full_recout_width;
+			src_hactive_odm = pipe_param->scale_ratio_depth.hscl_ratio * hactive_odm;
+		}
+
+		if (access_dir == 0) {
+			vp_width = dml_min(full_src_vp_width, src_hactive_odm);
+			dml_print("DML_DLG: %s: vp_width = %d\n", __func__, vp_width);
+		} else {
+			vp_height = dml_min(full_src_vp_width, src_hactive_odm);
+			dml_print("DML_DLG: %s: vp_height = %d\n", __func__, vp_height);
+
+		}
+		dml_print("DML_DLG: %s: full_src_vp_width = %d\n", __func__, full_src_vp_width);
+		dml_print("DML_DLG: %s: hactive_odm = %d\n", __func__, hactive_odm);
+		dml_print("DML_DLG: %s: src_hactive_odm = %d\n", __func__, src_hactive_odm);
+	}
+
+	rq_sizing_param->chunk_bytes = 8192;
+
+	if (is_alpha)
+		rq_sizing_param->chunk_bytes = 4096;
+
+	if (rq_sizing_param->chunk_bytes == 64 * 1024)
+		rq_sizing_param->min_chunk_bytes = 0;
+	else
+		rq_sizing_param->min_chunk_bytes = 1024;
+
+	rq_sizing_param->meta_chunk_bytes = 2048;
+	rq_sizing_param->min_meta_chunk_bytes = 256;
+
+	if (pipe_param->src.hostvm)
+		rq_sizing_param->mpte_group_bytes = 512;
+	else
+		rq_sizing_param->mpte_group_bytes = 2048;
+
+	get_meta_and_pte_attr(
+			mode_lib,
+			rq_dlg_param,
+			rq_misc_param,
+			rq_sizing_param,
+			vp_width,
+			vp_height,
+			data_pitch,
+			meta_pitch,
+			pipe_param->src.source_format,
+			pipe_param->src.sw_mode,
+			pipe_param->src.macro_tile_size,
+			pipe_param->src.source_scan,
+			pipe_param->src.hostvm,
+			is_chroma,
+			surface_height);
+}
+
+static void dml_rq_dlg_get_rq_params(struct display_mode_lib *mode_lib, display_rq_params_st *rq_param, const display_pipe_params_st *pipe_param)
+{
+	// get param for luma surface
+	rq_param->yuv420 = pipe_param->src.source_format == dm_420_8 || pipe_param->src.source_format == dm_420_10 || pipe_param->src.source_format == dm_rgbe_alpha
+			|| pipe_param->src.source_format == dm_420_12;
+
+	rq_param->yuv420_10bpc = pipe_param->src.source_format == dm_420_10;
+
+	rq_param->rgbe_alpha = (pipe_param->src.source_format == dm_rgbe_alpha) ? 1 : 0;
+
+	get_surf_rq_param(mode_lib, &(rq_param->sizing.rq_l), &(rq_param->dlg.rq_l), &(rq_param->misc.rq_l), pipe_param, 0, 0);
+
+	if (is_dual_plane((enum source_format_class) (pipe_param->src.source_format))) {
+		// get param for chroma surface
+		get_surf_rq_param(mode_lib, &(rq_param->sizing.rq_c), &(rq_param->dlg.rq_c), &(rq_param->misc.rq_c), pipe_param, 1, rq_param->rgbe_alpha);
+	}
+
+	// calculate how to split the det buffer space between luma and chroma
+	handle_det_buf_split(mode_lib, rq_param, &pipe_param->src);
+	print__rq_params_st(mode_lib, rq_param);
+}
+
+void dml314_rq_dlg_get_rq_reg(struct display_mode_lib *mode_lib, display_rq_regs_st *rq_regs, const display_pipe_params_st *pipe_param)
+{
+	display_rq_params_st rq_param = {0};
+
+	memset(rq_regs, 0, sizeof(*rq_regs));
+	dml_rq_dlg_get_rq_params(mode_lib, &rq_param, pipe_param);
+	extract_rq_regs(mode_lib, rq_regs, &rq_param);
+
+	print__rq_regs_st(mode_lib, rq_regs);
+}
+
+static void calculate_ttu_cursor(
+		struct display_mode_lib *mode_lib,
+		double *refcyc_per_req_delivery_pre_cur,
+		double *refcyc_per_req_delivery_cur,
+		double refclk_freq_in_mhz,
+		double ref_freq_to_pix_freq,
+		double hscale_pixel_rate_l,
+		double hscl_ratio,
+		double vratio_pre_l,
+		double vratio_l,
+		unsigned int cur_width,
+		enum cursor_bpp cur_bpp)
+{
+	unsigned int cur_src_width = cur_width;
+	unsigned int cur_req_size = 0;
+	unsigned int cur_req_width = 0;
+	double cur_width_ub = 0.0;
+	double cur_req_per_width = 0.0;
+	double hactive_cur = 0.0;
+
+	ASSERT(cur_src_width <= 256);
+
+	*refcyc_per_req_delivery_pre_cur = 0.0;
+	*refcyc_per_req_delivery_cur = 0.0;
+	if (cur_src_width > 0) {
+		unsigned int cur_bit_per_pixel = 0;
+
+		if (cur_bpp == dm_cur_2bit) {
+			cur_req_size = 64; // byte
+			cur_bit_per_pixel = 2;
+		} else { // 32bit
+			cur_bit_per_pixel = 32;
+			if (cur_src_width >= 1 && cur_src_width <= 16)
+				cur_req_size = 64;
+			else if (cur_src_width >= 17 && cur_src_width <= 31)
+				cur_req_size = 128;
+			else
+				cur_req_size = 256;
+		}
+
+		cur_req_width = (double) cur_req_size / ((double) cur_bit_per_pixel / 8.0);
+		cur_width_ub = dml_ceil((double) cur_src_width / (double) cur_req_width, 1) * (double) cur_req_width;
+		cur_req_per_width = cur_width_ub / (double) cur_req_width;
+		hactive_cur = (double) cur_src_width / hscl_ratio; // FIXME: oswin to think about what to do for cursor
+
+		if (vratio_pre_l <= 1.0)
+			*refcyc_per_req_delivery_pre_cur = hactive_cur * ref_freq_to_pix_freq / (double) cur_req_per_width;
+		else
+			*refcyc_per_req_delivery_pre_cur = (double) refclk_freq_in_mhz * (double) cur_src_width / hscale_pixel_rate_l / (double) cur_req_per_width;
+
+		ASSERT(*refcyc_per_req_delivery_pre_cur < dml_pow(2, 13));
+
+		if (vratio_l <= 1.0)
+			*refcyc_per_req_delivery_cur = hactive_cur * ref_freq_to_pix_freq / (double) cur_req_per_width;
+		else
+			*refcyc_per_req_delivery_cur = (double) refclk_freq_in_mhz * (double) cur_src_width / hscale_pixel_rate_l / (double) cur_req_per_width;
+
+		dml_print("DML_DLG: %s: cur_req_width                     = %d\n", __func__, cur_req_width);
+		dml_print("DML_DLG: %s: cur_width_ub                      = %3.2f\n", __func__, cur_width_ub);
+		dml_print("DML_DLG: %s: cur_req_per_width                 = %3.2f\n", __func__, cur_req_per_width);
+		dml_print("DML_DLG: %s: hactive_cur                       = %3.2f\n", __func__, hactive_cur);
+		dml_print("DML_DLG: %s: refcyc_per_req_delivery_pre_cur   = %3.2f\n", __func__, *refcyc_per_req_delivery_pre_cur);
+		dml_print("DML_DLG: %s: refcyc_per_req_delivery_cur       = %3.2f\n", __func__, *refcyc_per_req_delivery_cur);
+
+		ASSERT(*refcyc_per_req_delivery_cur < dml_pow(2, 13));
+	}
+}
+
+// Note: currently taken in as is.
+// Nice to decouple code from hw register implement and extract code that are repeated for luma and chroma.
+static void dml_rq_dlg_get_dlg_params(
+		struct display_mode_lib *mode_lib,
+		const display_e2e_pipe_params_st *e2e_pipe_param,
+		const unsigned int num_pipes,
+		const unsigned int pipe_idx,
+		display_dlg_regs_st *disp_dlg_regs,
+		display_ttu_regs_st *disp_ttu_regs,
+		const display_rq_dlg_params_st *rq_dlg_param,
+		const display_dlg_sys_params_st *dlg_sys_param,
+		const bool cstate_en,
+		const bool pstate_en,
+		const bool vm_en,
+		const bool ignore_viewport_pos,
+		const bool immediate_flip_support)
+{
+	const display_pipe_source_params_st *src = &e2e_pipe_param[pipe_idx].pipe.src;
+	const display_pipe_dest_params_st *dst = &e2e_pipe_param[pipe_idx].pipe.dest;
+	const display_clocks_and_cfg_st *clks = &e2e_pipe_param[pipe_idx].clks_cfg;
+	const scaler_ratio_depth_st *scl = &e2e_pipe_param[pipe_idx].pipe.scale_ratio_depth;
+	const scaler_taps_st *taps = &e2e_pipe_param[pipe_idx].pipe.scale_taps;
+	unsigned int pipe_index_in_combine[DC__NUM_PIPES__MAX];
+
+	// -------------------------
+	// Section 1.15.2.1: OTG dependent Params
+	// -------------------------
+	// Timing
+	unsigned int htotal = dst->htotal;
+	unsigned int hblank_end = dst->hblank_end;
+	unsigned int vblank_start = dst->vblank_start;
+	unsigned int vblank_end = dst->vblank_end;
+
+	double dppclk_freq_in_mhz = clks->dppclk_mhz;
+	double refclk_freq_in_mhz = clks->refclk_mhz;
+	double pclk_freq_in_mhz = dst->pixel_rate_mhz;
+	bool interlaced = dst->interlaced;
+	double ref_freq_to_pix_freq = refclk_freq_in_mhz / pclk_freq_in_mhz;
+	double min_ttu_vblank;
+	unsigned int dlg_vblank_start;
+	bool dual_plane;
+	unsigned int access_dir;
+	unsigned int vp_height_l;
+	unsigned int vp_width_l;
+	unsigned int vp_height_c;
+	unsigned int vp_width_c;
+
+	// Scaling
+	unsigned int htaps_l;
+	unsigned int htaps_c;
+	double hratio_l;
+	double hratio_c;
+	double vratio_l;
+	double vratio_c;
+
+	unsigned int swath_width_ub_l;
+	unsigned int dpte_groups_per_row_ub_l;
+	unsigned int swath_width_ub_c;
+	unsigned int dpte_groups_per_row_ub_c;
+
+	unsigned int meta_chunks_per_row_ub_l;
+	unsigned int meta_chunks_per_row_ub_c;
+	unsigned int vupdate_offset;
+	unsigned int vupdate_width;
+	unsigned int vready_offset;
+
+	unsigned int vstartup_start;
+	unsigned int dst_x_after_scaler;
+	unsigned int dst_y_after_scaler;
+	double dst_y_prefetch;
+	double dst_y_per_vm_vblank;
+	double dst_y_per_row_vblank;
+	double dst_y_per_vm_flip;
+	double dst_y_per_row_flip;
+	double max_dst_y_per_vm_vblank;
+	double max_dst_y_per_row_vblank;
+	double vratio_pre_l;
+	double vratio_pre_c;
+	unsigned int req_per_swath_ub_l;
+	unsigned int req_per_swath_ub_c;
+	unsigned int meta_row_height_l;
+	unsigned int meta_row_height_c;
+	unsigned int swath_width_pixels_ub_l;
+	unsigned int swath_width_pixels_ub_c;
+	unsigned int scaler_rec_in_width_l;
+	unsigned int scaler_rec_in_width_c;
+	unsigned int dpte_row_height_l;
+	unsigned int dpte_row_height_c;
+	double hscale_pixel_rate_l;
+	double hscale_pixel_rate_c;
+	double min_hratio_fact_l;
+	double min_hratio_fact_c;
+	double refcyc_per_line_delivery_pre_l;
+	double refcyc_per_line_delivery_pre_c;
+	double refcyc_per_line_delivery_l;
+	double refcyc_per_line_delivery_c;
+
+	double refcyc_per_req_delivery_pre_l;
+	double refcyc_per_req_delivery_pre_c;
+	double refcyc_per_req_delivery_l;
+	double refcyc_per_req_delivery_c;
+
+	unsigned int full_recout_width;
+	double refcyc_per_req_delivery_pre_cur0;
+	double refcyc_per_req_delivery_cur0;
+	double refcyc_per_req_delivery_pre_cur1;
+	double refcyc_per_req_delivery_cur1;
+	unsigned int vba__min_dst_y_next_start = get_min_dst_y_next_start(mode_lib, e2e_pipe_param, num_pipes, pipe_idx); // FROM VBA
+	unsigned int vba__vready_after_vcount0 = get_vready_at_or_after_vsync(mode_lib, e2e_pipe_param, num_pipes, pipe_idx); // From VBA
+
+	float vba__refcyc_per_line_delivery_pre_l = get_refcyc_per_line_delivery_pre_l_in_us(mode_lib, e2e_pipe_param, num_pipes, pipe_idx) * refclk_freq_in_mhz; // From VBA
+	float vba__refcyc_per_line_delivery_l = get_refcyc_per_line_delivery_l_in_us(mode_lib, e2e_pipe_param, num_pipes, pipe_idx) * refclk_freq_in_mhz; // From VBA
+
+	float vba__refcyc_per_req_delivery_pre_l = get_refcyc_per_req_delivery_pre_l_in_us(mode_lib, e2e_pipe_param, num_pipes, pipe_idx) * refclk_freq_in_mhz;  // From VBA
+	float vba__refcyc_per_req_delivery_l = get_refcyc_per_req_delivery_l_in_us(mode_lib, e2e_pipe_param, num_pipes, pipe_idx) * refclk_freq_in_mhz;  // From VBA
+
+	memset(disp_dlg_regs, 0, sizeof(*disp_dlg_regs));
+	memset(disp_ttu_regs, 0, sizeof(*disp_ttu_regs));
+
+	dml_print("DML_DLG: %s: cstate_en = %d\n", __func__, cstate_en);
+	dml_print("DML_DLG: %s: pstate_en = %d\n", __func__, pstate_en);
+	dml_print("DML_DLG: %s: vm_en     = %d\n", __func__, vm_en);
+	dml_print("DML_DLG: %s: ignore_viewport_pos  = %d\n", __func__, ignore_viewport_pos);
+	dml_print("DML_DLG: %s: immediate_flip_support  = %d\n", __func__, immediate_flip_support);
+
+	dml_print("DML_DLG: %s: dppclk_freq_in_mhz     = %3.2f\n", __func__, dppclk_freq_in_mhz);
+	dml_print("DML_DLG: %s: refclk_freq_in_mhz     = %3.2f\n", __func__, refclk_freq_in_mhz);
+	dml_print("DML_DLG: %s: pclk_freq_in_mhz       = %3.2f\n", __func__, pclk_freq_in_mhz);
+	dml_print("DML_DLG: %s: interlaced             = %d\n", __func__, interlaced); ASSERT(ref_freq_to_pix_freq < 4.0);
+
+	disp_dlg_regs->ref_freq_to_pix_freq = (unsigned int) (ref_freq_to_pix_freq * dml_pow(2, 19));
+	disp_dlg_regs->refcyc_per_htotal = (unsigned int) (ref_freq_to_pix_freq * (double) htotal * dml_pow(2, 8));
+	disp_dlg_regs->dlg_vblank_end = interlaced ? (vblank_end / 2) : vblank_end;	// 15 bits
+
+	//set_prefetch_mode(mode_lib, cstate_en, pstate_en, ignore_viewport_pos, immediate_flip_support);
+	min_ttu_vblank = get_min_ttu_vblank_in_us(mode_lib, e2e_pipe_param, num_pipes, pipe_idx);	// From VBA
+
+	dlg_vblank_start = interlaced ? (vblank_start / 2) : vblank_start;
+	disp_dlg_regs->min_dst_y_next_start_us =
+		(vba__min_dst_y_next_start * dst->hactive) / (unsigned int) dst->pixel_rate_mhz;
+	disp_dlg_regs->min_dst_y_next_start = vba__min_dst_y_next_start * dml_pow(2, 2);
+
+	ASSERT(disp_dlg_regs->min_dst_y_next_start < (unsigned int)dml_pow(2, 18));
+
+	dml_print("DML_DLG: %s: min_ttu_vblank (us)         = %3.2f\n", __func__, min_ttu_vblank);
+	dml_print("DML_DLG: %s: min_dst_y_next_start        = 0x%0x\n", __func__, disp_dlg_regs->min_dst_y_next_start);
+	dml_print("DML_DLG: %s: dlg_vblank_start            = 0x%0x\n", __func__, dlg_vblank_start);
+	dml_print("DML_DLG: %s: ref_freq_to_pix_freq        = %3.2f\n", __func__, ref_freq_to_pix_freq);
+	dml_print("DML_DLG: %s: vba__min_dst_y_next_start   = 0x%0x\n", __func__, vba__min_dst_y_next_start);
+
+	//old_impl_vs_vba_impl("min_dst_y_next_start", dlg_vblank_start, vba__min_dst_y_next_start);
+
+	// -------------------------
+	// Section 1.15.2.2: Prefetch, Active and TTU
+	// -------------------------
+	// Prefetch Calc
+	// Source
+	dual_plane = is_dual_plane((enum source_format_class) (src->source_format));
+	access_dir = (src->source_scan == dm_vert);	// vp access direction: horizontal or vertical accessed
+	vp_height_l = src->viewport_height;
+	vp_width_l = src->viewport_width;
+	vp_height_c = src->viewport_height_c;
+	vp_width_c = src->viewport_width_c;
+
+	// Scaling
+	htaps_l = taps->htaps;
+	htaps_c = taps->htaps_c;
+	hratio_l = scl->hscl_ratio;
+	hratio_c = scl->hscl_ratio_c;
+	vratio_l = scl->vscl_ratio;
+	vratio_c = scl->vscl_ratio_c;
+
+	swath_width_ub_l = rq_dlg_param->rq_l.swath_width_ub;
+	dpte_groups_per_row_ub_l = rq_dlg_param->rq_l.dpte_groups_per_row_ub;
+	swath_width_ub_c = rq_dlg_param->rq_c.swath_width_ub;
+	dpte_groups_per_row_ub_c = rq_dlg_param->rq_c.dpte_groups_per_row_ub;
+
+	meta_chunks_per_row_ub_l = rq_dlg_param->rq_l.meta_chunks_per_row_ub;
+	meta_chunks_per_row_ub_c = rq_dlg_param->rq_c.meta_chunks_per_row_ub;
+	vupdate_offset = dst->vupdate_offset;
+	vupdate_width = dst->vupdate_width;
+	vready_offset = dst->vready_offset;
+
+	vstartup_start = dst->vstartup_start;
+	if (interlaced) {
+		if (vstartup_start / 2.0 - (double) (vready_offset + vupdate_width + vupdate_offset) / htotal <= vblank_end / 2.0)
+			disp_dlg_regs->vready_after_vcount0 = 1;
+		else
+			disp_dlg_regs->vready_after_vcount0 = 0;
+	} else {
+		if (vstartup_start - (double) (vready_offset + vupdate_width + vupdate_offset) / htotal <= vblank_end)
+			disp_dlg_regs->vready_after_vcount0 = 1;
+		else
+			disp_dlg_regs->vready_after_vcount0 = 0;
+	}
+
+	dml_print("DML_DLG: %s: vready_after_vcount0 = %d\n", __func__, disp_dlg_regs->vready_after_vcount0);
+	dml_print("DML_DLG: %s: vba__vready_after_vcount0 = %d\n", __func__, vba__vready_after_vcount0);
+	//old_impl_vs_vba_impl("vready_after_vcount0", disp_dlg_regs->vready_after_vcount0, vba__vready_after_vcount0);
+
+	if (interlaced)
+		vstartup_start = vstartup_start / 2;
+
+	dst_x_after_scaler = get_dst_x_after_scaler(mode_lib, e2e_pipe_param, num_pipes, pipe_idx); // From VBA
+	dst_y_after_scaler = get_dst_y_after_scaler(mode_lib, e2e_pipe_param, num_pipes, pipe_idx); // From VBA
+
+	// do some adjustment on the dst_after scaler to account for odm combine mode
+	dml_print("DML_DLG: %s: input dst_x_after_scaler   = %d\n", __func__, dst_x_after_scaler);
+	dml_print("DML_DLG: %s: input dst_y_after_scaler   = %d\n", __func__, dst_y_after_scaler);
+
+	// need to figure out which side of odm combine we're in
+	if (dst->odm_combine) {
+		// figure out which pipes go together
+		bool visited[DC__NUM_PIPES__MAX];
+		unsigned int i, j, k;
+
+		for (k = 0; k < num_pipes; ++k) {
+			visited[k] = false;
+			pipe_index_in_combine[k] = 0;
+		}
+
+		for (i = 0; i < num_pipes; i++) {
+			if (e2e_pipe_param[i].pipe.src.is_hsplit && !visited[i]) {
+
+				unsigned int grp = e2e_pipe_param[i].pipe.src.hsplit_grp;
+				unsigned int grp_idx = 0;
+
+				for (j = i; j < num_pipes; j++) {
+					if (e2e_pipe_param[j].pipe.src.hsplit_grp == grp && e2e_pipe_param[j].pipe.src.is_hsplit && !visited[j]) {
+						pipe_index_in_combine[j] = grp_idx;
+						dml_print("DML_DLG: %s: pipe[%d] is in grp %d idx %d\n", __func__, j, grp, grp_idx);
+						grp_idx++;
+						visited[j] = true;
+					}
+				}
+			}
+		}
+
+	}
+
+	if (dst->odm_combine == dm_odm_combine_mode_disabled) {
+		disp_dlg_regs->refcyc_h_blank_end = (unsigned int) ((double) hblank_end * ref_freq_to_pix_freq);
+	} else {
+		unsigned int odm_combine_factor = (dst->odm_combine == dm_odm_combine_mode_2to1 ? 2 : 4); // TODO: We should really check that 4to1 is supported before setting it to 4
+		unsigned int odm_pipe_index = pipe_index_in_combine[pipe_idx];
+
+		disp_dlg_regs->refcyc_h_blank_end = (unsigned int) (((double) hblank_end + odm_pipe_index * (double) dst->hactive / odm_combine_factor) * ref_freq_to_pix_freq);
+	} ASSERT(disp_dlg_regs->refcyc_h_blank_end < (unsigned int)dml_pow(2, 13));
+
+	dml_print("DML_DLG: %s: htotal                     = %d\n", __func__, htotal);
+	dml_print("DML_DLG: %s: dst_x_after_scaler[%d]     = %d\n", __func__, pipe_idx, dst_x_after_scaler);
+	dml_print("DML_DLG: %s: dst_y_after_scaler[%d]     = %d\n", __func__, pipe_idx, dst_y_after_scaler);
+
+	dst_y_prefetch = get_dst_y_prefetch(mode_lib, e2e_pipe_param, num_pipes, pipe_idx);        // From VBA
+	dst_y_per_vm_vblank = get_dst_y_per_vm_vblank(mode_lib, e2e_pipe_param, num_pipes, pipe_idx);        // From VBA
+	dst_y_per_row_vblank = get_dst_y_per_row_vblank(mode_lib, e2e_pipe_param, num_pipes, pipe_idx);        // From VBA
+	dst_y_per_vm_flip = get_dst_y_per_vm_flip(mode_lib, e2e_pipe_param, num_pipes, pipe_idx);        // From VBA
+	dst_y_per_row_flip = get_dst_y_per_row_flip(mode_lib, e2e_pipe_param, num_pipes, pipe_idx);        // From VBA
+
+	max_dst_y_per_vm_vblank = 32.0;        //U5.2
+	max_dst_y_per_row_vblank = 16.0;        //U4.2
+
+	// magic!
+	if (htotal <= 75) {
+		max_dst_y_per_vm_vblank = 100.0;
+		max_dst_y_per_row_vblank = 100.0;
+	}
+
+	dml_print("DML_DLG: %s: dst_y_prefetch (after rnd) = %3.2f\n", __func__, dst_y_prefetch);
+	dml_print("DML_DLG: %s: dst_y_per_vm_flip    = %3.2f\n", __func__, dst_y_per_vm_flip);
+	dml_print("DML_DLG: %s: dst_y_per_row_flip   = %3.2f\n", __func__, dst_y_per_row_flip);
+	dml_print("DML_DLG: %s: dst_y_per_vm_vblank  = %3.2f\n", __func__, dst_y_per_vm_vblank);
+	dml_print("DML_DLG: %s: dst_y_per_row_vblank = %3.2f\n", __func__, dst_y_per_row_vblank);
+
+	ASSERT(dst_y_per_vm_vblank < max_dst_y_per_vm_vblank); ASSERT(dst_y_per_row_vblank < max_dst_y_per_row_vblank);
+
+	ASSERT(dst_y_prefetch > (dst_y_per_vm_vblank + dst_y_per_row_vblank));
+
+	vratio_pre_l = get_vratio_prefetch_l(mode_lib, e2e_pipe_param, num_pipes, pipe_idx);    // From VBA
+	vratio_pre_c = get_vratio_prefetch_c(mode_lib, e2e_pipe_param, num_pipes, pipe_idx);    // From VBA
+
+	dml_print("DML_DLG: %s: vratio_pre_l = %3.2f\n", __func__, vratio_pre_l);
+	dml_print("DML_DLG: %s: vratio_pre_c = %3.2f\n", __func__, vratio_pre_c);
+
+	// Active
+	req_per_swath_ub_l = rq_dlg_param->rq_l.req_per_swath_ub;
+	req_per_swath_ub_c = rq_dlg_param->rq_c.req_per_swath_ub;
+	meta_row_height_l = rq_dlg_param->rq_l.meta_row_height;
+	meta_row_height_c = rq_dlg_param->rq_c.meta_row_height;
+	swath_width_pixels_ub_l = 0;
+	swath_width_pixels_ub_c = 0;
+	scaler_rec_in_width_l = 0;
+	scaler_rec_in_width_c = 0;
+	dpte_row_height_l = rq_dlg_param->rq_l.dpte_row_height;
+	dpte_row_height_c = rq_dlg_param->rq_c.dpte_row_height;
+
+	swath_width_pixels_ub_l = swath_width_ub_l;
+	swath_width_pixels_ub_c = swath_width_ub_c;
+
+	if (hratio_l <= 1)
+		min_hratio_fact_l = 2.0;
+	else if (htaps_l <= 6) {
+		if ((hratio_l * 2.0) > 4.0)
+			min_hratio_fact_l = 4.0;
+		else
+			min_hratio_fact_l = hratio_l * 2.0;
+	} else {
+		if (hratio_l > 4.0)
+			min_hratio_fact_l = 4.0;
+		else
+			min_hratio_fact_l = hratio_l;
+	}
+
+	hscale_pixel_rate_l = min_hratio_fact_l * dppclk_freq_in_mhz;
+
+	dml_print("DML_DLG: %s: hratio_l = %3.2f\n", __func__, hratio_l);
+	dml_print("DML_DLG: %s: min_hratio_fact_l = %3.2f\n", __func__, min_hratio_fact_l);
+	dml_print("DML_DLG: %s: hscale_pixel_rate_l = %3.2f\n", __func__, hscale_pixel_rate_l);
+
+	if (hratio_c <= 1)
+		min_hratio_fact_c = 2.0;
+	else if (htaps_c <= 6) {
+		if ((hratio_c * 2.0) > 4.0)
+			min_hratio_fact_c = 4.0;
+		else
+			min_hratio_fact_c = hratio_c * 2.0;
+	} else {
+		if (hratio_c > 4.0)
+			min_hratio_fact_c = 4.0;
+		else
+			min_hratio_fact_c = hratio_c;
+	}
+
+	hscale_pixel_rate_c = min_hratio_fact_c * dppclk_freq_in_mhz;
+
+	refcyc_per_line_delivery_pre_l = 0.;
+	refcyc_per_line_delivery_pre_c = 0.;
+	refcyc_per_line_delivery_l = 0.;
+	refcyc_per_line_delivery_c = 0.;
+
+	refcyc_per_req_delivery_pre_l = 0.;
+	refcyc_per_req_delivery_pre_c = 0.;
+	refcyc_per_req_delivery_l = 0.;
+	refcyc_per_req_delivery_c = 0.;
+
+	full_recout_width = 0;
+	// In ODM
+	if (src->is_hsplit) {
+		// This "hack"  is only allowed (and valid) for MPC combine. In ODM
+		// combine, you MUST specify the full_recout_width...according to Oswin
+		if (dst->full_recout_width == 0 && !dst->odm_combine) {
+			dml_print("DML_DLG: %s: Warning: full_recout_width not set in hsplit mode\n", __func__);
+			full_recout_width = dst->recout_width * 2; // assume half split for dcn1
+		} else
+			full_recout_width = dst->full_recout_width;
+	} else
+		full_recout_width = dst->recout_width;
+
+	// As of DCN2, mpc_combine and odm_combine are mutually exclusive
+	refcyc_per_line_delivery_pre_l = get_refcyc_per_delivery(
+			mode_lib,
+			refclk_freq_in_mhz,
+			pclk_freq_in_mhz,
+			dst->odm_combine,
+			full_recout_width,
+			dst->hactive,
+			vratio_pre_l,
+			hscale_pixel_rate_l,
+			swath_width_pixels_ub_l,
+			1); // per line
+
+	refcyc_per_line_delivery_l = get_refcyc_per_delivery(
+			mode_lib,
+			refclk_freq_in_mhz,
+			pclk_freq_in_mhz,
+			dst->odm_combine,
+			full_recout_width,
+			dst->hactive,
+			vratio_l,
+			hscale_pixel_rate_l,
+			swath_width_pixels_ub_l,
+			1); // per line
+
+	dml_print("DML_DLG: %s: full_recout_width              = %d\n", __func__, full_recout_width);
+	dml_print("DML_DLG: %s: hscale_pixel_rate_l            = %3.2f\n", __func__, hscale_pixel_rate_l);
+	dml_print("DML_DLG: %s: refcyc_per_line_delivery_pre_l = %3.2f\n", __func__, refcyc_per_line_delivery_pre_l);
+	dml_print("DML_DLG: %s: refcyc_per_line_delivery_l     = %3.2f\n", __func__, refcyc_per_line_delivery_l);
+	dml_print("DML_DLG: %s: vba__refcyc_per_line_delivery_pre_l = %3.2f\n", __func__, vba__refcyc_per_line_delivery_pre_l);
+	dml_print("DML_DLG: %s: vba__refcyc_per_line_delivery_l     = %3.2f\n", __func__, vba__refcyc_per_line_delivery_l);
+
+	//old_impl_vs_vba_impl("refcyc_per_line_delivery_pre_l", refcyc_per_line_delivery_pre_l, vba__refcyc_per_line_delivery_pre_l);
+	//old_impl_vs_vba_impl("refcyc_per_line_delivery_l", refcyc_per_line_delivery_l, vba__refcyc_per_line_delivery_l);
+
+	if (dual_plane) {
+		float vba__refcyc_per_line_delivery_pre_c = get_refcyc_per_line_delivery_pre_c_in_us(mode_lib, e2e_pipe_param, num_pipes, pipe_idx) * refclk_freq_in_mhz; // From VBA
+		float vba__refcyc_per_line_delivery_c = get_refcyc_per_line_delivery_c_in_us(mode_lib, e2e_pipe_param, num_pipes, pipe_idx) * refclk_freq_in_mhz; // From VBA
+
+		refcyc_per_line_delivery_pre_c = get_refcyc_per_delivery(
+				mode_lib,
+				refclk_freq_in_mhz,
+				pclk_freq_in_mhz,
+				dst->odm_combine,
+				full_recout_width,
+				dst->hactive,
+				vratio_pre_c,
+				hscale_pixel_rate_c,
+				swath_width_pixels_ub_c,
+				1); // per line
+
+		refcyc_per_line_delivery_c = get_refcyc_per_delivery(
+				mode_lib,
+				refclk_freq_in_mhz,
+				pclk_freq_in_mhz,
+				dst->odm_combine,
+				full_recout_width,
+				dst->hactive,
+				vratio_c,
+				hscale_pixel_rate_c,
+				swath_width_pixels_ub_c,
+				1); // per line
+
+		dml_print("DML_DLG: %s: refcyc_per_line_delivery_pre_c = %3.2f\n", __func__, refcyc_per_line_delivery_pre_c);
+		dml_print("DML_DLG: %s: refcyc_per_line_delivery_c     = %3.2f\n", __func__, refcyc_per_line_delivery_c);
+		dml_print("DML_DLG: %s: vba__refcyc_per_line_delivery_pre_c = %3.2f\n", __func__, vba__refcyc_per_line_delivery_pre_c);
+		dml_print("DML_DLG: %s: vba__refcyc_per_line_delivery_c     = %3.2f\n", __func__, vba__refcyc_per_line_delivery_c);
+
+		//old_impl_vs_vba_impl("refcyc_per_line_delivery_pre_c", refcyc_per_line_delivery_pre_c, vba__refcyc_per_line_delivery_pre_c);
+		//old_impl_vs_vba_impl("refcyc_per_line_delivery_c", refcyc_per_line_delivery_c, vba__refcyc_per_line_delivery_c);
+	}
+
+	if (src->dynamic_metadata_enable && src->gpuvm)
+		disp_dlg_regs->refcyc_per_vm_dmdata = get_refcyc_per_vm_dmdata_in_us(mode_lib, e2e_pipe_param, num_pipes, pipe_idx) * refclk_freq_in_mhz; // From VBA
+
+	disp_dlg_regs->dmdata_dl_delta = get_dmdata_dl_delta_in_us(mode_lib, e2e_pipe_param, num_pipes, pipe_idx) * refclk_freq_in_mhz; // From VBA
+
+	// TTU - Luma / Chroma
+	if (access_dir) {  // vertical access
+		scaler_rec_in_width_l = vp_height_l;
+		scaler_rec_in_width_c = vp_height_c;
+	} else {
+		scaler_rec_in_width_l = vp_width_l;
+		scaler_rec_in_width_c = vp_width_c;
+	}
+
+	refcyc_per_req_delivery_pre_l = get_refcyc_per_delivery(
+			mode_lib,
+			refclk_freq_in_mhz,
+			pclk_freq_in_mhz,
+			dst->odm_combine,
+			full_recout_width,
+			dst->hactive,
+			vratio_pre_l,
+			hscale_pixel_rate_l,
+			scaler_rec_in_width_l,
+			req_per_swath_ub_l);  // per req
+
+	refcyc_per_req_delivery_l = get_refcyc_per_delivery(
+			mode_lib,
+			refclk_freq_in_mhz,
+			pclk_freq_in_mhz,
+			dst->odm_combine,
+			full_recout_width,
+			dst->hactive,
+			vratio_l,
+			hscale_pixel_rate_l,
+			scaler_rec_in_width_l,
+			req_per_swath_ub_l);  // per req
+
+	dml_print("DML_DLG: %s: refcyc_per_req_delivery_pre_l = %3.2f\n", __func__, refcyc_per_req_delivery_pre_l);
+	dml_print("DML_DLG: %s: refcyc_per_req_delivery_l     = %3.2f\n", __func__, refcyc_per_req_delivery_l);
+	dml_print("DML_DLG: %s: vba__refcyc_per_req_delivery_pre_l = %3.2f\n", __func__, vba__refcyc_per_req_delivery_pre_l);
+	dml_print("DML_DLG: %s: vba__refcyc_per_req_delivery_l     = %3.2f\n", __func__, vba__refcyc_per_req_delivery_l);
+
+	//old_impl_vs_vba_impl("refcyc_per_req_delivery_pre_l", refcyc_per_req_delivery_pre_l, vba__refcyc_per_req_delivery_pre_l);
+	//old_impl_vs_vba_impl("refcyc_per_req_delivery_l", refcyc_per_req_delivery_l, vba__refcyc_per_req_delivery_l);
+
+	ASSERT(refcyc_per_req_delivery_pre_l < dml_pow(2, 13)); ASSERT(refcyc_per_req_delivery_l < dml_pow(2, 13));
+
+	if (dual_plane) {
+		float vba__refcyc_per_req_delivery_pre_c = get_refcyc_per_req_delivery_pre_c_in_us(mode_lib, e2e_pipe_param, num_pipes, pipe_idx) * refclk_freq_in_mhz;  // From VBA
+		float vba__refcyc_per_req_delivery_c = get_refcyc_per_req_delivery_c_in_us(mode_lib, e2e_pipe_param, num_pipes, pipe_idx) * refclk_freq_in_mhz;  // From VBA
+
+		refcyc_per_req_delivery_pre_c = get_refcyc_per_delivery(
+				mode_lib,
+				refclk_freq_in_mhz,
+				pclk_freq_in_mhz,
+				dst->odm_combine,
+				full_recout_width,
+				dst->hactive,
+				vratio_pre_c,
+				hscale_pixel_rate_c,
+				scaler_rec_in_width_c,
+				req_per_swath_ub_c);  // per req
+		refcyc_per_req_delivery_c = get_refcyc_per_delivery(
+				mode_lib,
+				refclk_freq_in_mhz,
+				pclk_freq_in_mhz,
+				dst->odm_combine,
+				full_recout_width,
+				dst->hactive,
+				vratio_c,
+				hscale_pixel_rate_c,
+				scaler_rec_in_width_c,
+				req_per_swath_ub_c);  // per req
+
+		dml_print("DML_DLG: %s: refcyc_per_req_delivery_pre_c = %3.2f\n", __func__, refcyc_per_req_delivery_pre_c);
+		dml_print("DML_DLG: %s: refcyc_per_req_delivery_c     = %3.2f\n", __func__, refcyc_per_req_delivery_c);
+		dml_print("DML_DLG: %s: vba__refcyc_per_req_delivery_pre_c = %3.2f\n", __func__, vba__refcyc_per_req_delivery_pre_c);
+		dml_print("DML_DLG: %s: vba__refcyc_per_req_delivery_c     = %3.2f\n", __func__, vba__refcyc_per_req_delivery_c);
+
+		//old_impl_vs_vba_impl("refcyc_per_req_delivery_pre_c", refcyc_per_req_delivery_pre_c, vba__refcyc_per_req_delivery_pre_c);
+		//old_impl_vs_vba_impl("refcyc_per_req_delivery_c", refcyc_per_req_delivery_c, vba__refcyc_per_req_delivery_c);
+
+		ASSERT(refcyc_per_req_delivery_pre_c < dml_pow(2, 13)); ASSERT(refcyc_per_req_delivery_c < dml_pow(2, 13));
+	}
+
+	// TTU - Cursor
+	refcyc_per_req_delivery_pre_cur0 = 0.0;
+	refcyc_per_req_delivery_cur0 = 0.0;
+
+	ASSERT(src->num_cursors <= 1);
+
+	if (src->num_cursors > 0) {
+		float vba__refcyc_per_req_delivery_pre_cur0;
+		float vba__refcyc_per_req_delivery_cur0;
+
+		calculate_ttu_cursor(
+				mode_lib,
+				&refcyc_per_req_delivery_pre_cur0,
+				&refcyc_per_req_delivery_cur0,
+				refclk_freq_in_mhz,
+				ref_freq_to_pix_freq,
+				hscale_pixel_rate_l,
+				scl->hscl_ratio,
+				vratio_pre_l,
+				vratio_l,
+				src->cur0_src_width,
+				(enum cursor_bpp) (src->cur0_bpp));
+
+		vba__refcyc_per_req_delivery_pre_cur0 = get_refcyc_per_cursor_req_delivery_pre_in_us(mode_lib, e2e_pipe_param, num_pipes, pipe_idx) * refclk_freq_in_mhz; // From VBA
+		vba__refcyc_per_req_delivery_cur0 = get_refcyc_per_cursor_req_delivery_in_us(mode_lib, e2e_pipe_param, num_pipes, pipe_idx) * refclk_freq_in_mhz; // From VBA
+
+		dml_print("DML_DLG: %s: refcyc_per_req_delivery_pre_cur0 = %3.2f\n", __func__, refcyc_per_req_delivery_pre_cur0);
+		dml_print("DML_DLG: %s: refcyc_per_req_delivery_cur0     = %3.2f\n", __func__, refcyc_per_req_delivery_cur0);
+		dml_print("DML_DLG: %s: vba__refcyc_per_req_delivery_pre_cur0 = %3.2f\n", __func__, vba__refcyc_per_req_delivery_pre_cur0);
+		dml_print("DML_DLG: %s: vba__refcyc_per_req_delivery_cur0     = %3.2f\n", __func__, vba__refcyc_per_req_delivery_cur0);
+
+		//old_impl_vs_vba_impl("refcyc_per_req_delivery_pre_cur0", refcyc_per_req_delivery_pre_cur0, vba__refcyc_per_req_delivery_pre_cur0);
+		//old_impl_vs_vba_impl("refcyc_per_req_delivery_cur0", refcyc_per_req_delivery_cur0, vba__refcyc_per_req_delivery_cur0);
+	}
+
+	refcyc_per_req_delivery_pre_cur1 = 0.0;
+	refcyc_per_req_delivery_cur1 = 0.0;
+
+	// TTU - Misc
+	// all hard-coded
+
+	// Assignment to register structures
+	disp_dlg_regs->dst_y_after_scaler = dst_y_after_scaler;	// in terms of line
+	ASSERT(disp_dlg_regs->dst_y_after_scaler < 8);
+	disp_dlg_regs->refcyc_x_after_scaler = dst_x_after_scaler * ref_freq_to_pix_freq;	// in terms of refclk
+	ASSERT(disp_dlg_regs->refcyc_x_after_scaler < (unsigned int)dml_pow(2, 13));
+	disp_dlg_regs->dst_y_prefetch = (unsigned int) (dst_y_prefetch * dml_pow(2, 2));
+	disp_dlg_regs->dst_y_per_vm_vblank = (unsigned int) (dst_y_per_vm_vblank * dml_pow(2, 2));
+	disp_dlg_regs->dst_y_per_row_vblank = (unsigned int) (dst_y_per_row_vblank * dml_pow(2, 2));
+	disp_dlg_regs->dst_y_per_vm_flip = (unsigned int) (dst_y_per_vm_flip * dml_pow(2, 2));
+	disp_dlg_regs->dst_y_per_row_flip = (unsigned int) (dst_y_per_row_flip * dml_pow(2, 2));
+
+	disp_dlg_regs->vratio_prefetch = (unsigned int) (vratio_pre_l * dml_pow(2, 19));
+	disp_dlg_regs->vratio_prefetch_c = (unsigned int) (vratio_pre_c * dml_pow(2, 19));
+
+	dml_print("DML_DLG: %s: disp_dlg_regs->dst_y_per_vm_vblank  = 0x%x\n", __func__, disp_dlg_regs->dst_y_per_vm_vblank);
+	dml_print("DML_DLG: %s: disp_dlg_regs->dst_y_per_row_vblank = 0x%x\n", __func__, disp_dlg_regs->dst_y_per_row_vblank);
+	dml_print("DML_DLG: %s: disp_dlg_regs->dst_y_per_vm_flip    = 0x%x\n", __func__, disp_dlg_regs->dst_y_per_vm_flip);
+	dml_print("DML_DLG: %s: disp_dlg_regs->dst_y_per_row_flip   = 0x%x\n", __func__, disp_dlg_regs->dst_y_per_row_flip);
+
+	disp_dlg_regs->refcyc_per_pte_group_vblank_l = (unsigned int) (dst_y_per_row_vblank * (double) htotal * ref_freq_to_pix_freq / (double) dpte_groups_per_row_ub_l);
+	ASSERT(disp_dlg_regs->refcyc_per_pte_group_vblank_l < (unsigned int)dml_pow(2, 13));
+
+	if (dual_plane) {
+		disp_dlg_regs->refcyc_per_pte_group_vblank_c = (unsigned int) (dst_y_per_row_vblank * (double) htotal * ref_freq_to_pix_freq / (double) dpte_groups_per_row_ub_c);
+		ASSERT(disp_dlg_regs->refcyc_per_pte_group_vblank_c < (unsigned int)dml_pow(2, 13));
+	}
+
+	disp_dlg_regs->refcyc_per_meta_chunk_vblank_l = (unsigned int) (dst_y_per_row_vblank * (double) htotal * ref_freq_to_pix_freq / (double) meta_chunks_per_row_ub_l);
+	ASSERT(disp_dlg_regs->refcyc_per_meta_chunk_vblank_l < (unsigned int)dml_pow(2, 13));
+
+	disp_dlg_regs->refcyc_per_meta_chunk_vblank_c = disp_dlg_regs->refcyc_per_meta_chunk_vblank_l; // dcc for 4:2:0 is not supported in dcn1.0.  assigned to be the same as _l for now
+
+	disp_dlg_regs->refcyc_per_pte_group_flip_l = (unsigned int) (dst_y_per_row_flip * htotal * ref_freq_to_pix_freq) / dpte_groups_per_row_ub_l;
+	disp_dlg_regs->refcyc_per_meta_chunk_flip_l = (unsigned int) (dst_y_per_row_flip * htotal * ref_freq_to_pix_freq) / meta_chunks_per_row_ub_l;
+
+	if (dual_plane) {
+		disp_dlg_regs->refcyc_per_pte_group_flip_c = (unsigned int) (dst_y_per_row_flip * htotal * ref_freq_to_pix_freq) / dpte_groups_per_row_ub_c;
+		disp_dlg_regs->refcyc_per_meta_chunk_flip_c = (unsigned int) (dst_y_per_row_flip * htotal * ref_freq_to_pix_freq) / meta_chunks_per_row_ub_c;
+	}
+
+	disp_dlg_regs->refcyc_per_vm_group_vblank = get_refcyc_per_vm_group_vblank_in_us(mode_lib, e2e_pipe_param, num_pipes, pipe_idx) * refclk_freq_in_mhz;            // From VBA
+	disp_dlg_regs->refcyc_per_vm_group_flip = get_refcyc_per_vm_group_flip_in_us(mode_lib, e2e_pipe_param, num_pipes, pipe_idx) * refclk_freq_in_mhz;            // From VBA
+	disp_dlg_regs->refcyc_per_vm_req_vblank = get_refcyc_per_vm_req_vblank_in_us(mode_lib, e2e_pipe_param, num_pipes, pipe_idx) * refclk_freq_in_mhz * dml_pow(2, 10); // From VBA
+	disp_dlg_regs->refcyc_per_vm_req_flip = get_refcyc_per_vm_req_flip_in_us(mode_lib, e2e_pipe_param, num_pipes, pipe_idx) * refclk_freq_in_mhz * dml_pow(2, 10);   // From VBA
+
+	// Clamp to max for now
+	if (disp_dlg_regs->refcyc_per_vm_group_vblank >= (unsigned int) dml_pow(2, 23))
+		disp_dlg_regs->refcyc_per_vm_group_vblank = dml_pow(2, 23) - 1;
+
+	if (disp_dlg_regs->refcyc_per_vm_group_flip >= (unsigned int) dml_pow(2, 23))
+		disp_dlg_regs->refcyc_per_vm_group_flip = dml_pow(2, 23) - 1;
+
+	if (disp_dlg_regs->refcyc_per_vm_req_vblank >= (unsigned int) dml_pow(2, 23))
+		disp_dlg_regs->refcyc_per_vm_req_vblank = dml_pow(2, 23) - 1;
+
+	if (disp_dlg_regs->refcyc_per_vm_req_flip >= (unsigned int) dml_pow(2, 23))
+		disp_dlg_regs->refcyc_per_vm_req_flip = dml_pow(2, 23) - 1;
+
+	disp_dlg_regs->dst_y_per_pte_row_nom_l = (unsigned int) ((double) dpte_row_height_l / (double) vratio_l * dml_pow(2, 2));
+	ASSERT(disp_dlg_regs->dst_y_per_pte_row_nom_l < (unsigned int)dml_pow(2, 17));
+	if (dual_plane) {
+		disp_dlg_regs->dst_y_per_pte_row_nom_c = (unsigned int) ((double) dpte_row_height_c / (double) vratio_c * dml_pow(2, 2));
+		if (disp_dlg_regs->dst_y_per_pte_row_nom_c >= (unsigned int) dml_pow(2, 17)) {
+			dml_print(
+					"DML_DLG: %s: Warning dst_y_per_pte_row_nom_c %u larger than supported by register format U15.2 %u\n",
+					__func__,
+					disp_dlg_regs->dst_y_per_pte_row_nom_c,
+					(unsigned int) dml_pow(2, 17) - 1);
+		}
+	}
+
+	disp_dlg_regs->dst_y_per_meta_row_nom_l = (unsigned int) ((double) meta_row_height_l / (double) vratio_l * dml_pow(2, 2));
+	ASSERT(disp_dlg_regs->dst_y_per_meta_row_nom_l < (unsigned int)dml_pow(2, 17));
+
+	disp_dlg_regs->dst_y_per_meta_row_nom_c = (unsigned int) ((double) meta_row_height_c / (double) vratio_c * dml_pow(2, 2));
+	ASSERT(disp_dlg_regs->dst_y_per_meta_row_nom_c < (unsigned int)dml_pow(2, 17));
+
+	disp_dlg_regs->refcyc_per_pte_group_nom_l = (unsigned int) ((double) dpte_row_height_l / (double) vratio_l * (double) htotal * ref_freq_to_pix_freq
+			/ (double) dpte_groups_per_row_ub_l);
+	if (disp_dlg_regs->refcyc_per_pte_group_nom_l >= (unsigned int) dml_pow(2, 23))
+		disp_dlg_regs->refcyc_per_pte_group_nom_l = dml_pow(2, 23) - 1;
+	disp_dlg_regs->refcyc_per_meta_chunk_nom_l = (unsigned int) ((double) meta_row_height_l / (double) vratio_l * (double) htotal * ref_freq_to_pix_freq
+			/ (double) meta_chunks_per_row_ub_l);
+	if (disp_dlg_regs->refcyc_per_meta_chunk_nom_l >= (unsigned int) dml_pow(2, 23))
+		disp_dlg_regs->refcyc_per_meta_chunk_nom_l = dml_pow(2, 23) - 1;
+
+	if (dual_plane) {
+		disp_dlg_regs->refcyc_per_pte_group_nom_c = (unsigned int) ((double) dpte_row_height_c / (double) vratio_c * (double) htotal * ref_freq_to_pix_freq
+				/ (double) dpte_groups_per_row_ub_c);
+		if (disp_dlg_regs->refcyc_per_pte_group_nom_c >= (unsigned int) dml_pow(2, 23))
+			disp_dlg_regs->refcyc_per_pte_group_nom_c = dml_pow(2, 23) - 1;
+
+		// TODO: Is this the right calculation? Does htotal need to be halved?
+		disp_dlg_regs->refcyc_per_meta_chunk_nom_c = (unsigned int) ((double) meta_row_height_c / (double) vratio_c * (double) htotal * ref_freq_to_pix_freq
+				/ (double) meta_chunks_per_row_ub_c);
+		if (disp_dlg_regs->refcyc_per_meta_chunk_nom_c >= (unsigned int) dml_pow(2, 23))
+			disp_dlg_regs->refcyc_per_meta_chunk_nom_c = dml_pow(2, 23) - 1;
+	}
+
+	disp_dlg_regs->refcyc_per_line_delivery_pre_l = (unsigned int) dml_floor(refcyc_per_line_delivery_pre_l, 1);
+	disp_dlg_regs->refcyc_per_line_delivery_l = (unsigned int) dml_floor(refcyc_per_line_delivery_l, 1);
+	ASSERT(disp_dlg_regs->refcyc_per_line_delivery_pre_l < (unsigned int)dml_pow(2, 13)); ASSERT(disp_dlg_regs->refcyc_per_line_delivery_l < (unsigned int)dml_pow(2, 13));
+
+	disp_dlg_regs->refcyc_per_line_delivery_pre_c = (unsigned int) dml_floor(refcyc_per_line_delivery_pre_c, 1);
+	disp_dlg_regs->refcyc_per_line_delivery_c = (unsigned int) dml_floor(refcyc_per_line_delivery_c, 1);
+	ASSERT(disp_dlg_regs->refcyc_per_line_delivery_pre_c < (unsigned int)dml_pow(2, 13)); ASSERT(disp_dlg_regs->refcyc_per_line_delivery_c < (unsigned int)dml_pow(2, 13));
+
+	disp_dlg_regs->chunk_hdl_adjust_cur0 = 3;
+	disp_dlg_regs->dst_y_offset_cur0 = 0;
+	disp_dlg_regs->chunk_hdl_adjust_cur1 = 3;
+	disp_dlg_regs->dst_y_offset_cur1 = 0;
+
+	disp_dlg_regs->dst_y_delta_drq_limit = 0x7fff; // off
+
+	disp_ttu_regs->refcyc_per_req_delivery_pre_l = (unsigned int) (refcyc_per_req_delivery_pre_l * dml_pow(2, 10));
+	disp_ttu_regs->refcyc_per_req_delivery_l = (unsigned int) (refcyc_per_req_delivery_l * dml_pow(2, 10));
+	disp_ttu_regs->refcyc_per_req_delivery_pre_c = (unsigned int) (refcyc_per_req_delivery_pre_c * dml_pow(2, 10));
+	disp_ttu_regs->refcyc_per_req_delivery_c = (unsigned int) (refcyc_per_req_delivery_c * dml_pow(2, 10));
+	disp_ttu_regs->refcyc_per_req_delivery_pre_cur0 = (unsigned int) (refcyc_per_req_delivery_pre_cur0 * dml_pow(2, 10));
+	disp_ttu_regs->refcyc_per_req_delivery_cur0 = (unsigned int) (refcyc_per_req_delivery_cur0 * dml_pow(2, 10));
+	disp_ttu_regs->refcyc_per_req_delivery_pre_cur1 = (unsigned int) (refcyc_per_req_delivery_pre_cur1 * dml_pow(2, 10));
+	disp_ttu_regs->refcyc_per_req_delivery_cur1 = (unsigned int) (refcyc_per_req_delivery_cur1 * dml_pow(2, 10));
+
+	disp_ttu_regs->qos_level_low_wm = 0;
+	ASSERT(disp_ttu_regs->qos_level_low_wm < dml_pow(2, 14));
+
+	disp_ttu_regs->qos_level_high_wm = (unsigned int) (4.0 * (double) htotal * ref_freq_to_pix_freq);
+	ASSERT(disp_ttu_regs->qos_level_high_wm < dml_pow(2, 14));
+
+	disp_ttu_regs->qos_level_flip = 14;
+	disp_ttu_regs->qos_level_fixed_l = 8;
+	disp_ttu_regs->qos_level_fixed_c = 8;
+	disp_ttu_regs->qos_level_fixed_cur0 = 8;
+	disp_ttu_regs->qos_ramp_disable_l = 0;
+	disp_ttu_regs->qos_ramp_disable_c = 0;
+	disp_ttu_regs->qos_ramp_disable_cur0 = 0;
+
+	disp_ttu_regs->min_ttu_vblank = min_ttu_vblank * refclk_freq_in_mhz;
+	ASSERT(disp_ttu_regs->min_ttu_vblank < dml_pow(2, 24));
+
+	print__ttu_regs_st(mode_lib, disp_ttu_regs);
+	print__dlg_regs_st(mode_lib, disp_dlg_regs);
+}
+
+void dml314_rq_dlg_get_dlg_reg(
+		struct display_mode_lib *mode_lib,
+		display_dlg_regs_st *dlg_regs,
+		display_ttu_regs_st *ttu_regs,
+		const display_e2e_pipe_params_st *e2e_pipe_param,
+		const unsigned int num_pipes,
+		const unsigned int pipe_idx,
+		const bool cstate_en,
+		const bool pstate_en,
+		const bool vm_en,
+		const bool ignore_viewport_pos,
+		const bool immediate_flip_support)
+{
+	display_rq_params_st rq_param = {0};
+	display_dlg_sys_params_st dlg_sys_param = {0};
+
+	// Get watermark and Tex.
+	dlg_sys_param.t_urg_wm_us = get_wm_urgent(mode_lib, e2e_pipe_param, num_pipes);
+	dlg_sys_param.deepsleep_dcfclk_mhz = get_clk_dcf_deepsleep(mode_lib, e2e_pipe_param, num_pipes);
+	dlg_sys_param.t_extra_us = get_urgent_extra_latency(mode_lib, e2e_pipe_param, num_pipes);
+	dlg_sys_param.mem_trip_us = get_wm_memory_trip(mode_lib, e2e_pipe_param, num_pipes);
+	dlg_sys_param.t_mclk_wm_us = get_wm_dram_clock_change(mode_lib, e2e_pipe_param, num_pipes);
+	dlg_sys_param.t_sr_wm_us = get_wm_stutter_enter_exit(mode_lib, e2e_pipe_param, num_pipes);
+	dlg_sys_param.total_flip_bw = get_total_immediate_flip_bw(mode_lib, e2e_pipe_param, num_pipes);
+	dlg_sys_param.total_flip_bytes = get_total_immediate_flip_bytes(mode_lib, e2e_pipe_param, num_pipes);
+
+	print__dlg_sys_params_st(mode_lib, &dlg_sys_param);
+
+	// system parameter calculation done
+
+	dml_print("DML_DLG: Calculation for pipe[%d] start\n\n", pipe_idx);
+	dml_rq_dlg_get_rq_params(mode_lib, &rq_param, &e2e_pipe_param[pipe_idx].pipe);
+	dml_rq_dlg_get_dlg_params(
+			mode_lib,
+			e2e_pipe_param,
+			num_pipes,
+			pipe_idx,
+			dlg_regs,
+			ttu_regs,
+			&rq_param.dlg,
+			&dlg_sys_param,
+			cstate_en,
+			pstate_en,
+			vm_en,
+			ignore_viewport_pos,
+			immediate_flip_support);
+	dml_print("DML_DLG: Calculation for pipe[%d] end\n", pipe_idx);
+}
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn314/display_rq_dlg_calc_314.h b/drivers/gpu/drm/amd/display/dc/dml/dcn314/display_rq_dlg_calc_314.h
new file mode 100644
index 000000000000..49cb85d1056c
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn314/display_rq_dlg_calc_314.h
@@ -0,0 +1,70 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __DML314_DISPLAY_RQ_DLG_CALC_H__
+#define __DML314_DISPLAY_RQ_DLG_CALC_H__
+
+#include "../display_rq_dlg_helpers.h"
+
+struct display_mode_lib;
+
+// Function: dml_rq_dlg_get_rq_reg
+//  Main entry point for test to get the register values out of this DML class.
+//  This function calls <get_rq_param> and <extract_rq_regs> fucntions to calculate
+//  and then populate the rq_regs struct
+// Input:
+//  pipe_param - pipe source configuration (e.g. vp, pitch, scaling, dest, etc.)
+// Output:
+//  rq_regs - struct that holds all the RQ registers field value.
+//            See also: <display_rq_regs_st>
+void dml314_rq_dlg_get_rq_reg(struct display_mode_lib *mode_lib,
+		display_rq_regs_st *rq_regs,
+		const display_pipe_params_st *pipe_param);
+
+// Function: dml_rq_dlg_get_dlg_reg
+//   Calculate and return DLG and TTU register struct given the system setting
+// Output:
+//  dlg_regs - output DLG register struct
+//  ttu_regs - output DLG TTU register struct
+// Input:
+//  e2e_pipe_param - "compacted" array of e2e pipe param struct
+//  num_pipes - num of active "pipe" or "route"
+//  pipe_idx - index that identifies the e2e_pipe_param that corresponding to this dlg
+//  cstate - 0: when calculate min_ttu_vblank it is assumed cstate is not required. 1: Normal mode, cstate is considered.
+//           Added for legacy or unrealistic timing tests.
+void dml314_rq_dlg_get_dlg_reg(struct display_mode_lib             *mode_lib,
+		display_dlg_regs_st          *dlg_regs,
+		display_ttu_regs_st          *ttu_regs,
+		const display_e2e_pipe_params_st *e2e_pipe_param,
+		const unsigned int            num_pipes,
+		const unsigned int            pipe_idx,
+		const bool                    cstate_en,
+		const bool                    pstate_en,
+		const bool                    vm_en,
+		const bool                    ignore_viewport_pos,
+		const bool                    immediate_flip_support);
+
+#endif
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.c
new file mode 100644
index 000000000000..8a0f128722b0
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.c
@@ -0,0 +1,3613 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+#include "dcn32_fpu.h"
+#include "dcn32/dcn32_resource.h"
+#include "dcn20/dcn20_resource.h"
+#include "display_mode_vba_util_32.h"
+#include "dml/dcn32/display_mode_vba_32.h"
+// We need this includes for WATERMARKS_* defines
+#include "clk_mgr/dcn32/dcn32_smu13_driver_if.h"
+#include "dcn30/dcn30_resource.h"
+#include "link_service.h"
+#include "dc_state_priv.h"
+
+#define DC_LOGGER_INIT(logger)
+
+static const struct subvp_high_refresh_list subvp_high_refresh_list = {
+			.min_refresh = 120,
+			.max_refresh = 175,
+			.res = {
+				{.width = 3840, .height = 2160, },
+				{.width = 3440, .height = 1440, },
+				{.width = 2560, .height = 1440, },
+				{.width = 1920, .height = 1080, }},
+};
+
+static const struct subvp_active_margin_list subvp_active_margin_list = {
+			.min_refresh = 55,
+			.max_refresh = 65,
+			.res = {
+				{.width = 2560, .height = 1440, },
+				{.width = 1920, .height = 1080, }},
+};
+
+struct _vcs_dpi_ip_params_st dcn3_2_ip = {
+	.gpuvm_enable = 0,
+	.gpuvm_max_page_table_levels = 4,
+	.hostvm_enable = 0,
+	.rob_buffer_size_kbytes = 128,
+	.det_buffer_size_kbytes = DCN3_2_DEFAULT_DET_SIZE,
+	.config_return_buffer_size_in_kbytes = 1280,
+	.compressed_buffer_segment_size_in_kbytes = 64,
+	.meta_fifo_size_in_kentries = 22,
+	.zero_size_buffer_entries = 512,
+	.compbuf_reserved_space_64b = 256,
+	.compbuf_reserved_space_zs = 64,
+	.dpp_output_buffer_pixels = 2560,
+	.opp_output_buffer_lines = 1,
+	.pixel_chunk_size_kbytes = 8,
+	.alpha_pixel_chunk_size_kbytes = 4,
+	.min_pixel_chunk_size_bytes = 1024,
+	.dcc_meta_buffer_size_bytes = 6272,
+	.meta_chunk_size_kbytes = 2,
+	.min_meta_chunk_size_bytes = 256,
+	.writeback_chunk_size_kbytes = 8,
+	.ptoi_supported = false,
+	.num_dsc = 4,
+	.maximum_dsc_bits_per_component = 12,
+	.maximum_pixels_per_line_per_dsc_unit = 6016,
+	.dsc422_native_support = true,
+	.is_line_buffer_bpp_fixed = true,
+	.line_buffer_fixed_bpp = 57,
+	.line_buffer_size_bits = 1171920,
+	.max_line_buffer_lines = 32,
+	.writeback_interface_buffer_size_kbytes = 90,
+	.max_num_dpp = 4,
+	.max_num_otg = 4,
+	.max_num_hdmi_frl_outputs = 1,
+	.max_num_wb = 1,
+	.max_dchub_pscl_bw_pix_per_clk = 4,
+	.max_pscl_lb_bw_pix_per_clk = 2,
+	.max_lb_vscl_bw_pix_per_clk = 4,
+	.max_vscl_hscl_bw_pix_per_clk = 4,
+	.max_hscl_ratio = 6,
+	.max_vscl_ratio = 6,
+	.max_hscl_taps = 8,
+	.max_vscl_taps = 8,
+	.dpte_buffer_size_in_pte_reqs_luma = 64,
+	.dpte_buffer_size_in_pte_reqs_chroma = 34,
+	.dispclk_ramp_margin_percent = 1,
+	.max_inter_dcn_tile_repeaters = 8,
+	.cursor_buffer_size = 16,
+	.cursor_chunk_size = 2,
+	.writeback_line_buffer_buffer_size = 0,
+	.writeback_min_hscl_ratio = 1,
+	.writeback_min_vscl_ratio = 1,
+	.writeback_max_hscl_ratio = 1,
+	.writeback_max_vscl_ratio = 1,
+	.writeback_max_hscl_taps = 1,
+	.writeback_max_vscl_taps = 1,
+	.dppclk_delay_subtotal = 47,
+	.dppclk_delay_scl = 50,
+	.dppclk_delay_scl_lb_only = 16,
+	.dppclk_delay_cnvc_formatter = 28,
+	.dppclk_delay_cnvc_cursor = 6,
+	.dispclk_delay_subtotal = 125,
+	.dynamic_metadata_vm_enabled = false,
+	.odm_combine_4to1_supported = false,
+	.dcc_supported = true,
+	.max_num_dp2p0_outputs = 2,
+	.max_num_dp2p0_streams = 4,
+};
+
+struct _vcs_dpi_soc_bounding_box_st dcn3_2_soc = {
+	.clock_limits = {
+		{
+			.state = 0,
+			.dcfclk_mhz = 1564.0,
+			.fabricclk_mhz = 2500.0,
+			.dispclk_mhz = 2150.0,
+			.dppclk_mhz = 2150.0,
+			.phyclk_mhz = 810.0,
+			.phyclk_d18_mhz = 667.0,
+			.phyclk_d32_mhz = 625.0,
+			.socclk_mhz = 1200.0,
+			.dscclk_mhz = 716.667,
+			.dram_speed_mts = 18000.0,
+			.dtbclk_mhz = 1564.0,
+		},
+	},
+	.num_states = 1,
+	.sr_exit_time_us = 42.97,
+	.sr_enter_plus_exit_time_us = 49.94,
+	.sr_exit_z8_time_us = 285.0,
+	.sr_enter_plus_exit_z8_time_us = 320,
+	.writeback_latency_us = 12.0,
+	.round_trip_ping_latency_dcfclk_cycles = 263,
+	.urgent_latency_pixel_data_only_us = 4.0,
+	.urgent_latency_pixel_mixed_with_vm_data_us = 4.0,
+	.urgent_latency_vm_data_only_us = 4.0,
+	.fclk_change_latency_us = 25,
+	.usr_retraining_latency_us = 2,
+	.smn_latency_us = 2,
+	.mall_allocated_for_dcn_mbytes = 64,
+	.urgent_out_of_order_return_per_channel_pixel_only_bytes = 4096,
+	.urgent_out_of_order_return_per_channel_pixel_and_vm_bytes = 4096,
+	.urgent_out_of_order_return_per_channel_vm_only_bytes = 4096,
+	.pct_ideal_sdp_bw_after_urgent = 90.0,
+	.pct_ideal_fabric_bw_after_urgent = 67.0,
+	.pct_ideal_dram_sdp_bw_after_urgent_pixel_only = 20.0,
+	.pct_ideal_dram_sdp_bw_after_urgent_pixel_and_vm = 60.0,
+	.pct_ideal_dram_sdp_bw_after_urgent_vm_only = 30.0,
+	.pct_ideal_dram_bw_after_urgent_strobe = 67.0,
+	.max_avg_sdp_bw_use_normal_percent = 80.0,
+	.max_avg_fabric_bw_use_normal_percent = 60.0,
+	.max_avg_dram_bw_use_normal_strobe_percent = 50.0,
+	.max_avg_dram_bw_use_normal_percent = 15.0,
+	.num_chans = 24,
+	.dram_channel_width_bytes = 2,
+	.fabric_datapath_to_dcn_data_return_bytes = 64,
+	.return_bus_width_bytes = 64,
+	.downspread_percent = 0.38,
+	.dcn_downspread_percent = 0.5,
+	.dram_clock_change_latency_us = 400,
+	.dispclk_dppclk_vco_speed_mhz = 4300.0,
+	.do_urgent_latency_adjustment = true,
+	.urgent_latency_adjustment_fabric_clock_component_us = 1.0,
+	.urgent_latency_adjustment_fabric_clock_reference_mhz = 3000,
+};
+
+static bool dcn32_apply_merge_split_flags_helper(struct dc *dc, struct dc_state *context,
+	bool *repopulate_pipes, int *split, bool *merge);
+
+void dcn32_build_wm_range_table_fpu(struct clk_mgr_internal *clk_mgr)
+{
+	/* defaults */
+	double pstate_latency_us = clk_mgr->base.ctx->dc->dml.soc.dram_clock_change_latency_us;
+	double fclk_change_latency_us = clk_mgr->base.ctx->dc->dml.soc.fclk_change_latency_us;
+	double sr_exit_time_us = clk_mgr->base.ctx->dc->dml.soc.sr_exit_time_us;
+	double sr_enter_plus_exit_time_us = clk_mgr->base.ctx->dc->dml.soc.sr_enter_plus_exit_time_us;
+	/* For min clocks use as reported by PM FW and report those as min */
+	uint16_t min_uclk_mhz			= clk_mgr->base.bw_params->clk_table.entries[0].memclk_mhz;
+	uint16_t min_dcfclk_mhz			= clk_mgr->base.bw_params->clk_table.entries[0].dcfclk_mhz;
+	uint16_t setb_min_uclk_mhz		= min_uclk_mhz;
+	uint16_t dcfclk_mhz_for_the_second_state = clk_mgr->base.ctx->dc->dml.soc.clock_limits[2].dcfclk_mhz;
+
+	dc_assert_fp_enabled();
+
+	/* For Set B ranges use min clocks state 2 when available, and report those to PM FW */
+	if (dcfclk_mhz_for_the_second_state)
+		clk_mgr->base.bw_params->wm_table.nv_entries[WM_B].pmfw_breakdown.min_dcfclk = dcfclk_mhz_for_the_second_state;
+	else
+		clk_mgr->base.bw_params->wm_table.nv_entries[WM_B].pmfw_breakdown.min_dcfclk = clk_mgr->base.bw_params->clk_table.entries[0].dcfclk_mhz;
+
+	if (clk_mgr->base.bw_params->clk_table.entries[2].memclk_mhz)
+		setb_min_uclk_mhz = clk_mgr->base.bw_params->clk_table.entries[2].memclk_mhz;
+
+	/* Set A - Normal - default values */
+	clk_mgr->base.bw_params->wm_table.nv_entries[WM_A].valid = true;
+	clk_mgr->base.bw_params->wm_table.nv_entries[WM_A].dml_input.pstate_latency_us = pstate_latency_us;
+	clk_mgr->base.bw_params->wm_table.nv_entries[WM_A].dml_input.fclk_change_latency_us = fclk_change_latency_us;
+	clk_mgr->base.bw_params->wm_table.nv_entries[WM_A].dml_input.sr_exit_time_us = sr_exit_time_us;
+	clk_mgr->base.bw_params->wm_table.nv_entries[WM_A].dml_input.sr_enter_plus_exit_time_us = sr_enter_plus_exit_time_us;
+	clk_mgr->base.bw_params->wm_table.nv_entries[WM_A].pmfw_breakdown.wm_type = WATERMARKS_CLOCK_RANGE;
+	clk_mgr->base.bw_params->wm_table.nv_entries[WM_A].pmfw_breakdown.min_dcfclk = min_dcfclk_mhz;
+	clk_mgr->base.bw_params->wm_table.nv_entries[WM_A].pmfw_breakdown.max_dcfclk = 0xFFFF;
+	clk_mgr->base.bw_params->wm_table.nv_entries[WM_A].pmfw_breakdown.min_uclk = min_uclk_mhz;
+	clk_mgr->base.bw_params->wm_table.nv_entries[WM_A].pmfw_breakdown.max_uclk = 0xFFFF;
+
+	/* Set B - Performance - higher clocks, using DPM[2] DCFCLK and UCLK */
+	clk_mgr->base.bw_params->wm_table.nv_entries[WM_B].valid = true;
+	clk_mgr->base.bw_params->wm_table.nv_entries[WM_B].dml_input.pstate_latency_us = pstate_latency_us;
+	clk_mgr->base.bw_params->wm_table.nv_entries[WM_B].dml_input.fclk_change_latency_us = fclk_change_latency_us;
+	clk_mgr->base.bw_params->wm_table.nv_entries[WM_B].dml_input.sr_exit_time_us = sr_exit_time_us;
+	clk_mgr->base.bw_params->wm_table.nv_entries[WM_B].dml_input.sr_enter_plus_exit_time_us = sr_enter_plus_exit_time_us;
+	clk_mgr->base.bw_params->wm_table.nv_entries[WM_B].pmfw_breakdown.wm_type = WATERMARKS_CLOCK_RANGE;
+	clk_mgr->base.bw_params->wm_table.nv_entries[WM_B].pmfw_breakdown.max_dcfclk = 0xFFFF;
+	clk_mgr->base.bw_params->wm_table.nv_entries[WM_B].pmfw_breakdown.min_uclk = setb_min_uclk_mhz;
+	clk_mgr->base.bw_params->wm_table.nv_entries[WM_B].pmfw_breakdown.max_uclk = 0xFFFF;
+
+	/* Set C - Dummy P-State - P-State latency set to "dummy p-state" value */
+	/* 'DalDummyClockChangeLatencyNs' registry key option set to 0x7FFFFFFF can be used to disable Set C for dummy p-state */
+	if (clk_mgr->base.ctx->dc->bb_overrides.dummy_clock_change_latency_ns != 0x7FFFFFFF) {
+		clk_mgr->base.bw_params->wm_table.nv_entries[WM_C].valid = true;
+		clk_mgr->base.bw_params->wm_table.nv_entries[WM_C].dml_input.pstate_latency_us = 50;
+		clk_mgr->base.bw_params->wm_table.nv_entries[WM_C].dml_input.fclk_change_latency_us = fclk_change_latency_us;
+		clk_mgr->base.bw_params->wm_table.nv_entries[WM_C].dml_input.sr_exit_time_us = sr_exit_time_us;
+		clk_mgr->base.bw_params->wm_table.nv_entries[WM_C].dml_input.sr_enter_plus_exit_time_us = sr_enter_plus_exit_time_us;
+		clk_mgr->base.bw_params->wm_table.nv_entries[WM_C].pmfw_breakdown.wm_type = WATERMARKS_DUMMY_PSTATE;
+		clk_mgr->base.bw_params->wm_table.nv_entries[WM_C].pmfw_breakdown.min_dcfclk = min_dcfclk_mhz;
+		clk_mgr->base.bw_params->wm_table.nv_entries[WM_C].pmfw_breakdown.max_dcfclk = 0xFFFF;
+		clk_mgr->base.bw_params->wm_table.nv_entries[WM_C].pmfw_breakdown.min_uclk = min_uclk_mhz;
+		clk_mgr->base.bw_params->wm_table.nv_entries[WM_C].pmfw_breakdown.max_uclk = 0xFFFF;
+		clk_mgr->base.bw_params->dummy_pstate_table[0].dram_speed_mts = clk_mgr->base.bw_params->clk_table.entries[0].memclk_mhz * 16;
+		clk_mgr->base.bw_params->dummy_pstate_table[0].dummy_pstate_latency_us = 50;
+		clk_mgr->base.bw_params->dummy_pstate_table[1].dram_speed_mts = clk_mgr->base.bw_params->clk_table.entries[1].memclk_mhz * 16;
+		clk_mgr->base.bw_params->dummy_pstate_table[1].dummy_pstate_latency_us = 9;
+		clk_mgr->base.bw_params->dummy_pstate_table[2].dram_speed_mts = clk_mgr->base.bw_params->clk_table.entries[2].memclk_mhz * 16;
+		clk_mgr->base.bw_params->dummy_pstate_table[2].dummy_pstate_latency_us = 8;
+		clk_mgr->base.bw_params->dummy_pstate_table[3].dram_speed_mts = clk_mgr->base.bw_params->clk_table.entries[3].memclk_mhz * 16;
+		clk_mgr->base.bw_params->dummy_pstate_table[3].dummy_pstate_latency_us = 5;
+	}
+	/* Set D - MALL - SR enter and exit time specific to MALL, TBD after bringup or later phase for now use DRAM values / 2 */
+	/* For MALL DRAM clock change latency is N/A, for watermak calculations use lowest value dummy P state latency */
+	clk_mgr->base.bw_params->wm_table.nv_entries[WM_D].valid = true;
+	clk_mgr->base.bw_params->wm_table.nv_entries[WM_D].dml_input.pstate_latency_us = clk_mgr->base.bw_params->dummy_pstate_table[3].dummy_pstate_latency_us;
+	clk_mgr->base.bw_params->wm_table.nv_entries[WM_D].dml_input.fclk_change_latency_us = fclk_change_latency_us;
+	clk_mgr->base.bw_params->wm_table.nv_entries[WM_D].dml_input.sr_exit_time_us = sr_exit_time_us / 2; // TBD
+	clk_mgr->base.bw_params->wm_table.nv_entries[WM_D].dml_input.sr_enter_plus_exit_time_us = sr_enter_plus_exit_time_us / 2; // TBD
+	clk_mgr->base.bw_params->wm_table.nv_entries[WM_D].pmfw_breakdown.wm_type = WATERMARKS_MALL;
+	clk_mgr->base.bw_params->wm_table.nv_entries[WM_D].pmfw_breakdown.min_dcfclk = min_dcfclk_mhz;
+	clk_mgr->base.bw_params->wm_table.nv_entries[WM_D].pmfw_breakdown.max_dcfclk = 0xFFFF;
+	clk_mgr->base.bw_params->wm_table.nv_entries[WM_D].pmfw_breakdown.min_uclk = min_uclk_mhz;
+	clk_mgr->base.bw_params->wm_table.nv_entries[WM_D].pmfw_breakdown.max_uclk = 0xFFFF;
+}
+
+/*
+ * Finds dummy_latency_index when MCLK switching using firmware based
+ * vblank stretch is enabled. This function will iterate through the
+ * table of dummy pstate latencies until the lowest value that allows
+ * dm_allow_self_refresh_and_mclk_switch to happen is found
+ */
+int dcn32_find_dummy_latency_index_for_fw_based_mclk_switch(struct dc *dc,
+							    struct dc_state *context,
+							    display_e2e_pipe_params_st *pipes,
+							    int pipe_cnt,
+							    int vlevel)
+{
+	const int max_latency_table_entries = 4;
+	struct vba_vars_st *vba = &context->bw_ctx.dml.vba;
+	int dummy_latency_index = 0;
+	enum clock_change_support temp_clock_change_support = vba->DRAMClockChangeSupport[vlevel][context->bw_ctx.dml.vba.maxMpcComb];
+
+	dc_assert_fp_enabled();
+
+	while (dummy_latency_index < max_latency_table_entries) {
+		if (temp_clock_change_support != dm_dram_clock_change_unsupported)
+			vba->DRAMClockChangeSupport[vlevel][context->bw_ctx.dml.vba.maxMpcComb] = temp_clock_change_support;
+		context->bw_ctx.dml.soc.dram_clock_change_latency_us =
+				dc->clk_mgr->bw_params->dummy_pstate_table[dummy_latency_index].dummy_pstate_latency_us;
+		dcn32_internal_validate_bw(dc, context, pipes, &pipe_cnt, &vlevel, DC_VALIDATE_MODE_AND_PROGRAMMING);
+
+		/* for subvp + DRR case, if subvp pipes are still present we support pstate */
+		if (vba->DRAMClockChangeSupport[vlevel][vba->maxMpcComb] == dm_dram_clock_change_unsupported &&
+				dcn32_subvp_in_use(dc, context))
+			vba->DRAMClockChangeSupport[vlevel][context->bw_ctx.dml.vba.maxMpcComb] = temp_clock_change_support;
+
+		if (vlevel < context->bw_ctx.dml.vba.soc.num_states &&
+				vba->DRAMClockChangeSupport[vlevel][vba->maxMpcComb] != dm_dram_clock_change_unsupported)
+			break;
+
+		dummy_latency_index++;
+	}
+
+	if (dummy_latency_index == max_latency_table_entries) {
+		ASSERT(dummy_latency_index != max_latency_table_entries);
+		/* If the execution gets here, it means dummy p_states are
+		 * not possible. This should never happen and would mean
+		 * something is severely wrong.
+		 * Here we reset dummy_latency_index to 3, because it is
+		 * better to have underflows than system crashes.
+		 */
+		dummy_latency_index = max_latency_table_entries - 1;
+	}
+
+	return dummy_latency_index;
+}
+
+/**
+ * dcn32_helper_populate_phantom_dlg_params - Get DLG params for phantom pipes
+ * and populate pipe_ctx with those params.
+ * @dc: [in] current dc state
+ * @context: [in] new dc state
+ * @pipes: [in] DML pipe params array
+ * @pipe_cnt: [in] DML pipe count
+ *
+ * This function must be called AFTER the phantom pipes are added to context
+ * and run through DML (so that the DLG params for the phantom pipes can be
+ * populated), and BEFORE we program the timing for the phantom pipes.
+ */
+void dcn32_helper_populate_phantom_dlg_params(struct dc *dc,
+					      struct dc_state *context,
+					      display_e2e_pipe_params_st *pipes,
+					      int pipe_cnt)
+{
+	uint32_t i, pipe_idx;
+
+	dc_assert_fp_enabled();
+
+	for (i = 0, pipe_idx = 0; i < dc->res_pool->pipe_count; i++) {
+		struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
+
+		if (!pipe->stream)
+			continue;
+
+		if (pipe->plane_state && dc_state_get_pipe_subvp_type(context, pipe) == SUBVP_PHANTOM) {
+			pipes[pipe_idx].pipe.dest.vstartup_start =
+				get_vstartup(&context->bw_ctx.dml, pipes, pipe_cnt, pipe_idx);
+			pipes[pipe_idx].pipe.dest.vupdate_offset =
+				get_vupdate_offset(&context->bw_ctx.dml, pipes, pipe_cnt, pipe_idx);
+			pipes[pipe_idx].pipe.dest.vupdate_width =
+				get_vupdate_width(&context->bw_ctx.dml, pipes, pipe_cnt, pipe_idx);
+			pipes[pipe_idx].pipe.dest.vready_offset =
+				get_vready_offset(&context->bw_ctx.dml, pipes, pipe_cnt, pipe_idx);
+			pipe->pipe_dlg_param = pipes[pipe_idx].pipe.dest;
+		}
+		pipe_idx++;
+	}
+}
+
+static float calculate_net_bw_in_kbytes_sec(struct _vcs_dpi_voltage_scaling_st *entry)
+{
+	float memory_bw_kbytes_sec;
+	float fabric_bw_kbytes_sec;
+	float sdp_bw_kbytes_sec;
+	float limiting_bw_kbytes_sec;
+
+	memory_bw_kbytes_sec = entry->dram_speed_mts *
+				dcn3_2_soc.num_chans *
+				dcn3_2_soc.dram_channel_width_bytes *
+				((float)dcn3_2_soc.pct_ideal_dram_sdp_bw_after_urgent_pixel_only / 100);
+
+	fabric_bw_kbytes_sec = entry->fabricclk_mhz *
+				dcn3_2_soc.return_bus_width_bytes *
+				((float)dcn3_2_soc.pct_ideal_fabric_bw_after_urgent / 100);
+
+	sdp_bw_kbytes_sec = entry->dcfclk_mhz *
+				dcn3_2_soc.return_bus_width_bytes *
+				((float)dcn3_2_soc.pct_ideal_sdp_bw_after_urgent / 100);
+
+	limiting_bw_kbytes_sec = memory_bw_kbytes_sec;
+
+	if (fabric_bw_kbytes_sec < limiting_bw_kbytes_sec)
+		limiting_bw_kbytes_sec = fabric_bw_kbytes_sec;
+
+	if (sdp_bw_kbytes_sec < limiting_bw_kbytes_sec)
+		limiting_bw_kbytes_sec = sdp_bw_kbytes_sec;
+
+	return limiting_bw_kbytes_sec;
+}
+
+static void get_optimal_ntuple(struct _vcs_dpi_voltage_scaling_st *entry)
+{
+	if (entry->dcfclk_mhz > 0) {
+		float bw_on_sdp = entry->dcfclk_mhz * dcn3_2_soc.return_bus_width_bytes * ((float)dcn3_2_soc.pct_ideal_sdp_bw_after_urgent / 100);
+
+		entry->fabricclk_mhz = bw_on_sdp / (dcn3_2_soc.return_bus_width_bytes * ((float)dcn3_2_soc.pct_ideal_fabric_bw_after_urgent / 100));
+		entry->dram_speed_mts = bw_on_sdp / (dcn3_2_soc.num_chans *
+				dcn3_2_soc.dram_channel_width_bytes * ((float)dcn3_2_soc.pct_ideal_dram_sdp_bw_after_urgent_pixel_only / 100));
+	} else if (entry->fabricclk_mhz > 0) {
+		float bw_on_fabric = entry->fabricclk_mhz * dcn3_2_soc.return_bus_width_bytes * ((float)dcn3_2_soc.pct_ideal_fabric_bw_after_urgent / 100);
+
+		entry->dcfclk_mhz = bw_on_fabric / (dcn3_2_soc.return_bus_width_bytes * ((float)dcn3_2_soc.pct_ideal_sdp_bw_after_urgent / 100));
+		entry->dram_speed_mts = bw_on_fabric / (dcn3_2_soc.num_chans *
+				dcn3_2_soc.dram_channel_width_bytes * ((float)dcn3_2_soc.pct_ideal_dram_sdp_bw_after_urgent_pixel_only / 100));
+	} else if (entry->dram_speed_mts > 0) {
+		float bw_on_dram = entry->dram_speed_mts * dcn3_2_soc.num_chans *
+				dcn3_2_soc.dram_channel_width_bytes * ((float)dcn3_2_soc.pct_ideal_dram_sdp_bw_after_urgent_pixel_only / 100);
+
+		entry->fabricclk_mhz = bw_on_dram / (dcn3_2_soc.return_bus_width_bytes * ((float)dcn3_2_soc.pct_ideal_fabric_bw_after_urgent / 100));
+		entry->dcfclk_mhz = bw_on_dram / (dcn3_2_soc.return_bus_width_bytes * ((float)dcn3_2_soc.pct_ideal_sdp_bw_after_urgent / 100));
+	}
+}
+
+static void insert_entry_into_table_sorted(struct _vcs_dpi_voltage_scaling_st *table,
+				    unsigned int *num_entries,
+				    struct _vcs_dpi_voltage_scaling_st *entry)
+{
+	int i = 0;
+	int index = 0;
+
+	dc_assert_fp_enabled();
+
+	if (*num_entries == 0) {
+		table[0] = *entry;
+		(*num_entries)++;
+	} else {
+		while (entry->net_bw_in_kbytes_sec > table[index].net_bw_in_kbytes_sec) {
+			index++;
+			if (index >= *num_entries)
+				break;
+		}
+
+		for (i = *num_entries; i > index; i--)
+			table[i] = table[i - 1];
+
+		table[index] = *entry;
+		(*num_entries)++;
+	}
+}
+
+/**
+ * dcn32_set_phantom_stream_timing - Set timing params for the phantom stream
+ * @dc: current dc state
+ * @context: new dc state
+ * @ref_pipe: Main pipe for the phantom stream
+ * @phantom_stream: target phantom stream state
+ * @pipes: DML pipe params
+ * @pipe_cnt: number of DML pipes
+ * @dc_pipe_idx: DC pipe index for the main pipe (i.e. ref_pipe)
+ *
+ * Set timing params of the phantom stream based on calculated output from DML.
+ * This function first gets the DML pipe index using the DC pipe index, then
+ * calls into DML (get_subviewport_lines_needed_in_mall) to get the number of
+ * lines required for SubVP MCLK switching and assigns to the phantom stream
+ * accordingly.
+ *
+ * - The number of SubVP lines calculated in DML does not take into account
+ * FW processing delays and required pstate allow width, so we must include
+ * that separately.
+ *
+ * - Set phantom backporch = vstartup of main pipe
+ */
+void dcn32_set_phantom_stream_timing(struct dc *dc,
+				     struct dc_state *context,
+				     struct pipe_ctx *ref_pipe,
+				     struct dc_stream_state *phantom_stream,
+				     display_e2e_pipe_params_st *pipes,
+				     unsigned int pipe_cnt,
+				     unsigned int dc_pipe_idx)
+{
+	unsigned int i, pipe_idx;
+	struct pipe_ctx *pipe;
+	uint32_t phantom_vactive, phantom_bp, pstate_width_fw_delay_lines;
+	unsigned int num_dpp;
+	unsigned int vlevel = context->bw_ctx.dml.vba.VoltageLevel;
+	unsigned int dcfclk = context->bw_ctx.dml.vba.DCFCLKState[vlevel][context->bw_ctx.dml.vba.maxMpcComb];
+	unsigned int socclk = context->bw_ctx.dml.vba.SOCCLKPerState[vlevel];
+	struct vba_vars_st *vba = &context->bw_ctx.dml.vba;
+	struct dc_stream_state *main_stream = ref_pipe->stream;
+
+	dc_assert_fp_enabled();
+
+	// Find DML pipe index (pipe_idx) using dc_pipe_idx
+	for (i = 0, pipe_idx = 0; i < dc->res_pool->pipe_count; i++) {
+		pipe = &context->res_ctx.pipe_ctx[i];
+
+		if (!pipe->stream)
+			continue;
+
+		if (i == dc_pipe_idx)
+			break;
+
+		pipe_idx++;
+	}
+
+	// Calculate lines required for pstate allow width and FW processing delays
+	pstate_width_fw_delay_lines = ((double)(dc->caps.subvp_fw_processing_delay_us +
+			dc->caps.subvp_pstate_allow_width_us) / 1000000) *
+			(ref_pipe->stream->timing.pix_clk_100hz * 100) /
+			(double)ref_pipe->stream->timing.h_total;
+
+	// Update clks_cfg for calling into recalculate
+	pipes[0].clks_cfg.voltage = vlevel;
+	pipes[0].clks_cfg.dcfclk_mhz = dcfclk;
+	pipes[0].clks_cfg.socclk_mhz = socclk;
+
+	// DML calculation for MALL region doesn't take into account FW delay
+	// and required pstate allow width for multi-display cases
+	/* Add 16 lines margin to the MALL REGION because SUB_VP_START_LINE must be aligned
+	 * to 2 swaths (i.e. 16 lines)
+	 */
+	phantom_vactive = get_subviewport_lines_needed_in_mall(&context->bw_ctx.dml, pipes, pipe_cnt, pipe_idx) +
+				pstate_width_fw_delay_lines + dc->caps.subvp_swath_height_margin_lines;
+
+	// W/A for DCC corruption with certain high resolution timings.
+	// Determing if pipesplit is used. If so, add meta_row_height to the phantom vactive.
+	num_dpp = vba->NoOfDPP[vba->VoltageLevel][vba->maxMpcComb][vba->pipe_plane[pipe_idx]];
+	phantom_vactive += num_dpp > 1 ? vba->meta_row_height[vba->pipe_plane[pipe_idx]] : 0;
+
+	/* dc->debug.subvp_extra_lines 0 by default*/
+	phantom_vactive += dc->debug.subvp_extra_lines;
+
+	// For backporch of phantom pipe, use vstartup of the main pipe
+	phantom_bp = get_vstartup(&context->bw_ctx.dml, pipes, pipe_cnt, pipe_idx);
+
+	phantom_stream->dst.y = 0;
+	phantom_stream->dst.height = phantom_vactive;
+	/* When scaling, DML provides the end to end required number of lines for MALL.
+	 * dst.height is always correct for this case, but src.height is not which causes a
+	 * delta between main and phantom pipe scaling outputs. Need to adjust src.height on
+	 * phantom for this case.
+	 */
+	phantom_stream->src.y = 0;
+	phantom_stream->src.height = (double)phantom_vactive * (double)main_stream->src.height / (double)main_stream->dst.height;
+
+	phantom_stream->timing.v_addressable = phantom_vactive;
+	phantom_stream->timing.v_front_porch = 1;
+	phantom_stream->timing.v_total = phantom_stream->timing.v_addressable +
+						phantom_stream->timing.v_front_porch +
+						phantom_stream->timing.v_sync_width +
+						phantom_bp;
+	phantom_stream->timing.flags.DSC = 0; // Don't need DSC for phantom timing
+}
+
+/**
+ * dcn32_get_num_free_pipes - Calculate number of free pipes
+ * @dc: current dc state
+ * @context: new dc state
+ *
+ * This function assumes that a "used" pipe is a pipe that has
+ * both a stream and a plane assigned to it.
+ *
+ * Return: Number of free pipes available in the context
+ */
+static unsigned int dcn32_get_num_free_pipes(struct dc *dc, struct dc_state *context)
+{
+	unsigned int i;
+	unsigned int free_pipes = 0;
+	unsigned int num_pipes = 0;
+
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
+
+		if (pipe->stream && !pipe->top_pipe) {
+			while (pipe) {
+				num_pipes++;
+				pipe = pipe->bottom_pipe;
+			}
+		}
+	}
+
+	free_pipes = dc->res_pool->pipe_count - num_pipes;
+	return free_pipes;
+}
+
+/**
+ * dcn32_assign_subvp_pipe - Function to decide which pipe will use Sub-VP.
+ * @dc: current dc state
+ * @context: new dc state
+ * @index: [out] dc pipe index for the pipe chosen to have phantom pipes assigned
+ *
+ * We enter this function if we are Sub-VP capable (i.e. enough pipes available)
+ * and regular P-State switching (i.e. VACTIVE/VBLANK) is not supported, or if
+ * we are forcing SubVP P-State switching on the current config.
+ *
+ * The number of pipes used for the chosen surface must be less than or equal to the
+ * number of free pipes available.
+ *
+ * In general we choose surfaces with the longest frame time first (better for SubVP + VBLANK).
+ * For multi-display cases the ActiveDRAMClockChangeMargin doesn't provide enough info on its own
+ * for determining which should be the SubVP pipe (need a way to determine if a pipe / plane doesn't
+ * support MCLK switching naturally [i.e. ACTIVE or VBLANK]).
+ *
+ * Return: True if a valid pipe assignment was found for Sub-VP. Otherwise false.
+ */
+static bool dcn32_assign_subvp_pipe(struct dc *dc,
+				    struct dc_state *context,
+				    unsigned int *index)
+{
+	unsigned int i, pipe_idx;
+	unsigned int max_frame_time = 0;
+	bool valid_assignment_found = false;
+	unsigned int free_pipes = dcn32_get_num_free_pipes(dc, context);
+	struct vba_vars_st *vba = &context->bw_ctx.dml.vba;
+
+	for (i = 0, pipe_idx = 0; i < dc->res_pool->pipe_count; i++) {
+		struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
+		unsigned int num_pipes = 0;
+		unsigned int refresh_rate = 0;
+
+		if (!pipe->stream)
+			continue;
+
+		// Round up
+		refresh_rate = (pipe->stream->timing.pix_clk_100hz * 100 +
+				pipe->stream->timing.v_total * pipe->stream->timing.h_total - 1)
+				/ (double)(pipe->stream->timing.v_total * pipe->stream->timing.h_total);
+		/* SubVP pipe candidate requirements:
+		 * - Refresh rate < 120hz
+		 * - Not able to switch in vactive naturally (switching in active means the
+		 *   DET provides enough buffer to hide the P-State switch latency -- trying
+		 *   to combine this with SubVP can cause issues with the scheduling).
+		 * - Not TMZ surface
+		 */
+		if (pipe->plane_state && !pipe->top_pipe && !pipe->prev_odm_pipe && !dcn32_is_center_timing(pipe) &&
+				!pipe->stream->hw_cursor_req &&
+				!dc_state_get_stream_cursor_subvp_limit(pipe->stream, context) &&
+				!(pipe->stream->timing.pix_clk_100hz / 10000 > DCN3_2_MAX_SUBVP_PIXEL_RATE_MHZ) &&
+				(!dcn32_is_psr_capable(pipe) || (context->stream_count == 1 && dc->caps.dmub_caps.subvp_psr)) &&
+				dc_state_get_pipe_subvp_type(context, pipe) == SUBVP_NONE &&
+				(refresh_rate < 120 || dcn32_allow_subvp_high_refresh_rate(dc, context, pipe)) &&
+				!pipe->plane_state->address.tmz_surface &&
+				(vba->ActiveDRAMClockChangeLatencyMarginPerState[vba->VoltageLevel][vba->maxMpcComb][vba->pipe_plane[pipe_idx]] <= 0 ||
+				(vba->ActiveDRAMClockChangeLatencyMarginPerState[vba->VoltageLevel][vba->maxMpcComb][vba->pipe_plane[pipe_idx]] > 0 &&
+						dcn32_allow_subvp_with_active_margin(pipe)))) {
+			while (pipe) {
+				num_pipes++;
+				pipe = pipe->bottom_pipe;
+			}
+
+			pipe = &context->res_ctx.pipe_ctx[i];
+			if (num_pipes <= free_pipes) {
+				struct dc_stream_state *stream = pipe->stream;
+				unsigned int frame_us = (stream->timing.v_total * stream->timing.h_total /
+						(double)(stream->timing.pix_clk_100hz * 100)) * 1000000;
+				if (frame_us > max_frame_time) {
+					*index = i;
+					max_frame_time = frame_us;
+					valid_assignment_found = true;
+				}
+			}
+		}
+		pipe_idx++;
+	}
+	return valid_assignment_found;
+}
+
+/**
+ * dcn32_enough_pipes_for_subvp - Function to check if there are "enough" pipes for SubVP.
+ * @dc: current dc state
+ * @context: new dc state
+ *
+ * This function returns true if there are enough free pipes
+ * to create the required phantom pipes for any given stream
+ * (that does not already have phantom pipe assigned).
+ *
+ * e.g. For a 2 stream config where the first stream uses one
+ * pipe and the second stream uses 2 pipes (i.e. pipe split),
+ * this function will return true because there is 1 remaining
+ * pipe which can be used as the phantom pipe for the non pipe
+ * split pipe.
+ *
+ * Return:
+ * True if there are enough free pipes to assign phantom pipes to at least one
+ * stream that does not already have phantom pipes assigned. Otherwise false.
+ */
+static bool dcn32_enough_pipes_for_subvp(struct dc *dc, struct dc_state *context)
+{
+	unsigned int i, split_cnt, free_pipes;
+	unsigned int min_pipe_split = dc->res_pool->pipe_count + 1; // init as max number of pipes + 1
+	bool subvp_possible = false;
+
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
+
+		// Find the minimum pipe split count for non SubVP pipes
+		if (resource_is_pipe_type(pipe, OPP_HEAD) &&
+			dc_state_get_pipe_subvp_type(context, pipe) == SUBVP_NONE) {
+			split_cnt = 0;
+			while (pipe) {
+				split_cnt++;
+				pipe = pipe->bottom_pipe;
+			}
+
+			if (split_cnt < min_pipe_split)
+				min_pipe_split = split_cnt;
+		}
+	}
+
+	free_pipes = dcn32_get_num_free_pipes(dc, context);
+
+	// SubVP only possible if at least one pipe is being used (i.e. free_pipes
+	// should not equal to the pipe_count)
+	if (free_pipes >= min_pipe_split && free_pipes < dc->res_pool->pipe_count)
+		subvp_possible = true;
+
+	return subvp_possible;
+}
+
+/**
+ * subvp_subvp_schedulable - Determine if SubVP + SubVP config is schedulable
+ * @dc: current dc state
+ * @context: new dc state
+ *
+ * High level algorithm:
+ * 1. Find longest microschedule length (in us) between the two SubVP pipes
+ * 2. Check if the worst case overlap (VBLANK in middle of ACTIVE) for both
+ * pipes still allows for the maximum microschedule to fit in the active
+ * region for both pipes.
+ *
+ * Return: True if the SubVP + SubVP config is schedulable, false otherwise
+ */
+static bool subvp_subvp_schedulable(struct dc *dc, struct dc_state *context)
+{
+	struct pipe_ctx *subvp_pipes[2] = {0};
+	struct dc_stream_state *phantom = NULL;
+	uint32_t microschedule_lines = 0;
+	uint32_t index = 0;
+	uint32_t i;
+	uint32_t max_microschedule_us = 0;
+	int32_t vactive1_us, vactive2_us, vblank1_us, vblank2_us;
+
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
+		uint32_t time_us = 0;
+
+		/* Loop to calculate the maximum microschedule time between the two SubVP pipes,
+		 * and also to store the two main SubVP pipe pointers in subvp_pipes[2].
+		 */
+		phantom = dc_state_get_paired_subvp_stream(context, pipe->stream);
+		if (phantom && pipe->stream && pipe->plane_state && !pipe->top_pipe &&
+			dc_state_get_pipe_subvp_type(context, pipe) == SUBVP_MAIN) {
+			microschedule_lines = (phantom->timing.v_total - phantom->timing.v_front_porch) +
+					phantom->timing.v_addressable;
+
+			// Round up when calculating microschedule time (+ 1 at the end)
+			time_us = (microschedule_lines * phantom->timing.h_total) /
+					(double)(phantom->timing.pix_clk_100hz * 100) * 1000000 +
+						dc->caps.subvp_prefetch_end_to_mall_start_us +
+						dc->caps.subvp_fw_processing_delay_us + 1;
+			if (time_us > max_microschedule_us)
+				max_microschedule_us = time_us;
+
+			subvp_pipes[index] = pipe;
+			index++;
+
+			// Maximum 2 SubVP pipes
+			if (index == 2)
+				break;
+		}
+	}
+	vactive1_us = ((subvp_pipes[0]->stream->timing.v_addressable * subvp_pipes[0]->stream->timing.h_total) /
+			(double)(subvp_pipes[0]->stream->timing.pix_clk_100hz * 100)) * 1000000;
+	vactive2_us = ((subvp_pipes[1]->stream->timing.v_addressable * subvp_pipes[1]->stream->timing.h_total) /
+				(double)(subvp_pipes[1]->stream->timing.pix_clk_100hz * 100)) * 1000000;
+	vblank1_us = (((subvp_pipes[0]->stream->timing.v_total - subvp_pipes[0]->stream->timing.v_addressable) *
+			subvp_pipes[0]->stream->timing.h_total) /
+			(double)(subvp_pipes[0]->stream->timing.pix_clk_100hz * 100)) * 1000000;
+	vblank2_us = (((subvp_pipes[1]->stream->timing.v_total - subvp_pipes[1]->stream->timing.v_addressable) *
+			subvp_pipes[1]->stream->timing.h_total) /
+			(double)(subvp_pipes[1]->stream->timing.pix_clk_100hz * 100)) * 1000000;
+
+	if ((vactive1_us - vblank2_us) / 2 > max_microschedule_us &&
+	    (vactive2_us - vblank1_us) / 2 > max_microschedule_us)
+		return true;
+
+	return false;
+}
+
+/**
+ * subvp_drr_schedulable() - Determine if SubVP + DRR config is schedulable
+ * @dc: current dc state
+ * @context: new dc state
+ *
+ * High level algorithm:
+ * 1. Get timing for SubVP pipe, phantom pipe, and DRR pipe
+ * 2. Determine the frame time for the DRR display when adding required margin for MCLK switching
+ * (the margin is equal to the MALL region + DRR margin (500us))
+ * 3.If (SubVP Active - Prefetch > Stretched DRR frame + max(MALL region, Stretched DRR frame))
+ * then report the configuration as supported
+ *
+ * Return: True if the SubVP + DRR config is schedulable, false otherwise
+ */
+static bool subvp_drr_schedulable(struct dc *dc, struct dc_state *context)
+{
+	bool schedulable = false;
+	uint32_t i;
+	struct pipe_ctx *pipe = NULL;
+	struct pipe_ctx *drr_pipe = NULL;
+	struct dc_crtc_timing *main_timing = NULL;
+	struct dc_crtc_timing *phantom_timing = NULL;
+	struct dc_crtc_timing *drr_timing = NULL;
+	int16_t prefetch_us = 0;
+	int16_t mall_region_us = 0;
+	int16_t drr_frame_us = 0;	// nominal frame time
+	int16_t subvp_active_us = 0;
+	int16_t stretched_drr_us = 0;
+	int16_t drr_stretched_vblank_us = 0;
+	int16_t max_vblank_mallregion = 0;
+	struct dc_stream_state *phantom_stream;
+	bool subvp_found = false;
+	bool drr_found = false;
+
+	// Find SubVP pipe
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		pipe = &context->res_ctx.pipe_ctx[i];
+
+		// We check for master pipe, but it shouldn't matter since we only need
+		// the pipe for timing info (stream should be same for any pipe splits)
+		if (!resource_is_pipe_type(pipe, OTG_MASTER) ||
+				!resource_is_pipe_type(pipe, DPP_PIPE))
+			continue;
+
+		// Find the SubVP pipe
+		if (dc_state_get_pipe_subvp_type(context, pipe) == SUBVP_MAIN) {
+			subvp_found = true;
+			break;
+		}
+	}
+
+	// Find the DRR pipe
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		drr_pipe = &context->res_ctx.pipe_ctx[i];
+
+		// We check for master pipe only
+		if (!resource_is_pipe_type(drr_pipe, OTG_MASTER) ||
+				!resource_is_pipe_type(drr_pipe, DPP_PIPE))
+			continue;
+
+		if (dc_state_get_pipe_subvp_type(context, drr_pipe) == SUBVP_NONE && drr_pipe->stream->ignore_msa_timing_param &&
+				(drr_pipe->stream->allow_freesync || drr_pipe->stream->vrr_active_variable || drr_pipe->stream->vrr_active_fixed)) {
+			drr_found = true;
+			break;
+		}
+	}
+
+	phantom_stream = dc_state_get_paired_subvp_stream(context, pipe->stream);
+	if (phantom_stream && subvp_found && drr_found) {
+		main_timing = &pipe->stream->timing;
+		phantom_timing = &phantom_stream->timing;
+		drr_timing = &drr_pipe->stream->timing;
+		prefetch_us = (phantom_timing->v_total - phantom_timing->v_front_porch) * phantom_timing->h_total /
+				(double)(phantom_timing->pix_clk_100hz * 100) * 1000000 +
+				dc->caps.subvp_prefetch_end_to_mall_start_us;
+		subvp_active_us = main_timing->v_addressable * main_timing->h_total /
+				(double)(main_timing->pix_clk_100hz * 100) * 1000000;
+		drr_frame_us = drr_timing->v_total * drr_timing->h_total /
+				(double)(drr_timing->pix_clk_100hz * 100) * 1000000;
+		// P-State allow width and FW delays already included phantom_timing->v_addressable
+		mall_region_us = phantom_timing->v_addressable * phantom_timing->h_total /
+				(double)(phantom_timing->pix_clk_100hz * 100) * 1000000;
+		stretched_drr_us = drr_frame_us + mall_region_us + SUBVP_DRR_MARGIN_US;
+		drr_stretched_vblank_us = (drr_timing->v_total - drr_timing->v_addressable) * drr_timing->h_total /
+				(double)(drr_timing->pix_clk_100hz * 100) * 1000000 + (stretched_drr_us - drr_frame_us);
+		max_vblank_mallregion = drr_stretched_vblank_us > mall_region_us ? drr_stretched_vblank_us : mall_region_us;
+	}
+
+	/* We consider SubVP + DRR schedulable if the stretched frame duration of the DRR display (i.e. the
+	 * highest refresh rate + margin that can support UCLK P-State switch) passes the static analysis
+	 * for VBLANK: (VACTIVE region of the SubVP pipe can fit the MALL prefetch, VBLANK frame time,
+	 * and the max of (VBLANK blanking time, MALL region)).
+	 */
+	if (drr_timing &&
+	    stretched_drr_us < (1 / (double)drr_timing->min_refresh_in_uhz) * 1000000 * 1000000 &&
+	    subvp_active_us - prefetch_us - stretched_drr_us - max_vblank_mallregion > 0)
+		schedulable = true;
+
+	return schedulable;
+}
+
+
+/**
+ * subvp_vblank_schedulable - Determine if SubVP + VBLANK config is schedulable
+ * @dc: current dc state
+ * @context: new dc state
+ *
+ * High level algorithm:
+ * 1. Get timing for SubVP pipe, phantom pipe, and VBLANK pipe
+ * 2. If (SubVP Active - Prefetch > Vblank Frame Time + max(MALL region, Vblank blanking time))
+ * then report the configuration as supported
+ * 3. If the VBLANK display is DRR, then take the DRR static schedulability path
+ *
+ * Return: True if the SubVP + VBLANK/DRR config is schedulable, false otherwise
+ */
+static bool subvp_vblank_schedulable(struct dc *dc, struct dc_state *context)
+{
+	struct pipe_ctx *pipe = NULL;
+	struct pipe_ctx *subvp_pipe = NULL;
+	bool found = false;
+	bool schedulable = false;
+	uint32_t i = 0;
+	uint8_t vblank_index = 0;
+	uint16_t prefetch_us = 0;
+	uint16_t mall_region_us = 0;
+	uint16_t vblank_frame_us = 0;
+	uint16_t subvp_active_us = 0;
+	uint16_t vblank_blank_us = 0;
+	uint16_t max_vblank_mallregion = 0;
+	struct dc_crtc_timing *main_timing = NULL;
+	struct dc_crtc_timing *phantom_timing = NULL;
+	struct dc_crtc_timing *vblank_timing = NULL;
+	struct dc_stream_state *phantom_stream;
+	enum mall_stream_type pipe_mall_type;
+
+	/* For SubVP + VBLANK/DRR cases, we assume there can only be
+	 * a single VBLANK/DRR display. If DML outputs SubVP + VBLANK
+	 * is supported, it is either a single VBLANK case or two VBLANK
+	 * displays which are synchronized (in which case they have identical
+	 * timings).
+	 */
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		pipe = &context->res_ctx.pipe_ctx[i];
+		pipe_mall_type = dc_state_get_pipe_subvp_type(context, pipe);
+
+		// We check for master pipe, but it shouldn't matter since we only need
+		// the pipe for timing info (stream should be same for any pipe splits)
+		if (!resource_is_pipe_type(pipe, OTG_MASTER) ||
+				!resource_is_pipe_type(pipe, DPP_PIPE))
+			continue;
+
+		if (!found && pipe_mall_type == SUBVP_NONE) {
+			// Found pipe which is not SubVP or Phantom (i.e. the VBLANK pipe).
+			vblank_index = i;
+			found = true;
+		}
+
+		if (!subvp_pipe && pipe_mall_type == SUBVP_MAIN)
+			subvp_pipe = pipe;
+	}
+	if (found && subvp_pipe) {
+		phantom_stream = dc_state_get_paired_subvp_stream(context, subvp_pipe->stream);
+		main_timing = &subvp_pipe->stream->timing;
+		phantom_timing = &phantom_stream->timing;
+		vblank_timing = &context->res_ctx.pipe_ctx[vblank_index].stream->timing;
+		// Prefetch time is equal to VACTIVE + BP + VSYNC of the phantom pipe
+		// Also include the prefetch end to mallstart delay time
+		prefetch_us = (phantom_timing->v_total - phantom_timing->v_front_porch) * phantom_timing->h_total /
+				(double)(phantom_timing->pix_clk_100hz * 100) * 1000000 +
+				dc->caps.subvp_prefetch_end_to_mall_start_us;
+		// P-State allow width and FW delays already included phantom_timing->v_addressable
+		mall_region_us = phantom_timing->v_addressable * phantom_timing->h_total /
+				(double)(phantom_timing->pix_clk_100hz * 100) * 1000000;
+		vblank_frame_us = vblank_timing->v_total * vblank_timing->h_total /
+				(double)(vblank_timing->pix_clk_100hz * 100) * 1000000;
+		vblank_blank_us =  (vblank_timing->v_total - vblank_timing->v_addressable) * vblank_timing->h_total /
+				(double)(vblank_timing->pix_clk_100hz * 100) * 1000000;
+		subvp_active_us = main_timing->v_addressable * main_timing->h_total /
+				(double)(main_timing->pix_clk_100hz * 100) * 1000000;
+		max_vblank_mallregion = vblank_blank_us > mall_region_us ? vblank_blank_us : mall_region_us;
+
+		// Schedulable if VACTIVE region of the SubVP pipe can fit the MALL prefetch, VBLANK frame time,
+		// and the max of (VBLANK blanking time, MALL region)
+		// TODO: Possibly add some margin (i.e. the below conditions should be [...] > X instead of [...] > 0)
+		if (subvp_active_us - prefetch_us - vblank_frame_us - max_vblank_mallregion > 0)
+			schedulable = true;
+	}
+	return schedulable;
+}
+
+/**
+ * subvp_subvp_admissable() - Determine if subvp + subvp config is admissible
+ *
+ * @dc: Current DC state
+ * @context: New DC state to be programmed
+ *
+ * SubVP + SubVP is admissible under the following conditions:
+ * - All SubVP pipes are < 120Hz OR
+ * - All SubVP pipes are >= 120hz
+ *
+ * Return: True if admissible, false otherwise
+ */
+static bool subvp_subvp_admissable(struct dc *dc,
+				struct dc_state *context)
+{
+	bool result = false;
+	uint32_t i;
+	uint8_t subvp_count = 0;
+	uint32_t min_refresh = subvp_high_refresh_list.min_refresh, max_refresh = 0;
+	uint64_t refresh_rate = 0;
+
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
+
+		if (!pipe->stream)
+			continue;
+
+		if (pipe->plane_state && !pipe->top_pipe &&
+				dc_state_get_pipe_subvp_type(context, pipe) == SUBVP_MAIN) {
+			refresh_rate = (pipe->stream->timing.pix_clk_100hz * (uint64_t)100 +
+				pipe->stream->timing.v_total * (uint64_t)pipe->stream->timing.h_total - (uint64_t)1);
+			refresh_rate = div_u64(refresh_rate, pipe->stream->timing.v_total);
+			refresh_rate = div_u64(refresh_rate, pipe->stream->timing.h_total);
+
+			if ((uint32_t)refresh_rate < min_refresh)
+				min_refresh = (uint32_t)refresh_rate;
+			if ((uint32_t)refresh_rate > max_refresh)
+				max_refresh = (uint32_t)refresh_rate;
+			subvp_count++;
+		}
+	}
+
+	if (subvp_count == 2 && ((min_refresh < 120 && max_refresh < 120) ||
+		(min_refresh >= subvp_high_refresh_list.min_refresh &&
+				max_refresh <= subvp_high_refresh_list.max_refresh)))
+		result = true;
+
+	return result;
+}
+
+/**
+ * subvp_validate_static_schedulability - Check which SubVP case is calculated
+ * and handle static analysis based on the case.
+ * @dc: current dc state
+ * @context: new dc state
+ * @vlevel: Voltage level calculated by DML
+ *
+ * Three cases:
+ * 1. SubVP + SubVP
+ * 2. SubVP + VBLANK (DRR checked internally)
+ * 3. SubVP + VACTIVE (currently unsupported)
+ *
+ * Return: True if statically schedulable, false otherwise
+ */
+static bool subvp_validate_static_schedulability(struct dc *dc,
+				struct dc_state *context,
+				int vlevel)
+{
+	bool schedulable = false;
+	struct vba_vars_st *vba = &context->bw_ctx.dml.vba;
+	uint32_t i, pipe_idx;
+	uint8_t subvp_count = 0;
+	uint8_t vactive_count = 0;
+	uint8_t non_subvp_pipes = 0;
+
+	for (i = 0, pipe_idx = 0; i < dc->res_pool->pipe_count; i++) {
+		struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
+		enum mall_stream_type pipe_mall_type = dc_state_get_pipe_subvp_type(context, pipe);
+
+		if (!pipe->stream)
+			continue;
+
+		if (pipe->plane_state && !pipe->top_pipe) {
+			if (pipe_mall_type == SUBVP_MAIN)
+				subvp_count++;
+			if (pipe_mall_type == SUBVP_NONE)
+				non_subvp_pipes++;
+		}
+
+		// Count how many planes that aren't SubVP/phantom are capable of VACTIVE
+		// switching (SubVP + VACTIVE unsupported). In situations where we force
+		// SubVP for a VACTIVE plane, we don't want to increment the vactive_count.
+		if (vba->ActiveDRAMClockChangeLatencyMarginPerState[vlevel][vba->maxMpcComb][vba->pipe_plane[pipe_idx]] > 0 &&
+				pipe_mall_type == SUBVP_NONE) {
+			vactive_count++;
+		}
+		pipe_idx++;
+	}
+
+	if (subvp_count == 2) {
+		// Static schedulability check for SubVP + SubVP case
+		schedulable = subvp_subvp_admissable(dc, context) && subvp_subvp_schedulable(dc, context);
+	} else if (subvp_count == 1 && non_subvp_pipes == 0) {
+		// Single SubVP configs will be supported by default as long as it's suppported by DML
+		schedulable = true;
+	} else if (subvp_count == 1 && non_subvp_pipes == 1) {
+		if (dcn32_subvp_drr_admissable(dc, context))
+			schedulable = subvp_drr_schedulable(dc, context);
+		else if (dcn32_subvp_vblank_admissable(dc, context, vlevel))
+			schedulable = subvp_vblank_schedulable(dc, context);
+	} else if (vba->DRAMClockChangeSupport[vlevel][vba->maxMpcComb] == dm_dram_clock_change_vactive_w_mall_sub_vp &&
+			vactive_count > 0) {
+		// For single display SubVP cases, DML will output dm_dram_clock_change_vactive_w_mall_sub_vp by default.
+		// We tell the difference between SubVP vs. SubVP + VACTIVE by checking the vactive_count.
+		// SubVP + VACTIVE currently unsupported
+		schedulable = false;
+	}
+	return schedulable;
+}
+
+static void assign_subvp_index(struct dc *dc, struct dc_state *context)
+{
+	int i;
+	int index = 0;
+
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i];
+
+		if (resource_is_pipe_type(pipe_ctx, OTG_MASTER) &&
+				dc_state_get_pipe_subvp_type(context, pipe_ctx) == SUBVP_MAIN) {
+			pipe_ctx->subvp_index = index++;
+		} else {
+			pipe_ctx->subvp_index = 0;
+		}
+	}
+}
+
+struct pipe_slice_table {
+	struct {
+		struct dc_stream_state *stream;
+		int slice_count;
+	} odm_combines[MAX_STREAMS];
+	int odm_combine_count;
+
+	struct {
+		struct pipe_ctx *pri_pipe;
+		struct dc_plane_state *plane;
+		int slice_count;
+	} mpc_combines[MAX_PLANES];
+	int mpc_combine_count;
+};
+
+
+static void update_slice_table_for_stream(struct pipe_slice_table *table,
+		struct dc_stream_state *stream, int diff)
+{
+	int i;
+
+	for (i = 0; i < table->odm_combine_count; i++) {
+		if (table->odm_combines[i].stream == stream) {
+			table->odm_combines[i].slice_count += diff;
+			break;
+		}
+	}
+
+	if (i == table->odm_combine_count) {
+		table->odm_combine_count++;
+		table->odm_combines[i].stream = stream;
+		table->odm_combines[i].slice_count = diff;
+	}
+}
+
+static void update_slice_table_for_plane(struct pipe_slice_table *table,
+		struct pipe_ctx *dpp_pipe, struct dc_plane_state *plane, int diff)
+{
+	int i;
+	struct pipe_ctx *pri_dpp_pipe = resource_get_primary_dpp_pipe(dpp_pipe);
+
+	for (i = 0; i < table->mpc_combine_count; i++) {
+		if (table->mpc_combines[i].plane == plane &&
+				table->mpc_combines[i].pri_pipe == pri_dpp_pipe) {
+			table->mpc_combines[i].slice_count += diff;
+			break;
+		}
+	}
+
+	if (i == table->mpc_combine_count) {
+		table->mpc_combine_count++;
+		table->mpc_combines[i].plane = plane;
+		table->mpc_combines[i].pri_pipe = pri_dpp_pipe;
+		table->mpc_combines[i].slice_count = diff;
+	}
+}
+
+static void init_pipe_slice_table_from_context(
+		struct pipe_slice_table *table,
+		struct dc_state *context)
+{
+	int i, j;
+	struct pipe_ctx *otg_master;
+	struct pipe_ctx *dpp_pipes[MAX_PIPES];
+	struct dc_stream_state *stream;
+	int count;
+
+	memset(table, 0, sizeof(*table));
+
+	for (i = 0; i < context->stream_count; i++) {
+		stream = context->streams[i];
+		otg_master = resource_get_otg_master_for_stream(
+				&context->res_ctx, stream);
+		if (!otg_master)
+			continue;
+
+		count = resource_get_odm_slice_count(otg_master);
+		update_slice_table_for_stream(table, stream, count);
+
+		count = resource_get_dpp_pipes_for_opp_head(otg_master,
+				&context->res_ctx, dpp_pipes);
+		for (j = 0; j < count; j++)
+			if (dpp_pipes[j]->plane_state)
+				update_slice_table_for_plane(table, dpp_pipes[j],
+						dpp_pipes[j]->plane_state, 1);
+	}
+}
+
+static bool update_pipe_slice_table_with_split_flags(
+		struct pipe_slice_table *table,
+		struct dc *dc,
+		struct dc_state *context,
+		struct vba_vars_st *vba,
+		int split[MAX_PIPES],
+		bool merge[MAX_PIPES])
+{
+	/* NOTE: we are deprecating the support for the concept of pipe splitting
+	 * or pipe merging. Instead we append slices to the end and remove
+	 * slices from the end. The following code converts a pipe split or
+	 * merge to an append or remove operation.
+	 *
+	 * For example:
+	 * When split flags describe the following pipe connection transition
+	 *
+	 * from:
+	 *  pipe 0 (split=2) -> pipe 1 (split=2)
+	 * to: (old behavior)
+	 *  pipe 0 -> pipe 2 -> pipe 1 -> pipe 3
+	 *
+	 * the code below actually does:
+	 *  pipe 0 -> pipe 1 -> pipe 2 -> pipe 3
+	 *
+	 * This is the new intended behavior and for future DCNs we will retire
+	 * the old concept completely.
+	 */
+	struct pipe_ctx *pipe;
+	bool odm;
+	int dc_pipe_idx, dml_pipe_idx = 0;
+	bool updated = false;
+
+	for (dc_pipe_idx = 0;
+			dc_pipe_idx < dc->res_pool->pipe_count; dc_pipe_idx++) {
+		pipe = &context->res_ctx.pipe_ctx[dc_pipe_idx];
+		if (resource_is_pipe_type(pipe, FREE_PIPE))
+			continue;
+
+		if (merge[dc_pipe_idx]) {
+			if (resource_is_pipe_type(pipe, OPP_HEAD))
+				/* merging OPP head means reducing ODM slice
+				 * count by 1
+				 */
+				update_slice_table_for_stream(table, pipe->stream, -1);
+			else if (resource_is_pipe_type(pipe, DPP_PIPE) &&
+					resource_get_odm_slice_index(resource_get_opp_head(pipe)) == 0)
+				/* merging DPP pipe of the first ODM slice means
+				 * reducing MPC slice count by 1
+				 */
+				update_slice_table_for_plane(table, pipe, pipe->plane_state, -1);
+			updated = true;
+		}
+
+		if (split[dc_pipe_idx]) {
+			odm = vba->ODMCombineEnabled[vba->pipe_plane[dml_pipe_idx]] !=
+					dm_odm_combine_mode_disabled;
+			if (odm && resource_is_pipe_type(pipe, OPP_HEAD))
+				update_slice_table_for_stream(
+						table, pipe->stream, split[dc_pipe_idx] - 1);
+			else if (!odm && resource_is_pipe_type(pipe, DPP_PIPE))
+				update_slice_table_for_plane(table, pipe,
+						pipe->plane_state, split[dc_pipe_idx] - 1);
+			updated = true;
+		}
+		dml_pipe_idx++;
+	}
+	return updated;
+}
+
+static void update_pipes_with_slice_table(struct dc *dc, struct dc_state *context,
+		struct pipe_slice_table *table)
+{
+	int i;
+
+	for (i = 0; i < table->odm_combine_count; i++)
+		resource_update_pipes_for_stream_with_slice_count(context,
+				dc->current_state, dc->res_pool,
+				table->odm_combines[i].stream,
+				table->odm_combines[i].slice_count);
+
+	for (i = 0; i < table->mpc_combine_count; i++)
+		resource_update_pipes_for_plane_with_slice_count(context,
+				dc->current_state, dc->res_pool,
+				table->mpc_combines[i].plane,
+				table->mpc_combines[i].slice_count);
+}
+
+static bool update_pipes_with_split_flags(struct dc *dc, struct dc_state *context,
+		struct vba_vars_st *vba, int split[MAX_PIPES],
+		bool merge[MAX_PIPES])
+{
+	struct pipe_slice_table slice_table;
+	bool updated;
+
+	init_pipe_slice_table_from_context(&slice_table, context);
+	updated = update_pipe_slice_table_with_split_flags(
+			&slice_table, dc, context, vba,
+			split, merge);
+	update_pipes_with_slice_table(dc, context, &slice_table);
+	return updated;
+}
+
+static bool should_apply_odm_power_optimization(struct dc *dc,
+		struct dc_state *context, struct vba_vars_st *v, int *split,
+		bool *merge)
+{
+	struct dc_stream_state *stream = context->streams[0];
+	struct pipe_slice_table slice_table;
+	int i;
+
+	/*
+	 * this debug flag allows us to disable ODM power optimization feature
+	 * unconditionally. we force the feature off if this is set to false.
+	 */
+	if (!dc->debug.enable_single_display_2to1_odm_policy)
+		return false;
+
+	/* current design and test coverage is only limited to allow ODM power
+	 * optimization for single stream. Supporting it for multiple streams
+	 * use case would require additional algorithm to decide how to
+	 * optimize power consumption when there are not enough free pipes to
+	 * allocate for all the streams. This level of optimization would
+	 * require multiple attempts of revalidation to make an optimized
+	 * decision. Unfortunately We do not support revalidation flow in
+	 * current version of DML.
+	 */
+	if (context->stream_count != 1)
+		return false;
+
+	/*
+	 * Our hardware doesn't support ODM for HDMI TMDS
+	 */
+	if (dc_is_hdmi_signal(stream->signal))
+		return false;
+
+	/*
+	 * ODM Combine 2:1 requires horizontal timing divisible by 2 so each
+	 * ODM segment has the same size.
+	 */
+	if (!is_h_timing_divisible_by_2(stream))
+		return false;
+
+	/*
+	 * No power benefits if the timing's pixel clock is not high enough to
+	 * raise display clock from minimum power state.
+	 */
+	if (stream->timing.pix_clk_100hz * 100 <= DCN3_2_VMIN_DISPCLK_HZ)
+		return false;
+
+	if (dc->config.enable_windowed_mpo_odm) {
+		/*
+		 * ODM power optimization should only be allowed if the feature
+		 * can be seamlessly toggled off within an update. This would
+		 * require that the feature is applied on top of a minimal
+		 * state. A minimal state is defined as a state validated
+		 * without the need of pipe split. Therefore, when transition to
+		 * toggle the feature off, the same stream and plane
+		 * configuration can be supported by the pipe resource in the
+		 * first ODM slice alone without the need to acquire extra
+		 * resources.
+		 */
+		init_pipe_slice_table_from_context(&slice_table, context);
+		update_pipe_slice_table_with_split_flags(
+				&slice_table, dc, context, v,
+				split, merge);
+		for (i = 0; i < slice_table.mpc_combine_count; i++)
+			if (slice_table.mpc_combines[i].slice_count > 1)
+				return false;
+
+		for (i = 0; i < slice_table.odm_combine_count; i++)
+			if (slice_table.odm_combines[i].slice_count > 1)
+				return false;
+	} else {
+		/*
+		 * the new ODM power optimization feature reduces software
+		 * design limitation and allows ODM power optimization to be
+		 * supported even with presence of overlay planes. The new
+		 * feature is enabled based on enable_windowed_mpo_odm flag. If
+		 * the flag is not set, we limit our feature scope due to
+		 * previous software design limitation
+		 */
+		if (context->stream_status[0].plane_count != 1)
+			return false;
+
+		if (memcmp(&context->stream_status[0].plane_states[0]->clip_rect,
+				&stream->src, sizeof(struct rect)) != 0)
+			return false;
+
+		if (stream->src.width >= 5120 &&
+				stream->src.width > stream->dst.width)
+			return false;
+	}
+	return true;
+}
+
+static void try_odm_power_optimization_and_revalidate(
+		struct dc *dc,
+		struct dc_state *context,
+		display_e2e_pipe_params_st *pipes,
+		int *split,
+		bool *merge,
+		unsigned int *vlevel,
+		int pipe_cnt)
+{
+	int i;
+	unsigned int new_vlevel;
+	unsigned int cur_policy[MAX_PIPES];
+
+	for (i = 0; i < pipe_cnt; i++) {
+		cur_policy[i] = pipes[i].pipe.dest.odm_combine_policy;
+		pipes[i].pipe.dest.odm_combine_policy = dm_odm_combine_policy_2to1;
+	}
+
+	new_vlevel = dml_get_voltage_level(&context->bw_ctx.dml, pipes, pipe_cnt);
+
+	if (new_vlevel < context->bw_ctx.dml.soc.num_states) {
+		memset(split, 0, MAX_PIPES * sizeof(int));
+		memset(merge, 0, MAX_PIPES * sizeof(bool));
+		*vlevel = dcn20_validate_apply_pipe_split_flags(dc, context, new_vlevel, split, merge);
+		context->bw_ctx.dml.vba.VoltageLevel = *vlevel;
+	} else {
+		for (i = 0; i < pipe_cnt; i++)
+			pipes[i].pipe.dest.odm_combine_policy = cur_policy[i];
+	}
+}
+
+static bool is_test_pattern_enabled(
+		struct dc_state *context)
+{
+	int i;
+
+	for (i = 0; i < context->stream_count; i++) {
+		if (context->streams[i]->test_pattern.type != DP_TEST_PATTERN_VIDEO_MODE)
+			return true;
+	}
+
+	return false;
+}
+
+static bool dcn32_full_validate_bw_helper(struct dc *dc,
+				   struct dc_state *context,
+				   display_e2e_pipe_params_st *pipes,
+				   int *vlevel,
+				   int *split,
+				   bool *merge,
+				   int *pipe_cnt,
+				   bool *repopulate_pipes)
+{
+	struct vba_vars_st *vba = &context->bw_ctx.dml.vba;
+	unsigned int dc_pipe_idx = 0;
+	int i = 0;
+	bool found_supported_config = false;
+	int vlevel_temp = 0;
+
+	dc_assert_fp_enabled();
+
+	/*
+	 * DML favors voltage over p-state, but we're more interested in
+	 * supporting p-state over voltage. We can't support p-state in
+	 * prefetch mode > 0 so try capping the prefetch mode to start.
+	 * Override present for testing.
+	 */
+	if (dc->debug.dml_disallow_alternate_prefetch_modes)
+		context->bw_ctx.dml.soc.allow_for_pstate_or_stutter_in_vblank_final =
+			dm_prefetch_support_uclk_fclk_and_stutter;
+	else
+		context->bw_ctx.dml.soc.allow_for_pstate_or_stutter_in_vblank_final =
+			dm_prefetch_support_uclk_fclk_and_stutter_if_possible;
+
+	*vlevel = dml_get_voltage_level(&context->bw_ctx.dml, pipes, *pipe_cnt);
+	/* This may adjust vlevel and maxMpcComb */
+	if (*vlevel < context->bw_ctx.dml.soc.num_states) {
+		*vlevel = dcn20_validate_apply_pipe_split_flags(dc, context, *vlevel, split, merge);
+		vba->VoltageLevel = *vlevel;
+	}
+
+	/* Apply split and merge flags before checking for subvp */
+	if (!dcn32_apply_merge_split_flags_helper(dc, context, repopulate_pipes, split, merge))
+		return false;
+	memset(split, 0, MAX_PIPES * sizeof(int));
+	memset(merge, 0, MAX_PIPES * sizeof(bool));
+
+	/* Conditions for setting up phantom pipes for SubVP:
+	 * 1. Not force disable SubVP
+	 * 2. Full update (i.e. DC_VALIDATE_MODE_AND_PROGRAMMING)
+	 * 3. Enough pipes are available to support SubVP (TODO: Which pipes will use VACTIVE / VBLANK / SUBVP?)
+	 * 4. Display configuration passes validation
+	 * 5. (Config doesn't support MCLK in VACTIVE/VBLANK || dc->debug.force_subvp_mclk_switch)
+	 */
+	if (!dc->debug.force_disable_subvp && !dc->caps.dmub_caps.gecc_enable && dcn32_all_pipes_have_stream_and_plane(dc, context) &&
+	    !dcn32_mpo_in_use(context) && !dcn32_any_surfaces_rotated(dc, context) && !is_test_pattern_enabled(context) &&
+		(*vlevel == context->bw_ctx.dml.soc.num_states || (vba->DRAMSpeedPerState[*vlevel] != vba->DRAMSpeedPerState[0] &&
+				vba->DRAMClockChangeSupport[*vlevel][vba->maxMpcComb] != dm_dram_clock_change_unsupported) ||
+	    vba->DRAMClockChangeSupport[*vlevel][vba->maxMpcComb] == dm_dram_clock_change_unsupported ||
+	    dc->debug.force_subvp_mclk_switch)) {
+
+		vlevel_temp = *vlevel;
+
+		while (!found_supported_config && dcn32_enough_pipes_for_subvp(dc, context) &&
+			dcn32_assign_subvp_pipe(dc, context, &dc_pipe_idx)) {
+			/* For the case where *vlevel = num_states, bandwidth validation has failed for this config.
+			 * Adding phantom pipes won't change the validation result, so change the DML input param
+			 * for P-State support before adding phantom pipes and recalculating the DML result.
+			 * However, this case is only applicable for SubVP + DRR cases because the prefetch mode
+			 * will not allow for switch in VBLANK. The DRR display must have it's VBLANK stretched
+			 * enough to support MCLK switching.
+			 */
+			if (*vlevel == context->bw_ctx.dml.soc.num_states &&
+				context->bw_ctx.dml.soc.allow_for_pstate_or_stutter_in_vblank_final ==
+					dm_prefetch_support_uclk_fclk_and_stutter) {
+				context->bw_ctx.dml.soc.allow_for_pstate_or_stutter_in_vblank_final =
+								dm_prefetch_support_fclk_and_stutter;
+				/* There are params (such as FabricClock) that need to be recalculated
+				 * after validation fails (otherwise it will be 0). Calculation for
+				 * phantom vactive requires call into DML, so we must ensure all the
+				 * vba params are valid otherwise we'll get incorrect phantom vactive.
+				 */
+				*vlevel = dml_get_voltage_level(&context->bw_ctx.dml, pipes, *pipe_cnt);
+			}
+
+			dc->res_pool->funcs->add_phantom_pipes(dc, context, pipes, *pipe_cnt, dc_pipe_idx);
+
+			*pipe_cnt = dc->res_pool->funcs->populate_dml_pipes(dc, context, pipes,
+				DC_VALIDATE_MODE_AND_PROGRAMMING);
+			// Populate dppclk to trigger a recalculate in dml_get_voltage_level
+			// so the phantom pipe DLG params can be assigned correctly.
+			pipes[0].clks_cfg.dppclk_mhz = get_dppclk_calculated(&context->bw_ctx.dml, pipes, *pipe_cnt, 0);
+			*vlevel = dml_get_voltage_level(&context->bw_ctx.dml, pipes, *pipe_cnt);
+
+			/* Check that vlevel requested supports pstate or not
+			 * if not, select the lowest vlevel that supports it
+			 */
+			for (i = *vlevel; i < context->bw_ctx.dml.soc.num_states; i++) {
+				if (vba->DRAMClockChangeSupport[i][vba->maxMpcComb] != dm_dram_clock_change_unsupported) {
+					*vlevel = i;
+					break;
+				}
+			}
+
+			if (*vlevel < context->bw_ctx.dml.soc.num_states
+			    && subvp_validate_static_schedulability(dc, context, *vlevel))
+				found_supported_config = true;
+			if (found_supported_config) {
+				// For SubVP + DRR cases, we can force the lowest vlevel that supports the mode
+				if (dcn32_subvp_drr_admissable(dc, context) && subvp_drr_schedulable(dc, context)) {
+					/* find lowest vlevel that supports the config */
+					for (i = *vlevel; i >= 0; i--) {
+						if (vba->ModeSupport[i][vba->maxMpcComb]) {
+							*vlevel = i;
+						} else {
+							break;
+						}
+					}
+				}
+			}
+		}
+
+		if (vba->DRAMSpeedPerState[*vlevel] >= vba->DRAMSpeedPerState[vlevel_temp])
+			found_supported_config = false;
+
+		// If SubVP pipe config is unsupported (or cannot be used for UCLK switching)
+		// remove phantom pipes and repopulate dml pipes
+		if (!found_supported_config) {
+			dc_state_remove_phantom_streams_and_planes(dc, context);
+			dc_state_release_phantom_streams_and_planes(dc, context);
+			vba->DRAMClockChangeSupport[*vlevel][vba->maxMpcComb] = dm_dram_clock_change_unsupported;
+			*pipe_cnt = dc->res_pool->funcs->populate_dml_pipes(dc, context, pipes,
+				DC_VALIDATE_MODE_AND_PROGRAMMING);
+
+			*vlevel = dml_get_voltage_level(&context->bw_ctx.dml, pipes, *pipe_cnt);
+			/* This may adjust vlevel and maxMpcComb */
+			if (*vlevel < context->bw_ctx.dml.soc.num_states) {
+				*vlevel = dcn20_validate_apply_pipe_split_flags(dc, context, *vlevel, split, merge);
+				vba->VoltageLevel = *vlevel;
+			}
+		} else {
+			// Most populate phantom DLG params before programming hardware / timing for phantom pipe
+			dcn32_helper_populate_phantom_dlg_params(dc, context, pipes, *pipe_cnt);
+
+			/* Call validate_apply_pipe_split flags after calling DML getters for
+			 * phantom dlg params, or some of the VBA params indicating pipe split
+			 * can be overwritten by the getters.
+			 *
+			 * When setting up SubVP config, all pipes are merged before attempting to
+			 * add phantom pipes. If pipe split (ODM / MPC) is required, both the main
+			 * and phantom pipes will be split in the regular pipe splitting sequence.
+			 */
+			*vlevel = dcn20_validate_apply_pipe_split_flags(dc, context, *vlevel, split, merge);
+			vba->VoltageLevel = *vlevel;
+			// Note: We can't apply the phantom pipes to hardware at this time. We have to wait
+			// until driver has acquired the DMCUB lock to do it safely.
+			assign_subvp_index(dc, context);
+		}
+	}
+
+	if (should_apply_odm_power_optimization(dc, context, vba, split, merge))
+		try_odm_power_optimization_and_revalidate(
+				dc, context, pipes, split, merge, vlevel, *pipe_cnt);
+
+	return true;
+}
+
+static bool is_dtbclk_required(struct dc *dc, struct dc_state *context)
+{
+	int i;
+
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		if (!context->res_ctx.pipe_ctx[i].stream)
+			continue;
+		if (dc->link_srv->dp_is_128b_132b_signal(&context->res_ctx.pipe_ctx[i]))
+			return true;
+	}
+	return false;
+}
+
+static void dcn20_adjust_freesync_v_startup(const struct dc_crtc_timing *dc_crtc_timing, int *vstartup_start)
+{
+	struct dc_crtc_timing patched_crtc_timing;
+	uint32_t asic_blank_end   = 0;
+	uint32_t asic_blank_start = 0;
+	uint32_t newVstartup	  = 0;
+
+	patched_crtc_timing = *dc_crtc_timing;
+
+	if (patched_crtc_timing.flags.INTERLACE == 1) {
+		if (patched_crtc_timing.v_front_porch < 2)
+			patched_crtc_timing.v_front_porch = 2;
+	} else {
+		if (patched_crtc_timing.v_front_porch < 1)
+			patched_crtc_timing.v_front_porch = 1;
+	}
+
+	/* blank_start = frame end - front porch */
+	asic_blank_start = patched_crtc_timing.v_total -
+					patched_crtc_timing.v_front_porch;
+
+	/* blank_end = blank_start - active */
+	asic_blank_end = asic_blank_start -
+					patched_crtc_timing.v_border_bottom -
+					patched_crtc_timing.v_addressable -
+					patched_crtc_timing.v_border_top;
+
+	newVstartup = asic_blank_end + (patched_crtc_timing.v_total - asic_blank_start);
+
+	*vstartup_start = ((newVstartup > *vstartup_start) ? newVstartup : *vstartup_start);
+}
+
+static void dcn32_calculate_dlg_params(struct dc *dc, struct dc_state *context,
+				       display_e2e_pipe_params_st *pipes,
+				       int pipe_cnt, int vlevel)
+{
+	int i, pipe_idx, active_hubp_count = 0;
+	bool usr_retraining_support = false;
+	bool unbounded_req_enabled = false;
+	struct vba_vars_st *vba = &context->bw_ctx.dml.vba;
+
+	dc_assert_fp_enabled();
+
+	/* Writeback MCIF_WB arbitration parameters */
+	dc->res_pool->funcs->set_mcif_arb_params(dc, context, pipes, pipe_cnt);
+
+	context->bw_ctx.bw.dcn.clk.dispclk_khz = context->bw_ctx.dml.vba.DISPCLK * 1000;
+	context->bw_ctx.bw.dcn.clk.dcfclk_khz = context->bw_ctx.dml.vba.DCFCLK * 1000;
+	context->bw_ctx.bw.dcn.clk.socclk_khz = context->bw_ctx.dml.vba.SOCCLK * 1000;
+	context->bw_ctx.bw.dcn.clk.dramclk_khz = context->bw_ctx.dml.vba.DRAMSpeed * 1000 / 16;
+	context->bw_ctx.bw.dcn.clk.dcfclk_deep_sleep_khz = context->bw_ctx.dml.vba.DCFCLKDeepSleep * 1000;
+	context->bw_ctx.bw.dcn.clk.fclk_khz = context->bw_ctx.dml.vba.FabricClock * 1000;
+	context->bw_ctx.bw.dcn.clk.p_state_change_support =
+			context->bw_ctx.dml.vba.DRAMClockChangeSupport[vlevel][context->bw_ctx.dml.vba.maxMpcComb]
+					!= dm_dram_clock_change_unsupported;
+
+	/* Pstate change might not be supported by hardware, but it might be
+	 * possible with firmware driven vertical blank stretching.
+	 */
+	context->bw_ctx.bw.dcn.clk.p_state_change_support |= context->bw_ctx.bw.dcn.clk.fw_based_mclk_switching;
+
+	context->bw_ctx.bw.dcn.clk.dppclk_khz = 0;
+	context->bw_ctx.bw.dcn.clk.dtbclk_en = is_dtbclk_required(dc, context);
+	context->bw_ctx.bw.dcn.clk.ref_dtbclk_khz = context->bw_ctx.dml.vba.DTBCLKPerState[vlevel] * 1000;
+	if (context->bw_ctx.dml.vba.FCLKChangeSupport[vlevel][context->bw_ctx.dml.vba.maxMpcComb] == dm_fclock_change_unsupported)
+		context->bw_ctx.bw.dcn.clk.fclk_p_state_change_support = false;
+	else
+		context->bw_ctx.bw.dcn.clk.fclk_p_state_change_support = true;
+
+	usr_retraining_support = context->bw_ctx.dml.vba.USRRetrainingSupport[vlevel][context->bw_ctx.dml.vba.maxMpcComb];
+	ASSERT(usr_retraining_support);
+
+	if (context->bw_ctx.bw.dcn.clk.dispclk_khz < dc->debug.min_disp_clk_khz)
+		context->bw_ctx.bw.dcn.clk.dispclk_khz = dc->debug.min_disp_clk_khz;
+
+	unbounded_req_enabled = get_unbounded_request_enabled(&context->bw_ctx.dml, pipes, pipe_cnt);
+
+	if (unbounded_req_enabled && pipe_cnt > 1) {
+		// Unbounded requesting should not ever be used when more than 1 pipe is enabled.
+		ASSERT(false);
+		unbounded_req_enabled = false;
+	}
+
+	context->bw_ctx.bw.dcn.mall_ss_size_bytes = 0;
+	context->bw_ctx.bw.dcn.mall_ss_psr_active_size_bytes = 0;
+	context->bw_ctx.bw.dcn.mall_subvp_size_bytes = 0;
+
+	for (i = 0, pipe_idx = 0; i < dc->res_pool->pipe_count; i++) {
+		if (!context->res_ctx.pipe_ctx[i].stream)
+			continue;
+		if (context->res_ctx.pipe_ctx[i].plane_state)
+			active_hubp_count++;
+		pipes[pipe_idx].pipe.dest.vstartup_start = get_vstartup(&context->bw_ctx.dml, pipes, pipe_cnt,
+				pipe_idx);
+		pipes[pipe_idx].pipe.dest.vupdate_offset = get_vupdate_offset(&context->bw_ctx.dml, pipes, pipe_cnt,
+				pipe_idx);
+		pipes[pipe_idx].pipe.dest.vupdate_width = get_vupdate_width(&context->bw_ctx.dml, pipes, pipe_cnt,
+				pipe_idx);
+		pipes[pipe_idx].pipe.dest.vready_offset = get_vready_offset(&context->bw_ctx.dml, pipes, pipe_cnt,
+				pipe_idx);
+
+		if (dc_state_get_pipe_subvp_type(context, &context->res_ctx.pipe_ctx[i]) == SUBVP_PHANTOM) {
+			// Phantom pipe requires that DET_SIZE = 0 and no unbounded requests
+			context->res_ctx.pipe_ctx[i].det_buffer_size_kb = 0;
+			context->res_ctx.pipe_ctx[i].unbounded_req = false;
+		} else {
+			context->res_ctx.pipe_ctx[i].det_buffer_size_kb = get_det_buffer_size_kbytes(&context->bw_ctx.dml, pipes, pipe_cnt,
+							pipe_idx);
+			context->res_ctx.pipe_ctx[i].unbounded_req = unbounded_req_enabled;
+		}
+
+		if (context->bw_ctx.bw.dcn.clk.dppclk_khz < pipes[pipe_idx].clks_cfg.dppclk_mhz * 1000)
+			context->bw_ctx.bw.dcn.clk.dppclk_khz = pipes[pipe_idx].clks_cfg.dppclk_mhz * 1000;
+		if (context->res_ctx.pipe_ctx[i].plane_state)
+			context->res_ctx.pipe_ctx[i].plane_res.bw.dppclk_khz = pipes[pipe_idx].clks_cfg.dppclk_mhz * 1000;
+		else
+			context->res_ctx.pipe_ctx[i].plane_res.bw.dppclk_khz = 0;
+		context->res_ctx.pipe_ctx[i].pipe_dlg_param = pipes[pipe_idx].pipe.dest;
+
+		context->res_ctx.pipe_ctx[i].surface_size_in_mall_bytes = get_surface_size_in_mall(&context->bw_ctx.dml, pipes, pipe_cnt, pipe_idx);
+
+		if (vba->ActiveDRAMClockChangeLatencyMarginPerState[vba->VoltageLevel][vba->maxMpcComb][vba->pipe_plane[pipe_idx]] > 0)
+			context->res_ctx.pipe_ctx[i].has_vactive_margin = true;
+		else
+			context->res_ctx.pipe_ctx[i].has_vactive_margin = false;
+
+		/* MALL Allocation Sizes */
+		/* count from active, top pipes per plane only */
+		if (context->res_ctx.pipe_ctx[i].stream && context->res_ctx.pipe_ctx[i].plane_state &&
+				(context->res_ctx.pipe_ctx[i].top_pipe == NULL ||
+				context->res_ctx.pipe_ctx[i].plane_state != context->res_ctx.pipe_ctx[i].top_pipe->plane_state) &&
+				context->res_ctx.pipe_ctx[i].prev_odm_pipe == NULL) {
+			/* SS: all active surfaces stored in MALL */
+			if (dc_state_get_pipe_subvp_type(context, &context->res_ctx.pipe_ctx[i]) != SUBVP_PHANTOM) {
+				context->bw_ctx.bw.dcn.mall_ss_size_bytes += context->res_ctx.pipe_ctx[i].surface_size_in_mall_bytes;
+
+				if (context->res_ctx.pipe_ctx[i].stream->link->psr_settings.psr_version == DC_PSR_VERSION_UNSUPPORTED) {
+					/* SS PSR On: all active surfaces part of streams not supporting PSR stored in MALL */
+					context->bw_ctx.bw.dcn.mall_ss_psr_active_size_bytes += context->res_ctx.pipe_ctx[i].surface_size_in_mall_bytes;
+				}
+			} else {
+				/* SUBVP: phantom surfaces only stored in MALL */
+				context->bw_ctx.bw.dcn.mall_subvp_size_bytes += context->res_ctx.pipe_ctx[i].surface_size_in_mall_bytes;
+			}
+		}
+
+		if (context->res_ctx.pipe_ctx[i].stream->adaptive_sync_infopacket.valid)
+			dcn20_adjust_freesync_v_startup(
+				&context->res_ctx.pipe_ctx[i].stream->timing,
+				&context->res_ctx.pipe_ctx[i].pipe_dlg_param.vstartup_start);
+
+		pipe_idx++;
+	}
+	/* If DCN isn't making memory requests we can allow pstate change and lower clocks */
+	if (!active_hubp_count) {
+		context->bw_ctx.bw.dcn.clk.socclk_khz = 0;
+		context->bw_ctx.bw.dcn.clk.dppclk_khz = 0;
+		context->bw_ctx.bw.dcn.clk.dcfclk_khz = 0;
+		context->bw_ctx.bw.dcn.clk.dcfclk_deep_sleep_khz = 0;
+		context->bw_ctx.bw.dcn.clk.dramclk_khz = 0;
+		context->bw_ctx.bw.dcn.clk.fclk_khz = 0;
+		context->bw_ctx.bw.dcn.clk.p_state_change_support = true;
+		context->bw_ctx.bw.dcn.clk.fclk_p_state_change_support = true;
+	}
+	/*save a original dppclock copy*/
+	context->bw_ctx.bw.dcn.clk.bw_dppclk_khz = context->bw_ctx.bw.dcn.clk.dppclk_khz;
+	context->bw_ctx.bw.dcn.clk.bw_dispclk_khz = context->bw_ctx.bw.dcn.clk.dispclk_khz;
+	context->bw_ctx.bw.dcn.clk.max_supported_dppclk_khz = context->bw_ctx.dml.soc.clock_limits[vlevel].dppclk_mhz
+			* 1000;
+	context->bw_ctx.bw.dcn.clk.max_supported_dispclk_khz = context->bw_ctx.dml.soc.clock_limits[vlevel].dispclk_mhz
+			* 1000;
+
+	context->bw_ctx.bw.dcn.clk.num_ways = dcn32_helper_calculate_num_ways_for_subvp(dc, context);
+
+	context->bw_ctx.bw.dcn.compbuf_size_kb = context->bw_ctx.dml.ip.config_return_buffer_size_in_kbytes;
+
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		if (context->res_ctx.pipe_ctx[i].stream)
+			context->bw_ctx.bw.dcn.compbuf_size_kb -= context->res_ctx.pipe_ctx[i].det_buffer_size_kb;
+	}
+
+	for (i = 0, pipe_idx = 0; i < dc->res_pool->pipe_count; i++) {
+
+		if (!context->res_ctx.pipe_ctx[i].stream)
+			continue;
+
+		context->bw_ctx.dml.funcs.rq_dlg_get_dlg_reg_v2(&context->bw_ctx.dml,
+				&context->res_ctx.pipe_ctx[i].dlg_regs, &context->res_ctx.pipe_ctx[i].ttu_regs, pipes,
+				pipe_cnt, pipe_idx);
+
+		context->bw_ctx.dml.funcs.rq_dlg_get_rq_reg_v2(&context->res_ctx.pipe_ctx[i].rq_regs,
+				&context->bw_ctx.dml, pipes, pipe_cnt, pipe_idx);
+		pipe_idx++;
+	}
+}
+
+static struct pipe_ctx *dcn32_find_split_pipe(
+		struct dc *dc,
+		struct dc_state *context,
+		int old_index)
+{
+	struct pipe_ctx *pipe = NULL;
+	int i;
+
+	if (old_index >= 0 && context->res_ctx.pipe_ctx[old_index].stream == NULL) {
+		pipe = &context->res_ctx.pipe_ctx[old_index];
+		pipe->pipe_idx = old_index;
+	}
+
+	if (!pipe)
+		for (i = dc->res_pool->pipe_count - 1; i >= 0; i--) {
+			if (dc->current_state->res_ctx.pipe_ctx[i].top_pipe == NULL
+					&& dc->current_state->res_ctx.pipe_ctx[i].prev_odm_pipe == NULL) {
+				if (context->res_ctx.pipe_ctx[i].stream == NULL) {
+					pipe = &context->res_ctx.pipe_ctx[i];
+					pipe->pipe_idx = i;
+					break;
+				}
+			}
+		}
+
+	/*
+	 * May need to fix pipes getting tossed from 1 opp to another on flip
+	 * Add for debugging transient underflow during topology updates:
+	 * ASSERT(pipe);
+	 */
+	if (!pipe)
+		for (i = dc->res_pool->pipe_count - 1; i >= 0; i--) {
+			if (context->res_ctx.pipe_ctx[i].stream == NULL) {
+				pipe = &context->res_ctx.pipe_ctx[i];
+				pipe->pipe_idx = i;
+				break;
+			}
+		}
+
+	return pipe;
+}
+
+static bool dcn32_split_stream_for_mpc_or_odm(
+		const struct dc *dc,
+		struct resource_context *res_ctx,
+		struct pipe_ctx *pri_pipe,
+		struct pipe_ctx *sec_pipe,
+		bool odm)
+{
+	int pipe_idx = sec_pipe->pipe_idx;
+	const struct resource_pool *pool = dc->res_pool;
+
+	DC_LOGGER_INIT(dc->ctx->logger);
+
+	if (odm && pri_pipe->plane_state) {
+		/* ODM + window MPO, where MPO window is on left half only */
+		if (pri_pipe->plane_state->clip_rect.x + pri_pipe->plane_state->clip_rect.width <=
+				pri_pipe->stream->src.x + pri_pipe->stream->src.width/2) {
+
+			DC_LOG_SCALER("%s - ODM + window MPO(left). pri_pipe:%d\n",
+					__func__,
+					pri_pipe->pipe_idx);
+			return true;
+		}
+
+		/* ODM + window MPO, where MPO window is on right half only */
+		if (pri_pipe->plane_state->clip_rect.x >= pri_pipe->stream->src.x +  pri_pipe->stream->src.width/2) {
+
+			DC_LOG_SCALER("%s - ODM + window MPO(right). pri_pipe:%d\n",
+					__func__,
+					pri_pipe->pipe_idx);
+			return true;
+		}
+	}
+
+	*sec_pipe = *pri_pipe;
+
+	sec_pipe->pipe_idx = pipe_idx;
+	sec_pipe->plane_res.mi = pool->mis[pipe_idx];
+	sec_pipe->plane_res.hubp = pool->hubps[pipe_idx];
+	sec_pipe->plane_res.ipp = pool->ipps[pipe_idx];
+	sec_pipe->plane_res.xfm = pool->transforms[pipe_idx];
+	sec_pipe->plane_res.dpp = pool->dpps[pipe_idx];
+	sec_pipe->plane_res.mpcc_inst = pool->dpps[pipe_idx]->inst;
+	sec_pipe->stream_res.dsc = NULL;
+	if (odm) {
+		if (pri_pipe->next_odm_pipe) {
+			ASSERT(pri_pipe->next_odm_pipe != sec_pipe);
+			sec_pipe->next_odm_pipe = pri_pipe->next_odm_pipe;
+			sec_pipe->next_odm_pipe->prev_odm_pipe = sec_pipe;
+		}
+		if (pri_pipe->top_pipe && pri_pipe->top_pipe->next_odm_pipe) {
+			pri_pipe->top_pipe->next_odm_pipe->bottom_pipe = sec_pipe;
+			sec_pipe->top_pipe = pri_pipe->top_pipe->next_odm_pipe;
+		}
+		if (pri_pipe->bottom_pipe && pri_pipe->bottom_pipe->next_odm_pipe) {
+			pri_pipe->bottom_pipe->next_odm_pipe->top_pipe = sec_pipe;
+			sec_pipe->bottom_pipe = pri_pipe->bottom_pipe->next_odm_pipe;
+		}
+		pri_pipe->next_odm_pipe = sec_pipe;
+		sec_pipe->prev_odm_pipe = pri_pipe;
+		ASSERT(sec_pipe->top_pipe == NULL);
+
+		if (!sec_pipe->top_pipe)
+			sec_pipe->stream_res.opp = pool->opps[pipe_idx];
+		else
+			sec_pipe->stream_res.opp = sec_pipe->top_pipe->stream_res.opp;
+		if (sec_pipe->stream->timing.flags.DSC == 1) {
+			dcn20_acquire_dsc(dc, res_ctx, &sec_pipe->stream_res.dsc, pipe_idx);
+			ASSERT(sec_pipe->stream_res.dsc);
+			if (sec_pipe->stream_res.dsc == NULL)
+				return false;
+		}
+	} else {
+		if (pri_pipe->bottom_pipe) {
+			ASSERT(pri_pipe->bottom_pipe != sec_pipe);
+			sec_pipe->bottom_pipe = pri_pipe->bottom_pipe;
+			sec_pipe->bottom_pipe->top_pipe = sec_pipe;
+		}
+		pri_pipe->bottom_pipe = sec_pipe;
+		sec_pipe->top_pipe = pri_pipe;
+
+		ASSERT(pri_pipe->plane_state);
+	}
+
+	return true;
+}
+
+static bool dcn32_apply_merge_split_flags_helper(
+		struct dc *dc,
+		struct dc_state *context,
+		bool *repopulate_pipes,
+		int *split,
+		bool *merge)
+{
+	int i, pipe_idx;
+	bool newly_split[MAX_PIPES] = { false };
+	struct vba_vars_st *vba = &context->bw_ctx.dml.vba;
+
+	if (dc->config.enable_windowed_mpo_odm) {
+		if (update_pipes_with_split_flags(
+			dc, context, vba, split, merge))
+			*repopulate_pipes = true;
+	} else {
+
+		/* the code below will be removed once windowed mpo odm is fully
+		 * enabled.
+		 */
+		/* merge pipes if necessary */
+		for (i = 0; i < dc->res_pool->pipe_count; i++) {
+			struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
+
+			/*skip pipes that don't need merging*/
+			if (!merge[i])
+				continue;
+
+			/* if ODM merge we ignore mpc tree, mpo pipes will have their own flags */
+			if (pipe->prev_odm_pipe) {
+				/*split off odm pipe*/
+				pipe->prev_odm_pipe->next_odm_pipe = pipe->next_odm_pipe;
+				if (pipe->next_odm_pipe)
+					pipe->next_odm_pipe->prev_odm_pipe = pipe->prev_odm_pipe;
+
+				/*2:1ODM+MPC Split MPO to Single Pipe + MPC Split MPO*/
+				if (pipe->bottom_pipe) {
+					if (pipe->bottom_pipe->prev_odm_pipe || pipe->bottom_pipe->next_odm_pipe) {
+						/*MPC split rules will handle this case*/
+						pipe->bottom_pipe->top_pipe = NULL;
+					} else {
+						/* when merging an ODM pipes, the bottom MPC pipe must now point to
+						 * the previous ODM pipe and its associated stream assets
+						 */
+						if (pipe->prev_odm_pipe->bottom_pipe) {
+							/* 3 plane MPO*/
+							pipe->bottom_pipe->top_pipe = pipe->prev_odm_pipe->bottom_pipe;
+							pipe->prev_odm_pipe->bottom_pipe->bottom_pipe = pipe->bottom_pipe;
+						} else {
+							/* 2 plane MPO*/
+							pipe->bottom_pipe->top_pipe = pipe->prev_odm_pipe;
+							pipe->prev_odm_pipe->bottom_pipe = pipe->bottom_pipe;
+						}
+
+						memcpy(&pipe->bottom_pipe->stream_res, &pipe->bottom_pipe->top_pipe->stream_res, sizeof(struct stream_resource));
+					}
+				}
+
+				if (pipe->top_pipe) {
+					pipe->top_pipe->bottom_pipe = NULL;
+				}
+
+				pipe->bottom_pipe = NULL;
+				pipe->next_odm_pipe = NULL;
+				pipe->plane_state = NULL;
+				pipe->stream = NULL;
+				pipe->top_pipe = NULL;
+				pipe->prev_odm_pipe = NULL;
+				if (pipe->stream_res.dsc)
+					dcn20_release_dsc(&context->res_ctx, dc->res_pool, &pipe->stream_res.dsc);
+				memset(&pipe->plane_res, 0, sizeof(pipe->plane_res));
+				memset(&pipe->stream_res, 0, sizeof(pipe->stream_res));
+				memset(&pipe->link_res, 0, sizeof(pipe->link_res));
+				*repopulate_pipes = true;
+			} else if (pipe->top_pipe && pipe->top_pipe->plane_state == pipe->plane_state) {
+				struct pipe_ctx *top_pipe = pipe->top_pipe;
+				struct pipe_ctx *bottom_pipe = pipe->bottom_pipe;
+
+				top_pipe->bottom_pipe = bottom_pipe;
+				if (bottom_pipe)
+					bottom_pipe->top_pipe = top_pipe;
+
+				pipe->top_pipe = NULL;
+				pipe->bottom_pipe = NULL;
+				pipe->plane_state = NULL;
+				pipe->stream = NULL;
+				memset(&pipe->plane_res, 0, sizeof(pipe->plane_res));
+				memset(&pipe->stream_res, 0, sizeof(pipe->stream_res));
+				memset(&pipe->link_res, 0, sizeof(pipe->link_res));
+				*repopulate_pipes = true;
+			} else
+				ASSERT(0); /* Should never try to merge master pipe */
+
+		}
+
+		for (i = 0, pipe_idx = -1; i < dc->res_pool->pipe_count; i++) {
+			struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
+			struct pipe_ctx *old_pipe = &dc->current_state->res_ctx.pipe_ctx[i];
+			struct pipe_ctx *hsplit_pipe = NULL;
+			bool odm;
+			int old_index = -1;
+
+			if (!pipe->stream || newly_split[i])
+				continue;
+
+			pipe_idx++;
+			odm = vba->ODMCombineEnabled[vba->pipe_plane[pipe_idx]] != dm_odm_combine_mode_disabled;
+
+			if (!pipe->plane_state && !odm)
+				continue;
+
+			if (split[i]) {
+				if (odm) {
+					if (split[i] == 4 && old_pipe->next_odm_pipe && old_pipe->next_odm_pipe->next_odm_pipe)
+						old_index = old_pipe->next_odm_pipe->next_odm_pipe->pipe_idx;
+					else if (old_pipe->next_odm_pipe)
+						old_index = old_pipe->next_odm_pipe->pipe_idx;
+				} else {
+					if (split[i] == 4 && old_pipe->bottom_pipe && old_pipe->bottom_pipe->bottom_pipe &&
+							old_pipe->bottom_pipe->bottom_pipe->plane_state == old_pipe->plane_state)
+						old_index = old_pipe->bottom_pipe->bottom_pipe->pipe_idx;
+					else if (old_pipe->bottom_pipe &&
+							old_pipe->bottom_pipe->plane_state == old_pipe->plane_state)
+						old_index = old_pipe->bottom_pipe->pipe_idx;
+				}
+				hsplit_pipe = dcn32_find_split_pipe(dc, context, old_index);
+				ASSERT(hsplit_pipe);
+				if (!hsplit_pipe)
+					return false;
+
+				if (!dcn32_split_stream_for_mpc_or_odm(
+						dc, &context->res_ctx,
+						pipe, hsplit_pipe, odm))
+					return false;
+
+				newly_split[hsplit_pipe->pipe_idx] = true;
+				*repopulate_pipes = true;
+			}
+			if (split[i] == 4) {
+				struct pipe_ctx *pipe_4to1;
+
+				if (odm && old_pipe->next_odm_pipe)
+					old_index = old_pipe->next_odm_pipe->pipe_idx;
+				else if (!odm && old_pipe->bottom_pipe &&
+							old_pipe->bottom_pipe->plane_state == old_pipe->plane_state)
+					old_index = old_pipe->bottom_pipe->pipe_idx;
+				else
+					old_index = -1;
+				pipe_4to1 = dcn32_find_split_pipe(dc, context, old_index);
+				ASSERT(pipe_4to1);
+				if (!pipe_4to1)
+					return false;
+				if (!dcn32_split_stream_for_mpc_or_odm(
+						dc, &context->res_ctx,
+						pipe, pipe_4to1, odm))
+					return false;
+				newly_split[pipe_4to1->pipe_idx] = true;
+
+				if (odm && old_pipe->next_odm_pipe && old_pipe->next_odm_pipe->next_odm_pipe
+						&& old_pipe->next_odm_pipe->next_odm_pipe->next_odm_pipe)
+					old_index = old_pipe->next_odm_pipe->next_odm_pipe->next_odm_pipe->pipe_idx;
+				else if (!odm && old_pipe->bottom_pipe && old_pipe->bottom_pipe->bottom_pipe &&
+						old_pipe->bottom_pipe->bottom_pipe->bottom_pipe &&
+						old_pipe->bottom_pipe->bottom_pipe->bottom_pipe->plane_state == old_pipe->plane_state)
+					old_index = old_pipe->bottom_pipe->bottom_pipe->bottom_pipe->pipe_idx;
+				else
+					old_index = -1;
+				pipe_4to1 = dcn32_find_split_pipe(dc, context, old_index);
+				ASSERT(pipe_4to1);
+				if (!pipe_4to1)
+					return false;
+				if (!dcn32_split_stream_for_mpc_or_odm(
+						dc, &context->res_ctx,
+						hsplit_pipe, pipe_4to1, odm))
+					return false;
+				newly_split[pipe_4to1->pipe_idx] = true;
+			}
+			if (odm)
+				dcn20_build_mapped_resource(dc, context, pipe->stream);
+		}
+
+		for (i = 0; i < dc->res_pool->pipe_count; i++) {
+			struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
+
+			if (pipe->plane_state) {
+				if (!resource_build_scaling_params(pipe))
+					return false;
+			}
+		}
+
+		for (i = 0; i < context->stream_count; i++) {
+			struct pipe_ctx *otg_master = resource_get_otg_master_for_stream(&context->res_ctx,
+					context->streams[i]);
+
+			if (otg_master)
+				resource_build_test_pattern_params(&context->res_ctx, otg_master);
+		}
+	}
+	return true;
+}
+
+bool dcn32_internal_validate_bw(struct dc *dc,
+				struct dc_state *context,
+				display_e2e_pipe_params_st *pipes,
+				int *pipe_cnt_out,
+				int *vlevel_out,
+				enum dc_validate_mode validate_mode)
+{
+	bool out = false;
+	bool repopulate_pipes = false;
+	int split[MAX_PIPES] = { 0 };
+	bool merge[MAX_PIPES] = { false };
+	int pipe_cnt, i, pipe_idx;
+	int vlevel = context->bw_ctx.dml.soc.num_states;
+	struct vba_vars_st *vba = &context->bw_ctx.dml.vba;
+
+	dc_assert_fp_enabled();
+
+	ASSERT(pipes);
+	if (!pipes)
+		return false;
+
+	/* For each full update, remove all existing phantom pipes first */
+	dc_state_remove_phantom_streams_and_planes(dc, context);
+	dc_state_release_phantom_streams_and_planes(dc, context);
+
+	dc->res_pool->funcs->update_soc_for_wm_a(dc, context);
+
+	for (i = 0; i < context->stream_count; i++)
+		resource_update_pipes_for_stream_with_slice_count(context, dc->current_state, dc->res_pool, context->streams[i], 1);
+	pipe_cnt = dc->res_pool->funcs->populate_dml_pipes(dc, context, pipes, validate_mode);
+
+	if (!pipe_cnt) {
+		out = true;
+		goto validate_out;
+	}
+
+	dml_log_pipe_params(&context->bw_ctx.dml, pipes, pipe_cnt);
+	context->bw_ctx.dml.soc.max_vratio_pre = dcn32_determine_max_vratio_prefetch(dc, context);
+
+	if (validate_mode == DC_VALIDATE_MODE_AND_PROGRAMMING) {
+		if (!dcn32_full_validate_bw_helper(dc, context, pipes, &vlevel, split, merge,
+			&pipe_cnt, &repopulate_pipes))
+			goto validate_fail;
+	}
+
+	if (validate_mode != DC_VALIDATE_MODE_AND_PROGRAMMING ||
+			(dc->debug.dml_disallow_alternate_prefetch_modes &&
+			(vlevel == context->bw_ctx.dml.soc.num_states ||
+				vba->DRAMClockChangeSupport[vlevel][vba->maxMpcComb] == dm_dram_clock_change_unsupported))) {
+		/*
+		 * If dml_disallow_alternate_prefetch_modes is false, then we have already
+		 * tried alternate prefetch modes during full validation.
+		 *
+		 * If mode is unsupported or there is no p-state support, then
+		 * fall back to favouring voltage.
+		 *
+		 * If Prefetch mode 0 failed for this config, or passed with Max UCLK, then try
+		 * to support with Prefetch mode 1 (dm_prefetch_support_fclk_and_stutter == 2)
+		 */
+		context->bw_ctx.dml.soc.allow_for_pstate_or_stutter_in_vblank_final =
+			dm_prefetch_support_none;
+
+		context->bw_ctx.dml.validate_max_state = (validate_mode != DC_VALIDATE_MODE_AND_PROGRAMMING);
+		vlevel = dml_get_voltage_level(&context->bw_ctx.dml, pipes, pipe_cnt);
+
+		context->bw_ctx.dml.validate_max_state = false;
+
+		if (vlevel < context->bw_ctx.dml.soc.num_states) {
+			memset(split, 0, sizeof(split));
+			memset(merge, 0, sizeof(merge));
+			vlevel = dcn20_validate_apply_pipe_split_flags(dc, context, vlevel, split, merge);
+			/* dcn20_validate_apply_pipe_split_flags can modify voltage level outside of DML */
+			vba->VoltageLevel = vlevel;
+		}
+	}
+
+	dml_log_mode_support_params(&context->bw_ctx.dml);
+
+	if (vlevel == context->bw_ctx.dml.soc.num_states)
+		goto validate_fail;
+
+	for (i = 0, pipe_idx = 0; i < dc->res_pool->pipe_count; i++) {
+		struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
+		struct pipe_ctx *mpo_pipe = pipe->bottom_pipe;
+
+		if (!pipe->stream)
+			continue;
+
+		if (vba->ODMCombineEnabled[vba->pipe_plane[pipe_idx]] != dm_odm_combine_mode_disabled
+				&& !dc->config.enable_windowed_mpo_odm
+				&& pipe->plane_state && mpo_pipe
+				&& memcmp(&mpo_pipe->plane_state->clip_rect,
+						&pipe->stream->src,
+						sizeof(struct rect)) != 0) {
+			ASSERT(mpo_pipe->plane_state != pipe->plane_state);
+			goto validate_fail;
+		}
+		pipe_idx++;
+	}
+
+	if (!dcn32_apply_merge_split_flags_helper(dc, context, &repopulate_pipes, split, merge))
+		goto validate_fail;
+
+	/* Actual dsc count per stream dsc validation*/
+	if (!dcn20_validate_dsc(dc, context)) {
+		vba->ValidationStatus[vba->soc.num_states] = DML_FAIL_DSC_VALIDATION_FAILURE;
+		goto validate_fail;
+	}
+
+	if (repopulate_pipes) {
+		int flag_max_mpc_comb = vba->maxMpcComb;
+		int flag_vlevel = vlevel;
+		int i;
+
+		pipe_cnt = dc->res_pool->funcs->populate_dml_pipes(dc, context, pipes, validate_mode);
+		if (!dc->config.enable_windowed_mpo_odm)
+			dcn32_update_dml_pipes_odm_policy_based_on_context(dc, context, pipes);
+
+		/* repopulate_pipes = 1 means the pipes were either split or merged. In this case
+		 * we have to re-calculate the DET allocation and run through DML once more to
+		 * ensure all the params are calculated correctly. We do not need to run the
+		 * pipe split check again after this call (pipes are already split / merged).
+		 * */
+		context->bw_ctx.dml.soc.allow_for_pstate_or_stutter_in_vblank_final =
+					dm_prefetch_support_uclk_fclk_and_stutter_if_possible;
+
+		vlevel = dml_get_voltage_level(&context->bw_ctx.dml, pipes, pipe_cnt);
+
+		if (vlevel == context->bw_ctx.dml.soc.num_states) {
+			/* failed after DET size changes */
+			goto validate_fail;
+		} else if (flag_max_mpc_comb == 0 &&
+				flag_max_mpc_comb != context->bw_ctx.dml.vba.maxMpcComb) {
+			/* check the context constructed with pipe split flags is still valid*/
+			bool flags_valid = false;
+			for (i = flag_vlevel; i < context->bw_ctx.dml.soc.num_states; i++) {
+				if (vba->ModeSupport[i][flag_max_mpc_comb]) {
+					vba->maxMpcComb = flag_max_mpc_comb;
+					vba->VoltageLevel = i;
+					vlevel = i;
+					flags_valid = true;
+					break;
+				}
+			}
+
+			/* this should never happen */
+			if (!flags_valid)
+				goto validate_fail;
+		}
+	}
+	*vlevel_out = vlevel;
+	*pipe_cnt_out = pipe_cnt;
+
+	out = true;
+	goto validate_out;
+
+validate_fail:
+	out = false;
+
+validate_out:
+	return out;
+}
+
+
+void dcn32_calculate_wm_and_dlg_fpu(struct dc *dc, struct dc_state *context,
+				display_e2e_pipe_params_st *pipes,
+				int pipe_cnt,
+				int vlevel)
+{
+	int i, pipe_idx, vlevel_temp = 0;
+	double dcfclk = dcn3_2_soc.clock_limits[0].dcfclk_mhz;
+	double dcfclk_from_validation = context->bw_ctx.dml.vba.DCFCLKState[vlevel][context->bw_ctx.dml.vba.maxMpcComb];
+	double dram_speed_from_validation = context->bw_ctx.dml.vba.DRAMSpeed;
+	double dcfclk_from_fw_based_mclk_switching = dcfclk_from_validation;
+	bool pstate_en = context->bw_ctx.dml.vba.DRAMClockChangeSupport[vlevel][context->bw_ctx.dml.vba.maxMpcComb] !=
+			dm_dram_clock_change_unsupported;
+	unsigned int dummy_latency_index = 0;
+	int maxMpcComb = context->bw_ctx.dml.vba.maxMpcComb;
+	unsigned int min_dram_speed_mts = context->bw_ctx.dml.vba.DRAMSpeed;
+	bool subvp_in_use = dcn32_subvp_in_use(dc, context);
+	unsigned int min_dram_speed_mts_margin;
+	bool need_fclk_lat_as_dummy = false;
+	bool is_subvp_p_drr = false;
+	struct dc_stream_state *fpo_candidate_stream = NULL;
+	struct dc_stream_status *stream_status = NULL;
+
+	dc_assert_fp_enabled();
+
+	/* need to find dummy latency index for subvp */
+	if (subvp_in_use) {
+		/* Override DRAMClockChangeSupport for SubVP + DRR case where the DRR cannot switch without stretching it's VBLANK */
+		if (!pstate_en) {
+			context->bw_ctx.dml.vba.DRAMClockChangeSupport[vlevel][maxMpcComb] = dm_dram_clock_change_vblank_w_mall_sub_vp;
+			context->bw_ctx.dml.soc.allow_for_pstate_or_stutter_in_vblank_final = dm_prefetch_support_fclk_and_stutter;
+			pstate_en = true;
+			is_subvp_p_drr = true;
+		}
+		dummy_latency_index = dcn32_find_dummy_latency_index_for_fw_based_mclk_switch(dc,
+						context, pipes, pipe_cnt, vlevel);
+
+		/* For DCN32/321 need to validate with fclk pstate change latency equal to dummy so prefetch is
+		 * scheduled correctly to account for dummy pstate.
+		 */
+		if (context->bw_ctx.dml.soc.fclk_change_latency_us < dc->clk_mgr->bw_params->dummy_pstate_table[dummy_latency_index].dummy_pstate_latency_us) {
+			need_fclk_lat_as_dummy = true;
+			context->bw_ctx.dml.soc.fclk_change_latency_us =
+					dc->clk_mgr->bw_params->dummy_pstate_table[dummy_latency_index].dummy_pstate_latency_us;
+		}
+		context->bw_ctx.dml.soc.dram_clock_change_latency_us =
+							dc->clk_mgr->bw_params->wm_table.nv_entries[WM_A].dml_input.pstate_latency_us;
+		dcn32_internal_validate_bw(dc, context, pipes, &pipe_cnt, &vlevel, DC_VALIDATE_MODE_AND_PROGRAMMING);
+		maxMpcComb = context->bw_ctx.dml.vba.maxMpcComb;
+		if (is_subvp_p_drr) {
+			context->bw_ctx.dml.vba.DRAMClockChangeSupport[vlevel][maxMpcComb] = dm_dram_clock_change_vblank_w_mall_sub_vp;
+		}
+	}
+
+	context->bw_ctx.bw.dcn.clk.fw_based_mclk_switching = false;
+	for (i = 0; i < context->stream_count; i++) {
+		stream_status = NULL;
+		if (context->streams[i])
+			stream_status = dc_state_get_stream_status(context, context->streams[i]);
+		if (stream_status)
+			stream_status->fpo_in_use = false;
+	}
+
+	if (!pstate_en || (!dc->debug.disable_fpo_optimizations &&
+			pstate_en && vlevel != 0)) {
+		/* only when the mclk switch can not be natural, is the fw based vblank stretch attempted */
+		fpo_candidate_stream = dcn32_can_support_mclk_switch_using_fw_based_vblank_stretch(dc, context);
+		if (fpo_candidate_stream) {
+			stream_status = dc_state_get_stream_status(context, fpo_candidate_stream);
+			if (stream_status)
+				stream_status->fpo_in_use = true;
+			context->bw_ctx.bw.dcn.clk.fw_based_mclk_switching = true;
+		}
+
+		if (context->bw_ctx.bw.dcn.clk.fw_based_mclk_switching) {
+			dummy_latency_index = dcn32_find_dummy_latency_index_for_fw_based_mclk_switch(dc,
+				context, pipes, pipe_cnt, vlevel);
+
+			/* After calling dcn30_find_dummy_latency_index_for_fw_based_mclk_switch
+			 * we reinstate the original dram_clock_change_latency_us on the context
+			 * and all variables that may have changed up to this point, except the
+			 * newly found dummy_latency_index
+			 */
+			context->bw_ctx.dml.soc.dram_clock_change_latency_us =
+					dc->clk_mgr->bw_params->wm_table.nv_entries[WM_A].dml_input.pstate_latency_us;
+			/* For DCN32/321 need to validate with fclk pstate change latency equal to dummy so
+			 * prefetch is scheduled correctly to account for dummy pstate.
+			 */
+			if (context->bw_ctx.dml.soc.fclk_change_latency_us < dc->clk_mgr->bw_params->dummy_pstate_table[dummy_latency_index].dummy_pstate_latency_us) {
+				need_fclk_lat_as_dummy = true;
+				context->bw_ctx.dml.soc.fclk_change_latency_us =
+						dc->clk_mgr->bw_params->dummy_pstate_table[dummy_latency_index].dummy_pstate_latency_us;
+			}
+			dcn32_internal_validate_bw(dc, context, pipes, &pipe_cnt, &vlevel_temp,
+				DC_VALIDATE_MODE_AND_PROGRAMMING);
+			if (vlevel_temp < vlevel) {
+				vlevel = vlevel_temp;
+				maxMpcComb = context->bw_ctx.dml.vba.maxMpcComb;
+				dcfclk_from_fw_based_mclk_switching = context->bw_ctx.dml.vba.DCFCLKState[vlevel][context->bw_ctx.dml.vba.maxMpcComb];
+				pstate_en = true;
+				context->bw_ctx.dml.vba.DRAMClockChangeSupport[vlevel][maxMpcComb] = dm_dram_clock_change_vblank;
+			} else {
+				/* Restore FCLK latency and re-run validation to go back to original validation
+				 * output if we find that enabling FPO does not give us any benefit (i.e. lower
+				 * voltage level)
+				 */
+				context->bw_ctx.bw.dcn.clk.fw_based_mclk_switching = false;
+				for (i = 0; i < context->stream_count; i++) {
+					stream_status = NULL;
+					if (context->streams[i])
+						stream_status = dc_state_get_stream_status(context, context->streams[i]);
+					if (stream_status)
+						stream_status->fpo_in_use = false;
+				}
+				context->bw_ctx.dml.soc.fclk_change_latency_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_A].dml_input.fclk_change_latency_us;
+				dcn32_internal_validate_bw(dc, context, pipes, &pipe_cnt, &vlevel,
+					DC_VALIDATE_MODE_AND_PROGRAMMING);
+			}
+		}
+	}
+
+	/* Set B:
+	 * For Set B calculations use clocks from clock_limits[2] when available i.e. when SMU is present,
+	 * otherwise use arbitrary low value from spreadsheet for DCFCLK as lower is safer for watermark
+	 * calculations to cover bootup clocks.
+	 * DCFCLK: soc.clock_limits[2] when available
+	 * UCLK: soc.clock_limits[2] when available
+	 */
+	if (dcn3_2_soc.num_states > 2) {
+		vlevel_temp = 2;
+		dcfclk = dcn3_2_soc.clock_limits[2].dcfclk_mhz;
+	} else
+		dcfclk = 615; //DCFCLK Vmin_lv
+
+	pipes[0].clks_cfg.voltage = vlevel_temp;
+	pipes[0].clks_cfg.dcfclk_mhz = dcfclk;
+	pipes[0].clks_cfg.socclk_mhz = context->bw_ctx.dml.soc.clock_limits[vlevel_temp].socclk_mhz;
+
+	if (dc->clk_mgr->bw_params->wm_table.nv_entries[WM_B].valid) {
+		context->bw_ctx.dml.soc.dram_clock_change_latency_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_B].dml_input.pstate_latency_us;
+		context->bw_ctx.dml.soc.fclk_change_latency_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_B].dml_input.fclk_change_latency_us;
+		context->bw_ctx.dml.soc.sr_enter_plus_exit_time_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_B].dml_input.sr_enter_plus_exit_time_us;
+		context->bw_ctx.dml.soc.sr_exit_time_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_B].dml_input.sr_exit_time_us;
+	}
+	context->bw_ctx.bw.dcn.watermarks.b.urgent_ns = get_wm_urgent(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.b.cstate_pstate.cstate_enter_plus_exit_ns = get_wm_stutter_enter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.b.cstate_pstate.cstate_exit_ns = get_wm_stutter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.b.cstate_pstate.pstate_change_ns = get_wm_dram_clock_change(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.b.pte_meta_urgent_ns = get_wm_memory_trip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.b.frac_urg_bw_nom = get_fraction_of_urgent_bandwidth(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.b.frac_urg_bw_flip = get_fraction_of_urgent_bandwidth_imm_flip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.b.urgent_latency_ns = get_urgent_latency(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.b.cstate_pstate.fclk_pstate_change_ns = get_fclk_watermark(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.b.usr_retraining_ns = get_usr_retraining_watermark(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+
+	/* Set D:
+	 * All clocks min.
+	 * DCFCLK: Min, as reported by PM FW when available
+	 * UCLK  : Min, as reported by PM FW when available
+	 * sr_enter_exit/sr_exit should be lower than used for DRAM (TBD after bringup or later, use as decided in Clk Mgr)
+	 */
+
+	/*
+	if (dcn3_2_soc.num_states > 2) {
+		vlevel_temp = 0;
+		dcfclk = dc->clk_mgr->bw_params->clk_table.entries[0].dcfclk_mhz;
+	} else
+		dcfclk = 615; //DCFCLK Vmin_lv
+
+	pipes[0].clks_cfg.voltage = vlevel_temp;
+	pipes[0].clks_cfg.dcfclk_mhz = dcfclk;
+	pipes[0].clks_cfg.socclk_mhz = context->bw_ctx.dml.soc.clock_limits[vlevel_temp].socclk_mhz;
+
+	if (dc->clk_mgr->bw_params->wm_table.nv_entries[WM_D].valid) {
+		context->bw_ctx.dml.soc.dram_clock_change_latency_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_D].dml_input.pstate_latency_us;
+		context->bw_ctx.dml.soc.fclk_change_latency_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_D].dml_input.fclk_change_latency_us;
+		context->bw_ctx.dml.soc.sr_enter_plus_exit_time_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_D].dml_input.sr_enter_plus_exit_time_us;
+		context->bw_ctx.dml.soc.sr_exit_time_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_D].dml_input.sr_exit_time_us;
+	}
+	context->bw_ctx.bw.dcn.watermarks.d.urgent_ns = get_wm_urgent(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.d.cstate_pstate.cstate_enter_plus_exit_ns = get_wm_stutter_enter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.d.cstate_pstate.cstate_exit_ns = get_wm_stutter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.d.cstate_pstate.pstate_change_ns = get_wm_dram_clock_change(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.d.pte_meta_urgent_ns = get_wm_memory_trip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.d.frac_urg_bw_nom = get_fraction_of_urgent_bandwidth(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.d.frac_urg_bw_flip = get_fraction_of_urgent_bandwidth_imm_flip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.d.urgent_latency_ns = get_urgent_latency(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.d.cstate_pstate.fclk_pstate_change_ns = get_fclk_watermark(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.d.usr_retraining_ns = get_usr_retraining_watermark(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	*/
+
+	/* Set C, for Dummy P-State:
+	 * All clocks min.
+	 * DCFCLK: Min, as reported by PM FW, when available
+	 * UCLK  : Min,  as reported by PM FW, when available
+	 * pstate latency as per UCLK state dummy pstate latency
+	 */
+
+	// For Set A and Set C use values from validation
+	pipes[0].clks_cfg.voltage = vlevel;
+	pipes[0].clks_cfg.dcfclk_mhz = dcfclk_from_validation;
+	pipes[0].clks_cfg.socclk_mhz = context->bw_ctx.dml.soc.clock_limits[vlevel].socclk_mhz;
+
+	if (context->bw_ctx.bw.dcn.clk.fw_based_mclk_switching) {
+		pipes[0].clks_cfg.dcfclk_mhz = dcfclk_from_fw_based_mclk_switching;
+	}
+
+	if (dc->clk_mgr->bw_params->wm_table.nv_entries[WM_C].valid) {
+		min_dram_speed_mts = dram_speed_from_validation;
+		min_dram_speed_mts_margin = 160;
+
+		context->bw_ctx.dml.soc.dram_clock_change_latency_us =
+			dc->clk_mgr->bw_params->dummy_pstate_table[0].dummy_pstate_latency_us;
+
+		if (context->bw_ctx.dml.vba.DRAMClockChangeSupport[vlevel][maxMpcComb] ==
+			dm_dram_clock_change_unsupported) {
+			int min_dram_speed_mts_offset = dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_memclk_levels - 1;
+
+			min_dram_speed_mts =
+				dc->clk_mgr->bw_params->clk_table.entries[min_dram_speed_mts_offset].memclk_mhz * 16;
+		}
+
+		if (!context->bw_ctx.bw.dcn.clk.fw_based_mclk_switching && !subvp_in_use) {
+			/* find largest table entry that is lower than dram speed,
+			 * but lower than DPM0 still uses DPM0
+			 */
+			for (dummy_latency_index = 3; dummy_latency_index > 0; dummy_latency_index--)
+				if (min_dram_speed_mts + min_dram_speed_mts_margin >
+					dc->clk_mgr->bw_params->dummy_pstate_table[dummy_latency_index].dram_speed_mts)
+					break;
+		}
+
+		context->bw_ctx.dml.soc.dram_clock_change_latency_us =
+			dc->clk_mgr->bw_params->dummy_pstate_table[dummy_latency_index].dummy_pstate_latency_us;
+
+		context->bw_ctx.dml.soc.fclk_change_latency_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_C].dml_input.fclk_change_latency_us;
+		context->bw_ctx.dml.soc.sr_enter_plus_exit_time_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_C].dml_input.sr_enter_plus_exit_time_us;
+		context->bw_ctx.dml.soc.sr_exit_time_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_C].dml_input.sr_exit_time_us;
+	}
+
+	context->bw_ctx.bw.dcn.watermarks.c.urgent_ns = get_wm_urgent(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.c.cstate_pstate.cstate_enter_plus_exit_ns = get_wm_stutter_enter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.c.cstate_pstate.cstate_exit_ns = get_wm_stutter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.c.cstate_pstate.pstate_change_ns = get_wm_dram_clock_change(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.c.pte_meta_urgent_ns = get_wm_memory_trip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.c.frac_urg_bw_nom = get_fraction_of_urgent_bandwidth(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.c.frac_urg_bw_flip = get_fraction_of_urgent_bandwidth_imm_flip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.c.urgent_latency_ns = get_urgent_latency(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	/* On DCN32/321, PMFW will set PSTATE_CHANGE_TYPE = 1 (FCLK) for UCLK dummy p-state.
+	 * In this case we must program FCLK WM Set C to use the UCLK dummy p-state WM
+	 * value.
+	 */
+	context->bw_ctx.bw.dcn.watermarks.c.cstate_pstate.fclk_pstate_change_ns = get_wm_dram_clock_change(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	context->bw_ctx.bw.dcn.watermarks.c.usr_retraining_ns = get_usr_retraining_watermark(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+
+	if ((!pstate_en) && (dc->clk_mgr->bw_params->wm_table.nv_entries[WM_C].valid)) {
+		/* The only difference between A and C is p-state latency, if p-state is not supported
+		 * with full p-state latency we want to calculate DLG based on dummy p-state latency,
+		 * Set A p-state watermark set to 0 on DCN30, when p-state unsupported, for now keep as DCN30.
+		 */
+		context->bw_ctx.bw.dcn.watermarks.a = context->bw_ctx.bw.dcn.watermarks.c;
+		context->bw_ctx.bw.dcn.watermarks.a.cstate_pstate.pstate_change_ns = 0;
+		/* Calculate FCLK p-state change watermark based on FCLK pstate change latency in case
+		 * UCLK p-state is not supported, to avoid underflow in case FCLK pstate is supported
+		 */
+		context->bw_ctx.bw.dcn.watermarks.a.cstate_pstate.fclk_pstate_change_ns = get_fclk_watermark(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	} else {
+		/* Set A:
+		 * All clocks min.
+		 * DCFCLK: Min, as reported by PM FW, when available
+		 * UCLK: Min, as reported by PM FW, when available
+		 */
+
+		/* For set A set the correct latency values (i.e. non-dummy values) unconditionally
+		 */
+		context->bw_ctx.dml.soc.dram_clock_change_latency_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_A].dml_input.pstate_latency_us;
+		context->bw_ctx.dml.soc.sr_enter_plus_exit_time_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_A].dml_input.sr_enter_plus_exit_time_us;
+		context->bw_ctx.dml.soc.sr_exit_time_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_A].dml_input.sr_exit_time_us;
+
+		context->bw_ctx.bw.dcn.watermarks.a.urgent_ns = get_wm_urgent(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+		context->bw_ctx.bw.dcn.watermarks.a.cstate_pstate.cstate_enter_plus_exit_ns = get_wm_stutter_enter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+		context->bw_ctx.bw.dcn.watermarks.a.cstate_pstate.cstate_exit_ns = get_wm_stutter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+		context->bw_ctx.bw.dcn.watermarks.a.cstate_pstate.pstate_change_ns = get_wm_dram_clock_change(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+		context->bw_ctx.bw.dcn.watermarks.a.pte_meta_urgent_ns = get_wm_memory_trip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+		context->bw_ctx.bw.dcn.watermarks.a.frac_urg_bw_nom = get_fraction_of_urgent_bandwidth(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+		context->bw_ctx.bw.dcn.watermarks.a.frac_urg_bw_flip = get_fraction_of_urgent_bandwidth_imm_flip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+		context->bw_ctx.bw.dcn.watermarks.a.urgent_latency_ns = get_urgent_latency(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+		context->bw_ctx.bw.dcn.watermarks.a.cstate_pstate.fclk_pstate_change_ns = get_fclk_watermark(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+		context->bw_ctx.bw.dcn.watermarks.a.usr_retraining_ns = get_usr_retraining_watermark(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
+	}
+
+	/* Make set D = set A since we do not optimized watermarks for MALL */
+	context->bw_ctx.bw.dcn.watermarks.d = context->bw_ctx.bw.dcn.watermarks.a;
+
+	for (i = 0, pipe_idx = 0; i < dc->res_pool->pipe_count; i++) {
+		if (!context->res_ctx.pipe_ctx[i].stream)
+			continue;
+
+		pipes[pipe_idx].clks_cfg.dispclk_mhz = get_dispclk_calculated(&context->bw_ctx.dml, pipes, pipe_cnt);
+		pipes[pipe_idx].clks_cfg.dppclk_mhz = get_dppclk_calculated(&context->bw_ctx.dml, pipes, pipe_cnt, pipe_idx);
+
+		if (dc->config.forced_clocks) {
+			pipes[pipe_idx].clks_cfg.dispclk_mhz = context->bw_ctx.dml.soc.clock_limits[0].dispclk_mhz;
+			pipes[pipe_idx].clks_cfg.dppclk_mhz = context->bw_ctx.dml.soc.clock_limits[0].dppclk_mhz;
+		}
+		if (dc->debug.min_disp_clk_khz > pipes[pipe_idx].clks_cfg.dispclk_mhz * 1000)
+			pipes[pipe_idx].clks_cfg.dispclk_mhz = dc->debug.min_disp_clk_khz / 1000.0;
+		if (dc->debug.min_dpp_clk_khz > pipes[pipe_idx].clks_cfg.dppclk_mhz * 1000)
+			pipes[pipe_idx].clks_cfg.dppclk_mhz = dc->debug.min_dpp_clk_khz / 1000.0;
+
+		pipe_idx++;
+	}
+
+	context->perf_params.stutter_period_us = context->bw_ctx.dml.vba.StutterPeriod;
+
+	/* for proper prefetch calculations, if dummy lat > fclk lat, use fclk lat = dummy lat */
+	if (need_fclk_lat_as_dummy)
+		context->bw_ctx.dml.soc.fclk_change_latency_us =
+				dc->clk_mgr->bw_params->dummy_pstate_table[dummy_latency_index].dummy_pstate_latency_us;
+
+	dcn32_calculate_dlg_params(dc, context, pipes, pipe_cnt, vlevel);
+
+	if (!pstate_en)
+		/* Restore full p-state latency */
+		context->bw_ctx.dml.soc.dram_clock_change_latency_us =
+				dc->clk_mgr->bw_params->wm_table.nv_entries[WM_A].dml_input.pstate_latency_us;
+
+	/* revert fclk lat changes if required */
+	if (need_fclk_lat_as_dummy)
+		context->bw_ctx.dml.soc.fclk_change_latency_us =
+				dc->clk_mgr->bw_params->wm_table.nv_entries[WM_A].dml_input.fclk_change_latency_us;
+}
+
+static void dcn32_get_optimal_dcfclk_fclk_for_uclk(unsigned int uclk_mts,
+		unsigned int *optimal_dcfclk,
+		unsigned int *optimal_fclk)
+{
+	double bw_from_dram, bw_from_dram1, bw_from_dram2;
+
+	bw_from_dram1 = uclk_mts * dcn3_2_soc.num_chans *
+		dcn3_2_soc.dram_channel_width_bytes * (dcn3_2_soc.max_avg_dram_bw_use_normal_percent / 100);
+	bw_from_dram2 = uclk_mts * dcn3_2_soc.num_chans *
+		dcn3_2_soc.dram_channel_width_bytes * (dcn3_2_soc.max_avg_sdp_bw_use_normal_percent / 100);
+
+	bw_from_dram = (bw_from_dram1 < bw_from_dram2) ? bw_from_dram1 : bw_from_dram2;
+
+	if (optimal_fclk)
+		*optimal_fclk = bw_from_dram /
+		(dcn3_2_soc.fabric_datapath_to_dcn_data_return_bytes * (dcn3_2_soc.max_avg_sdp_bw_use_normal_percent / 100));
+
+	if (optimal_dcfclk)
+		*optimal_dcfclk =  bw_from_dram /
+		(dcn3_2_soc.return_bus_width_bytes * (dcn3_2_soc.max_avg_sdp_bw_use_normal_percent / 100));
+}
+
+static void remove_entry_from_table_at_index(struct _vcs_dpi_voltage_scaling_st *table, unsigned int *num_entries,
+		unsigned int index)
+{
+	int i;
+
+	if (*num_entries == 0)
+		return;
+
+	for (i = index; i < *num_entries - 1; i++) {
+		table[i] = table[i + 1];
+	}
+	memset(&table[--(*num_entries)], 0, sizeof(struct _vcs_dpi_voltage_scaling_st));
+}
+
+void dcn32_patch_dpm_table(struct clk_bw_params *bw_params)
+{
+	int i;
+	unsigned int max_dcfclk_mhz = 0, max_dispclk_mhz = 0, max_dppclk_mhz = 0,
+			max_phyclk_mhz = 0, max_dtbclk_mhz = 0, max_fclk_mhz = 0, max_uclk_mhz = 0;
+
+	for (i = 0; i < MAX_NUM_DPM_LVL; i++) {
+		if (bw_params->clk_table.entries[i].dcfclk_mhz > max_dcfclk_mhz)
+			max_dcfclk_mhz = bw_params->clk_table.entries[i].dcfclk_mhz;
+		if (bw_params->clk_table.entries[i].fclk_mhz > max_fclk_mhz)
+			max_fclk_mhz = bw_params->clk_table.entries[i].fclk_mhz;
+		if (bw_params->clk_table.entries[i].memclk_mhz > max_uclk_mhz)
+			max_uclk_mhz = bw_params->clk_table.entries[i].memclk_mhz;
+		if (bw_params->clk_table.entries[i].dispclk_mhz > max_dispclk_mhz)
+			max_dispclk_mhz = bw_params->clk_table.entries[i].dispclk_mhz;
+		if (bw_params->clk_table.entries[i].dppclk_mhz > max_dppclk_mhz)
+			max_dppclk_mhz = bw_params->clk_table.entries[i].dppclk_mhz;
+		if (bw_params->clk_table.entries[i].phyclk_mhz > max_phyclk_mhz)
+			max_phyclk_mhz = bw_params->clk_table.entries[i].phyclk_mhz;
+		if (bw_params->clk_table.entries[i].dtbclk_mhz > max_dtbclk_mhz)
+			max_dtbclk_mhz = bw_params->clk_table.entries[i].dtbclk_mhz;
+	}
+
+	/* Scan through clock values we currently have and if they are 0,
+	 *  then populate it with dcn3_2_soc.clock_limits[] value.
+	 *
+	 * Do it for DCFCLK, DISPCLK, DTBCLK and UCLK as any of those being
+	 *  0, will cause it to skip building the clock table.
+	 */
+	if (max_dcfclk_mhz == 0)
+		bw_params->clk_table.entries[0].dcfclk_mhz = dcn3_2_soc.clock_limits[0].dcfclk_mhz;
+	if (max_dispclk_mhz == 0)
+		bw_params->clk_table.entries[0].dispclk_mhz = dcn3_2_soc.clock_limits[0].dispclk_mhz;
+	if (max_dtbclk_mhz == 0)
+		bw_params->clk_table.entries[0].dtbclk_mhz = dcn3_2_soc.clock_limits[0].dtbclk_mhz;
+	if (max_uclk_mhz == 0)
+		bw_params->clk_table.entries[0].memclk_mhz = dcn3_2_soc.clock_limits[0].dram_speed_mts / 16;
+}
+
+static void swap_table_entries(struct _vcs_dpi_voltage_scaling_st *first_entry,
+		struct _vcs_dpi_voltage_scaling_st *second_entry)
+{
+	struct _vcs_dpi_voltage_scaling_st temp_entry = *first_entry;
+	*first_entry = *second_entry;
+	*second_entry = temp_entry;
+}
+
+/*
+ * sort_entries_with_same_bw - Sort entries sharing the same bandwidth by DCFCLK
+ */
+static void sort_entries_with_same_bw(struct _vcs_dpi_voltage_scaling_st *table, unsigned int *num_entries)
+{
+	unsigned int start_index = 0;
+	unsigned int end_index = 0;
+	unsigned int current_bw = 0;
+
+	for (int i = 0; i < (*num_entries - 1); i++) {
+		if (table[i].net_bw_in_kbytes_sec == table[i+1].net_bw_in_kbytes_sec) {
+			current_bw = table[i].net_bw_in_kbytes_sec;
+			start_index = i;
+			end_index = ++i;
+
+			while ((i < (*num_entries - 1)) && (table[i+1].net_bw_in_kbytes_sec == current_bw))
+				end_index = ++i;
+		}
+
+		if (start_index != end_index) {
+			for (int j = start_index; j < end_index; j++) {
+				for (int k = start_index; k < end_index; k++) {
+					if (table[k].dcfclk_mhz > table[k+1].dcfclk_mhz)
+						swap_table_entries(&table[k], &table[k+1]);
+				}
+			}
+		}
+
+		start_index = 0;
+		end_index = 0;
+
+	}
+}
+
+/*
+ * remove_inconsistent_entries - Ensure entries with the same bandwidth have MEMCLK and FCLK monotonically increasing
+ *                               and remove entries that do not
+ */
+static void remove_inconsistent_entries(struct _vcs_dpi_voltage_scaling_st *table, unsigned int *num_entries)
+{
+	for (int i = 0; i < (*num_entries - 1); i++) {
+		if (table[i].net_bw_in_kbytes_sec == table[i+1].net_bw_in_kbytes_sec) {
+			if ((table[i].dram_speed_mts > table[i+1].dram_speed_mts) ||
+				(table[i].fabricclk_mhz > table[i+1].fabricclk_mhz))
+				remove_entry_from_table_at_index(table, num_entries, i);
+		}
+	}
+}
+
+/*
+ * override_max_clk_values - Overwrite the max clock frequencies with the max DC mode timings
+ * Input:
+ *	max_clk_limit - struct containing the desired clock timings
+ * Output:
+ *	curr_clk_limit  - struct containing the timings that need to be overwritten
+ * Return: 0 upon success, non-zero for failure
+ */
+static int override_max_clk_values(struct clk_limit_table_entry *max_clk_limit,
+		struct clk_limit_table_entry *curr_clk_limit)
+{
+	if (NULL == max_clk_limit || NULL == curr_clk_limit)
+		return -1; //invalid parameters
+
+	//only overwrite if desired max clock frequency is initialized
+	if (max_clk_limit->dcfclk_mhz != 0)
+		curr_clk_limit->dcfclk_mhz = max_clk_limit->dcfclk_mhz;
+
+	if (max_clk_limit->fclk_mhz != 0)
+		curr_clk_limit->fclk_mhz = max_clk_limit->fclk_mhz;
+
+	if (max_clk_limit->memclk_mhz != 0)
+		curr_clk_limit->memclk_mhz = max_clk_limit->memclk_mhz;
+
+	if (max_clk_limit->socclk_mhz != 0)
+		curr_clk_limit->socclk_mhz = max_clk_limit->socclk_mhz;
+
+	if (max_clk_limit->dtbclk_mhz != 0)
+		curr_clk_limit->dtbclk_mhz = max_clk_limit->dtbclk_mhz;
+
+	if (max_clk_limit->dispclk_mhz != 0)
+		curr_clk_limit->dispclk_mhz = max_clk_limit->dispclk_mhz;
+
+	return 0;
+}
+
+static int build_synthetic_soc_states(bool disable_dc_mode_overwrite, struct clk_bw_params *bw_params,
+		struct _vcs_dpi_voltage_scaling_st *table, unsigned int *num_entries)
+{
+	int i, j;
+	struct _vcs_dpi_voltage_scaling_st entry = {0};
+	struct clk_limit_table_entry max_clk_data = {0};
+
+	unsigned int min_dcfclk_mhz = 199, min_fclk_mhz = 299;
+
+	static const unsigned int num_dcfclk_stas = 5;
+	unsigned int dcfclk_sta_targets[DC__VOLTAGE_STATES] = {199, 615, 906, 1324, 1564};
+
+	unsigned int num_uclk_dpms = 0;
+	unsigned int num_fclk_dpms = 0;
+	unsigned int num_dcfclk_dpms = 0;
+
+	unsigned int num_dc_uclk_dpms = 0;
+	unsigned int num_dc_fclk_dpms = 0;
+	unsigned int num_dc_dcfclk_dpms = 0;
+
+	for (i = 0; i < MAX_NUM_DPM_LVL; i++) {
+		if (bw_params->clk_table.entries[i].dcfclk_mhz > max_clk_data.dcfclk_mhz)
+			max_clk_data.dcfclk_mhz = bw_params->clk_table.entries[i].dcfclk_mhz;
+		if (bw_params->clk_table.entries[i].fclk_mhz > max_clk_data.fclk_mhz)
+			max_clk_data.fclk_mhz = bw_params->clk_table.entries[i].fclk_mhz;
+		if (bw_params->clk_table.entries[i].memclk_mhz > max_clk_data.memclk_mhz)
+			max_clk_data.memclk_mhz = bw_params->clk_table.entries[i].memclk_mhz;
+		if (bw_params->clk_table.entries[i].dispclk_mhz > max_clk_data.dispclk_mhz)
+			max_clk_data.dispclk_mhz = bw_params->clk_table.entries[i].dispclk_mhz;
+		if (bw_params->clk_table.entries[i].dppclk_mhz > max_clk_data.dppclk_mhz)
+			max_clk_data.dppclk_mhz = bw_params->clk_table.entries[i].dppclk_mhz;
+		if (bw_params->clk_table.entries[i].phyclk_mhz > max_clk_data.phyclk_mhz)
+			max_clk_data.phyclk_mhz = bw_params->clk_table.entries[i].phyclk_mhz;
+		if (bw_params->clk_table.entries[i].dtbclk_mhz > max_clk_data.dtbclk_mhz)
+			max_clk_data.dtbclk_mhz = bw_params->clk_table.entries[i].dtbclk_mhz;
+
+		if (bw_params->clk_table.entries[i].memclk_mhz > 0) {
+			num_uclk_dpms++;
+			if (bw_params->clk_table.entries[i].memclk_mhz <= bw_params->dc_mode_limit.memclk_mhz)
+				num_dc_uclk_dpms++;
+		}
+		if (bw_params->clk_table.entries[i].fclk_mhz > 0) {
+			num_fclk_dpms++;
+			if (bw_params->clk_table.entries[i].fclk_mhz <= bw_params->dc_mode_limit.fclk_mhz)
+				num_dc_fclk_dpms++;
+		}
+		if (bw_params->clk_table.entries[i].dcfclk_mhz > 0) {
+			num_dcfclk_dpms++;
+			if (bw_params->clk_table.entries[i].dcfclk_mhz <= bw_params->dc_mode_limit.dcfclk_mhz)
+				num_dc_dcfclk_dpms++;
+		}
+	}
+
+	if (!disable_dc_mode_overwrite) {
+		//Overwrite max frequencies with max DC mode frequencies for DC mode systems
+		override_max_clk_values(&bw_params->dc_mode_limit, &max_clk_data);
+		num_uclk_dpms = num_dc_uclk_dpms;
+		num_fclk_dpms = num_dc_fclk_dpms;
+		num_dcfclk_dpms = num_dc_dcfclk_dpms;
+		bw_params->clk_table.num_entries_per_clk.num_memclk_levels = num_uclk_dpms;
+		bw_params->clk_table.num_entries_per_clk.num_fclk_levels = num_fclk_dpms;
+	}
+
+	if (num_dcfclk_dpms > 0 && bw_params->clk_table.entries[0].fclk_mhz > min_fclk_mhz)
+		min_fclk_mhz = bw_params->clk_table.entries[0].fclk_mhz;
+
+	if (!max_clk_data.dcfclk_mhz || !max_clk_data.dispclk_mhz || !max_clk_data.dtbclk_mhz)
+		return -1;
+
+	if (max_clk_data.dppclk_mhz == 0)
+		max_clk_data.dppclk_mhz = max_clk_data.dispclk_mhz;
+
+	if (max_clk_data.fclk_mhz == 0)
+		max_clk_data.fclk_mhz = max_clk_data.dcfclk_mhz *
+				dcn3_2_soc.pct_ideal_sdp_bw_after_urgent /
+				dcn3_2_soc.pct_ideal_fabric_bw_after_urgent;
+
+	if (max_clk_data.phyclk_mhz == 0)
+		max_clk_data.phyclk_mhz = dcn3_2_soc.clock_limits[0].phyclk_mhz;
+
+	*num_entries = 0;
+	entry.dispclk_mhz = max_clk_data.dispclk_mhz;
+	entry.dscclk_mhz = max_clk_data.dispclk_mhz / 3;
+	entry.dppclk_mhz = max_clk_data.dppclk_mhz;
+	entry.dtbclk_mhz = max_clk_data.dtbclk_mhz;
+	entry.phyclk_mhz = max_clk_data.phyclk_mhz;
+	entry.phyclk_d18_mhz = dcn3_2_soc.clock_limits[0].phyclk_d18_mhz;
+	entry.phyclk_d32_mhz = dcn3_2_soc.clock_limits[0].phyclk_d32_mhz;
+
+	// Insert all the DCFCLK STAs
+	for (i = 0; i < num_dcfclk_stas; i++) {
+		entry.dcfclk_mhz = dcfclk_sta_targets[i];
+		entry.fabricclk_mhz = 0;
+		entry.dram_speed_mts = 0;
+
+		get_optimal_ntuple(&entry);
+		entry.net_bw_in_kbytes_sec = calculate_net_bw_in_kbytes_sec(&entry);
+		insert_entry_into_table_sorted(table, num_entries, &entry);
+	}
+
+	// Insert the max DCFCLK
+	entry.dcfclk_mhz = max_clk_data.dcfclk_mhz;
+	entry.fabricclk_mhz = 0;
+	entry.dram_speed_mts = 0;
+
+	get_optimal_ntuple(&entry);
+	entry.net_bw_in_kbytes_sec = calculate_net_bw_in_kbytes_sec(&entry);
+	insert_entry_into_table_sorted(table, num_entries, &entry);
+
+	// Insert the UCLK DPMS
+	for (i = 0; i < num_uclk_dpms; i++) {
+		entry.dcfclk_mhz = 0;
+		entry.fabricclk_mhz = 0;
+		entry.dram_speed_mts = bw_params->clk_table.entries[i].memclk_mhz * 16;
+
+		get_optimal_ntuple(&entry);
+		entry.net_bw_in_kbytes_sec = calculate_net_bw_in_kbytes_sec(&entry);
+		insert_entry_into_table_sorted(table, num_entries, &entry);
+	}
+
+	// If FCLK is coarse grained, insert individual DPMs.
+	if (num_fclk_dpms > 2) {
+		for (i = 0; i < num_fclk_dpms; i++) {
+			entry.dcfclk_mhz = 0;
+			entry.fabricclk_mhz = bw_params->clk_table.entries[i].fclk_mhz;
+			entry.dram_speed_mts = 0;
+
+			get_optimal_ntuple(&entry);
+			entry.net_bw_in_kbytes_sec = calculate_net_bw_in_kbytes_sec(&entry);
+			insert_entry_into_table_sorted(table, num_entries, &entry);
+		}
+	}
+	// If FCLK fine grained, only insert max
+	else {
+		entry.dcfclk_mhz = 0;
+		entry.fabricclk_mhz = max_clk_data.fclk_mhz;
+		entry.dram_speed_mts = 0;
+
+		get_optimal_ntuple(&entry);
+		entry.net_bw_in_kbytes_sec = calculate_net_bw_in_kbytes_sec(&entry);
+		insert_entry_into_table_sorted(table, num_entries, &entry);
+	}
+
+	// At this point, the table contains all "points of interest" based on
+	// DPMs from PMFW, and STAs.  Table is sorted by BW, and all clock
+	// ratios (by derate, are exact).
+
+	// Remove states that require higher clocks than are supported
+	for (i = *num_entries - 1; i >= 0 ; i--) {
+		if (table[i].dcfclk_mhz > max_clk_data.dcfclk_mhz ||
+				table[i].fabricclk_mhz > max_clk_data.fclk_mhz ||
+				table[i].dram_speed_mts > max_clk_data.memclk_mhz * 16)
+			remove_entry_from_table_at_index(table, num_entries, i);
+	}
+
+	// Insert entry with all max dc limits without bandwidth matching
+	if (!disable_dc_mode_overwrite) {
+		struct _vcs_dpi_voltage_scaling_st max_dc_limits_entry = entry;
+
+		max_dc_limits_entry.dcfclk_mhz = max_clk_data.dcfclk_mhz;
+		max_dc_limits_entry.fabricclk_mhz = max_clk_data.fclk_mhz;
+		max_dc_limits_entry.dram_speed_mts = max_clk_data.memclk_mhz * 16;
+
+		max_dc_limits_entry.net_bw_in_kbytes_sec = calculate_net_bw_in_kbytes_sec(&max_dc_limits_entry);
+		insert_entry_into_table_sorted(table, num_entries, &max_dc_limits_entry);
+
+		sort_entries_with_same_bw(table, num_entries);
+		remove_inconsistent_entries(table, num_entries);
+	}
+
+	// At this point, the table only contains supported points of interest
+	// it could be used as is, but some states may be redundant due to
+	// coarse grained nature of some clocks, so we want to round up to
+	// coarse grained DPMs and remove duplicates.
+
+	// Round up UCLKs
+	for (i = *num_entries - 1; i >= 0 ; i--) {
+		for (j = 0; j < num_uclk_dpms; j++) {
+			if (bw_params->clk_table.entries[j].memclk_mhz * 16 >= table[i].dram_speed_mts) {
+				table[i].dram_speed_mts = bw_params->clk_table.entries[j].memclk_mhz * 16;
+				break;
+			}
+		}
+	}
+
+	// If FCLK is coarse grained, round up to next DPMs
+	if (num_fclk_dpms > 2) {
+		for (i = *num_entries - 1; i >= 0 ; i--) {
+			for (j = 0; j < num_fclk_dpms; j++) {
+				if (bw_params->clk_table.entries[j].fclk_mhz >= table[i].fabricclk_mhz) {
+					table[i].fabricclk_mhz = bw_params->clk_table.entries[j].fclk_mhz;
+					break;
+				}
+			}
+		}
+	}
+	// Otherwise, round up to minimum.
+	else {
+		for (i = *num_entries - 1; i >= 0 ; i--) {
+			if (table[i].fabricclk_mhz < min_fclk_mhz) {
+				table[i].fabricclk_mhz = min_fclk_mhz;
+			}
+		}
+	}
+
+	// Round DCFCLKs up to minimum
+	for (i = *num_entries - 1; i >= 0 ; i--) {
+		if (table[i].dcfclk_mhz < min_dcfclk_mhz) {
+			table[i].dcfclk_mhz = min_dcfclk_mhz;
+		}
+	}
+
+	// Remove duplicate states, note duplicate states are always neighbouring since table is sorted.
+	i = 0;
+	while (i < *num_entries - 1) {
+		if (table[i].dcfclk_mhz == table[i + 1].dcfclk_mhz &&
+				table[i].fabricclk_mhz == table[i + 1].fabricclk_mhz &&
+				table[i].dram_speed_mts == table[i + 1].dram_speed_mts)
+			remove_entry_from_table_at_index(table, num_entries, i + 1);
+		else
+			i++;
+	}
+
+	// Fix up the state indicies
+	for (i = *num_entries - 1; i >= 0 ; i--) {
+		table[i].state = i;
+	}
+
+	return 0;
+}
+
+/*
+ * dcn32_update_bw_bounding_box
+ *
+ * This would override some dcn3_2 ip_or_soc initial parameters hardcoded from
+ * spreadsheet with actual values as per dGPU SKU:
+ * - with passed few options from dc->config
+ * - with dentist_vco_frequency from Clk Mgr (currently hardcoded, but might
+ *   need to get it from PM FW)
+ * - with passed latency values (passed in ns units) in dc-> bb override for
+ *   debugging purposes
+ * - with passed latencies from VBIOS (in 100_ns units) if available for
+ *   certain dGPU SKU
+ * - with number of DRAM channels from VBIOS (which differ for certain dGPU SKU
+ *   of the same ASIC)
+ * - clocks levels with passed clk_table entries from Clk Mgr as reported by PM
+ *   FW for different clocks (which might differ for certain dGPU SKU of the
+ *   same ASIC)
+ */
+void dcn32_update_bw_bounding_box_fpu(struct dc *dc, struct clk_bw_params *bw_params)
+{
+	dc_assert_fp_enabled();
+
+	/* Overrides from dc->config options */
+	dcn3_2_ip.clamp_min_dcfclk = dc->config.clamp_min_dcfclk;
+
+	/* Override from passed dc->bb_overrides if available*/
+	if ((int)(dcn3_2_soc.sr_exit_time_us * 1000) != dc->bb_overrides.sr_exit_time_ns
+			&& dc->bb_overrides.sr_exit_time_ns) {
+		dc->dml2_options.bbox_overrides.sr_exit_latency_us =
+		dcn3_2_soc.sr_exit_time_us = dc->bb_overrides.sr_exit_time_ns / 1000.0;
+	}
+
+	if ((int)(dcn3_2_soc.sr_enter_plus_exit_time_us * 1000)
+			!= dc->bb_overrides.sr_enter_plus_exit_time_ns
+			&& dc->bb_overrides.sr_enter_plus_exit_time_ns) {
+		dc->dml2_options.bbox_overrides.sr_enter_plus_exit_latency_us =
+		dcn3_2_soc.sr_enter_plus_exit_time_us =
+			dc->bb_overrides.sr_enter_plus_exit_time_ns / 1000.0;
+	}
+
+	if ((int)(dcn3_2_soc.urgent_latency_us * 1000) != dc->bb_overrides.urgent_latency_ns
+		&& dc->bb_overrides.urgent_latency_ns) {
+		dcn3_2_soc.urgent_latency_us = dc->bb_overrides.urgent_latency_ns / 1000.0;
+		dc->dml2_options.bbox_overrides.urgent_latency_us =
+		dcn3_2_soc.urgent_latency_pixel_data_only_us = dc->bb_overrides.urgent_latency_ns / 1000.0;
+	}
+
+	if ((int)(dcn3_2_soc.dram_clock_change_latency_us * 1000)
+			!= dc->bb_overrides.dram_clock_change_latency_ns
+			&& dc->bb_overrides.dram_clock_change_latency_ns) {
+		dc->dml2_options.bbox_overrides.dram_clock_change_latency_us =
+		dcn3_2_soc.dram_clock_change_latency_us =
+			dc->bb_overrides.dram_clock_change_latency_ns / 1000.0;
+	}
+
+	if ((int)(dcn3_2_soc.fclk_change_latency_us * 1000)
+			!= dc->bb_overrides.fclk_clock_change_latency_ns
+			&& dc->bb_overrides.fclk_clock_change_latency_ns) {
+		dc->dml2_options.bbox_overrides.fclk_change_latency_us =
+		dcn3_2_soc.fclk_change_latency_us =
+			dc->bb_overrides.fclk_clock_change_latency_ns / 1000;
+	}
+
+	if ((int)(dcn3_2_soc.dummy_pstate_latency_us * 1000)
+			!= dc->bb_overrides.dummy_clock_change_latency_ns
+			&& dc->bb_overrides.dummy_clock_change_latency_ns) {
+		dcn3_2_soc.dummy_pstate_latency_us =
+			dc->bb_overrides.dummy_clock_change_latency_ns / 1000.0;
+	}
+
+	/* Override from VBIOS if VBIOS bb_info available */
+	if (dc->ctx->dc_bios->funcs->get_soc_bb_info) {
+		struct bp_soc_bb_info bb_info = {0};
+
+		if (dc->ctx->dc_bios->funcs->get_soc_bb_info(dc->ctx->dc_bios, &bb_info) == BP_RESULT_OK) {
+			if (bb_info.dram_clock_change_latency_100ns > 0)
+				dc->dml2_options.bbox_overrides.dram_clock_change_latency_us =
+				dcn3_2_soc.dram_clock_change_latency_us =
+					bb_info.dram_clock_change_latency_100ns * 10;
+
+			if (bb_info.dram_sr_enter_exit_latency_100ns > 0)
+				dc->dml2_options.bbox_overrides.sr_enter_plus_exit_latency_us =
+				dcn3_2_soc.sr_enter_plus_exit_time_us =
+					bb_info.dram_sr_enter_exit_latency_100ns * 10;
+
+			if (bb_info.dram_sr_exit_latency_100ns > 0)
+				dc->dml2_options.bbox_overrides.sr_exit_latency_us =
+				dcn3_2_soc.sr_exit_time_us =
+					bb_info.dram_sr_exit_latency_100ns * 10;
+		}
+	}
+
+	/* Override from VBIOS for num_chan */
+	if (dc->ctx->dc_bios->vram_info.num_chans) {
+		dc->dml2_options.bbox_overrides.dram_num_chan =
+		dcn3_2_soc.num_chans = dc->ctx->dc_bios->vram_info.num_chans;
+		dcn3_2_soc.mall_allocated_for_dcn_mbytes = (double)(dcn32_calc_num_avail_chans_for_mall(dc,
+			dc->ctx->dc_bios->vram_info.num_chans) * dc->caps.mall_size_per_mem_channel);
+	}
+
+	if (dc->ctx->dc_bios->vram_info.dram_channel_width_bytes)
+		dc->dml2_options.bbox_overrides.dram_chanel_width_bytes =
+		dcn3_2_soc.dram_channel_width_bytes = dc->ctx->dc_bios->vram_info.dram_channel_width_bytes;
+
+	/* DML DSC delay factor workaround */
+	dcn3_2_ip.dsc_delay_factor_wa = dc->debug.dsc_delay_factor_wa_x1000 / 1000.0;
+
+	dcn3_2_ip.min_prefetch_in_strobe_us = dc->debug.min_prefetch_in_strobe_ns / 1000.0;
+
+	/* Override dispclk_dppclk_vco_speed_mhz from Clk Mgr */
+	dcn3_2_soc.dispclk_dppclk_vco_speed_mhz = dc->clk_mgr->dentist_vco_freq_khz / 1000.0;
+	dc->dml.soc.dispclk_dppclk_vco_speed_mhz = dc->clk_mgr->dentist_vco_freq_khz / 1000.0;
+	dc->dml2_options.bbox_overrides.disp_pll_vco_speed_mhz = dc->clk_mgr->dentist_vco_freq_khz / 1000.0;
+	dc->dml2_options.bbox_overrides.xtalclk_mhz = dc->ctx->dc_bios->fw_info.pll_info.crystal_frequency / 1000.0;
+	dc->dml2_options.bbox_overrides.dchub_refclk_mhz = dc->res_pool->ref_clocks.dchub_ref_clock_inKhz / 1000.0;
+	dc->dml2_options.bbox_overrides.dprefclk_mhz = dc->clk_mgr->dprefclk_khz / 1000.0;
+
+	/* Overrides Clock levelsfrom CLK Mgr table entries as reported by PM FW */
+	if (bw_params->clk_table.entries[0].memclk_mhz) {
+		if (dc->debug.use_legacy_soc_bb_mechanism) {
+			unsigned int i = 0, j = 0, num_states = 0;
+
+			unsigned int dcfclk_mhz[DC__VOLTAGE_STATES] = {0};
+			unsigned int dram_speed_mts[DC__VOLTAGE_STATES] = {0};
+			unsigned int optimal_uclk_for_dcfclk_sta_targets[DC__VOLTAGE_STATES] = {0};
+			unsigned int optimal_dcfclk_for_uclk[DC__VOLTAGE_STATES] = {0};
+			unsigned int min_dcfclk = UINT_MAX;
+			/* Set 199 as first value in STA target array to have a minimum DCFCLK value.
+			 * For DCN32 we set min to 199 so minimum FCLK DPM0 (300Mhz can be achieved) */
+			unsigned int dcfclk_sta_targets[DC__VOLTAGE_STATES] = {199, 615, 906, 1324, 1564};
+			unsigned int num_dcfclk_sta_targets = 4, num_uclk_states = 0;
+			unsigned int max_dcfclk_mhz = 0, max_dispclk_mhz = 0, max_dppclk_mhz = 0, max_phyclk_mhz = 0;
+
+			for (i = 0; i < MAX_NUM_DPM_LVL; i++) {
+				if (bw_params->clk_table.entries[i].dcfclk_mhz > max_dcfclk_mhz)
+					max_dcfclk_mhz = bw_params->clk_table.entries[i].dcfclk_mhz;
+				if (bw_params->clk_table.entries[i].dcfclk_mhz != 0 &&
+						bw_params->clk_table.entries[i].dcfclk_mhz < min_dcfclk)
+					min_dcfclk = bw_params->clk_table.entries[i].dcfclk_mhz;
+				if (bw_params->clk_table.entries[i].dispclk_mhz > max_dispclk_mhz)
+					max_dispclk_mhz = bw_params->clk_table.entries[i].dispclk_mhz;
+				if (bw_params->clk_table.entries[i].dppclk_mhz > max_dppclk_mhz)
+					max_dppclk_mhz = bw_params->clk_table.entries[i].dppclk_mhz;
+				if (bw_params->clk_table.entries[i].phyclk_mhz > max_phyclk_mhz)
+					max_phyclk_mhz = bw_params->clk_table.entries[i].phyclk_mhz;
+			}
+			if (min_dcfclk > dcfclk_sta_targets[0])
+				dcfclk_sta_targets[0] = min_dcfclk;
+			if (!max_dcfclk_mhz)
+				max_dcfclk_mhz = dcn3_2_soc.clock_limits[0].dcfclk_mhz;
+			if (!max_dispclk_mhz)
+				max_dispclk_mhz = dcn3_2_soc.clock_limits[0].dispclk_mhz;
+			if (!max_dppclk_mhz)
+				max_dppclk_mhz = dcn3_2_soc.clock_limits[0].dppclk_mhz;
+			if (!max_phyclk_mhz)
+				max_phyclk_mhz = dcn3_2_soc.clock_limits[0].phyclk_mhz;
+
+			if (max_dcfclk_mhz > dcfclk_sta_targets[num_dcfclk_sta_targets-1]) {
+				// If max DCFCLK is greater than the max DCFCLK STA target, insert into the DCFCLK STA target array
+				dcfclk_sta_targets[num_dcfclk_sta_targets] = max_dcfclk_mhz;
+				num_dcfclk_sta_targets++;
+			} else if (max_dcfclk_mhz < dcfclk_sta_targets[num_dcfclk_sta_targets-1]) {
+				// If max DCFCLK is less than the max DCFCLK STA target, cap values and remove duplicates
+				for (i = 0; i < num_dcfclk_sta_targets; i++) {
+					if (dcfclk_sta_targets[i] > max_dcfclk_mhz) {
+						dcfclk_sta_targets[i] = max_dcfclk_mhz;
+						break;
+					}
+				}
+				// Update size of array since we "removed" duplicates
+				num_dcfclk_sta_targets = i + 1;
+			}
+
+			num_uclk_states = bw_params->clk_table.num_entries;
+
+			// Calculate optimal dcfclk for each uclk
+			for (i = 0; i < num_uclk_states; i++) {
+				dcn32_get_optimal_dcfclk_fclk_for_uclk(bw_params->clk_table.entries[i].memclk_mhz * 16,
+						&optimal_dcfclk_for_uclk[i], NULL);
+				if (optimal_dcfclk_for_uclk[i] < bw_params->clk_table.entries[0].dcfclk_mhz) {
+					optimal_dcfclk_for_uclk[i] = bw_params->clk_table.entries[0].dcfclk_mhz;
+				}
+			}
+
+			// Calculate optimal uclk for each dcfclk sta target
+			for (i = 0; i < num_dcfclk_sta_targets; i++) {
+				for (j = 0; j < num_uclk_states; j++) {
+					if (dcfclk_sta_targets[i] < optimal_dcfclk_for_uclk[j]) {
+						optimal_uclk_for_dcfclk_sta_targets[i] =
+								bw_params->clk_table.entries[j].memclk_mhz * 16;
+						break;
+					}
+				}
+			}
+
+			i = 0;
+			j = 0;
+			// create the final dcfclk and uclk table
+			while (i < num_dcfclk_sta_targets && j < num_uclk_states && num_states < DC__VOLTAGE_STATES) {
+				if (dcfclk_sta_targets[i] < optimal_dcfclk_for_uclk[j]) {
+					dcfclk_mhz[num_states] = dcfclk_sta_targets[i];
+					dram_speed_mts[num_states++] = optimal_uclk_for_dcfclk_sta_targets[i++];
+				} else {
+					if (j < num_uclk_states && optimal_dcfclk_for_uclk[j] <= max_dcfclk_mhz) {
+						dcfclk_mhz[num_states] = optimal_dcfclk_for_uclk[j];
+						dram_speed_mts[num_states++] = bw_params->clk_table.entries[j++].memclk_mhz * 16;
+					} else {
+						j = num_uclk_states;
+					}
+				}
+			}
+
+			while (i < num_dcfclk_sta_targets && num_states < DC__VOLTAGE_STATES) {
+				dcfclk_mhz[num_states] = dcfclk_sta_targets[i];
+				dram_speed_mts[num_states++] = optimal_uclk_for_dcfclk_sta_targets[i++];
+			}
+
+			while (j < num_uclk_states && num_states < DC__VOLTAGE_STATES &&
+					optimal_dcfclk_for_uclk[j] <= max_dcfclk_mhz) {
+				dcfclk_mhz[num_states] = optimal_dcfclk_for_uclk[j];
+				dram_speed_mts[num_states++] = bw_params->clk_table.entries[j++].memclk_mhz * 16;
+			}
+
+			/* bw_params->clk_table.entries[MAX_NUM_DPM_LVL].
+			 * MAX_NUM_DPM_LVL is 8.
+			 * dcn3_02_soc.clock_limits[DC__VOLTAGE_STATES].
+			 * DC__VOLTAGE_STATES is 40.
+			 */
+			if (num_states > MAX_NUM_DPM_LVL) {
+				ASSERT(0);
+				return;
+			}
+
+			dcn3_2_soc.num_states = num_states;
+			for (i = 0; i < dcn3_2_soc.num_states; i++) {
+				dcn3_2_soc.clock_limits[i].state = i;
+				dcn3_2_soc.clock_limits[i].dcfclk_mhz = dcfclk_mhz[i];
+				dcn3_2_soc.clock_limits[i].fabricclk_mhz = dcfclk_mhz[i];
+
+				/* Fill all states with max values of all these clocks */
+				dcn3_2_soc.clock_limits[i].dispclk_mhz = max_dispclk_mhz;
+				dcn3_2_soc.clock_limits[i].dppclk_mhz  = max_dppclk_mhz;
+				dcn3_2_soc.clock_limits[i].phyclk_mhz  = max_phyclk_mhz;
+				dcn3_2_soc.clock_limits[i].dscclk_mhz  = max_dispclk_mhz / 3;
+
+				/* Populate from bw_params for DTBCLK, SOCCLK */
+				if (i > 0) {
+					if (!bw_params->clk_table.entries[i].dtbclk_mhz) {
+						dcn3_2_soc.clock_limits[i].dtbclk_mhz  = dcn3_2_soc.clock_limits[i-1].dtbclk_mhz;
+					} else {
+						dcn3_2_soc.clock_limits[i].dtbclk_mhz  = bw_params->clk_table.entries[i].dtbclk_mhz;
+					}
+				} else if (bw_params->clk_table.entries[i].dtbclk_mhz) {
+					dcn3_2_soc.clock_limits[i].dtbclk_mhz  = bw_params->clk_table.entries[i].dtbclk_mhz;
+				}
+
+				if (!bw_params->clk_table.entries[i].socclk_mhz && i > 0)
+					dcn3_2_soc.clock_limits[i].socclk_mhz = dcn3_2_soc.clock_limits[i-1].socclk_mhz;
+				else
+					dcn3_2_soc.clock_limits[i].socclk_mhz = bw_params->clk_table.entries[i].socclk_mhz;
+
+				if (!dram_speed_mts[i] && i > 0)
+					dcn3_2_soc.clock_limits[i].dram_speed_mts = dcn3_2_soc.clock_limits[i-1].dram_speed_mts;
+				else
+					dcn3_2_soc.clock_limits[i].dram_speed_mts = dram_speed_mts[i];
+
+				/* These clocks cannot come from bw_params, always fill from dcn3_2_soc[0] */
+				/* PHYCLK_D18, PHYCLK_D32 */
+				dcn3_2_soc.clock_limits[i].phyclk_d18_mhz = dcn3_2_soc.clock_limits[0].phyclk_d18_mhz;
+				dcn3_2_soc.clock_limits[i].phyclk_d32_mhz = dcn3_2_soc.clock_limits[0].phyclk_d32_mhz;
+			}
+		} else {
+			build_synthetic_soc_states(dc->debug.disable_dc_mode_overwrite, bw_params,
+					dcn3_2_soc.clock_limits, &dcn3_2_soc.num_states);
+		}
+
+		/* Re-init DML with updated bb */
+		dml_init_instance(&dc->dml, &dcn3_2_soc, &dcn3_2_ip, DML_PROJECT_DCN32);
+		if (dc->current_state)
+			dml_init_instance(&dc->current_state->bw_ctx.dml, &dcn3_2_soc, &dcn3_2_ip, DML_PROJECT_DCN32);
+	}
+
+	if (dc->clk_mgr->bw_params->clk_table.num_entries > 1) {
+		unsigned int i = 0;
+
+		dc->dml2_options.bbox_overrides.clks_table.num_states = dc->clk_mgr->bw_params->clk_table.num_entries;
+
+		dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_dcfclk_levels =
+			dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_dcfclk_levels;
+
+		dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_fclk_levels =
+			dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_fclk_levels;
+
+		dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_memclk_levels =
+			dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_memclk_levels;
+
+		dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_socclk_levels =
+			dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_socclk_levels;
+
+		dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_dtbclk_levels =
+			dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_dtbclk_levels;
+
+		dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_dispclk_levels =
+			dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_dispclk_levels;
+
+		dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_dppclk_levels =
+			dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_dppclk_levels;
+
+		for (i = 0; i < dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_dcfclk_levels; i++) {
+			if (dc->clk_mgr->bw_params->clk_table.entries[i].dcfclk_mhz)
+				dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].dcfclk_mhz =
+					dc->clk_mgr->bw_params->clk_table.entries[i].dcfclk_mhz;
+		}
+
+		for (i = 0; i < dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_fclk_levels; i++) {
+			if (dc->clk_mgr->bw_params->clk_table.entries[i].fclk_mhz)
+				dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].fclk_mhz =
+					dc->clk_mgr->bw_params->clk_table.entries[i].fclk_mhz;
+		}
+
+		for (i = 0; i < dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_memclk_levels; i++) {
+			if (dc->clk_mgr->bw_params->clk_table.entries[i].memclk_mhz)
+				dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].memclk_mhz =
+					dc->clk_mgr->bw_params->clk_table.entries[i].memclk_mhz;
+		}
+
+		for (i = 0; i < dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_socclk_levels; i++) {
+			if (dc->clk_mgr->bw_params->clk_table.entries[i].socclk_mhz)
+				dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].socclk_mhz =
+					dc->clk_mgr->bw_params->clk_table.entries[i].socclk_mhz;
+		}
+
+		for (i = 0; i < dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_dtbclk_levels; i++) {
+			if (dc->clk_mgr->bw_params->clk_table.entries[i].dtbclk_mhz)
+				dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].dtbclk_mhz =
+					dc->clk_mgr->bw_params->clk_table.entries[i].dtbclk_mhz;
+		}
+
+		for (i = 0; i < dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_dispclk_levels; i++) {
+			if (dc->clk_mgr->bw_params->clk_table.entries[i].dispclk_mhz) {
+				dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].dispclk_mhz =
+					dc->clk_mgr->bw_params->clk_table.entries[i].dispclk_mhz;
+				dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].dppclk_mhz =
+					dc->clk_mgr->bw_params->clk_table.entries[i].dispclk_mhz;
+			}
+		}
+	}
+}
+
+void dcn32_zero_pipe_dcc_fraction(display_e2e_pipe_params_st *pipes,
+				  int pipe_cnt)
+{
+	dc_assert_fp_enabled();
+
+	pipes[pipe_cnt].pipe.src.dcc_fraction_of_zs_req_luma = 0;
+	pipes[pipe_cnt].pipe.src.dcc_fraction_of_zs_req_chroma = 0;
+}
+
+bool dcn32_allow_subvp_with_active_margin(struct pipe_ctx *pipe)
+{
+	bool allow = false;
+	uint32_t refresh_rate = 0;
+	uint32_t min_refresh = subvp_active_margin_list.min_refresh;
+	uint32_t max_refresh = subvp_active_margin_list.max_refresh;
+	uint32_t i;
+
+	for (i = 0; i < SUBVP_ACTIVE_MARGIN_LIST_LEN; i++) {
+		uint32_t width = subvp_active_margin_list.res[i].width;
+		uint32_t height = subvp_active_margin_list.res[i].height;
+
+		refresh_rate = (pipe->stream->timing.pix_clk_100hz * (uint64_t)100 +
+			(uint64_t)pipe->stream->timing.v_total * pipe->stream->timing.h_total - (uint64_t)1);
+		refresh_rate = div_u64(refresh_rate, pipe->stream->timing.v_total);
+		refresh_rate = div_u64(refresh_rate, pipe->stream->timing.h_total);
+
+		if (refresh_rate >= min_refresh && refresh_rate <= max_refresh &&
+				dcn32_check_native_scaling_for_res(pipe, width, height)) {
+			allow = true;
+			break;
+		}
+	}
+	return allow;
+}
+
+/**
+ * dcn32_allow_subvp_high_refresh_rate: Determine if the high refresh rate config will allow subvp
+ *
+ * @dc: Current DC state
+ * @context: New DC state to be programmed
+ * @pipe: Pipe to be considered for use in subvp
+ *
+ * On high refresh rate display configs, we will allow subvp under the following conditions:
+ * 1. Resolution is 3840x2160, 3440x1440, or 2560x1440
+ * 2. Refresh rate is between 120hz - 165hz
+ * 3. No scaling
+ * 4. Freesync is inactive
+ * 5. For single display cases, freesync must be disabled
+ *
+ * Return: True if pipe can be used for subvp, false otherwise
+ */
+bool dcn32_allow_subvp_high_refresh_rate(struct dc *dc, struct dc_state *context, struct pipe_ctx *pipe)
+{
+	bool allow = false;
+	uint32_t refresh_rate = 0;
+	uint32_t subvp_min_refresh = subvp_high_refresh_list.min_refresh;
+	uint32_t subvp_max_refresh = subvp_high_refresh_list.max_refresh;
+	uint32_t min_refresh = subvp_max_refresh;
+	uint32_t i;
+
+	/* Only allow SubVP on high refresh displays if all connected displays
+	 * are considered "high refresh" (i.e. >= 120hz). We do not want to
+	 * allow combinations such as 120hz (SubVP) + 60hz (SubVP).
+	 */
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i];
+
+		if (!pipe_ctx->stream)
+			continue;
+		refresh_rate = (pipe_ctx->stream->timing.pix_clk_100hz * 100 +
+				pipe_ctx->stream->timing.v_total * pipe_ctx->stream->timing.h_total - 1)
+						/ (double)(pipe_ctx->stream->timing.v_total * pipe_ctx->stream->timing.h_total);
+
+		if (refresh_rate < min_refresh)
+			min_refresh = refresh_rate;
+	}
+
+	if (!dc->debug.disable_subvp_high_refresh && min_refresh >= subvp_min_refresh && pipe->stream &&
+			pipe->plane_state && !(pipe->stream->vrr_active_variable || pipe->stream->vrr_active_fixed)) {
+		refresh_rate = (pipe->stream->timing.pix_clk_100hz * 100 +
+						pipe->stream->timing.v_total * pipe->stream->timing.h_total - 1)
+						/ (double)(pipe->stream->timing.v_total * pipe->stream->timing.h_total);
+		if (refresh_rate >= subvp_min_refresh && refresh_rate <= subvp_max_refresh) {
+			for (i = 0; i < SUBVP_HIGH_REFRESH_LIST_LEN; i++) {
+				uint32_t width = subvp_high_refresh_list.res[i].width;
+				uint32_t height = subvp_high_refresh_list.res[i].height;
+
+				if (dcn32_check_native_scaling_for_res(pipe, width, height)) {
+					if ((context->stream_count == 1 && !pipe->stream->allow_freesync) || context->stream_count > 1) {
+						allow = true;
+						break;
+					}
+				}
+			}
+		}
+	}
+	return allow;
+}
+
+/**
+ * dcn32_determine_max_vratio_prefetch: Determine max Vratio for prefetch by driver policy
+ *
+ * @dc: Current DC state
+ * @context: New DC state to be programmed
+ *
+ * Return: Max vratio for prefetch
+ */
+double dcn32_determine_max_vratio_prefetch(struct dc *dc, struct dc_state *context)
+{
+	double max_vratio_pre = __DML_MAX_BW_RATIO_PRE__; // Default value is 4
+	int i;
+
+	/* For single display MPO configs, allow the max vratio to be 8
+	 * if any plane is YUV420 format
+	 */
+	if (context->stream_count == 1 && context->stream_status[0].plane_count > 1) {
+		for (i = 0; i < context->stream_status[0].plane_count; i++) {
+			if (context->stream_status[0].plane_states[i]->format == SURFACE_PIXEL_FORMAT_VIDEO_420_YCbCr ||
+					context->stream_status[0].plane_states[i]->format == SURFACE_PIXEL_FORMAT_VIDEO_420_YCrCb) {
+				max_vratio_pre = __DML_MAX_VRATIO_PRE__;
+			}
+		}
+	}
+	return max_vratio_pre;
+}
+
+/**
+ * dcn32_assign_fpo_vactive_candidate - Assign the FPO stream candidate for FPO + VActive case
+ *
+ * This function chooses the FPO candidate stream for FPO + VActive cases (2 stream config).
+ * For FPO + VAtive cases, the assumption is that one display has ActiveMargin > 0, and the
+ * other display has ActiveMargin <= 0. This function will choose the pipe/stream that has
+ * ActiveMargin <= 0 to be the FPO stream candidate if found.
+ *
+ *
+ * @dc: current dc state
+ * @context: new dc state
+ * @fpo_candidate_stream: pointer to FPO stream candidate if one is found
+ *
+ * Return: void
+ */
+void dcn32_assign_fpo_vactive_candidate(struct dc *dc, const struct dc_state *context, struct dc_stream_state **fpo_candidate_stream)
+{
+	unsigned int i, pipe_idx;
+	const struct vba_vars_st *vba = &context->bw_ctx.dml.vba;
+
+	for (i = 0, pipe_idx = 0; i < dc->res_pool->pipe_count; i++) {
+		const struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
+
+		/* In DCN32/321, FPO uses per-pipe P-State force.
+		 * If there's no planes, HUBP is power gated and
+		 * therefore programming UCLK_PSTATE_FORCE does
+		 * nothing (P-State will always be asserted naturally
+		 * on a pipe that has HUBP power gated. Therefore we
+		 * only want to enable FPO if the FPO pipe has both
+		 * a stream and a plane.
+		 */
+		if (!pipe->stream || !pipe->plane_state)
+			continue;
+
+		if (vba->ActiveDRAMClockChangeLatencyMarginPerState[vba->VoltageLevel][vba->maxMpcComb][vba->pipe_plane[pipe_idx]] <= 0) {
+			*fpo_candidate_stream = pipe->stream;
+			break;
+		}
+		pipe_idx++;
+	}
+}
+
+/**
+ * dcn32_find_vactive_pipe - Determines if the config has a pipe that can switch in VACTIVE
+ *
+ * @dc: current dc state
+ * @context: new dc state
+ * @fpo_candidate_stream: candidate stream to be chosen for FPO
+ * @vactive_margin_req_us: The vactive marign required for a vactive pipe to be considered "found"
+ *
+ * Return: True if VACTIVE display is found, false otherwise
+ */
+bool dcn32_find_vactive_pipe(struct dc *dc, const struct dc_state *context, struct dc_stream_state *fpo_candidate_stream, uint32_t vactive_margin_req_us)
+{
+	unsigned int i, pipe_idx;
+	const struct vba_vars_st *vba = &context->bw_ctx.dml.vba;
+	bool vactive_found = true;
+	unsigned int blank_us = 0;
+
+	for (i = 0, pipe_idx = 0; i < dc->res_pool->pipe_count; i++) {
+		const struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
+
+		if (!pipe->stream)
+			continue;
+
+		/* Don't need to check for vactive margin on the FPO candidate stream */
+		if (fpo_candidate_stream && pipe->stream == fpo_candidate_stream) {
+			pipe_idx++;
+			continue;
+		}
+
+		/* Every plane (apart from the ones driven by the FPO pipes) needs to have active margin
+		 * in order for us to have found a valid "vactive" config for FPO + Vactive
+		 */
+		blank_us = ((pipe->stream->timing.v_total - pipe->stream->timing.v_addressable) * pipe->stream->timing.h_total /
+				(double)(pipe->stream->timing.pix_clk_100hz * 100)) * 1000000;
+		if (vba->ActiveDRAMClockChangeLatencyMarginPerState[vba->VoltageLevel][vba->maxMpcComb][vba->pipe_plane[pipe_idx]] < vactive_margin_req_us ||
+				pipe->stream->vrr_active_variable || pipe->stream->vrr_active_fixed || blank_us >= dc->debug.fpo_vactive_max_blank_us) {
+			vactive_found = false;
+			break;
+		}
+		pipe_idx++;
+	}
+	return vactive_found;
+}
+
+void dcn32_set_clock_limits(const struct _vcs_dpi_soc_bounding_box_st *soc_bb)
+{
+	dc_assert_fp_enabled();
+	dcn3_2_soc.clock_limits[0].dcfclk_mhz = 1200.0;
+}
+
+void dcn32_override_min_req_memclk(struct dc *dc, struct dc_state *context)
+{
+	// WA: restrict FPO and SubVP to use first non-strobe mode (DCN32 BW issue)
+	if ((context->bw_ctx.bw.dcn.clk.fw_based_mclk_switching || dcn32_subvp_in_use(dc, context)) &&
+			dc->dml.soc.num_chans <= 8) {
+		int num_mclk_levels = dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_memclk_levels;
+
+		if (context->bw_ctx.dml.vba.DRAMSpeed <= dc->clk_mgr->bw_params->clk_table.entries[0].memclk_mhz * 16 &&
+				num_mclk_levels > 1) {
+			context->bw_ctx.dml.vba.DRAMSpeed = dc->clk_mgr->bw_params->clk_table.entries[1].memclk_mhz * 16;
+			context->bw_ctx.bw.dcn.clk.dramclk_khz = context->bw_ctx.dml.vba.DRAMSpeed * 1000 / 16;
+		}
+	}
+}
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.h b/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.h
new file mode 100644
index 000000000000..273d2bd79d85
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.h
@@ -0,0 +1,80 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __DCN32_FPU_H__
+#define __DCN32_FPU_H__
+
+#include "clk_mgr_internal.h"
+
+void dcn32_build_wm_range_table_fpu(struct clk_mgr_internal *clk_mgr);
+
+void dcn32_helper_populate_phantom_dlg_params(struct dc *dc,
+					      struct dc_state *context,
+					      display_e2e_pipe_params_st *pipes,
+					      int pipe_cnt);
+
+void dcn32_set_phantom_stream_timing(struct dc *dc,
+				     struct dc_state *context,
+				     struct pipe_ctx *ref_pipe,
+				     struct dc_stream_state *phantom_stream,
+				     display_e2e_pipe_params_st *pipes,
+				     unsigned int pipe_cnt,
+				     unsigned int dc_pipe_idx);
+
+bool dcn32_internal_validate_bw(struct dc *dc,
+				struct dc_state *context,
+				display_e2e_pipe_params_st *pipes,
+				int *pipe_cnt_out,
+				int *vlevel_out,
+				enum dc_validate_mode validate_mode);
+
+void dcn32_calculate_wm_and_dlg_fpu(struct dc *dc, struct dc_state *context,
+				display_e2e_pipe_params_st *pipes,
+				int pipe_cnt,
+				int vlevel);
+
+void dcn32_update_bw_bounding_box_fpu(struct dc *dc, struct clk_bw_params *bw_params);
+
+int dcn32_find_dummy_latency_index_for_fw_based_mclk_switch(struct dc *dc,
+							    struct dc_state *context,
+							    display_e2e_pipe_params_st *pipes,
+							    int pipe_cnt,
+							    int vlevel);
+
+void dcn32_patch_dpm_table(struct clk_bw_params *bw_params);
+
+void dcn32_zero_pipe_dcc_fraction(display_e2e_pipe_params_st *pipes,
+				  int pipe_cnt);
+
+void dcn32_assign_fpo_vactive_candidate(struct dc *dc, const struct dc_state *context, struct dc_stream_state **fpo_candidate_stream);
+
+bool dcn32_find_vactive_pipe(struct dc *dc, const struct dc_state *context, struct dc_stream_state *fpo_candidate_stream, uint32_t vactive_margin_req);
+
+void dcn32_override_min_req_memclk(struct dc *dc, struct dc_state *context);
+
+void dcn32_set_clock_limits(const struct _vcs_dpi_soc_bounding_box_st *soc_bb);
+
+#endif
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_32.c b/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_32.c
new file mode 100644
index 000000000000..0782a34689a0
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_32.c
@@ -0,0 +1,3762 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#include "dc.h"
+#include "../display_mode_lib.h"
+#include "display_mode_vba_32.h"
+#include "../dml_inline_defs.h"
+#include "display_mode_vba_util_32.h"
+
+void dml32_recalculate(struct display_mode_lib *mode_lib);
+static void DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation(
+		struct display_mode_lib *mode_lib);
+void dml32_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_lib);
+
+void dml32_recalculate(struct display_mode_lib *mode_lib)
+{
+	ModeSupportAndSystemConfiguration(mode_lib);
+
+	dml32_CalculateMaxDETAndMinCompressedBufferSize(mode_lib->vba.ConfigReturnBufferSizeInKByte,
+			mode_lib->vba.ROBBufferSizeInKByte,
+			DC__NUM_DPP,
+			false, //mode_lib->vba.override_setting.nomDETInKByteOverrideEnable,
+			0, //mode_lib->vba.override_setting.nomDETInKByteOverrideValue,
+
+			/* Output */
+			&mode_lib->vba.MaxTotalDETInKByte, &mode_lib->vba.nomDETInKByte,
+			&mode_lib->vba.MinCompressedBufferSizeInKByte);
+
+	PixelClockAdjustmentForProgressiveToInterlaceUnit(mode_lib);
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: Calling DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation\n", __func__);
+#endif
+	DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation(mode_lib);
+}
+
+static void DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation(
+		struct display_mode_lib *mode_lib)
+{
+	struct vba_vars_st *v = &mode_lib->vba;
+	unsigned int j, k;
+	bool ImmediateFlipRequirementFinal;
+	int iteration;
+	double MaxTotalRDBandwidth;
+	unsigned int NextPrefetchMode;
+	double MaxTotalRDBandwidthNoUrgentBurst = 0.0;
+	bool DestinationLineTimesForPrefetchLessThan2 = false;
+	bool VRatioPrefetchMoreThanMax = false;
+	double TWait;
+	double TotalWRBandwidth = 0;
+	double WRBandwidth = 0;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: --- START ---\n", __func__);
+	dml_print("DML::%s: mode_lib->vba.PrefetchMode = %d\n", __func__, mode_lib->vba.PrefetchMode);
+	dml_print("DML::%s: mode_lib->vba.ImmediateFlipSupport = %d\n", __func__, mode_lib->vba.ImmediateFlipSupport);
+	dml_print("DML::%s: mode_lib->vba.VoltageLevel = %d\n", __func__, mode_lib->vba.VoltageLevel);
+#endif
+
+	v->WritebackDISPCLK = 0.0;
+	v->GlobalDPPCLK = 0.0;
+
+	// DISPCLK and DPPCLK Calculation
+	for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+		if (mode_lib->vba.WritebackEnable[k]) {
+			v->WritebackDISPCLK = dml_max(v->WritebackDISPCLK,
+					dml32_CalculateWriteBackDISPCLK(
+							mode_lib->vba.WritebackPixelFormat[k],
+							mode_lib->vba.PixelClock[k], mode_lib->vba.WritebackHRatio[k],
+							mode_lib->vba.WritebackVRatio[k],
+							mode_lib->vba.WritebackHTaps[k],
+							mode_lib->vba.WritebackVTaps[k],
+							mode_lib->vba.WritebackSourceWidth[k],
+							mode_lib->vba.WritebackDestinationWidth[k],
+							mode_lib->vba.HTotal[k], mode_lib->vba.WritebackLineBufferSize,
+							mode_lib->vba.DISPCLKDPPCLKVCOSpeed));
+		}
+	}
+
+	v->DISPCLK_calculated = v->WritebackDISPCLK;
+
+	for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+		if (mode_lib->vba.BlendingAndTiming[k] == k) {
+			v->DISPCLK_calculated = dml_max(v->DISPCLK_calculated,
+					dml32_CalculateRequiredDispclk(
+							mode_lib->vba.ODMCombineEnabled[k],
+							mode_lib->vba.PixelClock[k],
+							mode_lib->vba.DISPCLKDPPCLKDSCCLKDownSpreading,
+							mode_lib->vba.DISPCLKRampingMargin,
+							mode_lib->vba.DISPCLKDPPCLKVCOSpeed,
+							mode_lib->vba.MaxDppclk[v->soc.num_states - 1]));
+		}
+	}
+
+	for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+		dml32_CalculateSinglePipeDPPCLKAndSCLThroughput(mode_lib->vba.HRatio[k],
+				mode_lib->vba.HRatioChroma[k],
+				mode_lib->vba.VRatio[k],
+				mode_lib->vba.VRatioChroma[k],
+				mode_lib->vba.MaxDCHUBToPSCLThroughput,
+				mode_lib->vba.MaxPSCLToLBThroughput,
+				mode_lib->vba.PixelClock[k],
+				mode_lib->vba.SourcePixelFormat[k],
+				mode_lib->vba.htaps[k],
+				mode_lib->vba.HTAPsChroma[k],
+				mode_lib->vba.vtaps[k],
+				mode_lib->vba.VTAPsChroma[k],
+
+				/* Output */
+				&v->PSCL_THROUGHPUT_LUMA[k], &v->PSCL_THROUGHPUT_CHROMA[k],
+				&v->DPPCLKUsingSingleDPP[k]);
+	}
+
+	dml32_CalculateDPPCLK(mode_lib->vba.NumberOfActiveSurfaces, mode_lib->vba.DISPCLKDPPCLKDSCCLKDownSpreading,
+			mode_lib->vba.DISPCLKDPPCLKVCOSpeed, v->DPPCLKUsingSingleDPP, mode_lib->vba.DPPPerPlane,
+			/* Output */
+			&v->GlobalDPPCLK, v->DPPCLK);
+
+	for (k = 0; k < v->NumberOfActiveSurfaces; ++k) {
+		v->DPPCLK_calculated[k] = v->DPPCLK[k];
+	}
+
+	for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+		dml32_CalculateBytePerPixelAndBlockSizes(
+				mode_lib->vba.SourcePixelFormat[k],
+				mode_lib->vba.SurfaceTiling[k],
+
+				/* Output */
+				&v->BytePerPixelY[k],
+				&v->BytePerPixelC[k],
+				&v->BytePerPixelDETY[k],
+				&v->BytePerPixelDETC[k],
+				&v->BlockHeight256BytesY[k],
+				&v->BlockHeight256BytesC[k],
+				&v->BlockWidth256BytesY[k],
+				&v->BlockWidth256BytesC[k],
+				&v->BlockHeightY[k],
+				&v->BlockHeightC[k],
+				&v->BlockWidthY[k],
+				&v->BlockWidthC[k]);
+	}
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: %d\n", __func__, __LINE__);
+#endif
+	dml32_CalculateSwathWidth(
+			false,  // ForceSingleDPP
+			mode_lib->vba.NumberOfActiveSurfaces,
+			mode_lib->vba.SourcePixelFormat,
+			mode_lib->vba.SourceRotation,
+			mode_lib->vba.ViewportStationary,
+			mode_lib->vba.ViewportWidth,
+			mode_lib->vba.ViewportHeight,
+			mode_lib->vba.ViewportXStartY,
+			mode_lib->vba.ViewportYStartY,
+			mode_lib->vba.ViewportXStartC,
+			mode_lib->vba.ViewportYStartC,
+			mode_lib->vba.SurfaceWidthY,
+			mode_lib->vba.SurfaceWidthC,
+			mode_lib->vba.SurfaceHeightY,
+			mode_lib->vba.SurfaceHeightC,
+			mode_lib->vba.ODMCombineEnabled,
+			v->BytePerPixelY,
+			v->BytePerPixelC,
+			v->BlockHeight256BytesY,
+			v->BlockHeight256BytesC,
+			v->BlockWidth256BytesY,
+			v->BlockWidth256BytesC,
+			mode_lib->vba.BlendingAndTiming,
+			mode_lib->vba.HActive,
+			mode_lib->vba.HRatio,
+			mode_lib->vba.DPPPerPlane,
+
+			/* Output */
+			v->SwathWidthSingleDPPY, v->SwathWidthSingleDPPC, v->SwathWidthY, v->SwathWidthC,
+			v->dummy_vars
+				.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation
+				.dummy_integer_array[0], // Integer             MaximumSwathHeightY[]
+			v->dummy_vars
+				.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation
+				.dummy_integer_array[1], // Integer             MaximumSwathHeightC[]
+			v->swath_width_luma_ub, v->swath_width_chroma_ub);
+
+	for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+		v->ReadBandwidthSurfaceLuma[k] = v->SwathWidthSingleDPPY[k] * v->BytePerPixelY[k]
+				/ (mode_lib->vba.HTotal[k] / mode_lib->vba.PixelClock[k]) * mode_lib->vba.VRatio[k];
+		v->ReadBandwidthSurfaceChroma[k] = v->SwathWidthSingleDPPC[k] * v->BytePerPixelC[k]
+				/ (mode_lib->vba.HTotal[k] / mode_lib->vba.PixelClock[k])
+				* mode_lib->vba.VRatioChroma[k];
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: ReadBandwidthSurfaceLuma[%i] = %fBps\n",
+				__func__, k, v->ReadBandwidthSurfaceLuma[k]);
+		dml_print("DML::%s: ReadBandwidthSurfaceChroma[%i] = %fBps\n",
+				__func__, k, v->ReadBandwidthSurfaceChroma[k]);
+#endif
+	}
+
+	{
+		// VBA_DELTA
+		// Calculate DET size, swath height
+		dml32_CalculateSwathAndDETConfiguration(
+				mode_lib->vba.DETSizeOverride,
+				mode_lib->vba.UsesMALLForPStateChange,
+				mode_lib->vba.ConfigReturnBufferSizeInKByte,
+				mode_lib->vba.MaxTotalDETInKByte,
+				mode_lib->vba.MinCompressedBufferSizeInKByte,
+				false, /* ForceSingleDPP */
+				mode_lib->vba.NumberOfActiveSurfaces,
+				mode_lib->vba.nomDETInKByte,
+				mode_lib->vba.UseUnboundedRequesting,
+				mode_lib->vba.DisableUnboundRequestIfCompBufReservedSpaceNeedAdjustment,
+				mode_lib->vba.ip.pixel_chunk_size_kbytes,
+				mode_lib->vba.ip.rob_buffer_size_kbytes,
+				mode_lib->vba.CompressedBufferSegmentSizeInkByteFinal,
+				v->dummy_vars
+					.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation
+					.dummy_output_encoder_array, /* output_encoder_class Output[] */
+				v->ReadBandwidthSurfaceLuma,
+				v->ReadBandwidthSurfaceChroma,
+				v->dummy_vars
+					.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation
+					.dummy_single_array[0], /* Single MaximumSwathWidthLuma[] */
+				v->dummy_vars
+					.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation
+					.dummy_single_array[1], /* Single MaximumSwathWidthChroma[] */
+				mode_lib->vba.SourceRotation,
+				mode_lib->vba.ViewportStationary,
+				mode_lib->vba.SourcePixelFormat,
+				mode_lib->vba.SurfaceTiling,
+				mode_lib->vba.ViewportWidth,
+				mode_lib->vba.ViewportHeight,
+				mode_lib->vba.ViewportXStartY,
+				mode_lib->vba.ViewportYStartY,
+				mode_lib->vba.ViewportXStartC,
+				mode_lib->vba.ViewportYStartC,
+				mode_lib->vba.SurfaceWidthY,
+				mode_lib->vba.SurfaceWidthC,
+				mode_lib->vba.SurfaceHeightY,
+				mode_lib->vba.SurfaceHeightC,
+				v->BlockHeight256BytesY,
+				v->BlockHeight256BytesC,
+				v->BlockWidth256BytesY,
+				v->BlockWidth256BytesC,
+				mode_lib->vba.ODMCombineEnabled,
+				mode_lib->vba.BlendingAndTiming,
+				v->BytePerPixelY,
+				v->BytePerPixelC,
+				v->BytePerPixelDETY,
+				v->BytePerPixelDETC,
+				mode_lib->vba.HActive,
+				mode_lib->vba.HRatio,
+				mode_lib->vba.HRatioChroma,
+				mode_lib->vba.DPPPerPlane,
+
+				/* Output */
+				v->dummy_vars
+					.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation
+					.dummy_long_array[0], /* Long swath_width_luma_ub[] */
+				v->dummy_vars
+					.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation
+					.dummy_long_array[1], /* Long swath_width_chroma_ub[] */
+				v->dummy_vars
+					.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation
+					.dummy_double_array[0], /* Long SwathWidth[] */
+				v->dummy_vars
+					.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation
+					.dummy_double_array[1], /* Long SwathWidthChroma[] */
+				mode_lib->vba.SwathHeightY,
+				mode_lib->vba.SwathHeightC,
+				mode_lib->vba.DETBufferSizeInKByte,
+				mode_lib->vba.DETBufferSizeY,
+				mode_lib->vba.DETBufferSizeC,
+				&v->UnboundedRequestEnabled,
+				&v->CompressedBufferSizeInkByte,
+				&v->CompBufReservedSpaceKBytes,
+				&v->dummy_vars
+					.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation
+					.dummy_boolean,       /* bool *CompBufReservedSpaceNeedAjustment */
+				v->dummy_vars
+					.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation
+					.dummy_boolean_array, /* bool ViewportSizeSupportPerSurface[] */
+				&v->dummy_vars
+					 .DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation
+					 .dummy_boolean); /* bool *ViewportSizeSupport */
+	}
+
+	v->CompBufReservedSpaceZs     = v->CompBufReservedSpaceKBytes * 1024.0 / 256.0;
+	v->CompBufReservedSpace64B    = v->CompBufReservedSpaceKBytes * 1024.0 / 64.0;
+
+	// DCFCLK Deep Sleep
+	dml32_CalculateDCFCLKDeepSleep(
+			mode_lib->vba.NumberOfActiveSurfaces,
+			v->BytePerPixelY,
+			v->BytePerPixelC,
+			mode_lib->vba.VRatio,
+			mode_lib->vba.VRatioChroma,
+			v->SwathWidthY,
+			v->SwathWidthC,
+			mode_lib->vba.DPPPerPlane,
+			mode_lib->vba.HRatio,
+			mode_lib->vba.HRatioChroma,
+			mode_lib->vba.PixelClock,
+			v->PSCL_THROUGHPUT_LUMA,
+			v->PSCL_THROUGHPUT_CHROMA,
+			mode_lib->vba.DPPCLK,
+			v->ReadBandwidthSurfaceLuma,
+			v->ReadBandwidthSurfaceChroma,
+			mode_lib->vba.ReturnBusWidth,
+
+			/* Output */
+			&v->DCFCLKDeepSleep);
+
+	// DSCCLK
+	for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+		if ((mode_lib->vba.BlendingAndTiming[k] != k) || !mode_lib->vba.DSCEnabled[k]) {
+			v->DSCCLK_calculated[k] = 0.0;
+		} else {
+			if (mode_lib->vba.OutputFormat[k] == dm_420)
+				mode_lib->vba.DSCFormatFactor = 2;
+			else if (mode_lib->vba.OutputFormat[k] == dm_444)
+				mode_lib->vba.DSCFormatFactor = 1;
+			else if (mode_lib->vba.OutputFormat[k] == dm_n422)
+				mode_lib->vba.DSCFormatFactor = 2;
+			else
+				mode_lib->vba.DSCFormatFactor = 1;
+			if (mode_lib->vba.ODMCombineEnabled[k] == dm_odm_combine_mode_4to1)
+				v->DSCCLK_calculated[k] = mode_lib->vba.PixelClockBackEnd[k] / 12
+						/ mode_lib->vba.DSCFormatFactor
+						/ (1 - mode_lib->vba.DISPCLKDPPCLKDSCCLKDownSpreading / 100);
+			else if (mode_lib->vba.ODMCombineEnabled[k] == dm_odm_combine_mode_2to1)
+				v->DSCCLK_calculated[k] = mode_lib->vba.PixelClockBackEnd[k] / 6
+						/ mode_lib->vba.DSCFormatFactor
+						/ (1 - mode_lib->vba.DISPCLKDPPCLKDSCCLKDownSpreading / 100);
+			else
+				v->DSCCLK_calculated[k] = mode_lib->vba.PixelClockBackEnd[k] / 3
+						/ mode_lib->vba.DSCFormatFactor
+						/ (1 - mode_lib->vba.DISPCLKDPPCLKDSCCLKDownSpreading / 100);
+		}
+	}
+
+	// DSC Delay
+	for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+		v->DSCDelay[k] = dml32_DSCDelayRequirement(mode_lib->vba.DSCEnabled[k],
+				mode_lib->vba.ODMCombineEnabled[k], mode_lib->vba.DSCInputBitPerComponent[k],
+				mode_lib->vba.OutputBppPerState[mode_lib->vba.VoltageLevel][k],
+				mode_lib->vba.HActive[k], mode_lib->vba.HTotal[k],
+				mode_lib->vba.NumberOfDSCSlices[k], mode_lib->vba.OutputFormat[k],
+				mode_lib->vba.Output[k], mode_lib->vba.PixelClock[k],
+				mode_lib->vba.PixelClockBackEnd[k], mode_lib->vba.ip.dsc_delay_factor_wa);
+	}
+
+	for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k)
+		for (j = 0; j < mode_lib->vba.NumberOfActiveSurfaces; ++j) // NumberOfSurfaces
+			if (j != k && mode_lib->vba.BlendingAndTiming[k] == j && mode_lib->vba.DSCEnabled[j])
+				v->DSCDelay[k] = v->DSCDelay[j];
+
+	//Immediate Flip
+	for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+		v->ImmediateFlipSupportedSurface[k] = mode_lib->vba.ImmediateFlipSupport
+				&& (mode_lib->vba.ImmediateFlipRequirement[k] != dm_immediate_flip_not_required);
+	}
+
+	// Prefetch
+	dml32_CalculateSurfaceSizeInMall(
+				mode_lib->vba.NumberOfActiveSurfaces,
+				mode_lib->vba.MALLAllocatedForDCNFinal,
+				mode_lib->vba.UseMALLForStaticScreen,
+				mode_lib->vba.UsesMALLForPStateChange,
+				mode_lib->vba.DCCEnable,
+				mode_lib->vba.ViewportStationary,
+				mode_lib->vba.ViewportXStartY,
+				mode_lib->vba.ViewportYStartY,
+				mode_lib->vba.ViewportXStartC,
+				mode_lib->vba.ViewportYStartC,
+				mode_lib->vba.ViewportWidth,
+				mode_lib->vba.ViewportHeight,
+				v->BytePerPixelY,
+				mode_lib->vba.ViewportWidthChroma,
+				mode_lib->vba.ViewportHeightChroma,
+				v->BytePerPixelC,
+				mode_lib->vba.SurfaceWidthY,
+				mode_lib->vba.SurfaceWidthC,
+				mode_lib->vba.SurfaceHeightY,
+				mode_lib->vba.SurfaceHeightC,
+				v->BlockWidth256BytesY,
+				v->BlockWidth256BytesC,
+				v->BlockHeight256BytesY,
+				v->BlockHeight256BytesC,
+				v->BlockWidthY,
+				v->BlockWidthC,
+				v->BlockHeightY,
+				v->BlockHeightC,
+				mode_lib->vba.DCCMetaPitchY,
+				mode_lib->vba.DCCMetaPitchC,
+
+				/* Output */
+				v->SurfaceSizeInMALL,
+				&v->dummy_vars.
+				DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation
+				.dummy_boolean2); /* Boolean *ExceededMALLSize */
+
+	for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.SurfaceParameters[k].PixelClock = mode_lib->vba.PixelClock[k];
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.SurfaceParameters[k].DPPPerSurface = mode_lib->vba.DPPPerPlane[k];
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.SurfaceParameters[k].SourceRotation = mode_lib->vba.SourceRotation[k];
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.SurfaceParameters[k].ViewportHeight = mode_lib->vba.ViewportHeight[k];
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.SurfaceParameters[k].ViewportHeightChroma = mode_lib->vba.ViewportHeightChroma[k];
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.SurfaceParameters[k].BlockWidth256BytesY = v->BlockWidth256BytesY[k];
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.SurfaceParameters[k].BlockHeight256BytesY = v->BlockHeight256BytesY[k];
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.SurfaceParameters[k].BlockWidth256BytesC = v->BlockWidth256BytesC[k];
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.SurfaceParameters[k].BlockHeight256BytesC = v->BlockHeight256BytesC[k];
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.SurfaceParameters[k].BlockWidthY = v->BlockWidthY[k];
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.SurfaceParameters[k].BlockHeightY = v->BlockHeightY[k];
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.SurfaceParameters[k].BlockWidthC = v->BlockWidthC[k];
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.SurfaceParameters[k].BlockHeightC = v->BlockHeightC[k];
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.SurfaceParameters[k].InterlaceEnable = mode_lib->vba.Interlace[k];
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.SurfaceParameters[k].HTotal = mode_lib->vba.HTotal[k];
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.SurfaceParameters[k].DCCEnable = mode_lib->vba.DCCEnable[k];
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.SurfaceParameters[k].SourcePixelFormat = mode_lib->vba.SourcePixelFormat[k];
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.SurfaceParameters[k].SurfaceTiling = mode_lib->vba.SurfaceTiling[k];
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.SurfaceParameters[k].BytePerPixelY = v->BytePerPixelY[k];
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.SurfaceParameters[k].BytePerPixelC = v->BytePerPixelC[k];
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.SurfaceParameters[k].ProgressiveToInterlaceUnitInOPP = mode_lib->vba.ProgressiveToInterlaceUnitInOPP;
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.SurfaceParameters[k].VRatio = mode_lib->vba.VRatio[k];
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.SurfaceParameters[k].VRatioChroma = mode_lib->vba.VRatioChroma[k];
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.SurfaceParameters[k].VTaps = mode_lib->vba.vtaps[k];
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.SurfaceParameters[k].VTapsChroma = mode_lib->vba.VTAPsChroma[k];
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.SurfaceParameters[k].PitchY = mode_lib->vba.PitchY[k];
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.SurfaceParameters[k].DCCMetaPitchY = mode_lib->vba.DCCMetaPitchY[k];
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.SurfaceParameters[k].PitchC = mode_lib->vba.PitchC[k];
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.SurfaceParameters[k].DCCMetaPitchC = mode_lib->vba.DCCMetaPitchC[k];
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.SurfaceParameters[k].ViewportStationary = mode_lib->vba.ViewportStationary[k];
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.SurfaceParameters[k].ViewportXStart = mode_lib->vba.ViewportXStartY[k];
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.SurfaceParameters[k].ViewportYStart = mode_lib->vba.ViewportYStartY[k];
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.SurfaceParameters[k].ViewportXStartC = mode_lib->vba.ViewportXStartC[k];
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.SurfaceParameters[k].ViewportYStartC = mode_lib->vba.ViewportYStartC[k];
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.SurfaceParameters[k].FORCE_ONE_ROW_FOR_FRAME = mode_lib->vba.ForceOneRowForFrame[k];
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.SurfaceParameters[k].SwathHeightY = mode_lib->vba.SwathHeightY[k];
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.SurfaceParameters[k].SwathHeightC = mode_lib->vba.SwathHeightC[k];
+	}
+
+	{
+
+		dml32_CalculateVMRowAndSwath(
+				mode_lib->vba.NumberOfActiveSurfaces,
+				v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.SurfaceParameters,
+				v->SurfaceSizeInMALL,
+				mode_lib->vba.PTEBufferSizeInRequestsLuma,
+				mode_lib->vba.PTEBufferSizeInRequestsChroma,
+				mode_lib->vba.DCCMetaBufferSizeBytes,
+				mode_lib->vba.UseMALLForStaticScreen,
+				mode_lib->vba.UsesMALLForPStateChange,
+				mode_lib->vba.MALLAllocatedForDCNFinal,
+				v->SwathWidthY,
+				v->SwathWidthC,
+				mode_lib->vba.GPUVMEnable,
+				mode_lib->vba.HostVMEnable,
+				mode_lib->vba.HostVMMaxNonCachedPageTableLevels,
+				mode_lib->vba.GPUVMMaxPageTableLevels,
+				mode_lib->vba.GPUVMMinPageSizeKBytes,
+				mode_lib->vba.HostVMMinPageSize,
+
+				/* Output */
+				v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.dummy_boolean_array2[0],  // Boolean PTEBufferSizeNotExceeded[]
+				v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.dummy_boolean_array2[1],  // Boolean DCCMetaBufferSizeNotExceeded[]
+				v->dpte_row_width_luma_ub,
+				v->dpte_row_width_chroma_ub,
+				v->dpte_row_height,
+				v->dpte_row_height_chroma,
+				v->dpte_row_height_linear,
+				v->dpte_row_height_linear_chroma,
+				v->meta_req_width,
+				v->meta_req_width_chroma,
+				v->meta_req_height,
+				v->meta_req_height_chroma,
+				v->meta_row_width,
+				v->meta_row_width_chroma,
+				v->meta_row_height,
+				v->meta_row_height_chroma,
+				v->vm_group_bytes,
+				v->dpte_group_bytes,
+				v->PixelPTEReqWidthY,
+				v->PixelPTEReqHeightY,
+				v->PTERequestSizeY,
+				v->PixelPTEReqWidthC,
+				v->PixelPTEReqHeightC,
+				v->PTERequestSizeC,
+				v->dpde0_bytes_per_frame_ub_l,
+				v->meta_pte_bytes_per_frame_ub_l,
+				v->dpde0_bytes_per_frame_ub_c,
+				v->meta_pte_bytes_per_frame_ub_c,
+				v->PrefetchSourceLinesY,
+				v->PrefetchSourceLinesC,
+				v->VInitPreFillY, v->VInitPreFillC,
+				v->MaxNumSwathY,
+				v->MaxNumSwathC,
+				v->meta_row_bw,
+				v->dpte_row_bw,
+				v->PixelPTEBytesPerRow,
+				v->PDEAndMetaPTEBytesFrame,
+				v->MetaRowByte,
+				v->Use_One_Row_For_Frame,
+				v->Use_One_Row_For_Frame_Flip,
+				v->UsesMALLForStaticScreen,
+				v->PTE_BUFFER_MODE,
+				v->BIGK_FRAGMENT_SIZE);
+	}
+
+
+	v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.ReorderBytes = mode_lib->vba.NumberOfChannels
+			* dml_max3(mode_lib->vba.UrgentOutOfOrderReturnPerChannelPixelDataOnly,
+					mode_lib->vba.UrgentOutOfOrderReturnPerChannelPixelMixedWithVMData,
+					mode_lib->vba.UrgentOutOfOrderReturnPerChannelVMDataOnly);
+
+	v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.VMDataOnlyReturnBW = dml32_get_return_bw_mbps_vm_only(
+			&mode_lib->vba.soc,
+			mode_lib->vba.VoltageLevel,
+			mode_lib->vba.DCFCLK,
+			mode_lib->vba.FabricClock,
+			mode_lib->vba.DRAMSpeed);
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: mode_lib->vba.ReturnBusWidth = %f\n", __func__, mode_lib->vba.ReturnBusWidth);
+	dml_print("DML::%s: mode_lib->vba.DCFCLK = %f\n", __func__, mode_lib->vba.DCFCLK);
+	dml_print("DML::%s: mode_lib->vba.FabricClock = %f\n", __func__, mode_lib->vba.FabricClock);
+	dml_print("DML::%s: mode_lib->vba.FabricDatapathToDCNDataReturn = %f\n", __func__,
+			mode_lib->vba.FabricDatapathToDCNDataReturn);
+	dml_print("DML::%s: mode_lib->vba.PercentOfIdealSDPPortBWReceivedAfterUrgLatency = %f\n",
+			__func__, mode_lib->vba.PercentOfIdealSDPPortBWReceivedAfterUrgLatency);
+	dml_print("DML::%s: mode_lib->vba.DRAMSpeed = %f\n", __func__, mode_lib->vba.DRAMSpeed);
+	dml_print("DML::%s: mode_lib->vba.NumberOfChannels = %f\n", __func__, mode_lib->vba.NumberOfChannels);
+	dml_print("DML::%s: mode_lib->vba.DRAMChannelWidth = %f\n", __func__, mode_lib->vba.DRAMChannelWidth);
+	dml_print("DML::%s: mode_lib->vba.PercentOfIdealDRAMBWReceivedAfterUrgLatencyVMDataOnly = %f\n",
+			__func__, mode_lib->vba.PercentOfIdealDRAMBWReceivedAfterUrgLatencyVMDataOnly);
+	dml_print("DML::%s: VMDataOnlyReturnBW = %f\n", __func__, VMDataOnlyReturnBW);
+	dml_print("DML::%s: ReturnBW = %f\n", __func__, mode_lib->vba.ReturnBW);
+#endif
+
+	v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.HostVMInefficiencyFactor = 1.0;
+
+	if (mode_lib->vba.GPUVMEnable && mode_lib->vba.HostVMEnable)
+		v->dummy_vars
+			.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation
+			.HostVMInefficiencyFactor =
+			mode_lib->vba.ReturnBW / v->dummy_vars
+				.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation
+				.VMDataOnlyReturnBW;
+
+	mode_lib->vba.TotalDCCActiveDPP = 0;
+	mode_lib->vba.TotalActiveDPP = 0;
+	for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+		mode_lib->vba.TotalActiveDPP = mode_lib->vba.TotalActiveDPP + mode_lib->vba.DPPPerPlane[k];
+		if (mode_lib->vba.DCCEnable[k])
+			mode_lib->vba.TotalDCCActiveDPP = mode_lib->vba.TotalDCCActiveDPP
+					+ mode_lib->vba.DPPPerPlane[k];
+	}
+
+	v->UrgentExtraLatency = dml32_CalculateExtraLatency(
+			mode_lib->vba.RoundTripPingLatencyCycles,
+			v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.ReorderBytes,
+			mode_lib->vba.DCFCLK,
+			mode_lib->vba.TotalActiveDPP,
+			mode_lib->vba.PixelChunkSizeInKByte,
+			mode_lib->vba.TotalDCCActiveDPP,
+			mode_lib->vba.MetaChunkSize,
+			mode_lib->vba.ReturnBW,
+			mode_lib->vba.GPUVMEnable,
+			mode_lib->vba.HostVMEnable,
+			mode_lib->vba.NumberOfActiveSurfaces,
+			mode_lib->vba.DPPPerPlane,
+			v->dpte_group_bytes,
+			v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.HostVMInefficiencyFactor,
+			mode_lib->vba.HostVMMinPageSize,
+			mode_lib->vba.HostVMMaxNonCachedPageTableLevels);
+
+	mode_lib->vba.TCalc = 24.0 / v->DCFCLKDeepSleep;
+
+	for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+		if (mode_lib->vba.BlendingAndTiming[k] == k) {
+			if (mode_lib->vba.WritebackEnable[k] == true) {
+				v->WritebackDelay[mode_lib->vba.VoltageLevel][k] = mode_lib->vba.WritebackLatency
+						+ dml32_CalculateWriteBackDelay(
+								mode_lib->vba.WritebackPixelFormat[k],
+								mode_lib->vba.WritebackHRatio[k],
+								mode_lib->vba.WritebackVRatio[k],
+								mode_lib->vba.WritebackVTaps[k],
+								mode_lib->vba.WritebackDestinationWidth[k],
+								mode_lib->vba.WritebackDestinationHeight[k],
+								mode_lib->vba.WritebackSourceHeight[k],
+								mode_lib->vba.HTotal[k]) / mode_lib->vba.DISPCLK;
+			} else
+				v->WritebackDelay[mode_lib->vba.VoltageLevel][k] = 0;
+			for (j = 0; j < mode_lib->vba.NumberOfActiveSurfaces; ++j) {
+				if (mode_lib->vba.BlendingAndTiming[j] == k &&
+					mode_lib->vba.WritebackEnable[j] == true) {
+					v->WritebackDelay[mode_lib->vba.VoltageLevel][k] =
+						dml_max(v->WritebackDelay[mode_lib->vba.VoltageLevel][k],
+						mode_lib->vba.WritebackLatency +
+						dml32_CalculateWriteBackDelay(
+								mode_lib->vba.WritebackPixelFormat[j],
+								mode_lib->vba.WritebackHRatio[j],
+								mode_lib->vba.WritebackVRatio[j],
+								mode_lib->vba.WritebackVTaps[j],
+								mode_lib->vba.WritebackDestinationWidth[j],
+								mode_lib->vba.WritebackDestinationHeight[j],
+								mode_lib->vba.WritebackSourceHeight[j],
+								mode_lib->vba.HTotal[k]) / mode_lib->vba.DISPCLK);
+				}
+			}
+		}
+	}
+
+	for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k)
+		for (j = 0; j < mode_lib->vba.NumberOfActiveSurfaces; ++j)
+			if (mode_lib->vba.BlendingAndTiming[k] == j)
+				v->WritebackDelay[mode_lib->vba.VoltageLevel][k] =
+						v->WritebackDelay[mode_lib->vba.VoltageLevel][j];
+
+	v->UrgentLatency = dml32_CalculateUrgentLatency(mode_lib->vba.UrgentLatencyPixelDataOnly,
+			mode_lib->vba.UrgentLatencyPixelMixedWithVMData,
+			mode_lib->vba.UrgentLatencyVMDataOnly,
+			mode_lib->vba.DoUrgentLatencyAdjustment,
+			mode_lib->vba.UrgentLatencyAdjustmentFabricClockComponent,
+			mode_lib->vba.UrgentLatencyAdjustmentFabricClockReference,
+			mode_lib->vba.FabricClock);
+
+	for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+		dml32_CalculateUrgentBurstFactor(mode_lib->vba.UsesMALLForPStateChange[k],
+				v->swath_width_luma_ub[k],
+				v->swath_width_chroma_ub[k],
+				mode_lib->vba.SwathHeightY[k],
+				mode_lib->vba.SwathHeightC[k],
+				mode_lib->vba.HTotal[k] / mode_lib->vba.PixelClock[k],
+				v->UrgentLatency,
+				mode_lib->vba.CursorBufferSize,
+				mode_lib->vba.CursorWidth[k][0],
+				mode_lib->vba.CursorBPP[k][0],
+				mode_lib->vba.VRatio[k],
+				mode_lib->vba.VRatioChroma[k],
+				v->BytePerPixelDETY[k],
+				v->BytePerPixelDETC[k],
+				mode_lib->vba.DETBufferSizeY[k],
+				mode_lib->vba.DETBufferSizeC[k],
+
+				/* output */
+				&v->UrgBurstFactorCursor[k],
+				&v->UrgBurstFactorLuma[k],
+				&v->UrgBurstFactorChroma[k],
+				&v->NoUrgentLatencyHiding[k]);
+
+		v->cursor_bw[k] = mode_lib->vba.NumberOfCursors[k] * mode_lib->vba.CursorWidth[k][0] * mode_lib->vba.CursorBPP[k][0] / 8 / (mode_lib->vba.HTotal[k] / mode_lib->vba.PixelClock[k]) * mode_lib->vba.VRatio[k];
+	}
+
+	v->NotEnoughDETSwathFillLatencyHiding = dml32_CalculateDETSwathFillLatencyHiding(
+						mode_lib->vba.NumberOfActiveSurfaces,
+						mode_lib->vba.ReturnBW,
+						v->UrgentLatency,
+						mode_lib->vba.SwathHeightY,
+						mode_lib->vba.SwathHeightC,
+						v->swath_width_luma_ub,
+						v->swath_width_chroma_ub,
+						v->BytePerPixelDETY,
+						v->BytePerPixelDETC,
+						mode_lib->vba.DETBufferSizeY,
+						mode_lib->vba.DETBufferSizeC,
+						mode_lib->vba.DPPPerPlane,
+						mode_lib->vba.HTotal,
+						mode_lib->vba.PixelClock,
+						mode_lib->vba.VRatio,
+						mode_lib->vba.VRatioChroma,
+						mode_lib->vba.UsesMALLForPStateChange,
+						mode_lib->vba.UseUnboundedRequesting);
+
+	for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+		v->MaxVStartupLines[k] = ((mode_lib->vba.Interlace[k] &&
+				!mode_lib->vba.ProgressiveToInterlaceUnitInOPP) ?
+				dml_floor((mode_lib->vba.VTotal[k] - mode_lib->vba.VActive[k]) / 2.0, 1.0) :
+				mode_lib->vba.VTotal[k] - mode_lib->vba.VActive[k]) - dml_max(1.0,
+				dml_ceil((double) v->WritebackDelay[mode_lib->vba.VoltageLevel][k]
+				/ (mode_lib->vba.HTotal[k] / mode_lib->vba.PixelClock[k]), 1));
+
+		// Clamp to max OTG vstartup register limit
+		if (v->MaxVStartupLines[k] > 1023)
+			v->MaxVStartupLines[k] = 1023;
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%d MaxVStartupLines = %d\n", __func__, k, v->MaxVStartupLines[k]);
+		dml_print("DML::%s: k=%d VoltageLevel = %d\n", __func__, k, mode_lib->vba.VoltageLevel);
+		dml_print("DML::%s: k=%d WritebackDelay = %f\n", __func__,
+				k, v->WritebackDelay[mode_lib->vba.VoltageLevel][k]);
+#endif
+	}
+
+	v->MaximumMaxVStartupLines = 0;
+	for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k)
+		v->MaximumMaxVStartupLines = dml_max(v->MaximumMaxVStartupLines, v->MaxVStartupLines[k]);
+
+	ImmediateFlipRequirementFinal = false;
+
+	for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+		ImmediateFlipRequirementFinal = ImmediateFlipRequirementFinal
+				|| (mode_lib->vba.ImmediateFlipRequirement[k] == dm_immediate_flip_required);
+	}
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: ImmediateFlipRequirementFinal = %d\n", __func__, ImmediateFlipRequirementFinal);
+#endif
+	// ModeProgramming will not repeat the schedule calculation using different prefetch mode,
+	//it is just calcualated once with given prefetch mode
+	dml32_CalculateMinAndMaxPrefetchMode(
+			mode_lib->vba.AllowForPStateChangeOrStutterInVBlankFinal,
+			&mode_lib->vba.MinPrefetchMode,
+			&mode_lib->vba.MaxPrefetchMode);
+
+	v->VStartupLines = __DML_VBA_MIN_VSTARTUP__;
+
+	iteration = 0;
+	MaxTotalRDBandwidth = 0;
+	NextPrefetchMode = mode_lib->vba.PrefetchModePerState[mode_lib->vba.VoltageLevel][mode_lib->vba.maxMpcComb];
+
+	do {
+		MaxTotalRDBandwidth = 0;
+		DestinationLineTimesForPrefetchLessThan2 = false;
+		VRatioPrefetchMoreThanMax = false;
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: Start loop: VStartup = %d\n", __func__, mode_lib->vba.VStartupLines);
+#endif
+		for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+			/* NOTE PerfetchMode variable is invalid in DAL as per the input received.
+			 * Hence the direction is to use PrefetchModePerState.
+			 */
+			TWait = dml32_CalculateTWait(
+				mode_lib->vba.PrefetchModePerState[mode_lib->vba.VoltageLevel][mode_lib->vba.maxMpcComb],
+				mode_lib->vba.UsesMALLForPStateChange[k],
+				mode_lib->vba.SynchronizeDRRDisplaysForUCLKPStateChangeFinal,
+				mode_lib->vba.DRRDisplay[k],
+				mode_lib->vba.DRAMClockChangeLatency,
+				mode_lib->vba.FCLKChangeLatency, v->UrgentLatency,
+				mode_lib->vba.SREnterPlusExitTime);
+
+			memset(&v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.myPipe, 0, sizeof(DmlPipe));
+
+			v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.myPipe.Dppclk = mode_lib->vba.DPPCLK[k];
+			v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.myPipe.Dispclk = mode_lib->vba.DISPCLK;
+			v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.myPipe.PixelClock = mode_lib->vba.PixelClock[k];
+			v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.myPipe.DCFClkDeepSleep = v->DCFCLKDeepSleep;
+			v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.myPipe.DPPPerSurface = mode_lib->vba.DPPPerPlane[k];
+			v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.myPipe.ScalerEnabled = mode_lib->vba.ScalerEnabled[k];
+			v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.myPipe.SourceRotation = mode_lib->vba.SourceRotation[k];
+			v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.myPipe.BlockWidth256BytesY = v->BlockWidth256BytesY[k];
+			v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.myPipe.BlockHeight256BytesY = v->BlockHeight256BytesY[k];
+			v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.myPipe.BlockWidth256BytesC = v->BlockWidth256BytesC[k];
+			v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.myPipe.BlockHeight256BytesC = v->BlockHeight256BytesC[k];
+			v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.myPipe.InterlaceEnable = mode_lib->vba.Interlace[k];
+			v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.myPipe.NumberOfCursors = mode_lib->vba.NumberOfCursors[k];
+			v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.myPipe.VBlank = mode_lib->vba.VTotal[k] - mode_lib->vba.VActive[k];
+			v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.myPipe.HTotal = mode_lib->vba.HTotal[k];
+			v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.myPipe.HActive = mode_lib->vba.HActive[k];
+			v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.myPipe.DCCEnable = mode_lib->vba.DCCEnable[k];
+			v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.myPipe.ODMMode = mode_lib->vba.ODMCombineEnabled[k];
+			v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.myPipe.SourcePixelFormat = mode_lib->vba.SourcePixelFormat[k];
+			v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.myPipe.BytePerPixelY = v->BytePerPixelY[k];
+			v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.myPipe.BytePerPixelC = v->BytePerPixelC[k];
+			v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.myPipe.ProgressiveToInterlaceUnitInOPP = mode_lib->vba.ProgressiveToInterlaceUnitInOPP;
+			v->ErrorResult[k] = dml32_CalculatePrefetchSchedule(
+					v,
+					k,
+					v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.HostVMInefficiencyFactor,
+					&v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.myPipe,
+					v->DSCDelay[k],
+					(unsigned int) (v->SwathWidthY[k] / v->HRatio[k]),
+					dml_min(v->VStartupLines, v->MaxVStartupLines[k]),
+					v->MaxVStartupLines[k],
+					v->UrgentLatency,
+					v->UrgentExtraLatency,
+					v->TCalc,
+					v->PDEAndMetaPTEBytesFrame[k],
+					v->MetaRowByte[k],
+					v->PixelPTEBytesPerRow[k],
+					v->PrefetchSourceLinesY[k],
+					v->SwathWidthY[k],
+					v->VInitPreFillY[k],
+					v->MaxNumSwathY[k],
+					v->PrefetchSourceLinesC[k],
+					v->SwathWidthC[k],
+					v->VInitPreFillC[k],
+					v->MaxNumSwathC[k],
+					v->swath_width_luma_ub[k],
+					v->swath_width_chroma_ub[k],
+					v->SwathHeightY[k],
+					v->SwathHeightC[k],
+					TWait,
+					(v->DRAMSpeedPerState[mode_lib->vba.VoltageLevel] <= MEM_STROBE_FREQ_MHZ ||
+						v->DCFCLKPerState[mode_lib->vba.VoltageLevel] <= DCFCLK_FREQ_EXTRA_PREFETCH_REQ_MHZ) ?
+							mode_lib->vba.ip.min_prefetch_in_strobe_us : 0,
+					mode_lib->vba.PrefetchModePerState[mode_lib->vba.VoltageLevel][mode_lib->vba.maxMpcComb] > 0 || mode_lib->vba.DRAMClockChangeRequirementFinal == false,
+
+					/* Output */
+					&v->DSTXAfterScaler[k],
+					&v->DSTYAfterScaler[k],
+					&v->DestinationLinesForPrefetch[k],
+					&v->PrefetchBandwidth[k],
+					&v->DestinationLinesToRequestVMInVBlank[k],
+					&v->DestinationLinesToRequestRowInVBlank[k],
+					&v->VRatioPrefetchY[k],
+					&v->VRatioPrefetchC[k],
+					&v->RequiredPrefetchPixDataBWLuma[k],
+					&v->RequiredPrefetchPixDataBWChroma[k],
+					&v->NotEnoughTimeForDynamicMetadata[k],
+					&v->Tno_bw[k], &v->prefetch_vmrow_bw[k],
+					&v->Tdmdl_vm[k],
+					&v->Tdmdl[k],
+					&v->TSetup[k],
+					&v->VUpdateOffsetPix[k],
+					&v->VUpdateWidthPix[k],
+					&v->VReadyOffsetPix[k]);
+
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: k=%0d Prefetch calculation errResult=%0d\n",
+					__func__, k, mode_lib->vba.ErrorResult[k]);
+#endif
+			v->VStartup[k] = dml_min(v->VStartupLines, v->MaxVStartupLines[k]);
+		}
+
+		for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+			dml32_CalculateUrgentBurstFactor(mode_lib->vba.UsesMALLForPStateChange[k],
+					v->swath_width_luma_ub[k],
+					v->swath_width_chroma_ub[k],
+					mode_lib->vba.SwathHeightY[k],
+					mode_lib->vba.SwathHeightC[k],
+					mode_lib->vba.HTotal[k] / mode_lib->vba.PixelClock[k],
+					v->UrgentLatency,
+					mode_lib->vba.CursorBufferSize,
+					mode_lib->vba.CursorWidth[k][0],
+					mode_lib->vba.CursorBPP[k][0],
+					v->VRatioPrefetchY[k],
+					v->VRatioPrefetchC[k],
+					v->BytePerPixelDETY[k],
+					v->BytePerPixelDETC[k],
+					mode_lib->vba.DETBufferSizeY[k],
+					mode_lib->vba.DETBufferSizeC[k],
+					/* Output */
+					&v->UrgBurstFactorCursorPre[k],
+					&v->UrgBurstFactorLumaPre[k],
+					&v->UrgBurstFactorChromaPre[k],
+					&v->NoUrgentLatencyHidingPre[k]);
+
+			v->cursor_bw_pre[k] = mode_lib->vba.NumberOfCursors[k] * mode_lib->vba.CursorWidth[k][0] * mode_lib->vba.CursorBPP[k][0] /
+					8.0 / (mode_lib->vba.HTotal[k] / mode_lib->vba.PixelClock[k]) * v->VRatioPrefetchY[k];
+
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: k=%0d DPPPerSurface=%d\n", __func__, k, mode_lib->vba.DPPPerPlane[k]);
+			dml_print("DML::%s: k=%0d UrgBurstFactorLuma=%f\n", __func__, k, v->UrgBurstFactorLuma[k]);
+			dml_print("DML::%s: k=%0d UrgBurstFactorChroma=%f\n", __func__, k, v->UrgBurstFactorChroma[k]);
+			dml_print("DML::%s: k=%0d UrgBurstFactorLumaPre=%f\n", __func__, k,
+					v->UrgBurstFactorLumaPre[k]);
+			dml_print("DML::%s: k=%0d UrgBurstFactorChromaPre=%f\n", __func__, k,
+					v->UrgBurstFactorChromaPre[k]);
+
+			dml_print("DML::%s: k=%0d VRatioPrefetchY=%f\n", __func__, k, v->VRatioPrefetchY[k]);
+			dml_print("DML::%s: k=%0d VRatioY=%f\n", __func__, k, mode_lib->vba.VRatio[k]);
+
+			dml_print("DML::%s: k=%0d prefetch_vmrow_bw=%f\n", __func__, k, v->prefetch_vmrow_bw[k]);
+			dml_print("DML::%s: k=%0d ReadBandwidthSurfaceLuma=%f\n", __func__, k,
+					v->ReadBandwidthSurfaceLuma[k]);
+			dml_print("DML::%s: k=%0d ReadBandwidthSurfaceChroma=%f\n", __func__, k,
+					v->ReadBandwidthSurfaceChroma[k]);
+			dml_print("DML::%s: k=%0d cursor_bw=%f\n", __func__, k, v->cursor_bw[k]);
+			dml_print("DML::%s: k=%0d meta_row_bw=%f\n", __func__, k, v->meta_row_bw[k]);
+			dml_print("DML::%s: k=%0d dpte_row_bw=%f\n", __func__, k, v->dpte_row_bw[k]);
+			dml_print("DML::%s: k=%0d RequiredPrefetchPixDataBWLuma=%f\n", __func__, k,
+					v->RequiredPrefetchPixDataBWLuma[k]);
+			dml_print("DML::%s: k=%0d RequiredPrefetchPixDataBWChroma=%f\n", __func__, k,
+					v->RequiredPrefetchPixDataBWChroma[k]);
+			dml_print("DML::%s: k=%0d cursor_bw_pre=%f\n", __func__, k, v->cursor_bw_pre[k]);
+			dml_print("DML::%s: k=%0d MaxTotalRDBandwidthNoUrgentBurst=%f\n", __func__, k,
+					MaxTotalRDBandwidthNoUrgentBurst);
+#endif
+			if (v->DestinationLinesForPrefetch[k] < 2)
+				DestinationLineTimesForPrefetchLessThan2 = true;
+
+			if (v->VRatioPrefetchY[k] > v->MaxVRatioPre
+					|| v->VRatioPrefetchC[k] > v->MaxVRatioPre)
+				VRatioPrefetchMoreThanMax = true;
+
+			//bool DestinationLinesToRequestVMInVBlankEqualOrMoreThan32 = false;
+			//bool DestinationLinesToRequestRowInVBlankEqualOrMoreThan16 = false;
+			//if (v->DestinationLinesToRequestVMInVBlank[k] >= 32) {
+			//    DestinationLinesToRequestVMInVBlankEqualOrMoreThan32 = true;
+			//}
+
+			//if (v->DestinationLinesToRequestRowInVBlank[k] >= 16) {
+			//    DestinationLinesToRequestRowInVBlankEqualOrMoreThan16 = true;
+			//}
+		}
+
+		v->FractionOfUrgentBandwidth = MaxTotalRDBandwidthNoUrgentBurst / mode_lib->vba.ReturnBW;
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: MaxTotalRDBandwidthNoUrgentBurst=%f\n",
+				__func__, MaxTotalRDBandwidthNoUrgentBurst);
+		dml_print("DML::%s: ReturnBW=%f\n", __func__, mode_lib->vba.ReturnBW);
+		dml_print("DML::%s: FractionOfUrgentBandwidth=%f\n",
+				__func__, mode_lib->vba.FractionOfUrgentBandwidth);
+#endif
+
+		{
+			dml32_CalculatePrefetchBandwithSupport(
+					mode_lib->vba.NumberOfActiveSurfaces,
+					mode_lib->vba.ReturnBW,
+					v->NoUrgentLatencyHidingPre,
+					v->ReadBandwidthSurfaceLuma,
+					v->ReadBandwidthSurfaceChroma,
+					v->RequiredPrefetchPixDataBWLuma,
+					v->RequiredPrefetchPixDataBWChroma,
+					v->cursor_bw,
+					v->meta_row_bw,
+					v->dpte_row_bw,
+					v->cursor_bw_pre,
+					v->prefetch_vmrow_bw,
+					mode_lib->vba.DPPPerPlane,
+					v->UrgBurstFactorLuma,
+					v->UrgBurstFactorChroma,
+					v->UrgBurstFactorCursor,
+					v->UrgBurstFactorLumaPre,
+					v->UrgBurstFactorChromaPre,
+					v->UrgBurstFactorCursorPre,
+					v->PrefetchBandwidth,
+					v->VRatio,
+					v->MaxVRatioPre,
+
+					/* output */
+					&MaxTotalRDBandwidth,
+					&v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.dummy_single[0],
+					&v->PrefetchModeSupported);
+		}
+
+		for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k)
+			v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.dummy_unit_vector[k] = 1.0;
+
+		{
+			dml32_CalculatePrefetchBandwithSupport(mode_lib->vba.NumberOfActiveSurfaces,
+					mode_lib->vba.ReturnBW,
+					v->NoUrgentLatencyHidingPre,
+					v->ReadBandwidthSurfaceLuma,
+					v->ReadBandwidthSurfaceChroma,
+					v->RequiredPrefetchPixDataBWLuma,
+					v->RequiredPrefetchPixDataBWChroma,
+					v->cursor_bw,
+					v->meta_row_bw,
+					v->dpte_row_bw,
+					v->cursor_bw_pre,
+					v->prefetch_vmrow_bw,
+					mode_lib->vba.DPPPerPlane,
+					v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.dummy_unit_vector,
+					v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.dummy_unit_vector,
+					v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.dummy_unit_vector,
+					v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.dummy_unit_vector,
+					v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.dummy_unit_vector,
+					v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.dummy_unit_vector,
+					v->PrefetchBandwidth,
+					v->VRatio,
+					v->MaxVRatioPre,
+
+					/* output */
+					&v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.dummy_single[0],
+					&v->FractionOfUrgentBandwidth,
+					&v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.dummy_boolean);
+		}
+
+		if (VRatioPrefetchMoreThanMax != false || DestinationLineTimesForPrefetchLessThan2 != false) {
+			v->PrefetchModeSupported = false;
+		}
+
+		for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+			if (v->ErrorResult[k] == true || v->NotEnoughTimeForDynamicMetadata[k]) {
+				v->PrefetchModeSupported = false;
+			}
+		}
+
+		if (v->PrefetchModeSupported == true && mode_lib->vba.ImmediateFlipSupport == true) {
+			mode_lib->vba.BandwidthAvailableForImmediateFlip = dml32_CalculateBandwidthAvailableForImmediateFlip(
+					mode_lib->vba.NumberOfActiveSurfaces,
+					mode_lib->vba.ReturnBW,
+					v->ReadBandwidthSurfaceLuma,
+					v->ReadBandwidthSurfaceChroma,
+					v->RequiredPrefetchPixDataBWLuma,
+					v->RequiredPrefetchPixDataBWChroma,
+					v->cursor_bw,
+					v->cursor_bw_pre,
+					mode_lib->vba.DPPPerPlane,
+					v->UrgBurstFactorLuma,
+					v->UrgBurstFactorChroma,
+					v->UrgBurstFactorCursor,
+					v->UrgBurstFactorLumaPre,
+					v->UrgBurstFactorChromaPre,
+					v->UrgBurstFactorCursorPre);
+
+			mode_lib->vba.TotImmediateFlipBytes = 0;
+			for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+				if (mode_lib->vba.ImmediateFlipRequirement[k] != dm_immediate_flip_not_required) {
+					mode_lib->vba.TotImmediateFlipBytes = mode_lib->vba.TotImmediateFlipBytes
+							+ mode_lib->vba.DPPPerPlane[k]
+									* (v->PDEAndMetaPTEBytesFrame[k]
+											+ v->MetaRowByte[k]);
+					if (v->use_one_row_for_frame_flip[k][0][0]) {
+						mode_lib->vba.TotImmediateFlipBytes =
+								mode_lib->vba.TotImmediateFlipBytes
+										+ 2 * v->PixelPTEBytesPerRow[k];
+					} else {
+						mode_lib->vba.TotImmediateFlipBytes =
+								mode_lib->vba.TotImmediateFlipBytes
+										+ v->PixelPTEBytesPerRow[k];
+					}
+				}
+			}
+			for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+				dml32_CalculateFlipSchedule(v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.HostVMInefficiencyFactor,
+						v->UrgentExtraLatency,
+						v->UrgentLatency,
+						mode_lib->vba.GPUVMMaxPageTableLevels,
+						mode_lib->vba.HostVMEnable,
+						mode_lib->vba.HostVMMaxNonCachedPageTableLevels,
+						mode_lib->vba.GPUVMEnable,
+						mode_lib->vba.HostVMMinPageSize,
+						v->PDEAndMetaPTEBytesFrame[k],
+						v->MetaRowByte[k],
+						v->PixelPTEBytesPerRow[k],
+						mode_lib->vba.BandwidthAvailableForImmediateFlip,
+						mode_lib->vba.TotImmediateFlipBytes,
+						mode_lib->vba.SourcePixelFormat[k],
+						mode_lib->vba.HTotal[k] / mode_lib->vba.PixelClock[k],
+						mode_lib->vba.VRatio[k],
+						mode_lib->vba.VRatioChroma[k],
+						v->Tno_bw[k],
+						mode_lib->vba.DCCEnable[k],
+						v->dpte_row_height[k],
+						v->meta_row_height[k],
+						v->dpte_row_height_chroma[k],
+						v->meta_row_height_chroma[k],
+						v->Use_One_Row_For_Frame_Flip[k],
+
+						/* Output */
+						&v->DestinationLinesToRequestVMInImmediateFlip[k],
+						&v->DestinationLinesToRequestRowInImmediateFlip[k],
+						&v->final_flip_bw[k],
+						&v->ImmediateFlipSupportedForPipe[k]);
+			}
+
+			{
+				dml32_CalculateImmediateFlipBandwithSupport(mode_lib->vba.NumberOfActiveSurfaces,
+						mode_lib->vba.ReturnBW,
+						mode_lib->vba.ImmediateFlipRequirement,
+						v->final_flip_bw,
+						v->ReadBandwidthSurfaceLuma,
+						v->ReadBandwidthSurfaceChroma,
+						v->RequiredPrefetchPixDataBWLuma,
+						v->RequiredPrefetchPixDataBWChroma,
+						v->cursor_bw,
+						v->meta_row_bw,
+						v->dpte_row_bw,
+						v->cursor_bw_pre,
+						v->prefetch_vmrow_bw,
+						mode_lib->vba.DPPPerPlane,
+						v->UrgBurstFactorLuma,
+						v->UrgBurstFactorChroma,
+						v->UrgBurstFactorCursor,
+						v->UrgBurstFactorLumaPre,
+						v->UrgBurstFactorChromaPre,
+						v->UrgBurstFactorCursorPre,
+
+						/* output */
+						&v->total_dcn_read_bw_with_flip,    // Single  *TotalBandwidth
+						&v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.dummy_single[0],                        // Single  *FractionOfUrgentBandwidth
+						&v->ImmediateFlipSupported);        // Boolean *ImmediateFlipBandwidthSupport
+
+				dml32_CalculateImmediateFlipBandwithSupport(mode_lib->vba.NumberOfActiveSurfaces,
+						mode_lib->vba.ReturnBW,
+						mode_lib->vba.ImmediateFlipRequirement,
+						v->final_flip_bw,
+						v->ReadBandwidthSurfaceLuma,
+						v->ReadBandwidthSurfaceChroma,
+						v->RequiredPrefetchPixDataBWLuma,
+						v->RequiredPrefetchPixDataBWChroma,
+						v->cursor_bw,
+						v->meta_row_bw,
+						v->dpte_row_bw,
+						v->cursor_bw_pre,
+						v->prefetch_vmrow_bw,
+						mode_lib->vba.DPPPerPlane,
+						v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.dummy_unit_vector,
+						v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.dummy_unit_vector,
+						v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.dummy_unit_vector,
+						v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.dummy_unit_vector,
+						v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.dummy_unit_vector,
+						v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.dummy_unit_vector,
+
+						/* output */
+						&v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.dummy_single[1],                                // Single  *TotalBandwidth
+						&v->FractionOfUrgentBandwidthImmediateFlip, // Single  *FractionOfUrgentBandwidth
+						&v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.dummy_boolean);                              // Boolean *ImmediateFlipBandwidthSupport
+			}
+
+			for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+				if (mode_lib->vba.ImmediateFlipRequirement[k] != dm_immediate_flip_not_required && v->ImmediateFlipSupportedForPipe[k] == false) {
+					v->ImmediateFlipSupported = false;
+#ifdef __DML_VBA_DEBUG__
+					dml_print("DML::%s: Pipe %0d not supporting iflip\n", __func__, k);
+#endif
+				}
+			}
+		} else {
+			v->ImmediateFlipSupported = false;
+		}
+
+		/* consider flip support is okay if the flip bw is ok or (when user does't require a iflip and there is no host vm) */
+		v->PrefetchAndImmediateFlipSupported = (v->PrefetchModeSupported == true &&
+				((!mode_lib->vba.ImmediateFlipSupport && !mode_lib->vba.HostVMEnable && !ImmediateFlipRequirementFinal) ||
+						v->ImmediateFlipSupported)) ? true : false;
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: PrefetchModeSupported = %d\n", __func__, locals->PrefetchModeSupported);
+		for (uint k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k)
+			dml_print("DML::%s: ImmediateFlipRequirement[%d] = %d\n", __func__, k,  mode_lib->vba.ImmediateFlipRequirement[k] == dm_immediate_flip_required);
+		dml_print("DML::%s: ImmediateFlipSupported = %d\n", __func__, locals->ImmediateFlipSupported);
+		dml_print("DML::%s: ImmediateFlipSupport = %d\n", __func__, mode_lib->vba.ImmediateFlipSupport);
+		dml_print("DML::%s: HostVMEnable = %d\n", __func__, mode_lib->vba.HostVMEnable);
+		dml_print("DML::%s: PrefetchAndImmediateFlipSupported = %d\n", __func__, locals->PrefetchAndImmediateFlipSupported);
+		dml_print("DML::%s: Done loop: Vstartup=%d, Max Vstartup=%d\n", __func__, locals->VStartupLines, locals->MaximumMaxVStartupLines);
+#endif
+
+		v->VStartupLines = v->VStartupLines + 1;
+
+		if (v->VStartupLines > v->MaximumMaxVStartupLines) {
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: Vstartup exceeds max vstartup, exiting loop\n", __func__);
+#endif
+			break; // VBA_DELTA: Implementation divergence! Gabe is *still* iterating across prefetch modes which we don't care to do
+		}
+		iteration++;
+		if (iteration > 2500) {
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: too many errors, exit now\n", __func__);
+			assert(0);
+#endif
+		}
+	} while (!(v->PrefetchAndImmediateFlipSupported || NextPrefetchMode > mode_lib->vba.MaxPrefetchMode));
+
+
+	if (v->VStartupLines <= v->MaximumMaxVStartupLines) {
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: Good, Prefetch and flip scheduling found solution at VStartupLines=%d\n", __func__, locals->VStartupLines-1);
+#endif
+	}
+
+
+	//Watermarks and NB P-State/DRAM Clock Change Support
+	{
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.mmSOCParameters.UrgentLatency = v->UrgentLatency;
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.mmSOCParameters.ExtraLatency = v->UrgentExtraLatency;
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.mmSOCParameters.WritebackLatency = mode_lib->vba.WritebackLatency;
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.mmSOCParameters.DRAMClockChangeLatency = mode_lib->vba.DRAMClockChangeLatency;
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.mmSOCParameters.FCLKChangeLatency = mode_lib->vba.FCLKChangeLatency;
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.mmSOCParameters.SRExitTime = mode_lib->vba.SRExitTime;
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.mmSOCParameters.SREnterPlusExitTime = mode_lib->vba.SREnterPlusExitTime;
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.mmSOCParameters.SRExitZ8Time = mode_lib->vba.SRExitZ8Time;
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.mmSOCParameters.SREnterPlusExitZ8Time = mode_lib->vba.SREnterPlusExitZ8Time;
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.mmSOCParameters.USRRetrainingLatency = mode_lib->vba.USRRetrainingLatency;
+		v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.mmSOCParameters.SMNLatency = mode_lib->vba.SMNLatency;
+
+		dml32_CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport(
+			v,
+			v->PrefetchModePerState[v->VoltageLevel][v->maxMpcComb],
+			v->DCFCLK,
+			v->ReturnBW,
+			v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.mmSOCParameters,
+			v->SOCCLK,
+			v->DCFCLKDeepSleep,
+			v->DETBufferSizeY,
+			v->DETBufferSizeC,
+			v->SwathHeightY,
+			v->SwathHeightC,
+			v->SwathWidthY,
+			v->SwathWidthC,
+			v->DPPPerPlane,
+			v->BytePerPixelDETY,
+			v->BytePerPixelDETC,
+			v->DSTXAfterScaler,
+			v->DSTYAfterScaler,
+			v->UnboundedRequestEnabled,
+			v->CompressedBufferSizeInkByte,
+
+			/* Output */
+			&v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.dummy_dramchange_support,
+			v->MaxActiveDRAMClockChangeLatencySupported,
+			v->SubViewportLinesNeededInMALL,
+			&v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.dummy_fclkchange_support,
+			&v->MinActiveFCLKChangeLatencySupported,
+			&v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.dummy_USRRetrainingSupport,
+			mode_lib->vba.ActiveDRAMClockChangeLatencyMargin);
+
+		/* DCN32 has a new struct Watermarks (typedef) which is used to store
+		 * calculated WM values. Copy over values from struct to vba varaibles
+		 * to ensure that the DCN32 getters return the correct value.
+		 */
+		v->UrgentWatermark = v->Watermark.UrgentWatermark;
+		v->WritebackUrgentWatermark = v->Watermark.WritebackUrgentWatermark;
+		v->DRAMClockChangeWatermark = v->Watermark.DRAMClockChangeWatermark;
+		v->WritebackDRAMClockChangeWatermark = v->Watermark.WritebackDRAMClockChangeWatermark;
+		v->StutterExitWatermark = v->Watermark.StutterExitWatermark;
+		v->StutterEnterPlusExitWatermark = v->Watermark.StutterEnterPlusExitWatermark;
+		v->Z8StutterExitWatermark = v->Watermark.Z8StutterExitWatermark;
+		v->Z8StutterEnterPlusExitWatermark = v->Watermark.Z8StutterEnterPlusExitWatermark;
+
+		for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+			if (mode_lib->vba.WritebackEnable[k] == true) {
+				v->WritebackAllowDRAMClockChangeEndPosition[k] = dml_max(0,
+						v->VStartup[k] * mode_lib->vba.HTotal[k] / mode_lib->vba.PixelClock[k]
+								- v->Watermark.WritebackDRAMClockChangeWatermark);
+				v->WritebackAllowFCLKChangeEndPosition[k] = dml_max(0,
+						v->VStartup[k] * mode_lib->vba.HTotal[k] / mode_lib->vba.PixelClock[k]
+								- v->Watermark.WritebackFCLKChangeWatermark);
+			} else {
+				v->WritebackAllowDRAMClockChangeEndPosition[k] = 0;
+				v->WritebackAllowFCLKChangeEndPosition[k] = 0;
+			}
+		}
+	}
+
+	//Display Pipeline Delivery Time in Prefetch, Groups
+	dml32_CalculatePixelDeliveryTimes(
+			mode_lib->vba.NumberOfActiveSurfaces,
+			mode_lib->vba.VRatio,
+			mode_lib->vba.VRatioChroma,
+			v->VRatioPrefetchY,
+			v->VRatioPrefetchC,
+			v->swath_width_luma_ub,
+			v->swath_width_chroma_ub,
+			mode_lib->vba.DPPPerPlane,
+			mode_lib->vba.HRatio,
+			mode_lib->vba.HRatioChroma,
+			mode_lib->vba.PixelClock,
+			v->PSCL_THROUGHPUT_LUMA,
+			v->PSCL_THROUGHPUT_CHROMA,
+			mode_lib->vba.DPPCLK,
+			v->BytePerPixelC,
+			mode_lib->vba.SourceRotation,
+			mode_lib->vba.NumberOfCursors,
+			mode_lib->vba.CursorWidth,
+			mode_lib->vba.CursorBPP,
+			v->BlockWidth256BytesY,
+			v->BlockHeight256BytesY,
+			v->BlockWidth256BytesC,
+			v->BlockHeight256BytesC,
+
+			/* Output */
+			v->DisplayPipeLineDeliveryTimeLuma,
+			v->DisplayPipeLineDeliveryTimeChroma,
+			v->DisplayPipeLineDeliveryTimeLumaPrefetch,
+			v->DisplayPipeLineDeliveryTimeChromaPrefetch,
+			v->DisplayPipeRequestDeliveryTimeLuma,
+			v->DisplayPipeRequestDeliveryTimeChroma,
+			v->DisplayPipeRequestDeliveryTimeLumaPrefetch,
+			v->DisplayPipeRequestDeliveryTimeChromaPrefetch,
+			v->CursorRequestDeliveryTime,
+			v->CursorRequestDeliveryTimePrefetch);
+
+	dml32_CalculateMetaAndPTETimes(v->Use_One_Row_For_Frame,
+			mode_lib->vba.NumberOfActiveSurfaces,
+			mode_lib->vba.GPUVMEnable,
+			mode_lib->vba.MetaChunkSize,
+			mode_lib->vba.MinMetaChunkSizeBytes,
+			mode_lib->vba.HTotal,
+			mode_lib->vba.VRatio,
+			mode_lib->vba.VRatioChroma,
+			v->DestinationLinesToRequestRowInVBlank,
+			v->DestinationLinesToRequestRowInImmediateFlip,
+			mode_lib->vba.DCCEnable,
+			mode_lib->vba.PixelClock,
+			v->BytePerPixelY,
+			v->BytePerPixelC,
+			mode_lib->vba.SourceRotation,
+			v->dpte_row_height,
+			v->dpte_row_height_chroma,
+			v->meta_row_width,
+			v->meta_row_width_chroma,
+			v->meta_row_height,
+			v->meta_row_height_chroma,
+			v->meta_req_width,
+			v->meta_req_width_chroma,
+			v->meta_req_height,
+			v->meta_req_height_chroma,
+			v->dpte_group_bytes,
+			v->PTERequestSizeY,
+			v->PTERequestSizeC,
+			v->PixelPTEReqWidthY,
+			v->PixelPTEReqHeightY,
+			v->PixelPTEReqWidthC,
+			v->PixelPTEReqHeightC,
+			v->dpte_row_width_luma_ub,
+			v->dpte_row_width_chroma_ub,
+
+			/* Output */
+			v->DST_Y_PER_PTE_ROW_NOM_L,
+			v->DST_Y_PER_PTE_ROW_NOM_C,
+			v->DST_Y_PER_META_ROW_NOM_L,
+			v->DST_Y_PER_META_ROW_NOM_C,
+			v->TimePerMetaChunkNominal,
+			v->TimePerChromaMetaChunkNominal,
+			v->TimePerMetaChunkVBlank,
+			v->TimePerChromaMetaChunkVBlank,
+			v->TimePerMetaChunkFlip,
+			v->TimePerChromaMetaChunkFlip,
+			v->time_per_pte_group_nom_luma,
+			v->time_per_pte_group_vblank_luma,
+			v->time_per_pte_group_flip_luma,
+			v->time_per_pte_group_nom_chroma,
+			v->time_per_pte_group_vblank_chroma,
+			v->time_per_pte_group_flip_chroma);
+
+	dml32_CalculateVMGroupAndRequestTimes(
+			mode_lib->vba.NumberOfActiveSurfaces,
+			mode_lib->vba.GPUVMEnable,
+			mode_lib->vba.GPUVMMaxPageTableLevels,
+			mode_lib->vba.HTotal,
+			v->BytePerPixelC,
+			v->DestinationLinesToRequestVMInVBlank,
+			v->DestinationLinesToRequestVMInImmediateFlip,
+			mode_lib->vba.DCCEnable,
+			mode_lib->vba.PixelClock,
+			v->dpte_row_width_luma_ub,
+			v->dpte_row_width_chroma_ub,
+			v->vm_group_bytes,
+			v->dpde0_bytes_per_frame_ub_l,
+			v->dpde0_bytes_per_frame_ub_c,
+			v->meta_pte_bytes_per_frame_ub_l,
+			v->meta_pte_bytes_per_frame_ub_c,
+
+			/* Output */
+			v->TimePerVMGroupVBlank,
+			v->TimePerVMGroupFlip,
+			v->TimePerVMRequestVBlank,
+			v->TimePerVMRequestFlip);
+
+	// Min TTUVBlank
+	for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+		if (mode_lib->vba.PrefetchModePerState[mode_lib->vba.VoltageLevel][mode_lib->vba.maxMpcComb] == 0) {
+			v->MinTTUVBlank[k] = dml_max4(v->Watermark.DRAMClockChangeWatermark,
+					v->Watermark.FCLKChangeWatermark, v->Watermark.StutterEnterPlusExitWatermark,
+					v->Watermark.UrgentWatermark);
+		} else if (mode_lib->vba.PrefetchModePerState[mode_lib->vba.VoltageLevel][mode_lib->vba.maxMpcComb]
+				== 1) {
+			v->MinTTUVBlank[k] = dml_max3(v->Watermark.FCLKChangeWatermark,
+					v->Watermark.StutterEnterPlusExitWatermark, v->Watermark.UrgentWatermark);
+		} else if (mode_lib->vba.PrefetchModePerState[mode_lib->vba.VoltageLevel][mode_lib->vba.maxMpcComb]
+				== 2) {
+			v->MinTTUVBlank[k] = dml_max(v->Watermark.StutterEnterPlusExitWatermark,
+					v->Watermark.UrgentWatermark);
+		} else {
+			v->MinTTUVBlank[k] = v->Watermark.UrgentWatermark;
+		}
+		if (!mode_lib->vba.DynamicMetadataEnable[k])
+			v->MinTTUVBlank[k] = mode_lib->vba.TCalc + v->MinTTUVBlank[k];
+	}
+
+	// DCC Configuration
+	for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: Calculate DCC configuration for surface k=%d\n", __func__, k);
+#endif
+		dml32_CalculateDCCConfiguration(
+				mode_lib->vba.DCCEnable[k],
+				mode_lib->vba.DCCProgrammingAssumesScanDirectionUnknownFinal,
+				mode_lib->vba.SourcePixelFormat[k], mode_lib->vba.SurfaceWidthY[k],
+				mode_lib->vba.SurfaceWidthC[k],
+				mode_lib->vba.SurfaceHeightY[k],
+				mode_lib->vba.SurfaceHeightC[k],
+				mode_lib->vba.nomDETInKByte,
+				v->BlockHeight256BytesY[k],
+				v->BlockHeight256BytesC[k],
+				mode_lib->vba.SurfaceTiling[k],
+				v->BytePerPixelY[k],
+				v->BytePerPixelC[k],
+				v->BytePerPixelDETY[k],
+				v->BytePerPixelDETC[k],
+				(enum dm_rotation_angle) mode_lib->vba.SourceScan[k],
+				/* Output */
+				&v->DCCYMaxUncompressedBlock[k],
+				&v->DCCCMaxUncompressedBlock[k],
+				&v->DCCYMaxCompressedBlock[k],
+				&v->DCCCMaxCompressedBlock[k],
+				&v->DCCYIndependentBlock[k],
+				&v->DCCCIndependentBlock[k]);
+	}
+
+	// VStartup Adjustment
+	for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+		bool isInterlaceTiming;
+		double Tvstartup_margin = (v->MaxVStartupLines[k] - v->VStartup[k]) * mode_lib->vba.HTotal[k]
+				/ mode_lib->vba.PixelClock[k];
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%d, MinTTUVBlank = %f (before vstartup margin)\n", __func__, k,
+				v->MinTTUVBlank[k]);
+#endif
+
+		v->MinTTUVBlank[k] = v->MinTTUVBlank[k] + Tvstartup_margin;
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%d, Tvstartup_margin = %f\n", __func__, k, Tvstartup_margin);
+		dml_print("DML::%s: k=%d, MaxVStartupLines = %d\n", __func__, k, v->MaxVStartupLines[k]);
+		dml_print("DML::%s: k=%d, VStartup = %d\n", __func__, k, v->VStartup[k]);
+		dml_print("DML::%s: k=%d, MinTTUVBlank = %f\n", __func__, k, v->MinTTUVBlank[k]);
+#endif
+
+		v->Tdmdl[k] = v->Tdmdl[k] + Tvstartup_margin;
+		if (mode_lib->vba.DynamicMetadataEnable[k] && mode_lib->vba.DynamicMetadataVMEnabled)
+			v->Tdmdl_vm[k] = v->Tdmdl_vm[k] + Tvstartup_margin;
+
+		isInterlaceTiming = (mode_lib->vba.Interlace[k] &&
+				!mode_lib->vba.ProgressiveToInterlaceUnitInOPP);
+
+		v->MIN_DST_Y_NEXT_START[k] = ((isInterlaceTiming ? dml_floor((mode_lib->vba.VTotal[k] -
+						mode_lib->vba.VFrontPorch[k]) / 2.0, 1.0) :
+						mode_lib->vba.VTotal[k]) - mode_lib->vba.VFrontPorch[k])
+						+ dml_max(1.0,
+						dml_ceil(v->WritebackDelay[mode_lib->vba.VoltageLevel][k]
+						/ (mode_lib->vba.HTotal[k] / mode_lib->vba.PixelClock[k]), 1.0))
+						+ dml_floor(4.0 * v->TSetup[k] / (mode_lib->vba.HTotal[k]
+						/ mode_lib->vba.PixelClock[k]), 1.0) / 4.0;
+
+		v->VStartup[k] = (isInterlaceTiming ? (2 * v->MaxVStartupLines[k]) : v->MaxVStartupLines[k]);
+
+		if (((v->VUpdateOffsetPix[k] + v->VUpdateWidthPix[k] + v->VReadyOffsetPix[k])
+			/ mode_lib->vba.HTotal[k]) <= (isInterlaceTiming ? dml_floor((mode_lib->vba.VTotal[k]
+			- mode_lib->vba.VActive[k] - mode_lib->vba.VFrontPorch[k] - v->VStartup[k]) / 2.0, 1.0) :
+			(int) (mode_lib->vba.VTotal[k] - mode_lib->vba.VActive[k]
+		       - mode_lib->vba.VFrontPorch[k] - v->VStartup[k]))) {
+			v->VREADY_AT_OR_AFTER_VSYNC[k] = true;
+		} else {
+			v->VREADY_AT_OR_AFTER_VSYNC[k] = false;
+		}
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%d, VStartup = %d (max)\n", __func__, k, v->VStartup[k]);
+		dml_print("DML::%s: k=%d, VUpdateOffsetPix = %d\n", __func__, k, v->VUpdateOffsetPix[k]);
+		dml_print("DML::%s: k=%d, VUpdateWidthPix = %d\n", __func__, k, v->VUpdateWidthPix[k]);
+		dml_print("DML::%s: k=%d, VReadyOffsetPix = %d\n", __func__, k, v->VReadyOffsetPix[k]);
+		dml_print("DML::%s: k=%d, HTotal = %d\n", __func__, k, mode_lib->vba.HTotal[k]);
+		dml_print("DML::%s: k=%d, VTotal = %d\n", __func__, k, mode_lib->vba.VTotal[k]);
+		dml_print("DML::%s: k=%d, VActive = %d\n", __func__, k, mode_lib->vba.VActive[k]);
+		dml_print("DML::%s: k=%d, VFrontPorch = %d\n", __func__, k, mode_lib->vba.VFrontPorch[k]);
+		dml_print("DML::%s: k=%d, VStartup = %d\n", __func__, k, v->VStartup[k]);
+		dml_print("DML::%s: k=%d, TSetup = %f\n", __func__, k, v->TSetup[k]);
+		dml_print("DML::%s: k=%d, MIN_DST_Y_NEXT_START = %f\n", __func__, k, v->MIN_DST_Y_NEXT_START[k]);
+		dml_print("DML::%s: k=%d, VREADY_AT_OR_AFTER_VSYNC = %d\n", __func__, k,
+				v->VREADY_AT_OR_AFTER_VSYNC[k]);
+#endif
+	}
+
+	{
+		//Maximum Bandwidth Used
+		for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+			if (mode_lib->vba.WritebackEnable[k] == true
+					&& mode_lib->vba.WritebackPixelFormat[k] == dm_444_32) {
+				WRBandwidth = mode_lib->vba.WritebackDestinationWidth[k]
+						* mode_lib->vba.WritebackDestinationHeight[k]
+						/ (mode_lib->vba.HTotal[k] * mode_lib->vba.WritebackSourceHeight[k]
+								/ mode_lib->vba.PixelClock[k]) * 4;
+			} else if (mode_lib->vba.WritebackEnable[k] == true) {
+				WRBandwidth = mode_lib->vba.WritebackDestinationWidth[k]
+						* mode_lib->vba.WritebackDestinationHeight[k]
+						/ (mode_lib->vba.HTotal[k] * mode_lib->vba.WritebackSourceHeight[k]
+								/ mode_lib->vba.PixelClock[k]) * 8;
+			}
+			TotalWRBandwidth = TotalWRBandwidth + WRBandwidth;
+		}
+
+		v->TotalDataReadBandwidth = 0;
+		for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+			v->TotalDataReadBandwidth = v->TotalDataReadBandwidth + v->ReadBandwidthSurfaceLuma[k]
+					+ v->ReadBandwidthSurfaceChroma[k];
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: k=%d, TotalDataReadBandwidth = %f\n",
+					__func__, k, v->TotalDataReadBandwidth);
+			dml_print("DML::%s: k=%d, ReadBandwidthSurfaceLuma = %f\n",
+					__func__, k, v->ReadBandwidthSurfaceLuma[k]);
+			dml_print("DML::%s: k=%d, ReadBandwidthSurfaceChroma = %f\n",
+					__func__, k, v->ReadBandwidthSurfaceChroma[k]);
+#endif
+		}
+	}
+
+	// Stutter Efficiency
+	dml32_CalculateStutterEfficiency(v->CompressedBufferSizeInkByte,
+			mode_lib->vba.UsesMALLForPStateChange,
+			v->UnboundedRequestEnabled,
+			mode_lib->vba.MetaFIFOSizeInKEntries,
+			mode_lib->vba.ZeroSizeBufferEntries,
+			mode_lib->vba.PixelChunkSizeInKByte,
+			mode_lib->vba.NumberOfActiveSurfaces,
+			mode_lib->vba.ROBBufferSizeInKByte,
+			v->TotalDataReadBandwidth,
+			mode_lib->vba.DCFCLK,
+			mode_lib->vba.ReturnBW,
+			v->CompbufReservedSpace64B,
+			v->CompbufReservedSpaceZs,
+			mode_lib->vba.SRExitTime,
+			mode_lib->vba.SRExitZ8Time,
+			mode_lib->vba.SynchronizeTimingsFinal,
+			mode_lib->vba.BlendingAndTiming,
+			v->Watermark.StutterEnterPlusExitWatermark,
+			v->Watermark.Z8StutterEnterPlusExitWatermark,
+			mode_lib->vba.ProgressiveToInterlaceUnitInOPP,
+			mode_lib->vba.Interlace,
+			v->MinTTUVBlank, mode_lib->vba.DPPPerPlane,
+			mode_lib->vba.DETBufferSizeY,
+			v->BytePerPixelY,
+			v->BytePerPixelDETY,
+			v->SwathWidthY,
+			mode_lib->vba.SwathHeightY,
+			mode_lib->vba.SwathHeightC,
+			mode_lib->vba.DCCRateLuma,
+			mode_lib->vba.DCCRateChroma,
+			mode_lib->vba.DCCFractionOfZeroSizeRequestsLuma,
+			mode_lib->vba.DCCFractionOfZeroSizeRequestsChroma,
+			mode_lib->vba.HTotal, mode_lib->vba.VTotal,
+			mode_lib->vba.PixelClock,
+			mode_lib->vba.VRatio,
+			mode_lib->vba.SourceRotation,
+			v->BlockHeight256BytesY,
+			v->BlockWidth256BytesY,
+			v->BlockHeight256BytesC,
+			v->BlockWidth256BytesC,
+			v->DCCYMaxUncompressedBlock,
+			v->DCCCMaxUncompressedBlock,
+			mode_lib->vba.VActive,
+			mode_lib->vba.DCCEnable,
+			mode_lib->vba.WritebackEnable,
+			v->ReadBandwidthSurfaceLuma,
+			v->ReadBandwidthSurfaceChroma,
+			v->meta_row_bw,
+			v->dpte_row_bw,
+			/* Output */
+			&v->StutterEfficiencyNotIncludingVBlank,
+			&v->StutterEfficiency,
+			&v->NumberOfStutterBurstsPerFrame,
+			&v->Z8StutterEfficiencyNotIncludingVBlank,
+			&v->Z8StutterEfficiency,
+			&v->Z8NumberOfStutterBurstsPerFrame,
+			&v->StutterPeriod,
+			&v->DCHUBBUB_ARB_CSTATE_MAX_CAP_MODE);
+
+#ifdef __DML_VBA_ALLOW_DELTA__
+	{
+		unsigned int dummy_integer[1];
+
+		// Calculate z8 stutter eff assuming 0 reserved space
+		dml32_CalculateStutterEfficiency(v->CompressedBufferSizeInkByte,
+				mode_lib->vba.UsesMALLForPStateChange,
+				v->UnboundedRequestEnabled,
+				mode_lib->vba.MetaFIFOSizeInKEntries,
+				mode_lib->vba.ZeroSizeBufferEntries,
+				mode_lib->vba.PixelChunkSizeInKByte,
+				mode_lib->vba.NumberOfActiveSurfaces,
+				mode_lib->vba.ROBBufferSizeInKByte,
+				v->TotalDataReadBandwidth,
+				mode_lib->vba.DCFCLK,
+				mode_lib->vba.ReturnBW,
+				0, //CompbufReservedSpace64B,
+				0, //CompbufReservedSpaceZs,
+				mode_lib->vba.SRExitTime,
+				mode_lib->vba.SRExitZ8Time,
+				mode_lib->vba.SynchronizeTimingsFinal,
+				mode_lib->vba.BlendingAndTiming,
+				v->Watermark.StutterEnterPlusExitWatermark,
+				v->Watermark.Z8StutterEnterPlusExitWatermark,
+				mode_lib->vba.ProgressiveToInterlaceUnitInOPP,
+				mode_lib->vba.Interlace,
+				v->MinTTUVBlank,
+				mode_lib->vba.DPPPerPlane,
+				mode_lib->vba.DETBufferSizeY,
+				v->BytePerPixelY, v->BytePerPixelDETY,
+				v->SwathWidthY, mode_lib->vba.SwathHeightY,
+				mode_lib->vba.SwathHeightC,
+				mode_lib->vba.DCCRateLuma,
+				mode_lib->vba.DCCRateChroma,
+				mode_lib->vba.DCCFractionOfZeroSizeRequestsLuma,
+				mode_lib->vba.DCCFractionOfZeroSizeRequestsChroma,
+				mode_lib->vba.HTotal,
+				mode_lib->vba.VTotal,
+				mode_lib->vba.PixelClock,
+				mode_lib->vba.VRatio,
+				mode_lib->vba.SourceRotation,
+				v->BlockHeight256BytesY,
+				v->BlockWidth256BytesY,
+				v->BlockHeight256BytesC,
+				v->BlockWidth256BytesC,
+				v->DCCYMaxUncompressedBlock,
+				v->DCCCMaxUncompressedBlock,
+				mode_lib->vba.VActive,
+				mode_lib->vba.DCCEnable,
+				mode_lib->vba.WritebackEnable,
+				v->ReadBandwidthSurfaceLuma,
+				v->ReadBandwidthSurfaceChroma,
+				v->meta_row_bw, v->dpte_row_bw,
+
+				/* Output */
+				&v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.dummy_single[0],
+				&v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.dummy_single[1],
+				&dummy_integer[0],
+				&v->Z8StutterEfficiencyNotIncludingVBlankBestCase,
+				&v->Z8StutterEfficiencyBestCase,
+				&v->Z8NumberOfStutterBurstsPerFrameBestCase,
+				&v->StutterPeriodBestCase,
+				&v->dummy_vars.DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.dummy_boolean);
+	}
+#else
+	v->Z8StutterEfficiencyNotIncludingVBlankBestCase = v->Z8StutterEfficiencyNotIncludingVBlank;
+	v->Z8StutterEfficiencyBestCase = v->Z8StutterEfficiency;
+	v->Z8NumberOfStutterBurstsPerFrameBestCase = v->Z8NumberOfStutterBurstsPerFrame;
+	v->StutterPeriodBestCase = v->StutterPeriod;
+#endif
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: --- END ---\n", __func__);
+#endif
+}
+
+static void mode_support_configuration(struct vba_vars_st *v,
+				  struct display_mode_lib *mode_lib)
+{
+	int i, j, start_state;
+
+	if (mode_lib->validate_max_state)
+		start_state = v->soc.num_states - 1;
+	else
+		start_state = 0;
+
+	for (i = v->soc.num_states - 1; i >= start_state; i--) {
+		for (j = 0; j < 2; j++) {
+			if (mode_lib->vba.ScaleRatioAndTapsSupport == true
+				&& mode_lib->vba.SourceFormatPixelAndScanSupport == true
+				&& mode_lib->vba.ViewportSizeSupport[i][j] == true
+				&& !mode_lib->vba.LinkRateDoesNotMatchDPVersion
+				&& !mode_lib->vba.LinkRateForMultistreamNotIndicated
+				&& !mode_lib->vba.BPPForMultistreamNotIndicated
+				&& !mode_lib->vba.MultistreamWithHDMIOreDP
+				&& !mode_lib->vba.ExceededMultistreamSlots[i]
+				&& !mode_lib->vba.MSOOrODMSplitWithNonDPLink
+				&& !mode_lib->vba.NotEnoughLanesForMSO
+				&& mode_lib->vba.LinkCapacitySupport[i] == true && !mode_lib->vba.P2IWith420
+				//&& !mode_lib->vba.DSCOnlyIfNecessaryWithBPP
+				&& !mode_lib->vba.DSC422NativeNotSupported
+				&& !mode_lib->vba.MPCCombineMethodIncompatible
+				&& mode_lib->vba.ODMCombine2To1SupportCheckOK[i] == true
+				&& mode_lib->vba.ODMCombine4To1SupportCheckOK[i] == true
+				&& mode_lib->vba.NotEnoughDSCUnits[i] == false
+				&& !mode_lib->vba.NotEnoughDSCSlices[i]
+				&& !mode_lib->vba.ImmediateFlipOrHostVMAndPStateWithMALLFullFrameOrPhantomPipe
+				&& !mode_lib->vba.InvalidCombinationOfMALLUseForPStateAndStaticScreen
+				&& mode_lib->vba.DSCCLKRequiredMoreThanSupported[i] == false
+				&& mode_lib->vba.PixelsPerLinePerDSCUnitSupport[i]
+				&& mode_lib->vba.DTBCLKRequiredMoreThanSupported[i] == false
+				&& !mode_lib->vba.InvalidCombinationOfMALLUseForPState
+				&& !mode_lib->vba.ImmediateFlipRequiredButTheRequirementForEachSurfaceIsNotSpecified
+				&& mode_lib->vba.ROBSupport[i][j] == true
+				&& mode_lib->vba.DISPCLK_DPPCLK_Support[i][j] == true
+				&& mode_lib->vba.TotalAvailablePipesSupport[i][j] == true
+				&& mode_lib->vba.NumberOfOTGSupport == true
+				&& mode_lib->vba.NumberOfHDMIFRLSupport == true
+				&& mode_lib->vba.EnoughWritebackUnits == true
+				&& mode_lib->vba.WritebackLatencySupport == true
+				&& mode_lib->vba.WritebackScaleRatioAndTapsSupport == true
+				&& mode_lib->vba.CursorSupport == true && mode_lib->vba.PitchSupport == true
+				&& mode_lib->vba.ViewportExceedsSurface == false
+				&& mode_lib->vba.PrefetchSupported[i][j] == true
+				&& mode_lib->vba.VActiveBandwithSupport[i][j] == true
+				&& mode_lib->vba.DynamicMetadataSupported[i][j] == true
+				&& mode_lib->vba.TotalVerticalActiveBandwidthSupport[i][j] == true
+				&& mode_lib->vba.VRatioInPrefetchSupported[i][j] == true
+				&& mode_lib->vba.PTEBufferSizeNotExceeded[i][j] == true
+				&& mode_lib->vba.DCCMetaBufferSizeNotExceeded[i][j] == true
+				&& mode_lib->vba.NonsupportedDSCInputBPC == false
+				&& !mode_lib->vba.ExceededMALLSize
+				&& (mode_lib->vba.NotEnoughDETSwathFillLatencyHidingPerState[i][j] == false
+				|| i == v->soc.num_states - 1)
+				&& ((mode_lib->vba.HostVMEnable == false
+				&& !mode_lib->vba.ImmediateFlipRequiredFinal)
+				|| mode_lib->vba.ImmediateFlipSupportedForState[i][j])
+				&& (!mode_lib->vba.DRAMClockChangeRequirementFinal
+				|| i == v->soc.num_states - 1
+				|| mode_lib->vba.DRAMClockChangeSupport[i][j] != dm_dram_clock_change_unsupported)
+				&& (!mode_lib->vba.FCLKChangeRequirementFinal || i == v->soc.num_states - 1
+				|| mode_lib->vba.FCLKChangeSupport[i][j] != dm_fclock_change_unsupported)
+				&& (!mode_lib->vba.USRRetrainingRequiredFinal
+				|| mode_lib->vba.USRRetrainingSupport[i][j])) {
+				mode_lib->vba.ModeSupport[i][j] = true;
+			} else {
+				mode_lib->vba.ModeSupport[i][j] = false;
+			}
+		}
+	}
+}
+
+void dml32_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_lib)
+{
+	struct vba_vars_st *v = &mode_lib->vba;
+	int i, j, start_state;
+	unsigned int k, m;
+	unsigned int MaximumMPCCombine;
+	unsigned int NumberOfNonCombinedSurfaceOfMaximumBandwidth;
+	unsigned int TotalSlots;
+	bool CompBufReservedSpaceNeedAdjustment;
+	bool CompBufReservedSpaceNeedAdjustmentSingleDPP;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: called\n", __func__);
+#endif
+
+	/*MODE SUPPORT, VOLTAGE STATE AND SOC CONFIGURATION*/
+	if (mode_lib->validate_max_state)
+		start_state = v->soc.num_states - 1;
+	else
+		start_state = 0;
+
+	/*Scale Ratio, taps Support Check*/
+
+	mode_lib->vba.ScaleRatioAndTapsSupport = true;
+	for (k = 0; k <= mode_lib->vba.NumberOfActiveSurfaces - 1; k++) {
+		if (mode_lib->vba.ScalerEnabled[k] == false
+				&& ((mode_lib->vba.SourcePixelFormat[k] != dm_444_64
+						&& mode_lib->vba.SourcePixelFormat[k] != dm_444_32
+						&& mode_lib->vba.SourcePixelFormat[k] != dm_444_16
+						&& mode_lib->vba.SourcePixelFormat[k] != dm_mono_16
+						&& mode_lib->vba.SourcePixelFormat[k] != dm_mono_8
+						&& mode_lib->vba.SourcePixelFormat[k] != dm_rgbe
+						&& mode_lib->vba.SourcePixelFormat[k] != dm_rgbe_alpha)
+						|| mode_lib->vba.HRatio[k] != 1.0 || mode_lib->vba.htaps[k] != 1.0
+						|| mode_lib->vba.VRatio[k] != 1.0 || mode_lib->vba.vtaps[k] != 1.0)) {
+			mode_lib->vba.ScaleRatioAndTapsSupport = false;
+		} else if (mode_lib->vba.vtaps[k] < 1.0 || mode_lib->vba.vtaps[k] > 8.0 || mode_lib->vba.htaps[k] < 1.0
+				|| mode_lib->vba.htaps[k] > 8.0
+				|| (mode_lib->vba.htaps[k] > 1.0 && (mode_lib->vba.htaps[k] % 2) == 1)
+				|| mode_lib->vba.HRatio[k] > mode_lib->vba.MaxHSCLRatio
+				|| mode_lib->vba.VRatio[k] > mode_lib->vba.MaxVSCLRatio
+				|| mode_lib->vba.HRatio[k] > mode_lib->vba.htaps[k]
+				|| mode_lib->vba.VRatio[k] > mode_lib->vba.vtaps[k]
+				|| (mode_lib->vba.SourcePixelFormat[k] != dm_444_64
+						&& mode_lib->vba.SourcePixelFormat[k] != dm_444_32
+						&& mode_lib->vba.SourcePixelFormat[k] != dm_444_16
+						&& mode_lib->vba.SourcePixelFormat[k] != dm_mono_16
+						&& mode_lib->vba.SourcePixelFormat[k] != dm_mono_8
+						&& mode_lib->vba.SourcePixelFormat[k] != dm_rgbe
+						&& (mode_lib->vba.VTAPsChroma[k] < 1
+								|| mode_lib->vba.VTAPsChroma[k] > 8
+								|| mode_lib->vba.HTAPsChroma[k] < 1
+								|| mode_lib->vba.HTAPsChroma[k] > 8
+								|| (mode_lib->vba.HTAPsChroma[k] > 1
+										&& mode_lib->vba.HTAPsChroma[k] % 2
+												== 1)
+								|| mode_lib->vba.HRatioChroma[k]
+										> mode_lib->vba.MaxHSCLRatio
+								|| mode_lib->vba.VRatioChroma[k]
+										> mode_lib->vba.MaxVSCLRatio
+								|| mode_lib->vba.HRatioChroma[k]
+										> mode_lib->vba.HTAPsChroma[k]
+								|| mode_lib->vba.VRatioChroma[k]
+										> mode_lib->vba.VTAPsChroma[k]))) {
+			mode_lib->vba.ScaleRatioAndTapsSupport = false;
+		}
+	}
+
+	/*Source Format, Pixel Format and Scan Support Check*/
+	mode_lib->vba.SourceFormatPixelAndScanSupport = true;
+	for (k = 0; k <= mode_lib->vba.NumberOfActiveSurfaces - 1; k++) {
+		if (mode_lib->vba.SurfaceTiling[k] == dm_sw_linear
+			&& (!(!IsVertical((enum dm_rotation_angle) mode_lib->vba.SourceScan[k]))
+				|| mode_lib->vba.DCCEnable[k] == true)) {
+			mode_lib->vba.SourceFormatPixelAndScanSupport = false;
+		}
+	}
+
+	for (k = 0; k <= mode_lib->vba.NumberOfActiveSurfaces - 1; k++) {
+		dml32_CalculateBytePerPixelAndBlockSizes(
+				mode_lib->vba.SourcePixelFormat[k],
+				mode_lib->vba.SurfaceTiling[k],
+
+				/* Output */
+				&mode_lib->vba.BytePerPixelY[k],
+				&mode_lib->vba.BytePerPixelC[k],
+				&mode_lib->vba.BytePerPixelInDETY[k],
+				&mode_lib->vba.BytePerPixelInDETC[k],
+				&mode_lib->vba.Read256BlockHeightY[k],
+				&mode_lib->vba.Read256BlockHeightC[k],
+				&mode_lib->vba.Read256BlockWidthY[k],
+				&mode_lib->vba.Read256BlockWidthC[k],
+				&mode_lib->vba.MacroTileHeightY[k],
+				&mode_lib->vba.MacroTileHeightC[k],
+				&mode_lib->vba.MacroTileWidthY[k],
+				&mode_lib->vba.MacroTileWidthC[k]);
+	}
+
+	/*Bandwidth Support Check*/
+	for (k = 0; k <= mode_lib->vba.NumberOfActiveSurfaces - 1; k++) {
+		if (!IsVertical(mode_lib->vba.SourceRotation[k])) {
+			v->SwathWidthYSingleDPP[k] = mode_lib->vba.ViewportWidth[k];
+			v->SwathWidthCSingleDPP[k] = mode_lib->vba.ViewportWidthChroma[k];
+		} else {
+			v->SwathWidthYSingleDPP[k] = mode_lib->vba.ViewportHeight[k];
+			v->SwathWidthCSingleDPP[k] = mode_lib->vba.ViewportHeightChroma[k];
+		}
+	}
+	for (k = 0; k <= mode_lib->vba.NumberOfActiveSurfaces - 1; k++) {
+		v->ReadBandwidthLuma[k] = v->SwathWidthYSingleDPP[k] * dml_ceil(v->BytePerPixelInDETY[k], 1.0)
+				/ (mode_lib->vba.HTotal[k] / mode_lib->vba.PixelClock[k]) * mode_lib->vba.VRatio[k];
+		v->ReadBandwidthChroma[k] = v->SwathWidthYSingleDPP[k] / 2 * dml_ceil(v->BytePerPixelInDETC[k], 2.0)
+				/ (mode_lib->vba.HTotal[k] / mode_lib->vba.PixelClock[k]) * mode_lib->vba.VRatio[k]
+				/ 2.0;
+	}
+	for (k = 0; k <= mode_lib->vba.NumberOfActiveSurfaces - 1; k++) {
+		if (mode_lib->vba.WritebackEnable[k] == true && mode_lib->vba.WritebackPixelFormat[k] == dm_444_64) {
+			v->WriteBandwidth[k] = mode_lib->vba.WritebackDestinationWidth[k]
+					* mode_lib->vba.WritebackDestinationHeight[k]
+					/ (mode_lib->vba.WritebackSourceHeight[k] * mode_lib->vba.HTotal[k]
+							/ mode_lib->vba.PixelClock[k]) * 8.0;
+		} else if (mode_lib->vba.WritebackEnable[k] == true) {
+			v->WriteBandwidth[k] = mode_lib->vba.WritebackDestinationWidth[k]
+					* mode_lib->vba.WritebackDestinationHeight[k]
+					/ (mode_lib->vba.WritebackSourceHeight[k] * mode_lib->vba.HTotal[k]
+							/ mode_lib->vba.PixelClock[k]) * 4.0;
+		} else {
+			v->WriteBandwidth[k] = 0.0;
+		}
+	}
+
+	/*Writeback Latency support check*/
+
+	mode_lib->vba.WritebackLatencySupport = true;
+	for (k = 0; k <= mode_lib->vba.NumberOfActiveSurfaces - 1; k++) {
+		if (mode_lib->vba.WritebackEnable[k] == true
+				&& (v->WriteBandwidth[k]
+						> mode_lib->vba.WritebackInterfaceBufferSize * 1024
+								/ mode_lib->vba.WritebackLatency)) {
+			mode_lib->vba.WritebackLatencySupport = false;
+		}
+	}
+
+	/*Writeback Mode Support Check*/
+	mode_lib->vba.EnoughWritebackUnits = true;
+	mode_lib->vba.TotalNumberOfActiveWriteback = 0;
+	for (k = 0; k <= mode_lib->vba.NumberOfActiveSurfaces - 1; k++) {
+		if (mode_lib->vba.WritebackEnable[k] == true)
+			mode_lib->vba.TotalNumberOfActiveWriteback = mode_lib->vba.TotalNumberOfActiveWriteback + 1;
+	}
+
+	if (mode_lib->vba.TotalNumberOfActiveWriteback > mode_lib->vba.MaxNumWriteback)
+		mode_lib->vba.EnoughWritebackUnits = false;
+
+	/*Writeback Scale Ratio and Taps Support Check*/
+	mode_lib->vba.WritebackScaleRatioAndTapsSupport = true;
+	for (k = 0; k <= mode_lib->vba.NumberOfActiveSurfaces - 1; k++) {
+		if (mode_lib->vba.WritebackEnable[k] == true) {
+			if (mode_lib->vba.WritebackHRatio[k] > mode_lib->vba.WritebackMaxHSCLRatio
+					|| mode_lib->vba.WritebackVRatio[k] > mode_lib->vba.WritebackMaxVSCLRatio
+					|| mode_lib->vba.WritebackHRatio[k] < mode_lib->vba.WritebackMinHSCLRatio
+					|| mode_lib->vba.WritebackVRatio[k] < mode_lib->vba.WritebackMinVSCLRatio
+					|| mode_lib->vba.WritebackHTaps[k] > mode_lib->vba.WritebackMaxHSCLTaps
+					|| mode_lib->vba.WritebackVTaps[k] > mode_lib->vba.WritebackMaxVSCLTaps
+					|| mode_lib->vba.WritebackHRatio[k] > mode_lib->vba.WritebackHTaps[k]
+					|| mode_lib->vba.WritebackVRatio[k] > mode_lib->vba.WritebackVTaps[k]
+					|| (mode_lib->vba.WritebackHTaps[k] > 2.0
+							&& ((mode_lib->vba.WritebackHTaps[k] % 2) == 1))) {
+				mode_lib->vba.WritebackScaleRatioAndTapsSupport = false;
+			}
+			if (2.0 * mode_lib->vba.WritebackDestinationWidth[k] * (mode_lib->vba.WritebackVTaps[k] - 1)
+					* 57 > mode_lib->vba.WritebackLineBufferSize) {
+				mode_lib->vba.WritebackScaleRatioAndTapsSupport = false;
+			}
+		}
+	}
+
+	for (k = 0; k <= mode_lib->vba.NumberOfActiveSurfaces - 1; k++) {
+		dml32_CalculateSinglePipeDPPCLKAndSCLThroughput(mode_lib->vba.HRatio[k], mode_lib->vba.HRatioChroma[k],
+				mode_lib->vba.VRatio[k], mode_lib->vba.VRatioChroma[k],
+				mode_lib->vba.MaxDCHUBToPSCLThroughput, mode_lib->vba.MaxPSCLToLBThroughput,
+				mode_lib->vba.PixelClock[k], mode_lib->vba.SourcePixelFormat[k],
+				mode_lib->vba.htaps[k], mode_lib->vba.HTAPsChroma[k], mode_lib->vba.vtaps[k],
+				mode_lib->vba.VTAPsChroma[k],
+				/* Output */
+				&mode_lib->vba.PSCL_FACTOR[k], &mode_lib->vba.PSCL_FACTOR_CHROMA[k],
+				&mode_lib->vba.MinDPPCLKUsingSingleDPP[k]);
+	}
+
+	for (k = 0; k <= mode_lib->vba.NumberOfActiveSurfaces - 1; k++) {
+
+		if (mode_lib->vba.SurfaceTiling[k] == dm_sw_linear) {
+			v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.MaximumSwathWidthSupportLuma = 8192;
+		} else if (!IsVertical(mode_lib->vba.SourceRotation[k]) && v->BytePerPixelC[k] > 0
+				&& mode_lib->vba.SourcePixelFormat[k] != dm_rgbe_alpha) {
+			v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.MaximumSwathWidthSupportLuma = 7680;
+		} else if (IsVertical(mode_lib->vba.SourceRotation[k]) && v->BytePerPixelC[k] > 0
+				&& mode_lib->vba.SourcePixelFormat[k] != dm_rgbe_alpha) {
+			v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.MaximumSwathWidthSupportLuma = 4320;
+		} else if (mode_lib->vba.SourcePixelFormat[k] == dm_rgbe_alpha) {
+			v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.MaximumSwathWidthSupportLuma = 3840;
+		} else if (IsVertical(mode_lib->vba.SourceRotation[k]) && v->BytePerPixelY[k] == 8 &&
+				mode_lib->vba.DCCEnable[k] == true) {
+			v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.MaximumSwathWidthSupportLuma = 3072;
+		} else {
+			v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.MaximumSwathWidthSupportLuma = 6144;
+		}
+
+		if (mode_lib->vba.SourcePixelFormat[k] == dm_420_8 || mode_lib->vba.SourcePixelFormat[k] == dm_420_10
+				|| mode_lib->vba.SourcePixelFormat[k] == dm_420_12) {
+			v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.MaximumSwathWidthSupportChroma = v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.MaximumSwathWidthSupportLuma / 2.0;
+		} else {
+			v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.MaximumSwathWidthSupportChroma = v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.MaximumSwathWidthSupportLuma;
+		}
+		v->MaximumSwathWidthInLineBufferLuma = mode_lib->vba.LineBufferSizeFinal
+				* dml_max(mode_lib->vba.HRatio[k], 1.0) / mode_lib->vba.LBBitPerPixel[k]
+				/ (mode_lib->vba.vtaps[k] + dml_max(dml_ceil(mode_lib->vba.VRatio[k], 1.0) - 2, 0.0));
+		if (v->BytePerPixelC[k] == 0.0) {
+			v->MaximumSwathWidthInLineBufferChroma = 0;
+		} else {
+			v->MaximumSwathWidthInLineBufferChroma = mode_lib->vba.LineBufferSizeFinal
+					* dml_max(mode_lib->vba.HRatioChroma[k], 1.0) / mode_lib->vba.LBBitPerPixel[k]
+					/ (mode_lib->vba.VTAPsChroma[k]
+							+ dml_max(dml_ceil(mode_lib->vba.VRatioChroma[k], 1.0) - 2,
+									0.0));
+		}
+		v->MaximumSwathWidthLuma[k] = dml_min(v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.MaximumSwathWidthSupportLuma,
+				v->MaximumSwathWidthInLineBufferLuma);
+		v->MaximumSwathWidthChroma[k] = dml_min(v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.MaximumSwathWidthSupportChroma,
+				v->MaximumSwathWidthInLineBufferChroma);
+	}
+
+	dml32_CalculateSwathAndDETConfiguration(
+			mode_lib->vba.DETSizeOverride,
+			mode_lib->vba.UsesMALLForPStateChange,
+			mode_lib->vba.ConfigReturnBufferSizeInKByte,
+			mode_lib->vba.MaxTotalDETInKByte,
+			mode_lib->vba.MinCompressedBufferSizeInKByte,
+			1, /* ForceSingleDPP */
+			mode_lib->vba.NumberOfActiveSurfaces,
+			mode_lib->vba.nomDETInKByte,
+			mode_lib->vba.UseUnboundedRequesting,
+			mode_lib->vba.DisableUnboundRequestIfCompBufReservedSpaceNeedAdjustment,
+			mode_lib->vba.ip.pixel_chunk_size_kbytes,
+			mode_lib->vba.ip.rob_buffer_size_kbytes,
+			mode_lib->vba.CompressedBufferSegmentSizeInkByteFinal,
+			mode_lib->vba.Output,
+			mode_lib->vba.ReadBandwidthLuma,
+			mode_lib->vba.ReadBandwidthChroma,
+			mode_lib->vba.MaximumSwathWidthLuma,
+			mode_lib->vba.MaximumSwathWidthChroma,
+			mode_lib->vba.SourceRotation,
+			mode_lib->vba.ViewportStationary,
+			mode_lib->vba.SourcePixelFormat,
+			mode_lib->vba.SurfaceTiling,
+			mode_lib->vba.ViewportWidth,
+			mode_lib->vba.ViewportHeight,
+			mode_lib->vba.ViewportXStartY,
+			mode_lib->vba.ViewportYStartY,
+			mode_lib->vba.ViewportXStartC,
+			mode_lib->vba.ViewportYStartC,
+			mode_lib->vba.SurfaceWidthY,
+			mode_lib->vba.SurfaceWidthC,
+			mode_lib->vba.SurfaceHeightY,
+			mode_lib->vba.SurfaceHeightC,
+			mode_lib->vba.Read256BlockHeightY,
+			mode_lib->vba.Read256BlockHeightC,
+			mode_lib->vba.Read256BlockWidthY,
+			mode_lib->vba.Read256BlockWidthC,
+			v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_odm_mode,
+			mode_lib->vba.BlendingAndTiming,
+			mode_lib->vba.BytePerPixelY,
+			mode_lib->vba.BytePerPixelC,
+			mode_lib->vba.BytePerPixelInDETY,
+			mode_lib->vba.BytePerPixelInDETC,
+			mode_lib->vba.HActive,
+			mode_lib->vba.HRatio,
+			mode_lib->vba.HRatioChroma,
+			v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_integer_array[0], /*  Integer DPPPerSurface[] */
+
+			/* Output */
+			v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_integer_array[1], /* Long            swath_width_luma_ub[] */
+			v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_integer_array[2], /* Long            swath_width_chroma_ub[]  */
+			v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_double_array[0], /* Long            SwathWidth[]  */
+			v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_double_array[1], /* Long            SwathWidthChroma[]  */
+			v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_integer_array[3], /* Integer         SwathHeightY[]  */
+			v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_integer_array[4], /* Integer         SwathHeightC[]  */
+			v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_integer_array[5], /* Long            DETBufferSizeInKByte[]  */
+			v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_integer_array[6], /* Long            DETBufferSizeY[]  */
+			v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_integer_array[7], /* Long            DETBufferSizeC[]  */
+			&v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_boolean_array[0][0], /* bool           *UnboundedRequestEnabled  */
+			&v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_integer_array[0][0], /* Long           *CompressedBufferSizeInkByte  */
+			&v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_integer_array[1][0], /* Long           *CompBufReservedSpaceKBytes */
+			&CompBufReservedSpaceNeedAdjustmentSingleDPP,
+			mode_lib->vba.SingleDPPViewportSizeSupportPerSurface,/* bool ViewportSizeSupportPerSurface[] */
+			&v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_boolean_array[1][0]); /* bool           *ViewportSizeSupport */
+
+	v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.MPCCombineMethodAsNeededForPStateChangeAndVoltage = false;
+	v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.MPCCombineMethodAsPossible = false;
+
+	for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+		if (mode_lib->vba.MPCCombineUse[k] == dm_mpc_reduce_voltage_and_clocks)
+			v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.MPCCombineMethodAsNeededForPStateChangeAndVoltage = true;
+		if (mode_lib->vba.MPCCombineUse[k] == dm_mpc_always_when_possible)
+			v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.MPCCombineMethodAsPossible = true;
+	}
+	mode_lib->vba.MPCCombineMethodIncompatible = v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.MPCCombineMethodAsNeededForPStateChangeAndVoltage
+			&& v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.MPCCombineMethodAsPossible;
+
+	for (i = start_state; i < v->soc.num_states; i++) {
+		for (j = 0; j < 2; j++) {
+			mode_lib->vba.TotalNumberOfActiveDPP[i][j] = 0;
+			mode_lib->vba.TotalAvailablePipesSupport[i][j] = true;
+			v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.ODMModeNoDSC = dm_odm_combine_mode_disabled;
+			v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.ODMModeDSC = dm_odm_combine_mode_disabled;
+
+			for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+				dml32_CalculateODMMode(
+						mode_lib->vba.MaximumPixelsPerLinePerDSCUnit,
+						mode_lib->vba.HActive[k],
+						mode_lib->vba.OutputFormat[k],
+						mode_lib->vba.Output[k],
+						mode_lib->vba.ODMUse[k],
+						mode_lib->vba.MaxDispclk[i],
+						mode_lib->vba.MaxDispclk[v->soc.num_states - 1],
+						false,
+						mode_lib->vba.TotalNumberOfActiveDPP[i][j],
+						mode_lib->vba.MaxNumDPP,
+						mode_lib->vba.PixelClock[k],
+						mode_lib->vba.DISPCLKDPPCLKDSCCLKDownSpreading,
+						mode_lib->vba.DISPCLKRampingMargin,
+						mode_lib->vba.DISPCLKDPPCLKVCOSpeed,
+						mode_lib->vba.NumberOfDSCSlices[k],
+
+						/* Output */
+						&v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.TotalAvailablePipesSupportNoDSC,
+						&v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.NumberOfDPPNoDSC,
+						&v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.ODMModeNoDSC,
+						&v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.RequiredDISPCLKPerSurfaceNoDSC);
+
+				dml32_CalculateODMMode(
+						mode_lib->vba.MaximumPixelsPerLinePerDSCUnit,
+						mode_lib->vba.HActive[k],
+						mode_lib->vba.OutputFormat[k],
+						mode_lib->vba.Output[k],
+						mode_lib->vba.ODMUse[k],
+						mode_lib->vba.MaxDispclk[i],
+						mode_lib->vba.MaxDispclk[v->soc.num_states - 1],
+						true,
+						mode_lib->vba.TotalNumberOfActiveDPP[i][j],
+						mode_lib->vba.MaxNumDPP,
+						mode_lib->vba.PixelClock[k],
+						mode_lib->vba.DISPCLKDPPCLKDSCCLKDownSpreading,
+						mode_lib->vba.DISPCLKRampingMargin,
+						mode_lib->vba.DISPCLKDPPCLKVCOSpeed,
+						mode_lib->vba.NumberOfDSCSlices[k],
+
+						/* Output */
+						&v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.TotalAvailablePipesSupportDSC,
+						&v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.NumberOfDPPDSC,
+						&v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.ODMModeDSC,
+						&v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.RequiredDISPCLKPerSurfaceDSC);
+
+				dml32_CalculateOutputLink(
+						mode_lib->vba.PHYCLKPerState[i],
+						mode_lib->vba.PHYCLKD18PerState[i],
+						mode_lib->vba.PHYCLKD32PerState[i],
+						mode_lib->vba.Downspreading,
+						(mode_lib->vba.BlendingAndTiming[k] == k),
+						mode_lib->vba.Output[k],
+						mode_lib->vba.OutputFormat[k],
+						mode_lib->vba.HTotal[k],
+						mode_lib->vba.HActive[k],
+						mode_lib->vba.PixelClockBackEnd[k],
+						mode_lib->vba.ForcedOutputLinkBPP[k],
+						mode_lib->vba.DSCInputBitPerComponent[k],
+						mode_lib->vba.NumberOfDSCSlices[k],
+						mode_lib->vba.AudioSampleRate[k],
+						mode_lib->vba.AudioSampleLayout[k],
+						v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.ODMModeNoDSC,
+						v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.ODMModeDSC,
+						mode_lib->vba.DSCEnable[k],
+						mode_lib->vba.OutputLinkDPLanes[k],
+						mode_lib->vba.OutputLinkDPRate[k],
+
+						/* Output */
+						&mode_lib->vba.RequiresDSC[i][k],
+						&mode_lib->vba.RequiresFEC[i][k],
+						&mode_lib->vba.OutputBppPerState[i][k],
+						&mode_lib->vba.OutputTypePerState[i][k],
+						&mode_lib->vba.OutputRatePerState[i][k],
+						&mode_lib->vba.RequiredSlots[i][k]);
+
+				if (mode_lib->vba.RequiresDSC[i][k] == false) {
+					mode_lib->vba.ODMCombineEnablePerState[i][k] = v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.ODMModeNoDSC;
+					mode_lib->vba.RequiredDISPCLKPerSurface[i][j][k] =
+							v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.RequiredDISPCLKPerSurfaceNoDSC;
+					if (!v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.TotalAvailablePipesSupportNoDSC)
+						mode_lib->vba.TotalAvailablePipesSupport[i][j] = false;
+					mode_lib->vba.TotalNumberOfActiveDPP[i][j] =
+							mode_lib->vba.TotalNumberOfActiveDPP[i][j] + v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.NumberOfDPPNoDSC;
+				} else {
+					mode_lib->vba.ODMCombineEnablePerState[i][k] = v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.ODMModeDSC;
+					mode_lib->vba.RequiredDISPCLKPerSurface[i][j][k] =
+							v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.RequiredDISPCLKPerSurfaceDSC;
+					if (!v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.TotalAvailablePipesSupportDSC)
+						mode_lib->vba.TotalAvailablePipesSupport[i][j] = false;
+					mode_lib->vba.TotalNumberOfActiveDPP[i][j] =
+							mode_lib->vba.TotalNumberOfActiveDPP[i][j] + v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.NumberOfDPPDSC;
+				}
+			}
+
+			for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+				if (mode_lib->vba.ODMCombineEnablePerState[i][k] == dm_odm_combine_mode_4to1) {
+					mode_lib->vba.MPCCombine[i][j][k] = false;
+					mode_lib->vba.NoOfDPP[i][j][k] = 4;
+				} else if (mode_lib->vba.ODMCombineEnablePerState[i][k] == dm_odm_combine_mode_2to1) {
+					mode_lib->vba.MPCCombine[i][j][k] = false;
+					mode_lib->vba.NoOfDPP[i][j][k] = 2;
+				} else if (mode_lib->vba.MPCCombineUse[k] == dm_mpc_never) {
+					mode_lib->vba.MPCCombine[i][j][k] = false;
+					mode_lib->vba.NoOfDPP[i][j][k] = 1;
+				} else if (dml32_RoundToDFSGranularity(
+						mode_lib->vba.MinDPPCLKUsingSingleDPP[k]
+								* (1 + mode_lib->vba.DISPCLKDPPCLKDSCCLKDownSpreading
+												/ 100), 1,
+						mode_lib->vba.DISPCLKDPPCLKVCOSpeed) <= mode_lib->vba.MaxDppclk[i] &&
+				mode_lib->vba.SingleDPPViewportSizeSupportPerSurface[k] == true) {
+					mode_lib->vba.MPCCombine[i][j][k] = false;
+					mode_lib->vba.NoOfDPP[i][j][k] = 1;
+				} else if (mode_lib->vba.TotalNumberOfActiveDPP[i][j] < mode_lib->vba.MaxNumDPP) {
+					mode_lib->vba.MPCCombine[i][j][k] = true;
+					mode_lib->vba.NoOfDPP[i][j][k] = 2;
+					mode_lib->vba.TotalNumberOfActiveDPP[i][j] =
+							mode_lib->vba.TotalNumberOfActiveDPP[i][j] + 1;
+				} else {
+					mode_lib->vba.MPCCombine[i][j][k] = false;
+					mode_lib->vba.NoOfDPP[i][j][k] = 1;
+					mode_lib->vba.TotalAvailablePipesSupport[i][j] = false;
+				}
+			}
+
+			mode_lib->vba.TotalNumberOfSingleDPPSurfaces[i][j] = 0;
+			v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.NoChroma = true;
+
+			for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+				if (mode_lib->vba.NoOfDPP[i][j][k] == 1)
+					mode_lib->vba.TotalNumberOfSingleDPPSurfaces[i][j] =
+							mode_lib->vba.TotalNumberOfSingleDPPSurfaces[i][j] + 1;
+				if (mode_lib->vba.SourcePixelFormat[k] == dm_420_8
+						|| mode_lib->vba.SourcePixelFormat[k] == dm_420_10
+						|| mode_lib->vba.SourcePixelFormat[k] == dm_420_12
+						|| mode_lib->vba.SourcePixelFormat[k] == dm_rgbe_alpha) {
+					v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.NoChroma = false;
+				}
+			}
+
+			// if TotalNumberOfActiveDPP is > 1, then there should be no unbounded req mode (hw limitation), the comp buf reserved adjustment is not needed regardless
+			// if TotalNumberOfActiveDPP is == 1, then will use the SingleDPP version of unbounded_req for the decision
+			CompBufReservedSpaceNeedAdjustment = (mode_lib->vba.TotalNumberOfActiveDPP[i][j] > 1) ? 0 : CompBufReservedSpaceNeedAdjustmentSingleDPP;
+
+
+
+			if (j == 1 && !dml32_UnboundedRequest(mode_lib->vba.UseUnboundedRequesting,
+					mode_lib->vba.TotalNumberOfActiveDPP[i][j], v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.NoChroma,
+					mode_lib->vba.Output[0],
+					mode_lib->vba.SurfaceTiling[0],
+					CompBufReservedSpaceNeedAdjustment,
+					mode_lib->vba.DisableUnboundRequestIfCompBufReservedSpaceNeedAdjustment)) {
+				while (!(mode_lib->vba.TotalNumberOfActiveDPP[i][j] >= mode_lib->vba.MaxNumDPP
+						|| mode_lib->vba.TotalNumberOfSingleDPPSurfaces[i][j] == 0)) {
+					v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.BWOfNonCombinedSurfaceOfMaximumBandwidth = 0;
+					NumberOfNonCombinedSurfaceOfMaximumBandwidth = 0;
+
+					for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+						if (mode_lib->vba.MPCCombineUse[k]
+							!= dm_mpc_never &&
+							mode_lib->vba.MPCCombineUse[k] != dm_mpc_reduce_voltage &&
+							mode_lib->vba.ReadBandwidthLuma[k] +
+							mode_lib->vba.ReadBandwidthChroma[k] >
+							v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.BWOfNonCombinedSurfaceOfMaximumBandwidth &&
+							(mode_lib->vba.ODMCombineEnablePerState[i][k] !=
+							dm_odm_combine_mode_2to1 &&
+							mode_lib->vba.ODMCombineEnablePerState[i][k] !=
+							dm_odm_combine_mode_4to1) &&
+								mode_lib->vba.MPCCombine[i][j][k] == false) {
+							v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.BWOfNonCombinedSurfaceOfMaximumBandwidth =
+								mode_lib->vba.ReadBandwidthLuma[k]
+								+ mode_lib->vba.ReadBandwidthChroma[k];
+							NumberOfNonCombinedSurfaceOfMaximumBandwidth = k;
+						}
+					}
+					mode_lib->vba.MPCCombine[i][j][NumberOfNonCombinedSurfaceOfMaximumBandwidth] =
+							true;
+					mode_lib->vba.NoOfDPP[i][j][NumberOfNonCombinedSurfaceOfMaximumBandwidth] = 2;
+					mode_lib->vba.TotalNumberOfActiveDPP[i][j] =
+							mode_lib->vba.TotalNumberOfActiveDPP[i][j] + 1;
+					mode_lib->vba.TotalNumberOfSingleDPPSurfaces[i][j] =
+							mode_lib->vba.TotalNumberOfSingleDPPSurfaces[i][j] - 1;
+				}
+			}
+
+			//DISPCLK/DPPCLK
+			mode_lib->vba.WritebackRequiredDISPCLK = 0;
+			for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+				if (mode_lib->vba.WritebackEnable[k]) {
+					mode_lib->vba.WritebackRequiredDISPCLK = dml_max(
+							mode_lib->vba.WritebackRequiredDISPCLK,
+							dml32_CalculateWriteBackDISPCLK(
+									mode_lib->vba.WritebackPixelFormat[k],
+									mode_lib->vba.PixelClock[k],
+									mode_lib->vba.WritebackHRatio[k],
+									mode_lib->vba.WritebackVRatio[k],
+									mode_lib->vba.WritebackHTaps[k],
+									mode_lib->vba.WritebackVTaps[k],
+									mode_lib->vba.WritebackSourceWidth[k],
+									mode_lib->vba.WritebackDestinationWidth[k],
+									mode_lib->vba.HTotal[k],
+									mode_lib->vba.WritebackLineBufferSize,
+									mode_lib->vba.DISPCLKDPPCLKVCOSpeed));
+				}
+			}
+
+			mode_lib->vba.RequiredDISPCLK[i][j] = mode_lib->vba.WritebackRequiredDISPCLK;
+			for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+				mode_lib->vba.RequiredDISPCLK[i][j] = dml_max(mode_lib->vba.RequiredDISPCLK[i][j],
+						mode_lib->vba.RequiredDISPCLKPerSurface[i][j][k]);
+			}
+
+			for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k)
+				mode_lib->vba.NoOfDPPThisState[k] = mode_lib->vba.NoOfDPP[i][j][k];
+
+			dml32_CalculateDPPCLK(mode_lib->vba.NumberOfActiveSurfaces,
+					mode_lib->vba.DISPCLKDPPCLKDSCCLKDownSpreading,
+					mode_lib->vba.DISPCLKDPPCLKVCOSpeed, mode_lib->vba.MinDPPCLKUsingSingleDPP,
+					mode_lib->vba.NoOfDPPThisState,
+					/* Output */
+					&mode_lib->vba.GlobalDPPCLK, mode_lib->vba.RequiredDPPCLKThisState);
+
+			for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k)
+				mode_lib->vba.RequiredDPPCLK[i][j][k] = mode_lib->vba.RequiredDPPCLKThisState[k];
+
+			mode_lib->vba.DISPCLK_DPPCLK_Support[i][j] = !((mode_lib->vba.RequiredDISPCLK[i][j]
+					> mode_lib->vba.MaxDispclk[i])
+					|| (mode_lib->vba.GlobalDPPCLK > mode_lib->vba.MaxDppclk[i]));
+
+			if (mode_lib->vba.TotalNumberOfActiveDPP[i][j] > mode_lib->vba.MaxNumDPP)
+				mode_lib->vba.TotalAvailablePipesSupport[i][j] = false;
+		} // j
+	} // i (VOLTAGE_STATE)
+
+	/* Total Available OTG, HDMIFRL, DP Support Check */
+	v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.TotalNumberOfActiveOTG = 0;
+	v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.TotalNumberOfActiveHDMIFRL = 0;
+	v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.TotalNumberOfActiveDP2p0 = 0;
+	v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.TotalNumberOfActiveDP2p0Outputs = 0;
+
+	for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+		if (mode_lib->vba.BlendingAndTiming[k] == k) {
+			v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.TotalNumberOfActiveOTG = v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.TotalNumberOfActiveOTG + 1;
+			if (mode_lib->vba.Output[k] == dm_dp2p0) {
+				v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.TotalNumberOfActiveDP2p0 = v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.TotalNumberOfActiveDP2p0 + 1;
+				if (mode_lib->vba.OutputMultistreamId[k]
+						== k || mode_lib->vba.OutputMultistreamEn[k] == false) {
+					v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.TotalNumberOfActiveDP2p0Outputs = v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.TotalNumberOfActiveDP2p0Outputs + 1;
+				}
+			}
+		}
+	}
+
+	mode_lib->vba.NumberOfOTGSupport = (v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.TotalNumberOfActiveOTG <= mode_lib->vba.MaxNumOTG);
+	mode_lib->vba.NumberOfHDMIFRLSupport = (v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.TotalNumberOfActiveHDMIFRL <= mode_lib->vba.MaxNumHDMIFRLOutputs);
+	mode_lib->vba.NumberOfDP2p0Support = (v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.TotalNumberOfActiveDP2p0 <= mode_lib->vba.MaxNumDP2p0Streams
+			&& v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.TotalNumberOfActiveDP2p0Outputs <= mode_lib->vba.MaxNumDP2p0Outputs);
+
+	/* Display IO and DSC Support Check */
+	mode_lib->vba.NonsupportedDSCInputBPC = false;
+	for (k = 0; k <= mode_lib->vba.NumberOfActiveSurfaces - 1; k++) {
+		if (!(mode_lib->vba.DSCInputBitPerComponent[k] == 12.0
+				|| mode_lib->vba.DSCInputBitPerComponent[k] == 10.0
+				|| mode_lib->vba.DSCInputBitPerComponent[k] == 8.0)
+				|| mode_lib->vba.DSCInputBitPerComponent[k] > mode_lib->vba.MaximumDSCBitsPerComponent) {
+			mode_lib->vba.NonsupportedDSCInputBPC = true;
+		}
+	}
+
+	for (i = start_state; i < v->soc.num_states; ++i) {
+		mode_lib->vba.ExceededMultistreamSlots[i] = false;
+		for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+			if (mode_lib->vba.OutputMultistreamEn[k] == true && mode_lib->vba.OutputMultistreamId[k] == k) {
+				TotalSlots = mode_lib->vba.RequiredSlots[i][k];
+				for (j = 0; j < mode_lib->vba.NumberOfActiveSurfaces; ++j) {
+					if (mode_lib->vba.OutputMultistreamId[j] == k)
+						TotalSlots = TotalSlots + mode_lib->vba.RequiredSlots[i][j];
+				}
+				if (mode_lib->vba.Output[k] == dm_dp && TotalSlots > 63)
+					mode_lib->vba.ExceededMultistreamSlots[i] = true;
+				if (mode_lib->vba.Output[k] == dm_dp2p0 && TotalSlots > 64)
+					mode_lib->vba.ExceededMultistreamSlots[i] = true;
+			}
+		}
+		mode_lib->vba.LinkCapacitySupport[i] = true;
+		for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+			if (mode_lib->vba.BlendingAndTiming[k] == k
+				&& (mode_lib->vba.Output[k] == dm_dp || mode_lib->vba.Output[k] == dm_dp2p0
+					|| mode_lib->vba.Output[k] == dm_edp
+					|| mode_lib->vba.Output[k] == dm_hdmi)
+				&& mode_lib->vba.OutputBppPerState[i][k] == 0 &&
+				(mode_lib->vba.UsesMALLForPStateChange[k] != dm_use_mall_pstate_change_phantom_pipe)) {
+				/* Phantom pipes don't consider DSC in DML, so it could fail link check.
+				 * However, we don't care about the link for phantom pipes.
+				 */
+				mode_lib->vba.LinkCapacitySupport[i] = false;
+			}
+		}
+	}
+
+	mode_lib->vba.P2IWith420 = false;
+	mode_lib->vba.DSCOnlyIfNecessaryWithBPP = false;
+	mode_lib->vba.DSC422NativeNotSupported = false;
+	mode_lib->vba.LinkRateDoesNotMatchDPVersion = false;
+	mode_lib->vba.LinkRateForMultistreamNotIndicated = false;
+	mode_lib->vba.BPPForMultistreamNotIndicated = false;
+	mode_lib->vba.MultistreamWithHDMIOreDP = false;
+	mode_lib->vba.MSOOrODMSplitWithNonDPLink = false;
+	mode_lib->vba.NotEnoughLanesForMSO = false;
+
+	for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+		if (mode_lib->vba.BlendingAndTiming[k] == k
+				&& (mode_lib->vba.Output[k] == dm_dp || mode_lib->vba.Output[k] == dm_dp2p0
+						|| mode_lib->vba.Output[k] == dm_edp
+						|| mode_lib->vba.Output[k] == dm_hdmi)) {
+			if (mode_lib->vba.OutputFormat[k]
+					== dm_420 && mode_lib->vba.Interlace[k] == 1 &&
+					mode_lib->vba.ProgressiveToInterlaceUnitInOPP == true)
+				mode_lib->vba.P2IWith420 = true;
+
+			if (mode_lib->vba.DSCEnable[k] && mode_lib->vba.ForcedOutputLinkBPP[k] != 0)
+				mode_lib->vba.DSCOnlyIfNecessaryWithBPP = true;
+			if (mode_lib->vba.DSCEnable[k] && mode_lib->vba.OutputFormat[k] == dm_n422
+					&& !mode_lib->vba.DSC422NativeSupport)
+				mode_lib->vba.DSC422NativeNotSupported = true;
+
+			if (((mode_lib->vba.OutputLinkDPRate[k] == dm_dp_rate_hbr
+					|| mode_lib->vba.OutputLinkDPRate[k] == dm_dp_rate_hbr2
+					|| mode_lib->vba.OutputLinkDPRate[k] == dm_dp_rate_hbr3)
+					&& mode_lib->vba.Output[k] != dm_dp && mode_lib->vba.Output[k] != dm_edp)
+					|| ((mode_lib->vba.OutputLinkDPRate[k] == dm_dp_rate_uhbr10
+							|| mode_lib->vba.OutputLinkDPRate[k] == dm_dp_rate_uhbr13p5
+							|| mode_lib->vba.OutputLinkDPRate[k] == dm_dp_rate_uhbr20)
+							&& mode_lib->vba.Output[k] != dm_dp2p0))
+				mode_lib->vba.LinkRateDoesNotMatchDPVersion = true;
+
+			if (mode_lib->vba.OutputMultistreamEn[k] == true) {
+				if (mode_lib->vba.OutputMultistreamId[k] == k
+					&& mode_lib->vba.OutputLinkDPRate[k] == dm_dp_rate_na)
+					mode_lib->vba.LinkRateForMultistreamNotIndicated = true;
+				if (mode_lib->vba.OutputMultistreamId[k] == k && mode_lib->vba.ForcedOutputLinkBPP[k] == 0)
+					mode_lib->vba.BPPForMultistreamNotIndicated = true;
+				for (j = 0; j < mode_lib->vba.NumberOfActiveSurfaces; ++j) {
+					if (mode_lib->vba.OutputMultistreamId[k] == j
+						&& mode_lib->vba.ForcedOutputLinkBPP[k] == 0)
+						mode_lib->vba.BPPForMultistreamNotIndicated = true;
+				}
+			}
+
+			if ((mode_lib->vba.Output[k] == dm_edp || mode_lib->vba.Output[k] == dm_hdmi)) {
+				if (mode_lib->vba.OutputMultistreamEn[k] == true && mode_lib->vba.OutputMultistreamId[k] == k)
+					mode_lib->vba.MultistreamWithHDMIOreDP = true;
+				for (j = 0; j < mode_lib->vba.NumberOfActiveSurfaces; ++j) {
+					if (mode_lib->vba.OutputMultistreamEn[k] == true && mode_lib->vba.OutputMultistreamId[k] == j)
+						mode_lib->vba.MultistreamWithHDMIOreDP = true;
+				}
+			}
+
+			if (mode_lib->vba.Output[k] != dm_dp
+					&& (mode_lib->vba.ODMUse[k] == dm_odm_split_policy_1to2
+							|| mode_lib->vba.ODMUse[k] == dm_odm_mso_policy_1to2
+							|| mode_lib->vba.ODMUse[k] == dm_odm_mso_policy_1to4))
+				mode_lib->vba.MSOOrODMSplitWithNonDPLink = true;
+
+			if ((mode_lib->vba.ODMUse[k] == dm_odm_mso_policy_1to2
+					&& mode_lib->vba.OutputLinkDPLanes[k] < 2)
+					|| (mode_lib->vba.ODMUse[k] == dm_odm_mso_policy_1to4
+							&& mode_lib->vba.OutputLinkDPLanes[k] < 4))
+				mode_lib->vba.NotEnoughLanesForMSO = true;
+		}
+	}
+
+	for (i = start_state; i < v->soc.num_states; ++i) {
+		mode_lib->vba.DTBCLKRequiredMoreThanSupported[i] = false;
+		for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+			if (mode_lib->vba.BlendingAndTiming[k] == k
+					&& dml32_RequiredDTBCLK(mode_lib->vba.RequiresDSC[i][k],
+							mode_lib->vba.PixelClockBackEnd[k],
+							mode_lib->vba.OutputFormat[k],
+							mode_lib->vba.OutputBppPerState[i][k],
+							mode_lib->vba.NumberOfDSCSlices[k], mode_lib->vba.HTotal[k],
+							mode_lib->vba.HActive[k], mode_lib->vba.AudioSampleRate[k],
+							mode_lib->vba.AudioSampleLayout[k])
+							> mode_lib->vba.DTBCLKPerState[i]) {
+				mode_lib->vba.DTBCLKRequiredMoreThanSupported[i] = true;
+			}
+		}
+	}
+
+	for (i = start_state; i < v->soc.num_states; ++i) {
+		mode_lib->vba.ODMCombine2To1SupportCheckOK[i] = true;
+		mode_lib->vba.ODMCombine4To1SupportCheckOK[i] = true;
+		for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+			if (mode_lib->vba.BlendingAndTiming[k] == k
+					&& mode_lib->vba.ODMCombineEnablePerState[i][k] == dm_odm_combine_mode_2to1
+					&& mode_lib->vba.Output[k] == dm_hdmi) {
+				mode_lib->vba.ODMCombine2To1SupportCheckOK[i] = false;
+			}
+			if (mode_lib->vba.BlendingAndTiming[k] == k
+					&& mode_lib->vba.ODMCombineEnablePerState[i][k] == dm_odm_combine_mode_4to1
+					&& (mode_lib->vba.Output[k] == dm_dp || mode_lib->vba.Output[k] == dm_edp
+							|| mode_lib->vba.Output[k] == dm_hdmi)) {
+				mode_lib->vba.ODMCombine4To1SupportCheckOK[i] = false;
+			}
+		}
+	}
+
+	for (i = start_state; i < v->soc.num_states; i++) {
+		mode_lib->vba.DSCCLKRequiredMoreThanSupported[i] = false;
+		for (k = 0; k <= mode_lib->vba.NumberOfActiveSurfaces - 1; k++) {
+			if (mode_lib->vba.BlendingAndTiming[k] == k) {
+				if (mode_lib->vba.Output[k] == dm_dp || mode_lib->vba.Output[k] == dm_dp2p0
+						|| mode_lib->vba.Output[k] == dm_edp) {
+					if (mode_lib->vba.OutputFormat[k] == dm_420) {
+						mode_lib->vba.DSCFormatFactor = 2;
+					} else if (mode_lib->vba.OutputFormat[k] == dm_444) {
+						mode_lib->vba.DSCFormatFactor = 1;
+					} else if (mode_lib->vba.OutputFormat[k] == dm_n422) {
+						mode_lib->vba.DSCFormatFactor = 2;
+					} else {
+						mode_lib->vba.DSCFormatFactor = 1;
+					}
+					if (mode_lib->vba.RequiresDSC[i][k] == true) {
+						if (mode_lib->vba.ODMCombineEnablePerState[i][k]
+								== dm_odm_combine_mode_4to1) {
+							if (mode_lib->vba.PixelClockBackEnd[k] / 12.0 / mode_lib->vba.DSCFormatFactor > (1.0 - mode_lib->vba.DISPCLKDPPCLKDSCCLKDownSpreading / 100.0) * mode_lib->vba.MaxDSCCLK[i])
+								mode_lib->vba.DSCCLKRequiredMoreThanSupported[i] = true;
+						} else if (mode_lib->vba.ODMCombineEnablePerState[i][k]
+								== dm_odm_combine_mode_2to1) {
+							if (mode_lib->vba.PixelClockBackEnd[k] / 6.0 / mode_lib->vba.DSCFormatFactor > (1.0 - mode_lib->vba.DISPCLKDPPCLKDSCCLKDownSpreading / 100.0) * mode_lib->vba.MaxDSCCLK[i])
+								mode_lib->vba.DSCCLKRequiredMoreThanSupported[i] = true;
+						} else {
+							if (mode_lib->vba.PixelClockBackEnd[k] / 3.0 / mode_lib->vba.DSCFormatFactor > (1.0 - mode_lib->vba.DISPCLKDPPCLKDSCCLKDownSpreading / 100.0) * mode_lib->vba.MaxDSCCLK[i])
+								mode_lib->vba.DSCCLKRequiredMoreThanSupported[i] = true;
+						}
+					}
+				}
+			}
+		}
+	}
+
+	/* Check DSC Unit and Slices Support */
+	v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.TotalDSCUnitsRequired = 0;
+
+	for (i = start_state; i < v->soc.num_states; ++i) {
+		mode_lib->vba.NotEnoughDSCUnits[i] = false;
+		mode_lib->vba.NotEnoughDSCSlices[i] = false;
+		v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.TotalDSCUnitsRequired = 0;
+		mode_lib->vba.PixelsPerLinePerDSCUnitSupport[i] = true;
+		for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+			if (mode_lib->vba.RequiresDSC[i][k] == true) {
+				if (mode_lib->vba.ODMCombineEnablePerState[i][k] == dm_odm_combine_mode_4to1) {
+					if (mode_lib->vba.HActive[k]
+							> 4 * mode_lib->vba.MaximumPixelsPerLinePerDSCUnit)
+						mode_lib->vba.PixelsPerLinePerDSCUnitSupport[i] = false;
+					v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.TotalDSCUnitsRequired = v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.TotalDSCUnitsRequired + 4;
+					if (mode_lib->vba.NumberOfDSCSlices[k] > 16)
+						mode_lib->vba.NotEnoughDSCSlices[i] = true;
+				} else if (mode_lib->vba.ODMCombineEnablePerState[i][k] == dm_odm_combine_mode_2to1) {
+					if (mode_lib->vba.HActive[k]
+							> 2 * mode_lib->vba.MaximumPixelsPerLinePerDSCUnit)
+						mode_lib->vba.PixelsPerLinePerDSCUnitSupport[i] = false;
+					v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.TotalDSCUnitsRequired = v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.TotalDSCUnitsRequired + 2;
+					if (mode_lib->vba.NumberOfDSCSlices[k] > 8)
+						mode_lib->vba.NotEnoughDSCSlices[i] = true;
+				} else {
+					if (mode_lib->vba.HActive[k] > mode_lib->vba.MaximumPixelsPerLinePerDSCUnit)
+						mode_lib->vba.PixelsPerLinePerDSCUnitSupport[i] = false;
+					v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.TotalDSCUnitsRequired = v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.TotalDSCUnitsRequired + 1;
+					if (mode_lib->vba.NumberOfDSCSlices[k] > 4)
+						mode_lib->vba.NotEnoughDSCSlices[i] = true;
+				}
+			}
+		}
+		if (v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.TotalDSCUnitsRequired > mode_lib->vba.NumberOfDSC)
+			mode_lib->vba.NotEnoughDSCUnits[i] = true;
+	}
+
+	/*DSC Delay per state*/
+	for (i = start_state; i < v->soc.num_states; ++i) {
+		for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+			mode_lib->vba.DSCDelayPerState[i][k] = dml32_DSCDelayRequirement(
+					mode_lib->vba.RequiresDSC[i][k], mode_lib->vba.ODMCombineEnablePerState[i][k],
+					mode_lib->vba.DSCInputBitPerComponent[k],
+					mode_lib->vba.OutputBppPerState[i][k], mode_lib->vba.HActive[k],
+					mode_lib->vba.HTotal[k], mode_lib->vba.NumberOfDSCSlices[k],
+					mode_lib->vba.OutputFormat[k], mode_lib->vba.Output[k],
+					mode_lib->vba.PixelClock[k], mode_lib->vba.PixelClockBackEnd[k],
+					mode_lib->vba.ip.dsc_delay_factor_wa);
+		}
+
+		for (k = 0; k <= mode_lib->vba.NumberOfActiveSurfaces - 1; k++) {
+			for (m = 0; m <= mode_lib->vba.NumberOfActiveSurfaces - 1; m++) {
+				for (j = 0; j <= mode_lib->vba.NumberOfActiveSurfaces - 1; j++) {
+					if (mode_lib->vba.BlendingAndTiming[k] == m &&
+							mode_lib->vba.RequiresDSC[i][m] == true) {
+						mode_lib->vba.DSCDelayPerState[i][k] =
+							mode_lib->vba.DSCDelayPerState[i][m];
+					}
+				}
+			}
+		}
+	}
+
+	//Calculate Swath, DET Configuration, DCFCLKDeepSleep
+	//
+	for (i = start_state; i < (int) v->soc.num_states; ++i) {
+		for (j = 0; j <= 1; ++j) {
+			for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+				mode_lib->vba.RequiredDPPCLKThisState[k] = mode_lib->vba.RequiredDPPCLK[i][j][k];
+				mode_lib->vba.NoOfDPPThisState[k] = mode_lib->vba.NoOfDPP[i][j][k];
+				mode_lib->vba.ODMCombineEnableThisState[k] =
+						mode_lib->vba.ODMCombineEnablePerState[i][k];
+			}
+
+			dml32_CalculateSwathAndDETConfiguration(
+					mode_lib->vba.DETSizeOverride,
+					mode_lib->vba.UsesMALLForPStateChange,
+					mode_lib->vba.ConfigReturnBufferSizeInKByte,
+					mode_lib->vba.MaxTotalDETInKByte,
+					mode_lib->vba.MinCompressedBufferSizeInKByte,
+					false, /* ForceSingleDPP */
+					mode_lib->vba.NumberOfActiveSurfaces,
+					mode_lib->vba.nomDETInKByte,
+					mode_lib->vba.UseUnboundedRequesting,
+					mode_lib->vba.DisableUnboundRequestIfCompBufReservedSpaceNeedAdjustment,
+					mode_lib->vba.ip.pixel_chunk_size_kbytes,
+					mode_lib->vba.ip.rob_buffer_size_kbytes,
+					mode_lib->vba.CompressedBufferSegmentSizeInkByteFinal,
+					mode_lib->vba.Output,
+					mode_lib->vba.ReadBandwidthLuma,
+					mode_lib->vba.ReadBandwidthChroma,
+					mode_lib->vba.MaximumSwathWidthLuma,
+					mode_lib->vba.MaximumSwathWidthChroma,
+					mode_lib->vba.SourceRotation,
+					mode_lib->vba.ViewportStationary,
+					mode_lib->vba.SourcePixelFormat,
+					mode_lib->vba.SurfaceTiling,
+					mode_lib->vba.ViewportWidth,
+					mode_lib->vba.ViewportHeight,
+					mode_lib->vba.ViewportXStartY,
+					mode_lib->vba.ViewportYStartY,
+					mode_lib->vba.ViewportXStartC,
+					mode_lib->vba.ViewportYStartC,
+					mode_lib->vba.SurfaceWidthY,
+					mode_lib->vba.SurfaceWidthC,
+					mode_lib->vba.SurfaceHeightY,
+					mode_lib->vba.SurfaceHeightC,
+					mode_lib->vba.Read256BlockHeightY,
+					mode_lib->vba.Read256BlockHeightC,
+					mode_lib->vba.Read256BlockWidthY,
+					mode_lib->vba.Read256BlockWidthC,
+					mode_lib->vba.ODMCombineEnableThisState,
+					mode_lib->vba.BlendingAndTiming,
+					mode_lib->vba.BytePerPixelY,
+					mode_lib->vba.BytePerPixelC,
+					mode_lib->vba.BytePerPixelInDETY,
+					mode_lib->vba.BytePerPixelInDETC,
+					mode_lib->vba.HActive,
+					mode_lib->vba.HRatio,
+					mode_lib->vba.HRatioChroma,
+					mode_lib->vba.NoOfDPPThisState,
+					/* Output */
+					mode_lib->vba.swath_width_luma_ub_this_state,
+					mode_lib->vba.swath_width_chroma_ub_this_state,
+					mode_lib->vba.SwathWidthYThisState,
+					mode_lib->vba.SwathWidthCThisState,
+					mode_lib->vba.SwathHeightYThisState,
+					mode_lib->vba.SwathHeightCThisState,
+					mode_lib->vba.DETBufferSizeInKByteThisState,
+					mode_lib->vba.DETBufferSizeYThisState,
+					mode_lib->vba.DETBufferSizeCThisState,
+					&mode_lib->vba.UnboundedRequestEnabledThisState,
+					&mode_lib->vba.CompressedBufferSizeInkByteThisState,
+					&v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_integer[0], /* Long CompBufReservedSpaceKBytes */
+					&v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_boolean[0], /* bool CompBufReservedSpaceNeedAdjustment */
+					v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_boolean_array[0],
+					&mode_lib->vba.ViewportSizeSupport[i][j]);
+
+			for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+				mode_lib->vba.swath_width_luma_ub_all_states[i][j][k] =
+						mode_lib->vba.swath_width_luma_ub_this_state[k];
+				mode_lib->vba.swath_width_chroma_ub_all_states[i][j][k] =
+						mode_lib->vba.swath_width_chroma_ub_this_state[k];
+				mode_lib->vba.SwathWidthYAllStates[i][j][k] = mode_lib->vba.SwathWidthYThisState[k];
+				mode_lib->vba.SwathWidthCAllStates[i][j][k] = mode_lib->vba.SwathWidthCThisState[k];
+				mode_lib->vba.SwathHeightYAllStates[i][j][k] = mode_lib->vba.SwathHeightYThisState[k];
+				mode_lib->vba.SwathHeightCAllStates[i][j][k] = mode_lib->vba.SwathHeightCThisState[k];
+				mode_lib->vba.UnboundedRequestEnabledAllStates[i][j] =
+						mode_lib->vba.UnboundedRequestEnabledThisState;
+				mode_lib->vba.CompressedBufferSizeInkByteAllStates[i][j] =
+						mode_lib->vba.CompressedBufferSizeInkByteThisState;
+				mode_lib->vba.DETBufferSizeInKByteAllStates[i][j][k] =
+						mode_lib->vba.DETBufferSizeInKByteThisState[k];
+				mode_lib->vba.DETBufferSizeYAllStates[i][j][k] =
+						mode_lib->vba.DETBufferSizeYThisState[k];
+				mode_lib->vba.DETBufferSizeCAllStates[i][j][k] =
+						mode_lib->vba.DETBufferSizeCThisState[k];
+			}
+		}
+	}
+
+	for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+		mode_lib->vba.cursor_bw[k] = mode_lib->vba.NumberOfCursors[k] * mode_lib->vba.CursorWidth[k][0]
+				* mode_lib->vba.CursorBPP[k][0] / 8.0
+				/ (mode_lib->vba.HTotal[k] / mode_lib->vba.PixelClock[k]) * mode_lib->vba.VRatio[k];
+	}
+
+	dml32_CalculateSurfaceSizeInMall(
+			mode_lib->vba.NumberOfActiveSurfaces,
+			mode_lib->vba.MALLAllocatedForDCNFinal,
+			mode_lib->vba.UseMALLForStaticScreen,
+			mode_lib->vba.UsesMALLForPStateChange,
+			mode_lib->vba.DCCEnable,
+			mode_lib->vba.ViewportStationary,
+			mode_lib->vba.ViewportXStartY,
+			mode_lib->vba.ViewportYStartY,
+			mode_lib->vba.ViewportXStartC,
+			mode_lib->vba.ViewportYStartC,
+			mode_lib->vba.ViewportWidth,
+			mode_lib->vba.ViewportHeight,
+			mode_lib->vba.BytePerPixelY,
+			mode_lib->vba.ViewportWidthChroma,
+			mode_lib->vba.ViewportHeightChroma,
+			mode_lib->vba.BytePerPixelC,
+			mode_lib->vba.SurfaceWidthY,
+			mode_lib->vba.SurfaceWidthC,
+			mode_lib->vba.SurfaceHeightY,
+			mode_lib->vba.SurfaceHeightC,
+			mode_lib->vba.Read256BlockWidthY,
+			mode_lib->vba.Read256BlockWidthC,
+			mode_lib->vba.Read256BlockHeightY,
+			mode_lib->vba.Read256BlockHeightC,
+			mode_lib->vba.MacroTileWidthY,
+			mode_lib->vba.MacroTileWidthC,
+			mode_lib->vba.MacroTileHeightY,
+			mode_lib->vba.MacroTileHeightC,
+			mode_lib->vba.DCCMetaPitchY,
+			mode_lib->vba.DCCMetaPitchC,
+
+			/* Output */
+			mode_lib->vba.SurfaceSizeInMALL,
+			&mode_lib->vba.ExceededMALLSize);
+
+	for (i = start_state; i < v->soc.num_states; i++) {
+		for (j = 0; j < 2; j++) {
+			for (k = 0; k <= mode_lib->vba.NumberOfActiveSurfaces - 1; k++) {
+				mode_lib->vba.swath_width_luma_ub_this_state[k] =
+						mode_lib->vba.swath_width_luma_ub_all_states[i][j][k];
+				mode_lib->vba.swath_width_chroma_ub_this_state[k] =
+						mode_lib->vba.swath_width_chroma_ub_all_states[i][j][k];
+				mode_lib->vba.SwathWidthYThisState[k] = mode_lib->vba.SwathWidthYAllStates[i][j][k];
+				mode_lib->vba.SwathWidthCThisState[k] = mode_lib->vba.SwathWidthCAllStates[i][j][k];
+				mode_lib->vba.SwathHeightYThisState[k] = mode_lib->vba.SwathHeightYAllStates[i][j][k];
+				mode_lib->vba.SwathHeightCThisState[k] = mode_lib->vba.SwathHeightCAllStates[i][j][k];
+				mode_lib->vba.DETBufferSizeInKByteThisState[k] =
+						mode_lib->vba.DETBufferSizeInKByteAllStates[i][j][k];
+				mode_lib->vba.DETBufferSizeYThisState[k] =
+						mode_lib->vba.DETBufferSizeYAllStates[i][j][k];
+				mode_lib->vba.DETBufferSizeCThisState[k] =
+						mode_lib->vba.DETBufferSizeCAllStates[i][j][k];
+				mode_lib->vba.RequiredDPPCLKThisState[k] = mode_lib->vba.RequiredDPPCLK[i][j][k];
+				mode_lib->vba.NoOfDPPThisState[k] = mode_lib->vba.NoOfDPP[i][j][k];
+			}
+
+			mode_lib->vba.TotalNumberOfDCCActiveDPP[i][j] = 0;
+			for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+				if (mode_lib->vba.DCCEnable[k] == true) {
+					mode_lib->vba.TotalNumberOfDCCActiveDPP[i][j] =
+							mode_lib->vba.TotalNumberOfDCCActiveDPP[i][j]
+									+ mode_lib->vba.NoOfDPP[i][j][k];
+				}
+			}
+
+
+			for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+				v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.SurfParameters[k].PixelClock = mode_lib->vba.PixelClock[k];
+				v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.SurfParameters[k].DPPPerSurface = mode_lib->vba.NoOfDPP[i][j][k];
+				v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.SurfParameters[k].SourceRotation = mode_lib->vba.SourceRotation[k];
+				v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.SurfParameters[k].ViewportHeight = mode_lib->vba.ViewportHeight[k];
+				v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.SurfParameters[k].ViewportHeightChroma = mode_lib->vba.ViewportHeightChroma[k];
+				v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.SurfParameters[k].BlockWidth256BytesY = mode_lib->vba.Read256BlockWidthY[k];
+				v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.SurfParameters[k].BlockHeight256BytesY = mode_lib->vba.Read256BlockHeightY[k];
+				v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.SurfParameters[k].BlockWidth256BytesC = mode_lib->vba.Read256BlockWidthC[k];
+				v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.SurfParameters[k].BlockHeight256BytesC = mode_lib->vba.Read256BlockHeightC[k];
+				v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.SurfParameters[k].BlockWidthY = mode_lib->vba.MacroTileWidthY[k];
+				v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.SurfParameters[k].BlockHeightY = mode_lib->vba.MacroTileHeightY[k];
+				v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.SurfParameters[k].BlockWidthC = mode_lib->vba.MacroTileWidthC[k];
+				v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.SurfParameters[k].BlockHeightC = mode_lib->vba.MacroTileHeightC[k];
+				v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.SurfParameters[k].InterlaceEnable = mode_lib->vba.Interlace[k];
+				v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.SurfParameters[k].HTotal = mode_lib->vba.HTotal[k];
+				v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.SurfParameters[k].DCCEnable = mode_lib->vba.DCCEnable[k];
+				v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.SurfParameters[k].SourcePixelFormat = mode_lib->vba.SourcePixelFormat[k];
+				v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.SurfParameters[k].SurfaceTiling = mode_lib->vba.SurfaceTiling[k];
+				v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.SurfParameters[k].BytePerPixelY = mode_lib->vba.BytePerPixelY[k];
+				v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.SurfParameters[k].BytePerPixelC = mode_lib->vba.BytePerPixelC[k];
+				v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.SurfParameters[k].ProgressiveToInterlaceUnitInOPP =
+				mode_lib->vba.ProgressiveToInterlaceUnitInOPP;
+				v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.SurfParameters[k].VRatio = mode_lib->vba.VRatio[k];
+				v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.SurfParameters[k].VRatioChroma = mode_lib->vba.VRatioChroma[k];
+				v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.SurfParameters[k].VTaps = mode_lib->vba.vtaps[k];
+				v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.SurfParameters[k].VTapsChroma = mode_lib->vba.VTAPsChroma[k];
+				v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.SurfParameters[k].PitchY = mode_lib->vba.PitchY[k];
+				v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.SurfParameters[k].DCCMetaPitchY = mode_lib->vba.DCCMetaPitchY[k];
+				v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.SurfParameters[k].PitchC = mode_lib->vba.PitchC[k];
+				v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.SurfParameters[k].DCCMetaPitchC = mode_lib->vba.DCCMetaPitchC[k];
+				v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.SurfParameters[k].ViewportStationary = mode_lib->vba.ViewportStationary[k];
+				v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.SurfParameters[k].ViewportXStart = mode_lib->vba.ViewportXStartY[k];
+				v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.SurfParameters[k].ViewportYStart = mode_lib->vba.ViewportYStartY[k];
+				v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.SurfParameters[k].ViewportXStartC = mode_lib->vba.ViewportXStartC[k];
+				v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.SurfParameters[k].ViewportYStartC = mode_lib->vba.ViewportYStartC[k];
+				v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.SurfParameters[k].FORCE_ONE_ROW_FOR_FRAME = mode_lib->vba.ForceOneRowForFrame[k];
+				v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.SurfParameters[k].SwathHeightY = mode_lib->vba.SwathHeightYThisState[k];
+				v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.SurfParameters[k].SwathHeightC = mode_lib->vba.SwathHeightCThisState[k];
+			}
+
+			{
+				dml32_CalculateVMRowAndSwath(
+						mode_lib->vba.NumberOfActiveSurfaces,
+						v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.SurfParameters,
+						mode_lib->vba.SurfaceSizeInMALL,
+						mode_lib->vba.PTEBufferSizeInRequestsLuma,
+						mode_lib->vba.PTEBufferSizeInRequestsChroma,
+						mode_lib->vba.DCCMetaBufferSizeBytes,
+						mode_lib->vba.UseMALLForStaticScreen,
+						mode_lib->vba.UsesMALLForPStateChange,
+						mode_lib->vba.MALLAllocatedForDCNFinal,
+						mode_lib->vba.SwathWidthYThisState,
+						mode_lib->vba.SwathWidthCThisState,
+						mode_lib->vba.GPUVMEnable,
+						mode_lib->vba.HostVMEnable,
+						mode_lib->vba.HostVMMaxNonCachedPageTableLevels,
+						mode_lib->vba.GPUVMMaxPageTableLevels,
+						mode_lib->vba.GPUVMMinPageSizeKBytes,
+						mode_lib->vba.HostVMMinPageSize,
+
+						/* Output */
+						mode_lib->vba.PTEBufferSizeNotExceededPerState,
+						mode_lib->vba.DCCMetaBufferSizeNotExceededPerState,
+						v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_integer_array[0],
+						v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_integer_array[1],
+						mode_lib->vba.dpte_row_height,
+						mode_lib->vba.dpte_row_height_chroma,
+						v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_integer_array[2],
+						v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_integer_array[3],
+						v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_integer_array[4],
+						v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_integer_array[5],
+						v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_integer_array[6],
+						v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_integer_array[7],
+						v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_integer_array[8],
+						v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_integer_array[9],
+						mode_lib->vba.meta_row_height,
+						mode_lib->vba.meta_row_height_chroma,
+						v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_integer_array[10],
+						mode_lib->vba.dpte_group_bytes,
+						v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_integer_array[11],
+						v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_integer_array[12],
+						v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_integer_array[13],
+						v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_integer_array[14],
+						v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_integer_array[15],
+						v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_integer_array[16],
+						v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_integer_array[17],
+						v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_integer_array[18],
+						v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_integer_array[19],
+						v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_integer_array[20],
+						mode_lib->vba.PrefetchLinesYThisState,
+						mode_lib->vba.PrefetchLinesCThisState,
+						mode_lib->vba.PrefillY,
+						mode_lib->vba.PrefillC,
+						mode_lib->vba.MaxNumSwY,
+						mode_lib->vba.MaxNumSwC,
+						mode_lib->vba.meta_row_bandwidth_this_state,
+						mode_lib->vba.dpte_row_bandwidth_this_state,
+						mode_lib->vba.DPTEBytesPerRowThisState,
+						mode_lib->vba.PDEAndMetaPTEBytesPerFrameThisState,
+						mode_lib->vba.MetaRowBytesThisState,
+						mode_lib->vba.use_one_row_for_frame_this_state,
+						mode_lib->vba.use_one_row_for_frame_flip_this_state,
+						v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_boolean_array[0], // Boolean UsesMALLForStaticScreen[]
+						v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_boolean_array[1], // Boolean PTE_BUFFER_MODE[]
+						v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_integer_array[21]); // Long BIGK_FRAGMENT_SIZE[]
+			}
+
+			for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+				mode_lib->vba.PrefetchLinesY[i][j][k] = mode_lib->vba.PrefetchLinesYThisState[k];
+				mode_lib->vba.PrefetchLinesC[i][j][k] = mode_lib->vba.PrefetchLinesCThisState[k];
+				mode_lib->vba.meta_row_bandwidth[i][j][k] =
+						mode_lib->vba.meta_row_bandwidth_this_state[k];
+				mode_lib->vba.dpte_row_bandwidth[i][j][k] =
+						mode_lib->vba.dpte_row_bandwidth_this_state[k];
+				mode_lib->vba.DPTEBytesPerRow[i][j][k] = mode_lib->vba.DPTEBytesPerRowThisState[k];
+				mode_lib->vba.PDEAndMetaPTEBytesPerFrame[i][j][k] =
+						mode_lib->vba.PDEAndMetaPTEBytesPerFrameThisState[k];
+				mode_lib->vba.MetaRowBytes[i][j][k] = mode_lib->vba.MetaRowBytesThisState[k];
+				mode_lib->vba.use_one_row_for_frame[i][j][k] =
+						mode_lib->vba.use_one_row_for_frame_this_state[k];
+				mode_lib->vba.use_one_row_for_frame_flip[i][j][k] =
+						mode_lib->vba.use_one_row_for_frame_flip_this_state[k];
+			}
+
+			mode_lib->vba.PTEBufferSizeNotExceeded[i][j] = true;
+			for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+				if (mode_lib->vba.PTEBufferSizeNotExceededPerState[k] == false)
+					mode_lib->vba.PTEBufferSizeNotExceeded[i][j] = false;
+			}
+
+			mode_lib->vba.DCCMetaBufferSizeNotExceeded[i][j] = true;
+			for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+				if (mode_lib->vba.DCCMetaBufferSizeNotExceededPerState[k] == false)
+					mode_lib->vba.DCCMetaBufferSizeNotExceeded[i][j] = false;
+			}
+
+			mode_lib->vba.UrgLatency[i] = dml32_CalculateUrgentLatency(
+					mode_lib->vba.UrgentLatencyPixelDataOnly,
+					mode_lib->vba.UrgentLatencyPixelMixedWithVMData,
+					mode_lib->vba.UrgentLatencyVMDataOnly, mode_lib->vba.DoUrgentLatencyAdjustment,
+					mode_lib->vba.UrgentLatencyAdjustmentFabricClockComponent,
+					mode_lib->vba.UrgentLatencyAdjustmentFabricClockReference,
+					mode_lib->vba.FabricClockPerState[i]);
+
+			//bool   NotUrgentLatencyHiding[DC__NUM_DPP__MAX];
+			for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+				dml32_CalculateUrgentBurstFactor(
+						mode_lib->vba.UsesMALLForPStateChange[k],
+						mode_lib->vba.swath_width_luma_ub_this_state[k],
+						mode_lib->vba.swath_width_chroma_ub_this_state[k],
+						mode_lib->vba.SwathHeightYThisState[k],
+						mode_lib->vba.SwathHeightCThisState[k],
+						(double) mode_lib->vba.HTotal[k] / mode_lib->vba.PixelClock[k],
+						mode_lib->vba.UrgLatency[i],
+						mode_lib->vba.CursorBufferSize,
+						mode_lib->vba.CursorWidth[k][0],
+						mode_lib->vba.CursorBPP[k][0],
+						mode_lib->vba.VRatio[k],
+						mode_lib->vba.VRatioChroma[k],
+						mode_lib->vba.BytePerPixelInDETY[k],
+						mode_lib->vba.BytePerPixelInDETC[k],
+						mode_lib->vba.DETBufferSizeYThisState[k],
+						mode_lib->vba.DETBufferSizeCThisState[k],
+						/* Output */
+						&mode_lib->vba.UrgentBurstFactorCursor[k],
+						&mode_lib->vba.UrgentBurstFactorLuma[k],
+						&mode_lib->vba.UrgentBurstFactorChroma[k],
+						&mode_lib->vba.NoUrgentLatencyHiding[k]);
+			}
+
+			dml32_CalculateDCFCLKDeepSleep(
+					mode_lib->vba.NumberOfActiveSurfaces,
+					mode_lib->vba.BytePerPixelY,
+					mode_lib->vba.BytePerPixelC,
+					mode_lib->vba.VRatio,
+					mode_lib->vba.VRatioChroma,
+					mode_lib->vba.SwathWidthYThisState,
+					mode_lib->vba.SwathWidthCThisState,
+					mode_lib->vba.NoOfDPPThisState,
+					mode_lib->vba.HRatio,
+					mode_lib->vba.HRatioChroma,
+					mode_lib->vba.PixelClock,
+					mode_lib->vba.PSCL_FACTOR,
+					mode_lib->vba.PSCL_FACTOR_CHROMA,
+					mode_lib->vba.RequiredDPPCLKThisState,
+					mode_lib->vba.ReadBandwidthLuma,
+					mode_lib->vba.ReadBandwidthChroma,
+					mode_lib->vba.ReturnBusWidth,
+
+					/* Output */
+					&mode_lib->vba.ProjectedDCFCLKDeepSleep[i][j]);
+		}
+	}
+
+	//Calculate Return BW
+	for (i = start_state; i < (int) v->soc.num_states; ++i) {
+		for (j = 0; j <= 1; ++j) {
+			for (k = 0; k <= mode_lib->vba.NumberOfActiveSurfaces - 1; k++) {
+				if (mode_lib->vba.BlendingAndTiming[k] == k) {
+					if (mode_lib->vba.WritebackEnable[k] == true) {
+						mode_lib->vba.WritebackDelayTime[k] =
+							mode_lib->vba.WritebackLatency
+						+ dml32_CalculateWriteBackDelay(
+							mode_lib->vba.WritebackPixelFormat[k],
+							mode_lib->vba.WritebackHRatio[k],
+							mode_lib->vba.WritebackVRatio[k],
+							mode_lib->vba.WritebackVTaps[k],
+							mode_lib->vba.WritebackDestinationWidth[k],
+							mode_lib->vba.WritebackDestinationHeight[k],
+							mode_lib->vba.WritebackSourceHeight[k],
+							mode_lib->vba.HTotal[k])
+							/ mode_lib->vba.RequiredDISPCLK[i][j];
+					} else {
+						mode_lib->vba.WritebackDelayTime[k] = 0.0;
+					}
+					for (m = 0; m <= mode_lib->vba.NumberOfActiveSurfaces - 1; m++) {
+						if (mode_lib->vba.BlendingAndTiming[m]
+								== k && mode_lib->vba.WritebackEnable[m] == true) {
+							mode_lib->vba.WritebackDelayTime[k] =
+								dml_max(mode_lib->vba.WritebackDelayTime[k],
+									mode_lib->vba.WritebackLatency
+								+ dml32_CalculateWriteBackDelay(
+									mode_lib->vba.WritebackPixelFormat[m],
+									mode_lib->vba.WritebackHRatio[m],
+									mode_lib->vba.WritebackVRatio[m],
+									mode_lib->vba.WritebackVTaps[m],
+									mode_lib->vba.WritebackDestinationWidth[m],
+									mode_lib->vba.WritebackDestinationHeight[m],
+									mode_lib->vba.WritebackSourceHeight[m],
+									mode_lib->vba.HTotal[m]) /
+									mode_lib->vba.RequiredDISPCLK[i][j]);
+						}
+					}
+				}
+			}
+			for (k = 0; k <= mode_lib->vba.NumberOfActiveSurfaces - 1; k++) {
+				for (m = 0; m <= mode_lib->vba.NumberOfActiveSurfaces - 1; m++) {
+					if (mode_lib->vba.BlendingAndTiming[k] == m) {
+						mode_lib->vba.WritebackDelayTime[k] =
+								mode_lib->vba.WritebackDelayTime[m];
+					}
+				}
+			}
+			mode_lib->vba.MaxMaxVStartup[i][j] = 0;
+			for (k = 0; k <= mode_lib->vba.NumberOfActiveSurfaces - 1; k++) {
+				mode_lib->vba.MaximumVStartup[i][j][k] = ((mode_lib->vba.Interlace[k] &&
+								!mode_lib->vba.ProgressiveToInterlaceUnitInOPP) ?
+								dml_floor((mode_lib->vba.VTotal[k] -
+									mode_lib->vba.VActive[k]) / 2.0, 1.0) :
+								mode_lib->vba.VTotal[k] - mode_lib->vba.VActive[k])
+								- dml_max(1.0, dml_ceil(1.0 *
+									mode_lib->vba.WritebackDelayTime[k] /
+									(mode_lib->vba.HTotal[k] /
+									mode_lib->vba.PixelClock[k]), 1.0));
+
+				// Clamp to max OTG vstartup register limit
+				if (mode_lib->vba.MaximumVStartup[i][j][k] > 1023)
+					mode_lib->vba.MaximumVStartup[i][j][k] = 1023;
+
+				mode_lib->vba.MaxMaxVStartup[i][j] = dml_max(mode_lib->vba.MaxMaxVStartup[i][j],
+						mode_lib->vba.MaximumVStartup[i][j][k]);
+			}
+		}
+	}
+
+	v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.ReorderingBytes = mode_lib->vba.NumberOfChannels
+			* dml_max3(mode_lib->vba.UrgentOutOfOrderReturnPerChannelPixelDataOnly,
+					mode_lib->vba.UrgentOutOfOrderReturnPerChannelPixelMixedWithVMData,
+					mode_lib->vba.UrgentOutOfOrderReturnPerChannelVMDataOnly);
+
+	dml32_CalculateMinAndMaxPrefetchMode(mode_lib->vba.AllowForPStateChangeOrStutterInVBlankFinal,
+			&mode_lib->vba.MinPrefetchMode,
+			&mode_lib->vba.MaxPrefetchMode);
+
+	for (i = start_state; i < (int) v->soc.num_states; ++i) {
+		for (j = 0; j <= 1; ++j)
+			mode_lib->vba.DCFCLKState[i][j] = mode_lib->vba.DCFCLKPerState[i];
+	}
+
+	/* Immediate Flip and MALL parameters */
+	mode_lib->vba.ImmediateFlipRequiredFinal = false;
+	for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+		mode_lib->vba.ImmediateFlipRequiredFinal = mode_lib->vba.ImmediateFlipRequiredFinal
+				|| (mode_lib->vba.ImmediateFlipRequirement[k] == dm_immediate_flip_required);
+	}
+
+	mode_lib->vba.ImmediateFlipRequiredButTheRequirementForEachSurfaceIsNotSpecified = false;
+	for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+		mode_lib->vba.ImmediateFlipRequiredButTheRequirementForEachSurfaceIsNotSpecified =
+				mode_lib->vba.ImmediateFlipRequiredButTheRequirementForEachSurfaceIsNotSpecified
+						|| ((mode_lib->vba.ImmediateFlipRequirement[k]
+								!= dm_immediate_flip_required)
+								&& (mode_lib->vba.ImmediateFlipRequirement[k]
+										!= dm_immediate_flip_not_required));
+	}
+	mode_lib->vba.ImmediateFlipRequiredButTheRequirementForEachSurfaceIsNotSpecified =
+			mode_lib->vba.ImmediateFlipRequiredButTheRequirementForEachSurfaceIsNotSpecified
+					&& mode_lib->vba.ImmediateFlipRequiredFinal;
+
+	mode_lib->vba.ImmediateFlipOrHostVMAndPStateWithMALLFullFrameOrPhantomPipe = false;
+	for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+		mode_lib->vba.ImmediateFlipOrHostVMAndPStateWithMALLFullFrameOrPhantomPipe =
+			mode_lib->vba.ImmediateFlipOrHostVMAndPStateWithMALLFullFrameOrPhantomPipe ||
+			((mode_lib->vba.HostVMEnable == true || mode_lib->vba.ImmediateFlipRequirement[k] !=
+					dm_immediate_flip_not_required) &&
+			(mode_lib->vba.UsesMALLForPStateChange[k] == dm_use_mall_pstate_change_full_frame ||
+			mode_lib->vba.UsesMALLForPStateChange[k] == dm_use_mall_pstate_change_phantom_pipe));
+	}
+
+	mode_lib->vba.InvalidCombinationOfMALLUseForPStateAndStaticScreen = false;
+	for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+		mode_lib->vba.InvalidCombinationOfMALLUseForPStateAndStaticScreen =
+			mode_lib->vba.InvalidCombinationOfMALLUseForPStateAndStaticScreen
+			|| ((mode_lib->vba.UseMALLForStaticScreen[k] == dm_use_mall_static_screen_enable
+			|| mode_lib->vba.UseMALLForStaticScreen[k] == dm_use_mall_static_screen_optimize)
+			&& (mode_lib->vba.UsesMALLForPStateChange[k] == dm_use_mall_pstate_change_phantom_pipe))
+			|| ((mode_lib->vba.UseMALLForStaticScreen[k] == dm_use_mall_static_screen_disable
+			|| mode_lib->vba.UseMALLForStaticScreen[k] == dm_use_mall_static_screen_optimize)
+			&& (mode_lib->vba.UsesMALLForPStateChange[k] == dm_use_mall_pstate_change_full_frame));
+	}
+
+	v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.FullFrameMALLPStateMethod = false;
+	v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.SubViewportMALLPStateMethod = false;
+	v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.PhantomPipeMALLPStateMethod = false;
+
+	for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+		if (mode_lib->vba.UsesMALLForPStateChange[k] == dm_use_mall_pstate_change_full_frame)
+			v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.FullFrameMALLPStateMethod = true;
+		if (mode_lib->vba.UsesMALLForPStateChange[k] == dm_use_mall_pstate_change_sub_viewport)
+			v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.SubViewportMALLPStateMethod = true;
+		if (mode_lib->vba.UsesMALLForPStateChange[k] == dm_use_mall_pstate_change_phantom_pipe)
+			v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.PhantomPipeMALLPStateMethod = true;
+	}
+	mode_lib->vba.InvalidCombinationOfMALLUseForPState = (v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.SubViewportMALLPStateMethod
+			!= v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.PhantomPipeMALLPStateMethod) || (v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.SubViewportMALLPStateMethod && v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.FullFrameMALLPStateMethod);
+
+	if (mode_lib->vba.UseMinimumRequiredDCFCLK == true) {
+		dml32_UseMinimumDCFCLK(
+				mode_lib->vba.UsesMALLForPStateChange,
+				mode_lib->vba.DRRDisplay,
+				mode_lib->vba.SynchronizeDRRDisplaysForUCLKPStateChangeFinal,
+				mode_lib->vba.MaxInterDCNTileRepeaters,
+				mode_lib->vba.MaxPrefetchMode,
+				mode_lib->vba.DRAMClockChangeLatency,
+				mode_lib->vba.FCLKChangeLatency,
+				mode_lib->vba.SREnterPlusExitTime,
+				mode_lib->vba.ReturnBusWidth,
+				mode_lib->vba.RoundTripPingLatencyCycles,
+				v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.ReorderingBytes,
+				mode_lib->vba.PixelChunkSizeInKByte,
+				mode_lib->vba.MetaChunkSize,
+				mode_lib->vba.GPUVMEnable,
+				mode_lib->vba.GPUVMMaxPageTableLevels,
+				mode_lib->vba.HostVMEnable,
+				mode_lib->vba.NumberOfActiveSurfaces,
+				mode_lib->vba.HostVMMinPageSize,
+				mode_lib->vba.HostVMMaxNonCachedPageTableLevels,
+				mode_lib->vba.DynamicMetadataVMEnabled,
+				mode_lib->vba.ImmediateFlipRequiredFinal,
+				mode_lib->vba.ProgressiveToInterlaceUnitInOPP,
+				mode_lib->vba.MaxAveragePercentOfIdealSDPPortBWDisplayCanUseInNormalSystemOperation,
+				mode_lib->vba.PercentOfIdealFabricAndSDPPortBWReceivedAfterUrgLatency,
+				mode_lib->vba.VTotal,
+				mode_lib->vba.VActive,
+				mode_lib->vba.DynamicMetadataTransmittedBytes,
+				mode_lib->vba.DynamicMetadataLinesBeforeActiveRequired,
+				mode_lib->vba.Interlace,
+				mode_lib->vba.RequiredDPPCLK,
+				mode_lib->vba.RequiredDISPCLK,
+				mode_lib->vba.UrgLatency,
+				mode_lib->vba.NoOfDPP,
+				mode_lib->vba.ProjectedDCFCLKDeepSleep,
+				mode_lib->vba.MaximumVStartup,
+				mode_lib->vba.TotalNumberOfActiveDPP,
+				mode_lib->vba.TotalNumberOfDCCActiveDPP,
+				mode_lib->vba.dpte_group_bytes,
+				mode_lib->vba.PrefetchLinesY,
+				mode_lib->vba.PrefetchLinesC,
+				mode_lib->vba.swath_width_luma_ub_all_states,
+				mode_lib->vba.swath_width_chroma_ub_all_states,
+				mode_lib->vba.BytePerPixelY,
+				mode_lib->vba.BytePerPixelC,
+				mode_lib->vba.HTotal,
+				mode_lib->vba.PixelClock,
+				mode_lib->vba.PDEAndMetaPTEBytesPerFrame,
+				mode_lib->vba.DPTEBytesPerRow,
+				mode_lib->vba.MetaRowBytes,
+				mode_lib->vba.DynamicMetadataEnable,
+				mode_lib->vba.ReadBandwidthLuma,
+				mode_lib->vba.ReadBandwidthChroma,
+				mode_lib->vba.DCFCLKPerState,
+
+				/* Output */
+				mode_lib->vba.DCFCLKState);
+	} // UseMinimumRequiredDCFCLK == true
+
+	for (i = start_state; i < (int) v->soc.num_states; ++i) {
+		for (j = 0; j <= 1; ++j) {
+			mode_lib->vba.ReturnBWPerState[i][j] = dml32_get_return_bw_mbps(&mode_lib->vba.soc, i,
+					mode_lib->vba.HostVMEnable, mode_lib->vba.DCFCLKState[i][j],
+					mode_lib->vba.FabricClockPerState[i], mode_lib->vba.DRAMSpeedPerState[i]);
+		}
+	}
+
+	//Re-ordering Buffer Support Check
+	for (i = start_state; i < (int) v->soc.num_states; ++i) {
+		for (j = 0; j <= 1; ++j) {
+			if ((mode_lib->vba.ROBBufferSizeInKByte - mode_lib->vba.PixelChunkSizeInKByte) * 1024
+					/ mode_lib->vba.ReturnBWPerState[i][j]
+					> (mode_lib->vba.RoundTripPingLatencyCycles + 32)
+							/ mode_lib->vba.DCFCLKState[i][j]
+							+ v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.ReorderingBytes / mode_lib->vba.ReturnBWPerState[i][j]) {
+				mode_lib->vba.ROBSupport[i][j] = true;
+			} else {
+				mode_lib->vba.ROBSupport[i][j] = false;
+			}
+		}
+	}
+
+	//Vertical Active BW support check
+	v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.MaxTotalVActiveRDBandwidth = 0;
+
+	for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+		v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.MaxTotalVActiveRDBandwidth += mode_lib->vba.ReadBandwidthLuma[k]
+				+ mode_lib->vba.ReadBandwidthChroma[k];
+	}
+
+	for (i = start_state; i < (int) v->soc.num_states; ++i) {
+		for (j = 0; j <= 1; ++j) {
+			mode_lib->vba.MaxTotalVerticalActiveAvailableBandwidth[i][j] =
+				dml_min3(mode_lib->vba.ReturnBusWidth * mode_lib->vba.DCFCLKState[i][j]
+					* mode_lib->vba.MaxAveragePercentOfIdealSDPPortBWDisplayCanUseInNormalSystemOperation / 100,
+					mode_lib->vba.FabricClockPerState[i]
+					* mode_lib->vba.FabricDatapathToDCNDataReturn
+					* mode_lib->vba.MaxAveragePercentOfIdealFabricBWDisplayCanUseInNormalSystemOperation / 100,
+					mode_lib->vba.DRAMSpeedPerState[i]
+					* mode_lib->vba.NumberOfChannels
+					* mode_lib->vba.DRAMChannelWidth
+					* (i < 2 ? mode_lib->vba.MaxAveragePercentOfIdealDRAMBWDisplayCanUseInNormalSystemOperationSTROBE : mode_lib->vba.MaxAveragePercentOfIdealDRAMBWDisplayCanUseInNormalSystemOperation) / 100);
+
+			if (v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.MaxTotalVActiveRDBandwidth
+					<= mode_lib->vba.MaxTotalVerticalActiveAvailableBandwidth[i][j]) {
+				mode_lib->vba.TotalVerticalActiveBandwidthSupport[i][j] = true;
+			} else {
+				mode_lib->vba.TotalVerticalActiveBandwidthSupport[i][j] = false;
+			}
+		}
+	}
+
+	/* Prefetch Check */
+
+	for (i = start_state; i < (int) v->soc.num_states; ++i) {
+		for (j = 0; j <= 1; ++j) {
+
+			mode_lib->vba.TimeCalc = 24 / mode_lib->vba.ProjectedDCFCLKDeepSleep[i][j];
+
+			for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+				mode_lib->vba.NoOfDPPThisState[k] = mode_lib->vba.NoOfDPP[i][j][k];
+				mode_lib->vba.swath_width_luma_ub_this_state[k] =
+						mode_lib->vba.swath_width_luma_ub_all_states[i][j][k];
+				mode_lib->vba.swath_width_chroma_ub_this_state[k] =
+						mode_lib->vba.swath_width_chroma_ub_all_states[i][j][k];
+				mode_lib->vba.SwathWidthYThisState[k] = mode_lib->vba.SwathWidthYAllStates[i][j][k];
+				mode_lib->vba.SwathWidthCThisState[k] = mode_lib->vba.SwathWidthCAllStates[i][j][k];
+				mode_lib->vba.SwathHeightYThisState[k] = mode_lib->vba.SwathHeightYAllStates[i][j][k];
+				mode_lib->vba.SwathHeightCThisState[k] = mode_lib->vba.SwathHeightCAllStates[i][j][k];
+				mode_lib->vba.UnboundedRequestEnabledThisState =
+						mode_lib->vba.UnboundedRequestEnabledAllStates[i][j];
+				mode_lib->vba.CompressedBufferSizeInkByteThisState =
+						mode_lib->vba.CompressedBufferSizeInkByteAllStates[i][j];
+				mode_lib->vba.DETBufferSizeInKByteThisState[k] =
+						mode_lib->vba.DETBufferSizeInKByteAllStates[i][j][k];
+				mode_lib->vba.DETBufferSizeYThisState[k] =
+						mode_lib->vba.DETBufferSizeYAllStates[i][j][k];
+				mode_lib->vba.DETBufferSizeCThisState[k] =
+						mode_lib->vba.DETBufferSizeCAllStates[i][j][k];
+			}
+
+			mode_lib->vba.VActiveBandwithSupport[i][j] = dml32_CalculateVActiveBandwithSupport(
+					mode_lib->vba.NumberOfActiveSurfaces,
+					mode_lib->vba.ReturnBWPerState[i][j],
+					mode_lib->vba.NoUrgentLatencyHiding,
+					mode_lib->vba.ReadBandwidthLuma,
+					mode_lib->vba.ReadBandwidthChroma,
+					mode_lib->vba.cursor_bw,
+					mode_lib->vba.meta_row_bandwidth_this_state,
+					mode_lib->vba.dpte_row_bandwidth_this_state,
+					mode_lib->vba.NoOfDPPThisState,
+					mode_lib->vba.UrgentBurstFactorLuma,
+					mode_lib->vba.UrgentBurstFactorChroma,
+					mode_lib->vba.UrgentBurstFactorCursor);
+
+			mode_lib->vba.NotEnoughDETSwathFillLatencyHidingPerState[i][j] = dml32_CalculateDETSwathFillLatencyHiding(
+					mode_lib->vba.NumberOfActiveSurfaces,
+					mode_lib->vba.ReturnBWPerState[i][j],
+					mode_lib->vba.UrgLatency[i],
+					mode_lib->vba.SwathHeightYThisState,
+					mode_lib->vba.SwathHeightCThisState,
+					mode_lib->vba.swath_width_luma_ub_this_state,
+					mode_lib->vba.swath_width_chroma_ub_this_state,
+					mode_lib->vba.BytePerPixelInDETY,
+					mode_lib->vba.BytePerPixelInDETC,
+					mode_lib->vba.DETBufferSizeYThisState,
+					mode_lib->vba.DETBufferSizeCThisState,
+					mode_lib->vba.NoOfDPPThisState,
+					mode_lib->vba.HTotal,
+					mode_lib->vba.PixelClock,
+					mode_lib->vba.VRatio,
+					mode_lib->vba.VRatioChroma,
+					mode_lib->vba.UsesMALLForPStateChange,
+					mode_lib->vba.UseUnboundedRequesting);
+
+			v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.VMDataOnlyReturnBWPerState = dml32_get_return_bw_mbps_vm_only(&mode_lib->vba.soc, i,
+					mode_lib->vba.DCFCLKState[i][j], mode_lib->vba.FabricClockPerState[i],
+					mode_lib->vba.DRAMSpeedPerState[i]);
+			v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.HostVMInefficiencyFactor = 1;
+
+			if (mode_lib->vba.GPUVMEnable && mode_lib->vba.HostVMEnable)
+				v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.HostVMInefficiencyFactor = mode_lib->vba.ReturnBWPerState[i][j]
+						/ v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.VMDataOnlyReturnBWPerState;
+
+			mode_lib->vba.ExtraLatency = dml32_CalculateExtraLatency(
+					mode_lib->vba.RoundTripPingLatencyCycles, v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.ReorderingBytes,
+					mode_lib->vba.DCFCLKState[i][j], mode_lib->vba.TotalNumberOfActiveDPP[i][j],
+					mode_lib->vba.PixelChunkSizeInKByte,
+					mode_lib->vba.TotalNumberOfDCCActiveDPP[i][j], mode_lib->vba.MetaChunkSize,
+					mode_lib->vba.ReturnBWPerState[i][j], mode_lib->vba.GPUVMEnable,
+					mode_lib->vba.HostVMEnable, mode_lib->vba.NumberOfActiveSurfaces,
+					mode_lib->vba.NoOfDPPThisState, mode_lib->vba.dpte_group_bytes,
+					v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.HostVMInefficiencyFactor, mode_lib->vba.HostVMMinPageSize,
+					mode_lib->vba.HostVMMaxNonCachedPageTableLevels);
+
+			v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.NextPrefetchModeState = mode_lib->vba.MinPrefetchMode;
+
+			mode_lib->vba.NextMaxVStartup = mode_lib->vba.MaxMaxVStartup[i][j];
+
+			do {
+				mode_lib->vba.PrefetchModePerState[i][j] = v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.NextPrefetchModeState;
+				mode_lib->vba.MaxVStartup = mode_lib->vba.NextMaxVStartup;
+
+				for (k = 0; k <= mode_lib->vba.NumberOfActiveSurfaces - 1; k++) {
+					mode_lib->vba.TWait = dml32_CalculateTWait(
+							mode_lib->vba.PrefetchModePerState[i][j],
+							mode_lib->vba.UsesMALLForPStateChange[k],
+							mode_lib->vba.SynchronizeDRRDisplaysForUCLKPStateChangeFinal,
+							mode_lib->vba.DRRDisplay[k],
+							mode_lib->vba.DRAMClockChangeLatency,
+							mode_lib->vba.FCLKChangeLatency, mode_lib->vba.UrgLatency[i],
+							mode_lib->vba.SREnterPlusExitTime);
+
+					memset(&v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull, 0, sizeof(DmlPipe));
+					v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.myPipe.Dppclk = mode_lib->vba.RequiredDPPCLK[i][j][k];
+					v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.myPipe.Dispclk = mode_lib->vba.RequiredDISPCLK[i][j];
+					v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.myPipe.PixelClock = mode_lib->vba.PixelClock[k];
+					v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.myPipe.DCFClkDeepSleep = mode_lib->vba.ProjectedDCFCLKDeepSleep[i][j];
+					v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.myPipe.DPPPerSurface = mode_lib->vba.NoOfDPP[i][j][k];
+					v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.myPipe.ScalerEnabled = mode_lib->vba.ScalerEnabled[k];
+					v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.myPipe.SourceRotation = mode_lib->vba.SourceRotation[k];
+					v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.myPipe.BlockWidth256BytesY = mode_lib->vba.Read256BlockWidthY[k];
+					v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.myPipe.BlockHeight256BytesY = mode_lib->vba.Read256BlockHeightY[k];
+					v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.myPipe.BlockWidth256BytesC = mode_lib->vba.Read256BlockWidthC[k];
+					v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.myPipe.BlockHeight256BytesC = mode_lib->vba.Read256BlockHeightC[k];
+					v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.myPipe.InterlaceEnable = mode_lib->vba.Interlace[k];
+					v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.myPipe.NumberOfCursors = mode_lib->vba.NumberOfCursors[k];
+					v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.myPipe.VBlank = mode_lib->vba.VTotal[k] - mode_lib->vba.VActive[k];
+					v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.myPipe.HTotal = mode_lib->vba.HTotal[k];
+					v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.myPipe.HActive = mode_lib->vba.HActive[k];
+					v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.myPipe.DCCEnable = mode_lib->vba.DCCEnable[k];
+					v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.myPipe.ODMMode = mode_lib->vba.ODMCombineEnablePerState[i][k];
+					v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.myPipe.SourcePixelFormat = mode_lib->vba.SourcePixelFormat[k];
+					v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.myPipe.BytePerPixelY = mode_lib->vba.BytePerPixelY[k];
+					v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.myPipe.BytePerPixelC = mode_lib->vba.BytePerPixelC[k];
+					v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.myPipe.ProgressiveToInterlaceUnitInOPP =
+							mode_lib->vba.ProgressiveToInterlaceUnitInOPP;
+
+					mode_lib->vba.NoTimeForPrefetch[i][j][k] =
+						dml32_CalculatePrefetchSchedule(
+							v,
+							k,
+							v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.HostVMInefficiencyFactor,
+							&v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.myPipe,
+							v->DSCDelayPerState[i][k],
+							v->SwathWidthYThisState[k] / v->HRatio[k],
+							dml_min(v->MaxVStartup, v->MaximumVStartup[i][j][k]),
+							v->MaximumVStartup[i][j][k],
+							v->UrgLatency[i],
+							v->ExtraLatency,
+							v->TimeCalc,
+							v->PDEAndMetaPTEBytesPerFrame[i][j][k],
+							v->MetaRowBytes[i][j][k],
+							v->DPTEBytesPerRow[i][j][k],
+							v->PrefetchLinesY[i][j][k],
+							v->SwathWidthYThisState[k],
+							v->PrefillY[k],
+							v->MaxNumSwY[k],
+							v->PrefetchLinesC[i][j][k],
+							v->SwathWidthCThisState[k],
+							v->PrefillC[k],
+							v->MaxNumSwC[k],
+							v->swath_width_luma_ub_this_state[k],
+							v->swath_width_chroma_ub_this_state[k],
+							v->SwathHeightYThisState[k],
+							v->SwathHeightCThisState[k], v->TWait,
+							(v->DRAMSpeedPerState[i] <= MEM_STROBE_FREQ_MHZ || v->DCFCLKState[i][j] <= DCFCLK_FREQ_EXTRA_PREFETCH_REQ_MHZ) ?
+									mode_lib->vba.ip.min_prefetch_in_strobe_us : 0,
+							mode_lib->vba.PrefetchModePerState[i][j] > 0 || mode_lib->vba.DRAMClockChangeRequirementFinal == false,
+
+							/* Output */
+							&v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.DSTXAfterScaler[k],
+							&v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.DSTYAfterScaler[k],
+							&v->LineTimesForPrefetch[k],
+							&v->PrefetchBW[k],
+							&v->LinesForMetaPTE[k],
+							&v->LinesForMetaAndDPTERow[k],
+							&v->VRatioPreY[i][j][k],
+							&v->VRatioPreC[i][j][k],
+							&v->RequiredPrefetchPixelDataBWLuma[0][0][k],
+							&v->RequiredPrefetchPixelDataBWChroma[0][0][k],
+							&v->NoTimeForDynamicMetadata[i][j][k],
+							&v->Tno_bw[k],
+							&v->prefetch_vmrow_bw[k],
+							&v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_single[0],         // double *Tdmdl_vm
+							&v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_single[1],         // double *Tdmdl
+							&v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_single[2],         // double *TSetup
+							&v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_integer[0],         							    // unsigned int   *VUpdateOffsetPix
+							&v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_single[3],         // unsigned int   *VUpdateWidthPix
+							&v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_single[4]);        // unsigned int   *VReadyOffsetPix
+				}
+
+				for (k = 0; k <= mode_lib->vba.NumberOfActiveSurfaces - 1; k++) {
+					dml32_CalculateUrgentBurstFactor(
+							mode_lib->vba.UsesMALLForPStateChange[k],
+							mode_lib->vba.swath_width_luma_ub_this_state[k],
+							mode_lib->vba.swath_width_chroma_ub_this_state[k],
+							mode_lib->vba.SwathHeightYThisState[k],
+							mode_lib->vba.SwathHeightCThisState[k],
+							mode_lib->vba.HTotal[k] / mode_lib->vba.PixelClock[k],
+							mode_lib->vba.UrgLatency[i], mode_lib->vba.CursorBufferSize,
+							mode_lib->vba.CursorWidth[k][0], mode_lib->vba.CursorBPP[k][0],
+							mode_lib->vba.VRatioPreY[i][j][k],
+							mode_lib->vba.VRatioPreC[i][j][k],
+							mode_lib->vba.BytePerPixelInDETY[k],
+							mode_lib->vba.BytePerPixelInDETC[k],
+							mode_lib->vba.DETBufferSizeYThisState[k],
+							mode_lib->vba.DETBufferSizeCThisState[k],
+							/* Output */
+							&mode_lib->vba.UrgentBurstFactorCursorPre[k],
+							&mode_lib->vba.UrgentBurstFactorLumaPre[k],
+							&mode_lib->vba.UrgentBurstFactorChromaPre[k],
+							&mode_lib->vba.NotUrgentLatencyHidingPre[k]);
+
+					v->cursor_bw_pre[k] = mode_lib->vba.NumberOfCursors[k] * mode_lib->vba.CursorWidth[k][0] * mode_lib->vba.CursorBPP[k][0] /
+							8.0 / (mode_lib->vba.HTotal[k] / mode_lib->vba.PixelClock[k]) * v->VRatioPreY[i][j][k];
+				}
+
+				{
+					dml32_CalculatePrefetchBandwithSupport(
+							mode_lib->vba.NumberOfActiveSurfaces,
+							mode_lib->vba.ReturnBWPerState[i][j],
+							mode_lib->vba.NotUrgentLatencyHidingPre,
+							mode_lib->vba.ReadBandwidthLuma,
+							mode_lib->vba.ReadBandwidthChroma,
+							mode_lib->vba.RequiredPrefetchPixelDataBWLuma[0][0],
+							mode_lib->vba.RequiredPrefetchPixelDataBWChroma[0][0],
+							mode_lib->vba.cursor_bw,
+							mode_lib->vba.meta_row_bandwidth_this_state,
+							mode_lib->vba.dpte_row_bandwidth_this_state,
+							mode_lib->vba.cursor_bw_pre,
+							mode_lib->vba.prefetch_vmrow_bw,
+							mode_lib->vba.NoOfDPPThisState,
+							mode_lib->vba.UrgentBurstFactorLuma,
+							mode_lib->vba.UrgentBurstFactorChroma,
+							mode_lib->vba.UrgentBurstFactorCursor,
+							mode_lib->vba.UrgentBurstFactorLumaPre,
+							mode_lib->vba.UrgentBurstFactorChromaPre,
+							mode_lib->vba.UrgentBurstFactorCursorPre,
+							v->PrefetchBW,
+							v->VRatio,
+							v->MaxVRatioPre,
+
+							/* output */
+							&v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_single[0],   // Single  *PrefetchBandwidth
+							&v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_single[1],   // Single  *FractionOfUrgentBandwidth
+							&mode_lib->vba.PrefetchSupported[i][j]);
+				}
+
+				for (k = 0; k <= mode_lib->vba.NumberOfActiveSurfaces - 1; k++) {
+					if (mode_lib->vba.LineTimesForPrefetch[k]
+							< 2.0 || mode_lib->vba.LinesForMetaPTE[k] >= 32.0
+							|| mode_lib->vba.LinesForMetaAndDPTERow[k] >= 16.0
+							|| mode_lib->vba.NoTimeForPrefetch[i][j][k] == true) {
+						mode_lib->vba.PrefetchSupported[i][j] = false;
+					}
+				}
+
+				mode_lib->vba.DynamicMetadataSupported[i][j] = true;
+				for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+					if (mode_lib->vba.NoTimeForDynamicMetadata[i][j][k] == true)
+						mode_lib->vba.DynamicMetadataSupported[i][j] = false;
+				}
+
+				mode_lib->vba.VRatioInPrefetchSupported[i][j] = true;
+				for (k = 0; k <= mode_lib->vba.NumberOfActiveSurfaces - 1; k++) {
+					if (mode_lib->vba.VRatioPreY[i][j][k] > mode_lib->vba.MaxVRatioPre
+							|| mode_lib->vba.VRatioPreC[i][j][k] > mode_lib->vba.MaxVRatioPre
+							|| mode_lib->vba.NoTimeForPrefetch[i][j][k] == true) {
+						mode_lib->vba.VRatioInPrefetchSupported[i][j] = false;
+					}
+				}
+				mode_lib->vba.AnyLinesForVMOrRowTooLarge = false;
+				for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+					if (mode_lib->vba.LinesForMetaAndDPTERow[k] >= 16
+							|| mode_lib->vba.LinesForMetaPTE[k] >= 32) {
+						mode_lib->vba.AnyLinesForVMOrRowTooLarge = true;
+					}
+				}
+
+				if (mode_lib->vba.PrefetchSupported[i][j] == true
+						&& mode_lib->vba.VRatioInPrefetchSupported[i][j] == true) {
+					mode_lib->vba.BandwidthAvailableForImmediateFlip =
+							dml32_CalculateBandwidthAvailableForImmediateFlip(
+							mode_lib->vba.NumberOfActiveSurfaces,
+							mode_lib->vba.ReturnBWPerState[i][j],
+							mode_lib->vba.ReadBandwidthLuma,
+							mode_lib->vba.ReadBandwidthChroma,
+							mode_lib->vba.RequiredPrefetchPixelDataBWLuma[0][0],
+							mode_lib->vba.RequiredPrefetchPixelDataBWChroma[0][0],
+							mode_lib->vba.cursor_bw,
+							mode_lib->vba.cursor_bw_pre,
+							mode_lib->vba.NoOfDPPThisState,
+							mode_lib->vba.UrgentBurstFactorLuma,
+							mode_lib->vba.UrgentBurstFactorChroma,
+							mode_lib->vba.UrgentBurstFactorCursor,
+							mode_lib->vba.UrgentBurstFactorLumaPre,
+							mode_lib->vba.UrgentBurstFactorChromaPre,
+							mode_lib->vba.UrgentBurstFactorCursorPre);
+
+					mode_lib->vba.TotImmediateFlipBytes = 0.0;
+					for (k = 0; k <= mode_lib->vba.NumberOfActiveSurfaces - 1; k++) {
+						if (!(mode_lib->vba.ImmediateFlipRequirement[k] ==
+								dm_immediate_flip_not_required)) {
+							mode_lib->vba.TotImmediateFlipBytes =
+									mode_lib->vba.TotImmediateFlipBytes
+								+ mode_lib->vba.NoOfDPP[i][j][k]
+								* mode_lib->vba.PDEAndMetaPTEBytesPerFrame[i][j][k]
+								+ mode_lib->vba.MetaRowBytes[i][j][k];
+							if (mode_lib->vba.use_one_row_for_frame_flip[i][j][k]) {
+								mode_lib->vba.TotImmediateFlipBytes =
+									mode_lib->vba.TotImmediateFlipBytes + 2
+								* mode_lib->vba.DPTEBytesPerRow[i][j][k];
+							} else {
+								mode_lib->vba.TotImmediateFlipBytes =
+									mode_lib->vba.TotImmediateFlipBytes
+								+ mode_lib->vba.DPTEBytesPerRow[i][j][k];
+							}
+						}
+					}
+
+					for (k = 0; k <= mode_lib->vba.NumberOfActiveSurfaces - 1; k++) {
+						dml32_CalculateFlipSchedule(v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.HostVMInefficiencyFactor,
+							mode_lib->vba.ExtraLatency,
+							mode_lib->vba.UrgLatency[i],
+							mode_lib->vba.GPUVMMaxPageTableLevels,
+							mode_lib->vba.HostVMEnable,
+							mode_lib->vba.HostVMMaxNonCachedPageTableLevels,
+							mode_lib->vba.GPUVMEnable,
+							mode_lib->vba.HostVMMinPageSize,
+							mode_lib->vba.PDEAndMetaPTEBytesPerFrame[i][j][k],
+							mode_lib->vba.MetaRowBytes[i][j][k],
+							mode_lib->vba.DPTEBytesPerRow[i][j][k],
+							mode_lib->vba.BandwidthAvailableForImmediateFlip,
+							mode_lib->vba.TotImmediateFlipBytes,
+							mode_lib->vba.SourcePixelFormat[k],
+							(mode_lib->vba.HTotal[k] / mode_lib->vba.PixelClock[k]),
+							mode_lib->vba.VRatio[k],
+							mode_lib->vba.VRatioChroma[k],
+							mode_lib->vba.Tno_bw[k],
+								mode_lib->vba.DCCEnable[k],
+							mode_lib->vba.dpte_row_height[k],
+							mode_lib->vba.meta_row_height[k],
+							mode_lib->vba.dpte_row_height_chroma[k],
+							mode_lib->vba.meta_row_height_chroma[k],
+							mode_lib->vba.use_one_row_for_frame_flip[i][j][k], // 24
+
+							/* Output */
+							&mode_lib->vba.DestinationLinesToRequestVMInImmediateFlip[k],
+							&mode_lib->vba.DestinationLinesToRequestRowInImmediateFlip[k],
+							&mode_lib->vba.final_flip_bw[k],
+							&mode_lib->vba.ImmediateFlipSupportedForPipe[k]);
+					}
+
+					{
+						dml32_CalculateImmediateFlipBandwithSupport(mode_lib->vba.NumberOfActiveSurfaces,
+								mode_lib->vba.ReturnBWPerState[i][j],
+								mode_lib->vba.ImmediateFlipRequirement,
+								mode_lib->vba.final_flip_bw,
+								mode_lib->vba.ReadBandwidthLuma,
+								mode_lib->vba.ReadBandwidthChroma,
+								mode_lib->vba.RequiredPrefetchPixelDataBWLuma[0][0],
+								mode_lib->vba.RequiredPrefetchPixelDataBWChroma[0][0],
+								mode_lib->vba.cursor_bw,
+								mode_lib->vba.meta_row_bandwidth_this_state,
+								mode_lib->vba.dpte_row_bandwidth_this_state,
+								mode_lib->vba.cursor_bw_pre,
+								mode_lib->vba.prefetch_vmrow_bw,
+								mode_lib->vba.DPPPerPlane,
+								mode_lib->vba.UrgentBurstFactorLuma,
+								mode_lib->vba.UrgentBurstFactorChroma,
+								mode_lib->vba.UrgentBurstFactorCursor,
+								mode_lib->vba.UrgentBurstFactorLumaPre,
+								mode_lib->vba.UrgentBurstFactorChromaPre,
+								mode_lib->vba.UrgentBurstFactorCursorPre,
+
+								/* output */
+								&v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_single[0], //  Single  *TotalBandwidth
+								&v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_single[1], //  Single  *FractionOfUrgentBandwidth
+								&mode_lib->vba.ImmediateFlipSupportedForState[i][j]); // Boolean *ImmediateFlipBandwidthSupport
+					}
+
+					for (k = 0; k <= mode_lib->vba.NumberOfActiveSurfaces - 1; k++) {
+						if (!(mode_lib->vba.ImmediateFlipRequirement[k]
+								== dm_immediate_flip_not_required)
+								&& (mode_lib->vba.ImmediateFlipSupportedForPipe[k]
+										== false))
+							mode_lib->vba.ImmediateFlipSupportedForState[i][j] = false;
+					}
+				} else { // if prefetch not support, assume iflip not supported
+					mode_lib->vba.ImmediateFlipSupportedForState[i][j] = false;
+				}
+
+				if (mode_lib->vba.MaxVStartup <= __DML_VBA_MIN_VSTARTUP__
+						|| mode_lib->vba.AnyLinesForVMOrRowTooLarge == false) {
+					mode_lib->vba.NextMaxVStartup = mode_lib->vba.MaxMaxVStartup[i][j];
+					v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.NextPrefetchModeState = v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.NextPrefetchModeState + 1;
+				} else {
+					mode_lib->vba.NextMaxVStartup = mode_lib->vba.NextMaxVStartup - 1;
+				}
+			} while (!((mode_lib->vba.PrefetchSupported[i][j] == true
+					&& mode_lib->vba.DynamicMetadataSupported[i][j] == true
+					&& mode_lib->vba.VRatioInPrefetchSupported[i][j] == true &&
+					// consider flip support is okay if when there is no hostvm and the
+					// user does't require a iflip OR the flip bw is ok
+					// If there is hostvm, DCN needs to support iflip for invalidation
+					((mode_lib->vba.HostVMEnable == false
+							&& !mode_lib->vba.ImmediateFlipRequiredFinal)
+							|| mode_lib->vba.ImmediateFlipSupportedForState[i][j] == true))
+					|| (mode_lib->vba.NextMaxVStartup == mode_lib->vba.MaxMaxVStartup[i][j]
+							&& v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.NextPrefetchModeState > mode_lib->vba.MaxPrefetchMode)));
+
+			for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k) {
+				mode_lib->vba.use_one_row_for_frame_this_state[k] =
+						mode_lib->vba.use_one_row_for_frame[i][j][k];
+			}
+
+
+			v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.mSOCParameters.UrgentLatency = mode_lib->vba.UrgLatency[i];
+			v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.mSOCParameters.ExtraLatency = mode_lib->vba.ExtraLatency;
+			v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.mSOCParameters.WritebackLatency = mode_lib->vba.WritebackLatency;
+			v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.mSOCParameters.DRAMClockChangeLatency = mode_lib->vba.DRAMClockChangeLatency;
+			v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.mSOCParameters.FCLKChangeLatency = mode_lib->vba.FCLKChangeLatency;
+			v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.mSOCParameters.SRExitTime = mode_lib->vba.SRExitTime;
+			v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.mSOCParameters.SREnterPlusExitTime = mode_lib->vba.SREnterPlusExitTime;
+			v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.mSOCParameters.SRExitZ8Time = mode_lib->vba.SRExitZ8Time;
+			v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.mSOCParameters.SREnterPlusExitZ8Time = mode_lib->vba.SREnterPlusExitZ8Time;
+			v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.mSOCParameters.USRRetrainingLatency = mode_lib->vba.USRRetrainingLatency;
+			v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.mSOCParameters.SMNLatency = mode_lib->vba.SMNLatency;
+
+			{
+				dml32_CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport(
+						v,
+						v->PrefetchModePerState[i][j],
+						v->DCFCLKState[i][j],
+						v->ReturnBWPerState[i][j],
+						v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.mSOCParameters,
+						v->SOCCLKPerState[i],
+						v->ProjectedDCFCLKDeepSleep[i][j],
+						v->DETBufferSizeYThisState,
+						v->DETBufferSizeCThisState,
+						v->SwathHeightYThisState,
+						v->SwathHeightCThisState,
+						v->SwathWidthYThisState, // 24
+						v->SwathWidthCThisState,
+						v->NoOfDPPThisState,
+						v->BytePerPixelInDETY,
+						v->BytePerPixelInDETC,
+						v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.DSTXAfterScaler,
+						v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.DSTYAfterScaler,
+						v->UnboundedRequestEnabledThisState,
+						v->CompressedBufferSizeInkByteThisState,
+
+						/* Output */
+						&v->DRAMClockChangeSupport[i][j],
+						&v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_single2[0], // double *MaxActiveDRAMClockChangeLatencySupported
+						&v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_integer[0], // Long SubViewportLinesNeededInMALL[]
+						&v->FCLKChangeSupport[i][j],
+						&v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.dummy_single2[1], // double *MinActiveFCLKChangeLatencySupported
+						&mode_lib->vba.USRRetrainingSupport[i][j],
+						mode_lib->vba.ActiveDRAMClockChangeLatencyMarginPerState[i][j]);
+			}
+		}
+	} // End of Prefetch Check
+
+	/*Cursor Support Check*/
+	mode_lib->vba.CursorSupport = true;
+	for (k = 0; k <= mode_lib->vba.NumberOfActiveSurfaces - 1; k++) {
+		if (mode_lib->vba.CursorWidth[k][0] > 0.0) {
+			if (mode_lib->vba.CursorBPP[k][0] == 64 && mode_lib->vba.Cursor64BppSupport == false)
+				mode_lib->vba.CursorSupport = false;
+		}
+	}
+
+	/*Valid Pitch Check*/
+	mode_lib->vba.PitchSupport = true;
+	for (k = 0; k <= mode_lib->vba.NumberOfActiveSurfaces - 1; k++) {
+		mode_lib->vba.AlignedYPitch[k] = dml_ceil(
+				dml_max(mode_lib->vba.PitchY[k], mode_lib->vba.SurfaceWidthY[k]),
+				mode_lib->vba.MacroTileWidthY[k]);
+		if (mode_lib->vba.DCCEnable[k] == true) {
+			mode_lib->vba.AlignedDCCMetaPitchY[k] = dml_ceil(
+					dml_max(mode_lib->vba.DCCMetaPitchY[k], mode_lib->vba.SurfaceWidthY[k]),
+					64.0 * mode_lib->vba.Read256BlockWidthY[k]);
+		} else {
+			mode_lib->vba.AlignedDCCMetaPitchY[k] = mode_lib->vba.DCCMetaPitchY[k];
+		}
+		if (mode_lib->vba.SourcePixelFormat[k] != dm_444_64 && mode_lib->vba.SourcePixelFormat[k] != dm_444_32
+				&& mode_lib->vba.SourcePixelFormat[k] != dm_444_16
+				&& mode_lib->vba.SourcePixelFormat[k] != dm_mono_16
+				&& mode_lib->vba.SourcePixelFormat[k] != dm_rgbe
+				&& mode_lib->vba.SourcePixelFormat[k] != dm_mono_8) {
+			mode_lib->vba.AlignedCPitch[k] = dml_ceil(
+					dml_max(mode_lib->vba.PitchC[k], mode_lib->vba.SurfaceWidthC[k]),
+					mode_lib->vba.MacroTileWidthC[k]);
+			if (mode_lib->vba.DCCEnable[k] == true) {
+				mode_lib->vba.AlignedDCCMetaPitchC[k] = dml_ceil(
+						dml_max(mode_lib->vba.DCCMetaPitchC[k],
+								mode_lib->vba.SurfaceWidthC[k]),
+						64.0 * mode_lib->vba.Read256BlockWidthC[k]);
+			} else {
+				mode_lib->vba.AlignedDCCMetaPitchC[k] = mode_lib->vba.DCCMetaPitchC[k];
+			}
+		} else {
+			mode_lib->vba.AlignedCPitch[k] = mode_lib->vba.PitchC[k];
+			mode_lib->vba.AlignedDCCMetaPitchC[k] = mode_lib->vba.DCCMetaPitchC[k];
+		}
+		if (mode_lib->vba.AlignedYPitch[k] > mode_lib->vba.PitchY[k]
+				|| mode_lib->vba.AlignedCPitch[k] > mode_lib->vba.PitchC[k]
+				|| mode_lib->vba.AlignedDCCMetaPitchY[k] > mode_lib->vba.DCCMetaPitchY[k]
+				|| mode_lib->vba.AlignedDCCMetaPitchC[k] > mode_lib->vba.DCCMetaPitchC[k]) {
+			mode_lib->vba.PitchSupport = false;
+		}
+	}
+
+	mode_lib->vba.ViewportExceedsSurface = false;
+	for (k = 0; k <= mode_lib->vba.NumberOfActiveSurfaces - 1; k++) {
+		if (mode_lib->vba.ViewportWidth[k] > mode_lib->vba.SurfaceWidthY[k]
+				|| mode_lib->vba.ViewportHeight[k] > mode_lib->vba.SurfaceHeightY[k]) {
+			mode_lib->vba.ViewportExceedsSurface = true;
+			if (mode_lib->vba.SourcePixelFormat[k] != dm_444_64
+					&& mode_lib->vba.SourcePixelFormat[k] != dm_444_32
+					&& mode_lib->vba.SourcePixelFormat[k] != dm_444_16
+					&& mode_lib->vba.SourcePixelFormat[k] != dm_444_8
+					&& mode_lib->vba.SourcePixelFormat[k] != dm_rgbe) {
+				if (mode_lib->vba.ViewportWidthChroma[k] > mode_lib->vba.SurfaceWidthC[k]
+						|| mode_lib->vba.ViewportHeightChroma[k]
+								> mode_lib->vba.SurfaceHeightC[k]) {
+					mode_lib->vba.ViewportExceedsSurface = true;
+				}
+			}
+		}
+	}
+
+	/*Mode Support, Voltage State and SOC Configuration*/
+	mode_support_configuration(v, mode_lib);
+
+	MaximumMPCCombine = 0;
+
+	for (i = v->soc.num_states; i >= start_state; i--) {
+		if (i == v->soc.num_states || mode_lib->vba.ModeSupport[i][0] == true ||
+				mode_lib->vba.ModeSupport[i][1] == true) {
+			mode_lib->vba.VoltageLevel = i;
+			mode_lib->vba.ModeIsSupported = mode_lib->vba.ModeSupport[i][0] == true
+					|| mode_lib->vba.ModeSupport[i][1] == true;
+
+			if (mode_lib->vba.ModeSupport[i][0] == true)
+				MaximumMPCCombine = 0;
+			else
+				MaximumMPCCombine = 1;
+		}
+	}
+
+	mode_lib->vba.ImmediateFlipSupport =
+			mode_lib->vba.ImmediateFlipSupportedForState[mode_lib->vba.VoltageLevel][MaximumMPCCombine];
+	mode_lib->vba.UnboundedRequestEnabled =
+			mode_lib->vba.UnboundedRequestEnabledAllStates[mode_lib->vba.VoltageLevel][MaximumMPCCombine];
+	mode_lib->vba.CompressedBufferSizeInkByte =
+			mode_lib->vba.CompressedBufferSizeInkByteAllStates[mode_lib->vba.VoltageLevel][MaximumMPCCombine]; // Not used, informational
+
+	for (k = 0; k <= mode_lib->vba.NumberOfActiveSurfaces - 1; k++) {
+		mode_lib->vba.MPCCombineEnable[k] =
+				mode_lib->vba.MPCCombine[mode_lib->vba.VoltageLevel][MaximumMPCCombine][k];
+		mode_lib->vba.DPPPerPlane[k] = mode_lib->vba.NoOfDPP[mode_lib->vba.VoltageLevel][MaximumMPCCombine][k];
+		mode_lib->vba.SwathHeightY[k] =
+				mode_lib->vba.SwathHeightYAllStates[mode_lib->vba.VoltageLevel][MaximumMPCCombine][k];
+		mode_lib->vba.SwathHeightC[k] =
+				mode_lib->vba.SwathHeightCAllStates[mode_lib->vba.VoltageLevel][MaximumMPCCombine][k];
+		mode_lib->vba.DETBufferSizeInKByte[k] =
+			mode_lib->vba.DETBufferSizeInKByteAllStates[mode_lib->vba.VoltageLevel][MaximumMPCCombine][k];
+		mode_lib->vba.DETBufferSizeY[k] =
+				mode_lib->vba.DETBufferSizeYAllStates[mode_lib->vba.VoltageLevel][MaximumMPCCombine][k];
+		mode_lib->vba.DETBufferSizeC[k] =
+				mode_lib->vba.DETBufferSizeCAllStates[mode_lib->vba.VoltageLevel][MaximumMPCCombine][k];
+		mode_lib->vba.OutputType[k] = mode_lib->vba.OutputTypePerState[mode_lib->vba.VoltageLevel][k];
+		mode_lib->vba.OutputRate[k] = mode_lib->vba.OutputRatePerState[mode_lib->vba.VoltageLevel][k];
+	}
+
+	mode_lib->vba.DCFCLK = mode_lib->vba.DCFCLKState[mode_lib->vba.VoltageLevel][MaximumMPCCombine];
+	mode_lib->vba.DRAMSpeed = mode_lib->vba.DRAMSpeedPerState[mode_lib->vba.VoltageLevel];
+	mode_lib->vba.FabricClock = mode_lib->vba.FabricClockPerState[mode_lib->vba.VoltageLevel];
+	mode_lib->vba.SOCCLK = mode_lib->vba.SOCCLKPerState[mode_lib->vba.VoltageLevel];
+	mode_lib->vba.ReturnBW = mode_lib->vba.ReturnBWPerState[mode_lib->vba.VoltageLevel][MaximumMPCCombine];
+	mode_lib->vba.DISPCLK = mode_lib->vba.RequiredDISPCLK[mode_lib->vba.VoltageLevel][MaximumMPCCombine];
+	mode_lib->vba.maxMpcComb = MaximumMPCCombine;
+
+	for (k = 0; k <= mode_lib->vba.NumberOfActiveSurfaces - 1; k++) {
+		if (mode_lib->vba.BlendingAndTiming[k] == k) {
+			mode_lib->vba.ODMCombineEnabled[k] =
+					mode_lib->vba.ODMCombineEnablePerState[mode_lib->vba.VoltageLevel][k];
+		} else {
+			mode_lib->vba.ODMCombineEnabled[k] = dm_odm_combine_mode_disabled;
+		}
+
+		mode_lib->vba.DSCEnabled[k] = mode_lib->vba.RequiresDSC[mode_lib->vba.VoltageLevel][k];
+		mode_lib->vba.FECEnable[k] = mode_lib->vba.RequiresFEC[mode_lib->vba.VoltageLevel][k];
+		mode_lib->vba.OutputBpp[k] = mode_lib->vba.OutputBppPerState[mode_lib->vba.VoltageLevel][k];
+	}
+
+	mode_lib->vba.UrgentWatermark = mode_lib->vba.Watermark.UrgentWatermark;
+	mode_lib->vba.StutterEnterPlusExitWatermark = mode_lib->vba.Watermark.StutterEnterPlusExitWatermark;
+	mode_lib->vba.StutterExitWatermark = mode_lib->vba.Watermark.StutterExitWatermark;
+	mode_lib->vba.WritebackDRAMClockChangeWatermark = mode_lib->vba.Watermark.WritebackDRAMClockChangeWatermark;
+	mode_lib->vba.DRAMClockChangeWatermark = mode_lib->vba.Watermark.DRAMClockChangeWatermark;
+	mode_lib->vba.UrgentLatency = mode_lib->vba.UrgLatency[mode_lib->vba.VoltageLevel];
+	mode_lib->vba.DCFCLKDeepSleep = mode_lib->vba.ProjectedDCFCLKDeepSleep[mode_lib->vba.VoltageLevel][MaximumMPCCombine];
+
+	/* VBA has Error type to Error Msg output here, but not necessary for DML-C */
+} // ModeSupportAndSystemConfigurationFull
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_32.h b/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_32.h
new file mode 100644
index 000000000000..c4745d63039b
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_32.h
@@ -0,0 +1,64 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __DML32_DISPLAY_MODE_VBA_H__
+#define __DML32_DISPLAY_MODE_VBA_H__
+
+#include "../display_mode_enums.h"
+
+// To enable a lot of debug msg
+//#define __DML_VBA_DEBUG__
+// For DML-C changes that hasn't been propagated to VBA yet
+//#define __DML_VBA_ALLOW_DELTA__
+
+// Move these to ip parameters/constant
+// At which vstartup the DML start to try if the mode can be supported
+#define __DML_VBA_MIN_VSTARTUP__    9
+
+// Delay in DCFCLK from ARB to DET (1st num is ARB to SDPIF, 2nd number is SDPIF to DET)
+#define __DML_ARB_TO_RET_DELAY__    7 + 95
+
+// fudge factor for min dcfclk calclation
+#define __DML_MIN_DCFCLK_FACTOR__   1.15
+
+// Prefetch schedule max vratio
+#define __DML_MAX_VRATIO_PRE__ 7.9
+#define __DML_MAX_BW_RATIO_PRE__ 4.0
+
+#define __DML_VBA_MAX_DST_Y_PRE__    63.75
+
+#define BPP_INVALID 0
+#define BPP_BLENDED_PIPE 0xffffffff
+
+#define MEM_STROBE_FREQ_MHZ 1600
+#define DCFCLK_FREQ_EXTRA_PREFETCH_REQ_MHZ 300
+#define MEM_STROBE_MAX_DELIVERY_TIME_US 60.0
+
+struct display_mode_lib;
+
+void dml32_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_lib);
+void dml32_recalculate(struct display_mode_lib *mode_lib);
+
+#endif
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_util_32.c b/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_util_32.c
new file mode 100644
index 000000000000..0748ef36a16a
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_util_32.c
@@ -0,0 +1,6350 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+#include "display_mode_vba_util_32.h"
+#include "../dml_inline_defs.h"
+#include "display_mode_vba_32.h"
+#include "../display_mode_lib.h"
+
+#define DCN32_MAX_FMT_420_BUFFER_WIDTH 4096
+
+unsigned int dml32_dscceComputeDelay(
+		unsigned int bpc,
+		double BPP,
+		unsigned int sliceWidth,
+		unsigned int numSlices,
+		enum output_format_class pixelFormat,
+		enum output_encoder_class Output)
+{
+	// valid bpc         = source bits per component in the set of {8, 10, 12}
+	// valid bpp         = increments of 1/16 of a bit
+	//                    min = 6/7/8 in N420/N422/444, respectively
+	//                    max = such that compression is 1:1
+	//valid sliceWidth  = number of pixels per slice line,
+	//	must be less than or equal to 5184/numSlices (or 4096/numSlices in 420 mode)
+	//valid numSlices   = number of slices in the horiziontal direction per DSC engine in the set of {1, 2, 3, 4}
+	//valid pixelFormat = pixel/color format in the set of {:N444_RGB, :S422, :N422, :N420}
+
+	// fixed value
+	unsigned int rcModelSize = 8192;
+
+	// N422/N420 operate at 2 pixels per clock
+	unsigned int pixelsPerClock, lstall, D, initalXmitDelay, w, s, ix, wx, p, l0, a, ax, L,
+	Delay, pixels;
+
+	if (pixelFormat == dm_420)
+		pixelsPerClock = 2;
+	else if (pixelFormat == dm_n422)
+		pixelsPerClock = 2;
+	// #all other modes operate at 1 pixel per clock
+	else
+		pixelsPerClock = 1;
+
+	//initial transmit delay as per PPS
+	initalXmitDelay = dml_round(rcModelSize / 2.0 / BPP / pixelsPerClock);
+
+	//compute ssm delay
+	if (bpc == 8)
+		D = 81;
+	else if (bpc == 10)
+		D = 89;
+	else
+		D = 113;
+
+	//divide by pixel per cycle to compute slice width as seen by DSC
+	w = sliceWidth / pixelsPerClock;
+
+	//422 mode has an additional cycle of delay
+	if (pixelFormat == dm_420 || pixelFormat == dm_444 || pixelFormat == dm_n422)
+		s = 0;
+	else
+		s = 1;
+
+	//main calculation for the dscce
+	ix = initalXmitDelay + 45;
+	wx = (w + 2) / 3;
+	p = 3 * wx - w;
+	l0 = ix / w;
+	a = ix + p * l0;
+	ax = (a + 2) / 3 + D + 6 + 1;
+	L = (ax + wx - 1) / wx;
+	if ((ix % w) == 0 && p != 0)
+		lstall = 1;
+	else
+		lstall = 0;
+	Delay = L * wx * (numSlices - 1) + ax + s + lstall + 22;
+
+	//dsc processes 3 pixel containers per cycle and a container can contain 1 or 2 pixels
+	pixels = Delay * 3 * pixelsPerClock;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: bpc: %d\n", __func__, bpc);
+	dml_print("DML::%s: BPP: %f\n", __func__, BPP);
+	dml_print("DML::%s: sliceWidth: %d\n", __func__, sliceWidth);
+	dml_print("DML::%s: numSlices: %d\n", __func__, numSlices);
+	dml_print("DML::%s: pixelFormat: %d\n", __func__, pixelFormat);
+	dml_print("DML::%s: Output: %d\n", __func__, Output);
+	dml_print("DML::%s: pixels: %d\n", __func__, pixels);
+#endif
+
+	return pixels;
+}
+
+unsigned int dml32_dscComputeDelay(enum output_format_class pixelFormat, enum output_encoder_class Output)
+{
+	unsigned int Delay = 0;
+
+	if (pixelFormat == dm_420) {
+		//   sfr
+		Delay = Delay + 2;
+		//   dsccif
+		Delay = Delay + 0;
+		//   dscc - input deserializer
+		Delay = Delay + 3;
+		//   dscc gets pixels every other cycle
+		Delay = Delay + 2;
+		//   dscc - input cdc fifo
+		Delay = Delay + 12;
+		//   dscc gets pixels every other cycle
+		Delay = Delay + 13;
+		//   dscc - cdc uncertainty
+		Delay = Delay + 2;
+		//   dscc - output cdc fifo
+		Delay = Delay + 7;
+		//   dscc gets pixels every other cycle
+		Delay = Delay + 3;
+		//   dscc - cdc uncertainty
+		Delay = Delay + 2;
+		//   dscc - output serializer
+		Delay = Delay + 1;
+		//   sft
+		Delay = Delay + 1;
+	} else if (pixelFormat == dm_n422 || (pixelFormat != dm_444)) {
+		//   sfr
+		Delay = Delay + 2;
+		//   dsccif
+		Delay = Delay + 1;
+		//   dscc - input deserializer
+		Delay = Delay + 5;
+		//  dscc - input cdc fifo
+		Delay = Delay + 25;
+		//   dscc - cdc uncertainty
+		Delay = Delay + 2;
+		//   dscc - output cdc fifo
+		Delay = Delay + 10;
+		//   dscc - cdc uncertainty
+		Delay = Delay + 2;
+		//   dscc - output serializer
+		Delay = Delay + 1;
+		//   sft
+		Delay = Delay + 1;
+	} else {
+		//   sfr
+		Delay = Delay + 2;
+		//   dsccif
+		Delay = Delay + 0;
+		//   dscc - input deserializer
+		Delay = Delay + 3;
+		//   dscc - input cdc fifo
+		Delay = Delay + 12;
+		//   dscc - cdc uncertainty
+		Delay = Delay + 2;
+		//   dscc - output cdc fifo
+		Delay = Delay + 7;
+		//   dscc - output serializer
+		Delay = Delay + 1;
+		//   dscc - cdc uncertainty
+		Delay = Delay + 2;
+		//   sft
+		Delay = Delay + 1;
+	}
+
+	return Delay;
+}
+
+
+bool IsVertical(enum dm_rotation_angle Scan)
+{
+	bool is_vert = false;
+
+	if (Scan == dm_rotation_90 || Scan == dm_rotation_90m || Scan == dm_rotation_270 || Scan == dm_rotation_270m)
+		is_vert = true;
+	else
+		is_vert = false;
+	return is_vert;
+}
+
+void dml32_CalculateSinglePipeDPPCLKAndSCLThroughput(
+		double HRatio,
+		double HRatioChroma,
+		double VRatio,
+		double VRatioChroma,
+		double MaxDCHUBToPSCLThroughput,
+		double MaxPSCLToLBThroughput,
+		double PixelClock,
+		enum source_format_class SourcePixelFormat,
+		unsigned int HTaps,
+		unsigned int HTapsChroma,
+		unsigned int VTaps,
+		unsigned int VTapsChroma,
+
+		/* output */
+		double *PSCL_THROUGHPUT,
+		double *PSCL_THROUGHPUT_CHROMA,
+		double *DPPCLKUsingSingleDPP)
+{
+	double DPPCLKUsingSingleDPPLuma;
+	double DPPCLKUsingSingleDPPChroma;
+
+	if (HRatio > 1) {
+		*PSCL_THROUGHPUT = dml_min(MaxDCHUBToPSCLThroughput, MaxPSCLToLBThroughput * HRatio /
+				dml_ceil((double) HTaps / 6.0, 1.0));
+	} else {
+		*PSCL_THROUGHPUT = dml_min(MaxDCHUBToPSCLThroughput, MaxPSCLToLBThroughput);
+	}
+
+	DPPCLKUsingSingleDPPLuma = PixelClock * dml_max3(VTaps / 6 * dml_min(1, HRatio), HRatio * VRatio /
+			*PSCL_THROUGHPUT, 1);
+
+	if ((HTaps > 6 || VTaps > 6) && DPPCLKUsingSingleDPPLuma < 2 * PixelClock)
+		DPPCLKUsingSingleDPPLuma = 2 * PixelClock;
+
+	if ((SourcePixelFormat != dm_420_8 && SourcePixelFormat != dm_420_10 && SourcePixelFormat != dm_420_12 &&
+			SourcePixelFormat != dm_rgbe_alpha)) {
+		*PSCL_THROUGHPUT_CHROMA = 0;
+		*DPPCLKUsingSingleDPP = DPPCLKUsingSingleDPPLuma;
+	} else {
+		if (HRatioChroma > 1) {
+			*PSCL_THROUGHPUT_CHROMA = dml_min(MaxDCHUBToPSCLThroughput, MaxPSCLToLBThroughput *
+					HRatioChroma / dml_ceil((double) HTapsChroma / 6.0, 1.0));
+		} else {
+			*PSCL_THROUGHPUT_CHROMA = dml_min(MaxDCHUBToPSCLThroughput, MaxPSCLToLBThroughput);
+		}
+		DPPCLKUsingSingleDPPChroma = PixelClock * dml_max3(VTapsChroma / 6 * dml_min(1, HRatioChroma),
+				HRatioChroma * VRatioChroma / *PSCL_THROUGHPUT_CHROMA, 1);
+		if ((HTapsChroma > 6 || VTapsChroma > 6) && DPPCLKUsingSingleDPPChroma < 2 * PixelClock)
+			DPPCLKUsingSingleDPPChroma = 2 * PixelClock;
+		*DPPCLKUsingSingleDPP = dml_max(DPPCLKUsingSingleDPPLuma, DPPCLKUsingSingleDPPChroma);
+	}
+}
+
+void dml32_CalculateBytePerPixelAndBlockSizes(
+		enum source_format_class SourcePixelFormat,
+		enum dm_swizzle_mode SurfaceTiling,
+
+		/* Output */
+		unsigned int *BytePerPixelY,
+		unsigned int *BytePerPixelC,
+		double  *BytePerPixelDETY,
+		double  *BytePerPixelDETC,
+		unsigned int *BlockHeight256BytesY,
+		unsigned int *BlockHeight256BytesC,
+		unsigned int *BlockWidth256BytesY,
+		unsigned int *BlockWidth256BytesC,
+		unsigned int *MacroTileHeightY,
+		unsigned int *MacroTileHeightC,
+		unsigned int *MacroTileWidthY,
+		unsigned int *MacroTileWidthC)
+{
+	if (SourcePixelFormat == dm_444_64) {
+		*BytePerPixelDETY = 8;
+		*BytePerPixelDETC = 0;
+		*BytePerPixelY = 8;
+		*BytePerPixelC = 0;
+	} else if (SourcePixelFormat == dm_444_32 || SourcePixelFormat == dm_rgbe) {
+		*BytePerPixelDETY = 4;
+		*BytePerPixelDETC = 0;
+		*BytePerPixelY = 4;
+		*BytePerPixelC = 0;
+	} else if (SourcePixelFormat == dm_444_16) {
+		*BytePerPixelDETY = 2;
+		*BytePerPixelDETC = 0;
+		*BytePerPixelY = 2;
+		*BytePerPixelC = 0;
+	} else if (SourcePixelFormat == dm_444_8) {
+		*BytePerPixelDETY = 1;
+		*BytePerPixelDETC = 0;
+		*BytePerPixelY = 1;
+		*BytePerPixelC = 0;
+	} else if (SourcePixelFormat == dm_rgbe_alpha) {
+		*BytePerPixelDETY = 4;
+		*BytePerPixelDETC = 1;
+		*BytePerPixelY = 4;
+		*BytePerPixelC = 1;
+	} else if (SourcePixelFormat == dm_420_8) {
+		*BytePerPixelDETY = 1;
+		*BytePerPixelDETC = 2;
+		*BytePerPixelY = 1;
+		*BytePerPixelC = 2;
+	} else if (SourcePixelFormat == dm_420_12) {
+		*BytePerPixelDETY = 2;
+		*BytePerPixelDETC = 4;
+		*BytePerPixelY = 2;
+		*BytePerPixelC = 4;
+	} else {
+		*BytePerPixelDETY = 4.0 / 3;
+		*BytePerPixelDETC = 8.0 / 3;
+		*BytePerPixelY = 2;
+		*BytePerPixelC = 4;
+	}
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: SourcePixelFormat = %d\n", __func__, SourcePixelFormat);
+	dml_print("DML::%s: BytePerPixelDETY = %f\n", __func__, *BytePerPixelDETY);
+	dml_print("DML::%s: BytePerPixelDETC = %f\n", __func__, *BytePerPixelDETC);
+	dml_print("DML::%s: BytePerPixelY    = %d\n", __func__, *BytePerPixelY);
+	dml_print("DML::%s: BytePerPixelC    = %d\n", __func__, *BytePerPixelC);
+#endif
+	if ((SourcePixelFormat == dm_444_64 || SourcePixelFormat == dm_444_32
+			|| SourcePixelFormat == dm_444_16
+			|| SourcePixelFormat == dm_444_8
+			|| SourcePixelFormat == dm_mono_16
+			|| SourcePixelFormat == dm_mono_8
+			|| SourcePixelFormat == dm_rgbe)) {
+		if (SurfaceTiling == dm_sw_linear)
+			*BlockHeight256BytesY = 1;
+		else if (SourcePixelFormat == dm_444_64)
+			*BlockHeight256BytesY = 4;
+		else if (SourcePixelFormat == dm_444_8)
+			*BlockHeight256BytesY = 16;
+		else
+			*BlockHeight256BytesY = 8;
+
+		*BlockWidth256BytesY = 256U / *BytePerPixelY / *BlockHeight256BytesY;
+		*BlockHeight256BytesC = 0;
+		*BlockWidth256BytesC = 0;
+	} else {
+		if (SurfaceTiling == dm_sw_linear) {
+			*BlockHeight256BytesY = 1;
+			*BlockHeight256BytesC = 1;
+		} else if (SourcePixelFormat == dm_rgbe_alpha) {
+			*BlockHeight256BytesY = 8;
+			*BlockHeight256BytesC = 16;
+		} else if (SourcePixelFormat == dm_420_8) {
+			*BlockHeight256BytesY = 16;
+			*BlockHeight256BytesC = 8;
+		} else {
+			*BlockHeight256BytesY = 8;
+			*BlockHeight256BytesC = 8;
+		}
+		*BlockWidth256BytesY = 256U / *BytePerPixelY / *BlockHeight256BytesY;
+		*BlockWidth256BytesC = 256U / *BytePerPixelC / *BlockHeight256BytesC;
+	}
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: BlockWidth256BytesY  = %d\n", __func__, *BlockWidth256BytesY);
+	dml_print("DML::%s: BlockHeight256BytesY = %d\n", __func__, *BlockHeight256BytesY);
+	dml_print("DML::%s: BlockWidth256BytesC  = %d\n", __func__, *BlockWidth256BytesC);
+	dml_print("DML::%s: BlockHeight256BytesC = %d\n", __func__, *BlockHeight256BytesC);
+#endif
+
+	if (SurfaceTiling == dm_sw_linear) {
+		*MacroTileHeightY = *BlockHeight256BytesY;
+		*MacroTileWidthY = 256 / *BytePerPixelY / *MacroTileHeightY;
+		*MacroTileHeightC = *BlockHeight256BytesC;
+		if (*MacroTileHeightC == 0)
+			*MacroTileWidthC = 0;
+		else
+			*MacroTileWidthC = 256 / *BytePerPixelC / *MacroTileHeightC;
+	} else if (SurfaceTiling == dm_sw_64kb_d || SurfaceTiling == dm_sw_64kb_d_t ||
+			SurfaceTiling == dm_sw_64kb_d_x || SurfaceTiling == dm_sw_64kb_r_x) {
+		*MacroTileHeightY = 16 * *BlockHeight256BytesY;
+		*MacroTileWidthY = 65536 / *BytePerPixelY / *MacroTileHeightY;
+		*MacroTileHeightC = 16 * *BlockHeight256BytesC;
+		if (*MacroTileHeightC == 0)
+			*MacroTileWidthC = 0;
+		else
+			*MacroTileWidthC = 65536 / *BytePerPixelC / *MacroTileHeightC;
+	} else {
+		*MacroTileHeightY = 32 * *BlockHeight256BytesY;
+		*MacroTileWidthY = 65536 * 4 / *BytePerPixelY / *MacroTileHeightY;
+		*MacroTileHeightC = 32 * *BlockHeight256BytesC;
+		if (*MacroTileHeightC == 0)
+			*MacroTileWidthC = 0;
+		else
+			*MacroTileWidthC = 65536 * 4 / *BytePerPixelC / *MacroTileHeightC;
+	}
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: MacroTileWidthY  = %d\n", __func__, *MacroTileWidthY);
+	dml_print("DML::%s: MacroTileHeightY = %d\n", __func__, *MacroTileHeightY);
+	dml_print("DML::%s: MacroTileWidthC  = %d\n", __func__, *MacroTileWidthC);
+	dml_print("DML::%s: MacroTileHeightC = %d\n", __func__, *MacroTileHeightC);
+#endif
+} // CalculateBytePerPixelAndBlockSizes
+
+void dml32_CalculateSwathAndDETConfiguration(
+		unsigned int DETSizeOverride[],
+		enum dm_use_mall_for_pstate_change_mode UseMALLForPStateChange[],
+		unsigned int ConfigReturnBufferSizeInKByte,
+		unsigned int MaxTotalDETInKByte,
+		unsigned int MinCompressedBufferSizeInKByte,
+		double ForceSingleDPP,
+		unsigned int NumberOfActiveSurfaces,
+		unsigned int nomDETInKByte,
+		enum unbounded_requesting_policy UseUnboundedRequestingFinal,
+		bool DisableUnboundRequestIfCompBufReservedSpaceNeedAdjustment,
+		unsigned int PixelChunkSizeKBytes,
+		unsigned int ROBSizeKBytes,
+		unsigned int CompressedBufferSegmentSizeInkByteFinal,
+		enum output_encoder_class Output[],
+		double ReadBandwidthLuma[],
+		double ReadBandwidthChroma[],
+		double MaximumSwathWidthLuma[],
+		double MaximumSwathWidthChroma[],
+		enum dm_rotation_angle SourceRotation[],
+		bool ViewportStationary[],
+		enum source_format_class SourcePixelFormat[],
+		enum dm_swizzle_mode SurfaceTiling[],
+		unsigned int ViewportWidth[],
+		unsigned int ViewportHeight[],
+		unsigned int ViewportXStart[],
+		unsigned int ViewportYStart[],
+		unsigned int ViewportXStartC[],
+		unsigned int ViewportYStartC[],
+		unsigned int SurfaceWidthY[],
+		unsigned int SurfaceWidthC[],
+		unsigned int SurfaceHeightY[],
+		unsigned int SurfaceHeightC[],
+		unsigned int Read256BytesBlockHeightY[],
+		unsigned int Read256BytesBlockHeightC[],
+		unsigned int Read256BytesBlockWidthY[],
+		unsigned int Read256BytesBlockWidthC[],
+		enum odm_combine_mode ODMMode[],
+		unsigned int BlendingAndTiming[],
+		unsigned int BytePerPixY[],
+		unsigned int BytePerPixC[],
+		double BytePerPixDETY[],
+		double BytePerPixDETC[],
+		unsigned int HActive[],
+		double HRatio[],
+		double HRatioChroma[],
+		unsigned int DPPPerSurface[],
+
+		/* Output */
+		unsigned int swath_width_luma_ub[],
+		unsigned int swath_width_chroma_ub[],
+		double SwathWidth[],
+		double SwathWidthChroma[],
+		unsigned int SwathHeightY[],
+		unsigned int SwathHeightC[],
+		unsigned int DETBufferSizeInKByte[],
+		unsigned int DETBufferSizeY[],
+		unsigned int DETBufferSizeC[],
+		bool *UnboundedRequestEnabled,
+		unsigned int *CompressedBufferSizeInkByte,
+		unsigned int *CompBufReservedSpaceKBytes,
+		bool *CompBufReservedSpaceNeedAdjustment,
+		bool ViewportSizeSupportPerSurface[],
+		bool *ViewportSizeSupport)
+{
+	unsigned int MaximumSwathHeightY[DC__NUM_DPP__MAX];
+	unsigned int MaximumSwathHeightC[DC__NUM_DPP__MAX];
+	unsigned int RoundedUpMaxSwathSizeBytesY[DC__NUM_DPP__MAX];
+	unsigned int RoundedUpMaxSwathSizeBytesC[DC__NUM_DPP__MAX];
+	unsigned int RoundedUpSwathSizeBytesY;
+	unsigned int RoundedUpSwathSizeBytesC;
+	double SwathWidthdoubleDPP[DC__NUM_DPP__MAX];
+	double SwathWidthdoubleDPPChroma[DC__NUM_DPP__MAX];
+	unsigned int k;
+	unsigned int TotalActiveDPP = 0;
+	bool NoChromaSurfaces = true;
+	unsigned int DETBufferSizeInKByteForSwathCalculation;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: ForceSingleDPP = %d\n", __func__, ForceSingleDPP);
+	dml_print("DML::%s: ROBSizeKBytes = %d\n", __func__, ROBSizeKBytes);
+	dml_print("DML::%s: PixelChunkSizeKBytes = %d\n", __func__, PixelChunkSizeKBytes);
+#endif
+	dml32_CalculateSwathWidth(ForceSingleDPP,
+			NumberOfActiveSurfaces,
+			SourcePixelFormat,
+			SourceRotation,
+			ViewportStationary,
+			ViewportWidth,
+			ViewportHeight,
+			ViewportXStart,
+			ViewportYStart,
+			ViewportXStartC,
+			ViewportYStartC,
+			SurfaceWidthY,
+			SurfaceWidthC,
+			SurfaceHeightY,
+			SurfaceHeightC,
+			ODMMode,
+			BytePerPixY,
+			BytePerPixC,
+			Read256BytesBlockHeightY,
+			Read256BytesBlockHeightC,
+			Read256BytesBlockWidthY,
+			Read256BytesBlockWidthC,
+			BlendingAndTiming,
+			HActive,
+			HRatio,
+			DPPPerSurface,
+
+			/* Output */
+			SwathWidthdoubleDPP,
+			SwathWidthdoubleDPPChroma,
+			SwathWidth,
+			SwathWidthChroma,
+			MaximumSwathHeightY,
+			MaximumSwathHeightC,
+			swath_width_luma_ub,
+			swath_width_chroma_ub);
+
+	for (k = 0; k < NumberOfActiveSurfaces; ++k) {
+		RoundedUpMaxSwathSizeBytesY[k] = swath_width_luma_ub[k] * BytePerPixDETY[k] * MaximumSwathHeightY[k];
+		RoundedUpMaxSwathSizeBytesC[k] = swath_width_chroma_ub[k] * BytePerPixDETC[k] * MaximumSwathHeightC[k];
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%0d DPPPerSurface = %d\n", __func__, k, DPPPerSurface[k]);
+		dml_print("DML::%s: k=%0d swath_width_luma_ub = %d\n", __func__, k, swath_width_luma_ub[k]);
+		dml_print("DML::%s: k=%0d BytePerPixDETY = %f\n", __func__, k, BytePerPixDETY[k]);
+		dml_print("DML::%s: k=%0d MaximumSwathHeightY = %d\n", __func__, k, MaximumSwathHeightY[k]);
+		dml_print("DML::%s: k=%0d RoundedUpMaxSwathSizeBytesY = %d\n", __func__, k,
+				RoundedUpMaxSwathSizeBytesY[k]);
+		dml_print("DML::%s: k=%0d swath_width_chroma_ub = %d\n", __func__, k, swath_width_chroma_ub[k]);
+		dml_print("DML::%s: k=%0d BytePerPixDETC = %f\n", __func__, k, BytePerPixDETC[k]);
+		dml_print("DML::%s: k=%0d MaximumSwathHeightC = %d\n", __func__, k, MaximumSwathHeightC[k]);
+		dml_print("DML::%s: k=%0d RoundedUpMaxSwathSizeBytesC = %d\n", __func__, k,
+				RoundedUpMaxSwathSizeBytesC[k]);
+#endif
+
+		if (SourcePixelFormat[k] == dm_420_10) {
+			RoundedUpMaxSwathSizeBytesY[k] = dml_ceil((unsigned int) RoundedUpMaxSwathSizeBytesY[k], 256);
+			RoundedUpMaxSwathSizeBytesC[k] = dml_ceil((unsigned int) RoundedUpMaxSwathSizeBytesC[k], 256);
+		}
+	}
+
+	for (k = 0; k < NumberOfActiveSurfaces; ++k) {
+		TotalActiveDPP = TotalActiveDPP + (ForceSingleDPP ? 1 : DPPPerSurface[k]);
+		if (SourcePixelFormat[k] == dm_420_8 || SourcePixelFormat[k] == dm_420_10 ||
+				SourcePixelFormat[k] == dm_420_12 || SourcePixelFormat[k] == dm_rgbe_alpha) {
+			NoChromaSurfaces = false;
+		}
+	}
+
+	// By default, just set the reserved space to 2 pixel chunks size
+	*CompBufReservedSpaceKBytes = PixelChunkSizeKBytes * 2;
+
+	// if unbounded req is enabled, program reserved space such that the ROB will not hold more than 8 swaths worth of data
+	// - assume worst-case compression rate of 4. [ROB size - 8 * swath_size / max_compression ratio]
+	// - assume for "narrow" vp case in which the ROB can fit 8 swaths, the DET should be big enough to do full size req
+	*CompBufReservedSpaceNeedAdjustment = ((int) ROBSizeKBytes - (int) *CompBufReservedSpaceKBytes) > (int) (RoundedUpMaxSwathSizeBytesY[0]/512);
+
+	if (*CompBufReservedSpaceNeedAdjustment == 1) {
+		*CompBufReservedSpaceKBytes = ROBSizeKBytes - RoundedUpMaxSwathSizeBytesY[0]/512;
+	}
+
+	#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: CompBufReservedSpaceKBytes          = %d\n",  __func__, *CompBufReservedSpaceKBytes);
+		dml_print("DML::%s: CompBufReservedSpaceNeedAdjustment  = %d\n",  __func__, *CompBufReservedSpaceNeedAdjustment);
+	#endif
+
+	*UnboundedRequestEnabled = dml32_UnboundedRequest(UseUnboundedRequestingFinal, TotalActiveDPP, NoChromaSurfaces, Output[0], SurfaceTiling[0], *CompBufReservedSpaceNeedAdjustment, DisableUnboundRequestIfCompBufReservedSpaceNeedAdjustment);
+
+	dml32_CalculateDETBufferSize(DETSizeOverride,
+			UseMALLForPStateChange,
+			ForceSingleDPP,
+			NumberOfActiveSurfaces,
+			*UnboundedRequestEnabled,
+			nomDETInKByte,
+			MaxTotalDETInKByte,
+			ConfigReturnBufferSizeInKByte,
+			MinCompressedBufferSizeInKByte,
+			CompressedBufferSegmentSizeInkByteFinal,
+			SourcePixelFormat,
+			ReadBandwidthLuma,
+			ReadBandwidthChroma,
+			RoundedUpMaxSwathSizeBytesY,
+			RoundedUpMaxSwathSizeBytesC,
+			DPPPerSurface,
+
+			/* Output */
+			DETBufferSizeInKByte,    // per hubp pipe
+			CompressedBufferSizeInkByte);
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: TotalActiveDPP = %d\n", __func__, TotalActiveDPP);
+	dml_print("DML::%s: nomDETInKByte = %d\n", __func__, nomDETInKByte);
+	dml_print("DML::%s: ConfigReturnBufferSizeInKByte = %d\n", __func__, ConfigReturnBufferSizeInKByte);
+	dml_print("DML::%s: UseUnboundedRequestingFinal = %d\n", __func__, UseUnboundedRequestingFinal);
+	dml_print("DML::%s: UnboundedRequestEnabled = %d\n", __func__, *UnboundedRequestEnabled);
+	dml_print("DML::%s: CompressedBufferSizeInkByte = %d\n", __func__, *CompressedBufferSizeInkByte);
+#endif
+
+	*ViewportSizeSupport = true;
+	for (k = 0; k < NumberOfActiveSurfaces; ++k) {
+
+		DETBufferSizeInKByteForSwathCalculation = (UseMALLForPStateChange[k] ==
+				dm_use_mall_pstate_change_phantom_pipe ? 1024 : DETBufferSizeInKByte[k]);
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%0d DETBufferSizeInKByteForSwathCalculation = %d\n", __func__, k,
+				DETBufferSizeInKByteForSwathCalculation);
+#endif
+
+		if (RoundedUpMaxSwathSizeBytesY[k] + RoundedUpMaxSwathSizeBytesC[k] <=
+				DETBufferSizeInKByteForSwathCalculation * 1024 / 2) {
+			SwathHeightY[k] = MaximumSwathHeightY[k];
+			SwathHeightC[k] = MaximumSwathHeightC[k];
+			RoundedUpSwathSizeBytesY = RoundedUpMaxSwathSizeBytesY[k];
+			RoundedUpSwathSizeBytesC = RoundedUpMaxSwathSizeBytesC[k];
+		} else if (RoundedUpMaxSwathSizeBytesY[k] >= 1.5 * RoundedUpMaxSwathSizeBytesC[k] &&
+				RoundedUpMaxSwathSizeBytesY[k] / 2 + RoundedUpMaxSwathSizeBytesC[k] <=
+				DETBufferSizeInKByteForSwathCalculation * 1024 / 2) {
+			SwathHeightY[k] = MaximumSwathHeightY[k] / 2;
+			SwathHeightC[k] = MaximumSwathHeightC[k];
+			RoundedUpSwathSizeBytesY = RoundedUpMaxSwathSizeBytesY[k] / 2;
+			RoundedUpSwathSizeBytesC = RoundedUpMaxSwathSizeBytesC[k];
+		} else if (RoundedUpMaxSwathSizeBytesY[k] < 1.5 * RoundedUpMaxSwathSizeBytesC[k] &&
+				RoundedUpMaxSwathSizeBytesY[k] + RoundedUpMaxSwathSizeBytesC[k] / 2 <=
+				DETBufferSizeInKByteForSwathCalculation * 1024 / 2) {
+			SwathHeightY[k] = MaximumSwathHeightY[k];
+			SwathHeightC[k] = MaximumSwathHeightC[k] / 2;
+			RoundedUpSwathSizeBytesY = RoundedUpMaxSwathSizeBytesY[k];
+			RoundedUpSwathSizeBytesC = RoundedUpMaxSwathSizeBytesC[k] / 2;
+		} else {
+			SwathHeightY[k] = MaximumSwathHeightY[k] / 2;
+			SwathHeightC[k] = MaximumSwathHeightC[k] / 2;
+			RoundedUpSwathSizeBytesY = RoundedUpMaxSwathSizeBytesY[k] / 2;
+			RoundedUpSwathSizeBytesC = RoundedUpMaxSwathSizeBytesC[k] / 2;
+		}
+
+		if ((RoundedUpMaxSwathSizeBytesY[k] / 2 + RoundedUpMaxSwathSizeBytesC[k] / 2 >
+				DETBufferSizeInKByteForSwathCalculation * 1024 / 2)
+				|| SwathWidth[k] > MaximumSwathWidthLuma[k] || (SwathHeightC[k] > 0 &&
+						SwathWidthChroma[k] > MaximumSwathWidthChroma[k])) {
+			*ViewportSizeSupport = false;
+			ViewportSizeSupportPerSurface[k] = false;
+		} else {
+			ViewportSizeSupportPerSurface[k] = true;
+		}
+
+		if (SwathHeightC[k] == 0) {
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: k=%0d All DET for plane0\n", __func__, k);
+#endif
+			DETBufferSizeY[k] = DETBufferSizeInKByte[k] * 1024;
+			DETBufferSizeC[k] = 0;
+		} else if (RoundedUpSwathSizeBytesY <= 1.5 * RoundedUpSwathSizeBytesC) {
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: k=%0d Half DET for plane0, half for plane1\n", __func__, k);
+#endif
+			DETBufferSizeY[k] = DETBufferSizeInKByte[k] * 1024 / 2;
+			DETBufferSizeC[k] = DETBufferSizeInKByte[k] * 1024 / 2;
+		} else {
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: k=%0d 2/3 DET for plane0, 1/3 for plane1\n", __func__, k);
+#endif
+			DETBufferSizeY[k] = dml_floor(DETBufferSizeInKByte[k] * 1024 * 2 / 3, 1024);
+			DETBufferSizeC[k] = DETBufferSizeInKByte[k] * 1024 - DETBufferSizeY[k];
+		}
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%0d SwathHeightY = %d\n", __func__, k, SwathHeightY[k]);
+		dml_print("DML::%s: k=%0d SwathHeightC = %d\n", __func__, k, SwathHeightC[k]);
+		dml_print("DML::%s: k=%0d RoundedUpMaxSwathSizeBytesY = %d\n", __func__,
+				k, RoundedUpMaxSwathSizeBytesY[k]);
+		dml_print("DML::%s: k=%0d RoundedUpMaxSwathSizeBytesC = %d\n", __func__,
+				k, RoundedUpMaxSwathSizeBytesC[k]);
+		dml_print("DML::%s: k=%0d RoundedUpSwathSizeBytesY = %d\n", __func__, k, RoundedUpSwathSizeBytesY);
+		dml_print("DML::%s: k=%0d RoundedUpSwathSizeBytesC = %d\n", __func__, k, RoundedUpSwathSizeBytesC);
+		dml_print("DML::%s: k=%0d DETBufferSizeInKByte = %d\n", __func__, k, DETBufferSizeInKByte[k]);
+		dml_print("DML::%s: k=%0d DETBufferSizeY = %d\n", __func__, k, DETBufferSizeY[k]);
+		dml_print("DML::%s: k=%0d DETBufferSizeC = %d\n", __func__, k, DETBufferSizeC[k]);
+		dml_print("DML::%s: k=%0d ViewportSizeSupportPerSurface = %d\n", __func__, k,
+				ViewportSizeSupportPerSurface[k]);
+#endif
+
+	}
+} // CalculateSwathAndDETConfiguration
+
+void dml32_CalculateSwathWidth(
+		bool				ForceSingleDPP,
+		unsigned int			NumberOfActiveSurfaces,
+		enum source_format_class	SourcePixelFormat[],
+		enum dm_rotation_angle		SourceRotation[],
+		bool				ViewportStationary[],
+		unsigned int			ViewportWidth[],
+		unsigned int			ViewportHeight[],
+		unsigned int			ViewportXStart[],
+		unsigned int			ViewportYStart[],
+		unsigned int			ViewportXStartC[],
+		unsigned int			ViewportYStartC[],
+		unsigned int			SurfaceWidthY[],
+		unsigned int			SurfaceWidthC[],
+		unsigned int			SurfaceHeightY[],
+		unsigned int			SurfaceHeightC[],
+		enum odm_combine_mode		ODMMode[],
+		unsigned int			BytePerPixY[],
+		unsigned int			BytePerPixC[],
+		unsigned int			Read256BytesBlockHeightY[],
+		unsigned int			Read256BytesBlockHeightC[],
+		unsigned int			Read256BytesBlockWidthY[],
+		unsigned int			Read256BytesBlockWidthC[],
+		unsigned int			BlendingAndTiming[],
+		unsigned int			HActive[],
+		double				HRatio[],
+		unsigned int			DPPPerSurface[],
+
+		/* Output */
+		double				SwathWidthdoubleDPPY[],
+		double				SwathWidthdoubleDPPC[],
+		double				SwathWidthY[], // per-pipe
+		double				SwathWidthC[], // per-pipe
+		unsigned int			MaximumSwathHeightY[],
+		unsigned int			MaximumSwathHeightC[],
+		unsigned int			swath_width_luma_ub[], // per-pipe
+		unsigned int			swath_width_chroma_ub[]) // per-pipe
+{
+	unsigned int k, j;
+	enum odm_combine_mode MainSurfaceODMMode;
+
+	unsigned int surface_width_ub_l;
+	unsigned int surface_height_ub_l;
+	unsigned int surface_width_ub_c = 0;
+	unsigned int surface_height_ub_c = 0;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: ForceSingleDPP = %d\n", __func__, ForceSingleDPP);
+	dml_print("DML::%s: NumberOfActiveSurfaces = %d\n", __func__, NumberOfActiveSurfaces);
+#endif
+
+	for (k = 0; k < NumberOfActiveSurfaces; ++k) {
+		if (!IsVertical(SourceRotation[k]))
+			SwathWidthdoubleDPPY[k] = ViewportWidth[k];
+		else
+			SwathWidthdoubleDPPY[k] = ViewportHeight[k];
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%d ViewportWidth=%d\n", __func__, k, ViewportWidth[k]);
+		dml_print("DML::%s: k=%d ViewportHeight=%d\n", __func__, k, ViewportHeight[k]);
+#endif
+
+		MainSurfaceODMMode = ODMMode[k];
+		for (j = 0; j < NumberOfActiveSurfaces; ++j) {
+			if (BlendingAndTiming[k] == j)
+				MainSurfaceODMMode = ODMMode[j];
+		}
+
+		if (ForceSingleDPP) {
+			SwathWidthY[k] = SwathWidthdoubleDPPY[k];
+		} else {
+			if (MainSurfaceODMMode == dm_odm_combine_mode_4to1) {
+				SwathWidthY[k] = dml_min(SwathWidthdoubleDPPY[k],
+						dml_round(HActive[k] / 4.0 * HRatio[k]));
+			} else if (MainSurfaceODMMode == dm_odm_combine_mode_2to1) {
+				SwathWidthY[k] = dml_min(SwathWidthdoubleDPPY[k],
+						dml_round(HActive[k] / 2.0 * HRatio[k]));
+			} else if (DPPPerSurface[k] == 2) {
+				SwathWidthY[k] = SwathWidthdoubleDPPY[k] / 2;
+			} else {
+				SwathWidthY[k] = SwathWidthdoubleDPPY[k];
+			}
+		}
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%d HActive=%d\n", __func__, k, HActive[k]);
+		dml_print("DML::%s: k=%d HRatio=%f\n", __func__, k, HRatio[k]);
+		dml_print("DML::%s: k=%d MainSurfaceODMMode=%d\n", __func__, k, MainSurfaceODMMode);
+		dml_print("DML::%s: k=%d SwathWidthdoubleDPPY=%d\n", __func__, k, SwathWidthdoubleDPPY[k]);
+		dml_print("DML::%s: k=%d SwathWidthY=%d\n", __func__, k, SwathWidthY[k]);
+#endif
+
+		if (SourcePixelFormat[k] == dm_420_8 || SourcePixelFormat[k] == dm_420_10 ||
+				SourcePixelFormat[k] == dm_420_12) {
+			SwathWidthC[k] = SwathWidthY[k] / 2;
+			SwathWidthdoubleDPPC[k] = SwathWidthdoubleDPPY[k] / 2;
+		} else {
+			SwathWidthC[k] = SwathWidthY[k];
+			SwathWidthdoubleDPPC[k] = SwathWidthdoubleDPPY[k];
+		}
+
+		if (ForceSingleDPP == true) {
+			SwathWidthY[k] = SwathWidthdoubleDPPY[k];
+			SwathWidthC[k] = SwathWidthdoubleDPPC[k];
+		}
+
+		surface_width_ub_l  = dml_ceil(SurfaceWidthY[k], Read256BytesBlockWidthY[k]);
+		surface_height_ub_l = dml_ceil(SurfaceHeightY[k], Read256BytesBlockHeightY[k]);
+
+		if (!IsVertical(SourceRotation[k])) {
+			MaximumSwathHeightY[k] = Read256BytesBlockHeightY[k];
+			MaximumSwathHeightC[k] = Read256BytesBlockHeightC[k];
+			if (ViewportStationary[k] && DPPPerSurface[k] == 1) {
+				swath_width_luma_ub[k] = dml_min(surface_width_ub_l,
+						dml_floor(ViewportXStart[k] +
+								SwathWidthY[k] +
+								Read256BytesBlockWidthY[k] - 1,
+								Read256BytesBlockWidthY[k]) -
+								dml_floor(ViewportXStart[k],
+								Read256BytesBlockWidthY[k]));
+			} else {
+				swath_width_luma_ub[k] = dml_min(surface_width_ub_l,
+						dml_ceil(SwathWidthY[k] - 1,
+								Read256BytesBlockWidthY[k]) +
+								Read256BytesBlockWidthY[k]);
+			}
+			if (BytePerPixC[k] > 0) {
+				surface_width_ub_c  = dml_ceil(SurfaceWidthC[k], Read256BytesBlockWidthC[k]);
+				if (ViewportStationary[k] && DPPPerSurface[k] == 1) {
+					swath_width_chroma_ub[k] = dml_min(surface_width_ub_c,
+							dml_floor(ViewportXStartC[k] + SwathWidthC[k] +
+									Read256BytesBlockWidthC[k] - 1,
+									Read256BytesBlockWidthC[k]) -
+									dml_floor(ViewportXStartC[k],
+									Read256BytesBlockWidthC[k]));
+				} else {
+					swath_width_chroma_ub[k] = dml_min(surface_width_ub_c,
+							dml_ceil(SwathWidthC[k] - 1,
+								Read256BytesBlockWidthC[k]) +
+								Read256BytesBlockWidthC[k]);
+				}
+			} else {
+				swath_width_chroma_ub[k] = 0;
+			}
+		} else {
+			MaximumSwathHeightY[k] = Read256BytesBlockWidthY[k];
+			MaximumSwathHeightC[k] = Read256BytesBlockWidthC[k];
+
+			if (ViewportStationary[k] && DPPPerSurface[k] == 1) {
+				swath_width_luma_ub[k] = dml_min(surface_height_ub_l, dml_floor(ViewportYStart[k] +
+						SwathWidthY[k] + Read256BytesBlockHeightY[k] - 1,
+						Read256BytesBlockHeightY[k]) -
+						dml_floor(ViewportYStart[k], Read256BytesBlockHeightY[k]));
+			} else {
+				swath_width_luma_ub[k] = dml_min(surface_height_ub_l, dml_ceil(SwathWidthY[k] - 1,
+						Read256BytesBlockHeightY[k]) + Read256BytesBlockHeightY[k]);
+			}
+			if (BytePerPixC[k] > 0) {
+				surface_height_ub_c = dml_ceil(SurfaceHeightC[k], Read256BytesBlockHeightC[k]);
+				if (ViewportStationary[k] && DPPPerSurface[k] == 1) {
+					swath_width_chroma_ub[k] = dml_min(surface_height_ub_c,
+							dml_floor(ViewportYStartC[k] + SwathWidthC[k] +
+									Read256BytesBlockHeightC[k] - 1,
+									Read256BytesBlockHeightC[k]) -
+									dml_floor(ViewportYStartC[k],
+											Read256BytesBlockHeightC[k]));
+				} else {
+					swath_width_chroma_ub[k] = dml_min(surface_height_ub_c,
+							dml_ceil(SwathWidthC[k] - 1, Read256BytesBlockHeightC[k]) +
+							Read256BytesBlockHeightC[k]);
+				}
+			} else {
+				swath_width_chroma_ub[k] = 0;
+			}
+		}
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%d surface_width_ub_l=%0d\n", __func__, k, surface_width_ub_l);
+		dml_print("DML::%s: k=%d surface_height_ub_l=%0d\n", __func__, k, surface_height_ub_l);
+		dml_print("DML::%s: k=%d surface_width_ub_c=%0d\n", __func__, k, surface_width_ub_c);
+		dml_print("DML::%s: k=%d surface_height_ub_c=%0d\n", __func__, k, surface_height_ub_c);
+		dml_print("DML::%s: k=%d Read256BytesBlockWidthY=%0d\n", __func__, k, Read256BytesBlockWidthY[k]);
+		dml_print("DML::%s: k=%d Read256BytesBlockHeightY=%0d\n", __func__, k, Read256BytesBlockHeightY[k]);
+		dml_print("DML::%s: k=%d Read256BytesBlockWidthC=%0d\n", __func__, k, Read256BytesBlockWidthC[k]);
+		dml_print("DML::%s: k=%d Read256BytesBlockHeightC=%0d\n", __func__, k, Read256BytesBlockHeightC[k]);
+		dml_print("DML::%s: k=%d ViewportStationary=%0d\n", __func__, k, ViewportStationary[k]);
+		dml_print("DML::%s: k=%d DPPPerSurface=%0d\n", __func__, k, DPPPerSurface[k]);
+		dml_print("DML::%s: k=%d swath_width_luma_ub=%0d\n", __func__, k, swath_width_luma_ub[k]);
+		dml_print("DML::%s: k=%d swath_width_chroma_ub=%0d\n", __func__, k, swath_width_chroma_ub[k]);
+		dml_print("DML::%s: k=%d MaximumSwathHeightY=%0d\n", __func__, k, MaximumSwathHeightY[k]);
+		dml_print("DML::%s: k=%d MaximumSwathHeightC=%0d\n", __func__, k, MaximumSwathHeightC[k]);
+#endif
+
+	}
+} // CalculateSwathWidth
+
+bool dml32_UnboundedRequest(enum unbounded_requesting_policy UseUnboundedRequestingFinal,
+			unsigned int TotalNumberOfActiveDPP,
+			bool NoChroma,
+			enum output_encoder_class Output,
+			enum dm_swizzle_mode SurfaceTiling,
+			bool CompBufReservedSpaceNeedAdjustment,
+			bool DisableUnboundRequestIfCompBufReservedSpaceNeedAdjustment)
+{
+	bool ret_val = false;
+
+	ret_val = (UseUnboundedRequestingFinal != dm_unbounded_requesting_disable &&
+			TotalNumberOfActiveDPP == 1 && NoChroma);
+	if (UseUnboundedRequestingFinal == dm_unbounded_requesting_edp_only && Output != dm_edp)
+		ret_val = false;
+
+	if (SurfaceTiling == dm_sw_linear)
+		ret_val = false;
+
+	if (CompBufReservedSpaceNeedAdjustment == 1 && DisableUnboundRequestIfCompBufReservedSpaceNeedAdjustment)
+		ret_val = false;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: CompBufReservedSpaceNeedAdjustment  = %d\n",  __func__, CompBufReservedSpaceNeedAdjustment);
+	dml_print("DML::%s: DisableUnboundRequestIfCompBufReservedSpaceNeedAdjustment  = %d\n",  __func__, DisableUnboundRequestIfCompBufReservedSpaceNeedAdjustment);
+	dml_print("DML::%s: ret_val = %d\n",  __func__, ret_val);
+#endif
+
+	return (ret_val);
+}
+
+void dml32_CalculateDETBufferSize(
+		unsigned int DETSizeOverride[],
+		enum dm_use_mall_for_pstate_change_mode UseMALLForPStateChange[],
+		bool ForceSingleDPP,
+		unsigned int NumberOfActiveSurfaces,
+		bool UnboundedRequestEnabled,
+		unsigned int nomDETInKByte,
+		unsigned int MaxTotalDETInKByte,
+		unsigned int ConfigReturnBufferSizeInKByte,
+		unsigned int MinCompressedBufferSizeInKByte,
+		unsigned int CompressedBufferSegmentSizeInkByteFinal,
+		enum source_format_class SourcePixelFormat[],
+		double ReadBandwidthLuma[],
+		double ReadBandwidthChroma[],
+		unsigned int RoundedUpMaxSwathSizeBytesY[],
+		unsigned int RoundedUpMaxSwathSizeBytesC[],
+		unsigned int DPPPerSurface[],
+		/* Output */
+		unsigned int DETBufferSizeInKByte[],
+		unsigned int *CompressedBufferSizeInkByte)
+{
+	unsigned int DETBufferSizePoolInKByte;
+	unsigned int NextDETBufferPieceInKByte;
+	bool DETPieceAssignedToThisSurfaceAlready[DC__NUM_DPP__MAX];
+	bool NextPotentialSurfaceToAssignDETPieceFound;
+	unsigned int NextSurfaceToAssignDETPiece;
+	double TotalBandwidth;
+	double BandwidthOfSurfacesNotAssignedDETPiece;
+	unsigned int max_minDET;
+	unsigned int minDET;
+	unsigned int minDET_pipe;
+	unsigned int j, k;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: ForceSingleDPP = %d\n", __func__, ForceSingleDPP);
+	dml_print("DML::%s: nomDETInKByte = %d\n", __func__, nomDETInKByte);
+	dml_print("DML::%s: NumberOfActiveSurfaces = %d\n", __func__, NumberOfActiveSurfaces);
+	dml_print("DML::%s: UnboundedRequestEnabled = %d\n", __func__, UnboundedRequestEnabled);
+	dml_print("DML::%s: MaxTotalDETInKByte = %d\n", __func__, MaxTotalDETInKByte);
+	dml_print("DML::%s: ConfigReturnBufferSizeInKByte = %d\n", __func__, ConfigReturnBufferSizeInKByte);
+	dml_print("DML::%s: MinCompressedBufferSizeInKByte = %d\n", __func__, MinCompressedBufferSizeInKByte);
+	dml_print("DML::%s: CompressedBufferSegmentSizeInkByteFinal = %d\n", __func__,
+			CompressedBufferSegmentSizeInkByteFinal);
+#endif
+
+	// Note: Will use default det size if that fits 2 swaths
+	if (UnboundedRequestEnabled) {
+		if (DETSizeOverride[0] > 0) {
+			DETBufferSizeInKByte[0] = DETSizeOverride[0];
+		} else {
+			DETBufferSizeInKByte[0] = dml_max(nomDETInKByte, dml_ceil(2.0 *
+					((double) RoundedUpMaxSwathSizeBytesY[0] +
+							(double) RoundedUpMaxSwathSizeBytesC[0]) / 1024.0, 64.0));
+		}
+		*CompressedBufferSizeInkByte = ConfigReturnBufferSizeInKByte - DETBufferSizeInKByte[0];
+	} else {
+		DETBufferSizePoolInKByte = MaxTotalDETInKByte;
+		for (k = 0; k < NumberOfActiveSurfaces; ++k) {
+			DETBufferSizeInKByte[k] = nomDETInKByte;
+			if (SourcePixelFormat[k] == dm_420_8 || SourcePixelFormat[k] == dm_420_10 ||
+					SourcePixelFormat[k] == dm_420_12) {
+				max_minDET = nomDETInKByte - 64;
+			} else {
+				max_minDET = nomDETInKByte;
+			}
+			minDET = 128;
+			minDET_pipe = 0;
+
+			// add DET resource until can hold 2 full swaths
+			while (minDET <= max_minDET && minDET_pipe == 0) {
+				if (2.0 * ((double) RoundedUpMaxSwathSizeBytesY[k] +
+						(double) RoundedUpMaxSwathSizeBytesC[k]) / 1024.0 <= minDET)
+					minDET_pipe = minDET;
+				minDET = minDET + 64;
+			}
+
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: k=%0d minDET        = %d\n", __func__, k, minDET);
+			dml_print("DML::%s: k=%0d max_minDET    = %d\n", __func__, k, max_minDET);
+			dml_print("DML::%s: k=%0d minDET_pipe   = %d\n", __func__, k, minDET_pipe);
+			dml_print("DML::%s: k=%0d RoundedUpMaxSwathSizeBytesY = %d\n", __func__, k,
+					RoundedUpMaxSwathSizeBytesY[k]);
+			dml_print("DML::%s: k=%0d RoundedUpMaxSwathSizeBytesC = %d\n", __func__, k,
+					RoundedUpMaxSwathSizeBytesC[k]);
+#endif
+
+			if (minDET_pipe == 0) {
+				minDET_pipe = dml_max(128, dml_ceil(((double)RoundedUpMaxSwathSizeBytesY[k] +
+						(double)RoundedUpMaxSwathSizeBytesC[k]) / 1024.0, 64));
+#ifdef __DML_VBA_DEBUG__
+				dml_print("DML::%s: k=%0d minDET_pipe = %d (assume each plane take half DET)\n",
+						__func__, k, minDET_pipe);
+#endif
+			}
+
+			if (UseMALLForPStateChange[k] == dm_use_mall_pstate_change_phantom_pipe) {
+				DETBufferSizeInKByte[k] = 0;
+			} else if (DETSizeOverride[k] > 0) {
+				DETBufferSizeInKByte[k] = DETSizeOverride[k];
+				DETBufferSizePoolInKByte = DETBufferSizePoolInKByte -
+						(ForceSingleDPP ? 1 : DPPPerSurface[k]) * DETSizeOverride[k];
+			} else if ((ForceSingleDPP ? 1 : DPPPerSurface[k]) * minDET_pipe <= DETBufferSizePoolInKByte) {
+				DETBufferSizeInKByte[k] = minDET_pipe;
+				DETBufferSizePoolInKByte = DETBufferSizePoolInKByte -
+						(ForceSingleDPP ? 1 : DPPPerSurface[k]) * minDET_pipe;
+			}
+
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: k=%d DPPPerSurface = %d\n", __func__, k, DPPPerSurface[k]);
+			dml_print("DML::%s: k=%d DETSizeOverride = %d\n", __func__, k, DETSizeOverride[k]);
+			dml_print("DML::%s: k=%d DETBufferSizeInKByte = %d\n", __func__, k, DETBufferSizeInKByte[k]);
+			dml_print("DML::%s: DETBufferSizePoolInKByte = %d\n", __func__, DETBufferSizePoolInKByte);
+#endif
+		}
+
+		TotalBandwidth = 0;
+		for (k = 0; k < NumberOfActiveSurfaces; ++k) {
+			if (UseMALLForPStateChange[k] != dm_use_mall_pstate_change_phantom_pipe)
+				TotalBandwidth = TotalBandwidth + ReadBandwidthLuma[k] + ReadBandwidthChroma[k];
+		}
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: --- Before bandwidth adjustment ---\n", __func__);
+		for (uint k = 0; k < NumberOfActiveSurfaces; ++k)
+			dml_print("DML::%s: k=%d DETBufferSizeInKByte   = %d\n", __func__, k, DETBufferSizeInKByte[k]);
+		dml_print("DML::%s: --- DET allocation with bandwidth ---\n", __func__);
+		dml_print("DML::%s: TotalBandwidth = %f\n", __func__, TotalBandwidth);
+#endif
+		BandwidthOfSurfacesNotAssignedDETPiece = TotalBandwidth;
+		for (k = 0; k < NumberOfActiveSurfaces; ++k) {
+
+			if (UseMALLForPStateChange[k] == dm_use_mall_pstate_change_phantom_pipe) {
+				DETPieceAssignedToThisSurfaceAlready[k] = true;
+			} else if (DETSizeOverride[k] > 0 || (((double) (ForceSingleDPP ? 1 : DPPPerSurface[k]) *
+					(double) DETBufferSizeInKByte[k] / (double) MaxTotalDETInKByte) >=
+					((ReadBandwidthLuma[k] + ReadBandwidthChroma[k]) / TotalBandwidth))) {
+				DETPieceAssignedToThisSurfaceAlready[k] = true;
+				BandwidthOfSurfacesNotAssignedDETPiece = BandwidthOfSurfacesNotAssignedDETPiece -
+						ReadBandwidthLuma[k] - ReadBandwidthChroma[k];
+			} else {
+				DETPieceAssignedToThisSurfaceAlready[k] = false;
+			}
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: k=%d DETPieceAssignedToThisSurfaceAlready = %d\n", __func__, k,
+					DETPieceAssignedToThisSurfaceAlready[k]);
+			dml_print("DML::%s: k=%d BandwidthOfSurfacesNotAssignedDETPiece = %f\n", __func__, k,
+					BandwidthOfSurfacesNotAssignedDETPiece);
+#endif
+		}
+
+		for (j = 0; j < NumberOfActiveSurfaces; ++j) {
+			NextPotentialSurfaceToAssignDETPieceFound = false;
+			NextSurfaceToAssignDETPiece = 0;
+
+			for (k = 0; k < NumberOfActiveSurfaces; ++k) {
+#ifdef __DML_VBA_DEBUG__
+				dml_print("DML::%s: j=%d k=%d, ReadBandwidthLuma[k] = %f\n", __func__, j, k,
+						ReadBandwidthLuma[k]);
+				dml_print("DML::%s: j=%d k=%d, ReadBandwidthChroma[k] = %f\n", __func__, j, k,
+						ReadBandwidthChroma[k]);
+				dml_print("DML::%s: j=%d k=%d, ReadBandwidthLuma[Next] = %f\n", __func__, j, k,
+						ReadBandwidthLuma[NextSurfaceToAssignDETPiece]);
+				dml_print("DML::%s: j=%d k=%d, ReadBandwidthChroma[Next] = %f\n", __func__, j, k,
+						ReadBandwidthChroma[NextSurfaceToAssignDETPiece]);
+				dml_print("DML::%s: j=%d k=%d, NextSurfaceToAssignDETPiece = %d\n", __func__, j, k,
+						NextSurfaceToAssignDETPiece);
+#endif
+				if (!DETPieceAssignedToThisSurfaceAlready[k] &&
+						(!NextPotentialSurfaceToAssignDETPieceFound ||
+						ReadBandwidthLuma[k] + ReadBandwidthChroma[k] <
+						ReadBandwidthLuma[NextSurfaceToAssignDETPiece] +
+						ReadBandwidthChroma[NextSurfaceToAssignDETPiece])) {
+					NextSurfaceToAssignDETPiece = k;
+					NextPotentialSurfaceToAssignDETPieceFound = true;
+				}
+#ifdef __DML_VBA_DEBUG__
+				dml_print("DML::%s: j=%d k=%d, DETPieceAssignedToThisSurfaceAlready = %d\n",
+						__func__, j, k, DETPieceAssignedToThisSurfaceAlready[k]);
+				dml_print("DML::%s: j=%d k=%d, NextPotentialSurfaceToAssignDETPieceFound = %d\n",
+						__func__, j, k, NextPotentialSurfaceToAssignDETPieceFound);
+#endif
+			}
+
+			if (NextPotentialSurfaceToAssignDETPieceFound) {
+				// Note: To show the banker's rounding behavior in VBA and also the fact
+				// that the DET buffer size varies due to precision issue
+				//
+				//double tmp1 =  ((double) DETBufferSizePoolInKByte *
+				// (ReadBandwidthLuma[NextSurfaceToAssignDETPiece] +
+				// ReadBandwidthChroma[NextSurfaceToAssignDETPiece]) /
+				// BandwidthOfSurfacesNotAssignedDETPiece /
+				// ((ForceSingleDPP ? 1 : DPPPerSurface[NextSurfaceToAssignDETPiece]) * 64.0));
+				//double tmp2 =  dml_round((double) DETBufferSizePoolInKByte *
+				// (ReadBandwidthLuma[NextSurfaceToAssignDETPiece] +
+				// ReadBandwidthChroma[NextSurfaceToAssignDETPiece]) /
+				 //BandwidthOfSurfacesNotAssignedDETPiece /
+				// ((ForceSingleDPP ? 1 : DPPPerSurface[NextSurfaceToAssignDETPiece]) * 64.0));
+				//
+				//dml_print("DML::%s: j=%d, tmp1 = %f\n", __func__, j, tmp1);
+				//dml_print("DML::%s: j=%d, tmp2 = %f\n", __func__, j, tmp2);
+
+				NextDETBufferPieceInKByte = dml_min(
+					dml_round((double) DETBufferSizePoolInKByte *
+						(ReadBandwidthLuma[NextSurfaceToAssignDETPiece] +
+						ReadBandwidthChroma[NextSurfaceToAssignDETPiece]) /
+						BandwidthOfSurfacesNotAssignedDETPiece /
+						((ForceSingleDPP ? 1 :
+								DPPPerSurface[NextSurfaceToAssignDETPiece]) * 64.0)) *
+						(ForceSingleDPP ? 1 :
+								DPPPerSurface[NextSurfaceToAssignDETPiece]) * 64.0,
+						dml_floor((double) DETBufferSizePoolInKByte,
+						(ForceSingleDPP ? 1 :
+								DPPPerSurface[NextSurfaceToAssignDETPiece]) * 64.0));
+
+				// Above calculation can assign the entire DET buffer allocation to a single pipe.
+				// We should limit the per-pipe DET size to the nominal / max per pipe.
+				if (NextDETBufferPieceInKByte > nomDETInKByte * (ForceSingleDPP ? 1 : DPPPerSurface[k])) {
+					if (DETBufferSizeInKByte[NextSurfaceToAssignDETPiece] <
+							nomDETInKByte * (ForceSingleDPP ? 1 : DPPPerSurface[k])) {
+						NextDETBufferPieceInKByte = nomDETInKByte * (ForceSingleDPP ? 1 : DPPPerSurface[k]) -
+								DETBufferSizeInKByte[NextSurfaceToAssignDETPiece];
+					} else {
+						// Case where DETBufferSizeInKByte[NextSurfaceToAssignDETPiece]
+						// already has the max per-pipe value
+						NextDETBufferPieceInKByte = 0;
+					}
+				}
+
+#ifdef __DML_VBA_DEBUG__
+				dml_print("DML::%s: j=%0d, DETBufferSizePoolInKByte = %d\n", __func__, j,
+					DETBufferSizePoolInKByte);
+				dml_print("DML::%s: j=%0d, NextSurfaceToAssignDETPiece = %d\n", __func__, j,
+					NextSurfaceToAssignDETPiece);
+				dml_print("DML::%s: j=%0d, ReadBandwidthLuma[%0d] = %f\n", __func__, j,
+					NextSurfaceToAssignDETPiece, ReadBandwidthLuma[NextSurfaceToAssignDETPiece]);
+				dml_print("DML::%s: j=%0d, ReadBandwidthChroma[%0d] = %f\n", __func__, j,
+					NextSurfaceToAssignDETPiece, ReadBandwidthChroma[NextSurfaceToAssignDETPiece]);
+				dml_print("DML::%s: j=%0d, BandwidthOfSurfacesNotAssignedDETPiece = %f\n",
+					__func__, j, BandwidthOfSurfacesNotAssignedDETPiece);
+				dml_print("DML::%s: j=%0d, NextDETBufferPieceInKByte = %d\n", __func__, j,
+					NextDETBufferPieceInKByte);
+				dml_print("DML::%s: j=%0d, DETBufferSizeInKByte[%0d] increases from %0d ",
+					__func__, j, NextSurfaceToAssignDETPiece,
+					DETBufferSizeInKByte[NextSurfaceToAssignDETPiece]);
+#endif
+
+				DETBufferSizeInKByte[NextSurfaceToAssignDETPiece] =
+						DETBufferSizeInKByte[NextSurfaceToAssignDETPiece]
+						+ NextDETBufferPieceInKByte
+						/ (ForceSingleDPP ? 1 : DPPPerSurface[NextSurfaceToAssignDETPiece]);
+#ifdef __DML_VBA_DEBUG__
+				dml_print("to %0d\n", DETBufferSizeInKByte[NextSurfaceToAssignDETPiece]);
+#endif
+
+				DETBufferSizePoolInKByte = DETBufferSizePoolInKByte - NextDETBufferPieceInKByte;
+				DETPieceAssignedToThisSurfaceAlready[NextSurfaceToAssignDETPiece] = true;
+				BandwidthOfSurfacesNotAssignedDETPiece = BandwidthOfSurfacesNotAssignedDETPiece -
+						(ReadBandwidthLuma[NextSurfaceToAssignDETPiece] +
+								ReadBandwidthChroma[NextSurfaceToAssignDETPiece]);
+			}
+		}
+		*CompressedBufferSizeInkByte = MinCompressedBufferSizeInKByte;
+	}
+	*CompressedBufferSizeInkByte = *CompressedBufferSizeInkByte * CompressedBufferSegmentSizeInkByteFinal / 64;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: --- After bandwidth adjustment ---\n", __func__);
+	dml_print("DML::%s: CompressedBufferSizeInkByte = %d\n", __func__, *CompressedBufferSizeInkByte);
+	for (uint k = 0; k < NumberOfActiveSurfaces; ++k) {
+		dml_print("DML::%s: k=%d DETBufferSizeInKByte = %d (TotalReadBandWidth=%f)\n",
+				__func__, k, DETBufferSizeInKByte[k], ReadBandwidthLuma[k] + ReadBandwidthChroma[k]);
+	}
+#endif
+} // CalculateDETBufferSize
+
+void dml32_CalculateODMMode(
+		unsigned int MaximumPixelsPerLinePerDSCUnit,
+		unsigned int HActive,
+		enum output_format_class OutFormat,
+		enum output_encoder_class Output,
+		enum odm_combine_policy ODMUse,
+		double StateDispclk,
+		double MaxDispclk,
+		bool DSCEnable,
+		unsigned int TotalNumberOfActiveDPP,
+		unsigned int MaxNumDPP,
+		double PixelClock,
+		double DISPCLKDPPCLKDSCCLKDownSpreading,
+		double DISPCLKRampingMargin,
+		double DISPCLKDPPCLKVCOSpeed,
+		unsigned int NumberOfDSCSlices,
+
+		/* Output */
+		bool *TotalAvailablePipesSupport,
+		unsigned int *NumberOfDPP,
+		enum odm_combine_mode *ODMMode,
+		double *RequiredDISPCLKPerSurface)
+{
+
+	double SurfaceRequiredDISPCLKWithoutODMCombine;
+	double SurfaceRequiredDISPCLKWithODMCombineTwoToOne;
+	double SurfaceRequiredDISPCLKWithODMCombineFourToOne;
+
+	SurfaceRequiredDISPCLKWithoutODMCombine = dml32_CalculateRequiredDispclk(dm_odm_combine_mode_disabled,
+			PixelClock, DISPCLKDPPCLKDSCCLKDownSpreading, DISPCLKRampingMargin, DISPCLKDPPCLKVCOSpeed,
+			MaxDispclk);
+	SurfaceRequiredDISPCLKWithODMCombineTwoToOne = dml32_CalculateRequiredDispclk(dm_odm_combine_mode_2to1,
+			PixelClock, DISPCLKDPPCLKDSCCLKDownSpreading, DISPCLKRampingMargin, DISPCLKDPPCLKVCOSpeed,
+			MaxDispclk);
+	SurfaceRequiredDISPCLKWithODMCombineFourToOne = dml32_CalculateRequiredDispclk(dm_odm_combine_mode_4to1,
+			PixelClock, DISPCLKDPPCLKDSCCLKDownSpreading, DISPCLKRampingMargin, DISPCLKDPPCLKVCOSpeed,
+			MaxDispclk);
+	*TotalAvailablePipesSupport = true;
+	*ODMMode = dm_odm_combine_mode_disabled; // initialize as disable
+
+	if (ODMUse == dm_odm_combine_policy_none)
+		*ODMMode = dm_odm_combine_mode_disabled;
+
+	*RequiredDISPCLKPerSurface = SurfaceRequiredDISPCLKWithoutODMCombine;
+	*NumberOfDPP = 0;
+
+	// FIXME check ODMUse == "" condition does it mean bypass or Gabriel means something like don't care??
+	// (ODMUse == "" || ODMUse == "CombineAsNeeded")
+
+	if (!(Output == dm_hdmi || Output == dm_dp || Output == dm_edp) && (ODMUse == dm_odm_combine_policy_4to1 ||
+			((SurfaceRequiredDISPCLKWithODMCombineTwoToOne > StateDispclk ||
+					(DSCEnable && (HActive > 2 * MaximumPixelsPerLinePerDSCUnit))
+					|| NumberOfDSCSlices > 8)))) {
+		if (TotalNumberOfActiveDPP + 4 <= MaxNumDPP) {
+			*ODMMode = dm_odm_combine_mode_4to1;
+			*RequiredDISPCLKPerSurface = SurfaceRequiredDISPCLKWithODMCombineFourToOne;
+			*NumberOfDPP = 4;
+		} else {
+			*TotalAvailablePipesSupport = false;
+		}
+	} else if (Output != dm_hdmi && (ODMUse == dm_odm_combine_policy_2to1 ||
+			(((SurfaceRequiredDISPCLKWithoutODMCombine > StateDispclk &&
+					SurfaceRequiredDISPCLKWithODMCombineTwoToOne <= StateDispclk) ||
+					(DSCEnable && (HActive > MaximumPixelsPerLinePerDSCUnit))
+					|| (NumberOfDSCSlices <= 8 && NumberOfDSCSlices > 4))))) {
+		if (TotalNumberOfActiveDPP + 2 <= MaxNumDPP) {
+			*ODMMode = dm_odm_combine_mode_2to1;
+			*RequiredDISPCLKPerSurface = SurfaceRequiredDISPCLKWithODMCombineTwoToOne;
+			*NumberOfDPP = 2;
+		} else {
+			*TotalAvailablePipesSupport = false;
+		}
+	} else {
+		if (TotalNumberOfActiveDPP + 1 <= MaxNumDPP)
+			*NumberOfDPP = 1;
+		else
+			*TotalAvailablePipesSupport = false;
+	}
+	if (OutFormat == dm_420 && HActive > DCN32_MAX_FMT_420_BUFFER_WIDTH &&
+			ODMUse != dm_odm_combine_policy_4to1) {
+		if (HActive > DCN32_MAX_FMT_420_BUFFER_WIDTH * 4) {
+			*ODMMode = dm_odm_combine_mode_disabled;
+			*NumberOfDPP = 0;
+			*TotalAvailablePipesSupport = false;
+		} else if (HActive > DCN32_MAX_FMT_420_BUFFER_WIDTH * 2 ||
+				*ODMMode == dm_odm_combine_mode_4to1) {
+			*ODMMode = dm_odm_combine_mode_4to1;
+			*RequiredDISPCLKPerSurface = SurfaceRequiredDISPCLKWithODMCombineFourToOne;
+			*NumberOfDPP = 4;
+		} else {
+			*ODMMode = dm_odm_combine_mode_2to1;
+			*RequiredDISPCLKPerSurface = SurfaceRequiredDISPCLKWithODMCombineTwoToOne;
+			*NumberOfDPP = 2;
+		}
+	}
+	if (Output == dm_hdmi && OutFormat == dm_420 &&
+			HActive > DCN32_MAX_FMT_420_BUFFER_WIDTH) {
+		*ODMMode = dm_odm_combine_mode_disabled;
+		*NumberOfDPP = 0;
+		*TotalAvailablePipesSupport = false;
+	}
+}
+
+double dml32_CalculateRequiredDispclk(
+		enum odm_combine_mode ODMMode,
+		double PixelClock,
+		double DISPCLKDPPCLKDSCCLKDownSpreading,
+		double DISPCLKRampingMargin,
+		double DISPCLKDPPCLKVCOSpeed,
+		double MaxDispclk)
+{
+	double RequiredDispclk = 0.;
+	double PixelClockAfterODM;
+	double DISPCLKWithRampingRoundedToDFSGranularity;
+	double DISPCLKWithoutRampingRoundedToDFSGranularity;
+	double MaxDispclkRoundedDownToDFSGranularity;
+
+	if (ODMMode == dm_odm_combine_mode_4to1)
+		PixelClockAfterODM = PixelClock / 4;
+	else if (ODMMode == dm_odm_combine_mode_2to1)
+		PixelClockAfterODM = PixelClock / 2;
+	else
+		PixelClockAfterODM = PixelClock;
+
+
+	DISPCLKWithRampingRoundedToDFSGranularity = dml32_RoundToDFSGranularity(
+			PixelClockAfterODM * (1 + DISPCLKDPPCLKDSCCLKDownSpreading / 100)
+					* (1 + DISPCLKRampingMargin / 100), 1, DISPCLKDPPCLKVCOSpeed);
+
+	DISPCLKWithoutRampingRoundedToDFSGranularity = dml32_RoundToDFSGranularity(
+			PixelClockAfterODM * (1 + DISPCLKDPPCLKDSCCLKDownSpreading / 100), 1, DISPCLKDPPCLKVCOSpeed);
+
+	MaxDispclkRoundedDownToDFSGranularity = dml32_RoundToDFSGranularity(MaxDispclk, 0, DISPCLKDPPCLKVCOSpeed);
+
+	if (DISPCLKWithoutRampingRoundedToDFSGranularity > MaxDispclkRoundedDownToDFSGranularity)
+		RequiredDispclk = DISPCLKWithoutRampingRoundedToDFSGranularity;
+	else if (DISPCLKWithRampingRoundedToDFSGranularity > MaxDispclkRoundedDownToDFSGranularity)
+		RequiredDispclk = MaxDispclkRoundedDownToDFSGranularity;
+	else
+		RequiredDispclk = DISPCLKWithRampingRoundedToDFSGranularity;
+
+	return RequiredDispclk;
+}
+
+double dml32_RoundToDFSGranularity(double Clock, bool round_up, double VCOSpeed)
+{
+	if (Clock <= 0.0)
+		return 0.0;
+
+	if (round_up)
+		return VCOSpeed * 4.0 / dml_floor(VCOSpeed * 4.0 / Clock, 1.0);
+	else
+		return VCOSpeed * 4.0 / dml_ceil(VCOSpeed * 4.0 / Clock, 1.0);
+}
+
+void dml32_CalculateOutputLink(
+		double PHYCLKPerState,
+		double PHYCLKD18PerState,
+		double PHYCLKD32PerState,
+		double Downspreading,
+		bool IsMainSurfaceUsingTheIndicatedTiming,
+		enum output_encoder_class Output,
+		enum output_format_class OutputFormat,
+		unsigned int HTotal,
+		unsigned int HActive,
+		double PixelClockBackEnd,
+		double ForcedOutputLinkBPP,
+		unsigned int DSCInputBitPerComponent,
+		unsigned int NumberOfDSCSlices,
+		double AudioSampleRate,
+		unsigned int AudioSampleLayout,
+		enum odm_combine_mode ODMModeNoDSC,
+		enum odm_combine_mode ODMModeDSC,
+		bool DSCEnable,
+		unsigned int OutputLinkDPLanes,
+		enum dm_output_link_dp_rate OutputLinkDPRate,
+
+		/* Output */
+		bool *RequiresDSC,
+		double *RequiresFEC,
+		double  *OutBpp,
+		enum dm_output_type *OutputType,
+		enum dm_output_rate *OutputRate,
+		unsigned int *RequiredSlots)
+{
+	bool LinkDSCEnable;
+	unsigned int dummy;
+	*RequiresDSC = false;
+	*RequiresFEC = false;
+	*OutBpp = 0;
+	*OutputType = dm_output_type_unknown;
+	*OutputRate = dm_output_rate_unknown;
+
+	if (IsMainSurfaceUsingTheIndicatedTiming) {
+		if (Output == dm_hdmi) {
+			*RequiresDSC = false;
+			*RequiresFEC = false;
+			*OutBpp = dml32_TruncToValidBPP(dml_min(600, PHYCLKPerState) * 10, 3, HTotal, HActive,
+					PixelClockBackEnd, ForcedOutputLinkBPP, false, Output, OutputFormat,
+					DSCInputBitPerComponent, NumberOfDSCSlices, AudioSampleRate, AudioSampleLayout,
+					ODMModeNoDSC, ODMModeDSC, &dummy);
+			//OutputTypeAndRate = "HDMI";
+			*OutputType = dm_output_type_hdmi;
+
+		} else if (Output == dm_dp || Output == dm_dp2p0 || Output == dm_edp) {
+			if (DSCEnable == true) {
+				*RequiresDSC = true;
+				LinkDSCEnable = true;
+				if (Output == dm_dp || Output == dm_dp2p0)
+					*RequiresFEC = true;
+				else
+					*RequiresFEC = false;
+			} else {
+				*RequiresDSC = false;
+				LinkDSCEnable = false;
+				if (Output == dm_dp2p0)
+					*RequiresFEC = true;
+				else
+					*RequiresFEC = false;
+			}
+			if (Output == dm_dp2p0) {
+				*OutBpp = 0;
+				if ((OutputLinkDPRate == dm_dp_rate_na || OutputLinkDPRate == dm_dp_rate_uhbr10) &&
+						PHYCLKD32PerState >= 10000.0 / 32) {
+					*OutBpp = dml32_TruncToValidBPP((1 - Downspreading / 100) * 10000,
+							OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd,
+							ForcedOutputLinkBPP, LinkDSCEnable, Output, OutputFormat,
+							DSCInputBitPerComponent, NumberOfDSCSlices, AudioSampleRate,
+							AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+					if (*OutBpp == 0 && PHYCLKD32PerState < 13500.0 / 32 && DSCEnable == true &&
+							ForcedOutputLinkBPP == 0) {
+						*RequiresDSC = true;
+						LinkDSCEnable = true;
+						*OutBpp = dml32_TruncToValidBPP((1 - Downspreading / 100) * 10000,
+								OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd,
+								ForcedOutputLinkBPP, LinkDSCEnable, Output,
+								OutputFormat, DSCInputBitPerComponent,
+								NumberOfDSCSlices, AudioSampleRate, AudioSampleLayout,
+								ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+					}
+					//OutputTypeAndRate = Output & " UHBR10";
+					*OutputType = dm_output_type_dp2p0;
+					*OutputRate = dm_output_rate_dp_rate_uhbr10;
+				}
+				if ((OutputLinkDPRate == dm_dp_rate_na || OutputLinkDPRate == dm_dp_rate_uhbr13p5) &&
+						*OutBpp == 0 && PHYCLKD32PerState >= 13500.0 / 32) {
+					*OutBpp = dml32_TruncToValidBPP((1 - Downspreading / 100) * 13500,
+							OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd,
+							ForcedOutputLinkBPP, LinkDSCEnable, Output, OutputFormat,
+							DSCInputBitPerComponent, NumberOfDSCSlices, AudioSampleRate,
+							AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+
+					if (*OutBpp == 0 && PHYCLKD32PerState < 20000 / 32 && DSCEnable == true &&
+							ForcedOutputLinkBPP == 0) {
+						*RequiresDSC = true;
+						LinkDSCEnable = true;
+						*OutBpp = dml32_TruncToValidBPP((1 - Downspreading / 100) * 13500,
+								OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd,
+								ForcedOutputLinkBPP, LinkDSCEnable, Output,
+								OutputFormat, DSCInputBitPerComponent,
+								NumberOfDSCSlices, AudioSampleRate, AudioSampleLayout,
+								ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+					}
+					//OutputTypeAndRate = Output & " UHBR13p5";
+					*OutputType = dm_output_type_dp2p0;
+					*OutputRate = dm_output_rate_dp_rate_uhbr13p5;
+				}
+				if ((OutputLinkDPRate == dm_dp_rate_na || OutputLinkDPRate == dm_dp_rate_uhbr20) &&
+						*OutBpp == 0 && PHYCLKD32PerState >= 20000 / 32) {
+					*OutBpp = dml32_TruncToValidBPP((1 - Downspreading / 100) * 20000,
+							OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd,
+							ForcedOutputLinkBPP, LinkDSCEnable, Output, OutputFormat,
+							DSCInputBitPerComponent, NumberOfDSCSlices, AudioSampleRate,
+							AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+					if (*OutBpp == 0 && DSCEnable == true && ForcedOutputLinkBPP == 0) {
+						*RequiresDSC = true;
+						LinkDSCEnable = true;
+						*OutBpp = dml32_TruncToValidBPP((1 - Downspreading / 100) * 20000,
+								OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd,
+								ForcedOutputLinkBPP, LinkDSCEnable, Output,
+								OutputFormat, DSCInputBitPerComponent,
+								NumberOfDSCSlices, AudioSampleRate, AudioSampleLayout,
+								ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+					}
+					//OutputTypeAndRate = Output & " UHBR20";
+					*OutputType = dm_output_type_dp2p0;
+					*OutputRate = dm_output_rate_dp_rate_uhbr20;
+				}
+			} else {
+				*OutBpp = 0;
+				if ((OutputLinkDPRate == dm_dp_rate_na || OutputLinkDPRate == dm_dp_rate_hbr) &&
+						PHYCLKPerState >= 270) {
+					*OutBpp = dml32_TruncToValidBPP((1 - Downspreading / 100) * 2700,
+							OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd,
+							ForcedOutputLinkBPP, LinkDSCEnable, Output, OutputFormat,
+							DSCInputBitPerComponent, NumberOfDSCSlices, AudioSampleRate,
+							AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+					if (*OutBpp == 0 && PHYCLKPerState < 540 && DSCEnable == true &&
+							ForcedOutputLinkBPP == 0) {
+						*RequiresDSC = true;
+						LinkDSCEnable = true;
+						if (Output == dm_dp)
+							*RequiresFEC = true;
+						*OutBpp = dml32_TruncToValidBPP((1 - Downspreading / 100) * 2700,
+								OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd,
+								ForcedOutputLinkBPP, LinkDSCEnable, Output,
+								OutputFormat, DSCInputBitPerComponent,
+								NumberOfDSCSlices, AudioSampleRate, AudioSampleLayout,
+								ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+					}
+					//OutputTypeAndRate = Output & " HBR";
+					*OutputType = (Output == dm_dp) ? dm_output_type_dp : dm_output_type_edp;
+					*OutputRate = dm_output_rate_dp_rate_hbr;
+				}
+				if ((OutputLinkDPRate == dm_dp_rate_na || OutputLinkDPRate == dm_dp_rate_hbr2) &&
+						*OutBpp == 0 && PHYCLKPerState >= 540) {
+					*OutBpp = dml32_TruncToValidBPP((1 - Downspreading / 100) * 5400,
+							OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd,
+							ForcedOutputLinkBPP, LinkDSCEnable, Output, OutputFormat,
+							DSCInputBitPerComponent, NumberOfDSCSlices, AudioSampleRate,
+							AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+
+					if (*OutBpp == 0 && PHYCLKPerState < 810 && DSCEnable == true &&
+							ForcedOutputLinkBPP == 0) {
+						*RequiresDSC = true;
+						LinkDSCEnable = true;
+						if (Output == dm_dp)
+							*RequiresFEC = true;
+
+						*OutBpp = dml32_TruncToValidBPP((1 - Downspreading / 100) * 5400,
+								OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd,
+								ForcedOutputLinkBPP, LinkDSCEnable, Output,
+								OutputFormat, DSCInputBitPerComponent,
+								NumberOfDSCSlices, AudioSampleRate, AudioSampleLayout,
+								ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+					}
+					//OutputTypeAndRate = Output & " HBR2";
+					*OutputType = (Output == dm_dp) ? dm_output_type_dp : dm_output_type_edp;
+					*OutputRate = dm_output_rate_dp_rate_hbr2;
+				}
+				if ((OutputLinkDPRate == dm_dp_rate_na || OutputLinkDPRate == dm_dp_rate_hbr3) && *OutBpp == 0 && PHYCLKPerState >= 810) {
+					*OutBpp = dml32_TruncToValidBPP((1 - Downspreading / 100) * 8100,
+							OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd,
+							ForcedOutputLinkBPP, LinkDSCEnable, Output,
+							OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices,
+							AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC,
+							RequiredSlots);
+
+					if (*OutBpp == 0 && DSCEnable == true && ForcedOutputLinkBPP == 0) {
+						*RequiresDSC = true;
+						LinkDSCEnable = true;
+						if (Output == dm_dp)
+							*RequiresFEC = true;
+
+						*OutBpp = dml32_TruncToValidBPP((1 - Downspreading / 100) * 8100,
+								OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd,
+								ForcedOutputLinkBPP, LinkDSCEnable, Output,
+								OutputFormat, DSCInputBitPerComponent,
+								NumberOfDSCSlices, AudioSampleRate, AudioSampleLayout,
+								ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+					}
+					//OutputTypeAndRate = Output & " HBR3";
+					*OutputType = (Output == dm_dp) ? dm_output_type_dp : dm_output_type_edp;
+					*OutputRate = dm_output_rate_dp_rate_hbr3;
+				}
+			}
+		}
+	}
+}
+
+void dml32_CalculateDPPCLK(
+		unsigned int NumberOfActiveSurfaces,
+		double DISPCLKDPPCLKDSCCLKDownSpreading,
+		double DISPCLKDPPCLKVCOSpeed,
+		double DPPCLKUsingSingleDPP[],
+		unsigned int DPPPerSurface[],
+
+		/* output */
+		double *GlobalDPPCLK,
+		double Dppclk[])
+{
+	unsigned int k;
+	*GlobalDPPCLK = 0;
+	for (k = 0; k < NumberOfActiveSurfaces; ++k) {
+		Dppclk[k] = DPPCLKUsingSingleDPP[k] / DPPPerSurface[k] * (1 + DISPCLKDPPCLKDSCCLKDownSpreading / 100);
+		*GlobalDPPCLK = dml_max(*GlobalDPPCLK, Dppclk[k]);
+	}
+	*GlobalDPPCLK = dml32_RoundToDFSGranularity(*GlobalDPPCLK, 1, DISPCLKDPPCLKVCOSpeed);
+	for (k = 0; k < NumberOfActiveSurfaces; ++k)
+		Dppclk[k] = *GlobalDPPCLK / 255 * dml_ceil(Dppclk[k] * 255.0 / *GlobalDPPCLK, 1.0);
+}
+
+double dml32_TruncToValidBPP(
+		double LinkBitRate,
+		unsigned int Lanes,
+		unsigned int HTotal,
+		unsigned int HActive,
+		double PixelClock,
+		double DesiredBPP,
+		bool DSCEnable,
+		enum output_encoder_class Output,
+		enum output_format_class Format,
+		unsigned int DSCInputBitPerComponent,
+		unsigned int DSCSlices,
+		unsigned int AudioRate,
+		unsigned int AudioLayout,
+		enum odm_combine_mode ODMModeNoDSC,
+		enum odm_combine_mode ODMModeDSC,
+		/* Output */
+		unsigned int *RequiredSlots)
+{
+	double    MaxLinkBPP;
+	unsigned int   MinDSCBPP;
+	double    MaxDSCBPP;
+	unsigned int   NonDSCBPP0;
+	unsigned int   NonDSCBPP1;
+	unsigned int   NonDSCBPP2;
+	unsigned int   NonDSCBPP3 = BPP_INVALID;
+
+	if (Format == dm_420) {
+		NonDSCBPP0 = 12;
+		NonDSCBPP1 = 15;
+		NonDSCBPP2 = 18;
+		MinDSCBPP = 6;
+		MaxDSCBPP = 1.5 * DSCInputBitPerComponent - 1.0 / 16;
+	} else if (Format == dm_444) {
+		NonDSCBPP3 = 18;
+		NonDSCBPP0 = 24;
+		NonDSCBPP1 = 30;
+		NonDSCBPP2 = 36;
+		MinDSCBPP = 8;
+		MaxDSCBPP = 3 * DSCInputBitPerComponent - 1.0 / 16;
+	} else {
+		if (Output == dm_hdmi) {
+			NonDSCBPP0 = 24;
+			NonDSCBPP1 = 24;
+			NonDSCBPP2 = 24;
+		} else {
+			NonDSCBPP0 = 16;
+			NonDSCBPP1 = 20;
+			NonDSCBPP2 = 24;
+		}
+		if (Format == dm_n422) {
+			MinDSCBPP = 7;
+			MaxDSCBPP = 2 * DSCInputBitPerComponent - 1.0 / 16.0;
+		} else {
+			MinDSCBPP = 8;
+			MaxDSCBPP = 3 * DSCInputBitPerComponent - 1.0 / 16.0;
+		}
+	}
+	if (Output == dm_dp2p0) {
+		MaxLinkBPP = LinkBitRate * Lanes / PixelClock * 128 / 132 * 383 / 384 * 65536 / 65540;
+	} else if (DSCEnable && Output == dm_dp) {
+		MaxLinkBPP = LinkBitRate / 10 * 8 * Lanes / PixelClock * (1 - 2.4 / 100);
+	} else {
+		MaxLinkBPP = LinkBitRate / 10 * 8 * Lanes / PixelClock;
+	}
+
+	if (DSCEnable) {
+		if (ODMModeDSC == dm_odm_combine_mode_4to1)
+			MaxLinkBPP = dml_min(MaxLinkBPP, 16);
+		else if (ODMModeDSC == dm_odm_combine_mode_2to1)
+			MaxLinkBPP = dml_min(MaxLinkBPP, 32);
+		else if (ODMModeDSC == dm_odm_split_mode_1to2)
+			MaxLinkBPP = 2 * MaxLinkBPP;
+	} else {
+		if (ODMModeNoDSC == dm_odm_combine_mode_4to1)
+			MaxLinkBPP = dml_min(MaxLinkBPP, 16);
+		else if (ODMModeNoDSC == dm_odm_combine_mode_2to1)
+			MaxLinkBPP = dml_min(MaxLinkBPP, 32);
+		else if (ODMModeNoDSC == dm_odm_split_mode_1to2)
+			MaxLinkBPP = 2 * MaxLinkBPP;
+	}
+
+	*RequiredSlots = dml_ceil(DesiredBPP / MaxLinkBPP * 64, 1);
+
+	if (DesiredBPP == 0) {
+		if (DSCEnable) {
+			if (MaxLinkBPP < MinDSCBPP)
+				return BPP_INVALID;
+			else if (MaxLinkBPP >= MaxDSCBPP)
+				return MaxDSCBPP;
+			else
+				return dml_floor(16.0 * MaxLinkBPP, 1.0) / 16.0;
+		} else {
+			if (MaxLinkBPP >= NonDSCBPP2)
+				return NonDSCBPP2;
+			else if (MaxLinkBPP >= NonDSCBPP1)
+				return NonDSCBPP1;
+			else if (MaxLinkBPP >= NonDSCBPP0)
+				return 16.0;
+			else if ((Output == dm_dp2p0 || Output == dm_dp) && NonDSCBPP3 != BPP_INVALID &&  MaxLinkBPP >= NonDSCBPP3)
+				return NonDSCBPP3; // Special case to allow 6bpc RGB for DP connections.
+			else
+				return BPP_INVALID;
+		}
+	} else {
+		if (!((DSCEnable == false && (DesiredBPP == NonDSCBPP2 || DesiredBPP == NonDSCBPP1 ||
+				DesiredBPP <= NonDSCBPP0)) ||
+				(DSCEnable && DesiredBPP >= MinDSCBPP && DesiredBPP <= MaxDSCBPP)))
+			return BPP_INVALID;
+		else
+			return DesiredBPP;
+	}
+} // TruncToValidBPP
+
+double dml32_RequiredDTBCLK(
+		bool              DSCEnable,
+		double               PixelClock,
+		enum output_format_class  OutputFormat,
+		double               OutputBpp,
+		unsigned int              DSCSlices,
+		unsigned int                 HTotal,
+		unsigned int                 HActive,
+		unsigned int              AudioRate,
+		unsigned int              AudioLayout)
+{
+	double PixelWordRate;
+	double HCActive;
+	double HCBlank;
+	double AverageTribyteRate;
+	double HActiveTribyteRate;
+
+	if (DSCEnable != true)
+		return dml_max(PixelClock / 4.0 * OutputBpp / 24.0, 25.0);
+
+	PixelWordRate = PixelClock /  (OutputFormat == dm_444 ? 1 : 2);
+	HCActive = dml_ceil(DSCSlices * dml_ceil(OutputBpp *
+			dml_ceil(HActive / DSCSlices, 1) / 8.0, 1) / 3.0, 1);
+	HCBlank = 64 + 32 *
+			dml_ceil(AudioRate *  (AudioLayout == 1 ? 1 : 0.25) * HTotal / (PixelClock * 1000), 1);
+	AverageTribyteRate = PixelWordRate * (HCActive + HCBlank) / HTotal;
+	HActiveTribyteRate = PixelWordRate * HCActive / HActive;
+	return dml_max4(PixelWordRate / 4.0, AverageTribyteRate / 4.0, HActiveTribyteRate / 4.0, 25.0) * 1.002;
+}
+
+unsigned int dml32_DSCDelayRequirement(bool DSCEnabled,
+		enum odm_combine_mode ODMMode,
+		unsigned int DSCInputBitPerComponent,
+		double OutputBpp,
+		unsigned int HActive,
+		unsigned int HTotal,
+		unsigned int NumberOfDSCSlices,
+		enum output_format_class  OutputFormat,
+		enum output_encoder_class Output,
+		double PixelClock,
+		double PixelClockBackEnd,
+		double dsc_delay_factor_wa)
+{
+	unsigned int DSCDelayRequirement_val;
+
+	if (DSCEnabled == true && OutputBpp != 0) {
+		if (ODMMode == dm_odm_combine_mode_4to1) {
+			DSCDelayRequirement_val = 4 * (dml32_dscceComputeDelay(DSCInputBitPerComponent, OutputBpp,
+					dml_ceil(HActive / NumberOfDSCSlices, 1), NumberOfDSCSlices / 4,
+					OutputFormat, Output) + dml32_dscComputeDelay(OutputFormat, Output));
+		} else if (ODMMode == dm_odm_combine_mode_2to1) {
+			DSCDelayRequirement_val = 2 * (dml32_dscceComputeDelay(DSCInputBitPerComponent, OutputBpp,
+					dml_ceil(HActive / NumberOfDSCSlices, 1), NumberOfDSCSlices / 2,
+					OutputFormat, Output) + dml32_dscComputeDelay(OutputFormat, Output));
+		} else {
+			DSCDelayRequirement_val = dml32_dscceComputeDelay(DSCInputBitPerComponent, OutputBpp,
+					dml_ceil(HActive / NumberOfDSCSlices, 1), NumberOfDSCSlices,
+					OutputFormat, Output) + dml32_dscComputeDelay(OutputFormat, Output);
+		}
+
+		DSCDelayRequirement_val = DSCDelayRequirement_val + (HTotal - HActive) *
+				dml_ceil((double)DSCDelayRequirement_val / HActive, 1);
+
+		DSCDelayRequirement_val = DSCDelayRequirement_val * PixelClock / PixelClockBackEnd;
+
+	} else {
+		DSCDelayRequirement_val = 0;
+	}
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: DSCEnabled              = %d\n", __func__, DSCEnabled);
+	dml_print("DML::%s: OutputBpp               = %f\n", __func__, OutputBpp);
+	dml_print("DML::%s: HActive                 = %d\n", __func__, HActive);
+	dml_print("DML::%s: OutputFormat            = %d\n", __func__, OutputFormat);
+	dml_print("DML::%s: DSCInputBitPerComponent = %d\n", __func__, DSCInputBitPerComponent);
+	dml_print("DML::%s: NumberOfDSCSlices       = %d\n", __func__, NumberOfDSCSlices);
+	dml_print("DML::%s: DSCDelayRequirement_val = %d\n", __func__, DSCDelayRequirement_val);
+#endif
+
+	return dml_ceil(DSCDelayRequirement_val * dsc_delay_factor_wa, 1);
+}
+
+void dml32_CalculateSurfaceSizeInMall(
+		unsigned int NumberOfActiveSurfaces,
+		unsigned int MALLAllocatedForDCN,
+		enum dm_use_mall_for_static_screen_mode UseMALLForStaticScreen[],
+		enum dm_use_mall_for_pstate_change_mode UsesMALLForPStateChange[],
+		bool DCCEnable[],
+		bool ViewportStationary[],
+		unsigned int ViewportXStartY[],
+		unsigned int ViewportYStartY[],
+		unsigned int ViewportXStartC[],
+		unsigned int ViewportYStartC[],
+		unsigned int ViewportWidthY[],
+		unsigned int ViewportHeightY[],
+		unsigned int BytesPerPixelY[],
+		unsigned int ViewportWidthC[],
+		unsigned int ViewportHeightC[],
+		unsigned int BytesPerPixelC[],
+		unsigned int SurfaceWidthY[],
+		unsigned int SurfaceWidthC[],
+		unsigned int SurfaceHeightY[],
+		unsigned int SurfaceHeightC[],
+		unsigned int Read256BytesBlockWidthY[],
+		unsigned int Read256BytesBlockWidthC[],
+		unsigned int Read256BytesBlockHeightY[],
+		unsigned int Read256BytesBlockHeightC[],
+		unsigned int ReadBlockWidthY[],
+		unsigned int ReadBlockWidthC[],
+		unsigned int ReadBlockHeightY[],
+		unsigned int ReadBlockHeightC[],
+		unsigned int DCCMetaPitchY[],
+		unsigned int DCCMetaPitchC[],
+
+		/* Output */
+		unsigned int    SurfaceSizeInMALL[],
+		bool *ExceededMALLSize)
+{
+	unsigned int k;
+	unsigned int TotalSurfaceSizeInMALLForSS = 0;
+	unsigned int TotalSurfaceSizeInMALLForSubVP = 0;
+	unsigned int MALLAllocatedForDCNInBytes = MALLAllocatedForDCN * 1024 * 1024;
+
+	for (k = 0; k < NumberOfActiveSurfaces; ++k) {
+		if (ViewportStationary[k]) {
+			SurfaceSizeInMALL[k] = dml_min(dml_ceil(SurfaceWidthY[k], ReadBlockWidthY[k]),
+					dml_floor(ViewportXStartY[k] + ViewportWidthY[k] + ReadBlockWidthY[k] - 1,
+						ReadBlockWidthY[k]) - dml_floor(ViewportXStartY[k],
+						ReadBlockWidthY[k])) * dml_min(dml_ceil(SurfaceHeightY[k],
+						ReadBlockHeightY[k]), dml_floor(ViewportYStartY[k] +
+						ViewportHeightY[k] + ReadBlockHeightY[k] - 1, ReadBlockHeightY[k]) -
+						dml_floor(ViewportYStartY[k], ReadBlockHeightY[k])) * BytesPerPixelY[k];
+
+			if (ReadBlockWidthC[k] > 0) {
+				SurfaceSizeInMALL[k] = SurfaceSizeInMALL[k] +
+						dml_min(dml_ceil(SurfaceWidthC[k], ReadBlockWidthC[k]),
+							dml_floor(ViewportXStartC[k] + ViewportWidthC[k] +
+							ReadBlockWidthC[k] - 1, ReadBlockWidthC[k]) -
+							dml_floor(ViewportXStartC[k], ReadBlockWidthC[k])) *
+							dml_min(dml_ceil(SurfaceHeightC[k], ReadBlockHeightC[k]),
+							dml_floor(ViewportYStartC[k] + ViewportHeightC[k] +
+							ReadBlockHeightC[k] - 1, ReadBlockHeightC[k]) -
+							dml_floor(ViewportYStartC[k], ReadBlockHeightC[k])) *
+							BytesPerPixelC[k];
+			}
+			if (DCCEnable[k] == true) {
+				SurfaceSizeInMALL[k] = SurfaceSizeInMALL[k] +
+						(dml_min(dml_ceil(DCCMetaPitchY[k], 8 * Read256BytesBlockWidthY[k]),
+							dml_floor(ViewportXStartY[k] + ViewportWidthY[k] + 8 *
+							Read256BytesBlockWidthY[k] - 1, 8 * Read256BytesBlockWidthY[k])
+							- dml_floor(ViewportXStartY[k], 8 * Read256BytesBlockWidthY[k]))
+							* dml_min(dml_ceil(SurfaceHeightY[k], 8 *
+							Read256BytesBlockHeightY[k]), dml_floor(ViewportYStartY[k] +
+							ViewportHeightY[k] + 8 * Read256BytesBlockHeightY[k] - 1, 8 *
+							Read256BytesBlockHeightY[k]) - dml_floor(ViewportYStartY[k], 8 *
+							Read256BytesBlockHeightY[k])) * BytesPerPixelY[k] / 256) + (64 * 1024);
+				if (Read256BytesBlockWidthC[k] > 0) {
+					SurfaceSizeInMALL[k] = SurfaceSizeInMALL[k] +
+							dml_min(dml_ceil(DCCMetaPitchC[k], 8 *
+								Read256BytesBlockWidthC[k]),
+								dml_floor(ViewportXStartC[k] + ViewportWidthC[k] + 8
+								* Read256BytesBlockWidthC[k] - 1, 8 *
+								Read256BytesBlockWidthC[k]) -
+								dml_floor(ViewportXStartC[k], 8 *
+								Read256BytesBlockWidthC[k])) *
+								dml_min(dml_ceil(SurfaceHeightC[k], 8 *
+								Read256BytesBlockHeightC[k]),
+								dml_floor(ViewportYStartC[k] + ViewportHeightC[k] +
+								8 * Read256BytesBlockHeightC[k] - 1, 8 *
+								Read256BytesBlockHeightC[k]) -
+								dml_floor(ViewportYStartC[k], 8 *
+								Read256BytesBlockHeightC[k])) *
+								BytesPerPixelC[k] / 256;
+				}
+			}
+		} else {
+			SurfaceSizeInMALL[k] = dml_ceil(dml_min(SurfaceWidthY[k], ViewportWidthY[k] +
+					ReadBlockWidthY[k] - 1), ReadBlockWidthY[k]) *
+					dml_ceil(dml_min(SurfaceHeightY[k], ViewportHeightY[k] +
+							ReadBlockHeightY[k] - 1), ReadBlockHeightY[k]) *
+							BytesPerPixelY[k];
+			if (ReadBlockWidthC[k] > 0) {
+				SurfaceSizeInMALL[k] = SurfaceSizeInMALL[k] +
+						dml_ceil(dml_min(SurfaceWidthC[k], ViewportWidthC[k] +
+								ReadBlockWidthC[k] - 1), ReadBlockWidthC[k]) *
+						dml_ceil(dml_min(SurfaceHeightC[k], ViewportHeightC[k] +
+								ReadBlockHeightC[k] - 1), ReadBlockHeightC[k]) *
+								BytesPerPixelC[k];
+			}
+			if (DCCEnable[k] == true) {
+				SurfaceSizeInMALL[k] = SurfaceSizeInMALL[k] +
+						(dml_ceil(dml_min(DCCMetaPitchY[k], ViewportWidthY[k] + 8 *
+								Read256BytesBlockWidthY[k] - 1), 8 *
+								Read256BytesBlockWidthY[k]) *
+						dml_ceil(dml_min(SurfaceHeightY[k], ViewportHeightY[k] + 8 *
+								Read256BytesBlockHeightY[k] - 1), 8 *
+								Read256BytesBlockHeightY[k]) * BytesPerPixelY[k] / 256) + (64 * 1024);
+
+				if (Read256BytesBlockWidthC[k] > 0) {
+					SurfaceSizeInMALL[k] = SurfaceSizeInMALL[k] +
+							dml_ceil(dml_min(DCCMetaPitchC[k], ViewportWidthC[k] + 8 *
+									Read256BytesBlockWidthC[k] - 1), 8 *
+									Read256BytesBlockWidthC[k]) *
+							dml_ceil(dml_min(SurfaceHeightC[k], ViewportHeightC[k] + 8 *
+									Read256BytesBlockHeightC[k] - 1), 8 *
+									Read256BytesBlockHeightC[k]) *
+									BytesPerPixelC[k] / 256;
+				}
+			}
+		}
+	}
+
+	for (k = 0; k < NumberOfActiveSurfaces; ++k) {
+		/* SS and Subvp counted separate as they are never used at the same time */
+		if (UsesMALLForPStateChange[k] == dm_use_mall_pstate_change_phantom_pipe)
+			TotalSurfaceSizeInMALLForSubVP = TotalSurfaceSizeInMALLForSubVP + SurfaceSizeInMALL[k];
+		else if (UseMALLForStaticScreen[k] == dm_use_mall_static_screen_enable)
+			TotalSurfaceSizeInMALLForSS = TotalSurfaceSizeInMALLForSS + SurfaceSizeInMALL[k];
+	}
+	*ExceededMALLSize =  (TotalSurfaceSizeInMALLForSS > MALLAllocatedForDCNInBytes) ||
+							(TotalSurfaceSizeInMALLForSubVP > MALLAllocatedForDCNInBytes);
+} // CalculateSurfaceSizeInMall
+
+void dml32_CalculateVMRowAndSwath(
+		unsigned int NumberOfActiveSurfaces,
+		DmlPipe myPipe[],
+		unsigned int SurfaceSizeInMALL[],
+		unsigned int PTEBufferSizeInRequestsLuma,
+		unsigned int PTEBufferSizeInRequestsChroma,
+		unsigned int DCCMetaBufferSizeBytes,
+		enum dm_use_mall_for_static_screen_mode UseMALLForStaticScreen[],
+		enum dm_use_mall_for_pstate_change_mode UseMALLForPStateChange[],
+		unsigned int MALLAllocatedForDCN,
+		double SwathWidthY[],
+		double SwathWidthC[],
+		bool GPUVMEnable,
+		bool HostVMEnable,
+		unsigned int HostVMMaxNonCachedPageTableLevels,
+		unsigned int GPUVMMaxPageTableLevels,
+		unsigned int GPUVMMinPageSizeKBytes[],
+		unsigned int HostVMMinPageSize,
+
+		/* Output */
+		bool PTEBufferSizeNotExceeded[],
+		bool DCCMetaBufferSizeNotExceeded[],
+		unsigned int dpte_row_width_luma_ub[],
+		unsigned int dpte_row_width_chroma_ub[],
+		unsigned int dpte_row_height_luma[],
+		unsigned int dpte_row_height_chroma[],
+		unsigned int dpte_row_height_linear_luma[],     // VBA_DELTA
+		unsigned int dpte_row_height_linear_chroma[],   // VBA_DELTA
+		unsigned int meta_req_width[],
+		unsigned int meta_req_width_chroma[],
+		unsigned int meta_req_height[],
+		unsigned int meta_req_height_chroma[],
+		unsigned int meta_row_width[],
+		unsigned int meta_row_width_chroma[],
+		unsigned int meta_row_height[],
+		unsigned int meta_row_height_chroma[],
+		unsigned int vm_group_bytes[],
+		unsigned int dpte_group_bytes[],
+		unsigned int PixelPTEReqWidthY[],
+		unsigned int PixelPTEReqHeightY[],
+		unsigned int PTERequestSizeY[],
+		unsigned int PixelPTEReqWidthC[],
+		unsigned int PixelPTEReqHeightC[],
+		unsigned int PTERequestSizeC[],
+		unsigned int dpde0_bytes_per_frame_ub_l[],
+		unsigned int meta_pte_bytes_per_frame_ub_l[],
+		unsigned int dpde0_bytes_per_frame_ub_c[],
+		unsigned int meta_pte_bytes_per_frame_ub_c[],
+		double PrefetchSourceLinesY[],
+		double PrefetchSourceLinesC[],
+		double VInitPreFillY[],
+		double VInitPreFillC[],
+		unsigned int MaxNumSwathY[],
+		unsigned int MaxNumSwathC[],
+		double meta_row_bw[],
+		double dpte_row_bw[],
+		double PixelPTEBytesPerRow[],
+		double PDEAndMetaPTEBytesFrame[],
+		double MetaRowByte[],
+		bool use_one_row_for_frame[],
+		bool use_one_row_for_frame_flip[],
+		bool UsesMALLForStaticScreen[],
+		bool PTE_BUFFER_MODE[],
+		unsigned int BIGK_FRAGMENT_SIZE[])
+{
+	unsigned int k;
+	unsigned int PTEBufferSizeInRequestsForLuma[DC__NUM_DPP__MAX];
+	unsigned int PTEBufferSizeInRequestsForChroma[DC__NUM_DPP__MAX];
+	unsigned int PDEAndMetaPTEBytesFrameY;
+	unsigned int PDEAndMetaPTEBytesFrameC;
+	unsigned int MetaRowByteY[DC__NUM_DPP__MAX] = {0};
+	unsigned int MetaRowByteC[DC__NUM_DPP__MAX] = {0};
+	unsigned int PixelPTEBytesPerRowY[DC__NUM_DPP__MAX];
+	unsigned int PixelPTEBytesPerRowC[DC__NUM_DPP__MAX];
+	unsigned int PixelPTEBytesPerRowY_one_row_per_frame[DC__NUM_DPP__MAX];
+	unsigned int PixelPTEBytesPerRowC_one_row_per_frame[DC__NUM_DPP__MAX];
+	unsigned int dpte_row_width_luma_ub_one_row_per_frame[DC__NUM_DPP__MAX];
+	unsigned int dpte_row_height_luma_one_row_per_frame[DC__NUM_DPP__MAX];
+	unsigned int dpte_row_width_chroma_ub_one_row_per_frame[DC__NUM_DPP__MAX];
+	unsigned int dpte_row_height_chroma_one_row_per_frame[DC__NUM_DPP__MAX];
+	bool one_row_per_frame_fits_in_buffer[DC__NUM_DPP__MAX];
+
+	for (k = 0; k < NumberOfActiveSurfaces; ++k) {
+		if (HostVMEnable == true) {
+			vm_group_bytes[k] = 512;
+			dpte_group_bytes[k] = 512;
+		} else if (GPUVMEnable == true) {
+			vm_group_bytes[k] = 2048;
+			if (GPUVMMinPageSizeKBytes[k] >= 64 && IsVertical(myPipe[k].SourceRotation))
+				dpte_group_bytes[k] = 512;
+			else
+				dpte_group_bytes[k] = 2048;
+		} else {
+			vm_group_bytes[k] = 0;
+			dpte_group_bytes[k] = 0;
+		}
+
+		if (myPipe[k].SourcePixelFormat == dm_420_8 || myPipe[k].SourcePixelFormat == dm_420_10 ||
+				myPipe[k].SourcePixelFormat == dm_420_12 ||
+				myPipe[k].SourcePixelFormat == dm_rgbe_alpha) {
+			if ((myPipe[k].SourcePixelFormat == dm_420_10 || myPipe[k].SourcePixelFormat == dm_420_12) &&
+					!IsVertical(myPipe[k].SourceRotation)) {
+				PTEBufferSizeInRequestsForLuma[k] =
+						(PTEBufferSizeInRequestsLuma + PTEBufferSizeInRequestsChroma) / 2;
+				PTEBufferSizeInRequestsForChroma[k] = PTEBufferSizeInRequestsForLuma[k];
+			} else {
+				PTEBufferSizeInRequestsForLuma[k] = PTEBufferSizeInRequestsLuma;
+				PTEBufferSizeInRequestsForChroma[k] = PTEBufferSizeInRequestsChroma;
+			}
+
+			PDEAndMetaPTEBytesFrameC = dml32_CalculateVMAndRowBytes(
+					myPipe[k].ViewportStationary,
+					myPipe[k].DCCEnable,
+					myPipe[k].DPPPerSurface,
+					myPipe[k].BlockHeight256BytesC,
+					myPipe[k].BlockWidth256BytesC,
+					myPipe[k].SourcePixelFormat,
+					myPipe[k].SurfaceTiling,
+					myPipe[k].BytePerPixelC,
+					myPipe[k].SourceRotation,
+					SwathWidthC[k],
+					myPipe[k].ViewportHeightChroma,
+					myPipe[k].ViewportXStartC,
+					myPipe[k].ViewportYStartC,
+					GPUVMEnable,
+					HostVMEnable,
+					HostVMMaxNonCachedPageTableLevels,
+					GPUVMMaxPageTableLevels,
+					GPUVMMinPageSizeKBytes[k],
+					HostVMMinPageSize,
+					PTEBufferSizeInRequestsForChroma[k],
+					myPipe[k].PitchC,
+					myPipe[k].DCCMetaPitchC,
+					myPipe[k].BlockWidthC,
+					myPipe[k].BlockHeightC,
+
+					/* Output */
+					&MetaRowByteC[k],
+					&PixelPTEBytesPerRowC[k],
+					&dpte_row_width_chroma_ub[k],
+					&dpte_row_height_chroma[k],
+					&dpte_row_height_linear_chroma[k],
+					&PixelPTEBytesPerRowC_one_row_per_frame[k],
+					&dpte_row_width_chroma_ub_one_row_per_frame[k],
+					&dpte_row_height_chroma_one_row_per_frame[k],
+					&meta_req_width_chroma[k],
+					&meta_req_height_chroma[k],
+					&meta_row_width_chroma[k],
+					&meta_row_height_chroma[k],
+					&PixelPTEReqWidthC[k],
+					&PixelPTEReqHeightC[k],
+					&PTERequestSizeC[k],
+					&dpde0_bytes_per_frame_ub_c[k],
+					&meta_pte_bytes_per_frame_ub_c[k]);
+
+			PrefetchSourceLinesC[k] = dml32_CalculatePrefetchSourceLines(
+					myPipe[k].VRatioChroma,
+					myPipe[k].VTapsChroma,
+					myPipe[k].InterlaceEnable,
+					myPipe[k].ProgressiveToInterlaceUnitInOPP,
+					myPipe[k].SwathHeightC,
+					myPipe[k].SourceRotation,
+					myPipe[k].ViewportStationary,
+					SwathWidthC[k],
+					myPipe[k].ViewportHeightChroma,
+					myPipe[k].ViewportXStartC,
+					myPipe[k].ViewportYStartC,
+
+					/* Output */
+					&VInitPreFillC[k],
+					&MaxNumSwathC[k]);
+		} else {
+			PTEBufferSizeInRequestsForLuma[k] = PTEBufferSizeInRequestsLuma + PTEBufferSizeInRequestsChroma;
+			PTEBufferSizeInRequestsForChroma[k] = 0;
+			PixelPTEBytesPerRowC[k] = 0;
+			PDEAndMetaPTEBytesFrameC = 0;
+			MetaRowByteC[k] = 0;
+			MaxNumSwathC[k] = 0;
+			PrefetchSourceLinesC[k] = 0;
+			dpte_row_height_chroma_one_row_per_frame[k] = 0;
+			dpte_row_width_chroma_ub_one_row_per_frame[k] = 0;
+			PixelPTEBytesPerRowC_one_row_per_frame[k] = 0;
+		}
+
+		PDEAndMetaPTEBytesFrameY = dml32_CalculateVMAndRowBytes(
+				myPipe[k].ViewportStationary,
+				myPipe[k].DCCEnable,
+				myPipe[k].DPPPerSurface,
+				myPipe[k].BlockHeight256BytesY,
+				myPipe[k].BlockWidth256BytesY,
+				myPipe[k].SourcePixelFormat,
+				myPipe[k].SurfaceTiling,
+				myPipe[k].BytePerPixelY,
+				myPipe[k].SourceRotation,
+				SwathWidthY[k],
+				myPipe[k].ViewportHeight,
+				myPipe[k].ViewportXStart,
+				myPipe[k].ViewportYStart,
+				GPUVMEnable,
+				HostVMEnable,
+				HostVMMaxNonCachedPageTableLevels,
+				GPUVMMaxPageTableLevels,
+				GPUVMMinPageSizeKBytes[k],
+				HostVMMinPageSize,
+				PTEBufferSizeInRequestsForLuma[k],
+				myPipe[k].PitchY,
+				myPipe[k].DCCMetaPitchY,
+				myPipe[k].BlockWidthY,
+				myPipe[k].BlockHeightY,
+
+				/* Output */
+				&MetaRowByteY[k],
+				&PixelPTEBytesPerRowY[k],
+				&dpte_row_width_luma_ub[k],
+				&dpte_row_height_luma[k],
+				&dpte_row_height_linear_luma[k],
+				&PixelPTEBytesPerRowY_one_row_per_frame[k],
+				&dpte_row_width_luma_ub_one_row_per_frame[k],
+				&dpte_row_height_luma_one_row_per_frame[k],
+				&meta_req_width[k],
+				&meta_req_height[k],
+				&meta_row_width[k],
+				&meta_row_height[k],
+				&PixelPTEReqWidthY[k],
+				&PixelPTEReqHeightY[k],
+				&PTERequestSizeY[k],
+				&dpde0_bytes_per_frame_ub_l[k],
+				&meta_pte_bytes_per_frame_ub_l[k]);
+
+		PrefetchSourceLinesY[k] = dml32_CalculatePrefetchSourceLines(
+				myPipe[k].VRatio,
+				myPipe[k].VTaps,
+				myPipe[k].InterlaceEnable,
+				myPipe[k].ProgressiveToInterlaceUnitInOPP,
+				myPipe[k].SwathHeightY,
+				myPipe[k].SourceRotation,
+				myPipe[k].ViewportStationary,
+				SwathWidthY[k],
+				myPipe[k].ViewportHeight,
+				myPipe[k].ViewportXStart,
+				myPipe[k].ViewportYStart,
+
+				/* Output */
+				&VInitPreFillY[k],
+				&MaxNumSwathY[k]);
+
+		PDEAndMetaPTEBytesFrame[k] = PDEAndMetaPTEBytesFrameY + PDEAndMetaPTEBytesFrameC;
+		MetaRowByte[k] = MetaRowByteY[k] + MetaRowByteC[k];
+
+		if (PixelPTEBytesPerRowY[k] <= 64 * PTEBufferSizeInRequestsForLuma[k] &&
+				PixelPTEBytesPerRowC[k] <= 64 * PTEBufferSizeInRequestsForChroma[k]) {
+			PTEBufferSizeNotExceeded[k] = true;
+		} else {
+			PTEBufferSizeNotExceeded[k] = false;
+		}
+
+		one_row_per_frame_fits_in_buffer[k] = (PixelPTEBytesPerRowY_one_row_per_frame[k] <= 64 * 2 *
+			PTEBufferSizeInRequestsForLuma[k] &&
+			PixelPTEBytesPerRowC_one_row_per_frame[k] <= 64 * 2 * PTEBufferSizeInRequestsForChroma[k]);
+	}
+
+	dml32_CalculateMALLUseForStaticScreen(
+			NumberOfActiveSurfaces,
+			MALLAllocatedForDCN,
+			UseMALLForStaticScreen,   // mode
+			SurfaceSizeInMALL,
+			one_row_per_frame_fits_in_buffer,
+			/* Output */
+			UsesMALLForStaticScreen); // boolen
+
+	for (k = 0; k < NumberOfActiveSurfaces; ++k) {
+		PTE_BUFFER_MODE[k] = myPipe[k].FORCE_ONE_ROW_FOR_FRAME || UsesMALLForStaticScreen[k] ||
+				(UseMALLForPStateChange[k] == dm_use_mall_pstate_change_sub_viewport) ||
+				(UseMALLForPStateChange[k] == dm_use_mall_pstate_change_phantom_pipe) ||
+				(GPUVMMinPageSizeKBytes[k] > 64);
+		BIGK_FRAGMENT_SIZE[k] = dml_log2(GPUVMMinPageSizeKBytes[k] * 1024) - 12;
+	}
+
+	for (k = 0; k < NumberOfActiveSurfaces; ++k) {
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%d, SurfaceSizeInMALL = %d\n",  __func__, k, SurfaceSizeInMALL[k]);
+		dml_print("DML::%s: k=%d, UsesMALLForStaticScreen = %d\n",  __func__, k, UsesMALLForStaticScreen[k]);
+#endif
+		use_one_row_for_frame[k] = myPipe[k].FORCE_ONE_ROW_FOR_FRAME || UsesMALLForStaticScreen[k] ||
+				(UseMALLForPStateChange[k] == dm_use_mall_pstate_change_sub_viewport) ||
+				(UseMALLForPStateChange[k] == dm_use_mall_pstate_change_phantom_pipe) ||
+				(GPUVMMinPageSizeKBytes[k] > 64 && IsVertical(myPipe[k].SourceRotation));
+
+		use_one_row_for_frame_flip[k] = use_one_row_for_frame[k] &&
+				!(UseMALLForPStateChange[k] == dm_use_mall_pstate_change_full_frame);
+
+		if (use_one_row_for_frame[k]) {
+			dpte_row_height_luma[k] = dpte_row_height_luma_one_row_per_frame[k];
+			dpte_row_width_luma_ub[k] = dpte_row_width_luma_ub_one_row_per_frame[k];
+			PixelPTEBytesPerRowY[k] = PixelPTEBytesPerRowY_one_row_per_frame[k];
+			dpte_row_height_chroma[k] = dpte_row_height_chroma_one_row_per_frame[k];
+			dpte_row_width_chroma_ub[k] = dpte_row_width_chroma_ub_one_row_per_frame[k];
+			PixelPTEBytesPerRowC[k] = PixelPTEBytesPerRowC_one_row_per_frame[k];
+			PTEBufferSizeNotExceeded[k] = one_row_per_frame_fits_in_buffer[k];
+		}
+
+		if (MetaRowByte[k] <= DCCMetaBufferSizeBytes)
+			DCCMetaBufferSizeNotExceeded[k] = true;
+		else
+			DCCMetaBufferSizeNotExceeded[k] = false;
+
+		PixelPTEBytesPerRow[k] = PixelPTEBytesPerRowY[k] + PixelPTEBytesPerRowC[k];
+		if (use_one_row_for_frame[k])
+			PixelPTEBytesPerRow[k] = PixelPTEBytesPerRow[k] / 2;
+
+		dml32_CalculateRowBandwidth(
+				GPUVMEnable,
+				myPipe[k].SourcePixelFormat,
+				myPipe[k].VRatio,
+				myPipe[k].VRatioChroma,
+				myPipe[k].DCCEnable,
+				myPipe[k].HTotal / myPipe[k].PixelClock,
+				MetaRowByteY[k], MetaRowByteC[k],
+				meta_row_height[k],
+				meta_row_height_chroma[k],
+				PixelPTEBytesPerRowY[k],
+				PixelPTEBytesPerRowC[k],
+				dpte_row_height_luma[k],
+				dpte_row_height_chroma[k],
+
+				/* Output */
+				&meta_row_bw[k],
+				&dpte_row_bw[k]);
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%d, use_one_row_for_frame        = %d\n",  __func__, k, use_one_row_for_frame[k]);
+		dml_print("DML::%s: k=%d, use_one_row_for_frame_flip   = %d\n",
+				__func__, k, use_one_row_for_frame_flip[k]);
+		dml_print("DML::%s: k=%d, UseMALLForPStateChange       = %d\n",
+				__func__, k, UseMALLForPStateChange[k]);
+		dml_print("DML::%s: k=%d, dpte_row_height_luma         = %d\n",  __func__, k, dpte_row_height_luma[k]);
+		dml_print("DML::%s: k=%d, dpte_row_width_luma_ub       = %d\n",
+				__func__, k, dpte_row_width_luma_ub[k]);
+		dml_print("DML::%s: k=%d, PixelPTEBytesPerRowY         = %d\n",  __func__, k, PixelPTEBytesPerRowY[k]);
+		dml_print("DML::%s: k=%d, dpte_row_height_chroma       = %d\n",
+				__func__, k, dpte_row_height_chroma[k]);
+		dml_print("DML::%s: k=%d, dpte_row_width_chroma_ub     = %d\n",
+				__func__, k, dpte_row_width_chroma_ub[k]);
+		dml_print("DML::%s: k=%d, PixelPTEBytesPerRowC         = %d\n",  __func__, k, PixelPTEBytesPerRowC[k]);
+		dml_print("DML::%s: k=%d, PixelPTEBytesPerRow          = %d\n",  __func__, k, PixelPTEBytesPerRow[k]);
+		dml_print("DML::%s: k=%d, PTEBufferSizeNotExceeded     = %d\n",
+				__func__, k, PTEBufferSizeNotExceeded[k]);
+		dml_print("DML::%s: k=%d, PTE_BUFFER_MODE              = %d\n", __func__, k, PTE_BUFFER_MODE[k]);
+		dml_print("DML::%s: k=%d, BIGK_FRAGMENT_SIZE           = %d\n", __func__, k, BIGK_FRAGMENT_SIZE[k]);
+#endif
+	}
+} // CalculateVMRowAndSwath
+
+unsigned int dml32_CalculateVMAndRowBytes(
+		bool ViewportStationary,
+		bool DCCEnable,
+		unsigned int NumberOfDPPs,
+		unsigned int BlockHeight256Bytes,
+		unsigned int BlockWidth256Bytes,
+		enum source_format_class SourcePixelFormat,
+		unsigned int SurfaceTiling,
+		unsigned int BytePerPixel,
+		enum dm_rotation_angle SourceRotation,
+		double SwathWidth,
+		unsigned int ViewportHeight,
+		unsigned int    ViewportXStart,
+		unsigned int    ViewportYStart,
+		bool GPUVMEnable,
+		bool HostVMEnable,
+		unsigned int HostVMMaxNonCachedPageTableLevels,
+		unsigned int GPUVMMaxPageTableLevels,
+		unsigned int GPUVMMinPageSizeKBytes,
+		unsigned int HostVMMinPageSize,
+		unsigned int PTEBufferSizeInRequests,
+		unsigned int Pitch,
+		unsigned int DCCMetaPitch,
+		unsigned int MacroTileWidth,
+		unsigned int MacroTileHeight,
+
+		/* Output */
+		unsigned int *MetaRowByte,
+		unsigned int *PixelPTEBytesPerRow,
+		unsigned int    *dpte_row_width_ub,
+		unsigned int *dpte_row_height,
+		unsigned int *dpte_row_height_linear,
+		unsigned int    *PixelPTEBytesPerRow_one_row_per_frame,
+		unsigned int    *dpte_row_width_ub_one_row_per_frame,
+		unsigned int    *dpte_row_height_one_row_per_frame,
+		unsigned int *MetaRequestWidth,
+		unsigned int *MetaRequestHeight,
+		unsigned int *meta_row_width,
+		unsigned int *meta_row_height,
+		unsigned int *PixelPTEReqWidth,
+		unsigned int *PixelPTEReqHeight,
+		unsigned int *PTERequestSize,
+		unsigned int    *DPDE0BytesFrame,
+		unsigned int    *MetaPTEBytesFrame)
+{
+	unsigned int MPDEBytesFrame;
+	unsigned int DCCMetaSurfaceBytes;
+	unsigned int ExtraDPDEBytesFrame;
+	unsigned int PDEAndMetaPTEBytesFrame;
+	unsigned int HostVMDynamicLevels = 0;
+	unsigned int    MacroTileSizeBytes;
+	unsigned int    vp_height_meta_ub;
+	unsigned int    vp_height_dpte_ub;
+	unsigned int PixelPTEReqWidth_linear = 0; // VBA_DELTA. VBA doesn't calculate this
+
+	if (GPUVMEnable == true && HostVMEnable == true) {
+		if (HostVMMinPageSize < 2048)
+			HostVMDynamicLevels = HostVMMaxNonCachedPageTableLevels;
+		else if (HostVMMinPageSize >= 2048 && HostVMMinPageSize < 1048576)
+			HostVMDynamicLevels = dml_max(0, (int) HostVMMaxNonCachedPageTableLevels - 1);
+		else
+			HostVMDynamicLevels = dml_max(0, (int) HostVMMaxNonCachedPageTableLevels - 2);
+	}
+
+	*MetaRequestHeight = 8 * BlockHeight256Bytes;
+	*MetaRequestWidth = 8 * BlockWidth256Bytes;
+	if (SurfaceTiling == dm_sw_linear) {
+		*meta_row_height = 32;
+		*meta_row_width = dml_floor(ViewportXStart + SwathWidth + *MetaRequestWidth - 1, *MetaRequestWidth)
+				- dml_floor(ViewportXStart, *MetaRequestWidth);
+	} else if (!IsVertical(SourceRotation)) {
+		*meta_row_height = *MetaRequestHeight;
+		if (ViewportStationary && NumberOfDPPs == 1) {
+			*meta_row_width = dml_floor(ViewportXStart + SwathWidth + *MetaRequestWidth - 1,
+					*MetaRequestWidth) - dml_floor(ViewportXStart, *MetaRequestWidth);
+		} else {
+			*meta_row_width = dml_ceil(SwathWidth - 1, *MetaRequestWidth) + *MetaRequestWidth;
+		}
+		*MetaRowByte = *meta_row_width * *MetaRequestHeight * BytePerPixel / 256.0;
+	} else {
+		*meta_row_height = *MetaRequestWidth;
+		if (ViewportStationary && NumberOfDPPs == 1) {
+			*meta_row_width = dml_floor(ViewportYStart + ViewportHeight + *MetaRequestHeight - 1,
+					*MetaRequestHeight) - dml_floor(ViewportYStart, *MetaRequestHeight);
+		} else {
+			*meta_row_width = dml_ceil(SwathWidth - 1, *MetaRequestHeight) + *MetaRequestHeight;
+		}
+		*MetaRowByte = *meta_row_width * *MetaRequestWidth * BytePerPixel / 256.0;
+	}
+
+	if (ViewportStationary && (NumberOfDPPs == 1 || !IsVertical(SourceRotation))) {
+		vp_height_meta_ub = dml_floor(ViewportYStart + ViewportHeight + 64 * BlockHeight256Bytes - 1,
+				64 * BlockHeight256Bytes) - dml_floor(ViewportYStart, 64 * BlockHeight256Bytes);
+	} else if (!IsVertical(SourceRotation)) {
+		vp_height_meta_ub = dml_ceil(ViewportHeight - 1, 64 * BlockHeight256Bytes) + 64 * BlockHeight256Bytes;
+	} else {
+		vp_height_meta_ub = dml_ceil(SwathWidth - 1, 64 * BlockHeight256Bytes) + 64 * BlockHeight256Bytes;
+	}
+
+	DCCMetaSurfaceBytes = DCCMetaPitch * vp_height_meta_ub * BytePerPixel / 256.0;
+
+	if (GPUVMEnable == true) {
+		*MetaPTEBytesFrame = (dml_ceil((double) (DCCMetaSurfaceBytes - 4.0 * 1024.0) /
+				(8 * 4.0 * 1024), 1) + 1) * 64;
+		MPDEBytesFrame = 128 * (GPUVMMaxPageTableLevels - 1);
+	} else {
+		*MetaPTEBytesFrame = 0;
+		MPDEBytesFrame = 0;
+	}
+
+	if (DCCEnable != true) {
+		*MetaPTEBytesFrame = 0;
+		MPDEBytesFrame = 0;
+		*MetaRowByte = 0;
+	}
+
+	MacroTileSizeBytes = MacroTileWidth * BytePerPixel * MacroTileHeight;
+
+	if (GPUVMEnable == true && GPUVMMaxPageTableLevels > 1) {
+		if (ViewportStationary && (NumberOfDPPs == 1 || !IsVertical(SourceRotation))) {
+			vp_height_dpte_ub = dml_floor(ViewportYStart + ViewportHeight +
+					MacroTileHeight - 1, MacroTileHeight) -
+					dml_floor(ViewportYStart, MacroTileHeight);
+		} else if (!IsVertical(SourceRotation)) {
+			vp_height_dpte_ub = dml_ceil(ViewportHeight - 1, MacroTileHeight) + MacroTileHeight;
+		} else {
+			vp_height_dpte_ub = dml_ceil(SwathWidth - 1, MacroTileHeight) + MacroTileHeight;
+		}
+		*DPDE0BytesFrame = 64 * (dml_ceil((Pitch * vp_height_dpte_ub * BytePerPixel - MacroTileSizeBytes) /
+				(8 * 2097152), 1) + 1);
+		ExtraDPDEBytesFrame = 128 * (GPUVMMaxPageTableLevels - 2);
+	} else {
+		*DPDE0BytesFrame = 0;
+		ExtraDPDEBytesFrame = 0;
+		vp_height_dpte_ub = 0;
+	}
+
+	PDEAndMetaPTEBytesFrame = *MetaPTEBytesFrame + MPDEBytesFrame + *DPDE0BytesFrame + ExtraDPDEBytesFrame;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: DCCEnable = %d\n", __func__, DCCEnable);
+	dml_print("DML::%s: GPUVMEnable = %d\n", __func__, GPUVMEnable);
+	dml_print("DML::%s: SwModeLinear = %d\n", __func__, SurfaceTiling == dm_sw_linear);
+	dml_print("DML::%s: BytePerPixel = %d\n", __func__, BytePerPixel);
+	dml_print("DML::%s: GPUVMMaxPageTableLevels = %d\n", __func__, GPUVMMaxPageTableLevels);
+	dml_print("DML::%s: BlockHeight256Bytes = %d\n", __func__, BlockHeight256Bytes);
+	dml_print("DML::%s: BlockWidth256Bytes = %d\n", __func__, BlockWidth256Bytes);
+	dml_print("DML::%s: MacroTileHeight = %d\n", __func__, MacroTileHeight);
+	dml_print("DML::%s: MacroTileWidth = %d\n", __func__, MacroTileWidth);
+	dml_print("DML::%s: MetaPTEBytesFrame = %d\n", __func__, *MetaPTEBytesFrame);
+	dml_print("DML::%s: MPDEBytesFrame = %d\n", __func__, MPDEBytesFrame);
+	dml_print("DML::%s: DPDE0BytesFrame = %d\n", __func__, *DPDE0BytesFrame);
+	dml_print("DML::%s: ExtraDPDEBytesFrame= %d\n", __func__, ExtraDPDEBytesFrame);
+	dml_print("DML::%s: PDEAndMetaPTEBytesFrame = %d\n", __func__, PDEAndMetaPTEBytesFrame);
+	dml_print("DML::%s: ViewportHeight = %d\n", __func__, ViewportHeight);
+	dml_print("DML::%s: SwathWidth = %d\n", __func__, SwathWidth);
+	dml_print("DML::%s: vp_height_dpte_ub = %d\n", __func__, vp_height_dpte_ub);
+#endif
+
+	if (HostVMEnable == true)
+		PDEAndMetaPTEBytesFrame = PDEAndMetaPTEBytesFrame * (1 + 8 * HostVMDynamicLevels);
+
+	if (SurfaceTiling == dm_sw_linear) {
+		*PixelPTEReqHeight = 1;
+		*PixelPTEReqWidth = GPUVMMinPageSizeKBytes * 1024 * 8 / BytePerPixel;
+		PixelPTEReqWidth_linear = GPUVMMinPageSizeKBytes * 1024 * 8 / BytePerPixel;
+		*PTERequestSize = 64;
+	} else if (GPUVMMinPageSizeKBytes == 4) {
+		*PixelPTEReqHeight = 16 * BlockHeight256Bytes;
+		*PixelPTEReqWidth = 16 * BlockWidth256Bytes;
+		*PTERequestSize = 128;
+	} else {
+		*PixelPTEReqHeight = MacroTileHeight;
+		*PixelPTEReqWidth = 8 *  1024 * GPUVMMinPageSizeKBytes / (MacroTileHeight * BytePerPixel);
+		*PTERequestSize = 64;
+	}
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: GPUVMMinPageSizeKBytes = %d\n", __func__, GPUVMMinPageSizeKBytes);
+	dml_print("DML::%s: PDEAndMetaPTEBytesFrame = %d (after HostVM factor)\n", __func__, PDEAndMetaPTEBytesFrame);
+	dml_print("DML::%s: PixelPTEReqHeight = %d\n", __func__, *PixelPTEReqHeight);
+	dml_print("DML::%s: PixelPTEReqWidth = %d\n", __func__, *PixelPTEReqWidth);
+	dml_print("DML::%s: PixelPTEReqWidth_linear = %d\n", __func__, PixelPTEReqWidth_linear);
+	dml_print("DML::%s: PTERequestSize = %d\n", __func__, *PTERequestSize);
+	dml_print("DML::%s: Pitch = %d\n", __func__, Pitch);
+#endif
+
+	*dpte_row_height_one_row_per_frame = vp_height_dpte_ub;
+	*dpte_row_width_ub_one_row_per_frame = (dml_ceil(((double)Pitch * (double)*dpte_row_height_one_row_per_frame /
+			(double) *PixelPTEReqHeight - 1) / (double) *PixelPTEReqWidth, 1) + 1) *
+					(double) *PixelPTEReqWidth;
+	*PixelPTEBytesPerRow_one_row_per_frame = *dpte_row_width_ub_one_row_per_frame / *PixelPTEReqWidth *
+			*PTERequestSize;
+
+	if (SurfaceTiling == dm_sw_linear) {
+		*dpte_row_height = dml_min(128, 1 << (unsigned int) dml_floor(dml_log2(PTEBufferSizeInRequests *
+				*PixelPTEReqWidth / Pitch), 1));
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: dpte_row_height = %d (1)\n", __func__,
+				PTEBufferSizeInRequests * *PixelPTEReqWidth / Pitch);
+		dml_print("DML::%s: dpte_row_height = %f (2)\n", __func__,
+				dml_log2(PTEBufferSizeInRequests * *PixelPTEReqWidth / Pitch));
+		dml_print("DML::%s: dpte_row_height = %f (3)\n", __func__,
+				dml_floor(dml_log2(PTEBufferSizeInRequests * *PixelPTEReqWidth / Pitch), 1));
+		dml_print("DML::%s: dpte_row_height = %d (4)\n", __func__,
+				1 << (unsigned int) dml_floor(dml_log2(PTEBufferSizeInRequests *
+						*PixelPTEReqWidth / Pitch), 1));
+		dml_print("DML::%s: dpte_row_height = %d\n", __func__, *dpte_row_height);
+#endif
+		*dpte_row_width_ub = dml_ceil(((double) Pitch * (double) *dpte_row_height - 1),
+				(double) *PixelPTEReqWidth) + *PixelPTEReqWidth;
+		*PixelPTEBytesPerRow = *dpte_row_width_ub / (double)*PixelPTEReqWidth * (double)*PTERequestSize;
+
+		// VBA_DELTA, VBA doesn't have programming value for pte row height linear.
+		*dpte_row_height_linear = 1 << (unsigned int) dml_floor(dml_log2(PTEBufferSizeInRequests *
+				PixelPTEReqWidth_linear / Pitch), 1);
+		if (*dpte_row_height_linear > 128)
+			*dpte_row_height_linear = 128;
+
+	} else if (!IsVertical(SourceRotation)) {
+		*dpte_row_height = *PixelPTEReqHeight;
+
+		if (GPUVMMinPageSizeKBytes > 64) {
+			*dpte_row_width_ub = (dml_ceil((Pitch * *dpte_row_height / *PixelPTEReqHeight - 1) /
+					*PixelPTEReqWidth, 1) + 1) * *PixelPTEReqWidth;
+		} else if (ViewportStationary && (NumberOfDPPs == 1)) {
+			*dpte_row_width_ub = dml_floor(ViewportXStart + SwathWidth +
+					*PixelPTEReqWidth - 1, *PixelPTEReqWidth) -
+					dml_floor(ViewportXStart, *PixelPTEReqWidth);
+		} else {
+			*dpte_row_width_ub = (dml_ceil((SwathWidth - 1) / *PixelPTEReqWidth, 1) + 1) *
+					*PixelPTEReqWidth;
+		}
+
+		*PixelPTEBytesPerRow = *dpte_row_width_ub / *PixelPTEReqWidth * *PTERequestSize;
+	} else {
+		*dpte_row_height = dml_min(*PixelPTEReqWidth, MacroTileWidth);
+
+		if (ViewportStationary && (NumberOfDPPs == 1)) {
+			*dpte_row_width_ub = dml_floor(ViewportYStart + ViewportHeight + *PixelPTEReqHeight - 1,
+					*PixelPTEReqHeight) - dml_floor(ViewportYStart, *PixelPTEReqHeight);
+		} else {
+			*dpte_row_width_ub = (dml_ceil((SwathWidth - 1) / *PixelPTEReqHeight, 1) + 1)
+					* *PixelPTEReqHeight;
+		}
+
+		*PixelPTEBytesPerRow = *dpte_row_width_ub / *PixelPTEReqHeight * *PTERequestSize;
+	}
+
+	if (GPUVMEnable != true)
+		*PixelPTEBytesPerRow = 0;
+	if (HostVMEnable == true)
+		*PixelPTEBytesPerRow = *PixelPTEBytesPerRow * (1 + 8 * HostVMDynamicLevels);
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: GPUVMMinPageSizeKBytes = %d\n", __func__, GPUVMMinPageSizeKBytes);
+	dml_print("DML::%s: dpte_row_height = %d\n", __func__, *dpte_row_height);
+	dml_print("DML::%s: dpte_row_height_linear = %d\n", __func__, *dpte_row_height_linear);
+	dml_print("DML::%s: dpte_row_width_ub = %d\n", __func__, *dpte_row_width_ub);
+	dml_print("DML::%s: PixelPTEBytesPerRow = %d\n", __func__, *PixelPTEBytesPerRow);
+	dml_print("DML::%s: PTEBufferSizeInRequests = %d\n", __func__, PTEBufferSizeInRequests);
+	dml_print("DML::%s: dpte_row_height_one_row_per_frame = %d\n", __func__, *dpte_row_height_one_row_per_frame);
+	dml_print("DML::%s: dpte_row_width_ub_one_row_per_frame = %d\n",
+			__func__, *dpte_row_width_ub_one_row_per_frame);
+	dml_print("DML::%s: PixelPTEBytesPerRow_one_row_per_frame = %d\n",
+			__func__, *PixelPTEBytesPerRow_one_row_per_frame);
+	dml_print("DML: vm_bytes = meta_pte_bytes_per_frame (per_pipe) = MetaPTEBytesFrame = : %i\n",
+			*MetaPTEBytesFrame);
+#endif
+
+	return PDEAndMetaPTEBytesFrame;
+} // CalculateVMAndRowBytes
+
+double dml32_CalculatePrefetchSourceLines(
+		double VRatio,
+		unsigned int VTaps,
+		bool Interlace,
+		bool ProgressiveToInterlaceUnitInOPP,
+		unsigned int SwathHeight,
+		enum dm_rotation_angle SourceRotation,
+		bool ViewportStationary,
+		double SwathWidth,
+		unsigned int ViewportHeight,
+		unsigned int ViewportXStart,
+		unsigned int ViewportYStart,
+
+		/* Output */
+		double *VInitPreFill,
+		unsigned int *MaxNumSwath)
+{
+
+	unsigned int vp_start_rot;
+	unsigned int sw0_tmp;
+	unsigned int MaxPartialSwath;
+	double numLines;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: VRatio = %f\n", __func__, VRatio);
+	dml_print("DML::%s: VTaps = %d\n", __func__, VTaps);
+	dml_print("DML::%s: ViewportXStart = %d\n", __func__, ViewportXStart);
+	dml_print("DML::%s: ViewportYStart = %d\n", __func__, ViewportYStart);
+	dml_print("DML::%s: ViewportStationary = %d\n", __func__, ViewportStationary);
+	dml_print("DML::%s: SwathHeight = %d\n", __func__, SwathHeight);
+#endif
+	if (ProgressiveToInterlaceUnitInOPP)
+		*VInitPreFill = dml_floor((VRatio + (double) VTaps + 1) / 2.0, 1);
+	else
+		*VInitPreFill = dml_floor((VRatio + (double) VTaps + 1 + Interlace * 0.5 * VRatio) / 2.0, 1);
+
+	if (ViewportStationary) {
+		if (SourceRotation == dm_rotation_180 || SourceRotation == dm_rotation_180m) {
+			vp_start_rot = SwathHeight -
+					(((unsigned int) (ViewportYStart + ViewportHeight - 1) % SwathHeight) + 1);
+		} else if (SourceRotation == dm_rotation_270 || SourceRotation == dm_rotation_90m) {
+			vp_start_rot = ViewportXStart;
+		} else if (SourceRotation == dm_rotation_90 || SourceRotation == dm_rotation_270m) {
+			vp_start_rot = SwathHeight -
+					(((unsigned int)(ViewportYStart + SwathWidth - 1) % SwathHeight) + 1);
+		} else {
+			vp_start_rot = ViewportYStart;
+		}
+		sw0_tmp = SwathHeight - (vp_start_rot % SwathHeight);
+		if (sw0_tmp < *VInitPreFill)
+			*MaxNumSwath = dml_ceil((*VInitPreFill - sw0_tmp) / SwathHeight, 1) + 1;
+		else
+			*MaxNumSwath = 1;
+		MaxPartialSwath = dml_max(1, (unsigned int) (vp_start_rot + *VInitPreFill - 1) % SwathHeight);
+	} else {
+		*MaxNumSwath = dml_ceil((*VInitPreFill - 1.0) / SwathHeight, 1) + 1;
+		if (*VInitPreFill > 1)
+			MaxPartialSwath = dml_max(1, (unsigned int) (*VInitPreFill - 2) % SwathHeight);
+		else
+			MaxPartialSwath = dml_max(1, (unsigned int) (*VInitPreFill + SwathHeight - 2) % SwathHeight);
+	}
+	numLines = *MaxNumSwath * SwathHeight + MaxPartialSwath;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: vp_start_rot = %d\n", __func__, vp_start_rot);
+	dml_print("DML::%s: VInitPreFill = %d\n", __func__, *VInitPreFill);
+	dml_print("DML::%s: MaxPartialSwath = %d\n", __func__, MaxPartialSwath);
+	dml_print("DML::%s: MaxNumSwath = %d\n", __func__, *MaxNumSwath);
+	dml_print("DML::%s: Prefetch source lines = %3.2f\n", __func__, numLines);
+#endif
+	return numLines;
+
+} // CalculatePrefetchSourceLines
+
+void dml32_CalculateMALLUseForStaticScreen(
+		unsigned int NumberOfActiveSurfaces,
+		unsigned int MALLAllocatedForDCNFinal,
+		enum dm_use_mall_for_static_screen_mode *UseMALLForStaticScreen,
+		unsigned int SurfaceSizeInMALL[],
+		bool one_row_per_frame_fits_in_buffer[],
+
+		/* output */
+		bool UsesMALLForStaticScreen[])
+{
+	unsigned int k;
+	unsigned int SurfaceToAddToMALL;
+	bool CanAddAnotherSurfaceToMALL;
+	unsigned int TotalSurfaceSizeInMALL;
+
+	TotalSurfaceSizeInMALL = 0;
+	for (k = 0; k < NumberOfActiveSurfaces; ++k) {
+		UsesMALLForStaticScreen[k] = (UseMALLForStaticScreen[k] == dm_use_mall_static_screen_enable);
+		if (UsesMALLForStaticScreen[k])
+			TotalSurfaceSizeInMALL = TotalSurfaceSizeInMALL + SurfaceSizeInMALL[k];
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%d, UsesMALLForStaticScreen = %d\n",  __func__, k, UsesMALLForStaticScreen[k]);
+		dml_print("DML::%s: k=%d, TotalSurfaceSizeInMALL = %d\n",  __func__, k, TotalSurfaceSizeInMALL);
+#endif
+	}
+
+	SurfaceToAddToMALL = 0;
+	CanAddAnotherSurfaceToMALL = true;
+	while (CanAddAnotherSurfaceToMALL) {
+		CanAddAnotherSurfaceToMALL = false;
+		for (k = 0; k < NumberOfActiveSurfaces; ++k) {
+			if (TotalSurfaceSizeInMALL + SurfaceSizeInMALL[k] <= MALLAllocatedForDCNFinal * 1024 * 1024 &&
+					!UsesMALLForStaticScreen[k] &&
+					UseMALLForStaticScreen[k] != dm_use_mall_static_screen_disable &&
+					one_row_per_frame_fits_in_buffer[k] &&
+					(!CanAddAnotherSurfaceToMALL ||
+					SurfaceSizeInMALL[k] < SurfaceSizeInMALL[SurfaceToAddToMALL])) {
+				CanAddAnotherSurfaceToMALL = true;
+				SurfaceToAddToMALL = k;
+#ifdef __DML_VBA_DEBUG__
+				dml_print("DML::%s: k=%d, UseMALLForStaticScreen = %d (dis, en, optimize)\n",
+						__func__, k, UseMALLForStaticScreen[k]);
+#endif
+			}
+		}
+		if (CanAddAnotherSurfaceToMALL) {
+			UsesMALLForStaticScreen[SurfaceToAddToMALL] = true;
+			TotalSurfaceSizeInMALL = TotalSurfaceSizeInMALL + SurfaceSizeInMALL[SurfaceToAddToMALL];
+
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: SurfaceToAddToMALL       = %d\n",  __func__, SurfaceToAddToMALL);
+			dml_print("DML::%s: TotalSurfaceSizeInMALL   = %d\n",  __func__, TotalSurfaceSizeInMALL);
+#endif
+
+		}
+	}
+}
+
+void dml32_CalculateRowBandwidth(
+		bool GPUVMEnable,
+		enum source_format_class SourcePixelFormat,
+		double VRatio,
+		double VRatioChroma,
+		bool DCCEnable,
+		double LineTime,
+		unsigned int MetaRowByteLuma,
+		unsigned int MetaRowByteChroma,
+		unsigned int meta_row_height_luma,
+		unsigned int meta_row_height_chroma,
+		unsigned int PixelPTEBytesPerRowLuma,
+		unsigned int PixelPTEBytesPerRowChroma,
+		unsigned int dpte_row_height_luma,
+		unsigned int dpte_row_height_chroma,
+		/* Output */
+		double *meta_row_bw,
+		double *dpte_row_bw)
+{
+	if (DCCEnable != true) {
+		*meta_row_bw = 0;
+	} else if (SourcePixelFormat == dm_420_8 || SourcePixelFormat == dm_420_10 || SourcePixelFormat == dm_420_12 ||
+			SourcePixelFormat == dm_rgbe_alpha) {
+		*meta_row_bw = VRatio * MetaRowByteLuma / (meta_row_height_luma * LineTime) + VRatioChroma *
+				MetaRowByteChroma / (meta_row_height_chroma * LineTime);
+	} else {
+		*meta_row_bw = VRatio * MetaRowByteLuma / (meta_row_height_luma * LineTime);
+	}
+
+	if (GPUVMEnable != true) {
+		*dpte_row_bw = 0;
+	} else if (SourcePixelFormat == dm_420_8 || SourcePixelFormat == dm_420_10 || SourcePixelFormat == dm_420_12 ||
+			SourcePixelFormat == dm_rgbe_alpha) {
+		*dpte_row_bw = VRatio * PixelPTEBytesPerRowLuma / (dpte_row_height_luma * LineTime) +
+				VRatioChroma * PixelPTEBytesPerRowChroma / (dpte_row_height_chroma * LineTime);
+	} else {
+		*dpte_row_bw = VRatio * PixelPTEBytesPerRowLuma / (dpte_row_height_luma * LineTime);
+	}
+}
+
+double dml32_CalculateUrgentLatency(
+		double UrgentLatencyPixelDataOnly,
+		double UrgentLatencyPixelMixedWithVMData,
+		double UrgentLatencyVMDataOnly,
+		bool   DoUrgentLatencyAdjustment,
+		double UrgentLatencyAdjustmentFabricClockComponent,
+		double UrgentLatencyAdjustmentFabricClockReference,
+		double FabricClock)
+{
+	double   ret;
+
+	ret = dml_max3(UrgentLatencyPixelDataOnly, UrgentLatencyPixelMixedWithVMData, UrgentLatencyVMDataOnly);
+	if (DoUrgentLatencyAdjustment == true) {
+		ret = ret + UrgentLatencyAdjustmentFabricClockComponent *
+				(UrgentLatencyAdjustmentFabricClockReference / FabricClock - 1);
+	}
+	return ret;
+}
+
+void dml32_CalculateUrgentBurstFactor(
+		enum dm_use_mall_for_pstate_change_mode UseMALLForPStateChange,
+		unsigned int    swath_width_luma_ub,
+		unsigned int    swath_width_chroma_ub,
+		unsigned int SwathHeightY,
+		unsigned int SwathHeightC,
+		double  LineTime,
+		double  UrgentLatency,
+		double  CursorBufferSize,
+		unsigned int CursorWidth,
+		unsigned int CursorBPP,
+		double  VRatio,
+		double  VRatioC,
+		double  BytePerPixelInDETY,
+		double  BytePerPixelInDETC,
+		unsigned int    DETBufferSizeY,
+		unsigned int    DETBufferSizeC,
+		/* Output */
+		double *UrgentBurstFactorCursor,
+		double *UrgentBurstFactorLuma,
+		double *UrgentBurstFactorChroma,
+		bool   *NotEnoughUrgentLatencyHiding)
+{
+	double       LinesInDETLuma;
+	double       LinesInDETChroma;
+	unsigned int LinesInCursorBuffer;
+	double       CursorBufferSizeInTime;
+	double       DETBufferSizeInTimeLuma;
+	double       DETBufferSizeInTimeChroma;
+
+	*NotEnoughUrgentLatencyHiding = 0;
+
+	if (CursorWidth > 0) {
+		LinesInCursorBuffer = 1 << (unsigned int) dml_floor(dml_log2(CursorBufferSize * 1024.0 /
+				(CursorWidth * CursorBPP / 8.0)), 1.0);
+		if (VRatio > 0) {
+			CursorBufferSizeInTime = LinesInCursorBuffer * LineTime / VRatio;
+			if (CursorBufferSizeInTime - UrgentLatency <= 0) {
+				*NotEnoughUrgentLatencyHiding = 1;
+				*UrgentBurstFactorCursor = 0;
+			} else {
+				*UrgentBurstFactorCursor = CursorBufferSizeInTime /
+						(CursorBufferSizeInTime - UrgentLatency);
+			}
+		} else {
+			*UrgentBurstFactorCursor = 1;
+		}
+	}
+
+	LinesInDETLuma = (UseMALLForPStateChange == dm_use_mall_pstate_change_phantom_pipe ? 1024*1024 :
+			DETBufferSizeY) / BytePerPixelInDETY / swath_width_luma_ub;
+
+	if (VRatio > 0) {
+		DETBufferSizeInTimeLuma = dml_floor(LinesInDETLuma, SwathHeightY) * LineTime / VRatio;
+		if (DETBufferSizeInTimeLuma - UrgentLatency <= 0) {
+			*NotEnoughUrgentLatencyHiding = 1;
+			*UrgentBurstFactorLuma = 0;
+		} else {
+			*UrgentBurstFactorLuma = DETBufferSizeInTimeLuma / (DETBufferSizeInTimeLuma - UrgentLatency);
+		}
+	} else {
+		*UrgentBurstFactorLuma = 1;
+	}
+
+	if (BytePerPixelInDETC > 0) {
+		LinesInDETChroma = (UseMALLForPStateChange == dm_use_mall_pstate_change_phantom_pipe ?
+					1024 * 1024 : DETBufferSizeC) / BytePerPixelInDETC
+					/ swath_width_chroma_ub;
+
+		if (VRatio > 0) {
+			DETBufferSizeInTimeChroma = dml_floor(LinesInDETChroma, SwathHeightC) * LineTime / VRatio;
+			if (DETBufferSizeInTimeChroma - UrgentLatency <= 0) {
+				*NotEnoughUrgentLatencyHiding = 1;
+				*UrgentBurstFactorChroma = 0;
+			} else {
+				*UrgentBurstFactorChroma = DETBufferSizeInTimeChroma
+						/ (DETBufferSizeInTimeChroma - UrgentLatency);
+			}
+		} else {
+			*UrgentBurstFactorChroma = 1;
+		}
+	}
+} // CalculateUrgentBurstFactor
+
+void dml32_CalculateDCFCLKDeepSleep(
+		unsigned int NumberOfActiveSurfaces,
+		unsigned int BytePerPixelY[],
+		unsigned int BytePerPixelC[],
+		double VRatio[],
+		double VRatioChroma[],
+		double SwathWidthY[],
+		double SwathWidthC[],
+		unsigned int DPPPerSurface[],
+		double HRatio[],
+		double HRatioChroma[],
+		double PixelClock[],
+		double PSCL_THROUGHPUT[],
+		double PSCL_THROUGHPUT_CHROMA[],
+		double Dppclk[],
+		double ReadBandwidthLuma[],
+		double ReadBandwidthChroma[],
+		unsigned int ReturnBusWidth,
+
+		/* Output */
+		double *DCFClkDeepSleep)
+{
+	unsigned int k;
+	double   DisplayPipeLineDeliveryTimeLuma;
+	double   DisplayPipeLineDeliveryTimeChroma;
+	double   DCFClkDeepSleepPerSurface[DC__NUM_DPP__MAX];
+	double ReadBandwidth = 0.0;
+
+	for (k = 0; k < NumberOfActiveSurfaces; ++k) {
+
+		if (VRatio[k] <= 1) {
+			DisplayPipeLineDeliveryTimeLuma = SwathWidthY[k] * DPPPerSurface[k] / HRatio[k]
+					/ PixelClock[k];
+		} else {
+			DisplayPipeLineDeliveryTimeLuma = SwathWidthY[k] / PSCL_THROUGHPUT[k] / Dppclk[k];
+		}
+		if (BytePerPixelC[k] == 0) {
+			DisplayPipeLineDeliveryTimeChroma = 0;
+		} else {
+			if (VRatioChroma[k] <= 1) {
+				DisplayPipeLineDeliveryTimeChroma = SwathWidthC[k] *
+						DPPPerSurface[k] / HRatioChroma[k] / PixelClock[k];
+			} else {
+				DisplayPipeLineDeliveryTimeChroma = SwathWidthC[k] / PSCL_THROUGHPUT_CHROMA[k]
+						/ Dppclk[k];
+			}
+		}
+
+		if (BytePerPixelC[k] > 0) {
+			DCFClkDeepSleepPerSurface[k] = dml_max(__DML_MIN_DCFCLK_FACTOR__ * SwathWidthY[k] *
+					BytePerPixelY[k] / 32.0 / DisplayPipeLineDeliveryTimeLuma,
+					__DML_MIN_DCFCLK_FACTOR__ * SwathWidthC[k] * BytePerPixelC[k] /
+					32.0 / DisplayPipeLineDeliveryTimeChroma);
+		} else {
+			DCFClkDeepSleepPerSurface[k] = __DML_MIN_DCFCLK_FACTOR__ * SwathWidthY[k] * BytePerPixelY[k] /
+					64.0 / DisplayPipeLineDeliveryTimeLuma;
+		}
+		DCFClkDeepSleepPerSurface[k] = dml_max(DCFClkDeepSleepPerSurface[k], PixelClock[k] / 16);
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%d, PixelClock = %f\n", __func__, k, PixelClock[k]);
+		dml_print("DML::%s: k=%d, DCFClkDeepSleepPerSurface = %f\n", __func__, k, DCFClkDeepSleepPerSurface[k]);
+#endif
+	}
+
+	for (k = 0; k < NumberOfActiveSurfaces; ++k)
+		ReadBandwidth = ReadBandwidth + ReadBandwidthLuma[k] + ReadBandwidthChroma[k];
+
+	*DCFClkDeepSleep = dml_max(8.0, __DML_MIN_DCFCLK_FACTOR__ * ReadBandwidth / (double) ReturnBusWidth);
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: __DML_MIN_DCFCLK_FACTOR__ = %f\n", __func__, __DML_MIN_DCFCLK_FACTOR__);
+	dml_print("DML::%s: ReadBandwidth = %f\n", __func__, ReadBandwidth);
+	dml_print("DML::%s: ReturnBusWidth = %d\n", __func__, ReturnBusWidth);
+	dml_print("DML::%s: DCFClkDeepSleep = %f\n", __func__, *DCFClkDeepSleep);
+#endif
+
+	for (k = 0; k < NumberOfActiveSurfaces; ++k)
+		*DCFClkDeepSleep = dml_max(*DCFClkDeepSleep, DCFClkDeepSleepPerSurface[k]);
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: DCFClkDeepSleep = %f (final)\n", __func__, *DCFClkDeepSleep);
+#endif
+} // CalculateDCFCLKDeepSleep
+
+double dml32_CalculateWriteBackDelay(
+		enum source_format_class WritebackPixelFormat,
+		double WritebackHRatio,
+		double WritebackVRatio,
+		unsigned int WritebackVTaps,
+		unsigned int         WritebackDestinationWidth,
+		unsigned int         WritebackDestinationHeight,
+		unsigned int         WritebackSourceHeight,
+		unsigned int HTotal)
+{
+	double CalculateWriteBackDelay;
+	double Line_length;
+	double Output_lines_last_notclamped;
+	double WritebackVInit;
+
+	WritebackVInit = (WritebackVRatio + WritebackVTaps + 1) / 2;
+	Line_length = dml_max((double) WritebackDestinationWidth,
+			dml_ceil((double)WritebackDestinationWidth / 6.0, 1.0) * WritebackVTaps);
+	Output_lines_last_notclamped = WritebackDestinationHeight - 1 -
+			dml_ceil(((double)WritebackSourceHeight -
+					(double) WritebackVInit) / (double)WritebackVRatio, 1.0);
+	if (Output_lines_last_notclamped < 0) {
+		CalculateWriteBackDelay = 0;
+	} else {
+		CalculateWriteBackDelay = Output_lines_last_notclamped * Line_length +
+				(HTotal - WritebackDestinationWidth) + 80;
+	}
+	return CalculateWriteBackDelay;
+}
+
+void dml32_UseMinimumDCFCLK(
+		enum dm_use_mall_for_pstate_change_mode UseMALLForPStateChange[],
+		bool DRRDisplay[],
+		bool SynchronizeDRRDisplaysForUCLKPStateChangeFinal,
+		unsigned int MaxInterDCNTileRepeaters,
+		unsigned int MaxPrefetchMode,
+		double DRAMClockChangeLatencyFinal,
+		double FCLKChangeLatency,
+		double SREnterPlusExitTime,
+		unsigned int ReturnBusWidth,
+		unsigned int RoundTripPingLatencyCycles,
+		unsigned int ReorderingBytes,
+		unsigned int PixelChunkSizeInKByte,
+		unsigned int MetaChunkSize,
+		bool GPUVMEnable,
+		unsigned int GPUVMMaxPageTableLevels,
+		bool HostVMEnable,
+		unsigned int NumberOfActiveSurfaces,
+		double HostVMMinPageSize,
+		unsigned int HostVMMaxNonCachedPageTableLevels,
+		bool DynamicMetadataVMEnabled,
+		bool ImmediateFlipRequirement,
+		bool ProgressiveToInterlaceUnitInOPP,
+		double MaxAveragePercentOfIdealSDPPortBWDisplayCanUseInNormalSystemOperation,
+		double PercentOfIdealSDPPortBWReceivedAfterUrgLatency,
+		unsigned int VTotal[],
+		unsigned int VActive[],
+		unsigned int DynamicMetadataTransmittedBytes[],
+		unsigned int DynamicMetadataLinesBeforeActiveRequired[],
+		bool Interlace[],
+		double RequiredDPPCLKPerSurface[][2][DC__NUM_DPP__MAX],
+		double RequiredDISPCLK[][2],
+		double UrgLatency[],
+		unsigned int NoOfDPP[][2][DC__NUM_DPP__MAX],
+		double ProjectedDCFClkDeepSleep[][2],
+		double MaximumVStartup[][2][DC__NUM_DPP__MAX],
+		unsigned int TotalNumberOfActiveDPP[][2],
+		unsigned int TotalNumberOfDCCActiveDPP[][2],
+		unsigned int dpte_group_bytes[],
+		double PrefetchLinesY[][2][DC__NUM_DPP__MAX],
+		double PrefetchLinesC[][2][DC__NUM_DPP__MAX],
+		unsigned int swath_width_luma_ub_all_states[][2][DC__NUM_DPP__MAX],
+		unsigned int swath_width_chroma_ub_all_states[][2][DC__NUM_DPP__MAX],
+		unsigned int BytePerPixelY[],
+		unsigned int BytePerPixelC[],
+		unsigned int HTotal[],
+		double PixelClock[],
+		double PDEAndMetaPTEBytesPerFrame[][2][DC__NUM_DPP__MAX],
+		double DPTEBytesPerRow[][2][DC__NUM_DPP__MAX],
+		double MetaRowBytes[][2][DC__NUM_DPP__MAX],
+		bool DynamicMetadataEnable[],
+		double ReadBandwidthLuma[],
+		double ReadBandwidthChroma[],
+		double DCFCLKPerState[],
+		/* Output */
+		double DCFCLKState[][2])
+{
+	unsigned int i, j, k;
+	unsigned int     dummy1;
+	double dummy2, dummy3;
+	double   NormalEfficiency;
+	double   TotalMaxPrefetchFlipDPTERowBandwidth[DC__VOLTAGE_STATES][2];
+
+	NormalEfficiency = PercentOfIdealSDPPortBWReceivedAfterUrgLatency / 100.0;
+	for  (i = 0; i < DC__VOLTAGE_STATES; ++i) {
+		for  (j = 0; j <= 1; ++j) {
+			double PixelDCFCLKCyclesRequiredInPrefetch[DC__NUM_DPP__MAX];
+			double PrefetchPixelLinesTime[DC__NUM_DPP__MAX];
+			double DCFCLKRequiredForPeakBandwidthPerSurface[DC__NUM_DPP__MAX];
+			double DynamicMetadataVMExtraLatency[DC__NUM_DPP__MAX];
+			double MinimumTWait = 0.0;
+			double DPTEBandwidth;
+			double DCFCLKRequiredForAverageBandwidth;
+			unsigned int ExtraLatencyBytes;
+			double ExtraLatencyCycles;
+			double DCFCLKRequiredForPeakBandwidth;
+			unsigned int NoOfDPPState[DC__NUM_DPP__MAX];
+			double MinimumTvmPlus2Tr0;
+
+			TotalMaxPrefetchFlipDPTERowBandwidth[i][j] = 0;
+			for (k = 0; k < NumberOfActiveSurfaces; ++k) {
+				TotalMaxPrefetchFlipDPTERowBandwidth[i][j] = TotalMaxPrefetchFlipDPTERowBandwidth[i][j]
+						+ NoOfDPP[i][j][k] * DPTEBytesPerRow[i][j][k]
+								/ (15.75 * HTotal[k] / PixelClock[k]);
+			}
+
+			for (k = 0; k <= NumberOfActiveSurfaces - 1; ++k)
+				NoOfDPPState[k] = NoOfDPP[i][j][k];
+
+			DPTEBandwidth = TotalMaxPrefetchFlipDPTERowBandwidth[i][j];
+			DCFCLKRequiredForAverageBandwidth = dml_max(ProjectedDCFClkDeepSleep[i][j], DPTEBandwidth / NormalEfficiency / ReturnBusWidth);
+
+			ExtraLatencyBytes = dml32_CalculateExtraLatencyBytes(ReorderingBytes,
+					TotalNumberOfActiveDPP[i][j], PixelChunkSizeInKByte,
+					TotalNumberOfDCCActiveDPP[i][j], MetaChunkSize, GPUVMEnable, HostVMEnable,
+					NumberOfActiveSurfaces, NoOfDPPState, dpte_group_bytes, 1, HostVMMinPageSize,
+					HostVMMaxNonCachedPageTableLevels);
+			ExtraLatencyCycles = RoundTripPingLatencyCycles + __DML_ARB_TO_RET_DELAY__
+					+ ExtraLatencyBytes / NormalEfficiency / ReturnBusWidth;
+			for (k = 0; k < NumberOfActiveSurfaces; ++k) {
+				double DCFCLKCyclesRequiredInPrefetch;
+				double PrefetchTime;
+
+				PixelDCFCLKCyclesRequiredInPrefetch[k] = (PrefetchLinesY[i][j][k]
+						* swath_width_luma_ub_all_states[i][j][k] * BytePerPixelY[k]
+						+ PrefetchLinesC[i][j][k] * swath_width_chroma_ub_all_states[i][j][k]
+								* BytePerPixelC[k]) / NormalEfficiency
+						/ ReturnBusWidth;
+				DCFCLKCyclesRequiredInPrefetch = 2 * ExtraLatencyCycles / NoOfDPPState[k]
+						+ PDEAndMetaPTEBytesPerFrame[i][j][k] / NormalEfficiency
+								/ NormalEfficiency / ReturnBusWidth
+								* (GPUVMMaxPageTableLevels > 2 ? 1 : 0)
+						+ 2 * DPTEBytesPerRow[i][j][k] / NormalEfficiency / NormalEfficiency
+								/ ReturnBusWidth
+						+ 2 * MetaRowBytes[i][j][k] / NormalEfficiency / ReturnBusWidth
+						+ PixelDCFCLKCyclesRequiredInPrefetch[k];
+				PrefetchPixelLinesTime[k] = dml_max(PrefetchLinesY[i][j][k], PrefetchLinesC[i][j][k])
+						* HTotal[k] / PixelClock[k];
+				DynamicMetadataVMExtraLatency[k] = (GPUVMEnable == true &&
+						DynamicMetadataEnable[k] == true && DynamicMetadataVMEnabled == true) ?
+						UrgLatency[i] * GPUVMMaxPageTableLevels *
+						(HostVMEnable == true ? HostVMMaxNonCachedPageTableLevels + 1 : 1) : 0;
+
+				MinimumTWait = dml32_CalculateTWait(MaxPrefetchMode,
+						UseMALLForPStateChange[k],
+						SynchronizeDRRDisplaysForUCLKPStateChangeFinal,
+						DRRDisplay[k],
+						DRAMClockChangeLatencyFinal,
+						FCLKChangeLatency,
+						UrgLatency[i],
+						SREnterPlusExitTime);
+
+				PrefetchTime = (MaximumVStartup[i][j][k] - 1) * HTotal[k] / PixelClock[k] -
+						MinimumTWait - UrgLatency[i] *
+						((GPUVMMaxPageTableLevels <= 2 ? GPUVMMaxPageTableLevels :
+						GPUVMMaxPageTableLevels - 2) *  (HostVMEnable == true ?
+						HostVMMaxNonCachedPageTableLevels + 1 : 1) - 1) -
+						DynamicMetadataVMExtraLatency[k];
+
+				if (PrefetchTime > 0) {
+					double ExpectedVRatioPrefetch;
+
+					ExpectedVRatioPrefetch = PrefetchPixelLinesTime[k] / (PrefetchTime *
+							PixelDCFCLKCyclesRequiredInPrefetch[k] /
+							DCFCLKCyclesRequiredInPrefetch);
+					DCFCLKRequiredForPeakBandwidthPerSurface[k] = NoOfDPPState[k] *
+							PixelDCFCLKCyclesRequiredInPrefetch[k] /
+							PrefetchPixelLinesTime[k] *
+							dml_max(1.0, ExpectedVRatioPrefetch) *
+							dml_max(1.0, ExpectedVRatioPrefetch / 4);
+					if (HostVMEnable == true || ImmediateFlipRequirement == true) {
+						DCFCLKRequiredForPeakBandwidthPerSurface[k] =
+								DCFCLKRequiredForPeakBandwidthPerSurface[k] +
+								NoOfDPPState[k] * DPTEBandwidth / NormalEfficiency /
+								NormalEfficiency / ReturnBusWidth;
+					}
+				} else {
+					DCFCLKRequiredForPeakBandwidthPerSurface[k] = DCFCLKPerState[i];
+				}
+				if (DynamicMetadataEnable[k] == true) {
+					double TSetupPipe;
+					double TdmbfPipe;
+					double TdmsksPipe;
+					double TdmecPipe;
+					double AllowedTimeForUrgentExtraLatency;
+
+					dml32_CalculateVUpdateAndDynamicMetadataParameters(
+							MaxInterDCNTileRepeaters,
+							RequiredDPPCLKPerSurface[i][j][k],
+							RequiredDISPCLK[i][j],
+							ProjectedDCFClkDeepSleep[i][j],
+							PixelClock[k],
+							HTotal[k],
+							VTotal[k] - VActive[k],
+							DynamicMetadataTransmittedBytes[k],
+							DynamicMetadataLinesBeforeActiveRequired[k],
+							Interlace[k],
+							ProgressiveToInterlaceUnitInOPP,
+
+							/* output */
+							&TSetupPipe,
+							&TdmbfPipe,
+							&TdmecPipe,
+							&TdmsksPipe,
+							&dummy1,
+							&dummy2,
+							&dummy3);
+					AllowedTimeForUrgentExtraLatency = MaximumVStartup[i][j][k] * HTotal[k] /
+							PixelClock[k] - MinimumTWait - TSetupPipe - TdmbfPipe -
+							TdmecPipe - TdmsksPipe - DynamicMetadataVMExtraLatency[k];
+					if (AllowedTimeForUrgentExtraLatency > 0)
+						DCFCLKRequiredForPeakBandwidthPerSurface[k] =
+								dml_max(DCFCLKRequiredForPeakBandwidthPerSurface[k],
+								ExtraLatencyCycles / AllowedTimeForUrgentExtraLatency);
+					else
+						DCFCLKRequiredForPeakBandwidthPerSurface[k] = DCFCLKPerState[i];
+				}
+			}
+			DCFCLKRequiredForPeakBandwidth = 0;
+			for (k = 0; k <= NumberOfActiveSurfaces - 1; ++k) {
+				DCFCLKRequiredForPeakBandwidth = DCFCLKRequiredForPeakBandwidth +
+						DCFCLKRequiredForPeakBandwidthPerSurface[k];
+			}
+			MinimumTvmPlus2Tr0 = UrgLatency[i] * (GPUVMEnable == true ?
+					(HostVMEnable == true ? (GPUVMMaxPageTableLevels + 2) *
+					(HostVMMaxNonCachedPageTableLevels + 1) - 1 : GPUVMMaxPageTableLevels + 1) : 0);
+			for (k = 0; k < NumberOfActiveSurfaces; ++k) {
+				double MaximumTvmPlus2Tr0PlusTsw;
+
+				MaximumTvmPlus2Tr0PlusTsw = (MaximumVStartup[i][j][k] - 2) * HTotal[k] /
+						PixelClock[k] - MinimumTWait - DynamicMetadataVMExtraLatency[k];
+				if (MaximumTvmPlus2Tr0PlusTsw <= MinimumTvmPlus2Tr0 + PrefetchPixelLinesTime[k] / 4) {
+					DCFCLKRequiredForPeakBandwidth = DCFCLKPerState[i];
+				} else {
+					DCFCLKRequiredForPeakBandwidth = dml_max3(DCFCLKRequiredForPeakBandwidth,
+							2 * ExtraLatencyCycles / (MaximumTvmPlus2Tr0PlusTsw -
+								MinimumTvmPlus2Tr0 -
+								PrefetchPixelLinesTime[k] / 4),
+							(2 * ExtraLatencyCycles +
+								PixelDCFCLKCyclesRequiredInPrefetch[k]) /
+								(MaximumTvmPlus2Tr0PlusTsw - MinimumTvmPlus2Tr0));
+				}
+			}
+			DCFCLKState[i][j] = dml_min(DCFCLKPerState[i], 1.05 *
+					dml_max(DCFCLKRequiredForAverageBandwidth, DCFCLKRequiredForPeakBandwidth));
+		}
+	}
+}
+
+unsigned int dml32_CalculateExtraLatencyBytes(unsigned int ReorderingBytes,
+		unsigned int TotalNumberOfActiveDPP,
+		unsigned int PixelChunkSizeInKByte,
+		unsigned int TotalNumberOfDCCActiveDPP,
+		unsigned int MetaChunkSize,
+		bool GPUVMEnable,
+		bool HostVMEnable,
+		unsigned int NumberOfActiveSurfaces,
+		unsigned int NumberOfDPP[],
+		unsigned int dpte_group_bytes[],
+		double HostVMInefficiencyFactor,
+		double HostVMMinPageSize,
+		unsigned int HostVMMaxNonCachedPageTableLevels)
+{
+	unsigned int k;
+	double   ret;
+	unsigned int  HostVMDynamicLevels;
+
+	if (GPUVMEnable == true && HostVMEnable == true) {
+		if (HostVMMinPageSize < 2048)
+			HostVMDynamicLevels = HostVMMaxNonCachedPageTableLevels;
+		else if (HostVMMinPageSize >= 2048 && HostVMMinPageSize < 1048576)
+			HostVMDynamicLevels = dml_max(0, (int) HostVMMaxNonCachedPageTableLevels - 1);
+		else
+			HostVMDynamicLevels = dml_max(0, (int) HostVMMaxNonCachedPageTableLevels - 2);
+	} else {
+		HostVMDynamicLevels = 0;
+	}
+
+	ret = ReorderingBytes + (TotalNumberOfActiveDPP * PixelChunkSizeInKByte +
+			TotalNumberOfDCCActiveDPP * MetaChunkSize) * 1024.0;
+
+	if (GPUVMEnable == true) {
+		for (k = 0; k < NumberOfActiveSurfaces; ++k) {
+			ret = ret + NumberOfDPP[k] * dpte_group_bytes[k] *
+					(1 + 8 * HostVMDynamicLevels) * HostVMInefficiencyFactor;
+		}
+	}
+	return ret;
+}
+
+void dml32_CalculateVUpdateAndDynamicMetadataParameters(
+		unsigned int MaxInterDCNTileRepeaters,
+		double Dppclk,
+		double Dispclk,
+		double DCFClkDeepSleep,
+		double PixelClock,
+		unsigned int HTotal,
+		unsigned int VBlank,
+		unsigned int DynamicMetadataTransmittedBytes,
+		unsigned int DynamicMetadataLinesBeforeActiveRequired,
+		unsigned int InterlaceEnable,
+		bool ProgressiveToInterlaceUnitInOPP,
+
+		/* output */
+		double *TSetup,
+		double *Tdmbf,
+		double *Tdmec,
+		double *Tdmsks,
+		unsigned int *VUpdateOffsetPix,
+		double *VUpdateWidthPix,
+		double *VReadyOffsetPix)
+{
+	double TotalRepeaterDelayTime;
+
+	TotalRepeaterDelayTime = MaxInterDCNTileRepeaters * (2 / Dppclk + 3 / Dispclk);
+	*VUpdateWidthPix  =
+			dml_ceil((14.0 / DCFClkDeepSleep + 12.0 / Dppclk + TotalRepeaterDelayTime) * PixelClock, 1.0);
+	*VReadyOffsetPix  = dml_ceil(dml_max(150.0 / Dppclk,
+			TotalRepeaterDelayTime + 20.0 / DCFClkDeepSleep + 10.0 / Dppclk) * PixelClock, 1.0);
+	*VUpdateOffsetPix = dml_ceil(HTotal / 4.0, 1.0);
+	*TSetup = (*VUpdateOffsetPix + *VUpdateWidthPix + *VReadyOffsetPix) / PixelClock;
+	*Tdmbf = DynamicMetadataTransmittedBytes / 4.0 / Dispclk;
+	*Tdmec = HTotal / PixelClock;
+
+	if (DynamicMetadataLinesBeforeActiveRequired == 0)
+		*Tdmsks = VBlank * HTotal / PixelClock / 2.0;
+	else
+		*Tdmsks = DynamicMetadataLinesBeforeActiveRequired * HTotal / PixelClock;
+
+	if (InterlaceEnable == 1 && ProgressiveToInterlaceUnitInOPP == false)
+		*Tdmsks = *Tdmsks / 2;
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: VUpdateWidthPix = %d\n", __func__, *VUpdateWidthPix);
+	dml_print("DML::%s: VReadyOffsetPix = %d\n", __func__, *VReadyOffsetPix);
+	dml_print("DML::%s: VUpdateOffsetPix = %d\n", __func__, *VUpdateOffsetPix);
+
+	dml_print("DML::%s: DynamicMetadataLinesBeforeActiveRequired = %d\n",
+			__func__, DynamicMetadataLinesBeforeActiveRequired);
+	dml_print("DML::%s: VBlank = %d\n", __func__, VBlank);
+	dml_print("DML::%s: HTotal = %d\n", __func__, HTotal);
+	dml_print("DML::%s: PixelClock = %f\n", __func__, PixelClock);
+	dml_print("DML::%s: Tdmsks = %f\n", __func__, *Tdmsks);
+#endif
+}
+
+double dml32_CalculateTWait(
+		unsigned int PrefetchMode,
+		enum dm_use_mall_for_pstate_change_mode UseMALLForPStateChange,
+		bool SynchronizeDRRDisplaysForUCLKPStateChangeFinal,
+		bool DRRDisplay,
+		double DRAMClockChangeLatency,
+		double FCLKChangeLatency,
+		double UrgentLatency,
+		double SREnterPlusExitTime)
+{
+	double TWait = 0.0;
+
+	if (PrefetchMode == 0 &&
+			!(UseMALLForPStateChange == dm_use_mall_pstate_change_full_frame) &&
+			!(UseMALLForPStateChange == dm_use_mall_pstate_change_sub_viewport) &&
+			!(UseMALLForPStateChange == dm_use_mall_pstate_change_phantom_pipe) &&
+			!(SynchronizeDRRDisplaysForUCLKPStateChangeFinal && DRRDisplay)) {
+		TWait = dml_max3(DRAMClockChangeLatency + UrgentLatency, SREnterPlusExitTime, UrgentLatency);
+	} else if (PrefetchMode <= 1 && !(UseMALLForPStateChange == dm_use_mall_pstate_change_phantom_pipe)) {
+		TWait = dml_max3(FCLKChangeLatency + UrgentLatency, SREnterPlusExitTime, UrgentLatency);
+	} else if (PrefetchMode <= 2 && !(UseMALLForPStateChange == dm_use_mall_pstate_change_phantom_pipe)) {
+		TWait = dml_max(SREnterPlusExitTime, UrgentLatency);
+	} else {
+		TWait = UrgentLatency;
+	}
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: PrefetchMode = %d\n", __func__, PrefetchMode);
+	dml_print("DML::%s: TWait = %f\n", __func__, TWait);
+#endif
+	return TWait;
+} // CalculateTWait
+
+// Function: get_return_bw_mbps
+// Megabyte per second
+double dml32_get_return_bw_mbps(const soc_bounding_box_st *soc,
+		const int VoltageLevel,
+		const bool HostVMEnable,
+		const double DCFCLK,
+		const double FabricClock,
+		const double DRAMSpeed)
+{
+	double ReturnBW = 0.;
+	double IdealSDPPortBandwidth    = soc->return_bus_width_bytes /*mode_lib->vba.ReturnBusWidth*/ * DCFCLK;
+	double IdealFabricBandwidth     = FabricClock * soc->fabric_datapath_to_dcn_data_return_bytes;
+	double IdealDRAMBandwidth       = DRAMSpeed * soc->num_chans * soc->dram_channel_width_bytes;
+	double PixelDataOnlyReturnBW    = dml_min3(IdealSDPPortBandwidth * soc->pct_ideal_sdp_bw_after_urgent / 100,
+			IdealFabricBandwidth * soc->pct_ideal_fabric_bw_after_urgent / 100,
+			IdealDRAMBandwidth * (VoltageLevel < 2 ? soc->pct_ideal_dram_bw_after_urgent_strobe  :
+					soc->pct_ideal_dram_sdp_bw_after_urgent_pixel_only) / 100);
+	double PixelMixedWithVMDataReturnBW = dml_min3(IdealSDPPortBandwidth * soc->pct_ideal_sdp_bw_after_urgent / 100,
+			IdealFabricBandwidth * soc->pct_ideal_fabric_bw_after_urgent / 100,
+			IdealDRAMBandwidth * (VoltageLevel < 2 ? soc->pct_ideal_dram_bw_after_urgent_strobe :
+					soc->pct_ideal_dram_sdp_bw_after_urgent_pixel_only) / 100);
+
+	if (HostVMEnable != true)
+		ReturnBW = PixelDataOnlyReturnBW;
+	else
+		ReturnBW = PixelMixedWithVMDataReturnBW;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: VoltageLevel = %d\n", __func__, VoltageLevel);
+	dml_print("DML::%s: HostVMEnable = %d\n", __func__, HostVMEnable);
+	dml_print("DML::%s: DCFCLK       = %f\n", __func__, DCFCLK);
+	dml_print("DML::%s: FabricClock  = %f\n", __func__, FabricClock);
+	dml_print("DML::%s: DRAMSpeed    = %f\n", __func__, DRAMSpeed);
+	dml_print("DML::%s: IdealSDPPortBandwidth        = %f\n", __func__, IdealSDPPortBandwidth);
+	dml_print("DML::%s: IdealFabricBandwidth         = %f\n", __func__, IdealFabricBandwidth);
+	dml_print("DML::%s: IdealDRAMBandwidth           = %f\n", __func__, IdealDRAMBandwidth);
+	dml_print("DML::%s: PixelDataOnlyReturnBW        = %f\n", __func__, PixelDataOnlyReturnBW);
+	dml_print("DML::%s: PixelMixedWithVMDataReturnBW = %f\n", __func__, PixelMixedWithVMDataReturnBW);
+	dml_print("DML::%s: ReturnBW                     = %f MBps\n", __func__, ReturnBW);
+#endif
+	return ReturnBW;
+}
+
+// Function: get_return_bw_mbps_vm_only
+// Megabyte per second
+double dml32_get_return_bw_mbps_vm_only(const soc_bounding_box_st *soc,
+		const int VoltageLevel,
+		const double DCFCLK,
+		const double FabricClock,
+		const double DRAMSpeed)
+{
+	double VMDataOnlyReturnBW = dml_min3(
+			soc->return_bus_width_bytes * DCFCLK * soc->pct_ideal_sdp_bw_after_urgent / 100.0,
+			FabricClock * soc->fabric_datapath_to_dcn_data_return_bytes
+					* soc->pct_ideal_sdp_bw_after_urgent / 100.0,
+			DRAMSpeed * soc->num_chans * soc->dram_channel_width_bytes
+					* (VoltageLevel < 2 ?
+							soc->pct_ideal_dram_bw_after_urgent_strobe :
+							soc->pct_ideal_dram_sdp_bw_after_urgent_vm_only) / 100.0);
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: VoltageLevel = %d\n", __func__, VoltageLevel);
+	dml_print("DML::%s: DCFCLK       = %f\n", __func__, DCFCLK);
+	dml_print("DML::%s: FabricClock  = %f\n", __func__, FabricClock);
+	dml_print("DML::%s: DRAMSpeed    = %f\n", __func__, DRAMSpeed);
+	dml_print("DML::%s: VMDataOnlyReturnBW = %f\n", __func__, VMDataOnlyReturnBW);
+#endif
+	return VMDataOnlyReturnBW;
+}
+
+double dml32_CalculateExtraLatency(
+		unsigned int RoundTripPingLatencyCycles,
+		unsigned int ReorderingBytes,
+		double DCFCLK,
+		unsigned int TotalNumberOfActiveDPP,
+		unsigned int PixelChunkSizeInKByte,
+		unsigned int TotalNumberOfDCCActiveDPP,
+		unsigned int MetaChunkSize,
+		double ReturnBW,
+		bool GPUVMEnable,
+		bool HostVMEnable,
+		unsigned int NumberOfActiveSurfaces,
+		unsigned int NumberOfDPP[],
+		unsigned int dpte_group_bytes[],
+		double HostVMInefficiencyFactor,
+		double HostVMMinPageSize,
+		unsigned int HostVMMaxNonCachedPageTableLevels)
+{
+	double ExtraLatencyBytes;
+	double ExtraLatency;
+
+	ExtraLatencyBytes = dml32_CalculateExtraLatencyBytes(
+			ReorderingBytes,
+			TotalNumberOfActiveDPP,
+			PixelChunkSizeInKByte,
+			TotalNumberOfDCCActiveDPP,
+			MetaChunkSize,
+			GPUVMEnable,
+			HostVMEnable,
+			NumberOfActiveSurfaces,
+			NumberOfDPP,
+			dpte_group_bytes,
+			HostVMInefficiencyFactor,
+			HostVMMinPageSize,
+			HostVMMaxNonCachedPageTableLevels);
+
+	ExtraLatency = (RoundTripPingLatencyCycles + __DML_ARB_TO_RET_DELAY__) / DCFCLK + ExtraLatencyBytes / ReturnBW;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: RoundTripPingLatencyCycles=%d\n", __func__, RoundTripPingLatencyCycles);
+	dml_print("DML::%s: DCFCLK=%f\n", __func__, DCFCLK);
+	dml_print("DML::%s: ExtraLatencyBytes=%f\n", __func__, ExtraLatencyBytes);
+	dml_print("DML::%s: ReturnBW=%f\n", __func__, ReturnBW);
+	dml_print("DML::%s: ExtraLatency=%f\n", __func__, ExtraLatency);
+#endif
+
+	return ExtraLatency;
+} // CalculateExtraLatency
+
+bool dml32_CalculatePrefetchSchedule(
+		struct vba_vars_st *v,
+		unsigned int k,
+		double HostVMInefficiencyFactor,
+		DmlPipe *myPipe,
+		unsigned int DSCDelay,
+		unsigned int DPP_RECOUT_WIDTH,
+		unsigned int VStartup,
+		unsigned int MaxVStartup,
+		double UrgentLatency,
+		double UrgentExtraLatency,
+		double TCalc,
+		unsigned int PDEAndMetaPTEBytesFrame,
+		unsigned int MetaRowByte,
+		unsigned int PixelPTEBytesPerRow,
+		double PrefetchSourceLinesY,
+		unsigned int SwathWidthY,
+		unsigned int VInitPreFillY,
+		unsigned int MaxNumSwathY,
+		double PrefetchSourceLinesC,
+		unsigned int SwathWidthC,
+		unsigned int VInitPreFillC,
+		unsigned int MaxNumSwathC,
+		unsigned int swath_width_luma_ub,
+		unsigned int swath_width_chroma_ub,
+		unsigned int SwathHeightY,
+		unsigned int SwathHeightC,
+		double TWait,
+		double TPreReq,
+		bool ExtendPrefetchIfPossible,
+		/* Output */
+		double   *DSTXAfterScaler,
+		double   *DSTYAfterScaler,
+		double *DestinationLinesForPrefetch,
+		double *PrefetchBandwidth,
+		double *DestinationLinesToRequestVMInVBlank,
+		double *DestinationLinesToRequestRowInVBlank,
+		double *VRatioPrefetchY,
+		double *VRatioPrefetchC,
+		double *RequiredPrefetchPixDataBWLuma,
+		double *RequiredPrefetchPixDataBWChroma,
+		bool   *NotEnoughTimeForDynamicMetadata,
+		double *Tno_bw,
+		double *prefetch_vmrow_bw,
+		double *Tdmdl_vm,
+		double *Tdmdl,
+		double *TSetup,
+		unsigned int   *VUpdateOffsetPix,
+		double   *VUpdateWidthPix,
+		double   *VReadyOffsetPix)
+{
+	double DPPCLKDelaySubtotalPlusCNVCFormater = v->DPPCLKDelaySubtotal + v->DPPCLKDelayCNVCFormater;
+	bool MyError = false;
+	unsigned int DPPCycles, DISPCLKCycles;
+	double DSTTotalPixelsAfterScaler;
+	double LineTime;
+	double dst_y_prefetch_equ;
+	double prefetch_bw_oto;
+	double Tvm_oto;
+	double Tr0_oto;
+	double Tvm_oto_lines;
+	double Tr0_oto_lines;
+	double dst_y_prefetch_oto;
+	double TimeForFetchingMetaPTE = 0;
+	double TimeForFetchingRowInVBlank = 0;
+	double LinesToRequestPrefetchPixelData = 0;
+	double LinesForPrefetchBandwidth = 0;
+	unsigned int HostVMDynamicLevelsTrips;
+	double  trip_to_mem;
+	double  Tvm_trips;
+	double  Tr0_trips;
+	double  Tvm_trips_rounded;
+	double  Tr0_trips_rounded;
+	double  Lsw_oto;
+	double  Tpre_rounded;
+	double  prefetch_bw_equ;
+	double  Tvm_equ;
+	double  Tr0_equ;
+	double  Tdmbf;
+	double  Tdmec;
+	double  Tdmsks;
+	double  prefetch_sw_bytes;
+	double  bytes_pp;
+	double  dep_bytes;
+	unsigned int max_vratio_pre = v->MaxVRatioPre;
+	double  min_Lsw;
+	double  Tsw_est1 = 0;
+	double  Tsw_est3 = 0;
+
+	if (v->GPUVMEnable == true && v->HostVMEnable == true)
+		HostVMDynamicLevelsTrips = v->HostVMMaxNonCachedPageTableLevels;
+	else
+		HostVMDynamicLevelsTrips = 0;
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: v->GPUVMEnable = %d\n", __func__, v->GPUVMEnable);
+	dml_print("DML::%s: v->GPUVMMaxPageTableLevels = %d\n", __func__, v->GPUVMMaxPageTableLevels);
+	dml_print("DML::%s: DCCEnable = %d\n", __func__, myPipe->DCCEnable);
+	dml_print("DML::%s: v->HostVMEnable=%d HostVMInefficiencyFactor=%f\n",
+			__func__, v->HostVMEnable, HostVMInefficiencyFactor);
+#endif
+	dml32_CalculateVUpdateAndDynamicMetadataParameters(
+			v->MaxInterDCNTileRepeaters,
+			myPipe->Dppclk,
+			myPipe->Dispclk,
+			myPipe->DCFClkDeepSleep,
+			myPipe->PixelClock,
+			myPipe->HTotal,
+			myPipe->VBlank,
+			v->DynamicMetadataTransmittedBytes[k],
+			v->DynamicMetadataLinesBeforeActiveRequired[k],
+			myPipe->InterlaceEnable,
+			myPipe->ProgressiveToInterlaceUnitInOPP,
+			TSetup,
+
+			/* output */
+			&Tdmbf,
+			&Tdmec,
+			&Tdmsks,
+			VUpdateOffsetPix,
+			VUpdateWidthPix,
+			VReadyOffsetPix);
+
+	LineTime = myPipe->HTotal / myPipe->PixelClock;
+	trip_to_mem = UrgentLatency;
+	Tvm_trips = UrgentExtraLatency + trip_to_mem * (v->GPUVMMaxPageTableLevels * (HostVMDynamicLevelsTrips + 1) - 1);
+
+	if (v->DynamicMetadataVMEnabled == true)
+		*Tdmdl = TWait + Tvm_trips + trip_to_mem;
+	else
+		*Tdmdl = TWait + UrgentExtraLatency;
+
+#ifdef __DML_VBA_ALLOW_DELTA__
+	if (v->DynamicMetadataEnable[k] == false)
+		*Tdmdl = 0.0;
+#endif
+
+	if (v->DynamicMetadataEnable[k] == true) {
+		if (VStartup * LineTime < *TSetup + *Tdmdl + Tdmbf + Tdmec + Tdmsks) {
+			*NotEnoughTimeForDynamicMetadata = true;
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: Not Enough Time for Dynamic Meta!\n", __func__);
+			dml_print("DML::%s: Tdmbf: %fus - time for dmd transfer from dchub to dio output buffer\n",
+					__func__, Tdmbf);
+			dml_print("DML::%s: Tdmec: %fus - time dio takes to transfer dmd\n", __func__, Tdmec);
+			dml_print("DML::%s: Tdmsks: %fus - time before active dmd must complete transmission at dio\n",
+					__func__, Tdmsks);
+			dml_print("DML::%s: Tdmdl: %fus - time for fabric to become ready and fetch dmd\n",
+					__func__, *Tdmdl);
+#endif
+		} else {
+			*NotEnoughTimeForDynamicMetadata = false;
+		}
+	} else {
+		*NotEnoughTimeForDynamicMetadata = false;
+	}
+
+	*Tdmdl_vm =  (v->DynamicMetadataEnable[k] == true && v->DynamicMetadataVMEnabled == true &&
+			v->GPUVMEnable == true ? TWait + Tvm_trips : 0);
+
+	if (myPipe->ScalerEnabled)
+		DPPCycles = DPPCLKDelaySubtotalPlusCNVCFormater + v->DPPCLKDelaySCL;
+	else
+		DPPCycles = DPPCLKDelaySubtotalPlusCNVCFormater + v->DPPCLKDelaySCLLBOnly;
+
+	DPPCycles = DPPCycles + myPipe->NumberOfCursors * v->DPPCLKDelayCNVCCursor;
+
+	DISPCLKCycles = v->DISPCLKDelaySubtotal;
+
+	if (myPipe->Dppclk == 0.0 || myPipe->Dispclk == 0.0)
+		return true;
+
+	*DSTXAfterScaler = DPPCycles * myPipe->PixelClock / myPipe->Dppclk + DISPCLKCycles *
+			myPipe->PixelClock / myPipe->Dispclk + DSCDelay;
+
+	*DSTXAfterScaler = *DSTXAfterScaler + (myPipe->ODMMode != dm_odm_combine_mode_disabled ? 18 : 0)
+			+ (myPipe->DPPPerSurface - 1) * DPP_RECOUT_WIDTH
+			+ ((myPipe->ODMMode == dm_odm_split_mode_1to2 || myPipe->ODMMode == dm_odm_mode_mso_1to2) ?
+					myPipe->HActive / 2 : 0)
+			+ ((myPipe->ODMMode == dm_odm_mode_mso_1to4) ? myPipe->HActive * 3 / 4 : 0);
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: DPPCycles: %d\n", __func__, DPPCycles);
+	dml_print("DML::%s: PixelClock: %f\n", __func__, myPipe->PixelClock);
+	dml_print("DML::%s: Dppclk: %f\n", __func__, myPipe->Dppclk);
+	dml_print("DML::%s: DISPCLKCycles: %d\n", __func__, DISPCLKCycles);
+	dml_print("DML::%s: DISPCLK: %f\n", __func__,  myPipe->Dispclk);
+	dml_print("DML::%s: DSCDelay: %d\n", __func__,  DSCDelay);
+	dml_print("DML::%s: ODMMode: %d\n", __func__,  myPipe->ODMMode);
+	dml_print("DML::%s: DPP_RECOUT_WIDTH: %d\n", __func__, DPP_RECOUT_WIDTH);
+	dml_print("DML::%s: DSTXAfterScaler: %d\n", __func__,  *DSTXAfterScaler);
+#endif
+
+	if (v->OutputFormat[k] == dm_420 || (myPipe->InterlaceEnable && myPipe->ProgressiveToInterlaceUnitInOPP))
+		*DSTYAfterScaler = 1;
+	else
+		*DSTYAfterScaler = 0;
+
+	DSTTotalPixelsAfterScaler = *DSTYAfterScaler * myPipe->HTotal + *DSTXAfterScaler;
+	*DSTYAfterScaler = dml_floor(DSTTotalPixelsAfterScaler / myPipe->HTotal, 1);
+	*DSTXAfterScaler = DSTTotalPixelsAfterScaler - ((double) (*DSTYAfterScaler * myPipe->HTotal));
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: DSTXAfterScaler: %d (final)\n", __func__,  *DSTXAfterScaler);
+	dml_print("DML::%s: DSTYAfterScaler: %d (final)\n", __func__, *DSTYAfterScaler);
+#endif
+
+	MyError = false;
+
+	Tr0_trips = trip_to_mem * (HostVMDynamicLevelsTrips + 1);
+
+	if (v->GPUVMEnable == true) {
+		Tvm_trips_rounded = dml_ceil(4.0 * Tvm_trips / LineTime, 1.0) / 4.0 * LineTime;
+		Tr0_trips_rounded = dml_ceil(4.0 * Tr0_trips / LineTime, 1.0) / 4.0 * LineTime;
+		if (v->GPUVMMaxPageTableLevels >= 3) {
+			*Tno_bw = UrgentExtraLatency + trip_to_mem *
+					(double) ((v->GPUVMMaxPageTableLevels - 2) * (HostVMDynamicLevelsTrips + 1) - 1);
+		} else if (v->GPUVMMaxPageTableLevels == 1 && myPipe->DCCEnable != true) {
+			Tr0_trips_rounded = dml_ceil(4.0 * UrgentExtraLatency / LineTime, 1.0) /
+					4.0 * LineTime; // VBA_ERROR
+			*Tno_bw = UrgentExtraLatency;
+		} else {
+			*Tno_bw = 0;
+		}
+	} else if (myPipe->DCCEnable == true) {
+		Tvm_trips_rounded = LineTime / 4.0;
+		Tr0_trips_rounded = dml_ceil(4.0 * Tr0_trips / LineTime, 1.0) / 4.0 * LineTime;
+		*Tno_bw = 0;
+	} else {
+		Tvm_trips_rounded = LineTime / 4.0;
+		Tr0_trips_rounded = LineTime / 2.0;
+		*Tno_bw = 0;
+	}
+	Tvm_trips_rounded = dml_max(Tvm_trips_rounded, LineTime / 4.0);
+	Tr0_trips_rounded = dml_max(Tr0_trips_rounded, LineTime / 4.0);
+
+	if (myPipe->SourcePixelFormat == dm_420_8 || myPipe->SourcePixelFormat == dm_420_10
+			|| myPipe->SourcePixelFormat == dm_420_12) {
+		bytes_pp = myPipe->BytePerPixelY + myPipe->BytePerPixelC / 4;
+	} else {
+		bytes_pp = myPipe->BytePerPixelY + myPipe->BytePerPixelC;
+	}
+
+	prefetch_sw_bytes = PrefetchSourceLinesY * swath_width_luma_ub * myPipe->BytePerPixelY
+			+ PrefetchSourceLinesC * swath_width_chroma_ub * myPipe->BytePerPixelC;
+	prefetch_bw_oto = dml_max(bytes_pp * myPipe->PixelClock / myPipe->DPPPerSurface,
+			prefetch_sw_bytes / (dml_max(PrefetchSourceLinesY, PrefetchSourceLinesC) * LineTime));
+
+	min_Lsw = dml_max(PrefetchSourceLinesY, PrefetchSourceLinesC) / max_vratio_pre;
+	min_Lsw = dml_max(min_Lsw, 1.0);
+	Lsw_oto = dml_ceil(4.0 * dml_max(prefetch_sw_bytes / prefetch_bw_oto / LineTime, min_Lsw), 1.0) / 4.0;
+
+	if (v->GPUVMEnable == true) {
+		Tvm_oto = dml_max3(
+				Tvm_trips,
+				*Tno_bw + PDEAndMetaPTEBytesFrame * HostVMInefficiencyFactor / prefetch_bw_oto,
+				LineTime / 4.0);
+	} else
+		Tvm_oto = LineTime / 4.0;
+
+	if ((v->GPUVMEnable == true || myPipe->DCCEnable == true)) {
+		Tr0_oto = dml_max4(
+				Tr0_trips,
+				(MetaRowByte + PixelPTEBytesPerRow * HostVMInefficiencyFactor) / prefetch_bw_oto,
+				(LineTime - Tvm_oto)/2.0,
+				LineTime / 4.0);
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: Tr0_oto max0 = %f\n", __func__,
+				(MetaRowByte + PixelPTEBytesPerRow * HostVMInefficiencyFactor) / prefetch_bw_oto);
+		dml_print("DML::%s: Tr0_oto max1 = %f\n", __func__, Tr0_trips);
+		dml_print("DML::%s: Tr0_oto max2 = %f\n", __func__, LineTime - Tvm_oto);
+		dml_print("DML::%s: Tr0_oto max3 = %f\n", __func__, LineTime / 4);
+#endif
+	} else
+		Tr0_oto = (LineTime - Tvm_oto) / 2.0;
+
+	Tvm_oto_lines = dml_ceil(4.0 * Tvm_oto / LineTime, 1) / 4.0;
+	Tr0_oto_lines = dml_ceil(4.0 * Tr0_oto / LineTime, 1) / 4.0;
+	dst_y_prefetch_oto = Tvm_oto_lines + 2 * Tr0_oto_lines + Lsw_oto;
+
+	dst_y_prefetch_equ = VStartup - (*TSetup + dml_max(TWait + TCalc, *Tdmdl)) / LineTime -
+			(*DSTYAfterScaler + (double) *DSTXAfterScaler / (double) myPipe->HTotal);
+
+	dst_y_prefetch_equ = dml_min(dst_y_prefetch_equ, __DML_VBA_MAX_DST_Y_PRE__);
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: HTotal = %d\n", __func__, myPipe->HTotal);
+	dml_print("DML::%s: min_Lsw = %f\n", __func__, min_Lsw);
+	dml_print("DML::%s: *Tno_bw = %f\n", __func__, *Tno_bw);
+	dml_print("DML::%s: UrgentExtraLatency = %f\n", __func__, UrgentExtraLatency);
+	dml_print("DML::%s: trip_to_mem = %f\n", __func__, trip_to_mem);
+	dml_print("DML::%s: BytePerPixelY = %d\n", __func__, myPipe->BytePerPixelY);
+	dml_print("DML::%s: PrefetchSourceLinesY = %f\n", __func__, PrefetchSourceLinesY);
+	dml_print("DML::%s: swath_width_luma_ub = %d\n", __func__, swath_width_luma_ub);
+	dml_print("DML::%s: BytePerPixelC = %d\n", __func__, myPipe->BytePerPixelC);
+	dml_print("DML::%s: PrefetchSourceLinesC = %f\n", __func__, PrefetchSourceLinesC);
+	dml_print("DML::%s: swath_width_chroma_ub = %d\n", __func__, swath_width_chroma_ub);
+	dml_print("DML::%s: prefetch_sw_bytes = %f\n", __func__, prefetch_sw_bytes);
+	dml_print("DML::%s: bytes_pp = %f\n", __func__, bytes_pp);
+	dml_print("DML::%s: PDEAndMetaPTEBytesFrame = %d\n", __func__, PDEAndMetaPTEBytesFrame);
+	dml_print("DML::%s: MetaRowByte = %d\n", __func__, MetaRowByte);
+	dml_print("DML::%s: PixelPTEBytesPerRow = %d\n", __func__, PixelPTEBytesPerRow);
+	dml_print("DML::%s: HostVMInefficiencyFactor = %f\n", __func__, HostVMInefficiencyFactor);
+	dml_print("DML::%s: Tvm_trips = %f\n", __func__, Tvm_trips);
+	dml_print("DML::%s: Tr0_trips = %f\n", __func__, Tr0_trips);
+	dml_print("DML::%s: prefetch_bw_oto = %f\n", __func__, prefetch_bw_oto);
+	dml_print("DML::%s: Tr0_oto = %f\n", __func__, Tr0_oto);
+	dml_print("DML::%s: Tvm_oto = %f\n", __func__, Tvm_oto);
+	dml_print("DML::%s: Tvm_oto_lines = %f\n", __func__, Tvm_oto_lines);
+	dml_print("DML::%s: Tr0_oto_lines = %f\n", __func__, Tr0_oto_lines);
+	dml_print("DML::%s: Lsw_oto = %f\n", __func__, Lsw_oto);
+	dml_print("DML::%s: dst_y_prefetch_oto = %f\n", __func__, dst_y_prefetch_oto);
+	dml_print("DML::%s: dst_y_prefetch_equ = %f\n", __func__, dst_y_prefetch_equ);
+#endif
+
+	dst_y_prefetch_equ = dml_floor(4.0 * (dst_y_prefetch_equ + 0.125), 1) / 4.0;
+	Tpre_rounded = dst_y_prefetch_equ * LineTime;
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: dst_y_prefetch_equ: %f (after round)\n", __func__, dst_y_prefetch_equ);
+	dml_print("DML::%s: LineTime: %f\n", __func__, LineTime);
+	dml_print("DML::%s: VStartup: %d\n", __func__, VStartup);
+	dml_print("DML::%s: Tvstartup: %fus - time between vstartup and first pixel of active\n",
+			__func__, VStartup * LineTime);
+	dml_print("DML::%s: TSetup: %fus - time from vstartup to vready\n", __func__, *TSetup);
+	dml_print("DML::%s: TCalc: %fus - time for calculations in dchub starting at vready\n", __func__, TCalc);
+	dml_print("DML::%s: Tdmbf: %fus - time for dmd transfer from dchub to dio output buffer\n", __func__, Tdmbf);
+	dml_print("DML::%s: Tdmec: %fus - time dio takes to transfer dmd\n", __func__, Tdmec);
+	dml_print("DML::%s: Tdmdl_vm: %fus - time for vm stages of dmd\n", __func__, *Tdmdl_vm);
+	dml_print("DML::%s: Tdmdl: %fus - time for fabric to become ready and fetch dmd\n", __func__, *Tdmdl);
+	dml_print("DML::%s: DSTYAfterScaler: %d lines - number of lines of pipeline and buffer delay after scaler\n",
+			__func__, *DSTYAfterScaler);
+#endif
+	dep_bytes = dml_max(PDEAndMetaPTEBytesFrame * HostVMInefficiencyFactor,
+			MetaRowByte + PixelPTEBytesPerRow * HostVMInefficiencyFactor);
+
+	if (prefetch_sw_bytes < dep_bytes)
+		prefetch_sw_bytes = 2 * dep_bytes;
+
+	*PrefetchBandwidth = 0;
+	*DestinationLinesToRequestVMInVBlank = 0;
+	*DestinationLinesToRequestRowInVBlank = 0;
+	*VRatioPrefetchY = 0;
+	*VRatioPrefetchC = 0;
+	*RequiredPrefetchPixDataBWLuma = 0;
+	if (dst_y_prefetch_equ > 1 &&
+			(Tpre_rounded >= TPreReq || dst_y_prefetch_equ == __DML_VBA_MAX_DST_Y_PRE__)) {
+		double PrefetchBandwidth1;
+		double PrefetchBandwidth2;
+		double PrefetchBandwidth3;
+		double PrefetchBandwidth4;
+
+		if (Tpre_rounded - *Tno_bw > 0) {
+			PrefetchBandwidth1 = (PDEAndMetaPTEBytesFrame * HostVMInefficiencyFactor + 2 * MetaRowByte
+					+ 2 * PixelPTEBytesPerRow * HostVMInefficiencyFactor
+					+ prefetch_sw_bytes) / (Tpre_rounded - *Tno_bw);
+			Tsw_est1 = prefetch_sw_bytes / PrefetchBandwidth1;
+		} else
+			PrefetchBandwidth1 = 0;
+
+		if (VStartup == MaxVStartup && (Tsw_est1 / LineTime < min_Lsw)
+				&& Tpre_rounded - min_Lsw * LineTime - 0.75 * LineTime - *Tno_bw > 0) {
+			PrefetchBandwidth1 = (PDEAndMetaPTEBytesFrame * HostVMInefficiencyFactor + 2 * MetaRowByte
+					+ 2 * PixelPTEBytesPerRow * HostVMInefficiencyFactor)
+					/ (Tpre_rounded - min_Lsw * LineTime - 0.75 * LineTime - *Tno_bw);
+		}
+
+		if (Tpre_rounded - *Tno_bw - 2 * Tr0_trips_rounded > 0)
+			PrefetchBandwidth2 = (PDEAndMetaPTEBytesFrame * HostVMInefficiencyFactor + prefetch_sw_bytes) /
+			(Tpre_rounded - *Tno_bw - 2 * Tr0_trips_rounded);
+		else
+			PrefetchBandwidth2 = 0;
+
+		if (Tpre_rounded - Tvm_trips_rounded > 0) {
+			PrefetchBandwidth3 = (2 * MetaRowByte + 2 * PixelPTEBytesPerRow * HostVMInefficiencyFactor
+					+ prefetch_sw_bytes) / (Tpre_rounded - Tvm_trips_rounded);
+			Tsw_est3 = prefetch_sw_bytes / PrefetchBandwidth3;
+		} else
+			PrefetchBandwidth3 = 0;
+
+
+		if (VStartup == MaxVStartup &&
+				(Tsw_est3 / LineTime < min_Lsw) && Tpre_rounded - min_Lsw * LineTime - 0.75 *
+				LineTime - Tvm_trips_rounded > 0) {
+			PrefetchBandwidth3 = (2 * MetaRowByte + 2 * PixelPTEBytesPerRow * HostVMInefficiencyFactor)
+					/ (Tpre_rounded - min_Lsw * LineTime - 0.75 * LineTime - Tvm_trips_rounded);
+		}
+
+		if (Tpre_rounded - Tvm_trips_rounded - 2 * Tr0_trips_rounded > 0) {
+			PrefetchBandwidth4 = prefetch_sw_bytes /
+					(Tpre_rounded - Tvm_trips_rounded - 2 * Tr0_trips_rounded);
+		} else {
+			PrefetchBandwidth4 = 0;
+		}
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: Tpre_rounded: %f\n", __func__, Tpre_rounded);
+		dml_print("DML::%s: Tno_bw: %f\n", __func__, *Tno_bw);
+		dml_print("DML::%s: Tvm_trips_rounded: %f\n", __func__, Tvm_trips_rounded);
+		dml_print("DML::%s: Tsw_est1: %f\n", __func__, Tsw_est1);
+		dml_print("DML::%s: Tsw_est3: %f\n", __func__, Tsw_est3);
+		dml_print("DML::%s: PrefetchBandwidth1: %f\n", __func__, PrefetchBandwidth1);
+		dml_print("DML::%s: PrefetchBandwidth2: %f\n", __func__, PrefetchBandwidth2);
+		dml_print("DML::%s: PrefetchBandwidth3: %f\n", __func__, PrefetchBandwidth3);
+		dml_print("DML::%s: PrefetchBandwidth4: %f\n", __func__, PrefetchBandwidth4);
+#endif
+		{
+			bool Case1OK;
+			bool Case2OK;
+			bool Case3OK;
+
+			if (PrefetchBandwidth1 > 0) {
+				if (*Tno_bw + PDEAndMetaPTEBytesFrame * HostVMInefficiencyFactor / PrefetchBandwidth1
+						>= Tvm_trips_rounded
+						&& (MetaRowByte + PixelPTEBytesPerRow * HostVMInefficiencyFactor)
+								/ PrefetchBandwidth1 >= Tr0_trips_rounded) {
+					Case1OK = true;
+				} else {
+					Case1OK = false;
+				}
+			} else {
+				Case1OK = false;
+			}
+
+			if (PrefetchBandwidth2 > 0) {
+				if (*Tno_bw + PDEAndMetaPTEBytesFrame * HostVMInefficiencyFactor / PrefetchBandwidth2
+						>= Tvm_trips_rounded
+						&& (MetaRowByte + PixelPTEBytesPerRow * HostVMInefficiencyFactor)
+						/ PrefetchBandwidth2 < Tr0_trips_rounded) {
+					Case2OK = true;
+				} else {
+					Case2OK = false;
+				}
+			} else {
+				Case2OK = false;
+			}
+
+			if (PrefetchBandwidth3 > 0) {
+				if (*Tno_bw + PDEAndMetaPTEBytesFrame * HostVMInefficiencyFactor / PrefetchBandwidth3 <
+						Tvm_trips_rounded && (MetaRowByte + PixelPTEBytesPerRow *
+								HostVMInefficiencyFactor) / PrefetchBandwidth3 >=
+								Tr0_trips_rounded) {
+					Case3OK = true;
+				} else {
+					Case3OK = false;
+				}
+			} else {
+				Case3OK = false;
+			}
+
+			if (Case1OK)
+				prefetch_bw_equ = PrefetchBandwidth1;
+			else if (Case2OK)
+				prefetch_bw_equ = PrefetchBandwidth2;
+			else if (Case3OK)
+				prefetch_bw_equ = PrefetchBandwidth3;
+			else
+				prefetch_bw_equ = PrefetchBandwidth4;
+
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: Case1OK: %d\n", __func__, Case1OK);
+			dml_print("DML::%s: Case2OK: %d\n", __func__, Case2OK);
+			dml_print("DML::%s: Case3OK: %d\n", __func__, Case3OK);
+			dml_print("DML::%s: prefetch_bw_equ: %f\n", __func__, prefetch_bw_equ);
+#endif
+
+			if (prefetch_bw_equ > 0) {
+				if (v->GPUVMEnable == true) {
+					Tvm_equ = dml_max3(*Tno_bw + PDEAndMetaPTEBytesFrame *
+							HostVMInefficiencyFactor / prefetch_bw_equ,
+							Tvm_trips, LineTime / 4);
+				} else {
+					Tvm_equ = LineTime / 4;
+				}
+
+				if ((v->GPUVMEnable == true || myPipe->DCCEnable == true)) {
+					Tr0_equ = dml_max4((MetaRowByte + PixelPTEBytesPerRow *
+							HostVMInefficiencyFactor) / prefetch_bw_equ, Tr0_trips,
+							(LineTime - Tvm_equ) / 2, LineTime / 4);
+				} else {
+					Tr0_equ = (LineTime - Tvm_equ) / 2;
+				}
+			} else {
+				Tvm_equ = 0;
+				Tr0_equ = 0;
+#ifdef __DML_VBA_DEBUG__
+				dml_print("DML: prefetch_bw_equ equals 0! %s:%d\n", __FILE__, __LINE__);
+#endif
+			}
+		}
+
+		if (dst_y_prefetch_oto < dst_y_prefetch_equ) {
+			if (dst_y_prefetch_oto * LineTime < TPreReq) {
+				*DestinationLinesForPrefetch = dst_y_prefetch_equ;
+			} else {
+				*DestinationLinesForPrefetch = dst_y_prefetch_oto;
+			}
+			TimeForFetchingMetaPTE = Tvm_oto;
+			TimeForFetchingRowInVBlank = Tr0_oto;
+			*PrefetchBandwidth = prefetch_bw_oto;
+			/* Clamp to oto for bandwidth calculation */
+			LinesForPrefetchBandwidth = dst_y_prefetch_oto;
+		} else {
+			/* For mode programming we want to extend the prefetch as much as possible
+			 * (up to oto, or as long as we can for equ) if we're not already applying
+			 * the 60us prefetch requirement. This is to avoid intermittent underflow
+			 * issues during prefetch.
+			 *
+			 * The prefetch extension is applied under the following scenarios:
+			 * 1. We're in prefetch mode > 0 (i.e. we don't support MCLK switch in blank)
+			 * 2. We're using subvp or drr methods of p-state switch, in which case we
+			 *    we don't care if prefetch takes up more of the blanking time
+			 *
+			 * Mode programming typically chooses the smallest prefetch time possible
+			 * (i.e. highest bandwidth during prefetch) presumably to create margin between
+			 * p-states / c-states that happen in vblank and prefetch. Therefore we only
+			 * apply this prefetch extension when p-state in vblank is not required (UCLK
+			 * p-states take up the most vblank time).
+			 */
+			if (ExtendPrefetchIfPossible && TPreReq == 0 && VStartup < MaxVStartup) {
+				MyError = true;
+			} else {
+				*DestinationLinesForPrefetch = dst_y_prefetch_equ;
+				TimeForFetchingMetaPTE = Tvm_equ;
+				TimeForFetchingRowInVBlank = Tr0_equ;
+				*PrefetchBandwidth = prefetch_bw_equ;
+				/* Clamp to equ for bandwidth calculation */
+				LinesForPrefetchBandwidth = dst_y_prefetch_equ;
+			}
+		}
+
+		*DestinationLinesToRequestVMInVBlank = dml_ceil(4.0 * TimeForFetchingMetaPTE / LineTime, 1.0) / 4.0;
+
+		*DestinationLinesToRequestRowInVBlank =
+				dml_ceil(4.0 * TimeForFetchingRowInVBlank / LineTime, 1.0) / 4.0;
+
+		LinesToRequestPrefetchPixelData = LinesForPrefetchBandwidth -
+				*DestinationLinesToRequestVMInVBlank - 2 * *DestinationLinesToRequestRowInVBlank;
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: DestinationLinesForPrefetch = %f\n", __func__, *DestinationLinesForPrefetch);
+		dml_print("DML::%s: DestinationLinesToRequestVMInVBlank = %f\n",
+				__func__, *DestinationLinesToRequestVMInVBlank);
+		dml_print("DML::%s: TimeForFetchingRowInVBlank = %f\n", __func__, TimeForFetchingRowInVBlank);
+		dml_print("DML::%s: LineTime = %f\n", __func__, LineTime);
+		dml_print("DML::%s: DestinationLinesToRequestRowInVBlank = %f\n",
+				__func__, *DestinationLinesToRequestRowInVBlank);
+		dml_print("DML::%s: PrefetchSourceLinesY = %f\n", __func__, PrefetchSourceLinesY);
+		dml_print("DML::%s: LinesToRequestPrefetchPixelData = %f\n", __func__, LinesToRequestPrefetchPixelData);
+#endif
+
+		if (LinesToRequestPrefetchPixelData >= 1 && prefetch_bw_equ > 0) {
+			*VRatioPrefetchY = (double) PrefetchSourceLinesY / LinesToRequestPrefetchPixelData;
+			*VRatioPrefetchY = dml_max(*VRatioPrefetchY, 1.0);
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: VRatioPrefetchY = %f\n", __func__, *VRatioPrefetchY);
+			dml_print("DML::%s: SwathHeightY = %d\n", __func__, SwathHeightY);
+			dml_print("DML::%s: VInitPreFillY = %d\n", __func__, VInitPreFillY);
+#endif
+			if ((SwathHeightY > 4) && (VInitPreFillY > 3)) {
+				if (LinesToRequestPrefetchPixelData > (VInitPreFillY - 3.0) / 2.0) {
+					*VRatioPrefetchY =
+							dml_max((double) PrefetchSourceLinesY /
+									LinesToRequestPrefetchPixelData,
+									(double) MaxNumSwathY * SwathHeightY /
+									(LinesToRequestPrefetchPixelData -
+									(VInitPreFillY - 3.0) / 2.0));
+					*VRatioPrefetchY = dml_max(*VRatioPrefetchY, 1.0);
+				} else {
+					MyError = true;
+					*VRatioPrefetchY = 0;
+				}
+#ifdef __DML_VBA_DEBUG__
+				dml_print("DML::%s: VRatioPrefetchY = %f\n", __func__, *VRatioPrefetchY);
+				dml_print("DML::%s: PrefetchSourceLinesY = %f\n", __func__, PrefetchSourceLinesY);
+				dml_print("DML::%s: MaxNumSwathY = %d\n", __func__, MaxNumSwathY);
+#endif
+			}
+
+			*VRatioPrefetchC = (double) PrefetchSourceLinesC / LinesToRequestPrefetchPixelData;
+			*VRatioPrefetchC = dml_max(*VRatioPrefetchC, 1.0);
+
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: VRatioPrefetchC = %f\n", __func__, *VRatioPrefetchC);
+			dml_print("DML::%s: SwathHeightC = %d\n", __func__, SwathHeightC);
+			dml_print("DML::%s: VInitPreFillC = %d\n", __func__, VInitPreFillC);
+#endif
+			if ((SwathHeightC > 4)) {
+				if (LinesToRequestPrefetchPixelData > (VInitPreFillC - 3.0) / 2.0) {
+					*VRatioPrefetchC =
+						dml_max(*VRatioPrefetchC,
+							(double) MaxNumSwathC * SwathHeightC /
+							(LinesToRequestPrefetchPixelData -
+							(VInitPreFillC - 3.0) / 2.0));
+					*VRatioPrefetchC = dml_max(*VRatioPrefetchC, 1.0);
+				} else {
+					MyError = true;
+					*VRatioPrefetchC = 0;
+				}
+#ifdef __DML_VBA_DEBUG__
+				dml_print("DML::%s: VRatioPrefetchC = %f\n", __func__, *VRatioPrefetchC);
+				dml_print("DML::%s: PrefetchSourceLinesC = %f\n", __func__, PrefetchSourceLinesC);
+				dml_print("DML::%s: MaxNumSwathC = %d\n", __func__, MaxNumSwathC);
+#endif
+			}
+
+			*RequiredPrefetchPixDataBWLuma = (double) PrefetchSourceLinesY
+					/ LinesToRequestPrefetchPixelData * myPipe->BytePerPixelY * swath_width_luma_ub
+					/ LineTime;
+
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: BytePerPixelY = %d\n", __func__, myPipe->BytePerPixelY);
+			dml_print("DML::%s: swath_width_luma_ub = %d\n", __func__, swath_width_luma_ub);
+			dml_print("DML::%s: LineTime = %f\n", __func__, LineTime);
+			dml_print("DML::%s: RequiredPrefetchPixDataBWLuma = %f\n",
+					__func__, *RequiredPrefetchPixDataBWLuma);
+#endif
+			*RequiredPrefetchPixDataBWChroma = (double) PrefetchSourceLinesC /
+					LinesToRequestPrefetchPixelData
+					* myPipe->BytePerPixelC
+					* swath_width_chroma_ub / LineTime;
+		} else {
+			MyError = true;
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML:%s: MyErr set. LinesToRequestPrefetchPixelData: %f, should be > 0\n",
+					__func__, LinesToRequestPrefetchPixelData);
+#endif
+			*VRatioPrefetchY = 0;
+			*VRatioPrefetchC = 0;
+			*RequiredPrefetchPixDataBWLuma = 0;
+			*RequiredPrefetchPixDataBWChroma = 0;
+		}
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML: Tpre: %fus - sum of time to request meta pte, 2 x data pte + meta data, swaths\n",
+			(double)LinesToRequestPrefetchPixelData * LineTime +
+			2.0*TimeForFetchingRowInVBlank + TimeForFetchingMetaPTE);
+		dml_print("DML:  Tvm: %fus - time to fetch page tables for meta surface\n", TimeForFetchingMetaPTE);
+		dml_print("DML: To: %fus - time for propagation from scaler to optc\n",
+			(*DSTYAfterScaler + ((double) (*DSTXAfterScaler) / (double) myPipe->HTotal)) * LineTime);
+		dml_print("DML: Tvstartup - TSetup - Tcalc - Twait - Tpre - To > 0\n");
+		dml_print("DML: Tslack(pre): %fus - time left over in schedule\n", VStartup * LineTime -
+			TimeForFetchingMetaPTE - 2*TimeForFetchingRowInVBlank - (*DSTYAfterScaler +
+			((double) (*DSTXAfterScaler) / (double) myPipe->HTotal)) * LineTime - TWait - TCalc - *TSetup);
+		dml_print("DML: row_bytes = dpte_row_bytes (per_pipe) = PixelPTEBytesPerRow = : %d\n",
+				PixelPTEBytesPerRow);
+#endif
+	} else {
+		MyError = true;
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: MyErr set, dst_y_prefetch_equ = %f (should be > 1)\n",
+				__func__, dst_y_prefetch_equ);
+#endif
+	}
+
+	{
+		double prefetch_vm_bw;
+		double prefetch_row_bw;
+
+		if (PDEAndMetaPTEBytesFrame == 0) {
+			prefetch_vm_bw = 0;
+		} else if (*DestinationLinesToRequestVMInVBlank > 0) {
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: PDEAndMetaPTEBytesFrame = %d\n", __func__, PDEAndMetaPTEBytesFrame);
+			dml_print("DML::%s: HostVMInefficiencyFactor = %f\n", __func__, HostVMInefficiencyFactor);
+			dml_print("DML::%s: DestinationLinesToRequestVMInVBlank = %f\n",
+					__func__, *DestinationLinesToRequestVMInVBlank);
+			dml_print("DML::%s: LineTime = %f\n", __func__, LineTime);
+#endif
+			prefetch_vm_bw = PDEAndMetaPTEBytesFrame * HostVMInefficiencyFactor /
+					(*DestinationLinesToRequestVMInVBlank * LineTime);
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: prefetch_vm_bw = %f\n", __func__, prefetch_vm_bw);
+#endif
+		} else {
+			prefetch_vm_bw = 0;
+			MyError = true;
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: MyErr set. DestinationLinesToRequestVMInVBlank=%f (should be > 0)\n",
+					__func__, *DestinationLinesToRequestVMInVBlank);
+#endif
+		}
+
+		if (MetaRowByte + PixelPTEBytesPerRow == 0) {
+			prefetch_row_bw = 0;
+		} else if (*DestinationLinesToRequestRowInVBlank > 0) {
+			prefetch_row_bw = (MetaRowByte + PixelPTEBytesPerRow * HostVMInefficiencyFactor) /
+					(*DestinationLinesToRequestRowInVBlank * LineTime);
+
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: MetaRowByte = %d\n", __func__, MetaRowByte);
+			dml_print("DML::%s: PixelPTEBytesPerRow = %d\n", __func__, PixelPTEBytesPerRow);
+			dml_print("DML::%s: DestinationLinesToRequestRowInVBlank = %f\n",
+					__func__, *DestinationLinesToRequestRowInVBlank);
+			dml_print("DML::%s: prefetch_row_bw = %f\n", __func__, prefetch_row_bw);
+#endif
+		} else {
+			prefetch_row_bw = 0;
+			MyError = true;
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: MyErr set. DestinationLinesToRequestRowInVBlank=%f (should be > 0)\n",
+					__func__, *DestinationLinesToRequestRowInVBlank);
+#endif
+		}
+
+		*prefetch_vmrow_bw = dml_max(prefetch_vm_bw, prefetch_row_bw);
+	}
+
+	if (MyError) {
+		*PrefetchBandwidth = 0;
+		*DestinationLinesToRequestVMInVBlank = 0;
+		*DestinationLinesToRequestRowInVBlank = 0;
+		*DestinationLinesForPrefetch = 0;
+		*VRatioPrefetchY = 0;
+		*VRatioPrefetchC = 0;
+		*RequiredPrefetchPixDataBWLuma = 0;
+		*RequiredPrefetchPixDataBWChroma = 0;
+	}
+
+	return MyError;
+} // CalculatePrefetchSchedule
+
+void dml32_CalculateFlipSchedule(
+		double HostVMInefficiencyFactor,
+		double UrgentExtraLatency,
+		double UrgentLatency,
+		unsigned int GPUVMMaxPageTableLevels,
+		bool HostVMEnable,
+		unsigned int HostVMMaxNonCachedPageTableLevels,
+		bool GPUVMEnable,
+		double HostVMMinPageSize,
+		double PDEAndMetaPTEBytesPerFrame,
+		double MetaRowBytes,
+		double DPTEBytesPerRow,
+		double BandwidthAvailableForImmediateFlip,
+		unsigned int TotImmediateFlipBytes,
+		enum source_format_class SourcePixelFormat,
+		double LineTime,
+		double VRatio,
+		double VRatioChroma,
+		double Tno_bw,
+		bool DCCEnable,
+		unsigned int dpte_row_height,
+		unsigned int meta_row_height,
+		unsigned int dpte_row_height_chroma,
+		unsigned int meta_row_height_chroma,
+		bool    use_one_row_for_frame_flip,
+
+		/* Output */
+		double *DestinationLinesToRequestVMInImmediateFlip,
+		double *DestinationLinesToRequestRowInImmediateFlip,
+		double *final_flip_bw,
+		bool *ImmediateFlipSupportedForPipe)
+{
+	double min_row_time = 0.0;
+	unsigned int HostVMDynamicLevelsTrips;
+	double TimeForFetchingMetaPTEImmediateFlip;
+	double TimeForFetchingRowInVBlankImmediateFlip;
+	double ImmediateFlipBW = 1.0;
+
+	if (GPUVMEnable == true && HostVMEnable == true)
+		HostVMDynamicLevelsTrips = HostVMMaxNonCachedPageTableLevels;
+	else
+		HostVMDynamicLevelsTrips = 0;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: TotImmediateFlipBytes = %d\n", __func__, TotImmediateFlipBytes);
+	dml_print("DML::%s: BandwidthAvailableForImmediateFlip = %f\n", __func__, BandwidthAvailableForImmediateFlip);
+#endif
+
+	if (TotImmediateFlipBytes > 0) {
+		if (use_one_row_for_frame_flip) {
+			ImmediateFlipBW = (PDEAndMetaPTEBytesPerFrame + MetaRowBytes + 2 * DPTEBytesPerRow) *
+					BandwidthAvailableForImmediateFlip / TotImmediateFlipBytes;
+		} else {
+			ImmediateFlipBW = (PDEAndMetaPTEBytesPerFrame + MetaRowBytes + DPTEBytesPerRow) *
+					BandwidthAvailableForImmediateFlip / TotImmediateFlipBytes;
+		}
+		if (GPUVMEnable == true) {
+			TimeForFetchingMetaPTEImmediateFlip = dml_max3(Tno_bw + PDEAndMetaPTEBytesPerFrame *
+					HostVMInefficiencyFactor / ImmediateFlipBW,
+					UrgentExtraLatency + UrgentLatency *
+					(GPUVMMaxPageTableLevels * (HostVMDynamicLevelsTrips + 1) - 1),
+					LineTime / 4.0);
+		} else {
+			TimeForFetchingMetaPTEImmediateFlip = 0;
+		}
+		if ((GPUVMEnable == true || DCCEnable == true)) {
+			TimeForFetchingRowInVBlankImmediateFlip = dml_max3(
+					(MetaRowBytes + DPTEBytesPerRow * HostVMInefficiencyFactor) / ImmediateFlipBW,
+					UrgentLatency * (HostVMDynamicLevelsTrips + 1), LineTime / 4.0);
+		} else {
+			TimeForFetchingRowInVBlankImmediateFlip = 0;
+		}
+
+		*DestinationLinesToRequestVMInImmediateFlip =
+				dml_ceil(4.0 * (TimeForFetchingMetaPTEImmediateFlip / LineTime), 1.0) / 4.0;
+		*DestinationLinesToRequestRowInImmediateFlip =
+				dml_ceil(4.0 * (TimeForFetchingRowInVBlankImmediateFlip / LineTime), 1.0) / 4.0;
+
+		if (GPUVMEnable == true) {
+			*final_flip_bw = dml_max(PDEAndMetaPTEBytesPerFrame * HostVMInefficiencyFactor /
+					(*DestinationLinesToRequestVMInImmediateFlip * LineTime),
+					(MetaRowBytes + DPTEBytesPerRow * HostVMInefficiencyFactor) /
+					(*DestinationLinesToRequestRowInImmediateFlip * LineTime));
+		} else if ((GPUVMEnable == true || DCCEnable == true)) {
+			*final_flip_bw = (MetaRowBytes + DPTEBytesPerRow * HostVMInefficiencyFactor) /
+					(*DestinationLinesToRequestRowInImmediateFlip * LineTime);
+		} else {
+			*final_flip_bw = 0;
+		}
+	} else {
+		TimeForFetchingMetaPTEImmediateFlip = 0;
+		TimeForFetchingRowInVBlankImmediateFlip = 0;
+		*DestinationLinesToRequestVMInImmediateFlip = 0;
+		*DestinationLinesToRequestRowInImmediateFlip = 0;
+		*final_flip_bw = 0;
+	}
+
+	if (SourcePixelFormat == dm_420_8 || SourcePixelFormat == dm_420_10 || SourcePixelFormat == dm_rgbe_alpha) {
+		if (GPUVMEnable == true && DCCEnable != true) {
+			min_row_time = dml_min(dpte_row_height *
+					LineTime / VRatio, dpte_row_height_chroma * LineTime / VRatioChroma);
+		} else if (GPUVMEnable != true && DCCEnable == true) {
+			min_row_time = dml_min(meta_row_height *
+					LineTime / VRatio, meta_row_height_chroma * LineTime / VRatioChroma);
+		} else {
+			min_row_time = dml_min4(dpte_row_height * LineTime / VRatio, meta_row_height *
+					LineTime / VRatio, dpte_row_height_chroma * LineTime /
+					VRatioChroma, meta_row_height_chroma * LineTime / VRatioChroma);
+		}
+	} else {
+		if (GPUVMEnable == true && DCCEnable != true) {
+			min_row_time = dpte_row_height * LineTime / VRatio;
+		} else if (GPUVMEnable != true && DCCEnable == true) {
+			min_row_time = meta_row_height * LineTime / VRatio;
+		} else {
+			min_row_time =
+				dml_min(dpte_row_height * LineTime / VRatio, meta_row_height * LineTime / VRatio);
+		}
+	}
+
+	if (*DestinationLinesToRequestVMInImmediateFlip >= 32 || *DestinationLinesToRequestRowInImmediateFlip >= 16
+			|| TimeForFetchingMetaPTEImmediateFlip + 2 * TimeForFetchingRowInVBlankImmediateFlip
+					> min_row_time) {
+		*ImmediateFlipSupportedForPipe = false;
+	} else {
+		*ImmediateFlipSupportedForPipe = true;
+	}
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: GPUVMEnable = %d\n", __func__, GPUVMEnable);
+	dml_print("DML::%s: DCCEnable = %d\n", __func__, DCCEnable);
+	dml_print("DML::%s: DestinationLinesToRequestVMInImmediateFlip = %f\n",
+			__func__, *DestinationLinesToRequestVMInImmediateFlip);
+	dml_print("DML::%s: DestinationLinesToRequestRowInImmediateFlip = %f\n",
+			__func__, *DestinationLinesToRequestRowInImmediateFlip);
+	dml_print("DML::%s: TimeForFetchingMetaPTEImmediateFlip = %f\n", __func__, TimeForFetchingMetaPTEImmediateFlip);
+	dml_print("DML::%s: TimeForFetchingRowInVBlankImmediateFlip = %f\n",
+			__func__, TimeForFetchingRowInVBlankImmediateFlip);
+	dml_print("DML::%s: min_row_time = %f\n", __func__, min_row_time);
+	dml_print("DML::%s: ImmediateFlipSupportedForPipe = %d\n", __func__, *ImmediateFlipSupportedForPipe);
+#endif
+} // CalculateFlipSchedule
+
+void dml32_CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport(
+		struct vba_vars_st *v,
+		unsigned int PrefetchMode,
+		double DCFCLK,
+		double ReturnBW,
+		SOCParametersList mmSOCParameters,
+		double SOCCLK,
+		double DCFClkDeepSleep,
+		unsigned int DETBufferSizeY[],
+		unsigned int DETBufferSizeC[],
+		unsigned int SwathHeightY[],
+		unsigned int SwathHeightC[],
+		double SwathWidthY[],
+		double SwathWidthC[],
+		unsigned int DPPPerSurface[],
+		double BytePerPixelDETY[],
+		double BytePerPixelDETC[],
+		double DSTXAfterScaler[],
+		double DSTYAfterScaler[],
+		bool UnboundedRequestEnabled,
+		unsigned int CompressedBufferSizeInkByte,
+
+		/* Output */
+		enum clock_change_support *DRAMClockChangeSupport,
+		double MaxActiveDRAMClockChangeLatencySupported[],
+		unsigned int SubViewportLinesNeededInMALL[],
+		enum dm_fclock_change_support *FCLKChangeSupport,
+		double *MinActiveFCLKChangeLatencySupported,
+		bool *USRRetrainingSupport,
+		double ActiveDRAMClockChangeLatencyMargin[])
+{
+	unsigned int i, j, k;
+	unsigned int SurfaceWithMinActiveFCLKChangeMargin = 0;
+	unsigned int DRAMClockChangeSupportNumber = 0;
+	unsigned int LastSurfaceWithoutMargin = 0;
+	unsigned int DRAMClockChangeMethod = 0;
+	bool FoundFirstSurfaceWithMinActiveFCLKChangeMargin = false;
+	double MinActiveFCLKChangeMargin = 0.;
+	double SecondMinActiveFCLKChangeMarginOneDisplayInVBLank = 0.;
+	double ActiveClockChangeLatencyHidingY;
+	double ActiveClockChangeLatencyHidingC;
+	double ActiveClockChangeLatencyHiding;
+	double EffectiveDETBufferSizeY;
+	double     ActiveFCLKChangeLatencyMargin[DC__NUM_DPP__MAX];
+	double     USRRetrainingLatencyMargin[DC__NUM_DPP__MAX];
+	double TotalPixelBW = 0.0;
+	bool    SynchronizedSurfaces[DC__NUM_DPP__MAX][DC__NUM_DPP__MAX];
+	double     EffectiveLBLatencyHidingY;
+	double     EffectiveLBLatencyHidingC;
+	double     LinesInDETY[DC__NUM_DPP__MAX];
+	double     LinesInDETC[DC__NUM_DPP__MAX];
+	unsigned int    LinesInDETYRoundedDownToSwath[DC__NUM_DPP__MAX];
+	unsigned int    LinesInDETCRoundedDownToSwath[DC__NUM_DPP__MAX];
+	double     FullDETBufferingTimeY;
+	double     FullDETBufferingTimeC;
+	double     WritebackDRAMClockChangeLatencyMargin;
+	double     WritebackFCLKChangeLatencyMargin;
+	double     WritebackLatencyHiding;
+	bool    SameTimingForFCLKChange;
+
+	unsigned int    TotalActiveWriteback = 0;
+	unsigned int LBLatencyHidingSourceLinesY[DC__NUM_DPP__MAX];
+	unsigned int LBLatencyHidingSourceLinesC[DC__NUM_DPP__MAX];
+
+	v->Watermark.UrgentWatermark = mmSOCParameters.UrgentLatency + mmSOCParameters.ExtraLatency;
+	v->Watermark.USRRetrainingWatermark = mmSOCParameters.UrgentLatency + mmSOCParameters.ExtraLatency
+			+ mmSOCParameters.USRRetrainingLatency + mmSOCParameters.SMNLatency;
+	v->Watermark.DRAMClockChangeWatermark = mmSOCParameters.DRAMClockChangeLatency + v->Watermark.UrgentWatermark;
+	v->Watermark.FCLKChangeWatermark = mmSOCParameters.FCLKChangeLatency + v->Watermark.UrgentWatermark;
+	v->Watermark.StutterExitWatermark = mmSOCParameters.SRExitTime + mmSOCParameters.ExtraLatency
+			+ 10 / DCFClkDeepSleep;
+	v->Watermark.StutterEnterPlusExitWatermark = mmSOCParameters.SREnterPlusExitTime + mmSOCParameters.ExtraLatency
+			+ 10 / DCFClkDeepSleep;
+	v->Watermark.Z8StutterExitWatermark = mmSOCParameters.SRExitZ8Time + mmSOCParameters.ExtraLatency
+			+ 10 / DCFClkDeepSleep;
+	v->Watermark.Z8StutterEnterPlusExitWatermark = mmSOCParameters.SREnterPlusExitZ8Time
+			+ mmSOCParameters.ExtraLatency + 10 / DCFClkDeepSleep;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: UrgentLatency = %f\n", __func__, mmSOCParameters.UrgentLatency);
+	dml_print("DML::%s: ExtraLatency = %f\n", __func__, mmSOCParameters.ExtraLatency);
+	dml_print("DML::%s: DRAMClockChangeLatency = %f\n", __func__, mmSOCParameters.DRAMClockChangeLatency);
+	dml_print("DML::%s: UrgentWatermark = %f\n", __func__, v->Watermark.UrgentWatermark);
+	dml_print("DML::%s: USRRetrainingWatermark = %f\n", __func__, v->Watermark.USRRetrainingWatermark);
+	dml_print("DML::%s: DRAMClockChangeWatermark = %f\n", __func__, v->Watermark.DRAMClockChangeWatermark);
+	dml_print("DML::%s: FCLKChangeWatermark = %f\n", __func__, v->Watermark.FCLKChangeWatermark);
+	dml_print("DML::%s: StutterExitWatermark = %f\n", __func__, v->Watermark.StutterExitWatermark);
+	dml_print("DML::%s: StutterEnterPlusExitWatermark = %f\n", __func__, v->Watermark.StutterEnterPlusExitWatermark);
+	dml_print("DML::%s: Z8StutterExitWatermark = %f\n", __func__, v->Watermark.Z8StutterExitWatermark);
+	dml_print("DML::%s: Z8StutterEnterPlusExitWatermark = %f\n",
+			__func__, v->Watermark.Z8StutterEnterPlusExitWatermark);
+#endif
+
+
+	TotalActiveWriteback = 0;
+	for (k = 0; k < v->NumberOfActiveSurfaces; ++k) {
+		if (v->WritebackEnable[k] == true)
+			TotalActiveWriteback = TotalActiveWriteback + 1;
+	}
+
+	if (TotalActiveWriteback <= 1) {
+		v->Watermark.WritebackUrgentWatermark = mmSOCParameters.WritebackLatency;
+	} else {
+		v->Watermark.WritebackUrgentWatermark = mmSOCParameters.WritebackLatency
+				+ v->WritebackChunkSize * 1024.0 / 32.0 / SOCCLK;
+	}
+	if (v->USRRetrainingRequiredFinal)
+		v->Watermark.WritebackDRAMClockChangeWatermark = v->Watermark.WritebackDRAMClockChangeWatermark
+				+ mmSOCParameters.USRRetrainingLatency;
+
+	if (TotalActiveWriteback <= 1) {
+		v->Watermark.WritebackDRAMClockChangeWatermark = mmSOCParameters.DRAMClockChangeLatency
+				+ mmSOCParameters.WritebackLatency;
+		v->Watermark.WritebackFCLKChangeWatermark = mmSOCParameters.FCLKChangeLatency
+				+ mmSOCParameters.WritebackLatency;
+	} else {
+		v->Watermark.WritebackDRAMClockChangeWatermark = mmSOCParameters.DRAMClockChangeLatency
+				+ mmSOCParameters.WritebackLatency + v->WritebackChunkSize * 1024.0 / 32.0 / SOCCLK;
+		v->Watermark.WritebackFCLKChangeWatermark = mmSOCParameters.FCLKChangeLatency
+				+ mmSOCParameters.WritebackLatency + v->WritebackChunkSize * 1024 / 32 / SOCCLK;
+	}
+
+	if (v->USRRetrainingRequiredFinal)
+		v->Watermark.WritebackDRAMClockChangeWatermark = v->Watermark.WritebackDRAMClockChangeWatermark
+				+ mmSOCParameters.USRRetrainingLatency;
+
+	if (v->USRRetrainingRequiredFinal)
+		v->Watermark.WritebackFCLKChangeWatermark = v->Watermark.WritebackFCLKChangeWatermark
+				+ mmSOCParameters.USRRetrainingLatency;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: WritebackDRAMClockChangeWatermark = %f\n",
+			__func__, v->Watermark.WritebackDRAMClockChangeWatermark);
+	dml_print("DML::%s: WritebackFCLKChangeWatermark = %f\n", __func__, v->Watermark.WritebackFCLKChangeWatermark);
+	dml_print("DML::%s: WritebackUrgentWatermark = %f\n", __func__, v->Watermark.WritebackUrgentWatermark);
+	dml_print("DML::%s: v->USRRetrainingRequiredFinal = %d\n", __func__, v->USRRetrainingRequiredFinal);
+	dml_print("DML::%s: USRRetrainingLatency = %f\n", __func__, mmSOCParameters.USRRetrainingLatency);
+#endif
+
+	for (k = 0; k < v->NumberOfActiveSurfaces; ++k) {
+		TotalPixelBW = TotalPixelBW + DPPPerSurface[k] * (SwathWidthY[k] * BytePerPixelDETY[k] * v->VRatio[k] +
+				SwathWidthC[k] * BytePerPixelDETC[k] * v->VRatioChroma[k]) / (v->HTotal[k] / v->PixelClock[k]);
+	}
+
+	for (k = 0; k < v->NumberOfActiveSurfaces; ++k) {
+
+		LBLatencyHidingSourceLinesY[k] = dml_min((double) v->MaxLineBufferLines, dml_floor(v->LineBufferSizeFinal / v->LBBitPerPixel[k] / (SwathWidthY[k] / dml_max(v->HRatio[k], 1.0)), 1)) - (v->vtaps[k] - 1);
+		LBLatencyHidingSourceLinesC[k] = dml_min((double) v->MaxLineBufferLines, dml_floor(v->LineBufferSizeFinal / v->LBBitPerPixel[k] / (SwathWidthC[k] / dml_max(v->HRatioChroma[k], 1.0)), 1)) - (v->VTAPsChroma[k] - 1);
+
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%d, v->MaxLineBufferLines = %d\n", __func__, k, v->MaxLineBufferLines);
+		dml_print("DML::%s: k=%d, v->LineBufferSizeFinal     = %d\n", __func__, k, v->LineBufferSizeFinal);
+		dml_print("DML::%s: k=%d, v->LBBitPerPixel      = %d\n", __func__, k, v->LBBitPerPixel[k]);
+		dml_print("DML::%s: k=%d, v->HRatio             = %f\n", __func__, k, v->HRatio[k]);
+		dml_print("DML::%s: k=%d, v->vtaps              = %d\n", __func__, k, v->vtaps[k]);
+#endif
+
+		EffectiveLBLatencyHidingY = LBLatencyHidingSourceLinesY[k] / v->VRatio[k] * (v->HTotal[k] / v->PixelClock[k]);
+		EffectiveLBLatencyHidingC = LBLatencyHidingSourceLinesC[k] / v->VRatioChroma[k] * (v->HTotal[k] / v->PixelClock[k]);
+		EffectiveDETBufferSizeY = DETBufferSizeY[k];
+
+		if (UnboundedRequestEnabled) {
+			EffectiveDETBufferSizeY = EffectiveDETBufferSizeY
+					+ CompressedBufferSizeInkByte * 1024
+							* (SwathWidthY[k] * BytePerPixelDETY[k] * v->VRatio[k])
+							/ (v->HTotal[k] / v->PixelClock[k]) / TotalPixelBW;
+		}
+
+		LinesInDETY[k] = (double) EffectiveDETBufferSizeY / BytePerPixelDETY[k] / SwathWidthY[k];
+		LinesInDETYRoundedDownToSwath[k] = dml_floor(LinesInDETY[k], SwathHeightY[k]);
+		FullDETBufferingTimeY = LinesInDETYRoundedDownToSwath[k] * (v->HTotal[k] / v->PixelClock[k]) / v->VRatio[k];
+
+		ActiveClockChangeLatencyHidingY = EffectiveLBLatencyHidingY + FullDETBufferingTimeY
+				- (DSTXAfterScaler[k] / v->HTotal[k] + DSTYAfterScaler[k]) * v->HTotal[k] / v->PixelClock[k];
+
+		if (v->NumberOfActiveSurfaces > 1) {
+			ActiveClockChangeLatencyHidingY = ActiveClockChangeLatencyHidingY
+					- (1.0 - 1.0 / v->NumberOfActiveSurfaces) * SwathHeightY[k] * v->HTotal[k]
+							/ v->PixelClock[k] / v->VRatio[k];
+		}
+
+		if (BytePerPixelDETC[k] > 0) {
+			LinesInDETC[k] = DETBufferSizeC[k] / BytePerPixelDETC[k] / SwathWidthC[k];
+			LinesInDETCRoundedDownToSwath[k] = dml_floor(LinesInDETC[k], SwathHeightC[k]);
+			FullDETBufferingTimeC = LinesInDETCRoundedDownToSwath[k] * (v->HTotal[k] / v->PixelClock[k])
+					/ v->VRatioChroma[k];
+			ActiveClockChangeLatencyHidingC = EffectiveLBLatencyHidingC + FullDETBufferingTimeC
+					- (DSTXAfterScaler[k] / v->HTotal[k] + DSTYAfterScaler[k]) * v->HTotal[k]
+							/ v->PixelClock[k];
+			if (v->NumberOfActiveSurfaces > 1) {
+				ActiveClockChangeLatencyHidingC = ActiveClockChangeLatencyHidingC
+						- (1 - 1 / v->NumberOfActiveSurfaces) * SwathHeightC[k] * v->HTotal[k]
+								/ v->PixelClock[k] / v->VRatioChroma[k];
+			}
+			ActiveClockChangeLatencyHiding = dml_min(ActiveClockChangeLatencyHidingY,
+					ActiveClockChangeLatencyHidingC);
+		} else {
+			ActiveClockChangeLatencyHiding = ActiveClockChangeLatencyHidingY;
+		}
+
+		ActiveDRAMClockChangeLatencyMargin[k] = ActiveClockChangeLatencyHiding - v->Watermark.UrgentWatermark
+				- v->Watermark.DRAMClockChangeWatermark;
+		ActiveFCLKChangeLatencyMargin[k] = ActiveClockChangeLatencyHiding - v->Watermark.UrgentWatermark
+				- v->Watermark.FCLKChangeWatermark;
+		USRRetrainingLatencyMargin[k] = ActiveClockChangeLatencyHiding - v->Watermark.USRRetrainingWatermark;
+
+		if (v->WritebackEnable[k]) {
+			WritebackLatencyHiding = v->WritebackInterfaceBufferSize * 1024
+					/ (v->WritebackDestinationWidth[k] * v->WritebackDestinationHeight[k]
+							/ (v->WritebackSourceHeight[k] * v->HTotal[k] / v->PixelClock[k]) * 4);
+			if (v->WritebackPixelFormat[k] == dm_444_64)
+				WritebackLatencyHiding = WritebackLatencyHiding / 2;
+
+			WritebackDRAMClockChangeLatencyMargin = WritebackLatencyHiding
+					- v->Watermark.WritebackDRAMClockChangeWatermark;
+
+			WritebackFCLKChangeLatencyMargin = WritebackLatencyHiding
+					- v->Watermark.WritebackFCLKChangeWatermark;
+
+			ActiveDRAMClockChangeLatencyMargin[k] = dml_min(ActiveDRAMClockChangeLatencyMargin[k],
+					WritebackFCLKChangeLatencyMargin);
+			ActiveFCLKChangeLatencyMargin[k] = dml_min(ActiveFCLKChangeLatencyMargin[k],
+					WritebackDRAMClockChangeLatencyMargin);
+		}
+		MaxActiveDRAMClockChangeLatencySupported[k] =
+				(v->UsesMALLForPStateChange[k] == dm_use_mall_pstate_change_phantom_pipe) ?
+						0 :
+						(ActiveDRAMClockChangeLatencyMargin[k]
+								+ mmSOCParameters.DRAMClockChangeLatency);
+	}
+
+	for (i = 0; i < v->NumberOfActiveSurfaces; ++i) {
+		for (j = 0; j < v->NumberOfActiveSurfaces; ++j) {
+			if (i == j ||
+					(v->BlendingAndTiming[i] == i && v->BlendingAndTiming[j] == i) ||
+					(v->BlendingAndTiming[j] == j && v->BlendingAndTiming[i] == j) ||
+					(v->BlendingAndTiming[i] == v->BlendingAndTiming[j] && v->BlendingAndTiming[i] != i) ||
+					(v->SynchronizeTimingsFinal && v->PixelClock[i] == v->PixelClock[j] &&
+					v->HTotal[i] == v->HTotal[j] && v->VTotal[i] == v->VTotal[j] &&
+					v->VActive[i] == v->VActive[j]) || (v->SynchronizeDRRDisplaysForUCLKPStateChangeFinal &&
+					(v->DRRDisplay[i] || v->DRRDisplay[j]))) {
+				SynchronizedSurfaces[i][j] = true;
+			} else {
+				SynchronizedSurfaces[i][j] = false;
+			}
+		}
+	}
+
+	for (k = 0; k < v->NumberOfActiveSurfaces; ++k) {
+		if ((v->UsesMALLForPStateChange[k] != dm_use_mall_pstate_change_phantom_pipe) &&
+				(!FoundFirstSurfaceWithMinActiveFCLKChangeMargin ||
+				ActiveFCLKChangeLatencyMargin[k] < MinActiveFCLKChangeMargin)) {
+			FoundFirstSurfaceWithMinActiveFCLKChangeMargin = true;
+			MinActiveFCLKChangeMargin = ActiveFCLKChangeLatencyMargin[k];
+			SurfaceWithMinActiveFCLKChangeMargin = k;
+		}
+	}
+
+	*MinActiveFCLKChangeLatencySupported = MinActiveFCLKChangeMargin + mmSOCParameters.FCLKChangeLatency;
+
+	SameTimingForFCLKChange = true;
+	for (k = 0; k < v->NumberOfActiveSurfaces; ++k) {
+		if (!SynchronizedSurfaces[k][SurfaceWithMinActiveFCLKChangeMargin]) {
+			if ((v->UsesMALLForPStateChange[k] != dm_use_mall_pstate_change_phantom_pipe) &&
+					(SameTimingForFCLKChange ||
+					ActiveFCLKChangeLatencyMargin[k] <
+					SecondMinActiveFCLKChangeMarginOneDisplayInVBLank)) {
+				SecondMinActiveFCLKChangeMarginOneDisplayInVBLank = ActiveFCLKChangeLatencyMargin[k];
+			}
+			SameTimingForFCLKChange = false;
+		}
+	}
+
+	if (MinActiveFCLKChangeMargin > 0) {
+		*FCLKChangeSupport = dm_fclock_change_vactive;
+	} else if ((SameTimingForFCLKChange || SecondMinActiveFCLKChangeMarginOneDisplayInVBLank > 0) &&
+			(PrefetchMode <= 1)) {
+		*FCLKChangeSupport = dm_fclock_change_vblank;
+	} else {
+		*FCLKChangeSupport = dm_fclock_change_unsupported;
+	}
+
+	*USRRetrainingSupport = true;
+	for (k = 0; k < v->NumberOfActiveSurfaces; ++k) {
+		if ((v->UsesMALLForPStateChange[k] != dm_use_mall_pstate_change_phantom_pipe) &&
+				(USRRetrainingLatencyMargin[k] < 0)) {
+			*USRRetrainingSupport = false;
+		}
+	}
+
+	for (k = 0; k < v->NumberOfActiveSurfaces; ++k) {
+		if (v->UsesMALLForPStateChange[k] != dm_use_mall_pstate_change_full_frame &&
+				v->UsesMALLForPStateChange[k] != dm_use_mall_pstate_change_sub_viewport &&
+				v->UsesMALLForPStateChange[k] != dm_use_mall_pstate_change_phantom_pipe &&
+				ActiveDRAMClockChangeLatencyMargin[k] < 0) {
+			if (PrefetchMode > 0) {
+				DRAMClockChangeSupportNumber = 2;
+			} else if (DRAMClockChangeSupportNumber == 0) {
+				DRAMClockChangeSupportNumber = 1;
+				LastSurfaceWithoutMargin = k;
+			} else if (DRAMClockChangeSupportNumber == 1 &&
+					!SynchronizedSurfaces[LastSurfaceWithoutMargin][k]) {
+				DRAMClockChangeSupportNumber = 2;
+			}
+		}
+	}
+
+	for (k = 0; k < v->NumberOfActiveSurfaces; ++k) {
+		if (v->UsesMALLForPStateChange[k] == dm_use_mall_pstate_change_full_frame)
+			DRAMClockChangeMethod = 1;
+		else if (v->UsesMALLForPStateChange[k] == dm_use_mall_pstate_change_sub_viewport)
+			DRAMClockChangeMethod = 2;
+	}
+
+	if (DRAMClockChangeMethod == 0) {
+		if (DRAMClockChangeSupportNumber == 0)
+			*DRAMClockChangeSupport = dm_dram_clock_change_vactive;
+		else if (DRAMClockChangeSupportNumber == 1)
+			*DRAMClockChangeSupport = dm_dram_clock_change_vblank;
+		else
+			*DRAMClockChangeSupport = dm_dram_clock_change_unsupported;
+	} else if (DRAMClockChangeMethod == 1) {
+		if (DRAMClockChangeSupportNumber == 0)
+			*DRAMClockChangeSupport = dm_dram_clock_change_vactive_w_mall_full_frame;
+		else if (DRAMClockChangeSupportNumber == 1)
+			*DRAMClockChangeSupport = dm_dram_clock_change_vblank_w_mall_full_frame;
+		else
+			*DRAMClockChangeSupport = dm_dram_clock_change_unsupported;
+	} else {
+		if (DRAMClockChangeSupportNumber == 0)
+			*DRAMClockChangeSupport = dm_dram_clock_change_vactive_w_mall_sub_vp;
+		else if (DRAMClockChangeSupportNumber == 1)
+			*DRAMClockChangeSupport = dm_dram_clock_change_vblank_w_mall_sub_vp;
+		else
+			*DRAMClockChangeSupport = dm_dram_clock_change_unsupported;
+	}
+
+	for (k = 0; k < v->NumberOfActiveSurfaces; ++k) {
+		unsigned int dst_y_pstate;
+		unsigned int src_y_pstate_l;
+		unsigned int src_y_pstate_c;
+		unsigned int src_y_ahead_l, src_y_ahead_c, sub_vp_lines_l, sub_vp_lines_c;
+
+		dst_y_pstate = dml_ceil((mmSOCParameters.DRAMClockChangeLatency + mmSOCParameters.UrgentLatency) / (v->HTotal[k] / v->PixelClock[k]), 1);
+		src_y_pstate_l = dml_ceil(dst_y_pstate * v->VRatio[k], SwathHeightY[k]);
+		src_y_ahead_l = dml_floor(DETBufferSizeY[k] / BytePerPixelDETY[k] / SwathWidthY[k], SwathHeightY[k]) + LBLatencyHidingSourceLinesY[k];
+		sub_vp_lines_l = src_y_pstate_l + src_y_ahead_l + v->meta_row_height[k];
+
+#ifdef __DML_VBA_DEBUG__
+dml_print("DML::%s: k=%d, DETBufferSizeY               = %d\n", __func__, k, DETBufferSizeY[k]);
+dml_print("DML::%s: k=%d, BytePerPixelDETY             = %f\n", __func__, k, BytePerPixelDETY[k]);
+dml_print("DML::%s: k=%d, SwathWidthY                  = %d\n", __func__, k, SwathWidthY[k]);
+dml_print("DML::%s: k=%d, SwathHeightY                 = %d\n", __func__, k, SwathHeightY[k]);
+dml_print("DML::%s: k=%d, LBLatencyHidingSourceLinesY  = %d\n", __func__, k, LBLatencyHidingSourceLinesY[k]);
+dml_print("DML::%s: k=%d, dst_y_pstate      = %d\n", __func__, k, dst_y_pstate);
+dml_print("DML::%s: k=%d, src_y_pstate_l    = %d\n", __func__, k, src_y_pstate_l);
+dml_print("DML::%s: k=%d, src_y_ahead_l     = %d\n", __func__, k, src_y_ahead_l);
+dml_print("DML::%s: k=%d, v->meta_row_height   = %d\n", __func__, k, v->meta_row_height[k]);
+dml_print("DML::%s: k=%d, sub_vp_lines_l    = %d\n", __func__, k, sub_vp_lines_l);
+#endif
+		SubViewportLinesNeededInMALL[k] = sub_vp_lines_l;
+
+		if (BytePerPixelDETC[k] > 0) {
+			src_y_pstate_c = dml_ceil(dst_y_pstate * v->VRatioChroma[k], SwathHeightC[k]);
+			src_y_ahead_c = dml_floor(DETBufferSizeC[k] / BytePerPixelDETC[k] / SwathWidthC[k], SwathHeightC[k]) + LBLatencyHidingSourceLinesC[k];
+			sub_vp_lines_c = src_y_pstate_c + src_y_ahead_c + v->meta_row_height_chroma[k];
+			SubViewportLinesNeededInMALL[k] = dml_max(sub_vp_lines_l, sub_vp_lines_c);
+
+#ifdef __DML_VBA_DEBUG__
+dml_print("DML::%s: k=%d, src_y_pstate_c            = %d\n", __func__, k, src_y_pstate_c);
+dml_print("DML::%s: k=%d, src_y_ahead_c             = %d\n", __func__, k, src_y_ahead_c);
+dml_print("DML::%s: k=%d, v->meta_row_height_chroma    = %d\n", __func__, k, v->meta_row_height_chroma[k]);
+dml_print("DML::%s: k=%d, sub_vp_lines_c            = %d\n", __func__, k, sub_vp_lines_c);
+#endif
+		}
+	}
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: DRAMClockChangeSupport = %d\n", __func__, *DRAMClockChangeSupport);
+	dml_print("DML::%s: FCLKChangeSupport = %d\n", __func__, *FCLKChangeSupport);
+	dml_print("DML::%s: MinActiveFCLKChangeLatencySupported = %f\n",
+			__func__, *MinActiveFCLKChangeLatencySupported);
+	dml_print("DML::%s: USRRetrainingSupport = %d\n", __func__, *USRRetrainingSupport);
+#endif
+} // CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport
+
+double dml32_CalculateWriteBackDISPCLK(
+		enum source_format_class WritebackPixelFormat,
+		double PixelClock,
+		double WritebackHRatio,
+		double WritebackVRatio,
+		unsigned int WritebackHTaps,
+		unsigned int WritebackVTaps,
+		unsigned int   WritebackSourceWidth,
+		unsigned int   WritebackDestinationWidth,
+		unsigned int HTotal,
+		unsigned int WritebackLineBufferSize,
+		double DISPCLKDPPCLKVCOSpeed)
+{
+	double DISPCLK_H, DISPCLK_V, DISPCLK_HB;
+
+	DISPCLK_H = PixelClock * dml_ceil(WritebackHTaps / 8.0, 1) / WritebackHRatio;
+	DISPCLK_V = PixelClock * (WritebackVTaps * dml_ceil(WritebackDestinationWidth / 6.0, 1) + 8.0) / HTotal;
+	DISPCLK_HB = PixelClock * WritebackVTaps * (WritebackDestinationWidth *
+			WritebackVTaps - WritebackLineBufferSize / 57.0) / 6.0 / WritebackSourceWidth;
+	return dml32_RoundToDFSGranularity(dml_max3(DISPCLK_H, DISPCLK_V, DISPCLK_HB), 1, DISPCLKDPPCLKVCOSpeed);
+}
+
+void dml32_CalculateMinAndMaxPrefetchMode(
+		enum dm_prefetch_modes   AllowForPStateChangeOrStutterInVBlankFinal,
+		unsigned int             *MinPrefetchMode,
+		unsigned int             *MaxPrefetchMode)
+{
+	if (AllowForPStateChangeOrStutterInVBlankFinal == dm_prefetch_support_none) {
+		*MinPrefetchMode = 3;
+		*MaxPrefetchMode = 3;
+	} else if (AllowForPStateChangeOrStutterInVBlankFinal == dm_prefetch_support_stutter) {
+		*MinPrefetchMode = 2;
+		*MaxPrefetchMode = 2;
+	} else if (AllowForPStateChangeOrStutterInVBlankFinal == dm_prefetch_support_fclk_and_stutter) {
+		*MinPrefetchMode = 1;
+		*MaxPrefetchMode = 1;
+	} else if (AllowForPStateChangeOrStutterInVBlankFinal == dm_prefetch_support_uclk_fclk_and_stutter) {
+		*MinPrefetchMode = 0;
+		*MaxPrefetchMode = 0;
+	} else {
+		*MinPrefetchMode = 0;
+		*MaxPrefetchMode = 3;
+	}
+} // CalculateMinAndMaxPrefetchMode
+
+void dml32_CalculatePixelDeliveryTimes(
+		unsigned int             NumberOfActiveSurfaces,
+		double              VRatio[],
+		double              VRatioChroma[],
+		double              VRatioPrefetchY[],
+		double              VRatioPrefetchC[],
+		unsigned int             swath_width_luma_ub[],
+		unsigned int             swath_width_chroma_ub[],
+		unsigned int             DPPPerSurface[],
+		double              HRatio[],
+		double              HRatioChroma[],
+		double              PixelClock[],
+		double              PSCL_THROUGHPUT[],
+		double              PSCL_THROUGHPUT_CHROMA[],
+		double              Dppclk[],
+		unsigned int             BytePerPixelC[],
+		enum dm_rotation_angle   SourceRotation[],
+		unsigned int             NumberOfCursors[],
+		unsigned int             CursorWidth[][DC__NUM_CURSOR__MAX],
+		unsigned int             CursorBPP[][DC__NUM_CURSOR__MAX],
+		unsigned int             BlockWidth256BytesY[],
+		unsigned int             BlockHeight256BytesY[],
+		unsigned int             BlockWidth256BytesC[],
+		unsigned int             BlockHeight256BytesC[],
+
+		/* Output */
+		double              DisplayPipeLineDeliveryTimeLuma[],
+		double              DisplayPipeLineDeliveryTimeChroma[],
+		double              DisplayPipeLineDeliveryTimeLumaPrefetch[],
+		double              DisplayPipeLineDeliveryTimeChromaPrefetch[],
+		double              DisplayPipeRequestDeliveryTimeLuma[],
+		double              DisplayPipeRequestDeliveryTimeChroma[],
+		double              DisplayPipeRequestDeliveryTimeLumaPrefetch[],
+		double              DisplayPipeRequestDeliveryTimeChromaPrefetch[],
+		double              CursorRequestDeliveryTime[],
+		double              CursorRequestDeliveryTimePrefetch[])
+{
+	double   req_per_swath_ub;
+	unsigned int k;
+
+	for (k = 0; k < NumberOfActiveSurfaces; ++k) {
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%d : HRatio = %f\n", __func__, k, HRatio[k]);
+		dml_print("DML::%s: k=%d : VRatio = %f\n", __func__, k, VRatio[k]);
+		dml_print("DML::%s: k=%d : HRatioChroma = %f\n", __func__, k, HRatioChroma[k]);
+		dml_print("DML::%s: k=%d : VRatioChroma = %f\n", __func__, k, VRatioChroma[k]);
+		dml_print("DML::%s: k=%d : swath_width_luma_ub = %d\n", __func__, k, swath_width_luma_ub[k]);
+		dml_print("DML::%s: k=%d : swath_width_chroma_ub = %d\n", __func__, k, swath_width_chroma_ub[k]);
+		dml_print("DML::%s: k=%d : PSCL_THROUGHPUT = %f\n", __func__, k, PSCL_THROUGHPUT[k]);
+		dml_print("DML::%s: k=%d : PSCL_THROUGHPUT_CHROMA = %f\n", __func__, k, PSCL_THROUGHPUT_CHROMA[k]);
+		dml_print("DML::%s: k=%d : DPPPerSurface = %d\n", __func__, k, DPPPerSurface[k]);
+		dml_print("DML::%s: k=%d : PixelClock = %f\n", __func__, k, PixelClock[k]);
+		dml_print("DML::%s: k=%d : Dppclk = %f\n", __func__, k, Dppclk[k]);
+#endif
+
+		if (VRatio[k] <= 1) {
+			DisplayPipeLineDeliveryTimeLuma[k] =
+					swath_width_luma_ub[k] * DPPPerSurface[k] / HRatio[k] / PixelClock[k];
+		} else {
+			DisplayPipeLineDeliveryTimeLuma[k] = swath_width_luma_ub[k] / PSCL_THROUGHPUT[k] / Dppclk[k];
+		}
+
+		if (BytePerPixelC[k] == 0) {
+			DisplayPipeLineDeliveryTimeChroma[k] = 0;
+		} else {
+			if (VRatioChroma[k] <= 1) {
+				DisplayPipeLineDeliveryTimeChroma[k] =
+					swath_width_chroma_ub[k] * DPPPerSurface[k] / HRatioChroma[k] / PixelClock[k];
+			} else {
+				DisplayPipeLineDeliveryTimeChroma[k] =
+					swath_width_chroma_ub[k] / PSCL_THROUGHPUT_CHROMA[k] / Dppclk[k];
+			}
+		}
+
+		if (VRatioPrefetchY[k] <= 1) {
+			DisplayPipeLineDeliveryTimeLumaPrefetch[k] =
+					swath_width_luma_ub[k] * DPPPerSurface[k] / HRatio[k] / PixelClock[k];
+		} else {
+			DisplayPipeLineDeliveryTimeLumaPrefetch[k] =
+					swath_width_luma_ub[k] / PSCL_THROUGHPUT[k] / Dppclk[k];
+		}
+
+		if (BytePerPixelC[k] == 0) {
+			DisplayPipeLineDeliveryTimeChromaPrefetch[k] = 0;
+		} else {
+			if (VRatioPrefetchC[k] <= 1) {
+				DisplayPipeLineDeliveryTimeChromaPrefetch[k] = swath_width_chroma_ub[k] *
+						DPPPerSurface[k] / HRatioChroma[k] / PixelClock[k];
+			} else {
+				DisplayPipeLineDeliveryTimeChromaPrefetch[k] =
+						swath_width_chroma_ub[k] / PSCL_THROUGHPUT_CHROMA[k] / Dppclk[k];
+			}
+		}
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%d : DisplayPipeLineDeliveryTimeLuma = %f\n",
+				__func__, k, DisplayPipeLineDeliveryTimeLuma[k]);
+		dml_print("DML::%s: k=%d : DisplayPipeLineDeliveryTimeLumaPrefetch = %f\n",
+				__func__, k, DisplayPipeLineDeliveryTimeLumaPrefetch[k]);
+		dml_print("DML::%s: k=%d : DisplayPipeLineDeliveryTimeChroma = %f\n",
+				__func__, k, DisplayPipeLineDeliveryTimeChroma[k]);
+		dml_print("DML::%s: k=%d : DisplayPipeLineDeliveryTimeChromaPrefetch = %f\n",
+				__func__, k, DisplayPipeLineDeliveryTimeChromaPrefetch[k]);
+#endif
+	}
+
+	for (k = 0; k < NumberOfActiveSurfaces; ++k) {
+		if (!IsVertical(SourceRotation[k]))
+			req_per_swath_ub = swath_width_luma_ub[k] / BlockWidth256BytesY[k];
+		else
+			req_per_swath_ub = swath_width_luma_ub[k] / BlockHeight256BytesY[k];
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%d : req_per_swath_ub = %f (Luma)\n", __func__, k, req_per_swath_ub);
+#endif
+
+		DisplayPipeRequestDeliveryTimeLuma[k] = DisplayPipeLineDeliveryTimeLuma[k] / req_per_swath_ub;
+		DisplayPipeRequestDeliveryTimeLumaPrefetch[k] =
+				DisplayPipeLineDeliveryTimeLumaPrefetch[k] / req_per_swath_ub;
+		if (BytePerPixelC[k] == 0) {
+			DisplayPipeRequestDeliveryTimeChroma[k] = 0;
+			DisplayPipeRequestDeliveryTimeChromaPrefetch[k] = 0;
+		} else {
+			if (!IsVertical(SourceRotation[k]))
+				req_per_swath_ub = swath_width_chroma_ub[k] / BlockWidth256BytesC[k];
+			else
+				req_per_swath_ub = swath_width_chroma_ub[k] / BlockHeight256BytesC[k];
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: k=%d : req_per_swath_ub = %f (Chroma)\n", __func__, k, req_per_swath_ub);
+#endif
+			DisplayPipeRequestDeliveryTimeChroma[k] =
+					DisplayPipeLineDeliveryTimeChroma[k] / req_per_swath_ub;
+			DisplayPipeRequestDeliveryTimeChromaPrefetch[k] =
+					DisplayPipeLineDeliveryTimeChromaPrefetch[k] / req_per_swath_ub;
+		}
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%d : DisplayPipeRequestDeliveryTimeLuma = %f\n",
+				__func__, k, DisplayPipeRequestDeliveryTimeLuma[k]);
+		dml_print("DML::%s: k=%d : DisplayPipeRequestDeliveryTimeLumaPrefetch = %f\n",
+				__func__, k, DisplayPipeRequestDeliveryTimeLumaPrefetch[k]);
+		dml_print("DML::%s: k=%d : DisplayPipeRequestDeliveryTimeChroma = %f\n",
+				__func__, k, DisplayPipeRequestDeliveryTimeChroma[k]);
+		dml_print("DML::%s: k=%d : DisplayPipeRequestDeliveryTimeChromaPrefetch = %f\n",
+				__func__, k, DisplayPipeRequestDeliveryTimeChromaPrefetch[k]);
+#endif
+	}
+
+	for (k = 0; k < NumberOfActiveSurfaces; ++k) {
+		unsigned int cursor_req_per_width;
+
+		cursor_req_per_width = dml_ceil((double) CursorWidth[k][0] * (double) CursorBPP[k][0] /
+				256.0 / 8.0, 1.0);
+		if (NumberOfCursors[k] > 0) {
+			if (VRatio[k] <= 1) {
+				CursorRequestDeliveryTime[k] = (double) CursorWidth[k][0] /
+						HRatio[k] / PixelClock[k] / cursor_req_per_width;
+			} else {
+				CursorRequestDeliveryTime[k] = (double) CursorWidth[k][0] /
+						PSCL_THROUGHPUT[k] / Dppclk[k] / cursor_req_per_width;
+			}
+			if (VRatioPrefetchY[k] <= 1) {
+				CursorRequestDeliveryTimePrefetch[k] = (double) CursorWidth[k][0] /
+						HRatio[k] / PixelClock[k] / cursor_req_per_width;
+			} else {
+				CursorRequestDeliveryTimePrefetch[k] = (double) CursorWidth[k][0] /
+						PSCL_THROUGHPUT[k] / Dppclk[k] / cursor_req_per_width;
+			}
+		} else {
+			CursorRequestDeliveryTime[k] = 0;
+			CursorRequestDeliveryTimePrefetch[k] = 0;
+		}
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%d : NumberOfCursors = %d\n",
+				__func__, k, NumberOfCursors[k]);
+		dml_print("DML::%s: k=%d : CursorRequestDeliveryTime = %f\n",
+				__func__, k, CursorRequestDeliveryTime[k]);
+		dml_print("DML::%s: k=%d : CursorRequestDeliveryTimePrefetch = %f\n",
+				__func__, k, CursorRequestDeliveryTimePrefetch[k]);
+#endif
+	}
+} // CalculatePixelDeliveryTimes
+
+void dml32_CalculateMetaAndPTETimes(
+		bool use_one_row_for_frame[],
+		unsigned int NumberOfActiveSurfaces,
+		bool GPUVMEnable,
+		unsigned int MetaChunkSize,
+		unsigned int MinMetaChunkSizeBytes,
+		unsigned int    HTotal[],
+		double  VRatio[],
+		double  VRatioChroma[],
+		double  DestinationLinesToRequestRowInVBlank[],
+		double  DestinationLinesToRequestRowInImmediateFlip[],
+		bool DCCEnable[],
+		double  PixelClock[],
+		unsigned int BytePerPixelY[],
+		unsigned int BytePerPixelC[],
+		enum dm_rotation_angle SourceRotation[],
+		unsigned int dpte_row_height[],
+		unsigned int dpte_row_height_chroma[],
+		unsigned int meta_row_width[],
+		unsigned int meta_row_width_chroma[],
+		unsigned int meta_row_height[],
+		unsigned int meta_row_height_chroma[],
+		unsigned int meta_req_width[],
+		unsigned int meta_req_width_chroma[],
+		unsigned int meta_req_height[],
+		unsigned int meta_req_height_chroma[],
+		unsigned int dpte_group_bytes[],
+		unsigned int    PTERequestSizeY[],
+		unsigned int    PTERequestSizeC[],
+		unsigned int    PixelPTEReqWidthY[],
+		unsigned int    PixelPTEReqHeightY[],
+		unsigned int    PixelPTEReqWidthC[],
+		unsigned int    PixelPTEReqHeightC[],
+		unsigned int    dpte_row_width_luma_ub[],
+		unsigned int    dpte_row_width_chroma_ub[],
+
+		/* Output */
+		double DST_Y_PER_PTE_ROW_NOM_L[],
+		double DST_Y_PER_PTE_ROW_NOM_C[],
+		double DST_Y_PER_META_ROW_NOM_L[],
+		double DST_Y_PER_META_ROW_NOM_C[],
+		double TimePerMetaChunkNominal[],
+		double TimePerChromaMetaChunkNominal[],
+		double TimePerMetaChunkVBlank[],
+		double TimePerChromaMetaChunkVBlank[],
+		double TimePerMetaChunkFlip[],
+		double TimePerChromaMetaChunkFlip[],
+		double time_per_pte_group_nom_luma[],
+		double time_per_pte_group_vblank_luma[],
+		double time_per_pte_group_flip_luma[],
+		double time_per_pte_group_nom_chroma[],
+		double time_per_pte_group_vblank_chroma[],
+		double time_per_pte_group_flip_chroma[])
+{
+	unsigned int   meta_chunk_width;
+	unsigned int   min_meta_chunk_width;
+	unsigned int   meta_chunk_per_row_int;
+	unsigned int   meta_row_remainder;
+	unsigned int   meta_chunk_threshold;
+	unsigned int   meta_chunks_per_row_ub;
+	unsigned int   meta_chunk_width_chroma;
+	unsigned int   min_meta_chunk_width_chroma;
+	unsigned int   meta_chunk_per_row_int_chroma;
+	unsigned int   meta_row_remainder_chroma;
+	unsigned int   meta_chunk_threshold_chroma;
+	unsigned int   meta_chunks_per_row_ub_chroma;
+	unsigned int   dpte_group_width_luma;
+	unsigned int   dpte_groups_per_row_luma_ub;
+	unsigned int   dpte_group_width_chroma;
+	unsigned int   dpte_groups_per_row_chroma_ub;
+	unsigned int k;
+
+	for (k = 0; k < NumberOfActiveSurfaces; ++k) {
+		DST_Y_PER_PTE_ROW_NOM_L[k] = dpte_row_height[k] / VRatio[k];
+		if (BytePerPixelC[k] == 0)
+			DST_Y_PER_PTE_ROW_NOM_C[k] = 0;
+		else
+			DST_Y_PER_PTE_ROW_NOM_C[k] = dpte_row_height_chroma[k] / VRatioChroma[k];
+		DST_Y_PER_META_ROW_NOM_L[k] = meta_row_height[k] / VRatio[k];
+		if (BytePerPixelC[k] == 0)
+			DST_Y_PER_META_ROW_NOM_C[k] = 0;
+		else
+			DST_Y_PER_META_ROW_NOM_C[k] = meta_row_height_chroma[k] / VRatioChroma[k];
+	}
+
+	for (k = 0; k < NumberOfActiveSurfaces; ++k) {
+		if (DCCEnable[k] == true) {
+			meta_chunk_width = MetaChunkSize * 1024 * 256 / BytePerPixelY[k] / meta_row_height[k];
+			min_meta_chunk_width = MinMetaChunkSizeBytes * 256 / BytePerPixelY[k] / meta_row_height[k];
+			meta_chunk_per_row_int = meta_row_width[k] / meta_chunk_width;
+			meta_row_remainder = meta_row_width[k] % meta_chunk_width;
+			if (!IsVertical(SourceRotation[k]))
+				meta_chunk_threshold = 2 * min_meta_chunk_width - meta_req_width[k];
+			else
+				meta_chunk_threshold = 2 * min_meta_chunk_width - meta_req_height[k];
+
+			if (meta_row_remainder <= meta_chunk_threshold)
+				meta_chunks_per_row_ub = meta_chunk_per_row_int + 1;
+			else
+				meta_chunks_per_row_ub = meta_chunk_per_row_int + 2;
+
+			TimePerMetaChunkNominal[k] = meta_row_height[k] / VRatio[k] *
+					HTotal[k] / PixelClock[k] / meta_chunks_per_row_ub;
+			TimePerMetaChunkVBlank[k] = DestinationLinesToRequestRowInVBlank[k] *
+					HTotal[k] / PixelClock[k] / meta_chunks_per_row_ub;
+			TimePerMetaChunkFlip[k] = DestinationLinesToRequestRowInImmediateFlip[k] *
+					HTotal[k] / PixelClock[k] / meta_chunks_per_row_ub;
+			if (BytePerPixelC[k] == 0) {
+				TimePerChromaMetaChunkNominal[k] = 0;
+				TimePerChromaMetaChunkVBlank[k] = 0;
+				TimePerChromaMetaChunkFlip[k] = 0;
+			} else {
+				meta_chunk_width_chroma = MetaChunkSize * 1024 * 256 / BytePerPixelC[k] /
+						meta_row_height_chroma[k];
+				min_meta_chunk_width_chroma = MinMetaChunkSizeBytes * 256 / BytePerPixelC[k] /
+						meta_row_height_chroma[k];
+				meta_chunk_per_row_int_chroma = (double) meta_row_width_chroma[k] /
+						meta_chunk_width_chroma;
+				meta_row_remainder_chroma = meta_row_width_chroma[k] % meta_chunk_width_chroma;
+				if (!IsVertical(SourceRotation[k])) {
+					meta_chunk_threshold_chroma = 2 * min_meta_chunk_width_chroma -
+							meta_req_width_chroma[k];
+				} else {
+					meta_chunk_threshold_chroma = 2 * min_meta_chunk_width_chroma -
+							meta_req_height_chroma[k];
+				}
+				if (meta_row_remainder_chroma <= meta_chunk_threshold_chroma)
+					meta_chunks_per_row_ub_chroma = meta_chunk_per_row_int_chroma + 1;
+				else
+					meta_chunks_per_row_ub_chroma = meta_chunk_per_row_int_chroma + 2;
+
+				TimePerChromaMetaChunkNominal[k] = meta_row_height_chroma[k] / VRatioChroma[k] *
+						HTotal[k] / PixelClock[k] / meta_chunks_per_row_ub_chroma;
+				TimePerChromaMetaChunkVBlank[k] = DestinationLinesToRequestRowInVBlank[k] *
+						HTotal[k] / PixelClock[k] / meta_chunks_per_row_ub_chroma;
+				TimePerChromaMetaChunkFlip[k] = DestinationLinesToRequestRowInImmediateFlip[k] *
+						HTotal[k] / PixelClock[k] / meta_chunks_per_row_ub_chroma;
+			}
+		} else {
+			TimePerMetaChunkNominal[k] = 0;
+			TimePerMetaChunkVBlank[k] = 0;
+			TimePerMetaChunkFlip[k] = 0;
+			TimePerChromaMetaChunkNominal[k] = 0;
+			TimePerChromaMetaChunkVBlank[k] = 0;
+			TimePerChromaMetaChunkFlip[k] = 0;
+		}
+	}
+
+	for (k = 0; k < NumberOfActiveSurfaces; ++k) {
+		if (GPUVMEnable == true) {
+			if (!IsVertical(SourceRotation[k])) {
+				dpte_group_width_luma = (double) dpte_group_bytes[k] /
+						(double) PTERequestSizeY[k] * PixelPTEReqWidthY[k];
+			} else {
+				dpte_group_width_luma = (double) dpte_group_bytes[k] /
+						(double) PTERequestSizeY[k] * PixelPTEReqHeightY[k];
+			}
+
+			if (use_one_row_for_frame[k]) {
+				dpte_groups_per_row_luma_ub = dml_ceil((double) dpte_row_width_luma_ub[k] /
+						(double) dpte_group_width_luma / 2.0, 1.0);
+			} else {
+				dpte_groups_per_row_luma_ub = dml_ceil((double) dpte_row_width_luma_ub[k] /
+						(double) dpte_group_width_luma, 1.0);
+			}
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: k=%0d, use_one_row_for_frame        = %d\n",
+					__func__, k, use_one_row_for_frame[k]);
+			dml_print("DML::%s: k=%0d, dpte_group_bytes             = %d\n",
+					__func__, k, dpte_group_bytes[k]);
+			dml_print("DML::%s: k=%0d, PTERequestSizeY              = %d\n",
+					__func__, k, PTERequestSizeY[k]);
+			dml_print("DML::%s: k=%0d, PixelPTEReqWidthY            = %d\n",
+					__func__, k, PixelPTEReqWidthY[k]);
+			dml_print("DML::%s: k=%0d, PixelPTEReqHeightY           = %d\n",
+					__func__, k, PixelPTEReqHeightY[k]);
+			dml_print("DML::%s: k=%0d, dpte_row_width_luma_ub       = %d\n",
+					__func__, k, dpte_row_width_luma_ub[k]);
+			dml_print("DML::%s: k=%0d, dpte_group_width_luma        = %d\n",
+					__func__, k, dpte_group_width_luma);
+			dml_print("DML::%s: k=%0d, dpte_groups_per_row_luma_ub  = %d\n",
+					__func__, k, dpte_groups_per_row_luma_ub);
+#endif
+
+			time_per_pte_group_nom_luma[k] = DST_Y_PER_PTE_ROW_NOM_L[k] *
+					HTotal[k] / PixelClock[k] / dpte_groups_per_row_luma_ub;
+			time_per_pte_group_vblank_luma[k] = DestinationLinesToRequestRowInVBlank[k] *
+					HTotal[k] / PixelClock[k] / dpte_groups_per_row_luma_ub;
+			time_per_pte_group_flip_luma[k] = DestinationLinesToRequestRowInImmediateFlip[k] *
+					HTotal[k] / PixelClock[k] / dpte_groups_per_row_luma_ub;
+			if (BytePerPixelC[k] == 0) {
+				time_per_pte_group_nom_chroma[k] = 0;
+				time_per_pte_group_vblank_chroma[k] = 0;
+				time_per_pte_group_flip_chroma[k] = 0;
+			} else {
+				if (!IsVertical(SourceRotation[k])) {
+					dpte_group_width_chroma = (double) dpte_group_bytes[k] /
+							(double) PTERequestSizeC[k] * PixelPTEReqWidthC[k];
+				} else {
+					dpte_group_width_chroma = (double) dpte_group_bytes[k] /
+							(double) PTERequestSizeC[k] * PixelPTEReqHeightC[k];
+				}
+
+				if (use_one_row_for_frame[k]) {
+					dpte_groups_per_row_chroma_ub = dml_ceil((double) dpte_row_width_chroma_ub[k] /
+							(double) dpte_group_width_chroma / 2.0, 1.0);
+				} else {
+					dpte_groups_per_row_chroma_ub = dml_ceil((double) dpte_row_width_chroma_ub[k] /
+							(double) dpte_group_width_chroma, 1.0);
+				}
+#ifdef __DML_VBA_DEBUG__
+				dml_print("DML::%s: k=%0d, dpte_row_width_chroma_ub        = %d\n",
+						__func__, k, dpte_row_width_chroma_ub[k]);
+				dml_print("DML::%s: k=%0d, dpte_group_width_chroma        = %d\n",
+						__func__, k, dpte_group_width_chroma);
+				dml_print("DML::%s: k=%0d, dpte_groups_per_row_chroma_ub  = %d\n",
+						__func__, k, dpte_groups_per_row_chroma_ub);
+#endif
+				time_per_pte_group_nom_chroma[k] = DST_Y_PER_PTE_ROW_NOM_C[k] *
+						HTotal[k] / PixelClock[k] / dpte_groups_per_row_chroma_ub;
+				time_per_pte_group_vblank_chroma[k] = DestinationLinesToRequestRowInVBlank[k] *
+						HTotal[k] / PixelClock[k] / dpte_groups_per_row_chroma_ub;
+				time_per_pte_group_flip_chroma[k] = DestinationLinesToRequestRowInImmediateFlip[k] *
+						HTotal[k] / PixelClock[k] / dpte_groups_per_row_chroma_ub;
+			}
+		} else {
+			time_per_pte_group_nom_luma[k] = 0;
+			time_per_pte_group_vblank_luma[k] = 0;
+			time_per_pte_group_flip_luma[k] = 0;
+			time_per_pte_group_nom_chroma[k] = 0;
+			time_per_pte_group_vblank_chroma[k] = 0;
+			time_per_pte_group_flip_chroma[k] = 0;
+		}
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%0d, DestinationLinesToRequestRowInVBlank         = %f\n",
+				__func__, k, DestinationLinesToRequestRowInVBlank[k]);
+		dml_print("DML::%s: k=%0d, DestinationLinesToRequestRowInImmediateFlip  = %f\n",
+				__func__, k, DestinationLinesToRequestRowInImmediateFlip[k]);
+		dml_print("DML::%s: k=%0d, DST_Y_PER_PTE_ROW_NOM_L                      = %f\n",
+				__func__, k, DST_Y_PER_PTE_ROW_NOM_L[k]);
+		dml_print("DML::%s: k=%0d, DST_Y_PER_PTE_ROW_NOM_C                      = %f\n",
+				__func__, k, DST_Y_PER_PTE_ROW_NOM_C[k]);
+		dml_print("DML::%s: k=%0d, DST_Y_PER_META_ROW_NOM_L                     = %f\n",
+				__func__, k, DST_Y_PER_META_ROW_NOM_L[k]);
+		dml_print("DML::%s: k=%0d, DST_Y_PER_META_ROW_NOM_C                     = %f\n",
+				__func__, k, DST_Y_PER_META_ROW_NOM_C[k]);
+		dml_print("DML::%s: k=%0d, TimePerMetaChunkNominal          = %f\n",
+				__func__, k, TimePerMetaChunkNominal[k]);
+		dml_print("DML::%s: k=%0d, TimePerMetaChunkVBlank           = %f\n",
+				__func__, k, TimePerMetaChunkVBlank[k]);
+		dml_print("DML::%s: k=%0d, TimePerMetaChunkFlip             = %f\n",
+				__func__, k, TimePerMetaChunkFlip[k]);
+		dml_print("DML::%s: k=%0d, TimePerChromaMetaChunkNominal    = %f\n",
+				__func__, k, TimePerChromaMetaChunkNominal[k]);
+		dml_print("DML::%s: k=%0d, TimePerChromaMetaChunkVBlank     = %f\n",
+				__func__, k, TimePerChromaMetaChunkVBlank[k]);
+		dml_print("DML::%s: k=%0d, TimePerChromaMetaChunkFlip       = %f\n",
+				__func__, k, TimePerChromaMetaChunkFlip[k]);
+		dml_print("DML::%s: k=%0d, time_per_pte_group_nom_luma      = %f\n",
+				__func__, k, time_per_pte_group_nom_luma[k]);
+		dml_print("DML::%s: k=%0d, time_per_pte_group_vblank_luma   = %f\n",
+				__func__, k, time_per_pte_group_vblank_luma[k]);
+		dml_print("DML::%s: k=%0d, time_per_pte_group_flip_luma     = %f\n",
+				__func__, k, time_per_pte_group_flip_luma[k]);
+		dml_print("DML::%s: k=%0d, time_per_pte_group_nom_chroma    = %f\n",
+				__func__, k, time_per_pte_group_nom_chroma[k]);
+		dml_print("DML::%s: k=%0d, time_per_pte_group_vblank_chroma = %f\n",
+				__func__, k, time_per_pte_group_vblank_chroma[k]);
+		dml_print("DML::%s: k=%0d, time_per_pte_group_flip_chroma   = %f\n",
+				__func__, k, time_per_pte_group_flip_chroma[k]);
+#endif
+	}
+} // CalculateMetaAndPTETimes
+
+void dml32_CalculateVMGroupAndRequestTimes(
+		unsigned int     NumberOfActiveSurfaces,
+		bool     GPUVMEnable,
+		unsigned int     GPUVMMaxPageTableLevels,
+		unsigned int     HTotal[],
+		unsigned int     BytePerPixelC[],
+		double      DestinationLinesToRequestVMInVBlank[],
+		double      DestinationLinesToRequestVMInImmediateFlip[],
+		bool     DCCEnable[],
+		double      PixelClock[],
+		unsigned int        dpte_row_width_luma_ub[],
+		unsigned int        dpte_row_width_chroma_ub[],
+		unsigned int     vm_group_bytes[],
+		unsigned int     dpde0_bytes_per_frame_ub_l[],
+		unsigned int     dpde0_bytes_per_frame_ub_c[],
+		unsigned int        meta_pte_bytes_per_frame_ub_l[],
+		unsigned int        meta_pte_bytes_per_frame_ub_c[],
+
+		/* Output */
+		double      TimePerVMGroupVBlank[],
+		double      TimePerVMGroupFlip[],
+		double      TimePerVMRequestVBlank[],
+		double      TimePerVMRequestFlip[])
+{
+	unsigned int k;
+	unsigned int   num_group_per_lower_vm_stage;
+	unsigned int   num_req_per_lower_vm_stage;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: NumberOfActiveSurfaces = %d\n", __func__, NumberOfActiveSurfaces);
+	dml_print("DML::%s: GPUVMEnable = %d\n", __func__, GPUVMEnable);
+#endif
+	for (k = 0; k < NumberOfActiveSurfaces; ++k) {
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%0d, DCCEnable = %d\n", __func__, k, DCCEnable[k]);
+		dml_print("DML::%s: k=%0d, vm_group_bytes = %d\n", __func__, k, vm_group_bytes[k]);
+		dml_print("DML::%s: k=%0d, dpde0_bytes_per_frame_ub_l = %d\n",
+				__func__, k, dpde0_bytes_per_frame_ub_l[k]);
+		dml_print("DML::%s: k=%0d, dpde0_bytes_per_frame_ub_c = %d\n",
+				__func__, k, dpde0_bytes_per_frame_ub_c[k]);
+		dml_print("DML::%s: k=%0d, meta_pte_bytes_per_frame_ub_l = %d\n",
+				__func__, k, meta_pte_bytes_per_frame_ub_l[k]);
+		dml_print("DML::%s: k=%0d, meta_pte_bytes_per_frame_ub_c = %d\n",
+				__func__, k, meta_pte_bytes_per_frame_ub_c[k]);
+#endif
+
+		if (GPUVMEnable == true && (DCCEnable[k] == true || GPUVMMaxPageTableLevels > 1)) {
+			if (DCCEnable[k] == false) {
+				if (BytePerPixelC[k] > 0) {
+					num_group_per_lower_vm_stage = dml_ceil(
+							(double) (dpde0_bytes_per_frame_ub_l[k]) /
+							(double) (vm_group_bytes[k]), 1.0) +
+							dml_ceil((double) (dpde0_bytes_per_frame_ub_c[k]) /
+							(double) (vm_group_bytes[k]), 1.0);
+				} else {
+					num_group_per_lower_vm_stage = dml_ceil(
+							(double) (dpde0_bytes_per_frame_ub_l[k]) /
+							(double) (vm_group_bytes[k]), 1.0);
+				}
+			} else {
+				if (GPUVMMaxPageTableLevels == 1) {
+					if (BytePerPixelC[k] > 0) {
+						num_group_per_lower_vm_stage = dml_ceil(
+							(double) (meta_pte_bytes_per_frame_ub_l[k]) /
+							(double) (vm_group_bytes[k]), 1.0) +
+							dml_ceil((double) (meta_pte_bytes_per_frame_ub_c[k]) /
+							(double) (vm_group_bytes[k]), 1.0);
+					} else {
+						num_group_per_lower_vm_stage = dml_ceil(
+								(double) (meta_pte_bytes_per_frame_ub_l[k]) /
+								(double) (vm_group_bytes[k]), 1.0);
+					}
+				} else {
+					if (BytePerPixelC[k] > 0) {
+						num_group_per_lower_vm_stage = 2 + dml_ceil(
+							(double) (dpde0_bytes_per_frame_ub_l[k]) /
+							(double) (vm_group_bytes[k]), 1) +
+							dml_ceil((double) (dpde0_bytes_per_frame_ub_c[k]) /
+							(double) (vm_group_bytes[k]), 1) +
+							dml_ceil((double) (meta_pte_bytes_per_frame_ub_l[k]) /
+							(double) (vm_group_bytes[k]), 1) +
+							dml_ceil((double) (meta_pte_bytes_per_frame_ub_c[k]) /
+							(double) (vm_group_bytes[k]), 1);
+					} else {
+						num_group_per_lower_vm_stage = 1 + dml_ceil(
+							(double) (dpde0_bytes_per_frame_ub_l[k]) /
+							(double) (vm_group_bytes[k]), 1) + dml_ceil(
+							(double) (meta_pte_bytes_per_frame_ub_l[k]) /
+							(double) (vm_group_bytes[k]), 1);
+					}
+				}
+			}
+
+			if (DCCEnable[k] == false) {
+				if (BytePerPixelC[k] > 0) {
+					num_req_per_lower_vm_stage = dpde0_bytes_per_frame_ub_l[k] / 64 +
+							dpde0_bytes_per_frame_ub_c[k] / 64;
+				} else {
+					num_req_per_lower_vm_stage = dpde0_bytes_per_frame_ub_l[k] / 64;
+				}
+			} else {
+				if (GPUVMMaxPageTableLevels == 1) {
+					if (BytePerPixelC[k] > 0) {
+						num_req_per_lower_vm_stage = meta_pte_bytes_per_frame_ub_l[k] / 64 +
+								meta_pte_bytes_per_frame_ub_c[k] / 64;
+					} else {
+						num_req_per_lower_vm_stage = meta_pte_bytes_per_frame_ub_l[k] / 64;
+					}
+				} else {
+					if (BytePerPixelC[k] > 0) {
+						num_req_per_lower_vm_stage = dpde0_bytes_per_frame_ub_l[k] /
+								64 + dpde0_bytes_per_frame_ub_c[k] / 64 +
+								meta_pte_bytes_per_frame_ub_l[k] / 64 +
+								meta_pte_bytes_per_frame_ub_c[k] / 64;
+					} else {
+						num_req_per_lower_vm_stage = dpde0_bytes_per_frame_ub_l[k] /
+								64 + meta_pte_bytes_per_frame_ub_l[k] / 64;
+					}
+				}
+			}
+
+			TimePerVMGroupVBlank[k] = DestinationLinesToRequestVMInVBlank[k] *
+					HTotal[k] / PixelClock[k] / num_group_per_lower_vm_stage;
+			TimePerVMGroupFlip[k] = DestinationLinesToRequestVMInImmediateFlip[k] *
+					HTotal[k] / PixelClock[k] / num_group_per_lower_vm_stage;
+			TimePerVMRequestVBlank[k] = DestinationLinesToRequestVMInVBlank[k] *
+					HTotal[k] / PixelClock[k] / num_req_per_lower_vm_stage;
+			TimePerVMRequestFlip[k] = DestinationLinesToRequestVMInImmediateFlip[k] *
+					HTotal[k] / PixelClock[k] / num_req_per_lower_vm_stage;
+
+			if (GPUVMMaxPageTableLevels > 2) {
+				TimePerVMGroupVBlank[k]    = TimePerVMGroupVBlank[k] / 2;
+				TimePerVMGroupFlip[k]      = TimePerVMGroupFlip[k] / 2;
+				TimePerVMRequestVBlank[k]  = TimePerVMRequestVBlank[k] / 2;
+				TimePerVMRequestFlip[k]    = TimePerVMRequestFlip[k] / 2;
+			}
+
+		} else {
+			TimePerVMGroupVBlank[k] = 0;
+			TimePerVMGroupFlip[k] = 0;
+			TimePerVMRequestVBlank[k] = 0;
+			TimePerVMRequestFlip[k] = 0;
+		}
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%0d, TimePerVMGroupVBlank = %f\n", __func__, k, TimePerVMGroupVBlank[k]);
+		dml_print("DML::%s: k=%0d, TimePerVMGroupFlip = %f\n", __func__, k, TimePerVMGroupFlip[k]);
+		dml_print("DML::%s: k=%0d, TimePerVMRequestVBlank = %f\n", __func__, k, TimePerVMRequestVBlank[k]);
+		dml_print("DML::%s: k=%0d, TimePerVMRequestFlip = %f\n", __func__, k, TimePerVMRequestFlip[k]);
+#endif
+	}
+} // CalculateVMGroupAndRequestTimes
+
+void dml32_CalculateDCCConfiguration(
+		bool             DCCEnabled,
+		bool             DCCProgrammingAssumesScanDirectionUnknown,
+		enum source_format_class SourcePixelFormat,
+		unsigned int             SurfaceWidthLuma,
+		unsigned int             SurfaceWidthChroma,
+		unsigned int             SurfaceHeightLuma,
+		unsigned int             SurfaceHeightChroma,
+		unsigned int                nomDETInKByte,
+		unsigned int             RequestHeight256ByteLuma,
+		unsigned int             RequestHeight256ByteChroma,
+		enum dm_swizzle_mode     TilingFormat,
+		unsigned int             BytePerPixelY,
+		unsigned int             BytePerPixelC,
+		double              BytePerPixelDETY,
+		double              BytePerPixelDETC,
+		enum dm_rotation_angle   SourceRotation,
+		/* Output */
+		unsigned int        *MaxUncompressedBlockLuma,
+		unsigned int        *MaxUncompressedBlockChroma,
+		unsigned int        *MaxCompressedBlockLuma,
+		unsigned int        *MaxCompressedBlockChroma,
+		unsigned int        *IndependentBlockLuma,
+		unsigned int        *IndependentBlockChroma)
+{
+	typedef enum {
+		REQ_256Bytes,
+		REQ_128BytesNonContiguous,
+		REQ_128BytesContiguous,
+		REQ_NA
+	} RequestType;
+
+	RequestType   RequestLuma;
+	RequestType   RequestChroma;
+
+	unsigned int   segment_order_horz_contiguous_luma;
+	unsigned int   segment_order_horz_contiguous_chroma;
+	unsigned int   segment_order_vert_contiguous_luma;
+	unsigned int   segment_order_vert_contiguous_chroma;
+	unsigned int req128_horz_wc_l;
+	unsigned int req128_horz_wc_c;
+	unsigned int req128_vert_wc_l;
+	unsigned int req128_vert_wc_c;
+	unsigned int MAS_vp_horz_limit;
+	unsigned int MAS_vp_vert_limit;
+	unsigned int max_vp_horz_width;
+	unsigned int max_vp_vert_height;
+	unsigned int eff_surf_width_l;
+	unsigned int eff_surf_width_c;
+	unsigned int eff_surf_height_l;
+	unsigned int eff_surf_height_c;
+	unsigned int full_swath_bytes_horz_wc_l;
+	unsigned int full_swath_bytes_horz_wc_c;
+	unsigned int full_swath_bytes_vert_wc_l;
+	unsigned int full_swath_bytes_vert_wc_c;
+	unsigned int DETBufferSizeForDCC = nomDETInKByte * 1024;
+
+	unsigned int   yuv420;
+	unsigned int   horz_div_l;
+	unsigned int   horz_div_c;
+	unsigned int   vert_div_l;
+	unsigned int   vert_div_c;
+
+	unsigned int     swath_buf_size;
+	double   detile_buf_vp_horz_limit;
+	double   detile_buf_vp_vert_limit;
+
+	yuv420 = ((SourcePixelFormat == dm_420_8 || SourcePixelFormat == dm_420_10 ||
+			SourcePixelFormat == dm_420_12) ? 1 : 0);
+	horz_div_l = 1;
+	horz_div_c = 1;
+	vert_div_l = 1;
+	vert_div_c = 1;
+
+	if (BytePerPixelY == 1)
+		vert_div_l = 0;
+	if (BytePerPixelC == 1)
+		vert_div_c = 0;
+
+	if (BytePerPixelC == 0) {
+		swath_buf_size = DETBufferSizeForDCC / 2 - 2 * 256;
+		detile_buf_vp_horz_limit = (double) swath_buf_size / ((double) RequestHeight256ByteLuma *
+				BytePerPixelY / (1 + horz_div_l));
+		detile_buf_vp_vert_limit = (double) swath_buf_size / (256.0 / RequestHeight256ByteLuma /
+				(1 + vert_div_l));
+	} else {
+		swath_buf_size = DETBufferSizeForDCC / 2 - 2 * 2 * 256;
+		detile_buf_vp_horz_limit = (double) swath_buf_size / ((double) RequestHeight256ByteLuma *
+				BytePerPixelY / (1 + horz_div_l) + (double) RequestHeight256ByteChroma *
+				BytePerPixelC / (1 + horz_div_c) / (1 + yuv420));
+		detile_buf_vp_vert_limit = (double) swath_buf_size / (256.0 / RequestHeight256ByteLuma /
+				(1 + vert_div_l) + 256.0 / RequestHeight256ByteChroma /
+				(1 + vert_div_c) / (1 + yuv420));
+	}
+
+	if (SourcePixelFormat == dm_420_10) {
+		detile_buf_vp_horz_limit = 1.5 * detile_buf_vp_horz_limit;
+		detile_buf_vp_vert_limit = 1.5 * detile_buf_vp_vert_limit;
+	}
+
+	detile_buf_vp_horz_limit = dml_floor(detile_buf_vp_horz_limit - 1, 16);
+	detile_buf_vp_vert_limit = dml_floor(detile_buf_vp_vert_limit - 1, 16);
+
+	MAS_vp_horz_limit = SourcePixelFormat == dm_rgbe_alpha ? 3840 : 6144;
+	MAS_vp_vert_limit = SourcePixelFormat == dm_rgbe_alpha ? 3840 : (BytePerPixelY == 8 ? 3072 : 6144);
+	max_vp_horz_width = dml_min((double) MAS_vp_horz_limit, detile_buf_vp_horz_limit);
+	max_vp_vert_height = dml_min((double) MAS_vp_vert_limit, detile_buf_vp_vert_limit);
+	eff_surf_width_l =  (SurfaceWidthLuma > max_vp_horz_width ? max_vp_horz_width : SurfaceWidthLuma);
+	eff_surf_width_c =  eff_surf_width_l / (1 + yuv420);
+	eff_surf_height_l =  (SurfaceHeightLuma > max_vp_vert_height ? max_vp_vert_height : SurfaceHeightLuma);
+	eff_surf_height_c =  eff_surf_height_l / (1 + yuv420);
+
+	full_swath_bytes_horz_wc_l = eff_surf_width_l * RequestHeight256ByteLuma * BytePerPixelY;
+	full_swath_bytes_vert_wc_l = eff_surf_height_l * 256 / RequestHeight256ByteLuma;
+	if (BytePerPixelC > 0) {
+		full_swath_bytes_horz_wc_c = eff_surf_width_c * RequestHeight256ByteChroma * BytePerPixelC;
+		full_swath_bytes_vert_wc_c = eff_surf_height_c * 256 / RequestHeight256ByteChroma;
+	} else {
+		full_swath_bytes_horz_wc_c = 0;
+		full_swath_bytes_vert_wc_c = 0;
+	}
+
+	if (SourcePixelFormat == dm_420_10) {
+		full_swath_bytes_horz_wc_l = dml_ceil((double) full_swath_bytes_horz_wc_l * 2.0 / 3.0, 256.0);
+		full_swath_bytes_horz_wc_c = dml_ceil((double) full_swath_bytes_horz_wc_c * 2.0 / 3.0, 256.0);
+		full_swath_bytes_vert_wc_l = dml_ceil((double) full_swath_bytes_vert_wc_l * 2.0 / 3.0, 256.0);
+		full_swath_bytes_vert_wc_c = dml_ceil((double) full_swath_bytes_vert_wc_c * 2.0 / 3.0, 256.0);
+	}
+
+	if (2 * full_swath_bytes_horz_wc_l + 2 * full_swath_bytes_horz_wc_c <= DETBufferSizeForDCC) {
+		req128_horz_wc_l = 0;
+		req128_horz_wc_c = 0;
+	} else if (full_swath_bytes_horz_wc_l < 1.5 * full_swath_bytes_horz_wc_c && 2 * full_swath_bytes_horz_wc_l +
+			full_swath_bytes_horz_wc_c <= DETBufferSizeForDCC) {
+		req128_horz_wc_l = 0;
+		req128_horz_wc_c = 1;
+	} else if (full_swath_bytes_horz_wc_l >= 1.5 * full_swath_bytes_horz_wc_c && full_swath_bytes_horz_wc_l + 2 *
+			full_swath_bytes_horz_wc_c <= DETBufferSizeForDCC) {
+		req128_horz_wc_l = 1;
+		req128_horz_wc_c = 0;
+	} else {
+		req128_horz_wc_l = 1;
+		req128_horz_wc_c = 1;
+	}
+
+	if (2 * full_swath_bytes_vert_wc_l + 2 * full_swath_bytes_vert_wc_c <= DETBufferSizeForDCC) {
+		req128_vert_wc_l = 0;
+		req128_vert_wc_c = 0;
+	} else if (full_swath_bytes_vert_wc_l < 1.5 * full_swath_bytes_vert_wc_c && 2 *
+			full_swath_bytes_vert_wc_l + full_swath_bytes_vert_wc_c <= DETBufferSizeForDCC) {
+		req128_vert_wc_l = 0;
+		req128_vert_wc_c = 1;
+	} else if (full_swath_bytes_vert_wc_l >= 1.5 * full_swath_bytes_vert_wc_c &&
+			full_swath_bytes_vert_wc_l + 2 * full_swath_bytes_vert_wc_c <= DETBufferSizeForDCC) {
+		req128_vert_wc_l = 1;
+		req128_vert_wc_c = 0;
+	} else {
+		req128_vert_wc_l = 1;
+		req128_vert_wc_c = 1;
+	}
+
+	if (BytePerPixelY == 2) {
+		segment_order_horz_contiguous_luma = 0;
+		segment_order_vert_contiguous_luma = 1;
+	} else {
+		segment_order_horz_contiguous_luma = 1;
+		segment_order_vert_contiguous_luma = 0;
+	}
+
+	if (BytePerPixelC == 2) {
+		segment_order_horz_contiguous_chroma = 0;
+		segment_order_vert_contiguous_chroma = 1;
+	} else {
+		segment_order_horz_contiguous_chroma = 1;
+		segment_order_vert_contiguous_chroma = 0;
+	}
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: DCCEnabled = %d\n", __func__, DCCEnabled);
+	dml_print("DML::%s: nomDETInKByte = %d\n", __func__, nomDETInKByte);
+	dml_print("DML::%s: DETBufferSizeForDCC = %d\n", __func__, DETBufferSizeForDCC);
+	dml_print("DML::%s: req128_horz_wc_l = %d\n", __func__, req128_horz_wc_l);
+	dml_print("DML::%s: req128_horz_wc_c = %d\n", __func__, req128_horz_wc_c);
+	dml_print("DML::%s: full_swath_bytes_horz_wc_l = %d\n", __func__, full_swath_bytes_horz_wc_l);
+	dml_print("DML::%s: full_swath_bytes_vert_wc_c = %d\n", __func__, full_swath_bytes_vert_wc_c);
+	dml_print("DML::%s: segment_order_horz_contiguous_luma = %d\n", __func__, segment_order_horz_contiguous_luma);
+	dml_print("DML::%s: segment_order_horz_contiguous_chroma = %d\n",
+			__func__, segment_order_horz_contiguous_chroma);
+#endif
+
+	if (DCCProgrammingAssumesScanDirectionUnknown == true) {
+		if (req128_horz_wc_l == 0 && req128_vert_wc_l == 0)
+			RequestLuma = REQ_256Bytes;
+		else if ((req128_horz_wc_l == 1 && segment_order_horz_contiguous_luma == 0) ||
+				(req128_vert_wc_l == 1 && segment_order_vert_contiguous_luma == 0))
+			RequestLuma = REQ_128BytesNonContiguous;
+		else
+			RequestLuma = REQ_128BytesContiguous;
+
+		if (req128_horz_wc_c == 0 && req128_vert_wc_c == 0)
+			RequestChroma = REQ_256Bytes;
+		else if ((req128_horz_wc_c == 1 && segment_order_horz_contiguous_chroma == 0) ||
+				(req128_vert_wc_c == 1 && segment_order_vert_contiguous_chroma == 0))
+			RequestChroma = REQ_128BytesNonContiguous;
+		else
+			RequestChroma = REQ_128BytesContiguous;
+
+	} else if (!IsVertical(SourceRotation)) {
+		if (req128_horz_wc_l == 0)
+			RequestLuma = REQ_256Bytes;
+		else if (segment_order_horz_contiguous_luma == 0)
+			RequestLuma = REQ_128BytesNonContiguous;
+		else
+			RequestLuma = REQ_128BytesContiguous;
+
+		if (req128_horz_wc_c == 0)
+			RequestChroma = REQ_256Bytes;
+		else if (segment_order_horz_contiguous_chroma == 0)
+			RequestChroma = REQ_128BytesNonContiguous;
+		else
+			RequestChroma = REQ_128BytesContiguous;
+
+	} else {
+		if (req128_vert_wc_l == 0)
+			RequestLuma = REQ_256Bytes;
+		else if (segment_order_vert_contiguous_luma == 0)
+			RequestLuma = REQ_128BytesNonContiguous;
+		else
+			RequestLuma = REQ_128BytesContiguous;
+
+		if (req128_vert_wc_c == 0)
+			RequestChroma = REQ_256Bytes;
+		else if (segment_order_vert_contiguous_chroma == 0)
+			RequestChroma = REQ_128BytesNonContiguous;
+		else
+			RequestChroma = REQ_128BytesContiguous;
+	}
+
+	if (RequestLuma == REQ_256Bytes) {
+		*MaxUncompressedBlockLuma = 256;
+		*MaxCompressedBlockLuma = 256;
+		*IndependentBlockLuma = 0;
+	} else if (RequestLuma == REQ_128BytesContiguous) {
+		*MaxUncompressedBlockLuma = 256;
+		*MaxCompressedBlockLuma = 128;
+		*IndependentBlockLuma = 128;
+	} else {
+		*MaxUncompressedBlockLuma = 256;
+		*MaxCompressedBlockLuma = 64;
+		*IndependentBlockLuma = 64;
+	}
+
+	if (RequestChroma == REQ_256Bytes) {
+		*MaxUncompressedBlockChroma = 256;
+		*MaxCompressedBlockChroma = 256;
+		*IndependentBlockChroma = 0;
+	} else if (RequestChroma == REQ_128BytesContiguous) {
+		*MaxUncompressedBlockChroma = 256;
+		*MaxCompressedBlockChroma = 128;
+		*IndependentBlockChroma = 128;
+	} else {
+		*MaxUncompressedBlockChroma = 256;
+		*MaxCompressedBlockChroma = 64;
+		*IndependentBlockChroma = 64;
+	}
+
+	if (DCCEnabled != true || BytePerPixelC == 0) {
+		*MaxUncompressedBlockChroma = 0;
+		*MaxCompressedBlockChroma = 0;
+		*IndependentBlockChroma = 0;
+	}
+
+	if (DCCEnabled != true) {
+		*MaxUncompressedBlockLuma = 0;
+		*MaxCompressedBlockLuma = 0;
+		*IndependentBlockLuma = 0;
+	}
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: MaxUncompressedBlockLuma = %d\n", __func__, *MaxUncompressedBlockLuma);
+	dml_print("DML::%s: MaxCompressedBlockLuma = %d\n", __func__, *MaxCompressedBlockLuma);
+	dml_print("DML::%s: IndependentBlockLuma = %d\n", __func__, *IndependentBlockLuma);
+	dml_print("DML::%s: MaxUncompressedBlockChroma = %d\n", __func__, *MaxUncompressedBlockChroma);
+	dml_print("DML::%s: MaxCompressedBlockChroma = %d\n", __func__, *MaxCompressedBlockChroma);
+	dml_print("DML::%s: IndependentBlockChroma = %d\n", __func__, *IndependentBlockChroma);
+#endif
+
+} // CalculateDCCConfiguration
+
+void dml32_CalculateStutterEfficiency(
+		unsigned int      CompressedBufferSizeInkByte,
+		enum dm_use_mall_for_pstate_change_mode UseMALLForPStateChange[],
+		bool   UnboundedRequestEnabled,
+		unsigned int      MetaFIFOSizeInKEntries,
+		unsigned int      ZeroSizeBufferEntries,
+		unsigned int      PixelChunkSizeInKByte,
+		unsigned int   NumberOfActiveSurfaces,
+		unsigned int      ROBBufferSizeInKByte,
+		double    TotalDataReadBandwidth,
+		double    DCFCLK,
+		double    ReturnBW,
+		unsigned int      CompbufReservedSpace64B,
+		unsigned int      CompbufReservedSpaceZs,
+		double    SRExitTime,
+		double    SRExitZ8Time,
+		bool   SynchronizeTimingsFinal,
+		unsigned int   BlendingAndTiming[],
+		double    StutterEnterPlusExitWatermark,
+		double    Z8StutterEnterPlusExitWatermark,
+		bool   ProgressiveToInterlaceUnitInOPP,
+		bool   Interlace[],
+		double    MinTTUVBlank[],
+		unsigned int   DPPPerSurface[],
+		unsigned int      DETBufferSizeY[],
+		unsigned int   BytePerPixelY[],
+		double    BytePerPixelDETY[],
+		double      SwathWidthY[],
+		unsigned int   SwathHeightY[],
+		unsigned int   SwathHeightC[],
+		double    NetDCCRateLuma[],
+		double    NetDCCRateChroma[],
+		double    DCCFractionOfZeroSizeRequestsLuma[],
+		double    DCCFractionOfZeroSizeRequestsChroma[],
+		unsigned int      HTotal[],
+		unsigned int      VTotal[],
+		double    PixelClock[],
+		double    VRatio[],
+		enum dm_rotation_angle SourceRotation[],
+		unsigned int   BlockHeight256BytesY[],
+		unsigned int   BlockWidth256BytesY[],
+		unsigned int   BlockHeight256BytesC[],
+		unsigned int   BlockWidth256BytesC[],
+		unsigned int   DCCYMaxUncompressedBlock[],
+		unsigned int   DCCCMaxUncompressedBlock[],
+		unsigned int      VActive[],
+		bool   DCCEnable[],
+		bool   WritebackEnable[],
+		double    ReadBandwidthSurfaceLuma[],
+		double    ReadBandwidthSurfaceChroma[],
+		double    meta_row_bw[],
+		double    dpte_row_bw[],
+
+		/* Output */
+		double   *StutterEfficiencyNotIncludingVBlank,
+		double   *StutterEfficiency,
+		unsigned int     *NumberOfStutterBurstsPerFrame,
+		double   *Z8StutterEfficiencyNotIncludingVBlank,
+		double   *Z8StutterEfficiency,
+		unsigned int     *Z8NumberOfStutterBurstsPerFrame,
+		double   *StutterPeriod,
+		bool  *DCHUBBUB_ARB_CSTATE_MAX_CAP_MODE)
+{
+
+	bool FoundCriticalSurface = false;
+	unsigned int SwathSizeCriticalSurface = 0;
+	unsigned int LastChunkOfSwathSize;
+	unsigned int MissingPartOfLastSwathOfDETSize;
+	double LastZ8StutterPeriod = 0.0;
+	double LastStutterPeriod = 0.0;
+	unsigned int TotalNumberOfActiveOTG = 0;
+	double doublePixelClock = 0;
+	unsigned int doubleHTotal = 0;
+	unsigned int doubleVTotal = 0;
+	bool SameTiming = true;
+	double DETBufferingTimeY;
+	double SwathWidthYCriticalSurface = 0.0;
+	double SwathHeightYCriticalSurface = 0.0;
+	double VActiveTimeCriticalSurface = 0.0;
+	double FrameTimeCriticalSurface = 0.0;
+	unsigned int BytePerPixelYCriticalSurface = 0;
+	double LinesToFinishSwathTransferStutterCriticalSurface = 0.0;
+	unsigned int DETBufferSizeYCriticalSurface = 0;
+	double MinTTUVBlankCriticalSurface = 0.0;
+	unsigned int BlockWidth256BytesYCriticalSurface = 0;
+	bool doublePlaneCriticalSurface = 0;
+	bool doublePipeCriticalSurface = 0;
+	double TotalCompressedReadBandwidth;
+	double TotalRowReadBandwidth;
+	double AverageDCCCompressionRate;
+	double EffectiveCompressedBufferSize;
+	double PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer;
+	double StutterBurstTime;
+	unsigned int TotalActiveWriteback;
+	double LinesInDETY;
+	double LinesInDETYRoundedDownToSwath;
+	double MaximumEffectiveCompressionLuma;
+	double MaximumEffectiveCompressionChroma;
+	double TotalZeroSizeRequestReadBandwidth;
+	double TotalZeroSizeCompressedReadBandwidth;
+	double AverageDCCZeroSizeFraction;
+	double AverageZeroSizeCompressionRate;
+	unsigned int k;
+
+	TotalZeroSizeRequestReadBandwidth = 0;
+	TotalZeroSizeCompressedReadBandwidth = 0;
+	TotalRowReadBandwidth = 0;
+	TotalCompressedReadBandwidth = 0;
+
+	for (k = 0; k < NumberOfActiveSurfaces; ++k) {
+		if (UseMALLForPStateChange[k] != dm_use_mall_pstate_change_phantom_pipe) {
+			if (DCCEnable[k] == true) {
+				if ((IsVertical(SourceRotation[k]) && BlockWidth256BytesY[k] > SwathHeightY[k])
+						|| (!IsVertical(SourceRotation[k])
+								&& BlockHeight256BytesY[k] > SwathHeightY[k])
+						|| DCCYMaxUncompressedBlock[k] < 256) {
+					MaximumEffectiveCompressionLuma = 2;
+				} else {
+					MaximumEffectiveCompressionLuma = 4;
+				}
+				TotalCompressedReadBandwidth = TotalCompressedReadBandwidth
+						+ ReadBandwidthSurfaceLuma[k]
+								/ dml_min(NetDCCRateLuma[k],
+										MaximumEffectiveCompressionLuma);
+#ifdef __DML_VBA_DEBUG__
+				dml_print("DML::%s: k=%0d, ReadBandwidthSurfaceLuma = %f\n",
+						__func__, k, ReadBandwidthSurfaceLuma[k]);
+				dml_print("DML::%s: k=%0d, NetDCCRateLuma = %f\n",
+						__func__, k, NetDCCRateLuma[k]);
+				dml_print("DML::%s: k=%0d, MaximumEffectiveCompressionLuma = %f\n",
+						__func__, k, MaximumEffectiveCompressionLuma);
+#endif
+				TotalZeroSizeRequestReadBandwidth = TotalZeroSizeRequestReadBandwidth
+						+ ReadBandwidthSurfaceLuma[k] * DCCFractionOfZeroSizeRequestsLuma[k];
+				TotalZeroSizeCompressedReadBandwidth = TotalZeroSizeCompressedReadBandwidth
+						+ ReadBandwidthSurfaceLuma[k] * DCCFractionOfZeroSizeRequestsLuma[k]
+								/ MaximumEffectiveCompressionLuma;
+
+				if (ReadBandwidthSurfaceChroma[k] > 0) {
+					if ((IsVertical(SourceRotation[k]) && BlockWidth256BytesC[k] > SwathHeightC[k])
+							|| (!IsVertical(SourceRotation[k])
+									&& BlockHeight256BytesC[k] > SwathHeightC[k])
+							|| DCCCMaxUncompressedBlock[k] < 256) {
+						MaximumEffectiveCompressionChroma = 2;
+					} else {
+						MaximumEffectiveCompressionChroma = 4;
+					}
+					TotalCompressedReadBandwidth =
+							TotalCompressedReadBandwidth
+							+ ReadBandwidthSurfaceChroma[k]
+							/ dml_min(NetDCCRateChroma[k],
+							MaximumEffectiveCompressionChroma);
+#ifdef __DML_VBA_DEBUG__
+					dml_print("DML::%s: k=%0d, ReadBandwidthSurfaceChroma = %f\n",
+							__func__, k, ReadBandwidthSurfaceChroma[k]);
+					dml_print("DML::%s: k=%0d, NetDCCRateChroma = %f\n",
+							__func__, k, NetDCCRateChroma[k]);
+					dml_print("DML::%s: k=%0d, MaximumEffectiveCompressionChroma = %f\n",
+							__func__, k, MaximumEffectiveCompressionChroma);
+#endif
+					TotalZeroSizeRequestReadBandwidth = TotalZeroSizeRequestReadBandwidth
+							+ ReadBandwidthSurfaceChroma[k]
+									* DCCFractionOfZeroSizeRequestsChroma[k];
+					TotalZeroSizeCompressedReadBandwidth = TotalZeroSizeCompressedReadBandwidth
+							+ ReadBandwidthSurfaceChroma[k]
+									* DCCFractionOfZeroSizeRequestsChroma[k]
+									/ MaximumEffectiveCompressionChroma;
+				}
+			} else {
+				TotalCompressedReadBandwidth = TotalCompressedReadBandwidth
+						+ ReadBandwidthSurfaceLuma[k] + ReadBandwidthSurfaceChroma[k];
+			}
+			TotalRowReadBandwidth = TotalRowReadBandwidth
+					+ DPPPerSurface[k] * (meta_row_bw[k] + dpte_row_bw[k]);
+		}
+	}
+
+	AverageDCCCompressionRate = TotalDataReadBandwidth / TotalCompressedReadBandwidth;
+	AverageDCCZeroSizeFraction = TotalZeroSizeRequestReadBandwidth / TotalDataReadBandwidth;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: UnboundedRequestEnabled = %d\n", __func__, UnboundedRequestEnabled);
+	dml_print("DML::%s: TotalCompressedReadBandwidth = %f\n", __func__, TotalCompressedReadBandwidth);
+	dml_print("DML::%s: TotalZeroSizeRequestReadBandwidth = %f\n", __func__, TotalZeroSizeRequestReadBandwidth);
+	dml_print("DML::%s: TotalZeroSizeCompressedReadBandwidth = %f\n",
+			__func__, TotalZeroSizeCompressedReadBandwidth);
+	dml_print("DML::%s: MaximumEffectiveCompressionLuma = %f\n", __func__, MaximumEffectiveCompressionLuma);
+	dml_print("DML::%s: MaximumEffectiveCompressionChroma = %f\n", __func__, MaximumEffectiveCompressionChroma);
+	dml_print("DML::%s: AverageDCCCompressionRate = %f\n", __func__, AverageDCCCompressionRate);
+	dml_print("DML::%s: AverageDCCZeroSizeFraction = %f\n", __func__, AverageDCCZeroSizeFraction);
+	dml_print("DML::%s: CompbufReservedSpace64B = %d\n", __func__, CompbufReservedSpace64B);
+	dml_print("DML::%s: CompbufReservedSpaceZs = %d\n", __func__, CompbufReservedSpaceZs);
+	dml_print("DML::%s: CompressedBufferSizeInkByte = %d\n", __func__, CompressedBufferSizeInkByte);
+#endif
+	if (AverageDCCZeroSizeFraction == 1) {
+		AverageZeroSizeCompressionRate = TotalZeroSizeRequestReadBandwidth
+				/ TotalZeroSizeCompressedReadBandwidth;
+		EffectiveCompressedBufferSize = (double) MetaFIFOSizeInKEntries * 1024 * 64
+				* AverageZeroSizeCompressionRate
+				+ ((double) ZeroSizeBufferEntries - CompbufReservedSpaceZs) * 64
+						* AverageZeroSizeCompressionRate;
+	} else if (AverageDCCZeroSizeFraction > 0) {
+		AverageZeroSizeCompressionRate = TotalZeroSizeRequestReadBandwidth
+				/ TotalZeroSizeCompressedReadBandwidth;
+		EffectiveCompressedBufferSize = dml_min(
+				(double) CompressedBufferSizeInkByte * 1024 * AverageDCCCompressionRate,
+				(double) MetaFIFOSizeInKEntries * 1024 * 64
+					/ (AverageDCCZeroSizeFraction / AverageZeroSizeCompressionRate
+					+ 1 / AverageDCCCompressionRate))
+					+ dml_min(((double) ROBBufferSizeInKByte * 1024 - CompbufReservedSpace64B * 64)
+					* AverageDCCCompressionRate,
+					((double) ZeroSizeBufferEntries - CompbufReservedSpaceZs) * 64
+					/ (AverageDCCZeroSizeFraction / AverageZeroSizeCompressionRate));
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: min 1 = %f\n", __func__,
+				CompressedBufferSizeInkByte * 1024 * AverageDCCCompressionRate);
+		dml_print("DML::%s: min 2 = %f\n", __func__, MetaFIFOSizeInKEntries * 1024 * 64 /
+				(AverageDCCZeroSizeFraction / AverageZeroSizeCompressionRate + 1 /
+						AverageDCCCompressionRate));
+		dml_print("DML::%s: min 3 = %f\n", __func__, (ROBBufferSizeInKByte * 1024 -
+				CompbufReservedSpace64B * 64) * AverageDCCCompressionRate);
+		dml_print("DML::%s: min 4 = %f\n", __func__, (ZeroSizeBufferEntries - CompbufReservedSpaceZs) * 64 /
+				(AverageDCCZeroSizeFraction / AverageZeroSizeCompressionRate));
+#endif
+	} else {
+		EffectiveCompressedBufferSize = dml_min(
+				(double) CompressedBufferSizeInkByte * 1024 * AverageDCCCompressionRate,
+				(double) MetaFIFOSizeInKEntries * 1024 * 64 * AverageDCCCompressionRate)
+				+ ((double) ROBBufferSizeInKByte * 1024 - CompbufReservedSpace64B * 64)
+						* AverageDCCCompressionRate;
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: min 1 = %f\n", __func__,
+				CompressedBufferSizeInkByte * 1024 * AverageDCCCompressionRate);
+		dml_print("DML::%s: min 2 = %f\n", __func__,
+				MetaFIFOSizeInKEntries * 1024 * 64 * AverageDCCCompressionRate);
+#endif
+	}
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: MetaFIFOSizeInKEntries = %d\n", __func__, MetaFIFOSizeInKEntries);
+	dml_print("DML::%s: AverageZeroSizeCompressionRate = %f\n", __func__, AverageZeroSizeCompressionRate);
+	dml_print("DML::%s: EffectiveCompressedBufferSize = %f\n", __func__, EffectiveCompressedBufferSize);
+#endif
+
+	*StutterPeriod = 0;
+
+	for (k = 0; k < NumberOfActiveSurfaces; ++k) {
+		if (UseMALLForPStateChange[k] != dm_use_mall_pstate_change_phantom_pipe) {
+			LinesInDETY = ((double) DETBufferSizeY[k]
+					+ (UnboundedRequestEnabled == true ? EffectiveCompressedBufferSize : 0)
+							* ReadBandwidthSurfaceLuma[k] / TotalDataReadBandwidth)
+					/ BytePerPixelDETY[k] / SwathWidthY[k];
+			LinesInDETYRoundedDownToSwath = dml_floor(LinesInDETY, SwathHeightY[k]);
+			DETBufferingTimeY = LinesInDETYRoundedDownToSwath * ((double) HTotal[k] / PixelClock[k])
+					/ VRatio[k];
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: k=%0d, DETBufferSizeY = %d\n", __func__, k, DETBufferSizeY[k]);
+			dml_print("DML::%s: k=%0d, BytePerPixelDETY = %f\n", __func__, k, BytePerPixelDETY[k]);
+			dml_print("DML::%s: k=%0d, SwathWidthY = %d\n", __func__, k, SwathWidthY[k]);
+			dml_print("DML::%s: k=%0d, ReadBandwidthSurfaceLuma = %f\n",
+					__func__, k, ReadBandwidthSurfaceLuma[k]);
+			dml_print("DML::%s: k=%0d, TotalDataReadBandwidth = %f\n", __func__, k, TotalDataReadBandwidth);
+			dml_print("DML::%s: k=%0d, LinesInDETY = %f\n", __func__, k, LinesInDETY);
+			dml_print("DML::%s: k=%0d, LinesInDETYRoundedDownToSwath = %f\n",
+					__func__, k, LinesInDETYRoundedDownToSwath);
+			dml_print("DML::%s: k=%0d, HTotal = %d\n", __func__, k, HTotal[k]);
+			dml_print("DML::%s: k=%0d, PixelClock = %f\n", __func__, k, PixelClock[k]);
+			dml_print("DML::%s: k=%0d, VRatio = %f\n", __func__, k, VRatio[k]);
+			dml_print("DML::%s: k=%0d, DETBufferingTimeY = %f\n", __func__, k, DETBufferingTimeY);
+			dml_print("DML::%s: k=%0d, PixelClock = %f\n", __func__, k, PixelClock[k]);
+#endif
+
+			if (!FoundCriticalSurface || DETBufferingTimeY < *StutterPeriod) {
+				bool isInterlaceTiming = Interlace[k] && !ProgressiveToInterlaceUnitInOPP;
+
+				FoundCriticalSurface = true;
+				*StutterPeriod = DETBufferingTimeY;
+				FrameTimeCriticalSurface = (
+						isInterlaceTiming ?
+								dml_floor((double) VTotal[k] / 2.0, 1.0) : VTotal[k])
+						* (double) HTotal[k] / PixelClock[k];
+				VActiveTimeCriticalSurface = (
+						isInterlaceTiming ?
+								dml_floor((double) VActive[k] / 2.0, 1.0) : VActive[k])
+						* (double) HTotal[k] / PixelClock[k];
+				BytePerPixelYCriticalSurface = BytePerPixelY[k];
+				SwathWidthYCriticalSurface = SwathWidthY[k];
+				SwathHeightYCriticalSurface = SwathHeightY[k];
+				BlockWidth256BytesYCriticalSurface = BlockWidth256BytesY[k];
+				LinesToFinishSwathTransferStutterCriticalSurface = SwathHeightY[k]
+						- (LinesInDETY - LinesInDETYRoundedDownToSwath);
+				DETBufferSizeYCriticalSurface = DETBufferSizeY[k];
+				MinTTUVBlankCriticalSurface = MinTTUVBlank[k];
+				doublePlaneCriticalSurface = (ReadBandwidthSurfaceChroma[k] == 0);
+				doublePipeCriticalSurface = (DPPPerSurface[k] == 1);
+
+#ifdef __DML_VBA_DEBUG__
+				dml_print("DML::%s: k=%0d, FoundCriticalSurface                = %d\n",
+						__func__, k, FoundCriticalSurface);
+				dml_print("DML::%s: k=%0d, StutterPeriod                       = %f\n",
+						__func__, k, *StutterPeriod);
+				dml_print("DML::%s: k=%0d, MinTTUVBlankCriticalSurface         = %f\n",
+						__func__, k, MinTTUVBlankCriticalSurface);
+				dml_print("DML::%s: k=%0d, FrameTimeCriticalSurface            = %f\n",
+						__func__, k, FrameTimeCriticalSurface);
+				dml_print("DML::%s: k=%0d, VActiveTimeCriticalSurface          = %f\n",
+						__func__, k, VActiveTimeCriticalSurface);
+				dml_print("DML::%s: k=%0d, BytePerPixelYCriticalSurface        = %d\n",
+						__func__, k, BytePerPixelYCriticalSurface);
+				dml_print("DML::%s: k=%0d, SwathWidthYCriticalSurface          = %f\n",
+						__func__, k, SwathWidthYCriticalSurface);
+				dml_print("DML::%s: k=%0d, SwathHeightYCriticalSurface         = %f\n",
+						__func__, k, SwathHeightYCriticalSurface);
+				dml_print("DML::%s: k=%0d, BlockWidth256BytesYCriticalSurface  = %d\n",
+						__func__, k, BlockWidth256BytesYCriticalSurface);
+				dml_print("DML::%s: k=%0d, doublePlaneCriticalSurface          = %d\n",
+						__func__, k, doublePlaneCriticalSurface);
+				dml_print("DML::%s: k=%0d, doublePipeCriticalSurface           = %d\n",
+						__func__, k, doublePipeCriticalSurface);
+				dml_print("DML::%s: k=%0d, LinesToFinishSwathTransferStutterCriticalSurface = %f\n",
+						__func__, k, LinesToFinishSwathTransferStutterCriticalSurface);
+#endif
+			}
+		}
+	}
+
+	PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer = dml_min(*StutterPeriod * TotalDataReadBandwidth,
+			EffectiveCompressedBufferSize);
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: ROBBufferSizeInKByte = %d\n", __func__, ROBBufferSizeInKByte);
+	dml_print("DML::%s: AverageDCCCompressionRate = %f\n", __func__, AverageDCCCompressionRate);
+	dml_print("DML::%s: StutterPeriod * TotalDataReadBandwidth = %f\n",
+			__func__, *StutterPeriod * TotalDataReadBandwidth);
+	dml_print("DML::%s: EffectiveCompressedBufferSize = %f\n", __func__, EffectiveCompressedBufferSize);
+	dml_print("DML::%s: PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer = %f\n", __func__,
+			PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer);
+	dml_print("DML::%s: ReturnBW = %f\n", __func__, ReturnBW);
+	dml_print("DML::%s: TotalDataReadBandwidth = %f\n", __func__, TotalDataReadBandwidth);
+	dml_print("DML::%s: TotalRowReadBandwidth = %f\n", __func__, TotalRowReadBandwidth);
+	dml_print("DML::%s: DCFCLK = %f\n", __func__, DCFCLK);
+#endif
+
+	StutterBurstTime = PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer / AverageDCCCompressionRate
+			/ ReturnBW
+			+ (*StutterPeriod * TotalDataReadBandwidth
+					- PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer) / (DCFCLK * 64)
+			+ *StutterPeriod * TotalRowReadBandwidth / ReturnBW;
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: Part 1 = %f\n", __func__, PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer /
+			AverageDCCCompressionRate / ReturnBW);
+	dml_print("DML::%s: StutterPeriod * TotalDataReadBandwidth = %f\n",
+			__func__, (*StutterPeriod * TotalDataReadBandwidth));
+	dml_print("DML::%s: Part 2 = %f\n", __func__, (*StutterPeriod * TotalDataReadBandwidth -
+			PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer) / (DCFCLK * 64));
+	dml_print("DML::%s: Part 3 = %f\n", __func__, *StutterPeriod * TotalRowReadBandwidth / ReturnBW);
+	dml_print("DML::%s: StutterBurstTime = %f\n", __func__, StutterBurstTime);
+#endif
+	StutterBurstTime = dml_max(StutterBurstTime,
+			LinesToFinishSwathTransferStutterCriticalSurface * BytePerPixelYCriticalSurface
+					* SwathWidthYCriticalSurface / ReturnBW);
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: Time to finish residue swath=%f\n",
+			__func__,
+			LinesToFinishSwathTransferStutterCriticalSurface *
+			BytePerPixelYCriticalSurface * SwathWidthYCriticalSurface / ReturnBW);
+#endif
+
+	TotalActiveWriteback = 0;
+	for (k = 0; k < NumberOfActiveSurfaces; ++k) {
+		if (WritebackEnable[k])
+			TotalActiveWriteback = TotalActiveWriteback + 1;
+	}
+
+	if (TotalActiveWriteback == 0) {
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: SRExitTime = %f\n", __func__, SRExitTime);
+		dml_print("DML::%s: SRExitZ8Time = %f\n", __func__, SRExitZ8Time);
+		dml_print("DML::%s: StutterBurstTime = %f (final)\n", __func__, StutterBurstTime);
+		dml_print("DML::%s: StutterPeriod = %f\n", __func__, *StutterPeriod);
+#endif
+		*StutterEfficiencyNotIncludingVBlank = dml_max(0.,
+				1 - (SRExitTime + StutterBurstTime) / *StutterPeriod) * 100;
+		*Z8StutterEfficiencyNotIncludingVBlank = dml_max(0.,
+				1 - (SRExitZ8Time + StutterBurstTime) / *StutterPeriod) * 100;
+		*NumberOfStutterBurstsPerFrame = (
+				*StutterEfficiencyNotIncludingVBlank > 0 ?
+						dml_ceil(VActiveTimeCriticalSurface / *StutterPeriod, 1) : 0);
+		*Z8NumberOfStutterBurstsPerFrame = (
+				*Z8StutterEfficiencyNotIncludingVBlank > 0 ?
+						dml_ceil(VActiveTimeCriticalSurface / *StutterPeriod, 1) : 0);
+	} else {
+		*StutterEfficiencyNotIncludingVBlank = 0.;
+		*Z8StutterEfficiencyNotIncludingVBlank = 0.;
+		*NumberOfStutterBurstsPerFrame = 0;
+		*Z8NumberOfStutterBurstsPerFrame = 0;
+	}
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: VActiveTimeCriticalSurface = %f\n", __func__, VActiveTimeCriticalSurface);
+	dml_print("DML::%s: StutterEfficiencyNotIncludingVBlank = %f\n",
+			__func__, *StutterEfficiencyNotIncludingVBlank);
+	dml_print("DML::%s: Z8StutterEfficiencyNotIncludingVBlank = %f\n",
+			__func__, *Z8StutterEfficiencyNotIncludingVBlank);
+	dml_print("DML::%s: NumberOfStutterBurstsPerFrame = %d\n", __func__, *NumberOfStutterBurstsPerFrame);
+	dml_print("DML::%s: Z8NumberOfStutterBurstsPerFrame = %d\n", __func__, *Z8NumberOfStutterBurstsPerFrame);
+#endif
+
+	for (k = 0; k < NumberOfActiveSurfaces; ++k) {
+		if (UseMALLForPStateChange[k] != dm_use_mall_pstate_change_phantom_pipe) {
+			if (BlendingAndTiming[k] == k) {
+				if (TotalNumberOfActiveOTG == 0) {
+					doublePixelClock = PixelClock[k];
+					doubleHTotal = HTotal[k];
+					doubleVTotal = VTotal[k];
+				} else if (doublePixelClock != PixelClock[k] || doubleHTotal != HTotal[k]
+						|| doubleVTotal != VTotal[k]) {
+					SameTiming = false;
+				}
+				TotalNumberOfActiveOTG = TotalNumberOfActiveOTG + 1;
+			}
+		}
+	}
+
+	if (*StutterEfficiencyNotIncludingVBlank > 0) {
+		LastStutterPeriod = VActiveTimeCriticalSurface - (*NumberOfStutterBurstsPerFrame - 1) * *StutterPeriod;
+
+		if ((SynchronizeTimingsFinal || TotalNumberOfActiveOTG == 1) && SameTiming
+				&& LastStutterPeriod + MinTTUVBlankCriticalSurface > StutterEnterPlusExitWatermark) {
+			*StutterEfficiency = (1 - (*NumberOfStutterBurstsPerFrame * SRExitTime
+						+ StutterBurstTime * VActiveTimeCriticalSurface
+						/ *StutterPeriod) / FrameTimeCriticalSurface) * 100;
+		} else {
+			*StutterEfficiency = *StutterEfficiencyNotIncludingVBlank;
+		}
+	} else {
+		*StutterEfficiency = 0;
+	}
+
+	if (*Z8StutterEfficiencyNotIncludingVBlank > 0) {
+		LastZ8StutterPeriod = VActiveTimeCriticalSurface
+				- (*NumberOfStutterBurstsPerFrame - 1) * *StutterPeriod;
+		if ((SynchronizeTimingsFinal || TotalNumberOfActiveOTG == 1) && SameTiming && LastZ8StutterPeriod +
+				MinTTUVBlankCriticalSurface > Z8StutterEnterPlusExitWatermark) {
+			*Z8StutterEfficiency = (1 - (*NumberOfStutterBurstsPerFrame * SRExitZ8Time + StutterBurstTime
+				* VActiveTimeCriticalSurface / *StutterPeriod) / FrameTimeCriticalSurface) * 100;
+		} else {
+			*Z8StutterEfficiency = *Z8StutterEfficiencyNotIncludingVBlank;
+		}
+	} else {
+		*Z8StutterEfficiency = 0.;
+	}
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: LastZ8StutterPeriod = %f\n", __func__, LastZ8StutterPeriod);
+	dml_print("DML::%s: Z8StutterEnterPlusExitWatermark = %f\n", __func__, Z8StutterEnterPlusExitWatermark);
+	dml_print("DML::%s: StutterBurstTime = %f\n", __func__, StutterBurstTime);
+	dml_print("DML::%s: StutterPeriod = %f\n", __func__, *StutterPeriod);
+	dml_print("DML::%s: StutterEfficiency = %f\n", __func__, *StutterEfficiency);
+	dml_print("DML::%s: Z8StutterEfficiency = %f\n", __func__, *Z8StutterEfficiency);
+	dml_print("DML::%s: StutterEfficiencyNotIncludingVBlank = %f\n",
+			__func__, *StutterEfficiencyNotIncludingVBlank);
+	dml_print("DML::%s: Z8NumberOfStutterBurstsPerFrame = %d\n", __func__, *Z8NumberOfStutterBurstsPerFrame);
+#endif
+
+	SwathSizeCriticalSurface = BytePerPixelYCriticalSurface * SwathHeightYCriticalSurface
+			* dml_ceil(SwathWidthYCriticalSurface, BlockWidth256BytesYCriticalSurface);
+	LastChunkOfSwathSize = SwathSizeCriticalSurface % (PixelChunkSizeInKByte * 1024);
+	MissingPartOfLastSwathOfDETSize = dml_ceil(DETBufferSizeYCriticalSurface, SwathSizeCriticalSurface)
+			- DETBufferSizeYCriticalSurface;
+
+	*DCHUBBUB_ARB_CSTATE_MAX_CAP_MODE = !(!UnboundedRequestEnabled && (NumberOfActiveSurfaces == 1)
+			&& doublePlaneCriticalSurface && doublePipeCriticalSurface && (LastChunkOfSwathSize > 0)
+			&& (LastChunkOfSwathSize <= 4096) && (MissingPartOfLastSwathOfDETSize > 0)
+			&& (MissingPartOfLastSwathOfDETSize <= LastChunkOfSwathSize));
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: SwathSizeCriticalSurface = %d\n", __func__, SwathSizeCriticalSurface);
+	dml_print("DML::%s: LastChunkOfSwathSize = %d\n", __func__, LastChunkOfSwathSize);
+	dml_print("DML::%s: MissingPartOfLastSwathOfDETSize = %d\n", __func__, MissingPartOfLastSwathOfDETSize);
+	dml_print("DML::%s: DCHUBBUB_ARB_CSTATE_MAX_CAP_MODE = %d\n", __func__, *DCHUBBUB_ARB_CSTATE_MAX_CAP_MODE);
+#endif
+} // CalculateStutterEfficiency
+
+void dml32_CalculateMaxDETAndMinCompressedBufferSize(
+		unsigned int    ConfigReturnBufferSizeInKByte,
+		unsigned int    ROBBufferSizeInKByte,
+		unsigned int MaxNumDPP,
+		bool nomDETInKByteOverrideEnable, // VBA_DELTA, allow DV to override default DET size
+		unsigned int nomDETInKByteOverrideValue,  // VBA_DELTA
+
+		/* Output */
+		unsigned int *MaxTotalDETInKByte,
+		unsigned int *nomDETInKByte,
+		unsigned int *MinCompressedBufferSizeInKByte)
+{
+	bool     det_buff_size_override_en  = nomDETInKByteOverrideEnable;
+	unsigned int        det_buff_size_override_val = nomDETInKByteOverrideValue;
+
+	*MaxTotalDETInKByte = dml_ceil(((double)ConfigReturnBufferSizeInKByte +
+			(double) ROBBufferSizeInKByte) * 4.0 / 5.0, 64);
+	*nomDETInKByte = dml_floor((double) *MaxTotalDETInKByte / (double) MaxNumDPP, 64);
+	*MinCompressedBufferSizeInKByte = ConfigReturnBufferSizeInKByte - *MaxTotalDETInKByte;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: ConfigReturnBufferSizeInKByte = %0d\n", __func__, ConfigReturnBufferSizeInKByte);
+	dml_print("DML::%s: ROBBufferSizeInKByte = %0d\n", __func__, ROBBufferSizeInKByte);
+	dml_print("DML::%s: MaxNumDPP = %0d\n", __func__, MaxNumDPP);
+	dml_print("DML::%s: MaxTotalDETInKByte = %0d\n", __func__, *MaxTotalDETInKByte);
+	dml_print("DML::%s: nomDETInKByte = %0d\n", __func__, *nomDETInKByte);
+	dml_print("DML::%s: MinCompressedBufferSizeInKByte = %0d\n", __func__, *MinCompressedBufferSizeInKByte);
+#endif
+
+	if (det_buff_size_override_en) {
+		*nomDETInKByte = det_buff_size_override_val;
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: nomDETInKByte = %0d (override)\n", __func__, *nomDETInKByte);
+#endif
+	}
+} // CalculateMaxDETAndMinCompressedBufferSize
+
+bool dml32_CalculateVActiveBandwithSupport(unsigned int NumberOfActiveSurfaces,
+		double ReturnBW,
+		bool NotUrgentLatencyHiding[],
+		double ReadBandwidthLuma[],
+		double ReadBandwidthChroma[],
+		double cursor_bw[],
+		double meta_row_bandwidth[],
+		double dpte_row_bandwidth[],
+		unsigned int NumberOfDPP[],
+		double UrgentBurstFactorLuma[],
+		double UrgentBurstFactorChroma[],
+		double UrgentBurstFactorCursor[])
+{
+	unsigned int k;
+	bool NotEnoughUrgentLatencyHiding = false;
+	bool CalculateVActiveBandwithSupport_val = false;
+	double VActiveBandwith = 0;
+
+	for (k = 0; k < NumberOfActiveSurfaces; ++k) {
+		if (NotUrgentLatencyHiding[k]) {
+			NotEnoughUrgentLatencyHiding = true;
+		}
+	}
+
+	for (k = 0; k < NumberOfActiveSurfaces; ++k) {
+		VActiveBandwith = VActiveBandwith + ReadBandwidthLuma[k] * UrgentBurstFactorLuma[k] + ReadBandwidthChroma[k] * UrgentBurstFactorChroma[k] + cursor_bw[k] * UrgentBurstFactorCursor[k] + NumberOfDPP[k] * meta_row_bandwidth[k] + NumberOfDPP[k] * dpte_row_bandwidth[k];
+	}
+
+	CalculateVActiveBandwithSupport_val = (VActiveBandwith <= ReturnBW) && !NotEnoughUrgentLatencyHiding;
+
+#ifdef __DML_VBA_DEBUG__
+dml_print("DML::%s: NotEnoughUrgentLatencyHiding        = %d\n", __func__, NotEnoughUrgentLatencyHiding);
+dml_print("DML::%s: VActiveBandwith                     = %f\n", __func__, VActiveBandwith);
+dml_print("DML::%s: ReturnBW                            = %f\n", __func__, ReturnBW);
+dml_print("DML::%s: CalculateVActiveBandwithSupport_val = %d\n", __func__, CalculateVActiveBandwithSupport_val);
+#endif
+	return CalculateVActiveBandwithSupport_val;
+}
+
+void dml32_CalculatePrefetchBandwithSupport(unsigned int NumberOfActiveSurfaces,
+		double ReturnBW,
+		bool NotUrgentLatencyHiding[],
+		double ReadBandwidthLuma[],
+		double ReadBandwidthChroma[],
+		double PrefetchBandwidthLuma[],
+		double PrefetchBandwidthChroma[],
+		double cursor_bw[],
+		double meta_row_bandwidth[],
+		double dpte_row_bandwidth[],
+		double cursor_bw_pre[],
+		double prefetch_vmrow_bw[],
+		unsigned int NumberOfDPP[],
+		double UrgentBurstFactorLuma[],
+		double UrgentBurstFactorChroma[],
+		double UrgentBurstFactorCursor[],
+		double UrgentBurstFactorLumaPre[],
+		double UrgentBurstFactorChromaPre[],
+		double UrgentBurstFactorCursorPre[],
+		double PrefetchBW[],
+		double VRatio[],
+		double MaxVRatioPre,
+
+		/* output */
+		double  *MaxPrefetchBandwidth,
+		double  *FractionOfUrgentBandwidth,
+		bool *PrefetchBandwidthSupport)
+{
+	unsigned int k;
+	double ActiveBandwidthPerSurface;
+	bool NotEnoughUrgentLatencyHiding = false;
+	double TotalActiveBandwidth = 0;
+	double TotalPrefetchBandwidth = 0;
+
+	for (k = 0; k < NumberOfActiveSurfaces; ++k) {
+		if (NotUrgentLatencyHiding[k]) {
+			NotEnoughUrgentLatencyHiding = true;
+		}
+	}
+
+	*MaxPrefetchBandwidth = 0;
+	for (k = 0; k < NumberOfActiveSurfaces; ++k) {
+		ActiveBandwidthPerSurface = ReadBandwidthLuma[k] * UrgentBurstFactorLuma[k] + ReadBandwidthChroma[k] * UrgentBurstFactorChroma[k] + cursor_bw[k] * UrgentBurstFactorCursor[k] + NumberOfDPP[k] * (meta_row_bandwidth[k] + dpte_row_bandwidth[k]);
+
+		TotalActiveBandwidth += ActiveBandwidthPerSurface;
+
+		TotalPrefetchBandwidth = TotalPrefetchBandwidth + PrefetchBW[k] * VRatio[k];
+
+		*MaxPrefetchBandwidth = *MaxPrefetchBandwidth + dml_max3(NumberOfDPP[k] * prefetch_vmrow_bw[k],
+				ActiveBandwidthPerSurface,
+				NumberOfDPP[k] * (PrefetchBandwidthLuma[k] * UrgentBurstFactorLumaPre[k] + PrefetchBandwidthChroma[k] * UrgentBurstFactorChromaPre[k]) + cursor_bw_pre[k] * UrgentBurstFactorCursorPre[k]);
+	}
+
+	if (MaxVRatioPre == __DML_MAX_VRATIO_PRE__)
+		*PrefetchBandwidthSupport = (*MaxPrefetchBandwidth <= ReturnBW) && (TotalPrefetchBandwidth <= TotalActiveBandwidth * __DML_MAX_BW_RATIO_PRE__) && !NotEnoughUrgentLatencyHiding;
+	else
+		*PrefetchBandwidthSupport = (*MaxPrefetchBandwidth <= ReturnBW) && !NotEnoughUrgentLatencyHiding;
+
+	*FractionOfUrgentBandwidth = *MaxPrefetchBandwidth / ReturnBW;
+}
+
+double dml32_CalculateBandwidthAvailableForImmediateFlip(unsigned int NumberOfActiveSurfaces,
+		double ReturnBW,
+		double ReadBandwidthLuma[],
+		double ReadBandwidthChroma[],
+		double PrefetchBandwidthLuma[],
+		double PrefetchBandwidthChroma[],
+		double cursor_bw[],
+		double cursor_bw_pre[],
+		unsigned int NumberOfDPP[],
+		double UrgentBurstFactorLuma[],
+		double UrgentBurstFactorChroma[],
+		double UrgentBurstFactorCursor[],
+		double UrgentBurstFactorLumaPre[],
+		double UrgentBurstFactorChromaPre[],
+		double UrgentBurstFactorCursorPre[])
+{
+	unsigned int k;
+	double CalculateBandwidthAvailableForImmediateFlip_val = ReturnBW;
+
+	for (k = 0; k < NumberOfActiveSurfaces; ++k) {
+		CalculateBandwidthAvailableForImmediateFlip_val = CalculateBandwidthAvailableForImmediateFlip_val - dml_max(ReadBandwidthLuma[k] * UrgentBurstFactorLuma[k] + ReadBandwidthChroma[k] * UrgentBurstFactorChroma[k] + cursor_bw[k] * UrgentBurstFactorCursor[k],
+				NumberOfDPP[k] * (PrefetchBandwidthLuma[k] * UrgentBurstFactorLumaPre[k] + PrefetchBandwidthChroma[k] * UrgentBurstFactorChromaPre[k]) + cursor_bw_pre[k] * UrgentBurstFactorCursorPre[k]);
+	}
+
+	return CalculateBandwidthAvailableForImmediateFlip_val;
+}
+
+void dml32_CalculateImmediateFlipBandwithSupport(unsigned int NumberOfActiveSurfaces,
+		double ReturnBW,
+		enum immediate_flip_requirement ImmediateFlipRequirement[],
+		double final_flip_bw[],
+		double ReadBandwidthLuma[],
+		double ReadBandwidthChroma[],
+		double PrefetchBandwidthLuma[],
+		double PrefetchBandwidthChroma[],
+		double cursor_bw[],
+		double meta_row_bandwidth[],
+		double dpte_row_bandwidth[],
+		double cursor_bw_pre[],
+		double prefetch_vmrow_bw[],
+		unsigned int NumberOfDPP[],
+		double UrgentBurstFactorLuma[],
+		double UrgentBurstFactorChroma[],
+		double UrgentBurstFactorCursor[],
+		double UrgentBurstFactorLumaPre[],
+		double UrgentBurstFactorChromaPre[],
+		double UrgentBurstFactorCursorPre[],
+
+		/* output */
+		double  *TotalBandwidth,
+		double  *FractionOfUrgentBandwidth,
+		bool *ImmediateFlipBandwidthSupport)
+{
+	unsigned int k;
+	*TotalBandwidth = 0;
+	for (k = 0; k < NumberOfActiveSurfaces; ++k) {
+		if (ImmediateFlipRequirement[k] != dm_immediate_flip_not_required) {
+			*TotalBandwidth = *TotalBandwidth + dml_max3(NumberOfDPP[k] * prefetch_vmrow_bw[k],
+					NumberOfDPP[k] * final_flip_bw[k] + ReadBandwidthLuma[k] * UrgentBurstFactorLuma[k] + ReadBandwidthChroma[k] * UrgentBurstFactorChroma[k] + cursor_bw[k] * UrgentBurstFactorCursor[k],
+					NumberOfDPP[k] * (final_flip_bw[k] + PrefetchBandwidthLuma[k] * UrgentBurstFactorLumaPre[k] + PrefetchBandwidthChroma[k] * UrgentBurstFactorChromaPre[k]) + cursor_bw_pre[k] * UrgentBurstFactorCursorPre[k]);
+		} else {
+			*TotalBandwidth = *TotalBandwidth + dml_max3(NumberOfDPP[k] * prefetch_vmrow_bw[k],
+					NumberOfDPP[k] * (meta_row_bandwidth[k] + dpte_row_bandwidth[k]) + ReadBandwidthLuma[k] * UrgentBurstFactorLuma[k] + ReadBandwidthChroma[k] * UrgentBurstFactorChroma[k] + cursor_bw[k] * UrgentBurstFactorCursor[k],
+					NumberOfDPP[k] * (PrefetchBandwidthLuma[k] * UrgentBurstFactorLumaPre[k] + PrefetchBandwidthChroma[k] * UrgentBurstFactorChromaPre[k]) + cursor_bw_pre[k] * UrgentBurstFactorCursorPre[k]);
+		}
+	}
+	*ImmediateFlipBandwidthSupport = (*TotalBandwidth <= ReturnBW);
+	*FractionOfUrgentBandwidth = *TotalBandwidth / ReturnBW;
+}
+
+bool dml32_CalculateDETSwathFillLatencyHiding(unsigned int NumberOfActiveSurfaces,
+		double ReturnBW,
+		double UrgentLatency,
+		unsigned int SwathHeightY[],
+		unsigned int SwathHeightC[],
+		unsigned int SwathWidthY[],
+		unsigned int SwathWidthC[],
+		double  BytePerPixelInDETY[],
+		double  BytePerPixelInDETC[],
+		unsigned int    DETBufferSizeY[],
+		unsigned int    DETBufferSizeC[],
+		unsigned int	NumOfDPP[],
+		unsigned int	HTotal[],
+		double	PixelClock[],
+		double	VRatioY[],
+		double	VRatioC[],
+		enum dm_use_mall_for_pstate_change_mode UsesMALLForPStateChange[],
+		enum unbounded_requesting_policy UseUnboundedRequesting)
+{
+	int k;
+	double SwathSizeAllSurfaces = 0;
+	double SwathSizeAllSurfacesInFetchTimeUs;
+	double DETSwathLatencyHidingUs;
+	double DETSwathLatencyHidingYUs;
+	double DETSwathLatencyHidingCUs;
+	double SwathSizePerSurfaceY[DC__NUM_DPP__MAX];
+	double SwathSizePerSurfaceC[DC__NUM_DPP__MAX];
+	bool NotEnoughDETSwathFillLatencyHiding = false;
+
+	if (UseUnboundedRequesting == dm_unbounded_requesting)
+		return false;
+
+	/* calculate sum of single swath size for all pipes in bytes */
+	for (k = 0; k < NumberOfActiveSurfaces; k++) {
+		SwathSizePerSurfaceY[k] = SwathHeightY[k] * SwathWidthY[k] * BytePerPixelInDETY[k] * NumOfDPP[k];
+
+		if (SwathHeightC[k] != 0)
+			SwathSizePerSurfaceC[k] = SwathHeightC[k] * SwathWidthC[k] * BytePerPixelInDETC[k] * NumOfDPP[k];
+		else
+			SwathSizePerSurfaceC[k] = 0;
+
+		SwathSizeAllSurfaces += SwathSizePerSurfaceY[k] + SwathSizePerSurfaceC[k];
+	}
+
+	SwathSizeAllSurfacesInFetchTimeUs = SwathSizeAllSurfaces / ReturnBW + UrgentLatency;
+
+	/* ensure all DET - 1 swath can hide a fetch for all surfaces */
+	for (k = 0; k < NumberOfActiveSurfaces; k++) {
+		double LineTime = HTotal[k] / PixelClock[k];
+
+		/* only care if surface is not phantom */
+		if (UsesMALLForPStateChange[k] != dm_use_mall_pstate_change_phantom_pipe) {
+			DETSwathLatencyHidingYUs = (dml_floor(DETBufferSizeY[k] / BytePerPixelInDETY[k] / SwathWidthY[k], 1.0) - SwathHeightY[k]) / VRatioY[k] * LineTime;
+
+			if (SwathHeightC[k] != 0) {
+				DETSwathLatencyHidingCUs = (dml_floor(DETBufferSizeC[k] / BytePerPixelInDETC[k] / SwathWidthC[k], 1.0) - SwathHeightC[k]) / VRatioC[k] * LineTime;
+
+				DETSwathLatencyHidingUs = dml_min(DETSwathLatencyHidingYUs, DETSwathLatencyHidingCUs);
+			} else {
+				DETSwathLatencyHidingUs = DETSwathLatencyHidingYUs;
+			}
+
+			/* DET must be able to hide time to fetch 1 swath for each surface */
+			if (DETSwathLatencyHidingUs < SwathSizeAllSurfacesInFetchTimeUs) {
+				NotEnoughDETSwathFillLatencyHiding = true;
+				break;
+			}
+		}
+	}
+
+	return NotEnoughDETSwathFillLatencyHiding;
+}
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_util_32.h b/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_util_32.h
new file mode 100644
index 000000000000..5d34735df83d
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_util_32.h
@@ -0,0 +1,1170 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __DML_DCN32_DISPLAY_MODE_VBA_UTIL_32_H__
+#define __DML_DCN32_DISPLAY_MODE_VBA_UTIL_32_H__
+
+#include "../display_mode_enums.h"
+#include "os_types.h"
+#include "../dc_features.h"
+#include "../display_mode_structs.h"
+#include "../display_mode_vba.h"
+
+unsigned int dml32_dscceComputeDelay(
+		unsigned int bpc,
+		double BPP,
+		unsigned int sliceWidth,
+		unsigned int numSlices,
+		enum output_format_class pixelFormat,
+		enum output_encoder_class Output);
+
+unsigned int dml32_dscComputeDelay(enum output_format_class pixelFormat, enum output_encoder_class Output);
+
+bool IsVertical(enum dm_rotation_angle Scan);
+
+void dml32_CalculateBytePerPixelAndBlockSizes(
+		enum source_format_class SourcePixelFormat,
+		enum dm_swizzle_mode SurfaceTiling,
+
+		/*Output*/
+		unsigned int *BytePerPixelY,
+		unsigned int *BytePerPixelC,
+		double           *BytePerPixelDETY,
+		double           *BytePerPixelDETC,
+		unsigned int *BlockHeight256BytesY,
+		unsigned int *BlockHeight256BytesC,
+		unsigned int *BlockWidth256BytesY,
+		unsigned int *BlockWidth256BytesC,
+		unsigned int *MacroTileHeightY,
+		unsigned int *MacroTileHeightC,
+		unsigned int *MacroTileWidthY,
+		unsigned int *MacroTileWidthC);
+
+void dml32_CalculateSinglePipeDPPCLKAndSCLThroughput(
+		double HRatio,
+		double HRatioChroma,
+		double VRatio,
+		double VRatioChroma,
+		double MaxDCHUBToPSCLThroughput,
+		double MaxPSCLToLBThroughput,
+		double PixelClock,
+		enum source_format_class SourcePixelFormat,
+		unsigned int HTaps,
+		unsigned int HTapsChroma,
+		unsigned int VTaps,
+		unsigned int VTapsChroma,
+
+		/* output */
+		double *PSCL_THROUGHPUT,
+		double *PSCL_THROUGHPUT_CHROMA,
+		double *DPPCLKUsingSingleDPP);
+
+void dml32_CalculateSwathAndDETConfiguration(
+		unsigned int DETSizeOverride[],
+		enum dm_use_mall_for_pstate_change_mode UseMALLForPStateChange[],
+		unsigned int ConfigReturnBufferSizeInKByte,
+		unsigned int MaxTotalDETInKByte,
+		unsigned int MinCompressedBufferSizeInKByte,
+		double ForceSingleDPP,
+		unsigned int NumberOfActiveSurfaces,
+		unsigned int nomDETInKByte,
+		enum unbounded_requesting_policy UseUnboundedRequestingFinal,
+		bool DisableUnboundRequestIfCompBufReservedSpaceNeedAdjustment,
+		unsigned int PixelChunkSizeKBytes,
+		unsigned int ROBSizeKBytes,
+		unsigned int CompressedBufferSegmentSizeInkByteFinal,
+		enum output_encoder_class Output[],
+		double ReadBandwidthLuma[],
+		double ReadBandwidthChroma[],
+		double MaximumSwathWidthLuma[],
+		double MaximumSwathWidthChroma[],
+		enum dm_rotation_angle SourceRotation[],
+		bool ViewportStationary[],
+		enum source_format_class SourcePixelFormat[],
+		enum dm_swizzle_mode SurfaceTiling[],
+		unsigned int ViewportWidth[],
+		unsigned int ViewportHeight[],
+		unsigned int ViewportXStart[],
+		unsigned int ViewportYStart[],
+		unsigned int ViewportXStartC[],
+		unsigned int ViewportYStartC[],
+		unsigned int SurfaceWidthY[],
+		unsigned int SurfaceWidthC[],
+		unsigned int SurfaceHeightY[],
+		unsigned int SurfaceHeightC[],
+		unsigned int Read256BytesBlockHeightY[],
+		unsigned int Read256BytesBlockHeightC[],
+		unsigned int Read256BytesBlockWidthY[],
+		unsigned int Read256BytesBlockWidthC[],
+		enum odm_combine_mode ODMMode[],
+		unsigned int BlendingAndTiming[],
+		unsigned int BytePerPixY[],
+		unsigned int BytePerPixC[],
+		double BytePerPixDETY[],
+		double BytePerPixDETC[],
+		unsigned int HActive[],
+		double HRatio[],
+		double HRatioChroma[],
+		unsigned int DPPPerSurface[],
+
+		/* Output */
+		unsigned int swath_width_luma_ub[],
+		unsigned int swath_width_chroma_ub[],
+		double SwathWidth[],
+		double SwathWidthChroma[],
+		unsigned int SwathHeightY[],
+		unsigned int SwathHeightC[],
+		unsigned int DETBufferSizeInKByte[],
+		unsigned int DETBufferSizeY[],
+		unsigned int DETBufferSizeC[],
+		bool *UnboundedRequestEnabled,
+		unsigned int *CompressedBufferSizeInkByte,
+		unsigned int *CompBufReservedSpaceKBytes,
+		bool *CompBufReservedSpaceNeedAdjustment,
+		bool ViewportSizeSupportPerSurface[],
+		bool *ViewportSizeSupport);
+
+void dml32_CalculateSwathWidth(
+		bool ForceSingleDPP,
+		unsigned int NumberOfActiveSurfaces,
+		enum source_format_class SourcePixelFormat[],
+		enum dm_rotation_angle SourceScan[],
+		bool ViewportStationary[],
+		unsigned int ViewportWidth[],
+		unsigned int ViewportHeight[],
+		unsigned int ViewportXStart[],
+		unsigned int ViewportYStart[],
+		unsigned int ViewportXStartC[],
+		unsigned int ViewportYStartC[],
+		unsigned int SurfaceWidthY[],
+		unsigned int SurfaceWidthC[],
+		unsigned int SurfaceHeightY[],
+		unsigned int SurfaceHeightC[],
+		enum odm_combine_mode ODMMode[],
+		unsigned int BytePerPixY[],
+		unsigned int BytePerPixC[],
+		unsigned int Read256BytesBlockHeightY[],
+		unsigned int Read256BytesBlockHeightC[],
+		unsigned int Read256BytesBlockWidthY[],
+		unsigned int Read256BytesBlockWidthC[],
+		unsigned int BlendingAndTiming[],
+		unsigned int HActive[],
+		double HRatio[],
+		unsigned int DPPPerSurface[],
+
+		/* Output */
+		double SwathWidthdoubleDPPY[],
+		double SwathWidthdoubleDPPC[],
+		double SwathWidthY[], // per-pipe
+		double SwathWidthC[], // per-pipe
+		unsigned int MaximumSwathHeightY[],
+		unsigned int MaximumSwathHeightC[],
+		unsigned int swath_width_luma_ub[], // per-pipe
+		unsigned int swath_width_chroma_ub[]);
+
+bool dml32_UnboundedRequest(enum unbounded_requesting_policy UseUnboundedRequestingFinal,
+		unsigned int TotalNumberOfActiveDPP,
+		bool NoChroma,
+		enum output_encoder_class Output,
+		enum dm_swizzle_mode SurfaceTiling,
+		bool CompBufReservedSpaceNeedAdjustment,
+		bool DisableUnboundRequestIfCompBufReservedSpaceNeedAdjustment);
+
+void dml32_CalculateDETBufferSize(
+		unsigned int DETSizeOverride[],
+		enum dm_use_mall_for_pstate_change_mode UseMALLForPStateChange[],
+		bool ForceSingleDPP,
+		unsigned int NumberOfActiveSurfaces,
+		bool UnboundedRequestEnabled,
+		unsigned int nomDETInKByte,
+		unsigned int MaxTotalDETInKByte,
+		unsigned int ConfigReturnBufferSizeInKByte,
+		unsigned int MinCompressedBufferSizeInKByte,
+		unsigned int CompressedBufferSegmentSizeInkByteFinal,
+		enum source_format_class SourcePixelFormat[],
+		double ReadBandwidthLuma[],
+		double ReadBandwidthChroma[],
+		unsigned int RoundedUpMaxSwathSizeBytesY[],
+		unsigned int RoundedUpMaxSwathSizeBytesC[],
+		unsigned int DPPPerSurface[],
+		/* Output */
+		unsigned int DETBufferSizeInKByte[],
+		unsigned int *CompressedBufferSizeInkByte);
+
+void dml32_CalculateODMMode(
+		unsigned int MaximumPixelsPerLinePerDSCUnit,
+		unsigned int HActive,
+		enum output_format_class OutFormat,
+		enum output_encoder_class Output,
+		enum odm_combine_policy ODMUse,
+		double StateDispclk,
+		double MaxDispclk,
+		bool DSCEnable,
+		unsigned int TotalNumberOfActiveDPP,
+		unsigned int MaxNumDPP,
+		double PixelClock,
+		double DISPCLKDPPCLKDSCCLKDownSpreading,
+		double DISPCLKRampingMargin,
+		double DISPCLKDPPCLKVCOSpeed,
+		unsigned int NumberOfDSCSlices,
+
+		/* Output */
+		bool *TotalAvailablePipesSupport,
+		unsigned int *NumberOfDPP,
+		enum odm_combine_mode *ODMMode,
+		double *RequiredDISPCLKPerSurface);
+
+double dml32_CalculateRequiredDispclk(
+		enum odm_combine_mode ODMMode,
+		double PixelClock,
+		double DISPCLKDPPCLKDSCCLKDownSpreading,
+		double DISPCLKRampingMargin,
+		double DISPCLKDPPCLKVCOSpeed,
+		double MaxDispclk);
+
+double dml32_RoundToDFSGranularity(double Clock, bool round_up, double VCOSpeed);
+
+void dml32_CalculateOutputLink(
+		double PHYCLKPerState,
+		double PHYCLKD18PerState,
+		double PHYCLKD32PerState,
+		double Downspreading,
+		bool IsMainSurfaceUsingTheIndicatedTiming,
+		enum output_encoder_class Output,
+		enum output_format_class OutputFormat,
+		unsigned int HTotal,
+		unsigned int HActive,
+		double PixelClockBackEnd,
+		double ForcedOutputLinkBPP,
+		unsigned int DSCInputBitPerComponent,
+		unsigned int NumberOfDSCSlices,
+		double AudioSampleRate,
+		unsigned int AudioSampleLayout,
+		enum odm_combine_mode ODMModeNoDSC,
+		enum odm_combine_mode ODMModeDSC,
+		bool DSCEnable,
+		unsigned int OutputLinkDPLanes,
+		enum dm_output_link_dp_rate OutputLinkDPRate,
+
+		/* Output */
+		bool *RequiresDSC,
+		double *RequiresFEC,
+		double  *OutBpp,
+		enum dm_output_type *OutputType,
+		enum dm_output_rate *OutputRate,
+		unsigned int *RequiredSlots);
+
+void dml32_CalculateDPPCLK(
+		unsigned int NumberOfActiveSurfaces,
+		double DISPCLKDPPCLKDSCCLKDownSpreading,
+		double DISPCLKDPPCLKVCOSpeed,
+		double DPPCLKUsingSingleDPP[],
+		unsigned int DPPPerSurface[],
+
+		/* output */
+		double *GlobalDPPCLK,
+		double Dppclk[]);
+
+double dml32_TruncToValidBPP(
+		double LinkBitRate,
+		unsigned int Lanes,
+		unsigned int HTotal,
+		unsigned int HActive,
+		double PixelClock,
+		double DesiredBPP,
+		bool DSCEnable,
+		enum output_encoder_class Output,
+		enum output_format_class Format,
+		unsigned int DSCInputBitPerComponent,
+		unsigned int DSCSlices,
+		unsigned int AudioRate,
+		unsigned int AudioLayout,
+		enum odm_combine_mode ODMModeNoDSC,
+		enum odm_combine_mode ODMModeDSC,
+		/* Output */
+		unsigned int *RequiredSlots);
+
+double dml32_RequiredDTBCLK(
+		bool              DSCEnable,
+		double               PixelClock,
+		enum output_format_class  OutputFormat,
+		double               OutputBpp,
+		unsigned int              DSCSlices,
+		unsigned int                 HTotal,
+		unsigned int                 HActive,
+		unsigned int              AudioRate,
+		unsigned int              AudioLayout);
+
+unsigned int dml32_DSCDelayRequirement(bool DSCEnabled,
+		enum odm_combine_mode ODMMode,
+		unsigned int DSCInputBitPerComponent,
+		double OutputBpp,
+		unsigned int HActive,
+		unsigned int HTotal,
+		unsigned int NumberOfDSCSlices,
+		enum output_format_class  OutputFormat,
+		enum output_encoder_class Output,
+		double PixelClock,
+		double PixelClockBackEnd,
+		double dsc_delay_factor_wa);
+
+void dml32_CalculateSurfaceSizeInMall(
+		unsigned int NumberOfActiveSurfaces,
+		unsigned int MALLAllocatedForDCN,
+		enum dm_use_mall_for_static_screen_mode UseMALLForStaticScreen[],
+		enum dm_use_mall_for_pstate_change_mode UsesMALLForPStateChange[],
+		bool DCCEnable[],
+		bool ViewportStationary[],
+		unsigned int ViewportXStartY[],
+		unsigned int ViewportYStartY[],
+		unsigned int ViewportXStartC[],
+		unsigned int ViewportYStartC[],
+		unsigned int ViewportWidthY[],
+		unsigned int ViewportHeightY[],
+		unsigned int BytesPerPixelY[],
+		unsigned int ViewportWidthC[],
+		unsigned int ViewportHeightC[],
+		unsigned int BytesPerPixelC[],
+		unsigned int SurfaceWidthY[],
+		unsigned int SurfaceWidthC[],
+		unsigned int SurfaceHeightY[],
+		unsigned int SurfaceHeightC[],
+		unsigned int Read256BytesBlockWidthY[],
+		unsigned int Read256BytesBlockWidthC[],
+		unsigned int Read256BytesBlockHeightY[],
+		unsigned int Read256BytesBlockHeightC[],
+		unsigned int ReadBlockWidthY[],
+		unsigned int ReadBlockWidthC[],
+		unsigned int ReadBlockHeightY[],
+		unsigned int ReadBlockHeightC[],
+		unsigned int DCCMetaPitchY[],
+		unsigned int DCCMetaPitchC[],
+
+		/* Output */
+		unsigned int    SurfaceSizeInMALL[],
+		bool *ExceededMALLSize);
+
+void dml32_CalculateVMRowAndSwath(
+		unsigned int NumberOfActiveSurfaces,
+		DmlPipe myPipe[],
+		unsigned int SurfaceSizeInMALL[],
+		unsigned int PTEBufferSizeInRequestsLuma,
+		unsigned int PTEBufferSizeInRequestsChroma,
+		unsigned int DCCMetaBufferSizeBytes,
+		enum dm_use_mall_for_static_screen_mode UseMALLForStaticScreen[],
+		enum dm_use_mall_for_pstate_change_mode UseMALLForPStateChange[],
+		unsigned int MALLAllocatedForDCN,
+		double SwathWidthY[],
+		double SwathWidthC[],
+		bool GPUVMEnable,
+		bool HostVMEnable,
+		unsigned int HostVMMaxNonCachedPageTableLevels,
+		unsigned int GPUVMMaxPageTableLevels,
+		unsigned int GPUVMMinPageSizeKBytes[],
+		unsigned int HostVMMinPageSize,
+
+		/* Output */
+		bool PTEBufferSizeNotExceeded[],
+		bool DCCMetaBufferSizeNotExceeded[],
+		unsigned int dpte_row_width_luma_ub[],
+		unsigned int dpte_row_width_chroma_ub[],
+		unsigned int dpte_row_height_luma[],
+		unsigned int dpte_row_height_chroma[],
+		unsigned int dpte_row_height_linear_luma[],     // VBA_DELTA
+		unsigned int dpte_row_height_linear_chroma[],   // VBA_DELTA
+		unsigned int meta_req_width[],
+		unsigned int meta_req_width_chroma[],
+		unsigned int meta_req_height[],
+		unsigned int meta_req_height_chroma[],
+		unsigned int meta_row_width[],
+		unsigned int meta_row_width_chroma[],
+		unsigned int meta_row_height[],
+		unsigned int meta_row_height_chroma[],
+		unsigned int vm_group_bytes[],
+		unsigned int dpte_group_bytes[],
+		unsigned int PixelPTEReqWidthY[],
+		unsigned int PixelPTEReqHeightY[],
+		unsigned int PTERequestSizeY[],
+		unsigned int PixelPTEReqWidthC[],
+		unsigned int PixelPTEReqHeightC[],
+		unsigned int PTERequestSizeC[],
+		unsigned int dpde0_bytes_per_frame_ub_l[],
+		unsigned int meta_pte_bytes_per_frame_ub_l[],
+		unsigned int dpde0_bytes_per_frame_ub_c[],
+		unsigned int meta_pte_bytes_per_frame_ub_c[],
+		double PrefetchSourceLinesY[],
+		double PrefetchSourceLinesC[],
+		double VInitPreFillY[],
+		double VInitPreFillC[],
+		unsigned int MaxNumSwathY[],
+		unsigned int MaxNumSwathC[],
+		double meta_row_bw[],
+		double dpte_row_bw[],
+		double PixelPTEBytesPerRow[],
+		double PDEAndMetaPTEBytesFrame[],
+		double MetaRowByte[],
+		bool use_one_row_for_frame[],
+		bool use_one_row_for_frame_flip[],
+		bool UsesMALLForStaticScreen[],
+		bool PTE_BUFFER_MODE[],
+		unsigned int BIGK_FRAGMENT_SIZE[]);
+
+unsigned int dml32_CalculateVMAndRowBytes(
+		bool ViewportStationary,
+		bool DCCEnable,
+		unsigned int NumberOfDPPs,
+		unsigned int BlockHeight256Bytes,
+		unsigned int BlockWidth256Bytes,
+		enum source_format_class SourcePixelFormat,
+		unsigned int SurfaceTiling,
+		unsigned int BytePerPixel,
+		enum dm_rotation_angle SourceScan,
+		double SwathWidth,
+		unsigned int ViewportHeight,
+		unsigned int    ViewportXStart,
+		unsigned int    ViewportYStart,
+		bool GPUVMEnable,
+		bool HostVMEnable,
+		unsigned int HostVMMaxNonCachedPageTableLevels,
+		unsigned int GPUVMMaxPageTableLevels,
+		unsigned int GPUVMMinPageSizeKBytes,
+		unsigned int HostVMMinPageSize,
+		unsigned int PTEBufferSizeInRequests,
+		unsigned int Pitch,
+		unsigned int DCCMetaPitch,
+		unsigned int MacroTileWidth,
+		unsigned int MacroTileHeight,
+
+		/* Output */
+		unsigned int *MetaRowByte,
+		unsigned int *PixelPTEBytesPerRow,
+		unsigned int    *dpte_row_width_ub,
+		unsigned int *dpte_row_height,
+		unsigned int *dpte_row_height_linear,
+		unsigned int    *PixelPTEBytesPerRow_one_row_per_frame,
+		unsigned int    *dpte_row_width_ub_one_row_per_frame,
+		unsigned int    *dpte_row_height_one_row_per_frame,
+		unsigned int *MetaRequestWidth,
+		unsigned int *MetaRequestHeight,
+		unsigned int *meta_row_width,
+		unsigned int *meta_row_height,
+		unsigned int *PixelPTEReqWidth,
+		unsigned int *PixelPTEReqHeight,
+		unsigned int *PTERequestSize,
+		unsigned int    *DPDE0BytesFrame,
+		unsigned int    *MetaPTEBytesFrame);
+
+double dml32_CalculatePrefetchSourceLines(
+		double VRatio,
+		unsigned int VTaps,
+		bool Interlace,
+		bool ProgressiveToInterlaceUnitInOPP,
+		unsigned int SwathHeight,
+		enum dm_rotation_angle SourceRotation,
+		bool ViewportStationary,
+		double SwathWidth,
+		unsigned int ViewportHeight,
+		unsigned int ViewportXStart,
+		unsigned int ViewportYStart,
+
+		/* Output */
+		double *VInitPreFill,
+		unsigned int *MaxNumSwath);
+
+void dml32_CalculateMALLUseForStaticScreen(
+		unsigned int NumberOfActiveSurfaces,
+		unsigned int MALLAllocatedForDCNFinal,
+		enum dm_use_mall_for_static_screen_mode *UseMALLForStaticScreen,
+		unsigned int SurfaceSizeInMALL[],
+		bool one_row_per_frame_fits_in_buffer[],
+
+		/* output */
+		bool UsesMALLForStaticScreen[]);
+
+void dml32_CalculateRowBandwidth(
+		bool GPUVMEnable,
+		enum source_format_class SourcePixelFormat,
+		double VRatio,
+		double VRatioChroma,
+		bool DCCEnable,
+		double LineTime,
+		unsigned int MetaRowByteLuma,
+		unsigned int MetaRowByteChroma,
+		unsigned int meta_row_height_luma,
+		unsigned int meta_row_height_chroma,
+		unsigned int PixelPTEBytesPerRowLuma,
+		unsigned int PixelPTEBytesPerRowChroma,
+		unsigned int dpte_row_height_luma,
+		unsigned int dpte_row_height_chroma,
+		/* Output */
+		double *meta_row_bw,
+		double *dpte_row_bw);
+
+double dml32_CalculateUrgentLatency(
+		double UrgentLatencyPixelDataOnly,
+		double UrgentLatencyPixelMixedWithVMData,
+		double UrgentLatencyVMDataOnly,
+		bool   DoUrgentLatencyAdjustment,
+		double UrgentLatencyAdjustmentFabricClockComponent,
+		double UrgentLatencyAdjustmentFabricClockReference,
+		double FabricClock);
+
+void dml32_CalculateUrgentBurstFactor(
+		enum dm_use_mall_for_pstate_change_mode UseMALLForPStateChange,
+		unsigned int    swath_width_luma_ub,
+		unsigned int    swath_width_chroma_ub,
+		unsigned int SwathHeightY,
+		unsigned int SwathHeightC,
+		double  LineTime,
+		double  UrgentLatency,
+		double  CursorBufferSize,
+		unsigned int CursorWidth,
+		unsigned int CursorBPP,
+		double  VRatio,
+		double  VRatioC,
+		double  BytePerPixelInDETY,
+		double  BytePerPixelInDETC,
+		unsigned int    DETBufferSizeY,
+		unsigned int    DETBufferSizeC,
+		/* Output */
+		double *UrgentBurstFactorCursor,
+		double *UrgentBurstFactorLuma,
+		double *UrgentBurstFactorChroma,
+		bool   *NotEnoughUrgentLatencyHiding);
+
+void dml32_CalculateDCFCLKDeepSleep(
+		unsigned int NumberOfActiveSurfaces,
+		unsigned int BytePerPixelY[],
+		unsigned int BytePerPixelC[],
+		double VRatio[],
+		double VRatioChroma[],
+		double SwathWidthY[],
+		double SwathWidthC[],
+		unsigned int DPPPerSurface[],
+		double HRatio[],
+		double HRatioChroma[],
+		double PixelClock[],
+		double PSCL_THROUGHPUT[],
+		double PSCL_THROUGHPUT_CHROMA[],
+		double Dppclk[],
+		double ReadBandwidthLuma[],
+		double ReadBandwidthChroma[],
+		unsigned int ReturnBusWidth,
+
+		/* Output */
+		double *DCFClkDeepSleep);
+
+double dml32_CalculateWriteBackDelay(
+		enum source_format_class WritebackPixelFormat,
+		double WritebackHRatio,
+		double WritebackVRatio,
+		unsigned int WritebackVTaps,
+		unsigned int         WritebackDestinationWidth,
+		unsigned int         WritebackDestinationHeight,
+		unsigned int         WritebackSourceHeight,
+		unsigned int HTotal);
+
+void dml32_UseMinimumDCFCLK(
+		enum dm_use_mall_for_pstate_change_mode UseMALLForPStateChange[],
+		bool DRRDisplay[],
+		bool SynchronizeDRRDisplaysForUCLKPStateChangeFinal,
+		unsigned int MaxInterDCNTileRepeaters,
+		unsigned int MaxPrefetchMode,
+		double DRAMClockChangeLatencyFinal,
+		double FCLKChangeLatency,
+		double SREnterPlusExitTime,
+		unsigned int ReturnBusWidth,
+		unsigned int RoundTripPingLatencyCycles,
+		unsigned int ReorderingBytes,
+		unsigned int PixelChunkSizeInKByte,
+		unsigned int MetaChunkSize,
+		bool GPUVMEnable,
+		unsigned int GPUVMMaxPageTableLevels,
+		bool HostVMEnable,
+		unsigned int NumberOfActiveSurfaces,
+		double HostVMMinPageSize,
+		unsigned int HostVMMaxNonCachedPageTableLevels,
+		bool DynamicMetadataVMEnabled,
+		bool ImmediateFlipRequirement,
+		bool ProgressiveToInterlaceUnitInOPP,
+		double MaxAveragePercentOfIdealSDPPortBWDisplayCanUseInNormalSystemOperation,
+		double PercentOfIdealSDPPortBWReceivedAfterUrgLatency,
+		unsigned int VTotal[],
+		unsigned int VActive[],
+		unsigned int DynamicMetadataTransmittedBytes[],
+		unsigned int DynamicMetadataLinesBeforeActiveRequired[],
+		bool Interlace[],
+		double RequiredDPPCLKPerSurface[][2][DC__NUM_DPP__MAX],
+		double RequiredDISPCLK[][2],
+		double UrgLatency[],
+		unsigned int NoOfDPP[][2][DC__NUM_DPP__MAX],
+		double ProjectedDCFClkDeepSleep[][2],
+		double MaximumVStartup[][2][DC__NUM_DPP__MAX],
+		unsigned int TotalNumberOfActiveDPP[][2],
+		unsigned int TotalNumberOfDCCActiveDPP[][2],
+		unsigned int dpte_group_bytes[],
+		double PrefetchLinesY[][2][DC__NUM_DPP__MAX],
+		double PrefetchLinesC[][2][DC__NUM_DPP__MAX],
+		unsigned int swath_width_luma_ub_all_states[][2][DC__NUM_DPP__MAX],
+		unsigned int swath_width_chroma_ub_all_states[][2][DC__NUM_DPP__MAX],
+		unsigned int BytePerPixelY[],
+		unsigned int BytePerPixelC[],
+		unsigned int HTotal[],
+		double PixelClock[],
+		double PDEAndMetaPTEBytesPerFrame[][2][DC__NUM_DPP__MAX],
+		double DPTEBytesPerRow[][2][DC__NUM_DPP__MAX],
+		double MetaRowBytes[][2][DC__NUM_DPP__MAX],
+		bool DynamicMetadataEnable[],
+		double ReadBandwidthLuma[],
+		double ReadBandwidthChroma[],
+		double DCFCLKPerState[],
+		/* Output */
+		double DCFCLKState[][2]);
+
+unsigned int dml32_CalculateExtraLatencyBytes(unsigned int ReorderingBytes,
+		unsigned int TotalNumberOfActiveDPP,
+		unsigned int PixelChunkSizeInKByte,
+		unsigned int TotalNumberOfDCCActiveDPP,
+		unsigned int MetaChunkSize,
+		bool GPUVMEnable,
+		bool HostVMEnable,
+		unsigned int NumberOfActiveSurfaces,
+		unsigned int NumberOfDPP[],
+		unsigned int dpte_group_bytes[],
+		double HostVMInefficiencyFactor,
+		double HostVMMinPageSize,
+		unsigned int HostVMMaxNonCachedPageTableLevels);
+
+void dml32_CalculateVUpdateAndDynamicMetadataParameters(
+		unsigned int MaxInterDCNTileRepeaters,
+		double Dppclk,
+		double Dispclk,
+		double DCFClkDeepSleep,
+		double PixelClock,
+		unsigned int HTotal,
+		unsigned int VBlank,
+		unsigned int DynamicMetadataTransmittedBytes,
+		unsigned int DynamicMetadataLinesBeforeActiveRequired,
+		unsigned int InterlaceEnable,
+		bool ProgressiveToInterlaceUnitInOPP,
+		double *TSetup,
+		double *Tdmbf,
+		double *Tdmec,
+		double *Tdmsks,
+		unsigned int *VUpdateOffsetPix,
+		double *VUpdateWidthPix,
+		double *VReadyOffsetPix);
+
+double dml32_CalculateTWait(
+		unsigned int PrefetchMode,
+		enum dm_use_mall_for_pstate_change_mode UseMALLForPStateChange,
+		bool SynchronizeDRRDisplaysForUCLKPStateChangeFinal,
+		bool DRRDisplay,
+		double DRAMClockChangeLatency,
+		double FCLKChangeLatency,
+		double UrgentLatency,
+		double SREnterPlusExitTime);
+
+double dml32_get_return_bw_mbps(const soc_bounding_box_st *soc,
+		const int VoltageLevel,
+		const bool HostVMEnable,
+		const double DCFCLK,
+		const double FabricClock,
+		const double DRAMSpeed);
+
+double dml32_get_return_bw_mbps_vm_only(const soc_bounding_box_st *soc,
+		const int VoltageLevel,
+		const double DCFCLK,
+		const double FabricClock,
+		const double DRAMSpeed);
+
+double dml32_CalculateExtraLatency(
+		unsigned int RoundTripPingLatencyCycles,
+		unsigned int ReorderingBytes,
+		double DCFCLK,
+		unsigned int TotalNumberOfActiveDPP,
+		unsigned int PixelChunkSizeInKByte,
+		unsigned int TotalNumberOfDCCActiveDPP,
+		unsigned int MetaChunkSize,
+		double ReturnBW,
+		bool GPUVMEnable,
+		bool HostVMEnable,
+		unsigned int NumberOfActiveSurfaces,
+		unsigned int NumberOfDPP[],
+		unsigned int dpte_group_bytes[],
+		double HostVMInefficiencyFactor,
+		double HostVMMinPageSize,
+		unsigned int HostVMMaxNonCachedPageTableLevels);
+
+bool dml32_CalculatePrefetchSchedule(
+		struct vba_vars_st *v,
+		unsigned int k,
+		double HostVMInefficiencyFactor,
+		DmlPipe *myPipe,
+		unsigned int DSCDelay,
+		unsigned int DPP_RECOUT_WIDTH,
+		unsigned int VStartup,
+		unsigned int MaxVStartup,
+		double UrgentLatency,
+		double UrgentExtraLatency,
+		double TCalc,
+		unsigned int PDEAndMetaPTEBytesFrame,
+		unsigned int MetaRowByte,
+		unsigned int PixelPTEBytesPerRow,
+		double PrefetchSourceLinesY,
+		unsigned int SwathWidthY,
+		unsigned int VInitPreFillY,
+		unsigned int MaxNumSwathY,
+		double PrefetchSourceLinesC,
+		unsigned int SwathWidthC,
+		unsigned int VInitPreFillC,
+		unsigned int MaxNumSwathC,
+		unsigned int swath_width_luma_ub,
+		unsigned int swath_width_chroma_ub,
+		unsigned int SwathHeightY,
+		unsigned int SwathHeightC,
+		double TWait,
+		double TPreReq,
+		bool ExtendPrefetchIfPossible,
+		/* Output */
+		double   *DSTXAfterScaler,
+		double   *DSTYAfterScaler,
+		double *DestinationLinesForPrefetch,
+		double *PrefetchBandwidth,
+		double *DestinationLinesToRequestVMInVBlank,
+		double *DestinationLinesToRequestRowInVBlank,
+		double *VRatioPrefetchY,
+		double *VRatioPrefetchC,
+		double *RequiredPrefetchPixDataBWLuma,
+		double *RequiredPrefetchPixDataBWChroma,
+		bool   *NotEnoughTimeForDynamicMetadata,
+		double *Tno_bw,
+		double *prefetch_vmrow_bw,
+		double *Tdmdl_vm,
+		double *Tdmdl,
+		double *TSetup,
+		unsigned int   *VUpdateOffsetPix,
+		double   *VUpdateWidthPix,
+		double   *VReadyOffsetPix);
+
+void dml32_CalculateFlipSchedule(
+		double HostVMInefficiencyFactor,
+		double UrgentExtraLatency,
+		double UrgentLatency,
+		unsigned int GPUVMMaxPageTableLevels,
+		bool HostVMEnable,
+		unsigned int HostVMMaxNonCachedPageTableLevels,
+		bool GPUVMEnable,
+		double HostVMMinPageSize,
+		double PDEAndMetaPTEBytesPerFrame,
+		double MetaRowBytes,
+		double DPTEBytesPerRow,
+		double BandwidthAvailableForImmediateFlip,
+		unsigned int TotImmediateFlipBytes,
+		enum source_format_class SourcePixelFormat,
+		double LineTime,
+		double VRatio,
+		double VRatioChroma,
+		double Tno_bw,
+		bool DCCEnable,
+		unsigned int dpte_row_height,
+		unsigned int meta_row_height,
+		unsigned int dpte_row_height_chroma,
+		unsigned int meta_row_height_chroma,
+		bool    use_one_row_for_frame_flip,
+
+		/* Output */
+		double *DestinationLinesToRequestVMInImmediateFlip,
+		double *DestinationLinesToRequestRowInImmediateFlip,
+		double *final_flip_bw,
+		bool *ImmediateFlipSupportedForPipe);
+
+void dml32_CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport(
+		struct vba_vars_st *v,
+		unsigned int PrefetchMode,
+		double DCFCLK,
+		double ReturnBW,
+		SOCParametersList mmSOCParameters,
+		double SOCCLK,
+		double DCFClkDeepSleep,
+		unsigned int DETBufferSizeY[],
+		unsigned int DETBufferSizeC[],
+		unsigned int SwathHeightY[],
+		unsigned int SwathHeightC[],
+		double SwathWidthY[],
+		double SwathWidthC[],
+		unsigned int DPPPerSurface[],
+		double BytePerPixelDETY[],
+		double BytePerPixelDETC[],
+		double DSTXAfterScaler[],
+		double DSTYAfterScaler[],
+		bool UnboundedRequestEnabled,
+		unsigned int CompressedBufferSizeInkByte,
+
+		/* Output */
+		enum clock_change_support *DRAMClockChangeSupport,
+		double MaxActiveDRAMClockChangeLatencySupported[],
+		unsigned int SubViewportLinesNeededInMALL[],
+		enum dm_fclock_change_support *FCLKChangeSupport,
+		double *MinActiveFCLKChangeLatencySupported,
+		bool *USRRetrainingSupport,
+		double ActiveDRAMClockChangeLatencyMargin[]);
+
+double dml32_CalculateWriteBackDISPCLK(
+		enum source_format_class WritebackPixelFormat,
+		double PixelClock,
+		double WritebackHRatio,
+		double WritebackVRatio,
+		unsigned int WritebackHTaps,
+		unsigned int WritebackVTaps,
+		unsigned int   WritebackSourceWidth,
+		unsigned int   WritebackDestinationWidth,
+		unsigned int HTotal,
+		unsigned int WritebackLineBufferSize,
+		double DISPCLKDPPCLKVCOSpeed);
+
+void dml32_CalculateMinAndMaxPrefetchMode(
+		enum dm_prefetch_modes   AllowForPStateChangeOrStutterInVBlankFinal,
+		unsigned int             *MinPrefetchMode,
+		unsigned int             *MaxPrefetchMode);
+
+void dml32_CalculatePixelDeliveryTimes(
+		unsigned int             NumberOfActiveSurfaces,
+		double              VRatio[],
+		double              VRatioChroma[],
+		double              VRatioPrefetchY[],
+		double              VRatioPrefetchC[],
+		unsigned int             swath_width_luma_ub[],
+		unsigned int             swath_width_chroma_ub[],
+		unsigned int             DPPPerSurface[],
+		double              HRatio[],
+		double              HRatioChroma[],
+		double              PixelClock[],
+		double              PSCL_THROUGHPUT[],
+		double              PSCL_THROUGHPUT_CHROMA[],
+		double              Dppclk[],
+		unsigned int             BytePerPixelC[],
+		enum dm_rotation_angle   SourceRotation[],
+		unsigned int             NumberOfCursors[],
+		unsigned int             CursorWidth[][DC__NUM_CURSOR__MAX],
+		unsigned int             CursorBPP[][DC__NUM_CURSOR__MAX],
+		unsigned int             BlockWidth256BytesY[],
+		unsigned int             BlockHeight256BytesY[],
+		unsigned int             BlockWidth256BytesC[],
+		unsigned int             BlockHeight256BytesC[],
+
+		/* Output */
+		double              DisplayPipeLineDeliveryTimeLuma[],
+		double              DisplayPipeLineDeliveryTimeChroma[],
+		double              DisplayPipeLineDeliveryTimeLumaPrefetch[],
+		double              DisplayPipeLineDeliveryTimeChromaPrefetch[],
+		double              DisplayPipeRequestDeliveryTimeLuma[],
+		double              DisplayPipeRequestDeliveryTimeChroma[],
+		double              DisplayPipeRequestDeliveryTimeLumaPrefetch[],
+		double              DisplayPipeRequestDeliveryTimeChromaPrefetch[],
+		double              CursorRequestDeliveryTime[],
+		double              CursorRequestDeliveryTimePrefetch[]);
+
+void dml32_CalculateMetaAndPTETimes(
+		bool use_one_row_for_frame[],
+		unsigned int NumberOfActiveSurfaces,
+		bool GPUVMEnable,
+		unsigned int MetaChunkSize,
+		unsigned int MinMetaChunkSizeBytes,
+		unsigned int    HTotal[],
+		double  VRatio[],
+		double  VRatioChroma[],
+		double  DestinationLinesToRequestRowInVBlank[],
+		double  DestinationLinesToRequestRowInImmediateFlip[],
+		bool DCCEnable[],
+		double  PixelClock[],
+		unsigned int BytePerPixelY[],
+		unsigned int BytePerPixelC[],
+		enum dm_rotation_angle SourceRotation[],
+		unsigned int dpte_row_height[],
+		unsigned int dpte_row_height_chroma[],
+		unsigned int meta_row_width[],
+		unsigned int meta_row_width_chroma[],
+		unsigned int meta_row_height[],
+		unsigned int meta_row_height_chroma[],
+		unsigned int meta_req_width[],
+		unsigned int meta_req_width_chroma[],
+		unsigned int meta_req_height[],
+		unsigned int meta_req_height_chroma[],
+		unsigned int dpte_group_bytes[],
+		unsigned int    PTERequestSizeY[],
+		unsigned int    PTERequestSizeC[],
+		unsigned int    PixelPTEReqWidthY[],
+		unsigned int    PixelPTEReqHeightY[],
+		unsigned int    PixelPTEReqWidthC[],
+		unsigned int    PixelPTEReqHeightC[],
+		unsigned int    dpte_row_width_luma_ub[],
+		unsigned int    dpte_row_width_chroma_ub[],
+
+		/* Output */
+		double DST_Y_PER_PTE_ROW_NOM_L[],
+		double DST_Y_PER_PTE_ROW_NOM_C[],
+		double DST_Y_PER_META_ROW_NOM_L[],
+		double DST_Y_PER_META_ROW_NOM_C[],
+		double TimePerMetaChunkNominal[],
+		double TimePerChromaMetaChunkNominal[],
+		double TimePerMetaChunkVBlank[],
+		double TimePerChromaMetaChunkVBlank[],
+		double TimePerMetaChunkFlip[],
+		double TimePerChromaMetaChunkFlip[],
+		double time_per_pte_group_nom_luma[],
+		double time_per_pte_group_vblank_luma[],
+		double time_per_pte_group_flip_luma[],
+		double time_per_pte_group_nom_chroma[],
+		double time_per_pte_group_vblank_chroma[],
+		double time_per_pte_group_flip_chroma[]);
+
+void dml32_CalculateVMGroupAndRequestTimes(
+		unsigned int     NumberOfActiveSurfaces,
+		bool     GPUVMEnable,
+		unsigned int     GPUVMMaxPageTableLevels,
+		unsigned int     HTotal[],
+		unsigned int     BytePerPixelC[],
+		double      DestinationLinesToRequestVMInVBlank[],
+		double      DestinationLinesToRequestVMInImmediateFlip[],
+		bool     DCCEnable[],
+		double      PixelClock[],
+		unsigned int        dpte_row_width_luma_ub[],
+		unsigned int        dpte_row_width_chroma_ub[],
+		unsigned int     vm_group_bytes[],
+		unsigned int     dpde0_bytes_per_frame_ub_l[],
+		unsigned int     dpde0_bytes_per_frame_ub_c[],
+		unsigned int        meta_pte_bytes_per_frame_ub_l[],
+		unsigned int        meta_pte_bytes_per_frame_ub_c[],
+
+		/* Output */
+		double      TimePerVMGroupVBlank[],
+		double      TimePerVMGroupFlip[],
+		double      TimePerVMRequestVBlank[],
+		double      TimePerVMRequestFlip[]);
+
+void dml32_CalculateDCCConfiguration(
+		bool             DCCEnabled,
+		bool             DCCProgrammingAssumesScanDirectionUnknown,
+		enum source_format_class SourcePixelFormat,
+		unsigned int             SurfaceWidthLuma,
+		unsigned int             SurfaceWidthChroma,
+		unsigned int             SurfaceHeightLuma,
+		unsigned int             SurfaceHeightChroma,
+		unsigned int                nomDETInKByte,
+		unsigned int             RequestHeight256ByteLuma,
+		unsigned int             RequestHeight256ByteChroma,
+		enum dm_swizzle_mode     TilingFormat,
+		unsigned int             BytePerPixelY,
+		unsigned int             BytePerPixelC,
+		double              BytePerPixelDETY,
+		double              BytePerPixelDETC,
+		enum dm_rotation_angle   SourceRotation,
+		/* Output */
+		unsigned int        *MaxUncompressedBlockLuma,
+		unsigned int        *MaxUncompressedBlockChroma,
+		unsigned int        *MaxCompressedBlockLuma,
+		unsigned int        *MaxCompressedBlockChroma,
+		unsigned int        *IndependentBlockLuma,
+		unsigned int        *IndependentBlockChroma);
+
+void dml32_CalculateStutterEfficiency(
+		unsigned int      CompressedBufferSizeInkByte,
+		enum dm_use_mall_for_pstate_change_mode UseMALLForPStateChange[],
+		bool   UnboundedRequestEnabled,
+		unsigned int      MetaFIFOSizeInKEntries,
+		unsigned int      ZeroSizeBufferEntries,
+		unsigned int      PixelChunkSizeInKByte,
+		unsigned int   NumberOfActiveSurfaces,
+		unsigned int      ROBBufferSizeInKByte,
+		double    TotalDataReadBandwidth,
+		double    DCFCLK,
+		double    ReturnBW,
+		unsigned int      CompbufReservedSpace64B,
+		unsigned int      CompbufReservedSpaceZs,
+		double    SRExitTime,
+		double    SRExitZ8Time,
+		bool   SynchronizeTimingsFinal,
+		unsigned int   BlendingAndTiming[],
+		double    StutterEnterPlusExitWatermark,
+		double    Z8StutterEnterPlusExitWatermark,
+		bool   ProgressiveToInterlaceUnitInOPP,
+		bool   Interlace[],
+		double    MinTTUVBlank[],
+		unsigned int   DPPPerSurface[],
+		unsigned int      DETBufferSizeY[],
+		unsigned int   BytePerPixelY[],
+		double    BytePerPixelDETY[],
+		double      SwathWidthY[],
+		unsigned int   SwathHeightY[],
+		unsigned int   SwathHeightC[],
+		double    NetDCCRateLuma[],
+		double    NetDCCRateChroma[],
+		double    DCCFractionOfZeroSizeRequestsLuma[],
+		double    DCCFractionOfZeroSizeRequestsChroma[],
+		unsigned int      HTotal[],
+		unsigned int      VTotal[],
+		double    PixelClock[],
+		double    VRatio[],
+		enum dm_rotation_angle SourceRotation[],
+		unsigned int   BlockHeight256BytesY[],
+		unsigned int   BlockWidth256BytesY[],
+		unsigned int   BlockHeight256BytesC[],
+		unsigned int   BlockWidth256BytesC[],
+		unsigned int   DCCYMaxUncompressedBlock[],
+		unsigned int   DCCCMaxUncompressedBlock[],
+		unsigned int      VActive[],
+		bool   DCCEnable[],
+		bool   WritebackEnable[],
+		double    ReadBandwidthSurfaceLuma[],
+		double    ReadBandwidthSurfaceChroma[],
+		double    meta_row_bw[],
+		double    dpte_row_bw[],
+
+		/* Output */
+		double   *StutterEfficiencyNotIncludingVBlank,
+		double   *StutterEfficiency,
+		unsigned int     *NumberOfStutterBurstsPerFrame,
+		double   *Z8StutterEfficiencyNotIncludingVBlank,
+		double   *Z8StutterEfficiency,
+		unsigned int     *Z8NumberOfStutterBurstsPerFrame,
+		double   *StutterPeriod,
+		bool  *DCHUBBUB_ARB_CSTATE_MAX_CAP_MODE);
+
+void dml32_CalculateMaxDETAndMinCompressedBufferSize(
+		unsigned int    ConfigReturnBufferSizeInKByte,
+		unsigned int    ROBBufferSizeInKByte,
+		unsigned int MaxNumDPP,
+		bool nomDETInKByteOverrideEnable, // VBA_DELTA, allow DV to override default DET size
+		unsigned int nomDETInKByteOverrideValue,  // VBA_DELTA
+
+		/* Output */
+		unsigned int *MaxTotalDETInKByte,
+		unsigned int *nomDETInKByte,
+		unsigned int *MinCompressedBufferSizeInKByte);
+
+bool dml32_CalculateVActiveBandwithSupport(unsigned int NumberOfActiveSurfaces,
+		double ReturnBW,
+		bool NotUrgentLatencyHiding[],
+		double ReadBandwidthLuma[],
+		double ReadBandwidthChroma[],
+		double cursor_bw[],
+		double meta_row_bandwidth[],
+		double dpte_row_bandwidth[],
+		unsigned int NumberOfDPP[],
+		double UrgentBurstFactorLuma[],
+		double UrgentBurstFactorChroma[],
+		double UrgentBurstFactorCursor[]);
+
+void dml32_CalculatePrefetchBandwithSupport(unsigned int NumberOfActiveSurfaces,
+		double ReturnBW,
+		bool NotUrgentLatencyHiding[],
+		double ReadBandwidthLuma[],
+		double ReadBandwidthChroma[],
+		double PrefetchBandwidthLuma[],
+		double PrefetchBandwidthChroma[],
+		double cursor_bw[],
+		double meta_row_bandwidth[],
+		double dpte_row_bandwidth[],
+		double cursor_bw_pre[],
+		double prefetch_vmrow_bw[],
+		unsigned int NumberOfDPP[],
+		double UrgentBurstFactorLuma[],
+		double UrgentBurstFactorChroma[],
+		double UrgentBurstFactorCursor[],
+		double UrgentBurstFactorLumaPre[],
+		double UrgentBurstFactorChromaPre[],
+		double UrgentBurstFactorCursorPre[],
+		double PrefetchBW[],
+		double VRatio[],
+		double MaxVRatioPre,
+
+		/* output */
+		double  *MaxPrefetchBandwidth,
+		double  *FractionOfUrgentBandwidth,
+		bool *PrefetchBandwidthSupport);
+
+double dml32_CalculateBandwidthAvailableForImmediateFlip(unsigned int NumberOfActiveSurfaces,
+		double ReturnBW,
+		double ReadBandwidthLuma[],
+		double ReadBandwidthChroma[],
+		double PrefetchBandwidthLuma[],
+		double PrefetchBandwidthChroma[],
+		double cursor_bw[],
+		double cursor_bw_pre[],
+		unsigned int NumberOfDPP[],
+		double UrgentBurstFactorLuma[],
+		double UrgentBurstFactorChroma[],
+		double UrgentBurstFactorCursor[],
+		double UrgentBurstFactorLumaPre[],
+		double UrgentBurstFactorChromaPre[],
+		double UrgentBurstFactorCursorPre[]);
+
+void dml32_CalculateImmediateFlipBandwithSupport(unsigned int NumberOfActiveSurfaces,
+		double ReturnBW,
+		enum immediate_flip_requirement ImmediateFlipRequirement[],
+		double final_flip_bw[],
+		double ReadBandwidthLuma[],
+		double ReadBandwidthChroma[],
+		double PrefetchBandwidthLuma[],
+		double PrefetchBandwidthChroma[],
+		double cursor_bw[],
+		double meta_row_bandwidth[],
+		double dpte_row_bandwidth[],
+		double cursor_bw_pre[],
+		double prefetch_vmrow_bw[],
+		unsigned int NumberOfDPP[],
+		double UrgentBurstFactorLuma[],
+		double UrgentBurstFactorChroma[],
+		double UrgentBurstFactorCursor[],
+		double UrgentBurstFactorLumaPre[],
+		double UrgentBurstFactorChromaPre[],
+		double UrgentBurstFactorCursorPre[],
+
+		/* output */
+		double  *TotalBandwidth,
+		double  *FractionOfUrgentBandwidth,
+		bool *ImmediateFlipBandwidthSupport);
+
+bool dml32_CalculateDETSwathFillLatencyHiding(unsigned int NumberOfActiveSurfaces,
+		double ReturnBW,
+		double UrgentLatency,
+		unsigned int SwathHeightY[],
+		unsigned int SwathHeightC[],
+		unsigned int SwathWidthY[],
+		unsigned int SwathWidthC[],
+		double  BytePerPixelInDETY[],
+		double  BytePerPixelInDETC[],
+		unsigned int    DETBufferSizeY[],
+		unsigned int    DETBufferSizeC[],
+		unsigned int	NumOfDPP[],
+		unsigned int	HTotal[],
+		double	PixelClock[],
+		double	VRatioY[],
+		double	VRatioC[],
+		enum dm_use_mall_for_pstate_change_mode UsesMALLForPStateChange[],
+		enum unbounded_requesting_policy UseUnboundedRequesting);
+
+#endif
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_rq_dlg_calc_32.c b/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_rq_dlg_calc_32.c
new file mode 100644
index 000000000000..6c75aa82327a
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_rq_dlg_calc_32.c
@@ -0,0 +1,614 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+
+#include "../display_mode_lib.h"
+#include "../display_mode_vba.h"
+#include "../dml_inline_defs.h"
+#include "display_rq_dlg_calc_32.h"
+
+static bool is_dual_plane(enum source_format_class source_format)
+{
+	bool ret_val = 0;
+
+	if ((source_format == dm_420_12) || (source_format == dm_420_8) || (source_format == dm_420_10)
+		|| (source_format == dm_rgbe_alpha))
+		ret_val = 1;
+
+	return ret_val;
+}
+
+void dml32_rq_dlg_get_rq_reg(display_rq_regs_st *rq_regs,
+		struct display_mode_lib *mode_lib,
+		const display_e2e_pipe_params_st *e2e_pipe_param,
+		const unsigned int num_pipes,
+		const unsigned int pipe_idx)
+{
+	const display_pipe_source_params_st *src = &e2e_pipe_param[pipe_idx].pipe.src;
+	bool dual_plane = is_dual_plane((enum source_format_class) (src->source_format));
+	double stored_swath_l_bytes;
+	double stored_swath_c_bytes;
+	bool is_phantom_pipe;
+	uint32_t pixel_chunk_bytes = 0;
+	uint32_t min_pixel_chunk_bytes = 0;
+	uint32_t meta_chunk_bytes = 0;
+	uint32_t min_meta_chunk_bytes = 0;
+	uint32_t dpte_group_bytes = 0;
+	uint32_t mpte_group_bytes = 0;
+
+	uint32_t p1_pixel_chunk_bytes = 0;
+	uint32_t p1_min_pixel_chunk_bytes = 0;
+	uint32_t p1_meta_chunk_bytes = 0;
+	uint32_t p1_min_meta_chunk_bytes = 0;
+	uint32_t p1_dpte_group_bytes = 0;
+	uint32_t p1_mpte_group_bytes = 0;
+
+	unsigned int detile_buf_size_in_bytes;
+	unsigned int detile_buf_plane1_addr;
+	unsigned int pte_row_height_linear;
+
+	memset(rq_regs, 0, sizeof(*rq_regs));
+
+	dml_print("DML_DLG::%s: Calculation for pipe[%d] start, num_pipes=%d\n", __func__, pipe_idx, num_pipes);
+
+	pixel_chunk_bytes = get_pixel_chunk_size_in_kbyte(mode_lib, e2e_pipe_param, num_pipes) * 1024; // From VBA
+	min_pixel_chunk_bytes = get_min_pixel_chunk_size_in_byte(mode_lib, e2e_pipe_param, num_pipes); // From VBA
+
+	if (pixel_chunk_bytes == 64 * 1024)
+		min_pixel_chunk_bytes = 0;
+
+	meta_chunk_bytes = get_meta_chunk_size_in_kbyte(mode_lib, e2e_pipe_param, num_pipes) * 1024; // From VBA
+	min_meta_chunk_bytes = get_min_meta_chunk_size_in_byte(mode_lib, e2e_pipe_param, num_pipes); // From VBA
+
+	dpte_group_bytes = get_dpte_group_size_in_bytes(mode_lib, e2e_pipe_param, num_pipes, pipe_idx); // From VBA
+	mpte_group_bytes = get_vm_group_size_in_bytes(mode_lib, e2e_pipe_param, num_pipes, pipe_idx); // From VBA
+
+	p1_pixel_chunk_bytes = pixel_chunk_bytes;
+	p1_min_pixel_chunk_bytes = min_pixel_chunk_bytes;
+	p1_meta_chunk_bytes = meta_chunk_bytes;
+	p1_min_meta_chunk_bytes = min_meta_chunk_bytes;
+	p1_dpte_group_bytes = dpte_group_bytes;
+	p1_mpte_group_bytes = mpte_group_bytes;
+
+	if ((enum source_format_class) src->source_format == dm_rgbe_alpha)
+		p1_pixel_chunk_bytes = get_alpha_pixel_chunk_size_in_kbyte(mode_lib, e2e_pipe_param, num_pipes) * 1024;
+
+	rq_regs->rq_regs_l.chunk_size = dml_log2(pixel_chunk_bytes) - 10;
+	rq_regs->rq_regs_c.chunk_size = dml_log2(p1_pixel_chunk_bytes) - 10;
+
+	if (min_pixel_chunk_bytes == 0)
+		rq_regs->rq_regs_l.min_chunk_size = 0;
+	else
+		rq_regs->rq_regs_l.min_chunk_size = dml_log2(min_pixel_chunk_bytes) - 8 + 1;
+
+	if (p1_min_pixel_chunk_bytes == 0)
+		rq_regs->rq_regs_c.min_chunk_size = 0;
+	else
+		rq_regs->rq_regs_c.min_chunk_size = dml_log2(p1_min_pixel_chunk_bytes) - 8 + 1;
+
+	rq_regs->rq_regs_l.meta_chunk_size = dml_log2(meta_chunk_bytes) - 10;
+	rq_regs->rq_regs_c.meta_chunk_size = dml_log2(p1_meta_chunk_bytes) - 10;
+
+	if (min_meta_chunk_bytes == 0)
+		rq_regs->rq_regs_l.min_meta_chunk_size = 0;
+	else
+		rq_regs->rq_regs_l.min_meta_chunk_size = dml_log2(min_meta_chunk_bytes) - 6 + 1;
+
+	if (p1_min_meta_chunk_bytes == 0)
+		rq_regs->rq_regs_c.min_meta_chunk_size = 0;
+	else
+		rq_regs->rq_regs_c.min_meta_chunk_size = dml_log2(p1_min_meta_chunk_bytes) - 6 + 1;
+
+	rq_regs->rq_regs_l.dpte_group_size = dml_log2(dpte_group_bytes) - 6;
+	rq_regs->rq_regs_l.mpte_group_size = dml_log2(mpte_group_bytes) - 6;
+	rq_regs->rq_regs_c.dpte_group_size = dml_log2(p1_dpte_group_bytes) - 6;
+	rq_regs->rq_regs_c.mpte_group_size = dml_log2(p1_mpte_group_bytes) - 6;
+
+	detile_buf_size_in_bytes = get_det_buffer_size_kbytes(mode_lib, e2e_pipe_param, num_pipes, pipe_idx) * 1024;
+	detile_buf_plane1_addr = 0;
+	pte_row_height_linear = get_dpte_row_height_linear_l(mode_lib, e2e_pipe_param, num_pipes,
+			pipe_idx);
+
+	if (src->sw_mode == dm_sw_linear)
+		ASSERT(pte_row_height_linear >= 8);
+
+	rq_regs->rq_regs_l.pte_row_height_linear = dml_floor(dml_log2(pte_row_height_linear), 1) - 3;
+
+	if (dual_plane) {
+		unsigned int p1_pte_row_height_linear = get_dpte_row_height_linear_c(mode_lib, e2e_pipe_param,
+				num_pipes, pipe_idx);
+		if (src->sw_mode == dm_sw_linear)
+			ASSERT(p1_pte_row_height_linear >= 8);
+
+		rq_regs->rq_regs_c.pte_row_height_linear = dml_floor(dml_log2(p1_pte_row_height_linear), 1) - 3;
+	}
+
+	rq_regs->rq_regs_l.swath_height = dml_log2(get_swath_height_l(mode_lib, e2e_pipe_param, num_pipes, pipe_idx));
+	rq_regs->rq_regs_c.swath_height = dml_log2(get_swath_height_c(mode_lib, e2e_pipe_param, num_pipes, pipe_idx));
+
+	// FIXME: take the max between luma, chroma chunk size?
+	// okay for now, as we are setting pixel_chunk_bytes to 8kb anyways
+	if (pixel_chunk_bytes >= 32 * 1024 || (dual_plane && p1_pixel_chunk_bytes >= 32 * 1024)) { //32kb
+		rq_regs->drq_expansion_mode = 0;
+	} else {
+		rq_regs->drq_expansion_mode = 2;
+	}
+	rq_regs->prq_expansion_mode = 1;
+	rq_regs->mrq_expansion_mode = 1;
+	rq_regs->crq_expansion_mode = 1;
+
+	stored_swath_l_bytes = get_det_stored_buffer_size_l_bytes(mode_lib, e2e_pipe_param, num_pipes,
+			pipe_idx);
+	stored_swath_c_bytes = get_det_stored_buffer_size_c_bytes(mode_lib, e2e_pipe_param, num_pipes,
+			pipe_idx);
+	is_phantom_pipe = get_is_phantom_pipe(mode_lib, e2e_pipe_param, num_pipes, pipe_idx);
+
+	// Note: detile_buf_plane1_addr is in unit of 1KB
+	if (dual_plane) {
+		if (is_phantom_pipe) {
+			detile_buf_plane1_addr = ((1024.0 * 1024.0) / 2.0 / 1024.0); // half to chroma
+		} else {
+			if (stored_swath_l_bytes / stored_swath_c_bytes <= 1.5) {
+				detile_buf_plane1_addr = (detile_buf_size_in_bytes / 2.0 / 1024.0); // half to chroma
+#ifdef __DML_RQ_DLG_CALC_DEBUG__
+				dml_print("DML_DLG: %s: detile_buf_plane1_addr = %d (1/2 to chroma)\n",
+						__func__, detile_buf_plane1_addr);
+#endif
+			} else {
+				detile_buf_plane1_addr =
+						dml_round_to_multiple(
+								(unsigned int) ((2.0 * detile_buf_size_in_bytes) / 3.0),
+								1024, 0) / 1024.0; // 2/3 to luma
+#ifdef __DML_RQ_DLG_CALC_DEBUG__
+				dml_print("DML_DLG: %s: detile_buf_plane1_addr = %d (1/3 chroma)\n",
+						__func__, detile_buf_plane1_addr);
+#endif
+			}
+		}
+	}
+	rq_regs->plane1_base_address = detile_buf_plane1_addr;
+
+#ifdef __DML_RQ_DLG_CALC_DEBUG__
+	dml_print("DML_DLG: %s: is_phantom_pipe = %d\n", __func__, is_phantom_pipe);
+	dml_print("DML_DLG: %s: stored_swath_l_bytes = %f\n", __func__, stored_swath_l_bytes);
+	dml_print("DML_DLG: %s: stored_swath_c_bytes = %f\n", __func__, stored_swath_c_bytes);
+	dml_print("DML_DLG: %s: detile_buf_size_in_bytes = %d\n", __func__, detile_buf_size_in_bytes);
+	dml_print("DML_DLG: %s: detile_buf_plane1_addr = %d\n", __func__, detile_buf_plane1_addr);
+	dml_print("DML_DLG: %s: plane1_base_address = %d\n", __func__, rq_regs->plane1_base_address);
+#endif
+	print__rq_regs_st(mode_lib, rq_regs);
+	dml_print("DML_DLG::%s: Calculation for pipe[%d] done, num_pipes=%d\n", __func__, pipe_idx, num_pipes);
+}
+
+void dml32_rq_dlg_get_dlg_reg(struct display_mode_lib *mode_lib,
+		display_dlg_regs_st *dlg_regs,
+		display_ttu_regs_st *ttu_regs,
+		display_e2e_pipe_params_st *e2e_pipe_param,
+		const unsigned int num_pipes,
+		const unsigned int pipe_idx)
+{
+	const display_pipe_source_params_st *src = &e2e_pipe_param[pipe_idx].pipe.src;
+	const display_pipe_dest_params_st *dst = &e2e_pipe_param[pipe_idx].pipe.dest;
+	const display_clocks_and_cfg_st *clks = &e2e_pipe_param[pipe_idx].clks_cfg;
+	double refcyc_per_req_delivery_pre_cur0 = 0.;
+	double refcyc_per_req_delivery_cur0 = 0.;
+	double refcyc_per_req_delivery_pre_c = 0.;
+	double refcyc_per_req_delivery_c = 0.;
+	double refcyc_per_req_delivery_pre_l;
+	double refcyc_per_req_delivery_l;
+	double refcyc_per_line_delivery_pre_c = 0.;
+	double refcyc_per_line_delivery_c = 0.;
+	double refcyc_per_line_delivery_pre_l;
+	double refcyc_per_line_delivery_l;
+	double min_ttu_vblank;
+	double vratio_pre_l;
+	double vratio_pre_c;
+	unsigned int min_dst_y_next_start;
+	unsigned int htotal = dst->htotal;
+	unsigned int hblank_end = dst->hblank_end;
+	unsigned int vblank_end = dst->vblank_end;
+	bool interlaced = dst->interlaced;
+	double pclk_freq_in_mhz = dst->pixel_rate_mhz;
+	unsigned int vready_after_vcount0;
+	double refclk_freq_in_mhz = clks->refclk_mhz;
+	double ref_freq_to_pix_freq = refclk_freq_in_mhz / pclk_freq_in_mhz;
+	bool dual_plane = 0;
+	unsigned int pipe_index_in_combine[DC__NUM_PIPES__MAX];
+	unsigned int dst_x_after_scaler;
+	unsigned int dst_y_after_scaler;
+	double dst_y_prefetch;
+	double dst_y_per_vm_vblank;
+	double dst_y_per_row_vblank;
+	double dst_y_per_vm_flip;
+	double dst_y_per_row_flip;
+	double max_dst_y_per_vm_vblank = 32.0;
+	double max_dst_y_per_row_vblank = 16.0;
+	double dst_y_per_pte_row_nom_l;
+	double dst_y_per_pte_row_nom_c;
+	double dst_y_per_meta_row_nom_l;
+	double dst_y_per_meta_row_nom_c;
+	double refcyc_per_pte_group_nom_l;
+	double refcyc_per_pte_group_nom_c;
+	double refcyc_per_pte_group_vblank_l;
+	double refcyc_per_pte_group_vblank_c;
+	double refcyc_per_pte_group_flip_l;
+	double refcyc_per_pte_group_flip_c;
+	double refcyc_per_meta_chunk_nom_l;
+	double refcyc_per_meta_chunk_nom_c;
+	double refcyc_per_meta_chunk_vblank_l;
+	double refcyc_per_meta_chunk_vblank_c;
+	double refcyc_per_meta_chunk_flip_l;
+	double refcyc_per_meta_chunk_flip_c;
+
+	memset(dlg_regs, 0, sizeof(*dlg_regs));
+	memset(ttu_regs, 0, sizeof(*ttu_regs));
+	dml_print("DML_DLG::%s: Calculation for pipe[%d] starts, num_pipes=%d\n", __func__, pipe_idx, num_pipes);
+	dml_print("DML_DLG: %s: refclk_freq_in_mhz     = %3.2f\n", __func__, refclk_freq_in_mhz);
+	dml_print("DML_DLG: %s: pclk_freq_in_mhz = %3.2f\n", __func__, pclk_freq_in_mhz);
+	dml_print("DML_DLG: %s: ref_freq_to_pix_freq   = %3.2f\n", __func__, ref_freq_to_pix_freq);
+	dml_print("DML_DLG: %s: interlaced = %d\n", __func__, interlaced);
+	ASSERT(ref_freq_to_pix_freq < 4.0);
+
+	dlg_regs->ref_freq_to_pix_freq = (unsigned int) (ref_freq_to_pix_freq * dml_pow(2, 19));
+	dlg_regs->refcyc_per_htotal = (unsigned int) (ref_freq_to_pix_freq * (double) htotal * dml_pow(2, 8));
+	dlg_regs->dlg_vblank_end = interlaced ? (vblank_end / 2) : vblank_end; // 15 bits
+
+	min_ttu_vblank = get_min_ttu_vblank_in_us(mode_lib, e2e_pipe_param, num_pipes, pipe_idx); // From VBA
+	min_dst_y_next_start = get_min_dst_y_next_start(mode_lib, e2e_pipe_param, num_pipes, pipe_idx);
+
+	dml_print("DML_DLG: %s: min_ttu_vblank (us)    = %3.2f\n", __func__, min_ttu_vblank);
+	dml_print("DML_DLG: %s: min_dst_y_next_start = %d\n", __func__, min_dst_y_next_start);
+	dml_print("DML_DLG: %s: ref_freq_to_pix_freq   = %3.2f\n", __func__, ref_freq_to_pix_freq);
+
+	dual_plane = is_dual_plane((enum source_format_class) (src->source_format));
+
+	vready_after_vcount0 = get_vready_at_or_after_vsync(mode_lib, e2e_pipe_param, num_pipes,
+			pipe_idx); // From VBA
+	dlg_regs->vready_after_vcount0 = vready_after_vcount0;
+
+	dml_print("DML_DLG: %s: vready_after_vcount0 = %d\n", __func__, dlg_regs->vready_after_vcount0);
+
+	dst_x_after_scaler = dml_ceil(get_dst_x_after_scaler(mode_lib, e2e_pipe_param, num_pipes, pipe_idx), 1);
+	dst_y_after_scaler = dml_ceil(get_dst_y_after_scaler(mode_lib, e2e_pipe_param, num_pipes, pipe_idx), 1);
+
+	// do some adjustment on the dst_after scaler to account for odm combine mode
+	dml_print("DML_DLG: %s: input dst_x_after_scaler   = %d\n", __func__, dst_x_after_scaler);
+	dml_print("DML_DLG: %s: input dst_y_after_scaler   = %d\n", __func__, dst_y_after_scaler);
+
+	// need to figure out which side of odm combine we're in
+	if (dst->odm_combine == dm_odm_combine_mode_2to1 || dst->odm_combine == dm_odm_combine_mode_4to1) {
+		// figure out which pipes go together
+		bool visited[DC__NUM_PIPES__MAX];
+		unsigned int i, j, k;
+
+		for (k = 0; k < num_pipes; ++k) {
+			visited[k] = false;
+			pipe_index_in_combine[k] = 0;
+		}
+
+		for (i = 0; i < num_pipes; i++) {
+			if (e2e_pipe_param[i].pipe.src.is_hsplit && !visited[i]) {
+
+				unsigned int grp = e2e_pipe_param[i].pipe.src.hsplit_grp;
+				unsigned int grp_idx = 0;
+
+				for (j = i; j < num_pipes; j++) {
+					if (e2e_pipe_param[j].pipe.src.hsplit_grp == grp
+						&& e2e_pipe_param[j].pipe.src.is_hsplit && !visited[j]) {
+						pipe_index_in_combine[j] = grp_idx;
+						dml_print("DML_DLG: %s: pipe[%d] is in grp %d idx %d\n",
+								__func__, j, grp, grp_idx);
+						grp_idx++;
+						visited[j] = true;
+					}
+				}
+			}
+		}
+	}
+
+	if (dst->odm_combine == dm_odm_combine_mode_disabled) {
+		// FIXME how about ODM split??
+		dlg_regs->refcyc_h_blank_end = (unsigned int) ((double) hblank_end * ref_freq_to_pix_freq);
+	} else {
+		if (dst->odm_combine == dm_odm_combine_mode_2to1 || dst->odm_combine == dm_odm_combine_mode_4to1) {
+			// TODO: We should really check that 4to1 is supported before setting it to 4
+			unsigned int odm_combine_factor = (dst->odm_combine == dm_odm_combine_mode_2to1 ? 2 : 4);
+			unsigned int odm_pipe_index = pipe_index_in_combine[pipe_idx];
+
+			dlg_regs->refcyc_h_blank_end = (unsigned int) (((double) hblank_end
+				+ odm_pipe_index * (double) dst->hactive / odm_combine_factor) * ref_freq_to_pix_freq);
+		}
+	}
+	ASSERT(dlg_regs->refcyc_h_blank_end < (unsigned int)dml_pow(2, 13));
+
+	dml_print("DML_DLG: %s: htotal= %d\n", __func__, htotal);
+	dml_print("DML_DLG: %s: dst_x_after_scaler[%d]= %d\n", __func__, pipe_idx, dst_x_after_scaler);
+	dml_print("DML_DLG: %s: dst_y_after_scaler[%d] = %d\n", __func__, pipe_idx, dst_y_after_scaler);
+
+	dst_y_prefetch = get_dst_y_prefetch(mode_lib, e2e_pipe_param, num_pipes, pipe_idx);        // From VBA
+	// From VBA
+	dst_y_per_vm_vblank = get_dst_y_per_vm_vblank(mode_lib, e2e_pipe_param, num_pipes, pipe_idx);
+	// From VBA
+	dst_y_per_row_vblank = get_dst_y_per_row_vblank(mode_lib, e2e_pipe_param, num_pipes, pipe_idx);
+	dst_y_per_vm_flip = get_dst_y_per_vm_flip(mode_lib, e2e_pipe_param, num_pipes, pipe_idx);    // From VBA
+	dst_y_per_row_flip = get_dst_y_per_row_flip(mode_lib, e2e_pipe_param, num_pipes, pipe_idx);  // From VBA
+
+	// magic!
+	if (htotal <= 75) {
+		max_dst_y_per_vm_vblank = 100.0;
+		max_dst_y_per_row_vblank = 100.0;
+	}
+
+	dml_print("DML_DLG: %s: dst_y_prefetch (after rnd) = %3.2f\n", __func__, dst_y_prefetch);
+	dml_print("DML_DLG: %s: dst_y_per_vm_flip    = %3.2f\n", __func__, dst_y_per_vm_flip);
+	dml_print("DML_DLG: %s: dst_y_per_row_flip   = %3.2f\n", __func__, dst_y_per_row_flip);
+	dml_print("DML_DLG: %s: dst_y_per_vm_vblank  = %3.2f\n", __func__, dst_y_per_vm_vblank);
+	dml_print("DML_DLG: %s: dst_y_per_row_vblank = %3.2f\n", __func__, dst_y_per_row_vblank);
+
+	ASSERT(dst_y_per_vm_vblank < max_dst_y_per_vm_vblank);
+	ASSERT(dst_y_per_row_vblank < max_dst_y_per_row_vblank);
+	ASSERT(dst_y_prefetch > (dst_y_per_vm_vblank + dst_y_per_row_vblank));
+
+	vratio_pre_l = get_vratio_prefetch_l(mode_lib, e2e_pipe_param, num_pipes, pipe_idx);    // From VBA
+	vratio_pre_c = get_vratio_prefetch_c(mode_lib, e2e_pipe_param, num_pipes, pipe_idx);    // From VBA
+
+	dml_print("DML_DLG: %s: vratio_pre_l = %3.2f\n", __func__, vratio_pre_l);
+	dml_print("DML_DLG: %s: vratio_pre_c = %3.2f\n", __func__, vratio_pre_c);
+
+	// Active
+	refcyc_per_line_delivery_pre_l = get_refcyc_per_line_delivery_pre_l_in_us(mode_lib, e2e_pipe_param, num_pipes, pipe_idx) * refclk_freq_in_mhz;   // From VBA
+	refcyc_per_line_delivery_l = get_refcyc_per_line_delivery_l_in_us(mode_lib, e2e_pipe_param, num_pipes,
+			pipe_idx) * refclk_freq_in_mhz;       // From VBA
+
+	dml_print("DML_DLG: %s: refcyc_per_line_delivery_pre_l = %3.2f\n", __func__, refcyc_per_line_delivery_pre_l);
+	dml_print("DML_DLG: %s: refcyc_per_line_delivery_l     = %3.2f\n", __func__, refcyc_per_line_delivery_l);
+
+	if (dual_plane) {
+		refcyc_per_line_delivery_pre_c = get_refcyc_per_line_delivery_pre_c_in_us(mode_lib, e2e_pipe_param,
+				num_pipes, pipe_idx) * refclk_freq_in_mhz;     // From VBA
+		refcyc_per_line_delivery_c = get_refcyc_per_line_delivery_c_in_us(mode_lib, e2e_pipe_param, num_pipes,
+				pipe_idx) * refclk_freq_in_mhz; // From VBA
+
+		dml_print("DML_DLG: %s: refcyc_per_line_delivery_pre_c = %3.2f\n",
+				__func__, refcyc_per_line_delivery_pre_c);
+		dml_print("DML_DLG: %s: refcyc_per_line_delivery_c     = %3.2f\n",
+				__func__, refcyc_per_line_delivery_c);
+	}
+
+	if (src->dynamic_metadata_enable && src->gpuvm)
+		dlg_regs->refcyc_per_vm_dmdata = get_refcyc_per_vm_dmdata_in_us(mode_lib, e2e_pipe_param, num_pipes,
+				pipe_idx) * refclk_freq_in_mhz; // From VBA
+
+	dlg_regs->dmdata_dl_delta = get_dmdata_dl_delta_in_us(mode_lib, e2e_pipe_param, num_pipes, pipe_idx)
+		* refclk_freq_in_mhz; // From VBA
+
+	refcyc_per_req_delivery_pre_l = get_refcyc_per_req_delivery_pre_l_in_us(mode_lib, e2e_pipe_param, num_pipes, pipe_idx) * refclk_freq_in_mhz; // From VBA
+	refcyc_per_req_delivery_l = get_refcyc_per_req_delivery_l_in_us(mode_lib, e2e_pipe_param, num_pipes,
+			pipe_idx) * refclk_freq_in_mhz;     // From VBA
+
+	dml_print("DML_DLG: %s: refcyc_per_req_delivery_pre_l = %3.2f\n", __func__, refcyc_per_req_delivery_pre_l);
+	dml_print("DML_DLG: %s: refcyc_per_req_delivery_l     = %3.2f\n", __func__, refcyc_per_req_delivery_l);
+
+	if (dual_plane) {
+		refcyc_per_req_delivery_pre_c = get_refcyc_per_req_delivery_pre_c_in_us(mode_lib, e2e_pipe_param,
+				num_pipes, pipe_idx) * refclk_freq_in_mhz;  // From VBA
+		refcyc_per_req_delivery_c = get_refcyc_per_req_delivery_c_in_us(mode_lib, e2e_pipe_param, num_pipes,
+				pipe_idx) * refclk_freq_in_mhz;      // From VBA
+
+		dml_print("DML_DLG: %s: refcyc_per_req_delivery_pre_c = %3.2f\n",
+				__func__, refcyc_per_req_delivery_pre_c);
+		dml_print("DML_DLG: %s: refcyc_per_req_delivery_c     = %3.2f\n", __func__, refcyc_per_req_delivery_c);
+	}
+
+	// TTU - Cursor
+	ASSERT(src->num_cursors <= 1);
+	if (src->num_cursors > 0) {
+		refcyc_per_req_delivery_pre_cur0 = get_refcyc_per_cursor_req_delivery_pre_in_us(mode_lib,
+				e2e_pipe_param, num_pipes, pipe_idx) * refclk_freq_in_mhz;  // From VBA
+		refcyc_per_req_delivery_cur0 = get_refcyc_per_cursor_req_delivery_in_us(mode_lib, e2e_pipe_param,
+				num_pipes, pipe_idx) * refclk_freq_in_mhz;      // From VBA
+
+		dml_print("DML_DLG: %s: refcyc_per_req_delivery_pre_cur0 = %3.2f\n",
+				__func__, refcyc_per_req_delivery_pre_cur0);
+		dml_print("DML_DLG: %s: refcyc_per_req_delivery_cur0     = %3.2f\n",
+				__func__, refcyc_per_req_delivery_cur0);
+	}
+
+	// Assign to register structures
+	dlg_regs->min_dst_y_next_start = min_dst_y_next_start * dml_pow(2, 2);
+	ASSERT(dlg_regs->min_dst_y_next_start < (unsigned int)dml_pow(2, 18));
+
+	dlg_regs->dst_y_after_scaler = dst_y_after_scaler; // in terms of line
+	dlg_regs->refcyc_x_after_scaler = dst_x_after_scaler * ref_freq_to_pix_freq; // in terms of refclk
+	dlg_regs->dst_y_prefetch = (unsigned int) (dst_y_prefetch * dml_pow(2, 2));
+	dlg_regs->dst_y_per_vm_vblank = (unsigned int) (dst_y_per_vm_vblank * dml_pow(2, 2));
+	dlg_regs->dst_y_per_row_vblank = (unsigned int) (dst_y_per_row_vblank * dml_pow(2, 2));
+	dlg_regs->dst_y_per_vm_flip = (unsigned int) (dst_y_per_vm_flip * dml_pow(2, 2));
+	dlg_regs->dst_y_per_row_flip = (unsigned int) (dst_y_per_row_flip * dml_pow(2, 2));
+
+	dlg_regs->vratio_prefetch = (unsigned int) (vratio_pre_l * dml_pow(2, 19));
+	dlg_regs->vratio_prefetch_c = (unsigned int) (vratio_pre_c * dml_pow(2, 19));
+
+	dml_print("DML_DLG: %s: dlg_regs->dst_y_per_vm_vblank  = 0x%x\n", __func__, dlg_regs->dst_y_per_vm_vblank);
+	dml_print("DML_DLG: %s: dlg_regs->dst_y_per_row_vblank = 0x%x\n", __func__, dlg_regs->dst_y_per_row_vblank);
+	dml_print("DML_DLG: %s: dlg_regs->dst_y_per_vm_flip    = 0x%x\n", __func__, dlg_regs->dst_y_per_vm_flip);
+	dml_print("DML_DLG: %s: dlg_regs->dst_y_per_row_flip   = 0x%x\n", __func__, dlg_regs->dst_y_per_row_flip);
+
+	dlg_regs->refcyc_per_vm_group_vblank = get_refcyc_per_vm_group_vblank_in_us(mode_lib, e2e_pipe_param,
+			num_pipes, pipe_idx) * refclk_freq_in_mhz;               // From VBA
+	dlg_regs->refcyc_per_vm_group_flip = get_refcyc_per_vm_group_flip_in_us(mode_lib, e2e_pipe_param, num_pipes,
+			pipe_idx) * refclk_freq_in_mhz;                 // From VBA
+	dlg_regs->refcyc_per_vm_req_vblank = get_refcyc_per_vm_req_vblank_in_us(mode_lib, e2e_pipe_param, num_pipes,
+			pipe_idx) * refclk_freq_in_mhz * dml_pow(2, 10);                 // From VBA
+	dlg_regs->refcyc_per_vm_req_flip = get_refcyc_per_vm_req_flip_in_us(mode_lib, e2e_pipe_param, num_pipes,
+			pipe_idx) * refclk_freq_in_mhz * dml_pow(2, 10);  // From VBA
+
+	// From VBA
+	dst_y_per_pte_row_nom_l = get_dst_y_per_pte_row_nom_l(mode_lib, e2e_pipe_param, num_pipes, pipe_idx);
+	// From VBA
+	dst_y_per_pte_row_nom_c = get_dst_y_per_pte_row_nom_c(mode_lib, e2e_pipe_param, num_pipes, pipe_idx);
+	// From VBA
+	dst_y_per_meta_row_nom_l = get_dst_y_per_meta_row_nom_l(mode_lib, e2e_pipe_param, num_pipes, pipe_idx);
+	// From VBA
+	dst_y_per_meta_row_nom_c = get_dst_y_per_meta_row_nom_c(mode_lib, e2e_pipe_param, num_pipes, pipe_idx);
+
+	refcyc_per_pte_group_nom_l = get_refcyc_per_pte_group_nom_l_in_us(mode_lib, e2e_pipe_param, num_pipes,
+			pipe_idx) * refclk_freq_in_mhz;         // From VBA
+	refcyc_per_pte_group_nom_c = get_refcyc_per_pte_group_nom_c_in_us(mode_lib, e2e_pipe_param, num_pipes,
+			pipe_idx) * refclk_freq_in_mhz;         // From VBA
+	refcyc_per_pte_group_vblank_l = get_refcyc_per_pte_group_vblank_l_in_us(mode_lib, e2e_pipe_param,
+			num_pipes, pipe_idx) * refclk_freq_in_mhz;      // From VBA
+	refcyc_per_pte_group_vblank_c = get_refcyc_per_pte_group_vblank_c_in_us(mode_lib, e2e_pipe_param,
+			num_pipes, pipe_idx) * refclk_freq_in_mhz;      // From VBA
+	refcyc_per_pte_group_flip_l = get_refcyc_per_pte_group_flip_l_in_us(mode_lib, e2e_pipe_param, num_pipes,
+			pipe_idx) * refclk_freq_in_mhz;        // From VBA
+	refcyc_per_pte_group_flip_c = get_refcyc_per_pte_group_flip_c_in_us(mode_lib, e2e_pipe_param, num_pipes,
+			pipe_idx) * refclk_freq_in_mhz;        // From VBA
+
+	refcyc_per_meta_chunk_nom_l = get_refcyc_per_meta_chunk_nom_l_in_us(mode_lib, e2e_pipe_param, num_pipes,
+			pipe_idx) * refclk_freq_in_mhz;        // From VBA
+	refcyc_per_meta_chunk_nom_c = get_refcyc_per_meta_chunk_nom_c_in_us(mode_lib, e2e_pipe_param, num_pipes,
+			pipe_idx) * refclk_freq_in_mhz;        // From VBA
+	refcyc_per_meta_chunk_vblank_l = get_refcyc_per_meta_chunk_vblank_l_in_us(mode_lib, e2e_pipe_param,
+			num_pipes, pipe_idx) * refclk_freq_in_mhz;     // From VBA
+	refcyc_per_meta_chunk_vblank_c = get_refcyc_per_meta_chunk_vblank_c_in_us(mode_lib, e2e_pipe_param,
+			num_pipes, pipe_idx) * refclk_freq_in_mhz;     // From VBA
+	refcyc_per_meta_chunk_flip_l = get_refcyc_per_meta_chunk_flip_l_in_us(mode_lib, e2e_pipe_param,
+			num_pipes, pipe_idx) * refclk_freq_in_mhz;       // From VBA
+	refcyc_per_meta_chunk_flip_c = get_refcyc_per_meta_chunk_flip_c_in_us(mode_lib, e2e_pipe_param,
+			num_pipes, pipe_idx) * refclk_freq_in_mhz;       // From VBA
+
+	dlg_regs->dst_y_per_pte_row_nom_l = dst_y_per_pte_row_nom_l * dml_pow(2, 2);
+	dlg_regs->dst_y_per_pte_row_nom_c = dst_y_per_pte_row_nom_c * dml_pow(2, 2);
+	dlg_regs->dst_y_per_meta_row_nom_l = dst_y_per_meta_row_nom_l * dml_pow(2, 2);
+	dlg_regs->dst_y_per_meta_row_nom_c = dst_y_per_meta_row_nom_c * dml_pow(2, 2);
+	dlg_regs->refcyc_per_pte_group_nom_l = refcyc_per_pte_group_nom_l;
+	dlg_regs->refcyc_per_pte_group_nom_c = refcyc_per_pte_group_nom_c;
+	dlg_regs->refcyc_per_pte_group_vblank_l = refcyc_per_pte_group_vblank_l;
+	dlg_regs->refcyc_per_pte_group_vblank_c = refcyc_per_pte_group_vblank_c;
+	dlg_regs->refcyc_per_pte_group_flip_l = refcyc_per_pte_group_flip_l;
+	dlg_regs->refcyc_per_pte_group_flip_c = refcyc_per_pte_group_flip_c;
+	dlg_regs->refcyc_per_meta_chunk_nom_l = refcyc_per_meta_chunk_nom_l;
+	dlg_regs->refcyc_per_meta_chunk_nom_c = refcyc_per_meta_chunk_nom_c;
+	dlg_regs->refcyc_per_meta_chunk_vblank_l = refcyc_per_meta_chunk_vblank_l;
+	dlg_regs->refcyc_per_meta_chunk_vblank_c = refcyc_per_meta_chunk_vblank_c;
+	dlg_regs->refcyc_per_meta_chunk_flip_l = refcyc_per_meta_chunk_flip_l;
+	dlg_regs->refcyc_per_meta_chunk_flip_c = refcyc_per_meta_chunk_flip_c;
+	dlg_regs->refcyc_per_line_delivery_pre_l = (unsigned int) dml_floor(refcyc_per_line_delivery_pre_l, 1);
+	dlg_regs->refcyc_per_line_delivery_l = (unsigned int) dml_floor(refcyc_per_line_delivery_l, 1);
+	dlg_regs->refcyc_per_line_delivery_pre_c = (unsigned int) dml_floor(refcyc_per_line_delivery_pre_c, 1);
+	dlg_regs->refcyc_per_line_delivery_c = (unsigned int) dml_floor(refcyc_per_line_delivery_c, 1);
+
+	dlg_regs->chunk_hdl_adjust_cur0 = 3;
+	dlg_regs->dst_y_offset_cur0 = 0;
+	dlg_regs->chunk_hdl_adjust_cur1 = 3;
+	dlg_regs->dst_y_offset_cur1 = 0;
+
+	dlg_regs->dst_y_delta_drq_limit = 0x7fff; // off
+
+	ttu_regs->refcyc_per_req_delivery_pre_l = (unsigned int) (refcyc_per_req_delivery_pre_l * dml_pow(2, 10));
+	ttu_regs->refcyc_per_req_delivery_l = (unsigned int) (refcyc_per_req_delivery_l * dml_pow(2, 10));
+	ttu_regs->refcyc_per_req_delivery_pre_c = (unsigned int) (refcyc_per_req_delivery_pre_c * dml_pow(2, 10));
+	ttu_regs->refcyc_per_req_delivery_c = (unsigned int) (refcyc_per_req_delivery_c * dml_pow(2, 10));
+	ttu_regs->refcyc_per_req_delivery_pre_cur0 =
+			(unsigned int) (refcyc_per_req_delivery_pre_cur0 * dml_pow(2, 10));
+	ttu_regs->refcyc_per_req_delivery_cur0 = (unsigned int) (refcyc_per_req_delivery_cur0 * dml_pow(2, 10));
+	ttu_regs->refcyc_per_req_delivery_pre_cur1 = 0;
+	ttu_regs->refcyc_per_req_delivery_cur1 = 0;
+	ttu_regs->qos_level_low_wm = 0;
+
+	ttu_regs->qos_level_high_wm = (unsigned int) (4.0 * (double) htotal * ref_freq_to_pix_freq);
+
+	ttu_regs->qos_level_flip = 14;
+	ttu_regs->qos_level_fixed_l = 8;
+	ttu_regs->qos_level_fixed_c = 8;
+	ttu_regs->qos_level_fixed_cur0 = 8;
+	ttu_regs->qos_ramp_disable_l = 0;
+	ttu_regs->qos_ramp_disable_c = 0;
+	ttu_regs->qos_ramp_disable_cur0 = 0;
+	ttu_regs->min_ttu_vblank = min_ttu_vblank * refclk_freq_in_mhz;
+
+	// CHECK for HW registers' range, assert or clamp
+	ASSERT(refcyc_per_req_delivery_pre_l < dml_pow(2, 13));
+	ASSERT(refcyc_per_req_delivery_l < dml_pow(2, 13));
+	ASSERT(refcyc_per_req_delivery_pre_c < dml_pow(2, 13));
+	ASSERT(refcyc_per_req_delivery_c < dml_pow(2, 13));
+	if (dlg_regs->refcyc_per_vm_group_vblank >= (unsigned int) dml_pow(2, 23))
+		dlg_regs->refcyc_per_vm_group_vblank = dml_pow(2, 23) - 1;
+
+	if (dlg_regs->refcyc_per_vm_group_flip >= (unsigned int) dml_pow(2, 23))
+		dlg_regs->refcyc_per_vm_group_flip = dml_pow(2, 23) - 1;
+
+	if (dlg_regs->refcyc_per_vm_req_vblank >= (unsigned int) dml_pow(2, 23))
+		dlg_regs->refcyc_per_vm_req_vblank = dml_pow(2, 23) - 1;
+
+	if (dlg_regs->refcyc_per_vm_req_flip >= (unsigned int) dml_pow(2, 23))
+		dlg_regs->refcyc_per_vm_req_flip = dml_pow(2, 23) - 1;
+
+	ASSERT(dlg_regs->dst_y_after_scaler < (unsigned int) 8);
+	ASSERT(dlg_regs->refcyc_x_after_scaler < (unsigned int)dml_pow(2, 13));
+	ASSERT(dlg_regs->dst_y_per_pte_row_nom_l < (unsigned int)dml_pow(2, 17));
+	if (dual_plane) {
+		if (dlg_regs->dst_y_per_pte_row_nom_c >= (unsigned int) dml_pow(2, 17)) {
+			// FIXME what so special about chroma, can we just assert?
+			dml_print("DML_DLG: %s: Warning dst_y_per_pte_row_nom_c %u > register max U15.2 %u\n",
+					__func__, dlg_regs->dst_y_per_pte_row_nom_c, (unsigned int)dml_pow(2, 17) - 1);
+		}
+	}
+	ASSERT(dlg_regs->dst_y_per_meta_row_nom_l < (unsigned int)dml_pow(2, 17));
+	ASSERT(dlg_regs->dst_y_per_meta_row_nom_c < (unsigned int)dml_pow(2, 17));
+
+	if (dlg_regs->refcyc_per_pte_group_nom_l >= (unsigned int) dml_pow(2, 23))
+		dlg_regs->refcyc_per_pte_group_nom_l = dml_pow(2, 23) - 1;
+	if (dual_plane) {
+		if (dlg_regs->refcyc_per_pte_group_nom_c >= (unsigned int) dml_pow(2, 23))
+			dlg_regs->refcyc_per_pte_group_nom_c = dml_pow(2, 23) - 1;
+	}
+	ASSERT(dlg_regs->refcyc_per_pte_group_vblank_l < (unsigned int)dml_pow(2, 13));
+	if (dual_plane) {
+		ASSERT(dlg_regs->refcyc_per_pte_group_vblank_c < (unsigned int)dml_pow(2, 13));
+	}
+
+	if (dlg_regs->refcyc_per_meta_chunk_nom_l >= (unsigned int) dml_pow(2, 23))
+		dlg_regs->refcyc_per_meta_chunk_nom_l = dml_pow(2, 23) - 1;
+	if (dual_plane) {
+		if (dlg_regs->refcyc_per_meta_chunk_nom_c >= (unsigned int) dml_pow(2, 23))
+			dlg_regs->refcyc_per_meta_chunk_nom_c = dml_pow(2, 23) - 1;
+	}
+	ASSERT(dlg_regs->refcyc_per_meta_chunk_vblank_l < (unsigned int)dml_pow(2, 13));
+	ASSERT(dlg_regs->refcyc_per_meta_chunk_vblank_c < (unsigned int)dml_pow(2, 13));
+	ASSERT(dlg_regs->refcyc_per_line_delivery_pre_l < (unsigned int)dml_pow(2, 13));
+	ASSERT(dlg_regs->refcyc_per_line_delivery_l < (unsigned int)dml_pow(2, 13));
+	ASSERT(dlg_regs->refcyc_per_line_delivery_pre_c < (unsigned int)dml_pow(2, 13));
+	ASSERT(dlg_regs->refcyc_per_line_delivery_c < (unsigned int)dml_pow(2, 13));
+	ASSERT(ttu_regs->qos_level_low_wm < dml_pow(2, 14));
+	ASSERT(ttu_regs->qos_level_high_wm < dml_pow(2, 14));
+	ASSERT(ttu_regs->min_ttu_vblank < dml_pow(2, 24));
+
+	print__ttu_regs_st(mode_lib, ttu_regs);
+	print__dlg_regs_st(mode_lib, dlg_regs);
+	dml_print("DML_DLG::%s: Calculation for pipe[%d] done, num_pipes=%d\n", __func__, pipe_idx, num_pipes);
+}
+
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_rq_dlg_calc_32.h b/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_rq_dlg_calc_32.h
new file mode 100644
index 000000000000..ebee365293cd
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_rq_dlg_calc_32.h
@@ -0,0 +1,70 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __DML32_DISPLAY_RQ_DLG_CALC_H__
+#define __DML32_DISPLAY_RQ_DLG_CALC_H__
+
+#include "../display_rq_dlg_helpers.h"
+
+struct display_mode_lib;
+
+/*
+* Function: dml_rq_dlg_get_rq_reg
+*  Main entry point for test to get the register values out of this DML class.
+*  This function calls <get_rq_param> and <extract_rq_regs> functions to calculate
+*  and then populate the rq_regs struct
+* Input:
+*  pipe_param - pipe source configuration (e.g. vp, pitch, scaling, dest, etc.)
+* Output:
+*  rq_regs - struct that holds all the RQ registers field value.
+*            See also: <display_rq_regs_st>
+*/
+void dml32_rq_dlg_get_rq_reg(display_rq_regs_st *rq_regs,
+		struct display_mode_lib *mode_lib,
+		const display_e2e_pipe_params_st *e2e_pipe_param,
+		const unsigned int num_pipes,
+		const unsigned int pipe_idx);
+
+/*
+* Function: dml_rq_dlg_get_dlg_reg
+*   Calculate and return DLG and TTU register struct given the system setting
+* Output:
+*  dlg_regs - output DLG register struct
+*  ttu_regs - output DLG TTU register struct
+* Input:
+*  e2e_pipe_param - "compacted" array of e2e pipe param struct
+*  num_pipes - num of active "pipe" or "route"
+*  pipe_idx - index that identifies the e2e_pipe_param that corresponding to this dlg
+*  cstate - 0: when calculate min_ttu_vblank it is assumed cstate is not required. 1: Normal mode, cstate is considered.
+*           Added for legacy or unrealistic timing tests.
+*/
+void dml32_rq_dlg_get_dlg_reg(struct display_mode_lib *mode_lib,
+		display_dlg_regs_st *dlg_regs,
+		display_ttu_regs_st *ttu_regs,
+		display_e2e_pipe_params_st *e2e_pipe_param,
+		const unsigned int num_pipes,
+		const unsigned int pipe_idx);
+
+#endif
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn321/dcn321_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn321/dcn321_fpu.c
new file mode 100644
index 000000000000..e0a1dc89ce43
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn321/dcn321_fpu.c
@@ -0,0 +1,931 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#include "clk_mgr.h"
+#include "resource.h"
+#include "dcn321_fpu.h"
+#include "dcn32/dcn32_resource.h"
+#include "dcn321/dcn321_resource.h"
+#include "dml/dcn32/display_mode_vba_util_32.h"
+
+#define DCN3_2_DEFAULT_DET_SIZE 256
+
+struct _vcs_dpi_ip_params_st dcn3_21_ip = {
+	.gpuvm_enable = 0,
+	.gpuvm_max_page_table_levels = 4,
+	.hostvm_enable = 0,
+	.rob_buffer_size_kbytes = 128,
+	.det_buffer_size_kbytes = DCN3_2_DEFAULT_DET_SIZE,
+	.config_return_buffer_size_in_kbytes = 1280,
+	.compressed_buffer_segment_size_in_kbytes = 64,
+	.meta_fifo_size_in_kentries = 22,
+	.zero_size_buffer_entries = 512,
+	.compbuf_reserved_space_64b = 256,
+	.compbuf_reserved_space_zs = 64,
+	.dpp_output_buffer_pixels = 2560,
+	.opp_output_buffer_lines = 1,
+	.pixel_chunk_size_kbytes = 8,
+	.alpha_pixel_chunk_size_kbytes = 4,
+	.min_pixel_chunk_size_bytes = 1024,
+	.dcc_meta_buffer_size_bytes = 6272,
+	.meta_chunk_size_kbytes = 2,
+	.min_meta_chunk_size_bytes = 256,
+	.writeback_chunk_size_kbytes = 8,
+	.ptoi_supported = false,
+	.num_dsc = 4,
+	.maximum_dsc_bits_per_component = 12,
+	.maximum_pixels_per_line_per_dsc_unit = 6016,
+	.dsc422_native_support = true,
+	.is_line_buffer_bpp_fixed = true,
+	.line_buffer_fixed_bpp = 57,
+	.line_buffer_size_bits = 1171920,
+	.max_line_buffer_lines = 32,
+	.writeback_interface_buffer_size_kbytes = 90,
+	.max_num_dpp = 4,
+	.max_num_otg = 4,
+	.max_num_hdmi_frl_outputs = 1,
+	.max_num_wb = 1,
+	.max_dchub_pscl_bw_pix_per_clk = 4,
+	.max_pscl_lb_bw_pix_per_clk = 2,
+	.max_lb_vscl_bw_pix_per_clk = 4,
+	.max_vscl_hscl_bw_pix_per_clk = 4,
+	.max_hscl_ratio = 6,
+	.max_vscl_ratio = 6,
+	.max_hscl_taps = 8,
+	.max_vscl_taps = 8,
+	.dpte_buffer_size_in_pte_reqs_luma = 64,
+	.dpte_buffer_size_in_pte_reqs_chroma = 34,
+	.dispclk_ramp_margin_percent = 1,
+	.max_inter_dcn_tile_repeaters = 8,
+	.cursor_buffer_size = 16,
+	.cursor_chunk_size = 2,
+	.writeback_line_buffer_buffer_size = 0,
+	.writeback_min_hscl_ratio = 1,
+	.writeback_min_vscl_ratio = 1,
+	.writeback_max_hscl_ratio = 1,
+	.writeback_max_vscl_ratio = 1,
+	.writeback_max_hscl_taps = 1,
+	.writeback_max_vscl_taps = 1,
+	.dppclk_delay_subtotal = 47,
+	.dppclk_delay_scl = 50,
+	.dppclk_delay_scl_lb_only = 16,
+	.dppclk_delay_cnvc_formatter = 28,
+	.dppclk_delay_cnvc_cursor = 6,
+	.dispclk_delay_subtotal = 125,
+	.dynamic_metadata_vm_enabled = false,
+	.odm_combine_4to1_supported = false,
+	.dcc_supported = true,
+	.max_num_dp2p0_outputs = 2,
+	.max_num_dp2p0_streams = 4,
+};
+
+struct _vcs_dpi_soc_bounding_box_st dcn3_21_soc = {
+	.clock_limits = {
+		{
+			.state = 0,
+			.dcfclk_mhz = 1434.0,
+			.fabricclk_mhz = 2250.0,
+			.dispclk_mhz = 1720.0,
+			.dppclk_mhz = 1720.0,
+			.phyclk_mhz = 810.0,
+			.phyclk_d18_mhz = 667.0,
+			.phyclk_d32_mhz = 313.0,
+			.socclk_mhz = 1200.0,
+			.dscclk_mhz = 573.333,
+			.dram_speed_mts = 16000.0,
+			.dtbclk_mhz = 1564.0,
+		},
+	},
+	.num_states = 1,
+	.sr_exit_time_us = 19.95,
+	.sr_enter_plus_exit_time_us = 24.36,
+	.sr_exit_z8_time_us = 285.0,
+	.sr_enter_plus_exit_z8_time_us = 320,
+	.writeback_latency_us = 12.0,
+	.round_trip_ping_latency_dcfclk_cycles = 207,
+	.urgent_latency_pixel_data_only_us = 4,
+	.urgent_latency_pixel_mixed_with_vm_data_us = 4,
+	.urgent_latency_vm_data_only_us = 4,
+	.fclk_change_latency_us = 7,
+	.usr_retraining_latency_us = 0,
+	.smn_latency_us = 0,
+	.mall_allocated_for_dcn_mbytes = 32,
+	.urgent_out_of_order_return_per_channel_pixel_only_bytes = 4096,
+	.urgent_out_of_order_return_per_channel_pixel_and_vm_bytes = 4096,
+	.urgent_out_of_order_return_per_channel_vm_only_bytes = 4096,
+	.pct_ideal_sdp_bw_after_urgent = 90.0,
+	.pct_ideal_fabric_bw_after_urgent = 67.0,
+	.pct_ideal_dram_sdp_bw_after_urgent_pixel_only = 20.0,
+	.pct_ideal_dram_sdp_bw_after_urgent_pixel_and_vm = 60.0,
+	.pct_ideal_dram_sdp_bw_after_urgent_vm_only = 30.0,
+	.pct_ideal_dram_bw_after_urgent_strobe = 67.0,
+	.max_avg_sdp_bw_use_normal_percent = 80.0,
+	.max_avg_fabric_bw_use_normal_percent = 60.0,
+	.max_avg_dram_bw_use_normal_strobe_percent = 50.0,
+	.max_avg_dram_bw_use_normal_percent = 15.0,
+	.num_chans = 8,
+	.dram_channel_width_bytes = 2,
+	.fabric_datapath_to_dcn_data_return_bytes = 64,
+	.return_bus_width_bytes = 64,
+	.downspread_percent = 0.38,
+	.dcn_downspread_percent = 0.5,
+	.dram_clock_change_latency_us = 400,
+	.dispclk_dppclk_vco_speed_mhz = 4300.0,
+	.do_urgent_latency_adjustment = true,
+	.urgent_latency_adjustment_fabric_clock_component_us = 1.0,
+	.urgent_latency_adjustment_fabric_clock_reference_mhz = 3000,
+};
+
+static void get_optimal_ntuple(struct _vcs_dpi_voltage_scaling_st *entry)
+{
+	if (entry->dcfclk_mhz > 0) {
+		float bw_on_sdp = entry->dcfclk_mhz * dcn3_21_soc.return_bus_width_bytes * ((float)dcn3_21_soc.pct_ideal_sdp_bw_after_urgent / 100);
+
+		entry->fabricclk_mhz = bw_on_sdp / (dcn3_21_soc.return_bus_width_bytes * ((float)dcn3_21_soc.pct_ideal_fabric_bw_after_urgent / 100));
+		entry->dram_speed_mts = bw_on_sdp / (dcn3_21_soc.num_chans *
+				dcn3_21_soc.dram_channel_width_bytes * ((float)dcn3_21_soc.pct_ideal_dram_sdp_bw_after_urgent_pixel_only / 100));
+	} else if (entry->fabricclk_mhz > 0) {
+		float bw_on_fabric = entry->fabricclk_mhz * dcn3_21_soc.return_bus_width_bytes * ((float)dcn3_21_soc.pct_ideal_fabric_bw_after_urgent / 100);
+
+		entry->dcfclk_mhz = bw_on_fabric / (dcn3_21_soc.return_bus_width_bytes * ((float)dcn3_21_soc.pct_ideal_sdp_bw_after_urgent / 100));
+		entry->dram_speed_mts = bw_on_fabric / (dcn3_21_soc.num_chans *
+				dcn3_21_soc.dram_channel_width_bytes * ((float)dcn3_21_soc.pct_ideal_dram_sdp_bw_after_urgent_pixel_only / 100));
+	} else if (entry->dram_speed_mts > 0) {
+		float bw_on_dram = entry->dram_speed_mts * dcn3_21_soc.num_chans *
+				dcn3_21_soc.dram_channel_width_bytes * ((float)dcn3_21_soc.pct_ideal_dram_sdp_bw_after_urgent_pixel_only / 100);
+
+		entry->fabricclk_mhz = bw_on_dram / (dcn3_21_soc.return_bus_width_bytes * ((float)dcn3_21_soc.pct_ideal_fabric_bw_after_urgent / 100));
+		entry->dcfclk_mhz = bw_on_dram / (dcn3_21_soc.return_bus_width_bytes * ((float)dcn3_21_soc.pct_ideal_sdp_bw_after_urgent / 100));
+	}
+}
+
+static float calculate_net_bw_in_kbytes_sec(struct _vcs_dpi_voltage_scaling_st *entry)
+{
+	float memory_bw_kbytes_sec;
+	float fabric_bw_kbytes_sec;
+	float sdp_bw_kbytes_sec;
+	float limiting_bw_kbytes_sec;
+
+	memory_bw_kbytes_sec = entry->dram_speed_mts * dcn3_21_soc.num_chans *
+			dcn3_21_soc.dram_channel_width_bytes * ((float)dcn3_21_soc.pct_ideal_dram_sdp_bw_after_urgent_pixel_only / 100);
+
+	fabric_bw_kbytes_sec = entry->fabricclk_mhz * dcn3_21_soc.return_bus_width_bytes * ((float)dcn3_21_soc.pct_ideal_fabric_bw_after_urgent / 100);
+
+	sdp_bw_kbytes_sec = entry->dcfclk_mhz * dcn3_21_soc.return_bus_width_bytes * ((float)dcn3_21_soc.pct_ideal_sdp_bw_after_urgent / 100);
+
+	limiting_bw_kbytes_sec = memory_bw_kbytes_sec;
+
+	if (fabric_bw_kbytes_sec < limiting_bw_kbytes_sec)
+		limiting_bw_kbytes_sec = fabric_bw_kbytes_sec;
+
+	if (sdp_bw_kbytes_sec < limiting_bw_kbytes_sec)
+		limiting_bw_kbytes_sec = sdp_bw_kbytes_sec;
+
+	return limiting_bw_kbytes_sec;
+}
+
+static void dcn321_insert_entry_into_table_sorted(struct _vcs_dpi_voltage_scaling_st *table,
+					   unsigned int *num_entries,
+					   struct _vcs_dpi_voltage_scaling_st *entry)
+{
+	int i = 0;
+	int index = 0;
+
+	dc_assert_fp_enabled();
+
+	if (*num_entries == 0) {
+		table[0] = *entry;
+		(*num_entries)++;
+	} else {
+		while (entry->net_bw_in_kbytes_sec > table[index].net_bw_in_kbytes_sec) {
+			index++;
+			if (index >= *num_entries)
+				break;
+		}
+
+		for (i = *num_entries; i > index; i--)
+			table[i] = table[i - 1];
+
+		table[index] = *entry;
+		(*num_entries)++;
+	}
+}
+
+static void remove_entry_from_table_at_index(struct _vcs_dpi_voltage_scaling_st *table, unsigned int *num_entries,
+		unsigned int index)
+{
+	int i;
+
+	if (*num_entries == 0)
+		return;
+
+	for (i = index; i < *num_entries - 1; i++) {
+		table[i] = table[i + 1];
+	}
+	memset(&table[--(*num_entries)], 0, sizeof(struct _vcs_dpi_voltage_scaling_st));
+}
+
+static void swap_table_entries(struct _vcs_dpi_voltage_scaling_st *first_entry,
+		struct _vcs_dpi_voltage_scaling_st *second_entry)
+{
+	struct _vcs_dpi_voltage_scaling_st temp_entry = *first_entry;
+	*first_entry = *second_entry;
+	*second_entry = temp_entry;
+}
+
+/*
+ * sort_entries_with_same_bw - Sort entries sharing the same bandwidth by DCFCLK
+ */
+static void sort_entries_with_same_bw(struct _vcs_dpi_voltage_scaling_st *table, unsigned int *num_entries)
+{
+	unsigned int start_index = 0;
+	unsigned int end_index = 0;
+	unsigned int current_bw = 0;
+
+	for (int i = 0; i < (*num_entries - 1); i++) {
+		if (table[i].net_bw_in_kbytes_sec == table[i+1].net_bw_in_kbytes_sec) {
+			current_bw = table[i].net_bw_in_kbytes_sec;
+			start_index = i;
+			end_index = ++i;
+
+			while ((i < (*num_entries - 1)) && (table[i+1].net_bw_in_kbytes_sec == current_bw))
+				end_index = ++i;
+		}
+
+		if (start_index != end_index) {
+			for (int j = start_index; j < end_index; j++) {
+				for (int k = start_index; k < end_index; k++) {
+					if (table[k].dcfclk_mhz > table[k+1].dcfclk_mhz)
+						swap_table_entries(&table[k], &table[k+1]);
+				}
+			}
+		}
+
+		start_index = 0;
+		end_index = 0;
+
+	}
+}
+
+/*
+ * remove_inconsistent_entries - Ensure entries with the same bandwidth have MEMCLK and FCLK monotonically increasing
+ *                               and remove entries that do not follow this order
+ */
+static void remove_inconsistent_entries(struct _vcs_dpi_voltage_scaling_st *table, unsigned int *num_entries)
+{
+	for (int i = 0; i < (*num_entries - 1); i++) {
+		if (table[i].net_bw_in_kbytes_sec == table[i+1].net_bw_in_kbytes_sec) {
+			if ((table[i].dram_speed_mts > table[i+1].dram_speed_mts) ||
+				(table[i].fabricclk_mhz > table[i+1].fabricclk_mhz))
+				remove_entry_from_table_at_index(table, num_entries, i);
+		}
+	}
+}
+
+/*
+ * override_max_clk_values - Overwrite the max clock frequencies with the max DC mode timings
+ * Input:
+ *	max_clk_limit - struct containing the desired clock timings
+ * Output:
+ *	curr_clk_limit  - struct containing the timings that need to be overwritten
+ * Return: 0 upon success, non-zero for failure
+ */
+static int override_max_clk_values(struct clk_limit_table_entry *max_clk_limit,
+		struct clk_limit_table_entry *curr_clk_limit)
+{
+	if (NULL == max_clk_limit || NULL == curr_clk_limit)
+		return -1; //invalid parameters
+
+	//only overwrite if desired max clock frequency is initialized
+	if (max_clk_limit->dcfclk_mhz != 0)
+		curr_clk_limit->dcfclk_mhz = max_clk_limit->dcfclk_mhz;
+
+	if (max_clk_limit->fclk_mhz != 0)
+		curr_clk_limit->fclk_mhz = max_clk_limit->fclk_mhz;
+
+	if (max_clk_limit->memclk_mhz != 0)
+		curr_clk_limit->memclk_mhz = max_clk_limit->memclk_mhz;
+
+	if (max_clk_limit->socclk_mhz != 0)
+		curr_clk_limit->socclk_mhz = max_clk_limit->socclk_mhz;
+
+	if (max_clk_limit->dtbclk_mhz != 0)
+		curr_clk_limit->dtbclk_mhz = max_clk_limit->dtbclk_mhz;
+
+	if (max_clk_limit->dispclk_mhz != 0)
+		curr_clk_limit->dispclk_mhz = max_clk_limit->dispclk_mhz;
+
+	return 0;
+}
+
+static int build_synthetic_soc_states(bool disable_dc_mode_overwrite, struct clk_bw_params *bw_params,
+		struct _vcs_dpi_voltage_scaling_st *table, unsigned int *num_entries)
+{
+	int i, j;
+	struct _vcs_dpi_voltage_scaling_st entry = {0};
+	struct clk_limit_table_entry max_clk_data = {0};
+
+	unsigned int min_dcfclk_mhz = 199, min_fclk_mhz = 299;
+
+	static const unsigned int num_dcfclk_stas = 5;
+	unsigned int dcfclk_sta_targets[DC__VOLTAGE_STATES] = {199, 615, 906, 1324, 1564};
+
+	unsigned int num_uclk_dpms = 0;
+	unsigned int num_fclk_dpms = 0;
+	unsigned int num_dcfclk_dpms = 0;
+
+	unsigned int num_dc_uclk_dpms = 0;
+	unsigned int num_dc_fclk_dpms = 0;
+	unsigned int num_dc_dcfclk_dpms = 0;
+
+	for (i = 0; i < MAX_NUM_DPM_LVL; i++) {
+		if (bw_params->clk_table.entries[i].dcfclk_mhz > max_clk_data.dcfclk_mhz)
+			max_clk_data.dcfclk_mhz = bw_params->clk_table.entries[i].dcfclk_mhz;
+		if (bw_params->clk_table.entries[i].fclk_mhz > max_clk_data.fclk_mhz)
+			max_clk_data.fclk_mhz = bw_params->clk_table.entries[i].fclk_mhz;
+		if (bw_params->clk_table.entries[i].memclk_mhz > max_clk_data.memclk_mhz)
+			max_clk_data.memclk_mhz = bw_params->clk_table.entries[i].memclk_mhz;
+		if (bw_params->clk_table.entries[i].dispclk_mhz > max_clk_data.dispclk_mhz)
+			max_clk_data.dispclk_mhz = bw_params->clk_table.entries[i].dispclk_mhz;
+		if (bw_params->clk_table.entries[i].dppclk_mhz > max_clk_data.dppclk_mhz)
+			max_clk_data.dppclk_mhz = bw_params->clk_table.entries[i].dppclk_mhz;
+		if (bw_params->clk_table.entries[i].phyclk_mhz > max_clk_data.phyclk_mhz)
+			max_clk_data.phyclk_mhz = bw_params->clk_table.entries[i].phyclk_mhz;
+		if (bw_params->clk_table.entries[i].dtbclk_mhz > max_clk_data.dtbclk_mhz)
+			max_clk_data.dtbclk_mhz = bw_params->clk_table.entries[i].dtbclk_mhz;
+
+		if (bw_params->clk_table.entries[i].memclk_mhz > 0) {
+			num_uclk_dpms++;
+			if (bw_params->clk_table.entries[i].memclk_mhz <= bw_params->dc_mode_limit.memclk_mhz)
+				num_dc_uclk_dpms++;
+		}
+		if (bw_params->clk_table.entries[i].fclk_mhz > 0) {
+			num_fclk_dpms++;
+			if (bw_params->clk_table.entries[i].fclk_mhz <= bw_params->dc_mode_limit.fclk_mhz)
+				num_dc_fclk_dpms++;
+		}
+		if (bw_params->clk_table.entries[i].dcfclk_mhz > 0) {
+			num_dcfclk_dpms++;
+			if (bw_params->clk_table.entries[i].dcfclk_mhz <= bw_params->dc_mode_limit.dcfclk_mhz)
+				num_dc_dcfclk_dpms++;
+		}
+	}
+
+	if (!disable_dc_mode_overwrite) {
+		//Overwrite max frequencies with max DC mode frequencies for DC mode systems
+		override_max_clk_values(&bw_params->dc_mode_limit, &max_clk_data);
+		num_uclk_dpms = num_dc_uclk_dpms;
+		num_fclk_dpms = num_dc_fclk_dpms;
+		num_dcfclk_dpms = num_dc_dcfclk_dpms;
+		bw_params->clk_table.num_entries_per_clk.num_memclk_levels = num_uclk_dpms;
+		bw_params->clk_table.num_entries_per_clk.num_fclk_levels = num_fclk_dpms;
+	}
+
+	if (num_dcfclk_dpms > 0 && bw_params->clk_table.entries[0].fclk_mhz > min_fclk_mhz)
+		min_fclk_mhz = bw_params->clk_table.entries[0].fclk_mhz;
+
+	if (!max_clk_data.dcfclk_mhz || !max_clk_data.dispclk_mhz || !max_clk_data.dtbclk_mhz)
+		return -1;
+
+	if (max_clk_data.dppclk_mhz == 0)
+		max_clk_data.dppclk_mhz = max_clk_data.dispclk_mhz;
+
+	if (max_clk_data.fclk_mhz == 0)
+		max_clk_data.fclk_mhz = max_clk_data.dcfclk_mhz *
+				dcn3_21_soc.pct_ideal_sdp_bw_after_urgent /
+				dcn3_21_soc.pct_ideal_fabric_bw_after_urgent;
+
+	if (max_clk_data.phyclk_mhz == 0)
+		max_clk_data.phyclk_mhz = dcn3_21_soc.clock_limits[0].phyclk_mhz;
+
+	*num_entries = 0;
+	entry.dispclk_mhz = max_clk_data.dispclk_mhz;
+	entry.dscclk_mhz = max_clk_data.dispclk_mhz / 3;
+	entry.dppclk_mhz = max_clk_data.dppclk_mhz;
+	entry.dtbclk_mhz = max_clk_data.dtbclk_mhz;
+	entry.phyclk_mhz = max_clk_data.phyclk_mhz;
+	entry.phyclk_d18_mhz = dcn3_21_soc.clock_limits[0].phyclk_d18_mhz;
+	entry.phyclk_d32_mhz = dcn3_21_soc.clock_limits[0].phyclk_d32_mhz;
+
+	// Insert all the DCFCLK STAs
+	for (i = 0; i < num_dcfclk_stas; i++) {
+		entry.dcfclk_mhz = dcfclk_sta_targets[i];
+		entry.fabricclk_mhz = 0;
+		entry.dram_speed_mts = 0;
+
+		get_optimal_ntuple(&entry);
+		entry.net_bw_in_kbytes_sec = calculate_net_bw_in_kbytes_sec(&entry);
+		dcn321_insert_entry_into_table_sorted(table, num_entries, &entry);
+	}
+
+	// Insert the max DCFCLK
+	entry.dcfclk_mhz = max_clk_data.dcfclk_mhz;
+	entry.fabricclk_mhz = 0;
+	entry.dram_speed_mts = 0;
+
+	get_optimal_ntuple(&entry);
+	entry.net_bw_in_kbytes_sec = calculate_net_bw_in_kbytes_sec(&entry);
+	dcn321_insert_entry_into_table_sorted(table, num_entries, &entry);
+
+	// Insert the UCLK DPMS
+	for (i = 0; i < num_uclk_dpms; i++) {
+		entry.dcfclk_mhz = 0;
+		entry.fabricclk_mhz = 0;
+		entry.dram_speed_mts = bw_params->clk_table.entries[i].memclk_mhz * 16;
+
+		get_optimal_ntuple(&entry);
+		entry.net_bw_in_kbytes_sec = calculate_net_bw_in_kbytes_sec(&entry);
+		dcn321_insert_entry_into_table_sorted(table, num_entries, &entry);
+	}
+
+	// If FCLK is coarse grained, insert individual DPMs.
+	if (num_fclk_dpms > 2) {
+		for (i = 0; i < num_fclk_dpms; i++) {
+			entry.dcfclk_mhz = 0;
+			entry.fabricclk_mhz = bw_params->clk_table.entries[i].fclk_mhz;
+			entry.dram_speed_mts = 0;
+
+			get_optimal_ntuple(&entry);
+			entry.net_bw_in_kbytes_sec = calculate_net_bw_in_kbytes_sec(&entry);
+			dcn321_insert_entry_into_table_sorted(table, num_entries, &entry);
+		}
+	}
+	// If FCLK fine grained, only insert max
+	else {
+		entry.dcfclk_mhz = 0;
+		entry.fabricclk_mhz = max_clk_data.fclk_mhz;
+		entry.dram_speed_mts = 0;
+
+		get_optimal_ntuple(&entry);
+		entry.net_bw_in_kbytes_sec = calculate_net_bw_in_kbytes_sec(&entry);
+		dcn321_insert_entry_into_table_sorted(table, num_entries, &entry);
+	}
+
+	// At this point, the table contains all "points of interest" based on
+	// DPMs from PMFW, and STAs.  Table is sorted by BW, and all clock
+	// ratios (by derate, are exact).
+
+	// Remove states that require higher clocks than are supported
+	for (i = *num_entries - 1; i >= 0 ; i--) {
+		if (table[i].dcfclk_mhz > max_clk_data.dcfclk_mhz ||
+				table[i].fabricclk_mhz > max_clk_data.fclk_mhz ||
+				table[i].dram_speed_mts > max_clk_data.memclk_mhz * 16)
+			remove_entry_from_table_at_index(table, num_entries, i);
+	}
+
+	// Insert entry with all max dc limits without bandwitch matching
+	if (!disable_dc_mode_overwrite) {
+		struct _vcs_dpi_voltage_scaling_st max_dc_limits_entry = entry;
+
+		max_dc_limits_entry.dcfclk_mhz = max_clk_data.dcfclk_mhz;
+		max_dc_limits_entry.fabricclk_mhz = max_clk_data.fclk_mhz;
+		max_dc_limits_entry.dram_speed_mts = max_clk_data.memclk_mhz * 16;
+
+		max_dc_limits_entry.net_bw_in_kbytes_sec = calculate_net_bw_in_kbytes_sec(&max_dc_limits_entry);
+		dcn321_insert_entry_into_table_sorted(table, num_entries, &max_dc_limits_entry);
+
+		sort_entries_with_same_bw(table, num_entries);
+		remove_inconsistent_entries(table, num_entries);
+	}
+
+
+
+	// At this point, the table only contains supported points of interest
+	// it could be used as is, but some states may be redundant due to
+	// coarse grained nature of some clocks, so we want to round up to
+	// coarse grained DPMs and remove duplicates.
+
+	// Round up UCLKs
+	for (i = *num_entries - 1; i >= 0 ; i--) {
+		for (j = 0; j < num_uclk_dpms; j++) {
+			if (bw_params->clk_table.entries[j].memclk_mhz * 16 >= table[i].dram_speed_mts) {
+				table[i].dram_speed_mts = bw_params->clk_table.entries[j].memclk_mhz * 16;
+				break;
+			}
+		}
+	}
+
+	// If FCLK is coarse grained, round up to next DPMs
+	if (num_fclk_dpms > 2) {
+		for (i = *num_entries - 1; i >= 0 ; i--) {
+			for (j = 0; j < num_fclk_dpms; j++) {
+				if (bw_params->clk_table.entries[j].fclk_mhz >= table[i].fabricclk_mhz) {
+					table[i].fabricclk_mhz = bw_params->clk_table.entries[j].fclk_mhz;
+					break;
+				}
+			}
+		}
+	}
+	// Otherwise, round up to minimum.
+	else {
+		for (i = *num_entries - 1; i >= 0 ; i--) {
+			if (table[i].fabricclk_mhz < min_fclk_mhz) {
+				table[i].fabricclk_mhz = min_fclk_mhz;
+			}
+		}
+	}
+
+	// Round DCFCLKs up to minimum
+	for (i = *num_entries - 1; i >= 0 ; i--) {
+		if (table[i].dcfclk_mhz < min_dcfclk_mhz) {
+			table[i].dcfclk_mhz = min_dcfclk_mhz;
+		}
+	}
+
+	// Remove duplicate states, note duplicate states are always neighbouring since table is sorted.
+	i = 0;
+	while (i < *num_entries - 1) {
+		if (table[i].dcfclk_mhz == table[i + 1].dcfclk_mhz &&
+				table[i].fabricclk_mhz == table[i + 1].fabricclk_mhz &&
+				table[i].dram_speed_mts == table[i + 1].dram_speed_mts)
+			remove_entry_from_table_at_index(table, num_entries, i + 1);
+		else
+			i++;
+	}
+
+	// Fix up the state indicies
+	for (i = *num_entries - 1; i >= 0 ; i--) {
+		table[i].state = i;
+	}
+
+	return 0;
+}
+
+static void dcn321_get_optimal_dcfclk_fclk_for_uclk(unsigned int uclk_mts,
+		unsigned int *optimal_dcfclk,
+		unsigned int *optimal_fclk)
+{
+	double bw_from_dram, bw_from_dram1, bw_from_dram2;
+
+	bw_from_dram1 = uclk_mts * dcn3_21_soc.num_chans *
+		dcn3_21_soc.dram_channel_width_bytes * (dcn3_21_soc.max_avg_dram_bw_use_normal_percent / 100);
+	bw_from_dram2 = uclk_mts * dcn3_21_soc.num_chans *
+		dcn3_21_soc.dram_channel_width_bytes * (dcn3_21_soc.max_avg_sdp_bw_use_normal_percent / 100);
+
+	bw_from_dram = (bw_from_dram1 < bw_from_dram2) ? bw_from_dram1 : bw_from_dram2;
+
+	if (optimal_fclk)
+		*optimal_fclk = bw_from_dram /
+		(dcn3_21_soc.fabric_datapath_to_dcn_data_return_bytes * (dcn3_21_soc.max_avg_sdp_bw_use_normal_percent / 100));
+
+	if (optimal_dcfclk)
+		*optimal_dcfclk =  bw_from_dram /
+		(dcn3_21_soc.return_bus_width_bytes * (dcn3_21_soc.max_avg_sdp_bw_use_normal_percent / 100));
+}
+
+/** dcn321_update_bw_bounding_box
+ * This would override some dcn3_2 ip_or_soc initial parameters hardcoded from spreadsheet
+ * with actual values as per dGPU SKU:
+ * -with passed few options from dc->config
+ * -with dentist_vco_frequency from Clk Mgr (currently hardcoded, but might need to get it from PM FW)
+ * -with passed latency values (passed in ns units) in dc-> bb override for debugging purposes
+ * -with passed latencies from VBIOS (in 100_ns units) if available for certain dGPU SKU
+ * -with number of DRAM channels from VBIOS (which differ for certain dGPU SKU of the same ASIC)
+ * -clocks levels with passed clk_table entries from Clk Mgr as reported by PM FW for different
+ *  clocks (which might differ for certain dGPU SKU of the same ASIC)
+ */
+void dcn321_update_bw_bounding_box_fpu(struct dc *dc, struct clk_bw_params *bw_params)
+{
+	dc_assert_fp_enabled();
+	/* Overrides from dc->config options */
+	dcn3_21_ip.clamp_min_dcfclk = dc->config.clamp_min_dcfclk;
+
+	/* Override from passed dc->bb_overrides if available*/
+	if ((int)(dcn3_21_soc.sr_exit_time_us * 1000) != dc->bb_overrides.sr_exit_time_ns
+			&& dc->bb_overrides.sr_exit_time_ns) {
+		dc->dml2_options.bbox_overrides.sr_exit_latency_us =
+		dcn3_21_soc.sr_exit_time_us = dc->bb_overrides.sr_exit_time_ns / 1000.0;
+	}
+
+	if ((int)(dcn3_21_soc.sr_enter_plus_exit_time_us * 1000)
+			!= dc->bb_overrides.sr_enter_plus_exit_time_ns
+			&& dc->bb_overrides.sr_enter_plus_exit_time_ns) {
+		dc->dml2_options.bbox_overrides.sr_enter_plus_exit_latency_us =
+		dcn3_21_soc.sr_enter_plus_exit_time_us =
+			dc->bb_overrides.sr_enter_plus_exit_time_ns / 1000.0;
+	}
+
+	if ((int)(dcn3_21_soc.urgent_latency_us * 1000) != dc->bb_overrides.urgent_latency_ns
+		&& dc->bb_overrides.urgent_latency_ns) {
+		dcn3_21_soc.urgent_latency_us = dc->bb_overrides.urgent_latency_ns / 1000.0;
+		dc->dml2_options.bbox_overrides.urgent_latency_us =
+		dcn3_21_soc.urgent_latency_pixel_data_only_us = dc->bb_overrides.urgent_latency_ns / 1000.0;
+	}
+
+	if ((int)(dcn3_21_soc.dram_clock_change_latency_us * 1000)
+			!= dc->bb_overrides.dram_clock_change_latency_ns
+			&& dc->bb_overrides.dram_clock_change_latency_ns) {
+		dc->dml2_options.bbox_overrides.dram_clock_change_latency_us =
+		dcn3_21_soc.dram_clock_change_latency_us =
+			dc->bb_overrides.dram_clock_change_latency_ns / 1000.0;
+	}
+
+	if ((int)(dcn3_21_soc.fclk_change_latency_us * 1000)
+			!= dc->bb_overrides.fclk_clock_change_latency_ns
+			&& dc->bb_overrides.fclk_clock_change_latency_ns) {
+		dc->dml2_options.bbox_overrides.fclk_change_latency_us =
+		dcn3_21_soc.fclk_change_latency_us =
+			dc->bb_overrides.fclk_clock_change_latency_ns / 1000;
+	}
+
+	if ((int)(dcn3_21_soc.dummy_pstate_latency_us * 1000)
+			!= dc->bb_overrides.dummy_clock_change_latency_ns
+			&& dc->bb_overrides.dummy_clock_change_latency_ns) {
+		dcn3_21_soc.dummy_pstate_latency_us =
+			dc->bb_overrides.dummy_clock_change_latency_ns / 1000.0;
+	}
+
+	/* Override from VBIOS if VBIOS bb_info available */
+	if (dc->ctx->dc_bios->funcs->get_soc_bb_info) {
+		struct bp_soc_bb_info bb_info = {0};
+
+		if (dc->ctx->dc_bios->funcs->get_soc_bb_info(dc->ctx->dc_bios, &bb_info) == BP_RESULT_OK) {
+			if (bb_info.dram_clock_change_latency_100ns > 0)
+				dc->dml2_options.bbox_overrides.dram_clock_change_latency_us =
+				dcn3_21_soc.dram_clock_change_latency_us =
+					bb_info.dram_clock_change_latency_100ns * 10;
+
+			if (bb_info.dram_sr_enter_exit_latency_100ns > 0)
+				dc->dml2_options.bbox_overrides.sr_enter_plus_exit_latency_us =
+				dcn3_21_soc.sr_enter_plus_exit_time_us =
+					bb_info.dram_sr_enter_exit_latency_100ns * 10;
+
+			if (bb_info.dram_sr_exit_latency_100ns > 0)
+				dc->dml2_options.bbox_overrides.sr_exit_latency_us =
+				dcn3_21_soc.sr_exit_time_us =
+					bb_info.dram_sr_exit_latency_100ns * 10;
+		}
+	}
+
+	/* Override from VBIOS for num_chan */
+	if (dc->ctx->dc_bios->vram_info.num_chans) {
+		dc->dml2_options.bbox_overrides.dram_num_chan =
+		dcn3_21_soc.num_chans = dc->ctx->dc_bios->vram_info.num_chans;
+		dcn3_21_soc.mall_allocated_for_dcn_mbytes = (double)(dcn32_calc_num_avail_chans_for_mall(dc,
+			dc->ctx->dc_bios->vram_info.num_chans) * dc->caps.mall_size_per_mem_channel);
+	}
+
+	if (dc->ctx->dc_bios->vram_info.dram_channel_width_bytes)
+		dc->dml2_options.bbox_overrides.dram_chanel_width_bytes =
+		dcn3_21_soc.dram_channel_width_bytes = dc->ctx->dc_bios->vram_info.dram_channel_width_bytes;
+
+	/* DML DSC delay factor workaround */
+	dcn3_21_ip.dsc_delay_factor_wa = dc->debug.dsc_delay_factor_wa_x1000 / 1000.0;
+
+	dcn3_21_ip.min_prefetch_in_strobe_us = dc->debug.min_prefetch_in_strobe_ns / 1000.0;
+
+	/* Override dispclk_dppclk_vco_speed_mhz from Clk Mgr */
+	dcn3_21_soc.dispclk_dppclk_vco_speed_mhz = dc->clk_mgr->dentist_vco_freq_khz / 1000.0;
+	dc->dml.soc.dispclk_dppclk_vco_speed_mhz = dc->clk_mgr->dentist_vco_freq_khz / 1000.0;
+	dc->dml2_options.bbox_overrides.disp_pll_vco_speed_mhz = dc->clk_mgr->dentist_vco_freq_khz / 1000.0;
+	dc->dml2_options.bbox_overrides.xtalclk_mhz = dc->ctx->dc_bios->fw_info.pll_info.crystal_frequency / 1000.0;
+	dc->dml2_options.bbox_overrides.dchub_refclk_mhz = dc->res_pool->ref_clocks.dchub_ref_clock_inKhz / 1000.0;
+	dc->dml2_options.bbox_overrides.dprefclk_mhz = dc->clk_mgr->dprefclk_khz / 1000.0;
+
+	/* Overrides Clock levelsfrom CLK Mgr table entries as reported by PM FW */
+	if (dc->debug.use_legacy_soc_bb_mechanism) {
+		unsigned int i = 0, j = 0, num_states = 0;
+
+		unsigned int dcfclk_mhz[DC__VOLTAGE_STATES] = {0};
+		unsigned int dram_speed_mts[DC__VOLTAGE_STATES] = {0};
+		unsigned int optimal_uclk_for_dcfclk_sta_targets[DC__VOLTAGE_STATES] = {0};
+		unsigned int optimal_dcfclk_for_uclk[DC__VOLTAGE_STATES] = {0};
+
+		unsigned int dcfclk_sta_targets[DC__VOLTAGE_STATES] = {615, 906, 1324, 1564};
+		unsigned int num_dcfclk_sta_targets = 4, num_uclk_states = 0;
+		unsigned int max_dcfclk_mhz = 0, max_dispclk_mhz = 0, max_dppclk_mhz = 0, max_phyclk_mhz = 0;
+
+		for (i = 0; i < MAX_NUM_DPM_LVL; i++) {
+			if (bw_params->clk_table.entries[i].dcfclk_mhz > max_dcfclk_mhz)
+				max_dcfclk_mhz = bw_params->clk_table.entries[i].dcfclk_mhz;
+			if (bw_params->clk_table.entries[i].dispclk_mhz > max_dispclk_mhz)
+				max_dispclk_mhz = bw_params->clk_table.entries[i].dispclk_mhz;
+			if (bw_params->clk_table.entries[i].dppclk_mhz > max_dppclk_mhz)
+				max_dppclk_mhz = bw_params->clk_table.entries[i].dppclk_mhz;
+			if (bw_params->clk_table.entries[i].phyclk_mhz > max_phyclk_mhz)
+				max_phyclk_mhz = bw_params->clk_table.entries[i].phyclk_mhz;
+		}
+		if (!max_dcfclk_mhz)
+			max_dcfclk_mhz = dcn3_21_soc.clock_limits[0].dcfclk_mhz;
+		if (!max_dispclk_mhz)
+			max_dispclk_mhz = dcn3_21_soc.clock_limits[0].dispclk_mhz;
+		if (!max_dppclk_mhz)
+			max_dppclk_mhz = dcn3_21_soc.clock_limits[0].dppclk_mhz;
+		if (!max_phyclk_mhz)
+			max_phyclk_mhz = dcn3_21_soc.clock_limits[0].phyclk_mhz;
+
+		if (max_dcfclk_mhz > dcfclk_sta_targets[num_dcfclk_sta_targets-1]) {
+			// If max DCFCLK is greater than the max DCFCLK STA target, insert into the DCFCLK STA target array
+			dcfclk_sta_targets[num_dcfclk_sta_targets] = max_dcfclk_mhz;
+			num_dcfclk_sta_targets++;
+		} else if (max_dcfclk_mhz < dcfclk_sta_targets[num_dcfclk_sta_targets-1]) {
+			// If max DCFCLK is less than the max DCFCLK STA target, cap values and remove duplicates
+			for (i = 0; i < num_dcfclk_sta_targets; i++) {
+				if (dcfclk_sta_targets[i] > max_dcfclk_mhz) {
+					dcfclk_sta_targets[i] = max_dcfclk_mhz;
+					break;
+				}
+			}
+			// Update size of array since we "removed" duplicates
+			num_dcfclk_sta_targets = i + 1;
+		}
+
+		num_uclk_states = bw_params->clk_table.num_entries;
+
+		// Calculate optimal dcfclk for each uclk
+		for (i = 0; i < num_uclk_states; i++) {
+			dcn321_get_optimal_dcfclk_fclk_for_uclk(bw_params->clk_table.entries[i].memclk_mhz * 16,
+					&optimal_dcfclk_for_uclk[i], NULL);
+			if (optimal_dcfclk_for_uclk[i] < bw_params->clk_table.entries[0].dcfclk_mhz) {
+				optimal_dcfclk_for_uclk[i] = bw_params->clk_table.entries[0].dcfclk_mhz;
+			}
+		}
+
+		// Calculate optimal uclk for each dcfclk sta target
+		for (i = 0; i < num_dcfclk_sta_targets; i++) {
+			for (j = 0; j < num_uclk_states; j++) {
+				if (dcfclk_sta_targets[i] < optimal_dcfclk_for_uclk[j]) {
+					optimal_uclk_for_dcfclk_sta_targets[i] =
+							bw_params->clk_table.entries[j].memclk_mhz * 16;
+					break;
+				}
+			}
+		}
+
+		i = 0;
+		j = 0;
+		// create the final dcfclk and uclk table
+		while (i < num_dcfclk_sta_targets && j < num_uclk_states && num_states < DC__VOLTAGE_STATES) {
+			if (dcfclk_sta_targets[i] < optimal_dcfclk_for_uclk[j]) {
+				dcfclk_mhz[num_states] = dcfclk_sta_targets[i];
+				dram_speed_mts[num_states++] = optimal_uclk_for_dcfclk_sta_targets[i++];
+			} else {
+				if (j < num_uclk_states && optimal_dcfclk_for_uclk[j] <= max_dcfclk_mhz) {
+					dcfclk_mhz[num_states] = optimal_dcfclk_for_uclk[j];
+					dram_speed_mts[num_states++] = bw_params->clk_table.entries[j++].memclk_mhz * 16;
+				} else {
+					j = num_uclk_states;
+				}
+			}
+		}
+
+		while (i < num_dcfclk_sta_targets && num_states < DC__VOLTAGE_STATES) {
+			dcfclk_mhz[num_states] = dcfclk_sta_targets[i];
+			dram_speed_mts[num_states++] = optimal_uclk_for_dcfclk_sta_targets[i++];
+		}
+
+		while (j < num_uclk_states && num_states < DC__VOLTAGE_STATES &&
+				optimal_dcfclk_for_uclk[j] <= max_dcfclk_mhz) {
+			dcfclk_mhz[num_states] = optimal_dcfclk_for_uclk[j];
+			dram_speed_mts[num_states++] = bw_params->clk_table.entries[j++].memclk_mhz * 16;
+		}
+
+		/* bw_params->clk_table.entries[MAX_NUM_DPM_LVL].
+		 * MAX_NUM_DPM_LVL is 8.
+		 * dcn3_02_soc.clock_limits[DC__VOLTAGE_STATES].
+		 * DC__VOLTAGE_STATES is 40.
+		 */
+		if (num_states > MAX_NUM_DPM_LVL) {
+			ASSERT(0);
+			return;
+		}
+
+		dcn3_21_soc.num_states = num_states;
+		for (i = 0; i < dcn3_21_soc.num_states; i++) {
+			dcn3_21_soc.clock_limits[i].state = i;
+			dcn3_21_soc.clock_limits[i].dcfclk_mhz = dcfclk_mhz[i];
+			dcn3_21_soc.clock_limits[i].fabricclk_mhz = dcfclk_mhz[i];
+
+			/* Fill all states with max values of all these clocks */
+			dcn3_21_soc.clock_limits[i].dispclk_mhz = max_dispclk_mhz;
+			dcn3_21_soc.clock_limits[i].dppclk_mhz  = max_dppclk_mhz;
+			dcn3_21_soc.clock_limits[i].phyclk_mhz  = max_phyclk_mhz;
+			dcn3_21_soc.clock_limits[i].dscclk_mhz  = max_dispclk_mhz / 3;
+
+			/* Populate from bw_params for DTBCLK, SOCCLK */
+			if (i > 0) {
+				if (!bw_params->clk_table.entries[i].dtbclk_mhz) {
+					dcn3_21_soc.clock_limits[i].dtbclk_mhz  = dcn3_21_soc.clock_limits[i-1].dtbclk_mhz;
+				} else {
+					dcn3_21_soc.clock_limits[i].dtbclk_mhz  = bw_params->clk_table.entries[i].dtbclk_mhz;
+				}
+			} else if (bw_params->clk_table.entries[i].dtbclk_mhz) {
+				dcn3_21_soc.clock_limits[i].dtbclk_mhz  = bw_params->clk_table.entries[i].dtbclk_mhz;
+			}
+
+			if (!bw_params->clk_table.entries[i].socclk_mhz && i > 0)
+				dcn3_21_soc.clock_limits[i].socclk_mhz = dcn3_21_soc.clock_limits[i-1].socclk_mhz;
+			else
+				dcn3_21_soc.clock_limits[i].socclk_mhz = bw_params->clk_table.entries[i].socclk_mhz;
+
+			if (!dram_speed_mts[i] && i > 0)
+				dcn3_21_soc.clock_limits[i].dram_speed_mts = dcn3_21_soc.clock_limits[i-1].dram_speed_mts;
+			else
+				dcn3_21_soc.clock_limits[i].dram_speed_mts = dram_speed_mts[i];
+
+			/* These clocks cannot come from bw_params, always fill from dcn3_21_soc[0] */
+			/* PHYCLK_D18, PHYCLK_D32 */
+			dcn3_21_soc.clock_limits[i].phyclk_d18_mhz = dcn3_21_soc.clock_limits[0].phyclk_d18_mhz;
+			dcn3_21_soc.clock_limits[i].phyclk_d32_mhz = dcn3_21_soc.clock_limits[0].phyclk_d32_mhz;
+		}
+	} else {
+		build_synthetic_soc_states(dc->debug.disable_dc_mode_overwrite, bw_params,
+			dcn3_21_soc.clock_limits, &dcn3_21_soc.num_states);
+	}
+
+	/* Re-init DML with updated bb */
+	dml_init_instance(&dc->dml, &dcn3_21_soc, &dcn3_21_ip, DML_PROJECT_DCN32);
+	if (dc->current_state)
+		dml_init_instance(&dc->current_state->bw_ctx.dml, &dcn3_21_soc, &dcn3_21_ip, DML_PROJECT_DCN32);
+
+	if (dc->clk_mgr->bw_params->clk_table.num_entries > 1) {
+		unsigned int i = 0;
+
+		dc->dml2_options.bbox_overrides.clks_table.num_states = dc->clk_mgr->bw_params->clk_table.num_entries;
+
+		dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_dcfclk_levels =
+			dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_dcfclk_levels;
+
+		dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_fclk_levels =
+			dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_fclk_levels;
+
+		dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_memclk_levels =
+			dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_memclk_levels;
+
+		dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_socclk_levels =
+			dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_socclk_levels;
+
+		dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_dtbclk_levels =
+			dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_dtbclk_levels;
+
+		dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_dispclk_levels =
+			dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_dispclk_levels;
+
+		dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_dppclk_levels =
+			dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_dppclk_levels;
+
+
+		for (i = 0; i < dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_dcfclk_levels; i++) {
+			if (dc->clk_mgr->bw_params->clk_table.entries[i].dcfclk_mhz)
+				dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].dcfclk_mhz =
+					dc->clk_mgr->bw_params->clk_table.entries[i].dcfclk_mhz;
+		}
+
+		for (i = 0; i < dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_fclk_levels; i++) {
+			if (dc->clk_mgr->bw_params->clk_table.entries[i].fclk_mhz)
+				dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].fclk_mhz =
+					dc->clk_mgr->bw_params->clk_table.entries[i].fclk_mhz;
+		}
+
+		for (i = 0; i < dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_memclk_levels; i++) {
+			if (dc->clk_mgr->bw_params->clk_table.entries[i].memclk_mhz)
+				dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].memclk_mhz =
+					dc->clk_mgr->bw_params->clk_table.entries[i].memclk_mhz;
+		}
+
+		for (i = 0; i < dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_socclk_levels; i++) {
+			if (dc->clk_mgr->bw_params->clk_table.entries[i].socclk_mhz)
+				dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].socclk_mhz =
+					dc->clk_mgr->bw_params->clk_table.entries[i].socclk_mhz;
+		}
+
+		for (i = 0; i < dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_dtbclk_levels; i++) {
+			if (dc->clk_mgr->bw_params->clk_table.entries[i].dtbclk_mhz)
+				dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].dtbclk_mhz =
+					dc->clk_mgr->bw_params->clk_table.entries[i].dtbclk_mhz;
+		}
+
+		for (i = 0; i < dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_dispclk_levels; i++) {
+			if (dc->clk_mgr->bw_params->clk_table.entries[i].dispclk_mhz) {
+				dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].dispclk_mhz =
+					dc->clk_mgr->bw_params->clk_table.entries[i].dispclk_mhz;
+				dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].dppclk_mhz =
+					dc->clk_mgr->bw_params->clk_table.entries[i].dispclk_mhz;
+			}
+		}
+	}
+}
+
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn321/dcn321_fpu.h b/drivers/gpu/drm/amd/display/dc/dml/dcn321/dcn321_fpu.h
new file mode 100644
index 000000000000..c6623b3705ca
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn321/dcn321_fpu.h
@@ -0,0 +1,34 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __DCN32_FPU_H__
+#define __DCN32_FPU_H__
+
+#include "dml/display_mode_vba.h"
+
+void dcn321_update_bw_bounding_box_fpu(struct dc *dc, struct clk_bw_params *bw_params);
+
+#endif
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn35/dcn35_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn35/dcn35_fpu.c
new file mode 100644
index 000000000000..817a370e80a7
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn35/dcn35_fpu.c
@@ -0,0 +1,620 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+#include "resource.h"
+#include "dcn35_fpu.h"
+#include "dcn31/dcn31_resource.h"
+#include "dcn32/dcn32_resource.h"
+#include "dcn35/dcn35_resource.h"
+#include "dml/dcn31/dcn31_fpu.h"
+#include "dml/dml_inline_defs.h"
+
+#include "link_service.h"
+
+#define DC_LOGGER_INIT(logger)
+
+struct _vcs_dpi_ip_params_st dcn3_5_ip = {
+	.VBlankNomDefaultUS = 668,
+	.gpuvm_enable = 1,
+	.gpuvm_max_page_table_levels = 1,
+	.hostvm_enable = 1,
+	.hostvm_max_page_table_levels = 2,
+	.rob_buffer_size_kbytes = 64,
+	.det_buffer_size_kbytes = 1536,
+	.config_return_buffer_size_in_kbytes = 1792,
+	.compressed_buffer_segment_size_in_kbytes = 64,
+	.meta_fifo_size_in_kentries = 32,
+	.zero_size_buffer_entries = 512,
+	.compbuf_reserved_space_64b = 256,
+	.compbuf_reserved_space_zs = 64,
+	.dpp_output_buffer_pixels = 2560,/*not used*/
+	.opp_output_buffer_lines = 1,/*not used*/
+	.pixel_chunk_size_kbytes = 8,
+	//.alpha_pixel_chunk_size_kbytes = 4;/*new*/
+	//.min_pixel_chunk_size_bytes = 1024;/*new*/
+	.meta_chunk_size_kbytes = 2,
+	.min_meta_chunk_size_bytes = 256,
+	.writeback_chunk_size_kbytes = 8,
+	.ptoi_supported = false,
+	.num_dsc = 4,
+	.maximum_dsc_bits_per_component = 12,/*delta from 10*/
+	.dsc422_native_support = true,/*delta from false*/
+	.is_line_buffer_bpp_fixed = true,/*new*/
+	.line_buffer_fixed_bpp = 32,/*delta from 48*/
+	.line_buffer_size_bits = 986880,/*delta from 789504*/
+	.max_line_buffer_lines = 32,/*delta from 12*/
+	.writeback_interface_buffer_size_kbytes = 90,
+	.max_num_dpp = 4,
+	.max_num_otg = 4,
+	.max_num_hdmi_frl_outputs = 1,
+	.max_num_wb = 1,
+	/*.max_num_hdmi_frl_outputs = 1; new in dml2*/
+	/*.max_num_dp2p0_outputs = 2; new in dml2*/
+	/*.max_num_dp2p0_streams = 4; new in dml2*/
+	.max_dchub_pscl_bw_pix_per_clk = 4,
+	.max_pscl_lb_bw_pix_per_clk = 2,
+	.max_lb_vscl_bw_pix_per_clk = 4,
+	.max_vscl_hscl_bw_pix_per_clk = 4,
+	.max_hscl_ratio = 6,
+	.max_vscl_ratio = 6,
+	.max_hscl_taps = 8,
+	.max_vscl_taps = 8,
+	.dpte_buffer_size_in_pte_reqs_luma = 68,/*changed from 64,*/
+	.dpte_buffer_size_in_pte_reqs_chroma = 36,/*changed from 34*/
+	/*.dcc_meta_buffer_size_bytes = 6272; new to dml2*/
+	.dispclk_ramp_margin_percent = 1.11,/*delta from 1*/
+	/*.dppclk_delay_subtotal = 47;
+	.dppclk_delay_scl = 50;
+	.dppclk_delay_scl_lb_only = 16;
+	.dppclk_delay_cnvc_formatter = 28;
+	.dppclk_delay_cnvc_cursor = 6;
+	.dispclk_delay_subtotal = 125;*/ /*new to dml2*/
+	.max_inter_dcn_tile_repeaters = 8,
+	.cursor_buffer_size = 16,
+	.cursor_chunk_size = 2,
+	.writeback_line_buffer_buffer_size = 0,
+	.writeback_min_hscl_ratio = 1,
+	.writeback_min_vscl_ratio = 1,
+	.writeback_max_hscl_ratio = 1,
+	.writeback_max_vscl_ratio = 1,
+	.writeback_max_hscl_taps = 1,
+	.writeback_max_vscl_taps = 1,
+	.dppclk_delay_subtotal = 47, /* changed from 46,*/
+	.dppclk_delay_scl = 50,
+	.dppclk_delay_scl_lb_only = 16,
+	.dppclk_delay_cnvc_formatter = 28,/*changed from 27,*/
+	.dppclk_delay_cnvc_cursor = 6,
+	.dispclk_delay_subtotal = 125, /*changed from 119,*/
+	.dynamic_metadata_vm_enabled = false,
+	.odm_combine_4to1_supported = false,
+	.dcc_supported = true,
+//	.config_return_buffer_segment_size_in_kbytes = 64;/*required, hard coded in dml2_translate_ip_params*/
+
+};
+
+struct _vcs_dpi_soc_bounding_box_st dcn3_5_soc = {
+		/*TODO: correct dispclk/dppclk voltage level determination*/
+	.clock_limits = {
+		{
+			.state = 0,
+			.dispclk_mhz = 1200.0,
+			.dppclk_mhz = 1200.0,
+			.phyclk_mhz = 600.0,
+			.phyclk_d18_mhz = 667.0,
+			.dscclk_mhz = 186.0,
+			.dtbclk_mhz = 600.0,
+		},
+		{
+			.state = 1,
+			.dispclk_mhz = 1200.0,
+			.dppclk_mhz = 1200.0,
+			.phyclk_mhz = 810.0,
+			.phyclk_d18_mhz = 667.0,
+			.dscclk_mhz = 209.0,
+			.dtbclk_mhz = 600.0,
+		},
+		{
+			.state = 2,
+			.dispclk_mhz = 1200.0,
+			.dppclk_mhz = 1200.0,
+			.phyclk_mhz = 810.0,
+			.phyclk_d18_mhz = 667.0,
+			.dscclk_mhz = 209.0,
+			.dtbclk_mhz = 600.0,
+		},
+		{
+			.state = 3,
+			.dispclk_mhz = 1200.0,
+			.dppclk_mhz = 1200.0,
+			.phyclk_mhz = 810.0,
+			.phyclk_d18_mhz = 667.0,
+			.dscclk_mhz = 371.0,
+			.dtbclk_mhz = 600.0,
+		},
+		{
+			.state = 4,
+			.dispclk_mhz = 1200.0,
+			.dppclk_mhz = 1200.0,
+			.phyclk_mhz = 810.0,
+			.phyclk_d18_mhz = 667.0,
+			.dscclk_mhz = 400.0,
+			.dtbclk_mhz = 600.0,
+		},
+	},
+	.num_states = 5,
+	.sr_exit_time_us = 28.0,
+	.sr_enter_plus_exit_time_us = 30.0,
+	.sr_exit_z8_time_us = 250.0,
+	.sr_enter_plus_exit_z8_time_us = 350.0,
+	.fclk_change_latency_us = 24.0,
+	.usr_retraining_latency_us = 2,
+	.writeback_latency_us = 12.0,
+
+	.dram_channel_width_bytes = 4,/*not exist in dml2*/
+	.round_trip_ping_latency_dcfclk_cycles = 106,/*not exist in dml2*/
+	.urgent_latency_pixel_data_only_us = 4.0,
+	.urgent_latency_pixel_mixed_with_vm_data_us = 4.0,
+	.urgent_latency_vm_data_only_us = 4.0,
+	.dram_clock_change_latency_us = 34.0,
+	.urgent_out_of_order_return_per_channel_pixel_only_bytes = 4096,
+	.urgent_out_of_order_return_per_channel_pixel_and_vm_bytes = 4096,
+	.urgent_out_of_order_return_per_channel_vm_only_bytes = 4096,
+
+	.pct_ideal_sdp_bw_after_urgent = 80.0,
+	.pct_ideal_fabric_bw_after_urgent = 80.0, /*new to dml2*/
+	.pct_ideal_dram_sdp_bw_after_urgent_pixel_only = 65.0,
+	.pct_ideal_dram_sdp_bw_after_urgent_pixel_and_vm = 60.0,
+	.pct_ideal_dram_sdp_bw_after_urgent_vm_only = 30.0,
+	.max_avg_sdp_bw_use_normal_percent = 60.0,
+	.max_avg_dram_bw_use_normal_percent = 60.0,
+	.fabric_datapath_to_dcn_data_return_bytes = 32,
+	.return_bus_width_bytes = 64,
+	.downspread_percent = 0.38,
+	.dcn_downspread_percent = 0.5,
+	.gpuvm_min_page_size_bytes = 4096,
+	.hostvm_min_page_size_bytes = 4096,
+	.do_urgent_latency_adjustment = 0,
+	.urgent_latency_adjustment_fabric_clock_component_us = 0,
+	.urgent_latency_adjustment_fabric_clock_reference_mhz = 0,
+};
+
+void dcn35_build_wm_range_table_fpu(struct clk_mgr *clk_mgr)
+{
+	//TODO
+}
+
+
+/*
+ * dcn35_update_bw_bounding_box
+ *
+ * This would override some dcn3_5 ip_or_soc initial parameters hardcoded from
+ * spreadsheet with actual values as per dGPU SKU:
+ * - with passed few options from dc->config
+ * - with dentist_vco_frequency from Clk Mgr (currently hardcoded, but might
+ *   need to get it from PM FW)
+ * - with passed latency values (passed in ns units) in dc-> bb override for
+ *   debugging purposes
+ * - with passed latencies from VBIOS (in 100_ns units) if available for
+ *   certain dGPU SKU
+ * - with number of DRAM channels from VBIOS (which differ for certain dGPU SKU
+ *   of the same ASIC)
+ * - clocks levels with passed clk_table entries from Clk Mgr as reported by PM
+ *   FW for different clocks (which might differ for certain dGPU SKU of the
+ *   same ASIC)
+ */
+void dcn35_update_bw_bounding_box_fpu(struct dc *dc,
+				      struct clk_bw_params *bw_params)
+{
+	unsigned int i, closest_clk_lvl;
+	int j;
+	struct clk_limit_table *clk_table = &bw_params->clk_table;
+	struct _vcs_dpi_voltage_scaling_st *clock_limits =
+		dc->scratch.update_bw_bounding_box.clock_limits;
+	int max_dispclk_mhz = 0, max_dppclk_mhz = 0;
+
+	dc_assert_fp_enabled();
+
+	dcn3_5_ip.max_num_otg =
+		dc->res_pool->res_cap->num_timing_generator;
+	dcn3_5_ip.max_num_dpp = dc->res_pool->pipe_count;
+	dcn3_5_soc.num_chans = bw_params->num_channels;
+
+	ASSERT(clk_table->num_entries);
+
+	/* Prepass to find max clocks independent of voltage level. */
+	for (i = 0; i < clk_table->num_entries; ++i) {
+		if (clk_table->entries[i].dispclk_mhz > max_dispclk_mhz)
+			max_dispclk_mhz = clk_table->entries[i].dispclk_mhz;
+		if (clk_table->entries[i].dppclk_mhz > max_dppclk_mhz)
+			max_dppclk_mhz = clk_table->entries[i].dppclk_mhz;
+	}
+
+	for (i = 0; i < clk_table->num_entries; i++) {
+		/* loop backwards*/
+		for (closest_clk_lvl = 0, j = dcn3_5_soc.num_states - 1;
+			j >= 0; j--) {
+			if (dcn3_5_soc.clock_limits[j].dcfclk_mhz <=
+				clk_table->entries[i].dcfclk_mhz) {
+				closest_clk_lvl = j;
+				break;
+			}
+		}
+		if (clk_table->num_entries == 1) {
+			/*smu gives one DPM level, let's take the highest one*/
+			closest_clk_lvl = dcn3_5_soc.num_states - 1;
+		}
+
+		clock_limits[i].state = i;
+
+		/* Clocks dependent on voltage level. */
+		clock_limits[i].dcfclk_mhz = clk_table->entries[i].dcfclk_mhz;
+		if (clk_table->num_entries == 1 &&
+			clock_limits[i].dcfclk_mhz <
+			dcn3_5_soc.clock_limits[closest_clk_lvl].dcfclk_mhz) {
+			/*SMU fix not released yet*/
+			clock_limits[i].dcfclk_mhz =
+				dcn3_5_soc.clock_limits[closest_clk_lvl].dcfclk_mhz;
+		}
+
+		clock_limits[i].fabricclk_mhz =
+			clk_table->entries[i].fclk_mhz;
+		clock_limits[i].socclk_mhz =
+			clk_table->entries[i].socclk_mhz;
+
+		if (clk_table->entries[i].memclk_mhz &&
+			clk_table->entries[i].wck_ratio)
+			clock_limits[i].dram_speed_mts =
+				clk_table->entries[i].memclk_mhz * 2 *
+				clk_table->entries[i].wck_ratio;
+
+		/* Clocks independent of voltage level. */
+		clock_limits[i].dispclk_mhz = max_dispclk_mhz ?
+			max_dispclk_mhz :
+			dcn3_5_soc.clock_limits[closest_clk_lvl].dispclk_mhz;
+
+		clock_limits[i].dppclk_mhz = max_dppclk_mhz ?
+			max_dppclk_mhz :
+			dcn3_5_soc.clock_limits[closest_clk_lvl].dppclk_mhz;
+
+		clock_limits[i].dram_bw_per_chan_gbps =
+			dcn3_5_soc.clock_limits[closest_clk_lvl].dram_bw_per_chan_gbps;
+		clock_limits[i].dscclk_mhz =
+			dcn3_5_soc.clock_limits[closest_clk_lvl].dscclk_mhz;
+		clock_limits[i].dtbclk_mhz =
+			dcn3_5_soc.clock_limits[closest_clk_lvl].dtbclk_mhz;
+		clock_limits[i].phyclk_d18_mhz =
+			dcn3_5_soc.clock_limits[closest_clk_lvl].phyclk_d18_mhz;
+		clock_limits[i].phyclk_mhz =
+			dcn3_5_soc.clock_limits[closest_clk_lvl].phyclk_mhz;
+	}
+
+	memcpy(dcn3_5_soc.clock_limits, clock_limits,
+		sizeof(dcn3_5_soc.clock_limits));
+
+	if (clk_table->num_entries)
+		dcn3_5_soc.num_states = clk_table->num_entries;
+
+	if (max_dispclk_mhz) {
+		dcn3_5_soc.dispclk_dppclk_vco_speed_mhz = max_dispclk_mhz * 2;
+		dc->dml.soc.dispclk_dppclk_vco_speed_mhz = max_dispclk_mhz * 2;
+	}
+	if ((int)(dcn3_5_soc.dram_clock_change_latency_us * 1000)
+				!= dc->debug.dram_clock_change_latency_ns
+			&& dc->debug.dram_clock_change_latency_ns) {
+		dcn3_5_soc.dram_clock_change_latency_us =
+			dc->debug.dram_clock_change_latency_ns / 1000.0;
+	}
+
+	if (dc->bb_overrides.dram_clock_change_latency_ns > 0)
+		dcn3_5_soc.dram_clock_change_latency_us =
+			dc->bb_overrides.dram_clock_change_latency_ns / 1000.0;
+
+	if (dc->bb_overrides.sr_exit_time_ns > 0)
+		dcn3_5_soc.sr_exit_time_us = dc->bb_overrides.sr_exit_time_ns / 1000.0;
+
+	if (dc->bb_overrides.sr_enter_plus_exit_time_ns > 0)
+		dcn3_5_soc.sr_enter_plus_exit_time_us =
+			dc->bb_overrides.sr_enter_plus_exit_time_ns / 1000.0;
+
+	if (dc->bb_overrides.sr_exit_z8_time_ns > 0)
+		dcn3_5_soc.sr_exit_z8_time_us = dc->bb_overrides.sr_exit_z8_time_ns / 1000.0;
+
+	if (dc->bb_overrides.sr_enter_plus_exit_z8_time_ns > 0)
+		dcn3_5_soc.sr_enter_plus_exit_z8_time_us =
+			dc->bb_overrides.sr_enter_plus_exit_z8_time_ns / 1000.0;
+
+	/*temp till dml2 fully work without dml1*/
+	dml_init_instance(&dc->dml, &dcn3_5_soc, &dcn3_5_ip,
+				DML_PROJECT_DCN31);
+
+	/*copy to dml2, before dml2_create*/
+	if (clk_table->num_entries > 2) {
+
+		for (i = 0; i < clk_table->num_entries; i++) {
+			dc->dml2_options.bbox_overrides.clks_table.num_states =
+				clk_table->num_entries;
+			dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].dcfclk_mhz =
+				clock_limits[i].dcfclk_mhz;
+			dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].fclk_mhz =
+				clock_limits[i].fabricclk_mhz;
+			dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].dispclk_mhz =
+				clock_limits[i].dispclk_mhz;
+			dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].dppclk_mhz =
+				clock_limits[i].dppclk_mhz;
+			dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].socclk_mhz =
+				clock_limits[i].socclk_mhz;
+			dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].memclk_mhz =
+				clk_table->entries[i].memclk_mhz * clk_table->entries[i].wck_ratio;
+
+			dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].dram_speed_mts = clock_limits[i].dram_speed_mts;
+			dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].dtbclk_mhz =
+				clock_limits[i].dtbclk_mhz;
+			dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_dcfclk_levels =
+				clk_table->num_entries;
+			dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_fclk_levels =
+				clk_table->num_entries;
+			dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_dispclk_levels =
+				clk_table->num_entries;
+			dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_dppclk_levels =
+				clk_table->num_entries;
+			dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_socclk_levels =
+				clk_table->num_entries;
+			dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_memclk_levels =
+				clk_table->num_entries;
+			dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_dtbclk_levels =
+				clk_table->num_entries;
+		}
+	}
+
+	/* Update latency values */
+	dc->dml2_options.bbox_overrides.dram_clock_change_latency_us = dcn3_5_soc.dram_clock_change_latency_us;
+
+	dc->dml2_options.bbox_overrides.sr_exit_latency_us = dcn3_5_soc.sr_exit_time_us;
+	dc->dml2_options.bbox_overrides.sr_enter_plus_exit_latency_us = dcn3_5_soc.sr_enter_plus_exit_time_us;
+
+	dc->dml2_options.bbox_overrides.sr_exit_z8_time_us = dcn3_5_soc.sr_exit_z8_time_us;
+	dc->dml2_options.bbox_overrides.sr_enter_plus_exit_z8_time_us = dcn3_5_soc.sr_enter_plus_exit_z8_time_us;
+}
+
+static bool is_dual_plane(enum surface_pixel_format format)
+{
+	return format >= SURFACE_PIXEL_FORMAT_VIDEO_BEGIN ||
+		format == SURFACE_PIXEL_FORMAT_GRPH_RGBE_ALPHA;
+}
+
+/*
+ * micro_sec_to_vert_lines () - converts time to number of vertical lines for a given timing
+ *
+ * @param: num_us: number of microseconds
+ * @return: number of vertical lines. If exact number of vertical lines is not found then
+ *          it will round up to next number of lines to guarantee num_us
+ */
+static unsigned int micro_sec_to_vert_lines(unsigned int num_us, struct dc_crtc_timing *timing)
+{
+	unsigned int num_lines = 0;
+	unsigned int lines_time_in_ns = 1000.0 *
+			(((float)timing->h_total * 1000.0) /
+			 ((float)timing->pix_clk_100hz / 10.0));
+
+	num_lines = dml_ceil(1000.0 * num_us / lines_time_in_ns, 1.0);
+
+	return num_lines;
+}
+
+static unsigned int get_vertical_back_porch(struct dc_crtc_timing *timing)
+{
+	unsigned int v_active = 0, v_blank = 0, v_back_porch = 0;
+
+	v_active = timing->v_border_top + timing->v_addressable + timing->v_border_bottom;
+	v_blank = timing->v_total - v_active;
+	v_back_porch = v_blank - timing->v_front_porch - timing->v_sync_width;
+
+	return v_back_porch;
+}
+
+int dcn35_populate_dml_pipes_from_context_fpu(struct dc *dc,
+					      struct dc_state *context,
+					      display_e2e_pipe_params_st *pipes,
+					      enum dc_validate_mode validate_mode)
+{
+	int i, pipe_cnt;
+	struct resource_context *res_ctx = &context->res_ctx;
+	struct pipe_ctx *pipe = 0;
+	bool upscaled = false;
+	const unsigned int max_allowed_vblank_nom = 1023;
+
+	dc_assert_fp_enabled();
+
+	dcn31_populate_dml_pipes_from_context(dc, context, pipes,
+					      validate_mode);
+
+	for (i = 0, pipe_cnt = 0; i < dc->res_pool->pipe_count; i++) {
+		struct dc_crtc_timing *timing;
+		unsigned int num_lines = 0;
+		unsigned int v_back_porch = 0;
+
+		if (!res_ctx->pipe_ctx[i].stream)
+			continue;
+
+		pipe = &res_ctx->pipe_ctx[i];
+		timing = &pipe->stream->timing;
+
+		num_lines = micro_sec_to_vert_lines(dcn3_5_ip.VBlankNomDefaultUS, timing);
+		v_back_porch  = get_vertical_back_porch(timing);
+
+		if (pipe->stream->adjust.v_total_max ==
+		    pipe->stream->adjust.v_total_min &&
+		    pipe->stream->adjust.v_total_min > timing->v_total) {
+			pipes[pipe_cnt].pipe.dest.vtotal =
+				pipe->stream->adjust.v_total_min;
+			pipes[pipe_cnt].pipe.dest.vblank_nom = timing->v_total -
+				pipes[pipe_cnt].pipe.dest.vactive;
+		}
+
+		pipes[pipe_cnt].pipe.dest.vblank_nom = timing->v_total - pipes[pipe_cnt].pipe.dest.vactive;
+		pipes[pipe_cnt].pipe.dest.vblank_nom = min(pipes[pipe_cnt].pipe.dest.vblank_nom, num_lines);
+		// vblank_nom should not smaller than (VSync (timing->v_sync_width + v_back_porch) + 2)
+		// + 2 is because
+		// 1 -> VStartup_start should be 1 line before VSync
+		// 1 -> always reserve 1 line between start of vblank to vstartup signal
+		pipes[pipe_cnt].pipe.dest.vblank_nom =
+			max(pipes[pipe_cnt].pipe.dest.vblank_nom, timing->v_sync_width + v_back_porch + 2);
+		pipes[pipe_cnt].pipe.dest.vblank_nom = min(pipes[pipe_cnt].pipe.dest.vblank_nom, max_allowed_vblank_nom);
+
+		if (pipe->plane_state &&
+		    (pipe->plane_state->src_rect.height <
+		     pipe->plane_state->dst_rect.height ||
+		     pipe->plane_state->src_rect.width <
+		     pipe->plane_state->dst_rect.width))
+			upscaled = true;
+
+		/*
+		 * Immediate flip can be set dynamically after enabling the
+		 * plane. We need to require support for immediate flip or
+		 * underflow can be intermittently experienced depending on peak
+		 * b/w requirements.
+		 */
+		pipes[pipe_cnt].pipe.src.immediate_flip = true;
+
+		pipes[pipe_cnt].pipe.src.unbounded_req_mode = false;
+
+		dcn31_zero_pipe_dcc_fraction(pipes, pipe_cnt);
+
+		pipes[pipe_cnt].pipe.dest.vfront_porch = timing->v_front_porch;
+		pipes[pipe_cnt].pipe.src.dcc_rate = 3;
+		pipes[pipe_cnt].dout.dsc_input_bpc = 0;
+		pipes[pipe_cnt].pipe.src.gpuvm_min_page_size_kbytes = 256;
+
+		if (pipes[pipe_cnt].dout.dsc_enable) {
+			switch (timing->display_color_depth) {
+			case COLOR_DEPTH_888:
+				pipes[pipe_cnt].dout.dsc_input_bpc = 8;
+				break;
+			case COLOR_DEPTH_101010:
+				pipes[pipe_cnt].dout.dsc_input_bpc = 10;
+				break;
+			case COLOR_DEPTH_121212:
+				pipes[pipe_cnt].dout.dsc_input_bpc = 12;
+				break;
+			default:
+				ASSERT(0);
+				break;
+			}
+		}
+
+		pipe_cnt++;
+	}
+
+	context->bw_ctx.dml.ip.det_buffer_size_kbytes = 384;/*per guide*/
+	dc->config.enable_4to1MPC = false;
+
+	if (pipe_cnt == 1 && pipe->plane_state && !dc->debug.disable_z9_mpc) {
+		if (is_dual_plane(pipe->plane_state->format)
+				&& pipe->plane_state->src_rect.width <= 1920 &&
+				pipe->plane_state->src_rect.height <= 1080) {
+			dc->config.enable_4to1MPC = true;
+		} else if (!is_dual_plane(pipe->plane_state->format) &&
+			   pipe->plane_state->src_rect.width <= 5120) {
+			/*
+			 * Limit to 5k max to avoid forced pipe split when there
+			 * is not enough detile for swath
+			 */
+			context->bw_ctx.dml.ip.det_buffer_size_kbytes = 192;
+			pipes[0].pipe.src.unbounded_req_mode = true;
+		}
+	} else if (context->stream_count >=
+		   dc->debug.crb_alloc_policy_min_disp_count &&
+		   dc->debug.crb_alloc_policy > DET_SIZE_DEFAULT) {
+		context->bw_ctx.dml.ip.det_buffer_size_kbytes =
+			dc->debug.crb_alloc_policy * 64;
+	} else if (context->stream_count >= 3 && upscaled) {
+		context->bw_ctx.dml.ip.det_buffer_size_kbytes = 192;
+	}
+
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
+
+		if (!pipe->stream)
+			continue;
+
+		if (pipe->stream->signal == SIGNAL_TYPE_EDP &&
+		    dc->debug.seamless_boot_odm_combine &&
+		    pipe->stream->apply_seamless_boot_optimization) {
+
+			if (pipe->stream->apply_boot_odm_mode ==
+			    dm_odm_combine_policy_2to1) {
+				context->bw_ctx.dml.vba.ODMCombinePolicy =
+					dm_odm_combine_policy_2to1;
+				break;
+			}
+		}
+	}
+
+	return pipe_cnt;
+}
+
+void dcn35_decide_zstate_support(struct dc *dc, struct dc_state *context)
+{
+	enum dcn_zstate_support_state support = DCN_ZSTATE_SUPPORT_DISALLOW;
+	unsigned int i, plane_count = 0;
+	DC_LOGGER_INIT(dc->ctx->logger);
+
+	dc_assert_fp_enabled();
+
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		if (context->res_ctx.pipe_ctx[i].plane_state)
+			plane_count++;
+	}
+
+	if (context->stream_count == 0 || plane_count == 0) {
+		support = DCN_ZSTATE_SUPPORT_ALLOW;
+	} else if (context->stream_count == 1 && context->streams[0]->signal == SIGNAL_TYPE_EDP) {
+		struct dc_link *link = context->streams[0]->sink->link;
+		bool is_pwrseq0 = link && link->link_index == 0;
+		bool is_psr = (link && (link->psr_settings.psr_version == DC_PSR_VERSION_1 ||
+								link->psr_settings.psr_version == DC_PSR_VERSION_SU_1) && !link->panel_config.psr.disable_psr);
+		bool is_replay = link && link->replay_settings.replay_feature_enabled;
+		int minmum_z8_residency =
+			dc->debug.minimum_z8_residency_time > 0 ? dc->debug.minimum_z8_residency_time : 1000;
+		bool allow_z8 = context->bw_ctx.dml.vba.StutterPeriod > (double)minmum_z8_residency;
+		int minmum_z10_residency =
+			dc->debug.minimum_z10_residency_time > 0 ? dc->debug.minimum_z10_residency_time : 5000;
+		bool allow_z10 = context->bw_ctx.dml.vba.StutterPeriod > (double)minmum_z10_residency;
+
+		/*for psr1/psr-su, we allow z8 and z10 based on latency, for replay with IPS enabled, it will enter ips2*/
+		if (is_pwrseq0 && allow_z10)
+			support = DCN_ZSTATE_SUPPORT_ALLOW;
+		else if (is_pwrseq0 && (is_psr || is_replay))
+			support = DCN_ZSTATE_SUPPORT_ALLOW_Z8_Z10_ONLY;
+		else if (allow_z8)
+			support = DCN_ZSTATE_SUPPORT_ALLOW_Z8_ONLY;
+
+	}
+
+	DC_LOG_SMU("zstate_support: %d, StutterPeriod: %d\n", support,
+		   (int)context->bw_ctx.dml.vba.StutterPeriod);
+
+	context->bw_ctx.bw.dcn.clk.zstate_support = support;
+}
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn35/dcn35_fpu.h b/drivers/gpu/drm/amd/display/dc/dml/dcn35/dcn35_fpu.h
new file mode 100644
index 000000000000..d121c5afce71
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn35/dcn35_fpu.h
@@ -0,0 +1,44 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __DCN35_FPU_H__
+#define __DCN35_FPU_H__
+
+#include "clk_mgr.h"
+
+void dcn35_build_wm_range_table_fpu(struct clk_mgr *clk_mgr);
+
+void dcn35_update_bw_bounding_box_fpu(struct dc *dc,
+				      struct clk_bw_params *bw_params);
+
+int dcn35_populate_dml_pipes_from_context_fpu(struct dc *dc,
+					      struct dc_state *context,
+					      display_e2e_pipe_params_st *pipes,
+					      enum dc_validate_mode validate_mode);
+
+void dcn35_decide_zstate_support(struct dc *dc, struct dc_state *context);
+
+#endif
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn351/dcn351_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn351/dcn351_fpu.c
new file mode 100644
index 000000000000..77023b619f1e
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn351/dcn351_fpu.c
@@ -0,0 +1,639 @@
+/* SPDX-License-Identifier: MIT */
+/* Copyright 2024 Advanced Micro Devices, Inc. */
+#include "resource.h"
+#include "dcn351_fpu.h"
+#include "dcn31/dcn31_resource.h"
+#include "dcn32/dcn32_resource.h"
+#include "dcn35/dcn35_resource.h"
+#include "dcn351/dcn351_resource.h"
+#include "dml/dcn31/dcn31_fpu.h"
+#include "dml/dcn35/dcn35_fpu.h"
+#include "dml/dml_inline_defs.h"
+
+#include "link_service.h"
+
+#define DC_LOGGER_INIT(logger)
+
+struct _vcs_dpi_ip_params_st dcn3_51_ip = {
+	.VBlankNomDefaultUS = 668,
+	.gpuvm_enable = 1,
+	.gpuvm_max_page_table_levels = 1,
+	.hostvm_enable = 1,
+	.hostvm_max_page_table_levels = 2,
+	.rob_buffer_size_kbytes = 64,
+	.det_buffer_size_kbytes = 1536,
+	.config_return_buffer_size_in_kbytes = 1792,
+	.compressed_buffer_segment_size_in_kbytes = 64,
+	.meta_fifo_size_in_kentries = 32,
+	.zero_size_buffer_entries = 512,
+	.compbuf_reserved_space_64b = 256,
+	.compbuf_reserved_space_zs = 64,
+	.dpp_output_buffer_pixels = 2560,/*not used*/
+	.opp_output_buffer_lines = 1,/*not used*/
+	.pixel_chunk_size_kbytes = 8,
+	//.alpha_pixel_chunk_size_kbytes = 4;/*new*/
+	//.min_pixel_chunk_size_bytes = 1024;/*new*/
+	.meta_chunk_size_kbytes = 2,
+	.min_meta_chunk_size_bytes = 256,
+	.writeback_chunk_size_kbytes = 8,
+	.ptoi_supported = false,
+	.num_dsc = 4,
+	.maximum_dsc_bits_per_component = 12,/*delta from 10*/
+	.dsc422_native_support = true,/*delta from false*/
+	.is_line_buffer_bpp_fixed = true,/*new*/
+	.line_buffer_fixed_bpp = 32,/*delta from 48*/
+	.line_buffer_size_bits = 986880,/*delta from 789504*/
+	.max_line_buffer_lines = 32,/*delta from 12*/
+	.writeback_interface_buffer_size_kbytes = 90,
+	.max_num_dpp = 4,
+	.max_num_otg = 4,
+	.max_num_hdmi_frl_outputs = 1,
+	.max_num_wb = 1,
+	/*.max_num_hdmi_frl_outputs = 1; new in dml2*/
+	/*.max_num_dp2p0_outputs = 2; new in dml2*/
+	/*.max_num_dp2p0_streams = 4; new in dml2*/
+	.max_dchub_pscl_bw_pix_per_clk = 4,
+	.max_pscl_lb_bw_pix_per_clk = 2,
+	.max_lb_vscl_bw_pix_per_clk = 4,
+	.max_vscl_hscl_bw_pix_per_clk = 4,
+	.max_hscl_ratio = 6,
+	.max_vscl_ratio = 6,
+	.max_hscl_taps = 8,
+	.max_vscl_taps = 8,
+	.dpte_buffer_size_in_pte_reqs_luma = 68,/*changed from 64,*/
+	.dpte_buffer_size_in_pte_reqs_chroma = 36,/*changed from 34*/
+	/*.dcc_meta_buffer_size_bytes = 6272; new to dml2*/
+	.dispclk_ramp_margin_percent = 1.11,/*delta from 1*/
+	/*.dppclk_delay_subtotal = 47;
+	.dppclk_delay_scl = 50;
+	.dppclk_delay_scl_lb_only = 16;
+	.dppclk_delay_cnvc_formatter = 28;
+	.dppclk_delay_cnvc_cursor = 6;
+	.dispclk_delay_subtotal = 125;*/ /*new to dml2*/
+	.max_inter_dcn_tile_repeaters = 8,
+	.cursor_buffer_size = 16,
+	.cursor_chunk_size = 2,
+	.writeback_line_buffer_buffer_size = 0,
+	.writeback_min_hscl_ratio = 1,
+	.writeback_min_vscl_ratio = 1,
+	.writeback_max_hscl_ratio = 1,
+	.writeback_max_vscl_ratio = 1,
+	.writeback_max_hscl_taps = 1,
+	.writeback_max_vscl_taps = 1,
+	.dppclk_delay_subtotal = 47, /* changed from 46,*/
+	.dppclk_delay_scl = 50,
+	.dppclk_delay_scl_lb_only = 16,
+	.dppclk_delay_cnvc_formatter = 28,/*changed from 27,*/
+	.dppclk_delay_cnvc_cursor = 6,
+	.dispclk_delay_subtotal = 125, /*changed from 119,*/
+	.dynamic_metadata_vm_enabled = false,
+	.odm_combine_4to1_supported = false,
+	.dcc_supported = true,
+//	.config_return_buffer_segment_size_in_kbytes = 64;/*required, hard coded in dml2_translate_ip_params*/
+
+};
+
+struct _vcs_dpi_soc_bounding_box_st dcn3_51_soc = {
+		/*TODO: correct dispclk/dppclk voltage level determination*/
+	.clock_limits = {
+		{
+			.state = 0,
+			.dcfclk_mhz = 400.0,
+			.fabricclk_mhz = 400.0,
+			.socclk_mhz = 600.0,
+			.dram_speed_mts = 3200.0,
+			.dispclk_mhz = 600.0,
+			.dppclk_mhz = 600.0,
+			.phyclk_mhz = 600.0,
+			.phyclk_d18_mhz = 667.0,
+			.dscclk_mhz = 200.0,
+			.dtbclk_mhz = 600.0,
+		},
+		{
+			.state = 1,
+			.dcfclk_mhz = 600.0,
+			.fabricclk_mhz = 1000.0,
+			.socclk_mhz = 733.0,
+			.dram_speed_mts = 6400.0,
+			.dispclk_mhz = 800.0,
+			.dppclk_mhz = 800.0,
+			.phyclk_mhz = 810.0,
+			.phyclk_d18_mhz = 667.0,
+			.dscclk_mhz = 266.7,
+			.dtbclk_mhz = 600.0,
+		},
+		{
+			.state = 2,
+			.dcfclk_mhz = 738.0,
+			.fabricclk_mhz = 1200.0,
+			.socclk_mhz = 880.0,
+			.dram_speed_mts = 7500.0,
+			.dispclk_mhz = 800.0,
+			.dppclk_mhz = 800.0,
+			.phyclk_mhz = 810.0,
+			.phyclk_d18_mhz = 667.0,
+			.dscclk_mhz = 266.7,
+			.dtbclk_mhz = 600.0,
+		},
+		{
+			.state = 3,
+			.dcfclk_mhz = 800.0,
+			.fabricclk_mhz = 1400.0,
+			.socclk_mhz = 978.0,
+			.dram_speed_mts = 7500.0,
+			.dispclk_mhz = 960.0,
+			.dppclk_mhz = 960.0,
+			.phyclk_mhz = 810.0,
+			.phyclk_d18_mhz = 667.0,
+			.dscclk_mhz = 320.0,
+			.dtbclk_mhz = 600.0,
+		},
+		{
+			.state = 4,
+			.dcfclk_mhz = 873.0,
+			.fabricclk_mhz = 1600.0,
+			.socclk_mhz = 1100.0,
+			.dram_speed_mts = 8533.0,
+			.dispclk_mhz = 1066.7,
+			.dppclk_mhz = 1066.7,
+			.phyclk_mhz = 810.0,
+			.phyclk_d18_mhz = 667.0,
+			.dscclk_mhz = 355.6,
+			.dtbclk_mhz = 600.0,
+		},
+		{
+			.state = 5,
+			.dcfclk_mhz = 960.0,
+			.fabricclk_mhz = 1700.0,
+			.socclk_mhz = 1257.0,
+			.dram_speed_mts = 8533.0,
+			.dispclk_mhz = 1200.0,
+			.dppclk_mhz = 1200.0,
+			.phyclk_mhz = 810.0,
+			.phyclk_d18_mhz = 667.0,
+			.dscclk_mhz = 400.0,
+			.dtbclk_mhz = 600.0,
+		},
+		{
+			.state = 6,
+			.dcfclk_mhz = 1067.0,
+			.fabricclk_mhz = 1850.0,
+			.socclk_mhz = 1257.0,
+			.dram_speed_mts = 8533.0,
+			.dispclk_mhz = 1371.4,
+			.dppclk_mhz = 1371.4,
+			.phyclk_mhz = 810.0,
+			.phyclk_d18_mhz = 667.0,
+			.dscclk_mhz = 457.1,
+			.dtbclk_mhz = 600.0,
+		},
+		{
+			.state = 7,
+			.dcfclk_mhz = 1200.0,
+			.fabricclk_mhz = 2000.0,
+			.socclk_mhz = 1467.0,
+			.dram_speed_mts = 8533.0,
+			.dispclk_mhz = 1600.0,
+			.dppclk_mhz = 1600.0,
+			.phyclk_mhz = 810.0,
+			.phyclk_d18_mhz = 667.0,
+			.dscclk_mhz = 533.3,
+			.dtbclk_mhz = 600.0,
+		},
+	},
+	.num_states = 8,
+	.sr_exit_time_us = 28.0,
+	.sr_enter_plus_exit_time_us = 30.0,
+	.sr_exit_z8_time_us = 263.0,
+	.sr_enter_plus_exit_z8_time_us = 363.0,
+	.fclk_change_latency_us = 24.0,
+	.usr_retraining_latency_us = 2,
+	.writeback_latency_us = 12.0,
+
+	.dram_channel_width_bytes = 4,/*not exist in dml2*/
+	.round_trip_ping_latency_dcfclk_cycles = 106,/*not exist in dml2*/
+	.urgent_latency_pixel_data_only_us = 4.0,
+	.urgent_latency_pixel_mixed_with_vm_data_us = 4.0,
+	.urgent_latency_vm_data_only_us = 4.0,
+	.dram_clock_change_latency_us = 34,
+	.urgent_out_of_order_return_per_channel_pixel_only_bytes = 4096,
+	.urgent_out_of_order_return_per_channel_pixel_and_vm_bytes = 4096,
+	.urgent_out_of_order_return_per_channel_vm_only_bytes = 4096,
+
+	.pct_ideal_sdp_bw_after_urgent = 80.0,
+	.pct_ideal_fabric_bw_after_urgent = 80.0, /*new to dml2*/
+	.pct_ideal_dram_sdp_bw_after_urgent_pixel_only = 65.0,
+	.pct_ideal_dram_sdp_bw_after_urgent_pixel_and_vm = 60.0,
+	.pct_ideal_dram_sdp_bw_after_urgent_vm_only = 30.0,
+	.max_avg_sdp_bw_use_normal_percent = 60.0,
+	.max_avg_dram_bw_use_normal_percent = 60.0,
+	.fabric_datapath_to_dcn_data_return_bytes = 32,
+	.return_bus_width_bytes = 64,
+	.downspread_percent = 0.38,
+	.dcn_downspread_percent = 0.5,
+	.gpuvm_min_page_size_bytes = 4096,
+	.hostvm_min_page_size_bytes = 4096,
+	.do_urgent_latency_adjustment = 0,
+	.urgent_latency_adjustment_fabric_clock_component_us = 0,
+	.urgent_latency_adjustment_fabric_clock_reference_mhz = 0,
+	.num_chans = 4,
+	.dispclk_dppclk_vco_speed_mhz = 2400.0,
+};
+
+/*
+ * dcn351_update_bw_bounding_box
+ *
+ * This would override some dcn3_51 ip_or_soc initial parameters hardcoded from
+ * spreadsheet with actual values as per dGPU SKU:
+ * - with passed few options from dc->config
+ * - with dentist_vco_frequency from Clk Mgr (currently hardcoded, but might
+ *   need to get it from PM FW)
+ * - with passed latency values (passed in ns units) in dc-> bb override for
+ *   debugging purposes
+ * - with passed latencies from VBIOS (in 100_ns units) if available for
+ *   certain dGPU SKU
+ * - with number of DRAM channels from VBIOS (which differ for certain dGPU SKU
+ *   of the same ASIC)
+ * - clocks levels with passed clk_table entries from Clk Mgr as reported by PM
+ *   FW for different clocks (which might differ for certain dGPU SKU of the
+ *   same ASIC)
+ */
+void dcn351_update_bw_bounding_box_fpu(struct dc *dc,
+				      struct clk_bw_params *bw_params)
+{
+	unsigned int i, closest_clk_lvl;
+	int j;
+	struct clk_limit_table *clk_table = &bw_params->clk_table;
+	struct _vcs_dpi_voltage_scaling_st *clock_limits =
+		dc->scratch.update_bw_bounding_box.clock_limits;
+	int max_dispclk_mhz = 0, max_dppclk_mhz = 0;
+
+	dc_assert_fp_enabled();
+
+	dcn3_51_ip.max_num_otg =
+		dc->res_pool->res_cap->num_timing_generator;
+	dcn3_51_ip.max_num_dpp = dc->res_pool->pipe_count;
+	dcn3_51_soc.num_chans = bw_params->num_channels;
+
+	ASSERT(clk_table->num_entries);
+
+	/* Prepass to find max clocks independent of voltage level. */
+	for (i = 0; i < clk_table->num_entries; ++i) {
+		if (clk_table->entries[i].dispclk_mhz > max_dispclk_mhz)
+			max_dispclk_mhz = clk_table->entries[i].dispclk_mhz;
+		if (clk_table->entries[i].dppclk_mhz > max_dppclk_mhz)
+			max_dppclk_mhz = clk_table->entries[i].dppclk_mhz;
+	}
+
+	for (i = 0; i < clk_table->num_entries; i++) {
+		/* loop backwards*/
+		for (closest_clk_lvl = 0, j = dcn3_51_soc.num_states - 1;
+			j >= 0; j--) {
+			if (dcn3_51_soc.clock_limits[j].dcfclk_mhz <=
+				clk_table->entries[i].dcfclk_mhz) {
+				closest_clk_lvl = j;
+				break;
+			}
+		}
+		if (clk_table->num_entries == 1) {
+			/*smu gives one DPM level, let's take the highest one*/
+			closest_clk_lvl = dcn3_51_soc.num_states - 1;
+		}
+
+		clock_limits[i].state = i;
+
+		/* Clocks dependent on voltage level. */
+		clock_limits[i].dcfclk_mhz = clk_table->entries[i].dcfclk_mhz;
+		if (clk_table->num_entries == 1 &&
+			clock_limits[i].dcfclk_mhz <
+			dcn3_51_soc.clock_limits[closest_clk_lvl].dcfclk_mhz) {
+			/*SMU fix not released yet*/
+			clock_limits[i].dcfclk_mhz =
+				dcn3_51_soc.clock_limits[closest_clk_lvl].dcfclk_mhz;
+		}
+
+		clock_limits[i].fabricclk_mhz =
+			clk_table->entries[i].fclk_mhz;
+		clock_limits[i].socclk_mhz =
+			clk_table->entries[i].socclk_mhz;
+
+		if (clk_table->entries[i].memclk_mhz &&
+			clk_table->entries[i].wck_ratio)
+			clock_limits[i].dram_speed_mts =
+				clk_table->entries[i].memclk_mhz * 2 *
+				clk_table->entries[i].wck_ratio;
+
+		/* Clocks independent of voltage level. */
+		clock_limits[i].dispclk_mhz = max_dispclk_mhz ?
+			max_dispclk_mhz :
+			dcn3_51_soc.clock_limits[closest_clk_lvl].dispclk_mhz;
+
+		clock_limits[i].dppclk_mhz = max_dppclk_mhz ?
+			max_dppclk_mhz :
+			dcn3_51_soc.clock_limits[closest_clk_lvl].dppclk_mhz;
+
+		clock_limits[i].dram_bw_per_chan_gbps =
+			dcn3_51_soc.clock_limits[closest_clk_lvl].dram_bw_per_chan_gbps;
+		clock_limits[i].dscclk_mhz =
+			dcn3_51_soc.clock_limits[closest_clk_lvl].dscclk_mhz;
+		clock_limits[i].dtbclk_mhz =
+			dcn3_51_soc.clock_limits[closest_clk_lvl].dtbclk_mhz;
+		clock_limits[i].phyclk_d18_mhz =
+			dcn3_51_soc.clock_limits[closest_clk_lvl].phyclk_d18_mhz;
+		clock_limits[i].phyclk_mhz =
+			dcn3_51_soc.clock_limits[closest_clk_lvl].phyclk_mhz;
+	}
+
+	memcpy(dcn3_51_soc.clock_limits, clock_limits,
+		sizeof(dcn3_51_soc.clock_limits));
+
+	if (clk_table->num_entries)
+		dcn3_51_soc.num_states = clk_table->num_entries;
+
+	if (max_dispclk_mhz) {
+		dcn3_51_soc.dispclk_dppclk_vco_speed_mhz = max_dispclk_mhz * 2;
+		dc->dml.soc.dispclk_dppclk_vco_speed_mhz = max_dispclk_mhz * 2;
+	}
+	if ((int)(dcn3_51_soc.dram_clock_change_latency_us * 1000)
+				!= dc->debug.dram_clock_change_latency_ns
+			&& dc->debug.dram_clock_change_latency_ns) {
+		dcn3_51_soc.dram_clock_change_latency_us =
+			dc->debug.dram_clock_change_latency_ns / 1000.0;
+	}
+
+	if (dc->bb_overrides.dram_clock_change_latency_ns > 0)
+		dcn3_51_soc.dram_clock_change_latency_us =
+			dc->bb_overrides.dram_clock_change_latency_ns / 1000.0;
+
+	if (dc->bb_overrides.sr_exit_time_ns > 0)
+		dcn3_51_soc.sr_exit_time_us = dc->bb_overrides.sr_exit_time_ns / 1000.0;
+
+	if (dc->bb_overrides.sr_enter_plus_exit_time_ns > 0)
+		dcn3_51_soc.sr_enter_plus_exit_time_us =
+			dc->bb_overrides.sr_enter_plus_exit_time_ns / 1000.0;
+
+	if (dc->bb_overrides.sr_exit_z8_time_ns > 0)
+		dcn3_51_soc.sr_exit_z8_time_us = dc->bb_overrides.sr_exit_z8_time_ns / 1000.0;
+
+	if (dc->bb_overrides.sr_enter_plus_exit_z8_time_ns > 0)
+		dcn3_51_soc.sr_enter_plus_exit_z8_time_us =
+			dc->bb_overrides.sr_enter_plus_exit_z8_time_ns / 1000.0;
+
+	/*temp till dml2 fully work without dml1*/
+	dml_init_instance(&dc->dml, &dcn3_51_soc, &dcn3_51_ip,
+			  DML_PROJECT_DCN31);
+
+	/*copy to dml2, before dml2_create*/
+	if (clk_table->num_entries > 2) {
+
+		for (i = 0; i < clk_table->num_entries; i++) {
+			dc->dml2_options.bbox_overrides.clks_table.num_states =
+				clk_table->num_entries;
+			dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].dcfclk_mhz =
+				clock_limits[i].dcfclk_mhz;
+			dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].fclk_mhz =
+				clock_limits[i].fabricclk_mhz;
+			dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].dispclk_mhz =
+				clock_limits[i].dispclk_mhz;
+			dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].dppclk_mhz =
+				clock_limits[i].dppclk_mhz;
+			dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].socclk_mhz =
+				clock_limits[i].socclk_mhz;
+			dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].memclk_mhz =
+				clk_table->entries[i].memclk_mhz * clk_table->entries[i].wck_ratio;
+			dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].dram_speed_mts = clock_limits[i].dram_speed_mts;
+			dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].dtbclk_mhz =
+				clock_limits[i].dtbclk_mhz;
+			dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_dcfclk_levels =
+				clk_table->num_entries;
+			dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_fclk_levels =
+				clk_table->num_entries;
+			dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_dispclk_levels =
+				clk_table->num_entries;
+			dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_dppclk_levels =
+				clk_table->num_entries;
+			dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_socclk_levels =
+				clk_table->num_entries;
+			dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_memclk_levels =
+				clk_table->num_entries;
+			dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_dtbclk_levels =
+				clk_table->num_entries;
+		}
+	}
+
+	/* Update latency values */
+	dc->dml2_options.bbox_overrides.dram_clock_change_latency_us = dcn3_51_soc.dram_clock_change_latency_us;
+
+	dc->dml2_options.bbox_overrides.sr_exit_latency_us = dcn3_51_soc.sr_exit_time_us;
+	dc->dml2_options.bbox_overrides.sr_enter_plus_exit_latency_us = dcn3_51_soc.sr_enter_plus_exit_time_us;
+
+	dc->dml2_options.bbox_overrides.sr_exit_z8_time_us = dcn3_51_soc.sr_exit_z8_time_us;
+	dc->dml2_options.bbox_overrides.sr_enter_plus_exit_z8_time_us = dcn3_51_soc.sr_enter_plus_exit_z8_time_us;
+}
+
+static bool is_dual_plane(enum surface_pixel_format format)
+{
+	return format >= SURFACE_PIXEL_FORMAT_VIDEO_BEGIN ||
+		format == SURFACE_PIXEL_FORMAT_GRPH_RGBE_ALPHA;
+}
+
+/*
+ * micro_sec_to_vert_lines () - converts time to number of vertical lines for a given timing
+ *
+ * @param: num_us: number of microseconds
+ * @return: number of vertical lines. If exact number of vertical lines is not found then
+ *          it will round up to next number of lines to guarantee num_us
+ */
+static unsigned int micro_sec_to_vert_lines(unsigned int num_us, struct dc_crtc_timing *timing)
+{
+	unsigned int num_lines = 0;
+	unsigned int lines_time_in_ns = 1000.0 *
+			(((float)timing->h_total * 1000.0) /
+			 ((float)timing->pix_clk_100hz / 10.0));
+
+	num_lines = dml_ceil(1000.0 * num_us / lines_time_in_ns, 1.0);
+
+	return num_lines;
+}
+
+static unsigned int get_vertical_back_porch(struct dc_crtc_timing *timing)
+{
+	unsigned int v_active = 0, v_blank = 0, v_back_porch = 0;
+
+	v_active = timing->v_border_top + timing->v_addressable + timing->v_border_bottom;
+	v_blank = timing->v_total - v_active;
+	v_back_porch = v_blank - timing->v_front_porch - timing->v_sync_width;
+
+	return v_back_porch;
+}
+
+int dcn351_populate_dml_pipes_from_context_fpu(struct dc *dc,
+					      struct dc_state *context,
+					      display_e2e_pipe_params_st *pipes,
+					      enum dc_validate_mode validate_mode)
+{
+	int i, pipe_cnt;
+	struct resource_context *res_ctx = &context->res_ctx;
+	struct pipe_ctx *pipe = 0;
+	bool upscaled = false;
+	const unsigned int max_allowed_vblank_nom = 1023;
+
+	dc_assert_fp_enabled();
+
+	dcn31_populate_dml_pipes_from_context(dc, context, pipes,
+					      validate_mode);
+
+	for (i = 0, pipe_cnt = 0; i < dc->res_pool->pipe_count; i++) {
+		struct dc_crtc_timing *timing;
+		unsigned int num_lines = 0;
+		unsigned int v_back_porch = 0;
+
+		if (!res_ctx->pipe_ctx[i].stream)
+			continue;
+
+		pipe = &res_ctx->pipe_ctx[i];
+		timing = &pipe->stream->timing;
+
+		num_lines = micro_sec_to_vert_lines(dcn3_51_ip.VBlankNomDefaultUS, timing);
+		v_back_porch  = get_vertical_back_porch(timing);
+
+		if (pipe->stream->adjust.v_total_max ==
+		    pipe->stream->adjust.v_total_min &&
+		    pipe->stream->adjust.v_total_min > timing->v_total) {
+			pipes[pipe_cnt].pipe.dest.vtotal =
+				pipe->stream->adjust.v_total_min;
+			pipes[pipe_cnt].pipe.dest.vblank_nom = timing->v_total -
+				pipes[pipe_cnt].pipe.dest.vactive;
+		}
+
+		pipes[pipe_cnt].pipe.dest.vblank_nom = timing->v_total - pipes[pipe_cnt].pipe.dest.vactive;
+		pipes[pipe_cnt].pipe.dest.vblank_nom = min(pipes[pipe_cnt].pipe.dest.vblank_nom, num_lines);
+		// vblank_nom should not smaller than (VSync (timing->v_sync_width + v_back_porch) + 2)
+		// + 2 is because
+		// 1 -> VStartup_start should be 1 line before VSync
+		// 1 -> always reserve 1 line between start of vblank to vstartup signal
+		pipes[pipe_cnt].pipe.dest.vblank_nom =
+			max(pipes[pipe_cnt].pipe.dest.vblank_nom, timing->v_sync_width + v_back_porch + 2);
+		pipes[pipe_cnt].pipe.dest.vblank_nom = min(pipes[pipe_cnt].pipe.dest.vblank_nom, max_allowed_vblank_nom);
+
+		if (pipe->plane_state &&
+		    (pipe->plane_state->src_rect.height <
+		     pipe->plane_state->dst_rect.height ||
+		     pipe->plane_state->src_rect.width <
+		     pipe->plane_state->dst_rect.width))
+			upscaled = true;
+
+		/*
+		 * Immediate flip can be set dynamically after enabling the
+		 * plane. We need to require support for immediate flip or
+		 * underflow can be intermittently experienced depending on peak
+		 * b/w requirements.
+		 */
+		pipes[pipe_cnt].pipe.src.immediate_flip = true;
+
+		pipes[pipe_cnt].pipe.src.unbounded_req_mode = false;
+
+		dcn31_zero_pipe_dcc_fraction(pipes, pipe_cnt);
+
+		pipes[pipe_cnt].pipe.dest.vfront_porch = timing->v_front_porch;
+		pipes[pipe_cnt].pipe.src.dcc_rate = 3;
+		pipes[pipe_cnt].dout.dsc_input_bpc = 0;
+		pipes[pipe_cnt].pipe.src.gpuvm_min_page_size_kbytes = 256;
+
+		if (pipes[pipe_cnt].dout.dsc_enable) {
+			switch (timing->display_color_depth) {
+			case COLOR_DEPTH_888:
+				pipes[pipe_cnt].dout.dsc_input_bpc = 8;
+				break;
+			case COLOR_DEPTH_101010:
+				pipes[pipe_cnt].dout.dsc_input_bpc = 10;
+				break;
+			case COLOR_DEPTH_121212:
+				pipes[pipe_cnt].dout.dsc_input_bpc = 12;
+				break;
+			default:
+				ASSERT(0);
+				break;
+			}
+		}
+
+		pipe_cnt++;
+	}
+
+	context->bw_ctx.dml.ip.det_buffer_size_kbytes = 384;/*per guide*/
+	dc->config.enable_4to1MPC = false;
+
+	if (pipe_cnt == 1 && pipe->plane_state && !dc->debug.disable_z9_mpc) {
+		if (is_dual_plane(pipe->plane_state->format)
+				&& pipe->plane_state->src_rect.width <= 1920 &&
+				pipe->plane_state->src_rect.height <= 1080) {
+			dc->config.enable_4to1MPC = true;
+		} else if (!is_dual_plane(pipe->plane_state->format) &&
+			   pipe->plane_state->src_rect.width <= 5120) {
+			/*
+			 * Limit to 5k max to avoid forced pipe split when there
+			 * is not enough detile for swath
+			 */
+			context->bw_ctx.dml.ip.det_buffer_size_kbytes = 192;
+			pipes[0].pipe.src.unbounded_req_mode = true;
+		}
+	} else if (context->stream_count >=
+		   dc->debug.crb_alloc_policy_min_disp_count &&
+		   dc->debug.crb_alloc_policy > DET_SIZE_DEFAULT) {
+		context->bw_ctx.dml.ip.det_buffer_size_kbytes =
+			dc->debug.crb_alloc_policy * 64;
+	} else if (context->stream_count >= 3 && upscaled) {
+		context->bw_ctx.dml.ip.det_buffer_size_kbytes = 192;
+	}
+
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
+
+		if (!pipe->stream)
+			continue;
+
+		if (pipe->stream->signal == SIGNAL_TYPE_EDP &&
+		    dc->debug.seamless_boot_odm_combine &&
+		    pipe->stream->apply_seamless_boot_optimization) {
+
+			if (pipe->stream->apply_boot_odm_mode ==
+			    dm_odm_combine_policy_2to1) {
+				context->bw_ctx.dml.vba.ODMCombinePolicy =
+					dm_odm_combine_policy_2to1;
+				break;
+			}
+		}
+	}
+
+	return pipe_cnt;
+}
+
+void dcn351_decide_zstate_support(struct dc *dc, struct dc_state *context)
+{
+	enum dcn_zstate_support_state support = DCN_ZSTATE_SUPPORT_DISALLOW;
+	unsigned int i, plane_count = 0;
+
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		if (context->res_ctx.pipe_ctx[i].plane_state)
+			plane_count++;
+	}
+
+	/*dcn351 does not support z9/z10*/
+	if (context->stream_count == 0 || plane_count == 0) {
+		support = DCN_ZSTATE_SUPPORT_ALLOW_Z8_ONLY;
+	} else if (context->stream_count == 1 && context->streams[0]->signal == SIGNAL_TYPE_EDP) {
+		struct dc_link *link = context->streams[0]->sink->link;
+		bool is_pwrseq0 = link && link->link_index == 0;
+		bool is_psr = (link && (link->psr_settings.psr_version == DC_PSR_VERSION_1 ||
+								link->psr_settings.psr_version == DC_PSR_VERSION_SU_1) && !link->panel_config.psr.disable_psr);
+		bool is_replay = link && link->replay_settings.replay_feature_enabled;
+		int minmum_z8_residency =
+			dc->debug.minimum_z8_residency_time > 0 ? dc->debug.minimum_z8_residency_time : 1000;
+		bool allow_z8 = context->bw_ctx.dml.vba.StutterPeriod > (double)minmum_z8_residency;
+
+		/*for psr1/psr-su, we allow z8 and z10 based on latency, for replay with IPS enabled, it will enter ips2*/
+		if (is_pwrseq0 && (is_psr || is_replay))
+			support = allow_z8 ? allow_z8 : DCN_ZSTATE_SUPPORT_DISALLOW;
+	}
+	context->bw_ctx.bw.dcn.clk.zstate_support = support;
+}
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn351/dcn351_fpu.h b/drivers/gpu/drm/amd/display/dc/dml/dcn351/dcn351_fpu.h
new file mode 100644
index 000000000000..f71d9d8d0759
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn351/dcn351_fpu.h
@@ -0,0 +1,19 @@
+/* SPDX-License-Identifier: MIT */
+/* Copyright 2024 Advanced Micro Devices, Inc. */
+
+#ifndef __DCN351_FPU_H__
+#define __DCN351_FPU_H__
+
+#include "clk_mgr.h"
+
+void dcn351_update_bw_bounding_box_fpu(struct dc *dc,
+				      struct clk_bw_params *bw_params);
+
+int dcn351_populate_dml_pipes_from_context_fpu(struct dc *dc,
+					      struct dc_state *context,
+					      display_e2e_pipe_params_st *pipes,
+					      enum dc_validate_mode validate_mode);
+
+void dcn351_decide_zstate_support(struct dc *dc, struct dc_state *context);
+
+#endif
diff --git a/drivers/gpu/drm/amd/display/dc/dml/display_mode_enums.h b/drivers/gpu/drm/amd/display/dc/dml/display_mode_enums.h
index edb9f7567d6d..d5831a34f5a1 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/display_mode_enums.h
+++ b/drivers/gpu/drm/amd/display/dc/dml/display_mode_enums.h
@@ -26,7 +26,11 @@
 #define __DISPLAY_MODE_ENUMS_H__
 
 enum output_encoder_class {
-	dm_dp = 0, dm_hdmi = 1, dm_wb = 2, dm_edp
+	dm_dp = 0,
+	dm_hdmi = 1,
+	dm_wb = 2,
+	dm_edp = 3,
+	dm_dp2p0 = 5,
 };
 enum output_format_class {
 	dm_444 = 0, dm_420 = 1, dm_n422, dm_s422
@@ -101,10 +105,39 @@ enum source_macro_tile_size {
 enum cursor_bpp {
 	dm_cur_2bit = 0, dm_cur_32bit = 1, dm_cur_64bit = 2
 };
+
+/**
+ * @enum clock_change_support - It represents possible reasons to change the DRAM clock.
+ *
+ * DC may change the DRAM clock during its execution, and this enum tracks all
+ * the available methods. Note that every ASIC has their specific way to deal
+ * with these clock switch.
+ */
 enum clock_change_support {
+	/**
+	 * @dm_dram_clock_change_uninitialized: If you see this, we might have
+	 * a code initialization issue
+	 */
 	dm_dram_clock_change_uninitialized = 0,
+
+	/**
+	 * @dm_dram_clock_change_vactive: Support DRAM switch in VActive
+	 */
 	dm_dram_clock_change_vactive,
+
+	/**
+	 * @dm_dram_clock_change_vblank: Support DRAM switch in VBlank
+	 */
 	dm_dram_clock_change_vblank,
+
+	dm_dram_clock_change_vactive_w_mall_full_frame,
+	dm_dram_clock_change_vactive_w_mall_sub_vp,
+	dm_dram_clock_change_vblank_w_mall_full_frame,
+	dm_dram_clock_change_vblank_w_mall_sub_vp,
+
+	/**
+	 * @dm_dram_clock_change_unsupported: Do not support DRAM switch
+	 */
 	dm_dram_clock_change_unsupported
 };
 
@@ -157,6 +190,14 @@ enum dm_validation_status {
 	DML_FAIL_DSC_INPUT_BPC,
 	DML_FAIL_PREFETCH_SUPPORT,
 	DML_FAIL_V_RATIO_PREFETCH,
+	DML_FAIL_P2I_WITH_420,
+	DML_FAIL_DSC_ONLY_IF_NECESSARY_WITH_BPP,
+	DML_FAIL_NOT_DSC422_NATIVE,
+	DML_FAIL_ODM_COMBINE4TO1,
+	DML_FAIL_ENOUGH_WRITEBACK_UNITS,
+	DML_FAIL_VIEWPORT_EXCEEDS_SURFACE,
+	DML_FAIL_DYNAMIC_METADATA,
+	DML_FAIL_FMT_BUFFER_EXCEEDED,
 };
 
 enum writeback_config {
@@ -169,6 +210,9 @@ enum odm_combine_mode {
 	dm_odm_combine_mode_disabled,
 	dm_odm_combine_mode_2to1,
 	dm_odm_combine_mode_4to1,
+	dm_odm_split_mode_1to2,
+	dm_odm_mode_mso_1to2,
+	dm_odm_mode_mso_1to4
 };
 
 enum odm_combine_policy {
@@ -176,11 +220,15 @@ enum odm_combine_policy {
 	dm_odm_combine_policy_none,
 	dm_odm_combine_policy_2to1,
 	dm_odm_combine_policy_4to1,
+	dm_odm_split_policy_1to2,
+	dm_odm_mso_policy_1to2,
+	dm_odm_mso_policy_1to4,
 };
 
 enum immediate_flip_requirement {
 	dm_immediate_flip_not_required,
 	dm_immediate_flip_required,
+	dm_immediate_flip_opportunistic,
 };
 
 enum unbounded_requesting_policy {
@@ -189,4 +237,75 @@ enum unbounded_requesting_policy {
 	dm_unbounded_requesting_disable
 };
 
+enum dm_rotation_angle {
+	dm_rotation_0,
+	dm_rotation_90,
+	dm_rotation_180,
+	dm_rotation_270,
+	dm_rotation_0m,
+	dm_rotation_90m,
+	dm_rotation_180m,
+	dm_rotation_270m,
+};
+
+enum dm_use_mall_for_pstate_change_mode {
+	dm_use_mall_pstate_change_disable,
+	dm_use_mall_pstate_change_full_frame,
+	dm_use_mall_pstate_change_sub_viewport,
+	dm_use_mall_pstate_change_phantom_pipe
+};
+
+enum dm_use_mall_for_static_screen_mode {
+	dm_use_mall_static_screen_disable,
+	dm_use_mall_static_screen_optimize,
+	dm_use_mall_static_screen_enable,
+};
+
+enum dm_output_link_dp_rate {
+	dm_dp_rate_na,
+	dm_dp_rate_hbr,
+	dm_dp_rate_hbr2,
+	dm_dp_rate_hbr3,
+	dm_dp_rate_uhbr10,
+	dm_dp_rate_uhbr13p5,
+	dm_dp_rate_uhbr20,
+};
+
+enum dm_fclock_change_support {
+	dm_fclock_change_vactive,
+	dm_fclock_change_vblank,
+	dm_fclock_change_unsupported,
+};
+
+enum dm_prefetch_modes {
+	dm_prefetch_support_uclk_fclk_and_stutter_if_possible,
+	dm_prefetch_support_uclk_fclk_and_stutter,
+	dm_prefetch_support_fclk_and_stutter,
+	dm_prefetch_support_stutter,
+	dm_prefetch_support_none,
+};
+enum dm_output_type {
+	dm_output_type_unknown,
+	dm_output_type_dp,
+	dm_output_type_edp,
+	dm_output_type_dp2p0,
+	dm_output_type_hdmi,
+	dm_output_type_hdmifrl,
+};
+
+enum dm_output_rate {
+	dm_output_rate_unknown,
+	dm_output_rate_dp_rate_hbr,
+	dm_output_rate_dp_rate_hbr2,
+	dm_output_rate_dp_rate_hbr3,
+	dm_output_rate_dp_rate_uhbr10,
+	dm_output_rate_dp_rate_uhbr13p5,
+	dm_output_rate_dp_rate_uhbr20,
+	dm_output_rate_hdmi_rate_3x3,
+	dm_output_rate_hdmi_rate_6x3,
+	dm_output_rate_hdmi_rate_6x4,
+	dm_output_rate_hdmi_rate_8x4,
+	dm_output_rate_hdmi_rate_10x4,
+	dm_output_rate_hdmi_rate_12x4,
+};
 #endif
diff --git a/drivers/gpu/drm/amd/display/dc/dml/display_mode_lib.c b/drivers/gpu/drm/amd/display/dc/dml/display_mode_lib.c
index 30db51fbd8cd..da0cfbb071e6 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/display_mode_lib.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/display_mode_lib.c
@@ -35,43 +35,61 @@
 #include "dcn30/display_rq_dlg_calc_30.h"
 #include "dcn31/display_mode_vba_31.h"
 #include "dcn31/display_rq_dlg_calc_31.h"
+#include "dcn314/display_mode_vba_314.h"
+#include "dcn314/display_rq_dlg_calc_314.h"
+#include "dcn32/display_mode_vba_32.h"
+#include "dcn32/display_rq_dlg_calc_32.h"
 #include "dml_logger.h"
 
-const struct dml_funcs dml20_funcs = {
+static const struct dml_funcs dml20_funcs = {
 	.validate = dml20_ModeSupportAndSystemConfigurationFull,
 	.recalculate = dml20_recalculate,
 	.rq_dlg_get_dlg_reg = dml20_rq_dlg_get_dlg_reg,
 	.rq_dlg_get_rq_reg = dml20_rq_dlg_get_rq_reg
 };
 
-const struct dml_funcs dml20v2_funcs = {
+static const struct dml_funcs dml20v2_funcs = {
 	.validate = dml20v2_ModeSupportAndSystemConfigurationFull,
 	.recalculate = dml20v2_recalculate,
 	.rq_dlg_get_dlg_reg = dml20v2_rq_dlg_get_dlg_reg,
 	.rq_dlg_get_rq_reg = dml20v2_rq_dlg_get_rq_reg
 };
 
-const struct dml_funcs dml21_funcs = {
-        .validate = dml21_ModeSupportAndSystemConfigurationFull,
-        .recalculate = dml21_recalculate,
-        .rq_dlg_get_dlg_reg = dml21_rq_dlg_get_dlg_reg,
-        .rq_dlg_get_rq_reg = dml21_rq_dlg_get_rq_reg
+static const struct dml_funcs dml21_funcs = {
+	.validate = dml21_ModeSupportAndSystemConfigurationFull,
+	.recalculate = dml21_recalculate,
+	.rq_dlg_get_dlg_reg = dml21_rq_dlg_get_dlg_reg,
+	.rq_dlg_get_rq_reg = dml21_rq_dlg_get_rq_reg
 };
 
-const struct dml_funcs dml30_funcs = {
+static const struct dml_funcs dml30_funcs = {
 	.validate = dml30_ModeSupportAndSystemConfigurationFull,
 	.recalculate = dml30_recalculate,
 	.rq_dlg_get_dlg_reg = dml30_rq_dlg_get_dlg_reg,
 	.rq_dlg_get_rq_reg = dml30_rq_dlg_get_rq_reg
 };
 
-const struct dml_funcs dml31_funcs = {
+static const struct dml_funcs dml31_funcs = {
 	.validate = dml31_ModeSupportAndSystemConfigurationFull,
 	.recalculate = dml31_recalculate,
 	.rq_dlg_get_dlg_reg = dml31_rq_dlg_get_dlg_reg,
 	.rq_dlg_get_rq_reg = dml31_rq_dlg_get_rq_reg
 };
 
+static const struct dml_funcs dml314_funcs = {
+	.validate = dml314_ModeSupportAndSystemConfigurationFull,
+	.recalculate = dml314_recalculate,
+	.rq_dlg_get_dlg_reg = dml314_rq_dlg_get_dlg_reg,
+	.rq_dlg_get_rq_reg = dml314_rq_dlg_get_rq_reg
+};
+
+static const struct dml_funcs dml32_funcs = {
+	.validate = dml32_ModeSupportAndSystemConfigurationFull,
+	.recalculate = dml32_recalculate,
+	.rq_dlg_get_dlg_reg_v2 = dml32_rq_dlg_get_dlg_reg,
+	.rq_dlg_get_rq_reg_v2 = dml32_rq_dlg_get_rq_reg
+};
+
 void dml_init_instance(struct display_mode_lib *lib,
 		const struct _vcs_dpi_soc_bounding_box_st *soc_bb,
 		const struct _vcs_dpi_ip_params_st *ip_params,
@@ -95,9 +113,15 @@ void dml_init_instance(struct display_mode_lib *lib,
 		lib->funcs = dml30_funcs;
 		break;
 	case DML_PROJECT_DCN31:
-	case DML_PROJECT_DCN31_FPGA:
+	case DML_PROJECT_DCN315:
 		lib->funcs = dml31_funcs;
 		break;
+	case DML_PROJECT_DCN314:
+		lib->funcs = dml314_funcs;
+		break;
+	case DML_PROJECT_DCN32:
+		lib->funcs = dml32_funcs;
+		break;
 
 	default:
 		break;
diff --git a/drivers/gpu/drm/amd/display/dc/dml/display_mode_lib.h b/drivers/gpu/drm/amd/display/dc/dml/display_mode_lib.h
index 6905ef1e75a6..5edf69fa40d1 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/display_mode_lib.h
+++ b/drivers/gpu/drm/amd/display/dc/dml/display_mode_lib.h
@@ -40,7 +40,9 @@ enum dml_project {
 	DML_PROJECT_DCN21,
 	DML_PROJECT_DCN30,
 	DML_PROJECT_DCN31,
-	DML_PROJECT_DCN31_FPGA,
+	DML_PROJECT_DCN315,
+	DML_PROJECT_DCN314,
+	DML_PROJECT_DCN32,
 };
 
 struct display_mode_lib;
@@ -62,6 +64,20 @@ struct dml_funcs {
 		struct display_mode_lib *mode_lib,
 		display_rq_regs_st *rq_regs,
 		const display_pipe_params_st *pipe_param);
+	// DLG interfaces have different function parameters in DCN32.
+	// Create new function pointers to address the changes
+	void (*rq_dlg_get_dlg_reg_v2)(
+			struct display_mode_lib *mode_lib,
+			display_dlg_regs_st *dlg_regs,
+			display_ttu_regs_st *ttu_regs,
+			display_e2e_pipe_params_st *e2e_pipe_param,
+			const unsigned int num_pipes,
+			const unsigned int pipe_idx);
+	void (*rq_dlg_get_rq_reg_v2)(display_rq_regs_st *rq_regs,
+			struct display_mode_lib *mode_lib,
+			const display_e2e_pipe_params_st *e2e_pipe_param,
+			const unsigned int num_pipes,
+			const unsigned int pipe_idx);
 	void (*recalculate)(struct display_mode_lib *mode_lib);
 	void (*validate)(struct display_mode_lib *mode_lib);
 };
@@ -73,6 +89,8 @@ struct display_mode_lib {
 	struct vba_vars_st vba;
 	struct dal_logger *logger;
 	struct dml_funcs funcs;
+	struct _vcs_dpi_display_e2e_pipe_params_st dml_pipe_state[6];
+	bool validate_max_state;
 };
 
 void dml_init_instance(struct display_mode_lib *lib,
diff --git a/drivers/gpu/drm/amd/display/dc/dml/display_mode_structs.h b/drivers/gpu/drm/amd/display/dc/dml/display_mode_structs.h
index d46a2733024c..641a8cd019cd 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/display_mode_structs.h
+++ b/drivers/gpu/drm/amd/display/dc/dml/display_mode_structs.h
@@ -26,6 +26,16 @@
 #include "dc_features.h"
 #include "display_mode_enums.h"
 
+/**
+ * DOC: overview
+ *
+ * Most of the DML code is automatically generated and tested via hardware
+ * description language. Usually, we use the reference _vcs_dpi in the code
+ * where VCS means "Verilog Compiled Simulator" and DPI stands for "Direct
+ * Programmer Interface". In other words, those structs can be used to
+ * interface with Verilog with other languages such as C.
+ */
+
 #ifndef __DISPLAY_MODE_STRUCTS_H__
 #define __DISPLAY_MODE_STRUCTS_H__
 
@@ -54,12 +64,102 @@ typedef struct _vcs_dpi_display_rq_regs_st display_rq_regs_st;
 typedef struct _vcs_dpi_display_dlg_sys_params_st display_dlg_sys_params_st;
 typedef struct _vcs_dpi_display_arb_params_st display_arb_params_st;
 
+typedef struct {
+	double UrgentWatermark;
+	double WritebackUrgentWatermark;
+	double DRAMClockChangeWatermark;
+	double FCLKChangeWatermark;
+	double WritebackDRAMClockChangeWatermark;
+	double WritebackFCLKChangeWatermark;
+	double StutterExitWatermark;
+	double StutterEnterPlusExitWatermark;
+	double Z8StutterExitWatermark;
+	double Z8StutterEnterPlusExitWatermark;
+	double USRRetrainingWatermark;
+} Watermarks;
+
+typedef struct {
+	double UrgentLatency;
+	double ExtraLatency;
+	double WritebackLatency;
+	double DRAMClockChangeLatency;
+	double FCLKChangeLatency;
+	double SRExitTime;
+	double SREnterPlusExitTime;
+	double SRExitZ8Time;
+	double SREnterPlusExitZ8Time;
+	double USRRetrainingLatencyPlusSMNLatency;
+} Latencies;
+
+typedef struct {
+	double Dppclk;
+	double Dispclk;
+	double PixelClock;
+	double DCFClkDeepSleep;
+	unsigned int DPPPerSurface;
+	bool ScalerEnabled;
+	enum dm_rotation_angle SourceRotation;
+	unsigned int ViewportHeight;
+	unsigned int ViewportHeightChroma;
+	unsigned int BlockWidth256BytesY;
+	unsigned int BlockHeight256BytesY;
+	unsigned int BlockWidth256BytesC;
+	unsigned int BlockHeight256BytesC;
+	unsigned int BlockWidthY;
+	unsigned int BlockHeightY;
+	unsigned int BlockWidthC;
+	unsigned int BlockHeightC;
+	unsigned int InterlaceEnable;
+	unsigned int NumberOfCursors;
+	unsigned int VBlank;
+	unsigned int HTotal;
+	unsigned int HActive;
+	bool DCCEnable;
+	enum odm_combine_mode ODMMode;
+	enum source_format_class SourcePixelFormat;
+	enum dm_swizzle_mode SurfaceTiling;
+	unsigned int BytePerPixelY;
+	unsigned int BytePerPixelC;
+	bool ProgressiveToInterlaceUnitInOPP;
+	double VRatio;
+	double VRatioChroma;
+	unsigned int VTaps;
+	unsigned int VTapsChroma;
+	unsigned int PitchY;
+	unsigned int DCCMetaPitchY;
+	unsigned int PitchC;
+	unsigned int DCCMetaPitchC;
+	bool ViewportStationary;
+	unsigned int ViewportXStart;
+	unsigned int ViewportYStart;
+	unsigned int ViewportXStartC;
+	unsigned int ViewportYStartC;
+	bool FORCE_ONE_ROW_FOR_FRAME;
+	unsigned int SwathHeightY;
+	unsigned int SwathHeightC;
+} DmlPipe;
+
+typedef struct {
+	double UrgentLatency;
+	double ExtraLatency;
+	double WritebackLatency;
+	double DRAMClockChangeLatency;
+	double FCLKChangeLatency;
+	double SRExitTime;
+	double SREnterPlusExitTime;
+	double SRExitZ8Time;
+	double SREnterPlusExitZ8Time;
+	double USRRetrainingLatency;
+	double SMNLatency;
+} SOCParametersList;
+
 struct _vcs_dpi_voltage_scaling_st {
 	int state;
 	double dscclk_mhz;
 	double dcfclk_mhz;
 	double socclk_mhz;
 	double phyclk_d18_mhz;
+	double phyclk_d32_mhz;
 	double dram_speed_mts;
 	double fabricclk_mhz;
 	double dispclk_mhz;
@@ -67,10 +167,23 @@ struct _vcs_dpi_voltage_scaling_st {
 	double phyclk_mhz;
 	double dppclk_mhz;
 	double dtbclk_mhz;
+	float net_bw_in_kbytes_sec;
 };
 
+/**
+ * _vcs_dpi_soc_bounding_box_st: SOC definitions
+ *
+ * This struct maintains the SOC Bounding Box information for the ASIC; it
+ * defines things such as clock, voltage, performance, etc. Usually, we load
+ * these values from VBIOS; if something goes wrong, we use some hard-coded
+ * values, which will enable the ASIC to light up with limitations.
+ */
 struct _vcs_dpi_soc_bounding_box_st {
 	struct _vcs_dpi_voltage_scaling_st clock_limits[DC__VOLTAGE_STATES];
+	/**
+	 * @num_states: It represents the total of Display Power Management
+	 * (DPM) supported by the specific ASIC.
+	 */
 	unsigned int num_states;
 	double sr_exit_time_us;
 	double sr_enter_plus_exit_time_us;
@@ -80,6 +193,16 @@ struct _vcs_dpi_soc_bounding_box_st {
 	double urgent_latency_pixel_data_only_us;
 	double urgent_latency_pixel_mixed_with_vm_data_us;
 	double urgent_latency_vm_data_only_us;
+	double usr_retraining_latency_us;
+	double smn_latency_us;
+	double fclk_change_latency_us;
+	double mall_allocated_for_dcn_mbytes;
+	double pct_ideal_fabric_bw_after_urgent;
+	double pct_ideal_dram_bw_after_urgent_strobe;
+	double max_avg_fabric_bw_use_normal_percent;
+	double max_avg_dram_bw_use_normal_strobe_percent;
+	enum dm_prefetch_modes allow_for_pstate_or_stutter_in_vblank_final;
+	bool dram_clock_change_requirement_final;
 	double writeback_latency_us;
 	double ideal_dram_bw_after_urgent_percent;
 	double pct_ideal_dram_sdp_bw_after_urgent_pixel_only; // PercentOfIdealDRAMFabricAndSDPPortBWReceivedAfterUrgLatencyPixelDataOnly
@@ -124,8 +247,17 @@ struct _vcs_dpi_soc_bounding_box_st {
 	bool disable_dram_clock_change_vactive_support;
 	bool allow_dram_clock_one_display_vactive;
 	enum self_refresh_affinity allow_dram_self_refresh_or_dram_clock_change_in_vblank;
+	double max_vratio_pre;
 };
 
+/**
+ * @_vcs_dpi_ip_params_st: IP configuraion for DCN blocks
+ *
+ * In this struct you can find the DCN configuration associated to the specific
+ * ASIC. For example, here we can save how many DPPs the ASIC is using and it
+ * is available.
+ *
+ */
 struct _vcs_dpi_ip_params_st {
 	bool use_min_dcfclk;
 	bool clamp_min_dcfclk;
@@ -141,12 +273,16 @@ struct _vcs_dpi_ip_params_st {
 	unsigned int odm_capable;
 	unsigned int rob_buffer_size_kbytes;
 	unsigned int det_buffer_size_kbytes;
+	unsigned int min_comp_buffer_size_kbytes;
 	unsigned int dpte_buffer_size_in_pte_reqs_luma;
 	unsigned int dpte_buffer_size_in_pte_reqs_chroma;
 	unsigned int pde_proc_buffer_size_64k_reqs;
 	unsigned int dpp_output_buffer_pixels;
 	unsigned int opp_output_buffer_lines;
 	unsigned int pixel_chunk_size_kbytes;
+	unsigned int alpha_pixel_chunk_size_kbytes;
+	unsigned int min_pixel_chunk_size_bytes;
+	unsigned int dcc_meta_buffer_size_bytes;
 	unsigned char pte_enable;
 	unsigned int pte_chunk_size_kbytes;
 	unsigned int meta_chunk_size_kbytes;
@@ -167,12 +303,16 @@ struct _vcs_dpi_ip_params_st {
 	double writeback_min_hscl_ratio;
 	double writeback_min_vscl_ratio;
 	unsigned int maximum_dsc_bits_per_component;
+	unsigned int maximum_pixels_per_line_per_dsc_unit;
 	unsigned int writeback_max_hscl_taps;
 	unsigned int writeback_max_vscl_taps;
 	unsigned int writeback_line_buffer_luma_buffer_size;
 	unsigned int writeback_line_buffer_chroma_buffer_size;
 
 	unsigned int max_page_table_levels;
+	/**
+	 * @max_num_dpp: Maximum number of DPP supported in the target ASIC.
+	 */
 	unsigned int max_num_dpp;
 	unsigned int max_num_otg;
 	unsigned int cursor_chunk_size;
@@ -223,6 +363,13 @@ struct _vcs_dpi_ip_params_st {
 	unsigned int can_vstartup_lines_exceed_vsync_plus_back_porch_lines_minus_one;
 	unsigned int bug_forcing_LC_req_same_size_fixed;
 	unsigned int number_of_cursors;
+	unsigned int max_num_dp2p0_outputs;
+	unsigned int max_num_dp2p0_streams;
+	unsigned int VBlankNomDefaultUS;
+
+	/* DM workarounds */
+	double dsc_delay_factor_wa; // TODO: Remove after implementing root cause fix
+	double min_prefetch_in_strobe_us;
 };
 
 struct _vcs_dpi_display_xfc_params_st {
@@ -249,6 +396,8 @@ struct _vcs_dpi_display_pipe_source_params_st {
 	bool hostvm_levels_force_en;
 	unsigned int hostvm_levels_force;
 	int source_scan;
+	int source_rotation; // new in dml32
+	unsigned int det_size_override; // use to populate DETSizeOverride in vba struct
 	int sw_mode;
 	int macro_tile_size;
 	unsigned int surface_width_y;
@@ -263,6 +412,15 @@ struct _vcs_dpi_display_pipe_source_params_st {
 	unsigned int viewport_height_c;
 	unsigned int viewport_width_max;
 	unsigned int viewport_height_max;
+	unsigned int viewport_x_y;
+	unsigned int viewport_x_c;
+	bool viewport_stationary;
+	unsigned int dcc_rate_luma;
+	unsigned int gpuvm_min_page_size_kbytes;
+	unsigned int use_mall_for_pstate_change;
+	unsigned int use_mall_for_static_screen;
+	bool force_one_row_for_frame;
+	bool pte_buffer_mode;
 	unsigned int data_pitch;
 	unsigned int data_pitch_c;
 	unsigned int meta_pitch;
@@ -295,10 +453,17 @@ struct writeback_st {
 	int wb_vtaps_luma;
 	int wb_htaps_chroma;
 	int wb_vtaps_chroma;
+	unsigned int wb_htaps;
+	unsigned int wb_vtaps;
 	double wb_hratio;
 	double wb_vratio;
 };
 
+struct display_audio_params_st {
+	unsigned int   audio_sample_rate_khz;
+	int    		   audio_sample_layout;
+};
+
 struct _vcs_dpi_display_output_params_st {
 	int dp_lanes;
 	double output_bpp;
@@ -312,6 +477,11 @@ struct _vcs_dpi_display_output_params_st {
 	int dsc_slices;
 	int max_audio_sample_rate;
 	struct writeback_st wb;
+	struct display_audio_params_st audio;
+	unsigned int output_bpc;
+	int dp_rate;
+	unsigned int dp_multistream_id;
+	bool dp_multistream_en;
 };
 
 struct _vcs_dpi_scaler_ratio_depth_st {
@@ -346,12 +516,14 @@ struct _vcs_dpi_display_pipe_dest_params_st {
 	unsigned int htotal;
 	unsigned int vtotal;
 	unsigned int vfront_porch;
+	unsigned int vblank_nom;
 	unsigned int vactive;
 	unsigned int hactive;
 	unsigned int vstartup_start;
 	unsigned int vupdate_offset;
 	unsigned int vupdate_width;
 	unsigned int vready_offset;
+	unsigned int pstate_keepout;
 	unsigned char interlaced;
 	double pixel_rate_mhz;
 	unsigned char synchronized_vblank_all_planes;
@@ -360,6 +532,10 @@ struct _vcs_dpi_display_pipe_dest_params_st {
 	unsigned char use_maximum_vstartup;
 	unsigned int vtotal_max;
 	unsigned int vtotal_min;
+	unsigned int refresh_rate;
+	bool synchronize_timings;
+	unsigned int odm_combine_policy;
+	bool drr_display;
 };
 
 struct _vcs_dpi_display_pipe_params_st {
@@ -445,6 +621,7 @@ struct _vcs_dpi_display_dlg_regs_st {
 	unsigned int refcyc_h_blank_end;
 	unsigned int dlg_vblank_end;
 	unsigned int min_dst_y_next_start;
+	unsigned int min_dst_y_next_start_us;
 	unsigned int refcyc_per_htotal;
 	unsigned int refcyc_x_after_scaler;
 	unsigned int dst_y_after_scaler;
@@ -456,6 +633,7 @@ struct _vcs_dpi_display_dlg_regs_st {
 	unsigned int ref_freq_to_pix_freq;
 	unsigned int vratio_prefetch;
 	unsigned int vratio_prefetch_c;
+	unsigned int refcyc_per_tdlut_group;
 	unsigned int refcyc_per_pte_group_vblank_l;
 	unsigned int refcyc_per_pte_group_vblank_c;
 	unsigned int refcyc_per_meta_chunk_vblank_l;
@@ -546,7 +724,6 @@ struct _vcs_dpi_display_dlg_sys_params_st {
 	double t_sr_wm_us;
 	double t_extra_us;
 	double mem_trip_us;
-	double t_srx_delay_us;
 	double deepsleep_dcfclk_mhz;
 	double total_flip_bw;
 	unsigned int total_flip_bytes;
@@ -556,6 +733,9 @@ struct _vcs_dpi_display_arb_params_st {
 	int max_req_outstanding;
 	int min_req_outstanding;
 	int sat_level_us;
+	int hvm_min_req_outstand_commit_threshold;
+	int hvm_max_qos_commit_threshold;
+	int compbuf_reserved_space_kbytes;
 };
 
 #endif /*__DISPLAY_MODE_STRUCTS_H__*/
diff --git a/drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.c b/drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.c
index 0fad15020c74..85453bbb4f9b 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.c
@@ -47,6 +47,7 @@ static void recalculate_params(
 		unsigned int num_pipes);
 
 static unsigned int CursorBppEnumToBits(enum cursor_bpp ebpp);
+static void cache_debug_params(struct display_mode_lib *mode_lib);
 
 unsigned int dml_get_voltage_level(
 		struct display_mode_lib *mode_lib,
@@ -73,6 +74,7 @@ unsigned int dml_get_voltage_level(
 		PixelClockAdjustmentForProgressiveToInterlaceUnit(mode_lib);
 	}
 	mode_lib->funcs.validate(mode_lib);
+	cache_debug_params(mode_lib);
 
 	return mode_lib->vba.VoltageLevel;
 }
@@ -109,6 +111,22 @@ dml_get_attr_func(tcalc, mode_lib->vba.TCalc);
 dml_get_attr_func(fraction_of_urgent_bandwidth, mode_lib->vba.FractionOfUrgentBandwidth);
 dml_get_attr_func(fraction_of_urgent_bandwidth_imm_flip, mode_lib->vba.FractionOfUrgentBandwidthImmediateFlip);
 
+
+dml_get_attr_func(cstate_max_cap_mode, mode_lib->vba.DCHUBBUB_ARB_CSTATE_MAX_CAP_MODE);
+dml_get_attr_func(comp_buffer_size_kbytes, mode_lib->vba.CompressedBufferSizeInkByte);
+dml_get_attr_func(pixel_chunk_size_in_kbyte, mode_lib->vba.PixelChunkSizeInKByte);
+dml_get_attr_func(alpha_pixel_chunk_size_in_kbyte, mode_lib->vba.AlphaPixelChunkSizeInKByte);
+dml_get_attr_func(meta_chunk_size_in_kbyte, mode_lib->vba.MetaChunkSize);
+dml_get_attr_func(min_pixel_chunk_size_in_byte, mode_lib->vba.MinPixelChunkSizeBytes);
+dml_get_attr_func(min_meta_chunk_size_in_byte, mode_lib->vba.MinMetaChunkSizeBytes);
+dml_get_attr_func(fclk_watermark, mode_lib->vba.Watermark.FCLKChangeWatermark);
+dml_get_attr_func(usr_retraining_watermark, mode_lib->vba.Watermark.USRRetrainingWatermark);
+
+dml_get_attr_func(comp_buffer_reserved_space_kbytes, mode_lib->vba.CompBufReservedSpaceKBytes);
+dml_get_attr_func(comp_buffer_reserved_space_64bytes, mode_lib->vba.CompBufReservedSpace64B);
+dml_get_attr_func(comp_buffer_reserved_space_zs, mode_lib->vba.CompBufReservedSpaceZs);
+dml_get_attr_func(unbounded_request_enabled, mode_lib->vba.UnboundedRequestEnabled);
+
 #define dml_get_pipe_attr_func(attr, var)  double get_##attr(struct display_mode_lib *mode_lib, const display_e2e_pipe_params_st *pipes, unsigned int num_pipes, unsigned int which_pipe) \
 {\
 	unsigned int which_plane; \
@@ -163,6 +181,28 @@ dml_get_pipe_attr_func(vupdate_width, mode_lib->vba.VUpdateWidthPix);
 dml_get_pipe_attr_func(vready_offset, mode_lib->vba.VReadyOffsetPix);
 dml_get_pipe_attr_func(vready_at_or_after_vsync, mode_lib->vba.VREADY_AT_OR_AFTER_VSYNC);
 dml_get_pipe_attr_func(min_dst_y_next_start, mode_lib->vba.MIN_DST_Y_NEXT_START);
+dml_get_pipe_attr_func(dst_y_per_pte_row_nom_l, mode_lib->vba.DST_Y_PER_PTE_ROW_NOM_L);
+dml_get_pipe_attr_func(dst_y_per_pte_row_nom_c, mode_lib->vba.DST_Y_PER_PTE_ROW_NOM_C);
+dml_get_pipe_attr_func(dst_y_per_meta_row_nom_l, mode_lib->vba.DST_Y_PER_META_ROW_NOM_L);
+dml_get_pipe_attr_func(dst_y_per_meta_row_nom_c, mode_lib->vba.DST_Y_PER_META_ROW_NOM_C);
+dml_get_pipe_attr_func(refcyc_per_pte_group_nom_l_in_us, mode_lib->vba.time_per_pte_group_nom_luma);
+dml_get_pipe_attr_func(refcyc_per_pte_group_nom_c_in_us, mode_lib->vba.time_per_pte_group_nom_chroma);
+dml_get_pipe_attr_func(refcyc_per_pte_group_vblank_l_in_us, mode_lib->vba.time_per_pte_group_vblank_luma);
+dml_get_pipe_attr_func(refcyc_per_pte_group_vblank_c_in_us, mode_lib->vba.time_per_pte_group_vblank_chroma);
+dml_get_pipe_attr_func(refcyc_per_pte_group_flip_l_in_us, mode_lib->vba.time_per_pte_group_flip_luma);
+dml_get_pipe_attr_func(refcyc_per_pte_group_flip_c_in_us, mode_lib->vba.time_per_pte_group_flip_chroma);
+dml_get_pipe_attr_func(vstartup_calculated, mode_lib->vba.VStartup);
+dml_get_pipe_attr_func(dpte_row_height_linear_c, mode_lib->vba.dpte_row_height_linear_chroma);
+dml_get_pipe_attr_func(swath_height_l, mode_lib->vba.SwathHeightY);
+dml_get_pipe_attr_func(swath_height_c, mode_lib->vba.SwathHeightC);
+dml_get_pipe_attr_func(det_stored_buffer_size_l_bytes, mode_lib->vba.DETBufferSizeY);
+dml_get_pipe_attr_func(det_stored_buffer_size_c_bytes, mode_lib->vba.DETBufferSizeC);
+dml_get_pipe_attr_func(dpte_group_size_in_bytes, mode_lib->vba.dpte_group_bytes);
+dml_get_pipe_attr_func(vm_group_size_in_bytes, mode_lib->vba.vm_group_bytes);
+dml_get_pipe_attr_func(dpte_row_height_linear_l, mode_lib->vba.dpte_row_height_linear);
+dml_get_pipe_attr_func(pte_buffer_mode, mode_lib->vba.PTE_BUFFER_MODE);
+dml_get_pipe_attr_func(subviewport_lines_needed_in_mall, mode_lib->vba.SubViewportLinesNeededInMALL);
+dml_get_pipe_attr_func(surface_size_in_mall, mode_lib->vba.SurfaceSizeInMALL)
 
 double get_total_immediate_flip_bytes(
 		struct display_mode_lib *mode_lib,
@@ -200,6 +240,67 @@ double get_total_prefetch_bw(
 	return total_prefetch_bw;
 }
 
+unsigned int get_total_surface_size_in_mall_bytes(
+		struct display_mode_lib *mode_lib,
+		const display_e2e_pipe_params_st *pipes,
+		unsigned int num_pipes)
+{
+	unsigned int k;
+	unsigned int size = 0.0;
+	recalculate_params(mode_lib, pipes, num_pipes);
+	for (k = 0; k < mode_lib->vba.NumberOfActiveSurfaces; ++k)
+		size += mode_lib->vba.SurfaceSizeInMALL[k];
+	return size;
+}
+
+static unsigned int get_pipe_idx(struct display_mode_lib *mode_lib, unsigned int plane_idx)
+{
+	int pipe_idx = -1;
+	int i;
+
+	ASSERT(plane_idx < DC__NUM_DPP__MAX);
+
+	for (i = 0; i < DC__NUM_DPP__MAX ; i++) {
+		if (plane_idx == mode_lib->vba.pipe_plane[i]) {
+			pipe_idx = i;
+			break;
+		}
+	}
+	ASSERT(pipe_idx >= 0);
+
+	return pipe_idx;
+}
+
+
+double get_det_buffer_size_kbytes(struct display_mode_lib *mode_lib, const display_e2e_pipe_params_st *pipes,
+		unsigned int num_pipes, unsigned int pipe_idx)
+{
+	unsigned int plane_idx;
+	double det_buf_size_kbytes;
+
+	recalculate_params(mode_lib, pipes, num_pipes);
+	plane_idx = mode_lib->vba.pipe_plane[pipe_idx];
+
+	dml_print("DML::%s: num_pipes=%d pipe_idx=%d plane_idx=%0d\n", __func__, num_pipes, pipe_idx, plane_idx);
+	det_buf_size_kbytes = mode_lib->vba.DETBufferSizeInKByte[plane_idx]; // per hubp DET buffer size
+
+	dml_print("DML::%s: det_buf_size_kbytes=%3.2f\n", __func__, det_buf_size_kbytes);
+
+	return det_buf_size_kbytes;
+}
+
+bool get_is_phantom_pipe(struct display_mode_lib *mode_lib, const display_e2e_pipe_params_st *pipes,
+		unsigned int num_pipes, unsigned int pipe_idx)
+{
+	unsigned int plane_idx;
+
+	recalculate_params(mode_lib, pipes, num_pipes);
+	plane_idx = mode_lib->vba.pipe_plane[pipe_idx];
+	dml_print("DML::%s: num_pipes=%d pipe_idx=%d UseMALLForPStateChange=%0d\n", __func__, num_pipes, pipe_idx,
+			mode_lib->vba.UsesMALLForPStateChange[plane_idx]);
+	return (mode_lib->vba.UsesMALLForPStateChange[plane_idx] == dm_use_mall_pstate_change_phantom_pipe);
+}
+
 static void fetch_socbb_params(struct display_mode_lib *mode_lib)
 {
 	soc_bounding_box_st *soc = &mode_lib->vba.soc;
@@ -239,6 +340,21 @@ static void fetch_socbb_params(struct display_mode_lib *mode_lib)
 			soc->max_avg_sdp_bw_use_normal_percent;
 	mode_lib->vba.SRExitZ8Time = soc->sr_exit_z8_time_us;
 	mode_lib->vba.SREnterPlusExitZ8Time = soc->sr_enter_plus_exit_z8_time_us;
+	mode_lib->vba.FCLKChangeLatency = soc->fclk_change_latency_us;
+	mode_lib->vba.USRRetrainingLatency = soc->usr_retraining_latency_us;
+	mode_lib->vba.SMNLatency = soc->smn_latency_us;
+	mode_lib->vba.MALLAllocatedForDCNFinal = soc->mall_allocated_for_dcn_mbytes;
+
+	mode_lib->vba.PercentOfIdealDRAMBWReceivedAfterUrgLatencySTROBE = soc->pct_ideal_dram_bw_after_urgent_strobe;
+	mode_lib->vba.MaxAveragePercentOfIdealFabricBWDisplayCanUseInNormalSystemOperation =
+			soc->max_avg_fabric_bw_use_normal_percent;
+	mode_lib->vba.MaxAveragePercentOfIdealDRAMBWDisplayCanUseInNormalSystemOperationSTROBE =
+			soc->max_avg_dram_bw_use_normal_strobe_percent;
+
+	mode_lib->vba.DRAMClockChangeRequirementFinal = soc->dram_clock_change_requirement_final;
+	mode_lib->vba.FCLKChangeRequirementFinal = 1;
+	mode_lib->vba.USRRetrainingRequiredFinal = 1;
+	mode_lib->vba.AllowForPStateChangeOrStutterInVBlankFinal = soc->allow_for_pstate_or_stutter_in_vblank_final;
 	mode_lib->vba.DRAMClockChangeLatency = soc->dram_clock_change_latency_us;
 	mode_lib->vba.DummyPStateCheck = soc->dram_clock_change_latency_us == soc->dummy_pstate_latency_us;
 	mode_lib->vba.DRAMClockChangeSupportsVActive = !soc->disable_dram_clock_change_vactive_support ||
@@ -281,6 +397,7 @@ static void fetch_socbb_params(struct display_mode_lib *mode_lib)
 		mode_lib->vba.SOCCLKPerState[i] = soc->clock_limits[i].socclk_mhz;
 		mode_lib->vba.PHYCLKPerState[i] = soc->clock_limits[i].phyclk_mhz;
 		mode_lib->vba.PHYCLKD18PerState[i] = soc->clock_limits[i].phyclk_d18_mhz;
+		mode_lib->vba.PHYCLKD32PerState[i] = soc->clock_limits[i].phyclk_d32_mhz;
 		mode_lib->vba.MaxDppclk[i] = soc->clock_limits[i].dppclk_mhz;
 		mode_lib->vba.MaxDSCCLK[i] = soc->clock_limits[i].dscclk_mhz;
 		mode_lib->vba.DRAMSpeedPerState[i] = soc->clock_limits[i].dram_speed_mts;
@@ -295,6 +412,7 @@ static void fetch_socbb_params(struct display_mode_lib *mode_lib)
 		soc->urgent_latency_adjustment_fabric_clock_component_us;
 	mode_lib->vba.UrgentLatencyAdjustmentFabricClockReference =
 		soc->urgent_latency_adjustment_fabric_clock_reference_mhz;
+	mode_lib->vba.MaxVRatioPre = soc->max_vratio_pre;
 }
 
 static void fetch_ip_params(struct display_mode_lib *mode_lib)
@@ -323,6 +441,18 @@ static void fetch_ip_params(struct display_mode_lib *mode_lib)
 	mode_lib->vba.COMPBUF_RESERVED_SPACE_ZS = ip->compbuf_reserved_space_zs;
 	mode_lib->vba.MaximumDSCBitsPerComponent = ip->maximum_dsc_bits_per_component;
 	mode_lib->vba.DSC422NativeSupport = ip->dsc422_native_support;
+    /* In DCN3.2, nomDETInKByte should be initialized correctly. */
+	mode_lib->vba.nomDETInKByte = ip->det_buffer_size_kbytes;
+	mode_lib->vba.CompbufReservedSpace64B  = ip->compbuf_reserved_space_64b;
+	mode_lib->vba.CompbufReservedSpaceZs = ip->compbuf_reserved_space_zs;
+	mode_lib->vba.CompressedBufferSegmentSizeInkByteFinal = ip->compressed_buffer_segment_size_in_kbytes;
+	mode_lib->vba.LineBufferSizeFinal = ip->line_buffer_size_bits;
+	mode_lib->vba.AlphaPixelChunkSizeInKByte = ip->alpha_pixel_chunk_size_kbytes; // not ysed
+	mode_lib->vba.MinPixelChunkSizeBytes = ip->min_pixel_chunk_size_bytes; // not used
+	mode_lib->vba.MaximumPixelsPerLinePerDSCUnit = ip->maximum_pixels_per_line_per_dsc_unit;
+	mode_lib->vba.MaxNumDP2p0Outputs = ip->max_num_dp2p0_outputs;
+	mode_lib->vba.MaxNumDP2p0Streams = ip->max_num_dp2p0_streams;
+	mode_lib->vba.DCCMetaBufferSizeBytes = ip->dcc_meta_buffer_size_bytes;
 
 	mode_lib->vba.PixelChunkSizeInKByte = ip->pixel_chunk_size_kbytes;
 	mode_lib->vba.MetaChunkSize = ip->meta_chunk_size_kbytes;
@@ -397,6 +527,7 @@ static void fetch_pipe_params(struct display_mode_lib *mode_lib)
 		visited[k] = false;
 
 	mode_lib->vba.NumberOfActivePlanes = 0;
+	mode_lib->vba.NumberOfActiveSurfaces = 0;
 	mode_lib->vba.ImmediateFlipSupport = false;
 	for (j = 0; j < mode_lib->vba.cache_num_pipes; ++j) {
 		display_pipe_source_params_st *src = &pipes[j].pipe.src;
@@ -427,6 +558,26 @@ static void fetch_pipe_params(struct display_mode_lib *mode_lib)
 				src->viewport_y_y;
 		mode_lib->vba.ViewportYStartC[mode_lib->vba.NumberOfActivePlanes] =
 				src->viewport_y_c;
+		mode_lib->vba.SourceRotation[mode_lib->vba.NumberOfActiveSurfaces] = src->source_rotation;
+		mode_lib->vba.ViewportXStartY[mode_lib->vba.NumberOfActiveSurfaces] = src->viewport_x_y;
+		mode_lib->vba.ViewportXStartC[mode_lib->vba.NumberOfActiveSurfaces] = src->viewport_x_c;
+		// TODO: Assign correct value to viewport_stationary
+		mode_lib->vba.ViewportStationary[mode_lib->vba.NumberOfActivePlanes] =
+				src->viewport_stationary;
+		mode_lib->vba.UsesMALLForPStateChange[mode_lib->vba.NumberOfActivePlanes] = src->use_mall_for_pstate_change;
+		mode_lib->vba.UseMALLForStaticScreen[mode_lib->vba.NumberOfActivePlanes] = src->use_mall_for_static_screen;
+		mode_lib->vba.GPUVMMinPageSizeKBytes[mode_lib->vba.NumberOfActivePlanes] = src->gpuvm_min_page_size_kbytes;
+		mode_lib->vba.RefreshRate[mode_lib->vba.NumberOfActivePlanes] = dst->refresh_rate; //todo remove this
+		mode_lib->vba.OutputLinkDPRate[mode_lib->vba.NumberOfActivePlanes] = dout->dp_rate;
+		mode_lib->vba.ODMUse[mode_lib->vba.NumberOfActivePlanes] = dst->odm_combine_policy;
+		mode_lib->vba.DETSizeOverride[mode_lib->vba.NumberOfActivePlanes] = src->det_size_override;
+		if (src->det_size_override)
+			mode_lib->vba.DETBufferSizeInKByte[mode_lib->vba.NumberOfActivePlanes] = src->det_size_override;
+		else
+			mode_lib->vba.DETBufferSizeInKByte[mode_lib->vba.NumberOfActivePlanes] = ip->det_buffer_size_kbytes;
+		//TODO: Need to assign correct values to dp_multistream vars
+		mode_lib->vba.OutputMultistreamEn[mode_lib->vba.NumberOfActiveSurfaces] = dout->dp_multistream_en;
+		mode_lib->vba.OutputMultistreamId[mode_lib->vba.NumberOfActiveSurfaces] = dout->dp_multistream_id;
 		mode_lib->vba.PitchY[mode_lib->vba.NumberOfActivePlanes] = src->data_pitch;
 		mode_lib->vba.SurfaceWidthY[mode_lib->vba.NumberOfActivePlanes] = src->surface_width_y;
 		mode_lib->vba.SurfaceHeightY[mode_lib->vba.NumberOfActivePlanes] = src->surface_height_y;
@@ -452,6 +603,7 @@ static void fetch_pipe_params(struct display_mode_lib *mode_lib)
 		mode_lib->vba.HTotal[mode_lib->vba.NumberOfActivePlanes] = dst->htotal;
 		mode_lib->vba.VTotal[mode_lib->vba.NumberOfActivePlanes] = dst->vtotal;
 		mode_lib->vba.VFrontPorch[mode_lib->vba.NumberOfActivePlanes] = dst->vfront_porch;
+		mode_lib->vba.VBlankNom[mode_lib->vba.NumberOfActivePlanes] = dst->vblank_nom;
 		mode_lib->vba.DCCFractionOfZeroSizeRequestsLuma[mode_lib->vba.NumberOfActivePlanes] = src->dcc_fraction_of_zs_req_luma;
 		mode_lib->vba.DCCFractionOfZeroSizeRequestsChroma[mode_lib->vba.NumberOfActivePlanes] = src->dcc_fraction_of_zs_req_chroma;
 		mode_lib->vba.DCCEnable[mode_lib->vba.NumberOfActivePlanes] =
@@ -479,7 +631,7 @@ static void fetch_pipe_params(struct display_mode_lib *mode_lib)
 		mode_lib->vba.skip_dio_check[mode_lib->vba.NumberOfActivePlanes] =
 				dout->is_virtual;
 
-		if (!dout->dsc_enable)
+		if (dout->dsc_enable)
 			mode_lib->vba.ForcedOutputLinkBPP[mode_lib->vba.NumberOfActivePlanes] = dout->output_bpp;
 		else
 			mode_lib->vba.ForcedOutputLinkBPP[mode_lib->vba.NumberOfActivePlanes] = 0.0;
@@ -553,6 +705,7 @@ static void fetch_pipe_params(struct display_mode_lib *mode_lib)
 		mode_lib->vba.PixelClock[mode_lib->vba.NumberOfActivePlanes] = dst->pixel_rate_mhz;
 		mode_lib->vba.PixelClockBackEnd[mode_lib->vba.NumberOfActivePlanes] = dst->pixel_rate_mhz;
 		mode_lib->vba.DPPCLK[mode_lib->vba.NumberOfActivePlanes] = clks->dppclk_mhz;
+		mode_lib->vba.DRRDisplay[mode_lib->vba.NumberOfActiveSurfaces] = dst->drr_display;
 		if (ip->is_line_buffer_bpp_fixed)
 			mode_lib->vba.LBBitPerPixel[mode_lib->vba.NumberOfActivePlanes] =
 					ip->line_buffer_fixed_bpp;
@@ -636,6 +789,8 @@ static void fetch_pipe_params(struct display_mode_lib *mode_lib)
 					mode_lib->vba.pipe_plane[k] =
 							mode_lib->vba.NumberOfActivePlanes;
 					mode_lib->vba.DPPPerPlane[mode_lib->vba.NumberOfActivePlanes]++;
+					if (src_k->det_size_override)
+						mode_lib->vba.DETBufferSizeInKByte[mode_lib->vba.NumberOfActivePlanes] = src_k->det_size_override;
 					if (mode_lib->vba.SourceScan[mode_lib->vba.NumberOfActivePlanes]
 							== dm_horz) {
 						mode_lib->vba.ViewportWidth[mode_lib->vba.NumberOfActivePlanes] +=
@@ -675,6 +830,7 @@ static void fetch_pipe_params(struct display_mode_lib *mode_lib)
 		}
 
 		mode_lib->vba.NumberOfActivePlanes++;
+		mode_lib->vba.NumberOfActiveSurfaces++;
 	}
 
 	// handle overlays through BlendingAndTiming
@@ -700,6 +856,11 @@ static void fetch_pipe_params(struct display_mode_lib *mode_lib)
 		}
 	}
 
+	mode_lib->vba.SynchronizeTimingsFinal = pipes[0].pipe.dest.synchronize_timings;
+	mode_lib->vba.DCCProgrammingAssumesScanDirectionUnknownFinal = false;
+
+	mode_lib->vba.DisableUnboundRequestIfCompBufReservedSpaceNeedAdjustment = 0;
+
 	mode_lib->vba.UseUnboundedRequesting = dm_unbounded_requesting;
 	for (k = 0; k < mode_lib->vba.cache_num_pipes; ++k) {
 		if (pipes[k].pipe.src.unbounded_req_mode == 0)
@@ -743,6 +904,52 @@ static void fetch_pipe_params(struct display_mode_lib *mode_lib)
 
 	mode_lib->vba.GPUVMEnable = mode_lib->vba.GPUVMEnable && !!ip->gpuvm_enable;
 	mode_lib->vba.HostVMEnable = mode_lib->vba.HostVMEnable && !!ip->hostvm_enable;
+
+	for (k = 0; k < mode_lib->vba.cache_num_pipes; ++k) {
+		mode_lib->vba.ForceOneRowForFrame[k] = pipes[k].pipe.src.force_one_row_for_frame;
+		mode_lib->vba.PteBufferMode[k] = pipes[k].pipe.src.pte_buffer_mode;
+
+		if (mode_lib->vba.PteBufferMode[k] == 0 && mode_lib->vba.GPUVMEnable) {
+			if (mode_lib->vba.ForceOneRowForFrame[k] ||
+				(mode_lib->vba.GPUVMMinPageSizeKBytes[k] > 64*1024) ||
+				(mode_lib->vba.UsesMALLForPStateChange[k] != dm_use_mall_pstate_change_disable) ||
+				(mode_lib->vba.UseMALLForStaticScreen[k] != dm_use_mall_static_screen_disable)) {
+#ifdef __DML_VBA_DEBUG__
+				dml_print("DML::%s: ERROR: Invalid PteBufferMode=%d for plane %0d!\n",
+						__func__, mode_lib->vba.PteBufferMode[k], k);
+				dml_print("DML::%s:  -  ForceOneRowForFrame     = %d\n",
+						__func__, mode_lib->vba.ForceOneRowForFrame[k]);
+				dml_print("DML::%s:  -  GPUVMMinPageSizeKBytes  = %d\n",
+						__func__, mode_lib->vba.GPUVMMinPageSizeKBytes[k]);
+				dml_print("DML::%s:  -  UseMALLForPStateChange  = %d\n",
+						__func__, (int) mode_lib->vba.UsesMALLForPStateChange[k]);
+				dml_print("DML::%s:  -  UseMALLForStaticScreen  = %d\n",
+						__func__, (int) mode_lib->vba.UseMALLForStaticScreen[k]);
+#endif
+				ASSERT(0);
+			}
+		}
+	}
+}
+
+/**
+ * cache_debug_params: Cache any params that needed to be maintained from the initial validation
+ * for debug purposes.
+ *
+ * The DML getters can modify some of the VBA params that we are interested in (for example when
+ * calculating with dummy p-state latency), so cache any params here that we want for debugging
+ *
+ * @mode_lib: mode_lib input/output of validate call
+ *
+ * Return: void
+ *
+ */
+static void cache_debug_params(struct display_mode_lib *mode_lib)
+{
+	int k = 0;
+
+	for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; k++)
+		mode_lib->vba.CachedActiveDRAMClockChangeLatencyMargin[k] = mode_lib->vba.ActiveDRAMClockChangeLatencyMargin[k];
 }
 
 // in wm mode we pull the parameters needed from the display_e2e_pipe_params_st structs
@@ -768,7 +975,7 @@ static void recalculate_params(
 	}
 }
 
-bool Calculate256BBlockSizes(
+void Calculate256BBlockSizes(
 		enum source_format_class SourcePixelFormat,
 		enum dm_swizzle_mode SurfaceTiling,
 		unsigned int BytePerPixelY,
@@ -806,7 +1013,6 @@ bool Calculate256BBlockSizes(
 		*BlockWidth256BytesY = 256 / BytePerPixelY / *BlockHeight256BytesY;
 		*BlockWidth256BytesC = 256 / BytePerPixelC / *BlockHeight256BytesC;
 	}
-	return true;
 }
 
 bool CalculateMinAndMaxPrefetchMode(
@@ -872,6 +1078,7 @@ void ModeSupportAndSystemConfiguration(struct display_mode_lib *mode_lib)
 	soc_bounding_box_st *soc = &mode_lib->vba.soc;
 	unsigned int k;
 	unsigned int total_pipes = 0;
+	unsigned int pipe_idx = 0;
 
 	mode_lib->vba.VoltageLevel = mode_lib->vba.cache_pipes[0].clks_cfg.voltage;
 	mode_lib->vba.ReturnBW = mode_lib->vba.ReturnBWPerState[mode_lib->vba.VoltageLevel][mode_lib->vba.maxMpcComb];
@@ -892,7 +1099,17 @@ void ModeSupportAndSystemConfiguration(struct display_mode_lib *mode_lib)
 
 	// Total Available Pipes Support Check
 	for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) {
+		pipe_idx = get_pipe_idx(mode_lib, k);
+		if (pipe_idx == -1) {
+			ASSERT(0);
+			continue; // skip inactive planes
+		}
 		total_pipes += mode_lib->vba.DPPPerPlane[k];
+
+		if (mode_lib->vba.cache_pipes[pipe_idx].clks_cfg.dppclk_mhz > 0.0)
+			mode_lib->vba.DPPCLK[k] = mode_lib->vba.cache_pipes[pipe_idx].clks_cfg.dppclk_mhz;
+		else
+			mode_lib->vba.DPPCLK[k] = soc->clock_limits[mode_lib->vba.VoltageLevel].dppclk_mhz;
 	}
 	ASSERT(total_pipes <= DC__NUM_DPP__MAX);
 }
diff --git a/drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.h b/drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.h
index 90e87961fe3e..07993741f5e6 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.h
+++ b/drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.h
@@ -58,6 +58,19 @@ dml_get_attr_decl(return_bw);
 dml_get_attr_decl(tcalc);
 dml_get_attr_decl(fraction_of_urgent_bandwidth);
 dml_get_attr_decl(fraction_of_urgent_bandwidth_imm_flip);
+dml_get_attr_decl(cstate_max_cap_mode);
+dml_get_attr_decl(comp_buffer_size_kbytes);
+dml_get_attr_decl(pixel_chunk_size_in_kbyte);
+dml_get_attr_decl(alpha_pixel_chunk_size_in_kbyte);
+dml_get_attr_decl(meta_chunk_size_in_kbyte);
+dml_get_attr_decl(min_pixel_chunk_size_in_byte);
+dml_get_attr_decl(min_meta_chunk_size_in_byte);
+dml_get_attr_decl(fclk_watermark);
+dml_get_attr_decl(usr_retraining_watermark);
+dml_get_attr_decl(comp_buffer_reserved_space_kbytes);
+dml_get_attr_decl(comp_buffer_reserved_space_64bytes);
+dml_get_attr_decl(comp_buffer_reserved_space_zs);
+dml_get_attr_decl(unbounded_request_enabled);
 
 #define dml_get_pipe_attr_decl(attr) double get_##attr(struct display_mode_lib *mode_lib, const display_e2e_pipe_params_st *pipes, unsigned int num_pipes, unsigned int which_pipe)
 
@@ -75,6 +88,26 @@ dml_get_pipe_attr_decl(dst_y_per_row_vblank);
 dml_get_pipe_attr_decl(dst_y_prefetch);
 dml_get_pipe_attr_decl(dst_y_per_vm_flip);
 dml_get_pipe_attr_decl(dst_y_per_row_flip);
+dml_get_pipe_attr_decl(dst_y_per_pte_row_nom_l);
+dml_get_pipe_attr_decl(dst_y_per_pte_row_nom_c);
+dml_get_pipe_attr_decl(dst_y_per_meta_row_nom_l);
+dml_get_pipe_attr_decl(dst_y_per_meta_row_nom_c);
+dml_get_pipe_attr_decl(dpte_row_height_linear_c);
+dml_get_pipe_attr_decl(swath_height_l);
+dml_get_pipe_attr_decl(swath_height_c);
+dml_get_pipe_attr_decl(det_stored_buffer_size_l_bytes);
+dml_get_pipe_attr_decl(det_stored_buffer_size_c_bytes);
+dml_get_pipe_attr_decl(dpte_group_size_in_bytes);
+dml_get_pipe_attr_decl(vm_group_size_in_bytes);
+dml_get_pipe_attr_decl(det_buffer_size_kbytes);
+dml_get_pipe_attr_decl(dpte_row_height_linear_l);
+dml_get_pipe_attr_decl(refcyc_per_pte_group_nom_l_in_us);
+dml_get_pipe_attr_decl(refcyc_per_pte_group_nom_c_in_us);
+dml_get_pipe_attr_decl(refcyc_per_pte_group_vblank_l_in_us);
+dml_get_pipe_attr_decl(refcyc_per_pte_group_vblank_c_in_us);
+dml_get_pipe_attr_decl(refcyc_per_pte_group_flip_l_in_us);
+dml_get_pipe_attr_decl(refcyc_per_pte_group_flip_c_in_us);
+dml_get_pipe_attr_decl(pte_buffer_mode);
 dml_get_pipe_attr_decl(refcyc_per_vm_group_vblank);
 dml_get_pipe_attr_decl(refcyc_per_vm_group_flip);
 dml_get_pipe_attr_decl(refcyc_per_vm_req_vblank);
@@ -108,6 +141,9 @@ dml_get_pipe_attr_decl(vupdate_width);
 dml_get_pipe_attr_decl(vready_offset);
 dml_get_pipe_attr_decl(vready_at_or_after_vsync);
 dml_get_pipe_attr_decl(min_dst_y_next_start);
+dml_get_pipe_attr_decl(vstartup_calculated);
+dml_get_pipe_attr_decl(subviewport_lines_needed_in_mall);
+dml_get_pipe_attr_decl(surface_size_in_mall);
 
 double get_total_immediate_flip_bytes(
 		struct display_mode_lib *mode_lib,
@@ -126,9 +162,18 @@ unsigned int dml_get_voltage_level(
 		const display_e2e_pipe_params_st *pipes,
 		unsigned int num_pipes);
 
+unsigned int get_total_surface_size_in_mall_bytes(
+		struct display_mode_lib *mode_lib,
+		const display_e2e_pipe_params_st *pipes,
+		unsigned int num_pipes);
+
+bool get_is_phantom_pipe(struct display_mode_lib *mode_lib,
+		const display_e2e_pipe_params_st *pipes,
+		unsigned int num_pipes,
+		unsigned int pipe_idx);
 void PixelClockAdjustmentForProgressiveToInterlaceUnit(struct display_mode_lib *mode_lib);
 
-bool Calculate256BBlockSizes(
+void Calculate256BBlockSizes(
 		enum source_format_class SourcePixelFormat,
 		enum dm_swizzle_mode SurfaceTiling,
 		unsigned int BytePerPixelY,
@@ -138,6 +183,79 @@ bool Calculate256BBlockSizes(
 		unsigned int *BlockWidth256BytesY,
 		unsigned int *BlockWidth256BytesC);
 
+struct DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation {
+	unsigned int dummy_integer_array[2][DC__NUM_DPP__MAX];
+	double dummy_single_array[2][DC__NUM_DPP__MAX];
+	unsigned int dummy_long_array[2][DC__NUM_DPP__MAX];
+	double dummy_double_array[2][DC__NUM_DPP__MAX];
+	bool dummy_boolean_array[DC__NUM_DPP__MAX];
+	bool dummy_boolean;
+	bool dummy_boolean2;
+	enum output_encoder_class dummy_output_encoder_array[DC__NUM_DPP__MAX];
+	DmlPipe SurfaceParameters[DC__NUM_DPP__MAX];
+	bool dummy_boolean_array2[2][DC__NUM_DPP__MAX];
+	unsigned int ReorderBytes;
+	unsigned int VMDataOnlyReturnBW;
+	double HostVMInefficiencyFactor;
+	DmlPipe myPipe;
+	SOCParametersList mmSOCParameters;
+	double dummy_unit_vector[DC__NUM_DPP__MAX];
+	double dummy_single[2];
+	enum clock_change_support dummy_dramchange_support;
+	enum dm_fclock_change_support dummy_fclkchange_support;
+	bool dummy_USRRetrainingSupport;
+};
+
+struct dml32_ModeSupportAndSystemConfigurationFull {
+	unsigned int dummy_integer_array[22][DC__NUM_DPP__MAX];
+	double dummy_double_array[2][DC__NUM_DPP__MAX];
+	DmlPipe SurfParameters[DC__NUM_DPP__MAX];
+	double dummy_single[5];
+	double dummy_single2[5];
+	SOCParametersList mSOCParameters;
+	unsigned int MaximumSwathWidthSupportLuma;
+	unsigned int MaximumSwathWidthSupportChroma;
+	double DSTYAfterScaler[DC__NUM_DPP__MAX];
+	double DSTXAfterScaler[DC__NUM_DPP__MAX];
+	double MaxTotalVActiveRDBandwidth;
+	bool dummy_boolean_array[2][DC__NUM_DPP__MAX];
+	enum odm_combine_mode dummy_odm_mode[DC__NUM_DPP__MAX];
+	DmlPipe myPipe;
+	unsigned int dummy_integer[4];
+	unsigned int TotalNumberOfActiveOTG;
+	unsigned int TotalNumberOfActiveHDMIFRL;
+	unsigned int TotalNumberOfActiveDP2p0;
+	unsigned int TotalNumberOfActiveDP2p0Outputs;
+	unsigned int TotalDSCUnitsRequired;
+	unsigned int ReorderingBytes;
+	unsigned int TotalSlots;
+	unsigned int NumberOfDPPDSC;
+	unsigned int NumberOfDPPNoDSC;
+	unsigned int NextPrefetchModeState;
+	bool MPCCombineMethodAsNeededForPStateChangeAndVoltage;
+	bool MPCCombineMethodAsPossible;
+	bool FullFrameMALLPStateMethod;
+	bool SubViewportMALLPStateMethod;
+	bool PhantomPipeMALLPStateMethod;
+	bool NoChroma;
+	bool TotalAvailablePipesSupportNoDSC;
+	bool TotalAvailablePipesSupportDSC;
+	enum odm_combine_mode ODMModeNoDSC;
+	enum odm_combine_mode ODMModeDSC;
+	double RequiredDISPCLKPerSurfaceNoDSC;
+	double RequiredDISPCLKPerSurfaceDSC;
+	double BWOfNonCombinedSurfaceOfMaximumBandwidth;
+	double VMDataOnlyReturnBWPerState;
+	double HostVMInefficiencyFactor;
+	bool dummy_boolean[2];
+};
+
+struct dummy_vars {
+	struct DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation
+	DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation;
+	struct dml32_ModeSupportAndSystemConfigurationFull dml32_ModeSupportAndSystemConfigurationFull;
+};
+
 struct vba_vars_st {
 	ip_params_st ip;
 	soc_bounding_box_st soc;
@@ -145,6 +263,7 @@ struct vba_vars_st {
 	int maxMpcComb;
 	bool UseMaximumVStartup;
 
+	double MaxVRatioPre;
 	double WritebackDISPCLK;
 	double DPPCLKUsingSingleDPPLuma;
 	double DPPCLKUsingSingleDPPChroma;
@@ -169,6 +288,7 @@ struct vba_vars_st {
 	double NextMaxVStartup;
 	double VBlankTime;
 	double SmallestVBlank;
+	enum dm_prefetch_modes AllowForPStateChangeOrStutterInVBlankFinal; // Mode Support only
 	double DCFCLKDeepSleepPerPlane[DC__NUM_DPP__MAX];
 	double EffectiveDETPlusLBLinesLuma;
 	double EffectiveDETPlusLBLinesChroma;
@@ -194,7 +314,9 @@ struct vba_vars_st {
 	unsigned int ActiveDPPs;
 	unsigned int LBLatencyHidingSourceLinesY;
 	unsigned int LBLatencyHidingSourceLinesC;
+	double ActiveDRAMClockChangeLatencyMarginPerState[DC__VOLTAGE_STATES][2][DC__NUM_DPP__MAX];// DML doesn't save active margin per state
 	double ActiveDRAMClockChangeLatencyMargin[DC__NUM_DPP__MAX];
+	double CachedActiveDRAMClockChangeLatencyMargin[DC__NUM_DPP__MAX]; // Cache in dml_get_voltage_level for debug purposes only
 	double MinActiveDRAMClockChangeMargin;
 	double InitFillLevel;
 	double FinalFillMargin;
@@ -211,6 +333,13 @@ struct vba_vars_st {
 	double UrgentLatencyPixelMixedWithVMData;
 	double UrgentLatencyVMDataOnly;
 	double UrgentLatency; // max of the above three
+	double USRRetrainingLatency;
+	double SMNLatency;
+	double FCLKChangeLatency;
+	unsigned int MALLAllocatedForDCNFinal;
+	double MaxAveragePercentOfIdealFabricBWDisplayCanUseInNormalSystemOperation;
+	double MaxAveragePercentOfIdealDRAMBWDisplayCanUseInNormalSystemOperationSTROBE;
+	double PercentOfIdealDRAMBWReceivedAfterUrgLatencySTROBE;
 	double WritebackLatency;
 	double PercentOfIdealDRAMFabricAndSDPPortBWReceivedAfterUrgLatencyPixelDataOnly; // Mode Support
 	double PercentOfIdealDRAMFabricAndSDPPortBWReceivedAfterUrgLatencyPixelMixedWithVMData; // Mode Support
@@ -283,13 +412,47 @@ struct vba_vars_st {
 	double DPPCLKDelayCNVCCursor;
 	double DISPCLKDelaySubtotal;
 	bool ProgressiveToInterlaceUnitInOPP;
+	unsigned int CompressedBufferSegmentSizeInkByteFinal;
+	unsigned int CompbufReservedSpace64B;
+	unsigned int CompbufReservedSpaceZs;
+	unsigned int LineBufferSizeFinal;
+	unsigned int MaximumPixelsPerLinePerDSCUnit;
+	unsigned int AlphaPixelChunkSizeInKByte;
+	double MinPixelChunkSizeBytes;
+	unsigned int DCCMetaBufferSizeBytes;
 	// Pipe/Plane Parameters
+
+	/** @VoltageLevel:
+	 * Every ASIC has a fixed number of DPM states, and some devices might
+	 * have some particular voltage configuration that does not map
+	 * directly to the DPM states. This field tells how many states the
+	 * target device supports; even though this field combines the DPM and
+	 * special SOC voltages, it mostly matches the total number of DPM
+	 * states.
+	 */
 	int VoltageLevel;
 	double FabricClock;
 	double DRAMSpeed;
 	double DISPCLK;
 	double SOCCLK;
 	double DCFCLK;
+	unsigned int MaxTotalDETInKByte;
+	unsigned int MinCompressedBufferSizeInKByte;
+	unsigned int NumberOfActiveSurfaces;
+	bool ViewportStationary[DC__NUM_DPP__MAX];
+	unsigned int RefreshRate[DC__NUM_DPP__MAX];
+	double       OutputBPP[DC__NUM_DPP__MAX];
+	unsigned int GPUVMMinPageSizeKBytes[DC__NUM_DPP__MAX];
+	bool SynchronizeTimingsFinal;
+	bool SynchronizeDRRDisplaysForUCLKPStateChangeFinal;
+	bool ForceOneRowForFrame[DC__NUM_DPP__MAX];
+	unsigned int ViewportXStartY[DC__NUM_DPP__MAX];
+	unsigned int ViewportXStartC[DC__NUM_DPP__MAX];
+	enum dm_rotation_angle SourceRotation[DC__NUM_DPP__MAX];
+	bool DRRDisplay[DC__NUM_DPP__MAX];
+	bool PteBufferMode[DC__NUM_DPP__MAX];
+	enum dm_output_type OutputType[DC__NUM_DPP__MAX];
+	enum dm_output_rate OutputRate[DC__NUM_DPP__MAX];
 
 	unsigned int NumberOfActivePlanes;
 	unsigned int NumberOfDSCSlices[DC__NUM_DPP__MAX];
@@ -354,6 +517,8 @@ struct vba_vars_st {
 	unsigned int CursorBPP[DC__NUM_DPP__MAX][DC__NUM_CURSOR__MAX];
 	bool XFCEnabled[DC__NUM_DPP__MAX];
 	bool ScalerEnabled[DC__NUM_DPP__MAX];
+	unsigned int VBlankNom[DC__NUM_DPP__MAX];
+	bool DisableUnboundRequestIfCompBufReservedSpaceNeedAdjustment;
 
 	// Intermediates/Informational
 	bool ImmediateFlipSupport;
@@ -391,6 +556,16 @@ struct vba_vars_st {
 	double StutterEfficiencyNotIncludingVBlank;
 	double NonUrgentLatencyTolerance;
 	double MinActiveDRAMClockChangeLatencySupported;
+	double Z8StutterEfficiencyBestCase;
+	unsigned int Z8NumberOfStutterBurstsPerFrameBestCase;
+	double Z8StutterEfficiencyNotIncludingVBlankBestCase;
+	double StutterPeriodBestCase;
+	Watermarks      Watermark;
+	bool DCHUBBUB_ARB_CSTATE_MAX_CAP_MODE;
+	unsigned int CompBufReservedSpaceKBytes;
+	unsigned int CompBufReservedSpace64B;
+	unsigned int CompBufReservedSpaceZs;
+	bool CompBufReservedSpaceNeedAdjustment;
 
 	// These are the clocks calcuated by the library but they are not actually
 	// used explicitly. They are fetched by tests and then possibly used. The
@@ -398,6 +573,10 @@ struct vba_vars_st {
 	double DISPCLK_calculated;
 	double DPPCLK_calculated[DC__NUM_DPP__MAX];
 
+	bool ImmediateFlipSupportedSurface[DC__NUM_DPP__MAX];
+
+	bool Use_One_Row_For_Frame[DC__NUM_DPP__MAX];
+	bool Use_One_Row_For_Frame_Flip[DC__NUM_DPP__MAX];
 	unsigned int VUpdateOffsetPix[DC__NUM_DPP__MAX];
 	double VUpdateWidthPix[DC__NUM_DPP__MAX];
 	double VReadyOffsetPix[DC__NUM_DPP__MAX];
@@ -428,6 +607,7 @@ struct vba_vars_st {
 	double DRAMSpeedPerState[DC__VOLTAGE_STATES];
 	double MaxDispclk[DC__VOLTAGE_STATES];
 	int VoltageOverrideLevel;
+	double PHYCLKD32PerState[DC__VOLTAGE_STATES];
 
 	/*outputs*/
 	bool ScaleRatioAndTapsSupport;
@@ -451,6 +631,51 @@ struct vba_vars_st {
 	bool PitchSupport;
 	enum dm_validation_status ValidationStatus[DC__VOLTAGE_STATES];
 
+	/* Mode Support Reason */
+	bool P2IWith420;
+	bool DSCOnlyIfNecessaryWithBPP;
+	bool DSC422NativeNotSupported;
+	bool LinkRateDoesNotMatchDPVersion;
+	bool LinkRateForMultistreamNotIndicated;
+	bool BPPForMultistreamNotIndicated;
+	bool MultistreamWithHDMIOreDP;
+	bool MSOOrODMSplitWithNonDPLink;
+	bool NotEnoughLanesForMSO;
+	bool ViewportExceedsSurface;
+
+	bool ImmediateFlipRequiredButTheRequirementForEachSurfaceIsNotSpecified;
+	bool ImmediateFlipOrHostVMAndPStateWithMALLFullFrameOrPhantomPipe;
+	bool InvalidCombinationOfMALLUseForPStateAndStaticScreen;
+	bool InvalidCombinationOfMALLUseForPState;
+
+	enum dm_output_link_dp_rate OutputLinkDPRate[DC__NUM_DPP__MAX];
+	double PrefetchLinesYThisState[DC__NUM_DPP__MAX];
+	double PrefetchLinesCThisState[DC__NUM_DPP__MAX];
+	double meta_row_bandwidth_this_state[DC__NUM_DPP__MAX];
+	double dpte_row_bandwidth_this_state[DC__NUM_DPP__MAX];
+	double DPTEBytesPerRowThisState[DC__NUM_DPP__MAX];
+	double PDEAndMetaPTEBytesPerFrameThisState[DC__NUM_DPP__MAX];
+	double MetaRowBytesThisState[DC__NUM_DPP__MAX];
+	bool use_one_row_for_frame[DC__VOLTAGE_STATES][2][DC__NUM_DPP__MAX];
+	bool use_one_row_for_frame_flip[DC__VOLTAGE_STATES][2][DC__NUM_DPP__MAX];
+	bool use_one_row_for_frame_this_state[DC__NUM_DPP__MAX];
+	bool use_one_row_for_frame_flip_this_state[DC__NUM_DPP__MAX];
+
+	unsigned int OutputTypeAndRatePerState[DC__VOLTAGE_STATES][DC__NUM_DPP__MAX];
+	double RequiredDISPCLKPerSurface[DC__VOLTAGE_STATES][2][DC__NUM_DPP__MAX];
+	unsigned int MacroTileHeightY[DC__NUM_DPP__MAX];
+	unsigned int MacroTileHeightC[DC__NUM_DPP__MAX];
+	unsigned int MacroTileWidthY[DC__NUM_DPP__MAX];
+	unsigned int MacroTileWidthC[DC__NUM_DPP__MAX];
+	bool ImmediateFlipRequiredFinal;
+	bool DCCProgrammingAssumesScanDirectionUnknownFinal;
+	bool EnoughWritebackUnits;
+	bool ODMCombine2To1SupportCheckOK[DC__VOLTAGE_STATES];
+	bool NumberOfDP2p0Support;
+	unsigned int MaxNumDP2p0Streams;
+	unsigned int MaxNumDP2p0Outputs;
+	enum dm_output_type OutputTypePerState[DC__VOLTAGE_STATES][DC__NUM_DPP__MAX];
+	enum dm_output_rate OutputRatePerState[DC__VOLTAGE_STATES][DC__NUM_DPP__MAX];
 	double WritebackLineBufferLumaBufferSize;
 	double WritebackLineBufferChromaBufferSize;
 	double WritebackMinHSCLRatio;
@@ -544,6 +769,8 @@ struct vba_vars_st {
 	bool DTBCLKRequiredMoreThanSupported[DC__VOLTAGE_STATES];
 	double UrgentRoundTripAndOutOfOrderLatencyPerState[DC__VOLTAGE_STATES];
 	bool ROBSupport[DC__VOLTAGE_STATES][2];
+	//based on rev 99: Dim DCCMetaBufferSizeSupport(NumberOfStates, 1) As Boolean
+	bool DCCMetaBufferSizeSupport[DC__VOLTAGE_STATES][2];
 	bool PTEBufferSizeNotExceeded[DC__VOLTAGE_STATES][2];
 	bool TotalVerticalActiveBandwidthSupport[DC__VOLTAGE_STATES][2];
 	double MaxTotalVerticalActiveAvailableBandwidth[DC__VOLTAGE_STATES][2];
@@ -585,8 +812,6 @@ struct vba_vars_st {
 	double PSCL_FACTOR[DC__NUM_DPP__MAX];
 	double PSCL_FACTOR_CHROMA[DC__NUM_DPP__MAX];
 	double MaximumVStartup[DC__VOLTAGE_STATES][2][DC__NUM_DPP__MAX];
-	unsigned int MacroTileWidthY[DC__NUM_DPP__MAX];
-	unsigned int MacroTileWidthC[DC__NUM_DPP__MAX];
 	double AlignedDCCMetaPitch[DC__NUM_DPP__MAX];
 	double AlignedYPitch[DC__NUM_DPP__MAX];
 	double AlignedCPitch[DC__NUM_DPP__MAX];
@@ -644,8 +869,7 @@ struct vba_vars_st {
 	double         dummy7[DC__NUM_DPP__MAX];
 	double         dummy8[DC__NUM_DPP__MAX];
 	double         dummy13[DC__NUM_DPP__MAX];
-	unsigned int        dummyinteger1ms[DC__NUM_DPP__MAX];
-	double        dummyinteger2ms[DC__NUM_DPP__MAX];
+	double         dummy_double_array[2][DC__NUM_DPP__MAX];
 	unsigned int        dummyinteger3[DC__NUM_DPP__MAX];
 	unsigned int        dummyinteger4[DC__NUM_DPP__MAX];
 	unsigned int        dummyinteger5;
@@ -655,14 +879,8 @@ struct vba_vars_st {
 	unsigned int        dummyinteger9;
 	unsigned int        dummyinteger10;
 	unsigned int        dummyinteger11;
-	unsigned int        dummyinteger12;
-	unsigned int        dummyinteger30;
-	unsigned int        dummyinteger31;
-	unsigned int        dummyinteger32;
-	unsigned int        dummyintegerarr1[DC__NUM_DPP__MAX];
-	unsigned int        dummyintegerarr2[DC__NUM_DPP__MAX];
-	unsigned int        dummyintegerarr3[DC__NUM_DPP__MAX];
-	unsigned int        dummyintegerarr4[DC__NUM_DPP__MAX];
+	unsigned int        dummy_integer_array[8][DC__NUM_DPP__MAX];
+
 	bool           dummysinglestring;
 	bool           SingleDPPViewportSizeSupportPerPlane[DC__NUM_DPP__MAX];
 	double         PlaneRequiredDISPCLKWithODMCombine2To1;
@@ -802,7 +1020,6 @@ struct vba_vars_st {
 	double TimePerChromaMetaChunkFlip[DC__NUM_DPP__MAX];
 	unsigned int DCCCMaxUncompressedBlock[DC__NUM_DPP__MAX];
 	unsigned int DCCCMaxCompressedBlock[DC__NUM_DPP__MAX];
-	unsigned int DCCCIndependent64ByteBlock[DC__NUM_DPP__MAX];
 	double VStartupMargin;
 	bool NotEnoughTimeForDynamicMetadata[DC__NUM_DPP__MAX];
 
@@ -835,6 +1052,7 @@ struct vba_vars_st {
 	double MinFullDETBufferingTime;
 	double AverageReadBandwidthGBytePerSecond;
 	bool   FirstMainPlane;
+	bool NotEnoughDETSwathFillLatencyHiding;
 
 	unsigned int ViewportWidthChroma[DC__NUM_DPP__MAX];
 	unsigned int ViewportHeightChroma[DC__NUM_DPP__MAX];
@@ -893,8 +1111,8 @@ struct vba_vars_st {
 	double meta_row_bandwidth[DC__VOLTAGE_STATES][2][DC__NUM_DPP__MAX];
 	double DETBufferSizeYAllStates[DC__VOLTAGE_STATES][2][DC__NUM_DPP__MAX];
 	double DETBufferSizeCAllStates[DC__VOLTAGE_STATES][2][DC__NUM_DPP__MAX];
-	int swath_width_luma_ub_all_states[DC__VOLTAGE_STATES][2][DC__NUM_DPP__MAX];
-	int swath_width_chroma_ub_all_states[DC__VOLTAGE_STATES][2][DC__NUM_DPP__MAX];
+	unsigned int swath_width_luma_ub_all_states[DC__VOLTAGE_STATES][2][DC__NUM_DPP__MAX];
+	unsigned int swath_width_chroma_ub_all_states[DC__VOLTAGE_STATES][2][DC__NUM_DPP__MAX];
 	bool NotUrgentLatencyHiding[DC__VOLTAGE_STATES][2];
 	unsigned int SwathHeightYAllStates[DC__VOLTAGE_STATES][2][DC__NUM_DPP__MAX];
 	unsigned int SwathHeightCAllStates[DC__VOLTAGE_STATES][2][DC__NUM_DPP__MAX];
@@ -907,8 +1125,6 @@ struct vba_vars_st {
 	double WritebackDelayTime[DC__NUM_DPP__MAX];
 	unsigned int DCCYIndependentBlock[DC__NUM_DPP__MAX];
 	unsigned int DCCCIndependentBlock[DC__NUM_DPP__MAX];
-	unsigned int dummyinteger15;
-	unsigned int dummyinteger16;
 	unsigned int dummyinteger17;
 	unsigned int dummyinteger18;
 	unsigned int dummyinteger19;
@@ -949,7 +1165,7 @@ struct vba_vars_st {
 	double UrgBurstFactorLumaPre[DC__NUM_DPP__MAX];
 	double UrgBurstFactorChromaPre[DC__NUM_DPP__MAX];
 	bool NotUrgentLatencyHidingPre[DC__NUM_DPP__MAX];
-	bool LinkCapacitySupport[DC__NUM_DPP__MAX];
+	bool LinkCapacitySupport[DC__VOLTAGE_STATES];
 	bool VREADY_AT_OR_AFTER_VSYNC[DC__NUM_DPP__MAX];
 	unsigned int MIN_DST_Y_NEXT_START[DC__NUM_DPP__MAX];
 	unsigned int VFrontPorch[DC__NUM_DPP__MAX];
@@ -969,6 +1185,59 @@ struct vba_vars_st {
 	int Z8NumberOfStutterBurstsPerFrame;
 	unsigned int MaximumDSCBitsPerComponent;
 	unsigned int NotEnoughUrgentLatencyHidingA[DC__VOLTAGE_STATES][2];
+	double ReadBandwidthSurfaceLuma[DC__NUM_DPP__MAX];
+	double ReadBandwidthSurfaceChroma[DC__NUM_DPP__MAX];
+	double SurfaceRequiredDISPCLKWithoutODMCombine;
+	double SurfaceRequiredDISPCLK;
+	double MinActiveFCLKChangeLatencySupported;
+	int MinVoltageLevel;
+	int MaxVoltageLevel;
+	unsigned int TotalNumberOfSingleDPPSurfaces[DC__VOLTAGE_STATES][2];
+	unsigned int CompressedBufferSizeInkByteAllStates[DC__VOLTAGE_STATES][2];
+	unsigned int DETBufferSizeInKByteAllStates[DC__VOLTAGE_STATES][2][DC__NUM_DPP__MAX];
+	unsigned int DETBufferSizeInKByteThisState[DC__NUM_DPP__MAX];
+	unsigned int SurfaceSizeInMALL[DC__NUM_DPP__MAX];
+	bool ExceededMALLSize;
+	bool PTE_BUFFER_MODE[DC__NUM_DPP__MAX];
+	unsigned int BIGK_FRAGMENT_SIZE[DC__NUM_DPP__MAX];
+	unsigned int CompressedBufferSizeInkByteThisState;
+	enum dm_fclock_change_support FCLKChangeSupport[DC__VOLTAGE_STATES][2];
+	bool USRRetrainingSupport[DC__VOLTAGE_STATES][2];
+	enum dm_use_mall_for_pstate_change_mode UsesMALLForPStateChange[DC__NUM_DPP__MAX];
+	bool UnboundedRequestEnabledAllStates[DC__VOLTAGE_STATES][2];
+	bool SingleDPPViewportSizeSupportPerSurface[DC__NUM_DPP__MAX];
+	enum dm_use_mall_for_static_screen_mode UseMALLForStaticScreen[DC__NUM_DPP__MAX];
+	bool UnboundedRequestEnabledThisState;
+	bool DRAMClockChangeRequirementFinal;
+	bool FCLKChangeRequirementFinal;
+	bool USRRetrainingRequiredFinal;
+	unsigned int DETSizeOverride[DC__NUM_DPP__MAX];
+	unsigned int nomDETInKByte;
+	enum mpc_combine_affinity  MPCCombineUse[DC__NUM_DPP__MAX];
+	bool MPCCombineMethodIncompatible;
+	unsigned int RequiredSlots[DC__VOLTAGE_STATES][DC__NUM_DPP__MAX];
+	bool ExceededMultistreamSlots[DC__VOLTAGE_STATES];
+	enum odm_combine_policy ODMUse[DC__NUM_DPP__MAX];
+	unsigned int OutputMultistreamId[DC__NUM_DPP__MAX];
+	bool OutputMultistreamEn[DC__NUM_DPP__MAX];
+	bool UsesMALLForStaticScreen[DC__NUM_DPP__MAX];
+	double MaxActiveDRAMClockChangeLatencySupported[DC__NUM_DPP__MAX];
+	double WritebackAllowFCLKChangeEndPosition[DC__NUM_DPP__MAX];
+	bool PTEBufferSizeNotExceededPerState[DC__NUM_DPP__MAX]; // new in DML32
+	bool DCCMetaBufferSizeNotExceededPerState[DC__NUM_DPP__MAX]; // new in DML32
+	bool NotEnoughDSCSlices[DC__VOLTAGE_STATES];
+	bool PixelsPerLinePerDSCUnitSupport[DC__VOLTAGE_STATES];
+	bool DCCMetaBufferSizeNotExceeded[DC__VOLTAGE_STATES][2];
+	unsigned int dpte_row_height_linear[DC__NUM_DPP__MAX];
+	unsigned int dpte_row_height_linear_chroma[DC__NUM_DPP__MAX];
+	unsigned int BlockHeightY[DC__NUM_DPP__MAX];
+	unsigned int BlockHeightC[DC__NUM_DPP__MAX];
+	unsigned int BlockWidthY[DC__NUM_DPP__MAX];
+	unsigned int BlockWidthC[DC__NUM_DPP__MAX];
+	unsigned int SubViewportLinesNeededInMALL[DC__NUM_DPP__MAX];
+	bool VActiveBandwithSupport[DC__VOLTAGE_STATES][2];
+	bool NotEnoughDETSwathFillLatencyHidingPerState[DC__VOLTAGE_STATES][2];
+	struct dummy_vars dummy_vars;
 };
 
 bool CalculateMinAndMaxPrefetchMode(
diff --git a/drivers/gpu/drm/amd/display/dc/dml/display_rq_dlg_helpers.c b/drivers/gpu/drm/amd/display/dc/dml/display_rq_dlg_helpers.c
index 71ea503cb32f..12ff65b6a7e5 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/display_rq_dlg_helpers.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/display_rq_dlg_helpers.c
@@ -122,17 +122,6 @@ void print__data_rq_misc_params_st(struct display_mode_lib *mode_lib, const stru
 	dml_print("DML_RQ_DLG_CALC: =====================================\n");
 }
 
-void print__rq_dlg_params_st(struct display_mode_lib *mode_lib, const struct _vcs_dpi_display_rq_dlg_params_st *rq_dlg_param)
-{
-	dml_print("DML_RQ_DLG_CALC: =====================================\n");
-	dml_print("DML_RQ_DLG_CALC: DISPLAY_RQ_DLG_PARAM_ST\n");
-	dml_print("DML_RQ_DLG_CALC:  <LUMA>\n");
-	print__data_rq_dlg_params_st(mode_lib, &rq_dlg_param->rq_l);
-	dml_print("DML_RQ_DLG_CALC:  <CHROMA>\n");
-	print__data_rq_dlg_params_st(mode_lib, &rq_dlg_param->rq_c);
-	dml_print("DML_RQ_DLG_CALC: =====================================\n");
-}
-
 void print__dlg_sys_params_st(struct display_mode_lib *mode_lib, const struct _vcs_dpi_display_dlg_sys_params_st *dlg_sys_param)
 {
 	dml_print("DML_RQ_DLG_CALC: =====================================\n");
@@ -142,9 +131,6 @@ void print__dlg_sys_params_st(struct display_mode_lib *mode_lib, const struct _v
 	dml_print("DML_RQ_DLG_CALC:    t_sr_wm_us           = %3.2f\n", dlg_sys_param->t_sr_wm_us);
 	dml_print("DML_RQ_DLG_CALC:    t_extra_us           = %3.2f\n", dlg_sys_param->t_extra_us);
 	dml_print(
-			"DML_RQ_DLG_CALC:    t_srx_delay_us       = %3.2f\n",
-			dlg_sys_param->t_srx_delay_us);
-	dml_print(
 			"DML_RQ_DLG_CALC:    deepsleep_dcfclk_mhz = %3.2f\n",
 			dlg_sys_param->deepsleep_dcfclk_mhz);
 	dml_print(
diff --git a/drivers/gpu/drm/amd/display/dc/dml/display_rq_dlg_helpers.h b/drivers/gpu/drm/amd/display/dc/dml/display_rq_dlg_helpers.h
index ebcd717744e5..2bc64c4081dc 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/display_rq_dlg_helpers.h
+++ b/drivers/gpu/drm/amd/display/dc/dml/display_rq_dlg_helpers.h
@@ -35,7 +35,6 @@ void print__rq_params_st(struct display_mode_lib *mode_lib, const struct _vcs_dp
 void print__data_rq_sizing_params_st(struct display_mode_lib *mode_lib, const struct _vcs_dpi_display_data_rq_sizing_params_st *rq_sizing);
 void print__data_rq_dlg_params_st(struct display_mode_lib *mode_lib, const struct _vcs_dpi_display_data_rq_dlg_params_st *rq_dlg_param);
 void print__data_rq_misc_params_st(struct display_mode_lib *mode_lib, const struct _vcs_dpi_display_data_rq_misc_params_st *rq_misc_param);
-void print__rq_dlg_params_st(struct display_mode_lib *mode_lib, const struct _vcs_dpi_display_rq_dlg_params_st *rq_dlg_param);
 void print__dlg_sys_params_st(struct display_mode_lib *mode_lib, const struct _vcs_dpi_display_dlg_sys_params_st *dlg_sys_param);
 
 void print__data_rq_regs_st(struct display_mode_lib *mode_lib, const struct _vcs_dpi_display_data_rq_regs_st *rq_regs);
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dml1_display_rq_dlg_calc.c b/drivers/gpu/drm/amd/display/dc/dml/dml1_display_rq_dlg_calc.c
index 59dc2c5b58dd..88dc2b97e7bf 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dml1_display_rq_dlg_calc.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/dml1_display_rq_dlg_calc.c
@@ -39,7 +39,7 @@
 
 static unsigned int get_bytes_per_element(enum source_format_class source_format, bool is_chroma)
 {
-	unsigned int ret_val = 0;
+	unsigned int ret_val = 1;
 
 	if (source_format == dm_444_16) {
 		if (!is_chroma)
@@ -559,12 +559,11 @@ static void get_surf_rq_param(
 		const struct _vcs_dpi_display_pipe_source_params_st *pipe_src_param,
 		bool is_chroma)
 {
-	bool mode_422 = 0;
 	unsigned int vp_width = 0;
 	unsigned int vp_height = 0;
 	unsigned int data_pitch = 0;
 	unsigned int meta_pitch = 0;
-	unsigned int ppe = mode_422 ? 2 : 1;
+	unsigned int ppe = 1;
 	bool surf_linear;
 	bool surf_vert;
 	unsigned int bytes_per_element;
@@ -1331,10 +1330,6 @@ void dml1_rq_dlg_get_dlg_params(
 	if (dual_plane)
 		DTRACE("DLG: %s: swath_height_c     = %d", __func__, swath_height_c);
 
-	DTRACE(
-			"DLG: %s: t_srx_delay_us     = %3.2f",
-			__func__,
-			(double) dlg_sys_param->t_srx_delay_us);
 	DTRACE("DLG: %s: line_time_in_us    = %3.2f", __func__, (double) line_time_in_us);
 	DTRACE("DLG: %s: vupdate_offset     = %d", __func__, vupdate_offset);
 	DTRACE("DLG: %s: vupdate_width      = %d", __func__, vupdate_width);
@@ -1600,11 +1595,6 @@ void dml1_rq_dlg_get_dlg_params(
 		swath_width_pixels_ub_c = swath_width_ub_c * 1;
 	}
 
-	hscale_pixel_rate_l = 0.;
-	hscale_pixel_rate_c = 0.;
-	min_hratio_fact_l = 1.0;
-	min_hratio_fact_c = 1.0;
-
 	if (htaps_l <= 1)
 		min_hratio_fact_l = 2.0;
 	else if (htaps_l <= 6) {
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dml_inline_defs.h b/drivers/gpu/drm/amd/display/dc/dml/dml_inline_defs.h
index 479d7d83220c..6b2ab4ec2b5f 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dml_inline_defs.h
+++ b/drivers/gpu/drm/amd/display/dc/dml/dml_inline_defs.h
@@ -66,24 +66,23 @@ static inline double dml_max5(double a, double b, double c, double d, double e)
 
 static inline double dml_ceil(double a, double granularity)
 {
+	if (granularity == 0)
+		return 0;
 	return (double) dcn_bw_ceil2(a, granularity);
 }
 
 static inline double dml_floor(double a, double granularity)
 {
+	if (granularity == 0)
+		return 0;
 	return (double) dcn_bw_floor2(a, granularity);
 }
 
 static inline double dml_round(double a)
 {
-	double round_pt = 0.5;
-	double ceil = dml_ceil(a, 1);
-	double floor = dml_floor(a, 1);
+	const double round_pt = 0.5;
 
-	if (a - floor >= round_pt)
-		return ceil;
-	else
-		return floor;
+	return dml_floor(a + round_pt, 1);
 }
 
 /* float
@@ -119,11 +118,15 @@ static inline double dml_ceil_2(double f)
 
 static inline double dml_ceil_ex(double x, double granularity)
 {
+	if (granularity == 0)
+		return 0;
 	return (double) dcn_bw_ceil2(x, granularity);
 }
 
 static inline double dml_floor_ex(double x, double granularity)
 {
+	if (granularity == 0)
+		return 0;
 	return (double) dcn_bw_floor2(x, granularity);
 }
 
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dsc/qp_tables.h b/drivers/gpu/drm/amd/display/dc/dml/dsc/qp_tables.h
index e5fac9f4181d..dcff0dd2b6a1 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dsc/qp_tables.h
+++ b/drivers/gpu/drm/amd/display/dc/dml/dsc/qp_tables.h
@@ -25,7 +25,7 @@
  */
 
 
-const qp_table   qp_table_422_10bpc_min = {
+static const qp_table   qp_table_422_10bpc_min = {
 	{   6, { 0, 4, 5, 6, 6, 6, 6, 7, 7, 8, 9, 9, 9, 12, 16} },
 	{ 6.5, { 0, 4, 5, 6, 6, 6, 6, 7, 7, 8, 9, 9, 9, 12, 16} },
 	{   7, { 0, 4, 5, 6, 6, 6, 6, 7, 7, 7, 9, 9, 9, 11, 15} },
@@ -58,7 +58,7 @@ const qp_table   qp_table_422_10bpc_min = {
 };
 
 
-const qp_table   qp_table_444_8bpc_max = {
+static const qp_table   qp_table_444_8bpc_max = {
 	{   6, { 4, 6, 8, 8, 9, 9, 9, 10, 11, 12, 12, 12, 12, 13, 15} },
 	{ 6.5, { 4, 6, 7, 8, 8, 8, 9, 10, 11, 11, 12, 12, 12, 13, 15} },
 	{   7, { 4, 5, 7, 7, 8, 8, 8, 9, 10, 11, 11, 12, 12, 13, 14} },
@@ -99,7 +99,7 @@ const qp_table   qp_table_444_8bpc_max = {
 };
 
 
-const qp_table   qp_table_420_12bpc_max = {
+static const qp_table   qp_table_420_12bpc_max = {
 	{   4, {11, 12, 13, 14, 15, 15, 15, 16, 17, 18, 18, 19, 19, 21, 22} },
 	{ 4.5, {10, 11, 12, 13, 14, 15, 15, 16, 17, 18, 18, 19, 19, 20, 21} },
 	{   5, { 9, 11, 12, 13, 14, 15, 15, 16, 17, 17, 18, 18, 19, 20, 21} },
@@ -132,7 +132,7 @@ const qp_table   qp_table_420_12bpc_max = {
 };
 
 
-const qp_table   qp_table_444_10bpc_min = {
+static const qp_table   qp_table_444_10bpc_min = {
 	{   6, { 0, 4, 7, 7, 9, 9, 9, 9, 9, 10, 10, 10, 10, 12, 18} },
 	{ 6.5, { 0, 4, 6, 7, 8, 8, 9, 9, 9, 9, 10, 10, 10, 12, 18} },
 	{   7, { 0, 4, 6, 6, 8, 8, 8, 8, 8, 9, 9, 10, 10, 12, 17} },
@@ -185,7 +185,7 @@ const qp_table   qp_table_444_10bpc_min = {
 };
 
 
-const qp_table   qp_table_420_8bpc_max = {
+static const qp_table   qp_table_420_8bpc_max = {
 	{   4, { 4, 4, 5, 6, 7, 7, 7, 8, 9, 10, 10, 11, 11, 13, 14} },
 	{ 4.5, { 4, 4, 5, 6, 7, 7, 7, 8, 9, 10, 10, 11, 11, 12, 13} },
 	{   5, { 3, 4, 5, 6, 7, 7, 7, 8, 9, 9, 10, 10, 11, 12, 13} },
@@ -206,7 +206,7 @@ const qp_table   qp_table_420_8bpc_max = {
 };
 
 
-const qp_table   qp_table_444_8bpc_min = {
+static const qp_table   qp_table_444_8bpc_min = {
 	{   6, { 0, 1, 3, 3, 5, 5, 5, 5, 5, 6, 6, 6, 6, 9, 14} },
 	{ 6.5, { 0, 1, 2, 3, 4, 4, 5, 5, 5, 5, 6, 6, 6, 9, 14} },
 	{   7, { 0, 0, 2, 2, 4, 4, 4, 4, 4, 5, 5, 6, 6, 9, 13} },
@@ -247,7 +247,7 @@ const qp_table   qp_table_444_8bpc_min = {
 };
 
 
-const qp_table   qp_table_444_12bpc_min = {
+static const qp_table   qp_table_444_12bpc_min = {
 	{   6, { 0, 5, 11, 11, 13, 13, 13, 13, 13, 14, 14, 14, 14, 17, 22} },
 	{ 6.5, { 0, 5, 10, 11, 12, 12, 13, 13, 13, 13, 14, 14, 14, 17, 22} },
 	{   7, { 0, 5, 10, 10, 12, 12, 12, 12, 12, 13, 13, 14, 14, 17, 21} },
@@ -312,7 +312,7 @@ const qp_table   qp_table_444_12bpc_min = {
 };
 
 
-const qp_table   qp_table_420_12bpc_min = {
+static const qp_table   qp_table_420_12bpc_min = {
 	{   4, { 0, 4, 9, 10, 11, 11, 11, 11, 11, 11, 13, 13, 13, 15, 21} },
 	{ 4.5, { 0, 4, 8, 9, 10, 11, 11, 11, 11, 11, 13, 13, 13, 15, 20} },
 	{   5, { 0, 4, 8, 9, 10, 11, 11, 11, 11, 11, 13, 13, 13, 15, 20} },
@@ -345,7 +345,7 @@ const qp_table   qp_table_420_12bpc_min = {
 };
 
 
-const qp_table   qp_table_422_12bpc_min = {
+static const qp_table   qp_table_422_12bpc_min = {
 	{   6, { 0, 4, 9, 10, 11, 11, 11, 11, 11, 11, 13, 13, 13, 16, 20} },
 	{ 6.5, { 0, 4, 9, 10, 11, 11, 11, 11, 11, 11, 13, 13, 13, 16, 20} },
 	{   7, { 0, 4, 9, 10, 11, 11, 11, 11, 11, 11, 13, 13, 13, 15, 19} },
@@ -386,7 +386,7 @@ const qp_table   qp_table_422_12bpc_min = {
 };
 
 
-const qp_table   qp_table_422_12bpc_max = {
+static const qp_table   qp_table_422_12bpc_max = {
 	{   6, {12, 12, 13, 14, 15, 15, 15, 16, 17, 18, 18, 19, 19, 20, 21} },
 	{ 6.5, {12, 12, 13, 14, 15, 15, 15, 16, 17, 18, 18, 19, 19, 20, 21} },
 	{   7, {11, 12, 13, 14, 15, 15, 15, 16, 17, 17, 18, 18, 19, 19, 20} },
@@ -427,7 +427,7 @@ const qp_table   qp_table_422_12bpc_max = {
 };
 
 
-const qp_table   qp_table_444_12bpc_max = {
+static const qp_table   qp_table_444_12bpc_max = {
 	{   6, {12, 14, 16, 16, 17, 17, 17, 18, 19, 20, 20, 20, 20, 21, 23} },
 	{ 6.5, {12, 14, 15, 16, 16, 16, 17, 18, 19, 19, 20, 20, 20, 21, 23} },
 	{   7, {12, 13, 15, 15, 16, 16, 16, 17, 18, 19, 19, 20, 20, 21, 22} },
@@ -492,7 +492,7 @@ const qp_table   qp_table_444_12bpc_max = {
 };
 
 
-const qp_table   qp_table_420_8bpc_min = {
+static const qp_table   qp_table_420_8bpc_min = {
 	{   4, { 0, 0, 1, 1, 3, 3, 3, 3, 3, 3, 5, 5, 5, 9, 13} },
 	{ 4.5, { 0, 0, 1, 1, 3, 3, 3, 3, 3, 3, 5, 5, 5, 8, 12} },
 	{   5, { 0, 0, 1, 1, 3, 3, 3, 3, 3, 3, 5, 5, 5, 8, 12} },
@@ -513,7 +513,7 @@ const qp_table   qp_table_420_8bpc_min = {
 };
 
 
-const qp_table   qp_table_422_8bpc_min = {
+static const qp_table   qp_table_422_8bpc_min = {
 	{   6, { 0, 0, 1, 2, 3, 3, 3, 3, 3, 3, 5, 5, 5, 8, 12} },
 	{ 6.5, { 0, 0, 1, 2, 3, 3, 3, 3, 3, 3, 5, 5, 5, 8, 12} },
 	{   7, { 0, 0, 1, 2, 3, 3, 3, 3, 3, 3, 5, 5, 5, 7, 11} },
@@ -538,7 +538,7 @@ const qp_table   qp_table_422_8bpc_min = {
 };
 
 
-const qp_table   qp_table_422_10bpc_max = {
+static const qp_table   qp_table_422_10bpc_max = {
 	{   6, { 8, 8, 9, 10, 11, 11, 11, 12, 13, 14, 14, 15, 15, 16, 17} },
 	{ 6.5, { 8, 8, 9, 10, 11, 11, 11, 12, 13, 14, 14, 15, 15, 16, 17} },
 	{   7, { 7, 8, 9, 10, 11, 11, 11, 12, 13, 13, 14, 14, 15, 15, 16} },
@@ -571,7 +571,7 @@ const qp_table   qp_table_422_10bpc_max = {
 };
 
 
-const qp_table qp_table_420_10bpc_max = {
+static const qp_table qp_table_420_10bpc_max = {
 	{   4, { 8, 8, 9, 10, 11, 11, 11, 12, 13, 14, 14, 15, 15, 17, 18} },
 	{ 4.5, { 8, 8, 9, 10, 11, 11, 11, 12, 13, 14, 14, 15, 15, 16, 17} },
 	{   5, { 7, 8, 9, 10, 11, 11, 11, 12, 13, 13, 14, 14, 15, 16, 17} },
@@ -598,7 +598,7 @@ const qp_table qp_table_420_10bpc_max = {
 };
 
 
-const qp_table   qp_table_420_10bpc_min = {
+static const qp_table   qp_table_420_10bpc_min = {
 	{   4, { 0, 4, 4, 5, 7, 7, 7, 7, 7, 7, 9, 9, 9, 13, 17} },
 	{ 4.5, { 0, 4, 4, 5, 7, 7, 7, 7, 7, 7, 9, 9, 9, 12, 16} },
 	{   5, { 0, 4, 4, 5, 7, 7, 7, 7, 7, 7, 9, 9, 9, 12, 16} },
@@ -625,7 +625,7 @@ const qp_table   qp_table_420_10bpc_min = {
 };
 
 
-const qp_table   qp_table_444_10bpc_max = {
+static const qp_table   qp_table_444_10bpc_max = {
 	{   6, { 8, 10, 12, 12, 13, 13, 13, 14, 15, 16, 16, 16, 16, 17, 19} },
 	{ 6.5, { 8, 10, 11, 12, 12, 12, 13, 14, 15, 15, 16, 16, 16, 17, 19} },
 	{   7, { 8, 9, 11, 11, 12, 12, 12, 13, 14, 15, 15, 16, 16, 17, 18} },
@@ -678,7 +678,7 @@ const qp_table   qp_table_444_10bpc_max = {
 };
 
 
-const qp_table   qp_table_422_8bpc_max = {
+static const qp_table   qp_table_422_8bpc_max = {
 	{   6, { 4, 4, 5, 6, 7, 7, 7, 8, 9, 10, 10, 11, 11, 12, 13} },
 	{ 6.5, { 4, 4, 5, 6, 7, 7, 7, 8, 9, 10, 10, 11, 11, 12, 13} },
 	{   7, { 3, 4, 5, 6, 7, 7, 7, 8, 9, 9, 10, 10, 11, 11, 12} },
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dsc/rc_calc_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dsc/rc_calc_fpu.c
index 3ee858f311d1..ef75eb7d5adc 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dsc/rc_calc_fpu.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/dsc/rc_calc_fpu.c
@@ -61,16 +61,6 @@ static double dsc_roundf(double num)
 	return (int)(num);
 }
 
-static double dsc_ceil(double num)
-{
-	double retval = (int)num;
-
-	if (retval != num && num > 0)
-		retval = num + 1;
-
-	return (int)retval;
-}
-
 static void get_qp_set(qp_set qps, enum colour_mode cm, enum bits_per_comp bpc,
 		       enum max_min max_min, float bpp)
 {
@@ -78,7 +68,7 @@ static void get_qp_set(qp_set qps, enum colour_mode cm, enum bits_per_comp bpc,
 	int sel = table_hash(mode, bpc, max_min);
 	int table_size = 0;
 	int index;
-	const struct qp_entry *table = 0L;
+	const struct qp_entry *table = NULL;
 
 	// alias enum
 	enum { min = DAL_MM_MIN, max = DAL_MM_MAX };
@@ -103,7 +93,7 @@ static void get_qp_set(qp_set qps, enum colour_mode cm, enum bits_per_comp bpc,
 		TABLE_CASE(420, 12, min);
 	}
 
-	if (table == 0)
+	if (!table)
 		return;
 
 	index = (bpp - table[0].bpp) * 2;
@@ -268,24 +258,3 @@ void _do_calc_rc_params(struct rc_params *rc,
 	rc->rc_buf_thresh[13] = 8064;
 }
 
-u32 _do_bytes_per_pixel_calc(int slice_width,
-		u16 drm_bpp,
-		bool is_navite_422_or_420)
-{
-	float bpp;
-	u32 bytes_per_pixel;
-	double d_bytes_per_pixel;
-
-	dc_assert_fp_enabled();
-
-	bpp = ((float)drm_bpp / 16.0);
-	d_bytes_per_pixel = dsc_ceil(bpp * slice_width / 8.0) / slice_width;
-	// TODO: Make sure the formula for calculating this is precise (ceiling
-	// vs. floor, and at what point they should be applied)
-	if (is_navite_422_or_420)
-		d_bytes_per_pixel /= 2;
-
-	bytes_per_pixel = (u32)dsc_ceil(d_bytes_per_pixel * 0x10000000);
-
-	return bytes_per_pixel;
-}
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dsc/rc_calc_fpu.h b/drivers/gpu/drm/amd/display/dc/dml/dsc/rc_calc_fpu.h
index b93b95409fbe..d7cd8cc24758 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dsc/rc_calc_fpu.h
+++ b/drivers/gpu/drm/amd/display/dc/dml/dsc/rc_calc_fpu.h
@@ -27,7 +27,7 @@
 #define __RC_CALC_FPU_H__
 
 #include "os_types.h"
-#include <drm/drm_dsc.h>
+#include <drm/display/drm_dsc.h>
 
 #define QP_SET_SIZE 15
 
@@ -78,10 +78,6 @@ struct qp_entry {
 
 typedef struct qp_entry qp_table[];
 
-u32 _do_bytes_per_pixel_calc(int slice_width,
-		u16 drm_bpp,
-		bool is_navite_422_or_420);
-
 void _do_calc_rc_params(struct rc_params *rc,
 		enum colour_mode cm,
 		enum bits_per_comp bpc,
diff --git a/drivers/gpu/drm/amd/display/dc/dml2_0/Makefile b/drivers/gpu/drm/amd/display/dc/dml2_0/Makefile
new file mode 100644
index 000000000000..97e068b6bf6b
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml2_0/Makefile
@@ -0,0 +1,140 @@
+# SPDX-License-Identifier: MIT */
+#
+# Copyright 2023 Advanced Micro Devices, Inc.
+#
+# Permission is hereby granted, free of charge, to any person obtaining a
+# copy of this software and associated documentation files (the "Software"),
+# to deal in the Software without restriction, including without limitation
+# the rights to use, copy, modify, merge, publish, distribute, sublicense,
+# and/or sell copies of the Software, and to permit persons to whom the
+# Software is furnished to do so, subject to the following conditions:
+#
+# The above copyright notice and this permission notice shall be included in
+# all copies or substantial portions of the Software.
+#
+# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+# THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+# OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+# ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+# OTHER DEALINGS IN THE SOFTWARE.
+#
+# Authors: AMD
+#
+# Makefile for dml2.
+
+dml2_ccflags := $(CC_FLAGS_FPU)
+dml2_rcflags := $(CC_FLAGS_NO_FPU)
+
+ifneq ($(CONFIG_FRAME_WARN),0)
+    ifeq ($(filter y,$(CONFIG_KASAN)$(CONFIG_KCSAN)),y)
+        ifeq ($(CONFIG_CC_IS_CLANG)$(CONFIG_COMPILE_TEST),yy)
+            frame_warn_limit := 4096
+        else
+            frame_warn_limit := 3072
+        endif
+    else
+        frame_warn_limit := 2056
+    endif
+
+    ifeq ($(call test-lt, $(CONFIG_FRAME_WARN), $(frame_warn_limit)),y)
+        frame_warn_flag := -Wframe-larger-than=$(frame_warn_limit)
+    endif
+endif
+
+subdir-ccflags-y += -I$(FULL_AMD_DISPLAY_PATH)/dc/dml2_0
+subdir-ccflags-y += -I$(FULL_AMD_DISPLAY_PATH)/dc/dml2_0/dml21/src/dml2_core
+subdir-ccflags-y += -I$(FULL_AMD_DISPLAY_PATH)/dc/dml2_0/dml21/src/dml2_mcg/
+subdir-ccflags-y += -I$(FULL_AMD_DISPLAY_PATH)/dc/dml2_0/dml21/src/dml2_dpmm/
+subdir-ccflags-y += -I$(FULL_AMD_DISPLAY_PATH)/dc/dml2_0/dml21/src/dml2_pmo/
+subdir-ccflags-y += -I$(FULL_AMD_DISPLAY_PATH)/dc/dml2_0/dml21/src/dml2_standalone_libraries/
+subdir-ccflags-y += -I$(FULL_AMD_DISPLAY_PATH)/dc/dml2_0/dml21/src/inc
+subdir-ccflags-y += -I$(FULL_AMD_DISPLAY_PATH)/dc/dml2_0/dml21/inc
+subdir-ccflags-y += -I$(FULL_AMD_DISPLAY_PATH)/dc/dml2_0/dml21/
+
+CFLAGS_$(AMDDALPATH)/dc/dml2_0/display_mode_core.o := $(dml2_ccflags) $(frame_warn_flag)
+CFLAGS_$(AMDDALPATH)/dc/dml2_0/display_mode_util.o := $(dml2_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml2_0/dml2_wrapper.o := $(dml2_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml2_0/dml2_utils.o := $(dml2_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml2_0/dml2_policy.o := $(dml2_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml2_0/dml2_translation_helper.o := $(dml2_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml2_0/dml2_mall_phantom.o := $(dml2_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml2_0/dml_display_rq_dlg_calc.o := $(dml2_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml2_0/dml2_dc_resource_mgmt.o := $(dml2_ccflags)
+
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2_0/display_mode_core.o := $(dml2_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2_0/display_mode_util.o := $(dml2_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2_0/dml2_wrapper.o := $(dml2_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2_0/dml2_utils.o := $(dml2_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2_0/dml2_policy.o := $(dml2_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2_0/dml2_translation_helper.o := $(dml2_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2_0/dml2_mall_phantom.o := $(dml2_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2_0/dml_display_rq_dlg_calc.o := $(dml2_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2_0/dml2_dc_resource_mgmt.o := $(dml2_rcflags)
+
+DML2 = display_mode_core.o display_mode_util.o dml2_wrapper.o \
+		dml2_utils.o dml2_policy.o dml2_translation_helper.o dml2_dc_resource_mgmt.o dml2_mall_phantom.o \
+		dml_display_rq_dlg_calc.o
+
+AMD_DAL_DML2 = $(addprefix $(AMDDALPATH)/dc/dml2_0/,$(DML2))
+
+AMD_DISPLAY_FILES += $(AMD_DAL_DML2)
+
+CFLAGS_$(AMDDALPATH)/dc/dml2_0/dml21/src/dml2_core/dml2_core_dcn4.o := $(dml2_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml2_0/dml21/src/dml2_core/dml2_core_dcn4_calcs.o := $(dml2_ccflags) $(frame_warn_flag)
+CFLAGS_$(AMDDALPATH)/dc/dml2_0/dml21/src/dml2_core/dml2_core_utils.o := $(dml2_ccflags) $(frame_warn_flag)
+CFLAGS_$(AMDDALPATH)/dc/dml2_0/dml21/src/dml2_top/dml2_top_interfaces.o := $(dml2_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml2_0/dml21/src/dml2_top/dml2_top_soc15.o := $(dml2_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml2_0/dml21/src/dml2_core/dml2_core_factory.o := $(dml2_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml2_0/dml21/src/dml2_dpmm/dml2_dpmm_dcn4.o := $(dml2_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml2_0/dml21/src/dml2_dpmm/dml2_dpmm_factory.o := $(dml2_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml2_0/dml21/src/dml2_mcg/dml2_mcg_dcn4.o := $(dml2_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml2_0/dml21/src/dml2_mcg/dml2_mcg_factory.o := $(dml2_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml2_0/dml21/src/dml2_pmo/dml2_pmo_dcn3.o := $(dml2_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml2_0/dml21/src/dml2_pmo/dml2_pmo_dcn4_fams2.o := $(dml2_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml2_0/dml21/src/dml2_pmo/dml2_pmo_factory.o := $(dml2_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml2_0/dml21/src/dml2_standalone_libraries/lib_float_math.o := $(dml2_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml2_0/dml21/src/dml21_wrapper.o := $(dml2_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml2_0/dml21/dml21_translation_helper.o := $(dml2_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml2_0/dml21/dml21_utils.o := $(dml2_ccflags)
+
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2_0/dml21/src/dml2_core/dml2_core_dcn4.o := $(dml2_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2_0/dml21/src/dml2_core/dml2_core_dcn4_calcs.o := $(dml2_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2_0/dml21/src/dml2_core/dml2_core_factory.o := $(dml2_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2_0/dml21/src/dml2_core/dml2_core_utils.o := $(dml2_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2_0/dml21/src/dml2_top/dml2_top_interfaces.o := $(dml2_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2_0/dml21/src/dml2_top/dml2_top_soc15.o := $(dml2_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2_0/dml21/src/dml2_dpmm/dml2_dpmm_dcn4.o := $(dml2_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2_0/dml21/src/dml2_dpmm/dml2_dpmm_factory.o := $(dml2_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2_0/dml21/src/dml2_mcg/dml2_mcg_dcn4.o := $(dml2_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2_0/dml21/src/dml2_mcg/dml2_mcg_factory.o := $(dml2_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2_0/dml21/src/dml2_pmo/dml2_pmo_dcn3.o := $(dml2_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2_0/dml21/src/dml2_pmo/dml2_pmo_dcn4_fams2.o := $(dml2_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2_0/dml21/src/dml2_pmo/dml2_pmo_factory.o := $(dml2_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2_0/dml21/src/dml2_standalone_libraries/lib_float_math.o := $(dml2_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2_0/dml21/src/dml21_wrapper.o := $(dml2_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2_0/dml21/dml21_translation_helper.o := $(dml2_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2_0/dml21/dml21_utils.o := $(dml2_rcflags)
+
+DML21 := src/dml2_top/dml2_top_interfaces.o
+DML21 += src/dml2_top/dml2_top_soc15.o
+DML21 += src/dml2_core/dml2_core_dcn4.o
+DML21 += src/dml2_core/dml2_core_utils.o
+DML21 += src/dml2_core/dml2_core_factory.o
+DML21 += src/dml2_core/dml2_core_dcn4_calcs.o
+DML21 += src/dml2_dpmm/dml2_dpmm_dcn4.o
+DML21 += src/dml2_dpmm/dml2_dpmm_factory.o
+DML21 += src/dml2_mcg/dml2_mcg_dcn4.o
+DML21 += src/dml2_mcg/dml2_mcg_factory.o
+DML21 += src/dml2_pmo/dml2_pmo_dcn3.o
+DML21 += src/dml2_pmo/dml2_pmo_factory.o
+DML21 += src/dml2_pmo/dml2_pmo_dcn4_fams2.o
+DML21 += src/dml2_standalone_libraries/lib_float_math.o
+DML21 += dml21_translation_helper.o
+DML21 += dml21_wrapper.o
+DML21 += dml21_utils.o
+
+AMD_DAL_DML21 = $(addprefix $(AMDDALPATH)/dc/dml2_0/dml21/,$(DML21))
+
+AMD_DISPLAY_FILES += $(AMD_DAL_DML21)
diff --git a/drivers/gpu/drm/amd/display/dc/dml2_0/cmntypes.h b/drivers/gpu/drm/amd/display/dc/dml2_0/cmntypes.h
new file mode 100644
index 000000000000..b954c9648fbe
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml2_0/cmntypes.h
@@ -0,0 +1,94 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __CMNTYPES_H__
+#define __CMNTYPES_H__
+
+#ifdef __GNUC__
+#if __GNUC__ == 4 && __GNUC_MINOR__ > 7
+typedef unsigned int uint;
+#endif
+#endif
+
+typedef signed char int8, *pint8;
+typedef signed short int16, *pint16;
+typedef signed int int32, *pint32;
+typedef signed int64, *pint64;
+
+typedef unsigned char uint8, *puint8;
+typedef unsigned short uint16, *puint16;
+typedef unsigned int uint32, *puint32;
+typedef unsigned uint64, *puint64;
+
+typedef unsigned long int ulong;
+typedef unsigned char uchar;
+typedef unsigned int uint;
+
+typedef void *pvoid;
+typedef char *pchar;
+typedef const void *const_pvoid;
+typedef const char *const_pchar;
+
+typedef struct rgba_struct {
+	uint8 a;
+	uint8 r;
+	uint8 g;
+	uint8 b;
+} rgba_t;
+
+typedef struct {
+	uint8 blue;
+	uint8 green;
+	uint8 red;
+	uint8 alpha;
+} gen_color_t;
+
+typedef union {
+	uint32		val;
+	gen_color_t f;
+} gen_color_u;
+
+//
+// Types to make it easy to get or set the bits of a float/double.
+// Avoids automatic casting from int to float and back.
+//
+#if 0
+typedef union {
+	uint32 i;
+	float f;
+} uintfloat32;
+
+typedef union {
+	uint64 i;
+	double f;
+} uintfloat64;
+
+#ifndef UNREFERENCED_PARAMETER
+#define UNREFERENCED_PARAMETER(x) (x = x)
+#endif
+#endif
+
+#endif  //__CMNTYPES_H__
diff --git a/drivers/gpu/drm/amd/display/dc/dml2_0/display_mode_core.c b/drivers/gpu/drm/amd/display/dc/dml2_0/display_mode_core.c
new file mode 100644
index 000000000000..c468f492b876
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml2_0/display_mode_core.c
@@ -0,0 +1,10337 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#include "display_mode_core.h"
+#include "display_mode_util.h"
+#include "display_mode_lib_defines.h"
+
+#include "dml_assert.h"
+
+#define DML2_MAX_FMT_420_BUFFER_WIDTH 4096
+#define TB_BORROWED_MAX 400
+#define DML_MAX_VSTARTUP_START 1023
+
+// ---------------------------
+//  Declaration Begins
+// ---------------------------
+static void CalculateBytePerPixelAndBlockSizes(
+	enum dml_source_format_class SourcePixelFormat,
+	enum dml_swizzle_mode SurfaceTiling,
+	// Output
+	dml_uint_t *BytePerPixelY,
+	dml_uint_t *BytePerPixelC,
+	dml_float_t *BytePerPixelDETY,
+	dml_float_t *BytePerPixelDETC,
+	dml_uint_t *BlockHeight256BytesY,
+	dml_uint_t *BlockHeight256BytesC,
+	dml_uint_t *BlockWidth256BytesY,
+	dml_uint_t *BlockWidth256BytesC,
+	dml_uint_t *MacroTileHeightY,
+	dml_uint_t *MacroTileHeightC,
+	dml_uint_t *MacroTileWidthY,
+	dml_uint_t *MacroTileWidthC);
+
+static dml_float_t CalculateWriteBackDISPCLK(
+	enum dml_source_format_class WritebackPixelFormat,
+	dml_float_t PixelClock,
+	dml_float_t WritebackHRatio,
+	dml_float_t WritebackVRatio,
+	dml_uint_t WritebackHTaps,
+	dml_uint_t WritebackVTaps,
+	dml_uint_t WritebackSourceWidth,
+	dml_uint_t WritebackDestinationWidth,
+	dml_uint_t HTotal,
+	dml_uint_t WritebackLineBufferSize,
+	dml_float_t DISPCLKDPPCLKVCOSpeed);
+
+static void CalculateVMRowAndSwath(
+	struct display_mode_lib_scratch_st *s,
+	struct CalculateVMRowAndSwath_params_st *p);
+
+static void CalculateOutputLink(
+	dml_float_t PHYCLKPerState,
+	dml_float_t PHYCLKD18PerState,
+	dml_float_t PHYCLKD32PerState,
+	dml_float_t Downspreading,
+	dml_bool_t IsMainSurfaceUsingTheIndicatedTiming,
+	enum dml_output_encoder_class Output,
+	enum dml_output_format_class OutputFormat,
+	dml_uint_t HTotal,
+	dml_uint_t HActive,
+	dml_float_t PixelClockBackEnd,
+	dml_float_t ForcedOutputLinkBPP,
+	dml_uint_t DSCInputBitPerComponent,
+	dml_uint_t NumberOfDSCSlices,
+	dml_float_t AudioSampleRate,
+	dml_uint_t AudioSampleLayout,
+	enum dml_odm_mode ODMModeNoDSC,
+	enum dml_odm_mode ODMModeDSC,
+	enum dml_dsc_enable DSCEnable,
+	dml_uint_t OutputLinkDPLanes,
+	enum dml_output_link_dp_rate OutputLinkDPRate,
+
+	// Output
+	dml_bool_t *RequiresDSC,
+	dml_bool_t *RequiresFEC,
+	dml_float_t *OutBpp,
+	enum dml_output_type_and_rate__type *OutputType,
+	enum dml_output_type_and_rate__rate *OutputRate,
+	dml_uint_t *RequiredSlots);
+
+static void CalculateODMMode(
+	dml_uint_t MaximumPixelsPerLinePerDSCUnit,
+	dml_uint_t HActive,
+	enum dml_output_encoder_class Output,
+	enum dml_output_format_class OutputFormat,
+	enum dml_odm_use_policy ODMUse,
+	dml_float_t StateDispclk,
+	dml_float_t MaxDispclk,
+	dml_bool_t DSCEnable,
+	dml_uint_t TotalNumberOfActiveDPP,
+	dml_uint_t MaxNumDPP,
+	dml_float_t PixelClock,
+	dml_float_t DISPCLKDPPCLKDSCCLKDownSpreading,
+	dml_float_t DISPCLKRampingMargin,
+	dml_float_t DISPCLKDPPCLKVCOSpeed,
+	dml_uint_t NumberOfDSCSlices,
+
+	// Output
+	dml_bool_t *TotalAvailablePipesSupport,
+	dml_uint_t *NumberOfDPP,
+	enum dml_odm_mode *ODMMode,
+	dml_float_t *RequiredDISPCLKPerSurface);
+
+static dml_float_t CalculateRequiredDispclk(
+	enum dml_odm_mode ODMMode,
+	dml_float_t PixelClock,
+	dml_float_t DISPCLKDPPCLKDSCCLKDownSpreading,
+	dml_float_t DISPCLKRampingMargin,
+	dml_float_t DISPCLKDPPCLKVCOSpeed,
+	dml_float_t MaxDispclkSingle);
+
+static void CalculateSinglePipeDPPCLKAndSCLThroughput(
+	dml_float_t HRatio,
+	dml_float_t HRatioChroma,
+	dml_float_t VRatio,
+	dml_float_t VRatioChroma,
+	dml_float_t MaxDCHUBToPSCLThroughput,
+	dml_float_t MaxPSCLToLBThroughput,
+	dml_float_t PixelClock,
+	enum dml_source_format_class SourcePixelFormat,
+	dml_uint_t HTaps,
+	dml_uint_t HTapsChroma,
+	dml_uint_t VTaps,
+	dml_uint_t VTapsChroma,
+
+	// Output
+	dml_float_t *PSCL_THROUGHPUT,
+	dml_float_t *PSCL_THROUGHPUT_CHROMA,
+	dml_float_t *DPPCLKUsingSingleDPP);
+
+static void CalculateDPPCLK(
+	dml_uint_t NumberOfActiveSurfaces,
+	dml_float_t DISPCLKDPPCLKDSCCLKDownSpreading,
+	dml_float_t DISPCLKDPPCLKVCOSpeed,
+	dml_float_t DPPCLKUsingSingleDPP[],
+	dml_uint_t DPPPerSurface[],
+
+	// Output
+	dml_float_t *GlobalDPPCLK,
+	dml_float_t Dppclk[]);
+
+static void CalculateMALLUseForStaticScreen(
+	dml_uint_t NumberOfActiveSurfaces,
+	dml_uint_t MALLAllocatedForDCNFinal,
+	enum dml_use_mall_for_static_screen_mode *UseMALLForStaticScreen,
+	dml_uint_t SurfaceSizeInMALL[],
+	dml_bool_t one_row_per_frame_fits_in_buffer[],
+
+	// Output
+	dml_bool_t UsesMALLForStaticScreen[]);
+
+static dml_uint_t dscceComputeDelay(
+	dml_uint_t bpc,
+	dml_float_t BPP,
+	dml_uint_t sliceWidth,
+	dml_uint_t numSlices,
+	enum dml_output_format_class pixelFormat,
+	enum dml_output_encoder_class Output);
+
+static dml_uint_t dscComputeDelay(enum dml_output_format_class pixelFormat,
+	enum dml_output_encoder_class Output);
+
+static dml_bool_t CalculatePrefetchSchedule(struct display_mode_lib_scratch_st *scratch,
+	struct CalculatePrefetchSchedule_params_st *p);
+
+static dml_float_t RoundToDFSGranularity(dml_float_t Clock, dml_bool_t round_up, dml_float_t VCOSpeed);
+
+static void CalculateDCCConfiguration(
+	dml_bool_t DCCEnabled,
+	dml_bool_t DCCProgrammingAssumesScanDirectionUnknown,
+	enum dml_source_format_class SourcePixelFormat,
+	dml_uint_t SurfaceWidthLuma,
+	dml_uint_t SurfaceWidthChroma,
+	dml_uint_t SurfaceHeightLuma,
+	dml_uint_t SurfaceHeightChroma,
+	dml_uint_t nomDETInKByte,
+	dml_uint_t RequestHeight256ByteLuma,
+	dml_uint_t RequestHeight256ByteChroma,
+	enum dml_swizzle_mode TilingFormat,
+	dml_uint_t BytePerPixelY,
+	dml_uint_t BytePerPixelC,
+	dml_float_t BytePerPixelDETY,
+	dml_float_t BytePerPixelDETC,
+	enum dml_rotation_angle SourceScan,
+	// Output
+	dml_uint_t *MaxUncompressedBlockLuma,
+	dml_uint_t *MaxUncompressedBlockChroma,
+	dml_uint_t *MaxCompressedBlockLuma,
+	dml_uint_t *MaxCompressedBlockChroma,
+	dml_uint_t *IndependentBlockLuma,
+	dml_uint_t *IndependentBlockChroma);
+
+static dml_uint_t CalculatePrefetchSourceLines(
+	dml_float_t VRatio,
+	dml_uint_t VTaps,
+	dml_bool_t Interlace,
+	dml_bool_t ProgressiveToInterlaceUnitInOPP,
+	dml_uint_t SwathHeight,
+	enum dml_rotation_angle SourceScan,
+	dml_bool_t ViewportStationary,
+	dml_uint_t SwathWidth,
+	dml_uint_t ViewportHeight,
+	dml_uint_t ViewportXStart,
+	dml_uint_t ViewportYStart,
+
+	// Output
+	dml_uint_t *VInitPreFill,
+	dml_uint_t *MaxNumSwath);
+
+static dml_uint_t CalculateVMAndRowBytes(
+	dml_bool_t ViewportStationary,
+	dml_bool_t DCCEnable,
+	dml_uint_t NumberOfDPPs,
+	dml_uint_t BlockHeight256Bytes,
+	dml_uint_t BlockWidth256Bytes,
+	enum dml_source_format_class SourcePixelFormat,
+	dml_uint_t SurfaceTiling,
+	dml_uint_t BytePerPixel,
+	enum dml_rotation_angle SourceScan,
+	dml_uint_t SwathWidth,
+	dml_uint_t ViewportHeight,
+	dml_uint_t ViewportXStart,
+	dml_uint_t ViewportYStart,
+	dml_bool_t GPUVMEnable,
+	dml_uint_t GPUVMMaxPageTableLevels,
+	dml_uint_t GPUVMMinPageSizeKBytes,
+	dml_uint_t PTEBufferSizeInRequests,
+	dml_uint_t Pitch,
+	dml_uint_t DCCMetaPitch,
+	dml_uint_t MacroTileWidth,
+	dml_uint_t MacroTileHeight,
+
+	// Output
+	dml_uint_t *MetaRowByte,
+	dml_uint_t *PixelPTEBytesPerRow,
+	dml_uint_t *PixelPTEBytesPerRowStorage, // for PTE buffer size check
+	dml_uint_t *dpte_row_width_ub,
+	dml_uint_t *dpte_row_height,
+	dml_uint_t *dpte_row_height_linear,
+	dml_uint_t *PixelPTEBytesPerRow_one_row_per_frame,
+	dml_uint_t *dpte_row_width_ub_one_row_per_frame,
+	dml_uint_t *dpte_row_height_one_row_per_frame,
+	dml_uint_t *MetaRequestWidth,
+	dml_uint_t *MetaRequestHeight,
+	dml_uint_t *meta_row_width,
+	dml_uint_t *meta_row_height,
+	dml_uint_t *PixelPTEReqWidth,
+	dml_uint_t *PixelPTEReqHeight,
+	dml_uint_t *PTERequestSize,
+	dml_uint_t *DPDE0BytesFrame,
+	dml_uint_t *MetaPTEBytesFrame);
+
+static dml_float_t CalculateTWait(
+	dml_uint_t PrefetchMode,
+	enum dml_use_mall_for_pstate_change_mode UseMALLForPStateChange,
+	dml_bool_t SynchronizeDRRDisplaysForUCLKPStateChangeFinal,
+	dml_bool_t DRRDisplay,
+	dml_float_t DRAMClockChangeLatency,
+	dml_float_t FCLKChangeLatency,
+	dml_float_t UrgentLatency,
+	dml_float_t SREnterPlusExitTime);
+
+static void CalculatePrefetchMode(
+	enum dml_prefetch_modes AllowForPStateChangeOrStutterInVBlank,
+	dml_uint_t *MinPrefetchMode,
+	dml_uint_t *MaxPrefetchMode);
+
+static void CalculateRowBandwidth(
+	dml_bool_t GPUVMEnable,
+	enum dml_source_format_class SourcePixelFormat,
+	dml_float_t VRatio,
+	dml_float_t VRatioChroma,
+	dml_bool_t DCCEnable,
+	dml_float_t LineTime,
+	dml_uint_t MetaRowByteLuma,
+	dml_uint_t MetaRowByteChroma,
+	dml_uint_t meta_row_height_luma,
+	dml_uint_t meta_row_height_chroma,
+	dml_uint_t PixelPTEBytesPerRowLuma,
+	dml_uint_t PixelPTEBytesPerRowChroma,
+	dml_uint_t dpte_row_height_luma,
+	dml_uint_t dpte_row_height_chroma,
+	// Output
+	dml_float_t *meta_row_bw,
+	dml_float_t *dpte_row_bw);
+
+static void CalculateFlipSchedule(
+	dml_float_t HostVMInefficiencyFactor,
+	dml_float_t UrgentExtraLatency,
+	dml_float_t UrgentLatency,
+	dml_uint_t GPUVMMaxPageTableLevels,
+	dml_bool_t HostVMEnable,
+	dml_uint_t HostVMMaxNonCachedPageTableLevels,
+	dml_bool_t GPUVMEnable,
+	dml_uint_t HostVMMinPageSize,
+	dml_float_t PDEAndMetaPTEBytesPerFrame,
+	dml_float_t MetaRowBytes,
+	dml_float_t DPTEBytesPerRow,
+	dml_float_t BandwidthAvailableForImmediateFlip,
+	dml_uint_t TotImmediateFlipBytes,
+	enum dml_source_format_class SourcePixelFormat,
+	dml_float_t LineTime,
+	dml_float_t VRatio,
+	dml_float_t VRatioChroma,
+	dml_float_t Tno_bw,
+	dml_bool_t DCCEnable,
+	dml_uint_t dpte_row_height,
+	dml_uint_t meta_row_height,
+	dml_uint_t dpte_row_height_chroma,
+	dml_uint_t meta_row_height_chroma,
+	dml_bool_t use_one_row_for_frame_flip,
+
+	// Output
+	dml_float_t *DestinationLinesToRequestVMInImmediateFlip,
+	dml_float_t *DestinationLinesToRequestRowInImmediateFlip,
+	dml_float_t *final_flip_bw,
+	dml_bool_t *ImmediateFlipSupportedForPipe);
+
+static dml_float_t CalculateWriteBackDelay(
+	enum dml_source_format_class WritebackPixelFormat,
+	dml_float_t WritebackHRatio,
+	dml_float_t WritebackVRatio,
+	dml_uint_t WritebackVTaps,
+	dml_uint_t WritebackDestinationWidth,
+	dml_uint_t WritebackDestinationHeight,
+	dml_uint_t WritebackSourceHeight,
+	dml_uint_t HTotal);
+
+static void CalculateVUpdateAndDynamicMetadataParameters(
+	dml_uint_t MaxInterDCNTileRepeaters,
+	dml_float_t Dppclk,
+	dml_float_t DISPCLK,
+	dml_float_t DCFClkDeepSleep,
+	dml_float_t PixelClock,
+	dml_uint_t HTotal,
+	dml_uint_t VBlank,
+	dml_uint_t DynamicMetadataTransmittedBytes,
+	dml_uint_t DynamicMetadataLinesBeforeActiveRequired,
+	dml_uint_t InterlaceEnable,
+	dml_bool_t ProgressiveToInterlaceUnitInOPP,
+	dml_float_t *TSetup,
+	dml_float_t *Tdmbf,
+	dml_float_t *Tdmec,
+	dml_float_t *Tdmsks,
+	dml_uint_t *VUpdateOffsetPix,
+	dml_uint_t *VUpdateWidthPix,
+	dml_uint_t *VReadyOffsetPix);
+
+static void PixelClockAdjustmentForProgressiveToInterlaceUnit(struct dml_display_cfg_st *display_cfg, dml_bool_t ptoi_supported);
+
+static dml_float_t TruncToValidBPP(
+	dml_float_t LinkBitRate,
+	dml_uint_t Lanes,
+	dml_uint_t HTotal,
+	dml_uint_t HActive,
+	dml_float_t PixelClock,
+	dml_float_t DesiredBPP,
+	dml_bool_t DSCEnable,
+	enum dml_output_encoder_class Output,
+	enum dml_output_format_class Format,
+	dml_uint_t DSCInputBitPerComponent,
+	dml_uint_t DSCSlices,
+	dml_uint_t AudioRate,
+	dml_uint_t AudioLayout,
+	enum dml_odm_mode ODMModeNoDSC,
+	enum dml_odm_mode ODMModeDSC,
+	// Output
+	dml_uint_t *RequiredSlotsSingle);
+
+static void CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport(
+	struct display_mode_lib_scratch_st *s,
+	struct CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_params_st *p);
+
+static void CalculateDCFCLKDeepSleep(
+	dml_uint_t NumberOfActiveSurfaces,
+	dml_uint_t BytePerPixelY[],
+	dml_uint_t BytePerPixelC[],
+	dml_float_t VRatio[],
+	dml_float_t VRatioChroma[],
+	dml_uint_t SwathWidthY[],
+	dml_uint_t SwathWidthC[],
+	dml_uint_t DPPPerSurface[],
+	dml_float_t HRatio[],
+	dml_float_t HRatioChroma[],
+	dml_float_t PixelClock[],
+	dml_float_t PSCL_THROUGHPUT[],
+	dml_float_t PSCL_THROUGHPUT_CHROMA[],
+	dml_float_t Dppclk[],
+	dml_float_t ReadBandwidthLuma[],
+	dml_float_t ReadBandwidthChroma[],
+	dml_uint_t ReturnBusWidth,
+
+	// Output
+	dml_float_t *DCFCLKDeepSleep);
+
+static void CalculateUrgentBurstFactor(
+	enum dml_use_mall_for_pstate_change_mode UseMALLForPStateChange,
+	dml_uint_t swath_width_luma_ub,
+	dml_uint_t swath_width_chroma_ub,
+	dml_uint_t SwathHeightY,
+	dml_uint_t SwathHeightC,
+	dml_float_t LineTime,
+	dml_float_t UrgentLatency,
+	dml_float_t CursorBufferSize,
+	dml_uint_t CursorWidth,
+	dml_uint_t CursorBPP,
+	dml_float_t VRatio,
+	dml_float_t VRatioC,
+	dml_float_t BytePerPixelInDETY,
+	dml_float_t BytePerPixelInDETC,
+	dml_uint_t DETBufferSizeY,
+	dml_uint_t DETBufferSizeC,
+	// Output
+	dml_float_t *UrgentBurstFactorCursor,
+	dml_float_t *UrgentBurstFactorLuma,
+	dml_float_t *UrgentBurstFactorChroma,
+	dml_bool_t *NotEnoughUrgentLatencyHiding);
+
+static dml_float_t RequiredDTBCLK(
+	dml_bool_t DSCEnable,
+	dml_float_t PixelClock,
+	enum dml_output_format_class OutputFormat,
+	dml_float_t OutputBpp,
+	dml_uint_t DSCSlices,
+	dml_uint_t HTotal,
+	dml_uint_t HActive,
+	dml_uint_t AudioRate,
+	dml_uint_t AudioLayoutSingle);
+
+static void UseMinimumDCFCLK(
+	struct display_mode_lib_scratch_st *scratch,
+	struct UseMinimumDCFCLK_params_st *p);
+
+static void CalculatePixelDeliveryTimes(
+	dml_uint_t NumberOfActiveSurfaces,
+	dml_float_t VRatio[],
+	dml_float_t VRatioChroma[],
+	dml_float_t VRatioPrefetchY[],
+	dml_float_t VRatioPrefetchC[],
+	dml_uint_t swath_width_luma_ub[],
+	dml_uint_t swath_width_chroma_ub[],
+	dml_uint_t DPPPerSurface[],
+	dml_float_t HRatio[],
+	dml_float_t HRatioChroma[],
+	dml_float_t PixelClock[],
+	dml_float_t PSCL_THROUGHPUT[],
+	dml_float_t PSCL_THROUGHPUT_CHROMA[],
+	dml_float_t Dppclk[],
+	dml_uint_t BytePerPixelC[],
+	enum dml_rotation_angle SourceScan[],
+	dml_uint_t NumberOfCursors[],
+	dml_uint_t CursorWidth[],
+	dml_uint_t CursorBPP[],
+	dml_uint_t BlockWidth256BytesY[],
+	dml_uint_t BlockHeight256BytesY[],
+	dml_uint_t BlockWidth256BytesC[],
+	dml_uint_t BlockHeight256BytesC[],
+
+	// Output
+	dml_float_t DisplayPipeLineDeliveryTimeLuma[],
+	dml_float_t DisplayPipeLineDeliveryTimeChroma[],
+	dml_float_t DisplayPipeLineDeliveryTimeLumaPrefetch[],
+	dml_float_t DisplayPipeLineDeliveryTimeChromaPrefetch[],
+	dml_float_t DisplayPipeRequestDeliveryTimeLuma[],
+	dml_float_t DisplayPipeRequestDeliveryTimeChroma[],
+	dml_float_t DisplayPipeRequestDeliveryTimeLumaPrefetch[],
+	dml_float_t DisplayPipeRequestDeliveryTimeChromaPrefetch[],
+	dml_float_t CursorRequestDeliveryTime[],
+	dml_float_t CursorRequestDeliveryTimePrefetch[]);
+
+static void CalculateMetaAndPTETimes(
+	dml_bool_t use_one_row_for_frame[],
+	dml_uint_t NumberOfActiveSurfaces,
+	dml_bool_t GPUVMEnable,
+	dml_uint_t MetaChunkSize,
+	dml_uint_t MinMetaChunkSizeBytes,
+	dml_uint_t HTotal[],
+	dml_float_t VRatio[],
+	dml_float_t VRatioChroma[],
+	dml_float_t DestinationLinesToRequestRowInVBlank[],
+	dml_float_t DestinationLinesToRequestRowInImmediateFlip[],
+	dml_bool_t DCCEnable[],
+	dml_float_t PixelClock[],
+	dml_uint_t BytePerPixelY[],
+	dml_uint_t BytePerPixelC[],
+	enum dml_rotation_angle SourceScan[],
+	dml_uint_t dpte_row_height[],
+	dml_uint_t dpte_row_height_chroma[],
+	dml_uint_t meta_row_width[],
+	dml_uint_t meta_row_width_chroma[],
+	dml_uint_t meta_row_height[],
+	dml_uint_t meta_row_height_chroma[],
+	dml_uint_t meta_req_width[],
+	dml_uint_t meta_req_width_chroma[],
+	dml_uint_t meta_req_height[],
+	dml_uint_t meta_req_height_chroma[],
+	dml_uint_t dpte_group_bytes[],
+	dml_uint_t PTERequestSizeY[],
+	dml_uint_t PTERequestSizeC[],
+	dml_uint_t PixelPTEReqWidthY[],
+	dml_uint_t PixelPTEReqHeightY[],
+	dml_uint_t PixelPTEReqWidthC[],
+	dml_uint_t PixelPTEReqHeightC[],
+	dml_uint_t dpte_row_width_luma_ub[],
+	dml_uint_t dpte_row_width_chroma_ub[],
+
+	// Output
+	dml_float_t DST_Y_PER_PTE_ROW_NOM_L[],
+	dml_float_t DST_Y_PER_PTE_ROW_NOM_C[],
+	dml_float_t DST_Y_PER_META_ROW_NOM_L[],
+	dml_float_t DST_Y_PER_META_ROW_NOM_C[],
+	dml_float_t TimePerMetaChunkNominal[],
+	dml_float_t TimePerChromaMetaChunkNominal[],
+	dml_float_t TimePerMetaChunkVBlank[],
+	dml_float_t TimePerChromaMetaChunkVBlank[],
+	dml_float_t TimePerMetaChunkFlip[],
+	dml_float_t TimePerChromaMetaChunkFlip[],
+	dml_float_t time_per_pte_group_nom_luma[],
+	dml_float_t time_per_pte_group_vblank_luma[],
+	dml_float_t time_per_pte_group_flip_luma[],
+	dml_float_t time_per_pte_group_nom_chroma[],
+	dml_float_t time_per_pte_group_vblank_chroma[],
+	dml_float_t time_per_pte_group_flip_chroma[]);
+
+static void CalculateVMGroupAndRequestTimes(
+	dml_uint_t NumberOfActiveSurfaces,
+	dml_bool_t GPUVMEnable,
+	dml_uint_t GPUVMMaxPageTableLevels,
+	dml_uint_t HTotal[],
+	dml_uint_t BytePerPixelC[],
+	dml_float_t DestinationLinesToRequestVMInVBlank[],
+	dml_float_t DestinationLinesToRequestVMInImmediateFlip[],
+	dml_bool_t DCCEnable[],
+	dml_float_t PixelClock[],
+	dml_uint_t dpte_row_width_luma_ub[],
+	dml_uint_t dpte_row_width_chroma_ub[],
+	dml_uint_t vm_group_bytes[],
+	dml_uint_t dpde0_bytes_per_frame_ub_l[],
+	dml_uint_t dpde0_bytes_per_frame_ub_c[],
+	dml_uint_t meta_pte_bytes_per_frame_ub_l[],
+	dml_uint_t meta_pte_bytes_per_frame_ub_c[],
+
+	// Output
+	dml_float_t TimePerVMGroupVBlank[],
+	dml_float_t TimePerVMGroupFlip[],
+	dml_float_t TimePerVMRequestVBlank[],
+	dml_float_t TimePerVMRequestFlip[]);
+
+static void CalculateStutterEfficiency(
+	struct display_mode_lib_scratch_st *scratch,
+	struct CalculateStutterEfficiency_params_st *p);
+
+static void CalculateSwathAndDETConfiguration(
+	struct display_mode_lib_scratch_st *scratch,
+	struct CalculateSwathAndDETConfiguration_params_st *p);
+
+static void CalculateSwathWidth(
+	dml_bool_t ForceSingleDPP,
+	dml_uint_t NumberOfActiveSurfaces,
+	enum dml_source_format_class SourcePixelFormat[],
+	enum dml_rotation_angle SourceScan[],
+	dml_bool_t ViewportStationary[],
+	dml_uint_t ViewportWidth[],
+	dml_uint_t ViewportHeight[],
+	dml_uint_t ViewportXStart[],
+	dml_uint_t ViewportYStart[],
+	dml_uint_t ViewportXStartC[],
+	dml_uint_t ViewportYStartC[],
+	dml_uint_t SurfaceWidthY[],
+	dml_uint_t SurfaceWidthC[],
+	dml_uint_t SurfaceHeightY[],
+	dml_uint_t SurfaceHeightC[],
+	enum dml_odm_mode ODMMode[],
+	dml_uint_t BytePerPixY[],
+	dml_uint_t BytePerPixC[],
+	dml_uint_t Read256BytesBlockHeightY[],
+	dml_uint_t Read256BytesBlockHeightC[],
+	dml_uint_t Read256BytesBlockWidthY[],
+	dml_uint_t Read256BytesBlockWidthC[],
+	dml_uint_t BlendingAndTiming[],
+	dml_uint_t HActive[],
+	dml_float_t HRatio[],
+	dml_uint_t DPPPerSurface[],
+
+	// Output
+	dml_uint_t SwathWidthSingleDPPY[],
+	dml_uint_t SwathWidthSingleDPPC[],
+	dml_uint_t SwathWidthY[],
+	dml_uint_t SwathWidthC[],
+	dml_uint_t MaximumSwathHeightY[],
+	dml_uint_t MaximumSwathHeightC[],
+	dml_uint_t swath_width_luma_ub[],
+	dml_uint_t swath_width_chroma_ub[]);
+
+static dml_float_t CalculateExtraLatency(
+	dml_uint_t RoundTripPingLatencyCycles,
+	dml_uint_t ReorderingBytes,
+	dml_float_t DCFCLK,
+	dml_uint_t TotalNumberOfActiveDPP,
+	dml_uint_t PixelChunkSizeInKByte,
+	dml_uint_t TotalNumberOfDCCActiveDPP,
+	dml_uint_t MetaChunkSize,
+	dml_float_t ReturnBW,
+	dml_bool_t GPUVMEnable,
+	dml_bool_t HostVMEnable,
+	dml_uint_t NumberOfActiveSurfaces,
+	dml_uint_t NumberOfDPP[],
+	dml_uint_t dpte_group_bytes[],
+	dml_float_t HostVMInefficiencyFactor,
+	dml_uint_t HostVMMinPageSize,
+	dml_uint_t HostVMMaxNonCachedPageTableLevels);
+
+static dml_uint_t CalculateExtraLatencyBytes(
+	dml_uint_t ReorderingBytes,
+	dml_uint_t TotalNumberOfActiveDPP,
+	dml_uint_t PixelChunkSizeInKByte,
+	dml_uint_t TotalNumberOfDCCActiveDPP,
+	dml_uint_t MetaChunkSize,
+	dml_bool_t GPUVMEnable,
+	dml_bool_t HostVMEnable,
+	dml_uint_t NumberOfActiveSurfaces,
+	dml_uint_t NumberOfDPP[],
+	dml_uint_t dpte_group_bytes[],
+	dml_float_t HostVMInefficiencyFactor,
+	dml_uint_t HostVMMinPageSize,
+	dml_uint_t HostVMMaxNonCachedPageTableLevels);
+
+static dml_float_t CalculateUrgentLatency(
+	dml_float_t UrgentLatencyPixelDataOnly,
+	dml_float_t UrgentLatencyPixelMixedWithVMData,
+	dml_float_t UrgentLatencyVMDataOnly,
+	dml_bool_t DoUrgentLatencyAdjustment,
+	dml_float_t UrgentLatencyAdjustmentFabricClockComponent,
+	dml_float_t UrgentLatencyAdjustmentFabricClockReference,
+	dml_float_t FabricClockSingle);
+
+static dml_bool_t UnboundedRequest(
+	enum dml_unbounded_requesting_policy UseUnboundedRequestingFinal,
+	dml_uint_t TotalNumberOfActiveDPP,
+	dml_bool_t NoChromaOrLinear,
+	enum dml_output_encoder_class Output);
+
+static void CalculateSurfaceSizeInMall(
+	dml_uint_t NumberOfActiveSurfaces,
+	dml_uint_t MALLAllocatedForDCN,
+	enum dml_use_mall_for_static_screen_mode UseMALLForStaticScreen[],
+	dml_bool_t DCCEnable[],
+	dml_bool_t ViewportStationary[],
+	dml_uint_t ViewportXStartY[],
+	dml_uint_t ViewportYStartY[],
+	dml_uint_t ViewportXStartC[],
+	dml_uint_t ViewportYStartC[],
+	dml_uint_t ViewportWidthY[],
+	dml_uint_t ViewportHeightY[],
+	dml_uint_t BytesPerPixelY[],
+	dml_uint_t ViewportWidthC[],
+	dml_uint_t ViewportHeightC[],
+	dml_uint_t BytesPerPixelC[],
+	dml_uint_t SurfaceWidthY[],
+	dml_uint_t SurfaceWidthC[],
+	dml_uint_t SurfaceHeightY[],
+	dml_uint_t SurfaceHeightC[],
+	dml_uint_t Read256BytesBlockWidthY[],
+	dml_uint_t Read256BytesBlockWidthC[],
+	dml_uint_t Read256BytesBlockHeightY[],
+	dml_uint_t Read256BytesBlockHeightC[],
+	dml_uint_t ReadBlockWidthY[],
+	dml_uint_t ReadBlockWidthC[],
+	dml_uint_t ReadBlockHeightY[],
+	dml_uint_t ReadBlockHeightC[],
+
+	// Output
+	dml_uint_t SurfaceSizeInMALL[],
+	dml_bool_t *ExceededMALLSize);
+
+static void CalculateDETBufferSize(
+	dml_uint_t DETSizeOverride[],
+	enum dml_use_mall_for_pstate_change_mode UseMALLForPStateChange[],
+	dml_bool_t ForceSingleDPP,
+	dml_uint_t NumberOfActiveSurfaces,
+	dml_bool_t UnboundedRequestEnabled,
+	dml_uint_t nomDETInKByte,
+	dml_uint_t MaxTotalDETInKByte,
+	dml_uint_t ConfigReturnBufferSizeInKByte,
+	dml_uint_t MinCompressedBufferSizeInKByte,
+	dml_uint_t ConfigReturnBufferSegmentSizeInkByte,
+	dml_uint_t CompressedBufferSegmentSizeInkByteFinal,
+	enum dml_source_format_class SourcePixelFormat[],
+	dml_float_t ReadBandwidthLuma[],
+	dml_float_t ReadBandwidthChroma[],
+	dml_uint_t RotesY[],
+	dml_uint_t RoundedUpMaxSwathSizeBytesC[],
+	dml_uint_t DPPPerSurface[],
+	// Output
+	dml_uint_t DETBufferSizeInKByte[],
+	dml_uint_t *CompressedBufferSizeInkByte);
+
+static void CalculateMaxDETAndMinCompressedBufferSize(
+	dml_uint_t ConfigReturnBufferSizeInKByte,
+	dml_uint_t ConfigReturnBufferSegmentSizeInKByte,
+	dml_uint_t ROBBufferSizeInKByte,
+	dml_uint_t MaxNumDPP,
+	dml_bool_t nomDETInKByteOverrideEnable,
+	dml_uint_t nomDETInKByteOverrideValue,
+
+	// Output
+	dml_uint_t *MaxTotalDETInKByte,
+	dml_uint_t *nomDETInKByte,
+	dml_uint_t *MinCompressedBufferSizeInKByte);
+
+static dml_uint_t DSCDelayRequirement(
+	dml_bool_t DSCEnabled,
+	enum dml_odm_mode ODMMode,
+	dml_uint_t DSCInputBitPerComponent,
+	dml_float_t OutputBpp,
+	dml_uint_t HActive,
+	dml_uint_t HTotal,
+	dml_uint_t NumberOfDSCSlices,
+	enum dml_output_format_class OutputFormat,
+	enum dml_output_encoder_class Output,
+	dml_float_t PixelClock,
+	dml_float_t PixelClockBackEnd);
+
+static dml_bool_t CalculateVActiveBandwithSupport(
+	dml_uint_t NumberOfActiveSurfaces,
+	dml_float_t ReturnBW,
+	dml_bool_t NotUrgentLatencyHiding[],
+	dml_float_t ReadBandwidthLuma[],
+	dml_float_t ReadBandwidthChroma[],
+	dml_float_t cursor_bw[],
+	dml_float_t meta_row_bandwidth[],
+	dml_float_t dpte_row_bandwidth[],
+	dml_uint_t NumberOfDPP[],
+	dml_float_t UrgentBurstFactorLuma[],
+	dml_float_t UrgentBurstFactorChroma[],
+	dml_float_t UrgentBurstFactorCursor[]);
+
+static void CalculatePrefetchBandwithSupport(
+	dml_uint_t NumberOfActiveSurfaces,
+	dml_float_t ReturnBW,
+	enum dml_use_mall_for_pstate_change_mode UseMALLForPStateChange[],
+	dml_bool_t NotUrgentLatencyHiding[],
+	dml_float_t ReadBandwidthLuma[],
+	dml_float_t ReadBandwidthChroma[],
+	dml_float_t PrefetchBandwidthLuma[],
+	dml_float_t PrefetchBandwidthChroma[],
+	dml_float_t cursor_bw[],
+	dml_float_t meta_row_bandwidth[],
+	dml_float_t dpte_row_bandwidth[],
+	dml_float_t cursor_bw_pre[],
+	dml_float_t prefetch_vmrow_bw[],
+	dml_uint_t NumberOfDPP[],
+	dml_float_t UrgentBurstFactorLuma[],
+	dml_float_t UrgentBurstFactorChroma[],
+	dml_float_t UrgentBurstFactorCursor[],
+	dml_float_t UrgentBurstFactorLumaPre[],
+	dml_float_t UrgentBurstFactorChromaPre[],
+	dml_float_t UrgentBurstFactorCursorPre[],
+
+	// Output
+	dml_float_t *PrefetchBandwidth,
+	dml_float_t *PrefetchBandwidthNotIncludingMALLPrefetch,
+	dml_float_t *FractionOfUrgentBandwidth,
+	dml_bool_t *PrefetchBandwidthSupport);
+
+static dml_float_t CalculateBandwidthAvailableForImmediateFlip(
+	dml_uint_t NumberOfActiveSurfaces,
+	dml_float_t ReturnBW,
+	dml_float_t ReadBandwidthLuma[],
+	dml_float_t ReadBandwidthChroma[],
+	dml_float_t PrefetchBandwidthLuma[],
+	dml_float_t PrefetchBandwidthChroma[],
+	dml_float_t cursor_bw[],
+	dml_float_t cursor_bw_pre[],
+	dml_uint_t NumberOfDPP[],
+	dml_float_t UrgentBurstFactorLuma[],
+	dml_float_t UrgentBurstFactorChroma[],
+	dml_float_t UrgentBurstFactorCursor[],
+	dml_float_t UrgentBurstFactorLumaPre[],
+	dml_float_t UrgentBurstFactorChromaPre[],
+	dml_float_t UrgentBurstFactorCursorPre[]);
+
+static void CalculateImmediateFlipBandwithSupport(
+	dml_uint_t NumberOfActiveSurfaces,
+	dml_float_t ReturnBW,
+	enum dml_use_mall_for_pstate_change_mode UseMALLForPStateChange[],
+	enum dml_immediate_flip_requirement ImmediateFlipRequirement[],
+	dml_float_t final_flip_bw[],
+	dml_float_t ReadBandwidthLuma[],
+	dml_float_t ReadBandwidthChroma[],
+	dml_float_t PrefetchBandwidthLuma[],
+	dml_float_t PrefetchBandwidthChroma[],
+	dml_float_t cursor_bw[],
+	dml_float_t meta_row_bandwidth[],
+	dml_float_t dpte_row_bandwidth[],
+	dml_float_t cursor_bw_pre[],
+	dml_float_t prefetch_vmrow_bw[],
+	dml_uint_t NumberOfDPP[],
+	dml_float_t UrgentBurstFactorLuma[],
+	dml_float_t UrgentBurstFactorChroma[],
+	dml_float_t UrgentBurstFactorCursor[],
+	dml_float_t UrgentBurstFactorLumaPre[],
+	dml_float_t UrgentBurstFactorChromaPre[],
+	dml_float_t UrgentBurstFactorCursorPre[],
+
+	// Output
+	dml_float_t *TotalBandwidth,
+	dml_float_t *TotalBandwidthNotIncludingMALLPrefetch,
+	dml_float_t *FractionOfUrgentBandwidth,
+	dml_bool_t *ImmediateFlipBandwidthSupport);
+
+// ---------------------------
+//  Declaration Ends
+// ---------------------------
+
+static dml_uint_t dscceComputeDelay(
+	dml_uint_t bpc,
+	dml_float_t BPP,
+	dml_uint_t sliceWidth,
+	dml_uint_t numSlices,
+	enum dml_output_format_class pixelFormat,
+	enum dml_output_encoder_class Output)
+{
+	// valid bpc = source bits per component in the set of {8, 10, 12}
+	// valid bpp = increments of 1/16 of a bit
+	// min = 6/7/8 in N420/N422/444, respectively
+	// max = such that compression is 1:1
+	//valid sliceWidth = number of pixels per slice line, must be less than or equal to 5184/numSlices (or 4096/numSlices in 420 mode)
+	//valid numSlices = number of slices in the horiziontal direction per DSC engine in the set of {1, 2, 3, 4}
+	//valid pixelFormat = pixel/color format in the set of {:N444_RGB, :S422, :N422, :N420}
+
+	// fixed value
+	dml_uint_t rcModelSize = 8192;
+
+	// N422/N420 operate at 2 pixels per clock
+	dml_uint_t pixelsPerClock, lstall, D, initalXmitDelay, w, s, ix, wx, p, l0, a, ax, L,
+		Delay, pixels;
+
+	if (pixelFormat == dml_420)
+		pixelsPerClock = 2;
+	// #all other modes operate at 1 pixel per clock
+	else if (pixelFormat == dml_444)
+		pixelsPerClock = 1;
+	else if (pixelFormat == dml_n422)
+		pixelsPerClock = 2;
+	else
+		pixelsPerClock = 1;
+
+	//initial transmit delay as per PPS
+	initalXmitDelay = (dml_uint_t)(dml_round(rcModelSize / 2.0 / BPP / pixelsPerClock, 1));
+
+	//compute ssm delay
+	if (bpc == 8)
+		D = 81;
+	else if (bpc == 10)
+		D = 89;
+	else
+		D = 113;
+
+	//divide by pixel per cycle to compute slice width as seen by DSC
+	w = sliceWidth / pixelsPerClock;
+
+	//422 mode has an additional cycle of delay
+	if (pixelFormat == dml_420 || pixelFormat == dml_444 || pixelFormat == dml_n422)
+		s = 0;
+	else
+		s = 1;
+
+	//main calculation for the dscce
+	ix = initalXmitDelay + 45;
+	wx = (w + 2) / 3;
+	p = 3 * wx - w;
+	l0 = ix / w;
+	a = ix + p * l0;
+	ax = (a + 2) / 3 + D + 6 + 1;
+	L = (ax + wx - 1) / wx;
+	if ((ix % w) == 0 && p != 0)
+		lstall = 1;
+	else
+		lstall = 0;
+	Delay = L * wx * (numSlices - 1) + ax + s + lstall + 22;
+
+	//dsc processes 3 pixel containers per cycle and a container can contain 1 or 2 pixels
+	pixels = Delay * 3 * pixelsPerClock;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: bpc: %u\n", __func__, bpc);
+	dml_print("DML::%s: BPP: %f\n", __func__, BPP);
+	dml_print("DML::%s: sliceWidth: %u\n", __func__, sliceWidth);
+	dml_print("DML::%s: numSlices: %u\n", __func__, numSlices);
+	dml_print("DML::%s: pixelFormat: %u\n", __func__, pixelFormat);
+	dml_print("DML::%s: Output: %u\n", __func__, Output);
+	dml_print("DML::%s: pixels: %u\n", __func__, pixels);
+#endif
+	return pixels;
+}
+
+static dml_uint_t dscComputeDelay(enum dml_output_format_class pixelFormat, enum dml_output_encoder_class Output)
+{
+	dml_uint_t Delay = 0;
+
+	if (pixelFormat == dml_420) {
+		// sfr
+		Delay = Delay + 2;
+		// dsccif
+		Delay = Delay + 0;
+		// dscc - input deserializer
+		Delay = Delay + 3;
+		// dscc gets pixels every other cycle
+		Delay = Delay + 2;
+		// dscc - input cdc fifo
+		Delay = Delay + 12;
+		// dscc gets pixels every other cycle
+		Delay = Delay + 13;
+		// dscc - cdc uncertainty
+		Delay = Delay + 2;
+		// dscc - output cdc fifo
+		Delay = Delay + 7;
+		// dscc gets pixels every other cycle
+		Delay = Delay + 3;
+		// dscc - cdc uncertainty
+		Delay = Delay + 2;
+		// dscc - output serializer
+		Delay = Delay + 1;
+		// sft
+		Delay = Delay + 1;
+	} else if (pixelFormat == dml_n422) {
+	// sfr
+	Delay = Delay + 2;
+	// dsccif
+	Delay = Delay + 1;
+	// dscc - input deserializer
+	Delay = Delay + 5;
+	// dscc - input cdc fifo
+	Delay = Delay + 25;
+	// dscc - cdc uncertainty
+	Delay = Delay + 2;
+	// dscc - output cdc fifo
+	Delay = Delay + 10;
+	// dscc - cdc uncertainty
+	Delay = Delay + 2;
+	// dscc - output serializer
+	Delay = Delay + 1;
+	// sft
+	Delay = Delay + 1;
+	} else {
+	// sfr
+	Delay = Delay + 2;
+	// dsccif
+	Delay = Delay + 0;
+	// dscc - input deserializer
+	Delay = Delay + 3;
+	// dscc - input cdc fifo
+	Delay = Delay + 12;
+	// dscc - cdc uncertainty
+	Delay = Delay + 2;
+	// dscc - output cdc fifo
+	Delay = Delay + 7;
+	// dscc - output serializer
+	Delay = Delay + 1;
+	// dscc - cdc uncertainty
+	Delay = Delay + 2;
+	// sft
+	Delay = Delay + 1;
+	}
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: pixelFormat = %u\n", __func__, pixelFormat);
+	dml_print("DML::%s: Delay = %u\n", __func__, Delay);
+#endif
+
+	return Delay;
+}
+
+static dml_bool_t CalculatePrefetchSchedule(struct display_mode_lib_scratch_st *scratch,
+	struct CalculatePrefetchSchedule_params_st *p)
+{
+	struct CalculatePrefetchSchedule_locals_st *s = &scratch->CalculatePrefetchSchedule_locals;
+
+	s->MyError = false;
+	s->DPPCycles = 0;
+	s->DISPCLKCycles = 0;
+	s->DSTTotalPixelsAfterScaler = 0.0;
+	s->LineTime = 0.0;
+	s->dst_y_prefetch_equ = 0.0;
+	s->prefetch_bw_oto = 0.0;
+	s->Tvm_oto = 0.0;
+	s->Tr0_oto = 0.0;
+	s->Tvm_oto_lines = 0.0;
+	s->Tr0_oto_lines = 0.0;
+	s->dst_y_prefetch_oto = 0.0;
+	s->TimeForFetchingMetaPTE = 0.0;
+	s->TimeForFetchingRowInVBlank = 0.0;
+	s->LinesToRequestPrefetchPixelData = 0.0;
+	s->HostVMDynamicLevelsTrips = 0;
+	s->trip_to_mem = 0.0;
+	s->Tvm_trips = 0.0;
+	s->Tr0_trips = 0.0;
+	s->Tvm_trips_rounded = 0.0;
+	s->Tr0_trips_rounded = 0.0;
+	s->max_Tsw = 0.0;
+	s->Lsw_oto = 0.0;
+	s->Tpre_rounded = 0.0;
+	s->prefetch_bw_equ = 0.0;
+	s->Tvm_equ = 0.0;
+	s->Tr0_equ = 0.0;
+	s->Tdmbf = 0.0;
+	s->Tdmec = 0.0;
+	s->Tdmsks = 0.0;
+	s->prefetch_sw_bytes = 0.0;
+	s->prefetch_bw_pr = 0.0;
+	s->bytes_pp = 0.0;
+	s->dep_bytes = 0.0;
+	s->min_Lsw_oto = 0.0;
+	s->Tsw_est1 = 0.0;
+	s->Tsw_est3 = 0.0;
+
+	if (p->GPUVMEnable == true && p->HostVMEnable == true) {
+		s->HostVMDynamicLevelsTrips = p->HostVMMaxNonCachedPageTableLevels;
+	} else {
+		s->HostVMDynamicLevelsTrips = 0;
+	}
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: GPUVMEnable = %u\n", __func__, p->GPUVMEnable);
+	dml_print("DML::%s: GPUVMPageTableLevels = %u\n", __func__, p->GPUVMPageTableLevels);
+	dml_print("DML::%s: DCCEnable = %u\n", __func__, p->myPipe->DCCEnable);
+	dml_print("DML::%s: VStartup = %u\n", __func__, p->VStartup);
+	dml_print("DML::%s: MaxVStartup = %u\n", __func__, p->MaxVStartup);
+	dml_print("DML::%s: HostVMEnable = %u\n", __func__, p->HostVMEnable);
+	dml_print("DML::%s: HostVMInefficiencyFactor= %f\n", __func__, p->HostVMInefficiencyFactor);
+	dml_print("DML::%s: myPipe->Dppclk = %f\n", __func__, p->myPipe->Dppclk);
+#endif
+	CalculateVUpdateAndDynamicMetadataParameters(
+	p->MaxInterDCNTileRepeaters,
+	p->myPipe->Dppclk,
+	p->myPipe->Dispclk,
+	p->myPipe->DCFClkDeepSleep,
+	p->myPipe->PixelClock,
+	p->myPipe->HTotal,
+	p->myPipe->VBlank,
+	p->DynamicMetadataTransmittedBytes,
+	p->DynamicMetadataLinesBeforeActiveRequired,
+	p->myPipe->InterlaceEnable,
+	p->myPipe->ProgressiveToInterlaceUnitInOPP,
+	p->TSetup,
+
+	// Output
+	&s->Tdmbf,
+	&s->Tdmec,
+	&s->Tdmsks,
+	p->VUpdateOffsetPix,
+	p->VUpdateWidthPix,
+	p->VReadyOffsetPix);
+
+	s->LineTime = p->myPipe->HTotal / p->myPipe->PixelClock;
+	s->trip_to_mem = p->UrgentLatency;
+	s->Tvm_trips = p->UrgentExtraLatency + s->trip_to_mem * (p->GPUVMPageTableLevels * (s->HostVMDynamicLevelsTrips + 1) - 1);
+
+	if (p->DynamicMetadataVMEnabled == true) {
+		*p->Tdmdl = p->TWait + s->Tvm_trips + s->trip_to_mem;
+	} else {
+		*p->Tdmdl = p->TWait + p->UrgentExtraLatency;
+	}
+
+#ifdef __DML_VBA_ALLOW_DELTA__
+	if (DynamicMetadataEnable == false) {
+		*Tdmdl = 0.0;
+	}
+#endif
+
+	if (p->DynamicMetadataEnable == true) {
+		if (p->VStartup * s->LineTime < *p->TSetup + *p->Tdmdl + s->Tdmbf + s->Tdmec + s->Tdmsks) {
+			*p->NotEnoughTimeForDynamicMetadata = true;
+			dml_print("DML::%s: Not Enough Time for Dynamic Meta!\n", __func__);
+			dml_print("DML::%s: Tdmbf: %fus - time for dmd transfer from dchub to dio output buffer\n", __func__, s->Tdmbf);
+			dml_print("DML::%s: Tdmec: %fus - time dio takes to transfer dmd\n", __func__, s->Tdmec);
+			dml_print("DML::%s: Tdmsks: %fus - time before active dmd must complete transmission at dio\n", __func__, s->Tdmsks);
+			dml_print("DML::%s: Tdmdl: %fus - time for fabric to become ready and fetch dmd \n", __func__, *p->Tdmdl);
+		} else {
+			*p->NotEnoughTimeForDynamicMetadata = false;
+		}
+	} else {
+		*p->NotEnoughTimeForDynamicMetadata = false;
+	}
+
+	*p->Tdmdl_vm = (p->DynamicMetadataEnable == true && p->DynamicMetadataVMEnabled == true && p->GPUVMEnable == true ? p->TWait + s->Tvm_trips : 0);
+
+	if (p->myPipe->ScalerEnabled)
+		s->DPPCycles = (dml_uint_t)(p->DPPCLKDelaySubtotalPlusCNVCFormater + p->DPPCLKDelaySCL);
+	else
+		s->DPPCycles = (dml_uint_t)(p->DPPCLKDelaySubtotalPlusCNVCFormater + p->DPPCLKDelaySCLLBOnly);
+
+	s->DPPCycles = (dml_uint_t)(s->DPPCycles + p->myPipe->NumberOfCursors * p->DPPCLKDelayCNVCCursor);
+
+	s->DISPCLKCycles = (dml_uint_t)p->DISPCLKDelaySubtotal;
+
+	if (p->myPipe->Dppclk == 0.0 || p->myPipe->Dispclk == 0.0)
+		return true;
+
+	*p->DSTXAfterScaler = (dml_uint_t) dml_round(s->DPPCycles * p->myPipe->PixelClock / p->myPipe->Dppclk + s->DISPCLKCycles * p->myPipe->PixelClock / p->myPipe->Dispclk + p->DSCDelay, 1.0);
+	*p->DSTXAfterScaler = (dml_uint_t) dml_round(*p->DSTXAfterScaler + (p->myPipe->ODMMode != dml_odm_mode_bypass ? 18 : 0) + (p->myPipe->DPPPerSurface - 1) * p->DPP_RECOUT_WIDTH +
+						((p->myPipe->ODMMode == dml_odm_mode_split_1to2 || p->myPipe->ODMMode == dml_odm_mode_mso_1to2) ? (dml_float_t)p->myPipe->HActive / 2.0 : 0) +
+						((p->myPipe->ODMMode == dml_odm_mode_mso_1to4) ? (dml_float_t)p->myPipe->HActive * 3.0 / 4.0 : 0), 1.0);
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: DPPCycles = %u\n", __func__, s->DPPCycles);
+	dml_print("DML::%s: PixelClock = %f\n", __func__, p->myPipe->PixelClock);
+	dml_print("DML::%s: Dppclk = %f\n", __func__, p->myPipe->Dppclk);
+	dml_print("DML::%s: DISPCLKCycles = %u\n", __func__, s->DISPCLKCycles);
+	dml_print("DML::%s: DISPCLK = %f\n", __func__, p->myPipe->Dispclk);
+	dml_print("DML::%s: DSCDelay = %u\n", __func__, p->DSCDelay);
+	dml_print("DML::%s: ODMMode = %u\n", __func__, p->myPipe->ODMMode);
+	dml_print("DML::%s: DPP_RECOUT_WIDTH = %u\n", __func__, p->DPP_RECOUT_WIDTH);
+	dml_print("DML::%s: DSTXAfterScaler = %u\n", __func__, *p->DSTXAfterScaler);
+#endif
+
+	if (p->OutputFormat == dml_420 || (p->myPipe->InterlaceEnable && p->myPipe->ProgressiveToInterlaceUnitInOPP))
+		*p->DSTYAfterScaler = 1;
+	else
+		*p->DSTYAfterScaler = 0;
+
+	s->DSTTotalPixelsAfterScaler = *p->DSTYAfterScaler * p->myPipe->HTotal + *p->DSTXAfterScaler;
+	*p->DSTYAfterScaler = (dml_uint_t)(dml_floor(s->DSTTotalPixelsAfterScaler / p->myPipe->HTotal, 1));
+	*p->DSTXAfterScaler = (dml_uint_t)(s->DSTTotalPixelsAfterScaler - ((dml_float_t) (*p->DSTYAfterScaler * p->myPipe->HTotal)));
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: DSTXAfterScaler = %u (final)\n", __func__,  *p->DSTXAfterScaler);
+	dml_print("DML::%s: DSTYAfterScaler = %u (final)\n", __func__, *p->DSTYAfterScaler);
+#endif
+
+	s->MyError = false;
+
+	s->Tr0_trips = s->trip_to_mem * (s->HostVMDynamicLevelsTrips + 1);
+
+	if (p->GPUVMEnable == true) {
+		s->Tvm_trips_rounded = dml_ceil(4.0 * s->Tvm_trips / s->LineTime, 1.0) / 4.0 * s->LineTime;
+		s->Tr0_trips_rounded = dml_ceil(4.0 * s->Tr0_trips / s->LineTime, 1.0) / 4.0 * s->LineTime;
+		if (p->GPUVMPageTableLevels >= 3) {
+			*p->Tno_bw = p->UrgentExtraLatency + s->trip_to_mem * (dml_float_t) ((p->GPUVMPageTableLevels - 2) * (s->HostVMDynamicLevelsTrips + 1) - 1);
+	} else if (p->GPUVMPageTableLevels == 1 && p->myPipe->DCCEnable != true) {
+			s->Tr0_trips_rounded = dml_ceil(4.0 * p->UrgentExtraLatency / s->LineTime, 1.0) / 4.0 * s->LineTime;
+			*p->Tno_bw = p->UrgentExtraLatency;
+		} else {
+			*p->Tno_bw = 0;
+		}
+	} else if (p->myPipe->DCCEnable == true) {
+		s->Tvm_trips_rounded = s->LineTime / 4.0;
+		s->Tr0_trips_rounded = dml_ceil(4.0 * s->Tr0_trips / s->LineTime, 1.0) / 4.0 * s->LineTime;
+		*p->Tno_bw = 0;
+	} else {
+		s->Tvm_trips_rounded = s->LineTime / 4.0;
+		s->Tr0_trips_rounded = s->LineTime / 2.0;
+		*p->Tno_bw = 0;
+	}
+	s->Tvm_trips_rounded = dml_max(s->Tvm_trips_rounded, s->LineTime / 4.0);
+	s->Tr0_trips_rounded = dml_max(s->Tr0_trips_rounded, s->LineTime / 4.0);
+
+	if (p->myPipe->SourcePixelFormat == dml_420_8 || p->myPipe->SourcePixelFormat == dml_420_10 || p->myPipe->SourcePixelFormat == dml_420_12) {
+		s->bytes_pp = p->myPipe->BytePerPixelY + p->myPipe->BytePerPixelC / 4;
+	} else {
+		s->bytes_pp = p->myPipe->BytePerPixelY + p->myPipe->BytePerPixelC;
+	}
+
+	s->prefetch_bw_pr = s->bytes_pp * p->myPipe->PixelClock / (dml_float_t)p->myPipe->DPPPerSurface;
+	if (p->myPipe->VRatio < 1.0)
+		s->prefetch_bw_pr = p->myPipe->VRatio * s->prefetch_bw_pr;
+
+	s->max_Tsw = (dml_max(p->PrefetchSourceLinesY, p->PrefetchSourceLinesC) * s->LineTime);
+
+	s->prefetch_sw_bytes = p->PrefetchSourceLinesY * p->swath_width_luma_ub * p->myPipe->BytePerPixelY + p->PrefetchSourceLinesC * p->swath_width_chroma_ub * p->myPipe->BytePerPixelC;
+	s->prefetch_bw_oto = dml_max(s->prefetch_bw_pr, s->prefetch_sw_bytes / s->max_Tsw);
+
+	s->min_Lsw_oto = dml_max(p->PrefetchSourceLinesY, p->PrefetchSourceLinesC) / __DML_MAX_VRATIO_PRE_OTO__;
+	s->min_Lsw_oto = dml_max(s->min_Lsw_oto, 1.0);
+	s->Lsw_oto = dml_ceil(4.0 * dml_max(s->prefetch_sw_bytes / s->prefetch_bw_oto / s->LineTime, s->min_Lsw_oto), 1.0) / 4.0;
+
+	if (p->GPUVMEnable == true) {
+		s->Tvm_oto = dml_max3(
+			s->Tvm_trips,
+			*p->Tno_bw + p->PDEAndMetaPTEBytesFrame * p->HostVMInefficiencyFactor / s->prefetch_bw_oto,
+			s->LineTime / 4.0);
+	} else
+		s->Tvm_oto = s->LineTime / 4.0;
+
+	if ((p->GPUVMEnable == true || p->myPipe->DCCEnable == true)) {
+		s->Tr0_oto = dml_max4(
+			s->Tr0_trips,
+			(p->MetaRowByte + p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor) / s->prefetch_bw_oto,
+			(s->LineTime - s->Tvm_oto)/2.0,
+			s->LineTime / 4.0);
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: Tr0_oto max0 = %f\n", __func__, (p->MetaRowByte + p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor) / s->prefetch_bw_oto);
+	dml_print("DML::%s: Tr0_oto max1 = %f\n", __func__, s->Tr0_trips);
+	dml_print("DML::%s: Tr0_oto max2 = %f\n", __func__, s->LineTime - s->Tvm_oto);
+	dml_print("DML::%s: Tr0_oto max3 = %f\n", __func__, s->LineTime / 4);
+#endif
+	} else
+		s->Tr0_oto = (s->LineTime - s->Tvm_oto) / 2.0;
+
+	s->Tvm_oto_lines = dml_ceil(4.0 * s->Tvm_oto / s->LineTime, 1) / 4.0;
+	s->Tr0_oto_lines = dml_ceil(4.0 * s->Tr0_oto / s->LineTime, 1) / 4.0;
+	s->dst_y_prefetch_oto = s->Tvm_oto_lines + 2 * s->Tr0_oto_lines + s->Lsw_oto;
+
+	s->dst_y_prefetch_equ = p->VStartup - (*p->TSetup + dml_max(p->TWait + p->TCalc, *p->Tdmdl)) / s->LineTime - (*p->DSTYAfterScaler + (dml_float_t) *p->DSTXAfterScaler / (dml_float_t)p->myPipe->HTotal);
+	s->dst_y_prefetch_equ = dml_min(s->dst_y_prefetch_equ, 63.75); // limit to the reg limit of U6.2 for DST_Y_PREFETCH
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: HTotal = %u\n", __func__, p->myPipe->HTotal);
+	dml_print("DML::%s: min_Lsw_oto = %f\n", __func__, s->min_Lsw_oto);
+	dml_print("DML::%s: *Tno_bw = %f\n", __func__, *p->Tno_bw);
+	dml_print("DML::%s: UrgentExtraLatency = %f\n", __func__, p->UrgentExtraLatency);
+	dml_print("DML::%s: trip_to_mem = %f\n", __func__, s->trip_to_mem);
+	dml_print("DML::%s: BytePerPixelY = %u\n", __func__, p->myPipe->BytePerPixelY);
+	dml_print("DML::%s: PrefetchSourceLinesY = %f\n", __func__, p->PrefetchSourceLinesY);
+	dml_print("DML::%s: swath_width_luma_ub = %u\n", __func__, p->swath_width_luma_ub);
+	dml_print("DML::%s: BytePerPixelC = %u\n", __func__, p->myPipe->BytePerPixelC);
+	dml_print("DML::%s: PrefetchSourceLinesC = %f\n", __func__, p->PrefetchSourceLinesC);
+	dml_print("DML::%s: swath_width_chroma_ub = %u\n", __func__, p->swath_width_chroma_ub);
+	dml_print("DML::%s: prefetch_sw_bytes = %f\n", __func__, s->prefetch_sw_bytes);
+	dml_print("DML::%s: bytes_pp = %f\n", __func__, s->bytes_pp);
+	dml_print("DML::%s: PDEAndMetaPTEBytesFrame = %u\n", __func__, p->PDEAndMetaPTEBytesFrame);
+	dml_print("DML::%s: MetaRowByte = %u\n", __func__, p->MetaRowByte);
+	dml_print("DML::%s: PixelPTEBytesPerRow = %u\n", __func__, p->PixelPTEBytesPerRow);
+	dml_print("DML::%s: HostVMInefficiencyFactor = %f\n", __func__, p->HostVMInefficiencyFactor);
+	dml_print("DML::%s: Tvm_trips = %f\n", __func__, s->Tvm_trips);
+	dml_print("DML::%s: Tr0_trips = %f\n", __func__, s->Tr0_trips);
+	dml_print("DML::%s: prefetch_bw_oto = %f\n", __func__, s->prefetch_bw_oto);
+	dml_print("DML::%s: Tr0_oto = %f\n", __func__, s->Tr0_oto);
+	dml_print("DML::%s: Tvm_oto = %f\n", __func__, s->Tvm_oto);
+	dml_print("DML::%s: Tvm_oto_lines = %f\n", __func__, s->Tvm_oto_lines);
+	dml_print("DML::%s: Tr0_oto_lines = %f\n", __func__, s->Tr0_oto_lines);
+	dml_print("DML::%s: Lsw_oto = %f\n", __func__, s->Lsw_oto);
+	dml_print("DML::%s: dst_y_prefetch_oto = %f\n", __func__, s->dst_y_prefetch_oto);
+	dml_print("DML::%s: dst_y_prefetch_equ = %f\n", __func__, s->dst_y_prefetch_equ);
+#endif
+
+	s->dst_y_prefetch_equ = dml_floor(4.0 * (s->dst_y_prefetch_equ + 0.125), 1) / 4.0;
+	s->Tpre_rounded = s->dst_y_prefetch_equ * s->LineTime;
+
+	dml_print("DML::%s: dst_y_prefetch_equ: %f (after round)\n", __func__, s->dst_y_prefetch_equ);
+
+	dml_print("DML::%s: LineTime: %f\n", __func__, s->LineTime);
+	dml_print("DML::%s: VStartup: %u\n", __func__, p->VStartup);
+	dml_print("DML::%s: Tvstartup: %fus - time between vstartup and first pixel of active\n", __func__, p->VStartup * s->LineTime);
+	dml_print("DML::%s: TSetup: %fus - time from vstartup to vready\n", __func__, *p->TSetup);
+	dml_print("DML::%s: TCalc: %fus - time for calculations in dchub starting at vready\n", __func__, p->TCalc);
+	dml_print("DML::%s: TWait: %fus - time for fabric to become ready max(pstate exit,cstate enter/exit, urgent latency) after TCalc\n", __func__, p->TWait);
+	dml_print("DML::%s: Tdmbf: %fus - time for dmd transfer from dchub to dio output buffer\n", __func__, s->Tdmbf);
+	dml_print("DML::%s: Tdmec: %fus - time dio takes to transfer dmd\n", __func__, s->Tdmec);
+	dml_print("DML::%s: Tdmsks: %fus - time before active dmd must complete transmission at dio\n", __func__, s->Tdmsks);
+	dml_print("DML::%s: Tdmdl_vm: %fus - time for vm stages of dmd \n", __func__, *p->Tdmdl_vm);
+	dml_print("DML::%s: Tdmdl: %fus - time for fabric to become ready and fetch dmd \n", __func__, *p->Tdmdl);
+	dml_print("DML::%s: DSTXAfterScaler: %u pixels - number of pixel clocks pipeline and buffer delay after scaler \n", __func__, *p->DSTXAfterScaler);
+	dml_print("DML::%s: DSTYAfterScaler: %u lines - number of lines of pipeline and buffer delay after scaler \n", __func__, *p->DSTYAfterScaler);
+
+	s->dep_bytes = dml_max(p->PDEAndMetaPTEBytesFrame * p->HostVMInefficiencyFactor, p->MetaRowByte + p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor);
+
+	if (s->prefetch_sw_bytes < s->dep_bytes) {
+		s->prefetch_sw_bytes = 2 * s->dep_bytes;
+	}
+
+	*p->DestinationLinesToRequestVMInVBlank = 0;
+	*p->DestinationLinesToRequestRowInVBlank = 0;
+	*p->VRatioPrefetchY = 0;
+	*p->VRatioPrefetchC = 0;
+	*p->RequiredPrefetchPixDataBWLuma = 0;
+	if (s->dst_y_prefetch_equ > 1) {
+
+		if (s->Tpre_rounded - *p->Tno_bw > 0) {
+		s->PrefetchBandwidth1 = (p->PDEAndMetaPTEBytesFrame * p->HostVMInefficiencyFactor + 2 * p->MetaRowByte
+					+ 2 * p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor
+					+ s->prefetch_sw_bytes)
+					/ (s->Tpre_rounded - *p->Tno_bw);
+			s->Tsw_est1 = s->prefetch_sw_bytes / s->PrefetchBandwidth1;
+		} else
+			s->PrefetchBandwidth1 = 0;
+
+		if (p->VStartup == p->MaxVStartup && (s->Tsw_est1 / s->LineTime < s->min_Lsw_oto) && s->Tpre_rounded - s->min_Lsw_oto * s->LineTime - 0.75 * s->LineTime - *p->Tno_bw > 0) {
+			s->PrefetchBandwidth1 = (p->PDEAndMetaPTEBytesFrame * p->HostVMInefficiencyFactor + 2 * p->MetaRowByte + 2 * p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor) /
+								(s->Tpre_rounded - s->min_Lsw_oto * s->LineTime - 0.75 * s->LineTime - *p->Tno_bw);
+		}
+
+		if (s->Tpre_rounded - *p->Tno_bw - 2 * s->Tr0_trips_rounded > 0)
+			s->PrefetchBandwidth2 = (p->PDEAndMetaPTEBytesFrame * p->HostVMInefficiencyFactor + s->prefetch_sw_bytes) /
+									(s->Tpre_rounded - *p->Tno_bw - 2 * s->Tr0_trips_rounded);
+		else
+			s->PrefetchBandwidth2 = 0;
+
+		if (s->Tpre_rounded - s->Tvm_trips_rounded > 0) {
+			s->PrefetchBandwidth3 = (2 * p->MetaRowByte + 2 * p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor + s->prefetch_sw_bytes) /
+									(s->Tpre_rounded - s->Tvm_trips_rounded);
+			s->Tsw_est3 = s->prefetch_sw_bytes / s->PrefetchBandwidth3;
+	}
+		else
+			s->PrefetchBandwidth3 = 0;
+
+
+		if (p->VStartup == p->MaxVStartup && (s->Tsw_est3 / s->LineTime < s->min_Lsw_oto) && s->Tpre_rounded - s->min_Lsw_oto * s->LineTime - 0.5 * s->LineTime - s->Tvm_trips_rounded > 0) {
+			s->PrefetchBandwidth3 = (2 * p->MetaRowByte + 2 * p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor) / (s->Tpre_rounded - s->min_Lsw_oto * s->LineTime - 0.5 * s->LineTime - s->Tvm_trips_rounded);
+		}
+
+		if (s->Tpre_rounded - s->Tvm_trips_rounded - 2 * s->Tr0_trips_rounded > 0)
+			s->PrefetchBandwidth4 = s->prefetch_sw_bytes / (s->Tpre_rounded - s->Tvm_trips_rounded - 2 * s->Tr0_trips_rounded);
+		else
+			s->PrefetchBandwidth4 = 0;
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: Tpre_rounded: %f\n", __func__, s->Tpre_rounded);
+		dml_print("DML::%s: Tno_bw: %f\n", __func__, *p->Tno_bw);
+		dml_print("DML::%s: Tvm_trips_rounded: %f\n", __func__, s->Tvm_trips_rounded);
+		dml_print("DML::%s: Tsw_est1: %f\n", __func__, s->Tsw_est1);
+		dml_print("DML::%s: Tsw_est3: %f\n", __func__, s->Tsw_est3);
+		dml_print("DML::%s: PrefetchBandwidth1: %f\n", __func__, s->PrefetchBandwidth1);
+		dml_print("DML::%s: PrefetchBandwidth2: %f\n", __func__, s->PrefetchBandwidth2);
+		dml_print("DML::%s: PrefetchBandwidth3: %f\n", __func__, s->PrefetchBandwidth3);
+		dml_print("DML::%s: PrefetchBandwidth4: %f\n", __func__, s->PrefetchBandwidth4);
+#endif
+		{
+			dml_bool_t Case1OK;
+			dml_bool_t Case2OK;
+			dml_bool_t Case3OK;
+
+			if (s->PrefetchBandwidth1 > 0) {
+				if (*p->Tno_bw + p->PDEAndMetaPTEBytesFrame * p->HostVMInefficiencyFactor / s->PrefetchBandwidth1 >= s->Tvm_trips_rounded && (p->MetaRowByte + p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor) / s->PrefetchBandwidth1 >= s->Tr0_trips_rounded) {
+					Case1OK = true;
+				} else {
+					Case1OK = false;
+				}
+			} else {
+				Case1OK = false;
+			}
+
+			if (s->PrefetchBandwidth2 > 0) {
+				if (*p->Tno_bw + p->PDEAndMetaPTEBytesFrame * p->HostVMInefficiencyFactor / s->PrefetchBandwidth2 >= s->Tvm_trips_rounded && (p->MetaRowByte + p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor) / s->PrefetchBandwidth2 < s->Tr0_trips_rounded) {
+					Case2OK = true;
+				} else {
+					Case2OK = false;
+				}
+			} else {
+				Case2OK = false;
+			}
+
+			if (s->PrefetchBandwidth3 > 0) {
+				if (*p->Tno_bw + p->PDEAndMetaPTEBytesFrame * p->HostVMInefficiencyFactor / s->PrefetchBandwidth3 < s->Tvm_trips_rounded && (p->MetaRowByte + p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor) / s->PrefetchBandwidth3 >= s->Tr0_trips_rounded) {
+					Case3OK = true;
+				} else {
+					Case3OK = false;
+				}
+			} else {
+				Case3OK = false;
+			}
+
+			if (Case1OK) {
+				s->prefetch_bw_equ = s->PrefetchBandwidth1;
+			} else if (Case2OK) {
+				s->prefetch_bw_equ = s->PrefetchBandwidth2;
+			} else if (Case3OK) {
+				s->prefetch_bw_equ = s->PrefetchBandwidth3;
+			} else {
+				s->prefetch_bw_equ = s->PrefetchBandwidth4;
+			}
+
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: Case1OK: %u\n", __func__, Case1OK);
+			dml_print("DML::%s: Case2OK: %u\n", __func__, Case2OK);
+			dml_print("DML::%s: Case3OK: %u\n", __func__, Case3OK);
+			dml_print("DML::%s: prefetch_bw_equ: %f\n", __func__, s->prefetch_bw_equ);
+#endif
+
+			if (s->prefetch_bw_equ > 0) {
+				if (p->GPUVMEnable == true) {
+					s->Tvm_equ = dml_max3(*p->Tno_bw + p->PDEAndMetaPTEBytesFrame * p->HostVMInefficiencyFactor / s->prefetch_bw_equ, s->Tvm_trips, s->LineTime / 4);
+				} else {
+					s->Tvm_equ = s->LineTime / 4;
+				}
+
+				if ((p->GPUVMEnable == true || p->myPipe->DCCEnable == true)) {
+					s->Tr0_equ = dml_max4((p->MetaRowByte + p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor) / s->prefetch_bw_equ, s->Tr0_trips, (s->LineTime - s->Tvm_equ) / 2, s->LineTime / 4);
+				} else {
+					s->Tr0_equ = (s->LineTime - s->Tvm_equ) / 2;
+				}
+			} else {
+				s->Tvm_equ = 0;
+				s->Tr0_equ = 0;
+				dml_print("DML::%s: prefetch_bw_equ equals 0!\n", __func__);
+			}
+		}
+
+
+		if (s->dst_y_prefetch_oto < s->dst_y_prefetch_equ) {
+			*p->DestinationLinesForPrefetch = s->dst_y_prefetch_oto;
+			s->TimeForFetchingMetaPTE = s->Tvm_oto;
+			s->TimeForFetchingRowInVBlank = s->Tr0_oto;
+
+			*p->DestinationLinesToRequestVMInVBlank = dml_ceil(4.0 * s->TimeForFetchingMetaPTE / s->LineTime, 1.0) / 4.0;
+			*p->DestinationLinesToRequestRowInVBlank = dml_ceil(4.0 * s->TimeForFetchingRowInVBlank / s->LineTime, 1.0) / 4.0;
+		} else {
+			*p->DestinationLinesForPrefetch = s->dst_y_prefetch_equ;
+			s->TimeForFetchingMetaPTE = s->Tvm_equ;
+			s->TimeForFetchingRowInVBlank = s->Tr0_equ;
+
+			if (p->VStartup == p->MaxVStartup && p->EnhancedPrefetchScheduleAccelerationFinal != 0) {
+				*p->DestinationLinesToRequestVMInVBlank = dml_floor(4.0 * s->TimeForFetchingMetaPTE / s->LineTime, 1.0) / 4.0;
+				*p->DestinationLinesToRequestRowInVBlank = dml_floor(4.0 * s->TimeForFetchingRowInVBlank / s->LineTime, 1.0) / 4.0;
+			} else {
+				*p->DestinationLinesToRequestVMInVBlank = dml_ceil(4.0 * s->TimeForFetchingMetaPTE / s->LineTime, 1.0) / 4.0;
+				*p->DestinationLinesToRequestRowInVBlank = dml_ceil(4.0 * s->TimeForFetchingRowInVBlank / s->LineTime, 1.0) / 4.0;
+			}
+		}
+
+		s->LinesToRequestPrefetchPixelData = *p->DestinationLinesForPrefetch - *p->DestinationLinesToRequestVMInVBlank - 2 * *p->DestinationLinesToRequestRowInVBlank;
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: DestinationLinesForPrefetch = %f\n", __func__, *p->DestinationLinesForPrefetch);
+		dml_print("DML::%s: DestinationLinesToRequestVMInVBlank = %f\n", __func__, *p->DestinationLinesToRequestVMInVBlank);
+		dml_print("DML::%s: TimeForFetchingRowInVBlank = %f\n", __func__, s->TimeForFetchingRowInVBlank);
+		dml_print("DML::%s: LineTime = %f\n", __func__, s->LineTime);
+		dml_print("DML::%s: DestinationLinesToRequestRowInVBlank = %f\n", __func__, *p->DestinationLinesToRequestRowInVBlank);
+		dml_print("DML::%s: PrefetchSourceLinesY = %f\n", __func__, p->PrefetchSourceLinesY);
+		dml_print("DML::%s: LinesToRequestPrefetchPixelData = %f\n", __func__, s->LinesToRequestPrefetchPixelData);
+#endif
+
+		if (s->LinesToRequestPrefetchPixelData >= 1 && s->prefetch_bw_equ > 0) {
+			*p->VRatioPrefetchY = (dml_float_t)p->PrefetchSourceLinesY / s->LinesToRequestPrefetchPixelData;
+			*p->VRatioPrefetchY = dml_max(*p->VRatioPrefetchY, 1.0);
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: VRatioPrefetchY = %f\n", __func__, *p->VRatioPrefetchY);
+			dml_print("DML::%s: SwathHeightY = %u\n", __func__, p->SwathHeightY);
+			dml_print("DML::%s: VInitPreFillY = %u\n", __func__, p->VInitPreFillY);
+#endif
+			if ((p->SwathHeightY > 4) && (p->VInitPreFillY > 3)) {
+				if (s->LinesToRequestPrefetchPixelData > (p->VInitPreFillY - 3.0) / 2.0) {
+					*p->VRatioPrefetchY = dml_max(*p->VRatioPrefetchY,
+										(dml_float_t)p->MaxNumSwathY * p->SwathHeightY / (s->LinesToRequestPrefetchPixelData - (p->VInitPreFillY - 3.0) / 2.0));
+			} else {
+				s->MyError = true;
+				dml_print("DML::%s: MyErr set. LinesToRequestPrefetchPixelData=%f VinitPreFillY=%u\n", __func__, s->LinesToRequestPrefetchPixelData, p->VInitPreFillY);
+				*p->VRatioPrefetchY = 0;
+			}
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: VRatioPrefetchY = %f\n", __func__, *p->VRatioPrefetchY);
+			dml_print("DML::%s: PrefetchSourceLinesY = %f\n", __func__, p->PrefetchSourceLinesY);
+			dml_print("DML::%s: MaxNumSwathY = %u\n", __func__, p->MaxNumSwathY);
+#endif
+			}
+
+			*p->VRatioPrefetchC = (dml_float_t)p->PrefetchSourceLinesC / s->LinesToRequestPrefetchPixelData;
+			*p->VRatioPrefetchC = dml_max(*p->VRatioPrefetchC, 1.0);
+
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: VRatioPrefetchC = %f\n", __func__, *p->VRatioPrefetchC);
+			dml_print("DML::%s: SwathHeightC = %u\n", __func__, p->SwathHeightC);
+			dml_print("DML::%s: VInitPreFillC = %u\n", __func__, p->VInitPreFillC);
+#endif
+			if ((p->SwathHeightC > 4) && (p->VInitPreFillC > 3)) {
+				if (s->LinesToRequestPrefetchPixelData > (p->VInitPreFillC - 3.0) / 2.0) {
+					*p->VRatioPrefetchC = dml_max(*p->VRatioPrefetchC, (dml_float_t)p->MaxNumSwathC * p->SwathHeightC / (s->LinesToRequestPrefetchPixelData - (p->VInitPreFillC - 3.0) / 2.0));
+				} else {
+					s->MyError = true;
+					dml_print("DML::%s: MyErr set. LinesToRequestPrefetchPixelData=%f VInitPreFillC=%u\n", __func__, s->LinesToRequestPrefetchPixelData, p->VInitPreFillC);
+					*p->VRatioPrefetchC = 0;
+				}
+#ifdef __DML_VBA_DEBUG__
+				dml_print("DML::%s: VRatioPrefetchC = %f\n", __func__, *p->VRatioPrefetchC);
+				dml_print("DML::%s: PrefetchSourceLinesC = %f\n", __func__, p->PrefetchSourceLinesC);
+				dml_print("DML::%s: MaxNumSwathC = %u\n", __func__, p->MaxNumSwathC);
+#endif
+			}
+
+			*p->RequiredPrefetchPixDataBWLuma = (dml_float_t)p->PrefetchSourceLinesY / s->LinesToRequestPrefetchPixelData
+				* p->myPipe->BytePerPixelY
+				* p->swath_width_luma_ub / s->LineTime;
+
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: BytePerPixelY = %u\n", __func__, p->myPipe->BytePerPixelY);
+			dml_print("DML::%s: swath_width_luma_ub = %u\n", __func__, p->swath_width_luma_ub);
+			dml_print("DML::%s: LineTime = %f\n", __func__, s->LineTime);
+			dml_print("DML::%s: RequiredPrefetchPixDataBWLuma = %f\n", __func__, *p->RequiredPrefetchPixDataBWLuma);
+#endif
+			*p->RequiredPrefetchPixDataBWChroma = (dml_float_t)p->PrefetchSourceLinesC / s->LinesToRequestPrefetchPixelData
+				*p->myPipe->BytePerPixelC
+				*p->swath_width_chroma_ub / s->LineTime;
+		} else {
+			s->MyError = true;
+			dml_print("DML:%s: MyErr set. LinesToRequestPrefetchPixelData: %f, should be > 0\n", __func__, s->LinesToRequestPrefetchPixelData);
+			*p->VRatioPrefetchY = 0;
+			*p->VRatioPrefetchC = 0;
+			*p->RequiredPrefetchPixDataBWLuma = 0;
+			*p->RequiredPrefetchPixDataBWChroma = 0;
+		}
+
+		dml_print("DML: Tpre: %fus - sum of time to request meta pte, 2 x data pte + meta data, swaths\n", (dml_float_t)s->LinesToRequestPrefetchPixelData * s->LineTime + 2.0 * s->TimeForFetchingRowInVBlank + s->TimeForFetchingMetaPTE);
+		dml_print("DML: Tvm: %fus - time to fetch page tables for meta surface\n", s->TimeForFetchingMetaPTE);
+		dml_print("DML: Tr0: %fus - time to fetch first row of data pagetables and first row of meta data (done in parallel)\n", s->TimeForFetchingRowInVBlank);
+		dml_print("DML: Tsw: %fus = time to fetch enough pixel data and cursor data to feed the scalers init position and detile\n", (dml_float_t)s->LinesToRequestPrefetchPixelData * s->LineTime);
+		dml_print("DML: To: %fus - time for propagation from scaler to optc\n", (*p->DSTYAfterScaler + ((dml_float_t) (*p->DSTXAfterScaler) / (dml_float_t)p->myPipe->HTotal)) * s->LineTime);
+		dml_print("DML: Tvstartup - TSetup - Tcalc - Twait - Tpre - To > 0\n");
+		dml_print("DML: Tslack(pre): %fus - time left over in schedule\n", p->VStartup * s->LineTime - s->TimeForFetchingMetaPTE - 2 * s->TimeForFetchingRowInVBlank - (*p->DSTYAfterScaler + ((dml_float_t) (*p->DSTXAfterScaler) / (dml_float_t)p->myPipe->HTotal)) * s->LineTime - p->TWait - p->TCalc - *p->TSetup);
+		dml_print("DML: row_bytes = dpte_row_bytes (per_pipe) = PixelPTEBytesPerRow = : %u\n", p->PixelPTEBytesPerRow);
+
+	} else {
+		s->MyError = true;
+		dml_print("DML::%s: MyErr set, dst_y_prefetch_equ = %f (should be > 1)\n", __func__, s->dst_y_prefetch_equ);
+		s->TimeForFetchingMetaPTE = 0;
+		s->TimeForFetchingRowInVBlank = 0;
+		*p->DestinationLinesToRequestVMInVBlank = 0;
+		*p->DestinationLinesToRequestRowInVBlank = 0;
+		s->LinesToRequestPrefetchPixelData = 0;
+		*p->VRatioPrefetchY = 0;
+		*p->VRatioPrefetchC = 0;
+		*p->RequiredPrefetchPixDataBWLuma = 0;
+		*p->RequiredPrefetchPixDataBWChroma = 0;
+	}
+
+	{
+		dml_float_t prefetch_vm_bw;
+		dml_float_t prefetch_row_bw;
+
+		if (p->PDEAndMetaPTEBytesFrame == 0) {
+			prefetch_vm_bw = 0;
+		} else if (*p->DestinationLinesToRequestVMInVBlank > 0) {
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: PDEAndMetaPTEBytesFrame = %u\n", __func__, p->PDEAndMetaPTEBytesFrame);
+			dml_print("DML::%s: HostVMInefficiencyFactor = %f\n", __func__, p->HostVMInefficiencyFactor);
+			dml_print("DML::%s: DestinationLinesToRequestVMInVBlank = %f\n", __func__, *p->DestinationLinesToRequestVMInVBlank);
+			dml_print("DML::%s: LineTime = %f\n", __func__, s->LineTime);
+#endif
+		prefetch_vm_bw = p->PDEAndMetaPTEBytesFrame * p->HostVMInefficiencyFactor / (*p->DestinationLinesToRequestVMInVBlank * s->LineTime);
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: prefetch_vm_bw = %f\n", __func__, prefetch_vm_bw);
+#endif
+		} else {
+			prefetch_vm_bw = 0;
+			s->MyError = true;
+			dml_print("DML::%s: MyErr set. DestinationLinesToRequestVMInVBlank=%f (should be > 0)\n", __func__, *p->DestinationLinesToRequestVMInVBlank);
+		}
+
+		if (p->MetaRowByte + p->PixelPTEBytesPerRow == 0) {
+			prefetch_row_bw = 0;
+		} else if (*p->DestinationLinesToRequestRowInVBlank > 0) {
+			prefetch_row_bw = (p->MetaRowByte + p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor) / (*p->DestinationLinesToRequestRowInVBlank * s->LineTime);
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: MetaRowByte = %u\n", __func__, p->MetaRowByte);
+		dml_print("DML::%s: PixelPTEBytesPerRow = %u\n", __func__, p->PixelPTEBytesPerRow);
+		dml_print("DML::%s: DestinationLinesToRequestRowInVBlank = %f\n", __func__, *p->DestinationLinesToRequestRowInVBlank);
+		dml_print("DML::%s: prefetch_row_bw = %f\n", __func__, prefetch_row_bw);
+#endif
+		} else {
+			prefetch_row_bw = 0;
+			s->MyError = true;
+			dml_print("DML::%s: MyErr set. DestinationLinesToRequestRowInVBlank=%f (should be > 0)\n", __func__, *p->DestinationLinesToRequestRowInVBlank);
+		}
+
+		*p->prefetch_vmrow_bw = dml_max(prefetch_vm_bw, prefetch_row_bw);
+	}
+
+	if (s->MyError) {
+		s->TimeForFetchingMetaPTE = 0;
+		s->TimeForFetchingRowInVBlank = 0;
+		*p->DestinationLinesToRequestVMInVBlank = 0;
+		*p->DestinationLinesToRequestRowInVBlank = 0;
+		*p->DestinationLinesForPrefetch = 0;
+		s->LinesToRequestPrefetchPixelData = 0;
+		*p->VRatioPrefetchY = 0;
+		*p->VRatioPrefetchC = 0;
+		*p->RequiredPrefetchPixDataBWLuma = 0;
+		*p->RequiredPrefetchPixDataBWChroma = 0;
+	}
+
+	return s->MyError;
+} // CalculatePrefetchSchedule
+
+static void CalculateBytePerPixelAndBlockSizes(
+	enum dml_source_format_class SourcePixelFormat,
+	enum dml_swizzle_mode SurfaceTiling,
+
+	// Output
+	dml_uint_t *BytePerPixelY,
+	dml_uint_t *BytePerPixelC,
+	dml_float_t *BytePerPixelDETY,
+	dml_float_t *BytePerPixelDETC,
+	dml_uint_t *BlockHeight256BytesY,
+	dml_uint_t *BlockHeight256BytesC,
+	dml_uint_t *BlockWidth256BytesY,
+	dml_uint_t *BlockWidth256BytesC,
+	dml_uint_t *MacroTileHeightY,
+	dml_uint_t *MacroTileHeightC,
+	dml_uint_t *MacroTileWidthY,
+	dml_uint_t *MacroTileWidthC)
+{
+	if (SourcePixelFormat == dml_444_64) {
+		*BytePerPixelDETY = 8;
+		*BytePerPixelDETC = 0;
+		*BytePerPixelY = 8;
+		*BytePerPixelC = 0;
+	} else if (SourcePixelFormat == dml_444_32 || SourcePixelFormat == dml_rgbe) {
+		*BytePerPixelDETY = 4;
+		*BytePerPixelDETC = 0;
+		*BytePerPixelY = 4;
+		*BytePerPixelC = 0;
+	} else if (SourcePixelFormat == dml_444_16 || SourcePixelFormat == dml_mono_16) {
+		*BytePerPixelDETY = 2;
+		*BytePerPixelDETC = 0;
+		*BytePerPixelY = 2;
+		*BytePerPixelC = 0;
+	} else if (SourcePixelFormat == dml_444_8 || SourcePixelFormat == dml_mono_8) {
+		*BytePerPixelDETY = 1;
+		*BytePerPixelDETC = 0;
+		*BytePerPixelY = 1;
+		*BytePerPixelC = 0;
+	} else if (SourcePixelFormat == dml_rgbe_alpha) {
+		*BytePerPixelDETY = 4;
+		*BytePerPixelDETC = 1;
+		*BytePerPixelY = 4;
+		*BytePerPixelC = 1;
+	} else if (SourcePixelFormat == dml_420_8) {
+		*BytePerPixelDETY = 1;
+		*BytePerPixelDETC = 2;
+		*BytePerPixelY = 1;
+		*BytePerPixelC = 2;
+	} else if (SourcePixelFormat == dml_420_12) {
+		*BytePerPixelDETY = 2;
+		*BytePerPixelDETC = 4;
+		*BytePerPixelY = 2;
+		*BytePerPixelC = 4;
+	} else {
+		*BytePerPixelDETY = (dml_float_t) (4.0 / 3);
+		*BytePerPixelDETC = (dml_float_t) (8.0 / 3);
+		*BytePerPixelY = 2;
+		*BytePerPixelC = 4;
+	}
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: SourcePixelFormat = %u\n", __func__, SourcePixelFormat);
+	dml_print("DML::%s: BytePerPixelDETY = %f\n", __func__, *BytePerPixelDETY);
+	dml_print("DML::%s: BytePerPixelDETC = %f\n", __func__, *BytePerPixelDETC);
+	dml_print("DML::%s: BytePerPixelY = %u\n", __func__, *BytePerPixelY);
+	dml_print("DML::%s: BytePerPixelC = %u\n", __func__, *BytePerPixelC);
+#endif
+	if ((SourcePixelFormat == dml_444_64 || SourcePixelFormat == dml_444_32
+		|| SourcePixelFormat == dml_444_16
+		|| SourcePixelFormat == dml_444_8
+		|| SourcePixelFormat == dml_mono_16
+		|| SourcePixelFormat == dml_mono_8
+		|| SourcePixelFormat == dml_rgbe)) {
+		if (SurfaceTiling == dml_sw_linear) {
+			*BlockHeight256BytesY = 1;
+		} else if (SourcePixelFormat == dml_444_64) {
+			*BlockHeight256BytesY = 4;
+		} else if (SourcePixelFormat == dml_444_8) {
+			*BlockHeight256BytesY = 16;
+		} else {
+			*BlockHeight256BytesY = 8;
+		}
+		*BlockWidth256BytesY = 256U / *BytePerPixelY / *BlockHeight256BytesY;
+		*BlockHeight256BytesC = 0;
+		*BlockWidth256BytesC = 0;
+	} else {
+		if (SurfaceTiling == dml_sw_linear) {
+			*BlockHeight256BytesY = 1;
+			*BlockHeight256BytesC = 1;
+		} else if (SourcePixelFormat == dml_rgbe_alpha) {
+			*BlockHeight256BytesY = 8;
+			*BlockHeight256BytesC = 16;
+		} else if (SourcePixelFormat == dml_420_8) {
+			*BlockHeight256BytesY = 16;
+			*BlockHeight256BytesC = 8;
+		} else {
+			*BlockHeight256BytesY = 8;
+			*BlockHeight256BytesC = 8;
+		}
+		*BlockWidth256BytesY = 256U / *BytePerPixelY / *BlockHeight256BytesY;
+		*BlockWidth256BytesC = 256U / *BytePerPixelC / *BlockHeight256BytesC;
+	}
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: BlockWidth256BytesY = %u\n", __func__, *BlockWidth256BytesY);
+	dml_print("DML::%s: BlockHeight256BytesY = %u\n", __func__, *BlockHeight256BytesY);
+	dml_print("DML::%s: BlockWidth256BytesC = %u\n", __func__, *BlockWidth256BytesC);
+	dml_print("DML::%s: BlockHeight256BytesC = %u\n", __func__, *BlockHeight256BytesC);
+#endif
+
+	if (SurfaceTiling == dml_sw_linear) {
+		*MacroTileHeightY = *BlockHeight256BytesY;
+		*MacroTileWidthY = 256 / *BytePerPixelY / *MacroTileHeightY;
+		*MacroTileHeightC = *BlockHeight256BytesC;
+		if (*MacroTileHeightC == 0) {
+			*MacroTileWidthC = 0;
+		} else {
+			*MacroTileWidthC = 256 / *BytePerPixelC / *MacroTileHeightC;
+	}
+	} else if (SurfaceTiling == dml_sw_64kb_d || SurfaceTiling == dml_sw_64kb_d_t || SurfaceTiling == dml_sw_64kb_d_x || SurfaceTiling == dml_sw_64kb_r_x) {
+		*MacroTileHeightY = 16 * *BlockHeight256BytesY;
+		*MacroTileWidthY = 65536 / *BytePerPixelY / *MacroTileHeightY;
+		*MacroTileHeightC = 16 * *BlockHeight256BytesC;
+		if (*MacroTileHeightC == 0) {
+			*MacroTileWidthC = 0;
+		} else {
+			*MacroTileWidthC = 65536 / *BytePerPixelC / *MacroTileHeightC;
+		}
+	} else {
+		*MacroTileHeightY = 32 * *BlockHeight256BytesY;
+		*MacroTileWidthY = 65536 * 4 / *BytePerPixelY / *MacroTileHeightY;
+		*MacroTileHeightC = 32 * *BlockHeight256BytesC;
+		if (*MacroTileHeightC == 0) {
+			*MacroTileWidthC = 0;
+		} else {
+			*MacroTileWidthC = 65536 * 4 / *BytePerPixelC / *MacroTileHeightC;
+		}
+	}
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: MacroTileWidthY = %u\n", __func__, *MacroTileWidthY);
+	dml_print("DML::%s: MacroTileHeightY = %u\n", __func__, *MacroTileHeightY);
+	dml_print("DML::%s: MacroTileWidthC = %u\n", __func__, *MacroTileWidthC);
+	dml_print("DML::%s: MacroTileHeightC = %u\n", __func__, *MacroTileHeightC);
+#endif
+} // CalculateBytePerPixelAndBlockSizes
+
+static noinline_for_stack dml_float_t CalculateTWait(
+		dml_uint_t PrefetchMode,
+		enum dml_use_mall_for_pstate_change_mode UseMALLForPStateChange,
+		dml_bool_t SynchronizeDRRDisplaysForUCLKPStateChangeFinal,
+		dml_bool_t DRRDisplay,
+		dml_float_t DRAMClockChangeLatency,
+		dml_float_t FCLKChangeLatency,
+		dml_float_t UrgentLatency,
+		dml_float_t SREnterPlusExitTime)
+{
+	dml_float_t TWait = 0.0;
+
+	if (PrefetchMode == 0 &&
+			!(UseMALLForPStateChange == dml_use_mall_pstate_change_full_frame) && !(UseMALLForPStateChange == dml_use_mall_pstate_change_sub_viewport) &&
+			!(UseMALLForPStateChange == dml_use_mall_pstate_change_phantom_pipe) && !(SynchronizeDRRDisplaysForUCLKPStateChangeFinal && DRRDisplay)) {
+		TWait = dml_max3(DRAMClockChangeLatency + UrgentLatency, SREnterPlusExitTime, UrgentLatency);
+	} else if (PrefetchMode <= 1 && !(UseMALLForPStateChange == dml_use_mall_pstate_change_phantom_pipe)) {
+		TWait = dml_max3(FCLKChangeLatency + UrgentLatency, SREnterPlusExitTime, UrgentLatency);
+	} else if (PrefetchMode <= 2 && !(UseMALLForPStateChange == dml_use_mall_pstate_change_phantom_pipe)) {
+		TWait = dml_max(SREnterPlusExitTime, UrgentLatency);
+	} else {
+		TWait = UrgentLatency;
+	}
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: PrefetchMode = %u\n", __func__, PrefetchMode);
+	dml_print("DML::%s: TWait = %f\n", __func__, TWait);
+#endif
+	return TWait;
+} // CalculateTWait
+
+
+/// @brief Calculate the "starting point" for prefetch calculation
+///  if AllowForPStateChangeOrStutterInVBlank is set as a particular requirement, then the mode evalulation
+///  will only be done at the given mode. If no specific requirement (i.e. *_if_possible), then will just go from
+///  try all the prefetch mode in decreasing order of "difficulty" (start from 0 which means all power saving
+///  features).
+static void CalculatePrefetchMode(
+		enum dml_prefetch_modes AllowForPStateChangeOrStutterInVBlank,
+		dml_uint_t *MinPrefetchMode,
+		dml_uint_t *MaxPrefetchMode)
+{
+	if (AllowForPStateChangeOrStutterInVBlank == dml_prefetch_support_uclk_fclk_and_stutter_if_possible) {
+		*MinPrefetchMode = 0;   // consider all pwr saving features
+		*MaxPrefetchMode = 3;   // consider just urgent latency
+	} else {
+		if (AllowForPStateChangeOrStutterInVBlank == dml_prefetch_support_none) {
+			*MinPrefetchMode = 3;
+		} else if (AllowForPStateChangeOrStutterInVBlank == dml_prefetch_support_stutter) {
+			*MinPrefetchMode = 2;
+		} else if (AllowForPStateChangeOrStutterInVBlank == dml_prefetch_support_fclk_and_stutter) {
+			*MinPrefetchMode = 1;
+		} else if (AllowForPStateChangeOrStutterInVBlank == dml_prefetch_support_uclk_fclk_and_stutter) {
+			*MinPrefetchMode = 0;
+		} else {
+			dml_print("ERROR: Invalid AllowForPStateChangeOrStutterInVBlank setting! val=%u\n", AllowForPStateChangeOrStutterInVBlank);
+			ASSERT(0);
+		}
+		*MaxPrefetchMode = *MinPrefetchMode;
+	}
+} // CalculatePrefetchMode
+
+static dml_float_t CalculateWriteBackDISPCLK(
+		enum dml_source_format_class WritebackPixelFormat,
+		dml_float_t PixelClock,
+		dml_float_t WritebackHRatio,
+		dml_float_t WritebackVRatio,
+		dml_uint_t WritebackHTaps,
+		dml_uint_t WritebackVTaps,
+		dml_uint_t WritebackSourceWidth,
+		dml_uint_t WritebackDestinationWidth,
+		dml_uint_t HTotal,
+		dml_uint_t WritebackLineBufferSize,
+		dml_float_t DISPCLKDPPCLKVCOSpeed)
+{
+	dml_float_t DISPCLK_H, DISPCLK_V, DISPCLK_HB;
+
+	DISPCLK_H = PixelClock * dml_ceil(WritebackHTaps / 8.0, 1) / WritebackHRatio;
+	DISPCLK_V = PixelClock * (WritebackVTaps * dml_ceil(WritebackDestinationWidth / 6.0, 1) + 8.0) / (dml_float_t) HTotal;
+	DISPCLK_HB = PixelClock * WritebackVTaps * (WritebackDestinationWidth * WritebackVTaps - WritebackLineBufferSize / 57.0) / 6.0 / (dml_float_t) WritebackSourceWidth;
+	return RoundToDFSGranularity(dml_max3(DISPCLK_H, DISPCLK_V, DISPCLK_HB), 1, DISPCLKDPPCLKVCOSpeed);
+}
+
+static dml_float_t CalculateWriteBackDelay(
+		enum dml_source_format_class WritebackPixelFormat,
+		dml_float_t WritebackHRatio,
+		dml_float_t WritebackVRatio,
+		dml_uint_t WritebackVTaps,
+		dml_uint_t WritebackDestinationWidth,
+		dml_uint_t WritebackDestinationHeight,
+		dml_uint_t WritebackSourceHeight,
+		dml_uint_t HTotal)
+{
+	dml_float_t CalculateWriteBackDelay;
+	dml_float_t Line_length;
+	dml_float_t Output_lines_last_notclamped;
+	dml_float_t WritebackVInit;
+
+	WritebackVInit = (WritebackVRatio + WritebackVTaps + 1) / 2;
+	Line_length = dml_max((dml_float_t) WritebackDestinationWidth, dml_ceil((dml_float_t)WritebackDestinationWidth / 6.0, 1.0) * WritebackVTaps);
+	Output_lines_last_notclamped = WritebackDestinationHeight - 1 - dml_ceil(((dml_float_t)WritebackSourceHeight - (dml_float_t) WritebackVInit) / (dml_float_t)WritebackVRatio, 1.0);
+	if (Output_lines_last_notclamped < 0) {
+		CalculateWriteBackDelay = 0;
+	} else {
+		CalculateWriteBackDelay = Output_lines_last_notclamped * Line_length + (HTotal - WritebackDestinationWidth) + 80;
+	}
+	return CalculateWriteBackDelay;
+}
+
+static void CalculateVUpdateAndDynamicMetadataParameters(
+		dml_uint_t MaxInterDCNTileRepeaters,
+		dml_float_t Dppclk,
+		dml_float_t Dispclk,
+		dml_float_t DCFClkDeepSleep,
+		dml_float_t PixelClock,
+		dml_uint_t HTotal,
+		dml_uint_t VBlank,
+		dml_uint_t DynamicMetadataTransmittedBytes,
+		dml_uint_t DynamicMetadataLinesBeforeActiveRequired,
+		dml_uint_t InterlaceEnable,
+		dml_bool_t ProgressiveToInterlaceUnitInOPP,
+
+		// Output
+		dml_float_t *TSetup,
+		dml_float_t *Tdmbf,
+		dml_float_t *Tdmec,
+		dml_float_t *Tdmsks,
+		dml_uint_t *VUpdateOffsetPix,
+		dml_uint_t *VUpdateWidthPix,
+		dml_uint_t *VReadyOffsetPix)
+{
+	dml_float_t TotalRepeaterDelayTime;
+	TotalRepeaterDelayTime = MaxInterDCNTileRepeaters * (2 / Dppclk + 3 / Dispclk);
+	*VUpdateWidthPix = (dml_uint_t)(dml_ceil((14.0 / DCFClkDeepSleep + 12.0 / Dppclk + TotalRepeaterDelayTime) * PixelClock, 1.0));
+	*VReadyOffsetPix = (dml_uint_t)(dml_ceil(dml_max(150.0 / Dppclk, TotalRepeaterDelayTime + 20.0 / DCFClkDeepSleep + 10.0 / Dppclk) * PixelClock, 1.0));
+	*VUpdateOffsetPix = (dml_uint_t)(dml_ceil(HTotal / 4.0, 1.0));
+	*TSetup = (*VUpdateOffsetPix + *VUpdateWidthPix + *VReadyOffsetPix) / PixelClock;
+	*Tdmbf = DynamicMetadataTransmittedBytes / 4.0 / Dispclk;
+	*Tdmec = HTotal / PixelClock;
+
+	if (DynamicMetadataLinesBeforeActiveRequired == 0) {
+		*Tdmsks = VBlank * HTotal / PixelClock / 2.0;
+	} else {
+		*Tdmsks = DynamicMetadataLinesBeforeActiveRequired * HTotal / PixelClock;
+	}
+	if (InterlaceEnable == 1 && ProgressiveToInterlaceUnitInOPP == false) {
+		*Tdmsks = *Tdmsks / 2;
+	}
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: DynamicMetadataLinesBeforeActiveRequired = %u\n", __func__, DynamicMetadataLinesBeforeActiveRequired);
+	dml_print("DML::%s: VBlank = %u\n", __func__, VBlank);
+	dml_print("DML::%s: HTotal = %u\n", __func__, HTotal);
+	dml_print("DML::%s: PixelClock = %f\n", __func__, PixelClock);
+	dml_print("DML::%s: Dppclk = %f\n", __func__, Dppclk);
+	dml_print("DML::%s: DCFClkDeepSleep = %f\n", __func__, DCFClkDeepSleep);
+	dml_print("DML::%s: MaxInterDCNTileRepeaters = %u\n", __func__, MaxInterDCNTileRepeaters);
+	dml_print("DML::%s: TotalRepeaterDelayTime = %f\n", __func__, TotalRepeaterDelayTime);
+
+	dml_print("DML::%s: VUpdateWidthPix = %u\n", __func__, *VUpdateWidthPix);
+	dml_print("DML::%s: VReadyOffsetPix = %u\n", __func__, *VReadyOffsetPix);
+	dml_print("DML::%s: VUpdateOffsetPix = %u\n", __func__, *VUpdateOffsetPix);
+
+	dml_print("DML::%s: Tdmsks = %f\n", __func__, *Tdmsks);
+#endif
+}
+
+static void CalculateRowBandwidth(
+		dml_bool_t GPUVMEnable,
+		enum dml_source_format_class SourcePixelFormat,
+		dml_float_t VRatio,
+		dml_float_t VRatioChroma,
+		dml_bool_t DCCEnable,
+		dml_float_t LineTime,
+		dml_uint_t MetaRowByteLuma,
+		dml_uint_t MetaRowByteChroma,
+		dml_uint_t meta_row_height_luma,
+		dml_uint_t meta_row_height_chroma,
+		dml_uint_t PixelPTEBytesPerRowLuma,
+		dml_uint_t PixelPTEBytesPerRowChroma,
+		dml_uint_t dpte_row_height_luma,
+		dml_uint_t dpte_row_height_chroma,
+		// Output
+		dml_float_t *meta_row_bw,
+		dml_float_t *dpte_row_bw)
+{
+	if (DCCEnable != true) {
+		*meta_row_bw = 0;
+	} else if (SourcePixelFormat == dml_420_8 || SourcePixelFormat == dml_420_10 || SourcePixelFormat == dml_420_12 || SourcePixelFormat == dml_rgbe_alpha) {
+		*meta_row_bw = VRatio * MetaRowByteLuma / (meta_row_height_luma * LineTime)
+				+ VRatioChroma * MetaRowByteChroma
+					/ (meta_row_height_chroma * LineTime);
+	} else {
+		*meta_row_bw = VRatio * MetaRowByteLuma / (meta_row_height_luma * LineTime);
+	}
+
+	if (GPUVMEnable != true) {
+		*dpte_row_bw = 0;
+	} else if (SourcePixelFormat == dml_420_8 || SourcePixelFormat == dml_420_10 || SourcePixelFormat == dml_420_12 || SourcePixelFormat == dml_rgbe_alpha) {
+		*dpte_row_bw = VRatio * PixelPTEBytesPerRowLuma / (dpte_row_height_luma * LineTime)
+				+ VRatioChroma * PixelPTEBytesPerRowChroma
+					/ (dpte_row_height_chroma * LineTime);
+	} else {
+		*dpte_row_bw = VRatio * PixelPTEBytesPerRowLuma / (dpte_row_height_luma * LineTime);
+	}
+}
+
+/// @brief Determine immediate flip schedule given bw remaining after considering the prefetch schedule
+/// @param BandwidthAvailableForImmediateFlip Bandwidth available for iflip for all planes
+static void CalculateFlipSchedule(
+		dml_float_t HostVMInefficiencyFactor,
+		dml_float_t UrgentExtraLatency,
+		dml_float_t UrgentLatency,
+		dml_uint_t GPUVMMaxPageTableLevels,
+		dml_bool_t HostVMEnable,
+		dml_uint_t HostVMMaxNonCachedPageTableLevels,
+		dml_bool_t GPUVMEnable,
+		dml_uint_t HostVMMinPageSize,
+		dml_float_t PDEAndMetaPTEBytesPerFrame,
+		dml_float_t MetaRowBytes,
+		dml_float_t DPTEBytesPerRow,
+		dml_float_t BandwidthAvailableForImmediateFlip,
+		dml_uint_t TotImmediateFlipBytes,
+		enum dml_source_format_class SourcePixelFormat,
+		dml_float_t LineTime,
+		dml_float_t VRatio,
+		dml_float_t VRatioChroma,
+		dml_float_t Tno_bw,
+		dml_bool_t DCCEnable,
+		dml_uint_t dpte_row_height,
+		dml_uint_t meta_row_height,
+		dml_uint_t dpte_row_height_chroma,
+		dml_uint_t meta_row_height_chroma,
+		dml_bool_t use_one_row_for_frame_flip,
+
+		// Output
+		dml_float_t *DestinationLinesToRequestVMInImmediateFlip,
+		dml_float_t *DestinationLinesToRequestRowInImmediateFlip,
+		dml_float_t *final_flip_bw,
+		dml_bool_t *ImmediateFlipSupportedForPipe)
+{
+	dml_float_t min_row_time = 0.0;
+	dml_uint_t HostVMDynamicLevelsTrips = 0;
+	dml_float_t TimeForFetchingMetaPTEImmediateFlip = 0;
+	dml_float_t TimeForFetchingRowInVBlankImmediateFlip = 0;
+	dml_float_t ImmediateFlipBW = 0; // @brief The immediate flip bandwidth for this pipe
+
+	if (GPUVMEnable == true && HostVMEnable == true) {
+		HostVMDynamicLevelsTrips = HostVMMaxNonCachedPageTableLevels;
+	} else {
+		HostVMDynamicLevelsTrips = 0;
+	}
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: TotImmediateFlipBytes = %u\n", __func__, TotImmediateFlipBytes);
+	dml_print("DML::%s: HostVMInefficiencyFactor = %f\n", __func__, HostVMInefficiencyFactor);
+	dml_print("DML::%s: UrgentLatency = %f\n", __func__, UrgentLatency);
+	dml_print("DML::%s: BandwidthAvailableForImmediateFlip = %f\n", __func__, BandwidthAvailableForImmediateFlip);
+#endif
+
+	if (TotImmediateFlipBytes > 0) {
+		if (use_one_row_for_frame_flip) {
+			ImmediateFlipBW = (PDEAndMetaPTEBytesPerFrame + MetaRowBytes + 2.0 * DPTEBytesPerRow) * BandwidthAvailableForImmediateFlip / (dml_float_t) TotImmediateFlipBytes;
+		} else {
+			ImmediateFlipBW = (PDEAndMetaPTEBytesPerFrame + MetaRowBytes + DPTEBytesPerRow) * BandwidthAvailableForImmediateFlip / (dml_float_t) TotImmediateFlipBytes;
+		}
+		if (GPUVMEnable == true) {
+			TimeForFetchingMetaPTEImmediateFlip = dml_max3(Tno_bw + PDEAndMetaPTEBytesPerFrame * HostVMInefficiencyFactor / ImmediateFlipBW,
+														UrgentExtraLatency + UrgentLatency * (GPUVMMaxPageTableLevels * (HostVMDynamicLevelsTrips + 1) - 1),
+														LineTime / 4.0);
+		} else {
+			TimeForFetchingMetaPTEImmediateFlip = 0;
+		}
+		if ((GPUVMEnable == true || DCCEnable == true)) {
+			TimeForFetchingRowInVBlankImmediateFlip = dml_max3((MetaRowBytes + DPTEBytesPerRow * HostVMInefficiencyFactor) / ImmediateFlipBW, UrgentLatency * (HostVMDynamicLevelsTrips + 1), LineTime / 4.0);
+		} else {
+			TimeForFetchingRowInVBlankImmediateFlip = 0;
+		}
+
+		*DestinationLinesToRequestVMInImmediateFlip = dml_ceil(4.0 * (TimeForFetchingMetaPTEImmediateFlip / LineTime), 1.0) / 4.0;
+		*DestinationLinesToRequestRowInImmediateFlip = dml_ceil(4.0 * (TimeForFetchingRowInVBlankImmediateFlip / LineTime), 1.0) / 4.0;
+
+		if (GPUVMEnable == true) {
+			*final_flip_bw = dml_max(PDEAndMetaPTEBytesPerFrame * HostVMInefficiencyFactor / (*DestinationLinesToRequestVMInImmediateFlip * LineTime),
+								(MetaRowBytes + DPTEBytesPerRow * HostVMInefficiencyFactor) / (*DestinationLinesToRequestRowInImmediateFlip * LineTime));
+		} else if ((GPUVMEnable == true || DCCEnable == true)) {
+			*final_flip_bw = (MetaRowBytes + DPTEBytesPerRow * HostVMInefficiencyFactor) / (*DestinationLinesToRequestRowInImmediateFlip * LineTime);
+		} else {
+			*final_flip_bw = 0;
+		}
+	} else {
+		TimeForFetchingMetaPTEImmediateFlip = 0;
+		TimeForFetchingRowInVBlankImmediateFlip = 0;
+		*DestinationLinesToRequestVMInImmediateFlip = 0;
+		*DestinationLinesToRequestRowInImmediateFlip = 0;
+		*final_flip_bw = 0;
+	}
+
+	if (SourcePixelFormat == dml_420_8 || SourcePixelFormat == dml_420_10 || SourcePixelFormat == dml_rgbe_alpha) {
+		if (GPUVMEnable == true && DCCEnable != true) {
+			min_row_time = dml_min(dpte_row_height * LineTime / VRatio, dpte_row_height_chroma * LineTime / VRatioChroma);
+		} else if (GPUVMEnable != true && DCCEnable == true) {
+			min_row_time = dml_min(meta_row_height * LineTime / VRatio, meta_row_height_chroma * LineTime / VRatioChroma);
+		} else {
+			min_row_time = dml_min4(dpte_row_height * LineTime / VRatio, meta_row_height * LineTime / VRatio, dpte_row_height_chroma * LineTime / VRatioChroma, meta_row_height_chroma * LineTime / VRatioChroma);
+		}
+	} else {
+		if (GPUVMEnable == true && DCCEnable != true) {
+			min_row_time = dpte_row_height * LineTime / VRatio;
+		} else if (GPUVMEnable != true && DCCEnable == true) {
+			min_row_time = meta_row_height * LineTime / VRatio;
+		} else {
+			min_row_time = dml_min(dpte_row_height * LineTime / VRatio, meta_row_height * LineTime / VRatio);
+		}
+	}
+
+	if (*DestinationLinesToRequestVMInImmediateFlip >= 32 || *DestinationLinesToRequestRowInImmediateFlip >= 16 || TimeForFetchingMetaPTEImmediateFlip + 2 * TimeForFetchingRowInVBlankImmediateFlip > min_row_time) {
+		*ImmediateFlipSupportedForPipe = false;
+	} else {
+		*ImmediateFlipSupportedForPipe = true;
+	}
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: GPUVMEnable = %u\n", __func__, GPUVMEnable);
+	dml_print("DML::%s: DCCEnable = %u\n", __func__, DCCEnable);
+
+	dml_print("DML::%s: MetaRowBytes = %f\n", __func__, MetaRowBytes);
+	dml_print("DML::%s: DPTEBytesPerRow = %f\n", __func__, DPTEBytesPerRow);
+	dml_print("DML::%s: BandwidthAvailableForImmediateFlip = %f\n", __func__, BandwidthAvailableForImmediateFlip);
+	dml_print("DML::%s: TotImmediateFlipBytes = %u\n", __func__, TotImmediateFlipBytes);
+	dml_print("DML::%s: ImmediateFlipBW = %f\n", __func__, ImmediateFlipBW);
+	dml_print("DML::%s: PDEAndMetaPTEBytesPerFrame = %f\n", __func__, PDEAndMetaPTEBytesPerFrame);
+	dml_print("DML::%s: HostVMInefficiencyFactor = %f\n", __func__, HostVMInefficiencyFactor);
+	dml_print("DML::%s: LineTime = %f\n", __func__, LineTime);
+	dml_print("DML::%s: final_flip_bw = %f\n", __func__, *final_flip_bw);
+
+	dml_print("DML::%s: DestinationLinesToRequestVMInImmediateFlip = %f\n", __func__, *DestinationLinesToRequestVMInImmediateFlip);
+	dml_print("DML::%s: DestinationLinesToRequestRowInImmediateFlip = %f\n", __func__, *DestinationLinesToRequestRowInImmediateFlip);
+	dml_print("DML::%s: TimeForFetchingMetaPTEImmediateFlip = %f\n", __func__, TimeForFetchingMetaPTEImmediateFlip);
+	dml_print("DML::%s: TimeForFetchingRowInVBlankImmediateFlip = %f\n", __func__, TimeForFetchingRowInVBlankImmediateFlip);
+	dml_print("DML::%s: min_row_time = %f\n", __func__, min_row_time);
+	dml_print("DML::%s: ImmediateFlipSupportedForPipe = %u\n", __func__, *ImmediateFlipSupportedForPipe);
+#endif
+} // CalculateFlipSchedule
+
+static dml_float_t RoundToDFSGranularity(dml_float_t Clock, dml_bool_t round_up, dml_float_t VCOSpeed)
+{
+	if (Clock <= 0.0)
+		return 0.0;
+	else {
+		if (round_up)
+			return VCOSpeed * 4.0 / dml_floor(VCOSpeed * 4.0 / Clock, 1.0);
+		else
+			return VCOSpeed * 4.0 / dml_ceil(VCOSpeed * 4.0 / Clock, 1.0);
+	}
+}
+
+static void CalculateDCCConfiguration(
+		dml_bool_t DCCEnabled,
+		dml_bool_t DCCProgrammingAssumesScanDirectionUnknown,
+		enum dml_source_format_class SourcePixelFormat,
+		dml_uint_t SurfaceWidthLuma,
+		dml_uint_t SurfaceWidthChroma,
+		dml_uint_t SurfaceHeightLuma,
+		dml_uint_t SurfaceHeightChroma,
+		dml_uint_t nomDETInKByte,
+		dml_uint_t RequestHeight256ByteLuma,
+		dml_uint_t RequestHeight256ByteChroma,
+		enum dml_swizzle_mode TilingFormat,
+		dml_uint_t BytePerPixelY,
+		dml_uint_t BytePerPixelC,
+		dml_float_t BytePerPixelDETY,
+		dml_float_t BytePerPixelDETC,
+		enum dml_rotation_angle SourceScan,
+		// Output
+		dml_uint_t *MaxUncompressedBlockLuma,
+		dml_uint_t *MaxUncompressedBlockChroma,
+		dml_uint_t *MaxCompressedBlockLuma,
+		dml_uint_t *MaxCompressedBlockChroma,
+		dml_uint_t *IndependentBlockLuma,
+		dml_uint_t *IndependentBlockChroma)
+{
+	dml_uint_t DETBufferSizeForDCC = nomDETInKByte * 1024;
+
+	dml_uint_t yuv420;
+	dml_uint_t horz_div_l;
+	dml_uint_t horz_div_c;
+	dml_uint_t vert_div_l;
+	dml_uint_t vert_div_c;
+
+	dml_uint_t swath_buf_size;
+	dml_float_t detile_buf_vp_horz_limit;
+	dml_float_t detile_buf_vp_vert_limit;
+
+	dml_uint_t MAS_vp_horz_limit;
+	dml_uint_t MAS_vp_vert_limit;
+	dml_uint_t max_vp_horz_width;
+	dml_uint_t max_vp_vert_height;
+	dml_uint_t eff_surf_width_l;
+	dml_uint_t eff_surf_width_c;
+	dml_uint_t eff_surf_height_l;
+	dml_uint_t eff_surf_height_c;
+
+	dml_uint_t full_swath_bytes_horz_wc_l;
+	dml_uint_t full_swath_bytes_horz_wc_c;
+	dml_uint_t full_swath_bytes_vert_wc_l;
+	dml_uint_t full_swath_bytes_vert_wc_c;
+
+	dml_uint_t req128_horz_wc_l;
+	dml_uint_t req128_horz_wc_c;
+	dml_uint_t req128_vert_wc_l;
+	dml_uint_t req128_vert_wc_c;
+
+	dml_uint_t   segment_order_horz_contiguous_luma;
+	dml_uint_t   segment_order_horz_contiguous_chroma;
+	dml_uint_t   segment_order_vert_contiguous_luma;
+	dml_uint_t   segment_order_vert_contiguous_chroma;
+
+	typedef enum{
+		REQ_256Bytes,
+		REQ_128BytesNonContiguous,
+		REQ_128BytesContiguous,
+		REQ_NA
+	} RequestType;
+
+	RequestType   RequestLuma;
+	RequestType   RequestChroma;
+
+	yuv420 = ((SourcePixelFormat == dml_420_8 || SourcePixelFormat == dml_420_10 || SourcePixelFormat == dml_420_12) ? 1 : 0);
+	horz_div_l = 1;
+	horz_div_c = 1;
+	vert_div_l = 1;
+	vert_div_c = 1;
+
+	if (BytePerPixelY == 1)
+		vert_div_l = 0;
+	if (BytePerPixelC == 1)
+		vert_div_c = 0;
+
+	if (BytePerPixelC == 0) {
+		swath_buf_size = DETBufferSizeForDCC / 2 - 2 * 256;
+		detile_buf_vp_horz_limit = (dml_float_t) swath_buf_size / ((dml_float_t) RequestHeight256ByteLuma * BytePerPixelY / (1 + horz_div_l));
+		detile_buf_vp_vert_limit = (dml_float_t) swath_buf_size / (256.0 / RequestHeight256ByteLuma / (1 + vert_div_l));
+	} else {
+		swath_buf_size = DETBufferSizeForDCC / 2 - 2 * 2 * 256;
+		detile_buf_vp_horz_limit = (dml_float_t) swath_buf_size / ((dml_float_t) RequestHeight256ByteLuma * BytePerPixelY / (1 + horz_div_l) + (dml_float_t) RequestHeight256ByteChroma * BytePerPixelC / (1 + horz_div_c) / (1 + yuv420));
+		detile_buf_vp_vert_limit = (dml_float_t) swath_buf_size / (256.0 / RequestHeight256ByteLuma / (1 + vert_div_l) + 256.0 / RequestHeight256ByteChroma / (1 + vert_div_c) / (1 + yuv420));
+	}
+
+	if (SourcePixelFormat == dml_420_10) {
+		detile_buf_vp_horz_limit = 1.5 * detile_buf_vp_horz_limit;
+		detile_buf_vp_vert_limit = 1.5 * detile_buf_vp_vert_limit;
+	}
+
+	detile_buf_vp_horz_limit = dml_floor(detile_buf_vp_horz_limit - 1, 16);
+	detile_buf_vp_vert_limit = dml_floor(detile_buf_vp_vert_limit - 1, 16);
+
+	MAS_vp_horz_limit = SourcePixelFormat == dml_rgbe_alpha ? 3840 : 6144;
+	MAS_vp_vert_limit = SourcePixelFormat == dml_rgbe_alpha ? 3840 : (BytePerPixelY == 8 ? 3072 : 6144);
+	max_vp_horz_width = (dml_uint_t)(dml_min((dml_float_t) MAS_vp_horz_limit, detile_buf_vp_horz_limit));
+	max_vp_vert_height = (dml_uint_t)(dml_min((dml_float_t) MAS_vp_vert_limit, detile_buf_vp_vert_limit));
+	eff_surf_width_l = (SurfaceWidthLuma > max_vp_horz_width ? max_vp_horz_width : SurfaceWidthLuma);
+	eff_surf_width_c = eff_surf_width_l / (1 + yuv420);
+	eff_surf_height_l = (SurfaceHeightLuma > max_vp_vert_height ? max_vp_vert_height : SurfaceHeightLuma);
+	eff_surf_height_c = eff_surf_height_l / (1 + yuv420);
+
+	full_swath_bytes_horz_wc_l = eff_surf_width_l * RequestHeight256ByteLuma * BytePerPixelY;
+	full_swath_bytes_vert_wc_l = eff_surf_height_l * 256 / RequestHeight256ByteLuma;
+	if (BytePerPixelC > 0) {
+		full_swath_bytes_horz_wc_c = eff_surf_width_c * RequestHeight256ByteChroma * BytePerPixelC;
+		full_swath_bytes_vert_wc_c = eff_surf_height_c * 256 / RequestHeight256ByteChroma;
+	} else {
+		full_swath_bytes_horz_wc_c = 0;
+		full_swath_bytes_vert_wc_c = 0;
+	}
+
+	if (SourcePixelFormat == dml_420_10) {
+		full_swath_bytes_horz_wc_l = (dml_uint_t)(dml_ceil((dml_float_t) full_swath_bytes_horz_wc_l * 2.0 / 3.0, 256.0));
+		full_swath_bytes_horz_wc_c = (dml_uint_t)(dml_ceil((dml_float_t) full_swath_bytes_horz_wc_c * 2.0 / 3.0, 256.0));
+		full_swath_bytes_vert_wc_l = (dml_uint_t)(dml_ceil((dml_float_t) full_swath_bytes_vert_wc_l * 2.0 / 3.0, 256.0));
+		full_swath_bytes_vert_wc_c = (dml_uint_t)(dml_ceil((dml_float_t) full_swath_bytes_vert_wc_c * 2.0 / 3.0, 256.0));
+	}
+
+	if (2 * full_swath_bytes_horz_wc_l + 2 * full_swath_bytes_horz_wc_c <= DETBufferSizeForDCC) {
+		req128_horz_wc_l = 0;
+		req128_horz_wc_c = 0;
+	} else if (full_swath_bytes_horz_wc_l < 1.5 * full_swath_bytes_horz_wc_c && 2 * full_swath_bytes_horz_wc_l + full_swath_bytes_horz_wc_c <= DETBufferSizeForDCC) {
+		req128_horz_wc_l = 0;
+		req128_horz_wc_c = 1;
+	} else if (full_swath_bytes_horz_wc_l >= 1.5 * full_swath_bytes_horz_wc_c && full_swath_bytes_horz_wc_l + 2 * full_swath_bytes_horz_wc_c <= DETBufferSizeForDCC) {
+		req128_horz_wc_l = 1;
+		req128_horz_wc_c = 0;
+	} else {
+		req128_horz_wc_l = 1;
+		req128_horz_wc_c = 1;
+	}
+
+	if (2 * full_swath_bytes_vert_wc_l + 2 * full_swath_bytes_vert_wc_c <= DETBufferSizeForDCC) {
+		req128_vert_wc_l = 0;
+		req128_vert_wc_c = 0;
+	} else if (full_swath_bytes_vert_wc_l < 1.5 * full_swath_bytes_vert_wc_c && 2 * full_swath_bytes_vert_wc_l + full_swath_bytes_vert_wc_c <= DETBufferSizeForDCC) {
+		req128_vert_wc_l = 0;
+		req128_vert_wc_c = 1;
+	} else if (full_swath_bytes_vert_wc_l >= 1.5 * full_swath_bytes_vert_wc_c && full_swath_bytes_vert_wc_l + 2 * full_swath_bytes_vert_wc_c <= DETBufferSizeForDCC) {
+		req128_vert_wc_l = 1;
+		req128_vert_wc_c = 0;
+	} else {
+		req128_vert_wc_l = 1;
+		req128_vert_wc_c = 1;
+	}
+
+	if (BytePerPixelY == 2) {
+		segment_order_horz_contiguous_luma = 0;
+		segment_order_vert_contiguous_luma = 1;
+	} else {
+		segment_order_horz_contiguous_luma = 1;
+		segment_order_vert_contiguous_luma = 0;
+	}
+
+	if (BytePerPixelC == 2) {
+		segment_order_horz_contiguous_chroma = 0;
+		segment_order_vert_contiguous_chroma = 1;
+	} else {
+		segment_order_horz_contiguous_chroma = 1;
+		segment_order_vert_contiguous_chroma = 0;
+	}
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: DCCEnabled = %u\n", __func__, DCCEnabled);
+	dml_print("DML::%s: nomDETInKByte = %u\n", __func__, nomDETInKByte);
+	dml_print("DML::%s: DETBufferSizeForDCC = %u\n", __func__, DETBufferSizeForDCC);
+	dml_print("DML::%s: req128_horz_wc_l = %u\n", __func__, req128_horz_wc_l);
+	dml_print("DML::%s: req128_horz_wc_c = %u\n", __func__, req128_horz_wc_c);
+	dml_print("DML::%s: full_swath_bytes_horz_wc_l = %u\n", __func__, full_swath_bytes_horz_wc_l);
+	dml_print("DML::%s: full_swath_bytes_vert_wc_c = %u\n", __func__, full_swath_bytes_vert_wc_c);
+	dml_print("DML::%s: segment_order_horz_contiguous_luma = %u\n", __func__, segment_order_horz_contiguous_luma);
+	dml_print("DML::%s: segment_order_horz_contiguous_chroma = %u\n", __func__, segment_order_horz_contiguous_chroma);
+#endif
+
+	if (DCCProgrammingAssumesScanDirectionUnknown == true) {
+		if (req128_horz_wc_l == 0 && req128_vert_wc_l == 0) {
+			RequestLuma = REQ_256Bytes;
+		} else if ((req128_horz_wc_l == 1 && segment_order_horz_contiguous_luma == 0) || (req128_vert_wc_l == 1 && segment_order_vert_contiguous_luma == 0)) {
+			RequestLuma = REQ_128BytesNonContiguous;
+		} else {
+			RequestLuma = REQ_128BytesContiguous;
+		}
+		if (req128_horz_wc_c == 0 && req128_vert_wc_c == 0) {
+			RequestChroma = REQ_256Bytes;
+		} else if ((req128_horz_wc_c == 1 && segment_order_horz_contiguous_chroma == 0) || (req128_vert_wc_c == 1 && segment_order_vert_contiguous_chroma == 0)) {
+			RequestChroma = REQ_128BytesNonContiguous;
+		} else {
+			RequestChroma = REQ_128BytesContiguous;
+		}
+	} else if (!dml_is_vertical_rotation(SourceScan)) {
+		if (req128_horz_wc_l == 0) {
+			RequestLuma = REQ_256Bytes;
+		} else if (segment_order_horz_contiguous_luma == 0) {
+			RequestLuma = REQ_128BytesNonContiguous;
+		} else {
+			RequestLuma = REQ_128BytesContiguous;
+		}
+		if (req128_horz_wc_c == 0) {
+			RequestChroma = REQ_256Bytes;
+		} else if (segment_order_horz_contiguous_chroma == 0) {
+			RequestChroma = REQ_128BytesNonContiguous;
+		} else {
+			RequestChroma = REQ_128BytesContiguous;
+		}
+	} else {
+		if (req128_vert_wc_l == 0) {
+			RequestLuma = REQ_256Bytes;
+		} else if (segment_order_vert_contiguous_luma == 0) {
+			RequestLuma = REQ_128BytesNonContiguous;
+		} else {
+			RequestLuma = REQ_128BytesContiguous;
+		}
+		if (req128_vert_wc_c == 0) {
+			RequestChroma = REQ_256Bytes;
+		} else if (segment_order_vert_contiguous_chroma == 0) {
+			RequestChroma = REQ_128BytesNonContiguous;
+		} else {
+			RequestChroma = REQ_128BytesContiguous;
+		}
+	}
+
+	if (RequestLuma == REQ_256Bytes) {
+		*MaxUncompressedBlockLuma = 256;
+		*MaxCompressedBlockLuma = 256;
+		*IndependentBlockLuma = 0;
+	} else if (RequestLuma == REQ_128BytesContiguous) {
+		*MaxUncompressedBlockLuma = 256;
+		*MaxCompressedBlockLuma = 128;
+		*IndependentBlockLuma = 128;
+	} else {
+		*MaxUncompressedBlockLuma = 256;
+		*MaxCompressedBlockLuma = 64;
+		*IndependentBlockLuma = 64;
+	}
+
+	if (RequestChroma == REQ_256Bytes) {
+		*MaxUncompressedBlockChroma = 256;
+		*MaxCompressedBlockChroma = 256;
+		*IndependentBlockChroma = 0;
+	} else if (RequestChroma == REQ_128BytesContiguous) {
+		*MaxUncompressedBlockChroma = 256;
+		*MaxCompressedBlockChroma = 128;
+		*IndependentBlockChroma = 128;
+	} else {
+		*MaxUncompressedBlockChroma = 256;
+		*MaxCompressedBlockChroma = 64;
+		*IndependentBlockChroma = 64;
+	}
+
+	if (DCCEnabled != true || BytePerPixelC == 0) {
+		*MaxUncompressedBlockChroma = 0;
+		*MaxCompressedBlockChroma = 0;
+		*IndependentBlockChroma = 0;
+	}
+
+	if (DCCEnabled != true) {
+		*MaxUncompressedBlockLuma = 0;
+		*MaxCompressedBlockLuma = 0;
+		*IndependentBlockLuma = 0;
+	}
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: MaxUncompressedBlockLuma = %u\n", __func__, *MaxUncompressedBlockLuma);
+	dml_print("DML::%s: MaxCompressedBlockLuma = %u\n", __func__, *MaxCompressedBlockLuma);
+	dml_print("DML::%s: IndependentBlockLuma = %u\n", __func__, *IndependentBlockLuma);
+	dml_print("DML::%s: MaxUncompressedBlockChroma = %u\n", __func__, *MaxUncompressedBlockChroma);
+	dml_print("DML::%s: MaxCompressedBlockChroma = %u\n", __func__, *MaxCompressedBlockChroma);
+	dml_print("DML::%s: IndependentBlockChroma = %u\n", __func__, *IndependentBlockChroma);
+#endif
+
+} // CalculateDCCConfiguration
+
+static dml_uint_t CalculatePrefetchSourceLines(
+		dml_float_t VRatio,
+		dml_uint_t VTaps,
+		dml_bool_t Interlace,
+		dml_bool_t ProgressiveToInterlaceUnitInOPP,
+		dml_uint_t SwathHeight,
+		enum dml_rotation_angle SourceScan,
+		dml_bool_t ViewportStationary,
+		dml_uint_t SwathWidth,
+		dml_uint_t ViewportHeight,
+		dml_uint_t ViewportXStart,
+		dml_uint_t ViewportYStart,
+
+		// Output
+		dml_uint_t *VInitPreFill,
+		dml_uint_t *MaxNumSwath)
+{
+
+	dml_uint_t vp_start_rot = 0;
+	dml_uint_t sw0_tmp = 0;
+	dml_uint_t MaxPartialSwath = 0;
+	dml_float_t numLines = 0;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: VRatio = %f\n", __func__, VRatio);
+	dml_print("DML::%s: VTaps = %u\n", __func__, VTaps);
+	dml_print("DML::%s: ViewportXStart = %u\n", __func__, ViewportXStart);
+	dml_print("DML::%s: ViewportYStart = %u\n", __func__, ViewportYStart);
+	dml_print("DML::%s: ViewportStationary = %u\n", __func__, ViewportStationary);
+	dml_print("DML::%s: SwathHeight = %u\n", __func__, SwathHeight);
+#endif
+	if (ProgressiveToInterlaceUnitInOPP)
+		*VInitPreFill = (dml_uint_t)(dml_floor((VRatio + (dml_float_t) VTaps + 1) / 2.0, 1));
+	else
+		*VInitPreFill = (dml_uint_t)(dml_floor((VRatio + (dml_float_t) VTaps + 1 + Interlace * 0.5 * VRatio) / 2.0, 1));
+
+	if (ViewportStationary) {
+		if (SourceScan == dml_rotation_180 || SourceScan == dml_rotation_180m) {
+			vp_start_rot = SwathHeight - (((dml_uint_t) (ViewportYStart + ViewportHeight - 1) % SwathHeight) + 1);
+		} else if (SourceScan == dml_rotation_270 || SourceScan == dml_rotation_90m) {
+			vp_start_rot = ViewportXStart;
+		} else if (SourceScan == dml_rotation_90 || SourceScan == dml_rotation_270m) {
+			vp_start_rot = SwathHeight - (((dml_uint_t)(ViewportYStart + SwathWidth - 1) % SwathHeight) + 1);
+		} else {
+			vp_start_rot = ViewportYStart;
+		}
+		sw0_tmp = SwathHeight - (vp_start_rot % SwathHeight);
+		if (sw0_tmp < *VInitPreFill) {
+			*MaxNumSwath = (dml_uint_t)(dml_ceil((*VInitPreFill - sw0_tmp) / (dml_float_t) SwathHeight, 1) + 1);
+		} else {
+			*MaxNumSwath = 1;
+		}
+		MaxPartialSwath = (dml_uint_t)(dml_max(1, (dml_uint_t) (vp_start_rot + *VInitPreFill - 1) % SwathHeight));
+	} else {
+		*MaxNumSwath = (dml_uint_t)(dml_ceil((*VInitPreFill - 1.0) / (dml_float_t) SwathHeight, 1) + 1);
+		if (*VInitPreFill > 1) {
+			MaxPartialSwath = (dml_uint_t)(dml_max(1, (dml_uint_t) (*VInitPreFill - 2) % SwathHeight));
+		} else {
+			MaxPartialSwath = (dml_uint_t)(dml_max(1, (dml_uint_t) (*VInitPreFill + SwathHeight - 2) % SwathHeight));
+		}
+	}
+	numLines = *MaxNumSwath * SwathHeight + MaxPartialSwath;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: vp_start_rot = %u\n", __func__, vp_start_rot);
+	dml_print("DML::%s: VInitPreFill = %u\n", __func__, *VInitPreFill);
+	dml_print("DML::%s: MaxPartialSwath = %u\n", __func__, MaxPartialSwath);
+	dml_print("DML::%s: MaxNumSwath = %u\n", __func__, *MaxNumSwath);
+	dml_print("DML::%s: Prefetch source lines = %3.2f\n", __func__, numLines);
+#endif
+	return (dml_uint_t)(numLines);
+
+} // CalculatePrefetchSourceLines
+
+static dml_uint_t CalculateVMAndRowBytes(
+		dml_bool_t ViewportStationary,
+		dml_bool_t DCCEnable,
+		dml_uint_t NumberOfDPPs,
+		dml_uint_t BlockHeight256Bytes,
+		dml_uint_t BlockWidth256Bytes,
+		enum dml_source_format_class SourcePixelFormat,
+		dml_uint_t SurfaceTiling,
+		dml_uint_t BytePerPixel,
+		enum dml_rotation_angle SourceScan,
+		dml_uint_t SwathWidth,
+		dml_uint_t ViewportHeight,
+		dml_uint_t ViewportXStart,
+		dml_uint_t ViewportYStart,
+		dml_bool_t GPUVMEnable,
+		dml_uint_t GPUVMMaxPageTableLevels,
+		dml_uint_t GPUVMMinPageSizeKBytes,
+		dml_uint_t PTEBufferSizeInRequests,
+		dml_uint_t Pitch,
+		dml_uint_t DCCMetaPitch,
+		dml_uint_t MacroTileWidth,
+		dml_uint_t MacroTileHeight,
+
+		// Output
+		dml_uint_t *MetaRowByte,
+		dml_uint_t *PixelPTEBytesPerRow, // for bandwidth calculation
+		dml_uint_t *PixelPTEBytesPerRowStorage, // for PTE buffer size check
+		dml_uint_t *dpte_row_width_ub,
+		dml_uint_t *dpte_row_height,
+		dml_uint_t *dpte_row_height_linear,
+		dml_uint_t *PixelPTEBytesPerRow_one_row_per_frame,
+		dml_uint_t *dpte_row_width_ub_one_row_per_frame,
+		dml_uint_t *dpte_row_height_one_row_per_frame,
+		dml_uint_t *MetaRequestWidth,
+		dml_uint_t *MetaRequestHeight,
+		dml_uint_t *meta_row_width,
+		dml_uint_t *meta_row_height,
+		dml_uint_t *PixelPTEReqWidth,
+		dml_uint_t *PixelPTEReqHeight,
+		dml_uint_t *PTERequestSize,
+		dml_uint_t *DPDE0BytesFrame,
+		dml_uint_t *MetaPTEBytesFrame)
+{
+	dml_uint_t MPDEBytesFrame;
+	dml_uint_t DCCMetaSurfaceBytes;
+	dml_uint_t ExtraDPDEBytesFrame;
+	dml_uint_t PDEAndMetaPTEBytesFrame;
+	dml_uint_t MacroTileSizeBytes;
+	dml_uint_t vp_height_meta_ub;
+	dml_uint_t vp_height_dpte_ub;
+
+	dml_uint_t PixelPTEReqWidth_linear = 0; // VBA_DELTA. VBA doesn't calculate this
+
+	*MetaRequestHeight = 8 * BlockHeight256Bytes;
+	*MetaRequestWidth = 8 * BlockWidth256Bytes;
+	if (SurfaceTiling == dml_sw_linear) {
+		*meta_row_height = 32;
+		*meta_row_width = (dml_uint_t)(dml_floor(ViewportXStart + SwathWidth + *MetaRequestWidth - 1, *MetaRequestWidth) - dml_floor(ViewportXStart, *MetaRequestWidth));
+	} else if (!dml_is_vertical_rotation(SourceScan)) {
+		*meta_row_height = *MetaRequestHeight;
+		if (ViewportStationary && NumberOfDPPs == 1) {
+			*meta_row_width = (dml_uint_t)(dml_floor(ViewportXStart + SwathWidth + *MetaRequestWidth - 1, *MetaRequestWidth) - dml_floor(ViewportXStart, *MetaRequestWidth));
+		} else {
+			*meta_row_width = (dml_uint_t)(dml_ceil(SwathWidth - 1, *MetaRequestWidth) + *MetaRequestWidth);
+		}
+		*MetaRowByte = (dml_uint_t)(*meta_row_width * *MetaRequestHeight * BytePerPixel / 256.0);
+	} else {
+		*meta_row_height = *MetaRequestWidth;
+		if (ViewportStationary && NumberOfDPPs == 1) {
+			*meta_row_width = (dml_uint_t)(dml_floor(ViewportYStart + ViewportHeight + *MetaRequestHeight - 1, *MetaRequestHeight) - dml_floor(ViewportYStart, *MetaRequestHeight));
+		} else {
+			*meta_row_width = (dml_uint_t)(dml_ceil(SwathWidth - 1, *MetaRequestHeight) + *MetaRequestHeight);
+		}
+		*MetaRowByte = (dml_uint_t)(*meta_row_width * *MetaRequestWidth * BytePerPixel / 256.0);
+	}
+
+	if (ViewportStationary && (NumberOfDPPs == 1 || !dml_is_vertical_rotation(SourceScan))) {
+		vp_height_meta_ub = (dml_uint_t)(dml_floor(ViewportYStart + ViewportHeight + 64 * BlockHeight256Bytes - 1, 64 * BlockHeight256Bytes) - dml_floor(ViewportYStart, 64 * BlockHeight256Bytes));
+	} else if (!dml_is_vertical_rotation(SourceScan)) {
+		vp_height_meta_ub = (dml_uint_t)(dml_ceil(ViewportHeight - 1, 64 * BlockHeight256Bytes) + 64 * BlockHeight256Bytes);
+	} else {
+		vp_height_meta_ub = (dml_uint_t)(dml_ceil(SwathWidth - 1, 64 * BlockHeight256Bytes) + 64 * BlockHeight256Bytes);
+	}
+
+	DCCMetaSurfaceBytes = (dml_uint_t)(DCCMetaPitch * vp_height_meta_ub * BytePerPixel / 256.0);
+
+	if (GPUVMEnable == true) {
+		*MetaPTEBytesFrame = (dml_uint_t)((dml_ceil((dml_float_t) (DCCMetaSurfaceBytes - 4.0 * 1024.0) / (8 * 4.0 * 1024), 1) + 1) * 64);
+		MPDEBytesFrame = 128 * (GPUVMMaxPageTableLevels - 1);
+	} else {
+		*MetaPTEBytesFrame = 0;
+		MPDEBytesFrame = 0;
+	}
+
+	if (DCCEnable != true) {
+		*MetaPTEBytesFrame = 0;
+		MPDEBytesFrame = 0;
+		*MetaRowByte = 0;
+	}
+
+	MacroTileSizeBytes = MacroTileWidth * BytePerPixel * MacroTileHeight;
+
+	if (ViewportStationary && (NumberOfDPPs == 1 || !dml_is_vertical_rotation(SourceScan))) {
+		vp_height_dpte_ub = (dml_uint_t)(dml_floor(ViewportYStart + ViewportHeight + MacroTileHeight - 1, MacroTileHeight) - dml_floor(ViewportYStart, MacroTileHeight));
+	} else if (!dml_is_vertical_rotation(SourceScan)) {
+		vp_height_dpte_ub = (dml_uint_t)(dml_ceil(ViewportHeight - 1, MacroTileHeight) + MacroTileHeight);
+	} else {
+		vp_height_dpte_ub = (dml_uint_t)(dml_ceil(SwathWidth - 1, MacroTileHeight) + MacroTileHeight);
+	}
+
+	if (GPUVMEnable == true && GPUVMMaxPageTableLevels > 1) {
+		*DPDE0BytesFrame = (dml_uint_t)(64 * (dml_ceil((dml_float_t) (Pitch * vp_height_dpte_ub * BytePerPixel - MacroTileSizeBytes) / (dml_float_t) (8 * 2097152), 1) + 1));
+		ExtraDPDEBytesFrame = 128 * (GPUVMMaxPageTableLevels - 2);
+	} else {
+		*DPDE0BytesFrame = 0;
+		ExtraDPDEBytesFrame = 0;
+	}
+
+	PDEAndMetaPTEBytesFrame = *MetaPTEBytesFrame + MPDEBytesFrame + *DPDE0BytesFrame + ExtraDPDEBytesFrame;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: DCCEnable = %u\n", __func__, DCCEnable);
+	dml_print("DML::%s: GPUVMEnable = %u\n", __func__, GPUVMEnable);
+	dml_print("DML::%s: SwModeLinear = %u\n", __func__, SurfaceTiling == dml_sw_linear);
+	dml_print("DML::%s: BytePerPixel = %u\n", __func__, BytePerPixel);
+	dml_print("DML::%s: GPUVMMaxPageTableLevels = %u\n", __func__, GPUVMMaxPageTableLevels);
+	dml_print("DML::%s: BlockHeight256Bytes = %u\n", __func__, BlockHeight256Bytes);
+	dml_print("DML::%s: BlockWidth256Bytes = %u\n", __func__, BlockWidth256Bytes);
+	dml_print("DML::%s: MacroTileHeight = %u\n", __func__, MacroTileHeight);
+	dml_print("DML::%s: MacroTileWidth = %u\n", __func__, MacroTileWidth);
+	dml_print("DML::%s: MetaPTEBytesFrame = %u\n", __func__, *MetaPTEBytesFrame);
+	dml_print("DML::%s: MPDEBytesFrame = %u\n", __func__, MPDEBytesFrame);
+	dml_print("DML::%s: DPDE0BytesFrame = %u\n", __func__, *DPDE0BytesFrame);
+	dml_print("DML::%s: ExtraDPDEBytesFrame= %u\n", __func__, ExtraDPDEBytesFrame);
+	dml_print("DML::%s: PDEAndMetaPTEBytesFrame = %u\n", __func__, PDEAndMetaPTEBytesFrame);
+	dml_print("DML::%s: ViewportHeight = %u\n", __func__, ViewportHeight);
+	dml_print("DML::%s: SwathWidth = %u\n", __func__, SwathWidth);
+	dml_print("DML::%s: vp_height_dpte_ub = %u\n", __func__, vp_height_dpte_ub);
+#endif
+
+	if (SurfaceTiling == dml_sw_linear) {
+		*PixelPTEReqHeight = 1;
+		*PixelPTEReqWidth = GPUVMMinPageSizeKBytes * 1024 * 8 / BytePerPixel;
+		PixelPTEReqWidth_linear = GPUVMMinPageSizeKBytes * 1024 * 8 / BytePerPixel;
+		*PTERequestSize = 64;
+	} else if (GPUVMMinPageSizeKBytes == 4) {
+		*PixelPTEReqHeight = 16 * BlockHeight256Bytes;
+		*PixelPTEReqWidth = 16 * BlockWidth256Bytes;
+		*PTERequestSize = 128;
+	} else {
+		*PixelPTEReqHeight = MacroTileHeight;
+		*PixelPTEReqWidth = 8 *  1024 * GPUVMMinPageSizeKBytes / (MacroTileHeight * BytePerPixel);
+		*PTERequestSize = 64;
+	}
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: GPUVMMinPageSizeKBytes = %u\n", __func__, GPUVMMinPageSizeKBytes);
+	dml_print("DML::%s: PDEAndMetaPTEBytesFrame = %u (after HostVM factor)\n", __func__, PDEAndMetaPTEBytesFrame);
+	dml_print("DML::%s: PixelPTEReqHeight = %u\n", __func__, *PixelPTEReqHeight);
+	dml_print("DML::%s: PixelPTEReqWidth = %u\n", __func__, *PixelPTEReqWidth);
+	dml_print("DML::%s: PixelPTEReqWidth_linear = %u\n", __func__, PixelPTEReqWidth_linear);
+	dml_print("DML::%s: PTERequestSize = %u\n", __func__, *PTERequestSize);
+	dml_print("DML::%s: Pitch = %u\n", __func__, Pitch);
+#endif
+
+	*dpte_row_height_one_row_per_frame = vp_height_dpte_ub;
+	*dpte_row_width_ub_one_row_per_frame = (dml_uint_t)((dml_ceil(((dml_float_t)Pitch * (dml_float_t) *dpte_row_height_one_row_per_frame / (dml_float_t) *PixelPTEReqHeight - 1) / (dml_float_t) *PixelPTEReqWidth, 1) + 1) * (dml_float_t) *PixelPTEReqWidth);
+	*PixelPTEBytesPerRow_one_row_per_frame = (dml_uint_t)((dml_float_t) *dpte_row_width_ub_one_row_per_frame / (dml_float_t) *PixelPTEReqWidth * *PTERequestSize);
+
+	if (SurfaceTiling == dml_sw_linear) {
+		*dpte_row_height = (dml_uint_t)(dml_min(128, 1 << (dml_uint_t) dml_floor(dml_log2(PTEBufferSizeInRequests * *PixelPTEReqWidth / Pitch), 1)));
+		dml_print("DML::%s: dpte_row_height term 1 = %u\n", __func__, PTEBufferSizeInRequests * *PixelPTEReqWidth / Pitch);
+		dml_print("DML::%s: dpte_row_height term 2 = %f\n", __func__, dml_log2(PTEBufferSizeInRequests * *PixelPTEReqWidth / Pitch));
+		dml_print("DML::%s: dpte_row_height term 3 = %f\n", __func__, dml_floor(dml_log2(PTEBufferSizeInRequests * *PixelPTEReqWidth / Pitch), 1));
+		dml_print("DML::%s: dpte_row_height term 4 = %u\n", __func__, 1 << (dml_uint_t) dml_floor(dml_log2(PTEBufferSizeInRequests * *PixelPTEReqWidth / Pitch), 1));
+		dml_print("DML::%s: dpte_row_height = %u\n", __func__, *dpte_row_height);
+
+		*dpte_row_width_ub = (dml_uint_t)(dml_ceil(((dml_float_t) Pitch * (dml_float_t) *dpte_row_height - 1), (dml_float_t) *PixelPTEReqWidth) + *PixelPTEReqWidth);
+		*PixelPTEBytesPerRow = (dml_uint_t)((dml_float_t) *dpte_row_width_ub / (dml_float_t) *PixelPTEReqWidth * *PTERequestSize);
+
+		// VBA_DELTA, VBA doesn't have programming value for pte row height linear.
+		*dpte_row_height_linear = 1 << (dml_uint_t) dml_floor(dml_log2(PTEBufferSizeInRequests * PixelPTEReqWidth_linear / Pitch), 1);
+		if (*dpte_row_height_linear > 128)
+			*dpte_row_height_linear = 128;
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: dpte_row_width_ub = %u (linear)\n", __func__, *dpte_row_width_ub);
+#endif
+
+	} else if (!dml_is_vertical_rotation(SourceScan)) {
+		*dpte_row_height = *PixelPTEReqHeight;
+
+		if (GPUVMMinPageSizeKBytes > 64) {
+			*dpte_row_width_ub = (dml_uint_t)((dml_ceil(((dml_float_t) Pitch * (dml_float_t) *dpte_row_height / (dml_float_t) *PixelPTEReqHeight - 1) / (dml_float_t) *PixelPTEReqWidth, 1) + 1) * *PixelPTEReqWidth);
+		} else if (ViewportStationary && (NumberOfDPPs == 1)) {
+			*dpte_row_width_ub = (dml_uint_t)(dml_floor(ViewportXStart + SwathWidth + *PixelPTEReqWidth - 1, *PixelPTEReqWidth) - dml_floor(ViewportXStart, *PixelPTEReqWidth));
+		} else {
+			*dpte_row_width_ub = (dml_uint_t)((dml_ceil((dml_float_t) (SwathWidth - 1) / (dml_float_t)*PixelPTEReqWidth, 1) + 1.0) * *PixelPTEReqWidth);
+		}
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: dpte_row_width_ub = %u (tiled horz)\n", __func__, *dpte_row_width_ub);
+#endif
+
+		ASSERT(*PixelPTEReqWidth);
+		if (*PixelPTEReqWidth != 0)
+			*PixelPTEBytesPerRow = *dpte_row_width_ub / *PixelPTEReqWidth * *PTERequestSize;
+	} else {
+		*dpte_row_height = (dml_uint_t)(dml_min(*PixelPTEReqWidth, MacroTileWidth));
+
+		if (ViewportStationary && (NumberOfDPPs == 1)) {
+			*dpte_row_width_ub = (dml_uint_t)(dml_floor(ViewportYStart + ViewportHeight + *PixelPTEReqHeight - 1, *PixelPTEReqHeight) - dml_floor(ViewportYStart, *PixelPTEReqHeight));
+		} else {
+			*dpte_row_width_ub = (dml_uint_t)((dml_ceil((dml_float_t) (SwathWidth - 1) / (dml_float_t) *PixelPTEReqHeight, 1) + 1) * *PixelPTEReqHeight);
+		}
+
+		*PixelPTEBytesPerRow = (dml_uint_t)((dml_float_t) *dpte_row_width_ub / (dml_float_t) *PixelPTEReqHeight * *PTERequestSize);
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: dpte_row_width_ub = %u (tiled vert)\n", __func__, *dpte_row_width_ub);
+#endif
+	}
+
+	if (GPUVMEnable != true)
+		*PixelPTEBytesPerRow = 0;
+
+	*PixelPTEBytesPerRowStorage = *PixelPTEBytesPerRow;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: GPUVMMinPageSizeKBytes = %u\n", __func__, GPUVMMinPageSizeKBytes);
+	dml_print("DML::%s: GPUVMEnable = %u\n", __func__, GPUVMEnable);
+	dml_print("DML::%s: dpte_row_height = %u\n", __func__, *dpte_row_height);
+	dml_print("DML::%s: dpte_row_height_linear = %u\n", __func__, *dpte_row_height_linear);
+	dml_print("DML::%s: dpte_row_width_ub = %u\n", __func__, *dpte_row_width_ub);
+	dml_print("DML::%s: PixelPTEBytesPerRow = %u\n", __func__, *PixelPTEBytesPerRow);
+	dml_print("DML::%s: PixelPTEBytesPerRowStorage = %u\n", __func__, *PixelPTEBytesPerRowStorage);
+	dml_print("DML::%s: PTEBufferSizeInRequests = %u\n", __func__, PTEBufferSizeInRequests);
+	dml_print("DML::%s: dpte_row_height_one_row_per_frame = %u\n", __func__, *dpte_row_height_one_row_per_frame);
+	dml_print("DML::%s: dpte_row_width_ub_one_row_per_frame = %u\n", __func__, *dpte_row_width_ub_one_row_per_frame);
+	dml_print("DML::%s: PixelPTEBytesPerRow_one_row_per_frame = %u\n", __func__, *PixelPTEBytesPerRow_one_row_per_frame);
+#endif
+
+	dml_print("DML: vm_bytes = meta_pte_bytes_per_frame (per_pipe) = MetaPTEBytesFrame = : %i\n", *MetaPTEBytesFrame);
+
+	return PDEAndMetaPTEBytesFrame;
+} // CalculateVMAndRowBytes
+
+static void PixelClockAdjustmentForProgressiveToInterlaceUnit(struct dml_display_cfg_st *display_cfg, dml_bool_t ptoi_supported)
+{
+	dml_uint_t num_active_planes = dml_get_num_active_planes(display_cfg);
+
+	//Progressive To Interlace Unit Effect
+	for (dml_uint_t k = 0; k < num_active_planes; ++k) {
+		display_cfg->output.PixelClockBackEnd[k] = display_cfg->timing.PixelClock[k];
+		if (display_cfg->timing.Interlace[k] == 1 && ptoi_supported == true) {
+			display_cfg->timing.PixelClock[k] = 2 * display_cfg->timing.PixelClock[k];
+		}
+	}
+}
+
+static dml_float_t TruncToValidBPP(
+		dml_float_t LinkBitRate,
+		dml_uint_t Lanes,
+		dml_uint_t HTotal,
+		dml_uint_t HActive,
+		dml_float_t PixelClock,
+		dml_float_t DesiredBPP,
+		dml_bool_t DSCEnable,
+		enum dml_output_encoder_class Output,
+		enum dml_output_format_class Format,
+		dml_uint_t DSCInputBitPerComponent,
+		dml_uint_t DSCSlices,
+		dml_uint_t AudioRate,
+		dml_uint_t AudioLayout,
+		enum dml_odm_mode ODMModeNoDSC,
+		enum dml_odm_mode ODMModeDSC,
+
+		// Output
+		dml_uint_t *RequiredSlots)
+{
+	dml_float_t MaxLinkBPP;
+	dml_uint_t MinDSCBPP;
+	dml_float_t MaxDSCBPP;
+	dml_uint_t NonDSCBPP0;
+	dml_uint_t NonDSCBPP1;
+	dml_uint_t NonDSCBPP2;
+
+	if (Format == dml_420) {
+		NonDSCBPP0 = 12;
+		NonDSCBPP1 = 15;
+		NonDSCBPP2 = 18;
+		MinDSCBPP = 6;
+		MaxDSCBPP = 1.5 * DSCInputBitPerComponent - 1.0 / 16;
+	} else if (Format == dml_444) {
+		NonDSCBPP0 = 24;
+		NonDSCBPP1 = 30;
+		NonDSCBPP2 = 36;
+		MinDSCBPP = 8;
+		MaxDSCBPP = 3 * DSCInputBitPerComponent - 1.0 / 16;
+	} else {
+		if (Output == dml_hdmi) {
+			NonDSCBPP0 = 24;
+			NonDSCBPP1 = 24;
+			NonDSCBPP2 = 24;
+		} else {
+			NonDSCBPP0 = 16;
+			NonDSCBPP1 = 20;
+			NonDSCBPP2 = 24;
+	}
+	if (Format == dml_n422) {
+		MinDSCBPP = 7;
+			MaxDSCBPP = 2 * DSCInputBitPerComponent - 1.0 / 16.0;
+		} else {
+			MinDSCBPP = 8;
+			MaxDSCBPP = 3 * DSCInputBitPerComponent - 1.0 / 16.0;
+		}
+	}
+
+	if (Output == dml_dp2p0) {
+		MaxLinkBPP = LinkBitRate * Lanes / PixelClock * 128.0 / 132.0 * 383.0 / 384.0 * 65536.0 / 65540.0;
+	} else if (DSCEnable && Output == dml_dp) {
+		MaxLinkBPP = LinkBitRate / 10.0 * 8.0 * Lanes / PixelClock * (1 - 2.4 / 100);
+	} else {
+		MaxLinkBPP = LinkBitRate / 10.0 * 8.0 * Lanes / PixelClock;
+	}
+
+	if (DSCEnable) {
+		if (ODMModeDSC == dml_odm_mode_combine_4to1) {
+			MaxLinkBPP = dml_min(MaxLinkBPP, 16);
+		} else if (ODMModeDSC == dml_odm_mode_combine_2to1) {
+			MaxLinkBPP = dml_min(MaxLinkBPP, 32);
+		} else if (ODMModeDSC == dml_odm_mode_split_1to2) {
+			MaxLinkBPP = 2 * MaxLinkBPP;
+		}
+	} else {
+		if (ODMModeNoDSC == dml_odm_mode_combine_4to1) {
+			MaxLinkBPP = dml_min(MaxLinkBPP, 16);
+		} else if (ODMModeNoDSC == dml_odm_mode_combine_2to1) {
+			MaxLinkBPP = dml_min(MaxLinkBPP, 32);
+		} else if (ODMModeNoDSC == dml_odm_mode_split_1to2) {
+			MaxLinkBPP = 2 * MaxLinkBPP;
+		}
+	}
+
+	*RequiredSlots = (dml_uint_t)(dml_ceil(DesiredBPP / MaxLinkBPP * 64, 1));
+
+	if (DesiredBPP == 0) {
+		if (DSCEnable) {
+			if (MaxLinkBPP < MinDSCBPP) {
+				return __DML_DPP_INVALID__;
+			} else if (MaxLinkBPP >= MaxDSCBPP) {
+				return MaxDSCBPP;
+			} else {
+				return dml_floor(16.0 * MaxLinkBPP, 1.0) / 16.0;
+			}
+		} else {
+			if (MaxLinkBPP >= NonDSCBPP2) {
+				return NonDSCBPP2;
+			} else if (MaxLinkBPP >= NonDSCBPP1) {
+				return NonDSCBPP1;
+			} else if (MaxLinkBPP >= NonDSCBPP0) {
+				return NonDSCBPP0;
+			} else {
+				return __DML_DPP_INVALID__;
+			}
+		}
+	} else {
+		if (!((DSCEnable == false && (DesiredBPP == NonDSCBPP2 || DesiredBPP == NonDSCBPP1 || DesiredBPP == NonDSCBPP0)) ||
+				(DSCEnable && DesiredBPP >= MinDSCBPP && DesiredBPP <= MaxDSCBPP))) {
+			return __DML_DPP_INVALID__;
+		} else {
+			return DesiredBPP;
+		}
+	}
+} // TruncToValidBPP
+
+static void CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport(
+		struct display_mode_lib_scratch_st *scratch,
+		struct CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_params_st *p)
+{
+	struct CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_locals_st *s = &scratch->CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_locals;
+
+	s->TotalActiveWriteback = 0;
+	p->Watermark->UrgentWatermark = p->mmSOCParameters.UrgentLatency + p->mmSOCParameters.ExtraLatency;
+	p->Watermark->USRRetrainingWatermark = p->mmSOCParameters.UrgentLatency + p->mmSOCParameters.ExtraLatency + p->mmSOCParameters.USRRetrainingLatency + p->mmSOCParameters.SMNLatency;
+	p->Watermark->DRAMClockChangeWatermark = p->mmSOCParameters.DRAMClockChangeLatency + p->Watermark->UrgentWatermark;
+	p->Watermark->FCLKChangeWatermark = p->mmSOCParameters.FCLKChangeLatency + p->Watermark->UrgentWatermark;
+	p->Watermark->StutterExitWatermark = p->mmSOCParameters.SRExitTime + p->mmSOCParameters.ExtraLatency + 10 / p->DCFClkDeepSleep;
+	p->Watermark->StutterEnterPlusExitWatermark = p->mmSOCParameters.SREnterPlusExitTime + p->mmSOCParameters.ExtraLatency + 10 / p->DCFClkDeepSleep;
+	p->Watermark->Z8StutterExitWatermark = p->mmSOCParameters.SRExitZ8Time + p->mmSOCParameters.ExtraLatency + 10 / p->DCFClkDeepSleep;
+	p->Watermark->Z8StutterEnterPlusExitWatermark = p->mmSOCParameters.SREnterPlusExitZ8Time + p->mmSOCParameters.ExtraLatency + 10 / p->DCFClkDeepSleep;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: UrgentLatency = %f\n", __func__, p->mmSOCParameters.UrgentLatency);
+	dml_print("DML::%s: ExtraLatency = %f\n", __func__, p->mmSOCParameters.ExtraLatency);
+	dml_print("DML::%s: DRAMClockChangeLatency = %f\n", __func__, p->mmSOCParameters.DRAMClockChangeLatency);
+	dml_print("DML::%s: UrgentWatermark = %f\n", __func__, p->Watermark->UrgentWatermark);
+	dml_print("DML::%s: USRRetrainingWatermark = %f\n", __func__, p->Watermark->USRRetrainingWatermark);
+	dml_print("DML::%s: DRAMClockChangeWatermark = %f\n", __func__, p->Watermark->DRAMClockChangeWatermark);
+	dml_print("DML::%s: FCLKChangeWatermark = %f\n", __func__, p->Watermark->FCLKChangeWatermark);
+	dml_print("DML::%s: StutterExitWatermark = %f\n", __func__, p->Watermark->StutterExitWatermark);
+	dml_print("DML::%s: StutterEnterPlusExitWatermark = %f\n", __func__, p->Watermark->StutterEnterPlusExitWatermark);
+	dml_print("DML::%s: Z8StutterExitWatermark = %f\n", __func__, p->Watermark->Z8StutterExitWatermark);
+	dml_print("DML::%s: Z8StutterEnterPlusExitWatermark = %f\n", __func__, p->Watermark->Z8StutterEnterPlusExitWatermark);
+#endif
+
+	s->TotalActiveWriteback = 0;
+	for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+		if (p->WritebackEnable[k] == true) {
+			s->TotalActiveWriteback = s->TotalActiveWriteback + 1;
+		}
+	}
+
+	if (s->TotalActiveWriteback <= 1) {
+		p->Watermark->WritebackUrgentWatermark = p->mmSOCParameters.WritebackLatency;
+	} else {
+		p->Watermark->WritebackUrgentWatermark = p->mmSOCParameters.WritebackLatency + p->WritebackChunkSize * 1024.0 / 32.0 / p->SOCCLK;
+	}
+	if (p->USRRetrainingRequiredFinal)
+		p->Watermark->WritebackUrgentWatermark = p->Watermark->WritebackUrgentWatermark + p->mmSOCParameters.USRRetrainingLatency;
+
+	if (s->TotalActiveWriteback <= 1) {
+		p->Watermark->WritebackDRAMClockChangeWatermark = p->mmSOCParameters.DRAMClockChangeLatency + p->mmSOCParameters.WritebackLatency;
+		p->Watermark->WritebackFCLKChangeWatermark = p->mmSOCParameters.FCLKChangeLatency + p->mmSOCParameters.WritebackLatency;
+	} else {
+		p->Watermark->WritebackDRAMClockChangeWatermark = p->mmSOCParameters.DRAMClockChangeLatency + p->mmSOCParameters.WritebackLatency + p->WritebackChunkSize * 1024.0 / 32.0 / p->SOCCLK;
+		p->Watermark->WritebackFCLKChangeWatermark = p->mmSOCParameters.FCLKChangeLatency + p->mmSOCParameters.WritebackLatency + p->WritebackChunkSize * 1024 / 32 / p->SOCCLK;
+	}
+
+	if (p->USRRetrainingRequiredFinal)
+		p->Watermark->WritebackDRAMClockChangeWatermark = p->Watermark->WritebackDRAMClockChangeWatermark + p->mmSOCParameters.USRRetrainingLatency;
+
+	if (p->USRRetrainingRequiredFinal)
+		p->Watermark->WritebackFCLKChangeWatermark = p->Watermark->WritebackFCLKChangeWatermark + p->mmSOCParameters.USRRetrainingLatency;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: WritebackDRAMClockChangeWatermark = %f\n", __func__, p->Watermark->WritebackDRAMClockChangeWatermark);
+	dml_print("DML::%s: WritebackFCLKChangeWatermark = %f\n", __func__, p->Watermark->WritebackFCLKChangeWatermark);
+	dml_print("DML::%s: WritebackUrgentWatermark = %f\n", __func__, p->Watermark->WritebackUrgentWatermark);
+	dml_print("DML::%s: USRRetrainingRequiredFinal = %u\n", __func__, p->USRRetrainingRequiredFinal);
+	dml_print("DML::%s: USRRetrainingLatency = %f\n", __func__, p->mmSOCParameters.USRRetrainingLatency);
+#endif
+
+	s->TotalPixelBW = 0.0;
+	for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+		s->TotalPixelBW = s->TotalPixelBW + p->DPPPerSurface[k]
+					* (p->SwathWidthY[k] * p->BytePerPixelDETY[k] * p->VRatio[k] + p->SwathWidthC[k] * p->BytePerPixelDETC[k] * p->VRatioChroma[k]) / (p->HTotal[k] / p->PixelClock[k]);
+	}
+
+	for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+
+		s->LBLatencyHidingSourceLinesY[k] = (dml_uint_t)(dml_min((dml_float_t)p->MaxLineBufferLines, dml_floor((dml_float_t)p->LineBufferSize / (dml_float_t)p->LBBitPerPixel[k] / ((dml_float_t)p->SwathWidthY[k] / dml_max(p->HRatio[k], 1.0)), 1)) - (p->VTaps[k] - 1));
+		s->LBLatencyHidingSourceLinesC[k] = (dml_uint_t)(dml_min((dml_float_t)p->MaxLineBufferLines, dml_floor((dml_float_t)p->LineBufferSize / (dml_float_t)p->LBBitPerPixel[k] / ((dml_float_t)p->SwathWidthC[k] / dml_max(p->HRatioChroma[k], 1.0)), 1)) - (p->VTapsChroma[k] - 1));
+
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%u, MaxLineBufferLines = %u\n", __func__, k, p->MaxLineBufferLines);
+		dml_print("DML::%s: k=%u, LineBufferSize = %u\n", __func__, k, p->LineBufferSize);
+		dml_print("DML::%s: k=%u, LBBitPerPixel = %u\n", __func__, k, p->LBBitPerPixel[k]);
+		dml_print("DML::%s: k=%u, HRatio = %f\n", __func__, k, p->HRatio[k]);
+		dml_print("DML::%s: k=%u, VTaps              = %u\n", __func__, k, p->VTaps[k]);
+#endif
+
+		s->EffectiveLBLatencyHidingY = s->LBLatencyHidingSourceLinesY[k] / p->VRatio[k] * (p->HTotal[k] / p->PixelClock[k]);
+		s->EffectiveLBLatencyHidingC = s->LBLatencyHidingSourceLinesC[k] / p->VRatioChroma[k] * (p->HTotal[k] / p->PixelClock[k]);
+
+		s->EffectiveDETBufferSizeY = p->DETBufferSizeY[k];
+		if (p->UnboundedRequestEnabled) {
+			s->EffectiveDETBufferSizeY = s->EffectiveDETBufferSizeY + p->CompressedBufferSizeInkByte * 1024 * (p->SwathWidthY[k] * p->BytePerPixelDETY[k] * p->VRatio[k]) / (p->HTotal[k] / p->PixelClock[k]) / s->TotalPixelBW;
+		}
+
+		s->LinesInDETY[k] = (dml_float_t)s->EffectiveDETBufferSizeY / p->BytePerPixelDETY[k] / p->SwathWidthY[k];
+		s->LinesInDETYRoundedDownToSwath[k] = (dml_uint_t)(dml_floor(s->LinesInDETY[k], p->SwathHeightY[k]));
+		s->FullDETBufferingTimeY = s->LinesInDETYRoundedDownToSwath[k] * (p->HTotal[k] / p->PixelClock[k]) / p->VRatio[k];
+
+		s->ActiveClockChangeLatencyHidingY = s->EffectiveLBLatencyHidingY + s->FullDETBufferingTimeY - ((dml_float_t)p->DSTXAfterScaler[k] / (dml_float_t)p->HTotal[k] + (dml_float_t)p->DSTYAfterScaler[k]) * (dml_float_t)p->HTotal[k] / p->PixelClock[k];
+
+		if (p->NumberOfActiveSurfaces > 1) {
+			s->ActiveClockChangeLatencyHidingY = s->ActiveClockChangeLatencyHidingY - (1.0 - 1.0 / (dml_float_t)p->NumberOfActiveSurfaces) * (dml_float_t)p->SwathHeightY[k] * (dml_float_t)p->HTotal[k] / p->PixelClock[k] / p->VRatio[k];
+	}
+
+		if (p->BytePerPixelDETC[k] > 0) {
+			s->LinesInDETC[k] = p->DETBufferSizeC[k] / p->BytePerPixelDETC[k] / p->SwathWidthC[k];
+			s->LinesInDETCRoundedDownToSwath[k] = (dml_uint_t)(dml_floor(s->LinesInDETC[k], p->SwathHeightC[k]));
+			s->FullDETBufferingTimeC = s->LinesInDETCRoundedDownToSwath[k] * (p->HTotal[k] / p->PixelClock[k]) / p->VRatioChroma[k];
+			s->ActiveClockChangeLatencyHidingC = s->EffectiveLBLatencyHidingC + s->FullDETBufferingTimeC - ((dml_float_t)p->DSTXAfterScaler[k] / (dml_float_t)p->HTotal[k] + (dml_float_t)p->DSTYAfterScaler[k]) * (dml_float_t)p->HTotal[k] / p->PixelClock[k];
+			if (p->NumberOfActiveSurfaces > 1) {
+			s->ActiveClockChangeLatencyHidingC = s->ActiveClockChangeLatencyHidingC - (1.0 - 1.0 / (dml_float_t)p->NumberOfActiveSurfaces) * (dml_float_t)p->SwathHeightC[k] * (dml_float_t)p->HTotal[k] / p->PixelClock[k] / p->VRatioChroma[k];
+			}
+			s->ActiveClockChangeLatencyHiding = dml_min(s->ActiveClockChangeLatencyHidingY, s->ActiveClockChangeLatencyHidingC);
+		} else {
+			s->ActiveClockChangeLatencyHiding = s->ActiveClockChangeLatencyHidingY;
+		}
+
+		s->ActiveDRAMClockChangeLatencyMargin[k] = s->ActiveClockChangeLatencyHiding - p->Watermark->UrgentWatermark - p->Watermark->DRAMClockChangeWatermark;
+		s->ActiveFCLKChangeLatencyMargin[k] = s->ActiveClockChangeLatencyHiding - p->Watermark->UrgentWatermark - p->Watermark->FCLKChangeWatermark;
+		s->USRRetrainingLatencyMargin[k] = s->ActiveClockChangeLatencyHiding - p->Watermark->USRRetrainingWatermark;
+
+		if (p->WritebackEnable[k]) {
+			s->WritebackLatencyHiding = (dml_float_t)p->WritebackInterfaceBufferSize * 1024.0 / ((dml_float_t)p->WritebackDestinationWidth[k] * (dml_float_t)p->WritebackDestinationHeight[k] / ((dml_float_t)p->WritebackSourceHeight[k] * (dml_float_t)p->HTotal[k] / p->PixelClock[k]) * 4.0);
+			if (p->WritebackPixelFormat[k] == dml_444_64) {
+				s->WritebackLatencyHiding = s->WritebackLatencyHiding / 2;
+			}
+			s->WritebackDRAMClockChangeLatencyMargin = s->WritebackLatencyHiding - p->Watermark->WritebackDRAMClockChangeWatermark;
+
+			s->WritebackFCLKChangeLatencyMargin = s->WritebackLatencyHiding - p->Watermark->WritebackFCLKChangeWatermark;
+
+			s->ActiveDRAMClockChangeLatencyMargin[k] = dml_min(s->ActiveDRAMClockChangeLatencyMargin[k], s->WritebackFCLKChangeLatencyMargin);
+			s->ActiveFCLKChangeLatencyMargin[k] = dml_min(s->ActiveFCLKChangeLatencyMargin[k], s->WritebackDRAMClockChangeLatencyMargin);
+		}
+		p->MaxActiveDRAMClockChangeLatencySupported[k] =  (p->UseMALLForPStateChange[k] == dml_use_mall_pstate_change_phantom_pipe) ? 0 : (s->ActiveDRAMClockChangeLatencyMargin[k] + p->mmSOCParameters.DRAMClockChangeLatency);
+		p->ActiveDRAMClockChangeLatencyMargin[k] = s->ActiveDRAMClockChangeLatencyMargin[k];
+	}
+
+	*p->USRRetrainingSupport = true;
+	for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+		if ((p->UseMALLForPStateChange[k] != dml_use_mall_pstate_change_phantom_pipe) && (s->USRRetrainingLatencyMargin[k] < 0)) {
+			*p->USRRetrainingSupport = false;
+		}
+	}
+
+	s->FoundCriticalSurface = false;
+	for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+		if ((p->UseMALLForPStateChange[k] != dml_use_mall_pstate_change_phantom_pipe) && ((!s->FoundCriticalSurface)
+			|| ((s->ActiveFCLKChangeLatencyMargin[k] + p->mmSOCParameters.FCLKChangeLatency) < *p->MaxActiveFCLKChangeLatencySupported))) {
+			s->FoundCriticalSurface = true;
+			*p->MaxActiveFCLKChangeLatencySupported = s->ActiveFCLKChangeLatencyMargin[k] + p->mmSOCParameters.FCLKChangeLatency;
+		}
+	}
+
+	for (dml_uint_t i = 0; i < p->NumberOfActiveSurfaces; ++i) {
+		for (dml_uint_t j = 0; j < p->NumberOfActiveSurfaces; ++j) {
+			if (i == j ||
+				(p->BlendingAndTiming[i] == i && p->BlendingAndTiming[j] == i) ||
+				(p->BlendingAndTiming[j] == j && p->BlendingAndTiming[i] == j) ||
+				(p->BlendingAndTiming[i] == p->BlendingAndTiming[j] && p->BlendingAndTiming[i] != i) ||
+				(p->SynchronizeTimingsFinal && p->PixelClock[i] == p->PixelClock[j] && p->HTotal[i] == p->HTotal[j] && p->VTotal[i] == p->VTotal[j] && p->VActive[i] == p->VActive[j]) ||
+				(p->SynchronizeDRRDisplaysForUCLKPStateChangeFinal && (p->DRRDisplay[i] || p->DRRDisplay[j]))) {
+				s->SynchronizedSurfaces[i][j] = true;
+			} else {
+				s->SynchronizedSurfaces[i][j] = false;
+			}
+		}
+	}
+
+	s->FCLKChangeSupportNumber = 0;
+	for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+		if ((p->UseMALLForPStateChange[k] != dml_use_mall_pstate_change_phantom_pipe) &&  (s->ActiveFCLKChangeLatencyMargin[k] < 0)) {
+			if (!(p->PrefetchMode[k] <= 1)) {
+				s->FCLKChangeSupportNumber = 3;
+			} else if (s->FCLKChangeSupportNumber == 0) {
+				s->FCLKChangeSupportNumber = ((p->SynchronizeDRRDisplaysForUCLKPStateChangeFinal && p->DRRDisplay[k]) ? 2 : 1);
+				s->LastSurfaceWithoutMargin = k;
+			} else if (((s->FCLKChangeSupportNumber == 1) && (p->DRRDisplay[k] || (!s->SynchronizedSurfaces[s->LastSurfaceWithoutMargin][k]))) || (s->FCLKChangeSupportNumber == 2))
+				s->FCLKChangeSupportNumber = 3;
+		}
+	}
+
+	if (s->FCLKChangeSupportNumber == 0) {
+		*p->FCLKChangeSupport = dml_fclock_change_vactive;
+	} else if ((s->FCLKChangeSupportNumber == 1) || (s->FCLKChangeSupportNumber == 2)) {
+		*p->FCLKChangeSupport = dml_fclock_change_vblank;
+	} else {
+		*p->FCLKChangeSupport = dml_fclock_change_unsupported;
+	}
+
+	s->DRAMClockChangeMethod = 0;
+	for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+		if (p->UseMALLForPStateChange[k] == dml_use_mall_pstate_change_full_frame)
+			s->DRAMClockChangeMethod = 1;
+		else if (p->UseMALLForPStateChange[k] == dml_use_mall_pstate_change_sub_viewport)
+			s->DRAMClockChangeMethod = 2;
+	}
+
+	s->DRAMClockChangeSupportNumber = 0;
+	for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+		if (((s->DRAMClockChangeMethod == 0) && (s->ActiveDRAMClockChangeLatencyMargin[k] < 0)) ||
+			((s->DRAMClockChangeMethod == 1) && (p->UseMALLForPStateChange[k] != dml_use_mall_pstate_change_full_frame)) ||
+			((s->DRAMClockChangeMethod == 2) && (p->UseMALLForPStateChange[k] != dml_use_mall_pstate_change_sub_viewport) && (p->UseMALLForPStateChange[k] != dml_use_mall_pstate_change_phantom_pipe))) {
+			if (p->PrefetchMode[k] != 0) { // Don't need to support DRAM clock change, PrefetchMode 0 means needs DRAM clock change support
+				s->DRAMClockChangeSupportNumber = 3;
+			} else if (s->DRAMClockChangeSupportNumber == 0) {
+				s->DRAMClockChangeSupportNumber =  (p->SynchronizeDRRDisplaysForUCLKPStateChangeFinal && p->DRRDisplay[k]) ? 2 : 1;
+				s->LastSurfaceWithoutMargin = k;
+			} else if (((s->DRAMClockChangeSupportNumber == 1) && (p->DRRDisplay[k] || !s->SynchronizedSurfaces[s->LastSurfaceWithoutMargin][k])) || (s->DRAMClockChangeSupportNumber == 2)) {
+				s->DRAMClockChangeSupportNumber = 3;
+			}
+		}
+	}
+
+	if (s->DRAMClockChangeMethod == 0) { // No MALL usage
+		if (s->DRAMClockChangeSupportNumber == 0) {
+			*p->DRAMClockChangeSupport = dml_dram_clock_change_vactive;
+		} else if (s->DRAMClockChangeSupportNumber == 1) {
+			*p->DRAMClockChangeSupport = dml_dram_clock_change_vblank;
+		} else if (s->DRAMClockChangeSupportNumber == 2) {
+			*p->DRAMClockChangeSupport = dml_dram_clock_change_vblank_drr;
+		} else {
+			*p->DRAMClockChangeSupport = dml_dram_clock_change_unsupported;
+		}
+	} else if (s->DRAMClockChangeMethod == 1) { // Any pipe using MALL full frame
+		if (s->DRAMClockChangeSupportNumber == 0) {
+			*p->DRAMClockChangeSupport = dml_dram_clock_change_vactive_w_mall_full_frame;
+		} else if (s->DRAMClockChangeSupportNumber == 1) {
+			*p->DRAMClockChangeSupport = dml_dram_clock_change_vblank_w_mall_full_frame;
+		} else if (s->DRAMClockChangeSupportNumber == 2) {
+			*p->DRAMClockChangeSupport = dml_dram_clock_change_vblank_drr_w_mall_full_frame;
+		} else {
+			*p->DRAMClockChangeSupport = dml_dram_clock_change_unsupported;
+		}
+	} else { // Any pipe using MALL subviewport
+		if (s->DRAMClockChangeSupportNumber == 0) {
+			*p->DRAMClockChangeSupport = dml_dram_clock_change_vactive_w_mall_sub_vp;
+		} else if (s->DRAMClockChangeSupportNumber == 1) {
+			*p->DRAMClockChangeSupport = dml_dram_clock_change_vblank_w_mall_sub_vp;
+		} else if (s->DRAMClockChangeSupportNumber == 2) {
+			*p->DRAMClockChangeSupport = dml_dram_clock_change_vblank_drr_w_mall_sub_vp;
+		} else {
+			*p->DRAMClockChangeSupport = dml_dram_clock_change_unsupported;
+		}
+	}
+
+	for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+		s->dst_y_pstate = (dml_uint_t)(dml_ceil((p->mmSOCParameters.DRAMClockChangeLatency + p->mmSOCParameters.UrgentLatency) / (p->HTotal[k] / p->PixelClock[k]), 1));
+		s->src_y_pstate_l = (dml_uint_t)(dml_ceil(s->dst_y_pstate * p->VRatio[k], p->SwathHeightY[k]));
+		s->src_y_ahead_l = (dml_uint_t)(dml_floor(p->DETBufferSizeY[k] / p->BytePerPixelDETY[k] / p->SwathWidthY[k], p->SwathHeightY[k]) + s->LBLatencyHidingSourceLinesY[k]);
+		s->sub_vp_lines_l = s->src_y_pstate_l + s->src_y_ahead_l + p->meta_row_height[k];
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%u, DETBufferSizeY = %u\n", __func__, k, p->DETBufferSizeY[k]);
+		dml_print("DML::%s: k=%u, BytePerPixelDETY = %f\n", __func__, k, p->BytePerPixelDETY[k]);
+		dml_print("DML::%s: k=%u, SwathWidthY = %u\n", __func__, k, p->SwathWidthY[k]);
+		dml_print("DML::%s: k=%u, SwathHeightY = %u\n", __func__, k, p->SwathHeightY[k]);
+		dml_print("DML::%s: k=%u, LBLatencyHidingSourceLinesY = %u\n", __func__, k, s->LBLatencyHidingSourceLinesY[k]);
+		dml_print("DML::%s: k=%u, dst_y_pstate = %u\n", __func__, k, s->dst_y_pstate);
+		dml_print("DML::%s: k=%u, src_y_pstate_l = %u\n", __func__, k, s->src_y_pstate_l);
+		dml_print("DML::%s: k=%u, src_y_ahead_l = %u\n", __func__, k, s->src_y_ahead_l);
+		dml_print("DML::%s: k=%u, meta_row_height = %u\n", __func__, k, p->meta_row_height[k]);
+		dml_print("DML::%s: k=%u, sub_vp_lines_l  = %u\n", __func__, k, s->sub_vp_lines_l);
+#endif
+		p->SubViewportLinesNeededInMALL[k] = s->sub_vp_lines_l;
+
+		if (p->BytePerPixelDETC[k] > 0) {
+		s->src_y_pstate_c = (dml_uint_t)(dml_ceil(s->dst_y_pstate * p->VRatioChroma[k], p->SwathHeightC[k]));
+		s->src_y_ahead_c = (dml_uint_t)(dml_floor(p->DETBufferSizeC[k] / p->BytePerPixelDETC[k] / p->SwathWidthC[k], p->SwathHeightC[k]) + s->LBLatencyHidingSourceLinesC[k]);
+		s->sub_vp_lines_c = s->src_y_pstate_c + s->src_y_ahead_c + p->meta_row_height_chroma[k];
+		p->SubViewportLinesNeededInMALL[k] = (dml_uint_t)(dml_max(s->sub_vp_lines_l, s->sub_vp_lines_c));
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%u, src_y_pstate_c = %u\n", __func__, k, s->src_y_pstate_c);
+		dml_print("DML::%s: k=%u, src_y_ahead_c = %u\n", __func__, k, s->src_y_ahead_c);
+		dml_print("DML::%s: k=%u, meta_row_height_chroma = %u\n", __func__, k, p->meta_row_height_chroma[k]);
+		dml_print("DML::%s: k=%u, sub_vp_lines_c            = %u\n", __func__, k, s->sub_vp_lines_c);
+#endif
+		}
+	}
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: DRAMClockChangeSupport = %u\n", __func__, *p->DRAMClockChangeSupport);
+	dml_print("DML::%s: FCLKChangeSupport = %u\n", __func__, *p->FCLKChangeSupport);
+	dml_print("DML::%s: MaxActiveFCLKChangeLatencySupported = %f\n", __func__, *p->MaxActiveFCLKChangeLatencySupported);
+	dml_print("DML::%s: USRRetrainingSupport                        = %u\n", __func__, *p->USRRetrainingSupport);
+#endif
+} // CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport
+
+static void CalculateDCFCLKDeepSleep(
+		dml_uint_t NumberOfActiveSurfaces,
+		dml_uint_t BytePerPixelY[],
+		dml_uint_t BytePerPixelC[],
+		dml_float_t VRatio[],
+		dml_float_t VRatioChroma[],
+		dml_uint_t SwathWidthY[],
+		dml_uint_t SwathWidthC[],
+		dml_uint_t DPPPerSurface[],
+		dml_float_t HRatio[],
+		dml_float_t HRatioChroma[],
+		dml_float_t PixelClock[],
+		dml_float_t PSCL_THROUGHPUT[],
+		dml_float_t PSCL_THROUGHPUT_CHROMA[],
+		dml_float_t Dppclk[],
+		dml_float_t ReadBandwidthLuma[],
+		dml_float_t ReadBandwidthChroma[],
+		dml_uint_t ReturnBusWidth,
+
+		// Output
+		dml_float_t *DCFClkDeepSleep)
+{
+	dml_float_t DisplayPipeLineDeliveryTimeLuma;
+	dml_float_t DisplayPipeLineDeliveryTimeChroma;
+	dml_float_t DCFClkDeepSleepPerSurface[__DML_NUM_PLANES__];
+	dml_float_t ReadBandwidth = 0.0;
+
+	for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+
+		if (VRatio[k] <= 1) {
+			DisplayPipeLineDeliveryTimeLuma = SwathWidthY[k] * DPPPerSurface[k] / HRatio[k] / PixelClock[k];
+		} else {
+			DisplayPipeLineDeliveryTimeLuma = SwathWidthY[k] / PSCL_THROUGHPUT[k] / Dppclk[k];
+		}
+		if (BytePerPixelC[k] == 0) {
+			DisplayPipeLineDeliveryTimeChroma = 0;
+		} else {
+			if (VRatioChroma[k] <= 1) {
+				DisplayPipeLineDeliveryTimeChroma = SwathWidthC[k] * DPPPerSurface[k] / HRatioChroma[k] / PixelClock[k];
+			} else {
+				DisplayPipeLineDeliveryTimeChroma = SwathWidthC[k] / PSCL_THROUGHPUT_CHROMA[k] / Dppclk[k];
+			}
+		}
+
+		if (BytePerPixelC[k] > 0) {
+			DCFClkDeepSleepPerSurface[k] = dml_max(__DML_MIN_DCFCLK_FACTOR__ * SwathWidthY[k] * BytePerPixelY[k] / 32.0 / DisplayPipeLineDeliveryTimeLuma,
+														__DML_MIN_DCFCLK_FACTOR__ * SwathWidthC[k] * BytePerPixelC[k] / 32.0 / DisplayPipeLineDeliveryTimeChroma);
+		} else {
+			DCFClkDeepSleepPerSurface[k] = __DML_MIN_DCFCLK_FACTOR__ * SwathWidthY[k] * BytePerPixelY[k] / 64.0 / DisplayPipeLineDeliveryTimeLuma;
+		}
+		DCFClkDeepSleepPerSurface[k] = dml_max(DCFClkDeepSleepPerSurface[k], PixelClock[k] / 16);
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%u, PixelClock = %f\n", __func__, k, PixelClock[k]);
+		dml_print("DML::%s: k=%u, DCFClkDeepSleepPerSurface = %f\n", __func__, k, DCFClkDeepSleepPerSurface[k]);
+#endif
+	}
+
+	for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+		ReadBandwidth = ReadBandwidth + ReadBandwidthLuma[k] + ReadBandwidthChroma[k];
+	}
+
+	*DCFClkDeepSleep = dml_max(8.0, __DML_MIN_DCFCLK_FACTOR__ * ReadBandwidth / (dml_float_t) ReturnBusWidth);
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: __DML_MIN_DCFCLK_FACTOR__ = %f\n", __func__, __DML_MIN_DCFCLK_FACTOR__);
+		dml_print("DML::%s: ReadBandwidth = %f\n", __func__, ReadBandwidth);
+		dml_print("DML::%s: ReturnBusWidth = %u\n", __func__, ReturnBusWidth);
+		dml_print("DML::%s: DCFClkDeepSleep = %f\n", __func__, *DCFClkDeepSleep);
+#endif
+
+	for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+		*DCFClkDeepSleep = dml_max(*DCFClkDeepSleep, DCFClkDeepSleepPerSurface[k]);
+	}
+	dml_print("DML::%s: DCFClkDeepSleep = %f (final)\n", __func__, *DCFClkDeepSleep);
+} // CalculateDCFCLKDeepSleep
+
+static void CalculateUrgentBurstFactor(
+		enum dml_use_mall_for_pstate_change_mode UseMALLForPStateChange,
+		dml_uint_t swath_width_luma_ub,
+		dml_uint_t swath_width_chroma_ub,
+		dml_uint_t SwathHeightY,
+		dml_uint_t SwathHeightC,
+		dml_float_t LineTime,
+		dml_float_t UrgentLatency,
+		dml_float_t CursorBufferSize,
+		dml_uint_t CursorWidth,
+		dml_uint_t CursorBPP,
+		dml_float_t VRatio,
+		dml_float_t VRatioC,
+		dml_float_t BytePerPixelInDETY,
+		dml_float_t BytePerPixelInDETC,
+		dml_uint_t DETBufferSizeY,
+		dml_uint_t DETBufferSizeC,
+		// Output
+		dml_float_t *UrgentBurstFactorCursor,
+		dml_float_t *UrgentBurstFactorLuma,
+		dml_float_t *UrgentBurstFactorChroma,
+		dml_bool_t *NotEnoughUrgentLatencyHiding)
+{
+	dml_float_t LinesInDETLuma;
+	dml_float_t LinesInDETChroma;
+	dml_uint_t LinesInCursorBuffer;
+	dml_float_t CursorBufferSizeInTime;
+	dml_float_t DETBufferSizeInTimeLuma;
+	dml_float_t DETBufferSizeInTimeChroma;
+
+	*NotEnoughUrgentLatencyHiding = 0;
+
+	if (CursorWidth > 0) {
+		LinesInCursorBuffer = 1 << (dml_uint_t) dml_floor(dml_log2(CursorBufferSize * 1024.0 / (CursorWidth * CursorBPP / 8.0)), 1.0);
+		if (VRatio > 0) {
+			CursorBufferSizeInTime = LinesInCursorBuffer * LineTime / VRatio;
+			if (CursorBufferSizeInTime - UrgentLatency <= 0) {
+				*NotEnoughUrgentLatencyHiding = 1;
+				*UrgentBurstFactorCursor = 0;
+			} else {
+				*UrgentBurstFactorCursor = CursorBufferSizeInTime / (CursorBufferSizeInTime - UrgentLatency);
+			}
+		} else {
+			*UrgentBurstFactorCursor = 1;
+		}
+	}
+
+	LinesInDETLuma = (UseMALLForPStateChange == dml_use_mall_pstate_change_phantom_pipe ? 1024*1024 : DETBufferSizeY) / BytePerPixelInDETY / swath_width_luma_ub;
+
+	if (VRatio > 0) {
+		DETBufferSizeInTimeLuma = dml_floor(LinesInDETLuma, SwathHeightY) * LineTime / VRatio;
+		if (DETBufferSizeInTimeLuma - UrgentLatency <= 0) {
+			*NotEnoughUrgentLatencyHiding = 1;
+			*UrgentBurstFactorLuma = 0;
+		} else {
+			*UrgentBurstFactorLuma = DETBufferSizeInTimeLuma / (DETBufferSizeInTimeLuma - UrgentLatency);
+		}
+	} else {
+		*UrgentBurstFactorLuma = 1;
+	}
+
+	if (BytePerPixelInDETC > 0) {
+		LinesInDETChroma = (UseMALLForPStateChange == dml_use_mall_pstate_change_phantom_pipe ? 1024*1024 : DETBufferSizeC) / BytePerPixelInDETC / swath_width_chroma_ub;
+
+		if (VRatioC > 0) {
+			DETBufferSizeInTimeChroma = dml_floor(LinesInDETChroma, SwathHeightC) * LineTime / VRatioC;
+			if (DETBufferSizeInTimeChroma - UrgentLatency <= 0) {
+				*NotEnoughUrgentLatencyHiding = 1;
+				*UrgentBurstFactorChroma = 0;
+			} else {
+				*UrgentBurstFactorChroma = DETBufferSizeInTimeChroma / (DETBufferSizeInTimeChroma - UrgentLatency);
+			}
+		} else {
+			*UrgentBurstFactorChroma = 1;
+		}
+	}
+} // CalculateUrgentBurstFactor
+
+static void CalculatePixelDeliveryTimes(
+		dml_uint_t NumberOfActiveSurfaces,
+		dml_float_t VRatio[],
+		dml_float_t VRatioChroma[],
+		dml_float_t VRatioPrefetchY[],
+		dml_float_t VRatioPrefetchC[],
+		dml_uint_t swath_width_luma_ub[],
+		dml_uint_t swath_width_chroma_ub[],
+		dml_uint_t DPPPerSurface[],
+		dml_float_t HRatio[],
+		dml_float_t HRatioChroma[],
+		dml_float_t PixelClock[],
+		dml_float_t PSCL_THROUGHPUT[],
+		dml_float_t PSCL_THROUGHPUT_CHROMA[],
+		dml_float_t Dppclk[],
+		dml_uint_t BytePerPixelC[],
+		enum dml_rotation_angle SourceScan[],
+		dml_uint_t NumberOfCursors[],
+		dml_uint_t CursorWidth[],
+		dml_uint_t CursorBPP[],
+		dml_uint_t BlockWidth256BytesY[],
+		dml_uint_t BlockHeight256BytesY[],
+		dml_uint_t BlockWidth256BytesC[],
+		dml_uint_t BlockHeight256BytesC[],
+
+		// Output
+		dml_float_t DisplayPipeLineDeliveryTimeLuma[],
+		dml_float_t DisplayPipeLineDeliveryTimeChroma[],
+		dml_float_t DisplayPipeLineDeliveryTimeLumaPrefetch[],
+		dml_float_t DisplayPipeLineDeliveryTimeChromaPrefetch[],
+		dml_float_t DisplayPipeRequestDeliveryTimeLuma[],
+		dml_float_t DisplayPipeRequestDeliveryTimeChroma[],
+		dml_float_t DisplayPipeRequestDeliveryTimeLumaPrefetch[],
+		dml_float_t DisplayPipeRequestDeliveryTimeChromaPrefetch[],
+		dml_float_t CursorRequestDeliveryTime[],
+		dml_float_t CursorRequestDeliveryTimePrefetch[])
+{
+		dml_float_t req_per_swath_ub;
+
+		for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%u : HRatio = %f\n", __func__, k, HRatio[k]);
+		dml_print("DML::%s: k=%u : VRatio = %f\n", __func__, k, VRatio[k]);
+		dml_print("DML::%s: k=%u : HRatioChroma = %f\n", __func__, k, HRatioChroma[k]);
+		dml_print("DML::%s: k=%u : VRatioChroma = %f\n", __func__, k, VRatioChroma[k]);
+		dml_print("DML::%s: k=%u : swath_width_luma_ub = %u\n", __func__, k, swath_width_luma_ub[k]);
+		dml_print("DML::%s: k=%u : swath_width_chroma_ub = %u\n", __func__, k, swath_width_chroma_ub[k]);
+		dml_print("DML::%s: k=%u : PSCL_THROUGHPUT = %f\n", __func__, k, PSCL_THROUGHPUT[k]);
+		dml_print("DML::%s: k=%u : PSCL_THROUGHPUT_CHROMA = %f\n", __func__, k, PSCL_THROUGHPUT_CHROMA[k]);
+		dml_print("DML::%s: k=%u : DPPPerSurface = %u\n", __func__, k, DPPPerSurface[k]);
+		dml_print("DML::%s: k=%u : PixelClock = %f\n", __func__, k, PixelClock[k]);
+		dml_print("DML::%s: k=%u : Dppclk = %f\n", __func__, k, Dppclk[k]);
+#endif
+
+		if (VRatio[k] <= 1) {
+			DisplayPipeLineDeliveryTimeLuma[k] = swath_width_luma_ub[k] * DPPPerSurface[k] / HRatio[k] / PixelClock[k];
+		} else {
+			DisplayPipeLineDeliveryTimeLuma[k] = swath_width_luma_ub[k] / PSCL_THROUGHPUT[k] / Dppclk[k];
+		}
+
+		if (BytePerPixelC[k] == 0) {
+			DisplayPipeLineDeliveryTimeChroma[k] = 0;
+		} else {
+			if (VRatioChroma[k] <= 1) {
+				DisplayPipeLineDeliveryTimeChroma[k] = swath_width_chroma_ub[k] * DPPPerSurface[k] / HRatioChroma[k] / PixelClock[k];
+			} else {
+				DisplayPipeLineDeliveryTimeChroma[k] = swath_width_chroma_ub[k] / PSCL_THROUGHPUT_CHROMA[k] / Dppclk[k];
+			}
+		}
+
+		if (VRatioPrefetchY[k] <= 1) {
+			DisplayPipeLineDeliveryTimeLumaPrefetch[k] = swath_width_luma_ub[k] * DPPPerSurface[k] / HRatio[k] / PixelClock[k];
+		} else {
+			DisplayPipeLineDeliveryTimeLumaPrefetch[k] = swath_width_luma_ub[k] / PSCL_THROUGHPUT[k] / Dppclk[k];
+		}
+
+		if (BytePerPixelC[k] == 0) {
+			DisplayPipeLineDeliveryTimeChromaPrefetch[k] = 0;
+		} else {
+			if (VRatioPrefetchC[k] <= 1) {
+				DisplayPipeLineDeliveryTimeChromaPrefetch[k] = swath_width_chroma_ub[k] * DPPPerSurface[k] / HRatioChroma[k] / PixelClock[k];
+			} else {
+				DisplayPipeLineDeliveryTimeChromaPrefetch[k] = swath_width_chroma_ub[k] / PSCL_THROUGHPUT_CHROMA[k] / Dppclk[k];
+			}
+		}
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%u : DisplayPipeLineDeliveryTimeLuma = %f\n", __func__, k, DisplayPipeLineDeliveryTimeLuma[k]);
+		dml_print("DML::%s: k=%u : DisplayPipeLineDeliveryTimeLumaPrefetch = %f\n", __func__, k, DisplayPipeLineDeliveryTimeLumaPrefetch[k]);
+		dml_print("DML::%s: k=%u : DisplayPipeLineDeliveryTimeChroma = %f\n", __func__, k, DisplayPipeLineDeliveryTimeChroma[k]);
+		dml_print("DML::%s: k=%u : DisplayPipeLineDeliveryTimeChromaPrefetch = %f\n", __func__, k, DisplayPipeLineDeliveryTimeChromaPrefetch[k]);
+#endif
+	}
+
+	for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+		if (!dml_is_vertical_rotation(SourceScan[k])) {
+			req_per_swath_ub = swath_width_luma_ub[k] / BlockWidth256BytesY[k];
+		} else {
+			req_per_swath_ub = swath_width_luma_ub[k] / BlockHeight256BytesY[k];
+		}
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%u : req_per_swath_ub = %f (Luma)\n", __func__, k, req_per_swath_ub);
+#endif
+
+		DisplayPipeRequestDeliveryTimeLuma[k] = DisplayPipeLineDeliveryTimeLuma[k] / req_per_swath_ub;
+		DisplayPipeRequestDeliveryTimeLumaPrefetch[k] = DisplayPipeLineDeliveryTimeLumaPrefetch[k] / req_per_swath_ub;
+		if (BytePerPixelC[k] == 0) {
+			DisplayPipeRequestDeliveryTimeChroma[k] = 0;
+			DisplayPipeRequestDeliveryTimeChromaPrefetch[k] = 0;
+		} else {
+			if (!dml_is_vertical_rotation(SourceScan[k])) {
+				req_per_swath_ub = swath_width_chroma_ub[k] / BlockWidth256BytesC[k];
+			} else {
+				req_per_swath_ub = swath_width_chroma_ub[k] / BlockHeight256BytesC[k];
+			}
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: k=%u : req_per_swath_ub = %f (Chroma)\n", __func__, k, req_per_swath_ub);
+#endif
+			DisplayPipeRequestDeliveryTimeChroma[k] = DisplayPipeLineDeliveryTimeChroma[k] / req_per_swath_ub;
+			DisplayPipeRequestDeliveryTimeChromaPrefetch[k] = DisplayPipeLineDeliveryTimeChromaPrefetch[k] / req_per_swath_ub;
+		}
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%u : DisplayPipeRequestDeliveryTimeLuma = %f\n", __func__, k, DisplayPipeRequestDeliveryTimeLuma[k]);
+		dml_print("DML::%s: k=%u : DisplayPipeRequestDeliveryTimeLumaPrefetch = %f\n", __func__, k, DisplayPipeRequestDeliveryTimeLumaPrefetch[k]);
+		dml_print("DML::%s: k=%u : DisplayPipeRequestDeliveryTimeChroma = %f\n", __func__, k, DisplayPipeRequestDeliveryTimeChroma[k]);
+		dml_print("DML::%s: k=%u : DisplayPipeRequestDeliveryTimeChromaPrefetch = %f\n", __func__, k, DisplayPipeRequestDeliveryTimeChromaPrefetch[k]);
+#endif
+	}
+
+	for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+		dml_uint_t cursor_req_per_width;
+		cursor_req_per_width = (dml_uint_t)(dml_ceil((dml_float_t) CursorWidth[k] * (dml_float_t) CursorBPP[k] / 256.0 / 8.0, 1.0));
+		if (NumberOfCursors[k] > 0) {
+			if (VRatio[k] <= 1) {
+				CursorRequestDeliveryTime[k] = (dml_float_t) CursorWidth[k] / HRatio[k] / PixelClock[k] / cursor_req_per_width;
+			} else {
+				CursorRequestDeliveryTime[k] = (dml_float_t) CursorWidth[k] / PSCL_THROUGHPUT[k] / Dppclk[k] / cursor_req_per_width;
+			}
+			if (VRatioPrefetchY[k] <= 1) {
+				CursorRequestDeliveryTimePrefetch[k] = (dml_float_t) CursorWidth[k] / HRatio[k] / PixelClock[k] / cursor_req_per_width;
+		} else {
+			CursorRequestDeliveryTimePrefetch[k] = (dml_float_t) CursorWidth[k] / PSCL_THROUGHPUT[k] / Dppclk[k] / cursor_req_per_width;
+			}
+		} else {
+			CursorRequestDeliveryTime[k] = 0;
+			CursorRequestDeliveryTimePrefetch[k] = 0;
+		}
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%u : NumberOfCursors = %u\n", __func__, k, NumberOfCursors[k]);
+		dml_print("DML::%s: k=%u : CursorRequestDeliveryTime = %f\n", __func__, k, CursorRequestDeliveryTime[k]);
+		dml_print("DML::%s: k=%u : CursorRequestDeliveryTimePrefetch = %f\n", __func__, k, CursorRequestDeliveryTimePrefetch[k]);
+#endif
+	}
+} // CalculatePixelDeliveryTimes
+
+static void CalculateMetaAndPTETimes(
+		dml_bool_t use_one_row_for_frame[],
+		dml_uint_t NumberOfActiveSurfaces,
+		dml_bool_t GPUVMEnable,
+		dml_uint_t MetaChunkSize,
+		dml_uint_t MinMetaChunkSizeBytes,
+		dml_uint_t HTotal[],
+		dml_float_t VRatio[],
+		dml_float_t VRatioChroma[],
+		dml_float_t DestinationLinesToRequestRowInVBlank[],
+		dml_float_t DestinationLinesToRequestRowInImmediateFlip[],
+		dml_bool_t DCCEnable[],
+		dml_float_t PixelClock[],
+		dml_uint_t BytePerPixelY[],
+		dml_uint_t BytePerPixelC[],
+		enum dml_rotation_angle SourceScan[],
+		dml_uint_t dpte_row_height[],
+		dml_uint_t dpte_row_height_chroma[],
+		dml_uint_t meta_row_width[],
+		dml_uint_t meta_row_width_chroma[],
+		dml_uint_t meta_row_height[],
+		dml_uint_t meta_row_height_chroma[],
+		dml_uint_t meta_req_width[],
+		dml_uint_t meta_req_width_chroma[],
+		dml_uint_t meta_req_height[],
+		dml_uint_t meta_req_height_chroma[],
+		dml_uint_t dpte_group_bytes[],
+		dml_uint_t PTERequestSizeY[],
+		dml_uint_t PTERequestSizeC[],
+		dml_uint_t PixelPTEReqWidthY[],
+		dml_uint_t PixelPTEReqHeightY[],
+		dml_uint_t PixelPTEReqWidthC[],
+		dml_uint_t PixelPTEReqHeightC[],
+		dml_uint_t dpte_row_width_luma_ub[],
+		dml_uint_t dpte_row_width_chroma_ub[],
+
+		// Output
+		dml_float_t DST_Y_PER_PTE_ROW_NOM_L[],
+		dml_float_t DST_Y_PER_PTE_ROW_NOM_C[],
+		dml_float_t DST_Y_PER_META_ROW_NOM_L[],
+		dml_float_t DST_Y_PER_META_ROW_NOM_C[],
+		dml_float_t TimePerMetaChunkNominal[],
+		dml_float_t TimePerChromaMetaChunkNominal[],
+		dml_float_t TimePerMetaChunkVBlank[],
+		dml_float_t TimePerChromaMetaChunkVBlank[],
+		dml_float_t TimePerMetaChunkFlip[],
+		dml_float_t TimePerChromaMetaChunkFlip[],
+		dml_float_t time_per_pte_group_nom_luma[],
+		dml_float_t time_per_pte_group_vblank_luma[],
+		dml_float_t time_per_pte_group_flip_luma[],
+		dml_float_t time_per_pte_group_nom_chroma[],
+		dml_float_t time_per_pte_group_vblank_chroma[],
+		dml_float_t time_per_pte_group_flip_chroma[])
+{
+	dml_uint_t meta_chunk_width;
+	dml_uint_t min_meta_chunk_width;
+	dml_uint_t meta_chunk_per_row_int;
+	dml_uint_t meta_row_remainder;
+	dml_uint_t meta_chunk_threshold;
+	dml_uint_t meta_chunks_per_row_ub;
+	dml_uint_t meta_chunk_width_chroma;
+	dml_uint_t min_meta_chunk_width_chroma;
+	dml_uint_t meta_chunk_per_row_int_chroma;
+	dml_uint_t meta_row_remainder_chroma;
+	dml_uint_t meta_chunk_threshold_chroma;
+	dml_uint_t meta_chunks_per_row_ub_chroma;
+	dml_uint_t dpte_group_width_luma;
+	dml_uint_t dpte_groups_per_row_luma_ub;
+	dml_uint_t dpte_group_width_chroma;
+	dml_uint_t dpte_groups_per_row_chroma_ub;
+
+	for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+		DST_Y_PER_PTE_ROW_NOM_L[k] = dpte_row_height[k] / VRatio[k];
+		if (BytePerPixelC[k] == 0) {
+			DST_Y_PER_PTE_ROW_NOM_C[k] = 0;
+		} else {
+			DST_Y_PER_PTE_ROW_NOM_C[k] = dpte_row_height_chroma[k] / VRatioChroma[k];
+		}
+		DST_Y_PER_META_ROW_NOM_L[k] = meta_row_height[k] / VRatio[k];
+		if (BytePerPixelC[k] == 0) {
+			DST_Y_PER_META_ROW_NOM_C[k] = 0;
+		} else {
+			DST_Y_PER_META_ROW_NOM_C[k] = meta_row_height_chroma[k] / VRatioChroma[k];
+		}
+	}
+
+	for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+		if (DCCEnable[k] == true) {
+			meta_chunk_width = MetaChunkSize * 1024 * 256 / BytePerPixelY[k] / meta_row_height[k];
+			min_meta_chunk_width = MinMetaChunkSizeBytes * 256 / BytePerPixelY[k] / meta_row_height[k];
+			meta_chunk_per_row_int = meta_row_width[k] / meta_chunk_width;
+			meta_row_remainder = meta_row_width[k] % meta_chunk_width;
+			if (!dml_is_vertical_rotation(SourceScan[k])) {
+				meta_chunk_threshold = 2 * min_meta_chunk_width - meta_req_width[k];
+			} else {
+				meta_chunk_threshold = 2 * min_meta_chunk_width - meta_req_height[k];
+			}
+			if (meta_row_remainder <= meta_chunk_threshold) {
+				meta_chunks_per_row_ub = meta_chunk_per_row_int + 1;
+			} else {
+				meta_chunks_per_row_ub = meta_chunk_per_row_int + 2;
+			}
+			TimePerMetaChunkNominal[k] = meta_row_height[k] / VRatio[k] * HTotal[k] / PixelClock[k] / meta_chunks_per_row_ub;
+			TimePerMetaChunkVBlank[k] = DestinationLinesToRequestRowInVBlank[k] * HTotal[k] / PixelClock[k] / meta_chunks_per_row_ub;
+			TimePerMetaChunkFlip[k] = DestinationLinesToRequestRowInImmediateFlip[k] * HTotal[k] / PixelClock[k] / meta_chunks_per_row_ub;
+			if (BytePerPixelC[k] == 0) {
+				TimePerChromaMetaChunkNominal[k] = 0;
+				TimePerChromaMetaChunkVBlank[k] = 0;
+				TimePerChromaMetaChunkFlip[k] = 0;
+			} else {
+				meta_chunk_width_chroma = MetaChunkSize * 1024 * 256 / BytePerPixelC[k] / meta_row_height_chroma[k];
+				min_meta_chunk_width_chroma = MinMetaChunkSizeBytes * 256 / BytePerPixelC[k] / meta_row_height_chroma[k];
+				meta_chunk_per_row_int_chroma = (dml_uint_t)((dml_float_t) meta_row_width_chroma[k] / meta_chunk_width_chroma);
+				meta_row_remainder_chroma = meta_row_width_chroma[k] % meta_chunk_width_chroma;
+				if (!dml_is_vertical_rotation(SourceScan[k])) {
+					meta_chunk_threshold_chroma = 2 * min_meta_chunk_width_chroma - meta_req_width_chroma[k];
+				} else {
+					meta_chunk_threshold_chroma = 2 * min_meta_chunk_width_chroma - meta_req_height_chroma[k];
+				}
+				if (meta_row_remainder_chroma <= meta_chunk_threshold_chroma) {
+					meta_chunks_per_row_ub_chroma = meta_chunk_per_row_int_chroma + 1;
+				} else {
+					meta_chunks_per_row_ub_chroma = meta_chunk_per_row_int_chroma + 2;
+				}
+				TimePerChromaMetaChunkNominal[k] = meta_row_height_chroma[k] / VRatioChroma[k] * HTotal[k] / PixelClock[k] / meta_chunks_per_row_ub_chroma;
+				TimePerChromaMetaChunkVBlank[k] = DestinationLinesToRequestRowInVBlank[k] * HTotal[k] / PixelClock[k] / meta_chunks_per_row_ub_chroma;
+				TimePerChromaMetaChunkFlip[k] = DestinationLinesToRequestRowInImmediateFlip[k] * HTotal[k] / PixelClock[k] / meta_chunks_per_row_ub_chroma;
+			}
+		} else {
+			TimePerMetaChunkNominal[k] = 0;
+			TimePerMetaChunkVBlank[k] = 0;
+			TimePerMetaChunkFlip[k] = 0;
+			TimePerChromaMetaChunkNominal[k] = 0;
+			TimePerChromaMetaChunkVBlank[k] = 0;
+			TimePerChromaMetaChunkFlip[k] = 0;
+		}
+	}
+
+	for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+		if (GPUVMEnable == true) {
+			if (!dml_is_vertical_rotation(SourceScan[k])) {
+				dpte_group_width_luma = (dml_uint_t)((dml_float_t) dpte_group_bytes[k] / (dml_float_t) PTERequestSizeY[k] * PixelPTEReqWidthY[k]);
+			} else {
+				dpte_group_width_luma = (dml_uint_t)((dml_float_t) dpte_group_bytes[k] / (dml_float_t) PTERequestSizeY[k] * PixelPTEReqHeightY[k]);
+			}
+
+			if (use_one_row_for_frame[k]) {
+				dpte_groups_per_row_luma_ub = (dml_uint_t)(dml_ceil((dml_float_t) dpte_row_width_luma_ub[k] / (dml_float_t) dpte_group_width_luma / 2.0, 1.0));
+			} else {
+				dpte_groups_per_row_luma_ub = (dml_uint_t)(dml_ceil((dml_float_t) dpte_row_width_luma_ub[k] / (dml_float_t) dpte_group_width_luma, 1.0));
+			}
+
+			dml_print("DML::%s: k=%u, use_one_row_for_frame = %u\n", __func__, k, use_one_row_for_frame[k]);
+			dml_print("DML::%s: k=%u, dpte_group_bytes = %u\n", __func__, k, dpte_group_bytes[k]);
+			dml_print("DML::%s: k=%u, PTERequestSizeY = %u\n", __func__, k, PTERequestSizeY[k]);
+			dml_print("DML::%s: k=%u, PixelPTEReqWidthY = %u\n", __func__, k, PixelPTEReqWidthY[k]);
+			dml_print("DML::%s: k=%u, PixelPTEReqHeightY = %u\n", __func__, k, PixelPTEReqHeightY[k]);
+			dml_print("DML::%s: k=%u, dpte_row_width_luma_ub = %u\n", __func__, k, dpte_row_width_luma_ub[k]);
+			dml_print("DML::%s: k=%u, dpte_group_width_luma = %u\n", __func__, k, dpte_group_width_luma);
+			dml_print("DML::%s: k=%u, dpte_groups_per_row_luma_ub = %u\n", __func__, k, dpte_groups_per_row_luma_ub);
+
+			time_per_pte_group_nom_luma[k] = DST_Y_PER_PTE_ROW_NOM_L[k] * HTotal[k] / PixelClock[k] / dpte_groups_per_row_luma_ub;
+			time_per_pte_group_vblank_luma[k] = DestinationLinesToRequestRowInVBlank[k] * HTotal[k] / PixelClock[k] / dpte_groups_per_row_luma_ub;
+			time_per_pte_group_flip_luma[k] = DestinationLinesToRequestRowInImmediateFlip[k] * HTotal[k] / PixelClock[k] / dpte_groups_per_row_luma_ub;
+			if (BytePerPixelC[k] == 0) {
+				time_per_pte_group_nom_chroma[k] = 0;
+				time_per_pte_group_vblank_chroma[k] = 0;
+				time_per_pte_group_flip_chroma[k] = 0;
+			} else {
+				if (!dml_is_vertical_rotation(SourceScan[k])) {
+					dpte_group_width_chroma = (dml_uint_t)((dml_float_t) dpte_group_bytes[k] / (dml_float_t) PTERequestSizeC[k] * PixelPTEReqWidthC[k]);
+				} else {
+					dpte_group_width_chroma = (dml_uint_t)((dml_float_t) dpte_group_bytes[k] / (dml_float_t) PTERequestSizeC[k] * PixelPTEReqHeightC[k]);
+				}
+
+				if (use_one_row_for_frame[k]) {
+					dpte_groups_per_row_chroma_ub = (dml_uint_t)(dml_ceil((dml_float_t) dpte_row_width_chroma_ub[k] / (dml_float_t) dpte_group_width_chroma / 2.0, 1.0));
+				} else {
+					dpte_groups_per_row_chroma_ub = (dml_uint_t)(dml_ceil((dml_float_t) dpte_row_width_chroma_ub[k] / (dml_float_t) dpte_group_width_chroma, 1.0));
+				}
+				dml_print("DML::%s: k=%u, dpte_row_width_chroma_ub = %u\n", __func__, k, dpte_row_width_chroma_ub[k]);
+				dml_print("DML::%s: k=%u, dpte_group_width_chroma = %u\n", __func__, k, dpte_group_width_chroma);
+				dml_print("DML::%s: k=%u, dpte_groups_per_row_chroma_ub = %u\n", __func__, k, dpte_groups_per_row_chroma_ub);
+
+				time_per_pte_group_nom_chroma[k] = DST_Y_PER_PTE_ROW_NOM_C[k] * HTotal[k] / PixelClock[k] / dpte_groups_per_row_chroma_ub;
+				time_per_pte_group_vblank_chroma[k] = DestinationLinesToRequestRowInVBlank[k] * HTotal[k] / PixelClock[k] / dpte_groups_per_row_chroma_ub;
+				time_per_pte_group_flip_chroma[k] = DestinationLinesToRequestRowInImmediateFlip[k] * HTotal[k] / PixelClock[k] / dpte_groups_per_row_chroma_ub;
+			}
+		} else {
+			time_per_pte_group_nom_luma[k] = 0;
+			time_per_pte_group_vblank_luma[k] = 0;
+			time_per_pte_group_flip_luma[k] = 0;
+			time_per_pte_group_nom_chroma[k] = 0;
+			time_per_pte_group_vblank_chroma[k] = 0;
+			time_per_pte_group_flip_chroma[k] = 0;
+		}
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%u, DestinationLinesToRequestRowInVBlank = %f\n", __func__, k, DestinationLinesToRequestRowInVBlank[k]);
+		dml_print("DML::%s: k=%u, DestinationLinesToRequestRowInImmediateFlip = %f\n", __func__, k, DestinationLinesToRequestRowInImmediateFlip[k]);
+
+		dml_print("DML::%s: k=%u, DST_Y_PER_PTE_ROW_NOM_L = %f\n", __func__, k, DST_Y_PER_PTE_ROW_NOM_L[k]);
+		dml_print("DML::%s: k=%u, DST_Y_PER_PTE_ROW_NOM_C = %f\n", __func__, k, DST_Y_PER_PTE_ROW_NOM_C[k]);
+		dml_print("DML::%s: k=%u, DST_Y_PER_META_ROW_NOM_L = %f\n", __func__, k, DST_Y_PER_META_ROW_NOM_L[k]);
+		dml_print("DML::%s: k=%u, DST_Y_PER_META_ROW_NOM_C = %f\n", __func__, k, DST_Y_PER_META_ROW_NOM_C[k]);
+		dml_print("DML::%s: k=%u, TimePerMetaChunkNominal = %f\n", __func__, k, TimePerMetaChunkNominal[k]);
+		dml_print("DML::%s: k=%u, TimePerMetaChunkVBlank = %f\n", __func__, k, TimePerMetaChunkVBlank[k]);
+		dml_print("DML::%s: k=%u, TimePerMetaChunkFlip = %f\n", __func__, k, TimePerMetaChunkFlip[k]);
+		dml_print("DML::%s: k=%u, TimePerChromaMetaChunkNominal = %f\n", __func__, k, TimePerChromaMetaChunkNominal[k]);
+		dml_print("DML::%s: k=%u, TimePerChromaMetaChunkVBlank = %f\n", __func__, k, TimePerChromaMetaChunkVBlank[k]);
+		dml_print("DML::%s: k=%u, TimePerChromaMetaChunkFlip = %f\n", __func__, k, TimePerChromaMetaChunkFlip[k]);
+		dml_print("DML::%s: k=%u, time_per_pte_group_nom_luma = %f\n", __func__, k, time_per_pte_group_nom_luma[k]);
+		dml_print("DML::%s: k=%u, time_per_pte_group_vblank_luma = %f\n", __func__, k, time_per_pte_group_vblank_luma[k]);
+		dml_print("DML::%s: k=%u, time_per_pte_group_flip_luma = %f\n", __func__, k, time_per_pte_group_flip_luma[k]);
+		dml_print("DML::%s: k=%u, time_per_pte_group_nom_chroma = %f\n", __func__, k, time_per_pte_group_nom_chroma[k]);
+		dml_print("DML::%s: k=%u, time_per_pte_group_vblank_chroma = %f\n", __func__, k, time_per_pte_group_vblank_chroma[k]);
+		dml_print("DML::%s: k=%u, time_per_pte_group_flip_chroma   = %f\n", __func__, k, time_per_pte_group_flip_chroma[k]);
+#endif
+	}
+} // CalculateMetaAndPTETimes
+
+static void CalculateVMGroupAndRequestTimes(
+		dml_uint_t NumberOfActiveSurfaces,
+		dml_bool_t GPUVMEnable,
+		dml_uint_t GPUVMMaxPageTableLevels,
+		dml_uint_t HTotal[],
+		dml_uint_t BytePerPixelC[],
+		dml_float_t DestinationLinesToRequestVMInVBlank[],
+		dml_float_t DestinationLinesToRequestVMInImmediateFlip[],
+		dml_bool_t DCCEnable[],
+		dml_float_t PixelClock[],
+		dml_uint_t dpte_row_width_luma_ub[],
+		dml_uint_t dpte_row_width_chroma_ub[],
+		dml_uint_t vm_group_bytes[],
+		dml_uint_t dpde0_bytes_per_frame_ub_l[],
+		dml_uint_t dpde0_bytes_per_frame_ub_c[],
+		dml_uint_t meta_pte_bytes_per_frame_ub_l[],
+		dml_uint_t meta_pte_bytes_per_frame_ub_c[],
+
+		// Output
+		dml_float_t TimePerVMGroupVBlank[],
+		dml_float_t TimePerVMGroupFlip[],
+		dml_float_t TimePerVMRequestVBlank[],
+		dml_float_t TimePerVMRequestFlip[])
+{
+	dml_uint_t num_group_per_lower_vm_stage;
+	dml_uint_t num_req_per_lower_vm_stage;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: NumberOfActiveSurfaces = %u\n", __func__, NumberOfActiveSurfaces);
+	dml_print("DML::%s: GPUVMEnable = %u\n", __func__, GPUVMEnable);
+#endif
+	for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%u, DCCEnable = %u\n", __func__, k, DCCEnable[k]);
+		dml_print("DML::%s: k=%u, vm_group_bytes = %u\n", __func__, k, vm_group_bytes[k]);
+		dml_print("DML::%s: k=%u, dpde0_bytes_per_frame_ub_l = %u\n", __func__, k, dpde0_bytes_per_frame_ub_l[k]);
+		dml_print("DML::%s: k=%u, dpde0_bytes_per_frame_ub_c = %u\n", __func__, k, dpde0_bytes_per_frame_ub_c[k]);
+		dml_print("DML::%s: k=%u, meta_pte_bytes_per_frame_ub_l = %u\n", __func__, k, meta_pte_bytes_per_frame_ub_l[k]);
+		dml_print("DML::%s: k=%u, meta_pte_bytes_per_frame_ub_c = %u\n", __func__, k, meta_pte_bytes_per_frame_ub_c[k]);
+#endif
+
+		if (GPUVMEnable == true && (DCCEnable[k] == true || GPUVMMaxPageTableLevels > 1)) {
+			if (DCCEnable[k] == false) {
+				if (BytePerPixelC[k] > 0) {
+					num_group_per_lower_vm_stage = (dml_uint_t) (dml_ceil((dml_float_t) dpde0_bytes_per_frame_ub_l[k] / (dml_float_t) vm_group_bytes[k], 1.0) +
+													dml_ceil((dml_float_t) dpde0_bytes_per_frame_ub_c[k] / (dml_float_t) vm_group_bytes[k], 1.0));
+				} else {
+					num_group_per_lower_vm_stage = (dml_uint_t) (dml_ceil((dml_float_t) dpde0_bytes_per_frame_ub_l[k] / (dml_float_t) vm_group_bytes[k], 1.0));
+				}
+			} else {
+				if (GPUVMMaxPageTableLevels == 1) {
+					if (BytePerPixelC[k] > 0) {
+						num_group_per_lower_vm_stage = (dml_uint_t)(dml_ceil((dml_float_t) (meta_pte_bytes_per_frame_ub_l[k]) / (dml_float_t) (vm_group_bytes[k]), 1.0) +
+																	dml_ceil((dml_float_t) (meta_pte_bytes_per_frame_ub_c[k]) / (dml_float_t) (vm_group_bytes[k]), 1.0));
+					} else {
+						num_group_per_lower_vm_stage = (dml_uint_t)(dml_ceil((dml_float_t) (meta_pte_bytes_per_frame_ub_l[k]) / (dml_float_t) (vm_group_bytes[k]), 1.0));
+					}
+				} else {
+					if (BytePerPixelC[k] > 0) {
+						num_group_per_lower_vm_stage = (dml_uint_t)(2.0 + dml_ceil((dml_float_t) (dpde0_bytes_per_frame_ub_l[k]) / (dml_float_t) (vm_group_bytes[k]), 1) +
+														dml_ceil((dml_float_t) (dpde0_bytes_per_frame_ub_c[k]) / (dml_float_t) (vm_group_bytes[k]), 1) +
+														dml_ceil((dml_float_t) (meta_pte_bytes_per_frame_ub_l[k]) / (dml_float_t) (vm_group_bytes[k]), 1) +
+														dml_ceil((dml_float_t) (meta_pte_bytes_per_frame_ub_c[k]) / (dml_float_t) (vm_group_bytes[k]), 1));
+					} else {
+						num_group_per_lower_vm_stage = (dml_uint_t)(1.0 + dml_ceil((dml_float_t) (dpde0_bytes_per_frame_ub_l[k]) / (dml_float_t) (vm_group_bytes[k]), 1) +
+																	dml_ceil((dml_float_t) (meta_pte_bytes_per_frame_ub_l[k]) / (dml_float_t) (vm_group_bytes[k]), 1));
+					}
+				}
+			}
+
+			if (DCCEnable[k] == false) {
+				if (BytePerPixelC[k] > 0) {
+					num_req_per_lower_vm_stage = dpde0_bytes_per_frame_ub_l[k] / 64 + dpde0_bytes_per_frame_ub_c[k] / 64;
+				} else {
+					num_req_per_lower_vm_stage = dpde0_bytes_per_frame_ub_l[k] / 64;
+				}
+			} else {
+				if (GPUVMMaxPageTableLevels == 1) {
+					if (BytePerPixelC[k] > 0) {
+						num_req_per_lower_vm_stage = meta_pte_bytes_per_frame_ub_l[k] / 64 + meta_pte_bytes_per_frame_ub_c[k] / 64;
+					} else {
+						num_req_per_lower_vm_stage = meta_pte_bytes_per_frame_ub_l[k] / 64;
+					}
+				} else {
+					if (BytePerPixelC[k] > 0) {
+						num_req_per_lower_vm_stage = dpde0_bytes_per_frame_ub_l[k] / 64 + dpde0_bytes_per_frame_ub_c[k] / 64 + meta_pte_bytes_per_frame_ub_l[k] / 64 + meta_pte_bytes_per_frame_ub_c[k] / 64;
+					} else {
+						num_req_per_lower_vm_stage = dpde0_bytes_per_frame_ub_l[k] / 64 + meta_pte_bytes_per_frame_ub_l[k] / 64;
+					}
+				}
+			}
+
+			TimePerVMGroupVBlank[k] = DestinationLinesToRequestVMInVBlank[k] * HTotal[k] / PixelClock[k] / num_group_per_lower_vm_stage;
+			TimePerVMGroupFlip[k] = DestinationLinesToRequestVMInImmediateFlip[k] * HTotal[k] / PixelClock[k] / num_group_per_lower_vm_stage;
+			TimePerVMRequestVBlank[k] = DestinationLinesToRequestVMInVBlank[k] * HTotal[k] / PixelClock[k] / num_req_per_lower_vm_stage;
+			TimePerVMRequestFlip[k] = DestinationLinesToRequestVMInImmediateFlip[k] * HTotal[k] / PixelClock[k] / num_req_per_lower_vm_stage;
+
+			if (GPUVMMaxPageTableLevels > 2) {
+				TimePerVMGroupVBlank[k] = TimePerVMGroupVBlank[k] / 2;
+				TimePerVMGroupFlip[k] = TimePerVMGroupFlip[k] / 2;
+				TimePerVMRequestVBlank[k] = TimePerVMRequestVBlank[k] / 2;
+				TimePerVMRequestFlip[k]    = TimePerVMRequestFlip[k] / 2;
+			}
+
+		} else {
+			TimePerVMGroupVBlank[k] = 0;
+			TimePerVMGroupFlip[k] = 0;
+			TimePerVMRequestVBlank[k] = 0;
+			TimePerVMRequestFlip[k] = 0;
+		}
+
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: k=%u, TimePerVMGroupVBlank = %f\n", __func__, k, TimePerVMGroupVBlank[k]);
+			dml_print("DML::%s: k=%u, TimePerVMGroupFlip = %f\n", __func__, k, TimePerVMGroupFlip[k]);
+			dml_print("DML::%s: k=%u, TimePerVMRequestVBlank = %f\n", __func__, k, TimePerVMRequestVBlank[k]);
+			dml_print("DML::%s: k=%u, TimePerVMRequestFlip = %f\n", __func__, k, TimePerVMRequestFlip[k]);
+#endif
+	}
+} // CalculateVMGroupAndRequestTimes
+
+static void CalculateStutterEfficiency(struct display_mode_lib_scratch_st *scratch,
+		struct CalculateStutterEfficiency_params_st *p)
+{
+	dml_float_t DETBufferingTimeY = 0;
+	dml_float_t SwathWidthYCriticalSurface = 0;
+	dml_float_t SwathHeightYCriticalSurface = 0;
+	dml_float_t VActiveTimeCriticalSurface = 0;
+	dml_float_t FrameTimeCriticalSurface = 0;
+	dml_uint_t BytePerPixelYCriticalSurface = 0;
+	dml_float_t LinesToFinishSwathTransferStutterCriticalSurface = 0;
+	dml_uint_t DETBufferSizeYCriticalSurface = 0;
+	dml_float_t MinTTUVBlankCriticalSurface = 0;
+	dml_uint_t BlockWidth256BytesYCriticalSurface = 0;
+	dml_bool_t SinglePlaneCriticalSurface = 0;
+	dml_bool_t SinglePipeCriticalSurface = 0;
+	dml_float_t TotalCompressedReadBandwidth = 0;
+	dml_float_t TotalRowReadBandwidth = 0;
+	dml_float_t AverageDCCCompressionRate = 0;
+	dml_float_t EffectiveCompressedBufferSize = 0;
+	dml_float_t PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer = 0;
+	dml_float_t StutterBurstTime = 0;
+	dml_uint_t TotalActiveWriteback = 0;
+	dml_float_t LinesInDETY = 0;
+	dml_float_t LinesInDETYRoundedDownToSwath = 0;
+	dml_float_t MaximumEffectiveCompressionLuma = 0;
+	dml_float_t MaximumEffectiveCompressionChroma = 0;
+	dml_float_t TotalZeroSizeRequestReadBandwidth = 0;
+	dml_float_t TotalZeroSizeCompressedReadBandwidth = 0;
+	dml_float_t AverageDCCZeroSizeFraction = 0;
+	dml_float_t AverageZeroSizeCompressionRate = 0;
+
+	dml_bool_t FoundCriticalSurface = false;
+
+	dml_uint_t TotalNumberOfActiveOTG = 0;
+	dml_float_t SinglePixelClock = 0;
+	dml_uint_t SingleHTotal = 0;
+	dml_uint_t SingleVTotal = 0;
+	dml_bool_t SameTiming = true;
+
+	dml_float_t LastStutterPeriod = 0.0;
+	dml_float_t LastZ8StutterPeriod = 0.0;
+
+	dml_uint_t SwathSizeCriticalSurface;
+	dml_uint_t LastChunkOfSwathSize;
+	dml_uint_t MissingPartOfLastSwathOfDETSize;
+
+	TotalZeroSizeRequestReadBandwidth = 0;
+	TotalZeroSizeCompressedReadBandwidth = 0;
+	TotalRowReadBandwidth = 0;
+	TotalCompressedReadBandwidth = 0;
+
+	for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+	if (p->UseMALLForPStateChange[k] != dml_use_mall_pstate_change_phantom_pipe) {
+		if (p->DCCEnable[k] == true) {
+			if ((dml_is_vertical_rotation(p->SourceScan[k]) && p->BlockWidth256BytesY[k] > p->SwathHeightY[k]) || (!dml_is_vertical_rotation(p->SourceScan[k]) && p->BlockHeight256BytesY[k] > p->SwathHeightY[k]) || p->DCCYMaxUncompressedBlock[k] < 256) {
+				MaximumEffectiveCompressionLuma = 2;
+			} else {
+				MaximumEffectiveCompressionLuma = 4;
+			}
+			TotalCompressedReadBandwidth = TotalCompressedReadBandwidth + p->ReadBandwidthSurfaceLuma[k] / dml_min(p->NetDCCRateLuma[k], MaximumEffectiveCompressionLuma);
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: k=%u, ReadBandwidthSurfaceLuma = %f\n", __func__, k, p->ReadBandwidthSurfaceLuma[k]);
+			dml_print("DML::%s: k=%u, NetDCCRateLuma = %f\n", __func__, k, p->NetDCCRateLuma[k]);
+			dml_print("DML::%s: k=%u, MaximumEffectiveCompressionLuma = %f\n", __func__, k, MaximumEffectiveCompressionLuma);
+#endif
+			TotalZeroSizeRequestReadBandwidth = TotalZeroSizeRequestReadBandwidth + p->ReadBandwidthSurfaceLuma[k] * p->DCCFractionOfZeroSizeRequestsLuma[k];
+			TotalZeroSizeCompressedReadBandwidth = TotalZeroSizeCompressedReadBandwidth + p->ReadBandwidthSurfaceLuma[k] * p->DCCFractionOfZeroSizeRequestsLuma[k] / MaximumEffectiveCompressionLuma;
+
+			if (p->ReadBandwidthSurfaceChroma[k] > 0) {
+				if ((dml_is_vertical_rotation(p->SourceScan[k]) && p->BlockWidth256BytesC[k] > p->SwathHeightC[k]) || (!dml_is_vertical_rotation(p->SourceScan[k]) && p->BlockHeight256BytesC[k] > p->SwathHeightC[k]) || p->DCCCMaxUncompressedBlock[k] < 256) {
+					MaximumEffectiveCompressionChroma = 2;
+				} else {
+					MaximumEffectiveCompressionChroma = 4;
+				}
+				TotalCompressedReadBandwidth = TotalCompressedReadBandwidth + p->ReadBandwidthSurfaceChroma[k] / dml_min(p->NetDCCRateChroma[k], MaximumEffectiveCompressionChroma);
+#ifdef __DML_VBA_DEBUG__
+				dml_print("DML::%s: k=%u, ReadBandwidthSurfaceChroma = %f\n", __func__, k, p->ReadBandwidthSurfaceChroma[k]);
+				dml_print("DML::%s: k=%u, NetDCCRateChroma = %f\n", __func__, k, p->NetDCCRateChroma[k]);
+				dml_print("DML::%s: k=%u, MaximumEffectiveCompressionChroma = %f\n", __func__, k, MaximumEffectiveCompressionChroma);
+#endif
+				TotalZeroSizeRequestReadBandwidth = TotalZeroSizeRequestReadBandwidth + p->ReadBandwidthSurfaceChroma[k] * p->DCCFractionOfZeroSizeRequestsChroma[k];
+				TotalZeroSizeCompressedReadBandwidth = TotalZeroSizeCompressedReadBandwidth + p->ReadBandwidthSurfaceChroma[k] * p->DCCFractionOfZeroSizeRequestsChroma[k] / MaximumEffectiveCompressionChroma;
+			}
+		} else {
+			TotalCompressedReadBandwidth = TotalCompressedReadBandwidth + p->ReadBandwidthSurfaceLuma[k] + p->ReadBandwidthSurfaceChroma[k];
+		}
+		TotalRowReadBandwidth = TotalRowReadBandwidth + p->DPPPerSurface[k] * (p->meta_row_bw[k] + p->dpte_row_bw[k]);
+	}
+	}
+
+	AverageDCCCompressionRate = p->TotalDataReadBandwidth / TotalCompressedReadBandwidth;
+	AverageDCCZeroSizeFraction = TotalZeroSizeRequestReadBandwidth / p->TotalDataReadBandwidth;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: UnboundedRequestEnabled = %u\n", __func__, p->UnboundedRequestEnabled);
+	dml_print("DML::%s: TotalCompressedReadBandwidth = %f\n", __func__, TotalCompressedReadBandwidth);
+	dml_print("DML::%s: TotalZeroSizeRequestReadBandwidth = %f\n", __func__, TotalZeroSizeRequestReadBandwidth);
+	dml_print("DML::%s: TotalZeroSizeCompressedReadBandwidth = %f\n", __func__, TotalZeroSizeCompressedReadBandwidth);
+	dml_print("DML::%s: MaximumEffectiveCompressionLuma = %f\n", __func__, MaximumEffectiveCompressionLuma);
+	dml_print("DML::%s: MaximumEffectiveCompressionChroma = %f\n", __func__, MaximumEffectiveCompressionChroma);
+	dml_print("DML::%s: AverageDCCCompressionRate = %f\n", __func__, AverageDCCCompressionRate);
+	dml_print("DML::%s: AverageDCCZeroSizeFraction = %f\n", __func__, AverageDCCZeroSizeFraction);
+	dml_print("DML::%s: CompbufReservedSpace64B = %u\n", __func__, p->CompbufReservedSpace64B);
+	dml_print("DML::%s: CompbufReservedSpaceZs = %u\n", __func__, p->CompbufReservedSpaceZs);
+	dml_print("DML::%s: CompressedBufferSizeInkByte = %u\n", __func__, p->CompressedBufferSizeInkByte);
+#endif
+	if (AverageDCCZeroSizeFraction == 1) {
+		AverageZeroSizeCompressionRate = TotalZeroSizeRequestReadBandwidth / TotalZeroSizeCompressedReadBandwidth;
+		EffectiveCompressedBufferSize = (dml_float_t)p->MetaFIFOSizeInKEntries * 1024 * 64 * AverageZeroSizeCompressionRate + ((dml_float_t)p->ZeroSizeBufferEntries - p->CompbufReservedSpaceZs) * 64 * AverageZeroSizeCompressionRate;
+	} 	else if (AverageDCCZeroSizeFraction > 0) {
+		AverageZeroSizeCompressionRate = TotalZeroSizeRequestReadBandwidth / TotalZeroSizeCompressedReadBandwidth;
+		EffectiveCompressedBufferSize = dml_min((dml_float_t)p->CompressedBufferSizeInkByte * 1024 * AverageDCCCompressionRate,
+											(dml_float_t)p->MetaFIFOSizeInKEntries * 1024 * 64 / (AverageDCCZeroSizeFraction / AverageZeroSizeCompressionRate + 1 / AverageDCCCompressionRate)) +
+										dml_min(((dml_float_t)p->ROBBufferSizeInKByte * 1024 - p->CompbufReservedSpace64B * 64) * AverageDCCCompressionRate,
+											((dml_float_t)p->ZeroSizeBufferEntries - p->CompbufReservedSpaceZs) * 64 / (AverageDCCZeroSizeFraction / AverageZeroSizeCompressionRate));
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: min 1 = %f\n", __func__, p->CompressedBufferSizeInkByte * 1024 * AverageDCCCompressionRate);
+		dml_print("DML::%s: min 2 = %f\n", __func__, p->MetaFIFOSizeInKEntries * 1024 * 64 / (AverageDCCZeroSizeFraction / AverageZeroSizeCompressionRate + 1 / AverageDCCCompressionRate));
+		dml_print("DML::%s: min 3 = %f\n", __func__, (p->ROBBufferSizeInKByte * 1024 - p->CompbufReservedSpace64B * 64) * AverageDCCCompressionRate);
+		dml_print("DML::%s: min 4 = %f\n", __func__, (p->ZeroSizeBufferEntries - p->CompbufReservedSpaceZs) * 64 / (AverageDCCZeroSizeFraction / AverageZeroSizeCompressionRate));
+#endif
+	} else {
+		EffectiveCompressedBufferSize = dml_min((dml_float_t)p->CompressedBufferSizeInkByte * 1024 * AverageDCCCompressionRate,
+												(dml_float_t)p->MetaFIFOSizeInKEntries * 1024 * 64 * AverageDCCCompressionRate) +
+												((dml_float_t)p->ROBBufferSizeInKByte * 1024 - p->CompbufReservedSpace64B * 64) * AverageDCCCompressionRate;
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: min 1 = %f\n", __func__, p->CompressedBufferSizeInkByte * 1024 * AverageDCCCompressionRate);
+		dml_print("DML::%s: min 2 = %f\n", __func__, p->MetaFIFOSizeInKEntries * 1024 * 64 * AverageDCCCompressionRate);
+#endif
+	}
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: MetaFIFOSizeInKEntries = %u\n", __func__, p->MetaFIFOSizeInKEntries);
+	dml_print("DML::%s: AverageZeroSizeCompressionRate = %f\n", __func__, AverageZeroSizeCompressionRate);
+	dml_print("DML::%s: EffectiveCompressedBufferSize = %f\n", __func__, EffectiveCompressedBufferSize);
+#endif
+
+	*p->StutterPeriod = 0;
+
+	for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+	if (p->UseMALLForPStateChange[k] != dml_use_mall_pstate_change_phantom_pipe) {
+		LinesInDETY = ((dml_float_t)p->DETBufferSizeY[k] + (p->UnboundedRequestEnabled == true ? EffectiveCompressedBufferSize : 0) * p->ReadBandwidthSurfaceLuma[k] / p->TotalDataReadBandwidth) / p->BytePerPixelDETY[k] / p->SwathWidthY[k];
+		LinesInDETYRoundedDownToSwath = dml_floor(LinesInDETY, p->SwathHeightY[k]);
+		DETBufferingTimeY = LinesInDETYRoundedDownToSwath * ((dml_float_t)p->HTotal[k] / p->PixelClock[k]) / p->VRatio[k];
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%u, DETBufferSizeY = %u\n", __func__, k, p->DETBufferSizeY[k]);
+		dml_print("DML::%s: k=%u, BytePerPixelDETY = %f\n", __func__, k, p->BytePerPixelDETY[k]);
+		dml_print("DML::%s: k=%u, SwathWidthY = %u\n", __func__, k, p->SwathWidthY[k]);
+		dml_print("DML::%s: k=%u, ReadBandwidthSurfaceLuma = %f\n", __func__, k, p->ReadBandwidthSurfaceLuma[k]);
+		dml_print("DML::%s: k=%u, TotalDataReadBandwidth = %f\n", __func__, k, p->TotalDataReadBandwidth);
+		dml_print("DML::%s: k=%u, LinesInDETY = %f\n", __func__, k, LinesInDETY);
+		dml_print("DML::%s: k=%u, LinesInDETYRoundedDownToSwath = %f\n", __func__, k, LinesInDETYRoundedDownToSwath);
+		dml_print("DML::%s: k=%u, HTotal = %u\n", __func__, k, p->HTotal[k]);
+		dml_print("DML::%s: k=%u, PixelClock = %f\n", __func__, k, p->PixelClock[k]);
+		dml_print("DML::%s: k=%u, VRatio = %f\n", __func__, k, p->VRatio[k]);
+		dml_print("DML::%s: k=%u, DETBufferingTimeY = %f\n", __func__, k, DETBufferingTimeY);
+		dml_print("DML::%s: k=%u,PixelClock = %f\n", __func__, k, p->PixelClock[k]);
+#endif
+
+		if (!FoundCriticalSurface || DETBufferingTimeY < *p->StutterPeriod) {
+			dml_bool_t isInterlaceTiming = p->Interlace[k] && !p->ProgressiveToInterlaceUnitInOPP;
+
+			FoundCriticalSurface = true;
+			*p->StutterPeriod = DETBufferingTimeY;
+			FrameTimeCriticalSurface = (isInterlaceTiming ? dml_floor((dml_float_t)p->VTotal[k]/2.0, 1.0) : p->VTotal[k]) * (dml_float_t)p->HTotal[k] / p->PixelClock[k];
+			VActiveTimeCriticalSurface = (isInterlaceTiming ? dml_floor((dml_float_t)p->VActive[k]/2.0, 1.0) : p->VActive[k]) * (dml_float_t)p->HTotal[k] / p->PixelClock[k];
+			BytePerPixelYCriticalSurface = p->BytePerPixelY[k];
+			SwathWidthYCriticalSurface = p->SwathWidthY[k];
+			SwathHeightYCriticalSurface = p->SwathHeightY[k];
+			BlockWidth256BytesYCriticalSurface = p->BlockWidth256BytesY[k];
+			LinesToFinishSwathTransferStutterCriticalSurface = p->SwathHeightY[k] - (LinesInDETY - LinesInDETYRoundedDownToSwath);
+			DETBufferSizeYCriticalSurface = p->DETBufferSizeY[k];
+			MinTTUVBlankCriticalSurface = p->MinTTUVBlank[k];
+			SinglePlaneCriticalSurface = (p->ReadBandwidthSurfaceChroma[k] == 0);
+			SinglePipeCriticalSurface = (p->DPPPerSurface[k] == 1);
+
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: k=%u, FoundCriticalSurface = %u\n", __func__, k, FoundCriticalSurface);
+			dml_print("DML::%s: k=%u, StutterPeriod = %f\n", __func__, k, *p->StutterPeriod);
+			dml_print("DML::%s: k=%u, MinTTUVBlankCriticalSurface = %f\n", __func__, k, MinTTUVBlankCriticalSurface);
+			dml_print("DML::%s: k=%u, FrameTimeCriticalSurface = %f\n", __func__, k, FrameTimeCriticalSurface);
+			dml_print("DML::%s: k=%u, VActiveTimeCriticalSurface = %f\n", __func__, k, VActiveTimeCriticalSurface);
+			dml_print("DML::%s: k=%u, BytePerPixelYCriticalSurface = %u\n", __func__, k, BytePerPixelYCriticalSurface);
+			dml_print("DML::%s: k=%u, SwathWidthYCriticalSurface = %f\n", __func__, k, SwathWidthYCriticalSurface);
+			dml_print("DML::%s: k=%u, SwathHeightYCriticalSurface = %f\n", __func__, k, SwathHeightYCriticalSurface);
+			dml_print("DML::%s: k=%u, BlockWidth256BytesYCriticalSurface = %u\n", __func__, k, BlockWidth256BytesYCriticalSurface);
+			dml_print("DML::%s: k=%u, SinglePlaneCriticalSurface = %u\n", __func__, k, SinglePlaneCriticalSurface);
+			dml_print("DML::%s: k=%u, SinglePipeCriticalSurface = %u\n", __func__, k, SinglePipeCriticalSurface);
+			dml_print("DML::%s: k=%u, LinesToFinishSwathTransferStutterCriticalSurface = %f\n", __func__, k, LinesToFinishSwathTransferStutterCriticalSurface);
+#endif
+		}
+	}
+	}
+
+	PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer = dml_min(*p->StutterPeriod * p->TotalDataReadBandwidth, EffectiveCompressedBufferSize);
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: ROBBufferSizeInKByte = %u\n", __func__, p->ROBBufferSizeInKByte);
+	dml_print("DML::%s: AverageDCCCompressionRate = %f\n", __func__, AverageDCCCompressionRate);
+	dml_print("DML::%s: StutterPeriod * TotalDataReadBandwidth = %f\n", __func__, *p->StutterPeriod * p->TotalDataReadBandwidth);
+	dml_print("DML::%s: ROBBufferSizeInKByte * 1024 * AverageDCCCompressionRate + EffectiveCompressedBufferSize = %f\n", __func__, p->ROBBufferSizeInKByte * 1024 * AverageDCCCompressionRate + EffectiveCompressedBufferSize);
+	dml_print("DML::%s: EffectiveCompressedBufferSize = %f\n", __func__, EffectiveCompressedBufferSize);
+	dml_print("DML::%s: PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer = %f\n", __func__, PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer);
+	dml_print("DML::%s: ReturnBW = %f\n", __func__, p->ReturnBW);
+	dml_print("DML::%s: TotalDataReadBandwidth = %f\n", __func__, p->TotalDataReadBandwidth);
+	dml_print("DML::%s: TotalRowReadBandwidth = %f\n", __func__, TotalRowReadBandwidth);
+	dml_print("DML::%s: DCFCLK = %f\n", __func__, p->DCFCLK);
+#endif
+
+	StutterBurstTime = PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer / AverageDCCCompressionRate / p->ReturnBW + (*p->StutterPeriod * p->TotalDataReadBandwidth - PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer) / (p->DCFCLK * 64) + *p->StutterPeriod * TotalRowReadBandwidth / p->ReturnBW;
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: Part 1 = %f\n", __func__, PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer / AverageDCCCompressionRate / p->ReturnBW);
+	dml_print("DML::%s: StutterPeriod * TotalDataReadBandwidth = %f\n", __func__, (*p->StutterPeriod * p->TotalDataReadBandwidth));
+	dml_print("DML::%s: Part 2 = %f\n", __func__, (*p->StutterPeriod * p->TotalDataReadBandwidth - PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer) / (p->DCFCLK * 64));
+	dml_print("DML::%s: Part 3 = %f\n", __func__, *p->StutterPeriod * TotalRowReadBandwidth / p->ReturnBW);
+	dml_print("DML::%s: StutterBurstTime = %f\n", __func__, StutterBurstTime);
+#endif
+	StutterBurstTime = dml_max(StutterBurstTime, LinesToFinishSwathTransferStutterCriticalSurface * BytePerPixelYCriticalSurface * SwathWidthYCriticalSurface / p->ReturnBW);
+
+	dml_print("DML::%s: Time to finish residue swath=%f\n", __func__, LinesToFinishSwathTransferStutterCriticalSurface * BytePerPixelYCriticalSurface * SwathWidthYCriticalSurface / p->ReturnBW);
+
+	TotalActiveWriteback = 0;
+	for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+		if (p->WritebackEnable[k]) {
+			TotalActiveWriteback = TotalActiveWriteback + 1;
+		}
+	}
+
+	if (TotalActiveWriteback == 0) {
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: SRExitTime = %f\n", __func__, p->SRExitTime);
+		dml_print("DML::%s: SRExitZ8Time = %f\n", __func__, p->SRExitZ8Time);
+		dml_print("DML::%s: StutterBurstTime = %f (final)\n", __func__, StutterBurstTime);
+		dml_print("DML::%s: StutterPeriod = %f\n", __func__, *p->StutterPeriod);
+#endif
+		*p->StutterEfficiencyNotIncludingVBlank = dml_max(0., 1 - (p->SRExitTime + StutterBurstTime) / *p->StutterPeriod) * 100;
+		*p->Z8StutterEfficiencyNotIncludingVBlank = dml_max(0., 1 - (p->SRExitZ8Time + StutterBurstTime) / *p->StutterPeriod) * 100;
+		*p->NumberOfStutterBurstsPerFrame = (*p->StutterEfficiencyNotIncludingVBlank > 0 ? (dml_uint_t)(dml_ceil(VActiveTimeCriticalSurface / *p->StutterPeriod, 1)) : 0);
+		*p->Z8NumberOfStutterBurstsPerFrame = (*p->Z8StutterEfficiencyNotIncludingVBlank > 0 ? (dml_uint_t)(dml_ceil(VActiveTimeCriticalSurface / *p->StutterPeriod, 1)) : 0);
+	} else {
+		*p->StutterEfficiencyNotIncludingVBlank = 0.;
+		*p->Z8StutterEfficiencyNotIncludingVBlank = 0.;
+		*p->NumberOfStutterBurstsPerFrame = 0;
+		*p->Z8NumberOfStutterBurstsPerFrame = 0;
+	}
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: VActiveTimeCriticalSurface = %f\n", __func__, VActiveTimeCriticalSurface);
+	dml_print("DML::%s: StutterEfficiencyNotIncludingVBlank = %f\n", __func__, *p->StutterEfficiencyNotIncludingVBlank);
+	dml_print("DML::%s: Z8StutterEfficiencyNotIncludingVBlank = %f\n", __func__, *p->Z8StutterEfficiencyNotIncludingVBlank);
+	dml_print("DML::%s: NumberOfStutterBurstsPerFrame = %u\n", __func__, *p->NumberOfStutterBurstsPerFrame);
+	dml_print("DML::%s: Z8NumberOfStutterBurstsPerFrame = %u\n", __func__, *p->Z8NumberOfStutterBurstsPerFrame);
+#endif
+
+	for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+		if (p->UseMALLForPStateChange[k] != dml_use_mall_pstate_change_phantom_pipe) {
+			if (p->BlendingAndTiming[k] == k) {
+				if (TotalNumberOfActiveOTG == 0) {
+					SinglePixelClock = p->PixelClock[k];
+					SingleHTotal = p->HTotal[k];
+					SingleVTotal = p->VTotal[k];
+				} else if (SinglePixelClock != p->PixelClock[k] || SingleHTotal != p->HTotal[k] || SingleVTotal != p->VTotal[k]) {
+					SameTiming = false;
+				}
+				TotalNumberOfActiveOTG = TotalNumberOfActiveOTG + 1;
+			}
+		}
+	}
+
+	if (*p->StutterEfficiencyNotIncludingVBlank > 0) {
+		LastStutterPeriod = VActiveTimeCriticalSurface - (*p->NumberOfStutterBurstsPerFrame - 1) * *p->StutterPeriod;
+
+		if ((p->SynchronizeTimingsFinal || TotalNumberOfActiveOTG == 1) && SameTiming &&
+			LastStutterPeriod + MinTTUVBlankCriticalSurface > p->StutterEnterPlusExitWatermark) {
+			*p->StutterEfficiency = (1 - (*p->NumberOfStutterBurstsPerFrame * p->SRExitTime + StutterBurstTime * VActiveTimeCriticalSurface / *p->StutterPeriod) / FrameTimeCriticalSurface) * 100;
+		} else {
+			*p->StutterEfficiency = *p->StutterEfficiencyNotIncludingVBlank;
+		}
+	} else {
+		*p->StutterEfficiency = 0;
+	}
+
+	if (*p->Z8StutterEfficiencyNotIncludingVBlank > 0) {
+		LastZ8StutterPeriod = VActiveTimeCriticalSurface - (*p->NumberOfStutterBurstsPerFrame - 1) * *p->StutterPeriod;
+		if ((p->SynchronizeTimingsFinal || TotalNumberOfActiveOTG == 1) && SameTiming && LastZ8StutterPeriod + MinTTUVBlankCriticalSurface > p->Z8StutterEnterPlusExitWatermark) {
+			*p->Z8StutterEfficiency = (1 - (*p->NumberOfStutterBurstsPerFrame * p->SRExitZ8Time + StutterBurstTime * VActiveTimeCriticalSurface / *p->StutterPeriod) / FrameTimeCriticalSurface) * 100;
+		} else {
+			*p->Z8StutterEfficiency = *p->Z8StutterEfficiencyNotIncludingVBlank;
+		}
+	} else {
+		*p->Z8StutterEfficiency = 0.;
+	}
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: LastZ8StutterPeriod = %f\n", __func__, LastZ8StutterPeriod);
+		dml_print("DML::%s: Z8StutterEnterPlusExitWatermark = %f\n", __func__, p->Z8StutterEnterPlusExitWatermark);
+		dml_print("DML::%s: StutterBurstTime = %f\n", __func__, StutterBurstTime);
+		dml_print("DML::%s: StutterPeriod = %f\n", __func__, *p->StutterPeriod);
+		dml_print("DML::%s: StutterEfficiency = %f\n", __func__, *p->StutterEfficiency);
+		dml_print("DML::%s: Z8StutterEfficiency = %f\n", __func__, *p->Z8StutterEfficiency);
+		dml_print("DML::%s: StutterEfficiencyNotIncludingVBlank = %f\n", __func__, *p->StutterEfficiencyNotIncludingVBlank);
+		dml_print("DML::%s: Z8NumberOfStutterBurstsPerFrame = %u\n", __func__, *p->Z8NumberOfStutterBurstsPerFrame);
+#endif
+
+		SwathSizeCriticalSurface = (dml_uint_t)(BytePerPixelYCriticalSurface * SwathHeightYCriticalSurface * dml_ceil(SwathWidthYCriticalSurface, BlockWidth256BytesYCriticalSurface));
+		LastChunkOfSwathSize = SwathSizeCriticalSurface % (p->PixelChunkSizeInKByte * 1024);
+		MissingPartOfLastSwathOfDETSize = (dml_uint_t)(dml_ceil(DETBufferSizeYCriticalSurface, SwathSizeCriticalSurface) - DETBufferSizeYCriticalSurface);
+
+		*p->DCHUBBUB_ARB_CSTATE_MAX_CAP_MODE = !(!p->UnboundedRequestEnabled && (p->NumberOfActiveSurfaces == 1) && SinglePlaneCriticalSurface && SinglePipeCriticalSurface && (LastChunkOfSwathSize > 0) &&
+			(LastChunkOfSwathSize <= 4096) && (MissingPartOfLastSwathOfDETSize > 0) && (MissingPartOfLastSwathOfDETSize <= LastChunkOfSwathSize));
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: SwathSizeCriticalSurface = %u\n", __func__, SwathSizeCriticalSurface);
+	dml_print("DML::%s: DETBufferSizeYCriticalSurface = %u\n", __func__, DETBufferSizeYCriticalSurface);
+	dml_print("DML::%s: PixelChunkSizeInKByte = %u\n", __func__, p->PixelChunkSizeInKByte);
+	dml_print("DML::%s: LastChunkOfSwathSize = %u\n", __func__, LastChunkOfSwathSize);
+	dml_print("DML::%s: MissingPartOfLastSwathOfDETSize = %u\n", __func__, MissingPartOfLastSwathOfDETSize);
+	dml_print("DML::%s: DCHUBBUB_ARB_CSTATE_MAX_CAP_MODE = %u\n", __func__, *p->DCHUBBUB_ARB_CSTATE_MAX_CAP_MODE);
+#endif
+} // CalculateStutterEfficiency
+
+/// \CalculateSwathAndDETConfiguration
+/// @brief Calculates swath width and different return buffers sizing (DET, CDB, etc.)
+static void CalculateSwathAndDETConfiguration(struct display_mode_lib_scratch_st *scratch,
+	struct CalculateSwathAndDETConfiguration_params_st *p)
+{
+	dml_uint_t MaximumSwathHeightY[__DML_NUM_PLANES__];
+	dml_uint_t MaximumSwathHeightC[__DML_NUM_PLANES__];
+	dml_uint_t RoundedUpMaxSwathSizeBytesY[__DML_NUM_PLANES__];
+	dml_uint_t RoundedUpMaxSwathSizeBytesC[__DML_NUM_PLANES__];
+	dml_uint_t RoundedUpSwathSizeBytesY[__DML_NUM_PLANES__];
+	dml_uint_t RoundedUpSwathSizeBytesC[__DML_NUM_PLANES__];
+	dml_uint_t SwathWidthSingleDPP[__DML_NUM_PLANES__];
+	dml_uint_t SwathWidthSingleDPPChroma[__DML_NUM_PLANES__];
+
+	dml_uint_t TotalActiveDPP = 0;
+	dml_bool_t NoChromaOrLinearSurfaces = true;
+	dml_uint_t SurfaceDoingUnboundedRequest = 0;
+
+	dml_uint_t DETBufferSizeInKByteForSwathCalculation;
+
+	const long TTUFIFODEPTH = 8;
+	const long MAXIMUMCOMPRESSION = 4;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: ForceSingleDPP = %u\n", __func__, p->ForceSingleDPP);
+	for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+		dml_print("DML::%s: DPPPerSurface[%u] = %u\n", __func__, k, p->DPPPerSurface[k]);
+	}
+#endif
+	CalculateSwathWidth(p->ForceSingleDPP,
+		p->NumberOfActiveSurfaces,
+		p->SourcePixelFormat,
+		p->SourceScan,
+		p->ViewportStationary,
+		p->ViewportWidth,
+		p->ViewportHeight,
+		p->ViewportXStart,
+		p->ViewportYStart,
+		p->ViewportXStartC,
+		p->ViewportYStartC,
+		p->SurfaceWidthY,
+		p->SurfaceWidthC,
+		p->SurfaceHeightY,
+		p->SurfaceHeightC,
+		p->ODMMode,
+		p->BytePerPixY,
+		p->BytePerPixC,
+		p->Read256BytesBlockHeightY,
+		p->Read256BytesBlockHeightC,
+		p->Read256BytesBlockWidthY,
+		p->Read256BytesBlockWidthC,
+		p->BlendingAndTiming,
+		p->HActive,
+		p->HRatio,
+		p->DPPPerSurface,
+
+		// Output
+		SwathWidthSingleDPP,
+		SwathWidthSingleDPPChroma,
+		p->SwathWidth,
+		p->SwathWidthChroma,
+		MaximumSwathHeightY,
+		MaximumSwathHeightC,
+		p->swath_width_luma_ub,
+		p->swath_width_chroma_ub);
+
+	for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+		RoundedUpMaxSwathSizeBytesY[k] = (dml_uint_t)(p->swath_width_luma_ub[k] * p->BytePerPixDETY[k] * MaximumSwathHeightY[k]);
+		RoundedUpMaxSwathSizeBytesC[k] = (dml_uint_t)(p->swath_width_chroma_ub[k] * p->BytePerPixDETC[k] * MaximumSwathHeightC[k]);
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%u DPPPerSurface = %u\n", __func__, k, p->DPPPerSurface[k]);
+		dml_print("DML::%s: k=%u swath_width_luma_ub = %u\n", __func__, k, p->swath_width_luma_ub[k]);
+		dml_print("DML::%s: k=%u BytePerPixDETY = %f\n", __func__, k, p->BytePerPixDETY[k]);
+		dml_print("DML::%s: k=%u MaximumSwathHeightY = %u\n", __func__, k, MaximumSwathHeightY[k]);
+		dml_print("DML::%s: k=%u RoundedUpMaxSwathSizeBytesY = %u\n", __func__, k, RoundedUpMaxSwathSizeBytesY[k]);
+		dml_print("DML::%s: k=%u swath_width_chroma_ub = %u\n", __func__, k, p->swath_width_chroma_ub[k]);
+		dml_print("DML::%s: k=%u BytePerPixDETC = %f\n", __func__, k, p->BytePerPixDETC[k]);
+		dml_print("DML::%s: k=%u MaximumSwathHeightC = %u\n", __func__, k, MaximumSwathHeightC[k]);
+		dml_print("DML::%s: k=%u RoundedUpMaxSwathSizeBytesC = %u\n", __func__, k, RoundedUpMaxSwathSizeBytesC[k]);
+#endif
+		if (p->SourcePixelFormat[k] == dml_420_10) {
+			RoundedUpMaxSwathSizeBytesY[k] = (dml_uint_t)(dml_ceil((dml_float_t) RoundedUpMaxSwathSizeBytesY[k], 256));
+			RoundedUpMaxSwathSizeBytesC[k] = (dml_uint_t)(dml_ceil((dml_float_t) RoundedUpMaxSwathSizeBytesC[k], 256));
+		}
+	}
+
+	for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+		TotalActiveDPP = TotalActiveDPP + (p->ForceSingleDPP ? 1 : p->DPPPerSurface[k]);
+		if (p->DPPPerSurface[k] > 0)
+			SurfaceDoingUnboundedRequest = k;
+		if (p->SourcePixelFormat[k] == dml_420_8 || p->SourcePixelFormat[k] == dml_420_10 ||
+				p->SourcePixelFormat[k] == dml_420_12 || p->SourcePixelFormat[k] == dml_rgbe_alpha
+				|| p->SurfaceTiling[k] == dml_sw_linear) {
+			NoChromaOrLinearSurfaces = false;
+		}
+	}
+
+	*p->UnboundedRequestEnabled = UnboundedRequest(p->UseUnboundedRequestingFinal, TotalActiveDPP,
+			NoChromaOrLinearSurfaces, p->Output[0]);
+
+	CalculateDETBufferSize(p->DETSizeOverride,
+		p->UseMALLForPStateChange,
+		p->ForceSingleDPP,
+		p->NumberOfActiveSurfaces,
+		*p->UnboundedRequestEnabled,
+		p->nomDETInKByte,
+		p->MaxTotalDETInKByte,
+		p->ConfigReturnBufferSizeInKByte,
+		p->MinCompressedBufferSizeInKByte,
+		p->ConfigReturnBufferSegmentSizeInkByte,
+		p->CompressedBufferSegmentSizeInkByteFinal,
+		p->SourcePixelFormat,
+		p->ReadBandwidthLuma,
+		p->ReadBandwidthChroma,
+		RoundedUpMaxSwathSizeBytesY,
+		RoundedUpMaxSwathSizeBytesC,
+		p->DPPPerSurface,
+
+		// Output
+		p->DETBufferSizeInKByte, // per hubp pipe
+		p->CompressedBufferSizeInkByte);
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: TotalActiveDPP = %u\n", __func__, TotalActiveDPP);
+		dml_print("DML::%s: nomDETInKByte = %u\n", __func__, p->nomDETInKByte);
+		dml_print("DML::%s: ConfigReturnBufferSizeInKByte = %u\n", __func__, p->ConfigReturnBufferSizeInKByte);
+		dml_print("DML::%s: UseUnboundedRequestingFinal = %u\n", __func__, p->UseUnboundedRequestingFinal);
+		dml_print("DML::%s: UnboundedRequestEnabled = %u\n", __func__, *p->UnboundedRequestEnabled);
+		dml_print("DML::%s: CompressedBufferSizeInkByte = %u\n", __func__, *p->CompressedBufferSizeInkByte);
+#endif
+
+	*p->ViewportSizeSupport = true;
+	for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+
+		DETBufferSizeInKByteForSwathCalculation =  (p->UseMALLForPStateChange[k] == dml_use_mall_pstate_change_phantom_pipe ? 1024 : p->DETBufferSizeInKByte[k]);
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: k=%u DETBufferSizeInKByteForSwathCalculation = %u\n", __func__, k, DETBufferSizeInKByteForSwathCalculation);
+#endif
+
+		if (RoundedUpMaxSwathSizeBytesY[k] + RoundedUpMaxSwathSizeBytesC[k] <= DETBufferSizeInKByteForSwathCalculation * 1024 / 2) {
+			p->SwathHeightY[k] = MaximumSwathHeightY[k];
+			p->SwathHeightC[k] = MaximumSwathHeightC[k];
+			RoundedUpSwathSizeBytesY[k] = RoundedUpMaxSwathSizeBytesY[k];
+			RoundedUpSwathSizeBytesC[k] = RoundedUpMaxSwathSizeBytesC[k];
+		} else if (RoundedUpMaxSwathSizeBytesY[k] >= 1.5 * RoundedUpMaxSwathSizeBytesC[k] && RoundedUpMaxSwathSizeBytesY[k] / 2 + RoundedUpMaxSwathSizeBytesC[k] <= DETBufferSizeInKByteForSwathCalculation * 1024 / 2) {
+			p->SwathHeightY[k] = MaximumSwathHeightY[k] / 2;
+			p->SwathHeightC[k] = MaximumSwathHeightC[k];
+			RoundedUpSwathSizeBytesY[k] = RoundedUpMaxSwathSizeBytesY[k] / 2;
+			RoundedUpSwathSizeBytesC[k] = RoundedUpMaxSwathSizeBytesC[k];
+		} else if (RoundedUpMaxSwathSizeBytesY[k] < 1.5 * RoundedUpMaxSwathSizeBytesC[k] && RoundedUpMaxSwathSizeBytesY[k] + RoundedUpMaxSwathSizeBytesC[k] / 2 <= DETBufferSizeInKByteForSwathCalculation * 1024 / 2) {
+			p->SwathHeightY[k] = MaximumSwathHeightY[k];
+			p->SwathHeightC[k] = MaximumSwathHeightC[k] / 2;
+			RoundedUpSwathSizeBytesY[k] = RoundedUpMaxSwathSizeBytesY[k];
+			RoundedUpSwathSizeBytesC[k] = RoundedUpMaxSwathSizeBytesC[k] / 2;
+		} else {
+			p->SwathHeightY[k] = MaximumSwathHeightY[k] / 2;
+			p->SwathHeightC[k] = MaximumSwathHeightC[k] / 2;
+			RoundedUpSwathSizeBytesY[k] = RoundedUpMaxSwathSizeBytesY[k] / 2;
+			RoundedUpSwathSizeBytesC[k] = RoundedUpMaxSwathSizeBytesC[k] / 2;
+		}
+
+		if ((RoundedUpMaxSwathSizeBytesY[k] / 2 + RoundedUpMaxSwathSizeBytesC[k] / 2 > DETBufferSizeInKByteForSwathCalculation * 1024 / 2) ||
+			p->SwathWidth[k] > p->MaximumSwathWidthLuma[k] || (p->SwathHeightC[k] > 0 && p->SwathWidthChroma[k] > p->MaximumSwathWidthChroma[k])) {
+			*p->ViewportSizeSupport = false;
+			p->ViewportSizeSupportPerSurface[k] = false;
+		} else {
+			p->ViewportSizeSupportPerSurface[k] = true;
+		}
+
+	if (p->SwathHeightC[k] == 0) {
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: k=%u All DET for plane0\n", __func__, k);
+#endif
+			p->DETBufferSizeY[k] = p->DETBufferSizeInKByte[k] * 1024;
+			p->DETBufferSizeC[k] = 0;
+		} else if (RoundedUpSwathSizeBytesY[k] <= 1.5 * RoundedUpSwathSizeBytesC[k]) {
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: k=%u Half DET for plane0, half for plane1\n", __func__, k);
+#endif
+			p->DETBufferSizeY[k] = p->DETBufferSizeInKByte[k] * 1024 / 2;
+			p->DETBufferSizeC[k] = p->DETBufferSizeInKByte[k] * 1024 / 2;
+		} else {
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: k=%u 2/3 DET for plane0, 1/3 for plane1\n", __func__, k);
+#endif
+			p->DETBufferSizeY[k] = (dml_uint_t)(dml_floor(p->DETBufferSizeInKByte[k] * 1024 * 2 / 3, 1024));
+			p->DETBufferSizeC[k] = p->DETBufferSizeInKByte[k] * 1024 - p->DETBufferSizeY[k];
+		}
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%u SwathHeightY = %u\n", __func__, k, p->SwathHeightY[k]);
+		dml_print("DML::%s: k=%u SwathHeightC = %u\n", __func__, k, p->SwathHeightC[k]);
+		dml_print("DML::%s: k=%u RoundedUpMaxSwathSizeBytesY = %u\n", __func__, k, RoundedUpMaxSwathSizeBytesY[k]);
+		dml_print("DML::%s: k=%u RoundedUpMaxSwathSizeBytesC = %u\n", __func__, k, RoundedUpMaxSwathSizeBytesC[k]);
+		dml_print("DML::%s: k=%u RoundedUpSwathSizeBytesY = %u\n", __func__, k, RoundedUpSwathSizeBytesY[k]);
+		dml_print("DML::%s: k=%u RoundedUpSwathSizeBytesC = %u\n", __func__, k, RoundedUpSwathSizeBytesC[k]);
+		dml_print("DML::%s: k=%u DETBufferSizeInKByte = %u\n", __func__, k, p->DETBufferSizeInKByte[k]);
+		dml_print("DML::%s: k=%u DETBufferSizeY = %u\n", __func__, k, p->DETBufferSizeY[k]);
+		dml_print("DML::%s: k=%u DETBufferSizeC = %u\n", __func__, k, p->DETBufferSizeC[k]);
+		dml_print("DML::%s: k=%u ViewportSizeSupportPerSurface = %u\n", __func__, k, p->ViewportSizeSupportPerSurface[k]);
+#endif
+
+	}
+
+	*p->compbuf_reserved_space_64b = 2 * p->PixelChunkSizeInKByte * 1024 / 64;
+	if (*p->UnboundedRequestEnabled) {
+		*p->compbuf_reserved_space_64b = dml_max(*p->compbuf_reserved_space_64b,
+				(dml_float_t)(p->ROBBufferSizeInKByte * 1024/64)
+				- (dml_float_t)(RoundedUpSwathSizeBytesY[SurfaceDoingUnboundedRequest] * TTUFIFODEPTH / MAXIMUMCOMPRESSION/64));
+	}
+	*p->compbuf_reserved_space_zs = 2 * p->PixelChunkSizeInKByte * 1024 / 256;
+} // CalculateSwathAndDETConfiguration
+
+static void CalculateSwathWidth(
+		dml_bool_t ForceSingleDPP,
+		dml_uint_t NumberOfActiveSurfaces,
+		enum dml_source_format_class SourcePixelFormat[],
+		enum dml_rotation_angle SourceScan[],
+		dml_bool_t ViewportStationary[],
+		dml_uint_t ViewportWidth[],
+		dml_uint_t ViewportHeight[],
+		dml_uint_t ViewportXStart[],
+		dml_uint_t ViewportYStart[],
+		dml_uint_t ViewportXStartC[],
+		dml_uint_t ViewportYStartC[],
+		dml_uint_t SurfaceWidthY[],
+		dml_uint_t SurfaceWidthC[],
+		dml_uint_t SurfaceHeightY[],
+		dml_uint_t SurfaceHeightC[],
+		enum dml_odm_mode ODMMode[],
+		dml_uint_t BytePerPixY[],
+		dml_uint_t BytePerPixC[],
+		dml_uint_t Read256BytesBlockHeightY[],
+		dml_uint_t Read256BytesBlockHeightC[],
+		dml_uint_t Read256BytesBlockWidthY[],
+		dml_uint_t Read256BytesBlockWidthC[],
+		dml_uint_t BlendingAndTiming[],
+		dml_uint_t HActive[],
+		dml_float_t HRatio[],
+		dml_uint_t DPPPerSurface[],
+
+		// Output
+		dml_uint_t SwathWidthSingleDPPY[],
+		dml_uint_t SwathWidthSingleDPPC[],
+		dml_uint_t SwathWidthY[], // per-pipe
+		dml_uint_t SwathWidthC[], // per-pipe
+		dml_uint_t MaximumSwathHeightY[],
+		dml_uint_t MaximumSwathHeightC[],
+		dml_uint_t swath_width_luma_ub[], // per-pipe
+		dml_uint_t swath_width_chroma_ub[]) // per-pipe
+{
+	enum dml_odm_mode   MainSurfaceODMMode;
+	dml_uint_t surface_width_ub_l;
+	dml_uint_t surface_height_ub_l;
+	dml_uint_t surface_width_ub_c = 0;
+	dml_uint_t surface_height_ub_c = 0;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: ForceSingleDPP = %u\n", __func__, ForceSingleDPP);
+	dml_print("DML::%s: NumberOfActiveSurfaces = %u\n", __func__, NumberOfActiveSurfaces);
+#endif
+
+	for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+		if (!dml_is_vertical_rotation(SourceScan[k])) {
+			SwathWidthSingleDPPY[k] = ViewportWidth[k];
+		} else {
+			SwathWidthSingleDPPY[k] = ViewportHeight[k];
+		}
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%u ViewportWidth=%u\n", __func__, k, ViewportWidth[k]);
+		dml_print("DML::%s: k=%u ViewportHeight=%u\n", __func__, k, ViewportHeight[k]);
+		dml_print("DML::%s: k=%u DPPPerSurface=%u\n", __func__, k, DPPPerSurface[k]);
+#endif
+
+		MainSurfaceODMMode = ODMMode[k];
+		for (dml_uint_t j = 0; j < NumberOfActiveSurfaces; ++j) {
+			if (BlendingAndTiming[k] == j) {
+				MainSurfaceODMMode = ODMMode[j];
+			}
+		}
+
+		if (ForceSingleDPP) {
+			SwathWidthY[k] = SwathWidthSingleDPPY[k];
+		} else {
+			if (MainSurfaceODMMode == dml_odm_mode_combine_4to1) {
+				SwathWidthY[k] = (dml_uint_t)(dml_min(SwathWidthSingleDPPY[k], dml_round(HActive[k] / 4.0 * HRatio[k], true)));
+			} else if (MainSurfaceODMMode == dml_odm_mode_combine_2to1) {
+				SwathWidthY[k] = (dml_uint_t)(dml_min(SwathWidthSingleDPPY[k], dml_round(HActive[k] / 2.0 * HRatio[k], true)));
+			} else if (DPPPerSurface[k] == 2) {
+				SwathWidthY[k] = SwathWidthSingleDPPY[k] / 2;
+			} else {
+				SwathWidthY[k] = SwathWidthSingleDPPY[k];
+			}
+		}
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%u HActive=%u\n", __func__, k, HActive[k]);
+		dml_print("DML::%s: k=%u HRatio=%f\n", __func__, k, HRatio[k]);
+		dml_print("DML::%s: k=%u MainSurfaceODMMode=%u\n", __func__, k, MainSurfaceODMMode);
+		dml_print("DML::%s: k=%u SwathWidthSingleDPPY=%u\n", __func__, k, SwathWidthSingleDPPY[k]);
+		dml_print("DML::%s: k=%u SwathWidthY=%u\n", __func__, k, SwathWidthY[k]);
+#endif
+
+		if (SourcePixelFormat[k] == dml_420_8 || SourcePixelFormat[k] == dml_420_10 || SourcePixelFormat[k] == dml_420_12) {
+			SwathWidthC[k] = SwathWidthY[k] / 2;
+			SwathWidthSingleDPPC[k] = SwathWidthSingleDPPY[k] / 2;
+		} else {
+			SwathWidthC[k] = SwathWidthY[k];
+			SwathWidthSingleDPPC[k] = SwathWidthSingleDPPY[k];
+		}
+
+		if (ForceSingleDPP == true) {
+			SwathWidthY[k] = SwathWidthSingleDPPY[k];
+			SwathWidthC[k] = SwathWidthSingleDPPC[k];
+		}
+
+		surface_width_ub_l = (dml_uint_t)dml_ceil(SurfaceWidthY[k], Read256BytesBlockWidthY[k]);
+		surface_height_ub_l = (dml_uint_t)dml_ceil(SurfaceHeightY[k], Read256BytesBlockHeightY[k]);
+
+		if (!dml_is_vertical_rotation(SourceScan[k])) {
+			MaximumSwathHeightY[k] = Read256BytesBlockHeightY[k];
+			MaximumSwathHeightC[k] = Read256BytesBlockHeightC[k];
+			if (ViewportStationary[k] && DPPPerSurface[k] == 1) {
+				swath_width_luma_ub[k] = (dml_uint_t)(dml_min(surface_width_ub_l, dml_floor(ViewportXStart[k] + SwathWidthY[k] + Read256BytesBlockWidthY[k] - 1, Read256BytesBlockWidthY[k]) - dml_floor(ViewportXStart[k], Read256BytesBlockWidthY[k])));
+			} else {
+				swath_width_luma_ub[k] = (dml_uint_t)(dml_min(surface_width_ub_l, dml_ceil(SwathWidthY[k] - 1, Read256BytesBlockWidthY[k]) + Read256BytesBlockWidthY[k]));
+			}
+			if (BytePerPixC[k] > 0) {
+				surface_width_ub_c = (dml_uint_t)dml_ceil(SurfaceWidthC[k], Read256BytesBlockWidthC[k]);
+				if (ViewportStationary[k] && DPPPerSurface[k] == 1) {
+					swath_width_chroma_ub[k] = (dml_uint_t)(dml_min(surface_width_ub_c, dml_floor(ViewportXStartC[k] + SwathWidthC[k] + Read256BytesBlockWidthC[k] - 1, Read256BytesBlockWidthC[k]) - dml_floor(ViewportXStartC[k], Read256BytesBlockWidthC[k])));
+				} else {
+					swath_width_chroma_ub[k] = (dml_uint_t)(dml_min(surface_width_ub_c, dml_ceil(SwathWidthC[k] - 1, Read256BytesBlockWidthC[k]) + Read256BytesBlockWidthC[k]));
+				}
+			} else {
+				swath_width_chroma_ub[k] = 0;
+			}
+		} else {
+			MaximumSwathHeightY[k] = Read256BytesBlockWidthY[k];
+			MaximumSwathHeightC[k] = Read256BytesBlockWidthC[k];
+
+			if (ViewportStationary[k] && DPPPerSurface[k] == 1) {
+				swath_width_luma_ub[k] = (dml_uint_t)(dml_min(surface_height_ub_l, dml_floor(ViewportYStart[k] + SwathWidthY[k] + Read256BytesBlockHeightY[k] - 1, Read256BytesBlockHeightY[k]) - dml_floor(ViewportYStart[k], Read256BytesBlockHeightY[k])));
+			} else {
+				swath_width_luma_ub[k] = (dml_uint_t)(dml_min(surface_height_ub_l, dml_ceil(SwathWidthY[k] - 1, Read256BytesBlockHeightY[k]) + Read256BytesBlockHeightY[k]));
+			}
+			if (BytePerPixC[k] > 0) {
+				surface_height_ub_c = (dml_uint_t)dml_ceil(SurfaceHeightC[k], Read256BytesBlockHeightC[k]);
+				if (ViewportStationary[k] && DPPPerSurface[k] == 1) {
+					swath_width_chroma_ub[k] = (dml_uint_t)(dml_min(surface_height_ub_c, dml_floor(ViewportYStartC[k] + SwathWidthC[k] + Read256BytesBlockHeightC[k] - 1, Read256BytesBlockHeightC[k]) - dml_floor(ViewportYStartC[k], Read256BytesBlockHeightC[k])));
+				} else {
+					swath_width_chroma_ub[k] = (dml_uint_t)(dml_min(surface_height_ub_c, dml_ceil(SwathWidthC[k] - 1, Read256BytesBlockHeightC[k]) + Read256BytesBlockHeightC[k]));
+				}
+			} else {
+				swath_width_chroma_ub[k] = 0;
+			}
+		}
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%u surface_width_ub_l=%u\n", __func__, k, surface_width_ub_l);
+		dml_print("DML::%s: k=%u surface_height_ub_l=%u\n", __func__, k, surface_height_ub_l);
+		dml_print("DML::%s: k=%u surface_width_ub_c=%u\n", __func__, k, surface_width_ub_c);
+		dml_print("DML::%s: k=%u surface_height_ub_c=%u\n", __func__, k, surface_height_ub_c);
+		dml_print("DML::%s: k=%u Read256BytesBlockWidthY=%u\n", __func__, k, Read256BytesBlockWidthY[k]);
+		dml_print("DML::%s: k=%u Read256BytesBlockHeightY=%u\n", __func__, k, Read256BytesBlockHeightY[k]);
+		dml_print("DML::%s: k=%u Read256BytesBlockWidthC=%u\n", __func__, k, Read256BytesBlockWidthC[k]);
+		dml_print("DML::%s: k=%u Read256BytesBlockHeightC=%u\n", __func__, k, Read256BytesBlockHeightC[k]);
+		dml_print("DML::%s: k=%u ViewportStationary=%u\n", __func__, k, ViewportStationary[k]);
+		dml_print("DML::%s: k=%u DPPPerSurface=%u\n", __func__, k, DPPPerSurface[k]);
+		dml_print("DML::%s: k=%u swath_width_luma_ub=%u\n", __func__, k, swath_width_luma_ub[k]);
+		dml_print("DML::%s: k=%u swath_width_chroma_ub=%u\n", __func__, k, swath_width_chroma_ub[k]);
+		dml_print("DML::%s: k=%u MaximumSwathHeightY=%u\n", __func__, k, MaximumSwathHeightY[k]);
+		dml_print("DML::%s: k=%u MaximumSwathHeightC=%u\n", __func__, k, MaximumSwathHeightC[k]);
+#endif
+
+	}
+} // CalculateSwathWidth
+
+static noinline_for_stack dml_float_t CalculateExtraLatency(
+		dml_uint_t RoundTripPingLatencyCycles,
+		dml_uint_t ReorderingBytes,
+		dml_float_t DCFCLK,
+		dml_uint_t TotalNumberOfActiveDPP,
+		dml_uint_t PixelChunkSizeInKByte,
+		dml_uint_t TotalNumberOfDCCActiveDPP,
+		dml_uint_t MetaChunkSize,
+		dml_float_t ReturnBW,
+		dml_bool_t GPUVMEnable,
+		dml_bool_t HostVMEnable,
+		dml_uint_t NumberOfActiveSurfaces,
+		dml_uint_t NumberOfDPP[],
+		dml_uint_t dpte_group_bytes[],
+		dml_float_t HostVMInefficiencyFactor,
+		dml_uint_t HostVMMinPageSize,
+		dml_uint_t HostVMMaxNonCachedPageTableLevels)
+{
+	dml_float_t ExtraLatencyBytes;
+	dml_float_t ExtraLatency;
+
+	ExtraLatencyBytes = CalculateExtraLatencyBytes(
+			ReorderingBytes,
+			TotalNumberOfActiveDPP,
+			PixelChunkSizeInKByte,
+			TotalNumberOfDCCActiveDPP,
+			MetaChunkSize,
+			GPUVMEnable,
+			HostVMEnable,
+			NumberOfActiveSurfaces,
+			NumberOfDPP,
+			dpte_group_bytes,
+			HostVMInefficiencyFactor,
+			HostVMMinPageSize,
+			HostVMMaxNonCachedPageTableLevels);
+
+	ExtraLatency = (RoundTripPingLatencyCycles + __DML_ARB_TO_RET_DELAY__) / DCFCLK + ExtraLatencyBytes / ReturnBW;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: RoundTripPingLatencyCycles=%u\n", __func__, RoundTripPingLatencyCycles);
+	dml_print("DML::%s: DCFCLK=%f\n", __func__, DCFCLK);
+	dml_print("DML::%s: ExtraLatencyBytes=%f\n", __func__, ExtraLatencyBytes);
+	dml_print("DML::%s: ReturnBW=%f\n", __func__, ReturnBW);
+	dml_print("DML::%s: ExtraLatency=%f\n", __func__, ExtraLatency);
+#endif
+
+	return ExtraLatency;
+} // CalculateExtraLatency
+
+static dml_uint_t CalculateHostVMDynamicLevels(
+									dml_bool_t GPUVMEnable,
+									dml_bool_t HostVMEnable,
+									dml_uint_t HostVMMinPageSize,
+									dml_uint_t HostVMMaxNonCachedPageTableLevels)
+{
+	dml_uint_t HostVMDynamicLevels = 0;
+
+	if (GPUVMEnable && HostVMEnable) {
+		if (HostVMMinPageSize < 2048)
+			HostVMDynamicLevels = HostVMMaxNonCachedPageTableLevels;
+		else if (HostVMMinPageSize >= 2048 && HostVMMinPageSize < 1048576)
+			HostVMDynamicLevels = (dml_uint_t) dml_max(0, (dml_float_t) HostVMMaxNonCachedPageTableLevels - 1);
+		else
+			HostVMDynamicLevels = (dml_uint_t) dml_max(0, (dml_float_t) HostVMMaxNonCachedPageTableLevels - 2);
+	} else {
+		HostVMDynamicLevels = 0;
+	}
+	return HostVMDynamicLevels;
+}
+
+static dml_uint_t CalculateExtraLatencyBytes(dml_uint_t ReorderingBytes,
+										dml_uint_t TotalNumberOfActiveDPP,
+										dml_uint_t PixelChunkSizeInKByte,
+										dml_uint_t TotalNumberOfDCCActiveDPP,
+										dml_uint_t MetaChunkSize,
+										dml_bool_t GPUVMEnable,
+										dml_bool_t HostVMEnable,
+										dml_uint_t NumberOfActiveSurfaces,
+										dml_uint_t NumberOfDPP[],
+										dml_uint_t dpte_group_bytes[],
+										dml_float_t HostVMInefficiencyFactor,
+										dml_uint_t HostVMMinPageSize,
+										dml_uint_t HostVMMaxNonCachedPageTableLevels)
+{
+	dml_uint_t  HostVMDynamicLevels = CalculateHostVMDynamicLevels(GPUVMEnable, HostVMEnable, HostVMMinPageSize, HostVMMaxNonCachedPageTableLevels);
+	dml_float_t ret                 = ReorderingBytes + (TotalNumberOfActiveDPP * PixelChunkSizeInKByte + TotalNumberOfDCCActiveDPP * MetaChunkSize) * 1024.0;
+
+	if (GPUVMEnable == true) {
+		for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+			ret = ret + NumberOfDPP[k] * dpte_group_bytes[k] * (1 + 8 * HostVMDynamicLevels) * HostVMInefficiencyFactor;
+		}
+	}
+	return (dml_uint_t)(ret);
+}
+
+static dml_float_t CalculateUrgentLatency(
+		dml_float_t UrgentLatencyPixelDataOnly,
+		dml_float_t UrgentLatencyPixelMixedWithVMData,
+		dml_float_t UrgentLatencyVMDataOnly,
+		dml_bool_t DoUrgentLatencyAdjustment,
+		dml_float_t UrgentLatencyAdjustmentFabricClockComponent,
+		dml_float_t UrgentLatencyAdjustmentFabricClockReference,
+		dml_float_t FabricClock)
+{
+	dml_float_t   ret;
+
+	ret = dml_max3(UrgentLatencyPixelDataOnly, UrgentLatencyPixelMixedWithVMData, UrgentLatencyVMDataOnly);
+	if (DoUrgentLatencyAdjustment == true) {
+		ret = ret + UrgentLatencyAdjustmentFabricClockComponent * (UrgentLatencyAdjustmentFabricClockReference / FabricClock - 1);
+	}
+	return ret;
+}
+
+static dml_float_t RequiredDTBCLK(
+		dml_bool_t DSCEnable,
+		dml_float_t PixelClock,
+		enum dml_output_format_class OutputFormat,
+		dml_float_t OutputBpp,
+		dml_uint_t DSCSlices,
+		dml_uint_t HTotal,
+		dml_uint_t HActive,
+		dml_uint_t AudioRate,
+		dml_uint_t AudioLayout)
+{
+	if (DSCEnable != true) {
+		return dml_max(PixelClock / 4.0 * OutputBpp / 24.0, 25.0);
+	} else {
+		dml_float_t PixelWordRate = PixelClock / (OutputFormat == dml_444 ? 1 : 2);
+		dml_float_t HCActive = dml_ceil(DSCSlices * dml_ceil(OutputBpp * dml_ceil(HActive / DSCSlices, 1) / 8.0, 1) / 3.0, 1);
+		dml_float_t HCBlank = 64 + 32 * dml_ceil(AudioRate * (AudioLayout == 1 ? 1 : 0.25) * HTotal / (PixelClock * 1000), 1);
+		dml_float_t AverageTribyteRate = PixelWordRate * (HCActive + HCBlank) / HTotal;
+		dml_float_t HActiveTribyteRate = PixelWordRate * HCActive / HActive;
+		return dml_max4(PixelWordRate / 4.0, AverageTribyteRate / 4.0, HActiveTribyteRate / 4.0, 25.0) * 1.002;
+	}
+}
+
+static void UseMinimumDCFCLK(struct display_mode_lib_scratch_st *scratch, struct UseMinimumDCFCLK_params_st *p)
+{
+	struct UseMinimumDCFCLK_locals_st *s = &scratch->UseMinimumDCFCLK_locals;
+
+	s->NormalEfficiency = p->PercentOfIdealSDPPortBWReceivedAfterUrgLatency / 100.0;
+	for (dml_uint_t j = 0; j < 2; ++j) {
+
+
+		s->TotalMaxPrefetchFlipDPTERowBandwidth[j] = 0;
+		for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+			s->TotalMaxPrefetchFlipDPTERowBandwidth[j] = s->TotalMaxPrefetchFlipDPTERowBandwidth[j] + p->NoOfDPP[j][k] * p->DPTEBytesPerRow[j][k] / (15.75 * p->HTotal[k] / p->PixelClock[k]);
+		}
+
+		for (dml_uint_t k = 0; k <= p->NumberOfActiveSurfaces - 1; ++k) {
+			s->NoOfDPPState[k] = p->NoOfDPP[j][k];
+		}
+
+		s->DPTEBandwidth = s->TotalMaxPrefetchFlipDPTERowBandwidth[j];
+
+		s->DCFCLKRequiredForAverageBandwidth = dml_max(p->ProjectedDCFCLKDeepSleep[j], s->DPTEBandwidth / s->NormalEfficiency / p->ReturnBusWidth);
+
+		s->ExtraLatencyBytes = CalculateExtraLatencyBytes(p->ReorderingBytes, p->TotalNumberOfActiveDPP[j], p->PixelChunkSizeInKByte, p->TotalNumberOfDCCActiveDPP[j],
+			p->MetaChunkSize, p->GPUVMEnable, p->HostVMEnable, p->NumberOfActiveSurfaces, s->NoOfDPPState, p->dpte_group_bytes,
+												1, p->HostVMMinPageSize, p->HostVMMaxNonCachedPageTableLevels);
+		s->ExtraLatencyCycles = p->RoundTripPingLatencyCycles + __DML_ARB_TO_RET_DELAY__ + s->ExtraLatencyBytes / s->NormalEfficiency / p->ReturnBusWidth;
+		for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+			dml_float_t DCFCLKCyclesRequiredInPrefetch;
+			dml_float_t PrefetchTime;
+
+			s->PixelDCFCLKCyclesRequiredInPrefetch[k] = (p->PrefetchLinesY[j][k] * p->swath_width_luma_ub_all_states[j][k] * p->BytePerPixelY[k] + p->PrefetchLinesC[j][k] * p->swath_width_chroma_ub_all_states[j][k] * p->BytePerPixelC[k]) / s->NormalEfficiency / p->ReturnBusWidth;
+			DCFCLKCyclesRequiredInPrefetch = 2 * s->ExtraLatencyCycles / s->NoOfDPPState[k] + p->PDEAndMetaPTEBytesPerFrame[j][k] / s->NormalEfficiency / s->NormalEfficiency / p->ReturnBusWidth * (p->GPUVMMaxPageTableLevels > 2 ? 1 : 0) + 2 * p->DPTEBytesPerRow[j][k] / s->NormalEfficiency / s->NormalEfficiency / p->ReturnBusWidth + 2 * p->MetaRowBytes[j][k] / s->NormalEfficiency / p->ReturnBusWidth + s->PixelDCFCLKCyclesRequiredInPrefetch[k];
+			s->PrefetchPixelLinesTime[k] = dml_max(p->PrefetchLinesY[j][k], p->PrefetchLinesC[j][k]) * p->HTotal[k] / p->PixelClock[k];
+			s->DynamicMetadataVMExtraLatency[k] = (p->GPUVMEnable == true && p->DynamicMetadataEnable[k] == true && p->DynamicMetadataVMEnabled == true) ? p->UrgLatency * p->GPUVMMaxPageTableLevels * (p->HostVMEnable == true ? p->HostVMMaxNonCachedPageTableLevels + 1 : 1) : 0;
+
+			s->MinimumTWait = CalculateTWait(p->MaxPrefetchMode,
+				p->UseMALLForPStateChange[k],
+				p->SynchronizeDRRDisplaysForUCLKPStateChangeFinal,
+				p->DRRDisplay[k],
+				p->DRAMClockChangeLatencyFinal,
+				p->FCLKChangeLatency,
+				p->UrgLatency,
+				p->SREnterPlusExitTime);
+
+			PrefetchTime = (p->MaximumVStartup[j][k] - 1) * p->HTotal[k] / p->PixelClock[k] - s->MinimumTWait - p->UrgLatency * ((p->GPUVMMaxPageTableLevels <= 2 ? p->GPUVMMaxPageTableLevels : p->GPUVMMaxPageTableLevels - 2) *  (p->HostVMEnable == true ? p->HostVMMaxNonCachedPageTableLevels + 1 : 1) - 1) - s->DynamicMetadataVMExtraLatency[k];
+
+			if (PrefetchTime > 0) {
+				dml_float_t ExpectedVRatioPrefetch;
+				ExpectedVRatioPrefetch = s->PrefetchPixelLinesTime[k] / (PrefetchTime * s->PixelDCFCLKCyclesRequiredInPrefetch[k] / DCFCLKCyclesRequiredInPrefetch);
+				s->DCFCLKRequiredForPeakBandwidthPerSurface[k] = s->NoOfDPPState[k] * s->PixelDCFCLKCyclesRequiredInPrefetch[k] / s->PrefetchPixelLinesTime[k] * dml_max(1.0, ExpectedVRatioPrefetch) * dml_max(1.0, ExpectedVRatioPrefetch / 4);
+				if (p->HostVMEnable == true || p->ImmediateFlipRequirement == true) {
+					s->DCFCLKRequiredForPeakBandwidthPerSurface[k] = s->DCFCLKRequiredForPeakBandwidthPerSurface[k] + s->NoOfDPPState[k] * s->DPTEBandwidth / s->NormalEfficiency / s->NormalEfficiency / p->ReturnBusWidth;
+				}
+			} else {
+				s->DCFCLKRequiredForPeakBandwidthPerSurface[k] = p->DCFCLKPerState;
+			}
+			if (p->DynamicMetadataEnable[k] == true) {
+				dml_float_t TSetupPipe;
+				dml_float_t TdmbfPipe;
+				dml_float_t TdmsksPipe;
+				dml_float_t TdmecPipe;
+				dml_float_t AllowedTimeForUrgentExtraLatency;
+
+				CalculateVUpdateAndDynamicMetadataParameters(
+					p->MaxInterDCNTileRepeaters,
+					p->RequiredDPPCLKPerSurface[j][k],
+					p->RequiredDISPCLK[j],
+					p->ProjectedDCFCLKDeepSleep[j],
+					p->PixelClock[k],
+					p->HTotal[k],
+					p->VTotal[k] - p->VActive[k],
+					p->DynamicMetadataTransmittedBytes[k],
+					p->DynamicMetadataLinesBeforeActiveRequired[k],
+					p->Interlace[k],
+					p->ProgressiveToInterlaceUnitInOPP,
+
+					// Output
+					&TSetupPipe,
+					&TdmbfPipe,
+					&TdmecPipe,
+					&TdmsksPipe,
+					&s->dummy1,
+					&s->dummy2,
+					&s->dummy3);
+
+				AllowedTimeForUrgentExtraLatency = p->MaximumVStartup[j][k] * p->HTotal[k] / p->PixelClock[k] - s->MinimumTWait - TSetupPipe - TdmbfPipe - TdmecPipe - TdmsksPipe - s->DynamicMetadataVMExtraLatency[k];
+				if (AllowedTimeForUrgentExtraLatency > 0) {
+					s->DCFCLKRequiredForPeakBandwidthPerSurface[k] = dml_max(s->DCFCLKRequiredForPeakBandwidthPerSurface[k], s->ExtraLatencyCycles / AllowedTimeForUrgentExtraLatency);
+				} else {
+					s->DCFCLKRequiredForPeakBandwidthPerSurface[k] = p->DCFCLKPerState;
+				}
+			}
+		}
+		s->DCFCLKRequiredForPeakBandwidth = 0;
+		for (dml_uint_t k = 0; k <= p->NumberOfActiveSurfaces - 1; ++k) {
+			s->DCFCLKRequiredForPeakBandwidth = s->DCFCLKRequiredForPeakBandwidth + s->DCFCLKRequiredForPeakBandwidthPerSurface[k];
+		}
+		s->MinimumTvmPlus2Tr0 = p->UrgLatency * (p->GPUVMEnable == true ? (p->HostVMEnable == true ? (p->GPUVMMaxPageTableLevels + 2) * (p->HostVMMaxNonCachedPageTableLevels + 1) - 1 : p->GPUVMMaxPageTableLevels + 1) : 0);
+		for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+			dml_float_t MaximumTvmPlus2Tr0PlusTsw;
+			MaximumTvmPlus2Tr0PlusTsw = (p->MaximumVStartup[j][k] - 2) * p->HTotal[k] / p->PixelClock[k] - s->MinimumTWait - s->DynamicMetadataVMExtraLatency[k];
+			if (MaximumTvmPlus2Tr0PlusTsw <= s->MinimumTvmPlus2Tr0 + s->PrefetchPixelLinesTime[k] / 4) {
+				s->DCFCLKRequiredForPeakBandwidth = p->DCFCLKPerState;
+			} else {
+				s->DCFCLKRequiredForPeakBandwidth = dml_max3(s->DCFCLKRequiredForPeakBandwidth,
+														2 * s->ExtraLatencyCycles / (MaximumTvmPlus2Tr0PlusTsw - s->MinimumTvmPlus2Tr0 - s->PrefetchPixelLinesTime[k] / 4),
+														(2 * s->ExtraLatencyCycles + s->PixelDCFCLKCyclesRequiredInPrefetch[k]) / (MaximumTvmPlus2Tr0PlusTsw - s->MinimumTvmPlus2Tr0));
+			}
+		}
+		p->DCFCLKState[j] = dml_min(p->DCFCLKPerState, 1.05 * dml_max(s->DCFCLKRequiredForAverageBandwidth, s->DCFCLKRequiredForPeakBandwidth));
+	}
+}
+
+
+static dml_bool_t UnboundedRequest(enum dml_unbounded_requesting_policy UseUnboundedRequestingFinal,
+						dml_uint_t TotalNumberOfActiveDPP,
+						dml_bool_t NoChromaOrLinear,
+						enum dml_output_encoder_class Output)
+{
+	dml_bool_t ret_val = false;
+
+	ret_val = (UseUnboundedRequestingFinal != dml_unbounded_requesting_disable
+			&& TotalNumberOfActiveDPP == 1 && NoChromaOrLinear);
+	if (UseUnboundedRequestingFinal == dml_unbounded_requesting_edp_only && Output != dml_edp) {
+		ret_val = false;
+	}
+	return (ret_val);
+}
+
+static void CalculateSurfaceSizeInMall(
+		dml_uint_t NumberOfActiveSurfaces,
+		dml_uint_t MALLAllocatedForDCN,
+		enum dml_use_mall_for_static_screen_mode UseMALLForStaticScreen[],
+		dml_bool_t DCCEnable[],
+		dml_bool_t ViewportStationary[],
+		dml_uint_t ViewportXStartY[],
+		dml_uint_t ViewportYStartY[],
+		dml_uint_t ViewportXStartC[],
+		dml_uint_t ViewportYStartC[],
+		dml_uint_t ViewportWidthY[],
+		dml_uint_t ViewportHeightY[],
+		dml_uint_t BytesPerPixelY[],
+		dml_uint_t ViewportWidthC[],
+		dml_uint_t ViewportHeightC[],
+		dml_uint_t BytesPerPixelC[],
+		dml_uint_t SurfaceWidthY[],
+		dml_uint_t SurfaceWidthC[],
+		dml_uint_t SurfaceHeightY[],
+		dml_uint_t SurfaceHeightC[],
+		dml_uint_t Read256BytesBlockWidthY[],
+		dml_uint_t Read256BytesBlockWidthC[],
+		dml_uint_t Read256BytesBlockHeightY[],
+		dml_uint_t Read256BytesBlockHeightC[],
+		dml_uint_t ReadBlockWidthY[],
+		dml_uint_t ReadBlockWidthC[],
+		dml_uint_t ReadBlockHeightY[],
+		dml_uint_t ReadBlockHeightC[],
+
+		// Output
+		dml_uint_t SurfaceSizeInMALL[],
+		dml_bool_t *ExceededMALLSize)
+{
+	dml_uint_t TotalSurfaceSizeInMALL  = 0;
+
+	for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+		if (ViewportStationary[k]) {
+			SurfaceSizeInMALL[k] = (dml_uint_t)(dml_min(dml_ceil(SurfaceWidthY[k], ReadBlockWidthY[k]), dml_floor(ViewportXStartY[k] + ViewportWidthY[k] + ReadBlockWidthY[k] - 1, ReadBlockWidthY[k]) - dml_floor(ViewportXStartY[k], ReadBlockWidthY[k])) *
+									dml_min(dml_ceil(SurfaceHeightY[k], ReadBlockHeightY[k]), dml_floor(ViewportYStartY[k] + ViewportHeightY[k] + ReadBlockHeightY[k] - 1, ReadBlockHeightY[k]) - dml_floor(ViewportYStartY[k], ReadBlockHeightY[k])) *
+									BytesPerPixelY[k]);
+
+			if (ReadBlockWidthC[k] > 0) {
+				SurfaceSizeInMALL[k] = (dml_uint_t)(SurfaceSizeInMALL[k] +
+										dml_min(dml_ceil(SurfaceWidthC[k], ReadBlockWidthC[k]), dml_floor(ViewportXStartC[k] + ViewportWidthC[k] + ReadBlockWidthC[k] - 1, ReadBlockWidthC[k]) - dml_floor(ViewportXStartC[k], ReadBlockWidthC[k])) *
+										dml_min(dml_ceil(SurfaceHeightC[k], ReadBlockHeightC[k]), dml_floor(ViewportYStartC[k] + ViewportHeightC[k] + ReadBlockHeightC[k] - 1, ReadBlockHeightC[k]) - dml_floor(ViewportYStartC[k], ReadBlockHeightC[k])) * BytesPerPixelC[k]);
+			}
+			if (DCCEnable[k] == true) {
+				SurfaceSizeInMALL[k] = (dml_uint_t)(SurfaceSizeInMALL[k] +
+					dml_min(dml_ceil(SurfaceWidthY[k], 8 * Read256BytesBlockWidthY[k]), dml_floor(ViewportXStartY[k] + ViewportWidthY[k] + 8 * Read256BytesBlockWidthY[k] - 1, 8 * Read256BytesBlockWidthY[k]) - dml_floor(ViewportXStartY[k], 8 * Read256BytesBlockWidthY[k])) *
+					dml_min(dml_ceil(SurfaceHeightY[k], 8 * Read256BytesBlockHeightY[k]), dml_floor(ViewportYStartY[k] + ViewportHeightY[k] + 8 * Read256BytesBlockHeightY[k] - 1, 8 * Read256BytesBlockHeightY[k]) - dml_floor(ViewportYStartY[k], 8 * Read256BytesBlockHeightY[k])) * BytesPerPixelY[k] / 256);
+				if (Read256BytesBlockWidthC[k] > 0) {
+					SurfaceSizeInMALL[k] = (dml_uint_t)(SurfaceSizeInMALL[k] +
+						dml_min(dml_ceil(SurfaceWidthC[k], 8 * Read256BytesBlockWidthC[k]), dml_floor(ViewportXStartC[k] + ViewportWidthC[k] + 8 * Read256BytesBlockWidthC[k] - 1, 8 * Read256BytesBlockWidthC[k]) - dml_floor(ViewportXStartC[k], 8 * Read256BytesBlockWidthC[k])) *
+						dml_min(dml_ceil(SurfaceHeightC[k], 8 * Read256BytesBlockHeightC[k]), dml_floor(ViewportYStartC[k] + ViewportHeightC[k] + 8 * Read256BytesBlockHeightC[k] - 1, 8 * Read256BytesBlockHeightC[k]) - dml_floor(ViewportYStartC[k], 8 * Read256BytesBlockHeightC[k])) * BytesPerPixelC[k] / 256);
+				}
+			}
+		} else {
+			SurfaceSizeInMALL[k] = (dml_uint_t)(dml_ceil(dml_min(SurfaceWidthY[k], ViewportWidthY[k] + ReadBlockWidthY[k] - 1), ReadBlockWidthY[k]) * dml_ceil(dml_min(SurfaceHeightY[k], ViewportHeightY[k] + ReadBlockHeightY[k] - 1), ReadBlockHeightY[k]) * BytesPerPixelY[k]);
+			if (ReadBlockWidthC[k] > 0) {
+				SurfaceSizeInMALL[k] = (dml_uint_t)(SurfaceSizeInMALL[k] +
+										dml_ceil(dml_min(SurfaceWidthC[k], ViewportWidthC[k] + ReadBlockWidthC[k] - 1), ReadBlockWidthC[k]) *
+										dml_ceil(dml_min(SurfaceHeightC[k], ViewportHeightC[k] + ReadBlockHeightC[k] - 1), ReadBlockHeightC[k]) * BytesPerPixelC[k]);
+		}
+		if (DCCEnable[k] == true) {
+			SurfaceSizeInMALL[k] = (dml_uint_t)(SurfaceSizeInMALL[k] +
+					dml_ceil(dml_min(SurfaceWidthY[k], ViewportWidthY[k] + 8 * Read256BytesBlockWidthY[k] - 1), 8 * Read256BytesBlockWidthY[k]) *
+					dml_ceil(dml_min(SurfaceHeightY[k], ViewportHeightY[k] + 8 * Read256BytesBlockHeightY[k] - 1), 8 * Read256BytesBlockHeightY[k]) * BytesPerPixelY[k] / 256);
+
+				if (Read256BytesBlockWidthC[k] > 0) {
+					SurfaceSizeInMALL[k] = (dml_uint_t)(SurfaceSizeInMALL[k] +
+						dml_ceil(dml_min(SurfaceWidthC[k], ViewportWidthC[k] + 8 * Read256BytesBlockWidthC[k] - 1), 8 * Read256BytesBlockWidthC[k]) *
+						dml_ceil(dml_min(SurfaceHeightC[k], ViewportHeightC[k] + 8 * Read256BytesBlockHeightC[k] - 1), 8 * Read256BytesBlockHeightC[k]) * BytesPerPixelC[k] / 256);
+				}
+			}
+		}
+	}
+
+	for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+		if (UseMALLForStaticScreen[k] == dml_use_mall_static_screen_enable)
+			TotalSurfaceSizeInMALL = TotalSurfaceSizeInMALL + SurfaceSizeInMALL[k];
+	}
+	*ExceededMALLSize = (TotalSurfaceSizeInMALL > MALLAllocatedForDCN * 1024 * 1024);
+} // CalculateSurfaceSizeInMall
+
+static void CalculateDETBufferSize(
+						dml_uint_t DETSizeOverride[],
+						enum dml_use_mall_for_pstate_change_mode UseMALLForPStateChange[],
+						dml_bool_t ForceSingleDPP,
+						dml_uint_t NumberOfActiveSurfaces,
+						dml_bool_t UnboundedRequestEnabled,
+						dml_uint_t nomDETInKByte,
+						dml_uint_t MaxTotalDETInKByte,
+						dml_uint_t ConfigReturnBufferSizeInKByte,
+						dml_uint_t MinCompressedBufferSizeInKByte,
+						dml_uint_t ConfigReturnBufferSegmentSizeInkByte,
+						dml_uint_t CompressedBufferSegmentSizeInkByteFinal,
+						enum dml_source_format_class SourcePixelFormat[],
+						dml_float_t ReadBandwidthLuma[],
+						dml_float_t ReadBandwidthChroma[],
+						dml_uint_t RoundedUpMaxSwathSizeBytesY[],
+						dml_uint_t RoundedUpMaxSwathSizeBytesC[],
+						dml_uint_t DPPPerSurface[],
+						// Output
+						dml_uint_t DETBufferSizeInKByte[],
+						dml_uint_t *CompressedBufferSizeInkByte)
+{
+	dml_uint_t DETBufferSizePoolInKByte;
+	dml_uint_t NextDETBufferPieceInKByte;
+	dml_bool_t DETPieceAssignedToThisSurfaceAlready[__DML_NUM_PLANES__];
+	dml_bool_t NextPotentialSurfaceToAssignDETPieceFound;
+	dml_uint_t NextSurfaceToAssignDETPiece;
+	dml_float_t TotalBandwidth;
+	dml_float_t BandwidthOfSurfacesNotAssignedDETPiece;
+	dml_uint_t max_minDET;
+	dml_uint_t minDET;
+	dml_uint_t minDET_pipe;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: ForceSingleDPP = %u\n", __func__, ForceSingleDPP);
+	dml_print("DML::%s: nomDETInKByte = %u\n", __func__, nomDETInKByte);
+	dml_print("DML::%s: NumberOfActiveSurfaces = %u\n", __func__, NumberOfActiveSurfaces);
+	dml_print("DML::%s: UnboundedRequestEnabled = %u\n", __func__, UnboundedRequestEnabled);
+	dml_print("DML::%s: MaxTotalDETInKByte = %u\n", __func__, MaxTotalDETInKByte);
+	dml_print("DML::%s: ConfigReturnBufferSizeInKByte = %u\n", __func__, ConfigReturnBufferSizeInKByte);
+	dml_print("DML::%s: MinCompressedBufferSizeInKByte = %u\n", __func__, MinCompressedBufferSizeInKByte);
+	dml_print("DML::%s: CompressedBufferSegmentSizeInkByteFinal = %u\n", __func__, CompressedBufferSegmentSizeInkByteFinal);
+#endif
+
+	// Note: Will use default det size if that fits 2 swaths
+	if (UnboundedRequestEnabled) {
+		if (DETSizeOverride[0] > 0) {
+			DETBufferSizeInKByte[0] = DETSizeOverride[0];
+		} else {
+			DETBufferSizeInKByte[0] = (dml_uint_t) dml_max(128.0, dml_ceil(2.0 * ((dml_float_t) RoundedUpMaxSwathSizeBytesY[0] + (dml_float_t) RoundedUpMaxSwathSizeBytesC[0]) / 1024.0, ConfigReturnBufferSegmentSizeInkByte));
+		}
+		*CompressedBufferSizeInkByte = ConfigReturnBufferSizeInKByte - DETBufferSizeInKByte[0];
+	} else {
+		DETBufferSizePoolInKByte = MaxTotalDETInKByte;
+		for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+			DETBufferSizeInKByte[k] = 0;
+			if (SourcePixelFormat[k] == dml_420_8 || SourcePixelFormat[k] == dml_420_10 || SourcePixelFormat[k] == dml_420_12) {
+				max_minDET = nomDETInKByte - ConfigReturnBufferSegmentSizeInkByte;
+			} else {
+				max_minDET = nomDETInKByte;
+			}
+			minDET = 128;
+			minDET_pipe = 0;
+
+			// add DET resource until can hold 2 full swaths
+			while (minDET <= max_minDET && minDET_pipe == 0) {
+				if (2.0 * ((dml_float_t) RoundedUpMaxSwathSizeBytesY[k] + (dml_float_t) RoundedUpMaxSwathSizeBytesC[k]) / 1024.0 <= minDET)
+					minDET_pipe = minDET;
+				minDET = minDET + ConfigReturnBufferSegmentSizeInkByte;
+			}
+
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: k=%u minDET = %u\n", __func__, k, minDET);
+			dml_print("DML::%s: k=%u max_minDET = %u\n", __func__, k, max_minDET);
+			dml_print("DML::%s: k=%u minDET_pipe = %u\n", __func__, k, minDET_pipe);
+			dml_print("DML::%s: k=%u RoundedUpMaxSwathSizeBytesY = %u\n", __func__, k, RoundedUpMaxSwathSizeBytesY[k]);
+			dml_print("DML::%s: k=%u RoundedUpMaxSwathSizeBytesC = %u\n", __func__, k, RoundedUpMaxSwathSizeBytesC[k]);
+#endif
+
+			if (minDET_pipe == 0) {
+				minDET_pipe = (dml_uint_t)(dml_max(128, dml_ceil(((dml_float_t)RoundedUpMaxSwathSizeBytesY[k] + (dml_float_t)RoundedUpMaxSwathSizeBytesC[k]) / 1024.0, ConfigReturnBufferSegmentSizeInkByte)));
+#ifdef __DML_VBA_DEBUG__
+				dml_print("DML::%s: k=%u minDET_pipe = %u (assume each plane take half DET)\n", __func__, k, minDET_pipe);
+#endif
+			}
+
+			if (UseMALLForPStateChange[k] == dml_use_mall_pstate_change_phantom_pipe) {
+				DETBufferSizeInKByte[k] = 0;
+			} else if (DETSizeOverride[k] > 0) {
+				DETBufferSizeInKByte[k] = DETSizeOverride[k];
+				DETBufferSizePoolInKByte = DETBufferSizePoolInKByte - (ForceSingleDPP ? 1 : DPPPerSurface[k]) * DETSizeOverride[k];
+			} else if ((ForceSingleDPP ? 1 : DPPPerSurface[k]) * minDET_pipe <= DETBufferSizePoolInKByte) {
+				DETBufferSizeInKByte[k] = minDET_pipe;
+				DETBufferSizePoolInKByte = DETBufferSizePoolInKByte - (ForceSingleDPP ? 1 : DPPPerSurface[k]) * minDET_pipe;
+			}
+
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: k=%u DPPPerSurface = %u\n", __func__, k, DPPPerSurface[k]);
+			dml_print("DML::%s: k=%u DETSizeOverride = %u\n", __func__, k, DETSizeOverride[k]);
+			dml_print("DML::%s: k=%u DETBufferSizeInKByte = %u\n", __func__, k, DETBufferSizeInKByte[k]);
+			dml_print("DML::%s: DETBufferSizePoolInKByte = %u\n", __func__, DETBufferSizePoolInKByte);
+#endif
+		}
+
+		TotalBandwidth = 0;
+		for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+			if (UseMALLForPStateChange[k] != dml_use_mall_pstate_change_phantom_pipe)
+				TotalBandwidth = TotalBandwidth + ReadBandwidthLuma[k] + ReadBandwidthChroma[k];
+		}
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: --- Before bandwidth adjustment ---\n", __func__);
+		for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+			dml_print("DML::%s: k=%u DETBufferSizeInKByte   = %u\n", __func__, k, DETBufferSizeInKByte[k]);
+		}
+		dml_print("DML::%s: --- DET allocation with bandwidth ---\n", __func__);
+#endif
+		dml_print("DML::%s: TotalBandwidth = %f\n", __func__, TotalBandwidth);
+		BandwidthOfSurfacesNotAssignedDETPiece = TotalBandwidth;
+		for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+
+			if (UseMALLForPStateChange[k] == dml_use_mall_pstate_change_phantom_pipe) {
+				DETPieceAssignedToThisSurfaceAlready[k] = true;
+			} else if (DETSizeOverride[k] > 0 || (((dml_float_t) (ForceSingleDPP ? 1 : DPPPerSurface[k]) * (dml_float_t) DETBufferSizeInKByte[k] / (dml_float_t) MaxTotalDETInKByte) >= ((ReadBandwidthLuma[k] + ReadBandwidthChroma[k]) / TotalBandwidth))) {
+				DETPieceAssignedToThisSurfaceAlready[k] = true;
+				BandwidthOfSurfacesNotAssignedDETPiece = BandwidthOfSurfacesNotAssignedDETPiece - ReadBandwidthLuma[k] - ReadBandwidthChroma[k];
+			} else {
+				DETPieceAssignedToThisSurfaceAlready[k] = false;
+			}
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: k=%u DETPieceAssignedToThisSurfaceAlready = %u\n", __func__, k, DETPieceAssignedToThisSurfaceAlready[k]);
+			dml_print("DML::%s: k=%u BandwidthOfSurfacesNotAssignedDETPiece = %f\n", __func__, k, BandwidthOfSurfacesNotAssignedDETPiece);
+#endif
+		}
+
+		for (dml_uint_t j = 0; j < NumberOfActiveSurfaces; ++j) {
+			NextPotentialSurfaceToAssignDETPieceFound = false;
+			NextSurfaceToAssignDETPiece = 0;
+
+			for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+#ifdef __DML_VBA_DEBUG__
+				dml_print("DML::%s: j=%u k=%u, ReadBandwidthLuma[k] = %f\n", __func__, j, k, ReadBandwidthLuma[k]);
+				dml_print("DML::%s: j=%u k=%u, ReadBandwidthChroma[k] = %f\n", __func__, j, k, ReadBandwidthChroma[k]);
+				dml_print("DML::%s: j=%u k=%u, ReadBandwidthLuma[Next] = %f\n", __func__, j, k, ReadBandwidthLuma[NextSurfaceToAssignDETPiece]);
+				dml_print("DML::%s: j=%u k=%u, ReadBandwidthChroma[Next] = %f\n", __func__, j, k, ReadBandwidthChroma[NextSurfaceToAssignDETPiece]);
+				dml_print("DML::%s: j=%u k=%u, NextSurfaceToAssignDETPiece = %u\n", __func__, j, k, NextSurfaceToAssignDETPiece);
+#endif
+				if (!DETPieceAssignedToThisSurfaceAlready[k] && (!NextPotentialSurfaceToAssignDETPieceFound ||
+					ReadBandwidthLuma[k] + ReadBandwidthChroma[k] < ReadBandwidthLuma[NextSurfaceToAssignDETPiece] + ReadBandwidthChroma[NextSurfaceToAssignDETPiece])) {
+					NextSurfaceToAssignDETPiece = k;
+					NextPotentialSurfaceToAssignDETPieceFound = true;
+				}
+#ifdef __DML_VBA_DEBUG__
+				dml_print("DML::%s: j=%u k=%u, DETPieceAssignedToThisSurfaceAlready = %u\n", __func__, j, k, DETPieceAssignedToThisSurfaceAlready[k]);
+				dml_print("DML::%s: j=%u k=%u, NextPotentialSurfaceToAssignDETPieceFound = %u\n", __func__, j, k, NextPotentialSurfaceToAssignDETPieceFound);
+#endif
+			}
+
+			if (NextPotentialSurfaceToAssignDETPieceFound) {
+				// Note: To show the banker's rounding behavior in VBA and also the fact that the DET buffer size varies due to precision issue
+				//
+				//dml_float_t tmp1 =  ((dml_float_t) DETBufferSizePoolInKByte * (ReadBandwidthLuma[NextSurfaceToAssignDETPiece] + ReadBandwidthChroma[NextSurfaceToAssignDETPiece]) / BandwidthOfSurfacesNotAssignedDETPiece /
+				//                         ((ForceSingleDPP ? 1 : DPPPerSurface[NextSurfaceToAssignDETPiece]) * 64.0));
+				//dml_float_t tmp2 =  dml_round((dml_float_t) DETBufferSizePoolInKByte * (ReadBandwidthLuma[NextSurfaceToAssignDETPiece] + ReadBandwidthChroma[NextSurfaceToAssignDETPiece]) / BandwidthOfSurfacesNotAssignedDETPiece /
+				//                         ((ForceSingleDPP ? 1 : DPPPerSurface[NextSurfaceToAssignDETPiece]) * 64.0));
+				//
+				//dml_print("DML::%s: j=%u, tmp1 = %f\n", __func__, j, tmp1);
+				//dml_print("DML::%s: j=%u, tmp2 = %f\n", __func__, j, tmp2);
+
+				NextDETBufferPieceInKByte = (dml_uint_t)(dml_min(
+											dml_round((dml_float_t) DETBufferSizePoolInKByte * (ReadBandwidthLuma[NextSurfaceToAssignDETPiece] + ReadBandwidthChroma[NextSurfaceToAssignDETPiece]) / BandwidthOfSurfacesNotAssignedDETPiece /
+												((ForceSingleDPP ? 1 : DPPPerSurface[NextSurfaceToAssignDETPiece]) * ConfigReturnBufferSegmentSizeInkByte), true)
+												* (ForceSingleDPP ? 1 : DPPPerSurface[NextSurfaceToAssignDETPiece]) * ConfigReturnBufferSegmentSizeInkByte,
+											dml_floor((dml_float_t) DETBufferSizePoolInKByte, (ForceSingleDPP ? 1 : DPPPerSurface[NextSurfaceToAssignDETPiece]) * ConfigReturnBufferSegmentSizeInkByte)));
+
+#ifdef __DML_VBA_DEBUG__
+				dml_print("DML::%s: j=%u, DETBufferSizePoolInKByte = %u\n", __func__, j, DETBufferSizePoolInKByte);
+				dml_print("DML::%s: j=%u, NextSurfaceToAssignDETPiece = %u\n", __func__, j, NextSurfaceToAssignDETPiece);
+				dml_print("DML::%s: j=%u, ReadBandwidthLuma[%u] = %f\n", __func__, j, NextSurfaceToAssignDETPiece, ReadBandwidthLuma[NextSurfaceToAssignDETPiece]);
+				dml_print("DML::%s: j=%u, ReadBandwidthChroma[%u] = %f\n", __func__, j, NextSurfaceToAssignDETPiece, ReadBandwidthChroma[NextSurfaceToAssignDETPiece]);
+				dml_print("DML::%s: j=%u, BandwidthOfSurfacesNotAssignedDETPiece = %f\n", __func__, j, BandwidthOfSurfacesNotAssignedDETPiece);
+				dml_print("DML::%s: j=%u, NextDETBufferPieceInKByte = %u\n", __func__, j, NextDETBufferPieceInKByte);
+				dml_print("DML::%s: j=%u, DETBufferSizeInKByte[%u] increases from %u ", __func__, j, NextSurfaceToAssignDETPiece, DETBufferSizeInKByte[NextSurfaceToAssignDETPiece]);
+#endif
+
+				DETBufferSizeInKByte[NextSurfaceToAssignDETPiece] = DETBufferSizeInKByte[NextSurfaceToAssignDETPiece] + NextDETBufferPieceInKByte / (ForceSingleDPP ? 1 : DPPPerSurface[NextSurfaceToAssignDETPiece]);
+#ifdef __DML_VBA_DEBUG__
+				dml_print("to %u\n", DETBufferSizeInKByte[NextSurfaceToAssignDETPiece]);
+#endif
+
+				DETBufferSizePoolInKByte = DETBufferSizePoolInKByte - NextDETBufferPieceInKByte;
+				DETPieceAssignedToThisSurfaceAlready[NextSurfaceToAssignDETPiece] = true;
+				BandwidthOfSurfacesNotAssignedDETPiece = BandwidthOfSurfacesNotAssignedDETPiece - (ReadBandwidthLuma[NextSurfaceToAssignDETPiece] + ReadBandwidthChroma[NextSurfaceToAssignDETPiece]);
+			}
+		}
+		*CompressedBufferSizeInkByte = MinCompressedBufferSizeInKByte;
+	}
+	*CompressedBufferSizeInkByte = *CompressedBufferSizeInkByte * CompressedBufferSegmentSizeInkByteFinal / ConfigReturnBufferSegmentSizeInkByte;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: --- After bandwidth adjustment ---\n", __func__);
+	dml_print("DML::%s: CompressedBufferSizeInkByte = %u\n", __func__, *CompressedBufferSizeInkByte);
+	for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+		dml_print("DML::%s: k=%u DETBufferSizeInKByte = %u (TotalReadBandWidth=%f)\n", __func__, k, DETBufferSizeInKByte[k], ReadBandwidthLuma[k] + ReadBandwidthChroma[k]);
+	}
+#endif
+} // CalculateDETBufferSize
+
+
+/// @brief Calculate the bound for return buffer sizing
+static void CalculateMaxDETAndMinCompressedBufferSize(
+		dml_uint_t  ConfigReturnBufferSizeInKByte,
+		dml_uint_t  ConfigReturnBufferSegmentSizeInKByte,
+		dml_uint_t  ROBBufferSizeInKByte,
+		dml_uint_t MaxNumDPP,
+		dml_bool_t nomDETInKByteOverrideEnable, // VBA_DELTA, allow DV to override default DET size
+		dml_uint_t nomDETInKByteOverrideValue,  // VBA_DELTA
+
+		// Output
+		dml_uint_t *MaxTotalDETInKByte,
+		dml_uint_t *nomDETInKByte,
+		dml_uint_t *MinCompressedBufferSizeInKByte)
+{
+	*MaxTotalDETInKByte = ConfigReturnBufferSizeInKByte - ConfigReturnBufferSegmentSizeInKByte;
+	*nomDETInKByte = (dml_uint_t)(dml_floor((dml_float_t) *MaxTotalDETInKByte / (dml_float_t) MaxNumDPP, ConfigReturnBufferSegmentSizeInKByte));
+	*MinCompressedBufferSizeInKByte = ConfigReturnBufferSizeInKByte - *MaxTotalDETInKByte;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: ConfigReturnBufferSizeInKByte = %u\n", __func__, ConfigReturnBufferSizeInKByte);
+	dml_print("DML::%s: ROBBufferSizeInKByte = %u\n", __func__, ROBBufferSizeInKByte);
+	dml_print("DML::%s: MaxNumDPP = %u\n", __func__, MaxNumDPP);
+	dml_print("DML::%s: MaxTotalDETInKByte = %u\n", __func__, *MaxTotalDETInKByte);
+	dml_print("DML::%s: nomDETInKByte = %u\n", __func__, *nomDETInKByte);
+	dml_print("DML::%s: MinCompressedBufferSizeInKByte = %u\n", __func__, *MinCompressedBufferSizeInKByte);
+#endif
+
+	if (nomDETInKByteOverrideEnable) {
+		*nomDETInKByte = nomDETInKByteOverrideValue;
+		dml_print("DML::%s: nomDETInKByte = %u (overrided)\n", __func__, *nomDETInKByte);
+	}
+} // CalculateMaxDETAndMinCompressedBufferSize
+
+/// @brief Calculate all the RQ request attributes, like row height and # swath
+static void CalculateVMRowAndSwath(struct display_mode_lib_scratch_st *scratch,
+		struct CalculateVMRowAndSwath_params_st *p)
+{
+	struct CalculateVMRowAndSwath_locals_st *s = &scratch->CalculateVMRowAndSwath_locals;
+
+	s->HostVMDynamicLevels = CalculateHostVMDynamicLevels(p->GPUVMEnable, p->HostVMEnable, p->HostVMMinPageSize, p->HostVMMaxNonCachedPageTableLevels);
+
+	for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+		if (p->HostVMEnable == true) {
+			p->vm_group_bytes[k] = 512;
+			p->dpte_group_bytes[k] = 512;
+		} else if (p->GPUVMEnable == true) {
+			p->vm_group_bytes[k] = 2048;
+			if (p->GPUVMMinPageSizeKBytes[k] >= 64 && dml_is_vertical_rotation(p->myPipe[k].SourceScan)) {
+				p->dpte_group_bytes[k] = 512;
+			} else {
+				p->dpte_group_bytes[k] = 2048;
+			}
+		} else {
+			p->vm_group_bytes[k] = 0;
+			p->dpte_group_bytes[k] = 0;
+		}
+
+		if (p->myPipe[k].SourcePixelFormat == dml_420_8 || p->myPipe[k].SourcePixelFormat == dml_420_10 ||
+			p->myPipe[k].SourcePixelFormat == dml_420_12 || p->myPipe[k].SourcePixelFormat == dml_rgbe_alpha) {
+			if ((p->myPipe[k].SourcePixelFormat == dml_420_10 || p->myPipe[k].SourcePixelFormat == dml_420_12) && !dml_is_vertical_rotation(p->myPipe[k].SourceScan)) {
+				s->PTEBufferSizeInRequestsForLuma[k] = (p->PTEBufferSizeInRequestsLuma + p->PTEBufferSizeInRequestsChroma) / 2;
+				s->PTEBufferSizeInRequestsForChroma[k] = s->PTEBufferSizeInRequestsForLuma[k];
+			} else {
+				s->PTEBufferSizeInRequestsForLuma[k] = p->PTEBufferSizeInRequestsLuma;
+				s->PTEBufferSizeInRequestsForChroma[k] = p->PTEBufferSizeInRequestsChroma;
+			}
+
+			s->PDEAndMetaPTEBytesFrameC = CalculateVMAndRowBytes(
+				p->myPipe[k].ViewportStationary,
+				p->myPipe[k].DCCEnable,
+				p->myPipe[k].DPPPerSurface,
+				p->myPipe[k].BlockHeight256BytesC,
+				p->myPipe[k].BlockWidth256BytesC,
+				p->myPipe[k].SourcePixelFormat,
+				p->myPipe[k].SurfaceTiling,
+				p->myPipe[k].BytePerPixelC,
+				p->myPipe[k].SourceScan,
+				p->SwathWidthC[k],
+				p->myPipe[k].ViewportHeightChroma,
+				p->myPipe[k].ViewportXStartC,
+				p->myPipe[k].ViewportYStartC,
+				p->GPUVMEnable,
+				p->GPUVMMaxPageTableLevels,
+				p->GPUVMMinPageSizeKBytes[k],
+				s->PTEBufferSizeInRequestsForChroma[k],
+				p->myPipe[k].PitchC,
+				p->myPipe[k].DCCMetaPitchC,
+				p->myPipe[k].BlockWidthC,
+				p->myPipe[k].BlockHeightC,
+
+				// Output
+				&s->MetaRowByteC[k],
+				&s->PixelPTEBytesPerRowC[k],
+				&s->PixelPTEBytesPerRowStorageC[k],
+				&p->dpte_row_width_chroma_ub[k],
+				&p->dpte_row_height_chroma[k],
+				&p->dpte_row_height_linear_chroma[k],
+				&s->PixelPTEBytesPerRowC_one_row_per_frame[k],
+				&s->dpte_row_width_chroma_ub_one_row_per_frame[k],
+				&s->dpte_row_height_chroma_one_row_per_frame[k],
+				&p->meta_req_width_chroma[k],
+				&p->meta_req_height_chroma[k],
+				&p->meta_row_width_chroma[k],
+				&p->meta_row_height_chroma[k],
+				&p->PixelPTEReqWidthC[k],
+				&p->PixelPTEReqHeightC[k],
+				&p->PTERequestSizeC[k],
+				&p->dpde0_bytes_per_frame_ub_c[k],
+				&p->meta_pte_bytes_per_frame_ub_c[k]);
+
+			p->PrefetchSourceLinesC[k] = CalculatePrefetchSourceLines (
+				p->myPipe[k].VRatioChroma,
+				p->myPipe[k].VTapsChroma,
+				p->myPipe[k].InterlaceEnable,
+				p->myPipe[k].ProgressiveToInterlaceUnitInOPP,
+				p->myPipe[k].SwathHeightC,
+				p->myPipe[k].SourceScan,
+				p->myPipe[k].ViewportStationary,
+				p->SwathWidthC[k],
+				p->myPipe[k].ViewportHeightChroma,
+				p->myPipe[k].ViewportXStartC,
+				p->myPipe[k].ViewportYStartC,
+
+				// Output
+				&p->VInitPreFillC[k],
+				&p->MaxNumSwathC[k]);
+		} else {
+			s->PTEBufferSizeInRequestsForLuma[k] = p->PTEBufferSizeInRequestsLuma + p->PTEBufferSizeInRequestsChroma;
+			s->PTEBufferSizeInRequestsForChroma[k] = 0;
+			s->PixelPTEBytesPerRowC[k] = 0;
+			s->PixelPTEBytesPerRowStorageC[k] = 0;
+			s->PDEAndMetaPTEBytesFrameC = 0;
+			s->MetaRowByteC[k] = 0;
+			p->MaxNumSwathC[k] = 0;
+			p->PrefetchSourceLinesC[k] = 0;
+			s->dpte_row_height_chroma_one_row_per_frame[k] = 0;
+			s->dpte_row_width_chroma_ub_one_row_per_frame[k] = 0;
+			s->PixelPTEBytesPerRowC_one_row_per_frame[k] = 0;
+		}
+
+		s->PDEAndMetaPTEBytesFrameY = CalculateVMAndRowBytes(
+			p->myPipe[k].ViewportStationary,
+			p->myPipe[k].DCCEnable,
+			p->myPipe[k].DPPPerSurface,
+			p->myPipe[k].BlockHeight256BytesY,
+			p->myPipe[k].BlockWidth256BytesY,
+			p->myPipe[k].SourcePixelFormat,
+			p->myPipe[k].SurfaceTiling,
+			p->myPipe[k].BytePerPixelY,
+			p->myPipe[k].SourceScan,
+			p->SwathWidthY[k],
+			p->myPipe[k].ViewportHeight,
+			p->myPipe[k].ViewportXStart,
+			p->myPipe[k].ViewportYStart,
+			p->GPUVMEnable,
+			p->GPUVMMaxPageTableLevels,
+			p->GPUVMMinPageSizeKBytes[k],
+			s->PTEBufferSizeInRequestsForLuma[k],
+			p->myPipe[k].PitchY,
+			p->myPipe[k].DCCMetaPitchY,
+			p->myPipe[k].BlockWidthY,
+			p->myPipe[k].BlockHeightY,
+
+			// Output
+			&s->MetaRowByteY[k],
+			&s->PixelPTEBytesPerRowY[k],
+			&s->PixelPTEBytesPerRowStorageY[k],
+			&p->dpte_row_width_luma_ub[k],
+			&p->dpte_row_height_luma[k],
+			&p->dpte_row_height_linear_luma[k],
+			&s->PixelPTEBytesPerRowY_one_row_per_frame[k],
+			&s->dpte_row_width_luma_ub_one_row_per_frame[k],
+			&s->dpte_row_height_luma_one_row_per_frame[k],
+			&p->meta_req_width[k],
+			&p->meta_req_height[k],
+			&p->meta_row_width[k],
+			&p->meta_row_height[k],
+			&p->PixelPTEReqWidthY[k],
+			&p->PixelPTEReqHeightY[k],
+			&p->PTERequestSizeY[k],
+			&p->dpde0_bytes_per_frame_ub_l[k],
+			&p->meta_pte_bytes_per_frame_ub_l[k]);
+
+			p->PrefetchSourceLinesY[k] = CalculatePrefetchSourceLines(
+			p->myPipe[k].VRatio,
+			p->myPipe[k].VTaps,
+			p->myPipe[k].InterlaceEnable,
+			p->myPipe[k].ProgressiveToInterlaceUnitInOPP,
+			p->myPipe[k].SwathHeightY,
+			p->myPipe[k].SourceScan,
+			p->myPipe[k].ViewportStationary,
+			p->SwathWidthY[k],
+			p->myPipe[k].ViewportHeight,
+			p->myPipe[k].ViewportXStart,
+			p->myPipe[k].ViewportYStart,
+
+			// Output
+			&p->VInitPreFillY[k],
+			&p->MaxNumSwathY[k]);
+
+		p->PDEAndMetaPTEBytesFrame[k] = (s->PDEAndMetaPTEBytesFrameY + s->PDEAndMetaPTEBytesFrameC) * (1 + 8 * s->HostVMDynamicLevels);
+		p->MetaRowByte[k] = s->MetaRowByteY[k] + s->MetaRowByteC[k];
+
+		if (s->PixelPTEBytesPerRowStorageY[k] <= 64 * s->PTEBufferSizeInRequestsForLuma[k] && s->PixelPTEBytesPerRowStorageC[k] <= 64 * s->PTEBufferSizeInRequestsForChroma[k]) {
+			p->PTEBufferSizeNotExceeded[k] = true;
+		} else {
+			p->PTEBufferSizeNotExceeded[k] = false;
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: k=%u, PixelPTEBytesPerRowY = %u\n", __func__, k, s->PixelPTEBytesPerRowY[k]);
+			dml_print("DML::%s: k=%u, PixelPTEBytesPerRowC = %u\n", __func__, k, s->PixelPTEBytesPerRowC[k]);
+			dml_print("DML::%s: k=%u, PixelPTEBytesPerRowStorageY = %u\n", __func__, k, s->PixelPTEBytesPerRowStorageY[k]);
+			dml_print("DML::%s: k=%u, PixelPTEBytesPerRowStorageC = %u\n", __func__, k, s->PixelPTEBytesPerRowStorageC[k]);
+			dml_print("DML::%s: k=%u, PTEBufferSizeInRequestsForLuma = %u\n", __func__, k, s->PTEBufferSizeInRequestsForLuma[k]);
+			dml_print("DML::%s: k=%u, PTEBufferSizeInRequestsForChroma = %u\n", __func__, k, s->PTEBufferSizeInRequestsForChroma[k]);
+			dml_print("DML::%s: k=%u, PTEBufferSizeNotExceeded          = %u\n",  __func__, k, p->PTEBufferSizeNotExceeded[k]);
+#endif
+		}
+		s->one_row_per_frame_fits_in_buffer[k] = (s->PixelPTEBytesPerRowY_one_row_per_frame[k] <= 64 * 2 * s->PTEBufferSizeInRequestsForLuma[k] &&
+			s->PixelPTEBytesPerRowC_one_row_per_frame[k] <= 64 * 2 * s->PTEBufferSizeInRequestsForChroma[k]);
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%u, PDEAndMetaPTEBytesFrame = %u\n", __func__, k, p->PDEAndMetaPTEBytesFrame[k]);
+		dml_print("DML::%s: k=%u, PDEAndMetaPTEBytesFrameY = %u\n", __func__, k, s->PDEAndMetaPTEBytesFrameY);
+		dml_print("DML::%s: k=%u, PDEAndMetaPTEBytesFrameC = %u\n", __func__, k, s->PDEAndMetaPTEBytesFrameC);
+		dml_print("DML::%s: k=%u, HostVMDynamicLevels = %u\n", __func__, k, s->HostVMDynamicLevels);
+		dml_print("DML::%s: k=%u, one_row_per_frame_fits_in_buffer = %u\n", __func__, k, s->one_row_per_frame_fits_in_buffer[k]);
+		dml_print("DML::%s: k=%u, PixelPTEBytesPerRowY_one_row_per_frame = %u\n", __func__, k, s->PixelPTEBytesPerRowY_one_row_per_frame[k]);
+		dml_print("DML::%s: k=%u, PixelPTEBytesPerRowC_one_row_per_frame    = %u\n",  __func__, k, s->PixelPTEBytesPerRowC_one_row_per_frame[k]);
+#endif
+	}
+
+	CalculateMALLUseForStaticScreen(
+		p->NumberOfActiveSurfaces,
+		p->MALLAllocatedForDCN,
+		p->UseMALLForStaticScreen,   // mode
+		p->SurfaceSizeInMALL,
+		s->one_row_per_frame_fits_in_buffer,
+		// Output
+		p->UsesMALLForStaticScreen); // boolen
+
+	for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+		if (p->PTEBufferModeOverrideEn[k] == 1) {
+			p->PTE_BUFFER_MODE[k] = p->PTEBufferModeOverrideVal[k];
+		}
+		p->PTE_BUFFER_MODE[k] = p->myPipe[k].FORCE_ONE_ROW_FOR_FRAME || p->UsesMALLForStaticScreen[k] || (p->UseMALLForPStateChange[k] == dml_use_mall_pstate_change_sub_viewport) ||
+			(p->UseMALLForPStateChange[k] == dml_use_mall_pstate_change_phantom_pipe) || (p->GPUVMMinPageSizeKBytes[k] > 64);
+		p->BIGK_FRAGMENT_SIZE[k] = (dml_uint_t)(dml_log2(p->GPUVMMinPageSizeKBytes[k] * 1024) - 12);
+	}
+
+	for (dml_uint_t k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%u, SurfaceSizeInMALL         = %u\n",  __func__, k, p->SurfaceSizeInMALL[k]);
+		dml_print("DML::%s: k=%u, UsesMALLForStaticScreen   = %u\n",  __func__, k, p->UsesMALLForStaticScreen[k]);
+#endif
+		p->use_one_row_for_frame[k] = p->myPipe[k].FORCE_ONE_ROW_FOR_FRAME || p->UsesMALLForStaticScreen[k] || (p->UseMALLForPStateChange[k] == dml_use_mall_pstate_change_sub_viewport) ||
+									(p->UseMALLForPStateChange[k] == dml_use_mall_pstate_change_phantom_pipe) || (p->GPUVMMinPageSizeKBytes[k] > 64 && dml_is_vertical_rotation(p->myPipe[k].SourceScan));
+
+		p->use_one_row_for_frame_flip[k] = p->use_one_row_for_frame[k] && !(p->UseMALLForPStateChange[k] == dml_use_mall_pstate_change_full_frame);
+
+		if (p->use_one_row_for_frame[k]) {
+			p->dpte_row_height_luma[k] = s->dpte_row_height_luma_one_row_per_frame[k];
+			p->dpte_row_width_luma_ub[k] = s->dpte_row_width_luma_ub_one_row_per_frame[k];
+			s->PixelPTEBytesPerRowY[k] = s->PixelPTEBytesPerRowY_one_row_per_frame[k];
+			p->dpte_row_height_chroma[k] = s->dpte_row_height_chroma_one_row_per_frame[k];
+			p->dpte_row_width_chroma_ub[k] = s->dpte_row_width_chroma_ub_one_row_per_frame[k];
+			s->PixelPTEBytesPerRowC[k] = s->PixelPTEBytesPerRowC_one_row_per_frame[k];
+			p->PTEBufferSizeNotExceeded[k] = s->one_row_per_frame_fits_in_buffer[k];
+		}
+
+		if (p->MetaRowByte[k] <= p->DCCMetaBufferSizeBytes) {
+			p->DCCMetaBufferSizeNotExceeded[k] = true;
+		} else {
+			p->DCCMetaBufferSizeNotExceeded[k] = false;
+
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: k=%u, MetaRowByte                   = %u\n",  __func__, k, p->MetaRowByte[k]);
+			dml_print("DML::%s: k=%u, DCCMetaBufferSizeBytes        = %u\n",  __func__, k, p->DCCMetaBufferSizeBytes);
+			dml_print("DML::%s: k=%u, DCCMetaBufferSizeNotExceeded  = %u\n",  __func__, k, p->DCCMetaBufferSizeNotExceeded[k]);
+#endif
+		}
+		s->PixelPTEBytesPerRowY[k] = s->PixelPTEBytesPerRowY[k] * (1 + 8 * s->HostVMDynamicLevels);
+		s->PixelPTEBytesPerRowC[k] = s->PixelPTEBytesPerRowC[k] * (1 + 8 * s->HostVMDynamicLevels);
+		p->PixelPTEBytesPerRow[k] = s->PixelPTEBytesPerRowY[k] + s->PixelPTEBytesPerRowC[k];
+		if (p->use_one_row_for_frame[k])
+			p->PixelPTEBytesPerRow[k] = p->PixelPTEBytesPerRow[k] / 2;
+
+		CalculateRowBandwidth(
+			p->GPUVMEnable,
+			p->myPipe[k].SourcePixelFormat,
+			p->myPipe[k].VRatio,
+			p->myPipe[k].VRatioChroma,
+			p->myPipe[k].DCCEnable,
+			p->myPipe[k].HTotal / p->myPipe[k].PixelClock,
+			s->MetaRowByteY[k],
+			s->MetaRowByteC[k],
+			p->meta_row_height[k],
+			p->meta_row_height_chroma[k],
+			s->PixelPTEBytesPerRowY[k],
+			s->PixelPTEBytesPerRowC[k],
+			p->dpte_row_height_luma[k],
+			p->dpte_row_height_chroma[k],
+
+			// Output
+			&p->meta_row_bw[k],
+			&p->dpte_row_bw[k]);
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%u, use_one_row_for_frame = %u\n", __func__, k, p->use_one_row_for_frame[k]);
+		dml_print("DML::%s: k=%u, use_one_row_for_frame_flip = %u\n", __func__, k, p->use_one_row_for_frame_flip[k]);
+		dml_print("DML::%s: k=%u, UseMALLForPStateChange = %u\n", __func__, k, p->UseMALLForPStateChange[k]);
+		dml_print("DML::%s: k=%u, dpte_row_height_luma = %u\n", __func__, k, p->dpte_row_height_luma[k]);
+		dml_print("DML::%s: k=%u, dpte_row_width_luma_ub = %u\n", __func__, k, p->dpte_row_width_luma_ub[k]);
+		dml_print("DML::%s: k=%u, PixelPTEBytesPerRowY = %u\n", __func__, k, s->PixelPTEBytesPerRowY[k]);
+		dml_print("DML::%s: k=%u, dpte_row_height_chroma = %u\n", __func__, k, p->dpte_row_height_chroma[k]);
+		dml_print("DML::%s: k=%u, dpte_row_width_chroma_ub = %u\n", __func__, k, p->dpte_row_width_chroma_ub[k]);
+		dml_print("DML::%s: k=%u, PixelPTEBytesPerRowC = %u\n", __func__, k, s->PixelPTEBytesPerRowC[k]);
+		dml_print("DML::%s: k=%u, PixelPTEBytesPerRow = %u\n", __func__, k, p->PixelPTEBytesPerRow[k]);
+		dml_print("DML::%s: k=%u, PTEBufferSizeNotExceeded = %u\n", __func__, k, p->PTEBufferSizeNotExceeded[k]);
+		dml_print("DML::%s: k=%u, PTE_BUFFER_MODE = %u\n", __func__, k, p->PTE_BUFFER_MODE[k]);
+		dml_print("DML::%s: k=%u, BIGK_FRAGMENT_SIZE     = %u\n", __func__, k, p->BIGK_FRAGMENT_SIZE[k]);
+#endif
+	}
+}
+
+static void CalculateOutputLink(
+		dml_float_t PHYCLKPerState,
+		dml_float_t PHYCLKD18PerState,
+		dml_float_t PHYCLKD32PerState,
+		dml_float_t Downspreading,
+		dml_bool_t IsMainSurfaceUsingTheIndicatedTiming,
+		enum dml_output_encoder_class Output,
+		enum dml_output_format_class OutputFormat,
+		dml_uint_t HTotal,
+		dml_uint_t HActive,
+		dml_float_t PixelClockBackEnd,
+		dml_float_t ForcedOutputLinkBPP,
+		dml_uint_t DSCInputBitPerComponent,
+		dml_uint_t NumberOfDSCSlices,
+		dml_float_t AudioSampleRate,
+		dml_uint_t AudioSampleLayout,
+		enum dml_odm_mode ODMModeNoDSC,
+		enum dml_odm_mode ODMModeDSC,
+		enum dml_dsc_enable DSCEnable,
+		dml_uint_t OutputLinkDPLanes,
+		enum dml_output_link_dp_rate OutputLinkDPRate,
+
+		// Output
+		dml_bool_t *RequiresDSC,
+		dml_bool_t *RequiresFEC,
+		dml_float_t *OutBpp,
+		enum dml_output_type_and_rate__type *OutputType,
+		enum dml_output_type_and_rate__rate *OutputRate,
+		dml_uint_t *RequiredSlots)
+{
+	dml_bool_t LinkDSCEnable;
+	dml_uint_t dummy;
+	*RequiresDSC = false;
+	*RequiresFEC = false;
+	*OutBpp = 0;
+
+	*OutputType = dml_output_type_unknown;
+	*OutputRate = dml_output_rate_unknown;
+
+	if (IsMainSurfaceUsingTheIndicatedTiming) {
+		if (Output == dml_hdmi) {
+			*RequiresDSC = false;
+			*RequiresFEC = false;
+			*OutBpp = TruncToValidBPP(dml_min(600, PHYCLKPerState) * 10, 3, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, false, Output,
+									OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (dml_uint_t)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, &dummy);
+			//OutputTypeAndRate = "HDMI";
+			*OutputType = dml_output_type_hdmi;
+
+		} else if (Output == dml_dp || Output == dml_dp2p0 || Output == dml_edp) {
+			if (DSCEnable == dml_dsc_enable) {
+				*RequiresDSC = true;
+				LinkDSCEnable = true;
+				if (Output == dml_dp || Output == dml_dp2p0) {
+					*RequiresFEC = true;
+				} else {
+					*RequiresFEC = false;
+				}
+			} else {
+				*RequiresDSC = false;
+				LinkDSCEnable = false;
+				if (Output == dml_dp2p0) {
+					*RequiresFEC = true;
+				} else {
+					*RequiresFEC = false;
+				}
+			}
+			if (Output == dml_dp2p0) {
+				*OutBpp = 0;
+				if ((OutputLinkDPRate == dml_dp_rate_na || OutputLinkDPRate == dml_dp_rate_uhbr10) && PHYCLKD32PerState >= 10000 / 32.0) {
+					*OutBpp = TruncToValidBPP((1 - Downspreading / 100) * 10000, OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output,
+												OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (dml_uint_t)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+					if (*OutBpp == 0 && PHYCLKD32PerState < 13500 / 32.0 && DSCEnable == dml_dsc_enable_if_necessary && ForcedOutputLinkBPP == 0) {
+						*RequiresDSC = true;
+						LinkDSCEnable = true;
+						*OutBpp = TruncToValidBPP((1 - Downspreading / 100) * 10000, OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output,
+													OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (dml_uint_t)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+					}
+					//OutputTypeAndRate = Output & " UHBR10";
+					*OutputType = dml_output_type_dp2p0;
+					*OutputRate = dml_output_rate_dp_rate_uhbr10;
+				}
+				if ((OutputLinkDPRate == dml_dp_rate_na || OutputLinkDPRate == dml_dp_rate_uhbr13p5) && *OutBpp == 0 && PHYCLKD32PerState >= 13500 / 32.0) {
+					*OutBpp = TruncToValidBPP((1 - Downspreading / 100) * 13500, OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output,
+												OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (dml_uint_t)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+
+					if (*OutBpp == 0 && PHYCLKD32PerState < 20000 / 32 && DSCEnable == dml_dsc_enable_if_necessary && ForcedOutputLinkBPP == 0) {
+						*RequiresDSC = true;
+						LinkDSCEnable = true;
+						*OutBpp = TruncToValidBPP((1 - Downspreading / 100) * 13500, OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output,
+												OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (dml_uint_t)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+					}
+					//OutputTypeAndRate = Output & " UHBR13p5";
+					*OutputType = dml_output_type_dp2p0;
+					*OutputRate = dml_output_rate_dp_rate_uhbr13p5;
+				}
+				if ((OutputLinkDPRate == dml_dp_rate_na || OutputLinkDPRate == dml_dp_rate_uhbr20) && *OutBpp == 0 && PHYCLKD32PerState >= 20000 / 32) {
+					*OutBpp = TruncToValidBPP((1 - Downspreading / 100) * 20000, OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output,
+											OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (dml_uint_t)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+					if (*OutBpp == 0 && DSCEnable == dml_dsc_enable_if_necessary && ForcedOutputLinkBPP == 0) {
+						*RequiresDSC = true;
+						LinkDSCEnable = true;
+						*OutBpp = TruncToValidBPP((1 - Downspreading / 100) * 20000, OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output,
+												OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (dml_uint_t)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+					}
+					//OutputTypeAndRate = Output & " UHBR20";
+					*OutputType = dml_output_type_dp2p0;
+					*OutputRate = dml_output_rate_dp_rate_uhbr20;
+				}
+			} else { // output is dp or edp
+				*OutBpp = 0;
+				if ((OutputLinkDPRate == dml_dp_rate_na || OutputLinkDPRate == dml_dp_rate_hbr) && PHYCLKPerState >= 270) {
+					*OutBpp = TruncToValidBPP((1 - Downspreading / 100) * 2700, OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output,
+											OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (dml_uint_t)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+					if (*OutBpp == 0 && PHYCLKPerState < 540 && DSCEnable == dml_dsc_enable_if_necessary && ForcedOutputLinkBPP == 0) {
+						*RequiresDSC = true;
+						LinkDSCEnable = true;
+						if (Output == dml_dp) {
+							*RequiresFEC = true;
+						}
+						*OutBpp = TruncToValidBPP((1 - Downspreading / 100) * 2700, OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output,
+												OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (dml_uint_t)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+					}
+					//OutputTypeAndRate = Output & " HBR";
+					*OutputType = (Output == dml_dp) ? dml_output_type_dp : dml_output_type_edp;
+					*OutputRate = dml_output_rate_dp_rate_hbr;
+				}
+				if ((OutputLinkDPRate == dml_dp_rate_na || OutputLinkDPRate == dml_dp_rate_hbr2) && *OutBpp == 0 && PHYCLKPerState >= 540) {
+					*OutBpp = TruncToValidBPP((1 - Downspreading / 100) * 5400, OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output,
+											OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (dml_uint_t)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+
+					if (*OutBpp == 0 && PHYCLKPerState < 810 && DSCEnable == dml_dsc_enable_if_necessary && ForcedOutputLinkBPP == 0) {
+						*RequiresDSC = true;
+						LinkDSCEnable = true;
+						if (Output == dml_dp) {
+							*RequiresFEC = true;
+						}
+						*OutBpp = TruncToValidBPP((1 - Downspreading / 100) * 5400, OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output,
+												OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (dml_uint_t)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+					}
+					//OutputTypeAndRate = Output & " HBR2";
+					*OutputType = (Output == dml_dp) ? dml_output_type_dp : dml_output_type_edp;
+					*OutputRate = dml_output_rate_dp_rate_hbr2;
+				}
+				if ((OutputLinkDPRate == dml_dp_rate_na || OutputLinkDPRate == dml_dp_rate_hbr3) && *OutBpp == 0 && PHYCLKPerState >= 810) { // VBA_ERROR, vba code doesn't have hbr3 check
+					*OutBpp = TruncToValidBPP((1 - Downspreading / 100) * 8100, OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output,
+											OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (dml_uint_t)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+
+					if (*OutBpp == 0 && DSCEnable == dml_dsc_enable_if_necessary && ForcedOutputLinkBPP == 0) {
+						*RequiresDSC = true;
+						LinkDSCEnable = true;
+						if (Output == dml_dp) {
+							*RequiresFEC = true;
+						}
+						*OutBpp = TruncToValidBPP((1 - Downspreading / 100) * 8100, OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output,
+												OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (dml_uint_t)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+					}
+					//OutputTypeAndRate = Output & " HBR3";
+					*OutputType = (Output == dml_dp) ? dml_output_type_dp : dml_output_type_edp;
+					*OutputRate = dml_output_rate_dp_rate_hbr3;
+				}
+			}
+		}
+	}
+}
+
+/// @brief Determine the ODM mode and number of DPP used per plane based on dispclk, dsc usage, odm usage policy
+static void CalculateODMMode(
+		dml_uint_t MaximumPixelsPerLinePerDSCUnit,
+		dml_uint_t HActive,
+		enum dml_output_encoder_class Output,
+		enum dml_output_format_class OutputFormat,
+		enum dml_odm_use_policy ODMUse,
+		dml_float_t StateDispclk,
+		dml_float_t MaxDispclk,
+		dml_bool_t DSCEnable,
+		dml_uint_t TotalNumberOfActiveDPP,
+		dml_uint_t MaxNumDPP,
+		dml_float_t PixelClock,
+		dml_float_t DISPCLKDPPCLKDSCCLKDownSpreading,
+		dml_float_t DISPCLKRampingMargin,
+		dml_float_t DISPCLKDPPCLKVCOSpeed,
+		dml_uint_t NumberOfDSCSlices,
+
+		// Output
+		dml_bool_t *TotalAvailablePipesSupport,
+		dml_uint_t *NumberOfDPP,
+		enum dml_odm_mode *ODMMode,
+		dml_float_t *RequiredDISPCLKPerSurface)
+{
+
+	dml_float_t SurfaceRequiredDISPCLKWithoutODMCombine;
+	dml_float_t SurfaceRequiredDISPCLKWithODMCombineTwoToOne;
+	dml_float_t SurfaceRequiredDISPCLKWithODMCombineFourToOne;
+
+	SurfaceRequiredDISPCLKWithoutODMCombine = CalculateRequiredDispclk(dml_odm_mode_bypass, PixelClock, DISPCLKDPPCLKDSCCLKDownSpreading, DISPCLKRampingMargin, DISPCLKDPPCLKVCOSpeed, MaxDispclk);
+	SurfaceRequiredDISPCLKWithODMCombineTwoToOne = CalculateRequiredDispclk(dml_odm_mode_combine_2to1, PixelClock, DISPCLKDPPCLKDSCCLKDownSpreading, DISPCLKRampingMargin, DISPCLKDPPCLKVCOSpeed, MaxDispclk);
+	SurfaceRequiredDISPCLKWithODMCombineFourToOne = CalculateRequiredDispclk(dml_odm_mode_combine_4to1, PixelClock, DISPCLKDPPCLKDSCCLKDownSpreading, DISPCLKRampingMargin, DISPCLKDPPCLKVCOSpeed, MaxDispclk);
+	*TotalAvailablePipesSupport = true;
+
+	if (OutputFormat == dml_420) {
+		if (HActive > 4 * DML2_MAX_FMT_420_BUFFER_WIDTH)
+			*TotalAvailablePipesSupport = false;
+		else if (HActive > 2 * DML2_MAX_FMT_420_BUFFER_WIDTH)
+			ODMUse = dml_odm_use_policy_combine_4to1;
+		else if (HActive > DML2_MAX_FMT_420_BUFFER_WIDTH && ODMUse != dml_odm_use_policy_combine_4to1)
+			ODMUse = dml_odm_use_policy_combine_2to1;
+		if (Output == dml_hdmi && ODMUse == dml_odm_use_policy_combine_2to1)
+			*TotalAvailablePipesSupport = false;
+		if ((Output == dml_hdmi || Output == dml_dp || Output == dml_edp) && ODMUse == dml_odm_use_policy_combine_4to1)
+			*TotalAvailablePipesSupport = false;
+	}
+
+	if (ODMUse == dml_odm_use_policy_bypass || ODMUse == dml_odm_use_policy_combine_as_needed)
+		*ODMMode = dml_odm_mode_bypass;
+	else if (ODMUse == dml_odm_use_policy_combine_2to1)
+		*ODMMode = dml_odm_mode_combine_2to1;
+	else if (ODMUse == dml_odm_use_policy_combine_4to1)
+		*ODMMode = dml_odm_mode_combine_4to1;
+	else if (ODMUse == dml_odm_use_policy_split_1to2)
+		*ODMMode = dml_odm_mode_split_1to2;
+	else if (ODMUse == dml_odm_use_policy_mso_1to2)
+		*ODMMode = dml_odm_mode_mso_1to2;
+	else if (ODMUse == dml_odm_use_policy_mso_1to4)
+		*ODMMode = dml_odm_mode_mso_1to4;
+
+	*RequiredDISPCLKPerSurface = SurfaceRequiredDISPCLKWithoutODMCombine;
+	*NumberOfDPP = 0;
+
+	if (!(Output == dml_hdmi || Output == dml_dp || Output == dml_edp) && (ODMUse == dml_odm_use_policy_combine_4to1 || (ODMUse == dml_odm_use_policy_combine_as_needed &&
+		(SurfaceRequiredDISPCLKWithODMCombineTwoToOne > StateDispclk || (DSCEnable && (HActive > 2 * MaximumPixelsPerLinePerDSCUnit)) || NumberOfDSCSlices > 8)))) {
+		if (TotalNumberOfActiveDPP + 4 <= MaxNumDPP) {
+			*ODMMode = dml_odm_mode_combine_4to1;
+			*RequiredDISPCLKPerSurface = SurfaceRequiredDISPCLKWithODMCombineFourToOne;
+			*NumberOfDPP = 4;
+		} else {
+			*TotalAvailablePipesSupport = false;
+		}
+	} else if (Output != dml_hdmi && (ODMUse == dml_odm_use_policy_combine_2to1 || (ODMUse == dml_odm_use_policy_combine_as_needed &&
+				((SurfaceRequiredDISPCLKWithoutODMCombine > StateDispclk && SurfaceRequiredDISPCLKWithODMCombineTwoToOne <= StateDispclk) ||
+				(DSCEnable && (HActive > MaximumPixelsPerLinePerDSCUnit)) || (NumberOfDSCSlices <= 8 && NumberOfDSCSlices > 4))))) {
+		if (TotalNumberOfActiveDPP + 2 <= MaxNumDPP) {
+			*ODMMode = dml_odm_mode_combine_2to1;
+			*RequiredDISPCLKPerSurface = SurfaceRequiredDISPCLKWithODMCombineTwoToOne;
+			*NumberOfDPP = 2;
+		} else {
+			*TotalAvailablePipesSupport = false;
+		}
+	} else {
+		if (TotalNumberOfActiveDPP + 1 <= MaxNumDPP) {
+			*NumberOfDPP = 1;
+		} else {
+			*TotalAvailablePipesSupport = false;
+		}
+	}
+}
+
+/// @brief Calculate the required DISPCLK given the odm mode and pixclk
+static dml_float_t CalculateRequiredDispclk(
+		enum dml_odm_mode ODMMode,
+		dml_float_t PixelClock,
+		dml_float_t DISPCLKDPPCLKDSCCLKDownSpreading,
+		dml_float_t DISPCLKRampingMargin,
+		dml_float_t DISPCLKDPPCLKVCOSpeed,
+		dml_float_t MaxDispclk)
+{
+	dml_float_t RequiredDispclk = 0.;
+	dml_float_t PixelClockAfterODM;
+
+	dml_float_t DISPCLKWithRampingRoundedToDFSGranularity;
+	dml_float_t DISPCLKWithoutRampingRoundedToDFSGranularity;
+	dml_float_t MaxDispclkRoundedDownToDFSGranularity;
+
+	if (ODMMode == dml_odm_mode_combine_4to1) {
+		PixelClockAfterODM = PixelClock / 4;
+	} else if (ODMMode == dml_odm_mode_combine_2to1) {
+		PixelClockAfterODM = PixelClock / 2;
+	} else {
+		PixelClockAfterODM = PixelClock;
+	}
+
+	DISPCLKWithRampingRoundedToDFSGranularity = RoundToDFSGranularity(PixelClockAfterODM * (1.0 + DISPCLKDPPCLKDSCCLKDownSpreading / 100.0) * (1 + DISPCLKRampingMargin / 100.0), 1, DISPCLKDPPCLKVCOSpeed);
+	DISPCLKWithoutRampingRoundedToDFSGranularity = RoundToDFSGranularity(PixelClockAfterODM * (1.0 + DISPCLKDPPCLKDSCCLKDownSpreading / 100.0), 1, DISPCLKDPPCLKVCOSpeed);
+	MaxDispclkRoundedDownToDFSGranularity = RoundToDFSGranularity(MaxDispclk, 0, DISPCLKDPPCLKVCOSpeed);
+
+	if (DISPCLKWithoutRampingRoundedToDFSGranularity > MaxDispclkRoundedDownToDFSGranularity) {
+		RequiredDispclk = DISPCLKWithoutRampingRoundedToDFSGranularity;
+	} else if (DISPCLKWithRampingRoundedToDFSGranularity > MaxDispclkRoundedDownToDFSGranularity) {
+		RequiredDispclk = MaxDispclkRoundedDownToDFSGranularity;
+	} else {
+		RequiredDispclk = DISPCLKWithRampingRoundedToDFSGranularity;
+	}
+
+	return RequiredDispclk;
+}
+
+/// @brief Determine DPPCLK if there only one DPP per plane, main factor is the pixel rate and DPP scaling parameter
+static void CalculateSinglePipeDPPCLKAndSCLThroughput(
+		dml_float_t HRatio,
+		dml_float_t HRatioChroma,
+		dml_float_t VRatio,
+		dml_float_t VRatioChroma,
+		dml_float_t MaxDCHUBToPSCLThroughput,
+		dml_float_t MaxPSCLToLBThroughput,
+		dml_float_t PixelClock,
+		enum dml_source_format_class SourcePixelFormat,
+		dml_uint_t HTaps,
+		dml_uint_t HTapsChroma,
+		dml_uint_t VTaps,
+		dml_uint_t VTapsChroma,
+
+		// Output
+		dml_float_t *PSCL_THROUGHPUT,
+		dml_float_t *PSCL_THROUGHPUT_CHROMA,
+		dml_float_t *DPPCLKUsingSingleDPP)
+{
+	dml_float_t DPPCLKUsingSingleDPPLuma;
+	dml_float_t DPPCLKUsingSingleDPPChroma;
+
+	if (HRatio > 1) {
+		*PSCL_THROUGHPUT = dml_min(MaxDCHUBToPSCLThroughput, MaxPSCLToLBThroughput * HRatio / dml_ceil((dml_float_t) HTaps / 6.0, 1.0));
+	} else {
+		*PSCL_THROUGHPUT = dml_min(MaxDCHUBToPSCLThroughput, MaxPSCLToLBThroughput);
+	}
+
+	DPPCLKUsingSingleDPPLuma = PixelClock * dml_max3(VTaps / 6 * dml_min(1, HRatio), HRatio * VRatio / *PSCL_THROUGHPUT, 1);
+
+	if ((HTaps > 6 || VTaps > 6) && DPPCLKUsingSingleDPPLuma < 2 * PixelClock)
+		DPPCLKUsingSingleDPPLuma = 2 * PixelClock;
+
+	if ((SourcePixelFormat != dml_420_8 && SourcePixelFormat != dml_420_10 && SourcePixelFormat != dml_420_12 && SourcePixelFormat != dml_rgbe_alpha)) {
+		*PSCL_THROUGHPUT_CHROMA = 0;
+		*DPPCLKUsingSingleDPP = DPPCLKUsingSingleDPPLuma;
+	} else {
+		if (HRatioChroma > 1) {
+			*PSCL_THROUGHPUT_CHROMA = dml_min(MaxDCHUBToPSCLThroughput, MaxPSCLToLBThroughput * HRatioChroma / dml_ceil((dml_float_t) HTapsChroma / 6.0, 1.0));
+		} else {
+			*PSCL_THROUGHPUT_CHROMA = dml_min(MaxDCHUBToPSCLThroughput, MaxPSCLToLBThroughput);
+		}
+		DPPCLKUsingSingleDPPChroma = PixelClock * dml_max3(VTapsChroma / 6 * dml_min(1, HRatioChroma),
+															HRatioChroma * VRatioChroma / *PSCL_THROUGHPUT_CHROMA, 1);
+		if ((HTapsChroma > 6 || VTapsChroma > 6) && DPPCLKUsingSingleDPPChroma < 2 * PixelClock)
+			DPPCLKUsingSingleDPPChroma = 2 * PixelClock;
+		*DPPCLKUsingSingleDPP = dml_max(DPPCLKUsingSingleDPPLuma, DPPCLKUsingSingleDPPChroma);
+	}
+}
+
+/// @brief Calculate the actual dppclk freq
+/// @param DPPCLKUsingSingleDPP DppClk freq required if there is only 1 DPP per plane
+/// @param DPPPerSurface Number of DPP for each plane
+static void CalculateDPPCLK(
+		dml_uint_t NumberOfActiveSurfaces,
+		dml_float_t DISPCLKDPPCLKDSCCLKDownSpreading,
+		dml_float_t DISPCLKDPPCLKVCOSpeed,
+		dml_float_t DPPCLKUsingSingleDPP[],
+		dml_uint_t DPPPerSurface[],
+
+		// Output
+		dml_float_t *GlobalDPPCLK,
+		dml_float_t Dppclk[])
+{
+	*GlobalDPPCLK = 0;
+	for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+		Dppclk[k] = DPPCLKUsingSingleDPP[k] / DPPPerSurface[k] * (1 + DISPCLKDPPCLKDSCCLKDownSpreading / 100.0);
+		*GlobalDPPCLK = dml_max(*GlobalDPPCLK, Dppclk[k]);
+	}
+	*GlobalDPPCLK = RoundToDFSGranularity(*GlobalDPPCLK, 1, DISPCLKDPPCLKVCOSpeed);
+
+	dml_print("DML::%s: GlobalDPPCLK = %f\n", __func__, *GlobalDPPCLK);
+	for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+		Dppclk[k] = *GlobalDPPCLK / 255.0 * dml_ceil(Dppclk[k] * 255.0 / *GlobalDPPCLK, 1.0);
+		dml_print("DML::%s: Dppclk[%0d] = %f\n", __func__, k, Dppclk[k]);
+	}
+}
+
+static void CalculateMALLUseForStaticScreen(
+		dml_uint_t NumberOfActiveSurfaces,
+		dml_uint_t MALLAllocatedForDCNFinal,
+		enum dml_use_mall_for_static_screen_mode *UseMALLForStaticScreen,
+		dml_uint_t SurfaceSizeInMALL[],
+		dml_bool_t one_row_per_frame_fits_in_buffer[],
+
+		// Output
+		dml_bool_t UsesMALLForStaticScreen[])
+{
+
+	dml_uint_t SurfaceToAddToMALL;
+	dml_bool_t CanAddAnotherSurfaceToMALL;
+	dml_uint_t TotalSurfaceSizeInMALL;
+
+	TotalSurfaceSizeInMALL = 0;
+	for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+		UsesMALLForStaticScreen[k] = (UseMALLForStaticScreen[k] == dml_use_mall_static_screen_enable);
+		if (UsesMALLForStaticScreen[k])
+			TotalSurfaceSizeInMALL = TotalSurfaceSizeInMALL + SurfaceSizeInMALL[k];
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%u, UsesMALLForStaticScreen = %u\n",  __func__, k, UsesMALLForStaticScreen[k]);
+		dml_print("DML::%s: k=%u, TotalSurfaceSizeInMALL = %u\n",  __func__, k, TotalSurfaceSizeInMALL);
+#endif
+	}
+
+	SurfaceToAddToMALL = 0;
+	CanAddAnotherSurfaceToMALL = true;
+	while (CanAddAnotherSurfaceToMALL) {
+		CanAddAnotherSurfaceToMALL = false;
+		for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+			if (TotalSurfaceSizeInMALL + SurfaceSizeInMALL[k] <= MALLAllocatedForDCNFinal * 1024 * 1024 &&
+				!UsesMALLForStaticScreen[k] && UseMALLForStaticScreen[k] != dml_use_mall_static_screen_disable && one_row_per_frame_fits_in_buffer[k] &&
+				(!CanAddAnotherSurfaceToMALL || SurfaceSizeInMALL[k] < SurfaceSizeInMALL[SurfaceToAddToMALL])) {
+				CanAddAnotherSurfaceToMALL = true;
+				SurfaceToAddToMALL = k;
+				dml_print("DML::%s: k=%u, UseMALLForStaticScreen = %u (dis, en, optimize)\n",  __func__, k, UseMALLForStaticScreen[k]);
+			}
+		}
+		if (CanAddAnotherSurfaceToMALL) {
+			UsesMALLForStaticScreen[SurfaceToAddToMALL] = true;
+			TotalSurfaceSizeInMALL = TotalSurfaceSizeInMALL + SurfaceSizeInMALL[SurfaceToAddToMALL];
+
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: SurfaceToAddToMALL       = %u\n",  __func__, SurfaceToAddToMALL);
+			dml_print("DML::%s: TotalSurfaceSizeInMALL   = %u\n",  __func__, TotalSurfaceSizeInMALL);
+#endif
+		}
+	}
+}
+
+// @brief Calculate return bw for VM only traffic
+dml_float_t dml_get_return_bw_mbps_vm_only(
+								const struct soc_bounding_box_st *soc,
+								dml_bool_t use_ideal_dram_bw_strobe,
+								dml_bool_t HostVMEnable,
+								dml_float_t DCFCLK,
+								dml_float_t FabricClock,
+								dml_float_t DRAMSpeed)
+{
+	dml_float_t VMDataOnlyReturnBW =
+		dml_min3(soc->return_bus_width_bytes * DCFCLK * soc->pct_ideal_sdp_bw_after_urgent / 100.0,
+				FabricClock * soc->fabric_datapath_to_dcn_data_return_bytes * soc->pct_ideal_sdp_bw_after_urgent / 100.0,
+				DRAMSpeed * soc->num_chans *  soc->dram_channel_width_bytes *
+				((use_ideal_dram_bw_strobe && !HostVMEnable) ? soc->pct_ideal_dram_bw_after_urgent_strobe :  soc->pct_ideal_dram_bw_after_urgent_vm_only) / 100.0);
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: use_ideal_dram_bw_strobe = %u\n", __func__, use_ideal_dram_bw_strobe);
+	dml_print("DML::%s: HostVMEnable = %u\n", __func__, HostVMEnable);
+	dml_print("DML::%s: DCFCLK = %f\n", __func__, DCFCLK);
+	dml_print("DML::%s: FabricClock = %f\n", __func__, FabricClock);
+	dml_print("DML::%s: DRAMSpeed = %f\n", __func__, DRAMSpeed);
+	dml_print("DML::%s: VMDataOnlyReturnBW = %f\n", __func__, VMDataOnlyReturnBW);
+#endif
+	return VMDataOnlyReturnBW;
+}
+
+// Function: dml_get_return_bw_mbps
+// Megabyte per second
+dml_float_t dml_get_return_bw_mbps(
+						const struct soc_bounding_box_st *soc,
+						dml_bool_t use_ideal_dram_bw_strobe,
+						dml_bool_t HostVMEnable,
+						dml_float_t DCFCLK,
+						dml_float_t FabricClock,
+						dml_float_t DRAMSpeed)
+{
+	dml_float_t ReturnBW = 0.;
+	dml_float_t IdealSDPPortBandwidth    = soc->return_bus_width_bytes * DCFCLK;
+	dml_float_t IdealFabricBandwidth     = FabricClock * soc->fabric_datapath_to_dcn_data_return_bytes;
+	dml_float_t IdealDRAMBandwidth       = DRAMSpeed * soc->num_chans * soc->dram_channel_width_bytes;
+	dml_float_t PixelDataOnlyReturnBW    = dml_min3(IdealSDPPortBandwidth * soc->pct_ideal_sdp_bw_after_urgent / 100,
+												IdealFabricBandwidth * soc->pct_ideal_fabric_bw_after_urgent / 100,
+												IdealDRAMBandwidth * ((use_ideal_dram_bw_strobe && !HostVMEnable) ? soc->pct_ideal_dram_bw_after_urgent_strobe :
+																						soc->pct_ideal_dram_bw_after_urgent_pixel_only) / 100);
+	dml_float_t PixelMixedWithVMDataReturnBW = dml_min3(IdealSDPPortBandwidth * soc->pct_ideal_sdp_bw_after_urgent / 100,
+													IdealFabricBandwidth * soc->pct_ideal_fabric_bw_after_urgent / 100,
+													IdealDRAMBandwidth * ((use_ideal_dram_bw_strobe && !HostVMEnable) ? soc->pct_ideal_dram_bw_after_urgent_strobe :
+																							soc->pct_ideal_dram_bw_after_urgent_pixel_and_vm) / 100);
+
+	if (HostVMEnable != true) {
+		ReturnBW = PixelDataOnlyReturnBW;
+	} else {
+		ReturnBW = PixelMixedWithVMDataReturnBW;
+	}
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: use_ideal_dram_bw_strobe = %u\n", __func__, use_ideal_dram_bw_strobe);
+	dml_print("DML::%s: HostVMEnable = %u\n", __func__, HostVMEnable);
+	dml_print("DML::%s: DCFCLK = %f\n", __func__, DCFCLK);
+	dml_print("DML::%s: FabricClock = %f\n", __func__, FabricClock);
+	dml_print("DML::%s: DRAMSpeed = %f\n", __func__, DRAMSpeed);
+	dml_print("DML::%s: IdealSDPPortBandwidth = %f\n", __func__, IdealSDPPortBandwidth);
+	dml_print("DML::%s: IdealFabricBandwidth = %f\n", __func__, IdealFabricBandwidth);
+	dml_print("DML::%s: IdealDRAMBandwidth = %f\n", __func__, IdealDRAMBandwidth);
+	dml_print("DML::%s: PixelDataOnlyReturnBW = %f\n", __func__, PixelDataOnlyReturnBW);
+	dml_print("DML::%s: PixelMixedWithVMDataReturnBW = %f\n", __func__, PixelMixedWithVMDataReturnBW);
+	dml_print("DML::%s: ReturnBW                  = %f MBps\n", __func__, ReturnBW);
+#endif
+	return ReturnBW;
+}
+
+// Function: dml_get_return_dram_bw_mbps
+// Megabyte per second
+static dml_float_t dml_get_return_dram_bw_mbps(
+						const struct soc_bounding_box_st *soc,
+						dml_bool_t use_ideal_dram_bw_strobe,
+						dml_bool_t HostVMEnable,
+						dml_float_t DRAMSpeed)
+{
+	dml_float_t ReturnDRAMBW = 0.;
+	dml_float_t IdealDRAMBandwidth       = DRAMSpeed * soc->num_chans * soc->dram_channel_width_bytes;
+	dml_float_t PixelDataOnlyReturnBW    = IdealDRAMBandwidth * ((use_ideal_dram_bw_strobe && !HostVMEnable) ? soc->pct_ideal_dram_bw_after_urgent_strobe :
+																						soc->pct_ideal_dram_bw_after_urgent_pixel_only) / 100;
+	dml_float_t PixelMixedWithVMDataReturnBW =  IdealDRAMBandwidth * ((use_ideal_dram_bw_strobe && !HostVMEnable) ? soc->pct_ideal_dram_bw_after_urgent_strobe :
+																							soc->pct_ideal_dram_bw_after_urgent_pixel_and_vm) / 100;
+
+	if (HostVMEnable != true) {
+		ReturnDRAMBW = PixelDataOnlyReturnBW;
+	} else {
+		ReturnDRAMBW = PixelMixedWithVMDataReturnBW;
+	}
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: use_ideal_dram_bw_strobe = %u\n", __func__, use_ideal_dram_bw_strobe);
+	dml_print("DML::%s: HostVMEnable = %u\n", __func__, HostVMEnable);
+	dml_print("DML::%s: DRAMSpeed = %f\n", __func__, DRAMSpeed);
+	dml_print("DML::%s: IdealDRAMBandwidth = %f\n", __func__, IdealDRAMBandwidth);
+	dml_print("DML::%s: PixelDataOnlyReturnBW = %f\n", __func__, PixelDataOnlyReturnBW);
+	dml_print("DML::%s: PixelMixedWithVMDataReturnBW = %f\n", __func__, PixelMixedWithVMDataReturnBW);
+	dml_print("DML::%s: ReturnDRAMBW                     = %f MBps\n", __func__, ReturnDRAMBW);
+#endif
+	return ReturnDRAMBW;
+}
+
+/// @brief BACKEND
+static dml_uint_t DSCDelayRequirement(
+						dml_bool_t DSCEnabled,
+						enum dml_odm_mode ODMMode,
+						dml_uint_t DSCInputBitPerComponent,
+						dml_float_t OutputBpp,
+						dml_uint_t HActive,
+						dml_uint_t HTotal,
+						dml_uint_t NumberOfDSCSlices,
+						enum dml_output_format_class OutputFormat,
+						enum dml_output_encoder_class Output,
+						dml_float_t PixelClock,
+						dml_float_t PixelClockBackEnd)
+{
+	dml_uint_t DSCDelayRequirement_val = 0;
+
+	if (DSCEnabled == true && OutputBpp != 0) {
+		if (ODMMode == dml_odm_mode_combine_4to1) {
+			DSCDelayRequirement_val = dscceComputeDelay(DSCInputBitPerComponent, OutputBpp, (dml_uint_t)(dml_ceil((dml_float_t) HActive / (dml_float_t) NumberOfDSCSlices, 1.0)),
+												(dml_uint_t) (NumberOfDSCSlices / 4.0), OutputFormat, Output) + dscComputeDelay(OutputFormat, Output);
+		} else if (ODMMode == dml_odm_mode_combine_2to1) {
+			DSCDelayRequirement_val = dscceComputeDelay(DSCInputBitPerComponent, OutputBpp, (dml_uint_t)(dml_ceil((dml_float_t) HActive / (dml_float_t) NumberOfDSCSlices, 1.0)),
+												(dml_uint_t) (NumberOfDSCSlices / 2.0), OutputFormat, Output) + dscComputeDelay(OutputFormat, Output);
+		} else {
+			DSCDelayRequirement_val = dscceComputeDelay(DSCInputBitPerComponent, OutputBpp, (dml_uint_t)((dml_float_t) dml_ceil(HActive / (dml_float_t) NumberOfDSCSlices, 1.0)),
+										NumberOfDSCSlices, OutputFormat, Output) + dscComputeDelay(OutputFormat, Output);
+		}
+		DSCDelayRequirement_val = (dml_uint_t)(DSCDelayRequirement_val + (HTotal - HActive) * dml_ceil((dml_float_t) DSCDelayRequirement_val / (dml_float_t) HActive, 1.0));
+		DSCDelayRequirement_val = (dml_uint_t)(DSCDelayRequirement_val * PixelClock / PixelClockBackEnd);
+
+	} else {
+		DSCDelayRequirement_val = 0;
+	}
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: DSCEnabled = %u\n", __func__, DSCEnabled);
+	dml_print("DML::%s: ODMMode = %u\n", __func__, ODMMode);
+	dml_print("DML::%s: OutputBpp = %f\n", __func__, OutputBpp);
+	dml_print("DML::%s: HActive = %u\n", __func__, HActive);
+	dml_print("DML::%s: HTotal = %u\n", __func__, HTotal);
+	dml_print("DML::%s: PixelClock = %f\n", __func__, PixelClock);
+	dml_print("DML::%s: PixelClockBackEnd = %f\n", __func__, PixelClockBackEnd);
+	dml_print("DML::%s: OutputFormat = %u\n", __func__, OutputFormat);
+	dml_print("DML::%s: DSCInputBitPerComponent = %u\n", __func__, DSCInputBitPerComponent);
+	dml_print("DML::%s: NumberOfDSCSlices = %u\n", __func__, NumberOfDSCSlices);
+	dml_print("DML::%s: DSCDelayRequirement_val = %u\n", __func__, DSCDelayRequirement_val);
+#endif
+
+	return DSCDelayRequirement_val;
+}
+
+static noinline_for_stack dml_bool_t CalculateVActiveBandwithSupport(dml_uint_t NumberOfActiveSurfaces,
+										dml_float_t ReturnBW,
+										dml_bool_t NotUrgentLatencyHiding[],
+										dml_float_t ReadBandwidthLuma[],
+										dml_float_t ReadBandwidthChroma[],
+										dml_float_t cursor_bw[],
+										dml_float_t meta_row_bandwidth[],
+										dml_float_t dpte_row_bandwidth[],
+										dml_uint_t NumberOfDPP[],
+										dml_float_t UrgentBurstFactorLuma[],
+										dml_float_t UrgentBurstFactorChroma[],
+										dml_float_t UrgentBurstFactorCursor[])
+{
+	dml_bool_t NotEnoughUrgentLatencyHiding = false;
+	dml_bool_t CalculateVActiveBandwithSupport_val = false;
+	dml_float_t VActiveBandwith = 0;
+
+	for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+		if (NotUrgentLatencyHiding[k]) {
+			NotEnoughUrgentLatencyHiding = true;
+		}
+	}
+
+	for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+		VActiveBandwith = VActiveBandwith + ReadBandwidthLuma[k] * UrgentBurstFactorLuma[k] + ReadBandwidthChroma[k] * UrgentBurstFactorChroma[k] + cursor_bw[k] * UrgentBurstFactorCursor[k] + NumberOfDPP[k] * meta_row_bandwidth[k] + NumberOfDPP[k] * dpte_row_bandwidth[k];
+	}
+
+	CalculateVActiveBandwithSupport_val = (VActiveBandwith <= ReturnBW) && !NotEnoughUrgentLatencyHiding;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: NotEnoughUrgentLatencyHiding        = %u\n", __func__, NotEnoughUrgentLatencyHiding);
+	dml_print("DML::%s: VActiveBandwith                     = %f\n", __func__, VActiveBandwith);
+	dml_print("DML::%s: ReturnBW                            = %f\n", __func__, ReturnBW);
+	dml_print("DML::%s: CalculateVActiveBandwithSupport_val = %u\n", __func__, CalculateVActiveBandwithSupport_val);
+#endif
+	return CalculateVActiveBandwithSupport_val;
+}
+
+static void CalculatePrefetchBandwithSupport(
+										dml_uint_t NumberOfActiveSurfaces,
+										dml_float_t ReturnBW,
+										enum dml_use_mall_for_pstate_change_mode UseMALLForPStateChange[],
+										dml_bool_t NotUrgentLatencyHiding[],
+										dml_float_t ReadBandwidthLuma[],
+										dml_float_t ReadBandwidthChroma[],
+										dml_float_t PrefetchBandwidthLuma[],
+										dml_float_t PrefetchBandwidthChroma[],
+										dml_float_t cursor_bw[],
+										dml_float_t meta_row_bandwidth[],
+										dml_float_t dpte_row_bandwidth[],
+										dml_float_t cursor_bw_pre[],
+										dml_float_t prefetch_vmrow_bw[],
+										dml_uint_t NumberOfDPP[],
+										dml_float_t UrgentBurstFactorLuma[],
+										dml_float_t UrgentBurstFactorChroma[],
+										dml_float_t UrgentBurstFactorCursor[],
+										dml_float_t UrgentBurstFactorLumaPre[],
+										dml_float_t UrgentBurstFactorChromaPre[],
+										dml_float_t UrgentBurstFactorCursorPre[],
+
+										// Output
+										dml_float_t *PrefetchBandwidth,
+										dml_float_t *PrefetchBandwidthNotIncludingMALLPrefetch,
+										dml_float_t *FractionOfUrgentBandwidth,
+										dml_bool_t *PrefetchBandwidthSupport)
+{
+	dml_bool_t NotEnoughUrgentLatencyHiding = false;
+	for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+		if (NotUrgentLatencyHiding[k]) {
+			NotEnoughUrgentLatencyHiding = true;
+		}
+	}
+
+	*PrefetchBandwidth = 0;
+	for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+		*PrefetchBandwidth = *PrefetchBandwidth + dml_max3(NumberOfDPP[k] * prefetch_vmrow_bw[k],
+														ReadBandwidthLuma[k] * UrgentBurstFactorLuma[k] + ReadBandwidthChroma[k] * UrgentBurstFactorChroma[k] + cursor_bw[k] * UrgentBurstFactorCursor[k] + NumberOfDPP[k] * (meta_row_bandwidth[k] + dpte_row_bandwidth[k]),
+														NumberOfDPP[k] * (PrefetchBandwidthLuma[k] * UrgentBurstFactorLumaPre[k] + PrefetchBandwidthChroma[k] * UrgentBurstFactorChromaPre[k]) + cursor_bw_pre[k] * UrgentBurstFactorCursorPre[k]);
+	}
+
+	*PrefetchBandwidthNotIncludingMALLPrefetch = 0;
+	for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+		if (UseMALLForPStateChange[k] != dml_use_mall_pstate_change_phantom_pipe)
+			*PrefetchBandwidthNotIncludingMALLPrefetch = *PrefetchBandwidthNotIncludingMALLPrefetch
+				+ dml_max3(NumberOfDPP[k] * prefetch_vmrow_bw[k],
+				ReadBandwidthLuma[k] * UrgentBurstFactorLuma[k] + ReadBandwidthChroma[k] * UrgentBurstFactorChroma[k]
+				+ cursor_bw[k] * UrgentBurstFactorCursor[k]
+				+ NumberOfDPP[k] * (meta_row_bandwidth[k] + dpte_row_bandwidth[k]),
+				NumberOfDPP[k] * (PrefetchBandwidthLuma[k] * UrgentBurstFactorLumaPre[k]
+				+ PrefetchBandwidthChroma[k] * UrgentBurstFactorChromaPre[k])
+				+ cursor_bw_pre[k] * UrgentBurstFactorCursorPre[k]);
+	}
+
+	*PrefetchBandwidthSupport = (*PrefetchBandwidth <= ReturnBW) && !NotEnoughUrgentLatencyHiding;
+	*FractionOfUrgentBandwidth = *PrefetchBandwidth / ReturnBW;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: ReturnBW = %f\n", __func__, ReturnBW);
+	dml_print("DML::%s: PrefetchBandwidth = %f\n", __func__, *PrefetchBandwidth);
+	dml_print("DML::%s: FractionOfUrgentBandwidth = %f\n", __func__, *FractionOfUrgentBandwidth);
+   dml_print("DML::%s: PrefetchBandwidthSupport = %u\n", __func__, *PrefetchBandwidthSupport);
+#endif
+}
+
+static noinline_for_stack dml_float_t CalculateBandwidthAvailableForImmediateFlip(
+													dml_uint_t NumberOfActiveSurfaces,
+													dml_float_t ReturnBW,
+													dml_float_t ReadBandwidthLuma[],
+													dml_float_t ReadBandwidthChroma[],
+													dml_float_t PrefetchBandwidthLuma[],
+													dml_float_t PrefetchBandwidthChroma[],
+													dml_float_t cursor_bw[],
+													dml_float_t cursor_bw_pre[],
+													dml_uint_t NumberOfDPP[],
+													dml_float_t UrgentBurstFactorLuma[],
+													dml_float_t UrgentBurstFactorChroma[],
+													dml_float_t UrgentBurstFactorCursor[],
+													dml_float_t UrgentBurstFactorLumaPre[],
+													dml_float_t UrgentBurstFactorChromaPre[],
+													dml_float_t UrgentBurstFactorCursorPre[])
+{
+	dml_float_t ret_val = ReturnBW;
+
+	for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+		ret_val = ret_val - dml_max(ReadBandwidthLuma[k] * UrgentBurstFactorLuma[k] + ReadBandwidthChroma[k] * UrgentBurstFactorChroma[k] + cursor_bw[k] * UrgentBurstFactorCursor[k],
+									NumberOfDPP[k] * (PrefetchBandwidthLuma[k] * UrgentBurstFactorLumaPre[k] + PrefetchBandwidthChroma[k] * UrgentBurstFactorChromaPre[k]) +
+									cursor_bw_pre[k] * UrgentBurstFactorCursorPre[k]);
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%u\n", __func__, k);
+		dml_print("DML::%s: NumberOfDPP = %u\n", __func__, NumberOfDPP[k]);
+		dml_print("DML::%s: ReadBandwidthLuma = %f\n", __func__, ReadBandwidthLuma[k]);
+		dml_print("DML::%s: UrgentBurstFactorLuma = %f\n", __func__, UrgentBurstFactorLuma[k]);
+		dml_print("DML::%s: ReadBandwidthChroma = %f\n", __func__, ReadBandwidthChroma[k]);
+		dml_print("DML::%s: UrgentBurstFactorChroma = %f\n", __func__, UrgentBurstFactorChroma[k]);
+		dml_print("DML::%s: cursor_bw = %f\n", __func__, cursor_bw[k]);
+		dml_print("DML::%s: UrgentBurstFactorCursor = %f\n", __func__, UrgentBurstFactorCursor[k]);
+
+		dml_print("DML::%s: PrefetchBandwidthLuma = %f\n", __func__, PrefetchBandwidthLuma[k]);
+		dml_print("DML::%s: UrgentBurstFactorLumaPre = %f\n", __func__, UrgentBurstFactorLumaPre[k]);
+		dml_print("DML::%s: PrefetchBandwidthChroma = %f\n", __func__, PrefetchBandwidthChroma[k]);
+		dml_print("DML::%s: UrgentBurstFactorChromaPre = %f\n", __func__, UrgentBurstFactorChromaPre[k]);
+		dml_print("DML::%s: cursor_bw_pre = %f\n", __func__, cursor_bw_pre[k]);
+		dml_print("DML::%s: UrgentBurstFactorCursorPre = %f\n", __func__, UrgentBurstFactorCursorPre[k]);
+		dml_print("DML::%s: ret_val              = %f\n", __func__, ret_val);
+#endif
+	}
+
+	return ret_val;
+}
+
+static void CalculateImmediateFlipBandwithSupport(
+											dml_uint_t NumberOfActiveSurfaces,
+											dml_float_t ReturnBW,
+											enum dml_use_mall_for_pstate_change_mode UseMALLForPStateChange[],
+											enum dml_immediate_flip_requirement ImmediateFlipRequirement[],
+											dml_float_t final_flip_bw[],
+											dml_float_t ReadBandwidthLuma[],
+											dml_float_t ReadBandwidthChroma[],
+											dml_float_t PrefetchBandwidthLuma[],
+											dml_float_t PrefetchBandwidthChroma[],
+											dml_float_t cursor_bw[],
+											dml_float_t meta_row_bandwidth[],
+											dml_float_t dpte_row_bandwidth[],
+											dml_float_t cursor_bw_pre[],
+											dml_float_t prefetch_vmrow_bw[],
+											dml_uint_t NumberOfDPP[],
+											dml_float_t UrgentBurstFactorLuma[],
+											dml_float_t UrgentBurstFactorChroma[],
+											dml_float_t UrgentBurstFactorCursor[],
+											dml_float_t UrgentBurstFactorLumaPre[],
+											dml_float_t UrgentBurstFactorChromaPre[],
+											dml_float_t UrgentBurstFactorCursorPre[],
+
+											// Output
+											dml_float_t *TotalBandwidth,
+											dml_float_t *TotalBandwidthNotIncludingMALLPrefetch,
+											dml_float_t *FractionOfUrgentBandwidth,
+											dml_bool_t *ImmediateFlipBandwidthSupport)
+{
+	*TotalBandwidth = 0;
+	for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+		if (ImmediateFlipRequirement[k] != dml_immediate_flip_not_required) {
+
+
+
+			*TotalBandwidth = *TotalBandwidth + dml_max3(NumberOfDPP[k] * prefetch_vmrow_bw[k],
+														NumberOfDPP[k] * final_flip_bw[k] + ReadBandwidthLuma[k] * UrgentBurstFactorLuma[k] + ReadBandwidthChroma[k] * UrgentBurstFactorChroma[k] + cursor_bw[k] * UrgentBurstFactorCursor[k],
+														NumberOfDPP[k] * (final_flip_bw[k] + PrefetchBandwidthLuma[k] * UrgentBurstFactorLumaPre[k] + PrefetchBandwidthChroma[k] * UrgentBurstFactorChromaPre[k]) + cursor_bw_pre[k] * UrgentBurstFactorCursorPre[k]);
+		} else {
+			*TotalBandwidth = *TotalBandwidth + dml_max3(NumberOfDPP[k] * prefetch_vmrow_bw[k],
+														NumberOfDPP[k] * (meta_row_bandwidth[k] + dpte_row_bandwidth[k]) + ReadBandwidthLuma[k] * UrgentBurstFactorLuma[k] + ReadBandwidthChroma[k] * UrgentBurstFactorChroma[k] + cursor_bw[k] * UrgentBurstFactorCursor[k],
+														NumberOfDPP[k] * (PrefetchBandwidthLuma[k] * UrgentBurstFactorLumaPre[k] + PrefetchBandwidthChroma[k] * UrgentBurstFactorChromaPre[k]) + cursor_bw_pre[k] * UrgentBurstFactorCursorPre[k]);
+		}
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k = %u\n", __func__, k);
+		dml_print("DML::%s: ImmediateFlipRequirement = %u\n", __func__, ImmediateFlipRequirement[k]);
+		dml_print("DML::%s: TotalBandwidth = %f\n", __func__, *TotalBandwidth);
+		dml_print("DML::%s: NumberOfDPP = %u\n", __func__, NumberOfDPP[k]);
+		dml_print("DML::%s: prefetch_vmrow_bw = %f\n", __func__, prefetch_vmrow_bw[k]);
+		dml_print("DML::%s: final_flip_bw = %f\n", __func__, final_flip_bw[k]);
+		dml_print("DML::%s: ReadBandwidthLuma = %f\n", __func__, ReadBandwidthLuma[k]);
+		dml_print("DML::%s: UrgentBurstFactorLuma = %f\n", __func__, UrgentBurstFactorLuma[k]);
+		dml_print("DML::%s: ReadBandwidthChroma = %f\n", __func__, ReadBandwidthChroma[k]);
+		dml_print("DML::%s: UrgentBurstFactorChroma = %f\n", __func__, UrgentBurstFactorChroma[k]);
+		dml_print("DML::%s: cursor_bw = %f\n", __func__, cursor_bw[k]);
+		dml_print("DML::%s: UrgentBurstFactorCursor = %f\n", __func__, UrgentBurstFactorCursor[k]);
+		dml_print("DML::%s: PrefetchBandwidthLuma = %f\n", __func__, PrefetchBandwidthLuma[k]);
+		dml_print("DML::%s: UrgentBurstFactorLumaPre = %f\n", __func__, UrgentBurstFactorLumaPre[k]);
+		dml_print("DML::%s: PrefetchBandwidthChroma = %f\n", __func__, PrefetchBandwidthChroma[k]);
+		dml_print("DML::%s: UrgentBurstFactorChromaPre = %f\n", __func__, UrgentBurstFactorChromaPre[k]);
+		dml_print("DML::%s: cursor_bw_pre = %f\n", __func__, cursor_bw_pre[k]);
+		dml_print("DML::%s: UrgentBurstFactorCursorPre = %f\n", __func__, UrgentBurstFactorCursorPre[k]);
+		dml_print("DML::%s: meta_row_bandwidth = %f\n", __func__, meta_row_bandwidth[k]);
+		dml_print("DML::%s: dpte_row_bandwidth          = %f\n", __func__, dpte_row_bandwidth[k]);
+#endif
+	}
+
+	*TotalBandwidthNotIncludingMALLPrefetch = 0;
+	for (dml_uint_t k = 0; k < NumberOfActiveSurfaces; ++k) {
+		if (UseMALLForPStateChange[k] != dml_use_mall_pstate_change_phantom_pipe) {
+			if (ImmediateFlipRequirement[k] != dml_immediate_flip_not_required)
+				*TotalBandwidthNotIncludingMALLPrefetch = *TotalBandwidthNotIncludingMALLPrefetch + dml_max3(NumberOfDPP[k] * prefetch_vmrow_bw[k],
+					NumberOfDPP[k] * final_flip_bw[k] + ReadBandwidthLuma[k] * UrgentBurstFactorLuma[k] + ReadBandwidthChroma[k] * UrgentBurstFactorChroma[k] + cursor_bw[k] * UrgentBurstFactorCursor[k],
+					NumberOfDPP[k] * (final_flip_bw[k] + PrefetchBandwidthLuma[k] * UrgentBurstFactorLumaPre[k] + PrefetchBandwidthChroma[k] * UrgentBurstFactorChromaPre[k])
+					+ cursor_bw_pre[k] * UrgentBurstFactorCursorPre[k]);
+			else
+				*TotalBandwidthNotIncludingMALLPrefetch = *TotalBandwidthNotIncludingMALLPrefetch + dml_max3(NumberOfDPP[k] * prefetch_vmrow_bw[k],
+					NumberOfDPP[k] * (meta_row_bandwidth[k] + dpte_row_bandwidth[k])
+					+ ReadBandwidthLuma[k] * UrgentBurstFactorLuma[k] + ReadBandwidthChroma[k] * UrgentBurstFactorChroma[k] + cursor_bw[k] * UrgentBurstFactorCursor[k],
+					NumberOfDPP[k] * (PrefetchBandwidthLuma[k] * UrgentBurstFactorLumaPre[k] + PrefetchBandwidthChroma[k] * UrgentBurstFactorChromaPre[k])
+					+ cursor_bw_pre[k] * UrgentBurstFactorCursorPre[k]);
+		}
+	}
+
+	*ImmediateFlipBandwidthSupport = (*TotalBandwidth <= ReturnBW);
+	*FractionOfUrgentBandwidth = *TotalBandwidth / ReturnBW;
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: ReturnBW = %f\n", __func__, ReturnBW);
+	dml_print("DML::%s: TotalBandwidth = %f\n", __func__, *TotalBandwidth);
+	dml_print("DML::%s: ImmediateFlipBandwidthSupport = %u\n", __func__, *ImmediateFlipBandwidthSupport);
+#endif
+}
+
+static dml_uint_t MicroSecToVertLines(dml_uint_t num_us, dml_uint_t h_total, dml_float_t pixel_clock)
+{
+	dml_uint_t lines_time_in_ns = 1000.0 * (h_total * 1000.0) / (pixel_clock * 1000.0);
+
+	return dml_ceil(1000.0 * num_us / lines_time_in_ns, 1.0);
+}
+
+/// @brief Calculate the maximum vstartup for mode support and mode programming consideration
+///         Bounded by min of actual vblank and input vblank_nom, dont want vstartup/ready to start too early if actual vbllank is huge
+static dml_uint_t CalculateMaxVStartup(
+			dml_uint_t          plane_idx,
+			dml_bool_t          ptoi_supported,
+			dml_uint_t          vblank_nom_default_us,
+			struct dml_timing_cfg_st  *timing,
+			dml_float_t         write_back_delay_us)
+{
+	dml_uint_t vblank_size = 0;
+	dml_uint_t max_vstartup_lines = 0;
+	const dml_uint_t max_allowed_vblank_nom = 1023;
+
+	dml_float_t line_time_us = (dml_float_t) timing->HTotal[plane_idx] / timing->PixelClock[plane_idx];
+	dml_uint_t vblank_actual = timing->VTotal[plane_idx] - timing->VActive[plane_idx];
+
+	dml_uint_t vblank_nom_default_in_line = MicroSecToVertLines(vblank_nom_default_us, timing->HTotal[plane_idx],
+			timing->PixelClock[plane_idx]);
+	dml_uint_t vblank_nom_input = (dml_uint_t)dml_min(vblank_actual, vblank_nom_default_in_line);
+
+	// vblank_nom should not be smaller than (VSync (VTotal - VActive - VFrontPorch) + 2)
+	// + 2 is because
+	// 1 -> VStartup_start should be 1 line before VSync
+	// 1 -> always reserve 1 line between start of VBlank to VStartup signal
+	dml_uint_t vblank_nom_vsync_capped = dml_max(vblank_nom_input,
+			timing->VTotal[plane_idx] - timing->VActive[plane_idx] - timing->VFrontPorch[plane_idx] + 2);
+	dml_uint_t vblank_nom_max_allowed_capped = dml_min(vblank_nom_vsync_capped, max_allowed_vblank_nom);
+	dml_uint_t vblank_avail = (vblank_nom_max_allowed_capped == 0) ?
+			vblank_nom_default_in_line : vblank_nom_max_allowed_capped;
+
+	vblank_size = (dml_uint_t) dml_min(vblank_actual, vblank_avail);
+
+   if (timing->Interlace[plane_idx] && !ptoi_supported)
+		max_vstartup_lines = (dml_uint_t) (dml_floor(vblank_size/2.0, 1.0));
+	else
+		max_vstartup_lines = vblank_size - (dml_uint_t) dml_max(1.0, dml_ceil(write_back_delay_us/line_time_us, 1.0));
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: plane_idx = %u\n", __func__, plane_idx);
+	dml_print("DML::%s: VBlankNom = %u\n", __func__, timing->VBlankNom[plane_idx]);
+	dml_print("DML::%s: vblank_nom_default_us = %u\n", __func__, vblank_nom_default_us);
+	dml_print("DML::%s: line_time_us = %f\n", __func__, line_time_us);
+	dml_print("DML::%s: vblank_actual = %u\n", __func__, vblank_actual);
+	dml_print("DML::%s: vblank_avail = %u\n", __func__, vblank_avail);
+	dml_print("DML::%s: max_vstartup_lines = %u\n", __func__, max_vstartup_lines);
+#endif
+	max_vstartup_lines = (dml_uint_t) dml_min(max_vstartup_lines, DML_MAX_VSTARTUP_START);
+	return max_vstartup_lines;
+}
+
+static noinline_for_stack void set_calculate_prefetch_schedule_params(struct display_mode_lib_st *mode_lib,
+						   struct CalculatePrefetchSchedule_params_st *CalculatePrefetchSchedule_params,
+						   dml_uint_t j,
+						   dml_uint_t k)
+{
+				CalculatePrefetchSchedule_params->DSCDelay = mode_lib->ms.DSCDelayPerState[k];
+				CalculatePrefetchSchedule_params->EnhancedPrefetchScheduleAccelerationFinal = mode_lib->ms.policy.EnhancedPrefetchScheduleAccelerationFinal;
+				CalculatePrefetchSchedule_params->DPPCLKDelaySubtotalPlusCNVCFormater = mode_lib->ms.ip.dppclk_delay_subtotal + mode_lib->ms.ip.dppclk_delay_cnvc_formatter;
+				CalculatePrefetchSchedule_params->DPPCLKDelaySCL = mode_lib->ms.ip.dppclk_delay_scl;
+				CalculatePrefetchSchedule_params->DPPCLKDelaySCLLBOnly = mode_lib->ms.ip.dppclk_delay_scl_lb_only;
+				CalculatePrefetchSchedule_params->DPPCLKDelayCNVCCursor = mode_lib->ms.ip.dppclk_delay_cnvc_cursor;
+				CalculatePrefetchSchedule_params->DISPCLKDelaySubtotal = mode_lib->ms.ip.dispclk_delay_subtotal;
+				CalculatePrefetchSchedule_params->DPP_RECOUT_WIDTH = (dml_uint_t)(mode_lib->ms.SwathWidthYThisState[k] / mode_lib->ms.cache_display_cfg.plane.HRatio[k]);
+				CalculatePrefetchSchedule_params->OutputFormat = mode_lib->ms.cache_display_cfg.output.OutputFormat[k];
+				CalculatePrefetchSchedule_params->MaxInterDCNTileRepeaters = mode_lib->ms.ip.max_inter_dcn_tile_repeaters;
+				CalculatePrefetchSchedule_params->GPUVMPageTableLevels = mode_lib->ms.cache_display_cfg.plane.GPUVMMaxPageTableLevels;
+				CalculatePrefetchSchedule_params->GPUVMEnable = mode_lib->ms.cache_display_cfg.plane.GPUVMEnable;
+				CalculatePrefetchSchedule_params->HostVMEnable = mode_lib->ms.cache_display_cfg.plane.HostVMEnable;
+				CalculatePrefetchSchedule_params->HostVMMaxNonCachedPageTableLevels = mode_lib->ms.cache_display_cfg.plane.HostVMMaxPageTableLevels;
+				CalculatePrefetchSchedule_params->HostVMMinPageSize = mode_lib->ms.soc.hostvm_min_page_size_kbytes * 1024;
+				CalculatePrefetchSchedule_params->DynamicMetadataEnable = mode_lib->ms.cache_display_cfg.plane.DynamicMetadataEnable[k];
+				CalculatePrefetchSchedule_params->DynamicMetadataVMEnabled = mode_lib->ms.ip.dynamic_metadata_vm_enabled;
+				CalculatePrefetchSchedule_params->DynamicMetadataLinesBeforeActiveRequired = mode_lib->ms.cache_display_cfg.plane.DynamicMetadataLinesBeforeActiveRequired[k];
+				CalculatePrefetchSchedule_params->DynamicMetadataTransmittedBytes = mode_lib->ms.cache_display_cfg.plane.DynamicMetadataTransmittedBytes[k];
+				CalculatePrefetchSchedule_params->UrgentLatency = mode_lib->ms.UrgLatency;
+				CalculatePrefetchSchedule_params->UrgentExtraLatency = mode_lib->ms.ExtraLatency;
+				CalculatePrefetchSchedule_params->TCalc = mode_lib->ms.TimeCalc;
+				CalculatePrefetchSchedule_params->PDEAndMetaPTEBytesFrame = mode_lib->ms.PDEAndMetaPTEBytesPerFrame[j][k];
+				CalculatePrefetchSchedule_params->MetaRowByte = mode_lib->ms.MetaRowBytes[j][k];
+				CalculatePrefetchSchedule_params->PixelPTEBytesPerRow = mode_lib->ms.DPTEBytesPerRow[j][k];
+				CalculatePrefetchSchedule_params->PrefetchSourceLinesY = mode_lib->ms.PrefetchLinesY[j][k];
+				CalculatePrefetchSchedule_params->VInitPreFillY = mode_lib->ms.PrefillY[k];
+				CalculatePrefetchSchedule_params->MaxNumSwathY = mode_lib->ms.MaxNumSwY[k];
+				CalculatePrefetchSchedule_params->PrefetchSourceLinesC = mode_lib->ms.PrefetchLinesC[j][k];
+				CalculatePrefetchSchedule_params->VInitPreFillC = mode_lib->ms.PrefillC[k];
+				CalculatePrefetchSchedule_params->MaxNumSwathC = mode_lib->ms.MaxNumSwC[k];
+				CalculatePrefetchSchedule_params->swath_width_luma_ub = mode_lib->ms.swath_width_luma_ub_this_state[k];
+				CalculatePrefetchSchedule_params->swath_width_chroma_ub = mode_lib->ms.swath_width_chroma_ub_this_state[k];
+				CalculatePrefetchSchedule_params->SwathHeightY = mode_lib->ms.SwathHeightYThisState[k];
+				CalculatePrefetchSchedule_params->SwathHeightC = mode_lib->ms.SwathHeightCThisState[k];
+				CalculatePrefetchSchedule_params->TWait = mode_lib->ms.TWait;
+				CalculatePrefetchSchedule_params->DestinationLinesForPrefetch = &mode_lib->ms.LineTimesForPrefetch[k];
+				CalculatePrefetchSchedule_params->DestinationLinesToRequestVMInVBlank = &mode_lib->ms.LinesForMetaPTE[k];
+				CalculatePrefetchSchedule_params->DestinationLinesToRequestRowInVBlank = &mode_lib->ms.LinesForMetaAndDPTERow[k];
+				CalculatePrefetchSchedule_params->VRatioPrefetchY = &mode_lib->ms.VRatioPreY[j][k];
+				CalculatePrefetchSchedule_params->VRatioPrefetchC = &mode_lib->ms.VRatioPreC[j][k];
+				CalculatePrefetchSchedule_params->RequiredPrefetchPixDataBWLuma = &mode_lib->ms.RequiredPrefetchPixelDataBWLuma[k];
+				CalculatePrefetchSchedule_params->RequiredPrefetchPixDataBWChroma = &mode_lib->ms.RequiredPrefetchPixelDataBWChroma[k];
+				CalculatePrefetchSchedule_params->NotEnoughTimeForDynamicMetadata = &mode_lib->ms.support.NoTimeForDynamicMetadata[j][k];
+				CalculatePrefetchSchedule_params->Tno_bw = &mode_lib->ms.Tno_bw[k];
+}
+
+static noinline_for_stack void dml_prefetch_check(struct display_mode_lib_st *mode_lib)
+{
+	struct dml_core_mode_support_locals_st *s = &mode_lib->scratch.dml_core_mode_support_locals;
+	struct CalculatePrefetchSchedule_params_st *CalculatePrefetchSchedule_params = &mode_lib->scratch.CalculatePrefetchSchedule_params;
+	struct CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_params_st *CalculateWatermarks_params = &mode_lib->scratch.CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_params;
+	struct DmlPipe *myPipe;
+	dml_uint_t j, k;
+
+	for (j = 0; j < 2; ++j) {
+		mode_lib->ms.TimeCalc = 24 / mode_lib->ms.ProjectedDCFCLKDeepSleep[j];
+
+		for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+			mode_lib->ms.NoOfDPPThisState[k] = mode_lib->ms.NoOfDPP[j][k];
+			mode_lib->ms.swath_width_luma_ub_this_state[k] = mode_lib->ms.swath_width_luma_ub_all_states[j][k];
+			mode_lib->ms.swath_width_chroma_ub_this_state[k] = mode_lib->ms.swath_width_chroma_ub_all_states[j][k];
+			mode_lib->ms.SwathWidthYThisState[k] = mode_lib->ms.SwathWidthYAllStates[j][k];
+			mode_lib->ms.SwathWidthCThisState[k] = mode_lib->ms.SwathWidthCAllStates[j][k];
+			mode_lib->ms.SwathHeightYThisState[k] = mode_lib->ms.SwathHeightYAllStates[j][k];
+			mode_lib->ms.SwathHeightCThisState[k] = mode_lib->ms.SwathHeightCAllStates[j][k];
+			mode_lib->ms.UnboundedRequestEnabledThisState = mode_lib->ms.UnboundedRequestEnabledAllStates[j];
+			mode_lib->ms.CompressedBufferSizeInkByteThisState = mode_lib->ms.CompressedBufferSizeInkByteAllStates[j];
+			mode_lib->ms.DETBufferSizeInKByteThisState[k] = mode_lib->ms.DETBufferSizeInKByteAllStates[j][k];
+			mode_lib->ms.DETBufferSizeYThisState[k] = mode_lib->ms.DETBufferSizeYAllStates[j][k];
+			mode_lib->ms.DETBufferSizeCThisState[k] = mode_lib->ms.DETBufferSizeCAllStates[j][k];
+		}
+
+		mode_lib->ms.support.VActiveBandwithSupport[j] = CalculateVActiveBandwithSupport(
+			mode_lib->ms.num_active_planes,
+			mode_lib->ms.ReturnBWPerState[j],
+			mode_lib->ms.NotUrgentLatencyHiding,
+			mode_lib->ms.ReadBandwidthLuma,
+			mode_lib->ms.ReadBandwidthChroma,
+			mode_lib->ms.cursor_bw,
+			mode_lib->ms.meta_row_bandwidth_this_state,
+			mode_lib->ms.dpte_row_bandwidth_this_state,
+			mode_lib->ms.NoOfDPPThisState,
+			mode_lib->ms.UrgentBurstFactorLuma[j],
+			mode_lib->ms.UrgentBurstFactorChroma[j],
+			mode_lib->ms.UrgentBurstFactorCursor[j]);
+
+		s->VMDataOnlyReturnBWPerState = dml_get_return_bw_mbps_vm_only(
+																	&mode_lib->ms.soc,
+																	mode_lib->ms.state.use_ideal_dram_bw_strobe,
+																	mode_lib->ms.cache_display_cfg.plane.HostVMEnable,
+																	mode_lib->ms.DCFCLKState[j],
+																	mode_lib->ms.state.fabricclk_mhz,
+																	mode_lib->ms.state.dram_speed_mts);
+
+		s->HostVMInefficiencyFactor = 1;
+		if (mode_lib->ms.cache_display_cfg.plane.GPUVMEnable && mode_lib->ms.cache_display_cfg.plane.HostVMEnable)
+			s->HostVMInefficiencyFactor = mode_lib->ms.ReturnBWPerState[j] / s->VMDataOnlyReturnBWPerState;
+
+		mode_lib->ms.ExtraLatency = CalculateExtraLatency(
+				mode_lib->ms.soc.round_trip_ping_latency_dcfclk_cycles,
+				s->ReorderingBytes,
+				mode_lib->ms.DCFCLKState[j],
+				mode_lib->ms.TotalNumberOfActiveDPP[j],
+				mode_lib->ms.ip.pixel_chunk_size_kbytes,
+				mode_lib->ms.TotalNumberOfDCCActiveDPP[j],
+				mode_lib->ms.ip.meta_chunk_size_kbytes,
+				mode_lib->ms.ReturnBWPerState[j],
+				mode_lib->ms.cache_display_cfg.plane.GPUVMEnable,
+				mode_lib->ms.cache_display_cfg.plane.HostVMEnable,
+				mode_lib->ms.num_active_planes,
+				mode_lib->ms.NoOfDPPThisState,
+				mode_lib->ms.dpte_group_bytes,
+				s->HostVMInefficiencyFactor,
+				mode_lib->ms.soc.hostvm_min_page_size_kbytes * 1024,
+				mode_lib->ms.cache_display_cfg.plane.HostVMMaxPageTableLevels);
+
+		s->NextMaxVStartup = s->MaxVStartupAllPlanes[j];
+		s->MaxVStartup = 0;
+		s->AllPrefetchModeTested = true;
+		for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+			CalculatePrefetchMode(mode_lib->ms.policy.AllowForPStateChangeOrStutterInVBlank[k], &s->MinPrefetchMode[k], &s->MaxPrefetchMode[k]);
+			s->NextPrefetchMode[k] = s->MinPrefetchMode[k];
+		}
+
+		do {
+			s->MaxVStartup = s->NextMaxVStartup;
+			s->AllPrefetchModeTested = true;
+
+			for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+				mode_lib->ms.PrefetchMode[k] = s->NextPrefetchMode[k];
+				mode_lib->ms.TWait = CalculateTWait(
+								mode_lib->ms.PrefetchMode[k],
+								mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange[k],
+								mode_lib->ms.policy.SynchronizeDRRDisplaysForUCLKPStateChangeFinal,
+								mode_lib->ms.cache_display_cfg.timing.DRRDisplay[k],
+								mode_lib->ms.state.dram_clock_change_latency_us,
+								mode_lib->ms.state.fclk_change_latency_us,
+								mode_lib->ms.UrgLatency,
+								mode_lib->ms.state.sr_enter_plus_exit_time_us);
+
+				myPipe = &s->myPipe;
+				myPipe->Dppclk = mode_lib->ms.RequiredDPPCLKPerSurface[j][k];
+				myPipe->Dispclk = mode_lib->ms.RequiredDISPCLK[j];
+				myPipe->PixelClock = mode_lib->ms.cache_display_cfg.timing.PixelClock[k];
+				myPipe->DCFClkDeepSleep = mode_lib->ms.ProjectedDCFCLKDeepSleep[j];
+				myPipe->DPPPerSurface = mode_lib->ms.NoOfDPP[j][k];
+				myPipe->ScalerEnabled = mode_lib->ms.cache_display_cfg.plane.ScalerEnabled[k];
+				myPipe->SourceScan = mode_lib->ms.cache_display_cfg.plane.SourceScan[k];
+				myPipe->BlockWidth256BytesY = mode_lib->ms.Read256BlockWidthY[k];
+				myPipe->BlockHeight256BytesY = mode_lib->ms.Read256BlockHeightY[k];
+				myPipe->BlockWidth256BytesC = mode_lib->ms.Read256BlockWidthC[k];
+				myPipe->BlockHeight256BytesC = mode_lib->ms.Read256BlockHeightC[k];
+				myPipe->InterlaceEnable = mode_lib->ms.cache_display_cfg.timing.Interlace[k];
+				myPipe->NumberOfCursors = mode_lib->ms.cache_display_cfg.plane.NumberOfCursors[k];
+				myPipe->VBlank = mode_lib->ms.cache_display_cfg.timing.VTotal[k] - mode_lib->ms.cache_display_cfg.timing.VActive[k];
+				myPipe->HTotal = mode_lib->ms.cache_display_cfg.timing.HTotal[k];
+				myPipe->HActive = mode_lib->ms.cache_display_cfg.timing.HActive[k];
+				myPipe->DCCEnable = mode_lib->ms.cache_display_cfg.surface.DCCEnable[k];
+				myPipe->ODMMode = mode_lib->ms.ODMModePerState[k];
+				myPipe->SourcePixelFormat = mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k];
+				myPipe->BytePerPixelY = mode_lib->ms.BytePerPixelY[k];
+				myPipe->BytePerPixelC = mode_lib->ms.BytePerPixelC[k];
+				myPipe->ProgressiveToInterlaceUnitInOPP = mode_lib->ms.ip.ptoi_supported;
+
+#ifdef __DML_VBA_DEBUG__
+				dml_print("DML::%s: Calling CalculatePrefetchSchedule for j=%u, k=%u\n", __func__, j, k);
+				dml_print("DML::%s: MaximumVStartup = %u\n", __func__, s->MaximumVStartup[j][k]);
+				dml_print("DML::%s: MaxVStartup = %u\n", __func__, s->MaxVStartup);
+				dml_print("DML::%s: NextPrefetchMode = %u\n", __func__, s->NextPrefetchMode[k]);
+				dml_print("DML::%s: AllowForPStateChangeOrStutterInVBlank = %u\n", __func__, mode_lib->ms.policy.AllowForPStateChangeOrStutterInVBlank[k]);
+				dml_print("DML::%s: PrefetchMode = %u\n", __func__, mode_lib->ms.PrefetchMode[k]);
+#endif
+
+				CalculatePrefetchSchedule_params->HostVMInefficiencyFactor = s->HostVMInefficiencyFactor;
+				CalculatePrefetchSchedule_params->myPipe = myPipe;
+				CalculatePrefetchSchedule_params->VStartup = (dml_uint_t)(dml_min(s->MaxVStartup, s->MaximumVStartup[j][k]));
+				CalculatePrefetchSchedule_params->MaxVStartup = s->MaximumVStartup[j][k];
+				CalculatePrefetchSchedule_params->DSTXAfterScaler = &s->DSTXAfterScaler[k];
+				CalculatePrefetchSchedule_params->DSTYAfterScaler = &s->DSTYAfterScaler[k];
+				CalculatePrefetchSchedule_params->prefetch_vmrow_bw = &mode_lib->ms.prefetch_vmrow_bw[k];
+				CalculatePrefetchSchedule_params->Tdmdl_vm = &s->dummy_single[0];
+				CalculatePrefetchSchedule_params->Tdmdl = &s->dummy_single[1];
+				CalculatePrefetchSchedule_params->TSetup = &s->dummy_single[2];
+				CalculatePrefetchSchedule_params->VUpdateOffsetPix = &s->dummy_integer[0];
+				CalculatePrefetchSchedule_params->VUpdateWidthPix = &s->dummy_integer[1];
+				CalculatePrefetchSchedule_params->VReadyOffsetPix = &s->dummy_integer[2];
+
+				set_calculate_prefetch_schedule_params(mode_lib, CalculatePrefetchSchedule_params, j, k);
+
+				mode_lib->ms.support.NoTimeForPrefetch[j][k] =
+								CalculatePrefetchSchedule(&mode_lib->scratch,
+								CalculatePrefetchSchedule_params);
+			}
+
+			for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+					CalculateUrgentBurstFactor(
+							mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange[k],
+							mode_lib->ms.swath_width_luma_ub_this_state[k],
+							mode_lib->ms.swath_width_chroma_ub_this_state[k],
+							mode_lib->ms.SwathHeightYThisState[k],
+							mode_lib->ms.SwathHeightCThisState[k],
+							mode_lib->ms.cache_display_cfg.timing.HTotal[k] / mode_lib->ms.cache_display_cfg.timing.PixelClock[k],
+							mode_lib->ms.UrgLatency,
+							mode_lib->ms.ip.cursor_buffer_size,
+							mode_lib->ms.cache_display_cfg.plane.CursorWidth[k],
+							mode_lib->ms.cache_display_cfg.plane.CursorBPP[k],
+							mode_lib->ms.VRatioPreY[j][k],
+							mode_lib->ms.VRatioPreC[j][k],
+							mode_lib->ms.BytePerPixelInDETY[k],
+							mode_lib->ms.BytePerPixelInDETC[k],
+							mode_lib->ms.DETBufferSizeYThisState[k],
+							mode_lib->ms.DETBufferSizeCThisState[k],
+							/* Output */
+							&mode_lib->ms.UrgentBurstFactorCursorPre[k],
+							&mode_lib->ms.UrgentBurstFactorLumaPre[k],
+							&mode_lib->ms.UrgentBurstFactorChromaPre[k],
+							&mode_lib->ms.NotUrgentLatencyHidingPre[k]);
+
+					mode_lib->ms.cursor_bw_pre[k] = mode_lib->ms.cache_display_cfg.plane.NumberOfCursors[k] * mode_lib->ms.cache_display_cfg.plane.CursorWidth[k] *
+													mode_lib->ms.cache_display_cfg.plane.CursorBPP[k] / 8.0 / (mode_lib->ms.cache_display_cfg.timing.HTotal[k] /
+													mode_lib->ms.cache_display_cfg.timing.PixelClock[k]) * mode_lib->ms.VRatioPreY[j][k];
+			}
+
+			{
+			CalculatePrefetchBandwithSupport(
+				mode_lib->ms.num_active_planes,
+				mode_lib->ms.ReturnBWPerState[j],
+				mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange,
+				mode_lib->ms.NotUrgentLatencyHidingPre,
+				mode_lib->ms.ReadBandwidthLuma,
+				mode_lib->ms.ReadBandwidthChroma,
+				mode_lib->ms.RequiredPrefetchPixelDataBWLuma,
+				mode_lib->ms.RequiredPrefetchPixelDataBWChroma,
+				mode_lib->ms.cursor_bw,
+				mode_lib->ms.meta_row_bandwidth_this_state,
+				mode_lib->ms.dpte_row_bandwidth_this_state,
+				mode_lib->ms.cursor_bw_pre,
+				mode_lib->ms.prefetch_vmrow_bw,
+				mode_lib->ms.NoOfDPPThisState,
+				mode_lib->ms.UrgentBurstFactorLuma[j],
+				mode_lib->ms.UrgentBurstFactorChroma[j],
+				mode_lib->ms.UrgentBurstFactorCursor[j],
+				mode_lib->ms.UrgentBurstFactorLumaPre,
+				mode_lib->ms.UrgentBurstFactorChromaPre,
+				mode_lib->ms.UrgentBurstFactorCursorPre,
+
+				/* output */
+				&s->dummy_single[0], // dml_float_t *PrefetchBandwidth
+				&s->dummy_single[1], // dml_float_t *PrefetchBandwidthNotIncludingMALLPrefetch
+				&mode_lib->mp.FractionOfUrgentBandwidth, // dml_float_t *FractionOfUrgentBandwidth
+				&mode_lib->ms.support.PrefetchSupported[j]);
+			}
+
+			for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+				if (mode_lib->ms.LineTimesForPrefetch[k] < 2.0
+					|| mode_lib->ms.LinesForMetaPTE[k] >= 32.0
+					|| mode_lib->ms.LinesForMetaAndDPTERow[k] >= 16.0
+					|| mode_lib->ms.support.NoTimeForPrefetch[j][k] == true) {
+						mode_lib->ms.support.PrefetchSupported[j] = false;
+				}
+			}
+
+			mode_lib->ms.support.DynamicMetadataSupported[j] = true;
+			for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+				if (mode_lib->ms.support.NoTimeForDynamicMetadata[j][k] == true) {
+					mode_lib->ms.support.DynamicMetadataSupported[j] = false;
+				}
+			}
+
+			mode_lib->ms.support.VRatioInPrefetchSupported[j] = true;
+			for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+				if (mode_lib->ms.support.NoTimeForPrefetch[j][k] == true ||
+					mode_lib->ms.VRatioPreY[j][k] > __DML_MAX_VRATIO_PRE_ENHANCE_PREFETCH_ACC__ ||
+					mode_lib->ms.VRatioPreC[j][k] > __DML_MAX_VRATIO_PRE_ENHANCE_PREFETCH_ACC__ ||
+					((s->MaxVStartup < s->MaximumVStartup[j][k] || mode_lib->ms.policy.EnhancedPrefetchScheduleAccelerationFinal == 0) &&
+						(mode_lib->ms.VRatioPreY[j][k] > __DML_MAX_VRATIO_PRE__ || mode_lib->ms.VRatioPreC[j][k] > __DML_MAX_VRATIO_PRE__))) {
+							mode_lib->ms.support.VRatioInPrefetchSupported[j] = false;
+				}
+			}
+
+			s->AnyLinesForVMOrRowTooLarge = false;
+			for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+				if (mode_lib->ms.LinesForMetaAndDPTERow[k] >= 16 || mode_lib->ms.LinesForMetaPTE[k] >= 32) {
+					s->AnyLinesForVMOrRowTooLarge = true;
+				}
+			}
+
+			if (mode_lib->ms.support.PrefetchSupported[j] == true && mode_lib->ms.support.VRatioInPrefetchSupported[j] == true) {
+				mode_lib->ms.BandwidthAvailableForImmediateFlip = CalculateBandwidthAvailableForImmediateFlip(
+						mode_lib->ms.num_active_planes,
+						mode_lib->ms.ReturnBWPerState[j],
+						mode_lib->ms.ReadBandwidthLuma,
+						mode_lib->ms.ReadBandwidthChroma,
+						mode_lib->ms.RequiredPrefetchPixelDataBWLuma,
+						mode_lib->ms.RequiredPrefetchPixelDataBWChroma,
+						mode_lib->ms.cursor_bw,
+						mode_lib->ms.cursor_bw_pre,
+						mode_lib->ms.NoOfDPPThisState,
+						mode_lib->ms.UrgentBurstFactorLuma[j],
+						mode_lib->ms.UrgentBurstFactorChroma[j],
+						mode_lib->ms.UrgentBurstFactorCursor[j],
+						mode_lib->ms.UrgentBurstFactorLumaPre,
+						mode_lib->ms.UrgentBurstFactorChromaPre,
+						mode_lib->ms.UrgentBurstFactorCursorPre);
+
+				mode_lib->ms.TotImmediateFlipBytes = 0;
+				for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+					if (!(mode_lib->ms.policy.ImmediateFlipRequirement[k] == dml_immediate_flip_not_required)) {
+						mode_lib->ms.TotImmediateFlipBytes = mode_lib->ms.TotImmediateFlipBytes + mode_lib->ms.NoOfDPP[j][k] * (mode_lib->ms.PDEAndMetaPTEBytesPerFrame[j][k] + mode_lib->ms.MetaRowBytes[j][k]);
+						if (mode_lib->ms.use_one_row_for_frame_flip[j][k]) {
+							mode_lib->ms.TotImmediateFlipBytes = mode_lib->ms.TotImmediateFlipBytes + mode_lib->ms.NoOfDPP[j][k] * (2 * mode_lib->ms.DPTEBytesPerRow[j][k]);
+						} else {
+							mode_lib->ms.TotImmediateFlipBytes = mode_lib->ms.TotImmediateFlipBytes + mode_lib->ms.NoOfDPP[j][k] * mode_lib->ms.DPTEBytesPerRow[j][k];
+						}
+					}
+				}
+
+				for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+					CalculateFlipSchedule(
+						s->HostVMInefficiencyFactor,
+						mode_lib->ms.ExtraLatency,
+						mode_lib->ms.UrgLatency,
+						mode_lib->ms.cache_display_cfg.plane.GPUVMMaxPageTableLevels,
+						mode_lib->ms.cache_display_cfg.plane.HostVMEnable,
+						mode_lib->ms.cache_display_cfg.plane.HostVMMaxPageTableLevels,
+						mode_lib->ms.cache_display_cfg.plane.GPUVMEnable,
+						mode_lib->ms.soc.hostvm_min_page_size_kbytes * 1024,
+						mode_lib->ms.PDEAndMetaPTEBytesPerFrame[j][k],
+						mode_lib->ms.MetaRowBytes[j][k],
+						mode_lib->ms.DPTEBytesPerRow[j][k],
+						mode_lib->ms.BandwidthAvailableForImmediateFlip,
+						mode_lib->ms.TotImmediateFlipBytes,
+						mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k],
+						(mode_lib->ms.cache_display_cfg.timing.HTotal[k] / mode_lib->ms.cache_display_cfg.timing.PixelClock[k]),
+						mode_lib->ms.cache_display_cfg.plane.VRatio[k],
+						mode_lib->ms.cache_display_cfg.plane.VRatioChroma[k],
+						mode_lib->ms.Tno_bw[k],
+						mode_lib->ms.cache_display_cfg.surface.DCCEnable[k],
+						mode_lib->ms.dpte_row_height[k],
+						mode_lib->ms.meta_row_height[k],
+						mode_lib->ms.dpte_row_height_chroma[k],
+						mode_lib->ms.meta_row_height_chroma[k],
+						mode_lib->ms.use_one_row_for_frame_flip[j][k], // 24
+
+						/* Output */
+						&mode_lib->ms.DestinationLinesToRequestVMInImmediateFlip[k],
+						&mode_lib->ms.DestinationLinesToRequestRowInImmediateFlip[k],
+						&mode_lib->ms.final_flip_bw[k],
+						&mode_lib->ms.ImmediateFlipSupportedForPipe[k]);
+				}
+
+				{
+				CalculateImmediateFlipBandwithSupport(mode_lib->ms.num_active_planes,
+													mode_lib->ms.ReturnBWPerState[j],
+													mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange,
+													mode_lib->ms.policy.ImmediateFlipRequirement,
+													mode_lib->ms.final_flip_bw,
+													mode_lib->ms.ReadBandwidthLuma,
+													mode_lib->ms.ReadBandwidthChroma,
+													mode_lib->ms.RequiredPrefetchPixelDataBWLuma,
+													mode_lib->ms.RequiredPrefetchPixelDataBWChroma,
+													mode_lib->ms.cursor_bw,
+													mode_lib->ms.meta_row_bandwidth_this_state,
+													mode_lib->ms.dpte_row_bandwidth_this_state,
+													mode_lib->ms.cursor_bw_pre,
+													mode_lib->ms.prefetch_vmrow_bw,
+													mode_lib->ms.NoOfDPP[j], // VBA_ERROR DPPPerSurface is not assigned at this point, should use NoOfDpp here
+													mode_lib->ms.UrgentBurstFactorLuma[j],
+													mode_lib->ms.UrgentBurstFactorChroma[j],
+													mode_lib->ms.UrgentBurstFactorCursor[j],
+													mode_lib->ms.UrgentBurstFactorLumaPre,
+													mode_lib->ms.UrgentBurstFactorChromaPre,
+													mode_lib->ms.UrgentBurstFactorCursorPre,
+
+													/* output */
+													&s->dummy_single[0], // dml_float_t *TotalBandwidth
+													&s->dummy_single[1], // dml_float_t *TotalBandwidthNotIncludingMALLPrefetch
+													&s->dummy_single[2], // dml_float_t *FractionOfUrgentBandwidth
+													&mode_lib->ms.support.ImmediateFlipSupportedForState[j]); // dml_bool_t *ImmediateFlipBandwidthSupport
+				}
+
+				for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+					if (!(mode_lib->ms.policy.ImmediateFlipRequirement[k] == dml_immediate_flip_not_required) && (mode_lib->ms.ImmediateFlipSupportedForPipe[k] == false))
+						mode_lib->ms.support.ImmediateFlipSupportedForState[j] = false;
+				}
+
+			} else { // if prefetch not support, assume iflip not supported
+				mode_lib->ms.support.ImmediateFlipSupportedForState[j] = false;
+			}
+
+			if (s->MaxVStartup <= __DML_VBA_MIN_VSTARTUP__ || s->AnyLinesForVMOrRowTooLarge == false) {
+				s->NextMaxVStartup = s->MaxVStartupAllPlanes[j];
+				for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+					s->NextPrefetchMode[k] = s->NextPrefetchMode[k] + 1;
+
+					if (s->NextPrefetchMode[k] <= s->MaxPrefetchMode[k])
+						s->AllPrefetchModeTested = false;
+				}
+			} else {
+				s->NextMaxVStartup = s->NextMaxVStartup - 1;
+			}
+		} while (!((mode_lib->ms.support.PrefetchSupported[j] == true && mode_lib->ms.support.DynamicMetadataSupported[j] == true &&
+					mode_lib->ms.support.VRatioInPrefetchSupported[j] == true &&
+					// consider flip support is okay if when there is no hostvm and the user does't require a iflip OR the flip bw is ok
+					// If there is hostvm, DCN needs to support iflip for invalidation
+					((s->ImmediateFlipRequiredFinal) || mode_lib->ms.support.ImmediateFlipSupportedForState[j] == true)) ||
+					(s->NextMaxVStartup == s->MaxVStartupAllPlanes[j] && s->AllPrefetchModeTested)));
+
+		for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+			mode_lib->ms.use_one_row_for_frame_this_state[k] = mode_lib->ms.use_one_row_for_frame[j][k];
+		}
+
+		s->mSOCParameters.UrgentLatency = mode_lib->ms.UrgLatency;
+		s->mSOCParameters.ExtraLatency = mode_lib->ms.ExtraLatency;
+		s->mSOCParameters.WritebackLatency = mode_lib->ms.state.writeback_latency_us;
+		s->mSOCParameters.DRAMClockChangeLatency = mode_lib->ms.state.dram_clock_change_latency_us;
+		s->mSOCParameters.FCLKChangeLatency = mode_lib->ms.state.fclk_change_latency_us;
+		s->mSOCParameters.SRExitTime = mode_lib->ms.state.sr_exit_time_us;
+		s->mSOCParameters.SREnterPlusExitTime = mode_lib->ms.state.sr_enter_plus_exit_time_us;
+		s->mSOCParameters.SRExitZ8Time = mode_lib->ms.state.sr_exit_z8_time_us;
+		s->mSOCParameters.SREnterPlusExitZ8Time = mode_lib->ms.state.sr_enter_plus_exit_z8_time_us;
+		s->mSOCParameters.USRRetrainingLatency = mode_lib->ms.state.usr_retraining_latency_us;
+		s->mSOCParameters.SMNLatency = mode_lib->ms.soc.smn_latency_us;
+
+		CalculateWatermarks_params->USRRetrainingRequiredFinal = mode_lib->ms.policy.USRRetrainingRequiredFinal;
+		CalculateWatermarks_params->UseMALLForPStateChange = mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange;
+		CalculateWatermarks_params->PrefetchMode = mode_lib->ms.PrefetchMode;
+		CalculateWatermarks_params->NumberOfActiveSurfaces = mode_lib->ms.num_active_planes;
+		CalculateWatermarks_params->MaxLineBufferLines = mode_lib->ms.ip.max_line_buffer_lines;
+		CalculateWatermarks_params->LineBufferSize = mode_lib->ms.ip.line_buffer_size_bits;
+		CalculateWatermarks_params->WritebackInterfaceBufferSize = mode_lib->ms.ip.writeback_interface_buffer_size_kbytes;
+		CalculateWatermarks_params->DCFCLK = mode_lib->ms.DCFCLKState[j];
+		CalculateWatermarks_params->ReturnBW = mode_lib->ms.ReturnBWPerState[j];
+		CalculateWatermarks_params->SynchronizeTimingsFinal = mode_lib->ms.policy.SynchronizeTimingsFinal;
+		CalculateWatermarks_params->SynchronizeDRRDisplaysForUCLKPStateChangeFinal = mode_lib->ms.policy.SynchronizeDRRDisplaysForUCLKPStateChangeFinal;
+		CalculateWatermarks_params->DRRDisplay = mode_lib->ms.cache_display_cfg.timing.DRRDisplay;
+		CalculateWatermarks_params->dpte_group_bytes = mode_lib->ms.dpte_group_bytes;
+		CalculateWatermarks_params->meta_row_height = mode_lib->ms.meta_row_height;
+		CalculateWatermarks_params->meta_row_height_chroma = mode_lib->ms.meta_row_height_chroma;
+		CalculateWatermarks_params->mmSOCParameters = s->mSOCParameters;
+		CalculateWatermarks_params->WritebackChunkSize = mode_lib->ms.ip.writeback_chunk_size_kbytes;
+		CalculateWatermarks_params->SOCCLK = mode_lib->ms.state.socclk_mhz;
+		CalculateWatermarks_params->DCFClkDeepSleep = mode_lib->ms.ProjectedDCFCLKDeepSleep[j];
+		CalculateWatermarks_params->DETBufferSizeY = mode_lib->ms.DETBufferSizeYThisState;
+		CalculateWatermarks_params->DETBufferSizeC = mode_lib->ms.DETBufferSizeCThisState;
+		CalculateWatermarks_params->SwathHeightY = mode_lib->ms.SwathHeightYThisState;
+		CalculateWatermarks_params->SwathHeightC = mode_lib->ms.SwathHeightCThisState;
+		CalculateWatermarks_params->LBBitPerPixel = mode_lib->ms.cache_display_cfg.plane.LBBitPerPixel;
+		CalculateWatermarks_params->SwathWidthY = mode_lib->ms.SwathWidthYThisState;
+		CalculateWatermarks_params->SwathWidthC = mode_lib->ms.SwathWidthCThisState;
+		CalculateWatermarks_params->HRatio = mode_lib->ms.cache_display_cfg.plane.HRatio;
+		CalculateWatermarks_params->HRatioChroma = mode_lib->ms.cache_display_cfg.plane.HRatioChroma;
+		CalculateWatermarks_params->VTaps = mode_lib->ms.cache_display_cfg.plane.VTaps;
+		CalculateWatermarks_params->VTapsChroma = mode_lib->ms.cache_display_cfg.plane.VTapsChroma;
+		CalculateWatermarks_params->VRatio = mode_lib->ms.cache_display_cfg.plane.VRatio;
+		CalculateWatermarks_params->VRatioChroma = mode_lib->ms.cache_display_cfg.plane.VRatioChroma;
+		CalculateWatermarks_params->HTotal = mode_lib->ms.cache_display_cfg.timing.HTotal;
+		CalculateWatermarks_params->VTotal = mode_lib->ms.cache_display_cfg.timing.VTotal;
+		CalculateWatermarks_params->VActive = mode_lib->ms.cache_display_cfg.timing.VActive;
+		CalculateWatermarks_params->PixelClock = mode_lib->ms.cache_display_cfg.timing.PixelClock;
+		CalculateWatermarks_params->BlendingAndTiming = mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming;
+		CalculateWatermarks_params->DPPPerSurface = mode_lib->ms.NoOfDPPThisState;
+		CalculateWatermarks_params->BytePerPixelDETY = mode_lib->ms.BytePerPixelInDETY;
+		CalculateWatermarks_params->BytePerPixelDETC = mode_lib->ms.BytePerPixelInDETC;
+		CalculateWatermarks_params->DSTXAfterScaler = s->DSTXAfterScaler;
+		CalculateWatermarks_params->DSTYAfterScaler = s->DSTYAfterScaler;
+		CalculateWatermarks_params->WritebackEnable = mode_lib->ms.cache_display_cfg.writeback.WritebackEnable;
+		CalculateWatermarks_params->WritebackPixelFormat = mode_lib->ms.cache_display_cfg.writeback.WritebackPixelFormat;
+		CalculateWatermarks_params->WritebackDestinationWidth = mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationWidth;
+		CalculateWatermarks_params->WritebackDestinationHeight = mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationHeight;
+		CalculateWatermarks_params->WritebackSourceHeight = mode_lib->ms.cache_display_cfg.writeback.WritebackSourceHeight;
+		CalculateWatermarks_params->UnboundedRequestEnabled = mode_lib->ms.UnboundedRequestEnabledThisState;
+		CalculateWatermarks_params->CompressedBufferSizeInkByte = mode_lib->ms.CompressedBufferSizeInkByteThisState;
+
+		// Output
+		CalculateWatermarks_params->Watermark = &s->dummy_watermark; // Watermarks *Watermark
+		CalculateWatermarks_params->DRAMClockChangeSupport = &mode_lib->ms.support.DRAMClockChangeSupport[j];
+		CalculateWatermarks_params->MaxActiveDRAMClockChangeLatencySupported = &s->dummy_single_array[0]; // dml_float_t *MaxActiveDRAMClockChangeLatencySupported[]
+		CalculateWatermarks_params->SubViewportLinesNeededInMALL = &mode_lib->ms.SubViewportLinesNeededInMALL[j]; // dml_uint_t SubViewportLinesNeededInMALL[]
+		CalculateWatermarks_params->FCLKChangeSupport = &mode_lib->ms.support.FCLKChangeSupport[j];
+		CalculateWatermarks_params->MaxActiveFCLKChangeLatencySupported = &s->dummy_single[0]; // dml_float_t *MaxActiveFCLKChangeLatencySupported
+		CalculateWatermarks_params->USRRetrainingSupport = &mode_lib->ms.support.USRRetrainingSupport[j];
+		CalculateWatermarks_params->ActiveDRAMClockChangeLatencyMargin = mode_lib->ms.support.ActiveDRAMClockChangeLatencyMargin;
+
+		CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport(&mode_lib->scratch,
+			CalculateWatermarks_params);
+
+	} // for j
+}
+
+/// @brief The Mode Support function.
+dml_bool_t dml_core_mode_support(struct display_mode_lib_st *mode_lib)
+{
+	struct dml_core_mode_support_locals_st *s = &mode_lib->scratch.dml_core_mode_support_locals;
+	struct UseMinimumDCFCLK_params_st *UseMinimumDCFCLK_params = &mode_lib->scratch.UseMinimumDCFCLK_params;
+	struct CalculateSwathAndDETConfiguration_params_st *CalculateSwathAndDETConfiguration_params = &mode_lib->scratch.CalculateSwathAndDETConfiguration_params;
+	struct CalculateVMRowAndSwath_params_st *CalculateVMRowAndSwath_params = &mode_lib->scratch.CalculateVMRowAndSwath_params;
+
+	dml_uint_t j, k, m;
+
+	mode_lib->ms.num_active_planes = dml_get_num_active_planes(&mode_lib->ms.cache_display_cfg);
+	dml_print("DML::%s: num_active_planes = %u\n", __func__, mode_lib->ms.num_active_planes);
+
+	CalculateMaxDETAndMinCompressedBufferSize(
+								mode_lib->ms.ip.config_return_buffer_size_in_kbytes,
+								mode_lib->ms.ip.config_return_buffer_segment_size_in_kbytes,
+								mode_lib->ms.ip.rob_buffer_size_kbytes,
+								mode_lib->ms.ip.max_num_dpp,
+								mode_lib->ms.policy.NomDETInKByteOverrideEnable,   // VBA_DELTA
+								mode_lib->ms.policy.NomDETInKByteOverrideValue,    // VBA_DELTA
+
+								/* Output */
+								&mode_lib->ms.MaxTotalDETInKByte,
+								&mode_lib->ms.NomDETInKByte,
+								&mode_lib->ms.MinCompressedBufferSizeInKByte);
+
+	PixelClockAdjustmentForProgressiveToInterlaceUnit(&mode_lib->ms.cache_display_cfg, mode_lib->ms.ip.ptoi_supported);
+
+
+	/*MODE SUPPORT, VOLTAGE STATE AND SOC CONFIGURATION*/
+
+	/*Scale Ratio, taps Support Check*/
+	mode_lib->ms.support.ScaleRatioAndTapsSupport = true;
+	for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+		if (mode_lib->ms.cache_display_cfg.plane.ScalerEnabled[k] == false
+				&& ((mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_444_64
+						&& mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_444_32
+						&& mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_444_16
+						&& mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_mono_16
+						&& mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_mono_8
+						&& mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_rgbe
+						&& mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_rgbe_alpha)
+						|| mode_lib->ms.cache_display_cfg.plane.HRatio[k] != 1.0
+						|| mode_lib->ms.cache_display_cfg.plane.HTaps[k] != 1.0
+						|| mode_lib->ms.cache_display_cfg.plane.VRatio[k] != 1.0
+						|| mode_lib->ms.cache_display_cfg.plane.VTaps[k] != 1.0)) {
+			mode_lib->ms.support.ScaleRatioAndTapsSupport = false;
+		} else if (mode_lib->ms.cache_display_cfg.plane.VTaps[k] < 1.0 || mode_lib->ms.cache_display_cfg.plane.VTaps[k] > 8.0
+				|| mode_lib->ms.cache_display_cfg.plane.HTaps[k] < 1.0 || mode_lib->ms.cache_display_cfg.plane.HTaps[k] > 8.0
+				|| (mode_lib->ms.cache_display_cfg.plane.HTaps[k] > 1.0 && (mode_lib->ms.cache_display_cfg.plane.HTaps[k] % 2) == 1)
+				|| mode_lib->ms.cache_display_cfg.plane.HRatio[k] > mode_lib->ms.ip.max_hscl_ratio
+				|| mode_lib->ms.cache_display_cfg.plane.VRatio[k] > mode_lib->ms.ip.max_vscl_ratio
+				|| mode_lib->ms.cache_display_cfg.plane.HRatio[k] > mode_lib->ms.cache_display_cfg.plane.HTaps[k]
+				|| mode_lib->ms.cache_display_cfg.plane.VRatio[k] > mode_lib->ms.cache_display_cfg.plane.VTaps[k]
+				|| (mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_444_64
+					&& mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_444_32
+					&& mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_444_16
+					&& mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_mono_16
+					&& mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_mono_8
+					&& mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_rgbe
+					&& (mode_lib->ms.cache_display_cfg.plane.VTapsChroma[k] < 1 || mode_lib->ms.cache_display_cfg.plane.VTapsChroma[k] > 8 || mode_lib->ms.cache_display_cfg.plane.HTapsChroma[k] < 1 || mode_lib->ms.cache_display_cfg.plane.HTapsChroma[k] > 8 ||
+						(mode_lib->ms.cache_display_cfg.plane.HTapsChroma[k] > 1 && mode_lib->ms.cache_display_cfg.plane.HTapsChroma[k] % 2 == 1) ||
+					mode_lib->ms.cache_display_cfg.plane.HRatioChroma[k] > mode_lib->ms.ip.max_hscl_ratio ||
+					mode_lib->ms.cache_display_cfg.plane.VRatioChroma[k] > mode_lib->ms.ip.max_vscl_ratio ||
+					mode_lib->ms.cache_display_cfg.plane.HRatioChroma[k] > mode_lib->ms.cache_display_cfg.plane.HTapsChroma[k] ||
+					mode_lib->ms.cache_display_cfg.plane.VRatioChroma[k] > mode_lib->ms.cache_display_cfg.plane.VTapsChroma[k]))) {
+			mode_lib->ms.support.ScaleRatioAndTapsSupport = false;
+		}
+	}
+
+	/*Source Format, Pixel Format and Scan Support Check*/
+	mode_lib->ms.support.SourceFormatPixelAndScanSupport = true;
+	for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+			if (mode_lib->ms.cache_display_cfg.surface.SurfaceTiling[k] == dml_sw_linear && (!(!dml_is_vertical_rotation(mode_lib->ms.cache_display_cfg.plane.SourceScan[k])) || mode_lib->ms.cache_display_cfg.surface.DCCEnable[k] == true)) {
+			mode_lib->ms.support.SourceFormatPixelAndScanSupport = false;
+		}
+	}
+
+	for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+		CalculateBytePerPixelAndBlockSizes(
+								mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k],
+								mode_lib->ms.cache_display_cfg.surface.SurfaceTiling[k],
+
+								/* Output */
+								&mode_lib->ms.BytePerPixelY[k],
+								&mode_lib->ms.BytePerPixelC[k],
+								&mode_lib->ms.BytePerPixelInDETY[k],
+								&mode_lib->ms.BytePerPixelInDETC[k],
+								&mode_lib->ms.Read256BlockHeightY[k],
+								&mode_lib->ms.Read256BlockHeightC[k],
+								&mode_lib->ms.Read256BlockWidthY[k],
+								&mode_lib->ms.Read256BlockWidthC[k],
+								&mode_lib->ms.MacroTileHeightY[k],
+								&mode_lib->ms.MacroTileHeightC[k],
+								&mode_lib->ms.MacroTileWidthY[k],
+								&mode_lib->ms.MacroTileWidthC[k]);
+	}
+
+	/*Bandwidth Support Check*/
+	for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+		if (!dml_is_vertical_rotation(mode_lib->ms.cache_display_cfg.plane.SourceScan[k])) {
+			mode_lib->ms.SwathWidthYSingleDPP[k] = mode_lib->ms.cache_display_cfg.plane.ViewportWidth[k];
+			mode_lib->ms.SwathWidthCSingleDPP[k] = mode_lib->ms.cache_display_cfg.plane.ViewportWidthChroma[k];
+		} else {
+			mode_lib->ms.SwathWidthYSingleDPP[k] = mode_lib->ms.cache_display_cfg.plane.ViewportHeight[k];
+			mode_lib->ms.SwathWidthCSingleDPP[k] = mode_lib->ms.cache_display_cfg.plane.ViewportHeightChroma[k];
+		}
+	}
+	for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+		mode_lib->ms.ReadBandwidthLuma[k] = mode_lib->ms.SwathWidthYSingleDPP[k] * dml_ceil(mode_lib->ms.BytePerPixelInDETY[k], 1.0) / (mode_lib->ms.cache_display_cfg.timing.HTotal[k] / mode_lib->ms.cache_display_cfg.timing.PixelClock[k]) * mode_lib->ms.cache_display_cfg.plane.VRatio[k];
+		mode_lib->ms.ReadBandwidthChroma[k] = mode_lib->ms.SwathWidthYSingleDPP[k] / 2 * dml_ceil(mode_lib->ms.BytePerPixelInDETC[k], 2.0) / (mode_lib->ms.cache_display_cfg.timing.HTotal[k] / mode_lib->ms.cache_display_cfg.timing.PixelClock[k]) * mode_lib->ms.cache_display_cfg.plane.VRatio[k] / 2.0;
+	}
+	for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+		if (mode_lib->ms.cache_display_cfg.writeback.WritebackEnable[k] == true
+				&& mode_lib->ms.cache_display_cfg.writeback.WritebackPixelFormat[k] == dml_444_64) {
+			mode_lib->ms.WriteBandwidth[k] = mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationWidth[k]
+					* mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationHeight[k]
+					/ (mode_lib->ms.cache_display_cfg.writeback.WritebackSourceHeight[k]
+							* mode_lib->ms.cache_display_cfg.timing.HTotal[k]
+							/ mode_lib->ms.cache_display_cfg.timing.PixelClock[k]) * 8.0;
+		} else if (mode_lib->ms.cache_display_cfg.writeback.WritebackEnable[k] == true) {
+			mode_lib->ms.WriteBandwidth[k] = mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationWidth[k]
+					* mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationHeight[k]
+					/ (mode_lib->ms.cache_display_cfg.writeback.WritebackSourceHeight[k]
+							* mode_lib->ms.cache_display_cfg.timing.HTotal[k]
+							/ mode_lib->ms.cache_display_cfg.timing.PixelClock[k]) * 4.0;
+		} else {
+			mode_lib->ms.WriteBandwidth[k] = 0.0;
+		}
+	}
+
+	/*Writeback Latency support check*/
+	mode_lib->ms.support.WritebackLatencySupport = true;
+	for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+		if (mode_lib->ms.cache_display_cfg.writeback.WritebackEnable[k] == true &&
+			(mode_lib->ms.WriteBandwidth[k] > mode_lib->ms.ip.writeback_interface_buffer_size_kbytes * 1024 / mode_lib->ms.state.writeback_latency_us)) {
+			mode_lib->ms.support.WritebackLatencySupport = false;
+		}
+	}
+
+	/*Writeback Mode Support Check*/
+	s->TotalNumberOfActiveWriteback = 0;
+	for (k = 0; k <= (dml_uint_t) mode_lib->ms.num_active_planes - 1; k++) {
+		if (mode_lib->ms.cache_display_cfg.writeback.WritebackEnable[k] == true) {
+			s->TotalNumberOfActiveWriteback = s->TotalNumberOfActiveWriteback + 1;
+		}
+   }
+
+	mode_lib->ms.support.EnoughWritebackUnits = 1;
+	if (s->TotalNumberOfActiveWriteback > (dml_uint_t) mode_lib->ms.ip.max_num_wb) {
+		mode_lib->ms.support.EnoughWritebackUnits = false;
+	}
+
+	/*Writeback Scale Ratio and Taps Support Check*/
+	mode_lib->ms.support.WritebackScaleRatioAndTapsSupport = true;
+	for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+		if (mode_lib->ms.cache_display_cfg.writeback.WritebackEnable[k] == true) {
+			if (mode_lib->ms.cache_display_cfg.writeback.WritebackHRatio[k] > mode_lib->ms.ip.writeback_max_hscl_ratio
+				|| mode_lib->ms.cache_display_cfg.writeback.WritebackVRatio[k] > mode_lib->ms.ip.writeback_max_vscl_ratio
+				|| mode_lib->ms.cache_display_cfg.writeback.WritebackHRatio[k] < mode_lib->ms.ip.writeback_min_hscl_ratio
+				|| mode_lib->ms.cache_display_cfg.writeback.WritebackVRatio[k] < mode_lib->ms.ip.writeback_min_vscl_ratio
+				|| mode_lib->ms.cache_display_cfg.writeback.WritebackHTaps[k] > (dml_uint_t) mode_lib->ms.ip.writeback_max_hscl_taps
+				|| mode_lib->ms.cache_display_cfg.writeback.WritebackVTaps[k] > (dml_uint_t) mode_lib->ms.ip.writeback_max_vscl_taps
+				|| mode_lib->ms.cache_display_cfg.writeback.WritebackHRatio[k] > (dml_uint_t) mode_lib->ms.cache_display_cfg.writeback.WritebackHTaps[k]
+				|| mode_lib->ms.cache_display_cfg.writeback.WritebackVRatio[k] > (dml_uint_t) mode_lib->ms.cache_display_cfg.writeback.WritebackVTaps[k]
+				|| (mode_lib->ms.cache_display_cfg.writeback.WritebackHTaps[k] > 2.0 && ((mode_lib->ms.cache_display_cfg.writeback.WritebackHTaps[k] % 2) == 1))) {
+				mode_lib->ms.support.WritebackScaleRatioAndTapsSupport = false;
+			}
+			if (2.0 * mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationWidth[k] * (mode_lib->ms.cache_display_cfg.writeback.WritebackVTaps[k] - 1) * 57 > mode_lib->ms.ip.writeback_line_buffer_buffer_size) {
+				mode_lib->ms.support.WritebackScaleRatioAndTapsSupport = false;
+			}
+		}
+	}
+
+	for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+		CalculateSinglePipeDPPCLKAndSCLThroughput(
+				mode_lib->ms.cache_display_cfg.plane.HRatio[k],
+				mode_lib->ms.cache_display_cfg.plane.HRatioChroma[k],
+				mode_lib->ms.cache_display_cfg.plane.VRatio[k],
+				mode_lib->ms.cache_display_cfg.plane.VRatioChroma[k],
+				mode_lib->ms.ip.max_dchub_pscl_bw_pix_per_clk,
+				mode_lib->ms.ip.max_pscl_lb_bw_pix_per_clk,
+				mode_lib->ms.cache_display_cfg.timing.PixelClock[k],
+				mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k],
+				mode_lib->ms.cache_display_cfg.plane.HTaps[k],
+				mode_lib->ms.cache_display_cfg.plane.HTapsChroma[k],
+				mode_lib->ms.cache_display_cfg.plane.VTaps[k],
+				mode_lib->ms.cache_display_cfg.plane.VTapsChroma[k],
+				/* Output */
+				&mode_lib->ms.PSCL_FACTOR[k],
+				&mode_lib->ms.PSCL_FACTOR_CHROMA[k],
+				&mode_lib->ms.MinDPPCLKUsingSingleDPP[k]);
+	}
+
+	for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+		if (mode_lib->ms.cache_display_cfg.surface.SurfaceTiling[k] == dml_sw_linear) {
+			s->MaximumSwathWidthSupportLuma = 8192;
+		} else if (!dml_is_vertical_rotation(mode_lib->ms.cache_display_cfg.plane.SourceScan[k]) && mode_lib->ms.BytePerPixelC[k] > 0 && mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_rgbe_alpha) {
+			s->MaximumSwathWidthSupportLuma = 7680;
+		} else if (dml_is_vertical_rotation(mode_lib->ms.cache_display_cfg.plane.SourceScan[k]) && mode_lib->ms.BytePerPixelC[k] > 0 && mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_rgbe_alpha) {
+			s->MaximumSwathWidthSupportLuma = 4320;
+		} else if (mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] == dml_rgbe_alpha) {
+			s->MaximumSwathWidthSupportLuma = 3840;
+		} else if (dml_is_vertical_rotation(mode_lib->ms.cache_display_cfg.plane.SourceScan[k]) && mode_lib->ms.BytePerPixelY[k] == 8 && mode_lib->ms.cache_display_cfg.surface.DCCEnable[k] == true) {
+			s->MaximumSwathWidthSupportLuma = 3072;
+		} else {
+			s->MaximumSwathWidthSupportLuma = 6144;
+		}
+
+       if (mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] == dml_420_8 || mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] == dml_420_10 || mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] == dml_420_12) {
+			s->MaximumSwathWidthSupportChroma = (dml_uint_t)(s->MaximumSwathWidthSupportLuma / 2.0);
+		} else {
+			s->MaximumSwathWidthSupportChroma = s->MaximumSwathWidthSupportLuma;
+       }
+		mode_lib->ms.MaximumSwathWidthInLineBufferLuma = mode_lib->ms.ip.line_buffer_size_bits * dml_max(mode_lib->ms.cache_display_cfg.plane.HRatio[k], 1.0) / mode_lib->ms.cache_display_cfg.plane.LBBitPerPixel[k] /
+															(mode_lib->ms.cache_display_cfg.plane.VTaps[k] + dml_max(dml_ceil(mode_lib->ms.cache_display_cfg.plane.VRatio[k], 1.0) - 2, 0.0));
+		if (mode_lib->ms.BytePerPixelC[k] == 0.0) {
+			mode_lib->ms.MaximumSwathWidthInLineBufferChroma = 0;
+		} else {
+			mode_lib->ms.MaximumSwathWidthInLineBufferChroma =
+							mode_lib->ms.ip.line_buffer_size_bits
+									* dml_max(mode_lib->ms.cache_display_cfg.plane.HRatioChroma[k], 1.0)
+									/ mode_lib->ms.cache_display_cfg.plane.LBBitPerPixel[k]
+									/ (mode_lib->ms.cache_display_cfg.plane.VTapsChroma[k]
+									+ dml_max(dml_ceil(mode_lib->ms.cache_display_cfg.plane.VRatioChroma[k], 1.0) - 2, 0.0));
+		}
+		mode_lib->ms.MaximumSwathWidthLuma[k] = dml_min(s->MaximumSwathWidthSupportLuma, mode_lib->ms.MaximumSwathWidthInLineBufferLuma);
+		mode_lib->ms.MaximumSwathWidthChroma[k] = dml_min(s->MaximumSwathWidthSupportChroma, mode_lib->ms.MaximumSwathWidthInLineBufferChroma);
+	}
+
+	/*Number Of DSC Slices*/
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		if (mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming[k] == k &&
+			mode_lib->ms.cache_display_cfg.output.DSCEnable[k] != dml_dsc_disable) {
+			mode_lib->ms.support.NumberOfDSCSlices[k] = mode_lib->ms.cache_display_cfg.output.DSCSlices[k];
+
+			if (mode_lib->ms.support.NumberOfDSCSlices[k] == 0) {
+				if (mode_lib->ms.cache_display_cfg.output.PixelClockBackEnd[k] > 4800) {
+					mode_lib->ms.support.NumberOfDSCSlices[k] = (dml_uint_t)(dml_ceil(mode_lib->ms.cache_display_cfg.output.PixelClockBackEnd[k] / 600, 4));
+				} else if (mode_lib->ms.cache_display_cfg.output.PixelClockBackEnd[k] > 2400) {
+					mode_lib->ms.support.NumberOfDSCSlices[k] = 8;
+				} else if (mode_lib->ms.cache_display_cfg.output.PixelClockBackEnd[k] > 1200) {
+					mode_lib->ms.support.NumberOfDSCSlices[k] = 4;
+				} else if (mode_lib->ms.cache_display_cfg.output.PixelClockBackEnd[k] > 340) {
+					mode_lib->ms.support.NumberOfDSCSlices[k] = 2;
+				} else {
+					mode_lib->ms.support.NumberOfDSCSlices[k] = 1;
+				}
+			}
+		} else {
+			mode_lib->ms.support.NumberOfDSCSlices[k] = 1;
+		}
+	}
+
+	CalculateSwathAndDETConfiguration_params->DETSizeOverride = mode_lib->ms.cache_display_cfg.plane.DETSizeOverride;
+	CalculateSwathAndDETConfiguration_params->UseMALLForPStateChange = mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange;
+	CalculateSwathAndDETConfiguration_params->ConfigReturnBufferSizeInKByte = mode_lib->ms.ip.config_return_buffer_size_in_kbytes;
+	CalculateSwathAndDETConfiguration_params->ROBBufferSizeInKByte = mode_lib->ms.ip.rob_buffer_size_kbytes;
+	CalculateSwathAndDETConfiguration_params->MaxTotalDETInKByte = mode_lib->ms.MaxTotalDETInKByte;
+	CalculateSwathAndDETConfiguration_params->MinCompressedBufferSizeInKByte = mode_lib->ms.MinCompressedBufferSizeInKByte;
+	CalculateSwathAndDETConfiguration_params->PixelChunkSizeInKByte = mode_lib->ms.ip.pixel_chunk_size_kbytes;
+	CalculateSwathAndDETConfiguration_params->ForceSingleDPP = 1;
+	CalculateSwathAndDETConfiguration_params->NumberOfActiveSurfaces = mode_lib->ms.num_active_planes;
+	CalculateSwathAndDETConfiguration_params->nomDETInKByte = mode_lib->ms.NomDETInKByte;
+	CalculateSwathAndDETConfiguration_params->UseUnboundedRequestingFinal = mode_lib->ms.policy.UseUnboundedRequesting;
+	CalculateSwathAndDETConfiguration_params->ConfigReturnBufferSegmentSizeInkByte = mode_lib->ms.ip.config_return_buffer_segment_size_in_kbytes;
+	CalculateSwathAndDETConfiguration_params->CompressedBufferSegmentSizeInkByteFinal = mode_lib->ms.ip.compressed_buffer_segment_size_in_kbytes;
+	CalculateSwathAndDETConfiguration_params->Output = mode_lib->ms.cache_display_cfg.output.OutputEncoder;
+	CalculateSwathAndDETConfiguration_params->ReadBandwidthLuma = mode_lib->ms.ReadBandwidthLuma;
+	CalculateSwathAndDETConfiguration_params->ReadBandwidthChroma = mode_lib->ms.ReadBandwidthChroma;
+	CalculateSwathAndDETConfiguration_params->MaximumSwathWidthLuma = mode_lib->ms.MaximumSwathWidthLuma;
+	CalculateSwathAndDETConfiguration_params->MaximumSwathWidthChroma = mode_lib->ms.MaximumSwathWidthChroma;
+	CalculateSwathAndDETConfiguration_params->SourceScan = mode_lib->ms.cache_display_cfg.plane.SourceScan;
+	CalculateSwathAndDETConfiguration_params->ViewportStationary = mode_lib->ms.cache_display_cfg.plane.ViewportStationary;
+	CalculateSwathAndDETConfiguration_params->SourcePixelFormat = mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat;
+	CalculateSwathAndDETConfiguration_params->SurfaceTiling = mode_lib->ms.cache_display_cfg.surface.SurfaceTiling;
+	CalculateSwathAndDETConfiguration_params->ViewportWidth = mode_lib->ms.cache_display_cfg.plane.ViewportWidth;
+	CalculateSwathAndDETConfiguration_params->ViewportHeight = mode_lib->ms.cache_display_cfg.plane.ViewportHeight;
+	CalculateSwathAndDETConfiguration_params->ViewportXStart = mode_lib->ms.cache_display_cfg.plane.ViewportXStart;
+	CalculateSwathAndDETConfiguration_params->ViewportYStart = mode_lib->ms.cache_display_cfg.plane.ViewportYStart;
+	CalculateSwathAndDETConfiguration_params->ViewportXStartC = mode_lib->ms.cache_display_cfg.plane.ViewportXStartC;
+	CalculateSwathAndDETConfiguration_params->ViewportYStartC = mode_lib->ms.cache_display_cfg.plane.ViewportYStartC;
+	CalculateSwathAndDETConfiguration_params->SurfaceWidthY = mode_lib->ms.cache_display_cfg.surface.SurfaceWidthY;
+	CalculateSwathAndDETConfiguration_params->SurfaceWidthC = mode_lib->ms.cache_display_cfg.surface.SurfaceWidthC;
+	CalculateSwathAndDETConfiguration_params->SurfaceHeightY = mode_lib->ms.cache_display_cfg.surface.SurfaceHeightY;
+	CalculateSwathAndDETConfiguration_params->SurfaceHeightC = mode_lib->ms.cache_display_cfg.surface.SurfaceHeightC;
+	CalculateSwathAndDETConfiguration_params->Read256BytesBlockHeightY = mode_lib->ms.Read256BlockHeightY;
+	CalculateSwathAndDETConfiguration_params->Read256BytesBlockHeightC = mode_lib->ms.Read256BlockHeightC;
+	CalculateSwathAndDETConfiguration_params->Read256BytesBlockWidthY = mode_lib->ms.Read256BlockWidthY;
+	CalculateSwathAndDETConfiguration_params->Read256BytesBlockWidthC = mode_lib->ms.Read256BlockWidthC;
+	CalculateSwathAndDETConfiguration_params->ODMMode = s->dummy_odm_mode;
+	CalculateSwathAndDETConfiguration_params->BlendingAndTiming = mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming;
+	CalculateSwathAndDETConfiguration_params->BytePerPixY = mode_lib->ms.BytePerPixelY;
+	CalculateSwathAndDETConfiguration_params->BytePerPixC = mode_lib->ms.BytePerPixelC;
+	CalculateSwathAndDETConfiguration_params->BytePerPixDETY = mode_lib->ms.BytePerPixelInDETY;
+	CalculateSwathAndDETConfiguration_params->BytePerPixDETC = mode_lib->ms.BytePerPixelInDETC;
+	CalculateSwathAndDETConfiguration_params->HActive = mode_lib->ms.cache_display_cfg.timing.HActive;
+	CalculateSwathAndDETConfiguration_params->HRatio = mode_lib->ms.cache_display_cfg.plane.HRatio;
+	CalculateSwathAndDETConfiguration_params->HRatioChroma = mode_lib->ms.cache_display_cfg.plane.HRatioChroma;
+	CalculateSwathAndDETConfiguration_params->DPPPerSurface = s->dummy_integer_array[0];
+	CalculateSwathAndDETConfiguration_params->swath_width_luma_ub = s->dummy_integer_array[1];
+	CalculateSwathAndDETConfiguration_params->swath_width_chroma_ub = s->dummy_integer_array[2];
+	CalculateSwathAndDETConfiguration_params->SwathWidth = s->dummy_integer_array[3];
+	CalculateSwathAndDETConfiguration_params->SwathWidthChroma = s->dummy_integer_array[4];
+	CalculateSwathAndDETConfiguration_params->SwathHeightY = s->dummy_integer_array[5];
+	CalculateSwathAndDETConfiguration_params->SwathHeightC = s->dummy_integer_array[6];
+	CalculateSwathAndDETConfiguration_params->DETBufferSizeInKByte = s->dummy_integer_array[7];
+	CalculateSwathAndDETConfiguration_params->DETBufferSizeY = mode_lib->ms.DETBufferSizeY;
+	CalculateSwathAndDETConfiguration_params->DETBufferSizeC = mode_lib->ms.DETBufferSizeC;
+	CalculateSwathAndDETConfiguration_params->UnboundedRequestEnabled = &s->dummy_boolean[0];
+	CalculateSwathAndDETConfiguration_params->compbuf_reserved_space_64b = &s->dummy_integer[2];
+	CalculateSwathAndDETConfiguration_params->compbuf_reserved_space_zs = &s->dummy_integer[1];
+	CalculateSwathAndDETConfiguration_params->CompressedBufferSizeInkByte = &s->dummy_integer[0];
+	CalculateSwathAndDETConfiguration_params->ViewportSizeSupportPerSurface = mode_lib->ms.SingleDPPViewportSizeSupportPerSurface;
+	CalculateSwathAndDETConfiguration_params->ViewportSizeSupport = &s->dummy_boolean[1];
+
+	CalculateSwathAndDETConfiguration(&mode_lib->scratch,
+	CalculateSwathAndDETConfiguration_params); /* dml_bool_t *ViewportSizeSupport */
+
+	s->MPCCombineMethodAsNeededForPStateChangeAndVoltage = false;
+	s->MPCCombineMethodAsPossible = false;
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		if (mode_lib->ms.policy.MPCCombineUse[k] == dml_mpc_as_needed_for_pstate_and_voltage)
+			s->MPCCombineMethodAsNeededForPStateChangeAndVoltage = true;
+		if (mode_lib->ms.policy.MPCCombineUse[k] == dml_mpc_as_possible)
+			s->MPCCombineMethodAsPossible = true;
+	}
+	mode_lib->ms.support.MPCCombineMethodIncompatible = s->MPCCombineMethodAsNeededForPStateChangeAndVoltage && s->MPCCombineMethodAsPossible;
+
+	for (j = 0; j < 2; j++) {
+		mode_lib->ms.TotalNumberOfActiveDPP[j] = 0;
+		mode_lib->ms.support.TotalAvailablePipesSupport[j] = true;
+
+		for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+			CalculateODMMode(
+					mode_lib->ms.ip.maximum_pixels_per_line_per_dsc_unit,
+					mode_lib->ms.cache_display_cfg.timing.HActive[k],
+					mode_lib->ms.cache_display_cfg.output.OutputEncoder[k],
+					mode_lib->ms.cache_display_cfg.output.OutputFormat[k],
+					mode_lib->ms.policy.ODMUse[k],
+					mode_lib->ms.state.dispclk_mhz,
+					mode_lib->ms.max_state.dispclk_mhz,
+					false, // DSCEnable
+					mode_lib->ms.TotalNumberOfActiveDPP[j],
+					mode_lib->ms.ip.max_num_dpp,
+					mode_lib->ms.cache_display_cfg.timing.PixelClock[k],
+					mode_lib->ms.soc.dcn_downspread_percent,
+					mode_lib->ms.ip.dispclk_ramp_margin_percent,
+					mode_lib->ms.soc.dispclk_dppclk_vco_speed_mhz,
+					mode_lib->ms.support.NumberOfDSCSlices[k],
+
+					/* Output */
+					&s->TotalAvailablePipesSupportNoDSC,
+					&s->NumberOfDPPNoDSC,
+					&s->ODMModeNoDSC,
+					&s->RequiredDISPCLKPerSurfaceNoDSC);
+
+			CalculateODMMode(
+					mode_lib->ms.ip.maximum_pixels_per_line_per_dsc_unit,
+					mode_lib->ms.cache_display_cfg.timing.HActive[k],
+					mode_lib->ms.cache_display_cfg.output.OutputEncoder[k],
+					mode_lib->ms.cache_display_cfg.output.OutputFormat[k],
+					mode_lib->ms.policy.ODMUse[k],
+					mode_lib->ms.state.dispclk_mhz,
+					mode_lib->ms.max_state.dispclk_mhz,
+					true, // DSCEnable
+					mode_lib->ms.TotalNumberOfActiveDPP[j],
+					mode_lib->ms.ip.max_num_dpp,
+					mode_lib->ms.cache_display_cfg.timing.PixelClock[k],
+					mode_lib->ms.soc.dcn_downspread_percent,
+					mode_lib->ms.ip.dispclk_ramp_margin_percent,
+					mode_lib->ms.soc.dispclk_dppclk_vco_speed_mhz,
+					mode_lib->ms.support.NumberOfDSCSlices[k],
+
+					/* Output */
+					&s->TotalAvailablePipesSupportDSC,
+					&s->NumberOfDPPDSC,
+					&s->ODMModeDSC,
+					&s->RequiredDISPCLKPerSurfaceDSC);
+
+			CalculateOutputLink(
+					mode_lib->ms.state.phyclk_mhz,
+					mode_lib->ms.state.phyclk_d18_mhz,
+					mode_lib->ms.state.phyclk_d32_mhz,
+					mode_lib->ms.soc.phy_downspread_percent,
+					(mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming[k] == k),
+					mode_lib->ms.cache_display_cfg.output.OutputEncoder[k],
+					mode_lib->ms.cache_display_cfg.output.OutputFormat[k],
+					mode_lib->ms.cache_display_cfg.timing.HTotal[k],
+					mode_lib->ms.cache_display_cfg.timing.HActive[k],
+					mode_lib->ms.cache_display_cfg.output.PixelClockBackEnd[k],
+					mode_lib->ms.cache_display_cfg.output.ForcedOutputLinkBPP[k],
+					mode_lib->ms.cache_display_cfg.output.DSCInputBitPerComponent[k],
+					mode_lib->ms.support.NumberOfDSCSlices[k],
+					mode_lib->ms.cache_display_cfg.output.AudioSampleRate[k],
+					mode_lib->ms.cache_display_cfg.output.AudioSampleLayout[k],
+					s->ODMModeNoDSC,
+					s->ODMModeDSC,
+					mode_lib->ms.cache_display_cfg.output.DSCEnable[k],
+					mode_lib->ms.cache_display_cfg.output.OutputLinkDPLanes[k],
+					mode_lib->ms.cache_display_cfg.output.OutputLinkDPRate[k],
+
+					/* Output */
+					&mode_lib->ms.RequiresDSC[k],
+					&mode_lib->ms.RequiresFEC[k],
+					&mode_lib->ms.OutputBppPerState[k],
+					&mode_lib->ms.OutputTypePerState[k], // VBA_DELTA, VBA uses a string to represent type and rate, but DML uses enum, don't want to rely on strng
+					&mode_lib->ms.OutputRatePerState[k],
+					&mode_lib->ms.RequiredSlots[k]);
+
+			if (mode_lib->ms.RequiresDSC[k] == false) {
+				mode_lib->ms.ODMModePerState[k] = s->ODMModeNoDSC;
+				mode_lib->ms.RequiredDISPCLKPerSurface[j][k] = s->RequiredDISPCLKPerSurfaceNoDSC;
+				if (!s->TotalAvailablePipesSupportNoDSC)
+					mode_lib->ms.support.TotalAvailablePipesSupport[j] = false;
+				mode_lib->ms.TotalNumberOfActiveDPP[j] = mode_lib->ms.TotalNumberOfActiveDPP[j] + s->NumberOfDPPNoDSC;
+			} else {
+				mode_lib->ms.ODMModePerState[k] = s->ODMModeDSC;
+				mode_lib->ms.RequiredDISPCLKPerSurface[j][k] = s->RequiredDISPCLKPerSurfaceDSC;
+				if (!s->TotalAvailablePipesSupportDSC)
+					mode_lib->ms.support.TotalAvailablePipesSupport[j] = false;
+				mode_lib->ms.TotalNumberOfActiveDPP[j] = mode_lib->ms.TotalNumberOfActiveDPP[j] + s->NumberOfDPPDSC;
+			}
+		}
+
+		for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+			if (mode_lib->ms.ODMModePerState[k] == dml_odm_mode_combine_4to1) {
+				mode_lib->ms.MPCCombine[j][k] = false;
+				mode_lib->ms.NoOfDPP[j][k] = 4;
+			} else if (mode_lib->ms.ODMModePerState[k] == dml_odm_mode_combine_2to1) {
+				mode_lib->ms.MPCCombine[j][k] = false;
+				mode_lib->ms.NoOfDPP[j][k] = 2;
+			} else if (mode_lib->ms.policy.MPCCombineUse[k] == dml_mpc_disabled) {
+				mode_lib->ms.MPCCombine[j][k] = false;
+				mode_lib->ms.NoOfDPP[j][k] = 1;
+			} else if (RoundToDFSGranularity(mode_lib->ms.MinDPPCLKUsingSingleDPP[k] * (1 + mode_lib->ms.soc.dcn_downspread_percent / 100),
+											1, mode_lib->ms.soc.dispclk_dppclk_vco_speed_mhz) <= mode_lib->ms.state.dppclk_mhz &&
+											mode_lib->ms.SingleDPPViewportSizeSupportPerSurface[k] == true) {
+				mode_lib->ms.MPCCombine[j][k] = false;
+				mode_lib->ms.NoOfDPP[j][k] = 1;
+			} else if (mode_lib->ms.TotalNumberOfActiveDPP[j] < (dml_uint_t) mode_lib->ms.ip.max_num_dpp) {
+				mode_lib->ms.MPCCombine[j][k] = true;
+				mode_lib->ms.NoOfDPP[j][k] = 2;
+				mode_lib->ms.TotalNumberOfActiveDPP[j] = (dml_uint_t) mode_lib->ms.TotalNumberOfActiveDPP[j] + 1;
+			} else {
+				mode_lib->ms.MPCCombine[j][k] = false;
+				mode_lib->ms.NoOfDPP[j][k] = 1;
+				mode_lib->ms.support.TotalAvailablePipesSupport[j] = false;
+			}
+		}
+
+		mode_lib->ms.TotalNumberOfSingleDPPSurfaces[j] = 0;
+		s->NoChromaOrLinear = true;
+		for (k = 0; k < (dml_uint_t) mode_lib->ms.num_active_planes; ++k) {
+			if (mode_lib->ms.NoOfDPP[j][k] == 1)
+				mode_lib->ms.TotalNumberOfSingleDPPSurfaces[j] = mode_lib->ms.TotalNumberOfSingleDPPSurfaces[j] + 1;
+			if (mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] == dml_420_8
+					|| mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] == dml_420_10
+					|| mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] == dml_420_12
+					|| mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] == dml_rgbe_alpha
+					|| mode_lib->ms.cache_display_cfg.surface.SurfaceTiling[k] == dml_sw_linear) {
+				s->NoChromaOrLinear = false;
+			}
+		}
+
+		if (j == 1 && !UnboundedRequest(mode_lib->ms.policy.UseUnboundedRequesting,
+				mode_lib->ms.TotalNumberOfActiveDPP[j], s->NoChromaOrLinear,
+				mode_lib->ms.cache_display_cfg.output.OutputEncoder[0])) {
+			while (!(mode_lib->ms.TotalNumberOfActiveDPP[j] >= (dml_uint_t) mode_lib->ms.ip.max_num_dpp || mode_lib->ms.TotalNumberOfSingleDPPSurfaces[j] == 0)) {
+				s->BWOfNonCombinedSurfaceOfMaximumBandwidth = 0;
+				s->NumberOfNonCombinedSurfaceOfMaximumBandwidth = 0;
+				for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+					if (mode_lib->ms.policy.MPCCombineUse[k] != dml_mpc_disabled && mode_lib->ms.policy.MPCCombineUse[k] != dml_mpc_as_needed_for_voltage &&
+						mode_lib->ms.ReadBandwidthLuma[k] + mode_lib->ms.ReadBandwidthChroma[k] > s->BWOfNonCombinedSurfaceOfMaximumBandwidth &&
+						(mode_lib->ms.ODMModePerState[k] != dml_odm_mode_combine_2to1 && mode_lib->ms.ODMModePerState[k] != dml_odm_mode_combine_4to1) &&
+						mode_lib->ms.MPCCombine[j][k] == false) {
+						s->BWOfNonCombinedSurfaceOfMaximumBandwidth = mode_lib->ms.ReadBandwidthLuma[k] + mode_lib->ms.ReadBandwidthChroma[k];
+						s->NumberOfNonCombinedSurfaceOfMaximumBandwidth = k;
+					}
+				}
+				mode_lib->ms.MPCCombine[j][s->NumberOfNonCombinedSurfaceOfMaximumBandwidth] = true;
+				mode_lib->ms.NoOfDPP[j][s->NumberOfNonCombinedSurfaceOfMaximumBandwidth] = 2;
+				mode_lib->ms.TotalNumberOfActiveDPP[j] = mode_lib->ms.TotalNumberOfActiveDPP[j] + 1;
+				mode_lib->ms.TotalNumberOfSingleDPPSurfaces[j] = mode_lib->ms.TotalNumberOfSingleDPPSurfaces[j] - 1;
+			}
+		}
+
+		//DISPCLK/DPPCLK
+		mode_lib->ms.WritebackRequiredDISPCLK = 0;
+		for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+			if (mode_lib->ms.cache_display_cfg.writeback.WritebackEnable[k]) {
+				mode_lib->ms.WritebackRequiredDISPCLK = dml_max(mode_lib->ms.WritebackRequiredDISPCLK,
+																	CalculateWriteBackDISPCLK(mode_lib->ms.cache_display_cfg.writeback.WritebackPixelFormat[k],
+																							mode_lib->ms.cache_display_cfg.timing.PixelClock[k],
+																							mode_lib->ms.cache_display_cfg.writeback.WritebackHRatio[k],
+																							mode_lib->ms.cache_display_cfg.writeback.WritebackVRatio[k],
+																							mode_lib->ms.cache_display_cfg.writeback.WritebackHTaps[k],
+																							mode_lib->ms.cache_display_cfg.writeback.WritebackVTaps[k],
+																							mode_lib->ms.cache_display_cfg.writeback.WritebackSourceWidth[k],
+																							mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationWidth[k],
+																							mode_lib->ms.cache_display_cfg.timing.HTotal[k],
+																							mode_lib->ms.ip.writeback_line_buffer_buffer_size,
+																							mode_lib->ms.soc.dispclk_dppclk_vco_speed_mhz));
+			}
+		}
+
+		 mode_lib->ms.RequiredDISPCLK[j] = mode_lib->ms.WritebackRequiredDISPCLK;
+		for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+			mode_lib->ms.RequiredDISPCLK[j] = dml_max(mode_lib->ms.RequiredDISPCLK[j], mode_lib->ms.RequiredDISPCLKPerSurface[j][k]);
+		}
+
+		for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+			mode_lib->ms.NoOfDPPThisState[k] = mode_lib->ms.NoOfDPP[j][k];
+		}
+
+		CalculateDPPCLK(mode_lib->ms.num_active_planes,
+					mode_lib->ms.soc.dcn_downspread_percent,
+					mode_lib->ms.soc.dispclk_dppclk_vco_speed_mhz,
+					mode_lib->ms.MinDPPCLKUsingSingleDPP,
+					mode_lib->ms.NoOfDPPThisState,
+					/* Output */
+					&mode_lib->ms.GlobalDPPCLK,
+					mode_lib->ms.RequiredDPPCLKThisState);
+
+		for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+			mode_lib->ms.RequiredDPPCLKPerSurface[j][k] = mode_lib->ms.RequiredDPPCLKThisState[k];
+		}
+
+		mode_lib->ms.support.DISPCLK_DPPCLK_Support[j] = !((mode_lib->ms.RequiredDISPCLK[j] > mode_lib->ms.state.dispclk_mhz) || (mode_lib->ms.GlobalDPPCLK > mode_lib->ms.state.dppclk_mhz));
+
+		if (mode_lib->ms.TotalNumberOfActiveDPP[j] > (dml_uint_t) mode_lib->ms.ip.max_num_dpp) {
+			mode_lib->ms.support.TotalAvailablePipesSupport[j] = false;
+		}
+	} // j
+
+	/* Total Available OTG, HDMIFRL, DP Support Check */
+	s->TotalNumberOfActiveOTG = 0;
+	s->TotalNumberOfActiveHDMIFRL = 0;
+	s->TotalNumberOfActiveDP2p0 = 0;
+	s->TotalNumberOfActiveDP2p0Outputs = 0;
+
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		if (mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming[k] == k) {
+			s->TotalNumberOfActiveOTG = s->TotalNumberOfActiveOTG + 1;
+			if (mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_hdmifrl)
+				s->TotalNumberOfActiveHDMIFRL = s->TotalNumberOfActiveHDMIFRL + 1;
+			if (mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_dp2p0) {
+				s->TotalNumberOfActiveDP2p0 = s->TotalNumberOfActiveDP2p0 + 1;
+				if (mode_lib->ms.cache_display_cfg.output.OutputMultistreamId[k] == k || mode_lib->ms.cache_display_cfg.output.OutputMultistreamEn[k] == false) {
+					s->TotalNumberOfActiveDP2p0Outputs = s->TotalNumberOfActiveDP2p0Outputs + 1;
+				}
+			}
+		}
+	}
+
+	mode_lib->ms.support.NumberOfOTGSupport      = (s->TotalNumberOfActiveOTG <= (dml_uint_t) mode_lib->ms.ip.max_num_otg);
+	mode_lib->ms.support.NumberOfHDMIFRLSupport  = (s->TotalNumberOfActiveHDMIFRL <= (dml_uint_t) mode_lib->ms.ip.max_num_hdmi_frl_outputs);
+	mode_lib->ms.support.NumberOfDP2p0Support    = (s->TotalNumberOfActiveDP2p0 <= (dml_uint_t) mode_lib->ms.ip.max_num_dp2p0_streams && s->TotalNumberOfActiveDP2p0Outputs <= (dml_uint_t) mode_lib->ms.ip.max_num_dp2p0_outputs);
+
+	/* Display IO and DSC Support Check */
+	mode_lib->ms.support.NonsupportedDSCInputBPC = false;
+	for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+		if (mode_lib->ms.cache_display_cfg.output.OutputDisabled[k] == false &&
+			!(mode_lib->ms.cache_display_cfg.output.DSCInputBitPerComponent[k] == 12.0
+				|| mode_lib->ms.cache_display_cfg.output.DSCInputBitPerComponent[k] == 10.0
+				|| mode_lib->ms.cache_display_cfg.output.DSCInputBitPerComponent[k] == 8.0
+				|| mode_lib->ms.cache_display_cfg.output.DSCInputBitPerComponent[k] > (dml_uint_t) mode_lib->ms.ip.maximum_dsc_bits_per_component
+				)) {
+			mode_lib->ms.support.NonsupportedDSCInputBPC = true;
+		}
+	}
+
+	mode_lib->ms.support.ExceededMultistreamSlots = false;
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		if (mode_lib->ms.cache_display_cfg.output.OutputMultistreamId[k] == k) {
+			s->TotalSlots = mode_lib->ms.RequiredSlots[k];
+			for (j = 0; j < mode_lib->ms.num_active_planes; ++j) {
+				if (mode_lib->ms.cache_display_cfg.output.OutputMultistreamId[j] == k)
+					s->TotalSlots = s->TotalSlots + mode_lib->ms.RequiredSlots[j];
+			}
+			if (mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_dp && s->TotalSlots > 63)
+				mode_lib->ms.support.ExceededMultistreamSlots = true;
+			if (mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_dp2p0 && s->TotalSlots > 64)
+				mode_lib->ms.support.ExceededMultistreamSlots = true;
+		}
+	}
+	mode_lib->ms.support.LinkCapacitySupport = true;
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		if (mode_lib->ms.cache_display_cfg.output.OutputDisabled[k] == false &&
+			mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming[k] == k && (mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_dp || mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_dp2p0 || mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_edp ||
+			mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_hdmi || mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_hdmifrl) && mode_lib->ms.OutputBppPerState[k] == 0) {
+			mode_lib->ms.support.LinkCapacitySupport = false;
+		}
+	}
+
+	mode_lib->ms.support.P2IWith420 = false;
+	mode_lib->ms.support.DSCOnlyIfNecessaryWithBPP = false;
+	mode_lib->ms.support.DSC422NativeNotSupported = false;
+	mode_lib->ms.support.LinkRateDoesNotMatchDPVersion = false;
+	mode_lib->ms.support.LinkRateForMultistreamNotIndicated = false;
+	mode_lib->ms.support.BPPForMultistreamNotIndicated = false;
+	mode_lib->ms.support.MultistreamWithHDMIOreDP = false;
+	mode_lib->ms.support.MSOOrODMSplitWithNonDPLink = false;
+	mode_lib->ms.support.NotEnoughLanesForMSO = false;
+
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		if (mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming[k] == k && (mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_dp || mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_dp2p0 || mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_edp ||
+														mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_hdmi || mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_hdmifrl)) {
+			if (mode_lib->ms.cache_display_cfg.output.OutputFormat[k] == dml_420 && mode_lib->ms.cache_display_cfg.timing.Interlace[k] == 1 && mode_lib->ms.ip.ptoi_supported == true)
+				mode_lib->ms.support.P2IWith420 = true;
+
+			if (mode_lib->ms.cache_display_cfg.output.DSCEnable[k] == dml_dsc_enable_if_necessary && mode_lib->ms.cache_display_cfg.output.ForcedOutputLinkBPP[k] != 0)
+				mode_lib->ms.support.DSCOnlyIfNecessaryWithBPP = true;
+			if ((mode_lib->ms.cache_display_cfg.output.DSCEnable[k] == dml_dsc_enable || mode_lib->ms.cache_display_cfg.output.DSCEnable[k] == dml_dsc_enable_if_necessary) && mode_lib->ms.cache_display_cfg.output.OutputFormat[k] == dml_n422 && !mode_lib->ms.ip.dsc422_native_support)
+				mode_lib->ms.support.DSC422NativeNotSupported = true;
+
+			if (((mode_lib->ms.cache_display_cfg.output.OutputLinkDPRate[k] == dml_dp_rate_hbr || mode_lib->ms.cache_display_cfg.output.OutputLinkDPRate[k] == dml_dp_rate_hbr2 || mode_lib->ms.cache_display_cfg.output.OutputLinkDPRate[k] == dml_dp_rate_hbr3) &&
+					mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] != dml_dp && mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] != dml_edp) ||
+					((mode_lib->ms.cache_display_cfg.output.OutputLinkDPRate[k] == dml_dp_rate_uhbr10 || mode_lib->ms.cache_display_cfg.output.OutputLinkDPRate[k] == dml_dp_rate_uhbr13p5 || mode_lib->ms.cache_display_cfg.output.OutputLinkDPRate[k] == dml_dp_rate_uhbr20) &&
+					mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] != dml_dp2p0))
+				mode_lib->ms.support.LinkRateDoesNotMatchDPVersion = true;
+
+			if (mode_lib->ms.cache_display_cfg.output.OutputMultistreamEn[k] == 1) {
+				if (mode_lib->ms.cache_display_cfg.output.OutputMultistreamId[k] == k && mode_lib->ms.cache_display_cfg.output.OutputLinkDPRate[k] == dml_dp_rate_na)
+					mode_lib->ms.support.LinkRateForMultistreamNotIndicated = true;
+				if (mode_lib->ms.cache_display_cfg.output.OutputMultistreamId[k] == k && mode_lib->ms.cache_display_cfg.output.ForcedOutputLinkBPP[k] == 0)
+					mode_lib->ms.support.BPPForMultistreamNotIndicated = true;
+				for (j = 0; j < mode_lib->ms.num_active_planes; ++j) {
+					if (mode_lib->ms.cache_display_cfg.output.OutputMultistreamId[k] == j && mode_lib->ms.cache_display_cfg.output.ForcedOutputLinkBPP[k] == 0)
+						mode_lib->ms.support.BPPForMultistreamNotIndicated = true;
+				}
+			}
+
+			if ((mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_edp || mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_hdmi || mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_hdmifrl)) {
+				if (mode_lib->ms.cache_display_cfg.output.OutputMultistreamEn[k] == 1 && mode_lib->ms.cache_display_cfg.output.OutputMultistreamId[k] == k)
+					mode_lib->ms.support.MultistreamWithHDMIOreDP = true;
+				for (j = 0; j < mode_lib->ms.num_active_planes; ++j) {
+					if (mode_lib->ms.cache_display_cfg.output.OutputMultistreamEn[k] == 1 && mode_lib->ms.cache_display_cfg.output.OutputMultistreamId[k] == j)
+						mode_lib->ms.support.MultistreamWithHDMIOreDP = true;
+				}
+			}
+			if (mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] != dml_dp && (mode_lib->ms.policy.ODMUse[k] == dml_odm_use_policy_split_1to2 ||
+				mode_lib->ms.policy.ODMUse[k] == dml_odm_use_policy_mso_1to2 || mode_lib->ms.policy.ODMUse[k] == dml_odm_use_policy_mso_1to4))
+				mode_lib->ms.support.MSOOrODMSplitWithNonDPLink = true;
+
+			if ((mode_lib->ms.policy.ODMUse[k] == dml_odm_use_policy_mso_1to2 && mode_lib->ms.cache_display_cfg.output.OutputLinkDPLanes[k] < 2) ||
+				(mode_lib->ms.policy.ODMUse[k] == dml_odm_use_policy_mso_1to4 && mode_lib->ms.cache_display_cfg.output.OutputLinkDPLanes[k] < 4))
+				mode_lib->ms.support.NotEnoughLanesForMSO = true;
+		}
+	}
+
+	mode_lib->ms.support.DTBCLKRequiredMoreThanSupported = false;
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		if (mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming[k] == k &&
+				mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_hdmifrl &&
+				RequiredDTBCLK(
+							mode_lib->ms.RequiresDSC[k],
+							mode_lib->ms.cache_display_cfg.output.PixelClockBackEnd[k],
+							mode_lib->ms.cache_display_cfg.output.OutputFormat[k],
+							mode_lib->ms.OutputBppPerState[k],
+							mode_lib->ms.support.NumberOfDSCSlices[k],
+							mode_lib->ms.cache_display_cfg.timing.HTotal[k],
+							mode_lib->ms.cache_display_cfg.timing.HActive[k],
+							mode_lib->ms.cache_display_cfg.output.AudioSampleRate[k],
+							mode_lib->ms.cache_display_cfg.output.AudioSampleLayout[k]) > mode_lib->ms.state.dtbclk_mhz) {
+								mode_lib->ms.support.DTBCLKRequiredMoreThanSupported = true;
+							}
+	}
+
+	mode_lib->ms.support.ODMCombineTwoToOneSupportCheckOK = true;
+	mode_lib->ms.support.ODMCombineFourToOneSupportCheckOK = true;
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		if (mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming[k] == k && mode_lib->ms.ODMModePerState[k] == dml_odm_mode_combine_2to1 && mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_hdmi) {
+			mode_lib->ms.support.ODMCombineTwoToOneSupportCheckOK = false;
+		}
+		if (mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming[k] == k && mode_lib->ms.ODMModePerState[k] == dml_odm_mode_combine_4to1 && (mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_dp ||
+			mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_edp || mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_hdmi)) {
+			mode_lib->ms.support.ODMCombineFourToOneSupportCheckOK = false;
+		}
+	}
+
+	mode_lib->ms.support.DSCCLKRequiredMoreThanSupported = false;
+	for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+		if (mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming[k] == k) {
+			if (mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_dp ||
+				mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_dp2p0 ||
+				mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_edp ||
+				mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_hdmifrl) {
+				if (mode_lib->ms.cache_display_cfg.output.OutputFormat[k] == dml_420) {
+					s->DSCFormatFactor = 2;
+				} else if (mode_lib->ms.cache_display_cfg.output.OutputFormat[k] == dml_444) {
+					s->DSCFormatFactor = 1;
+				} else if (mode_lib->ms.cache_display_cfg.output.OutputFormat[k] == dml_n422 || mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_hdmifrl) {
+					s->DSCFormatFactor = 2;
+				} else {
+					s->DSCFormatFactor = 1;
+				}
+#ifdef __DML_VBA_DEBUG__
+				dml_print("DML::%s: k=%u, RequiresDSC = %u\n",  __func__, k, mode_lib->ms.RequiresDSC[k]);
+#endif
+				if (mode_lib->ms.RequiresDSC[k] == true) {
+					if (mode_lib->ms.ODMModePerState[k] == dml_odm_mode_combine_4to1) {
+						if (mode_lib->ms.cache_display_cfg.output.PixelClockBackEnd[k] / 12.0 / (dml_float_t)s->DSCFormatFactor > (1.0 - mode_lib->ms.soc.dcn_downspread_percent / 100.0) * mode_lib->ms.state.dscclk_mhz) {
+#ifdef __DML_VBA_DEBUG__
+							dml_print("DML::%s: k=%u, PixelClockBackEnd     = %f\n",  __func__, k, mode_lib->ms.cache_display_cfg.output.PixelClockBackEnd[k]);
+							dml_print("DML::%s: k=%u, DSCCLKPerState        = %f\n",  __func__, k, mode_lib->ms.state.dscclk_mhz);
+							dml_print("DML::%s: k=%u, DSCFormatFactor       = %u\n",  __func__, k, s->DSCFormatFactor);
+#endif
+							mode_lib->ms.support.DSCCLKRequiredMoreThanSupported = true;
+						}
+					} else if (mode_lib->ms.ODMModePerState[k] == dml_odm_mode_combine_2to1) {
+						if (mode_lib->ms.cache_display_cfg.output.PixelClockBackEnd[k] / 6.0 / (dml_float_t)s->DSCFormatFactor > (1.0 - mode_lib->ms.soc.dcn_downspread_percent / 100.0) * mode_lib->ms.state.dscclk_mhz) {
+							mode_lib->ms.support.DSCCLKRequiredMoreThanSupported = true;
+						}
+					} else {
+						if (mode_lib->ms.cache_display_cfg.output.PixelClockBackEnd[k] / 3.0 / (dml_float_t)s->DSCFormatFactor > (1.0 - mode_lib->ms.soc.dcn_downspread_percent / 100.0) * mode_lib->ms.state.dscclk_mhz) {
+							mode_lib->ms.support.DSCCLKRequiredMoreThanSupported = true;
+						}
+					}
+				}
+			}
+		}
+	}
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: DSCCLKRequiredMoreThanSupported = %u\n",  __func__, mode_lib->ms.support.DSCCLKRequiredMoreThanSupported);
+#endif
+
+	/* Check DSC Unit and Slices Support */
+	mode_lib->ms.support.NotEnoughDSCUnits = false;
+	mode_lib->ms.support.NotEnoughDSCSlices = false;
+	s->TotalDSCUnitsRequired = 0;
+	mode_lib->ms.support.PixelsPerLinePerDSCUnitSupport = true;
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		if (mode_lib->ms.RequiresDSC[k] == true) {
+			if (mode_lib->ms.ODMModePerState[k] == dml_odm_mode_combine_4to1) {
+				if (mode_lib->ms.cache_display_cfg.timing.HActive[k] > 4 * (dml_uint_t) mode_lib->ms.ip.maximum_pixels_per_line_per_dsc_unit)
+					mode_lib->ms.support.PixelsPerLinePerDSCUnitSupport = false;
+				s->TotalDSCUnitsRequired = s->TotalDSCUnitsRequired + 4;
+				if (mode_lib->ms.support.NumberOfDSCSlices[k] > 16)
+					mode_lib->ms.support.NotEnoughDSCSlices = true;
+			} else if (mode_lib->ms.ODMModePerState[k] == dml_odm_mode_combine_2to1) {
+				if (mode_lib->ms.cache_display_cfg.timing.HActive[k] > 2 * (dml_uint_t) mode_lib->ms.ip.maximum_pixels_per_line_per_dsc_unit)
+					mode_lib->ms.support.PixelsPerLinePerDSCUnitSupport = false;
+				s->TotalDSCUnitsRequired = s->TotalDSCUnitsRequired + 2;
+				if (mode_lib->ms.support.NumberOfDSCSlices[k] > 8)
+					mode_lib->ms.support.NotEnoughDSCSlices = true;
+			} else {
+				if (mode_lib->ms.cache_display_cfg.timing.HActive[k] > (dml_uint_t) mode_lib->ms.ip.maximum_pixels_per_line_per_dsc_unit)
+					mode_lib->ms.support.PixelsPerLinePerDSCUnitSupport = false;
+				s->TotalDSCUnitsRequired = s->TotalDSCUnitsRequired + 1;
+				if (mode_lib->ms.support.NumberOfDSCSlices[k] > 4)
+					mode_lib->ms.support.NotEnoughDSCSlices = true;
+			}
+		}
+	}
+   if (s->TotalDSCUnitsRequired > (dml_uint_t) mode_lib->ms.ip.num_dsc) {
+		mode_lib->ms.support.NotEnoughDSCUnits = true;
+	}
+
+	/*DSC Delay per state*/
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		mode_lib->ms.DSCDelayPerState[k] = DSCDelayRequirement(mode_lib->ms.RequiresDSC[k],
+													mode_lib->ms.ODMModePerState[k],
+													mode_lib->ms.cache_display_cfg.output.DSCInputBitPerComponent[k],
+													mode_lib->ms.OutputBppPerState[k],
+													mode_lib->ms.cache_display_cfg.timing.HActive[k],
+													mode_lib->ms.cache_display_cfg.timing.HTotal[k],
+													mode_lib->ms.support.NumberOfDSCSlices[k],
+													mode_lib->ms.cache_display_cfg.output.OutputFormat[k],
+													mode_lib->ms.cache_display_cfg.output.OutputEncoder[k],
+													mode_lib->ms.cache_display_cfg.timing.PixelClock[k],
+													mode_lib->ms.cache_display_cfg.output.PixelClockBackEnd[k]);
+	}
+
+	for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+		for (m = 0; m <= mode_lib->ms.num_active_planes - 1; m++) {
+			for (j = 0; j <= mode_lib->ms.num_active_planes - 1; j++) {
+				if (mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming[k] == m && mode_lib->ms.RequiresDSC[m] == true) {
+					mode_lib->ms.DSCDelayPerState[k] = mode_lib->ms.DSCDelayPerState[m];
+				}
+			}
+		}
+	}
+
+	//Calculate Swath, DET Configuration, DCFCLKDeepSleep
+	//
+	for (j = 0; j < 2; ++j) {
+		for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+			mode_lib->ms.RequiredDPPCLKThisState[k] = mode_lib->ms.RequiredDPPCLKPerSurface[j][k];
+			mode_lib->ms.NoOfDPPThisState[k] = mode_lib->ms.NoOfDPP[j][k];
+			mode_lib->ms.ODMModeThisState[k] = mode_lib->ms.ODMModePerState[k];
+		}
+
+		CalculateSwathAndDETConfiguration_params->DETSizeOverride = mode_lib->ms.cache_display_cfg.plane.DETSizeOverride;
+		CalculateSwathAndDETConfiguration_params->UseMALLForPStateChange = mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange;
+		CalculateSwathAndDETConfiguration_params->ConfigReturnBufferSizeInKByte = mode_lib->ms.ip.config_return_buffer_size_in_kbytes;
+		CalculateSwathAndDETConfiguration_params->ROBBufferSizeInKByte = mode_lib->ms.ip.rob_buffer_size_kbytes;
+		CalculateSwathAndDETConfiguration_params->MaxTotalDETInKByte = mode_lib->ms.MaxTotalDETInKByte;
+		CalculateSwathAndDETConfiguration_params->MinCompressedBufferSizeInKByte = mode_lib->ms.MinCompressedBufferSizeInKByte;
+		CalculateSwathAndDETConfiguration_params->PixelChunkSizeInKByte = mode_lib->ms.ip.pixel_chunk_size_kbytes;
+		CalculateSwathAndDETConfiguration_params->ForceSingleDPP = false;
+		CalculateSwathAndDETConfiguration_params->NumberOfActiveSurfaces = mode_lib->ms.num_active_planes;
+		CalculateSwathAndDETConfiguration_params->nomDETInKByte = mode_lib->ms.NomDETInKByte;
+		CalculateSwathAndDETConfiguration_params->UseUnboundedRequestingFinal = mode_lib->ms.policy.UseUnboundedRequesting;
+		CalculateSwathAndDETConfiguration_params->ConfigReturnBufferSegmentSizeInkByte = mode_lib->ms.ip.config_return_buffer_segment_size_in_kbytes;
+		CalculateSwathAndDETConfiguration_params->CompressedBufferSegmentSizeInkByteFinal = mode_lib->ms.ip.compressed_buffer_segment_size_in_kbytes;
+		CalculateSwathAndDETConfiguration_params->Output = mode_lib->ms.cache_display_cfg.output.OutputEncoder;
+		CalculateSwathAndDETConfiguration_params->ReadBandwidthLuma = mode_lib->ms.ReadBandwidthLuma;
+		CalculateSwathAndDETConfiguration_params->ReadBandwidthChroma = mode_lib->ms.ReadBandwidthChroma;
+		CalculateSwathAndDETConfiguration_params->MaximumSwathWidthLuma = mode_lib->ms.MaximumSwathWidthLuma;
+		CalculateSwathAndDETConfiguration_params->MaximumSwathWidthChroma = mode_lib->ms.MaximumSwathWidthChroma;
+		CalculateSwathAndDETConfiguration_params->SourceScan = mode_lib->ms.cache_display_cfg.plane.SourceScan;
+		CalculateSwathAndDETConfiguration_params->ViewportStationary = mode_lib->ms.cache_display_cfg.plane.ViewportStationary;
+		CalculateSwathAndDETConfiguration_params->SourcePixelFormat = mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat;
+		CalculateSwathAndDETConfiguration_params->SurfaceTiling = mode_lib->ms.cache_display_cfg.surface.SurfaceTiling;
+		CalculateSwathAndDETConfiguration_params->ViewportWidth = mode_lib->ms.cache_display_cfg.plane.ViewportWidth;
+		CalculateSwathAndDETConfiguration_params->ViewportHeight = mode_lib->ms.cache_display_cfg.plane.ViewportHeight;
+		CalculateSwathAndDETConfiguration_params->ViewportXStart = mode_lib->ms.cache_display_cfg.plane.ViewportXStart;
+		CalculateSwathAndDETConfiguration_params->ViewportYStart = mode_lib->ms.cache_display_cfg.plane.ViewportYStart;
+		CalculateSwathAndDETConfiguration_params->ViewportXStartC = mode_lib->ms.cache_display_cfg.plane.ViewportXStartC;
+		CalculateSwathAndDETConfiguration_params->ViewportYStartC = mode_lib->ms.cache_display_cfg.plane.ViewportYStartC;
+		CalculateSwathAndDETConfiguration_params->SurfaceWidthY = mode_lib->ms.cache_display_cfg.surface.SurfaceWidthY;
+		CalculateSwathAndDETConfiguration_params->SurfaceWidthC = mode_lib->ms.cache_display_cfg.surface.SurfaceWidthC;
+		CalculateSwathAndDETConfiguration_params->SurfaceHeightY = mode_lib->ms.cache_display_cfg.surface.SurfaceHeightY;
+		CalculateSwathAndDETConfiguration_params->SurfaceHeightC = mode_lib->ms.cache_display_cfg.surface.SurfaceHeightC;
+		CalculateSwathAndDETConfiguration_params->Read256BytesBlockHeightY = mode_lib->ms.Read256BlockHeightY;
+		CalculateSwathAndDETConfiguration_params->Read256BytesBlockHeightC = mode_lib->ms.Read256BlockHeightC;
+		CalculateSwathAndDETConfiguration_params->Read256BytesBlockWidthY = mode_lib->ms.Read256BlockWidthY;
+		CalculateSwathAndDETConfiguration_params->Read256BytesBlockWidthC = mode_lib->ms.Read256BlockWidthC;
+		CalculateSwathAndDETConfiguration_params->ODMMode = mode_lib->ms.ODMModeThisState;
+		CalculateSwathAndDETConfiguration_params->BlendingAndTiming = mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming;
+		CalculateSwathAndDETConfiguration_params->BytePerPixY = mode_lib->ms.BytePerPixelY;
+		CalculateSwathAndDETConfiguration_params->BytePerPixC = mode_lib->ms.BytePerPixelC;
+		CalculateSwathAndDETConfiguration_params->BytePerPixDETY = mode_lib->ms.BytePerPixelInDETY;
+		CalculateSwathAndDETConfiguration_params->BytePerPixDETC = mode_lib->ms.BytePerPixelInDETC;
+		CalculateSwathAndDETConfiguration_params->HActive = mode_lib->ms.cache_display_cfg.timing.HActive;
+		CalculateSwathAndDETConfiguration_params->HRatio = mode_lib->ms.cache_display_cfg.plane.HRatio;
+		CalculateSwathAndDETConfiguration_params->HRatioChroma = mode_lib->ms.cache_display_cfg.plane.HRatioChroma;
+		CalculateSwathAndDETConfiguration_params->DPPPerSurface = mode_lib->ms.NoOfDPPThisState;
+		CalculateSwathAndDETConfiguration_params->swath_width_luma_ub = mode_lib->ms.swath_width_luma_ub_this_state;
+		CalculateSwathAndDETConfiguration_params->swath_width_chroma_ub = mode_lib->ms.swath_width_chroma_ub_this_state;
+		CalculateSwathAndDETConfiguration_params->SwathWidth = mode_lib->ms.SwathWidthYThisState;
+		CalculateSwathAndDETConfiguration_params->SwathWidthChroma = mode_lib->ms.SwathWidthCThisState;
+		CalculateSwathAndDETConfiguration_params->SwathHeightY = mode_lib->ms.SwathHeightYThisState;
+		CalculateSwathAndDETConfiguration_params->SwathHeightC = mode_lib->ms.SwathHeightCThisState;
+		CalculateSwathAndDETConfiguration_params->DETBufferSizeInKByte = mode_lib->ms.DETBufferSizeInKByteThisState;
+		CalculateSwathAndDETConfiguration_params->DETBufferSizeY = mode_lib->ms.DETBufferSizeYThisState;
+		CalculateSwathAndDETConfiguration_params->DETBufferSizeC = mode_lib->ms.DETBufferSizeCThisState;
+		CalculateSwathAndDETConfiguration_params->UnboundedRequestEnabled = &mode_lib->ms.UnboundedRequestEnabledThisState;
+		CalculateSwathAndDETConfiguration_params->compbuf_reserved_space_64b = &s->dummy_integer[2];
+		CalculateSwathAndDETConfiguration_params->compbuf_reserved_space_zs = &s->dummy_integer[1];
+		CalculateSwathAndDETConfiguration_params->CompressedBufferSizeInkByte = &mode_lib->ms.CompressedBufferSizeInkByteThisState;
+		CalculateSwathAndDETConfiguration_params->ViewportSizeSupportPerSurface = s->dummy_boolean_array[0];
+		CalculateSwathAndDETConfiguration_params->ViewportSizeSupport = &mode_lib->ms.support.ViewportSizeSupport[j];
+
+		CalculateSwathAndDETConfiguration(&mode_lib->scratch,
+		CalculateSwathAndDETConfiguration_params);
+
+		for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+			mode_lib->ms.swath_width_luma_ub_all_states[j][k] = mode_lib->ms.swath_width_luma_ub_this_state[k];
+			mode_lib->ms.swath_width_chroma_ub_all_states[j][k] = mode_lib->ms.swath_width_chroma_ub_this_state[k];
+			mode_lib->ms.SwathWidthYAllStates[j][k] = mode_lib->ms.SwathWidthYThisState[k];
+			mode_lib->ms.SwathWidthCAllStates[j][k] = mode_lib->ms.SwathWidthCThisState[k];
+			mode_lib->ms.SwathHeightYAllStates[j][k] = mode_lib->ms.SwathHeightYThisState[k];
+			mode_lib->ms.SwathHeightCAllStates[j][k] = mode_lib->ms.SwathHeightCThisState[k];
+			mode_lib->ms.UnboundedRequestEnabledAllStates[j] = mode_lib->ms.UnboundedRequestEnabledThisState;
+			mode_lib->ms.CompressedBufferSizeInkByteAllStates[j] = mode_lib->ms.CompressedBufferSizeInkByteThisState;
+			mode_lib->ms.DETBufferSizeInKByteAllStates[j][k] = mode_lib->ms.DETBufferSizeInKByteThisState[k];
+			mode_lib->ms.DETBufferSizeYAllStates[j][k] = mode_lib->ms.DETBufferSizeYThisState[k];
+			mode_lib->ms.DETBufferSizeCAllStates[j][k] = mode_lib->ms.DETBufferSizeCThisState[k];
+		}
+	}
+
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		mode_lib->ms.cursor_bw[k] = mode_lib->ms.cache_display_cfg.plane.NumberOfCursors[k] * mode_lib->ms.cache_display_cfg.plane.CursorWidth[k] * mode_lib->ms.cache_display_cfg.plane.CursorBPP[k] / 8.0 / (mode_lib->ms.cache_display_cfg.timing.HTotal[k] / mode_lib->ms.cache_display_cfg.timing.PixelClock[k]) * mode_lib->ms.cache_display_cfg.plane.VRatio[k];
+	}
+
+	CalculateSurfaceSizeInMall(
+			mode_lib->ms.num_active_planes,
+			mode_lib->ms.soc.mall_allocated_for_dcn_mbytes,
+			mode_lib->ms.cache_display_cfg.plane.UseMALLForStaticScreen,
+			mode_lib->ms.cache_display_cfg.surface.DCCEnable,
+			mode_lib->ms.cache_display_cfg.plane.ViewportStationary,
+			mode_lib->ms.cache_display_cfg.plane.ViewportXStart,
+			mode_lib->ms.cache_display_cfg.plane.ViewportYStart,
+			mode_lib->ms.cache_display_cfg.plane.ViewportXStartC,
+			mode_lib->ms.cache_display_cfg.plane.ViewportYStartC,
+			mode_lib->ms.cache_display_cfg.plane.ViewportWidth,
+			mode_lib->ms.cache_display_cfg.plane.ViewportHeight,
+			mode_lib->ms.BytePerPixelY,
+			mode_lib->ms.cache_display_cfg.plane.ViewportWidthChroma,
+			mode_lib->ms.cache_display_cfg.plane.ViewportHeightChroma,
+			mode_lib->ms.BytePerPixelC,
+			mode_lib->ms.cache_display_cfg.surface.SurfaceWidthY,
+			mode_lib->ms.cache_display_cfg.surface.SurfaceWidthC,
+			mode_lib->ms.cache_display_cfg.surface.SurfaceHeightY,
+			mode_lib->ms.cache_display_cfg.surface.SurfaceHeightC,
+			mode_lib->ms.Read256BlockWidthY,
+			mode_lib->ms.Read256BlockWidthC,
+			mode_lib->ms.Read256BlockHeightY,
+			mode_lib->ms.Read256BlockHeightC,
+			mode_lib->ms.MacroTileWidthY,
+			mode_lib->ms.MacroTileWidthC,
+			mode_lib->ms.MacroTileHeightY,
+			mode_lib->ms.MacroTileHeightC,
+
+			/* Output */
+			mode_lib->ms.SurfaceSizeInMALL,
+			&mode_lib->ms.support.ExceededMALLSize);
+
+	for (j = 0; j < 2; j++) {
+		for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+			mode_lib->ms.swath_width_luma_ub_this_state[k] = mode_lib->ms.swath_width_luma_ub_all_states[j][k];
+			mode_lib->ms.swath_width_chroma_ub_this_state[k] = mode_lib->ms.swath_width_chroma_ub_all_states[j][k];
+			mode_lib->ms.SwathWidthYThisState[k] = mode_lib->ms.SwathWidthYAllStates[j][k];
+			mode_lib->ms.SwathWidthCThisState[k] = mode_lib->ms.SwathWidthCAllStates[j][k];
+			mode_lib->ms.SwathHeightYThisState[k] = mode_lib->ms.SwathHeightYAllStates[j][k];
+			mode_lib->ms.SwathHeightCThisState[k] = mode_lib->ms.SwathHeightCAllStates[j][k];
+			mode_lib->ms.DETBufferSizeInKByteThisState[k] = mode_lib->ms.DETBufferSizeInKByteAllStates[j][k];
+			mode_lib->ms.DETBufferSizeYThisState[k] = mode_lib->ms.DETBufferSizeYAllStates[j][k];
+			mode_lib->ms.DETBufferSizeCThisState[k] = mode_lib->ms.DETBufferSizeCAllStates[j][k];
+			mode_lib->ms.RequiredDPPCLKThisState[k] = mode_lib->ms.RequiredDPPCLKPerSurface[j][k];
+			mode_lib->ms.NoOfDPPThisState[k] = mode_lib->ms.NoOfDPP[j][k];
+		}
+
+		mode_lib->ms.TotalNumberOfDCCActiveDPP[j] = 0;
+		for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+			if (mode_lib->ms.cache_display_cfg.surface.DCCEnable[k] == true) {
+				mode_lib->ms.TotalNumberOfDCCActiveDPP[j] = mode_lib->ms.TotalNumberOfDCCActiveDPP[j] + mode_lib->ms.NoOfDPP[j][k];
+			}
+		}
+
+		for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+			s->SurfParameters[k].PixelClock = mode_lib->ms.cache_display_cfg.timing.PixelClock[k];
+			s->SurfParameters[k].DPPPerSurface = mode_lib->ms.NoOfDPP[j][k];
+			s->SurfParameters[k].SourceScan = mode_lib->ms.cache_display_cfg.plane.SourceScan[k];
+			s->SurfParameters[k].ViewportHeight = mode_lib->ms.cache_display_cfg.plane.ViewportHeight[k];
+			s->SurfParameters[k].ViewportHeightChroma = mode_lib->ms.cache_display_cfg.plane.ViewportHeightChroma[k];
+			s->SurfParameters[k].BlockWidth256BytesY = mode_lib->ms.Read256BlockWidthY[k];
+			s->SurfParameters[k].BlockHeight256BytesY = mode_lib->ms.Read256BlockHeightY[k];
+			s->SurfParameters[k].BlockWidth256BytesC = mode_lib->ms.Read256BlockWidthC[k];
+			s->SurfParameters[k].BlockHeight256BytesC = mode_lib->ms.Read256BlockHeightC[k];
+			s->SurfParameters[k].BlockWidthY = mode_lib->ms.MacroTileWidthY[k];
+			s->SurfParameters[k].BlockHeightY = mode_lib->ms.MacroTileHeightY[k];
+			s->SurfParameters[k].BlockWidthC = mode_lib->ms.MacroTileWidthC[k];
+			s->SurfParameters[k].BlockHeightC = mode_lib->ms.MacroTileHeightC[k];
+			s->SurfParameters[k].InterlaceEnable = mode_lib->ms.cache_display_cfg.timing.Interlace[k];
+			s->SurfParameters[k].HTotal = mode_lib->ms.cache_display_cfg.timing.HTotal[k];
+			s->SurfParameters[k].DCCEnable = mode_lib->ms.cache_display_cfg.surface.DCCEnable[k];
+			s->SurfParameters[k].SourcePixelFormat = mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k];
+			s->SurfParameters[k].SurfaceTiling = mode_lib->ms.cache_display_cfg.surface.SurfaceTiling[k];
+			s->SurfParameters[k].BytePerPixelY = mode_lib->ms.BytePerPixelY[k];
+			s->SurfParameters[k].BytePerPixelC = mode_lib->ms.BytePerPixelC[k];
+			s->SurfParameters[k].ProgressiveToInterlaceUnitInOPP = mode_lib->ms.ip.ptoi_supported;
+			s->SurfParameters[k].VRatio = mode_lib->ms.cache_display_cfg.plane.VRatio[k];
+			s->SurfParameters[k].VRatioChroma = mode_lib->ms.cache_display_cfg.plane.VRatioChroma[k];
+			s->SurfParameters[k].VTaps = mode_lib->ms.cache_display_cfg.plane.VTaps[k];
+			s->SurfParameters[k].VTapsChroma = mode_lib->ms.cache_display_cfg.plane.VTapsChroma[k];
+			s->SurfParameters[k].PitchY = mode_lib->ms.cache_display_cfg.surface.PitchY[k];
+			s->SurfParameters[k].DCCMetaPitchY = mode_lib->ms.cache_display_cfg.surface.DCCMetaPitchY[k];
+			s->SurfParameters[k].PitchC = mode_lib->ms.cache_display_cfg.surface.PitchC[k];
+			s->SurfParameters[k].DCCMetaPitchC = mode_lib->ms.cache_display_cfg.surface.DCCMetaPitchC[k];
+			s->SurfParameters[k].ViewportStationary = mode_lib->ms.cache_display_cfg.plane.ViewportStationary[k];
+			s->SurfParameters[k].ViewportXStart = mode_lib->ms.cache_display_cfg.plane.ViewportXStart[k];
+			s->SurfParameters[k].ViewportYStart = mode_lib->ms.cache_display_cfg.plane.ViewportYStart[k];
+			s->SurfParameters[k].ViewportXStartC = mode_lib->ms.cache_display_cfg.plane.ViewportXStartC[k];
+			s->SurfParameters[k].ViewportYStartC = mode_lib->ms.cache_display_cfg.plane.ViewportYStartC[k];
+			s->SurfParameters[k].FORCE_ONE_ROW_FOR_FRAME = mode_lib->ms.cache_display_cfg.plane.ForceOneRowForFrame[k];
+			s->SurfParameters[k].SwathHeightY = mode_lib->ms.SwathHeightYThisState[k];
+			s->SurfParameters[k].SwathHeightC = mode_lib->ms.SwathHeightCThisState[k];
+		}
+
+		CalculateVMRowAndSwath_params->NumberOfActiveSurfaces = mode_lib->ms.num_active_planes;
+		CalculateVMRowAndSwath_params->myPipe = s->SurfParameters;
+		CalculateVMRowAndSwath_params->SurfaceSizeInMALL = mode_lib->ms.SurfaceSizeInMALL;
+		CalculateVMRowAndSwath_params->PTEBufferSizeInRequestsLuma = mode_lib->ms.ip.dpte_buffer_size_in_pte_reqs_luma;
+		CalculateVMRowAndSwath_params->PTEBufferSizeInRequestsChroma = mode_lib->ms.ip.dpte_buffer_size_in_pte_reqs_chroma;
+		CalculateVMRowAndSwath_params->DCCMetaBufferSizeBytes = mode_lib->ms.ip.dcc_meta_buffer_size_bytes;
+		CalculateVMRowAndSwath_params->UseMALLForStaticScreen = mode_lib->ms.cache_display_cfg.plane.UseMALLForStaticScreen;
+		CalculateVMRowAndSwath_params->UseMALLForPStateChange = mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange;
+		CalculateVMRowAndSwath_params->MALLAllocatedForDCN = mode_lib->ms.soc.mall_allocated_for_dcn_mbytes;
+		CalculateVMRowAndSwath_params->SwathWidthY = mode_lib->ms.SwathWidthYThisState;
+		CalculateVMRowAndSwath_params->SwathWidthC = mode_lib->ms.SwathWidthCThisState;
+		CalculateVMRowAndSwath_params->GPUVMEnable = mode_lib->ms.cache_display_cfg.plane.GPUVMEnable;
+		CalculateVMRowAndSwath_params->HostVMEnable = mode_lib->ms.cache_display_cfg.plane.HostVMEnable;
+		CalculateVMRowAndSwath_params->HostVMMaxNonCachedPageTableLevels = mode_lib->ms.cache_display_cfg.plane.HostVMMaxPageTableLevels;
+		CalculateVMRowAndSwath_params->GPUVMMaxPageTableLevels = mode_lib->ms.cache_display_cfg.plane.GPUVMMaxPageTableLevels;
+		CalculateVMRowAndSwath_params->GPUVMMinPageSizeKBytes = mode_lib->ms.cache_display_cfg.plane.GPUVMMinPageSizeKBytes;
+		CalculateVMRowAndSwath_params->HostVMMinPageSize = mode_lib->ms.soc.hostvm_min_page_size_kbytes * 1024;
+		CalculateVMRowAndSwath_params->PTEBufferModeOverrideEn = mode_lib->ms.cache_display_cfg.plane.PTEBufferModeOverrideEn;
+		CalculateVMRowAndSwath_params->PTEBufferModeOverrideVal = mode_lib->ms.cache_display_cfg.plane.PTEBufferMode;
+		CalculateVMRowAndSwath_params->PTEBufferSizeNotExceeded = mode_lib->ms.PTEBufferSizeNotExceededPerState;
+		CalculateVMRowAndSwath_params->DCCMetaBufferSizeNotExceeded = mode_lib->ms.DCCMetaBufferSizeNotExceededPerState;
+		CalculateVMRowAndSwath_params->dpte_row_width_luma_ub = s->dummy_integer_array[0];
+		CalculateVMRowAndSwath_params->dpte_row_width_chroma_ub = s->dummy_integer_array[1];
+		CalculateVMRowAndSwath_params->dpte_row_height_luma = mode_lib->ms.dpte_row_height;
+		CalculateVMRowAndSwath_params->dpte_row_height_chroma = mode_lib->ms.dpte_row_height_chroma;
+		CalculateVMRowAndSwath_params->dpte_row_height_linear_luma = s->dummy_integer_array[2]; // VBA_DELTA
+		CalculateVMRowAndSwath_params->dpte_row_height_linear_chroma = s->dummy_integer_array[3]; // VBA_DELTA
+		CalculateVMRowAndSwath_params->meta_req_width = s->dummy_integer_array[4];
+		CalculateVMRowAndSwath_params->meta_req_width_chroma = s->dummy_integer_array[5];
+		CalculateVMRowAndSwath_params->meta_req_height = s->dummy_integer_array[6];
+		CalculateVMRowAndSwath_params->meta_req_height_chroma = s->dummy_integer_array[7];
+		CalculateVMRowAndSwath_params->meta_row_width = s->dummy_integer_array[8];
+		CalculateVMRowAndSwath_params->meta_row_width_chroma = s->dummy_integer_array[9];
+		CalculateVMRowAndSwath_params->meta_row_height = mode_lib->ms.meta_row_height;
+		CalculateVMRowAndSwath_params->meta_row_height_chroma = mode_lib->ms.meta_row_height_chroma;
+		CalculateVMRowAndSwath_params->vm_group_bytes = s->dummy_integer_array[10];
+		CalculateVMRowAndSwath_params->dpte_group_bytes = mode_lib->ms.dpte_group_bytes;
+		CalculateVMRowAndSwath_params->PixelPTEReqWidthY = s->dummy_integer_array[11];
+		CalculateVMRowAndSwath_params->PixelPTEReqHeightY = s->dummy_integer_array[12];
+		CalculateVMRowAndSwath_params->PTERequestSizeY = s->dummy_integer_array[13];
+		CalculateVMRowAndSwath_params->PixelPTEReqWidthC = s->dummy_integer_array[14];
+		CalculateVMRowAndSwath_params->PixelPTEReqHeightC = s->dummy_integer_array[15];
+		CalculateVMRowAndSwath_params->PTERequestSizeC = s->dummy_integer_array[16];
+		CalculateVMRowAndSwath_params->dpde0_bytes_per_frame_ub_l = s->dummy_integer_array[17];
+		CalculateVMRowAndSwath_params->meta_pte_bytes_per_frame_ub_l = s->dummy_integer_array[18];
+		CalculateVMRowAndSwath_params->dpde0_bytes_per_frame_ub_c = s->dummy_integer_array[19];
+		CalculateVMRowAndSwath_params->meta_pte_bytes_per_frame_ub_c = s->dummy_integer_array[20];
+		CalculateVMRowAndSwath_params->PrefetchSourceLinesY = mode_lib->ms.PrefetchLinesYThisState;
+		CalculateVMRowAndSwath_params->PrefetchSourceLinesC = mode_lib->ms.PrefetchLinesCThisState;
+		CalculateVMRowAndSwath_params->VInitPreFillY = mode_lib->ms.PrefillY;
+		CalculateVMRowAndSwath_params->VInitPreFillC = mode_lib->ms.PrefillC;
+		CalculateVMRowAndSwath_params->MaxNumSwathY = mode_lib->ms.MaxNumSwY;
+		CalculateVMRowAndSwath_params->MaxNumSwathC = mode_lib->ms.MaxNumSwC;
+		CalculateVMRowAndSwath_params->meta_row_bw = mode_lib->ms.meta_row_bandwidth_this_state;
+		CalculateVMRowAndSwath_params->dpte_row_bw = mode_lib->ms.dpte_row_bandwidth_this_state;
+		CalculateVMRowAndSwath_params->PixelPTEBytesPerRow = mode_lib->ms.DPTEBytesPerRowThisState;
+		CalculateVMRowAndSwath_params->PDEAndMetaPTEBytesFrame = mode_lib->ms.PDEAndMetaPTEBytesPerFrameThisState;
+		CalculateVMRowAndSwath_params->MetaRowByte = mode_lib->ms.MetaRowBytesThisState;
+		CalculateVMRowAndSwath_params->use_one_row_for_frame = mode_lib->ms.use_one_row_for_frame_this_state;
+		CalculateVMRowAndSwath_params->use_one_row_for_frame_flip = mode_lib->ms.use_one_row_for_frame_flip_this_state;
+		CalculateVMRowAndSwath_params->UsesMALLForStaticScreen = s->dummy_boolean_array[0];
+		CalculateVMRowAndSwath_params->PTE_BUFFER_MODE = s->dummy_boolean_array[1];
+		CalculateVMRowAndSwath_params->BIGK_FRAGMENT_SIZE = s->dummy_integer_array[21];
+
+		CalculateVMRowAndSwath(&mode_lib->scratch,
+			CalculateVMRowAndSwath_params);
+
+		for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+			mode_lib->ms.PrefetchLinesY[j][k] = mode_lib->ms.PrefetchLinesYThisState[k];
+			mode_lib->ms.PrefetchLinesC[j][k] = mode_lib->ms.PrefetchLinesCThisState[k];
+			mode_lib->ms.meta_row_bandwidth[j][k] = mode_lib->ms.meta_row_bandwidth_this_state[k];
+			mode_lib->ms.dpte_row_bandwidth[j][k] = mode_lib->ms.dpte_row_bandwidth_this_state[k];
+			mode_lib->ms.DPTEBytesPerRow[j][k] = mode_lib->ms.DPTEBytesPerRowThisState[k];
+			mode_lib->ms.PDEAndMetaPTEBytesPerFrame[j][k] = mode_lib->ms.PDEAndMetaPTEBytesPerFrameThisState[k];
+			mode_lib->ms.MetaRowBytes[j][k] = mode_lib->ms.MetaRowBytesThisState[k];
+			mode_lib->ms.use_one_row_for_frame[j][k] = mode_lib->ms.use_one_row_for_frame_this_state[k];
+			mode_lib->ms.use_one_row_for_frame_flip[j][k] = mode_lib->ms.use_one_row_for_frame_flip_this_state[k];
+		}
+
+		mode_lib->ms.support.PTEBufferSizeNotExceeded[j] = true;
+
+		for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+			if (mode_lib->ms.PTEBufferSizeNotExceededPerState[k] == false)
+				mode_lib->ms.support.PTEBufferSizeNotExceeded[j] = false;
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: j=%u k=%u, PTEBufferSizeNotExceededPerState[%u] = %u\n",  __func__, j, k, k, mode_lib->ms.PTEBufferSizeNotExceededPerState[k]);
+#endif
+		}
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: PTEBufferSizeNotExceeded[%u] = %u\n",  __func__, j, mode_lib->ms.support.PTEBufferSizeNotExceeded[j]);
+#endif
+
+		mode_lib->ms.support.DCCMetaBufferSizeNotExceeded[j] = true;
+		for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+			if (mode_lib->ms.DCCMetaBufferSizeNotExceededPerState[k] == false)
+				mode_lib->ms.support.DCCMetaBufferSizeNotExceeded[j] = false;
+		}
+
+		mode_lib->ms.UrgLatency = CalculateUrgentLatency(mode_lib->ms.state.urgent_latency_pixel_data_only_us,
+													mode_lib->ms.state.urgent_latency_pixel_mixed_with_vm_data_us,
+													mode_lib->ms.state.urgent_latency_vm_data_only_us,
+													mode_lib->ms.soc.do_urgent_latency_adjustment,
+													mode_lib->ms.state.urgent_latency_adjustment_fabric_clock_component_us,
+													mode_lib->ms.state.urgent_latency_adjustment_fabric_clock_reference_mhz,
+													mode_lib->ms.state.fabricclk_mhz);
+
+		/* Getter functions work at mp interface so copy the urgent latency to mp*/
+		mode_lib->mp.UrgentLatency = mode_lib->ms.UrgLatency;
+
+		for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+			CalculateUrgentBurstFactor(
+				mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange[k],
+				mode_lib->ms.swath_width_luma_ub_this_state[k],
+				mode_lib->ms.swath_width_chroma_ub_this_state[k],
+				mode_lib->ms.SwathHeightYThisState[k],
+				mode_lib->ms.SwathHeightCThisState[k],
+				(dml_float_t) mode_lib->ms.cache_display_cfg.timing.HTotal[k] / mode_lib->ms.cache_display_cfg.timing.PixelClock[k],
+				mode_lib->ms.UrgLatency,
+				mode_lib->ms.ip.cursor_buffer_size,
+				mode_lib->ms.cache_display_cfg.plane.CursorWidth[k],
+				mode_lib->ms.cache_display_cfg.plane.CursorBPP[k],
+				mode_lib->ms.cache_display_cfg.plane.VRatio[k],
+				mode_lib->ms.cache_display_cfg.plane.VRatioChroma[k],
+				mode_lib->ms.BytePerPixelInDETY[k],
+				mode_lib->ms.BytePerPixelInDETC[k],
+				mode_lib->ms.DETBufferSizeYThisState[k],
+				mode_lib->ms.DETBufferSizeCThisState[k],
+				/* Output */
+				&mode_lib->ms.UrgentBurstFactorCursor[j][k],
+				&mode_lib->ms.UrgentBurstFactorLuma[j][k],
+				&mode_lib->ms.UrgentBurstFactorChroma[j][k],
+				&mode_lib->ms.NotUrgentLatencyHiding[k]);
+		}
+
+		CalculateDCFCLKDeepSleep(
+				mode_lib->ms.num_active_planes,
+				mode_lib->ms.BytePerPixelY,
+				mode_lib->ms.BytePerPixelC,
+				mode_lib->ms.cache_display_cfg.plane.VRatio,
+				mode_lib->ms.cache_display_cfg.plane.VRatioChroma,
+				mode_lib->ms.SwathWidthYThisState,
+				mode_lib->ms.SwathWidthCThisState,
+				mode_lib->ms.NoOfDPPThisState,
+				mode_lib->ms.cache_display_cfg.plane.HRatio,
+				mode_lib->ms.cache_display_cfg.plane.HRatioChroma,
+				mode_lib->ms.cache_display_cfg.timing.PixelClock,
+				mode_lib->ms.PSCL_FACTOR,
+				mode_lib->ms.PSCL_FACTOR_CHROMA,
+				mode_lib->ms.RequiredDPPCLKThisState,
+				mode_lib->ms.ReadBandwidthLuma,
+				mode_lib->ms.ReadBandwidthChroma,
+				mode_lib->ms.soc.return_bus_width_bytes,
+
+				/* Output */
+				&mode_lib->ms.ProjectedDCFCLKDeepSleep[j]);
+	}
+
+	//Calculate Return BW
+	for (j = 0; j < 2; ++j) {
+		for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+			if (mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming[k] == k) {
+				if (mode_lib->ms.cache_display_cfg.writeback.WritebackEnable[k] == true) {
+					mode_lib->ms.WritebackDelayTime[k] = mode_lib->ms.state.writeback_latency_us + CalculateWriteBackDelay(
+									mode_lib->ms.cache_display_cfg.writeback.WritebackPixelFormat[k],
+									mode_lib->ms.cache_display_cfg.writeback.WritebackHRatio[k],
+									mode_lib->ms.cache_display_cfg.writeback.WritebackVRatio[k],
+									mode_lib->ms.cache_display_cfg.writeback.WritebackVTaps[k],
+									mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationWidth[k],
+									mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationHeight[k],
+									mode_lib->ms.cache_display_cfg.writeback.WritebackSourceHeight[k],
+									mode_lib->ms.cache_display_cfg.timing.HTotal[k]) / mode_lib->ms.RequiredDISPCLK[j];
+				} else {
+					mode_lib->ms.WritebackDelayTime[k] = 0.0;
+				}
+				for (m = 0; m <= mode_lib->ms.num_active_planes - 1; m++) {
+					if (mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming[m] == k && mode_lib->ms.cache_display_cfg.writeback.WritebackEnable[m] == true) {
+						mode_lib->ms.WritebackDelayTime[k] = dml_max(mode_lib->ms.WritebackDelayTime[k],
+											mode_lib->ms.state.writeback_latency_us + CalculateWriteBackDelay(
+											mode_lib->ms.cache_display_cfg.writeback.WritebackPixelFormat[m],
+											mode_lib->ms.cache_display_cfg.writeback.WritebackHRatio[m],
+											mode_lib->ms.cache_display_cfg.writeback.WritebackVRatio[m],
+											mode_lib->ms.cache_display_cfg.writeback.WritebackVTaps[m],
+											mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationWidth[m],
+											mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationHeight[m],
+											mode_lib->ms.cache_display_cfg.writeback.WritebackSourceHeight[m],
+											mode_lib->ms.cache_display_cfg.timing.HTotal[m]) / mode_lib->ms.RequiredDISPCLK[j]);
+					}
+				}
+			}
+		}
+		for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+			for (m = 0; m <= mode_lib->ms.num_active_planes - 1; m++) {
+				if (mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming[k] == m) {
+					mode_lib->ms.WritebackDelayTime[k] = mode_lib->ms.WritebackDelayTime[m];
+				}
+			}
+		}
+       s->MaxVStartupAllPlanes[j] = 0;  // max vstartup among all planes
+
+		for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+			s->MaximumVStartup[j][k] = CalculateMaxVStartup(k,
+														mode_lib->ms.ip.ptoi_supported,
+														mode_lib->ms.ip.vblank_nom_default_us,
+														&mode_lib->ms.cache_display_cfg.timing,
+														mode_lib->ms.WritebackDelayTime[k]);
+
+			s->MaxVStartupAllPlanes[j] = (dml_uint_t)(dml_max(s->MaxVStartupAllPlanes[j], s->MaximumVStartup[j][k]));
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: k=%u, MaxVStartupAllPlanes[%u] = %u\n", __func__, k, j, s->MaxVStartupAllPlanes[j]);
+			dml_print("DML::%s: k=%u, MaximumVStartup[%u][%u] = %u\n", __func__, k, j, k, s->MaximumVStartup[j][k]);
+#endif
+		}
+	}
+
+	s->ReorderingBytes = (dml_uint_t)(mode_lib->ms.soc.num_chans * dml_max3(mode_lib->ms.soc.urgent_out_of_order_return_per_channel_pixel_only_bytes,
+																mode_lib->ms.soc.urgent_out_of_order_return_per_channel_pixel_and_vm_bytes,
+																mode_lib->ms.soc.urgent_out_of_order_return_per_channel_vm_only_bytes));
+
+	for (j = 0; j < 2; ++j) {
+		mode_lib->ms.DCFCLKState[j] = mode_lib->ms.state.dcfclk_mhz;
+	}
+
+	/* Immediate Flip and MALL parameters */
+	s->ImmediateFlipRequiredFinal = false;
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		s->ImmediateFlipRequiredFinal = s->ImmediateFlipRequiredFinal || (mode_lib->ms.policy.ImmediateFlipRequirement[k] == dml_immediate_flip_required);
+	}
+
+	mode_lib->ms.support.ImmediateFlipRequiredButTheRequirementForEachSurfaceIsNotSpecified = false;
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		mode_lib->ms.support.ImmediateFlipRequiredButTheRequirementForEachSurfaceIsNotSpecified = mode_lib->ms.support.ImmediateFlipRequiredButTheRequirementForEachSurfaceIsNotSpecified ||
+																							((mode_lib->ms.policy.ImmediateFlipRequirement[k] != dml_immediate_flip_required) &&
+																							(mode_lib->ms.policy.ImmediateFlipRequirement[k] != dml_immediate_flip_not_required));
+	}
+	mode_lib->ms.support.ImmediateFlipRequiredButTheRequirementForEachSurfaceIsNotSpecified = mode_lib->ms.support.ImmediateFlipRequiredButTheRequirementForEachSurfaceIsNotSpecified && s->ImmediateFlipRequiredFinal;
+
+	mode_lib->ms.support.ImmediateFlipOrHostVMAndPStateWithMALLFullFrameOrPhantomPipe = false;
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		mode_lib->ms.support.ImmediateFlipOrHostVMAndPStateWithMALLFullFrameOrPhantomPipe =
+										mode_lib->ms.support.ImmediateFlipOrHostVMAndPStateWithMALLFullFrameOrPhantomPipe || ((mode_lib->ms.cache_display_cfg.plane.HostVMEnable == true || mode_lib->ms.policy.ImmediateFlipRequirement[k] != dml_immediate_flip_not_required) &&
+										(mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange[k] == dml_use_mall_pstate_change_full_frame || mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange[k] == dml_use_mall_pstate_change_phantom_pipe));
+	}
+
+	mode_lib->ms.support.InvalidCombinationOfMALLUseForPStateAndStaticScreen = false;
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		mode_lib->ms.support.InvalidCombinationOfMALLUseForPStateAndStaticScreen = mode_lib->ms.support.InvalidCombinationOfMALLUseForPStateAndStaticScreen ||
+																((mode_lib->ms.cache_display_cfg.plane.UseMALLForStaticScreen[k] == dml_use_mall_static_screen_enable || mode_lib->ms.cache_display_cfg.plane.UseMALLForStaticScreen[k] == dml_use_mall_static_screen_optimize) && (mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange[k] == dml_use_mall_pstate_change_phantom_pipe)) ||
+																((mode_lib->ms.cache_display_cfg.plane.UseMALLForStaticScreen[k] == dml_use_mall_static_screen_disable || mode_lib->ms.cache_display_cfg.plane.UseMALLForStaticScreen[k] == dml_use_mall_static_screen_optimize) && (mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange[k] == dml_use_mall_pstate_change_full_frame));
+	}
+
+	s->FullFrameMALLPStateMethod = false;
+	s->SubViewportMALLPStateMethod = false;
+	s->PhantomPipeMALLPStateMethod = false;
+	s->SubViewportMALLRefreshGreaterThan120Hz = false;
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		if (mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange[k] == dml_use_mall_pstate_change_full_frame)
+			s->FullFrameMALLPStateMethod = true;
+		if (mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange[k] == dml_use_mall_pstate_change_sub_viewport) {
+			s->SubViewportMALLPStateMethod = true;
+			if (mode_lib->ms.cache_display_cfg.timing.RefreshRate[k] > 120)
+				s->SubViewportMALLRefreshGreaterThan120Hz = true;
+		}
+		if (mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange[k] == dml_use_mall_pstate_change_phantom_pipe)
+			s->PhantomPipeMALLPStateMethod = true;
+	}
+	mode_lib->ms.support.InvalidCombinationOfMALLUseForPState = (s->SubViewportMALLPStateMethod != s->PhantomPipeMALLPStateMethod)
+	|| (s->SubViewportMALLPStateMethod && s->FullFrameMALLPStateMethod) || s->SubViewportMALLRefreshGreaterThan120Hz;
+
+    if (mode_lib->ms.policy.UseMinimumRequiredDCFCLK == true) {
+		UseMinimumDCFCLK_params->UseMALLForPStateChange = mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange;
+		UseMinimumDCFCLK_params->DRRDisplay = mode_lib->ms.cache_display_cfg.timing.DRRDisplay;
+		UseMinimumDCFCLK_params->SynchronizeDRRDisplaysForUCLKPStateChangeFinal = mode_lib->ms.policy.SynchronizeDRRDisplaysForUCLKPStateChangeFinal;
+		UseMinimumDCFCLK_params->MaxInterDCNTileRepeaters = mode_lib->ms.ip.max_inter_dcn_tile_repeaters;
+		UseMinimumDCFCLK_params->MaxPrefetchMode = dml_prefetch_support_stutter;
+		UseMinimumDCFCLK_params->DRAMClockChangeLatencyFinal = mode_lib->ms.state.dram_clock_change_latency_us;
+		UseMinimumDCFCLK_params->FCLKChangeLatency = mode_lib->ms.state.fclk_change_latency_us;
+		UseMinimumDCFCLK_params->SREnterPlusExitTime = mode_lib->ms.state.sr_enter_plus_exit_time_us;
+		UseMinimumDCFCLK_params->ReturnBusWidth = mode_lib->ms.soc.return_bus_width_bytes;
+		UseMinimumDCFCLK_params->RoundTripPingLatencyCycles = mode_lib->ms.soc.round_trip_ping_latency_dcfclk_cycles;
+		UseMinimumDCFCLK_params->ReorderingBytes = s->ReorderingBytes;
+		UseMinimumDCFCLK_params->PixelChunkSizeInKByte = mode_lib->ms.ip.pixel_chunk_size_kbytes;
+		UseMinimumDCFCLK_params->MetaChunkSize = mode_lib->ms.ip.meta_chunk_size_kbytes;
+		UseMinimumDCFCLK_params->GPUVMEnable = mode_lib->ms.cache_display_cfg.plane.GPUVMEnable;
+		UseMinimumDCFCLK_params->GPUVMMaxPageTableLevels = mode_lib->ms.cache_display_cfg.plane.GPUVMMaxPageTableLevels;
+		UseMinimumDCFCLK_params->HostVMEnable = mode_lib->ms.cache_display_cfg.plane.HostVMEnable;
+		UseMinimumDCFCLK_params->NumberOfActiveSurfaces = mode_lib->ms.num_active_planes;
+		UseMinimumDCFCLK_params->HostVMMinPageSize = mode_lib->ms.soc.hostvm_min_page_size_kbytes * 1024;
+		UseMinimumDCFCLK_params->HostVMMaxNonCachedPageTableLevels = mode_lib->ms.cache_display_cfg.plane.HostVMMaxPageTableLevels;
+		UseMinimumDCFCLK_params->DynamicMetadataVMEnabled = mode_lib->ms.ip.dynamic_metadata_vm_enabled;
+		UseMinimumDCFCLK_params->ImmediateFlipRequirement = s->ImmediateFlipRequiredFinal;
+		UseMinimumDCFCLK_params->ProgressiveToInterlaceUnitInOPP = mode_lib->ms.ip.ptoi_supported;
+		UseMinimumDCFCLK_params->MaxAveragePercentOfIdealSDPPortBWDisplayCanUseInNormalSystemOperation = mode_lib->ms.soc.max_avg_sdp_bw_use_normal_percent;
+		UseMinimumDCFCLK_params->PercentOfIdealSDPPortBWReceivedAfterUrgLatency = mode_lib->ms.soc.pct_ideal_sdp_bw_after_urgent;
+		UseMinimumDCFCLK_params->VTotal = mode_lib->ms.cache_display_cfg.timing.VTotal;
+		UseMinimumDCFCLK_params->VActive = mode_lib->ms.cache_display_cfg.timing.VActive;
+		UseMinimumDCFCLK_params->DynamicMetadataTransmittedBytes = mode_lib->ms.cache_display_cfg.plane.DynamicMetadataTransmittedBytes;
+		UseMinimumDCFCLK_params->DynamicMetadataLinesBeforeActiveRequired = mode_lib->ms.cache_display_cfg.plane.DynamicMetadataLinesBeforeActiveRequired;
+		UseMinimumDCFCLK_params->Interlace = mode_lib->ms.cache_display_cfg.timing.Interlace;
+		UseMinimumDCFCLK_params->RequiredDPPCLKPerSurface = mode_lib->ms.RequiredDPPCLKPerSurface;
+		UseMinimumDCFCLK_params->RequiredDISPCLK = mode_lib->ms.RequiredDISPCLK;
+		UseMinimumDCFCLK_params->UrgLatency = mode_lib->ms.UrgLatency;
+		UseMinimumDCFCLK_params->NoOfDPP = mode_lib->ms.NoOfDPP;
+		UseMinimumDCFCLK_params->ProjectedDCFCLKDeepSleep = mode_lib->ms.ProjectedDCFCLKDeepSleep;
+		UseMinimumDCFCLK_params->MaximumVStartup = s->MaximumVStartup;
+		UseMinimumDCFCLK_params->TotalNumberOfActiveDPP = mode_lib->ms.TotalNumberOfActiveDPP;
+		UseMinimumDCFCLK_params->TotalNumberOfDCCActiveDPP = mode_lib->ms.TotalNumberOfDCCActiveDPP;
+		UseMinimumDCFCLK_params->dpte_group_bytes = mode_lib->ms.dpte_group_bytes;
+		UseMinimumDCFCLK_params->PrefetchLinesY = mode_lib->ms.PrefetchLinesY;
+		UseMinimumDCFCLK_params->PrefetchLinesC = mode_lib->ms.PrefetchLinesC;
+		UseMinimumDCFCLK_params->swath_width_luma_ub_all_states = mode_lib->ms.swath_width_luma_ub_all_states;
+		UseMinimumDCFCLK_params->swath_width_chroma_ub_all_states = mode_lib->ms.swath_width_chroma_ub_all_states;
+		UseMinimumDCFCLK_params->BytePerPixelY = mode_lib->ms.BytePerPixelY;
+		UseMinimumDCFCLK_params->BytePerPixelC = mode_lib->ms.BytePerPixelC;
+		UseMinimumDCFCLK_params->HTotal = mode_lib->ms.cache_display_cfg.timing.HTotal;
+		UseMinimumDCFCLK_params->PixelClock = mode_lib->ms.cache_display_cfg.timing.PixelClock;
+		UseMinimumDCFCLK_params->PDEAndMetaPTEBytesPerFrame = mode_lib->ms.PDEAndMetaPTEBytesPerFrame;
+		UseMinimumDCFCLK_params->DPTEBytesPerRow = mode_lib->ms.DPTEBytesPerRow;
+		UseMinimumDCFCLK_params->MetaRowBytes = mode_lib->ms.MetaRowBytes;
+		UseMinimumDCFCLK_params->DynamicMetadataEnable = mode_lib->ms.cache_display_cfg.plane.DynamicMetadataEnable;
+		UseMinimumDCFCLK_params->ReadBandwidthLuma = mode_lib->ms.ReadBandwidthLuma;
+		UseMinimumDCFCLK_params->ReadBandwidthChroma = mode_lib->ms.ReadBandwidthChroma;
+		UseMinimumDCFCLK_params->DCFCLKPerState = mode_lib->ms.state.dcfclk_mhz;
+		UseMinimumDCFCLK_params->DCFCLKState = mode_lib->ms.DCFCLKState;
+
+		UseMinimumDCFCLK(&mode_lib->scratch,
+		UseMinimumDCFCLK_params);
+
+	 } // UseMinimumRequiredDCFCLK == true
+
+	for (j = 0; j < 2; ++j) {
+		mode_lib->ms.ReturnBWPerState[j] = dml_get_return_bw_mbps(&mode_lib->ms.soc, mode_lib->ms.state.use_ideal_dram_bw_strobe,
+																mode_lib->ms.cache_display_cfg.plane.HostVMEnable, mode_lib->ms.DCFCLKState[j], mode_lib->ms.state.fabricclk_mhz,
+																mode_lib->ms.state.dram_speed_mts);
+		mode_lib->ms.ReturnDRAMBWPerState[j] = dml_get_return_dram_bw_mbps(&mode_lib->ms.soc, mode_lib->ms.state.use_ideal_dram_bw_strobe,
+																mode_lib->ms.cache_display_cfg.plane.HostVMEnable,
+																mode_lib->ms.state.dram_speed_mts);
+	}
+
+	//Re-ordering Buffer Support Check
+	for (j = 0; j < 2; ++j) {
+		if ((mode_lib->ms.ip.rob_buffer_size_kbytes - mode_lib->ms.ip.pixel_chunk_size_kbytes) * 1024 / mode_lib->ms.ReturnBWPerState[j] >
+			(mode_lib->ms.soc.round_trip_ping_latency_dcfclk_cycles + 32) / mode_lib->ms.DCFCLKState[j] + s->ReorderingBytes / mode_lib->ms.ReturnBWPerState[j]) {
+			mode_lib->ms.support.ROBSupport[j] = true;
+		} else {
+			mode_lib->ms.support.ROBSupport[j] = false;
+		}
+		dml_print("DML::%s: DEBUG ROBSupport[%u] = %u (%u)\n",  __func__, j, mode_lib->ms.support.ROBSupport[j], __LINE__);
+	}
+
+	//Vertical Active BW support check
+	s->MaxTotalVActiveRDBandwidth = 0;
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		s->MaxTotalVActiveRDBandwidth = s->MaxTotalVActiveRDBandwidth + mode_lib->ms.ReadBandwidthLuma[k] + mode_lib->ms.ReadBandwidthChroma[k];
+	}
+
+	for (j = 0; j < 2; ++j) {
+		mode_lib->ms.support.MaxTotalVerticalActiveAvailableBandwidth[j] = dml_min3(mode_lib->ms.soc.return_bus_width_bytes * mode_lib->ms.DCFCLKState[j] * mode_lib->ms.soc.max_avg_sdp_bw_use_normal_percent / 100.0,
+																	mode_lib->ms.state.fabricclk_mhz * mode_lib->ms.soc.fabric_datapath_to_dcn_data_return_bytes * mode_lib->ms.soc.max_avg_fabric_bw_use_normal_percent / 100.0,
+																	mode_lib->ms.state.dram_speed_mts * mode_lib->ms.soc.num_chans * mode_lib->ms.soc.dram_channel_width_bytes *
+																	((mode_lib->ms.state.use_ideal_dram_bw_strobe && !mode_lib->ms.cache_display_cfg.plane.HostVMEnable) ?
+																	mode_lib->ms.soc.max_avg_dram_bw_use_normal_strobe_percent : mode_lib->ms.soc.max_avg_dram_bw_use_normal_percent) / 100.0);
+
+		if (s->MaxTotalVActiveRDBandwidth <= mode_lib->ms.support.MaxTotalVerticalActiveAvailableBandwidth[j]) {
+			mode_lib->ms.support.TotalVerticalActiveBandwidthSupport[j] = true;
+		} else {
+			mode_lib->ms.support.TotalVerticalActiveBandwidthSupport[j] = false;
+		}
+	}
+
+	/* Prefetch Check */
+	dml_prefetch_check(mode_lib);
+
+	// End of Prefetch Check
+	dml_print("DML::%s: Done prefetch calculation\n", __func__);
+
+	/*Cursor Support Check*/
+	mode_lib->ms.support.CursorSupport = true;
+	for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+		if (mode_lib->ms.cache_display_cfg.plane.CursorWidth[k] > 0.0) {
+			if (mode_lib->ms.cache_display_cfg.plane.CursorBPP[k] == 64 && mode_lib->ms.ip.cursor_64bpp_support == false) {
+				mode_lib->ms.support.CursorSupport = false;
+			}
+		}
+	}
+
+	/*Valid Pitch Check*/
+	mode_lib->ms.support.PitchSupport = true;
+	for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+		mode_lib->ms.support.AlignedYPitch[k] = dml_ceil(
+				dml_max(mode_lib->ms.cache_display_cfg.surface.PitchY[k], mode_lib->ms.cache_display_cfg.surface.SurfaceWidthY[k]),
+				mode_lib->ms.MacroTileWidthY[k]);
+		if (mode_lib->ms.cache_display_cfg.surface.DCCEnable[k] == true) {
+			mode_lib->ms.support.AlignedDCCMetaPitchY[k] = dml_ceil(dml_max(mode_lib->ms.cache_display_cfg.surface.DCCMetaPitchY[k], mode_lib->ms.cache_display_cfg.surface.SurfaceWidthY[k]), 64.0 * mode_lib->ms.Read256BlockWidthY[k]);
+		} else {
+			mode_lib->ms.support.AlignedDCCMetaPitchY[k] = mode_lib->ms.cache_display_cfg.surface.DCCMetaPitchY[k];
+		}
+		if (mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_444_64
+			&& mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_444_32
+			&& mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_444_16
+			&& mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_mono_16
+			&& mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_rgbe
+			&& mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_mono_8) {
+			mode_lib->ms.support.AlignedCPitch[k] = dml_ceil(dml_max(mode_lib->ms.cache_display_cfg.surface.PitchC[k], mode_lib->ms.cache_display_cfg.surface.SurfaceWidthC[k]), mode_lib->ms.MacroTileWidthC[k]);
+			if (mode_lib->ms.cache_display_cfg.surface.DCCEnable[k] == true) {
+				mode_lib->ms.support.AlignedDCCMetaPitchC[k] = dml_ceil(dml_max(mode_lib->ms.cache_display_cfg.surface.DCCMetaPitchC[k], mode_lib->ms.cache_display_cfg.surface.SurfaceWidthC[k]), 64.0 * mode_lib->ms.Read256BlockWidthC[k]);
+			} else {
+				mode_lib->ms.support.AlignedDCCMetaPitchC[k] = mode_lib->ms.cache_display_cfg.surface.DCCMetaPitchC[k];
+			}
+		} else {
+			mode_lib->ms.support.AlignedCPitch[k] = mode_lib->ms.cache_display_cfg.surface.PitchC[k];
+			mode_lib->ms.support.AlignedDCCMetaPitchC[k] = mode_lib->ms.cache_display_cfg.surface.DCCMetaPitchC[k];
+		}
+		if (mode_lib->ms.support.AlignedYPitch[k] > mode_lib->ms.cache_display_cfg.surface.PitchY[k] || mode_lib->ms.support.AlignedCPitch[k] > mode_lib->ms.cache_display_cfg.surface.PitchC[k] ||
+			mode_lib->ms.support.AlignedDCCMetaPitchY[k] > mode_lib->ms.cache_display_cfg.surface.DCCMetaPitchY[k] || mode_lib->ms.support.AlignedDCCMetaPitchC[k] > mode_lib->ms.cache_display_cfg.surface.DCCMetaPitchC[k]) {
+			mode_lib->ms.support.PitchSupport = false;
+		}
+	}
+
+	mode_lib->ms.support.ViewportExceedsSurface = false;
+	for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+		if (mode_lib->ms.cache_display_cfg.plane.ViewportWidth[k] > mode_lib->ms.cache_display_cfg.surface.SurfaceWidthY[k] || mode_lib->ms.cache_display_cfg.plane.ViewportHeight[k] > mode_lib->ms.cache_display_cfg.surface.SurfaceHeightY[k]) {
+			mode_lib->ms.support.ViewportExceedsSurface = true;
+			if (mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_444_64 && mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_444_32 &&
+				mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_444_16 && mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_444_8 && mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k] != dml_rgbe) {
+				if (mode_lib->ms.cache_display_cfg.plane.ViewportWidthChroma[k] > mode_lib->ms.cache_display_cfg.surface.SurfaceWidthC[k] || mode_lib->ms.cache_display_cfg.plane.ViewportHeightChroma[k] > mode_lib->ms.cache_display_cfg.surface.SurfaceHeightC[k]) {
+					mode_lib->ms.support.ViewportExceedsSurface = true;
+				}
+			}
+		}
+	}
+
+	/*Mode Support, Voltage State and SOC Configuration*/
+	for (j = 0; j < 2; j++) { // j iterator is for the combine mode off or on
+		dml_print("DML::%s: checking support for j=%u\n", __func__, j);
+		dml_print("DML::%s: state_idx=%0d max_state_idx=%0d\n", __func__, mode_lib->ms.state_idx, mode_lib->ms.max_state_idx);
+
+		s->is_max_pwr_state = (mode_lib->ms.max_state_idx == mode_lib->ms.state_idx);
+		s->is_max_dram_pwr_state = (mode_lib->ms.max_state.dram_speed_mts == mode_lib->ms.state.dram_speed_mts);
+
+		s->dram_clock_change_support = (!mode_lib->ms.policy.DRAMClockChangeRequirementFinal ||
+											(s->is_max_dram_pwr_state && mode_lib->policy.AssumeModeSupportAtMaxPwrStateEvenDRAMClockChangeNotSupported) ||
+											mode_lib->ms.support.DRAMClockChangeSupport[j] != dml_dram_clock_change_unsupported);
+		s->f_clock_change_support = (!mode_lib->ms.policy.FCLKChangeRequirementFinal ||
+											(s->is_max_pwr_state && mode_lib->policy.AssumeModeSupportAtMaxPwrStateEvenFClockChangeNotSupported) ||
+											mode_lib->ms.support.FCLKChangeSupport[j] != dml_fclock_change_unsupported);
+
+		if (mode_lib->ms.support.ScaleRatioAndTapsSupport == true
+			&& mode_lib->ms.support.SourceFormatPixelAndScanSupport == true
+			&& mode_lib->ms.support.ViewportSizeSupport[j] == true
+			&& !mode_lib->ms.support.LinkRateDoesNotMatchDPVersion
+			&& !mode_lib->ms.support.LinkRateForMultistreamNotIndicated
+			&& !mode_lib->ms.support.BPPForMultistreamNotIndicated
+			&& !mode_lib->ms.support.MultistreamWithHDMIOreDP
+			&& !mode_lib->ms.support.ExceededMultistreamSlots
+			&& !mode_lib->ms.support.MSOOrODMSplitWithNonDPLink
+			&& !mode_lib->ms.support.NotEnoughLanesForMSO
+			&& mode_lib->ms.support.LinkCapacitySupport == true
+			&& !mode_lib->ms.support.P2IWith420
+			&& !mode_lib->ms.support.DSCOnlyIfNecessaryWithBPP
+			&& !mode_lib->ms.support.DSC422NativeNotSupported
+			&& !mode_lib->ms.support.MPCCombineMethodIncompatible
+			&& mode_lib->ms.support.ODMCombineTwoToOneSupportCheckOK == true
+			&& mode_lib->ms.support.ODMCombineFourToOneSupportCheckOK == true
+			&& mode_lib->ms.support.NotEnoughDSCUnits == false
+			&& !mode_lib->ms.support.NotEnoughDSCSlices
+			&& !mode_lib->ms.support.ImmediateFlipOrHostVMAndPStateWithMALLFullFrameOrPhantomPipe
+			&& !mode_lib->ms.support.InvalidCombinationOfMALLUseForPStateAndStaticScreen
+			&& mode_lib->ms.support.DSCCLKRequiredMoreThanSupported == false
+			&& mode_lib->ms.support.PixelsPerLinePerDSCUnitSupport
+			&& mode_lib->ms.support.DTBCLKRequiredMoreThanSupported == false
+			&& !mode_lib->ms.support.InvalidCombinationOfMALLUseForPState
+			&& !mode_lib->ms.support.ImmediateFlipRequiredButTheRequirementForEachSurfaceIsNotSpecified
+			&& mode_lib->ms.support.ROBSupport[j] == true
+			&& mode_lib->ms.support.DISPCLK_DPPCLK_Support[j] == true
+			&& mode_lib->ms.support.TotalAvailablePipesSupport[j] == true
+			&& mode_lib->ms.support.NumberOfOTGSupport == true
+			&& mode_lib->ms.support.NumberOfHDMIFRLSupport == true
+			&& mode_lib->ms.support.NumberOfDP2p0Support == true
+			&& mode_lib->ms.support.EnoughWritebackUnits == true
+			&& mode_lib->ms.support.WritebackLatencySupport == true
+			&& mode_lib->ms.support.WritebackScaleRatioAndTapsSupport == true
+			&& mode_lib->ms.support.CursorSupport == true
+			&& mode_lib->ms.support.PitchSupport == true
+			&& mode_lib->ms.support.ViewportExceedsSurface == false
+			&& mode_lib->ms.support.PrefetchSupported[j] == true
+			&& mode_lib->ms.support.VActiveBandwithSupport[j] == true
+			&& mode_lib->ms.support.DynamicMetadataSupported[j] == true
+			&& mode_lib->ms.support.TotalVerticalActiveBandwidthSupport[j] == true
+			&& mode_lib->ms.support.VRatioInPrefetchSupported[j] == true
+			&& mode_lib->ms.support.PTEBufferSizeNotExceeded[j] == true
+			&& mode_lib->ms.support.DCCMetaBufferSizeNotExceeded[j] == true
+			&& mode_lib->ms.support.NonsupportedDSCInputBPC == false
+			&& !mode_lib->ms.support.ExceededMALLSize
+			&& ((mode_lib->ms.cache_display_cfg.plane.HostVMEnable == false && !s->ImmediateFlipRequiredFinal) || mode_lib->ms.support.ImmediateFlipSupportedForState[j])
+			&& s->dram_clock_change_support == true
+			&& s->f_clock_change_support == true
+			&& (!mode_lib->ms.policy.USRRetrainingRequiredFinal || mode_lib->ms.support.USRRetrainingSupport[j])) {
+			dml_print("DML::%s: mode is supported\n", __func__);
+			mode_lib->ms.support.ModeSupport[j] = true;
+		} else {
+			dml_print("DML::%s: mode is NOT supported\n", __func__);
+			mode_lib->ms.support.ModeSupport[j] = false;
+			dml_print_mode_support(mode_lib, j);
+		}
+	}
+
+	mode_lib->ms.support.MaximumMPCCombine = 0;
+	mode_lib->ms.support.ModeIsSupported = 0;
+	if (mode_lib->ms.support.ModeSupport[0] == true || mode_lib->ms.support.ModeSupport[1] == true) {  // if the mode is supported by either no combine or mpccombine
+		mode_lib->ms.support.ModeIsSupported = mode_lib->ms.support.ModeSupport[0] == true || mode_lib->ms.support.ModeSupport[1] == true;
+
+		// Determine if MPC combine is necessary, depends on if using MPC combine will help dram clock change or fclk change, etc.
+		if ((mode_lib->ms.support.ModeSupport[0] == false && mode_lib->ms.support.ModeSupport[1] == true) || s->MPCCombineMethodAsPossible ||
+			(s->MPCCombineMethodAsNeededForPStateChangeAndVoltage && mode_lib->ms.policy.DRAMClockChangeRequirementFinal &&
+			(((mode_lib->ms.support.DRAMClockChangeSupport[1] == dml_dram_clock_change_vactive || mode_lib->ms.support.DRAMClockChangeSupport[1] == dml_dram_clock_change_vactive_w_mall_full_frame || mode_lib->ms.support.DRAMClockChangeSupport[1] == dml_dram_clock_change_vactive_w_mall_sub_vp) &&
+			!(mode_lib->ms.support.DRAMClockChangeSupport[0] == dml_dram_clock_change_vactive || mode_lib->ms.support.DRAMClockChangeSupport[0] == dml_dram_clock_change_vactive_w_mall_full_frame || mode_lib->ms.support.DRAMClockChangeSupport[0] == dml_dram_clock_change_vactive_w_mall_sub_vp)) ||
+			((mode_lib->ms.support.DRAMClockChangeSupport[1] == dml_dram_clock_change_vblank || mode_lib->ms.support.DRAMClockChangeSupport[1] == dml_dram_clock_change_vblank_drr
+		|| mode_lib->ms.support.DRAMClockChangeSupport[1] == dml_dram_clock_change_vblank_w_mall_full_frame || mode_lib->ms.support.DRAMClockChangeSupport[1] == dml_dram_clock_change_vblank_drr_w_mall_full_frame
+		|| mode_lib->ms.support.DRAMClockChangeSupport[1] == dml_dram_clock_change_vblank_w_mall_sub_vp || mode_lib->ms.support.DRAMClockChangeSupport[1] == dml_dram_clock_change_vblank_drr_w_mall_sub_vp
+		) &&
+				mode_lib->ms.support.DRAMClockChangeSupport[0] == dml_dram_clock_change_unsupported)))
+			|| (s->MPCCombineMethodAsNeededForPStateChangeAndVoltage && mode_lib->ms.policy.FCLKChangeRequirementFinal &&
+				((mode_lib->ms.support.FCLKChangeSupport[1] == dml_fclock_change_vactive && mode_lib->ms.support.FCLKChangeSupport[0] != dml_fclock_change_vactive) ||
+				(mode_lib->ms.support.FCLKChangeSupport[1] == dml_fclock_change_vblank && mode_lib->ms.support.FCLKChangeSupport[0] == dml_fclock_change_unsupported)))) {
+			mode_lib->ms.support.MaximumMPCCombine = 1;
+		} else {
+			mode_lib->ms.support.MaximumMPCCombine = 0;
+		}
+	}
+
+	// Since now the mode_support work on 1 particular power state, so there is only 1 state idx (index 0).
+	mode_lib->ms.support.ImmediateFlipSupport          = mode_lib->ms.support.ImmediateFlipSupportedForState[mode_lib->ms.support.MaximumMPCCombine];   // Consider flip support if max combine support imm flip
+	mode_lib->ms.support.UnboundedRequestEnabled       = mode_lib->ms.UnboundedRequestEnabledAllStates[mode_lib->ms.support.MaximumMPCCombine];         // Not used, informational
+	mode_lib->ms.support.CompressedBufferSizeInkByte   = mode_lib->ms.CompressedBufferSizeInkByteAllStates[mode_lib->ms.support.MaximumMPCCombine];     // Not used, informational
+
+	dml_print("DML::%s: ModeIsSupported                = %u\n", __func__, mode_lib->ms.support.ModeIsSupported);
+	dml_print("DML::%s: MaximumMPCCombine              = %u\n", __func__, mode_lib->ms.support.MaximumMPCCombine);
+	dml_print("DML::%s: ImmediateFlipSupport           = %u\n", __func__, mode_lib->ms.support.ImmediateFlipSupport);
+	dml_print("DML::%s: UnboundedRequestEnabled        = %u\n", __func__, mode_lib->ms.support.UnboundedRequestEnabled);
+	dml_print("DML::%s: CompressedBufferSizeInkByte    = %u\n", __func__, mode_lib->ms.support.CompressedBufferSizeInkByte);
+
+	for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+		mode_lib->ms.support.MPCCombineEnable[k]   = mode_lib->ms.MPCCombine[mode_lib->ms.support.MaximumMPCCombine][k];
+		mode_lib->ms.support.DPPPerSurface[k]      = mode_lib->ms.NoOfDPP[mode_lib->ms.support.MaximumMPCCombine][k];
+		mode_lib->ms.SwathHeightY[k]               = mode_lib->ms.SwathHeightYAllStates[mode_lib->ms.support.MaximumMPCCombine][k];
+		mode_lib->ms.SwathHeightC[k]               = mode_lib->ms.SwathHeightCAllStates[mode_lib->ms.support.MaximumMPCCombine][k];
+		mode_lib->ms.DETBufferSizeInKByte[k]       = mode_lib->ms.DETBufferSizeInKByteAllStates[mode_lib->ms.support.MaximumMPCCombine][k];
+		mode_lib->ms.DETBufferSizeY[k]             = mode_lib->ms.DETBufferSizeYAllStates[mode_lib->ms.support.MaximumMPCCombine][k];
+		mode_lib->ms.DETBufferSizeC[k]             = mode_lib->ms.DETBufferSizeCAllStates[mode_lib->ms.support.MaximumMPCCombine][k];
+	}
+
+	mode_lib->ms.DRAMSpeed     = mode_lib->ms.state.dram_speed_mts;
+	mode_lib->ms.FabricClock   = mode_lib->ms.state.fabricclk_mhz;
+	mode_lib->ms.SOCCLK        = mode_lib->ms.state.socclk_mhz;
+	mode_lib->ms.DCFCLK        = mode_lib->ms.DCFCLKState[mode_lib->ms.support.MaximumMPCCombine];
+	mode_lib->ms.ReturnBW      = mode_lib->ms.ReturnBWPerState[mode_lib->ms.support.MaximumMPCCombine];
+	mode_lib->ms.ReturnDRAMBW  = mode_lib->ms.ReturnDRAMBWPerState[mode_lib->ms.support.MaximumMPCCombine];
+
+	for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+		if (mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming[k] == k) {
+			mode_lib->ms.support.ODMMode[k] = mode_lib->ms.ODMModePerState[k];
+		} else {
+			mode_lib->ms.support.ODMMode[k] = dml_odm_mode_bypass;
+		}
+
+		mode_lib->ms.support.DSCEnabled[k] = mode_lib->ms.RequiresDSC[k];
+		mode_lib->ms.support.FECEnabled[k] = mode_lib->ms.RequiresFEC[k];
+		mode_lib->ms.support.OutputBpp[k] = mode_lib->ms.OutputBppPerState[k];
+		mode_lib->ms.support.OutputType[k] = mode_lib->ms.OutputTypePerState[k];
+		mode_lib->ms.support.OutputRate[k] = mode_lib->ms.OutputRatePerState[k];
+		mode_lib->ms.support.SubViewportLinesNeededInMALL[k] = mode_lib->ms.SubViewportLinesNeededInMALL[k];
+	}
+
+	return mode_lib->ms.support.ModeIsSupported;
+} // dml_core_mode_support
+
+/// @brief This function calculates some parameters thats are needed ahead of the mode programming function all
+void dml_core_mode_support_partial(struct display_mode_lib_st *mode_lib)
+{
+	CalculateMaxDETAndMinCompressedBufferSize(
+								mode_lib->ms.ip.config_return_buffer_size_in_kbytes,
+								mode_lib->ms.ip.config_return_buffer_segment_size_in_kbytes,
+								mode_lib->ms.ip.rob_buffer_size_kbytes,
+								mode_lib->ms.ip.max_num_dpp,
+								mode_lib->ms.policy.NomDETInKByteOverrideEnable,
+								mode_lib->ms.policy.NomDETInKByteOverrideValue,
+
+								/* Output */
+								&mode_lib->ms.MaxTotalDETInKByte,
+								&mode_lib->ms.NomDETInKByte,
+								&mode_lib->ms.MinCompressedBufferSizeInKByte);
+
+	PixelClockAdjustmentForProgressiveToInterlaceUnit(&mode_lib->ms.cache_display_cfg, mode_lib->ms.ip.ptoi_supported);
+
+	mode_lib->ms.ReturnBW = dml_get_return_bw_mbps(&mode_lib->ms.soc,
+													mode_lib->ms.state.use_ideal_dram_bw_strobe,
+													mode_lib->ms.cache_display_cfg.plane.HostVMEnable,
+													mode_lib->ms.DCFCLK,
+													mode_lib->ms.FabricClock,
+													mode_lib->ms.DRAMSpeed);
+	dml_print("DML::%s: ReturnBW = %f\n", __func__, mode_lib->ms.ReturnBW);
+
+} // dml_core_mode_support_partial
+
+/// @brief This is the mode programming function. It is assumed the display cfg is support at the given power state
+void dml_core_mode_programming(struct display_mode_lib_st *mode_lib, const struct dml_clk_cfg_st *clk_cfg)
+{
+	struct dml_core_mode_programming_locals_st *s = &mode_lib->scratch.dml_core_mode_programming_locals;
+	struct CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_params_st *CalculateWatermarks_params = &mode_lib->scratch.CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_params;
+	struct CalculateVMRowAndSwath_params_st *CalculateVMRowAndSwath_params = &mode_lib->scratch.CalculateVMRowAndSwath_params;
+	struct CalculateSwathAndDETConfiguration_params_st *CalculateSwathAndDETConfiguration_params = &mode_lib->scratch.CalculateSwathAndDETConfiguration_params;
+	struct CalculateStutterEfficiency_params_st *CalculateStutterEfficiency_params = &mode_lib->scratch.CalculateStutterEfficiency_params;
+	struct CalculatePrefetchSchedule_params_st *CalculatePrefetchSchedule_params = &mode_lib->scratch.CalculatePrefetchSchedule_params;
+
+	struct mode_program_st   *locals    = &mode_lib->mp;
+	struct DmlPipe *myPipe;
+	dml_uint_t j = 0, k = 0;
+	dml_float_t TWait;
+	dml_bool_t isInterlaceTiming;
+
+	mode_lib->ms.num_active_planes = dml_get_num_active_planes(&mode_lib->ms.cache_display_cfg);
+	mode_lib->mp.num_active_pipes = dml_get_num_active_pipes(&mode_lib->ms.cache_display_cfg);
+	dml_calc_pipe_plane_mapping(&mode_lib->ms.cache_display_cfg.hw, mode_lib->mp.pipe_plane);
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: --- START --- \n",  __func__);
+	dml_print("DML::%s: num_active_planes = %u\n", __func__, mode_lib->ms.num_active_planes);
+	dml_print("DML::%s: num_active_pipes = %u\n", __func__, mode_lib->mp.num_active_pipes);
+#endif
+
+	s->DSCFormatFactor          = 0;
+
+	// Unlike dppclk and dispclk which can be calculated in mode_programming
+	// DCFCLK is calculated in mode_support (which is the state bbox dcfclk or min dcfclk if min dcfclk option is used in mode support calculation)
+	if (clk_cfg->dcfclk_option != dml_use_override_freq)
+		locals->Dcfclk = mode_lib->ms.DCFCLK;
+	else
+		locals->Dcfclk = clk_cfg->dcfclk_mhz;
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print_dml_policy(&mode_lib->ms.policy);
+	dml_print_soc_state_bounding_box(&mode_lib->ms.state);
+	dml_print_soc_bounding_box(&mode_lib->ms.soc);
+	dml_print_clk_cfg(clk_cfg);
+
+	dml_print("DML::%s: ImmediateFlipSupport = %u\n", __func__, mode_lib->ms.support.ImmediateFlipSupport);
+	dml_print("DML::%s: Using DCFCLK = %f\n", __func__, locals->Dcfclk);
+	dml_print("DML::%s: Using SOCCLK = %f\n", __func__, mode_lib->ms.SOCCLK);
+#endif
+
+	locals->WritebackDISPCLK = 0.0;
+	locals->GlobalDPPCLK = 0.0;
+
+	// DISPCLK and DPPCLK Calculation
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		if (mode_lib->ms.cache_display_cfg.writeback.WritebackEnable[k]) {
+			locals->WritebackDISPCLK =
+					dml_max(
+							locals->WritebackDISPCLK,
+							CalculateWriteBackDISPCLK(
+									mode_lib->ms.cache_display_cfg.writeback.WritebackPixelFormat[k],
+									mode_lib->ms.cache_display_cfg.timing.PixelClock[k],
+									mode_lib->ms.cache_display_cfg.writeback.WritebackHRatio[k],
+									mode_lib->ms.cache_display_cfg.writeback.WritebackVRatio[k],
+									mode_lib->ms.cache_display_cfg.writeback.WritebackHTaps[k],
+									mode_lib->ms.cache_display_cfg.writeback.WritebackVTaps[k],
+									mode_lib->ms.cache_display_cfg.writeback.WritebackSourceWidth[k],
+									mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationWidth[k],
+									mode_lib->ms.cache_display_cfg.timing.HTotal[k],
+									mode_lib->ms.ip.writeback_line_buffer_buffer_size,
+									mode_lib->ms.soc.dispclk_dppclk_vco_speed_mhz));
+		}
+	}
+
+	locals->Dispclk_calculated = locals->WritebackDISPCLK;
+
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		if (mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming[k] == k) {
+				locals->Dispclk_calculated = dml_max(locals->Dispclk_calculated, CalculateRequiredDispclk(
+																					mode_lib->ms.cache_display_cfg.hw.ODMMode[k],
+																					mode_lib->ms.cache_display_cfg.timing.PixelClock[k],
+																					mode_lib->ms.soc.dcn_downspread_percent,
+																					mode_lib->ms.ip.dispclk_ramp_margin_percent,
+																					mode_lib->ms.soc.dispclk_dppclk_vco_speed_mhz,
+																					mode_lib->ms.max_state.dispclk_mhz));
+		}
+	}
+	if (clk_cfg->dispclk_option == dml_use_required_freq)
+		locals->Dispclk = locals->Dispclk_calculated;
+	else if (clk_cfg->dispclk_option == dml_use_override_freq)
+		locals->Dispclk = clk_cfg->dispclk_mhz;
+	else
+		locals->Dispclk = mode_lib->ms.state.dispclk_mhz;
+#ifdef __DML_VBA_DEBUG__
+	 dml_print("DML::%s: Using Dispclk = %f\n", __func__, locals->Dispclk);
+#endif
+
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		CalculateSinglePipeDPPCLKAndSCLThroughput(
+				mode_lib->ms.cache_display_cfg.plane.HRatio[k],
+				mode_lib->ms.cache_display_cfg.plane.HRatioChroma[k],
+				mode_lib->ms.cache_display_cfg.plane.VRatio[k],
+				mode_lib->ms.cache_display_cfg.plane.VRatioChroma[k],
+				mode_lib->ms.ip.max_dchub_pscl_bw_pix_per_clk,
+				mode_lib->ms.ip.max_pscl_lb_bw_pix_per_clk,
+				mode_lib->ms.cache_display_cfg.timing.PixelClock[k],
+				mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k],
+				mode_lib->ms.cache_display_cfg.plane.HTaps[k],
+				mode_lib->ms.cache_display_cfg.plane.HTapsChroma[k],
+				mode_lib->ms.cache_display_cfg.plane.VTaps[k],
+				mode_lib->ms.cache_display_cfg.plane.VTapsChroma[k],
+
+				/* Output */
+				&locals->PSCL_THROUGHPUT[k],
+				&locals->PSCL_THROUGHPUT_CHROMA[k],
+				&locals->DPPCLKUsingSingleDPP[k]);
+	}
+
+	CalculateDPPCLK(mode_lib->ms.num_active_planes,
+					mode_lib->ms.soc.dcn_downspread_percent,
+					mode_lib->ms.soc.dispclk_dppclk_vco_speed_mhz,
+					locals->DPPCLKUsingSingleDPP,
+					mode_lib->ms.cache_display_cfg.hw.DPPPerSurface,
+					/* Output */
+					&locals->GlobalDPPCLK,
+					locals->Dppclk_calculated);
+
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		if (clk_cfg->dppclk_option[k] == dml_use_required_freq)
+			locals->Dppclk[k] = locals->Dppclk_calculated[k];
+		else if (clk_cfg->dppclk_option[k] == dml_use_override_freq)
+			locals->Dppclk[k] = clk_cfg->dppclk_mhz[k];
+		else
+			locals->Dppclk[k] = mode_lib->ms.state.dppclk_mhz;
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: Using Dppclk[%0d] = %f\n", __func__, k, locals->Dppclk[k]);
+#endif
+	}
+
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		CalculateBytePerPixelAndBlockSizes(
+				mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k],
+				mode_lib->ms.cache_display_cfg.surface.SurfaceTiling[k],
+
+				/* Output */
+				&locals->BytePerPixelY[k],
+				&locals->BytePerPixelC[k],
+				&locals->BytePerPixelDETY[k],
+				&locals->BytePerPixelDETC[k],
+				&locals->BlockHeight256BytesY[k],
+				&locals->BlockHeight256BytesC[k],
+				&locals->BlockWidth256BytesY[k],
+				&locals->BlockWidth256BytesC[k],
+				&locals->BlockHeightY[k],
+				&locals->BlockHeightC[k],
+				&locals->BlockWidthY[k],
+				&locals->BlockWidthC[k]);
+	}
+
+
+	dml_print("DML::%s: %u\n", __func__, __LINE__);
+	CalculateSwathWidth(
+		false,  // ForceSingleDPP
+		mode_lib->ms.num_active_planes,
+		mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat,
+		mode_lib->ms.cache_display_cfg.plane.SourceScan,
+		mode_lib->ms.cache_display_cfg.plane.ViewportStationary,
+		mode_lib->ms.cache_display_cfg.plane.ViewportWidth,
+		mode_lib->ms.cache_display_cfg.plane.ViewportHeight,
+		mode_lib->ms.cache_display_cfg.plane.ViewportXStart,
+		mode_lib->ms.cache_display_cfg.plane.ViewportYStart,
+		mode_lib->ms.cache_display_cfg.plane.ViewportXStartC,
+		mode_lib->ms.cache_display_cfg.plane.ViewportYStartC,
+		mode_lib->ms.cache_display_cfg.surface.SurfaceWidthY,
+		mode_lib->ms.cache_display_cfg.surface.SurfaceWidthC,
+		mode_lib->ms.cache_display_cfg.surface.SurfaceHeightY,
+		mode_lib->ms.cache_display_cfg.surface.SurfaceHeightC,
+		mode_lib->ms.cache_display_cfg.hw.ODMMode,
+		locals->BytePerPixelY,
+		locals->BytePerPixelC,
+		locals->BlockHeight256BytesY,
+		locals->BlockHeight256BytesC,
+		locals->BlockWidth256BytesY,
+		locals->BlockWidth256BytesC,
+		mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming,
+		mode_lib->ms.cache_display_cfg.timing.HActive,
+		mode_lib->ms.cache_display_cfg.plane.HRatio,
+		mode_lib->ms.cache_display_cfg.hw.DPPPerSurface,
+
+		/* Output */
+		locals->SwathWidthSingleDPPY,
+		locals->SwathWidthSingleDPPC,
+		locals->SwathWidthY,
+		locals->SwathWidthC,
+		s->dummy_integer_array[0], // dml_uint_t MaximumSwathHeightY[]
+		s->dummy_integer_array[1], // dml_uint_t MaximumSwathHeightC[]
+		locals->swath_width_luma_ub,
+		locals->swath_width_chroma_ub);
+
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		locals->ReadBandwidthSurfaceLuma[k] = locals->SwathWidthSingleDPPY[k] * locals->BytePerPixelY[k] / (mode_lib->ms.cache_display_cfg.timing.HTotal[k] / mode_lib->ms.cache_display_cfg.timing.PixelClock[k]) * mode_lib->ms.cache_display_cfg.plane.VRatio[k];
+		locals->ReadBandwidthSurfaceChroma[k] = locals->SwathWidthSingleDPPC[k] * locals->BytePerPixelC[k] / (mode_lib->ms.cache_display_cfg.timing.HTotal[k] / mode_lib->ms.cache_display_cfg.timing.PixelClock[k]) * mode_lib->ms.cache_display_cfg.plane.VRatioChroma[k];
+		dml_print("DML::%s: ReadBandwidthSurfaceLuma[%i] = %fBps\n", __func__, k, locals->ReadBandwidthSurfaceLuma[k]);
+		dml_print("DML::%s: ReadBandwidthSurfaceChroma[%i] = %fBps\n", __func__, k, locals->ReadBandwidthSurfaceChroma[k]);
+	}
+
+	CalculateSwathAndDETConfiguration_params->DETSizeOverride = mode_lib->ms.cache_display_cfg.plane.DETSizeOverride;
+	CalculateSwathAndDETConfiguration_params->UseMALLForPStateChange = mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange;
+	CalculateSwathAndDETConfiguration_params->ConfigReturnBufferSizeInKByte = mode_lib->ms.ip.config_return_buffer_size_in_kbytes;
+	CalculateSwathAndDETConfiguration_params->ROBBufferSizeInKByte = mode_lib->ms.ip.rob_buffer_size_kbytes;
+	CalculateSwathAndDETConfiguration_params->MaxTotalDETInKByte = mode_lib->ms.MaxTotalDETInKByte;
+	CalculateSwathAndDETConfiguration_params->MinCompressedBufferSizeInKByte = mode_lib->ms.MinCompressedBufferSizeInKByte;
+	CalculateSwathAndDETConfiguration_params->PixelChunkSizeInKByte = mode_lib->ms.ip.pixel_chunk_size_kbytes;
+	CalculateSwathAndDETConfiguration_params->ForceSingleDPP = false;
+	CalculateSwathAndDETConfiguration_params->NumberOfActiveSurfaces = mode_lib->ms.num_active_planes;
+	CalculateSwathAndDETConfiguration_params->nomDETInKByte = mode_lib->ms.NomDETInKByte;
+	CalculateSwathAndDETConfiguration_params->UseUnboundedRequestingFinal = mode_lib->ms.policy.UseUnboundedRequesting;
+	CalculateSwathAndDETConfiguration_params->ConfigReturnBufferSegmentSizeInkByte = mode_lib->ms.ip.config_return_buffer_segment_size_in_kbytes;
+	CalculateSwathAndDETConfiguration_params->CompressedBufferSegmentSizeInkByteFinal = mode_lib->ms.ip.compressed_buffer_segment_size_in_kbytes;
+	CalculateSwathAndDETConfiguration_params->Output = s->dummy_output_encoder_array;
+	CalculateSwathAndDETConfiguration_params->ReadBandwidthLuma = locals->ReadBandwidthSurfaceLuma;
+	CalculateSwathAndDETConfiguration_params->ReadBandwidthChroma = locals->ReadBandwidthSurfaceChroma;
+	CalculateSwathAndDETConfiguration_params->MaximumSwathWidthLuma = s->dummy_single_array[0];
+	CalculateSwathAndDETConfiguration_params->MaximumSwathWidthChroma = s->dummy_single_array[1];
+	CalculateSwathAndDETConfiguration_params->SourceScan = mode_lib->ms.cache_display_cfg.plane.SourceScan;
+	CalculateSwathAndDETConfiguration_params->ViewportStationary = mode_lib->ms.cache_display_cfg.plane.ViewportStationary;
+	CalculateSwathAndDETConfiguration_params->SourcePixelFormat = mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat;
+	CalculateSwathAndDETConfiguration_params->SurfaceTiling = mode_lib->ms.cache_display_cfg.surface.SurfaceTiling;
+	CalculateSwathAndDETConfiguration_params->ViewportWidth = mode_lib->ms.cache_display_cfg.plane.ViewportWidth;
+	CalculateSwathAndDETConfiguration_params->ViewportHeight = mode_lib->ms.cache_display_cfg.plane.ViewportHeight;
+	CalculateSwathAndDETConfiguration_params->ViewportXStart = mode_lib->ms.cache_display_cfg.plane.ViewportXStart;
+	CalculateSwathAndDETConfiguration_params->ViewportYStart = mode_lib->ms.cache_display_cfg.plane.ViewportYStart;
+	CalculateSwathAndDETConfiguration_params->ViewportXStartC = mode_lib->ms.cache_display_cfg.plane.ViewportXStartC;
+	CalculateSwathAndDETConfiguration_params->ViewportYStartC = mode_lib->ms.cache_display_cfg.plane.ViewportYStartC;
+	CalculateSwathAndDETConfiguration_params->SurfaceWidthY = mode_lib->ms.cache_display_cfg.surface.SurfaceWidthY;
+	CalculateSwathAndDETConfiguration_params->SurfaceWidthC = mode_lib->ms.cache_display_cfg.surface.SurfaceWidthC;
+	CalculateSwathAndDETConfiguration_params->SurfaceHeightY = mode_lib->ms.cache_display_cfg.surface.SurfaceHeightY;
+	CalculateSwathAndDETConfiguration_params->SurfaceHeightC = mode_lib->ms.cache_display_cfg.surface.SurfaceHeightC;
+	CalculateSwathAndDETConfiguration_params->Read256BytesBlockHeightY = locals->BlockHeight256BytesY;
+	CalculateSwathAndDETConfiguration_params->Read256BytesBlockHeightC = locals->BlockHeight256BytesC;
+	CalculateSwathAndDETConfiguration_params->Read256BytesBlockWidthY = locals->BlockWidth256BytesY;
+	CalculateSwathAndDETConfiguration_params->Read256BytesBlockWidthC = locals->BlockWidth256BytesC;
+	CalculateSwathAndDETConfiguration_params->ODMMode = mode_lib->ms.cache_display_cfg.hw.ODMMode;
+	CalculateSwathAndDETConfiguration_params->BlendingAndTiming = mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming;
+	CalculateSwathAndDETConfiguration_params->BytePerPixY = locals->BytePerPixelY;
+	CalculateSwathAndDETConfiguration_params->BytePerPixC = locals->BytePerPixelC;
+	CalculateSwathAndDETConfiguration_params->BytePerPixDETY = locals->BytePerPixelDETY;
+	CalculateSwathAndDETConfiguration_params->BytePerPixDETC = locals->BytePerPixelDETC;
+	CalculateSwathAndDETConfiguration_params->HActive = mode_lib->ms.cache_display_cfg.timing.HActive;
+	CalculateSwathAndDETConfiguration_params->HRatio = mode_lib->ms.cache_display_cfg.plane.HRatio;
+	CalculateSwathAndDETConfiguration_params->HRatioChroma = mode_lib->ms.cache_display_cfg.plane.HRatioChroma;
+	CalculateSwathAndDETConfiguration_params->DPPPerSurface = mode_lib->ms.cache_display_cfg.hw.DPPPerSurface;
+	CalculateSwathAndDETConfiguration_params->swath_width_luma_ub = s->dummy_long_array[0];
+	CalculateSwathAndDETConfiguration_params->swath_width_chroma_ub = s->dummy_long_array[1];
+	CalculateSwathAndDETConfiguration_params->SwathWidth = s->dummy_long_array[2];
+	CalculateSwathAndDETConfiguration_params->SwathWidthChroma = s->dummy_long_array[3];
+	CalculateSwathAndDETConfiguration_params->SwathHeightY = locals->SwathHeightY;
+	CalculateSwathAndDETConfiguration_params->SwathHeightC = locals->SwathHeightC;
+	CalculateSwathAndDETConfiguration_params->DETBufferSizeInKByte = locals->DETBufferSizeInKByte;
+	CalculateSwathAndDETConfiguration_params->DETBufferSizeY = locals->DETBufferSizeY;
+	CalculateSwathAndDETConfiguration_params->DETBufferSizeC = locals->DETBufferSizeC;
+	CalculateSwathAndDETConfiguration_params->UnboundedRequestEnabled = &locals->UnboundedRequestEnabled;
+	CalculateSwathAndDETConfiguration_params->compbuf_reserved_space_64b = &locals->compbuf_reserved_space_64b;
+	CalculateSwathAndDETConfiguration_params->compbuf_reserved_space_zs = &locals->compbuf_reserved_space_zs;
+	CalculateSwathAndDETConfiguration_params->CompressedBufferSizeInkByte = &locals->CompressedBufferSizeInkByte;
+	CalculateSwathAndDETConfiguration_params->ViewportSizeSupportPerSurface = &s->dummy_boolean_array[0][0];
+	CalculateSwathAndDETConfiguration_params->ViewportSizeSupport = &s->dummy_boolean[0];
+
+	// VBA_DELTA
+	// Calculate DET size, swath height here. In VBA, they are calculated in mode check stage
+	CalculateSwathAndDETConfiguration(&mode_lib->scratch,
+		CalculateSwathAndDETConfiguration_params);
+
+	// DCFCLK Deep Sleep
+	CalculateDCFCLKDeepSleep(
+			mode_lib->ms.num_active_planes,
+			locals->BytePerPixelY,
+			locals->BytePerPixelC,
+			mode_lib->ms.cache_display_cfg.plane.VRatio,
+			mode_lib->ms.cache_display_cfg.plane.VRatioChroma,
+			locals->SwathWidthY,
+			locals->SwathWidthC,
+			mode_lib->ms.cache_display_cfg.hw.DPPPerSurface,
+			mode_lib->ms.cache_display_cfg.plane.HRatio,
+			mode_lib->ms.cache_display_cfg.plane.HRatioChroma,
+			mode_lib->ms.cache_display_cfg.timing.PixelClock,
+			locals->PSCL_THROUGHPUT,
+			locals->PSCL_THROUGHPUT_CHROMA,
+			locals->Dppclk,
+			locals->ReadBandwidthSurfaceLuma,
+			locals->ReadBandwidthSurfaceChroma,
+			mode_lib->ms.soc.return_bus_width_bytes,
+
+			/* Output */
+			&locals->DCFCLKDeepSleep);
+
+	// DSCCLK
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		if ((mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming[k] != k) || !mode_lib->ms.cache_display_cfg.hw.DSCEnabled[k]) {
+			locals->DSCCLK_calculated[k] = 0.0;
+		} else {
+			if (mode_lib->ms.cache_display_cfg.output.OutputFormat[k] == dml_420)
+				s->DSCFormatFactor = 2;
+			else if (mode_lib->ms.cache_display_cfg.output.OutputFormat[k] == dml_444)
+				s->DSCFormatFactor = 1;
+			else if (mode_lib->ms.cache_display_cfg.output.OutputFormat[k] == dml_n422 || mode_lib->ms.cache_display_cfg.output.OutputEncoder[k] == dml_hdmifrl)
+				s->DSCFormatFactor = 2;
+			else
+				s->DSCFormatFactor = 1;
+			if (mode_lib->ms.cache_display_cfg.hw.ODMMode[k] == dml_odm_mode_combine_4to1)
+				locals->DSCCLK_calculated[k] = mode_lib->ms.cache_display_cfg.output.PixelClockBackEnd[k] / 12 / s->DSCFormatFactor / (1 - mode_lib->ms.soc.dcn_downspread_percent / 100);
+			else if (mode_lib->ms.cache_display_cfg.hw.ODMMode[k] == dml_odm_mode_combine_2to1)
+				locals->DSCCLK_calculated[k] = mode_lib->ms.cache_display_cfg.output.PixelClockBackEnd[k] / 6 / s->DSCFormatFactor / (1 - mode_lib->ms.soc.dcn_downspread_percent / 100);
+			else
+				locals->DSCCLK_calculated[k] = mode_lib->ms.cache_display_cfg.output.PixelClockBackEnd[k] / 3 / s->DSCFormatFactor / (1 - mode_lib->ms.soc.dcn_downspread_percent / 100);
+		}
+	}
+
+	// DSC Delay
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		locals->DSCDelay[k] = DSCDelayRequirement(mode_lib->ms.cache_display_cfg.hw.DSCEnabled[k],
+												mode_lib->ms.cache_display_cfg.hw.ODMMode[k],
+												mode_lib->ms.cache_display_cfg.output.DSCInputBitPerComponent[k],
+												mode_lib->ms.cache_display_cfg.output.OutputBpp[k],
+												mode_lib->ms.cache_display_cfg.timing.HActive[k],
+												mode_lib->ms.cache_display_cfg.timing.HTotal[k],
+												mode_lib->ms.cache_display_cfg.hw.NumberOfDSCSlices[k],
+												mode_lib->ms.cache_display_cfg.output.OutputFormat[k],
+												mode_lib->ms.cache_display_cfg.output.OutputEncoder[k],
+												mode_lib->ms.cache_display_cfg.timing.PixelClock[k],
+												mode_lib->ms.cache_display_cfg.output.PixelClockBackEnd[k]);
+	}
+
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k)
+		for (j = 0; j < mode_lib->ms.num_active_planes; ++j) // NumberOfSurfaces
+			if (j != k && mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming[k] == j && mode_lib->ms.cache_display_cfg.hw.DSCEnabled[j])
+				locals->DSCDelay[k] = locals->DSCDelay[j];
+
+	// Prefetch
+	CalculateSurfaceSizeInMall(
+		mode_lib->ms.num_active_planes,
+		mode_lib->ms.soc.mall_allocated_for_dcn_mbytes,
+		mode_lib->ms.cache_display_cfg.plane.UseMALLForStaticScreen,
+		mode_lib->ms.cache_display_cfg.surface.DCCEnable,
+		mode_lib->ms.cache_display_cfg.plane.ViewportStationary,
+		mode_lib->ms.cache_display_cfg.plane.ViewportXStart,
+		mode_lib->ms.cache_display_cfg.plane.ViewportYStart,
+		mode_lib->ms.cache_display_cfg.plane.ViewportXStartC,
+		mode_lib->ms.cache_display_cfg.plane.ViewportYStartC,
+		mode_lib->ms.cache_display_cfg.plane.ViewportWidth,
+		mode_lib->ms.cache_display_cfg.plane.ViewportHeight,
+		locals->BytePerPixelY,
+		mode_lib->ms.cache_display_cfg.plane.ViewportWidthChroma,
+		mode_lib->ms.cache_display_cfg.plane.ViewportHeightChroma,
+		locals->BytePerPixelC,
+		mode_lib->ms.cache_display_cfg.surface.SurfaceWidthY,
+		mode_lib->ms.cache_display_cfg.surface.SurfaceWidthC,
+		mode_lib->ms.cache_display_cfg.surface.SurfaceHeightY,
+		mode_lib->ms.cache_display_cfg.surface.SurfaceHeightC,
+		locals->BlockWidth256BytesY,
+		locals->BlockWidth256BytesC,
+		locals->BlockHeight256BytesY,
+		locals->BlockHeight256BytesC,
+		locals->BlockWidthY,
+		locals->BlockWidthC,
+		locals->BlockHeightY,
+		locals->BlockHeightC,
+
+		/* Output */
+		locals->SurfaceSizeInTheMALL,
+		&s->dummy_boolean[0]); /* dml_bool_t *ExceededMALLSize */
+
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		s->SurfaceParameters[k].PixelClock = mode_lib->ms.cache_display_cfg.timing.PixelClock[k];
+		s->SurfaceParameters[k].DPPPerSurface = mode_lib->ms.cache_display_cfg.hw.DPPPerSurface[k];
+		s->SurfaceParameters[k].SourceScan = mode_lib->ms.cache_display_cfg.plane.SourceScan[k];
+		s->SurfaceParameters[k].ViewportHeight = mode_lib->ms.cache_display_cfg.plane.ViewportHeight[k];
+		s->SurfaceParameters[k].ViewportHeightChroma = mode_lib->ms.cache_display_cfg.plane.ViewportHeightChroma[k];
+		s->SurfaceParameters[k].BlockWidth256BytesY = locals->BlockWidth256BytesY[k];
+		s->SurfaceParameters[k].BlockHeight256BytesY = locals->BlockHeight256BytesY[k];
+		s->SurfaceParameters[k].BlockWidth256BytesC = locals->BlockWidth256BytesC[k];
+		s->SurfaceParameters[k].BlockHeight256BytesC = locals->BlockHeight256BytesC[k];
+		s->SurfaceParameters[k].BlockWidthY = locals->BlockWidthY[k];
+		s->SurfaceParameters[k].BlockHeightY = locals->BlockHeightY[k];
+		s->SurfaceParameters[k].BlockWidthC = locals->BlockWidthC[k];
+		s->SurfaceParameters[k].BlockHeightC = locals->BlockHeightC[k];
+		s->SurfaceParameters[k].InterlaceEnable = mode_lib->ms.cache_display_cfg.timing.Interlace[k];
+		s->SurfaceParameters[k].HTotal = mode_lib->ms.cache_display_cfg.timing.HTotal[k];
+		s->SurfaceParameters[k].DCCEnable = mode_lib->ms.cache_display_cfg.surface.DCCEnable[k];
+		s->SurfaceParameters[k].SourcePixelFormat = mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k];
+		s->SurfaceParameters[k].SurfaceTiling = mode_lib->ms.cache_display_cfg.surface.SurfaceTiling[k];
+		s->SurfaceParameters[k].BytePerPixelY = locals->BytePerPixelY[k];
+		s->SurfaceParameters[k].BytePerPixelC = locals->BytePerPixelC[k];
+		s->SurfaceParameters[k].ProgressiveToInterlaceUnitInOPP = mode_lib->ms.ip.ptoi_supported;
+		s->SurfaceParameters[k].VRatio = mode_lib->ms.cache_display_cfg.plane.VRatio[k];
+		s->SurfaceParameters[k].VRatioChroma = mode_lib->ms.cache_display_cfg.plane.VRatioChroma[k];
+		s->SurfaceParameters[k].VTaps = mode_lib->ms.cache_display_cfg.plane.VTaps[k];
+		s->SurfaceParameters[k].VTapsChroma = mode_lib->ms.cache_display_cfg.plane.VTapsChroma[k];
+		s->SurfaceParameters[k].PitchY = mode_lib->ms.cache_display_cfg.surface.PitchY[k];
+		s->SurfaceParameters[k].DCCMetaPitchY = mode_lib->ms.cache_display_cfg.surface.DCCMetaPitchY[k];
+		s->SurfaceParameters[k].PitchC = mode_lib->ms.cache_display_cfg.surface.PitchC[k];
+		s->SurfaceParameters[k].DCCMetaPitchC = mode_lib->ms.cache_display_cfg.surface.DCCMetaPitchC[k];
+		s->SurfaceParameters[k].ViewportStationary = mode_lib->ms.cache_display_cfg.plane.ViewportStationary[k];
+		s->SurfaceParameters[k].ViewportXStart = mode_lib->ms.cache_display_cfg.plane.ViewportXStart[k];
+		s->SurfaceParameters[k].ViewportYStart = mode_lib->ms.cache_display_cfg.plane.ViewportYStart[k];
+		s->SurfaceParameters[k].ViewportXStartC = mode_lib->ms.cache_display_cfg.plane.ViewportXStartC[k];
+		s->SurfaceParameters[k].ViewportYStartC = mode_lib->ms.cache_display_cfg.plane.ViewportYStartC[k];
+		s->SurfaceParameters[k].FORCE_ONE_ROW_FOR_FRAME = mode_lib->ms.cache_display_cfg.plane.ForceOneRowForFrame[k];
+		s->SurfaceParameters[k].SwathHeightY = locals->SwathHeightY[k];
+		s->SurfaceParameters[k].SwathHeightC = locals->SwathHeightC[k];
+	}
+
+	CalculateVMRowAndSwath_params->NumberOfActiveSurfaces = mode_lib->ms.num_active_planes;
+	CalculateVMRowAndSwath_params->myPipe = s->SurfaceParameters;
+	CalculateVMRowAndSwath_params->SurfaceSizeInMALL = locals->SurfaceSizeInTheMALL;
+	CalculateVMRowAndSwath_params->PTEBufferSizeInRequestsLuma = mode_lib->ms.ip.dpte_buffer_size_in_pte_reqs_luma;
+	CalculateVMRowAndSwath_params->PTEBufferSizeInRequestsChroma = mode_lib->ms.ip.dpte_buffer_size_in_pte_reqs_chroma;
+	CalculateVMRowAndSwath_params->DCCMetaBufferSizeBytes = mode_lib->ms.ip.dcc_meta_buffer_size_bytes;
+	CalculateVMRowAndSwath_params->UseMALLForStaticScreen = mode_lib->ms.cache_display_cfg.plane.UseMALLForStaticScreen;
+	CalculateVMRowAndSwath_params->UseMALLForPStateChange = mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange;
+	CalculateVMRowAndSwath_params->MALLAllocatedForDCN = mode_lib->ms.soc.mall_allocated_for_dcn_mbytes;
+	CalculateVMRowAndSwath_params->SwathWidthY = locals->SwathWidthY;
+	CalculateVMRowAndSwath_params->SwathWidthC = locals->SwathWidthC;
+	CalculateVMRowAndSwath_params->GPUVMEnable = mode_lib->ms.cache_display_cfg.plane.GPUVMEnable;
+	CalculateVMRowAndSwath_params->HostVMEnable = mode_lib->ms.cache_display_cfg.plane.HostVMEnable;
+	CalculateVMRowAndSwath_params->HostVMMaxNonCachedPageTableLevels = mode_lib->ms.cache_display_cfg.plane.HostVMMaxPageTableLevels;
+	CalculateVMRowAndSwath_params->GPUVMMaxPageTableLevels = mode_lib->ms.cache_display_cfg.plane.GPUVMMaxPageTableLevels;
+	CalculateVMRowAndSwath_params->GPUVMMinPageSizeKBytes = mode_lib->ms.cache_display_cfg.plane.GPUVMMinPageSizeKBytes;
+	CalculateVMRowAndSwath_params->HostVMMinPageSize = mode_lib->ms.soc.hostvm_min_page_size_kbytes * 1024;
+	CalculateVMRowAndSwath_params->PTEBufferModeOverrideEn = mode_lib->ms.cache_display_cfg.plane.PTEBufferModeOverrideEn;
+	CalculateVMRowAndSwath_params->PTEBufferModeOverrideVal = mode_lib->ms.cache_display_cfg.plane.PTEBufferMode;
+	CalculateVMRowAndSwath_params->PTEBufferSizeNotExceeded = s->dummy_boolean_array[0];
+	CalculateVMRowAndSwath_params->DCCMetaBufferSizeNotExceeded = s->dummy_boolean_array[1];
+	CalculateVMRowAndSwath_params->dpte_row_width_luma_ub = locals->dpte_row_width_luma_ub;
+	CalculateVMRowAndSwath_params->dpte_row_width_chroma_ub = locals->dpte_row_width_chroma_ub;
+	CalculateVMRowAndSwath_params->dpte_row_height_luma = locals->dpte_row_height;
+	CalculateVMRowAndSwath_params->dpte_row_height_chroma = locals->dpte_row_height_chroma;
+	CalculateVMRowAndSwath_params->dpte_row_height_linear_luma = locals->dpte_row_height_linear;
+	CalculateVMRowAndSwath_params->dpte_row_height_linear_chroma = locals->dpte_row_height_linear_chroma;
+	CalculateVMRowAndSwath_params->meta_req_width = locals->meta_req_width;
+	CalculateVMRowAndSwath_params->meta_req_width_chroma = locals->meta_req_width_chroma;
+	CalculateVMRowAndSwath_params->meta_req_height = locals->meta_req_height;
+	CalculateVMRowAndSwath_params->meta_req_height_chroma = locals->meta_req_height_chroma;
+	CalculateVMRowAndSwath_params->meta_row_width = locals->meta_row_width;
+	CalculateVMRowAndSwath_params->meta_row_width_chroma = locals->meta_row_width_chroma;
+	CalculateVMRowAndSwath_params->meta_row_height = locals->meta_row_height;
+	CalculateVMRowAndSwath_params->meta_row_height_chroma = locals->meta_row_height_chroma;
+	CalculateVMRowAndSwath_params->vm_group_bytes = locals->vm_group_bytes;
+	CalculateVMRowAndSwath_params->dpte_group_bytes = locals->dpte_group_bytes;
+	CalculateVMRowAndSwath_params->PixelPTEReqWidthY = locals->PixelPTEReqWidthY;
+	CalculateVMRowAndSwath_params->PixelPTEReqHeightY = locals->PixelPTEReqHeightY;
+	CalculateVMRowAndSwath_params->PTERequestSizeY = locals->PTERequestSizeY;
+	CalculateVMRowAndSwath_params->PixelPTEReqWidthC = locals->PixelPTEReqWidthC;
+	CalculateVMRowAndSwath_params->PixelPTEReqHeightC = locals->PixelPTEReqHeightC;
+	CalculateVMRowAndSwath_params->PTERequestSizeC = locals->PTERequestSizeC;
+	CalculateVMRowAndSwath_params->dpde0_bytes_per_frame_ub_l = locals->dpde0_bytes_per_frame_ub_l;
+	CalculateVMRowAndSwath_params->meta_pte_bytes_per_frame_ub_l = locals->meta_pte_bytes_per_frame_ub_l;
+	CalculateVMRowAndSwath_params->dpde0_bytes_per_frame_ub_c = locals->dpde0_bytes_per_frame_ub_c;
+	CalculateVMRowAndSwath_params->meta_pte_bytes_per_frame_ub_c = locals->meta_pte_bytes_per_frame_ub_c;
+	CalculateVMRowAndSwath_params->PrefetchSourceLinesY = locals->PrefetchSourceLinesY;
+	CalculateVMRowAndSwath_params->PrefetchSourceLinesC = locals->PrefetchSourceLinesC;
+	CalculateVMRowAndSwath_params->VInitPreFillY = locals->VInitPreFillY;
+	CalculateVMRowAndSwath_params->VInitPreFillC = locals->VInitPreFillC;
+	CalculateVMRowAndSwath_params->MaxNumSwathY = locals->MaxNumSwathY;
+	CalculateVMRowAndSwath_params->MaxNumSwathC = locals->MaxNumSwathC;
+	CalculateVMRowAndSwath_params->meta_row_bw = locals->meta_row_bw;
+	CalculateVMRowAndSwath_params->dpte_row_bw = locals->dpte_row_bw;
+	CalculateVMRowAndSwath_params->PixelPTEBytesPerRow = locals->PixelPTEBytesPerRow;
+	CalculateVMRowAndSwath_params->PDEAndMetaPTEBytesFrame = locals->PDEAndMetaPTEBytesFrame;
+	CalculateVMRowAndSwath_params->MetaRowByte = locals->MetaRowByte;
+	CalculateVMRowAndSwath_params->use_one_row_for_frame = locals->use_one_row_for_frame;
+	CalculateVMRowAndSwath_params->use_one_row_for_frame_flip = locals->use_one_row_for_frame_flip;
+	CalculateVMRowAndSwath_params->UsesMALLForStaticScreen = locals->UsesMALLForStaticScreen;
+	CalculateVMRowAndSwath_params->PTE_BUFFER_MODE = locals->PTE_BUFFER_MODE;
+	CalculateVMRowAndSwath_params->BIGK_FRAGMENT_SIZE = locals->BIGK_FRAGMENT_SIZE;
+
+	CalculateVMRowAndSwath(&mode_lib->scratch,
+	CalculateVMRowAndSwath_params);
+
+	s->ReorderBytes = (dml_uint_t)(mode_lib->ms.soc.num_chans * dml_max3(
+			mode_lib->ms.soc.urgent_out_of_order_return_per_channel_pixel_only_bytes,
+			mode_lib->ms.soc.urgent_out_of_order_return_per_channel_pixel_and_vm_bytes,
+			mode_lib->ms.soc.urgent_out_of_order_return_per_channel_vm_only_bytes));
+
+	s->VMDataOnlyReturnBW = dml_get_return_bw_mbps_vm_only(&mode_lib->ms.soc,
+																	mode_lib->ms.state.use_ideal_dram_bw_strobe,
+																	mode_lib->ms.cache_display_cfg.plane.HostVMEnable,
+																	locals->Dcfclk,
+																	mode_lib->ms.FabricClock,
+																	mode_lib->ms.DRAMSpeed);
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: locals->Dcfclk = %f\n", __func__, locals->Dcfclk);
+	dml_print("DML::%s: mode_lib->ms.soc.return_bus_width_bytes = %u\n", __func__, mode_lib->ms.soc.return_bus_width_bytes);
+	dml_print("DML::%s: mode_lib->ms.FabricClock = %f\n", __func__, mode_lib->ms.FabricClock);
+	dml_print("DML::%s: mode_lib->ms.soc.fabric_datapath_to_dcn_data_return_bytes = %u\n", __func__, mode_lib->ms.soc.fabric_datapath_to_dcn_data_return_bytes);
+	dml_print("DML::%s: mode_lib->ms.soc.pct_ideal_sdp_bw_after_urgent = %f\n", __func__, mode_lib->ms.soc.pct_ideal_sdp_bw_after_urgent);
+	dml_print("DML::%s: mode_lib->ms.DRAMSpeed = %f\n", __func__, mode_lib->ms.DRAMSpeed);
+	dml_print("DML::%s: mode_lib->ms.soc.num_chans = %u\n", __func__, mode_lib->ms.soc.num_chans);
+	dml_print("DML::%s: mode_lib->ms.soc.dram_channel_width_bytes = %u\n", __func__, mode_lib->ms.soc.dram_channel_width_bytes);
+	dml_print("DML::%s: mode_lib->ms.state_idx = %u\n", __func__, mode_lib->ms.state_idx);
+	dml_print("DML::%s: mode_lib->ms.max_state_idx = %u\n", __func__, mode_lib->ms.max_state_idx);
+	dml_print("DML::%s: mode_lib->ms.state.use_ideal_dram_bw_strobe = %u\n", __func__, mode_lib->ms.state.use_ideal_dram_bw_strobe);
+	dml_print("DML::%s: VMDataOnlyReturnBW = %f\n", __func__, s->VMDataOnlyReturnBW);
+	dml_print("DML::%s: ReturnBW = %f\n",  __func__, mode_lib->ms.ReturnBW);
+#endif
+
+	s->HostVMInefficiencyFactor = 1.0;
+	if (mode_lib->ms.cache_display_cfg.plane.GPUVMEnable && mode_lib->ms.cache_display_cfg.plane.HostVMEnable)
+		s->HostVMInefficiencyFactor = mode_lib->ms.ReturnBW / s->VMDataOnlyReturnBW;
+
+	s->TotalDCCActiveDPP = 0;
+	s->TotalActiveDPP = 0;
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		s->TotalActiveDPP = s->TotalActiveDPP + mode_lib->ms.cache_display_cfg.hw.DPPPerSurface[k];
+		if (mode_lib->ms.cache_display_cfg.surface.DCCEnable[k])
+			s->TotalDCCActiveDPP = s->TotalDCCActiveDPP + mode_lib->ms.cache_display_cfg.hw.DPPPerSurface[k];
+	}
+
+	locals->UrgentExtraLatency = CalculateExtraLatency(
+			mode_lib->ms.soc.round_trip_ping_latency_dcfclk_cycles,
+			s->ReorderBytes,
+			locals->Dcfclk,
+			s->TotalActiveDPP,
+			mode_lib->ms.ip.pixel_chunk_size_kbytes,
+			s->TotalDCCActiveDPP,
+			mode_lib->ms.ip.meta_chunk_size_kbytes,
+			mode_lib->ms.ReturnBW,
+			mode_lib->ms.cache_display_cfg.plane.GPUVMEnable,
+			mode_lib->ms.cache_display_cfg.plane.HostVMEnable,
+			mode_lib->ms.num_active_planes,
+			mode_lib->ms.cache_display_cfg.hw.DPPPerSurface,
+			locals->dpte_group_bytes,
+			s->HostVMInefficiencyFactor,
+			mode_lib->ms.soc.hostvm_min_page_size_kbytes * 1024,
+			mode_lib->ms.cache_display_cfg.plane.HostVMMaxPageTableLevels);
+
+	locals->TCalc = 24.0 / locals->DCFCLKDeepSleep;
+
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		if (mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming[k] == k) {
+			if (mode_lib->ms.cache_display_cfg.writeback.WritebackEnable[k] == true) {
+				locals->WritebackDelay[k] =
+						mode_lib->ms.state.writeback_latency_us
+								+ CalculateWriteBackDelay(
+										mode_lib->ms.cache_display_cfg.writeback.WritebackPixelFormat[k],
+										mode_lib->ms.cache_display_cfg.writeback.WritebackHRatio[k],
+										mode_lib->ms.cache_display_cfg.writeback.WritebackVRatio[k],
+										mode_lib->ms.cache_display_cfg.writeback.WritebackVTaps[k],
+										mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationWidth[k],
+										mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationHeight[k],
+										mode_lib->ms.cache_display_cfg.writeback.WritebackSourceHeight[k],
+										mode_lib->ms.cache_display_cfg.timing.HTotal[k]) / locals->Dispclk;
+			} else
+				locals->WritebackDelay[k] = 0;
+			for (j = 0; j < mode_lib->ms.num_active_planes; ++j) {
+				if (mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming[j] == k
+						&& mode_lib->ms.cache_display_cfg.writeback.WritebackEnable[j] == true) {
+					locals->WritebackDelay[k] =
+							dml_max(
+									locals->WritebackDelay[k],
+									mode_lib->ms.state.writeback_latency_us
+											+ CalculateWriteBackDelay(
+													mode_lib->ms.cache_display_cfg.writeback.WritebackPixelFormat[j],
+													mode_lib->ms.cache_display_cfg.writeback.WritebackHRatio[j],
+													mode_lib->ms.cache_display_cfg.writeback.WritebackVRatio[j],
+													mode_lib->ms.cache_display_cfg.writeback.WritebackVTaps[j],
+													mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationWidth[j],
+													mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationHeight[j],
+													mode_lib->ms.cache_display_cfg.writeback.WritebackSourceHeight[j],
+													mode_lib->ms.cache_display_cfg.timing.HTotal[k]) / locals->Dispclk);
+				}
+			}
+		}
+	}
+
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k)
+		for (j = 0; j < mode_lib->ms.num_active_planes; ++j)
+			if (mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming[k] == j)
+				locals->WritebackDelay[k] = locals->WritebackDelay[j];
+
+	locals->UrgentLatency = CalculateUrgentLatency(mode_lib->ms.state.urgent_latency_pixel_data_only_us,
+												mode_lib->ms.state.urgent_latency_pixel_mixed_with_vm_data_us,
+												mode_lib->ms.state.urgent_latency_vm_data_only_us,
+												mode_lib->ms.soc.do_urgent_latency_adjustment,
+												mode_lib->ms.state.urgent_latency_adjustment_fabric_clock_component_us,
+												mode_lib->ms.state.urgent_latency_adjustment_fabric_clock_reference_mhz,
+												mode_lib->ms.FabricClock);
+
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		CalculateUrgentBurstFactor(mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange[k],
+									locals->swath_width_luma_ub[k],
+									locals->swath_width_chroma_ub[k],
+									locals->SwathHeightY[k],
+									locals->SwathHeightC[k],
+									mode_lib->ms.cache_display_cfg.timing.HTotal[k] / mode_lib->ms.cache_display_cfg.timing.PixelClock[k],
+									locals->UrgentLatency,
+									mode_lib->ms.ip.cursor_buffer_size,
+									mode_lib->ms.cache_display_cfg.plane.CursorWidth[k],
+									mode_lib->ms.cache_display_cfg.plane.CursorBPP[k],
+									mode_lib->ms.cache_display_cfg.plane.VRatio[k],
+									mode_lib->ms.cache_display_cfg.plane.VRatioChroma[k],
+									locals->BytePerPixelDETY[k],
+									locals->BytePerPixelDETC[k],
+									locals->DETBufferSizeY[k],
+									locals->DETBufferSizeC[k],
+
+									/* output */
+									&locals->UrgBurstFactorCursor[k],
+									&locals->UrgBurstFactorLuma[k],
+									&locals->UrgBurstFactorChroma[k],
+									&locals->NoUrgentLatencyHiding[k]);
+
+		locals->cursor_bw[k] = mode_lib->ms.cache_display_cfg.plane.NumberOfCursors[k] * mode_lib->ms.cache_display_cfg.plane.CursorWidth[k] * mode_lib->ms.cache_display_cfg.plane.CursorBPP[k] / 8.0 /
+								((dml_float_t) mode_lib->ms.cache_display_cfg.timing.HTotal[k] / mode_lib->ms.cache_display_cfg.timing.PixelClock[k]) * mode_lib->ms.cache_display_cfg.plane.VRatio[k];
+	}
+
+	s->VStartupLines = __DML_VBA_MIN_VSTARTUP__;
+	s->MaxVStartupAllPlanes = 0;
+
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		s->MaxVStartupLines[k] = CalculateMaxVStartup(k,
+													mode_lib->ms.ip.ptoi_supported,
+													mode_lib->ms.ip.vblank_nom_default_us,
+													&mode_lib->ms.cache_display_cfg.timing,
+													locals->WritebackDelay[k]);
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%u MaxVStartupLines   = %u\n", __func__, k, s->MaxVStartupLines[k]);
+		dml_print("DML::%s: k=%u WritebackDelay     = %f\n", __func__, k, locals->WritebackDelay[k]);
+#endif
+	}
+
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k)
+		s->MaxVStartupAllPlanes = (dml_uint_t)(dml_max(s->MaxVStartupAllPlanes, s->MaxVStartupLines[k]));
+
+	s->ImmediateFlipRequirementFinal = false;
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		s->ImmediateFlipRequirementFinal = s->ImmediateFlipRequirementFinal || (mode_lib->ms.policy.ImmediateFlipRequirement[k] == dml_immediate_flip_required);
+	}
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: ImmediateFlipRequirementFinal = %u\n", __func__, s->ImmediateFlipRequirementFinal);
+#endif
+
+	// The prefetch scheduling should only be calculated once as per AllowForPStateChangeOrStutterInVBlank requirement
+	// If the AllowForPStateChangeOrStutterInVBlank requirement is not strict (i.e. only try those power saving feature
+	// if possible, then will try to program for the best power saving features in order of difficulty (dram, fclk, stutter)
+	s->iteration = 0;
+	s->MaxTotalRDBandwidth = 0;
+	s->AllPrefetchModeTested = false;
+	for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+		CalculatePrefetchMode(mode_lib->ms.policy.AllowForPStateChangeOrStutterInVBlank[k], &s->MinPrefetchMode[k], &s->MaxPrefetchMode[k]);
+		s->NextPrefetchMode[k] = s->MinPrefetchMode[k];
+	}
+
+	do {
+		s->MaxTotalRDBandwidthNoUrgentBurst = 0.0;
+		s->DestinationLineTimesForPrefetchLessThan2 = false;
+		s->VRatioPrefetchMoreThanMax = false;
+
+		dml_print("DML::%s: Start one iteration: VStartupLines = %u\n", __func__, s->VStartupLines);
+
+		s->AllPrefetchModeTested = true;
+		s->MaxTotalRDBandwidth = 0;
+		for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+			locals->PrefetchMode[k] = s->NextPrefetchMode[k];
+			TWait = CalculateTWait(
+					locals->PrefetchMode[k],
+					mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange[k],
+					mode_lib->ms.policy.SynchronizeDRRDisplaysForUCLKPStateChangeFinal,
+					mode_lib->ms.cache_display_cfg.timing.DRRDisplay[k],
+					mode_lib->ms.state.dram_clock_change_latency_us,
+					mode_lib->ms.state.fclk_change_latency_us,
+					locals->UrgentLatency,
+					mode_lib->ms.state.sr_enter_plus_exit_time_us);
+
+			myPipe = &s->myPipe;
+			myPipe->Dppclk = locals->Dppclk[k];
+			myPipe->Dispclk = locals->Dispclk;
+			myPipe->PixelClock = mode_lib->ms.cache_display_cfg.timing.PixelClock[k];
+			myPipe->DCFClkDeepSleep = locals->DCFCLKDeepSleep;
+			myPipe->DPPPerSurface = mode_lib->ms.cache_display_cfg.hw.DPPPerSurface[k];
+			myPipe->ScalerEnabled = mode_lib->ms.cache_display_cfg.plane.ScalerEnabled[k];
+			myPipe->SourceScan = mode_lib->ms.cache_display_cfg.plane.SourceScan[k];
+			myPipe->BlockWidth256BytesY = locals->BlockWidth256BytesY[k];
+			myPipe->BlockHeight256BytesY = locals->BlockHeight256BytesY[k];
+			myPipe->BlockWidth256BytesC = locals->BlockWidth256BytesC[k];
+			myPipe->BlockHeight256BytesC = locals->BlockHeight256BytesC[k];
+			myPipe->InterlaceEnable = mode_lib->ms.cache_display_cfg.timing.Interlace[k];
+			myPipe->NumberOfCursors = mode_lib->ms.cache_display_cfg.plane.NumberOfCursors[k];
+			myPipe->VBlank = mode_lib->ms.cache_display_cfg.timing.VTotal[k] - mode_lib->ms.cache_display_cfg.timing.VActive[k];
+			myPipe->HTotal = mode_lib->ms.cache_display_cfg.timing.HTotal[k];
+			myPipe->HActive = mode_lib->ms.cache_display_cfg.timing.HActive[k];
+			myPipe->DCCEnable = mode_lib->ms.cache_display_cfg.surface.DCCEnable[k];
+			myPipe->ODMMode = mode_lib->ms.cache_display_cfg.hw.ODMMode[k];
+			myPipe->SourcePixelFormat = mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k];
+			myPipe->BytePerPixelY = locals->BytePerPixelY[k];
+			myPipe->BytePerPixelC = locals->BytePerPixelC[k];
+			myPipe->ProgressiveToInterlaceUnitInOPP = mode_lib->ms.ip.ptoi_supported;
+
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: Calling CalculatePrefetchSchedule for k=%u\n", __func__, k);
+			dml_print("DML::%s: AllowForPStateChangeOrStutterInVBlank = %u\n", __func__, mode_lib->ms.policy.AllowForPStateChangeOrStutterInVBlank[k]);
+			dml_print("DML::%s: PrefetchMode[k] = %u (Min=%u Max=%u)\n", __func__, locals->PrefetchMode[k], s->MinPrefetchMode[k], s->MaxPrefetchMode[k]);
+#endif
+
+			CalculatePrefetchSchedule_params->EnhancedPrefetchScheduleAccelerationFinal = mode_lib->ms.policy.EnhancedPrefetchScheduleAccelerationFinal;
+			CalculatePrefetchSchedule_params->HostVMInefficiencyFactor = s->HostVMInefficiencyFactor;
+			CalculatePrefetchSchedule_params->myPipe = myPipe;
+			CalculatePrefetchSchedule_params->DSCDelay = locals->DSCDelay[k];
+			CalculatePrefetchSchedule_params->DPPCLKDelaySubtotalPlusCNVCFormater = mode_lib->ms.ip.dppclk_delay_subtotal + mode_lib->ms.ip.dppclk_delay_cnvc_formatter;
+			CalculatePrefetchSchedule_params->DPPCLKDelaySCL = mode_lib->ms.ip.dppclk_delay_scl;
+			CalculatePrefetchSchedule_params->DPPCLKDelaySCLLBOnly = mode_lib->ms.ip.dppclk_delay_scl_lb_only;
+			CalculatePrefetchSchedule_params->DPPCLKDelayCNVCCursor = mode_lib->ms.ip.dppclk_delay_cnvc_cursor;
+			CalculatePrefetchSchedule_params->DISPCLKDelaySubtotal = mode_lib->ms.ip.dispclk_delay_subtotal;
+			CalculatePrefetchSchedule_params->DPP_RECOUT_WIDTH = (dml_uint_t)(locals->SwathWidthY[k] / mode_lib->ms.cache_display_cfg.plane.HRatio[k]);
+			CalculatePrefetchSchedule_params->OutputFormat = mode_lib->ms.cache_display_cfg.output.OutputFormat[k];
+			CalculatePrefetchSchedule_params->MaxInterDCNTileRepeaters = mode_lib->ms.ip.max_inter_dcn_tile_repeaters;
+			CalculatePrefetchSchedule_params->VStartup = (dml_uint_t)(dml_min(s->VStartupLines, s->MaxVStartupLines[k]));
+			CalculatePrefetchSchedule_params->MaxVStartup = s->MaxVStartupLines[k];
+			CalculatePrefetchSchedule_params->GPUVMPageTableLevels = mode_lib->ms.cache_display_cfg.plane.GPUVMMaxPageTableLevels;
+			CalculatePrefetchSchedule_params->GPUVMEnable = mode_lib->ms.cache_display_cfg.plane.GPUVMEnable;
+			CalculatePrefetchSchedule_params->HostVMEnable = mode_lib->ms.cache_display_cfg.plane.HostVMEnable;
+			CalculatePrefetchSchedule_params->HostVMMaxNonCachedPageTableLevels = mode_lib->ms.cache_display_cfg.plane.HostVMMaxPageTableLevels;
+			CalculatePrefetchSchedule_params->HostVMMinPageSize = mode_lib->ms.soc.hostvm_min_page_size_kbytes * 1024;
+			CalculatePrefetchSchedule_params->DynamicMetadataEnable = mode_lib->ms.cache_display_cfg.plane.DynamicMetadataEnable[k];
+			CalculatePrefetchSchedule_params->DynamicMetadataVMEnabled = mode_lib->ms.ip.dynamic_metadata_vm_enabled;
+			CalculatePrefetchSchedule_params->DynamicMetadataLinesBeforeActiveRequired = mode_lib->ms.cache_display_cfg.plane.DynamicMetadataLinesBeforeActiveRequired[k];
+			CalculatePrefetchSchedule_params->DynamicMetadataTransmittedBytes = mode_lib->ms.cache_display_cfg.plane.DynamicMetadataTransmittedBytes[k];
+			CalculatePrefetchSchedule_params->UrgentLatency = locals->UrgentLatency;
+			CalculatePrefetchSchedule_params->UrgentExtraLatency = locals->UrgentExtraLatency;
+			CalculatePrefetchSchedule_params->TCalc = locals->TCalc;
+			CalculatePrefetchSchedule_params->PDEAndMetaPTEBytesFrame = locals->PDEAndMetaPTEBytesFrame[k];
+			CalculatePrefetchSchedule_params->MetaRowByte = locals->MetaRowByte[k];
+			CalculatePrefetchSchedule_params->PixelPTEBytesPerRow = locals->PixelPTEBytesPerRow[k];
+			CalculatePrefetchSchedule_params->PrefetchSourceLinesY = locals->PrefetchSourceLinesY[k];
+			CalculatePrefetchSchedule_params->VInitPreFillY = locals->VInitPreFillY[k];
+			CalculatePrefetchSchedule_params->MaxNumSwathY = locals->MaxNumSwathY[k];
+			CalculatePrefetchSchedule_params->PrefetchSourceLinesC = locals->PrefetchSourceLinesC[k];
+			CalculatePrefetchSchedule_params->VInitPreFillC = locals->VInitPreFillC[k];
+			CalculatePrefetchSchedule_params->MaxNumSwathC = locals->MaxNumSwathC[k];
+			CalculatePrefetchSchedule_params->swath_width_luma_ub = locals->swath_width_luma_ub[k];
+			CalculatePrefetchSchedule_params->swath_width_chroma_ub = locals->swath_width_chroma_ub[k];
+			CalculatePrefetchSchedule_params->SwathHeightY = locals->SwathHeightY[k];
+			CalculatePrefetchSchedule_params->SwathHeightC = locals->SwathHeightC[k];
+			CalculatePrefetchSchedule_params->TWait = TWait;
+			CalculatePrefetchSchedule_params->DSTXAfterScaler = &locals->DSTXAfterScaler[k];
+			CalculatePrefetchSchedule_params->DSTYAfterScaler = &locals->DSTYAfterScaler[k];
+			CalculatePrefetchSchedule_params->DestinationLinesForPrefetch = &locals->DestinationLinesForPrefetch[k];
+			CalculatePrefetchSchedule_params->DestinationLinesToRequestVMInVBlank = &locals->DestinationLinesToRequestVMInVBlank[k];
+			CalculatePrefetchSchedule_params->DestinationLinesToRequestRowInVBlank = &locals->DestinationLinesToRequestRowInVBlank[k];
+			CalculatePrefetchSchedule_params->VRatioPrefetchY = &locals->VRatioPrefetchY[k];
+			CalculatePrefetchSchedule_params->VRatioPrefetchC = &locals->VRatioPrefetchC[k];
+			CalculatePrefetchSchedule_params->RequiredPrefetchPixDataBWLuma = &locals->RequiredPrefetchPixDataBWLuma[k];
+			CalculatePrefetchSchedule_params->RequiredPrefetchPixDataBWChroma = &locals->RequiredPrefetchPixDataBWChroma[k];
+			CalculatePrefetchSchedule_params->NotEnoughTimeForDynamicMetadata = &locals->NotEnoughTimeForDynamicMetadata[k];
+			CalculatePrefetchSchedule_params->Tno_bw = &locals->Tno_bw[k];
+			CalculatePrefetchSchedule_params->prefetch_vmrow_bw = &locals->prefetch_vmrow_bw[k];
+			CalculatePrefetchSchedule_params->Tdmdl_vm = &locals->Tdmdl_vm[k];
+			CalculatePrefetchSchedule_params->Tdmdl = &locals->Tdmdl[k];
+			CalculatePrefetchSchedule_params->TSetup = &locals->TSetup[k];
+			CalculatePrefetchSchedule_params->VUpdateOffsetPix = &locals->VUpdateOffsetPix[k];
+			CalculatePrefetchSchedule_params->VUpdateWidthPix = &locals->VUpdateWidthPix[k];
+			CalculatePrefetchSchedule_params->VReadyOffsetPix = &locals->VReadyOffsetPix[k];
+
+			locals->NoTimeToPrefetch[k] =
+				CalculatePrefetchSchedule(&mode_lib->scratch,
+					CalculatePrefetchSchedule_params);
+
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: k=%0u NoTimeToPrefetch=%0d\n", __func__, k, locals->NoTimeToPrefetch[k]);
+#endif
+			locals->VStartup[k] = (dml_uint_t)(dml_min(s->VStartupLines, s->MaxVStartupLines[k]));
+			locals->VStartupMin[k] = locals->VStartup[k];
+		}
+
+		for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+			CalculateUrgentBurstFactor(
+				mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange[k],
+				locals->swath_width_luma_ub[k],
+				locals->swath_width_chroma_ub[k],
+				locals->SwathHeightY[k],
+				locals->SwathHeightC[k],
+				mode_lib->ms.cache_display_cfg.timing.HTotal[k] / mode_lib->ms.cache_display_cfg.timing.PixelClock[k],
+				locals->UrgentLatency,
+				mode_lib->ms.ip.cursor_buffer_size,
+				mode_lib->ms.cache_display_cfg.plane.CursorWidth[k],
+				mode_lib->ms.cache_display_cfg.plane.CursorBPP[k],
+				locals->VRatioPrefetchY[k],
+				locals->VRatioPrefetchC[k],
+				locals->BytePerPixelDETY[k],
+				locals->BytePerPixelDETC[k],
+				locals->DETBufferSizeY[k],
+				locals->DETBufferSizeC[k],
+				/* Output */
+				&locals->UrgBurstFactorCursorPre[k],
+				&locals->UrgBurstFactorLumaPre[k],
+				&locals->UrgBurstFactorChromaPre[k],
+				&locals->NoUrgentLatencyHidingPre[k]);
+
+			locals->cursor_bw_pre[k] = mode_lib->ms.cache_display_cfg.plane.NumberOfCursors[k] * mode_lib->ms.cache_display_cfg.plane.CursorWidth[k] * mode_lib->ms.cache_display_cfg.plane.CursorBPP[k] / 8.0 / (mode_lib->ms.cache_display_cfg.timing.HTotal[k] / mode_lib->ms.cache_display_cfg.timing.PixelClock[k]) * locals->VRatioPrefetchY[k];
+
+#ifdef __DML_VBA_DEBUG__
+			dml_print("DML::%s: k=%0u DPPPerSurface=%u\n", __func__, k, mode_lib->ms.cache_display_cfg.hw.DPPPerSurface[k]);
+			dml_print("DML::%s: k=%0u UrgBurstFactorLuma=%f\n", __func__, k, locals->UrgBurstFactorLuma[k]);
+			dml_print("DML::%s: k=%0u UrgBurstFactorChroma=%f\n", __func__, k, locals->UrgBurstFactorChroma[k]);
+			dml_print("DML::%s: k=%0u UrgBurstFactorLumaPre=%f\n", __func__, k, locals->UrgBurstFactorLumaPre[k]);
+			dml_print("DML::%s: k=%0u UrgBurstFactorChromaPre=%f\n", __func__, k, locals->UrgBurstFactorChromaPre[k]);
+
+			dml_print("DML::%s: k=%0u VRatioPrefetchY=%f\n", __func__, k, locals->VRatioPrefetchY[k]);
+			dml_print("DML::%s: k=%0u VRatioY=%f\n", __func__, k, mode_lib->ms.cache_display_cfg.plane.VRatio[k]);
+
+			dml_print("DML::%s: k=%0u prefetch_vmrow_bw=%f\n", __func__, k, locals->prefetch_vmrow_bw[k]);
+			dml_print("DML::%s: k=%0u ReadBandwidthSurfaceLuma=%f\n", __func__, k, locals->ReadBandwidthSurfaceLuma[k]);
+			dml_print("DML::%s: k=%0u ReadBandwidthSurfaceChroma=%f\n", __func__, k, locals->ReadBandwidthSurfaceChroma[k]);
+			dml_print("DML::%s: k=%0u cursor_bw=%f\n", __func__, k, locals->cursor_bw[k]);
+			dml_print("DML::%s: k=%0u meta_row_bw=%f\n", __func__, k, locals->meta_row_bw[k]);
+			dml_print("DML::%s: k=%0u dpte_row_bw=%f\n", __func__, k, locals->dpte_row_bw[k]);
+			dml_print("DML::%s: k=%0u RequiredPrefetchPixDataBWLuma=%f\n", __func__, k, locals->RequiredPrefetchPixDataBWLuma[k]);
+			dml_print("DML::%s: k=%0u RequiredPrefetchPixDataBWChroma=%f\n", __func__, k, locals->RequiredPrefetchPixDataBWChroma[k]);
+			dml_print("DML::%s: k=%0u cursor_bw_pre=%f\n", __func__, k, locals->cursor_bw_pre[k]);
+			dml_print("DML::%s: k=%0u MaxTotalRDBandwidthNoUrgentBurst=%f\n", __func__, k, s->MaxTotalRDBandwidthNoUrgentBurst);
+#endif
+			if (locals->DestinationLinesForPrefetch[k] < 2)
+				s->DestinationLineTimesForPrefetchLessThan2 = true;
+
+			if (locals->VRatioPrefetchY[k] > __DML_MAX_VRATIO_PRE_ENHANCE_PREFETCH_ACC__ ||
+				locals->VRatioPrefetchC[k] > __DML_MAX_VRATIO_PRE_ENHANCE_PREFETCH_ACC__ ||
+				((s->VStartupLines < s->MaxVStartupLines[k] || mode_lib->ms.policy.EnhancedPrefetchScheduleAccelerationFinal == 0) &&
+					(locals->VRatioPrefetchY[k] > __DML_MAX_VRATIO_PRE__ || locals->VRatioPrefetchC[k] > __DML_MAX_VRATIO_PRE__)))
+				s->VRatioPrefetchMoreThanMax = true;
+
+			//dml_bool_t DestinationLinesToRequestVMInVBlankEqualOrMoreThan32 = false;
+			//dml_bool_t DestinationLinesToRequestRowInVBlankEqualOrMoreThan16 = false;
+			//if (locals->DestinationLinesToRequestVMInVBlank[k] >= 32) {
+			//    DestinationLinesToRequestVMInVBlankEqualOrMoreThan32 = true;
+			//}
+
+			//if (locals->DestinationLinesToRequestRowInVBlank[k] >= 16) {
+			//    DestinationLinesToRequestRowInVBlankEqualOrMoreThan16 = true;
+			//}
+		}
+
+		locals->FractionOfUrgentBandwidth = s->MaxTotalRDBandwidthNoUrgentBurst / mode_lib->ms.ReturnBW;
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: MaxTotalRDBandwidthNoUrgentBurst=%f \n", __func__, s->MaxTotalRDBandwidthNoUrgentBurst);
+		dml_print("DML::%s: ReturnBW=%f \n", __func__, mode_lib->ms.ReturnBW);
+		dml_print("DML::%s: FractionOfUrgentBandwidth=%f \n", __func__, locals->FractionOfUrgentBandwidth);
+#endif
+
+		CalculatePrefetchBandwithSupport(
+			mode_lib->ms.num_active_planes,
+			mode_lib->ms.ReturnBW,
+			mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange,
+			locals->NoUrgentLatencyHidingPre,
+			locals->ReadBandwidthSurfaceLuma,
+			locals->ReadBandwidthSurfaceChroma,
+			locals->RequiredPrefetchPixDataBWLuma,
+			locals->RequiredPrefetchPixDataBWChroma,
+			locals->cursor_bw,
+			locals->meta_row_bw,
+			locals->dpte_row_bw,
+			locals->cursor_bw_pre,
+			locals->prefetch_vmrow_bw,
+			mode_lib->ms.cache_display_cfg.hw.DPPPerSurface,
+			locals->UrgBurstFactorLuma,
+			locals->UrgBurstFactorChroma,
+			locals->UrgBurstFactorCursor,
+			locals->UrgBurstFactorLumaPre,
+			locals->UrgBurstFactorChromaPre,
+			locals->UrgBurstFactorCursorPre,
+
+			/* output */
+			&s->MaxTotalRDBandwidth, // dml_float_t *PrefetchBandwidth
+			&s->MaxTotalRDBandwidthNotIncludingMALLPrefetch, // dml_float_t *PrefetchBandwidthNotIncludingMALLPrefetch
+			&s->dummy_single[0], // dml_float_t *FractionOfUrgentBandwidth
+			&locals->PrefetchModeSupported);
+
+		for (k = 0; k < mode_lib->ms.num_active_planes; ++k)
+			s->dummy_unit_vector[k] = 1.0;
+
+		CalculatePrefetchBandwithSupport(mode_lib->ms.num_active_planes,
+			mode_lib->ms.ReturnBW,
+			mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange,
+			locals->NoUrgentLatencyHidingPre,
+			locals->ReadBandwidthSurfaceLuma,
+			locals->ReadBandwidthSurfaceChroma,
+			locals->RequiredPrefetchPixDataBWLuma,
+			locals->RequiredPrefetchPixDataBWChroma,
+			locals->cursor_bw,
+			locals->meta_row_bw,
+			locals->dpte_row_bw,
+			locals->cursor_bw_pre,
+			locals->prefetch_vmrow_bw,
+			mode_lib->ms.cache_display_cfg.hw.DPPPerSurface,
+			s->dummy_unit_vector,
+			s->dummy_unit_vector,
+			s->dummy_unit_vector,
+			s->dummy_unit_vector,
+			s->dummy_unit_vector,
+			s->dummy_unit_vector,
+
+			/* output */
+			&s->NonUrgentMaxTotalRDBandwidth, // dml_float_t *PrefetchBandwidth
+			&s->NonUrgentMaxTotalRDBandwidthNotIncludingMALLPrefetch, // dml_float_t *PrefetchBandwidthNotIncludingMALLPrefetch
+			&locals->FractionOfUrgentBandwidth,
+			&s->dummy_boolean[0]); // dml_bool_t *PrefetchBandwidthSupport
+
+
+
+		if (s->VRatioPrefetchMoreThanMax != false || s->DestinationLineTimesForPrefetchLessThan2 != false) {
+			dml_print("DML::%s: VRatioPrefetchMoreThanMax                   = %u\n", __func__, s->VRatioPrefetchMoreThanMax);
+			dml_print("DML::%s: DestinationLineTimesForPrefetchLessThan2    = %u\n", __func__, s->DestinationLineTimesForPrefetchLessThan2);
+			locals->PrefetchModeSupported = false;
+		}
+
+		for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+			if (locals->NoTimeToPrefetch[k] == true || locals->NotEnoughTimeForDynamicMetadata[k]) {
+				dml_print("DML::%s: k=%u, NoTimeToPrefetch = %0d\n", __func__, k, locals->NoTimeToPrefetch[k]);
+				dml_print("DML::%s: k=%u, NotEnoughTimeForDynamicMetadata=%u\n", __func__, k, locals->NotEnoughTimeForDynamicMetadata[k]);
+				locals->PrefetchModeSupported = false;
+			}
+		}
+
+
+		if (locals->PrefetchModeSupported == true && mode_lib->ms.support.ImmediateFlipSupport == true) {
+			locals->BandwidthAvailableForImmediateFlip = CalculateBandwidthAvailableForImmediateFlip(
+																	mode_lib->ms.num_active_planes,
+																	mode_lib->ms.ReturnBW,
+																	locals->ReadBandwidthSurfaceLuma,
+																	locals->ReadBandwidthSurfaceChroma,
+																	locals->RequiredPrefetchPixDataBWLuma,
+																	locals->RequiredPrefetchPixDataBWChroma,
+																	locals->cursor_bw,
+																	locals->cursor_bw_pre,
+																	mode_lib->ms.cache_display_cfg.hw.DPPPerSurface,
+																	locals->UrgBurstFactorLuma,
+																	locals->UrgBurstFactorChroma,
+																	locals->UrgBurstFactorCursor,
+																	locals->UrgBurstFactorLumaPre,
+																	locals->UrgBurstFactorChromaPre,
+																	locals->UrgBurstFactorCursorPre);
+
+			locals->TotImmediateFlipBytes = 0;
+			for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+				if (mode_lib->ms.policy.ImmediateFlipRequirement[k] != dml_immediate_flip_not_required) {
+					locals->TotImmediateFlipBytes = locals->TotImmediateFlipBytes + mode_lib->ms.cache_display_cfg.hw.DPPPerSurface[k] * (locals->PDEAndMetaPTEBytesFrame[k] + locals->MetaRowByte[k]);
+					if (locals->use_one_row_for_frame_flip[k]) {
+						locals->TotImmediateFlipBytes = locals->TotImmediateFlipBytes + mode_lib->ms.cache_display_cfg.hw.DPPPerSurface[k] * (2 * locals->PixelPTEBytesPerRow[k]);
+					} else {
+						locals->TotImmediateFlipBytes = locals->TotImmediateFlipBytes + mode_lib->ms.cache_display_cfg.hw.DPPPerSurface[k] * locals->PixelPTEBytesPerRow[k];
+					}
+#ifdef __DML_VBA_DEBUG__
+					dml_print("DML::%s: k = %u\n", __func__, k);
+					dml_print("DML::%s: DPPPerSurface = %u\n", __func__, mode_lib->ms.cache_display_cfg.hw.DPPPerSurface[k]);
+					dml_print("DML::%s: PDEAndMetaPTEBytesFrame = %u\n", __func__, locals->PDEAndMetaPTEBytesFrame[k]);
+					dml_print("DML::%s: MetaRowByte = %u\n", __func__, locals->MetaRowByte[k]);
+					dml_print("DML::%s: PixelPTEBytesPerRow = %u\n", __func__, locals->PixelPTEBytesPerRow[k]);
+					dml_print("DML::%s: TotImmediateFlipBytes = %u\n", __func__, locals->TotImmediateFlipBytes);
+#endif
+				}
+			}
+			for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+				CalculateFlipSchedule(
+						s->HostVMInefficiencyFactor,
+						locals->UrgentExtraLatency,
+						locals->UrgentLatency,
+						mode_lib->ms.cache_display_cfg.plane.GPUVMMaxPageTableLevels,
+						mode_lib->ms.cache_display_cfg.plane.HostVMEnable,
+						mode_lib->ms.cache_display_cfg.plane.HostVMMaxPageTableLevels,
+						mode_lib->ms.cache_display_cfg.plane.GPUVMEnable,
+						mode_lib->ms.soc.hostvm_min_page_size_kbytes * 1024,
+						locals->PDEAndMetaPTEBytesFrame[k],
+						locals->MetaRowByte[k],
+						locals->PixelPTEBytesPerRow[k],
+						locals->BandwidthAvailableForImmediateFlip,
+						locals->TotImmediateFlipBytes,
+						mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k],
+						mode_lib->ms.cache_display_cfg.timing.HTotal[k] / mode_lib->ms.cache_display_cfg.timing.PixelClock[k],
+						mode_lib->ms.cache_display_cfg.plane.VRatio[k],
+						mode_lib->ms.cache_display_cfg.plane.VRatioChroma[k],
+						locals->Tno_bw[k],
+						mode_lib->ms.cache_display_cfg.surface.DCCEnable[k],
+						locals->dpte_row_height[k],
+						locals->meta_row_height[k],
+						locals->dpte_row_height_chroma[k],
+						locals->meta_row_height_chroma[k],
+						locals->use_one_row_for_frame_flip[k],
+
+						/* Output */
+						&locals->DestinationLinesToRequestVMInImmediateFlip[k],
+						&locals->DestinationLinesToRequestRowInImmediateFlip[k],
+						&locals->final_flip_bw[k],
+						&locals->ImmediateFlipSupportedForPipe[k]);
+			}
+
+			CalculateImmediateFlipBandwithSupport(mode_lib->ms.num_active_planes,
+												mode_lib->ms.ReturnBW,
+												mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange,
+												mode_lib->ms.policy.ImmediateFlipRequirement,
+												locals->final_flip_bw,
+												locals->ReadBandwidthSurfaceLuma,
+												locals->ReadBandwidthSurfaceChroma,
+												locals->RequiredPrefetchPixDataBWLuma,
+												locals->RequiredPrefetchPixDataBWChroma,
+												locals->cursor_bw,
+												locals->meta_row_bw,
+												locals->dpte_row_bw,
+												locals->cursor_bw_pre,
+												locals->prefetch_vmrow_bw,
+												mode_lib->ms.cache_display_cfg.hw.DPPPerSurface,
+												locals->UrgBurstFactorLuma,
+												locals->UrgBurstFactorChroma,
+												locals->UrgBurstFactorCursor,
+												locals->UrgBurstFactorLumaPre,
+												locals->UrgBurstFactorChromaPre,
+												locals->UrgBurstFactorCursorPre,
+
+												/* output */
+												&locals->total_dcn_read_bw_with_flip, // dml_float_t *TotalBandwidth
+												&locals->total_dcn_read_bw_with_flip_not_including_MALL_prefetch, // dml_float_t TotalBandwidthNotIncludingMALLPrefetch
+												&s->dummy_single[0], // dml_float_t *FractionOfUrgentBandwidth
+												&locals->ImmediateFlipSupported); // dml_bool_t *ImmediateFlipBandwidthSupport
+
+			CalculateImmediateFlipBandwithSupport(mode_lib->ms.num_active_planes,
+				mode_lib->ms.ReturnBW,
+				mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange,
+				mode_lib->ms.policy.ImmediateFlipRequirement,
+				locals->final_flip_bw,
+				locals->ReadBandwidthSurfaceLuma,
+				locals->ReadBandwidthSurfaceChroma,
+				locals->RequiredPrefetchPixDataBWLuma,
+				locals->RequiredPrefetchPixDataBWChroma,
+				locals->cursor_bw,
+				locals->meta_row_bw,
+				locals->dpte_row_bw,
+				locals->cursor_bw_pre,
+				locals->prefetch_vmrow_bw,
+				mode_lib->ms.cache_display_cfg.hw.DPPPerSurface,
+				s->dummy_unit_vector,
+				s->dummy_unit_vector,
+				s->dummy_unit_vector,
+				s->dummy_unit_vector,
+				s->dummy_unit_vector,
+				s->dummy_unit_vector,
+
+				/* output */
+				&locals->non_urgent_total_dcn_read_bw_with_flip, // dml_float_t *TotalBandwidth
+				&locals->non_urgent_total_dcn_read_bw_with_flip_not_including_MALL_prefetch, // dml_float_t TotalBandwidthNotIncludingMALLPrefetch
+				&locals->FractionOfUrgentBandwidthImmediateFlip, // dml_float_t *FractionOfUrgentBandwidth
+				&s->dummy_boolean[0]); // dml_bool_t *ImmediateFlipBandwidthSupport
+
+			for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+				if (mode_lib->ms.policy.ImmediateFlipRequirement[k] != dml_immediate_flip_not_required && locals->ImmediateFlipSupportedForPipe[k] == false) {
+					locals->ImmediateFlipSupported = false;
+#ifdef __DML_VBA_DEBUG__
+					dml_print("DML::%s: Pipe %0d not supporting iflip\n", __func__, k);
+#endif
+				}
+			}
+		} else {
+			locals->ImmediateFlipSupported = false;
+			locals->total_dcn_read_bw_with_flip = s->MaxTotalRDBandwidth;
+			locals->total_dcn_read_bw_with_flip_not_including_MALL_prefetch = s->MaxTotalRDBandwidthNotIncludingMALLPrefetch;
+			locals->non_urgent_total_dcn_read_bw_with_flip = s->NonUrgentMaxTotalRDBandwidth;
+			locals->non_urgent_total_dcn_read_bw_with_flip_not_including_MALL_prefetch = s->NonUrgentMaxTotalRDBandwidthNotIncludingMALLPrefetch;
+		}
+
+		/* consider flip support is okay if the flip bw is ok or (when user does't require a iflip and there is no host vm) */
+		locals->PrefetchAndImmediateFlipSupported = (locals->PrefetchModeSupported == true &&
+													((!mode_lib->ms.support.ImmediateFlipSupport && !mode_lib->ms.cache_display_cfg.plane.HostVMEnable && !s->ImmediateFlipRequirementFinal) ||
+													locals->ImmediateFlipSupported)) ? true : false;
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: PrefetchModeSupported = %u\n", __func__, locals->PrefetchModeSupported);
+		for (k = 0; k < mode_lib->ms.num_active_planes; ++k)
+		dml_print("DML::%s: ImmediateFlipRequirement[%u] = %u\n", __func__, k, mode_lib->ms.policy.ImmediateFlipRequirement[k] == dml_immediate_flip_required);
+		dml_print("DML::%s: HostVMEnable = %u\n", __func__, mode_lib->ms.cache_display_cfg.plane.HostVMEnable);
+		dml_print("DML::%s: ImmediateFlipSupport = %u (from mode_support)\n", __func__, mode_lib->ms.support.ImmediateFlipSupport);
+		dml_print("DML::%s: ImmediateFlipSupported = %u\n", __func__, locals->ImmediateFlipSupported);
+		dml_print("DML::%s: PrefetchAndImmediateFlipSupported = %u\n", __func__, locals->PrefetchAndImmediateFlipSupported);
+#endif
+		dml_print("DML::%s: Done one iteration: VStartupLines=%u, MaxVStartupAllPlanes=%u\n", __func__, s->VStartupLines, s->MaxVStartupAllPlanes);
+
+		s->VStartupLines = s->VStartupLines + 1;
+
+		if (s->VStartupLines > s->MaxVStartupAllPlanes) {
+			s->VStartupLines = __DML_VBA_MIN_VSTARTUP__;
+
+			for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+				s->NextPrefetchMode[k] = s->NextPrefetchMode[k] + 1;
+
+				if (s->NextPrefetchMode[k] <= s->MaxPrefetchMode[k])
+					s->AllPrefetchModeTested = false;
+				dml_print("DML::%s: VStartupLines=%u, reaches max vstartup, try next prefetch mode=%u\n", __func__, s->VStartupLines-1, s->AllPrefetchModeTested);
+			}
+		} else {
+			s->AllPrefetchModeTested = false;
+		}
+		s->iteration++;
+		if (s->iteration > 2500) {
+			dml_print("ERROR: DML::%s: Too many errors, exit now\n", __func__);
+			ASSERT(0);
+		}
+	} while (!(locals->PrefetchAndImmediateFlipSupported || s->AllPrefetchModeTested));
+
+	if (locals->PrefetchAndImmediateFlipSupported) {
+		dml_print("DML::%s: Good, Prefetch and flip scheduling solution found at VStartupLines=%u (MaxVStartupAllPlanes=%u)\n", __func__, s->VStartupLines-1, s->MaxVStartupAllPlanes);
+	} else {
+		dml_print("DML::%s: Bad, Prefetch and flip scheduling solution did NOT find solution! (MaxVStartupAllPlanes=%u)\n", __func__, s->MaxVStartupAllPlanes);
+	}
+
+	//Watermarks and NB P-State/DRAM Clock Change Support
+	{
+		s->mmSOCParameters.UrgentLatency = locals->UrgentLatency;
+		s->mmSOCParameters.ExtraLatency = locals->UrgentExtraLatency;
+		s->mmSOCParameters.WritebackLatency = mode_lib->ms.state.writeback_latency_us;
+		s->mmSOCParameters.DRAMClockChangeLatency = mode_lib->ms.state.dram_clock_change_latency_us;
+		s->mmSOCParameters.FCLKChangeLatency = mode_lib->ms.state.fclk_change_latency_us;
+		s->mmSOCParameters.SRExitTime = mode_lib->ms.state.sr_exit_time_us;
+		s->mmSOCParameters.SREnterPlusExitTime = mode_lib->ms.state.sr_enter_plus_exit_time_us;
+		s->mmSOCParameters.SRExitZ8Time = mode_lib->ms.state.sr_exit_z8_time_us;
+		s->mmSOCParameters.SREnterPlusExitZ8Time = mode_lib->ms.state.sr_enter_plus_exit_z8_time_us;
+		s->mmSOCParameters.USRRetrainingLatency = mode_lib->ms.state.usr_retraining_latency_us;
+		s->mmSOCParameters.SMNLatency = mode_lib->ms.soc.smn_latency_us;
+
+		CalculateWatermarks_params->USRRetrainingRequiredFinal = mode_lib->ms.policy.USRRetrainingRequiredFinal;
+		CalculateWatermarks_params->UseMALLForPStateChange = mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange;
+		CalculateWatermarks_params->PrefetchMode = locals->PrefetchMode;
+		CalculateWatermarks_params->NumberOfActiveSurfaces = mode_lib->ms.num_active_planes;
+		CalculateWatermarks_params->MaxLineBufferLines = mode_lib->ms.ip.max_line_buffer_lines;
+		CalculateWatermarks_params->LineBufferSize = mode_lib->ms.ip.line_buffer_size_bits;
+		CalculateWatermarks_params->WritebackInterfaceBufferSize = mode_lib->ms.ip.writeback_interface_buffer_size_kbytes;
+		CalculateWatermarks_params->DCFCLK = locals->Dcfclk;
+		CalculateWatermarks_params->ReturnBW = mode_lib->ms.ReturnBW;
+		CalculateWatermarks_params->SynchronizeTimingsFinal = mode_lib->ms.policy.SynchronizeTimingsFinal;
+		CalculateWatermarks_params->SynchronizeDRRDisplaysForUCLKPStateChangeFinal = mode_lib->ms.policy.SynchronizeDRRDisplaysForUCLKPStateChangeFinal;
+		CalculateWatermarks_params->DRRDisplay = mode_lib->ms.cache_display_cfg.timing.DRRDisplay;
+		CalculateWatermarks_params->dpte_group_bytes = locals->dpte_group_bytes;
+		CalculateWatermarks_params->meta_row_height = locals->meta_row_height;
+		CalculateWatermarks_params->meta_row_height_chroma = locals->meta_row_height_chroma;
+		CalculateWatermarks_params->mmSOCParameters = s->mmSOCParameters;
+		CalculateWatermarks_params->WritebackChunkSize = mode_lib->ms.ip.writeback_chunk_size_kbytes;
+		CalculateWatermarks_params->SOCCLK = mode_lib->ms.SOCCLK;
+		CalculateWatermarks_params->DCFClkDeepSleep = locals->DCFCLKDeepSleep;
+		CalculateWatermarks_params->DETBufferSizeY = locals->DETBufferSizeY;
+		CalculateWatermarks_params->DETBufferSizeC = locals->DETBufferSizeC;
+		CalculateWatermarks_params->SwathHeightY = locals->SwathHeightY;
+		CalculateWatermarks_params->SwathHeightC = locals->SwathHeightC;
+		CalculateWatermarks_params->LBBitPerPixel = mode_lib->ms.cache_display_cfg.plane.LBBitPerPixel;
+		CalculateWatermarks_params->SwathWidthY = locals->SwathWidthY;
+		CalculateWatermarks_params->SwathWidthC = locals->SwathWidthC;
+		CalculateWatermarks_params->HRatio = mode_lib->ms.cache_display_cfg.plane.HRatio;
+		CalculateWatermarks_params->HRatioChroma = mode_lib->ms.cache_display_cfg.plane.HRatioChroma;
+		CalculateWatermarks_params->VTaps = mode_lib->ms.cache_display_cfg.plane.VTaps;
+		CalculateWatermarks_params->VTapsChroma = mode_lib->ms.cache_display_cfg.plane.VTapsChroma;
+		CalculateWatermarks_params->VRatio = mode_lib->ms.cache_display_cfg.plane.VRatio;
+		CalculateWatermarks_params->VRatioChroma = mode_lib->ms.cache_display_cfg.plane.VRatioChroma;
+		CalculateWatermarks_params->HTotal = mode_lib->ms.cache_display_cfg.timing.HTotal;
+		CalculateWatermarks_params->VTotal = mode_lib->ms.cache_display_cfg.timing.VTotal;
+		CalculateWatermarks_params->VActive = mode_lib->ms.cache_display_cfg.timing.VActive;
+		CalculateWatermarks_params->PixelClock = mode_lib->ms.cache_display_cfg.timing.PixelClock;
+		CalculateWatermarks_params->BlendingAndTiming = mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming;
+		CalculateWatermarks_params->DPPPerSurface = mode_lib->ms.cache_display_cfg.hw.DPPPerSurface;
+		CalculateWatermarks_params->BytePerPixelDETY = locals->BytePerPixelDETY;
+		CalculateWatermarks_params->BytePerPixelDETC = locals->BytePerPixelDETC;
+		CalculateWatermarks_params->DSTXAfterScaler = locals->DSTXAfterScaler;
+		CalculateWatermarks_params->DSTYAfterScaler = locals->DSTYAfterScaler;
+		CalculateWatermarks_params->WritebackEnable = mode_lib->ms.cache_display_cfg.writeback.WritebackEnable;
+		CalculateWatermarks_params->WritebackPixelFormat = mode_lib->ms.cache_display_cfg.writeback.WritebackPixelFormat;
+		CalculateWatermarks_params->WritebackDestinationWidth = mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationWidth;
+		CalculateWatermarks_params->WritebackDestinationHeight = mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationHeight;
+		CalculateWatermarks_params->WritebackSourceHeight = mode_lib->ms.cache_display_cfg.writeback.WritebackSourceHeight;
+		CalculateWatermarks_params->UnboundedRequestEnabled = locals->UnboundedRequestEnabled;
+		CalculateWatermarks_params->CompressedBufferSizeInkByte = locals->CompressedBufferSizeInkByte;
+
+		// Output
+		CalculateWatermarks_params->Watermark = &locals->Watermark; // Watermarks *Watermark
+		CalculateWatermarks_params->DRAMClockChangeSupport = &locals->DRAMClockChangeSupport;
+		CalculateWatermarks_params->MaxActiveDRAMClockChangeLatencySupported = locals->MaxActiveDRAMClockChangeLatencySupported; // dml_float_t *MaxActiveDRAMClockChangeLatencySupported[]
+		CalculateWatermarks_params->SubViewportLinesNeededInMALL = locals->SubViewportLinesNeededInMALL; // dml_uint_t SubViewportLinesNeededInMALL[]
+		CalculateWatermarks_params->FCLKChangeSupport = &locals->FCLKChangeSupport;
+		CalculateWatermarks_params->MaxActiveFCLKChangeLatencySupported = &locals->MaxActiveFCLKChangeLatencySupported; // dml_float_t *MaxActiveFCLKChangeLatencySupported
+		CalculateWatermarks_params->USRRetrainingSupport = &locals->USRRetrainingSupport;
+
+		CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport(
+			&mode_lib->scratch,
+			CalculateWatermarks_params);
+
+		/* Copy the calculated watermarks to mp.Watermark as the getter functions are
+		 * implemented by the DML team to copy the calculated values from the mp.Watermark interface.
+		 * &mode_lib->mp.Watermark and &locals->Watermark are the same address, memcpy may lead to
+		 * unexpected behavior. memmove should be used.
+		 */
+		memmove(&mode_lib->mp.Watermark, CalculateWatermarks_params->Watermark, sizeof(struct Watermarks));
+
+		for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+			if (mode_lib->ms.cache_display_cfg.writeback.WritebackEnable[k] == true) {
+				locals->WritebackAllowDRAMClockChangeEndPosition[k] = dml_max(0, locals->VStartupMin[k] * mode_lib->ms.cache_display_cfg.timing.HTotal[k] /
+																			mode_lib->ms.cache_display_cfg.timing.PixelClock[k] - locals->Watermark.WritebackDRAMClockChangeWatermark);
+				locals->WritebackAllowFCLKChangeEndPosition[k] = dml_max(0, locals->VStartupMin[k] * mode_lib->ms.cache_display_cfg.timing.HTotal[k] /
+																			mode_lib->ms.cache_display_cfg.timing.PixelClock[k] - locals->Watermark.WritebackFCLKChangeWatermark);
+			} else {
+				locals->WritebackAllowDRAMClockChangeEndPosition[k] = 0;
+				locals->WritebackAllowFCLKChangeEndPosition[k] = 0;
+			}
+		}
+	}
+
+	//Display Pipeline Delivery Time in Prefetch, Groups
+	CalculatePixelDeliveryTimes(
+			mode_lib->ms.num_active_planes,
+			mode_lib->ms.cache_display_cfg.plane.VRatio,
+			mode_lib->ms.cache_display_cfg.plane.VRatioChroma,
+			locals->VRatioPrefetchY,
+			locals->VRatioPrefetchC,
+			locals->swath_width_luma_ub,
+			locals->swath_width_chroma_ub,
+			mode_lib->ms.cache_display_cfg.hw.DPPPerSurface,
+			mode_lib->ms.cache_display_cfg.plane.HRatio,
+			mode_lib->ms.cache_display_cfg.plane.HRatioChroma,
+			mode_lib->ms.cache_display_cfg.timing.PixelClock,
+			locals->PSCL_THROUGHPUT,
+			locals->PSCL_THROUGHPUT_CHROMA,
+			locals->Dppclk,
+			locals->BytePerPixelC,
+			mode_lib->ms.cache_display_cfg.plane.SourceScan,
+			mode_lib->ms.cache_display_cfg.plane.NumberOfCursors,
+			mode_lib->ms.cache_display_cfg.plane.CursorWidth,
+			mode_lib->ms.cache_display_cfg.plane.CursorBPP,
+			locals->BlockWidth256BytesY,
+			locals->BlockHeight256BytesY,
+			locals->BlockWidth256BytesC,
+			locals->BlockHeight256BytesC,
+
+			/* Output */
+			locals->DisplayPipeLineDeliveryTimeLuma,
+			locals->DisplayPipeLineDeliveryTimeChroma,
+			locals->DisplayPipeLineDeliveryTimeLumaPrefetch,
+			locals->DisplayPipeLineDeliveryTimeChromaPrefetch,
+			locals->DisplayPipeRequestDeliveryTimeLuma,
+			locals->DisplayPipeRequestDeliveryTimeChroma,
+			locals->DisplayPipeRequestDeliveryTimeLumaPrefetch,
+			locals->DisplayPipeRequestDeliveryTimeChromaPrefetch,
+			locals->CursorRequestDeliveryTime,
+			locals->CursorRequestDeliveryTimePrefetch);
+
+	CalculateMetaAndPTETimes(
+			locals->use_one_row_for_frame,
+			mode_lib->ms.num_active_planes,
+			mode_lib->ms.cache_display_cfg.plane.GPUVMEnable,
+			mode_lib->ms.ip.meta_chunk_size_kbytes,
+			mode_lib->ms.ip.min_meta_chunk_size_bytes,
+			mode_lib->ms.cache_display_cfg.timing.HTotal,
+			mode_lib->ms.cache_display_cfg.plane.VRatio,
+			mode_lib->ms.cache_display_cfg.plane.VRatioChroma,
+			locals->DestinationLinesToRequestRowInVBlank,
+			locals->DestinationLinesToRequestRowInImmediateFlip,
+			mode_lib->ms.cache_display_cfg.surface.DCCEnable,
+			mode_lib->ms.cache_display_cfg.timing.PixelClock,
+			locals->BytePerPixelY,
+			locals->BytePerPixelC,
+			mode_lib->ms.cache_display_cfg.plane.SourceScan,
+			locals->dpte_row_height,
+			locals->dpte_row_height_chroma,
+			locals->meta_row_width,
+			locals->meta_row_width_chroma,
+			locals->meta_row_height,
+			locals->meta_row_height_chroma,
+			locals->meta_req_width,
+			locals->meta_req_width_chroma,
+			locals->meta_req_height,
+			locals->meta_req_height_chroma,
+			locals->dpte_group_bytes,
+			locals->PTERequestSizeY,
+			locals->PTERequestSizeC,
+			locals->PixelPTEReqWidthY,
+			locals->PixelPTEReqHeightY,
+			locals->PixelPTEReqWidthC,
+			locals->PixelPTEReqHeightC,
+			locals->dpte_row_width_luma_ub,
+			locals->dpte_row_width_chroma_ub,
+
+			/* Output */
+			locals->DST_Y_PER_PTE_ROW_NOM_L,
+			locals->DST_Y_PER_PTE_ROW_NOM_C,
+			locals->DST_Y_PER_META_ROW_NOM_L,
+			locals->DST_Y_PER_META_ROW_NOM_C,
+			locals->TimePerMetaChunkNominal,
+			locals->TimePerChromaMetaChunkNominal,
+			locals->TimePerMetaChunkVBlank,
+			locals->TimePerChromaMetaChunkVBlank,
+			locals->TimePerMetaChunkFlip,
+			locals->TimePerChromaMetaChunkFlip,
+			locals->time_per_pte_group_nom_luma,
+			locals->time_per_pte_group_vblank_luma,
+			locals->time_per_pte_group_flip_luma,
+			locals->time_per_pte_group_nom_chroma,
+			locals->time_per_pte_group_vblank_chroma,
+			locals->time_per_pte_group_flip_chroma);
+
+	CalculateVMGroupAndRequestTimes(
+			mode_lib->ms.num_active_planes,
+			mode_lib->ms.cache_display_cfg.plane.GPUVMEnable,
+			mode_lib->ms.cache_display_cfg.plane.GPUVMMaxPageTableLevels,
+			mode_lib->ms.cache_display_cfg.timing.HTotal,
+			locals->BytePerPixelC,
+			locals->DestinationLinesToRequestVMInVBlank,
+			locals->DestinationLinesToRequestVMInImmediateFlip,
+			mode_lib->ms.cache_display_cfg.surface.DCCEnable,
+			mode_lib->ms.cache_display_cfg.timing.PixelClock,
+			locals->dpte_row_width_luma_ub,
+			locals->dpte_row_width_chroma_ub,
+			locals->vm_group_bytes,
+			locals->dpde0_bytes_per_frame_ub_l,
+			locals->dpde0_bytes_per_frame_ub_c,
+			locals->meta_pte_bytes_per_frame_ub_l,
+			locals->meta_pte_bytes_per_frame_ub_c,
+
+			/* Output */
+			locals->TimePerVMGroupVBlank,
+			locals->TimePerVMGroupFlip,
+			locals->TimePerVMRequestVBlank,
+			locals->TimePerVMRequestFlip);
+
+	// Min TTUVBlank
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		if (locals->PrefetchMode[k] == 0) {
+			locals->MinTTUVBlank[k] = dml_max4(
+					locals->Watermark.DRAMClockChangeWatermark,
+					locals->Watermark.FCLKChangeWatermark,
+					locals->Watermark.StutterEnterPlusExitWatermark,
+					locals->Watermark.UrgentWatermark);
+		} else if (locals->PrefetchMode[k] == 1) {
+			locals->MinTTUVBlank[k] = dml_max3(
+					locals->Watermark.FCLKChangeWatermark,
+					locals->Watermark.StutterEnterPlusExitWatermark,
+					locals->Watermark.UrgentWatermark);
+		} else if (locals->PrefetchMode[k] == 2) {
+			locals->MinTTUVBlank[k] = dml_max(
+					locals->Watermark.StutterEnterPlusExitWatermark,
+					locals->Watermark.UrgentWatermark);
+		} else {
+			locals->MinTTUVBlank[k] = locals->Watermark.UrgentWatermark;
+		}
+		if (!mode_lib->ms.cache_display_cfg.plane.DynamicMetadataEnable[k])
+			locals->MinTTUVBlank[k] = locals->TCalc + locals->MinTTUVBlank[k];
+	}
+
+	// DCC Configuration
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: Calculate DCC configuration for surface k=%u\n", __func__, k);
+#endif
+		CalculateDCCConfiguration(
+			mode_lib->ms.cache_display_cfg.surface.DCCEnable[k],
+			mode_lib->ms.policy.DCCProgrammingAssumesScanDirectionUnknownFinal,
+			mode_lib->ms.cache_display_cfg.surface.SourcePixelFormat[k],
+			mode_lib->ms.cache_display_cfg.surface.SurfaceWidthY[k],
+			mode_lib->ms.cache_display_cfg.surface.SurfaceWidthC[k],
+			mode_lib->ms.cache_display_cfg.surface.SurfaceHeightY[k],
+			mode_lib->ms.cache_display_cfg.surface.SurfaceHeightC[k],
+			mode_lib->ms.NomDETInKByte,
+			locals->BlockHeight256BytesY[k],
+			locals->BlockHeight256BytesC[k],
+			mode_lib->ms.cache_display_cfg.surface.SurfaceTiling[k],
+			locals->BytePerPixelY[k],
+			locals->BytePerPixelC[k],
+			locals->BytePerPixelDETY[k],
+			locals->BytePerPixelDETC[k],
+			mode_lib->ms.cache_display_cfg.plane.SourceScan[k],
+			/* Output */
+			&locals->DCCYMaxUncompressedBlock[k],
+			&locals->DCCCMaxUncompressedBlock[k],
+			&locals->DCCYMaxCompressedBlock[k],
+			&locals->DCCCMaxCompressedBlock[k],
+			&locals->DCCYIndependentBlock[k],
+			&locals->DCCCIndependentBlock[k]);
+	}
+
+	// VStartup Adjustment
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		s->Tvstartup_margin = (s->MaxVStartupLines[k] - locals->VStartupMin[k]) * mode_lib->ms.cache_display_cfg.timing.HTotal[k] / mode_lib->ms.cache_display_cfg.timing.PixelClock[k];
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%u, MinTTUVBlank = %f (before vstartup margin)\n", __func__, k, locals->MinTTUVBlank[k]);
+#endif
+
+		locals->MinTTUVBlank[k] = locals->MinTTUVBlank[k] + s->Tvstartup_margin;
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%u, Tvstartup_margin = %f\n", __func__, k, s->Tvstartup_margin);
+		dml_print("DML::%s: k=%u, MaxVStartupLines = %u\n", __func__, k, s->MaxVStartupLines[k]);
+		dml_print("DML::%s: k=%u, MinTTUVBlank = %f\n", __func__, k, locals->MinTTUVBlank[k]);
+#endif
+
+		locals->Tdmdl[k] = locals->Tdmdl[k] + s->Tvstartup_margin;
+		if (mode_lib->ms.cache_display_cfg.plane.DynamicMetadataEnable[k] && mode_lib->ms.ip.dynamic_metadata_vm_enabled) {
+			locals->Tdmdl_vm[k] = locals->Tdmdl_vm[k] + s->Tvstartup_margin;
+		}
+
+		isInterlaceTiming        = (mode_lib->ms.cache_display_cfg.timing.Interlace[k] && !mode_lib->ms.ip.ptoi_supported);
+
+		// The actual positioning of the vstartup
+		locals->VStartup[k] = (isInterlaceTiming ? (2 * s->MaxVStartupLines[k]) : s->MaxVStartupLines[k]);
+
+		s->dlg_vblank_start        =  ((isInterlaceTiming ? dml_floor((mode_lib->ms.cache_display_cfg.timing.VTotal[k] - mode_lib->ms.cache_display_cfg.timing.VFrontPorch[k]) / 2.0, 1.0) :
+																	mode_lib->ms.cache_display_cfg.timing.VTotal[k]) - mode_lib->ms.cache_display_cfg.timing.VFrontPorch[k]);
+		s->LSetup                  = dml_floor(4.0 * locals->TSetup[k] / ((dml_float_t) mode_lib->ms.cache_display_cfg.timing.HTotal[k] / mode_lib->ms.cache_display_cfg.timing.PixelClock[k]), 1.0) / 4.0;
+		s->blank_lines_remaining   = (mode_lib->ms.cache_display_cfg.timing.VTotal[k] - mode_lib->ms.cache_display_cfg.timing.VActive[k]) - locals->VStartup[k];
+
+		if (s->blank_lines_remaining < 0) {
+			dml_print("ERROR: Vstartup is larger than vblank!?\n");
+			s->blank_lines_remaining = 0;
+			ASSERT(0);
+		}
+		locals->MIN_DST_Y_NEXT_START[k] = s->dlg_vblank_start + s->blank_lines_remaining + s->LSetup;
+
+		// debug only
+		s->old_MIN_DST_Y_NEXT_START = ((isInterlaceTiming ? dml_floor((mode_lib->ms.cache_display_cfg.timing.VTotal[k] - mode_lib->ms.cache_display_cfg.timing.VFrontPorch[k]) / 2.0, 1.0) :
+																	mode_lib->ms.cache_display_cfg.timing.VTotal[k]) - mode_lib->ms.cache_display_cfg.timing.VFrontPorch[k])
+											+ dml_max(1.0, dml_ceil((dml_float_t) locals->WritebackDelay[k] / ((dml_float_t) mode_lib->ms.cache_display_cfg.timing.HTotal[k] / mode_lib->ms.cache_display_cfg.timing.PixelClock[k]), 1.0))
+											+ dml_floor(4.0 * locals->TSetup[k] / ((dml_float_t) mode_lib->ms.cache_display_cfg.timing.HTotal[k] / mode_lib->ms.cache_display_cfg.timing.PixelClock[k]), 1.0) / 4.0;
+
+		if (((locals->VUpdateOffsetPix[k] + locals->VUpdateWidthPix[k] + locals->VReadyOffsetPix[k]) / (double) mode_lib->ms.cache_display_cfg.timing.HTotal[k]) <=
+			(isInterlaceTiming ?
+				dml_floor((mode_lib->ms.cache_display_cfg.timing.VTotal[k] - mode_lib->ms.cache_display_cfg.timing.VActive[k] - mode_lib->ms.cache_display_cfg.timing.VFrontPorch[k] - locals->VStartup[k]) / 2.0, 1.0) :
+				(int) (mode_lib->ms.cache_display_cfg.timing.VTotal[k] - mode_lib->ms.cache_display_cfg.timing.VActive[k] - mode_lib->ms.cache_display_cfg.timing.VFrontPorch[k] - locals->VStartup[k]))) {
+			locals->VREADY_AT_OR_AFTER_VSYNC[k] = true;
+		} else {
+			locals->VREADY_AT_OR_AFTER_VSYNC[k] = false;
+		}
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%u, VStartup = %u (max)\n", __func__, k, locals->VStartup[k]);
+		dml_print("DML::%s: k=%u, VStartupMin = %u (max)\n", __func__, k, locals->VStartupMin[k]);
+		dml_print("DML::%s: k=%u, VUpdateOffsetPix = %u\n", __func__, k, locals->VUpdateOffsetPix[k]);
+		dml_print("DML::%s: k=%u, VUpdateWidthPix = %u\n", __func__, k, locals->VUpdateWidthPix[k]);
+		dml_print("DML::%s: k=%u, VReadyOffsetPix = %u\n", __func__, k, locals->VReadyOffsetPix[k]);
+		dml_print("DML::%s: k=%u, HTotal = %u\n", __func__, k, mode_lib->ms.cache_display_cfg.timing.HTotal[k]);
+		dml_print("DML::%s: k=%u, VTotal = %u\n", __func__, k, mode_lib->ms.cache_display_cfg.timing.VTotal[k]);
+		dml_print("DML::%s: k=%u, VActive = %u\n", __func__, k, mode_lib->ms.cache_display_cfg.timing.VActive[k]);
+		dml_print("DML::%s: k=%u, VFrontPorch = %u\n", __func__, k, mode_lib->ms.cache_display_cfg.timing.VFrontPorch[k]);
+		dml_print("DML::%s: k=%u, TSetup = %f\n", __func__, k, locals->TSetup[k]);
+		dml_print("DML::%s: k=%u, MIN_DST_Y_NEXT_START = %f\n", __func__, k, locals->MIN_DST_Y_NEXT_START[k]);
+		dml_print("DML::%s: k=%u, MIN_DST_Y_NEXT_START = %f (old)\n", __func__, k, s->old_MIN_DST_Y_NEXT_START);
+		dml_print("DML::%s: k=%u, VREADY_AT_OR_AFTER_VSYNC = %u\n", __func__, k, locals->VREADY_AT_OR_AFTER_VSYNC[k]);
+#endif
+	}
+
+	//Maximum Bandwidth Used
+	s->TotalWRBandwidth = 0;
+	s->WRBandwidth = 0;
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		if (mode_lib->ms.cache_display_cfg.writeback.WritebackEnable[k] == true && mode_lib->ms.cache_display_cfg.writeback.WritebackPixelFormat[k] == dml_444_32) {
+			s->WRBandwidth = mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationWidth[k] * mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationHeight[k] /
+							(mode_lib->ms.cache_display_cfg.timing.HTotal[k] * mode_lib->ms.cache_display_cfg.writeback.WritebackSourceHeight[k] / mode_lib->ms.cache_display_cfg.timing.PixelClock[k]) * 4;
+		} else if (mode_lib->ms.cache_display_cfg.writeback.WritebackEnable[k] == true) {
+			s->WRBandwidth = mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationWidth[k] * mode_lib->ms.cache_display_cfg.writeback.WritebackDestinationHeight[k] /
+							(mode_lib->ms.cache_display_cfg.timing.HTotal[k] * mode_lib->ms.cache_display_cfg.writeback.WritebackSourceHeight[k] / mode_lib->ms.cache_display_cfg.timing.PixelClock[k]) * 8;
+		}
+		s->TotalWRBandwidth = s->TotalWRBandwidth + s->WRBandwidth;
+	}
+
+	locals->TotalDataReadBandwidth = 0;
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		locals->TotalDataReadBandwidth = locals->TotalDataReadBandwidth + locals->ReadBandwidthSurfaceLuma[k] + locals->ReadBandwidthSurfaceChroma[k];
+
+#ifdef __DML_VBA_DEBUG__
+		dml_print("DML::%s: k=%u, TotalDataReadBandwidth = %f\n", __func__, k, locals->TotalDataReadBandwidth);
+		dml_print("DML::%s: k=%u, ReadBandwidthSurfaceLuma = %f\n", __func__, k, locals->ReadBandwidthSurfaceLuma[k]);
+		dml_print("DML::%s: k=%u, ReadBandwidthSurfaceChroma = %f\n", __func__, k, locals->ReadBandwidthSurfaceChroma[k]);
+#endif
+	}
+
+	locals->TotalDataReadBandwidthNotIncludingMALLPrefetch = 0;
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		if (mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange[k] != dml_use_mall_pstate_change_phantom_pipe) {
+			locals->TotalDataReadBandwidthNotIncludingMALLPrefetch = locals->TotalDataReadBandwidthNotIncludingMALLPrefetch
+		+ locals->ReadBandwidthSurfaceLuma[k] + locals->ReadBandwidthSurfaceChroma[k];
+		}
+	}
+
+	CalculateStutterEfficiency_params->CompressedBufferSizeInkByte = locals->CompressedBufferSizeInkByte;
+	CalculateStutterEfficiency_params->UseMALLForPStateChange = mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange;
+	CalculateStutterEfficiency_params->UnboundedRequestEnabled = locals->UnboundedRequestEnabled;
+	CalculateStutterEfficiency_params->MetaFIFOSizeInKEntries = mode_lib->ms.ip.meta_fifo_size_in_kentries;
+	CalculateStutterEfficiency_params->ZeroSizeBufferEntries = mode_lib->ms.ip.zero_size_buffer_entries;
+	CalculateStutterEfficiency_params->PixelChunkSizeInKByte = mode_lib->ms.ip.pixel_chunk_size_kbytes;
+	CalculateStutterEfficiency_params->NumberOfActiveSurfaces = mode_lib->ms.num_active_planes;
+	CalculateStutterEfficiency_params->ROBBufferSizeInKByte = mode_lib->ms.ip.rob_buffer_size_kbytes;
+	CalculateStutterEfficiency_params->TotalDataReadBandwidth = locals->TotalDataReadBandwidth;
+	CalculateStutterEfficiency_params->DCFCLK = locals->Dcfclk;
+	CalculateStutterEfficiency_params->ReturnBW = mode_lib->ms.ReturnBW;
+	CalculateStutterEfficiency_params->CompbufReservedSpace64B = locals->compbuf_reserved_space_64b;
+	CalculateStutterEfficiency_params->CompbufReservedSpaceZs = locals->compbuf_reserved_space_zs;
+	CalculateStutterEfficiency_params->SRExitTime = mode_lib->ms.state.sr_exit_time_us;
+	CalculateStutterEfficiency_params->SRExitZ8Time = mode_lib->ms.state.sr_exit_z8_time_us;
+	CalculateStutterEfficiency_params->SynchronizeTimingsFinal = mode_lib->ms.policy.SynchronizeTimingsFinal;
+	CalculateStutterEfficiency_params->BlendingAndTiming = mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming;
+	CalculateStutterEfficiency_params->StutterEnterPlusExitWatermark = locals->Watermark.StutterEnterPlusExitWatermark;
+	CalculateStutterEfficiency_params->Z8StutterEnterPlusExitWatermark = locals->Watermark.Z8StutterEnterPlusExitWatermark;
+	CalculateStutterEfficiency_params->ProgressiveToInterlaceUnitInOPP = mode_lib->ms.ip.ptoi_supported;
+	CalculateStutterEfficiency_params->Interlace = mode_lib->ms.cache_display_cfg.timing.Interlace;
+	CalculateStutterEfficiency_params->MinTTUVBlank = locals->MinTTUVBlank;
+	CalculateStutterEfficiency_params->DPPPerSurface = mode_lib->ms.cache_display_cfg.hw.DPPPerSurface;
+	CalculateStutterEfficiency_params->DETBufferSizeY = locals->DETBufferSizeY;
+	CalculateStutterEfficiency_params->BytePerPixelY = locals->BytePerPixelY;
+	CalculateStutterEfficiency_params->BytePerPixelDETY = locals->BytePerPixelDETY;
+	CalculateStutterEfficiency_params->SwathWidthY = locals->SwathWidthY;
+	CalculateStutterEfficiency_params->SwathHeightY = locals->SwathHeightY;
+	CalculateStutterEfficiency_params->SwathHeightC = locals->SwathHeightC;
+	CalculateStutterEfficiency_params->NetDCCRateLuma = mode_lib->ms.cache_display_cfg.surface.DCCRateLuma;
+	CalculateStutterEfficiency_params->NetDCCRateChroma = mode_lib->ms.cache_display_cfg.surface.DCCRateChroma;
+	CalculateStutterEfficiency_params->DCCFractionOfZeroSizeRequestsLuma = mode_lib->ms.cache_display_cfg.surface.DCCFractionOfZeroSizeRequestsLuma;
+	CalculateStutterEfficiency_params->DCCFractionOfZeroSizeRequestsChroma = mode_lib->ms.cache_display_cfg.surface.DCCFractionOfZeroSizeRequestsChroma;
+	CalculateStutterEfficiency_params->HTotal = mode_lib->ms.cache_display_cfg.timing.HTotal;
+	CalculateStutterEfficiency_params->VTotal = mode_lib->ms.cache_display_cfg.timing.VTotal;
+	CalculateStutterEfficiency_params->PixelClock = mode_lib->ms.cache_display_cfg.timing.PixelClock;
+	CalculateStutterEfficiency_params->VRatio = mode_lib->ms.cache_display_cfg.plane.VRatio;
+	CalculateStutterEfficiency_params->SourceScan = mode_lib->ms.cache_display_cfg.plane.SourceScan;
+	CalculateStutterEfficiency_params->BlockHeight256BytesY = locals->BlockHeight256BytesY;
+	CalculateStutterEfficiency_params->BlockWidth256BytesY = locals->BlockWidth256BytesY;
+	CalculateStutterEfficiency_params->BlockHeight256BytesC = locals->BlockHeight256BytesC;
+	CalculateStutterEfficiency_params->BlockWidth256BytesC = locals->BlockWidth256BytesC;
+	CalculateStutterEfficiency_params->DCCYMaxUncompressedBlock = locals->DCCYMaxUncompressedBlock;
+	CalculateStutterEfficiency_params->DCCCMaxUncompressedBlock = locals->DCCCMaxUncompressedBlock;
+	CalculateStutterEfficiency_params->VActive = mode_lib->ms.cache_display_cfg.timing.VActive;
+	CalculateStutterEfficiency_params->DCCEnable = mode_lib->ms.cache_display_cfg.surface.DCCEnable;
+	CalculateStutterEfficiency_params->WritebackEnable = mode_lib->ms.cache_display_cfg.writeback.WritebackEnable;
+	CalculateStutterEfficiency_params->ReadBandwidthSurfaceLuma = locals->ReadBandwidthSurfaceLuma;
+	CalculateStutterEfficiency_params->ReadBandwidthSurfaceChroma = locals->ReadBandwidthSurfaceChroma;
+	CalculateStutterEfficiency_params->meta_row_bw = locals->meta_row_bw;
+	CalculateStutterEfficiency_params->dpte_row_bw = locals->dpte_row_bw;
+	CalculateStutterEfficiency_params->StutterEfficiencyNotIncludingVBlank = &locals->StutterEfficiencyNotIncludingVBlank;
+	CalculateStutterEfficiency_params->StutterEfficiency = &locals->StutterEfficiency;
+	CalculateStutterEfficiency_params->NumberOfStutterBurstsPerFrame = &locals->NumberOfStutterBurstsPerFrame;
+	CalculateStutterEfficiency_params->Z8StutterEfficiencyNotIncludingVBlank = &locals->Z8StutterEfficiencyNotIncludingVBlank;
+	CalculateStutterEfficiency_params->Z8StutterEfficiency = &locals->Z8StutterEfficiency;
+	CalculateStutterEfficiency_params->Z8NumberOfStutterBurstsPerFrame = &locals->Z8NumberOfStutterBurstsPerFrame;
+	CalculateStutterEfficiency_params->StutterPeriod = &locals->StutterPeriod;
+	CalculateStutterEfficiency_params->DCHUBBUB_ARB_CSTATE_MAX_CAP_MODE = &locals->DCHUBBUB_ARB_CSTATE_MAX_CAP_MODE;
+
+	// Stutter Efficiency
+	CalculateStutterEfficiency(&mode_lib->scratch,
+		CalculateStutterEfficiency_params);
+
+#ifdef __DML_VBA_ALLOW_DELTA__
+	{
+	dml_float_t dummy_single[2];
+	dml_uint_t dummy_integer[1];
+	dml_bool_t dummy_boolean[1];
+
+	// Calculate z8 stutter eff assuming 0 reserved space
+	CalculateStutterEfficiency(
+			locals->CompressedBufferSizeInkByte,
+			mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange,
+			locals->UnboundedRequestEnabled,
+			mode_lib->ms.ip.meta_fifo_size_in_kentries,
+			mode_lib->ms.ip.zero_size_buffer_entries,
+			mode_lib->ms.ip.pixel_chunk_size_kbytes,
+			mode_lib->ms.num_active_planes,
+			mode_lib->ms.ip.rob_buffer_size_kbytes,
+			locals->TotalDataReadBandwidth,
+			locals->Dcfclk,
+			mode_lib->ms.ReturnBW,
+			0, //mode_lib->ms.ip.compbuf_reserved_space_64b,
+			0, //mode_lib->ms.ip.compbuf_reserved_space_zs,
+			mode_lib->ms.state.sr_exit_time_us,
+			mode_lib->ms.state.sr_exit_z8_time_us,
+			mode_lib->ms.policy.SynchronizeTimingsFinal,
+			mode_lib->ms.cache_display_cfg.plane.BlendingAndTiming,
+			locals->Watermark.StutterEnterPlusExitWatermark,
+			locals->Watermark.Z8StutterEnterPlusExitWatermark,
+			mode_lib->ms.ip.ptoi_supported,
+			mode_lib->ms.cache_display_cfg.timing.Interlace,
+			locals->MinTTUVBlank,
+			mode_lib->ms.cache_display_cfg.hw.DPPPerSurface,
+			mode_lib->ms.DETBufferSizeY,
+			locals->BytePerPixelY,
+			locals->BytePerPixelDETY,
+			locals->SwathWidthY,
+			mode_lib->ms.SwathHeightY,
+			mode_lib->ms.SwathHeightC,
+			mode_lib->ms.cache_display_cfg.surface.DCCRateLuma,
+			mode_lib->ms.cache_display_cfg.surface.DCCRateChroma,
+			mode_lib->ms.cache_display_cfg.surface.DCCFractionOfZeroSizeRequestsLuma,
+			mode_lib->ms.cache_display_cfg.surface.DCCFractionOfZeroSizeRequestsChroma,
+			mode_lib->ms.cache_display_cfg.timing.HTotal,
+			mode_lib->ms.cache_display_cfg.timing.VTotal,
+			mode_lib->ms.cache_display_cfg.timing.PixelClock,
+			mode_lib->ms.cache_display_cfg.plane.VRatio,
+			mode_lib->ms.cache_display_cfg.plane.SourceScan,
+			locals->BlockHeight256BytesY,
+			locals->BlockWidth256BytesY,
+			locals->BlockHeight256BytesC,
+			locals->BlockWidth256BytesC,
+			locals->DCCYMaxUncompressedBlock,
+			locals->DCCCMaxUncompressedBlock,
+			mode_lib->ms.cache_display_cfg.timing.VActive,
+			mode_lib->ms.cache_display_cfg.surface.DCCEnable,
+			mode_lib->ms.cache_display_cfg.writeback.WritebackEnable,
+			locals->ReadBandwidthSurfaceLuma,
+			locals->ReadBandwidthSurfaceChroma,
+			locals->meta_row_bw,
+			locals->dpte_row_bw,
+
+			/* Output */
+			&dummy_single[0],
+			&dummy_single[1],
+			&dummy_integer[0],
+			&locals->Z8StutterEfficiencyNotIncludingVBlankBestCase,
+			&locals->Z8StutterEfficiencyBestCase,
+			&locals->Z8NumberOfStutterBurstsPerFrameBestCase,
+			&locals->StutterPeriodBestCase,
+			&dummy_boolean[0]);
+	}
+#else
+	locals->Z8StutterEfficiencyNotIncludingVBlankBestCase = locals->Z8StutterEfficiencyNotIncludingVBlank;
+	locals->Z8StutterEfficiencyBestCase                   = locals->Z8StutterEfficiency;
+	locals->Z8NumberOfStutterBurstsPerFrameBestCase       = locals->Z8NumberOfStutterBurstsPerFrame;
+	locals->StutterPeriodBestCase                         = locals->StutterPeriod;
+#endif
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: --- END --- \n",  __func__);
+#endif
+} // dml_core_mode_programming
+
+/// Function: dml_core_get_row_heights
+/// @brief Get row height for DPTE and META with minimal input.
+void dml_core_get_row_heights(
+						dml_uint_t                         *dpte_row_height,
+						dml_uint_t                         *meta_row_height,
+						const struct display_mode_lib_st   *mode_lib,
+						dml_bool_t                         is_plane1,
+						enum dml_source_format_class        SourcePixelFormat,
+						enum dml_swizzle_mode            SurfaceTiling,
+						enum dml_rotation_angle          ScanDirection,
+						dml_uint_t                         pitch,
+						dml_uint_t                         GPUVMMinPageSizeKBytes)
+{
+	dml_uint_t BytePerPixelY;
+	dml_uint_t BytePerPixelC;
+	dml_float_t BytePerPixelInDETY;
+	dml_float_t BytePerPixelInDETC;
+	dml_uint_t BlockHeight256BytesY;
+	dml_uint_t BlockHeight256BytesC;
+	dml_uint_t BlockWidth256BytesY;
+	dml_uint_t BlockWidth256BytesC;
+	dml_uint_t MacroTileWidthY;
+	dml_uint_t MacroTileWidthC;
+	dml_uint_t MacroTileHeightY;
+	dml_uint_t MacroTileHeightC;
+
+	dml_uint_t BytePerPixel;
+	dml_uint_t BlockHeight256Bytes;
+	dml_uint_t BlockWidth256Bytes;
+	dml_uint_t MacroTileWidth;
+	dml_uint_t MacroTileHeight;
+	dml_uint_t PTEBufferSizeInRequests;
+
+	dml_uint_t dummy_integer[16];
+
+	CalculateBytePerPixelAndBlockSizes(
+		SourcePixelFormat,
+		SurfaceTiling,
+
+		/* Output */
+		&BytePerPixelY,
+		&BytePerPixelC,
+		&BytePerPixelInDETY,
+		&BytePerPixelInDETC,
+		&BlockHeight256BytesY,
+		&BlockHeight256BytesC,
+		&BlockWidth256BytesY,
+		&BlockWidth256BytesC,
+		&MacroTileHeightY,
+		&MacroTileHeightC,
+		&MacroTileWidthY,
+		&MacroTileWidthC);
+
+	BytePerPixel = is_plane1 ? BytePerPixelC : BytePerPixelY;
+	BlockHeight256Bytes = is_plane1 ? BlockHeight256BytesC : BlockHeight256BytesY;
+	BlockWidth256Bytes = is_plane1 ? BlockWidth256BytesC : BlockWidth256BytesY;
+	MacroTileWidth = is_plane1 ? MacroTileWidthC : MacroTileWidthY;
+	MacroTileHeight = is_plane1 ? MacroTileHeightC : MacroTileHeightY;
+	PTEBufferSizeInRequests = is_plane1 ? mode_lib->ip.dpte_buffer_size_in_pte_reqs_chroma : mode_lib->ip.dpte_buffer_size_in_pte_reqs_luma;
+#ifdef __DML_RQ_DLG_CALC_DEBUG__
+	dml_print("DML_DLG: %s: is_plane1 = %u\n", __func__, is_plane1);
+	dml_print("DML_DLG: %s: BytePerPixel = %u\n", __func__, BytePerPixel);
+	dml_print("DML_DLG: %s: BlockHeight256Bytes = %u\n", __func__, BlockHeight256Bytes);
+	dml_print("DML_DLG: %s: BlockWidth256Bytes = %u\n", __func__, BlockWidth256Bytes);
+	dml_print("DML_DLG: %s: MacroTileWidth = %u\n", __func__, MacroTileWidth);
+	dml_print("DML_DLG: %s: MacroTileHeight = %u\n", __func__, MacroTileHeight);
+	dml_print("DML_DLG: %s: PTEBufferSizeInRequests = %u\n", __func__, PTEBufferSizeInRequests);
+	dml_print("DML_DLG: %s: dpte_buffer_size_in_pte_reqs_luma = %u\n", __func__, mode_lib->ip.dpte_buffer_size_in_pte_reqs_luma);
+	dml_print("DML_DLG: %s: dpte_buffer_size_in_pte_reqs_chroma = %u\n", __func__, mode_lib->ip.dpte_buffer_size_in_pte_reqs_chroma);
+	dml_print("DML_DLG: %s: GPUVMMinPageSizeKBytes = %u\n", __func__, GPUVMMinPageSizeKBytes);
+#endif
+
+	// just supply with enough parameters to calculate meta and dte
+	CalculateVMAndRowBytes(
+			0, // dml_bool_t ViewportStationary,
+			1, // dml_bool_t DCCEnable,
+			1, // dml_uint_t NumberOfDPPs,
+			BlockHeight256Bytes,
+			BlockWidth256Bytes,
+			SourcePixelFormat,
+			SurfaceTiling,
+			BytePerPixel,
+			ScanDirection,
+			0, // dml_uint_t SwathWidth,
+			0, // dml_uint_t ViewportHeight, (Note: DML calculates one_row_for_frame height regardless, would need test input if that height is useful)
+			0, // dml_uint_t ViewportXStart,
+			0, // dml_uint_t ViewportYStart,
+			1, // dml_bool_t GPUVMEnable,
+			4, // dml_uint_t GPUVMMaxPageTableLevels,
+			GPUVMMinPageSizeKBytes,
+			PTEBufferSizeInRequests,
+			pitch,
+			0, // dml_uint_t DCCMetaPitch,
+			MacroTileWidth,
+			MacroTileHeight,
+
+			// /* Output */
+			&dummy_integer[0], // dml_uint_t *MetaRowByte,
+			&dummy_integer[1], // dml_uint_t *PixelPTEBytesPerRow,
+			&dummy_integer[2], // dml_uint_t *PixelPTEBytesPerRowStorage,
+			&dummy_integer[3], // dml_uint_t *dpte_row_width_ub,
+			dpte_row_height,
+			&dummy_integer[4], // dml_uint_t *dpte_row_height_linear
+			&dummy_integer[5], // dml_uint_t *PixelPTEBytesPerRow_one_row_per_frame,
+			&dummy_integer[6], // dml_uint_t *dpte_row_width_ub_one_row_per_frame,
+			&dummy_integer[7], // dml_uint_t *dpte_row_height_one_row_per_frame,
+			&dummy_integer[8], // dml_uint_t *MetaRequestWidth,
+			&dummy_integer[9], // dml_uint_t *MetaRequestHeight,
+			&dummy_integer[10], // dml_uint_t *meta_row_width,
+			meta_row_height,
+			&dummy_integer[11], // dml_uint_t *PixelPTEReqWidth,
+			&dummy_integer[12], // dml_uint_t *PixelPTEReqHeight,
+			&dummy_integer[13], // dml_uint_t *PTERequestSize,
+			&dummy_integer[14], // dml_uint_t *DPDE0BytesFrame,
+			&dummy_integer[15]); // dml_uint_t *MetaPTEBytesFrame)
+
+#ifdef __DML_RQ_DLG_CALC_DEBUG__
+	dml_print("DML_DLG: %s: dpte_row_height = %u\n", __func__, *dpte_row_height);
+	dml_print("DML_DLG: %s: meta_row_height = %u\n", __func__, *meta_row_height);
+#endif
+}
+
+static struct soc_state_bounding_box_st dml_get_soc_state_bounding_box(
+	const struct soc_states_st *states,
+	dml_uint_t state_idx)
+{
+	dml_print("DML::%s: state_idx=%u (num_states=%u)\n", __func__, state_idx, states->num_states);
+
+	if (state_idx >= (dml_uint_t)states->num_states) {
+		dml_print("DML::%s: ERROR: Invalid state_idx=%u! num_states=%u\n", __func__, state_idx, states->num_states);
+		ASSERT(0);
+	}
+	return (states->state_array[state_idx]);
+}
+
+/// @brief Copy the parameters to a calculation struct, it actually only need when the DML needs to have
+///        the intelligence to re-calculate when any of display cfg, bbox, or policy changes since last calculated.
+///
+static void cache_ip_soc_cfg(struct display_mode_lib_st *mode_lib,
+						dml_uint_t state_idx)
+{
+	mode_lib->ms.state_idx = state_idx;
+	mode_lib->ms.max_state_idx = mode_lib->states.num_states - 1;
+	mode_lib->ms.soc = mode_lib->soc;
+	mode_lib->ms.ip = mode_lib->ip;
+	mode_lib->ms.policy = mode_lib->policy;
+	mode_lib->ms.state = dml_get_soc_state_bounding_box(&mode_lib->states, state_idx);
+	mode_lib->ms.max_state = dml_get_soc_state_bounding_box(&mode_lib->states, mode_lib->states.num_states - 1);
+}
+
+static void cache_display_cfg(struct display_mode_lib_st *mode_lib,
+	const struct dml_display_cfg_st *display_cfg)
+{
+	mode_lib->ms.cache_display_cfg = *display_cfg;
+}
+
+static void fetch_socbb_params(struct display_mode_lib_st *mode_lib)
+{
+	struct soc_state_bounding_box_st *state = &mode_lib->ms.state;
+
+	// Default values, SOCCLK, DRAMSpeed, and FabricClock will be reassigned to the same state value in mode_check step
+	// If UseMinimumRequiredDCFCLK is used, the DCFCLK will be the min dcflk for the mode support
+	mode_lib->ms.SOCCLK = (dml_float_t)state->socclk_mhz;
+	mode_lib->ms.DRAMSpeed = (dml_float_t)state->dram_speed_mts;
+	mode_lib->ms.FabricClock = (dml_float_t)state->fabricclk_mhz;
+	mode_lib->ms.DCFCLK = (dml_float_t)state->dcfclk_mhz;
+}
+
+/// @brief Use display_cfg directly for mode_support calculation
+///        Calculated values and informational output are stored in mode_lib.vba data struct
+///        The display configuration is described with pipes struct and num_pipes
+///        This function is used when physical resource mapping is not finalized (for example,
+///        don't know how many pipes to represent a surface)
+/// @param mode_lib Contains the bounding box and policy setting.
+/// @param state_idx Power state index
+/// @param display_cfg Display configurations. A display
+dml_bool_t dml_mode_support(
+	struct display_mode_lib_st *mode_lib,
+	dml_uint_t                        state_idx,
+	const struct dml_display_cfg_st *display_cfg)
+{
+	dml_bool_t is_mode_support;
+
+	dml_print("DML::%s: ------------- START ----------\n", __func__);
+	cache_ip_soc_cfg(mode_lib, state_idx);
+	cache_display_cfg(mode_lib, display_cfg);
+
+	fetch_socbb_params(mode_lib);
+
+	dml_print("DML::%s: state_idx          = %u\n", __func__, state_idx);
+
+	is_mode_support = dml_core_mode_support(mode_lib);
+
+	dml_print("DML::%s: is_mode_support = %u\n", __func__, is_mode_support);
+	dml_print("DML::%s: ------------- DONE ----------\n", __func__);
+	return is_mode_support;
+}
+
+/// @Brief A function to calculate the programming values for DCN DCHUB (Assume mode is supported)
+/// The output will be stored in the mode_lib.mp (mode_program_st) data struct and those can be accessed via the getter functions
+/// Calculated values include: watermarks, dlg, rq reg, different clock frequency
+/// This function returns 1 when there is no error.
+/// Note: In this function, it is assumed that DCFCLK, SOCCLK freq are the state values, and mode_program will just use the DML calculated DPPCLK and DISPCLK
+/// @param mode_lib mode_lib data struct that house all the input/output/bbox and calculation values.
+/// @param state_idx Power state idx chosen
+/// @param display_cfg Display Configuration
+/// @param call_standalone Calling mode_programming without calling mode support.  Some of the "support" struct member will be pre-calculated before doing mode programming
+/// TODO: Add clk_cfg input, could be useful for standalone mode
+dml_bool_t dml_mode_programming(
+	struct display_mode_lib_st *mode_lib,
+	dml_uint_t                         state_idx,
+	const struct dml_display_cfg_st *display_cfg,
+	bool                               call_standalone)
+{
+	struct dml_clk_cfg_st clk_cfg;
+	memset(&clk_cfg, 0, sizeof(clk_cfg));
+
+	clk_cfg.dcfclk_option = dml_use_required_freq;
+	clk_cfg.dispclk_option = dml_use_required_freq;
+	for (dml_uint_t k = 0; k < __DML_NUM_PLANES__; ++k)
+		clk_cfg.dppclk_option[k] = dml_use_required_freq;
+
+	dml_print("DML::%s: ------------- START ----------\n", __func__);
+	dml_print("DML::%s: state_idx       = %u\n", __func__, state_idx);
+	dml_print("DML::%s: call_standalone = %u\n", __func__, call_standalone);
+
+	cache_ip_soc_cfg(mode_lib, state_idx);
+	cache_display_cfg(mode_lib, display_cfg);
+
+	fetch_socbb_params(mode_lib);
+	if (call_standalone) {
+		mode_lib->ms.support.ImmediateFlipSupport = 1; // assume mode support say immediate flip ok at max state/combine
+		dml_core_mode_support_partial(mode_lib);
+	}
+
+	dml_core_mode_programming(mode_lib, &clk_cfg);
+
+	dml_print("DML::%s: ------------- DONE ----------\n", __func__);
+	dml_print("DML::%s: PrefetchAndImmediateFlipSupported = %0d\n", __func__, mode_lib->mp.PrefetchAndImmediateFlipSupported);
+	return mode_lib->mp.PrefetchAndImmediateFlipSupported;
+}
+
+static dml_uint_t mode_support_pwr_states(
+	dml_uint_t *lowest_state_idx,
+	struct display_mode_lib_st *mode_lib,
+	const struct dml_display_cfg_st *display_cfg,
+	dml_uint_t start_state_idx,
+	dml_uint_t end_state_idx)
+{
+	dml_uint_t state_idx = 0;
+	dml_bool_t mode_is_supported = 0;
+	*lowest_state_idx = end_state_idx;
+
+	if (end_state_idx < start_state_idx)
+		ASSERT(0);
+
+	if (end_state_idx >= mode_lib->states.num_states) // idx is 0-based
+		ASSERT(0);
+
+	for (state_idx = start_state_idx; state_idx <= end_state_idx; state_idx++) {
+		if (dml_mode_support(mode_lib, state_idx, display_cfg)) {
+			dml_print("DML::%s: Mode is supported at power state_idx = %u\n", __func__, state_idx);
+			mode_is_supported = 1;
+			*lowest_state_idx = state_idx;
+			break;
+		}
+	}
+
+	return mode_is_supported;
+}
+
+dml_uint_t dml_mode_support_ex(struct dml_mode_support_ex_params_st *in_out_params)
+{
+	dml_uint_t result;
+
+	result = mode_support_pwr_states(&in_out_params->out_lowest_state_idx,
+		in_out_params->mode_lib,
+		in_out_params->in_display_cfg,
+		in_out_params->in_start_state_idx,
+		in_out_params->mode_lib->states.num_states - 1);
+
+	if (result)
+		*in_out_params->out_evaluation_info = in_out_params->mode_lib->ms.support;
+
+	return result;
+}
+
+dml_bool_t dml_get_is_phantom_pipe(struct display_mode_lib_st *mode_lib, dml_uint_t pipe_idx)
+{
+	dml_uint_t plane_idx = mode_lib->mp.pipe_plane[pipe_idx];
+	dml_print("DML::%s: pipe_idx=%d UseMALLForPStateChange=%0d\n", __func__, pipe_idx, mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange[plane_idx]);
+	return (mode_lib->ms.cache_display_cfg.plane.UseMALLForPStateChange[plane_idx] == dml_use_mall_pstate_change_phantom_pipe);
+}
+
+
+#define dml_get_per_surface_var_func(variable, type, interval_var) type dml_get_##variable(struct display_mode_lib_st *mode_lib, dml_uint_t surface_idx) \
+{ \
+	dml_uint_t plane_idx; \
+	plane_idx = mode_lib->mp.pipe_plane[surface_idx]; \
+	return (type) interval_var[plane_idx]; \
+}
+
+#define dml_get_var_func(var, type, internal_var)  type dml_get_##var(struct display_mode_lib_st *mode_lib) \
+{ \
+	return (type) internal_var; \
+}
+
+dml_get_var_func(wm_urgent, dml_float_t, mode_lib->mp.Watermark.UrgentWatermark);
+dml_get_var_func(wm_stutter_exit, dml_float_t, mode_lib->mp.Watermark.StutterExitWatermark);
+dml_get_var_func(wm_stutter_enter_exit, dml_float_t, mode_lib->mp.Watermark.StutterEnterPlusExitWatermark);
+dml_get_var_func(wm_memory_trip, dml_float_t, mode_lib->mp.UrgentLatency);
+dml_get_var_func(wm_fclk_change, dml_float_t, mode_lib->mp.Watermark.FCLKChangeWatermark);
+dml_get_var_func(wm_usr_retraining, dml_float_t, mode_lib->mp.Watermark.USRRetrainingWatermark);
+dml_get_var_func(wm_dram_clock_change, dml_float_t, mode_lib->mp.Watermark.DRAMClockChangeWatermark);
+dml_get_var_func(wm_z8_stutter_enter_exit, dml_float_t, mode_lib->mp.Watermark.Z8StutterEnterPlusExitWatermark);
+dml_get_var_func(wm_z8_stutter, dml_float_t, mode_lib->mp.Watermark.Z8StutterExitWatermark);
+dml_get_var_func(fraction_of_urgent_bandwidth, dml_float_t, mode_lib->mp.FractionOfUrgentBandwidth);
+dml_get_var_func(fraction_of_urgent_bandwidth_imm_flip, dml_float_t, mode_lib->mp.FractionOfUrgentBandwidthImmediateFlip);
+dml_get_var_func(urgent_latency, dml_float_t, mode_lib->mp.UrgentLatency);
+dml_get_var_func(clk_dcf_deepsleep, dml_float_t, mode_lib->mp.DCFCLKDeepSleep);
+dml_get_var_func(wm_writeback_dram_clock_change, dml_float_t, mode_lib->mp.Watermark.WritebackDRAMClockChangeWatermark);
+dml_get_var_func(wm_writeback_urgent, dml_float_t, mode_lib->mp.Watermark.WritebackUrgentWatermark);
+dml_get_var_func(stutter_efficiency, dml_float_t, mode_lib->mp.StutterEfficiency);
+dml_get_var_func(stutter_efficiency_no_vblank, dml_float_t, mode_lib->mp.StutterEfficiencyNotIncludingVBlank);
+dml_get_var_func(stutter_efficiency_z8, dml_float_t, mode_lib->mp.Z8StutterEfficiency);
+dml_get_var_func(stutter_num_bursts_z8, dml_float_t, mode_lib->mp.Z8NumberOfStutterBurstsPerFrame);
+dml_get_var_func(stutter_period, dml_float_t, mode_lib->mp.StutterPeriod);
+dml_get_var_func(stutter_efficiency_z8_bestcase, dml_float_t, mode_lib->mp.Z8StutterEfficiencyBestCase);
+dml_get_var_func(stutter_num_bursts_z8_bestcase, dml_float_t, mode_lib->mp.Z8NumberOfStutterBurstsPerFrameBestCase);
+dml_get_var_func(stutter_period_bestcase, dml_float_t, mode_lib->mp.StutterPeriodBestCase);
+dml_get_var_func(urgent_extra_latency, dml_float_t, mode_lib->mp.UrgentExtraLatency);
+dml_get_var_func(fclk_change_latency, dml_float_t, mode_lib->mp.MaxActiveFCLKChangeLatencySupported);
+dml_get_var_func(dispclk_calculated, dml_float_t, mode_lib->mp.Dispclk_calculated);
+dml_get_var_func(total_data_read_bw, dml_float_t, mode_lib->mp.TotalDataReadBandwidth);
+dml_get_var_func(return_bw, dml_float_t, mode_lib->ms.ReturnBW);
+dml_get_var_func(return_dram_bw, dml_float_t, mode_lib->ms.ReturnDRAMBW);
+dml_get_var_func(tcalc, dml_float_t, mode_lib->mp.TCalc);
+dml_get_var_func(comp_buffer_size_kbytes, dml_uint_t, mode_lib->mp.CompressedBufferSizeInkByte);
+dml_get_var_func(pixel_chunk_size_in_kbyte, dml_uint_t, mode_lib->ms.ip.pixel_chunk_size_kbytes);
+dml_get_var_func(alpha_pixel_chunk_size_in_kbyte, dml_uint_t, mode_lib->ms.ip.alpha_pixel_chunk_size_kbytes);
+dml_get_var_func(meta_chunk_size_in_kbyte, dml_uint_t, mode_lib->ms.ip.meta_chunk_size_kbytes);
+dml_get_var_func(min_pixel_chunk_size_in_byte, dml_uint_t, mode_lib->ms.ip.min_pixel_chunk_size_bytes);
+dml_get_var_func(min_meta_chunk_size_in_byte, dml_uint_t, mode_lib->ms.ip.min_meta_chunk_size_bytes);
+dml_get_var_func(total_immediate_flip_bytes, dml_uint_t, mode_lib->mp.TotImmediateFlipBytes);
+
+dml_get_per_surface_var_func(dsc_delay, dml_uint_t, mode_lib->mp.DSCDelay); // this is the dsc latency
+dml_get_per_surface_var_func(dppclk_calculated, dml_float_t, mode_lib->mp.Dppclk_calculated);
+dml_get_per_surface_var_func(dscclk_calculated, dml_float_t, mode_lib->mp.DSCCLK_calculated);
+dml_get_per_surface_var_func(min_ttu_vblank_in_us, dml_float_t, mode_lib->mp.MinTTUVBlank);
+dml_get_per_surface_var_func(vratio_prefetch_l, dml_float_t, mode_lib->mp.VRatioPrefetchY);
+dml_get_per_surface_var_func(vratio_prefetch_c, dml_float_t, mode_lib->mp.VRatioPrefetchC);
+dml_get_per_surface_var_func(dst_x_after_scaler, dml_uint_t, mode_lib->mp.DSTXAfterScaler);
+dml_get_per_surface_var_func(dst_y_after_scaler, dml_uint_t, mode_lib->mp.DSTYAfterScaler);
+dml_get_per_surface_var_func(dst_y_per_vm_vblank, dml_float_t, mode_lib->mp.DestinationLinesToRequestVMInVBlank);
+dml_get_per_surface_var_func(dst_y_per_row_vblank, dml_float_t, mode_lib->mp.DestinationLinesToRequestRowInVBlank);
+dml_get_per_surface_var_func(dst_y_prefetch, dml_float_t, mode_lib->mp.DestinationLinesForPrefetch);
+dml_get_per_surface_var_func(dst_y_per_vm_flip, dml_float_t, mode_lib->mp.DestinationLinesToRequestVMInImmediateFlip);
+dml_get_per_surface_var_func(dst_y_per_row_flip, dml_float_t, mode_lib->mp.DestinationLinesToRequestRowInImmediateFlip);
+dml_get_per_surface_var_func(dst_y_per_pte_row_nom_l, dml_float_t, mode_lib->mp.DST_Y_PER_PTE_ROW_NOM_L);
+dml_get_per_surface_var_func(dst_y_per_pte_row_nom_c, dml_float_t, mode_lib->mp.DST_Y_PER_PTE_ROW_NOM_C);
+dml_get_per_surface_var_func(dst_y_per_meta_row_nom_l, dml_float_t, mode_lib->mp.DST_Y_PER_META_ROW_NOM_L);
+dml_get_per_surface_var_func(dst_y_per_meta_row_nom_c, dml_float_t, mode_lib->mp.DST_Y_PER_META_ROW_NOM_C);
+dml_get_per_surface_var_func(refcyc_per_vm_group_vblank_in_us, dml_float_t, mode_lib->mp.TimePerVMGroupVBlank);
+dml_get_per_surface_var_func(refcyc_per_vm_group_flip_in_us, dml_float_t, mode_lib->mp.TimePerVMGroupFlip);
+dml_get_per_surface_var_func(refcyc_per_vm_req_vblank_in_us, dml_float_t, mode_lib->mp.TimePerVMRequestVBlank);
+dml_get_per_surface_var_func(refcyc_per_vm_req_flip_in_us, dml_float_t, mode_lib->mp.TimePerVMRequestFlip);
+dml_get_per_surface_var_func(refcyc_per_vm_dmdata_in_us, dml_float_t, mode_lib->mp.Tdmdl_vm);
+dml_get_per_surface_var_func(dmdata_dl_delta_in_us, dml_float_t, mode_lib->mp.Tdmdl);
+dml_get_per_surface_var_func(refcyc_per_line_delivery_l_in_us, dml_float_t, mode_lib->mp.DisplayPipeLineDeliveryTimeLuma);
+dml_get_per_surface_var_func(refcyc_per_line_delivery_c_in_us, dml_float_t, mode_lib->mp.DisplayPipeLineDeliveryTimeChroma);
+dml_get_per_surface_var_func(refcyc_per_line_delivery_pre_l_in_us, dml_float_t, mode_lib->mp.DisplayPipeLineDeliveryTimeLumaPrefetch);
+dml_get_per_surface_var_func(refcyc_per_line_delivery_pre_c_in_us, dml_float_t, mode_lib->mp.DisplayPipeLineDeliveryTimeChromaPrefetch);
+dml_get_per_surface_var_func(refcyc_per_req_delivery_l_in_us, dml_float_t, mode_lib->mp.DisplayPipeRequestDeliveryTimeLuma);
+dml_get_per_surface_var_func(refcyc_per_req_delivery_c_in_us, dml_float_t, mode_lib->mp.DisplayPipeRequestDeliveryTimeChroma);
+dml_get_per_surface_var_func(refcyc_per_req_delivery_pre_l_in_us, dml_float_t, mode_lib->mp.DisplayPipeRequestDeliveryTimeLumaPrefetch);
+dml_get_per_surface_var_func(refcyc_per_req_delivery_pre_c_in_us, dml_float_t, mode_lib->mp.DisplayPipeRequestDeliveryTimeChromaPrefetch);
+dml_get_per_surface_var_func(refcyc_per_cursor_req_delivery_in_us, dml_float_t, mode_lib->mp.CursorRequestDeliveryTime);
+dml_get_per_surface_var_func(refcyc_per_cursor_req_delivery_pre_in_us, dml_float_t, mode_lib->mp.CursorRequestDeliveryTimePrefetch);
+dml_get_per_surface_var_func(refcyc_per_meta_chunk_nom_l_in_us, dml_float_t, mode_lib->mp.TimePerMetaChunkNominal);
+dml_get_per_surface_var_func(refcyc_per_meta_chunk_nom_c_in_us, dml_float_t, mode_lib->mp.TimePerChromaMetaChunkNominal);
+dml_get_per_surface_var_func(refcyc_per_meta_chunk_vblank_l_in_us, dml_float_t, mode_lib->mp.TimePerMetaChunkVBlank);
+dml_get_per_surface_var_func(refcyc_per_meta_chunk_vblank_c_in_us, dml_float_t, mode_lib->mp.TimePerChromaMetaChunkVBlank);
+dml_get_per_surface_var_func(refcyc_per_meta_chunk_flip_l_in_us, dml_float_t, mode_lib->mp.TimePerMetaChunkFlip);
+dml_get_per_surface_var_func(refcyc_per_meta_chunk_flip_c_in_us, dml_float_t, mode_lib->mp.TimePerChromaMetaChunkFlip);
+dml_get_per_surface_var_func(refcyc_per_pte_group_nom_l_in_us, dml_float_t, mode_lib->mp.time_per_pte_group_nom_luma);
+dml_get_per_surface_var_func(refcyc_per_pte_group_nom_c_in_us, dml_float_t, mode_lib->mp.time_per_pte_group_nom_chroma);
+dml_get_per_surface_var_func(refcyc_per_pte_group_vblank_l_in_us, dml_float_t, mode_lib->mp.time_per_pte_group_vblank_luma);
+dml_get_per_surface_var_func(refcyc_per_pte_group_vblank_c_in_us, dml_float_t, mode_lib->mp.time_per_pte_group_vblank_chroma);
+dml_get_per_surface_var_func(refcyc_per_pte_group_flip_l_in_us, dml_float_t, mode_lib->mp.time_per_pte_group_flip_luma);
+dml_get_per_surface_var_func(refcyc_per_pte_group_flip_c_in_us, dml_float_t, mode_lib->mp.time_per_pte_group_flip_chroma);
+dml_get_per_surface_var_func(dpte_group_size_in_bytes, dml_uint_t, mode_lib->mp.dpte_group_bytes);
+dml_get_per_surface_var_func(vm_group_size_in_bytes, dml_uint_t, mode_lib->mp.vm_group_bytes);
+dml_get_per_surface_var_func(swath_height_l, dml_uint_t, mode_lib->ms.SwathHeightY);
+dml_get_per_surface_var_func(swath_height_c, dml_uint_t, mode_lib->ms.SwathHeightC);
+dml_get_per_surface_var_func(dpte_row_height_l, dml_uint_t, mode_lib->mp.dpte_row_height);
+dml_get_per_surface_var_func(dpte_row_height_c, dml_uint_t, mode_lib->mp.dpte_row_height_chroma);
+dml_get_per_surface_var_func(dpte_row_height_linear_l, dml_uint_t, mode_lib->mp.dpte_row_height_linear);
+dml_get_per_surface_var_func(dpte_row_height_linear_c, dml_uint_t, mode_lib->mp.dpte_row_height_linear_chroma);
+dml_get_per_surface_var_func(meta_row_height_l, dml_uint_t, mode_lib->mp.meta_row_height);
+dml_get_per_surface_var_func(meta_row_height_c, dml_uint_t, mode_lib->mp.meta_row_height_chroma);
+
+dml_get_per_surface_var_func(vstartup_calculated, dml_uint_t, mode_lib->mp.VStartup);
+dml_get_per_surface_var_func(vupdate_offset, dml_uint_t, mode_lib->mp.VUpdateOffsetPix);
+dml_get_per_surface_var_func(vupdate_width, dml_uint_t, mode_lib->mp.VUpdateWidthPix);
+dml_get_per_surface_var_func(vready_offset, dml_uint_t, mode_lib->mp.VReadyOffsetPix);
+dml_get_per_surface_var_func(vready_at_or_after_vsync, dml_uint_t, mode_lib->mp.VREADY_AT_OR_AFTER_VSYNC);
+dml_get_per_surface_var_func(min_dst_y_next_start, dml_uint_t, mode_lib->mp.MIN_DST_Y_NEXT_START);
+dml_get_per_surface_var_func(det_stored_buffer_size_l_bytes, dml_uint_t, mode_lib->ms.DETBufferSizeY);
+dml_get_per_surface_var_func(det_stored_buffer_size_c_bytes, dml_uint_t, mode_lib->ms.DETBufferSizeC);
+dml_get_per_surface_var_func(use_mall_for_static_screen, dml_uint_t, mode_lib->mp.UsesMALLForStaticScreen);
+dml_get_per_surface_var_func(surface_size_for_mall, dml_uint_t, mode_lib->mp.SurfaceSizeInTheMALL);
+dml_get_per_surface_var_func(dcc_max_uncompressed_block_l, dml_uint_t, mode_lib->mp.DCCYMaxUncompressedBlock);
+dml_get_per_surface_var_func(dcc_max_compressed_block_l, dml_uint_t, mode_lib->mp.DCCYMaxCompressedBlock);
+dml_get_per_surface_var_func(dcc_independent_block_l, dml_uint_t, mode_lib->mp.DCCYIndependentBlock);
+dml_get_per_surface_var_func(dcc_max_uncompressed_block_c, dml_uint_t, mode_lib->mp.DCCCMaxUncompressedBlock);
+dml_get_per_surface_var_func(dcc_max_compressed_block_c, dml_uint_t, mode_lib->mp.DCCCMaxCompressedBlock);
+dml_get_per_surface_var_func(dcc_independent_block_c, dml_uint_t, mode_lib->mp.DCCCIndependentBlock);
+dml_get_per_surface_var_func(max_active_dram_clock_change_latency_supported, dml_uint_t, mode_lib->mp.MaxActiveDRAMClockChangeLatencySupported);
+dml_get_per_surface_var_func(pte_buffer_mode, dml_uint_t, mode_lib->mp.PTE_BUFFER_MODE);
+dml_get_per_surface_var_func(bigk_fragment_size, dml_uint_t, mode_lib->mp.BIGK_FRAGMENT_SIZE);
+dml_get_per_surface_var_func(dpte_bytes_per_row, dml_uint_t, mode_lib->mp.PixelPTEBytesPerRow);
+dml_get_per_surface_var_func(meta_bytes_per_row, dml_uint_t, mode_lib->mp.MetaRowByte);
+dml_get_per_surface_var_func(det_buffer_size_kbytes, dml_uint_t, mode_lib->ms.DETBufferSizeInKByte);
+
diff --git a/drivers/gpu/drm/amd/display/dc/dml2_0/display_mode_core.h b/drivers/gpu/drm/amd/display/dc/dml2_0/display_mode_core.h
new file mode 100644
index 000000000000..a38ed89c47a9
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml2_0/display_mode_core.h
@@ -0,0 +1,204 @@
+/*
+ * Copyright 2022 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __DISPLAY_MODE_CORE_H__
+#define __DISPLAY_MODE_CORE_H__
+
+#include "display_mode_core_structs.h"
+
+struct display_mode_lib_st;
+
+dml_bool_t dml_core_mode_support(struct display_mode_lib_st *mode_lib);
+void dml_core_mode_support_partial(struct display_mode_lib_st *mode_lib);
+void dml_core_mode_programming(struct display_mode_lib_st *mode_lib, const struct dml_clk_cfg_st *clk_cfg);
+
+void dml_core_get_row_heights(
+						dml_uint_t *dpte_row_height,
+						dml_uint_t *meta_row_height,
+						const struct display_mode_lib_st *mode_lib,
+						dml_bool_t is_plane1,
+						enum dml_source_format_class SourcePixelFormat,
+						enum dml_swizzle_mode SurfaceTiling,
+						enum dml_rotation_angle ScanDirection,
+						dml_uint_t pitch,
+						dml_uint_t GPUVMMinPageSizeKBytes);
+
+dml_float_t dml_get_return_bw_mbps_vm_only(
+									const struct soc_bounding_box_st *soc,
+									dml_bool_t use_ideal_dram_bw_strobe,
+									dml_bool_t HostVMEnable,
+									dml_float_t DCFCLK,
+									dml_float_t FabricClock,
+									dml_float_t DRAMSpeed);
+
+dml_float_t dml_get_return_bw_mbps(
+							const struct soc_bounding_box_st *soc,
+							dml_bool_t use_ideal_dram_bw_strobe,
+							dml_bool_t HostVMEnable,
+							dml_float_t DCFCLK,
+							dml_float_t FabricClock,
+							dml_float_t DRAMSpeed);
+
+dml_bool_t dml_mode_support(
+	struct display_mode_lib_st *mode_lib,
+	dml_uint_t state_idx,
+	const struct dml_display_cfg_st *display_cfg);
+
+dml_bool_t dml_mode_programming(
+	struct display_mode_lib_st *mode_lib,
+	dml_uint_t state_idx,
+	const struct dml_display_cfg_st *display_cfg,
+	bool call_standalone);
+
+dml_uint_t dml_mode_support_ex(
+	struct dml_mode_support_ex_params_st *in_out_params);
+
+dml_bool_t dml_get_is_phantom_pipe(struct display_mode_lib_st *mode_lib, dml_uint_t pipe_idx);
+
+#define dml_get_per_surface_var_decl(variable, type) type dml_get_##variable(struct display_mode_lib_st *mode_lib, dml_uint_t surface_idx)
+#define dml_get_var_decl(var, type) type dml_get_##var(struct display_mode_lib_st *mode_lib)
+
+dml_get_var_decl(wm_urgent, dml_float_t);
+dml_get_var_decl(wm_stutter_exit, dml_float_t);
+dml_get_var_decl(wm_stutter_enter_exit, dml_float_t);
+dml_get_var_decl(wm_memory_trip, dml_float_t);
+dml_get_var_decl(wm_dram_clock_change, dml_float_t);
+dml_get_var_decl(wm_z8_stutter_enter_exit, dml_float_t);
+dml_get_var_decl(wm_z8_stutter, dml_float_t);
+dml_get_var_decl(urgent_latency, dml_float_t);
+dml_get_var_decl(clk_dcf_deepsleep, dml_float_t);
+dml_get_var_decl(wm_fclk_change, dml_float_t);
+dml_get_var_decl(wm_usr_retraining, dml_float_t);
+dml_get_var_decl(urgent_latency, dml_float_t);
+
+dml_get_var_decl(wm_writeback_dram_clock_change, dml_float_t);
+dml_get_var_decl(wm_writeback_urgent, dml_float_t);
+dml_get_var_decl(stutter_efficiency_no_vblank, dml_float_t);
+dml_get_var_decl(stutter_efficiency, dml_float_t);
+dml_get_var_decl(stutter_efficiency_z8, dml_float_t);
+dml_get_var_decl(stutter_num_bursts_z8, dml_float_t);
+dml_get_var_decl(stutter_period, dml_float_t);
+dml_get_var_decl(stutter_efficiency_z8_bestcase, dml_float_t);
+dml_get_var_decl(stutter_num_bursts_z8_bestcase, dml_float_t);
+dml_get_var_decl(stutter_period_bestcase, dml_float_t);
+dml_get_var_decl(urgent_latency, dml_float_t);
+dml_get_var_decl(urgent_extra_latency, dml_float_t);
+dml_get_var_decl(fclk_change_latency, dml_float_t);
+dml_get_var_decl(nonurgent_latency, dml_float_t);
+dml_get_var_decl(dispclk_calculated, dml_float_t);
+dml_get_var_decl(total_data_read_bw, dml_float_t);
+dml_get_var_decl(return_bw, dml_float_t);
+dml_get_var_decl(return_dram_bw, dml_float_t);
+dml_get_var_decl(tcalc, dml_float_t);
+dml_get_var_decl(fraction_of_urgent_bandwidth, dml_float_t);
+dml_get_var_decl(fraction_of_urgent_bandwidth_imm_flip, dml_float_t);
+dml_get_var_decl(comp_buffer_size_kbytes, dml_uint_t);
+dml_get_var_decl(pixel_chunk_size_in_kbyte, dml_uint_t);
+dml_get_var_decl(alpha_pixel_chunk_size_in_kbyte, dml_uint_t);
+dml_get_var_decl(meta_chunk_size_in_kbyte, dml_uint_t);
+dml_get_var_decl(min_pixel_chunk_size_in_byte, dml_uint_t);
+dml_get_var_decl(min_meta_chunk_size_in_byte, dml_uint_t);
+dml_get_var_decl(total_immediate_flip_bytes, dml_uint_t);
+
+dml_get_per_surface_var_decl(dsc_delay, dml_uint_t);
+dml_get_per_surface_var_decl(dppclk_calculated, dml_float_t);
+dml_get_per_surface_var_decl(dscclk_calculated, dml_float_t);
+dml_get_per_surface_var_decl(min_ttu_vblank_in_us, dml_float_t);
+dml_get_per_surface_var_decl(vratio_prefetch_l, dml_float_t);
+dml_get_per_surface_var_decl(vratio_prefetch_c, dml_float_t);
+dml_get_per_surface_var_decl(dst_x_after_scaler, dml_uint_t);
+dml_get_per_surface_var_decl(dst_y_after_scaler, dml_uint_t);
+dml_get_per_surface_var_decl(dst_y_per_vm_vblank, dml_float_t);
+dml_get_per_surface_var_decl(dst_y_per_row_vblank, dml_float_t);
+dml_get_per_surface_var_decl(dst_y_prefetch, dml_float_t);
+dml_get_per_surface_var_decl(dst_y_per_vm_flip, dml_float_t);
+dml_get_per_surface_var_decl(dst_y_per_row_flip, dml_float_t);
+dml_get_per_surface_var_decl(dst_y_per_pte_row_nom_l, dml_float_t);
+dml_get_per_surface_var_decl(dst_y_per_pte_row_nom_c, dml_float_t);
+dml_get_per_surface_var_decl(dst_y_per_meta_row_nom_l, dml_float_t);
+dml_get_per_surface_var_decl(dst_y_per_meta_row_nom_c, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_vm_group_vblank_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_vm_group_flip_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_vm_req_vblank_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_vm_req_flip_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_vm_dmdata_in_us, dml_float_t);
+dml_get_per_surface_var_decl(dmdata_dl_delta_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_line_delivery_l_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_line_delivery_c_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_line_delivery_pre_l_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_line_delivery_pre_c_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_req_delivery_l_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_req_delivery_c_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_req_delivery_pre_l_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_req_delivery_pre_c_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_cursor_req_delivery_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_cursor_req_delivery_pre_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_meta_chunk_nom_l_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_meta_chunk_nom_c_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_meta_chunk_vblank_l_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_meta_chunk_vblank_c_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_meta_chunk_flip_l_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_meta_chunk_flip_c_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_pte_group_nom_l_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_pte_group_nom_c_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_pte_group_vblank_l_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_pte_group_vblank_c_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_pte_group_flip_l_in_us, dml_float_t);
+dml_get_per_surface_var_decl(refcyc_per_pte_group_flip_c_in_us, dml_float_t);
+
+dml_get_per_surface_var_decl(dpte_group_size_in_bytes, dml_uint_t);
+dml_get_per_surface_var_decl(vm_group_size_in_bytes, dml_uint_t);
+dml_get_per_surface_var_decl(swath_height_l, dml_uint_t);
+dml_get_per_surface_var_decl(swath_height_c, dml_uint_t);
+dml_get_per_surface_var_decl(dpte_row_height_l, dml_uint_t);
+dml_get_per_surface_var_decl(dpte_row_height_c, dml_uint_t);
+dml_get_per_surface_var_decl(dpte_row_height_linear_l, dml_uint_t);
+dml_get_per_surface_var_decl(dpte_row_height_linear_c, dml_uint_t);
+dml_get_per_surface_var_decl(meta_row_height_l, dml_uint_t);
+dml_get_per_surface_var_decl(meta_row_height_c, dml_uint_t);
+dml_get_per_surface_var_decl(vstartup_calculated, dml_uint_t);
+dml_get_per_surface_var_decl(vupdate_offset, dml_uint_t);
+dml_get_per_surface_var_decl(vupdate_width, dml_uint_t);
+dml_get_per_surface_var_decl(vready_offset, dml_uint_t);
+dml_get_per_surface_var_decl(vready_at_or_after_vsync, dml_uint_t);
+dml_get_per_surface_var_decl(min_dst_y_next_start, dml_uint_t);
+dml_get_per_surface_var_decl(det_stored_buffer_size_l_bytes, dml_uint_t);
+dml_get_per_surface_var_decl(det_stored_buffer_size_c_bytes, dml_uint_t);
+dml_get_per_surface_var_decl(use_mall_for_static_screen, dml_uint_t);
+dml_get_per_surface_var_decl(surface_size_for_mall, dml_uint_t);
+dml_get_per_surface_var_decl(dcc_max_uncompressed_block_l, dml_uint_t);
+dml_get_per_surface_var_decl(dcc_max_uncompressed_block_c, dml_uint_t);
+dml_get_per_surface_var_decl(dcc_max_compressed_block_l, dml_uint_t);
+dml_get_per_surface_var_decl(dcc_max_compressed_block_c, dml_uint_t);
+dml_get_per_surface_var_decl(dcc_independent_block_l, dml_uint_t);
+dml_get_per_surface_var_decl(dcc_independent_block_c, dml_uint_t);
+dml_get_per_surface_var_decl(max_active_dram_clock_change_latency_supported, dml_uint_t);
+dml_get_per_surface_var_decl(pte_buffer_mode, dml_uint_t);
+dml_get_per_surface_var_decl(bigk_fragment_size, dml_uint_t);
+dml_get_per_surface_var_decl(dpte_bytes_per_row, dml_uint_t);
+dml_get_per_surface_var_decl(meta_bytes_per_row, dml_uint_t);
+dml_get_per_surface_var_decl(det_buffer_size_kbytes, dml_uint_t);
+
+#endif
diff --git a/drivers/gpu/drm/amd/display/dc/dml2_0/display_mode_core_structs.h b/drivers/gpu/drm/amd/display/dc/dml2_0/display_mode_core_structs.h
new file mode 100644
index 000000000000..5b40dcdc4406
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml2_0/display_mode_core_structs.h
@@ -0,0 +1,2032 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __DISPLAY_MODE_CORE_STRUCT_H__
+#define __DISPLAY_MODE_CORE_STRUCT_H__
+
+#include "display_mode_lib_defines.h"
+#include "dml_top_display_cfg_types.h"
+
+enum dml_project_id {
+	dml_project_invalid = 0,
+	dml_project_default = 1,
+	dml_project_dcn32 = dml_project_default,
+	dml_project_dcn321 = 2,
+	dml_project_dcn35 = 3,
+	dml_project_dcn351 = 4,
+	dml_project_dcn401 = 5,
+	dml_project_dcn36 = 6,
+};
+enum dml_prefetch_modes {
+	dml_prefetch_support_uclk_fclk_and_stutter_if_possible = 0,
+	dml_prefetch_support_uclk_fclk_and_stutter = 1,
+	dml_prefetch_support_fclk_and_stutter = 2,
+	dml_prefetch_support_stutter = 3,
+	dml_prefetch_support_none = 4
+};
+enum dml_use_mall_for_pstate_change_mode {
+	dml_use_mall_pstate_change_disable = 0,
+	dml_use_mall_pstate_change_full_frame = 1,
+	dml_use_mall_pstate_change_sub_viewport = 2,
+	dml_use_mall_pstate_change_phantom_pipe = 3,
+	dml_use_mall_pstate_change_phantom_pipe_no_data_return = 4,
+	dml_use_mall_pstate_change_imall = 5
+};
+enum dml_use_mall_for_static_screen_mode {
+	dml_use_mall_static_screen_disable = 0,
+	dml_use_mall_static_screen_enable = 1,
+	dml_use_mall_static_screen_optimize = 2
+};
+enum dml_output_encoder_class {
+	dml_dp = 0,
+	dml_edp = 1,
+	dml_dp2p0 = 2,
+	dml_hdmi = 3,
+	dml_hdmifrl = 4,
+	dml_none = 5
+};
+enum dml_output_link_dp_rate{
+	dml_dp_rate_na = 0,
+	dml_dp_rate_hbr = 1,
+	dml_dp_rate_hbr2 = 2,
+	dml_dp_rate_hbr3 = 3,
+	dml_dp_rate_uhbr10 = 4,
+	dml_dp_rate_uhbr13p5 = 5,
+	dml_dp_rate_uhbr20 = 6
+};
+enum dml_output_type_and_rate__type{
+	dml_output_type_unknown = 0,
+	dml_output_type_dp = 1,
+	dml_output_type_edp = 2,
+	dml_output_type_dp2p0 = 3,
+	dml_output_type_hdmi = 4,
+	dml_output_type_hdmifrl = 5
+};
+enum dml_output_type_and_rate__rate {
+	dml_output_rate_unknown = 0,
+	dml_output_rate_dp_rate_hbr = 1,
+	dml_output_rate_dp_rate_hbr2 = 2,
+	dml_output_rate_dp_rate_hbr3 = 3,
+	dml_output_rate_dp_rate_uhbr10 = 4,
+	dml_output_rate_dp_rate_uhbr13p5 = 5,
+	dml_output_rate_dp_rate_uhbr20 = 6,
+	dml_output_rate_hdmi_rate_3x3 = 7,
+	dml_output_rate_hdmi_rate_6x3 = 8,
+	dml_output_rate_hdmi_rate_6x4 = 9,
+	dml_output_rate_hdmi_rate_8x4 = 10,
+	dml_output_rate_hdmi_rate_10x4 = 11,
+	dml_output_rate_hdmi_rate_12x4 = 12
+};
+enum dml_output_format_class {
+	dml_444 = 0,
+	dml_s422 = 1,
+	dml_n422 = 2,
+	dml_420 = 3
+};
+enum dml_source_format_class {
+	dml_444_8 = 0,
+	dml_444_16 = 1,
+	dml_444_32 = 2,
+	dml_444_64 = 3,
+	dml_420_8 = 4,
+	dml_420_10 = 5,
+	dml_420_12 = 6,
+	dml_422_8 = 7,
+	dml_422_10 = 8,
+	dml_rgbe_alpha = 9,
+	dml_rgbe = 10,
+	dml_mono_8 = 11,
+	dml_mono_16 = 12
+};
+enum dml_output_bpc_class {
+	dml_out_6 = 0,
+	dml_out_8 = 1,
+	dml_out_10 = 2,
+	dml_out_12 = 3,
+	dml_out_16 = 4
+};
+enum dml_output_standard_class {
+	dml_std_cvt = 0,
+	dml_std_cea = 1,
+	dml_std_cvtr2 = 2
+};
+enum dml_rotation_angle {
+	dml_rotation_0 = 0,
+	dml_rotation_90 = 1,
+	dml_rotation_180 = 2,
+	dml_rotation_270 = 3,
+	dml_rotation_0m = 4,
+	dml_rotation_90m = 5,
+	dml_rotation_180m = 6,
+	dml_rotation_270m = 7
+};
+enum dml_swizzle_mode {
+	dml_sw_linear = 0,
+	dml_sw_256b_s = 1,
+	dml_sw_256b_d = 2,
+	dml_sw_256b_r = 3,
+	dml_sw_4kb_z = 4,
+	dml_sw_4kb_s = 5,
+	dml_sw_4kb_d = 6,
+	dml_sw_4kb_r = 7,
+	dml_sw_64kb_z = 8,
+	dml_sw_64kb_s = 9,
+	dml_sw_64kb_d = 10,
+	dml_sw_64kb_r = 11,
+	dml_sw_256kb_z = 12,
+	dml_sw_256kb_s = 13,
+	dml_sw_256kb_d = 14,
+	dml_sw_256kb_r = 15,
+	dml_sw_64kb_z_t = 16,
+	dml_sw_64kb_s_t = 17,
+	dml_sw_64kb_d_t = 18,
+	dml_sw_64kb_r_t = 19,
+	dml_sw_4kb_z_x = 20,
+	dml_sw_4kb_s_x = 21,
+	dml_sw_4kb_d_x = 22,
+	dml_sw_4kb_r_x = 23,
+	dml_sw_64kb_z_x = 24,
+	dml_sw_64kb_s_x = 25,
+	dml_sw_64kb_d_x = 26,
+	dml_sw_64kb_r_x = 27,
+	dml_sw_256kb_z_x = 28,
+	dml_sw_256kb_s_x = 29,
+	dml_sw_256kb_d_x = 30,
+	dml_sw_256kb_r_x = 31,
+	dml_sw_256b_2d = 32,
+	dml_sw_4kb_2d = 33,
+	dml_sw_64kb_2d = 34,
+	dml_sw_256kb_2d = 35
+};
+enum dml_lb_depth {
+	dml_lb_6 = 0,
+	dml_lb_8 = 1,
+	dml_lb_10 = 2,
+	dml_lb_12 = 3,
+	dml_lb_16 = 4
+};
+enum dml_voltage_state {
+	dml_vmin_lv = 0,
+	dml_vmin = 1,
+	dml_vmid = 2,
+	dml_vnom = 3,
+	dml_vmax = 4
+};
+enum dml_source_macro_tile_size {
+	dml_4k_tile = 0,
+	dml_64k_tile = 1,
+	dml_256k_tile = 2
+};
+enum dml_cursor_bpp {
+	dml_cur_2bit = 0,
+	dml_cur_32bit = 1,
+	dml_cur_64bit = 2
+};
+enum dml_dram_clock_change_support {
+	dml_dram_clock_change_vactive = 0,
+	dml_dram_clock_change_vblank = 1,
+	dml_dram_clock_change_vblank_drr = 2,
+	dml_dram_clock_change_vactive_w_mall_full_frame = 3,
+	dml_dram_clock_change_vactive_w_mall_sub_vp = 4,
+	dml_dram_clock_change_vblank_w_mall_full_frame = 5,
+	dml_dram_clock_change_vblank_drr_w_mall_full_frame = 6,
+	dml_dram_clock_change_vblank_w_mall_sub_vp = 7,
+	dml_dram_clock_change_vblank_drr_w_mall_sub_vp = 8,
+	dml_dram_clock_change_unsupported = 9
+};
+enum dml_fclock_change_support {
+	dml_fclock_change_vactive = 0,
+	dml_fclock_change_vblank = 1,
+	dml_fclock_change_unsupported = 2
+};
+enum dml_dsc_enable {
+	dml_dsc_disable = 0,
+	dml_dsc_enable = 1,
+	dml_dsc_enable_if_necessary = 2
+};
+enum dml_mpc_use_policy {
+	dml_mpc_disabled = 0,
+	dml_mpc_as_possible = 1,
+	dml_mpc_as_needed_for_voltage = 2,
+	dml_mpc_as_needed_for_pstate_and_voltage = 3,
+	dml_mpc_as_needed = 4,
+	dml_mpc_2to1 = 5
+};
+enum dml_odm_use_policy {
+	dml_odm_use_policy_bypass = 0,
+	dml_odm_use_policy_combine_as_needed = 1,
+	dml_odm_use_policy_combine_2to1 = 2,
+	dml_odm_use_policy_combine_3to1 = 3,
+	dml_odm_use_policy_combine_4to1 = 4,
+	dml_odm_use_policy_split_1to2 = 5,
+	dml_odm_use_policy_mso_1to2 = 6,
+	dml_odm_use_policy_mso_1to4 = 7
+};
+enum dml_odm_mode {
+	dml_odm_mode_bypass = 0,
+	dml_odm_mode_combine_2to1 = 1,
+	dml_odm_mode_combine_3to1 = 2,
+	dml_odm_mode_combine_4to1 = 3,
+	dml_odm_mode_split_1to2 = 4,
+	dml_odm_mode_mso_1to2 = 5,
+	dml_odm_mode_mso_1to4 = 6
+};
+enum dml_writeback_configuration {
+	dml_whole_buffer_for_single_stream_no_interleave = 0,
+	dml_whole_buffer_for_single_stream_interleave = 1
+};
+enum dml_immediate_flip_requirement {
+	dml_immediate_flip_not_required = 0,
+	dml_immediate_flip_required = 1,
+	dml_immediate_flip_if_possible = 2
+};
+enum dml_unbounded_requesting_policy {
+	dml_unbounded_requesting_enable = 0,
+	dml_unbounded_requesting_edp_only = 1,
+	dml_unbounded_requesting_disable = 2
+};
+enum dml_clk_cfg_policy {
+	dml_use_required_freq = 0,
+	dml_use_override_freq = 1,
+	dml_use_state_freq = 2
+};
+
+struct soc_state_bounding_box_st {
+	dml_float_t socclk_mhz;
+	dml_float_t dscclk_mhz;
+	dml_float_t phyclk_mhz;
+	dml_float_t phyclk_d18_mhz;
+	dml_float_t phyclk_d32_mhz;
+	dml_float_t dtbclk_mhz;
+	dml_float_t fabricclk_mhz;
+	dml_float_t dcfclk_mhz;
+	dml_float_t dispclk_mhz;
+	dml_float_t dppclk_mhz;
+	dml_float_t dram_speed_mts;
+	dml_float_t urgent_latency_pixel_data_only_us;
+	dml_float_t urgent_latency_pixel_mixed_with_vm_data_us;
+	dml_float_t urgent_latency_vm_data_only_us;
+	dml_float_t writeback_latency_us;
+	dml_float_t urgent_latency_adjustment_fabric_clock_component_us;
+	dml_float_t urgent_latency_adjustment_fabric_clock_reference_mhz;
+	dml_float_t sr_exit_time_us;
+	dml_float_t sr_enter_plus_exit_time_us;
+	dml_float_t sr_exit_z8_time_us;
+	dml_float_t sr_enter_plus_exit_z8_time_us;
+	dml_float_t dram_clock_change_latency_us;
+	dml_float_t fclk_change_latency_us;
+	dml_float_t usr_retraining_latency_us;
+	dml_bool_t use_ideal_dram_bw_strobe;
+	dml_float_t g6_temp_read_blackout_us;
+
+	struct {
+		dml_uint_t urgent_ramp_uclk_cycles;
+		dml_uint_t trip_to_memory_uclk_cycles;
+		dml_uint_t meta_trip_to_memory_uclk_cycles;
+		dml_uint_t maximum_latency_when_urgent_uclk_cycles;
+		dml_uint_t average_latency_when_urgent_uclk_cycles;
+		dml_uint_t maximum_latency_when_non_urgent_uclk_cycles;
+		dml_uint_t average_latency_when_non_urgent_uclk_cycles;
+	}  dml_dcn401_uclk_dpm_dependent_soc_qos_params;
+};
+
+struct soc_bounding_box_st {
+	dml_float_t dprefclk_mhz;
+	dml_float_t xtalclk_mhz;
+	dml_float_t pcierefclk_mhz;
+	dml_float_t refclk_mhz;
+	dml_float_t amclk_mhz;
+	dml_uint_t max_outstanding_reqs;
+	dml_float_t pct_ideal_sdp_bw_after_urgent;
+	dml_float_t pct_ideal_fabric_bw_after_urgent;
+	dml_float_t pct_ideal_dram_bw_after_urgent_pixel_only;
+	dml_float_t pct_ideal_dram_bw_after_urgent_pixel_and_vm;
+	dml_float_t pct_ideal_dram_bw_after_urgent_vm_only;
+	dml_float_t pct_ideal_dram_bw_after_urgent_strobe;
+	dml_float_t max_avg_sdp_bw_use_normal_percent;
+	dml_float_t max_avg_fabric_bw_use_normal_percent;
+	dml_float_t max_avg_dram_bw_use_normal_percent;
+	dml_float_t max_avg_dram_bw_use_normal_strobe_percent;
+
+	dml_float_t svp_prefetch_pct_ideal_sdp_bw_after_urgent;
+	dml_float_t svp_prefetch_pct_ideal_fabric_bw_after_urgent;
+	dml_float_t svp_prefetch_pct_ideal_dram_bw_after_urgent_pixel_only;
+	dml_float_t svp_prefetch_pct_ideal_dram_bw_after_urgent_pixel_and_vm;
+	dml_float_t svp_prefetch_pct_ideal_dram_bw_after_urgent_vm_only;
+	dml_float_t svp_prefetch_max_avg_sdp_bw_use_normal_percent;
+	dml_float_t svp_prefetch_max_avg_fabric_bw_use_normal_percent;
+	dml_float_t svp_prefetch_max_avg_dram_bw_use_normal_percent;
+
+	dml_uint_t round_trip_ping_latency_dcfclk_cycles;
+	dml_uint_t urgent_out_of_order_return_per_channel_pixel_only_bytes;
+	dml_uint_t urgent_out_of_order_return_per_channel_pixel_and_vm_bytes;
+	dml_uint_t urgent_out_of_order_return_per_channel_vm_only_bytes;
+	dml_uint_t num_chans;
+	dml_uint_t return_bus_width_bytes;
+	dml_uint_t dram_channel_width_bytes;
+	dml_uint_t fabric_datapath_to_dcn_data_return_bytes;
+	dml_uint_t hostvm_min_page_size_kbytes;
+	dml_uint_t gpuvm_min_page_size_kbytes;
+	dml_float_t phy_downspread_percent;
+	dml_float_t dcn_downspread_percent;
+	dml_float_t smn_latency_us;
+	dml_uint_t mall_allocated_for_dcn_mbytes;
+	dml_float_t dispclk_dppclk_vco_speed_mhz;
+	dml_bool_t do_urgent_latency_adjustment;
+
+	dml_uint_t mem_word_bytes;
+	dml_uint_t num_dcc_mcaches;
+	dml_uint_t mcache_size_bytes;
+	dml_uint_t mcache_line_size_bytes;
+
+	struct {
+		dml_bool_t UseNewDCN401SOCParameters;
+		dml_uint_t df_qos_response_time_fclk_cycles;
+		dml_uint_t max_round_trip_to_furthest_cs_fclk_cycles;
+		dml_uint_t mall_overhead_fclk_cycles;
+		dml_uint_t meta_trip_adder_fclk_cycles;
+		dml_uint_t average_transport_distance_fclk_cycles;
+		dml_float_t umc_urgent_ramp_latency_margin;
+		dml_float_t umc_max_latency_margin;
+		dml_float_t umc_average_latency_margin;
+		dml_float_t fabric_max_transport_latency_margin;
+		dml_float_t fabric_average_transport_latency_margin;
+	}  dml_dcn401_soc_qos_params;
+
+};
+
+struct ip_params_st {
+	dml_uint_t vblank_nom_default_us;
+	dml_uint_t rob_buffer_size_kbytes;
+	dml_uint_t config_return_buffer_size_in_kbytes;
+	dml_uint_t config_return_buffer_segment_size_in_kbytes;
+	dml_uint_t compressed_buffer_segment_size_in_kbytes;
+	dml_uint_t meta_fifo_size_in_kentries;
+	dml_uint_t zero_size_buffer_entries;
+	dml_uint_t dpte_buffer_size_in_pte_reqs_luma;
+	dml_uint_t dpte_buffer_size_in_pte_reqs_chroma;
+	dml_uint_t dcc_meta_buffer_size_bytes;
+	dml_bool_t gpuvm_enable;
+	dml_bool_t hostvm_enable;
+	dml_uint_t gpuvm_max_page_table_levels;
+	dml_uint_t hostvm_max_page_table_levels;
+	dml_uint_t pixel_chunk_size_kbytes;
+	dml_uint_t alpha_pixel_chunk_size_kbytes;
+	dml_uint_t min_pixel_chunk_size_bytes;
+	dml_uint_t meta_chunk_size_kbytes;
+	dml_uint_t min_meta_chunk_size_bytes;
+	dml_uint_t writeback_chunk_size_kbytes;
+	dml_uint_t line_buffer_size_bits;
+	dml_uint_t max_line_buffer_lines;
+	dml_uint_t writeback_interface_buffer_size_kbytes;
+	dml_uint_t max_num_dpp;
+	dml_uint_t max_num_otg;
+	dml_uint_t max_num_wb;
+	dml_uint_t max_dchub_pscl_bw_pix_per_clk;
+	dml_uint_t max_pscl_lb_bw_pix_per_clk;
+	dml_uint_t max_lb_vscl_bw_pix_per_clk;
+	dml_uint_t max_vscl_hscl_bw_pix_per_clk;
+	dml_float_t max_hscl_ratio;
+	dml_float_t max_vscl_ratio;
+	dml_uint_t max_hscl_taps;
+	dml_uint_t max_vscl_taps;
+	dml_uint_t num_dsc;
+	dml_uint_t maximum_dsc_bits_per_component;
+	dml_uint_t maximum_pixels_per_line_per_dsc_unit;
+	dml_bool_t dsc422_native_support;
+	dml_bool_t cursor_64bpp_support;
+	dml_float_t dispclk_ramp_margin_percent;
+	dml_uint_t dppclk_delay_subtotal;
+	dml_uint_t dppclk_delay_scl;
+	dml_uint_t dppclk_delay_scl_lb_only;
+	dml_uint_t dppclk_delay_cnvc_formatter;
+	dml_uint_t dppclk_delay_cnvc_cursor;
+	dml_uint_t cursor_buffer_size;
+	dml_uint_t cursor_chunk_size;
+	dml_uint_t dispclk_delay_subtotal;
+	dml_bool_t dynamic_metadata_vm_enabled;
+	dml_uint_t max_inter_dcn_tile_repeaters;
+	dml_uint_t max_num_hdmi_frl_outputs;
+	dml_uint_t max_num_dp2p0_outputs;
+	dml_uint_t max_num_dp2p0_streams;
+	dml_bool_t dcc_supported;
+	dml_bool_t ptoi_supported;
+	dml_float_t writeback_max_hscl_ratio;
+	dml_float_t writeback_max_vscl_ratio;
+	dml_float_t writeback_min_hscl_ratio;
+	dml_float_t writeback_min_vscl_ratio;
+	dml_uint_t writeback_max_hscl_taps;
+	dml_uint_t writeback_max_vscl_taps;
+	dml_uint_t writeback_line_buffer_buffer_size;
+};
+
+struct DmlPipe {
+	dml_float_t Dppclk;
+	dml_float_t Dispclk;
+	dml_float_t PixelClock;
+	dml_float_t DCFClkDeepSleep;
+	dml_uint_t DPPPerSurface;
+	dml_bool_t ScalerEnabled;
+	enum dml_rotation_angle SourceScan;
+	dml_uint_t ViewportHeight;
+	dml_uint_t ViewportHeightChroma;
+	dml_uint_t BlockWidth256BytesY;
+	dml_uint_t BlockHeight256BytesY;
+	dml_uint_t BlockWidth256BytesC;
+	dml_uint_t BlockHeight256BytesC;
+	dml_uint_t BlockWidthY;
+	dml_uint_t BlockHeightY;
+	dml_uint_t BlockWidthC;
+	dml_uint_t BlockHeightC;
+	dml_uint_t InterlaceEnable;
+	dml_uint_t NumberOfCursors;
+	dml_uint_t VBlank;
+	dml_uint_t HTotal;
+	dml_uint_t HActive;
+	dml_bool_t DCCEnable;
+	enum dml_odm_mode ODMMode;
+	enum dml_source_format_class SourcePixelFormat;
+	enum dml_swizzle_mode SurfaceTiling;
+	dml_uint_t BytePerPixelY;
+	dml_uint_t BytePerPixelC;
+	dml_bool_t ProgressiveToInterlaceUnitInOPP;
+	dml_float_t VRatio;
+	dml_float_t VRatioChroma;
+	dml_uint_t VTaps;
+	dml_uint_t VTapsChroma;
+	dml_uint_t PitchY;
+	dml_uint_t DCCMetaPitchY;
+	dml_uint_t PitchC;
+	dml_uint_t DCCMetaPitchC;
+	dml_bool_t ViewportStationary;
+	dml_uint_t ViewportXStart;
+	dml_uint_t ViewportYStart;
+	dml_uint_t ViewportXStartC;
+	dml_uint_t ViewportYStartC;
+	dml_bool_t FORCE_ONE_ROW_FOR_FRAME;
+	dml_uint_t SwathHeightY;
+	dml_uint_t SwathHeightC;
+};
+
+struct Watermarks {
+	dml_float_t UrgentWatermark;
+	dml_float_t WritebackUrgentWatermark;
+	dml_float_t DRAMClockChangeWatermark;
+	dml_float_t FCLKChangeWatermark;
+	dml_float_t WritebackDRAMClockChangeWatermark;
+	dml_float_t WritebackFCLKChangeWatermark;
+	dml_float_t StutterExitWatermark;
+	dml_float_t StutterEnterPlusExitWatermark;
+	dml_float_t Z8StutterExitWatermark;
+	dml_float_t Z8StutterEnterPlusExitWatermark;
+	dml_float_t USRRetrainingWatermark;
+};
+
+struct SOCParametersList {
+	dml_float_t UrgentLatency;
+	dml_float_t ExtraLatency;
+	dml_float_t WritebackLatency;
+	dml_float_t DRAMClockChangeLatency;
+	dml_float_t FCLKChangeLatency;
+	dml_float_t SRExitTime;
+	dml_float_t SREnterPlusExitTime;
+	dml_float_t SRExitZ8Time;
+	dml_float_t SREnterPlusExitZ8Time;
+	dml_float_t USRRetrainingLatency;
+	dml_float_t SMNLatency;
+};
+
+/// @brief Struct that represent Plane configration of a display cfg
+struct dml_plane_cfg_st {
+	//
+	// Pipe/Surface Parameters
+	//
+	dml_bool_t GPUVMEnable; /// <brief Set if any pipe has GPUVM enable
+	dml_bool_t HostVMEnable; /// <brief Set if any pipe has HostVM enable
+
+	dml_uint_t GPUVMMaxPageTableLevels; /// <brief GPUVM level; max of all pipes'
+	dml_uint_t HostVMMaxPageTableLevels; /// <brief HostVM level; max of all pipes'; that is the number of non-cache HVM level
+
+	dml_uint_t GPUVMMinPageSizeKBytes[__DML_NUM_PLANES__];
+	dml_bool_t ForceOneRowForFrame[__DML_NUM_PLANES__];
+	dml_bool_t PTEBufferModeOverrideEn[__DML_NUM_PLANES__]; //< brief when override enable; the DML will only check the given pte buffer and will use the pte buffer mode as is
+	dml_bool_t PTEBufferMode[__DML_NUM_PLANES__];
+	dml_uint_t ViewportWidth[__DML_NUM_PLANES__];
+	dml_uint_t ViewportHeight[__DML_NUM_PLANES__];
+	dml_uint_t ViewportWidthChroma[__DML_NUM_PLANES__];
+	dml_uint_t ViewportHeightChroma[__DML_NUM_PLANES__];
+	dml_uint_t ViewportXStart[__DML_NUM_PLANES__];
+	dml_uint_t ViewportXStartC[__DML_NUM_PLANES__];
+	dml_uint_t ViewportYStart[__DML_NUM_PLANES__];
+	dml_uint_t ViewportYStartC[__DML_NUM_PLANES__];
+	dml_bool_t ViewportStationary[__DML_NUM_PLANES__];
+
+	dml_bool_t ScalerEnabled[__DML_NUM_PLANES__];
+	dml_float_t HRatio[__DML_NUM_PLANES__];
+	dml_float_t VRatio[__DML_NUM_PLANES__];
+	dml_float_t HRatioChroma[__DML_NUM_PLANES__];
+	dml_float_t VRatioChroma[__DML_NUM_PLANES__];
+	dml_uint_t HTaps[__DML_NUM_PLANES__];
+	dml_uint_t VTaps[__DML_NUM_PLANES__];
+	dml_uint_t HTapsChroma[__DML_NUM_PLANES__];
+	dml_uint_t VTapsChroma[__DML_NUM_PLANES__];
+	dml_uint_t LBBitPerPixel[__DML_NUM_PLANES__];
+
+	enum dml_rotation_angle SourceScan[__DML_NUM_PLANES__];
+	dml_uint_t ScalerRecoutWidth[__DML_NUM_PLANES__];
+
+	dml_bool_t DynamicMetadataEnable[__DML_NUM_PLANES__];
+	dml_uint_t DynamicMetadataLinesBeforeActiveRequired[__DML_NUM_PLANES__];
+	dml_uint_t DynamicMetadataTransmittedBytes[__DML_NUM_PLANES__];
+	dml_uint_t DETSizeOverride[__DML_NUM_PLANES__]; /// <brief user can specify the desire DET buffer usage per-plane
+
+	dml_uint_t NumberOfCursors[__DML_NUM_PLANES__];
+	dml_uint_t CursorWidth[__DML_NUM_PLANES__];
+	dml_uint_t CursorBPP[__DML_NUM_PLANES__];
+
+	dml_bool_t setup_for_tdlut[__DML_NUM_PLANES__];
+	enum dml2_tdlut_addressing_mode tdlut_addressing_mode[__DML_NUM_PLANES__];
+	enum dml2_tdlut_width_mode tdlut_width_mode[__DML_NUM_PLANES__];
+
+	enum dml_use_mall_for_static_screen_mode UseMALLForStaticScreen[__DML_NUM_PLANES__];
+	enum dml_use_mall_for_pstate_change_mode UseMALLForPStateChange[__DML_NUM_PLANES__];
+
+	dml_uint_t BlendingAndTiming[__DML_NUM_PLANES__]; /// <brief From which timing group (like OTG) that this plane is getting its timing from. Mode check also need this info for example to check num OTG; encoder; dsc etc.
+}; // dml_plane_cfg_st;
+
+/// @brief Surface Parameters
+struct dml_surface_cfg_st {
+	enum dml_swizzle_mode SurfaceTiling[__DML_NUM_PLANES__];
+	enum dml_source_format_class SourcePixelFormat[__DML_NUM_PLANES__];
+	dml_uint_t PitchY[__DML_NUM_PLANES__];
+	dml_uint_t SurfaceWidthY[__DML_NUM_PLANES__];
+	dml_uint_t SurfaceHeightY[__DML_NUM_PLANES__];
+	dml_uint_t PitchC[__DML_NUM_PLANES__];
+	dml_uint_t SurfaceWidthC[__DML_NUM_PLANES__];
+	dml_uint_t SurfaceHeightC[__DML_NUM_PLANES__];
+
+	dml_bool_t DCCEnable[__DML_NUM_PLANES__];
+	dml_uint_t DCCMetaPitchY[__DML_NUM_PLANES__];
+	dml_uint_t DCCMetaPitchC[__DML_NUM_PLANES__];
+
+	dml_float_t DCCRateLuma[__DML_NUM_PLANES__];
+	dml_float_t DCCRateChroma[__DML_NUM_PLANES__];
+	dml_float_t DCCFractionOfZeroSizeRequestsLuma[__DML_NUM_PLANES__];
+	dml_float_t DCCFractionOfZeroSizeRequestsChroma[__DML_NUM_PLANES__];
+}; // dml_surface_cfg_st
+
+/// @brief structure that represents the timing configuration
+struct dml_timing_cfg_st {
+	dml_uint_t HTotal[__DML_NUM_PLANES__];
+	dml_uint_t VTotal[__DML_NUM_PLANES__];
+	dml_uint_t HBlankEnd[__DML_NUM_PLANES__];
+	dml_uint_t VBlankEnd[__DML_NUM_PLANES__];
+	dml_uint_t RefreshRate[__DML_NUM_PLANES__];
+	dml_uint_t VFrontPorch[__DML_NUM_PLANES__];
+	dml_float_t PixelClock[__DML_NUM_PLANES__];
+	dml_uint_t HActive[__DML_NUM_PLANES__];
+	dml_uint_t VActive[__DML_NUM_PLANES__];
+	dml_bool_t Interlace[__DML_NUM_PLANES__];
+	dml_bool_t DRRDisplay[__DML_NUM_PLANES__];
+	dml_uint_t VBlankNom[__DML_NUM_PLANES__];
+}; // dml_timing_cfg_st;
+
+/// @brief structure that represents the output stream
+struct dml_output_cfg_st {
+	// Output Setting
+	dml_uint_t DSCInputBitPerComponent[__DML_NUM_PLANES__];
+	enum dml_output_format_class OutputFormat[__DML_NUM_PLANES__];
+	enum dml_output_encoder_class OutputEncoder[__DML_NUM_PLANES__];
+	dml_uint_t OutputMultistreamId[__DML_NUM_PLANES__];
+	dml_bool_t OutputMultistreamEn[__DML_NUM_PLANES__];
+	dml_float_t OutputBpp[__DML_NUM_PLANES__]; //< brief Use by mode_programming to specify a output bpp; user can use the output from mode_support (support.OutputBpp)
+	dml_float_t PixelClockBackEnd[__DML_NUM_PLANES__];
+	enum dml_dsc_enable DSCEnable[__DML_NUM_PLANES__]; //< brief for mode support check; use to determine if dsc is required
+	dml_uint_t OutputLinkDPLanes[__DML_NUM_PLANES__];
+	enum dml_output_link_dp_rate OutputLinkDPRate[__DML_NUM_PLANES__];
+	dml_float_t ForcedOutputLinkBPP[__DML_NUM_PLANES__];
+	dml_uint_t AudioSampleRate[__DML_NUM_PLANES__];
+	dml_uint_t AudioSampleLayout[__DML_NUM_PLANES__];
+	dml_bool_t OutputDisabled[__DML_NUM_PLANES__];
+	dml_uint_t DSCSlices[__DML_NUM_PLANES__];
+}; // dml_timing_cfg_st;
+
+/// @brief Writeback Setting
+struct dml_writeback_cfg_st {
+	enum dml_source_format_class WritebackPixelFormat[__DML_NUM_PLANES__];
+	dml_bool_t WritebackEnable[__DML_NUM_PLANES__];
+	dml_uint_t ActiveWritebacksPerSurface[__DML_NUM_PLANES__];
+	dml_uint_t WritebackDestinationWidth[__DML_NUM_PLANES__];
+	dml_uint_t WritebackDestinationHeight[__DML_NUM_PLANES__];
+	dml_uint_t WritebackSourceWidth[__DML_NUM_PLANES__];
+	dml_uint_t WritebackSourceHeight[__DML_NUM_PLANES__];
+	dml_uint_t WritebackHTaps[__DML_NUM_PLANES__];
+	dml_uint_t WritebackVTaps[__DML_NUM_PLANES__];
+	dml_float_t WritebackHRatio[__DML_NUM_PLANES__];
+	dml_float_t WritebackVRatio[__DML_NUM_PLANES__];
+}; // dml_writeback_cfg_st;
+
+/// @brief Hardware resource specific; mainly used by mode_programming when test/sw wants to do some specific setting
+///        which are not the same as what the mode support stage derive.  When call mode_support with mode_programm; the hw-specific
+//         resource will be set to what the mode_support layer recommends
+struct dml_hw_resource_st {
+	enum dml_odm_mode ODMMode[__DML_NUM_PLANES__]; /// <brief ODM mode that is chosen in the mode check stage and will be used in mode programming stage
+	dml_uint_t DPPPerSurface[__DML_NUM_PLANES__]; /// <brief How many DPPs are needed drive the surface to output. If MPCC or ODMC could be 2 or 4.
+	dml_bool_t DSCEnabled[__DML_NUM_PLANES__]; /// <brief Indicate if the DSC is enabled; used in mode_programming
+	dml_uint_t NumberOfDSCSlices[__DML_NUM_PLANES__]; /// <brief Indicate how many slices needed to support the given mode
+	dml_float_t DLGRefClkFreqMHz; /// <brief DLG Global Reference timer
+};
+
+/// @brief To control the clk usage for model programming
+struct dml_clk_cfg_st {
+	enum dml_clk_cfg_policy dcfclk_option; ///< brief Use for mode_program; user can select between use the min require clk req as calculated by DML or use the test-specific freq
+	enum dml_clk_cfg_policy dispclk_option; ///< brief Use for mode_program; user can select between use the min require clk req as calculated by DML or use the test-specific freq
+	enum dml_clk_cfg_policy dppclk_option[__DML_NUM_PLANES__];
+
+	dml_float_t dcfclk_mhz;
+	dml_float_t dispclk_mhz;
+	dml_float_t dppclk_mhz[__DML_NUM_PLANES__];
+}; // dml_clk_cfg_st
+
+/// @brief DML display configuration.
+///        Describe how to display a surface in multi-plane setup and output to different output and writeback using the specified timgin
+struct dml_display_cfg_st {
+	struct dml_surface_cfg_st surface;
+	struct dml_plane_cfg_st plane;
+	struct dml_timing_cfg_st timing;
+	struct dml_output_cfg_st output;
+	struct dml_writeback_cfg_st writeback;
+	unsigned int num_surfaces;
+	unsigned int num_timings;
+
+	struct dml_hw_resource_st hw; //< brief for mode programming
+	struct dml_clk_cfg_st clk_overrides;   //< brief for mode programming clk override
+}; // dml_display_cfg_st
+
+/// @brief DML mode evaluation and programming policy
+/// Those knobs that affect mode support and mode programming
+struct dml_mode_eval_policy_st {
+	// -------------------
+	// Policy
+	// -------------------
+	enum dml_mpc_use_policy MPCCombineUse[__DML_NUM_PLANES__]; /// <brief MPC Combine mode as selected by the user; used in mode check stage
+	enum dml_odm_use_policy ODMUse[__DML_NUM_PLANES__]; /// <brief ODM mode as selected by the user; used in mode check stage
+	enum dml_unbounded_requesting_policy UseUnboundedRequesting; ///< brief Unbounded request mode preference
+	enum dml_immediate_flip_requirement ImmediateFlipRequirement[__DML_NUM_PLANES__]; /// <brief Is immediate flip a requirement for this plane. When host vm is present iflip is needed regardless
+	enum dml_prefetch_modes AllowForPStateChangeOrStutterInVBlank[__DML_NUM_PLANES__]; /// <brief To specify if the DML should calculate the values for support different pwr saving features (cstate; pstate; etc.) during vblank
+
+	enum dml_prefetch_modes AllowForPStateChangeOrStutterInVBlankFinal;
+	bool UseOnlyMaxPrefetchModes;
+	dml_bool_t UseMinimumRequiredDCFCLK; //<brief When set the mode_check stage will figure the min DCFCLK freq to support the given display configuration. User can tell use the output DCFCLK for mode programming.
+	dml_bool_t DRAMClockChangeRequirementFinal;
+	dml_bool_t FCLKChangeRequirementFinal;
+	dml_bool_t USRRetrainingRequiredFinal;
+	dml_bool_t EnhancedPrefetchScheduleAccelerationFinal;
+
+	dml_bool_t NomDETInKByteOverrideEnable; //<brief Nomimal DET buffer size for a pipe. If this size fit the required 2 swathes; DML will use this DET size
+	dml_uint_t NomDETInKByteOverrideValue;
+
+	dml_bool_t DCCProgrammingAssumesScanDirectionUnknownFinal;
+	dml_bool_t SynchronizeTimingsFinal;
+	dml_bool_t SynchronizeDRRDisplaysForUCLKPStateChangeFinal;
+	dml_bool_t AssumeModeSupportAtMaxPwrStateEvenDRAMClockChangeNotSupported; //<brief if set; the mode support will say mode is supported even though the DRAM clock change is not support (assuming the soc will be stay in max power state)
+	dml_bool_t AssumeModeSupportAtMaxPwrStateEvenFClockChangeNotSupported; //<brief if set; the mode support will say mode is supported even though the Fabric clock change is not support (assuming the soc will be stay in max power state
+};
+
+/// @brief Contains important information after the mode support steps. Also why a mode is not supported.
+struct dml_mode_support_info_st {
+	//-----------------
+	// Mode Support Information
+	//-----------------
+	dml_bool_t ModeIsSupported; //<brief Is the mode support any voltage and combine setting
+	dml_bool_t ImmediateFlipSupport; //<brief Means mode support immediate flip at the max combine setting; determine in mode support and used in mode programming
+	dml_uint_t MaximumMPCCombine; //<brief If using MPC combine helps the power saving support; then this will be set to 1
+	dml_bool_t UnboundedRequestEnabled;
+	dml_uint_t CompressedBufferSizeInkByte;
+
+	/* Mode Support Reason */
+	dml_bool_t WritebackLatencySupport;
+	dml_bool_t ScaleRatioAndTapsSupport;
+	dml_bool_t SourceFormatPixelAndScanSupport;
+	dml_bool_t MPCCombineMethodIncompatible;
+	dml_bool_t P2IWith420;
+	dml_bool_t DSCOnlyIfNecessaryWithBPP;
+	dml_bool_t DSC422NativeNotSupported;
+	dml_bool_t LinkRateDoesNotMatchDPVersion;
+	dml_bool_t LinkRateForMultistreamNotIndicated;
+	dml_bool_t BPPForMultistreamNotIndicated;
+	dml_bool_t MultistreamWithHDMIOreDP;
+	dml_bool_t MSOOrODMSplitWithNonDPLink;
+	dml_bool_t NotEnoughLanesForMSO;
+	dml_bool_t NumberOfOTGSupport;
+	dml_bool_t NumberOfHDMIFRLSupport;
+	dml_bool_t NumberOfDP2p0Support;
+	dml_bool_t NonsupportedDSCInputBPC;
+	dml_bool_t WritebackScaleRatioAndTapsSupport;
+	dml_bool_t CursorSupport;
+	dml_bool_t PitchSupport;
+	dml_bool_t ViewportExceedsSurface;
+	dml_bool_t ImmediateFlipRequiredButTheRequirementForEachSurfaceIsNotSpecified;
+	dml_bool_t ImmediateFlipOrHostVMAndPStateWithMALLFullFrameOrPhantomPipe;
+	dml_bool_t InvalidCombinationOfMALLUseForPStateAndStaticScreen;
+	dml_bool_t InvalidCombinationOfMALLUseForPState;
+	dml_bool_t ExceededMALLSize;
+	dml_bool_t EnoughWritebackUnits;
+
+	dml_bool_t ExceededMultistreamSlots;
+	dml_bool_t ODMCombineTwoToOneSupportCheckOK;
+	dml_bool_t ODMCombineFourToOneSupportCheckOK;
+	dml_bool_t NotEnoughDSCUnits;
+	dml_bool_t NotEnoughDSCSlices;
+	dml_bool_t PixelsPerLinePerDSCUnitSupport;
+	dml_bool_t DSCCLKRequiredMoreThanSupported;
+	dml_bool_t DTBCLKRequiredMoreThanSupported;
+	dml_bool_t LinkCapacitySupport;
+
+	dml_bool_t ROBSupport[2];
+	dml_bool_t PTEBufferSizeNotExceeded[2];
+	dml_bool_t DCCMetaBufferSizeNotExceeded[2];
+	dml_bool_t TotalVerticalActiveBandwidthSupport[2];
+	enum dml_dram_clock_change_support DRAMClockChangeSupport[2];
+	dml_float_t ActiveDRAMClockChangeLatencyMargin[__DML_NUM_PLANES__];
+	dml_uint_t SubViewportLinesNeededInMALL[__DML_NUM_PLANES__];
+	enum dml_fclock_change_support FCLKChangeSupport[2];
+	dml_bool_t USRRetrainingSupport[2];
+	dml_bool_t VActiveBandwithSupport[2];
+	dml_bool_t PrefetchSupported[2];
+	dml_bool_t DynamicMetadataSupported[2];
+	dml_bool_t VRatioInPrefetchSupported[2];
+	dml_bool_t DISPCLK_DPPCLK_Support[2];
+	dml_bool_t TotalAvailablePipesSupport[2];
+	dml_bool_t ModeSupport[2];
+	dml_bool_t ViewportSizeSupport[2];
+	dml_bool_t ImmediateFlipSupportedForState[2];
+
+	dml_bool_t NoTimeForPrefetch[2][__DML_NUM_PLANES__];
+	dml_bool_t NoTimeForDynamicMetadata[2][__DML_NUM_PLANES__];
+
+	dml_bool_t MPCCombineEnable[__DML_NUM_PLANES__]; /// <brief Indicate if the MPC Combine enable in the given state and optimize mpc combine setting
+	enum dml_odm_mode ODMMode[__DML_NUM_PLANES__]; /// <brief ODM mode that is chosen in the mode check stage and will be used in mode programming stage
+	dml_uint_t DPPPerSurface[__DML_NUM_PLANES__]; /// <brief How many DPPs are needed drive the surface to output. If MPCC or ODMC could be 2 or 4.
+	dml_bool_t DSCEnabled[__DML_NUM_PLANES__]; /// <brief Indicate if the DSC is actually required; used in mode_programming
+	dml_bool_t FECEnabled[__DML_NUM_PLANES__]; /// <brief Indicate if the FEC is actually required
+	dml_uint_t NumberOfDSCSlices[__DML_NUM_PLANES__]; /// <brief Indicate how many slices needed to support the given mode
+
+	dml_float_t OutputBpp[__DML_NUM_PLANES__];
+	enum dml_output_type_and_rate__type OutputType[__DML_NUM_PLANES__];
+	enum dml_output_type_and_rate__rate OutputRate[__DML_NUM_PLANES__];
+
+	dml_float_t AlignedDCCMetaPitchY[__DML_NUM_PLANES__]; /// <brief Pitch value that is aligned to tiling setting
+	dml_float_t AlignedDCCMetaPitchC[__DML_NUM_PLANES__];
+	dml_float_t AlignedYPitch[__DML_NUM_PLANES__];
+	dml_float_t AlignedCPitch[__DML_NUM_PLANES__];
+	dml_float_t MaxTotalVerticalActiveAvailableBandwidth[2]; /// <brief nominal bw available for display
+}; // dml_mode_support_info_st
+
+/// @brief Treat this as the intermediate values and outputs of mode check function. User can query the content of the struct to know more about the result of mode evaluation.
+struct mode_support_st {
+	struct ip_params_st ip;
+	struct soc_bounding_box_st soc;
+	struct soc_state_bounding_box_st state; //<brief Per-state bbox values; only 1 state per compute
+	struct dml_mode_eval_policy_st policy;
+
+	dml_uint_t state_idx; //<brief The power state idx for the power state under this computation
+	dml_uint_t max_state_idx; //<brief The MAX power state idx
+	struct soc_state_bounding_box_st max_state; //<brief The MAX power state; some algo needs to know the max state info to determine if
+	struct dml_display_cfg_st cache_display_cfg; // <brief A copy of the current display cfg in consideration
+
+	// Physical info; only using for programming
+	dml_uint_t num_active_planes; // <brief As determined by either e2e_pipe_param or display_cfg
+
+	// Calculated Clocks
+	dml_float_t RequiredDISPCLK[2]; /// <brief Required DISPCLK; depends on pixel rate; odm mode etc.
+	dml_float_t RequiredDPPCLKThisState[__DML_NUM_PLANES__];
+	dml_float_t DCFCLKState[2]; /// <brief recommended DCFCLK freq; calculated by DML. If UseMinimumRequiredDCFCLK is not set; then it will be just the state DCFCLK; else it will min DCFCLK for support
+	dml_float_t RequiredDISPCLKPerSurface[2][__DML_NUM_PLANES__];
+	dml_float_t RequiredDPPCLKPerSurface[2][__DML_NUM_PLANES__];
+
+	dml_float_t FabricClock; /// <brief Basically just the clock freq at the min (or given) state
+	dml_float_t DRAMSpeed; /// <brief Basically just the clock freq at the min (or given) state
+	dml_float_t SOCCLK; /// <brief Basically just the clock freq at the min (or given) state
+	dml_float_t DCFCLK; /// <brief Basically just the clock freq at the min (or given) state and max combine setting
+	dml_float_t GlobalDPPCLK; /// <brief the Max DPPCLK freq out of all pipes
+
+	// ----------------------------------
+	// Mode Support Info and fail reason
+	// ----------------------------------
+	struct dml_mode_support_info_st support;
+
+	// These are calculated before the ModeSupport and ModeProgram step
+	// They represent the bound for the return buffer sizing
+	dml_uint_t MaxTotalDETInKByte;
+	dml_uint_t NomDETInKByte;
+	dml_uint_t MinCompressedBufferSizeInKByte;
+
+	// Info obtained at the end of mode support calculations
+	// The reported info is at the "optimal" state and combine setting
+	dml_float_t ReturnBW;
+	dml_float_t ReturnDRAMBW;
+	dml_uint_t DETBufferSizeInKByte[__DML_NUM_PLANES__]; // <brief Recommended DET size configuration for this plane. All pipes under this plane should program the DET buffer size to the calculated value.
+	dml_uint_t DETBufferSizeY[__DML_NUM_PLANES__];
+	dml_uint_t DETBufferSizeC[__DML_NUM_PLANES__];
+	dml_uint_t SwathHeightY[__DML_NUM_PLANES__];
+	dml_uint_t SwathHeightC[__DML_NUM_PLANES__];
+
+	// ----------------------------------
+	// Intermediates/Informational
+	// ----------------------------------
+	dml_uint_t TotImmediateFlipBytes;
+	dml_bool_t DCCEnabledInAnySurface;
+	dml_float_t WritebackRequiredDISPCLK;
+	dml_float_t TimeCalc;
+	dml_float_t TWait;
+
+	dml_uint_t SwathWidthYAllStates[2][__DML_NUM_PLANES__];
+	dml_uint_t SwathWidthCAllStates[2][__DML_NUM_PLANES__];
+	dml_uint_t SwathHeightYAllStates[2][__DML_NUM_PLANES__];
+	dml_uint_t SwathHeightCAllStates[2][__DML_NUM_PLANES__];
+	dml_uint_t SwathWidthYThisState[__DML_NUM_PLANES__];
+	dml_uint_t SwathWidthCThisState[__DML_NUM_PLANES__];
+	dml_uint_t SwathHeightYThisState[__DML_NUM_PLANES__];
+	dml_uint_t SwathHeightCThisState[__DML_NUM_PLANES__];
+	dml_uint_t DETBufferSizeInKByteAllStates[2][__DML_NUM_PLANES__];
+	dml_uint_t DETBufferSizeYAllStates[2][__DML_NUM_PLANES__];
+	dml_uint_t DETBufferSizeCAllStates[2][__DML_NUM_PLANES__];
+	dml_bool_t UnboundedRequestEnabledAllStates[2];
+	dml_uint_t CompressedBufferSizeInkByteAllStates[2];
+	dml_bool_t UnboundedRequestEnabledThisState;
+	dml_uint_t CompressedBufferSizeInkByteThisState;
+	dml_uint_t DETBufferSizeInKByteThisState[__DML_NUM_PLANES__];
+	dml_uint_t DETBufferSizeYThisState[__DML_NUM_PLANES__];
+	dml_uint_t DETBufferSizeCThisState[__DML_NUM_PLANES__];
+	dml_float_t VRatioPreY[2][__DML_NUM_PLANES__];
+	dml_float_t VRatioPreC[2][__DML_NUM_PLANES__];
+	dml_uint_t swath_width_luma_ub_all_states[2][__DML_NUM_PLANES__];
+	dml_uint_t swath_width_chroma_ub_all_states[2][__DML_NUM_PLANES__];
+	dml_uint_t swath_width_luma_ub_this_state[__DML_NUM_PLANES__];
+	dml_uint_t swath_width_chroma_ub_this_state[__DML_NUM_PLANES__];
+	dml_uint_t RequiredSlots[__DML_NUM_PLANES__];
+	dml_uint_t PDEAndMetaPTEBytesPerFrame[2][__DML_NUM_PLANES__];
+	dml_uint_t MetaRowBytes[2][__DML_NUM_PLANES__];
+	dml_uint_t DPTEBytesPerRow[2][__DML_NUM_PLANES__];
+	dml_uint_t PrefetchLinesY[2][__DML_NUM_PLANES__];
+	dml_uint_t PrefetchLinesC[2][__DML_NUM_PLANES__];
+	dml_uint_t MaxNumSwY[__DML_NUM_PLANES__]; /// <brief Max number of swath for prefetch
+	dml_uint_t MaxNumSwC[__DML_NUM_PLANES__]; /// <brief Max number of swath for prefetch
+	dml_uint_t PrefillY[__DML_NUM_PLANES__];
+	dml_uint_t PrefillC[__DML_NUM_PLANES__];
+
+	dml_uint_t PrefetchLinesYThisState[__DML_NUM_PLANES__];
+	dml_uint_t PrefetchLinesCThisState[__DML_NUM_PLANES__];
+	dml_uint_t DPTEBytesPerRowThisState[__DML_NUM_PLANES__];
+	dml_uint_t PDEAndMetaPTEBytesPerFrameThisState[__DML_NUM_PLANES__];
+	dml_uint_t MetaRowBytesThisState[__DML_NUM_PLANES__];
+	dml_bool_t use_one_row_for_frame[2][__DML_NUM_PLANES__];
+	dml_bool_t use_one_row_for_frame_flip[2][__DML_NUM_PLANES__];
+	dml_bool_t use_one_row_for_frame_this_state[__DML_NUM_PLANES__];
+	dml_bool_t use_one_row_for_frame_flip_this_state[__DML_NUM_PLANES__];
+
+	dml_float_t LineTimesForPrefetch[__DML_NUM_PLANES__];
+	dml_float_t LinesForMetaPTE[__DML_NUM_PLANES__];
+	dml_float_t LinesForMetaAndDPTERow[__DML_NUM_PLANES__];
+	dml_float_t SwathWidthYSingleDPP[__DML_NUM_PLANES__];
+	dml_float_t SwathWidthCSingleDPP[__DML_NUM_PLANES__];
+	dml_uint_t BytePerPixelY[__DML_NUM_PLANES__];
+	dml_uint_t BytePerPixelC[__DML_NUM_PLANES__];
+	dml_float_t BytePerPixelInDETY[__DML_NUM_PLANES__];
+	dml_float_t BytePerPixelInDETC[__DML_NUM_PLANES__];
+
+	dml_uint_t Read256BlockHeightY[__DML_NUM_PLANES__];
+	dml_uint_t Read256BlockWidthY[__DML_NUM_PLANES__];
+	dml_uint_t Read256BlockHeightC[__DML_NUM_PLANES__];
+	dml_uint_t Read256BlockWidthC[__DML_NUM_PLANES__];
+	dml_uint_t MacroTileHeightY[__DML_NUM_PLANES__];
+	dml_uint_t MacroTileHeightC[__DML_NUM_PLANES__];
+	dml_uint_t MacroTileWidthY[__DML_NUM_PLANES__];
+	dml_uint_t MacroTileWidthC[__DML_NUM_PLANES__];
+	dml_float_t PSCL_FACTOR[__DML_NUM_PLANES__];
+	dml_float_t PSCL_FACTOR_CHROMA[__DML_NUM_PLANES__];
+	dml_float_t MaximumSwathWidthLuma[__DML_NUM_PLANES__];
+	dml_float_t MaximumSwathWidthChroma[__DML_NUM_PLANES__];
+	dml_float_t Tno_bw[__DML_NUM_PLANES__];
+	dml_float_t DestinationLinesToRequestVMInImmediateFlip[__DML_NUM_PLANES__];
+	dml_float_t DestinationLinesToRequestRowInImmediateFlip[__DML_NUM_PLANES__];
+	dml_float_t WritebackDelayTime[__DML_NUM_PLANES__];
+	dml_uint_t dpte_group_bytes[__DML_NUM_PLANES__];
+	dml_uint_t dpte_row_height[__DML_NUM_PLANES__];
+	dml_uint_t dpte_row_height_chroma[__DML_NUM_PLANES__];
+	dml_uint_t meta_row_height[__DML_NUM_PLANES__];
+	dml_uint_t meta_row_height_chroma[__DML_NUM_PLANES__];
+	dml_float_t UrgLatency;
+	dml_float_t UrgentBurstFactorCursor[2][__DML_NUM_PLANES__];
+	dml_float_t UrgentBurstFactorCursorPre[__DML_NUM_PLANES__];
+	dml_float_t UrgentBurstFactorLuma[2][__DML_NUM_PLANES__];
+	dml_float_t UrgentBurstFactorLumaPre[__DML_NUM_PLANES__];
+	dml_float_t UrgentBurstFactorChroma[2][__DML_NUM_PLANES__];
+	dml_float_t UrgentBurstFactorChromaPre[__DML_NUM_PLANES__];
+	dml_float_t MaximumSwathWidthInLineBufferLuma;
+	dml_float_t MaximumSwathWidthInLineBufferChroma;
+	dml_float_t ExtraLatency;
+
+	// Backend
+	dml_bool_t RequiresDSC[__DML_NUM_PLANES__];
+	dml_bool_t RequiresFEC[__DML_NUM_PLANES__];
+	dml_float_t OutputBppPerState[__DML_NUM_PLANES__];
+	dml_uint_t DSCDelayPerState[__DML_NUM_PLANES__];
+	enum dml_output_type_and_rate__type OutputTypePerState[__DML_NUM_PLANES__];
+	enum dml_output_type_and_rate__rate OutputRatePerState[__DML_NUM_PLANES__];
+
+	// Bandwidth Related Info
+	dml_float_t BandwidthAvailableForImmediateFlip;
+	dml_float_t ReadBandwidthLuma[__DML_NUM_PLANES__];
+	dml_float_t ReadBandwidthChroma[__DML_NUM_PLANES__];
+	dml_float_t WriteBandwidth[__DML_NUM_PLANES__];
+	dml_float_t RequiredPrefetchPixelDataBWLuma[__DML_NUM_PLANES__];
+	dml_float_t RequiredPrefetchPixelDataBWChroma[__DML_NUM_PLANES__];
+	dml_float_t cursor_bw[__DML_NUM_PLANES__];
+	dml_float_t cursor_bw_pre[__DML_NUM_PLANES__];
+	dml_float_t prefetch_vmrow_bw[__DML_NUM_PLANES__];
+	dml_float_t final_flip_bw[__DML_NUM_PLANES__];
+	dml_float_t meta_row_bandwidth_this_state[__DML_NUM_PLANES__];
+	dml_float_t dpte_row_bandwidth_this_state[__DML_NUM_PLANES__];
+	dml_float_t ReturnBWPerState[2];
+	dml_float_t ReturnDRAMBWPerState[2];
+	dml_float_t meta_row_bandwidth[2][__DML_NUM_PLANES__];
+	dml_float_t dpte_row_bandwidth[2][__DML_NUM_PLANES__];
+
+	// Something that should be feedback to caller
+	enum dml_odm_mode ODMModePerState[__DML_NUM_PLANES__];
+	enum dml_odm_mode ODMModeThisState[__DML_NUM_PLANES__];
+	dml_uint_t SurfaceSizeInMALL[__DML_NUM_PLANES__];
+	dml_uint_t NoOfDPP[2][__DML_NUM_PLANES__];
+	dml_uint_t NoOfDPPThisState[__DML_NUM_PLANES__];
+	dml_bool_t MPCCombine[2][__DML_NUM_PLANES__];
+	dml_bool_t MPCCombineThisState[__DML_NUM_PLANES__];
+	dml_float_t ProjectedDCFCLKDeepSleep[2];
+	dml_float_t MinDPPCLKUsingSingleDPP[__DML_NUM_PLANES__];
+	dml_bool_t SingleDPPViewportSizeSupportPerSurface[__DML_NUM_PLANES__];
+	dml_bool_t ImmediateFlipSupportedForPipe[__DML_NUM_PLANES__];
+	dml_bool_t NotUrgentLatencyHiding[__DML_NUM_PLANES__];
+	dml_bool_t NotUrgentLatencyHidingPre[__DML_NUM_PLANES__];
+	dml_bool_t PTEBufferSizeNotExceededPerState[__DML_NUM_PLANES__];
+	dml_bool_t DCCMetaBufferSizeNotExceededPerState[__DML_NUM_PLANES__];
+	dml_uint_t PrefetchMode[__DML_NUM_PLANES__];
+	dml_uint_t TotalNumberOfActiveDPP[2];
+	dml_uint_t TotalNumberOfSingleDPPSurfaces[2];
+	dml_uint_t TotalNumberOfDCCActiveDPP[2];
+
+	dml_uint_t SubViewportLinesNeededInMALL[__DML_NUM_PLANES__];
+
+}; // mode_support_st
+
+/// @brief A mega structure that houses various info for model programming step.
+struct mode_program_st {
+
+	//-------------
+	// Intermediate/Informational
+	//-------------
+	dml_float_t UrgentLatency;
+	dml_float_t UrgentLatencyWithUSRRetraining;
+	dml_uint_t VInitPreFillY[__DML_NUM_PLANES__];
+	dml_uint_t VInitPreFillC[__DML_NUM_PLANES__];
+	dml_uint_t MaxNumSwathY[__DML_NUM_PLANES__];
+	dml_uint_t MaxNumSwathC[__DML_NUM_PLANES__];
+
+	dml_float_t BytePerPixelDETY[__DML_NUM_PLANES__];
+	dml_float_t BytePerPixelDETC[__DML_NUM_PLANES__];
+	dml_uint_t BytePerPixelY[__DML_NUM_PLANES__];
+	dml_uint_t BytePerPixelC[__DML_NUM_PLANES__];
+	dml_uint_t SwathWidthY[__DML_NUM_PLANES__];
+	dml_uint_t SwathWidthC[__DML_NUM_PLANES__];
+	dml_uint_t SwathWidthSingleDPPY[__DML_NUM_PLANES__];
+	dml_uint_t SwathWidthSingleDPPC[__DML_NUM_PLANES__];
+	dml_float_t ReadBandwidthSurfaceLuma[__DML_NUM_PLANES__];
+	dml_float_t ReadBandwidthSurfaceChroma[__DML_NUM_PLANES__];
+
+	dml_uint_t PixelPTEBytesPerRow[__DML_NUM_PLANES__];
+	dml_uint_t PDEAndMetaPTEBytesFrame[__DML_NUM_PLANES__];
+	dml_uint_t MetaRowByte[__DML_NUM_PLANES__];
+	dml_uint_t PrefetchSourceLinesY[__DML_NUM_PLANES__];
+	dml_float_t RequiredPrefetchPixDataBWLuma[__DML_NUM_PLANES__];
+	dml_float_t RequiredPrefetchPixDataBWChroma[__DML_NUM_PLANES__];
+	dml_uint_t PrefetchSourceLinesC[__DML_NUM_PLANES__];
+	dml_float_t PSCL_THROUGHPUT[__DML_NUM_PLANES__];
+	dml_float_t PSCL_THROUGHPUT_CHROMA[__DML_NUM_PLANES__];
+	dml_uint_t DSCDelay[__DML_NUM_PLANES__];
+	dml_float_t DPPCLKUsingSingleDPP[__DML_NUM_PLANES__];
+
+	dml_uint_t MacroTileWidthY[__DML_NUM_PLANES__];
+	dml_uint_t MacroTileWidthC[__DML_NUM_PLANES__];
+	dml_uint_t BlockHeight256BytesY[__DML_NUM_PLANES__];
+	dml_uint_t BlockHeight256BytesC[__DML_NUM_PLANES__];
+	dml_uint_t BlockWidth256BytesY[__DML_NUM_PLANES__];
+	dml_uint_t BlockWidth256BytesC[__DML_NUM_PLANES__];
+
+	dml_uint_t BlockHeightY[__DML_NUM_PLANES__];
+	dml_uint_t BlockHeightC[__DML_NUM_PLANES__];
+	dml_uint_t BlockWidthY[__DML_NUM_PLANES__];
+	dml_uint_t BlockWidthC[__DML_NUM_PLANES__];
+
+	dml_uint_t SurfaceSizeInTheMALL[__DML_NUM_PLANES__];
+	dml_float_t VRatioPrefetchY[__DML_NUM_PLANES__];
+	dml_float_t VRatioPrefetchC[__DML_NUM_PLANES__];
+	dml_float_t Tno_bw[__DML_NUM_PLANES__];
+	dml_float_t final_flip_bw[__DML_NUM_PLANES__];
+	dml_float_t prefetch_vmrow_bw[__DML_NUM_PLANES__];
+	dml_float_t cursor_bw[__DML_NUM_PLANES__];
+	dml_float_t cursor_bw_pre[__DML_NUM_PLANES__];
+	dml_float_t WritebackDelay[__DML_NUM_PLANES__];
+	dml_uint_t dpte_row_height[__DML_NUM_PLANES__];
+	dml_uint_t dpte_row_height_linear[__DML_NUM_PLANES__];
+	dml_uint_t meta_req_width[__DML_NUM_PLANES__];
+	dml_uint_t meta_req_height[__DML_NUM_PLANES__];
+	dml_uint_t meta_row_width[__DML_NUM_PLANES__];
+	dml_uint_t meta_row_height[__DML_NUM_PLANES__];
+	dml_uint_t dpte_row_width_luma_ub[__DML_NUM_PLANES__];
+	dml_uint_t dpte_row_width_chroma_ub[__DML_NUM_PLANES__];
+	dml_uint_t dpte_row_height_chroma[__DML_NUM_PLANES__];
+	dml_uint_t dpte_row_height_linear_chroma[__DML_NUM_PLANES__];
+	dml_uint_t meta_req_width_chroma[__DML_NUM_PLANES__];
+	dml_uint_t meta_req_height_chroma[__DML_NUM_PLANES__];
+	dml_uint_t meta_row_width_chroma[__DML_NUM_PLANES__];
+	dml_uint_t meta_row_height_chroma[__DML_NUM_PLANES__];
+	dml_uint_t vm_group_bytes[__DML_NUM_PLANES__];
+	dml_uint_t dpte_group_bytes[__DML_NUM_PLANES__];
+	dml_float_t meta_row_bw[__DML_NUM_PLANES__];
+	dml_float_t dpte_row_bw[__DML_NUM_PLANES__];
+	dml_float_t UrgBurstFactorCursor[__DML_NUM_PLANES__];
+	dml_float_t UrgBurstFactorCursorPre[__DML_NUM_PLANES__];
+	dml_float_t UrgBurstFactorLuma[__DML_NUM_PLANES__];
+	dml_float_t UrgBurstFactorLumaPre[__DML_NUM_PLANES__];
+	dml_float_t UrgBurstFactorChroma[__DML_NUM_PLANES__];
+	dml_float_t UrgBurstFactorChromaPre[__DML_NUM_PLANES__];
+
+	dml_uint_t swath_width_luma_ub[__DML_NUM_PLANES__];
+	dml_uint_t swath_width_chroma_ub[__DML_NUM_PLANES__];
+	dml_uint_t PixelPTEReqWidthY[__DML_NUM_PLANES__];
+	dml_uint_t PixelPTEReqHeightY[__DML_NUM_PLANES__];
+	dml_uint_t PTERequestSizeY[__DML_NUM_PLANES__];
+	dml_uint_t PixelPTEReqWidthC[__DML_NUM_PLANES__];
+	dml_uint_t PixelPTEReqHeightC[__DML_NUM_PLANES__];
+	dml_uint_t PTERequestSizeC[__DML_NUM_PLANES__];
+
+	dml_float_t Tdmdl_vm[__DML_NUM_PLANES__];
+	dml_float_t Tdmdl[__DML_NUM_PLANES__];
+	dml_float_t TSetup[__DML_NUM_PLANES__];
+	dml_uint_t dpde0_bytes_per_frame_ub_l[__DML_NUM_PLANES__];
+	dml_uint_t meta_pte_bytes_per_frame_ub_l[__DML_NUM_PLANES__];
+	dml_uint_t dpde0_bytes_per_frame_ub_c[__DML_NUM_PLANES__];
+	dml_uint_t meta_pte_bytes_per_frame_ub_c[__DML_NUM_PLANES__];
+
+	dml_bool_t UnboundedRequestEnabled;
+	dml_uint_t compbuf_reserved_space_64b;
+	dml_uint_t compbuf_reserved_space_zs;
+	dml_uint_t CompressedBufferSizeInkByte;
+
+	dml_bool_t NoUrgentLatencyHiding[__DML_NUM_PLANES__];
+	dml_bool_t NoUrgentLatencyHidingPre[__DML_NUM_PLANES__];
+	dml_float_t UrgentExtraLatency;
+	dml_bool_t PrefetchAndImmediateFlipSupported;
+	dml_float_t TotalDataReadBandwidth;
+	dml_float_t BandwidthAvailableForImmediateFlip;
+	dml_bool_t NotEnoughTimeForDynamicMetadata[__DML_NUM_PLANES__];
+
+	dml_float_t ReadBandwidthLuma[__DML_NUM_PLANES__];
+	dml_float_t ReadBandwidthChroma[__DML_NUM_PLANES__];
+
+	dml_float_t total_dcn_read_bw_with_flip;
+	dml_float_t total_dcn_read_bw_with_flip_no_urgent_burst;
+	dml_float_t TotalDataReadBandwidthNotIncludingMALLPrefetch;
+	dml_float_t total_dcn_read_bw_with_flip_not_including_MALL_prefetch;
+	dml_float_t non_urgent_total_dcn_read_bw_with_flip;
+	dml_float_t non_urgent_total_dcn_read_bw_with_flip_not_including_MALL_prefetch;
+
+	dml_bool_t use_one_row_for_frame[__DML_NUM_PLANES__];
+	dml_bool_t use_one_row_for_frame_flip[__DML_NUM_PLANES__];
+
+	dml_float_t TCalc;
+	dml_uint_t TotImmediateFlipBytes;
+
+	// -------------------
+	// Output
+	// -------------------
+	dml_uint_t pipe_plane[__DML_NUM_PLANES__]; // <brief used mainly by dv to map the pipe inst to plane index within DML core; the plane idx of a pipe
+	dml_uint_t num_active_pipes;
+
+	dml_bool_t NoTimeToPrefetch[__DML_NUM_PLANES__]; /// <brief Prefetch schedule calculation result
+
+	// Support
+	dml_uint_t PrefetchMode[__DML_NUM_PLANES__]; /// <brief prefetch mode used for prefetch support check in mode programming step
+	dml_bool_t PrefetchModeSupported; /// <brief Is the prefetch mode (bandwidth and latency) supported
+	dml_bool_t ImmediateFlipSupported;
+	dml_bool_t ImmediateFlipSupportedForPipe[__DML_NUM_PLANES__];
+
+	// Clock
+	dml_float_t Dcfclk;
+	dml_float_t Dispclk; /// <brief dispclk being used in mode programming
+	dml_float_t Dppclk[__DML_NUM_PLANES__]; /// <brief dppclk being used in mode programming
+	dml_float_t WritebackDISPCLK;
+	dml_float_t GlobalDPPCLK;
+
+	//@ brief These "calculated" dispclk and dppclk clocks are calculated in the mode programming step.
+	// Depends on the dml_clk_cfg_st option; these calculated values may not used in subsequent calculation.
+	// Possible DV usage: Calculated values fetched by test once after mode_programming step and then possibly
+	// use the values as min and adjust the actual freq used for the 2nd pass
+	dml_float_t Dispclk_calculated;
+	dml_float_t Dppclk_calculated[__DML_NUM_PLANES__];
+
+	dml_float_t DSCCLK_calculated[__DML_NUM_PLANES__]; //< brief Required DSCCLK freq. Backend; not used in any subsequent calculations for now
+	dml_float_t DCFCLKDeepSleep;
+
+	// ARB reg
+	dml_bool_t DCHUBBUB_ARB_CSTATE_MAX_CAP_MODE;
+	struct Watermarks Watermark;
+
+	// DCC compression control
+	dml_uint_t DCCYMaxUncompressedBlock[__DML_NUM_PLANES__];
+	dml_uint_t DCCYMaxCompressedBlock[__DML_NUM_PLANES__];
+	dml_uint_t DCCYIndependentBlock[__DML_NUM_PLANES__];
+	dml_uint_t DCCCMaxUncompressedBlock[__DML_NUM_PLANES__];
+	dml_uint_t DCCCMaxCompressedBlock[__DML_NUM_PLANES__];
+	dml_uint_t DCCCIndependentBlock[__DML_NUM_PLANES__];
+
+	// Stutter Efficiency
+	dml_float_t StutterEfficiency;
+	dml_float_t StutterEfficiencyNotIncludingVBlank;
+	dml_uint_t NumberOfStutterBurstsPerFrame;
+	dml_float_t Z8StutterEfficiency;
+	dml_uint_t Z8NumberOfStutterBurstsPerFrame;
+	dml_float_t Z8StutterEfficiencyNotIncludingVBlank;
+	dml_float_t StutterPeriod;
+	dml_float_t Z8StutterEfficiencyBestCase;
+	dml_uint_t Z8NumberOfStutterBurstsPerFrameBestCase;
+	dml_float_t Z8StutterEfficiencyNotIncludingVBlankBestCase;
+	dml_float_t StutterPeriodBestCase;
+
+	// DLG TTU reg
+	dml_float_t MIN_DST_Y_NEXT_START[__DML_NUM_PLANES__];
+	dml_bool_t VREADY_AT_OR_AFTER_VSYNC[__DML_NUM_PLANES__];
+	dml_uint_t DSTYAfterScaler[__DML_NUM_PLANES__];
+	dml_uint_t DSTXAfterScaler[__DML_NUM_PLANES__];
+	dml_float_t DestinationLinesForPrefetch[__DML_NUM_PLANES__];
+	dml_float_t DestinationLinesToRequestVMInVBlank[__DML_NUM_PLANES__];
+	dml_float_t DestinationLinesToRequestRowInVBlank[__DML_NUM_PLANES__];
+	dml_float_t DestinationLinesToRequestVMInImmediateFlip[__DML_NUM_PLANES__];
+	dml_float_t DestinationLinesToRequestRowInImmediateFlip[__DML_NUM_PLANES__];
+	dml_float_t MinTTUVBlank[__DML_NUM_PLANES__];
+	dml_float_t DisplayPipeLineDeliveryTimeLuma[__DML_NUM_PLANES__];
+	dml_float_t DisplayPipeLineDeliveryTimeChroma[__DML_NUM_PLANES__];
+	dml_float_t DisplayPipeLineDeliveryTimeLumaPrefetch[__DML_NUM_PLANES__];
+	dml_float_t DisplayPipeLineDeliveryTimeChromaPrefetch[__DML_NUM_PLANES__];
+	dml_float_t DisplayPipeRequestDeliveryTimeLuma[__DML_NUM_PLANES__];
+	dml_float_t DisplayPipeRequestDeliveryTimeChroma[__DML_NUM_PLANES__];
+	dml_float_t DisplayPipeRequestDeliveryTimeLumaPrefetch[__DML_NUM_PLANES__];
+	dml_float_t DisplayPipeRequestDeliveryTimeChromaPrefetch[__DML_NUM_PLANES__];
+	dml_float_t CursorRequestDeliveryTime[__DML_NUM_PLANES__];
+	dml_float_t CursorRequestDeliveryTimePrefetch[__DML_NUM_PLANES__];
+
+	dml_float_t DST_Y_PER_PTE_ROW_NOM_L[__DML_NUM_PLANES__];
+	dml_float_t DST_Y_PER_PTE_ROW_NOM_C[__DML_NUM_PLANES__];
+	dml_float_t DST_Y_PER_META_ROW_NOM_L[__DML_NUM_PLANES__];
+	dml_float_t DST_Y_PER_META_ROW_NOM_C[__DML_NUM_PLANES__];
+	dml_float_t TimePerMetaChunkNominal[__DML_NUM_PLANES__];
+	dml_float_t TimePerChromaMetaChunkNominal[__DML_NUM_PLANES__];
+	dml_float_t TimePerMetaChunkVBlank[__DML_NUM_PLANES__];
+	dml_float_t TimePerChromaMetaChunkVBlank[__DML_NUM_PLANES__];
+	dml_float_t TimePerMetaChunkFlip[__DML_NUM_PLANES__];
+	dml_float_t TimePerChromaMetaChunkFlip[__DML_NUM_PLANES__];
+	dml_float_t time_per_pte_group_nom_luma[__DML_NUM_PLANES__];
+	dml_float_t time_per_pte_group_nom_chroma[__DML_NUM_PLANES__];
+	dml_float_t time_per_pte_group_vblank_luma[__DML_NUM_PLANES__];
+	dml_float_t time_per_pte_group_vblank_chroma[__DML_NUM_PLANES__];
+	dml_float_t time_per_pte_group_flip_luma[__DML_NUM_PLANES__];
+	dml_float_t time_per_pte_group_flip_chroma[__DML_NUM_PLANES__];
+	dml_float_t TimePerVMGroupVBlank[__DML_NUM_PLANES__];
+	dml_float_t TimePerVMGroupFlip[__DML_NUM_PLANES__];
+	dml_float_t TimePerVMRequestVBlank[__DML_NUM_PLANES__];
+	dml_float_t TimePerVMRequestFlip[__DML_NUM_PLANES__];
+
+	dml_float_t FractionOfUrgentBandwidth;
+	dml_float_t FractionOfUrgentBandwidthImmediateFlip;
+
+	// RQ registers
+	dml_bool_t PTE_BUFFER_MODE[__DML_NUM_PLANES__];
+	dml_uint_t BIGK_FRAGMENT_SIZE[__DML_NUM_PLANES__];
+
+	dml_uint_t SubViewportLinesNeededInMALL[__DML_NUM_PLANES__];
+	dml_bool_t UsesMALLForStaticScreen[__DML_NUM_PLANES__];
+
+	// OTG
+	dml_uint_t VStartupMin[__DML_NUM_PLANES__]; /// <brief Minimum vstartup to meet the prefetch schedule (i.e. the prefetch solution can be found at this vstartup time); not the actual global sync vstartup pos.
+	dml_uint_t VStartup[__DML_NUM_PLANES__]; /// <brief The vstartup value for OTG programming (will set to max vstartup; but now bounded by min(vblank_nom. actual vblank))
+	dml_uint_t VUpdateOffsetPix[__DML_NUM_PLANES__];
+	dml_uint_t VUpdateWidthPix[__DML_NUM_PLANES__];
+	dml_uint_t VReadyOffsetPix[__DML_NUM_PLANES__];
+
+	// Latency and Support
+	dml_float_t MaxActiveFCLKChangeLatencySupported;
+	dml_bool_t USRRetrainingSupport;
+	enum dml_fclock_change_support FCLKChangeSupport;
+	enum dml_dram_clock_change_support DRAMClockChangeSupport;
+	dml_float_t MaxActiveDRAMClockChangeLatencySupported[__DML_NUM_PLANES__];
+	dml_float_t WritebackAllowFCLKChangeEndPosition[__DML_NUM_PLANES__];
+	dml_float_t WritebackAllowDRAMClockChangeEndPosition[__DML_NUM_PLANES__];
+
+	// buffer sizing
+	dml_uint_t DETBufferSizeInKByte[__DML_NUM_PLANES__];  // <brief Recommended DET size configuration for this plane.  All pipes under this plane should program the DET buffer size to the calculated value.
+	dml_uint_t DETBufferSizeY[__DML_NUM_PLANES__];
+	dml_uint_t DETBufferSizeC[__DML_NUM_PLANES__];
+	dml_uint_t SwathHeightY[__DML_NUM_PLANES__];
+	dml_uint_t SwathHeightC[__DML_NUM_PLANES__];
+}; // mode_program_st
+
+struct soc_states_st {
+	dml_uint_t num_states; /// <brief num of soc pwr states
+	struct soc_state_bounding_box_st state_array[__DML_MAX_STATE_ARRAY_SIZE__]; /// <brief fixed size array that holds states struct
+};
+
+struct UseMinimumDCFCLK_params_st {
+	enum dml_use_mall_for_pstate_change_mode *UseMALLForPStateChange;
+	dml_bool_t *DRRDisplay;
+	dml_bool_t SynchronizeDRRDisplaysForUCLKPStateChangeFinal;
+	dml_uint_t MaxInterDCNTileRepeaters;
+	dml_uint_t MaxPrefetchMode;
+	dml_float_t DRAMClockChangeLatencyFinal;
+	dml_float_t FCLKChangeLatency;
+	dml_float_t SREnterPlusExitTime;
+	dml_uint_t ReturnBusWidth;
+	dml_uint_t RoundTripPingLatencyCycles;
+	dml_uint_t ReorderingBytes;
+	dml_uint_t PixelChunkSizeInKByte;
+	dml_uint_t MetaChunkSize;
+	dml_bool_t GPUVMEnable;
+	dml_uint_t GPUVMMaxPageTableLevels;
+	dml_bool_t HostVMEnable;
+	dml_uint_t NumberOfActiveSurfaces;
+	dml_uint_t HostVMMinPageSize;
+	dml_uint_t HostVMMaxNonCachedPageTableLevels;
+	dml_bool_t DynamicMetadataVMEnabled;
+	dml_bool_t ImmediateFlipRequirement;
+	dml_bool_t ProgressiveToInterlaceUnitInOPP;
+	dml_float_t MaxAveragePercentOfIdealSDPPortBWDisplayCanUseInNormalSystemOperation;
+	dml_float_t PercentOfIdealSDPPortBWReceivedAfterUrgLatency;
+	dml_uint_t *VTotal;
+	dml_uint_t *VActive;
+	dml_uint_t *DynamicMetadataTransmittedBytes;
+	dml_uint_t *DynamicMetadataLinesBeforeActiveRequired;
+	dml_bool_t *Interlace;
+	dml_float_t (*RequiredDPPCLKPerSurface)[__DML_NUM_PLANES__];
+	dml_float_t *RequiredDISPCLK;
+	dml_float_t UrgLatency;
+	dml_uint_t (*NoOfDPP)[__DML_NUM_PLANES__];
+	dml_float_t *ProjectedDCFCLKDeepSleep;
+	dml_uint_t (*MaximumVStartup)[__DML_NUM_PLANES__];
+	dml_uint_t *TotalNumberOfActiveDPP;
+	dml_uint_t *TotalNumberOfDCCActiveDPP;
+	dml_uint_t *dpte_group_bytes;
+	dml_uint_t (*PrefetchLinesY)[__DML_NUM_PLANES__];
+	dml_uint_t (*PrefetchLinesC)[__DML_NUM_PLANES__];
+	dml_uint_t (*swath_width_luma_ub_all_states)[__DML_NUM_PLANES__];
+	dml_uint_t (*swath_width_chroma_ub_all_states)[__DML_NUM_PLANES__];
+	dml_uint_t *BytePerPixelY;
+	dml_uint_t *BytePerPixelC;
+	dml_uint_t *HTotal;
+	dml_float_t *PixelClock;
+	dml_uint_t (*PDEAndMetaPTEBytesPerFrame)[__DML_NUM_PLANES__];
+	dml_uint_t (*DPTEBytesPerRow)[__DML_NUM_PLANES__];
+	dml_uint_t (*MetaRowBytes)[__DML_NUM_PLANES__];
+	dml_bool_t *DynamicMetadataEnable;
+	dml_float_t *ReadBandwidthLuma;
+	dml_float_t *ReadBandwidthChroma;
+	dml_float_t DCFCLKPerState;
+	dml_float_t *DCFCLKState;
+};
+
+struct CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_params_st {
+	dml_bool_t USRRetrainingRequiredFinal;
+	enum dml_use_mall_for_pstate_change_mode *UseMALLForPStateChange;
+	dml_uint_t *PrefetchMode;
+	dml_uint_t NumberOfActiveSurfaces;
+	dml_uint_t MaxLineBufferLines;
+	dml_uint_t LineBufferSize;
+	dml_uint_t WritebackInterfaceBufferSize;
+	dml_float_t DCFCLK;
+	dml_float_t ReturnBW;
+	dml_bool_t SynchronizeTimingsFinal;
+	dml_bool_t SynchronizeDRRDisplaysForUCLKPStateChangeFinal;
+	dml_bool_t *DRRDisplay;
+	dml_uint_t *dpte_group_bytes;
+	dml_uint_t *meta_row_height;
+	dml_uint_t *meta_row_height_chroma;
+	struct SOCParametersList mmSOCParameters;
+	dml_uint_t WritebackChunkSize;
+	dml_float_t SOCCLK;
+	dml_float_t DCFClkDeepSleep;
+	dml_uint_t *DETBufferSizeY;
+	dml_uint_t *DETBufferSizeC;
+	dml_uint_t *SwathHeightY;
+	dml_uint_t *SwathHeightC;
+	dml_uint_t *LBBitPerPixel;
+	dml_uint_t *SwathWidthY;
+	dml_uint_t *SwathWidthC;
+	dml_float_t *HRatio;
+	dml_float_t *HRatioChroma;
+	dml_uint_t *VTaps;
+	dml_uint_t *VTapsChroma;
+	dml_float_t *VRatio;
+	dml_float_t *VRatioChroma;
+	dml_uint_t *HTotal;
+	dml_uint_t *VTotal;
+	dml_uint_t *VActive;
+	dml_float_t *PixelClock;
+	dml_uint_t *BlendingAndTiming;
+	dml_uint_t *DPPPerSurface;
+	dml_float_t *BytePerPixelDETY;
+	dml_float_t *BytePerPixelDETC;
+	dml_uint_t *DSTXAfterScaler;
+	dml_uint_t *DSTYAfterScaler;
+	dml_bool_t *WritebackEnable;
+	enum dml_source_format_class *WritebackPixelFormat;
+	dml_uint_t *WritebackDestinationWidth;
+	dml_uint_t *WritebackDestinationHeight;
+	dml_uint_t *WritebackSourceHeight;
+	dml_bool_t UnboundedRequestEnabled;
+	dml_uint_t CompressedBufferSizeInkByte;
+
+	// Output
+	struct Watermarks *Watermark;
+	enum dml_dram_clock_change_support *DRAMClockChangeSupport;
+	dml_float_t *MaxActiveDRAMClockChangeLatencySupported;
+	dml_uint_t *SubViewportLinesNeededInMALL;
+	enum dml_fclock_change_support *FCLKChangeSupport;
+	dml_float_t *MaxActiveFCLKChangeLatencySupported;
+	dml_bool_t *USRRetrainingSupport;
+	dml_float_t *ActiveDRAMClockChangeLatencyMargin;
+};
+
+struct CalculateVMRowAndSwath_params_st {
+	dml_uint_t NumberOfActiveSurfaces;
+	struct DmlPipe *myPipe;
+	dml_uint_t *SurfaceSizeInMALL;
+	dml_uint_t PTEBufferSizeInRequestsLuma;
+	dml_uint_t PTEBufferSizeInRequestsChroma;
+	dml_uint_t DCCMetaBufferSizeBytes;
+	enum dml_use_mall_for_static_screen_mode *UseMALLForStaticScreen;
+	enum dml_use_mall_for_pstate_change_mode *UseMALLForPStateChange;
+	dml_uint_t MALLAllocatedForDCN;
+	dml_uint_t *SwathWidthY;
+	dml_uint_t *SwathWidthC;
+	dml_bool_t GPUVMEnable;
+	dml_bool_t HostVMEnable;
+	dml_uint_t HostVMMaxNonCachedPageTableLevels;
+	dml_uint_t GPUVMMaxPageTableLevels;
+	dml_uint_t *GPUVMMinPageSizeKBytes;
+	dml_uint_t HostVMMinPageSize;
+	dml_bool_t *PTEBufferModeOverrideEn;
+	dml_bool_t *PTEBufferModeOverrideVal;
+
+	// Output
+	dml_bool_t *PTEBufferSizeNotExceeded;
+	dml_bool_t *DCCMetaBufferSizeNotExceeded;
+	dml_uint_t *dpte_row_width_luma_ub;
+	dml_uint_t *dpte_row_width_chroma_ub;
+	dml_uint_t *dpte_row_height_luma;
+	dml_uint_t *dpte_row_height_chroma;
+	dml_uint_t *dpte_row_height_linear_luma; // VBA_DELTA
+	dml_uint_t *dpte_row_height_linear_chroma; // VBA_DELTA
+	dml_uint_t *meta_req_width;
+	dml_uint_t *meta_req_width_chroma;
+	dml_uint_t *meta_req_height;
+	dml_uint_t *meta_req_height_chroma;
+	dml_uint_t *meta_row_width;
+	dml_uint_t *meta_row_width_chroma;
+	dml_uint_t *meta_row_height;
+	dml_uint_t *meta_row_height_chroma;
+	dml_uint_t *vm_group_bytes;
+	dml_uint_t *dpte_group_bytes;
+	dml_uint_t *PixelPTEReqWidthY;
+	dml_uint_t *PixelPTEReqHeightY;
+	dml_uint_t *PTERequestSizeY;
+	dml_uint_t *PixelPTEReqWidthC;
+	dml_uint_t *PixelPTEReqHeightC;
+	dml_uint_t *PTERequestSizeC;
+	dml_uint_t *dpde0_bytes_per_frame_ub_l;
+	dml_uint_t *meta_pte_bytes_per_frame_ub_l;
+	dml_uint_t *dpde0_bytes_per_frame_ub_c;
+	dml_uint_t *meta_pte_bytes_per_frame_ub_c;
+	dml_uint_t *PrefetchSourceLinesY;
+	dml_uint_t *PrefetchSourceLinesC;
+	dml_uint_t *VInitPreFillY;
+	dml_uint_t *VInitPreFillC;
+	dml_uint_t *MaxNumSwathY;
+	dml_uint_t *MaxNumSwathC;
+	dml_float_t *meta_row_bw;
+	dml_float_t *dpte_row_bw;
+	dml_uint_t *PixelPTEBytesPerRow;
+	dml_uint_t *PDEAndMetaPTEBytesFrame;
+	dml_uint_t *MetaRowByte;
+	dml_bool_t *use_one_row_for_frame;
+	dml_bool_t *use_one_row_for_frame_flip;
+	dml_bool_t *UsesMALLForStaticScreen;
+	dml_bool_t *PTE_BUFFER_MODE;
+	dml_uint_t *BIGK_FRAGMENT_SIZE;
+};
+
+struct CalculateSwathAndDETConfiguration_params_st {
+	dml_uint_t *DETSizeOverride;
+	enum dml_use_mall_for_pstate_change_mode *UseMALLForPStateChange;
+	dml_uint_t ConfigReturnBufferSizeInKByte;
+	dml_uint_t ROBBufferSizeInKByte;
+	dml_uint_t MaxTotalDETInKByte;
+	dml_uint_t MinCompressedBufferSizeInKByte;
+	dml_uint_t PixelChunkSizeInKByte;
+	dml_bool_t ForceSingleDPP;
+	dml_uint_t NumberOfActiveSurfaces;
+	dml_uint_t nomDETInKByte;
+	enum dml_unbounded_requesting_policy UseUnboundedRequestingFinal;
+	dml_uint_t ConfigReturnBufferSegmentSizeInkByte;
+	dml_uint_t CompressedBufferSegmentSizeInkByteFinal;
+	enum dml_output_encoder_class *Output;
+	dml_float_t *ReadBandwidthLuma;
+	dml_float_t *ReadBandwidthChroma;
+	dml_float_t *MaximumSwathWidthLuma;
+	dml_float_t *MaximumSwathWidthChroma;
+	enum dml_rotation_angle *SourceScan;
+	dml_bool_t *ViewportStationary;
+	enum dml_source_format_class *SourcePixelFormat;
+	enum dml_swizzle_mode *SurfaceTiling;
+	dml_uint_t *ViewportWidth;
+	dml_uint_t *ViewportHeight;
+	dml_uint_t *ViewportXStart;
+	dml_uint_t *ViewportYStart;
+	dml_uint_t *ViewportXStartC;
+	dml_uint_t *ViewportYStartC;
+	dml_uint_t *SurfaceWidthY;
+	dml_uint_t *SurfaceWidthC;
+	dml_uint_t *SurfaceHeightY;
+	dml_uint_t *SurfaceHeightC;
+	dml_uint_t *Read256BytesBlockHeightY;
+	dml_uint_t *Read256BytesBlockHeightC;
+	dml_uint_t *Read256BytesBlockWidthY;
+	dml_uint_t *Read256BytesBlockWidthC;
+	enum dml_odm_mode *ODMMode;
+	dml_uint_t *BlendingAndTiming;
+	dml_uint_t *BytePerPixY;
+	dml_uint_t *BytePerPixC;
+	dml_float_t *BytePerPixDETY;
+	dml_float_t *BytePerPixDETC;
+	dml_uint_t *HActive;
+	dml_float_t *HRatio;
+	dml_float_t *HRatioChroma;
+	dml_uint_t *DPPPerSurface;
+	dml_uint_t *swath_width_luma_ub;
+	dml_uint_t *swath_width_chroma_ub;
+	dml_uint_t *SwathWidth;
+	dml_uint_t *SwathWidthChroma;
+	dml_uint_t *SwathHeightY;
+	dml_uint_t *SwathHeightC;
+	dml_uint_t *DETBufferSizeInKByte;
+	dml_uint_t *DETBufferSizeY;
+	dml_uint_t *DETBufferSizeC;
+	dml_bool_t *UnboundedRequestEnabled;
+	dml_uint_t *compbuf_reserved_space_64b;
+	dml_uint_t *compbuf_reserved_space_zs;
+	dml_uint_t *CompressedBufferSizeInkByte;
+	dml_bool_t *ViewportSizeSupportPerSurface;
+	dml_bool_t *ViewportSizeSupport;
+};
+
+struct CalculateStutterEfficiency_params_st {
+	dml_uint_t CompressedBufferSizeInkByte;
+	enum dml_use_mall_for_pstate_change_mode *UseMALLForPStateChange;
+	dml_bool_t UnboundedRequestEnabled;
+	dml_uint_t MetaFIFOSizeInKEntries;
+	dml_uint_t ZeroSizeBufferEntries;
+	dml_uint_t PixelChunkSizeInKByte;
+	dml_uint_t NumberOfActiveSurfaces;
+	dml_uint_t ROBBufferSizeInKByte;
+	dml_float_t TotalDataReadBandwidth;
+	dml_float_t DCFCLK;
+	dml_float_t ReturnBW;
+	dml_uint_t CompbufReservedSpace64B;
+	dml_uint_t CompbufReservedSpaceZs;
+	dml_float_t SRExitTime;
+	dml_float_t SRExitZ8Time;
+	dml_bool_t SynchronizeTimingsFinal;
+	dml_uint_t *BlendingAndTiming;
+	dml_float_t StutterEnterPlusExitWatermark;
+	dml_float_t Z8StutterEnterPlusExitWatermark;
+	dml_bool_t ProgressiveToInterlaceUnitInOPP;
+	dml_bool_t *Interlace;
+	dml_float_t *MinTTUVBlank;
+	dml_uint_t *DPPPerSurface;
+	dml_uint_t *DETBufferSizeY;
+	dml_uint_t *BytePerPixelY;
+	dml_float_t *BytePerPixelDETY;
+	dml_uint_t *SwathWidthY;
+	dml_uint_t *SwathHeightY;
+	dml_uint_t *SwathHeightC;
+	dml_float_t *NetDCCRateLuma;
+	dml_float_t *NetDCCRateChroma;
+	dml_float_t *DCCFractionOfZeroSizeRequestsLuma;
+	dml_float_t *DCCFractionOfZeroSizeRequestsChroma;
+	dml_uint_t *HTotal;
+	dml_uint_t *VTotal;
+	dml_float_t *PixelClock;
+	dml_float_t *VRatio;
+	enum dml_rotation_angle *SourceScan;
+	dml_uint_t *BlockHeight256BytesY;
+	dml_uint_t *BlockWidth256BytesY;
+	dml_uint_t *BlockHeight256BytesC;
+	dml_uint_t *BlockWidth256BytesC;
+	dml_uint_t *DCCYMaxUncompressedBlock;
+	dml_uint_t *DCCCMaxUncompressedBlock;
+	dml_uint_t *VActive;
+	dml_bool_t *DCCEnable;
+	dml_bool_t *WritebackEnable;
+	dml_float_t *ReadBandwidthSurfaceLuma;
+	dml_float_t *ReadBandwidthSurfaceChroma;
+	dml_float_t *meta_row_bw;
+	dml_float_t *dpte_row_bw;
+	dml_float_t *StutterEfficiencyNotIncludingVBlank;
+	dml_float_t *StutterEfficiency;
+	dml_uint_t *NumberOfStutterBurstsPerFrame;
+	dml_float_t *Z8StutterEfficiencyNotIncludingVBlank;
+	dml_float_t *Z8StutterEfficiency;
+	dml_uint_t *Z8NumberOfStutterBurstsPerFrame;
+	dml_float_t *StutterPeriod;
+	dml_bool_t *DCHUBBUB_ARB_CSTATE_MAX_CAP_MODE;
+};
+
+struct CalculatePrefetchSchedule_params_st {
+	dml_bool_t EnhancedPrefetchScheduleAccelerationFinal;
+	dml_float_t HostVMInefficiencyFactor;
+	struct DmlPipe *myPipe;
+	dml_uint_t DSCDelay;
+	dml_float_t DPPCLKDelaySubtotalPlusCNVCFormater;
+	dml_float_t DPPCLKDelaySCL;
+	dml_float_t DPPCLKDelaySCLLBOnly;
+	dml_float_t DPPCLKDelayCNVCCursor;
+	dml_float_t DISPCLKDelaySubtotal;
+	dml_uint_t DPP_RECOUT_WIDTH;
+	enum dml_output_format_class OutputFormat;
+	dml_uint_t MaxInterDCNTileRepeaters;
+	dml_uint_t VStartup;
+	dml_uint_t MaxVStartup;
+	dml_uint_t GPUVMPageTableLevels;
+	dml_bool_t GPUVMEnable;
+	dml_bool_t HostVMEnable;
+	dml_uint_t HostVMMaxNonCachedPageTableLevels;
+	dml_uint_t HostVMMinPageSize;
+	dml_bool_t DynamicMetadataEnable;
+	dml_bool_t DynamicMetadataVMEnabled;
+	int DynamicMetadataLinesBeforeActiveRequired;
+	dml_uint_t DynamicMetadataTransmittedBytes;
+	dml_float_t UrgentLatency;
+	dml_float_t UrgentExtraLatency;
+	dml_float_t TCalc;
+	dml_uint_t PDEAndMetaPTEBytesFrame;
+	dml_uint_t MetaRowByte;
+	dml_uint_t PixelPTEBytesPerRow;
+	dml_float_t PrefetchSourceLinesY;
+	dml_uint_t VInitPreFillY;
+	dml_uint_t MaxNumSwathY;
+	dml_float_t PrefetchSourceLinesC;
+	dml_uint_t VInitPreFillC;
+	dml_uint_t MaxNumSwathC;
+	dml_uint_t swath_width_luma_ub;
+	dml_uint_t swath_width_chroma_ub;
+	dml_uint_t SwathHeightY;
+	dml_uint_t SwathHeightC;
+	dml_float_t TWait;
+	dml_uint_t *DSTXAfterScaler;
+	dml_uint_t *DSTYAfterScaler;
+	dml_float_t *DestinationLinesForPrefetch;
+	dml_float_t *DestinationLinesToRequestVMInVBlank;
+	dml_float_t *DestinationLinesToRequestRowInVBlank;
+	dml_float_t *VRatioPrefetchY;
+	dml_float_t *VRatioPrefetchC;
+	dml_float_t *RequiredPrefetchPixDataBWLuma;
+	dml_float_t *RequiredPrefetchPixDataBWChroma;
+	dml_bool_t *NotEnoughTimeForDynamicMetadata;
+	dml_float_t *Tno_bw;
+	dml_float_t *prefetch_vmrow_bw;
+	dml_float_t *Tdmdl_vm;
+	dml_float_t *Tdmdl;
+	dml_float_t *TSetup;
+	dml_uint_t *VUpdateOffsetPix;
+	dml_uint_t *VUpdateWidthPix;
+	dml_uint_t *VReadyOffsetPix;
+};
+
+struct dml_core_mode_support_locals_st {
+	dml_bool_t dummy_boolean[2];
+	dml_uint_t dummy_integer[3];
+	dml_uint_t dummy_integer_array[22][__DML_NUM_PLANES__];
+	enum dml_odm_mode dummy_odm_mode[__DML_NUM_PLANES__];
+	dml_bool_t dummy_boolean_array[2][__DML_NUM_PLANES__];
+	dml_uint_t MaxVStartupAllPlanes[2];
+	dml_uint_t MaximumVStartup[2][__DML_NUM_PLANES__];
+	dml_uint_t DSTYAfterScaler[__DML_NUM_PLANES__];
+	dml_uint_t DSTXAfterScaler[__DML_NUM_PLANES__];
+	dml_uint_t NextPrefetchMode[__DML_NUM_PLANES__];
+	dml_uint_t MinPrefetchMode[__DML_NUM_PLANES__];
+	dml_uint_t MaxPrefetchMode[__DML_NUM_PLANES__];
+	dml_float_t dummy_single[3];
+	dml_float_t dummy_single_array[__DML_NUM_PLANES__];
+	struct Watermarks dummy_watermark;
+	struct SOCParametersList mSOCParameters;
+	struct DmlPipe myPipe;
+	struct DmlPipe SurfParameters[__DML_NUM_PLANES__];
+	dml_uint_t TotalNumberOfActiveWriteback;
+	dml_uint_t MaximumSwathWidthSupportLuma;
+	dml_uint_t MaximumSwathWidthSupportChroma;
+	dml_bool_t MPCCombineMethodAsNeededForPStateChangeAndVoltage;
+	dml_bool_t MPCCombineMethodAsPossible;
+	dml_bool_t TotalAvailablePipesSupportNoDSC;
+	dml_uint_t NumberOfDPPNoDSC;
+	enum dml_odm_mode ODMModeNoDSC;
+	dml_float_t RequiredDISPCLKPerSurfaceNoDSC;
+	dml_bool_t TotalAvailablePipesSupportDSC;
+	dml_uint_t NumberOfDPPDSC;
+	enum dml_odm_mode ODMModeDSC;
+	dml_float_t RequiredDISPCLKPerSurfaceDSC;
+	dml_bool_t NoChromaOrLinear;
+	dml_float_t BWOfNonCombinedSurfaceOfMaximumBandwidth;
+	dml_uint_t NumberOfNonCombinedSurfaceOfMaximumBandwidth;
+	dml_uint_t TotalNumberOfActiveOTG;
+	dml_uint_t TotalNumberOfActiveHDMIFRL;
+	dml_uint_t TotalNumberOfActiveDP2p0;
+	dml_uint_t TotalNumberOfActiveDP2p0Outputs;
+	dml_uint_t TotalSlots;
+	dml_uint_t DSCFormatFactor;
+	dml_uint_t TotalDSCUnitsRequired;
+	dml_uint_t ReorderingBytes;
+	dml_bool_t ImmediateFlipRequiredFinal;
+	dml_bool_t FullFrameMALLPStateMethod;
+	dml_bool_t SubViewportMALLPStateMethod;
+	dml_bool_t PhantomPipeMALLPStateMethod;
+	dml_bool_t SubViewportMALLRefreshGreaterThan120Hz;
+	dml_float_t MaxTotalVActiveRDBandwidth;
+	dml_float_t VMDataOnlyReturnBWPerState;
+	dml_float_t HostVMInefficiencyFactor;
+	dml_uint_t NextMaxVStartup;
+	dml_uint_t MaxVStartup;
+	dml_bool_t AllPrefetchModeTested;
+	dml_bool_t AnyLinesForVMOrRowTooLarge;
+	dml_bool_t is_max_pwr_state;
+	dml_bool_t is_max_dram_pwr_state;
+	dml_bool_t dram_clock_change_support;
+	dml_bool_t f_clock_change_support;
+};
+
+struct dml_core_mode_programming_locals_st {
+	dml_uint_t DSCFormatFactor;
+	dml_uint_t dummy_integer_array[2][__DML_NUM_PLANES__];
+	enum dml_output_encoder_class dummy_output_encoder_array[__DML_NUM_PLANES__];
+	dml_float_t dummy_single_array[2][__DML_NUM_PLANES__];
+	dml_uint_t dummy_long_array[4][__DML_NUM_PLANES__];
+	dml_bool_t dummy_boolean_array[2][__DML_NUM_PLANES__];
+	dml_bool_t dummy_boolean[1];
+	struct DmlPipe SurfaceParameters[__DML_NUM_PLANES__];
+	dml_uint_t ReorderBytes;
+	dml_float_t VMDataOnlyReturnBW;
+	dml_float_t HostVMInefficiencyFactor;
+	dml_uint_t TotalDCCActiveDPP;
+	dml_uint_t TotalActiveDPP;
+	dml_uint_t VStartupLines;
+	dml_uint_t MaxVStartupLines[__DML_NUM_PLANES__]; /// <brief more like vblank for the plane's OTG
+	dml_uint_t MaxVStartupAllPlanes;
+	dml_bool_t ImmediateFlipRequirementFinal;
+	int iteration;
+	dml_float_t MaxTotalRDBandwidth;
+	dml_float_t MaxTotalRDBandwidthNoUrgentBurst;
+	dml_bool_t DestinationLineTimesForPrefetchLessThan2;
+	dml_bool_t VRatioPrefetchMoreThanMax;
+	dml_float_t MaxTotalRDBandwidthNotIncludingMALLPrefetch;
+	dml_uint_t NextPrefetchMode[__DML_NUM_PLANES__];
+	dml_uint_t MinPrefetchMode[__DML_NUM_PLANES__];
+	dml_uint_t MaxPrefetchMode[__DML_NUM_PLANES__];
+	dml_bool_t AllPrefetchModeTested;
+	dml_float_t dummy_unit_vector[__DML_NUM_PLANES__];
+	dml_float_t NonUrgentMaxTotalRDBandwidth;
+	dml_float_t NonUrgentMaxTotalRDBandwidthNotIncludingMALLPrefetch;
+	dml_float_t dummy_single[2];
+	struct SOCParametersList mmSOCParameters;
+	dml_float_t Tvstartup_margin;
+	dml_float_t dlg_vblank_start;
+	dml_float_t LSetup;
+	dml_float_t blank_lines_remaining;
+	dml_float_t old_MIN_DST_Y_NEXT_START;
+	dml_float_t TotalWRBandwidth;
+	dml_float_t WRBandwidth;
+	struct Watermarks dummy_watermark;
+	struct DmlPipe myPipe;
+};
+
+struct CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_locals_st {
+	dml_float_t ActiveDRAMClockChangeLatencyMargin[__DML_NUM_PLANES__];
+	dml_float_t ActiveFCLKChangeLatencyMargin[__DML_NUM_PLANES__];
+	dml_float_t USRRetrainingLatencyMargin[__DML_NUM_PLANES__];
+
+	dml_bool_t SynchronizedSurfaces[__DML_NUM_PLANES__][__DML_NUM_PLANES__];
+	dml_float_t EffectiveLBLatencyHidingY;
+	dml_float_t EffectiveLBLatencyHidingC;
+	dml_float_t LinesInDETY[__DML_NUM_PLANES__];
+	dml_float_t LinesInDETC[__DML_NUM_PLANES__];
+	dml_uint_t LinesInDETYRoundedDownToSwath[__DML_NUM_PLANES__];
+	dml_uint_t LinesInDETCRoundedDownToSwath[__DML_NUM_PLANES__];
+	dml_float_t FullDETBufferingTimeY;
+	dml_float_t FullDETBufferingTimeC;
+	dml_float_t WritebackDRAMClockChangeLatencyMargin;
+	dml_float_t WritebackFCLKChangeLatencyMargin;
+	dml_float_t WritebackLatencyHiding;
+
+	dml_uint_t TotalActiveWriteback;
+	dml_uint_t LBLatencyHidingSourceLinesY[__DML_NUM_PLANES__];
+	dml_uint_t LBLatencyHidingSourceLinesC[__DML_NUM_PLANES__];
+	dml_float_t TotalPixelBW;
+	dml_float_t EffectiveDETBufferSizeY;
+	dml_float_t ActiveClockChangeLatencyHidingY;
+	dml_float_t ActiveClockChangeLatencyHidingC;
+	dml_float_t ActiveClockChangeLatencyHiding;
+	dml_bool_t FoundCriticalSurface;
+	dml_uint_t LastSurfaceWithoutMargin;
+	dml_uint_t FCLKChangeSupportNumber;
+	dml_uint_t DRAMClockChangeMethod;
+	dml_uint_t DRAMClockChangeSupportNumber;
+	dml_uint_t dst_y_pstate;
+	dml_uint_t src_y_pstate_l;
+	dml_uint_t src_y_pstate_c;
+	dml_uint_t src_y_ahead_l;
+	dml_uint_t src_y_ahead_c;
+	dml_uint_t sub_vp_lines_l;
+	dml_uint_t sub_vp_lines_c;
+};
+
+struct CalculateVMRowAndSwath_locals_st {
+	dml_uint_t PTEBufferSizeInRequestsForLuma[__DML_NUM_PLANES__];
+	dml_uint_t PTEBufferSizeInRequestsForChroma[__DML_NUM_PLANES__];
+	dml_uint_t PDEAndMetaPTEBytesFrameY;
+	dml_uint_t PDEAndMetaPTEBytesFrameC;
+	dml_uint_t MetaRowByteY[__DML_NUM_PLANES__];
+	dml_uint_t MetaRowByteC[__DML_NUM_PLANES__];
+	dml_uint_t PixelPTEBytesPerRowY[__DML_NUM_PLANES__];
+	dml_uint_t PixelPTEBytesPerRowC[__DML_NUM_PLANES__];
+	dml_uint_t PixelPTEBytesPerRowStorageY[__DML_NUM_PLANES__];
+	dml_uint_t PixelPTEBytesPerRowStorageC[__DML_NUM_PLANES__];
+	dml_uint_t PixelPTEBytesPerRowY_one_row_per_frame[__DML_NUM_PLANES__];
+	dml_uint_t PixelPTEBytesPerRowC_one_row_per_frame[__DML_NUM_PLANES__];
+	dml_uint_t dpte_row_width_luma_ub_one_row_per_frame[__DML_NUM_PLANES__];
+	dml_uint_t dpte_row_height_luma_one_row_per_frame[__DML_NUM_PLANES__];
+	dml_uint_t dpte_row_width_chroma_ub_one_row_per_frame[__DML_NUM_PLANES__];
+	dml_uint_t dpte_row_height_chroma_one_row_per_frame[__DML_NUM_PLANES__];
+	dml_bool_t one_row_per_frame_fits_in_buffer[__DML_NUM_PLANES__];
+
+	dml_uint_t HostVMDynamicLevels;
+};
+
+struct UseMinimumDCFCLK_locals_st {
+	dml_uint_t dummy1;
+	dml_uint_t dummy2;
+	dml_uint_t dummy3;
+	dml_float_t NormalEfficiency;
+	dml_float_t TotalMaxPrefetchFlipDPTERowBandwidth[2];
+
+	dml_float_t PixelDCFCLKCyclesRequiredInPrefetch[__DML_NUM_PLANES__];
+	dml_float_t PrefetchPixelLinesTime[__DML_NUM_PLANES__];
+	dml_float_t DCFCLKRequiredForPeakBandwidthPerSurface[__DML_NUM_PLANES__];
+	dml_float_t DynamicMetadataVMExtraLatency[__DML_NUM_PLANES__];
+	dml_float_t MinimumTWait;
+	dml_float_t DPTEBandwidth;
+	dml_float_t DCFCLKRequiredForAverageBandwidth;
+	dml_uint_t ExtraLatencyBytes;
+	dml_float_t ExtraLatencyCycles;
+	dml_float_t DCFCLKRequiredForPeakBandwidth;
+	dml_uint_t NoOfDPPState[__DML_NUM_PLANES__];
+	dml_float_t MinimumTvmPlus2Tr0;
+};
+
+struct CalculatePrefetchSchedule_locals_st {
+	dml_bool_t MyError;
+	dml_uint_t DPPCycles;
+	dml_uint_t DISPCLKCycles;
+	dml_float_t DSTTotalPixelsAfterScaler;
+	dml_float_t LineTime;
+	dml_float_t dst_y_prefetch_equ;
+	dml_float_t prefetch_bw_oto;
+	dml_float_t Tvm_oto;
+	dml_float_t Tr0_oto;
+	dml_float_t Tvm_oto_lines;
+	dml_float_t Tr0_oto_lines;
+	dml_float_t dst_y_prefetch_oto;
+	dml_float_t TimeForFetchingMetaPTE;
+	dml_float_t TimeForFetchingRowInVBlank;
+	dml_float_t LinesToRequestPrefetchPixelData;
+	dml_uint_t HostVMDynamicLevelsTrips;
+	dml_float_t trip_to_mem;
+	dml_float_t Tvm_trips;
+	dml_float_t Tr0_trips;
+	dml_float_t Tvm_trips_rounded;
+	dml_float_t Tr0_trips_rounded;
+	dml_float_t max_Tsw;
+	dml_float_t Lsw_oto;
+	dml_float_t Tpre_rounded;
+	dml_float_t prefetch_bw_equ;
+	dml_float_t Tvm_equ;
+	dml_float_t Tr0_equ;
+	dml_float_t Tdmbf;
+	dml_float_t Tdmec;
+	dml_float_t Tdmsks;
+	dml_float_t prefetch_sw_bytes;
+	dml_float_t prefetch_bw_pr;
+	dml_float_t bytes_pp;
+	dml_float_t dep_bytes;
+	dml_float_t min_Lsw_oto;
+	dml_float_t Tsw_est1;
+	dml_float_t Tsw_est3;
+	dml_float_t PrefetchBandwidth1;
+	dml_float_t PrefetchBandwidth2;
+	dml_float_t PrefetchBandwidth3;
+	dml_float_t PrefetchBandwidth4;
+};
+
+/// @brief To minimize stack usage; function locals are instead placed into this scratch structure which is allocated per context
+struct display_mode_lib_scratch_st {
+	// Scratch space for function locals
+	struct dml_core_mode_support_locals_st dml_core_mode_support_locals;
+	struct dml_core_mode_programming_locals_st dml_core_mode_programming_locals;
+	struct CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_locals_st CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_locals;
+	struct CalculateVMRowAndSwath_locals_st CalculateVMRowAndSwath_locals;
+	struct UseMinimumDCFCLK_locals_st UseMinimumDCFCLK_locals;
+	struct CalculatePrefetchSchedule_locals_st CalculatePrefetchSchedule_locals;
+
+	// Scratch space for function params
+	struct CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_params_st CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_params;
+	struct CalculateVMRowAndSwath_params_st CalculateVMRowAndSwath_params;
+	struct UseMinimumDCFCLK_params_st UseMinimumDCFCLK_params;
+	struct CalculateSwathAndDETConfiguration_params_st CalculateSwathAndDETConfiguration_params;
+	struct CalculateStutterEfficiency_params_st CalculateStutterEfficiency_params;
+	struct CalculatePrefetchSchedule_params_st CalculatePrefetchSchedule_params;
+};
+
+/// @brief Represent the overall soc/ip environment. It contains data structure represent the soc/ip characteristic and also structures that hold calculation output
+struct display_mode_lib_st {
+	dml_uint_t project;
+
+	//@brief Mode evaluation and programming policy
+	struct dml_mode_eval_policy_st policy;
+
+	//@brief IP/SOC characteristic
+	struct ip_params_st ip;
+	struct soc_bounding_box_st soc;
+	struct soc_states_st states;
+
+	//@brief Mode Support and Mode programming struct
+	// Used to hold input; intermediate and output of the calculations
+	struct mode_support_st ms; // struct for mode support
+	struct mode_program_st mp; // struct for mode programming
+
+	struct display_mode_lib_scratch_st scratch;
+};
+
+struct dml_mode_support_ex_params_st {
+	struct display_mode_lib_st *mode_lib;
+	const struct dml_display_cfg_st *in_display_cfg;
+	dml_uint_t in_start_state_idx;
+	dml_uint_t out_lowest_state_idx;
+	struct dml_mode_support_info_st *out_evaluation_info;
+};
+
+typedef struct _vcs_dpi_dml_display_rq_regs_st  dml_display_rq_regs_st;
+typedef struct _vcs_dpi_dml_display_dlg_regs_st dml_display_dlg_regs_st;
+typedef struct _vcs_dpi_dml_display_ttu_regs_st dml_display_ttu_regs_st;
+typedef struct _vcs_dpi_dml_display_arb_params_st   dml_display_arb_params_st;
+typedef struct _vcs_dpi_dml_display_plane_rq_regs_st    dml_display_plane_rq_regs_st;
+
+struct  _vcs_dpi_dml_display_dlg_regs_st {
+	dml_uint_t  refcyc_h_blank_end;
+	dml_uint_t  dlg_vblank_end;
+	dml_uint_t  min_dst_y_next_start;
+	dml_uint_t  refcyc_per_htotal;
+	dml_uint_t  refcyc_x_after_scaler;
+	dml_uint_t  dst_y_after_scaler;
+	dml_uint_t  dst_y_prefetch;
+	dml_uint_t  dst_y_per_vm_vblank;
+	dml_uint_t  dst_y_per_row_vblank;
+	dml_uint_t  dst_y_per_vm_flip;
+	dml_uint_t  dst_y_per_row_flip;
+	dml_uint_t  ref_freq_to_pix_freq;
+	dml_uint_t  vratio_prefetch;
+	dml_uint_t  vratio_prefetch_c;
+	dml_uint_t  refcyc_per_pte_group_vblank_l;
+	dml_uint_t  refcyc_per_pte_group_vblank_c;
+	dml_uint_t  refcyc_per_meta_chunk_vblank_l;
+	dml_uint_t  refcyc_per_meta_chunk_vblank_c;
+	dml_uint_t  refcyc_per_pte_group_flip_l;
+	dml_uint_t  refcyc_per_pte_group_flip_c;
+	dml_uint_t  refcyc_per_meta_chunk_flip_l;
+	dml_uint_t  refcyc_per_meta_chunk_flip_c;
+	dml_uint_t  dst_y_per_pte_row_nom_l;
+	dml_uint_t  dst_y_per_pte_row_nom_c;
+	dml_uint_t  refcyc_per_pte_group_nom_l;
+	dml_uint_t  refcyc_per_pte_group_nom_c;
+	dml_uint_t  dst_y_per_meta_row_nom_l;
+	dml_uint_t  dst_y_per_meta_row_nom_c;
+	dml_uint_t  refcyc_per_meta_chunk_nom_l;
+	dml_uint_t  refcyc_per_meta_chunk_nom_c;
+	dml_uint_t  refcyc_per_line_delivery_pre_l;
+	dml_uint_t  refcyc_per_line_delivery_pre_c;
+	dml_uint_t  refcyc_per_line_delivery_l;
+	dml_uint_t  refcyc_per_line_delivery_c;
+	dml_uint_t  refcyc_per_vm_group_vblank;
+	dml_uint_t  refcyc_per_vm_group_flip;
+	dml_uint_t  refcyc_per_vm_req_vblank;
+	dml_uint_t  refcyc_per_vm_req_flip;
+	dml_uint_t  dst_y_offset_cur0;
+	dml_uint_t  chunk_hdl_adjust_cur0;
+	dml_uint_t  dst_y_offset_cur1;
+	dml_uint_t  chunk_hdl_adjust_cur1;
+	dml_uint_t  vready_after_vcount0;
+	dml_uint_t  dst_y_delta_drq_limit;
+	dml_uint_t  refcyc_per_vm_dmdata;
+	dml_uint_t  dmdata_dl_delta;
+};
+
+struct  _vcs_dpi_dml_display_ttu_regs_st {
+	dml_uint_t  qos_level_low_wm;
+	dml_uint_t  qos_level_high_wm;
+	dml_uint_t  min_ttu_vblank;
+	dml_uint_t  qos_level_flip;
+	dml_uint_t  refcyc_per_req_delivery_l;
+	dml_uint_t  refcyc_per_req_delivery_c;
+	dml_uint_t  refcyc_per_req_delivery_cur0;
+	dml_uint_t  refcyc_per_req_delivery_cur1;
+	dml_uint_t  refcyc_per_req_delivery_pre_l;
+	dml_uint_t  refcyc_per_req_delivery_pre_c;
+	dml_uint_t  refcyc_per_req_delivery_pre_cur0;
+	dml_uint_t  refcyc_per_req_delivery_pre_cur1;
+	dml_uint_t  qos_level_fixed_l;
+	dml_uint_t  qos_level_fixed_c;
+	dml_uint_t  qos_level_fixed_cur0;
+	dml_uint_t  qos_level_fixed_cur1;
+	dml_uint_t  qos_ramp_disable_l;
+	dml_uint_t  qos_ramp_disable_c;
+	dml_uint_t  qos_ramp_disable_cur0;
+	dml_uint_t  qos_ramp_disable_cur1;
+};
+
+struct  _vcs_dpi_dml_display_arb_params_st {
+	dml_uint_t  max_req_outstanding;
+	dml_uint_t  min_req_outstanding;
+	dml_uint_t  sat_level_us;
+	dml_uint_t  hvm_max_qos_commit_threshold;
+	dml_uint_t  hvm_min_req_outstand_commit_threshold;
+	dml_uint_t  compbuf_reserved_space_kbytes;
+};
+
+struct  _vcs_dpi_dml_display_plane_rq_regs_st {
+	dml_uint_t  chunk_size;
+	dml_uint_t  min_chunk_size;
+	dml_uint_t  meta_chunk_size;
+	dml_uint_t  min_meta_chunk_size;
+	dml_uint_t  dpte_group_size;
+	dml_uint_t  mpte_group_size;
+	dml_uint_t  swath_height;
+	dml_uint_t  pte_row_height_linear;
+};
+
+struct  _vcs_dpi_dml_display_rq_regs_st {
+	dml_display_plane_rq_regs_st    rq_regs_l;
+	dml_display_plane_rq_regs_st    rq_regs_c;
+	dml_uint_t  drq_expansion_mode;
+	dml_uint_t  prq_expansion_mode;
+	dml_uint_t  mrq_expansion_mode;
+	dml_uint_t  crq_expansion_mode;
+	dml_uint_t  plane1_base_address;
+};
+
+#endif
diff --git a/drivers/gpu/drm/amd/display/dc/dml2_0/display_mode_lib_defines.h b/drivers/gpu/drm/amd/display/dc/dml2_0/display_mode_lib_defines.h
new file mode 100644
index 000000000000..e574c81edf5e
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml2_0/display_mode_lib_defines.h
@@ -0,0 +1,79 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __DISPLAY_MODE_LIB_DEFINES_H__
+#define __DISPLAY_MODE_LIB_DEFINES_H__
+
+#define DCN_DML__DML_STANDALONE 1
+#define DCN_DML__DML_STANDALONE__1 1
+#define DCN_DML__PRESENT 1
+#define DCN_DML__PRESENT__1 1
+#define DCN_DML__NUM_PLANE 8
+#define DCN_DML__NUM_PLANE__8 1
+#define DCN_DML__NUM_CURSOR 1
+#define DCN_DML__NUM_CURSOR__1 1
+#define DCN_DML__NUM_PWR_STATE 30
+#define DCN_DML__NUM_PWR_STATE__30 1
+#define DCN_DML__VM_PRESENT 1
+#define DCN_DML__VM_PRESENT__1 1
+#define DCN_DML__HOST_VM_PRESENT 1
+#define DCN_DML__HOST_VM_PRESENT__1 1
+#define DCN_DML__DWB 1
+
+#include "dml_depedencies.h"
+
+#include "dml_logging.h"
+#include "dml_assert.h"
+
+// To enable a lot of debug msg
+#define __DML_VBA_DEBUG__
+#define __DML_VBA_ENABLE_INLINE_CHECK_                  0
+#define __DML_VBA_MIN_VSTARTUP__                        9       //<brief At which vstartup the DML start to try if the mode can be supported
+#define __DML_ARB_TO_RET_DELAY__                        (7 + 95)  //<brief Delay in DCFCLK from ARB to DET (1st num is ARB to SDPIF, 2nd number is SDPIF to DET)
+#define __DML_MIN_DCFCLK_FACTOR__                       1.15    //<brief fudge factor for min dcfclk calclation
+#define __DML_MAX_VRATIO_PRE__                          4.0     //<brief Prefetch schedule max vratio
+#define __DML_MAX_VRATIO_PRE_OTO__                      4.0     //<brief Prefetch schedule max vratio for one to one scheduling calculation for prefetch
+#define __DML_MAX_VRATIO_PRE_ENHANCE_PREFETCH_ACC__     6.0     //<brief Prefetch schedule max vratio when enhance prefetch schedule acceleration is enabled and vstartup is earliest possible already
+#define __DML_NUM_PLANES__                              DCN_DML__NUM_PLANE
+#define __DML_NUM_CURSORS__                             DCN_DML__NUM_CURSOR
+#define __DML_DPP_INVALID__                             0
+#define __DML_NUM_DMB__                                 DCN_DML__DWB
+#define __DML_PIPE_NO_PLANE__                           99
+
+#define __DML_MAX_STATE_ARRAY_SIZE__        DCN_DML__NUM_PWR_STATE
+
+// Compilation define
+#define __DML_DLL_EXPORT__
+
+typedef          int   dml_int_t;   // int is 32-bit in C/C++, but Integer datatype is 16-bit in VBA. this should map to Long in VBA
+typedef unsigned int   dml_uint_t;
+typedef double         dml_float_t;
+
+// Note: bool is 8-bit in C/C++, but Boolean is 16-bit in VBA, use "short" in C/C++ DLL so the struct work when vba uses DLL
+// Or the VBA side don't use Boolean, just use "Byte", then C side can use bool
+typedef bool          dml_bool_t;
+
+#endif
diff --git a/drivers/gpu/drm/amd/display/dc/dml2_0/display_mode_util.c b/drivers/gpu/drm/amd/display/dc/dml2_0/display_mode_util.c
new file mode 100644
index 000000000000..89890c88fd66
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml2_0/display_mode_util.c
@@ -0,0 +1,798 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#include "display_mode_util.h"
+
+static dml_float_t _log(float in)
+{
+	int * const exp_ptr = (int *)(&in);
+	int x = *exp_ptr;
+	const int log_2 = ((x >> 23) & 255) - 128;
+
+	x &= ~(255 << 23);
+	x += 127 << 23;
+	*exp_ptr = x;
+
+	in = ((-1.0f / 3) * in + 2) * in - 2.0f / 3;
+
+	return (in + log_2);
+}
+
+dml_bool_t dml_util_is_420(enum dml_source_format_class source_format)
+{
+	dml_bool_t val = false;
+
+	switch (source_format) {
+	case dml_444_16:
+		val = 0;
+		break;
+	case dml_444_32:
+		val = 0;
+		break;
+	case dml_444_64:
+		val = 0;
+		break;
+	case dml_420_8:
+		val = 1;
+		break;
+	case dml_420_10:
+		val = 1;
+		break;
+	case dml_422_8:
+		val = 0;
+		break;
+	case dml_422_10:
+		val = 0;
+		break;
+	default:
+		ASSERT(0);
+		break;
+	}
+	return val;
+}
+
+static inline float dcn_bw_pow(float a, float exp)
+{
+	float temp;
+	/*ASSERT(exp == (int)exp);*/
+	if ((int)exp == 0)
+		return 1;
+	temp = dcn_bw_pow(a, (int)(exp / 2));
+	if (((int)exp % 2) == 0) {
+		return temp * temp;
+	} else {
+		if ((int)exp > 0)
+			return a * temp * temp;
+		else
+			return (temp * temp) / a;
+	}
+}
+
+static inline float dcn_bw_ceil2(const float arg, const float significance)
+{
+	ASSERT(significance != 0);
+
+	return ((int)(arg / significance + 0.99999)) * significance;
+}
+
+static inline float dcn_bw_floor2(const float arg, const float significance)
+{
+	ASSERT(significance != 0);
+
+	return ((int)(arg / significance)) * significance;
+}
+
+dml_float_t dml_ceil(dml_float_t x, dml_float_t granularity)
+{
+	if (granularity == 0)
+		return 0;
+	//return (dml_float_t) (ceil(x / granularity) * granularity);
+	return (dml_float_t)dcn_bw_ceil2(x, granularity);
+}
+
+dml_float_t dml_floor(dml_float_t x, dml_float_t granularity)
+{
+	if (granularity == 0)
+	return 0;
+	//return (dml_float_t) (floor(x / granularity) * granularity);
+	return (dml_float_t)dcn_bw_floor2(x, granularity);
+}
+
+dml_float_t dml_min(dml_float_t x, dml_float_t y)
+{
+	if (x != x)
+		return y;
+	if (y != y)
+		return x;
+	if (x < y)
+		return x;
+	else
+		return y;
+}
+
+dml_float_t dml_min3(dml_float_t x, dml_float_t y, dml_float_t z)
+{
+	return dml_min(dml_min(x, y), z);
+}
+
+dml_float_t dml_min4(dml_float_t x, dml_float_t y, dml_float_t z, dml_float_t w)
+{
+	return dml_min(dml_min(dml_min(x, y), z), w);
+}
+
+dml_float_t dml_max(dml_float_t x, dml_float_t y)
+{
+	if (x != x)
+		return y;
+	if (y != y)
+		return x;
+if (x > y)
+		return x;
+	else
+		return y;
+}
+dml_float_t dml_max3(dml_float_t x, dml_float_t y, dml_float_t z)
+{
+	return dml_max(dml_max(x, y), z);
+}
+dml_float_t dml_max4(dml_float_t a, dml_float_t b, dml_float_t c, dml_float_t d)
+{
+	return dml_max(dml_max(a, b), dml_max(c, d));
+}
+dml_float_t dml_max5(dml_float_t a, dml_float_t b, dml_float_t c, dml_float_t d, dml_float_t e)
+{
+	return dml_max(dml_max4(a, b, c, d), e);
+}
+dml_float_t dml_log(dml_float_t x, dml_float_t base)
+{
+	return (dml_float_t) (_log(x) / _log(base));
+}
+
+dml_float_t dml_log2(dml_float_t x)
+{
+	return (dml_float_t) (_log(x) / _log(2));
+}
+
+dml_float_t dml_round(dml_float_t val, dml_bool_t bankers_rounding)
+{
+//	if (bankers_rounding)
+//			return (dml_float_t) lrint(val);
+//	else {
+//		return round(val);
+		double round_pt = 0.5;
+		double ceil = dml_ceil(val, 1);
+		double floor = dml_floor(val, 1);
+
+		if (val - floor >= round_pt)
+			return ceil;
+		else
+			return floor;
+//	}
+}
+
+dml_float_t dml_pow(dml_float_t base, int exp)
+{
+	return (dml_float_t) dcn_bw_pow(base, exp);
+}
+
+dml_uint_t dml_round_to_multiple(dml_uint_t num, dml_uint_t multiple, dml_bool_t up)
+{
+	dml_uint_t remainder;
+
+	if (multiple == 0)
+		return num;
+
+	remainder = num % multiple;
+	if (remainder == 0)
+		return num;
+
+	if (up)
+		return (num + multiple - remainder);
+	else
+		return (num - remainder);
+}
+
+void dml_print_data_rq_regs_st(const dml_display_plane_rq_regs_st *rq_regs)
+{
+	dml_print("DML: ===================================== \n");
+	dml_print("DML: DISPLAY_PLANE_RQ_REGS_ST\n");
+	dml_print("DML: chunk_size = 0x%x\n", rq_regs->chunk_size);
+	dml_print("DML: min_chunk_size = 0x%x\n", rq_regs->min_chunk_size);
+	dml_print("DML: meta_chunk_size = 0x%x\n", rq_regs->meta_chunk_size);
+	dml_print("DML: min_meta_chunk_size = 0x%x\n", rq_regs->min_meta_chunk_size);
+	dml_print("DML: dpte_group_size = 0x%x\n", rq_regs->dpte_group_size);
+	dml_print("DML: mpte_group_size = 0x%x\n", rq_regs->mpte_group_size);
+	dml_print("DML: swath_height = 0x%x\n", rq_regs->swath_height);
+	dml_print("DML: pte_row_height_linear = 0x%x\n", rq_regs->pte_row_height_linear);
+	dml_print("DML: ===================================== \n");
+}
+
+void dml_print_rq_regs_st(const dml_display_rq_regs_st *rq_regs)
+{
+	dml_print("DML: ===================================== \n");
+	dml_print("DML: DISPLAY_RQ_REGS_ST\n");
+	dml_print("DML: <LUMA> \n");
+	dml_print_data_rq_regs_st(&rq_regs->rq_regs_l);
+	dml_print("DML: <CHROMA> \n");
+	dml_print_data_rq_regs_st(&rq_regs->rq_regs_c);
+	dml_print("DML: drq_expansion_mode = 0x%x\n", rq_regs->drq_expansion_mode);
+	dml_print("DML: prq_expansion_mode = 0x%x\n", rq_regs->prq_expansion_mode);
+	dml_print("DML: mrq_expansion_mode = 0x%x\n", rq_regs->mrq_expansion_mode);
+	dml_print("DML: crq_expansion_mode = 0x%x\n", rq_regs->crq_expansion_mode);
+	dml_print("DML: plane1_base_address = 0x%x\n", rq_regs->plane1_base_address);
+	dml_print("DML: ===================================== \n");
+}
+
+void dml_print_dlg_regs_st(const dml_display_dlg_regs_st *dlg_regs)
+{
+	dml_print("DML: ===================================== \n");
+	dml_print("DML: DISPLAY_DLG_REGS_ST \n");
+	dml_print("DML: refcyc_h_blank_end = 0x%x\n", dlg_regs->refcyc_h_blank_end);
+	dml_print("DML: dlg_vblank_end = 0x%x\n", dlg_regs->dlg_vblank_end);
+	dml_print("DML: min_dst_y_next_start = 0x%x\n", dlg_regs->min_dst_y_next_start);
+	dml_print("DML: refcyc_per_htotal = 0x%x\n", dlg_regs->refcyc_per_htotal);
+	dml_print("DML: refcyc_x_after_scaler = 0x%x\n", dlg_regs->refcyc_x_after_scaler);
+	dml_print("DML: dst_y_after_scaler = 0x%x\n", dlg_regs->dst_y_after_scaler);
+	dml_print("DML: dst_y_prefetch = 0x%x\n", dlg_regs->dst_y_prefetch);
+	dml_print("DML: dst_y_per_vm_vblank = 0x%x\n", dlg_regs->dst_y_per_vm_vblank);
+	dml_print("DML: dst_y_per_row_vblank = 0x%x\n", dlg_regs->dst_y_per_row_vblank);
+	dml_print("DML: dst_y_per_vm_flip = 0x%x\n", dlg_regs->dst_y_per_vm_flip);
+	dml_print("DML: dst_y_per_row_flip = 0x%x\n", dlg_regs->dst_y_per_row_flip);
+	dml_print("DML: ref_freq_to_pix_freq = 0x%x\n", dlg_regs->ref_freq_to_pix_freq);
+	dml_print("DML: vratio_prefetch = 0x%x\n", dlg_regs->vratio_prefetch);
+	dml_print("DML: vratio_prefetch_c = 0x%x\n", dlg_regs->vratio_prefetch_c);
+	dml_print("DML: refcyc_per_pte_group_vblank_l = 0x%x\n", dlg_regs->refcyc_per_pte_group_vblank_l);
+	dml_print("DML: refcyc_per_pte_group_vblank_c = 0x%x\n", dlg_regs->refcyc_per_pte_group_vblank_c);
+	dml_print("DML: refcyc_per_meta_chunk_vblank_l = 0x%x\n", dlg_regs->refcyc_per_meta_chunk_vblank_l);
+	dml_print("DML: refcyc_per_meta_chunk_vblank_c = 0x%x\n", dlg_regs->refcyc_per_meta_chunk_vblank_c);
+	dml_print("DML: refcyc_per_pte_group_flip_l = 0x%x\n", dlg_regs->refcyc_per_pte_group_flip_l);
+	dml_print("DML: refcyc_per_pte_group_flip_c = 0x%x\n", dlg_regs->refcyc_per_pte_group_flip_c);
+	dml_print("DML: refcyc_per_meta_chunk_flip_l = 0x%x\n", dlg_regs->refcyc_per_meta_chunk_flip_l);
+	dml_print("DML: refcyc_per_meta_chunk_flip_c = 0x%x\n", dlg_regs->refcyc_per_meta_chunk_flip_c);
+	dml_print("DML: dst_y_per_pte_row_nom_l = 0x%x\n", dlg_regs->dst_y_per_pte_row_nom_l);
+	dml_print("DML: dst_y_per_pte_row_nom_c = 0x%x\n", dlg_regs->dst_y_per_pte_row_nom_c);
+	dml_print("DML: refcyc_per_pte_group_nom_l = 0x%x\n", dlg_regs->refcyc_per_pte_group_nom_l);
+	dml_print("DML: refcyc_per_pte_group_nom_c = 0x%x\n", dlg_regs->refcyc_per_pte_group_nom_c);
+	dml_print("DML: dst_y_per_meta_row_nom_l = 0x%x\n", dlg_regs->dst_y_per_meta_row_nom_l);
+	dml_print("DML: dst_y_per_meta_row_nom_c = 0x%x\n", dlg_regs->dst_y_per_meta_row_nom_c);
+	dml_print("DML: refcyc_per_meta_chunk_nom_l = 0x%x\n", dlg_regs->refcyc_per_meta_chunk_nom_l);
+	dml_print("DML: refcyc_per_meta_chunk_nom_c = 0x%x\n", dlg_regs->refcyc_per_meta_chunk_nom_c);
+	dml_print("DML: refcyc_per_line_delivery_pre_l = 0x%x\n", dlg_regs->refcyc_per_line_delivery_pre_l);
+	dml_print("DML: refcyc_per_line_delivery_pre_c = 0x%x\n", dlg_regs->refcyc_per_line_delivery_pre_c);
+	dml_print("DML: refcyc_per_line_delivery_l = 0x%x\n", dlg_regs->refcyc_per_line_delivery_l);
+	dml_print("DML: refcyc_per_line_delivery_c = 0x%x\n", dlg_regs->refcyc_per_line_delivery_c);
+	dml_print("DML: refcyc_per_vm_group_vblank = 0x%x\n", dlg_regs->refcyc_per_vm_group_vblank);
+	dml_print("DML: refcyc_per_vm_group_flip = 0x%x\n", dlg_regs->refcyc_per_vm_group_flip);
+	dml_print("DML: refcyc_per_vm_req_vblank = 0x%x\n", dlg_regs->refcyc_per_vm_req_vblank);
+	dml_print("DML: refcyc_per_vm_req_flip = 0x%x\n", dlg_regs->refcyc_per_vm_req_flip);
+	dml_print("DML: chunk_hdl_adjust_cur0 = 0x%x\n", dlg_regs->chunk_hdl_adjust_cur0);
+	dml_print("DML: dst_y_offset_cur1 = 0x%x\n", dlg_regs->dst_y_offset_cur1);
+	dml_print("DML: chunk_hdl_adjust_cur1 = 0x%x\n", dlg_regs->chunk_hdl_adjust_cur1);
+	dml_print("DML: vready_after_vcount0 = 0x%x\n", dlg_regs->vready_after_vcount0);
+	dml_print("DML: dst_y_delta_drq_limit = 0x%x\n", dlg_regs->dst_y_delta_drq_limit);
+	dml_print("DML: refcyc_per_vm_dmdata = 0x%x\n", dlg_regs->refcyc_per_vm_dmdata);
+	dml_print("DML: ===================================== \n");
+}
+
+void dml_print_ttu_regs_st(const dml_display_ttu_regs_st *ttu_regs)
+{
+	dml_print("DML: ===================================== \n");
+	dml_print("DML: DISPLAY_TTU_REGS_ST \n");
+	dml_print("DML: qos_level_low_wm = 0x%x\n", ttu_regs->qos_level_low_wm);
+	dml_print("DML: qos_level_high_wm = 0x%x\n", ttu_regs->qos_level_high_wm);
+	dml_print("DML: min_ttu_vblank = 0x%x\n", ttu_regs->min_ttu_vblank);
+	dml_print("DML: qos_level_flip = 0x%x\n", ttu_regs->qos_level_flip);
+	dml_print("DML: refcyc_per_req_delivery_pre_l = 0x%x\n", ttu_regs->refcyc_per_req_delivery_pre_l);
+	dml_print("DML: refcyc_per_req_delivery_l = 0x%x\n", ttu_regs->refcyc_per_req_delivery_l);
+	dml_print("DML: refcyc_per_req_delivery_pre_c = 0x%x\n", ttu_regs->refcyc_per_req_delivery_pre_c);
+	dml_print("DML: refcyc_per_req_delivery_c = 0x%x\n", ttu_regs->refcyc_per_req_delivery_c);
+	dml_print("DML: refcyc_per_req_delivery_cur0 = 0x%x\n", ttu_regs->refcyc_per_req_delivery_cur0);
+	dml_print("DML: refcyc_per_req_delivery_pre_cur0 = 0x%x\n", ttu_regs->refcyc_per_req_delivery_pre_cur0);
+	dml_print("DML: refcyc_per_req_delivery_cur1 = 0x%x\n", ttu_regs->refcyc_per_req_delivery_cur1);
+	dml_print("DML: refcyc_per_req_delivery_pre_cur1 = 0x%x\n", ttu_regs->refcyc_per_req_delivery_pre_cur1);
+	dml_print("DML: qos_level_fixed_l = 0x%x\n", ttu_regs->qos_level_fixed_l);
+	dml_print("DML: qos_ramp_disable_l = 0x%x\n", ttu_regs->qos_ramp_disable_l);
+	dml_print("DML: qos_level_fixed_c = 0x%x\n", ttu_regs->qos_level_fixed_c);
+	dml_print("DML: qos_ramp_disable_c = 0x%x\n", ttu_regs->qos_ramp_disable_c);
+	dml_print("DML: qos_level_fixed_cur0 = 0x%x\n", ttu_regs->qos_level_fixed_cur0);
+	dml_print("DML: qos_ramp_disable_cur0 = 0x%x\n", ttu_regs->qos_ramp_disable_cur0);
+	dml_print("DML: qos_level_fixed_cur1 = 0x%x\n", ttu_regs->qos_level_fixed_cur1);
+	dml_print("DML: qos_ramp_disable_cur1 = 0x%x\n", ttu_regs->qos_ramp_disable_cur1);
+	dml_print("DML: ===================================== \n");
+}
+
+void dml_print_dml_policy(const struct dml_mode_eval_policy_st *policy)
+{
+	dml_print("DML: ===================================== \n");
+	dml_print("DML: DML_MODE_EVAL_POLICY_ST\n");
+	dml_print("DML: Policy: UseUnboundedRequesting = 0x%x\n", policy->UseUnboundedRequesting);
+	dml_print("DML: Policy: UseMinimumRequiredDCFCLK = 0x%x\n", policy->UseMinimumRequiredDCFCLK);
+	dml_print("DML: Policy: DRAMClockChangeRequirementFinal = 0x%x\n", policy->DRAMClockChangeRequirementFinal);
+	dml_print("DML: Policy: FCLKChangeRequirementFinal = 0x%x\n", policy->FCLKChangeRequirementFinal);
+	dml_print("DML: Policy: USRRetrainingRequiredFinal = 0x%x\n", policy->USRRetrainingRequiredFinal);
+	dml_print("DML: Policy: EnhancedPrefetchScheduleAccelerationFinal = 0x%x\n", policy->EnhancedPrefetchScheduleAccelerationFinal);
+	dml_print("DML: Policy: NomDETInKByteOverrideEnable = 0x%x\n", policy->NomDETInKByteOverrideEnable);
+	dml_print("DML: Policy: NomDETInKByteOverrideValue = 0x%x\n", policy->NomDETInKByteOverrideValue);
+	dml_print("DML: Policy: DCCProgrammingAssumesScanDirectionUnknownFinal = 0x%x\n", policy->DCCProgrammingAssumesScanDirectionUnknownFinal);
+	dml_print("DML: Policy: SynchronizeTimingsFinal = 0x%x\n", policy->SynchronizeTimingsFinal);
+	dml_print("DML: Policy: SynchronizeDRRDisplaysForUCLKPStateChangeFinal = 0x%x\n", policy->SynchronizeDRRDisplaysForUCLKPStateChangeFinal);
+	dml_print("DML: Policy: AssumeModeSupportAtMaxPwrStateEvenDRAMClockChangeNotSupported = 0x%x\n", policy->AssumeModeSupportAtMaxPwrStateEvenDRAMClockChangeNotSupported);
+	dml_print("DML: Policy: AssumeModeSupportAtMaxPwrStateEvenFClockChangeNotSupported = 0x%x\n", policy->AssumeModeSupportAtMaxPwrStateEvenFClockChangeNotSupported);
+
+	for (dml_uint_t i = 0; i < DCN_DML__NUM_PLANE; i++) {
+		dml_print("DML: i=%0d, Policy: MPCCombineUse = 0x%x\n", i, policy->MPCCombineUse[i]);
+		dml_print("DML: i=%0d, Policy: ODMUse = 0x%x\n", i, policy->ODMUse[i]);
+		dml_print("DML: i=%0d, Policy: ImmediateFlipRequirement = 0x%x\n", i, policy->ImmediateFlipRequirement[i]);
+		dml_print("DML: i=%0d, Policy: AllowForPStateChangeOrStutterInVBlank = 0x%x\n", i, policy->AllowForPStateChangeOrStutterInVBlank[i]);
+	}
+	dml_print("DML: ===================================== \n");
+}
+
+void dml_print_mode_support(struct display_mode_lib_st *mode_lib, dml_uint_t j)
+{
+	dml_print("DML: MODE SUPPORT: ===============================================\n");
+	dml_print("DML: MODE SUPPORT: Voltage State %d\n", j);
+	dml_print("DML: MODE SUPPORT:     Mode Supported              : %s\n", mode_lib->ms.support.ModeSupport[j] == true ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     Scale Ratio And Taps                : %s\n", mode_lib->ms.support.ScaleRatioAndTapsSupport == true ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     Source Format Pixel And Scan        : %s\n", mode_lib->ms.support.SourceFormatPixelAndScanSupport == true ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     Viewport Size                       : %s\n", mode_lib->ms.support.ViewportSizeSupport[j] == true ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     Link Rate Does Not Match DP Version        : %s\n", mode_lib->ms.support.LinkRateDoesNotMatchDPVersion == false ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     Link Rate For Multistream Not Indicated    : %s\n", mode_lib->ms.support.LinkRateForMultistreamNotIndicated == false ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     BPP For Multi stream Not Indicated         : %s\n", mode_lib->ms.support.BPPForMultistreamNotIndicated == false ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     Multistream With HDMI Or eDP               : %s\n", mode_lib->ms.support.MultistreamWithHDMIOreDP == false ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     Exceeded Multistream Slots                 : %s\n", mode_lib->ms.support.ExceededMultistreamSlots == false ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     MSO Or ODM Split With Non DP Link          : %s\n", mode_lib->ms.support.MSOOrODMSplitWithNonDPLink == false ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     Not Enough Lanes For MSO                   : %s\n", mode_lib->ms.support.NotEnoughLanesForMSO == false ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     LinkCapacitySupport                        : %s\n", mode_lib->ms.support.LinkCapacitySupport == true ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     P2IWith420                                 : %s\n", mode_lib->ms.support.P2IWith420 == false ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     DSCOnlyIfNecessaryWithBPP                  : %s\n", mode_lib->ms.support.DSCOnlyIfNecessaryWithBPP == false ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     DSC422NativeNotSupported                   : %s\n", mode_lib->ms.support.DSC422NativeNotSupported == false ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     MPCCombineMethodIncompatible               : %s\n", mode_lib->ms.support.MPCCombineMethodIncompatible == false ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     ODMCombineTwoToOneSupportCheckOK           : %s\n", mode_lib->ms.support.ODMCombineTwoToOneSupportCheckOK == true ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     ODMCombineFourToOneSupportCheckOK          : %s\n", mode_lib->ms.support.ODMCombineFourToOneSupportCheckOK == true ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     NotEnoughDSCUnits                          : %s\n", mode_lib->ms.support.NotEnoughDSCUnits == false ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     NotEnoughDSCSlices                         : %s\n", mode_lib->ms.support.NotEnoughDSCSlices == false ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     ImmediateFlipOrHostVMAndPStateWithMALLFullFrameOrPhantomPipe : %s\n", mode_lib->ms.support.ImmediateFlipOrHostVMAndPStateWithMALLFullFrameOrPhantomPipe == false ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     InvalidCombinationOfMALLUseForPStateAndStaticScreen          : %s\n", mode_lib->ms.support.InvalidCombinationOfMALLUseForPStateAndStaticScreen == false ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     DSCCLKRequiredMoreThanSupported            : %s\n", mode_lib->ms.support.DSCCLKRequiredMoreThanSupported == false ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     PixelsPerLinePerDSCUnitSupport             : %s\n", mode_lib->ms.support.PixelsPerLinePerDSCUnitSupport == true ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     DTBCLKRequiredMoreThanSupported            : %s\n", mode_lib->ms.support.DTBCLKRequiredMoreThanSupported == false ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     InvalidCombinationOfMALLUseForPState       : %s\n", mode_lib->ms.support.InvalidCombinationOfMALLUseForPState == false ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     ImmediateFlipRequiredButTheRequirementForEachSurfaceIsNotSpecified : %s\n", mode_lib->ms.support.ImmediateFlipRequiredButTheRequirementForEachSurfaceIsNotSpecified == false ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     ROB Support                                : %s\n", mode_lib->ms.support.ROBSupport[j] == true ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     DISPCLK DPPCLK Support                     : %s\n", mode_lib->ms.support.DISPCLK_DPPCLK_Support[j] == true ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     Total Available Pipes Support              : %s\n", mode_lib->ms.support.TotalAvailablePipesSupport[j] == true ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     Number Of OTG Support                      : %s\n", mode_lib->ms.support.NumberOfOTGSupport == true ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     Number Of DP2p0 Support                    : %s\n", mode_lib->ms.support.NumberOfDP2p0Support == true ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     Writeback Latency Support                  : %s\n", mode_lib->ms.support.WritebackLatencySupport == true ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     Writeback Scale Ratio And Taps Support     : %s\n", mode_lib->ms.support.WritebackScaleRatioAndTapsSupport == true ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     Cursor Support                             : %s\n", mode_lib->ms.support.CursorSupport == true ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     Pitch Support                              : %s\n", mode_lib->ms.support.PitchSupport == true ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     Viewport Exceeds Surface                   : %s\n", mode_lib->ms.support.ViewportExceedsSurface == false ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     Prefetch Supported                         : %s\n", mode_lib->ms.support.PrefetchSupported[j] == true ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     VActive Bandwith Support                   : %s\n", mode_lib->ms.support.VActiveBandwithSupport[j] == true ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     Dynamic Metadata Supported                 : %s\n", mode_lib->ms.support.DynamicMetadataSupported[j] == true ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     Total Vertical Active Bandwidth Support    : %s\n", mode_lib->ms.support.TotalVerticalActiveBandwidthSupport[j] == true ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     VRatio In Prefetch Supported               : %s\n", mode_lib->ms.support.VRatioInPrefetchSupported[j] == true ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     PTE Buffer Size Not Exceeded               : %s\n", mode_lib->ms.support.PTEBufferSizeNotExceeded[j] == true ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     DCC Meta Buffer Size Not Exceeded          : %s\n", mode_lib->ms.support.DCCMetaBufferSizeNotExceeded[j] == true ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     Non supported DSC Input BPC                : %s\n", mode_lib->ms.support.NonsupportedDSCInputBPC == false ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     Exceeded MALL Size                         : %s\n", mode_lib->ms.support.ExceededMALLSize == false ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     Host VM or Immediate Flip Supported        : %s\n", ((mode_lib->ms.cache_display_cfg.plane.HostVMEnable == false && !mode_lib->scratch.dml_core_mode_support_locals.ImmediateFlipRequiredFinal) || mode_lib->ms.support.ImmediateFlipSupportedForState[j]) ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     dram clock change support                  : %s\n", mode_lib->scratch.dml_core_mode_support_locals.dram_clock_change_support == true ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     f_clock change support                     : %s\n", mode_lib->scratch.dml_core_mode_support_locals.f_clock_change_support == true ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT:     USR Retraining Support                     : %s\n", (!mode_lib->ms.policy.USRRetrainingRequiredFinal || &mode_lib->ms.support.USRRetrainingSupport[j]) ? "Supported" : "NOT Supported");
+	dml_print("DML: MODE SUPPORT: ===============================================\n");
+}
+
+void dml_print_dml_mode_support_info(const struct dml_mode_support_info_st *support, dml_bool_t fail_only)
+{
+	dml_print("DML: ===================================== \n");
+	dml_print("DML: DML_MODE_SUPPORT_INFO_ST\n");
+	if (!fail_only || support->ModeIsSupported == 0)
+		dml_print("DML: support: ModeIsSupported = 0x%x\n", support->ModeIsSupported);
+	if (!fail_only || support->ImmediateFlipSupport == 0)
+		dml_print("DML: support: ImmediateFlipSupport = 0x%x\n", support->ImmediateFlipSupport);
+	if (!fail_only || support->WritebackLatencySupport == 0)
+		dml_print("DML: support: WritebackLatencySupport = 0x%x\n", support->WritebackLatencySupport);
+	if (!fail_only || support->ScaleRatioAndTapsSupport == 0)
+		dml_print("DML: support: ScaleRatioAndTapsSupport = 0x%x\n", support->ScaleRatioAndTapsSupport);
+	if (!fail_only || support->SourceFormatPixelAndScanSupport == 0)
+		dml_print("DML: support: SourceFormatPixelAndScanSupport = 0x%x\n", support->SourceFormatPixelAndScanSupport);
+	if (!fail_only || support->MPCCombineMethodIncompatible == 1)
+		dml_print("DML: support: MPCCombineMethodIncompatible = 0x%x\n", support->MPCCombineMethodIncompatible);
+	if (!fail_only || support->P2IWith420 == 1)
+		dml_print("DML: support: P2IWith420 = 0x%x\n", support->P2IWith420);
+	if (!fail_only || support->DSCOnlyIfNecessaryWithBPP == 1)
+		dml_print("DML: support: DSCOnlyIfNecessaryWithBPP = 0x%x\n", support->DSCOnlyIfNecessaryWithBPP);
+	if (!fail_only || support->DSC422NativeNotSupported == 1)
+		dml_print("DML: support: DSC422NativeNotSupported = 0x%x\n", support->DSC422NativeNotSupported);
+	if (!fail_only || support->LinkRateDoesNotMatchDPVersion == 1)
+		dml_print("DML: support: LinkRateDoesNotMatchDPVersion = 0x%x\n", support->LinkRateDoesNotMatchDPVersion);
+	if (!fail_only || support->LinkRateForMultistreamNotIndicated == 1)
+		dml_print("DML: support: LinkRateForMultistreamNotIndicated = 0x%x\n", support->LinkRateForMultistreamNotIndicated);
+	if (!fail_only || support->BPPForMultistreamNotIndicated == 1)
+		dml_print("DML: support: BPPForMultistreamNotIndicated = 0x%x\n", support->BPPForMultistreamNotIndicated);
+	if (!fail_only || support->MultistreamWithHDMIOreDP == 1)
+		dml_print("DML: support: MultistreamWithHDMIOreDP = 0x%x\n", support->MultistreamWithHDMIOreDP);
+	if (!fail_only || support->MSOOrODMSplitWithNonDPLink == 1)
+		dml_print("DML: support: MSOOrODMSplitWithNonDPLink = 0x%x\n", support->MSOOrODMSplitWithNonDPLink);
+	if (!fail_only || support->NotEnoughLanesForMSO == 1)
+		dml_print("DML: support: NotEnoughLanesForMSO = 0x%x\n", support->NotEnoughLanesForMSO);
+	if (!fail_only || support->NumberOfOTGSupport == 0)
+		dml_print("DML: support: NumberOfOTGSupport = 0x%x\n", support->NumberOfOTGSupport);
+	if (!fail_only || support->NumberOfDP2p0Support == 0)
+		dml_print("DML: support: NumberOfDP2p0Support = 0x%x\n", support->NumberOfDP2p0Support);
+	if (!fail_only || support->NonsupportedDSCInputBPC == 1)
+		dml_print("DML: support: NonsupportedDSCInputBPC = 0x%x\n", support->NonsupportedDSCInputBPC);
+	if (!fail_only || support->WritebackScaleRatioAndTapsSupport == 0)
+		dml_print("DML: support: WritebackScaleRatioAndTapsSupport = 0x%x\n", support->WritebackScaleRatioAndTapsSupport);
+	if (!fail_only || support->CursorSupport == 0)
+		dml_print("DML: support: CursorSupport = 0x%x\n", support->CursorSupport);
+	if (!fail_only || support->PitchSupport == 0)
+		dml_print("DML: support: PitchSupport = 0x%x\n", support->PitchSupport);
+	if (!fail_only || support->ViewportExceedsSurface == 1)
+		dml_print("DML: support: ViewportExceedsSurface = 0x%x\n", support->ViewportExceedsSurface);
+	if (!fail_only || support->ExceededMALLSize == 1)
+		dml_print("DML: support: ExceededMALLSize = 0x%x\n", support->ExceededMALLSize);
+	if (!fail_only || support->EnoughWritebackUnits == 0)
+		dml_print("DML: support: EnoughWritebackUnits = 0x%x\n", support->EnoughWritebackUnits);
+	if (!fail_only || support->ImmediateFlipRequiredButTheRequirementForEachSurfaceIsNotSpecified == 1)
+		dml_print("DML: support: ImmediateFlipRequiredButTheRequirementForEachSurfaceIsNotSpecified = 0x%x\n", support->ImmediateFlipRequiredButTheRequirementForEachSurfaceIsNotSpecified);
+	if (!fail_only || support->ImmediateFlipOrHostVMAndPStateWithMALLFullFrameOrPhantomPipe == 1)
+		dml_print("DML: support: ImmediateFlipOrHostVMAndPStateWithMALLFullFrameOrPhantomPipe = 0x%x\n", support->ImmediateFlipOrHostVMAndPStateWithMALLFullFrameOrPhantomPipe);
+	if (!fail_only || support->InvalidCombinationOfMALLUseForPStateAndStaticScreen == 1)
+		dml_print("DML: support: InvalidCombinationOfMALLUseForPStateAndStaticScreen = 0x%x\n", support->InvalidCombinationOfMALLUseForPStateAndStaticScreen);
+	if (!fail_only || support->InvalidCombinationOfMALLUseForPState == 1)
+		dml_print("DML: support: InvalidCombinationOfMALLUseForPState = 0x%x\n", support->InvalidCombinationOfMALLUseForPState);
+
+	if (!fail_only || support->ExceededMultistreamSlots == 1)
+		dml_print("DML: support: ExceededMultistreamSlots = 0x%x\n", support->ExceededMultistreamSlots);
+	if (!fail_only || support->ODMCombineTwoToOneSupportCheckOK == 0)
+		dml_print("DML: support: ODMCombineTwoToOneSupportCheckOK = 0x%x\n", support->ODMCombineTwoToOneSupportCheckOK);
+	if (!fail_only || support->ODMCombineFourToOneSupportCheckOK == 0)
+		dml_print("DML: support: ODMCombineFourToOneSupportCheckOK = 0x%x\n", support->ODMCombineFourToOneSupportCheckOK);
+	if (!fail_only || support->NotEnoughDSCUnits == 1)
+		dml_print("DML: support: NotEnoughDSCUnits = 0x%x\n", support->NotEnoughDSCUnits);
+	if (!fail_only || support->NotEnoughDSCSlices == 1)
+		dml_print("DML: support: NotEnoughDSCSlices = 0x%x\n", support->NotEnoughDSCSlices);
+	if (!fail_only || support->PixelsPerLinePerDSCUnitSupport == 0)
+		dml_print("DML: support: PixelsPerLinePerDSCUnitSupport = 0x%x\n", support->PixelsPerLinePerDSCUnitSupport);
+	if (!fail_only || support->DSCCLKRequiredMoreThanSupported == 1)
+		dml_print("DML: support: DSCCLKRequiredMoreThanSupported = 0x%x\n", support->DSCCLKRequiredMoreThanSupported);
+	if (!fail_only || support->DTBCLKRequiredMoreThanSupported == 1)
+		dml_print("DML: support: DTBCLKRequiredMoreThanSupported = 0x%x\n", support->DTBCLKRequiredMoreThanSupported);
+	if (!fail_only || support->LinkCapacitySupport == 0)
+		dml_print("DML: support: LinkCapacitySupport = 0x%x\n", support->LinkCapacitySupport);
+
+	for (dml_uint_t j = 0; j < 2; j++) {
+		if (!fail_only || support->DRAMClockChangeSupport[j] == dml_dram_clock_change_unsupported)
+			dml_print("DML: support: combine=%d, DRAMClockChangeSupport = %d\n", j, support->DRAMClockChangeSupport[j]);
+		if (!fail_only || support->FCLKChangeSupport[j] == dml_fclock_change_unsupported)
+			dml_print("DML: support: combine=%d, FCLKChangeSupport = %d\n", j, support->FCLKChangeSupport[j]);
+		if (!fail_only || support->ROBSupport[j] == 0)
+			dml_print("DML: support: combine=%d, ROBSupport = %d\n", j, support->ROBSupport[j]);
+		if (!fail_only || support->PTEBufferSizeNotExceeded[j] == 0)
+			dml_print("DML: support: combine=%d, PTEBufferSizeNotExceeded = %d\n", j, support->PTEBufferSizeNotExceeded[j]);
+		if (!fail_only || support->DCCMetaBufferSizeNotExceeded[j] == 0)
+			dml_print("DML: support: combine=%d, DCCMetaBufferSizeNotExceeded = %d\n", j, support->DCCMetaBufferSizeNotExceeded[j]);
+		if (!fail_only || support->TotalVerticalActiveBandwidthSupport[j] == 0)
+			dml_print("DML: support: combine=%d, TotalVerticalActiveBandwidthSupport = %d\n", j, support->TotalVerticalActiveBandwidthSupport[j]);
+		if (!fail_only || support->USRRetrainingSupport[j] == 0)
+			dml_print("DML: support: combine=%d, USRRetrainingSupport = %d\n", j, support->USRRetrainingSupport[j]);
+		if (!fail_only || support->VActiveBandwithSupport[j] == 0)
+			dml_print("DML: support: combine=%d, VActiveBandwithSupport = %d\n", j, support->VActiveBandwithSupport[j]);
+		if (!fail_only || support->PrefetchSupported[j] == 0)
+			dml_print("DML: support: combine=%d, PrefetchSupported = %d\n", j, support->PrefetchSupported[j]);
+		if (!fail_only || support->DynamicMetadataSupported[j] == 0)
+			dml_print("DML: support: combine=%d, DynamicMetadataSupported = %d\n", j, support->DynamicMetadataSupported[j]);
+		if (!fail_only || support->VRatioInPrefetchSupported[j] == 0)
+			dml_print("DML: support: combine=%d, VRatioInPrefetchSupported = %d\n", j, support->VRatioInPrefetchSupported[j]);
+		if (!fail_only || support->DISPCLK_DPPCLK_Support[j] == 0)
+			dml_print("DML: support: combine=%d, DISPCLK_DPPCLK_Support = %d\n", j, support->DISPCLK_DPPCLK_Support[j]);
+		if (!fail_only || support->TotalAvailablePipesSupport[j] == 0)
+			dml_print("DML: support: combine=%d, TotalAvailablePipesSupport = %d\n", j, support->TotalAvailablePipesSupport[j]);
+		if (!fail_only || support->ModeSupport[j] == 0)
+			dml_print("DML: support: combine=%d, ModeSupport = %d\n", j, support->ModeSupport[j]);
+		if (!fail_only || support->ViewportSizeSupport[j] == 0)
+			dml_print("DML: support: combine=%d, ViewportSizeSupport = %d\n", j, support->ViewportSizeSupport[j]);
+		if (!fail_only || support->ImmediateFlipSupportedForState[j] == 0)
+			dml_print("DML: support: combine=%d, ImmediateFlipSupportedForState = %d\n", j, support->ImmediateFlipSupportedForState[j]);
+	}
+}
+
+void dml_print_dml_display_cfg_timing(const struct dml_timing_cfg_st *timing, dml_uint_t num_plane)
+{
+	for (dml_uint_t i = 0; i < num_plane; i++) {
+		dml_print("DML: timing_cfg: plane=%d, HTotal = %d\n", i, timing->HTotal[i]);
+		dml_print("DML: timing_cfg: plane=%d, VTotal = %d\n", i, timing->VTotal[i]);
+		dml_print("DML: timing_cfg: plane=%d, HActive = %d\n", i, timing->HActive[i]);
+		dml_print("DML: timing_cfg: plane=%d, VActive = %d\n", i, timing->VActive[i]);
+		dml_print("DML: timing_cfg: plane=%d, VFrontPorch = %d\n", i, timing->VFrontPorch[i]);
+		dml_print("DML: timing_cfg: plane=%d, VBlankNom = %d\n", i, timing->VBlankNom[i]);
+		dml_print("DML: timing_cfg: plane=%d, RefreshRate = %d\n", i, timing->RefreshRate[i]);
+		dml_print("DML: timing_cfg: plane=%d, PixelClock = %f\n", i, timing->PixelClock[i]);
+		dml_print("DML: timing_cfg: plane=%d, Interlace = %d\n", i, timing->Interlace[i]);
+		dml_print("DML: timing_cfg: plane=%d, DRRDisplay = %d\n", i, timing->DRRDisplay[i]);
+	}
+}
+
+void dml_print_dml_display_cfg_plane(const struct dml_plane_cfg_st *plane, dml_uint_t num_plane)
+{
+	dml_print("DML: plane_cfg: num_plane = %d\n", num_plane);
+	dml_print("DML: plane_cfg: GPUVMEnable = %d\n", plane->GPUVMEnable);
+	dml_print("DML: plane_cfg: HostVMEnable = %d\n", plane->HostVMEnable);
+	dml_print("DML: plane_cfg: GPUVMMaxPageTableLevels = %d\n", plane->GPUVMMaxPageTableLevels);
+	dml_print("DML: plane_cfg: HostVMMaxPageTableLevels = %d\n", plane->HostVMMaxPageTableLevels);
+
+	for (dml_uint_t i = 0; i < num_plane; i++) {
+		dml_print("DML: plane_cfg: plane=%d, GPUVMMinPageSizeKBytes = %d\n", i, plane->GPUVMMinPageSizeKBytes[i]);
+		dml_print("DML: plane_cfg: plane=%d, ForceOneRowForFrame = %d\n", i, plane->ForceOneRowForFrame[i]);
+		dml_print("DML: plane_cfg: plane=%d, PTEBufferModeOverrideEn = %d\n", i, plane->PTEBufferModeOverrideEn[i]);
+		dml_print("DML: plane_cfg: plane=%d, PTEBufferMode = %d\n", i, plane->PTEBufferMode[i]);
+		dml_print("DML: plane_cfg: plane=%d, DETSizeOverride = %d\n", i, plane->DETSizeOverride[i]);
+		dml_print("DML: plane_cfg: plane=%d, UseMALLForStaticScreen = %d\n", i, plane->UseMALLForStaticScreen[i]);
+		dml_print("DML: plane_cfg: plane=%d, UseMALLForPStateChange = %d\n", i, plane->UseMALLForPStateChange[i]);
+		dml_print("DML: plane_cfg: plane=%d, BlendingAndTiming = %d\n", i, plane->BlendingAndTiming[i]);
+		dml_print("DML: plane_cfg: plane=%d, ViewportWidth = %d\n", i, plane->ViewportWidth[i]);
+		dml_print("DML: plane_cfg: plane=%d, ViewportHeight = %d\n", i, plane->ViewportHeight[i]);
+		dml_print("DML: plane_cfg: plane=%d, ViewportWidthChroma = %d\n", i, plane->ViewportWidthChroma[i]);
+		dml_print("DML: plane_cfg: plane=%d, ViewportHeightChroma = %d\n", i, plane->ViewportHeightChroma[i]);
+		dml_print("DML: plane_cfg: plane=%d, ViewportXStart = %d\n", i, plane->ViewportXStart[i]);
+		dml_print("DML: plane_cfg: plane=%d, ViewportXStartC = %d\n", i, plane->ViewportXStartC[i]);
+		dml_print("DML: plane_cfg: plane=%d, ViewportYStart = %d\n", i, plane->ViewportYStart[i]);
+		dml_print("DML: plane_cfg: plane=%d, ViewportYStartC = %d\n", i, plane->ViewportYStartC[i]);
+		dml_print("DML: plane_cfg: plane=%d, ViewportStationary = %d\n", i, plane->ViewportStationary[i]);
+		dml_print("DML: plane_cfg: plane=%d, ScalerEnabled = %d\n", i, plane->ScalerEnabled[i]);
+		dml_print("DML: plane_cfg: plane=%d, HRatio = %3.2f\n", i, plane->HRatio[i]);
+		dml_print("DML: plane_cfg: plane=%d, VRatio = %3.2f\n", i, plane->VRatio[i]);
+		dml_print("DML: plane_cfg: plane=%d, HRatioChroma = %3.2f\n", i, plane->HRatioChroma[i]);
+		dml_print("DML: plane_cfg: plane=%d, VRatioChroma = %3.2f\n", i, plane->VRatioChroma[i]);
+		dml_print("DML: plane_cfg: plane=%d, HTaps = %d\n", i, plane->HTaps[i]);
+		dml_print("DML: plane_cfg: plane=%d, VTaps = %d\n", i, plane->VTaps[i]);
+		dml_print("DML: plane_cfg: plane=%d, HTapsChroma = %d\n", i, plane->HTapsChroma[i]);
+		dml_print("DML: plane_cfg: plane=%d, VTapsChroma = %d\n", i, plane->VTapsChroma[i]);
+		dml_print("DML: plane_cfg: plane=%d, LBBitPerPixel = %d\n", i, plane->LBBitPerPixel[i]);
+		dml_print("DML: plane_cfg: plane=%d, SourceScan = %d\n", i, plane->SourceScan[i]);
+		dml_print("DML: plane_cfg: plane=%d, ScalerRecoutWidth = %d\n", i, plane->ScalerRecoutWidth[i]);
+		dml_print("DML: plane_cfg: plane=%d, NumberOfCursors = %d\n", i, plane->NumberOfCursors[i]);
+		dml_print("DML: plane_cfg: plane=%d, CursorWidth = %d\n", i, plane->CursorWidth[i]);
+		dml_print("DML: plane_cfg: plane=%d, CursorBPP = %d\n", i, plane->CursorBPP[i]);
+
+		dml_print("DML: plane_cfg: plane=%d, DynamicMetadataEnable = %d\n", i, plane->DynamicMetadataEnable[i]);
+		dml_print("DML: plane_cfg: plane=%d, DynamicMetadataLinesBeforeActiveRequired = %d\n", i, plane->DynamicMetadataLinesBeforeActiveRequired[i]);
+		dml_print("DML: plane_cfg: plane=%d, DynamicMetadataTransmittedBytes = %d\n", i, plane->DynamicMetadataTransmittedBytes[i]);
+	}
+}
+
+void dml_print_dml_display_cfg_surface(const struct dml_surface_cfg_st *surface, dml_uint_t num_plane)
+{
+	for (dml_uint_t i = 0; i < num_plane; i++) {
+		dml_print("DML: surface_cfg: plane=%d, PitchY = %d\n", i, surface->PitchY[i]);
+		dml_print("DML: surface_cfg: plane=%d, SurfaceWidthY = %d\n", i, surface->SurfaceWidthY[i]);
+		dml_print("DML: surface_cfg: plane=%d, SurfaceHeightY = %d\n", i, surface->SurfaceHeightY[i]);
+		dml_print("DML: surface_cfg: plane=%d, PitchC = %d\n", i, surface->PitchC[i]);
+		dml_print("DML: surface_cfg: plane=%d, SurfaceWidthC = %d\n", i, surface->SurfaceWidthC[i]);
+		dml_print("DML: surface_cfg: plane=%d, SurfaceHeightC = %d\n", i, surface->SurfaceHeightC[i]);
+		dml_print("DML: surface_cfg: plane=%d, DCCEnable = %d\n", i, surface->DCCEnable[i]);
+		dml_print("DML: surface_cfg: plane=%d, DCCMetaPitchY = %d\n", i, surface->DCCMetaPitchY[i]);
+		dml_print("DML: surface_cfg: plane=%d, DCCMetaPitchC = %d\n", i, surface->DCCMetaPitchC[i]);
+		dml_print("DML: surface_cfg: plane=%d, DCCRateLuma = %f\n", i, surface->DCCRateLuma[i]);
+		dml_print("DML: surface_cfg: plane=%d, DCCRateChroma = %f\n", i, surface->DCCRateChroma[i]);
+		dml_print("DML: surface_cfg: plane=%d, DCCFractionOfZeroSizeRequestsLuma = %f\n", i, surface->DCCFractionOfZeroSizeRequestsLuma[i]);
+		dml_print("DML: surface_cfg: plane=%d, DCCFractionOfZeroSizeRequestsChroma= %f\n", i, surface->DCCFractionOfZeroSizeRequestsChroma[i]);
+	}
+}
+
+void dml_print_dml_display_cfg_hw_resource(const struct dml_hw_resource_st *hw, dml_uint_t num_plane)
+{
+	for (dml_uint_t i = 0; i < num_plane; i++) {
+		dml_print("DML: hw_resource: plane=%d, ODMMode = %d\n", i, hw->ODMMode[i]);
+		dml_print("DML: hw_resource: plane=%d, DPPPerSurface = %d\n", i, hw->DPPPerSurface[i]);
+		dml_print("DML: hw_resource: plane=%d, DSCEnabled = %d\n", i, hw->DSCEnabled[i]);
+		dml_print("DML: hw_resource: plane=%d, NumberOfDSCSlices = %d\n", i, hw->NumberOfDSCSlices[i]);
+	}
+	dml_print("DML: hw_resource: DLGRefClkFreqMHz   = %f\n", hw->DLGRefClkFreqMHz);
+}
+
+__DML_DLL_EXPORT__ void dml_print_soc_state_bounding_box(const struct soc_state_bounding_box_st *state)
+{
+	dml_print("DML: state_bbox: socclk_mhz = %f\n", state->socclk_mhz);
+	dml_print("DML: state_bbox: dscclk_mhz = %f\n", state->dscclk_mhz);
+	dml_print("DML: state_bbox: phyclk_mhz = %f\n", state->phyclk_mhz);
+	dml_print("DML: state_bbox: phyclk_d18_mhz = %f\n", state->phyclk_d18_mhz);
+	dml_print("DML: state_bbox: phyclk_d32_mhz = %f\n", state->phyclk_d32_mhz);
+	dml_print("DML: state_bbox: dtbclk_mhz = %f\n", state->dtbclk_mhz);
+	dml_print("DML: state_bbox: dispclk_mhz = %f\n", state->dispclk_mhz);
+	dml_print("DML: state_bbox: dppclk_mhz = %f\n", state->dppclk_mhz);
+	dml_print("DML: state_bbox: fabricclk_mhz = %f\n", state->fabricclk_mhz);
+	dml_print("DML: state_bbox: dcfclk_mhz = %f\n", state->dcfclk_mhz);
+	dml_print("DML: state_bbox: dram_speed_mts = %f\n", state->dram_speed_mts);
+	dml_print("DML: state_bbox: urgent_latency_pixel_data_only_us = %f\n", state->urgent_latency_pixel_data_only_us);
+	dml_print("DML: state_bbox: urgent_latency_pixel_mixed_with_vm_data_us = %f\n", state->urgent_latency_pixel_mixed_with_vm_data_us);
+	dml_print("DML: state_bbox: urgent_latency_vm_data_only_us = %f\n", state->urgent_latency_vm_data_only_us);
+	dml_print("DML: state_bbox: writeback_latency_us = %f\n", state->writeback_latency_us);
+	dml_print("DML: state_bbox: urgent_latency_adjustment_fabric_clock_component_us = %f\n", state->urgent_latency_adjustment_fabric_clock_component_us);
+	dml_print("DML: state_bbox: urgent_latency_adjustment_fabric_clock_reference_mhz= %f\n", state->urgent_latency_adjustment_fabric_clock_reference_mhz);
+	dml_print("DML: state_bbox: sr_exit_time_us = %f\n", state->sr_exit_time_us);
+	dml_print("DML: state_bbox: sr_enter_plus_exit_time_us = %f\n", state->sr_enter_plus_exit_time_us);
+	dml_print("DML: state_bbox: sr_exit_z8_time_us = %f\n", state->sr_exit_z8_time_us);
+	dml_print("DML: state_bbox: sr_enter_plus_exit_z8_time_us = %f\n", state->sr_enter_plus_exit_z8_time_us);
+	dml_print("DML: state_bbox: dram_clock_change_latency_us = %f\n", state->dram_clock_change_latency_us);
+	dml_print("DML: state_bbox: fclk_change_latency_us = %f\n", state->fclk_change_latency_us);
+	dml_print("DML: state_bbox: usr_retraining_latency_us = %f\n", state->usr_retraining_latency_us);
+	dml_print("DML: state_bbox: use_ideal_dram_bw_strobe = %d\n", state->use_ideal_dram_bw_strobe);
+}
+
+__DML_DLL_EXPORT__ void dml_print_soc_bounding_box(const struct soc_bounding_box_st *soc)
+{
+	dml_print("DML: soc_bbox: dprefclk_mhz = %f\n", soc->dprefclk_mhz);
+	dml_print("DML: soc_bbox: xtalclk_mhz = %f\n", soc->xtalclk_mhz);
+	dml_print("DML: soc_bbox: pcierefclk_mhz = %f\n", soc->pcierefclk_mhz);
+	dml_print("DML: soc_bbox: refclk_mhz = %f\n", soc->refclk_mhz);
+	dml_print("DML: soc_bbox: amclk_mhz = %f\n", soc->amclk_mhz);
+
+	dml_print("DML: soc_bbox: max_outstanding_reqs = %f\n", soc->max_outstanding_reqs);
+	dml_print("DML: soc_bbox: pct_ideal_sdp_bw_after_urgent = %f\n", soc->pct_ideal_sdp_bw_after_urgent);
+	dml_print("DML: soc_bbox: pct_ideal_fabric_bw_after_urgent = %f\n", soc->pct_ideal_fabric_bw_after_urgent);
+	dml_print("DML: soc_bbox: pct_ideal_dram_bw_after_urgent_pixel_only = %f\n", soc->pct_ideal_dram_bw_after_urgent_pixel_only);
+	dml_print("DML: soc_bbox: pct_ideal_dram_bw_after_urgent_pixel_and_vm = %f\n", soc->pct_ideal_dram_bw_after_urgent_pixel_and_vm);
+	dml_print("DML: soc_bbox: pct_ideal_dram_bw_after_urgent_vm_only = %f\n", soc->pct_ideal_dram_bw_after_urgent_vm_only);
+	dml_print("DML: soc_bbox: pct_ideal_dram_bw_after_urgent_strobe = %f\n", soc->pct_ideal_dram_bw_after_urgent_strobe);
+	dml_print("DML: soc_bbox: max_avg_sdp_bw_use_normal_percent = %f\n", soc->max_avg_sdp_bw_use_normal_percent);
+	dml_print("DML: soc_bbox: max_avg_fabric_bw_use_normal_percent = %f\n", soc->max_avg_fabric_bw_use_normal_percent);
+	dml_print("DML: soc_bbox: max_avg_dram_bw_use_normal_percent = %f\n", soc->max_avg_dram_bw_use_normal_percent);
+	dml_print("DML: soc_bbox: max_avg_dram_bw_use_normal_strobe_percent = %f\n", soc->max_avg_dram_bw_use_normal_strobe_percent);
+	dml_print("DML: soc_bbox: round_trip_ping_latency_dcfclk_cycles = %d\n", soc->round_trip_ping_latency_dcfclk_cycles);
+	dml_print("DML: soc_bbox: urgent_out_of_order_return_per_channel_pixel_only_bytes = %d\n", soc->urgent_out_of_order_return_per_channel_pixel_only_bytes);
+	dml_print("DML: soc_bbox: urgent_out_of_order_return_per_channel_pixel_and_vm_bytes = %d\n", soc->urgent_out_of_order_return_per_channel_pixel_and_vm_bytes);
+	dml_print("DML: soc_bbox: urgent_out_of_order_return_per_channel_vm_only_bytes = %d\n", soc->urgent_out_of_order_return_per_channel_vm_only_bytes);
+	dml_print("DML: soc_bbox: num_chans = %d\n", soc->num_chans);
+	dml_print("DML: soc_bbox: return_bus_width_bytes = %d\n", soc->return_bus_width_bytes);
+	dml_print("DML: soc_bbox: dram_channel_width_bytes = %d\n", soc->dram_channel_width_bytes);
+	dml_print("DML: soc_bbox: fabric_datapath_to_dcn_data_return_bytes = %d\n", soc->fabric_datapath_to_dcn_data_return_bytes);
+	dml_print("DML: soc_bbox: hostvm_min_page_size_kbytes = %d\n", soc->hostvm_min_page_size_kbytes);
+	dml_print("DML: soc_bbox: gpuvm_min_page_size_kbytes = %d\n", soc->gpuvm_min_page_size_kbytes);
+	dml_print("DML: soc_bbox: phy_downspread_percent = %f\n", soc->phy_downspread_percent);
+	dml_print("DML: soc_bbox: dcn_downspread_percent = %f\n", soc->dcn_downspread_percent);
+	dml_print("DML: soc_bbox: smn_latency_us = %f\n", soc->smn_latency_us);
+	dml_print("DML: soc_bbox: mall_allocated_for_dcn_mbytes = %d\n", soc->mall_allocated_for_dcn_mbytes);
+	dml_print("DML: soc_bbox: dispclk_dppclk_vco_speed_mhz = %f\n", soc->dispclk_dppclk_vco_speed_mhz);
+	dml_print("DML: soc_bbox: do_urgent_latency_adjustment = %d\n", soc->do_urgent_latency_adjustment);
+}
+
+__DML_DLL_EXPORT__ void dml_print_clk_cfg(const struct dml_clk_cfg_st *clk_cfg)
+{
+	dml_print("DML: clk_cfg: 0-use_required, 1-use pipe.clks_cfg, 2-use state bbox\n");
+	dml_print("DML: clk_cfg: dcfclk_option = %d\n", clk_cfg->dcfclk_option);
+	dml_print("DML: clk_cfg: dispclk_option = %d\n", clk_cfg->dispclk_option);
+
+	dml_print("DML: clk_cfg: dcfclk_mhz = %f\n", clk_cfg->dcfclk_mhz);
+	dml_print("DML: clk_cfg: dispclk_mhz = %f\n", clk_cfg->dispclk_mhz);
+
+	for (dml_uint_t i = 0; i < DCN_DML__NUM_PLANE; i++) {
+		dml_print("DML: clk_cfg: i=%d, dppclk_option = %d\n", i, clk_cfg->dppclk_option[i]);
+		dml_print("DML: clk_cfg: i=%d, dppclk_mhz = %f\n", i, clk_cfg->dppclk_mhz[i]);
+	}
+}
+
+dml_bool_t dml_is_vertical_rotation(enum dml_rotation_angle Scan)
+{
+	dml_bool_t is_vert = false;
+	if (Scan == dml_rotation_90 || Scan == dml_rotation_90m || Scan == dml_rotation_270 || Scan == dml_rotation_270m) {
+		is_vert = true;
+	} else {
+		is_vert = false;
+	}
+	return is_vert;
+} // dml_is_vertical_rotation
+
+dml_uint_t dml_get_cursor_bit_per_pixel(enum dml_cursor_bpp ebpp)
+{
+	switch (ebpp) {
+	case dml_cur_2bit:
+		return 2;
+	case dml_cur_32bit:
+		return 32;
+	case dml_cur_64bit:
+		return 64;
+	default:
+		return 0;
+	}
+}
+
+/// @brief Determine the physical pipe to logical plane mapping using the display_cfg
+dml_uint_t dml_get_num_active_planes(const struct dml_display_cfg_st *display_cfg)
+{
+	dml_uint_t num_active_planes = 0;
+
+	for (dml_uint_t k = 0; k < __DML_NUM_PLANES__; k++) {
+		if (display_cfg->plane.ViewportWidth[k] > 0)
+		num_active_planes = num_active_planes + 1;
+	}
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: num_active_planes = %d\n", __func__, num_active_planes);
+#endif
+	return num_active_planes;
+}
+
+/// @brief Determine the physical pipe to logical plane mapping using the display_cfg
+dml_uint_t dml_get_num_active_pipes(const struct dml_display_cfg_st *display_cfg)
+{
+	dml_uint_t num_active_pipes = 0;
+
+	for (dml_uint_t j = 0; j < dml_get_num_active_planes(display_cfg); j++) {
+		num_active_pipes = num_active_pipes + display_cfg->hw.DPPPerSurface[j];
+	}
+
+#ifdef __DML_VBA_DEBUG__
+	dml_print("DML::%s: num_active_pipes    = %d\n", __func__, num_active_pipes);
+#endif
+	return num_active_pipes;
+}
+
+dml_uint_t dml_get_plane_idx(const struct display_mode_lib_st *mode_lib, dml_uint_t pipe_idx)
+{
+	dml_uint_t plane_idx = mode_lib->mp.pipe_plane[pipe_idx];
+	return plane_idx;
+}
+
+dml_uint_t dml_get_pipe_idx(const struct display_mode_lib_st *mode_lib, dml_uint_t plane_idx)
+{
+	dml_uint_t     pipe_idx = 0;
+	dml_bool_t     pipe_found = 0;
+
+	ASSERT(plane_idx < __DML_NUM_PLANES__);
+
+	for (dml_uint_t i = 0; i < __DML_NUM_PLANES__; i++) {
+		if (plane_idx == mode_lib->mp.pipe_plane[i]) {
+			pipe_idx = i;
+			pipe_found = 1;
+			break;
+		}
+	}
+	ASSERT(pipe_found != 0);
+
+	return pipe_idx;
+}
+
+void dml_calc_pipe_plane_mapping(const struct dml_hw_resource_st *hw, dml_uint_t *pipe_plane)
+{
+	dml_uint_t pipe_idx = 0;
+
+	for (dml_uint_t k = 0; k < __DML_NUM_PLANES__; ++k) {
+		pipe_plane[k] = __DML_PIPE_NO_PLANE__;
+	}
+
+	for (dml_uint_t plane_idx = 0; plane_idx < __DML_NUM_PLANES__; plane_idx++) {
+		for (dml_uint_t i = 0; i < hw->DPPPerSurface[plane_idx]; i++) {
+			pipe_plane[pipe_idx] = plane_idx;
+			pipe_idx++;
+		}
+	}
+}
+
+
diff --git a/drivers/gpu/drm/amd/display/dc/dml2_0/display_mode_util.h b/drivers/gpu/drm/amd/display/dc/dml2_0/display_mode_util.h
new file mode 100644
index 000000000000..a82b49cf7fb0
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml2_0/display_mode_util.h
@@ -0,0 +1,74 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __DISPLAY_MODE_UTIL_H__
+#define __DISPLAY_MODE_UTIL_H__
+
+#include "display_mode_core_structs.h"
+#include "cmntypes.h"
+
+#include "dml_assert.h"
+#include "dml_logging.h"
+
+__DML_DLL_EXPORT__ dml_bool_t dml_util_is_420(enum dml_source_format_class source_format);
+__DML_DLL_EXPORT__ dml_float_t dml_ceil(dml_float_t x, dml_float_t granularity);
+__DML_DLL_EXPORT__ dml_float_t dml_floor(dml_float_t x, dml_float_t granularity);
+__DML_DLL_EXPORT__ dml_float_t dml_min(dml_float_t x, dml_float_t y);
+__DML_DLL_EXPORT__ dml_float_t dml_min3(dml_float_t x, dml_float_t y, dml_float_t z);
+__DML_DLL_EXPORT__ dml_float_t dml_min4(dml_float_t x, dml_float_t y, dml_float_t z, dml_float_t w);
+__DML_DLL_EXPORT__ dml_float_t dml_max(dml_float_t x, dml_float_t y);
+__DML_DLL_EXPORT__ dml_float_t dml_max3(dml_float_t x, dml_float_t y, dml_float_t z);
+__DML_DLL_EXPORT__ dml_float_t dml_max4(dml_float_t a, dml_float_t b, dml_float_t c, dml_float_t d);
+__DML_DLL_EXPORT__ dml_float_t dml_max5(dml_float_t a, dml_float_t b, dml_float_t c, dml_float_t d, dml_float_t e);
+__DML_DLL_EXPORT__ dml_float_t dml_log(dml_float_t x, dml_float_t base);
+__DML_DLL_EXPORT__ dml_float_t dml_log2(dml_float_t x);
+__DML_DLL_EXPORT__ dml_float_t dml_round(dml_float_t val, dml_bool_t bankers_rounding);
+__DML_DLL_EXPORT__ dml_float_t dml_pow(dml_float_t base, int exp);
+__DML_DLL_EXPORT__ dml_uint_t dml_round_to_multiple(dml_uint_t num, dml_uint_t multiple, dml_bool_t up);
+__DML_DLL_EXPORT__ dml_bool_t dml_is_vertical_rotation(enum dml_rotation_angle scan);
+__DML_DLL_EXPORT__ dml_uint_t  dml_get_cursor_bit_per_pixel(enum dml_cursor_bpp ebpp);
+__DML_DLL_EXPORT__ void dml_print_data_rq_regs_st(const dml_display_plane_rq_regs_st *data_rq_regs);
+__DML_DLL_EXPORT__ void dml_print_rq_regs_st(const dml_display_rq_regs_st *rq_regs);
+__DML_DLL_EXPORT__ void dml_print_dlg_regs_st(const dml_display_dlg_regs_st *dlg_regs);
+__DML_DLL_EXPORT__ void dml_print_ttu_regs_st(const dml_display_ttu_regs_st *ttu_regs);
+__DML_DLL_EXPORT__ void dml_print_dml_policy(const struct dml_mode_eval_policy_st *policy);
+__DML_DLL_EXPORT__ void dml_print_mode_support(struct display_mode_lib_st *mode_lib, dml_uint_t j);
+__DML_DLL_EXPORT__ void dml_print_dml_mode_support_info(const struct dml_mode_support_info_st *support, dml_bool_t fail_only);
+__DML_DLL_EXPORT__ void dml_print_dml_display_cfg_timing(const struct dml_timing_cfg_st *timing, dml_uint_t num_plane);
+__DML_DLL_EXPORT__ void dml_print_dml_display_cfg_plane(const struct dml_plane_cfg_st *plane, dml_uint_t num_plane);
+__DML_DLL_EXPORT__ void dml_print_dml_display_cfg_surface(const struct dml_surface_cfg_st *surface, dml_uint_t num_plane);
+__DML_DLL_EXPORT__ void dml_print_dml_display_cfg_hw_resource(const struct dml_hw_resource_st *hw, dml_uint_t num_plane);
+__DML_DLL_EXPORT__ void dml_print_soc_state_bounding_box(const struct soc_state_bounding_box_st *state);
+__DML_DLL_EXPORT__ void dml_print_soc_bounding_box(const struct soc_bounding_box_st *soc);
+__DML_DLL_EXPORT__ void dml_print_clk_cfg(const struct dml_clk_cfg_st *clk_cfg);
+
+__DML_DLL_EXPORT__ dml_uint_t dml_get_num_active_planes(const struct dml_display_cfg_st *display_cfg);
+__DML_DLL_EXPORT__ dml_uint_t dml_get_num_active_pipes(const struct dml_display_cfg_st *display_cfg);
+__DML_DLL_EXPORT__ dml_uint_t dml_get_plane_idx(const struct display_mode_lib_st *mode_lib, dml_uint_t pipe_idx);
+__DML_DLL_EXPORT__ dml_uint_t dml_get_pipe_idx(const struct display_mode_lib_st *mode_lib, dml_uint_t plane_idx);
+__DML_DLL_EXPORT__ void       dml_calc_pipe_plane_mapping(const struct dml_hw_resource_st *hw, dml_uint_t *pipe_plane);
+
+#endif
diff --git a/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/dml21_translation_helper.c b/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/dml21_translation_helper.c
new file mode 100644
index 000000000000..bf5e7f4e0416
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/dml21_translation_helper.c
@@ -0,0 +1,929 @@
+// SPDX-License-Identifier: MIT
+//
+// Copyright 2024 Advanced Micro Devices, Inc.
+
+#include "dml21_wrapper.h"
+#include "dml2_core_dcn4_calcs.h"
+#include "dml2_internal_shared_types.h"
+#include "dml2_internal_types.h"
+#include "dml21_utils.h"
+#include "dml21_translation_helper.h"
+#include "soc_and_ip_translator.h"
+
+static void dml21_populate_pmo_options(struct dml2_pmo_options *pmo_options,
+		const struct dc *in_dc,
+		const struct dml2_configuration_options *config)
+{
+	bool disable_fams2 = !in_dc->debug.fams2_config.bits.enable;
+
+	/* ODM options */
+	pmo_options->disable_dyn_odm = !config->minimize_dispclk_using_odm;
+	pmo_options->disable_dyn_odm_for_multi_stream = true;
+	pmo_options->disable_dyn_odm_for_stream_with_svp = true;
+
+	pmo_options->disable_vblank = ((in_dc->debug.dml21_disable_pstate_method_mask >> 1) & 1);
+
+	/* NOTE: DRR and SubVP Require FAMS2 */
+	pmo_options->disable_svp = ((in_dc->debug.dml21_disable_pstate_method_mask >> 2) & 1) ||
+			in_dc->debug.force_disable_subvp ||
+			disable_fams2;
+	pmo_options->disable_drr_clamped = ((in_dc->debug.dml21_disable_pstate_method_mask >> 3) & 1) ||
+			disable_fams2;
+	pmo_options->disable_drr_var = ((in_dc->debug.dml21_disable_pstate_method_mask >> 4) & 1) ||
+			disable_fams2;
+	pmo_options->disable_fams2 = disable_fams2;
+
+	pmo_options->disable_drr_var_when_var_active = in_dc->debug.disable_fams_gaming == INGAME_FAMS_DISABLE ||
+			in_dc->debug.disable_fams_gaming == INGAME_FAMS_MULTI_DISP_CLAMPED_ONLY;
+	pmo_options->disable_drr_clamped_when_var_active = in_dc->debug.disable_fams_gaming == INGAME_FAMS_DISABLE;
+}
+
+static enum dml2_project_id dml21_dcn_revision_to_dml2_project_id(enum dce_version dcn_version)
+{
+	enum dml2_project_id project_id;
+	switch (dcn_version) {
+	case DCN_VERSION_4_01:
+		project_id = dml2_project_dcn4x_stage2_auto_drr_svp;
+		break;
+	default:
+		project_id = dml2_project_invalid;
+		DC_ERR("unsupported dcn version for DML21!");
+		break;
+	}
+
+	return project_id;
+}
+
+void dml21_populate_dml_init_params(struct dml2_initialize_instance_in_out *dml_init,
+		const struct dml2_configuration_options *config,
+		const struct dc *in_dc)
+{
+	dml_init->options.project_id = dml21_dcn_revision_to_dml2_project_id(in_dc->ctx->dce_version);
+
+	if (config->use_native_soc_bb_construction) {
+		in_dc->soc_and_ip_translator->translator_funcs->get_soc_bb(&dml_init->soc_bb, in_dc, config);
+		in_dc->soc_and_ip_translator->translator_funcs->get_ip_caps(&dml_init->ip_caps);
+	} else {
+		dml_init->soc_bb = config->external_socbb_ip_params->soc_bb;
+		dml_init->ip_caps = config->external_socbb_ip_params->ip_params;
+	}
+
+	dml21_populate_pmo_options(&dml_init->options.pmo_options, in_dc, config);
+}
+
+static unsigned int calc_max_hardware_v_total(const struct dc_stream_state *stream)
+{
+	unsigned int max_hw_v_total = stream->ctx->dc->caps.max_v_total;
+
+	if (stream->ctx->dc->caps.vtotal_limited_by_fp2) {
+		max_hw_v_total -= stream->timing.v_front_porch + 1;
+	}
+
+	return max_hw_v_total;
+}
+
+static void populate_dml21_timing_config_from_stream_state(struct dml2_timing_cfg *timing,
+		struct dc_stream_state *stream,
+		struct pipe_ctx *pipe_ctx,
+		struct dml2_context *dml_ctx)
+{
+	unsigned int hblank_start, vblank_start, min_hardware_refresh_in_uhz;
+	uint32_t pix_clk_100hz;
+
+	timing->h_active = stream->timing.h_addressable + stream->timing.h_border_left + stream->timing.h_border_right + pipe_ctx->dsc_padding_params.dsc_hactive_padding;
+	timing->v_active = stream->timing.v_addressable + stream->timing.v_border_bottom + stream->timing.v_border_top;
+	timing->h_front_porch = stream->timing.h_front_porch;
+	timing->v_front_porch = stream->timing.v_front_porch;
+	timing->pixel_clock_khz = stream->timing.pix_clk_100hz / 10;
+	if (pipe_ctx->dsc_padding_params.dsc_hactive_padding != 0)
+		timing->pixel_clock_khz = pipe_ctx->dsc_padding_params.dsc_pix_clk_100hz / 10;
+	if (stream->timing.timing_3d_format == TIMING_3D_FORMAT_HW_FRAME_PACKING)
+		timing->pixel_clock_khz *= 2;
+	timing->h_total = stream->timing.h_total + pipe_ctx->dsc_padding_params.dsc_htotal_padding;
+	timing->v_total = stream->timing.v_total;
+	timing->h_sync_width = stream->timing.h_sync_width;
+	timing->interlaced = stream->timing.flags.INTERLACE;
+
+	hblank_start = stream->timing.h_total - stream->timing.h_front_porch;
+
+	timing->h_blank_end = hblank_start - stream->timing.h_addressable - pipe_ctx->dsc_padding_params.dsc_hactive_padding
+		- stream->timing.h_border_left - stream->timing.h_border_right;
+
+	if (hblank_start < stream->timing.h_addressable)
+		timing->h_blank_end = 0;
+
+	vblank_start = stream->timing.v_total - stream->timing.v_front_porch;
+
+	timing->v_blank_end = vblank_start - stream->timing.v_addressable
+		- stream->timing.v_border_top - stream->timing.v_border_bottom;
+
+	timing->drr_config.enabled = stream->ignore_msa_timing_param;
+	timing->drr_config.drr_active_variable = stream->vrr_active_variable;
+	timing->drr_config.drr_active_fixed = stream->vrr_active_fixed;
+	timing->drr_config.disallowed = !stream->allow_freesync;
+
+	/* limit min refresh rate to DC cap */
+	min_hardware_refresh_in_uhz = stream->timing.min_refresh_in_uhz;
+	if (stream->ctx->dc->caps.max_v_total != 0) {
+		if (pipe_ctx->dsc_padding_params.dsc_hactive_padding != 0) {
+			pix_clk_100hz = pipe_ctx->dsc_padding_params.dsc_pix_clk_100hz;
+		} else {
+			pix_clk_100hz = stream->timing.pix_clk_100hz;
+		}
+		min_hardware_refresh_in_uhz = div64_u64((pix_clk_100hz * 100000000ULL),
+				(timing->h_total * (long long)calc_max_hardware_v_total(stream)));
+	}
+
+	timing->drr_config.min_refresh_uhz = max(stream->timing.min_refresh_in_uhz, min_hardware_refresh_in_uhz);
+
+	if (dml_ctx->config.callbacks.get_max_flickerless_instant_vtotal_increase &&
+			stream->ctx->dc->config.enable_fpo_flicker_detection == 1)
+		timing->drr_config.max_instant_vtotal_delta = dml_ctx->config.callbacks.get_max_flickerless_instant_vtotal_increase(stream, false);
+	else
+		timing->drr_config.max_instant_vtotal_delta = 0;
+
+	if (stream->timing.flags.DSC) {
+		timing->dsc.enable = dml2_dsc_enable;
+		timing->dsc.overrides.num_slices = stream->timing.dsc_cfg.num_slices_h;
+		timing->dsc.dsc_compressed_bpp_x16 = stream->timing.dsc_cfg.bits_per_pixel;
+	} else
+		timing->dsc.enable = dml2_dsc_disable;
+
+	switch (stream->timing.display_color_depth) {
+	case COLOR_DEPTH_666:
+		timing->bpc = 6;
+		break;
+	case COLOR_DEPTH_888:
+		timing->bpc = 8;
+		break;
+	case COLOR_DEPTH_101010:
+		timing->bpc = 10;
+		break;
+	case COLOR_DEPTH_121212:
+		timing->bpc = 12;
+		break;
+	case COLOR_DEPTH_141414:
+		timing->bpc = 14;
+		break;
+	case COLOR_DEPTH_161616:
+		timing->bpc = 16;
+		break;
+	case COLOR_DEPTH_999:
+		timing->bpc = 9;
+		break;
+	case COLOR_DEPTH_111111:
+		timing->bpc = 11;
+		break;
+	default:
+		timing->bpc = 8;
+		break;
+	}
+
+	timing->vblank_nom = timing->v_total - timing->v_active;
+}
+
+static void populate_dml21_output_config_from_stream_state(struct dml2_link_output_cfg *output,
+		struct dc_stream_state *stream, const struct pipe_ctx *pipe)
+{
+	output->output_dp_lane_count = 4;
+
+	switch (stream->signal) {
+	case SIGNAL_TYPE_DISPLAY_PORT_MST:
+	case SIGNAL_TYPE_DISPLAY_PORT:
+		output->output_encoder = dml2_dp;
+		if (check_dp2p0_output_encoder(pipe))
+			output->output_encoder = dml2_dp2p0;
+		break;
+	case SIGNAL_TYPE_EDP:
+		output->output_encoder = dml2_edp;
+		break;
+	case SIGNAL_TYPE_HDMI_TYPE_A:
+	case SIGNAL_TYPE_DVI_SINGLE_LINK:
+	case SIGNAL_TYPE_DVI_DUAL_LINK:
+		output->output_encoder = dml2_hdmi;
+		break;
+	default:
+			output->output_encoder = dml2_dp;
+	}
+
+	switch (stream->timing.pixel_encoding) {
+	case PIXEL_ENCODING_RGB:
+	case PIXEL_ENCODING_YCBCR444:
+		output->output_format = dml2_444;
+		break;
+	case PIXEL_ENCODING_YCBCR420:
+		output->output_format = dml2_420;
+		break;
+	case PIXEL_ENCODING_YCBCR422:
+		if (stream->timing.flags.DSC && !stream->timing.dsc_cfg.ycbcr422_simple)
+			output->output_format = dml2_n422;
+		else
+			output->output_format = dml2_s422;
+		break;
+	default:
+		output->output_format = dml2_444;
+		break;
+	}
+
+	switch (stream->signal) {
+	case SIGNAL_TYPE_NONE:
+	case SIGNAL_TYPE_DVI_SINGLE_LINK:
+	case SIGNAL_TYPE_DVI_DUAL_LINK:
+	case SIGNAL_TYPE_HDMI_TYPE_A:
+	case SIGNAL_TYPE_LVDS:
+	case SIGNAL_TYPE_RGB:
+	case SIGNAL_TYPE_DISPLAY_PORT:
+	case SIGNAL_TYPE_DISPLAY_PORT_MST:
+	case SIGNAL_TYPE_EDP:
+	case SIGNAL_TYPE_VIRTUAL:
+	default:
+		output->output_dp_link_rate = dml2_dp_rate_na;
+		break;
+	}
+
+	output->audio_sample_layout = stream->audio_info.modes->sample_size;
+	output->audio_sample_rate = stream->audio_info.modes->max_bit_rate;
+	output->output_disabled = true;
+
+	//TODO : New to DML2.1. How do we populate this ?
+	// output->validate_output
+}
+
+static void populate_dml21_stream_overrides_from_stream_state(
+		struct dml2_stream_parameters *stream_desc,
+		struct dc_stream_state *stream,
+		struct dc_stream_status *stream_status)
+{
+	switch (stream->debug.force_odm_combine_segments) {
+	case 0:
+		stream_desc->overrides.odm_mode = dml2_odm_mode_auto;
+		break;
+	case 1:
+		stream_desc->overrides.odm_mode = dml2_odm_mode_bypass;
+		break;
+	case 2:
+		stream_desc->overrides.odm_mode = dml2_odm_mode_combine_2to1;
+		break;
+	case 3:
+		stream_desc->overrides.odm_mode = dml2_odm_mode_combine_3to1;
+		break;
+	case 4:
+		stream_desc->overrides.odm_mode = dml2_odm_mode_combine_4to1;
+		break;
+	default:
+		stream_desc->overrides.odm_mode =  dml2_odm_mode_auto;
+		break;
+	}
+	if (!stream->ctx->dc->debug.enable_single_display_2to1_odm_policy ||
+			stream->debug.force_odm_combine_segments > 0)
+		stream_desc->overrides.disable_dynamic_odm = true;
+	stream_desc->overrides.disable_subvp = stream->ctx->dc->debug.force_disable_subvp ||
+			stream->hw_cursor_req ||
+			stream_status->mall_stream_config.cursor_size_limit_subvp;
+}
+
+static enum dml2_swizzle_mode gfx_addr3_to_dml2_swizzle_mode(enum swizzle_mode_addr3_values addr3_mode)
+{
+	enum dml2_swizzle_mode dml2_mode = dml2_sw_linear;
+
+	switch (addr3_mode) {
+	case DC_ADDR3_SW_LINEAR:
+		dml2_mode = dml2_sw_linear;
+		break;
+	case DC_ADDR3_SW_256B_2D:
+		dml2_mode = dml2_sw_256b_2d;
+		break;
+	case DC_ADDR3_SW_4KB_2D:
+		dml2_mode = dml2_sw_4kb_2d;
+		break;
+	case DC_ADDR3_SW_64KB_2D:
+		dml2_mode = dml2_sw_64kb_2d;
+		break;
+	case DC_ADDR3_SW_256KB_2D:
+		dml2_mode = dml2_sw_256kb_2d;
+		break;
+	default:
+		/* invalid swizzle mode for DML2.1 */
+		ASSERT(false);
+		dml2_mode = dml2_sw_linear;
+	}
+
+	return dml2_mode;
+}
+
+static enum dml2_swizzle_mode gfx9_to_dml2_swizzle_mode(enum swizzle_mode_values gfx9_mode)
+{
+	enum dml2_swizzle_mode dml2_mode = dml2_sw_64kb_2d;
+
+	switch (gfx9_mode) {
+	case DC_SW_LINEAR:
+		dml2_mode = dml2_sw_linear;
+		break;
+	case DC_SW_256_D:
+	case DC_SW_256_R:
+		dml2_mode = dml2_sw_256b_2d;
+		break;
+	case DC_SW_4KB_D:
+	case DC_SW_4KB_R:
+	case DC_SW_4KB_R_X:
+		dml2_mode = dml2_sw_4kb_2d;
+		break;
+	case DC_SW_64KB_D:
+	case DC_SW_64KB_D_X:
+	case DC_SW_64KB_R:
+	case DC_SW_64KB_R_X:
+		dml2_mode = dml2_sw_64kb_2d;
+		break;
+	case DC_SW_256B_S:
+	case DC_SW_4KB_S:
+	case DC_SW_64KB_S:
+	case DC_SW_VAR_S:
+	case DC_SW_VAR_D:
+	case DC_SW_VAR_R:
+	case DC_SW_64KB_S_T:
+	case DC_SW_64KB_D_T:
+	case DC_SW_4KB_S_X:
+	case DC_SW_4KB_D_X:
+	case DC_SW_64KB_S_X:
+	case DC_SW_VAR_S_X:
+	case DC_SW_VAR_D_X:
+	case DC_SW_VAR_R_X:
+	default:
+		/*
+		 * invalid swizzle mode for DML2.1. This could happen because
+		 * DML21 is not intended to be used by N-1 in production. To
+		 * properly filter out unsupported swizzle modes, we will need
+		 * to fix capability reporting when DML2.1 is used for N-1 in
+		 * dc. So DML will only receive DML21 supported swizzle modes.
+		 * This implementation is not added and has a low value because
+		 * the supported swizzle modes should already cover most of our
+		 * N-1 test cases.
+		 */
+		return dml2_sw_64kb_2d;
+	}
+
+	return dml2_mode;
+}
+
+static void populate_dml21_dummy_surface_cfg(struct dml2_surface_cfg *surface, const struct dc_stream_state *stream)
+{
+	surface->plane0.width = stream->timing.h_addressable;
+	surface->plane0.height = stream->timing.v_addressable;
+	surface->plane1.width = stream->timing.h_addressable;
+	surface->plane1.height = stream->timing.v_addressable;
+	surface->plane0.pitch = ((surface->plane0.width + 127) / 128) * 128;
+	surface->plane1.pitch = 0;
+	surface->dcc.enable = false;
+	surface->dcc.informative.dcc_rate_plane0 = 1.0;
+	surface->dcc.informative.dcc_rate_plane1 = 1.0;
+	surface->dcc.informative.fraction_of_zero_size_request_plane0 = 0;
+	surface->dcc.informative.fraction_of_zero_size_request_plane1 = 0;
+	surface->tiling = dml2_sw_64kb_2d;
+}
+
+static void populate_dml21_dummy_plane_cfg(struct dml2_plane_parameters *plane, const struct dc_stream_state *stream)
+{
+	unsigned int width, height;
+
+	if (stream->timing.h_addressable > 3840)
+		width = 3840;
+	else
+		width = stream->timing.h_addressable;	// 4K max
+
+	if (stream->timing.v_addressable > 2160)
+		height = 2160;
+	else
+		height = stream->timing.v_addressable;	// 4K max
+
+	plane->cursor.cursor_bpp = 32;
+
+	plane->cursor.cursor_width = 256;
+	plane->cursor.num_cursors = 1;
+
+	plane->composition.viewport.plane0.width = width;
+	plane->composition.viewport.plane0.height = height;
+	plane->composition.viewport.plane1.width = 0;
+	plane->composition.viewport.plane1.height = 0;
+
+	plane->composition.viewport.stationary = false;
+	plane->composition.viewport.plane0.x_start = 0;
+	plane->composition.viewport.plane0.y_start = 0;
+	plane->composition.viewport.plane1.x_start = 0;
+	plane->composition.viewport.plane1.y_start = 0;
+
+	plane->composition.scaler_info.enabled = false;
+	plane->composition.rotation_angle = dml2_rotation_0;
+	plane->composition.scaler_info.plane0.h_ratio = 1.0;
+	plane->composition.scaler_info.plane0.v_ratio = 1.0;
+	plane->composition.scaler_info.plane1.h_ratio = 0;
+	plane->composition.scaler_info.plane1.v_ratio = 0;
+	plane->composition.scaler_info.plane0.h_taps = 1;
+	plane->composition.scaler_info.plane0.v_taps = 1;
+	plane->composition.scaler_info.plane1.h_taps = 0;
+	plane->composition.scaler_info.plane1.v_taps = 0;
+	plane->composition.scaler_info.rect_out_width = width;
+	plane->pixel_format = dml2_444_32;
+
+	plane->dynamic_meta_data.enable = false;
+	plane->overrides.gpuvm_min_page_size_kbytes = 256;
+}
+
+static void populate_dml21_surface_config_from_plane_state(
+		const struct dc *in_dc,
+		struct dml2_surface_cfg *surface,
+		const struct dc_plane_state *plane_state)
+{
+	surface->plane0.pitch = plane_state->plane_size.surface_pitch;
+	surface->plane1.pitch = plane_state->plane_size.chroma_pitch;
+	surface->plane0.height = plane_state->plane_size.surface_size.height;
+	surface->plane0.width = plane_state->plane_size.surface_size.width;
+	surface->plane1.height = plane_state->plane_size.chroma_size.height;
+	surface->plane1.width = plane_state->plane_size.chroma_size.width;
+	surface->dcc.enable = plane_state->dcc.enable;
+	surface->dcc.informative.dcc_rate_plane0 = 1.0;
+	surface->dcc.informative.dcc_rate_plane1 = 1.0;
+	surface->dcc.informative.fraction_of_zero_size_request_plane0 = plane_state->dcc.independent_64b_blks;
+	surface->dcc.informative.fraction_of_zero_size_request_plane1 = plane_state->dcc.independent_64b_blks_c;
+	surface->dcc.plane0.pitch = plane_state->dcc.meta_pitch;
+	surface->dcc.plane1.pitch = plane_state->dcc.meta_pitch_c;
+
+	// Update swizzle / array mode based on the gfx_format
+	switch (plane_state->tiling_info.gfxversion) {
+	case DcGfxVersion7:
+	case DcGfxVersion8:
+		break;
+	case DcGfxVersion9:
+	case DcGfxVersion10:
+	case DcGfxVersion11:
+		surface->tiling = gfx9_to_dml2_swizzle_mode(plane_state->tiling_info.gfx9.swizzle);
+		break;
+	case DcGfxAddr3:
+		surface->tiling = gfx_addr3_to_dml2_swizzle_mode(plane_state->tiling_info.gfx_addr3.swizzle);
+		break;
+	}
+}
+
+static const struct scaler_data *get_scaler_data_for_plane(
+		struct dml2_context *dml_ctx,
+		const struct dc_plane_state *in,
+		const struct dc_state *context)
+{
+	int i;
+	struct pipe_ctx *temp_pipe = &dml_ctx->v21.scratch.temp_pipe;
+
+	memset(temp_pipe, 0, sizeof(struct pipe_ctx));
+
+	for (i = 0; i < MAX_PIPES; i++)	{
+		const struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
+
+		if (pipe->plane_state == in && !pipe->prev_odm_pipe) {
+			temp_pipe->stream = pipe->stream;
+			temp_pipe->plane_state = pipe->plane_state;
+			temp_pipe->plane_res.scl_data.taps = pipe->plane_res.scl_data.taps;
+			temp_pipe->stream_res = pipe->stream_res;
+			temp_pipe->dsc_padding_params.dsc_hactive_padding = pipe->dsc_padding_params.dsc_hactive_padding;
+			temp_pipe->dsc_padding_params.dsc_htotal_padding = pipe->dsc_padding_params.dsc_htotal_padding;
+			temp_pipe->dsc_padding_params.dsc_pix_clk_100hz = pipe->dsc_padding_params.dsc_pix_clk_100hz;
+			dml_ctx->config.callbacks.build_scaling_params(temp_pipe);
+			break;
+		}
+	}
+
+	ASSERT(i < MAX_PIPES);
+	return &temp_pipe->plane_res.scl_data;
+}
+
+static void populate_dml21_plane_config_from_plane_state(struct dml2_context *dml_ctx,
+		struct dml2_plane_parameters *plane, const struct dc_plane_state *plane_state,
+		const struct dc_state *context, unsigned int stream_index)
+{
+	const struct scaler_data *scaler_data = get_scaler_data_for_plane(dml_ctx, plane_state, context);
+	struct dc_stream_state *stream = context->streams[stream_index];
+
+	plane->cursor.cursor_bpp = 32;
+	plane->cursor.cursor_width = 256;
+	plane->cursor.num_cursors = 1;
+
+	switch (plane_state->format) {
+	case SURFACE_PIXEL_FORMAT_VIDEO_420_YCbCr:
+	case SURFACE_PIXEL_FORMAT_VIDEO_420_YCrCb:
+		plane->pixel_format = dml2_420_8;
+		break;
+	case SURFACE_PIXEL_FORMAT_VIDEO_420_10bpc_YCbCr:
+	case SURFACE_PIXEL_FORMAT_VIDEO_420_10bpc_YCrCb:
+		plane->pixel_format = dml2_420_10;
+		break;
+	case SURFACE_PIXEL_FORMAT_GRPH_ARGB16161616:
+	case SURFACE_PIXEL_FORMAT_GRPH_ABGR16161616:
+	case SURFACE_PIXEL_FORMAT_GRPH_ARGB16161616F:
+	case SURFACE_PIXEL_FORMAT_GRPH_ABGR16161616F:
+		plane->pixel_format = dml2_444_64;
+		break;
+	case SURFACE_PIXEL_FORMAT_GRPH_ARGB1555:
+	case SURFACE_PIXEL_FORMAT_GRPH_RGB565:
+		plane->pixel_format = dml2_444_16;
+		break;
+	case SURFACE_PIXEL_FORMAT_GRPH_PALETA_256_COLORS:
+		plane->pixel_format = dml2_444_8;
+		break;
+	case SURFACE_PIXEL_FORMAT_GRPH_RGBE_ALPHA:
+		plane->pixel_format = dml2_rgbe_alpha;
+		break;
+	default:
+		plane->pixel_format = dml2_444_32;
+		break;
+	}
+
+	plane->composition.viewport.plane0.height = scaler_data->viewport.height;
+	plane->composition.viewport.plane0.width = scaler_data->viewport.width;
+	plane->composition.viewport.plane1.height = scaler_data->viewport_c.height;
+	plane->composition.viewport.plane1.width = scaler_data->viewport_c.width;
+	plane->composition.viewport.plane0.x_start = scaler_data->viewport.x;
+	plane->composition.viewport.plane0.y_start = scaler_data->viewport.y;
+	plane->composition.viewport.plane1.x_start = scaler_data->viewport_c.x;
+	plane->composition.viewport.plane1.y_start = scaler_data->viewport_c.y;
+	plane->composition.viewport.stationary = false;
+	plane->composition.scaler_info.enabled = scaler_data->ratios.horz.value != dc_fixpt_one.value ||
+		scaler_data->ratios.horz_c.value != dc_fixpt_one.value ||
+		scaler_data->ratios.vert.value != dc_fixpt_one.value ||
+		scaler_data->ratios.vert_c.value != dc_fixpt_one.value;
+
+	if (!scaler_data->taps.h_taps) {
+		/* Above logic determines scaling should be enabled even when there are no taps for
+		 * certain cases. Hence do corrective active and disable scaling.
+		 */
+		plane->composition.scaler_info.enabled = false;
+	} else if ((plane_state->ctx->dc->config.use_spl == true) &&
+		(plane->composition.scaler_info.enabled == false)) {
+		/* To enable sharpener for 1:1, scaler must be enabled.  If use_spl is set, then
+		 *  allow case where ratio is 1 but taps > 1
+		 */
+		if ((scaler_data->taps.h_taps > 1) || (scaler_data->taps.v_taps > 1) ||
+			(scaler_data->taps.h_taps_c > 1) || (scaler_data->taps.v_taps_c > 1))
+			plane->composition.scaler_info.enabled = true;
+	}
+
+	/* always_scale is only used for debug purposes not used in production but has to be
+	 * maintained for certain complainces. */
+	if (plane_state->ctx->dc->debug.always_scale == true) {
+		plane->composition.scaler_info.enabled = true;
+	}
+
+	if (plane->composition.scaler_info.enabled == false) {
+		plane->composition.scaler_info.plane0.h_ratio = 1.0;
+		plane->composition.scaler_info.plane0.v_ratio = 1.0;
+		plane->composition.scaler_info.plane1.h_ratio = 1.0;
+		plane->composition.scaler_info.plane1.v_ratio = 1.0;
+	} else {
+		plane->composition.scaler_info.plane0.h_ratio = (double)scaler_data->ratios.horz.value / (1ULL << 32);
+		plane->composition.scaler_info.plane0.v_ratio = (double)scaler_data->ratios.vert.value / (1ULL << 32);
+		plane->composition.scaler_info.plane1.h_ratio = (double)scaler_data->ratios.horz_c.value / (1ULL << 32);
+		plane->composition.scaler_info.plane1.v_ratio = (double)scaler_data->ratios.vert_c.value / (1ULL << 32);
+	}
+
+	if (!scaler_data->taps.h_taps) {
+		plane->composition.scaler_info.plane0.h_taps = 1;
+		plane->composition.scaler_info.plane1.h_taps = 1;
+	} else {
+		plane->composition.scaler_info.plane0.h_taps = scaler_data->taps.h_taps;
+		plane->composition.scaler_info.plane1.h_taps = scaler_data->taps.h_taps_c;
+	}
+	if (!scaler_data->taps.v_taps) {
+		plane->composition.scaler_info.plane0.v_taps = 1;
+		plane->composition.scaler_info.plane1.v_taps = 1;
+	} else {
+		plane->composition.scaler_info.plane0.v_taps = scaler_data->taps.v_taps;
+		plane->composition.scaler_info.plane1.v_taps = scaler_data->taps.v_taps_c;
+	}
+
+	plane->composition.viewport.stationary = false;
+
+	if (plane_state->mcm_luts.lut3d_data.lut3d_src == DC_CM2_TRANSFER_FUNC_SOURCE_VIDMEM) {
+		plane->tdlut.setup_for_tdlut = true;
+
+		switch (plane_state->mcm_luts.lut3d_data.gpu_mem_params.layout) {
+		case DC_CM2_GPU_MEM_LAYOUT_3D_SWIZZLE_LINEAR_RGB:
+		case DC_CM2_GPU_MEM_LAYOUT_3D_SWIZZLE_LINEAR_BGR:
+			plane->tdlut.tdlut_addressing_mode = dml2_tdlut_sw_linear;
+			break;
+		case DC_CM2_GPU_MEM_LAYOUT_1D_PACKED_LINEAR:
+			plane->tdlut.tdlut_addressing_mode = dml2_tdlut_simple_linear;
+			break;
+		}
+
+		switch (plane_state->mcm_luts.lut3d_data.gpu_mem_params.size) {
+		case DC_CM2_GPU_MEM_SIZE_171717:
+			plane->tdlut.tdlut_width_mode = dml2_tdlut_width_17_cube;
+			break;
+		case DC_CM2_GPU_MEM_SIZE_TRANSFORMED:
+		default:
+			//plane->tdlut.tdlut_width_mode = dml2_tdlut_width_flatten; // dml2_tdlut_width_flatten undefined
+			break;
+		}
+	}
+	plane->tdlut.setup_for_tdlut |= dml_ctx->config.force_tdlut_enable;
+
+	plane->dynamic_meta_data.enable = false;
+	plane->dynamic_meta_data.lines_before_active_required = 0;
+	plane->dynamic_meta_data.transmitted_bytes = 0;
+
+	plane->composition.scaler_info.rect_out_width = plane_state->dst_rect.width;
+	plane->composition.rotation_angle = (enum dml2_rotation_angle) plane_state->rotation;
+	plane->stream_index = stream_index;
+
+	plane->overrides.gpuvm_min_page_size_kbytes = 256;
+
+	plane->immediate_flip = plane_state->flip_immediate;
+
+	plane->composition.rect_out_height_spans_vactive =
+		plane_state->dst_rect.height >= stream->src.height &&
+		stream->dst.height >= stream->timing.v_addressable;
+}
+
+//TODO : Could be possibly moved to a common helper layer.
+static bool dml21_wrapper_get_plane_id(const struct dc_state *context, unsigned int stream_id, const struct dc_plane_state *plane, unsigned int *plane_id)
+{
+	int i, j;
+
+	if (!plane_id)
+		return false;
+
+	for (i = 0; i < context->stream_count; i++) {
+		if (context->streams[i]->stream_id == stream_id) {
+			for (j = 0; j < context->stream_status[i].plane_count; j++) {
+				if (context->stream_status[i].plane_states[j] == plane) {
+					*plane_id = (i << 16) | j;
+					return true;
+				}
+			}
+		}
+	}
+
+	return false;
+}
+
+static unsigned int map_stream_to_dml21_display_cfg(const struct dml2_context *dml_ctx, const struct dc_stream_state *stream)
+{
+	int i = 0;
+	int location = -1;
+
+	for (i = 0; i < __DML2_WRAPPER_MAX_STREAMS_PLANES__; i++) {
+		if (dml_ctx->v21.dml_to_dc_pipe_mapping.disp_cfg_to_stream_id_valid[i] && dml_ctx->v21.dml_to_dc_pipe_mapping.disp_cfg_to_stream_id[i] == stream->stream_id) {
+			location = i;
+			break;
+		}
+	}
+
+	return location;
+}
+
+unsigned int map_plane_to_dml21_display_cfg(const struct dml2_context *dml_ctx, unsigned int stream_id,
+		const struct dc_plane_state *plane, const struct dc_state *context)
+{
+	unsigned int plane_id;
+	int i = 0;
+	int location = -1;
+
+	if (!dml21_wrapper_get_plane_id(context, stream_id, plane, &plane_id)) {
+		ASSERT(false);
+		return -1;
+	}
+
+	for (i = 0; i < __DML2_WRAPPER_MAX_STREAMS_PLANES__; i++) {
+		if (dml_ctx->v21.dml_to_dc_pipe_mapping.disp_cfg_to_plane_id_valid[i] && dml_ctx->v21.dml_to_dc_pipe_mapping.disp_cfg_to_plane_id[i] == plane_id) {
+			location = i;
+			break;
+		}
+	}
+
+	return location;
+}
+
+static enum dml2_uclk_pstate_change_strategy dml21_force_pstate_method_to_uclk_state_change_strategy(enum dml2_force_pstate_methods force_pstate_method)
+{
+	enum dml2_uclk_pstate_change_strategy val = dml2_uclk_pstate_change_strategy_auto;
+
+	switch (force_pstate_method) {
+	case dml2_force_pstate_method_vactive:
+		val = dml2_uclk_pstate_change_strategy_force_vactive;
+		break;
+	case dml2_force_pstate_method_vblank:
+		val = dml2_uclk_pstate_change_strategy_force_vblank;
+		break;
+	case dml2_force_pstate_method_drr:
+		val = dml2_uclk_pstate_change_strategy_force_drr;
+		break;
+	case dml2_force_pstate_method_subvp:
+		val = dml2_uclk_pstate_change_strategy_force_mall_svp;
+		break;
+	case dml2_force_pstate_method_auto:
+	default:
+		val = dml2_uclk_pstate_change_strategy_auto;
+	}
+
+	return val;
+}
+
+bool dml21_map_dc_state_into_dml_display_cfg(const struct dc *in_dc, struct dc_state *context, struct dml2_context *dml_ctx)
+{
+	int stream_index, plane_index;
+	int disp_cfg_stream_location, disp_cfg_plane_location;
+	struct dml2_display_cfg *dml_dispcfg = &dml_ctx->v21.display_config;
+	unsigned int plane_count = 0;
+
+	memset(&dml_ctx->v21.dml_to_dc_pipe_mapping, 0, sizeof(struct dml2_dml_to_dc_pipe_mapping));
+
+	dml_dispcfg->gpuvm_enable = dml_ctx->config.gpuvm_enable;
+	dml_dispcfg->gpuvm_max_page_table_levels = 4;
+	dml_dispcfg->hostvm_enable = false;
+	dml_dispcfg->minimize_det_reallocation = true;
+	dml_dispcfg->overrides.enable_subvp_implicit_pmo = true;
+
+	if (in_dc->debug.disable_unbounded_requesting) {
+		dml_dispcfg->overrides.hw.force_unbounded_requesting.enable = true;
+		dml_dispcfg->overrides.hw.force_unbounded_requesting.value = false;
+	}
+
+	for (stream_index = 0; stream_index < context->stream_count; stream_index++) {
+		disp_cfg_stream_location = map_stream_to_dml21_display_cfg(dml_ctx, context->streams[stream_index]);
+
+		if (disp_cfg_stream_location < 0)
+			disp_cfg_stream_location = dml_dispcfg->num_streams++;
+
+		ASSERT(disp_cfg_stream_location >= 0 && disp_cfg_stream_location < __DML2_WRAPPER_MAX_STREAMS_PLANES__);
+		populate_dml21_timing_config_from_stream_state(&dml_dispcfg->stream_descriptors[disp_cfg_stream_location].timing, context->streams[stream_index], &context->res_ctx.pipe_ctx[stream_index], dml_ctx);
+		populate_dml21_output_config_from_stream_state(&dml_dispcfg->stream_descriptors[disp_cfg_stream_location].output, context->streams[stream_index], &context->res_ctx.pipe_ctx[stream_index]);
+		populate_dml21_stream_overrides_from_stream_state(&dml_dispcfg->stream_descriptors[disp_cfg_stream_location], context->streams[stream_index], &context->stream_status[stream_index]);
+
+		dml_dispcfg->stream_descriptors[disp_cfg_stream_location].overrides.hw.twait_budgeting.fclk_pstate = dml2_twait_budgeting_setting_if_needed;
+		dml_dispcfg->stream_descriptors[disp_cfg_stream_location].overrides.hw.twait_budgeting.uclk_pstate = dml2_twait_budgeting_setting_if_needed;
+		dml_dispcfg->stream_descriptors[disp_cfg_stream_location].overrides.hw.twait_budgeting.stutter_enter_exit = dml2_twait_budgeting_setting_if_needed;
+
+		dml_ctx->v21.dml_to_dc_pipe_mapping.disp_cfg_to_stream_id[disp_cfg_stream_location] = context->streams[stream_index]->stream_id;
+		dml_ctx->v21.dml_to_dc_pipe_mapping.disp_cfg_to_stream_id_valid[disp_cfg_stream_location] = true;
+
+		if (context->stream_status[stream_index].plane_count == 0) {
+			disp_cfg_plane_location = dml_dispcfg->num_planes++;
+			populate_dml21_dummy_surface_cfg(&dml_dispcfg->plane_descriptors[disp_cfg_plane_location].surface, context->streams[stream_index]);
+			populate_dml21_dummy_plane_cfg(&dml_dispcfg->plane_descriptors[disp_cfg_plane_location], context->streams[stream_index]);
+			dml_dispcfg->plane_descriptors[disp_cfg_plane_location].stream_index = disp_cfg_stream_location;
+		} else {
+			for (plane_index = 0; plane_index < context->stream_status[stream_index].plane_count; plane_index++) {
+				disp_cfg_plane_location = map_plane_to_dml21_display_cfg(dml_ctx, context->streams[stream_index]->stream_id, context->stream_status[stream_index].plane_states[plane_index], context);
+
+				if (disp_cfg_plane_location < 0)
+					disp_cfg_plane_location = dml_dispcfg->num_planes++;
+
+				ASSERT(disp_cfg_plane_location >= 0 && disp_cfg_plane_location < __DML2_WRAPPER_MAX_STREAMS_PLANES__);
+
+				populate_dml21_surface_config_from_plane_state(in_dc, &dml_dispcfg->plane_descriptors[disp_cfg_plane_location].surface, context->stream_status[stream_index].plane_states[plane_index]);
+				populate_dml21_plane_config_from_plane_state(dml_ctx, &dml_dispcfg->plane_descriptors[disp_cfg_plane_location], context->stream_status[stream_index].plane_states[plane_index], context, stream_index);
+				dml_dispcfg->plane_descriptors[disp_cfg_plane_location].stream_index = disp_cfg_stream_location;
+
+				if (dml21_wrapper_get_plane_id(context, context->streams[stream_index]->stream_id, context->stream_status[stream_index].plane_states[plane_index], &dml_ctx->v21.dml_to_dc_pipe_mapping.disp_cfg_to_plane_id[disp_cfg_plane_location]))
+					dml_ctx->v21.dml_to_dc_pipe_mapping.disp_cfg_to_plane_id_valid[disp_cfg_plane_location] = true;
+
+				/* apply forced pstate policy */
+				if (dml_ctx->config.pmo.force_pstate_method_enable) {
+					dml_dispcfg->plane_descriptors[disp_cfg_plane_location].overrides.uclk_pstate_change_strategy =
+							dml21_force_pstate_method_to_uclk_state_change_strategy(dml_ctx->config.pmo.force_pstate_method_values[stream_index]);
+				}
+
+				plane_count++;
+			}
+		}
+	}
+
+	if (plane_count == 0) {
+		dml_dispcfg->overrides.all_streams_blanked = true;
+	}
+
+	return true;
+}
+
+void dml21_copy_clocks_to_dc_state(struct dml2_context *in_ctx, struct dc_state *context)
+{
+	/* TODO these should be the max of active, svp prefetch and idle should be tracked seperately */
+	context->bw_ctx.bw.dcn.clk.dispclk_khz = in_ctx->v21.mode_programming.programming->min_clocks.dcn4x.dispclk_khz;
+	context->bw_ctx.bw.dcn.clk.dcfclk_khz = in_ctx->v21.mode_programming.programming->min_clocks.dcn4x.active.dcfclk_khz;
+	context->bw_ctx.bw.dcn.clk.dramclk_khz = in_ctx->v21.mode_programming.programming->min_clocks.dcn4x.active.uclk_khz;
+	context->bw_ctx.bw.dcn.clk.fclk_khz = in_ctx->v21.mode_programming.programming->min_clocks.dcn4x.active.fclk_khz;
+	context->bw_ctx.bw.dcn.clk.idle_dramclk_khz = in_ctx->v21.mode_programming.programming->min_clocks.dcn4x.idle.uclk_khz;
+	context->bw_ctx.bw.dcn.clk.idle_fclk_khz = in_ctx->v21.mode_programming.programming->min_clocks.dcn4x.idle.fclk_khz;
+	context->bw_ctx.bw.dcn.clk.dcfclk_deep_sleep_khz = in_ctx->v21.mode_programming.programming->min_clocks.dcn4x.deepsleep_dcfclk_khz;
+	context->bw_ctx.bw.dcn.clk.fclk_p_state_change_support = in_ctx->v21.mode_programming.programming->fclk_pstate_supported;
+	context->bw_ctx.bw.dcn.clk.p_state_change_support = in_ctx->v21.mode_programming.programming->uclk_pstate_supported;
+	context->bw_ctx.bw.dcn.clk.dtbclk_en = in_ctx->v21.mode_programming.programming->min_clocks.dcn4x.dtbrefclk_khz > 0;
+	context->bw_ctx.bw.dcn.clk.ref_dtbclk_khz = in_ctx->v21.mode_programming.programming->min_clocks.dcn4x.dtbrefclk_khz;
+	context->bw_ctx.bw.dcn.clk.socclk_khz = in_ctx->v21.mode_programming.programming->min_clocks.dcn4x.socclk_khz;
+	context->bw_ctx.bw.dcn.clk.subvp_prefetch_dramclk_khz = in_ctx->v21.mode_programming.programming->min_clocks.dcn4x.svp_prefetch_no_throttle.uclk_khz;
+	context->bw_ctx.bw.dcn.clk.subvp_prefetch_fclk_khz = in_ctx->v21.mode_programming.programming->min_clocks.dcn4x.svp_prefetch_no_throttle.fclk_khz;
+	context->bw_ctx.bw.dcn.clk.stutter_efficiency.base_efficiency = in_ctx->v21.mode_programming.programming->stutter.base_percent_efficiency;
+	context->bw_ctx.bw.dcn.clk.stutter_efficiency.low_power_efficiency = in_ctx->v21.mode_programming.programming->stutter.low_power_percent_efficiency;
+}
+
+static struct dml2_dchub_watermark_regs *wm_set_index_to_dc_wm_set(union dcn_watermark_set *watermarks, const enum dml2_dchub_watermark_reg_set_index wm_index)
+{
+	struct dml2_dchub_watermark_regs *wm_regs = NULL;
+
+	switch (wm_index) {
+	case DML2_DCHUB_WATERMARK_SET_A:
+		wm_regs = &watermarks->dcn4x.a;
+		break;
+	case DML2_DCHUB_WATERMARK_SET_B:
+		wm_regs = &watermarks->dcn4x.b;
+		break;
+	case DML2_DCHUB_WATERMARK_SET_C:
+		wm_regs = &watermarks->dcn4x.c;
+		break;
+	case DML2_DCHUB_WATERMARK_SET_D:
+		wm_regs = &watermarks->dcn4x.d;
+		break;
+	case DML2_DCHUB_WATERMARK_SET_NUM:
+	default:
+		/* invalid wm set index */
+		wm_regs = NULL;
+	}
+
+	return wm_regs;
+}
+
+void dml21_extract_watermark_sets(const struct dc *in_dc, union dcn_watermark_set *watermarks, struct dml2_context *in_ctx)
+{
+	const struct dml2_display_cfg_programming *programming = in_ctx->v21.mode_programming.programming;
+
+	unsigned int wm_index;
+
+	/* copy watermark sets from DML */
+	for (wm_index = 0; wm_index < programming->global_regs.num_watermark_sets; wm_index++) {
+		struct dml2_dchub_watermark_regs *wm_regs = wm_set_index_to_dc_wm_set(watermarks, wm_index);
+
+		if (wm_regs)
+			memcpy(wm_regs,
+				&programming->global_regs.wm_regs[wm_index],
+				sizeof(struct dml2_dchub_watermark_regs));
+	}
+}
+
+void dml21_map_hw_resources(struct dml2_context *dml_ctx)
+{
+	unsigned int i = 0;
+
+	for (i = 0; i < __DML2_WRAPPER_MAX_STREAMS_PLANES__; i++) {
+		dml_ctx->v21.dml_to_dc_pipe_mapping.dml_pipe_idx_to_stream_id[i] = dml_ctx->v21.dml_to_dc_pipe_mapping.disp_cfg_to_stream_id[i];
+		dml_ctx->v21.dml_to_dc_pipe_mapping.dml_pipe_idx_to_stream_id_valid[i] = true;
+		dml_ctx->v21.dml_to_dc_pipe_mapping.dml_pipe_idx_to_plane_id[i] = dml_ctx->v21.dml_to_dc_pipe_mapping.disp_cfg_to_plane_id[i];
+		dml_ctx->v21.dml_to_dc_pipe_mapping.dml_pipe_idx_to_plane_id_valid[i] = true;
+	}
+
+}
+
+void dml21_get_pipe_mcache_config(
+	struct dc_state *context,
+	struct pipe_ctx *pipe_ctx,
+	struct dml2_per_plane_programming *pln_prog,
+	struct dml2_pipe_configuration_descriptor *mcache_pipe_config)
+{
+	mcache_pipe_config->plane0.viewport_x_start = pipe_ctx->plane_res.scl_data.viewport.x;
+	mcache_pipe_config->plane0.viewport_width = pipe_ctx->plane_res.scl_data.viewport.width;
+
+	mcache_pipe_config->plane1.viewport_x_start = pipe_ctx->plane_res.scl_data.viewport_c.x;
+	mcache_pipe_config->plane1.viewport_width = pipe_ctx->plane_res.scl_data.viewport_c.width;
+
+	mcache_pipe_config->plane1_enabled =
+			dml21_is_plane1_enabled(pln_prog->plane_descriptor->pixel_format);
+}
+
+void dml21_set_dc_p_state_type(
+		struct pipe_ctx *pipe_ctx,
+		struct dml2_per_stream_programming *stream_programming,
+		bool sub_vp_enabled)
+{
+	switch (stream_programming->uclk_pstate_method) {
+	case dml2_pstate_method_vactive:
+	case dml2_pstate_method_fw_vactive_drr:
+		pipe_ctx->p_state_type = P_STATE_V_ACTIVE;
+		break;
+	case dml2_pstate_method_vblank:
+	case dml2_pstate_method_fw_vblank_drr:
+		if (sub_vp_enabled)
+			pipe_ctx->p_state_type = P_STATE_V_BLANK_SUB_VP;
+		else
+			pipe_ctx->p_state_type = P_STATE_V_BLANK;
+		break;
+	case dml2_pstate_method_fw_svp:
+	case dml2_pstate_method_fw_svp_drr:
+		pipe_ctx->p_state_type = P_STATE_SUB_VP;
+		break;
+	case dml2_pstate_method_fw_drr:
+		if (sub_vp_enabled)
+			pipe_ctx->p_state_type = P_STATE_DRR_SUB_VP;
+		else
+			pipe_ctx->p_state_type = P_STATE_FPO;
+		break;
+	default:
+		pipe_ctx->p_state_type = P_STATE_UNKNOWN;
+		break;
+	}
+}
+
diff --git a/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/dml21_translation_helper.h b/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/dml21_translation_helper.h
new file mode 100644
index 000000000000..9880d3e0398e
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/dml21_translation_helper.h
@@ -0,0 +1,28 @@
+// SPDX-License-Identifier: MIT
+//
+// Copyright 2024 Advanced Micro Devices, Inc.
+
+
+#ifndef _DML21_TRANSLATION_HELPER_H_
+#define _DML21_TRANSLATION_HELPER_H_
+
+struct dc;
+struct dc_state;
+struct dcn_watermarks;
+union dcn_watermark_set;
+struct pipe_ctx;
+struct dc_plane_state;
+
+struct dml2_context;
+struct dml2_configuration_options;
+struct dml2_initialize_instance_in_out;
+
+void dml21_populate_dml_init_params(struct dml2_initialize_instance_in_out *dml_init, const struct dml2_configuration_options *config, const struct dc *in_dc);
+bool dml21_map_dc_state_into_dml_display_cfg(const struct dc *in_dc, struct dc_state *context, struct dml2_context *dml_ctx);
+void dml21_copy_clocks_to_dc_state(struct dml2_context *in_ctx, struct dc_state *context);
+void dml21_extract_watermark_sets(const struct dc *in_dc, union dcn_watermark_set *watermarks, struct dml2_context *in_ctx);
+void dml21_map_hw_resources(struct dml2_context *dml_ctx);
+void dml21_get_pipe_mcache_config(struct dc_state *context, struct pipe_ctx *pipe_ctx, struct dml2_per_plane_programming *pln_prog, struct dml2_pipe_configuration_descriptor *mcache_pipe_config);
+void dml21_set_dc_p_state_type(struct pipe_ctx *pipe_ctx, struct dml2_per_stream_programming *stream_programming, bool sub_vp_enabled);
+unsigned int map_plane_to_dml21_display_cfg(const struct dml2_context *dml_ctx, unsigned int stream_id, const struct dc_plane_state *plane, const struct dc_state *context);
+#endif
diff --git a/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/dml21_utils.c b/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/dml21_utils.c
new file mode 100644
index 000000000000..ee721606b883
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/dml21_utils.c
@@ -0,0 +1,516 @@
+// SPDX-License-Identifier: MIT
+//
+// Copyright 2024 Advanced Micro Devices, Inc.
+
+
+#include "dml2_internal_shared_types.h"
+#include "dml21_translation_helper.h"
+#include "dml2_internal_types.h"
+#include "dml21_utils.h"
+#include "dml2_dc_resource_mgmt.h"
+
+#include "dml2_core_dcn4_calcs.h"
+
+int dml21_helper_find_dml_pipe_idx_by_stream_id(struct dml2_context *ctx, unsigned int stream_id)
+{
+	int i;
+	for (i = 0; i < __DML2_WRAPPER_MAX_STREAMS_PLANES__; i++) {
+		if (ctx->v21.dml_to_dc_pipe_mapping.dml_pipe_idx_to_stream_id_valid[i] && ctx->v21.dml_to_dc_pipe_mapping.dml_pipe_idx_to_stream_id[i] == stream_id)
+			return  i;
+	}
+
+	return -1;
+}
+
+int dml21_find_dml_pipe_idx_by_plane_id(struct dml2_context *ctx, unsigned int plane_id)
+{
+	int i;
+	for (i = 0; i < __DML2_WRAPPER_MAX_STREAMS_PLANES__; i++) {
+		if (ctx->v21.dml_to_dc_pipe_mapping.dml_pipe_idx_to_plane_id_valid[i] && ctx->v21.dml_to_dc_pipe_mapping.dml_pipe_idx_to_plane_id[i] == plane_id)
+			return  i;
+	}
+
+	return -1;
+}
+
+bool dml21_get_plane_id(const struct dc_state *state, const struct dc_plane_state *plane, unsigned int *plane_id)
+{
+	int i, j;
+
+	if (!plane_id)
+		return false;
+
+	for (i = 0; i < state->stream_count; i++) {
+		for (j = 0; j < state->stream_status[i].plane_count; j++) {
+			if (state->stream_status[i].plane_states[j] == plane) {
+				*plane_id = (i << 16) | j;
+				return true;
+			}
+		}
+	}
+
+	return false;
+}
+
+unsigned int dml21_get_dc_plane_idx_from_plane_id(unsigned int plane_id)
+{
+	return 0xffff & plane_id;
+}
+
+void find_valid_pipe_idx_for_stream_index(const struct dml2_context *dml_ctx, unsigned int *dml_pipe_idx, unsigned int stream_index)
+{
+	unsigned int i = 0;
+
+	for (i = 0; i < __DML2_WRAPPER_MAX_STREAMS_PLANES__; i++) {
+		if (dml_ctx->v21.mode_programming.programming->plane_programming[i].plane_descriptor->stream_index == stream_index) {
+			*dml_pipe_idx = i;
+			return;
+		}
+	}
+}
+
+void find_pipe_regs_idx(const struct dml2_context *dml_ctx,
+		struct pipe_ctx *pipe, unsigned int *pipe_regs_idx)
+{
+	struct pipe_ctx *opp_head = dml_ctx->config.callbacks.get_opp_head(pipe);
+
+	*pipe_regs_idx = dml_ctx->config.callbacks.get_odm_slice_index(opp_head);
+
+	if (pipe->plane_state)
+		*pipe_regs_idx += dml_ctx->config.callbacks.get_mpc_slice_index(pipe);
+}
+
+/* places pipe references into pipes arrays and returns number of pipes */
+int dml21_find_dc_pipes_for_plane(const struct dc *in_dc,
+		struct dc_state *context,
+		struct dml2_context *dml_ctx,
+		struct pipe_ctx *dc_main_pipes[__DML2_WRAPPER_MAX_STREAMS_PLANES__],
+		struct pipe_ctx *dc_phantom_pipes[__DML2_WRAPPER_MAX_STREAMS_PLANES__],
+		int dml_plane_idx)
+{
+	unsigned int dml_stream_index;
+	unsigned int main_stream_id;
+	unsigned int dc_plane_index;
+	struct dc_stream_state *dc_main_stream;
+	struct dc_stream_status *dc_main_stream_status;
+	struct dc_plane_state *dc_main_plane;
+	struct dc_stream_state *dc_phantom_stream;
+	struct dc_stream_status *dc_phantom_stream_status;
+	struct dc_plane_state *dc_phantom_plane;
+	int num_pipes = 0;
+
+	memset(dc_main_pipes, 0, sizeof(struct pipe_ctx *) * __DML2_WRAPPER_MAX_STREAMS_PLANES__);
+	memset(dc_phantom_pipes, 0, sizeof(struct pipe_ctx *) * __DML2_WRAPPER_MAX_STREAMS_PLANES__);
+
+	dml_stream_index = dml_ctx->v21.mode_programming.programming->plane_programming[dml_plane_idx].plane_descriptor->stream_index;
+	main_stream_id = dml_ctx->v21.dml_to_dc_pipe_mapping.dml_pipe_idx_to_stream_id[dml_stream_index];
+
+	dc_main_stream = dml_ctx->config.callbacks.get_stream_from_id(context, main_stream_id);
+	dc_main_stream_status = dml_ctx->config.callbacks.get_stream_status(context, dc_main_stream);
+	if (!dc_main_stream_status)
+		return num_pipes;
+
+	/* find main plane based on id */
+	dc_plane_index = dml21_get_dc_plane_idx_from_plane_id(dml_ctx->v21.dml_to_dc_pipe_mapping.dml_pipe_idx_to_plane_id[dml_plane_idx]);
+	dc_main_plane = dc_main_stream_status->plane_states[dc_plane_index];
+
+	if (dc_main_plane) {
+		num_pipes = dml_ctx->config.callbacks.get_dpp_pipes_for_plane(dc_main_plane, &context->res_ctx, dc_main_pipes);
+	} else {
+		/* stream was configured with dummy plane, so get pipes from opp head */
+		struct pipe_ctx *otg_master_pipe = dml_ctx->config.callbacks.get_otg_master_for_stream(&context->res_ctx, dc_main_stream);
+		if (otg_master_pipe != NULL)
+			num_pipes = dml_ctx->config.callbacks.get_opp_heads_for_otg_master(otg_master_pipe, &context->res_ctx, dc_main_pipes);
+	}
+
+	/* if phantom exists, find associated pipes */
+	dc_phantom_stream = dml_ctx->config.svp_pstate.callbacks.get_paired_subvp_stream(context, dc_main_stream);
+	if (dc_phantom_stream && num_pipes > 0) {
+		dc_phantom_stream_status = dml_ctx->config.callbacks.get_stream_status(context, dc_phantom_stream);
+
+		if (dc_phantom_stream_status) {
+			/* phantom plane will have same index as main */
+			dc_phantom_plane = dc_phantom_stream_status->plane_states[dc_plane_index];
+
+			if (dc_phantom_plane) {
+				/* only care about phantom pipes if they contain the phantom plane */
+				dml_ctx->config.callbacks.get_dpp_pipes_for_plane(dc_phantom_plane, &context->res_ctx, dc_phantom_pipes);
+			}
+		}
+	}
+
+	return num_pipes;
+}
+
+void dml21_pipe_populate_global_sync(struct dml2_context *dml_ctx,
+	struct dc_state *context,
+	struct pipe_ctx *pipe_ctx,
+	struct dml2_per_stream_programming *stream_programming)
+{
+	union dml2_global_sync_programming *global_sync = &stream_programming->global_sync;
+
+	if (dml_ctx->config.svp_pstate.callbacks.get_pipe_subvp_type(context, pipe_ctx) == SUBVP_PHANTOM) {
+		/* phantom has its own global sync */
+		global_sync = &stream_programming->phantom_stream.global_sync;
+	}
+
+	memcpy(&pipe_ctx->global_sync,
+		global_sync,
+		sizeof(union dml2_global_sync_programming));
+}
+
+void dml21_populate_mall_allocation_size(struct dc_state *context,
+		struct dml2_context *in_ctx,
+		struct dml2_per_plane_programming *pln_prog,
+		struct pipe_ctx *dc_pipe)
+{
+
+	/* Reuse MALL Allocation Sizes logic from dcn32_fpu.c */
+	/* Count from active, top pipes per plane only. Only add mall_ss_size_bytes for each unique plane. */
+	if (dc_pipe->stream && dc_pipe->plane_state &&
+			(dc_pipe->top_pipe == NULL ||
+			dc_pipe->plane_state != dc_pipe->top_pipe->plane_state) &&
+			dc_pipe->prev_odm_pipe == NULL) {
+		/* SS: all active surfaces stored in MALL */
+		if (in_ctx->config.svp_pstate.callbacks.get_pipe_subvp_type(context, dc_pipe) != SUBVP_PHANTOM) {
+			dc_pipe->surface_size_in_mall_bytes = pln_prog->surface_size_mall_bytes;
+			context->bw_ctx.bw.dcn.mall_ss_size_bytes += dc_pipe->surface_size_in_mall_bytes;
+		} else {
+			/* SUBVP: phantom surfaces only stored in MALL */
+			dc_pipe->surface_size_in_mall_bytes = pln_prog->svp_size_mall_bytes;
+			context->bw_ctx.bw.dcn.mall_subvp_size_bytes += dc_pipe->surface_size_in_mall_bytes;
+		}
+	}
+}
+
+bool check_dp2p0_output_encoder(const struct pipe_ctx *pipe_ctx)
+{
+	/* If this assert is hit then we have a link encoder dynamic management issue */
+	ASSERT(pipe_ctx->stream_res.hpo_dp_stream_enc ? pipe_ctx->link_res.hpo_dp_link_enc != NULL : true);
+	return (pipe_ctx->stream_res.hpo_dp_stream_enc &&
+		pipe_ctx->link_res.hpo_dp_link_enc &&
+		dc_is_dp_signal(pipe_ctx->stream->signal));
+}
+
+
+static bool is_sub_vp_enabled(struct dc *dc, struct dc_state *context)
+{
+	int i;
+
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i];
+
+		if (pipe_ctx->stream && dc_state_get_paired_subvp_stream(context, pipe_ctx->stream) &&
+							dc_state_get_pipe_subvp_type(context, pipe_ctx) == SUBVP_MAIN) {
+			return true;
+		}
+	}
+	return false;
+}
+
+
+void dml21_program_dc_pipe(struct dml2_context *dml_ctx, struct dc_state *context, struct pipe_ctx *pipe_ctx, struct dml2_per_plane_programming *pln_prog,
+		struct dml2_per_stream_programming *stream_prog)
+{
+	unsigned int pipe_reg_index = 0;
+
+	dml21_pipe_populate_global_sync(dml_ctx, context, pipe_ctx, stream_prog);
+	find_pipe_regs_idx(dml_ctx, pipe_ctx, &pipe_reg_index);
+
+	if (dml_ctx->config.svp_pstate.callbacks.get_pipe_subvp_type(context, pipe_ctx) == SUBVP_PHANTOM) {
+		memcpy(&pipe_ctx->hubp_regs, pln_prog->phantom_plane.pipe_regs[pipe_reg_index], sizeof(struct dml2_dchub_per_pipe_register_set));
+		pipe_ctx->unbounded_req = false;
+		pipe_ctx->det_buffer_size_kb = 0;
+	} else {
+		memcpy(&pipe_ctx->hubp_regs, pln_prog->pipe_regs[pipe_reg_index], sizeof(struct dml2_dchub_per_pipe_register_set));
+		pipe_ctx->unbounded_req = pln_prog->pipe_regs[pipe_reg_index]->rq_regs.unbounded_request_enabled;
+		pipe_ctx->det_buffer_size_kb = pln_prog->pipe_regs[pipe_reg_index]->det_size * 64;
+	}
+
+	pipe_ctx->plane_res.bw.dppclk_khz = pln_prog->min_clocks.dcn4x.dppclk_khz;
+	if (context->bw_ctx.bw.dcn.clk.dppclk_khz < pipe_ctx->plane_res.bw.dppclk_khz)
+		context->bw_ctx.bw.dcn.clk.dppclk_khz = pipe_ctx->plane_res.bw.dppclk_khz;
+
+	dml21_populate_mall_allocation_size(context, dml_ctx, pln_prog, pipe_ctx);
+
+	bool sub_vp_enabled = is_sub_vp_enabled(pipe_ctx->stream->ctx->dc, context);
+
+	dml21_set_dc_p_state_type(pipe_ctx, stream_prog, sub_vp_enabled);
+}
+
+static struct dc_stream_state *dml21_add_phantom_stream(struct dml2_context *dml_ctx,
+	const struct dc *dc,
+	struct dc_state *context,
+	struct dc_stream_state *main_stream,
+	struct dml2_per_stream_programming *stream_programming)
+{
+	struct dc_stream_state *phantom_stream;
+	struct dml2_stream_parameters *phantom_stream_descriptor = &stream_programming->phantom_stream.descriptor;
+
+	phantom_stream = dml_ctx->config.svp_pstate.callbacks.create_phantom_stream(dc, context, main_stream);
+	if (!phantom_stream)
+		return NULL;
+
+	/* copy details of phantom stream from main */
+	memcpy(&phantom_stream->timing, &main_stream->timing, sizeof(phantom_stream->timing));
+	memcpy(&phantom_stream->src, &main_stream->src, sizeof(phantom_stream->src));
+	memcpy(&phantom_stream->dst, &main_stream->dst, sizeof(phantom_stream->dst));
+
+	/* modify timing for phantom */
+	phantom_stream->timing.v_front_porch = phantom_stream_descriptor->timing.v_front_porch;
+	phantom_stream->timing.v_addressable = phantom_stream_descriptor->timing.v_active;
+	phantom_stream->timing.v_total = phantom_stream_descriptor->timing.v_total;
+	phantom_stream->timing.flags.DSC = 0; // phantom always has DSC disabled
+
+	phantom_stream->dst.y = 0;
+	phantom_stream->dst.height = stream_programming->phantom_stream.descriptor.timing.v_active;
+
+	phantom_stream->src.y = 0;
+	phantom_stream->src.height = (double)phantom_stream_descriptor->timing.v_active * (double)main_stream->src.height / (double)main_stream->dst.height;
+
+	phantom_stream->use_dynamic_meta = false;
+
+	dml_ctx->config.svp_pstate.callbacks.add_phantom_stream(dc, context, phantom_stream, main_stream);
+
+	return phantom_stream;
+}
+
+static struct dc_plane_state *dml21_add_phantom_plane(struct dml2_context *dml_ctx,
+	const struct dc *dc,
+	struct dc_state *context,
+	struct dc_stream_state *phantom_stream,
+	struct dc_plane_state *main_plane,
+	struct dml2_per_plane_programming *plane_programming)
+{
+	struct dc_plane_state *phantom_plane;
+
+	phantom_plane = dml_ctx->config.svp_pstate.callbacks.create_phantom_plane(dc, context, main_plane);
+	if (!phantom_plane)
+		return NULL;
+
+	phantom_plane->format = main_plane->format;
+	phantom_plane->rotation = main_plane->rotation;
+	phantom_plane->visible = main_plane->visible;
+
+	memcpy(&phantom_plane->address, &main_plane->address, sizeof(phantom_plane->address));
+	memcpy(&phantom_plane->scaling_quality, &main_plane->scaling_quality,
+		sizeof(phantom_plane->scaling_quality));
+	memcpy(&phantom_plane->src_rect, &main_plane->src_rect, sizeof(phantom_plane->src_rect));
+	memcpy(&phantom_plane->dst_rect, &main_plane->dst_rect, sizeof(phantom_plane->dst_rect));
+	memcpy(&phantom_plane->clip_rect, &main_plane->clip_rect, sizeof(phantom_plane->clip_rect));
+	memcpy(&phantom_plane->plane_size, &main_plane->plane_size,
+		sizeof(phantom_plane->plane_size));
+	memcpy(&phantom_plane->tiling_info, &main_plane->tiling_info,
+		sizeof(phantom_plane->tiling_info));
+	memcpy(&phantom_plane->dcc, &main_plane->dcc, sizeof(phantom_plane->dcc));
+
+	phantom_plane->format = main_plane->format;
+	phantom_plane->rotation = main_plane->rotation;
+	phantom_plane->visible = main_plane->visible;
+
+	/* Shadow pipe has small viewport. */
+	phantom_plane->clip_rect.y = 0;
+	phantom_plane->clip_rect.height = phantom_stream->src.height;
+
+	dml_ctx->config.svp_pstate.callbacks.add_phantom_plane(dc, phantom_stream, phantom_plane, context);
+
+	return phantom_plane;
+}
+
+void dml21_handle_phantom_streams_planes(const struct dc *dc, struct dc_state *context, struct dml2_context *dml_ctx)
+{
+	unsigned int dml_stream_index, dml_plane_index, dc_plane_index;
+	struct dc_stream_state *main_stream;
+	struct dc_stream_status *main_stream_status;
+	struct dc_stream_state *phantom_stream;
+	struct dc_plane_state *main_plane;
+	bool phantoms_added = false;
+
+	/* create phantom streams and planes and add to context */
+	for (dml_stream_index = 0; dml_stream_index < dml_ctx->v21.mode_programming.programming->display_config.num_streams; dml_stream_index++) {
+		/* iterate through DML streams looking for phantoms */
+		if (dml_ctx->v21.mode_programming.programming->stream_programming[dml_stream_index].phantom_stream.enabled) {
+			/* find associated dc stream */
+			main_stream = dml_ctx->config.callbacks.get_stream_from_id(context,
+					dml_ctx->v21.dml_to_dc_pipe_mapping.dml_pipe_idx_to_stream_id[dml_stream_index]);
+
+			main_stream_status = dml_ctx->config.callbacks.get_stream_status(context, main_stream);
+
+			if (!main_stream_status || main_stream_status->plane_count == 0)
+				continue;
+
+			/* create phantom stream for subvp enabled stream */
+			phantom_stream = dml21_add_phantom_stream(dml_ctx,
+					dc,
+					context,
+					main_stream,
+					&dml_ctx->v21.mode_programming.programming->stream_programming[dml_stream_index]);
+
+			if (!phantom_stream)
+				continue;
+
+			/* iterate through DML planes associated with this stream */
+			for (dml_plane_index = 0; dml_plane_index < dml_ctx->v21.mode_programming.programming->display_config.num_planes; dml_plane_index++) {
+				if (dml_ctx->v21.mode_programming.programming->plane_programming[dml_plane_index].plane_descriptor->stream_index == dml_stream_index) {
+					/* find associated dc plane */
+					dc_plane_index = dml21_get_dc_plane_idx_from_plane_id(dml_ctx->v21.dml_to_dc_pipe_mapping.dml_pipe_idx_to_plane_id[dml_plane_index]);
+					main_plane = main_stream_status->plane_states[dc_plane_index];
+
+					/* create phantom planes for subvp enabled plane */
+					dml21_add_phantom_plane(dml_ctx,
+							dc,
+							context,
+							phantom_stream,
+							main_plane,
+							&dml_ctx->v21.mode_programming.programming->plane_programming[dml_plane_index]);
+
+					phantoms_added = true;
+				}
+			}
+		}
+	}
+
+	if (phantoms_added)
+		dml2_map_dc_pipes(dml_ctx, context, NULL, &dml_ctx->v21.dml_to_dc_pipe_mapping, dc->current_state);
+}
+
+void dml21_build_fams2_programming(const struct dc *dc,
+		struct dc_state *context,
+		struct dml2_context *dml_ctx)
+{
+	int i, j, k;
+	unsigned int num_fams2_streams = 0;
+
+	/* reset fams2 data */
+	memset(&context->bw_ctx.bw.dcn.fams2_stream_base_params, 0, sizeof(union dmub_cmd_fams2_config) * DML2_MAX_PLANES);
+	memset(&context->bw_ctx.bw.dcn.fams2_stream_sub_params, 0, sizeof(union dmub_cmd_fams2_config) * DML2_MAX_PLANES);
+	memset(&context->bw_ctx.bw.dcn.fams2_stream_sub_params_v2, 0, sizeof(union dmub_fams2_stream_static_sub_state_v2) * DML2_MAX_PLANES);
+	memset(&context->bw_ctx.bw.dcn.fams2_global_config, 0, sizeof(struct dmub_cmd_fams2_global_config));
+
+	if (dml_ctx->v21.mode_programming.programming->fams2_required) {
+		for (i = 0; i < context->stream_count; i++) {
+			int dml_stream_idx;
+			struct dc_stream_state *phantom_stream;
+			struct dc_stream_status *phantom_status;
+			enum fams2_stream_type type = 0;
+
+			union dmub_cmd_fams2_config *static_base_state = &context->bw_ctx.bw.dcn.fams2_stream_base_params[num_fams2_streams];
+			union dmub_cmd_fams2_config *static_sub_state = &context->bw_ctx.bw.dcn.fams2_stream_sub_params[num_fams2_streams];
+
+			struct dc_stream_state *stream = context->streams[i];
+
+			if (context->stream_status[i].plane_count == 0 ||
+					dml_ctx->config.svp_pstate.callbacks.get_stream_subvp_type(context, stream) == SUBVP_PHANTOM) {
+				/* can ignore blanked or phantom streams */
+				continue;
+			}
+
+			dml_stream_idx = dml21_helper_find_dml_pipe_idx_by_stream_id(dml_ctx, stream->stream_id);
+			if (dml_stream_idx < 0) {
+				ASSERT(dml_stream_idx >= 0);
+				continue;
+			}
+
+			/* copy static state from PMO */
+			memcpy(static_base_state,
+					&dml_ctx->v21.mode_programming.programming->stream_programming[dml_stream_idx].fams2_base_params,
+					sizeof(union dmub_cmd_fams2_config));
+
+			if (dc->debug.fams_version.major == 3) {
+				memcpy(&context->bw_ctx.bw.dcn.fams2_stream_sub_params_v2[num_fams2_streams],
+						&dml_ctx->v21.mode_programming.programming->stream_programming[dml_stream_idx].fams2_sub_params_v2,
+						sizeof(union dmub_fams2_stream_static_sub_state_v2));
+			} else {
+				memcpy(static_sub_state,
+						&dml_ctx->v21.mode_programming.programming->stream_programming[dml_stream_idx].fams2_sub_params,
+						sizeof(union dmub_cmd_fams2_config));
+			}
+
+			switch (dc->debug.fams_version.minor) {
+			case 1:
+			default:
+				type = static_base_state->stream_v1.base.type;
+
+				/* get information from context */
+				static_base_state->stream_v1.base.num_planes = context->stream_status[i].plane_count;
+				static_base_state->stream_v1.base.otg_inst = context->stream_status[i].primary_otg_inst;
+
+				/* populate pipe masks for planes */
+				for (j = 0; j < context->stream_status[i].plane_count; j++) {
+					for (k = 0; k < dc->res_pool->pipe_count; k++) {
+						if (context->res_ctx.pipe_ctx[k].stream &&
+								context->res_ctx.pipe_ctx[k].stream->stream_id == stream->stream_id &&
+								context->res_ctx.pipe_ctx[k].plane_state == context->stream_status[i].plane_states[j]) {
+							static_base_state->stream_v1.base.pipe_mask |= (1 << k);
+							static_base_state->stream_v1.base.plane_pipe_masks[j] |= (1 << k);
+						}
+					}
+				}
+			}
+
+
+			/* get per method programming */
+			switch (type) {
+			case FAMS2_STREAM_TYPE_VBLANK:
+			case FAMS2_STREAM_TYPE_VACTIVE:
+			case FAMS2_STREAM_TYPE_DRR:
+				break;
+			case FAMS2_STREAM_TYPE_SUBVP:
+				phantom_stream = dml_ctx->config.svp_pstate.callbacks.get_paired_subvp_stream(context, stream);
+				if (!phantom_stream)
+					break;
+
+				phantom_status = dml_ctx->config.callbacks.get_stream_status(context, phantom_stream);
+
+				/* phantom status should always be present */
+				ASSERT(phantom_status);
+				if (!phantom_status)
+					break;
+
+				switch (dc->debug.fams_version.minor) {
+				case 1:
+				default:
+					static_sub_state->stream_v1.sub_state.subvp.phantom_otg_inst = phantom_status->primary_otg_inst;
+
+					/* populate pipe masks for phantom planes */
+					for (j = 0; j < phantom_status->plane_count; j++) {
+						for (k = 0; k < dc->res_pool->pipe_count; k++) {
+							if (context->res_ctx.pipe_ctx[k].stream &&
+									context->res_ctx.pipe_ctx[k].stream->stream_id == phantom_stream->stream_id &&
+									context->res_ctx.pipe_ctx[k].plane_state == phantom_status->plane_states[j]) {
+								switch (dc->debug.fams_version.minor) {
+								case 1:
+								default:
+									static_sub_state->stream_v1.sub_state.subvp.phantom_pipe_mask |= (1 << k);
+									static_sub_state->stream_v1.sub_state.subvp.phantom_plane_pipe_masks[j] |= (1 << k);
+								}
+							}
+						}
+					}
+				}
+				break;
+			default:
+				ASSERT(false);
+				break;
+			}
+
+			num_fams2_streams++;
+		}
+	}
+
+	if (num_fams2_streams > 0) {
+		/* copy FAMS2 configuration */
+		memcpy(&context->bw_ctx.bw.dcn.fams2_global_config,
+				&dml_ctx->v21.mode_programming.programming->fams2_global_config,
+				sizeof(struct dmub_cmd_fams2_global_config));
+
+		context->bw_ctx.bw.dcn.fams2_global_config.num_streams = num_fams2_streams;
+	}
+
+	context->bw_ctx.bw.dcn.clk.fw_based_mclk_switching = context->bw_ctx.bw.dcn.fams2_global_config.features.bits.enable;
+}
+
+bool dml21_is_plane1_enabled(enum dml2_source_format_class source_format)
+{
+	return source_format >= dml2_420_8 && source_format <= dml2_rgbe_alpha;
+}
diff --git a/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/dml21_utils.h b/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/dml21_utils.h
new file mode 100644
index 000000000000..4bff52eaaef8
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/dml21_utils.h
@@ -0,0 +1,50 @@
+// SPDX-License-Identifier: MIT
+//
+// Copyright 2024 Advanced Micro Devices, Inc.
+
+
+#ifndef _DML21_UTILS_H_
+#define _DML21_UTILS_H_
+
+struct dc_state;
+struct dc_plane_state;
+struct pipe_ctx;
+
+struct dml2_context;
+struct dml2_display_rq_regs;
+struct dml2_display_dlg_regs;
+struct dml2_display_ttu_regs;
+
+int dml21_helper_find_dml_pipe_idx_by_stream_id(struct dml2_context *ctx, unsigned int stream_id);
+int dml21_find_dml_pipe_idx_by_plane_id(struct dml2_context *ctx, unsigned int plane_id);
+bool dml21_get_plane_id(const struct dc_state *state, const struct dc_plane_state *plane, unsigned int *plane_id);
+void dml21_pipe_populate_global_sync(struct dml2_context *dml_ctx,
+		struct dc_state *context,
+		struct pipe_ctx *pipe_ctx,
+		struct dml2_per_stream_programming *stream_programming);
+void dml21_populate_mall_allocation_size(struct dc_state *context,
+		struct dml2_context *in_ctx,
+		struct dml2_per_plane_programming *pln_prog,
+		struct pipe_ctx *dc_pipe);
+bool check_dp2p0_output_encoder(const struct pipe_ctx *pipe_ctx);
+void find_valid_pipe_idx_for_stream_index(const struct dml2_context *dml_ctx, unsigned int *dml_pipe_idx, unsigned int stream_index);
+void find_pipe_regs_idx(const struct dml2_context *dml_ctx,
+		struct pipe_ctx *pipe, unsigned int *pipe_regs_idx);
+int dml21_find_dc_pipes_for_plane(const struct dc *in_dc,
+		struct dc_state *context,
+		struct dml2_context *dml_ctx,
+		struct pipe_ctx *dc_main_pipes[__DML2_WRAPPER_MAX_STREAMS_PLANES__],
+		struct pipe_ctx *dc_phantom_pipes[__DML2_WRAPPER_MAX_STREAMS_PLANES__],
+		int dml_plane_idx);
+void dml21_program_dc_pipe(struct dml2_context *dml_ctx,
+		struct dc_state *context,
+		struct pipe_ctx *pipe_ctx,
+		struct dml2_per_plane_programming *pln_prog,
+		struct dml2_per_stream_programming *stream_prog);
+void dml21_handle_phantom_streams_planes(const struct dc *in_dc, struct dc_state *context, struct dml2_context *dml_ctx);
+unsigned int dml21_get_dc_plane_idx_from_plane_id(unsigned int plane_id);
+void dml21_build_fams2_programming(const struct dc *dc,
+		struct dc_state *context,
+		struct dml2_context *dml_ctx);
+bool dml21_is_plane1_enabled(enum dml2_source_format_class source_format);
+#endif
diff --git a/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/dml21_wrapper.c b/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/dml21_wrapper.c
new file mode 100644
index 000000000000..798abb2b2e67
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/dml21_wrapper.c
@@ -0,0 +1,466 @@
+// SPDX-License-Identifier: MIT
+//
+// Copyright 2024 Advanced Micro Devices, Inc.
+
+#include "dml2_internal_types.h"
+#include "dml_top.h"
+#include "dml2_core_dcn4_calcs.h"
+#include "dml2_internal_shared_types.h"
+#include "dml21_utils.h"
+#include "dml21_translation_helper.h"
+#include "dml2_dc_resource_mgmt.h"
+
+#define INVALID -1
+
+static bool dml21_allocate_memory(struct dml2_context **dml_ctx)
+{
+	*dml_ctx = vzalloc(sizeof(struct dml2_context));
+	if (!(*dml_ctx))
+		return false;
+
+	(*dml_ctx)->v21.dml_init.dml2_instance = vzalloc(sizeof(struct dml2_instance));
+	if (!((*dml_ctx)->v21.dml_init.dml2_instance))
+		return false;
+
+	(*dml_ctx)->v21.mode_support.dml2_instance = (*dml_ctx)->v21.dml_init.dml2_instance;
+	(*dml_ctx)->v21.mode_programming.dml2_instance = (*dml_ctx)->v21.dml_init.dml2_instance;
+
+	(*dml_ctx)->v21.mode_support.display_config = &(*dml_ctx)->v21.display_config;
+	(*dml_ctx)->v21.mode_programming.display_config = (*dml_ctx)->v21.mode_support.display_config;
+
+	(*dml_ctx)->v21.mode_programming.programming = vzalloc(sizeof(struct dml2_display_cfg_programming));
+	if (!((*dml_ctx)->v21.mode_programming.programming))
+		return false;
+
+	return true;
+}
+
+static void dml21_populate_configuration_options(const struct dc *in_dc,
+		struct dml2_context *dml_ctx,
+		const struct dml2_configuration_options *config)
+{
+	dml_ctx->config = *config;
+
+	/* UCLK P-State options */
+	if (in_dc->debug.dml21_force_pstate_method) {
+		dml_ctx->config.pmo.force_pstate_method_enable = true;
+		for (int i = 0; i < MAX_PIPES; i++)
+			dml_ctx->config.pmo.force_pstate_method_values[i] = in_dc->debug.dml21_force_pstate_method_values[i];
+	} else {
+		dml_ctx->config.pmo.force_pstate_method_enable = false;
+	}
+}
+
+static void dml21_init(const struct dc *in_dc, struct dml2_context *dml_ctx, const struct dml2_configuration_options *config)
+{
+
+	dml_ctx->architecture = dml2_architecture_21;
+
+	dml21_populate_configuration_options(in_dc, dml_ctx, config);
+
+	DC_FP_START();
+
+	dml21_populate_dml_init_params(&dml_ctx->v21.dml_init, &dml_ctx->config, in_dc);
+
+	dml2_initialize_instance(&dml_ctx->v21.dml_init);
+
+	DC_FP_END();
+}
+
+bool dml21_create(const struct dc *in_dc, struct dml2_context **dml_ctx, const struct dml2_configuration_options *config)
+{
+	/* Allocate memory for initializing DML21 instance */
+	if (!dml21_allocate_memory(dml_ctx))
+		return false;
+
+	dml21_init(in_dc, *dml_ctx, config);
+
+	return true;
+}
+
+void dml21_destroy(struct dml2_context *dml2)
+{
+	vfree(dml2->v21.dml_init.dml2_instance);
+	vfree(dml2->v21.mode_programming.programming);
+}
+
+static void dml21_calculate_rq_and_dlg_params(const struct dc *dc, struct dc_state *context, struct resource_context *out_new_hw_state,
+	struct dml2_context *in_ctx, unsigned int pipe_cnt)
+{
+	unsigned int dml_prog_idx = 0, dc_pipe_index = 0, num_dpps_required = 0;
+	struct dml2_per_plane_programming *pln_prog = NULL;
+	struct dml2_per_stream_programming *stream_prog = NULL;
+	struct pipe_ctx *dc_main_pipes[__DML2_WRAPPER_MAX_STREAMS_PLANES__];
+	struct pipe_ctx *dc_phantom_pipes[__DML2_WRAPPER_MAX_STREAMS_PLANES__] = {0};
+	int num_pipes;
+	unsigned int dml_phantom_prog_idx;
+
+	context->bw_ctx.bw.dcn.clk.dppclk_khz = 0;
+
+	/* copy global DCHUBBUB arbiter registers */
+	memcpy(&context->bw_ctx.bw.dcn.arb_regs, &in_ctx->v21.mode_programming.programming->global_regs.arb_regs, sizeof(struct dml2_display_arb_regs));
+
+	/* legacy only */
+	context->bw_ctx.bw.dcn.compbuf_size_kb = (int)in_ctx->v21.mode_programming.programming->global_regs.arb_regs.compbuf_size * 64;
+
+	context->bw_ctx.bw.dcn.mall_ss_size_bytes = 0;
+	context->bw_ctx.bw.dcn.mall_ss_psr_active_size_bytes = 0;
+	context->bw_ctx.bw.dcn.mall_subvp_size_bytes = 0;
+
+	/* phantom's start after main planes */
+	dml_phantom_prog_idx = in_ctx->v21.mode_programming.programming->display_config.num_planes;
+
+	for (dml_prog_idx = 0; dml_prog_idx < DML2_MAX_PLANES; dml_prog_idx++) {
+		pln_prog = &in_ctx->v21.mode_programming.programming->plane_programming[dml_prog_idx];
+
+		if (!pln_prog->plane_descriptor)
+			continue;
+
+		stream_prog = &in_ctx->v21.mode_programming.programming->stream_programming[pln_prog->plane_descriptor->stream_index];
+		num_dpps_required = pln_prog->num_dpps_required;
+
+		if (num_dpps_required == 0) {
+			continue;
+		}
+		num_pipes = dml21_find_dc_pipes_for_plane(dc, context, in_ctx, dc_main_pipes, dc_phantom_pipes, dml_prog_idx);
+
+		if (num_pipes <= 0)
+			continue;
+
+		/* program each pipe */
+		for (dc_pipe_index = 0; dc_pipe_index < num_pipes; dc_pipe_index++) {
+			dml21_program_dc_pipe(in_ctx, context, dc_main_pipes[dc_pipe_index], pln_prog, stream_prog);
+
+			if (pln_prog->phantom_plane.valid && dc_phantom_pipes[dc_pipe_index]) {
+				dml21_program_dc_pipe(in_ctx, context, dc_phantom_pipes[dc_pipe_index], pln_prog, stream_prog);
+			}
+		}
+
+		/* copy per plane mcache allocation */
+		memcpy(&context->bw_ctx.bw.dcn.mcache_allocations[dml_prog_idx], &pln_prog->mcache_allocation, sizeof(struct dml2_mcache_surface_allocation));
+		if (pln_prog->phantom_plane.valid) {
+			memcpy(&context->bw_ctx.bw.dcn.mcache_allocations[dml_phantom_prog_idx],
+					&pln_prog->phantom_plane.mcache_allocation,
+					sizeof(struct dml2_mcache_surface_allocation));
+
+			dml_phantom_prog_idx++;
+		}
+	}
+
+	/* assign global clocks */
+	context->bw_ctx.bw.dcn.clk.bw_dppclk_khz = context->bw_ctx.bw.dcn.clk.dppclk_khz;
+	context->bw_ctx.bw.dcn.clk.bw_dispclk_khz = context->bw_ctx.bw.dcn.clk.dispclk_khz;
+	if (in_ctx->v21.dml_init.soc_bb.clk_table.dispclk.num_clk_values > 1) {
+		context->bw_ctx.bw.dcn.clk.max_supported_dispclk_khz =
+			in_ctx->v21.dml_init.soc_bb.clk_table.dispclk.clk_values_khz[in_ctx->v21.dml_init.soc_bb.clk_table.dispclk.num_clk_values] * 1000;
+	} else {
+		context->bw_ctx.bw.dcn.clk.max_supported_dispclk_khz = in_ctx->v21.dml_init.soc_bb.clk_table.dispclk.clk_values_khz[0] * 1000;
+	}
+
+	if (in_ctx->v21.dml_init.soc_bb.clk_table.dppclk.num_clk_values > 1) {
+		context->bw_ctx.bw.dcn.clk.max_supported_dppclk_khz =
+			in_ctx->v21.dml_init.soc_bb.clk_table.dppclk.clk_values_khz[in_ctx->v21.dml_init.soc_bb.clk_table.dppclk.num_clk_values] * 1000;
+	} else {
+		context->bw_ctx.bw.dcn.clk.max_supported_dppclk_khz = in_ctx->v21.dml_init.soc_bb.clk_table.dppclk.clk_values_khz[0] * 1000;
+	}
+
+	/* get global mall allocation */
+	if (dc->res_pool->funcs->calculate_mall_ways_from_bytes) {
+		context->bw_ctx.bw.dcn.clk.num_ways = dc->res_pool->funcs->calculate_mall_ways_from_bytes(dc, context->bw_ctx.bw.dcn.mall_subvp_size_bytes);
+	} else {
+		context->bw_ctx.bw.dcn.clk.num_ways = 0;
+	}
+}
+
+static void dml21_prepare_mcache_params(struct dml2_context *dml_ctx, struct dc_state *context, struct dc_mcache_params *mcache_params)
+{
+	int dc_plane_idx = 0;
+	int dml_prog_idx, stream_idx, plane_idx;
+	struct dml2_per_plane_programming *pln_prog = NULL;
+
+	for (stream_idx = 0; stream_idx < context->stream_count; stream_idx++) {
+		for (plane_idx = 0; plane_idx < context->stream_status[stream_idx].plane_count; plane_idx++) {
+			dml_prog_idx = map_plane_to_dml21_display_cfg(dml_ctx, context->streams[stream_idx]->stream_id, context->stream_status[stream_idx].plane_states[plane_idx], context);
+			if (dml_prog_idx == INVALID) {
+				continue;
+			}
+			pln_prog = &dml_ctx->v21.mode_programming.programming->plane_programming[dml_prog_idx];
+			mcache_params[dc_plane_idx].valid = pln_prog->mcache_allocation.valid;
+			mcache_params[dc_plane_idx].num_mcaches_plane0 = pln_prog->mcache_allocation.num_mcaches_plane0;
+			mcache_params[dc_plane_idx].num_mcaches_plane1 = pln_prog->mcache_allocation.num_mcaches_plane1;
+			mcache_params[dc_plane_idx].requires_dedicated_mall_mcache = pln_prog->mcache_allocation.requires_dedicated_mall_mcache;
+			mcache_params[dc_plane_idx].last_slice_sharing.plane0_plane1 = pln_prog->mcache_allocation.last_slice_sharing.plane0_plane1;
+			memcpy(mcache_params[dc_plane_idx].mcache_x_offsets_plane0,
+				pln_prog->mcache_allocation.mcache_x_offsets_plane0,
+				sizeof(int) * (DML2_MAX_MCACHES + 1));
+			memcpy(mcache_params[dc_plane_idx].mcache_x_offsets_plane1,
+				pln_prog->mcache_allocation.mcache_x_offsets_plane1,
+				sizeof(int) * (DML2_MAX_MCACHES + 1));
+			dc_plane_idx++;
+		}
+	}
+}
+
+static bool dml21_mode_check_and_programming(const struct dc *in_dc, struct dc_state *context, struct dml2_context *dml_ctx)
+{
+	bool result = false;
+	struct dml2_build_mode_programming_in_out *mode_programming = &dml_ctx->v21.mode_programming;
+	struct dc_mcache_params mcache_params[MAX_PLANES] = {0};
+
+	memset(&dml_ctx->v21.display_config, 0, sizeof(struct dml2_display_cfg));
+	memset(&dml_ctx->v21.dml_to_dc_pipe_mapping, 0, sizeof(struct dml2_dml_to_dc_pipe_mapping));
+	memset(&dml_ctx->v21.mode_programming.dml2_instance->scratch.build_mode_programming_locals.mode_programming_params, 0, sizeof(struct dml2_core_mode_programming_in_out));
+
+	if (!context)
+		return true;
+
+	if (context->stream_count == 0) {
+		dml21_build_fams2_programming(in_dc, context, dml_ctx);
+		return true;
+	}
+
+	/* scrub phantom's from current dc_state */
+	dml_ctx->config.svp_pstate.callbacks.remove_phantom_streams_and_planes(in_dc, context);
+	dml_ctx->config.svp_pstate.callbacks.release_phantom_streams_and_planes(in_dc, context);
+
+	/* Populate stream, plane mappings and other fields in display config. */
+	result = dml21_map_dc_state_into_dml_display_cfg(in_dc, context, dml_ctx);
+	if (!result)
+		return false;
+
+	DC_FP_START();
+	result = dml2_build_mode_programming(mode_programming);
+	DC_FP_END();
+	if (!result)
+		return false;
+
+	/* Check and map HW resources */
+	if (result && !dml_ctx->config.skip_hw_state_mapping) {
+		dml21_map_hw_resources(dml_ctx);
+		dml2_map_dc_pipes(dml_ctx, context, NULL, &dml_ctx->v21.dml_to_dc_pipe_mapping, in_dc->current_state);
+		/* if subvp phantoms are present, expand them into dc context */
+		dml21_handle_phantom_streams_planes(in_dc, context, dml_ctx);
+
+		if (in_dc->res_pool->funcs->program_mcache_pipe_config) {
+			//Prepare mcache params for each plane based on mcache output from DML
+			dml21_prepare_mcache_params(dml_ctx, context, mcache_params);
+
+			//populate mcache regs to each pipe
+			dml_ctx->config.callbacks.allocate_mcache(context, mcache_params);
+		}
+	}
+
+	/* Copy DML CLK, WM and REG outputs to bandwidth context */
+	if (result && !dml_ctx->config.skip_hw_state_mapping) {
+		dml21_calculate_rq_and_dlg_params(in_dc, context, &context->res_ctx, dml_ctx, in_dc->res_pool->pipe_count);
+		dml21_copy_clocks_to_dc_state(dml_ctx, context);
+		dml21_extract_watermark_sets(in_dc, &context->bw_ctx.bw.dcn.watermarks, dml_ctx);
+		dml21_build_fams2_programming(in_dc, context, dml_ctx);
+	}
+
+	return true;
+}
+
+static bool dml21_check_mode_support(const struct dc *in_dc, struct dc_state *context, struct dml2_context *dml_ctx)
+{
+	bool is_supported = false;
+	struct dml2_initialize_instance_in_out *dml_init = &dml_ctx->v21.dml_init;
+	struct dml2_check_mode_supported_in_out *mode_support = &dml_ctx->v21.mode_support;
+
+	memset(&dml_ctx->v21.display_config, 0, sizeof(struct dml2_display_cfg));
+	memset(&dml_ctx->v21.dml_to_dc_pipe_mapping, 0, sizeof(struct dml2_dml_to_dc_pipe_mapping));
+	memset(&dml_ctx->v21.mode_programming.dml2_instance->scratch.check_mode_supported_locals.mode_support_params, 0, sizeof(struct dml2_core_mode_support_in_out));
+
+	if (!context || context->stream_count == 0)
+		return true;
+
+	/* Scrub phantom's from current dc_state */
+	dml_ctx->config.svp_pstate.callbacks.remove_phantom_streams_and_planes(in_dc, context);
+	dml_ctx->config.svp_pstate.callbacks.release_phantom_streams_and_planes(in_dc, context);
+
+	mode_support->dml2_instance = dml_init->dml2_instance;
+	dml21_map_dc_state_into_dml_display_cfg(in_dc, context, dml_ctx);
+	dml_ctx->v21.mode_programming.dml2_instance->scratch.build_mode_programming_locals.mode_programming_params.programming = dml_ctx->v21.mode_programming.programming;
+	DC_FP_START();
+	is_supported = dml2_check_mode_supported(mode_support);
+	DC_FP_END();
+	if (!is_supported)
+		return false;
+
+	return true;
+}
+
+bool dml21_validate(const struct dc *in_dc, struct dc_state *context, struct dml2_context *dml_ctx,
+	enum dc_validate_mode validate_mode)
+{
+	bool out = false;
+
+	/* Use dml21_check_mode_support for DC_VALIDATE_MODE_ONLY and DC_VALIDATE_MODE_AND_STATE_INDEX path */
+	if (validate_mode != DC_VALIDATE_MODE_AND_PROGRAMMING)
+		out = dml21_check_mode_support(in_dc, context, dml_ctx);
+	else
+		out = dml21_mode_check_and_programming(in_dc, context, dml_ctx);
+
+	return out;
+}
+
+void dml21_prepare_mcache_programming(struct dc *in_dc, struct dc_state *context, struct dml2_context *dml_ctx)
+{
+	unsigned int dml_prog_idx, dml_phantom_prog_idx, dc_pipe_index;
+	int num_pipes;
+	struct pipe_ctx *dc_main_pipes[__DML2_WRAPPER_MAX_STREAMS_PLANES__];
+	struct pipe_ctx *dc_phantom_pipes[__DML2_WRAPPER_MAX_STREAMS_PLANES__] = {0};
+
+	struct dml2_per_plane_programming *pln_prog = NULL;
+	struct dml2_plane_mcache_configuration_descriptor *mcache_config = NULL;
+	struct prepare_mcache_programming_locals *l = &dml_ctx->v21.scratch.prepare_mcache_locals;
+
+	if (context->stream_count == 0) {
+		return;
+	}
+
+	memset(&l->build_mcache_programming_params, 0, sizeof(struct dml2_build_mcache_programming_in_out));
+	l->build_mcache_programming_params.dml2_instance = dml_ctx->v21.dml_init.dml2_instance;
+
+	/* phantom's start after main planes */
+	dml_phantom_prog_idx = dml_ctx->v21.mode_programming.programming->display_config.num_planes;
+
+	/* Build mcache programming parameters per plane per pipe */
+	for (dml_prog_idx = 0; dml_prog_idx < dml_ctx->v21.mode_programming.programming->display_config.num_planes; dml_prog_idx++) {
+		pln_prog = &dml_ctx->v21.mode_programming.programming->plane_programming[dml_prog_idx];
+
+		mcache_config = &l->build_mcache_programming_params.mcache_configurations[dml_prog_idx];
+		memset(mcache_config, 0, sizeof(struct dml2_plane_mcache_configuration_descriptor));
+		mcache_config->plane_descriptor = pln_prog->plane_descriptor;
+		mcache_config->mcache_allocation = &context->bw_ctx.bw.dcn.mcache_allocations[dml_prog_idx];
+		mcache_config->num_pipes = pln_prog->num_dpps_required;
+		l->build_mcache_programming_params.num_configurations++;
+
+		if (pln_prog->num_dpps_required == 0) {
+			continue;
+		}
+
+		num_pipes = dml21_find_dc_pipes_for_plane(in_dc, context, dml_ctx, dc_main_pipes, dc_phantom_pipes, dml_prog_idx);
+		if (num_pipes <= 0 || dc_main_pipes[0]->stream == NULL ||
+		    dc_main_pipes[0]->plane_state == NULL)
+			continue;
+
+		/* get config for each pipe */
+		for (dc_pipe_index = 0; dc_pipe_index < num_pipes; dc_pipe_index++) {
+			ASSERT(dc_main_pipes[dc_pipe_index]);
+			dml21_get_pipe_mcache_config(context, dc_main_pipes[dc_pipe_index], pln_prog, &mcache_config->pipe_configurations[dc_pipe_index]);
+		}
+
+		/* get config for each phantom pipe */
+		if (pln_prog->phantom_plane.valid &&
+				dc_phantom_pipes[0] &&
+				dc_main_pipes[0]->stream &&
+				dc_phantom_pipes[0]->plane_state) {
+			mcache_config = &l->build_mcache_programming_params.mcache_configurations[dml_phantom_prog_idx];
+			memset(mcache_config, 0, sizeof(struct dml2_plane_mcache_configuration_descriptor));
+			mcache_config->plane_descriptor = pln_prog->plane_descriptor;
+			mcache_config->mcache_allocation = &context->bw_ctx.bw.dcn.mcache_allocations[dml_phantom_prog_idx];
+			mcache_config->num_pipes = pln_prog->num_dpps_required;
+			l->build_mcache_programming_params.num_configurations++;
+
+			for (dc_pipe_index = 0; dc_pipe_index < num_pipes; dc_pipe_index++) {
+				ASSERT(dc_phantom_pipes[dc_pipe_index]);
+				dml21_get_pipe_mcache_config(context, dc_phantom_pipes[dc_pipe_index], pln_prog, &mcache_config->pipe_configurations[dc_pipe_index]);
+			}
+
+			/* increment phantom index */
+			dml_phantom_prog_idx++;
+		}
+	}
+
+	/* Call to generate mcache programming per plane per pipe for the given display configuration */
+	dml2_build_mcache_programming(&l->build_mcache_programming_params);
+
+	/* get per plane per pipe mcache programming */
+	for (dml_prog_idx = 0; dml_prog_idx < dml_ctx->v21.mode_programming.programming->display_config.num_planes; dml_prog_idx++) {
+		pln_prog = &dml_ctx->v21.mode_programming.programming->plane_programming[dml_prog_idx];
+
+		num_pipes = dml21_find_dc_pipes_for_plane(in_dc, context, dml_ctx, dc_main_pipes, dc_phantom_pipes, dml_prog_idx);
+		if (num_pipes <= 0 || dc_main_pipes[0]->stream == NULL ||
+		    dc_main_pipes[0]->plane_state == NULL)
+			continue;
+
+		/* get config for each pipe */
+		for (dc_pipe_index = 0; dc_pipe_index < num_pipes; dc_pipe_index++) {
+			ASSERT(dc_main_pipes[dc_pipe_index]);
+			if (l->build_mcache_programming_params.per_plane_pipe_mcache_regs[dml_prog_idx][dc_pipe_index]) {
+				memcpy(&dc_main_pipes[dc_pipe_index]->mcache_regs,
+						l->build_mcache_programming_params.per_plane_pipe_mcache_regs[dml_prog_idx][dc_pipe_index],
+						sizeof(struct dml2_hubp_pipe_mcache_regs));
+			}
+		}
+
+		/* get config for each phantom pipe */
+		if (pln_prog->phantom_plane.valid &&
+				dc_phantom_pipes[0] &&
+				dc_main_pipes[0]->stream &&
+				dc_phantom_pipes[0]->plane_state) {
+			for (dc_pipe_index = 0; dc_pipe_index < num_pipes; dc_pipe_index++) {
+				ASSERT(dc_phantom_pipes[dc_pipe_index]);
+				if (l->build_mcache_programming_params.per_plane_pipe_mcache_regs[dml_phantom_prog_idx][dc_pipe_index]) {
+					memcpy(&dc_phantom_pipes[dc_pipe_index]->mcache_regs,
+							l->build_mcache_programming_params.per_plane_pipe_mcache_regs[dml_phantom_prog_idx][dc_pipe_index],
+							sizeof(struct dml2_hubp_pipe_mcache_regs));
+				}
+			}
+			/* increment phantom index */
+			dml_phantom_prog_idx++;
+		}
+	}
+}
+
+void dml21_copy(struct dml2_context *dst_dml_ctx,
+	struct dml2_context *src_dml_ctx)
+{
+	/* Preserve references to internals */
+	struct dml2_instance *dst_dml2_instance = dst_dml_ctx->v21.dml_init.dml2_instance;
+	struct dml2_display_cfg_programming *dst_dml2_programming = dst_dml_ctx->v21.mode_programming.programming;
+
+	/* Copy context */
+	memcpy(dst_dml_ctx, src_dml_ctx, sizeof(struct dml2_context));
+
+	/* Copy Internals */
+	memcpy(dst_dml2_instance, src_dml_ctx->v21.dml_init.dml2_instance, sizeof(struct dml2_instance));
+	memcpy(dst_dml2_programming, src_dml_ctx->v21.mode_programming.programming, sizeof(struct dml2_display_cfg_programming));
+
+	/* Restore references to internals */
+	dst_dml_ctx->v21.dml_init.dml2_instance = dst_dml2_instance;
+
+	dst_dml_ctx->v21.mode_support.dml2_instance = dst_dml2_instance;
+	dst_dml_ctx->v21.mode_programming.dml2_instance = dst_dml2_instance;
+
+	dst_dml_ctx->v21.mode_support.display_config = &dst_dml_ctx->v21.display_config;
+	dst_dml_ctx->v21.mode_programming.display_config = dst_dml_ctx->v21.mode_support.display_config;
+
+	dst_dml_ctx->v21.mode_programming.programming = dst_dml2_programming;
+
+	DC_FP_START();
+
+	/* need to initialize copied instance for internal references to be correct */
+	dml2_initialize_instance(&dst_dml_ctx->v21.dml_init);
+
+	DC_FP_END();
+}
+
+bool dml21_create_copy(struct dml2_context **dst_dml_ctx,
+	struct dml2_context *src_dml_ctx)
+{
+	/* Allocate memory for initializing DML21 instance */
+	if (!dml21_allocate_memory(dst_dml_ctx))
+		return false;
+
+	dml21_copy(*dst_dml_ctx, src_dml_ctx);
+
+	return true;
+}
+
+void dml21_reinit(const struct dc *in_dc, struct dml2_context *dml_ctx, const struct dml2_configuration_options *config)
+{
+	dml21_init(in_dc, dml_ctx, config);
+}
+
diff --git a/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/dml21_wrapper.h b/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/dml21_wrapper.h
new file mode 100644
index 000000000000..15f92029d2e5
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/dml21_wrapper.h
@@ -0,0 +1,135 @@
+// SPDX-License-Identifier: MIT
+//
+// Copyright 2024 Advanced Micro Devices, Inc.
+
+
+#ifndef _DML21_WRAPPER_H_
+#define _DML21_WRAPPER_H_
+
+#include "os_types.h"
+#include "dml_top_soc_parameter_types.h"
+#include "dml_top_display_cfg_types.h"
+
+struct dc;
+struct dc_state;
+struct dml2_configuration_options;
+struct dml2_context;
+enum dc_validate_mode;
+
+/**
+ * dml2_create - Creates dml21_context.
+ * @in_dc: dc.
+ * @dml2: Created dml21 context.
+ * @config: dml21 configuration options.
+ *
+ * Create of DML21 is done as part of dc_state creation.
+ * DML21 IP, SOC and STATES are initialized at
+ * creation time.
+ *
+ * Return: True if dml2 is successfully created, false otherwise.
+ */
+bool dml21_create(const struct dc *in_dc, struct dml2_context **dml_ctx, const struct dml2_configuration_options *config);
+void dml21_destroy(struct dml2_context *dml2);
+void dml21_copy(struct dml2_context *dst_dml_ctx,
+	struct dml2_context *src_dml_ctx);
+bool dml21_create_copy(struct dml2_context **dst_dml_ctx,
+	struct dml2_context *src_dml_ctx);
+void dml21_reinit(const struct dc *in_dc, struct dml2_context *dml_ctx, const struct dml2_configuration_options *config);
+
+/**
+ * dml21_validate - Determines if a display configuration is supported or not.
+ * @in_dc: dc.
+ * @context: dc_state to be validated.
+ * @validate_mode: DC_VALIDATE_MODE_ONLY and DC_VALIDATE_MODE_AND_STATE_INDEX
+ *           will not populate context.res_ctx.
+ *
+ * Based on fast_validate option internally would call:
+ *
+ * -dml21_mode_check_and_programming - for DC_VALIDATE_MODE_AND_PROGRAMMING option
+ * Calculates if dc_state can be supported on the input display
+ * configuration. If supported, generates the necessary HW
+ * programming for the new dc_state.
+ *
+ * -dml21_check_mode_support - for DC_VALIDATE_MODE_ONLY and DC_VALIDATE_MODE_AND_STATE_INDEX option
+ * Calculates if dc_state can be supported for the input display
+ * config.
+
+ * Context: Two threads may not invoke this function concurrently unless they reference
+ *          separate dc_states for validation.
+ * Return: True if mode is supported, false otherwise.
+ */
+bool dml21_validate(const struct dc *in_dc, struct dc_state *context, struct dml2_context *dml_ctx,
+	enum dc_validate_mode validate_mode);
+
+/* Prepare hubp mcache_regs for hubp mcache ID and split coordinate programming */
+void dml21_prepare_mcache_programming(struct dc *in_dc, struct dc_state *context, struct dml2_context *dml_ctx);
+
+/* Structure for inputting external SOCBB and DCNIP values for tool based debugging. */
+struct socbb_ip_params_external {
+	struct dml2_ip_capabilities ip_params;
+	struct dml2_soc_bb soc_bb;
+};
+
+/*mcache parameters decided by dml*/
+struct dc_mcache_params {
+	bool valid;
+	/*
+	* For iMALL, dedicated mall mcaches are required (sharing of last
+	* slice possible), for legacy phantom or phantom without return
+	* the only mall mcaches need to be valid.
+	*/
+	bool requires_dedicated_mall_mcache;
+	unsigned int num_mcaches_plane0;
+	unsigned int num_mcaches_plane1;
+	/*
+	* Generally, plane0/1 slices must use a disjoint set of caches
+	* but in some cases the final segement of the two planes can
+	* use the same cache. If plane0_plane1 is set, then this is
+	* allowed.
+	*
+	* Similarly, the caches allocated to MALL prefetcher are generally
+	* disjoint, but if mall_prefetch is set, then the final segment
+	* between the main and the mall pixel requestor can use the same
+	* cache.
+	*
+	* Note that both bits may be set at the same time.
+	*/
+	struct {
+		bool mall_comb_mcache_p0;
+		bool mall_comb_mcache_p1;
+		bool plane0_plane1;
+	} last_slice_sharing;
+	/*
+	* A plane is divided into vertical slices of mcaches,
+	* which wrap on the surface width.
+	*
+	* For example, if the surface width is 7680, and split into
+	* three slices of equal width, the boundary array would contain
+	* [2560, 5120, 7680]
+	*
+	* The assignments are
+	* 0 = [0 .. 2559]
+	* 1 = [2560 .. 5119]
+	* 2 = [5120 .. 7679]
+	* 0 = [7680 .. INF]
+	* The final element implicitly is the same as the first, and
+	* at first seems invalid since it is never referenced (since)
+	* it is outside the surface. However, its useful when shifting
+	* (see below).
+	*
+	* For any given valid mcache assignment, a shifted version, wrapped
+	* on the surface width boundary is also assumed to be valid.
+	*
+	* For example, shifting [2560, 5120, 7680] by -50 results in
+	* [2510, 5170, 7630].
+	*
+	* The assignments are now:
+	* 0 = [0 .. 2509]
+	* 1 = [2510 .. 5169]
+	* 2 = [5170 .. 7629]
+	* 0 = [7630 .. INF]
+	*/
+	int mcache_x_offsets_plane0[DML2_MAX_MCACHES + 1];
+	int mcache_x_offsets_plane1[DML2_MAX_MCACHES + 1];
+};
+#endif
diff --git a/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/inc/bounding_boxes/dcn4_soc_bb.h b/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/inc/bounding_boxes/dcn4_soc_bb.h
new file mode 100644
index 000000000000..16a4f97bca4e
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/inc/bounding_boxes/dcn4_soc_bb.h
@@ -0,0 +1,372 @@
+// SPDX-License-Identifier: MIT
+//
+// Copyright 2024 Advanced Micro Devices, Inc.
+
+#ifndef __DML_DML_DCN4_SOC_BB__
+#define __DML_DML_DCN4_SOC_BB__
+
+#include "dml_top_soc_parameter_types.h"
+
+static const struct dml2_soc_qos_parameters dml_dcn4_variant_a_soc_qos_params = {
+	.derate_table = {
+		.system_active_urgent = {
+			.dram_derate_percent_pixel = 22,
+			.dram_derate_percent_vm = 0,
+			.dram_derate_percent_pixel_and_vm = 0,
+			.fclk_derate_percent = 76,
+			.dcfclk_derate_percent = 100,
+		},
+		.system_active_average = {
+			.dram_derate_percent_pixel = 17,
+			.dram_derate_percent_vm = 0,
+			.dram_derate_percent_pixel_and_vm = 0,
+			.fclk_derate_percent = 57,
+			.dcfclk_derate_percent = 75,
+		},
+		.dcn_mall_prefetch_urgent = {
+			.dram_derate_percent_pixel = 40,
+			.dram_derate_percent_vm = 0,
+			.dram_derate_percent_pixel_and_vm = 0,
+			.fclk_derate_percent = 83,
+			.dcfclk_derate_percent = 100,
+		},
+		.dcn_mall_prefetch_average = {
+			.dram_derate_percent_pixel = 33,
+			.dram_derate_percent_vm = 0,
+			.dram_derate_percent_pixel_and_vm = 0,
+			.fclk_derate_percent = 62,
+			.dcfclk_derate_percent = 83,
+		},
+		.system_idle_average = {
+			.dram_derate_percent_pixel = 70,
+			.dram_derate_percent_vm = 0,
+			.dram_derate_percent_pixel_and_vm = 0,
+			.fclk_derate_percent = 83,
+			.dcfclk_derate_percent = 100,
+		},
+	},
+	.writeback = {
+		.base_latency_us = 12,
+		.scaling_factor_us = 0,
+		.scaling_factor_mhz = 0,
+	},
+	.qos_params = {
+		.dcn4x = {
+			.df_qos_response_time_fclk_cycles = 300,
+			.max_round_trip_to_furthest_cs_fclk_cycles = 350,
+			.mall_overhead_fclk_cycles = 50,
+			.meta_trip_adder_fclk_cycles = 36,
+			.average_transport_distance_fclk_cycles = 257,
+			.umc_urgent_ramp_latency_margin = 50,
+			.umc_max_latency_margin = 30,
+			.umc_average_latency_margin = 20,
+			.fabric_max_transport_latency_margin = 20,
+			.fabric_average_transport_latency_margin = 10,
+
+			.per_uclk_dpm_params = {
+				{
+					.minimum_uclk_khz = 97 * 1000,
+					.urgent_ramp_uclk_cycles = 472,
+					.trip_to_memory_uclk_cycles = 827,
+					.meta_trip_to_memory_uclk_cycles = 827,
+					.maximum_latency_when_urgent_uclk_cycles = 72,
+					.average_latency_when_urgent_uclk_cycles = 61,
+					.maximum_latency_when_non_urgent_uclk_cycles = 827,
+					.average_latency_when_non_urgent_uclk_cycles = 118,
+				},
+			},
+		},
+	},
+	.qos_type = dml2_qos_param_type_dcn4x,
+};
+
+static const struct dml2_soc_bb dml2_socbb_dcn401 = {
+	.clk_table = {
+		.uclk = {
+				.clk_values_khz = {97000},
+				.num_clk_values = 1,
+		},
+		.fclk = {
+				.clk_values_khz = {300000, 2500000},
+				.num_clk_values = 2,
+		},
+		.dcfclk = {
+				.clk_values_khz = {200000, 1564000},
+				.num_clk_values = 2,
+		},
+		.dispclk = {
+				.clk_values_khz = {100000, 2000000},
+				.num_clk_values = 2,
+		},
+		.dppclk = {
+				.clk_values_khz = {100000, 2000000},
+				.num_clk_values = 2,
+		},
+		.dtbclk = {
+				.clk_values_khz = {100000, 1564000},
+				.num_clk_values = 2,
+		},
+		.phyclk = {
+				.clk_values_khz = {810000, 810000},
+				.num_clk_values = 2,
+		},
+		.socclk = {
+				.clk_values_khz = {300000, 1200000},
+				.num_clk_values = 2,
+		},
+		.dscclk = {
+				.clk_values_khz = {666667, 666667},
+				.num_clk_values = 2,
+		},
+		.phyclk_d18 = {
+				.clk_values_khz = {625000, 625000},
+				.num_clk_values = 2,
+		},
+		.phyclk_d32 = {
+				.clk_values_khz = {625000, 625000},
+				.num_clk_values = 2,
+		},
+		.dram_config = {
+			.channel_width_bytes = 2,
+			.channel_count = 16,
+			.transactions_per_clock = 16,
+		},
+	},
+
+	.qos_parameters = {
+		.derate_table = {
+			.system_active_urgent = {
+				.dram_derate_percent_pixel = 22,
+				.dram_derate_percent_vm = 0,
+				.dram_derate_percent_pixel_and_vm = 0,
+				.fclk_derate_percent = 76,
+				.dcfclk_derate_percent = 100,
+			},
+			.system_active_average = {
+				.dram_derate_percent_pixel = 15,
+				.dram_derate_percent_vm = 0,
+				.dram_derate_percent_pixel_and_vm = 0,
+				.fclk_derate_percent = 57,
+				.dcfclk_derate_percent = 75,
+			},
+			.dcn_mall_prefetch_urgent = {
+				.dram_derate_percent_pixel = 40,
+				.dram_derate_percent_vm = 0,
+				.dram_derate_percent_pixel_and_vm = 0,
+				.fclk_derate_percent = 83,
+				.dcfclk_derate_percent = 100,
+			},
+			.dcn_mall_prefetch_average = {
+				.dram_derate_percent_pixel = 30,
+				.dram_derate_percent_vm = 0,
+				.dram_derate_percent_pixel_and_vm = 0,
+				.fclk_derate_percent = 62,
+				.dcfclk_derate_percent = 83,
+			},
+			.system_idle_average = {
+				.dram_derate_percent_pixel = 70,
+				.dram_derate_percent_vm = 0,
+				.dram_derate_percent_pixel_and_vm = 0,
+				.fclk_derate_percent = 83,
+				.dcfclk_derate_percent = 100,
+			},
+		},
+		.writeback = {
+			.base_latency_us = 0,
+			.scaling_factor_us = 0,
+			.scaling_factor_mhz = 0,
+		},
+		.qos_params = {
+			.dcn4x = {
+				.df_qos_response_time_fclk_cycles = 300,
+				.max_round_trip_to_furthest_cs_fclk_cycles = 350,
+				.mall_overhead_fclk_cycles = 50,
+				.meta_trip_adder_fclk_cycles = 36,
+				.average_transport_distance_fclk_cycles = 260,
+				.umc_urgent_ramp_latency_margin = 50,
+				.umc_max_latency_margin = 30,
+				.umc_average_latency_margin = 20,
+				.fabric_max_transport_latency_margin = 20,
+				.fabric_average_transport_latency_margin = 10,
+
+				.per_uclk_dpm_params = {
+					{
+						// State 1
+						.minimum_uclk_khz = 0,
+						.urgent_ramp_uclk_cycles = 472,
+						.trip_to_memory_uclk_cycles = 827,
+						.meta_trip_to_memory_uclk_cycles = 827,
+						.maximum_latency_when_urgent_uclk_cycles = 72,
+						.average_latency_when_urgent_uclk_cycles = 72,
+						.maximum_latency_when_non_urgent_uclk_cycles = 827,
+						.average_latency_when_non_urgent_uclk_cycles = 117,
+					},
+					{
+						// State 2
+						.minimum_uclk_khz = 0,
+						.urgent_ramp_uclk_cycles = 546,
+						.trip_to_memory_uclk_cycles = 848,
+						.meta_trip_to_memory_uclk_cycles = 848,
+						.maximum_latency_when_urgent_uclk_cycles = 146,
+						.average_latency_when_urgent_uclk_cycles = 146,
+						.maximum_latency_when_non_urgent_uclk_cycles = 848,
+						.average_latency_when_non_urgent_uclk_cycles = 133,
+					},
+					{
+						// State 3
+						.minimum_uclk_khz = 0,
+						.urgent_ramp_uclk_cycles = 564,
+						.trip_to_memory_uclk_cycles = 853,
+						.meta_trip_to_memory_uclk_cycles = 853,
+						.maximum_latency_when_urgent_uclk_cycles = 164,
+						.average_latency_when_urgent_uclk_cycles = 164,
+						.maximum_latency_when_non_urgent_uclk_cycles = 853,
+						.average_latency_when_non_urgent_uclk_cycles = 136,
+					},
+					{
+						// State 4
+						.minimum_uclk_khz = 0,
+						.urgent_ramp_uclk_cycles = 613,
+						.trip_to_memory_uclk_cycles = 869,
+						.meta_trip_to_memory_uclk_cycles = 869,
+						.maximum_latency_when_urgent_uclk_cycles = 213,
+						.average_latency_when_urgent_uclk_cycles = 213,
+						.maximum_latency_when_non_urgent_uclk_cycles = 869,
+						.average_latency_when_non_urgent_uclk_cycles = 149,
+					},
+					{
+						// State 5
+						.minimum_uclk_khz = 0,
+						.urgent_ramp_uclk_cycles = 632,
+						.trip_to_memory_uclk_cycles = 874,
+						.meta_trip_to_memory_uclk_cycles = 874,
+						.maximum_latency_when_urgent_uclk_cycles = 232,
+						.average_latency_when_urgent_uclk_cycles = 232,
+						.maximum_latency_when_non_urgent_uclk_cycles = 874,
+						.average_latency_when_non_urgent_uclk_cycles = 153,
+					},
+					{
+						// State 6
+						.minimum_uclk_khz = 0,
+						.urgent_ramp_uclk_cycles = 665,
+						.trip_to_memory_uclk_cycles = 885,
+						.meta_trip_to_memory_uclk_cycles = 885,
+						.maximum_latency_when_urgent_uclk_cycles = 265,
+						.average_latency_when_urgent_uclk_cycles = 265,
+						.maximum_latency_when_non_urgent_uclk_cycles = 885,
+						.average_latency_when_non_urgent_uclk_cycles = 161,
+					},
+					{
+						// State 7
+						.minimum_uclk_khz = 0,
+						.urgent_ramp_uclk_cycles = 689,
+						.trip_to_memory_uclk_cycles = 895,
+						.meta_trip_to_memory_uclk_cycles = 895,
+						.maximum_latency_when_urgent_uclk_cycles = 289,
+						.average_latency_when_urgent_uclk_cycles = 289,
+						.maximum_latency_when_non_urgent_uclk_cycles = 895,
+						.average_latency_when_non_urgent_uclk_cycles = 167,
+					},
+					{
+						// State 8
+						.minimum_uclk_khz = 0,
+						.urgent_ramp_uclk_cycles = 716,
+						.trip_to_memory_uclk_cycles = 902,
+						.meta_trip_to_memory_uclk_cycles = 902,
+						.maximum_latency_when_urgent_uclk_cycles = 316,
+						.average_latency_when_urgent_uclk_cycles = 316,
+						.maximum_latency_when_non_urgent_uclk_cycles = 902,
+						.average_latency_when_non_urgent_uclk_cycles = 174,
+					},
+				},
+			},
+		},
+		.qos_type = dml2_qos_param_type_dcn4x,
+	},
+
+	.power_management_parameters = {
+		.dram_clk_change_blackout_us = 400,
+		.fclk_change_blackout_us = 0,
+		.g7_ppt_blackout_us = 0,
+		.stutter_enter_plus_exit_latency_us = 54,
+		.stutter_exit_latency_us = 41,
+		.z8_stutter_enter_plus_exit_latency_us = 0,
+		.z8_stutter_exit_latency_us = 0,
+		/*
+		.g6_temp_read_blackout_us = {
+			23.00,
+			10.00,
+			10.00,
+			8.00,
+			8.00,
+			5.00,
+			5.00,
+			5.00,
+		},
+		*/
+	},
+
+	 .vmin_limit = {
+		.dispclk_khz = 600 * 1000,
+	 },
+
+	.dprefclk_mhz = 720,
+	.xtalclk_mhz = 100,
+	.pcie_refclk_mhz = 100,
+	.dchub_refclk_mhz = 50,
+	.mall_allocated_for_dcn_mbytes = 64,
+	.max_outstanding_reqs = 512,
+	.fabric_datapath_to_dcn_data_return_bytes = 64,
+	.return_bus_width_bytes = 64,
+	.hostvm_min_page_size_kbytes = 0,
+	.gpuvm_min_page_size_kbytes = 256,
+	.phy_downspread_percent = 0.38,
+	.dcn_downspread_percent = 0.38,
+	.dispclk_dppclk_vco_speed_mhz = 4500,
+	.do_urgent_latency_adjustment = 0,
+	.mem_word_bytes = 32,
+	.num_dcc_mcaches = 8,
+	.mcache_size_bytes = 2048,
+	.mcache_line_size_bytes = 32,
+	.max_fclk_for_uclk_dpm_khz = 1250 * 1000,
+};
+
+static const struct dml2_ip_capabilities dml2_dcn401_max_ip_caps = {
+	.pipe_count = 4,
+	.otg_count = 4,
+	.num_dsc = 4,
+	.max_num_dp2p0_streams = 4,
+	.max_num_hdmi_frl_outputs = 1,
+	.max_num_dp2p0_outputs = 4,
+	.rob_buffer_size_kbytes = 192,
+	.config_return_buffer_size_in_kbytes = 1344,
+	.config_return_buffer_segment_size_in_kbytes = 64,
+	.meta_fifo_size_in_kentries = 22,
+	.compressed_buffer_segment_size_in_kbytes = 64,
+	.cursor_buffer_size = 24,
+	.max_flip_time_us = 80,
+	.max_flip_time_lines = 32,
+	.hostvm_mode = 0,
+	.subvp_drr_scheduling_margin_us = 100,
+	.subvp_prefetch_end_to_mall_start_us = 15,
+	.subvp_fw_processing_delay = 15,
+	.max_vactive_det_fill_delay_us = 400,
+
+	.fams2 = {
+		.max_allow_delay_us = 100 * 1000,
+		.scheduling_delay_us = 550,
+		.vertical_interrupt_ack_delay_us = 40,
+		.allow_programming_delay_us = 18,
+		.min_allow_width_us = 20,
+		.subvp_df_throttle_delay_us = 100,
+		.subvp_programming_delay_us = 200,
+		.subvp_prefetch_to_mall_delay_us = 18,
+		.drr_programming_delay_us = 35,
+
+		.lock_timeout_us = 5000,
+		.recovery_timeout_us = 5000,
+		.flip_programming_delay_us = 300,
+	},
+};
+
+#endif
diff --git a/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/inc/dml2_external_lib_deps.h b/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/inc/dml2_external_lib_deps.h
new file mode 100644
index 000000000000..281d7ad230d8
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/inc/dml2_external_lib_deps.h
@@ -0,0 +1,10 @@
+// SPDX-License-Identifier: MIT
+//
+// Copyright 2024 Advanced Micro Devices, Inc.
+
+#ifndef __DML2_EXTERNAL_LIB_DEPS__
+#define __DML2_EXTERNAL_LIB_DEPS__
+
+#include "os_types.h"
+
+#endif
diff --git a/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/inc/dml_top.h b/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/inc/dml_top.h
new file mode 100644
index 000000000000..a64ec4dcf11a
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/inc/dml_top.h
@@ -0,0 +1,46 @@
+// SPDX-License-Identifier: MIT
+//
+// Copyright 2024 Advanced Micro Devices, Inc.
+
+#ifndef __DML_TOP_H__
+#define __DML_TOP_H__
+
+#include "dml_top_types.h"
+
+/*
+ * Top Level Interface for DML2
+ */
+
+/*
+ * Returns the size of the DML instance for the caller to allocate
+ */
+unsigned int dml2_get_instance_size_bytes(void);
+
+/*
+ * Initializes the DML instance (i.e. with configuration, soc BB, IP params, etc...)
+ */
+bool dml2_initialize_instance(struct dml2_initialize_instance_in_out *in_out);
+
+/*
+ * Determines if the input mode is supported (boolean) on the SoC at all.  Does not return
+ * information on how mode should be programmed.
+ */
+bool dml2_check_mode_supported(struct dml2_check_mode_supported_in_out *in_out);
+
+/*
+ * Determines the full (optimized) programming for the input mode.  Returns minimum
+ * clocks as well as dchub register programming values for all pipes, additional meta
+ * such as ODM or MPCC combine factors.
+ */
+bool dml2_build_mode_programming(struct dml2_build_mode_programming_in_out *in_out);
+
+/*
+ * Determines the correct per pipe mcache register programming for a valid mode.
+ * The mcache allocation must have been calculated (successfully) in a previous
+ * call to dml2_build_mode_programming.
+ * The actual hubp viewport dimensions be what the actual registers will be
+ * programmed to (i.e. based on scaler setup).
+ */
+bool dml2_build_mcache_programming(struct dml2_build_mcache_programming_in_out *in_out);
+
+#endif
diff --git a/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/inc/dml_top_dchub_registers.h b/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/inc/dml_top_dchub_registers.h
new file mode 100644
index 000000000000..bf57df42d1d9
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/inc/dml_top_dchub_registers.h
@@ -0,0 +1,191 @@
+// SPDX-License-Identifier: MIT
+//
+// Copyright 2024 Advanced Micro Devices, Inc.
+
+#ifndef __dml2_TOP_DCHUB_REGISTERS_H__
+#define __dml2_TOP_DCHUB_REGISTERS_H__
+
+#include "dml2_external_lib_deps.h"
+// These types are uint32_t as they represent actual calculated register values for HW
+
+struct dml2_display_dlg_regs {
+	uint32_t refcyc_h_blank_end;
+	uint32_t dlg_vblank_end;
+	uint32_t min_dst_y_next_start;
+	uint32_t refcyc_per_htotal;
+	uint32_t refcyc_x_after_scaler;
+	uint32_t dst_y_after_scaler;
+	uint32_t dst_y_prefetch;
+	uint32_t dst_y_per_vm_vblank;
+	uint32_t dst_y_per_row_vblank;
+	uint32_t dst_y_per_vm_flip;
+	uint32_t dst_y_per_row_flip;
+	uint32_t ref_freq_to_pix_freq;
+	uint32_t vratio_prefetch;
+	uint32_t vratio_prefetch_c;
+	uint32_t refcyc_per_tdlut_group;
+	uint32_t refcyc_per_pte_group_vblank_l;
+	uint32_t refcyc_per_pte_group_vblank_c;
+	uint32_t refcyc_per_pte_group_flip_l;
+	uint32_t refcyc_per_pte_group_flip_c;
+	uint32_t dst_y_per_pte_row_nom_l;
+	uint32_t dst_y_per_pte_row_nom_c;
+	uint32_t refcyc_per_pte_group_nom_l;
+	uint32_t refcyc_per_pte_group_nom_c;
+	uint32_t refcyc_per_line_delivery_pre_l;
+	uint32_t refcyc_per_line_delivery_pre_c;
+	uint32_t refcyc_per_line_delivery_l;
+	uint32_t refcyc_per_line_delivery_c;
+	uint32_t refcyc_per_vm_group_vblank;
+	uint32_t refcyc_per_vm_group_flip;
+	uint32_t refcyc_per_vm_req_vblank;
+	uint32_t refcyc_per_vm_req_flip;
+	uint32_t dst_y_offset_cur0;
+	uint32_t chunk_hdl_adjust_cur0;
+	uint32_t vready_after_vcount0;
+	uint32_t dst_y_delta_drq_limit;
+	uint32_t refcyc_per_vm_dmdata;
+	uint32_t dmdata_dl_delta;
+
+	// MRQ
+	uint32_t refcyc_per_meta_chunk_vblank_l;
+	uint32_t refcyc_per_meta_chunk_vblank_c;
+	uint32_t refcyc_per_meta_chunk_flip_l;
+	uint32_t refcyc_per_meta_chunk_flip_c;
+	uint32_t dst_y_per_meta_row_nom_l;
+	uint32_t dst_y_per_meta_row_nom_c;
+	uint32_t refcyc_per_meta_chunk_nom_l;
+	uint32_t refcyc_per_meta_chunk_nom_c;
+};
+
+struct dml2_display_ttu_regs {
+	uint32_t qos_level_low_wm;
+	uint32_t qos_level_high_wm;
+	uint32_t min_ttu_vblank;
+	uint32_t qos_level_flip;
+	uint32_t refcyc_per_req_delivery_l;
+	uint32_t refcyc_per_req_delivery_c;
+	uint32_t refcyc_per_req_delivery_cur0;
+	uint32_t refcyc_per_req_delivery_pre_l;
+	uint32_t refcyc_per_req_delivery_pre_c;
+	uint32_t refcyc_per_req_delivery_pre_cur0;
+	uint32_t qos_level_fixed_l;
+	uint32_t qos_level_fixed_c;
+	uint32_t qos_level_fixed_cur0;
+	uint32_t qos_ramp_disable_l;
+	uint32_t qos_ramp_disable_c;
+	uint32_t qos_ramp_disable_cur0;
+};
+
+struct dml2_display_arb_regs {
+	uint32_t max_req_outstanding;
+	uint32_t min_req_outstanding;
+	uint32_t sat_level_us;
+	uint32_t hvm_max_qos_commit_threshold;
+	uint32_t hvm_min_req_outstand_commit_threshold;
+	uint32_t compbuf_reserved_space_kbytes;
+	uint32_t compbuf_size;
+	uint32_t sdpif_request_rate_limit;
+	uint32_t allow_sdpif_rate_limit_when_cstate_req;
+	uint32_t dcfclk_deep_sleep_hysteresis;
+	uint32_t pstate_stall_threshold;
+};
+
+struct dml2_cursor_dlg_regs{
+	uint32_t dst_x_offset;			   // CURSOR0_DST_X_OFFSET
+	uint32_t dst_y_offset;			   // CURSOR0_DST_Y_OFFSET
+	uint32_t chunk_hdl_adjust;		   // CURSOR0_CHUNK_HDL_ADJUST
+
+	uint32_t qos_level_fixed;
+	uint32_t qos_ramp_disable;
+};
+
+struct dml2_display_plane_rq_regs {
+	uint32_t chunk_size;
+	uint32_t min_chunk_size;
+	uint32_t dpte_group_size;
+	uint32_t mpte_group_size;
+	uint32_t swath_height;
+	uint32_t pte_row_height_linear;
+
+	// MRQ
+	uint32_t meta_chunk_size;
+	uint32_t min_meta_chunk_size;
+};
+
+struct dml2_display_rq_regs {
+	struct dml2_display_plane_rq_regs rq_regs_l;
+	struct dml2_display_plane_rq_regs rq_regs_c;
+	uint32_t drq_expansion_mode;
+	uint32_t prq_expansion_mode;
+	uint32_t crq_expansion_mode;
+	uint32_t plane1_base_address;
+	uint32_t unbounded_request_enabled;
+	bool pte_buffer_mode;
+	bool force_one_row_for_frame;
+
+	// MRQ
+	uint32_t mrq_expansion_mode;
+};
+
+struct dml2_display_mcache_regs {
+	uint32_t mcache_id_first;
+	uint32_t mcache_id_second;
+	uint32_t split_location;
+};
+
+struct dml2_hubp_pipe_mcache_regs {
+	struct {
+		struct dml2_display_mcache_regs p0;
+		struct dml2_display_mcache_regs p1;
+	} main;
+	struct {
+		struct dml2_display_mcache_regs p0;
+		struct dml2_display_mcache_regs p1;
+	} mall;
+};
+
+struct dml2_dchub_per_pipe_register_set {
+	struct dml2_display_rq_regs rq_regs;
+	struct dml2_display_ttu_regs ttu_regs;
+	struct dml2_display_dlg_regs dlg_regs;
+
+	uint32_t det_size;
+};
+
+struct dml2_dchub_watermark_regs {
+	/* watermarks */
+	uint32_t urgent;
+	uint32_t sr_enter;
+	uint32_t sr_exit;
+	uint32_t sr_enter_z8;
+	uint32_t sr_exit_z8;
+	uint32_t sr_enter_low_power;
+	uint32_t sr_exit_low_power;
+	uint32_t uclk_pstate;
+	uint32_t fclk_pstate;
+	uint32_t temp_read_or_ppt;
+	uint32_t usr;
+	/* qos */
+	uint32_t refcyc_per_trip_to_mem;
+	uint32_t refcyc_per_meta_trip_to_mem;
+	uint32_t frac_urg_bw_flip;
+	uint32_t frac_urg_bw_nom;
+	uint32_t frac_urg_bw_mall;
+};
+
+enum dml2_dchub_watermark_reg_set_index {
+	DML2_DCHUB_WATERMARK_SET_A = 0,
+	DML2_DCHUB_WATERMARK_SET_B = 1,
+	DML2_DCHUB_WATERMARK_SET_C = 2,
+	DML2_DCHUB_WATERMARK_SET_D = 3,
+	DML2_DCHUB_WATERMARK_SET_NUM = 4,
+};
+
+struct dml2_dchub_global_register_set {
+	struct dml2_display_arb_regs arb_regs;
+	struct dml2_dchub_watermark_regs wm_regs[DML2_DCHUB_WATERMARK_SET_NUM];
+	unsigned int num_watermark_sets;
+};
+
+#endif
diff --git a/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/inc/dml_top_display_cfg_types.h b/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/inc/dml_top_display_cfg_types.h
new file mode 100644
index 000000000000..35aa954248cd
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/inc/dml_top_display_cfg_types.h
@@ -0,0 +1,526 @@
+// SPDX-License-Identifier: MIT
+//
+// Copyright 2024 Advanced Micro Devices, Inc.
+
+#ifndef __DML_TOP_DISPLAY_CFG_TYPES_H__
+#define __DML_TOP_DISPLAY_CFG_TYPES_H__
+
+#include "dml2_external_lib_deps.h"
+
+#define DML2_MAX_PLANES 8
+#define DML2_MAX_DCN_PIPES 8
+#define DML2_MAX_MCACHES 8 // assume plane is going to be supported by a max of 8 mcaches
+#define DML2_MAX_WRITEBACK 3
+
+enum dml2_swizzle_mode {
+	dml2_sw_linear, // SW_LINEAR accepts 256 byte aligned pitch and also 128 byte aligned pitch if DCC is not enabled
+	dml2_sw_256b_2d,
+	dml2_sw_4kb_2d,
+	dml2_sw_64kb_2d,
+	dml2_sw_256kb_2d,
+
+	dml2_gfx11_sw_linear,
+	dml2_gfx11_sw_64kb_d,
+	dml2_gfx11_sw_64kb_d_t,
+	dml2_gfx11_sw_64kb_d_x,
+	dml2_gfx11_sw_64kb_r_x,
+	dml2_gfx11_sw_256kb_d_x,
+	dml2_gfx11_sw_256kb_r_x,
+
+};
+
+enum dml2_source_format_class {
+	dml2_444_8 = 0,
+	dml2_444_16 = 1,
+	dml2_444_32 = 2,
+	dml2_444_64 = 3,
+	dml2_420_8 = 4,
+	dml2_420_10 = 5,
+	dml2_420_12 = 6,
+	dml2_rgbe_alpha = 9,
+	dml2_rgbe = 10,
+	dml2_mono_8 = 11,
+	dml2_mono_16 = 12,
+	dml2_422_planar_8 = 13,
+	dml2_422_planar_10 = 14,
+	dml2_422_planar_12 = 15,
+	dml2_422_packed_8 = 16,
+	dml2_422_packed_10 = 17,
+	dml2_422_packed_12 = 18
+};
+
+enum dml2_sample_positioning {
+	dml2_interstitial = 0,
+	dml2_cosited = 1
+};
+
+enum dml2_rotation_angle {
+	dml2_rotation_0 = 0,
+	dml2_rotation_90 = 1,
+	dml2_rotation_180 = 2,
+	dml2_rotation_270 = 3
+};
+
+enum dml2_output_format_class {
+	dml2_444 = 0,
+	dml2_s422 = 1,
+	dml2_n422 = 2,
+	dml2_420 = 3
+};
+
+enum dml2_output_encoder_class {
+	dml2_dp = 0,
+	dml2_edp = 1,
+	dml2_dp2p0 = 2,
+	dml2_hdmi = 3,
+	dml2_hdmifrl = 4,
+	dml2_none = 5
+};
+
+enum dml2_output_link_dp_rate {
+	dml2_dp_rate_na = 0,
+	dml2_dp_rate_hbr = 1,
+	dml2_dp_rate_hbr2 = 2,
+	dml2_dp_rate_hbr3 = 3,
+	dml2_dp_rate_uhbr10 = 4,
+	dml2_dp_rate_uhbr13p5 = 5,
+	dml2_dp_rate_uhbr20 = 6
+};
+
+enum dml2_pstate_type {
+	dml2_pstate_type_uclk = 0,
+	dml2_pstate_type_fclk = 1,
+	dml2_pstate_type_ppt = 2,
+	dml2_pstate_type_temp_read = 3,
+	dml2_pstate_type_dummy_pstate = 4,
+	dml2_pstate_type_count = 5
+};
+
+enum dml2_uclk_pstate_change_strategy {
+	dml2_uclk_pstate_change_strategy_auto = 0,
+	dml2_uclk_pstate_change_strategy_force_vactive = 1,
+	dml2_uclk_pstate_change_strategy_force_vblank = 2,
+	dml2_uclk_pstate_change_strategy_force_drr = 3,
+	dml2_uclk_pstate_change_strategy_force_mall_svp = 4,
+	dml2_uclk_pstate_change_strategy_force_mall_full_frame = 5,
+};
+
+enum dml2_svp_mode_override {
+	dml2_svp_mode_override_auto = 0,
+	dml2_svp_mode_override_main_pipe = 1,
+	dml2_svp_mode_override_phantom_pipe = 2, //does not need to be defined explicitly, main overrides result in implicit phantom additions
+	dml2_svp_mode_override_phantom_pipe_no_data_return = 3,
+	dml2_svp_mode_override_imall = 4
+};
+
+enum dml2_refresh_from_mall_mode_override {
+	dml2_refresh_from_mall_mode_override_auto = 0,
+	dml2_refresh_from_mall_mode_override_force_disable = 1,
+	dml2_refresh_from_mall_mode_override_force_enable = 2
+};
+
+enum dml2_odm_mode {
+	dml2_odm_mode_auto = 0,
+	dml2_odm_mode_bypass,
+	dml2_odm_mode_combine_2to1,
+	dml2_odm_mode_combine_3to1,
+	dml2_odm_mode_combine_4to1,
+	dml2_odm_mode_split_1to2,
+	dml2_odm_mode_mso_1to2,
+	dml2_odm_mode_mso_1to4
+};
+
+enum dml2_scaling_transform {
+	dml2_scaling_transform_explicit = 0,
+	dml2_scaling_transform_fullscreen,
+	dml2_scaling_transform_aspect_ratio,
+	dml2_scaling_transform_centered
+};
+
+enum dml2_dsc_enable_option {
+	dml2_dsc_disable = 0,
+	dml2_dsc_enable = 1,
+	dml2_dsc_enable_if_necessary = 2
+};
+
+enum dml2_tdlut_addressing_mode {
+	dml2_tdlut_sw_linear = 0,
+	dml2_tdlut_simple_linear = 1
+};
+
+enum dml2_tdlut_width_mode {
+	dml2_tdlut_width_17_cube = 0,
+	dml2_tdlut_width_33_cube = 1
+};
+
+enum dml2_twait_budgeting_setting {
+	dml2_twait_budgeting_setting_ignore = 0,// Ignore this budget in twait
+
+	dml2_twait_budgeting_setting_if_needed,         // Budget for it only if needed
+											//(i.e. UCLK/FCLK DPM cannot be supported in active)
+
+	dml2_twait_budgeting_setting_try,	   // Budget for it as long as there is an SoC state that
+											// can support it
+};
+
+struct dml2_get_cursor_dlg_reg{
+	unsigned int cursor_x_position;
+	unsigned int cursor_hotspot_x;
+	unsigned int cursor_primary_offset;
+	unsigned int cursor_secondary_offset;
+	bool cursor_stereo_en;
+	bool cursor_2x_magnify;
+	double hratio;
+	double pixel_rate_mhz;
+	double dlg_refclk_mhz;
+};
+
+/// @brief Surface Parameters
+struct dml2_surface_cfg {
+	enum dml2_swizzle_mode tiling;
+
+	struct {
+		unsigned long pitch; // In elements, two pixels per element in 422 packed format
+		unsigned long width;
+		unsigned long height;
+	} plane0;
+
+
+	struct {
+		unsigned long pitch;
+		unsigned long width;
+		unsigned long height;
+	} plane1;
+
+	struct {
+		bool enable;
+		struct {
+			unsigned long pitch;
+		} plane0;
+		struct {
+			unsigned long pitch;
+		} plane1;
+
+		struct {
+			double dcc_rate_plane0;
+			double dcc_rate_plane1;
+			double fraction_of_zero_size_request_plane0;
+			double fraction_of_zero_size_request_plane1;
+		} informative;
+	} dcc;
+};
+
+
+struct dml2_composition_cfg {
+	enum dml2_rotation_angle rotation_angle;
+	bool mirrored;
+	enum dml2_scaling_transform scaling_transform;
+	bool rect_out_height_spans_vactive;
+
+	struct {
+		bool stationary;
+		struct {
+			unsigned long width;
+			unsigned long height;
+			unsigned long x_start;
+			unsigned long y_start;
+		} plane0;
+
+		struct {
+			unsigned long width;
+			unsigned long height;
+			unsigned long x_start;
+			unsigned long y_start;
+		} plane1;
+	} viewport;
+
+	struct {
+		bool enabled;
+		bool easf_enabled;
+		bool isharp_enabled;
+		bool upsp_enabled;
+		enum dml2_sample_positioning upsp_sample_positioning;
+		unsigned int upsp_vtaps;
+		struct {
+			double h_ratio;
+			double v_ratio;
+			unsigned int h_taps;
+			unsigned int v_taps;
+		} plane0;
+
+		struct {
+			double h_ratio;
+			double v_ratio;
+			unsigned int h_taps;
+			unsigned int v_taps;
+		} plane1;
+
+		unsigned long rect_out_width;
+	} scaler_info;
+};
+
+struct dml2_timing_cfg {
+	unsigned long h_total;
+	unsigned long v_total;
+	unsigned long h_blank_end;
+	unsigned long v_blank_end;
+	unsigned long h_front_porch;
+	unsigned long v_front_porch;
+	unsigned long h_sync_width;
+	unsigned long pixel_clock_khz;
+	unsigned long h_active;
+	unsigned long v_active;
+	unsigned int bpc; //FIXME: review with Jun
+	struct {
+		enum dml2_dsc_enable_option enable;
+		unsigned int dsc_compressed_bpp_x16;
+		struct {
+			// for dv to specify num dsc slices to use
+			unsigned int num_slices;
+		} overrides;
+	} dsc;
+	bool interlaced;
+	struct {
+		/* static */
+		bool enabled;
+		unsigned long min_refresh_uhz;
+		unsigned int max_instant_vtotal_delta;
+		/* dynamic */
+		bool disallowed;
+		bool drr_active_variable;
+		bool drr_active_fixed;
+	} drr_config;
+	unsigned long vblank_nom;
+};
+
+struct dml2_link_output_cfg {
+	enum dml2_output_format_class output_format;
+	enum dml2_output_encoder_class output_encoder;
+	unsigned int output_dp_lane_count;
+	enum dml2_output_link_dp_rate output_dp_link_rate;
+	unsigned long audio_sample_rate;
+	unsigned long audio_sample_layout;
+	bool output_disabled; // The stream does not go to a backend for output to a physical
+						  //connector (e.g. writeback only, phantom pipe) goes to writeback
+	bool validate_output; // Do not validate the link configuration for this display stream.
+};
+
+struct dml2_writeback_info {
+	enum dml2_source_format_class pixel_format;
+	unsigned long input_width;
+	unsigned long input_height;
+	unsigned long output_width;
+	unsigned long output_height;
+	unsigned long v_taps;
+	unsigned long h_taps;
+	unsigned long v_taps_chroma;
+	unsigned long h_taps_chroma;
+	double h_ratio;
+	double v_ratio;
+};
+
+struct dml2_writeback_cfg {
+	unsigned int active_writebacks_per_stream;
+	struct dml2_writeback_info writeback_stream[DML2_MAX_WRITEBACK];
+};
+
+struct dml2_plane_parameters {
+	unsigned int stream_index; // Identifies which plane will be composed
+
+	enum dml2_source_format_class pixel_format;
+	/*
+	 * The surface and composition structures use
+	 * the terms plane0 and plane1.  These planes
+	 * are expected to hold the following data based
+	 * on the pixel format.
+	 *
+	 * RGB or YUV Non-Planar Types:
+	 *  dml2_444_8
+	 *	dml2_444_16
+	 *	dml2_444_32
+	 *	dml2_444_64
+	 *	dml2_rgbe
+	 *
+	 * plane0 = argb or rgbe
+	 * plane1 = not used
+	 *
+	 * YUV Planar-Types:
+	 *	dml2_420_8
+	 *	dml2_420_10
+	 *	dml2_420_12
+	 *
+	 * plane0 = luma
+	 * plane1 = chroma
+	 *
+	 * RGB Planar Types:
+	 *	dml2_rgbe_alpha
+	 *
+	 * plane0 = rgbe
+	 * plane1 = alpha
+	 *
+	 * Mono Non-Planar Types:
+	 *	dml2_mono_8
+	 *	dml2_mono_16
+	 *
+	 * plane0 = luma
+	 * plane1 = not used
+	 */
+
+	struct dml2_surface_cfg surface;
+	struct dml2_composition_cfg composition;
+
+	struct {
+		bool enable;
+		unsigned long lines_before_active_required;
+		unsigned long transmitted_bytes;
+	} dynamic_meta_data;
+
+	struct {
+		unsigned int num_cursors;
+		unsigned long cursor_width;
+		unsigned long cursor_bpp;
+	} cursor;
+
+	// For TDLUT, SW would assume TDLUT is setup and enable all the time and
+	// budget for worst case addressing/width mode
+	struct {
+		bool setup_for_tdlut;
+		enum dml2_tdlut_addressing_mode tdlut_addressing_mode;
+		enum dml2_tdlut_width_mode tdlut_width_mode;
+		bool tdlut_mpc_width_flag;
+	} tdlut;
+
+	bool immediate_flip;
+
+	struct {
+		// Logical overrides to power management policies (usually)
+		enum dml2_uclk_pstate_change_strategy uclk_pstate_change_strategy;
+		enum dml2_refresh_from_mall_mode_override refresh_from_mall;
+		unsigned int det_size_override_kb;
+		unsigned int mpcc_combine_factor;
+
+		// reserved_vblank_time_ns is the minimum time to reserve in vblank for Twait
+		// The actual reserved vblank time used for the corresponding stream in mode_programming would be at least as much as this per-plane override.
+		long reserved_vblank_time_ns;
+		unsigned int max_vactive_det_fill_delay_us[dml2_pstate_type_count]; // 0 = no reserved time, +ve = explicit max delay
+		unsigned int gpuvm_min_page_size_kbytes;
+		unsigned int hostvm_min_page_size_kbytes;
+
+		enum dml2_svp_mode_override legacy_svp_config; //TODO remove in favor of svp_config
+
+		struct {
+			// HW specific overrides, there's almost no reason to mess with these
+			// generally used for debugging or simulation
+			bool force_one_row_for_frame;
+			struct {
+				bool enable;
+				bool value;
+			} force_pte_buffer_mode;
+			double dppclk_mhz;
+		} hw;
+	} overrides;
+};
+
+struct dml2_stream_parameters {
+	struct dml2_timing_cfg timing;
+	struct dml2_link_output_cfg output;
+	struct dml2_writeback_cfg writeback;
+
+	struct {
+		enum dml2_odm_mode odm_mode;
+		bool disable_dynamic_odm;
+		bool disable_subvp;
+		int minimum_vblank_idle_requirement_us;
+
+		struct {
+			struct {
+				enum dml2_twait_budgeting_setting uclk_pstate;
+				enum dml2_twait_budgeting_setting fclk_pstate;
+				enum dml2_twait_budgeting_setting stutter_enter_exit;
+			} twait_budgeting;
+		} hw;
+	} overrides;
+};
+
+struct dml2_display_cfg {
+	bool gpuvm_enable;
+	bool ffbm_enable;
+	bool hostvm_enable;
+
+	// Allocate DET proportionally between streams based on pixel rate
+	// and then allocate proportionally between planes.
+	bool minimize_det_reallocation;
+
+	unsigned int gpuvm_max_page_table_levels;
+	unsigned int hostvm_max_non_cached_page_table_levels;
+
+	struct dml2_plane_parameters plane_descriptors[DML2_MAX_PLANES];
+	struct dml2_stream_parameters stream_descriptors[DML2_MAX_PLANES];
+
+	unsigned int num_planes;
+	unsigned int num_streams;
+
+	struct {
+		struct {
+			// HW specific overrides, there's almost no reason to mess with these
+			// generally used for debugging or simulation
+			struct {
+				bool enable;
+				bool value;
+			} force_unbounded_requesting;
+
+			struct {
+				bool enable;
+				bool value;
+			} force_nom_det_size_kbytes;
+
+			bool mode_support_check_disable;
+			bool mcache_admissibility_check_disable;
+			bool surface_viewport_size_check_disable;
+			double dlg_ref_clk_mhz;
+			double dispclk_mhz;
+			double dcfclk_mhz;
+			bool optimize_tdlut_scheduling; // TBD: for DV, will set this to 1, to ensure tdlut schedule is calculated based on address/width mode
+		} hw;
+
+		struct {
+			bool uclk_pstate_change_disable;
+			bool fclk_pstate_change_disable;
+			bool g6_temp_read_pstate_disable;
+			bool g7_ppt_pstate_disable;
+		} power_management;
+
+		bool enhanced_prefetch_schedule_acceleration;
+		bool dcc_programming_assumes_scan_direction_unknown;
+		bool synchronize_timings;
+		bool synchronize_ddr_displays_for_uclk_pstate_change;
+		bool max_outstanding_when_urgent_expected_disable;
+		bool enable_subvp_implicit_pmo; //enables PMO to switch pipe uclk strategy to subvp, and generate phantom programming
+		bool all_streams_blanked;
+	} overrides;
+};
+
+struct dml2_pipe_configuration_descriptor {
+	struct {
+		unsigned int viewport_x_start;
+		unsigned int viewport_width;
+	} plane0;
+
+	struct {
+		unsigned int viewport_x_start;
+		unsigned int viewport_width;
+	} plane1;
+
+	bool plane1_enabled;
+	bool imall_enabled;
+};
+
+struct dml2_plane_mcache_configuration_descriptor {
+	const struct dml2_plane_parameters *plane_descriptor;
+	const struct dml2_mcache_surface_allocation *mcache_allocation;
+
+	struct dml2_pipe_configuration_descriptor pipe_configurations[DML2_MAX_DCN_PIPES];
+	char num_pipes;
+};
+
+#endif
diff --git a/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/inc/dml_top_policy_types.h b/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/inc/dml_top_policy_types.h
new file mode 100644
index 000000000000..8f624a912e78
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/inc/dml_top_policy_types.h
@@ -0,0 +1,13 @@
+// SPDX-License-Identifier: MIT
+//
+// Copyright 2024 Advanced Micro Devices, Inc.
+
+#ifndef __DML_TOP_POLICY_TYPES_H__
+#define __DML_TOP_POLICY_TYPES_H__
+
+struct dml2_policy_parameters {
+	unsigned long odm_combine_dispclk_threshold_khz;
+	unsigned int max_immediate_flip_latency;
+};
+
+#endif
diff --git a/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/inc/dml_top_soc_parameter_types.h b/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/inc/dml_top_soc_parameter_types.h
new file mode 100644
index 000000000000..1fbc520c2540
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/inc/dml_top_soc_parameter_types.h
@@ -0,0 +1,215 @@
+// SPDX-License-Identifier: MIT
+//
+// Copyright 2024 Advanced Micro Devices, Inc.
+
+#ifndef __DML_TOP_SOC_PARAMETER_TYPES_H__
+#define __DML_TOP_SOC_PARAMETER_TYPES_H__
+
+#include "dml2_external_lib_deps.h"
+
+#define DML_MAX_CLK_TABLE_SIZE 20
+
+struct dml2_soc_derate_values {
+	unsigned int dram_derate_percent_pixel;
+	unsigned int dram_derate_percent_vm;
+	unsigned int dram_derate_percent_pixel_and_vm;
+
+	unsigned int fclk_derate_percent;
+	unsigned int dcfclk_derate_percent;
+};
+
+struct dml2_soc_derates {
+	struct dml2_soc_derate_values system_active_urgent;
+	struct dml2_soc_derate_values system_active_average;
+	struct dml2_soc_derate_values dcn_mall_prefetch_urgent;
+	struct dml2_soc_derate_values dcn_mall_prefetch_average;
+	struct dml2_soc_derate_values system_idle_average;
+};
+
+struct dml2_dcn32x_soc_qos_params {
+	struct {
+		unsigned int base_latency_us;
+		unsigned int base_latency_pixel_vm_us;
+		unsigned int base_latency_vm_us;
+		unsigned int scaling_factor_fclk_us;
+		unsigned int scaling_factor_mhz;
+	} urgent_latency_us;
+
+	unsigned int loaded_round_trip_latency_fclk_cycles;
+	unsigned int urgent_out_of_order_return_per_channel_pixel_only_bytes;
+	unsigned int urgent_out_of_order_return_per_channel_pixel_and_vm_bytes;
+	unsigned int urgent_out_of_order_return_per_channel_vm_only_bytes;
+};
+
+struct dml2_dcn4_uclk_dpm_dependent_qos_params {
+	unsigned long minimum_uclk_khz;
+	unsigned int urgent_ramp_uclk_cycles;
+	unsigned int trip_to_memory_uclk_cycles;
+	unsigned int meta_trip_to_memory_uclk_cycles;
+	unsigned int maximum_latency_when_urgent_uclk_cycles;
+	unsigned int average_latency_when_urgent_uclk_cycles;
+	unsigned int maximum_latency_when_non_urgent_uclk_cycles;
+	unsigned int average_latency_when_non_urgent_uclk_cycles;
+};
+
+struct dml2_dcn4x_soc_qos_params {
+	unsigned int df_qos_response_time_fclk_cycles;
+	unsigned int max_round_trip_to_furthest_cs_fclk_cycles;
+	unsigned int mall_overhead_fclk_cycles;
+	unsigned int meta_trip_adder_fclk_cycles;
+	unsigned int average_transport_distance_fclk_cycles;
+	double umc_urgent_ramp_latency_margin;
+	double umc_max_latency_margin;
+	double umc_average_latency_margin;
+	double fabric_max_transport_latency_margin;
+	double fabric_average_transport_latency_margin;
+	struct dml2_dcn4_uclk_dpm_dependent_qos_params per_uclk_dpm_params[DML_MAX_CLK_TABLE_SIZE];
+};
+
+enum dml2_qos_param_type {
+	dml2_qos_param_type_dcn3,
+	dml2_qos_param_type_dcn4x
+};
+
+struct dml2_soc_qos_parameters {
+	struct dml2_soc_derates derate_table;
+	struct {
+		unsigned int base_latency_us;
+		unsigned int scaling_factor_us;
+		unsigned int scaling_factor_mhz;
+	} writeback;
+
+	union {
+		struct dml2_dcn32x_soc_qos_params dcn32x;
+		struct dml2_dcn4x_soc_qos_params dcn4x;
+	} qos_params;
+
+	enum dml2_qos_param_type qos_type;
+};
+
+struct dml2_soc_power_management_parameters {
+	double dram_clk_change_blackout_us;
+	double dram_clk_change_read_only_us; // deprecated
+	double dram_clk_change_write_only_us; // deprecated
+	double fclk_change_blackout_us;
+	double g7_ppt_blackout_us;
+	double g7_temperature_read_blackout_us;
+	double stutter_enter_plus_exit_latency_us;
+	double stutter_exit_latency_us;
+	double low_power_stutter_enter_plus_exit_latency_us;
+	double low_power_stutter_exit_latency_us;
+	double z8_stutter_enter_plus_exit_latency_us;
+	double z8_stutter_exit_latency_us;
+	double z8_min_idle_time;
+	double g6_temp_read_blackout_us[DML_MAX_CLK_TABLE_SIZE];
+	double type_b_dram_clk_change_blackout_us;
+	double type_b_ppt_blackout_us;
+};
+
+struct dml2_clk_table {
+	unsigned long clk_values_khz[DML_MAX_CLK_TABLE_SIZE];
+	unsigned char num_clk_values;
+};
+
+struct dml2_dram_params {
+	unsigned int channel_width_bytes;
+	unsigned int channel_count;
+	unsigned int transactions_per_clock;
+};
+
+struct dml2_soc_state_table {
+	struct dml2_clk_table uclk;
+	struct dml2_clk_table fclk;
+	struct dml2_clk_table dcfclk;
+	struct dml2_clk_table dispclk;
+	struct dml2_clk_table dppclk;
+	struct dml2_clk_table dtbclk;
+	struct dml2_clk_table phyclk;
+	struct dml2_clk_table socclk;
+	struct dml2_clk_table dscclk;
+	struct dml2_clk_table phyclk_d18;
+	struct dml2_clk_table phyclk_d32;
+
+	struct dml2_dram_params dram_config;
+};
+
+struct dml2_soc_vmin_clock_limits {
+	unsigned long dispclk_khz;
+	unsigned long dcfclk_khz;
+};
+
+struct dml2_soc_bb {
+	struct dml2_soc_state_table clk_table;
+	struct dml2_soc_qos_parameters qos_parameters;
+	struct dml2_soc_power_management_parameters power_management_parameters;
+	struct dml2_soc_vmin_clock_limits vmin_limit;
+
+	double lower_bound_bandwidth_dchub;
+	double fraction_of_urgent_bandwidth_nominal_target;
+	double fraction_of_urgent_bandwidth_flip_target;
+	unsigned int dprefclk_mhz;
+	unsigned int xtalclk_mhz;
+	unsigned int pcie_refclk_mhz;
+	unsigned int dchub_refclk_mhz;
+	unsigned int mall_allocated_for_dcn_mbytes;
+	unsigned int max_outstanding_reqs;
+	unsigned long fabric_datapath_to_dcn_data_return_bytes;
+	unsigned long return_bus_width_bytes;
+	unsigned long hostvm_min_page_size_kbytes;
+	unsigned long gpuvm_min_page_size_kbytes;
+	double phy_downspread_percent;
+	double dcn_downspread_percent;
+	double dispclk_dppclk_vco_speed_mhz;
+	bool no_dfs;
+	bool do_urgent_latency_adjustment;
+	unsigned int mem_word_bytes;
+	unsigned int num_dcc_mcaches;
+	unsigned int mcache_size_bytes;
+	unsigned int mcache_line_size_bytes;
+	unsigned long max_fclk_for_uclk_dpm_khz;
+};
+
+struct dml2_ip_capabilities {
+	unsigned int pipe_count;
+	unsigned int otg_count;
+	unsigned int TDLUT_33cube_count;
+	unsigned int num_dsc;
+	unsigned int max_num_dp2p0_streams;
+	unsigned int max_num_hdmi_frl_outputs;
+	unsigned int max_num_dp2p0_outputs;
+	unsigned int max_num_wb;
+	unsigned int rob_buffer_size_kbytes;
+	unsigned int config_return_buffer_size_in_kbytes;
+	unsigned int config_return_buffer_segment_size_in_kbytes;
+	unsigned int meta_fifo_size_in_kentries;
+	unsigned int compressed_buffer_segment_size_in_kbytes;
+	unsigned int cursor_buffer_size;
+	unsigned int max_flip_time_us;
+	unsigned int max_flip_time_lines;
+	unsigned int hostvm_mode;
+	unsigned int subvp_drr_scheduling_margin_us;
+	unsigned int subvp_prefetch_end_to_mall_start_us;
+	unsigned int subvp_fw_processing_delay;
+	unsigned int max_vactive_det_fill_delay_us;
+	unsigned int ppt_max_allow_delay_us;
+	unsigned int temp_read_max_allow_delay_us;
+	unsigned int dummy_pstate_max_allow_delay_us;
+	/* FAMS2 delays */
+	struct {
+		unsigned int max_allow_delay_us;
+		unsigned int scheduling_delay_us;
+		unsigned int vertical_interrupt_ack_delay_us; // delay to acknowledge vline int
+		unsigned int allow_programming_delay_us; // time requires to program allow
+		unsigned int min_allow_width_us;
+		unsigned int subvp_df_throttle_delay_us;
+		unsigned int subvp_programming_delay_us;
+		unsigned int subvp_prefetch_to_mall_delay_us;
+		unsigned int drr_programming_delay_us;
+
+		unsigned int lock_timeout_us;
+		unsigned int recovery_timeout_us;
+		unsigned int flip_programming_delay_us;
+	} fams2;
+};
+
+#endif
diff --git a/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/inc/dml_top_types.h b/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/inc/dml_top_types.h
new file mode 100644
index 000000000000..452e4a2e72c0
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/inc/dml_top_types.h
@@ -0,0 +1,744 @@
+// SPDX-License-Identifier: MIT
+//
+// Copyright 2024 Advanced Micro Devices, Inc.
+
+#ifndef __DML_TOP_TYPES_H__
+#define __DML_TOP_TYPES_H__
+
+#include "dml_top_display_cfg_types.h"
+#include "dml_top_soc_parameter_types.h"
+#include "dml_top_policy_types.h"
+#include "dml_top_dchub_registers.h"
+
+#include "dmub_cmd.h"
+
+struct dml2_instance;
+
+enum dml2_project_id {
+	dml2_project_invalid = 0,
+	dml2_project_dcn4x_stage1,
+	dml2_project_dcn4x_stage2,
+	dml2_project_dcn4x_stage2_auto_drr_svp,
+};
+
+enum dml2_pstate_change_support {
+	dml2_pstate_change_vactive = 0,
+	dml2_pstate_change_vblank = 1,
+	dml2_pstate_change_vblank_and_vactive = 2,
+	dml2_pstate_change_drr = 3,
+	dml2_pstate_change_mall_svp = 4,
+	dml2_pstate_change_mall_full_frame = 6,
+	dml2_pstate_change_unsupported = 7
+};
+
+enum dml2_output_type_and_rate__type {
+	dml2_output_type_unknown = 0,
+	dml2_output_type_dp = 1,
+	dml2_output_type_edp = 2,
+	dml2_output_type_dp2p0 = 3,
+	dml2_output_type_hdmi = 4,
+	dml2_output_type_hdmifrl = 5
+};
+
+enum dml2_output_type_and_rate__rate {
+	dml2_output_rate_unknown = 0,
+	dml2_output_rate_dp_rate_hbr = 1,
+	dml2_output_rate_dp_rate_hbr2 = 2,
+	dml2_output_rate_dp_rate_hbr3 = 3,
+	dml2_output_rate_dp_rate_uhbr10 = 4,
+	dml2_output_rate_dp_rate_uhbr13p5 = 5,
+	dml2_output_rate_dp_rate_uhbr20 = 6,
+	dml2_output_rate_hdmi_rate_3x3 = 7,
+	dml2_output_rate_hdmi_rate_6x3 = 8,
+	dml2_output_rate_hdmi_rate_6x4 = 9,
+	dml2_output_rate_hdmi_rate_8x4 = 10,
+	dml2_output_rate_hdmi_rate_10x4 = 11,
+	dml2_output_rate_hdmi_rate_12x4 = 12,
+	dml2_output_rate_hdmi_rate_16x4 = 13,
+	dml2_output_rate_hdmi_rate_20x4 = 14
+};
+
+struct dml2_pmo_options {
+	bool disable_vblank;
+	bool disable_svp;
+	bool disable_drr_var;
+	bool disable_drr_clamped;
+	bool disable_drr_var_when_var_active;
+	bool disable_drr_clamped_when_var_active;
+	bool disable_fams2;
+	bool disable_vactive_det_fill_bw_pad; /* dml2_project_dcn4x_stage2_auto_drr_svp and above only */
+	bool disable_dyn_odm;
+	bool disable_dyn_odm_for_multi_stream;
+	bool disable_dyn_odm_for_stream_with_svp;
+	struct dml2_pmo_pstate_strategy *override_strategy_lists[DML2_MAX_PLANES];
+	unsigned int num_override_strategies_per_list[DML2_MAX_PLANES];
+};
+
+struct dml2_options {
+	enum dml2_project_id project_id;
+	struct dml2_pmo_options pmo_options;
+};
+
+struct dml2_initialize_instance_in_out {
+	struct dml2_instance *dml2_instance;
+	struct dml2_options options;
+	struct dml2_soc_bb soc_bb;
+	struct dml2_ip_capabilities ip_caps;
+
+	struct {
+		void *explicit_ip_bb;
+		unsigned int explicit_ip_bb_size;
+	} overrides;
+};
+
+struct dml2_reset_instance_in_out {
+	struct dml2_instance *dml2_instance;
+};
+
+struct dml2_check_mode_supported_in_out {
+	/*
+	* Inputs
+	*/
+	struct dml2_instance *dml2_instance;
+	const struct dml2_display_cfg *display_config;
+
+	/*
+	* Outputs
+	*/
+	bool is_supported;
+};
+
+struct dml2_mcache_surface_allocation {
+	bool valid;
+	/*
+	* For iMALL, dedicated mall mcaches are required (sharing of last
+	* slice possible), for legacy phantom or phantom without return
+	* the only mall mcaches need to be valid.
+	*/
+	bool requires_dedicated_mall_mcache;
+
+	unsigned int num_mcaches_plane0;
+	unsigned int num_mcaches_plane1;
+	/*
+	* A plane is divided into vertical slices of mcaches,
+	* which wrap on the surface width.
+	*
+	* For example, if the surface width is 7680, and split into
+	* three slices of equal width, the boundary array would contain
+	* [2560, 5120, 7680]
+	*
+	* The assignments are
+	* 0 = [0 .. 2559]
+	* 1 = [2560 .. 5119]
+	* 2 = [5120 .. 7679]
+	* 0 = [7680 .. INF]
+	* The final element implicitly is the same as the first, and
+	* at first seems invalid since it is never referenced (since)
+	* it is outside the surface. However, its useful when shifting
+	* (see below).
+	*
+	* For any given valid mcache assignment, a shifted version, wrapped
+	* on the surface width boundary is also assumed to be valid.
+	*
+	* For example, shifting [2560, 5120, 7680] by -50 results in
+	* [2510, 5170, 7630].
+	*
+	* The assignments are now:
+	* 0 = [0 .. 2509]
+	* 1 = [2510 .. 5169]
+	* 2 = [5170 .. 7629]
+	* 0 = [7630 .. INF]
+	*/
+	int mcache_x_offsets_plane0[DML2_MAX_MCACHES + 1];
+	int mcache_x_offsets_plane1[DML2_MAX_MCACHES + 1];
+
+	/*
+	* Shift grainularity is not necessarily 1
+	*/
+	struct {
+		int p0;
+		int p1;
+	} shift_granularity;
+
+	/*
+	* MCacheIDs have global scope in the SoC, and they are stored here.
+	* These IDs are generally not valid until all planes in a display
+	* configuration have had their mcache requirements calculated.
+	*/
+	int global_mcache_ids_plane0[DML2_MAX_MCACHES + 1];
+	int global_mcache_ids_plane1[DML2_MAX_MCACHES + 1];
+	int global_mcache_ids_mall_plane0[DML2_MAX_MCACHES + 1];
+	int global_mcache_ids_mall_plane1[DML2_MAX_MCACHES + 1];
+
+	/*
+	* Generally, plane0/1 slices must use a disjoint set of caches
+	* but in some cases the final segement of the two planes can
+	* use the same cache. If plane0_plane1 is set, then this is
+	* allowed.
+	*
+	* Similarly, the caches allocated to MALL prefetcher are generally
+	* disjoint, but if mall_prefetch is set, then the final segment
+	* between the main and the mall pixel requestor can use the same
+	* cache.
+	*
+	* Note that both bits may be set at the same time.
+	*/
+	struct {
+		bool mall_comb_mcache_p0;
+		bool mall_comb_mcache_p1;
+		bool plane0_plane1;
+	} last_slice_sharing;
+
+	struct {
+		int meta_row_bytes_plane0;
+		int meta_row_bytes_plane1;
+	} informative;
+};
+
+enum dml2_pstate_method {
+	dml2_pstate_method_na = 0,
+	/* hw exclusive modes */
+	dml2_pstate_method_vactive = 1,
+	dml2_pstate_method_vblank = 2,
+	dml2_pstate_method_reserved_hw = 5,
+	/* fw assisted exclusive modes */
+	dml2_pstate_method_fw_svp = 6,
+	dml2_pstate_method_reserved_fw = 10,
+	/* fw assisted modes requiring drr modulation */
+	dml2_pstate_method_fw_vactive_drr = 11,
+	dml2_pstate_method_fw_vblank_drr = 12,
+	dml2_pstate_method_fw_svp_drr = 13,
+	dml2_pstate_method_reserved_fw_drr_clamped = 20,
+	dml2_pstate_method_fw_drr = 21,
+	dml2_pstate_method_reserved_fw_drr_var = 22,
+	dml2_pstate_method_count
+};
+
+struct dml2_per_plane_programming {
+	const struct dml2_plane_parameters *plane_descriptor;
+
+	union {
+		struct {
+			unsigned long dppclk_khz;
+		} dcn4x;
+	} min_clocks;
+
+	struct dml2_mcache_surface_allocation mcache_allocation;
+
+	// If a stream is using automatic or forced odm combine
+	// and the stream for this plane has num_odms_required > 1
+	// num_dpps_required is always equal to num_odms_required for
+	// ALL planes of the stream
+
+	// If a stream is using odm split, then this value is always 1
+	unsigned int num_dpps_required;
+
+	enum dml2_pstate_method uclk_pstate_support_method;
+
+	// MALL size requirements for MALL SS and SubVP
+	unsigned int surface_size_mall_bytes;
+	unsigned int svp_size_mall_bytes;
+
+	struct dml2_dchub_per_pipe_register_set *pipe_regs[DML2_MAX_PLANES];
+
+	struct {
+		bool valid;
+		struct dml2_plane_parameters descriptor;
+		struct dml2_mcache_surface_allocation mcache_allocation;
+		struct dml2_dchub_per_pipe_register_set *pipe_regs[DML2_MAX_PLANES];
+	} phantom_plane;
+};
+
+union dml2_global_sync_programming {
+	struct {
+		unsigned int vstartup_lines;
+		unsigned int vupdate_offset_pixels;
+		unsigned int vupdate_vupdate_width_pixels;
+		unsigned int vready_offset_pixels;
+		unsigned int pstate_keepout_start_lines;
+	} dcn4x;
+};
+
+struct dml2_per_stream_programming {
+	const struct dml2_stream_parameters *stream_descriptor;
+
+	union {
+		struct {
+			unsigned long dscclk_khz;
+			unsigned long dtbclk_khz;
+			unsigned long phyclk_khz;
+		} dcn4x;
+	} min_clocks;
+
+	union dml2_global_sync_programming global_sync;
+
+	unsigned int num_odms_required;
+
+	enum dml2_pstate_method uclk_pstate_method;
+
+	struct {
+		bool enabled;
+		struct dml2_stream_parameters descriptor;
+		union dml2_global_sync_programming global_sync;
+	} phantom_stream;
+
+	union dmub_cmd_fams2_config fams2_base_params;
+	union {
+		union dmub_cmd_fams2_config fams2_sub_params;
+		union dmub_fams2_stream_static_sub_state_v2 fams2_sub_params_v2;
+	};
+};
+
+//-----------------
+// Mode Support Information
+//-----------------
+
+struct dml2_mode_support_info {
+	bool ModeIsSupported; //<brief Is the mode support any voltage and combine setting
+	bool ImmediateFlipSupport; //<brief Means mode support immediate flip at the max combine setting; determine in mode support and used in mode programming
+	// Mode Support Reason
+	bool WritebackLatencySupport;
+	bool ScaleRatioAndTapsSupport;
+	bool SourceFormatPixelAndScanSupport;
+	bool P2IWith420;
+	bool DSCOnlyIfNecessaryWithBPP;
+	bool DSC422NativeNotSupported;
+	bool LinkRateDoesNotMatchDPVersion;
+	bool LinkRateForMultistreamNotIndicated;
+	bool BPPForMultistreamNotIndicated;
+	bool MultistreamWithHDMIOreDP;
+	bool MSOOrODMSplitWithNonDPLink;
+	bool NotEnoughLanesForMSO;
+	bool NumberOfOTGSupport;
+	bool NumberOfHDMIFRLSupport;
+	bool NumberOfDP2p0Support;
+	bool NumberOfTDLUT33cubeSupport;
+	bool WritebackScaleRatioAndTapsSupport;
+	bool CursorSupport;
+	bool PitchSupport;
+	bool ViewportExceedsSurface;
+	bool ImmediateFlipRequiredButTheRequirementForEachSurfaceIsNotSpecified;
+	bool ImmediateFlipOrHostVMAndPStateWithMALLFullFrameOrPhantomPipe;
+	bool InvalidCombinationOfMALLUseForPStateAndStaticScreen;
+	bool InvalidCombinationOfMALLUseForPState;
+	bool ExceededMALLSize;
+	bool EnoughWritebackUnits;
+	bool ExceededMultistreamSlots;
+	bool NotEnoughDSCUnits;
+	bool NotEnoughDSCSlices;
+	bool PixelsPerLinePerDSCUnitSupport;
+	bool DSCCLKRequiredMoreThanSupported;
+	bool DTBCLKRequiredMoreThanSupported;
+	bool LinkCapacitySupport;
+	bool ROBSupport;
+	bool OutstandingRequestsSupport;
+	bool OutstandingRequestsUrgencyAvoidance;
+	bool PTEBufferSizeNotExceeded;
+	bool DCCMetaBufferSizeNotExceeded;
+	bool TotalVerticalActiveBandwidthSupport;
+	bool VActiveBandwidthSupport;
+	enum dml2_pstate_change_support FCLKChangeSupport[DML2_MAX_PLANES];
+	bool USRRetrainingSupport;
+	bool PrefetchSupported;
+	bool DynamicMetadataSupported;
+	bool VRatioInPrefetchSupported;
+	bool DISPCLK_DPPCLK_Support;
+	bool TotalAvailablePipesSupport;
+	bool ViewportSizeSupport;
+	bool ImmediateFlipSupportedForState;
+	double MaxTotalVerticalActiveAvailableBandwidth;
+	bool MPCCombineEnable[DML2_MAX_PLANES]; /// <brief Indicate if the MPC Combine enable in the given state and optimize mpc combine setting
+	enum dml2_odm_mode ODMMode[DML2_MAX_PLANES]; /// <brief ODM mode that is chosen in the mode check stage and will be used in mode programming stage
+	unsigned int DPPPerSurface[DML2_MAX_PLANES]; /// <brief How many DPPs are needed drive the surface to output. If MPCC or ODMC could be 2 or 4.
+	bool DSCEnabled[DML2_MAX_PLANES]; /// <brief Indicate if the DSC is actually required; used in mode_programming
+	bool FECEnabled[DML2_MAX_PLANES]; /// <brief Indicate if the FEC is actually required
+	unsigned int NumberOfDSCSlices[DML2_MAX_PLANES]; /// <brief Indicate how many slices needed to support the given mode
+	double OutputBpp[DML2_MAX_PLANES];
+	enum dml2_output_type_and_rate__type OutputType[DML2_MAX_PLANES];
+	enum dml2_output_type_and_rate__rate OutputRate[DML2_MAX_PLANES];
+	unsigned int AlignedYPitch[DML2_MAX_PLANES];
+	unsigned int AlignedCPitch[DML2_MAX_PLANES];
+	bool g6_temp_read_support;
+	bool temp_read_or_ppt_support;
+	bool qos_bandwidth_support;
+	bool dcfclk_support;
+}; // dml2_mode_support_info
+
+struct dml2_display_cfg_programming {
+	struct dml2_display_cfg display_config;
+
+	union {
+		struct {
+			unsigned long dcfclk_khz;
+			unsigned long fclk_khz;
+			unsigned long uclk_khz;
+			unsigned long socclk_khz;
+			unsigned long dispclk_khz;
+			unsigned long dcfclk_deepsleep_khz;
+			unsigned long dpp_ref_khz;
+		} dcn32x;
+		struct {
+			struct {
+				unsigned long uclk_khz;
+				unsigned long fclk_khz;
+				unsigned long dcfclk_khz;
+			} active;
+			struct {
+				unsigned long uclk_khz;
+				unsigned long fclk_khz;
+				unsigned long dcfclk_khz;
+			} idle;
+			struct {
+				unsigned long uclk_khz;
+				unsigned long fclk_khz;
+				unsigned long dcfclk_khz;
+			} svp_prefetch;
+			struct {
+				unsigned long uclk_khz;
+				unsigned long fclk_khz;
+				unsigned long dcfclk_khz;
+			} svp_prefetch_no_throttle;
+
+			unsigned long deepsleep_dcfclk_khz;
+			unsigned long dispclk_khz;
+			unsigned long dpprefclk_khz;
+			unsigned long dtbrefclk_khz;
+			unsigned long socclk_khz;
+
+			struct {
+				uint32_t dispclk_did;
+				uint32_t dpprefclk_did;
+				uint32_t dtbrefclk_did;
+			} divider_ids;
+		} dcn4x;
+	} min_clocks;
+
+	bool uclk_pstate_supported;
+	bool fclk_pstate_supported;
+
+	/* indicates this configuration requires FW to support */
+	bool fams2_required;
+	struct dmub_cmd_fams2_global_config fams2_global_config;
+
+	struct {
+		bool supported_in_blank; // Changing to configurations where this is false requires stutter to be disabled during the transition
+		uint8_t base_percent_efficiency; //LP1
+		uint8_t low_power_percent_efficiency; //LP2
+	} stutter;
+
+	struct {
+		bool meets_eco; // Stutter cycles will meet Z8 ECO criteria
+		bool supported_in_blank; // Changing to configurations where this is false requires Z8 to be disabled during the transition
+	} z8_stutter;
+
+	struct dml2_dchub_global_register_set global_regs;
+
+	struct dml2_per_plane_programming plane_programming[DML2_MAX_PLANES];
+	struct dml2_per_stream_programming stream_programming[DML2_MAX_PLANES];
+
+	// Don't access this structure directly, access it through plane_programming.pipe_regs
+	struct dml2_dchub_per_pipe_register_set pipe_regs[DML2_MAX_PLANES];
+
+	struct {
+		struct {
+			double urgent_us;
+			double writeback_urgent_us;
+			double writeback_pstate_us;
+			double writeback_fclk_pstate_us;
+			double cstate_exit_us;
+			double cstate_enter_plus_exit_us;
+			double z8_cstate_exit_us;
+			double z8_cstate_enter_plus_exit_us;
+			double pstate_change_us;
+			double fclk_pstate_change_us;
+			double usr_retraining_us;
+			double temp_read_or_ppt_watermark_us;
+		} watermarks;
+
+		struct {
+			unsigned int swath_width_plane0;
+			unsigned int swath_height_plane0;
+			unsigned int swath_height_plane1;
+			unsigned int dpte_row_height_plane0;
+			unsigned int dpte_row_height_plane1;
+			unsigned int meta_row_height_plane0;
+			unsigned int meta_row_height_plane1;
+		} plane_info[DML2_MAX_PLANES];
+
+		struct {
+			unsigned int total_num_dpps_required;
+		} dpp;
+
+		struct {
+			unsigned long long total_surface_size_in_mall_bytes;
+			unsigned int subviewport_lines_needed_in_mall[DML2_MAX_PLANES];
+		} mall;
+
+		struct {
+			double urgent_latency_us; // urgent ramp latency
+			double max_non_urgent_latency_us;
+			double max_urgent_latency_us;
+			double avg_non_urgent_latency_us;
+			double avg_urgent_latency_us;
+			double wm_memory_trip_us;
+			double meta_trip_memory_us;
+			double fraction_of_urgent_bandwidth; // nom
+			double fraction_of_urgent_bandwidth_immediate_flip;
+			double fraction_of_urgent_bandwidth_mall;
+			double max_active_fclk_change_latency_supported;
+			unsigned int min_return_latency_in_dcfclk;
+
+			struct {
+				struct {
+					double sdp_bw_mbps;
+					double dram_bw_mbps;
+					double dram_vm_only_bw_mbps;
+				} svp_prefetch;
+
+				struct {
+					double sdp_bw_mbps;
+					double dram_bw_mbps;
+					double dram_vm_only_bw_mbps;
+				} sys_active;
+			} urg_bw_available;
+
+			struct {
+				struct {
+					double sdp_bw_mbps;
+					double dram_bw_mbps;
+				} svp_prefetch;
+
+				struct {
+					double sdp_bw_mbps;
+					double dram_bw_mbps;
+				} sys_active;
+			} avg_bw_available;
+
+			struct {
+				struct {
+					double sdp_bw_mbps;
+					double dram_bw_mbps;
+				} svp_prefetch;
+
+				struct {
+					double sdp_bw_mbps;
+					double dram_bw_mbps;
+				} sys_active;
+			} non_urg_bw_required;
+
+			struct {
+				struct {
+					double sdp_bw_mbps;
+					double dram_bw_mbps;
+				} svp_prefetch;
+
+				struct {
+					double sdp_bw_mbps;
+					double dram_bw_mbps;
+				} sys_active;
+			} non_urg_bw_required_with_flip;
+
+			struct {
+				struct {
+					double sdp_bw_mbps;
+					double dram_bw_mbps;
+				} svp_prefetch;
+
+				struct {
+					double sdp_bw_mbps;
+					double dram_bw_mbps;
+				} sys_active;
+
+			} urg_bw_required;
+
+			struct {
+				struct {
+					double sdp_bw_mbps;
+					double dram_bw_mbps;
+				} svp_prefetch;
+
+				struct {
+					double sdp_bw_mbps;
+					double dram_bw_mbps;
+				} sys_active;
+			} urg_bw_required_with_flip;
+
+			struct {
+				struct {
+					double sdp_bw_mbps;
+					double dram_bw_mbps;
+				} svp_prefetch;
+
+				struct {
+					double sdp_bw_mbps;
+					double dram_bw_mbps;
+				} sys_active;
+			} avg_bw_required;
+		} qos;
+
+		struct {
+			unsigned long long det_size_in_kbytes[DML2_MAX_PLANES];
+			unsigned long long DETBufferSizeY[DML2_MAX_PLANES];
+			unsigned long long comp_buffer_size_kbytes;
+			bool UnboundedRequestEnabled;
+			unsigned int compbuf_reserved_space_64b;
+		} crb;
+
+		struct {
+			unsigned int max_uncompressed_block_plane0;
+			unsigned int max_compressed_block_plane0;
+			unsigned int independent_block_plane0;
+			unsigned int max_uncompressed_block_plane1;
+			unsigned int max_compressed_block_plane1;
+			unsigned int independent_block_plane1;
+		} dcc_control[DML2_MAX_PLANES];
+
+		struct {
+			double stutter_efficiency;
+			double stutter_efficiency_with_vblank;
+			double stutter_num_bursts;
+
+			struct {
+				double stutter_efficiency;
+				double stutter_efficiency_with_vblank;
+				double stutter_num_bursts;
+				double stutter_period;
+
+				struct {
+					double stutter_efficiency;
+					double stutter_num_bursts;
+					double stutter_period;
+				} bestcase;
+			} z8;
+		} power_management;
+
+		struct {
+			double min_ttu_vblank_us[DML2_MAX_PLANES];
+			bool vready_at_or_after_vsync[DML2_MAX_PLANES];
+			double min_dst_y_next_start[DML2_MAX_PLANES];
+			bool cstate_max_cap_mode;
+			bool hw_debug5;
+			unsigned int dcfclk_deep_sleep_hysteresis;
+			unsigned int dst_x_after_scaler[DML2_MAX_PLANES];
+			unsigned int dst_y_after_scaler[DML2_MAX_PLANES];
+			unsigned int prefetch_source_lines_plane0[DML2_MAX_PLANES];
+			unsigned int prefetch_source_lines_plane1[DML2_MAX_PLANES];
+			bool ImmediateFlipSupportedForPipe[DML2_MAX_PLANES];
+			bool UsesMALLForStaticScreen[DML2_MAX_PLANES];
+			unsigned int CursorDstXOffset[DML2_MAX_PLANES];
+			unsigned int CursorDstYOffset[DML2_MAX_PLANES];
+			unsigned int CursorChunkHDLAdjust[DML2_MAX_PLANES];
+			unsigned int dpte_group_bytes[DML2_MAX_PLANES];
+			unsigned int vm_group_bytes[DML2_MAX_PLANES];
+			double DisplayPipeRequestDeliveryTimeLuma[DML2_MAX_PLANES];
+			double DisplayPipeRequestDeliveryTimeChroma[DML2_MAX_PLANES];
+			double DisplayPipeRequestDeliveryTimeLumaPrefetch[DML2_MAX_PLANES];
+			double DisplayPipeRequestDeliveryTimeChromaPrefetch[DML2_MAX_PLANES];
+			double TimePerVMGroupVBlank[DML2_MAX_PLANES];
+			double TimePerVMGroupFlip[DML2_MAX_PLANES];
+			double TimePerVMRequestVBlank[DML2_MAX_PLANES];
+			double TimePerVMRequestFlip[DML2_MAX_PLANES];
+			double Tdmdl_vm[DML2_MAX_PLANES];
+			double Tdmdl[DML2_MAX_PLANES];
+			unsigned int VStartup[DML2_MAX_PLANES];
+			unsigned int VUpdateOffsetPix[DML2_MAX_PLANES];
+			unsigned int VUpdateWidthPix[DML2_MAX_PLANES];
+			unsigned int VReadyOffsetPix[DML2_MAX_PLANES];
+
+			double DST_Y_PER_PTE_ROW_NOM_L[DML2_MAX_PLANES];
+			double DST_Y_PER_PTE_ROW_NOM_C[DML2_MAX_PLANES];
+			double time_per_pte_group_nom_luma[DML2_MAX_PLANES];
+			double time_per_pte_group_nom_chroma[DML2_MAX_PLANES];
+			double time_per_pte_group_vblank_luma[DML2_MAX_PLANES];
+			double time_per_pte_group_vblank_chroma[DML2_MAX_PLANES];
+			double time_per_pte_group_flip_luma[DML2_MAX_PLANES];
+			double time_per_pte_group_flip_chroma[DML2_MAX_PLANES];
+			double VRatioPrefetchY[DML2_MAX_PLANES];
+			double VRatioPrefetchC[DML2_MAX_PLANES];
+			double DestinationLinesForPrefetch[DML2_MAX_PLANES];
+			double DestinationLinesToRequestVMInVBlank[DML2_MAX_PLANES];
+			double DestinationLinesToRequestRowInVBlank[DML2_MAX_PLANES];
+			double DestinationLinesToRequestVMInImmediateFlip[DML2_MAX_PLANES];
+			double DestinationLinesToRequestRowInImmediateFlip[DML2_MAX_PLANES];
+			double DisplayPipeLineDeliveryTimeLuma[DML2_MAX_PLANES];
+			double DisplayPipeLineDeliveryTimeChroma[DML2_MAX_PLANES];
+			double DisplayPipeLineDeliveryTimeLumaPrefetch[DML2_MAX_PLANES];
+			double DisplayPipeLineDeliveryTimeChromaPrefetch[DML2_MAX_PLANES];
+
+			double WritebackRequiredBandwidth;
+			double WritebackAllowDRAMClockChangeEndPosition[DML2_MAX_PLANES];
+			double WritebackAllowFCLKChangeEndPosition[DML2_MAX_PLANES];
+			double DSCCLK_calculated[DML2_MAX_PLANES];
+			unsigned int BIGK_FRAGMENT_SIZE[DML2_MAX_PLANES];
+			bool PTE_BUFFER_MODE[DML2_MAX_PLANES];
+			double DSCDelay[DML2_MAX_PLANES];
+			double MaxActiveDRAMClockChangeLatencySupported[DML2_MAX_PLANES];
+			unsigned int PrefetchMode[DML2_MAX_PLANES]; // LEGACY_ONLY
+			bool ROBUrgencyAvoidance;
+			double LowestPrefetchMargin;
+
+			unsigned int pstate_recout_reduction_lines[DML2_MAX_PLANES];
+		} misc;
+
+		struct dml2_mode_support_info mode_support_info;
+		unsigned int voltage_level; // LEGACY_ONLY
+
+		// For DV only
+		// This is what dml core calculated, only on the full_vp width and assume we have
+		// unlimited # of mcache
+		struct dml2_mcache_surface_allocation non_optimized_mcache_allocation[DML2_MAX_PLANES];
+
+		bool failed_prefetch;
+		bool failed_uclk_pstate;
+		bool failed_mcache_validation;
+		bool failed_dpmm;
+		bool failed_mode_programming;
+		bool failed_mode_programming_dcfclk;
+		bool failed_mode_programming_prefetch;
+		bool failed_mode_programming_flip;
+		bool failed_map_watermarks;
+	} informative;
+};
+
+struct dml2_build_mode_programming_in_out {
+	/*
+	* Inputs
+	*/
+	struct dml2_instance *dml2_instance;
+	const struct dml2_display_cfg *display_config;
+
+	/*
+	* Outputs
+	*/
+	struct dml2_display_cfg_programming *programming;
+};
+
+struct dml2_build_mcache_programming_in_out {
+	/*
+	* Inputs
+	*/
+	struct dml2_instance *dml2_instance;
+
+	struct dml2_plane_mcache_configuration_descriptor mcache_configurations[DML2_MAX_PLANES];
+	char num_configurations;
+
+	/*
+	* Outputs
+	*/
+	// per_plane_pipe_mcache_regs[i][j] refers to the proper programming for the j-th pipe of the
+	// i-th plane (from mcache_configurations)
+	struct dml2_hubp_pipe_mcache_regs *per_plane_pipe_mcache_regs[DML2_MAX_PLANES][DML2_MAX_DCN_PIPES];
+
+	// It's not a good idea to reference this directly, better to use the pointer structure above instead
+	struct dml2_hubp_pipe_mcache_regs mcache_regs_set[DML2_MAX_DCN_PIPES];
+};
+
+struct dml2_unit_test_in_out {
+	/*
+	* Inputs
+	*/
+	struct dml2_instance *dml2_instance;
+};
+
+
+#endif
diff --git a/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/src/dml2_core/dml2_core_dcn4.c b/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/src/dml2_core/dml2_core_dcn4.c
new file mode 100644
index 000000000000..eba948e187c1
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/src/dml2_core/dml2_core_dcn4.c
@@ -0,0 +1,661 @@
+// SPDX-License-Identifier: MIT
+//
+// Copyright 2024 Advanced Micro Devices, Inc.
+
+#include "dml2_internal_shared_types.h"
+#include "dml2_core_shared_types.h"
+#include "dml2_core_dcn4.h"
+#include "dml2_core_dcn4_calcs.h"
+#include "dml2_debug.h"
+#include "lib_float_math.h"
+
+struct dml2_core_ip_params core_dcn4_ip_caps_base = {
+	// Hardcoded values for DCN3x
+	.vblank_nom_default_us = 668,
+	.remote_iommu_outstanding_translations = 256,
+	.rob_buffer_size_kbytes = 128,
+	.config_return_buffer_size_in_kbytes = 1280,
+	.config_return_buffer_segment_size_in_kbytes = 64,
+	.compressed_buffer_segment_size_in_kbytes = 64,
+	.dpte_buffer_size_in_pte_reqs_luma = 68,
+	.dpte_buffer_size_in_pte_reqs_chroma = 36,
+	.pixel_chunk_size_kbytes = 8,
+	.alpha_pixel_chunk_size_kbytes = 4,
+	.min_pixel_chunk_size_bytes = 1024,
+	.writeback_chunk_size_kbytes = 8,
+	.line_buffer_size_bits = 1171920,
+	.max_line_buffer_lines = 32,
+	.writeback_interface_buffer_size_kbytes = 90,
+	//Number of pipes after DCN Pipe harvesting
+	.max_num_dpp = 4,
+	.max_num_opp = 4,
+	.max_num_otg = 4,
+	.max_num_wb = 1,
+	.max_dchub_pscl_bw_pix_per_clk = 4,
+	.max_pscl_lb_bw_pix_per_clk = 2,
+	.max_lb_vscl_bw_pix_per_clk = 4,
+	.max_vscl_hscl_bw_pix_per_clk = 4,
+	.max_hscl_ratio = 6,
+	.max_vscl_ratio = 6,
+	.max_hscl_taps = 8,
+	.max_vscl_taps = 8,
+	.dispclk_ramp_margin_percent = 1,
+	.dppclk_delay_subtotal = 47,
+	.dppclk_delay_scl = 50,
+	.dppclk_delay_scl_lb_only = 16,
+	.dppclk_delay_cnvc_formatter = 28,
+	.dppclk_delay_cnvc_cursor = 6,
+	.cursor_buffer_size = 24,
+	.cursor_chunk_size = 2,
+	.dispclk_delay_subtotal = 125,
+	.max_inter_dcn_tile_repeaters = 8,
+	.writeback_max_hscl_ratio = 1,
+	.writeback_max_vscl_ratio = 1,
+	.writeback_min_hscl_ratio = 1,
+	.writeback_min_vscl_ratio = 1,
+	.writeback_max_hscl_taps = 1,
+	.writeback_max_vscl_taps = 1,
+	.writeback_line_buffer_buffer_size = 0,
+	.num_dsc = 4,
+	.maximum_dsc_bits_per_component = 12,
+	.maximum_pixels_per_line_per_dsc_unit = 5760,
+	.dsc422_native_support = true,
+	.dcc_supported = true,
+	.ptoi_supported = false,
+
+	.cursor_64bpp_support = true,
+	.dynamic_metadata_vm_enabled = false,
+
+	.max_num_dp2p0_outputs = 4,
+	.max_num_dp2p0_streams = 4,
+	.imall_supported = 1,
+	.max_flip_time_us = 80,
+	.max_flip_time_lines = 32,
+	.words_per_channel = 16,
+
+	.subvp_fw_processing_delay_us = 15,
+	.subvp_pstate_allow_width_us = 20,
+	.subvp_swath_height_margin_lines = 16,
+};
+
+static void patch_ip_caps_with_explicit_ip_params(struct dml2_ip_capabilities *ip_caps, const struct dml2_core_ip_params *ip_params)
+{
+	ip_caps->pipe_count = ip_params->max_num_dpp;
+	ip_caps->otg_count = ip_params->max_num_otg;
+	ip_caps->num_dsc = ip_params->num_dsc;
+	ip_caps->max_num_dp2p0_streams = ip_params->max_num_dp2p0_streams;
+	ip_caps->max_num_dp2p0_outputs = ip_params->max_num_dp2p0_outputs;
+	ip_caps->max_num_hdmi_frl_outputs = ip_params->max_num_hdmi_frl_outputs;
+	ip_caps->rob_buffer_size_kbytes = ip_params->rob_buffer_size_kbytes;
+	ip_caps->config_return_buffer_size_in_kbytes = ip_params->config_return_buffer_size_in_kbytes;
+	ip_caps->config_return_buffer_segment_size_in_kbytes = ip_params->config_return_buffer_segment_size_in_kbytes;
+	ip_caps->meta_fifo_size_in_kentries = ip_params->meta_fifo_size_in_kentries;
+	ip_caps->compressed_buffer_segment_size_in_kbytes = ip_params->compressed_buffer_segment_size_in_kbytes;
+	ip_caps->cursor_buffer_size = ip_params->cursor_buffer_size;
+	ip_caps->max_flip_time_us = ip_params->max_flip_time_us;
+	ip_caps->max_flip_time_lines = ip_params->max_flip_time_lines;
+	ip_caps->hostvm_mode = ip_params->hostvm_mode;
+
+	// FIXME_STAGE2: cleanup after adding all dv override to ip_caps
+	ip_caps->subvp_drr_scheduling_margin_us = 100;
+	ip_caps->subvp_prefetch_end_to_mall_start_us = 15;
+	ip_caps->subvp_fw_processing_delay = 16;
+
+}
+
+static void patch_ip_params_with_ip_caps(struct dml2_core_ip_params *ip_params, const struct dml2_ip_capabilities *ip_caps)
+{
+	ip_params->max_num_dpp = ip_caps->pipe_count;
+	ip_params->max_num_otg = ip_caps->otg_count;
+	ip_params->num_dsc = ip_caps->num_dsc;
+	ip_params->max_num_dp2p0_streams = ip_caps->max_num_dp2p0_streams;
+	ip_params->max_num_dp2p0_outputs = ip_caps->max_num_dp2p0_outputs;
+	ip_params->max_num_hdmi_frl_outputs = ip_caps->max_num_hdmi_frl_outputs;
+	ip_params->rob_buffer_size_kbytes = ip_caps->rob_buffer_size_kbytes;
+	ip_params->config_return_buffer_size_in_kbytes = ip_caps->config_return_buffer_size_in_kbytes;
+	ip_params->config_return_buffer_segment_size_in_kbytes = ip_caps->config_return_buffer_segment_size_in_kbytes;
+	ip_params->meta_fifo_size_in_kentries = ip_caps->meta_fifo_size_in_kentries;
+	ip_params->compressed_buffer_segment_size_in_kbytes = ip_caps->compressed_buffer_segment_size_in_kbytes;
+	ip_params->cursor_buffer_size = ip_caps->cursor_buffer_size;
+	ip_params->max_flip_time_us = ip_caps->max_flip_time_us;
+	ip_params->max_flip_time_lines = ip_caps->max_flip_time_lines;
+	ip_params->hostvm_mode = ip_caps->hostvm_mode;
+}
+
+bool core_dcn4_initialize(struct dml2_core_initialize_in_out *in_out)
+{
+	struct dml2_core_instance *core = in_out->instance;
+
+	if (!in_out->minimum_clock_table)
+		return false;
+	else
+		core->minimum_clock_table = in_out->minimum_clock_table;
+
+	if (in_out->explicit_ip_bb && in_out->explicit_ip_bb_size > 0) {
+		memcpy(&core->clean_me_up.mode_lib.ip, in_out->explicit_ip_bb, in_out->explicit_ip_bb_size);
+
+		// FIXME_STAGE2:
+		// DV still uses stage1 ip_param_st for each variant, need to patch the ip_caps with ip_param info
+		// Should move DV to use ip_caps but need move more overrides to ip_caps
+		patch_ip_caps_with_explicit_ip_params(in_out->ip_caps, in_out->explicit_ip_bb);
+		core->clean_me_up.mode_lib.ip.subvp_pstate_allow_width_us = core_dcn4_ip_caps_base.subvp_pstate_allow_width_us;
+		core->clean_me_up.mode_lib.ip.subvp_fw_processing_delay_us = core_dcn4_ip_caps_base.subvp_pstate_allow_width_us;
+		core->clean_me_up.mode_lib.ip.subvp_swath_height_margin_lines = core_dcn4_ip_caps_base.subvp_swath_height_margin_lines;
+	} else {
+		memcpy(&core->clean_me_up.mode_lib.ip, &core_dcn4_ip_caps_base, sizeof(struct dml2_core_ip_params));
+		patch_ip_params_with_ip_caps(&core->clean_me_up.mode_lib.ip, in_out->ip_caps);
+		core->clean_me_up.mode_lib.ip.imall_supported = false;
+	}
+
+	memcpy(&core->clean_me_up.mode_lib.soc, in_out->soc_bb, sizeof(struct dml2_soc_bb));
+	memcpy(&core->clean_me_up.mode_lib.ip_caps, in_out->ip_caps, sizeof(struct dml2_ip_capabilities));
+
+	return true;
+}
+
+static void create_phantom_stream_from_main_stream(struct dml2_stream_parameters *phantom, const struct dml2_stream_parameters *main,
+	const struct dml2_implicit_svp_meta *meta)
+{
+	memcpy(phantom, main, sizeof(struct dml2_stream_parameters));
+
+	phantom->timing.v_total = meta->v_total;
+	phantom->timing.v_active = meta->v_active;
+	phantom->timing.v_front_porch = meta->v_front_porch;
+	phantom->timing.v_blank_end = phantom->timing.v_total - phantom->timing.v_front_porch - phantom->timing.v_active;
+	phantom->timing.vblank_nom = phantom->timing.v_total - phantom->timing.v_active;
+	phantom->timing.drr_config.enabled = false;
+}
+
+static void create_phantom_plane_from_main_plane(struct dml2_plane_parameters *phantom, const struct dml2_plane_parameters *main,
+	const struct dml2_stream_parameters *phantom_stream, int phantom_stream_index, const struct dml2_stream_parameters *main_stream)
+{
+	memcpy(phantom, main, sizeof(struct dml2_plane_parameters));
+
+	phantom->stream_index = phantom_stream_index;
+	phantom->overrides.refresh_from_mall = dml2_refresh_from_mall_mode_override_force_disable;
+	phantom->overrides.legacy_svp_config = dml2_svp_mode_override_phantom_pipe_no_data_return;
+	phantom->composition.viewport.plane0.height = (long int unsigned) math_min2(math_ceil2(
+		(double)main->composition.scaler_info.plane0.v_ratio * (double)phantom_stream->timing.v_active, 16.0),
+		(double)main->composition.viewport.plane0.height);
+	phantom->composition.viewport.plane1.height = (long int unsigned) math_min2(math_ceil2(
+		(double)main->composition.scaler_info.plane1.v_ratio * (double)phantom_stream->timing.v_active, 16.0),
+		(double)main->composition.viewport.plane1.height);
+	phantom->immediate_flip = false;
+	phantom->dynamic_meta_data.enable = false;
+	phantom->cursor.num_cursors = 0;
+	phantom->cursor.cursor_width = 0;
+	phantom->tdlut.setup_for_tdlut = false;
+}
+
+static void expand_implict_subvp(const struct display_configuation_with_meta *display_cfg, struct dml2_display_cfg *svp_expanded_display_cfg,
+	struct dml2_core_scratch *scratch)
+{
+	unsigned int stream_index, plane_index;
+	const struct dml2_plane_parameters *main_plane;
+	const struct dml2_stream_parameters *main_stream;
+	const struct dml2_stream_parameters *phantom_stream;
+
+	memcpy(svp_expanded_display_cfg, &display_cfg->display_config, sizeof(struct dml2_display_cfg));
+	memset(scratch->main_stream_index_from_svp_stream_index, 0, sizeof(int) * DML2_MAX_PLANES);
+	memset(scratch->svp_stream_index_from_main_stream_index, 0, sizeof(int) * DML2_MAX_PLANES);
+	memset(scratch->main_plane_index_to_phantom_plane_index, 0, sizeof(int) * DML2_MAX_PLANES);
+
+	if (!display_cfg->display_config.overrides.enable_subvp_implicit_pmo)
+		return;
+
+	/* disable unbounded requesting for all planes until stage 3 has been performed */
+	if (!display_cfg->stage3.performed) {
+		svp_expanded_display_cfg->overrides.hw.force_unbounded_requesting.enable = true;
+		svp_expanded_display_cfg->overrides.hw.force_unbounded_requesting.value = false;
+	}
+	// Create the phantom streams
+	for (stream_index = 0; stream_index < display_cfg->display_config.num_streams; stream_index++) {
+		main_stream = &display_cfg->display_config.stream_descriptors[stream_index];
+		scratch->main_stream_index_from_svp_stream_index[stream_index] = stream_index;
+		scratch->svp_stream_index_from_main_stream_index[stream_index] = stream_index;
+
+		if (display_cfg->stage3.stream_svp_meta[stream_index].valid) {
+			// Create the phantom stream
+			create_phantom_stream_from_main_stream(&svp_expanded_display_cfg->stream_descriptors[svp_expanded_display_cfg->num_streams],
+				main_stream, &display_cfg->stage3.stream_svp_meta[stream_index]);
+
+			// Associate this phantom stream to the main stream
+			scratch->main_stream_index_from_svp_stream_index[svp_expanded_display_cfg->num_streams] = stream_index;
+			scratch->svp_stream_index_from_main_stream_index[stream_index] = svp_expanded_display_cfg->num_streams;
+
+			// Increment num streams
+			svp_expanded_display_cfg->num_streams++;
+		}
+	}
+
+	// Create the phantom planes
+	for (plane_index = 0; plane_index < display_cfg->display_config.num_planes; plane_index++) {
+		main_plane = &display_cfg->display_config.plane_descriptors[plane_index];
+
+		if (display_cfg->stage3.stream_svp_meta[main_plane->stream_index].valid) {
+			main_stream = &display_cfg->display_config.stream_descriptors[main_plane->stream_index];
+			phantom_stream = &svp_expanded_display_cfg->stream_descriptors[scratch->svp_stream_index_from_main_stream_index[main_plane->stream_index]];
+			create_phantom_plane_from_main_plane(&svp_expanded_display_cfg->plane_descriptors[svp_expanded_display_cfg->num_planes],
+				main_plane, phantom_stream, scratch->svp_stream_index_from_main_stream_index[main_plane->stream_index], main_stream);
+
+			// Associate this phantom plane to the main plane
+			scratch->phantom_plane_index_to_main_plane_index[svp_expanded_display_cfg->num_planes] = plane_index;
+			scratch->main_plane_index_to_phantom_plane_index[plane_index] = svp_expanded_display_cfg->num_planes;
+
+			// Increment num planes
+			svp_expanded_display_cfg->num_planes++;
+
+			// Adjust the main plane settings
+			svp_expanded_display_cfg->plane_descriptors[plane_index].overrides.legacy_svp_config = dml2_svp_mode_override_main_pipe;
+		}
+	}
+}
+
+static void pack_mode_programming_params_with_implicit_subvp(struct dml2_core_instance *core, const struct display_configuation_with_meta *display_cfg,
+	const struct dml2_display_cfg *svp_expanded_display_cfg, struct dml2_display_cfg_programming *programming, struct dml2_core_scratch *scratch)
+{
+	unsigned int stream_index, plane_index, pipe_offset, stream_already_populated_mask, main_plane_index, mcache_index;
+	unsigned int total_main_mcaches_required = 0;
+	int total_pipe_regs_copied = 0;
+	int dml_internal_pipe_index = 0;
+	const struct dml2_plane_parameters *main_plane;
+	const struct dml2_plane_parameters *phantom_plane;
+	const struct dml2_stream_parameters *main_stream;
+	const struct dml2_stream_parameters *phantom_stream;
+
+	// Copy the unexpanded display config to output
+	memcpy(&programming->display_config, &display_cfg->display_config, sizeof(struct dml2_display_cfg));
+
+	// Set the global register values
+	dml2_core_calcs_get_arb_params(&display_cfg->display_config, &core->clean_me_up.mode_lib, &programming->global_regs.arb_regs);
+	// Get watermarks uses display config for ref clock override, so it doesn't matter whether we pass the pre or post expansion
+	// display config
+	dml2_core_calcs_get_watermarks(&display_cfg->display_config, &core->clean_me_up.mode_lib, &programming->global_regs.wm_regs[0]);
+
+	// Check if FAMS2 is required
+	if (display_cfg->stage3.performed && display_cfg->stage3.success) {
+		programming->fams2_required = display_cfg->stage3.fams2_required;
+
+		dml2_core_calcs_get_global_fams2_programming(&core->clean_me_up.mode_lib, display_cfg, &programming->fams2_global_config);
+	}
+
+	// Only loop over all the main streams (the implicit svp streams will be packed as part of the main stream)
+	for (stream_index = 0; stream_index < programming->display_config.num_streams; stream_index++) {
+		main_stream = &svp_expanded_display_cfg->stream_descriptors[stream_index];
+		phantom_stream = &svp_expanded_display_cfg->stream_descriptors[scratch->svp_stream_index_from_main_stream_index[stream_index]];
+
+		// Set the descriptor
+		programming->stream_programming[stream_index].stream_descriptor = &programming->display_config.stream_descriptors[stream_index];
+
+		// Set the odm combine factor
+		programming->stream_programming[stream_index].num_odms_required = display_cfg->mode_support_result.cfg_support_info.stream_support_info[stream_index].odms_used;
+
+		// Check if the stream has implicit SVP enabled
+		if (main_stream != phantom_stream) {
+			// If so, copy the phantom stream descriptor
+			programming->stream_programming[stream_index].phantom_stream.enabled = true;
+			memcpy(&programming->stream_programming[stream_index].phantom_stream.descriptor, phantom_stream, sizeof(struct dml2_stream_parameters));
+		} else {
+			programming->stream_programming[stream_index].phantom_stream.enabled = false;
+		}
+
+		// Due to the way DML indexes data internally, it's easier to populate the rest of the display
+		// stream programming in the next stage
+	}
+
+	dml_internal_pipe_index = 0;
+	total_pipe_regs_copied = 0;
+	stream_already_populated_mask = 0x0;
+
+	// Loop over all main planes
+	for (plane_index = 0; plane_index < programming->display_config.num_planes; plane_index++) {
+		main_plane = &svp_expanded_display_cfg->plane_descriptors[plane_index];
+
+		// Set the descriptor
+		programming->plane_programming[plane_index].plane_descriptor = &programming->display_config.plane_descriptors[plane_index];
+
+		// Set the mpc combine factor
+		programming->plane_programming[plane_index].num_dpps_required = core->clean_me_up.mode_lib.mp.NoOfDPP[plane_index];
+
+		// Setup the appropriate p-state strategy
+		if (display_cfg->stage3.performed && display_cfg->stage3.success) {
+			programming->plane_programming[plane_index].uclk_pstate_support_method = display_cfg->stage3.pstate_switch_modes[plane_index];
+		} else {
+			programming->plane_programming[plane_index].uclk_pstate_support_method = dml2_pstate_method_na;
+		}
+
+		dml2_core_calcs_get_mall_allocation(&core->clean_me_up.mode_lib, &programming->plane_programming[plane_index].surface_size_mall_bytes, dml_internal_pipe_index);
+
+		memcpy(&programming->plane_programming[plane_index].mcache_allocation,
+			&display_cfg->stage2.mcache_allocations[plane_index],
+			sizeof(struct dml2_mcache_surface_allocation));
+		total_main_mcaches_required += programming->plane_programming[plane_index].mcache_allocation.num_mcaches_plane0 +
+			programming->plane_programming[plane_index].mcache_allocation.num_mcaches_plane1 -
+			(programming->plane_programming[plane_index].mcache_allocation.last_slice_sharing.plane0_plane1 ? 1 : 0);
+
+		for (pipe_offset = 0; pipe_offset < programming->plane_programming[plane_index].num_dpps_required; pipe_offset++) {
+			// Assign storage for this pipe's register values
+			programming->plane_programming[plane_index].pipe_regs[pipe_offset] = &programming->pipe_regs[total_pipe_regs_copied];
+			memset(programming->plane_programming[plane_index].pipe_regs[pipe_offset], 0, sizeof(struct dml2_dchub_per_pipe_register_set));
+			total_pipe_regs_copied++;
+
+			// Populate the main plane regs
+			dml2_core_calcs_get_pipe_regs(svp_expanded_display_cfg, &core->clean_me_up.mode_lib, programming->plane_programming[plane_index].pipe_regs[pipe_offset], dml_internal_pipe_index);
+
+			// Multiple planes can refer to the same stream index, so it's only necessary to populate it once
+			if (!(stream_already_populated_mask & (0x1 << main_plane->stream_index))) {
+				dml2_core_calcs_get_stream_programming(&core->clean_me_up.mode_lib, &programming->stream_programming[main_plane->stream_index], dml_internal_pipe_index);
+
+				programming->stream_programming[main_plane->stream_index].uclk_pstate_method = programming->plane_programming[plane_index].uclk_pstate_support_method;
+
+				/* unconditionally populate fams2 params */
+				dml2_core_calcs_get_stream_fams2_programming(&core->clean_me_up.mode_lib,
+					display_cfg,
+					&programming->stream_programming[main_plane->stream_index].fams2_base_params,
+					&programming->stream_programming[main_plane->stream_index].fams2_sub_params,
+					programming->stream_programming[main_plane->stream_index].uclk_pstate_method,
+					plane_index);
+
+				stream_already_populated_mask |= (0x1 << main_plane->stream_index);
+			}
+			dml_internal_pipe_index++;
+		}
+	}
+
+	for (plane_index = programming->display_config.num_planes; plane_index < svp_expanded_display_cfg->num_planes; plane_index++) {
+		phantom_plane = &svp_expanded_display_cfg->plane_descriptors[plane_index];
+		main_plane_index = scratch->phantom_plane_index_to_main_plane_index[plane_index];
+		main_plane = &svp_expanded_display_cfg->plane_descriptors[main_plane_index];
+
+		programming->plane_programming[main_plane_index].phantom_plane.valid = true;
+		memcpy(&programming->plane_programming[main_plane_index].phantom_plane.descriptor, phantom_plane, sizeof(struct dml2_plane_parameters));
+
+		dml2_core_calcs_get_mall_allocation(&core->clean_me_up.mode_lib, &programming->plane_programming[main_plane_index].svp_size_mall_bytes, dml_internal_pipe_index);
+
+		/* generate mcache allocation, phantoms use identical mcache configuration, but in the MALL set and unique mcache ID's beginning after all main ID's */
+		memcpy(&programming->plane_programming[main_plane_index].phantom_plane.mcache_allocation,
+			&programming->plane_programming[main_plane_index].mcache_allocation,
+			sizeof(struct dml2_mcache_surface_allocation));
+		for (mcache_index = 0; mcache_index < programming->plane_programming[main_plane_index].phantom_plane.mcache_allocation.num_mcaches_plane0; mcache_index++) {
+			programming->plane_programming[main_plane_index].phantom_plane.mcache_allocation.global_mcache_ids_plane0[mcache_index] += total_main_mcaches_required;
+			programming->plane_programming[main_plane_index].phantom_plane.mcache_allocation.global_mcache_ids_mall_plane0[mcache_index] =
+				programming->plane_programming[main_plane_index].phantom_plane.mcache_allocation.global_mcache_ids_plane0[mcache_index];
+		}
+		for (mcache_index = 0; mcache_index < programming->plane_programming[main_plane_index].phantom_plane.mcache_allocation.num_mcaches_plane1; mcache_index++) {
+			programming->plane_programming[main_plane_index].phantom_plane.mcache_allocation.global_mcache_ids_plane1[mcache_index] += total_main_mcaches_required;
+			programming->plane_programming[main_plane_index].phantom_plane.mcache_allocation.global_mcache_ids_mall_plane1[mcache_index] =
+				programming->plane_programming[main_plane_index].phantom_plane.mcache_allocation.global_mcache_ids_plane1[mcache_index];
+		}
+
+		for (pipe_offset = 0; pipe_offset < programming->plane_programming[main_plane_index].num_dpps_required; pipe_offset++) {
+			// Assign storage for this pipe's register values
+			programming->plane_programming[main_plane_index].phantom_plane.pipe_regs[pipe_offset] = &programming->pipe_regs[total_pipe_regs_copied];
+			memset(programming->plane_programming[main_plane_index].phantom_plane.pipe_regs[pipe_offset], 0, sizeof(struct dml2_dchub_per_pipe_register_set));
+			total_pipe_regs_copied++;
+
+			// Populate the phantom plane regs
+			dml2_core_calcs_get_pipe_regs(svp_expanded_display_cfg, &core->clean_me_up.mode_lib, programming->plane_programming[main_plane_index].phantom_plane.pipe_regs[pipe_offset], dml_internal_pipe_index);
+			// Populate the phantom stream specific programming
+			if (!(stream_already_populated_mask & (0x1 << phantom_plane->stream_index))) {
+				dml2_core_calcs_get_global_sync_programming(&core->clean_me_up.mode_lib, &programming->stream_programming[main_plane->stream_index].phantom_stream.global_sync, dml_internal_pipe_index);
+
+				stream_already_populated_mask |= (0x1 << phantom_plane->stream_index);
+			}
+
+			dml_internal_pipe_index++;
+		}
+	}
+}
+
+bool core_dcn4_mode_support(struct dml2_core_mode_support_in_out *in_out)
+{
+	struct dml2_core_instance *core = (struct dml2_core_instance *)in_out->instance;
+	struct dml2_core_mode_support_locals *l = &core->scratch.mode_support_locals;
+
+	bool result;
+	unsigned int i, stream_index, stream_bitmask;
+	int unsigned odm_count, num_odm_output_segments, dpp_count;
+
+	expand_implict_subvp(in_out->display_cfg, &l->svp_expanded_display_cfg, &core->scratch);
+
+	l->mode_support_ex_params.mode_lib = &core->clean_me_up.mode_lib;
+	l->mode_support_ex_params.in_display_cfg = &l->svp_expanded_display_cfg;
+	l->mode_support_ex_params.min_clk_table = in_out->min_clk_table;
+	l->mode_support_ex_params.min_clk_index = in_out->min_clk_index;
+	l->mode_support_ex_params.out_evaluation_info = &in_out->mode_support_result.cfg_support_info.clean_me_up.support_info;
+
+	result = dml2_core_calcs_mode_support_ex(&l->mode_support_ex_params);
+
+	in_out->mode_support_result.cfg_support_info.is_supported = result;
+
+	if (result) {
+		in_out->mode_support_result.global.dispclk_khz = (unsigned int)(core->clean_me_up.mode_lib.ms.RequiredDISPCLK * 1000);
+		in_out->mode_support_result.global.dcfclk_deepsleep_khz = (unsigned int)(core->clean_me_up.mode_lib.ms.dcfclk_deepsleep * 1000);
+		in_out->mode_support_result.global.socclk_khz = (unsigned int)(core->clean_me_up.mode_lib.ms.SOCCLK * 1000);
+
+		in_out->mode_support_result.global.fclk_pstate_supported = l->mode_support_ex_params.out_evaluation_info->global_fclk_change_supported;
+		in_out->mode_support_result.global.uclk_pstate_supported = l->mode_support_ex_params.out_evaluation_info->global_dram_clock_change_supported;
+
+		in_out->mode_support_result.global.active.fclk_khz = (unsigned long)(core->clean_me_up.mode_lib.ms.FabricClock * 1000);
+		in_out->mode_support_result.global.active.dcfclk_khz = (unsigned long)(core->clean_me_up.mode_lib.ms.DCFCLK * 1000);
+
+
+		in_out->mode_support_result.global.svp_prefetch.fclk_khz = (unsigned long)core->clean_me_up.mode_lib.ms.FabricClock * 1000;
+		in_out->mode_support_result.global.svp_prefetch.dcfclk_khz = (unsigned long)core->clean_me_up.mode_lib.ms.DCFCLK * 1000;
+
+		in_out->mode_support_result.global.active.average_bw_sdp_kbps = 0;
+		in_out->mode_support_result.global.active.urgent_bw_dram_kbps = 0;
+		in_out->mode_support_result.global.svp_prefetch.average_bw_sdp_kbps = 0;
+		in_out->mode_support_result.global.svp_prefetch.urgent_bw_dram_kbps = 0;
+
+		in_out->mode_support_result.global.active.average_bw_sdp_kbps = (unsigned long)math_ceil2((l->mode_support_ex_params.out_evaluation_info->avg_bandwidth_required[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_sdp] * 1000), 1.0);
+		in_out->mode_support_result.global.active.urgent_bw_sdp_kbps = (unsigned long)math_ceil2((l->mode_support_ex_params.out_evaluation_info->urg_bandwidth_required_flip[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_sdp] * 1000), 1.0);
+		in_out->mode_support_result.global.svp_prefetch.average_bw_sdp_kbps = (unsigned long)math_ceil2((l->mode_support_ex_params.out_evaluation_info->avg_bandwidth_required[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_sdp] * 1000), 1.0);
+		in_out->mode_support_result.global.svp_prefetch.urgent_bw_sdp_kbps = (unsigned long)math_ceil2((l->mode_support_ex_params.out_evaluation_info->urg_bandwidth_required_flip[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_sdp] * 1000), 1.0);
+
+		in_out->mode_support_result.global.active.average_bw_dram_kbps = (unsigned long)math_ceil2((l->mode_support_ex_params.out_evaluation_info->avg_bandwidth_required[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_dram] * 1000), 1.0);
+		in_out->mode_support_result.global.active.urgent_bw_dram_kbps = (unsigned long)math_ceil2((l->mode_support_ex_params.out_evaluation_info->urg_bandwidth_required_flip[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_dram] * 1000), 1.0);
+		in_out->mode_support_result.global.svp_prefetch.average_bw_dram_kbps = (unsigned long)math_ceil2((l->mode_support_ex_params.out_evaluation_info->avg_bandwidth_required[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_dram] * 1000), 1.0);
+		in_out->mode_support_result.global.svp_prefetch.urgent_bw_dram_kbps = (unsigned long)math_ceil2((l->mode_support_ex_params.out_evaluation_info->urg_bandwidth_required_flip[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_dram] * 1000), 1.0);
+		DML_LOG_VERBOSE("DML::%s: in_out->mode_support_result.global.active.urgent_bw_sdp_kbps = %ld\n", __func__, in_out->mode_support_result.global.active.urgent_bw_sdp_kbps);
+		DML_LOG_VERBOSE("DML::%s: in_out->mode_support_result.global.svp_prefetch.urgent_bw_sdp_kbps = %ld\n", __func__, in_out->mode_support_result.global.svp_prefetch.urgent_bw_sdp_kbps);
+		DML_LOG_VERBOSE("DML::%s: in_out->mode_support_result.global.active.urgent_bw_dram_kbps = %ld\n", __func__, in_out->mode_support_result.global.active.urgent_bw_dram_kbps);
+		DML_LOG_VERBOSE("DML::%s: in_out->mode_support_result.global.svp_prefetch.urgent_bw_dram_kbps = %ld\n", __func__, in_out->mode_support_result.global.svp_prefetch.urgent_bw_dram_kbps);
+
+		for (i = 0; i < l->svp_expanded_display_cfg.num_planes; i++) {
+			in_out->mode_support_result.per_plane[i].dppclk_khz = (unsigned int)(core->clean_me_up.mode_lib.ms.RequiredDPPCLK[i] * 1000);
+		}
+
+		stream_bitmask = 0;
+		for (i = 0; i < l->svp_expanded_display_cfg.num_planes; i++) {
+			odm_count = 1;
+			dpp_count = l->mode_support_ex_params.out_evaluation_info->DPPPerSurface[i];
+			num_odm_output_segments = 1;
+
+			switch (l->mode_support_ex_params.out_evaluation_info->ODMMode[i]) {
+			case dml2_odm_mode_bypass:
+				odm_count = 1;
+				dpp_count = l->mode_support_ex_params.out_evaluation_info->DPPPerSurface[i];
+				break;
+			case dml2_odm_mode_combine_2to1:
+				odm_count = 2;
+				dpp_count = 2;
+				break;
+			case dml2_odm_mode_combine_3to1:
+				odm_count = 3;
+				dpp_count = 3;
+				break;
+			case dml2_odm_mode_combine_4to1:
+				odm_count = 4;
+				dpp_count = 4;
+				break;
+			case dml2_odm_mode_split_1to2:
+			case dml2_odm_mode_mso_1to2:
+				num_odm_output_segments = 2;
+				break;
+			case dml2_odm_mode_mso_1to4:
+				num_odm_output_segments = 4;
+				break;
+			case dml2_odm_mode_auto:
+			default:
+				odm_count = 1;
+				dpp_count = l->mode_support_ex_params.out_evaluation_info->DPPPerSurface[i];
+				break;
+			}
+
+			in_out->mode_support_result.cfg_support_info.plane_support_info[i].dpps_used = dpp_count;
+
+			dml2_core_calcs_get_plane_support_info(&l->svp_expanded_display_cfg, &core->clean_me_up.mode_lib, &in_out->mode_support_result.cfg_support_info.plane_support_info[i], i);
+
+			stream_index = l->svp_expanded_display_cfg.plane_descriptors[i].stream_index;
+
+			in_out->mode_support_result.per_stream[stream_index].dscclk_khz = (unsigned int)core->clean_me_up.mode_lib.ms.required_dscclk_freq_mhz[i] * 1000;
+			DML_LOG_VERBOSE("CORE_DCN4::%s: i=%d stream_index=%d, in_out->mode_support_result.per_stream[stream_index].dscclk_khz = %u\n", __func__, i, stream_index, in_out->mode_support_result.per_stream[stream_index].dscclk_khz);
+
+			if (!((stream_bitmask >> stream_index) & 0x1)) {
+				in_out->mode_support_result.cfg_support_info.stream_support_info[stream_index].odms_used = odm_count;
+				in_out->mode_support_result.cfg_support_info.stream_support_info[stream_index].num_odm_output_segments = num_odm_output_segments;
+				in_out->mode_support_result.cfg_support_info.stream_support_info[stream_index].dsc_enable = l->mode_support_ex_params.out_evaluation_info->DSCEnabled[i];
+				in_out->mode_support_result.cfg_support_info.stream_support_info[stream_index].num_dsc_slices = l->mode_support_ex_params.out_evaluation_info->NumberOfDSCSlices[i];
+				dml2_core_calcs_get_stream_support_info(&l->svp_expanded_display_cfg, &core->clean_me_up.mode_lib, &in_out->mode_support_result.cfg_support_info.stream_support_info[stream_index], i);
+				in_out->mode_support_result.per_stream[stream_index].dtbclk_khz = (unsigned int)(core->clean_me_up.mode_lib.ms.RequiredDTBCLK[i] * 1000);
+				stream_bitmask |= 0x1 << stream_index;
+			}
+		}
+	}
+
+	return result;
+}
+
+static int lookup_uclk_dpm_index_by_freq(unsigned long uclk_freq_khz, struct dml2_soc_bb *soc_bb)
+{
+	int i;
+
+	for (i = 0; i < soc_bb->clk_table.uclk.num_clk_values; i++) {
+		if (uclk_freq_khz == soc_bb->clk_table.uclk.clk_values_khz[i])
+			return i;
+	}
+	return 0;
+}
+
+bool core_dcn4_mode_programming(struct dml2_core_mode_programming_in_out *in_out)
+{
+	struct dml2_core_instance *core = (struct dml2_core_instance *)in_out->instance;
+	struct dml2_core_mode_programming_locals *l = &core->scratch.mode_programming_locals;
+
+	bool result = false;
+	unsigned int pipe_offset;
+	int dml_internal_pipe_index;
+	int total_pipe_regs_copied = 0;
+	int stream_already_populated_mask = 0;
+
+	int main_stream_index;
+	unsigned int plane_index;
+
+	expand_implict_subvp(in_out->display_cfg, &l->svp_expanded_display_cfg, &core->scratch);
+
+	l->mode_programming_ex_params.mode_lib = &core->clean_me_up.mode_lib;
+	l->mode_programming_ex_params.in_display_cfg = &l->svp_expanded_display_cfg;
+	l->mode_programming_ex_params.min_clk_table = in_out->instance->minimum_clock_table;
+	l->mode_programming_ex_params.cfg_support_info = in_out->cfg_support_info;
+	l->mode_programming_ex_params.programming = in_out->programming;
+	l->mode_programming_ex_params.min_clk_index = lookup_uclk_dpm_index_by_freq(in_out->programming->min_clocks.dcn4x.active.uclk_khz,
+		&core->clean_me_up.mode_lib.soc);
+
+	result = dml2_core_calcs_mode_programming_ex(&l->mode_programming_ex_params);
+
+	if (result) {
+		// If the input display configuration contains implict SVP, we need to use a special packer
+		if (in_out->display_cfg->display_config.overrides.enable_subvp_implicit_pmo) {
+			pack_mode_programming_params_with_implicit_subvp(core, in_out->display_cfg, &l->svp_expanded_display_cfg, in_out->programming, &core->scratch);
+		} else {
+			memcpy(&in_out->programming->display_config, in_out->display_cfg, sizeof(struct dml2_display_cfg));
+
+			dml2_core_calcs_get_arb_params(&l->svp_expanded_display_cfg, &core->clean_me_up.mode_lib, &in_out->programming->global_regs.arb_regs);
+			dml2_core_calcs_get_watermarks(&l->svp_expanded_display_cfg, &core->clean_me_up.mode_lib, &in_out->programming->global_regs.wm_regs[0]);
+
+			dml_internal_pipe_index = 0;
+
+			for (plane_index = 0; plane_index < in_out->programming->display_config.num_planes; plane_index++) {
+				in_out->programming->plane_programming[plane_index].num_dpps_required = core->clean_me_up.mode_lib.mp.NoOfDPP[plane_index];
+
+				if (in_out->programming->display_config.plane_descriptors[plane_index].overrides.legacy_svp_config == dml2_svp_mode_override_main_pipe)
+					in_out->programming->plane_programming[plane_index].uclk_pstate_support_method = dml2_pstate_method_fw_svp;
+				else if (in_out->programming->display_config.plane_descriptors[plane_index].overrides.legacy_svp_config == dml2_svp_mode_override_phantom_pipe)
+					in_out->programming->plane_programming[plane_index].uclk_pstate_support_method = dml2_pstate_method_fw_svp;
+				else if (in_out->programming->display_config.plane_descriptors[plane_index].overrides.legacy_svp_config == dml2_svp_mode_override_phantom_pipe_no_data_return)
+					in_out->programming->plane_programming[plane_index].uclk_pstate_support_method = dml2_pstate_method_fw_svp;
+				else {
+					if (core->clean_me_up.mode_lib.mp.MaxActiveDRAMClockChangeLatencySupported[plane_index] >= core->clean_me_up.mode_lib.soc.power_management_parameters.dram_clk_change_blackout_us)
+						in_out->programming->plane_programming[plane_index].uclk_pstate_support_method = dml2_pstate_method_vactive;
+					else if (core->clean_me_up.mode_lib.mp.TWait[plane_index] >= core->clean_me_up.mode_lib.soc.power_management_parameters.dram_clk_change_blackout_us)
+						in_out->programming->plane_programming[plane_index].uclk_pstate_support_method = dml2_pstate_method_vblank;
+					else
+						in_out->programming->plane_programming[plane_index].uclk_pstate_support_method = dml2_pstate_method_na;
+				}
+
+				dml2_core_calcs_get_mall_allocation(&core->clean_me_up.mode_lib, &in_out->programming->plane_programming[plane_index].surface_size_mall_bytes, dml_internal_pipe_index);
+
+				memcpy(&in_out->programming->plane_programming[plane_index].mcache_allocation,
+					&in_out->display_cfg->stage2.mcache_allocations[plane_index],
+					sizeof(struct dml2_mcache_surface_allocation));
+
+				for (pipe_offset = 0; pipe_offset < in_out->programming->plane_programming[plane_index].num_dpps_required; pipe_offset++) {
+					in_out->programming->plane_programming[plane_index].plane_descriptor = &in_out->programming->display_config.plane_descriptors[plane_index];
+
+					// Assign storage for this pipe's register values
+					in_out->programming->plane_programming[plane_index].pipe_regs[pipe_offset] = &in_out->programming->pipe_regs[total_pipe_regs_copied];
+					memset(in_out->programming->plane_programming[plane_index].pipe_regs[pipe_offset], 0, sizeof(struct dml2_dchub_per_pipe_register_set));
+					total_pipe_regs_copied++;
+
+					// Populate
+					dml2_core_calcs_get_pipe_regs(&l->svp_expanded_display_cfg, &core->clean_me_up.mode_lib, in_out->programming->plane_programming[plane_index].pipe_regs[pipe_offset], dml_internal_pipe_index);
+
+					main_stream_index = in_out->programming->display_config.plane_descriptors[plane_index].stream_index;
+
+					// Multiple planes can refer to the same stream index, so it's only necessary to populate it once
+					if (!(stream_already_populated_mask & (0x1 << main_stream_index))) {
+						in_out->programming->stream_programming[main_stream_index].stream_descriptor = &in_out->programming->display_config.stream_descriptors[main_stream_index];
+						in_out->programming->stream_programming[main_stream_index].num_odms_required = in_out->cfg_support_info->stream_support_info[main_stream_index].odms_used;
+						dml2_core_calcs_get_stream_programming(&core->clean_me_up.mode_lib, &in_out->programming->stream_programming[main_stream_index], dml_internal_pipe_index);
+
+						stream_already_populated_mask |= (0x1 << main_stream_index);
+					}
+					dml_internal_pipe_index++;
+				}
+			}
+		}
+	}
+
+	return result;
+}
+
+bool core_dcn4_populate_informative(struct dml2_core_populate_informative_in_out *in_out)
+{
+	struct dml2_core_internal_display_mode_lib *mode_lib = &in_out->instance->clean_me_up.mode_lib;
+
+	if (in_out->mode_is_supported)
+		in_out->programming->informative.voltage_level = in_out->instance->scratch.mode_programming_locals.mode_programming_ex_params.min_clk_index;
+	else
+		in_out->programming->informative.voltage_level = in_out->instance->scratch.mode_support_locals.mode_support_ex_params.min_clk_index;
+
+	dml2_core_calcs_get_informative(mode_lib, in_out->programming);
+	return true;
+}
+
+bool core_dcn4_calculate_mcache_allocation(struct dml2_calculate_mcache_allocation_in_out *in_out)
+{
+	memset(in_out->mcache_allocation, 0, sizeof(struct dml2_mcache_surface_allocation));
+
+	dml2_core_calcs_get_mcache_allocation(&in_out->instance->clean_me_up.mode_lib, in_out->mcache_allocation, in_out->plane_index);
+
+	if (in_out->mcache_allocation->num_mcaches_plane0 > 0)
+		in_out->mcache_allocation->mcache_x_offsets_plane0[in_out->mcache_allocation->num_mcaches_plane0 - 1] = in_out->plane_descriptor->surface.plane0.width;
+
+	if (in_out->mcache_allocation->num_mcaches_plane1 > 0)
+		in_out->mcache_allocation->mcache_x_offsets_plane1[in_out->mcache_allocation->num_mcaches_plane1 - 1] = in_out->plane_descriptor->surface.plane1.width;
+
+	in_out->mcache_allocation->requires_dedicated_mall_mcache = false;
+
+	return true;
+}
diff --git a/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/src/dml2_core/dml2_core_dcn4.h b/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/src/dml2_core/dml2_core_dcn4.h
new file mode 100644
index 000000000000..a68bb001a346
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/src/dml2_core/dml2_core_dcn4.h
@@ -0,0 +1,12 @@
+// SPDX-License-Identifier: MIT
+//
+// Copyright 2024 Advanced Micro Devices, Inc.
+
+#ifndef __DML2_CORE_DCN4_H__
+#define __DML2_CORE_DCN4_H__
+bool core_dcn4_initialize(struct dml2_core_initialize_in_out *in_out);
+bool core_dcn4_mode_support(struct dml2_core_mode_support_in_out *in_out);
+bool core_dcn4_mode_programming(struct dml2_core_mode_programming_in_out *in_out);
+bool core_dcn4_populate_informative(struct dml2_core_populate_informative_in_out *in_out);
+bool core_dcn4_calculate_mcache_allocation(struct dml2_calculate_mcache_allocation_in_out *in_out);
+#endif
diff --git a/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/src/dml2_core/dml2_core_dcn4_calcs.c b/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/src/dml2_core/dml2_core_dcn4_calcs.c
new file mode 100644
index 000000000000..a02e9fd6b5ca
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/src/dml2_core/dml2_core_dcn4_calcs.c
@@ -0,0 +1,13342 @@
+// SPDX-License-Identifier: MIT
+//
+// Copyright 2024 Advanced Micro Devices, Inc.
+
+
+#include "dml2_internal_shared_types.h"
+#include "dml2_core_dcn4_calcs.h"
+#include "dml2_debug.h"
+#include "lib_float_math.h"
+#include "dml_top_types.h"
+
+#define DML2_MAX_FMT_420_BUFFER_WIDTH 4096
+#define DML_MAX_NUM_OF_SLICES_PER_DSC 4
+#define DML_MAX_COMPRESSION_RATIO 4
+//#define DML_MODE_SUPPORT_USE_DPM_DRAM_BW
+//#define DML_GLOBAL_PREFETCH_CHECK
+#define ALLOW_SDPIF_RATE_LIMIT_PRE_CSTATE
+#define DML_MAX_VSTARTUP_START 1023
+
+const char *dml2_core_internal_bw_type_str(enum dml2_core_internal_bw_type bw_type)
+{
+	switch (bw_type) {
+	case (dml2_core_internal_bw_sdp):
+		return("dml2_core_internal_bw_sdp");
+	case (dml2_core_internal_bw_dram):
+		return("dml2_core_internal_bw_dram");
+	case (dml2_core_internal_bw_max):
+		return("dml2_core_internal_bw_max");
+	default:
+		return("dml2_core_internal_bw_unknown");
+	}
+}
+
+const char *dml2_core_internal_soc_state_type_str(enum dml2_core_internal_soc_state_type dml2_core_internal_soc_state_type)
+{
+	switch (dml2_core_internal_soc_state_type) {
+	case (dml2_core_internal_soc_state_sys_idle):
+		return("dml2_core_internal_soc_state_sys_idle");
+	case (dml2_core_internal_soc_state_sys_active):
+		return("dml2_core_internal_soc_state_sys_active");
+	case (dml2_core_internal_soc_state_svp_prefetch):
+		return("dml2_core_internal_soc_state_svp_prefetch");
+	case dml2_core_internal_soc_state_max:
+	default:
+		return("dml2_core_internal_soc_state_unknown");
+	}
+}
+
+static double dml2_core_div_rem(double dividend, unsigned int divisor, unsigned int *remainder)
+{
+	*remainder = ((dividend / divisor) - (int)(dividend / divisor) > 0);
+	return dividend / divisor;
+}
+
+static void dml2_print_mode_support_info(const struct dml2_core_internal_mode_support_info *support, bool fail_only)
+{
+	DML_LOG_VERBOSE("DML: ===================================== \n");
+	DML_LOG_VERBOSE("DML: DML_MODE_SUPPORT_INFO_ST\n");
+	if (!fail_only || support->ScaleRatioAndTapsSupport == 0)
+		DML_LOG_VERBOSE("DML: support: ScaleRatioAndTapsSupport = %d\n", support->ScaleRatioAndTapsSupport);
+	if (!fail_only || support->SourceFormatPixelAndScanSupport == 0)
+		DML_LOG_VERBOSE("DML: support: SourceFormatPixelAndScanSupport = %d\n", support->SourceFormatPixelAndScanSupport);
+	if (!fail_only || support->ViewportSizeSupport == 0)
+		DML_LOG_VERBOSE("DML: support: ViewportSizeSupport = %d\n", support->ViewportSizeSupport);
+	if (!fail_only || support->LinkRateDoesNotMatchDPVersion == 1)
+		DML_LOG_VERBOSE("DML: support: LinkRateDoesNotMatchDPVersion = %d\n", support->LinkRateDoesNotMatchDPVersion);
+	if (!fail_only || support->LinkRateForMultistreamNotIndicated == 1)
+		DML_LOG_VERBOSE("DML: support: LinkRateForMultistreamNotIndicated = %d\n", support->LinkRateForMultistreamNotIndicated);
+	if (!fail_only || support->BPPForMultistreamNotIndicated == 1)
+		DML_LOG_VERBOSE("DML: support: BPPForMultistreamNotIndicated = %d\n", support->BPPForMultistreamNotIndicated);
+	if (!fail_only || support->MultistreamWithHDMIOreDP == 1)
+		DML_LOG_VERBOSE("DML: support: MultistreamWithHDMIOreDP = %d\n", support->MultistreamWithHDMIOreDP);
+	if (!fail_only || support->ExceededMultistreamSlots == 1)
+		DML_LOG_VERBOSE("DML: support: ExceededMultistreamSlots = %d\n", support->ExceededMultistreamSlots);
+	if (!fail_only || support->MSOOrODMSplitWithNonDPLink == 1)
+		DML_LOG_VERBOSE("DML: support: MSOOrODMSplitWithNonDPLink = %d\n", support->MSOOrODMSplitWithNonDPLink);
+	if (!fail_only || support->NotEnoughLanesForMSO == 1)
+		DML_LOG_VERBOSE("DML: support: NotEnoughLanesForMSO = %d\n", support->NotEnoughLanesForMSO);
+	if (!fail_only || support->P2IWith420 == 1)
+		DML_LOG_VERBOSE("DML: support: P2IWith420 = %d\n", support->P2IWith420);
+	if (!fail_only || support->DSC422NativeNotSupported == 1)
+		DML_LOG_VERBOSE("DML: support: DSC422NativeNotSupported = %d\n", support->DSC422NativeNotSupported);
+	if (!fail_only || support->DSCSlicesODMModeSupported == 0)
+		DML_LOG_VERBOSE("DML: support: DSCSlicesODMModeSupported = %d\n", support->DSCSlicesODMModeSupported);
+	if (!fail_only || support->NotEnoughDSCUnits == 1)
+		DML_LOG_VERBOSE("DML: support: NotEnoughDSCUnits = %d\n", support->NotEnoughDSCUnits);
+	if (!fail_only || support->NotEnoughDSCSlices == 1)
+		DML_LOG_VERBOSE("DML: support: NotEnoughDSCSlices = %d\n", support->NotEnoughDSCSlices);
+	if (!fail_only || support->ImmediateFlipOrHostVMAndPStateWithMALLFullFrameOrPhantomPipe == 1)
+		DML_LOG_VERBOSE("DML: support: ImmediateFlipOrHostVMAndPStateWithMALLFullFrameOrPhantomPipe = %d\n", support->ImmediateFlipOrHostVMAndPStateWithMALLFullFrameOrPhantomPipe);
+	if (!fail_only || support->InvalidCombinationOfMALLUseForPStateAndStaticScreen == 1)
+		DML_LOG_VERBOSE("DML: support: InvalidCombinationOfMALLUseForPStateAndStaticScreen = %d\n", support->InvalidCombinationOfMALLUseForPStateAndStaticScreen);
+	if (!fail_only || support->DSCCLKRequiredMoreThanSupported == 1)
+		DML_LOG_VERBOSE("DML: support: DSCCLKRequiredMoreThanSupported = %d\n", support->DSCCLKRequiredMoreThanSupported);
+	if (!fail_only || support->PixelsPerLinePerDSCUnitSupport == 0)
+		DML_LOG_VERBOSE("DML: support: PixelsPerLinePerDSCUnitSupport = %d\n", support->PixelsPerLinePerDSCUnitSupport);
+	if (!fail_only || support->DTBCLKRequiredMoreThanSupported == 1)
+		DML_LOG_VERBOSE("DML: support: DTBCLKRequiredMoreThanSupported = %d\n", support->DTBCLKRequiredMoreThanSupported);
+	if (!fail_only || support->InvalidCombinationOfMALLUseForPState == 1)
+		DML_LOG_VERBOSE("DML: support: InvalidCombinationOfMALLUseForPState = %d\n", support->InvalidCombinationOfMALLUseForPState);
+	if (!fail_only || support->ROBSupport == 0)
+		DML_LOG_VERBOSE("DML: support: ROBSupport = %d\n", support->ROBSupport);
+	if (!fail_only || support->OutstandingRequestsSupport == 0)
+		DML_LOG_VERBOSE("DML: support: OutstandingRequestsSupport = %d\n", support->OutstandingRequestsSupport);
+	if (!fail_only || support->OutstandingRequestsUrgencyAvoidance == 0)
+		DML_LOG_VERBOSE("DML: support: OutstandingRequestsUrgencyAvoidance = %d\n", support->OutstandingRequestsUrgencyAvoidance);
+	if (!fail_only || support->DISPCLK_DPPCLK_Support == 0)
+		DML_LOG_VERBOSE("DML: support: DISPCLK_DPPCLK_Support = %d\n", support->DISPCLK_DPPCLK_Support);
+	if (!fail_only || support->TotalAvailablePipesSupport == 0)
+		DML_LOG_VERBOSE("DML: support: TotalAvailablePipesSupport = %d\n", support->TotalAvailablePipesSupport);
+	if (!fail_only || support->NumberOfOTGSupport == 0)
+		DML_LOG_VERBOSE("DML: support: NumberOfOTGSupport = %d\n", support->NumberOfOTGSupport);
+	if (!fail_only || support->NumberOfHDMIFRLSupport == 0)
+		DML_LOG_VERBOSE("DML: support: NumberOfHDMIFRLSupport = %d\n", support->NumberOfHDMIFRLSupport);
+	if (!fail_only || support->NumberOfDP2p0Support == 0)
+		DML_LOG_VERBOSE("DML: support: NumberOfDP2p0Support = %d\n", support->NumberOfDP2p0Support);
+	if (!fail_only || support->EnoughWritebackUnits == 0)
+		DML_LOG_VERBOSE("DML: support: EnoughWritebackUnits = %d\n", support->EnoughWritebackUnits);
+	if (!fail_only || support->WritebackScaleRatioAndTapsSupport == 0)
+		DML_LOG_VERBOSE("DML: support: WritebackScaleRatioAndTapsSupport = %d\n", support->WritebackScaleRatioAndTapsSupport);
+	if (!fail_only || support->WritebackLatencySupport == 0)
+		DML_LOG_VERBOSE("DML: support: WritebackLatencySupport = %d\n", support->WritebackLatencySupport);
+	if (!fail_only || support->CursorSupport == 0)
+		DML_LOG_VERBOSE("DML: support: CursorSupport = %d\n", support->CursorSupport);
+	if (!fail_only || support->PitchSupport == 0)
+		DML_LOG_VERBOSE("DML: support: PitchSupport = %d\n", support->PitchSupport);
+	if (!fail_only || support->ViewportExceedsSurface == 1)
+		DML_LOG_VERBOSE("DML: support: ViewportExceedsSurface = %d\n", support->ViewportExceedsSurface);
+	if (!fail_only || support->PrefetchSupported == 0)
+		DML_LOG_VERBOSE("DML: support: PrefetchSupported = %d\n", support->PrefetchSupported);
+	if (!fail_only || support->EnoughUrgentLatencyHidingSupport == 0)
+		DML_LOG_VERBOSE("DML: support: EnoughUrgentLatencyHidingSupport = %d\n", support->EnoughUrgentLatencyHidingSupport);
+	if (!fail_only || support->AvgBandwidthSupport == 0)
+		DML_LOG_VERBOSE("DML: support: AvgBandwidthSupport = %d\n", support->AvgBandwidthSupport);
+	if (!fail_only || support->DynamicMetadataSupported == 0)
+		DML_LOG_VERBOSE("DML: support: DynamicMetadataSupported = %d\n", support->DynamicMetadataSupported);
+	if (!fail_only || support->VRatioInPrefetchSupported == 0)
+		DML_LOG_VERBOSE("DML: support: VRatioInPrefetchSupported = %d\n", support->VRatioInPrefetchSupported);
+	if (!fail_only || support->PTEBufferSizeNotExceeded == 0)
+		DML_LOG_VERBOSE("DML: support: PTEBufferSizeNotExceeded = %d\n", support->PTEBufferSizeNotExceeded);
+	if (!fail_only || support->DCCMetaBufferSizeNotExceeded == 0)
+		DML_LOG_VERBOSE("DML: support: DCCMetaBufferSizeNotExceeded = %d\n", support->DCCMetaBufferSizeNotExceeded);
+	if (!fail_only || support->ExceededMALLSize == 1)
+		DML_LOG_VERBOSE("DML: support: ExceededMALLSize = %d\n", support->ExceededMALLSize);
+	if (!fail_only || support->g6_temp_read_support == 0)
+		DML_LOG_VERBOSE("DML: support: g6_temp_read_support = %d\n", support->g6_temp_read_support);
+	if (!fail_only || support->ImmediateFlipSupport == 0)
+		DML_LOG_VERBOSE("DML: support: ImmediateFlipSupport = %d\n", support->ImmediateFlipSupport);
+	if (!fail_only || support->LinkCapacitySupport == 0)
+		DML_LOG_VERBOSE("DML: support: LinkCapacitySupport = %d\n", support->LinkCapacitySupport);
+
+	if (!fail_only || support->ModeSupport == 0)
+		DML_LOG_VERBOSE("DML: support: ModeSupport = %d\n", support->ModeSupport);
+	DML_LOG_VERBOSE("DML: ===================================== \n");
+}
+
+static void get_stream_output_bpp(double *out_bpp, const struct dml2_display_cfg *display_cfg)
+{
+	for (unsigned int k = 0; k < display_cfg->num_planes; k++) {
+		double bpc = (double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.bpc;
+		if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.dsc.enable == dml2_dsc_disable) {
+			switch (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_format) {
+			case dml2_444:
+				out_bpp[k] = bpc * 3;
+				break;
+			case dml2_s422:
+				out_bpp[k] = bpc * 2;
+				break;
+			case dml2_n422:
+				out_bpp[k] = bpc * 2;
+				break;
+			case dml2_420:
+			default:
+				out_bpp[k] = bpc * 1.5;
+				break;
+			}
+		} else if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.dsc.enable == dml2_dsc_enable) {
+			out_bpp[k] = (double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.dsc.dsc_compressed_bpp_x16 / 16;
+		} else {
+			out_bpp[k] = 0;
+		}
+		DML_LOG_VERBOSE("DML::%s: k=%d bpc=%f\n", __func__, k, bpc);
+		DML_LOG_VERBOSE("DML::%s: k=%d dsc.enable=%d\n", __func__, k, display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.dsc.enable);
+		DML_LOG_VERBOSE("DML::%s: k=%d out_bpp=%f\n", __func__, k, out_bpp[k]);
+	}
+}
+
+static unsigned int dml_round_to_multiple(unsigned int num, unsigned int multiple, bool up)
+{
+	unsigned int remainder;
+
+	if (multiple == 0)
+		return num;
+
+	remainder = num % multiple;
+	if (remainder == 0)
+		return num;
+
+	if (up)
+		return (num + multiple - remainder);
+	else
+		return (num - remainder);
+}
+
+static unsigned int dml_get_num_active_pipes(int unsigned num_planes, const struct core_display_cfg_support_info *cfg_support_info)
+{
+	unsigned int num_active_pipes = 0;
+
+	for (unsigned int k = 0; k < num_planes; k++) {
+		num_active_pipes = num_active_pipes + (unsigned int)cfg_support_info->plane_support_info[k].dpps_used;
+	}
+
+	DML_LOG_VERBOSE("DML::%s: num_active_pipes = %d\n", __func__, num_active_pipes);
+	return num_active_pipes;
+}
+
+static void dml_calc_pipe_plane_mapping(const struct core_display_cfg_support_info *cfg_support_info, unsigned int *pipe_plane)
+{
+	unsigned int pipe_idx = 0;
+
+	for (unsigned int k = 0; k < DML2_MAX_PLANES; ++k) {
+		pipe_plane[k] = __DML2_CALCS_PIPE_NO_PLANE__;
+	}
+
+	for (unsigned int plane_idx = 0; plane_idx < DML2_MAX_PLANES; plane_idx++) {
+		for (int i = 0; i < cfg_support_info->plane_support_info[plane_idx].dpps_used; i++) {
+			pipe_plane[pipe_idx] = plane_idx;
+			pipe_idx++;
+		}
+	}
+}
+
+static bool dml_is_phantom_pipe(const struct dml2_plane_parameters *plane_cfg)
+{
+	bool is_phantom = false;
+
+	if (plane_cfg->overrides.legacy_svp_config == dml2_svp_mode_override_phantom_pipe ||
+		plane_cfg->overrides.legacy_svp_config == dml2_svp_mode_override_phantom_pipe_no_data_return) {
+		is_phantom = true;
+	}
+
+	return is_phantom;
+}
+
+static bool dml_get_is_phantom_pipe(const struct dml2_display_cfg *display_cfg, const struct dml2_core_internal_display_mode_lib *mode_lib, unsigned int pipe_idx)
+{
+	unsigned int plane_idx = mode_lib->mp.pipe_plane[pipe_idx];
+
+	bool is_phantom = dml_is_phantom_pipe(&display_cfg->plane_descriptors[plane_idx]);
+	DML_LOG_VERBOSE("DML::%s: pipe_idx=%d legacy_svp_config=%0d is_phantom=%d\n", __func__, pipe_idx, display_cfg->plane_descriptors[plane_idx].overrides.legacy_svp_config, is_phantom);
+	return is_phantom;
+}
+
+#define dml_get_per_pipe_var_func(variable, type, interval_var) static type dml_get_##variable(const struct dml2_core_internal_display_mode_lib *mode_lib, unsigned int pipe_idx) \
+{ \
+unsigned int plane_idx; \
+plane_idx = mode_lib->mp.pipe_plane[pipe_idx]; \
+return (type) interval_var[plane_idx]; \
+}
+
+dml_get_per_pipe_var_func(dpte_group_size_in_bytes, unsigned int, mode_lib->mp.dpte_group_bytes);
+dml_get_per_pipe_var_func(vm_group_size_in_bytes, unsigned int, mode_lib->mp.vm_group_bytes);
+dml_get_per_pipe_var_func(swath_height_l, unsigned int, mode_lib->mp.SwathHeightY);
+dml_get_per_pipe_var_func(swath_height_c, unsigned int, mode_lib->mp.SwathHeightC);
+dml_get_per_pipe_var_func(dpte_row_height_linear_l, unsigned int, mode_lib->mp.dpte_row_height_linear);
+dml_get_per_pipe_var_func(dpte_row_height_linear_c, unsigned int, mode_lib->mp.dpte_row_height_linear_chroma);
+
+dml_get_per_pipe_var_func(vstartup_calculated, unsigned int, mode_lib->mp.VStartup);
+dml_get_per_pipe_var_func(vupdate_offset, unsigned int, mode_lib->mp.VUpdateOffsetPix);
+dml_get_per_pipe_var_func(vupdate_width, unsigned int, mode_lib->mp.VUpdateWidthPix);
+dml_get_per_pipe_var_func(vready_offset, unsigned int, mode_lib->mp.VReadyOffsetPix);
+dml_get_per_pipe_var_func(pstate_keepout_dst_lines, unsigned int, mode_lib->mp.pstate_keepout_dst_lines);
+dml_get_per_pipe_var_func(det_stored_buffer_size_l_bytes, unsigned int, mode_lib->mp.DETBufferSizeY);
+dml_get_per_pipe_var_func(det_stored_buffer_size_c_bytes, unsigned int, mode_lib->mp.DETBufferSizeC);
+dml_get_per_pipe_var_func(det_buffer_size_kbytes, unsigned int, mode_lib->mp.DETBufferSizeInKByte);
+dml_get_per_pipe_var_func(surface_size_in_mall_bytes, unsigned int, mode_lib->mp.SurfaceSizeInTheMALL);
+
+#define dml_get_per_plane_var_func(variable, type, interval_var) static type dml_get_plane_##variable(const struct dml2_core_internal_display_mode_lib *mode_lib, unsigned int plane_idx) \
+{ \
+return (type) interval_var[plane_idx]; \
+}
+
+dml_get_per_plane_var_func(num_mcaches_plane0, unsigned int, mode_lib->ms.num_mcaches_l);
+dml_get_per_plane_var_func(mcache_row_bytes_plane0, unsigned int, mode_lib->ms.mcache_row_bytes_l);
+dml_get_per_plane_var_func(mcache_shift_granularity_plane0, unsigned int, mode_lib->ms.mcache_shift_granularity_l);
+dml_get_per_plane_var_func(num_mcaches_plane1, unsigned int, mode_lib->ms.num_mcaches_c);
+dml_get_per_plane_var_func(mcache_row_bytes_plane1, unsigned int, mode_lib->ms.mcache_row_bytes_c);
+dml_get_per_plane_var_func(mcache_shift_granularity_plane1, unsigned int, mode_lib->ms.mcache_shift_granularity_c);
+dml_get_per_plane_var_func(mall_comb_mcache_l, unsigned int, mode_lib->ms.mall_comb_mcache_l);
+dml_get_per_plane_var_func(mall_comb_mcache_c, unsigned int, mode_lib->ms.mall_comb_mcache_c);
+dml_get_per_plane_var_func(lc_comb_mcache, unsigned int, mode_lib->ms.lc_comb_mcache);
+dml_get_per_plane_var_func(subviewport_lines_needed_in_mall, unsigned int, mode_lib->ms.SubViewportLinesNeededInMALL);
+dml_get_per_plane_var_func(max_vstartup_lines, unsigned int, mode_lib->ms.MaxVStartupLines);
+
+#define dml_get_per_plane_array_var_func(variable, type, interval_var) static type dml_get_plane_array_##variable(const struct dml2_core_internal_display_mode_lib *mode_lib, unsigned int plane_idx, unsigned int array_idx) \
+{ \
+return (type) interval_var[plane_idx][array_idx]; \
+}
+
+dml_get_per_plane_array_var_func(mcache_offsets_plane0, unsigned int, mode_lib->ms.mcache_offsets_l);
+dml_get_per_plane_array_var_func(mcache_offsets_plane1, unsigned int, mode_lib->ms.mcache_offsets_c);
+
+#define dml_get_var_func(var, type, internal_var) static type dml_get_##var(const struct dml2_core_internal_display_mode_lib *mode_lib) \
+{ \
+return (type) internal_var; \
+}
+
+dml_get_var_func(wm_urgent, double, mode_lib->mp.Watermark.UrgentWatermark);
+dml_get_var_func(wm_stutter_exit, double, mode_lib->mp.Watermark.StutterExitWatermark);
+dml_get_var_func(wm_stutter_enter_exit, double, mode_lib->mp.Watermark.StutterEnterPlusExitWatermark);
+dml_get_var_func(wm_z8_stutter_exit, double, mode_lib->mp.Watermark.Z8StutterExitWatermark);
+dml_get_var_func(wm_z8_stutter_enter_exit, double, mode_lib->mp.Watermark.Z8StutterEnterPlusExitWatermark);
+dml_get_var_func(wm_memory_trip, double, mode_lib->mp.UrgentLatency);
+dml_get_var_func(meta_trip_memory_us, double, mode_lib->mp.MetaTripToMemory);
+
+dml_get_var_func(wm_fclk_change, double, mode_lib->mp.Watermark.FCLKChangeWatermark);
+dml_get_var_func(wm_usr_retraining, double, mode_lib->mp.Watermark.USRRetrainingWatermark);
+dml_get_var_func(wm_temp_read_or_ppt, double, mode_lib->mp.Watermark.temp_read_or_ppt_watermark_us);
+dml_get_var_func(wm_dram_clock_change, double, mode_lib->mp.Watermark.DRAMClockChangeWatermark);
+dml_get_var_func(fraction_of_urgent_bandwidth, double, mode_lib->mp.FractionOfUrgentBandwidth);
+dml_get_var_func(fraction_of_urgent_bandwidth_imm_flip, double, mode_lib->mp.FractionOfUrgentBandwidthImmediateFlip);
+dml_get_var_func(fraction_of_urgent_bandwidth_mall, double, mode_lib->mp.FractionOfUrgentBandwidthMALL);
+dml_get_var_func(wm_writeback_dram_clock_change, double, mode_lib->mp.Watermark.WritebackDRAMClockChangeWatermark);
+dml_get_var_func(wm_writeback_fclk_change, double, mode_lib->mp.Watermark.WritebackFCLKChangeWatermark);
+dml_get_var_func(stutter_efficiency, double, mode_lib->mp.StutterEfficiency);
+dml_get_var_func(stutter_efficiency_no_vblank, double, mode_lib->mp.StutterEfficiencyNotIncludingVBlank);
+dml_get_var_func(stutter_num_bursts, double, mode_lib->mp.NumberOfStutterBurstsPerFrame);
+dml_get_var_func(stutter_efficiency_z8, double, mode_lib->mp.Z8StutterEfficiency);
+dml_get_var_func(stutter_num_bursts_z8, double, mode_lib->mp.Z8NumberOfStutterBurstsPerFrame);
+dml_get_var_func(stutter_period, double, mode_lib->mp.StutterPeriod);
+dml_get_var_func(stutter_efficiency_z8_bestcase, double, mode_lib->mp.Z8StutterEfficiencyBestCase);
+dml_get_var_func(stutter_num_bursts_z8_bestcase, double, mode_lib->mp.Z8NumberOfStutterBurstsPerFrameBestCase);
+dml_get_var_func(stutter_period_bestcase, double, mode_lib->mp.StutterPeriodBestCase);
+dml_get_var_func(fclk_change_latency, double, mode_lib->mp.MaxActiveFCLKChangeLatencySupported);
+dml_get_var_func(global_dppclk_khz, double, mode_lib->mp.GlobalDPPCLK * 1000.0);
+
+dml_get_var_func(sys_active_avg_bw_required_sdp, double, mode_lib->ms.support.avg_bandwidth_required[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_sdp]);
+dml_get_var_func(sys_active_avg_bw_required_dram, double, mode_lib->ms.support.avg_bandwidth_required[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_dram]);
+
+dml_get_var_func(svp_prefetch_avg_bw_required_sdp, double, mode_lib->ms.support.avg_bandwidth_required[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_sdp]);
+dml_get_var_func(svp_prefetch_avg_bw_required_dram, double, mode_lib->ms.support.avg_bandwidth_required[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_dram]);
+
+dml_get_var_func(sys_active_avg_bw_available_sdp, double, mode_lib->mp.avg_bandwidth_available[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_sdp]);
+dml_get_var_func(sys_active_avg_bw_available_dram, double, mode_lib->mp.avg_bandwidth_available[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_dram]);
+
+dml_get_var_func(svp_prefetch_avg_bw_available_sdp, double, mode_lib->mp.avg_bandwidth_available[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_sdp]);
+dml_get_var_func(svp_prefetch_avg_bw_available_dram, double, mode_lib->mp.avg_bandwidth_available[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_dram]);
+
+dml_get_var_func(sys_active_urg_bw_available_sdp, double, mode_lib->mp.urg_bandwidth_available[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_sdp]);
+dml_get_var_func(sys_active_urg_bw_available_dram, double, mode_lib->mp.urg_bandwidth_available[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_dram]);
+dml_get_var_func(sys_active_urg_bw_available_dram_vm_only, double, mode_lib->mp.urg_bandwidth_available_vm_only[dml2_core_internal_soc_state_sys_active]);
+
+dml_get_var_func(svp_prefetch_urg_bw_available_sdp, double, mode_lib->mp.urg_bandwidth_available[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_sdp]);
+dml_get_var_func(svp_prefetch_urg_bw_available_dram, double, mode_lib->mp.urg_bandwidth_available[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_dram]);
+dml_get_var_func(svp_prefetch_urg_bw_available_dram_vm_only, double, mode_lib->mp.urg_bandwidth_available_vm_only[dml2_core_internal_soc_state_svp_prefetch]);
+
+dml_get_var_func(urgent_latency, double, mode_lib->mp.UrgentLatency);
+dml_get_var_func(max_urgent_latency_us, double, mode_lib->ms.support.max_urgent_latency_us);
+dml_get_var_func(max_non_urgent_latency_us, double, mode_lib->ms.support.max_non_urgent_latency_us);
+dml_get_var_func(avg_non_urgent_latency_us, double, mode_lib->ms.support.avg_non_urgent_latency_us);
+dml_get_var_func(avg_urgent_latency_us, double, mode_lib->ms.support.avg_urgent_latency_us);
+
+dml_get_var_func(sys_active_urg_bw_required_sdp, double, mode_lib->mp.urg_bandwidth_required[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_sdp]);
+dml_get_var_func(sys_active_urg_bw_required_dram, double, mode_lib->mp.urg_bandwidth_required[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_dram]);
+dml_get_var_func(svp_prefetch_urg_bw_required_sdp, double, mode_lib->mp.urg_bandwidth_required[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_sdp]);
+dml_get_var_func(svp_prefetch_urg_bw_required_dram, double, mode_lib->mp.urg_bandwidth_required[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_dram]);
+
+dml_get_var_func(sys_active_non_urg_required_sdp, double, mode_lib->mp.non_urg_bandwidth_required[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_sdp]);
+dml_get_var_func(sys_active_non_urg_required_dram, double, mode_lib->mp.non_urg_bandwidth_required[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_dram]);
+dml_get_var_func(svp_prefetch_non_urg_bw_required_sdp, double, mode_lib->mp.non_urg_bandwidth_required[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_sdp]);
+dml_get_var_func(svp_prefetch_non_urg_bw_required_dram, double, mode_lib->mp.non_urg_bandwidth_required[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_dram]);
+
+dml_get_var_func(sys_active_urg_bw_required_sdp_flip, double, mode_lib->mp.urg_bandwidth_required_flip[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_sdp]);
+dml_get_var_func(sys_active_urg_bw_required_dram_flip, double, mode_lib->mp.urg_bandwidth_required_flip[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_dram]);
+dml_get_var_func(svp_prefetch_urg_bw_required_sdp_flip, double, mode_lib->mp.urg_bandwidth_required_flip[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_sdp]);
+dml_get_var_func(svp_prefetch_urg_bw_required_dram_flip, double, mode_lib->mp.urg_bandwidth_required_flip[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_dram]);
+
+dml_get_var_func(sys_active_non_urg_required_sdp_flip, double, mode_lib->mp.non_urg_bandwidth_required_flip[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_sdp]);
+dml_get_var_func(sys_active_non_urg_required_dram_flip, double, mode_lib->mp.non_urg_bandwidth_required_flip[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_dram]);
+dml_get_var_func(svp_prefetch_non_urg_bw_required_sdp_flip, double, mode_lib->mp.non_urg_bandwidth_required_flip[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_sdp]);
+dml_get_var_func(svp_prefetch_non_urg_bw_required_dram_flip, double, mode_lib->mp.non_urg_bandwidth_required_flip[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_dram]);
+
+dml_get_var_func(comp_buffer_size_kbytes, unsigned int, mode_lib->mp.CompressedBufferSizeInkByte);
+
+dml_get_var_func(unbounded_request_enabled, bool, mode_lib->mp.UnboundedRequestEnabled);
+dml_get_var_func(wm_writeback_urgent, double, mode_lib->mp.Watermark.WritebackUrgentWatermark);
+dml_get_var_func(cstate_max_cap_mode, bool, mode_lib->mp.DCHUBBUB_ARB_CSTATE_MAX_CAP_MODE);
+dml_get_var_func(compbuf_reserved_space_64b, unsigned int, mode_lib->mp.compbuf_reserved_space_64b);
+dml_get_var_func(hw_debug5, bool, mode_lib->mp.hw_debug5);
+dml_get_var_func(dcfclk_deep_sleep_hysteresis, unsigned int, mode_lib->mp.dcfclk_deep_sleep_hysteresis);
+
+static void CalculateMaxDETAndMinCompressedBufferSize(
+	unsigned int ConfigReturnBufferSizeInKByte,
+	unsigned int ConfigReturnBufferSegmentSizeInKByte,
+	unsigned int ROBBufferSizeInKByte,
+	unsigned int MaxNumDPP,
+	unsigned int nomDETInKByteOverrideEnable, // VBA_DELTA, allow DV to override default DET size
+	unsigned int nomDETInKByteOverrideValue, // VBA_DELTA
+	bool is_mrq_present,
+
+	// Output
+	unsigned int *MaxTotalDETInKByte,
+	unsigned int *nomDETInKByte,
+	unsigned int *MinCompressedBufferSizeInKByte)
+{
+	if (is_mrq_present)
+		*MaxTotalDETInKByte = (unsigned int) math_ceil2((double)(ConfigReturnBufferSizeInKByte + ROBBufferSizeInKByte)*4/5, 64);
+	else
+		*MaxTotalDETInKByte = ConfigReturnBufferSizeInKByte - ConfigReturnBufferSegmentSizeInKByte;
+
+	*nomDETInKByte = (unsigned int)(math_floor2((double)*MaxTotalDETInKByte / (double)MaxNumDPP, ConfigReturnBufferSegmentSizeInKByte));
+	*MinCompressedBufferSizeInKByte = ConfigReturnBufferSizeInKByte - *MaxTotalDETInKByte;
+
+	DML_LOG_VERBOSE("DML::%s: is_mrq_present = %u\n", __func__, is_mrq_present);
+	DML_LOG_VERBOSE("DML::%s: ConfigReturnBufferSizeInKByte = %u\n", __func__, ConfigReturnBufferSizeInKByte);
+	DML_LOG_VERBOSE("DML::%s: ROBBufferSizeInKByte = %u\n", __func__, ROBBufferSizeInKByte);
+	DML_LOG_VERBOSE("DML::%s: MaxNumDPP = %u\n", __func__, MaxNumDPP);
+	DML_LOG_VERBOSE("DML::%s: MaxTotalDETInKByte = %u\n", __func__, *MaxTotalDETInKByte);
+	DML_LOG_VERBOSE("DML::%s: nomDETInKByte = %u\n", __func__, *nomDETInKByte);
+	DML_LOG_VERBOSE("DML::%s: MinCompressedBufferSizeInKByte = %u\n", __func__, *MinCompressedBufferSizeInKByte);
+
+	if (nomDETInKByteOverrideEnable) {
+		*nomDETInKByte = nomDETInKByteOverrideValue;
+		DML_LOG_VERBOSE("DML::%s: nomDETInKByte = %u (overrided)\n", __func__, *nomDETInKByte);
+	}
+}
+
+static void PixelClockAdjustmentForProgressiveToInterlaceUnit(const struct dml2_display_cfg *display_cfg, bool ptoi_supported, double *PixelClockBackEnd)
+{
+	//unsigned int num_active_planes = display_cfg->num_planes;
+
+	//Progressive To Interlace Unit Effect
+	for (unsigned int k = 0; k < display_cfg->num_planes; ++k) {
+		PixelClockBackEnd[k] = ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000);
+		if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.interlaced == 1 && ptoi_supported == true) {
+			// FIXME_STAGE2... can sw pass the pixel rate for interlaced directly
+			//display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz = 2 * display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz;
+		}
+	}
+}
+
+static bool dml_is_420(enum dml2_source_format_class source_format)
+{
+	bool val = false;
+
+	switch (source_format) {
+	case dml2_444_8:
+		val = 0;
+		break;
+	case dml2_444_16:
+		val = 0;
+		break;
+	case dml2_444_32:
+		val = 0;
+		break;
+	case dml2_444_64:
+		val = 0;
+		break;
+	case dml2_420_8:
+		val = 1;
+		break;
+	case dml2_420_10:
+		val = 1;
+		break;
+	case dml2_420_12:
+		val = 1;
+		break;
+	case dml2_422_planar_8:
+		val = 0;
+		break;
+	case dml2_422_planar_10:
+		val = 0;
+		break;
+	case dml2_422_planar_12:
+		val = 0;
+		break;
+	case dml2_422_packed_8:
+		val = 0;
+		break;
+	case dml2_422_packed_10:
+		val = 0;
+		break;
+	case dml2_422_packed_12:
+		val = 0;
+		break;
+	case dml2_rgbe_alpha:
+		val = 0;
+		break;
+	case dml2_rgbe:
+		val = 0;
+		break;
+	case dml2_mono_8:
+		val = 0;
+		break;
+	case dml2_mono_16:
+		val = 0;
+		break;
+	default:
+		DML_ASSERT(0);
+		break;
+	}
+	return val;
+}
+
+static unsigned int dml_get_tile_block_size_bytes(enum dml2_swizzle_mode sw_mode)
+{
+	if (sw_mode == dml2_sw_linear)
+		return 256;
+	else if (sw_mode == dml2_sw_256b_2d)
+		return 256;
+	else if (sw_mode == dml2_sw_4kb_2d)
+		return 4096;
+	else if (sw_mode == dml2_sw_64kb_2d)
+		return 65536;
+	else if (sw_mode == dml2_sw_256kb_2d)
+		return 262144;
+	else if (sw_mode == dml2_gfx11_sw_linear)
+		return 256;
+	else if (sw_mode == dml2_gfx11_sw_64kb_d)
+		return 65536;
+	else if (sw_mode == dml2_gfx11_sw_64kb_d_t)
+		return 65536;
+	else if (sw_mode == dml2_gfx11_sw_64kb_d_x)
+		return 65536;
+	else if (sw_mode == dml2_gfx11_sw_64kb_r_x)
+		return 65536;
+	else if (sw_mode == dml2_gfx11_sw_256kb_d_x)
+		return 262144;
+	else if (sw_mode == dml2_gfx11_sw_256kb_r_x)
+		return 262144;
+	else {
+		DML_ASSERT(0);
+		return 256;
+	}
+}
+
+static bool dml_is_vertical_rotation(enum dml2_rotation_angle Scan)
+{
+	bool is_vert = false;
+	if (Scan == dml2_rotation_90 || Scan == dml2_rotation_270) {
+		is_vert = true;
+	} else {
+		is_vert = false;
+	}
+	return is_vert;
+}
+
+static int unsigned dml_get_gfx_version(enum dml2_swizzle_mode sw_mode)
+{
+	int unsigned version = 0;
+
+	if (sw_mode == dml2_sw_linear ||
+		sw_mode == dml2_sw_256b_2d ||
+		sw_mode == dml2_sw_4kb_2d ||
+		sw_mode == dml2_sw_64kb_2d ||
+		sw_mode == dml2_sw_256kb_2d) {
+		version = 12;
+	} else if (sw_mode == dml2_gfx11_sw_linear ||
+		sw_mode == dml2_gfx11_sw_64kb_d ||
+		sw_mode == dml2_gfx11_sw_64kb_d_t ||
+		sw_mode == dml2_gfx11_sw_64kb_d_x ||
+		sw_mode == dml2_gfx11_sw_64kb_r_x ||
+		sw_mode == dml2_gfx11_sw_256kb_d_x ||
+		sw_mode == dml2_gfx11_sw_256kb_r_x) {
+		version = 11;
+	} else {
+		DML_LOG_VERBOSE("ERROR: Invalid sw_mode setting! val=%u\n", sw_mode);
+		DML_ASSERT(0);
+	}
+
+	return version;
+}
+
+static void CalculateBytePerPixelAndBlockSizes(
+	enum dml2_source_format_class SourcePixelFormat,
+	enum dml2_swizzle_mode SurfaceTiling,
+	unsigned int pitch_y,
+	unsigned int pitch_c,
+
+	// Output
+	unsigned int *BytePerPixelY,
+	unsigned int *BytePerPixelC,
+	double *BytePerPixelDETY,
+	double *BytePerPixelDETC,
+	unsigned int *BlockHeight256BytesY,
+	unsigned int *BlockHeight256BytesC,
+	unsigned int *BlockWidth256BytesY,
+	unsigned int *BlockWidth256BytesC,
+	unsigned int *MacroTileHeightY,
+	unsigned int *MacroTileHeightC,
+	unsigned int *MacroTileWidthY,
+	unsigned int *MacroTileWidthC,
+	bool *surf_linear128_l,
+	bool *surf_linear128_c)
+{
+	*BytePerPixelDETY = 0;
+	*BytePerPixelDETC = 0;
+	*BytePerPixelY = 1;
+	*BytePerPixelC = 1;
+
+	if (SourcePixelFormat == dml2_444_64) {
+		*BytePerPixelDETY = 8;
+		*BytePerPixelDETC = 0;
+		*BytePerPixelY = 8;
+		*BytePerPixelC = 0;
+	} else if (SourcePixelFormat == dml2_444_32 || SourcePixelFormat == dml2_rgbe) {
+		*BytePerPixelDETY = 4;
+		*BytePerPixelDETC = 0;
+		*BytePerPixelY = 4;
+		*BytePerPixelC = 0;
+	} else if (SourcePixelFormat == dml2_444_16 || SourcePixelFormat == dml2_mono_16) {
+		*BytePerPixelDETY = 2;
+		*BytePerPixelDETC = 0;
+		*BytePerPixelY = 2;
+		*BytePerPixelC = 0;
+	} else if (SourcePixelFormat == dml2_444_8 || SourcePixelFormat == dml2_mono_8) {
+		*BytePerPixelDETY = 1;
+		*BytePerPixelDETC = 0;
+		*BytePerPixelY = 1;
+		*BytePerPixelC = 0;
+	} else if (SourcePixelFormat == dml2_rgbe_alpha) {
+		*BytePerPixelDETY = 4;
+		*BytePerPixelDETC = 1;
+		*BytePerPixelY = 4;
+		*BytePerPixelC = 1;
+	} else if (SourcePixelFormat == dml2_420_8) {
+		*BytePerPixelDETY = 1;
+		*BytePerPixelDETC = 2;
+		*BytePerPixelY = 1;
+		*BytePerPixelC = 2;
+	} else if (SourcePixelFormat == dml2_420_12) {
+		*BytePerPixelDETY = 2;
+		*BytePerPixelDETC = 4;
+		*BytePerPixelY = 2;
+		*BytePerPixelC = 4;
+	} else if (SourcePixelFormat == dml2_420_10) {
+		*BytePerPixelDETY = (double)(4.0 / 3);
+		*BytePerPixelDETC = (double)(8.0 / 3);
+		*BytePerPixelY = 2;
+		*BytePerPixelC = 4;
+	} else {
+		DML_LOG_VERBOSE("ERROR: DML::%s: SourcePixelFormat = %u not supported!\n", __func__, SourcePixelFormat);
+		DML_ASSERT(0);
+	}
+
+	DML_LOG_VERBOSE("DML::%s: SourcePixelFormat = %u\n", __func__, SourcePixelFormat);
+	DML_LOG_VERBOSE("DML::%s: BytePerPixelDETY = %f\n", __func__, *BytePerPixelDETY);
+	DML_LOG_VERBOSE("DML::%s: BytePerPixelDETC = %f\n", __func__, *BytePerPixelDETC);
+	DML_LOG_VERBOSE("DML::%s: BytePerPixelY = %u\n", __func__, *BytePerPixelY);
+	DML_LOG_VERBOSE("DML::%s: BytePerPixelC = %u\n", __func__, *BytePerPixelC);
+	DML_LOG_VERBOSE("DML::%s: pitch_y = %u\n", __func__, pitch_y);
+	DML_LOG_VERBOSE("DML::%s: pitch_c = %u\n", __func__, pitch_c);
+	DML_LOG_VERBOSE("DML::%s: surf_linear128_l = %u\n", __func__, *surf_linear128_l);
+	DML_LOG_VERBOSE("DML::%s: surf_linear128_c = %u\n", __func__, *surf_linear128_c);
+
+	if (dml_get_gfx_version(SurfaceTiling) == 11) {
+		*surf_linear128_l = 0;
+		*surf_linear128_c = 0;
+	} else {
+		if (SurfaceTiling == dml2_sw_linear) {
+			*surf_linear128_l = (((pitch_y * *BytePerPixelY) % 256) != 0);
+
+			if (dml_is_420(SourcePixelFormat) || SourcePixelFormat == dml2_rgbe_alpha)
+				*surf_linear128_c = (((pitch_c * *BytePerPixelC) % 256) != 0);
+		}
+	}
+
+	if (!(dml_is_420(SourcePixelFormat) || SourcePixelFormat == dml2_rgbe_alpha)) {
+		if (SurfaceTiling == dml2_sw_linear) {
+			*BlockHeight256BytesY = 1;
+		} else if (SourcePixelFormat == dml2_444_64) {
+			*BlockHeight256BytesY = 4;
+		} else if (SourcePixelFormat == dml2_444_8) {
+			*BlockHeight256BytesY = 16;
+		} else {
+			*BlockHeight256BytesY = 8;
+		}
+		*BlockWidth256BytesY = 256U / *BytePerPixelY / *BlockHeight256BytesY;
+		*BlockHeight256BytesC = 0;
+		*BlockWidth256BytesC = 0;
+	} else { // dual plane
+		if (SurfaceTiling == dml2_sw_linear) {
+			*BlockHeight256BytesY = 1;
+			*BlockHeight256BytesC = 1;
+		} else if (SourcePixelFormat == dml2_rgbe_alpha) {
+			*BlockHeight256BytesY = 8;
+			*BlockHeight256BytesC = 16;
+		} else if (SourcePixelFormat == dml2_420_8) {
+			*BlockHeight256BytesY = 16;
+			*BlockHeight256BytesC = 8;
+		} else {
+			*BlockHeight256BytesY = 8;
+			*BlockHeight256BytesC = 8;
+		}
+		*BlockWidth256BytesY = 256U / *BytePerPixelY / *BlockHeight256BytesY;
+		*BlockWidth256BytesC = 256U / *BytePerPixelC / *BlockHeight256BytesC;
+	}
+	DML_LOG_VERBOSE("DML::%s: BlockWidth256BytesY = %u\n", __func__, *BlockWidth256BytesY);
+	DML_LOG_VERBOSE("DML::%s: BlockHeight256BytesY = %u\n", __func__, *BlockHeight256BytesY);
+	DML_LOG_VERBOSE("DML::%s: BlockWidth256BytesC = %u\n", __func__, *BlockWidth256BytesC);
+	DML_LOG_VERBOSE("DML::%s: BlockHeight256BytesC = %u\n", __func__, *BlockHeight256BytesC);
+
+	if (dml_get_gfx_version(SurfaceTiling) == 11) {
+		if (SurfaceTiling == dml2_gfx11_sw_linear) {
+			*MacroTileHeightY = *BlockHeight256BytesY;
+			*MacroTileWidthY = 256 / *BytePerPixelY / *MacroTileHeightY;
+			*MacroTileHeightC = *BlockHeight256BytesC;
+			if (*MacroTileHeightC == 0) {
+				*MacroTileWidthC = 0;
+			} else {
+				*MacroTileWidthC = 256 / *BytePerPixelC / *MacroTileHeightC;
+			}
+		} else if (SurfaceTiling == dml2_gfx11_sw_64kb_d || SurfaceTiling == dml2_gfx11_sw_64kb_d_t || SurfaceTiling == dml2_gfx11_sw_64kb_d_x || SurfaceTiling == dml2_gfx11_sw_64kb_r_x) {
+			*MacroTileHeightY = 16 * *BlockHeight256BytesY;
+			*MacroTileWidthY = 65536 / *BytePerPixelY / *MacroTileHeightY;
+			*MacroTileHeightC = 16 * *BlockHeight256BytesC;
+			if (*MacroTileHeightC == 0) {
+				*MacroTileWidthC = 0;
+			} else {
+				*MacroTileWidthC = 65536 / *BytePerPixelC / *MacroTileHeightC;
+			}
+		} else {
+			*MacroTileHeightY = 32 * *BlockHeight256BytesY;
+			*MacroTileWidthY = 65536 * 4 / *BytePerPixelY / *MacroTileHeightY;
+			*MacroTileHeightC = 32 * *BlockHeight256BytesC;
+			if (*MacroTileHeightC == 0) {
+				*MacroTileWidthC = 0;
+			} else {
+				*MacroTileWidthC = 65536 * 4 / *BytePerPixelC / *MacroTileHeightC;
+			}
+		}
+	} else {
+		unsigned int macro_tile_size_bytes = dml_get_tile_block_size_bytes(SurfaceTiling);
+		unsigned int macro_tile_scale = 1; // macro tile to 256B req scaling
+
+		if (SurfaceTiling == dml2_sw_linear) {
+			macro_tile_scale = 1;
+		} else if (SurfaceTiling == dml2_sw_4kb_2d) {
+			macro_tile_scale = 4;
+		} else if (SurfaceTiling == dml2_sw_64kb_2d) {
+			macro_tile_scale = 16;
+		} else if (SurfaceTiling == dml2_sw_256kb_2d) {
+			macro_tile_scale = 32;
+		} else {
+			DML_LOG_VERBOSE("ERROR: Invalid SurfaceTiling setting! val=%u\n", SurfaceTiling);
+			DML_ASSERT(0);
+		}
+
+		*MacroTileHeightY = macro_tile_scale * *BlockHeight256BytesY;
+		*MacroTileWidthY = macro_tile_size_bytes / *BytePerPixelY / *MacroTileHeightY;
+		*MacroTileHeightC = macro_tile_scale * *BlockHeight256BytesC;
+		if (*MacroTileHeightC == 0) {
+			*MacroTileWidthC = 0;
+		} else {
+			*MacroTileWidthC = macro_tile_size_bytes / *BytePerPixelC / *MacroTileHeightC;
+		}
+	}
+
+	DML_LOG_VERBOSE("DML::%s: MacroTileWidthY = %u\n", __func__, *MacroTileWidthY);
+	DML_LOG_VERBOSE("DML::%s: MacroTileHeightY = %u\n", __func__, *MacroTileHeightY);
+	DML_LOG_VERBOSE("DML::%s: MacroTileWidthC = %u\n", __func__, *MacroTileWidthC);
+	DML_LOG_VERBOSE("DML::%s: MacroTileHeightC = %u\n", __func__, *MacroTileHeightC);
+}
+
+static void CalculateSinglePipeDPPCLKAndSCLThroughput(
+	double HRatio,
+	double HRatioChroma,
+	double VRatio,
+	double VRatioChroma,
+	double MaxDCHUBToPSCLThroughput,
+	double MaxPSCLToLBThroughput,
+	double PixelClock,
+	enum dml2_source_format_class SourcePixelFormat,
+	unsigned int HTaps,
+	unsigned int HTapsChroma,
+	unsigned int VTaps,
+	unsigned int VTapsChroma,
+
+	// Output
+	double *PSCL_THROUGHPUT,
+	double *PSCL_THROUGHPUT_CHROMA,
+	double *DPPCLKUsingSingleDPP)
+{
+	double DPPCLKUsingSingleDPPLuma;
+	double DPPCLKUsingSingleDPPChroma;
+
+	if (HRatio > 1) {
+		*PSCL_THROUGHPUT = math_min2(MaxDCHUBToPSCLThroughput, MaxPSCLToLBThroughput * HRatio / math_ceil2((double)HTaps / 6.0, 1.0));
+	} else {
+		*PSCL_THROUGHPUT = math_min2(MaxDCHUBToPSCLThroughput, MaxPSCLToLBThroughput);
+	}
+
+	DPPCLKUsingSingleDPPLuma = PixelClock * math_max3(VTaps / 6 * math_min2(1, HRatio), HRatio * VRatio / *PSCL_THROUGHPUT, 1);
+
+	if ((HTaps > 6 || VTaps > 6) && DPPCLKUsingSingleDPPLuma < 2 * PixelClock)
+		DPPCLKUsingSingleDPPLuma = 2 * PixelClock;
+
+	if (!dml_is_420(SourcePixelFormat) && SourcePixelFormat != dml2_rgbe_alpha) {
+		*PSCL_THROUGHPUT_CHROMA = 0;
+		*DPPCLKUsingSingleDPP = DPPCLKUsingSingleDPPLuma;
+	} else {
+		if (HRatioChroma > 1) {
+			*PSCL_THROUGHPUT_CHROMA = math_min2(MaxDCHUBToPSCLThroughput, MaxPSCLToLBThroughput * HRatioChroma / math_ceil2((double)HTapsChroma / 6.0, 1.0));
+		} else {
+			*PSCL_THROUGHPUT_CHROMA = math_min2(MaxDCHUBToPSCLThroughput, MaxPSCLToLBThroughput);
+		}
+		DPPCLKUsingSingleDPPChroma = PixelClock * math_max3(VTapsChroma / 6 * math_min2(1, HRatioChroma),
+			HRatioChroma * VRatioChroma / *PSCL_THROUGHPUT_CHROMA, 1);
+		if ((HTapsChroma > 6 || VTapsChroma > 6) && DPPCLKUsingSingleDPPChroma < 2 * PixelClock)
+			DPPCLKUsingSingleDPPChroma = 2 * PixelClock;
+		*DPPCLKUsingSingleDPP = math_max2(DPPCLKUsingSingleDPPLuma, DPPCLKUsingSingleDPPChroma);
+	}
+}
+
+static void CalculateSwathWidth(
+	const struct dml2_display_cfg *display_cfg,
+	bool ForceSingleDPP,
+	unsigned int NumberOfActiveSurfaces,
+	enum dml2_odm_mode ODMMode[],
+	unsigned int BytePerPixY[],
+	unsigned int BytePerPixC[],
+	unsigned int Read256BytesBlockHeightY[],
+	unsigned int Read256BytesBlockHeightC[],
+	unsigned int Read256BytesBlockWidthY[],
+	unsigned int Read256BytesBlockWidthC[],
+	bool surf_linear128_l[],
+	bool surf_linear128_c[],
+	unsigned int DPPPerSurface[],
+
+	// Output
+	unsigned int req_per_swath_ub_l[],
+	unsigned int req_per_swath_ub_c[],
+	unsigned int SwathWidthSingleDPPY[], // post-rotated plane width
+	unsigned int SwathWidthSingleDPPC[],
+	unsigned int SwathWidthY[], // per-pipe
+	unsigned int SwathWidthC[], // per-pipe
+	unsigned int MaximumSwathHeightY[],
+	unsigned int MaximumSwathHeightC[],
+	unsigned int swath_width_luma_ub[], // per-pipe
+	unsigned int swath_width_chroma_ub[]) // per-pipe
+{
+	enum dml2_odm_mode MainSurfaceODMMode;
+	double odm_hactive_factor = 1.0;
+	unsigned int req_width_horz_y;
+	unsigned int req_width_horz_c;
+	unsigned int surface_width_ub_l;
+	unsigned int surface_height_ub_l;
+	unsigned int surface_width_ub_c;
+	unsigned int surface_height_ub_c;
+
+	DML_LOG_VERBOSE("DML::%s: ForceSingleDPP = %u\n", __func__, ForceSingleDPP);
+	DML_LOG_VERBOSE("DML::%s: NumberOfActiveSurfaces = %u\n", __func__, NumberOfActiveSurfaces);
+
+	for (unsigned int k = 0; k < NumberOfActiveSurfaces; ++k) {
+		if (!dml_is_vertical_rotation(display_cfg->plane_descriptors[k].composition.rotation_angle)) {
+			SwathWidthSingleDPPY[k] = (unsigned int)display_cfg->plane_descriptors[k].composition.viewport.plane0.width;
+		} else {
+			SwathWidthSingleDPPY[k] = (unsigned int)display_cfg->plane_descriptors[k].composition.viewport.plane0.height;
+		}
+
+		DML_LOG_VERBOSE("DML::%s: k=%u ViewportWidth=%lu\n", __func__, k, display_cfg->plane_descriptors[k].composition.viewport.plane0.width);
+		DML_LOG_VERBOSE("DML::%s: k=%u ViewportHeight=%lu\n", __func__, k, display_cfg->plane_descriptors[k].composition.viewport.plane0.height);
+		DML_LOG_VERBOSE("DML::%s: k=%u DPPPerSurface=%u\n", __func__, k, DPPPerSurface[k]);
+
+		MainSurfaceODMMode = ODMMode[k];
+
+		if (ForceSingleDPP) {
+			SwathWidthY[k] = SwathWidthSingleDPPY[k];
+		} else {
+			if (MainSurfaceODMMode == dml2_odm_mode_combine_4to1)
+				odm_hactive_factor = 4.0;
+			else if (MainSurfaceODMMode == dml2_odm_mode_combine_3to1)
+				odm_hactive_factor = 3.0;
+			else if (MainSurfaceODMMode == dml2_odm_mode_combine_2to1)
+				odm_hactive_factor = 2.0;
+
+			if (MainSurfaceODMMode == dml2_odm_mode_combine_4to1 || MainSurfaceODMMode == dml2_odm_mode_combine_3to1 || MainSurfaceODMMode == dml2_odm_mode_combine_2to1) {
+				SwathWidthY[k] = (unsigned int)(math_min2((double)SwathWidthSingleDPPY[k], math_round((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_active / odm_hactive_factor * display_cfg->plane_descriptors[k].composition.scaler_info.plane0.h_ratio)));
+			} else if (DPPPerSurface[k] == 2) {
+				SwathWidthY[k] = SwathWidthSingleDPPY[k] / 2;
+			} else {
+				SwathWidthY[k] = SwathWidthSingleDPPY[k];
+			}
+		}
+
+		DML_LOG_VERBOSE("DML::%s: k=%u HActive=%lu\n", __func__, k, display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_active);
+		DML_LOG_VERBOSE("DML::%s: k=%u HRatio=%f\n", __func__, k, display_cfg->plane_descriptors[k].composition.scaler_info.plane0.h_ratio);
+		DML_LOG_VERBOSE("DML::%s: k=%u MainSurfaceODMMode=%u\n", __func__, k, MainSurfaceODMMode);
+		DML_LOG_VERBOSE("DML::%s: k=%u SwathWidthSingleDPPY=%u\n", __func__, k, SwathWidthSingleDPPY[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u SwathWidthY=%u\n", __func__, k, SwathWidthY[k]);
+
+		if (dml_is_420(display_cfg->plane_descriptors[k].pixel_format)) {
+			SwathWidthC[k] = SwathWidthY[k] / 2;
+			SwathWidthSingleDPPC[k] = SwathWidthSingleDPPY[k] / 2;
+		} else {
+			SwathWidthC[k] = SwathWidthY[k];
+			SwathWidthSingleDPPC[k] = SwathWidthSingleDPPY[k];
+		}
+
+		if (ForceSingleDPP == true) {
+			SwathWidthY[k] = SwathWidthSingleDPPY[k];
+			SwathWidthC[k] = SwathWidthSingleDPPC[k];
+		}
+
+		req_width_horz_y = Read256BytesBlockWidthY[k];
+		req_width_horz_c = Read256BytesBlockWidthC[k];
+
+		if (surf_linear128_l[k])
+			req_width_horz_y = req_width_horz_y / 2;
+
+		if (surf_linear128_c[k])
+			req_width_horz_c = req_width_horz_c / 2;
+
+		surface_width_ub_l = (unsigned int)math_ceil2((double)display_cfg->plane_descriptors[k].surface.plane0.width, req_width_horz_y);
+		surface_height_ub_l = (unsigned int)math_ceil2((double)display_cfg->plane_descriptors[k].surface.plane0.height, Read256BytesBlockHeightY[k]);
+		surface_width_ub_c = (unsigned int)math_ceil2((double)display_cfg->plane_descriptors[k].surface.plane1.width, req_width_horz_c);
+		surface_height_ub_c = (unsigned int)math_ceil2((double)display_cfg->plane_descriptors[k].surface.plane1.height, Read256BytesBlockHeightC[k]);
+
+		DML_LOG_VERBOSE("DML::%s: k=%u surface_width_ub_l=%u\n", __func__, k, surface_width_ub_l);
+		DML_LOG_VERBOSE("DML::%s: k=%u surface_height_ub_l=%u\n", __func__, k, surface_height_ub_l);
+		DML_LOG_VERBOSE("DML::%s: k=%u surface_width_ub_c=%u\n", __func__, k, surface_width_ub_c);
+		DML_LOG_VERBOSE("DML::%s: k=%u surface_height_ub_c=%u\n", __func__, k, surface_height_ub_c);
+		DML_LOG_VERBOSE("DML::%s: k=%u req_width_horz_y=%u\n", __func__, k, req_width_horz_y);
+		DML_LOG_VERBOSE("DML::%s: k=%u req_width_horz_c=%u\n", __func__, k, req_width_horz_c);
+		DML_LOG_VERBOSE("DML::%s: k=%u Read256BytesBlockWidthY=%u\n", __func__, k, Read256BytesBlockWidthY[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u Read256BytesBlockHeightY=%u\n", __func__, k, Read256BytesBlockHeightY[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u Read256BytesBlockWidthC=%u\n", __func__, k, Read256BytesBlockWidthC[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u Read256BytesBlockHeightC=%u\n", __func__, k, Read256BytesBlockHeightC[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u req_width_horz_y=%u\n", __func__, k, req_width_horz_y);
+		DML_LOG_VERBOSE("DML::%s: k=%u req_width_horz_c=%u\n", __func__, k, req_width_horz_c);
+		DML_LOG_VERBOSE("DML::%s: k=%u ViewportStationary=%u\n", __func__, k, display_cfg->plane_descriptors[k].composition.viewport.stationary);
+		DML_LOG_VERBOSE("DML::%s: k=%u DPPPerSurface=%u\n", __func__, k, DPPPerSurface[k]);
+
+		req_per_swath_ub_l[k] = 0;
+		req_per_swath_ub_c[k] = 0;
+		if (!dml_is_vertical_rotation(display_cfg->plane_descriptors[k].composition.rotation_angle)) {
+			MaximumSwathHeightY[k] = Read256BytesBlockHeightY[k];
+			MaximumSwathHeightC[k] = Read256BytesBlockHeightC[k];
+			if (display_cfg->plane_descriptors[k].composition.viewport.stationary && DPPPerSurface[k] == 1) {
+				swath_width_luma_ub[k] = (unsigned int)(math_min2(surface_width_ub_l, math_floor2(display_cfg->plane_descriptors[k].composition.viewport.plane0.x_start + SwathWidthY[k] + req_width_horz_y - 1, req_width_horz_y) - math_floor2(display_cfg->plane_descriptors[k].composition.viewport.plane0.x_start, req_width_horz_y)));
+			} else {
+				swath_width_luma_ub[k] = (unsigned int)(math_min2(surface_width_ub_l, math_ceil2((double)SwathWidthY[k] - 1, req_width_horz_y) + req_width_horz_y));
+			}
+			req_per_swath_ub_l[k] = swath_width_luma_ub[k] / req_width_horz_y;
+
+			if (BytePerPixC[k] > 0) {
+				if (display_cfg->plane_descriptors[k].composition.viewport.stationary && DPPPerSurface[k] == 1) {
+					swath_width_chroma_ub[k] = (unsigned int)(math_min2(surface_width_ub_c, math_floor2(display_cfg->plane_descriptors[k].composition.viewport.plane1.y_start + SwathWidthC[k] + req_width_horz_c - 1, req_width_horz_c) - math_floor2(display_cfg->plane_descriptors[k].composition.viewport.plane1.y_start, req_width_horz_c)));
+				} else {
+					swath_width_chroma_ub[k] = (unsigned int)(math_min2(surface_width_ub_c, math_ceil2((double)SwathWidthC[k] - 1, req_width_horz_c) + req_width_horz_c));
+				}
+				req_per_swath_ub_c[k] = swath_width_chroma_ub[k] / req_width_horz_c;
+			} else {
+				swath_width_chroma_ub[k] = 0;
+			}
+		} else {
+			MaximumSwathHeightY[k] = Read256BytesBlockWidthY[k];
+			MaximumSwathHeightC[k] = Read256BytesBlockWidthC[k];
+
+			if (display_cfg->plane_descriptors[k].composition.viewport.stationary && DPPPerSurface[k] == 1) {
+				swath_width_luma_ub[k] = (unsigned int)(math_min2(surface_height_ub_l, math_floor2(display_cfg->plane_descriptors[k].composition.viewport.plane0.y_start + SwathWidthY[k] + Read256BytesBlockHeightY[k] - 1, Read256BytesBlockHeightY[k]) - math_floor2(display_cfg->plane_descriptors[k].composition.viewport.plane0.y_start, Read256BytesBlockHeightY[k])));
+			} else {
+				swath_width_luma_ub[k] = (unsigned int)(math_min2(surface_height_ub_l, math_ceil2((double)SwathWidthY[k] - 1, Read256BytesBlockHeightY[k]) + Read256BytesBlockHeightY[k]));
+			}
+			req_per_swath_ub_l[k] = swath_width_luma_ub[k] / Read256BytesBlockHeightY[k];
+			if (BytePerPixC[k] > 0) {
+				if (display_cfg->plane_descriptors[k].composition.viewport.stationary && DPPPerSurface[k] == 1) {
+					swath_width_chroma_ub[k] = (unsigned int)(math_min2(surface_height_ub_c, math_floor2(display_cfg->plane_descriptors[k].composition.viewport.plane1.y_start + SwathWidthC[k] + Read256BytesBlockHeightC[k] - 1, Read256BytesBlockHeightC[k]) - math_floor2(display_cfg->plane_descriptors[k].composition.viewport.plane1.y_start, Read256BytesBlockHeightC[k])));
+				} else {
+					swath_width_chroma_ub[k] = (unsigned int)(math_min2(surface_height_ub_c, math_ceil2((double)SwathWidthC[k] - 1, Read256BytesBlockHeightC[k]) + Read256BytesBlockHeightC[k]));
+				}
+				req_per_swath_ub_c[k] = swath_width_chroma_ub[k] / Read256BytesBlockHeightC[k];
+			} else {
+				swath_width_chroma_ub[k] = 0;
+			}
+		}
+
+		DML_LOG_VERBOSE("DML::%s: k=%u swath_width_luma_ub=%u\n", __func__, k, swath_width_luma_ub[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u swath_width_chroma_ub=%u\n", __func__, k, swath_width_chroma_ub[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u MaximumSwathHeightY=%u\n", __func__, k, MaximumSwathHeightY[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u MaximumSwathHeightC=%u\n", __func__, k, MaximumSwathHeightC[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u req_per_swath_ub_l=%u\n", __func__, k, req_per_swath_ub_l[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u req_per_swath_ub_c=%u\n", __func__, k, req_per_swath_ub_c[k]);
+	}
+}
+
+static bool UnboundedRequest(bool unb_req_force_en, bool unb_req_force_val, unsigned int TotalNumberOfActiveDPP, bool NoChromaOrLinear)
+{
+	bool unb_req_ok = false;
+	bool unb_req_en = false;
+
+	unb_req_ok = (TotalNumberOfActiveDPP == 1 && NoChromaOrLinear);
+	unb_req_en = unb_req_ok;
+
+	if (unb_req_force_en) {
+		unb_req_en = unb_req_force_val && unb_req_ok;
+	}
+	DML_LOG_VERBOSE("DML::%s: unb_req_force_en = %u\n", __func__, unb_req_force_en);
+	DML_LOG_VERBOSE("DML::%s: unb_req_force_val = %u\n", __func__, unb_req_force_val);
+	DML_LOG_VERBOSE("DML::%s: unb_req_ok = %u\n", __func__, unb_req_ok);
+	DML_LOG_VERBOSE("DML::%s: unb_req_en = %u\n", __func__, unb_req_en);
+	return unb_req_en;
+}
+
+static void CalculateDETBufferSize(
+	struct dml2_core_shared_CalculateDETBufferSize_locals *l,
+	const struct dml2_display_cfg *display_cfg,
+	bool ForceSingleDPP,
+	unsigned int NumberOfActiveSurfaces,
+	bool UnboundedRequestEnabled,
+	unsigned int nomDETInKByte,
+	unsigned int MaxTotalDETInKByte,
+	unsigned int ConfigReturnBufferSizeInKByte,
+	unsigned int MinCompressedBufferSizeInKByte,
+	unsigned int ConfigReturnBufferSegmentSizeInkByte,
+	unsigned int CompressedBufferSegmentSizeInkByte,
+	double ReadBandwidthLuma[],
+	double ReadBandwidthChroma[],
+	unsigned int full_swath_bytes_l[],
+	unsigned int full_swath_bytes_c[],
+	unsigned int DPPPerSurface[],
+	// Output
+	unsigned int DETBufferSizeInKByte[],
+	unsigned int *CompressedBufferSizeInkByte)
+{
+	memset(l, 0, sizeof(struct dml2_core_shared_CalculateDETBufferSize_locals));
+
+	bool DETPieceAssignedToThisSurfaceAlready[DML2_MAX_PLANES];
+	bool NextPotentialSurfaceToAssignDETPieceFound;
+	bool MinimizeReallocationSuccess = false;
+
+	DML_LOG_VERBOSE("DML::%s: ForceSingleDPP = %u\n", __func__, ForceSingleDPP);
+	DML_LOG_VERBOSE("DML::%s: nomDETInKByte = %u\n", __func__, nomDETInKByte);
+	DML_LOG_VERBOSE("DML::%s: NumberOfActiveSurfaces = %u\n", __func__, NumberOfActiveSurfaces);
+	DML_LOG_VERBOSE("DML::%s: UnboundedRequestEnabled = %u\n", __func__, UnboundedRequestEnabled);
+	DML_LOG_VERBOSE("DML::%s: MaxTotalDETInKByte = %u\n", __func__, MaxTotalDETInKByte);
+	DML_LOG_VERBOSE("DML::%s: ConfigReturnBufferSizeInKByte = %u\n", __func__, ConfigReturnBufferSizeInKByte);
+	DML_LOG_VERBOSE("DML::%s: MinCompressedBufferSizeInKByte = %u\n", __func__, MinCompressedBufferSizeInKByte);
+	DML_LOG_VERBOSE("DML::%s: CompressedBufferSegmentSizeInkByte = %u\n", __func__, CompressedBufferSegmentSizeInkByte);
+
+	// Note: Will use default det size if that fits 2 swaths
+	if (UnboundedRequestEnabled) {
+		if (display_cfg->plane_descriptors[0].overrides.det_size_override_kb > 0) {
+			DETBufferSizeInKByte[0] = display_cfg->plane_descriptors[0].overrides.det_size_override_kb;
+		} else {
+			DETBufferSizeInKByte[0] = (unsigned int)math_max2(128.0, math_ceil2(2.0 * ((double)full_swath_bytes_l[0] + (double)full_swath_bytes_c[0]) / 1024.0, ConfigReturnBufferSegmentSizeInkByte));
+		}
+		*CompressedBufferSizeInkByte = ConfigReturnBufferSizeInKByte - DETBufferSizeInKByte[0];
+	} else {
+		l->DETBufferSizePoolInKByte = MaxTotalDETInKByte;
+		for (unsigned int k = 0; k < NumberOfActiveSurfaces; ++k) {
+			DETBufferSizeInKByte[k] = 0;
+			if (dml_is_420(display_cfg->plane_descriptors[k].pixel_format)) {
+				l->max_minDET = nomDETInKByte - ConfigReturnBufferSegmentSizeInkByte;
+			} else {
+				l->max_minDET = nomDETInKByte;
+			}
+			l->minDET = 128;
+			l->minDET_pipe = 0;
+
+			// add DET resource until can hold 2 full swaths
+			while (l->minDET <= l->max_minDET && l->minDET_pipe == 0) {
+				if (2.0 * ((double)full_swath_bytes_l[k] + (double)full_swath_bytes_c[k]) / 1024.0 <= l->minDET)
+					l->minDET_pipe = l->minDET;
+				l->minDET = l->minDET + ConfigReturnBufferSegmentSizeInkByte;
+			}
+
+			DML_LOG_VERBOSE("DML::%s: k=%u minDET = %u\n", __func__, k, l->minDET);
+			DML_LOG_VERBOSE("DML::%s: k=%u max_minDET = %u\n", __func__, k, l->max_minDET);
+			DML_LOG_VERBOSE("DML::%s: k=%u minDET_pipe = %u\n", __func__, k, l->minDET_pipe);
+			DML_LOG_VERBOSE("DML::%s: k=%u full_swath_bytes_l = %u\n", __func__, k, full_swath_bytes_l[k]);
+			DML_LOG_VERBOSE("DML::%s: k=%u full_swath_bytes_c = %u\n", __func__, k, full_swath_bytes_c[k]);
+
+			if (l->minDET_pipe == 0) {
+				l->minDET_pipe = (unsigned int)(math_max2(128, math_ceil2(((double)full_swath_bytes_l[k] + (double)full_swath_bytes_c[k]) / 1024.0, ConfigReturnBufferSegmentSizeInkByte)));
+				DML_LOG_VERBOSE("DML::%s: k=%u minDET_pipe = %u (assume each plane take half DET)\n", __func__, k, l->minDET_pipe);
+			}
+
+			if (dml_is_phantom_pipe(&display_cfg->plane_descriptors[k])) {
+				DETBufferSizeInKByte[k] = 0;
+			} else if (display_cfg->plane_descriptors[k].overrides.det_size_override_kb > 0) {
+				DETBufferSizeInKByte[k] = display_cfg->plane_descriptors[k].overrides.det_size_override_kb;
+				l->DETBufferSizePoolInKByte = l->DETBufferSizePoolInKByte - (ForceSingleDPP ? 1 : DPPPerSurface[k]) * display_cfg->plane_descriptors[k].overrides.det_size_override_kb;
+			} else if ((ForceSingleDPP ? 1 : DPPPerSurface[k]) * l->minDET_pipe <= l->DETBufferSizePoolInKByte) {
+				DETBufferSizeInKByte[k] = l->minDET_pipe;
+				l->DETBufferSizePoolInKByte = l->DETBufferSizePoolInKByte - (ForceSingleDPP ? 1 : DPPPerSurface[k]) * l->minDET_pipe;
+			}
+
+			DML_LOG_VERBOSE("DML::%s: k=%u DPPPerSurface = %u\n", __func__, k, DPPPerSurface[k]);
+			DML_LOG_VERBOSE("DML::%s: k=%u DETSizeOverride = %u\n", __func__, k, display_cfg->plane_descriptors[k].overrides.det_size_override_kb);
+			DML_LOG_VERBOSE("DML::%s: k=%u DETBufferSizeInKByte = %u\n", __func__, k, DETBufferSizeInKByte[k]);
+			DML_LOG_VERBOSE("DML::%s: DETBufferSizePoolInKByte = %u\n", __func__, l->DETBufferSizePoolInKByte);
+		}
+
+		if (display_cfg->minimize_det_reallocation) {
+			MinimizeReallocationSuccess = true;
+			// To minimize det reallocation, we don't distribute based on each surfaces bandwidth proportional to the global
+			// but rather distribute DET across streams proportionally based on pixel rate, and only distribute based on
+			// bandwidth between the planes on the same stream.  This ensures that large scale re-distribution only on a
+			// stream count and/or pixel rate change, which is must less likely then general bandwidth changes per plane.
+
+			// Calculate total pixel rate
+			for (unsigned int k = 0; k < display_cfg->num_streams; ++k) {
+				l->TotalPixelRate += display_cfg->stream_descriptors[k].timing.pixel_clock_khz;
+			}
+
+			// Calculate per stream DET budget
+			for (unsigned int k = 0; k < display_cfg->num_streams; ++k) {
+				l->DETBudgetPerStream[k] = (unsigned int)((double) display_cfg->stream_descriptors[k].timing.pixel_clock_khz * MaxTotalDETInKByte / l->TotalPixelRate);
+				l->RemainingDETBudgetPerStream[k] = l->DETBudgetPerStream[k];
+			}
+
+			// Calculate the per stream total bandwidth
+			for (unsigned int k = 0; k < NumberOfActiveSurfaces; ++k) {
+				if (!dml_is_phantom_pipe(&display_cfg->plane_descriptors[k])) {
+					l->TotalBandwidthPerStream[display_cfg->plane_descriptors[k].stream_index] += (unsigned int)(ReadBandwidthLuma[k] + ReadBandwidthChroma[k]);
+
+					// Check the minimum can be satisfied by budget
+					if (l->RemainingDETBudgetPerStream[display_cfg->plane_descriptors[k].stream_index] >= DETBufferSizeInKByte[k] * (ForceSingleDPP ? 1 : DPPPerSurface[k])) {
+						l->RemainingDETBudgetPerStream[display_cfg->plane_descriptors[k].stream_index] -= DETBufferSizeInKByte[k] * (ForceSingleDPP ? 1 : DPPPerSurface[k]);
+					} else {
+						MinimizeReallocationSuccess = false;
+						break;
+					}
+				}
+			}
+
+			if (MinimizeReallocationSuccess) {
+				// Since a fixed budget per stream is sufficient to satisfy the minimums, just re-distribute each streams
+				// budget proportionally across its planes
+				l->ResidualDETAfterRounding = MaxTotalDETInKByte;
+
+				for (unsigned int k = 0; k < NumberOfActiveSurfaces; ++k) {
+					if (!dml_is_phantom_pipe(&display_cfg->plane_descriptors[k])) {
+						l->IdealDETBudget = (unsigned int)(((ReadBandwidthLuma[k] + ReadBandwidthChroma[k]) / l->TotalBandwidthPerStream[display_cfg->plane_descriptors[k].stream_index])
+							* l->DETBudgetPerStream[display_cfg->plane_descriptors[k].stream_index]);
+
+						if (l->IdealDETBudget > DETBufferSizeInKByte[k]) {
+							l->DeltaDETBudget = l->IdealDETBudget - DETBufferSizeInKByte[k];
+							if (l->DeltaDETBudget > l->RemainingDETBudgetPerStream[display_cfg->plane_descriptors[k].stream_index])
+								l->DeltaDETBudget = l->RemainingDETBudgetPerStream[display_cfg->plane_descriptors[k].stream_index];
+
+							/* split the additional budgeted DET among the pipes per plane */
+							DETBufferSizeInKByte[k] += (unsigned int)((double)l->DeltaDETBudget / (ForceSingleDPP ? 1 : DPPPerSurface[k]));
+							l->RemainingDETBudgetPerStream[display_cfg->plane_descriptors[k].stream_index] -= l->DeltaDETBudget;
+						}
+
+						// Round down to segment size
+						DETBufferSizeInKByte[k] = (DETBufferSizeInKByte[k] / ConfigReturnBufferSegmentSizeInkByte) * ConfigReturnBufferSegmentSizeInkByte;
+
+						l->ResidualDETAfterRounding -= DETBufferSizeInKByte[k] * (ForceSingleDPP ? 1 : DPPPerSurface[k]);
+					}
+				}
+			}
+		}
+
+		if (!MinimizeReallocationSuccess) {
+			l->TotalBandwidth = 0;
+			for (unsigned int k = 0; k < NumberOfActiveSurfaces; ++k) {
+				if (!dml_is_phantom_pipe(&display_cfg->plane_descriptors[k])) {
+					l->TotalBandwidth = l->TotalBandwidth + ReadBandwidthLuma[k] + ReadBandwidthChroma[k];
+				}
+			}
+			DML_LOG_VERBOSE("DML::%s: --- Before bandwidth adjustment ---\n", __func__);
+			for (unsigned int k = 0; k < NumberOfActiveSurfaces; ++k) {
+				DML_LOG_VERBOSE("DML::%s: k=%u DETBufferSizeInKByte = %u\n", __func__, k, DETBufferSizeInKByte[k]);
+			}
+			DML_LOG_VERBOSE("DML::%s: --- DET allocation with bandwidth ---\n", __func__);
+			DML_LOG_VERBOSE("DML::%s: TotalBandwidth = %f\n", __func__, l->TotalBandwidth);
+			l->BandwidthOfSurfacesNotAssignedDETPiece = l->TotalBandwidth;
+			for (unsigned int k = 0; k < NumberOfActiveSurfaces; ++k) {
+
+				if (dml_is_phantom_pipe(&display_cfg->plane_descriptors[k])) {
+					DETPieceAssignedToThisSurfaceAlready[k] = true;
+				} else if (display_cfg->plane_descriptors[k].overrides.det_size_override_kb > 0 || (((double)(ForceSingleDPP ? 1 : DPPPerSurface[k]) * (double)DETBufferSizeInKByte[k] / (double)MaxTotalDETInKByte) >= ((ReadBandwidthLuma[k] + ReadBandwidthChroma[k]) / l->TotalBandwidth))) {
+					DETPieceAssignedToThisSurfaceAlready[k] = true;
+					l->BandwidthOfSurfacesNotAssignedDETPiece = l->BandwidthOfSurfacesNotAssignedDETPiece - ReadBandwidthLuma[k] - ReadBandwidthChroma[k];
+				} else {
+					DETPieceAssignedToThisSurfaceAlready[k] = false;
+				}
+				DML_LOG_VERBOSE("DML::%s: k=%u DETPieceAssignedToThisSurfaceAlready = %u\n", __func__, k, DETPieceAssignedToThisSurfaceAlready[k]);
+				DML_LOG_VERBOSE("DML::%s: k=%u BandwidthOfSurfacesNotAssignedDETPiece = %f\n", __func__, k, l->BandwidthOfSurfacesNotAssignedDETPiece);
+			}
+
+			for (unsigned int j = 0; j < NumberOfActiveSurfaces; ++j) {
+				NextPotentialSurfaceToAssignDETPieceFound = false;
+				l->NextSurfaceToAssignDETPiece = 0;
+
+				for (unsigned int k = 0; k < NumberOfActiveSurfaces; ++k) {
+					DML_LOG_VERBOSE("DML::%s: j=%u k=%u, ReadBandwidthLuma[k] = %f\n", __func__, j, k, ReadBandwidthLuma[k]);
+					DML_LOG_VERBOSE("DML::%s: j=%u k=%u, ReadBandwidthChroma[k] = %f\n", __func__, j, k, ReadBandwidthChroma[k]);
+					DML_LOG_VERBOSE("DML::%s: j=%u k=%u, ReadBandwidthLuma[Next] = %f\n", __func__, j, k, ReadBandwidthLuma[l->NextSurfaceToAssignDETPiece]);
+					DML_LOG_VERBOSE("DML::%s: j=%u k=%u, ReadBandwidthChroma[Next] = %f\n", __func__, j, k, ReadBandwidthChroma[l->NextSurfaceToAssignDETPiece]);
+					DML_LOG_VERBOSE("DML::%s: j=%u k=%u, NextSurfaceToAssignDETPiece = %u\n", __func__, j, k, l->NextSurfaceToAssignDETPiece);
+					if (!DETPieceAssignedToThisSurfaceAlready[k] && (!NextPotentialSurfaceToAssignDETPieceFound ||
+						ReadBandwidthLuma[k] + ReadBandwidthChroma[k] < ReadBandwidthLuma[l->NextSurfaceToAssignDETPiece] + ReadBandwidthChroma[l->NextSurfaceToAssignDETPiece])) {
+						l->NextSurfaceToAssignDETPiece = k;
+						NextPotentialSurfaceToAssignDETPieceFound = true;
+					}
+					DML_LOG_VERBOSE("DML::%s: j=%u k=%u, DETPieceAssignedToThisSurfaceAlready = %u\n", __func__, j, k, DETPieceAssignedToThisSurfaceAlready[k]);
+					DML_LOG_VERBOSE("DML::%s: j=%u k=%u, NextPotentialSurfaceToAssignDETPieceFound = %u\n", __func__, j, k, NextPotentialSurfaceToAssignDETPieceFound);
+				}
+
+				if (NextPotentialSurfaceToAssignDETPieceFound) {
+					l->NextDETBufferPieceInKByte = (unsigned int)(math_min2(
+						math_round((double)l->DETBufferSizePoolInKByte * (ReadBandwidthLuma[l->NextSurfaceToAssignDETPiece] + ReadBandwidthChroma[l->NextSurfaceToAssignDETPiece]) / l->BandwidthOfSurfacesNotAssignedDETPiece /
+							((ForceSingleDPP ? 1 : DPPPerSurface[l->NextSurfaceToAssignDETPiece]) * ConfigReturnBufferSegmentSizeInkByte))
+						* (ForceSingleDPP ? 1 : DPPPerSurface[l->NextSurfaceToAssignDETPiece]) * ConfigReturnBufferSegmentSizeInkByte,
+						math_floor2((double)l->DETBufferSizePoolInKByte, (ForceSingleDPP ? 1 : DPPPerSurface[l->NextSurfaceToAssignDETPiece]) * ConfigReturnBufferSegmentSizeInkByte)));
+
+					DML_LOG_VERBOSE("DML::%s: j=%u, DETBufferSizePoolInKByte = %u\n", __func__, j, l->DETBufferSizePoolInKByte);
+					DML_LOG_VERBOSE("DML::%s: j=%u, NextSurfaceToAssignDETPiece = %u\n", __func__, j, l->NextSurfaceToAssignDETPiece);
+					DML_LOG_VERBOSE("DML::%s: j=%u, ReadBandwidthLuma[%u] = %f\n", __func__, j, l->NextSurfaceToAssignDETPiece, ReadBandwidthLuma[l->NextSurfaceToAssignDETPiece]);
+					DML_LOG_VERBOSE("DML::%s: j=%u, ReadBandwidthChroma[%u] = %f\n", __func__, j, l->NextSurfaceToAssignDETPiece, ReadBandwidthChroma[l->NextSurfaceToAssignDETPiece]);
+					DML_LOG_VERBOSE("DML::%s: j=%u, BandwidthOfSurfacesNotAssignedDETPiece = %f\n", __func__, j, l->BandwidthOfSurfacesNotAssignedDETPiece);
+					DML_LOG_VERBOSE("DML::%s: j=%u, NextDETBufferPieceInKByte = %u\n", __func__, j, l->NextDETBufferPieceInKByte);
+					DML_LOG_VERBOSE("DML::%s: j=%u, DETBufferSizeInKByte[%u] increases from %u ", __func__, j, l->NextSurfaceToAssignDETPiece, DETBufferSizeInKByte[l->NextSurfaceToAssignDETPiece]);
+
+					DETBufferSizeInKByte[l->NextSurfaceToAssignDETPiece] = DETBufferSizeInKByte[l->NextSurfaceToAssignDETPiece] + l->NextDETBufferPieceInKByte / (ForceSingleDPP ? 1 : DPPPerSurface[l->NextSurfaceToAssignDETPiece]);
+					DML_LOG_VERBOSE("to %u\n", DETBufferSizeInKByte[l->NextSurfaceToAssignDETPiece]);
+
+					l->DETBufferSizePoolInKByte = l->DETBufferSizePoolInKByte - l->NextDETBufferPieceInKByte;
+					DETPieceAssignedToThisSurfaceAlready[l->NextSurfaceToAssignDETPiece] = true;
+					l->BandwidthOfSurfacesNotAssignedDETPiece = l->BandwidthOfSurfacesNotAssignedDETPiece - (ReadBandwidthLuma[l->NextSurfaceToAssignDETPiece] + ReadBandwidthChroma[l->NextSurfaceToAssignDETPiece]);
+				}
+			}
+		}
+		*CompressedBufferSizeInkByte = MinCompressedBufferSizeInKByte;
+	}
+	*CompressedBufferSizeInkByte = *CompressedBufferSizeInkByte * CompressedBufferSegmentSizeInkByte / ConfigReturnBufferSegmentSizeInkByte;
+
+	DML_LOG_VERBOSE("DML::%s: --- After bandwidth adjustment ---\n", __func__);
+	DML_LOG_VERBOSE("DML::%s: CompressedBufferSizeInkByte = %u\n", __func__, *CompressedBufferSizeInkByte);
+	for (unsigned int k = 0; k < NumberOfActiveSurfaces; ++k) {
+		DML_LOG_VERBOSE("DML::%s: k=%u DETBufferSizeInKByte = %u (TotalReadBandWidth=%f)\n", __func__, k, DETBufferSizeInKByte[k], ReadBandwidthLuma[k] + ReadBandwidthChroma[k]);
+	}
+}
+
+static double CalculateRequiredDispclk(
+	enum dml2_odm_mode ODMMode,
+	double PixelClock,
+	bool isTMDS420)
+{
+	double DispClk;
+
+	if (ODMMode == dml2_odm_mode_combine_4to1) {
+		DispClk = PixelClock / 4.0;
+	} else if (ODMMode == dml2_odm_mode_combine_3to1) {
+		DispClk = PixelClock / 3.0;
+	} else if (ODMMode == dml2_odm_mode_combine_2to1) {
+		DispClk = PixelClock / 2.0;
+	} else {
+		DispClk = PixelClock;
+	}
+
+	if (isTMDS420) {
+		double TMDS420MinPixClock = PixelClock / 2.0;
+		DispClk = math_max2(DispClk, TMDS420MinPixClock);
+	}
+
+	return DispClk;
+}
+
+static double TruncToValidBPP(
+	struct dml2_core_shared_TruncToValidBPP_locals *l,
+	double LinkBitRate,
+	unsigned int Lanes,
+	unsigned int HTotal,
+	unsigned int HActive,
+	double PixelClock,
+	double DesiredBPP,
+	bool DSCEnable,
+	enum dml2_output_encoder_class Output,
+	enum dml2_output_format_class Format,
+	unsigned int DSCInputBitPerComponent,
+	unsigned int DSCSlices,
+	unsigned int AudioRate,
+	unsigned int AudioLayout,
+	enum dml2_odm_mode ODMModeNoDSC,
+	enum dml2_odm_mode ODMModeDSC,
+
+	// Output
+	unsigned int *RequiredSlots)
+{
+	double MaxLinkBPP;
+	unsigned int MinDSCBPP;
+	double MaxDSCBPP;
+	unsigned int NonDSCBPP0;
+	unsigned int NonDSCBPP1;
+	unsigned int NonDSCBPP2;
+	enum dml2_odm_mode ODMMode;
+
+	if (Format == dml2_420) {
+		NonDSCBPP0 = 12;
+		NonDSCBPP1 = 15;
+		NonDSCBPP2 = 18;
+		MinDSCBPP = 6;
+		MaxDSCBPP = 16;
+	} else if (Format == dml2_444) {
+		NonDSCBPP0 = 24;
+		NonDSCBPP1 = 30;
+		NonDSCBPP2 = 36;
+		MinDSCBPP = 8;
+		MaxDSCBPP = 16;
+	} else {
+
+		if (Output == dml2_hdmi || Output == dml2_hdmifrl) {
+			NonDSCBPP0 = 24;
+			NonDSCBPP1 = 24;
+			NonDSCBPP2 = 24;
+		} else {
+			NonDSCBPP0 = 16;
+			NonDSCBPP1 = 20;
+			NonDSCBPP2 = 24;
+		}
+		if (Format == dml2_n422 || Output == dml2_hdmifrl) {
+			MinDSCBPP = 7;
+			MaxDSCBPP = 16;
+		} else {
+			MinDSCBPP = 8;
+			MaxDSCBPP = 16;
+		}
+	}
+
+	if (Output == dml2_dp2p0) {
+		MaxLinkBPP = LinkBitRate * Lanes / PixelClock * 128.0 / 132.0 * 383.0 / 384.0 * 65536.0 / 65540.0;
+	} else if (DSCEnable && Output == dml2_dp) {
+		MaxLinkBPP = LinkBitRate / 10.0 * 8.0 * Lanes / PixelClock * (1 - 2.4 / 100);
+	} else {
+		MaxLinkBPP = LinkBitRate / 10.0 * 8.0 * Lanes / PixelClock;
+	}
+
+	ODMMode = DSCEnable ? ODMModeDSC : ODMModeNoDSC;
+
+	if (ODMMode == dml2_odm_mode_split_1to2) {
+		MaxLinkBPP = 2 * MaxLinkBPP;
+	}
+
+	if (DesiredBPP == 0) {
+		if (DSCEnable) {
+			if (MaxLinkBPP < MinDSCBPP) {
+				return __DML2_CALCS_DPP_INVALID__;
+			} else if (MaxLinkBPP >= MaxDSCBPP) {
+				return MaxDSCBPP;
+			} else {
+				return math_floor2(16.0 * MaxLinkBPP, 1.0) / 16.0;
+			}
+		} else {
+			if (MaxLinkBPP >= NonDSCBPP2) {
+				return NonDSCBPP2;
+			} else if (MaxLinkBPP >= NonDSCBPP1) {
+				return NonDSCBPP1;
+			} else if (MaxLinkBPP >= NonDSCBPP0) {
+				return NonDSCBPP0;
+			} else {
+				return __DML2_CALCS_DPP_INVALID__;
+			}
+		}
+	} else {
+		if (!((DSCEnable == false && (DesiredBPP == NonDSCBPP2 || DesiredBPP == NonDSCBPP1 || DesiredBPP == NonDSCBPP0)) ||
+			(DSCEnable && DesiredBPP >= MinDSCBPP && DesiredBPP <= MaxDSCBPP))) {
+			return __DML2_CALCS_DPP_INVALID__;
+		} else {
+			return DesiredBPP;
+		}
+	}
+}
+
+// updated for dcn4
+static unsigned int dscceComputeDelay(
+	unsigned int bpc,
+	double BPP,
+	unsigned int sliceWidth,
+	unsigned int numSlices,
+	enum dml2_output_format_class pixelFormat,
+	enum dml2_output_encoder_class Output)
+{
+	// valid bpc = source bits per component in the set of {8, 10, 12}
+	// valid bpp = increments of 1/16 of a bit
+	// min = 6/7/8 in N420/N422/444, respectively
+	// max = such that compression is 1:1
+	//valid sliceWidth = number of pixels per slice line, must be less than or equal to 5184/numSlices (or 4096/numSlices in 420 mode)
+	//valid numSlices = number of slices in the horiziontal direction per DSC engine in the set of {1, 2, 3, 4}
+	//valid pixelFormat = pixel/color format in the set of {:N444_RGB, :S422, :N422, :N420}
+
+	// fixed value
+	unsigned int rcModelSize = 8192;
+
+	// N422/N420 operate at 2 pixels per clock
+	unsigned int pixelsPerClock, padding_pixels, ssm_group_priming_delay, ssm_pipeline_delay, obsm_pipeline_delay, slice_padded_pixels, ixd_plus_padding, ixd_plus_padding_groups, cycles_per_group, group_delay, pipeline_delay, pixels, additional_group_delay, lines_to_reach_ixd, groups_to_reach_ixd, slice_width_groups, initial_xmit_delay, number_of_lines_to_reach_ixd, slice_width_modified;
+
+	if (pixelFormat == dml2_420)
+		pixelsPerClock = 2;
+	// #all other modes operate at 1 pixel per clock
+	else if (pixelFormat == dml2_444)
+		pixelsPerClock = 1;
+	else if (pixelFormat == dml2_n422 || Output == dml2_hdmifrl)
+		pixelsPerClock = 2;
+	else
+		pixelsPerClock = 1;
+
+	//initial transmit delay as per PPS
+	initial_xmit_delay = (unsigned int)(math_round(rcModelSize / 2.0 / BPP / pixelsPerClock));
+
+	//slice width as seen by dscc_bcl in pixels or pixels pairs (depending on number of pixels per pixel container based on pixel format)
+	slice_width_modified = (pixelFormat == dml2_444 || pixelFormat == dml2_420 || Output == dml2_hdmifrl) ? sliceWidth / 2 : sliceWidth;
+
+	padding_pixels = ((slice_width_modified % 3) != 0) ? (3 - (slice_width_modified % 3)) * (initial_xmit_delay / slice_width_modified) : 0;
+
+	if ((3.0 * pixelsPerClock * BPP) >= ((double)((initial_xmit_delay + 2) / 3) * (double)(3 + (pixelFormat == dml2_n422)))) {
+		if ((initial_xmit_delay + padding_pixels) % 3 == 1) {
+			initial_xmit_delay++;
+		}
+	}
+
+	//sub-stream multiplexer balance fifo priming delay in groups as per dsc standard
+	if (bpc == 8)
+		ssm_group_priming_delay = 83;
+	else if (bpc == 10)
+		ssm_group_priming_delay = 91;
+	else if (bpc == 12)
+		ssm_group_priming_delay = 115;
+	else if (bpc == 14)
+		ssm_group_priming_delay = 123;
+	else
+		ssm_group_priming_delay = 128;
+
+	//slice width in groups is rounded up to the nearest group as DSC adds padded pixels such that there are an integer number of groups per slice
+	slice_width_groups = (slice_width_modified + 2) / 3;
+
+	//determine number of padded pixels in the last group of a slice line, computed as
+	slice_padded_pixels = 3 * slice_width_groups - slice_width_modified;
+
+	//determine integer number of complete slice lines required to reach initial transmit delay without ssm delay considered
+	number_of_lines_to_reach_ixd = initial_xmit_delay / slice_width_modified;
+
+	//increase initial transmit delay by the number of padded pixels added to a slice line multipled by the integer number of complete lines to reach initial transmit delay
+	//this step is necessary as each padded pixel added takes up a clock cycle and, therefore, adds to the overall delay
+	ixd_plus_padding = initial_xmit_delay + slice_padded_pixels * number_of_lines_to_reach_ixd;
+
+	//convert the padded initial transmit delay from pixels to groups by rounding up to the nearest group as DSC processes in groups of pixels
+	ixd_plus_padding_groups = (ixd_plus_padding + 2) / 3;
+
+	//number of groups required for a slice to reach initial transmit delay is the sum of the padded initial transmit delay plus the ssm group priming delay
+	groups_to_reach_ixd = ixd_plus_padding_groups + ssm_group_priming_delay;
+
+	//number of lines required to reach padded initial transmit delay in groups in slices to the left of the last horizontal slice
+	//needs to be rounded up as a complete slice lines are buffered prior to initial transmit delay being reached in the last horizontal slice
+	lines_to_reach_ixd = (groups_to_reach_ixd + slice_width_groups - 1) / slice_width_groups; //round up lines to reach ixd to next
+
+	//determine if there are non-zero number of pixels reached in the group where initial transmit delay is reached
+	//an additional group time (i.e., 3 pixel times) is required before the first output if there are no additional pixels beyond initial transmit delay
+	additional_group_delay = ((initial_xmit_delay - number_of_lines_to_reach_ixd * slice_width_modified) % 3) == 0 ? 1 : 0;
+
+	//number of pipeline delay cycles in the ssm block (can be determined empirically or analytically by inspecting the ssm block)
+	ssm_pipeline_delay = 2;
+
+	//number of pipe delay cycles in the obsm block (can be determined empirically or analytically by inspecting the obsm block)
+	obsm_pipeline_delay = 1;
+
+	//a group of pixels is worth 6 pixels in N422/N420 mode or 3 pixels in all other modes
+	if (pixelFormat == dml2_420 || pixelFormat == dml2_444 || pixelFormat == dml2_n422 || Output == dml2_hdmifrl)
+		cycles_per_group = 6;
+	else
+		cycles_per_group = 3;
+	//delay of the bit stream contruction layer in pixels is the sum of:
+	//1. number of pixel containers in a slice line multipled by the number of lines required to reach initial transmit delay multipled by number of slices to the left of the last horizontal slice
+	//2. number of pixel containers required to reach initial transmit delay (specifically, in the last horizontal slice)
+	//3. additional group of delay if initial transmit delay is reached exactly in a group
+	//4. ssm and obsm pipeline delay (i.e., clock cycles of delay)
+	group_delay = (lines_to_reach_ixd * slice_width_groups * (numSlices - 1)) + groups_to_reach_ixd + additional_group_delay;
+	pipeline_delay = ssm_pipeline_delay + obsm_pipeline_delay;
+
+	//pixel delay is group_delay (converted to pixels) + pipeline, however, first group is a special case since it is processed as soon as it arrives (i.e., in 3 cycles regardless of pixel format)
+	pixels = (group_delay - 1) * cycles_per_group + 3 + pipeline_delay;
+
+	DML_LOG_VERBOSE("DML::%s: bpc: %u\n", __func__, bpc);
+	DML_LOG_VERBOSE("DML::%s: BPP: %f\n", __func__, BPP);
+	DML_LOG_VERBOSE("DML::%s: sliceWidth: %u\n", __func__, sliceWidth);
+	DML_LOG_VERBOSE("DML::%s: numSlices: %u\n", __func__, numSlices);
+	DML_LOG_VERBOSE("DML::%s: pixelFormat: %u\n", __func__, pixelFormat);
+	DML_LOG_VERBOSE("DML::%s: Output: %u\n", __func__, Output);
+	DML_LOG_VERBOSE("DML::%s: pixels: %u\n", __func__, pixels);
+	return pixels;
+}
+
+//updated in dcn4
+static unsigned int dscComputeDelay(enum dml2_output_format_class pixelFormat, enum dml2_output_encoder_class Output)
+{
+	unsigned int Delay = 0;
+	unsigned int dispclk_per_dscclk = 3;
+
+	// sfr
+	Delay = Delay + 2;
+
+	if (pixelFormat == dml2_420 || pixelFormat == dml2_n422 || (Output == dml2_hdmifrl && pixelFormat != dml2_444)) {
+		dispclk_per_dscclk = 3 * 2;
+	}
+
+	if (pixelFormat == dml2_420) {
+		//dscc top delay for pixel compression layer
+		Delay = Delay + 16 * dispclk_per_dscclk;
+
+		// dscc - input deserializer
+		Delay = Delay + 5;
+
+		// dscc - input cdc fifo
+		Delay = Delay + 1 + 4 * dispclk_per_dscclk;
+
+		// dscc - output cdc fifo
+		Delay = Delay + 3 + 1 * dispclk_per_dscclk;
+
+		// dscc - cdc uncertainty
+		Delay = Delay + 3 + 3 * dispclk_per_dscclk;
+	} else if (pixelFormat == dml2_n422 || (Output == dml2_hdmifrl && pixelFormat != dml2_444)) {
+		//dscc top delay for pixel compression layer
+		Delay = Delay + 16 * dispclk_per_dscclk;
+		// dsccif
+		Delay = Delay + 1;
+		// dscc - input deserializer
+		Delay = Delay + 5;
+		// dscc - input cdc fifo
+		Delay = Delay + 1 + 4 * dispclk_per_dscclk;
+
+
+		// dscc - output cdc fifo
+		Delay = Delay + 3 + 1 * dispclk_per_dscclk;
+		// dscc - cdc uncertainty
+		Delay = Delay + 3 + 3 * dispclk_per_dscclk;
+	} else if (pixelFormat == dml2_s422) {
+		//dscc top delay for pixel compression layer
+		Delay = Delay + 17 * dispclk_per_dscclk;
+
+		// dscc - input deserializer
+		Delay = Delay + 3;
+		// dscc - input cdc fifo
+		Delay = Delay + 1 + 4 * dispclk_per_dscclk;
+		// dscc - output cdc fifo
+		Delay = Delay + 3 + 1 * dispclk_per_dscclk;
+		// dscc - cdc uncertainty
+		Delay = Delay + 3 + 3 * dispclk_per_dscclk;
+	} else {
+		//dscc top delay for pixel compression layer
+		Delay = Delay + 16 * dispclk_per_dscclk;
+		// dscc - input deserializer
+		Delay = Delay + 3;
+		// dscc - input cdc fifo
+		Delay = Delay + 1 + 4 * dispclk_per_dscclk;
+		// dscc - output cdc fifo
+		Delay = Delay + 3 + 1 * dispclk_per_dscclk;
+
+		// dscc - cdc uncertainty
+		Delay = Delay + 3 + 3 * dispclk_per_dscclk;
+	}
+
+	// sft
+	Delay = Delay + 1;
+	DML_LOG_VERBOSE("DML::%s: pixelFormat = %u\n", __func__, pixelFormat);
+	DML_LOG_VERBOSE("DML::%s: Delay = %u\n", __func__, Delay);
+
+	return Delay;
+}
+
+static unsigned int CalculateHostVMDynamicLevels(
+	bool GPUVMEnable,
+	bool HostVMEnable,
+	unsigned int HostVMMinPageSize,
+	unsigned int HostVMMaxNonCachedPageTableLevels)
+{
+	unsigned int HostVMDynamicLevels = 0;
+
+	if (GPUVMEnable && HostVMEnable) {
+		if (HostVMMinPageSize < 2048)
+			HostVMDynamicLevels = HostVMMaxNonCachedPageTableLevels;
+		else if (HostVMMinPageSize >= 2048 && HostVMMinPageSize < 1048576)
+			HostVMDynamicLevels = (unsigned int)math_max2(0, (double)HostVMMaxNonCachedPageTableLevels - 1);
+		else
+			HostVMDynamicLevels = (unsigned int)math_max2(0, (double)HostVMMaxNonCachedPageTableLevels - 2);
+	} else {
+		HostVMDynamicLevels = 0;
+	}
+	return HostVMDynamicLevels;
+}
+
+static unsigned int CalculateVMAndRowBytes(struct dml2_core_shared_calculate_vm_and_row_bytes_params *p)
+{
+	unsigned int extra_dpde_bytes;
+	unsigned int extra_mpde_bytes;
+	unsigned int MacroTileSizeBytes;
+	unsigned int vp_height_dpte_ub;
+
+	unsigned int meta_surface_bytes;
+	unsigned int vm_bytes;
+	unsigned int vp_height_meta_ub;
+	unsigned int PixelPTEReqWidth_linear = 0; // VBA_DELTA. VBA doesn't calculate this
+
+	*p->MetaRequestHeight = 8 * p->BlockHeight256Bytes;
+	*p->MetaRequestWidth = 8 * p->BlockWidth256Bytes;
+	if (p->SurfaceTiling == dml2_sw_linear) {
+		*p->meta_row_height = 32;
+		*p->meta_row_width = (unsigned int)(math_floor2(p->ViewportXStart + p->SwathWidth + *p->MetaRequestWidth - 1, *p->MetaRequestWidth) - math_floor2(p->ViewportXStart, *p->MetaRequestWidth));
+		*p->meta_row_bytes = (unsigned int)(*p->meta_row_width * *p->MetaRequestHeight * p->BytePerPixel / 256.0); // FIXME_DCN4SW missing in old code but no dcc for linear anyways?
+	} else if (!dml_is_vertical_rotation(p->RotationAngle)) {
+		*p->meta_row_height = *p->MetaRequestHeight;
+		if (p->ViewportStationary && p->NumberOfDPPs == 1) {
+			*p->meta_row_width = (unsigned int)(math_floor2(p->ViewportXStart + p->SwathWidth + *p->MetaRequestWidth - 1, *p->MetaRequestWidth) - math_floor2(p->ViewportXStart, *p->MetaRequestWidth));
+		} else {
+			*p->meta_row_width = (unsigned int)(math_ceil2(p->SwathWidth - 1, *p->MetaRequestWidth) + *p->MetaRequestWidth);
+		}
+		*p->meta_row_bytes = (unsigned int)(*p->meta_row_width * *p->MetaRequestHeight * p->BytePerPixel / 256.0);
+	} else {
+		*p->meta_row_height = *p->MetaRequestWidth;
+		if (p->ViewportStationary && p->NumberOfDPPs == 1) {
+			*p->meta_row_width = (unsigned int)(math_floor2(p->ViewportYStart + p->ViewportHeight + *p->MetaRequestHeight - 1, *p->MetaRequestHeight) - math_floor2(p->ViewportYStart, *p->MetaRequestHeight));
+		} else {
+			*p->meta_row_width = (unsigned int)(math_ceil2(p->SwathWidth - 1, *p->MetaRequestHeight) + *p->MetaRequestHeight);
+		}
+		*p->meta_row_bytes = (unsigned int)(*p->meta_row_width * *p->MetaRequestWidth * p->BytePerPixel / 256.0);
+	}
+
+	if (p->ViewportStationary && p->is_phantom && (p->NumberOfDPPs == 1 || !dml_is_vertical_rotation(p->RotationAngle))) {
+		vp_height_meta_ub = (unsigned int)(math_floor2(p->ViewportYStart + p->ViewportHeight + 64 * p->BlockHeight256Bytes - 1, 64 * p->BlockHeight256Bytes) - math_floor2(p->ViewportYStart, 64 * p->BlockHeight256Bytes));
+	} else if (!dml_is_vertical_rotation(p->RotationAngle)) {
+		vp_height_meta_ub = (unsigned int)(math_ceil2(p->ViewportHeight - 1, 64 * p->BlockHeight256Bytes) + 64 * p->BlockHeight256Bytes);
+	} else {
+		vp_height_meta_ub = (unsigned int)(math_ceil2(p->SwathWidth - 1, 64 * p->BlockHeight256Bytes) + 64 * p->BlockHeight256Bytes);
+	}
+
+	meta_surface_bytes = (unsigned int)(p->DCCMetaPitch * vp_height_meta_ub * p->BytePerPixel / 256.0);
+	DML_LOG_VERBOSE("DML::%s: DCCMetaPitch = %u\n", __func__, p->DCCMetaPitch);
+	DML_LOG_VERBOSE("DML::%s: meta_surface_bytes = %u\n", __func__, meta_surface_bytes);
+	if (p->GPUVMEnable == true) {
+		double meta_vmpg_bytes = 4.0 * 1024.0;
+		*p->meta_pte_bytes_per_frame_ub = (unsigned int)((math_ceil2((double) (meta_surface_bytes - meta_vmpg_bytes) / (8 * meta_vmpg_bytes), 1) + 1) * 64);
+		extra_mpde_bytes = 128 * (p->GPUVMMaxPageTableLevels - 1);
+	} else {
+		*p->meta_pte_bytes_per_frame_ub = 0;
+		extra_mpde_bytes = 0;
+	}
+
+	if (!p->DCCEnable || !p->mrq_present) {
+		*p->meta_pte_bytes_per_frame_ub = 0;
+		extra_mpde_bytes = 0;
+		*p->meta_row_bytes = 0;
+	}
+
+	if (!p->GPUVMEnable) {
+		*p->PixelPTEBytesPerRow = 0;
+		*p->PixelPTEBytesPerRowStorage = 0;
+		*p->dpte_row_width_ub = 0;
+		*p->dpte_row_height = 0;
+		*p->dpte_row_height_linear = 0;
+		*p->PixelPTEBytesPerRow_one_row_per_frame = 0;
+		*p->dpte_row_width_ub_one_row_per_frame = 0;
+		*p->dpte_row_height_one_row_per_frame = 0;
+		*p->vmpg_width = 0;
+		*p->vmpg_height = 0;
+		*p->PixelPTEReqWidth = 0;
+		*p->PixelPTEReqHeight = 0;
+		*p->PTERequestSize = 0;
+		*p->dpde0_bytes_per_frame_ub = 0;
+		return 0;
+	}
+
+	MacroTileSizeBytes = p->MacroTileWidth * p->BytePerPixel * p->MacroTileHeight;
+
+	if (p->ViewportStationary && p->is_phantom && (p->NumberOfDPPs == 1 || !dml_is_vertical_rotation(p->RotationAngle))) {
+		vp_height_dpte_ub = (unsigned int)(math_floor2(p->ViewportYStart + p->ViewportHeight + p->MacroTileHeight - 1, p->MacroTileHeight) - math_floor2(p->ViewportYStart, p->MacroTileHeight));
+	} else if (!dml_is_vertical_rotation(p->RotationAngle)) {
+		vp_height_dpte_ub = (unsigned int)(math_ceil2((double)p->ViewportHeight - 1, p->MacroTileHeight) + p->MacroTileHeight);
+	} else {
+		vp_height_dpte_ub = (unsigned int)(math_ceil2((double)p->SwathWidth - 1, p->MacroTileHeight) + p->MacroTileHeight);
+	}
+
+	if (p->GPUVMEnable == true && p->GPUVMMaxPageTableLevels > 1) {
+		*p->dpde0_bytes_per_frame_ub = (unsigned int)(64 * (math_ceil2((double)(p->Pitch * vp_height_dpte_ub * p->BytePerPixel - MacroTileSizeBytes) / (double)(8 * 2097152), 1) + 1));
+		extra_dpde_bytes = 128 * (p->GPUVMMaxPageTableLevels - 2);
+	} else {
+		*p->dpde0_bytes_per_frame_ub = 0;
+		extra_dpde_bytes = 0;
+	}
+
+	vm_bytes = *p->meta_pte_bytes_per_frame_ub + extra_mpde_bytes + *p->dpde0_bytes_per_frame_ub + extra_dpde_bytes;
+
+	DML_LOG_VERBOSE("DML::%s: DCCEnable = %u\n", __func__, p->DCCEnable);
+	DML_LOG_VERBOSE("DML::%s: GPUVMEnable = %u\n", __func__, p->GPUVMEnable);
+	DML_LOG_VERBOSE("DML::%s: SwModeLinear = %u\n", __func__, p->SurfaceTiling == dml2_sw_linear);
+	DML_LOG_VERBOSE("DML::%s: BytePerPixel = %u\n", __func__, p->BytePerPixel);
+	DML_LOG_VERBOSE("DML::%s: GPUVMMaxPageTableLevels = %u\n", __func__, p->GPUVMMaxPageTableLevels);
+	DML_LOG_VERBOSE("DML::%s: BlockHeight256Bytes = %u\n", __func__, p->BlockHeight256Bytes);
+	DML_LOG_VERBOSE("DML::%s: BlockWidth256Bytes = %u\n", __func__, p->BlockWidth256Bytes);
+	DML_LOG_VERBOSE("DML::%s: MacroTileHeight = %u\n", __func__, p->MacroTileHeight);
+	DML_LOG_VERBOSE("DML::%s: MacroTileWidth = %u\n", __func__, p->MacroTileWidth);
+	DML_LOG_VERBOSE("DML::%s: meta_pte_bytes_per_frame_ub = %u\n", __func__, *p->meta_pte_bytes_per_frame_ub);
+	DML_LOG_VERBOSE("DML::%s: dpde0_bytes_per_frame_ub = %u\n", __func__, *p->dpde0_bytes_per_frame_ub);
+	DML_LOG_VERBOSE("DML::%s: extra_mpde_bytes = %u\n", __func__, extra_mpde_bytes);
+	DML_LOG_VERBOSE("DML::%s: extra_dpde_bytes = %u\n", __func__, extra_dpde_bytes);
+	DML_LOG_VERBOSE("DML::%s: vm_bytes = %u\n", __func__, vm_bytes);
+	DML_LOG_VERBOSE("DML::%s: ViewportHeight = %u\n", __func__, p->ViewportHeight);
+	DML_LOG_VERBOSE("DML::%s: SwathWidth = %u\n", __func__, p->SwathWidth);
+	DML_LOG_VERBOSE("DML::%s: vp_height_dpte_ub = %u\n", __func__, vp_height_dpte_ub);
+
+	if (p->SurfaceTiling == dml2_sw_linear) {
+		*p->PixelPTEReqHeight = 1;
+		*p->PixelPTEReqWidth = p->GPUVMMinPageSizeKBytes * 1024 * 8 / p->BytePerPixel;
+		PixelPTEReqWidth_linear = p->GPUVMMinPageSizeKBytes * 1024 * 8 / p->BytePerPixel;
+		*p->PTERequestSize = 64;
+
+		*p->vmpg_height = 1;
+		*p->vmpg_width = p->GPUVMMinPageSizeKBytes * 1024 / p->BytePerPixel;
+	} else if (p->GPUVMMinPageSizeKBytes * 1024 >= dml_get_tile_block_size_bytes(p->SurfaceTiling)) { // 1 64B 8x1 PTE
+		*p->PixelPTEReqHeight = p->MacroTileHeight;
+		*p->PixelPTEReqWidth = 8 * 1024 * p->GPUVMMinPageSizeKBytes / (p->MacroTileHeight * p->BytePerPixel);
+		*p->PTERequestSize = 64;
+
+		*p->vmpg_height = p->MacroTileHeight;
+		*p->vmpg_width = 1024 * p->GPUVMMinPageSizeKBytes / (p->MacroTileHeight * p->BytePerPixel);
+
+	} else if (p->GPUVMMinPageSizeKBytes == 4 && dml_get_tile_block_size_bytes(p->SurfaceTiling) == 65536) { // 2 64B PTE requests to get 16 PTEs to cover the 64K tile
+		// one 64KB tile, is 16x16x256B req
+		*p->PixelPTEReqHeight = 16 * p->BlockHeight256Bytes;
+		*p->PixelPTEReqWidth = 16 * p->BlockWidth256Bytes;
+		*p->PTERequestSize = 128;
+
+		*p->vmpg_height = *p->PixelPTEReqHeight;
+		*p->vmpg_width = *p->PixelPTEReqWidth;
+	} else {
+		// default for rest of calculation to go through, when vm is disable, the calulated pte related values shouldnt be used anyways
+		*p->PixelPTEReqHeight = p->MacroTileHeight;
+		*p->PixelPTEReqWidth = 8 * 1024 * p->GPUVMMinPageSizeKBytes / (p->MacroTileHeight * p->BytePerPixel);
+		*p->PTERequestSize = 64;
+
+		*p->vmpg_height = p->MacroTileHeight;
+		*p->vmpg_width = 1024 * p->GPUVMMinPageSizeKBytes / (p->MacroTileHeight * p->BytePerPixel);
+
+		if (p->GPUVMEnable == true) {
+			DML_LOG_VERBOSE("DML::%s: GPUVMMinPageSizeKBytes=%u and sw_mode=%u (tile_size=%d) not supported!\n",
+				__func__, p->GPUVMMinPageSizeKBytes, p->SurfaceTiling, dml_get_tile_block_size_bytes(p->SurfaceTiling));
+			DML_ASSERT(0);
+		}
+	}
+
+	DML_LOG_VERBOSE("DML::%s: GPUVMMinPageSizeKBytes = %u\n", __func__, p->GPUVMMinPageSizeKBytes);
+	DML_LOG_VERBOSE("DML::%s: PixelPTEReqHeight = %u\n", __func__, *p->PixelPTEReqHeight);
+	DML_LOG_VERBOSE("DML::%s: PixelPTEReqWidth = %u\n", __func__, *p->PixelPTEReqWidth);
+	DML_LOG_VERBOSE("DML::%s: PixelPTEReqWidth_linear = %u\n", __func__, PixelPTEReqWidth_linear);
+	DML_LOG_VERBOSE("DML::%s: PTERequestSize = %u\n", __func__, *p->PTERequestSize);
+	DML_LOG_VERBOSE("DML::%s: Pitch = %u\n", __func__, p->Pitch);
+	DML_LOG_VERBOSE("DML::%s: vmpg_width = %u\n", __func__, *p->vmpg_width);
+	DML_LOG_VERBOSE("DML::%s: vmpg_height = %u\n", __func__, *p->vmpg_height);
+
+	*p->dpte_row_height_one_row_per_frame = vp_height_dpte_ub;
+	*p->dpte_row_width_ub_one_row_per_frame = (unsigned int)((math_ceil2(((double)p->Pitch * (double)*p->dpte_row_height_one_row_per_frame / (double)*p->PixelPTEReqHeight - 1) / (double)*p->PixelPTEReqWidth, 1) + 1) * (double)*p->PixelPTEReqWidth);
+	*p->PixelPTEBytesPerRow_one_row_per_frame = (unsigned int)((double)*p->dpte_row_width_ub_one_row_per_frame / (double)*p->PixelPTEReqWidth * *p->PTERequestSize);
+	*p->dpte_row_height_linear = 0;
+
+	if (p->SurfaceTiling == dml2_sw_linear) {
+		*p->dpte_row_height = (unsigned int)(math_min2(128, (double)(1ULL << (unsigned int)math_floor2(math_log((float)(p->PTEBufferSizeInRequests * *p->PixelPTEReqWidth / p->Pitch), 2.0), 1))));
+		*p->dpte_row_width_ub = (unsigned int)(math_ceil2(((double)p->Pitch * (double)*p->dpte_row_height - 1), (double)*p->PixelPTEReqWidth) + *p->PixelPTEReqWidth);
+		*p->PixelPTEBytesPerRow = (unsigned int)((double)*p->dpte_row_width_ub / (double)*p->PixelPTEReqWidth * *p->PTERequestSize);
+
+		// VBA_DELTA, VBA doesn't have programming value for pte row height linear.
+		*p->dpte_row_height_linear = (unsigned int)1 << (unsigned int)math_floor2(math_log((float)(p->PTEBufferSizeInRequests * PixelPTEReqWidth_linear / p->Pitch), 2.0), 1);
+		if (*p->dpte_row_height_linear > 128)
+			*p->dpte_row_height_linear = 128;
+
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: dpte_row_width_ub = %u (linear)\n", __func__, *p->dpte_row_width_ub);
+#endif
+
+	} else if (!dml_is_vertical_rotation(p->RotationAngle)) {
+		*p->dpte_row_height = *p->PixelPTEReqHeight;
+
+		if (p->GPUVMMinPageSizeKBytes > 64) {
+			*p->dpte_row_width_ub = (unsigned int)((math_ceil2(((double)p->Pitch * (double)*p->dpte_row_height / (double)*p->PixelPTEReqHeight - 1) / (double)*p->PixelPTEReqWidth, 1) + 1) * *p->PixelPTEReqWidth);
+		} else if (p->ViewportStationary && (p->NumberOfDPPs == 1)) {
+			*p->dpte_row_width_ub = (unsigned int)(math_floor2(p->ViewportXStart + p->SwathWidth + *p->PixelPTEReqWidth - 1, *p->PixelPTEReqWidth) - math_floor2(p->ViewportXStart, *p->PixelPTEReqWidth));
+		} else {
+			*p->dpte_row_width_ub = (unsigned int)((math_ceil2((double)(p->SwathWidth - 1) / (double)*p->PixelPTEReqWidth, 1) + 1.0) * *p->PixelPTEReqWidth);
+		}
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: dpte_row_width_ub = %u (tiled horz)\n", __func__, *p->dpte_row_width_ub);
+#endif
+
+		*p->PixelPTEBytesPerRow = *p->dpte_row_width_ub / *p->PixelPTEReqWidth * *p->PTERequestSize;
+	} else {
+		*p->dpte_row_height = (unsigned int)(math_min2(*p->PixelPTEReqWidth, p->MacroTileWidth));
+
+		if (p->ViewportStationary && (p->NumberOfDPPs == 1)) {
+			*p->dpte_row_width_ub = (unsigned int)(math_floor2(p->ViewportYStart + p->ViewportHeight + *p->PixelPTEReqHeight - 1, *p->PixelPTEReqHeight) - math_floor2(p->ViewportYStart, *p->PixelPTEReqHeight));
+		} else {
+			*p->dpte_row_width_ub = (unsigned int)((math_ceil2((double)(p->SwathWidth - 1) / (double)*p->PixelPTEReqHeight, 1) + 1) * *p->PixelPTEReqHeight);
+		}
+
+		*p->PixelPTEBytesPerRow = (unsigned int)((double)*p->dpte_row_width_ub / (double)*p->PixelPTEReqHeight * *p->PTERequestSize);
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: dpte_row_width_ub = %u (tiled vert)\n", __func__, *p->dpte_row_width_ub);
+#endif
+	}
+
+	if (p->GPUVMEnable != true) {
+		*p->PixelPTEBytesPerRow = 0;
+		*p->PixelPTEBytesPerRow_one_row_per_frame = 0;
+	}
+
+	*p->PixelPTEBytesPerRowStorage = *p->PixelPTEBytesPerRow;
+
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: GPUVMMinPageSizeKBytes = %u\n", __func__, p->GPUVMMinPageSizeKBytes);
+	DML_LOG_VERBOSE("DML::%s: GPUVMEnable = %u\n", __func__, p->GPUVMEnable);
+	DML_LOG_VERBOSE("DML::%s: meta_row_height = %u\n", __func__, *p->meta_row_height);
+	DML_LOG_VERBOSE("DML::%s: dpte_row_height = %u\n", __func__, *p->dpte_row_height);
+	DML_LOG_VERBOSE("DML::%s: dpte_row_height_linear = %u\n", __func__, *p->dpte_row_height_linear);
+	DML_LOG_VERBOSE("DML::%s: dpte_row_width_ub = %u\n", __func__, *p->dpte_row_width_ub);
+	DML_LOG_VERBOSE("DML::%s: PixelPTEBytesPerRow = %u\n", __func__, *p->PixelPTEBytesPerRow);
+	DML_LOG_VERBOSE("DML::%s: PixelPTEBytesPerRowStorage = %u\n", __func__, *p->PixelPTEBytesPerRowStorage);
+	DML_LOG_VERBOSE("DML::%s: PTEBufferSizeInRequests = %u\n", __func__, p->PTEBufferSizeInRequests);
+	DML_LOG_VERBOSE("DML::%s: dpte_row_height_one_row_per_frame = %u\n", __func__, *p->dpte_row_height_one_row_per_frame);
+	DML_LOG_VERBOSE("DML::%s: dpte_row_width_ub_one_row_per_frame = %u\n", __func__, *p->dpte_row_width_ub_one_row_per_frame);
+	DML_LOG_VERBOSE("DML::%s: PixelPTEBytesPerRow_one_row_per_frame = %u\n", __func__, *p->PixelPTEBytesPerRow_one_row_per_frame);
+#endif
+
+	return vm_bytes;
+} // CalculateVMAndRowBytes
+
+static unsigned int CalculatePrefetchSourceLines(
+	double VRatio,
+	unsigned int VTaps,
+	bool Interlace,
+	bool ProgressiveToInterlaceUnitInOPP,
+	unsigned int SwathHeight,
+	enum dml2_rotation_angle RotationAngle,
+	bool mirrored,
+	bool ViewportStationary,
+	unsigned int SwathWidth,
+	unsigned int ViewportHeight,
+	unsigned int ViewportXStart,
+	unsigned int ViewportYStart,
+
+	// Output
+	unsigned int *VInitPreFill,
+	unsigned int *MaxNumSwath)
+{
+
+	unsigned int vp_start_rot = 0;
+	unsigned int sw0_tmp = 0;
+	unsigned int MaxPartialSwath = 0;
+	double numLines = 0;
+
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: VRatio = %f\n", __func__, VRatio);
+	DML_LOG_VERBOSE("DML::%s: VTaps = %u\n", __func__, VTaps);
+	DML_LOG_VERBOSE("DML::%s: ViewportXStart = %u\n", __func__, ViewportXStart);
+	DML_LOG_VERBOSE("DML::%s: ViewportYStart = %u\n", __func__, ViewportYStart);
+	DML_LOG_VERBOSE("DML::%s: ViewportStationary = %u\n", __func__, ViewportStationary);
+	DML_LOG_VERBOSE("DML::%s: SwathHeight = %u\n", __func__, SwathHeight);
+#endif
+	if (ProgressiveToInterlaceUnitInOPP)
+		*VInitPreFill = (unsigned int)(math_floor2((VRatio + (double)VTaps + 1) / 2.0, 1));
+	else
+		*VInitPreFill = (unsigned int)(math_floor2((VRatio + (double)VTaps + 1 + (Interlace ? 1 : 0) * 0.5 * VRatio) / 2.0, 1));
+
+	if (ViewportStationary) {
+		if (RotationAngle == dml2_rotation_180) {
+			vp_start_rot = SwathHeight - (((unsigned int)(ViewportYStart + ViewportHeight - 1) % SwathHeight) + 1);
+		} else if ((RotationAngle == dml2_rotation_270 && !mirrored) || (RotationAngle == dml2_rotation_90 && mirrored)) {
+			vp_start_rot = ViewportXStart;
+		} else if ((RotationAngle == dml2_rotation_90 && !mirrored) || (RotationAngle == dml2_rotation_270 && mirrored)) {
+			vp_start_rot = SwathHeight - (((unsigned int)(ViewportYStart + SwathWidth - 1) % SwathHeight) + 1);
+		} else {
+			vp_start_rot = ViewportYStart;
+		}
+		sw0_tmp = SwathHeight - (vp_start_rot % SwathHeight);
+		if (sw0_tmp < *VInitPreFill) {
+			*MaxNumSwath = (unsigned int)(math_ceil2((*VInitPreFill - sw0_tmp) / (double)SwathHeight, 1) + 1);
+		} else {
+			*MaxNumSwath = 1;
+		}
+		MaxPartialSwath = (unsigned int)(math_max2(1, (unsigned int)(vp_start_rot + *VInitPreFill - 1) % SwathHeight));
+	} else {
+		*MaxNumSwath = (unsigned int)(math_ceil2((*VInitPreFill - 1.0) / (double)SwathHeight, 1) + 1);
+		if (*VInitPreFill > 1) {
+			MaxPartialSwath = (unsigned int)(math_max2(1, (unsigned int)(*VInitPreFill - 2) % SwathHeight));
+		} else {
+			MaxPartialSwath = (unsigned int)(math_max2(1, (unsigned int)(*VInitPreFill + SwathHeight - 2) % SwathHeight));
+		}
+	}
+	numLines = *MaxNumSwath * SwathHeight + MaxPartialSwath;
+
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: vp_start_rot = %u\n", __func__, vp_start_rot);
+	DML_LOG_VERBOSE("DML::%s: VInitPreFill = %u\n", __func__, *VInitPreFill);
+	DML_LOG_VERBOSE("DML::%s: MaxPartialSwath = %u\n", __func__, MaxPartialSwath);
+	DML_LOG_VERBOSE("DML::%s: MaxNumSwath = %u\n", __func__, *MaxNumSwath);
+	DML_LOG_VERBOSE("DML::%s: Prefetch source lines = %3.2f\n", __func__, numLines);
+#endif
+	return (unsigned int)(numLines);
+
+}
+
+static void CalculateRowBandwidth(
+	bool GPUVMEnable,
+	bool use_one_row_for_frame,
+	enum dml2_source_format_class SourcePixelFormat,
+	double VRatio,
+	double VRatioChroma,
+	bool DCCEnable,
+	double LineTime,
+	unsigned int PixelPTEBytesPerRowLuma,
+	unsigned int PixelPTEBytesPerRowChroma,
+	unsigned int dpte_row_height_luma,
+	unsigned int dpte_row_height_chroma,
+
+	bool mrq_present,
+	unsigned int meta_row_bytes_per_row_ub_l,
+	unsigned int meta_row_bytes_per_row_ub_c,
+	unsigned int meta_row_height_luma,
+	unsigned int meta_row_height_chroma,
+
+	// Output
+	double *dpte_row_bw,
+	double *meta_row_bw)
+{
+	if (!DCCEnable || !mrq_present) {
+		*meta_row_bw = 0;
+	} else if (dml_is_420(SourcePixelFormat) || SourcePixelFormat == dml2_rgbe_alpha) {
+		*meta_row_bw = VRatio * meta_row_bytes_per_row_ub_l / (meta_row_height_luma * LineTime)
+				+ VRatioChroma * meta_row_bytes_per_row_ub_c / (meta_row_height_chroma * LineTime);
+	} else {
+		*meta_row_bw = VRatio * meta_row_bytes_per_row_ub_l / (meta_row_height_luma * LineTime);
+	}
+
+	if (GPUVMEnable != true) {
+		*dpte_row_bw = 0;
+	} else if (dml_is_420(SourcePixelFormat) || SourcePixelFormat == dml2_rgbe_alpha) {
+		*dpte_row_bw = VRatio * PixelPTEBytesPerRowLuma / (dpte_row_height_luma * LineTime)
+			+ VRatioChroma * PixelPTEBytesPerRowChroma / (dpte_row_height_chroma * LineTime);
+	} else {
+		*dpte_row_bw = VRatio * PixelPTEBytesPerRowLuma / (dpte_row_height_luma * LineTime);
+	}
+}
+
+static void CalculateMALLUseForStaticScreen(
+	const struct dml2_display_cfg *display_cfg,
+	unsigned int NumberOfActiveSurfaces,
+	unsigned int MALLAllocatedForDCN,
+	unsigned int SurfaceSizeInMALL[],
+	bool one_row_per_frame_fits_in_buffer[],
+
+	// Output
+	bool is_using_mall_for_ss[])
+{
+
+	unsigned int SurfaceToAddToMALL;
+	bool CanAddAnotherSurfaceToMALL;
+	unsigned int TotalSurfaceSizeInMALL;
+
+	TotalSurfaceSizeInMALL = 0;
+	for (unsigned int k = 0; k < NumberOfActiveSurfaces; ++k) {
+		is_using_mall_for_ss[k] = (display_cfg->plane_descriptors[k].overrides.refresh_from_mall == dml2_refresh_from_mall_mode_override_force_enable);
+		if (is_using_mall_for_ss[k])
+			TotalSurfaceSizeInMALL = TotalSurfaceSizeInMALL + SurfaceSizeInMALL[k];
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: k=%u, is_using_mall_for_ss = %u\n", __func__, k, is_using_mall_for_ss[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u, TotalSurfaceSizeInMALL = %u\n", __func__, k, TotalSurfaceSizeInMALL);
+#endif
+	}
+
+	SurfaceToAddToMALL = 0;
+	CanAddAnotherSurfaceToMALL = true;
+	while (CanAddAnotherSurfaceToMALL) {
+		CanAddAnotherSurfaceToMALL = false;
+		for (unsigned int k = 0; k < NumberOfActiveSurfaces; ++k) {
+			if (TotalSurfaceSizeInMALL + SurfaceSizeInMALL[k] <= MALLAllocatedForDCN * 1024 * 1024 &&
+				!is_using_mall_for_ss[k] && display_cfg->plane_descriptors[k].overrides.refresh_from_mall != dml2_refresh_from_mall_mode_override_force_disable && one_row_per_frame_fits_in_buffer[k] &&
+				(!CanAddAnotherSurfaceToMALL || SurfaceSizeInMALL[k] < SurfaceSizeInMALL[SurfaceToAddToMALL])) {
+				CanAddAnotherSurfaceToMALL = true;
+				SurfaceToAddToMALL = k;
+				DML_LOG_VERBOSE("DML::%s: k=%u, UseMALLForStaticScreen = %u (dis, en, optimize)\n", __func__, k, display_cfg->plane_descriptors[k].overrides.refresh_from_mall);
+			}
+		}
+		if (CanAddAnotherSurfaceToMALL) {
+			is_using_mall_for_ss[SurfaceToAddToMALL] = true;
+			TotalSurfaceSizeInMALL = TotalSurfaceSizeInMALL + SurfaceSizeInMALL[SurfaceToAddToMALL];
+
+#ifdef __DML_VBA_DEBUG__
+			DML_LOG_VERBOSE("DML::%s: SurfaceToAddToMALL = %u\n", __func__, SurfaceToAddToMALL);
+			DML_LOG_VERBOSE("DML::%s: TotalSurfaceSizeInMALL = %u\n", __func__, TotalSurfaceSizeInMALL);
+#endif
+		}
+	}
+}
+
+static void CalculateDCCConfiguration(
+	bool DCCEnabled,
+	bool DCCProgrammingAssumesScanDirectionUnknown,
+	enum dml2_source_format_class SourcePixelFormat,
+	unsigned int SurfaceWidthLuma,
+	unsigned int SurfaceWidthChroma,
+	unsigned int SurfaceHeightLuma,
+	unsigned int SurfaceHeightChroma,
+	unsigned int nomDETInKByte,
+	unsigned int RequestHeight256ByteLuma,
+	unsigned int RequestHeight256ByteChroma,
+	enum dml2_swizzle_mode TilingFormat,
+	unsigned int BytePerPixelY,
+	unsigned int BytePerPixelC,
+	double BytePerPixelDETY,
+	double BytePerPixelDETC,
+	enum dml2_rotation_angle RotationAngle,
+
+	// Output
+	enum dml2_core_internal_request_type *RequestLuma,
+	enum dml2_core_internal_request_type *RequestChroma,
+	unsigned int *MaxUncompressedBlockLuma,
+	unsigned int *MaxUncompressedBlockChroma,
+	unsigned int *MaxCompressedBlockLuma,
+	unsigned int *MaxCompressedBlockChroma,
+	unsigned int *IndependentBlockLuma,
+	unsigned int *IndependentBlockChroma)
+{
+	unsigned int DETBufferSizeForDCC = nomDETInKByte * 1024;
+
+	unsigned int segment_order_horz_contiguous_luma;
+	unsigned int segment_order_horz_contiguous_chroma;
+	unsigned int segment_order_vert_contiguous_luma;
+	unsigned int segment_order_vert_contiguous_chroma;
+
+	unsigned int req128_horz_wc_l;
+	unsigned int req128_horz_wc_c;
+	unsigned int req128_vert_wc_l;
+	unsigned int req128_vert_wc_c;
+
+	unsigned int yuv420;
+	unsigned int horz_div_l;
+	unsigned int horz_div_c;
+	unsigned int vert_div_l;
+	unsigned int vert_div_c;
+
+	unsigned int swath_buf_size;
+	double detile_buf_vp_horz_limit;
+	double detile_buf_vp_vert_limit;
+
+	unsigned int MAS_vp_horz_limit;
+	unsigned int MAS_vp_vert_limit;
+	unsigned int max_vp_horz_width;
+	unsigned int max_vp_vert_height;
+	unsigned int eff_surf_width_l;
+	unsigned int eff_surf_width_c;
+	unsigned int eff_surf_height_l;
+	unsigned int eff_surf_height_c;
+
+	unsigned int full_swath_bytes_horz_wc_l;
+	unsigned int full_swath_bytes_horz_wc_c;
+	unsigned int full_swath_bytes_vert_wc_l;
+	unsigned int full_swath_bytes_vert_wc_c;
+
+	if (dml_is_420(SourcePixelFormat))
+		yuv420 = 1;
+	else
+		yuv420 = 0;
+	horz_div_l = 1;
+	horz_div_c = 1;
+	vert_div_l = 1;
+	vert_div_c = 1;
+
+	if (BytePerPixelY == 1)
+		vert_div_l = 0;
+	if (BytePerPixelC == 1)
+		vert_div_c = 0;
+
+	if (BytePerPixelC == 0) {
+		swath_buf_size = DETBufferSizeForDCC / 2 - 2 * 256;
+		detile_buf_vp_horz_limit = (double)swath_buf_size / ((double)RequestHeight256ByteLuma * BytePerPixelY / (1 + horz_div_l));
+		detile_buf_vp_vert_limit = (double)swath_buf_size / (256.0 / RequestHeight256ByteLuma / (1 + vert_div_l));
+	} else {
+		swath_buf_size = DETBufferSizeForDCC / 2 - 2 * 2 * 256;
+		detile_buf_vp_horz_limit = (double)swath_buf_size / ((double)RequestHeight256ByteLuma * BytePerPixelY / (1 + horz_div_l) + (double)RequestHeight256ByteChroma * BytePerPixelC / (1 + horz_div_c) / (1 + yuv420));
+		detile_buf_vp_vert_limit = (double)swath_buf_size / (256.0 / RequestHeight256ByteLuma / (1 + vert_div_l) + 256.0 / RequestHeight256ByteChroma / (1 + vert_div_c) / (1 + yuv420));
+	}
+
+	if (SourcePixelFormat == dml2_420_10) {
+		detile_buf_vp_horz_limit = 1.5 * detile_buf_vp_horz_limit;
+		detile_buf_vp_vert_limit = 1.5 * detile_buf_vp_vert_limit;
+	}
+
+	detile_buf_vp_horz_limit = math_floor2(detile_buf_vp_horz_limit - 1, 16);
+	detile_buf_vp_vert_limit = math_floor2(detile_buf_vp_vert_limit - 1, 16);
+
+	MAS_vp_horz_limit = SourcePixelFormat == dml2_rgbe_alpha ? 3840 : 6144;
+	MAS_vp_vert_limit = SourcePixelFormat == dml2_rgbe_alpha ? 3840 : (BytePerPixelY == 8 ? 3072 : 6144);
+	max_vp_horz_width = (unsigned int)(math_min2((double)MAS_vp_horz_limit, detile_buf_vp_horz_limit));
+	max_vp_vert_height = (unsigned int)(math_min2((double)MAS_vp_vert_limit, detile_buf_vp_vert_limit));
+	eff_surf_width_l = (SurfaceWidthLuma > max_vp_horz_width ? max_vp_horz_width : SurfaceWidthLuma);
+	eff_surf_width_c = eff_surf_width_l / (1 + yuv420);
+	eff_surf_height_l = (SurfaceHeightLuma > max_vp_vert_height ? max_vp_vert_height : SurfaceHeightLuma);
+	eff_surf_height_c = eff_surf_height_l / (1 + yuv420);
+
+	full_swath_bytes_horz_wc_l = eff_surf_width_l * RequestHeight256ByteLuma * BytePerPixelY;
+	full_swath_bytes_vert_wc_l = eff_surf_height_l * 256 / RequestHeight256ByteLuma;
+	if (BytePerPixelC > 0) {
+		full_swath_bytes_horz_wc_c = eff_surf_width_c * RequestHeight256ByteChroma * BytePerPixelC;
+		full_swath_bytes_vert_wc_c = eff_surf_height_c * 256 / RequestHeight256ByteChroma;
+	} else {
+		full_swath_bytes_horz_wc_c = 0;
+		full_swath_bytes_vert_wc_c = 0;
+	}
+
+	if (SourcePixelFormat == dml2_420_10) {
+		full_swath_bytes_horz_wc_l = (unsigned int)(math_ceil2((double)full_swath_bytes_horz_wc_l * 2.0 / 3.0, 256.0));
+		full_swath_bytes_horz_wc_c = (unsigned int)(math_ceil2((double)full_swath_bytes_horz_wc_c * 2.0 / 3.0, 256.0));
+		full_swath_bytes_vert_wc_l = (unsigned int)(math_ceil2((double)full_swath_bytes_vert_wc_l * 2.0 / 3.0, 256.0));
+		full_swath_bytes_vert_wc_c = (unsigned int)(math_ceil2((double)full_swath_bytes_vert_wc_c * 2.0 / 3.0, 256.0));
+	}
+
+	if (2 * full_swath_bytes_horz_wc_l + 2 * full_swath_bytes_horz_wc_c <= DETBufferSizeForDCC) {
+		req128_horz_wc_l = 0;
+		req128_horz_wc_c = 0;
+	} else if (full_swath_bytes_horz_wc_l < 1.5 * full_swath_bytes_horz_wc_c && 2 * full_swath_bytes_horz_wc_l + full_swath_bytes_horz_wc_c <= DETBufferSizeForDCC) {
+		req128_horz_wc_l = 0;
+		req128_horz_wc_c = 1;
+	} else if (full_swath_bytes_horz_wc_l >= 1.5 * full_swath_bytes_horz_wc_c && full_swath_bytes_horz_wc_l + 2 * full_swath_bytes_horz_wc_c <= DETBufferSizeForDCC) {
+		req128_horz_wc_l = 1;
+		req128_horz_wc_c = 0;
+	} else {
+		req128_horz_wc_l = 1;
+		req128_horz_wc_c = 1;
+	}
+
+	if (2 * full_swath_bytes_vert_wc_l + 2 * full_swath_bytes_vert_wc_c <= DETBufferSizeForDCC) {
+		req128_vert_wc_l = 0;
+		req128_vert_wc_c = 0;
+	} else if (full_swath_bytes_vert_wc_l < 1.5 * full_swath_bytes_vert_wc_c && 2 * full_swath_bytes_vert_wc_l + full_swath_bytes_vert_wc_c <= DETBufferSizeForDCC) {
+		req128_vert_wc_l = 0;
+		req128_vert_wc_c = 1;
+	} else if (full_swath_bytes_vert_wc_l >= 1.5 * full_swath_bytes_vert_wc_c && full_swath_bytes_vert_wc_l + 2 * full_swath_bytes_vert_wc_c <= DETBufferSizeForDCC) {
+		req128_vert_wc_l = 1;
+		req128_vert_wc_c = 0;
+	} else {
+		req128_vert_wc_l = 1;
+		req128_vert_wc_c = 1;
+	}
+
+	if (BytePerPixelY == 2) {
+		segment_order_horz_contiguous_luma = 0;
+		segment_order_vert_contiguous_luma = 1;
+	} else {
+		segment_order_horz_contiguous_luma = 1;
+		segment_order_vert_contiguous_luma = 0;
+	}
+
+	if (BytePerPixelC == 2) {
+		segment_order_horz_contiguous_chroma = 0;
+		segment_order_vert_contiguous_chroma = 1;
+	} else {
+		segment_order_horz_contiguous_chroma = 1;
+		segment_order_vert_contiguous_chroma = 0;
+	}
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: DCCEnabled = %u\n", __func__, DCCEnabled);
+	DML_LOG_VERBOSE("DML::%s: nomDETInKByte = %u\n", __func__, nomDETInKByte);
+	DML_LOG_VERBOSE("DML::%s: DETBufferSizeForDCC = %u\n", __func__, DETBufferSizeForDCC);
+	DML_LOG_VERBOSE("DML::%s: req128_horz_wc_l = %u\n", __func__, req128_horz_wc_l);
+	DML_LOG_VERBOSE("DML::%s: req128_horz_wc_c = %u\n", __func__, req128_horz_wc_c);
+	DML_LOG_VERBOSE("DML::%s: full_swath_bytes_horz_wc_l = %u\n", __func__, full_swath_bytes_horz_wc_l);
+	DML_LOG_VERBOSE("DML::%s: full_swath_bytes_vert_wc_c = %u\n", __func__, full_swath_bytes_vert_wc_c);
+	DML_LOG_VERBOSE("DML::%s: segment_order_horz_contiguous_luma = %u\n", __func__, segment_order_horz_contiguous_luma);
+	DML_LOG_VERBOSE("DML::%s: segment_order_horz_contiguous_chroma = %u\n", __func__, segment_order_horz_contiguous_chroma);
+#endif
+	if (DCCProgrammingAssumesScanDirectionUnknown == true) {
+		if (req128_horz_wc_l == 0 && req128_vert_wc_l == 0) {
+			*RequestLuma = dml2_core_internal_request_type_256_bytes;
+		} else if ((req128_horz_wc_l == 1 && segment_order_horz_contiguous_luma == 0) || (req128_vert_wc_l == 1 && segment_order_vert_contiguous_luma == 0)) {
+			*RequestLuma = dml2_core_internal_request_type_128_bytes_non_contiguous;
+		} else {
+			*RequestLuma = dml2_core_internal_request_type_128_bytes_contiguous;
+		}
+		if (req128_horz_wc_c == 0 && req128_vert_wc_c == 0) {
+			*RequestChroma = dml2_core_internal_request_type_256_bytes;
+		} else if ((req128_horz_wc_c == 1 && segment_order_horz_contiguous_chroma == 0) || (req128_vert_wc_c == 1 && segment_order_vert_contiguous_chroma == 0)) {
+			*RequestChroma = dml2_core_internal_request_type_128_bytes_non_contiguous;
+		} else {
+			*RequestChroma = dml2_core_internal_request_type_128_bytes_contiguous;
+		}
+	} else if (!dml_is_vertical_rotation(RotationAngle)) {
+		if (req128_horz_wc_l == 0) {
+			*RequestLuma = dml2_core_internal_request_type_256_bytes;
+		} else if (segment_order_horz_contiguous_luma == 0) {
+			*RequestLuma = dml2_core_internal_request_type_128_bytes_non_contiguous;
+		} else {
+			*RequestLuma = dml2_core_internal_request_type_128_bytes_contiguous;
+		}
+		if (req128_horz_wc_c == 0) {
+			*RequestChroma = dml2_core_internal_request_type_256_bytes;
+		} else if (segment_order_horz_contiguous_chroma == 0) {
+			*RequestChroma = dml2_core_internal_request_type_128_bytes_non_contiguous;
+		} else {
+			*RequestChroma = dml2_core_internal_request_type_128_bytes_contiguous;
+		}
+	} else {
+		if (req128_vert_wc_l == 0) {
+			*RequestLuma = dml2_core_internal_request_type_256_bytes;
+		} else if (segment_order_vert_contiguous_luma == 0) {
+			*RequestLuma = dml2_core_internal_request_type_128_bytes_non_contiguous;
+		} else {
+			*RequestLuma = dml2_core_internal_request_type_128_bytes_contiguous;
+		}
+		if (req128_vert_wc_c == 0) {
+			*RequestChroma = dml2_core_internal_request_type_256_bytes;
+		} else if (segment_order_vert_contiguous_chroma == 0) {
+			*RequestChroma = dml2_core_internal_request_type_128_bytes_non_contiguous;
+		} else {
+			*RequestChroma = dml2_core_internal_request_type_128_bytes_contiguous;
+		}
+	}
+
+	if (*RequestLuma == dml2_core_internal_request_type_256_bytes) {
+		*MaxUncompressedBlockLuma = 256;
+		*MaxCompressedBlockLuma = 256;
+		*IndependentBlockLuma = 0;
+	} else if (*RequestLuma == dml2_core_internal_request_type_128_bytes_contiguous) {
+		*MaxUncompressedBlockLuma = 256;
+		*MaxCompressedBlockLuma = 128;
+		*IndependentBlockLuma = 128;
+	} else {
+		*MaxUncompressedBlockLuma = 256;
+		*MaxCompressedBlockLuma = 64;
+		*IndependentBlockLuma = 64;
+	}
+
+	if (*RequestChroma == dml2_core_internal_request_type_256_bytes) {
+		*MaxUncompressedBlockChroma = 256;
+		*MaxCompressedBlockChroma = 256;
+		*IndependentBlockChroma = 0;
+	} else if (*RequestChroma == dml2_core_internal_request_type_128_bytes_contiguous) {
+		*MaxUncompressedBlockChroma = 256;
+		*MaxCompressedBlockChroma = 128;
+		*IndependentBlockChroma = 128;
+	} else {
+		*MaxUncompressedBlockChroma = 256;
+		*MaxCompressedBlockChroma = 64;
+		*IndependentBlockChroma = 64;
+	}
+
+	if (DCCEnabled != true || BytePerPixelC == 0) {
+		*MaxUncompressedBlockChroma = 0;
+		*MaxCompressedBlockChroma = 0;
+		*IndependentBlockChroma = 0;
+	}
+
+	if (DCCEnabled != true) {
+		*MaxUncompressedBlockLuma = 0;
+		*MaxCompressedBlockLuma = 0;
+		*IndependentBlockLuma = 0;
+	}
+
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: MaxUncompressedBlockLuma = %u\n", __func__, *MaxUncompressedBlockLuma);
+	DML_LOG_VERBOSE("DML::%s: MaxCompressedBlockLuma = %u\n", __func__, *MaxCompressedBlockLuma);
+	DML_LOG_VERBOSE("DML::%s: IndependentBlockLuma = %u\n", __func__, *IndependentBlockLuma);
+	DML_LOG_VERBOSE("DML::%s: MaxUncompressedBlockChroma = %u\n", __func__, *MaxUncompressedBlockChroma);
+	DML_LOG_VERBOSE("DML::%s: MaxCompressedBlockChroma = %u\n", __func__, *MaxCompressedBlockChroma);
+	DML_LOG_VERBOSE("DML::%s: IndependentBlockChroma = %u\n", __func__, *IndependentBlockChroma);
+#endif
+
+}
+
+static void calculate_mcache_row_bytes(
+	struct dml2_core_internal_scratch *scratch,
+	struct dml2_core_calcs_calculate_mcache_row_bytes_params *p)
+{
+	unsigned int vmpg_bytes = 0;
+	unsigned int blk_bytes = 0;
+	float meta_per_mvmpg_per_channel = 0;
+	unsigned int est_blk_per_vmpg = 2;
+	unsigned int mvmpg_per_row_ub = 0;
+	unsigned int full_vp_width_mvmpg_aligned = 0;
+	unsigned int full_vp_height_mvmpg_aligned = 0;
+	unsigned int meta_per_mvmpg_per_channel_ub = 0;
+	unsigned int mvmpg_per_mcache;
+
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: num_chans = %u\n", __func__, p->num_chans);
+	DML_LOG_VERBOSE("DML::%s: mem_word_bytes = %u\n", __func__, p->mem_word_bytes);
+	DML_LOG_VERBOSE("DML::%s: mcache_line_size_bytes = %u\n", __func__, p->mcache_line_size_bytes);
+	DML_LOG_VERBOSE("DML::%s: mcache_size_bytes = %u\n", __func__, p->mcache_size_bytes);
+	DML_LOG_VERBOSE("DML::%s: gpuvm_enable = %u\n", __func__, p->gpuvm_enable);
+	DML_LOG_VERBOSE("DML::%s: gpuvm_page_size_kbytes = %u\n", __func__, p->gpuvm_page_size_kbytes);
+	DML_LOG_VERBOSE("DML::%s: vp_stationary = %u\n", __func__, p->vp_stationary);
+	DML_LOG_VERBOSE("DML::%s: tiling_mode = %u\n", __func__, p->tiling_mode);
+	DML_LOG_VERBOSE("DML::%s: vp_start_x = %u\n", __func__, p->vp_start_x);
+	DML_LOG_VERBOSE("DML::%s: vp_start_y = %u\n", __func__, p->vp_start_y);
+	DML_LOG_VERBOSE("DML::%s: full_vp_width = %u\n", __func__, p->full_vp_width);
+	DML_LOG_VERBOSE("DML::%s: full_vp_height = %u\n", __func__, p->full_vp_height);
+	DML_LOG_VERBOSE("DML::%s: blk_width = %u\n", __func__, p->blk_width);
+	DML_LOG_VERBOSE("DML::%s: blk_height = %u\n", __func__, p->blk_height);
+	DML_LOG_VERBOSE("DML::%s: vmpg_width = %u\n", __func__, p->vmpg_width);
+	DML_LOG_VERBOSE("DML::%s: vmpg_height = %u\n", __func__, p->vmpg_height);
+	DML_LOG_VERBOSE("DML::%s: full_swath_bytes = %u\n", __func__, p->full_swath_bytes);
+#endif
+	DML_ASSERT(p->mcache_line_size_bytes != 0);
+	DML_ASSERT(p->mcache_size_bytes != 0);
+
+	*p->mvmpg_width = 0;
+	*p->mvmpg_height = 0;
+
+	if (p->full_vp_height == 0 && p->full_vp_width == 0) {
+		*p->num_mcaches = 0;
+		*p->mcache_row_bytes = 0;
+		*p->mcache_row_bytes_per_channel = 0;
+	} else {
+		blk_bytes = dml_get_tile_block_size_bytes(p->tiling_mode);
+
+		// if gpuvm is not enable, the alignment boundary should be in terms of tiling block size
+		vmpg_bytes = p->gpuvm_page_size_kbytes * 1024;
+
+		//With vmpg_bytes >= tile blk_bytes, the meta_row_width alignment equations are relative to the vmpg_width/height.
+		// But for 4KB page with 64KB tile block, we need the meta for all pages in the tile block.
+		// Therefore, the alignment is relative to the blk_width/height. The factor of 16 vmpg per 64KB tile block is applied at the end.
+		*p->mvmpg_width = p->blk_width;
+		*p->mvmpg_height = p->blk_height;
+		if (p->gpuvm_enable) {
+			if (vmpg_bytes >= blk_bytes) {
+				*p->mvmpg_width = p->vmpg_width;
+				*p->mvmpg_height = p->vmpg_height;
+			} else if (!((blk_bytes == 65536) && (vmpg_bytes == 4096))) {
+				DML_LOG_VERBOSE("ERROR: DML::%s: Tiling size and vm page size combination not supported\n", __func__);
+				DML_ASSERT(0);
+			}
+		}
+
+		//For plane0 & 1, first calculate full_vp_width/height_l/c aligned to vmpg_width/height_l/c
+		full_vp_width_mvmpg_aligned = (unsigned int)(math_floor2((p->vp_start_x + p->full_vp_width) + *p->mvmpg_width - 1, *p->mvmpg_width) - math_floor2(p->vp_start_x, *p->mvmpg_width));
+		full_vp_height_mvmpg_aligned = (unsigned int)(math_floor2((p->vp_start_y + p->full_vp_height) + *p->mvmpg_height - 1, *p->mvmpg_height) - math_floor2(p->vp_start_y, *p->mvmpg_height));
+
+		*p->full_vp_access_width_mvmpg_aligned = p->surf_vert ? full_vp_height_mvmpg_aligned : full_vp_width_mvmpg_aligned;
+
+		//Use the equation for the exact alignment when possible. Note that the exact alignment cannot be used for horizontal access if vmpg_bytes > blk_bytes.
+		if (!p->surf_vert) { //horizontal access
+			if (p->vp_stationary == 1 && vmpg_bytes <= blk_bytes)
+				*p->meta_row_width_ub = full_vp_width_mvmpg_aligned;
+			else
+				*p->meta_row_width_ub = (unsigned int)math_ceil2((double)p->full_vp_width - 1, *p->mvmpg_width) + *p->mvmpg_width;
+			mvmpg_per_row_ub = *p->meta_row_width_ub / *p->mvmpg_width;
+		} else { //vertical access
+			if (p->vp_stationary == 1)
+				*p->meta_row_width_ub = full_vp_height_mvmpg_aligned;
+			else
+				*p->meta_row_width_ub = (unsigned int)math_ceil2((double)p->full_vp_height - 1, *p->mvmpg_height) + *p->mvmpg_height;
+			mvmpg_per_row_ub = *p->meta_row_width_ub / *p->mvmpg_height;
+		}
+
+		if (p->gpuvm_enable) {
+			meta_per_mvmpg_per_channel = (float)vmpg_bytes / (float)256 / p->num_chans;
+
+			//but using the est_blk_per_vmpg between 2 and 4, to be not as pessimestic
+			if (p->surf_vert && vmpg_bytes > blk_bytes) {
+				meta_per_mvmpg_per_channel = (float)est_blk_per_vmpg * blk_bytes / (float)256 / p->num_chans;
+			}
+
+			*p->dcc_dram_bw_nom_overhead_factor = 1 + math_max2(1.0 / 256.0, math_ceil2(meta_per_mvmpg_per_channel, p->mem_word_bytes) / (256 * meta_per_mvmpg_per_channel)); // dcc_dr_oh_nom
+		} else {
+			meta_per_mvmpg_per_channel = (float) blk_bytes / (float)256 / p->num_chans;
+
+			if (!p->surf_vert)
+				*p->dcc_dram_bw_nom_overhead_factor = 1 + 1.0 / 256.0;
+			else
+				 *p->dcc_dram_bw_nom_overhead_factor = 1 + math_max2(1.0 / 256.0, math_ceil2(meta_per_mvmpg_per_channel, p->mem_word_bytes) / (256 * meta_per_mvmpg_per_channel));
+		}
+
+		meta_per_mvmpg_per_channel_ub = (unsigned int)math_ceil2((double)meta_per_mvmpg_per_channel, p->mcache_line_size_bytes);
+
+		//but for 4KB vmpg with 64KB tile blk
+		if (p->gpuvm_enable && (blk_bytes == 65536) && (vmpg_bytes == 4096))
+			meta_per_mvmpg_per_channel_ub = 16 * meta_per_mvmpg_per_channel_ub;
+
+		// If this mcache_row_bytes for the full viewport of the surface is less than or equal to mcache_bytes,
+		// then one mcache can be used for this request stream. If not, it is useful to know the width of the viewport that can be supported in the mcache_bytes.
+		if (p->gpuvm_enable || p->surf_vert) {
+			*p->mcache_row_bytes_per_channel = mvmpg_per_row_ub * meta_per_mvmpg_per_channel_ub;
+			*p->mcache_row_bytes = *p->mcache_row_bytes_per_channel * p->num_chans;
+		} else { // horizontal and gpuvm disable
+			*p->mcache_row_bytes = *p->meta_row_width_ub * p->blk_height * p->bytes_per_pixel / 256;
+			if (p->mcache_line_size_bytes != 0)
+				*p->mcache_row_bytes_per_channel = (unsigned int)math_ceil2((double)*p->mcache_row_bytes / p->num_chans, p->mcache_line_size_bytes);
+		}
+
+		*p->dcc_dram_bw_pref_overhead_factor = 1 + math_max2(1.0 / 256.0, *p->mcache_row_bytes / p->full_swath_bytes); // dcc_dr_oh_pref
+		if (p->mcache_size_bytes != 0)
+			*p->num_mcaches = (unsigned int)math_ceil2((double)*p->mcache_row_bytes_per_channel / p->mcache_size_bytes, 1);
+
+		mvmpg_per_mcache = p->mcache_size_bytes / meta_per_mvmpg_per_channel_ub;
+		*p->mvmpg_per_mcache_lb = (unsigned int)math_floor2(mvmpg_per_mcache, 1);
+
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: gpuvm_enable = %u\n", __func__, p->gpuvm_enable);
+		DML_LOG_VERBOSE("DML::%s: vmpg_bytes = %u\n", __func__, vmpg_bytes);
+		DML_LOG_VERBOSE("DML::%s: blk_bytes = %u\n", __func__, blk_bytes);
+		DML_LOG_VERBOSE("DML::%s: meta_per_mvmpg_per_channel = %f\n", __func__, meta_per_mvmpg_per_channel);
+		DML_LOG_VERBOSE("DML::%s: mvmpg_per_row_ub = %u\n", __func__, mvmpg_per_row_ub);
+		DML_LOG_VERBOSE("DML::%s: meta_row_width_ub = %u\n", __func__, *p->meta_row_width_ub);
+		DML_LOG_VERBOSE("DML::%s: mvmpg_width = %u\n", __func__, *p->mvmpg_width);
+		DML_LOG_VERBOSE("DML::%s: mvmpg_height = %u\n", __func__, *p->mvmpg_height);
+		DML_LOG_VERBOSE("DML::%s: dcc_dram_bw_nom_overhead_factor = %f\n", __func__, *p->dcc_dram_bw_nom_overhead_factor);
+		DML_LOG_VERBOSE("DML::%s: dcc_dram_bw_pref_overhead_factor = %f\n", __func__, *p->dcc_dram_bw_pref_overhead_factor);
+#endif
+	}
+
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: mcache_row_bytes = %u\n", __func__, *p->mcache_row_bytes);
+	DML_LOG_VERBOSE("DML::%s: mcache_row_bytes_per_channel = %u\n", __func__, *p->mcache_row_bytes_per_channel);
+	DML_LOG_VERBOSE("DML::%s: num_mcaches = %u\n", __func__, *p->num_mcaches);
+#endif
+	DML_ASSERT(*p->num_mcaches > 0);
+}
+
+static void calculate_mcache_setting(
+	struct dml2_core_internal_scratch *scratch,
+	struct dml2_core_calcs_calculate_mcache_setting_params *p)
+{
+	unsigned int n;
+
+	struct dml2_core_shared_calculate_mcache_setting_locals *l = &scratch->calculate_mcache_setting_locals;
+	memset(l, 0, sizeof(struct dml2_core_shared_calculate_mcache_setting_locals));
+
+	*p->num_mcaches_l = 0;
+	*p->mcache_row_bytes_l = 0;
+	*p->mcache_row_bytes_per_channel_l = 0;
+	*p->dcc_dram_bw_nom_overhead_factor_l = 1.0;
+	*p->dcc_dram_bw_pref_overhead_factor_l = 1.0;
+
+	*p->num_mcaches_c = 0;
+	*p->mcache_row_bytes_c = 0;
+	*p->mcache_row_bytes_per_channel_c = 0;
+	*p->dcc_dram_bw_nom_overhead_factor_c = 1.0;
+	*p->dcc_dram_bw_pref_overhead_factor_c = 1.0;
+
+	*p->mall_comb_mcache_l = 0;
+	*p->mall_comb_mcache_c = 0;
+	*p->lc_comb_mcache = 0;
+
+	if (!p->dcc_enable)
+		return;
+
+	l->is_dual_plane = dml_is_420(p->source_format) || p->source_format == dml2_rgbe_alpha;
+
+	l->l_p.num_chans = p->num_chans;
+	l->l_p.mem_word_bytes = p->mem_word_bytes;
+	l->l_p.mcache_size_bytes = p->mcache_size_bytes;
+	l->l_p.mcache_line_size_bytes = p->mcache_line_size_bytes;
+	l->l_p.gpuvm_enable = p->gpuvm_enable;
+	l->l_p.gpuvm_page_size_kbytes = p->gpuvm_page_size_kbytes;
+	l->l_p.surf_vert = p->surf_vert;
+	l->l_p.vp_stationary = p->vp_stationary;
+	l->l_p.tiling_mode = p->tiling_mode;
+	l->l_p.vp_start_x = p->vp_start_x_l;
+	l->l_p.vp_start_y = p->vp_start_y_l;
+	l->l_p.full_vp_width = p->full_vp_width_l;
+	l->l_p.full_vp_height = p->full_vp_height_l;
+	l->l_p.blk_width = p->blk_width_l;
+	l->l_p.blk_height = p->blk_height_l;
+	l->l_p.vmpg_width = p->vmpg_width_l;
+	l->l_p.vmpg_height = p->vmpg_height_l;
+	l->l_p.full_swath_bytes = p->full_swath_bytes_l;
+	l->l_p.bytes_per_pixel = p->bytes_per_pixel_l;
+
+	// output
+	l->l_p.num_mcaches = p->num_mcaches_l;
+	l->l_p.mcache_row_bytes = p->mcache_row_bytes_l;
+	l->l_p.mcache_row_bytes_per_channel = p->mcache_row_bytes_per_channel_l;
+	l->l_p.dcc_dram_bw_nom_overhead_factor = p->dcc_dram_bw_nom_overhead_factor_l;
+	l->l_p.dcc_dram_bw_pref_overhead_factor = p->dcc_dram_bw_pref_overhead_factor_l;
+	l->l_p.mvmpg_width = &l->mvmpg_width_l;
+	l->l_p.mvmpg_height = &l->mvmpg_height_l;
+	l->l_p.full_vp_access_width_mvmpg_aligned = &l->full_vp_access_width_mvmpg_aligned_l;
+	l->l_p.meta_row_width_ub = &l->meta_row_width_l;
+	l->l_p.mvmpg_per_mcache_lb = &l->mvmpg_per_mcache_lb_l;
+
+	calculate_mcache_row_bytes(scratch, &l->l_p);
+	DML_ASSERT(*p->num_mcaches_l > 0);
+
+	if (l->is_dual_plane) {
+		l->c_p.num_chans = p->num_chans;
+		l->c_p.mem_word_bytes = p->mem_word_bytes;
+		l->c_p.mcache_size_bytes = p->mcache_size_bytes;
+		l->c_p.mcache_line_size_bytes = p->mcache_line_size_bytes;
+		l->c_p.gpuvm_enable = p->gpuvm_enable;
+		l->c_p.gpuvm_page_size_kbytes = p->gpuvm_page_size_kbytes;
+		l->c_p.surf_vert = p->surf_vert;
+		l->c_p.vp_stationary = p->vp_stationary;
+		l->c_p.tiling_mode = p->tiling_mode;
+		l->c_p.vp_start_x = p->vp_start_x_c;
+		l->c_p.vp_start_y = p->vp_start_y_c;
+		l->c_p.full_vp_width = p->full_vp_width_c;
+		l->c_p.full_vp_height = p->full_vp_height_c;
+		l->c_p.blk_width = p->blk_width_c;
+		l->c_p.blk_height = p->blk_height_c;
+		l->c_p.vmpg_width = p->vmpg_width_c;
+		l->c_p.vmpg_height = p->vmpg_height_c;
+		l->c_p.full_swath_bytes = p->full_swath_bytes_c;
+		l->c_p.bytes_per_pixel = p->bytes_per_pixel_c;
+
+		// output
+		l->c_p.num_mcaches = p->num_mcaches_c;
+		l->c_p.mcache_row_bytes = p->mcache_row_bytes_c;
+		l->c_p.mcache_row_bytes_per_channel = p->mcache_row_bytes_per_channel_c;
+		l->c_p.dcc_dram_bw_nom_overhead_factor = p->dcc_dram_bw_nom_overhead_factor_c;
+		l->c_p.dcc_dram_bw_pref_overhead_factor = p->dcc_dram_bw_pref_overhead_factor_c;
+		l->c_p.mvmpg_width = &l->mvmpg_width_c;
+		l->c_p.mvmpg_height = &l->mvmpg_height_c;
+		l->c_p.full_vp_access_width_mvmpg_aligned = &l->full_vp_access_width_mvmpg_aligned_c;
+		l->c_p.meta_row_width_ub = &l->meta_row_width_c;
+		l->c_p.mvmpg_per_mcache_lb = &l->mvmpg_per_mcache_lb_c;
+
+		calculate_mcache_row_bytes(scratch, &l->c_p);
+		DML_ASSERT(*p->num_mcaches_c > 0);
+	}
+
+	// Sharing for iMALL access
+	l->mcache_remainder_l = *p->mcache_row_bytes_per_channel_l % p->mcache_size_bytes;
+	l->mcache_remainder_c = *p->mcache_row_bytes_per_channel_c % p->mcache_size_bytes;
+	l->mvmpg_access_width_l = p->surf_vert ? l->mvmpg_height_l : l->mvmpg_width_l;
+	l->mvmpg_access_width_c = p->surf_vert ? l->mvmpg_height_c : l->mvmpg_width_c;
+
+	if (p->imall_enable) {
+		*p->mall_comb_mcache_l = (2 * l->mcache_remainder_l <= p->mcache_size_bytes);
+
+		if (l->is_dual_plane)
+			*p->mall_comb_mcache_c = (2 * l->mcache_remainder_c <= p->mcache_size_bytes);
+	}
+
+	if (!p->surf_vert) // horizonatal access
+		l->luma_time_factor = (double)l->mvmpg_height_c / l->mvmpg_height_l * 2;
+	else // vertical access
+		l->luma_time_factor = (double)l->mvmpg_width_c / l->mvmpg_width_l * 2;
+
+	// The algorithm starts with computing a non-integer, avg_mcache_element_size_l/c:
+	if (*p->num_mcaches_l) {
+		l->avg_mcache_element_size_l = l->meta_row_width_l / *p->num_mcaches_l;
+	}
+	if (l->is_dual_plane) {
+		l->avg_mcache_element_size_c = l->meta_row_width_c / *p->num_mcaches_c;
+
+		/* if either remainder is 0, then mcache sharing is not needed or not possible due to full utilization */
+		if (l->mcache_remainder_l && l->mcache_remainder_c) {
+			if (!p->imall_enable || (*p->mall_comb_mcache_l == *p->mall_comb_mcache_c)) {
+				l->lc_comb_last_mcache_size = (unsigned int)((l->mcache_remainder_l * (*p->mall_comb_mcache_l ? 2 : 1) * l->luma_time_factor) +
+					(l->mcache_remainder_c * (*p->mall_comb_mcache_c ? 2 : 1)));
+			}
+			*p->lc_comb_mcache = (l->lc_comb_last_mcache_size <= p->mcache_size_bytes) && (*p->mall_comb_mcache_l == *p->mall_comb_mcache_c);
+		}
+	}
+
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: imall_enable = %u\n", __func__, p->imall_enable);
+	DML_LOG_VERBOSE("DML::%s: is_dual_plane = %u\n", __func__, l->is_dual_plane);
+	DML_LOG_VERBOSE("DML::%s: surf_vert = %u\n", __func__, p->surf_vert);
+	DML_LOG_VERBOSE("DML::%s: mvmpg_width_l = %u\n", __func__, l->mvmpg_width_l);
+	DML_LOG_VERBOSE("DML::%s: mvmpg_height_l = %u\n", __func__, l->mvmpg_height_l);
+	DML_LOG_VERBOSE("DML::%s: mcache_remainder_l = %f\n", __func__, l->mcache_remainder_l);
+	DML_LOG_VERBOSE("DML::%s: num_mcaches_l = %u\n", __func__, *p->num_mcaches_l);
+	DML_LOG_VERBOSE("DML::%s: avg_mcache_element_size_l = %u\n", __func__, l->avg_mcache_element_size_l);
+	DML_LOG_VERBOSE("DML::%s: mvmpg_access_width_l = %u\n", __func__, l->mvmpg_access_width_l);
+	DML_LOG_VERBOSE("DML::%s: mall_comb_mcache_l = %u\n", __func__, *p->mall_comb_mcache_l);
+
+	if (l->is_dual_plane) {
+		DML_LOG_VERBOSE("DML::%s: mvmpg_width_c = %u\n", __func__, l->mvmpg_width_c);
+		DML_LOG_VERBOSE("DML::%s: mvmpg_height_c = %u\n", __func__, l->mvmpg_height_c);
+		DML_LOG_VERBOSE("DML::%s: mcache_remainder_c = %f\n", __func__, l->mcache_remainder_c);
+		DML_LOG_VERBOSE("DML::%s: luma_time_factor = %f\n", __func__, l->luma_time_factor);
+		DML_LOG_VERBOSE("DML::%s: num_mcaches_c = %u\n", __func__, *p->num_mcaches_c);
+		DML_LOG_VERBOSE("DML::%s: avg_mcache_element_size_c = %u\n", __func__, l->avg_mcache_element_size_c);
+		DML_LOG_VERBOSE("DML::%s: mvmpg_access_width_c = %u\n", __func__, l->mvmpg_access_width_c);
+		DML_LOG_VERBOSE("DML::%s: mall_comb_mcache_c = %u\n", __func__, *p->mall_comb_mcache_c);
+		DML_LOG_VERBOSE("DML::%s: lc_comb_last_mcache_size = %u\n", __func__, l->lc_comb_last_mcache_size);
+		DML_LOG_VERBOSE("DML::%s: lc_comb_mcache = %u\n", __func__, *p->lc_comb_mcache);
+	}
+#endif
+	// calculate split_coordinate
+	l->full_vp_access_width_l = p->surf_vert ? p->full_vp_height_l : p->full_vp_width_l;
+	l->full_vp_access_width_c = p->surf_vert ? p->full_vp_height_c : p->full_vp_width_c;
+
+	for (n = 0; n < *p->num_mcaches_l - 1; n++) {
+		p->mcache_offsets_l[n] = (unsigned int)(math_floor2((n + 1) * l->avg_mcache_element_size_l / l->mvmpg_access_width_l, 1)) * l->mvmpg_access_width_l;
+	}
+	p->mcache_offsets_l[*p->num_mcaches_l - 1] = l->full_vp_access_width_l;
+
+	if (l->is_dual_plane) {
+		for (n = 0; n < *p->num_mcaches_c - 1; n++) {
+			p->mcache_offsets_c[n] = (unsigned int)(math_floor2((n + 1) * l->avg_mcache_element_size_c / l->mvmpg_access_width_c, 1)) * l->mvmpg_access_width_c;
+		}
+		p->mcache_offsets_c[*p->num_mcaches_c - 1] = l->full_vp_access_width_c;
+	}
+#ifdef __DML_VBA_DEBUG__
+	for (n = 0; n < *p->num_mcaches_l; n++)
+		DML_LOG_VERBOSE("DML::%s: mcache_offsets_l[%u] = %u\n", __func__, n, p->mcache_offsets_l[n]);
+
+	if (l->is_dual_plane) {
+		for (n = 0; n < *p->num_mcaches_c; n++)
+			DML_LOG_VERBOSE("DML::%s: mcache_offsets_c[%u] = %u\n", __func__, n, p->mcache_offsets_c[n]);
+	}
+#endif
+
+	// Luma/Chroma combine in the last mcache
+	// In the case of Luma/Chroma combine-mCache (with lc_comb_mcache==1), all mCaches except the last segment are filled as much as possible, when stay aligned to mvmpg boundary
+	if (*p->lc_comb_mcache && l->is_dual_plane) {
+		for (n = 0; n < *p->num_mcaches_l - 1; n++)
+			p->mcache_offsets_l[n] = (n + 1) * l->mvmpg_per_mcache_lb_l * l->mvmpg_access_width_l;
+		p->mcache_offsets_l[*p->num_mcaches_l - 1] = l->full_vp_access_width_l;
+
+		for (n = 0; n < *p->num_mcaches_c - 1; n++)
+			p->mcache_offsets_c[n] = (n + 1) * l->mvmpg_per_mcache_lb_c * l->mvmpg_access_width_c;
+		p->mcache_offsets_c[*p->num_mcaches_c - 1] = l->full_vp_access_width_c;
+
+#ifdef __DML_VBA_DEBUG__
+		for (n = 0; n < *p->num_mcaches_l; n++)
+			DML_LOG_VERBOSE("DML::%s: mcache_offsets_l[%u] = %u\n", __func__, n, p->mcache_offsets_l[n]);
+
+		for (n = 0; n < *p->num_mcaches_c; n++)
+			DML_LOG_VERBOSE("DML::%s: mcache_offsets_c[%u] = %u\n", __func__, n, p->mcache_offsets_c[n]);
+#endif
+	}
+
+	*p->mcache_shift_granularity_l = l->mvmpg_access_width_l;
+	*p->mcache_shift_granularity_c = l->mvmpg_access_width_c;
+}
+
+static void calculate_mall_bw_overhead_factor(
+	double mall_prefetch_sdp_overhead_factor[], //mall_sdp_oh_nom/pref
+	double mall_prefetch_dram_overhead_factor[], //mall_dram_oh_nom/pref
+
+	// input
+	const struct dml2_display_cfg *display_cfg,
+	unsigned int num_active_planes)
+{
+	for (unsigned int k = 0; k < num_active_planes; ++k) {
+		mall_prefetch_sdp_overhead_factor[k] = 1.0;
+		mall_prefetch_dram_overhead_factor[k] = 1.0;
+
+		// SDP - on the return side
+		if (display_cfg->plane_descriptors[k].overrides.legacy_svp_config == dml2_svp_mode_override_imall) // always no data return
+			mall_prefetch_sdp_overhead_factor[k] = 1.25;
+		else if (display_cfg->plane_descriptors[k].overrides.legacy_svp_config == dml2_svp_mode_override_phantom_pipe_no_data_return)
+			mall_prefetch_sdp_overhead_factor[k] = 0.25;
+
+		// DRAM
+		if (display_cfg->plane_descriptors[k].overrides.legacy_svp_config == dml2_svp_mode_override_imall)
+			mall_prefetch_dram_overhead_factor[k] = 2.0;
+
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: k=%u, mall_prefetch_sdp_overhead_factor = %f\n", __func__, k, mall_prefetch_sdp_overhead_factor[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u, mall_prefetch_dram_overhead_factor = %f\n", __func__, k, mall_prefetch_dram_overhead_factor[k]);
+#endif
+	}
+}
+
+static double dml_get_return_bandwidth_available(
+	const struct dml2_soc_bb *soc,
+	enum dml2_core_internal_soc_state_type state_type,
+	enum dml2_core_internal_bw_type bw_type,
+	bool is_avg_bw,
+	bool is_hvm_en,
+	bool is_hvm_only,
+	double dcfclk_mhz,
+	double fclk_mhz,
+	double dram_bw_mbps)
+{
+	double return_bw_mbps = 0.;
+	double ideal_sdp_bandwidth = (double)soc->return_bus_width_bytes * dcfclk_mhz;
+	double ideal_fabric_bandwidth = fclk_mhz * (double)soc->fabric_datapath_to_dcn_data_return_bytes;
+	double ideal_dram_bandwidth = dram_bw_mbps; //dram_speed_mts * soc->clk_table.dram_config.channel_count * soc->clk_table.dram_config.channel_width_bytes;
+
+	double derate_sdp_factor;
+	double derate_fabric_factor;
+	double derate_dram_factor;
+
+	double derate_sdp_bandwidth;
+	double derate_fabric_bandwidth;
+	double derate_dram_bandwidth;
+
+	if (is_avg_bw) {
+		if (state_type == dml2_core_internal_soc_state_svp_prefetch) {
+			derate_sdp_factor = soc->qos_parameters.derate_table.dcn_mall_prefetch_average.dcfclk_derate_percent / 100.0;
+			derate_fabric_factor = soc->qos_parameters.derate_table.dcn_mall_prefetch_average.fclk_derate_percent / 100.0;
+			derate_dram_factor = soc->qos_parameters.derate_table.dcn_mall_prefetch_average.dram_derate_percent_pixel / 100.0;
+		} else { // just assume sys_active
+			derate_sdp_factor = soc->qos_parameters.derate_table.system_active_average.dcfclk_derate_percent / 100.0;
+			derate_fabric_factor = soc->qos_parameters.derate_table.system_active_average.fclk_derate_percent / 100.0;
+			derate_dram_factor = soc->qos_parameters.derate_table.system_active_average.dram_derate_percent_pixel / 100.0;
+		}
+	} else { // urgent bw
+		if (state_type == dml2_core_internal_soc_state_svp_prefetch) {
+			derate_sdp_factor = soc->qos_parameters.derate_table.dcn_mall_prefetch_urgent.dcfclk_derate_percent / 100.0;
+			derate_fabric_factor = soc->qos_parameters.derate_table.dcn_mall_prefetch_urgent.fclk_derate_percent / 100.0;
+			derate_dram_factor = soc->qos_parameters.derate_table.dcn_mall_prefetch_urgent.dram_derate_percent_pixel / 100.0;
+
+			if (is_hvm_en) {
+				if (is_hvm_only)
+					derate_dram_factor = soc->qos_parameters.derate_table.dcn_mall_prefetch_urgent.dram_derate_percent_vm / 100.0;
+				else
+					derate_dram_factor = soc->qos_parameters.derate_table.dcn_mall_prefetch_urgent.dram_derate_percent_pixel_and_vm / 100.0;
+			} else {
+				derate_dram_factor = soc->qos_parameters.derate_table.dcn_mall_prefetch_urgent.dram_derate_percent_pixel / 100.0;
+			}
+		} else { // just assume sys_active
+			derate_sdp_factor = soc->qos_parameters.derate_table.system_active_urgent.dcfclk_derate_percent / 100.0;
+			derate_fabric_factor = soc->qos_parameters.derate_table.system_active_urgent.fclk_derate_percent / 100.0;
+
+			if (is_hvm_en) {
+				if (is_hvm_only)
+					derate_dram_factor = soc->qos_parameters.derate_table.system_active_urgent.dram_derate_percent_vm / 100.0;
+				else
+					derate_dram_factor = soc->qos_parameters.derate_table.system_active_urgent.dram_derate_percent_pixel_and_vm / 100.0;
+			} else {
+				derate_dram_factor = soc->qos_parameters.derate_table.system_active_urgent.dram_derate_percent_pixel / 100.0;
+			}
+		}
+	}
+
+	derate_sdp_bandwidth = ideal_sdp_bandwidth * derate_sdp_factor;
+	derate_fabric_bandwidth = ideal_fabric_bandwidth * derate_fabric_factor;
+	derate_dram_bandwidth = ideal_dram_bandwidth * derate_dram_factor;
+
+	if (bw_type == dml2_core_internal_bw_sdp)
+		return_bw_mbps = math_min2(derate_sdp_bandwidth, derate_fabric_bandwidth);
+	else // dml2_core_internal_bw_dram
+		return_bw_mbps = derate_dram_bandwidth;
+
+	DML_LOG_VERBOSE("DML::%s: is_avg_bw = %u\n", __func__, is_avg_bw);
+	DML_LOG_VERBOSE("DML::%s: is_hvm_en = %u\n", __func__, is_hvm_en);
+	DML_LOG_VERBOSE("DML::%s: is_hvm_only = %u\n", __func__, is_hvm_only);
+	DML_LOG_VERBOSE("DML::%s: state_type = %s\n", __func__, dml2_core_internal_soc_state_type_str(state_type));
+	DML_LOG_VERBOSE("DML::%s: bw_type = %s\n", __func__, dml2_core_internal_bw_type_str(bw_type));
+	DML_LOG_VERBOSE("DML::%s: dcfclk_mhz = %f\n", __func__, dcfclk_mhz);
+	DML_LOG_VERBOSE("DML::%s: fclk_mhz = %f\n", __func__, fclk_mhz);
+	DML_LOG_VERBOSE("DML::%s: ideal_sdp_bandwidth = %f\n", __func__, ideal_sdp_bandwidth);
+	DML_LOG_VERBOSE("DML::%s: ideal_fabric_bandwidth = %f\n", __func__, ideal_fabric_bandwidth);
+	DML_LOG_VERBOSE("DML::%s: ideal_dram_bandwidth = %f\n", __func__, ideal_dram_bandwidth);
+	DML_LOG_VERBOSE("DML::%s: derate_sdp_bandwidth = %f (derate %f)\n", __func__, derate_sdp_bandwidth, derate_sdp_factor);
+	DML_LOG_VERBOSE("DML::%s: derate_fabric_bandwidth = %f (derate %f)\n", __func__, derate_fabric_bandwidth, derate_fabric_factor);
+	DML_LOG_VERBOSE("DML::%s: derate_dram_bandwidth = %f (derate %f)\n", __func__, derate_dram_bandwidth, derate_dram_factor);
+	DML_LOG_VERBOSE("DML::%s: return_bw_mbps = %f\n", __func__, return_bw_mbps);
+	return return_bw_mbps;
+}
+
+static noinline_for_stack void calculate_bandwidth_available(
+	double avg_bandwidth_available_min[dml2_core_internal_soc_state_max],
+	double avg_bandwidth_available[dml2_core_internal_soc_state_max][dml2_core_internal_bw_max],
+	double urg_bandwidth_available_min[dml2_core_internal_soc_state_max], // min between SDP and DRAM
+	double urg_bandwidth_available[dml2_core_internal_soc_state_max][dml2_core_internal_bw_max],
+	double urg_bandwidth_available_vm_only[dml2_core_internal_soc_state_max],
+	double urg_bandwidth_available_pixel_and_vm[dml2_core_internal_soc_state_max],
+
+	const struct dml2_soc_bb *soc,
+	bool HostVMEnable,
+	double dcfclk_mhz,
+	double fclk_mhz,
+	double dram_bw_mbps)
+{
+	unsigned int n, m;
+
+	DML_LOG_VERBOSE("DML::%s: dcfclk_mhz = %f\n", __func__, dcfclk_mhz);
+	DML_LOG_VERBOSE("DML::%s: fclk_mhz = %f\n", __func__, fclk_mhz);
+	DML_LOG_VERBOSE("DML::%s: dram_bw_mbps = %f\n", __func__, dram_bw_mbps);
+
+	// Calculate all the bandwidth availabe
+	for (m = 0; m < dml2_core_internal_soc_state_max; m++) {
+		for (n = 0; n < dml2_core_internal_bw_max; n++) {
+			avg_bandwidth_available[m][n] = dml_get_return_bandwidth_available(soc,
+				m, // soc_state
+				n, // bw_type
+				1, // avg_bw
+				HostVMEnable,
+				0, // hvm_only
+				dcfclk_mhz,
+				fclk_mhz,
+				dram_bw_mbps);
+
+			urg_bandwidth_available[m][n] = dml_get_return_bandwidth_available(soc, m, n, 0, HostVMEnable, 0, dcfclk_mhz, fclk_mhz, dram_bw_mbps);
+
+
+#ifdef __DML_VBA_DEBUG__
+			DML_LOG_VERBOSE("DML::%s: avg_bandwidth_available[%s][%s]=%f\n", __func__, dml2_core_internal_soc_state_type_str(m), dml2_core_internal_bw_type_str(n), avg_bandwidth_available[m][n]);
+			DML_LOG_VERBOSE("DML::%s: urg_bandwidth_available[%s][%s]=%f\n", __func__, dml2_core_internal_soc_state_type_str(m), dml2_core_internal_bw_type_str(n), urg_bandwidth_available[m][n]);
+#endif
+
+			// urg_bandwidth_available_vm_only is indexed by soc_state
+			if (n == dml2_core_internal_bw_dram) {
+				urg_bandwidth_available_vm_only[m] = dml_get_return_bandwidth_available(soc, m, n, 0, HostVMEnable, 1, dcfclk_mhz, fclk_mhz, dram_bw_mbps);
+				urg_bandwidth_available_pixel_and_vm[m] = dml_get_return_bandwidth_available(soc, m, n, 0, HostVMEnable, 0, dcfclk_mhz, fclk_mhz, dram_bw_mbps);
+			}
+		}
+
+		avg_bandwidth_available_min[m] = math_min2(avg_bandwidth_available[m][dml2_core_internal_bw_dram], avg_bandwidth_available[m][dml2_core_internal_bw_sdp]);
+		urg_bandwidth_available_min[m] = math_min2(urg_bandwidth_available[m][dml2_core_internal_bw_dram], urg_bandwidth_available[m][dml2_core_internal_bw_sdp]);
+
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: avg_bandwidth_available_min[%s]=%f\n", __func__, dml2_core_internal_soc_state_type_str(m), avg_bandwidth_available_min[m]);
+		DML_LOG_VERBOSE("DML::%s: urg_bandwidth_available_min[%s]=%f\n", __func__, dml2_core_internal_soc_state_type_str(m), urg_bandwidth_available_min[m]);
+		DML_LOG_VERBOSE("DML::%s: urg_bandwidth_available_vm_only[%s]=%f\n", __func__, dml2_core_internal_soc_state_type_str(m), urg_bandwidth_available_vm_only[n]);
+#endif
+	}
+}
+
+static void calculate_avg_bandwidth_required(
+	double avg_bandwidth_required[dml2_core_internal_soc_state_max][dml2_core_internal_bw_max],
+
+	// input
+	const struct dml2_display_cfg *display_cfg,
+	unsigned int num_active_planes,
+	double ReadBandwidthLuma[],
+	double ReadBandwidthChroma[],
+	double cursor_bw[],
+	double dcc_dram_bw_nom_overhead_factor_p0[],
+	double dcc_dram_bw_nom_overhead_factor_p1[],
+	double mall_prefetch_dram_overhead_factor[],
+	double mall_prefetch_sdp_overhead_factor[])
+{
+	unsigned int n, m, k;
+	double sdp_overhead_factor;
+	double dram_overhead_factor_p0;
+	double dram_overhead_factor_p1;
+
+	// Average BW support check
+	for (m = 0; m < dml2_core_internal_soc_state_max; m++) {
+		for (n = 0; n < dml2_core_internal_bw_max; n++) { // sdp, dram
+			avg_bandwidth_required[m][n] = 0;
+		}
+	}
+
+	// SysActive and SVP Prefetch AVG bandwidth Check
+	for (k = 0; k < num_active_planes; ++k) {
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: plane %0d\n", __func__, k);
+		DML_LOG_VERBOSE("DML::%s: ReadBandwidthLuma=%f\n", __func__, ReadBandwidthLuma[k]);
+		DML_LOG_VERBOSE("DML::%s: ReadBandwidthChroma=%f\n", __func__, ReadBandwidthChroma[k]);
+		DML_LOG_VERBOSE("DML::%s: dcc_dram_bw_nom_overhead_factor_p0=%f\n", __func__, dcc_dram_bw_nom_overhead_factor_p0[k]);
+		DML_LOG_VERBOSE("DML::%s: dcc_dram_bw_nom_overhead_factor_p1=%f\n", __func__, dcc_dram_bw_nom_overhead_factor_p1[k]);
+		DML_LOG_VERBOSE("DML::%s: mall_prefetch_dram_overhead_factor=%f\n", __func__, mall_prefetch_dram_overhead_factor[k]);
+		DML_LOG_VERBOSE("DML::%s: mall_prefetch_sdp_overhead_factor=%f\n", __func__, mall_prefetch_sdp_overhead_factor[k]);
+#endif
+
+		sdp_overhead_factor = mall_prefetch_sdp_overhead_factor[k];
+		dram_overhead_factor_p0 = dcc_dram_bw_nom_overhead_factor_p0[k] * mall_prefetch_dram_overhead_factor[k];
+		dram_overhead_factor_p1 = dcc_dram_bw_nom_overhead_factor_p1[k] * mall_prefetch_dram_overhead_factor[k];
+
+		// FIXME_DCN4, was missing cursor_bw in here, but do I actually need that and tdlut bw for average bandwidth calculation?
+		// active avg bw not include phantom, but svp_prefetch avg bw should include phantom pipes
+		if (!dml_is_phantom_pipe(&display_cfg->plane_descriptors[k])) {
+			avg_bandwidth_required[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_sdp] += sdp_overhead_factor * (ReadBandwidthLuma[k] + ReadBandwidthChroma[k]) + cursor_bw[k];
+			avg_bandwidth_required[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_dram] += dram_overhead_factor_p0 * ReadBandwidthLuma[k] + dram_overhead_factor_p1 * ReadBandwidthChroma[k] + cursor_bw[k];
+		}
+		avg_bandwidth_required[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_sdp] += sdp_overhead_factor * (ReadBandwidthLuma[k] + ReadBandwidthChroma[k]) + cursor_bw[k];
+		avg_bandwidth_required[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_dram] += dram_overhead_factor_p0 * ReadBandwidthLuma[k] + dram_overhead_factor_p1 * ReadBandwidthChroma[k] + cursor_bw[k];
+
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: avg_bandwidth_required[%s][%s]=%f\n", __func__, dml2_core_internal_soc_state_type_str(dml2_core_internal_soc_state_sys_active), dml2_core_internal_bw_type_str(dml2_core_internal_bw_sdp), avg_bandwidth_required[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_sdp]);
+		DML_LOG_VERBOSE("DML::%s: avg_bandwidth_required[%s][%s]=%f\n", __func__, dml2_core_internal_soc_state_type_str(dml2_core_internal_soc_state_sys_active), dml2_core_internal_bw_type_str(dml2_core_internal_bw_dram), avg_bandwidth_required[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_dram]);
+		DML_LOG_VERBOSE("DML::%s: avg_bandwidth_required[%s][%s]=%f\n", __func__, dml2_core_internal_soc_state_type_str(dml2_core_internal_soc_state_svp_prefetch), dml2_core_internal_bw_type_str(dml2_core_internal_bw_sdp), avg_bandwidth_required[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_sdp]);
+		DML_LOG_VERBOSE("DML::%s: avg_bandwidth_required[%s][%s]=%f\n", __func__, dml2_core_internal_soc_state_type_str(dml2_core_internal_soc_state_svp_prefetch), dml2_core_internal_bw_type_str(dml2_core_internal_bw_dram), avg_bandwidth_required[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_dram]);
+#endif
+	}
+}
+
+static void CalculateVMRowAndSwath(struct dml2_core_internal_scratch *scratch,
+	struct dml2_core_calcs_CalculateVMRowAndSwath_params *p)
+{
+	struct dml2_core_calcs_CalculateVMRowAndSwath_locals *s = &scratch->CalculateVMRowAndSwath_locals;
+
+	s->HostVMDynamicLevels = CalculateHostVMDynamicLevels(p->display_cfg->gpuvm_enable, p->display_cfg->hostvm_enable, p->HostVMMinPageSize, p->display_cfg->hostvm_max_non_cached_page_table_levels);
+
+	for (unsigned int k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+		if (p->display_cfg->gpuvm_enable == true) {
+			p->vm_group_bytes[k] = 512;
+			p->dpte_group_bytes[k] = 512;
+		} else {
+			p->vm_group_bytes[k] = 0;
+			p->dpte_group_bytes[k] = 0;
+		}
+
+		if (dml_is_420(p->myPipe[k].SourcePixelFormat) || p->myPipe[k].SourcePixelFormat == dml2_rgbe_alpha) {
+			if ((p->myPipe[k].SourcePixelFormat == dml2_420_10 || p->myPipe[k].SourcePixelFormat == dml2_420_12) && !dml_is_vertical_rotation(p->myPipe[k].RotationAngle)) {
+				s->PTEBufferSizeInRequestsForLuma[k] = (p->PTEBufferSizeInRequestsLuma + p->PTEBufferSizeInRequestsChroma) / 2;
+				s->PTEBufferSizeInRequestsForChroma[k] = s->PTEBufferSizeInRequestsForLuma[k];
+			} else {
+				s->PTEBufferSizeInRequestsForLuma[k] = p->PTEBufferSizeInRequestsLuma;
+				s->PTEBufferSizeInRequestsForChroma[k] = p->PTEBufferSizeInRequestsChroma;
+			}
+
+			scratch->calculate_vm_and_row_bytes_params.ViewportStationary = p->myPipe[k].ViewportStationary;
+			scratch->calculate_vm_and_row_bytes_params.DCCEnable = p->myPipe[k].DCCEnable;
+			scratch->calculate_vm_and_row_bytes_params.NumberOfDPPs = p->myPipe[k].DPPPerSurface;
+			scratch->calculate_vm_and_row_bytes_params.BlockHeight256Bytes = p->myPipe[k].BlockHeight256BytesC;
+			scratch->calculate_vm_and_row_bytes_params.BlockWidth256Bytes = p->myPipe[k].BlockWidth256BytesC;
+			scratch->calculate_vm_and_row_bytes_params.SourcePixelFormat = p->myPipe[k].SourcePixelFormat;
+			scratch->calculate_vm_and_row_bytes_params.SurfaceTiling = p->myPipe[k].SurfaceTiling;
+			scratch->calculate_vm_and_row_bytes_params.BytePerPixel = p->myPipe[k].BytePerPixelC;
+			scratch->calculate_vm_and_row_bytes_params.RotationAngle = p->myPipe[k].RotationAngle;
+			scratch->calculate_vm_and_row_bytes_params.SwathWidth = p->SwathWidthC[k];
+			scratch->calculate_vm_and_row_bytes_params.ViewportHeight = p->myPipe[k].ViewportHeightC;
+			scratch->calculate_vm_and_row_bytes_params.ViewportXStart = p->myPipe[k].ViewportXStartC;
+			scratch->calculate_vm_and_row_bytes_params.ViewportYStart = p->myPipe[k].ViewportYStartC;
+			scratch->calculate_vm_and_row_bytes_params.GPUVMEnable = p->display_cfg->gpuvm_enable;
+			scratch->calculate_vm_and_row_bytes_params.GPUVMMaxPageTableLevels = p->display_cfg->gpuvm_max_page_table_levels;
+			scratch->calculate_vm_and_row_bytes_params.GPUVMMinPageSizeKBytes = p->display_cfg->plane_descriptors[k].overrides.gpuvm_min_page_size_kbytes;
+			scratch->calculate_vm_and_row_bytes_params.PTEBufferSizeInRequests = s->PTEBufferSizeInRequestsForChroma[k];
+			scratch->calculate_vm_and_row_bytes_params.Pitch = p->myPipe[k].PitchC;
+			scratch->calculate_vm_and_row_bytes_params.MacroTileWidth = p->myPipe[k].BlockWidthC;
+			scratch->calculate_vm_and_row_bytes_params.MacroTileHeight = p->myPipe[k].BlockHeightC;
+			scratch->calculate_vm_and_row_bytes_params.is_phantom = dml_is_phantom_pipe(&p->display_cfg->plane_descriptors[k]);
+			scratch->calculate_vm_and_row_bytes_params.DCCMetaPitch = p->myPipe[k].DCCMetaPitchC;
+			scratch->calculate_vm_and_row_bytes_params.mrq_present = p->mrq_present;
+
+			scratch->calculate_vm_and_row_bytes_params.PixelPTEBytesPerRow = &s->PixelPTEBytesPerRowC[k];
+			scratch->calculate_vm_and_row_bytes_params.PixelPTEBytesPerRowStorage = &s->PixelPTEBytesPerRowStorageC[k];
+			scratch->calculate_vm_and_row_bytes_params.dpte_row_width_ub = &p->dpte_row_width_chroma_ub[k];
+			scratch->calculate_vm_and_row_bytes_params.dpte_row_height = &p->dpte_row_height_chroma[k];
+			scratch->calculate_vm_and_row_bytes_params.dpte_row_height_linear = &p->dpte_row_height_linear_chroma[k];
+			scratch->calculate_vm_and_row_bytes_params.PixelPTEBytesPerRow_one_row_per_frame = &s->PixelPTEBytesPerRowC_one_row_per_frame[k];
+			scratch->calculate_vm_and_row_bytes_params.dpte_row_width_ub_one_row_per_frame = &s->dpte_row_width_chroma_ub_one_row_per_frame[k];
+			scratch->calculate_vm_and_row_bytes_params.dpte_row_height_one_row_per_frame = &s->dpte_row_height_chroma_one_row_per_frame[k];
+			scratch->calculate_vm_and_row_bytes_params.vmpg_width = &p->vmpg_width_c[k];
+			scratch->calculate_vm_and_row_bytes_params.vmpg_height = &p->vmpg_height_c[k];
+			scratch->calculate_vm_and_row_bytes_params.PixelPTEReqWidth = &p->PixelPTEReqWidthC[k];
+			scratch->calculate_vm_and_row_bytes_params.PixelPTEReqHeight = &p->PixelPTEReqHeightC[k];
+			scratch->calculate_vm_and_row_bytes_params.PTERequestSize = &p->PTERequestSizeC[k];
+			scratch->calculate_vm_and_row_bytes_params.dpde0_bytes_per_frame_ub = &p->dpde0_bytes_per_frame_ub_c[k];
+
+			scratch->calculate_vm_and_row_bytes_params.meta_row_bytes = &s->meta_row_bytes_per_row_ub_c[k];
+			scratch->calculate_vm_and_row_bytes_params.MetaRequestWidth = &p->meta_req_width_chroma[k];
+			scratch->calculate_vm_and_row_bytes_params.MetaRequestHeight = &p->meta_req_height_chroma[k];
+			scratch->calculate_vm_and_row_bytes_params.meta_row_width = &p->meta_row_width_chroma[k];
+			scratch->calculate_vm_and_row_bytes_params.meta_row_height = &p->meta_row_height_chroma[k];
+			scratch->calculate_vm_and_row_bytes_params.meta_pte_bytes_per_frame_ub = &p->meta_pte_bytes_per_frame_ub_c[k];
+
+			s->vm_bytes_c = CalculateVMAndRowBytes(&scratch->calculate_vm_and_row_bytes_params);
+
+			p->PrefetchSourceLinesC[k] = CalculatePrefetchSourceLines(
+				p->myPipe[k].VRatioChroma,
+				p->myPipe[k].VTapsChroma,
+				p->myPipe[k].InterlaceEnable,
+				p->myPipe[k].ProgressiveToInterlaceUnitInOPP,
+				p->myPipe[k].SwathHeightC,
+				p->myPipe[k].RotationAngle,
+				p->myPipe[k].mirrored,
+				p->myPipe[k].ViewportStationary,
+				p->SwathWidthC[k],
+				p->myPipe[k].ViewportHeightC,
+				p->myPipe[k].ViewportXStartC,
+				p->myPipe[k].ViewportYStartC,
+
+				// Output
+				&p->VInitPreFillC[k],
+				&p->MaxNumSwathC[k]);
+		} else {
+			s->PTEBufferSizeInRequestsForLuma[k] = p->PTEBufferSizeInRequestsLuma + p->PTEBufferSizeInRequestsChroma;
+			s->PTEBufferSizeInRequestsForChroma[k] = 0;
+			s->PixelPTEBytesPerRowC[k] = 0;
+			s->PixelPTEBytesPerRowStorageC[k] = 0;
+			s->vm_bytes_c = 0;
+			p->MaxNumSwathC[k] = 0;
+			p->PrefetchSourceLinesC[k] = 0;
+			s->dpte_row_height_chroma_one_row_per_frame[k] = 0;
+			s->dpte_row_width_chroma_ub_one_row_per_frame[k] = 0;
+			s->PixelPTEBytesPerRowC_one_row_per_frame[k] = 0;
+		}
+
+		scratch->calculate_vm_and_row_bytes_params.ViewportStationary = p->myPipe[k].ViewportStationary;
+		scratch->calculate_vm_and_row_bytes_params.DCCEnable = p->myPipe[k].DCCEnable;
+		scratch->calculate_vm_and_row_bytes_params.NumberOfDPPs = p->myPipe[k].DPPPerSurface;
+		scratch->calculate_vm_and_row_bytes_params.BlockHeight256Bytes = p->myPipe[k].BlockHeight256BytesY;
+		scratch->calculate_vm_and_row_bytes_params.BlockWidth256Bytes = p->myPipe[k].BlockWidth256BytesY;
+		scratch->calculate_vm_and_row_bytes_params.SourcePixelFormat = p->myPipe[k].SourcePixelFormat;
+		scratch->calculate_vm_and_row_bytes_params.SurfaceTiling = p->myPipe[k].SurfaceTiling;
+		scratch->calculate_vm_and_row_bytes_params.BytePerPixel = p->myPipe[k].BytePerPixelY;
+		scratch->calculate_vm_and_row_bytes_params.RotationAngle = p->myPipe[k].RotationAngle;
+		scratch->calculate_vm_and_row_bytes_params.SwathWidth = p->SwathWidthY[k];
+		scratch->calculate_vm_and_row_bytes_params.ViewportHeight = p->myPipe[k].ViewportHeight;
+		scratch->calculate_vm_and_row_bytes_params.ViewportXStart = p->myPipe[k].ViewportXStart;
+		scratch->calculate_vm_and_row_bytes_params.ViewportYStart = p->myPipe[k].ViewportYStart;
+		scratch->calculate_vm_and_row_bytes_params.GPUVMEnable = p->display_cfg->gpuvm_enable;
+		scratch->calculate_vm_and_row_bytes_params.GPUVMMaxPageTableLevels = p->display_cfg->gpuvm_max_page_table_levels;
+		scratch->calculate_vm_and_row_bytes_params.GPUVMMinPageSizeKBytes = p->display_cfg->plane_descriptors[k].overrides.gpuvm_min_page_size_kbytes;
+		scratch->calculate_vm_and_row_bytes_params.PTEBufferSizeInRequests = s->PTEBufferSizeInRequestsForLuma[k];
+		scratch->calculate_vm_and_row_bytes_params.Pitch = p->myPipe[k].PitchY;
+		scratch->calculate_vm_and_row_bytes_params.MacroTileWidth = p->myPipe[k].BlockWidthY;
+		scratch->calculate_vm_and_row_bytes_params.MacroTileHeight = p->myPipe[k].BlockHeightY;
+		scratch->calculate_vm_and_row_bytes_params.is_phantom = dml_is_phantom_pipe(&p->display_cfg->plane_descriptors[k]);
+		scratch->calculate_vm_and_row_bytes_params.DCCMetaPitch = p->myPipe[k].DCCMetaPitchY;
+		scratch->calculate_vm_and_row_bytes_params.mrq_present = p->mrq_present;
+
+		scratch->calculate_vm_and_row_bytes_params.PixelPTEBytesPerRow = &s->PixelPTEBytesPerRowY[k];
+		scratch->calculate_vm_and_row_bytes_params.PixelPTEBytesPerRowStorage = &s->PixelPTEBytesPerRowStorageY[k];
+		scratch->calculate_vm_and_row_bytes_params.dpte_row_width_ub = &p->dpte_row_width_luma_ub[k];
+		scratch->calculate_vm_and_row_bytes_params.dpte_row_height = &p->dpte_row_height_luma[k];
+		scratch->calculate_vm_and_row_bytes_params.dpte_row_height_linear = &p->dpte_row_height_linear_luma[k];
+		scratch->calculate_vm_and_row_bytes_params.PixelPTEBytesPerRow_one_row_per_frame = &s->PixelPTEBytesPerRowY_one_row_per_frame[k];
+		scratch->calculate_vm_and_row_bytes_params.dpte_row_width_ub_one_row_per_frame = &s->dpte_row_width_luma_ub_one_row_per_frame[k];
+		scratch->calculate_vm_and_row_bytes_params.dpte_row_height_one_row_per_frame = &s->dpte_row_height_luma_one_row_per_frame[k];
+		scratch->calculate_vm_and_row_bytes_params.vmpg_width = &p->vmpg_width_y[k];
+		scratch->calculate_vm_and_row_bytes_params.vmpg_height = &p->vmpg_height_y[k];
+		scratch->calculate_vm_and_row_bytes_params.PixelPTEReqWidth = &p->PixelPTEReqWidthY[k];
+		scratch->calculate_vm_and_row_bytes_params.PixelPTEReqHeight = &p->PixelPTEReqHeightY[k];
+		scratch->calculate_vm_and_row_bytes_params.PTERequestSize = &p->PTERequestSizeY[k];
+		scratch->calculate_vm_and_row_bytes_params.dpde0_bytes_per_frame_ub = &p->dpde0_bytes_per_frame_ub_l[k];
+
+		scratch->calculate_vm_and_row_bytes_params.meta_row_bytes = &s->meta_row_bytes_per_row_ub_l[k];
+		scratch->calculate_vm_and_row_bytes_params.MetaRequestWidth = &p->meta_req_width_luma[k];
+		scratch->calculate_vm_and_row_bytes_params.MetaRequestHeight = &p->meta_req_height_luma[k];
+		scratch->calculate_vm_and_row_bytes_params.meta_row_width = &p->meta_row_width_luma[k];
+		scratch->calculate_vm_and_row_bytes_params.meta_row_height = &p->meta_row_height_luma[k];
+		scratch->calculate_vm_and_row_bytes_params.meta_pte_bytes_per_frame_ub = &p->meta_pte_bytes_per_frame_ub_l[k];
+
+		s->vm_bytes_l = CalculateVMAndRowBytes(&scratch->calculate_vm_and_row_bytes_params);
+
+		p->PrefetchSourceLinesY[k] = CalculatePrefetchSourceLines(
+			p->myPipe[k].VRatio,
+			p->myPipe[k].VTaps,
+			p->myPipe[k].InterlaceEnable,
+			p->myPipe[k].ProgressiveToInterlaceUnitInOPP,
+			p->myPipe[k].SwathHeightY,
+			p->myPipe[k].RotationAngle,
+			p->myPipe[k].mirrored,
+			p->myPipe[k].ViewportStationary,
+			p->SwathWidthY[k],
+			p->myPipe[k].ViewportHeight,
+			p->myPipe[k].ViewportXStart,
+			p->myPipe[k].ViewportYStart,
+
+			// Output
+			&p->VInitPreFillY[k],
+			&p->MaxNumSwathY[k]);
+
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: k=%u, vm_bytes_l = %u (before hvm level)\n", __func__, k, s->vm_bytes_l);
+		DML_LOG_VERBOSE("DML::%s: k=%u, vm_bytes_c = %u (before hvm level)\n", __func__, k, s->vm_bytes_c);
+		DML_LOG_VERBOSE("DML::%s: k=%u, meta_row_bytes_per_row_ub_l = %u\n", __func__, k, s->meta_row_bytes_per_row_ub_l[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u, meta_row_bytes_per_row_ub_c = %u\n", __func__, k, s->meta_row_bytes_per_row_ub_c[k]);
+#endif
+		p->vm_bytes[k] = (s->vm_bytes_l + s->vm_bytes_c) * (1 + 8 * s->HostVMDynamicLevels);
+		p->meta_row_bytes[k] = s->meta_row_bytes_per_row_ub_l[k] + s->meta_row_bytes_per_row_ub_c[k];
+		p->meta_row_bytes_per_row_ub_l[k] = s->meta_row_bytes_per_row_ub_l[k];
+		p->meta_row_bytes_per_row_ub_c[k] = s->meta_row_bytes_per_row_ub_c[k];
+
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: k=%u, meta_row_bytes = %u\n", __func__, k, p->meta_row_bytes[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u, vm_bytes = %u (after hvm level)\n", __func__, k, p->vm_bytes[k]);
+#endif
+		if (s->PixelPTEBytesPerRowStorageY[k] <= 64 * s->PTEBufferSizeInRequestsForLuma[k] && s->PixelPTEBytesPerRowStorageC[k] <= 64 * s->PTEBufferSizeInRequestsForChroma[k]) {
+			p->PTEBufferSizeNotExceeded[k] = true;
+		} else {
+			p->PTEBufferSizeNotExceeded[k] = false;
+		}
+
+		s->one_row_per_frame_fits_in_buffer[k] = (s->PixelPTEBytesPerRowY_one_row_per_frame[k] <= 64 * 2 * s->PTEBufferSizeInRequestsForLuma[k] &&
+												s->PixelPTEBytesPerRowC_one_row_per_frame[k] <= 64 * 2 * s->PTEBufferSizeInRequestsForChroma[k]);
+#ifdef __DML_VBA_DEBUG__
+		if (p->PTEBufferSizeNotExceeded[k] == 0 || s->one_row_per_frame_fits_in_buffer[k] == 0) {
+			DML_LOG_VERBOSE("DML::%s: k=%u, PixelPTEBytesPerRowY = %u (before hvm level)\n", __func__, k, s->PixelPTEBytesPerRowY[k]);
+			DML_LOG_VERBOSE("DML::%s: k=%u, PixelPTEBytesPerRowC = %u (before hvm level)\n", __func__, k, s->PixelPTEBytesPerRowC[k]);
+			DML_LOG_VERBOSE("DML::%s: k=%u, PixelPTEBytesPerRowStorageY = %u\n", __func__, k, s->PixelPTEBytesPerRowStorageY[k]);
+			DML_LOG_VERBOSE("DML::%s: k=%u, PixelPTEBytesPerRowStorageC = %u\n", __func__, k, s->PixelPTEBytesPerRowStorageC[k]);
+			DML_LOG_VERBOSE("DML::%s: k=%u, PTEBufferSizeInRequestsForLuma = %u\n", __func__, k, s->PTEBufferSizeInRequestsForLuma[k]);
+			DML_LOG_VERBOSE("DML::%s: k=%u, PTEBufferSizeInRequestsForChroma = %u\n", __func__, k, s->PTEBufferSizeInRequestsForChroma[k]);
+			DML_LOG_VERBOSE("DML::%s: k=%u, PTEBufferSizeNotExceeded (not one_row_per_frame) = %u\n", __func__, k, p->PTEBufferSizeNotExceeded[k]);
+
+			DML_LOG_VERBOSE("DML::%s: k=%u, HostVMDynamicLevels = %u\n", __func__, k, s->HostVMDynamicLevels);
+			DML_LOG_VERBOSE("DML::%s: k=%u, PixelPTEBytesPerRowY_one_row_per_frame = %u\n", __func__, k, s->PixelPTEBytesPerRowY_one_row_per_frame[k]);
+			DML_LOG_VERBOSE("DML::%s: k=%u, PixelPTEBytesPerRowC_one_row_per_frame = %u\n", __func__, k, s->PixelPTEBytesPerRowC_one_row_per_frame[k]);
+			DML_LOG_VERBOSE("DML::%s: k=%u, one_row_per_frame_fits_in_buffer = %u\n", __func__, k, s->one_row_per_frame_fits_in_buffer[k]);
+		}
+#endif
+	}
+
+	CalculateMALLUseForStaticScreen(
+		p->display_cfg,
+		p->NumberOfActiveSurfaces,
+		p->MALLAllocatedForDCN,
+		p->SurfaceSizeInMALL,
+		s->one_row_per_frame_fits_in_buffer,
+		// Output
+		p->is_using_mall_for_ss);
+
+	for (unsigned int k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+		if (p->display_cfg->gpuvm_enable) {
+			if (p->display_cfg->plane_descriptors[k].overrides.hw.force_pte_buffer_mode.enable == 1) {
+				p->PTE_BUFFER_MODE[k] = p->display_cfg->plane_descriptors[k].overrides.hw.force_pte_buffer_mode.value;
+			}
+			p->PTE_BUFFER_MODE[k] = p->myPipe[k].FORCE_ONE_ROW_FOR_FRAME || p->is_using_mall_for_ss[k] || (p->display_cfg->plane_descriptors[k].overrides.legacy_svp_config == dml2_svp_mode_override_main_pipe) ||
+				dml_is_phantom_pipe(&p->display_cfg->plane_descriptors[k]) || (p->display_cfg->plane_descriptors[k].overrides.gpuvm_min_page_size_kbytes > 64);
+			p->BIGK_FRAGMENT_SIZE[k] = (unsigned int)(math_log((float)p->display_cfg->plane_descriptors[k].overrides.gpuvm_min_page_size_kbytes * 1024, 2) - 12);
+		} else {
+			p->PTE_BUFFER_MODE[k] = 0;
+			p->BIGK_FRAGMENT_SIZE[k] = 0;
+		}
+	}
+
+	for (unsigned int k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+		p->DCCMetaBufferSizeNotExceeded[k] = true;
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: k=%u, SurfaceSizeInMALL = %u\n", __func__, k, p->SurfaceSizeInMALL[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u, is_using_mall_for_ss = %u\n", __func__, k, p->is_using_mall_for_ss[k]);
+#endif
+		p->use_one_row_for_frame[k] = p->myPipe[k].FORCE_ONE_ROW_FOR_FRAME || p->is_using_mall_for_ss[k] || (p->display_cfg->plane_descriptors[k].overrides.legacy_svp_config == dml2_svp_mode_override_main_pipe) ||
+			(dml_is_phantom_pipe(&p->display_cfg->plane_descriptors[k])) || (p->display_cfg->plane_descriptors[k].overrides.gpuvm_min_page_size_kbytes > 64 && dml_is_vertical_rotation(p->myPipe[k].RotationAngle));
+
+		p->use_one_row_for_frame_flip[k] = p->use_one_row_for_frame[k] && !(p->display_cfg->plane_descriptors[k].overrides.uclk_pstate_change_strategy == dml2_uclk_pstate_change_strategy_force_mall_full_frame);
+
+		if (p->use_one_row_for_frame[k]) {
+			p->dpte_row_height_luma[k] = s->dpte_row_height_luma_one_row_per_frame[k];
+			p->dpte_row_width_luma_ub[k] = s->dpte_row_width_luma_ub_one_row_per_frame[k];
+			s->PixelPTEBytesPerRowY[k] = s->PixelPTEBytesPerRowY_one_row_per_frame[k];
+			p->dpte_row_height_chroma[k] = s->dpte_row_height_chroma_one_row_per_frame[k];
+			p->dpte_row_width_chroma_ub[k] = s->dpte_row_width_chroma_ub_one_row_per_frame[k];
+			s->PixelPTEBytesPerRowC[k] = s->PixelPTEBytesPerRowC_one_row_per_frame[k];
+			p->PTEBufferSizeNotExceeded[k] = s->one_row_per_frame_fits_in_buffer[k];
+		}
+
+		if (p->meta_row_bytes[k] <= p->DCCMetaBufferSizeBytes) {
+			p->DCCMetaBufferSizeNotExceeded[k] = true;
+		} else {
+			p->DCCMetaBufferSizeNotExceeded[k] = false;
+
+#ifdef __DML_VBA_DEBUG__
+			DML_LOG_VERBOSE("DML::%s: k=%d, meta_row_bytes = %d\n",  __func__, k, p->meta_row_bytes[k]);
+			DML_LOG_VERBOSE("DML::%s: k=%d, DCCMetaBufferSizeBytes = %d\n",  __func__, k, p->DCCMetaBufferSizeBytes);
+			DML_LOG_VERBOSE("DML::%s: k=%d, DCCMetaBufferSizeNotExceeded = %d\n",  __func__, k, p->DCCMetaBufferSizeNotExceeded[k]);
+#endif
+		}
+
+		s->PixelPTEBytesPerRowY[k] = s->PixelPTEBytesPerRowY[k] * (1 + 8 * s->HostVMDynamicLevels);
+		s->PixelPTEBytesPerRowC[k] = s->PixelPTEBytesPerRowC[k] * (1 + 8 * s->HostVMDynamicLevels);
+		p->PixelPTEBytesPerRow[k] = s->PixelPTEBytesPerRowY[k] + s->PixelPTEBytesPerRowC[k];
+		p->dpte_row_bytes_per_row_l[k] = s->PixelPTEBytesPerRowY[k];
+		p->dpte_row_bytes_per_row_c[k] = s->PixelPTEBytesPerRowC[k];
+
+		// if one row of dPTEs is meant to span the entire frame, then for these calculations, we will pretend like that one big row is fetched in two halfs
+		if (p->use_one_row_for_frame[k])
+			p->PixelPTEBytesPerRow[k] = p->PixelPTEBytesPerRow[k] / 2;
+
+		CalculateRowBandwidth(
+			p->display_cfg->gpuvm_enable,
+			p->use_one_row_for_frame[k],
+			p->myPipe[k].SourcePixelFormat,
+			p->myPipe[k].VRatio,
+			p->myPipe[k].VRatioChroma,
+			p->myPipe[k].DCCEnable,
+			p->myPipe[k].HTotal / p->myPipe[k].PixelClock,
+			s->PixelPTEBytesPerRowY[k],
+			s->PixelPTEBytesPerRowC[k],
+			p->dpte_row_height_luma[k],
+			p->dpte_row_height_chroma[k],
+
+			p->mrq_present,
+			p->meta_row_bytes_per_row_ub_l[k],
+			p->meta_row_bytes_per_row_ub_c[k],
+			p->meta_row_height_luma[k],
+			p->meta_row_height_chroma[k],
+
+			// Output
+			&p->dpte_row_bw[k],
+			&p->meta_row_bw[k]);
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: k=%u, use_one_row_for_frame = %u\n", __func__, k, p->use_one_row_for_frame[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u, use_one_row_for_frame_flip = %u\n", __func__, k, p->use_one_row_for_frame_flip[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u, UseMALLForPStateChange = %u\n", __func__, k, p->display_cfg->plane_descriptors[k].overrides.legacy_svp_config);
+		DML_LOG_VERBOSE("DML::%s: k=%u, dpte_row_height_luma = %u\n", __func__, k, p->dpte_row_height_luma[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u, dpte_row_width_luma_ub = %u\n", __func__, k, p->dpte_row_width_luma_ub[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u, PixelPTEBytesPerRowY = %u (after hvm level)\n", __func__, k, s->PixelPTEBytesPerRowY[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u, dpte_row_height_chroma = %u\n", __func__, k, p->dpte_row_height_chroma[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u, dpte_row_width_chroma_ub = %u\n", __func__, k, p->dpte_row_width_chroma_ub[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u, PixelPTEBytesPerRowC = %u (after hvm level)\n", __func__, k, s->PixelPTEBytesPerRowC[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u, PixelPTEBytesPerRow = %u\n", __func__, k, p->PixelPTEBytesPerRow[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u, PTEBufferSizeNotExceeded = %u\n", __func__, k, p->PTEBufferSizeNotExceeded[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u, gpuvm_enable = %u\n", __func__, k, p->display_cfg->gpuvm_enable);
+		DML_LOG_VERBOSE("DML::%s: k=%u, PTE_BUFFER_MODE = %u\n", __func__, k, p->PTE_BUFFER_MODE[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u, BIGK_FRAGMENT_SIZE = %u\n", __func__, k, p->BIGK_FRAGMENT_SIZE[k]);
+#endif
+	}
+}
+
+static double CalculateUrgentLatency(
+	double UrgentLatencyPixelDataOnly,
+	double UrgentLatencyPixelMixedWithVMData,
+	double UrgentLatencyVMDataOnly,
+	bool DoUrgentLatencyAdjustment,
+	double UrgentLatencyAdjustmentFabricClockComponent,
+	double UrgentLatencyAdjustmentFabricClockReference,
+	double FabricClock,
+	double uclk_freq_mhz,
+	enum dml2_qos_param_type qos_type,
+	unsigned int urgent_ramp_uclk_cycles,
+	unsigned int df_qos_response_time_fclk_cycles,
+	unsigned int max_round_trip_to_furthest_cs_fclk_cycles,
+	unsigned int mall_overhead_fclk_cycles,
+	double umc_urgent_ramp_latency_margin,
+	double fabric_max_transport_latency_margin)
+{
+	double urgent_latency = 0;
+	if (qos_type == dml2_qos_param_type_dcn4x) {
+		urgent_latency = (df_qos_response_time_fclk_cycles + mall_overhead_fclk_cycles) / FabricClock
+			+ max_round_trip_to_furthest_cs_fclk_cycles / FabricClock * (1 + fabric_max_transport_latency_margin / 100.0)
+			+ urgent_ramp_uclk_cycles / uclk_freq_mhz * (1 + umc_urgent_ramp_latency_margin / 100.0);
+	} else {
+		urgent_latency = math_max3(UrgentLatencyPixelDataOnly, UrgentLatencyPixelMixedWithVMData, UrgentLatencyVMDataOnly);
+		if (DoUrgentLatencyAdjustment == true) {
+			urgent_latency = urgent_latency + UrgentLatencyAdjustmentFabricClockComponent * (UrgentLatencyAdjustmentFabricClockReference / FabricClock - 1);
+		}
+	}
+#ifdef __DML_VBA_DEBUG__
+	if (qos_type == dml2_qos_param_type_dcn4x) {
+		DML_LOG_VERBOSE("DML::%s: qos_type = %d\n", __func__, qos_type);
+		DML_LOG_VERBOSE("DML::%s: urgent_ramp_uclk_cycles = %d\n", __func__, urgent_ramp_uclk_cycles);
+		DML_LOG_VERBOSE("DML::%s: uclk_freq_mhz = %f\n", __func__, uclk_freq_mhz);
+		DML_LOG_VERBOSE("DML::%s: umc_urgent_ramp_latency_margin = %f\n", __func__, umc_urgent_ramp_latency_margin);
+	} else {
+		DML_LOG_VERBOSE("DML::%s: UrgentLatencyPixelDataOnly = %f\n", __func__, UrgentLatencyPixelDataOnly);
+		DML_LOG_VERBOSE("DML::%s: UrgentLatencyPixelMixedWithVMData = %f\n", __func__, UrgentLatencyPixelMixedWithVMData);
+		DML_LOG_VERBOSE("DML::%s: UrgentLatencyVMDataOnly = %f\n", __func__, UrgentLatencyVMDataOnly);
+		DML_LOG_VERBOSE("DML::%s: UrgentLatencyAdjustmentFabricClockComponent = %f\n", __func__, UrgentLatencyAdjustmentFabricClockComponent);
+		DML_LOG_VERBOSE("DML::%s: UrgentLatencyAdjustmentFabricClockReference = %f\n", __func__, UrgentLatencyAdjustmentFabricClockReference);
+	}
+	DML_LOG_VERBOSE("DML::%s: FabricClock = %f\n", __func__, FabricClock);
+	DML_LOG_VERBOSE("DML::%s: UrgentLatency = %f\n", __func__, urgent_latency);
+#endif
+	return urgent_latency;
+}
+
+static double CalculateTripToMemory(
+	double UrgLatency,
+	double FabricClock,
+	double uclk_freq_mhz,
+	enum dml2_qos_param_type qos_type,
+	unsigned int trip_to_memory_uclk_cycles,
+	unsigned int max_round_trip_to_furthest_cs_fclk_cycles,
+	unsigned int mall_overhead_fclk_cycles,
+	double umc_max_latency_margin,
+	double fabric_max_transport_latency_margin)
+{
+	double trip_to_memory_us;
+	if (qos_type == dml2_qos_param_type_dcn4x) {
+		trip_to_memory_us = mall_overhead_fclk_cycles / FabricClock
+			+ max_round_trip_to_furthest_cs_fclk_cycles / FabricClock * (1.0 + fabric_max_transport_latency_margin / 100.0)
+			+ trip_to_memory_uclk_cycles / uclk_freq_mhz * (1.0 + umc_max_latency_margin / 100.0);
+	} else {
+		trip_to_memory_us = UrgLatency;
+	}
+
+#ifdef __DML_VBA_DEBUG__
+	if (qos_type == dml2_qos_param_type_dcn4x) {
+		DML_LOG_VERBOSE("DML::%s: qos_type = %d\n", __func__, qos_type);
+		DML_LOG_VERBOSE("DML::%s: max_round_trip_to_furthest_cs_fclk_cycles = %d\n", __func__, max_round_trip_to_furthest_cs_fclk_cycles);
+		DML_LOG_VERBOSE("DML::%s: mall_overhead_fclk_cycles = %d\n", __func__, mall_overhead_fclk_cycles);
+		DML_LOG_VERBOSE("DML::%s: trip_to_memory_uclk_cycles = %d\n", __func__, trip_to_memory_uclk_cycles);
+		DML_LOG_VERBOSE("DML::%s: uclk_freq_mhz = %f\n", __func__, uclk_freq_mhz);
+		DML_LOG_VERBOSE("DML::%s: FabricClock = %f\n", __func__, FabricClock);
+		DML_LOG_VERBOSE("DML::%s: fabric_max_transport_latency_margin = %f\n", __func__, fabric_max_transport_latency_margin);
+		DML_LOG_VERBOSE("DML::%s: umc_max_latency_margin = %f\n", __func__, umc_max_latency_margin);
+	} else {
+		DML_LOG_VERBOSE("DML::%s: UrgLatency = %f\n", __func__, UrgLatency);
+	}
+	DML_LOG_VERBOSE("DML::%s: trip_to_memory_us = %f\n", __func__, trip_to_memory_us);
+#endif
+
+
+	return trip_to_memory_us;
+}
+
+static double CalculateMetaTripToMemory(
+	double UrgLatency,
+	double FabricClock,
+	double uclk_freq_mhz,
+	enum dml2_qos_param_type qos_type,
+	unsigned int meta_trip_to_memory_uclk_cycles,
+	unsigned int meta_trip_to_memory_fclk_cycles,
+	double umc_max_latency_margin,
+	double fabric_max_transport_latency_margin)
+{
+	double meta_trip_to_memory_us;
+	if (qos_type == dml2_qos_param_type_dcn4x) {
+		meta_trip_to_memory_us = meta_trip_to_memory_fclk_cycles / FabricClock * (1.0 + fabric_max_transport_latency_margin / 100.0)
+			+ meta_trip_to_memory_uclk_cycles / uclk_freq_mhz * (1.0 + umc_max_latency_margin / 100.0);
+	} else {
+		meta_trip_to_memory_us = UrgLatency;
+	}
+
+#ifdef __DML_VBA_DEBUG__
+	if (qos_type == dml2_qos_param_type_dcn4x) {
+		DML_LOG_VERBOSE("DML::%s: qos_type = %d\n", __func__, qos_type);
+		DML_LOG_VERBOSE("DML::%s: meta_trip_to_memory_fclk_cycles = %d\n", __func__, meta_trip_to_memory_fclk_cycles);
+		DML_LOG_VERBOSE("DML::%s: meta_trip_to_memory_uclk_cycles = %d\n", __func__, meta_trip_to_memory_uclk_cycles);
+		DML_LOG_VERBOSE("DML::%s: uclk_freq_mhz = %f\n", __func__, uclk_freq_mhz);
+	} else {
+		DML_LOG_VERBOSE("DML::%s: UrgLatency = %f\n", __func__, UrgLatency);
+	}
+	DML_LOG_VERBOSE("DML::%s: meta_trip_to_memory_us = %f\n", __func__, meta_trip_to_memory_us);
+#endif
+
+
+	return meta_trip_to_memory_us;
+}
+
+static void calculate_cursor_req_attributes(
+	unsigned int cursor_width,
+	unsigned int cursor_bpp,
+
+	// output
+	unsigned int *cursor_lines_per_chunk,
+	unsigned int *cursor_bytes_per_line,
+	unsigned int *cursor_bytes_per_chunk,
+	unsigned int *cursor_bytes)
+{
+	unsigned int cursor_bytes_per_req = 0;
+	unsigned int cursor_width_bytes = 0;
+	unsigned int cursor_height = 0;
+
+	//SW determines the cursor pitch to support the maximum cursor_width that will be used but the following restrictions apply.
+	//- For 2bpp, cursor_pitch = 256 pixels due to min cursor request size of 64B
+	//- For 32 or 64 bpp, cursor_pitch = 64, 128 or 256 pixels depending on the cursor width
+
+	//The cursor requestor uses a cursor request size of 64B, 128B, or 256B depending on the cursor_width and cursor_bpp as follows.
+
+	cursor_width_bytes = (unsigned int)math_ceil2((double)cursor_width * cursor_bpp / 8, 1);
+	if (cursor_width_bytes <= 64)
+		cursor_bytes_per_req = 64;
+	else if (cursor_width_bytes <= 128)
+		cursor_bytes_per_req = 128;
+	else
+		cursor_bytes_per_req = 256;
+
+	//If cursor_width_bytes is greater than 256B, then multiple 256B requests are issued to fetch the entire cursor line.
+	*cursor_bytes_per_line = (unsigned int)math_ceil2((double)cursor_width_bytes, cursor_bytes_per_req);
+
+	//Nominally, the cursor chunk is 1KB or 2KB but it is restricted to a power of 2 number of lines with a maximum of 16 lines.
+	if (cursor_bpp == 2) {
+		*cursor_lines_per_chunk = 16;
+	} else if (cursor_bpp == 32) {
+		if (cursor_width <= 32)
+			*cursor_lines_per_chunk = 16;
+		else if (cursor_width <= 64)
+			*cursor_lines_per_chunk = 8;
+		else if (cursor_width <= 128)
+			*cursor_lines_per_chunk = 4;
+		else
+			*cursor_lines_per_chunk = 2;
+	} else if (cursor_bpp == 64) {
+		if (cursor_width <= 16)
+			*cursor_lines_per_chunk = 16;
+		else if (cursor_width <= 32)
+			*cursor_lines_per_chunk = 8;
+		else if (cursor_width <= 64)
+			*cursor_lines_per_chunk = 4;
+		else if (cursor_width <= 128)
+			*cursor_lines_per_chunk = 2;
+		else
+			*cursor_lines_per_chunk = 1;
+	} else {
+		if (cursor_width > 0) {
+			DML_LOG_VERBOSE("DML::%s: Invalid cursor_bpp = %d\n", __func__, cursor_bpp);
+			DML_ASSERT(0);
+		}
+	}
+
+	*cursor_bytes_per_chunk = *cursor_bytes_per_line * *cursor_lines_per_chunk;
+
+	// For the cursor implementation, all requested data is stored in the return buffer. Given this fact, the cursor_bytes can be directly compared with the CursorBufferSize.
+	// Only cursor_width is provided for worst case sizing so assume that the cursor is square
+	cursor_height = cursor_width;
+	*cursor_bytes = *cursor_bytes_per_line * cursor_height;
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: cursor_bpp = %d\n", __func__, cursor_bpp);
+	DML_LOG_VERBOSE("DML::%s: cursor_width = %d\n", __func__, cursor_width);
+	DML_LOG_VERBOSE("DML::%s: cursor_width_bytes = %d\n", __func__, cursor_width_bytes);
+	DML_LOG_VERBOSE("DML::%s: cursor_bytes_per_req = %d\n", __func__, cursor_bytes_per_req);
+	DML_LOG_VERBOSE("DML::%s: cursor_lines_per_chunk = %d\n", __func__, *cursor_lines_per_chunk);
+	DML_LOG_VERBOSE("DML::%s: cursor_bytes_per_line = %d\n", __func__, *cursor_bytes_per_line);
+	DML_LOG_VERBOSE("DML::%s: cursor_bytes_per_chunk = %d\n", __func__, *cursor_bytes_per_chunk);
+	DML_LOG_VERBOSE("DML::%s: cursor_bytes = %d\n", __func__, *cursor_bytes);
+	DML_LOG_VERBOSE("DML::%s: cursor_pitch = %d\n", __func__, cursor_bpp == 2 ? 256 : (unsigned int)1 << (unsigned int)math_ceil2(math_log((float)cursor_width, 2), 1));
+#endif
+}
+
+static void calculate_cursor_urgent_burst_factor(
+	unsigned int CursorBufferSize,
+	unsigned int CursorWidth,
+	unsigned int cursor_bytes_per_chunk,
+	unsigned int cursor_lines_per_chunk,
+	double LineTime,
+	double UrgentLatency,
+
+	double *UrgentBurstFactorCursor,
+	bool *NotEnoughUrgentLatencyHiding)
+{
+	unsigned int LinesInCursorBuffer = 0;
+	double CursorBufferSizeInTime = 0;
+
+	if (CursorWidth > 0) {
+		LinesInCursorBuffer = (unsigned int)math_floor2(CursorBufferSize * 1024.0 / (double)cursor_bytes_per_chunk, 1) * cursor_lines_per_chunk;
+
+		CursorBufferSizeInTime = LinesInCursorBuffer * LineTime;
+		if (CursorBufferSizeInTime - UrgentLatency <= 0) {
+			*NotEnoughUrgentLatencyHiding = 1;
+			*UrgentBurstFactorCursor = 1;
+		} else {
+			*NotEnoughUrgentLatencyHiding = 0;
+			*UrgentBurstFactorCursor = CursorBufferSizeInTime / (CursorBufferSizeInTime - UrgentLatency);
+		}
+
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: LinesInCursorBuffer = %u\n", __func__, LinesInCursorBuffer);
+		DML_LOG_VERBOSE("DML::%s: CursorBufferSizeInTime = %f\n", __func__, CursorBufferSizeInTime);
+		DML_LOG_VERBOSE("DML::%s: CursorBufferSize = %u (kbytes)\n", __func__, CursorBufferSize);
+		DML_LOG_VERBOSE("DML::%s: cursor_bytes_per_chunk = %u\n", __func__, cursor_bytes_per_chunk);
+		DML_LOG_VERBOSE("DML::%s: cursor_lines_per_chunk = %u\n", __func__, cursor_lines_per_chunk);
+		DML_LOG_VERBOSE("DML::%s: UrgentBurstFactorCursor = %f\n", __func__, *UrgentBurstFactorCursor);
+		DML_LOG_VERBOSE("DML::%s: NotEnoughUrgentLatencyHiding = %d\n", __func__, *NotEnoughUrgentLatencyHiding);
+#endif
+
+	}
+}
+
+static void CalculateUrgentBurstFactor(
+	const struct dml2_plane_parameters *plane_cfg,
+	unsigned int swath_width_luma_ub,
+	unsigned int swath_width_chroma_ub,
+	unsigned int SwathHeightY,
+	unsigned int SwathHeightC,
+	double LineTime,
+	double UrgentLatency,
+	double VRatio,
+	double VRatioC,
+	double BytePerPixelInDETY,
+	double BytePerPixelInDETC,
+	unsigned int DETBufferSizeY,
+	unsigned int DETBufferSizeC,
+	// Output
+	double *UrgentBurstFactorLuma,
+	double *UrgentBurstFactorChroma,
+	bool *NotEnoughUrgentLatencyHiding)
+{
+	double LinesInDETLuma;
+	double LinesInDETChroma;
+	double DETBufferSizeInTimeLuma;
+	double DETBufferSizeInTimeChroma;
+
+	*NotEnoughUrgentLatencyHiding = 0;
+	*UrgentBurstFactorLuma = 0;
+	*UrgentBurstFactorChroma = 0;
+
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: VRatio = %f\n", __func__, VRatio);
+	DML_LOG_VERBOSE("DML::%s: VRatioC = %f\n", __func__, VRatioC);
+	DML_LOG_VERBOSE("DML::%s: DETBufferSizeY = %d\n", __func__, DETBufferSizeY);
+	DML_LOG_VERBOSE("DML::%s: DETBufferSizeC = %d\n", __func__, DETBufferSizeC);
+	DML_LOG_VERBOSE("DML::%s: BytePerPixelInDETY = %f\n", __func__, BytePerPixelInDETY);
+	DML_LOG_VERBOSE("DML::%s: swath_width_luma_ub = %d\n", __func__, swath_width_luma_ub);
+	DML_LOG_VERBOSE("DML::%s: LineTime = %f\n", __func__, LineTime);
+#endif
+	DML_ASSERT(VRatio > 0);
+
+	LinesInDETLuma = (dml_is_phantom_pipe(plane_cfg) ? 1024 * 1024 : DETBufferSizeY) / BytePerPixelInDETY / swath_width_luma_ub;
+
+	DETBufferSizeInTimeLuma = math_floor2(LinesInDETLuma, SwathHeightY) * LineTime / VRatio;
+	if (DETBufferSizeInTimeLuma - UrgentLatency <= 0) {
+		*NotEnoughUrgentLatencyHiding = 1;
+		*UrgentBurstFactorLuma = 1;
+	} else {
+		*UrgentBurstFactorLuma = DETBufferSizeInTimeLuma / (DETBufferSizeInTimeLuma - UrgentLatency);
+	}
+
+	if (BytePerPixelInDETC > 0) {
+		LinesInDETChroma = (dml_is_phantom_pipe(plane_cfg) ? 1024 * 1024 : DETBufferSizeC) / BytePerPixelInDETC / swath_width_chroma_ub;
+
+		DETBufferSizeInTimeChroma = math_floor2(LinesInDETChroma, SwathHeightC) * LineTime / VRatioC;
+		if (DETBufferSizeInTimeChroma - UrgentLatency <= 0) {
+			*NotEnoughUrgentLatencyHiding = 1;
+			*UrgentBurstFactorChroma = 1;
+		} else {
+			*UrgentBurstFactorChroma = DETBufferSizeInTimeChroma / (DETBufferSizeInTimeChroma - UrgentLatency);
+		}
+	}
+
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: LinesInDETLuma = %f\n", __func__, LinesInDETLuma);
+	DML_LOG_VERBOSE("DML::%s: UrgentLatency = %f\n", __func__, UrgentLatency);
+	DML_LOG_VERBOSE("DML::%s: DETBufferSizeInTimeLuma = %f\n", __func__, DETBufferSizeInTimeLuma);
+	DML_LOG_VERBOSE("DML::%s: UrgentBurstFactorLuma = %f\n", __func__, *UrgentBurstFactorLuma);
+	DML_LOG_VERBOSE("DML::%s: UrgentBurstFactorChroma = %f\n", __func__, *UrgentBurstFactorChroma);
+	DML_LOG_VERBOSE("DML::%s: NotEnoughUrgentLatencyHiding = %d\n", __func__, *NotEnoughUrgentLatencyHiding);
+#endif
+}
+
+static void CalculateDCFCLKDeepSleepTdlut(
+	const struct dml2_display_cfg *display_cfg,
+	unsigned int NumberOfActiveSurfaces,
+	unsigned int BytePerPixelY[],
+	unsigned int BytePerPixelC[],
+	unsigned int SwathWidthY[],
+	unsigned int SwathWidthC[],
+	unsigned int DPPPerSurface[],
+	double PSCL_THROUGHPUT[],
+	double PSCL_THROUGHPUT_CHROMA[],
+	double Dppclk[],
+	double ReadBandwidthLuma[],
+	double ReadBandwidthChroma[],
+	unsigned int ReturnBusWidth,
+
+	double dispclk,
+	unsigned int tdlut_bytes_to_deliver[],
+	double prefetch_swath_time_us[],
+
+	// Output
+	double *DCFClkDeepSleep)
+{
+	double DisplayPipeLineDeliveryTimeLuma;
+	double DisplayPipeLineDeliveryTimeChroma;
+	double DCFClkDeepSleepPerSurface[DML2_MAX_PLANES];
+	double ReadBandwidth = 0.0;
+
+	for (unsigned int k = 0; k < NumberOfActiveSurfaces; ++k) {
+		double pixel_rate_mhz = ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000);
+
+		if (display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio <= 1) {
+			DisplayPipeLineDeliveryTimeLuma = SwathWidthY[k] * DPPPerSurface[k] / display_cfg->plane_descriptors[k].composition.scaler_info.plane0.h_ratio / pixel_rate_mhz;
+		} else {
+			DisplayPipeLineDeliveryTimeLuma = SwathWidthY[k] / PSCL_THROUGHPUT[k] / Dppclk[k];
+		}
+		if (BytePerPixelC[k] == 0) {
+			DisplayPipeLineDeliveryTimeChroma = 0;
+		} else {
+			if (display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_ratio <= 1) {
+				DisplayPipeLineDeliveryTimeChroma = SwathWidthC[k] * DPPPerSurface[k] / display_cfg->plane_descriptors[k].composition.scaler_info.plane1.h_ratio / pixel_rate_mhz;
+			} else {
+				DisplayPipeLineDeliveryTimeChroma = SwathWidthC[k] / PSCL_THROUGHPUT_CHROMA[k] / Dppclk[k];
+			}
+		}
+
+		if (BytePerPixelC[k] > 0) {
+			DCFClkDeepSleepPerSurface[k] = math_max2(__DML2_CALCS_DCFCLK_FACTOR__ * SwathWidthY[k] * BytePerPixelY[k] / 32.0 / DisplayPipeLineDeliveryTimeLuma,
+				__DML2_CALCS_DCFCLK_FACTOR__ * SwathWidthC[k] * BytePerPixelC[k] / 32.0 / DisplayPipeLineDeliveryTimeChroma);
+		} else {
+			DCFClkDeepSleepPerSurface[k] = __DML2_CALCS_DCFCLK_FACTOR__ * SwathWidthY[k] * BytePerPixelY[k] / 64.0 / DisplayPipeLineDeliveryTimeLuma;
+		}
+		DCFClkDeepSleepPerSurface[k] = math_max2(DCFClkDeepSleepPerSurface[k], pixel_rate_mhz / 16);
+
+		// adjust for 3dlut delivery time
+		if (display_cfg->plane_descriptors[k].tdlut.setup_for_tdlut && tdlut_bytes_to_deliver[k] > 0) {
+			double tdlut_required_deepsleep_dcfclk = (double) tdlut_bytes_to_deliver[k] / 64.0 / prefetch_swath_time_us[k];
+
+			DML_LOG_VERBOSE("DML::%s: k=%d, DCFClkDeepSleepPerSurface = %f\n", __func__, k, DCFClkDeepSleepPerSurface[k]);
+			DML_LOG_VERBOSE("DML::%s: k=%d, tdlut_bytes_to_deliver = %d\n", __func__, k, tdlut_bytes_to_deliver[k]);
+			DML_LOG_VERBOSE("DML::%s: k=%d, prefetch_swath_time_us = %f\n", __func__, k, prefetch_swath_time_us[k]);
+			DML_LOG_VERBOSE("DML::%s: k=%d, tdlut_required_deepsleep_dcfclk = %f\n", __func__, k, tdlut_required_deepsleep_dcfclk);
+
+			// increase the deepsleep dcfclk to match the original dispclk throughput rate
+			if (tdlut_required_deepsleep_dcfclk > DCFClkDeepSleepPerSurface[k]) {
+				DCFClkDeepSleepPerSurface[k] = math_max2(DCFClkDeepSleepPerSurface[k], tdlut_required_deepsleep_dcfclk);
+				DCFClkDeepSleepPerSurface[k] = math_max2(DCFClkDeepSleepPerSurface[k], dispclk / 4.0);
+			}
+		}
+
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: k=%u, PixelClock = %f\n", __func__, k, pixel_rate_mhz);
+		DML_LOG_VERBOSE("DML::%s: k=%u, DCFClkDeepSleepPerSurface = %f\n", __func__, k, DCFClkDeepSleepPerSurface[k]);
+#endif
+	}
+
+	for (unsigned int k = 0; k < NumberOfActiveSurfaces; ++k) {
+		ReadBandwidth = ReadBandwidth + ReadBandwidthLuma[k] + ReadBandwidthChroma[k];
+	}
+
+	*DCFClkDeepSleep = math_max2(8.0, __DML2_CALCS_DCFCLK_FACTOR__ * ReadBandwidth / (double)ReturnBusWidth);
+
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: __DML2_CALCS_DCFCLK_FACTOR__ = %f\n", __func__, __DML2_CALCS_DCFCLK_FACTOR__);
+	DML_LOG_VERBOSE("DML::%s: ReadBandwidth = %f\n", __func__, ReadBandwidth);
+	DML_LOG_VERBOSE("DML::%s: ReturnBusWidth = %u\n", __func__, ReturnBusWidth);
+	DML_LOG_VERBOSE("DML::%s: DCFClkDeepSleep = %f\n", __func__, *DCFClkDeepSleep);
+#endif
+
+	for (unsigned int k = 0; k < NumberOfActiveSurfaces; ++k) {
+		*DCFClkDeepSleep = math_max2(*DCFClkDeepSleep, DCFClkDeepSleepPerSurface[k]);
+	}
+
+	DML_LOG_VERBOSE("DML::%s: DCFClkDeepSleep = %f (final)\n", __func__, *DCFClkDeepSleep);
+}
+
+static noinline_for_stack void CalculateDCFCLKDeepSleep(
+	const struct dml2_display_cfg *display_cfg,
+	unsigned int NumberOfActiveSurfaces,
+	unsigned int BytePerPixelY[],
+	unsigned int BytePerPixelC[],
+	unsigned int SwathWidthY[],
+	unsigned int SwathWidthC[],
+	unsigned int DPPPerSurface[],
+	double PSCL_THROUGHPUT[],
+	double PSCL_THROUGHPUT_CHROMA[],
+	double Dppclk[],
+	double ReadBandwidthLuma[],
+	double ReadBandwidthChroma[],
+	unsigned int ReturnBusWidth,
+
+	// Output
+	double *DCFClkDeepSleep)
+{
+	double zero_double[DML2_MAX_PLANES];
+	unsigned int zero_integer[DML2_MAX_PLANES];
+
+	memset(zero_double, 0, DML2_MAX_PLANES * sizeof(double));
+	memset(zero_integer, 0, DML2_MAX_PLANES * sizeof(unsigned int));
+
+	CalculateDCFCLKDeepSleepTdlut(
+		display_cfg,
+		NumberOfActiveSurfaces,
+		BytePerPixelY,
+		BytePerPixelC,
+		SwathWidthY,
+		SwathWidthC,
+		DPPPerSurface,
+		PSCL_THROUGHPUT,
+		PSCL_THROUGHPUT_CHROMA,
+		Dppclk,
+		ReadBandwidthLuma,
+		ReadBandwidthChroma,
+		ReturnBusWidth,
+		0,
+		zero_integer, //tdlut_bytes_to_deliver,
+		zero_double, //prefetch_swath_time_us,
+
+		// Output
+		DCFClkDeepSleep);
+}
+
+static double CalculateWriteBackDelay(
+	enum dml2_source_format_class WritebackPixelFormat,
+	double WritebackHRatio,
+	double WritebackVRatio,
+	unsigned int WritebackVTaps,
+	unsigned int WritebackDestinationWidth,
+	unsigned int WritebackDestinationHeight,
+	unsigned int WritebackSourceHeight,
+	unsigned int HTotal)
+{
+	double CalculateWriteBackDelay;
+	double Line_length;
+	double Output_lines_last_notclamped;
+	double WritebackVInit;
+
+	WritebackVInit = (WritebackVRatio + WritebackVTaps + 1) / 2;
+	Line_length = math_max2((double)WritebackDestinationWidth, math_ceil2((double)WritebackDestinationWidth / 6.0, 1.0) * WritebackVTaps);
+	Output_lines_last_notclamped = WritebackDestinationHeight - 1 - math_ceil2(((double)WritebackSourceHeight - (double)WritebackVInit) / (double)WritebackVRatio, 1.0);
+	if (Output_lines_last_notclamped < 0) {
+		CalculateWriteBackDelay = 0;
+	} else {
+		CalculateWriteBackDelay = Output_lines_last_notclamped * Line_length + (HTotal - WritebackDestinationWidth) + 80;
+	}
+	return CalculateWriteBackDelay;
+}
+
+static unsigned int CalculateMaxVStartup(
+	bool ptoi_supported,
+	unsigned int vblank_nom_default_us,
+	const struct dml2_timing_cfg *timing,
+	double write_back_delay_us)
+{
+	unsigned int vblank_size = 0;
+	unsigned int max_vstartup_lines = 0;
+
+	double line_time_us = (double)timing->h_total / ((double)timing->pixel_clock_khz / 1000);
+	unsigned int vblank_actual = timing->v_total - timing->v_active;
+	unsigned int vblank_nom_default_in_line = (unsigned int)math_floor2((double)vblank_nom_default_us / line_time_us, 1.0);
+	unsigned int vblank_avail = (timing->vblank_nom == 0) ? vblank_nom_default_in_line : (unsigned int)timing->vblank_nom;
+
+	vblank_size = (unsigned int)math_min2(vblank_actual, vblank_avail);
+
+	if (timing->interlaced && !ptoi_supported)
+		max_vstartup_lines = (unsigned int)(math_floor2((vblank_size - 1) / 2.0, 1.0));
+	else
+		max_vstartup_lines = vblank_size - (unsigned int)math_max2(1.0, math_ceil2(write_back_delay_us / line_time_us, 1.0));
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: VBlankNom = %lu\n", __func__, timing->vblank_nom);
+	DML_LOG_VERBOSE("DML::%s: vblank_nom_default_us = %u\n", __func__, vblank_nom_default_us);
+	DML_LOG_VERBOSE("DML::%s: line_time_us = %f\n", __func__, line_time_us);
+	DML_LOG_VERBOSE("DML::%s: vblank_actual = %u\n", __func__, vblank_actual);
+	DML_LOG_VERBOSE("DML::%s: vblank_avail = %u\n", __func__, vblank_avail);
+	DML_LOG_VERBOSE("DML::%s: max_vstartup_lines = %u\n", __func__, max_vstartup_lines);
+#endif
+	max_vstartup_lines = (unsigned int)math_min2(max_vstartup_lines, DML_MAX_VSTARTUP_START);
+	return max_vstartup_lines;
+}
+
+static void CalculateSwathAndDETConfiguration(struct dml2_core_internal_scratch *scratch,
+	struct dml2_core_calcs_CalculateSwathAndDETConfiguration_params *p)
+{
+	unsigned int MaximumSwathHeightY[DML2_MAX_PLANES] = { 0 };
+	unsigned int MaximumSwathHeightC[DML2_MAX_PLANES] = { 0 };
+	unsigned int RoundedUpSwathSizeBytesY[DML2_MAX_PLANES] = { 0 };
+	unsigned int RoundedUpSwathSizeBytesC[DML2_MAX_PLANES] = { 0 };
+	unsigned int SwathWidthSingleDPP[DML2_MAX_PLANES] = { 0 };
+	unsigned int SwathWidthSingleDPPChroma[DML2_MAX_PLANES] = { 0 };
+
+	unsigned int TotalActiveDPP = 0;
+	bool NoChromaOrLinear = true;
+	unsigned int SurfaceDoingUnboundedRequest = 0;
+	unsigned int DETBufferSizeInKByteForSwathCalculation;
+
+	const long TTUFIFODEPTH = 8;
+	const long MAXIMUMCOMPRESSION = 4;
+
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: ForceSingleDPP = %u\n", __func__, p->ForceSingleDPP);
+	for (unsigned int k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+		DML_LOG_VERBOSE("DML::%s: DPPPerSurface[%u] = %u\n", __func__, k, p->DPPPerSurface[k]);
+	}
+#endif
+	CalculateSwathWidth(
+		p->display_cfg,
+		p->ForceSingleDPP,
+		p->NumberOfActiveSurfaces,
+		p->ODMMode,
+		p->BytePerPixY,
+		p->BytePerPixC,
+		p->Read256BytesBlockHeightY,
+		p->Read256BytesBlockHeightC,
+		p->Read256BytesBlockWidthY,
+		p->Read256BytesBlockWidthC,
+		p->surf_linear128_l,
+		p->surf_linear128_c,
+		p->DPPPerSurface,
+
+		// Output
+		p->req_per_swath_ub_l,
+		p->req_per_swath_ub_c,
+		SwathWidthSingleDPP,
+		SwathWidthSingleDPPChroma,
+		p->SwathWidth,
+		p->SwathWidthChroma,
+		MaximumSwathHeightY,
+		MaximumSwathHeightC,
+		p->swath_width_luma_ub,
+		p->swath_width_chroma_ub);
+
+	for (unsigned int k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+		p->full_swath_bytes_l[k] = (unsigned int)(p->swath_width_luma_ub[k] * p->BytePerPixDETY[k] * MaximumSwathHeightY[k]);
+		p->full_swath_bytes_c[k] = (unsigned int)(p->swath_width_chroma_ub[k] * p->BytePerPixDETC[k] * MaximumSwathHeightC[k]);
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: k=%u DPPPerSurface = %u\n", __func__, k, p->DPPPerSurface[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u swath_width_luma_ub = %u\n", __func__, k, p->swath_width_luma_ub[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u BytePerPixDETY = %f\n", __func__, k, p->BytePerPixDETY[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u MaximumSwathHeightY = %u\n", __func__, k, MaximumSwathHeightY[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u full_swath_bytes_l = %u\n", __func__, k, p->full_swath_bytes_l[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u swath_width_chroma_ub = %u\n", __func__, k, p->swath_width_chroma_ub[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u BytePerPixDETC = %f\n", __func__, k, p->BytePerPixDETC[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u MaximumSwathHeightC = %u\n", __func__, k, MaximumSwathHeightC[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u full_swath_bytes_c = %u\n", __func__, k, p->full_swath_bytes_c[k]);
+#endif
+		if (p->display_cfg->plane_descriptors[k].pixel_format == dml2_420_10) {
+			p->full_swath_bytes_l[k] = (unsigned int)(math_ceil2((double)p->full_swath_bytes_l[k], 256));
+			p->full_swath_bytes_c[k] = (unsigned int)(math_ceil2((double)p->full_swath_bytes_c[k], 256));
+		}
+	}
+
+	for (unsigned int k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+		TotalActiveDPP = TotalActiveDPP + (p->ForceSingleDPP ? 1 : p->DPPPerSurface[k]);
+		if (p->DPPPerSurface[k] > 0)
+			SurfaceDoingUnboundedRequest = k;
+		if (dml_is_420(p->display_cfg->plane_descriptors[k].pixel_format) || p->display_cfg->plane_descriptors[k].pixel_format == dml2_rgbe_alpha
+			|| p->display_cfg->plane_descriptors[k].surface.tiling == dml2_sw_linear) {
+			NoChromaOrLinear = false;
+		}
+	}
+
+	*p->UnboundedRequestEnabled = UnboundedRequest(p->display_cfg->overrides.hw.force_unbounded_requesting.enable, p->display_cfg->overrides.hw.force_unbounded_requesting.value, TotalActiveDPP, NoChromaOrLinear);
+
+	CalculateDETBufferSize(
+		&scratch->CalculateDETBufferSize_locals,
+		p->display_cfg,
+		p->ForceSingleDPP,
+		p->NumberOfActiveSurfaces,
+		*p->UnboundedRequestEnabled,
+		p->nomDETInKByte,
+		p->MaxTotalDETInKByte,
+		p->ConfigReturnBufferSizeInKByte,
+		p->MinCompressedBufferSizeInKByte,
+		p->ConfigReturnBufferSegmentSizeInkByte,
+		p->CompressedBufferSegmentSizeInkByte,
+		p->ReadBandwidthLuma,
+		p->ReadBandwidthChroma,
+		p->full_swath_bytes_l,
+		p->full_swath_bytes_c,
+		p->DPPPerSurface,
+
+		// Output
+		p->DETBufferSizeInKByte, // per hubp pipe
+		p->CompressedBufferSizeInkByte);
+
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: TotalActiveDPP = %u\n", __func__, TotalActiveDPP);
+	DML_LOG_VERBOSE("DML::%s: nomDETInKByte = %u\n", __func__, p->nomDETInKByte);
+	DML_LOG_VERBOSE("DML::%s: ConfigReturnBufferSizeInKByte = %u\n", __func__, p->ConfigReturnBufferSizeInKByte);
+	DML_LOG_VERBOSE("DML::%s: UnboundedRequestEnabled = %u\n", __func__, *p->UnboundedRequestEnabled);
+	DML_LOG_VERBOSE("DML::%s: CompressedBufferSizeInkByte = %u\n", __func__, *p->CompressedBufferSizeInkByte);
+#endif
+
+	*p->ViewportSizeSupport = true;
+	for (unsigned int k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+
+		DETBufferSizeInKByteForSwathCalculation = (dml_is_phantom_pipe(&p->display_cfg->plane_descriptors[k]) ? 1024 : p->DETBufferSizeInKByte[k]);
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: k=%u DETBufferSizeInKByteForSwathCalculation = %u\n", __func__, k, DETBufferSizeInKByteForSwathCalculation);
+#endif
+		if (p->display_cfg->plane_descriptors[k].surface.tiling == dml2_sw_linear) {
+			p->SwathHeightY[k] = MaximumSwathHeightY[k];
+			p->SwathHeightC[k] = MaximumSwathHeightC[k];
+			RoundedUpSwathSizeBytesY[k] = p->full_swath_bytes_l[k];
+			RoundedUpSwathSizeBytesC[k] = p->full_swath_bytes_c[k];
+
+			if (p->surf_linear128_l[k])
+				p->request_size_bytes_luma[k] = 128;
+			else
+				p->request_size_bytes_luma[k] = 256;
+
+			if (p->surf_linear128_c[k])
+				p->request_size_bytes_chroma[k] = 128;
+			else
+				p->request_size_bytes_chroma[k] = 256;
+
+		} else if (p->full_swath_bytes_l[k] + p->full_swath_bytes_c[k] <= DETBufferSizeInKByteForSwathCalculation * 1024 / 2) {
+			p->SwathHeightY[k] = MaximumSwathHeightY[k];
+			p->SwathHeightC[k] = MaximumSwathHeightC[k];
+			RoundedUpSwathSizeBytesY[k] = p->full_swath_bytes_l[k];
+			RoundedUpSwathSizeBytesC[k] = p->full_swath_bytes_c[k];
+			p->request_size_bytes_luma[k] = 256;
+			p->request_size_bytes_chroma[k] = 256;
+
+		} else if (p->full_swath_bytes_l[k] >= 1.5 * p->full_swath_bytes_c[k] && p->full_swath_bytes_l[k] / 2 + p->full_swath_bytes_c[k] <= DETBufferSizeInKByteForSwathCalculation * 1024 / 2) {
+			p->SwathHeightY[k] = MaximumSwathHeightY[k] / 2;
+			p->SwathHeightC[k] = MaximumSwathHeightC[k];
+			RoundedUpSwathSizeBytesY[k] = p->full_swath_bytes_l[k] / 2;
+			RoundedUpSwathSizeBytesC[k] = p->full_swath_bytes_c[k];
+			p->request_size_bytes_luma[k] = ((p->BytePerPixY[k] == 2) == dml_is_vertical_rotation(p->display_cfg->plane_descriptors[k].composition.rotation_angle)) ? 128 : 64;
+			p->request_size_bytes_chroma[k] = 256;
+
+		} else if (p->full_swath_bytes_l[k] < 1.5 * p->full_swath_bytes_c[k] && p->full_swath_bytes_l[k] + p->full_swath_bytes_c[k] / 2 <= DETBufferSizeInKByteForSwathCalculation * 1024 / 2) {
+			p->SwathHeightY[k] = MaximumSwathHeightY[k];
+			p->SwathHeightC[k] = MaximumSwathHeightC[k] / 2;
+			RoundedUpSwathSizeBytesY[k] = p->full_swath_bytes_l[k];
+			RoundedUpSwathSizeBytesC[k] = p->full_swath_bytes_c[k] / 2;
+			p->request_size_bytes_luma[k] = 256;
+			p->request_size_bytes_chroma[k] = ((p->BytePerPixC[k] == 2) == dml_is_vertical_rotation(p->display_cfg->plane_descriptors[k].composition.rotation_angle)) ? 128 : 64;
+
+		} else {
+			p->SwathHeightY[k] = MaximumSwathHeightY[k] / 2;
+			p->SwathHeightC[k] = MaximumSwathHeightC[k] / 2;
+			RoundedUpSwathSizeBytesY[k] = p->full_swath_bytes_l[k] / 2;
+			RoundedUpSwathSizeBytesC[k] = p->full_swath_bytes_c[k] / 2;
+			p->request_size_bytes_luma[k] = ((p->BytePerPixY[k] == 2) == dml_is_vertical_rotation(p->display_cfg->plane_descriptors[k].composition.rotation_angle)) ? 128 : 64;
+			p->request_size_bytes_chroma[k] = ((p->BytePerPixC[k] == 2) == dml_is_vertical_rotation(p->display_cfg->plane_descriptors[k].composition.rotation_angle)) ? 128 : 64;
+		}
+
+		if (p->SwathHeightC[k] == 0)
+			p->request_size_bytes_chroma[k] = 0;
+
+		if ((p->full_swath_bytes_l[k] / 2 + p->full_swath_bytes_c[k] / 2 > DETBufferSizeInKByteForSwathCalculation * 1024 / 2) ||
+			p->SwathWidth[k] > p->MaximumSwathWidthLuma[k] || (p->SwathHeightC[k] > 0 && p->SwathWidthChroma[k] > p->MaximumSwathWidthChroma[k])) {
+			*p->ViewportSizeSupport = false;
+			DML_LOG_VERBOSE("DML::%s: k=%u full_swath_bytes_l=%u\n", __func__, k, p->full_swath_bytes_l[k]);
+			DML_LOG_VERBOSE("DML::%s: k=%u full_swath_bytes_c=%u\n", __func__, k, p->full_swath_bytes_c[k]);
+			DML_LOG_VERBOSE("DML::%s: k=%u DETBufferSizeInKByteForSwathCalculation=%u\n", __func__, k, DETBufferSizeInKByteForSwathCalculation);
+			DML_LOG_VERBOSE("DML::%s: k=%u SwathWidth=%u\n", __func__, k, p->SwathWidth[k]);
+			DML_LOG_VERBOSE("DML::%s: k=%u MaximumSwathWidthLuma=%f\n", __func__, k, p->MaximumSwathWidthLuma[k]);
+			DML_LOG_VERBOSE("DML::%s: k=%u SwathWidthChroma=%d\n", __func__, k, p->SwathWidthChroma[k]);
+			DML_LOG_VERBOSE("DML::%s: k=%u MaximumSwathWidthChroma=%f\n", __func__, k, p->MaximumSwathWidthChroma[k]);
+			p->ViewportSizeSupportPerSurface[k] = false;
+		} else {
+			p->ViewportSizeSupportPerSurface[k] = true;
+		}
+
+		if (p->SwathHeightC[k] == 0) {
+#ifdef __DML_VBA_DEBUG__
+			DML_LOG_VERBOSE("DML::%s: k=%u, All DET will be used for plane0\n", __func__, k);
+#endif
+			p->DETBufferSizeY[k] = p->DETBufferSizeInKByte[k] * 1024;
+			p->DETBufferSizeC[k] = 0;
+		} else if (RoundedUpSwathSizeBytesY[k] <= 1.5 * RoundedUpSwathSizeBytesC[k]) {
+#ifdef __DML_VBA_DEBUG__
+			DML_LOG_VERBOSE("DML::%s: k=%u, Half DET will be used for plane0, and half for plane1\n", __func__, k);
+#endif
+			p->DETBufferSizeY[k] = p->DETBufferSizeInKByte[k] * 1024 / 2;
+			p->DETBufferSizeC[k] = p->DETBufferSizeInKByte[k] * 1024 / 2;
+		} else {
+#ifdef __DML_VBA_DEBUG__
+			DML_LOG_VERBOSE("DML::%s: k=%u, 2/3 DET will be used for plane0, and 1/3 for plane1\n", __func__, k);
+#endif
+			p->DETBufferSizeY[k] = (unsigned int)(math_floor2(p->DETBufferSizeInKByte[k] * 1024 * 2 / 3, 1024));
+			p->DETBufferSizeC[k] = p->DETBufferSizeInKByte[k] * 1024 - p->DETBufferSizeY[k];
+		}
+
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: k=%u SwathHeightY = %u\n", __func__, k, p->SwathHeightY[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u SwathHeightC = %u\n", __func__, k, p->SwathHeightC[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u full_swath_bytes_l = %u\n", __func__, k, p->full_swath_bytes_l[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u full_swath_bytes_c = %u\n", __func__, k, p->full_swath_bytes_c[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u RoundedUpSwathSizeBytesY = %u\n", __func__, k, RoundedUpSwathSizeBytesY[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u RoundedUpSwathSizeBytesC = %u\n", __func__, k, RoundedUpSwathSizeBytesC[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u DETBufferSizeInKByte = %u\n", __func__, k, p->DETBufferSizeInKByte[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u DETBufferSizeY = %u\n", __func__, k, p->DETBufferSizeY[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u DETBufferSizeC = %u\n", __func__, k, p->DETBufferSizeC[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u ViewportSizeSupportPerSurface = %u\n", __func__, k, p->ViewportSizeSupportPerSurface[k]);
+#endif
+
+	}
+
+	*p->compbuf_reserved_space_64b = 2 * p->pixel_chunk_size_kbytes * 1024 / 64;
+	if (*p->UnboundedRequestEnabled) {
+		*p->compbuf_reserved_space_64b = (unsigned int)math_ceil2(math_max2(*p->compbuf_reserved_space_64b,
+			(double)(p->rob_buffer_size_kbytes * 1024 / 64) - (double)(RoundedUpSwathSizeBytesY[SurfaceDoingUnboundedRequest] * TTUFIFODEPTH / (p->mrq_present ? MAXIMUMCOMPRESSION : 1) / 64)), 1.0);
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: RoundedUpSwathSizeBytesY[%d] = %u\n", __func__, SurfaceDoingUnboundedRequest, RoundedUpSwathSizeBytesY[SurfaceDoingUnboundedRequest]);
+		DML_LOG_VERBOSE("DML::%s: rob_buffer_size_kbytes = %u\n", __func__, p->rob_buffer_size_kbytes);
+#endif
+	}
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: compbuf_reserved_space_64b = %u\n", __func__, *p->compbuf_reserved_space_64b);
+#endif
+
+	*p->hw_debug5 = false;
+#ifdef ALLOW_SDPIF_RATE_LIMIT_PRE_CSTATE
+	if (p->NumberOfActiveSurfaces > 1)
+		*p->hw_debug5 = true;
+#else
+	for (unsigned int k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+		if (!(p->mrq_present) && (!(*p->UnboundedRequestEnabled)) && (TotalActiveDPP == 1)
+			&& p->display_cfg->plane_descriptors[k].surface.dcc.enable
+			&& ((p->rob_buffer_size_kbytes * 1024 * (p->mrq_present ? MAXIMUMCOMPRESSION : 1)
+				+ *p->CompressedBufferSizeInkByte * MAXIMUMCOMPRESSION * 1024) > TTUFIFODEPTH * (RoundedUpSwathSizeBytesY[k] + RoundedUpSwathSizeBytesC[k])))
+			*p->hw_debug5 = true;
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: k=%u UnboundedRequestEnabled = %u\n", __func__, k, *p->UnboundedRequestEnabled);
+		DML_LOG_VERBOSE("DML::%s: k=%u MAXIMUMCOMPRESSION = %lu\n", __func__, k, MAXIMUMCOMPRESSION);
+		DML_LOG_VERBOSE("DML::%s: k=%u TTUFIFODEPTH = %lu\n", __func__, k, TTUFIFODEPTH);
+		DML_LOG_VERBOSE("DML::%s: k=%u CompressedBufferSizeInkByte = %u\n", __func__, k, *p->CompressedBufferSizeInkByte);
+		DML_LOG_VERBOSE("DML::%s: k=%u RoundedUpSwathSizeBytesC = %u\n", __func__, k, RoundedUpSwathSizeBytesC[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u hw_debug5 = %u\n", __func__, k, *p->hw_debug5);
+#endif
+	}
+#endif
+}
+
+static enum dml2_odm_mode DecideODMMode(unsigned int HActive,
+	double MaxDispclk,
+	unsigned int MaximumPixelsPerLinePerDSCUnit,
+	enum dml2_output_format_class OutFormat,
+	bool UseDSC,
+	unsigned int NumberOfDSCSlices,
+	double SurfaceRequiredDISPCLKWithoutODMCombine,
+	double SurfaceRequiredDISPCLKWithODMCombineTwoToOne,
+	double SurfaceRequiredDISPCLKWithODMCombineThreeToOne,
+	double SurfaceRequiredDISPCLKWithODMCombineFourToOne)
+{
+	enum dml2_odm_mode MinimumRequiredODMModeForMaxDispClock;
+	enum dml2_odm_mode MinimumRequiredODMModeForMaxDSCHActive;
+	enum dml2_odm_mode MinimumRequiredODMModeForMax420HActive;
+	enum dml2_odm_mode ODMMode = dml2_odm_mode_bypass;
+
+	MinimumRequiredODMModeForMaxDispClock =
+			(SurfaceRequiredDISPCLKWithoutODMCombine <= MaxDispclk) ? dml2_odm_mode_bypass :
+			(SurfaceRequiredDISPCLKWithODMCombineTwoToOne <= MaxDispclk) ? dml2_odm_mode_combine_2to1 :
+			(SurfaceRequiredDISPCLKWithODMCombineThreeToOne <= MaxDispclk) ? dml2_odm_mode_combine_3to1 : dml2_odm_mode_combine_4to1;
+	if (ODMMode < MinimumRequiredODMModeForMaxDispClock)
+		ODMMode = MinimumRequiredODMModeForMaxDispClock;
+
+	if (UseDSC) {
+		MinimumRequiredODMModeForMaxDSCHActive =
+				(HActive <= 1 * MaximumPixelsPerLinePerDSCUnit) ? dml2_odm_mode_bypass :
+				(HActive <= 2 * MaximumPixelsPerLinePerDSCUnit) ? dml2_odm_mode_combine_2to1 :
+				(HActive <= 3 * MaximumPixelsPerLinePerDSCUnit) ? dml2_odm_mode_combine_3to1 : dml2_odm_mode_combine_4to1;
+		if (ODMMode < MinimumRequiredODMModeForMaxDSCHActive)
+			ODMMode = MinimumRequiredODMModeForMaxDSCHActive;
+	}
+
+	if (OutFormat == dml2_420) {
+		MinimumRequiredODMModeForMax420HActive =
+				(HActive <= 1 * DML2_MAX_FMT_420_BUFFER_WIDTH) ? dml2_odm_mode_bypass :
+				(HActive <= 2 * DML2_MAX_FMT_420_BUFFER_WIDTH) ? dml2_odm_mode_combine_2to1 :
+				(HActive <= 3 * DML2_MAX_FMT_420_BUFFER_WIDTH) ? dml2_odm_mode_combine_3to1 : dml2_odm_mode_combine_4to1;
+		if (ODMMode < MinimumRequiredODMModeForMax420HActive)
+			ODMMode = MinimumRequiredODMModeForMax420HActive;
+	}
+
+	if (UseDSC) {
+		if (ODMMode == dml2_odm_mode_bypass && NumberOfDSCSlices > 4)
+			ODMMode = dml2_odm_mode_combine_2to1;
+		if (ODMMode == dml2_odm_mode_combine_2to1 && NumberOfDSCSlices > 8)
+			ODMMode = dml2_odm_mode_combine_3to1;
+		if (ODMMode == dml2_odm_mode_combine_3to1 && NumberOfDSCSlices != 12)
+			ODMMode = dml2_odm_mode_combine_4to1;
+	}
+
+	return ODMMode;
+}
+
+static void CalculateODMConstraints(
+	enum dml2_odm_mode ODMUse,
+	double SurfaceRequiredDISPCLKWithoutODMCombine,
+	double SurfaceRequiredDISPCLKWithODMCombineTwoToOne,
+	double SurfaceRequiredDISPCLKWithODMCombineThreeToOne,
+	double SurfaceRequiredDISPCLKWithODMCombineFourToOne,
+	unsigned int MaximumPixelsPerLinePerDSCUnit,
+	/* Output */
+	double *DISPCLKRequired,
+	unsigned int *NumberOfDPPRequired,
+	unsigned int *MaxHActiveForDSC,
+	unsigned int *MaxDSCSlices,
+	unsigned int *MaxHActiveFor420)
+{
+	switch (ODMUse) {
+	case dml2_odm_mode_combine_2to1:
+		*DISPCLKRequired = SurfaceRequiredDISPCLKWithODMCombineTwoToOne;
+		*NumberOfDPPRequired = 2;
+		break;
+	case dml2_odm_mode_combine_3to1:
+		*DISPCLKRequired = SurfaceRequiredDISPCLKWithODMCombineThreeToOne;
+		*NumberOfDPPRequired = 3;
+		break;
+	case dml2_odm_mode_combine_4to1:
+		*DISPCLKRequired = SurfaceRequiredDISPCLKWithODMCombineFourToOne;
+		*NumberOfDPPRequired = 4;
+		break;
+	case dml2_odm_mode_auto:
+	case dml2_odm_mode_split_1to2:
+	case dml2_odm_mode_mso_1to2:
+	case dml2_odm_mode_mso_1to4:
+	case dml2_odm_mode_bypass:
+	default:
+		*DISPCLKRequired = SurfaceRequiredDISPCLKWithoutODMCombine;
+		*NumberOfDPPRequired = 1;
+		break;
+	}
+	*MaxHActiveForDSC = *NumberOfDPPRequired * MaximumPixelsPerLinePerDSCUnit;
+	*MaxDSCSlices = *NumberOfDPPRequired * DML_MAX_NUM_OF_SLICES_PER_DSC;
+	*MaxHActiveFor420 = *NumberOfDPPRequired * DML2_MAX_FMT_420_BUFFER_WIDTH;
+}
+
+static bool ValidateODMMode(enum dml2_odm_mode ODMMode,
+	double MaxDispclk,
+	unsigned int HActive,
+	enum dml2_output_format_class OutFormat,
+	bool UseDSC,
+	unsigned int NumberOfDSCSlices,
+	unsigned int TotalNumberOfActiveDPP,
+	unsigned int TotalNumberOfActiveOPP,
+	unsigned int MaxNumDPP,
+	unsigned int MaxNumOPP,
+	double DISPCLKRequired,
+	unsigned int NumberOfDPPRequired,
+	unsigned int MaxHActiveForDSC,
+	unsigned int MaxDSCSlices,
+	unsigned int MaxHActiveFor420)
+{
+	bool are_odm_segments_symmetrical = (ODMMode == dml2_odm_mode_combine_3to1) ? UseDSC : true;
+	bool is_max_dsc_slice_required = (ODMMode == dml2_odm_mode_combine_3to1);
+	unsigned int pixels_per_clock_cycle = (OutFormat == dml2_420 || OutFormat == dml2_n422) ? 2 : 1;
+	unsigned int h_timing_div_mode =
+			(ODMMode == dml2_odm_mode_combine_4to1 || ODMMode == dml2_odm_mode_combine_3to1) ? 4 :
+			(ODMMode == dml2_odm_mode_combine_2to1) ? 2 : pixels_per_clock_cycle;
+
+	if (DISPCLKRequired > MaxDispclk)
+		return false;
+	if ((TotalNumberOfActiveDPP + NumberOfDPPRequired) > MaxNumDPP || (TotalNumberOfActiveOPP + NumberOfDPPRequired) > MaxNumOPP)
+		return false;
+	if (are_odm_segments_symmetrical) {
+		if (HActive % (NumberOfDPPRequired * pixels_per_clock_cycle))
+			return false;
+	}
+	if (HActive % h_timing_div_mode)
+		/*
+		 * TODO - OTG_H_TOTAL, OTG_H_BLANK_START/END and
+		 * OTG_H_SYNC_A_START/END all need to be visible by h timing div
+		 * mode. This logic only checks H active.
+		 */
+		return false;
+
+	if (UseDSC) {
+		if (HActive > MaxHActiveForDSC)
+			return false;
+		if (NumberOfDSCSlices > MaxDSCSlices)
+			return false;
+		if (HActive % NumberOfDSCSlices)
+			return false;
+		if (NumberOfDSCSlices % NumberOfDPPRequired)
+			return false;
+		if (is_max_dsc_slice_required) {
+			if (NumberOfDSCSlices != MaxDSCSlices)
+				return false;
+		}
+	}
+
+	if (OutFormat == dml2_420) {
+		if (HActive > MaxHActiveFor420)
+			return false;
+	}
+
+	return true;
+}
+
+static noinline_for_stack void CalculateODMMode(
+	unsigned int MaximumPixelsPerLinePerDSCUnit,
+	unsigned int HActive,
+	enum dml2_output_format_class OutFormat,
+	enum dml2_output_encoder_class Output,
+	enum dml2_odm_mode ODMUse,
+	double MaxDispclk,
+	bool DSCEnable,
+	unsigned int TotalNumberOfActiveDPP,
+	unsigned int TotalNumberOfActiveOPP,
+	unsigned int MaxNumDPP,
+	unsigned int MaxNumOPP,
+	double PixelClock,
+	unsigned int NumberOfDSCSlices,
+
+	// Output
+	bool *TotalAvailablePipesSupport,
+	unsigned int *NumberOfDPP,
+	enum dml2_odm_mode *ODMMode,
+	double *RequiredDISPCLKPerSurface)
+{
+	double SurfaceRequiredDISPCLKWithoutODMCombine;
+	double SurfaceRequiredDISPCLKWithODMCombineTwoToOne;
+	double SurfaceRequiredDISPCLKWithODMCombineThreeToOne;
+	double SurfaceRequiredDISPCLKWithODMCombineFourToOne;
+	double DISPCLKRequired;
+	unsigned int NumberOfDPPRequired;
+	unsigned int MaxHActiveForDSC;
+	unsigned int MaxDSCSlices;
+	unsigned int MaxHActiveFor420;
+	bool success;
+	bool UseDSC = DSCEnable && (NumberOfDSCSlices > 0);
+	enum dml2_odm_mode DecidedODMMode;
+	bool isTMDS420 = (OutFormat == dml2_420 && Output == dml2_hdmi);
+
+	SurfaceRequiredDISPCLKWithoutODMCombine = CalculateRequiredDispclk(dml2_odm_mode_bypass, PixelClock, isTMDS420);
+	SurfaceRequiredDISPCLKWithODMCombineTwoToOne = CalculateRequiredDispclk(dml2_odm_mode_combine_2to1, PixelClock, isTMDS420);
+	SurfaceRequiredDISPCLKWithODMCombineThreeToOne = CalculateRequiredDispclk(dml2_odm_mode_combine_3to1, PixelClock, isTMDS420);
+	SurfaceRequiredDISPCLKWithODMCombineFourToOne = CalculateRequiredDispclk(dml2_odm_mode_combine_4to1, PixelClock, isTMDS420);
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: ODMUse = %d\n", __func__, ODMUse);
+	DML_LOG_VERBOSE("DML::%s: Output = %d\n", __func__, Output);
+	DML_LOG_VERBOSE("DML::%s: DSCEnable = %d\n", __func__, DSCEnable);
+	DML_LOG_VERBOSE("DML::%s: MaxDispclk = %f\n", __func__, MaxDispclk);
+	DML_LOG_VERBOSE("DML::%s: MaximumPixelsPerLinePerDSCUnit = %d\n", __func__, MaximumPixelsPerLinePerDSCUnit);
+	DML_LOG_VERBOSE("DML::%s: SurfaceRequiredDISPCLKWithoutODMCombine = %f\n", __func__, SurfaceRequiredDISPCLKWithoutODMCombine);
+	DML_LOG_VERBOSE("DML::%s: SurfaceRequiredDISPCLKWithODMCombineTwoToOne = %f\n", __func__, SurfaceRequiredDISPCLKWithODMCombineTwoToOne);
+	DML_LOG_VERBOSE("DML::%s: SurfaceRequiredDISPCLKWithODMCombineThreeToOne = %f\n", __func__, SurfaceRequiredDISPCLKWithODMCombineThreeToOne);
+	DML_LOG_VERBOSE("DML::%s: SurfaceRequiredDISPCLKWithODMCombineFourToOne = %f\n", __func__, SurfaceRequiredDISPCLKWithODMCombineFourToOne);
+#endif
+	if (ODMUse == dml2_odm_mode_auto)
+		DecidedODMMode = DecideODMMode(HActive,
+				MaxDispclk,
+				MaximumPixelsPerLinePerDSCUnit,
+				OutFormat,
+				UseDSC,
+				NumberOfDSCSlices,
+				SurfaceRequiredDISPCLKWithoutODMCombine,
+				SurfaceRequiredDISPCLKWithODMCombineTwoToOne,
+				SurfaceRequiredDISPCLKWithODMCombineThreeToOne,
+				SurfaceRequiredDISPCLKWithODMCombineFourToOne);
+	else
+		DecidedODMMode = ODMUse;
+	CalculateODMConstraints(DecidedODMMode,
+			SurfaceRequiredDISPCLKWithoutODMCombine,
+			SurfaceRequiredDISPCLKWithODMCombineTwoToOne,
+			SurfaceRequiredDISPCLKWithODMCombineThreeToOne,
+			SurfaceRequiredDISPCLKWithODMCombineFourToOne,
+			MaximumPixelsPerLinePerDSCUnit,
+			&DISPCLKRequired,
+			&NumberOfDPPRequired,
+			&MaxHActiveForDSC,
+			&MaxDSCSlices,
+			&MaxHActiveFor420);
+	success = ValidateODMMode(DecidedODMMode,
+			MaxDispclk,
+			HActive,
+			OutFormat,
+			UseDSC,
+			NumberOfDSCSlices,
+			TotalNumberOfActiveDPP,
+			TotalNumberOfActiveOPP,
+			MaxNumDPP,
+			MaxNumOPP,
+			DISPCLKRequired,
+			NumberOfDPPRequired,
+			MaxHActiveForDSC,
+			MaxDSCSlices,
+			MaxHActiveFor420);
+
+	*ODMMode = DecidedODMMode;
+	*TotalAvailablePipesSupport = success;
+	*NumberOfDPP = NumberOfDPPRequired;
+	*RequiredDISPCLKPerSurface = success ? DISPCLKRequired : 0;
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: ODMMode = %d\n", __func__, *ODMMode);
+	DML_LOG_VERBOSE("DML::%s: NumberOfDPP = %d\n", __func__, *NumberOfDPP);
+	DML_LOG_VERBOSE("DML::%s: TotalAvailablePipesSupport = %d\n", __func__, *TotalAvailablePipesSupport);
+	DML_LOG_VERBOSE("DML::%s: RequiredDISPCLKPerSurface = %f\n", __func__, *RequiredDISPCLKPerSurface);
+#endif
+}
+
+static noinline_for_stack void CalculateOutputLink(
+	struct dml2_core_internal_scratch *s,
+	double PHYCLK,
+	double PHYCLKD18,
+	double PHYCLKD32,
+	double Downspreading,
+	enum dml2_output_encoder_class Output,
+	enum dml2_output_format_class OutputFormat,
+	unsigned int HTotal,
+	unsigned int HActive,
+	double PixelClockBackEnd,
+	double ForcedOutputLinkBPP,
+	unsigned int DSCInputBitPerComponent,
+	unsigned int NumberOfDSCSlices,
+	double AudioSampleRate,
+	unsigned int AudioSampleLayout,
+	enum dml2_odm_mode ODMModeNoDSC,
+	enum dml2_odm_mode ODMModeDSC,
+	enum dml2_dsc_enable_option DSCEnable,
+	unsigned int OutputLinkDPLanes,
+	enum dml2_output_link_dp_rate OutputLinkDPRate,
+
+	// Output
+	bool *RequiresDSC,
+	bool *RequiresFEC,
+	double *OutBpp,
+	enum dml2_core_internal_output_type *OutputType,
+	enum dml2_core_internal_output_type_rate *OutputRate,
+	unsigned int *RequiredSlots)
+{
+	bool LinkDSCEnable;
+	unsigned int dummy;
+	*RequiresDSC = false;
+	*RequiresFEC = false;
+	*OutBpp = 0;
+
+	*OutputType = dml2_core_internal_output_type_unknown;
+	*OutputRate = dml2_core_internal_output_rate_unknown;
+
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: DSCEnable = %u (dis, en, en_if_necessary)\n", __func__, DSCEnable);
+	DML_LOG_VERBOSE("DML::%s: PHYCLK = %f\n", __func__, PHYCLK);
+	DML_LOG_VERBOSE("DML::%s: PixelClockBackEnd = %f\n", __func__, PixelClockBackEnd);
+	DML_LOG_VERBOSE("DML::%s: AudioSampleRate = %f\n", __func__, AudioSampleRate);
+	DML_LOG_VERBOSE("DML::%s: HActive = %u\n", __func__, HActive);
+	DML_LOG_VERBOSE("DML::%s: HTotal = %u\n", __func__, HTotal);
+	DML_LOG_VERBOSE("DML::%s: ODMModeNoDSC = %u\n", __func__, ODMModeNoDSC);
+	DML_LOG_VERBOSE("DML::%s: ODMModeDSC = %u\n", __func__, ODMModeDSC);
+	DML_LOG_VERBOSE("DML::%s: ForcedOutputLinkBPP = %f\n", __func__, ForcedOutputLinkBPP);
+	DML_LOG_VERBOSE("DML::%s: Output (encoder) = %u\n", __func__, Output);
+	DML_LOG_VERBOSE("DML::%s: OutputLinkDPRate = %u\n", __func__, OutputLinkDPRate);
+#endif
+	{
+		if (Output == dml2_hdmi) {
+			*RequiresDSC = false;
+			*RequiresFEC = false;
+			*OutBpp = TruncToValidBPP(&s->TruncToValidBPP_locals, math_min2(600, PHYCLK) * 10, 3, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, false, Output,
+				OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (unsigned int)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, &dummy);
+			//OutputTypeAndRate = "HDMI";
+			*OutputType = dml2_core_internal_output_type_hdmi;
+		} else if (Output == dml2_dp || Output == dml2_dp2p0 || Output == dml2_edp) {
+			if (DSCEnable == dml2_dsc_enable) {
+				*RequiresDSC = true;
+				LinkDSCEnable = true;
+				if (Output == dml2_dp || Output == dml2_dp2p0) {
+					*RequiresFEC = true;
+				} else {
+					*RequiresFEC = false;
+				}
+			} else {
+				*RequiresDSC = false;
+				LinkDSCEnable = false;
+				if (Output == dml2_dp2p0) {
+					*RequiresFEC = true;
+				} else {
+					*RequiresFEC = false;
+				}
+			}
+			if (Output == dml2_dp2p0) {
+				*OutBpp = 0;
+				if ((OutputLinkDPRate == dml2_dp_rate_na || OutputLinkDPRate == dml2_dp_rate_uhbr10) && PHYCLKD32 >= 10000.0 / 32) {
+					*OutBpp = TruncToValidBPP(&s->TruncToValidBPP_locals, (1 - Downspreading / 100) * 10000, OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output,
+						OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (unsigned int)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+					if (*OutBpp == 0 && PHYCLKD32 < 13500.0 / 32 && DSCEnable == dml2_dsc_enable_if_necessary && ForcedOutputLinkBPP == 0) {
+						*RequiresDSC = true;
+						LinkDSCEnable = true;
+						*OutBpp = TruncToValidBPP(&s->TruncToValidBPP_locals, (1 - Downspreading / 100) * 10000, OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output,
+							OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (unsigned int)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+					}
+					//OutputTypeAndRate = Output & " UHBR10";
+					*OutputType = dml2_core_internal_output_type_dp2p0;
+					*OutputRate = dml2_core_internal_output_rate_dp_rate_uhbr10;
+				}
+				if ((OutputLinkDPRate == dml2_dp_rate_na || OutputLinkDPRate == dml2_dp_rate_uhbr13p5) && *OutBpp == 0 && PHYCLKD32 >= 13500.0 / 32) {
+					*OutBpp = TruncToValidBPP(&s->TruncToValidBPP_locals, (1 - Downspreading / 100) * 13500, OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output,
+						OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (unsigned int)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+
+					if (*OutBpp == 0 && PHYCLKD32 < 20000.0 / 32 && DSCEnable == dml2_dsc_enable_if_necessary && ForcedOutputLinkBPP == 0) {
+						*RequiresDSC = true;
+						LinkDSCEnable = true;
+						*OutBpp = TruncToValidBPP(&s->TruncToValidBPP_locals, (1 - Downspreading / 100) * 13500, OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output,
+							OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (unsigned int)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+					}
+					//OutputTypeAndRate = Output & " UHBR13p5";
+					*OutputType = dml2_core_internal_output_type_dp2p0;
+					*OutputRate = dml2_core_internal_output_rate_dp_rate_uhbr13p5;
+				}
+				if ((OutputLinkDPRate == dml2_dp_rate_na || OutputLinkDPRate == dml2_dp_rate_uhbr20) && *OutBpp == 0 && PHYCLKD32 >= 20000.0 / 32) {
+					*OutBpp = TruncToValidBPP(&s->TruncToValidBPP_locals, (1 - Downspreading / 100) * 20000, OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output,
+						OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (unsigned int)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+					if (*OutBpp == 0 && DSCEnable == dml2_dsc_enable_if_necessary && ForcedOutputLinkBPP == 0) {
+						*RequiresDSC = true;
+						LinkDSCEnable = true;
+						*OutBpp = TruncToValidBPP(&s->TruncToValidBPP_locals, (1 - Downspreading / 100) * 20000, OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output,
+							OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (unsigned int)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+					}
+					//OutputTypeAndRate = Output & " UHBR20";
+					*OutputType = dml2_core_internal_output_type_dp2p0;
+					*OutputRate = dml2_core_internal_output_rate_dp_rate_uhbr20;
+				}
+			} else { // output is dp or edp
+				*OutBpp = 0;
+				if ((OutputLinkDPRate == dml2_dp_rate_na || OutputLinkDPRate == dml2_dp_rate_hbr) && PHYCLK >= 270) {
+					*OutBpp = TruncToValidBPP(&s->TruncToValidBPP_locals, (1 - Downspreading / 100) * 2700, OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output,
+						OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (unsigned int)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+					if (*OutBpp == 0 && PHYCLK < 540 && DSCEnable == dml2_dsc_enable_if_necessary && ForcedOutputLinkBPP == 0) {
+						*RequiresDSC = true;
+						LinkDSCEnable = true;
+						if (Output == dml2_dp) {
+							*RequiresFEC = true;
+						}
+						*OutBpp = TruncToValidBPP(&s->TruncToValidBPP_locals, (1 - Downspreading / 100) * 2700, OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output,
+							OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (unsigned int)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+					}
+					//OutputTypeAndRate = Output & " HBR";
+					*OutputType = (Output == dml2_dp) ? dml2_core_internal_output_type_dp : dml2_core_internal_output_type_edp;
+					*OutputRate = dml2_core_internal_output_rate_dp_rate_hbr;
+				}
+				if ((OutputLinkDPRate == dml2_dp_rate_na || OutputLinkDPRate == dml2_dp_rate_hbr2) && *OutBpp == 0 && PHYCLK >= 540) {
+					*OutBpp = TruncToValidBPP(&s->TruncToValidBPP_locals, (1 - Downspreading / 100) * 5400, OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output,
+						OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (unsigned int)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+
+					if (*OutBpp == 0 && PHYCLK < 810 && DSCEnable == dml2_dsc_enable_if_necessary && ForcedOutputLinkBPP == 0) {
+						*RequiresDSC = true;
+						LinkDSCEnable = true;
+						if (Output == dml2_dp) {
+							*RequiresFEC = true;
+						}
+						*OutBpp = TruncToValidBPP(&s->TruncToValidBPP_locals, (1 - Downspreading / 100) * 5400, OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output,
+							OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (unsigned int)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+					}
+					//OutputTypeAndRate = Output & " HBR2";
+					*OutputType = (Output == dml2_dp) ? dml2_core_internal_output_type_dp : dml2_core_internal_output_type_edp;
+					*OutputRate = dml2_core_internal_output_rate_dp_rate_hbr2;
+				}
+				if ((OutputLinkDPRate == dml2_dp_rate_na || OutputLinkDPRate == dml2_dp_rate_hbr3) && *OutBpp == 0 && PHYCLK >= 810) { // VBA_ERROR, vba code doesn't have hbr3 check
+					*OutBpp = TruncToValidBPP(&s->TruncToValidBPP_locals, (1 - Downspreading / 100) * 8100, OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output,
+						OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (unsigned int)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+
+					if (*OutBpp == 0 && DSCEnable == dml2_dsc_enable_if_necessary && ForcedOutputLinkBPP == 0) {
+						*RequiresDSC = true;
+						LinkDSCEnable = true;
+						if (Output == dml2_dp) {
+							*RequiresFEC = true;
+						}
+						*OutBpp = TruncToValidBPP(&s->TruncToValidBPP_locals, (1 - Downspreading / 100) * 8100, OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output,
+							OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (unsigned int)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, RequiredSlots);
+					}
+					//OutputTypeAndRate = Output & " HBR3";
+					*OutputType = (Output == dml2_dp) ? dml2_core_internal_output_type_dp : dml2_core_internal_output_type_edp;
+					*OutputRate = dml2_core_internal_output_rate_dp_rate_hbr3;
+				}
+			}
+		} else if (Output == dml2_hdmifrl) {
+			if (DSCEnable == dml2_dsc_enable) {
+				*RequiresDSC = true;
+				LinkDSCEnable = true;
+				*RequiresFEC = true;
+			} else {
+				*RequiresDSC = false;
+				LinkDSCEnable = false;
+				*RequiresFEC = false;
+			}
+			*OutBpp = 0;
+			if (PHYCLKD18 >= 3000.0 / 18) {
+				*OutBpp = TruncToValidBPP(&s->TruncToValidBPP_locals, 3000, 3, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output, OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (unsigned int)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, &dummy);
+				//OutputTypeAndRate = Output & "3x3";
+				*OutputType = dml2_core_internal_output_type_hdmifrl;
+				*OutputRate = dml2_core_internal_output_rate_hdmi_rate_3x3;
+			}
+			if (*OutBpp == 0 && PHYCLKD18 >= 6000.0 / 18) {
+				*OutBpp = TruncToValidBPP(&s->TruncToValidBPP_locals, 6000, 3, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output, OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (unsigned int)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, &dummy);
+				//OutputTypeAndRate = Output & "6x3";
+				*OutputType = dml2_core_internal_output_type_hdmifrl;
+				*OutputRate = dml2_core_internal_output_rate_hdmi_rate_6x3;
+			}
+			if (*OutBpp == 0 && PHYCLKD18 >= 6000.0 / 18) {
+				*OutBpp = TruncToValidBPP(&s->TruncToValidBPP_locals, 6000, 4, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output, OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (unsigned int)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, &dummy);
+				//OutputTypeAndRate = Output & "6x4";
+				*OutputType = dml2_core_internal_output_type_hdmifrl;
+				*OutputRate = dml2_core_internal_output_rate_hdmi_rate_6x4;
+			}
+			if (*OutBpp == 0 && PHYCLKD18 >= 8000.0 / 18) {
+				*OutBpp = TruncToValidBPP(&s->TruncToValidBPP_locals, 8000, 4, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output, OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (unsigned int)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, &dummy);
+				//OutputTypeAndRate = Output & "8x4";
+				*OutputType = dml2_core_internal_output_type_hdmifrl;
+				*OutputRate = dml2_core_internal_output_rate_hdmi_rate_8x4;
+			}
+			if (*OutBpp == 0 && PHYCLKD18 >= 10000.0 / 18) {
+				*OutBpp = TruncToValidBPP(&s->TruncToValidBPP_locals, 10000, 4, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output, OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (unsigned int)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, &dummy);
+				if (*OutBpp == 0 && DSCEnable == dml2_dsc_enable_if_necessary && ForcedOutputLinkBPP == 0 && PHYCLKD18 < 12000.0 / 18) {
+					*RequiresDSC = true;
+					LinkDSCEnable = true;
+					*RequiresFEC = true;
+					*OutBpp = TruncToValidBPP(&s->TruncToValidBPP_locals, 10000, 4, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output, OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (unsigned int)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, &dummy);
+				}
+				//OutputTypeAndRate = Output & "10x4";
+				*OutputType = dml2_core_internal_output_type_hdmifrl;
+				*OutputRate = dml2_core_internal_output_rate_hdmi_rate_10x4;
+			}
+			if (*OutBpp == 0 && PHYCLKD18 >= 12000.0 / 18) {
+				*OutBpp = TruncToValidBPP(&s->TruncToValidBPP_locals, 12000, 4, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output, OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (unsigned int)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, &dummy);
+				if (*OutBpp == 0 && DSCEnable == dml2_dsc_enable_if_necessary && ForcedOutputLinkBPP == 0) {
+					*RequiresDSC = true;
+					LinkDSCEnable = true;
+					*RequiresFEC = true;
+					*OutBpp = TruncToValidBPP(&s->TruncToValidBPP_locals, 12000, 4, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output, OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (unsigned int)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, &dummy);
+				}
+				//OutputTypeAndRate = Output & "12x4";
+				*OutputType = dml2_core_internal_output_type_hdmifrl;
+				*OutputRate = dml2_core_internal_output_rate_hdmi_rate_12x4;
+			}
+		}
+	}
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: RequiresDSC = %u\n", __func__, *RequiresDSC);
+	DML_LOG_VERBOSE("DML::%s: RequiresFEC = %u\n", __func__, *RequiresFEC);
+	DML_LOG_VERBOSE("DML::%s: OutBpp = %f\n", __func__, *OutBpp);
+#endif
+}
+
+static double CalculateWriteBackDISPCLK(
+	enum dml2_source_format_class WritebackPixelFormat,
+	double PixelClock,
+	double WritebackHRatio,
+	double WritebackVRatio,
+	unsigned int WritebackHTaps,
+	unsigned int WritebackVTaps,
+	unsigned int WritebackSourceWidth,
+	unsigned int WritebackDestinationWidth,
+	unsigned int HTotal,
+	unsigned int WritebackLineBufferSize)
+{
+	double DISPCLK_H, DISPCLK_V, DISPCLK_HB;
+
+	DISPCLK_H = PixelClock * math_ceil2((double)WritebackHTaps / 8.0, 1) / WritebackHRatio;
+	DISPCLK_V = PixelClock * (WritebackVTaps * math_ceil2((double)WritebackDestinationWidth / 6.0, 1) + 8.0) / (double)HTotal;
+	DISPCLK_HB = PixelClock * WritebackVTaps * (WritebackDestinationWidth * WritebackVTaps - WritebackLineBufferSize / 57.0) / 6.0 / (double)WritebackSourceWidth;
+	return math_max3(DISPCLK_H, DISPCLK_V, DISPCLK_HB);
+}
+
+static double RequiredDTBCLK(
+	bool DSCEnable,
+	double PixelClock,
+	enum dml2_output_format_class OutputFormat,
+	double OutputBpp,
+	unsigned int DSCSlices,
+	unsigned int HTotal,
+	unsigned int HActive,
+	unsigned int AudioRate,
+	unsigned int AudioLayout)
+{
+	if (DSCEnable != true) {
+		return math_max2(PixelClock / 4.0 * OutputBpp / 24.0, 25.0);
+	} else {
+		double PixelWordRate = PixelClock / (OutputFormat == dml2_444 ? 1 : 2);
+		double HCActive = math_ceil2(DSCSlices * math_ceil2(OutputBpp * math_ceil2(HActive / DSCSlices, 1) / 8.0, 1) / 3.0, 1);
+		double HCBlank = 64 + 32 * math_ceil2(AudioRate * (AudioLayout == 1 ? 1 : 0.25) * HTotal / (PixelClock * 1000), 1);
+		double AverageTribyteRate = PixelWordRate * (HCActive + HCBlank) / HTotal;
+		double HActiveTribyteRate = PixelWordRate * HCActive / HActive;
+		return math_max4(PixelWordRate / 4.0, AverageTribyteRate / 4.0, HActiveTribyteRate / 4.0, 25.0) * 1.002;
+	}
+}
+
+static unsigned int DSCDelayRequirement(
+	bool DSCEnabled,
+	enum dml2_odm_mode ODMMode,
+	unsigned int DSCInputBitPerComponent,
+	double OutputBpp,
+	unsigned int HActive,
+	unsigned int HTotal,
+	unsigned int NumberOfDSCSlices,
+	enum dml2_output_format_class OutputFormat,
+	enum dml2_output_encoder_class Output,
+	double PixelClock,
+	double PixelClockBackEnd)
+{
+	unsigned int DSCDelayRequirement_val = 0;
+	unsigned int NumberOfDSCSlicesFactor = 1;
+
+	if (DSCEnabled == true && OutputBpp != 0) {
+
+		if (ODMMode == dml2_odm_mode_combine_4to1)
+			NumberOfDSCSlicesFactor = 4;
+		else if (ODMMode == dml2_odm_mode_combine_3to1)
+			NumberOfDSCSlicesFactor = 3;
+		else if (ODMMode == dml2_odm_mode_combine_2to1)
+			NumberOfDSCSlicesFactor = 2;
+
+		DSCDelayRequirement_val = NumberOfDSCSlicesFactor * (dscceComputeDelay(DSCInputBitPerComponent, OutputBpp, (unsigned int)(math_ceil2((double)HActive / (double)NumberOfDSCSlices, 1.0)),
+			(NumberOfDSCSlices / NumberOfDSCSlicesFactor), OutputFormat, Output) + dscComputeDelay(OutputFormat, Output));
+
+		DSCDelayRequirement_val = (unsigned int)(DSCDelayRequirement_val + (HTotal - HActive) * math_ceil2((double)DSCDelayRequirement_val / (double)HActive, 1.0));
+		DSCDelayRequirement_val = (unsigned int)(DSCDelayRequirement_val * PixelClock / PixelClockBackEnd);
+
+	} else {
+		DSCDelayRequirement_val = 0;
+	}
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: DSCEnabled= %u\n", __func__, DSCEnabled);
+	DML_LOG_VERBOSE("DML::%s: ODMMode = %u\n", __func__, ODMMode);
+	DML_LOG_VERBOSE("DML::%s: OutputBpp = %f\n", __func__, OutputBpp);
+	DML_LOG_VERBOSE("DML::%s: HActive = %u\n", __func__, HActive);
+	DML_LOG_VERBOSE("DML::%s: HTotal = %u\n", __func__, HTotal);
+	DML_LOG_VERBOSE("DML::%s: PixelClock = %f\n", __func__, PixelClock);
+	DML_LOG_VERBOSE("DML::%s: PixelClockBackEnd = %f\n", __func__, PixelClockBackEnd);
+	DML_LOG_VERBOSE("DML::%s: OutputFormat = %u\n", __func__, OutputFormat);
+	DML_LOG_VERBOSE("DML::%s: DSCInputBitPerComponent = %u\n", __func__, DSCInputBitPerComponent);
+	DML_LOG_VERBOSE("DML::%s: NumberOfDSCSlices = %u\n", __func__, NumberOfDSCSlices);
+	DML_LOG_VERBOSE("DML::%s: DSCDelayRequirement_val = %u\n", __func__, DSCDelayRequirement_val);
+#endif
+
+	return DSCDelayRequirement_val;
+}
+
+static void CalculateSurfaceSizeInMall(
+	const struct dml2_display_cfg *display_cfg,
+	unsigned int NumberOfActiveSurfaces,
+	unsigned int MALLAllocatedForDCN,
+	unsigned int BytesPerPixelY[],
+	unsigned int BytesPerPixelC[],
+	unsigned int Read256BytesBlockWidthY[],
+	unsigned int Read256BytesBlockWidthC[],
+	unsigned int Read256BytesBlockHeightY[],
+	unsigned int Read256BytesBlockHeightC[],
+	unsigned int ReadBlockWidthY[],
+	unsigned int ReadBlockWidthC[],
+	unsigned int ReadBlockHeightY[],
+	unsigned int ReadBlockHeightC[],
+
+	// Output
+	unsigned int SurfaceSizeInMALL[],
+	bool *ExceededMALLSize)
+{
+	unsigned int TotalSurfaceSizeInMALLForSS = 0;
+	unsigned int TotalSurfaceSizeInMALLForSubVP = 0;
+	unsigned int MALLAllocatedForDCNInBytes = MALLAllocatedForDCN * 1024 * 1024;
+
+	for (unsigned int k = 0; k < NumberOfActiveSurfaces; ++k) {
+		const struct dml2_composition_cfg *composition = &display_cfg->plane_descriptors[k].composition;
+		const struct dml2_surface_cfg *surface = &display_cfg->plane_descriptors[k].surface;
+
+		if (composition->viewport.stationary) {
+			SurfaceSizeInMALL[k] = (unsigned int)(math_min2(math_ceil2((double)surface->plane0.width, ReadBlockWidthY[k]),
+				math_floor2(composition->viewport.plane0.x_start + composition->viewport.plane0.width + ReadBlockWidthY[k] - 1, ReadBlockWidthY[k]) -
+				math_floor2((double)composition->viewport.plane0.x_start, ReadBlockWidthY[k])) *
+				math_min2(math_ceil2((double)surface->plane0.height, ReadBlockHeightY[k]),
+					math_floor2((double)composition->viewport.plane0.y_start + composition->viewport.plane0.height + ReadBlockHeightY[k] - 1, ReadBlockHeightY[k]) -
+					math_floor2((double)composition->viewport.plane0.y_start, ReadBlockHeightY[k])) * BytesPerPixelY[k]);
+
+			if (ReadBlockWidthC[k] > 0) {
+				SurfaceSizeInMALL[k] = (unsigned int)(SurfaceSizeInMALL[k] +
+					math_min2(math_ceil2((double)surface->plane1.width, ReadBlockWidthC[k]),
+						math_floor2((double)composition->viewport.plane1.y_start + composition->viewport.plane1.width + ReadBlockWidthC[k] - 1, ReadBlockWidthC[k]) -
+						math_floor2((double)composition->viewport.plane1.y_start, ReadBlockWidthC[k])) *
+					math_min2(math_ceil2((double)surface->plane1.height, ReadBlockHeightC[k]),
+						math_floor2((double)composition->viewport.plane1.y_start + composition->viewport.plane1.height + ReadBlockHeightC[k] - 1, ReadBlockHeightC[k]) -
+						math_floor2(composition->viewport.plane1.y_start, ReadBlockHeightC[k])) * BytesPerPixelC[k]);
+			}
+		} else {
+			SurfaceSizeInMALL[k] = (unsigned int)(math_ceil2(math_min2(surface->plane0.width, composition->viewport.plane0.width + ReadBlockWidthY[k] - 1), ReadBlockWidthY[k]) *
+				math_ceil2(math_min2(surface->plane0.height, composition->viewport.plane0.height + ReadBlockHeightY[k] - 1), ReadBlockHeightY[k]) * BytesPerPixelY[k]);
+			if (ReadBlockWidthC[k] > 0) {
+				SurfaceSizeInMALL[k] = (unsigned int)(SurfaceSizeInMALL[k] +
+					math_ceil2(math_min2(surface->plane1.width, composition->viewport.plane1.width + ReadBlockWidthC[k] - 1), ReadBlockWidthC[k]) *
+					math_ceil2(math_min2(surface->plane1.height, composition->viewport.plane1.height + ReadBlockHeightC[k] - 1), ReadBlockHeightC[k]) * BytesPerPixelC[k]);
+			}
+		}
+	}
+
+	for (unsigned int k = 0; k < NumberOfActiveSurfaces; ++k) {
+		/* SS and Subvp counted separate as they are never used at the same time */
+		if (dml_is_phantom_pipe(&display_cfg->plane_descriptors[k]))
+			TotalSurfaceSizeInMALLForSubVP += SurfaceSizeInMALL[k];
+		else if (display_cfg->plane_descriptors[k].overrides.refresh_from_mall == dml2_refresh_from_mall_mode_override_force_enable)
+			TotalSurfaceSizeInMALLForSS += SurfaceSizeInMALL[k];
+	}
+
+	*ExceededMALLSize = (TotalSurfaceSizeInMALLForSS > MALLAllocatedForDCNInBytes) ||
+		(TotalSurfaceSizeInMALLForSubVP > MALLAllocatedForDCNInBytes);
+
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: MALLAllocatedForDCN = %u\n", __func__, MALLAllocatedForDCN * 1024 * 1024);
+	DML_LOG_VERBOSE("DML::%s: TotalSurfaceSizeInMALLForSubVP = %u\n", __func__, TotalSurfaceSizeInMALLForSubVP);
+	DML_LOG_VERBOSE("DML::%s: TotalSurfaceSizeInMALLForSS = %u\n", __func__, TotalSurfaceSizeInMALLForSS);
+	DML_LOG_VERBOSE("DML::%s: ExceededMALLSize = %u\n", __func__, *ExceededMALLSize);
+#endif
+}
+
+static void calculate_tdlut_setting(
+		struct dml2_core_internal_scratch *scratch,
+		struct dml2_core_calcs_calculate_tdlut_setting_params *p)
+{
+	// locals
+	unsigned int tdlut_bpe = 8;
+	unsigned int tdlut_width;
+	unsigned int tdlut_pitch_bytes;
+	unsigned int tdlut_footprint_bytes;
+	unsigned int vmpg_bytes;
+	unsigned int tdlut_vmpg_per_frame;
+	unsigned int tdlut_pte_req_per_frame;
+	unsigned int tdlut_bytes_per_line;
+	double tdlut_drain_rate;
+	unsigned int tdlut_mpc_width;
+	unsigned int tdlut_bytes_per_group_simple;
+
+	if (!p->setup_for_tdlut) {
+		*p->tdlut_groups_per_2row_ub = 0;
+		*p->tdlut_opt_time = 0;
+		*p->tdlut_drain_time = 0;
+		*p->tdlut_bytes_to_deliver = 0;
+		*p->tdlut_bytes_per_group = 0;
+		*p->tdlut_pte_bytes_per_frame = 0;
+		*p->tdlut_bytes_per_frame = 0;
+		return;
+	}
+
+	if (p->tdlut_mpc_width_flag) {
+		tdlut_mpc_width = 33;
+		tdlut_bytes_per_group_simple = 39*256;
+	} else {
+		tdlut_mpc_width = 17;
+		tdlut_bytes_per_group_simple = 10*256;
+	}
+
+	vmpg_bytes = p->gpuvm_page_size_kbytes * 1024;
+
+	if (p->tdlut_addressing_mode == dml2_tdlut_simple_linear) {
+		if (p->tdlut_width_mode == dml2_tdlut_width_17_cube)
+			tdlut_width = 4916;
+		else
+			tdlut_width = 35940;
+	} else {
+		if (p->tdlut_width_mode == dml2_tdlut_width_17_cube)
+			tdlut_width = 17;
+		else // dml2_tdlut_width_33_cube
+			tdlut_width = 33;
+	}
+
+	if (p->is_gfx11)
+		tdlut_pitch_bytes = (unsigned int)math_ceil2(tdlut_width * tdlut_bpe, 256); //256B alignment
+	else
+		tdlut_pitch_bytes = (unsigned int)math_ceil2(tdlut_width * tdlut_bpe, 128); //128B alignment
+
+	if (p->tdlut_addressing_mode == dml2_tdlut_sw_linear)
+		tdlut_footprint_bytes = tdlut_pitch_bytes * tdlut_width * tdlut_width;
+	else
+		tdlut_footprint_bytes = tdlut_pitch_bytes;
+
+	if (!p->gpuvm_enable) {
+		tdlut_vmpg_per_frame = 0;
+		tdlut_pte_req_per_frame = 0;
+	} else {
+		tdlut_vmpg_per_frame = (unsigned int)math_ceil2(tdlut_footprint_bytes - 1, vmpg_bytes) / vmpg_bytes + 1;
+		tdlut_pte_req_per_frame = (unsigned int)math_ceil2(tdlut_vmpg_per_frame - 1, 8) / 8 + 1;
+	}
+	tdlut_bytes_per_line = (unsigned int)math_ceil2(tdlut_width * tdlut_bpe, 64); //64b request
+	*p->tdlut_pte_bytes_per_frame = tdlut_pte_req_per_frame * 64;
+
+	if (p->tdlut_addressing_mode == dml2_tdlut_sw_linear) {
+		//the tdlut_width is either 17 or 33 but the 33x33x33 is subsampled every other line/slice
+		*p->tdlut_bytes_per_frame = tdlut_bytes_per_line * tdlut_mpc_width * tdlut_mpc_width;
+		*p->tdlut_bytes_per_group = tdlut_bytes_per_line * tdlut_mpc_width;
+		//the delivery cycles is DispClk cycles per line * number of lines * number of slices
+		//tdlut_delivery_cycles = (unsigned int)math_ceil2(tdlut_mpc_width/2.0, 1) * tdlut_mpc_width * tdlut_mpc_width;
+		tdlut_drain_rate = tdlut_bytes_per_line * p->dispclk_mhz / math_ceil2(tdlut_mpc_width/2.0, 1);
+	} else {
+		//tdlut_addressing_mode = tdlut_simple_linear, 3dlut width should be 4*1229=4916 elements
+		*p->tdlut_bytes_per_frame = (unsigned int)math_ceil2(tdlut_width * tdlut_bpe, 256);
+		*p->tdlut_bytes_per_group = tdlut_bytes_per_group_simple;
+		//tdlut_delivery_cycles = (unsigned int)math_ceil2(tdlut_width/2.0, 1);
+		tdlut_drain_rate = 2 * tdlut_bpe * p->dispclk_mhz;
+	}
+
+	//the tdlut is fetched during the 2 row times of prefetch.
+	if (p->setup_for_tdlut) {
+		*p->tdlut_groups_per_2row_ub = (unsigned int)math_ceil2((double) *p->tdlut_bytes_per_frame / *p->tdlut_bytes_per_group, 1);
+		if (*p->tdlut_bytes_per_frame > p->cursor_buffer_size * 1024)
+			*p->tdlut_opt_time = (*p->tdlut_bytes_per_frame - p->cursor_buffer_size * 1024) / tdlut_drain_rate;
+		else
+			*p->tdlut_opt_time = 0;
+		*p->tdlut_drain_time = p->cursor_buffer_size * 1024 / tdlut_drain_rate;
+		*p->tdlut_bytes_to_deliver = (unsigned int) (p->cursor_buffer_size * 1024.0);
+	}
+
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: gpuvm_enable = %d\n", __func__, p->gpuvm_enable);
+	DML_LOG_VERBOSE("DML::%s: vmpg_bytes = %d\n", __func__, vmpg_bytes);
+	DML_LOG_VERBOSE("DML::%s: tdlut_vmpg_per_frame = %d\n", __func__, tdlut_vmpg_per_frame);
+	DML_LOG_VERBOSE("DML::%s: tdlut_pte_req_per_frame = %d\n", __func__, tdlut_pte_req_per_frame);
+
+	DML_LOG_VERBOSE("DML::%s: dispclk_mhz = %f\n", __func__, p->dispclk_mhz);
+	DML_LOG_VERBOSE("DML::%s: tdlut_width = %u\n", __func__, tdlut_width);
+	DML_LOG_VERBOSE("DML::%s: tdlut_addressing_mode = %s\n", __func__, (p->tdlut_addressing_mode == dml2_tdlut_sw_linear) ? "sw_linear" : "simple_linear");
+	DML_LOG_VERBOSE("DML::%s: tdlut_pitch_bytes = %u\n", __func__, tdlut_pitch_bytes);
+	DML_LOG_VERBOSE("DML::%s: tdlut_footprint_bytes = %u\n", __func__, tdlut_footprint_bytes);
+	DML_LOG_VERBOSE("DML::%s: tdlut_bytes_per_frame = %u\n", __func__, *p->tdlut_bytes_per_frame);
+	DML_LOG_VERBOSE("DML::%s: tdlut_bytes_per_line = %u\n", __func__, tdlut_bytes_per_line);
+	DML_LOG_VERBOSE("DML::%s: tdlut_bytes_per_group = %u\n", __func__, *p->tdlut_bytes_per_group);
+	DML_LOG_VERBOSE("DML::%s: tdlut_drain_rate = %f\n", __func__, tdlut_drain_rate);
+	DML_LOG_VERBOSE("DML::%s: tdlut_delivery_cycles = %u\n", __func__, p->tdlut_addressing_mode == dml2_tdlut_sw_linear ? (unsigned int)math_ceil2(tdlut_mpc_width/2.0, 1) * tdlut_mpc_width * tdlut_mpc_width : (unsigned int)math_ceil2(tdlut_width/2.0, 1));
+	DML_LOG_VERBOSE("DML::%s: tdlut_opt_time = %f\n", __func__, *p->tdlut_opt_time);
+	DML_LOG_VERBOSE("DML::%s: tdlut_drain_time = %f\n", __func__, *p->tdlut_drain_time);
+	DML_LOG_VERBOSE("DML::%s: tdlut_bytes_to_deliver = %d\n", __func__, *p->tdlut_bytes_to_deliver);
+	DML_LOG_VERBOSE("DML::%s: tdlut_groups_per_2row_ub = %d\n", __func__, *p->tdlut_groups_per_2row_ub);
+#endif
+}
+
+static void CalculateTarb(
+	const struct dml2_display_cfg *display_cfg,
+	unsigned int PixelChunkSizeInKByte,
+	unsigned int NumberOfActiveSurfaces,
+	unsigned int NumberOfDPP[],
+	unsigned int dpte_group_bytes[],
+	unsigned int tdlut_bytes_per_group[],
+	double HostVMInefficiencyFactor,
+	double HostVMInefficiencyFactorPrefetch,
+	unsigned int HostVMMinPageSize,
+	double ReturnBW,
+	unsigned int MetaChunkSize,
+
+	// output
+	double *Tarb,
+	double *Tarb_prefetch)
+{
+	double extra_bytes = 0;
+	double extra_bytes_prefetch = 0;
+	double HostVMDynamicLevels = CalculateHostVMDynamicLevels(display_cfg->gpuvm_enable, display_cfg->hostvm_enable, HostVMMinPageSize, display_cfg->hostvm_max_non_cached_page_table_levels);
+
+	for (unsigned int k = 0; k < NumberOfActiveSurfaces; ++k) {
+		extra_bytes = extra_bytes + (NumberOfDPP[k] * PixelChunkSizeInKByte * 1024);
+
+		if (display_cfg->plane_descriptors[k].surface.dcc.enable)
+			extra_bytes = extra_bytes + (MetaChunkSize * 1024);
+
+		if (display_cfg->plane_descriptors[k].tdlut.setup_for_tdlut)
+			extra_bytes = extra_bytes + tdlut_bytes_per_group[k];
+	}
+
+	extra_bytes_prefetch = extra_bytes;
+
+	for (unsigned int k = 0; k < NumberOfActiveSurfaces; ++k) {
+		if (display_cfg->gpuvm_enable == true) {
+			extra_bytes = extra_bytes + NumberOfDPP[k] * dpte_group_bytes[k] * (1 + 8 * HostVMDynamicLevels) * HostVMInefficiencyFactor;
+			extra_bytes_prefetch = extra_bytes_prefetch + NumberOfDPP[k] * dpte_group_bytes[k] * (1 + 8 * HostVMDynamicLevels) * HostVMInefficiencyFactorPrefetch;
+		}
+	}
+	*Tarb = extra_bytes / ReturnBW;
+	*Tarb_prefetch = extra_bytes_prefetch / ReturnBW;
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: PixelChunkSizeInKByte = %d\n", __func__, PixelChunkSizeInKByte);
+	DML_LOG_VERBOSE("DML::%s: MetaChunkSize = %d\n", __func__, MetaChunkSize);
+	DML_LOG_VERBOSE("DML::%s: extra_bytes = %f\n", __func__, extra_bytes);
+	DML_LOG_VERBOSE("DML::%s: extra_bytes_prefetch = %f\n", __func__, extra_bytes_prefetch);
+#endif
+}
+
+static double CalculateTWait(
+	long reserved_vblank_time_ns,
+	double UrgentLatency,
+	double Ttrip,
+	double g6_temp_read_blackout_us)
+{
+	double TWait;
+	double t_urg_trip = math_max2(UrgentLatency, Ttrip);
+	TWait = math_max2(reserved_vblank_time_ns/1000.0, g6_temp_read_blackout_us) + t_urg_trip;
+
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: reserved_vblank_time_ns = %ld\n", __func__, reserved_vblank_time_ns);
+	DML_LOG_VERBOSE("DML::%s: UrgentLatency = %f\n", __func__, UrgentLatency);
+	DML_LOG_VERBOSE("DML::%s: Ttrip = %f\n", __func__, Ttrip);
+	DML_LOG_VERBOSE("DML::%s: TWait = %f\n", __func__, TWait);
+#endif
+	return TWait;
+}
+
+
+static void CalculateVUpdateAndDynamicMetadataParameters(
+	unsigned int MaxInterDCNTileRepeaters,
+	double Dppclk,
+	double Dispclk,
+	double DCFClkDeepSleep,
+	double PixelClock,
+	unsigned int HTotal,
+	unsigned int VBlank,
+	unsigned int DynamicMetadataTransmittedBytes,
+	unsigned int DynamicMetadataLinesBeforeActiveRequired,
+	unsigned int InterlaceEnable,
+	bool ProgressiveToInterlaceUnitInOPP,
+
+	// Output
+	double *TSetup,
+	double *Tdmbf,
+	double *Tdmec,
+	double *Tdmsks,
+	unsigned int *VUpdateOffsetPix,
+	unsigned int *VUpdateWidthPix,
+	unsigned int *VReadyOffsetPix)
+{
+	double TotalRepeaterDelayTime;
+	TotalRepeaterDelayTime = MaxInterDCNTileRepeaters * (2 / Dppclk + 3 / Dispclk);
+	*VUpdateWidthPix = (unsigned int)(math_ceil2((14.0 / DCFClkDeepSleep + 12.0 / Dppclk + TotalRepeaterDelayTime) * PixelClock, 1.0));
+	*VReadyOffsetPix = (unsigned int)(math_ceil2(math_max2(150.0 / Dppclk, TotalRepeaterDelayTime + 20.0 / DCFClkDeepSleep + 10.0 / Dppclk) * PixelClock, 1.0));
+	*VUpdateOffsetPix = (unsigned int)(math_ceil2(HTotal / 4.0, 1.0));
+	*TSetup = (*VUpdateOffsetPix + *VUpdateWidthPix + *VReadyOffsetPix) / PixelClock;
+	*Tdmbf = DynamicMetadataTransmittedBytes / 4.0 / Dispclk;
+	*Tdmec = HTotal / PixelClock;
+
+	if (DynamicMetadataLinesBeforeActiveRequired == 0) {
+		*Tdmsks = VBlank * HTotal / PixelClock / 2.0;
+	} else {
+		*Tdmsks = DynamicMetadataLinesBeforeActiveRequired * HTotal / PixelClock;
+	}
+	if (InterlaceEnable == 1 && ProgressiveToInterlaceUnitInOPP == false) {
+		*Tdmsks = *Tdmsks / 2;
+	}
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: DynamicMetadataLinesBeforeActiveRequired = %u\n", __func__, DynamicMetadataLinesBeforeActiveRequired);
+	DML_LOG_VERBOSE("DML::%s: VBlank = %u\n", __func__, VBlank);
+	DML_LOG_VERBOSE("DML::%s: HTotal = %u\n", __func__, HTotal);
+	DML_LOG_VERBOSE("DML::%s: PixelClock = %f\n", __func__, PixelClock);
+	DML_LOG_VERBOSE("DML::%s: Dppclk = %f\n", __func__, Dppclk);
+	DML_LOG_VERBOSE("DML::%s: DCFClkDeepSleep = %f\n", __func__, DCFClkDeepSleep);
+	DML_LOG_VERBOSE("DML::%s: MaxInterDCNTileRepeaters = %u\n", __func__, MaxInterDCNTileRepeaters);
+	DML_LOG_VERBOSE("DML::%s: TotalRepeaterDelayTime = %f\n", __func__, TotalRepeaterDelayTime);
+
+	DML_LOG_VERBOSE("DML::%s: VUpdateWidthPix = %u\n", __func__, *VUpdateWidthPix);
+	DML_LOG_VERBOSE("DML::%s: VReadyOffsetPix = %u\n", __func__, *VReadyOffsetPix);
+	DML_LOG_VERBOSE("DML::%s: VUpdateOffsetPix = %u\n", __func__, *VUpdateOffsetPix);
+
+	DML_LOG_VERBOSE("DML::%s: Tdmsks = %f\n", __func__, *Tdmsks);
+#endif
+}
+
+static double get_urgent_bandwidth_required(
+	struct dml2_core_shared_get_urgent_bandwidth_required_locals *l,
+	const struct dml2_display_cfg *display_cfg,
+	enum dml2_core_internal_soc_state_type state_type,
+	enum dml2_core_internal_bw_type bw_type,
+	bool inc_flip_bw, // including flip bw
+	bool use_qual_row_bw,
+	unsigned int NumberOfActiveSurfaces,
+	unsigned int NumberOfDPP[],
+	double dcc_dram_bw_nom_overhead_factor_p0[],
+	double dcc_dram_bw_nom_overhead_factor_p1[],
+	double dcc_dram_bw_pref_overhead_factor_p0[],
+	double dcc_dram_bw_pref_overhead_factor_p1[],
+	double mall_prefetch_sdp_overhead_factor[],
+	double mall_prefetch_dram_overhead_factor[],
+	double ReadBandwidthLuma[],
+	double ReadBandwidthChroma[],
+	double PrefetchBandwidthLuma[],
+	double PrefetchBandwidthChroma[],
+	double PrefetchBandwidthMax[],
+	double excess_vactive_fill_bw_l[],
+	double excess_vactive_fill_bw_c[],
+	double cursor_bw[],
+	double dpte_row_bw[],
+	double meta_row_bw[],
+	double prefetch_cursor_bw[],
+	double prefetch_vmrow_bw[],
+	double flip_bw[],
+	double UrgentBurstFactorLuma[],
+	double UrgentBurstFactorChroma[],
+	double UrgentBurstFactorCursor[],
+	double UrgentBurstFactorLumaPre[],
+	double UrgentBurstFactorChromaPre[],
+	double UrgentBurstFactorCursorPre[],
+	/* outputs */
+	double surface_required_bw[],
+	double surface_peak_required_bw[])
+{
+	// set inc_flip_bw = 0 for total_dchub_urgent_read_bw_noflip calculation, 1 for total_dchub_urgent_read_bw as described in the MAS
+	// set use_qual_row_bw = 1 to calculate using qualified row bandwidth, used for total_flip_bw calculation
+
+	memset(l, 0, sizeof(struct dml2_core_shared_get_urgent_bandwidth_required_locals));
+
+	for (unsigned int k = 0; k < NumberOfActiveSurfaces; ++k) {
+		l->mall_svp_prefetch_factor = (state_type == dml2_core_internal_soc_state_svp_prefetch) ? (bw_type == dml2_core_internal_bw_dram ? mall_prefetch_dram_overhead_factor[k] : mall_prefetch_sdp_overhead_factor[k]) : 1.0;
+		l->tmp_nom_adj_factor_p0 = (bw_type == dml2_core_internal_bw_dram ? dcc_dram_bw_nom_overhead_factor_p0[k] : 1.0) * l->mall_svp_prefetch_factor;
+		l->tmp_nom_adj_factor_p1 = (bw_type == dml2_core_internal_bw_dram ? dcc_dram_bw_nom_overhead_factor_p1[k] : 1.0) * l->mall_svp_prefetch_factor;
+		l->tmp_pref_adj_factor_p0 = (bw_type == dml2_core_internal_bw_dram ? dcc_dram_bw_pref_overhead_factor_p0[k] : 1.0) * l->mall_svp_prefetch_factor;
+		l->tmp_pref_adj_factor_p1 = (bw_type == dml2_core_internal_bw_dram ? dcc_dram_bw_pref_overhead_factor_p1[k] : 1.0) * l->mall_svp_prefetch_factor;
+
+		l->adj_factor_p0 = UrgentBurstFactorLuma[k] * l->tmp_nom_adj_factor_p0;
+		l->adj_factor_p1 = UrgentBurstFactorChroma[k] * l->tmp_nom_adj_factor_p1;
+		l->adj_factor_cur = UrgentBurstFactorCursor[k];
+		l->adj_factor_p0_pre = UrgentBurstFactorLumaPre[k] * l->tmp_pref_adj_factor_p0;
+		l->adj_factor_p1_pre = UrgentBurstFactorChromaPre[k] * l->tmp_pref_adj_factor_p1;
+		l->adj_factor_cur_pre = UrgentBurstFactorCursorPre[k];
+
+		bool is_phantom = dml_is_phantom_pipe(&display_cfg->plane_descriptors[k]);
+		bool exclude_this_plane = false;
+
+		// Exclude phantom pipe in bw calculation for non svp prefetch state
+		if (state_type != dml2_core_internal_soc_state_svp_prefetch && is_phantom)
+			exclude_this_plane = true;
+
+		// The qualified row bandwidth, qual_row_bw, accounts for the regular non-flip row bandwidth when there is no possible immediate flip or HostVM invalidation flip.
+		// The qual_row_bw is zero if HostVM is possible and only non-zero and equal to row_bw(i) if immediate flip is not allowed for that pipe.
+		if (use_qual_row_bw) {
+			if (display_cfg->hostvm_enable)
+				l->per_plane_flip_bw[k] = 0; // qual_row_bw
+			else if (!display_cfg->plane_descriptors[k].immediate_flip)
+				l->per_plane_flip_bw[k] = NumberOfDPP[k] * (dpte_row_bw[k] + meta_row_bw[k]);
+		} else {
+			// the final_flip_bw includes the regular row_bw when immediate flip is disallowed (and no HostVM)
+			if ((!display_cfg->plane_descriptors[k].immediate_flip && !display_cfg->hostvm_enable) || !inc_flip_bw)
+				l->per_plane_flip_bw[k] = NumberOfDPP[k] * (dpte_row_bw[k] + meta_row_bw[k]);
+			else
+				l->per_plane_flip_bw[k] = NumberOfDPP[k] * flip_bw[k];
+		}
+
+		if (!exclude_this_plane) {
+			l->vm_row_bw = NumberOfDPP[k] * prefetch_vmrow_bw[k];
+			l->flip_and_active_bw = l->per_plane_flip_bw[k] + ReadBandwidthLuma[k] * l->adj_factor_p0 + ReadBandwidthChroma[k] * l->adj_factor_p1 + cursor_bw[k] * l->adj_factor_cur;
+			l->flip_and_prefetch_bw = l->per_plane_flip_bw[k] + NumberOfDPP[k] * (PrefetchBandwidthLuma[k] * l->adj_factor_p0_pre + PrefetchBandwidthChroma[k] * l->adj_factor_p1_pre) + prefetch_cursor_bw[k] * l->adj_factor_cur_pre;
+			l->flip_and_prefetch_bw_max = l->per_plane_flip_bw[k] + NumberOfDPP[k] * (PrefetchBandwidthMax[k] * l->adj_factor_p0_pre + PrefetchBandwidthChroma[k] * l->adj_factor_p1_pre) + prefetch_cursor_bw[k] * l->adj_factor_cur_pre;
+			l->active_and_excess_bw = (ReadBandwidthLuma[k] + excess_vactive_fill_bw_l[k]) * l->tmp_nom_adj_factor_p0 + (ReadBandwidthChroma[k] + excess_vactive_fill_bw_c[k]) * l->tmp_nom_adj_factor_p1 + dpte_row_bw[k] + meta_row_bw[k];
+			surface_required_bw[k] = math_max5(l->vm_row_bw, l->flip_and_active_bw, l->flip_and_prefetch_bw, l->active_and_excess_bw, l->flip_and_prefetch_bw_max);
+
+			/* export peak required bandwidth for the surface */
+			surface_peak_required_bw[k] = math_max2(surface_required_bw[k], surface_peak_required_bw[k]);
+
+#ifdef __DML_VBA_DEBUG__
+			DML_LOG_VERBOSE("DML::%s: k=%d, max1: vm_row_bw=%f\n", __func__, k, l->vm_row_bw);
+			DML_LOG_VERBOSE("DML::%s: k=%d, max2: flip_and_active_bw=%f\n", __func__, k, l->flip_and_active_bw);
+			DML_LOG_VERBOSE("DML::%s: k=%d, max3: flip_and_prefetch_bw=%f\n", __func__, k, l->flip_and_prefetch_bw);
+			DML_LOG_VERBOSE("DML::%s: k=%d, max4: active_and_excess_bw=%f\n", __func__, k, l->active_and_excess_bw);
+			DML_LOG_VERBOSE("DML::%s: k=%d, surface_required_bw=%f\n", __func__, k, surface_required_bw[k]);
+			DML_LOG_VERBOSE("DML::%s: k=%d, surface_peak_required_bw=%f\n", __func__, k, surface_peak_required_bw[k]);
+#endif
+		} else {
+			surface_required_bw[k] = 0.0;
+		}
+
+		l->required_bandwidth_mbps += surface_required_bw[k];
+
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: k=%d, NumberOfDPP=%d\n", __func__, k, NumberOfDPP[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%d, use_qual_row_bw=%d\n", __func__, k, use_qual_row_bw);
+		DML_LOG_VERBOSE("DML::%s: k=%d, immediate_flip=%d\n", __func__, k, display_cfg->plane_descriptors[k].immediate_flip);
+		DML_LOG_VERBOSE("DML::%s: k=%d, mall_svp_prefetch_factor=%f\n", __func__, k, l->mall_svp_prefetch_factor);
+		DML_LOG_VERBOSE("DML::%s: k=%d, adj_factor_p0=%f\n", __func__, k, l->adj_factor_p0);
+		DML_LOG_VERBOSE("DML::%s: k=%d, adj_factor_p1=%f\n", __func__, k, l->adj_factor_p1);
+		DML_LOG_VERBOSE("DML::%s: k=%d, adj_factor_cur=%f\n", __func__, k, l->adj_factor_cur);
+
+		DML_LOG_VERBOSE("DML::%s: k=%d, adj_factor_p0_pre=%f\n", __func__, k, l->adj_factor_p0_pre);
+		DML_LOG_VERBOSE("DML::%s: k=%d, adj_factor_p1_pre=%f\n", __func__, k, l->adj_factor_p1_pre);
+		DML_LOG_VERBOSE("DML::%s: k=%d, adj_factor_cur_pre=%f\n", __func__, k, l->adj_factor_cur_pre);
+
+		DML_LOG_VERBOSE("DML::%s: k=%d, per_plane_flip_bw=%f\n", __func__, k, l->per_plane_flip_bw[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%d, prefetch_vmrow_bw=%f\n", __func__, k, prefetch_vmrow_bw[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%d, ReadBandwidthLuma=%f\n", __func__, k, ReadBandwidthLuma[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%d, ReadBandwidthChroma=%f\n", __func__, k, ReadBandwidthChroma[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%d, excess_vactive_fill_bw_l=%f\n", __func__, k, excess_vactive_fill_bw_l[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%d, excess_vactive_fill_bw_c=%f\n", __func__, k, excess_vactive_fill_bw_c[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%d, cursor_bw=%f\n", __func__, k, cursor_bw[k]);
+
+		DML_LOG_VERBOSE("DML::%s: k=%d, meta_row_bw=%f\n", __func__, k, meta_row_bw[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%d, dpte_row_bw=%f\n", __func__, k, dpte_row_bw[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%d, PrefetchBandwidthLuma=%f\n", __func__, k, PrefetchBandwidthLuma[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%d, PrefetchBandwidthChroma=%f\n", __func__, k, PrefetchBandwidthChroma[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%d, prefetch_cursor_bw=%f\n", __func__, k, prefetch_cursor_bw[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%d, required_bandwidth_mbps=%f (total), inc_flip_bw=%d, is_phantom=%d exclude_this_plane=%d\n", __func__, k, l->required_bandwidth_mbps, inc_flip_bw, is_phantom, exclude_this_plane);
+		DML_LOG_VERBOSE("DML::%s: k=%d, required_bandwidth_mbps=%f (total), soc_state=%s, inc_flip_bw=%d, is_phantom=%d exclude_this_plane=%d\n", __func__, k, l->required_bandwidth_mbps, dml2_core_internal_soc_state_type_str(state_type), inc_flip_bw, is_phantom, exclude_this_plane);
+		DML_LOG_VERBOSE("DML::%s: k=%d, required_bandwidth_mbps=%f (total), inc_flip_bw=%d, is_phantom=%d exclude_this_plane=%d\n", __func__, k, l->required_bandwidth_mbps, inc_flip_bw, is_phantom, exclude_this_plane);
+#endif
+	}
+
+	return l->required_bandwidth_mbps;
+}
+
+static void CalculateExtraLatency(
+	const struct dml2_display_cfg *display_cfg,
+	unsigned int ROBBufferSizeInKByte,
+	unsigned int RoundTripPingLatencyCycles,
+	unsigned int ReorderingBytes,
+	double DCFCLK,
+	double FabricClock,
+	unsigned int PixelChunkSizeInKByte,
+	double ReturnBW,
+	unsigned int NumberOfActiveSurfaces,
+	unsigned int NumberOfDPP[],
+	unsigned int dpte_group_bytes[],
+	unsigned int tdlut_bytes_per_group[],
+	double HostVMInefficiencyFactor,
+	double HostVMInefficiencyFactorPrefetch,
+	unsigned int HostVMMinPageSize,
+	enum dml2_qos_param_type qos_type,
+	bool max_outstanding_when_urgent_expected,
+	unsigned int max_outstanding_requests,
+	unsigned int request_size_bytes_luma[],
+	unsigned int request_size_bytes_chroma[],
+	unsigned int MetaChunkSize,
+	unsigned int dchub_arb_to_ret_delay,
+	double Ttrip,
+	unsigned int hostvm_mode,
+
+	// output
+	double *ExtraLatency, // Tex
+	double *ExtraLatency_sr, // Tex_sr
+	double *ExtraLatencyPrefetch)
+
+{
+	double Tarb;
+	double Tarb_prefetch;
+	double Tex_trips;
+	unsigned int max_request_size_bytes = 0;
+
+	CalculateTarb(
+		display_cfg,
+		PixelChunkSizeInKByte,
+		NumberOfActiveSurfaces,
+		NumberOfDPP,
+		dpte_group_bytes,
+		tdlut_bytes_per_group,
+		HostVMInefficiencyFactor,
+		HostVMInefficiencyFactorPrefetch,
+		HostVMMinPageSize,
+		ReturnBW,
+		MetaChunkSize,
+		// output
+		&Tarb,
+		&Tarb_prefetch);
+
+	Tex_trips = (display_cfg->hostvm_enable && hostvm_mode == 1) ? (2.0 * Ttrip) : 0.0;
+
+	for (unsigned int k = 0; k < NumberOfActiveSurfaces; ++k) {
+		if (request_size_bytes_luma[k] > max_request_size_bytes)
+			max_request_size_bytes = request_size_bytes_luma[k];
+		if (request_size_bytes_chroma[k] > max_request_size_bytes)
+			max_request_size_bytes = request_size_bytes_chroma[k];
+	}
+
+	if (qos_type == dml2_qos_param_type_dcn4x) {
+		*ExtraLatency_sr = dchub_arb_to_ret_delay / DCFCLK;
+		*ExtraLatency = *ExtraLatency_sr;
+		if (max_outstanding_when_urgent_expected)
+			*ExtraLatency = *ExtraLatency + (ROBBufferSizeInKByte * 1024 - max_outstanding_requests * max_request_size_bytes) / ReturnBW;
+	} else {
+		*ExtraLatency_sr = dchub_arb_to_ret_delay / DCFCLK + RoundTripPingLatencyCycles / FabricClock + ReorderingBytes / ReturnBW;
+		*ExtraLatency = *ExtraLatency_sr;
+	}
+	*ExtraLatency = *ExtraLatency + Tex_trips;
+	*ExtraLatencyPrefetch = *ExtraLatency + Tarb_prefetch;
+	*ExtraLatency = *ExtraLatency + Tarb;
+	*ExtraLatency_sr = *ExtraLatency_sr + Tarb;
+
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: qos_type=%u\n", __func__, qos_type);
+	DML_LOG_VERBOSE("DML::%s: hostvm_mode=%u\n", __func__, hostvm_mode);
+	DML_LOG_VERBOSE("DML::%s: Tex_trips=%f\n", __func__, Tex_trips);
+	DML_LOG_VERBOSE("DML::%s: max_outstanding_when_urgent_expected=%u\n", __func__, max_outstanding_when_urgent_expected);
+	DML_LOG_VERBOSE("DML::%s: FabricClock=%f\n", __func__, FabricClock);
+	DML_LOG_VERBOSE("DML::%s: DCFCLK=%f\n", __func__, DCFCLK);
+	DML_LOG_VERBOSE("DML::%s: ReturnBW=%f\n", __func__, ReturnBW);
+	DML_LOG_VERBOSE("DML::%s: RoundTripPingLatencyCycles=%u\n", __func__, RoundTripPingLatencyCycles);
+	DML_LOG_VERBOSE("DML::%s: ReorderingBytes=%u\n", __func__, ReorderingBytes);
+	DML_LOG_VERBOSE("DML::%s: Tarb=%f\n", __func__, Tarb);
+	DML_LOG_VERBOSE("DML::%s: ExtraLatency=%f\n", __func__, *ExtraLatency);
+	DML_LOG_VERBOSE("DML::%s: ExtraLatency_sr=%f\n", __func__, *ExtraLatency_sr);
+	DML_LOG_VERBOSE("DML::%s: ExtraLatencyPrefetch=%f\n", __func__, *ExtraLatencyPrefetch);
+#endif
+}
+
+static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch, struct dml2_core_calcs_CalculatePrefetchSchedule_params *p)
+{
+	struct dml2_core_calcs_CalculatePrefetchSchedule_locals *s = &scratch->CalculatePrefetchSchedule_locals;
+	bool dcc_mrq_enable;
+
+	unsigned int vm_bytes;
+	unsigned int extra_tdpe_bytes;
+	unsigned int tdlut_row_bytes;
+	unsigned int Lo;
+
+	s->NoTimeToPrefetch = false;
+	s->DPPCycles = 0;
+	s->DISPCLKCycles = 0;
+	s->DSTTotalPixelsAfterScaler = 0.0;
+	s->LineTime = 0.0;
+	s->dst_y_prefetch_equ = 0.0;
+	s->prefetch_bw_oto = 0.0;
+	s->Tvm_oto = 0.0;
+	s->Tr0_oto = 0.0;
+	s->Tvm_oto_lines = 0.0;
+	s->Tr0_oto_lines = 0.0;
+	s->dst_y_prefetch_oto = 0.0;
+	s->TimeForFetchingVM = 0.0;
+	s->TimeForFetchingRowInVBlank = 0.0;
+	s->LinesToRequestPrefetchPixelData = 0.0;
+	s->HostVMDynamicLevelsTrips = 0;
+	s->trip_to_mem = 0.0;
+	*p->Tvm_trips = 0.0;
+	*p->Tr0_trips = 0.0;
+	s->Tvm_trips_rounded = 0.0;
+	s->Tr0_trips_rounded = 0.0;
+	s->max_Tsw = 0.0;
+	s->Lsw_oto = 0.0;
+	*p->Tpre_rounded = 0.0;
+	s->prefetch_bw_equ = 0.0;
+	s->Tvm_equ = 0.0;
+	s->Tr0_equ = 0.0;
+	s->Tdmbf = 0.0;
+	s->Tdmec = 0.0;
+	s->Tdmsks = 0.0;
+	*p->prefetch_sw_bytes = 0.0;
+	s->prefetch_bw_pr = 0.0;
+	s->bytes_pp = 0.0;
+	s->dep_bytes = 0.0;
+	s->min_Lsw_oto = 0.0;
+	s->min_Lsw_equ = 0.0;
+	s->Tsw_est1 = 0.0;
+	s->Tsw_est2 = 0.0;
+	s->Tsw_est3 = 0.0;
+	s->cursor_prefetch_bytes = 0;
+	*p->prefetch_cursor_bw = 0;
+	*p->RequiredPrefetchBWMax = 0.0;
+
+	dcc_mrq_enable = (p->dcc_enable && p->mrq_present);
+
+	s->TWait_p = p->TWait - p->Ttrip; // TWait includes max(Turg, Ttrip) and Ttrip here is already max(Turg, Ttrip)
+
+	if (p->display_cfg->gpuvm_enable == true && p->display_cfg->hostvm_enable == true) {
+		s->HostVMDynamicLevelsTrips = p->display_cfg->hostvm_max_non_cached_page_table_levels;
+	} else {
+		s->HostVMDynamicLevelsTrips = 0;
+	}
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: dcc_enable = %u\n", __func__, p->dcc_enable);
+	DML_LOG_VERBOSE("DML::%s: mrq_present = %u\n", __func__, p->mrq_present);
+	DML_LOG_VERBOSE("DML::%s: dcc_mrq_enable = %u\n", __func__, dcc_mrq_enable);
+	DML_LOG_VERBOSE("DML::%s: GPUVMEnable = %u\n", __func__, p->display_cfg->gpuvm_enable);
+	DML_LOG_VERBOSE("DML::%s: GPUVMPageTableLevels = %u\n", __func__, p->display_cfg->gpuvm_max_page_table_levels);
+	DML_LOG_VERBOSE("DML::%s: DCCEnable = %u\n", __func__, p->myPipe->DCCEnable);
+	DML_LOG_VERBOSE("DML::%s: VStartup = %u\n", __func__, p->VStartup);
+	DML_LOG_VERBOSE("DML::%s: HostVMEnable = %u\n", __func__, p->display_cfg->hostvm_enable);
+	DML_LOG_VERBOSE("DML::%s: HostVMInefficiencyFactor = %f\n", __func__, p->HostVMInefficiencyFactor);
+	DML_LOG_VERBOSE("DML::%s: TWait = %f\n", __func__, p->TWait);
+	DML_LOG_VERBOSE("DML::%s: TWait_p = %f\n", __func__, s->TWait_p);
+	DML_LOG_VERBOSE("DML::%s: Ttrip = %f\n", __func__, p->Ttrip);
+	DML_LOG_VERBOSE("DML::%s: myPipe->Dppclk = %f\n", __func__, p->myPipe->Dppclk);
+	DML_LOG_VERBOSE("DML::%s: myPipe->Dispclk = %f\n", __func__, p->myPipe->Dispclk);
+#endif
+	CalculateVUpdateAndDynamicMetadataParameters(
+		p->MaxInterDCNTileRepeaters,
+		p->myPipe->Dppclk,
+		p->myPipe->Dispclk,
+		p->myPipe->DCFClkDeepSleep,
+		p->myPipe->PixelClock,
+		p->myPipe->HTotal,
+		p->myPipe->VBlank,
+		p->DynamicMetadataTransmittedBytes,
+		p->DynamicMetadataLinesBeforeActiveRequired,
+		p->myPipe->InterlaceEnable,
+		p->myPipe->ProgressiveToInterlaceUnitInOPP,
+		p->TSetup,
+
+		// Output
+		&s->Tdmbf,
+		&s->Tdmec,
+		&s->Tdmsks,
+		p->VUpdateOffsetPix,
+		p->VUpdateWidthPix,
+		p->VReadyOffsetPix);
+
+	s->LineTime = p->myPipe->HTotal / p->myPipe->PixelClock;
+	s->trip_to_mem = p->Ttrip;
+	*p->Tvm_trips = p->ExtraLatencyPrefetch + math_max2(s->trip_to_mem * (p->display_cfg->gpuvm_max_page_table_levels * (s->HostVMDynamicLevelsTrips + 1)), p->Turg);
+	if (dcc_mrq_enable)
+		*p->Tvm_trips_flip = *p->Tvm_trips;
+	else
+		*p->Tvm_trips_flip = *p->Tvm_trips - s->trip_to_mem;
+
+	*p->Tr0_trips_flip = s->trip_to_mem * (s->HostVMDynamicLevelsTrips + 1);
+	*p->Tr0_trips = math_max2(*p->Tr0_trips_flip, p->tdlut_opt_time / 2);
+
+	if (p->DynamicMetadataVMEnabled == true) {
+		*p->Tdmdl_vm = s->TWait_p + *p->Tvm_trips;
+		*p->Tdmdl = *p->Tdmdl_vm + p->Ttrip;
+	} else {
+		*p->Tdmdl_vm = 0;
+		*p->Tdmdl = s->TWait_p + p->ExtraLatencyPrefetch + p->Ttrip; // Tex
+	}
+
+	if (p->DynamicMetadataEnable == true) {
+		if (p->VStartup * s->LineTime < *p->TSetup + *p->Tdmdl + s->Tdmbf + s->Tdmec + s->Tdmsks) {
+			*p->NotEnoughTimeForDynamicMetadata = true;
+			DML_LOG_VERBOSE("DML::%s: Not Enough Time for Dynamic Meta!\n", __func__);
+			DML_LOG_VERBOSE("DML::%s: Tdmbf: %fus - time for dmd transfer from dchub to dio output buffer\n", __func__, s->Tdmbf);
+			DML_LOG_VERBOSE("DML::%s: Tdmec: %fus - time dio takes to transfer dmd\n", __func__, s->Tdmec);
+			DML_LOG_VERBOSE("DML::%s: Tdmsks: %fus - time before active dmd must complete transmission at dio\n", __func__, s->Tdmsks);
+			DML_LOG_VERBOSE("DML::%s: Tdmdl: %fus - time for fabric to become ready and fetch dmd \n", __func__, *p->Tdmdl);
+		} else {
+			*p->NotEnoughTimeForDynamicMetadata = false;
+		}
+	} else {
+		*p->NotEnoughTimeForDynamicMetadata = false;
+	}
+
+	if (p->myPipe->ScalerEnabled)
+		s->DPPCycles = (unsigned int)(p->DPPCLKDelaySubtotalPlusCNVCFormater + p->DPPCLKDelaySCL);
+	else
+		s->DPPCycles = (unsigned int)(p->DPPCLKDelaySubtotalPlusCNVCFormater + p->DPPCLKDelaySCLLBOnly);
+
+	s->DPPCycles = (unsigned int)(s->DPPCycles + p->myPipe->NumberOfCursors * p->DPPCLKDelayCNVCCursor);
+
+	s->DISPCLKCycles = (unsigned int)p->DISPCLKDelaySubtotal;
+
+	if (p->myPipe->Dppclk == 0.0 || p->myPipe->Dispclk == 0.0)
+		return true;
+
+	*p->DSTXAfterScaler = (unsigned int)math_round(s->DPPCycles * p->myPipe->PixelClock / p->myPipe->Dppclk + s->DISPCLKCycles * p->myPipe->PixelClock / p->myPipe->Dispclk + p->DSCDelay);
+	*p->DSTXAfterScaler = (unsigned int)math_round(*p->DSTXAfterScaler + (p->myPipe->ODMMode != dml2_odm_mode_bypass ? 18 : 0) + (p->myPipe->DPPPerSurface - 1) * p->DPP_RECOUT_WIDTH +
+		((p->myPipe->ODMMode == dml2_odm_mode_split_1to2 || p->myPipe->ODMMode == dml2_odm_mode_mso_1to2) ? (double)p->myPipe->HActive / 2.0 : 0) +
+		((p->myPipe->ODMMode == dml2_odm_mode_mso_1to4) ? (double)p->myPipe->HActive * 3.0 / 4.0 : 0));
+
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: DynamicMetadataVMEnabled = %u\n", __func__, p->DynamicMetadataVMEnabled);
+	DML_LOG_VERBOSE("DML::%s: DPPCycles = %u\n", __func__, s->DPPCycles);
+	DML_LOG_VERBOSE("DML::%s: PixelClock = %f\n", __func__, p->myPipe->PixelClock);
+	DML_LOG_VERBOSE("DML::%s: Dppclk = %f\n", __func__, p->myPipe->Dppclk);
+	DML_LOG_VERBOSE("DML::%s: DISPCLKCycles = %u\n", __func__, s->DISPCLKCycles);
+	DML_LOG_VERBOSE("DML::%s: DISPCLK = %f\n", __func__, p->myPipe->Dispclk);
+	DML_LOG_VERBOSE("DML::%s: DSCDelay = %u\n", __func__, p->DSCDelay);
+	DML_LOG_VERBOSE("DML::%s: ODMMode = %u\n", __func__, p->myPipe->ODMMode);
+	DML_LOG_VERBOSE("DML::%s: DPP_RECOUT_WIDTH = %u\n", __func__, p->DPP_RECOUT_WIDTH);
+	DML_LOG_VERBOSE("DML::%s: DSTXAfterScaler = %u\n", __func__, *p->DSTXAfterScaler);
+
+	DML_LOG_VERBOSE("DML::%s: setup_for_tdlut = %u\n", __func__, p->setup_for_tdlut);
+	DML_LOG_VERBOSE("DML::%s: tdlut_opt_time = %f\n", __func__, p->tdlut_opt_time);
+	DML_LOG_VERBOSE("DML::%s: tdlut_pte_bytes_per_frame = %u\n", __func__, p->tdlut_pte_bytes_per_frame);
+	DML_LOG_VERBOSE("DML::%s: tdlut_drain_time = %f\n", __func__, p->tdlut_drain_time);
+#endif
+
+	if (p->OutputFormat == dml2_420 || (p->myPipe->InterlaceEnable && p->myPipe->ProgressiveToInterlaceUnitInOPP))
+		*p->DSTYAfterScaler = 1;
+	else
+		*p->DSTYAfterScaler = 0;
+
+	s->DSTTotalPixelsAfterScaler = *p->DSTYAfterScaler * p->myPipe->HTotal + *p->DSTXAfterScaler;
+	*p->DSTYAfterScaler = (unsigned int)(math_floor2(s->DSTTotalPixelsAfterScaler / p->myPipe->HTotal, 1));
+	*p->DSTXAfterScaler = (unsigned int)(s->DSTTotalPixelsAfterScaler - ((double)(*p->DSTYAfterScaler * p->myPipe->HTotal)));
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: DSTXAfterScaler = %u (final)\n", __func__, *p->DSTXAfterScaler);
+	DML_LOG_VERBOSE("DML::%s: DSTYAfterScaler = %u (final)\n", __func__, *p->DSTYAfterScaler);
+#endif
+
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: Tr0_trips = %f\n", __func__, *p->Tr0_trips);
+	DML_LOG_VERBOSE("DML::%s: Tvm_trips = %f\n", __func__, *p->Tvm_trips);
+	DML_LOG_VERBOSE("DML::%s: trip_to_mem = %f\n", __func__, s->trip_to_mem);
+	DML_LOG_VERBOSE("DML::%s: ExtraLatencyPrefetch = %f\n", __func__, p->ExtraLatencyPrefetch);
+	DML_LOG_VERBOSE("DML::%s: GPUVMPageTableLevels = %u\n", __func__, p->display_cfg->gpuvm_max_page_table_levels);
+	DML_LOG_VERBOSE("DML::%s: HostVMDynamicLevelsTrips = %u\n", __func__, s->HostVMDynamicLevelsTrips);
+#endif
+	if (p->display_cfg->gpuvm_enable) {
+		s->Tvm_trips_rounded = math_ceil2(4.0 * *p->Tvm_trips / s->LineTime, 1.0) / 4.0 * s->LineTime;
+		*p->Tvm_trips_flip_rounded = math_ceil2(4.0 * *p->Tvm_trips_flip / s->LineTime, 1.0) / 4.0 * s->LineTime;
+	} else {
+		if (p->DynamicMetadataEnable || dcc_mrq_enable || p->setup_for_tdlut)
+			s->Tvm_trips_rounded = math_max2(s->LineTime * math_ceil2(4.0*math_max3(p->ExtraLatencyPrefetch, p->Turg, s->trip_to_mem)/s->LineTime, 1)/4, s->LineTime/4.0);
+		else
+			s->Tvm_trips_rounded = s->LineTime / 4.0;
+		*p->Tvm_trips_flip_rounded = s->LineTime / 4.0;
+	}
+
+	s->Tvm_trips_rounded = math_max2(s->Tvm_trips_rounded, s->LineTime / 4.0);
+	*p->Tvm_trips_flip_rounded = math_max2(*p->Tvm_trips_flip_rounded, s->LineTime / 4.0);
+
+	if (p->display_cfg->gpuvm_enable == true || p->setup_for_tdlut || dcc_mrq_enable) {
+		s->Tr0_trips_rounded = math_ceil2(4.0 * *p->Tr0_trips / s->LineTime, 1.0) / 4.0 * s->LineTime;
+		*p->Tr0_trips_flip_rounded = math_ceil2(4.0 * *p->Tr0_trips_flip / s->LineTime, 1.0) / 4.0 * s->LineTime;
+	} else {
+		s->Tr0_trips_rounded = s->LineTime / 4.0;
+		*p->Tr0_trips_flip_rounded = s->LineTime / 4.0;
+	}
+	s->Tr0_trips_rounded = math_max2(s->Tr0_trips_rounded, s->LineTime / 4.0);
+	*p->Tr0_trips_flip_rounded = math_max2(*p->Tr0_trips_flip_rounded, s->LineTime / 4.0);
+
+	if (p->display_cfg->gpuvm_enable == true) {
+		if (p->display_cfg->gpuvm_max_page_table_levels >= 3) {
+			*p->Tno_bw = p->ExtraLatencyPrefetch + s->trip_to_mem * (double)((p->display_cfg->gpuvm_max_page_table_levels - 2) * (s->HostVMDynamicLevelsTrips + 1));
+		} else if (p->display_cfg->gpuvm_max_page_table_levels == 1 && !dcc_mrq_enable && !p->setup_for_tdlut) {
+			*p->Tno_bw = p->ExtraLatencyPrefetch;
+		} else {
+			*p->Tno_bw = 0;
+		}
+	} else {
+		*p->Tno_bw = 0;
+	}
+
+	if (p->mrq_present || p->display_cfg->gpuvm_max_page_table_levels >= 3)
+		*p->Tno_bw_flip = *p->Tno_bw;
+	else
+		*p->Tno_bw_flip = 0; //because there is no 3DLUT for iFlip
+
+	if (dml_is_420(p->myPipe->SourcePixelFormat)) {
+		s->bytes_pp = p->myPipe->BytePerPixelY + p->myPipe->BytePerPixelC / 4.0;
+	} else {
+		s->bytes_pp = p->myPipe->BytePerPixelY + p->myPipe->BytePerPixelC;
+	}
+
+	*p->prefetch_sw_bytes = p->PrefetchSourceLinesY * p->swath_width_luma_ub * p->myPipe->BytePerPixelY + p->PrefetchSourceLinesC * p->swath_width_chroma_ub * p->myPipe->BytePerPixelC;
+	*p->prefetch_sw_bytes = *p->prefetch_sw_bytes * p->mall_prefetch_sdp_overhead_factor;
+
+	vm_bytes = p->vm_bytes; // vm_bytes is dpde0_bytes_per_frame_ub_l + dpde0_bytes_per_frame_ub_c + 2*extra_dpde_bytes;
+	extra_tdpe_bytes = (unsigned int)math_max2(0, (p->display_cfg->gpuvm_max_page_table_levels - 1) * 128);
+
+	if (p->setup_for_tdlut)
+		vm_bytes = vm_bytes + p->tdlut_pte_bytes_per_frame + (p->display_cfg->gpuvm_enable ? extra_tdpe_bytes : 0);
+
+	tdlut_row_bytes = (unsigned long) math_ceil2(p->tdlut_bytes_per_frame/2.0, 1.0);
+
+	s->min_Lsw_oto = math_max2(p->PrefetchSourceLinesY, p->PrefetchSourceLinesC) / __DML2_CALCS_MAX_VRATIO_PRE_OTO__;
+	s->min_Lsw_oto = math_max2(s->min_Lsw_oto, p->tdlut_drain_time / s->LineTime);
+	s->min_Lsw_oto = math_max2(s->min_Lsw_oto, 2.0);
+
+	// use vactive swath bw for prefetch oto and also cap prefetch_bw_oto to max_vratio_oto
+	// Note: in prefetch calculation, acounting is done mostly per-pipe.
+	// vactive swath bw represents the per-surface (aka per dml plane) bw to move vratio_l/c lines of bytes_l/c per line time
+	s->per_pipe_vactive_sw_bw = p->vactive_sw_bw_l / (double)p->myPipe->DPPPerSurface;
+
+	// one-to-one prefetch bw as one line of bytes per line time (as per vratio_pre_l/c = 1)
+	s->prefetch_bw_oto = (p->swath_width_luma_ub * p->myPipe->BytePerPixelY) / s->LineTime;
+
+	if (p->myPipe->BytePerPixelC > 0) {
+		s->per_pipe_vactive_sw_bw += p->vactive_sw_bw_c / (double)p->myPipe->DPPPerSurface;
+		s->prefetch_bw_oto += (p->swath_width_chroma_ub * p->myPipe->BytePerPixelC) / s->LineTime;
+	}
+
+	/* oto prefetch bw should be always be less than total vactive bw */
+	//DML_ASSERT(s->prefetch_bw_oto < s->per_pipe_vactive_sw_bw * p->myPipe->DPPPerSurface);
+
+	s->prefetch_bw_oto = math_max2(s->per_pipe_vactive_sw_bw, s->prefetch_bw_oto) * p->mall_prefetch_sdp_overhead_factor;
+
+	s->prefetch_bw_oto = math_min2(s->prefetch_bw_oto, *p->prefetch_sw_bytes/(s->min_Lsw_oto*s->LineTime));
+
+	s->Lsw_oto = math_ceil2(4.0 * *p->prefetch_sw_bytes / s->prefetch_bw_oto / s->LineTime, 1.0) / 4.0;
+
+	s->prefetch_bw_oto = math_max3(s->prefetch_bw_oto,
+					p->vm_bytes * p->HostVMInefficiencyFactor / (31 * s->LineTime) - *p->Tno_bw,
+					(p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor + p->meta_row_bytes + tdlut_row_bytes) / (15 * s->LineTime));
+
+	/* oto bw needs to be outputted even if the oto schedule isn't being used to avoid ms/mp mismatch.
+	 * mp will fail if ms decides to use equ schedule and mp decides to use oto schedule
+	 * and the required bandwidth increases when going from ms to mp
+	 */
+	*p->RequiredPrefetchBWMax = s->prefetch_bw_oto;
+
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: vactive_sw_bw_l = %f\n", __func__, p->vactive_sw_bw_l);
+	DML_LOG_VERBOSE("DML::%s: vactive_sw_bw_c = %f\n", __func__, p->vactive_sw_bw_c);
+	DML_LOG_VERBOSE("DML::%s: per_pipe_vactive_sw_bw = %f\n", __func__, s->per_pipe_vactive_sw_bw);
+#endif
+
+	if (p->display_cfg->gpuvm_enable == true) {
+		s->Tvm_oto = math_max3(
+			*p->Tvm_trips,
+			*p->Tno_bw + vm_bytes * p->HostVMInefficiencyFactor / s->prefetch_bw_oto,
+			s->LineTime / 4.0);
+
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: Tvm_oto max0 = %f\n", __func__, *p->Tvm_trips);
+		DML_LOG_VERBOSE("DML::%s: Tvm_oto max1 = %f\n", __func__, *p->Tno_bw + vm_bytes * p->HostVMInefficiencyFactor / s->prefetch_bw_oto);
+		DML_LOG_VERBOSE("DML::%s: Tvm_oto max2 = %f\n", __func__, s->LineTime / 4.0);
+#endif
+	} else {
+		s->Tvm_oto = s->Tvm_trips_rounded;
+	}
+
+	if ((p->display_cfg->gpuvm_enable == true || p->setup_for_tdlut || dcc_mrq_enable)) {
+		s->Tr0_oto = math_max3(
+			*p->Tr0_trips,
+			(p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor + p->meta_row_bytes + tdlut_row_bytes) / s->prefetch_bw_oto,
+			s->LineTime / 4.0);
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: Tr0_oto max0 = %f\n", __func__, *p->Tr0_trips);
+		DML_LOG_VERBOSE("DML::%s: Tr0_oto max1 = %f\n", __func__, (p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor + p->meta_row_bytes + tdlut_row_bytes) / s->prefetch_bw_oto);
+		DML_LOG_VERBOSE("DML::%s: Tr0_oto max2 = %f\n", __func__, s->LineTime / 4);
+#endif
+	} else
+		s->Tr0_oto = s->LineTime / 4.0;
+
+	s->Tvm_oto_lines = math_ceil2(4.0 * s->Tvm_oto / s->LineTime, 1) / 4.0;
+	s->Tr0_oto_lines = math_ceil2(4.0 * s->Tr0_oto / s->LineTime, 1) / 4.0;
+	s->dst_y_prefetch_oto = s->Tvm_oto_lines + 2 * s->Tr0_oto_lines + s->Lsw_oto;
+
+#ifdef DML_GLOBAL_PREFETCH_CHECK
+	DML_LOG_VERBOSE("DML::%s: impacted_Tpre = %f\n", __func__, p->impacted_dst_y_pre);
+	if (p->impacted_dst_y_pre > 0) {
+		DML_LOG_VERBOSE("DML::%s: dst_y_prefetch_oto = %f\n", __func__, s->dst_y_prefetch_oto);
+		s->dst_y_prefetch_oto = math_max2(s->dst_y_prefetch_oto, p->impacted_dst_y_pre);
+		DML_LOG_VERBOSE("DML::%s: dst_y_prefetch_oto = %f (impacted)\n", __func__, s->dst_y_prefetch_oto);
+	}
+#endif
+	*p->Tpre_oto = s->dst_y_prefetch_oto * s->LineTime;
+
+	//To (time for delay after scaler) in line time
+	Lo = (unsigned int)(*p->DSTYAfterScaler + (double)*p->DSTXAfterScaler / (double)p->myPipe->HTotal);
+
+	s->min_Lsw_equ = math_max2(p->PrefetchSourceLinesY, p->PrefetchSourceLinesC) / __DML2_CALCS_MAX_VRATIO_PRE_EQU__;
+	s->min_Lsw_equ = math_max2(s->min_Lsw_equ, p->tdlut_drain_time / s->LineTime);
+	s->min_Lsw_equ = math_max2(s->min_Lsw_equ, 2.0);
+	//Tpre_equ in line time
+	if (p->DynamicMetadataVMEnabled && p->DynamicMetadataEnable)
+		s->dst_y_prefetch_equ = p->VStartup - (*p->TSetup + math_max2(p->TCalc, *p->Tvm_trips) + s->TWait_p) / s->LineTime - Lo;
+	else
+		s->dst_y_prefetch_equ = p->VStartup - (*p->TSetup + math_max2(p->TCalc, p->ExtraLatencyPrefetch) + s->TWait_p) / s->LineTime - Lo;
+
+#ifdef DML_GLOBAL_PREFETCH_CHECK
+	s->dst_y_prefetch_equ_impacted = math_max2(p->impacted_dst_y_pre, s->dst_y_prefetch_equ);
+
+	s->dst_y_prefetch_equ_impacted = math_min2(s->dst_y_prefetch_equ_impacted, 63.75); // limit to the reg limit of U6.2 for DST_Y_PREFETCH
+
+	if (s->dst_y_prefetch_equ_impacted > s->dst_y_prefetch_equ)
+		s->dst_y_prefetch_equ -= s->dst_y_prefetch_equ_impacted - s->dst_y_prefetch_equ;
+#endif
+
+	s->dst_y_prefetch_equ = math_min2(s->dst_y_prefetch_equ, 63.75); // limit to the reg limit of U6.2 for DST_Y_PREFETCH
+
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: HTotal = %u\n", __func__, p->myPipe->HTotal);
+	DML_LOG_VERBOSE("DML::%s: min_Lsw_oto = %f\n", __func__, s->min_Lsw_oto);
+	DML_LOG_VERBOSE("DML::%s: min_Lsw_equ = %f\n", __func__, s->min_Lsw_equ);
+	DML_LOG_VERBOSE("DML::%s: Tno_bw = %f\n", __func__, *p->Tno_bw);
+	DML_LOG_VERBOSE("DML::%s: Tno_bw_flip = %f\n", __func__, *p->Tno_bw_flip);
+	DML_LOG_VERBOSE("DML::%s: ExtraLatencyPrefetch = %f\n", __func__, p->ExtraLatencyPrefetch);
+	DML_LOG_VERBOSE("DML::%s: trip_to_mem = %f\n", __func__, s->trip_to_mem);
+	DML_LOG_VERBOSE("DML::%s: mall_prefetch_sdp_overhead_factor = %f\n", __func__, p->mall_prefetch_sdp_overhead_factor);
+	DML_LOG_VERBOSE("DML::%s: BytePerPixelY = %u\n", __func__, p->myPipe->BytePerPixelY);
+	DML_LOG_VERBOSE("DML::%s: PrefetchSourceLinesY = %f\n", __func__, p->PrefetchSourceLinesY);
+	DML_LOG_VERBOSE("DML::%s: swath_width_luma_ub = %u\n", __func__, p->swath_width_luma_ub);
+	DML_LOG_VERBOSE("DML::%s: BytePerPixelC = %u\n", __func__, p->myPipe->BytePerPixelC);
+	DML_LOG_VERBOSE("DML::%s: PrefetchSourceLinesC = %f\n", __func__, p->PrefetchSourceLinesC);
+	DML_LOG_VERBOSE("DML::%s: swath_width_chroma_ub = %u\n", __func__, p->swath_width_chroma_ub);
+	DML_LOG_VERBOSE("DML::%s: prefetch_sw_bytes = %f\n", __func__, *p->prefetch_sw_bytes);
+	DML_LOG_VERBOSE("DML::%s: max_Tsw = %f\n", __func__, s->max_Tsw);
+	DML_LOG_VERBOSE("DML::%s: bytes_pp = %f\n", __func__, s->bytes_pp);
+	DML_LOG_VERBOSE("DML::%s: vm_bytes = %u\n", __func__, vm_bytes);
+	DML_LOG_VERBOSE("DML::%s: PixelPTEBytesPerRow = %u\n", __func__, p->PixelPTEBytesPerRow);
+	DML_LOG_VERBOSE("DML::%s: HostVMInefficiencyFactor = %f\n", __func__, p->HostVMInefficiencyFactor);
+	DML_LOG_VERBOSE("DML::%s: Tvm_trips = %f\n", __func__, *p->Tvm_trips);
+	DML_LOG_VERBOSE("DML::%s: Tr0_trips = %f\n", __func__, *p->Tr0_trips);
+	DML_LOG_VERBOSE("DML::%s: Tvm_trips_flip = %f\n", __func__, *p->Tvm_trips_flip);
+	DML_LOG_VERBOSE("DML::%s: Tr0_trips_flip = %f\n", __func__, *p->Tr0_trips_flip);
+	DML_LOG_VERBOSE("DML::%s: prefetch_bw_pr = %f\n", __func__, s->prefetch_bw_pr);
+	DML_LOG_VERBOSE("DML::%s: prefetch_bw_oto = %f\n", __func__, s->prefetch_bw_oto);
+	DML_LOG_VERBOSE("DML::%s: Tr0_oto = %f\n", __func__, s->Tr0_oto);
+	DML_LOG_VERBOSE("DML::%s: Tvm_oto = %f\n", __func__, s->Tvm_oto);
+	DML_LOG_VERBOSE("DML::%s: Tvm_oto_lines = %f\n", __func__, s->Tvm_oto_lines);
+	DML_LOG_VERBOSE("DML::%s: Tr0_oto_lines = %f\n", __func__, s->Tr0_oto_lines);
+	DML_LOG_VERBOSE("DML::%s: Lsw_oto = %f\n", __func__, s->Lsw_oto);
+	DML_LOG_VERBOSE("DML::%s: dst_y_prefetch_oto = %f\n", __func__, s->dst_y_prefetch_oto);
+	DML_LOG_VERBOSE("DML::%s: dst_y_prefetch_equ = %f\n", __func__, s->dst_y_prefetch_equ);
+	DML_LOG_VERBOSE("DML::%s: tdlut_row_bytes = %d\n", __func__, tdlut_row_bytes);
+	DML_LOG_VERBOSE("DML::%s: meta_row_bytes = %d\n", __func__, p->meta_row_bytes);
+#endif
+	s->dst_y_prefetch_equ = math_floor2(4.0 * (s->dst_y_prefetch_equ + 0.125), 1) / 4.0;
+	*p->Tpre_rounded = s->dst_y_prefetch_equ * s->LineTime;
+
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: dst_y_prefetch_equ: %f (after round)\n", __func__, s->dst_y_prefetch_equ);
+	DML_LOG_VERBOSE("DML::%s: LineTime: %f\n", __func__, s->LineTime);
+	DML_LOG_VERBOSE("DML::%s: VStartup: %u\n", __func__, p->VStartup);
+	DML_LOG_VERBOSE("DML::%s: Tvstartup: %fus - time between vstartup and first pixel of active\n", __func__, p->VStartup * s->LineTime);
+	DML_LOG_VERBOSE("DML::%s: TSetup: %fus - time from vstartup to vready\n", __func__, *p->TSetup);
+	DML_LOG_VERBOSE("DML::%s: TCalc: %fus - time for calculations in dchub starting at vready\n", __func__, p->TCalc);
+	DML_LOG_VERBOSE("DML::%s: TWait: %fus - time for fabric to become ready max(pstate exit,cstate enter/exit, urgent latency) after TCalc\n", __func__, p->TWait);
+	DML_LOG_VERBOSE("DML::%s: Tdmbf: %fus - time for dmd transfer from dchub to dio output buffer\n", __func__, s->Tdmbf);
+	DML_LOG_VERBOSE("DML::%s: Tdmec: %fus - time dio takes to transfer dmd\n", __func__, s->Tdmec);
+	DML_LOG_VERBOSE("DML::%s: Tdmsks: %fus - time before active dmd must complete transmission at dio\n", __func__, s->Tdmsks);
+	DML_LOG_VERBOSE("DML::%s: TWait = %f\n", __func__, p->TWait);
+	DML_LOG_VERBOSE("DML::%s: TWait_p = %f\n", __func__, s->TWait_p);
+	DML_LOG_VERBOSE("DML::%s: Ttrip = %f\n", __func__, p->Ttrip);
+	DML_LOG_VERBOSE("DML::%s: Tex = %f\n", __func__, p->ExtraLatencyPrefetch);
+	DML_LOG_VERBOSE("DML::%s: Tdmdl_vm: %fus - time for vm stages of dmd \n", __func__, *p->Tdmdl_vm);
+	DML_LOG_VERBOSE("DML::%s: Tdmdl: %fus - time for fabric to become ready and fetch dmd \n", __func__, *p->Tdmdl);
+	DML_LOG_VERBOSE("DML::%s: TWait_p: %fus\n", __func__, s->TWait_p);
+	DML_LOG_VERBOSE("DML::%s: Ttrip: %fus\n", __func__, p->Ttrip);
+	DML_LOG_VERBOSE("DML::%s: DSTXAfterScaler: %u pixels - number of pixel clocks pipeline and buffer delay after scaler \n", __func__, *p->DSTXAfterScaler);
+	DML_LOG_VERBOSE("DML::%s: DSTYAfterScaler: %u lines - number of lines of pipeline and buffer delay after scaler \n", __func__, *p->DSTYAfterScaler);
+	DML_LOG_VERBOSE("DML::%s: vm_bytes: %f (hvm inefficiency scaled)\n", __func__, vm_bytes*p->HostVMInefficiencyFactor);
+	DML_LOG_VERBOSE("DML::%s: row_bytes: %f (hvm inefficiency scaled, 1 row)\n", __func__, p->PixelPTEBytesPerRow*p->HostVMInefficiencyFactor+p->meta_row_bytes+tdlut_row_bytes);
+	DML_LOG_VERBOSE("DML::%s: Tno_bw: %f\n", __func__, *p->Tno_bw);
+	DML_LOG_VERBOSE("DML::%s: Tpre=%f Tpre_rounded: %f, delta=%f\n", __func__, (s->dst_y_prefetch_equ * s->LineTime), *p->Tpre_rounded, (*p->Tpre_rounded - (s->dst_y_prefetch_equ * s->LineTime)));
+	DML_LOG_VERBOSE("DML::%s: Tvm_trips=%f Tvm_trips_rounded: %f, delta=%f\n", __func__, *p->Tvm_trips, s->Tvm_trips_rounded, (s->Tvm_trips_rounded - *p->Tvm_trips));
+#endif
+
+	*p->dst_y_per_vm_vblank = 0;
+	*p->dst_y_per_row_vblank = 0;
+	*p->VRatioPrefetchY = 0;
+	*p->VRatioPrefetchC = 0;
+	*p->RequiredPrefetchPixelDataBWLuma = 0;
+
+	// Derive bandwidth by finding how much data to move within the time constraint
+	// Tpre_rounded is Tpre rounding to 2-bit fraction
+	// Tvm_trips_rounded is Tvm_trips ceiling to 1/4 line time
+	// Tr0_trips_rounded is Tr0_trips ceiling to 1/4 line time
+	// So that means prefetch bw calculated can be higher since the total time available for prefetch is less
+	bool min_Lsw_equ_ok = *p->Tpre_rounded >= s->Tvm_trips_rounded + 2.0*s->Tr0_trips_rounded + s->min_Lsw_equ*s->LineTime;
+	bool tpre_gt_req_latency = true;
+#if 0
+	// Check that Tpre_rounded is big enough if all of the stages of the prefetch are time constrained.
+	// The terms Tvm_trips_rounded and Tr0_trips_rounded represent the min time constraints for the VM and row stages.
+	// Normally, these terms cover the overall time constraint for Tpre >= (Tex + max{Ttrip, Turg}), but if these terms are at their minimum, an explicit check is necessary.
+	tpre_gt_req_latency = *p->Tpre_rounded > (math_max2(p->Turg, s->trip_to_mem) + p->ExtraLatencyPrefetch);
+#endif
+
+	if (s->dst_y_prefetch_equ > 1 && min_Lsw_equ_ok && tpre_gt_req_latency) {
+		s->prefetch_bw1 = 0.;
+		s->prefetch_bw2 = 0.;
+		s->prefetch_bw3 = 0.;
+		s->prefetch_bw4 = 0.;
+
+		// prefetch_bw1: VM + 2*R0 + SW
+		if (*p->Tpre_rounded - *p->Tno_bw > 0) {
+			s->prefetch_bw1 = (vm_bytes * p->HostVMInefficiencyFactor
+				+ 2 * (p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor + p->meta_row_bytes + tdlut_row_bytes)
+				+ *p->prefetch_sw_bytes)
+				/ (*p->Tpre_rounded - *p->Tno_bw);
+			s->Tsw_est1 = *p->prefetch_sw_bytes / s->prefetch_bw1;
+		} else
+			s->prefetch_bw1 = 0;
+
+		DML_LOG_VERBOSE("DML::%s: prefetch_bw1: %f\n", __func__, s->prefetch_bw1);
+		if ((s->Tsw_est1 < s->min_Lsw_equ * s->LineTime) && (*p->Tpre_rounded - s->min_Lsw_equ * s->LineTime - 0.75 * s->LineTime - *p->Tno_bw > 0)) {
+			s->prefetch_bw1 = (vm_bytes * p->HostVMInefficiencyFactor + 2 * (p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor + p->meta_row_bytes + tdlut_row_bytes)) /
+				(*p->Tpre_rounded - s->min_Lsw_equ * s->LineTime - 0.75 * s->LineTime - *p->Tno_bw);
+#ifdef __DML_VBA_DEBUG__
+			DML_LOG_VERBOSE("DML::%s: vm and 2 rows bytes = %f\n", __func__, (vm_bytes * p->HostVMInefficiencyFactor + 2 * (p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor + p->meta_row_bytes + tdlut_row_bytes)));
+			DML_LOG_VERBOSE("DML::%s: Tpre_rounded = %f\n", __func__, *p->Tpre_rounded);
+			DML_LOG_VERBOSE("DML::%s: minus term = %f\n", __func__, s->min_Lsw_equ * s->LineTime + 0.75 * s->LineTime + *p->Tno_bw);
+			DML_LOG_VERBOSE("DML::%s: min_Lsw_equ = %f\n", __func__, s->min_Lsw_equ);
+			DML_LOG_VERBOSE("DML::%s: LineTime = %f\n", __func__, s->LineTime);
+			DML_LOG_VERBOSE("DML::%s: Tno_bw = %f\n", __func__, *p->Tno_bw);
+			DML_LOG_VERBOSE("DML::%s: Time to fetch vm and 2 rows = %f\n", __func__, (*p->Tpre_rounded - s->min_Lsw_equ * s->LineTime - 0.75 * s->LineTime - *p->Tno_bw));
+			DML_LOG_VERBOSE("DML::%s: prefetch_bw1: %f (updated)\n", __func__, s->prefetch_bw1);
+#endif
+		}
+
+		// prefetch_bw2: VM + SW
+		if (*p->Tpre_rounded - *p->Tno_bw - 2.0 * s->Tr0_trips_rounded > 0) {
+			s->prefetch_bw2 = (vm_bytes * p->HostVMInefficiencyFactor + *p->prefetch_sw_bytes) /
+			(*p->Tpre_rounded - *p->Tno_bw - 2.0 * s->Tr0_trips_rounded);
+			s->Tsw_est2 = *p->prefetch_sw_bytes / s->prefetch_bw2;
+		} else
+			s->prefetch_bw2 = 0;
+
+		DML_LOG_VERBOSE("DML::%s: prefetch_bw2: %f\n", __func__, s->prefetch_bw2);
+		if ((s->Tsw_est2 < s->min_Lsw_equ * s->LineTime) && ((*p->Tpre_rounded - *p->Tno_bw - 2.0 * s->Tr0_trips_rounded - s->min_Lsw_equ * s->LineTime - 0.25 * s->LineTime) > 0)) {
+			s->prefetch_bw2 = vm_bytes * p->HostVMInefficiencyFactor / (*p->Tpre_rounded - *p->Tno_bw - 2.0 * s->Tr0_trips_rounded - s->min_Lsw_equ * s->LineTime - 0.25 * s->LineTime);
+			DML_LOG_VERBOSE("DML::%s: prefetch_bw2: %f (updated)\n", __func__, s->prefetch_bw2);
+		}
+
+		// prefetch_bw3: 2*R0 + SW
+		if (*p->Tpre_rounded - s->Tvm_trips_rounded > 0) {
+			s->prefetch_bw3 = (2 * (p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor + p->meta_row_bytes + tdlut_row_bytes) + *p->prefetch_sw_bytes) /
+				(*p->Tpre_rounded - s->Tvm_trips_rounded);
+			s->Tsw_est3 = *p->prefetch_sw_bytes / s->prefetch_bw3;
+		} else
+			s->prefetch_bw3 = 0;
+
+		DML_LOG_VERBOSE("DML::%s: prefetch_bw3: %f\n", __func__, s->prefetch_bw3);
+		if ((s->Tsw_est3 < s->min_Lsw_equ * s->LineTime) && ((*p->Tpre_rounded - s->min_Lsw_equ * s->LineTime - 0.5 * s->LineTime - s->Tvm_trips_rounded) > 0)) {
+			s->prefetch_bw3 = (2 * (p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor + p->meta_row_bytes + tdlut_row_bytes)) / (*p->Tpre_rounded - s->min_Lsw_equ * s->LineTime - 0.5 * s->LineTime - s->Tvm_trips_rounded);
+			DML_LOG_VERBOSE("DML::%s: prefetch_bw3: %f (updated)\n", __func__, s->prefetch_bw3);
+		}
+
+		// prefetch_bw4: SW
+		if (*p->Tpre_rounded - s->Tvm_trips_rounded - 2 * s->Tr0_trips_rounded > 0)
+			s->prefetch_bw4 = *p->prefetch_sw_bytes / (*p->Tpre_rounded - s->Tvm_trips_rounded - 2 * s->Tr0_trips_rounded);
+		else
+			s->prefetch_bw4 = 0;
+
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: Tno_bw: %f\n", __func__, *p->Tno_bw);
+		DML_LOG_VERBOSE("DML::%s: Tpre=%f Tpre_rounded: %f, delta=%f\n", __func__, s->dst_y_prefetch_equ * s->LineTime, *p->Tpre_rounded, (*p->Tpre_rounded - (s->dst_y_prefetch_equ * s->LineTime)));
+		DML_LOG_VERBOSE("DML::%s: Tvm_trips=%f Tvm_trips_rounded: %f, delta=%f\n", __func__, *p->Tvm_trips, s->Tvm_trips_rounded, (s->Tvm_trips_rounded - *p->Tvm_trips));
+		DML_LOG_VERBOSE("DML::%s: Tr0_trips=%f Tr0_trips_rounded: %f, delta=%f\n", __func__, *p->Tr0_trips, s->Tr0_trips_rounded, (s->Tr0_trips_rounded - *p->Tr0_trips));
+		DML_LOG_VERBOSE("DML::%s: Tsw_est1: %f\n", __func__, s->Tsw_est1);
+		DML_LOG_VERBOSE("DML::%s: Tsw_est2: %f\n", __func__, s->Tsw_est2);
+		DML_LOG_VERBOSE("DML::%s: Tsw_est3: %f\n", __func__, s->Tsw_est3);
+		DML_LOG_VERBOSE("DML::%s: prefetch_bw1: %f (final)\n", __func__, s->prefetch_bw1);
+		DML_LOG_VERBOSE("DML::%s: prefetch_bw2: %f (final)\n", __func__, s->prefetch_bw2);
+		DML_LOG_VERBOSE("DML::%s: prefetch_bw3: %f (final)\n", __func__, s->prefetch_bw3);
+		DML_LOG_VERBOSE("DML::%s: prefetch_bw4: %f (final)\n", __func__, s->prefetch_bw4);
+#endif
+		{
+			bool Case1OK = false;
+			bool Case2OK = false;
+			bool Case3OK = false;
+
+			// get "equalized" bw among all stages (vm, r0, sw), so based is all 3 stages are just above the latency-based requirement
+			// so it is not too dis-portionally favor a particular stage, next is either r0 more agressive and next is vm more agressive, the worst is all are agressive
+			// vs the latency based number
+
+			// prefetch_bw1: VM + 2*R0 + SW
+			// so prefetch_bw1 will have enough bw to transfer the necessary data within Tpre_rounded - Tno_bw (Tpre is the the worst-case latency based time to fetch the data)
+			// here is to make sure equ bw wont be more agressive than the latency-based requirement.
+			// check vm time >= vm_trips
+			// check r0 time >= r0_trips
+
+			double total_row_bytes = (p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor + p->meta_row_bytes + tdlut_row_bytes);
+
+			DML_LOG_VERBOSE("DML::%s: Tvm_trips_rounded = %f\n", __func__, s->Tvm_trips_rounded);
+			DML_LOG_VERBOSE("DML::%s: Tr0_trips_rounded = %f\n", __func__, s->Tr0_trips_rounded);
+
+			if (s->prefetch_bw1 > 0) {
+				double vm_transfer_time = *p->Tno_bw + vm_bytes * p->HostVMInefficiencyFactor / s->prefetch_bw1;
+				double row_transfer_time = total_row_bytes / s->prefetch_bw1;
+				DML_LOG_VERBOSE("DML::%s: Case1: vm_transfer_time  = %f\n", __func__, vm_transfer_time);
+				DML_LOG_VERBOSE("DML::%s: Case1: row_transfer_time = %f\n", __func__, row_transfer_time);
+				if (vm_transfer_time >= s->Tvm_trips_rounded && row_transfer_time >= s->Tr0_trips_rounded) {
+					Case1OK = true;
+				}
+			}
+
+			// prefetch_bw2: VM + SW
+			// prefetch_bw2 will be enough bw to transfer VM and SW data within (Tpre_rounded - Tr0_trips_rounded - Tno_bw)
+			// check vm time >= vm_trips
+			// check r0 time < r0_trips
+			if (s->prefetch_bw2 > 0) {
+				double vm_transfer_time = *p->Tno_bw + vm_bytes * p->HostVMInefficiencyFactor / s->prefetch_bw2;
+				double row_transfer_time = total_row_bytes / s->prefetch_bw2;
+				DML_LOG_VERBOSE("DML::%s: Case2: vm_transfer_time  = %f\n", __func__, vm_transfer_time);
+				DML_LOG_VERBOSE("DML::%s: Case2: row_transfer_time = %f\n", __func__, row_transfer_time);
+				if (vm_transfer_time >= s->Tvm_trips_rounded && row_transfer_time < s->Tr0_trips_rounded) {
+					Case2OK = true;
+				}
+			}
+
+			// prefetch_bw3: VM + 2*R0
+			// check vm time < vm_trips
+			// check r0 time >= r0_trips
+			if (s->prefetch_bw3 > 0) {
+				double vm_transfer_time = *p->Tno_bw + vm_bytes * p->HostVMInefficiencyFactor / s->prefetch_bw3;
+				double row_transfer_time = total_row_bytes / s->prefetch_bw3;
+				DML_LOG_VERBOSE("DML::%s: Case3: vm_transfer_time  = %f\n", __func__, vm_transfer_time);
+				DML_LOG_VERBOSE("DML::%s: Case3: row_transfer_time = %f\n", __func__, row_transfer_time);
+				if (vm_transfer_time < s->Tvm_trips_rounded && row_transfer_time >= s->Tr0_trips_rounded) {
+					Case3OK = true;
+				}
+			}
+
+			if (Case1OK) {
+				s->prefetch_bw_equ = s->prefetch_bw1;
+			} else if (Case2OK) {
+				s->prefetch_bw_equ = s->prefetch_bw2;
+			} else if (Case3OK) {
+				s->prefetch_bw_equ = s->prefetch_bw3;
+			} else {
+				s->prefetch_bw_equ = s->prefetch_bw4;
+			}
+
+			s->prefetch_bw_equ = math_max3(s->prefetch_bw_equ,
+							p->vm_bytes * p->HostVMInefficiencyFactor / (31 * s->LineTime) - *p->Tno_bw,
+							(p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor + p->meta_row_bytes + tdlut_row_bytes) / (15 * s->LineTime));
+#ifdef __DML_VBA_DEBUG__
+			DML_LOG_VERBOSE("DML::%s: Case1OK: %u\n", __func__, Case1OK);
+			DML_LOG_VERBOSE("DML::%s: Case2OK: %u\n", __func__, Case2OK);
+			DML_LOG_VERBOSE("DML::%s: Case3OK: %u\n", __func__, Case3OK);
+			DML_LOG_VERBOSE("DML::%s: prefetch_bw_equ: %f\n", __func__, s->prefetch_bw_equ);
+#endif
+
+			if (s->prefetch_bw_equ > 0) {
+				if (p->display_cfg->gpuvm_enable == true) {
+					s->Tvm_equ = math_max3(*p->Tno_bw + vm_bytes * p->HostVMInefficiencyFactor / s->prefetch_bw_equ, *p->Tvm_trips, s->LineTime / 4);
+				} else {
+					s->Tvm_equ = s->LineTime / 4;
+				}
+
+				if (p->display_cfg->gpuvm_enable == true || dcc_mrq_enable || p->setup_for_tdlut) {
+					s->Tr0_equ = math_max3((p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor + p->meta_row_bytes + tdlut_row_bytes) / s->prefetch_bw_equ, // PixelPTEBytesPerRow is dpte_row_bytes
+						*p->Tr0_trips,
+						s->LineTime / 4);
+				} else {
+					s->Tr0_equ = s->LineTime / 4;
+				}
+			} else {
+				s->Tvm_equ = 0;
+				s->Tr0_equ = 0;
+				DML_LOG_VERBOSE("DML::%s: prefetch_bw_equ equals 0!\n", __func__);
+			}
+		}
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: Tvm_equ = %f\n", __func__, s->Tvm_equ);
+		DML_LOG_VERBOSE("DML::%s: Tr0_equ = %f\n", __func__, s->Tr0_equ);
+#endif
+		// Use the more stressful prefetch schedule
+		if (s->dst_y_prefetch_oto < s->dst_y_prefetch_equ) {
+			*p->dst_y_prefetch = s->dst_y_prefetch_oto;
+			s->TimeForFetchingVM = s->Tvm_oto;
+			s->TimeForFetchingRowInVBlank = s->Tr0_oto;
+
+			*p->dst_y_per_vm_vblank = math_ceil2(4.0 * s->TimeForFetchingVM / s->LineTime, 1.0) / 4.0;
+			*p->dst_y_per_row_vblank = math_ceil2(4.0 * s->TimeForFetchingRowInVBlank / s->LineTime, 1.0) / 4.0;
+#ifdef __DML_VBA_DEBUG__
+			DML_LOG_VERBOSE("DML::%s: Using oto scheduling for prefetch\n", __func__);
+#endif
+
+		} else {
+			*p->dst_y_prefetch = s->dst_y_prefetch_equ;
+
+			if (s->dst_y_prefetch_equ < s->dst_y_prefetch_equ_impacted)
+				*p->dst_y_prefetch = s->dst_y_prefetch_equ_impacted;
+
+			s->TimeForFetchingVM = s->Tvm_equ;
+			s->TimeForFetchingRowInVBlank = s->Tr0_equ;
+
+			*p->dst_y_per_vm_vblank = math_ceil2(4.0 * s->TimeForFetchingVM / s->LineTime, 1.0) / 4.0;
+			*p->dst_y_per_row_vblank = math_ceil2(4.0 * s->TimeForFetchingRowInVBlank / s->LineTime, 1.0) / 4.0;
+
+			/* equ bw should be propagated so a ceiling of the equ bw is accounted for prior to mode programming.
+			 * Overall bandwidth may be lower when going from mode support to mode programming but final pixel data
+			 * bandwidth may end up higher than what was calculated in mode support.
+			 */
+			*p->RequiredPrefetchBWMax = math_max2(s->prefetch_bw_equ, *p->RequiredPrefetchBWMax);
+
+#ifdef __DML_VBA_DEBUG__
+			DML_LOG_VERBOSE("DML::%s: Using equ bw scheduling for prefetch\n", __func__);
+#endif
+		}
+
+		// Lsw = dst_y_prefetch - (dst_y_per_vm_vblank + 2*dst_y_per_row_vblank)
+		s->LinesToRequestPrefetchPixelData = *p->dst_y_prefetch - *p->dst_y_per_vm_vblank - 2 * *p->dst_y_per_row_vblank; // Lsw
+
+		s->cursor_prefetch_bytes = (unsigned int)math_max2(p->cursor_bytes_per_chunk, 4 * p->cursor_bytes_per_line);
+		*p->prefetch_cursor_bw = p->num_cursors * s->cursor_prefetch_bytes / (s->LinesToRequestPrefetchPixelData * s->LineTime);
+		*p->prefetch_swath_time_us = (s->LinesToRequestPrefetchPixelData * s->LineTime);
+
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: TimeForFetchingVM = %f\n", __func__, s->TimeForFetchingVM);
+		DML_LOG_VERBOSE("DML::%s: TimeForFetchingRowInVBlank = %f\n", __func__, s->TimeForFetchingRowInVBlank);
+		DML_LOG_VERBOSE("DML::%s: LineTime = %f\n", __func__, s->LineTime);
+		DML_LOG_VERBOSE("DML::%s: dst_y_prefetch = %f\n", __func__, *p->dst_y_prefetch);
+		DML_LOG_VERBOSE("DML::%s: dst_y_per_vm_vblank = %f\n", __func__, *p->dst_y_per_vm_vblank);
+		DML_LOG_VERBOSE("DML::%s: dst_y_per_row_vblank = %f\n", __func__, *p->dst_y_per_row_vblank);
+		DML_LOG_VERBOSE("DML::%s: LinesToRequestPrefetchPixelData = %f\n", __func__, s->LinesToRequestPrefetchPixelData);
+		DML_LOG_VERBOSE("DML::%s: PrefetchSourceLinesY = %f\n", __func__, p->PrefetchSourceLinesY);
+		DML_LOG_VERBOSE("DML::%s: prefetch_swath_time_us = %f\n", __func__, *p->prefetch_swath_time_us);
+
+		DML_LOG_VERBOSE("DML::%s: cursor_bytes_per_chunk = %d\n", __func__, p->cursor_bytes_per_chunk);
+		DML_LOG_VERBOSE("DML::%s: cursor_bytes_per_line = %d\n", __func__, p->cursor_bytes_per_line);
+		DML_LOG_VERBOSE("DML::%s: cursor_prefetch_bytes = %d\n", __func__, s->cursor_prefetch_bytes);
+		DML_LOG_VERBOSE("DML::%s: prefetch_cursor_bw = %f\n", __func__, *p->prefetch_cursor_bw);
+#endif
+		DML_ASSERT(*p->dst_y_prefetch < 64);
+
+		unsigned int min_lsw_required = (unsigned int)math_max2(2, p->tdlut_drain_time / s->LineTime);
+		if (s->LinesToRequestPrefetchPixelData >= min_lsw_required && s->prefetch_bw_equ > 0) {
+			*p->VRatioPrefetchY = (double)p->PrefetchSourceLinesY / s->LinesToRequestPrefetchPixelData;
+			*p->VRatioPrefetchY = math_max2(*p->VRatioPrefetchY, 1.0);
+#ifdef __DML_VBA_DEBUG__
+			DML_LOG_VERBOSE("DML::%s: VRatioPrefetchY = %f\n", __func__, *p->VRatioPrefetchY);
+			DML_LOG_VERBOSE("DML::%s: SwathHeightY = %u\n", __func__, p->SwathHeightY);
+			DML_LOG_VERBOSE("DML::%s: VInitPreFillY = %u\n", __func__, p->VInitPreFillY);
+#endif
+			if ((p->SwathHeightY > 4) && (p->VInitPreFillY > 3)) {
+				if (s->LinesToRequestPrefetchPixelData > (p->VInitPreFillY - 3.0) / 2.0) {
+					*p->VRatioPrefetchY = math_max2(*p->VRatioPrefetchY,
+						(double)p->MaxNumSwathY * p->SwathHeightY / (s->LinesToRequestPrefetchPixelData - (p->VInitPreFillY - 3.0) / 2.0));
+				} else {
+					s->NoTimeToPrefetch = true;
+					DML_LOG_VERBOSE("DML::%s: No time to prefetch!. LinesToRequestPrefetchPixelData=%f VinitPreFillY=%u\n", __func__, s->LinesToRequestPrefetchPixelData, p->VInitPreFillY);
+					*p->VRatioPrefetchY = 0;
+				}
+#ifdef __DML_VBA_DEBUG__
+				DML_LOG_VERBOSE("DML::%s: VRatioPrefetchY = %f\n", __func__, *p->VRatioPrefetchY);
+				DML_LOG_VERBOSE("DML::%s: PrefetchSourceLinesY = %f\n", __func__, p->PrefetchSourceLinesY);
+				DML_LOG_VERBOSE("DML::%s: MaxNumSwathY = %u\n", __func__, p->MaxNumSwathY);
+#endif
+			}
+
+			*p->VRatioPrefetchC = (double)p->PrefetchSourceLinesC / s->LinesToRequestPrefetchPixelData;
+			*p->VRatioPrefetchC = math_max2(*p->VRatioPrefetchC, 1.0);
+
+#ifdef __DML_VBA_DEBUG__
+			DML_LOG_VERBOSE("DML::%s: VRatioPrefetchC = %f\n", __func__, *p->VRatioPrefetchC);
+			DML_LOG_VERBOSE("DML::%s: SwathHeightC = %u\n", __func__, p->SwathHeightC);
+			DML_LOG_VERBOSE("DML::%s: VInitPreFillC = %u\n", __func__, p->VInitPreFillC);
+#endif
+			if ((p->SwathHeightC > 4) && (p->VInitPreFillC > 3)) {
+				if (s->LinesToRequestPrefetchPixelData > (p->VInitPreFillC - 3.0) / 2.0) {
+					*p->VRatioPrefetchC = math_max2(*p->VRatioPrefetchC, (double)p->MaxNumSwathC * p->SwathHeightC / (s->LinesToRequestPrefetchPixelData - (p->VInitPreFillC - 3.0) / 2.0));
+				} else {
+					s->NoTimeToPrefetch = true;
+					DML_LOG_VERBOSE("DML::%s: No time to prefetch!. LinesToRequestPrefetchPixelData=%f VInitPreFillC=%u\n", __func__, s->LinesToRequestPrefetchPixelData, p->VInitPreFillC);
+					*p->VRatioPrefetchC = 0;
+				}
+#ifdef __DML_VBA_DEBUG__
+				DML_LOG_VERBOSE("DML::%s: VRatioPrefetchC = %f\n", __func__, *p->VRatioPrefetchC);
+				DML_LOG_VERBOSE("DML::%s: PrefetchSourceLinesC = %f\n", __func__, p->PrefetchSourceLinesC);
+				DML_LOG_VERBOSE("DML::%s: MaxNumSwathC = %u\n", __func__, p->MaxNumSwathC);
+#endif
+			}
+
+			*p->RequiredPrefetchPixelDataBWLuma = (double)p->PrefetchSourceLinesY / s->LinesToRequestPrefetchPixelData * p->myPipe->BytePerPixelY * p->swath_width_luma_ub / s->LineTime;
+			*p->RequiredPrefetchPixelDataBWChroma = (double)p->PrefetchSourceLinesC / s->LinesToRequestPrefetchPixelData * p->myPipe->BytePerPixelC * p->swath_width_chroma_ub / s->LineTime;
+
+#ifdef __DML_VBA_DEBUG__
+			DML_LOG_VERBOSE("DML::%s: BytePerPixelY = %u\n", __func__, p->myPipe->BytePerPixelY);
+			DML_LOG_VERBOSE("DML::%s: swath_width_luma_ub = %u\n", __func__, p->swath_width_luma_ub);
+			DML_LOG_VERBOSE("DML::%s: LineTime = %f\n", __func__, s->LineTime);
+			DML_LOG_VERBOSE("DML::%s: RequiredPrefetchPixelDataBWLuma = %f\n", __func__, *p->RequiredPrefetchPixelDataBWLuma);
+			DML_LOG_VERBOSE("DML::%s: RequiredPrefetchPixelDataBWChroma = %f\n", __func__, *p->RequiredPrefetchPixelDataBWChroma);
+#endif
+		} else {
+			s->NoTimeToPrefetch = true;
+			DML_LOG_VERBOSE("DML::%s: No time to prefetch!, LinesToRequestPrefetchPixelData: %f, should be >= %d\n", __func__, s->LinesToRequestPrefetchPixelData, min_lsw_required);
+			DML_LOG_VERBOSE("DML::%s: No time to prefetch!, prefetch_bw_equ: %f, should be > 0\n", __func__, s->prefetch_bw_equ);
+			*p->VRatioPrefetchY = 0;
+			*p->VRatioPrefetchC = 0;
+			*p->RequiredPrefetchPixelDataBWLuma = 0;
+			*p->RequiredPrefetchPixelDataBWChroma = 0;
+		}
+		DML_LOG_VERBOSE("DML: Tpre: %fus - sum of time to request 2 x data pte, swaths\n", (double)s->LinesToRequestPrefetchPixelData * s->LineTime + 2.0 * s->TimeForFetchingRowInVBlank + s->TimeForFetchingVM);
+		DML_LOG_VERBOSE("DML: Tvm: %fus - time to fetch vm\n", s->TimeForFetchingVM);
+		DML_LOG_VERBOSE("DML: Tr0: %fus - time to fetch first row of data pagetables\n", s->TimeForFetchingRowInVBlank);
+		DML_LOG_VERBOSE("DML: Tsw: %fus = time to fetch enough pixel data and cursor data to feed the scalers init position and detile\n", (double)s->LinesToRequestPrefetchPixelData * s->LineTime);
+		DML_LOG_VERBOSE("DML: To: %fus - time for propagation from scaler to optc\n", (*p->DSTYAfterScaler + ((double)(*p->DSTXAfterScaler) / (double)p->myPipe->HTotal)) * s->LineTime);
+		DML_LOG_VERBOSE("DML: Tvstartup - TSetup - Tcalc - TWait - Tpre - To > 0\n");
+		DML_LOG_VERBOSE("DML: Tslack(pre): %fus - time left over in schedule\n", p->VStartup * s->LineTime - s->TimeForFetchingVM - 2 * s->TimeForFetchingRowInVBlank - (*p->DSTYAfterScaler + ((double)(*p->DSTXAfterScaler) / (double)p->myPipe->HTotal)) * s->LineTime - p->TWait - p->TCalc - *p->TSetup);
+		DML_LOG_VERBOSE("DML: row_bytes = dpte_row_bytes (per_pipe) = PixelPTEBytesPerRow = : %u\n", p->PixelPTEBytesPerRow);
+
+	} else {
+		DML_LOG_VERBOSE("DML::%s: No time to prefetch! dst_y_prefetch_equ = %f (should be > 1)\n", __func__, s->dst_y_prefetch_equ);
+		DML_LOG_VERBOSE("DML::%s: No time to prefetch! min_Lsw_equ_ok = %d, Tpre_rounded (%f) should be >= Tvm_trips_rounded (%f) + 2.0*Tr0_trips_rounded (%f) + min_Tsw_equ (%f)\n",
+				__func__, min_Lsw_equ_ok, *p->Tpre_rounded, s->Tvm_trips_rounded, 2.0*s->Tr0_trips_rounded, s->min_Lsw_equ*s->LineTime);
+		s->NoTimeToPrefetch = true;
+		s->TimeForFetchingVM = 0;
+		s->TimeForFetchingRowInVBlank = 0;
+		*p->dst_y_per_vm_vblank = 0;
+		*p->dst_y_per_row_vblank = 0;
+		s->LinesToRequestPrefetchPixelData = 0;
+		*p->VRatioPrefetchY = 0;
+		*p->VRatioPrefetchC = 0;
+		*p->RequiredPrefetchPixelDataBWLuma = 0;
+		*p->RequiredPrefetchPixelDataBWChroma = 0;
+	}
+
+	{
+		double prefetch_vm_bw;
+		double prefetch_row_bw;
+
+		if (vm_bytes == 0) {
+			prefetch_vm_bw = 0;
+		} else if (*p->dst_y_per_vm_vblank > 0) {
+#ifdef __DML_VBA_DEBUG__
+			DML_LOG_VERBOSE("DML::%s: HostVMInefficiencyFactor = %f\n", __func__, p->HostVMInefficiencyFactor);
+			DML_LOG_VERBOSE("DML::%s: dst_y_per_vm_vblank = %f\n", __func__, *p->dst_y_per_vm_vblank);
+			DML_LOG_VERBOSE("DML::%s: LineTime = %f\n", __func__, s->LineTime);
+#endif
+			prefetch_vm_bw = vm_bytes * p->HostVMInefficiencyFactor / (*p->dst_y_per_vm_vblank * s->LineTime);
+#ifdef __DML_VBA_DEBUG__
+			DML_LOG_VERBOSE("DML::%s: prefetch_vm_bw = %f\n", __func__, prefetch_vm_bw);
+#endif
+		} else {
+			prefetch_vm_bw = 0;
+			s->NoTimeToPrefetch = true;
+			DML_LOG_VERBOSE("DML::%s: No time to prefetch!. dst_y_per_vm_vblank=%f (should be > 0)\n", __func__, *p->dst_y_per_vm_vblank);
+		}
+
+		if (p->PixelPTEBytesPerRow == 0 && tdlut_row_bytes == 0) {
+			prefetch_row_bw = 0;
+		} else if (*p->dst_y_per_row_vblank > 0) {
+			prefetch_row_bw = (p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor + tdlut_row_bytes) / (*p->dst_y_per_row_vblank * s->LineTime);
+
+#ifdef __DML_VBA_DEBUG__
+			DML_LOG_VERBOSE("DML::%s: PixelPTEBytesPerRow = %u\n", __func__, p->PixelPTEBytesPerRow);
+			DML_LOG_VERBOSE("DML::%s: dst_y_per_row_vblank = %f\n", __func__, *p->dst_y_per_row_vblank);
+			DML_LOG_VERBOSE("DML::%s: prefetch_row_bw = %f\n", __func__, prefetch_row_bw);
+#endif
+		} else {
+			prefetch_row_bw = 0;
+			s->NoTimeToPrefetch = true;
+			DML_LOG_VERBOSE("DML::%s: No time to prefetch!. dst_y_per_row_vblank=%f (should be > 0)\n", __func__, *p->dst_y_per_row_vblank);
+		}
+
+		*p->prefetch_vmrow_bw = math_max2(prefetch_vm_bw, prefetch_row_bw);
+	}
+
+	if (s->NoTimeToPrefetch) {
+		s->TimeForFetchingVM = 0;
+		s->TimeForFetchingRowInVBlank = 0;
+		*p->dst_y_per_vm_vblank = 0;
+		*p->dst_y_per_row_vblank = 0;
+		*p->dst_y_prefetch = 0;
+		s->LinesToRequestPrefetchPixelData = 0;
+		*p->VRatioPrefetchY = 0;
+		*p->VRatioPrefetchC = 0;
+		*p->RequiredPrefetchPixelDataBWLuma = 0;
+		*p->RequiredPrefetchPixelDataBWChroma = 0;
+		*p->prefetch_vmrow_bw = 0;
+	}
+
+	DML_LOG_VERBOSE("DML::%s: dst_y_per_vm_vblank = %f (final)\n", __func__, *p->dst_y_per_vm_vblank);
+	DML_LOG_VERBOSE("DML::%s: dst_y_per_row_vblank = %f (final)\n", __func__, *p->dst_y_per_row_vblank);
+	DML_LOG_VERBOSE("DML::%s: prefetch_vmrow_bw = %f (final)\n", __func__, *p->prefetch_vmrow_bw);
+	DML_LOG_VERBOSE("DML::%s: RequiredPrefetchPixelDataBWLuma = %f (final)\n", __func__, *p->RequiredPrefetchPixelDataBWLuma);
+	DML_LOG_VERBOSE("DML::%s: RequiredPrefetchPixelDataBWChroma = %f (final)\n", __func__, *p->RequiredPrefetchPixelDataBWChroma);
+	DML_LOG_VERBOSE("DML::%s: NoTimeToPrefetch=%d\n", __func__, s->NoTimeToPrefetch);
+
+	return s->NoTimeToPrefetch;
+}
+
+static unsigned int get_num_lb_source_lines(unsigned int max_line_buffer_lines,
+									unsigned int line_buffer_size_bits,
+									unsigned int num_pipes,
+									unsigned int vp_width,
+									unsigned int vp_height,
+									double h_ratio,
+									enum dml2_rotation_angle rotation_angle)
+{
+	unsigned int num_lb_source_lines = 0;
+	double lb_bit_per_pixel = 57.0;
+	unsigned recin_width = vp_width/num_pipes;
+
+	if (dml_is_vertical_rotation(rotation_angle))
+		recin_width = vp_height/num_pipes;
+
+	num_lb_source_lines = (unsigned int) math_min2((double) max_line_buffer_lines,
+								math_floor2(line_buffer_size_bits / lb_bit_per_pixel / (recin_width / math_max2(h_ratio, 1.0)), 1.0));
+
+	return num_lb_source_lines;
+}
+
+static unsigned int find_max_impact_plane(unsigned int this_plane_idx, unsigned int num_planes, unsigned int Trpd_dcfclk_cycles[])
+{
+	int max_value = -1;
+	int max_idx = -1;
+	for (unsigned int i = 0; i < num_planes; i++) {
+		if (i != this_plane_idx && (int) Trpd_dcfclk_cycles[i] > max_value) {
+			max_value = Trpd_dcfclk_cycles[i];
+			max_idx = i;
+		}
+	}
+	if (max_idx <= 0) {
+		DML_ASSERT(max_idx >= 0);
+		max_idx = this_plane_idx;
+	}
+
+	return max_idx;
+}
+
+static double calculate_impacted_Tsw(unsigned int exclude_plane_idx, unsigned int num_planes, double *prefetch_swath_bytes, double bw_mbps)
+{
+	double sum = 0.;
+	for (unsigned int i = 0; i < num_planes; i++) {
+		if (i != exclude_plane_idx) {
+			sum += prefetch_swath_bytes[i];
+		}
+	}
+	return sum / bw_mbps;
+}
+
+// a global check against the aggregate effect of the per plane prefetch schedule
+static noinline_for_stack bool CheckGlobalPrefetchAdmissibility(struct dml2_core_internal_scratch *scratch,
+											 struct dml2_core_calcs_CheckGlobalPrefetchAdmissibility_params *p)
+{
+	struct dml2_core_calcs_CheckGlobalPrefetchAdmissibility_locals *s = &scratch->CheckGlobalPrefetchAdmissibility_locals;
+	unsigned int i, k;
+
+	memset(s, 0, sizeof(struct dml2_core_calcs_CheckGlobalPrefetchAdmissibility_locals));
+
+	*p->recalc_prefetch_schedule = 0;
+	s->prefetch_global_check_passed = 1;
+	// worst case if the rob and cdb is fully hogged
+	s->max_Trpd_dcfclk_cycles = (unsigned int) math_ceil2((p->rob_buffer_size_kbytes*1024 + p->compressed_buffer_size_kbytes*DML_MAX_COMPRESSION_RATIO*1024)/64.0, 1.0);
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: num_active_planes = %d\n", __func__, p->num_active_planes);
+	DML_LOG_VERBOSE("DML::%s: rob_buffer_size_kbytes = %d\n", __func__, p->rob_buffer_size_kbytes);
+	DML_LOG_VERBOSE("DML::%s: compressed_buffer_size_kbytes = %d\n", __func__, p->compressed_buffer_size_kbytes);
+	DML_LOG_VERBOSE("DML::%s: estimated_urg_bandwidth_required_mbps = %f\n", __func__, p->estimated_urg_bandwidth_required_mbps);
+	DML_LOG_VERBOSE("DML::%s: estimated_dcfclk_mhz = %f\n", __func__, p->estimated_dcfclk_mhz);
+	DML_LOG_VERBOSE("DML::%s: max_Trpd_dcfclk_cycles = %u\n", __func__, s->max_Trpd_dcfclk_cycles);
+#endif
+
+	// calculate the return impact from each plane, request is 256B per dcfclk
+	for (i = 0; i < p->num_active_planes; i++) {
+		s->src_detile_buf_size_bytes_l[i] = p->detile_buffer_size_bytes_l[i];
+		s->src_detile_buf_size_bytes_c[i] = p->detile_buffer_size_bytes_c[i];
+		s->src_swath_bytes_l[i] = p->full_swath_bytes_l[i];
+		s->src_swath_bytes_c[i] = p->full_swath_bytes_c[i];
+
+		if (p->pixel_format[i] == dml2_420_10) {
+			s->src_detile_buf_size_bytes_l[i] = (unsigned int) (s->src_detile_buf_size_bytes_l[i] * 1.5);
+			s->src_detile_buf_size_bytes_c[i] = (unsigned int) (s->src_detile_buf_size_bytes_c[i] * 1.5);
+			s->src_swath_bytes_l[i] = (unsigned int) (s->src_swath_bytes_l[i] * 1.5);
+			s->src_swath_bytes_c[i] = (unsigned int) (s->src_swath_bytes_c[i] * 1.5);
+		}
+
+		s->burst_bytes_to_fill_det = (unsigned int) (math_floor2(s->src_detile_buf_size_bytes_l[i] / p->chunk_bytes_l, 1) * p->chunk_bytes_l);
+		s->burst_bytes_to_fill_det += (unsigned int) (math_floor2(p->lb_source_lines_l[i] / p->swath_height_l[i], 1) * s->src_swath_bytes_l[i]);
+
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: i=%u pixel_format = %d\n", __func__, i, p->pixel_format[i]);
+		DML_LOG_VERBOSE("DML::%s: i=%u chunk_bytes_l = %d\n", __func__, i, p->chunk_bytes_l);
+		DML_LOG_VERBOSE("DML::%s: i=%u lb_source_lines_l = %d\n", __func__, i, p->lb_source_lines_l[i]);
+		DML_LOG_VERBOSE("DML::%s: i=%u src_detile_buf_size_bytes_l=%d\n", __func__, i, s->src_detile_buf_size_bytes_l[i]);
+		DML_LOG_VERBOSE("DML::%s: i=%u src_swath_bytes_l=%d\n", __func__, i, s->src_swath_bytes_l[i]);
+		DML_LOG_VERBOSE("DML::%s: i=%u burst_bytes_to_fill_det=%d (luma)\n", __func__, i, s->burst_bytes_to_fill_det);
+#endif
+
+		if (s->src_swath_bytes_c[i] > 0) { // dual_plane
+			s->burst_bytes_to_fill_det += (unsigned int) (math_floor2(s->src_detile_buf_size_bytes_c[i] / p->chunk_bytes_c, 1) * p->chunk_bytes_c);
+
+			if (p->pixel_format[i] == dml2_422_planar_8 || p->pixel_format[i] == dml2_422_planar_10 || p->pixel_format[i] == dml2_422_planar_12) {
+				s->burst_bytes_to_fill_det += (unsigned int) (math_floor2(p->lb_source_lines_c[i] / p->swath_height_c[i], 1) * s->src_swath_bytes_c[i]);
+			}
+
+#ifdef __DML_VBA_DEBUG__
+			DML_LOG_VERBOSE("DML::%s: i=%u chunk_bytes_c = %d\n", __func__, i, p->chunk_bytes_c);
+			DML_LOG_VERBOSE("DML::%s: i=%u lb_source_lines_c = %d\n", __func__, i, p->lb_source_lines_c[i]);
+			DML_LOG_VERBOSE("DML::%s: i=%u src_detile_buf_size_bytes_c=%d\n", __func__, i, s->src_detile_buf_size_bytes_c[i]);
+			DML_LOG_VERBOSE("DML::%s: i=%u src_swath_bytes_c=%d\n", __func__, i, s->src_swath_bytes_c[i]);
+#endif
+		}
+
+		s->time_to_fill_det_us = (double) s->burst_bytes_to_fill_det / (256 * p->estimated_dcfclk_mhz); // fill time assume full burst at request rate
+		s->accumulated_return_path_dcfclk_cycles[i] = (unsigned int) math_ceil2(((DML_MAX_COMPRESSION_RATIO-1) * 64 * p->estimated_dcfclk_mhz) * s->time_to_fill_det_us / 64.0, 1.0); //for 64B per DCFClk
+
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: i=%u burst_bytes_to_fill_det=%d\n", __func__, i, s->burst_bytes_to_fill_det);
+		DML_LOG_VERBOSE("DML::%s: i=%u time_to_fill_det_us=%f\n", __func__, i, s->time_to_fill_det_us);
+		DML_LOG_VERBOSE("DML::%s: i=%u accumulated_return_path_dcfclk_cycles=%u\n", __func__, i, s->accumulated_return_path_dcfclk_cycles[i]);
+#endif
+		// clamping to worst case delay which is one which occupy the full rob+cdb
+		if (s->accumulated_return_path_dcfclk_cycles[i] > s->max_Trpd_dcfclk_cycles)
+			s->accumulated_return_path_dcfclk_cycles[i] = s->max_Trpd_dcfclk_cycles;
+	}
+
+	// Figure out the impacted prefetch time for each plane
+	// if impacted_Tre is > equ bw Tpre, we need to fail the prefetch schedule as we need a higher state to support the bw
+	for (i = 0; i < p->num_active_planes; i++) {
+		k = find_max_impact_plane(i, p->num_active_planes, s->accumulated_return_path_dcfclk_cycles); // plane k causes most impact to plane i
+		// the rest of planes (except for k) complete for bw
+		p->impacted_dst_y_pre[i] = s->accumulated_return_path_dcfclk_cycles[k]/p->estimated_dcfclk_mhz;
+		p->impacted_dst_y_pre[i] += calculate_impacted_Tsw(k, p->num_active_planes, p->prefetch_sw_bytes, p->estimated_urg_bandwidth_required_mbps);
+		p->impacted_dst_y_pre[i] = math_ceil2(p->impacted_dst_y_pre[i] / p->line_time[i], 0.25);
+
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: i=%u impacted_Tpre=%f (k=%u)\n", __func__, i, p->impacted_dst_y_pre[i], k);
+#endif
+	}
+
+	if (p->Tpre_rounded != NULL && p->Tpre_oto != NULL) {
+		for (i = 0; i < p->num_active_planes; i++) {
+			if (p->impacted_dst_y_pre[i] > p->dst_y_prefetch[i]) {
+				s->prefetch_global_check_passed = 0;
+				*p->recalc_prefetch_schedule = 1;
+			}
+#ifdef __DML_VBA_DEBUG__
+			DML_LOG_VERBOSE("DML::%s: i=%u Tpre_rounded=%f\n", __func__, i, p->Tpre_rounded[i]);
+			DML_LOG_VERBOSE("DML::%s: i=%u Tpre_oto=%f\n", __func__, i, p->Tpre_oto[i]);
+#endif
+		}
+	} else {
+		// likely a mode programming calls, assume support, and no recalc - not used anyways
+		s->prefetch_global_check_passed = 1;
+		*p->recalc_prefetch_schedule = 0;
+	}
+
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: prefetch_global_check_passed=%u\n", __func__, s->prefetch_global_check_passed);
+	DML_LOG_VERBOSE("DML::%s: recalc_prefetch_schedule=%u\n", __func__, *p->recalc_prefetch_schedule);
+#endif
+
+	return s->prefetch_global_check_passed;
+}
+
+static void calculate_peak_bandwidth_required(
+	struct dml2_core_internal_scratch *s,
+	struct dml2_core_calcs_calculate_peak_bandwidth_required_params *p)
+{
+	unsigned int n;
+	unsigned int m;
+
+	struct dml2_core_shared_calculate_peak_bandwidth_required_locals *l = &s->calculate_peak_bandwidth_required_locals;
+
+	memset(l, 0, sizeof(struct dml2_core_shared_calculate_peak_bandwidth_required_locals));
+
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: inc_flip_bw = %d\n", __func__, p->inc_flip_bw);
+	DML_LOG_VERBOSE("DML::%s: NumberOfActiveSurfaces = %d\n", __func__, p->num_active_planes);
+#endif
+
+	for (unsigned int k = 0; k < p->num_active_planes; ++k) {
+		l->unity_array[k] = 1.0;
+		l->zero_array[k] = 0.0;
+	}
+
+	for (m = 0; m < dml2_core_internal_soc_state_max; m++) {
+		for (n = 0; n < dml2_core_internal_bw_max; n++) {
+			get_urgent_bandwidth_required(
+				&s->get_urgent_bandwidth_required_locals,
+				p->display_cfg,
+				m,
+				n,
+				0, //inc_flip_bw,
+				0, //use_qual_row_bw
+				p->num_active_planes,
+				p->num_of_dpp,
+				p->dcc_dram_bw_nom_overhead_factor_p0,
+				p->dcc_dram_bw_nom_overhead_factor_p1,
+				p->dcc_dram_bw_pref_overhead_factor_p0,
+				p->dcc_dram_bw_pref_overhead_factor_p1,
+				p->mall_prefetch_sdp_overhead_factor,
+				p->mall_prefetch_dram_overhead_factor,
+				p->surface_read_bandwidth_l,
+				p->surface_read_bandwidth_c,
+				l->zero_array, //PrefetchBandwidthLuma,
+				l->zero_array, //PrefetchBandwidthChroma,
+				l->zero_array, //PrefetchBWMax
+				l->zero_array,
+				l->zero_array,
+				l->zero_array,
+				p->dpte_row_bw,
+				p->meta_row_bw,
+				l->zero_array, //prefetch_cursor_bw,
+				l->zero_array, //prefetch_vmrow_bw,
+				l->zero_array, //flip_bw,
+				l->zero_array,
+				l->zero_array,
+				l->zero_array,
+				l->zero_array,
+				l->zero_array,
+				l->zero_array,
+				p->surface_avg_vactive_required_bw[m][n],
+				p->surface_peak_required_bw[m][n]);
+
+			p->urg_vactive_bandwidth_required[m][n] = get_urgent_bandwidth_required(
+				&s->get_urgent_bandwidth_required_locals,
+				p->display_cfg,
+				m,
+				n,
+				0, //inc_flip_bw,
+				0, //use_qual_row_bw
+				p->num_active_planes,
+				p->num_of_dpp,
+				p->dcc_dram_bw_nom_overhead_factor_p0,
+				p->dcc_dram_bw_nom_overhead_factor_p1,
+				p->dcc_dram_bw_pref_overhead_factor_p0,
+				p->dcc_dram_bw_pref_overhead_factor_p1,
+				p->mall_prefetch_sdp_overhead_factor,
+				p->mall_prefetch_dram_overhead_factor,
+				p->surface_read_bandwidth_l,
+				p->surface_read_bandwidth_c,
+				l->zero_array, //PrefetchBandwidthLuma,
+				l->zero_array, //PrefetchBandwidthChroma,
+				l->zero_array, //PrefetchBWMax
+				p->excess_vactive_fill_bw_l,
+				p->excess_vactive_fill_bw_c,
+				p->cursor_bw,
+				p->dpte_row_bw,
+				p->meta_row_bw,
+				l->zero_array, //prefetch_cursor_bw,
+				l->zero_array, //prefetch_vmrow_bw,
+				l->zero_array, //flip_bw,
+				p->urgent_burst_factor_l,
+				p->urgent_burst_factor_c,
+				p->urgent_burst_factor_cursor,
+				p->urgent_burst_factor_prefetch_l,
+				p->urgent_burst_factor_prefetch_c,
+				p->urgent_burst_factor_prefetch_cursor,
+				l->surface_dummy_bw,
+				p->surface_peak_required_bw[m][n]);
+
+			p->urg_bandwidth_required[m][n] = get_urgent_bandwidth_required(
+				&s->get_urgent_bandwidth_required_locals,
+				p->display_cfg,
+				m,
+				n,
+				p->inc_flip_bw,
+				0, //use_qual_row_bw
+				p->num_active_planes,
+				p->num_of_dpp,
+				p->dcc_dram_bw_nom_overhead_factor_p0,
+				p->dcc_dram_bw_nom_overhead_factor_p1,
+				p->dcc_dram_bw_pref_overhead_factor_p0,
+				p->dcc_dram_bw_pref_overhead_factor_p1,
+				p->mall_prefetch_sdp_overhead_factor,
+				p->mall_prefetch_dram_overhead_factor,
+				p->surface_read_bandwidth_l,
+				p->surface_read_bandwidth_c,
+				p->prefetch_bandwidth_l,
+				p->prefetch_bandwidth_c,
+				p->prefetch_bandwidth_max, // to prevent ms/mp mismatches where mp prefetch bw > ms prefetch bw
+				p->excess_vactive_fill_bw_l,
+				p->excess_vactive_fill_bw_c,
+				p->cursor_bw,
+				p->dpte_row_bw,
+				p->meta_row_bw,
+				p->prefetch_cursor_bw,
+				p->prefetch_vmrow_bw,
+				p->flip_bw,
+				p->urgent_burst_factor_l,
+				p->urgent_burst_factor_c,
+				p->urgent_burst_factor_cursor,
+				p->urgent_burst_factor_prefetch_l,
+				p->urgent_burst_factor_prefetch_c,
+				p->urgent_burst_factor_prefetch_cursor,
+				l->surface_dummy_bw,
+				p->surface_peak_required_bw[m][n]);
+
+			p->urg_bandwidth_required_qual[m][n] = get_urgent_bandwidth_required(
+				&s->get_urgent_bandwidth_required_locals,
+				p->display_cfg,
+				m,
+				n,
+				0, //inc_flip_bw
+				1, //use_qual_row_bw
+				p->num_active_planes,
+				p->num_of_dpp,
+				p->dcc_dram_bw_nom_overhead_factor_p0,
+				p->dcc_dram_bw_nom_overhead_factor_p1,
+				p->dcc_dram_bw_pref_overhead_factor_p0,
+				p->dcc_dram_bw_pref_overhead_factor_p1,
+				p->mall_prefetch_sdp_overhead_factor,
+				p->mall_prefetch_dram_overhead_factor,
+				p->surface_read_bandwidth_l,
+				p->surface_read_bandwidth_c,
+				p->prefetch_bandwidth_l,
+				p->prefetch_bandwidth_c,
+				p->prefetch_bandwidth_max, // to prevent ms/mp mismatch where mp prefetch bw > ms prefetch bw
+				p->excess_vactive_fill_bw_l,
+				p->excess_vactive_fill_bw_c,
+				p->cursor_bw,
+				p->dpte_row_bw,
+				p->meta_row_bw,
+				p->prefetch_cursor_bw,
+				p->prefetch_vmrow_bw,
+				p->flip_bw,
+				p->urgent_burst_factor_l,
+				p->urgent_burst_factor_c,
+				p->urgent_burst_factor_cursor,
+				p->urgent_burst_factor_prefetch_l,
+				p->urgent_burst_factor_prefetch_c,
+				p->urgent_burst_factor_prefetch_cursor,
+				l->surface_dummy_bw,
+				p->surface_peak_required_bw[m][n]);
+
+			p->non_urg_bandwidth_required[m][n] = get_urgent_bandwidth_required(
+				&s->get_urgent_bandwidth_required_locals,
+				p->display_cfg,
+				m,
+				n,
+				p->inc_flip_bw,
+				0, //use_qual_row_bw
+				p->num_active_planes,
+				p->num_of_dpp,
+				p->dcc_dram_bw_nom_overhead_factor_p0,
+				p->dcc_dram_bw_nom_overhead_factor_p1,
+				p->dcc_dram_bw_pref_overhead_factor_p0,
+				p->dcc_dram_bw_pref_overhead_factor_p1,
+				p->mall_prefetch_sdp_overhead_factor,
+				p->mall_prefetch_dram_overhead_factor,
+				p->surface_read_bandwidth_l,
+				p->surface_read_bandwidth_c,
+				p->prefetch_bandwidth_l,
+				p->prefetch_bandwidth_c,
+				p->prefetch_bandwidth_max, // to prevent ms/mp mismatches where mp prefetch bw > ms prefetch bw
+				p->excess_vactive_fill_bw_l,
+				p->excess_vactive_fill_bw_c,
+				p->cursor_bw,
+				p->dpte_row_bw,
+				p->meta_row_bw,
+				p->prefetch_cursor_bw,
+				p->prefetch_vmrow_bw,
+				p->flip_bw,
+				l->unity_array,
+				l->unity_array,
+				l->unity_array,
+				l->unity_array,
+				l->unity_array,
+				l->unity_array,
+				l->surface_dummy_bw,
+				p->surface_peak_required_bw[m][n]);
+
+#ifdef __DML_VBA_DEBUG__
+			DML_LOG_VERBOSE("DML::%s: urg_vactive_bandwidth_required%s[%s][%s]=%f\n", __func__, (p->inc_flip_bw ? "_flip" : ""), dml2_core_internal_soc_state_type_str(m), dml2_core_internal_bw_type_str(n), p->urg_vactive_bandwidth_required[m][n]);
+			DML_LOG_VERBOSE("DML::%s: urg_bandwidth_required%s[%s][%s]=%f\n", __func__, (p->inc_flip_bw ? "_flip" : ""), dml2_core_internal_soc_state_type_str(m), dml2_core_internal_bw_type_str(n), p->urg_bandwidth_required[m][n]);
+			DML_LOG_VERBOSE("DML::%s: urg_bandwidth_required_qual[%s][%s]=%f\n", __func__, dml2_core_internal_soc_state_type_str(m), dml2_core_internal_bw_type_str(n), p->urg_bandwidth_required[m][n]);
+			DML_LOG_VERBOSE("DML::%s: non_urg_bandwidth_required%s[%s][%s]=%f\n", __func__, (p->inc_flip_bw ? "_flip" : ""), dml2_core_internal_soc_state_type_str(m), dml2_core_internal_bw_type_str(n), p->non_urg_bandwidth_required[m][n]);
+#endif
+			DML_ASSERT(p->urg_bandwidth_required[m][n] >= p->non_urg_bandwidth_required[m][n]);
+		}
+	}
+}
+
+static void check_urgent_bandwidth_support(
+	double *frac_urg_bandwidth_nom,
+	double *frac_urg_bandwidth_mall,
+	bool *vactive_bandwidth_support_ok, // vactive ok
+	bool *bandwidth_support_ok,// max of vm, prefetch, vactive all ok
+
+	unsigned int mall_allocated_for_dcn_mbytes,
+	double non_urg_bandwidth_required[dml2_core_internal_soc_state_max][dml2_core_internal_bw_max],
+	double urg_vactive_bandwidth_required[dml2_core_internal_soc_state_max][dml2_core_internal_bw_max],
+	double urg_bandwidth_required[dml2_core_internal_soc_state_max][dml2_core_internal_bw_max],
+	double urg_bandwidth_available[dml2_core_internal_soc_state_max][dml2_core_internal_bw_max])
+{
+	double frac_urg_bandwidth_nom_sdp = non_urg_bandwidth_required[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_sdp] / urg_bandwidth_available[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_sdp];
+	double frac_urg_bandwidth_nom_dram = non_urg_bandwidth_required[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_dram] / urg_bandwidth_available[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_dram];
+	double frac_urg_bandwidth_mall_sdp;
+	double frac_urg_bandwidth_mall_dram;
+	if (urg_bandwidth_available[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_sdp] > 0)
+		frac_urg_bandwidth_mall_sdp = non_urg_bandwidth_required[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_sdp] / urg_bandwidth_available[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_sdp];
+	else
+		frac_urg_bandwidth_mall_sdp = 0.0;
+	if (urg_bandwidth_available[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_dram] > 0)
+		frac_urg_bandwidth_mall_dram = non_urg_bandwidth_required[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_dram] / urg_bandwidth_available[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_dram];
+	else
+		frac_urg_bandwidth_mall_dram = 0.0;
+
+	*bandwidth_support_ok = 1;
+	*vactive_bandwidth_support_ok = 1;
+
+	// Check urgent bandwidth required at sdp vs urgent bandwidth avail at sdp -> FractionOfUrgentBandwidth
+	// Check urgent bandwidth required at dram vs urgent bandwidth avail at dram
+	// Check urgent bandwidth required at sdp vs urgent bandwidth avail at sdp, svp_prefetch -> FractionOfUrgentBandwidthMALL
+	// Check urgent bandwidth required at dram vs urgent bandwidth avail at dram, svp_prefetch
+
+	*bandwidth_support_ok &= urg_bandwidth_required[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_sdp] <= urg_bandwidth_available[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_sdp];
+	*bandwidth_support_ok &= urg_bandwidth_required[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_dram] <= urg_bandwidth_available[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_dram];
+
+	if (mall_allocated_for_dcn_mbytes > 0) {
+		*bandwidth_support_ok &= urg_bandwidth_required[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_sdp] <= urg_bandwidth_available[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_sdp];
+		*bandwidth_support_ok &= urg_bandwidth_required[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_dram] <= urg_bandwidth_available[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_dram];
+	}
+
+	*frac_urg_bandwidth_nom = math_max2(frac_urg_bandwidth_nom_sdp, frac_urg_bandwidth_nom_dram);
+	*frac_urg_bandwidth_mall = math_max2(frac_urg_bandwidth_mall_sdp, frac_urg_bandwidth_mall_dram);
+
+	*bandwidth_support_ok &= (*frac_urg_bandwidth_nom <= 1.0);
+
+	if (mall_allocated_for_dcn_mbytes > 0)
+		*bandwidth_support_ok &= (*frac_urg_bandwidth_mall <= 1.0);
+
+	*vactive_bandwidth_support_ok &= urg_vactive_bandwidth_required[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_sdp] <= urg_bandwidth_available[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_sdp];
+	*vactive_bandwidth_support_ok &= urg_vactive_bandwidth_required[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_dram] <= urg_bandwidth_available[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_dram];
+	if (mall_allocated_for_dcn_mbytes > 0) {
+		*vactive_bandwidth_support_ok &= urg_vactive_bandwidth_required[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_sdp] <= urg_bandwidth_available[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_sdp];
+		*vactive_bandwidth_support_ok &= urg_vactive_bandwidth_required[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_dram] <= urg_bandwidth_available[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_dram];
+	}
+
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: frac_urg_bandwidth_nom_sdp = %f\n", __func__, frac_urg_bandwidth_nom_sdp);
+	DML_LOG_VERBOSE("DML::%s: frac_urg_bandwidth_nom_dram = %f\n", __func__, frac_urg_bandwidth_nom_dram);
+	DML_LOG_VERBOSE("DML::%s: frac_urg_bandwidth_nom = %f\n", __func__, *frac_urg_bandwidth_nom);
+
+	DML_LOG_VERBOSE("DML::%s: frac_urg_bandwidth_mall_sdp = %f\n", __func__, frac_urg_bandwidth_mall_sdp);
+	DML_LOG_VERBOSE("DML::%s: frac_urg_bandwidth_mall_dram = %f\n", __func__, frac_urg_bandwidth_mall_dram);
+	DML_LOG_VERBOSE("DML::%s: frac_urg_bandwidth_mall = %f\n", __func__, *frac_urg_bandwidth_mall);
+	DML_LOG_VERBOSE("DML::%s: bandwidth_support_ok = %d\n", __func__, *bandwidth_support_ok);
+
+	for (unsigned int m = 0; m < dml2_core_internal_soc_state_max; m++) {
+		for (unsigned int n = 0; n < dml2_core_internal_bw_max; n++) {
+			DML_LOG_VERBOSE("DML::%s: state:%s bw_type:%s urg_bandwidth_available=%f %s urg_bandwidth_required=%f\n",
+			__func__, dml2_core_internal_soc_state_type_str(m), dml2_core_internal_bw_type_str(n),
+			urg_bandwidth_available[m][n], (urg_bandwidth_available[m][n] < urg_bandwidth_required[m][n]) ? "<" : ">=", urg_bandwidth_required[m][n]);
+		}
+	}
+#endif
+}
+
+static double get_bandwidth_available_for_immediate_flip(enum dml2_core_internal_soc_state_type eval_state,
+	double urg_bandwidth_required[dml2_core_internal_soc_state_max][dml2_core_internal_bw_max], // no flip
+	double urg_bandwidth_available[dml2_core_internal_soc_state_max][dml2_core_internal_bw_max])
+{
+	double flip_bw_available_mbps;
+	double flip_bw_available_sdp_mbps;
+	double flip_bw_available_dram_mbps;
+
+	flip_bw_available_sdp_mbps = urg_bandwidth_available[eval_state][dml2_core_internal_bw_sdp] - urg_bandwidth_required[eval_state][dml2_core_internal_bw_sdp];
+	flip_bw_available_dram_mbps = urg_bandwidth_available[eval_state][dml2_core_internal_bw_dram] - urg_bandwidth_required[eval_state][dml2_core_internal_bw_dram];
+	flip_bw_available_mbps = flip_bw_available_sdp_mbps < flip_bw_available_dram_mbps ? flip_bw_available_sdp_mbps : flip_bw_available_dram_mbps;
+
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: eval_state = %s\n", __func__, dml2_core_internal_soc_state_type_str(eval_state));
+	DML_LOG_VERBOSE("DML::%s: urg_bandwidth_available_sdp_mbps = %f\n", __func__, urg_bandwidth_available[eval_state][dml2_core_internal_bw_sdp]);
+	DML_LOG_VERBOSE("DML::%s: urg_bandwidth_available_dram_mbps = %f\n", __func__, urg_bandwidth_available[eval_state][dml2_core_internal_bw_dram]);
+	DML_LOG_VERBOSE("DML::%s: urg_bandwidth_required_sdp_mbps = %f\n", __func__, urg_bandwidth_required[eval_state][dml2_core_internal_bw_sdp]);
+	DML_LOG_VERBOSE("DML::%s: urg_bandwidth_required_dram_mbps = %f\n", __func__, urg_bandwidth_required[eval_state][dml2_core_internal_bw_dram]);
+	DML_LOG_VERBOSE("DML::%s: flip_bw_available_sdp_mbps = %f\n", __func__, flip_bw_available_sdp_mbps);
+	DML_LOG_VERBOSE("DML::%s: flip_bw_available_dram_mbps = %f\n", __func__, flip_bw_available_dram_mbps);
+	DML_LOG_VERBOSE("DML::%s: flip_bw_available_mbps = %f\n", __func__, flip_bw_available_mbps);
+#endif
+
+	return flip_bw_available_mbps;
+}
+
+static void calculate_immediate_flip_bandwidth_support(
+	// Output
+	double *frac_urg_bandwidth_flip,
+	bool *flip_bandwidth_support_ok,
+
+	// Input
+	enum dml2_core_internal_soc_state_type eval_state,
+	double urg_bandwidth_required_flip[dml2_core_internal_soc_state_max][dml2_core_internal_bw_max],
+	double non_urg_bandwidth_required_flip[dml2_core_internal_soc_state_max][dml2_core_internal_bw_max],
+	double urg_bandwidth_available[dml2_core_internal_soc_state_max][dml2_core_internal_bw_max])
+{
+	double frac_urg_bw_flip_sdp = non_urg_bandwidth_required_flip[eval_state][dml2_core_internal_bw_sdp] / urg_bandwidth_available[eval_state][dml2_core_internal_bw_sdp];
+	double frac_urg_bw_flip_dram = non_urg_bandwidth_required_flip[eval_state][dml2_core_internal_bw_dram] / urg_bandwidth_available[eval_state][dml2_core_internal_bw_dram];
+
+	*flip_bandwidth_support_ok = true;
+	for (unsigned int n = 0; n < dml2_core_internal_bw_max; n++) { // check sdp and dram
+		*flip_bandwidth_support_ok &= urg_bandwidth_available[eval_state][n] >= urg_bandwidth_required_flip[eval_state][n];
+
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: n = %s\n", __func__, dml2_core_internal_bw_type_str(n));
+		DML_LOG_VERBOSE("DML::%s: urg_bandwidth_available = %f\n", __func__, urg_bandwidth_available[eval_state][n]);
+		DML_LOG_VERBOSE("DML::%s: non_urg_bandwidth_required_flip = %f\n", __func__, non_urg_bandwidth_required_flip[eval_state][n]);
+		DML_LOG_VERBOSE("DML::%s: urg_bandwidth_required_flip = %f\n", __func__, urg_bandwidth_required_flip[eval_state][n]);
+		DML_LOG_VERBOSE("DML::%s: flip_bandwidth_support_ok = %d\n", __func__, *flip_bandwidth_support_ok);
+#endif
+		DML_ASSERT(urg_bandwidth_required_flip[eval_state][n] >= non_urg_bandwidth_required_flip[eval_state][n]);
+	}
+
+	*frac_urg_bandwidth_flip = (frac_urg_bw_flip_sdp > frac_urg_bw_flip_dram) ? frac_urg_bw_flip_sdp : frac_urg_bw_flip_dram;
+	*flip_bandwidth_support_ok &= (*frac_urg_bandwidth_flip <= 1.0);
+
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: eval_state = %s\n", __func__, dml2_core_internal_soc_state_type_str(eval_state));
+	DML_LOG_VERBOSE("DML::%s: frac_urg_bw_flip_sdp = %f\n", __func__, frac_urg_bw_flip_sdp);
+	DML_LOG_VERBOSE("DML::%s: frac_urg_bw_flip_dram = %f\n", __func__, frac_urg_bw_flip_dram);
+	DML_LOG_VERBOSE("DML::%s: frac_urg_bandwidth_flip = %f\n", __func__, *frac_urg_bandwidth_flip);
+	DML_LOG_VERBOSE("DML::%s: flip_bandwidth_support_ok = %d\n", __func__, *flip_bandwidth_support_ok);
+
+	for (unsigned int m = 0; m < dml2_core_internal_soc_state_max; m++) {
+		for (unsigned int n = 0; n < dml2_core_internal_bw_max; n++) {
+			DML_LOG_VERBOSE("DML::%s: state:%s bw_type:%s, urg_bandwidth_available=%f %s urg_bandwidth_required=%f\n",
+			__func__, dml2_core_internal_soc_state_type_str(m), dml2_core_internal_bw_type_str(n),
+			urg_bandwidth_available[m][n], (urg_bandwidth_available[m][n] < urg_bandwidth_required_flip[m][n]) ? "<" : ">=", urg_bandwidth_required_flip[m][n]);
+		}
+	}
+#endif
+}
+
+static void CalculateFlipSchedule(
+	struct dml2_core_internal_scratch *s,
+	bool iflip_enable,
+	bool use_lb_flip_bw,
+	double HostVMInefficiencyFactor,
+	double Tvm_trips_flip,
+	double Tr0_trips_flip,
+	double Tvm_trips_flip_rounded,
+	double Tr0_trips_flip_rounded,
+	bool GPUVMEnable,
+	double vm_bytes, // vm_bytes
+	double DPTEBytesPerRow, // dpte_row_bytes
+	double BandwidthAvailableForImmediateFlip,
+	unsigned int TotImmediateFlipBytes,
+	enum dml2_source_format_class SourcePixelFormat,
+	double LineTime,
+	double VRatio,
+	double VRatioChroma,
+	double Tno_bw_flip,
+	unsigned int dpte_row_height,
+	unsigned int dpte_row_height_chroma,
+	bool use_one_row_for_frame_flip,
+	unsigned int max_flip_time_us,
+	unsigned int max_flip_time_lines,
+	unsigned int per_pipe_flip_bytes,
+	unsigned int meta_row_bytes,
+	unsigned int meta_row_height,
+	unsigned int meta_row_height_chroma,
+	bool dcc_mrq_enable,
+
+	// Output
+	double *dst_y_per_vm_flip,
+	double *dst_y_per_row_flip,
+	double *final_flip_bw,
+	bool *ImmediateFlipSupportedForPipe)
+{
+	struct dml2_core_shared_CalculateFlipSchedule_locals *l = &s->CalculateFlipSchedule_locals;
+
+	l->dual_plane = dml_is_420(SourcePixelFormat) || SourcePixelFormat == dml2_rgbe_alpha;
+	l->dpte_row_bytes = DPTEBytesPerRow;
+
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: GPUVMEnable = %u\n", __func__, GPUVMEnable);
+	DML_LOG_VERBOSE("DML::%s: ip.max_flip_time_us = %d\n", __func__, max_flip_time_us);
+	DML_LOG_VERBOSE("DML::%s: ip.max_flip_time_lines = %d\n", __func__, max_flip_time_lines);
+	DML_LOG_VERBOSE("DML::%s: BandwidthAvailableForImmediateFlip = %f\n", __func__, BandwidthAvailableForImmediateFlip);
+	DML_LOG_VERBOSE("DML::%s: TotImmediateFlipBytes = %u\n", __func__, TotImmediateFlipBytes);
+	DML_LOG_VERBOSE("DML::%s: use_lb_flip_bw = %u\n", __func__, use_lb_flip_bw);
+	DML_LOG_VERBOSE("DML::%s: iflip_enable = %u\n", __func__, iflip_enable);
+	DML_LOG_VERBOSE("DML::%s: HostVMInefficiencyFactor = %f\n", __func__, HostVMInefficiencyFactor);
+	DML_LOG_VERBOSE("DML::%s: LineTime = %f\n", __func__, LineTime);
+	DML_LOG_VERBOSE("DML::%s: Tno_bw_flip = %f\n", __func__, Tno_bw_flip);
+	DML_LOG_VERBOSE("DML::%s: Tvm_trips_flip = %f\n", __func__, Tvm_trips_flip);
+	DML_LOG_VERBOSE("DML::%s: Tr0_trips_flip = %f\n", __func__, Tr0_trips_flip);
+	DML_LOG_VERBOSE("DML::%s: Tvm_trips_flip_rounded = %f\n", __func__, Tvm_trips_flip_rounded);
+	DML_LOG_VERBOSE("DML::%s: Tr0_trips_flip_rounded = %f\n", __func__, Tr0_trips_flip_rounded);
+	DML_LOG_VERBOSE("DML::%s: vm_bytes = %f\n", __func__, vm_bytes);
+	DML_LOG_VERBOSE("DML::%s: DPTEBytesPerRow = %f\n", __func__, DPTEBytesPerRow);
+	DML_LOG_VERBOSE("DML::%s: meta_row_bytes = %d\n", __func__, meta_row_bytes);
+	DML_LOG_VERBOSE("DML::%s: dpte_row_bytes = %f\n", __func__, l->dpte_row_bytes);
+	DML_LOG_VERBOSE("DML::%s: dpte_row_height = %d\n", __func__, dpte_row_height);
+	DML_LOG_VERBOSE("DML::%s: meta_row_height = %d\n", __func__, meta_row_height);
+	DML_LOG_VERBOSE("DML::%s: VRatio = %f\n", __func__, VRatio);
+#endif
+
+	if (TotImmediateFlipBytes > 0 && (GPUVMEnable || dcc_mrq_enable)) {
+		if (l->dual_plane) {
+			if (dcc_mrq_enable & GPUVMEnable) {
+				l->min_row_height = math_min2(dpte_row_height, meta_row_height);
+				l->min_row_height_chroma = math_min2(dpte_row_height_chroma, meta_row_height_chroma);
+			} else if (GPUVMEnable) {
+				l->min_row_height = dpte_row_height;
+				l->min_row_height_chroma = dpte_row_height_chroma;
+			} else {
+				l->min_row_height = meta_row_height;
+				l->min_row_height_chroma = meta_row_height_chroma;
+			}
+			l->min_row_time = math_min2(l->min_row_height * LineTime / VRatio, l->min_row_height_chroma * LineTime / VRatioChroma);
+		} else {
+			if (dcc_mrq_enable & GPUVMEnable)
+				l->min_row_height = math_min2(dpte_row_height, meta_row_height);
+			else if (GPUVMEnable)
+				l->min_row_height = dpte_row_height;
+			else
+				l->min_row_height = meta_row_height;
+
+			l->min_row_time = l->min_row_height * LineTime / VRatio;
+		}
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: min_row_time = %f\n", __func__, l->min_row_time);
+#endif
+		DML_ASSERT(l->min_row_time > 0);
+
+		if (use_lb_flip_bw) {
+			// For mode check, calculation the flip bw requirement with worst case flip time
+			l->max_flip_time = math_min2(math_min2(l->min_row_time, (double)max_flip_time_lines * LineTime / VRatio),
+				math_max2(Tvm_trips_flip_rounded + 2 * Tr0_trips_flip_rounded, (double)max_flip_time_us));
+
+			//The lower bound on flip bandwidth
+			// Note: The get_urgent_bandwidth_required already consider dpte_row_bw and meta_row_bw in bandwidth calculation, so leave final_flip_bw = 0 if iflip not required
+			l->lb_flip_bw = 0;
+
+			if (iflip_enable) {
+				l->hvm_scaled_vm_bytes = vm_bytes * HostVMInefficiencyFactor;
+				l->num_rows = 2;
+				l->hvm_scaled_row_bytes = (l->num_rows * l->dpte_row_bytes * HostVMInefficiencyFactor + l->num_rows * meta_row_bytes);
+				l->hvm_scaled_vm_row_bytes = l->hvm_scaled_vm_bytes + l->hvm_scaled_row_bytes;
+				l->lb_flip_bw = math_max3(
+					l->hvm_scaled_vm_row_bytes / (l->max_flip_time - Tno_bw_flip),
+					l->hvm_scaled_vm_bytes / (l->max_flip_time - Tno_bw_flip - 2 * Tr0_trips_flip_rounded),
+					l->hvm_scaled_row_bytes / (l->max_flip_time - Tvm_trips_flip_rounded));
+#ifdef __DML_VBA_DEBUG__
+				DML_LOG_VERBOSE("DML::%s: max_flip_time = %f\n", __func__, l->max_flip_time);
+				DML_LOG_VERBOSE("DML::%s: total vm bytes (hvm ineff scaled) = %f\n", __func__, l->hvm_scaled_vm_bytes);
+				DML_LOG_VERBOSE("DML::%s: total row bytes (%f row, hvm ineff scaled) = %f\n", __func__, l->num_rows, l->hvm_scaled_row_bytes);
+				DML_LOG_VERBOSE("DML::%s: total vm+row bytes (hvm ineff scaled) = %f\n", __func__, l->hvm_scaled_vm_row_bytes);
+				DML_LOG_VERBOSE("DML::%s: lb_flip_bw for vm and row = %f\n", __func__, l->hvm_scaled_vm_row_bytes / (l->max_flip_time - Tno_bw_flip));
+				DML_LOG_VERBOSE("DML::%s: lb_flip_bw for vm = %f\n", __func__, l->hvm_scaled_vm_bytes / (l->max_flip_time - Tno_bw_flip - 2 * Tr0_trips_flip_rounded));
+				DML_LOG_VERBOSE("DML::%s: lb_flip_bw for row = %f\n", __func__, l->hvm_scaled_row_bytes / (l->max_flip_time - Tvm_trips_flip_rounded));
+
+				if (l->lb_flip_bw > 0) {
+					DML_LOG_VERBOSE("DML::%s: mode_support est Tvm_flip = %f (bw-based)\n", __func__, Tno_bw_flip + l->hvm_scaled_vm_bytes / l->lb_flip_bw);
+					DML_LOG_VERBOSE("DML::%s: mode_support est Tr0_flip = %f (bw-based)\n", __func__, l->hvm_scaled_row_bytes / l->lb_flip_bw / l->num_rows);
+					DML_LOG_VERBOSE("DML::%s: mode_support est dst_y_per_vm_flip = %f (bw-based)\n", __func__, Tno_bw_flip + l->hvm_scaled_vm_bytes / l->lb_flip_bw / LineTime);
+					DML_LOG_VERBOSE("DML::%s: mode_support est dst_y_per_row_flip = %f (bw-based)\n", __func__, l->hvm_scaled_row_bytes / l->lb_flip_bw / LineTime / l->num_rows);
+					DML_LOG_VERBOSE("DML::%s: Tvm_trips_flip_rounded + 2*Tr0_trips_flip_rounded = %f\n", __func__, (Tvm_trips_flip_rounded + 2 * Tr0_trips_flip_rounded));
+				}
+#endif
+				l->lb_flip_bw = math_max3(l->lb_flip_bw,
+						l->hvm_scaled_vm_bytes / (31 * LineTime) - Tno_bw_flip,
+						(l->dpte_row_bytes * HostVMInefficiencyFactor + meta_row_bytes) / (15 * LineTime));
+
+#ifdef __DML_VBA_DEBUG__
+				DML_LOG_VERBOSE("DML::%s: lb_flip_bw for vm reg limit = %f\n", __func__, l->hvm_scaled_vm_bytes / (31 * LineTime) - Tno_bw_flip);
+				DML_LOG_VERBOSE("DML::%s: lb_flip_bw for row reg limit = %f\n", __func__, (l->dpte_row_bytes * HostVMInefficiencyFactor + meta_row_bytes) / (15 * LineTime));
+#endif
+			}
+
+			*final_flip_bw = l->lb_flip_bw;
+
+			*dst_y_per_vm_flip = 1; // not used
+			*dst_y_per_row_flip = 1; // not used
+			*ImmediateFlipSupportedForPipe = l->min_row_time >= (Tvm_trips_flip_rounded + 2 * Tr0_trips_flip_rounded);
+		} else {
+			if (iflip_enable) {
+				l->ImmediateFlipBW = (double)per_pipe_flip_bytes * BandwidthAvailableForImmediateFlip / (double)TotImmediateFlipBytes; // flip_bw(i)
+
+#ifdef __DML_VBA_DEBUG__
+				DML_LOG_VERBOSE("DML::%s: per_pipe_flip_bytes = %d\n", __func__, per_pipe_flip_bytes);
+				DML_LOG_VERBOSE("DML::%s: BandwidthAvailableForImmediateFlip = %f\n", __func__, BandwidthAvailableForImmediateFlip);
+				DML_LOG_VERBOSE("DML::%s: ImmediateFlipBW = %f\n", __func__, l->ImmediateFlipBW);
+				DML_LOG_VERBOSE("DML::%s: portion of flip bw = %f\n", __func__, (double)per_pipe_flip_bytes / (double)TotImmediateFlipBytes);
+#endif
+				if (l->ImmediateFlipBW == 0) {
+					l->Tvm_flip = 0;
+					l->Tr0_flip = 0;
+				} else {
+					l->Tvm_flip = math_max3(Tvm_trips_flip,
+						Tno_bw_flip + vm_bytes * HostVMInefficiencyFactor / l->ImmediateFlipBW,
+						LineTime / 4.0);
+
+					l->Tr0_flip = math_max3(Tr0_trips_flip,
+						(l->dpte_row_bytes * HostVMInefficiencyFactor + meta_row_bytes) / l->ImmediateFlipBW,
+						LineTime / 4.0);
+				}
+#ifdef __DML_VBA_DEBUG__
+				DML_LOG_VERBOSE("DML::%s: total vm bytes (hvm ineff scaled) = %f\n", __func__, vm_bytes * HostVMInefficiencyFactor);
+				DML_LOG_VERBOSE("DML::%s: total row bytes (hvm ineff scaled, one row) = %f\n", __func__, (l->dpte_row_bytes * HostVMInefficiencyFactor + meta_row_bytes));
+
+				DML_LOG_VERBOSE("DML::%s: Tvm_flip = %f (bw-based), Tvm_trips_flip = %f (latency-based)\n", __func__, Tno_bw_flip + vm_bytes * HostVMInefficiencyFactor / l->ImmediateFlipBW, Tvm_trips_flip);
+				DML_LOG_VERBOSE("DML::%s: Tr0_flip = %f (bw-based), Tr0_trips_flip = %f (latency-based)\n", __func__, (l->dpte_row_bytes * HostVMInefficiencyFactor + meta_row_bytes) / l->ImmediateFlipBW, Tr0_trips_flip);
+#endif
+				*dst_y_per_vm_flip = math_ceil2(4.0 * (l->Tvm_flip / LineTime), 1.0) / 4.0;
+				*dst_y_per_row_flip = math_ceil2(4.0 * (l->Tr0_flip / LineTime), 1.0) / 4.0;
+
+				*final_flip_bw = math_max2(vm_bytes * HostVMInefficiencyFactor / (*dst_y_per_vm_flip * LineTime),
+					(l->dpte_row_bytes * HostVMInefficiencyFactor + meta_row_bytes) / (*dst_y_per_row_flip * LineTime));
+
+				if (*dst_y_per_vm_flip >= 32 || *dst_y_per_row_flip >= 16 || l->Tvm_flip + 2 * l->Tr0_flip > l->min_row_time) {
+					*ImmediateFlipSupportedForPipe = false;
+				} else {
+					*ImmediateFlipSupportedForPipe = iflip_enable;
+				}
+			} else {
+				l->Tvm_flip = 0;
+				l->Tr0_flip = 0;
+				*dst_y_per_vm_flip = 0;
+				*dst_y_per_row_flip = 0;
+				*final_flip_bw = 0;
+				*ImmediateFlipSupportedForPipe = iflip_enable;
+			}
+		}
+	} else {
+		l->Tvm_flip = 0;
+		l->Tr0_flip = 0;
+		*dst_y_per_vm_flip = 0;
+		*dst_y_per_row_flip = 0;
+		*final_flip_bw = 0;
+		*ImmediateFlipSupportedForPipe = iflip_enable;
+	}
+
+#ifdef __DML_VBA_DEBUG__
+	if (!use_lb_flip_bw) {
+		DML_LOG_VERBOSE("DML::%s: dst_y_per_vm_flip = %f (should be < 32)\n", __func__, *dst_y_per_vm_flip);
+		DML_LOG_VERBOSE("DML::%s: dst_y_per_row_flip = %f (should be < 16)\n", __func__, *dst_y_per_row_flip);
+		DML_LOG_VERBOSE("DML::%s: Tvm_flip = %f (final)\n", __func__, l->Tvm_flip);
+		DML_LOG_VERBOSE("DML::%s: Tr0_flip = %f (final)\n", __func__, l->Tr0_flip);
+		DML_LOG_VERBOSE("DML::%s: Tvm_flip + 2*Tr0_flip = %f (should be <= min_row_time=%f)\n", __func__, l->Tvm_flip + 2 * l->Tr0_flip, l->min_row_time);
+	}
+	DML_LOG_VERBOSE("DML::%s: final_flip_bw = %f\n", __func__, *final_flip_bw);
+	DML_LOG_VERBOSE("DML::%s: ImmediateFlipSupportedForPipe = %u\n", __func__, *ImmediateFlipSupportedForPipe);
+#endif
+}
+
+static void CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport(
+	struct dml2_core_internal_scratch *scratch,
+	struct dml2_core_calcs_CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_params *p)
+{
+	struct dml2_core_calcs_CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_locals *s = &scratch->CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_locals;
+
+	enum dml2_uclk_pstate_change_strategy uclk_pstate_change_strategy;
+	double reserved_vblank_time_us;
+	bool FoundCriticalSurface = false;
+
+	s->TotalActiveWriteback = 0;
+	p->Watermark->UrgentWatermark = p->mmSOCParameters.UrgentLatency + p->mmSOCParameters.ExtraLatency;
+
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: UrgentWatermark = %f\n", __func__, p->Watermark->UrgentWatermark);
+#endif
+
+	p->Watermark->USRRetrainingWatermark = p->mmSOCParameters.UrgentLatency + p->mmSOCParameters.ExtraLatency + p->mmSOCParameters.USRRetrainingLatency + p->mmSOCParameters.SMNLatency;
+	p->Watermark->DRAMClockChangeWatermark = p->mmSOCParameters.DRAMClockChangeLatency + p->Watermark->UrgentWatermark;
+	p->Watermark->FCLKChangeWatermark = p->mmSOCParameters.FCLKChangeLatency + p->Watermark->UrgentWatermark;
+	p->Watermark->StutterExitWatermark = p->mmSOCParameters.SRExitTime + p->mmSOCParameters.ExtraLatency_sr + 10 / p->DCFClkDeepSleep;
+	p->Watermark->StutterEnterPlusExitWatermark = p->mmSOCParameters.SREnterPlusExitTime + p->mmSOCParameters.ExtraLatency_sr + 10 / p->DCFClkDeepSleep;
+	p->Watermark->Z8StutterExitWatermark = p->mmSOCParameters.SRExitZ8Time + p->mmSOCParameters.ExtraLatency_sr + 10 / p->DCFClkDeepSleep;
+	p->Watermark->Z8StutterEnterPlusExitWatermark = p->mmSOCParameters.SREnterPlusExitZ8Time + p->mmSOCParameters.ExtraLatency_sr + 10 / p->DCFClkDeepSleep;
+	if (p->mmSOCParameters.qos_type == dml2_qos_param_type_dcn4x) {
+		p->Watermark->StutterExitWatermark += p->mmSOCParameters.max_urgent_latency_us + p->mmSOCParameters.df_response_time_us;
+		p->Watermark->StutterEnterPlusExitWatermark += p->mmSOCParameters.max_urgent_latency_us + p->mmSOCParameters.df_response_time_us;
+		p->Watermark->Z8StutterExitWatermark += p->mmSOCParameters.max_urgent_latency_us + p->mmSOCParameters.df_response_time_us;
+		p->Watermark->Z8StutterEnterPlusExitWatermark += p->mmSOCParameters.max_urgent_latency_us + p->mmSOCParameters.df_response_time_us;
+	}
+	p->Watermark->temp_read_or_ppt_watermark_us = p->mmSOCParameters.g6_temp_read_blackout_us + p->Watermark->UrgentWatermark;
+
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: UrgentLatency = %f\n", __func__, p->mmSOCParameters.UrgentLatency);
+	DML_LOG_VERBOSE("DML::%s: ExtraLatency = %f\n", __func__, p->mmSOCParameters.ExtraLatency);
+	DML_LOG_VERBOSE("DML::%s: DRAMClockChangeLatency = %f\n", __func__, p->mmSOCParameters.DRAMClockChangeLatency);
+	DML_LOG_VERBOSE("DML::%s: SREnterPlusExitZ8Time = %f\n", __func__, p->mmSOCParameters.SREnterPlusExitZ8Time);
+	DML_LOG_VERBOSE("DML::%s: SREnterPlusExitTime = %f\n", __func__, p->mmSOCParameters.SREnterPlusExitTime);
+	DML_LOG_VERBOSE("DML::%s: UrgentWatermark = %f\n", __func__, p->Watermark->UrgentWatermark);
+	DML_LOG_VERBOSE("DML::%s: USRRetrainingWatermark = %f\n", __func__, p->Watermark->USRRetrainingWatermark);
+	DML_LOG_VERBOSE("DML::%s: DRAMClockChangeWatermark = %f\n", __func__, p->Watermark->DRAMClockChangeWatermark);
+	DML_LOG_VERBOSE("DML::%s: FCLKChangeWatermark = %f\n", __func__, p->Watermark->FCLKChangeWatermark);
+	DML_LOG_VERBOSE("DML::%s: StutterExitWatermark = %f\n", __func__, p->Watermark->StutterExitWatermark);
+	DML_LOG_VERBOSE("DML::%s: StutterEnterPlusExitWatermark = %f\n", __func__, p->Watermark->StutterEnterPlusExitWatermark);
+	DML_LOG_VERBOSE("DML::%s: Z8StutterExitWatermark = %f\n", __func__, p->Watermark->Z8StutterExitWatermark);
+	DML_LOG_VERBOSE("DML::%s: Z8StutterEnterPlusExitWatermark = %f\n", __func__, p->Watermark->Z8StutterEnterPlusExitWatermark);
+	DML_LOG_VERBOSE("DML::%s: temp_read_or_ppt_watermark_us = %f\n", __func__, p->Watermark->temp_read_or_ppt_watermark_us);
+#endif
+
+	s->TotalActiveWriteback = 0;
+	for (unsigned int k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+		if (p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].writeback.active_writebacks_per_stream > 0) {
+			s->TotalActiveWriteback = s->TotalActiveWriteback + 1;
+		}
+	}
+
+	if (s->TotalActiveWriteback <= 1) {
+		p->Watermark->WritebackUrgentWatermark = p->mmSOCParameters.WritebackLatency;
+	} else {
+		p->Watermark->WritebackUrgentWatermark = p->mmSOCParameters.WritebackLatency + p->WritebackChunkSize * 1024.0 / 32.0 / p->SOCCLK;
+	}
+	if (p->USRRetrainingRequired)
+		p->Watermark->WritebackUrgentWatermark = p->Watermark->WritebackUrgentWatermark + p->mmSOCParameters.USRRetrainingLatency;
+
+	if (s->TotalActiveWriteback <= 1) {
+		p->Watermark->WritebackDRAMClockChangeWatermark = p->mmSOCParameters.DRAMClockChangeLatency + p->mmSOCParameters.WritebackLatency;
+		p->Watermark->WritebackFCLKChangeWatermark = p->mmSOCParameters.FCLKChangeLatency + p->mmSOCParameters.WritebackLatency;
+	} else {
+		p->Watermark->WritebackDRAMClockChangeWatermark = p->mmSOCParameters.DRAMClockChangeLatency + p->mmSOCParameters.WritebackLatency + p->WritebackChunkSize * 1024.0 / 32.0 / p->SOCCLK;
+		p->Watermark->WritebackFCLKChangeWatermark = p->mmSOCParameters.FCLKChangeLatency + p->mmSOCParameters.WritebackLatency + p->WritebackChunkSize * 1024 / 32 / p->SOCCLK;
+	}
+
+	if (p->USRRetrainingRequired)
+		p->Watermark->WritebackDRAMClockChangeWatermark = p->Watermark->WritebackDRAMClockChangeWatermark + p->mmSOCParameters.USRRetrainingLatency;
+
+	if (p->USRRetrainingRequired)
+		p->Watermark->WritebackFCLKChangeWatermark = p->Watermark->WritebackFCLKChangeWatermark + p->mmSOCParameters.USRRetrainingLatency;
+
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: WritebackDRAMClockChangeWatermark = %f\n", __func__, p->Watermark->WritebackDRAMClockChangeWatermark);
+	DML_LOG_VERBOSE("DML::%s: WritebackFCLKChangeWatermark = %f\n", __func__, p->Watermark->WritebackFCLKChangeWatermark);
+	DML_LOG_VERBOSE("DML::%s: WritebackUrgentWatermark = %f\n", __func__, p->Watermark->WritebackUrgentWatermark);
+	DML_LOG_VERBOSE("DML::%s: USRRetrainingRequired = %u\n", __func__, p->USRRetrainingRequired);
+	DML_LOG_VERBOSE("DML::%s: USRRetrainingLatency = %f\n", __func__, p->mmSOCParameters.USRRetrainingLatency);
+#endif
+
+	s->TotalPixelBW = 0.0;
+	for (unsigned int k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+		double h_total = (double)p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.h_total;
+		double pixel_clock_mhz = p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000.0;
+		double v_ratio = p->display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio;
+		double v_ratio_c = p->display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_ratio;
+		s->TotalPixelBW = s->TotalPixelBW + p->DPPPerSurface[k]
+			* (p->SwathWidthY[k] * p->BytePerPixelDETY[k] * v_ratio + p->SwathWidthC[k] * p->BytePerPixelDETC[k] * v_ratio_c) / (h_total / pixel_clock_mhz);
+	}
+
+	*p->global_fclk_change_supported = true;
+	*p->global_dram_clock_change_supported = true;
+
+	for (unsigned int k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+		double h_total = (double)p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.h_total;
+		double pixel_clock_mhz = p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000.0;
+		double v_ratio = p->display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio;
+		double v_ratio_c = p->display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_ratio;
+		double v_taps = p->display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_taps;
+		double v_taps_c = p->display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_taps;
+		double h_ratio = p->display_cfg->plane_descriptors[k].composition.scaler_info.plane0.h_ratio;
+		double h_ratio_c = p->display_cfg->plane_descriptors[k].composition.scaler_info.plane1.h_ratio;
+		double LBBitPerPixel = 57;
+
+		s->LBLatencyHidingSourceLinesY[k] = (unsigned int)(math_min2((double)p->MaxLineBufferLines, math_floor2((double)p->LineBufferSize / LBBitPerPixel / ((double)p->SwathWidthY[k] / math_max2(h_ratio, 1.0)), 1)) - (v_taps - 1));
+		s->LBLatencyHidingSourceLinesC[k] = (unsigned int)(math_min2((double)p->MaxLineBufferLines, math_floor2((double)p->LineBufferSize / LBBitPerPixel / ((double)p->SwathWidthC[k] / math_max2(h_ratio_c, 1.0)), 1)) - (v_taps_c - 1));
+
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: k=%u, MaxLineBufferLines = %u\n", __func__, k, p->MaxLineBufferLines);
+		DML_LOG_VERBOSE("DML::%s: k=%u, LineBufferSize = %u\n", __func__, k, p->LineBufferSize);
+		DML_LOG_VERBOSE("DML::%s: k=%u, LBBitPerPixel = %f\n", __func__, k, LBBitPerPixel);
+		DML_LOG_VERBOSE("DML::%s: k=%u, HRatio = %f\n", __func__, k, h_ratio);
+		DML_LOG_VERBOSE("DML::%s: k=%u, VTaps = %f\n", __func__, k, v_taps);
+#endif
+
+		s->EffectiveLBLatencyHidingY = s->LBLatencyHidingSourceLinesY[k] / v_ratio * (h_total / pixel_clock_mhz);
+		s->EffectiveLBLatencyHidingC = s->LBLatencyHidingSourceLinesC[k] / v_ratio_c * (h_total / pixel_clock_mhz);
+
+		s->EffectiveDETBufferSizeY = p->DETBufferSizeY[k];
+		if (p->UnboundedRequestEnabled) {
+			s->EffectiveDETBufferSizeY = s->EffectiveDETBufferSizeY + p->CompressedBufferSizeInkByte * 1024 * (p->SwathWidthY[k] * p->BytePerPixelDETY[k] * v_ratio) / (h_total / pixel_clock_mhz) / s->TotalPixelBW;
+		}
+
+		s->LinesInDETY[k] = (double)s->EffectiveDETBufferSizeY / p->BytePerPixelDETY[k] / p->SwathWidthY[k];
+		s->LinesInDETYRoundedDownToSwath[k] = (unsigned int)(math_floor2(s->LinesInDETY[k], p->SwathHeightY[k]));
+		s->FullDETBufferingTimeY = s->LinesInDETYRoundedDownToSwath[k] * (h_total / pixel_clock_mhz) / v_ratio;
+
+		s->ActiveClockChangeLatencyHidingY = s->EffectiveLBLatencyHidingY + s->FullDETBufferingTimeY - ((double)p->DSTXAfterScaler[k] / h_total + (double)p->DSTYAfterScaler[k]) * h_total / pixel_clock_mhz;
+
+		if (p->NumberOfActiveSurfaces > 1) {
+			s->ActiveClockChangeLatencyHidingY = s->ActiveClockChangeLatencyHidingY - (1.0 - 1.0 / (double)p->NumberOfActiveSurfaces) * (double)p->SwathHeightY[k] * (double)h_total / pixel_clock_mhz / v_ratio;
+		}
+
+		if (p->BytePerPixelDETC[k] > 0) {
+			s->LinesInDETC[k] = p->DETBufferSizeC[k] / p->BytePerPixelDETC[k] / p->SwathWidthC[k];
+			s->LinesInDETCRoundedDownToSwath[k] = (unsigned int)(math_floor2(s->LinesInDETC[k], p->SwathHeightC[k]));
+			s->FullDETBufferingTimeC = s->LinesInDETCRoundedDownToSwath[k] * (h_total / pixel_clock_mhz) / v_ratio_c;
+			s->ActiveClockChangeLatencyHidingC = s->EffectiveLBLatencyHidingC + s->FullDETBufferingTimeC - ((double)p->DSTXAfterScaler[k] / (double)h_total + (double)p->DSTYAfterScaler[k]) * (double)h_total / pixel_clock_mhz;
+			if (p->NumberOfActiveSurfaces > 1) {
+				s->ActiveClockChangeLatencyHidingC = s->ActiveClockChangeLatencyHidingC - (1.0 - 1.0 / (double)p->NumberOfActiveSurfaces) * (double)p->SwathHeightC[k] * (double)h_total / pixel_clock_mhz / v_ratio_c;
+			}
+			s->ActiveClockChangeLatencyHiding = math_min2(s->ActiveClockChangeLatencyHidingY, s->ActiveClockChangeLatencyHidingC);
+		} else {
+			s->ActiveClockChangeLatencyHiding = s->ActiveClockChangeLatencyHidingY;
+		}
+
+		s->ActiveDRAMClockChangeLatencyMargin[k] = s->ActiveClockChangeLatencyHiding - p->Watermark->DRAMClockChangeWatermark;
+		s->ActiveFCLKChangeLatencyMargin[k] = s->ActiveClockChangeLatencyHiding - p->Watermark->FCLKChangeWatermark;
+		s->USRRetrainingLatencyMargin[k] = s->ActiveClockChangeLatencyHiding - p->Watermark->USRRetrainingWatermark;
+		s->g6_temp_read_latency_margin[k] = s->ActiveClockChangeLatencyHiding - p->Watermark->temp_read_or_ppt_watermark_us;
+
+		if (p->VActiveLatencyHidingMargin)
+			p->VActiveLatencyHidingMargin[k] = s->ActiveDRAMClockChangeLatencyMargin[k];
+
+		if (p->VActiveLatencyHidingUs)
+			p->VActiveLatencyHidingUs[k] = s->ActiveClockChangeLatencyHiding;
+
+		if (p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].writeback.active_writebacks_per_stream > 0) {
+			s->WritebackLatencyHiding = (double)p->WritebackInterfaceBufferSize * 1024.0
+				/ ((double)p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].output_height
+					* (double)p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].output_width
+					/ ((double)p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].input_height * (double)h_total / pixel_clock_mhz) * 4.0);
+			if (p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].pixel_format == dml2_444_64) {
+				s->WritebackLatencyHiding = s->WritebackLatencyHiding / 2;
+			}
+			s->WritebackDRAMClockChangeLatencyMargin = s->WritebackLatencyHiding - p->Watermark->WritebackDRAMClockChangeWatermark;
+
+			s->WritebackFCLKChangeLatencyMargin = s->WritebackLatencyHiding - p->Watermark->WritebackFCLKChangeWatermark;
+
+			s->ActiveDRAMClockChangeLatencyMargin[k] = math_min2(s->ActiveDRAMClockChangeLatencyMargin[k], s->WritebackDRAMClockChangeLatencyMargin);
+			s->ActiveFCLKChangeLatencyMargin[k] = math_min2(s->ActiveFCLKChangeLatencyMargin[k], s->WritebackFCLKChangeLatencyMargin);
+		}
+		p->MaxActiveDRAMClockChangeLatencySupported[k] = dml_is_phantom_pipe(&p->display_cfg->plane_descriptors[k]) ? 0 : (s->ActiveDRAMClockChangeLatencyMargin[k] + p->mmSOCParameters.DRAMClockChangeLatency);
+
+		uclk_pstate_change_strategy = p->display_cfg->plane_descriptors[k].overrides.uclk_pstate_change_strategy;
+		reserved_vblank_time_us = (double)p->display_cfg->plane_descriptors[k].overrides.reserved_vblank_time_ns / 1000;
+
+		p->FCLKChangeSupport[k] = dml2_pstate_change_unsupported;
+		if (s->ActiveFCLKChangeLatencyMargin[k] > 0)
+			p->FCLKChangeSupport[k] = dml2_pstate_change_vactive;
+		else if (reserved_vblank_time_us >= p->mmSOCParameters.FCLKChangeLatency)
+			p->FCLKChangeSupport[k] = dml2_pstate_change_vblank;
+
+		if (p->FCLKChangeSupport[k] == dml2_pstate_change_unsupported)
+			*p->global_fclk_change_supported = false;
+
+		p->DRAMClockChangeSupport[k] = dml2_pstate_change_unsupported;
+		if (uclk_pstate_change_strategy == dml2_uclk_pstate_change_strategy_auto) {
+			if (p->display_cfg->overrides.all_streams_blanked ||
+					(s->ActiveDRAMClockChangeLatencyMargin[k] > 0 && reserved_vblank_time_us >= p->mmSOCParameters.DRAMClockChangeLatency))
+				p->DRAMClockChangeSupport[k] = dml2_pstate_change_vblank_and_vactive;
+			else if (s->ActiveDRAMClockChangeLatencyMargin[k] > 0)
+				p->DRAMClockChangeSupport[k] = dml2_pstate_change_vactive;
+			else if (reserved_vblank_time_us >= p->mmSOCParameters.DRAMClockChangeLatency)
+				p->DRAMClockChangeSupport[k] = dml2_pstate_change_vblank;
+		} else if (uclk_pstate_change_strategy == dml2_uclk_pstate_change_strategy_force_vactive && s->ActiveDRAMClockChangeLatencyMargin[k] > 0)
+			p->DRAMClockChangeSupport[k] = dml2_pstate_change_vactive;
+		else if (uclk_pstate_change_strategy == dml2_uclk_pstate_change_strategy_force_vblank && reserved_vblank_time_us >= p->mmSOCParameters.DRAMClockChangeLatency)
+			p->DRAMClockChangeSupport[k] = dml2_pstate_change_vblank;
+		else if (uclk_pstate_change_strategy == dml2_uclk_pstate_change_strategy_force_drr)
+			p->DRAMClockChangeSupport[k] = dml2_pstate_change_drr;
+		else if (uclk_pstate_change_strategy == dml2_uclk_pstate_change_strategy_force_mall_svp)
+			p->DRAMClockChangeSupport[k] = dml2_pstate_change_mall_svp;
+		else if (uclk_pstate_change_strategy == dml2_uclk_pstate_change_strategy_force_mall_full_frame)
+			p->DRAMClockChangeSupport[k] = dml2_pstate_change_mall_full_frame;
+
+		if (p->DRAMClockChangeSupport[k] == dml2_pstate_change_unsupported)
+			*p->global_dram_clock_change_supported = false;
+
+		s->dst_y_pstate = (unsigned int)(math_ceil2((p->mmSOCParameters.DRAMClockChangeLatency + p->mmSOCParameters.UrgentLatency) / (h_total / pixel_clock_mhz), 1));
+		s->src_y_pstate_l = (unsigned int)(math_ceil2(s->dst_y_pstate * v_ratio, p->SwathHeightY[k]));
+		s->src_y_ahead_l = (unsigned int)(math_floor2(p->DETBufferSizeY[k] / p->BytePerPixelDETY[k] / p->SwathWidthY[k], p->SwathHeightY[k]) + s->LBLatencyHidingSourceLinesY[k]);
+		s->sub_vp_lines_l = s->src_y_pstate_l + s->src_y_ahead_l + p->meta_row_height_l[k];
+
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: k=%u, DETBufferSizeY = %u\n", __func__, k, p->DETBufferSizeY[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u, BytePerPixelDETY = %f\n", __func__, k, p->BytePerPixelDETY[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u, SwathWidthY = %u\n", __func__, k, p->SwathWidthY[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u, SwathHeightY = %u\n", __func__, k, p->SwathHeightY[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u, LBLatencyHidingSourceLinesY = %u\n", __func__, k, s->LBLatencyHidingSourceLinesY[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u, dst_y_pstate = %u\n", __func__, k, s->dst_y_pstate);
+		DML_LOG_VERBOSE("DML::%s: k=%u, src_y_pstate_l = %u\n", __func__, k, s->src_y_pstate_l);
+		DML_LOG_VERBOSE("DML::%s: k=%u, src_y_ahead_l = %u\n", __func__, k, s->src_y_ahead_l);
+		DML_LOG_VERBOSE("DML::%s: k=%u, meta_row_height_l = %u\n", __func__, k, p->meta_row_height_l[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u, sub_vp_lines_l = %u\n", __func__, k, s->sub_vp_lines_l);
+#endif
+		p->SubViewportLinesNeededInMALL[k] = s->sub_vp_lines_l;
+
+		if (p->BytePerPixelDETC[k] > 0) {
+			s->src_y_pstate_c = (unsigned int)(math_ceil2(s->dst_y_pstate * v_ratio_c, p->SwathHeightC[k]));
+			s->src_y_ahead_c = (unsigned int)(math_floor2(p->DETBufferSizeC[k] / p->BytePerPixelDETC[k] / p->SwathWidthC[k], p->SwathHeightC[k]) + s->LBLatencyHidingSourceLinesC[k]);
+			s->sub_vp_lines_c = s->src_y_pstate_c + s->src_y_ahead_c + p->meta_row_height_c[k];
+
+			if (dml_is_420(p->display_cfg->plane_descriptors[k].pixel_format))
+				p->SubViewportLinesNeededInMALL[k] = (unsigned int)(math_max2(s->sub_vp_lines_l, 2 * s->sub_vp_lines_c));
+			else
+				p->SubViewportLinesNeededInMALL[k] = (unsigned int)(math_max2(s->sub_vp_lines_l, s->sub_vp_lines_c));
+
+#ifdef __DML_VBA_DEBUG__
+			DML_LOG_VERBOSE("DML::%s: k=%u, meta_row_height_c = %u\n", __func__, k, p->meta_row_height_c[k]);
+			DML_LOG_VERBOSE("DML::%s: k=%u, src_y_pstate_c = %u\n", __func__, k, s->src_y_pstate_c);
+			DML_LOG_VERBOSE("DML::%s: k=%u, src_y_ahead_c = %u\n", __func__, k, s->src_y_ahead_c);
+			DML_LOG_VERBOSE("DML::%s: k=%u, sub_vp_lines_c = %u\n", __func__, k, s->sub_vp_lines_c);
+#endif
+		}
+	}
+
+	*p->g6_temp_read_support = true;
+	for (unsigned int k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+		if ((!dml_is_phantom_pipe(&p->display_cfg->plane_descriptors[k])) &&
+			(s->g6_temp_read_latency_margin[k] < 0)) {
+			*p->g6_temp_read_support = false;
+		}
+	}
+
+	for (unsigned int k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+		if ((!dml_is_phantom_pipe(&p->display_cfg->plane_descriptors[k])) && ((!FoundCriticalSurface)
+			|| ((s->ActiveFCLKChangeLatencyMargin[k] + p->mmSOCParameters.FCLKChangeLatency) < *p->MaxActiveFCLKChangeLatencySupported))) {
+			FoundCriticalSurface = true;
+			*p->MaxActiveFCLKChangeLatencySupported = s->ActiveFCLKChangeLatencyMargin[k] + p->mmSOCParameters.FCLKChangeLatency;
+		}
+	}
+
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: DRAMClockChangeSupport = %u\n", __func__, *p->global_dram_clock_change_supported);
+	DML_LOG_VERBOSE("DML::%s: FCLKChangeSupport = %u\n", __func__, *p->global_fclk_change_supported);
+	DML_LOG_VERBOSE("DML::%s: MaxActiveFCLKChangeLatencySupported = %f\n", __func__, *p->MaxActiveFCLKChangeLatencySupported);
+	DML_LOG_VERBOSE("DML::%s: USRRetrainingSupport = %u\n", __func__, *p->USRRetrainingSupport);
+#endif
+}
+
+static void calculate_bytes_to_fetch_required_to_hide_latency(
+		struct dml2_core_calcs_calculate_bytes_to_fetch_required_to_hide_latency_params *p)
+{
+	unsigned int dst_lines_to_hide;
+	unsigned int src_lines_to_hide_l;
+	unsigned int src_lines_to_hide_c;
+	unsigned int plane_index;
+	unsigned int stream_index;
+
+	for (plane_index = 0; plane_index < p->num_active_planes; plane_index++) {
+		if (dml_is_phantom_pipe(&p->display_cfg->plane_descriptors[plane_index]))
+			continue;
+
+		stream_index = p->display_cfg->plane_descriptors[plane_index].stream_index;
+
+		dst_lines_to_hide = (unsigned int)math_ceil(p->latency_to_hide_us[0] /
+			((double)p->display_cfg->stream_descriptors[stream_index].timing.h_total /
+				(double)p->display_cfg->stream_descriptors[stream_index].timing.pixel_clock_khz * 1000.0));
+
+		src_lines_to_hide_l = (unsigned int)math_ceil2(p->display_cfg->plane_descriptors[plane_index].composition.scaler_info.plane0.v_ratio * dst_lines_to_hide,
+			p->swath_height_l[plane_index]);
+		p->bytes_required_l[plane_index] = src_lines_to_hide_l * p->num_of_dpp[plane_index] * p->swath_width_l[plane_index] * p->byte_per_pix_l[plane_index];
+
+		src_lines_to_hide_c = (unsigned int)math_ceil2(p->display_cfg->plane_descriptors[plane_index].composition.scaler_info.plane1.v_ratio * dst_lines_to_hide,
+			p->swath_height_c[plane_index]);
+		p->bytes_required_c[plane_index] = src_lines_to_hide_c * p->num_of_dpp[plane_index] * p->swath_width_c[plane_index] * p->byte_per_pix_c[plane_index];
+
+		if (p->display_cfg->plane_descriptors[plane_index].surface.dcc.enable && p->mrq_present) {
+			p->bytes_required_l[plane_index] += (unsigned int)math_ceil((double)src_lines_to_hide_l / p->meta_row_height_l[plane_index]) * p->meta_row_bytes_per_row_ub_l[plane_index];
+			if (p->meta_row_height_c[plane_index]) {
+				p->bytes_required_c[plane_index] += (unsigned int)math_ceil((double)src_lines_to_hide_c / p->meta_row_height_c[plane_index]) * p->meta_row_bytes_per_row_ub_c[plane_index];
+			}
+		}
+
+		if (p->display_cfg->gpuvm_enable == true) {
+			p->bytes_required_l[plane_index] += (unsigned int)math_ceil((double)src_lines_to_hide_l / p->dpte_row_height_l[plane_index]) * p->dpte_bytes_per_row_l[plane_index];
+			if (p->dpte_row_height_c[plane_index]) {
+				p->bytes_required_c[plane_index] += (unsigned int)math_ceil((double)src_lines_to_hide_c / p->dpte_row_height_c[plane_index]) * p->dpte_bytes_per_row_c[plane_index];
+			}
+		}
+	}
+}
+
+static noinline_for_stack void calculate_vactive_det_fill_latency(
+		const struct dml2_display_cfg *display_cfg,
+		unsigned int num_active_planes,
+		unsigned int bytes_required_l[],
+		unsigned int bytes_required_c[],
+		double dcc_dram_bw_nom_overhead_factor_p0[],
+		double dcc_dram_bw_nom_overhead_factor_p1[],
+		double surface_read_bw_l[],
+		double surface_read_bw_c[],
+		double (*surface_avg_vactive_required_bw)[dml2_core_internal_bw_max][DML2_MAX_PLANES],
+		double (*surface_peak_required_bw)[dml2_core_internal_bw_max][DML2_MAX_PLANES],
+		/* output */
+		double vactive_det_fill_delay_us[])
+{
+	double effective_excess_bandwidth;
+	double effective_excess_bandwidth_l;
+	double effective_excess_bandwidth_c;
+	double adj_factor;
+	unsigned int plane_index;
+	unsigned int soc_state;
+	unsigned int bw_type;
+
+	for (plane_index = 0; plane_index < num_active_planes; plane_index++) {
+		if (dml_is_phantom_pipe(&display_cfg->plane_descriptors[plane_index]))
+			continue;
+
+		vactive_det_fill_delay_us[plane_index] = 0.0;
+		for (soc_state = 0; soc_state < dml2_core_internal_soc_state_max; soc_state++) {
+			for (bw_type = 0; bw_type < dml2_core_internal_bw_max; bw_type++) {
+				effective_excess_bandwidth = (surface_peak_required_bw[soc_state][bw_type][plane_index] - surface_avg_vactive_required_bw[soc_state][bw_type][plane_index]);
+
+				/* luma */
+				adj_factor = bw_type == dml2_core_internal_bw_dram ? dcc_dram_bw_nom_overhead_factor_p0[plane_index] : 1.0;
+
+				effective_excess_bandwidth_l = effective_excess_bandwidth * surface_read_bw_l[plane_index] / (surface_read_bw_l[plane_index] + surface_read_bw_c[plane_index]) / adj_factor;
+				if (effective_excess_bandwidth_l > 0.0) {
+					vactive_det_fill_delay_us[plane_index] = math_max2(vactive_det_fill_delay_us[plane_index], bytes_required_l[plane_index] / effective_excess_bandwidth_l);
+				}
+
+				/* chroma */
+				adj_factor = bw_type == dml2_core_internal_bw_dram ? dcc_dram_bw_nom_overhead_factor_p1[plane_index] : 1.0;
+
+				effective_excess_bandwidth_c = effective_excess_bandwidth * surface_read_bw_c[plane_index] / (surface_read_bw_l[plane_index] + surface_read_bw_c[plane_index]) / adj_factor;
+				if (effective_excess_bandwidth_c > 0.0) {
+					vactive_det_fill_delay_us[plane_index] = math_max2(vactive_det_fill_delay_us[plane_index], bytes_required_c[plane_index] / effective_excess_bandwidth_c);
+				}
+			}
+		}
+	}
+}
+
+static void calculate_excess_vactive_bandwidth_required(
+	const struct dml2_display_cfg *display_cfg,
+	unsigned int num_active_planes,
+	unsigned int bytes_required_l[],
+	unsigned int bytes_required_c[],
+	/* outputs */
+	double excess_vactive_fill_bw_l[],
+	double excess_vactive_fill_bw_c[])
+{
+	unsigned int plane_index;
+
+	for (plane_index = 0; plane_index < num_active_planes; plane_index++) {
+		if (dml_is_phantom_pipe(&display_cfg->plane_descriptors[plane_index]))
+			continue;
+
+		excess_vactive_fill_bw_l[plane_index] = 0.0;
+		excess_vactive_fill_bw_c[plane_index] = 0.0;
+
+		if (display_cfg->plane_descriptors[plane_index].overrides.max_vactive_det_fill_delay_us[dml2_pstate_type_uclk] > 0) {
+			excess_vactive_fill_bw_l[plane_index] = (double)bytes_required_l[plane_index] / (double)display_cfg->plane_descriptors[plane_index].overrides.max_vactive_det_fill_delay_us[dml2_pstate_type_uclk];
+			excess_vactive_fill_bw_c[plane_index] = (double)bytes_required_c[plane_index] / (double)display_cfg->plane_descriptors[plane_index].overrides.max_vactive_det_fill_delay_us[dml2_pstate_type_uclk];
+		}
+	}
+}
+
+static double uclk_khz_to_dram_bw_mbps(unsigned long uclk_khz, const struct dml2_dram_params *dram_config)
+{
+	double bw_mbps = 0;
+	bw_mbps = ((double)uclk_khz * dram_config->channel_count * dram_config->channel_width_bytes * dram_config->transactions_per_clock) / 1000.0;
+
+	return bw_mbps;
+}
+
+static double dram_bw_kbps_to_uclk_mhz(unsigned long long bw_kbps, const struct dml2_dram_params *dram_config)
+{
+	double uclk_mhz = 0;
+
+	uclk_mhz = (double)bw_kbps / (dram_config->channel_count * dram_config->channel_width_bytes * dram_config->transactions_per_clock) / 1000.0;
+
+	return uclk_mhz;
+}
+
+static unsigned int get_qos_param_index(unsigned long uclk_freq_khz, const struct dml2_dcn4_uclk_dpm_dependent_qos_params *per_uclk_dpm_params)
+{
+	unsigned int i;
+	unsigned int index = 0;
+
+	for (i = 0; i < DML_MAX_CLK_TABLE_SIZE; i++) {
+		DML_LOG_VERBOSE("DML::%s: per_uclk_dpm_params[%d].minimum_uclk_khz = %ld\n", __func__, i, per_uclk_dpm_params[i].minimum_uclk_khz);
+
+		if (i == 0)
+			index = 0;
+		else
+			index = i - 1;
+
+		if (uclk_freq_khz < per_uclk_dpm_params[i].minimum_uclk_khz ||
+			per_uclk_dpm_params[i].minimum_uclk_khz == 0) {
+			break;
+		}
+	}
+	DML_LOG_VERBOSE("DML::%s: uclk_freq_khz = %ld\n", __func__, uclk_freq_khz);
+	DML_LOG_VERBOSE("DML::%s: index = %d\n", __func__, index);
+	return index;
+}
+
+static unsigned int get_active_min_uclk_dpm_index(unsigned long uclk_freq_khz, const struct dml2_soc_state_table *clk_table)
+{
+	unsigned int i;
+	bool clk_entry_found = false;
+
+	for (i = 0; i < clk_table->uclk.num_clk_values; i++) {
+		DML_LOG_VERBOSE("DML::%s: clk_table.uclk.clk_values_khz[%d] = %ld\n", __func__, i, clk_table->uclk.clk_values_khz[i]);
+
+		if (uclk_freq_khz == clk_table->uclk.clk_values_khz[i]) {
+			clk_entry_found = true;
+			break;
+		}
+	}
+
+	if (!clk_entry_found)
+		DML_ASSERT(clk_entry_found);
+#if defined(__DML_VBA_DEBUG__)
+	DML_LOG_VERBOSE("DML::%s: uclk_freq_khz = %ld\n", __func__, uclk_freq_khz);
+	DML_LOG_VERBOSE("DML::%s: index = %d\n", __func__, i);
+#endif
+	return i;
+}
+
+static unsigned int get_pipe_flip_bytes(
+		double hostvm_inefficiency_factor,
+		unsigned int vm_bytes,
+		unsigned int dpte_row_bytes,
+		unsigned int meta_row_bytes)
+{
+	unsigned int flip_bytes = 0;
+
+	flip_bytes += (unsigned int) ((vm_bytes * hostvm_inefficiency_factor) + 2*meta_row_bytes);
+	flip_bytes += (unsigned int) (2*dpte_row_bytes * hostvm_inefficiency_factor);
+
+	return flip_bytes;
+}
+
+static void calculate_hostvm_inefficiency_factor(
+		double *HostVMInefficiencyFactor,
+		double *HostVMInefficiencyFactorPrefetch,
+
+		bool gpuvm_enable,
+		bool hostvm_enable,
+		unsigned int remote_iommu_outstanding_translations,
+		unsigned int max_outstanding_reqs,
+		double urg_bandwidth_avail_active_pixel_and_vm,
+		double urg_bandwidth_avail_active_vm_only)
+{
+		*HostVMInefficiencyFactor = 1;
+		*HostVMInefficiencyFactorPrefetch = 1;
+
+		if (gpuvm_enable && hostvm_enable) {
+			*HostVMInefficiencyFactor = urg_bandwidth_avail_active_pixel_and_vm / urg_bandwidth_avail_active_vm_only;
+			*HostVMInefficiencyFactorPrefetch = *HostVMInefficiencyFactor;
+
+			if ((*HostVMInefficiencyFactorPrefetch < 4) && (remote_iommu_outstanding_translations < max_outstanding_reqs))
+				*HostVMInefficiencyFactorPrefetch = 4;
+#ifdef __DML_VBA_DEBUG__
+			DML_LOG_VERBOSE("DML::%s: urg_bandwidth_avail_active_pixel_and_vm = %f\n", __func__, urg_bandwidth_avail_active_pixel_and_vm);
+			DML_LOG_VERBOSE("DML::%s: urg_bandwidth_avail_active_vm_only = %f\n", __func__, urg_bandwidth_avail_active_vm_only);
+			DML_LOG_VERBOSE("DML::%s: HostVMInefficiencyFactor = %f\n", __func__, *HostVMInefficiencyFactor);
+			DML_LOG_VERBOSE("DML::%s: HostVMInefficiencyFactorPrefetch = %f\n", __func__, *HostVMInefficiencyFactorPrefetch);
+#endif
+		}
+}
+
+struct dml2_core_internal_g6_temp_read_blackouts_table {
+	struct {
+		unsigned int uclk_khz;
+		unsigned int blackout_us;
+	} entries[DML_MAX_CLK_TABLE_SIZE];
+};
+
+struct dml2_core_internal_g6_temp_read_blackouts_table core_dcn4_g6_temp_read_blackout_table = {
+	.entries = {
+		{
+			.uclk_khz = 96000,
+			.blackout_us = 23,
+		},
+		{
+			.uclk_khz = 435000,
+			.blackout_us = 10,
+		},
+		{
+			.uclk_khz = 521000,
+			.blackout_us = 10,
+		},
+		{
+			.uclk_khz = 731000,
+			.blackout_us = 8,
+		},
+		{
+			.uclk_khz = 822000,
+			.blackout_us = 8,
+		},
+		{
+			.uclk_khz = 962000,
+			.blackout_us = 5,
+		},
+		{
+			.uclk_khz = 1069000,
+			.blackout_us = 5,
+		},
+		{
+			.uclk_khz = 1187000,
+			.blackout_us = 5,
+		},
+	},
+};
+
+static double get_g6_temp_read_blackout_us(
+	struct dml2_soc_bb *soc,
+	unsigned int uclk_freq_khz,
+	unsigned int min_clk_index)
+{
+	unsigned int i;
+	unsigned int blackout_us = core_dcn4_g6_temp_read_blackout_table.entries[0].blackout_us;
+
+	if (soc->power_management_parameters.g6_temp_read_blackout_us[0] > 0.0) {
+		/* overrides are present in the SoC BB */
+		return soc->power_management_parameters.g6_temp_read_blackout_us[min_clk_index];
+	}
+
+	/* use internal table */
+	blackout_us = core_dcn4_g6_temp_read_blackout_table.entries[0].blackout_us;
+
+	for (i = 0; i < DML_MAX_CLK_TABLE_SIZE; i++) {
+		if (uclk_freq_khz < core_dcn4_g6_temp_read_blackout_table.entries[i].uclk_khz ||
+			core_dcn4_g6_temp_read_blackout_table.entries[i].uclk_khz == 0) {
+			break;
+		}
+
+		blackout_us = core_dcn4_g6_temp_read_blackout_table.entries[i].blackout_us;
+	}
+
+	return (double)blackout_us;
+}
+
+static double get_max_urgent_latency_us(
+	struct dml2_dcn4x_soc_qos_params *dcn4x,
+	double uclk_freq_mhz,
+	double FabricClock,
+	unsigned int min_clk_index)
+{
+	double latency;
+	latency = dcn4x->per_uclk_dpm_params[min_clk_index].maximum_latency_when_urgent_uclk_cycles / uclk_freq_mhz
+		* (1 + dcn4x->umc_max_latency_margin / 100.0)
+		+ dcn4x->mall_overhead_fclk_cycles / FabricClock
+		+ dcn4x->max_round_trip_to_furthest_cs_fclk_cycles / FabricClock
+		* (1 + dcn4x->fabric_max_transport_latency_margin / 100.0);
+	return latency;
+}
+
+static void calculate_pstate_keepout_dst_lines(
+		const struct dml2_display_cfg *display_cfg,
+		const struct dml2_core_internal_watermarks *watermarks,
+		unsigned int pstate_keepout_dst_lines[])
+{
+	const struct dml2_stream_parameters *stream_descriptor;
+	unsigned int i;
+
+	for (i = 0; i < display_cfg->num_planes; i++) {
+		if (!dml_is_phantom_pipe(&display_cfg->plane_descriptors[i])) {
+			stream_descriptor = &display_cfg->stream_descriptors[display_cfg->plane_descriptors[i].stream_index];
+
+			pstate_keepout_dst_lines[i] =
+					(unsigned int)math_ceil(watermarks->DRAMClockChangeWatermark / ((double)stream_descriptor->timing.h_total * 1000.0 / (double)stream_descriptor->timing.pixel_clock_khz));
+
+			if (pstate_keepout_dst_lines[i] > stream_descriptor->timing.v_total - 1) {
+				pstate_keepout_dst_lines[i] = stream_descriptor->timing.v_total - 1;
+			}
+		}
+	}
+}
+
+static noinline_for_stack void dml_core_ms_prefetch_check(struct dml2_core_internal_display_mode_lib *mode_lib,
+	const struct dml2_display_cfg *display_cfg)
+{
+	struct dml2_core_calcs_mode_support_locals *s = &mode_lib->scratch.dml_core_mode_support_locals;
+	struct dml2_core_calcs_calculate_tdlut_setting_params *calculate_tdlut_setting_params = &mode_lib->scratch.calculate_tdlut_setting_params;
+	struct dml2_core_calcs_CalculatePrefetchSchedule_params *CalculatePrefetchSchedule_params = &mode_lib->scratch.CalculatePrefetchSchedule_params;
+	struct dml2_core_calcs_calculate_peak_bandwidth_required_params *calculate_peak_bandwidth_params = &mode_lib->scratch.calculate_peak_bandwidth_params;
+#ifdef DML_GLOBAL_PREFETCH_CHECK
+	struct dml2_core_calcs_CheckGlobalPrefetchAdmissibility_params *CheckGlobalPrefetchAdmissibility_params = &mode_lib->scratch.CheckGlobalPrefetchAdmissibility_params;
+#endif
+	struct dml2_core_calcs_CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_params *CalculateWatermarks_params = &mode_lib->scratch.CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_params;
+
+	double min_return_bw_for_latency;
+	unsigned int k;
+
+	mode_lib->ms.TimeCalc = 24 / mode_lib->ms.dcfclk_deepsleep;
+
+	calculate_hostvm_inefficiency_factor(
+			&s->HostVMInefficiencyFactor,
+			&s->HostVMInefficiencyFactorPrefetch,
+
+			display_cfg->gpuvm_enable,
+			display_cfg->hostvm_enable,
+			mode_lib->ip.remote_iommu_outstanding_translations,
+			mode_lib->soc.max_outstanding_reqs,
+			mode_lib->ms.support.urg_bandwidth_available_pixel_and_vm[dml2_core_internal_soc_state_sys_active],
+			mode_lib->ms.support.urg_bandwidth_available_vm_only[dml2_core_internal_soc_state_sys_active]);
+
+	mode_lib->ms.Total3dlutActive = 0;
+	for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+		if (display_cfg->plane_descriptors[k].tdlut.setup_for_tdlut)
+			mode_lib->ms.Total3dlutActive = mode_lib->ms.Total3dlutActive + 1;
+
+		// Calculate tdlut schedule related terms
+		calculate_tdlut_setting_params->dispclk_mhz = mode_lib->ms.RequiredDISPCLK;
+		calculate_tdlut_setting_params->setup_for_tdlut = display_cfg->plane_descriptors[k].tdlut.setup_for_tdlut;
+		calculate_tdlut_setting_params->tdlut_width_mode = display_cfg->plane_descriptors[k].tdlut.tdlut_width_mode;
+		calculate_tdlut_setting_params->tdlut_addressing_mode = display_cfg->plane_descriptors[k].tdlut.tdlut_addressing_mode;
+		calculate_tdlut_setting_params->cursor_buffer_size = mode_lib->ip.cursor_buffer_size;
+		calculate_tdlut_setting_params->gpuvm_enable = display_cfg->gpuvm_enable;
+		calculate_tdlut_setting_params->gpuvm_page_size_kbytes = display_cfg->plane_descriptors[k].overrides.gpuvm_min_page_size_kbytes;
+		calculate_tdlut_setting_params->tdlut_mpc_width_flag = display_cfg->plane_descriptors[k].tdlut.tdlut_mpc_width_flag;
+		calculate_tdlut_setting_params->is_gfx11 = dml_get_gfx_version(display_cfg->plane_descriptors[k].surface.tiling);
+
+		// output
+		calculate_tdlut_setting_params->tdlut_pte_bytes_per_frame = &s->tdlut_pte_bytes_per_frame[k];
+		calculate_tdlut_setting_params->tdlut_bytes_per_frame = &s->tdlut_bytes_per_frame[k];
+		calculate_tdlut_setting_params->tdlut_groups_per_2row_ub = &s->tdlut_groups_per_2row_ub[k];
+		calculate_tdlut_setting_params->tdlut_opt_time = &s->tdlut_opt_time[k];
+		calculate_tdlut_setting_params->tdlut_drain_time = &s->tdlut_drain_time[k];
+		calculate_tdlut_setting_params->tdlut_bytes_to_deliver = &s->tdlut_bytes_to_deliver[k];
+		calculate_tdlut_setting_params->tdlut_bytes_per_group = &s->tdlut_bytes_per_group[k];
+
+		calculate_tdlut_setting(&mode_lib->scratch, calculate_tdlut_setting_params);
+	}
+
+	min_return_bw_for_latency = mode_lib->ms.support.urg_bandwidth_available_min_latency[dml2_core_internal_soc_state_sys_active];
+
+	if (mode_lib->soc.qos_parameters.qos_type == dml2_qos_param_type_dcn3)
+		s->ReorderingBytes = (unsigned int)(mode_lib->soc.clk_table.dram_config.channel_count * math_max3(mode_lib->soc.qos_parameters.qos_params.dcn32x.urgent_out_of_order_return_per_channel_pixel_only_bytes,
+										mode_lib->soc.qos_parameters.qos_params.dcn32x.urgent_out_of_order_return_per_channel_pixel_and_vm_bytes,
+										mode_lib->soc.qos_parameters.qos_params.dcn32x.urgent_out_of_order_return_per_channel_vm_only_bytes));
+
+	CalculateExtraLatency(
+		display_cfg,
+		mode_lib->ip.rob_buffer_size_kbytes,
+		mode_lib->soc.qos_parameters.qos_params.dcn32x.loaded_round_trip_latency_fclk_cycles,
+		s->ReorderingBytes,
+		mode_lib->ms.DCFCLK,
+		mode_lib->ms.FabricClock,
+		mode_lib->ip.pixel_chunk_size_kbytes,
+		min_return_bw_for_latency,
+		mode_lib->ms.num_active_planes,
+		mode_lib->ms.NoOfDPP,
+		mode_lib->ms.dpte_group_bytes,
+		s->tdlut_bytes_per_group,
+		s->HostVMInefficiencyFactor,
+		s->HostVMInefficiencyFactorPrefetch,
+		mode_lib->soc.hostvm_min_page_size_kbytes,
+		mode_lib->soc.qos_parameters.qos_type,
+		!(display_cfg->overrides.max_outstanding_when_urgent_expected_disable),
+		mode_lib->soc.max_outstanding_reqs,
+		mode_lib->ms.support.request_size_bytes_luma,
+		mode_lib->ms.support.request_size_bytes_chroma,
+		mode_lib->ip.meta_chunk_size_kbytes,
+		mode_lib->ip.dchub_arb_to_ret_delay,
+		mode_lib->ms.TripToMemory,
+		mode_lib->ip.hostvm_mode,
+
+		// output
+		&mode_lib->ms.ExtraLatency,
+		&mode_lib->ms.ExtraLatency_sr,
+		&mode_lib->ms.ExtraLatencyPrefetch);
+
+	for (k = 0; k < mode_lib->ms.num_active_planes; k++)
+		s->impacted_dst_y_pre[k] = 0;
+
+	s->recalc_prefetch_schedule = 0;
+	s->recalc_prefetch_done = 0;
+	do {
+		mode_lib->ms.support.PrefetchSupported = true;
+
+		for (k = 0; k < mode_lib->ms.num_active_planes; k++) {
+			s->line_times[k] = display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000);
+			s->pixel_format[k] = display_cfg->plane_descriptors[k].pixel_format;
+
+			s->lb_source_lines_l[k] = get_num_lb_source_lines(mode_lib->ip.max_line_buffer_lines, mode_lib->ip.line_buffer_size_bits,
+																mode_lib->ms.NoOfDPP[k],
+																display_cfg->plane_descriptors[k].composition.viewport.plane0.width,
+																display_cfg->plane_descriptors[k].composition.viewport.plane0.height,
+																display_cfg->plane_descriptors[k].composition.scaler_info.plane0.h_ratio,
+																display_cfg->plane_descriptors[k].composition.rotation_angle);
+
+			s->lb_source_lines_c[k] = get_num_lb_source_lines(mode_lib->ip.max_line_buffer_lines, mode_lib->ip.line_buffer_size_bits,
+																mode_lib->ms.NoOfDPP[k],
+																display_cfg->plane_descriptors[k].composition.viewport.plane1.width,
+																display_cfg->plane_descriptors[k].composition.viewport.plane1.height,
+																display_cfg->plane_descriptors[k].composition.scaler_info.plane1.h_ratio,
+																display_cfg->plane_descriptors[k].composition.rotation_angle);
+
+			struct dml2_core_internal_DmlPipe *myPipe = &s->myPipe;
+
+			mode_lib->ms.TWait[k] = CalculateTWait(
+				display_cfg->plane_descriptors[k].overrides.reserved_vblank_time_ns,
+				mode_lib->ms.UrgLatency,
+				mode_lib->ms.TripToMemory,
+				!dml_is_phantom_pipe(&display_cfg->plane_descriptors[k]) && display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.drr_config.enabled ?
+				get_g6_temp_read_blackout_us(&mode_lib->soc, (unsigned int)(mode_lib->ms.uclk_freq_mhz * 1000), mode_lib->ms.state_idx) : 0.0);
+
+			myPipe->Dppclk = mode_lib->ms.RequiredDPPCLK[k];
+			myPipe->Dispclk = mode_lib->ms.RequiredDISPCLK;
+			myPipe->PixelClock = ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000);
+			myPipe->DCFClkDeepSleep = mode_lib->ms.dcfclk_deepsleep;
+			myPipe->DPPPerSurface = mode_lib->ms.NoOfDPP[k];
+			myPipe->ScalerEnabled = display_cfg->plane_descriptors[k].composition.scaler_info.enabled;
+			myPipe->VRatio = display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio;
+			myPipe->VRatioChroma = display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_ratio;
+			myPipe->VTaps = display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_taps;
+			myPipe->VTapsChroma = display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_taps;
+			myPipe->RotationAngle = display_cfg->plane_descriptors[k].composition.rotation_angle;
+			myPipe->mirrored = display_cfg->plane_descriptors[k].composition.mirrored;
+			myPipe->BlockWidth256BytesY = mode_lib->ms.Read256BlockWidthY[k];
+			myPipe->BlockHeight256BytesY = mode_lib->ms.Read256BlockHeightY[k];
+			myPipe->BlockWidth256BytesC = mode_lib->ms.Read256BlockWidthC[k];
+			myPipe->BlockHeight256BytesC = mode_lib->ms.Read256BlockHeightC[k];
+			myPipe->InterlaceEnable = display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.interlaced;
+			myPipe->NumberOfCursors = display_cfg->plane_descriptors[k].cursor.num_cursors;
+			myPipe->VBlank = display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.v_total - display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.v_active;
+			myPipe->HTotal = display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total;
+			myPipe->HActive = display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_active;
+			myPipe->DCCEnable = display_cfg->plane_descriptors[k].surface.dcc.enable;
+			myPipe->ODMMode = mode_lib->ms.ODMMode[k];
+			myPipe->SourcePixelFormat = display_cfg->plane_descriptors[k].pixel_format;
+			myPipe->BytePerPixelY = mode_lib->ms.BytePerPixelY[k];
+			myPipe->BytePerPixelC = mode_lib->ms.BytePerPixelC[k];
+			myPipe->ProgressiveToInterlaceUnitInOPP = mode_lib->ip.ptoi_supported;
+
+#ifdef __DML_VBA_DEBUG__
+			DML_LOG_VERBOSE("DML::%s: Calling CalculatePrefetchSchedule for k=%u\n", __func__, k);
+			DML_LOG_VERBOSE("DML::%s: MaximumVStartup = %u\n", __func__, s->MaximumVStartup[k]);
+#endif
+			CalculatePrefetchSchedule_params->display_cfg = display_cfg;
+			CalculatePrefetchSchedule_params->HostVMInefficiencyFactor = s->HostVMInefficiencyFactorPrefetch;
+			CalculatePrefetchSchedule_params->myPipe = myPipe;
+			CalculatePrefetchSchedule_params->DSCDelay = mode_lib->ms.DSCDelay[k];
+			CalculatePrefetchSchedule_params->DPPCLKDelaySubtotalPlusCNVCFormater = mode_lib->ip.dppclk_delay_subtotal + mode_lib->ip.dppclk_delay_cnvc_formatter;
+			CalculatePrefetchSchedule_params->DPPCLKDelaySCL = mode_lib->ip.dppclk_delay_scl;
+			CalculatePrefetchSchedule_params->DPPCLKDelaySCLLBOnly = mode_lib->ip.dppclk_delay_scl_lb_only;
+			CalculatePrefetchSchedule_params->DPPCLKDelayCNVCCursor = mode_lib->ip.dppclk_delay_cnvc_cursor;
+			CalculatePrefetchSchedule_params->DISPCLKDelaySubtotal = mode_lib->ip.dispclk_delay_subtotal;
+			CalculatePrefetchSchedule_params->DPP_RECOUT_WIDTH = (unsigned int)(mode_lib->ms.SwathWidthY[k] / display_cfg->plane_descriptors[k].composition.scaler_info.plane0.h_ratio);
+			CalculatePrefetchSchedule_params->OutputFormat = display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_format;
+			CalculatePrefetchSchedule_params->MaxInterDCNTileRepeaters = mode_lib->ip.max_inter_dcn_tile_repeaters;
+			CalculatePrefetchSchedule_params->VStartup = s->MaximumVStartup[k];
+			CalculatePrefetchSchedule_params->HostVMMinPageSize = mode_lib->soc.hostvm_min_page_size_kbytes;
+			CalculatePrefetchSchedule_params->DynamicMetadataEnable = display_cfg->plane_descriptors[k].dynamic_meta_data.enable;
+			CalculatePrefetchSchedule_params->DynamicMetadataVMEnabled = mode_lib->ip.dynamic_metadata_vm_enabled;
+			CalculatePrefetchSchedule_params->DynamicMetadataLinesBeforeActiveRequired = display_cfg->plane_descriptors[k].dynamic_meta_data.lines_before_active_required;
+			CalculatePrefetchSchedule_params->DynamicMetadataTransmittedBytes = display_cfg->plane_descriptors[k].dynamic_meta_data.transmitted_bytes;
+			CalculatePrefetchSchedule_params->UrgentLatency = mode_lib->ms.UrgLatency;
+			CalculatePrefetchSchedule_params->ExtraLatencyPrefetch = mode_lib->ms.ExtraLatencyPrefetch;
+			CalculatePrefetchSchedule_params->TCalc = mode_lib->ms.TimeCalc;
+			CalculatePrefetchSchedule_params->vm_bytes = mode_lib->ms.vm_bytes[k];
+			CalculatePrefetchSchedule_params->PixelPTEBytesPerRow = mode_lib->ms.DPTEBytesPerRow[k];
+			CalculatePrefetchSchedule_params->PrefetchSourceLinesY = mode_lib->ms.PrefetchLinesY[k];
+			CalculatePrefetchSchedule_params->VInitPreFillY = mode_lib->ms.PrefillY[k];
+			CalculatePrefetchSchedule_params->MaxNumSwathY = mode_lib->ms.MaxNumSwathY[k];
+			CalculatePrefetchSchedule_params->PrefetchSourceLinesC = mode_lib->ms.PrefetchLinesC[k];
+			CalculatePrefetchSchedule_params->VInitPreFillC = mode_lib->ms.PrefillC[k];
+			CalculatePrefetchSchedule_params->MaxNumSwathC = mode_lib->ms.MaxNumSwathC[k];
+			CalculatePrefetchSchedule_params->swath_width_luma_ub = mode_lib->ms.swath_width_luma_ub[k];
+			CalculatePrefetchSchedule_params->swath_width_chroma_ub = mode_lib->ms.swath_width_chroma_ub[k];
+			CalculatePrefetchSchedule_params->SwathHeightY = mode_lib->ms.SwathHeightY[k];
+			CalculatePrefetchSchedule_params->SwathHeightC = mode_lib->ms.SwathHeightC[k];
+			CalculatePrefetchSchedule_params->TWait = mode_lib->ms.TWait[k];
+			CalculatePrefetchSchedule_params->Ttrip = mode_lib->ms.TripToMemory;
+			CalculatePrefetchSchedule_params->Turg = mode_lib->ms.UrgLatency;
+			CalculatePrefetchSchedule_params->setup_for_tdlut = display_cfg->plane_descriptors[k].tdlut.setup_for_tdlut;
+			CalculatePrefetchSchedule_params->tdlut_pte_bytes_per_frame = s->tdlut_pte_bytes_per_frame[k];
+			CalculatePrefetchSchedule_params->tdlut_bytes_per_frame = s->tdlut_bytes_per_frame[k];
+			CalculatePrefetchSchedule_params->tdlut_opt_time = s->tdlut_opt_time[k];
+			CalculatePrefetchSchedule_params->tdlut_drain_time = s->tdlut_drain_time[k];
+			CalculatePrefetchSchedule_params->num_cursors = (display_cfg->plane_descriptors[k].cursor.cursor_width > 0);
+			CalculatePrefetchSchedule_params->cursor_bytes_per_chunk = s->cursor_bytes_per_chunk[k];
+			CalculatePrefetchSchedule_params->cursor_bytes_per_line = s->cursor_bytes_per_line[k];
+			CalculatePrefetchSchedule_params->dcc_enable = display_cfg->plane_descriptors[k].surface.dcc.enable;
+			CalculatePrefetchSchedule_params->mrq_present = mode_lib->ip.dcn_mrq_present;
+			CalculatePrefetchSchedule_params->meta_row_bytes = mode_lib->ms.meta_row_bytes[k];
+			CalculatePrefetchSchedule_params->mall_prefetch_sdp_overhead_factor = mode_lib->ms.mall_prefetch_sdp_overhead_factor[k];
+			CalculatePrefetchSchedule_params->impacted_dst_y_pre = s->impacted_dst_y_pre[k];
+			CalculatePrefetchSchedule_params->vactive_sw_bw_l = mode_lib->ms.vactive_sw_bw_l[k];
+			CalculatePrefetchSchedule_params->vactive_sw_bw_c = mode_lib->ms.vactive_sw_bw_c[k];
+
+			// output
+			CalculatePrefetchSchedule_params->DSTXAfterScaler = &s->DSTXAfterScaler[k];
+			CalculatePrefetchSchedule_params->DSTYAfterScaler = &s->DSTYAfterScaler[k];
+			CalculatePrefetchSchedule_params->dst_y_prefetch = &mode_lib->ms.dst_y_prefetch[k];
+			CalculatePrefetchSchedule_params->dst_y_per_vm_vblank = &mode_lib->ms.LinesForVM[k];
+			CalculatePrefetchSchedule_params->dst_y_per_row_vblank = &mode_lib->ms.LinesForDPTERow[k];
+			CalculatePrefetchSchedule_params->VRatioPrefetchY = &mode_lib->ms.VRatioPreY[k];
+			CalculatePrefetchSchedule_params->VRatioPrefetchC = &mode_lib->ms.VRatioPreC[k];
+			CalculatePrefetchSchedule_params->RequiredPrefetchPixelDataBWLuma = &mode_lib->ms.RequiredPrefetchPixelDataBWLuma[k]; // prefetch_sw_bw_l
+			CalculatePrefetchSchedule_params->RequiredPrefetchPixelDataBWChroma = &mode_lib->ms.RequiredPrefetchPixelDataBWChroma[k]; // prefetch_sw_bw_c
+			CalculatePrefetchSchedule_params->RequiredPrefetchBWMax = &mode_lib->ms.RequiredPrefetchBWMax[k];
+			CalculatePrefetchSchedule_params->NotEnoughTimeForDynamicMetadata = &mode_lib->ms.NoTimeForDynamicMetadata[k];
+			CalculatePrefetchSchedule_params->Tno_bw = &mode_lib->ms.Tno_bw[k];
+			CalculatePrefetchSchedule_params->Tno_bw_flip = &mode_lib->ms.Tno_bw_flip[k];
+			CalculatePrefetchSchedule_params->prefetch_vmrow_bw = &mode_lib->ms.prefetch_vmrow_bw[k];
+			CalculatePrefetchSchedule_params->Tdmdl_vm = &s->dummy_single[0];
+			CalculatePrefetchSchedule_params->Tdmdl = &s->dummy_single[1];
+			CalculatePrefetchSchedule_params->TSetup = &s->dummy_single[2];
+			CalculatePrefetchSchedule_params->Tvm_trips = &s->Tvm_trips[k];
+			CalculatePrefetchSchedule_params->Tr0_trips = &s->Tr0_trips[k];
+			CalculatePrefetchSchedule_params->Tvm_trips_flip = &s->Tvm_trips_flip[k];
+			CalculatePrefetchSchedule_params->Tr0_trips_flip = &s->Tr0_trips_flip[k];
+			CalculatePrefetchSchedule_params->Tvm_trips_flip_rounded = &s->Tvm_trips_flip_rounded[k];
+			CalculatePrefetchSchedule_params->Tr0_trips_flip_rounded = &s->Tr0_trips_flip_rounded[k];
+			CalculatePrefetchSchedule_params->VUpdateOffsetPix = &s->dummy_integer[0];
+			CalculatePrefetchSchedule_params->VUpdateWidthPix = &s->dummy_integer[1];
+			CalculatePrefetchSchedule_params->VReadyOffsetPix = &s->dummy_integer[2];
+			CalculatePrefetchSchedule_params->prefetch_cursor_bw = &mode_lib->ms.prefetch_cursor_bw[k];
+			CalculatePrefetchSchedule_params->prefetch_sw_bytes = &s->prefetch_sw_bytes[k];
+			CalculatePrefetchSchedule_params->Tpre_rounded = &s->Tpre_rounded[k];
+			CalculatePrefetchSchedule_params->Tpre_oto = &s->Tpre_oto[k];
+			CalculatePrefetchSchedule_params->prefetch_swath_time_us = &s->prefetch_swath_time_us[k];
+
+			mode_lib->ms.NoTimeForPrefetch[k] = CalculatePrefetchSchedule(&mode_lib->scratch, CalculatePrefetchSchedule_params);
+
+			mode_lib->ms.support.PrefetchSupported &= !mode_lib->ms.NoTimeForPrefetch[k];
+			DML_LOG_VERBOSE("DML::%s: k=%d, dst_y_per_vm_vblank = %f\n", __func__, k, *CalculatePrefetchSchedule_params->dst_y_per_vm_vblank);
+			DML_LOG_VERBOSE("DML::%s: k=%d, dst_y_per_row_vblank = %f\n", __func__, k, *CalculatePrefetchSchedule_params->dst_y_per_row_vblank);
+		} // for k num_planes
+
+		CalculateDCFCLKDeepSleepTdlut(
+			display_cfg,
+			mode_lib->ms.num_active_planes,
+			mode_lib->ms.BytePerPixelY,
+			mode_lib->ms.BytePerPixelC,
+			mode_lib->ms.SwathWidthY,
+			mode_lib->ms.SwathWidthC,
+			mode_lib->ms.NoOfDPP,
+			mode_lib->ms.PSCL_FACTOR,
+			mode_lib->ms.PSCL_FACTOR_CHROMA,
+			mode_lib->ms.RequiredDPPCLK,
+			mode_lib->ms.vactive_sw_bw_l,
+			mode_lib->ms.vactive_sw_bw_c,
+			mode_lib->soc.return_bus_width_bytes,
+			mode_lib->ms.RequiredDISPCLK,
+			s->tdlut_bytes_to_deliver,
+			s->prefetch_swath_time_us,
+
+			/* Output */
+			&mode_lib->ms.dcfclk_deepsleep);
+
+		for (k = 0; k < mode_lib->ms.num_active_planes; k++) {
+			if (mode_lib->ms.dst_y_prefetch[k] < 2.0
+				|| mode_lib->ms.LinesForVM[k] >= 32.0
+				|| mode_lib->ms.LinesForDPTERow[k] >= 16.0
+				|| mode_lib->ms.NoTimeForPrefetch[k] == true
+				|| s->DSTYAfterScaler[k] > 8) {
+				mode_lib->ms.support.PrefetchSupported = false;
+				DML_LOG_VERBOSE("DML::%s: k=%d, dst_y_prefetch=%f (should not be < 2)\n", __func__, k, mode_lib->ms.dst_y_prefetch[k]);
+				DML_LOG_VERBOSE("DML::%s: k=%d, LinesForVM=%f (should not be >= 32)\n", __func__, k, mode_lib->ms.LinesForVM[k]);
+				DML_LOG_VERBOSE("DML::%s: k=%d, LinesForDPTERow=%f (should not be >= 16)\n", __func__, k, mode_lib->ms.LinesForDPTERow[k]);
+				DML_LOG_VERBOSE("DML::%s: k=%d, DSTYAfterScaler=%d (should be <= 8)\n", __func__, k, s->DSTYAfterScaler[k]);
+				DML_LOG_VERBOSE("DML::%s: k=%d, NoTimeForPrefetch=%d\n", __func__, k, mode_lib->ms.NoTimeForPrefetch[k]);
+			}
+		}
+
+		mode_lib->ms.support.DynamicMetadataSupported = true;
+		for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+			if (mode_lib->ms.NoTimeForDynamicMetadata[k] == true) {
+				mode_lib->ms.support.DynamicMetadataSupported = false;
+			}
+		}
+
+		mode_lib->ms.support.VRatioInPrefetchSupported = true;
+		for (k = 0; k < mode_lib->ms.num_active_planes; k++) {
+			if (mode_lib->ms.VRatioPreY[k] > __DML2_CALCS_MAX_VRATIO_PRE__ ||
+				mode_lib->ms.VRatioPreC[k] > __DML2_CALCS_MAX_VRATIO_PRE__) {
+				mode_lib->ms.support.VRatioInPrefetchSupported = false;
+				DML_LOG_VERBOSE("DML::%s: k=%d VRatioPreY = %f (should be <= %f)\n", __func__, k, mode_lib->ms.VRatioPreY[k], __DML2_CALCS_MAX_VRATIO_PRE__);
+				DML_LOG_VERBOSE("DML::%s: k=%d VRatioPreC = %f (should be <= %f)\n", __func__, k, mode_lib->ms.VRatioPreC[k], __DML2_CALCS_MAX_VRATIO_PRE__);
+				DML_LOG_VERBOSE("DML::%s: VRatioInPrefetchSupported = %u\n", __func__, mode_lib->ms.support.VRatioInPrefetchSupported);
+			}
+		}
+
+		mode_lib->ms.support.PrefetchSupported &= mode_lib->ms.support.VRatioInPrefetchSupported;
+
+		// By default, do not recalc prefetch schedule
+		s->recalc_prefetch_schedule = 0;
+
+		// Only do urg vs prefetch bandwidth check, flip schedule check, power saving feature support check IF the Prefetch Schedule Check is ok
+		if (mode_lib->ms.support.PrefetchSupported) {
+			for (k = 0; k < mode_lib->ms.num_active_planes; k++) {
+				// Calculate Urgent burst factor for prefetch
+#ifdef __DML_VBA_DEBUG__
+				DML_LOG_VERBOSE("DML::%s: k=%d, Calling CalculateUrgentBurstFactor (for prefetch)\n", __func__, k);
+				DML_LOG_VERBOSE("DML::%s: k=%d, VRatioPreY=%f\n", __func__, k, mode_lib->ms.VRatioPreY[k]);
+				DML_LOG_VERBOSE("DML::%s: k=%d, VRatioPreC=%f\n", __func__, k, mode_lib->ms.VRatioPreC[k]);
+#endif
+				CalculateUrgentBurstFactor(
+					&display_cfg->plane_descriptors[k],
+					mode_lib->ms.swath_width_luma_ub[k],
+					mode_lib->ms.swath_width_chroma_ub[k],
+					mode_lib->ms.SwathHeightY[k],
+					mode_lib->ms.SwathHeightC[k],
+					s->line_times[k],
+					mode_lib->ms.UrgLatency,
+					mode_lib->ms.VRatioPreY[k],
+					mode_lib->ms.VRatioPreC[k],
+					mode_lib->ms.BytePerPixelInDETY[k],
+					mode_lib->ms.BytePerPixelInDETC[k],
+					mode_lib->ms.DETBufferSizeY[k],
+					mode_lib->ms.DETBufferSizeC[k],
+					/* Output */
+					&mode_lib->ms.UrgentBurstFactorLumaPre[k],
+					&mode_lib->ms.UrgentBurstFactorChromaPre[k],
+					&mode_lib->ms.NotEnoughUrgentLatencyHidingPre[k]);
+			}
+
+			// Calculate urgent bandwidth required, both urg and non urg peak bandwidth
+			// assume flip bw is 0 at this point
+			for (k = 0; k < mode_lib->ms.num_active_planes; k++)
+				mode_lib->ms.final_flip_bw[k] = 0;
+
+			calculate_peak_bandwidth_params->urg_vactive_bandwidth_required = mode_lib->ms.support.urg_vactive_bandwidth_required;
+			calculate_peak_bandwidth_params->urg_bandwidth_required = mode_lib->ms.support.urg_bandwidth_required;
+			calculate_peak_bandwidth_params->urg_bandwidth_required_qual = mode_lib->ms.support.urg_bandwidth_required_qual;
+			calculate_peak_bandwidth_params->non_urg_bandwidth_required = mode_lib->ms.support.non_urg_bandwidth_required;
+			calculate_peak_bandwidth_params->surface_avg_vactive_required_bw = mode_lib->ms.surface_avg_vactive_required_bw;
+			calculate_peak_bandwidth_params->surface_peak_required_bw = mode_lib->ms.surface_peak_required_bw;
+
+			calculate_peak_bandwidth_params->display_cfg = display_cfg;
+			calculate_peak_bandwidth_params->inc_flip_bw = 0;
+			calculate_peak_bandwidth_params->num_active_planes =  mode_lib->ms.num_active_planes;
+			calculate_peak_bandwidth_params->num_of_dpp = mode_lib->ms.NoOfDPP;
+			calculate_peak_bandwidth_params->dcc_dram_bw_nom_overhead_factor_p0 = mode_lib->ms.dcc_dram_bw_nom_overhead_factor_p0;
+			calculate_peak_bandwidth_params->dcc_dram_bw_nom_overhead_factor_p1 = mode_lib->ms.dcc_dram_bw_nom_overhead_factor_p1;
+			calculate_peak_bandwidth_params->dcc_dram_bw_pref_overhead_factor_p0 = mode_lib->ms.dcc_dram_bw_pref_overhead_factor_p0;
+			calculate_peak_bandwidth_params->dcc_dram_bw_pref_overhead_factor_p1 = mode_lib->ms.dcc_dram_bw_pref_overhead_factor_p1;
+			calculate_peak_bandwidth_params->mall_prefetch_sdp_overhead_factor = mode_lib->ms.mall_prefetch_sdp_overhead_factor;
+			calculate_peak_bandwidth_params->mall_prefetch_dram_overhead_factor = mode_lib->ms.mall_prefetch_dram_overhead_factor;
+
+			calculate_peak_bandwidth_params->surface_read_bandwidth_l = mode_lib->ms.vactive_sw_bw_l;
+			calculate_peak_bandwidth_params->surface_read_bandwidth_c = mode_lib->ms.vactive_sw_bw_c;
+			calculate_peak_bandwidth_params->prefetch_bandwidth_l = mode_lib->ms.RequiredPrefetchPixelDataBWLuma;
+			calculate_peak_bandwidth_params->prefetch_bandwidth_c = mode_lib->ms.RequiredPrefetchPixelDataBWChroma;
+			calculate_peak_bandwidth_params->prefetch_bandwidth_max = mode_lib->ms.RequiredPrefetchBWMax;
+			calculate_peak_bandwidth_params->excess_vactive_fill_bw_l = mode_lib->ms.excess_vactive_fill_bw_l;
+			calculate_peak_bandwidth_params->excess_vactive_fill_bw_c = mode_lib->ms.excess_vactive_fill_bw_c;
+			calculate_peak_bandwidth_params->cursor_bw = mode_lib->ms.cursor_bw;
+			calculate_peak_bandwidth_params->dpte_row_bw = mode_lib->ms.dpte_row_bw;
+			calculate_peak_bandwidth_params->meta_row_bw = mode_lib->ms.meta_row_bw;
+			calculate_peak_bandwidth_params->prefetch_cursor_bw = mode_lib->ms.prefetch_cursor_bw;
+			calculate_peak_bandwidth_params->prefetch_vmrow_bw = mode_lib->ms.prefetch_vmrow_bw;
+			calculate_peak_bandwidth_params->flip_bw = mode_lib->ms.final_flip_bw;
+			calculate_peak_bandwidth_params->urgent_burst_factor_l = mode_lib->ms.UrgentBurstFactorLuma;
+			calculate_peak_bandwidth_params->urgent_burst_factor_c = mode_lib->ms.UrgentBurstFactorChroma;
+			calculate_peak_bandwidth_params->urgent_burst_factor_cursor = mode_lib->ms.UrgentBurstFactorCursor;
+			calculate_peak_bandwidth_params->urgent_burst_factor_prefetch_l = mode_lib->ms.UrgentBurstFactorLumaPre;
+			calculate_peak_bandwidth_params->urgent_burst_factor_prefetch_c = mode_lib->ms.UrgentBurstFactorChromaPre;
+			calculate_peak_bandwidth_params->urgent_burst_factor_prefetch_cursor = mode_lib->ms.UrgentBurstFactorCursorPre;
+
+			calculate_peak_bandwidth_required(
+					&mode_lib->scratch,
+					calculate_peak_bandwidth_params);
+
+			// Check urg peak bandwidth against available urg bw
+			// check at SDP and DRAM, for all soc states (SVP prefetch an Sys Active)
+			check_urgent_bandwidth_support(
+				&s->dummy_single[0], // double* frac_urg_bandwidth
+				&s->dummy_single[1], // double* frac_urg_bandwidth_mall
+				&mode_lib->ms.support.UrgVactiveBandwidthSupport,
+				&mode_lib->ms.support.PrefetchBandwidthSupported,
+
+				mode_lib->soc.mall_allocated_for_dcn_mbytes,
+				mode_lib->ms.support.non_urg_bandwidth_required,
+				mode_lib->ms.support.urg_vactive_bandwidth_required,
+				mode_lib->ms.support.urg_bandwidth_required,
+				mode_lib->ms.support.urg_bandwidth_available);
+
+			mode_lib->ms.support.PrefetchSupported &= mode_lib->ms.support.PrefetchBandwidthSupported;
+			DML_LOG_VERBOSE("DML::%s: PrefetchBandwidthSupported=%0d\n", __func__, mode_lib->ms.support.PrefetchBandwidthSupported);
+
+			for (k = 0; k < mode_lib->ms.num_active_planes; k++) {
+				if (mode_lib->ms.NotEnoughUrgentLatencyHidingPre[k]) {
+					mode_lib->ms.support.PrefetchSupported = false;
+					DML_LOG_VERBOSE("DML::%s: k=%d, NotEnoughUrgentLatencyHidingPre=%d\n", __func__, k, mode_lib->ms.NotEnoughUrgentLatencyHidingPre[k]);
+				}
+			}
+
+#ifdef DML_GLOBAL_PREFETCH_CHECK
+			if (mode_lib->ms.support.PrefetchSupported && mode_lib->ms.num_active_planes > 1 && s->recalc_prefetch_done == 0) {
+				CheckGlobalPrefetchAdmissibility_params->num_active_planes =  mode_lib->ms.num_active_planes;
+				CheckGlobalPrefetchAdmissibility_params->pixel_format = s->pixel_format;
+				CheckGlobalPrefetchAdmissibility_params->chunk_bytes_l = mode_lib->ip.pixel_chunk_size_kbytes * 1024;
+				CheckGlobalPrefetchAdmissibility_params->chunk_bytes_c = mode_lib->ip.pixel_chunk_size_kbytes * 1024;
+				CheckGlobalPrefetchAdmissibility_params->lb_source_lines_l = s->lb_source_lines_l;
+				CheckGlobalPrefetchAdmissibility_params->lb_source_lines_c = s->lb_source_lines_c;
+				CheckGlobalPrefetchAdmissibility_params->swath_height_l =  mode_lib->ms.SwathHeightY;
+				CheckGlobalPrefetchAdmissibility_params->swath_height_c =  mode_lib->ms.SwathHeightC;
+				CheckGlobalPrefetchAdmissibility_params->rob_buffer_size_kbytes = mode_lib->ip.rob_buffer_size_kbytes;
+				CheckGlobalPrefetchAdmissibility_params->compressed_buffer_size_kbytes = mode_lib->ms.CompressedBufferSizeInkByte;
+				CheckGlobalPrefetchAdmissibility_params->detile_buffer_size_bytes_l = mode_lib->ms.DETBufferSizeY;
+				CheckGlobalPrefetchAdmissibility_params->detile_buffer_size_bytes_c = mode_lib->ms.DETBufferSizeC;
+				CheckGlobalPrefetchAdmissibility_params->full_swath_bytes_l = s->full_swath_bytes_l;
+				CheckGlobalPrefetchAdmissibility_params->full_swath_bytes_c = s->full_swath_bytes_c;
+				CheckGlobalPrefetchAdmissibility_params->prefetch_sw_bytes = s->prefetch_sw_bytes;
+				CheckGlobalPrefetchAdmissibility_params->Tpre_rounded = s->Tpre_rounded;
+				CheckGlobalPrefetchAdmissibility_params->Tpre_oto = s->Tpre_oto;
+				CheckGlobalPrefetchAdmissibility_params->estimated_urg_bandwidth_required_mbps = mode_lib->ms.support.urg_bandwidth_required[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_sdp];
+				CheckGlobalPrefetchAdmissibility_params->line_time = s->line_times;
+				CheckGlobalPrefetchAdmissibility_params->dst_y_prefetch = mode_lib->ms.dst_y_prefetch;
+				if (CheckGlobalPrefetchAdmissibility_params->estimated_urg_bandwidth_required_mbps < 10 * 1024)
+					CheckGlobalPrefetchAdmissibility_params->estimated_urg_bandwidth_required_mbps = 10 * 1024;
+
+				CheckGlobalPrefetchAdmissibility_params->estimated_dcfclk_mhz = (CheckGlobalPrefetchAdmissibility_params->estimated_urg_bandwidth_required_mbps / (double) mode_lib->soc.return_bus_width_bytes) /
+																				((double)mode_lib->soc.qos_parameters.derate_table.system_active_urgent.dcfclk_derate_percent / 100.0);
+
+				// if recalc_prefetch_schedule is set, recalculate the prefetch schedule with the new impacted_Tpre, prefetch should be possible
+				CheckGlobalPrefetchAdmissibility_params->recalc_prefetch_schedule = &s->recalc_prefetch_schedule;
+				CheckGlobalPrefetchAdmissibility_params->impacted_dst_y_pre = s->impacted_dst_y_pre;
+				mode_lib->ms.support.PrefetchSupported = CheckGlobalPrefetchAdmissibility(&mode_lib->scratch, CheckGlobalPrefetchAdmissibility_params);
+				s->recalc_prefetch_done = 1;
+				s->recalc_prefetch_schedule = 1;
+			}
+#endif
+		} // prefetch schedule ok, do urg bw and flip schedule
+	} while (s->recalc_prefetch_schedule);
+
+	// Flip Schedule
+	// Both prefetch schedule and BW okay
+	if (mode_lib->ms.support.PrefetchSupported == true) {
+		mode_lib->ms.BandwidthAvailableForImmediateFlip =
+			get_bandwidth_available_for_immediate_flip(
+				dml2_core_internal_soc_state_sys_active,
+				mode_lib->ms.support.urg_bandwidth_required_qual, // no flip
+				mode_lib->ms.support.urg_bandwidth_available);
+
+		mode_lib->ms.TotImmediateFlipBytes = 0;
+		for (k = 0; k < mode_lib->ms.num_active_planes; k++) {
+			if (display_cfg->plane_descriptors[k].immediate_flip) {
+				s->per_pipe_flip_bytes[k] = get_pipe_flip_bytes(
+								s->HostVMInefficiencyFactor,
+								mode_lib->ms.vm_bytes[k],
+								mode_lib->ms.DPTEBytesPerRow[k],
+								mode_lib->ms.meta_row_bytes[k]);
+			} else {
+				s->per_pipe_flip_bytes[k] = 0;
+			}
+			mode_lib->ms.TotImmediateFlipBytes += s->per_pipe_flip_bytes[k] * mode_lib->ms.NoOfDPP[k];
+
+		}
+
+		for (k = 0; k < mode_lib->ms.num_active_planes; k++) {
+			CalculateFlipSchedule(
+				&mode_lib->scratch,
+				display_cfg->plane_descriptors[k].immediate_flip,
+				1, // use_lb_flip_bw
+				s->HostVMInefficiencyFactor,
+				s->Tvm_trips_flip[k],
+				s->Tr0_trips_flip[k],
+				s->Tvm_trips_flip_rounded[k],
+				s->Tr0_trips_flip_rounded[k],
+				display_cfg->gpuvm_enable,
+				mode_lib->ms.vm_bytes[k],
+				mode_lib->ms.DPTEBytesPerRow[k],
+				mode_lib->ms.BandwidthAvailableForImmediateFlip,
+				mode_lib->ms.TotImmediateFlipBytes,
+				display_cfg->plane_descriptors[k].pixel_format,
+				(display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000)),
+				display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio,
+				display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_ratio,
+				mode_lib->ms.Tno_bw_flip[k],
+				mode_lib->ms.dpte_row_height[k],
+				mode_lib->ms.dpte_row_height_chroma[k],
+				mode_lib->ms.use_one_row_for_frame_flip[k],
+				mode_lib->ip.max_flip_time_us,
+				mode_lib->ip.max_flip_time_lines,
+				s->per_pipe_flip_bytes[k],
+				mode_lib->ms.meta_row_bytes[k],
+				s->meta_row_height_luma[k],
+				s->meta_row_height_chroma[k],
+				mode_lib->ip.dcn_mrq_present && display_cfg->plane_descriptors[k].surface.dcc.enable,
+
+				/* Output */
+				&mode_lib->ms.dst_y_per_vm_flip[k],
+				&mode_lib->ms.dst_y_per_row_flip[k],
+				&mode_lib->ms.final_flip_bw[k],
+				&mode_lib->ms.ImmediateFlipSupportedForPipe[k]);
+		}
+
+		calculate_peak_bandwidth_params->urg_vactive_bandwidth_required = s->dummy_bw;
+		calculate_peak_bandwidth_params->urg_bandwidth_required = mode_lib->ms.support.urg_bandwidth_required_flip;
+		calculate_peak_bandwidth_params->urg_bandwidth_required_qual = s->dummy_bw;
+		calculate_peak_bandwidth_params->non_urg_bandwidth_required = mode_lib->ms.support.non_urg_bandwidth_required_flip;
+		calculate_peak_bandwidth_params->surface_avg_vactive_required_bw = s->surface_dummy_bw;
+		calculate_peak_bandwidth_params->surface_peak_required_bw = mode_lib->ms.surface_peak_required_bw;
+
+		calculate_peak_bandwidth_params->display_cfg = display_cfg;
+		calculate_peak_bandwidth_params->inc_flip_bw = 1;
+		calculate_peak_bandwidth_params->num_active_planes = mode_lib->ms.num_active_planes;
+		calculate_peak_bandwidth_params->num_of_dpp = mode_lib->ms.NoOfDPP;
+		calculate_peak_bandwidth_params->dcc_dram_bw_nom_overhead_factor_p0 = mode_lib->ms.dcc_dram_bw_nom_overhead_factor_p0;
+		calculate_peak_bandwidth_params->dcc_dram_bw_nom_overhead_factor_p1 = mode_lib->ms.dcc_dram_bw_nom_overhead_factor_p1;
+		calculate_peak_bandwidth_params->dcc_dram_bw_pref_overhead_factor_p0 = mode_lib->ms.dcc_dram_bw_pref_overhead_factor_p0;
+		calculate_peak_bandwidth_params->dcc_dram_bw_pref_overhead_factor_p1 = mode_lib->ms.dcc_dram_bw_pref_overhead_factor_p1;
+		calculate_peak_bandwidth_params->mall_prefetch_sdp_overhead_factor = mode_lib->ms.mall_prefetch_sdp_overhead_factor;
+		calculate_peak_bandwidth_params->mall_prefetch_dram_overhead_factor = mode_lib->ms.mall_prefetch_dram_overhead_factor;
+
+		calculate_peak_bandwidth_params->surface_read_bandwidth_l = mode_lib->ms.vactive_sw_bw_l;
+		calculate_peak_bandwidth_params->surface_read_bandwidth_c = mode_lib->ms.vactive_sw_bw_c;
+		calculate_peak_bandwidth_params->prefetch_bandwidth_l = mode_lib->ms.RequiredPrefetchPixelDataBWLuma;
+		calculate_peak_bandwidth_params->prefetch_bandwidth_c = mode_lib->ms.RequiredPrefetchPixelDataBWChroma;
+		calculate_peak_bandwidth_params->prefetch_bandwidth_max = mode_lib->ms.RequiredPrefetchBWMax;
+		calculate_peak_bandwidth_params->excess_vactive_fill_bw_l = mode_lib->ms.excess_vactive_fill_bw_l;
+		calculate_peak_bandwidth_params->excess_vactive_fill_bw_c = mode_lib->ms.excess_vactive_fill_bw_c;
+		calculate_peak_bandwidth_params->cursor_bw = mode_lib->ms.cursor_bw;
+		calculate_peak_bandwidth_params->dpte_row_bw = mode_lib->ms.dpte_row_bw;
+		calculate_peak_bandwidth_params->meta_row_bw = mode_lib->ms.meta_row_bw;
+		calculate_peak_bandwidth_params->prefetch_cursor_bw = mode_lib->ms.prefetch_cursor_bw;
+		calculate_peak_bandwidth_params->prefetch_vmrow_bw = mode_lib->ms.prefetch_vmrow_bw;
+		calculate_peak_bandwidth_params->flip_bw = mode_lib->ms.final_flip_bw;
+		calculate_peak_bandwidth_params->urgent_burst_factor_l = mode_lib->ms.UrgentBurstFactorLuma;
+		calculate_peak_bandwidth_params->urgent_burst_factor_c = mode_lib->ms.UrgentBurstFactorChroma;
+		calculate_peak_bandwidth_params->urgent_burst_factor_cursor = mode_lib->ms.UrgentBurstFactorCursor;
+		calculate_peak_bandwidth_params->urgent_burst_factor_prefetch_l = mode_lib->ms.UrgentBurstFactorLumaPre;
+		calculate_peak_bandwidth_params->urgent_burst_factor_prefetch_c = mode_lib->ms.UrgentBurstFactorChromaPre;
+		calculate_peak_bandwidth_params->urgent_burst_factor_prefetch_cursor = mode_lib->ms.UrgentBurstFactorCursorPre;
+
+		calculate_peak_bandwidth_required(
+				&mode_lib->scratch,
+				calculate_peak_bandwidth_params);
+
+		calculate_immediate_flip_bandwidth_support(
+			&s->dummy_single[0], // double* frac_urg_bandwidth_flip
+			&mode_lib->ms.support.ImmediateFlipSupport,
+
+			dml2_core_internal_soc_state_sys_active,
+			mode_lib->ms.support.urg_bandwidth_required_flip,
+			mode_lib->ms.support.non_urg_bandwidth_required_flip,
+			mode_lib->ms.support.urg_bandwidth_available);
+
+		for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+			if (display_cfg->plane_descriptors[k].immediate_flip == true && mode_lib->ms.ImmediateFlipSupportedForPipe[k] == false)
+				mode_lib->ms.support.ImmediateFlipSupport = false;
+		}
+
+	} else { // if prefetch not support, assume iflip is not supported too
+		mode_lib->ms.support.ImmediateFlipSupport = false;
+	}
+
+	s->mSOCParameters.UrgentLatency = mode_lib->ms.UrgLatency;
+	s->mSOCParameters.ExtraLatency = mode_lib->ms.ExtraLatency;
+	s->mSOCParameters.ExtraLatency_sr = mode_lib->ms.ExtraLatency_sr;
+	s->mSOCParameters.WritebackLatency = mode_lib->soc.qos_parameters.writeback.base_latency_us;
+	s->mSOCParameters.DRAMClockChangeLatency = mode_lib->soc.power_management_parameters.dram_clk_change_blackout_us;
+	s->mSOCParameters.FCLKChangeLatency = mode_lib->soc.power_management_parameters.fclk_change_blackout_us;
+	s->mSOCParameters.SRExitTime = mode_lib->soc.power_management_parameters.stutter_exit_latency_us;
+	s->mSOCParameters.SREnterPlusExitTime = mode_lib->soc.power_management_parameters.stutter_enter_plus_exit_latency_us;
+	s->mSOCParameters.SRExitZ8Time = mode_lib->soc.power_management_parameters.z8_stutter_exit_latency_us;
+	s->mSOCParameters.SREnterPlusExitZ8Time = mode_lib->soc.power_management_parameters.z8_stutter_enter_plus_exit_latency_us;
+	s->mSOCParameters.USRRetrainingLatency = 0;
+	s->mSOCParameters.SMNLatency = 0;
+	s->mSOCParameters.g6_temp_read_blackout_us = get_g6_temp_read_blackout_us(&mode_lib->soc, (unsigned int)(mode_lib->ms.uclk_freq_mhz * 1000), mode_lib->ms.state_idx);
+	s->mSOCParameters.max_urgent_latency_us = get_max_urgent_latency_us(&mode_lib->soc.qos_parameters.qos_params.dcn4x, mode_lib->ms.uclk_freq_mhz, mode_lib->ms.FabricClock, mode_lib->ms.state_idx);
+	s->mSOCParameters.df_response_time_us = mode_lib->soc.qos_parameters.qos_params.dcn4x.df_qos_response_time_fclk_cycles / mode_lib->ms.FabricClock;
+	s->mSOCParameters.qos_type = mode_lib->soc.qos_parameters.qos_type;
+
+	CalculateWatermarks_params->display_cfg = display_cfg;
+	CalculateWatermarks_params->USRRetrainingRequired = false;
+	CalculateWatermarks_params->NumberOfActiveSurfaces = mode_lib->ms.num_active_planes;
+	CalculateWatermarks_params->MaxLineBufferLines = mode_lib->ip.max_line_buffer_lines;
+	CalculateWatermarks_params->LineBufferSize = mode_lib->ip.line_buffer_size_bits;
+	CalculateWatermarks_params->WritebackInterfaceBufferSize = mode_lib->ip.writeback_interface_buffer_size_kbytes;
+	CalculateWatermarks_params->DCFCLK = mode_lib->ms.DCFCLK;
+	CalculateWatermarks_params->SynchronizeTimings = display_cfg->overrides.synchronize_timings;
+	CalculateWatermarks_params->SynchronizeDRRDisplaysForUCLKPStateChange = display_cfg->overrides.synchronize_ddr_displays_for_uclk_pstate_change;
+	CalculateWatermarks_params->dpte_group_bytes = mode_lib->ms.dpte_group_bytes;
+	CalculateWatermarks_params->mmSOCParameters = s->mSOCParameters;
+	CalculateWatermarks_params->WritebackChunkSize = mode_lib->ip.writeback_chunk_size_kbytes;
+	CalculateWatermarks_params->SOCCLK = mode_lib->ms.SOCCLK;
+	CalculateWatermarks_params->DCFClkDeepSleep = mode_lib->ms.dcfclk_deepsleep;
+	CalculateWatermarks_params->DETBufferSizeY = mode_lib->ms.DETBufferSizeY;
+	CalculateWatermarks_params->DETBufferSizeC = mode_lib->ms.DETBufferSizeC;
+	CalculateWatermarks_params->SwathHeightY = mode_lib->ms.SwathHeightY;
+	CalculateWatermarks_params->SwathHeightC = mode_lib->ms.SwathHeightC;
+	CalculateWatermarks_params->SwathWidthY = mode_lib->ms.SwathWidthY;
+	CalculateWatermarks_params->SwathWidthC = mode_lib->ms.SwathWidthC;
+	CalculateWatermarks_params->DPPPerSurface = mode_lib->ms.NoOfDPP;
+	CalculateWatermarks_params->BytePerPixelDETY = mode_lib->ms.BytePerPixelInDETY;
+	CalculateWatermarks_params->BytePerPixelDETC = mode_lib->ms.BytePerPixelInDETC;
+	CalculateWatermarks_params->DSTXAfterScaler = s->DSTXAfterScaler;
+	CalculateWatermarks_params->DSTYAfterScaler = s->DSTYAfterScaler;
+	CalculateWatermarks_params->UnboundedRequestEnabled = mode_lib->ms.UnboundedRequestEnabled;
+	CalculateWatermarks_params->CompressedBufferSizeInkByte = mode_lib->ms.CompressedBufferSizeInkByte;
+	CalculateWatermarks_params->meta_row_height_l = s->meta_row_height_luma;
+	CalculateWatermarks_params->meta_row_height_c = s->meta_row_height_chroma;
+
+	// Output
+	CalculateWatermarks_params->Watermark = &mode_lib->ms.support.watermarks; // Watermarks *Watermark
+	CalculateWatermarks_params->DRAMClockChangeSupport = mode_lib->ms.support.DRAMClockChangeSupport;
+	CalculateWatermarks_params->global_dram_clock_change_supported = &mode_lib->ms.support.global_dram_clock_change_supported;
+	CalculateWatermarks_params->MaxActiveDRAMClockChangeLatencySupported = &s->dummy_single_array[0]; // double *MaxActiveDRAMClockChangeLatencySupported[]
+	CalculateWatermarks_params->SubViewportLinesNeededInMALL = mode_lib->ms.SubViewportLinesNeededInMALL; // unsigned int SubViewportLinesNeededInMALL[]
+	CalculateWatermarks_params->FCLKChangeSupport = mode_lib->ms.support.FCLKChangeSupport;
+	CalculateWatermarks_params->global_fclk_change_supported = &mode_lib->ms.support.global_fclk_change_supported;
+	CalculateWatermarks_params->MaxActiveFCLKChangeLatencySupported = &s->dummy_single[0]; // double *MaxActiveFCLKChangeLatencySupported
+	CalculateWatermarks_params->USRRetrainingSupport = &mode_lib->ms.support.USRRetrainingSupport;
+	CalculateWatermarks_params->g6_temp_read_support = &mode_lib->ms.support.g6_temp_read_support;
+	CalculateWatermarks_params->VActiveLatencyHidingMargin = mode_lib->ms.VActiveLatencyHidingMargin;
+	CalculateWatermarks_params->VActiveLatencyHidingUs = mode_lib->ms.VActiveLatencyHidingUs;
+
+	CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport(&mode_lib->scratch, CalculateWatermarks_params);
+
+	calculate_pstate_keepout_dst_lines(display_cfg, &mode_lib->ms.support.watermarks, s->dummy_integer_array[0]);
+	DML_LOG_VERBOSE("DML::%s: Done prefetch calculation\n", __func__);
+
+}
+
+
+static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out_params)
+{
+	struct dml2_core_internal_display_mode_lib *mode_lib = in_out_params->mode_lib;
+	const struct dml2_display_cfg *display_cfg = in_out_params->in_display_cfg;
+	const struct dml2_mcg_min_clock_table *min_clk_table = in_out_params->min_clk_table;
+
+	double outstanding_latency_us = 0;
+
+	struct dml2_core_calcs_mode_support_locals *s = &mode_lib->scratch.dml_core_mode_support_locals;
+	struct dml2_core_calcs_CalculateVMRowAndSwath_params *CalculateVMRowAndSwath_params = &mode_lib->scratch.CalculateVMRowAndSwath_params;
+	struct dml2_core_calcs_CalculateSwathAndDETConfiguration_params *CalculateSwathAndDETConfiguration_params = &mode_lib->scratch.CalculateSwathAndDETConfiguration_params;
+	struct dml2_core_calcs_calculate_mcache_setting_params *calculate_mcache_setting_params = &mode_lib->scratch.calculate_mcache_setting_params;
+	struct dml2_core_calcs_calculate_bytes_to_fetch_required_to_hide_latency_params *calculate_bytes_to_fetch_required_to_hide_latency_params = &mode_lib->scratch.calculate_bytes_to_fetch_required_to_hide_latency_params;
+	unsigned int k, m, n;
+
+	memset(&mode_lib->scratch, 0, sizeof(struct dml2_core_internal_scratch));
+	memset(&mode_lib->ms, 0, sizeof(struct dml2_core_internal_mode_support));
+
+	mode_lib->ms.num_active_planes = display_cfg->num_planes;
+	get_stream_output_bpp(s->OutputBpp, display_cfg);
+
+	mode_lib->ms.state_idx = in_out_params->min_clk_index;
+	mode_lib->ms.SOCCLK = ((double)mode_lib->soc.clk_table.socclk.clk_values_khz[0] / 1000);
+	mode_lib->ms.DCFCLK = ((double)min_clk_table->dram_bw_table.entries[in_out_params->min_clk_index].min_dcfclk_khz / 1000);
+	mode_lib->ms.FabricClock = ((double)min_clk_table->dram_bw_table.entries[in_out_params->min_clk_index].min_fclk_khz / 1000);
+	mode_lib->ms.MaxDCFCLK = (double)min_clk_table->max_clocks_khz.dcfclk / 1000;
+	mode_lib->ms.MaxFabricClock = (double)min_clk_table->max_clocks_khz.fclk / 1000;
+	mode_lib->ms.max_dispclk_freq_mhz = (double)min_clk_table->max_ss_clocks_khz.dispclk / 1000;
+	mode_lib->ms.max_dscclk_freq_mhz = (double)min_clk_table->max_clocks_khz.dscclk / 1000;
+	mode_lib->ms.max_dppclk_freq_mhz = (double)min_clk_table->max_ss_clocks_khz.dppclk / 1000;
+	mode_lib->ms.uclk_freq_mhz = dram_bw_kbps_to_uclk_mhz(min_clk_table->dram_bw_table.entries[in_out_params->min_clk_index].pre_derate_dram_bw_kbps, &mode_lib->soc.clk_table.dram_config);
+	mode_lib->ms.dram_bw_mbps = ((double)min_clk_table->dram_bw_table.entries[in_out_params->min_clk_index].pre_derate_dram_bw_kbps / 1000);
+	mode_lib->ms.max_dram_bw_mbps = ((double)min_clk_table->dram_bw_table.entries[min_clk_table->dram_bw_table.num_entries - 1].pre_derate_dram_bw_kbps / 1000);
+	mode_lib->ms.qos_param_index = get_qos_param_index((unsigned int) (mode_lib->ms.uclk_freq_mhz * 1000.0), mode_lib->soc.qos_parameters.qos_params.dcn4x.per_uclk_dpm_params);
+	mode_lib->ms.active_min_uclk_dpm_index = get_active_min_uclk_dpm_index((unsigned int) (mode_lib->ms.uclk_freq_mhz * 1000.0), &mode_lib->soc.clk_table);
+
+#if defined(__DML_VBA_DEBUG__)
+	DML_LOG_VERBOSE("DML::%s: --- START --- \n", __func__);
+	DML_LOG_VERBOSE("DML::%s: num_active_planes = %u\n", __func__, mode_lib->ms.num_active_planes);
+	DML_LOG_VERBOSE("DML::%s: min_clk_index = %0d\n", __func__, in_out_params->min_clk_index);
+	DML_LOG_VERBOSE("DML::%s: qos_param_index = %0d\n", __func__, mode_lib->ms.qos_param_index);
+	DML_LOG_VERBOSE("DML::%s: SOCCLK = %f\n", __func__, mode_lib->ms.SOCCLK);
+	DML_LOG_VERBOSE("DML::%s: dram_bw_mbps = %f\n", __func__, mode_lib->ms.dram_bw_mbps);
+	DML_LOG_VERBOSE("DML::%s: uclk_freq_mhz = %f\n", __func__, mode_lib->ms.uclk_freq_mhz);
+	DML_LOG_VERBOSE("DML::%s: DCFCLK = %f\n", __func__, mode_lib->ms.DCFCLK);
+	DML_LOG_VERBOSE("DML::%s: FabricClock = %f\n", __func__, mode_lib->ms.FabricClock);
+	DML_LOG_VERBOSE("DML::%s: MaxDCFCLK = %f\n", __func__, mode_lib->ms.MaxDCFCLK);
+	DML_LOG_VERBOSE("DML::%s: max_dispclk_freq_mhz = %f\n", __func__, mode_lib->ms.max_dispclk_freq_mhz);
+	DML_LOG_VERBOSE("DML::%s: max_dscclk_freq_mhz = %f\n", __func__, mode_lib->ms.max_dscclk_freq_mhz);
+	DML_LOG_VERBOSE("DML::%s: max_dppclk_freq_mhz = %f\n", __func__, mode_lib->ms.max_dppclk_freq_mhz);
+	DML_LOG_VERBOSE("DML::%s: MaxFabricClock = %f\n", __func__, mode_lib->ms.MaxFabricClock);
+	DML_LOG_VERBOSE("DML::%s: ip.compressed_buffer_segment_size_in_kbytes = %u\n", __func__, mode_lib->ip.compressed_buffer_segment_size_in_kbytes);
+	DML_LOG_VERBOSE("DML::%s: ip.dcn_mrq_present = %u\n", __func__, mode_lib->ip.dcn_mrq_present);
+
+	for (k = 0; k < mode_lib->ms.num_active_planes; k++)
+		DML_LOG_VERBOSE("DML::%s: plane_%d: reserved_vblank_time_ns = %lu\n", __func__, k, display_cfg->plane_descriptors[k].overrides.reserved_vblank_time_ns);
+#endif
+
+	CalculateMaxDETAndMinCompressedBufferSize(
+		mode_lib->ip.config_return_buffer_size_in_kbytes,
+		mode_lib->ip.config_return_buffer_segment_size_in_kbytes,
+		mode_lib->ip.rob_buffer_size_kbytes,
+		mode_lib->ip.max_num_dpp,
+		display_cfg->overrides.hw.force_nom_det_size_kbytes.enable,
+		display_cfg->overrides.hw.force_nom_det_size_kbytes.value,
+		mode_lib->ip.dcn_mrq_present,
+
+		/* Output */
+		&mode_lib->ms.MaxTotalDETInKByte,
+		&mode_lib->ms.NomDETInKByte,
+		&mode_lib->ms.MinCompressedBufferSizeInKByte);
+
+	PixelClockAdjustmentForProgressiveToInterlaceUnit(display_cfg, mode_lib->ip.ptoi_supported, s->PixelClockBackEnd);
+
+	/*MODE SUPPORT, VOLTAGE STATE AND SOC CONFIGURATION*/
+
+	/*Scale Ratio, taps Support Check*/
+	mode_lib->ms.support.ScaleRatioAndTapsSupport = true;
+	// Many core tests are still setting scaling parameters "incorrectly"
+	for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+		if (display_cfg->plane_descriptors[k].composition.scaler_info.enabled == false
+			&& (dml_is_420(display_cfg->plane_descriptors[k].pixel_format)
+				|| display_cfg->plane_descriptors[k].composition.scaler_info.plane0.h_ratio != 1.0
+				|| display_cfg->plane_descriptors[k].composition.scaler_info.plane0.h_taps != 1.0
+				|| display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio != 1.0
+				|| display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_taps != 1.0)) {
+			mode_lib->ms.support.ScaleRatioAndTapsSupport = false;
+		} else if (display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_taps < 1.0 || display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_taps > 8.0
+			|| display_cfg->plane_descriptors[k].composition.scaler_info.plane0.h_taps < 1.0 || display_cfg->plane_descriptors[k].composition.scaler_info.plane0.h_taps > 8.0
+			|| (display_cfg->plane_descriptors[k].composition.scaler_info.plane0.h_taps > 1.0 && (display_cfg->plane_descriptors[k].composition.scaler_info.plane0.h_taps % 2) == 1)
+			|| display_cfg->plane_descriptors[k].composition.scaler_info.plane0.h_ratio > mode_lib->ip.max_hscl_ratio
+			|| display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio > mode_lib->ip.max_vscl_ratio
+			|| display_cfg->plane_descriptors[k].composition.scaler_info.plane0.h_ratio > display_cfg->plane_descriptors[k].composition.scaler_info.plane0.h_taps
+			|| display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio > display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_taps
+			|| (dml_is_420(display_cfg->plane_descriptors[k].pixel_format)
+				&& (display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_taps < 1 || display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_taps > 8 ||
+					display_cfg->plane_descriptors[k].composition.scaler_info.plane1.h_taps < 1 || display_cfg->plane_descriptors[k].composition.scaler_info.plane1.h_taps > 8 ||
+					(display_cfg->plane_descriptors[k].composition.scaler_info.plane1.h_taps > 1 && display_cfg->plane_descriptors[k].composition.scaler_info.plane1.h_taps % 2 == 1) ||
+					display_cfg->plane_descriptors[k].composition.scaler_info.plane1.h_ratio > mode_lib->ip.max_hscl_ratio ||
+					display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_ratio > mode_lib->ip.max_vscl_ratio ||
+					display_cfg->plane_descriptors[k].composition.scaler_info.plane1.h_ratio > display_cfg->plane_descriptors[k].composition.scaler_info.plane1.h_taps ||
+					display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_ratio > display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_taps))) {
+			mode_lib->ms.support.ScaleRatioAndTapsSupport = false;
+		}
+	}
+
+	/*Source Format, Pixel Format and Scan Support Check*/
+	mode_lib->ms.support.SourceFormatPixelAndScanSupport = true;
+	for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+		if (display_cfg->plane_descriptors[k].surface.tiling == dml2_sw_linear && dml_is_vertical_rotation(display_cfg->plane_descriptors[k].composition.rotation_angle)) {
+			mode_lib->ms.support.SourceFormatPixelAndScanSupport = false;
+		}
+	}
+
+	for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+		CalculateBytePerPixelAndBlockSizes(
+			display_cfg->plane_descriptors[k].pixel_format,
+			display_cfg->plane_descriptors[k].surface.tiling,
+			display_cfg->plane_descriptors[k].surface.plane0.pitch,
+			display_cfg->plane_descriptors[k].surface.plane1.pitch,
+
+			/* Output */
+			&mode_lib->ms.BytePerPixelY[k],
+			&mode_lib->ms.BytePerPixelC[k],
+			&mode_lib->ms.BytePerPixelInDETY[k],
+			&mode_lib->ms.BytePerPixelInDETC[k],
+			&mode_lib->ms.Read256BlockHeightY[k],
+			&mode_lib->ms.Read256BlockHeightC[k],
+			&mode_lib->ms.Read256BlockWidthY[k],
+			&mode_lib->ms.Read256BlockWidthC[k],
+			&mode_lib->ms.MacroTileHeightY[k],
+			&mode_lib->ms.MacroTileHeightC[k],
+			&mode_lib->ms.MacroTileWidthY[k],
+			&mode_lib->ms.MacroTileWidthC[k],
+			&mode_lib->ms.surf_linear128_l[k],
+			&mode_lib->ms.surf_linear128_c[k]);
+	}
+
+	/*Bandwidth Support Check*/
+	for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+		if (!dml_is_vertical_rotation(display_cfg->plane_descriptors[k].composition.rotation_angle)) {
+			mode_lib->ms.SwathWidthYSingleDPP[k] = display_cfg->plane_descriptors[k].composition.viewport.plane0.width;
+			mode_lib->ms.SwathWidthCSingleDPP[k] = display_cfg->plane_descriptors[k].composition.viewport.plane1.width;
+		} else {
+			mode_lib->ms.SwathWidthYSingleDPP[k] = display_cfg->plane_descriptors[k].composition.viewport.plane0.height;
+			mode_lib->ms.SwathWidthCSingleDPP[k] = display_cfg->plane_descriptors[k].composition.viewport.plane1.height;
+		}
+	}
+
+	for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+		mode_lib->ms.vactive_sw_bw_l[k] = mode_lib->ms.SwathWidthYSingleDPP[k] * math_ceil2(mode_lib->ms.BytePerPixelY[k], 1.0) / (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000)) * display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio;
+		mode_lib->ms.vactive_sw_bw_c[k] = mode_lib->ms.SwathWidthCSingleDPP[k] * math_ceil2(mode_lib->ms.BytePerPixelC[k], 2.0) / (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000)) * display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_ratio;
+
+		mode_lib->ms.cursor_bw[k] = display_cfg->plane_descriptors[k].cursor.num_cursors * display_cfg->plane_descriptors[k].cursor.cursor_width *
+			display_cfg->plane_descriptors[k].cursor.cursor_bpp / 8.0 / (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000));
+
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: k=%u, old_ReadBandwidthLuma = %f\n", __func__, k, mode_lib->ms.SwathWidthYSingleDPP[k] * math_ceil2(mode_lib->ms.BytePerPixelInDETY[k], 1.0) / (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000)) * display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio);
+		DML_LOG_VERBOSE("DML::%s: k=%u, old_ReadBandwidthChroma = %f\n", __func__, k, mode_lib->ms.SwathWidthYSingleDPP[k] / 2 * math_ceil2(mode_lib->ms.BytePerPixelInDETC[k], 2.0) / (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000)) * display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio / 2.0);
+		DML_LOG_VERBOSE("DML::%s: k=%u, vactive_sw_bw_l = %f\n", __func__, k, mode_lib->ms.vactive_sw_bw_l[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u, vactive_sw_bw_c = %f\n", __func__, k, mode_lib->ms.vactive_sw_bw_c[k]);
+#endif
+	}
+
+	// Writeback bandwidth
+	for (k = 0; k < mode_lib->ms.num_active_planes; k++) {
+		if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.active_writebacks_per_stream > 0 && display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].pixel_format == dml2_444_64) {
+			mode_lib->ms.WriteBandwidth[k][0] = display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].output_height
+				* display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].output_width
+				/ (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].input_height
+					* display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total
+					/ ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000)) * 8.0;
+		} else if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.active_writebacks_per_stream > 0) {
+			mode_lib->ms.WriteBandwidth[k][0] = display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].output_height
+				* display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].output_width
+				/ (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].input_height
+					* display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total
+					/ ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000)) * 4.0;
+		} else {
+			mode_lib->ms.WriteBandwidth[k][0] = 0.0;
+		}
+	}
+
+	/*Writeback Latency support check*/
+	mode_lib->ms.support.WritebackLatencySupport = true;
+	for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+		if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.active_writebacks_per_stream > 0 &&
+			(mode_lib->ms.WriteBandwidth[k][0] > mode_lib->ip.writeback_interface_buffer_size_kbytes * 1024 / ((double)mode_lib->soc.qos_parameters.writeback.base_latency_us))) {
+			mode_lib->ms.support.WritebackLatencySupport = false;
+		}
+	}
+
+
+	/* Writeback Scale Ratio and Taps Support Check */
+	mode_lib->ms.support.WritebackScaleRatioAndTapsSupport = true;
+	for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+		if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.active_writebacks_per_stream > 0) {
+			if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].h_ratio > mode_lib->ip.writeback_max_hscl_ratio
+				|| display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].v_ratio > mode_lib->ip.writeback_max_vscl_ratio
+				|| display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].h_ratio < mode_lib->ip.writeback_min_hscl_ratio
+				|| display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].v_ratio < mode_lib->ip.writeback_min_vscl_ratio
+				|| display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].h_taps > (unsigned int) mode_lib->ip.writeback_max_hscl_taps
+				|| display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].v_taps > (unsigned int) mode_lib->ip.writeback_max_vscl_taps
+				|| display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].h_ratio > (unsigned int)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].h_taps
+				|| display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].v_ratio > (unsigned int)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].v_taps
+				|| (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].h_taps > 2.0 && ((display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].h_taps % 2) == 1))) {
+				mode_lib->ms.support.WritebackScaleRatioAndTapsSupport = false;
+			}
+			if (2.0 * display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].output_height * (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].v_taps - 1) * 57 > mode_lib->ip.writeback_line_buffer_buffer_size) {
+				mode_lib->ms.support.WritebackScaleRatioAndTapsSupport = false;
+			}
+		}
+	}
+
+	for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+		CalculateSinglePipeDPPCLKAndSCLThroughput(
+			display_cfg->plane_descriptors[k].composition.scaler_info.plane0.h_ratio,
+			display_cfg->plane_descriptors[k].composition.scaler_info.plane1.h_ratio,
+			display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio,
+			display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_ratio,
+			mode_lib->ip.max_dchub_pscl_bw_pix_per_clk,
+			mode_lib->ip.max_pscl_lb_bw_pix_per_clk,
+			((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000),
+			display_cfg->plane_descriptors[k].pixel_format,
+			display_cfg->plane_descriptors[k].composition.scaler_info.plane0.h_taps,
+			display_cfg->plane_descriptors[k].composition.scaler_info.plane1.h_taps,
+			display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_taps,
+			display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_taps,
+			/* Output */
+			&mode_lib->ms.PSCL_FACTOR[k],
+			&mode_lib->ms.PSCL_FACTOR_CHROMA[k],
+			&mode_lib->ms.MinDPPCLKUsingSingleDPP[k]);
+	}
+
+	// Max Viewport Size support
+	for (k = 0; k < mode_lib->ms.num_active_planes; k++) {
+		if (display_cfg->plane_descriptors[k].surface.tiling == dml2_sw_linear) {
+			s->MaximumSwathWidthSupportLuma = 15360;
+		} else if (!dml_is_vertical_rotation(display_cfg->plane_descriptors[k].composition.rotation_angle) && mode_lib->ms.BytePerPixelC[k] > 0 && display_cfg->plane_descriptors[k].pixel_format != dml2_rgbe_alpha) { // horz video
+			s->MaximumSwathWidthSupportLuma = 7680 + 16;
+		} else if (dml_is_vertical_rotation(display_cfg->plane_descriptors[k].composition.rotation_angle) && mode_lib->ms.BytePerPixelC[k] > 0 && display_cfg->plane_descriptors[k].pixel_format != dml2_rgbe_alpha) { // vert video
+			s->MaximumSwathWidthSupportLuma = 4320 + 16;
+		} else if (display_cfg->plane_descriptors[k].pixel_format == dml2_rgbe_alpha) { // rgbe + alpha
+			s->MaximumSwathWidthSupportLuma = 5120 + 16;
+		} else if (dml_is_vertical_rotation(display_cfg->plane_descriptors[k].composition.rotation_angle) && mode_lib->ms.BytePerPixelY[k] == 8 && display_cfg->plane_descriptors[k].surface.dcc.enable == true) { // vert 64bpp
+			s->MaximumSwathWidthSupportLuma = 3072 + 16;
+		} else {
+			s->MaximumSwathWidthSupportLuma = 6144 + 16;
+		}
+
+		if (dml_is_420(display_cfg->plane_descriptors[k].pixel_format)) {
+			s->MaximumSwathWidthSupportChroma = (unsigned int)(s->MaximumSwathWidthSupportLuma / 2.0);
+		} else {
+			s->MaximumSwathWidthSupportChroma = s->MaximumSwathWidthSupportLuma;
+		}
+
+		unsigned lb_buffer_size_bits_luma = mode_lib->ip.line_buffer_size_bits;
+		unsigned lb_buffer_size_bits_chroma = mode_lib->ip.line_buffer_size_bits;
+
+/*
+#if defined(DV_BUILD)
+		// Assume a memory config setting of 3 in 420 mode or get a new ip parameter that reflects the programming.
+		if (mode_lib->ms.BytePerPixelC[k] != 0.0 && display_cfg->plane_descriptors[k].pixel_format != dml2_rgbe_alpha) {
+			lb_buffer_size_bits_luma = 34620 * 57;
+			lb_buffer_size_bits_chroma = 13560 * 57;
+		}
+#endif
+*/
+		mode_lib->ms.MaximumSwathWidthInLineBufferLuma = lb_buffer_size_bits_luma * math_max2(display_cfg->plane_descriptors[k].composition.scaler_info.plane0.h_ratio, 1.0) / 57 /
+			(display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_taps + math_max2(math_ceil2(display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio, 1.0) - 2, 0.0));
+		if (mode_lib->ms.BytePerPixelC[k] == 0.0) {
+			mode_lib->ms.MaximumSwathWidthInLineBufferChroma = 0;
+		} else {
+			mode_lib->ms.MaximumSwathWidthInLineBufferChroma = lb_buffer_size_bits_chroma * math_max2(display_cfg->plane_descriptors[k].composition.scaler_info.plane1.h_ratio, 1.0) / 57 /
+				(display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_taps + math_max2(math_ceil2(display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_ratio, 1.0) - 2, 0.0));
+		}
+
+		mode_lib->ms.MaximumSwathWidthLuma[k] = math_min2(s->MaximumSwathWidthSupportLuma, mode_lib->ms.MaximumSwathWidthInLineBufferLuma);
+		mode_lib->ms.MaximumSwathWidthChroma[k] = math_min2(s->MaximumSwathWidthSupportChroma, mode_lib->ms.MaximumSwathWidthInLineBufferChroma);
+
+		DML_LOG_VERBOSE("DML::%s: k=%u MaximumSwathWidthLuma=%f\n", __func__, k, mode_lib->ms.MaximumSwathWidthLuma[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u MaximumSwathWidthSupportLuma=%u\n", __func__, k, s->MaximumSwathWidthSupportLuma);
+		DML_LOG_VERBOSE("DML::%s: k=%u MaximumSwathWidthInLineBufferLuma=%f\n", __func__, k, mode_lib->ms.MaximumSwathWidthInLineBufferLuma);
+
+		DML_LOG_VERBOSE("DML::%s: k=%u MaximumSwathWidthChroma=%f\n", __func__, k, mode_lib->ms.MaximumSwathWidthChroma[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u MaximumSwathWidthSupportChroma=%u\n", __func__, k, s->MaximumSwathWidthSupportChroma);
+		DML_LOG_VERBOSE("DML::%s: k=%u MaximumSwathWidthInLineBufferChroma=%f\n", __func__, k, mode_lib->ms.MaximumSwathWidthInLineBufferChroma);
+	}
+
+	/* Cursor Support Check */
+	mode_lib->ms.support.CursorSupport = true;
+	for (k = 0; k < mode_lib->ms.num_active_planes; k++) {
+		if (display_cfg->plane_descriptors[k].cursor.num_cursors > 0) {
+			if (display_cfg->plane_descriptors[k].cursor.cursor_bpp == 64 && mode_lib->ip.cursor_64bpp_support == false)
+				mode_lib->ms.support.CursorSupport = false;
+		}
+	}
+
+	/* Valid Pitch Check */
+	mode_lib->ms.support.PitchSupport = true;
+	for (k = 0; k < mode_lib->ms.num_active_planes; k++) {
+
+		// data pitch
+		unsigned int alignment_l = mode_lib->ms.MacroTileWidthY[k];
+
+		if (mode_lib->ms.surf_linear128_l[k])
+			alignment_l = alignment_l / 2;
+
+		mode_lib->ms.support.AlignedYPitch[k] = (unsigned int)math_ceil2(math_max2(display_cfg->plane_descriptors[k].surface.plane0.pitch, display_cfg->plane_descriptors[k].surface.plane0.width), alignment_l);
+		if (dml_is_420(display_cfg->plane_descriptors[k].pixel_format) || display_cfg->plane_descriptors[k].pixel_format == dml2_rgbe_alpha) {
+			unsigned int alignment_c = mode_lib->ms.MacroTileWidthC[k];
+
+			if (mode_lib->ms.surf_linear128_c[k])
+				alignment_c = alignment_c / 2;
+			mode_lib->ms.support.AlignedCPitch[k] = (unsigned int)math_ceil2(math_max2(display_cfg->plane_descriptors[k].surface.plane1.pitch, display_cfg->plane_descriptors[k].surface.plane1.width), alignment_c);
+		} else {
+			mode_lib->ms.support.AlignedCPitch[k] = display_cfg->plane_descriptors[k].surface.plane1.pitch;
+		}
+
+		if (mode_lib->ms.support.AlignedYPitch[k] > display_cfg->plane_descriptors[k].surface.plane0.pitch ||
+			mode_lib->ms.support.AlignedCPitch[k] > display_cfg->plane_descriptors[k].surface.plane1.pitch) {
+			mode_lib->ms.support.PitchSupport = false;
+#if defined(__DML_VBA_DEBUG__)
+			DML_LOG_VERBOSE("DML::%s: k=%u AlignedYPitch = %d\n", __func__, k, mode_lib->ms.support.AlignedYPitch[k]);
+			DML_LOG_VERBOSE("DML::%s: k=%u PitchY = %ld\n", __func__, k, display_cfg->plane_descriptors[k].surface.plane0.pitch);
+			DML_LOG_VERBOSE("DML::%s: k=%u AlignedCPitch = %d\n", __func__, k, mode_lib->ms.support.AlignedCPitch[k]);
+			DML_LOG_VERBOSE("DML::%s: k=%u PitchC = %ld\n", __func__, k, display_cfg->plane_descriptors[k].surface.plane1.pitch);
+			DML_LOG_VERBOSE("DML::%s: k=%u PitchSupport = %d\n", __func__, k, mode_lib->ms.support.PitchSupport);
+#endif
+		}
+
+		// meta pitch
+		if (mode_lib->ip.dcn_mrq_present && display_cfg->plane_descriptors[k].surface.dcc.enable) {
+			mode_lib->ms.support.AlignedDCCMetaPitchY[k] = (unsigned int)math_ceil2(math_max2(display_cfg->plane_descriptors[k].surface.dcc.plane0.pitch,
+															display_cfg->plane_descriptors[k].surface.plane0.width), 64.0 * mode_lib->ms.Read256BlockWidthY[k]);
+
+			if (mode_lib->ms.support.AlignedDCCMetaPitchY[k] > display_cfg->plane_descriptors[k].surface.dcc.plane0.pitch)
+				mode_lib->ms.support.PitchSupport = false;
+
+			if (dml_is_420(display_cfg->plane_descriptors[k].pixel_format) || display_cfg->plane_descriptors[k].pixel_format == dml2_rgbe_alpha) {
+				mode_lib->ms.support.AlignedDCCMetaPitchC[k] = (unsigned int)math_ceil2(math_max2(display_cfg->plane_descriptors[k].surface.dcc.plane1.pitch,
+																display_cfg->plane_descriptors[k].surface.plane1.width), 64.0 * mode_lib->ms.Read256BlockWidthC[k]);
+
+				if (mode_lib->ms.support.AlignedDCCMetaPitchC[k] > display_cfg->plane_descriptors[k].surface.dcc.plane1.pitch)
+					mode_lib->ms.support.PitchSupport = false;
+			}
+		} else {
+			mode_lib->ms.support.AlignedDCCMetaPitchY[k] = 0;
+			mode_lib->ms.support.AlignedDCCMetaPitchC[k] = 0;
+		}
+	}
+
+	mode_lib->ms.support.ViewportExceedsSurface = false;
+	if (!display_cfg->overrides.hw.surface_viewport_size_check_disable) {
+		for (k = 0; k < mode_lib->ms.num_active_planes; k++) {
+			if (display_cfg->plane_descriptors[k].composition.viewport.plane0.width > display_cfg->plane_descriptors[k].surface.plane0.width ||
+				display_cfg->plane_descriptors[k].composition.viewport.plane0.height > display_cfg->plane_descriptors[k].surface.plane0.height) {
+				mode_lib->ms.support.ViewportExceedsSurface = true;
+#if defined(__DML_VBA_DEBUG__)
+				DML_LOG_VERBOSE("DML::%s: k=%u ViewportWidth = %ld\n", __func__, k, display_cfg->plane_descriptors[k].composition.viewport.plane0.width);
+				DML_LOG_VERBOSE("DML::%s: k=%u SurfaceWidthY = %ld\n", __func__, k, display_cfg->plane_descriptors[k].surface.plane0.width);
+				DML_LOG_VERBOSE("DML::%s: k=%u ViewportHeight = %ld\n", __func__, k, display_cfg->plane_descriptors[k].composition.viewport.plane0.height);
+				DML_LOG_VERBOSE("DML::%s: k=%u SurfaceHeightY = %ld\n", __func__, k, display_cfg->plane_descriptors[k].surface.plane0.height);
+				DML_LOG_VERBOSE("DML::%s: k=%u ViewportExceedsSurface = %d\n", __func__, k, mode_lib->ms.support.ViewportExceedsSurface);
+#endif
+			}
+			if (dml_is_420(display_cfg->plane_descriptors[k].pixel_format) || display_cfg->plane_descriptors[k].pixel_format == dml2_rgbe_alpha) {
+				if (display_cfg->plane_descriptors[k].composition.viewport.plane1.width > display_cfg->plane_descriptors[k].surface.plane1.width ||
+					display_cfg->plane_descriptors[k].composition.viewport.plane1.height > display_cfg->plane_descriptors[k].surface.plane1.height) {
+					mode_lib->ms.support.ViewportExceedsSurface = true;
+				}
+			}
+		}
+	}
+
+	CalculateSwathAndDETConfiguration_params->display_cfg = display_cfg;
+	CalculateSwathAndDETConfiguration_params->ConfigReturnBufferSizeInKByte = mode_lib->ip.config_return_buffer_size_in_kbytes;
+	CalculateSwathAndDETConfiguration_params->MaxTotalDETInKByte = mode_lib->ms.MaxTotalDETInKByte;
+	CalculateSwathAndDETConfiguration_params->MinCompressedBufferSizeInKByte = mode_lib->ms.MinCompressedBufferSizeInKByte;
+	CalculateSwathAndDETConfiguration_params->rob_buffer_size_kbytes = mode_lib->ip.rob_buffer_size_kbytes;
+	CalculateSwathAndDETConfiguration_params->pixel_chunk_size_kbytes = mode_lib->ip.pixel_chunk_size_kbytes;
+	CalculateSwathAndDETConfiguration_params->rob_buffer_size_kbytes = mode_lib->ip.rob_buffer_size_kbytes;
+	CalculateSwathAndDETConfiguration_params->pixel_chunk_size_kbytes = mode_lib->ip.pixel_chunk_size_kbytes;
+	CalculateSwathAndDETConfiguration_params->ForceSingleDPP = 1;
+	CalculateSwathAndDETConfiguration_params->NumberOfActiveSurfaces = mode_lib->ms.num_active_planes;
+	CalculateSwathAndDETConfiguration_params->nomDETInKByte = mode_lib->ms.NomDETInKByte;
+	CalculateSwathAndDETConfiguration_params->ConfigReturnBufferSegmentSizeInkByte = mode_lib->ip.config_return_buffer_segment_size_in_kbytes;
+	CalculateSwathAndDETConfiguration_params->CompressedBufferSegmentSizeInkByte = mode_lib->ip.compressed_buffer_segment_size_in_kbytes;
+	CalculateSwathAndDETConfiguration_params->ReadBandwidthLuma = mode_lib->ms.vactive_sw_bw_l;
+	CalculateSwathAndDETConfiguration_params->ReadBandwidthChroma = mode_lib->ms.vactive_sw_bw_c;
+	CalculateSwathAndDETConfiguration_params->MaximumSwathWidthLuma = mode_lib->ms.MaximumSwathWidthLuma;
+	CalculateSwathAndDETConfiguration_params->MaximumSwathWidthChroma = mode_lib->ms.MaximumSwathWidthChroma;
+	CalculateSwathAndDETConfiguration_params->Read256BytesBlockHeightY = mode_lib->ms.Read256BlockHeightY;
+	CalculateSwathAndDETConfiguration_params->Read256BytesBlockHeightC = mode_lib->ms.Read256BlockHeightC;
+	CalculateSwathAndDETConfiguration_params->Read256BytesBlockWidthY = mode_lib->ms.Read256BlockWidthY;
+	CalculateSwathAndDETConfiguration_params->Read256BytesBlockWidthC = mode_lib->ms.Read256BlockWidthC;
+	CalculateSwathAndDETConfiguration_params->surf_linear128_l = mode_lib->ms.surf_linear128_l;
+	CalculateSwathAndDETConfiguration_params->surf_linear128_c = mode_lib->ms.surf_linear128_c;
+	CalculateSwathAndDETConfiguration_params->ODMMode = s->dummy_odm_mode;
+	CalculateSwathAndDETConfiguration_params->BytePerPixY = mode_lib->ms.BytePerPixelY;
+	CalculateSwathAndDETConfiguration_params->BytePerPixC = mode_lib->ms.BytePerPixelC;
+	CalculateSwathAndDETConfiguration_params->BytePerPixDETY = mode_lib->ms.BytePerPixelInDETY;
+	CalculateSwathAndDETConfiguration_params->BytePerPixDETC = mode_lib->ms.BytePerPixelInDETC;
+	CalculateSwathAndDETConfiguration_params->DPPPerSurface = s->dummy_integer_array[2];
+	CalculateSwathAndDETConfiguration_params->mrq_present = mode_lib->ip.dcn_mrq_present;
+
+	// output
+	CalculateSwathAndDETConfiguration_params->req_per_swath_ub_l = s->dummy_integer_array[0];
+	CalculateSwathAndDETConfiguration_params->req_per_swath_ub_c = s->dummy_integer_array[1];
+	CalculateSwathAndDETConfiguration_params->swath_width_luma_ub = s->dummy_integer_array[3];
+	CalculateSwathAndDETConfiguration_params->swath_width_chroma_ub = s->dummy_integer_array[4];
+	CalculateSwathAndDETConfiguration_params->SwathWidth = s->dummy_integer_array[5];
+	CalculateSwathAndDETConfiguration_params->SwathWidthChroma = s->dummy_integer_array[6];
+	CalculateSwathAndDETConfiguration_params->SwathHeightY = s->dummy_integer_array[7];
+	CalculateSwathAndDETConfiguration_params->SwathHeightC = s->dummy_integer_array[8];
+	CalculateSwathAndDETConfiguration_params->request_size_bytes_luma = s->dummy_integer_array[26];
+	CalculateSwathAndDETConfiguration_params->request_size_bytes_chroma = s->dummy_integer_array[27];
+	CalculateSwathAndDETConfiguration_params->DETBufferSizeInKByte = s->dummy_integer_array[9];
+	CalculateSwathAndDETConfiguration_params->DETBufferSizeY = s->dummy_integer_array[10];
+	CalculateSwathAndDETConfiguration_params->DETBufferSizeC = s->dummy_integer_array[11];
+	CalculateSwathAndDETConfiguration_params->full_swath_bytes_l = s->full_swath_bytes_l;
+	CalculateSwathAndDETConfiguration_params->full_swath_bytes_c = s->full_swath_bytes_c;
+	CalculateSwathAndDETConfiguration_params->UnboundedRequestEnabled = &s->dummy_boolean[0];
+	CalculateSwathAndDETConfiguration_params->compbuf_reserved_space_64b = &s->dummy_integer[1];
+	CalculateSwathAndDETConfiguration_params->hw_debug5 = &s->dummy_boolean[2];
+	CalculateSwathAndDETConfiguration_params->CompressedBufferSizeInkByte = &s->dummy_integer[0];
+	CalculateSwathAndDETConfiguration_params->ViewportSizeSupportPerSurface = mode_lib->ms.SingleDPPViewportSizeSupportPerSurface;
+	CalculateSwathAndDETConfiguration_params->ViewportSizeSupport = &s->dummy_boolean[1];
+
+	// This calls is just to find out if there is enough DET space to support full vp in 1 pipe.
+	CalculateSwathAndDETConfiguration(&mode_lib->scratch, CalculateSwathAndDETConfiguration_params);
+
+	mode_lib->ms.TotalNumberOfActiveDPP = 0;
+	mode_lib->ms.TotalNumberOfActiveOPP = 0;
+	mode_lib->ms.support.TotalAvailablePipesSupport = true;
+
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		/*Number Of DSC Slices*/
+		if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.dsc.enable == dml2_dsc_enable ||
+			display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.dsc.enable == dml2_dsc_enable_if_necessary) {
+
+			if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.dsc.overrides.num_slices != 0)
+				mode_lib->ms.support.NumberOfDSCSlices[k] = display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.dsc.overrides.num_slices;
+			else {
+				if (s->PixelClockBackEnd[k] > 4800) {
+					mode_lib->ms.support.NumberOfDSCSlices[k] = (unsigned int)(math_ceil2(s->PixelClockBackEnd[k] / 600, 4));
+				} else if (s->PixelClockBackEnd[k] > 2400) {
+					mode_lib->ms.support.NumberOfDSCSlices[k] = 8;
+				} else if (s->PixelClockBackEnd[k] > 1200) {
+					mode_lib->ms.support.NumberOfDSCSlices[k] = 4;
+				} else if (s->PixelClockBackEnd[k] > 340) {
+					mode_lib->ms.support.NumberOfDSCSlices[k] = 2;
+				} else {
+					mode_lib->ms.support.NumberOfDSCSlices[k] = 1;
+				}
+			}
+		} else {
+			mode_lib->ms.support.NumberOfDSCSlices[k] = 0;
+		}
+
+		CalculateODMMode(
+			mode_lib->ip.maximum_pixels_per_line_per_dsc_unit,
+			display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_active,
+			display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_format,
+			display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_encoder,
+			display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].overrides.odm_mode,
+			mode_lib->ms.max_dispclk_freq_mhz,
+			false, // DSCEnable
+			mode_lib->ms.TotalNumberOfActiveDPP,
+			mode_lib->ms.TotalNumberOfActiveOPP,
+			mode_lib->ip.max_num_dpp,
+			mode_lib->ip.max_num_opp,
+			((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000),
+			mode_lib->ms.support.NumberOfDSCSlices[k],
+
+			/* Output */
+			&s->TotalAvailablePipesSupportNoDSC,
+			&s->NumberOfDPPNoDSC,
+			&s->ODMModeNoDSC,
+			&s->RequiredDISPCLKPerSurfaceNoDSC);
+
+		CalculateODMMode(
+			mode_lib->ip.maximum_pixels_per_line_per_dsc_unit,
+			display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_active,
+			display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_format,
+			display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_encoder,
+			display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].overrides.odm_mode,
+			mode_lib->ms.max_dispclk_freq_mhz,
+			true, // DSCEnable
+			mode_lib->ms.TotalNumberOfActiveDPP,
+			mode_lib->ms.TotalNumberOfActiveOPP,
+			mode_lib->ip.max_num_dpp,
+			mode_lib->ip.max_num_opp,
+			((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000),
+			mode_lib->ms.support.NumberOfDSCSlices[k],
+
+			/* Output */
+			&s->TotalAvailablePipesSupportDSC,
+			&s->NumberOfDPPDSC,
+			&s->ODMModeDSC,
+			&s->RequiredDISPCLKPerSurfaceDSC);
+
+		CalculateOutputLink(
+			&mode_lib->scratch,
+			((double)mode_lib->soc.clk_table.phyclk.clk_values_khz[0] / 1000),
+			((double)mode_lib->soc.clk_table.phyclk_d18.clk_values_khz[0] / 1000),
+			((double)mode_lib->soc.clk_table.phyclk_d32.clk_values_khz[0] / 1000),
+			mode_lib->soc.phy_downspread_percent,
+			display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_encoder,
+			display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_format,
+			display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total,
+			display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_active,
+			s->PixelClockBackEnd[k],
+			s->OutputBpp[k],
+			mode_lib->ip.maximum_dsc_bits_per_component,
+			mode_lib->ms.support.NumberOfDSCSlices[k],
+			display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.audio_sample_rate,
+			display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.audio_sample_layout,
+			s->ODMModeNoDSC,
+			s->ODMModeDSC,
+			display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.dsc.enable,
+			display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_dp_lane_count,
+			display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_dp_link_rate,
+
+			/* Output */
+			&mode_lib->ms.RequiresDSC[k],
+			&mode_lib->ms.RequiresFEC[k],
+			&mode_lib->ms.OutputBpp[k],
+			&mode_lib->ms.OutputType[k],
+			&mode_lib->ms.OutputRate[k],
+			&mode_lib->ms.RequiredSlots[k]);
+
+		if (s->OutputBpp[k] == 0.0) {
+			s->OutputBpp[k] = mode_lib->ms.OutputBpp[k];
+		}
+
+		if (mode_lib->ms.RequiresDSC[k] == false) {
+			mode_lib->ms.ODMMode[k] = s->ODMModeNoDSC;
+			mode_lib->ms.RequiredDISPCLKPerSurface[k] = s->RequiredDISPCLKPerSurfaceNoDSC;
+			if (!s->TotalAvailablePipesSupportNoDSC)
+				mode_lib->ms.support.TotalAvailablePipesSupport = false;
+			mode_lib->ms.TotalNumberOfActiveDPP = mode_lib->ms.TotalNumberOfActiveDPP + s->NumberOfDPPNoDSC;
+		} else {
+			mode_lib->ms.ODMMode[k] = s->ODMModeDSC;
+			mode_lib->ms.RequiredDISPCLKPerSurface[k] = s->RequiredDISPCLKPerSurfaceDSC;
+			if (!s->TotalAvailablePipesSupportDSC)
+				mode_lib->ms.support.TotalAvailablePipesSupport = false;
+			mode_lib->ms.TotalNumberOfActiveDPP = mode_lib->ms.TotalNumberOfActiveDPP + s->NumberOfDPPDSC;
+		}
+#if defined(__DML_VBA_DEBUG__)
+		DML_LOG_VERBOSE("DML::%s: k=%d RequiresDSC = %d\n", __func__, k, mode_lib->ms.RequiresDSC[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%d ODMMode = %d\n", __func__, k, mode_lib->ms.ODMMode[k]);
+#endif
+
+		// ensure the number dsc slices is integer multiple based on ODM mode
+		mode_lib->ms.support.DSCSlicesODMModeSupported = true;
+		if (mode_lib->ms.RequiresDSC[k]) {
+			// fail a ms check if the override num_slices doesn't align with odm mode setting
+			if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.dsc.overrides.num_slices != 0) {
+				if (mode_lib->ms.ODMMode[k] == dml2_odm_mode_combine_2to1)
+					mode_lib->ms.support.DSCSlicesODMModeSupported = ((mode_lib->ms.support.NumberOfDSCSlices[k] % 2) == 0);
+				else if (mode_lib->ms.ODMMode[k] == dml2_odm_mode_combine_3to1)
+					mode_lib->ms.support.DSCSlicesODMModeSupported = (mode_lib->ms.support.NumberOfDSCSlices[k] == 12);
+				else if (mode_lib->ms.ODMMode[k] == dml2_odm_mode_combine_4to1)
+					mode_lib->ms.support.DSCSlicesODMModeSupported = ((mode_lib->ms.support.NumberOfDSCSlices[k] % 4) == 0);
+#if defined(__DML_VBA_DEBUG__)
+				if (!mode_lib->ms.support.DSCSlicesODMModeSupported) {
+					DML_LOG_VERBOSE("DML::%s: k=%d Invalid dsc num_slices and ODM mode setting\n", __func__, k);
+					DML_LOG_VERBOSE("DML::%s: k=%d num_slices = %d\n", __func__, k, display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.dsc.overrides.num_slices);
+					DML_LOG_VERBOSE("DML::%s: k=%d ODMMode = %d\n", __func__, k, mode_lib->ms.ODMMode[k]);
+				}
+#endif
+			} else {
+				// safe guard to ensure the dml derived dsc slices and odm setting are compatible
+				if (mode_lib->ms.ODMMode[k] == dml2_odm_mode_combine_2to1)
+					mode_lib->ms.support.NumberOfDSCSlices[k] = 2 * (unsigned int)math_ceil2(mode_lib->ms.support.NumberOfDSCSlices[k] / 2.0, 1.0);
+				else if (mode_lib->ms.ODMMode[k] == dml2_odm_mode_combine_3to1)
+					mode_lib->ms.support.NumberOfDSCSlices[k] = 12;
+				else if (mode_lib->ms.ODMMode[k] == dml2_odm_mode_combine_4to1)
+					mode_lib->ms.support.NumberOfDSCSlices[k] = 4 * (unsigned int)math_ceil2(mode_lib->ms.support.NumberOfDSCSlices[k] / 4.0, 1.0);
+			}
+
+		} else {
+			mode_lib->ms.support.NumberOfDSCSlices[k] = 0;
+		}
+	}
+
+	mode_lib->ms.support.incorrect_imall_usage = 0;
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		if (mode_lib->ip.imall_supported && display_cfg->plane_descriptors[k].overrides.legacy_svp_config == dml2_svp_mode_override_imall)
+			mode_lib->ms.support.incorrect_imall_usage = 1;
+	}
+
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		mode_lib->ms.MPCCombine[k] = false;
+		mode_lib->ms.NoOfDPP[k] = 1;
+		mode_lib->ms.NoOfOPP[k] = 1;
+
+		if (mode_lib->ms.ODMMode[k] == dml2_odm_mode_combine_4to1) {
+			mode_lib->ms.MPCCombine[k] = false;
+			mode_lib->ms.NoOfDPP[k] = 4;
+			mode_lib->ms.NoOfOPP[k] = 4;
+		} else if (mode_lib->ms.ODMMode[k] == dml2_odm_mode_combine_3to1) {
+			mode_lib->ms.MPCCombine[k] = false;
+			mode_lib->ms.NoOfDPP[k] = 3;
+			mode_lib->ms.NoOfOPP[k] = 3;
+		} else if (mode_lib->ms.ODMMode[k] == dml2_odm_mode_combine_2to1) {
+			mode_lib->ms.MPCCombine[k] = false;
+			mode_lib->ms.NoOfDPP[k] = 2;
+			mode_lib->ms.NoOfOPP[k] = 2;
+		} else if (display_cfg->plane_descriptors[k].overrides.mpcc_combine_factor == 2) {
+			mode_lib->ms.MPCCombine[k] = true;
+			mode_lib->ms.NoOfDPP[k] = 2;
+		} else if (display_cfg->plane_descriptors[k].overrides.mpcc_combine_factor == 1) {
+			mode_lib->ms.MPCCombine[k] = false;
+			mode_lib->ms.NoOfDPP[k] = 1;
+			if (!mode_lib->ms.SingleDPPViewportSizeSupportPerSurface[k]) {
+				DML_LOG_VERBOSE("WARNING: DML::%s: MPCC is override to disable but viewport is too large to be supported with single pipe!\n", __func__);
+			}
+		} else {
+			if ((mode_lib->ms.MinDPPCLKUsingSingleDPP[k] > mode_lib->ms.max_dppclk_freq_mhz) || !mode_lib->ms.SingleDPPViewportSizeSupportPerSurface[k]) {
+				mode_lib->ms.MPCCombine[k] = true;
+				mode_lib->ms.NoOfDPP[k] = 2;
+			}
+		}
+#if defined(__DML_VBA_DEBUG__)
+		DML_LOG_VERBOSE("DML::%s: k=%d, NoOfDPP = %d\n", __func__, k, mode_lib->ms.NoOfDPP[k]);
+#endif
+	}
+
+	mode_lib->ms.TotalNumberOfActiveDPP = 0;
+	mode_lib->ms.TotalNumberOfActiveOPP = 0;
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		mode_lib->ms.TotalNumberOfActiveDPP += mode_lib->ms.NoOfDPP[k];
+		mode_lib->ms.TotalNumberOfActiveOPP += mode_lib->ms.NoOfOPP[k];
+	}
+	if (mode_lib->ms.TotalNumberOfActiveDPP > (unsigned int)mode_lib->ip.max_num_dpp)
+		mode_lib->ms.support.TotalAvailablePipesSupport = false;
+	if (mode_lib->ms.TotalNumberOfActiveOPP > (unsigned int)mode_lib->ip.max_num_opp)
+		mode_lib->ms.support.TotalAvailablePipesSupport = false;
+
+
+	mode_lib->ms.TotalNumberOfSingleDPPSurfaces = 0;
+	for (k = 0; k < (unsigned int)mode_lib->ms.num_active_planes; ++k) {
+		if (mode_lib->ms.NoOfDPP[k] == 1)
+			mode_lib->ms.TotalNumberOfSingleDPPSurfaces = mode_lib->ms.TotalNumberOfSingleDPPSurfaces + 1;
+	}
+
+	//DISPCLK/DPPCLK
+	mode_lib->ms.WritebackRequiredDISPCLK = 0;
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.active_writebacks_per_stream > 0) {
+			mode_lib->ms.WritebackRequiredDISPCLK = math_max2(mode_lib->ms.WritebackRequiredDISPCLK,
+				CalculateWriteBackDISPCLK(display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].pixel_format,
+					((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000),
+					display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].h_ratio,
+					display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].v_ratio,
+					display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].h_taps,
+					display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].v_taps,
+					display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].input_width,
+					display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].output_width,
+					display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total,
+					mode_lib->ip.writeback_line_buffer_buffer_size));
+		}
+	}
+
+	mode_lib->ms.RequiredDISPCLK = mode_lib->ms.WritebackRequiredDISPCLK;
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		mode_lib->ms.RequiredDISPCLK = math_max2(mode_lib->ms.RequiredDISPCLK, mode_lib->ms.RequiredDISPCLKPerSurface[k]);
+	}
+
+	mode_lib->ms.GlobalDPPCLK = 0;
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		mode_lib->ms.RequiredDPPCLK[k] = mode_lib->ms.MinDPPCLKUsingSingleDPP[k] / mode_lib->ms.NoOfDPP[k];
+		mode_lib->ms.GlobalDPPCLK = math_max2(mode_lib->ms.GlobalDPPCLK, mode_lib->ms.RequiredDPPCLK[k]);
+	}
+
+	mode_lib->ms.support.DISPCLK_DPPCLK_Support = !((mode_lib->ms.RequiredDISPCLK > mode_lib->ms.max_dispclk_freq_mhz) || (mode_lib->ms.GlobalDPPCLK > mode_lib->ms.max_dppclk_freq_mhz));
+
+	/* Total Available OTG, Writeback, HDMIFRL, DP Support Check */
+	s->TotalNumberOfActiveOTG = 0;
+	s->TotalNumberOfActiveHDMIFRL = 0;
+	s->TotalNumberOfActiveDP2p0 = 0;
+	s->TotalNumberOfActiveDP2p0Outputs = 0;
+	s->TotalNumberOfActiveWriteback = 0;
+	memset(s->stream_visited, 0, DML2_MAX_PLANES * sizeof(bool));
+
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		if (!dml_is_phantom_pipe(&display_cfg->plane_descriptors[k])) {
+			if (!s->stream_visited[display_cfg->plane_descriptors[k].stream_index]) {
+				s->stream_visited[display_cfg->plane_descriptors[k].stream_index] = 1;
+
+				if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.active_writebacks_per_stream > 0)
+					s->TotalNumberOfActiveWriteback = s->TotalNumberOfActiveWriteback + 1;
+
+				s->TotalNumberOfActiveOTG = s->TotalNumberOfActiveOTG + 1;
+				if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_encoder == dml2_hdmifrl)
+					s->TotalNumberOfActiveHDMIFRL = s->TotalNumberOfActiveHDMIFRL + 1;
+				if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_encoder == dml2_dp2p0) {
+					s->TotalNumberOfActiveDP2p0 = s->TotalNumberOfActiveDP2p0 + 1;
+					// FIXME_STAGE2: SW not using backend related stuff, need mapping for mst setup
+					//if (display_cfg->output.OutputMultistreamId[k] == k || display_cfg->output.OutputMultistreamEn[k] == false) {
+					s->TotalNumberOfActiveDP2p0Outputs = s->TotalNumberOfActiveDP2p0Outputs + 1;
+					//}
+				}
+			}
+		}
+	}
+
+	/* Writeback Mode Support Check */
+	mode_lib->ms.support.EnoughWritebackUnits = 1;
+	if (s->TotalNumberOfActiveWriteback > (unsigned int)mode_lib->ip.max_num_wb) {
+		mode_lib->ms.support.EnoughWritebackUnits = false;
+	}
+	mode_lib->ms.support.NumberOfOTGSupport = (s->TotalNumberOfActiveOTG <= (unsigned int)mode_lib->ip.max_num_otg);
+	mode_lib->ms.support.NumberOfHDMIFRLSupport = (s->TotalNumberOfActiveHDMIFRL <= (unsigned int)mode_lib->ip.max_num_hdmi_frl_outputs);
+	mode_lib->ms.support.NumberOfDP2p0Support = (s->TotalNumberOfActiveDP2p0 <= (unsigned int)mode_lib->ip.max_num_dp2p0_streams && s->TotalNumberOfActiveDP2p0Outputs <= (unsigned int)mode_lib->ip.max_num_dp2p0_outputs);
+
+
+	mode_lib->ms.support.ExceededMultistreamSlots = false;
+	mode_lib->ms.support.LinkCapacitySupport = true;
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_disabled == false &&
+				(display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_encoder == dml2_dp || display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_encoder == dml2_dp2p0 || display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_encoder == dml2_edp ||
+				display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_encoder == dml2_hdmi || display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_encoder == dml2_hdmifrl) && mode_lib->ms.OutputBpp[k] == 0) {
+			mode_lib->ms.support.LinkCapacitySupport = false;
+		}
+	}
+
+	mode_lib->ms.support.P2IWith420 = false;
+	mode_lib->ms.support.DSCOnlyIfNecessaryWithBPP = false;
+	mode_lib->ms.support.DSC422NativeNotSupported = false;
+	mode_lib->ms.support.LinkRateDoesNotMatchDPVersion = false;
+	mode_lib->ms.support.LinkRateForMultistreamNotIndicated = false;
+	mode_lib->ms.support.BPPForMultistreamNotIndicated = false;
+	mode_lib->ms.support.MultistreamWithHDMIOreDP = false;
+	mode_lib->ms.support.MSOOrODMSplitWithNonDPLink = false;
+	mode_lib->ms.support.NotEnoughLanesForMSO = false;
+
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_encoder == dml2_dp || display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_encoder == dml2_dp2p0 || display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_encoder == dml2_edp ||
+			display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_encoder == dml2_hdmi || display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_encoder == dml2_hdmifrl) {
+			if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_format == dml2_420 && display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.interlaced == 1 && mode_lib->ip.ptoi_supported == true)
+				mode_lib->ms.support.P2IWith420 = true;
+
+			if ((display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.dsc.enable == dml2_dsc_enable || display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.dsc.enable == dml2_dsc_enable_if_necessary) && display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_format == dml2_n422 && !mode_lib->ip.dsc422_native_support)
+				mode_lib->ms.support.DSC422NativeNotSupported = true;
+
+			if (((display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_dp_link_rate == dml2_dp_rate_hbr || display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_dp_link_rate == dml2_dp_rate_hbr2 ||
+				display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_dp_link_rate == dml2_dp_rate_hbr3) &&
+				display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_encoder != dml2_dp && display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_encoder != dml2_edp) ||
+				((display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_dp_link_rate == dml2_dp_rate_uhbr10 || display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_dp_link_rate == dml2_dp_rate_uhbr13p5 ||
+					display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_dp_link_rate == dml2_dp_rate_uhbr20) &&
+					display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_encoder != dml2_dp2p0))
+				mode_lib->ms.support.LinkRateDoesNotMatchDPVersion = true;
+
+			// FIXME_STAGE2
+			//if (display_cfg->output.OutputMultistreamEn[k] == 1) {
+			// if (display_cfg->output.OutputMultistreamId[k] == k && display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_dp_link_rate == dml2_dp_rate_na)
+			// mode_lib->ms.support.LinkRateForMultistreamNotIndicated = true;
+			// if (display_cfg->output.OutputMultistreamId[k] == k && s->OutputBpp[k] == 0)
+			// mode_lib->ms.support.BPPForMultistreamNotIndicated = true;
+			// for (n = 0; n < mode_lib->ms.num_active_planes; ++n) {
+			// if (display_cfg->output.OutputMultistreamId[k] == n && s->OutputBpp[k] == 0)
+			// mode_lib->ms.support.BPPForMultistreamNotIndicated = true;
+			// }
+			//}
+
+			if ((display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_encoder == dml2_edp ||
+				display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_encoder == dml2_hdmi ||
+				display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_encoder == dml2_hdmifrl)) {
+				// FIXME_STAGE2
+				//if (display_cfg->output.OutputMultistreamEn[k] == 1 && display_cfg->output.OutputMultistreamId[k] == k)
+				// mode_lib->ms.support.MultistreamWithHDMIOreDP = true;
+				//for (n = 0; n < mode_lib->ms.num_active_planes; ++n) {
+				// if (display_cfg->output.OutputMultistreamEn[k] == 1 && display_cfg->output.OutputMultistreamId[k] == n)
+				// mode_lib->ms.support.MultistreamWithHDMIOreDP = true;
+				//}
+			}
+			if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_encoder != dml2_dp && (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].overrides.odm_mode == dml2_odm_mode_split_1to2 ||
+				display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].overrides.odm_mode == dml2_odm_mode_mso_1to2 || display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].overrides.odm_mode == dml2_odm_mode_mso_1to4))
+				mode_lib->ms.support.MSOOrODMSplitWithNonDPLink = true;
+
+			if ((display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].overrides.odm_mode == dml2_odm_mode_mso_1to2 && display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_dp_lane_count < 2) ||
+				(display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].overrides.odm_mode == dml2_odm_mode_mso_1to4 && display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_dp_lane_count < 4))
+				mode_lib->ms.support.NotEnoughLanesForMSO = true;
+		}
+	}
+
+	mode_lib->ms.support.DTBCLKRequiredMoreThanSupported = false;
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_encoder == dml2_hdmifrl &&
+				!dml_is_phantom_pipe(&display_cfg->plane_descriptors[k])) {
+			mode_lib->ms.RequiredDTBCLK[k] = RequiredDTBCLK(
+				mode_lib->ms.RequiresDSC[k],
+				s->PixelClockBackEnd[k],
+				display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_format,
+				mode_lib->ms.OutputBpp[k],
+				mode_lib->ms.support.NumberOfDSCSlices[k],
+				display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total,
+				display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_active,
+				display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.audio_sample_rate,
+				display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.audio_sample_layout);
+
+			if (mode_lib->ms.RequiredDTBCLK[k] > ((double)min_clk_table->max_ss_clocks_khz.dtbclk / 1000)) {
+				mode_lib->ms.support.DTBCLKRequiredMoreThanSupported = true;
+			}
+		} else {
+			/* Phantom DTBCLK can be calculated different from main because phantom has no DSC and thus
+			 * will have a different output BPP. Ignore phantom DTBCLK requirement and only consider
+			 * non-phantom DTBCLK requirements. In map_mode_to_soc_dpm we choose the highest DTBCLK
+			 * required - by setting phantom dtbclk to 0 we ignore it.
+			 */
+			mode_lib->ms.RequiredDTBCLK[k] = 0;
+		}
+	}
+
+	mode_lib->ms.support.DSCCLKRequiredMoreThanSupported = false;
+	for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+		if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_encoder == dml2_dp ||
+			display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_encoder == dml2_dp2p0 ||
+			display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_encoder == dml2_edp ||
+			display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_encoder == dml2_hdmifrl) {
+			if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_format == dml2_420) {
+				s->DSCFormatFactor = 2;
+			} else if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_format == dml2_444) {
+				s->DSCFormatFactor = 1;
+			} else if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_format == dml2_n422 || display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_encoder == dml2_hdmifrl) {
+				s->DSCFormatFactor = 2;
+			} else {
+				s->DSCFormatFactor = 1;
+			}
+#ifdef __DML_VBA_DEBUG__
+			DML_LOG_VERBOSE("DML::%s: k=%u, RequiresDSC = %u\n", __func__, k, mode_lib->ms.RequiresDSC[k]);
+#endif
+			if (mode_lib->ms.RequiresDSC[k] == true) {
+				s->PixelClockBackEndFactor = 3.0;
+
+				if (mode_lib->ms.ODMMode[k] == dml2_odm_mode_combine_4to1)
+					s->PixelClockBackEndFactor = 12.0;
+				else if (mode_lib->ms.ODMMode[k] == dml2_odm_mode_combine_3to1)
+					s->PixelClockBackEndFactor = 9.0;
+				else if (mode_lib->ms.ODMMode[k] == dml2_odm_mode_combine_2to1)
+					s->PixelClockBackEndFactor = 6.0;
+
+				mode_lib->ms.required_dscclk_freq_mhz[k] = s->PixelClockBackEnd[k] / s->PixelClockBackEndFactor / (double)s->DSCFormatFactor;
+				if (mode_lib->ms.required_dscclk_freq_mhz[k] > mode_lib->ms.max_dscclk_freq_mhz) {
+					mode_lib->ms.support.DSCCLKRequiredMoreThanSupported = true;
+				}
+
+#ifdef __DML_VBA_DEBUG__
+				DML_LOG_VERBOSE("DML::%s: k=%u, PixelClockBackEnd = %f\n", __func__, k, s->PixelClockBackEnd[k]);
+				DML_LOG_VERBOSE("DML::%s: k=%u, required_dscclk_freq_mhz = %f\n", __func__, k, mode_lib->ms.required_dscclk_freq_mhz[k]);
+				DML_LOG_VERBOSE("DML::%s: k=%u, DSCFormatFactor = %u\n", __func__, k, s->DSCFormatFactor);
+				DML_LOG_VERBOSE("DML::%s: k=%u, DSCCLKRequiredMoreThanSupported = %u\n", __func__, k, mode_lib->ms.support.DSCCLKRequiredMoreThanSupported);
+#endif
+			}
+		}
+	}
+
+	/* Check DSC Unit and Slices Support */
+	mode_lib->ms.support.NotEnoughDSCSlices = false;
+	s->TotalDSCUnitsRequired = 0;
+	mode_lib->ms.support.PixelsPerLinePerDSCUnitSupport = true;
+	memset(s->stream_visited, 0, DML2_MAX_PLANES * sizeof(bool));
+
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		if (mode_lib->ms.RequiresDSC[k] == true && !s->stream_visited[display_cfg->plane_descriptors[k].stream_index]) {
+			s->NumDSCUnitRequired = 1;
+
+			if (mode_lib->ms.ODMMode[k] == dml2_odm_mode_combine_4to1)
+				s->NumDSCUnitRequired = 4;
+			else if (mode_lib->ms.ODMMode[k] == dml2_odm_mode_combine_3to1)
+				s->NumDSCUnitRequired = 3;
+			else if (mode_lib->ms.ODMMode[k] == dml2_odm_mode_combine_2to1)
+				s->NumDSCUnitRequired = 2;
+
+			if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_active > s->NumDSCUnitRequired * (unsigned int)mode_lib->ip.maximum_pixels_per_line_per_dsc_unit)
+				mode_lib->ms.support.PixelsPerLinePerDSCUnitSupport = false;
+			s->TotalDSCUnitsRequired = s->TotalDSCUnitsRequired + s->NumDSCUnitRequired;
+
+			if (mode_lib->ms.support.NumberOfDSCSlices[k] > 4 * s->NumDSCUnitRequired)
+				mode_lib->ms.support.NotEnoughDSCSlices = true;
+		}
+		s->stream_visited[display_cfg->plane_descriptors[k].stream_index] = 1;
+	}
+
+	mode_lib->ms.support.NotEnoughDSCUnits = false;
+	if (s->TotalDSCUnitsRequired > (unsigned int)mode_lib->ip.num_dsc) {
+		mode_lib->ms.support.NotEnoughDSCUnits = true;
+	}
+
+	/*DSC Delay per state*/
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		mode_lib->ms.DSCDelay[k] = DSCDelayRequirement(mode_lib->ms.RequiresDSC[k],
+			mode_lib->ms.ODMMode[k],
+			mode_lib->ip.maximum_dsc_bits_per_component,
+			s->OutputBpp[k],
+			display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_active,
+			display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total,
+			mode_lib->ms.support.NumberOfDSCSlices[k],
+			display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_format,
+			display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_encoder,
+			((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000),
+			s->PixelClockBackEnd[k]);
+	}
+
+	// Figure out the swath and DET configuration after the num dpp per plane is figured out
+	CalculateSwathAndDETConfiguration_params->ForceSingleDPP = false;
+	CalculateSwathAndDETConfiguration_params->ODMMode = mode_lib->ms.ODMMode;
+	CalculateSwathAndDETConfiguration_params->DPPPerSurface = mode_lib->ms.NoOfDPP;
+
+	// output
+	CalculateSwathAndDETConfiguration_params->req_per_swath_ub_l = s->dummy_integer_array[0];
+	CalculateSwathAndDETConfiguration_params->req_per_swath_ub_c = s->dummy_integer_array[1];
+	CalculateSwathAndDETConfiguration_params->swath_width_luma_ub = mode_lib->ms.swath_width_luma_ub;
+	CalculateSwathAndDETConfiguration_params->swath_width_chroma_ub = mode_lib->ms.swath_width_chroma_ub;
+	CalculateSwathAndDETConfiguration_params->SwathWidth = mode_lib->ms.SwathWidthY;
+	CalculateSwathAndDETConfiguration_params->SwathWidthChroma = mode_lib->ms.SwathWidthC;
+	CalculateSwathAndDETConfiguration_params->SwathHeightY = mode_lib->ms.SwathHeightY;
+	CalculateSwathAndDETConfiguration_params->SwathHeightC = mode_lib->ms.SwathHeightC;
+	CalculateSwathAndDETConfiguration_params->request_size_bytes_luma = mode_lib->ms.support.request_size_bytes_luma;
+	CalculateSwathAndDETConfiguration_params->request_size_bytes_chroma = mode_lib->ms.support.request_size_bytes_chroma;
+	CalculateSwathAndDETConfiguration_params->DETBufferSizeInKByte = mode_lib->ms.DETBufferSizeInKByte; // FIXME: This is per pipe but the pipes in plane will use that
+	CalculateSwathAndDETConfiguration_params->DETBufferSizeY = mode_lib->ms.DETBufferSizeY;
+	CalculateSwathAndDETConfiguration_params->DETBufferSizeC = mode_lib->ms.DETBufferSizeC;
+	CalculateSwathAndDETConfiguration_params->UnboundedRequestEnabled = &mode_lib->ms.UnboundedRequestEnabled;
+	CalculateSwathAndDETConfiguration_params->compbuf_reserved_space_64b = s->dummy_integer_array[3];
+	CalculateSwathAndDETConfiguration_params->hw_debug5 = s->dummy_boolean_array[1];
+	CalculateSwathAndDETConfiguration_params->CompressedBufferSizeInkByte = &mode_lib->ms.CompressedBufferSizeInkByte;
+	CalculateSwathAndDETConfiguration_params->ViewportSizeSupportPerSurface = s->dummy_boolean_array[0];
+	CalculateSwathAndDETConfiguration_params->ViewportSizeSupport = &mode_lib->ms.support.ViewportSizeSupport;
+
+	CalculateSwathAndDETConfiguration(&mode_lib->scratch, CalculateSwathAndDETConfiguration_params);
+
+	if (mode_lib->soc.mall_allocated_for_dcn_mbytes == 0) {
+		for (k = 0; k < mode_lib->ms.num_active_planes; k++)
+			mode_lib->ms.SurfaceSizeInMALL[k] = 0;
+		mode_lib->ms.support.ExceededMALLSize = 0;
+	} else {
+		CalculateSurfaceSizeInMall(
+			display_cfg,
+			mode_lib->ms.num_active_planes,
+			mode_lib->soc.mall_allocated_for_dcn_mbytes,
+
+			mode_lib->ms.BytePerPixelY,
+			mode_lib->ms.BytePerPixelC,
+			mode_lib->ms.Read256BlockWidthY,
+			mode_lib->ms.Read256BlockWidthC,
+			mode_lib->ms.Read256BlockHeightY,
+			mode_lib->ms.Read256BlockHeightC,
+			mode_lib->ms.MacroTileWidthY,
+			mode_lib->ms.MacroTileWidthC,
+			mode_lib->ms.MacroTileHeightY,
+			mode_lib->ms.MacroTileHeightC,
+
+			/* Output */
+			mode_lib->ms.SurfaceSizeInMALL,
+			&mode_lib->ms.support.ExceededMALLSize);
+	}
+
+	mode_lib->ms.TotalNumberOfDCCActiveDPP = 0;
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		if (display_cfg->plane_descriptors[k].surface.dcc.enable == true) {
+			mode_lib->ms.TotalNumberOfDCCActiveDPP = mode_lib->ms.TotalNumberOfDCCActiveDPP + mode_lib->ms.NoOfDPP[k];
+		}
+	}
+
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		s->SurfParameters[k].PixelClock = ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000);
+		s->SurfParameters[k].DPPPerSurface = mode_lib->ms.NoOfDPP[k];
+		s->SurfParameters[k].RotationAngle = display_cfg->plane_descriptors[k].composition.rotation_angle;
+		s->SurfParameters[k].ViewportHeight = display_cfg->plane_descriptors[k].composition.viewport.plane0.height;
+		s->SurfParameters[k].ViewportHeightC = display_cfg->plane_descriptors[k].composition.viewport.plane1.height;
+		s->SurfParameters[k].BlockWidth256BytesY = mode_lib->ms.Read256BlockWidthY[k];
+		s->SurfParameters[k].BlockHeight256BytesY = mode_lib->ms.Read256BlockHeightY[k];
+		s->SurfParameters[k].BlockWidth256BytesC = mode_lib->ms.Read256BlockWidthC[k];
+		s->SurfParameters[k].BlockHeight256BytesC = mode_lib->ms.Read256BlockHeightC[k];
+		s->SurfParameters[k].BlockWidthY = mode_lib->ms.MacroTileWidthY[k];
+		s->SurfParameters[k].BlockHeightY = mode_lib->ms.MacroTileHeightY[k];
+		s->SurfParameters[k].BlockWidthC = mode_lib->ms.MacroTileWidthC[k];
+		s->SurfParameters[k].BlockHeightC = mode_lib->ms.MacroTileHeightC[k];
+		s->SurfParameters[k].InterlaceEnable = display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.interlaced;
+		s->SurfParameters[k].HTotal = display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total;
+		s->SurfParameters[k].DCCEnable = display_cfg->plane_descriptors[k].surface.dcc.enable;
+		s->SurfParameters[k].SourcePixelFormat = display_cfg->plane_descriptors[k].pixel_format;
+		s->SurfParameters[k].SurfaceTiling = display_cfg->plane_descriptors[k].surface.tiling;
+		s->SurfParameters[k].BytePerPixelY = mode_lib->ms.BytePerPixelY[k];
+		s->SurfParameters[k].BytePerPixelC = mode_lib->ms.BytePerPixelC[k];
+		s->SurfParameters[k].ProgressiveToInterlaceUnitInOPP = mode_lib->ip.ptoi_supported;
+		s->SurfParameters[k].VRatio = display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio;
+		s->SurfParameters[k].VRatioChroma = display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_ratio;
+		s->SurfParameters[k].VTaps = display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_taps;
+		s->SurfParameters[k].VTapsChroma = display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_taps;
+		s->SurfParameters[k].PitchY = display_cfg->plane_descriptors[k].surface.plane0.pitch;
+		s->SurfParameters[k].PitchC = display_cfg->plane_descriptors[k].surface.plane1.pitch;
+		s->SurfParameters[k].ViewportStationary = display_cfg->plane_descriptors[k].composition.viewport.stationary;
+		s->SurfParameters[k].ViewportXStart = display_cfg->plane_descriptors[k].composition.viewport.plane0.x_start;
+		s->SurfParameters[k].ViewportYStart = display_cfg->plane_descriptors[k].composition.viewport.plane0.y_start;
+		s->SurfParameters[k].ViewportXStartC = display_cfg->plane_descriptors[k].composition.viewport.plane1.y_start;
+		s->SurfParameters[k].ViewportYStartC = display_cfg->plane_descriptors[k].composition.viewport.plane1.y_start;
+		s->SurfParameters[k].FORCE_ONE_ROW_FOR_FRAME = display_cfg->plane_descriptors[k].overrides.hw.force_one_row_for_frame;
+		s->SurfParameters[k].SwathHeightY = mode_lib->ms.SwathHeightY[k];
+		s->SurfParameters[k].SwathHeightC = mode_lib->ms.SwathHeightC[k];
+
+		s->SurfParameters[k].DCCMetaPitchY = display_cfg->plane_descriptors[k].surface.dcc.plane0.pitch;
+		s->SurfParameters[k].DCCMetaPitchC = display_cfg->plane_descriptors[k].surface.dcc.plane1.pitch;
+	}
+
+	CalculateVMRowAndSwath_params->display_cfg = display_cfg;
+	CalculateVMRowAndSwath_params->NumberOfActiveSurfaces = mode_lib->ms.num_active_planes;
+	CalculateVMRowAndSwath_params->myPipe = s->SurfParameters;
+	CalculateVMRowAndSwath_params->SurfaceSizeInMALL = mode_lib->ms.SurfaceSizeInMALL;
+	CalculateVMRowAndSwath_params->PTEBufferSizeInRequestsLuma = mode_lib->ip.dpte_buffer_size_in_pte_reqs_luma;
+	CalculateVMRowAndSwath_params->PTEBufferSizeInRequestsChroma = mode_lib->ip.dpte_buffer_size_in_pte_reqs_chroma;
+	CalculateVMRowAndSwath_params->MALLAllocatedForDCN = mode_lib->soc.mall_allocated_for_dcn_mbytes;
+	CalculateVMRowAndSwath_params->SwathWidthY = mode_lib->ms.SwathWidthY;
+	CalculateVMRowAndSwath_params->SwathWidthC = mode_lib->ms.SwathWidthC;
+	CalculateVMRowAndSwath_params->HostVMMinPageSize = mode_lib->soc.hostvm_min_page_size_kbytes;
+	CalculateVMRowAndSwath_params->DCCMetaBufferSizeBytes = mode_lib->ip.dcc_meta_buffer_size_bytes;
+	CalculateVMRowAndSwath_params->mrq_present = mode_lib->ip.dcn_mrq_present;
+
+	// output
+	CalculateVMRowAndSwath_params->PTEBufferSizeNotExceeded = mode_lib->ms.PTEBufferSizeNotExceeded;
+	CalculateVMRowAndSwath_params->dpte_row_width_luma_ub = s->dummy_integer_array[12];
+	CalculateVMRowAndSwath_params->dpte_row_width_chroma_ub = s->dummy_integer_array[13];
+	CalculateVMRowAndSwath_params->dpte_row_height_luma = mode_lib->ms.dpte_row_height;
+	CalculateVMRowAndSwath_params->dpte_row_height_chroma = mode_lib->ms.dpte_row_height_chroma;
+	CalculateVMRowAndSwath_params->dpte_row_height_linear_luma = s->dummy_integer_array[14]; // VBA_DELTA
+	CalculateVMRowAndSwath_params->dpte_row_height_linear_chroma = s->dummy_integer_array[15]; // VBA_DELTA
+	CalculateVMRowAndSwath_params->vm_group_bytes = s->dummy_integer_array[16];
+	CalculateVMRowAndSwath_params->dpte_group_bytes = mode_lib->ms.dpte_group_bytes;
+	CalculateVMRowAndSwath_params->PixelPTEReqWidthY = s->dummy_integer_array[17];
+	CalculateVMRowAndSwath_params->PixelPTEReqHeightY = s->dummy_integer_array[18];
+	CalculateVMRowAndSwath_params->PTERequestSizeY = s->dummy_integer_array[19];
+	CalculateVMRowAndSwath_params->PixelPTEReqWidthC = s->dummy_integer_array[20];
+	CalculateVMRowAndSwath_params->PixelPTEReqHeightC = s->dummy_integer_array[21];
+	CalculateVMRowAndSwath_params->PTERequestSizeC = s->dummy_integer_array[22];
+	CalculateVMRowAndSwath_params->vmpg_width_y = s->vmpg_width_y;
+	CalculateVMRowAndSwath_params->vmpg_height_y = s->vmpg_height_y;
+	CalculateVMRowAndSwath_params->vmpg_width_c = s->vmpg_width_c;
+	CalculateVMRowAndSwath_params->vmpg_height_c = s->vmpg_height_c;
+	CalculateVMRowAndSwath_params->dpde0_bytes_per_frame_ub_l = s->dummy_integer_array[23];
+	CalculateVMRowAndSwath_params->dpde0_bytes_per_frame_ub_c = s->dummy_integer_array[24];
+	CalculateVMRowAndSwath_params->PrefetchSourceLinesY = mode_lib->ms.PrefetchLinesY;
+	CalculateVMRowAndSwath_params->PrefetchSourceLinesC = mode_lib->ms.PrefetchLinesC;
+	CalculateVMRowAndSwath_params->VInitPreFillY = mode_lib->ms.PrefillY;
+	CalculateVMRowAndSwath_params->VInitPreFillC = mode_lib->ms.PrefillC;
+	CalculateVMRowAndSwath_params->MaxNumSwathY = mode_lib->ms.MaxNumSwathY;
+	CalculateVMRowAndSwath_params->MaxNumSwathC = mode_lib->ms.MaxNumSwathC;
+	CalculateVMRowAndSwath_params->dpte_row_bw = mode_lib->ms.dpte_row_bw;
+	CalculateVMRowAndSwath_params->PixelPTEBytesPerRow = mode_lib->ms.DPTEBytesPerRow;
+	CalculateVMRowAndSwath_params->dpte_row_bytes_per_row_l = s->dpte_row_bytes_per_row_l;
+	CalculateVMRowAndSwath_params->dpte_row_bytes_per_row_c = s->dpte_row_bytes_per_row_c;
+	CalculateVMRowAndSwath_params->vm_bytes = mode_lib->ms.vm_bytes;
+	CalculateVMRowAndSwath_params->use_one_row_for_frame = mode_lib->ms.use_one_row_for_frame;
+	CalculateVMRowAndSwath_params->use_one_row_for_frame_flip = mode_lib->ms.use_one_row_for_frame_flip;
+	CalculateVMRowAndSwath_params->is_using_mall_for_ss = s->dummy_boolean_array[0];
+	CalculateVMRowAndSwath_params->PTE_BUFFER_MODE = s->dummy_boolean_array[1];
+	CalculateVMRowAndSwath_params->BIGK_FRAGMENT_SIZE = s->dummy_integer_array[25];
+	CalculateVMRowAndSwath_params->DCCMetaBufferSizeNotExceeded = mode_lib->ms.DCCMetaBufferSizeNotExceeded;
+	CalculateVMRowAndSwath_params->meta_row_bw = mode_lib->ms.meta_row_bw;
+	CalculateVMRowAndSwath_params->meta_row_bytes = mode_lib->ms.meta_row_bytes;
+	CalculateVMRowAndSwath_params->meta_row_bytes_per_row_ub_l = s->meta_row_bytes_per_row_ub_l;
+	CalculateVMRowAndSwath_params->meta_row_bytes_per_row_ub_c = s->meta_row_bytes_per_row_ub_c;
+	CalculateVMRowAndSwath_params->meta_req_width_luma = s->dummy_integer_array[26];
+	CalculateVMRowAndSwath_params->meta_req_height_luma = s->dummy_integer_array[27];
+	CalculateVMRowAndSwath_params->meta_row_width_luma = s->dummy_integer_array[28];
+	CalculateVMRowAndSwath_params->meta_row_height_luma = s->meta_row_height_luma;
+	CalculateVMRowAndSwath_params->meta_pte_bytes_per_frame_ub_l = s->dummy_integer_array[29];
+	CalculateVMRowAndSwath_params->meta_req_width_chroma = s->dummy_integer_array[30];
+	CalculateVMRowAndSwath_params->meta_req_height_chroma = s->dummy_integer_array[31];
+	CalculateVMRowAndSwath_params->meta_row_width_chroma = s->dummy_integer_array[32];
+	CalculateVMRowAndSwath_params->meta_row_height_chroma = s->meta_row_height_chroma;
+	CalculateVMRowAndSwath_params->meta_pte_bytes_per_frame_ub_c = s->dummy_integer_array[33];
+
+	CalculateVMRowAndSwath(&mode_lib->scratch, CalculateVMRowAndSwath_params);
+
+	mode_lib->ms.support.PTEBufferSizeNotExceeded = true;
+	mode_lib->ms.support.DCCMetaBufferSizeNotExceeded = true;
+
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		if (mode_lib->ms.PTEBufferSizeNotExceeded[k] == false)
+			mode_lib->ms.support.PTEBufferSizeNotExceeded = false;
+
+		if (mode_lib->ms.DCCMetaBufferSizeNotExceeded[k] == false)
+			mode_lib->ms.support.DCCMetaBufferSizeNotExceeded = false;
+
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: k=%u, PTEBufferSizeNotExceeded = %u\n", __func__, k, mode_lib->ms.PTEBufferSizeNotExceeded[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u, DCCMetaBufferSizeNotExceeded = %u\n", __func__, k, mode_lib->ms.DCCMetaBufferSizeNotExceeded[k]);
+#endif
+	}
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: PTEBufferSizeNotExceeded = %u\n", __func__, mode_lib->ms.support.PTEBufferSizeNotExceeded);
+	DML_LOG_VERBOSE("DML::%s: DCCMetaBufferSizeNotExceeded = %u\n", __func__, mode_lib->ms.support.DCCMetaBufferSizeNotExceeded);
+#endif
+
+	/* VActive bytes to fetch for UCLK P-State */
+	calculate_bytes_to_fetch_required_to_hide_latency_params->display_cfg = display_cfg;
+	calculate_bytes_to_fetch_required_to_hide_latency_params->mrq_present = mode_lib->ip.dcn_mrq_present;
+
+	calculate_bytes_to_fetch_required_to_hide_latency_params->num_active_planes = mode_lib->ms.num_active_planes;
+	calculate_bytes_to_fetch_required_to_hide_latency_params->num_of_dpp = mode_lib->ms.NoOfDPP;
+	calculate_bytes_to_fetch_required_to_hide_latency_params->meta_row_height_l = s->meta_row_height_luma;
+	calculate_bytes_to_fetch_required_to_hide_latency_params->meta_row_height_c = s->meta_row_height_chroma;
+	calculate_bytes_to_fetch_required_to_hide_latency_params->meta_row_bytes_per_row_ub_l = s->meta_row_bytes_per_row_ub_l;
+	calculate_bytes_to_fetch_required_to_hide_latency_params->meta_row_bytes_per_row_ub_c = s->meta_row_bytes_per_row_ub_c;
+	calculate_bytes_to_fetch_required_to_hide_latency_params->dpte_row_height_l = mode_lib->ms.dpte_row_height;
+	calculate_bytes_to_fetch_required_to_hide_latency_params->dpte_row_height_c = mode_lib->ms.dpte_row_height_chroma;
+	calculate_bytes_to_fetch_required_to_hide_latency_params->dpte_bytes_per_row_l = s->dpte_row_bytes_per_row_l;
+	calculate_bytes_to_fetch_required_to_hide_latency_params->dpte_bytes_per_row_c = s->dpte_row_bytes_per_row_c;
+	calculate_bytes_to_fetch_required_to_hide_latency_params->byte_per_pix_l = mode_lib->ms.BytePerPixelY;
+	calculate_bytes_to_fetch_required_to_hide_latency_params->byte_per_pix_c = mode_lib->ms.BytePerPixelC;
+	calculate_bytes_to_fetch_required_to_hide_latency_params->swath_width_l = mode_lib->ms.SwathWidthY;
+	calculate_bytes_to_fetch_required_to_hide_latency_params->swath_width_c = mode_lib->ms.SwathWidthC;
+	calculate_bytes_to_fetch_required_to_hide_latency_params->swath_height_l = mode_lib->ms.SwathHeightY;
+	calculate_bytes_to_fetch_required_to_hide_latency_params->swath_height_c = mode_lib->ms.SwathHeightC;
+	calculate_bytes_to_fetch_required_to_hide_latency_params->latency_to_hide_us[0] = mode_lib->soc.power_management_parameters.dram_clk_change_blackout_us;
+
+	/* outputs */
+	calculate_bytes_to_fetch_required_to_hide_latency_params->bytes_required_l = s->pstate_bytes_required_l[dml2_pstate_type_uclk];
+	calculate_bytes_to_fetch_required_to_hide_latency_params->bytes_required_c = s->pstate_bytes_required_c[dml2_pstate_type_uclk];
+
+	calculate_bytes_to_fetch_required_to_hide_latency(calculate_bytes_to_fetch_required_to_hide_latency_params);
+
+	/* Excess VActive bandwidth required to fill DET */
+	calculate_excess_vactive_bandwidth_required(
+			display_cfg,
+			mode_lib->ms.num_active_planes,
+			s->pstate_bytes_required_l[dml2_pstate_type_uclk],
+			s->pstate_bytes_required_c[dml2_pstate_type_uclk],
+			/* outputs */
+			mode_lib->ms.excess_vactive_fill_bw_l,
+			mode_lib->ms.excess_vactive_fill_bw_c);
+
+	mode_lib->ms.UrgLatency = CalculateUrgentLatency(
+		mode_lib->soc.qos_parameters.qos_params.dcn32x.urgent_latency_us.base_latency_us,
+		mode_lib->soc.qos_parameters.qos_params.dcn32x.urgent_latency_us.base_latency_pixel_vm_us,
+		mode_lib->soc.qos_parameters.qos_params.dcn32x.urgent_latency_us.base_latency_vm_us,
+		mode_lib->soc.do_urgent_latency_adjustment,
+		mode_lib->soc.qos_parameters.qos_params.dcn32x.urgent_latency_us.scaling_factor_fclk_us,
+		mode_lib->soc.qos_parameters.qos_params.dcn32x.urgent_latency_us.scaling_factor_mhz,
+		mode_lib->ms.FabricClock,
+		mode_lib->ms.uclk_freq_mhz,
+		mode_lib->soc.qos_parameters.qos_type,
+		mode_lib->soc.qos_parameters.qos_params.dcn4x.per_uclk_dpm_params[mode_lib->ms.qos_param_index].urgent_ramp_uclk_cycles,
+		mode_lib->soc.qos_parameters.qos_params.dcn4x.df_qos_response_time_fclk_cycles,
+		mode_lib->soc.qos_parameters.qos_params.dcn4x.max_round_trip_to_furthest_cs_fclk_cycles,
+		mode_lib->soc.qos_parameters.qos_params.dcn4x.mall_overhead_fclk_cycles,
+		mode_lib->soc.qos_parameters.qos_params.dcn4x.umc_urgent_ramp_latency_margin,
+		mode_lib->soc.qos_parameters.qos_params.dcn4x.fabric_max_transport_latency_margin);
+
+	mode_lib->ms.TripToMemory = CalculateTripToMemory(
+		mode_lib->ms.UrgLatency,
+		mode_lib->ms.FabricClock,
+		mode_lib->ms.uclk_freq_mhz,
+		mode_lib->soc.qos_parameters.qos_type,
+		mode_lib->soc.qos_parameters.qos_params.dcn4x.per_uclk_dpm_params[mode_lib->ms.qos_param_index].trip_to_memory_uclk_cycles,
+		mode_lib->soc.qos_parameters.qos_params.dcn4x.max_round_trip_to_furthest_cs_fclk_cycles,
+		mode_lib->soc.qos_parameters.qos_params.dcn4x.mall_overhead_fclk_cycles,
+		mode_lib->soc.qos_parameters.qos_params.dcn4x.umc_max_latency_margin,
+		mode_lib->soc.qos_parameters.qos_params.dcn4x.fabric_max_transport_latency_margin);
+
+	mode_lib->ms.TripToMemory = math_max2(mode_lib->ms.UrgLatency, mode_lib->ms.TripToMemory);
+
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		double line_time_us = (double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000);
+		bool cursor_not_enough_urgent_latency_hiding = false;
+
+		if (display_cfg->plane_descriptors[k].cursor.num_cursors > 0) {
+			calculate_cursor_req_attributes(
+				display_cfg->plane_descriptors[k].cursor.cursor_width,
+				display_cfg->plane_descriptors[k].cursor.cursor_bpp,
+
+				// output
+				&s->cursor_lines_per_chunk[k],
+				&s->cursor_bytes_per_line[k],
+				&s->cursor_bytes_per_chunk[k],
+				&s->cursor_bytes[k]);
+
+			calculate_cursor_urgent_burst_factor(
+				mode_lib->ip.cursor_buffer_size,
+				display_cfg->plane_descriptors[k].cursor.cursor_width,
+				s->cursor_bytes_per_chunk[k],
+				s->cursor_lines_per_chunk[k],
+				line_time_us,
+				mode_lib->ms.UrgLatency,
+
+				// output
+				&mode_lib->ms.UrgentBurstFactorCursor[k],
+				&cursor_not_enough_urgent_latency_hiding);
+		}
+
+		mode_lib->ms.UrgentBurstFactorCursorPre[k] = mode_lib->ms.UrgentBurstFactorCursor[k];
+
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: k=%d, Calling CalculateUrgentBurstFactor\n", __func__, k);
+		DML_LOG_VERBOSE("DML::%s: k=%d, VRatio=%f\n", __func__, k, display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio);
+		DML_LOG_VERBOSE("DML::%s: k=%d, VRatioChroma=%f\n", __func__, k, display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_ratio);
+#endif
+
+		CalculateUrgentBurstFactor(
+			&display_cfg->plane_descriptors[k],
+			mode_lib->ms.swath_width_luma_ub[k],
+			mode_lib->ms.swath_width_chroma_ub[k],
+			mode_lib->ms.SwathHeightY[k],
+			mode_lib->ms.SwathHeightC[k],
+			line_time_us,
+			mode_lib->ms.UrgLatency,
+			display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio,
+			display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_ratio,
+			mode_lib->ms.BytePerPixelInDETY[k],
+			mode_lib->ms.BytePerPixelInDETC[k],
+			mode_lib->ms.DETBufferSizeY[k],
+			mode_lib->ms.DETBufferSizeC[k],
+
+			// Output
+			&mode_lib->ms.UrgentBurstFactorLuma[k],
+			&mode_lib->ms.UrgentBurstFactorChroma[k],
+			&mode_lib->ms.NotEnoughUrgentLatencyHiding[k]);
+
+		mode_lib->ms.NotEnoughUrgentLatencyHiding[k] = mode_lib->ms.NotEnoughUrgentLatencyHiding[k] || cursor_not_enough_urgent_latency_hiding;
+	}
+
+	CalculateDCFCLKDeepSleep(
+		display_cfg,
+		mode_lib->ms.num_active_planes,
+		mode_lib->ms.BytePerPixelY,
+		mode_lib->ms.BytePerPixelC,
+		mode_lib->ms.SwathWidthY,
+		mode_lib->ms.SwathWidthC,
+		mode_lib->ms.NoOfDPP,
+		mode_lib->ms.PSCL_FACTOR,
+		mode_lib->ms.PSCL_FACTOR_CHROMA,
+		mode_lib->ms.RequiredDPPCLK,
+		mode_lib->ms.vactive_sw_bw_l,
+		mode_lib->ms.vactive_sw_bw_c,
+		mode_lib->soc.return_bus_width_bytes,
+
+		/* Output */
+		&mode_lib->ms.dcfclk_deepsleep);
+
+	for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+		if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.active_writebacks_per_stream > 0) {
+			mode_lib->ms.WritebackDelayTime[k] = mode_lib->soc.qos_parameters.writeback.base_latency_us + CalculateWriteBackDelay(
+				display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].pixel_format,
+				display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].h_ratio,
+				display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].v_ratio,
+				display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].v_taps,
+				display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].output_width,
+				display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].output_height,
+				display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].input_height,
+				display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total) / mode_lib->ms.RequiredDISPCLK;
+		} else {
+			mode_lib->ms.WritebackDelayTime[k] = 0.0;
+		}
+	}
+
+	// MaximumVStartup is actually Tvstartup_min in DCN4 programming guide
+	for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+		bool isInterlaceTiming = (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.interlaced && !mode_lib->ip.ptoi_supported);
+		s->MaximumVStartup[k] = CalculateMaxVStartup(
+			mode_lib->ip.ptoi_supported,
+			mode_lib->ip.vblank_nom_default_us,
+			&display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing,
+			mode_lib->ms.WritebackDelayTime[k]);
+		mode_lib->ms.MaxVStartupLines[k] = (isInterlaceTiming ? (2 * s->MaximumVStartup[k]) : s->MaximumVStartup[k]);
+	}
+
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: k=%u, MaximumVStartup = %u\n", __func__, k, s->MaximumVStartup[k]);
+#endif
+
+	/* Immediate Flip and MALL parameters */
+	s->ImmediateFlipRequired = false;
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		s->ImmediateFlipRequired = s->ImmediateFlipRequired || display_cfg->plane_descriptors[k].immediate_flip;
+	}
+
+	mode_lib->ms.support.ImmediateFlipOrHostVMAndPStateWithMALLFullFrameOrPhantomPipe = false;
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		mode_lib->ms.support.ImmediateFlipOrHostVMAndPStateWithMALLFullFrameOrPhantomPipe =
+			mode_lib->ms.support.ImmediateFlipOrHostVMAndPStateWithMALLFullFrameOrPhantomPipe ||
+			((display_cfg->hostvm_enable == true || display_cfg->plane_descriptors[k].immediate_flip == true) &&
+				(display_cfg->plane_descriptors[k].overrides.uclk_pstate_change_strategy == dml2_uclk_pstate_change_strategy_force_mall_full_frame || dml_is_phantom_pipe(&display_cfg->plane_descriptors[k])));
+	}
+
+	mode_lib->ms.support.InvalidCombinationOfMALLUseForPStateAndStaticScreen = false;
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		mode_lib->ms.support.InvalidCombinationOfMALLUseForPStateAndStaticScreen = mode_lib->ms.support.InvalidCombinationOfMALLUseForPStateAndStaticScreen ||
+			((display_cfg->plane_descriptors[k].overrides.refresh_from_mall == dml2_refresh_from_mall_mode_override_force_enable || display_cfg->plane_descriptors[k].overrides.refresh_from_mall == dml2_refresh_from_mall_mode_override_auto) && (dml_is_phantom_pipe(&display_cfg->plane_descriptors[k]))) ||
+			((display_cfg->plane_descriptors[k].overrides.refresh_from_mall == dml2_refresh_from_mall_mode_override_force_disable || display_cfg->plane_descriptors[k].overrides.refresh_from_mall == dml2_refresh_from_mall_mode_override_auto) && (display_cfg->plane_descriptors[k].overrides.uclk_pstate_change_strategy == dml2_uclk_pstate_change_strategy_force_mall_full_frame));
+	}
+
+	s->FullFrameMALLPStateMethod = false;
+	s->SubViewportMALLPStateMethod = false;
+	s->PhantomPipeMALLPStateMethod = false;
+	s->SubViewportMALLRefreshGreaterThan120Hz = false;
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		if (display_cfg->plane_descriptors[k].overrides.uclk_pstate_change_strategy == dml2_uclk_pstate_change_strategy_force_mall_full_frame)
+			s->FullFrameMALLPStateMethod = true;
+		if (display_cfg->plane_descriptors[k].overrides.legacy_svp_config == dml2_svp_mode_override_main_pipe) {
+			s->SubViewportMALLPStateMethod = true;
+			if (!display_cfg->overrides.enable_subvp_implicit_pmo) {
+				// For dv, small frame tests will have very high refresh rate
+				unsigned long long refresh_rate = (unsigned long long) ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz * 1000 /
+					(double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total /
+					(double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.v_total);
+				if (refresh_rate > 120)
+					s->SubViewportMALLRefreshGreaterThan120Hz = true;
+			}
+		}
+		if (dml_is_phantom_pipe(&display_cfg->plane_descriptors[k]))
+			s->PhantomPipeMALLPStateMethod = true;
+	}
+	mode_lib->ms.support.InvalidCombinationOfMALLUseForPState = (s->SubViewportMALLPStateMethod != s->PhantomPipeMALLPStateMethod) ||
+		(s->SubViewportMALLPStateMethod && s->FullFrameMALLPStateMethod) || s->SubViewportMALLRefreshGreaterThan120Hz;
+
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: SubViewportMALLPStateMethod = %u\n", __func__, s->SubViewportMALLPStateMethod);
+	DML_LOG_VERBOSE("DML::%s: PhantomPipeMALLPStateMethod = %u\n", __func__, s->PhantomPipeMALLPStateMethod);
+	DML_LOG_VERBOSE("DML::%s: FullFrameMALLPStateMethod = %u\n", __func__, s->FullFrameMALLPStateMethod);
+	DML_LOG_VERBOSE("DML::%s: SubViewportMALLRefreshGreaterThan120Hz = %u\n", __func__, s->SubViewportMALLRefreshGreaterThan120Hz);
+	DML_LOG_VERBOSE("DML::%s: InvalidCombinationOfMALLUseForPState = %u\n", __func__, mode_lib->ms.support.InvalidCombinationOfMALLUseForPState);
+	DML_LOG_VERBOSE("DML::%s: in_out_params->min_clk_index = %u\n", __func__, in_out_params->min_clk_index);
+	DML_LOG_VERBOSE("DML::%s: mode_lib->ms.DCFCLK = %f\n", __func__, mode_lib->ms.DCFCLK);
+	DML_LOG_VERBOSE("DML::%s: mode_lib->ms.FabricClock = %f\n", __func__, mode_lib->ms.FabricClock);
+	DML_LOG_VERBOSE("DML::%s: mode_lib->ms.uclk_freq_mhz = %f\n", __func__, mode_lib->ms.uclk_freq_mhz);
+#endif
+
+	mode_lib->ms.support.OutstandingRequestsSupport = true;
+	mode_lib->ms.support.OutstandingRequestsUrgencyAvoidance = true;
+
+	mode_lib->ms.support.avg_urgent_latency_us
+		= (mode_lib->soc.qos_parameters.qos_params.dcn4x.per_uclk_dpm_params[mode_lib->ms.qos_param_index].average_latency_when_urgent_uclk_cycles / mode_lib->ms.uclk_freq_mhz
+			* (1 + mode_lib->soc.qos_parameters.qos_params.dcn4x.umc_average_latency_margin / 100.0)
+			+ mode_lib->soc.qos_parameters.qos_params.dcn4x.average_transport_distance_fclk_cycles / mode_lib->ms.FabricClock)
+		* (1 + mode_lib->soc.qos_parameters.qos_params.dcn4x.fabric_average_transport_latency_margin / 100.0);
+
+	mode_lib->ms.support.avg_non_urgent_latency_us
+		= (mode_lib->soc.qos_parameters.qos_params.dcn4x.per_uclk_dpm_params[mode_lib->ms.qos_param_index].average_latency_when_non_urgent_uclk_cycles / mode_lib->ms.uclk_freq_mhz
+			* (1 + mode_lib->soc.qos_parameters.qos_params.dcn4x.umc_average_latency_margin / 100.0)
+			+ mode_lib->soc.qos_parameters.qos_params.dcn4x.average_transport_distance_fclk_cycles / mode_lib->ms.FabricClock)
+		* (1 + mode_lib->soc.qos_parameters.qos_params.dcn4x.fabric_average_transport_latency_margin / 100.0);
+
+	mode_lib->ms.support.max_non_urgent_latency_us
+		= mode_lib->soc.qos_parameters.qos_params.dcn4x.per_uclk_dpm_params[mode_lib->ms.qos_param_index].maximum_latency_when_non_urgent_uclk_cycles
+		/ mode_lib->ms.uclk_freq_mhz * (1 + mode_lib->soc.qos_parameters.qos_params.dcn4x.umc_max_latency_margin / 100.0)
+		+ mode_lib->soc.qos_parameters.qos_params.dcn4x.mall_overhead_fclk_cycles / mode_lib->ms.FabricClock
+		+ mode_lib->soc.qos_parameters.qos_params.dcn4x.max_round_trip_to_furthest_cs_fclk_cycles / mode_lib->ms.FabricClock
+		* (1 + mode_lib->soc.qos_parameters.qos_params.dcn4x.fabric_max_transport_latency_margin / 100.0);
+
+	for (k = 0; k < mode_lib->ms.num_active_planes; k++) {
+
+		if (mode_lib->soc.qos_parameters.qos_type == dml2_qos_param_type_dcn4x) {
+			outstanding_latency_us = (mode_lib->soc.max_outstanding_reqs * mode_lib->ms.support.request_size_bytes_luma[k]
+				/ (mode_lib->ms.DCFCLK * mode_lib->soc.return_bus_width_bytes));
+
+			if (outstanding_latency_us < mode_lib->ms.support.avg_urgent_latency_us) {
+				mode_lib->ms.support.OutstandingRequestsSupport = false;
+			}
+
+			if (outstanding_latency_us < mode_lib->ms.support.avg_non_urgent_latency_us) {
+				mode_lib->ms.support.OutstandingRequestsUrgencyAvoidance = false;
+			}
+
+#ifdef __DML_VBA_DEBUG__
+			DML_LOG_VERBOSE("DML::%s: avg_urgent_latency_us = %f\n", __func__, mode_lib->ms.support.avg_urgent_latency_us);
+			DML_LOG_VERBOSE("DML::%s: avg_non_urgent_latency_us = %f\n", __func__, mode_lib->ms.support.avg_non_urgent_latency_us);
+			DML_LOG_VERBOSE("DML::%s: k=%d, request_size_bytes_luma = %d\n", __func__, k, mode_lib->ms.support.request_size_bytes_luma[k]);
+			DML_LOG_VERBOSE("DML::%s: k=%d, outstanding_latency_us = %f (luma)\n", __func__, k, outstanding_latency_us);
+#endif
+		}
+
+		if (mode_lib->soc.qos_parameters.qos_type == dml2_qos_param_type_dcn4x && mode_lib->ms.BytePerPixelC[k] > 0) {
+			outstanding_latency_us = (mode_lib->soc.max_outstanding_reqs * mode_lib->ms.support.request_size_bytes_chroma[k]
+				/ (mode_lib->ms.DCFCLK * mode_lib->soc.return_bus_width_bytes));
+
+			if (outstanding_latency_us < mode_lib->ms.support.avg_urgent_latency_us) {
+				mode_lib->ms.support.OutstandingRequestsSupport = false;
+			}
+
+			if (outstanding_latency_us < mode_lib->ms.support.avg_non_urgent_latency_us) {
+				mode_lib->ms.support.OutstandingRequestsUrgencyAvoidance = false;
+			}
+#ifdef __DML_VBA_DEBUG__
+			DML_LOG_VERBOSE("DML::%s: k=%d, request_size_bytes_chroma = %d\n", __func__, k, mode_lib->ms.support.request_size_bytes_chroma[k]);
+			DML_LOG_VERBOSE("DML::%s: k=%d, outstanding_latency_us = %f (chroma)\n", __func__, k, outstanding_latency_us);
+#endif
+		}
+	}
+
+	memset(calculate_mcache_setting_params, 0, sizeof(struct dml2_core_calcs_calculate_mcache_setting_params));
+	if (mode_lib->soc.mcache_size_bytes == 0 || mode_lib->ip.dcn_mrq_present) {
+		for (k = 0; k < mode_lib->ms.num_active_planes; k++) {
+			mode_lib->ms.mall_prefetch_sdp_overhead_factor[k] = 1.0;
+			mode_lib->ms.mall_prefetch_dram_overhead_factor[k] = 1.0;
+			mode_lib->ms.dcc_dram_bw_nom_overhead_factor_p0[k] = 1.0;
+			mode_lib->ms.dcc_dram_bw_pref_overhead_factor_p0[k] = 1.0;
+			mode_lib->ms.dcc_dram_bw_nom_overhead_factor_p1[k] = 1.0;
+			mode_lib->ms.dcc_dram_bw_pref_overhead_factor_p1[k] = 1.0;
+		}
+	} else {
+		for (k = 0; k < mode_lib->ms.num_active_planes; k++) {
+			calculate_mcache_setting_params->dcc_enable = display_cfg->plane_descriptors[k].surface.dcc.enable;
+			calculate_mcache_setting_params->num_chans = mode_lib->soc.clk_table.dram_config.channel_count;
+			calculate_mcache_setting_params->mem_word_bytes = mode_lib->soc.mem_word_bytes;
+			calculate_mcache_setting_params->mcache_size_bytes = mode_lib->soc.mcache_size_bytes;
+			calculate_mcache_setting_params->mcache_line_size_bytes = mode_lib->soc.mcache_line_size_bytes;
+			calculate_mcache_setting_params->gpuvm_enable = display_cfg->gpuvm_enable;
+			calculate_mcache_setting_params->gpuvm_page_size_kbytes = display_cfg->plane_descriptors[k].overrides.gpuvm_min_page_size_kbytes;
+
+			calculate_mcache_setting_params->source_format = display_cfg->plane_descriptors[k].pixel_format;
+			calculate_mcache_setting_params->surf_vert = dml_is_vertical_rotation(display_cfg->plane_descriptors[k].composition.rotation_angle);
+			calculate_mcache_setting_params->vp_stationary = display_cfg->plane_descriptors[k].composition.viewport.stationary;
+			calculate_mcache_setting_params->tiling_mode = display_cfg->plane_descriptors[k].surface.tiling;
+			calculate_mcache_setting_params->imall_enable = mode_lib->ip.imall_supported && display_cfg->plane_descriptors[k].overrides.legacy_svp_config == dml2_svp_mode_override_imall;
+
+			calculate_mcache_setting_params->vp_start_x_l = display_cfg->plane_descriptors[k].composition.viewport.plane0.x_start;
+			calculate_mcache_setting_params->vp_start_y_l = display_cfg->plane_descriptors[k].composition.viewport.plane0.y_start;
+			calculate_mcache_setting_params->full_vp_width_l = display_cfg->plane_descriptors[k].composition.viewport.plane0.width;
+			calculate_mcache_setting_params->full_vp_height_l = display_cfg->plane_descriptors[k].composition.viewport.plane0.height;
+			calculate_mcache_setting_params->blk_width_l = mode_lib->ms.MacroTileWidthY[k];
+			calculate_mcache_setting_params->blk_height_l = mode_lib->ms.MacroTileHeightY[k];
+			calculate_mcache_setting_params->vmpg_width_l = s->vmpg_width_y[k];
+			calculate_mcache_setting_params->vmpg_height_l = s->vmpg_height_y[k];
+			calculate_mcache_setting_params->full_swath_bytes_l = s->full_swath_bytes_l[k];
+			calculate_mcache_setting_params->bytes_per_pixel_l = mode_lib->ms.BytePerPixelY[k];
+
+			calculate_mcache_setting_params->vp_start_x_c = display_cfg->plane_descriptors[k].composition.viewport.plane1.x_start;
+			calculate_mcache_setting_params->vp_start_y_c = display_cfg->plane_descriptors[k].composition.viewport.plane1.y_start;
+			calculate_mcache_setting_params->full_vp_width_c = display_cfg->plane_descriptors[k].composition.viewport.plane1.width;
+			calculate_mcache_setting_params->full_vp_height_c = display_cfg->plane_descriptors[k].composition.viewport.plane1.height;
+			calculate_mcache_setting_params->blk_width_c = mode_lib->ms.MacroTileWidthC[k];
+			calculate_mcache_setting_params->blk_height_c = mode_lib->ms.MacroTileHeightC[k];
+			calculate_mcache_setting_params->vmpg_width_c = s->vmpg_width_c[k];
+			calculate_mcache_setting_params->vmpg_height_c = s->vmpg_height_c[k];
+			calculate_mcache_setting_params->full_swath_bytes_c = s->full_swath_bytes_c[k];
+			calculate_mcache_setting_params->bytes_per_pixel_c = mode_lib->ms.BytePerPixelC[k];
+
+			// output
+			calculate_mcache_setting_params->dcc_dram_bw_nom_overhead_factor_l = &mode_lib->ms.dcc_dram_bw_nom_overhead_factor_p0[k];
+			calculate_mcache_setting_params->dcc_dram_bw_pref_overhead_factor_l = &mode_lib->ms.dcc_dram_bw_pref_overhead_factor_p0[k];
+			calculate_mcache_setting_params->dcc_dram_bw_nom_overhead_factor_c = &mode_lib->ms.dcc_dram_bw_nom_overhead_factor_p1[k];
+			calculate_mcache_setting_params->dcc_dram_bw_pref_overhead_factor_c = &mode_lib->ms.dcc_dram_bw_pref_overhead_factor_p1[k];
+
+			calculate_mcache_setting_params->num_mcaches_l = &mode_lib->ms.num_mcaches_l[k];
+			calculate_mcache_setting_params->mcache_row_bytes_l = &mode_lib->ms.mcache_row_bytes_l[k];
+			calculate_mcache_setting_params->mcache_row_bytes_per_channel_l = &mode_lib->ms.mcache_row_bytes_per_channel_l[k];
+			calculate_mcache_setting_params->mcache_offsets_l = mode_lib->ms.mcache_offsets_l[k];
+			calculate_mcache_setting_params->mcache_shift_granularity_l = &mode_lib->ms.mcache_shift_granularity_l[k];
+
+			calculate_mcache_setting_params->num_mcaches_c = &mode_lib->ms.num_mcaches_c[k];
+			calculate_mcache_setting_params->mcache_row_bytes_c = &mode_lib->ms.mcache_row_bytes_c[k];
+			calculate_mcache_setting_params->mcache_row_bytes_per_channel_c = &mode_lib->ms.mcache_row_bytes_per_channel_c[k];
+			calculate_mcache_setting_params->mcache_offsets_c = mode_lib->ms.mcache_offsets_c[k];
+			calculate_mcache_setting_params->mcache_shift_granularity_c = &mode_lib->ms.mcache_shift_granularity_c[k];
+
+			calculate_mcache_setting_params->mall_comb_mcache_l = &mode_lib->ms.mall_comb_mcache_l[k];
+			calculate_mcache_setting_params->mall_comb_mcache_c = &mode_lib->ms.mall_comb_mcache_c[k];
+			calculate_mcache_setting_params->lc_comb_mcache = &mode_lib->ms.lc_comb_mcache[k];
+
+			calculate_mcache_setting(&mode_lib->scratch, calculate_mcache_setting_params);
+		}
+
+		calculate_mall_bw_overhead_factor(
+				mode_lib->ms.mall_prefetch_sdp_overhead_factor,
+				mode_lib->ms.mall_prefetch_dram_overhead_factor,
+
+				// input
+				display_cfg,
+				mode_lib->ms.num_active_planes);
+	}
+
+	// Calculate all the bandwidth available
+	// Need anothe bw for latency evaluation
+	calculate_bandwidth_available(
+		mode_lib->ms.support.avg_bandwidth_available_min, // not used
+		mode_lib->ms.support.avg_bandwidth_available, // not used
+		mode_lib->ms.support.urg_bandwidth_available_min_latency,
+		mode_lib->ms.support.urg_bandwidth_available, // not used
+		mode_lib->ms.support.urg_bandwidth_available_vm_only, // not used
+		mode_lib->ms.support.urg_bandwidth_available_pixel_and_vm, // not used
+
+		&mode_lib->soc,
+		display_cfg->hostvm_enable,
+		mode_lib->ms.DCFCLK,
+		mode_lib->ms.FabricClock,
+		mode_lib->ms.dram_bw_mbps);
+
+	calculate_bandwidth_available(
+		mode_lib->ms.support.avg_bandwidth_available_min,
+		mode_lib->ms.support.avg_bandwidth_available,
+		mode_lib->ms.support.urg_bandwidth_available_min,
+		mode_lib->ms.support.urg_bandwidth_available,
+		mode_lib->ms.support.urg_bandwidth_available_vm_only,
+		mode_lib->ms.support.urg_bandwidth_available_pixel_and_vm,
+
+		&mode_lib->soc,
+		display_cfg->hostvm_enable,
+		mode_lib->ms.MaxDCFCLK,
+		mode_lib->ms.MaxFabricClock,
+#ifdef DML_MODE_SUPPORT_USE_DPM_DRAM_BW
+		mode_lib->ms.dram_bw_mbps);
+#else
+		mode_lib->ms.max_dram_bw_mbps);
+#endif
+
+	// Average BW support check
+	calculate_avg_bandwidth_required(
+		mode_lib->ms.support.avg_bandwidth_required,
+		// input
+		display_cfg,
+		mode_lib->ms.num_active_planes,
+		mode_lib->ms.vactive_sw_bw_l,
+		mode_lib->ms.vactive_sw_bw_c,
+		mode_lib->ms.cursor_bw,
+		mode_lib->ms.dcc_dram_bw_nom_overhead_factor_p0,
+		mode_lib->ms.dcc_dram_bw_nom_overhead_factor_p1,
+		mode_lib->ms.mall_prefetch_dram_overhead_factor,
+		mode_lib->ms.mall_prefetch_sdp_overhead_factor);
+
+	for (m = 0; m < dml2_core_internal_bw_max; m++) { // check sdp and dram
+		mode_lib->ms.support.avg_bandwidth_support_ok[dml2_core_internal_soc_state_sys_idle][m] = 1;
+		mode_lib->ms.support.avg_bandwidth_support_ok[dml2_core_internal_soc_state_sys_active][m] = (mode_lib->ms.support.avg_bandwidth_required[dml2_core_internal_soc_state_sys_active][m] <= mode_lib->ms.support.avg_bandwidth_available[dml2_core_internal_soc_state_sys_active][m]);
+		mode_lib->ms.support.avg_bandwidth_support_ok[dml2_core_internal_soc_state_svp_prefetch][m] = (mode_lib->ms.support.avg_bandwidth_required[dml2_core_internal_soc_state_svp_prefetch][m] <= mode_lib->ms.support.avg_bandwidth_available[dml2_core_internal_soc_state_svp_prefetch][m]);
+	}
+
+	mode_lib->ms.support.AvgBandwidthSupport = true;
+	mode_lib->ms.support.EnoughUrgentLatencyHidingSupport = true;
+	for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
+		if (mode_lib->ms.NotEnoughUrgentLatencyHiding[k]) {
+			mode_lib->ms.support.EnoughUrgentLatencyHidingSupport = false;
+			DML_LOG_VERBOSE("DML::%s: k=%u NotEnoughUrgentLatencyHiding set\n", __func__, k);
+
+		}
+	}
+	for (m = 0; m < dml2_core_internal_soc_state_max; m++) {
+		for (n = 0; n < dml2_core_internal_bw_max; n++) { // check sdp and dram
+			if (!mode_lib->ms.support.avg_bandwidth_support_ok[m][n] && (m == dml2_core_internal_soc_state_sys_active || mode_lib->soc.mall_allocated_for_dcn_mbytes > 0)) {
+				mode_lib->ms.support.AvgBandwidthSupport = false;
+#ifdef __DML_VBA_DEBUG__
+				DML_LOG_VERBOSE("DML::%s: avg_bandwidth_support_ok[%s][%s] not ok\n", __func__, dml2_core_internal_soc_state_type_str(m), dml2_core_internal_bw_type_str(n));
+#endif
+			}
+		}
+	}
+
+	dml_core_ms_prefetch_check(mode_lib, display_cfg);
+
+	mode_lib->ms.support.max_urgent_latency_us = s->mSOCParameters.max_urgent_latency_us;
+
+	//Re-ordering Buffer Support Check
+	if (mode_lib->soc.qos_parameters.qos_type == dml2_qos_param_type_dcn4x) {
+		if (((mode_lib->ip.rob_buffer_size_kbytes - mode_lib->ip.pixel_chunk_size_kbytes) * 1024
+			/ mode_lib->ms.support.non_urg_bandwidth_required_flip[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_sdp]) >= s->mSOCParameters.max_urgent_latency_us) {
+			mode_lib->ms.support.ROBSupport = true;
+		} else {
+			mode_lib->ms.support.ROBSupport = false;
+		}
+	} else {
+		if (mode_lib->ip.rob_buffer_size_kbytes * 1024 >= mode_lib->soc.qos_parameters.qos_params.dcn32x.loaded_round_trip_latency_fclk_cycles * mode_lib->soc.fabric_datapath_to_dcn_data_return_bytes) {
+			mode_lib->ms.support.ROBSupport = true;
+		} else {
+			mode_lib->ms.support.ROBSupport = false;
+		}
+	}
+
+	/* VActive fill time calculations (informative) */
+	calculate_vactive_det_fill_latency(
+			display_cfg,
+			mode_lib->ms.num_active_planes,
+			s->pstate_bytes_required_l[dml2_pstate_type_uclk],
+			s->pstate_bytes_required_c[dml2_pstate_type_uclk],
+			mode_lib->ms.dcc_dram_bw_nom_overhead_factor_p0,
+			mode_lib->ms.dcc_dram_bw_nom_overhead_factor_p1,
+			mode_lib->ms.vactive_sw_bw_l,
+			mode_lib->ms.vactive_sw_bw_c,
+			mode_lib->ms.surface_avg_vactive_required_bw,
+			mode_lib->ms.surface_peak_required_bw,
+			/* outputs */
+			mode_lib->ms.pstate_vactive_det_fill_delay_us[dml2_pstate_type_uclk]);
+
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: max_urgent_latency_us = %f\n", __func__, s->mSOCParameters.max_urgent_latency_us);
+	DML_LOG_VERBOSE("DML::%s: ROBSupport = %u\n", __func__, mode_lib->ms.support.ROBSupport);
+#endif
+
+	/*Mode Support, Voltage State and SOC Configuration*/
+	{
+		if (mode_lib->ms.support.ScaleRatioAndTapsSupport
+			&& mode_lib->ms.support.SourceFormatPixelAndScanSupport
+			&& mode_lib->ms.support.ViewportSizeSupport
+			&& !mode_lib->ms.support.LinkRateDoesNotMatchDPVersion
+			&& !mode_lib->ms.support.LinkRateForMultistreamNotIndicated
+			&& !mode_lib->ms.support.BPPForMultistreamNotIndicated
+			&& !mode_lib->ms.support.MultistreamWithHDMIOreDP
+			&& !mode_lib->ms.support.ExceededMultistreamSlots
+			&& !mode_lib->ms.support.MSOOrODMSplitWithNonDPLink
+			&& !mode_lib->ms.support.NotEnoughLanesForMSO
+			&& !mode_lib->ms.support.P2IWith420
+			&& !mode_lib->ms.support.DSC422NativeNotSupported
+			&& mode_lib->ms.support.DSCSlicesODMModeSupported
+			&& !mode_lib->ms.support.NotEnoughDSCUnits
+			&& !mode_lib->ms.support.NotEnoughDSCSlices
+			&& !mode_lib->ms.support.ImmediateFlipOrHostVMAndPStateWithMALLFullFrameOrPhantomPipe
+			&& !mode_lib->ms.support.InvalidCombinationOfMALLUseForPStateAndStaticScreen
+			&& !mode_lib->ms.support.DSCCLKRequiredMoreThanSupported
+			&& mode_lib->ms.support.PixelsPerLinePerDSCUnitSupport
+			&& !mode_lib->ms.support.DTBCLKRequiredMoreThanSupported
+			&& !mode_lib->ms.support.InvalidCombinationOfMALLUseForPState
+			&& mode_lib->ms.support.ROBSupport
+			&& mode_lib->ms.support.OutstandingRequestsSupport
+			&& mode_lib->ms.support.OutstandingRequestsUrgencyAvoidance
+			&& mode_lib->ms.support.DISPCLK_DPPCLK_Support
+			&& mode_lib->ms.support.TotalAvailablePipesSupport
+			&& mode_lib->ms.support.NumberOfOTGSupport
+			&& mode_lib->ms.support.NumberOfHDMIFRLSupport
+			&& mode_lib->ms.support.NumberOfDP2p0Support
+			&& mode_lib->ms.support.EnoughWritebackUnits
+			&& mode_lib->ms.support.WritebackLatencySupport
+			&& mode_lib->ms.support.WritebackScaleRatioAndTapsSupport
+			&& mode_lib->ms.support.CursorSupport
+			&& mode_lib->ms.support.PitchSupport
+			&& !mode_lib->ms.support.ViewportExceedsSurface
+			&& mode_lib->ms.support.PrefetchSupported
+			&& mode_lib->ms.support.EnoughUrgentLatencyHidingSupport
+			&& mode_lib->ms.support.AvgBandwidthSupport
+			&& mode_lib->ms.support.DynamicMetadataSupported
+			&& mode_lib->ms.support.VRatioInPrefetchSupported
+			&& mode_lib->ms.support.PTEBufferSizeNotExceeded
+			&& mode_lib->ms.support.DCCMetaBufferSizeNotExceeded
+			&& !mode_lib->ms.support.ExceededMALLSize
+			&& mode_lib->ms.support.g6_temp_read_support
+			&& ((!display_cfg->hostvm_enable && !s->ImmediateFlipRequired) || mode_lib->ms.support.ImmediateFlipSupport)) {
+			DML_LOG_VERBOSE("DML::%s: mode is supported\n", __func__);
+			mode_lib->ms.support.ModeSupport = true;
+		} else {
+			DML_LOG_VERBOSE("DML::%s: mode is NOT supported\n", __func__);
+			mode_lib->ms.support.ModeSupport = false;
+		}
+	}
+
+	// Since now the mode_support work on 1 particular power state, so there is only 1 state idx (index 0).
+	DML_LOG_VERBOSE("DML::%s: ModeSupport = %u\n", __func__, mode_lib->ms.support.ModeSupport);
+	DML_LOG_VERBOSE("DML::%s: ImmediateFlipSupport = %u\n", __func__, mode_lib->ms.support.ImmediateFlipSupport);
+
+	for (k = 0; k < mode_lib->ms.num_active_planes; k++) {
+		mode_lib->ms.support.MPCCombineEnable[k] = mode_lib->ms.MPCCombine[k];
+		mode_lib->ms.support.DPPPerSurface[k] = mode_lib->ms.NoOfDPP[k];
+	}
+
+	for (k = 0; k < mode_lib->ms.num_active_planes; k++) {
+		mode_lib->ms.support.ODMMode[k] = mode_lib->ms.ODMMode[k];
+		mode_lib->ms.support.DSCEnabled[k] = mode_lib->ms.RequiresDSC[k];
+		mode_lib->ms.support.FECEnabled[k] = mode_lib->ms.RequiresFEC[k];
+		mode_lib->ms.support.OutputBpp[k] = mode_lib->ms.OutputBpp[k];
+		mode_lib->ms.support.OutputType[k] = mode_lib->ms.OutputType[k];
+		mode_lib->ms.support.OutputRate[k] = mode_lib->ms.OutputRate[k];
+
+#if defined(__DML_VBA_DEBUG__)
+		DML_LOG_VERBOSE("DML::%s: k=%d, ODMMode = %u\n", __func__, k, mode_lib->ms.support.ODMMode[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%d, DSCEnabled = %u\n", __func__, k, mode_lib->ms.support.DSCEnabled[k]);
+#endif
+	}
+
+#if defined(__DML_VBA_DEBUG__)
+	if (!mode_lib->ms.support.ModeSupport)
+		dml2_print_mode_support_info(&mode_lib->ms.support, true);
+
+	DML_LOG_VERBOSE("DML::%s: --- DONE --- \n", __func__);
+#endif
+
+	return mode_lib->ms.support.ModeSupport;
+}
+
+unsigned int dml2_core_calcs_mode_support_ex(struct dml2_core_calcs_mode_support_ex *in_out_params)
+{
+	unsigned int result;
+
+	DML_LOG_VERBOSE("DML::%s: ------------- START ----------\n", __func__);
+	result = dml_core_mode_support(in_out_params);
+
+	if (result)
+		*in_out_params->out_evaluation_info = in_out_params->mode_lib->ms.support;
+
+	DML_LOG_VERBOSE("DML::%s: is_mode_support = %u (min_clk_index=%d)\n", __func__, result, in_out_params->min_clk_index);
+
+	for (unsigned int k = 0; k < in_out_params->in_display_cfg->num_planes; k++)
+		DML_LOG_VERBOSE("DML::%s: plane_%d: reserved_vblank_time_ns = %lu\n", __func__, k, in_out_params->in_display_cfg->plane_descriptors[k].overrides.reserved_vblank_time_ns);
+
+	DML_LOG_VERBOSE("DML::%s: ------------- DONE ----------\n", __func__);
+
+	return result;
+}
+
+static void CalculatePixelDeliveryTimes(
+	const struct dml2_display_cfg *display_cfg,
+	const struct core_display_cfg_support_info *cfg_support_info,
+	unsigned int NumberOfActiveSurfaces,
+	double VRatioPrefetchY[],
+	double VRatioPrefetchC[],
+	unsigned int swath_width_luma_ub[],
+	unsigned int swath_width_chroma_ub[],
+	double PSCL_THROUGHPUT[],
+	double PSCL_THROUGHPUT_CHROMA[],
+	double Dppclk[],
+	unsigned int BytePerPixelC[],
+	unsigned int req_per_swath_ub_l[],
+	unsigned int req_per_swath_ub_c[],
+
+	// Output
+	double DisplayPipeLineDeliveryTimeLuma[],
+	double DisplayPipeLineDeliveryTimeChroma[],
+	double DisplayPipeLineDeliveryTimeLumaPrefetch[],
+	double DisplayPipeLineDeliveryTimeChromaPrefetch[],
+	double DisplayPipeRequestDeliveryTimeLuma[],
+	double DisplayPipeRequestDeliveryTimeChroma[],
+	double DisplayPipeRequestDeliveryTimeLumaPrefetch[],
+	double DisplayPipeRequestDeliveryTimeChromaPrefetch[])
+{
+	for (unsigned int k = 0; k < NumberOfActiveSurfaces; ++k) {
+		double pixel_clock_mhz = ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000);
+
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: k=%u : HRatio = %f\n", __func__, k, display_cfg->plane_descriptors[k].composition.scaler_info.plane0.h_ratio);
+		DML_LOG_VERBOSE("DML::%s: k=%u : VRatio = %f\n", __func__, k, display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio);
+		DML_LOG_VERBOSE("DML::%s: k=%u : HRatioChroma = %f\n", __func__, k, display_cfg->plane_descriptors[k].composition.scaler_info.plane1.h_ratio);
+		DML_LOG_VERBOSE("DML::%s: k=%u : VRatioChroma = %f\n", __func__, k, display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_ratio);
+		DML_LOG_VERBOSE("DML::%s: k=%u : VRatioPrefetchY = %f\n", __func__, k, VRatioPrefetchY[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u : VRatioPrefetchC = %f\n", __func__, k, VRatioPrefetchC[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u : swath_width_luma_ub = %u\n", __func__, k, swath_width_luma_ub[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u : swath_width_chroma_ub = %u\n", __func__, k, swath_width_chroma_ub[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u : PSCL_THROUGHPUT = %f\n", __func__, k, PSCL_THROUGHPUT[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u : PSCL_THROUGHPUT_CHROMA = %f\n", __func__, k, PSCL_THROUGHPUT_CHROMA[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u : DPPPerSurface = %u\n", __func__, k, cfg_support_info->plane_support_info[k].dpps_used);
+		DML_LOG_VERBOSE("DML::%s: k=%u : pixel_clock_mhz = %f\n", __func__, k, pixel_clock_mhz);
+		DML_LOG_VERBOSE("DML::%s: k=%u : Dppclk = %f\n", __func__, k, Dppclk[k]);
+#endif
+		if (display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio <= 1) {
+			DisplayPipeLineDeliveryTimeLuma[k] = swath_width_luma_ub[k] * cfg_support_info->plane_support_info[k].dpps_used / display_cfg->plane_descriptors[k].composition.scaler_info.plane0.h_ratio / pixel_clock_mhz;
+		} else {
+			DisplayPipeLineDeliveryTimeLuma[k] = swath_width_luma_ub[k] / PSCL_THROUGHPUT[k] / Dppclk[k];
+		}
+
+		if (BytePerPixelC[k] == 0) {
+			DisplayPipeLineDeliveryTimeChroma[k] = 0;
+		} else {
+			if (display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_ratio <= 1) {
+				DisplayPipeLineDeliveryTimeChroma[k] = swath_width_chroma_ub[k] * cfg_support_info->plane_support_info[k].dpps_used / display_cfg->plane_descriptors[k].composition.scaler_info.plane1.h_ratio / pixel_clock_mhz;
+			} else {
+				DisplayPipeLineDeliveryTimeChroma[k] = swath_width_chroma_ub[k] / PSCL_THROUGHPUT_CHROMA[k] / Dppclk[k];
+			}
+		}
+
+		if (VRatioPrefetchY[k] <= 1) {
+			DisplayPipeLineDeliveryTimeLumaPrefetch[k] = swath_width_luma_ub[k] * cfg_support_info->plane_support_info[k].dpps_used / display_cfg->plane_descriptors[k].composition.scaler_info.plane0.h_ratio / pixel_clock_mhz;
+		} else {
+			DisplayPipeLineDeliveryTimeLumaPrefetch[k] = swath_width_luma_ub[k] / PSCL_THROUGHPUT[k] / Dppclk[k];
+		}
+
+		if (BytePerPixelC[k] == 0) {
+			DisplayPipeLineDeliveryTimeChromaPrefetch[k] = 0;
+		} else {
+			if (VRatioPrefetchC[k] <= 1) {
+				DisplayPipeLineDeliveryTimeChromaPrefetch[k] = swath_width_chroma_ub[k] * cfg_support_info->plane_support_info[k].dpps_used / display_cfg->plane_descriptors[k].composition.scaler_info.plane1.h_ratio / pixel_clock_mhz;
+			} else {
+				DisplayPipeLineDeliveryTimeChromaPrefetch[k] = swath_width_chroma_ub[k] / PSCL_THROUGHPUT_CHROMA[k] / Dppclk[k];
+			}
+		}
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: k=%u : DisplayPipeLineDeliveryTimeLuma = %f\n", __func__, k, DisplayPipeLineDeliveryTimeLuma[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u : DisplayPipeLineDeliveryTimeLumaPrefetch = %f\n", __func__, k, DisplayPipeLineDeliveryTimeLumaPrefetch[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u : DisplayPipeLineDeliveryTimeChroma = %f\n", __func__, k, DisplayPipeLineDeliveryTimeChroma[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u : DisplayPipeLineDeliveryTimeChromaPrefetch = %f\n", __func__, k, DisplayPipeLineDeliveryTimeChromaPrefetch[k]);
+#endif
+	}
+
+	for (unsigned int k = 0; k < NumberOfActiveSurfaces; ++k) {
+
+		DisplayPipeRequestDeliveryTimeLuma[k] = DisplayPipeLineDeliveryTimeLuma[k] / req_per_swath_ub_l[k];
+		DisplayPipeRequestDeliveryTimeLumaPrefetch[k] = DisplayPipeLineDeliveryTimeLumaPrefetch[k] / req_per_swath_ub_l[k];
+		if (BytePerPixelC[k] == 0) {
+			DisplayPipeRequestDeliveryTimeChroma[k] = 0;
+			DisplayPipeRequestDeliveryTimeChromaPrefetch[k] = 0;
+		} else {
+			DisplayPipeRequestDeliveryTimeChroma[k] = DisplayPipeLineDeliveryTimeChroma[k] / req_per_swath_ub_c[k];
+			DisplayPipeRequestDeliveryTimeChromaPrefetch[k] = DisplayPipeLineDeliveryTimeChromaPrefetch[k] / req_per_swath_ub_c[k];
+		}
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: k=%u : DisplayPipeRequestDeliveryTimeLuma = %f\n", __func__, k, DisplayPipeRequestDeliveryTimeLuma[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u : DisplayPipeRequestDeliveryTimeLumaPrefetch = %f\n", __func__, k, DisplayPipeRequestDeliveryTimeLumaPrefetch[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u : req_per_swath_ub_l = %d\n", __func__, k, req_per_swath_ub_l[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u : DisplayPipeRequestDeliveryTimeChroma = %f\n", __func__, k, DisplayPipeRequestDeliveryTimeChroma[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u : DisplayPipeRequestDeliveryTimeChromaPrefetch = %f\n", __func__, k, DisplayPipeRequestDeliveryTimeChromaPrefetch[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u : req_per_swath_ub_c = %d\n", __func__, k, req_per_swath_ub_c[k]);
+#endif
+	}
+}
+
+static void CalculateMetaAndPTETimes(struct dml2_core_shared_CalculateMetaAndPTETimes_params *p)
+{
+	unsigned int meta_chunk_width;
+	unsigned int min_meta_chunk_width;
+	unsigned int meta_chunk_per_row_int;
+	unsigned int meta_row_remainder;
+	unsigned int meta_chunk_threshold;
+	unsigned int meta_chunks_per_row_ub;
+	unsigned int meta_chunk_width_chroma;
+	unsigned int min_meta_chunk_width_chroma;
+	unsigned int meta_chunk_per_row_int_chroma;
+	unsigned int meta_row_remainder_chroma;
+	unsigned int meta_chunk_threshold_chroma;
+	unsigned int meta_chunks_per_row_ub_chroma;
+	unsigned int dpte_group_width_luma;
+	unsigned int dpte_groups_per_row_luma_ub;
+	unsigned int dpte_group_width_chroma;
+	unsigned int dpte_groups_per_row_chroma_ub;
+	double pixel_clock_mhz;
+
+	for (unsigned int k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+		p->DST_Y_PER_PTE_ROW_NOM_L[k] = p->dpte_row_height[k] / p->display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio;
+		if (p->BytePerPixelC[k] == 0) {
+			p->DST_Y_PER_PTE_ROW_NOM_C[k] = 0;
+		} else {
+			p->DST_Y_PER_PTE_ROW_NOM_C[k] = p->dpte_row_height_chroma[k] / p->display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_ratio;
+		}
+		p->DST_Y_PER_META_ROW_NOM_L[k] = p->meta_row_height[k] / p->display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio;
+		if (p->BytePerPixelC[k] == 0) {
+			p->DST_Y_PER_META_ROW_NOM_C[k] = 0;
+		} else {
+			p->DST_Y_PER_META_ROW_NOM_C[k] = p->meta_row_height_chroma[k] / p->display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_ratio;
+		}
+	}
+
+	for (unsigned int k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+		if (p->display_cfg->plane_descriptors[k].surface.dcc.enable == true && p->mrq_present) {
+			meta_chunk_width = p->MetaChunkSize * 1024 * 256 / p->BytePerPixelY[k] / p->meta_row_height[k];
+			min_meta_chunk_width = p->MinMetaChunkSizeBytes * 256 / p->BytePerPixelY[k] / p->meta_row_height[k];
+			meta_chunk_per_row_int = p->meta_row_width[k] / meta_chunk_width;
+			meta_row_remainder = p->meta_row_width[k] % meta_chunk_width;
+			if (!dml_is_vertical_rotation(p->display_cfg->plane_descriptors[k].composition.rotation_angle)) {
+				meta_chunk_threshold = 2 * min_meta_chunk_width - p->meta_req_width[k];
+			} else {
+				meta_chunk_threshold = 2 * min_meta_chunk_width - p->meta_req_height[k];
+			}
+			if (meta_row_remainder <= meta_chunk_threshold) {
+				meta_chunks_per_row_ub = meta_chunk_per_row_int + 1;
+			} else {
+				meta_chunks_per_row_ub = meta_chunk_per_row_int + 2;
+			}
+			p->TimePerMetaChunkNominal[k] = p->meta_row_height[k] / p->display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio *
+				p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.h_total /
+				(p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000) / meta_chunks_per_row_ub;
+			p->TimePerMetaChunkVBlank[k] = p->dst_y_per_row_vblank[k] * p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.h_total /
+				(p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000) / meta_chunks_per_row_ub;
+			p->TimePerMetaChunkFlip[k] = p->dst_y_per_row_flip[k] * p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.h_total /
+				(p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000) / meta_chunks_per_row_ub;
+			if (p->BytePerPixelC[k] == 0) {
+				p->TimePerChromaMetaChunkNominal[k] = 0;
+				p->TimePerChromaMetaChunkVBlank[k] = 0;
+				p->TimePerChromaMetaChunkFlip[k] = 0;
+			} else {
+				meta_chunk_width_chroma = p->MetaChunkSize * 1024 * 256 / p->BytePerPixelC[k] / p->meta_row_height_chroma[k];
+				min_meta_chunk_width_chroma = p->MinMetaChunkSizeBytes * 256 / p->BytePerPixelC[k] / p->meta_row_height_chroma[k];
+				meta_chunk_per_row_int_chroma = (unsigned int)((double)p->meta_row_width_chroma[k] / meta_chunk_width_chroma);
+				meta_row_remainder_chroma = p->meta_row_width_chroma[k] % meta_chunk_width_chroma;
+				if (!dml_is_vertical_rotation(p->display_cfg->plane_descriptors[k].composition.rotation_angle)) {
+					meta_chunk_threshold_chroma = 2 * min_meta_chunk_width_chroma - p->meta_req_width_chroma[k];
+				} else {
+					meta_chunk_threshold_chroma = 2 * min_meta_chunk_width_chroma - p->meta_req_height_chroma[k];
+				}
+				if (meta_row_remainder_chroma <= meta_chunk_threshold_chroma) {
+					meta_chunks_per_row_ub_chroma = meta_chunk_per_row_int_chroma + 1;
+				} else {
+					meta_chunks_per_row_ub_chroma = meta_chunk_per_row_int_chroma + 2;
+				}
+				p->TimePerChromaMetaChunkNominal[k] = p->meta_row_height_chroma[k] / p->display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_ratio * p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.h_total / (p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000) / meta_chunks_per_row_ub_chroma;
+				p->TimePerChromaMetaChunkVBlank[k] = p->dst_y_per_row_vblank[k] * p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.h_total / (p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000) / meta_chunks_per_row_ub_chroma;
+				p->TimePerChromaMetaChunkFlip[k] = p->dst_y_per_row_flip[k] * p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.h_total / (p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000) / meta_chunks_per_row_ub_chroma;
+			}
+		} else {
+			p->TimePerMetaChunkNominal[k] = 0;
+			p->TimePerMetaChunkVBlank[k] = 0;
+			p->TimePerMetaChunkFlip[k] = 0;
+			p->TimePerChromaMetaChunkNominal[k] = 0;
+			p->TimePerChromaMetaChunkVBlank[k] = 0;
+			p->TimePerChromaMetaChunkFlip[k] = 0;
+		}
+
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: k=%d, DST_Y_PER_META_ROW_NOM_L = %f\n", __func__, k, p->DST_Y_PER_META_ROW_NOM_L[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%d, DST_Y_PER_META_ROW_NOM_C = %f\n", __func__, k, p->DST_Y_PER_META_ROW_NOM_C[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%d, TimePerMetaChunkNominal		  = %f\n", __func__, k, p->TimePerMetaChunkNominal[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%d, TimePerMetaChunkVBlank		   = %f\n", __func__, k, p->TimePerMetaChunkVBlank[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%d, TimePerMetaChunkFlip			 = %f\n", __func__, k, p->TimePerMetaChunkFlip[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%d, TimePerChromaMetaChunkNominal	= %f\n", __func__, k, p->TimePerChromaMetaChunkNominal[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%d, TimePerChromaMetaChunkVBlank	 = %f\n", __func__, k, p->TimePerChromaMetaChunkVBlank[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%d, TimePerChromaMetaChunkFlip	   = %f\n", __func__, k, p->TimePerChromaMetaChunkFlip[k]);
+#endif
+	}
+
+	for (unsigned int k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+		p->DST_Y_PER_PTE_ROW_NOM_L[k] = p->dpte_row_height[k] / p->display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio;
+		if (p->BytePerPixelC[k] == 0) {
+			p->DST_Y_PER_PTE_ROW_NOM_C[k] = 0;
+		} else {
+			p->DST_Y_PER_PTE_ROW_NOM_C[k] = p->dpte_row_height_chroma[k] / p->display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_ratio;
+		}
+	}
+
+	for (unsigned int k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+		pixel_clock_mhz = ((double)p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000);
+
+		if (p->display_cfg->plane_descriptors[k].tdlut.setup_for_tdlut)
+			p->time_per_tdlut_group[k] = 2 * p->dst_y_per_row_vblank[k] * p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.h_total / pixel_clock_mhz / p->tdlut_groups_per_2row_ub[k];
+		else
+			p->time_per_tdlut_group[k] = 0;
+
+		DML_LOG_VERBOSE("DML::%s: k=%u, time_per_tdlut_group = %f\n", __func__, k, p->time_per_tdlut_group[k]);
+
+		if (p->display_cfg->gpuvm_enable == true) {
+			if (!dml_is_vertical_rotation(p->display_cfg->plane_descriptors[k].composition.rotation_angle)) {
+				dpte_group_width_luma = (unsigned int)((double)p->dpte_group_bytes[k] / (double)p->PTERequestSizeY[k] * p->PixelPTEReqWidthY[k]);
+			} else {
+				dpte_group_width_luma = (unsigned int)((double)p->dpte_group_bytes[k] / (double)p->PTERequestSizeY[k] * p->PixelPTEReqHeightY[k]);
+			}
+			if (p->use_one_row_for_frame[k]) {
+				dpte_groups_per_row_luma_ub = (unsigned int)(math_ceil2((double)p->dpte_row_width_luma_ub[k] / (double)dpte_group_width_luma / 2.0, 1.0));
+			} else {
+				dpte_groups_per_row_luma_ub = (unsigned int)(math_ceil2((double)p->dpte_row_width_luma_ub[k] / (double)dpte_group_width_luma, 1.0));
+			}
+			if (dpte_groups_per_row_luma_ub <= 2) {
+				dpte_groups_per_row_luma_ub = dpte_groups_per_row_luma_ub + 1;
+			}
+			DML_LOG_VERBOSE("DML::%s: k=%u, use_one_row_for_frame = %u\n", __func__, k, p->use_one_row_for_frame[k]);
+			DML_LOG_VERBOSE("DML::%s: k=%u, dpte_group_bytes = %u\n", __func__, k, p->dpte_group_bytes[k]);
+			DML_LOG_VERBOSE("DML::%s: k=%u, PTERequestSizeY = %u\n", __func__, k, p->PTERequestSizeY[k]);
+			DML_LOG_VERBOSE("DML::%s: k=%u, PixelPTEReqWidthY = %u\n", __func__, k, p->PixelPTEReqWidthY[k]);
+			DML_LOG_VERBOSE("DML::%s: k=%u, PixelPTEReqHeightY = %u\n", __func__, k, p->PixelPTEReqHeightY[k]);
+			DML_LOG_VERBOSE("DML::%s: k=%u, dpte_row_width_luma_ub = %u\n", __func__, k, p->dpte_row_width_luma_ub[k]);
+			DML_LOG_VERBOSE("DML::%s: k=%u, dpte_group_width_luma = %u\n", __func__, k, dpte_group_width_luma);
+			DML_LOG_VERBOSE("DML::%s: k=%u, dpte_groups_per_row_luma_ub = %u\n", __func__, k, dpte_groups_per_row_luma_ub);
+
+			p->time_per_pte_group_nom_luma[k] = p->DST_Y_PER_PTE_ROW_NOM_L[k] * p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.h_total / pixel_clock_mhz / dpte_groups_per_row_luma_ub;
+			p->time_per_pte_group_vblank_luma[k] = p->dst_y_per_row_vblank[k] * p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.h_total / pixel_clock_mhz / dpte_groups_per_row_luma_ub;
+			p->time_per_pte_group_flip_luma[k] = p->dst_y_per_row_flip[k] * p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.h_total / pixel_clock_mhz / dpte_groups_per_row_luma_ub;
+			if (p->BytePerPixelC[k] == 0) {
+				p->time_per_pte_group_nom_chroma[k] = 0;
+				p->time_per_pte_group_vblank_chroma[k] = 0;
+				p->time_per_pte_group_flip_chroma[k] = 0;
+			} else {
+				if (!dml_is_vertical_rotation(p->display_cfg->plane_descriptors[k].composition.rotation_angle)) {
+					dpte_group_width_chroma = (unsigned int)((double)p->dpte_group_bytes[k] / (double)p->PTERequestSizeC[k] * p->PixelPTEReqWidthC[k]);
+				} else {
+					dpte_group_width_chroma = (unsigned int)((double)p->dpte_group_bytes[k] / (double)p->PTERequestSizeC[k] * p->PixelPTEReqHeightC[k]);
+				}
+
+				if (p->use_one_row_for_frame[k]) {
+					dpte_groups_per_row_chroma_ub = (unsigned int)(math_ceil2((double)p->dpte_row_width_chroma_ub[k] / (double)dpte_group_width_chroma / 2.0, 1.0));
+				} else {
+					dpte_groups_per_row_chroma_ub = (unsigned int)(math_ceil2((double)p->dpte_row_width_chroma_ub[k] / (double)dpte_group_width_chroma, 1.0));
+				}
+				if (dpte_groups_per_row_chroma_ub <= 2) {
+					dpte_groups_per_row_chroma_ub = dpte_groups_per_row_chroma_ub + 1;
+				}
+				DML_LOG_VERBOSE("DML::%s: k=%u, dpte_row_width_chroma_ub = %u\n", __func__, k, p->dpte_row_width_chroma_ub[k]);
+				DML_LOG_VERBOSE("DML::%s: k=%u, dpte_group_width_chroma = %u\n", __func__, k, dpte_group_width_chroma);
+				DML_LOG_VERBOSE("DML::%s: k=%u, dpte_groups_per_row_chroma_ub = %u\n", __func__, k, dpte_groups_per_row_chroma_ub);
+
+				p->time_per_pte_group_nom_chroma[k] = p->DST_Y_PER_PTE_ROW_NOM_C[k] * p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.h_total / pixel_clock_mhz / dpte_groups_per_row_chroma_ub;
+				p->time_per_pte_group_vblank_chroma[k] = p->dst_y_per_row_vblank[k] * p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.h_total / pixel_clock_mhz / dpte_groups_per_row_chroma_ub;
+				p->time_per_pte_group_flip_chroma[k] = p->dst_y_per_row_flip[k] * p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.h_total / pixel_clock_mhz / dpte_groups_per_row_chroma_ub;
+			}
+		} else {
+			p->time_per_pte_group_nom_luma[k] = 0;
+			p->time_per_pte_group_vblank_luma[k] = 0;
+			p->time_per_pte_group_flip_luma[k] = 0;
+			p->time_per_pte_group_nom_chroma[k] = 0;
+			p->time_per_pte_group_vblank_chroma[k] = 0;
+			p->time_per_pte_group_flip_chroma[k] = 0;
+		}
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: k=%u, dst_y_per_row_vblank = %f\n", __func__, k, p->dst_y_per_row_vblank[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u, dst_y_per_row_flip = %f\n", __func__, k, p->dst_y_per_row_flip[k]);
+
+		DML_LOG_VERBOSE("DML::%s: k=%u, DST_Y_PER_PTE_ROW_NOM_L = %f\n", __func__, k, p->DST_Y_PER_PTE_ROW_NOM_L[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u, DST_Y_PER_PTE_ROW_NOM_C = %f\n", __func__, k, p->DST_Y_PER_PTE_ROW_NOM_C[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u, time_per_pte_group_nom_luma = %f\n", __func__, k, p->time_per_pte_group_nom_luma[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u, time_per_pte_group_vblank_luma = %f\n", __func__, k, p->time_per_pte_group_vblank_luma[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u, time_per_pte_group_flip_luma = %f\n", __func__, k, p->time_per_pte_group_flip_luma[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u, time_per_pte_group_nom_chroma = %f\n", __func__, k, p->time_per_pte_group_nom_chroma[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u, time_per_pte_group_vblank_chroma = %f\n", __func__, k, p->time_per_pte_group_vblank_chroma[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u, time_per_pte_group_flip_chroma = %f\n", __func__, k, p->time_per_pte_group_flip_chroma[k]);
+#endif
+	}
+} // CalculateMetaAndPTETimes
+
+static void CalculateVMGroupAndRequestTimes(
+	const struct dml2_display_cfg *display_cfg,
+	unsigned int NumberOfActiveSurfaces,
+	unsigned int BytePerPixelC[],
+	double dst_y_per_vm_vblank[],
+	double dst_y_per_vm_flip[],
+	unsigned int dpte_row_width_luma_ub[],
+	unsigned int dpte_row_width_chroma_ub[],
+	unsigned int vm_group_bytes[],
+	unsigned int dpde0_bytes_per_frame_ub_l[],
+	unsigned int dpde0_bytes_per_frame_ub_c[],
+	unsigned int tdlut_pte_bytes_per_frame[],
+	unsigned int meta_pte_bytes_per_frame_ub_l[],
+	unsigned int meta_pte_bytes_per_frame_ub_c[],
+	bool mrq_present,
+
+	// Output
+	double TimePerVMGroupVBlank[],
+	double TimePerVMGroupFlip[],
+	double TimePerVMRequestVBlank[],
+	double TimePerVMRequestFlip[])
+{
+	unsigned int num_group_per_lower_vm_stage = 0;
+	unsigned int num_req_per_lower_vm_stage = 0;
+	unsigned int num_group_per_lower_vm_stage_flip;
+	unsigned int num_group_per_lower_vm_stage_pref;
+	unsigned int num_req_per_lower_vm_stage_flip;
+	unsigned int num_req_per_lower_vm_stage_pref;
+	double line_time;
+
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: NumberOfActiveSurfaces = %u\n", __func__, NumberOfActiveSurfaces);
+#endif
+	for (unsigned int k = 0; k < NumberOfActiveSurfaces; ++k) {
+		double pixel_clock_mhz = ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000);
+		bool dcc_mrq_enable = display_cfg->plane_descriptors[k].surface.dcc.enable && mrq_present;
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: k=%u, dcc_mrq_enable = %u\n", __func__, k, dcc_mrq_enable);
+		DML_LOG_VERBOSE("DML::%s: k=%u, vm_group_bytes = %u\n", __func__, k, vm_group_bytes[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u, dpde0_bytes_per_frame_ub_l = %u\n", __func__, k, dpde0_bytes_per_frame_ub_l[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u, dpde0_bytes_per_frame_ub_c = %u\n", __func__, k, dpde0_bytes_per_frame_ub_c[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%d, meta_pte_bytes_per_frame_ub_l = %d\n", __func__, k, meta_pte_bytes_per_frame_ub_l[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%d, meta_pte_bytes_per_frame_ub_c = %d\n", __func__, k, meta_pte_bytes_per_frame_ub_c[k]);
+#endif
+
+		if (display_cfg->gpuvm_enable) {
+			if (display_cfg->gpuvm_max_page_table_levels >= 2) {
+				num_group_per_lower_vm_stage += (unsigned int) math_ceil2((double) (dpde0_bytes_per_frame_ub_l[k]) / (double) (vm_group_bytes[k]), 1);
+
+				if (BytePerPixelC[k] > 0)
+					num_group_per_lower_vm_stage += (unsigned int) math_ceil2((double) (dpde0_bytes_per_frame_ub_c[k]) / (double) (vm_group_bytes[k]), 1);
+			}
+
+			if (dcc_mrq_enable) {
+				if (BytePerPixelC[k] > 0) {
+					num_group_per_lower_vm_stage += (unsigned int)(2.0 /*for each mpde0 group*/ + math_ceil2((double) (meta_pte_bytes_per_frame_ub_l[k]) / (double) (vm_group_bytes[k]), 1) +
+																math_ceil2((double) (meta_pte_bytes_per_frame_ub_c[k]) / (double) (vm_group_bytes[k]), 1));
+				} else {
+					num_group_per_lower_vm_stage += (unsigned int)(1.0 + math_ceil2((double) (meta_pte_bytes_per_frame_ub_l[k]) / (double) (vm_group_bytes[k]), 1));
+				}
+			}
+
+			num_group_per_lower_vm_stage_flip = num_group_per_lower_vm_stage;
+			num_group_per_lower_vm_stage_pref = num_group_per_lower_vm_stage;
+
+			if (display_cfg->plane_descriptors[k].tdlut.setup_for_tdlut && display_cfg->gpuvm_enable) {
+				num_group_per_lower_vm_stage_pref += (unsigned int) math_ceil2(tdlut_pte_bytes_per_frame[k] / vm_group_bytes[k], 1);
+				if (display_cfg->gpuvm_max_page_table_levels >= 2)
+					num_group_per_lower_vm_stage_pref += 1; // tdpe0 group
+			}
+
+			if (display_cfg->gpuvm_max_page_table_levels >= 2) {
+				num_req_per_lower_vm_stage += dpde0_bytes_per_frame_ub_l[k] / 64;
+				if (BytePerPixelC[k] > 0)
+					num_req_per_lower_vm_stage += dpde0_bytes_per_frame_ub_c[k];
+			}
+
+			if (dcc_mrq_enable) {
+				num_req_per_lower_vm_stage += meta_pte_bytes_per_frame_ub_l[k] / 64;
+				if (BytePerPixelC[k] > 0)
+					num_req_per_lower_vm_stage += meta_pte_bytes_per_frame_ub_c[k] / 64;
+			}
+
+			num_req_per_lower_vm_stage_flip = num_req_per_lower_vm_stage;
+			num_req_per_lower_vm_stage_pref = num_req_per_lower_vm_stage;
+
+			if (display_cfg->plane_descriptors[k].tdlut.setup_for_tdlut && display_cfg->gpuvm_enable) {
+				num_req_per_lower_vm_stage_pref += tdlut_pte_bytes_per_frame[k] / 64;
+			}
+
+			line_time = display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total / pixel_clock_mhz;
+
+			if (num_group_per_lower_vm_stage_pref > 0)
+				TimePerVMGroupVBlank[k] = dst_y_per_vm_vblank[k] * line_time / num_group_per_lower_vm_stage_pref;
+			else
+				TimePerVMGroupVBlank[k] = 0;
+
+			if (num_group_per_lower_vm_stage_flip > 0)
+				TimePerVMGroupFlip[k] = dst_y_per_vm_flip[k] * line_time / num_group_per_lower_vm_stage_flip;
+			else
+				TimePerVMGroupFlip[k] = 0;
+
+			if (num_req_per_lower_vm_stage_pref > 0)
+				TimePerVMRequestVBlank[k] = dst_y_per_vm_vblank[k] * line_time / num_req_per_lower_vm_stage_pref;
+			else
+				TimePerVMRequestVBlank[k] = 0.0;
+			if (num_req_per_lower_vm_stage_flip > 0)
+				TimePerVMRequestFlip[k] = dst_y_per_vm_flip[k] * line_time / num_req_per_lower_vm_stage_flip;
+			else
+				TimePerVMRequestFlip[k] = 0.0;
+
+			DML_LOG_VERBOSE("DML::%s: k=%u, dst_y_per_vm_vblank = %f\n", __func__, k, dst_y_per_vm_vblank[k]);
+			DML_LOG_VERBOSE("DML::%s: k=%u, dst_y_per_vm_flip = %f\n", __func__, k, dst_y_per_vm_flip[k]);
+			DML_LOG_VERBOSE("DML::%s: k=%u, line_time = %f\n", __func__, k, line_time);
+			DML_LOG_VERBOSE("DML::%s: k=%u, num_group_per_lower_vm_stage_pref = %d\n", __func__, k, num_group_per_lower_vm_stage_pref);
+			DML_LOG_VERBOSE("DML::%s: k=%u, num_group_per_lower_vm_stage_flip = %d\n", __func__, k, num_group_per_lower_vm_stage_flip);
+			DML_LOG_VERBOSE("DML::%s: k=%u, num_req_per_lower_vm_stage_pref = %d\n", __func__, k, num_req_per_lower_vm_stage_pref);
+			DML_LOG_VERBOSE("DML::%s: k=%u, num_req_per_lower_vm_stage_flip = %d\n", __func__, k, num_req_per_lower_vm_stage_flip);
+
+			if (display_cfg->gpuvm_max_page_table_levels > 2) {
+				TimePerVMGroupVBlank[k] = TimePerVMGroupVBlank[k] / 2;
+				TimePerVMGroupFlip[k] = TimePerVMGroupFlip[k] / 2;
+				TimePerVMRequestVBlank[k] = TimePerVMRequestVBlank[k] / 2;
+				TimePerVMRequestFlip[k] = TimePerVMRequestFlip[k] / 2;
+			}
+
+		} else {
+			TimePerVMGroupVBlank[k] = 0;
+			TimePerVMGroupFlip[k] = 0;
+			TimePerVMRequestVBlank[k] = 0;
+			TimePerVMRequestFlip[k] = 0;
+		}
+
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: k=%u, TimePerVMGroupVBlank = %f\n", __func__, k, TimePerVMGroupVBlank[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u, TimePerVMGroupFlip = %f\n", __func__, k, TimePerVMGroupFlip[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u, TimePerVMRequestVBlank = %f\n", __func__, k, TimePerVMRequestVBlank[k]);
+		DML_LOG_VERBOSE("DML::%s: k=%u, TimePerVMRequestFlip = %f\n", __func__, k, TimePerVMRequestFlip[k]);
+#endif
+	}
+}
+
+static void CalculateStutterEfficiency(struct dml2_core_internal_scratch *scratch,
+	struct dml2_core_calcs_CalculateStutterEfficiency_params *p)
+{
+	struct dml2_core_calcs_CalculateStutterEfficiency_locals *l = &scratch->CalculateStutterEfficiency_locals;
+
+	unsigned int TotalNumberOfActiveOTG = 0;
+	double SinglePixelClock = 0;
+	unsigned int SingleHTotal = 0;
+	unsigned int SingleVTotal = 0;
+	bool SameTiming = true;
+	bool FoundCriticalSurface = false;
+
+	memset(l, 0, sizeof(struct dml2_core_calcs_CalculateStutterEfficiency_locals));
+
+	for (unsigned int k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+		if (!dml_is_phantom_pipe(&p->display_cfg->plane_descriptors[k])) {
+			if (p->display_cfg->plane_descriptors[k].surface.dcc.enable == true) {
+				if ((dml_is_vertical_rotation(p->display_cfg->plane_descriptors[k].composition.rotation_angle) && p->BlockWidth256BytesY[k] > p->SwathHeightY[k]) || (!dml_is_vertical_rotation(p->display_cfg->plane_descriptors[k].composition.rotation_angle) && p->BlockHeight256BytesY[k] > p->SwathHeightY[k]) || p->DCCYMaxUncompressedBlock[k] < 256) {
+					l->MaximumEffectiveCompressionLuma = 2;
+				} else {
+					l->MaximumEffectiveCompressionLuma = 4;
+				}
+				l->TotalCompressedReadBandwidth = l->TotalCompressedReadBandwidth + p->ReadBandwidthSurfaceLuma[k] / math_min2(p->display_cfg->plane_descriptors[k].surface.dcc.informative.dcc_rate_plane0, l->MaximumEffectiveCompressionLuma);
+#ifdef __DML_VBA_DEBUG__
+				DML_LOG_VERBOSE("DML::%s: k=%u, ReadBandwidthSurfaceLuma = %f\n", __func__, k, p->ReadBandwidthSurfaceLuma[k]);
+				DML_LOG_VERBOSE("DML::%s: k=%u, NetDCCRateLuma = %f\n", __func__, k, p->display_cfg->plane_descriptors[k].surface.dcc.informative.dcc_rate_plane0);
+				DML_LOG_VERBOSE("DML::%s: k=%u, MaximumEffectiveCompressionLuma = %f\n", __func__, k, l->MaximumEffectiveCompressionLuma);
+#endif
+				l->TotalZeroSizeRequestReadBandwidth = l->TotalZeroSizeRequestReadBandwidth + p->ReadBandwidthSurfaceLuma[k] * p->display_cfg->plane_descriptors[k].surface.dcc.informative.fraction_of_zero_size_request_plane0;
+				l->TotalZeroSizeCompressedReadBandwidth = l->TotalZeroSizeCompressedReadBandwidth + p->ReadBandwidthSurfaceLuma[k] * p->display_cfg->plane_descriptors[k].surface.dcc.informative.fraction_of_zero_size_request_plane0 / l->MaximumEffectiveCompressionLuma;
+
+				if (p->ReadBandwidthSurfaceChroma[k] > 0) {
+					if ((dml_is_vertical_rotation(p->display_cfg->plane_descriptors[k].composition.rotation_angle) && p->BlockWidth256BytesC[k] > p->SwathHeightC[k]) || (!dml_is_vertical_rotation(p->display_cfg->plane_descriptors[k].composition.rotation_angle) && p->BlockHeight256BytesC[k] > p->SwathHeightC[k]) || p->DCCCMaxUncompressedBlock[k] < 256) {
+						l->MaximumEffectiveCompressionChroma = 2;
+					} else {
+						l->MaximumEffectiveCompressionChroma = 4;
+					}
+					l->TotalCompressedReadBandwidth = l->TotalCompressedReadBandwidth + p->ReadBandwidthSurfaceChroma[k] / math_min2(p->display_cfg->plane_descriptors[k].surface.dcc.informative.dcc_rate_plane1, l->MaximumEffectiveCompressionChroma);
+#ifdef __DML_VBA_DEBUG__
+					DML_LOG_VERBOSE("DML::%s: k=%u, ReadBandwidthSurfaceChroma = %f\n", __func__, k, p->ReadBandwidthSurfaceChroma[k]);
+					DML_LOG_VERBOSE("DML::%s: k=%u, NetDCCRateChroma = %f\n", __func__, k, p->display_cfg->plane_descriptors[k].surface.dcc.informative.dcc_rate_plane1);
+					DML_LOG_VERBOSE("DML::%s: k=%u, MaximumEffectiveCompressionChroma = %f\n", __func__, k, l->MaximumEffectiveCompressionChroma);
+#endif
+					l->TotalZeroSizeRequestReadBandwidth = l->TotalZeroSizeRequestReadBandwidth + p->ReadBandwidthSurfaceChroma[k] * p->display_cfg->plane_descriptors[k].surface.dcc.informative.fraction_of_zero_size_request_plane1;
+					l->TotalZeroSizeCompressedReadBandwidth = l->TotalZeroSizeCompressedReadBandwidth + p->ReadBandwidthSurfaceChroma[k] * p->display_cfg->plane_descriptors[k].surface.dcc.informative.fraction_of_zero_size_request_plane1 / l->MaximumEffectiveCompressionChroma;
+				}
+			} else {
+				l->TotalCompressedReadBandwidth = l->TotalCompressedReadBandwidth + p->ReadBandwidthSurfaceLuma[k] + p->ReadBandwidthSurfaceChroma[k];
+			}
+			l->TotalRowReadBandwidth = l->TotalRowReadBandwidth + p->DPPPerSurface[k] * (p->meta_row_bw[k] + p->dpte_row_bw[k]);
+		}
+	}
+
+	l->AverageDCCCompressionRate = p->TotalDataReadBandwidth / l->TotalCompressedReadBandwidth;
+	l->AverageDCCZeroSizeFraction = l->TotalZeroSizeRequestReadBandwidth / p->TotalDataReadBandwidth;
+
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: UnboundedRequestEnabled = %u\n", __func__, p->UnboundedRequestEnabled);
+	DML_LOG_VERBOSE("DML::%s: TotalCompressedReadBandwidth = %f\n", __func__, l->TotalCompressedReadBandwidth);
+	DML_LOG_VERBOSE("DML::%s: TotalZeroSizeRequestReadBandwidth = %f\n", __func__, l->TotalZeroSizeRequestReadBandwidth);
+	DML_LOG_VERBOSE("DML::%s: TotalZeroSizeCompressedReadBandwidth = %f\n", __func__, l->TotalZeroSizeCompressedReadBandwidth);
+	DML_LOG_VERBOSE("DML::%s: MaximumEffectiveCompressionLuma = %f\n", __func__, l->MaximumEffectiveCompressionLuma);
+	DML_LOG_VERBOSE("DML::%s: MaximumEffectiveCompressionChroma = %f\n", __func__, l->MaximumEffectiveCompressionChroma);
+	DML_LOG_VERBOSE("DML::%s: AverageDCCCompressionRate = %f\n", __func__, l->AverageDCCCompressionRate);
+	DML_LOG_VERBOSE("DML::%s: AverageDCCZeroSizeFraction = %f\n", __func__, l->AverageDCCZeroSizeFraction);
+
+	DML_LOG_VERBOSE("DML::%s: CompbufReservedSpace64B = %u (%f kbytes)\n", __func__, p->CompbufReservedSpace64B, p->CompbufReservedSpace64B * 64 / 1024.0);
+	DML_LOG_VERBOSE("DML::%s: CompbufReservedSpaceZs = %u\n", __func__, p->CompbufReservedSpaceZs);
+	DML_LOG_VERBOSE("DML::%s: CompressedBufferSizeInkByte = %u kbytes\n", __func__, p->CompressedBufferSizeInkByte);
+	DML_LOG_VERBOSE("DML::%s: ROBBufferSizeInKByte = %u kbytes\n", __func__, p->ROBBufferSizeInKByte);
+#endif
+	if (l->AverageDCCZeroSizeFraction == 1) {
+		l->AverageZeroSizeCompressionRate = l->TotalZeroSizeRequestReadBandwidth / l->TotalZeroSizeCompressedReadBandwidth;
+		l->EffectiveCompressedBufferSize = (double)p->MetaFIFOSizeInKEntries * 1024 * 64 * l->AverageZeroSizeCompressionRate + ((double)p->ZeroSizeBufferEntries - p->CompbufReservedSpaceZs) * 64 * l->AverageZeroSizeCompressionRate;
+
+
+	} else if (l->AverageDCCZeroSizeFraction > 0) {
+		l->AverageZeroSizeCompressionRate = l->TotalZeroSizeRequestReadBandwidth / l->TotalZeroSizeCompressedReadBandwidth;
+		l->EffectiveCompressedBufferSize = math_min2((double)p->CompressedBufferSizeInkByte * 1024 * l->AverageDCCCompressionRate,
+			(double)p->MetaFIFOSizeInKEntries * 1024 * 64 / (l->AverageDCCZeroSizeFraction / l->AverageZeroSizeCompressionRate + 1 / l->AverageDCCCompressionRate)) +
+			(p->rob_alloc_compressed ? math_min2(((double)p->ROBBufferSizeInKByte * 1024 - p->CompbufReservedSpace64B * 64) * l->AverageDCCCompressionRate,
+				((double)p->ZeroSizeBufferEntries - p->CompbufReservedSpaceZs) * 64 / (l->AverageDCCZeroSizeFraction / l->AverageZeroSizeCompressionRate))
+				: ((double)p->ROBBufferSizeInKByte * 1024 - p->CompbufReservedSpace64B * 64));
+
+
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: min 1 = %f\n", __func__, p->CompressedBufferSizeInkByte * 1024 * l->AverageDCCCompressionRate);
+		DML_LOG_VERBOSE("DML::%s: min 2 = %f\n", __func__, p->MetaFIFOSizeInKEntries * 1024 * 64 / (l->AverageDCCZeroSizeFraction / l->AverageZeroSizeCompressionRate + 1 / l->AverageDCCCompressionRate));
+		DML_LOG_VERBOSE("DML::%s: min 3 = %d\n", __func__, (p->ROBBufferSizeInKByte * 1024 - p->CompbufReservedSpace64B * 64));
+		DML_LOG_VERBOSE("DML::%s: min 4 = %f\n", __func__, (p->ZeroSizeBufferEntries - p->CompbufReservedSpaceZs) * 64 / (l->AverageDCCZeroSizeFraction / l->AverageZeroSizeCompressionRate));
+#endif
+	} else {
+		l->EffectiveCompressedBufferSize = math_min2((double)p->CompressedBufferSizeInkByte * 1024 * l->AverageDCCCompressionRate,
+			(double)p->MetaFIFOSizeInKEntries * 1024 * 64 * l->AverageDCCCompressionRate) +
+			((double)p->ROBBufferSizeInKByte * 1024 - p->CompbufReservedSpace64B * 64) * (p->rob_alloc_compressed ? l->AverageDCCCompressionRate : 1.0);
+
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: min 1 = %f\n", __func__, p->CompressedBufferSizeInkByte * 1024 * l->AverageDCCCompressionRate);
+		DML_LOG_VERBOSE("DML::%s: min 2 = %f\n", __func__, p->MetaFIFOSizeInKEntries * 1024 * 64 * l->AverageDCCCompressionRate);
+#endif
+	}
+
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: MetaFIFOSizeInKEntries = %u\n", __func__, p->MetaFIFOSizeInKEntries);
+	DML_LOG_VERBOSE("DML::%s: ZeroSizeBufferEntries = %u\n", __func__, p->ZeroSizeBufferEntries);
+	DML_LOG_VERBOSE("DML::%s: AverageZeroSizeCompressionRate = %f\n", __func__, l->AverageZeroSizeCompressionRate);
+	DML_LOG_VERBOSE("DML::%s: EffectiveCompressedBufferSize = %f (%f kbytes)\n", __func__, l->EffectiveCompressedBufferSize, l->EffectiveCompressedBufferSize / 1024.0);
+#endif
+
+	*p->StutterPeriod = 0;
+
+	for (unsigned int k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+		if (!dml_is_phantom_pipe(&p->display_cfg->plane_descriptors[k])) {
+			l->LinesInDETY = ((double)p->DETBufferSizeY[k] + (p->UnboundedRequestEnabled == true ? l->EffectiveCompressedBufferSize : 0) * p->ReadBandwidthSurfaceLuma[k] / p->TotalDataReadBandwidth) / p->BytePerPixelDETY[k] / p->SwathWidthY[k];
+			l->LinesInDETYRoundedDownToSwath = math_floor2(l->LinesInDETY, p->SwathHeightY[k]);
+			l->DETBufferingTimeY = l->LinesInDETYRoundedDownToSwath * ((double)p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000)) / p->display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio;
+#ifdef __DML_VBA_DEBUG__
+			DML_LOG_VERBOSE("DML::%s: k=%u, DETBufferSizeY = %u (%u kbytes)\n", __func__, k, p->DETBufferSizeY[k], p->DETBufferSizeY[k] / 1024);
+			DML_LOG_VERBOSE("DML::%s: k=%u, BytePerPixelDETY = %f\n", __func__, k, p->BytePerPixelDETY[k]);
+			DML_LOG_VERBOSE("DML::%s: k=%u, SwathWidthY = %u\n", __func__, k, p->SwathWidthY[k]);
+			DML_LOG_VERBOSE("DML::%s: k=%u, ReadBandwidthSurfaceLuma = %f\n", __func__, k, p->ReadBandwidthSurfaceLuma[k]);
+			DML_LOG_VERBOSE("DML::%s: k=%u, TotalDataReadBandwidth = %f\n", __func__, k, p->TotalDataReadBandwidth);
+			DML_LOG_VERBOSE("DML::%s: k=%u, LinesInDETY = %f\n", __func__, k, l->LinesInDETY);
+			DML_LOG_VERBOSE("DML::%s: k=%u, LinesInDETYRoundedDownToSwath = %f\n", __func__, k, l->LinesInDETYRoundedDownToSwath);
+			DML_LOG_VERBOSE("DML::%s: k=%u, VRatio = %f\n", __func__, k, p->display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio);
+			DML_LOG_VERBOSE("DML::%s: k=%u, DETBufferingTimeY = %f\n", __func__, k, l->DETBufferingTimeY);
+#endif
+
+			if (!FoundCriticalSurface || l->DETBufferingTimeY < *p->StutterPeriod) {
+				bool isInterlaceTiming = p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.interlaced && !p->ProgressiveToInterlaceUnitInOPP;
+
+				FoundCriticalSurface = true;
+				*p->StutterPeriod = l->DETBufferingTimeY;
+				l->FrameTimeCriticalSurface = (isInterlaceTiming ? math_floor2((double)p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.v_total / 2.0, 1.0) : p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.v_total) * (double)p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000);
+				l->VActiveTimeCriticalSurface = (isInterlaceTiming ? math_floor2((double)p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.v_active / 2.0, 1.0) : p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.v_active) * (double)p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000);
+				l->BytePerPixelYCriticalSurface = p->BytePerPixelY[k];
+				l->SwathWidthYCriticalSurface = p->SwathWidthY[k];
+				l->SwathHeightYCriticalSurface = p->SwathHeightY[k];
+				l->BlockWidth256BytesYCriticalSurface = p->BlockWidth256BytesY[k];
+				l->DETBufferSizeYCriticalSurface = p->DETBufferSizeY[k];
+				l->MinTTUVBlankCriticalSurface = p->MinTTUVBlank[k];
+				l->SinglePlaneCriticalSurface = (p->ReadBandwidthSurfaceChroma[k] == 0);
+				l->SinglePipeCriticalSurface = (p->DPPPerSurface[k] == 1);
+
+#ifdef __DML_VBA_DEBUG__
+				DML_LOG_VERBOSE("DML::%s: k=%u, FoundCriticalSurface = %u\n", __func__, k, FoundCriticalSurface);
+				DML_LOG_VERBOSE("DML::%s: k=%u, StutterPeriod = %f\n", __func__, k, *p->StutterPeriod);
+				DML_LOG_VERBOSE("DML::%s: k=%u, MinTTUVBlankCriticalSurface = %f\n", __func__, k, l->MinTTUVBlankCriticalSurface);
+				DML_LOG_VERBOSE("DML::%s: k=%u, FrameTimeCriticalSurface= %f\n", __func__, k, l->FrameTimeCriticalSurface);
+				DML_LOG_VERBOSE("DML::%s: k=%u, VActiveTimeCriticalSurface = %f\n", __func__, k, l->VActiveTimeCriticalSurface);
+				DML_LOG_VERBOSE("DML::%s: k=%u, BytePerPixelYCriticalSurface = %u\n", __func__, k, l->BytePerPixelYCriticalSurface);
+				DML_LOG_VERBOSE("DML::%s: k=%u, SwathWidthYCriticalSurface = %f\n", __func__, k, l->SwathWidthYCriticalSurface);
+				DML_LOG_VERBOSE("DML::%s: k=%u, SwathHeightYCriticalSurface = %f\n", __func__, k, l->SwathHeightYCriticalSurface);
+				DML_LOG_VERBOSE("DML::%s: k=%u, BlockWidth256BytesYCriticalSurface = %u\n", __func__, k, l->BlockWidth256BytesYCriticalSurface);
+				DML_LOG_VERBOSE("DML::%s: k=%u, SinglePlaneCriticalSurface = %u\n", __func__, k, l->SinglePlaneCriticalSurface);
+				DML_LOG_VERBOSE("DML::%s: k=%u, SinglePipeCriticalSurface = %u\n", __func__, k, l->SinglePipeCriticalSurface);
+#endif
+			}
+		}
+	}
+
+	// for bounded req, the stutter period is calculated only based on DET size, but during burst there can be some return inside ROB/compressed buffer
+	// stutter period is calculated only on the det sizing
+	// if (cdb + rob >= det) the stutter burst will be absorbed by the cdb + rob which is before decompress
+	// else
+	// the cdb + rob part will be in compressed rate with urg bw (idea bw)
+	// the det part will be return at uncompressed rate with 64B/dcfclk
+	//
+	// for unbounded req, the stutter period should be calculated as total of CDB+ROB+DET, so the term "PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer"
+	// should be == EffectiveCompressedBufferSize which will returned a compressed rate, the rest of stutter period is from the DET will be returned at uncompressed rate with 64B/dcfclk
+
+	l->PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer = math_min2(*p->StutterPeriod * p->TotalDataReadBandwidth, l->EffectiveCompressedBufferSize);
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: AverageDCCCompressionRate = %f\n", __func__, l->AverageDCCCompressionRate);
+	DML_LOG_VERBOSE("DML::%s: StutterPeriod*TotalDataReadBandwidth = %f (%f kbytes)\n", __func__, *p->StutterPeriod * p->TotalDataReadBandwidth, (*p->StutterPeriod * p->TotalDataReadBandwidth) / 1024.0);
+	DML_LOG_VERBOSE("DML::%s: EffectiveCompressedBufferSize = %f (%f kbytes)\n", __func__, l->EffectiveCompressedBufferSize, l->EffectiveCompressedBufferSize / 1024.0);
+	DML_LOG_VERBOSE("DML::%s: PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer = %f (%f kbytes)\n", __func__, l->PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer, l->PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer / 1024);
+	DML_LOG_VERBOSE("DML::%s: ReturnBW = %f\n", __func__, p->ReturnBW);
+	DML_LOG_VERBOSE("DML::%s: TotalDataReadBandwidth = %f\n", __func__, p->TotalDataReadBandwidth);
+	DML_LOG_VERBOSE("DML::%s: TotalRowReadBandwidth = %f\n", __func__, l->TotalRowReadBandwidth);
+	DML_LOG_VERBOSE("DML::%s: DCFCLK = %f\n", __func__, p->DCFCLK);
+#endif
+
+	l->StutterBurstTime = l->PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer
+		/ (p->ReturnBW * (p->hw_debug5 ? 1 : l->AverageDCCCompressionRate)) +
+		(*p->StutterPeriod * p->TotalDataReadBandwidth - l->PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer)
+		/ math_min2(p->DCFCLK * 64, p->ReturnBW * (p->hw_debug5 ? 1 : l->AverageDCCCompressionRate)) +
+		*p->StutterPeriod * l->TotalRowReadBandwidth / p->ReturnBW;
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: Part 1 = %f\n", __func__, l->PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer / p->ReturnBW / (p->hw_debug5 ? 1 : l->AverageDCCCompressionRate));
+	DML_LOG_VERBOSE("DML::%s: Part 2 = %f\n", __func__, (*p->StutterPeriod * p->TotalDataReadBandwidth - l->PartOfUncompressedPixelBurstThatFitsInROBAndCompressedBuffer) / (p->DCFCLK * 64));
+	DML_LOG_VERBOSE("DML::%s: Part 3 = %f\n", __func__, *p->StutterPeriod * l->TotalRowReadBandwidth / p->ReturnBW);
+	DML_LOG_VERBOSE("DML::%s: StutterBurstTime = %f\n", __func__, l->StutterBurstTime);
+#endif
+	l->TotalActiveWriteback = 0;
+	memset(l->stream_visited, 0, DML2_MAX_PLANES * sizeof(bool));
+
+	for (unsigned int k = 0; k < p->NumberOfActiveSurfaces; ++k) {
+		if (!dml_is_phantom_pipe(&p->display_cfg->plane_descriptors[k])) {
+			if (!l->stream_visited[p->display_cfg->plane_descriptors[k].stream_index]) {
+
+				if (p->display_cfg->stream_descriptors[k].writeback.active_writebacks_per_stream > 0)
+					l->TotalActiveWriteback = l->TotalActiveWriteback + 1;
+
+				if (TotalNumberOfActiveOTG == 0) { // first otg
+					SinglePixelClock = ((double)p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000);
+					SingleHTotal = p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.h_total;
+					SingleVTotal = p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.v_total;
+				} else if (SinglePixelClock != ((double)p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000) ||
+							SingleHTotal != p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.h_total ||
+							SingleVTotal != p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].timing.v_total) {
+					SameTiming = false;
+				}
+				TotalNumberOfActiveOTG = TotalNumberOfActiveOTG + 1;
+				l->stream_visited[p->display_cfg->plane_descriptors[k].stream_index] = 1;
+			}
+		}
+	}
+
+	if (l->TotalActiveWriteback == 0) {
+#ifdef __DML_VBA_DEBUG__
+		DML_LOG_VERBOSE("DML::%s: SRExitTime = %f\n", __func__, p->SRExitTime);
+		DML_LOG_VERBOSE("DML::%s: SRExitZ8Time = %f\n", __func__, p->SRExitZ8Time);
+		DML_LOG_VERBOSE("DML::%s: StutterPeriod = %f\n", __func__, *p->StutterPeriod);
+#endif
+		*p->StutterEfficiencyNotIncludingVBlank = math_max2(0., 1 - (p->SRExitTime + l->StutterBurstTime) / *p->StutterPeriod) * 100;
+		*p->Z8StutterEfficiencyNotIncludingVBlank = math_max2(0., 1 - (p->SRExitZ8Time + l->StutterBurstTime) / *p->StutterPeriod) * 100;
+		*p->NumberOfStutterBurstsPerFrame = (*p->StutterEfficiencyNotIncludingVBlank > 0 ? (unsigned int)(math_ceil2(l->VActiveTimeCriticalSurface / *p->StutterPeriod, 1)) : 0);
+		*p->Z8NumberOfStutterBurstsPerFrame = (*p->Z8StutterEfficiencyNotIncludingVBlank > 0 ? (unsigned int)(math_ceil2(l->VActiveTimeCriticalSurface / *p->StutterPeriod, 1)) : 0);
+	} else {
+		*p->StutterEfficiencyNotIncludingVBlank = 0.;
+		*p->Z8StutterEfficiencyNotIncludingVBlank = 0.;
+		*p->NumberOfStutterBurstsPerFrame = 0;
+		*p->Z8NumberOfStutterBurstsPerFrame = 0;
+	}
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: VActiveTimeCriticalSurface = %f\n", __func__, l->VActiveTimeCriticalSurface);
+	DML_LOG_VERBOSE("DML::%s: StutterEfficiencyNotIncludingVBlank = %f\n", __func__, *p->StutterEfficiencyNotIncludingVBlank);
+	DML_LOG_VERBOSE("DML::%s: Z8StutterEfficiencyNotIncludingVBlank = %f\n", __func__, *p->Z8StutterEfficiencyNotIncludingVBlank);
+	DML_LOG_VERBOSE("DML::%s: NumberOfStutterBurstsPerFrame = %u\n", __func__, *p->NumberOfStutterBurstsPerFrame);
+	DML_LOG_VERBOSE("DML::%s: Z8NumberOfStutterBurstsPerFrame = %u\n", __func__, *p->Z8NumberOfStutterBurstsPerFrame);
+#endif
+
+	if (*p->StutterEfficiencyNotIncludingVBlank > 0) {
+		if (!((p->SynchronizeTimings || TotalNumberOfActiveOTG == 1) && SameTiming)) {
+			*p->StutterEfficiency = *p->StutterEfficiencyNotIncludingVBlank;
+		} else {
+			*p->StutterEfficiency = (1 - (*p->NumberOfStutterBurstsPerFrame * p->SRExitTime + l->StutterBurstTime * l->VActiveTimeCriticalSurface / *p->StutterPeriod) / l->FrameTimeCriticalSurface) * 100;
+		}
+	} else {
+		*p->StutterEfficiency = 0;
+		*p->NumberOfStutterBurstsPerFrame = 0;
+	}
+
+	if (*p->Z8StutterEfficiencyNotIncludingVBlank > 0) {
+		//LastZ8StutterPeriod = l->VActiveTimeCriticalSurface - (*p->Z8NumberOfStutterBurstsPerFrame - 1) * *p->StutterPeriod;
+		if (!((p->SynchronizeTimings || TotalNumberOfActiveOTG == 1) && SameTiming)) {
+			*p->Z8StutterEfficiency = *p->Z8StutterEfficiencyNotIncludingVBlank;
+		} else {
+			*p->Z8StutterEfficiency = (1 - (*p->Z8NumberOfStutterBurstsPerFrame * p->SRExitZ8Time + l->StutterBurstTime * l->VActiveTimeCriticalSurface / *p->StutterPeriod) / l->FrameTimeCriticalSurface) * 100;
+		}
+	} else {
+		*p->Z8StutterEfficiency = 0.;
+		*p->Z8NumberOfStutterBurstsPerFrame = 0;
+	}
+
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: TotalNumberOfActiveOTG = %u\n", __func__, TotalNumberOfActiveOTG);
+	DML_LOG_VERBOSE("DML::%s: SameTiming = %u\n", __func__, SameTiming);
+	DML_LOG_VERBOSE("DML::%s: SynchronizeTimings = %u\n", __func__, p->SynchronizeTimings);
+	DML_LOG_VERBOSE("DML::%s: LastZ8StutterPeriod = %f\n", __func__, *p->Z8StutterEfficiencyNotIncludingVBlank > 0 ? l->VActiveTimeCriticalSurface - (*p->Z8NumberOfStutterBurstsPerFrame - 1) * *p->StutterPeriod : 0);
+	DML_LOG_VERBOSE("DML::%s: Z8StutterEnterPlusExitWatermark = %f\n", __func__, p->Z8StutterEnterPlusExitWatermark);
+	DML_LOG_VERBOSE("DML::%s: StutterBurstTime = %f\n", __func__, l->StutterBurstTime);
+	DML_LOG_VERBOSE("DML::%s: StutterPeriod = %f\n", __func__, *p->StutterPeriod);
+	DML_LOG_VERBOSE("DML::%s: StutterEfficiency = %f\n", __func__, *p->StutterEfficiency);
+	DML_LOG_VERBOSE("DML::%s: Z8StutterEfficiency = %f\n", __func__, *p->Z8StutterEfficiency);
+	DML_LOG_VERBOSE("DML::%s: StutterEfficiencyNotIncludingVBlank = %f\n", __func__, *p->StutterEfficiencyNotIncludingVBlank);
+	DML_LOG_VERBOSE("DML::%s: Z8NumberOfStutterBurstsPerFrame = %u\n", __func__, *p->Z8NumberOfStutterBurstsPerFrame);
+#endif
+
+	*p->DCHUBBUB_ARB_CSTATE_MAX_CAP_MODE = !(!p->UnboundedRequestEnabled && (p->NumberOfActiveSurfaces == 1) && l->SinglePlaneCriticalSurface && l->SinglePipeCriticalSurface);
+
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: DETBufferSizeYCriticalSurface = %u\n", __func__, l->DETBufferSizeYCriticalSurface);
+	DML_LOG_VERBOSE("DML::%s: PixelChunkSizeInKByte = %u\n", __func__, p->PixelChunkSizeInKByte);
+	DML_LOG_VERBOSE("DML::%s: DCHUBBUB_ARB_CSTATE_MAX_CAP_MODE = %u\n", __func__, *p->DCHUBBUB_ARB_CSTATE_MAX_CAP_MODE);
+#endif
+}
+
+static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex *in_out_params)
+{
+	const struct dml2_display_cfg *display_cfg = in_out_params->in_display_cfg;
+	const struct dml2_mcg_min_clock_table *min_clk_table = in_out_params->min_clk_table;
+	const struct core_display_cfg_support_info *cfg_support_info = in_out_params->cfg_support_info;
+	struct dml2_core_internal_display_mode_lib *mode_lib = in_out_params->mode_lib;
+	struct dml2_display_cfg_programming *programming = in_out_params->programming;
+
+	struct dml2_core_calcs_mode_programming_locals *s = &mode_lib->scratch.dml_core_mode_programming_locals;
+	struct dml2_core_calcs_CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_params *CalculateWatermarks_params = &mode_lib->scratch.CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_params;
+	struct dml2_core_calcs_CalculateVMRowAndSwath_params *CalculateVMRowAndSwath_params = &mode_lib->scratch.CalculateVMRowAndSwath_params;
+	struct dml2_core_calcs_CalculateSwathAndDETConfiguration_params *CalculateSwathAndDETConfiguration_params = &mode_lib->scratch.CalculateSwathAndDETConfiguration_params;
+	struct dml2_core_calcs_CalculateStutterEfficiency_params *CalculateStutterEfficiency_params = &mode_lib->scratch.CalculateStutterEfficiency_params;
+	struct dml2_core_calcs_CalculatePrefetchSchedule_params *CalculatePrefetchSchedule_params = &mode_lib->scratch.CalculatePrefetchSchedule_params;
+	struct dml2_core_calcs_CheckGlobalPrefetchAdmissibility_params *CheckGlobalPrefetchAdmissibility_params = &mode_lib->scratch.CheckGlobalPrefetchAdmissibility_params;
+	struct dml2_core_calcs_calculate_mcache_setting_params *calculate_mcache_setting_params = &mode_lib->scratch.calculate_mcache_setting_params;
+	struct dml2_core_calcs_calculate_tdlut_setting_params *calculate_tdlut_setting_params = &mode_lib->scratch.calculate_tdlut_setting_params;
+	struct dml2_core_shared_CalculateMetaAndPTETimes_params *CalculateMetaAndPTETimes_params = &mode_lib->scratch.CalculateMetaAndPTETimes_params;
+	struct dml2_core_calcs_calculate_peak_bandwidth_required_params *calculate_peak_bandwidth_params = &mode_lib->scratch.calculate_peak_bandwidth_params;
+	struct dml2_core_calcs_calculate_bytes_to_fetch_required_to_hide_latency_params *calculate_bytes_to_fetch_required_to_hide_latency_params = &mode_lib->scratch.calculate_bytes_to_fetch_required_to_hide_latency_params;
+
+	unsigned int k;
+	bool must_support_iflip;
+	const long min_return_uclk_cycles = 83;
+	const long min_return_fclk_cycles = 75;
+	const double max_fclk_mhz = min_clk_table->max_clocks_khz.fclk / 1000.0;
+	double hard_minimum_dcfclk_mhz = (double)min_clk_table->dram_bw_table.entries[0].min_dcfclk_khz / 1000.0;
+	double max_uclk_mhz = 0;
+	double min_return_latency_in_DCFCLK_cycles = 0;
+
+	DML_LOG_VERBOSE("DML::%s: --- START --- \n", __func__);
+
+	memset(&mode_lib->scratch, 0, sizeof(struct dml2_core_internal_scratch));
+	memset(&mode_lib->mp, 0, sizeof(struct dml2_core_internal_mode_program));
+
+	s->num_active_planes = display_cfg->num_planes;
+	get_stream_output_bpp(s->OutputBpp, display_cfg);
+
+	mode_lib->mp.num_active_pipes = dml_get_num_active_pipes(display_cfg->num_planes, cfg_support_info);
+	dml_calc_pipe_plane_mapping(cfg_support_info, mode_lib->mp.pipe_plane);
+
+	mode_lib->mp.Dcfclk = programming->min_clocks.dcn4x.active.dcfclk_khz / 1000.0;
+	mode_lib->mp.FabricClock = programming->min_clocks.dcn4x.active.fclk_khz / 1000.0;
+	mode_lib->mp.dram_bw_mbps = uclk_khz_to_dram_bw_mbps(programming->min_clocks.dcn4x.active.uclk_khz, &mode_lib->soc.clk_table.dram_config);
+	mode_lib->mp.uclk_freq_mhz = programming->min_clocks.dcn4x.active.uclk_khz / 1000.0;
+	mode_lib->mp.GlobalDPPCLK = programming->min_clocks.dcn4x.dpprefclk_khz / 1000.0;
+	s->SOCCLK = (double)programming->min_clocks.dcn4x.socclk_khz / 1000;
+	mode_lib->mp.qos_param_index = get_qos_param_index(programming->min_clocks.dcn4x.active.uclk_khz, mode_lib->soc.qos_parameters.qos_params.dcn4x.per_uclk_dpm_params);
+	mode_lib->mp.active_min_uclk_dpm_index = get_active_min_uclk_dpm_index(programming->min_clocks.dcn4x.active.uclk_khz, &mode_lib->soc.clk_table);
+
+	for (k = 0; k < s->num_active_planes; ++k) {
+		unsigned int stream_index = display_cfg->plane_descriptors[k].stream_index;
+		DML_ASSERT(cfg_support_info->stream_support_info[stream_index].odms_used <= 4);
+		DML_ASSERT(cfg_support_info->stream_support_info[stream_index].num_odm_output_segments == 4 ||
+					cfg_support_info->stream_support_info[stream_index].num_odm_output_segments == 2 ||
+					cfg_support_info->stream_support_info[stream_index].num_odm_output_segments == 1);
+
+		if (cfg_support_info->stream_support_info[stream_index].odms_used > 1)
+			DML_ASSERT(cfg_support_info->stream_support_info[stream_index].num_odm_output_segments == 1);
+
+		switch (cfg_support_info->stream_support_info[stream_index].odms_used) {
+		case (4):
+			mode_lib->mp.ODMMode[k] = dml2_odm_mode_combine_4to1;
+			break;
+		case (3):
+			mode_lib->mp.ODMMode[k] = dml2_odm_mode_combine_3to1;
+			break;
+		case (2):
+			mode_lib->mp.ODMMode[k] = dml2_odm_mode_combine_2to1;
+			break;
+		default:
+			if (cfg_support_info->stream_support_info[stream_index].num_odm_output_segments == 4)
+				mode_lib->mp.ODMMode[k] = dml2_odm_mode_mso_1to4;
+			else if (cfg_support_info->stream_support_info[stream_index].num_odm_output_segments == 2)
+				mode_lib->mp.ODMMode[k] = dml2_odm_mode_mso_1to2;
+			else
+				mode_lib->mp.ODMMode[k] = dml2_odm_mode_bypass;
+			break;
+		}
+	}
+
+	for (k = 0; k < s->num_active_planes; ++k) {
+		mode_lib->mp.NoOfDPP[k] = cfg_support_info->plane_support_info[k].dpps_used;
+		mode_lib->mp.Dppclk[k] = programming->plane_programming[k].min_clocks.dcn4x.dppclk_khz / 1000.0;
+		DML_ASSERT(mode_lib->mp.Dppclk[k] > 0);
+	}
+
+	for (k = 0; k < s->num_active_planes; ++k) {
+		unsigned int stream_index = display_cfg->plane_descriptors[k].stream_index;
+		mode_lib->mp.DSCCLK[k] = programming->stream_programming[stream_index].min_clocks.dcn4x.dscclk_khz / 1000.0;
+		DML_LOG_VERBOSE("DML::%s: k=%d stream_index=%d, mode_lib->mp.DSCCLK = %f\n", __func__, k, stream_index, mode_lib->mp.DSCCLK[k]);
+	}
+
+	mode_lib->mp.Dispclk = programming->min_clocks.dcn4x.dispclk_khz / 1000.0;
+	mode_lib->mp.DCFCLKDeepSleep = programming->min_clocks.dcn4x.deepsleep_dcfclk_khz / 1000.0;
+
+	DML_ASSERT(mode_lib->mp.Dcfclk > 0);
+	DML_ASSERT(mode_lib->mp.FabricClock > 0);
+	DML_ASSERT(mode_lib->mp.dram_bw_mbps > 0);
+	DML_ASSERT(mode_lib->mp.uclk_freq_mhz > 0);
+	DML_ASSERT(mode_lib->mp.GlobalDPPCLK > 0);
+	DML_ASSERT(mode_lib->mp.Dispclk > 0);
+	DML_ASSERT(mode_lib->mp.DCFCLKDeepSleep > 0);
+	DML_ASSERT(s->SOCCLK > 0);
+
+#ifdef __DML_VBA_DEBUG__
+	DML_LOG_VERBOSE("DML::%s: num_active_planes = %u\n", __func__, s->num_active_planes);
+	DML_LOG_VERBOSE("DML::%s: num_active_pipes = %u\n", __func__, mode_lib->mp.num_active_pipes);
+	DML_LOG_VERBOSE("DML::%s: Dcfclk = %f\n", __func__, mode_lib->mp.Dcfclk);
+	DML_LOG_VERBOSE("DML::%s: FabricClock = %f\n", __func__, mode_lib->mp.FabricClock);
+	DML_LOG_VERBOSE("DML::%s: dram_bw_mbps = %f\n", __func__, mode_lib->mp.dram_bw_mbps);
+	DML_LOG_VERBOSE("DML::%s: uclk_freq_mhz = %f\n", __func__, mode_lib->mp.uclk_freq_mhz);
+	DML_LOG_VERBOSE("DML::%s: Dispclk = %f\n", __func__, mode_lib->mp.Dispclk);
+	for (k = 0; k < s->num_active_planes; ++k) {
+		DML_LOG_VERBOSE("DML::%s: Dppclk[%0d] = %f\n", __func__, k, mode_lib->mp.Dppclk[k]);
+	}
+	DML_LOG_VERBOSE("DML::%s: GlobalDPPCLK = %f\n", __func__, mode_lib->mp.GlobalDPPCLK);
+	DML_LOG_VERBOSE("DML::%s: DCFCLKDeepSleep = %f\n", __func__, mode_lib->mp.DCFCLKDeepSleep);
+	DML_LOG_VERBOSE("DML::%s: SOCCLK = %f\n", __func__, s->SOCCLK);
+	DML_LOG_VERBOSE("DML::%s: min_clk_index = %0d\n", __func__, in_out_params->min_clk_index);
+	DML_LOG_VERBOSE("DML::%s: min_clk_table min_fclk_khz = %ld\n", __func__, min_clk_table->dram_bw_table.entries[in_out_params->min_clk_index].min_fclk_khz);
+	DML_LOG_VERBOSE("DML::%s: min_clk_table uclk_mhz = %f\n", __func__, dram_bw_kbps_to_uclk_mhz(min_clk_table->dram_bw_table.entries[in_out_params->min_clk_index].pre_derate_dram_bw_kbps, &mode_lib->soc.clk_table.dram_config));
+	for (k = 0; k < mode_lib->mp.num_active_pipes; ++k) {
+		DML_LOG_VERBOSE("DML::%s: pipe=%d is in plane=%d\n", __func__, k, mode_lib->mp.pipe_plane[k]);
+		DML_LOG_VERBOSE("DML::%s: Per-plane DPPPerSurface[%0d] = %d\n", __func__, k, mode_lib->mp.NoOfDPP[k]);
+	}
+
+	for (k = 0; k < s->num_active_planes; k++)
+		DML_LOG_VERBOSE("DML::%s: plane_%d: reserved_vblank_time_ns = %lu\n", __func__, k, display_cfg->plane_descriptors[k].overrides.reserved_vblank_time_ns);
+#endif
+
+	CalculateMaxDETAndMinCompressedBufferSize(
+		mode_lib->ip.config_return_buffer_size_in_kbytes,
+		mode_lib->ip.config_return_buffer_segment_size_in_kbytes,
+		mode_lib->ip.rob_buffer_size_kbytes,
+		mode_lib->ip.max_num_dpp,
+		display_cfg->overrides.hw.force_nom_det_size_kbytes.enable,
+		display_cfg->overrides.hw.force_nom_det_size_kbytes.value,
+		mode_lib->ip.dcn_mrq_present,
+
+		/* Output */
+		&s->MaxTotalDETInKByte,
+		&s->NomDETInKByte,
+		&s->MinCompressedBufferSizeInKByte);
+
+
+	PixelClockAdjustmentForProgressiveToInterlaceUnit(display_cfg, mode_lib->ip.ptoi_supported, s->PixelClockBackEnd);
+
+	for (k = 0; k < s->num_active_planes; ++k) {
+		CalculateSinglePipeDPPCLKAndSCLThroughput(
+			display_cfg->plane_descriptors[k].composition.scaler_info.plane0.h_ratio,
+			display_cfg->plane_descriptors[k].composition.scaler_info.plane1.h_ratio,
+			display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio,
+			display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_ratio,
+			mode_lib->ip.max_dchub_pscl_bw_pix_per_clk,
+			mode_lib->ip.max_pscl_lb_bw_pix_per_clk,
+			((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000),
+			display_cfg->plane_descriptors[k].pixel_format,
+			display_cfg->plane_descriptors[k].composition.scaler_info.plane0.h_taps,
+			display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_taps,
+			display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_taps,
+			display_cfg->plane_descriptors[k].composition.scaler_info.plane1.h_taps,
+
+			/* Output */
+			&mode_lib->mp.PSCL_THROUGHPUT[k],
+			&mode_lib->mp.PSCL_THROUGHPUT_CHROMA[k],
+			&mode_lib->mp.DPPCLKUsingSingleDPP[k]);
+	}
+
+	for (k = 0; k < s->num_active_planes; ++k) {
+		CalculateBytePerPixelAndBlockSizes(
+			display_cfg->plane_descriptors[k].pixel_format,
+			display_cfg->plane_descriptors[k].surface.tiling,
+			display_cfg->plane_descriptors[k].surface.plane0.pitch,
+			display_cfg->plane_descriptors[k].surface.plane1.pitch,
+
+			// Output
+			&mode_lib->mp.BytePerPixelY[k],
+			&mode_lib->mp.BytePerPixelC[k],
+			&mode_lib->mp.BytePerPixelInDETY[k],
+			&mode_lib->mp.BytePerPixelInDETC[k],
+			&mode_lib->mp.Read256BlockHeightY[k],
+			&mode_lib->mp.Read256BlockHeightC[k],
+			&mode_lib->mp.Read256BlockWidthY[k],
+			&mode_lib->mp.Read256BlockWidthC[k],
+			&mode_lib->mp.MacroTileHeightY[k],
+			&mode_lib->mp.MacroTileHeightC[k],
+			&mode_lib->mp.MacroTileWidthY[k],
+			&mode_lib->mp.MacroTileWidthC[k],
+			&mode_lib->mp.surf_linear128_l[k],
+			&mode_lib->mp.surf_linear128_c[k]);
+	}
+
+	CalculateSwathWidth(
+		display_cfg,
+		false, // ForceSingleDPP
+		s->num_active_planes,
+		mode_lib->mp.ODMMode,
+		mode_lib->mp.BytePerPixelY,
+		mode_lib->mp.BytePerPixelC,
+		mode_lib->mp.Read256BlockHeightY,
+		mode_lib->mp.Read256BlockHeightC,
+		mode_lib->mp.Read256BlockWidthY,
+		mode_lib->mp.Read256BlockWidthC,
+		mode_lib->mp.surf_linear128_l,
+		mode_lib->mp.surf_linear128_c,
+		mode_lib->mp.NoOfDPP,
+
+		/* Output */
+		mode_lib->mp.req_per_swath_ub_l,
+		mode_lib->mp.req_per_swath_ub_c,
+		mode_lib->mp.SwathWidthSingleDPPY,
+		mode_lib->mp.SwathWidthSingleDPPC,
+		mode_lib->mp.SwathWidthY,
+		mode_lib->mp.SwathWidthC,
+		s->dummy_integer_array[0], // unsigned int MaximumSwathHeightY[]
+		s->dummy_integer_array[1], // unsigned int MaximumSwathHeightC[]
+		mode_lib->mp.swath_width_luma_ub,
+		mode_lib->mp.swath_width_chroma_ub);
+
+	for (k = 0; k < s->num_active_planes; ++k) {
+		mode_lib->mp.cursor_bw[k] = display_cfg->plane_descriptors[k].cursor.num_cursors * display_cfg->plane_descriptors[k].cursor.cursor_width * display_cfg->plane_descriptors[k].cursor.cursor_bpp / 8.0 /
+			((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000));
+		mode_lib->mp.vactive_sw_bw_l[k] = mode_lib->mp.SwathWidthSingleDPPY[k] * mode_lib->mp.BytePerPixelY[k] / (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000)) * display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio;
+		mode_lib->mp.vactive_sw_bw_c[k] = mode_lib->mp.SwathWidthSingleDPPC[k] * mode_lib->mp.BytePerPixelC[k] / (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000)) * display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_ratio;
+		DML_LOG_VERBOSE("DML::%s: vactive_sw_bw_l[%i] = %fBps\n", __func__, k, mode_lib->mp.vactive_sw_bw_l[k]);
+		DML_LOG_VERBOSE("DML::%s: vactive_sw_bw_c[%i] = %fBps\n", __func__, k, mode_lib->mp.vactive_sw_bw_c[k]);
+	}
+
+	CalculateSwathAndDETConfiguration_params->display_cfg = display_cfg;
+	CalculateSwathAndDETConfiguration_params->ConfigReturnBufferSizeInKByte = mode_lib->ip.config_return_buffer_size_in_kbytes;
+	CalculateSwathAndDETConfiguration_params->MaxTotalDETInKByte = s->MaxTotalDETInKByte;
+	CalculateSwathAndDETConfiguration_params->MinCompressedBufferSizeInKByte = s->MinCompressedBufferSizeInKByte;
+	CalculateSwathAndDETConfiguration_params->rob_buffer_size_kbytes = mode_lib->ip.rob_buffer_size_kbytes;
+	CalculateSwathAndDETConfiguration_params->pixel_chunk_size_kbytes = mode_lib->ip.pixel_chunk_size_kbytes;
+	CalculateSwathAndDETConfiguration_params->rob_buffer_size_kbytes = mode_lib->ip.rob_buffer_size_kbytes;
+	CalculateSwathAndDETConfiguration_params->pixel_chunk_size_kbytes = mode_lib->ip.pixel_chunk_size_kbytes;
+	CalculateSwathAndDETConfiguration_params->ForceSingleDPP = false;
+	CalculateSwathAndDETConfiguration_params->NumberOfActiveSurfaces = s->num_active_planes;
+	CalculateSwathAndDETConfiguration_params->nomDETInKByte = s->NomDETInKByte;
+	CalculateSwathAndDETConfiguration_params->ConfigReturnBufferSegmentSizeInkByte = mode_lib->ip.config_return_buffer_segment_size_in_kbytes;
+	CalculateSwathAndDETConfiguration_params->CompressedBufferSegmentSizeInkByte = mode_lib->ip.compressed_buffer_segment_size_in_kbytes;
+	CalculateSwathAndDETConfiguration_params->ReadBandwidthLuma = mode_lib->mp.vactive_sw_bw_l;
+	CalculateSwathAndDETConfiguration_params->ReadBandwidthChroma = mode_lib->mp.vactive_sw_bw_c;
+	CalculateSwathAndDETConfiguration_params->MaximumSwathWidthLuma = s->dummy_single_array[0];
+	CalculateSwathAndDETConfiguration_params->MaximumSwathWidthChroma = s->dummy_single_array[1];
+	CalculateSwathAndDETConfiguration_params->Read256BytesBlockHeightY = mode_lib->mp.Read256BlockHeightY;
+	CalculateSwathAndDETConfiguration_params->Read256BytesBlockHeightC = mode_lib->mp.Read256BlockHeightC;
+	CalculateSwathAndDETConfiguration_params->Read256BytesBlockWidthY = mode_lib->mp.Read256BlockWidthY;
+	CalculateSwathAndDETConfiguration_params->Read256BytesBlockWidthC = mode_lib->mp.Read256BlockWidthC;
+	CalculateSwathAndDETConfiguration_params->surf_linear128_l = mode_lib->mp.surf_linear128_l;
+	CalculateSwathAndDETConfiguration_params->surf_linear128_c = mode_lib->mp.surf_linear128_c;
+	CalculateSwathAndDETConfiguration_params->ODMMode = mode_lib->mp.ODMMode;
+	CalculateSwathAndDETConfiguration_params->DPPPerSurface = mode_lib->mp.NoOfDPP;
+	CalculateSwathAndDETConfiguration_params->BytePerPixY = mode_lib->mp.BytePerPixelY;
+	CalculateSwathAndDETConfiguration_params->BytePerPixC = mode_lib->mp.BytePerPixelC;
+	CalculateSwathAndDETConfiguration_params->BytePerPixDETY = mode_lib->mp.BytePerPixelInDETY;
+	CalculateSwathAndDETConfiguration_params->BytePerPixDETC = mode_lib->mp.BytePerPixelInDETC;
+	CalculateSwathAndDETConfiguration_params->mrq_present = mode_lib->ip.dcn_mrq_present;
+
+	// output
+	CalculateSwathAndDETConfiguration_params->req_per_swath_ub_l = mode_lib->mp.req_per_swath_ub_l;
+	CalculateSwathAndDETConfiguration_params->req_per_swath_ub_c = mode_lib->mp.req_per_swath_ub_c;
+	CalculateSwathAndDETConfiguration_params->swath_width_luma_ub = s->dummy_long_array[0];
+	CalculateSwathAndDETConfiguration_params->swath_width_chroma_ub = s->dummy_long_array[1];
+	CalculateSwathAndDETConfiguration_params->SwathWidth = s->dummy_long_array[2];
+	CalculateSwathAndDETConfiguration_params->SwathWidthChroma = s->dummy_long_array[3];
+	CalculateSwathAndDETConfiguration_params->SwathHeightY = mode_lib->mp.SwathHeightY;
+	CalculateSwathAndDETConfiguration_params->SwathHeightC = mode_lib->mp.SwathHeightC;
+	CalculateSwathAndDETConfiguration_params->request_size_bytes_luma = mode_lib->mp.request_size_bytes_luma;
+	CalculateSwathAndDETConfiguration_params->request_size_bytes_chroma = mode_lib->mp.request_size_bytes_chroma;
+	CalculateSwathAndDETConfiguration_params->DETBufferSizeInKByte = mode_lib->mp.DETBufferSizeInKByte;
+	CalculateSwathAndDETConfiguration_params->DETBufferSizeY = mode_lib->mp.DETBufferSizeY;
+	CalculateSwathAndDETConfiguration_params->DETBufferSizeC = mode_lib->mp.DETBufferSizeC;
+	CalculateSwathAndDETConfiguration_params->full_swath_bytes_l = s->full_swath_bytes_l;
+	CalculateSwathAndDETConfiguration_params->full_swath_bytes_c = s->full_swath_bytes_c;
+	CalculateSwathAndDETConfiguration_params->UnboundedRequestEnabled = &mode_lib->mp.UnboundedRequestEnabled;
+	CalculateSwathAndDETConfiguration_params->compbuf_reserved_space_64b = &mode_lib->mp.compbuf_reserved_space_64b;
+	CalculateSwathAndDETConfiguration_params->hw_debug5 = &mode_lib->mp.hw_debug5;
+	CalculateSwathAndDETConfiguration_params->CompressedBufferSizeInkByte = &mode_lib->mp.CompressedBufferSizeInkByte;
+	CalculateSwathAndDETConfiguration_params->ViewportSizeSupportPerSurface = &s->dummy_boolean_array[0][0];
+	CalculateSwathAndDETConfiguration_params->ViewportSizeSupport = &s->dummy_boolean[0];
+
+	// Calculate DET size, swath height here.
+	CalculateSwathAndDETConfiguration(&mode_lib->scratch, CalculateSwathAndDETConfiguration_params);
+
+	// DSC Delay
+	for (k = 0; k < s->num_active_planes; ++k) {
+		mode_lib->mp.DSCDelay[k] = DSCDelayRequirement(cfg_support_info->stream_support_info[display_cfg->plane_descriptors[k].stream_index].dsc_enable,
+			mode_lib->mp.ODMMode[k],
+			mode_lib->ip.maximum_dsc_bits_per_component,
+			s->OutputBpp[k],
+			display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_active,
+			display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total,
+			cfg_support_info->stream_support_info[display_cfg->plane_descriptors[k].stream_index].num_dsc_slices,
+			display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_format,
+			display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].output.output_encoder,
+			((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000),
+			s->PixelClockBackEnd[k]);
+	}
+
+	// Prefetch
+	if (mode_lib->soc.mall_allocated_for_dcn_mbytes == 0) {
+		for (k = 0; k < s->num_active_planes; ++k)
+			mode_lib->mp.SurfaceSizeInTheMALL[k] = 0;
+	} else {
+		CalculateSurfaceSizeInMall(
+			display_cfg,
+			s->num_active_planes,
+			mode_lib->soc.mall_allocated_for_dcn_mbytes,
+			mode_lib->mp.BytePerPixelY,
+			mode_lib->mp.BytePerPixelC,
+			mode_lib->mp.Read256BlockWidthY,
+			mode_lib->mp.Read256BlockWidthC,
+			mode_lib->mp.Read256BlockHeightY,
+			mode_lib->mp.Read256BlockHeightC,
+			mode_lib->mp.MacroTileWidthY,
+			mode_lib->mp.MacroTileWidthC,
+			mode_lib->mp.MacroTileHeightY,
+			mode_lib->mp.MacroTileHeightC,
+
+			/* Output */
+			mode_lib->mp.SurfaceSizeInTheMALL,
+			&s->dummy_boolean[0]); /* bool *ExceededMALLSize */
+	}
+
+	for (k = 0; k < s->num_active_planes; ++k) {
+		s->SurfaceParameters[k].PixelClock = ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000);
+		s->SurfaceParameters[k].DPPPerSurface = mode_lib->mp.NoOfDPP[k];
+		s->SurfaceParameters[k].RotationAngle = display_cfg->plane_descriptors[k].composition.rotation_angle;
+		s->SurfaceParameters[k].ViewportHeight = display_cfg->plane_descriptors[k].composition.viewport.plane0.height;
+		s->SurfaceParameters[k].ViewportHeightC = display_cfg->plane_descriptors[k].composition.viewport.plane1.height;
+		s->SurfaceParameters[k].BlockWidth256BytesY = mode_lib->mp.Read256BlockWidthY[k];
+		s->SurfaceParameters[k].BlockHeight256BytesY = mode_lib->mp.Read256BlockHeightY[k];
+		s->SurfaceParameters[k].BlockWidth256BytesC = mode_lib->mp.Read256BlockWidthC[k];
+		s->SurfaceParameters[k].BlockHeight256BytesC = mode_lib->mp.Read256BlockHeightC[k];
+		s->SurfaceParameters[k].BlockWidthY = mode_lib->mp.MacroTileWidthY[k];
+		s->SurfaceParameters[k].BlockHeightY = mode_lib->mp.MacroTileHeightY[k];
+		s->SurfaceParameters[k].BlockWidthC = mode_lib->mp.MacroTileWidthC[k];
+		s->SurfaceParameters[k].BlockHeightC = mode_lib->mp.MacroTileHeightC[k];
+		s->SurfaceParameters[k].InterlaceEnable = display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.interlaced;
+		s->SurfaceParameters[k].HTotal = display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total;
+		s->SurfaceParameters[k].DCCEnable = display_cfg->plane_descriptors[k].surface.dcc.enable;
+		s->SurfaceParameters[k].SourcePixelFormat = display_cfg->plane_descriptors[k].pixel_format;
+		s->SurfaceParameters[k].SurfaceTiling = display_cfg->plane_descriptors[k].surface.tiling;
+		s->SurfaceParameters[k].BytePerPixelY = mode_lib->mp.BytePerPixelY[k];
+		s->SurfaceParameters[k].BytePerPixelC = mode_lib->mp.BytePerPixelC[k];
+		s->SurfaceParameters[k].ProgressiveToInterlaceUnitInOPP = mode_lib->ip.ptoi_supported;
+		s->SurfaceParameters[k].VRatio = display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio;
+		s->SurfaceParameters[k].VRatioChroma = display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_ratio;
+		s->SurfaceParameters[k].VTaps = display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_taps;
+		s->SurfaceParameters[k].VTapsChroma = display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_taps;
+		s->SurfaceParameters[k].PitchY = display_cfg->plane_descriptors[k].surface.plane0.pitch;
+		s->SurfaceParameters[k].PitchC = display_cfg->plane_descriptors[k].surface.plane1.pitch;
+		s->SurfaceParameters[k].ViewportStationary = display_cfg->plane_descriptors[k].composition.viewport.stationary;
+		s->SurfaceParameters[k].ViewportXStart = display_cfg->plane_descriptors[k].composition.viewport.plane0.x_start;
+		s->SurfaceParameters[k].ViewportYStart = display_cfg->plane_descriptors[k].composition.viewport.plane0.y_start;
+		s->SurfaceParameters[k].ViewportXStartC = display_cfg->plane_descriptors[k].composition.viewport.plane1.y_start;
+		s->SurfaceParameters[k].ViewportYStartC = display_cfg->plane_descriptors[k].composition.viewport.plane1.y_start;
+		s->SurfaceParameters[k].FORCE_ONE_ROW_FOR_FRAME = display_cfg->plane_descriptors[k].overrides.hw.force_one_row_for_frame;
+		s->SurfaceParameters[k].SwathHeightY = mode_lib->mp.SwathHeightY[k];
+		s->SurfaceParameters[k].SwathHeightC = mode_lib->mp.SwathHeightC[k];
+		s->SurfaceParameters[k].DCCMetaPitchY = display_cfg->plane_descriptors[k].surface.dcc.plane0.pitch;
+		s->SurfaceParameters[k].DCCMetaPitchC = display_cfg->plane_descriptors[k].surface.dcc.plane1.pitch;
+	}
+
+	CalculateVMRowAndSwath_params->display_cfg = display_cfg;
+	CalculateVMRowAndSwath_params->NumberOfActiveSurfaces = s->num_active_planes;
+	CalculateVMRowAndSwath_params->myPipe = s->SurfaceParameters;
+	CalculateVMRowAndSwath_params->SurfaceSizeInMALL = mode_lib->mp.SurfaceSizeInTheMALL;
+	CalculateVMRowAndSwath_params->PTEBufferSizeInRequestsLuma = mode_lib->ip.dpte_buffer_size_in_pte_reqs_luma;
+	CalculateVMRowAndSwath_params->PTEBufferSizeInRequestsChroma = mode_lib->ip.dpte_buffer_size_in_pte_reqs_chroma;
+	CalculateVMRowAndSwath_params->MALLAllocatedForDCN = mode_lib->soc.mall_allocated_for_dcn_mbytes;
+	CalculateVMRowAndSwath_params->SwathWidthY = mode_lib->mp.SwathWidthY;
+	CalculateVMRowAndSwath_params->SwathWidthC = mode_lib->mp.SwathWidthC;
+	CalculateVMRowAndSwath_params->HostVMMinPageSize = mode_lib->soc.hostvm_min_page_size_kbytes;
+	CalculateVMRowAndSwath_params->DCCMetaBufferSizeBytes = mode_lib->ip.dcc_meta_buffer_size_bytes;
+	CalculateVMRowAndSwath_params->mrq_present = mode_lib->ip.dcn_mrq_present;
+
+	// output
+	CalculateVMRowAndSwath_params->PTEBufferSizeNotExceeded = s->dummy_boolean_array[0];
+	CalculateVMRowAndSwath_params->dpte_row_width_luma_ub = mode_lib->mp.dpte_row_width_luma_ub;
+	CalculateVMRowAndSwath_params->dpte_row_width_chroma_ub = mode_lib->mp.dpte_row_width_chroma_ub;
+	CalculateVMRowAndSwath_params->dpte_row_height_luma = mode_lib->mp.dpte_row_height;
+	CalculateVMRowAndSwath_params->dpte_row_height_chroma = mode_lib->mp.dpte_row_height_chroma;
+	CalculateVMRowAndSwath_params->dpte_row_height_linear_luma = mode_lib->mp.dpte_row_height_linear;
+	CalculateVMRowAndSwath_params->dpte_row_height_linear_chroma = mode_lib->mp.dpte_row_height_linear_chroma;
+	CalculateVMRowAndSwath_params->vm_group_bytes = mode_lib->mp.vm_group_bytes;
+	CalculateVMRowAndSwath_params->dpte_group_bytes = mode_lib->mp.dpte_group_bytes;
+	CalculateVMRowAndSwath_params->PixelPTEReqWidthY = mode_lib->mp.PixelPTEReqWidthY;
+	CalculateVMRowAndSwath_params->PixelPTEReqHeightY = mode_lib->mp.PixelPTEReqHeightY;
+	CalculateVMRowAndSwath_params->PTERequestSizeY = mode_lib->mp.PTERequestSizeY;
+	CalculateVMRowAndSwath_params->PixelPTEReqWidthC = mode_lib->mp.PixelPTEReqWidthC;
+	CalculateVMRowAndSwath_params->PixelPTEReqHeightC = mode_lib->mp.PixelPTEReqHeightC;
+	CalculateVMRowAndSwath_params->PTERequestSizeC = mode_lib->mp.PTERequestSizeC;
+	CalculateVMRowAndSwath_params->vmpg_width_y = s->vmpg_width_y;
+	CalculateVMRowAndSwath_params->vmpg_height_y = s->vmpg_height_y;
+	CalculateVMRowAndSwath_params->vmpg_width_c = s->vmpg_width_c;
+	CalculateVMRowAndSwath_params->vmpg_height_c = s->vmpg_height_c;
+	CalculateVMRowAndSwath_params->dpde0_bytes_per_frame_ub_l = mode_lib->mp.dpde0_bytes_per_frame_ub_l;
+	CalculateVMRowAndSwath_params->dpde0_bytes_per_frame_ub_c = mode_lib->mp.dpde0_bytes_per_frame_ub_c;
+	CalculateVMRowAndSwath_params->PrefetchSourceLinesY = mode_lib->mp.PrefetchSourceLinesY;
+	CalculateVMRowAndSwath_params->PrefetchSourceLinesC = mode_lib->mp.PrefetchSourceLinesC;
+	CalculateVMRowAndSwath_params->VInitPreFillY = mode_lib->mp.VInitPreFillY;
+	CalculateVMRowAndSwath_params->VInitPreFillC = mode_lib->mp.VInitPreFillC;
+	CalculateVMRowAndSwath_params->MaxNumSwathY = mode_lib->mp.MaxNumSwathY;
+	CalculateVMRowAndSwath_params->MaxNumSwathC = mode_lib->mp.MaxNumSwathC;
+	CalculateVMRowAndSwath_params->dpte_row_bw = mode_lib->mp.dpte_row_bw;
+	CalculateVMRowAndSwath_params->PixelPTEBytesPerRow = mode_lib->mp.PixelPTEBytesPerRow;
+	CalculateVMRowAndSwath_params->dpte_row_bytes_per_row_l = s->dpte_row_bytes_per_row_l;
+	CalculateVMRowAndSwath_params->dpte_row_bytes_per_row_c = s->dpte_row_bytes_per_row_c;
+	CalculateVMRowAndSwath_params->vm_bytes = mode_lib->mp.vm_bytes;
+	CalculateVMRowAndSwath_params->use_one_row_for_frame = mode_lib->mp.use_one_row_for_frame;
+	CalculateVMRowAndSwath_params->use_one_row_for_frame_flip = mode_lib->mp.use_one_row_for_frame_flip;
+	CalculateVMRowAndSwath_params->is_using_mall_for_ss = mode_lib->mp.is_using_mall_for_ss;
+	CalculateVMRowAndSwath_params->PTE_BUFFER_MODE = mode_lib->mp.PTE_BUFFER_MODE;
+	CalculateVMRowAndSwath_params->BIGK_FRAGMENT_SIZE = mode_lib->mp.BIGK_FRAGMENT_SIZE;
+	CalculateVMRowAndSwath_params->DCCMetaBufferSizeNotExceeded = s->dummy_boolean_array[1];
+	CalculateVMRowAndSwath_params->meta_row_bw = mode_lib->mp.meta_row_bw;
+	CalculateVMRowAndSwath_params->meta_row_bytes = mode_lib->mp.meta_row_bytes;
+	CalculateVMRowAndSwath_params->meta_row_bytes_per_row_ub_l = s->meta_row_bytes_per_row_ub_l;
+	CalculateVMRowAndSwath_params->meta_row_bytes_per_row_ub_c = s->meta_row_bytes_per_row_ub_c;
+	CalculateVMRowAndSwath_params->meta_req_width_luma = mode_lib->mp.meta_req_width;
+	CalculateVMRowAndSwath_params->meta_req_height_luma = mode_lib->mp.meta_req_height;
+	CalculateVMRowAndSwath_params->meta_row_width_luma = mode_lib->mp.meta_row_width;
+	CalculateVMRowAndSwath_params->meta_row_height_luma = mode_lib->mp.meta_row_height;
+	CalculateVMRowAndSwath_params->meta_pte_bytes_per_frame_ub_l = mode_lib->mp.meta_pte_bytes_per_frame_ub_l;
+	CalculateVMRowAndSwath_params->meta_req_width_chroma = mode_lib->mp.meta_req_width_chroma;
+	CalculateVMRowAndSwath_params->meta_row_height_chroma = mode_lib->mp.meta_row_height_chroma;
+	CalculateVMRowAndSwath_params->meta_row_width_chroma = mode_lib->mp.meta_row_width_chroma;
+	CalculateVMRowAndSwath_params->meta_req_height_chroma = mode_lib->mp.meta_req_height_chroma;
+	CalculateVMRowAndSwath_params->meta_pte_bytes_per_frame_ub_c = mode_lib->mp.meta_pte_bytes_per_frame_ub_c;
+
+	CalculateVMRowAndSwath(&mode_lib->scratch, CalculateVMRowAndSwath_params);
+
+	memset(calculate_mcache_setting_params, 0, sizeof(struct dml2_core_calcs_calculate_mcache_setting_params));
+	if (mode_lib->soc.mall_allocated_for_dcn_mbytes == 0 || mode_lib->ip.dcn_mrq_present) {
+		for (k = 0; k < s->num_active_planes; k++) {
+			mode_lib->mp.mall_prefetch_sdp_overhead_factor[k] = 1.0;
+			mode_lib->mp.mall_prefetch_dram_overhead_factor[k] = 1.0;
+			mode_lib->mp.dcc_dram_bw_nom_overhead_factor_p0[k] = 1.0;
+			mode_lib->mp.dcc_dram_bw_pref_overhead_factor_p0[k] = 1.0;
+			mode_lib->mp.dcc_dram_bw_nom_overhead_factor_p1[k] = 1.0;
+			mode_lib->mp.dcc_dram_bw_pref_overhead_factor_p1[k] = 1.0;
+		}
+	} else {
+		for (k = 0; k < s->num_active_planes; k++) {
+			calculate_mcache_setting_params->dcc_enable = display_cfg->plane_descriptors[k].surface.dcc.enable;
+			calculate_mcache_setting_params->num_chans = mode_lib->soc.clk_table.dram_config.channel_count;
+			calculate_mcache_setting_params->mem_word_bytes = mode_lib->soc.mem_word_bytes;
+			calculate_mcache_setting_params->mcache_size_bytes = mode_lib->soc.mcache_size_bytes;
+			calculate_mcache_setting_params->mcache_line_size_bytes = mode_lib->soc.mcache_line_size_bytes;
+			calculate_mcache_setting_params->gpuvm_enable = display_cfg->gpuvm_enable;
+			calculate_mcache_setting_params->gpuvm_page_size_kbytes = display_cfg->plane_descriptors[k].overrides.gpuvm_min_page_size_kbytes;
+
+			calculate_mcache_setting_params->source_format = display_cfg->plane_descriptors[k].pixel_format;
+			calculate_mcache_setting_params->surf_vert = dml_is_vertical_rotation(display_cfg->plane_descriptors[k].composition.rotation_angle);
+			calculate_mcache_setting_params->vp_stationary = display_cfg->plane_descriptors[k].composition.viewport.stationary;
+			calculate_mcache_setting_params->tiling_mode = display_cfg->plane_descriptors[k].surface.tiling;
+			calculate_mcache_setting_params->imall_enable = mode_lib->ip.imall_supported && display_cfg->plane_descriptors[k].overrides.legacy_svp_config == dml2_svp_mode_override_imall;
+
+			calculate_mcache_setting_params->vp_start_x_l = display_cfg->plane_descriptors[k].composition.viewport.plane0.x_start;
+			calculate_mcache_setting_params->vp_start_y_l = display_cfg->plane_descriptors[k].composition.viewport.plane0.y_start;
+			calculate_mcache_setting_params->full_vp_width_l = display_cfg->plane_descriptors[k].composition.viewport.plane0.width;
+			calculate_mcache_setting_params->full_vp_height_l = display_cfg->plane_descriptors[k].composition.viewport.plane0.height;
+			calculate_mcache_setting_params->blk_width_l = mode_lib->mp.MacroTileWidthY[k];
+			calculate_mcache_setting_params->blk_height_l = mode_lib->mp.MacroTileHeightY[k];
+			calculate_mcache_setting_params->vmpg_width_l = s->vmpg_width_y[k];
+			calculate_mcache_setting_params->vmpg_height_l = s->vmpg_height_y[k];
+			calculate_mcache_setting_params->full_swath_bytes_l = s->full_swath_bytes_l[k];
+			calculate_mcache_setting_params->bytes_per_pixel_l = mode_lib->mp.BytePerPixelY[k];
+
+			calculate_mcache_setting_params->vp_start_x_c = display_cfg->plane_descriptors[k].composition.viewport.plane1.y_start;
+			calculate_mcache_setting_params->vp_start_y_c = display_cfg->plane_descriptors[k].composition.viewport.plane1.y_start;
+			calculate_mcache_setting_params->full_vp_width_c = display_cfg->plane_descriptors[k].composition.viewport.plane1.width;
+			calculate_mcache_setting_params->full_vp_height_c = display_cfg->plane_descriptors[k].composition.viewport.plane1.height;
+			calculate_mcache_setting_params->blk_width_c = mode_lib->mp.MacroTileWidthC[k];
+			calculate_mcache_setting_params->blk_height_c = mode_lib->mp.MacroTileHeightC[k];
+			calculate_mcache_setting_params->vmpg_width_c = s->vmpg_width_c[k];
+			calculate_mcache_setting_params->vmpg_height_c = s->vmpg_height_c[k];
+			calculate_mcache_setting_params->full_swath_bytes_c = s->full_swath_bytes_c[k];
+			calculate_mcache_setting_params->bytes_per_pixel_c = mode_lib->mp.BytePerPixelC[k];
+
+			// output
+			calculate_mcache_setting_params->dcc_dram_bw_nom_overhead_factor_l = &mode_lib->mp.dcc_dram_bw_nom_overhead_factor_p0[k];
+			calculate_mcache_setting_params->dcc_dram_bw_pref_overhead_factor_l = &mode_lib->mp.dcc_dram_bw_pref_overhead_factor_p0[k];
+			calculate_mcache_setting_params->dcc_dram_bw_nom_overhead_factor_c = &mode_lib->mp.dcc_dram_bw_nom_overhead_factor_p1[k];
+			calculate_mcache_setting_params->dcc_dram_bw_pref_overhead_factor_c = &mode_lib->mp.dcc_dram_bw_pref_overhead_factor_p1[k];
+
+			calculate_mcache_setting_params->num_mcaches_l = &mode_lib->mp.num_mcaches_l[k];
+			calculate_mcache_setting_params->mcache_row_bytes_l = &mode_lib->mp.mcache_row_bytes_l[k];
+			calculate_mcache_setting_params->mcache_row_bytes_per_channel_l = &mode_lib->mp.mcache_row_bytes_per_channel_l[k];
+			calculate_mcache_setting_params->mcache_offsets_l = mode_lib->mp.mcache_offsets_l[k];
+			calculate_mcache_setting_params->mcache_shift_granularity_l = &mode_lib->mp.mcache_shift_granularity_l[k];
+
+			calculate_mcache_setting_params->num_mcaches_c = &mode_lib->mp.num_mcaches_c[k];
+			calculate_mcache_setting_params->mcache_row_bytes_c = &mode_lib->mp.mcache_row_bytes_c[k];
+			calculate_mcache_setting_params->mcache_row_bytes_per_channel_c = &mode_lib->mp.mcache_row_bytes_per_channel_c[k];
+			calculate_mcache_setting_params->mcache_offsets_c = mode_lib->mp.mcache_offsets_c[k];
+			calculate_mcache_setting_params->mcache_shift_granularity_c = &mode_lib->mp.mcache_shift_granularity_c[k];
+
+			calculate_mcache_setting_params->mall_comb_mcache_l = &mode_lib->mp.mall_comb_mcache_l[k];
+			calculate_mcache_setting_params->mall_comb_mcache_c = &mode_lib->mp.mall_comb_mcache_c[k];
+			calculate_mcache_setting_params->lc_comb_mcache = &mode_lib->mp.lc_comb_mcache[k];
+			calculate_mcache_setting(&mode_lib->scratch, calculate_mcache_setting_params);
+		}
+
+		calculate_mall_bw_overhead_factor(
+			mode_lib->mp.mall_prefetch_sdp_overhead_factor,
+			mode_lib->mp.mall_prefetch_dram_overhead_factor,
+
+			// input
+			display_cfg,
+			s->num_active_planes);
+	}
+
+	// Calculate all the bandwidth availabe
+	calculate_bandwidth_available(
+		mode_lib->mp.avg_bandwidth_available_min,
+		mode_lib->mp.avg_bandwidth_available,
+		mode_lib->mp.urg_bandwidth_available_min,
+		mode_lib->mp.urg_bandwidth_available,
+		mode_lib->mp.urg_bandwidth_available_vm_only,
+		mode_lib->mp.urg_bandwidth_available_pixel_and_vm,
+
+		&mode_lib->soc,
+		display_cfg->hostvm_enable,
+		mode_lib->mp.Dcfclk,
+		mode_lib->mp.FabricClock,
+		mode_lib->mp.dram_bw_mbps);
+
+
+	calculate_hostvm_inefficiency_factor(
+		&s->HostVMInefficiencyFactor,
+		&s->HostVMInefficiencyFactorPrefetch,
+
+		display_cfg->gpuvm_enable,
+		display_cfg->hostvm_enable,
+		mode_lib->ip.remote_iommu_outstanding_translations,
+		mode_lib->soc.max_outstanding_reqs,
+		mode_lib->mp.urg_bandwidth_available_pixel_and_vm[dml2_core_internal_soc_state_sys_active],
+		mode_lib->mp.urg_bandwidth_available_vm_only[dml2_core_internal_soc_state_sys_active]);
+
+	s->TotalDCCActiveDPP = 0;
+	s->TotalActiveDPP = 0;
+	for (k = 0; k < s->num_active_planes; ++k) {
+		s->TotalActiveDPP = s->TotalActiveDPP + mode_lib->mp.NoOfDPP[k];
+		if (display_cfg->plane_descriptors[k].surface.dcc.enable)
+			s->TotalDCCActiveDPP = s->TotalDCCActiveDPP + mode_lib->mp.NoOfDPP[k];
+	}
+	// Calculate tdlut schedule related terms
+	for (k = 0; k <= s->num_active_planes - 1; k++) {
+		calculate_tdlut_setting_params->dispclk_mhz = mode_lib->mp.Dispclk;
+		calculate_tdlut_setting_params->setup_for_tdlut = display_cfg->plane_descriptors[k].tdlut.setup_for_tdlut;
+		calculate_tdlut_setting_params->tdlut_width_mode = display_cfg->plane_descriptors[k].tdlut.tdlut_width_mode;
+		calculate_tdlut_setting_params->tdlut_addressing_mode = display_cfg->plane_descriptors[k].tdlut.tdlut_addressing_mode;
+		calculate_tdlut_setting_params->cursor_buffer_size = mode_lib->ip.cursor_buffer_size;
+		calculate_tdlut_setting_params->gpuvm_enable = display_cfg->gpuvm_enable;
+		calculate_tdlut_setting_params->gpuvm_page_size_kbytes = display_cfg->plane_descriptors[k].overrides.gpuvm_min_page_size_kbytes;
+
+		// output
+		calculate_tdlut_setting_params->tdlut_pte_bytes_per_frame = &s->tdlut_pte_bytes_per_frame[k];
+		calculate_tdlut_setting_params->tdlut_bytes_per_frame = &s->tdlut_bytes_per_frame[k];
+		calculate_tdlut_setting_params->tdlut_groups_per_2row_ub = &s->tdlut_groups_per_2row_ub[k];
+		calculate_tdlut_setting_params->tdlut_opt_time = &s->tdlut_opt_time[k];
+		calculate_tdlut_setting_params->tdlut_drain_time = &s->tdlut_drain_time[k];
+		calculate_tdlut_setting_params->tdlut_bytes_to_deliver = &s->tdlut_bytes_to_deliver[k];
+		calculate_tdlut_setting_params->tdlut_bytes_per_group = &s->tdlut_bytes_per_group[k];
+		calculate_tdlut_setting(&mode_lib->scratch, calculate_tdlut_setting_params);
+	}
+
+	if (mode_lib->soc.qos_parameters.qos_type == dml2_qos_param_type_dcn3)
+		s->ReorderingBytes = (unsigned int)(mode_lib->soc.clk_table.dram_config.channel_count * math_max3(mode_lib->soc.qos_parameters.qos_params.dcn32x.urgent_out_of_order_return_per_channel_pixel_only_bytes,
+										mode_lib->soc.qos_parameters.qos_params.dcn32x.urgent_out_of_order_return_per_channel_pixel_and_vm_bytes,
+										mode_lib->soc.qos_parameters.qos_params.dcn32x.urgent_out_of_order_return_per_channel_vm_only_bytes));
+
+	CalculateExtraLatency(
+		display_cfg,
+		mode_lib->ip.rob_buffer_size_kbytes,
+		mode_lib->soc.qos_parameters.qos_params.dcn32x.loaded_round_trip_latency_fclk_cycles,
+		s->ReorderingBytes,
+		mode_lib->mp.Dcfclk,
+		mode_lib->mp.FabricClock,
+		mode_lib->ip.pixel_chunk_size_kbytes,
+		mode_lib->mp.urg_bandwidth_available_min[dml2_core_internal_soc_state_sys_active],
+		s->num_active_planes,
+		mode_lib->mp.NoOfDPP,
+		mode_lib->mp.dpte_group_bytes,
+		s->tdlut_bytes_per_group,
+		s->HostVMInefficiencyFactor,
+		s->HostVMInefficiencyFactorPrefetch,
+		mode_lib->soc.hostvm_min_page_size_kbytes,
+		mode_lib->soc.qos_parameters.qos_type,
+		!(display_cfg->overrides.max_outstanding_when_urgent_expected_disable),
+		mode_lib->soc.max_outstanding_reqs,
+		mode_lib->mp.request_size_bytes_luma,
+		mode_lib->mp.request_size_bytes_chroma,
+		mode_lib->ip.meta_chunk_size_kbytes,
+		mode_lib->ip.dchub_arb_to_ret_delay,
+		mode_lib->mp.TripToMemory,
+		mode_lib->ip.hostvm_mode,
+
+		// output
+		&mode_lib->mp.ExtraLatency,
+		&mode_lib->mp.ExtraLatency_sr,
+		&mode_lib->mp.ExtraLatencyPrefetch);
+
+	mode_lib->mp.TCalc = 24.0 / mode_lib->mp.DCFCLKDeepSleep;
+
+	for (k = 0; k < s->num_active_planes; ++k) {
+		if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.active_writebacks_per_stream > 0) {
+			mode_lib->mp.WritebackDelay[k] =
+				mode_lib->soc.qos_parameters.writeback.base_latency_us
+				+ CalculateWriteBackDelay(
+					display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].pixel_format,
+					display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].h_ratio,
+					display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].v_ratio,
+					display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].v_taps,
+					display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].output_width,
+					display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].output_height,
+					display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].input_height,
+					display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total) / mode_lib->mp.Dispclk;
+		} else
+			mode_lib->mp.WritebackDelay[k] = 0;
+	}
+
+	/* VActive bytes to fetch for UCLK P-State */
+	calculate_bytes_to_fetch_required_to_hide_latency_params->display_cfg = display_cfg;
+	calculate_bytes_to_fetch_required_to_hide_latency_params->mrq_present = mode_lib->ip.dcn_mrq_present;
+
+	calculate_bytes_to_fetch_required_to_hide_latency_params->num_active_planes = s->num_active_planes;
+	calculate_bytes_to_fetch_required_to_hide_latency_params->num_of_dpp = mode_lib->mp.NoOfDPP;
+	calculate_bytes_to_fetch_required_to_hide_latency_params->meta_row_height_l = mode_lib->mp.meta_row_height;
+	calculate_bytes_to_fetch_required_to_hide_latency_params->meta_row_height_c = mode_lib->mp.meta_row_height_chroma;
+	calculate_bytes_to_fetch_required_to_hide_latency_params->meta_row_bytes_per_row_ub_l = s->meta_row_bytes_per_row_ub_l;
+	calculate_bytes_to_fetch_required_to_hide_latency_params->meta_row_bytes_per_row_ub_c = s->meta_row_bytes_per_row_ub_c;
+	calculate_bytes_to_fetch_required_to_hide_latency_params->dpte_row_height_l = mode_lib->mp.dpte_row_height;
+	calculate_bytes_to_fetch_required_to_hide_latency_params->dpte_row_height_c = mode_lib->mp.dpte_row_height_chroma;
+	calculate_bytes_to_fetch_required_to_hide_latency_params->dpte_bytes_per_row_l = s->dpte_row_bytes_per_row_l;
+	calculate_bytes_to_fetch_required_to_hide_latency_params->dpte_bytes_per_row_c = s->dpte_row_bytes_per_row_c;
+	calculate_bytes_to_fetch_required_to_hide_latency_params->byte_per_pix_l = mode_lib->mp.BytePerPixelY;
+	calculate_bytes_to_fetch_required_to_hide_latency_params->byte_per_pix_c = mode_lib->mp.BytePerPixelC;
+	calculate_bytes_to_fetch_required_to_hide_latency_params->swath_width_l = mode_lib->mp.SwathWidthY;
+	calculate_bytes_to_fetch_required_to_hide_latency_params->swath_width_c = mode_lib->mp.SwathWidthC;
+	calculate_bytes_to_fetch_required_to_hide_latency_params->swath_height_l = mode_lib->mp.SwathHeightY;
+	calculate_bytes_to_fetch_required_to_hide_latency_params->swath_height_c = mode_lib->mp.SwathHeightC;
+	calculate_bytes_to_fetch_required_to_hide_latency_params->latency_to_hide_us[0] = mode_lib->soc.power_management_parameters.dram_clk_change_blackout_us;
+
+	/* outputs */
+	calculate_bytes_to_fetch_required_to_hide_latency_params->bytes_required_l = s->pstate_bytes_required_l[dml2_pstate_type_uclk];
+	calculate_bytes_to_fetch_required_to_hide_latency_params->bytes_required_c = s->pstate_bytes_required_c[dml2_pstate_type_uclk];
+
+	calculate_bytes_to_fetch_required_to_hide_latency(calculate_bytes_to_fetch_required_to_hide_latency_params);
+
+	/* Excess VActive bandwidth required to fill DET */
+	calculate_excess_vactive_bandwidth_required(
+			display_cfg,
+			s->num_active_planes,
+			s->pstate_bytes_required_l[dml2_pstate_type_uclk],
+			s->pstate_bytes_required_c[dml2_pstate_type_uclk],
+			/* outputs */
+			mode_lib->mp.excess_vactive_fill_bw_l,
+			mode_lib->mp.excess_vactive_fill_bw_c);
+
+	mode_lib->mp.UrgentLatency = CalculateUrgentLatency(
+		mode_lib->soc.qos_parameters.qos_params.dcn32x.urgent_latency_us.base_latency_us,
+		mode_lib->soc.qos_parameters.qos_params.dcn32x.urgent_latency_us.base_latency_pixel_vm_us,
+		mode_lib->soc.qos_parameters.qos_params.dcn32x.urgent_latency_us.base_latency_vm_us,
+		mode_lib->soc.do_urgent_latency_adjustment,
+		mode_lib->soc.qos_parameters.qos_params.dcn32x.urgent_latency_us.scaling_factor_fclk_us,
+		mode_lib->soc.qos_parameters.qos_params.dcn32x.urgent_latency_us.scaling_factor_mhz,
+		mode_lib->mp.FabricClock,
+		mode_lib->mp.uclk_freq_mhz,
+		mode_lib->soc.qos_parameters.qos_type,
+		mode_lib->soc.qos_parameters.qos_params.dcn4x.per_uclk_dpm_params[mode_lib->mp.qos_param_index].urgent_ramp_uclk_cycles,
+		mode_lib->soc.qos_parameters.qos_params.dcn4x.df_qos_response_time_fclk_cycles,
+		mode_lib->soc.qos_parameters.qos_params.dcn4x.max_round_trip_to_furthest_cs_fclk_cycles,
+		mode_lib->soc.qos_parameters.qos_params.dcn4x.mall_overhead_fclk_cycles,
+		mode_lib->soc.qos_parameters.qos_params.dcn4x.umc_urgent_ramp_latency_margin,
+		mode_lib->soc.qos_parameters.qos_params.dcn4x.fabric_max_transport_latency_margin);
+
+	mode_lib->mp.TripToMemory = CalculateTripToMemory(
+		mode_lib->mp.UrgentLatency,
+		mode_lib->mp.FabricClock,
+		mode_lib->mp.uclk_freq_mhz,
+		mode_lib->soc.qos_parameters.qos_type,
+		mode_lib->soc.qos_parameters.qos_params.dcn4x.per_uclk_dpm_params[mode_lib->mp.qos_param_index].trip_to_memory_uclk_cycles,
+		mode_lib->soc.qos_parameters.qos_params.dcn4x.max_round_trip_to_furthest_cs_fclk_cycles,
+		mode_lib->soc.qos_parameters.qos_params.dcn4x.mall_overhead_fclk_cycles,
+		mode_lib->soc.qos_parameters.qos_params.dcn4x.umc_max_latency_margin,
+		mode_lib->soc.qos_parameters.qos_params.dcn4x.fabric_max_transport_latency_margin);
+
+	mode_lib->mp.TripToMemory = math_max2(mode_lib->mp.UrgentLatency, mode_lib->mp.TripToMemory);
+
+	mode_lib->mp.MetaTripToMemory = CalculateMetaTripToMemory(
+		mode_lib->mp.UrgentLatency,
+		mode_lib->mp.FabricClock,
+		mode_lib->mp.uclk_freq_mhz,
+		mode_lib->soc.qos_parameters.qos_type,
+		mode_lib->soc.qos_parameters.qos_params.dcn4x.per_uclk_dpm_params[mode_lib->mp.qos_param_index].meta_trip_to_memory_uclk_cycles,
+		mode_lib->soc.qos_parameters.qos_params.dcn4x.meta_trip_adder_fclk_cycles,
+		mode_lib->soc.qos_parameters.qos_params.dcn4x.umc_max_latency_margin,
+		mode_lib->soc.qos_parameters.qos_params.dcn4x.fabric_max_transport_latency_margin);
+
+	for (k = 0; k < s->num_active_planes; ++k) {
+		bool cursor_not_enough_urgent_latency_hiding = false;
+		s->line_times[k] = display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total /
+			((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000);
+
+		s->pixel_format[k] = display_cfg->plane_descriptors[k].pixel_format;
+
+		s->lb_source_lines_l[k] = get_num_lb_source_lines(mode_lib->ip.max_line_buffer_lines, mode_lib->ip.line_buffer_size_bits,
+															mode_lib->mp.NoOfDPP[k],
+															display_cfg->plane_descriptors[k].composition.viewport.plane0.width,
+															display_cfg->plane_descriptors[k].composition.viewport.plane0.height,
+															display_cfg->plane_descriptors[k].composition.scaler_info.plane0.h_ratio,
+															display_cfg->plane_descriptors[k].composition.rotation_angle);
+
+		s->lb_source_lines_c[k] = get_num_lb_source_lines(mode_lib->ip.max_line_buffer_lines, mode_lib->ip.line_buffer_size_bits,
+															mode_lib->mp.NoOfDPP[k],
+															display_cfg->plane_descriptors[k].composition.viewport.plane1.width,
+															display_cfg->plane_descriptors[k].composition.viewport.plane1.height,
+															display_cfg->plane_descriptors[k].composition.scaler_info.plane1.h_ratio,
+															display_cfg->plane_descriptors[k].composition.rotation_angle);
+
+		if (display_cfg->plane_descriptors[k].cursor.num_cursors > 0) {
+			calculate_cursor_req_attributes(
+				display_cfg->plane_descriptors[k].cursor.cursor_width,
+				display_cfg->plane_descriptors[k].cursor.cursor_bpp,
+
+				// output
+				&s->cursor_lines_per_chunk[k],
+				&s->cursor_bytes_per_line[k],
+				&s->cursor_bytes_per_chunk[k],
+				&s->cursor_bytes[k]);
+
+			calculate_cursor_urgent_burst_factor(
+				mode_lib->ip.cursor_buffer_size,
+				display_cfg->plane_descriptors[k].cursor.cursor_width,
+				s->cursor_bytes_per_chunk[k],
+				s->cursor_lines_per_chunk[k],
+				s->line_times[k],
+				mode_lib->mp.UrgentLatency,
+
+				// output
+
author	Michal Kalderon <michal.kalderon@marvell.com>	2020-01-27 15:26:15 +0200
committer	David S. Miller <davem@davemloft.net>	2020-01-27 14:35:32 +0100
commit	0500a70d6e071040ffdaadebb966986afa83c5e9 (patch)
tree	0905d51352b4a10aa351092ad5c55ead9b481991 /tools/perf/scripts/python/netdev-times.py
parent	6459d93619b5bc21f775e7eb12bc4d051743d7aa (diff)